diff --git "a/checkpoint-50354/trainer_state.json" "b/checkpoint-50354/trainer_state.json" new file mode 100644--- /dev/null +++ "b/checkpoint-50354/trainer_state.json" @@ -0,0 +1,363641 @@ +{ + "best_metric": null, + "best_model_checkpoint": null, + "epoch": 1.0, + "global_step": 50354, + "is_hyper_param_search": false, + "is_local_process_zero": true, + "is_world_process_zero": true, + "log_history": [ + { + "epoch": 0.0, + "learning_rate": 1.5873015873015873e-06, + "loss": 10.5427, + "theoretical_loss": 20.841284642178216, + "tokens_seen": 65536 + }, + { + "epoch": 0.0, + "learning_rate": 3.1746031746031746e-06, + "loss": 10.5418, + "theoretical_loss": 17.59466794495971, + "tokens_seen": 131072 + }, + { + "epoch": 0.0, + "learning_rate": 4.7619047619047615e-06, + "loss": 10.4482, + "theoretical_loss": 15.967943933467634, + "tokens_seen": 196608 + }, + { + "epoch": 0.0, + "learning_rate": 6.349206349206349e-06, + "loss": 10.2263, + "theoretical_loss": 14.920783596619636, + "tokens_seen": 262144 + }, + { + "epoch": 0.0, + "learning_rate": 7.936507936507936e-06, + "loss": 9.9479, + "theoretical_loss": 14.164683778394041, + "tokens_seen": 327680 + }, + { + "epoch": 0.0, + "learning_rate": 9.523809523809523e-06, + "loss": 9.6891, + "theoretical_loss": 13.581028313181289, + "tokens_seen": 393216 + }, + { + "epoch": 0.0, + "learning_rate": 1.1111111111111112e-05, + "loss": 9.4516, + "theoretical_loss": 13.110275842032255, + "tokens_seen": 458752 + }, + { + "epoch": 0.0, + "learning_rate": 1.2698412698412699e-05, + "loss": 9.276, + "theoretical_loss": 12.71859646611439, + "tokens_seen": 524288 + }, + { + "epoch": 0.0, + "learning_rate": 1.4285714285714285e-05, + "loss": 9.1605, + "theoretical_loss": 12.38505930450469, + "tokens_seen": 589824 + }, + { + "epoch": 0.0, + "learning_rate": 1.5873015873015872e-05, + "loss": 9.0261, + "theoretical_loss": 12.095879447666144, + "tokens_seen": 655360 + }, + { + "epoch": 0.0, + "learning_rate": 1.746031746031746e-05, + "loss": 8.9446, + "theoretical_loss": 11.841533071732742, + "tokens_seen": 720896 + }, + { + "epoch": 0.0, + "learning_rate": 1.9047619047619046e-05, + "loss": 8.8253, + "theoretical_loss": 11.615186049337796, + "tokens_seen": 786432 + }, + { + "epoch": 0.0, + "learning_rate": 2.0634920634920636e-05, + "loss": 8.7592, + "theoretical_loss": 11.411780980817854, + "tokens_seen": 851968 + }, + { + "epoch": 0.0, + "learning_rate": 2.2222222222222223e-05, + "loss": 8.6786, + "theoretical_loss": 11.227478542742938, + "tokens_seen": 917504 + }, + { + "epoch": 0.0, + "learning_rate": 2.380952380952381e-05, + "loss": 8.5982, + "theoretical_loss": 11.059300758608979, + "tokens_seen": 983040 + }, + { + "epoch": 0.0, + "learning_rate": 2.5396825396825397e-05, + "loss": 8.5607, + "theoretical_loss": 10.904894927088016, + "tokens_seen": 1048576 + }, + { + "epoch": 0.0, + "learning_rate": 2.6984126984126984e-05, + "loss": 8.6523, + "theoretical_loss": 10.76237258069379, + "tokens_seen": 1114112 + }, + { + "epoch": 0.0, + "learning_rate": 2.857142857142857e-05, + "loss": 8.4586, + "theoretical_loss": 10.630196716861345, + "tokens_seen": 1179648 + }, + { + "epoch": 0.0, + "learning_rate": 3.0158730158730158e-05, + "loss": 8.4275, + "theoretical_loss": 10.507101012987707, + "tokens_seen": 1245184 + }, + { + "epoch": 0.0, + "learning_rate": 3.1746031746031745e-05, + "loss": 8.378, + "theoretical_loss": 10.392030784394397, + "tokens_seen": 1310720 + }, + { + "epoch": 0.0, + "learning_rate": 3.3333333333333335e-05, + "loss": 8.1728, + "theoretical_loss": 10.284099061527812, + "tokens_seen": 1376256 + }, + { + "epoch": 0.0, + "learning_rate": 3.492063492063492e-05, + "loss": 8.2605, + "theoretical_loss": 10.182553393901085, + "tokens_seen": 1441792 + }, + { + "epoch": 0.0, + "learning_rate": 3.650793650793651e-05, + "loss": 8.2034, + "theoretical_loss": 10.086750402363664, + "tokens_seen": 1507328 + }, + { + "epoch": 0.0, + "learning_rate": 3.809523809523809e-05, + "loss": 8.1691, + "theoretical_loss": 9.996136019471344, + "tokens_seen": 1572864 + }, + { + "epoch": 0.0, + "learning_rate": 3.968253968253968e-05, + "loss": 8.0089, + "theoretical_loss": 9.910229967024176, + "tokens_seen": 1638400 + }, + { + "epoch": 0.0, + "learning_rate": 4.126984126984127e-05, + "loss": 8.1184, + "theoretical_loss": 9.828613432171625, + "tokens_seen": 1703936 + }, + { + "epoch": 0.0, + "learning_rate": 4.2857142857142856e-05, + "loss": 8.0347, + "theoretical_loss": 9.75091918753851, + "tokens_seen": 1769472 + }, + { + "epoch": 0.0, + "learning_rate": 4.4444444444444447e-05, + "loss": 8.0262, + "theoretical_loss": 9.676823599712613, + "tokens_seen": 1835008 + }, + { + "epoch": 0.0, + "learning_rate": 4.603174603174603e-05, + "loss": 7.9121, + "theoretical_loss": 9.606040111776705, + "tokens_seen": 1900544 + }, + { + "epoch": 0.0, + "learning_rate": 4.761904761904762e-05, + "loss": 7.9007, + "theoretical_loss": 9.538313887395919, + "tokens_seen": 1966080 + }, + { + "epoch": 0.0, + "learning_rate": 4.9206349206349204e-05, + "loss": 7.867, + "theoretical_loss": 9.473417378263001, + "tokens_seen": 2031616 + }, + { + "epoch": 0.0, + "learning_rate": 5.0793650793650794e-05, + "loss": 7.8194, + "theoretical_loss": 9.411146631541524, + "tokens_seen": 2097152 + }, + { + "epoch": 0.0, + "learning_rate": 5.2380952380952384e-05, + "loss": 7.7465, + "theoretical_loss": 9.351318194865048, + "tokens_seen": 2162688 + }, + { + "epoch": 0.0, + "learning_rate": 5.396825396825397e-05, + "loss": 7.6701, + "theoretical_loss": 9.293766507291341, + "tokens_seen": 2228224 + }, + { + "epoch": 0.0, + "learning_rate": 5.555555555555556e-05, + "loss": 7.8228, + "theoretical_loss": 9.238341688075822, + "tokens_seen": 2293760 + }, + { + "epoch": 0.0, + "learning_rate": 5.714285714285714e-05, + "loss": 7.5608, + "theoretical_loss": 9.184907653139359, + "tokens_seen": 2359296 + }, + { + "epoch": 0.0, + "learning_rate": 5.873015873015873e-05, + "loss": 7.5692, + "theoretical_loss": 9.133340503044415, + "tokens_seen": 2424832 + }, + { + "epoch": 0.0, + "learning_rate": 6.0317460317460316e-05, + "loss": 7.4204, + "theoretical_loss": 9.0835271371648, + "tokens_seen": 2490368 + }, + { + "epoch": 0.0, + "learning_rate": 6.19047619047619e-05, + "loss": 7.4244, + "theoretical_loss": 9.035364057274894, + "tokens_seen": 2555904 + }, + { + "epoch": 0.0, + "learning_rate": 6.349206349206349e-05, + "loss": 7.4974, + "theoretical_loss": 8.988756330540422, + "tokens_seen": 2621440 + }, + { + "epoch": 0.0, + "learning_rate": 6.507936507936509e-05, + "loss": 7.3008, + "theoretical_loss": 8.943616687271678, + "tokens_seen": 2686976 + }, + { + "epoch": 0.0, + "learning_rate": 6.666666666666667e-05, + "loss": 7.3011, + "theoretical_loss": 8.89986473310929, + "tokens_seen": 2752512 + }, + { + "epoch": 0.0, + "learning_rate": 6.825396825396825e-05, + "loss": 7.1055, + "theoretical_loss": 8.857426258784564, + "tokens_seen": 2818048 + }, + { + "epoch": 0.0, + "learning_rate": 6.984126984126984e-05, + "loss": 7.0704, + "theoretical_loss": 8.816232633409479, + "tokens_seen": 2883584 + }, + { + "epoch": 0.0, + "learning_rate": 7.142857142857143e-05, + "loss": 7.2394, + "theoretical_loss": 8.776220269542522, + "tokens_seen": 2949120 + }, + { + "epoch": 0.0, + "learning_rate": 7.301587301587302e-05, + "loss": 7.0576, + "theoretical_loss": 8.737330150151898, + "tokens_seen": 3014656 + }, + { + "epoch": 0.0, + "learning_rate": 7.460317460317461e-05, + "loss": 7.0347, + "theoretical_loss": 8.699507409139912, + "tokens_seen": 3080192 + }, + { + "epoch": 0.0, + "learning_rate": 7.619047619047618e-05, + "loss": 7.0122, + "theoretical_loss": 8.662700958366539, + "tokens_seen": 3145728 + }, + { + "epoch": 0.0, + "learning_rate": 7.777777777777778e-05, + "loss": 6.9111, + "theoretical_loss": 8.62686315516739, + "tokens_seen": 3211264 + }, + { + "epoch": 0.0, + "objective/train/docs_used": 8371, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 6.933422565460205, + "objective/train/theoretical_loss": 8.591949505242134, + "objective/train/tokens_used": 23736800, + "theoretical_loss": 8.591949505242134, + "tokens_seen": 3276800 + }, + { + "epoch": 0.0, + "learning_rate": 7.936507936507937e-05, + "loss": 6.8487, + "theoretical_loss": 8.591949505242134, + "tokens_seen": 3276800 + }, + { + "epoch": 0.0, + "learning_rate": 8.095238095238096e-05, + "loss": 6.7315, + "theoretical_loss": 8.557918396526182, + "tokens_seen": 3342336 + }, + { + "epoch": 0.0, + "learning_rate": 8.253968253968255e-05, + "loss": 6.6564, + "theoretical_loss": 8.524730860277067, + "tokens_seen": 3407872 + }, + { + "epoch": 0.0, + "learning_rate": 8.412698412698413e-05, + "loss": 6.6878, + "theoretical_loss": 8.492350356128252, + "tokens_seen": 3473408 + }, + { + "epoch": 0.0, + "learning_rate": 8.571428571428571e-05, + "loss": 6.5932, + "theoretical_loss": 8.460742578303845, + "tokens_seen": 3538944 + }, + { + "epoch": 0.0, + "learning_rate": 8.730158730158731e-05, + "loss": 6.5965, + "theoretical_loss": 8.429875280561783, + "tokens_seen": 3604480 + }, + { + "epoch": 0.0, + "learning_rate": 8.888888888888889e-05, + "loss": 6.5835, + "theoretical_loss": 8.399718117751275, + "tokens_seen": 3670016 + }, + { + "epoch": 0.0, + "learning_rate": 9.047619047619048e-05, + "loss": 6.499, + "theoretical_loss": 8.370242502142142, + "tokens_seen": 3735552 + }, + { + "epoch": 0.0, + "learning_rate": 9.206349206349206e-05, + "loss": 6.4711, + "theoretical_loss": 8.341421472916394, + "tokens_seen": 3801088 + }, + { + "epoch": 0.0, + "learning_rate": 9.365079365079366e-05, + "loss": 6.2799, + "theoretical_loss": 8.313229577412248, + "tokens_seen": 3866624 + }, + { + "epoch": 0.0, + "learning_rate": 9.523809523809524e-05, + "loss": 6.3054, + "theoretical_loss": 8.28564276288293, + "tokens_seen": 3932160 + }, + { + "epoch": 0.0, + "learning_rate": 9.682539682539682e-05, + "loss": 6.3407, + "theoretical_loss": 8.258638277681179, + "tokens_seen": 3997696 + }, + { + "epoch": 0.0, + "learning_rate": 9.841269841269841e-05, + "loss": 6.1946, + "theoretical_loss": 8.232194580909036, + "tokens_seen": 4063232 + }, + { + "epoch": 0.0, + "learning_rate": 0.0001, + "loss": 6.228, + "theoretical_loss": 8.20629125968411, + "tokens_seen": 4128768 + }, + { + "epoch": 0.0, + "learning_rate": 0.00010158730158730159, + "loss": 6.2701, + "theoretical_loss": 8.180908953270682, + "tokens_seen": 4194304 + }, + { + "epoch": 0.0, + "learning_rate": 0.00010317460317460319, + "loss": 6.1144, + "theoretical_loss": 8.156029283408564, + "tokens_seen": 4259840 + }, + { + "epoch": 0.0, + "learning_rate": 0.00010476190476190477, + "loss": 6.1101, + "theoretical_loss": 8.131634790246775, + "tokens_seen": 4325376 + }, + { + "epoch": 0.0, + "learning_rate": 0.00010634920634920635, + "loss": 6.1424, + "theoretical_loss": 8.107708873353706, + "tokens_seen": 4390912 + }, + { + "epoch": 0.0, + "learning_rate": 0.00010793650793650794, + "loss": 6.0276, + "theoretical_loss": 8.084235737332481, + "tokens_seen": 4456448 + }, + { + "epoch": 0.0, + "learning_rate": 0.00010952380952380953, + "loss": 5.9829, + "theoretical_loss": 8.061200341620074, + "tokens_seen": 4521984 + }, + { + "epoch": 0.0, + "learning_rate": 0.00011111111111111112, + "loss": 5.9884, + "theoretical_loss": 8.038588354092902, + "tokens_seen": 4587520 + }, + { + "epoch": 0.0, + "learning_rate": 0.0001126984126984127, + "loss": 5.8108, + "theoretical_loss": 8.016386108140445, + "tokens_seen": 4653056 + }, + { + "epoch": 0.0, + "learning_rate": 0.00011428571428571428, + "loss": 5.9521, + "theoretical_loss": 7.994580562902867, + "tokens_seen": 4718592 + }, + { + "epoch": 0.0, + "learning_rate": 0.0001158730158730159, + "loss": 6.0278, + "theoretical_loss": 7.973159266399054, + "tokens_seen": 4784128 + }, + { + "epoch": 0.0, + "learning_rate": 0.00011746031746031746, + "loss": 5.8612, + "theoretical_loss": 7.952110321298584, + "tokens_seen": 4849664 + }, + { + "epoch": 0.0, + "learning_rate": 0.00011904761904761905, + "loss": 5.9738, + "theoretical_loss": 7.931422353115133, + "tokens_seen": 4915200 + }, + { + "epoch": 0.0, + "learning_rate": 0.00012063492063492063, + "loss": 5.8284, + "theoretical_loss": 7.911084480620269, + "tokens_seen": 4980736 + }, + { + "epoch": 0.0, + "learning_rate": 0.00012222222222222224, + "loss": 5.7613, + "theoretical_loss": 7.891086288295694, + "tokens_seen": 5046272 + }, + { + "epoch": 0.0, + "learning_rate": 0.0001238095238095238, + "loss": 5.863, + "theoretical_loss": 7.871417800659003, + "tokens_seen": 5111808 + }, + { + "epoch": 0.0, + "learning_rate": 0.0001253968253968254, + "loss": 5.8605, + "theoretical_loss": 7.8520694583133785, + "tokens_seen": 5177344 + }, + { + "epoch": 0.0, + "learning_rate": 0.00012698412698412698, + "loss": 5.6894, + "theoretical_loss": 7.833032095585231, + "tokens_seen": 5242880 + }, + { + "epoch": 0.0, + "learning_rate": 0.00012857142857142858, + "loss": 5.735, + "theoretical_loss": 7.8142969196261385, + "tokens_seen": 5308416 + }, + { + "epoch": 0.0, + "learning_rate": 0.00013015873015873017, + "loss": 5.6694, + "theoretical_loss": 7.79585549086643, + "tokens_seen": 5373952 + }, + { + "epoch": 0.0, + "learning_rate": 0.00013174603174603174, + "loss": 5.8253, + "theoretical_loss": 7.777699704717685, + "tokens_seen": 5439488 + }, + { + "epoch": 0.0, + "learning_rate": 0.00013333333333333334, + "loss": 5.8306, + "theoretical_loss": 7.759821774430384, + "tokens_seen": 5505024 + }, + { + "epoch": 0.0, + "learning_rate": 0.00013492063492063494, + "loss": 5.7255, + "theoretical_loss": 7.742214215020986, + "tokens_seen": 5570560 + }, + { + "epoch": 0.0, + "learning_rate": 0.0001365079365079365, + "loss": 5.6653, + "theoretical_loss": 7.72486982819, + "tokens_seen": 5636096 + }, + { + "epoch": 0.0, + "learning_rate": 0.0001380952380952381, + "loss": 5.6049, + "theoretical_loss": 7.707781688159247, + "tokens_seen": 5701632 + }, + { + "epoch": 0.0, + "learning_rate": 0.00013968253968253967, + "loss": 5.6135, + "theoretical_loss": 7.690943128362394, + "tokens_seen": 5767168 + }, + { + "epoch": 0.0, + "learning_rate": 0.0001412698412698413, + "loss": 5.6258, + "theoretical_loss": 7.674347728928345, + "tokens_seen": 5832704 + }, + { + "epoch": 0.0, + "learning_rate": 0.00014285714285714287, + "loss": 5.5429, + "theoretical_loss": 7.6579893049019265, + "tokens_seen": 5898240 + }, + { + "epoch": 0.0, + "learning_rate": 0.00014444444444444444, + "loss": 5.6666, + "theoretical_loss": 7.641861895150762, + "tokens_seen": 5963776 + }, + { + "epoch": 0.0, + "learning_rate": 0.00014603174603174603, + "loss": 5.6324, + "theoretical_loss": 7.625959751911346, + "tokens_seen": 6029312 + }, + { + "epoch": 0.0, + "learning_rate": 0.00014761904761904763, + "loss": 5.6327, + "theoretical_loss": 7.610277330930954, + "tokens_seen": 6094848 + }, + { + "epoch": 0.0, + "learning_rate": 0.00014920634920634923, + "loss": 5.5506, + "theoretical_loss": 7.5948092821654365, + "tokens_seen": 6160384 + }, + { + "epoch": 0.0, + "learning_rate": 0.0001507936507936508, + "loss": 5.538, + "theoretical_loss": 7.579550440996015, + "tokens_seen": 6225920 + }, + { + "epoch": 0.0, + "learning_rate": 0.00015238095238095237, + "loss": 5.6126, + "theoretical_loss": 7.564495819930985, + "tokens_seen": 6291456 + }, + { + "epoch": 0.0, + "learning_rate": 0.000153968253968254, + "loss": 5.5712, + "theoretical_loss": 7.549640600760828, + "tokens_seen": 6356992 + }, + { + "epoch": 0.0, + "learning_rate": 0.00015555555555555556, + "loss": 5.4686, + "theoretical_loss": 7.534980127137563, + "tokens_seen": 6422528 + }, + { + "epoch": 0.0, + "learning_rate": 0.00015714285714285716, + "loss": 5.474, + "theoretical_loss": 7.520509897551315, + "tokens_seen": 6488064 + }, + { + "epoch": 0.0, + "objective/train/docs_used": 9704, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 5.4859185218811035, + "objective/train/theoretical_loss": 7.5062255586790965, + "objective/train/tokens_used": 27013600, + "theoretical_loss": 7.5062255586790965, + "tokens_seen": 6553600 + }, + { + "epoch": 0.0, + "learning_rate": 0.00015873015873015873, + "loss": 5.5409, + "theoretical_loss": 7.5062255586790965, + "tokens_seen": 6553600 + }, + { + "epoch": 0.0, + "learning_rate": 0.00016031746031746033, + "loss": 5.5717, + "theoretical_loss": 7.492122899082528, + "tokens_seen": 6619136 + }, + { + "epoch": 0.0, + "learning_rate": 0.00016190476190476192, + "loss": 5.317, + "theoretical_loss": 7.478197843232982, + "tokens_seen": 6684672 + }, + { + "epoch": 0.0, + "learning_rate": 0.0001634920634920635, + "loss": 5.4912, + "theoretical_loss": 7.464446445844085, + "tokens_seen": 6750208 + }, + { + "epoch": 0.0, + "learning_rate": 0.0001650793650793651, + "loss": 5.5295, + "theoretical_loss": 7.450864886492971, + "tokens_seen": 6815744 + }, + { + "epoch": 0.0, + "learning_rate": 0.0001666666666666667, + "loss": 5.2981, + "theoretical_loss": 7.437449464512941, + "tokens_seen": 6881280 + }, + { + "epoch": 0.0, + "learning_rate": 0.00016825396825396826, + "loss": 5.3466, + "theoretical_loss": 7.424196594141411, + "tokens_seen": 6946816 + }, + { + "epoch": 0.0, + "learning_rate": 0.00016984126984126986, + "loss": 5.4795, + "theoretical_loss": 7.4111027999080985, + "tokens_seen": 7012352 + }, + { + "epoch": 0.0, + "learning_rate": 0.00017142857142857143, + "loss": 5.5522, + "theoretical_loss": 7.398164712249439, + "tokens_seen": 7077888 + }, + { + "epoch": 0.0, + "learning_rate": 0.00017301587301587302, + "loss": 5.37, + "theoretical_loss": 7.3853790633361545, + "tokens_seen": 7143424 + }, + { + "epoch": 0.0, + "learning_rate": 0.00017460317460317462, + "loss": 5.3812, + "theoretical_loss": 7.3727426831017535, + "tokens_seen": 7208960 + }, + { + "epoch": 0.0, + "learning_rate": 0.0001761904761904762, + "loss": 5.4132, + "theoretical_loss": 7.360252495460585, + "tokens_seen": 7274496 + }, + { + "epoch": 0.0, + "learning_rate": 0.00017777777777777779, + "loss": 5.3813, + "theoretical_loss": 7.347905514704743, + "tokens_seen": 7340032 + }, + { + "epoch": 0.0, + "learning_rate": 0.00017936507936507938, + "loss": 5.3924, + "theoretical_loss": 7.33569884206989, + "tokens_seen": 7405568 + }, + { + "epoch": 0.0, + "learning_rate": 0.00018095238095238095, + "loss": 5.4972, + "theoretical_loss": 7.32362966246063, + "tokens_seen": 7471104 + }, + { + "epoch": 0.0, + "learning_rate": 0.00018253968253968255, + "loss": 5.2111, + "theoretical_loss": 7.311695241326718, + "tokens_seen": 7536640 + }, + { + "epoch": 0.0, + "learning_rate": 0.00018412698412698412, + "loss": 5.3073, + "theoretical_loss": 7.299892921681898, + "tokens_seen": 7602176 + }, + { + "epoch": 0.0, + "learning_rate": 0.00018571428571428572, + "loss": 5.4158, + "theoretical_loss": 7.288220121257689, + "tokens_seen": 7667712 + }, + { + "epoch": 0.0, + "learning_rate": 0.00018730158730158731, + "loss": 5.3593, + "theoretical_loss": 7.276674329784936, + "tokens_seen": 7733248 + }, + { + "epoch": 0.0, + "learning_rate": 0.00018888888888888888, + "loss": 5.3053, + "theoretical_loss": 7.265253106396328, + "tokens_seen": 7798784 + }, + { + "epoch": 0.0, + "learning_rate": 0.00019047619047619048, + "loss": 5.4652, + "theoretical_loss": 7.253954077143563, + "tokens_seen": 7864320 + }, + { + "epoch": 0.0, + "learning_rate": 0.00019206349206349208, + "loss": 5.3082, + "theoretical_loss": 7.242774932623186, + "tokens_seen": 7929856 + }, + { + "epoch": 0.0, + "learning_rate": 0.00019365079365079365, + "loss": 5.4536, + "theoretical_loss": 7.2317134257054665, + "tokens_seen": 7995392 + }, + { + "epoch": 0.0, + "learning_rate": 0.00019523809523809525, + "loss": 5.1461, + "theoretical_loss": 7.220767369361081, + "tokens_seen": 8060928 + }, + { + "epoch": 0.0, + "learning_rate": 0.00019682539682539682, + "loss": 5.2898, + "theoretical_loss": 7.2099346345805815, + "tokens_seen": 8126464 + }, + { + "epoch": 0.0, + "learning_rate": 0.00019841269841269844, + "loss": 5.326, + "theoretical_loss": 7.1992131483820305, + "tokens_seen": 8192000 + }, + { + "epoch": 0.0, + "learning_rate": 0.0002, + "loss": 5.2216, + "theoretical_loss": 7.188600891902338, + "tokens_seen": 8257536 + }, + { + "epoch": 0.0, + "learning_rate": 0.00020158730158730158, + "loss": 5.2558, + "theoretical_loss": 7.178095898568204, + "tokens_seen": 8323072 + }, + { + "epoch": 0.0, + "learning_rate": 0.00020317460317460318, + "loss": 5.1693, + "theoretical_loss": 7.167696252342704, + "tokens_seen": 8388608 + }, + { + "epoch": 0.0, + "learning_rate": 0.00020476190476190475, + "loss": 5.1991, + "theoretical_loss": 7.15740008604386, + "tokens_seen": 8454144 + }, + { + "epoch": 0.0, + "learning_rate": 0.00020634920634920637, + "loss": 5.2315, + "theoretical_loss": 7.147205579731682, + "tokens_seen": 8519680 + }, + { + "epoch": 0.0, + "learning_rate": 0.00020793650793650797, + "loss": 5.2062, + "theoretical_loss": 7.1371109591604105, + "tokens_seen": 8585216 + }, + { + "epoch": 0.0, + "learning_rate": 0.00020952380952380954, + "loss": 5.1778, + "theoretical_loss": 7.127114494292836, + "tokens_seen": 8650752 + }, + { + "epoch": 0.0, + "learning_rate": 0.00021111111111111113, + "loss": 5.0834, + "theoretical_loss": 7.117214497873759, + "tokens_seen": 8716288 + }, + { + "epoch": 0.0, + "learning_rate": 0.0002126984126984127, + "loss": 5.0543, + "theoretical_loss": 7.107409324059815, + "tokens_seen": 8781824 + }, + { + "epoch": 0.0, + "learning_rate": 0.0002142857142857143, + "loss": 5.3091, + "theoretical_loss": 7.097697367103032, + "tokens_seen": 8847360 + }, + { + "epoch": 0.0, + "learning_rate": 0.00021587301587301587, + "loss": 5.1001, + "theoretical_loss": 7.0880770600856335, + "tokens_seen": 8912896 + }, + { + "epoch": 0.0, + "learning_rate": 0.00021746031746031744, + "loss": 4.9668, + "theoretical_loss": 7.0785468737037345, + "tokens_seen": 8978432 + }, + { + "epoch": 0.0, + "learning_rate": 0.00021904761904761907, + "loss": 5.2415, + "theoretical_loss": 7.069105315097691, + "tokens_seen": 9043968 + }, + { + "epoch": 0.0, + "learning_rate": 0.00022063492063492066, + "loss": 5.0942, + "theoretical_loss": 7.059750926726999, + "tokens_seen": 9109504 + }, + { + "epoch": 0.0, + "learning_rate": 0.00022222222222222223, + "loss": 5.1902, + "theoretical_loss": 7.050482285287744, + "tokens_seen": 9175040 + }, + { + "epoch": 0.0, + "learning_rate": 0.00022380952380952383, + "loss": 5.1624, + "theoretical_loss": 7.041298000670675, + "tokens_seen": 9240576 + }, + { + "epoch": 0.0, + "learning_rate": 0.0002253968253968254, + "loss": 5.1738, + "theoretical_loss": 7.032196714958137, + "tokens_seen": 9306112 + }, + { + "epoch": 0.0, + "learning_rate": 0.000226984126984127, + "loss": 5.1272, + "theoretical_loss": 7.023177101458123, + "tokens_seen": 9371648 + }, + { + "epoch": 0.0, + "learning_rate": 0.00022857142857142857, + "loss": 5.1345, + "theoretical_loss": 7.014237863773846, + "tokens_seen": 9437184 + }, + { + "epoch": 0.0, + "learning_rate": 0.00023015873015873016, + "loss": 5.0756, + "theoretical_loss": 7.005377734907255, + "tokens_seen": 9502720 + }, + { + "epoch": 0.0, + "learning_rate": 0.0002317460317460318, + "loss": 5.1274, + "theoretical_loss": 6.996595476395081, + "tokens_seen": 9568256 + }, + { + "epoch": 0.0, + "learning_rate": 0.00023333333333333336, + "loss": 5.1703, + "theoretical_loss": 6.98788987747596, + "tokens_seen": 9633792 + }, + { + "epoch": 0.0, + "learning_rate": 0.00023492063492063493, + "loss": 5.1119, + "theoretical_loss": 6.979259754287375, + "tokens_seen": 9699328 + }, + { + "epoch": 0.0, + "learning_rate": 0.00023650793650793653, + "loss": 5.0673, + "theoretical_loss": 6.970703949091095, + "tokens_seen": 9764864 + }, + { + "epoch": 0.0, + "objective/train/docs_used": 11015, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 5.158774375915527, + "objective/train/theoretical_loss": 6.962221329525966, + "objective/train/tokens_used": 30290400, + "theoretical_loss": 6.962221329525966, + "tokens_seen": 9830400 + }, + { + "epoch": 0.0, + "learning_rate": 0.0002380952380952381, + "loss": 5.141, + "theoretical_loss": 6.962221329525966, + "tokens_seen": 9830400 + }, + { + "epoch": 0.0, + "learning_rate": 0.0002396825396825397, + "loss": 5.119, + "theoretical_loss": 6.953810787886883, + "tokens_seen": 9895936 + }, + { + "epoch": 0.0, + "learning_rate": 0.00024126984126984126, + "loss": 5.0928, + "theoretical_loss": 6.945471240428864, + "tokens_seen": 9961472 + }, + { + "epoch": 0.0, + "learning_rate": 0.00024285714285714286, + "loss": 5.1011, + "theoretical_loss": 6.937201626695213, + "tokens_seen": 10027008 + }, + { + "epoch": 0.0, + "learning_rate": 0.0002444444444444445, + "loss": 5.1427, + "theoretical_loss": 6.929000908868754, + "tokens_seen": 10092544 + }, + { + "epoch": 0.0, + "learning_rate": 0.00024603174603174605, + "loss": 5.0633, + "theoretical_loss": 6.920868071145226, + "tokens_seen": 10158080 + }, + { + "epoch": 0.0, + "learning_rate": 0.0002476190476190476, + "loss": 5.1099, + "theoretical_loss": 6.912802119127937, + "tokens_seen": 10223616 + }, + { + "epoch": 0.0, + "learning_rate": 0.00024920634920634925, + "loss": 5.2568, + "theoretical_loss": 6.904802079242816, + "tokens_seen": 10289152 + }, + { + "epoch": 0.0, + "learning_rate": 0.0002507936507936508, + "loss": 4.9779, + "theoretical_loss": 6.896866998173063, + "tokens_seen": 10354688 + }, + { + "epoch": 0.0, + "learning_rate": 0.0002523809523809524, + "loss": 5.091, + "theoretical_loss": 6.888995942312617, + "tokens_seen": 10420224 + }, + { + "epoch": 0.0, + "learning_rate": 0.00025396825396825396, + "loss": 5.0945, + "theoretical_loss": 6.881187997237696, + "tokens_seen": 10485760 + }, + { + "epoch": 0.0, + "learning_rate": 0.00025555555555555553, + "loss": 5.0332, + "theoretical_loss": 6.873442267195708, + "tokens_seen": 10551296 + }, + { + "epoch": 0.0, + "learning_rate": 0.00025714285714285715, + "loss": 5.2141, + "theoretical_loss": 6.8657578746108605, + "tokens_seen": 10616832 + }, + { + "epoch": 0.0, + "learning_rate": 0.0002587301587301588, + "loss": 4.865, + "theoretical_loss": 6.858133959605815, + "tokens_seen": 10682368 + }, + { + "epoch": 0.0, + "learning_rate": 0.00026031746031746035, + "loss": 4.9942, + "theoretical_loss": 6.8505696795387845, + "tokens_seen": 10747904 + }, + { + "epoch": 0.0, + "learning_rate": 0.0002619047619047619, + "loss": 4.996, + "theoretical_loss": 6.843064208555461, + "tokens_seen": 10813440 + }, + { + "epoch": 0.0, + "learning_rate": 0.0002634920634920635, + "loss": 5.0363, + "theoretical_loss": 6.835616737155247, + "tokens_seen": 10878976 + }, + { + "epoch": 0.0, + "learning_rate": 0.0002650793650793651, + "loss": 4.8561, + "theoretical_loss": 6.828226471771211, + "tokens_seen": 10944512 + }, + { + "epoch": 0.0, + "learning_rate": 0.0002666666666666667, + "loss": 4.9455, + "theoretical_loss": 6.8208926343632905, + "tokens_seen": 11010048 + }, + { + "epoch": 0.0, + "learning_rate": 0.00026825396825396825, + "loss": 4.7407, + "theoretical_loss": 6.813614462024214, + "tokens_seen": 11075584 + }, + { + "epoch": 0.0, + "learning_rate": 0.0002698412698412699, + "loss": 5.0034, + "theoretical_loss": 6.806391206597704, + "tokens_seen": 11141120 + }, + { + "epoch": 0.0, + "learning_rate": 0.00027142857142857144, + "loss": 4.9295, + "theoretical_loss": 6.799222134308489, + "tokens_seen": 11206656 + }, + { + "epoch": 0.0, + "learning_rate": 0.000273015873015873, + "loss": 5.0309, + "theoretical_loss": 6.792106525403691, + "tokens_seen": 11272192 + }, + { + "epoch": 0.0, + "learning_rate": 0.00027460317460317464, + "loss": 4.8622, + "theoretical_loss": 6.785043673805189, + "tokens_seen": 11337728 + }, + { + "epoch": 0.0, + "learning_rate": 0.0002761904761904762, + "loss": 4.9395, + "theoretical_loss": 6.778032886772552, + "tokens_seen": 11403264 + }, + { + "epoch": 0.0, + "learning_rate": 0.0002777777777777778, + "loss": 5.0467, + "theoretical_loss": 6.771073484576156, + "tokens_seen": 11468800 + }, + { + "epoch": 0.0, + "learning_rate": 0.00027936507936507935, + "loss": 4.9171, + "theoretical_loss": 6.764164800180141, + "tokens_seen": 11534336 + }, + { + "epoch": 0.0, + "learning_rate": 0.00028095238095238097, + "loss": 4.843, + "theoretical_loss": 6.757306178934827, + "tokens_seen": 11599872 + }, + { + "epoch": 0.0, + "learning_rate": 0.0002825396825396826, + "loss": 4.9291, + "theoretical_loss": 6.750496978278292, + "tokens_seen": 11665408 + }, + { + "epoch": 0.0, + "learning_rate": 0.00028412698412698417, + "loss": 4.8907, + "theoretical_loss": 6.743736567446755, + "tokens_seen": 11730944 + }, + { + "epoch": 0.0, + "learning_rate": 0.00028571428571428574, + "loss": 4.7503, + "theoretical_loss": 6.737024327193478, + "tokens_seen": 11796480 + }, + { + "epoch": 0.0, + "learning_rate": 0.0002873015873015873, + "loss": 4.7646, + "theoretical_loss": 6.73035964951589, + "tokens_seen": 11862016 + }, + { + "epoch": 0.0, + "learning_rate": 0.0002888888888888889, + "loss": 4.9551, + "theoretical_loss": 6.723741937390625, + "tokens_seen": 11927552 + }, + { + "epoch": 0.0, + "learning_rate": 0.0002904761904761905, + "loss": 5.0033, + "theoretical_loss": 6.717170604516246, + "tokens_seen": 11993088 + }, + { + "epoch": 0.0, + "learning_rate": 0.00029206349206349207, + "loss": 4.8948, + "theoretical_loss": 6.71064507506334, + "tokens_seen": 12058624 + }, + { + "epoch": 0.0, + "learning_rate": 0.00029365079365079364, + "loss": 4.8931, + "theoretical_loss": 6.704164783431781, + "tokens_seen": 12124160 + }, + { + "epoch": 0.0, + "learning_rate": 0.00029523809523809526, + "loss": 4.8743, + "theoretical_loss": 6.6977291740148805, + "tokens_seen": 12189696 + }, + { + "epoch": 0.0, + "learning_rate": 0.00029682539682539683, + "loss": 4.7657, + "theoretical_loss": 6.691337700970224, + "tokens_seen": 12255232 + }, + { + "epoch": 0.0, + "learning_rate": 0.00029841269841269846, + "loss": 5.0346, + "theoretical_loss": 6.684989827996945, + "tokens_seen": 12320768 + }, + { + "epoch": 0.0, + "learning_rate": 0.00030000000000000003, + "loss": 5.011, + "theoretical_loss": 6.6786850281192365, + "tokens_seen": 12386304 + }, + { + "epoch": 0.0, + "learning_rate": 0.0003015873015873016, + "loss": 4.8876, + "theoretical_loss": 6.672422783475897, + "tokens_seen": 12451840 + }, + { + "epoch": 0.0, + "learning_rate": 0.00030317460317460317, + "loss": 5.0622, + "theoretical_loss": 6.666202585115692, + "tokens_seen": 12517376 + }, + { + "epoch": 0.0, + "learning_rate": 0.00030476190476190474, + "loss": 4.9138, + "theoretical_loss": 6.66002393279837, + "tokens_seen": 12582912 + }, + { + "epoch": 0.0, + "learning_rate": 0.00030634920634920636, + "loss": 4.8948, + "theoretical_loss": 6.653886334801115, + "tokens_seen": 12648448 + }, + { + "epoch": 0.0, + "learning_rate": 0.000307936507936508, + "loss": 4.7958, + "theoretical_loss": 6.647789307730289, + "tokens_seen": 12713984 + }, + { + "epoch": 0.0, + "learning_rate": 0.00030952380952380956, + "loss": 4.8338, + "theoretical_loss": 6.641732376338277, + "tokens_seen": 12779520 + }, + { + "epoch": 0.0, + "learning_rate": 0.0003111111111111111, + "loss": 4.8447, + "theoretical_loss": 6.6357150733452785, + "tokens_seen": 12845056 + }, + { + "epoch": 0.0, + "learning_rate": 0.0003126984126984127, + "loss": 4.8269, + "theoretical_loss": 6.629736939265876, + "tokens_seen": 12910592 + }, + { + "epoch": 0.0, + "learning_rate": 0.0003142857142857143, + "loss": 4.8123, + "theoretical_loss": 6.623797522240246, + "tokens_seen": 12976128 + }, + { + "epoch": 0.0, + "learning_rate": 0.0003158730158730159, + "loss": 4.8286, + "theoretical_loss": 6.617896377869851, + "tokens_seen": 13041664 + }, + { + "epoch": 0.0, + "objective/train/docs_used": 12170, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 4.938109397888184, + "objective/train/theoretical_loss": 6.61203306905748, + "objective/train/tokens_used": 33567200, + "theoretical_loss": 6.61203306905748, + "tokens_seen": 13107200 + }, + { + "epoch": 0.0, + "learning_rate": 0.00031746031746031746, + "loss": 4.8842, + "theoretical_loss": 6.61203306905748, + "tokens_seen": 13107200 + }, + { + "epoch": 0.0, + "learning_rate": 0.0003190476190476191, + "loss": 4.712, + "theoretical_loss": 6.6062071658514965, + "tokens_seen": 13172736 + }, + { + "epoch": 0.0, + "learning_rate": 0.00032063492063492065, + "loss": 4.9187, + "theoretical_loss": 6.600418245294164, + "tokens_seen": 13238272 + }, + { + "epoch": 0.0, + "learning_rate": 0.0003222222222222222, + "loss": 4.771, + "theoretical_loss": 6.594665891273929, + "tokens_seen": 13303808 + }, + { + "epoch": 0.0, + "learning_rate": 0.00032380952380952385, + "loss": 4.9402, + "theoretical_loss": 6.588949694381529, + "tokens_seen": 13369344 + }, + { + "epoch": 0.0, + "learning_rate": 0.0003253968253968254, + "loss": 4.7856, + "theoretical_loss": 6.583269251769811, + "tokens_seen": 13434880 + }, + { + "epoch": 0.0, + "learning_rate": 0.000326984126984127, + "loss": 4.8638, + "theoretical_loss": 6.577624167017157, + "tokens_seen": 13500416 + }, + { + "epoch": 0.0, + "learning_rate": 0.00032857142857142856, + "loss": 4.7116, + "theoretical_loss": 6.5720140499943955, + "tokens_seen": 13565952 + }, + { + "epoch": 0.0, + "learning_rate": 0.0003301587301587302, + "loss": 4.7373, + "theoretical_loss": 6.566438516735092, + "tokens_seen": 13631488 + }, + { + "epoch": 0.0, + "learning_rate": 0.0003317460317460318, + "loss": 4.724, + "theoretical_loss": 6.560897189309134, + "tokens_seen": 13697024 + }, + { + "epoch": 0.0, + "learning_rate": 0.0003333333333333334, + "loss": 4.6804, + "theoretical_loss": 6.5553896956994855, + "tokens_seen": 13762560 + }, + { + "epoch": 0.0, + "learning_rate": 0.00033492063492063495, + "loss": 4.7364, + "theoretical_loss": 6.5499156696820435, + "tokens_seen": 13828096 + }, + { + "epoch": 0.0, + "learning_rate": 0.0003365079365079365, + "loss": 4.7471, + "theoretical_loss": 6.544474750708482, + "tokens_seen": 13893632 + }, + { + "epoch": 0.0, + "learning_rate": 0.0003380952380952381, + "loss": 4.7404, + "theoretical_loss": 6.539066583792016, + "tokens_seen": 13959168 + }, + { + "epoch": 0.0, + "learning_rate": 0.0003396825396825397, + "loss": 4.8469, + "theoretical_loss": 6.533690819395976, + "tokens_seen": 14024704 + }, + { + "epoch": 0.0, + "learning_rate": 0.0003412698412698413, + "loss": 4.7664, + "theoretical_loss": 6.528347113325134, + "tokens_seen": 14090240 + }, + { + "epoch": 0.0, + "learning_rate": 0.00034285714285714285, + "loss": 4.7533, + "theoretical_loss": 6.523035126619684, + "tokens_seen": 14155776 + }, + { + "epoch": 0.0, + "learning_rate": 0.0003444444444444445, + "loss": 4.7784, + "theoretical_loss": 6.517754525451811, + "tokens_seen": 14221312 + }, + { + "epoch": 0.0, + "learning_rate": 0.00034603174603174604, + "loss": 4.7662, + "theoretical_loss": 6.512504981024765, + "tokens_seen": 14286848 + }, + { + "epoch": 0.0, + "learning_rate": 0.00034761904761904767, + "loss": 4.6805, + "theoretical_loss": 6.507286169474376, + "tokens_seen": 14352384 + }, + { + "epoch": 0.0, + "learning_rate": 0.00034920634920634924, + "loss": 4.7612, + "theoretical_loss": 6.502097771772935, + "tokens_seen": 14417920 + }, + { + "epoch": 0.0, + "learning_rate": 0.0003507936507936508, + "loss": 4.6998, + "theoretical_loss": 6.496939473635374, + "tokens_seen": 14483456 + }, + { + "epoch": 0.0, + "learning_rate": 0.0003523809523809524, + "loss": 4.5995, + "theoretical_loss": 6.491810965427682, + "tokens_seen": 14548992 + }, + { + "epoch": 0.0, + "learning_rate": 0.00035396825396825395, + "loss": 4.7017, + "theoretical_loss": 6.486711942077496, + "tokens_seen": 14614528 + }, + { + "epoch": 0.0, + "learning_rate": 0.00035555555555555557, + "loss": 4.7104, + "theoretical_loss": 6.481642102986796, + "tokens_seen": 14680064 + }, + { + "epoch": 0.0, + "learning_rate": 0.0003571428571428572, + "loss": 4.8033, + "theoretical_loss": 6.476601151946659, + "tokens_seen": 14745600 + }, + { + "epoch": 0.0, + "learning_rate": 0.00035873015873015877, + "loss": 4.6994, + "theoretical_loss": 6.471588797054005, + "tokens_seen": 14811136 + }, + { + "epoch": 0.0, + "learning_rate": 0.00036031746031746034, + "loss": 4.6984, + "theoretical_loss": 6.46660475063028, + "tokens_seen": 14876672 + }, + { + "epoch": 0.0, + "learning_rate": 0.0003619047619047619, + "loss": 4.7571, + "theoretical_loss": 6.46164872914203, + "tokens_seen": 14942208 + }, + { + "epoch": 0.0, + "learning_rate": 0.00036349206349206353, + "loss": 4.8314, + "theoretical_loss": 6.456720453123305, + "tokens_seen": 15007744 + }, + { + "epoch": 0.0, + "learning_rate": 0.0003650793650793651, + "loss": 4.7367, + "theoretical_loss": 6.4518196470998515, + "tokens_seen": 15073280 + }, + { + "epoch": 0.0, + "learning_rate": 0.00036666666666666667, + "loss": 4.6006, + "theoretical_loss": 6.446946039515035, + "tokens_seen": 15138816 + }, + { + "epoch": 0.0, + "learning_rate": 0.00036825396825396824, + "loss": 4.5752, + "theoretical_loss": 6.442099362657457, + "tokens_seen": 15204352 + }, + { + "epoch": 0.0, + "learning_rate": 0.00036984126984126986, + "loss": 4.4604, + "theoretical_loss": 6.437279352590206, + "tokens_seen": 15269888 + }, + { + "epoch": 0.0, + "learning_rate": 0.00037142857142857143, + "loss": 4.5994, + "theoretical_loss": 6.432485749081721, + "tokens_seen": 15335424 + }, + { + "epoch": 0.0, + "learning_rate": 0.00037301587301587306, + "loss": 4.6021, + "theoretical_loss": 6.427718295538199, + "tokens_seen": 15400960 + }, + { + "epoch": 0.0, + "learning_rate": 0.00037460317460317463, + "loss": 4.5965, + "theoretical_loss": 6.422976738937518, + "tokens_seen": 15466496 + }, + { + "epoch": 0.0, + "learning_rate": 0.0003761904761904762, + "loss": 4.7128, + "theoretical_loss": 6.418260829764641, + "tokens_seen": 15532032 + }, + { + "epoch": 0.0, + "learning_rate": 0.00037777777777777777, + "loss": 4.8108, + "theoretical_loss": 6.413570321948453, + "tokens_seen": 15597568 + }, + { + "epoch": 0.0, + "learning_rate": 0.0003793650793650794, + "loss": 4.5699, + "theoretical_loss": 6.40890497279999, + "tokens_seen": 15663104 + }, + { + "epoch": 0.0, + "learning_rate": 0.00038095238095238096, + "loss": 4.7258, + "theoretical_loss": 6.404264542952033, + "tokens_seen": 15728640 + }, + { + "epoch": 0.0, + "learning_rate": 0.0003825396825396826, + "loss": 4.591, + "theoretical_loss": 6.399648796300032, + "tokens_seen": 15794176 + }, + { + "epoch": 0.0, + "learning_rate": 0.00038412698412698416, + "loss": 4.5724, + "theoretical_loss": 6.395057499944313, + "tokens_seen": 15859712 + }, + { + "epoch": 0.0, + "learning_rate": 0.0003857142857142857, + "loss": 4.7205, + "theoretical_loss": 6.390490424133546, + "tokens_seen": 15925248 + }, + { + "epoch": 0.0, + "learning_rate": 0.0003873015873015873, + "loss": 4.7254, + "theoretical_loss": 6.3859473422094375, + "tokens_seen": 15990784 + }, + { + "epoch": 0.0, + "learning_rate": 0.0003888888888888889, + "loss": 4.7821, + "theoretical_loss": 6.381428030552619, + "tokens_seen": 16056320 + }, + { + "epoch": 0.0, + "learning_rate": 0.0003904761904761905, + "loss": 4.7415, + "theoretical_loss": 6.376932268529695, + "tokens_seen": 16121856 + }, + { + "epoch": 0.0, + "learning_rate": 0.00039206349206349206, + "loss": 4.7275, + "theoretical_loss": 6.372459838441432, + "tokens_seen": 16187392 + }, + { + "epoch": 0.0, + "learning_rate": 0.00039365079365079363, + "loss": 4.658, + "theoretical_loss": 6.368010525472036, + "tokens_seen": 16252928 + }, + { + "epoch": 0.0, + "learning_rate": 0.00039523809523809526, + "loss": 4.6939, + "theoretical_loss": 6.363584117639526, + "tokens_seen": 16318464 + }, + { + "epoch": 0.0, + "objective/train/docs_used": 13382, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 4.386369228363037, + "objective/train/theoretical_loss": 6.35918040574715, + "objective/train/tokens_used": 36844000, + "theoretical_loss": 6.35918040574715, + "tokens_seen": 16384000 + }, + { + "epoch": 0.0, + "learning_rate": 0.0003968253968253969, + "loss": 4.5742, + "theoretical_loss": 6.35918040574715, + "tokens_seen": 16384000 + }, + { + "epoch": 0.0, + "learning_rate": 0.00039841269841269845, + "loss": 4.5963, + "theoretical_loss": 6.354799183335816, + "tokens_seen": 16449536 + }, + { + "epoch": 0.01, + "learning_rate": 0.0004, + "loss": 4.4627, + "theoretical_loss": 6.350440246637536, + "tokens_seen": 16515072 + }, + { + "epoch": 0.01, + "learning_rate": 0.0004015873015873016, + "loss": 4.4934, + "theoretical_loss": 6.3461033945298295, + "tokens_seen": 16580608 + }, + { + "epoch": 0.01, + "learning_rate": 0.00040317460317460316, + "loss": 4.5596, + "theoretical_loss": 6.341788428491087, + "tokens_seen": 16646144 + }, + { + "epoch": 0.01, + "learning_rate": 0.0004047619047619048, + "loss": 4.5483, + "theoretical_loss": 6.337495152556851, + "tokens_seen": 16711680 + }, + { + "epoch": 0.01, + "learning_rate": 0.00040634920634920635, + "loss": 4.5704, + "theoretical_loss": 6.333223373277006, + "tokens_seen": 16777216 + }, + { + "epoch": 0.01, + "learning_rate": 0.0004079365079365079, + "loss": 4.7543, + "theoretical_loss": 6.32897289967384, + "tokens_seen": 16842752 + }, + { + "epoch": 0.01, + "learning_rate": 0.0004095238095238095, + "loss": 4.5719, + "theoretical_loss": 6.324743543200984, + "tokens_seen": 16908288 + }, + { + "epoch": 0.01, + "learning_rate": 0.0004111111111111111, + "loss": 4.5897, + "theoretical_loss": 6.320535117703164, + "tokens_seen": 16973824 + }, + { + "epoch": 0.01, + "learning_rate": 0.00041269841269841274, + "loss": 4.5878, + "theoretical_loss": 6.316347439376797, + "tokens_seen": 17039360 + }, + { + "epoch": 0.01, + "learning_rate": 0.00041428571428571437, + "loss": 4.5525, + "theoretical_loss": 6.312180326731362, + "tokens_seen": 17104896 + }, + { + "epoch": 0.01, + "learning_rate": 0.00041587301587301594, + "loss": 4.5865, + "theoretical_loss": 6.308033600551573, + "tokens_seen": 17170432 + }, + { + "epoch": 0.01, + "learning_rate": 0.0004174603174603175, + "loss": 4.3291, + "theoretical_loss": 6.303907083860294, + "tokens_seen": 17235968 + }, + { + "epoch": 0.01, + "learning_rate": 0.0004190476190476191, + "loss": 4.7171, + "theoretical_loss": 6.2998006018822075, + "tokens_seen": 17301504 + }, + { + "epoch": 0.01, + "learning_rate": 0.00042063492063492065, + "loss": 4.499, + "theoretical_loss": 6.295713982008203, + "tokens_seen": 17367040 + }, + { + "epoch": 0.01, + "learning_rate": 0.00042222222222222227, + "loss": 4.5934, + "theoretical_loss": 6.2916470537604745, + "tokens_seen": 17432576 + }, + { + "epoch": 0.01, + "learning_rate": 0.00042380952380952384, + "loss": 4.4748, + "theoretical_loss": 6.287599648758312, + "tokens_seen": 17498112 + }, + { + "epoch": 0.01, + "learning_rate": 0.0004253968253968254, + "loss": 4.3349, + "theoretical_loss": 6.283571600684564, + "tokens_seen": 17563648 + }, + { + "epoch": 0.01, + "learning_rate": 0.000426984126984127, + "loss": 4.2795, + "theoretical_loss": 6.279562745252756, + "tokens_seen": 17629184 + }, + { + "epoch": 0.01, + "learning_rate": 0.0004285714285714286, + "loss": 4.498, + "theoretical_loss": 6.2755729201748665, + "tokens_seen": 17694720 + }, + { + "epoch": 0.01, + "learning_rate": 0.0004301587301587302, + "loss": 4.62, + "theoretical_loss": 6.271601965129717, + "tokens_seen": 17760256 + }, + { + "epoch": 0.01, + "learning_rate": 0.00043174603174603174, + "loss": 4.5682, + "theoretical_loss": 6.267649721731981, + "tokens_seen": 17825792 + }, + { + "epoch": 0.01, + "learning_rate": 0.0004333333333333333, + "loss": 4.4555, + "theoretical_loss": 6.263716033501797, + "tokens_seen": 17891328 + }, + { + "epoch": 0.01, + "learning_rate": 0.0004349206349206349, + "loss": 4.3683, + "theoretical_loss": 6.259800745834964, + "tokens_seen": 17956864 + }, + { + "epoch": 0.01, + "learning_rate": 0.0004365079365079365, + "loss": 4.4385, + "theoretical_loss": 6.255903705973709, + "tokens_seen": 18022400 + }, + { + "epoch": 0.01, + "learning_rate": 0.00043809523809523813, + "loss": 4.497, + "theoretical_loss": 6.252024762978021, + "tokens_seen": 18087936 + }, + { + "epoch": 0.01, + "learning_rate": 0.00043968253968253976, + "loss": 4.4003, + "theoretical_loss": 6.248163767697523, + "tokens_seen": 18153472 + }, + { + "epoch": 0.01, + "learning_rate": 0.0004412698412698413, + "loss": 4.5891, + "theoretical_loss": 6.244320572743888, + "tokens_seen": 18219008 + }, + { + "epoch": 0.01, + "learning_rate": 0.0004428571428571429, + "loss": 4.4552, + "theoretical_loss": 6.240495032463768, + "tokens_seen": 18284544 + }, + { + "epoch": 0.01, + "learning_rate": 0.00044444444444444447, + "loss": 4.4253, + "theoretical_loss": 6.236687002912243, + "tokens_seen": 18350080 + }, + { + "epoch": 0.01, + "learning_rate": 0.0004460317460317461, + "loss": 4.4212, + "theoretical_loss": 6.232896341826766, + "tokens_seen": 18415616 + }, + { + "epoch": 0.01, + "learning_rate": 0.00044761904761904766, + "loss": 4.2641, + "theoretical_loss": 6.229122908601596, + "tokens_seen": 18481152 + }, + { + "epoch": 0.01, + "learning_rate": 0.00044920634920634923, + "loss": 4.5705, + "theoretical_loss": 6.225366564262705, + "tokens_seen": 18546688 + }, + { + "epoch": 0.01, + "learning_rate": 0.0004507936507936508, + "loss": 4.445, + "theoretical_loss": 6.2216271714431635, + "tokens_seen": 18612224 + }, + { + "epoch": 0.01, + "learning_rate": 0.00045238095238095237, + "loss": 4.4382, + "theoretical_loss": 6.217904594358968, + "tokens_seen": 18677760 + }, + { + "epoch": 0.01, + "learning_rate": 0.000453968253968254, + "loss": 4.4608, + "theoretical_loss": 6.214198698785328, + "tokens_seen": 18743296 + }, + { + "epoch": 0.01, + "learning_rate": 0.00045555555555555556, + "loss": 4.4226, + "theoretical_loss": 6.210509352033379, + "tokens_seen": 18808832 + }, + { + "epoch": 0.01, + "learning_rate": 0.00045714285714285713, + "loss": 4.5968, + "theoretical_loss": 6.206836422927337, + "tokens_seen": 18874368 + }, + { + "epoch": 0.01, + "learning_rate": 0.0004587301587301587, + "loss": 4.5992, + "theoretical_loss": 6.203179781782049, + "tokens_seen": 18939904 + }, + { + "epoch": 0.01, + "learning_rate": 0.00046031746031746033, + "loss": 4.5135, + "theoretical_loss": 6.199539300380979, + "tokens_seen": 19005440 + }, + { + "epoch": 0.01, + "learning_rate": 0.0004619047619047619, + "loss": 4.5009, + "theoretical_loss": 6.195914851954575, + "tokens_seen": 19070976 + }, + { + "epoch": 0.01, + "learning_rate": 0.0004634920634920636, + "loss": 4.4208, + "theoretical_loss": 6.19230631115903, + "tokens_seen": 19136512 + }, + { + "epoch": 0.01, + "learning_rate": 0.00046507936507936515, + "loss": 4.3844, + "theoretical_loss": 6.188713554055438, + "tokens_seen": 19202048 + }, + { + "epoch": 0.01, + "learning_rate": 0.0004666666666666667, + "loss": 4.2517, + "theoretical_loss": 6.185136458089309, + "tokens_seen": 19267584 + }, + { + "epoch": 0.01, + "learning_rate": 0.0004682539682539683, + "loss": 4.5069, + "theoretical_loss": 6.181574902070463, + "tokens_seen": 19333120 + }, + { + "epoch": 0.01, + "learning_rate": 0.00046984126984126986, + "loss": 4.4657, + "theoretical_loss": 6.178028766153277, + "tokens_seen": 19398656 + }, + { + "epoch": 0.01, + "learning_rate": 0.0004714285714285715, + "loss": 4.4703, + "theoretical_loss": 6.174497931817287, + "tokens_seen": 19464192 + }, + { + "epoch": 0.01, + "learning_rate": 0.00047301587301587305, + "loss": 4.4331, + "theoretical_loss": 6.17098228184813, + "tokens_seen": 19529728 + }, + { + "epoch": 0.01, + "learning_rate": 0.0004746031746031746, + "loss": 4.3217, + "theoretical_loss": 6.167481700318827, + "tokens_seen": 19595264 + }, + { + "epoch": 0.01, + "objective/train/docs_used": 14656, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 4.349937438964844, + "objective/train/theoretical_loss": 6.163996072571392, + "objective/train/tokens_used": 40120800, + "theoretical_loss": 6.163996072571392, + "tokens_seen": 19660800 + }, + { + "epoch": 0.01, + "learning_rate": 0.0004761904761904762, + "loss": 4.4227, + "theoretical_loss": 6.163996072571392, + "tokens_seen": 19660800 + }, + { + "epoch": 0.01, + "learning_rate": 0.0004777777777777778, + "loss": 4.4241, + "theoretical_loss": 6.160525285198771, + "tokens_seen": 19726336 + }, + { + "epoch": 0.01, + "learning_rate": 0.0004793650793650794, + "loss": 4.3427, + "theoretical_loss": 6.157069226027087, + "tokens_seen": 19791872 + }, + { + "epoch": 0.01, + "learning_rate": 0.00048095238095238095, + "loss": 4.2634, + "theoretical_loss": 6.153627784098204, + "tokens_seen": 19857408 + }, + { + "epoch": 0.01, + "learning_rate": 0.0004825396825396825, + "loss": 4.3541, + "theoretical_loss": 6.150200849652586, + "tokens_seen": 19922944 + }, + { + "epoch": 0.01, + "learning_rate": 0.0004841269841269841, + "loss": 4.3712, + "theoretical_loss": 6.146788314112464, + "tokens_seen": 19988480 + }, + { + "epoch": 0.01, + "learning_rate": 0.0004857142857142857, + "loss": 4.3159, + "theoretical_loss": 6.143390070065279, + "tokens_seen": 20054016 + }, + { + "epoch": 0.01, + "learning_rate": 0.0004873015873015873, + "loss": 4.4434, + "theoretical_loss": 6.140006011247426, + "tokens_seen": 20119552 + }, + { + "epoch": 0.01, + "learning_rate": 0.000488888888888889, + "loss": 4.5034, + "theoretical_loss": 6.136636032528262, + "tokens_seen": 20185088 + }, + { + "epoch": 0.01, + "learning_rate": 0.0004904761904761905, + "loss": 4.4359, + "theoretical_loss": 6.133280029894401, + "tokens_seen": 20250624 + }, + { + "epoch": 0.01, + "learning_rate": 0.0004920634920634921, + "loss": 4.4147, + "theoretical_loss": 6.129937900434269, + "tokens_seen": 20316160 + }, + { + "epoch": 0.01, + "learning_rate": 0.0004936507936507937, + "loss": 4.3241, + "theoretical_loss": 6.126609542322921, + "tokens_seen": 20381696 + }, + { + "epoch": 0.01, + "learning_rate": 0.0004952380952380952, + "loss": 4.313, + "theoretical_loss": 6.123294854807119, + "tokens_seen": 20447232 + }, + { + "epoch": 0.01, + "learning_rate": 0.0004968253968253969, + "loss": 4.4732, + "theoretical_loss": 6.11999373819066, + "tokens_seen": 20512768 + }, + { + "epoch": 0.01, + "learning_rate": 0.0004984126984126985, + "loss": 4.1775, + "theoretical_loss": 6.116706093819951, + "tokens_seen": 20578304 + }, + { + "epoch": 0.01, + "learning_rate": 0.0005, + "loss": 4.3969, + "theoretical_loss": 6.113431824069829, + "tokens_seen": 20643840 + }, + { + "epoch": 0.01, + "learning_rate": 0.0005015873015873016, + "loss": 4.3395, + "theoretical_loss": 6.110170832329612, + "tokens_seen": 20709376 + }, + { + "epoch": 0.01, + "learning_rate": 0.0005031746031746032, + "loss": 4.3461, + "theoretical_loss": 6.10692302298939, + "tokens_seen": 20774912 + }, + { + "epoch": 0.01, + "learning_rate": 0.0005047619047619048, + "loss": 4.3346, + "theoretical_loss": 6.103688301426538, + "tokens_seen": 20840448 + }, + { + "epoch": 0.01, + "learning_rate": 0.0005063492063492064, + "loss": 4.4503, + "theoretical_loss": 6.100466573992453, + "tokens_seen": 20905984 + }, + { + "epoch": 0.01, + "learning_rate": 0.0005079365079365079, + "loss": 4.4352, + "theoretical_loss": 6.097257747999514, + "tokens_seen": 20971520 + }, + { + "epoch": 0.01, + "learning_rate": 0.0005095238095238095, + "loss": 4.3703, + "theoretical_loss": 6.094061731708257, + "tokens_seen": 21037056 + }, + { + "epoch": 0.01, + "learning_rate": 0.0005111111111111111, + "loss": 4.2974, + "theoretical_loss": 6.090878434314753, + "tokens_seen": 21102592 + }, + { + "epoch": 0.01, + "learning_rate": 0.0005126984126984128, + "loss": 4.2911, + "theoretical_loss": 6.0877077659382, + "tokens_seen": 21168128 + }, + { + "epoch": 0.01, + "learning_rate": 0.0005142857142857143, + "loss": 4.3841, + "theoretical_loss": 6.084549637608715, + "tokens_seen": 21233664 + }, + { + "epoch": 0.01, + "learning_rate": 0.0005158730158730159, + "loss": 4.3726, + "theoretical_loss": 6.081403961255324, + "tokens_seen": 21299200 + }, + { + "epoch": 0.01, + "learning_rate": 0.0005174603174603176, + "loss": 4.3931, + "theoretical_loss": 6.078270649694149, + "tokens_seen": 21364736 + }, + { + "epoch": 0.01, + "learning_rate": 0.0005190476190476191, + "loss": 4.3284, + "theoretical_loss": 6.07514961661678, + "tokens_seen": 21430272 + }, + { + "epoch": 0.01, + "learning_rate": 0.0005206349206349207, + "loss": 4.217, + "theoretical_loss": 6.072040776578845, + "tokens_seen": 21495808 + }, + { + "epoch": 0.01, + "learning_rate": 0.0005222222222222222, + "loss": 4.4046, + "theoretical_loss": 6.068944044988752, + "tokens_seen": 21561344 + }, + { + "epoch": 0.01, + "learning_rate": 0.0005238095238095238, + "loss": 4.3049, + "theoretical_loss": 6.065859338096619, + "tokens_seen": 21626880 + }, + { + "epoch": 0.01, + "learning_rate": 0.0005253968253968255, + "loss": 4.2159, + "theoretical_loss": 6.0627865729833745, + "tokens_seen": 21692416 + }, + { + "epoch": 0.01, + "learning_rate": 0.000526984126984127, + "loss": 4.4085, + "theoretical_loss": 6.059725667550044, + "tokens_seen": 21757952 + }, + { + "epoch": 0.01, + "learning_rate": 0.0005285714285714286, + "loss": 4.2093, + "theoretical_loss": 6.0566765405071905, + "tokens_seen": 21823488 + }, + { + "epoch": 0.01, + "learning_rate": 0.0005301587301587302, + "loss": 4.2315, + "theoretical_loss": 6.0536391113645305, + "tokens_seen": 21889024 + }, + { + "epoch": 0.01, + "learning_rate": 0.0005317460317460317, + "loss": 4.3177, + "theoretical_loss": 6.050613300420713, + "tokens_seen": 21954560 + }, + { + "epoch": 0.01, + "learning_rate": 0.0005333333333333334, + "loss": 4.1934, + "theoretical_loss": 6.047599028753265, + "tokens_seen": 22020096 + }, + { + "epoch": 0.01, + "learning_rate": 0.0005349206349206349, + "loss": 4.3796, + "theoretical_loss": 6.04459621820868, + "tokens_seen": 22085632 + }, + { + "epoch": 0.01, + "learning_rate": 0.0005365079365079365, + "loss": 4.2781, + "theoretical_loss": 6.041604791392675, + "tokens_seen": 22151168 + }, + { + "epoch": 0.01, + "learning_rate": 0.0005380952380952381, + "loss": 4.2331, + "theoretical_loss": 6.038624671660598, + "tokens_seen": 22216704 + }, + { + "epoch": 0.01, + "learning_rate": 0.0005396825396825397, + "loss": 4.1207, + "theoretical_loss": 6.035655783107973, + "tokens_seen": 22282240 + }, + { + "epoch": 0.01, + "learning_rate": 0.0005412698412698414, + "loss": 4.1665, + "theoretical_loss": 6.0326980505612084, + "tokens_seen": 22347776 + }, + { + "epoch": 0.01, + "learning_rate": 0.0005428571428571429, + "loss": 4.1452, + "theoretical_loss": 6.029751399568434, + "tokens_seen": 22413312 + }, + { + "epoch": 0.01, + "learning_rate": 0.0005444444444444445, + "loss": 4.1987, + "theoretical_loss": 6.026815756390491, + "tokens_seen": 22478848 + }, + { + "epoch": 0.01, + "learning_rate": 0.000546031746031746, + "loss": 4.2997, + "theoretical_loss": 6.023891047992053, + "tokens_seen": 22544384 + }, + { + "epoch": 0.01, + "learning_rate": 0.0005476190476190477, + "loss": 4.2866, + "theoretical_loss": 6.020977202032884, + "tokens_seen": 22609920 + }, + { + "epoch": 0.01, + "learning_rate": 0.0005492063492063493, + "loss": 4.146, + "theoretical_loss": 6.018074146859234, + "tokens_seen": 22675456 + }, + { + "epoch": 0.01, + "learning_rate": 0.0005507936507936508, + "loss": 4.2995, + "theoretical_loss": 6.015181811495358, + "tokens_seen": 22740992 + }, + { + "epoch": 0.01, + "learning_rate": 0.0005523809523809524, + "loss": 4.1463, + "theoretical_loss": 6.012300125635177, + "tokens_seen": 22806528 + }, + { + "epoch": 0.01, + "learning_rate": 0.0005539682539682539, + "loss": 4.1924, + "theoretical_loss": 6.009429019634049, + "tokens_seen": 22872064 + }, + { + "epoch": 0.01, + "objective/train/docs_used": 15907, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 4.320045471191406, + "objective/train/theoretical_loss": 6.006568424500674, + "objective/train/tokens_used": 43397600, + "theoretical_loss": 6.006568424500674, + "tokens_seen": 22937600 + }, + { + "epoch": 0.01, + "learning_rate": 0.0005555555555555556, + "loss": 4.2643, + "theoretical_loss": 6.006568424500674, + "tokens_seen": 22937600 + }, + { + "epoch": 0.01, + "learning_rate": 0.0005571428571428572, + "loss": 3.9606, + "theoretical_loss": 6.003718271889123, + "tokens_seen": 23003136 + }, + { + "epoch": 0.01, + "learning_rate": 0.0005587301587301587, + "loss": 4.1423, + "theoretical_loss": 6.000878494090979, + "tokens_seen": 23068672 + }, + { + "epoch": 0.01, + "learning_rate": 0.0005603174603174603, + "loss": 4.2877, + "theoretical_loss": 5.998049024027599, + "tokens_seen": 23134208 + }, + { + "epoch": 0.01, + "learning_rate": 0.0005619047619047619, + "loss": 4.2917, + "theoretical_loss": 5.995229795242498, + "tokens_seen": 23199744 + }, + { + "epoch": 0.01, + "learning_rate": 0.0005634920634920636, + "loss": 4.1756, + "theoretical_loss": 5.99242074189383, + "tokens_seen": 23265280 + }, + { + "epoch": 0.01, + "learning_rate": 0.0005650793650793652, + "loss": 4.0492, + "theoretical_loss": 5.989621798747002, + "tokens_seen": 23330816 + }, + { + "epoch": 0.01, + "learning_rate": 0.0005666666666666667, + "loss": 4.0727, + "theoretical_loss": 5.98683290116738, + "tokens_seen": 23396352 + }, + { + "epoch": 0.01, + "learning_rate": 0.0005682539682539683, + "loss": 4.0493, + "theoretical_loss": 5.9840539851131105, + "tokens_seen": 23461888 + }, + { + "epoch": 0.01, + "learning_rate": 0.0005698412698412698, + "loss": 4.1448, + "theoretical_loss": 5.981284987128044, + "tokens_seen": 23527424 + }, + { + "epoch": 0.01, + "learning_rate": 0.0005714285714285715, + "loss": 4.0902, + "theoretical_loss": 5.97852584433477, + "tokens_seen": 23592960 + }, + { + "epoch": 0.01, + "learning_rate": 0.0005730158730158731, + "loss": 4.1207, + "theoretical_loss": 5.975776494427746, + "tokens_seen": 23658496 + }, + { + "epoch": 0.01, + "learning_rate": 0.0005746031746031746, + "loss": 4.1424, + "theoretical_loss": 5.973036875666525, + "tokens_seen": 23724032 + }, + { + "epoch": 0.01, + "learning_rate": 0.0005761904761904762, + "loss": 4.1671, + "theoretical_loss": 5.970306926869093, + "tokens_seen": 23789568 + }, + { + "epoch": 0.01, + "learning_rate": 0.0005777777777777778, + "loss": 4.2791, + "theoretical_loss": 5.967586587405295, + "tokens_seen": 23855104 + }, + { + "epoch": 0.01, + "learning_rate": 0.0005793650793650794, + "loss": 4.0813, + "theoretical_loss": 5.96487579719035, + "tokens_seen": 23920640 + }, + { + "epoch": 0.01, + "learning_rate": 0.000580952380952381, + "loss": 4.0538, + "theoretical_loss": 5.962174496678481, + "tokens_seen": 23986176 + }, + { + "epoch": 0.01, + "learning_rate": 0.0005825396825396825, + "loss": 3.9648, + "theoretical_loss": 5.959482626856606, + "tokens_seen": 24051712 + }, + { + "epoch": 0.01, + "learning_rate": 0.0005841269841269841, + "loss": 4.2193, + "theoretical_loss": 5.956800129238152, + "tokens_seen": 24117248 + }, + { + "epoch": 0.01, + "learning_rate": 0.0005857142857142857, + "loss": 4.043, + "theoretical_loss": 5.954126945856933, + "tokens_seen": 24182784 + }, + { + "epoch": 0.01, + "learning_rate": 0.0005873015873015873, + "loss": 4.0895, + "theoretical_loss": 5.951463019261125, + "tokens_seen": 24248320 + }, + { + "epoch": 0.01, + "learning_rate": 0.0005888888888888889, + "loss": 3.9968, + "theoretical_loss": 5.948808292507333, + "tokens_seen": 24313856 + }, + { + "epoch": 0.01, + "learning_rate": 0.0005904761904761905, + "loss": 4.0092, + "theoretical_loss": 5.946162709154724, + "tokens_seen": 24379392 + }, + { + "epoch": 0.01, + "learning_rate": 0.0005920634920634922, + "loss": 4.0991, + "theoretical_loss": 5.943526213259268, + "tokens_seen": 24444928 + }, + { + "epoch": 0.01, + "learning_rate": 0.0005936507936507937, + "loss": 4.063, + "theoretical_loss": 5.940898749368037, + "tokens_seen": 24510464 + }, + { + "epoch": 0.01, + "learning_rate": 0.0005952380952380953, + "loss": 4.2183, + "theoretical_loss": 5.9382802625136, + "tokens_seen": 24576000 + }, + { + "epoch": 0.01, + "learning_rate": 0.0005968253968253969, + "loss": 4.0664, + "theoretical_loss": 5.935670698208488, + "tokens_seen": 24641536 + }, + { + "epoch": 0.01, + "learning_rate": 0.0005984126984126984, + "loss": 4.0469, + "theoretical_loss": 5.933070002439743, + "tokens_seen": 24707072 + }, + { + "epoch": 0.01, + "learning_rate": 0.0006000000000000001, + "loss": 4.0508, + "theoretical_loss": 5.930478121663538, + "tokens_seen": 24772608 + }, + { + "epoch": 0.01, + "learning_rate": 0.0006015873015873016, + "loss": 4.0802, + "theoretical_loss": 5.927895002799874, + "tokens_seen": 24838144 + }, + { + "epoch": 0.01, + "learning_rate": 0.0006031746031746032, + "loss": 4.1526, + "theoretical_loss": 5.925320593227351, + "tokens_seen": 24903680 + }, + { + "epoch": 0.01, + "learning_rate": 0.0006047619047619048, + "loss": 4.0217, + "theoretical_loss": 5.922754840778015, + "tokens_seen": 24969216 + }, + { + "epoch": 0.01, + "learning_rate": 0.0006063492063492063, + "loss": 3.996, + "theoretical_loss": 5.9201976937322645, + "tokens_seen": 25034752 + }, + { + "epoch": 0.01, + "learning_rate": 0.000607936507936508, + "loss": 4.1476, + "theoretical_loss": 5.917649100813846, + "tokens_seen": 25100288 + }, + { + "epoch": 0.01, + "learning_rate": 0.0006095238095238095, + "loss": 4.0247, + "theoretical_loss": 5.915109011184899, + "tokens_seen": 25165824 + }, + { + "epoch": 0.01, + "learning_rate": 0.0006111111111111111, + "loss": 4.07, + "theoretical_loss": 5.9125773744410814, + "tokens_seen": 25231360 + }, + { + "epoch": 0.01, + "learning_rate": 0.0006126984126984127, + "loss": 4.0942, + "theoretical_loss": 5.91005414060676, + "tokens_seen": 25296896 + }, + { + "epoch": 0.01, + "learning_rate": 0.0006142857142857143, + "loss": 4.1634, + "theoretical_loss": 5.90753926013026, + "tokens_seen": 25362432 + }, + { + "epoch": 0.01, + "learning_rate": 0.000615873015873016, + "loss": 4.098, + "theoretical_loss": 5.90503268387919, + "tokens_seen": 25427968 + }, + { + "epoch": 0.01, + "learning_rate": 0.0006174603174603175, + "loss": 4.0062, + "theoretical_loss": 5.9025343631358185, + "tokens_seen": 25493504 + }, + { + "epoch": 0.01, + "learning_rate": 0.0006190476190476191, + "loss": 4.2125, + "theoretical_loss": 5.900044249592523, + "tokens_seen": 25559040 + }, + { + "epoch": 0.01, + "learning_rate": 0.0006206349206349206, + "loss": 4.1743, + "theoretical_loss": 5.897562295347297, + "tokens_seen": 25624576 + }, + { + "epoch": 0.01, + "learning_rate": 0.0006222222222222223, + "loss": 4.0068, + "theoretical_loss": 5.895088452899312, + "tokens_seen": 25690112 + }, + { + "epoch": 0.01, + "learning_rate": 0.0006238095238095239, + "loss": 4.138, + "theoretical_loss": 5.892622675144553, + "tokens_seen": 25755648 + }, + { + "epoch": 0.01, + "learning_rate": 0.0006253968253968254, + "loss": 4.0475, + "theoretical_loss": 5.890164915371495, + "tokens_seen": 25821184 + }, + { + "epoch": 0.01, + "learning_rate": 0.000626984126984127, + "loss": 4.083, + "theoretical_loss": 5.887715127256854, + "tokens_seen": 25886720 + }, + { + "epoch": 0.01, + "learning_rate": 0.0006285714285714286, + "loss": 4.1093, + "theoretical_loss": 5.885273264861381, + "tokens_seen": 25952256 + }, + { + "epoch": 0.01, + "learning_rate": 0.0006301587301587302, + "loss": 4.1541, + "theoretical_loss": 5.882839282625721, + "tokens_seen": 26017792 + }, + { + "epoch": 0.01, + "learning_rate": 0.0006317460317460318, + "loss": 3.9766, + "theoretical_loss": 5.880413135366323, + "tokens_seen": 26083328 + }, + { + "epoch": 0.01, + "learning_rate": 0.0006333333333333333, + "loss": 3.8962, + "theoretical_loss": 5.877994778271411, + "tokens_seen": 26148864 + }, + { + "epoch": 0.01, + "objective/train/docs_used": 17033, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 4.103891372680664, + "objective/train/theoretical_loss": 5.8755841668969895, + "objective/train/tokens_used": 46674400, + "theoretical_loss": 5.8755841668969895, + "tokens_seen": 26214400 + }, + { + "epoch": 0.01, + "learning_rate": 0.0006349206349206349, + "loss": 4.1687, + "theoretical_loss": 5.8755841668969895, + "tokens_seen": 26214400 + }, + { + "epoch": 0.01, + "learning_rate": 0.0006365079365079365, + "loss": 3.8758, + "theoretical_loss": 5.873181257162926, + "tokens_seen": 26279936 + }, + { + "epoch": 0.01, + "learning_rate": 0.0006380952380952382, + "loss": 3.9489, + "theoretical_loss": 5.870786005349071, + "tokens_seen": 26345472 + }, + { + "epoch": 0.01, + "learning_rate": 0.0006396825396825398, + "loss": 3.98, + "theoretical_loss": 5.868398368091427, + "tokens_seen": 26411008 + }, + { + "epoch": 0.01, + "learning_rate": 0.0006412698412698413, + "loss": 4.0264, + "theoretical_loss": 5.866018302378377, + "tokens_seen": 26476544 + }, + { + "epoch": 0.01, + "learning_rate": 0.0006428571428571429, + "loss": 3.8847, + "theoretical_loss": 5.863645765546956, + "tokens_seen": 26542080 + }, + { + "epoch": 0.01, + "learning_rate": 0.0006444444444444444, + "loss": 3.9805, + "theoretical_loss": 5.861280715279168, + "tokens_seen": 26607616 + }, + { + "epoch": 0.01, + "learning_rate": 0.0006460317460317461, + "loss": 3.894, + "theoretical_loss": 5.858923109598366, + "tokens_seen": 26673152 + }, + { + "epoch": 0.01, + "learning_rate": 0.0006476190476190477, + "loss": 3.8437, + "theoretical_loss": 5.8565729068656545, + "tokens_seen": 26738688 + }, + { + "epoch": 0.01, + "learning_rate": 0.0006492063492063492, + "loss": 3.9946, + "theoretical_loss": 5.85423006577636, + "tokens_seen": 26804224 + }, + { + "epoch": 0.01, + "learning_rate": 0.0006507936507936508, + "loss": 4.0519, + "theoretical_loss": 5.851894545356539, + "tokens_seen": 26869760 + }, + { + "epoch": 0.01, + "learning_rate": 0.0006523809523809524, + "loss": 3.896, + "theoretical_loss": 5.849566304959527, + "tokens_seen": 26935296 + }, + { + "epoch": 0.01, + "learning_rate": 0.000653968253968254, + "loss": 3.7162, + "theoretical_loss": 5.84724530426254, + "tokens_seen": 27000832 + }, + { + "epoch": 0.01, + "learning_rate": 0.0006555555555555556, + "loss": 3.8662, + "theoretical_loss": 5.844931503263309, + "tokens_seen": 27066368 + }, + { + "epoch": 0.01, + "learning_rate": 0.0006571428571428571, + "loss": 3.966, + "theoretical_loss": 5.842624862276773, + "tokens_seen": 27131904 + }, + { + "epoch": 0.01, + "learning_rate": 0.0006587301587301587, + "loss": 3.8366, + "theoretical_loss": 5.840325341931795, + "tokens_seen": 27197440 + }, + { + "epoch": 0.01, + "learning_rate": 0.0006603174603174604, + "loss": 3.6861, + "theoretical_loss": 5.838032903167934, + "tokens_seen": 27262976 + }, + { + "epoch": 0.01, + "learning_rate": 0.0006619047619047619, + "loss": 4.0173, + "theoretical_loss": 5.835747507232257, + "tokens_seen": 27328512 + }, + { + "epoch": 0.01, + "learning_rate": 0.0006634920634920636, + "loss": 3.9618, + "theoretical_loss": 5.833469115676177, + "tokens_seen": 27394048 + }, + { + "epoch": 0.01, + "learning_rate": 0.0006650793650793651, + "loss": 3.962, + "theoretical_loss": 5.8311976903523535, + "tokens_seen": 27459584 + }, + { + "epoch": 0.01, + "learning_rate": 0.0006666666666666668, + "loss": 3.9711, + "theoretical_loss": 5.828933193411611, + "tokens_seen": 27525120 + }, + { + "epoch": 0.01, + "learning_rate": 0.0006682539682539683, + "loss": 3.8676, + "theoretical_loss": 5.826675587299913, + "tokens_seen": 27590656 + }, + { + "epoch": 0.01, + "learning_rate": 0.0006698412698412699, + "loss": 3.9769, + "theoretical_loss": 5.824424834755357, + "tokens_seen": 27656192 + }, + { + "epoch": 0.01, + "learning_rate": 0.0006714285714285715, + "loss": 3.852, + "theoretical_loss": 5.8221808988052235, + "tokens_seen": 27721728 + }, + { + "epoch": 0.01, + "learning_rate": 0.000673015873015873, + "loss": 3.8931, + "theoretical_loss": 5.819943742763051, + "tokens_seen": 27787264 + }, + { + "epoch": 0.01, + "learning_rate": 0.0006746031746031747, + "loss": 3.8953, + "theoretical_loss": 5.817713330225754, + "tokens_seen": 27852800 + }, + { + "epoch": 0.01, + "learning_rate": 0.0006761904761904762, + "loss": 3.8834, + "theoretical_loss": 5.815489625070766, + "tokens_seen": 27918336 + }, + { + "epoch": 0.01, + "learning_rate": 0.0006777777777777778, + "loss": 3.8205, + "theoretical_loss": 5.813272591453236, + "tokens_seen": 27983872 + }, + { + "epoch": 0.01, + "learning_rate": 0.0006793650793650794, + "loss": 3.7655, + "theoretical_loss": 5.811062193803242, + "tokens_seen": 28049408 + }, + { + "epoch": 0.01, + "learning_rate": 0.0006809523809523809, + "loss": 3.8621, + "theoretical_loss": 5.808858396823043, + "tokens_seen": 28114944 + }, + { + "epoch": 0.01, + "learning_rate": 0.0006825396825396826, + "loss": 3.7632, + "theoretical_loss": 5.806661165484377, + "tokens_seen": 28180480 + }, + { + "epoch": 0.01, + "learning_rate": 0.0006841269841269841, + "loss": 3.9688, + "theoretical_loss": 5.804470465025777, + "tokens_seen": 28246016 + }, + { + "epoch": 0.01, + "learning_rate": 0.0006857142857142857, + "loss": 3.912, + "theoretical_loss": 5.802286260949924, + "tokens_seen": 28311552 + }, + { + "epoch": 0.01, + "learning_rate": 0.0006873015873015873, + "loss": 3.8936, + "theoretical_loss": 5.8001085190210375, + "tokens_seen": 28377088 + }, + { + "epoch": 0.01, + "learning_rate": 0.000688888888888889, + "loss": 3.9773, + "theoretical_loss": 5.797937205262292, + "tokens_seen": 28442624 + }, + { + "epoch": 0.01, + "learning_rate": 0.0006904761904761906, + "loss": 3.7441, + "theoretical_loss": 5.795772285953266, + "tokens_seen": 28508160 + }, + { + "epoch": 0.01, + "learning_rate": 0.0006920634920634921, + "loss": 3.8072, + "theoretical_loss": 5.793613727627429, + "tokens_seen": 28573696 + }, + { + "epoch": 0.01, + "learning_rate": 0.0006936507936507937, + "loss": 3.6394, + "theoretical_loss": 5.791461497069644, + "tokens_seen": 28639232 + }, + { + "epoch": 0.01, + "learning_rate": 0.0006952380952380953, + "loss": 3.8893, + "theoretical_loss": 5.789315561313717, + "tokens_seen": 28704768 + }, + { + "epoch": 0.01, + "learning_rate": 0.0006968253968253969, + "loss": 3.8561, + "theoretical_loss": 5.787175887639959, + "tokens_seen": 28770304 + }, + { + "epoch": 0.01, + "learning_rate": 0.0006984126984126985, + "loss": 3.89, + "theoretical_loss": 5.785042443572798, + "tokens_seen": 28835840 + }, + { + "epoch": 0.01, + "learning_rate": 0.0007, + "loss": 3.7668, + "theoretical_loss": 5.782915196878396, + "tokens_seen": 28901376 + }, + { + "epoch": 0.01, + "learning_rate": 0.0007015873015873016, + "loss": 3.8187, + "theoretical_loss": 5.780794115562314, + "tokens_seen": 28966912 + }, + { + "epoch": 0.01, + "learning_rate": 0.0007031746031746032, + "loss": 3.9142, + "theoretical_loss": 5.778679167867196, + "tokens_seen": 29032448 + }, + { + "epoch": 0.01, + "learning_rate": 0.0007047619047619048, + "loss": 3.795, + "theoretical_loss": 5.7765703222704765, + "tokens_seen": 29097984 + }, + { + "epoch": 0.01, + "learning_rate": 0.0007063492063492064, + "loss": 3.856, + "theoretical_loss": 5.7744675474821285, + "tokens_seen": 29163520 + }, + { + "epoch": 0.01, + "learning_rate": 0.0007079365079365079, + "loss": 3.913, + "theoretical_loss": 5.772370812442426, + "tokens_seen": 29229056 + }, + { + "epoch": 0.01, + "learning_rate": 0.0007095238095238095, + "loss": 3.7888, + "theoretical_loss": 5.770280086319733, + "tokens_seen": 29294592 + }, + { + "epoch": 0.01, + "learning_rate": 0.0007111111111111111, + "loss": 3.8747, + "theoretical_loss": 5.768195338508333, + "tokens_seen": 29360128 + }, + { + "epoch": 0.01, + "learning_rate": 0.0007126984126984128, + "loss": 3.7361, + "theoretical_loss": 5.766116538626263, + "tokens_seen": 29425664 + }, + { + "epoch": 0.01, + "objective/train/docs_used": 18174, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 3.9875965118408203, + "objective/train/theoretical_loss": 5.7640436565131905, + "objective/train/tokens_used": 49951200, + "theoretical_loss": 5.7640436565131905, + "tokens_seen": 29491200 + }, + { + "epoch": 0.01, + "learning_rate": 0.0007142857142857144, + "loss": 3.8826, + "theoretical_loss": 5.7640436565131905, + "tokens_seen": 29491200 + }, + { + "epoch": 0.01, + "learning_rate": 0.0007158730158730159, + "loss": 3.8191, + "theoretical_loss": 5.761976662228307, + "tokens_seen": 29556736 + }, + { + "epoch": 0.01, + "learning_rate": 0.0007174603174603175, + "loss": 3.7279, + "theoretical_loss": 5.759915526048243, + "tokens_seen": 29622272 + }, + { + "epoch": 0.01, + "learning_rate": 0.000719047619047619, + "loss": 3.8132, + "theoretical_loss": 5.757860218465023, + "tokens_seen": 29687808 + }, + { + "epoch": 0.01, + "learning_rate": 0.0007206349206349207, + "loss": 3.8766, + "theoretical_loss": 5.755810710184019, + "tokens_seen": 29753344 + }, + { + "epoch": 0.01, + "learning_rate": 0.0007222222222222223, + "loss": 3.7341, + "theoretical_loss": 5.753766972121954, + "tokens_seen": 29818880 + }, + { + "epoch": 0.01, + "learning_rate": 0.0007238095238095238, + "loss": 3.8926, + "theoretical_loss": 5.751728975404911, + "tokens_seen": 29884416 + }, + { + "epoch": 0.01, + "learning_rate": 0.0007253968253968254, + "loss": 3.9156, + "theoretical_loss": 5.7496966913663705, + "tokens_seen": 29949952 + }, + { + "epoch": 0.01, + "learning_rate": 0.0007269841269841271, + "loss": 3.7597, + "theoretical_loss": 5.747670091545274, + "tokens_seen": 30015488 + }, + { + "epoch": 0.01, + "learning_rate": 0.0007285714285714286, + "loss": 3.8255, + "theoretical_loss": 5.745649147684103, + "tokens_seen": 30081024 + }, + { + "epoch": 0.01, + "learning_rate": 0.0007301587301587302, + "loss": 3.9182, + "theoretical_loss": 5.743633831726987, + "tokens_seen": 30146560 + }, + { + "epoch": 0.01, + "learning_rate": 0.0007317460317460317, + "loss": 3.8177, + "theoretical_loss": 5.741624115817827, + "tokens_seen": 30212096 + }, + { + "epoch": 0.01, + "learning_rate": 0.0007333333333333333, + "loss": 3.8092, + "theoretical_loss": 5.739619972298446, + "tokens_seen": 30277632 + }, + { + "epoch": 0.01, + "learning_rate": 0.000734920634920635, + "loss": 3.8899, + "theoretical_loss": 5.737621373706753, + "tokens_seen": 30343168 + }, + { + "epoch": 0.01, + "learning_rate": 0.0007365079365079365, + "loss": 3.7504, + "theoretical_loss": 5.735628292774945, + "tokens_seen": 30408704 + }, + { + "epoch": 0.01, + "learning_rate": 0.0007380952380952382, + "loss": 3.7239, + "theoretical_loss": 5.7336407024276985, + "tokens_seen": 30474240 + }, + { + "epoch": 0.01, + "learning_rate": 0.0007396825396825397, + "loss": 3.7393, + "theoretical_loss": 5.731658575780418, + "tokens_seen": 30539776 + }, + { + "epoch": 0.01, + "learning_rate": 0.0007412698412698414, + "loss": 3.7214, + "theoretical_loss": 5.729681886137476, + "tokens_seen": 30605312 + }, + { + "epoch": 0.01, + "learning_rate": 0.0007428571428571429, + "loss": 3.638, + "theoretical_loss": 5.727710606990488, + "tokens_seen": 30670848 + }, + { + "epoch": 0.01, + "learning_rate": 0.0007444444444444445, + "loss": 3.8293, + "theoretical_loss": 5.725744712016601, + "tokens_seen": 30736384 + }, + { + "epoch": 0.01, + "learning_rate": 0.0007460317460317461, + "loss": 3.7691, + "theoretical_loss": 5.723784175076802, + "tokens_seen": 30801920 + }, + { + "epoch": 0.01, + "learning_rate": 0.0007476190476190476, + "loss": 3.7248, + "theoretical_loss": 5.72182897021425, + "tokens_seen": 30867456 + }, + { + "epoch": 0.01, + "learning_rate": 0.0007492063492063493, + "loss": 3.6, + "theoretical_loss": 5.719879071652615, + "tokens_seen": 30932992 + }, + { + "epoch": 0.01, + "learning_rate": 0.0007507936507936508, + "loss": 3.7567, + "theoretical_loss": 5.717934453794454, + "tokens_seen": 30998528 + }, + { + "epoch": 0.01, + "learning_rate": 0.0007523809523809524, + "loss": 3.7718, + "theoretical_loss": 5.715995091219584, + "tokens_seen": 31064064 + }, + { + "epoch": 0.01, + "learning_rate": 0.000753968253968254, + "loss": 3.6864, + "theoretical_loss": 5.7140609586834925, + "tokens_seen": 31129600 + }, + { + "epoch": 0.01, + "learning_rate": 0.0007555555555555555, + "loss": 3.6952, + "theoretical_loss": 5.712132031115749, + "tokens_seen": 31195136 + }, + { + "epoch": 0.01, + "learning_rate": 0.0007571428571428572, + "loss": 3.6507, + "theoretical_loss": 5.710208283618447, + "tokens_seen": 31260672 + }, + { + "epoch": 0.01, + "learning_rate": 0.0007587301587301588, + "loss": 3.8584, + "theoretical_loss": 5.708289691464657, + "tokens_seen": 31326208 + }, + { + "epoch": 0.01, + "learning_rate": 0.0007603174603174603, + "loss": 3.8918, + "theoretical_loss": 5.706376230096897, + "tokens_seen": 31391744 + }, + { + "epoch": 0.01, + "learning_rate": 0.0007619047619047619, + "loss": 3.6634, + "theoretical_loss": 5.704467875125619, + "tokens_seen": 31457280 + }, + { + "epoch": 0.01, + "learning_rate": 0.0007634920634920635, + "loss": 3.7427, + "theoretical_loss": 5.702564602327719, + "tokens_seen": 31522816 + }, + { + "epoch": 0.01, + "learning_rate": 0.0007650793650793652, + "loss": 3.6043, + "theoretical_loss": 5.700666387645049, + "tokens_seen": 31588352 + }, + { + "epoch": 0.01, + "learning_rate": 0.0007666666666666667, + "loss": 3.884, + "theoretical_loss": 5.698773207182963, + "tokens_seen": 31653888 + }, + { + "epoch": 0.01, + "learning_rate": 0.0007682539682539683, + "loss": 3.7503, + "theoretical_loss": 5.696885037208865, + "tokens_seen": 31719424 + }, + { + "epoch": 0.01, + "learning_rate": 0.0007698412698412699, + "loss": 3.699, + "theoretical_loss": 5.695001854150782, + "tokens_seen": 31784960 + }, + { + "epoch": 0.01, + "learning_rate": 0.0007714285714285715, + "loss": 3.6369, + "theoretical_loss": 5.693123634595938, + "tokens_seen": 31850496 + }, + { + "epoch": 0.01, + "learning_rate": 0.0007730158730158731, + "loss": 3.7626, + "theoretical_loss": 5.69125035528937, + "tokens_seen": 31916032 + }, + { + "epoch": 0.01, + "learning_rate": 0.0007746031746031746, + "loss": 3.6916, + "theoretical_loss": 5.689381993132532, + "tokens_seen": 31981568 + }, + { + "epoch": 0.01, + "learning_rate": 0.0007761904761904762, + "loss": 3.7321, + "theoretical_loss": 5.687518525181924, + "tokens_seen": 32047104 + }, + { + "epoch": 0.01, + "learning_rate": 0.0007777777777777778, + "loss": 3.6983, + "theoretical_loss": 5.685659928647743, + "tokens_seen": 32112640 + }, + { + "epoch": 0.01, + "learning_rate": 0.0007793650793650794, + "loss": 3.8384, + "theoretical_loss": 5.68380618089254, + "tokens_seen": 32178176 + }, + { + "epoch": 0.01, + "learning_rate": 0.000780952380952381, + "loss": 3.6331, + "theoretical_loss": 5.6819572594298915, + "tokens_seen": 32243712 + }, + { + "epoch": 0.01, + "learning_rate": 0.0007825396825396825, + "loss": 3.6316, + "theoretical_loss": 5.680113141923088, + "tokens_seen": 32309248 + }, + { + "epoch": 0.01, + "learning_rate": 0.0007841269841269841, + "loss": 3.7945, + "theoretical_loss": 5.67827380618384, + "tokens_seen": 32374784 + }, + { + "epoch": 0.01, + "learning_rate": 0.0007857142857142857, + "loss": 3.7029, + "theoretical_loss": 5.6764392301709865, + "tokens_seen": 32440320 + }, + { + "epoch": 0.01, + "learning_rate": 0.0007873015873015873, + "loss": 3.5931, + "theoretical_loss": 5.674609391989232, + "tokens_seen": 32505856 + }, + { + "epoch": 0.01, + "learning_rate": 0.000788888888888889, + "loss": 3.8317, + "theoretical_loss": 5.672784269887885, + "tokens_seen": 32571392 + }, + { + "epoch": 0.01, + "learning_rate": 0.0007904761904761905, + "loss": 3.818, + "theoretical_loss": 5.670963842259615, + "tokens_seen": 32636928 + }, + { + "epoch": 0.01, + "learning_rate": 0.0007920634920634921, + "loss": 3.8296, + "theoretical_loss": 5.669148087639222, + "tokens_seen": 32702464 + }, + { + "epoch": 0.01, + "objective/train/docs_used": 19239, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 3.7655105590820312, + "objective/train/theoretical_loss": 5.66733698470242, + "objective/train/tokens_used": 53228000, + "theoretical_loss": 5.66733698470242, + "tokens_seen": 32768000 + }, + { + "epoch": 0.01, + "learning_rate": 0.0007936507936507938, + "loss": 3.7708, + "theoretical_loss": 5.66733698470242, + "tokens_seen": 32768000 + }, + { + "epoch": 0.01, + "learning_rate": 0.0007952380952380953, + "loss": 3.6941, + "theoretical_loss": 5.665530512264631, + "tokens_seen": 32833536 + }, + { + "epoch": 0.01, + "learning_rate": 0.0007968253968253969, + "loss": 3.7219, + "theoretical_loss": 5.663728649279793, + "tokens_seen": 32899072 + }, + { + "epoch": 0.01, + "learning_rate": 0.0007984126984126984, + "loss": 3.5495, + "theoretical_loss": 5.6619313748391855, + "tokens_seen": 32964608 + }, + { + "epoch": 0.01, + "learning_rate": 0.0008, + "loss": 3.7722, + "theoretical_loss": 5.6601386681702515, + "tokens_seen": 33030144 + }, + { + "epoch": 0.01, + "learning_rate": 0.0007999839518555668, + "loss": 3.6901, + "theoretical_loss": 5.658350508635454, + "tokens_seen": 33095680 + }, + { + "epoch": 0.01, + "learning_rate": 0.0007999679037111334, + "loss": 3.5906, + "theoretical_loss": 5.656566875731127, + "tokens_seen": 33161216 + }, + { + "epoch": 0.01, + "learning_rate": 0.0007999518555667002, + "loss": 3.5962, + "theoretical_loss": 5.654787749086345, + "tokens_seen": 33226752 + }, + { + "epoch": 0.01, + "learning_rate": 0.0007999358074222668, + "loss": 3.653, + "theoretical_loss": 5.6530131084618045, + "tokens_seen": 33292288 + }, + { + "epoch": 0.01, + "learning_rate": 0.0007999197592778336, + "loss": 3.6413, + "theoretical_loss": 5.651242933748718, + "tokens_seen": 33357824 + }, + { + "epoch": 0.01, + "learning_rate": 0.0007999037111334002, + "loss": 3.6472, + "theoretical_loss": 5.649477204967717, + "tokens_seen": 33423360 + }, + { + "epoch": 0.01, + "learning_rate": 0.000799887662988967, + "loss": 3.7624, + "theoretical_loss": 5.647715902267764, + "tokens_seen": 33488896 + }, + { + "epoch": 0.01, + "learning_rate": 0.0007998716148445336, + "loss": 3.6623, + "theoretical_loss": 5.645959005925086, + "tokens_seen": 33554432 + }, + { + "epoch": 0.01, + "learning_rate": 0.0007998555667001004, + "loss": 3.6399, + "theoretical_loss": 5.644206496342103, + "tokens_seen": 33619968 + }, + { + "epoch": 0.01, + "learning_rate": 0.000799839518555667, + "loss": 3.6901, + "theoretical_loss": 5.642458354046386, + "tokens_seen": 33685504 + }, + { + "epoch": 0.01, + "learning_rate": 0.0007998234704112338, + "loss": 3.6232, + "theoretical_loss": 5.6407145596896076, + "tokens_seen": 33751040 + }, + { + "epoch": 0.01, + "learning_rate": 0.0007998074222668004, + "loss": 3.7036, + "theoretical_loss": 5.638975094046518, + "tokens_seen": 33816576 + }, + { + "epoch": 0.01, + "learning_rate": 0.0007997913741223672, + "loss": 3.5183, + "theoretical_loss": 5.637239938013924, + "tokens_seen": 33882112 + }, + { + "epoch": 0.01, + "learning_rate": 0.0007997753259779338, + "loss": 3.5563, + "theoretical_loss": 5.6355090726096755, + "tokens_seen": 33947648 + }, + { + "epoch": 0.01, + "learning_rate": 0.0007997592778335006, + "loss": 3.5341, + "theoretical_loss": 5.633782478971672, + "tokens_seen": 34013184 + }, + { + "epoch": 0.01, + "learning_rate": 0.0007997432296890672, + "loss": 3.5986, + "theoretical_loss": 5.632060138356873, + "tokens_seen": 34078720 + }, + { + "epoch": 0.01, + "learning_rate": 0.000799727181544634, + "loss": 3.5259, + "theoretical_loss": 5.630342032140315, + "tokens_seen": 34144256 + }, + { + "epoch": 0.01, + "learning_rate": 0.0007997111334002006, + "loss": 3.7063, + "theoretical_loss": 5.628628141814151, + "tokens_seen": 34209792 + }, + { + "epoch": 0.01, + "learning_rate": 0.0007996950852557674, + "loss": 3.5641, + "theoretical_loss": 5.62691844898668, + "tokens_seen": 34275328 + }, + { + "epoch": 0.01, + "learning_rate": 0.000799679037111334, + "loss": 3.506, + "theoretical_loss": 5.625212935381407, + "tokens_seen": 34340864 + }, + { + "epoch": 0.01, + "learning_rate": 0.0007996629889669008, + "loss": 3.5209, + "theoretical_loss": 5.6235115828361035, + "tokens_seen": 34406400 + }, + { + "epoch": 0.01, + "learning_rate": 0.0007996469408224675, + "loss": 3.4175, + "theoretical_loss": 5.621814373301866, + "tokens_seen": 34471936 + }, + { + "epoch": 0.01, + "learning_rate": 0.0007996308926780342, + "loss": 3.6197, + "theoretical_loss": 5.620121288842209, + "tokens_seen": 34537472 + }, + { + "epoch": 0.01, + "learning_rate": 0.0007996148445336009, + "loss": 3.5816, + "theoretical_loss": 5.618432311632143, + "tokens_seen": 34603008 + }, + { + "epoch": 0.01, + "learning_rate": 0.0007995987963891676, + "loss": 3.5379, + "theoretical_loss": 5.616747423957278, + "tokens_seen": 34668544 + }, + { + "epoch": 0.01, + "learning_rate": 0.0007995827482447343, + "loss": 3.5478, + "theoretical_loss": 5.615066608212926, + "tokens_seen": 34734080 + }, + { + "epoch": 0.01, + "learning_rate": 0.000799566700100301, + "loss": 3.6232, + "theoretical_loss": 5.613389846903216, + "tokens_seen": 34799616 + }, + { + "epoch": 0.01, + "learning_rate": 0.0007995506519558677, + "loss": 3.5677, + "theoretical_loss": 5.611717122640225, + "tokens_seen": 34865152 + }, + { + "epoch": 0.01, + "learning_rate": 0.0007995346038114344, + "loss": 3.4754, + "theoretical_loss": 5.610048418143098, + "tokens_seen": 34930688 + }, + { + "epoch": 0.01, + "learning_rate": 0.0007995185556670011, + "loss": 3.5103, + "theoretical_loss": 5.6083837162372, + "tokens_seen": 34996224 + }, + { + "epoch": 0.01, + "learning_rate": 0.0007995025075225678, + "loss": 3.5373, + "theoretical_loss": 5.606722999853261, + "tokens_seen": 35061760 + }, + { + "epoch": 0.01, + "learning_rate": 0.0007994864593781345, + "loss": 3.4716, + "theoretical_loss": 5.605066252026537, + "tokens_seen": 35127296 + }, + { + "epoch": 0.01, + "learning_rate": 0.0007994704112337012, + "loss": 3.5505, + "theoretical_loss": 5.603413455895973, + "tokens_seen": 35192832 + }, + { + "epoch": 0.01, + "learning_rate": 0.0007994543630892679, + "loss": 3.5079, + "theoretical_loss": 5.601764594703377, + "tokens_seen": 35258368 + }, + { + "epoch": 0.01, + "learning_rate": 0.0007994383149448346, + "loss": 3.2808, + "theoretical_loss": 5.600119651792607, + "tokens_seen": 35323904 + }, + { + "epoch": 0.01, + "learning_rate": 0.0007994222668004013, + "loss": 3.4043, + "theoretical_loss": 5.598478610608758, + "tokens_seen": 35389440 + }, + { + "epoch": 0.01, + "learning_rate": 0.000799406218655968, + "loss": 3.4338, + "theoretical_loss": 5.596841454697365, + "tokens_seen": 35454976 + }, + { + "epoch": 0.01, + "learning_rate": 0.0007993901705115347, + "loss": 3.5769, + "theoretical_loss": 5.595208167703601, + "tokens_seen": 35520512 + }, + { + "epoch": 0.01, + "learning_rate": 0.0007993741223671014, + "loss": 3.4669, + "theoretical_loss": 5.5935787333714995, + "tokens_seen": 35586048 + }, + { + "epoch": 0.01, + "learning_rate": 0.0007993580742226681, + "loss": 3.769, + "theoretical_loss": 5.59195313554317, + "tokens_seen": 35651584 + }, + { + "epoch": 0.01, + "learning_rate": 0.0007993420260782347, + "loss": 3.6233, + "theoretical_loss": 5.590331358158036, + "tokens_seen": 35717120 + }, + { + "epoch": 0.01, + "learning_rate": 0.0007993259779338015, + "loss": 3.6649, + "theoretical_loss": 5.58871338525206, + "tokens_seen": 35782656 + }, + { + "epoch": 0.01, + "learning_rate": 0.0007993099297893681, + "loss": 3.4736, + "theoretical_loss": 5.587099200956995, + "tokens_seen": 35848192 + }, + { + "epoch": 0.01, + "learning_rate": 0.0007992938816449349, + "loss": 3.653, + "theoretical_loss": 5.585488789499637, + "tokens_seen": 35913728 + }, + { + "epoch": 0.01, + "learning_rate": 0.0007992778335005015, + "loss": 3.635, + "theoretical_loss": 5.583882135201078, + "tokens_seen": 35979264 + }, + { + "epoch": 0.01, + "objective/train/docs_used": 20470, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 3.4506337642669678, + "objective/train/theoretical_loss": 5.582279222475973, + "objective/train/tokens_used": 56504800, + "theoretical_loss": 5.582279222475973, + "tokens_seen": 36044800 + }, + { + "epoch": 0.01, + "learning_rate": 0.0007992617853560683, + "loss": 3.4134, + "theoretical_loss": 5.582279222475973, + "tokens_seen": 36044800 + }, + { + "epoch": 0.01, + "learning_rate": 0.0007992457372116349, + "loss": 3.7112, + "theoretical_loss": 5.580680035831821, + "tokens_seen": 36110336 + }, + { + "epoch": 0.01, + "learning_rate": 0.0007992296890672017, + "loss": 3.525, + "theoretical_loss": 5.579084559868232, + "tokens_seen": 36175872 + }, + { + "epoch": 0.01, + "learning_rate": 0.0007992136409227683, + "loss": 3.6194, + "theoretical_loss": 5.577492779276222, + "tokens_seen": 36241408 + }, + { + "epoch": 0.01, + "learning_rate": 0.000799197592778335, + "loss": 3.4336, + "theoretical_loss": 5.575904678837501, + "tokens_seen": 36306944 + }, + { + "epoch": 0.01, + "learning_rate": 0.0007991815446339017, + "loss": 3.6176, + "theoretical_loss": 5.574320243423783, + "tokens_seen": 36372480 + }, + { + "epoch": 0.01, + "learning_rate": 0.0007991654964894684, + "loss": 3.3777, + "theoretical_loss": 5.572739457996079, + "tokens_seen": 36438016 + }, + { + "epoch": 0.01, + "learning_rate": 0.0007991494483450351, + "loss": 3.3934, + "theoretical_loss": 5.5711623076040215, + "tokens_seen": 36503552 + }, + { + "epoch": 0.01, + "learning_rate": 0.0007991334002006018, + "loss": 3.65, + "theoretical_loss": 5.569588777385177, + "tokens_seen": 36569088 + }, + { + "epoch": 0.01, + "learning_rate": 0.0007991173520561685, + "loss": 3.519, + "theoretical_loss": 5.568018852564379, + "tokens_seen": 36634624 + }, + { + "epoch": 0.01, + "learning_rate": 0.0007991013039117352, + "loss": 3.5048, + "theoretical_loss": 5.566452518453051, + "tokens_seen": 36700160 + }, + { + "epoch": 0.01, + "learning_rate": 0.0007990852557673019, + "loss": 3.3481, + "theoretical_loss": 5.564889760448561, + "tokens_seen": 36765696 + }, + { + "epoch": 0.01, + "learning_rate": 0.0007990692076228686, + "loss": 3.5093, + "theoretical_loss": 5.5633305640335475, + "tokens_seen": 36831232 + }, + { + "epoch": 0.01, + "learning_rate": 0.0007990531594784353, + "loss": 3.5026, + "theoretical_loss": 5.561774914775286, + "tokens_seen": 36896768 + }, + { + "epoch": 0.01, + "learning_rate": 0.000799037111334002, + "loss": 3.5222, + "theoretical_loss": 5.5602227983250385, + "tokens_seen": 36962304 + }, + { + "epoch": 0.01, + "learning_rate": 0.0007990210631895687, + "loss": 3.4603, + "theoretical_loss": 5.558674200417419, + "tokens_seen": 37027840 + }, + { + "epoch": 0.01, + "learning_rate": 0.0007990050150451354, + "loss": 3.6149, + "theoretical_loss": 5.557129106869764, + "tokens_seen": 37093376 + }, + { + "epoch": 0.01, + "learning_rate": 0.0007989889669007021, + "loss": 3.4941, + "theoretical_loss": 5.555587503581506, + "tokens_seen": 37158912 + }, + { + "epoch": 0.01, + "learning_rate": 0.0007989729187562688, + "loss": 3.5612, + "theoretical_loss": 5.554049376533554, + "tokens_seen": 37224448 + }, + { + "epoch": 0.01, + "learning_rate": 0.0007989568706118355, + "loss": 3.5255, + "theoretical_loss": 5.552514711787683, + "tokens_seen": 37289984 + }, + { + "epoch": 0.01, + "learning_rate": 0.0007989408224674022, + "loss": 3.4369, + "theoretical_loss": 5.550983495485925, + "tokens_seen": 37355520 + }, + { + "epoch": 0.01, + "learning_rate": 0.0007989247743229689, + "loss": 3.6268, + "theoretical_loss": 5.549455713849963, + "tokens_seen": 37421056 + }, + { + "epoch": 0.01, + "learning_rate": 0.0007989087261785356, + "loss": 3.6116, + "theoretical_loss": 5.547931353180543, + "tokens_seen": 37486592 + }, + { + "epoch": 0.01, + "learning_rate": 0.0007988926780341023, + "loss": 3.4586, + "theoretical_loss": 5.546410399856875, + "tokens_seen": 37552128 + }, + { + "epoch": 0.01, + "learning_rate": 0.000798876629889669, + "loss": 3.4418, + "theoretical_loss": 5.544892840336053, + "tokens_seen": 37617664 + }, + { + "epoch": 0.01, + "learning_rate": 0.0007988605817452357, + "loss": 3.4787, + "theoretical_loss": 5.543378661152472, + "tokens_seen": 37683200 + }, + { + "epoch": 0.01, + "learning_rate": 0.0007988445336008024, + "loss": 3.4459, + "theoretical_loss": 5.5418678489172555, + "tokens_seen": 37748736 + }, + { + "epoch": 0.01, + "learning_rate": 0.0007988284854563691, + "loss": 3.5373, + "theoretical_loss": 5.540360390317682, + "tokens_seen": 37814272 + }, + { + "epoch": 0.01, + "learning_rate": 0.0007988124373119358, + "loss": 3.3507, + "theoretical_loss": 5.538856272116626, + "tokens_seen": 37879808 + }, + { + "epoch": 0.01, + "learning_rate": 0.0007987963891675025, + "loss": 3.3229, + "theoretical_loss": 5.5373554811519945, + "tokens_seen": 37945344 + }, + { + "epoch": 0.01, + "learning_rate": 0.0007987803410230693, + "loss": 3.3885, + "theoretical_loss": 5.535858004336175, + "tokens_seen": 38010880 + }, + { + "epoch": 0.01, + "learning_rate": 0.0007987642928786359, + "loss": 3.4684, + "theoretical_loss": 5.534363828655488, + "tokens_seen": 38076416 + }, + { + "epoch": 0.01, + "learning_rate": 0.0007987482447342027, + "loss": 3.4267, + "theoretical_loss": 5.532872941169639, + "tokens_seen": 38141952 + }, + { + "epoch": 0.01, + "learning_rate": 0.0007987321965897693, + "loss": 3.4052, + "theoretical_loss": 5.531385329011181, + "tokens_seen": 38207488 + }, + { + "epoch": 0.01, + "learning_rate": 0.0007987161484453361, + "loss": 3.4792, + "theoretical_loss": 5.529900979384985, + "tokens_seen": 38273024 + }, + { + "epoch": 0.01, + "learning_rate": 0.0007987001003009027, + "loss": 3.3322, + "theoretical_loss": 5.5284198795677, + "tokens_seen": 38338560 + }, + { + "epoch": 0.01, + "learning_rate": 0.0007986840521564695, + "loss": 3.2621, + "theoretical_loss": 5.526942016907242, + "tokens_seen": 38404096 + }, + { + "epoch": 0.01, + "learning_rate": 0.0007986680040120361, + "loss": 3.4628, + "theoretical_loss": 5.5254673788222615, + "tokens_seen": 38469632 + }, + { + "epoch": 0.01, + "learning_rate": 0.0007986519558676029, + "loss": 3.408, + "theoretical_loss": 5.523995952801634, + "tokens_seen": 38535168 + }, + { + "epoch": 0.01, + "learning_rate": 0.0007986359077231695, + "loss": 3.2654, + "theoretical_loss": 5.522527726403953, + "tokens_seen": 38600704 + }, + { + "epoch": 0.01, + "learning_rate": 0.0007986198595787363, + "loss": 3.5291, + "theoretical_loss": 5.521062687257016, + "tokens_seen": 38666240 + }, + { + "epoch": 0.01, + "learning_rate": 0.0007986038114343029, + "loss": 3.4476, + "theoretical_loss": 5.519600823057332, + "tokens_seen": 38731776 + }, + { + "epoch": 0.01, + "learning_rate": 0.0007985877632898697, + "loss": 3.3845, + "theoretical_loss": 5.518142121569613, + "tokens_seen": 38797312 + }, + { + "epoch": 0.01, + "learning_rate": 0.0007985717151454363, + "loss": 3.3291, + "theoretical_loss": 5.516686570626293, + "tokens_seen": 38862848 + }, + { + "epoch": 0.01, + "learning_rate": 0.0007985556670010031, + "loss": 3.5068, + "theoretical_loss": 5.515234158127032, + "tokens_seen": 38928384 + }, + { + "epoch": 0.01, + "learning_rate": 0.0007985396188565697, + "loss": 3.5176, + "theoretical_loss": 5.5137848720382365, + "tokens_seen": 38993920 + }, + { + "epoch": 0.01, + "learning_rate": 0.0007985235707121365, + "loss": 3.375, + "theoretical_loss": 5.5123387003925775, + "tokens_seen": 39059456 + }, + { + "epoch": 0.01, + "learning_rate": 0.0007985075225677031, + "loss": 3.3799, + "theoretical_loss": 5.510895631288515, + "tokens_seen": 39124992 + }, + { + "epoch": 0.01, + "learning_rate": 0.0007984914744232699, + "loss": 3.3877, + "theoretical_loss": 5.50945565288983, + "tokens_seen": 39190528 + }, + { + "epoch": 0.01, + "learning_rate": 0.0007984754262788365, + "loss": 3.3537, + "theoretical_loss": 5.508018753425152, + "tokens_seen": 39256064 + }, + { + "epoch": 0.01, + "objective/train/docs_used": 21760, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 3.1137375831604004, + "objective/train/theoretical_loss": 5.506584921187503, + "objective/train/tokens_used": 59781600, + "theoretical_loss": 5.506584921187503, + "tokens_seen": 39321600 + }, + { + "epoch": 0.01, + "learning_rate": 0.0007984593781344033, + "loss": 3.2481, + "theoretical_loss": 5.506584921187503, + "tokens_seen": 39321600 + }, + { + "epoch": 0.01, + "learning_rate": 0.0007984433299899699, + "loss": 3.3811, + "theoretical_loss": 5.505154144533835, + "tokens_seen": 39387136 + }, + { + "epoch": 0.01, + "learning_rate": 0.0007984272818455367, + "loss": 3.4658, + "theoretical_loss": 5.503726411884568, + "tokens_seen": 39452672 + }, + { + "epoch": 0.01, + "learning_rate": 0.0007984112337011033, + "loss": 3.4491, + "theoretical_loss": 5.502301711723156, + "tokens_seen": 39518208 + }, + { + "epoch": 0.01, + "learning_rate": 0.0007983951855566701, + "loss": 3.2945, + "theoretical_loss": 5.5008800325956235, + "tokens_seen": 39583744 + }, + { + "epoch": 0.01, + "learning_rate": 0.0007983791374122367, + "loss": 3.4917, + "theoretical_loss": 5.499461363110136, + "tokens_seen": 39649280 + }, + { + "epoch": 0.01, + "learning_rate": 0.0007983630892678035, + "loss": 3.2398, + "theoretical_loss": 5.498045691936548, + "tokens_seen": 39714816 + }, + { + "epoch": 0.01, + "learning_rate": 0.0007983470411233701, + "loss": 3.4307, + "theoretical_loss": 5.496633007805977, + "tokens_seen": 39780352 + }, + { + "epoch": 0.01, + "learning_rate": 0.0007983309929789369, + "loss": 3.3324, + "theoretical_loss": 5.495223299510372, + "tokens_seen": 39845888 + }, + { + "epoch": 0.01, + "learning_rate": 0.0007983149448345036, + "loss": 3.3503, + "theoretical_loss": 5.493816555902079, + "tokens_seen": 39911424 + }, + { + "epoch": 0.01, + "learning_rate": 0.0007982988966900703, + "loss": 3.474, + "theoretical_loss": 5.492412765893421, + "tokens_seen": 39976960 + }, + { + "epoch": 0.01, + "learning_rate": 0.000798282848545637, + "loss": 3.4197, + "theoretical_loss": 5.49101191845628, + "tokens_seen": 40042496 + }, + { + "epoch": 0.01, + "learning_rate": 0.0007982668004012037, + "loss": 3.2609, + "theoretical_loss": 5.489614002621677, + "tokens_seen": 40108032 + }, + { + "epoch": 0.01, + "learning_rate": 0.0007982507522567704, + "loss": 3.3257, + "theoretical_loss": 5.4882190074793575, + "tokens_seen": 40173568 + }, + { + "epoch": 0.01, + "learning_rate": 0.0007982347041123371, + "loss": 3.3909, + "theoretical_loss": 5.486826922177387, + "tokens_seen": 40239104 + }, + { + "epoch": 0.01, + "learning_rate": 0.0007982186559679038, + "loss": 3.3093, + "theoretical_loss": 5.4854377359217406, + "tokens_seen": 40304640 + }, + { + "epoch": 0.01, + "learning_rate": 0.0007982026078234705, + "loss": 3.3669, + "theoretical_loss": 5.484051437975901, + "tokens_seen": 40370176 + }, + { + "epoch": 0.01, + "learning_rate": 0.0007981865596790372, + "loss": 3.2858, + "theoretical_loss": 5.482668017660462, + "tokens_seen": 40435712 + }, + { + "epoch": 0.01, + "learning_rate": 0.0007981705115346039, + "loss": 3.5071, + "theoretical_loss": 5.481287464352727, + "tokens_seen": 40501248 + }, + { + "epoch": 0.01, + "learning_rate": 0.0007981544633901706, + "loss": 3.3696, + "theoretical_loss": 5.479909767486326, + "tokens_seen": 40566784 + }, + { + "epoch": 0.01, + "learning_rate": 0.0007981384152457373, + "loss": 3.3582, + "theoretical_loss": 5.478534916550817, + "tokens_seen": 40632320 + }, + { + "epoch": 0.01, + "learning_rate": 0.000798122367101304, + "loss": 3.5942, + "theoretical_loss": 5.477162901091306, + "tokens_seen": 40697856 + }, + { + "epoch": 0.01, + "learning_rate": 0.0007981063189568707, + "loss": 3.3191, + "theoretical_loss": 5.475793710708061, + "tokens_seen": 40763392 + }, + { + "epoch": 0.01, + "learning_rate": 0.0007980902708124374, + "loss": 3.4508, + "theoretical_loss": 5.474427335056138, + "tokens_seen": 40828928 + }, + { + "epoch": 0.01, + "learning_rate": 0.0007980742226680041, + "loss": 3.3342, + "theoretical_loss": 5.473063763844998, + "tokens_seen": 40894464 + }, + { + "epoch": 0.01, + "learning_rate": 0.0007980581745235708, + "loss": 3.4699, + "theoretical_loss": 5.471702986838139, + "tokens_seen": 40960000 + }, + { + "epoch": 0.01, + "learning_rate": 0.0007980421263791375, + "loss": 3.4539, + "theoretical_loss": 5.470344993852729, + "tokens_seen": 41025536 + }, + { + "epoch": 0.01, + "learning_rate": 0.0007980260782347042, + "loss": 3.3324, + "theoretical_loss": 5.46898977475923, + "tokens_seen": 41091072 + }, + { + "epoch": 0.01, + "learning_rate": 0.0007980100300902709, + "loss": 3.31, + "theoretical_loss": 5.467637319481042, + "tokens_seen": 41156608 + }, + { + "epoch": 0.01, + "learning_rate": 0.0007979939819458376, + "loss": 3.3149, + "theoretical_loss": 5.466287617994144, + "tokens_seen": 41222144 + }, + { + "epoch": 0.01, + "learning_rate": 0.0007979779338014043, + "loss": 3.5059, + "theoretical_loss": 5.464940660326731, + "tokens_seen": 41287680 + }, + { + "epoch": 0.01, + "learning_rate": 0.000797961885656971, + "loss": 3.3213, + "theoretical_loss": 5.463596436558861, + "tokens_seen": 41353216 + }, + { + "epoch": 0.01, + "learning_rate": 0.0007979458375125376, + "loss": 3.2666, + "theoretical_loss": 5.462254936822105, + "tokens_seen": 41418752 + }, + { + "epoch": 0.01, + "learning_rate": 0.0007979297893681044, + "loss": 3.1943, + "theoretical_loss": 5.4609161512992, + "tokens_seen": 41484288 + }, + { + "epoch": 0.01, + "learning_rate": 0.000797913741223671, + "loss": 3.3354, + "theoretical_loss": 5.4595800702237005, + "tokens_seen": 41549824 + }, + { + "epoch": 0.01, + "learning_rate": 0.0007978976930792378, + "loss": 3.2864, + "theoretical_loss": 5.458246683879636, + "tokens_seen": 41615360 + }, + { + "epoch": 0.01, + "learning_rate": 0.0007978816449348044, + "loss": 3.3117, + "theoretical_loss": 5.456915982601173, + "tokens_seen": 41680896 + }, + { + "epoch": 0.01, + "learning_rate": 0.0007978655967903712, + "loss": 3.3061, + "theoretical_loss": 5.455587956772277, + "tokens_seen": 41746432 + }, + { + "epoch": 0.01, + "learning_rate": 0.0007978495486459378, + "loss": 3.3569, + "theoretical_loss": 5.454262596826377, + "tokens_seen": 41811968 + }, + { + "epoch": 0.01, + "learning_rate": 0.0007978335005015046, + "loss": 3.4201, + "theoretical_loss": 5.452939893246034, + "tokens_seen": 41877504 + }, + { + "epoch": 0.01, + "learning_rate": 0.0007978174523570712, + "loss": 3.2255, + "theoretical_loss": 5.451619836562617, + "tokens_seen": 41943040 + }, + { + "epoch": 0.01, + "learning_rate": 0.0007978014042126379, + "loss": 3.1569, + "theoretical_loss": 5.450302417355973, + "tokens_seen": 42008576 + }, + { + "epoch": 0.01, + "learning_rate": 0.0007977853560682046, + "loss": 3.2409, + "theoretical_loss": 5.448987626254097, + "tokens_seen": 42074112 + }, + { + "epoch": 0.01, + "learning_rate": 0.0007977693079237713, + "loss": 3.4047, + "theoretical_loss": 5.447675453932829, + "tokens_seen": 42139648 + }, + { + "epoch": 0.01, + "learning_rate": 0.000797753259779338, + "loss": 3.2751, + "theoretical_loss": 5.446365891115516, + "tokens_seen": 42205184 + }, + { + "epoch": 0.01, + "learning_rate": 0.0007977372116349047, + "loss": 3.2969, + "theoretical_loss": 5.445058928572712, + "tokens_seen": 42270720 + }, + { + "epoch": 0.01, + "learning_rate": 0.0007977211634904714, + "loss": 3.0706, + "theoretical_loss": 5.443754557121853, + "tokens_seen": 42336256 + }, + { + "epoch": 0.01, + "learning_rate": 0.0007977051153460381, + "loss": 3.3292, + "theoretical_loss": 5.442452767626957, + "tokens_seen": 42401792 + }, + { + "epoch": 0.01, + "learning_rate": 0.0007976890672016048, + "loss": 3.2338, + "theoretical_loss": 5.441153550998311, + "tokens_seen": 42467328 + }, + { + "epoch": 0.01, + "learning_rate": 0.0007976730190571715, + "loss": 3.4458, + "theoretical_loss": 5.439856898192163, + "tokens_seen": 42532864 + }, + { + "epoch": 0.01, + "objective/train/docs_used": 22938, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 3.327707529067993, + "objective/train/theoretical_loss": 5.438562800210427, + "objective/train/tokens_used": 63058400, + "theoretical_loss": 5.438562800210427, + "tokens_seen": 42598400 + }, + { + "epoch": 0.01, + "learning_rate": 0.0007976569709127382, + "loss": 3.3737, + "theoretical_loss": 5.438562800210427, + "tokens_seen": 42598400 + }, + { + "epoch": 0.01, + "learning_rate": 0.0007976409227683049, + "loss": 3.2308, + "theoretical_loss": 5.437271248100375, + "tokens_seen": 42663936 + }, + { + "epoch": 0.01, + "learning_rate": 0.0007976248746238716, + "loss": 3.167, + "theoretical_loss": 5.435982232954342, + "tokens_seen": 42729472 + }, + { + "epoch": 0.01, + "learning_rate": 0.0007976088264794383, + "loss": 3.232, + "theoretical_loss": 5.434695745909433, + "tokens_seen": 42795008 + }, + { + "epoch": 0.01, + "learning_rate": 0.000797592778335005, + "loss": 3.2576, + "theoretical_loss": 5.433411778147226, + "tokens_seen": 42860544 + }, + { + "epoch": 0.01, + "learning_rate": 0.0007975767301905718, + "loss": 3.1416, + "theoretical_loss": 5.432130320893483, + "tokens_seen": 42926080 + }, + { + "epoch": 0.01, + "learning_rate": 0.0007975606820461384, + "loss": 3.2612, + "theoretical_loss": 5.430851365417862, + "tokens_seen": 42991616 + }, + { + "epoch": 0.01, + "learning_rate": 0.0007975446339017052, + "loss": 3.1622, + "theoretical_loss": 5.429574903033627, + "tokens_seen": 43057152 + }, + { + "epoch": 0.01, + "learning_rate": 0.0007975285857572718, + "loss": 3.2983, + "theoretical_loss": 5.428300925097373, + "tokens_seen": 43122688 + }, + { + "epoch": 0.01, + "learning_rate": 0.0007975125376128386, + "loss": 3.2018, + "theoretical_loss": 5.427029423008732, + "tokens_seen": 43188224 + }, + { + "epoch": 0.01, + "learning_rate": 0.0007974964894684052, + "loss": 3.1671, + "theoretical_loss": 5.425760388210108, + "tokens_seen": 43253760 + }, + { + "epoch": 0.01, + "learning_rate": 0.000797480441323972, + "loss": 3.2674, + "theoretical_loss": 5.424493812186386, + "tokens_seen": 43319296 + }, + { + "epoch": 0.01, + "learning_rate": 0.0007974643931795386, + "loss": 3.1448, + "theoretical_loss": 5.423229686464667, + "tokens_seen": 43384832 + }, + { + "epoch": 0.01, + "learning_rate": 0.0007974483450351054, + "loss": 3.2137, + "theoretical_loss": 5.421968002613992, + "tokens_seen": 43450368 + }, + { + "epoch": 0.01, + "learning_rate": 0.000797432296890672, + "loss": 3.4475, + "theoretical_loss": 5.420708752245071, + "tokens_seen": 43515904 + }, + { + "epoch": 0.01, + "learning_rate": 0.0007974162487462388, + "loss": 3.2962, + "theoretical_loss": 5.419451927010014, + "tokens_seen": 43581440 + }, + { + "epoch": 0.01, + "learning_rate": 0.0007974002006018054, + "loss": 3.3231, + "theoretical_loss": 5.418197518602068, + "tokens_seen": 43646976 + }, + { + "epoch": 0.01, + "learning_rate": 0.0007973841524573722, + "loss": 3.2739, + "theoretical_loss": 5.41694551875535, + "tokens_seen": 43712512 + }, + { + "epoch": 0.01, + "learning_rate": 0.0007973681043129388, + "loss": 3.2956, + "theoretical_loss": 5.415695919244586, + "tokens_seen": 43778048 + }, + { + "epoch": 0.01, + "learning_rate": 0.0007973520561685056, + "loss": 3.2201, + "theoretical_loss": 5.41444871188485, + "tokens_seen": 43843584 + }, + { + "epoch": 0.01, + "learning_rate": 0.0007973360080240722, + "loss": 3.0638, + "theoretical_loss": 5.413203888531308, + "tokens_seen": 43909120 + }, + { + "epoch": 0.01, + "learning_rate": 0.000797319959879639, + "loss": 3.1342, + "theoretical_loss": 5.411961441078963, + "tokens_seen": 43974656 + }, + { + "epoch": 0.01, + "learning_rate": 0.0007973039117352056, + "loss": 3.2203, + "theoretical_loss": 5.410721361462395, + "tokens_seen": 44040192 + }, + { + "epoch": 0.01, + "learning_rate": 0.0007972878635907724, + "loss": 3.4047, + "theoretical_loss": 5.409483641655516, + "tokens_seen": 44105728 + }, + { + "epoch": 0.01, + "learning_rate": 0.000797271815446339, + "loss": 3.1766, + "theoretical_loss": 5.408248273671317, + "tokens_seen": 44171264 + }, + { + "epoch": 0.01, + "learning_rate": 0.0007972557673019058, + "loss": 3.0235, + "theoretical_loss": 5.407015249561624, + "tokens_seen": 44236800 + }, + { + "epoch": 0.01, + "learning_rate": 0.0007972397191574724, + "loss": 3.2561, + "theoretical_loss": 5.405784561416844, + "tokens_seen": 44302336 + }, + { + "epoch": 0.01, + "learning_rate": 0.0007972236710130392, + "loss": 3.0968, + "theoretical_loss": 5.404556201365727, + "tokens_seen": 44367872 + }, + { + "epoch": 0.01, + "learning_rate": 0.0007972076228686058, + "loss": 3.1796, + "theoretical_loss": 5.403330161575123, + "tokens_seen": 44433408 + }, + { + "epoch": 0.01, + "learning_rate": 0.0007971915747241726, + "loss": 3.1669, + "theoretical_loss": 5.402106434249742, + "tokens_seen": 44498944 + }, + { + "epoch": 0.01, + "learning_rate": 0.0007971755265797392, + "loss": 3.1565, + "theoretical_loss": 5.400885011631912, + "tokens_seen": 44564480 + }, + { + "epoch": 0.01, + "learning_rate": 0.000797159478435306, + "loss": 3.2563, + "theoretical_loss": 5.399665886001349, + "tokens_seen": 44630016 + }, + { + "epoch": 0.01, + "learning_rate": 0.0007971434302908726, + "loss": 3.1613, + "theoretical_loss": 5.398449049674916, + "tokens_seen": 44695552 + }, + { + "epoch": 0.01, + "learning_rate": 0.0007971273821464394, + "loss": 3.1838, + "theoretical_loss": 5.397234495006392, + "tokens_seen": 44761088 + }, + { + "epoch": 0.01, + "learning_rate": 0.000797111334002006, + "loss": 3.2226, + "theoretical_loss": 5.396022214386246, + "tokens_seen": 44826624 + }, + { + "epoch": 0.01, + "learning_rate": 0.0007970952858575728, + "loss": 2.9938, + "theoretical_loss": 5.394812200241393, + "tokens_seen": 44892160 + }, + { + "epoch": 0.01, + "learning_rate": 0.0007970792377131394, + "loss": 3.0679, + "theoretical_loss": 5.393604445034988, + "tokens_seen": 44957696 + }, + { + "epoch": 0.01, + "learning_rate": 0.0007970631895687062, + "loss": 3.2004, + "theoretical_loss": 5.392398941266183, + "tokens_seen": 45023232 + }, + { + "epoch": 0.01, + "learning_rate": 0.0007970471414242728, + "loss": 3.0204, + "theoretical_loss": 5.391195681469908, + "tokens_seen": 45088768 + }, + { + "epoch": 0.01, + "learning_rate": 0.0007970310932798396, + "loss": 3.2206, + "theoretical_loss": 5.389994658216651, + "tokens_seen": 45154304 + }, + { + "epoch": 0.01, + "learning_rate": 0.0007970150451354062, + "loss": 3.1509, + "theoretical_loss": 5.388795864112235, + "tokens_seen": 45219840 + }, + { + "epoch": 0.01, + "learning_rate": 0.000796998996990973, + "loss": 3.0982, + "theoretical_loss": 5.387599291797599, + "tokens_seen": 45285376 + }, + { + "epoch": 0.01, + "learning_rate": 0.0007969829488465396, + "loss": 3.1327, + "theoretical_loss": 5.386404933948584, + "tokens_seen": 45350912 + }, + { + "epoch": 0.01, + "learning_rate": 0.0007969669007021064, + "loss": 3.0638, + "theoretical_loss": 5.38521278327571, + "tokens_seen": 45416448 + }, + { + "epoch": 0.01, + "learning_rate": 0.000796950852557673, + "loss": 2.9757, + "theoretical_loss": 5.384022832523971, + "tokens_seen": 45481984 + }, + { + "epoch": 0.01, + "learning_rate": 0.0007969348044132398, + "loss": 3.1146, + "theoretical_loss": 5.382835074472616, + "tokens_seen": 45547520 + }, + { + "epoch": 0.01, + "learning_rate": 0.0007969187562688065, + "loss": 3.2602, + "theoretical_loss": 5.381649501934939, + "tokens_seen": 45613056 + }, + { + "epoch": 0.01, + "learning_rate": 0.0007969027081243732, + "loss": 3.2194, + "theoretical_loss": 5.3804661077580755, + "tokens_seen": 45678592 + }, + { + "epoch": 0.01, + "learning_rate": 0.0007968866599799399, + "loss": 3.1288, + "theoretical_loss": 5.379284884822787, + "tokens_seen": 45744128 + }, + { + "epoch": 0.01, + "learning_rate": 0.0007968706118355066, + "loss": 3.2619, + "theoretical_loss": 5.37810582604326, + "tokens_seen": 45809664 + }, + { + "epoch": 0.01, + "objective/train/docs_used": 24173, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 3.0289993286132812, + "objective/train/theoretical_loss": 5.376928924366901, + "objective/train/tokens_used": 66335200, + "theoretical_loss": 5.376928924366901, + "tokens_seen": 45875200 + }, + { + "epoch": 0.01, + "learning_rate": 0.0007968545636910733, + "loss": 3.0563, + "theoretical_loss": 5.376928924366901, + "tokens_seen": 45875200 + }, + { + "epoch": 0.01, + "learning_rate": 0.00079683851554664, + "loss": 3.1021, + "theoretical_loss": 5.37575417277413, + "tokens_seen": 45940736 + }, + { + "epoch": 0.01, + "learning_rate": 0.0007968224674022067, + "loss": 3.2264, + "theoretical_loss": 5.374581564278184, + "tokens_seen": 46006272 + }, + { + "epoch": 0.01, + "learning_rate": 0.0007968064192577734, + "loss": 3.0505, + "theoretical_loss": 5.373411091924918, + "tokens_seen": 46071808 + }, + { + "epoch": 0.01, + "learning_rate": 0.0007967903711133401, + "loss": 3.0466, + "theoretical_loss": 5.372242748792598, + "tokens_seen": 46137344 + }, + { + "epoch": 0.01, + "learning_rate": 0.0007967743229689068, + "loss": 3.1208, + "theoretical_loss": 5.371076527991711, + "tokens_seen": 46202880 + }, + { + "epoch": 0.01, + "learning_rate": 0.0007967582748244735, + "loss": 3.1449, + "theoretical_loss": 5.36991242266477, + "tokens_seen": 46268416 + }, + { + "epoch": 0.01, + "learning_rate": 0.0007967422266800402, + "loss": 2.9019, + "theoretical_loss": 5.36875042598612, + "tokens_seen": 46333952 + }, + { + "epoch": 0.01, + "learning_rate": 0.0007967261785356069, + "loss": 3.0713, + "theoretical_loss": 5.3675905311617385, + "tokens_seen": 46399488 + }, + { + "epoch": 0.01, + "learning_rate": 0.0007967101303911736, + "loss": 3.1251, + "theoretical_loss": 5.366432731429054, + "tokens_seen": 46465024 + }, + { + "epoch": 0.01, + "learning_rate": 0.0007966940822467403, + "loss": 3.3322, + "theoretical_loss": 5.365277020056752, + "tokens_seen": 46530560 + }, + { + "epoch": 0.01, + "learning_rate": 0.000796678034102307, + "loss": 2.9947, + "theoretical_loss": 5.364123390344582, + "tokens_seen": 46596096 + }, + { + "epoch": 0.01, + "learning_rate": 0.0007966619859578737, + "loss": 3.2612, + "theoretical_loss": 5.362971835623181, + "tokens_seen": 46661632 + }, + { + "epoch": 0.01, + "learning_rate": 0.0007966459378134404, + "loss": 3.1102, + "theoretical_loss": 5.361822349253881, + "tokens_seen": 46727168 + }, + { + "epoch": 0.01, + "learning_rate": 0.0007966298896690071, + "loss": 3.1557, + "theoretical_loss": 5.360674924628524, + "tokens_seen": 46792704 + }, + { + "epoch": 0.01, + "learning_rate": 0.0007966138415245738, + "loss": 3.1385, + "theoretical_loss": 5.359529555169287, + "tokens_seen": 46858240 + }, + { + "epoch": 0.01, + "learning_rate": 0.0007965977933801405, + "loss": 3.1181, + "theoretical_loss": 5.358386234328487, + "tokens_seen": 46923776 + }, + { + "epoch": 0.01, + "learning_rate": 0.0007965817452357072, + "loss": 3.0314, + "theoretical_loss": 5.357244955588417, + "tokens_seen": 46989312 + }, + { + "epoch": 0.01, + "learning_rate": 0.0007965656970912739, + "loss": 3.1929, + "theoretical_loss": 5.356105712461154, + "tokens_seen": 47054848 + }, + { + "epoch": 0.01, + "learning_rate": 0.0007965496489468405, + "loss": 3.226, + "theoretical_loss": 5.354968498488391, + "tokens_seen": 47120384 + }, + { + "epoch": 0.01, + "learning_rate": 0.0007965336008024073, + "loss": 2.9977, + "theoretical_loss": 5.353833307241256, + "tokens_seen": 47185920 + }, + { + "epoch": 0.01, + "learning_rate": 0.0007965175526579739, + "loss": 3.1227, + "theoretical_loss": 5.352700132320135, + "tokens_seen": 47251456 + }, + { + "epoch": 0.01, + "learning_rate": 0.0007965015045135407, + "loss": 3.1262, + "theoretical_loss": 5.351568967354504, + "tokens_seen": 47316992 + }, + { + "epoch": 0.01, + "learning_rate": 0.0007964854563691073, + "loss": 3.2971, + "theoretical_loss": 5.350439806002757, + "tokens_seen": 47382528 + }, + { + "epoch": 0.01, + "learning_rate": 0.0007964694082246741, + "loss": 2.9265, + "theoretical_loss": 5.349312641952025, + "tokens_seen": 47448064 + }, + { + "epoch": 0.01, + "learning_rate": 0.0007964533600802407, + "loss": 2.9574, + "theoretical_loss": 5.348187468918022, + "tokens_seen": 47513600 + }, + { + "epoch": 0.01, + "learning_rate": 0.0007964373119358075, + "loss": 3.1343, + "theoretical_loss": 5.347064280644861, + "tokens_seen": 47579136 + }, + { + "epoch": 0.01, + "learning_rate": 0.0007964212637913741, + "loss": 3.0708, + "theoretical_loss": 5.345943070904897, + "tokens_seen": 47644672 + }, + { + "epoch": 0.01, + "learning_rate": 0.0007964052156469408, + "loss": 3.1461, + "theoretical_loss": 5.344823833498557, + "tokens_seen": 47710208 + }, + { + "epoch": 0.01, + "learning_rate": 0.0007963891675025075, + "loss": 2.9919, + "theoretical_loss": 5.343706562254177, + "tokens_seen": 47775744 + }, + { + "epoch": 0.01, + "learning_rate": 0.0007963731193580743, + "loss": 3.0731, + "theoretical_loss": 5.342591251027833, + "tokens_seen": 47841280 + }, + { + "epoch": 0.01, + "learning_rate": 0.0007963570712136409, + "loss": 3.0833, + "theoretical_loss": 5.3414778937031855, + "tokens_seen": 47906816 + }, + { + "epoch": 0.01, + "learning_rate": 0.0007963410230692077, + "loss": 2.9718, + "theoretical_loss": 5.340366484191316, + "tokens_seen": 47972352 + }, + { + "epoch": 0.01, + "learning_rate": 0.0007963249749247743, + "loss": 2.9305, + "theoretical_loss": 5.339257016430563, + "tokens_seen": 48037888 + }, + { + "epoch": 0.01, + "learning_rate": 0.0007963089267803411, + "loss": 2.8745, + "theoretical_loss": 5.338149484386367, + "tokens_seen": 48103424 + }, + { + "epoch": 0.01, + "learning_rate": 0.0007962928786359077, + "loss": 2.8049, + "theoretical_loss": 5.337043882051111, + "tokens_seen": 48168960 + }, + { + "epoch": 0.01, + "learning_rate": 0.0007962768304914745, + "loss": 3.1916, + "theoretical_loss": 5.3359402034439665, + "tokens_seen": 48234496 + }, + { + "epoch": 0.01, + "learning_rate": 0.0007962607823470411, + "loss": 3.0542, + "theoretical_loss": 5.334838442610732, + "tokens_seen": 48300032 + }, + { + "epoch": 0.01, + "learning_rate": 0.0007962447342026079, + "loss": 2.9216, + "theoretical_loss": 5.333738593623686, + "tokens_seen": 48365568 + }, + { + "epoch": 0.01, + "learning_rate": 0.0007962286860581745, + "loss": 2.8883, + "theoretical_loss": 5.332640650581424, + "tokens_seen": 48431104 + }, + { + "epoch": 0.01, + "learning_rate": 0.0007962126379137413, + "loss": 2.9486, + "theoretical_loss": 5.331544607608718, + "tokens_seen": 48496640 + }, + { + "epoch": 0.01, + "learning_rate": 0.0007961965897693079, + "loss": 3.0114, + "theoretical_loss": 5.330450458856356, + "tokens_seen": 48562176 + }, + { + "epoch": 0.01, + "learning_rate": 0.0007961805416248747, + "loss": 3.0676, + "theoretical_loss": 5.329358198500994, + "tokens_seen": 48627712 + }, + { + "epoch": 0.01, + "learning_rate": 0.0007961644934804413, + "loss": 3.0856, + "theoretical_loss": 5.328267820745011, + "tokens_seen": 48693248 + }, + { + "epoch": 0.01, + "learning_rate": 0.0007961484453360081, + "loss": 2.9656, + "theoretical_loss": 5.327179319816354, + "tokens_seen": 48758784 + }, + { + "epoch": 0.01, + "learning_rate": 0.0007961323971915747, + "loss": 2.9607, + "theoretical_loss": 5.326092689968397, + "tokens_seen": 48824320 + }, + { + "epoch": 0.01, + "learning_rate": 0.0007961163490471415, + "loss": 3.0658, + "theoretical_loss": 5.3250079254797935, + "tokens_seen": 48889856 + }, + { + "epoch": 0.01, + "learning_rate": 0.0007961003009027081, + "loss": 3.0627, + "theoretical_loss": 5.323925020654329, + "tokens_seen": 48955392 + }, + { + "epoch": 0.01, + "learning_rate": 0.0007960842527582749, + "loss": 3.0791, + "theoretical_loss": 5.322843969820781, + "tokens_seen": 49020928 + }, + { + "epoch": 0.01, + "learning_rate": 0.0007960682046138415, + "loss": 3.1406, + "theoretical_loss": 5.321764767332772, + "tokens_seen": 49086464 + }, + { + "epoch": 0.01, + "objective/train/docs_used": 25394, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 3.053786277770996, + "objective/train/theoretical_loss": 5.320687407568633, + "objective/train/tokens_used": 69612000, + "theoretical_loss": 5.320687407568633, + "tokens_seen": 49152000 + }, + { + "epoch": 0.01, + "learning_rate": 0.0007960521564694083, + "loss": 3.0162, + "theoretical_loss": 5.320687407568633, + "tokens_seen": 49152000 + }, + { + "epoch": 0.01, + "learning_rate": 0.0007960361083249749, + "loss": 2.9864, + "theoretical_loss": 5.3196118849312555, + "tokens_seen": 49217536 + }, + { + "epoch": 0.01, + "learning_rate": 0.0007960200601805417, + "loss": 2.8737, + "theoretical_loss": 5.31853819384796, + "tokens_seen": 49283072 + }, + { + "epoch": 0.01, + "learning_rate": 0.0007960040120361083, + "loss": 2.9983, + "theoretical_loss": 5.31746632877035, + "tokens_seen": 49348608 + }, + { + "epoch": 0.01, + "learning_rate": 0.0007959879638916751, + "loss": 2.9766, + "theoretical_loss": 5.316396284174177, + "tokens_seen": 49414144 + }, + { + "epoch": 0.01, + "learning_rate": 0.0007959719157472417, + "loss": 2.7321, + "theoretical_loss": 5.3153280545592025, + "tokens_seen": 49479680 + }, + { + "epoch": 0.02, + "learning_rate": 0.0007959558676028085, + "loss": 2.913, + "theoretical_loss": 5.314261634449066, + "tokens_seen": 49545216 + }, + { + "epoch": 0.02, + "learning_rate": 0.0007959398194583751, + "loss": 2.9795, + "theoretical_loss": 5.313197018391143, + "tokens_seen": 49610752 + }, + { + "epoch": 0.02, + "learning_rate": 0.0007959237713139419, + "loss": 2.8927, + "theoretical_loss": 5.312134200956416, + "tokens_seen": 49676288 + }, + { + "epoch": 0.02, + "learning_rate": 0.0007959077231695085, + "loss": 2.8683, + "theoretical_loss": 5.3110731767393435, + "tokens_seen": 49741824 + }, + { + "epoch": 0.02, + "learning_rate": 0.0007958916750250753, + "loss": 2.8908, + "theoretical_loss": 5.31001394035772, + "tokens_seen": 49807360 + }, + { + "epoch": 0.02, + "learning_rate": 0.0007958756268806419, + "loss": 2.7939, + "theoretical_loss": 5.308956486452549, + "tokens_seen": 49872896 + }, + { + "epoch": 0.02, + "learning_rate": 0.0007958595787362087, + "loss": 2.9764, + "theoretical_loss": 5.307900809687913, + "tokens_seen": 49938432 + }, + { + "epoch": 0.02, + "learning_rate": 0.0007958435305917753, + "loss": 2.9051, + "theoretical_loss": 5.306846904750847, + "tokens_seen": 50003968 + }, + { + "epoch": 0.02, + "learning_rate": 0.0007958274824473421, + "loss": 2.8669, + "theoretical_loss": 5.305794766351202, + "tokens_seen": 50069504 + }, + { + "epoch": 0.02, + "learning_rate": 0.0007958114343029087, + "loss": 2.9247, + "theoretical_loss": 5.304744389221523, + "tokens_seen": 50135040 + }, + { + "epoch": 0.02, + "learning_rate": 0.0007957953861584755, + "loss": 2.9377, + "theoretical_loss": 5.303695768116921, + "tokens_seen": 50200576 + }, + { + "epoch": 0.02, + "learning_rate": 0.0007957793380140421, + "loss": 2.7793, + "theoretical_loss": 5.302648897814946, + "tokens_seen": 50266112 + }, + { + "epoch": 0.02, + "learning_rate": 0.0007957632898696089, + "loss": 2.8698, + "theoretical_loss": 5.301603773115465, + "tokens_seen": 50331648 + }, + { + "epoch": 0.02, + "learning_rate": 0.0007957472417251755, + "loss": 2.924, + "theoretical_loss": 5.300560388840532, + "tokens_seen": 50397184 + }, + { + "epoch": 0.02, + "learning_rate": 0.0007957311935807423, + "loss": 2.8245, + "theoretical_loss": 5.299518739834273, + "tokens_seen": 50462720 + }, + { + "epoch": 0.02, + "learning_rate": 0.0007957151454363089, + "loss": 2.8056, + "theoretical_loss": 5.298478820962751, + "tokens_seen": 50528256 + }, + { + "epoch": 0.02, + "learning_rate": 0.0007956990972918757, + "loss": 2.8558, + "theoretical_loss": 5.297440627113859, + "tokens_seen": 50593792 + }, + { + "epoch": 0.02, + "learning_rate": 0.0007956830491474423, + "loss": 2.8927, + "theoretical_loss": 5.296404153197189, + "tokens_seen": 50659328 + }, + { + "epoch": 0.02, + "learning_rate": 0.0007956670010030091, + "loss": 2.8758, + "theoretical_loss": 5.295369394143913, + "tokens_seen": 50724864 + }, + { + "epoch": 0.02, + "learning_rate": 0.0007956509528585757, + "loss": 3.0202, + "theoretical_loss": 5.294336344906668, + "tokens_seen": 50790400 + }, + { + "epoch": 0.02, + "learning_rate": 0.0007956349047141425, + "loss": 2.9869, + "theoretical_loss": 5.293305000459435, + "tokens_seen": 50855936 + }, + { + "epoch": 0.02, + "learning_rate": 0.0007956188565697091, + "loss": 2.8233, + "theoretical_loss": 5.292275355797421, + "tokens_seen": 50921472 + }, + { + "epoch": 0.02, + "learning_rate": 0.0007956028084252759, + "loss": 2.9185, + "theoretical_loss": 5.291247405936941, + "tokens_seen": 50987008 + }, + { + "epoch": 0.02, + "learning_rate": 0.0007955867602808426, + "loss": 2.7764, + "theoretical_loss": 5.290221145915307, + "tokens_seen": 51052544 + }, + { + "epoch": 0.02, + "learning_rate": 0.0007955707121364093, + "loss": 2.783, + "theoretical_loss": 5.289196570790708, + "tokens_seen": 51118080 + }, + { + "epoch": 0.02, + "learning_rate": 0.000795554663991976, + "loss": 2.973, + "theoretical_loss": 5.288173675642094, + "tokens_seen": 51183616 + }, + { + "epoch": 0.02, + "learning_rate": 0.0007955386158475427, + "loss": 2.7109, + "theoretical_loss": 5.287152455569069, + "tokens_seen": 51249152 + }, + { + "epoch": 0.02, + "learning_rate": 0.0007955225677031094, + "loss": 2.6985, + "theoretical_loss": 5.286132905691774, + "tokens_seen": 51314688 + }, + { + "epoch": 0.02, + "learning_rate": 0.0007955065195586761, + "loss": 2.8212, + "theoretical_loss": 5.285115021150775, + "tokens_seen": 51380224 + }, + { + "epoch": 0.02, + "learning_rate": 0.0007954904714142428, + "loss": 2.8413, + "theoretical_loss": 5.284098797106948, + "tokens_seen": 51445760 + }, + { + "epoch": 0.02, + "learning_rate": 0.0007954744232698095, + "loss": 2.9257, + "theoretical_loss": 5.283084228741377, + "tokens_seen": 51511296 + }, + { + "epoch": 0.02, + "learning_rate": 0.0007954583751253762, + "loss": 2.8072, + "theoretical_loss": 5.282071311255235, + "tokens_seen": 51576832 + }, + { + "epoch": 0.02, + "learning_rate": 0.0007954423269809429, + "loss": 2.6292, + "theoretical_loss": 5.2810600398696845, + "tokens_seen": 51642368 + }, + { + "epoch": 0.02, + "learning_rate": 0.0007954262788365096, + "loss": 2.7969, + "theoretical_loss": 5.2800504098257575, + "tokens_seen": 51707904 + }, + { + "epoch": 0.02, + "learning_rate": 0.0007954102306920763, + "loss": 2.7126, + "theoretical_loss": 5.279042416384257, + "tokens_seen": 51773440 + }, + { + "epoch": 0.02, + "learning_rate": 0.000795394182547643, + "loss": 2.8646, + "theoretical_loss": 5.278036054825645, + "tokens_seen": 51838976 + }, + { + "epoch": 0.02, + "learning_rate": 0.0007953781344032097, + "loss": 2.8047, + "theoretical_loss": 5.277031320449941, + "tokens_seen": 51904512 + }, + { + "epoch": 0.02, + "learning_rate": 0.0007953620862587764, + "loss": 2.8045, + "theoretical_loss": 5.27602820857661, + "tokens_seen": 51970048 + }, + { + "epoch": 0.02, + "learning_rate": 0.0007953460381143431, + "loss": 2.8097, + "theoretical_loss": 5.275026714544463, + "tokens_seen": 52035584 + }, + { + "epoch": 0.02, + "learning_rate": 0.0007953299899699098, + "loss": 2.7058, + "theoretical_loss": 5.27402683371155, + "tokens_seen": 52101120 + }, + { + "epoch": 0.02, + "learning_rate": 0.0007953139418254765, + "loss": 2.6665, + "theoretical_loss": 5.273028561455056, + "tokens_seen": 52166656 + }, + { + "epoch": 0.02, + "learning_rate": 0.0007952978936810432, + "loss": 2.8235, + "theoretical_loss": 5.272031893171203, + "tokens_seen": 52232192 + }, + { + "epoch": 0.02, + "learning_rate": 0.0007952818455366099, + "loss": 2.7788, + "theoretical_loss": 5.27103682427514, + "tokens_seen": 52297728 + }, + { + "epoch": 0.02, + "learning_rate": 0.0007952657973921766, + "loss": 2.8237, + "theoretical_loss": 5.270043350200851, + "tokens_seen": 52363264 + }, + { + "epoch": 0.02, + "objective/train/docs_used": 26541, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 2.6444597244262695, + "objective/train/theoretical_loss": 5.269051466401043, + "objective/train/tokens_used": 72888800, + "theoretical_loss": 5.269051466401043, + "tokens_seen": 52428800 + }, + { + "epoch": 0.02, + "learning_rate": 0.0007952497492477433, + "loss": 2.6213, + "theoretical_loss": 5.269051466401043, + "tokens_seen": 52428800 + }, + { + "epoch": 0.02, + "learning_rate": 0.00079523370110331, + "loss": 2.7597, + "theoretical_loss": 5.268061168347058, + "tokens_seen": 52494336 + }, + { + "epoch": 0.02, + "learning_rate": 0.0007952176529588767, + "loss": 2.8027, + "theoretical_loss": 5.267072451528764, + "tokens_seen": 52559872 + }, + { + "epoch": 0.02, + "learning_rate": 0.0007952016048144434, + "loss": 2.7992, + "theoretical_loss": 5.266085311454461, + "tokens_seen": 52625408 + }, + { + "epoch": 0.02, + "learning_rate": 0.0007951855566700101, + "loss": 2.7699, + "theoretical_loss": 5.265099743650779, + "tokens_seen": 52690944 + }, + { + "epoch": 0.02, + "learning_rate": 0.0007951695085255768, + "loss": 2.8732, + "theoretical_loss": 5.264115743662588, + "tokens_seen": 52756480 + }, + { + "epoch": 0.02, + "learning_rate": 0.0007951534603811434, + "loss": 2.8061, + "theoretical_loss": 5.26313330705289, + "tokens_seen": 52822016 + }, + { + "epoch": 0.02, + "learning_rate": 0.0007951374122367102, + "loss": 2.6897, + "theoretical_loss": 5.262152429402734, + "tokens_seen": 52887552 + }, + { + "epoch": 0.02, + "learning_rate": 0.0007951213640922768, + "loss": 2.6925, + "theoretical_loss": 5.261173106311116, + "tokens_seen": 52953088 + }, + { + "epoch": 0.02, + "learning_rate": 0.0007951053159478436, + "loss": 2.8578, + "theoretical_loss": 5.2601953333948765, + "tokens_seen": 53018624 + }, + { + "epoch": 0.02, + "learning_rate": 0.0007950892678034102, + "loss": 2.7215, + "theoretical_loss": 5.25921910628862, + "tokens_seen": 53084160 + }, + { + "epoch": 0.02, + "learning_rate": 0.000795073219658977, + "loss": 2.6609, + "theoretical_loss": 5.258244420644611, + "tokens_seen": 53149696 + }, + { + "epoch": 0.02, + "learning_rate": 0.0007950571715145436, + "loss": 2.7235, + "theoretical_loss": 5.257271272132685, + "tokens_seen": 53215232 + }, + { + "epoch": 0.02, + "learning_rate": 0.0007950411233701104, + "loss": 2.6561, + "theoretical_loss": 5.256299656440154, + "tokens_seen": 53280768 + }, + { + "epoch": 0.02, + "learning_rate": 0.000795025075225677, + "loss": 2.6849, + "theoretical_loss": 5.255329569271717, + "tokens_seen": 53346304 + }, + { + "epoch": 0.02, + "learning_rate": 0.0007950090270812438, + "loss": 2.6087, + "theoretical_loss": 5.254361006349367, + "tokens_seen": 53411840 + }, + { + "epoch": 0.02, + "learning_rate": 0.0007949929789368104, + "loss": 2.7259, + "theoretical_loss": 5.2533939634122975, + "tokens_seen": 53477376 + }, + { + "epoch": 0.02, + "learning_rate": 0.0007949769307923772, + "loss": 2.6714, + "theoretical_loss": 5.252428436216821, + "tokens_seen": 53542912 + }, + { + "epoch": 0.02, + "learning_rate": 0.0007949608826479438, + "loss": 2.7025, + "theoretical_loss": 5.251464420536269, + "tokens_seen": 53608448 + }, + { + "epoch": 0.02, + "learning_rate": 0.0007949448345035106, + "loss": 2.6972, + "theoretical_loss": 5.250501912160911, + "tokens_seen": 53673984 + }, + { + "epoch": 0.02, + "learning_rate": 0.0007949287863590772, + "loss": 2.7932, + "theoretical_loss": 5.24954090689786, + "tokens_seen": 53739520 + }, + { + "epoch": 0.02, + "learning_rate": 0.000794912738214644, + "loss": 2.9039, + "theoretical_loss": 5.248581400570989, + "tokens_seen": 53805056 + }, + { + "epoch": 0.02, + "learning_rate": 0.0007948966900702106, + "loss": 2.764, + "theoretical_loss": 5.247623389020841, + "tokens_seen": 53870592 + }, + { + "epoch": 0.02, + "learning_rate": 0.0007948806419257774, + "loss": 2.6855, + "theoretical_loss": 5.246666868104547, + "tokens_seen": 53936128 + }, + { + "epoch": 0.02, + "learning_rate": 0.000794864593781344, + "loss": 2.7238, + "theoretical_loss": 5.245711833695731, + "tokens_seen": 54001664 + }, + { + "epoch": 0.02, + "learning_rate": 0.0007948485456369108, + "loss": 2.7811, + "theoretical_loss": 5.244758281684433, + "tokens_seen": 54067200 + }, + { + "epoch": 0.02, + "learning_rate": 0.0007948324974924774, + "loss": 2.7522, + "theoretical_loss": 5.24380620797702, + "tokens_seen": 54132736 + }, + { + "epoch": 0.02, + "learning_rate": 0.0007948164493480442, + "loss": 2.4551, + "theoretical_loss": 5.242855608496102, + "tokens_seen": 54198272 + }, + { + "epoch": 0.02, + "learning_rate": 0.0007948004012036108, + "loss": 2.7775, + "theoretical_loss": 5.2419064791804475, + "tokens_seen": 54263808 + }, + { + "epoch": 0.02, + "learning_rate": 0.0007947843530591776, + "loss": 2.7152, + "theoretical_loss": 5.240958815984901, + "tokens_seen": 54329344 + }, + { + "epoch": 0.02, + "learning_rate": 0.0007947683049147442, + "loss": 2.6175, + "theoretical_loss": 5.2400126148803, + "tokens_seen": 54394880 + }, + { + "epoch": 0.02, + "learning_rate": 0.000794752256770311, + "loss": 2.6674, + "theoretical_loss": 5.23906787185339, + "tokens_seen": 54460416 + }, + { + "epoch": 0.02, + "learning_rate": 0.0007947362086258776, + "loss": 2.6822, + "theoretical_loss": 5.238124582906748, + "tokens_seen": 54525952 + }, + { + "epoch": 0.02, + "learning_rate": 0.0007947201604814444, + "loss": 2.7534, + "theoretical_loss": 5.237182744058696, + "tokens_seen": 54591488 + }, + { + "epoch": 0.02, + "learning_rate": 0.000794704112337011, + "loss": 2.6713, + "theoretical_loss": 5.236242351343224, + "tokens_seen": 54657024 + }, + { + "epoch": 0.02, + "learning_rate": 0.0007946880641925778, + "loss": 2.6033, + "theoretical_loss": 5.235303400809904, + "tokens_seen": 54722560 + }, + { + "epoch": 0.02, + "learning_rate": 0.0007946720160481444, + "loss": 2.6995, + "theoretical_loss": 5.2343658885238185, + "tokens_seen": 54788096 + }, + { + "epoch": 0.02, + "learning_rate": 0.0007946559679037112, + "loss": 2.6862, + "theoretical_loss": 5.2334298105654735, + "tokens_seen": 54853632 + }, + { + "epoch": 0.02, + "learning_rate": 0.0007946399197592778, + "loss": 2.7475, + "theoretical_loss": 5.232495163030723, + "tokens_seen": 54919168 + }, + { + "epoch": 0.02, + "learning_rate": 0.0007946238716148446, + "loss": 2.745, + "theoretical_loss": 5.231561942030694, + "tokens_seen": 54984704 + }, + { + "epoch": 0.02, + "learning_rate": 0.0007946078234704112, + "loss": 2.6859, + "theoretical_loss": 5.230630143691698, + "tokens_seen": 55050240 + }, + { + "epoch": 0.02, + "learning_rate": 0.000794591775325978, + "loss": 2.8485, + "theoretical_loss": 5.22969976415517, + "tokens_seen": 55115776 + }, + { + "epoch": 0.02, + "learning_rate": 0.0007945757271815446, + "loss": 2.4169, + "theoretical_loss": 5.228770799577576, + "tokens_seen": 55181312 + }, + { + "epoch": 0.02, + "learning_rate": 0.0007945596790371114, + "loss": 2.6858, + "theoretical_loss": 5.227843246130343, + "tokens_seen": 55246848 + }, + { + "epoch": 0.02, + "learning_rate": 0.000794543630892678, + "loss": 2.5811, + "theoretical_loss": 5.2269170999997865, + "tokens_seen": 55312384 + }, + { + "epoch": 0.02, + "learning_rate": 0.0007945275827482448, + "loss": 2.6851, + "theoretical_loss": 5.225992357387031, + "tokens_seen": 55377920 + }, + { + "epoch": 0.02, + "learning_rate": 0.0007945115346038114, + "loss": 2.6441, + "theoretical_loss": 5.225069014507933, + "tokens_seen": 55443456 + }, + { + "epoch": 0.02, + "learning_rate": 0.0007944954864593782, + "loss": 2.4684, + "theoretical_loss": 5.224147067593012, + "tokens_seen": 55508992 + }, + { + "epoch": 0.02, + "learning_rate": 0.0007944794383149448, + "loss": 2.6989, + "theoretical_loss": 5.223226512887375, + "tokens_seen": 55574528 + }, + { + "epoch": 0.02, + "learning_rate": 0.0007944633901705116, + "loss": 2.6736, + "theoretical_loss": 5.222307346650636, + "tokens_seen": 55640064 + }, + { + "epoch": 0.02, + "objective/train/docs_used": 27758, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 2.528623580932617, + "objective/train/theoretical_loss": 5.221389565156855, + "objective/train/tokens_used": 76165600, + "theoretical_loss": 5.221389565156855, + "tokens_seen": 55705600 + }, + { + "epoch": 0.02, + "learning_rate": 0.0007944473420260782, + "loss": 2.5681, + "theoretical_loss": 5.221389565156855, + "tokens_seen": 55705600 + }, + { + "epoch": 0.02, + "learning_rate": 0.000794431293881645, + "loss": 2.8749, + "theoretical_loss": 5.2204731646944555, + "tokens_seen": 55771136 + }, + { + "epoch": 0.02, + "learning_rate": 0.0007944152457372116, + "loss": 2.5533, + "theoretical_loss": 5.219558141566156, + "tokens_seen": 55836672 + }, + { + "epoch": 0.02, + "learning_rate": 0.0007943991975927784, + "loss": 2.6242, + "theoretical_loss": 5.218644492088899, + "tokens_seen": 55902208 + }, + { + "epoch": 0.02, + "learning_rate": 0.000794383149448345, + "loss": 2.5736, + "theoretical_loss": 5.217732212593778, + "tokens_seen": 55967744 + }, + { + "epoch": 0.02, + "learning_rate": 0.0007943671013039118, + "loss": 2.6834, + "theoretical_loss": 5.216821299425968, + "tokens_seen": 56033280 + }, + { + "epoch": 0.02, + "learning_rate": 0.0007943510531594784, + "loss": 2.6184, + "theoretical_loss": 5.215911748944652, + "tokens_seen": 56098816 + }, + { + "epoch": 0.02, + "learning_rate": 0.0007943350050150452, + "loss": 2.6754, + "theoretical_loss": 5.215003557522959, + "tokens_seen": 56164352 + }, + { + "epoch": 0.02, + "learning_rate": 0.0007943189568706118, + "loss": 2.6391, + "theoretical_loss": 5.21409672154788, + "tokens_seen": 56229888 + }, + { + "epoch": 0.02, + "learning_rate": 0.0007943029087261786, + "loss": 2.4654, + "theoretical_loss": 5.213191237420215, + "tokens_seen": 56295424 + }, + { + "epoch": 0.02, + "learning_rate": 0.0007942868605817452, + "loss": 2.6377, + "theoretical_loss": 5.212287101554496, + "tokens_seen": 56360960 + }, + { + "epoch": 0.02, + "learning_rate": 0.000794270812437312, + "loss": 2.4726, + "theoretical_loss": 5.211384310378916, + "tokens_seen": 56426496 + }, + { + "epoch": 0.02, + "learning_rate": 0.0007942547642928786, + "loss": 2.6869, + "theoretical_loss": 5.210482860335269, + "tokens_seen": 56492032 + }, + { + "epoch": 0.02, + "learning_rate": 0.0007942387161484454, + "loss": 2.5533, + "theoretical_loss": 5.2095827478788745, + "tokens_seen": 56557568 + }, + { + "epoch": 0.02, + "learning_rate": 0.000794222668004012, + "loss": 2.6459, + "theoretical_loss": 5.208683969478518, + "tokens_seen": 56623104 + }, + { + "epoch": 0.02, + "learning_rate": 0.0007942066198595788, + "loss": 2.5763, + "theoretical_loss": 5.2077865216163755, + "tokens_seen": 56688640 + }, + { + "epoch": 0.02, + "learning_rate": 0.0007941905717151455, + "loss": 2.4978, + "theoretical_loss": 5.206890400787959, + "tokens_seen": 56754176 + }, + { + "epoch": 0.02, + "learning_rate": 0.0007941745235707122, + "loss": 2.5383, + "theoretical_loss": 5.205995603502039, + "tokens_seen": 56819712 + }, + { + "epoch": 0.02, + "learning_rate": 0.0007941584754262789, + "loss": 2.5882, + "theoretical_loss": 5.205102126280586, + "tokens_seen": 56885248 + }, + { + "epoch": 0.02, + "learning_rate": 0.0007941424272818456, + "loss": 2.5623, + "theoretical_loss": 5.204209965658704, + "tokens_seen": 56950784 + }, + { + "epoch": 0.02, + "learning_rate": 0.0007941263791374123, + "loss": 2.5851, + "theoretical_loss": 5.203319118184565, + "tokens_seen": 57016320 + }, + { + "epoch": 0.02, + "learning_rate": 0.000794110330992979, + "loss": 2.5279, + "theoretical_loss": 5.202429580419343, + "tokens_seen": 57081856 + }, + { + "epoch": 0.02, + "learning_rate": 0.0007940942828485457, + "loss": 2.5693, + "theoretical_loss": 5.201541348937157, + "tokens_seen": 57147392 + }, + { + "epoch": 0.02, + "learning_rate": 0.0007940782347041124, + "loss": 2.4777, + "theoretical_loss": 5.200654420324998, + "tokens_seen": 57212928 + }, + { + "epoch": 0.02, + "learning_rate": 0.0007940621865596791, + "loss": 2.4306, + "theoretical_loss": 5.199768791182677, + "tokens_seen": 57278464 + }, + { + "epoch": 0.02, + "learning_rate": 0.0007940461384152458, + "loss": 2.6016, + "theoretical_loss": 5.198884458122748, + "tokens_seen": 57344000 + }, + { + "epoch": 0.02, + "learning_rate": 0.0007940300902708125, + "loss": 2.6248, + "theoretical_loss": 5.198001417770461, + "tokens_seen": 57409536 + }, + { + "epoch": 0.02, + "learning_rate": 0.0007940140421263792, + "loss": 2.5662, + "theoretical_loss": 5.1971196667636885, + "tokens_seen": 57475072 + }, + { + "epoch": 0.02, + "learning_rate": 0.0007939979939819459, + "loss": 2.5387, + "theoretical_loss": 5.196239201752871, + "tokens_seen": 57540608 + }, + { + "epoch": 0.02, + "learning_rate": 0.0007939819458375126, + "loss": 2.5532, + "theoretical_loss": 5.19536001940095, + "tokens_seen": 57606144 + }, + { + "epoch": 0.02, + "learning_rate": 0.0007939658976930794, + "loss": 2.4779, + "theoretical_loss": 5.1944821163833135, + "tokens_seen": 57671680 + }, + { + "epoch": 0.02, + "learning_rate": 0.000793949849548646, + "loss": 2.5137, + "theoretical_loss": 5.19360548938773, + "tokens_seen": 57737216 + }, + { + "epoch": 0.02, + "learning_rate": 0.0007939338014042128, + "loss": 2.492, + "theoretical_loss": 5.192730135114292, + "tokens_seen": 57802752 + }, + { + "epoch": 0.02, + "learning_rate": 0.0007939177532597794, + "loss": 2.6245, + "theoretical_loss": 5.191856050275353, + "tokens_seen": 57868288 + }, + { + "epoch": 0.02, + "learning_rate": 0.0007939017051153462, + "loss": 2.6765, + "theoretical_loss": 5.190983231595473, + "tokens_seen": 57933824 + }, + { + "epoch": 0.02, + "learning_rate": 0.0007938856569709128, + "loss": 2.4179, + "theoretical_loss": 5.1901116758113535, + "tokens_seen": 57999360 + }, + { + "epoch": 0.02, + "learning_rate": 0.0007938696088264795, + "loss": 2.42, + "theoretical_loss": 5.189241379671782, + "tokens_seen": 58064896 + }, + { + "epoch": 0.02, + "learning_rate": 0.0007938535606820462, + "loss": 2.6298, + "theoretical_loss": 5.188372339937576, + "tokens_seen": 58130432 + }, + { + "epoch": 0.02, + "learning_rate": 0.0007938375125376129, + "loss": 2.61, + "theoretical_loss": 5.187504553381521, + "tokens_seen": 58195968 + }, + { + "epoch": 0.02, + "learning_rate": 0.0007938214643931796, + "loss": 2.3955, + "theoretical_loss": 5.1866380167883115, + "tokens_seen": 58261504 + }, + { + "epoch": 0.02, + "learning_rate": 0.0007938054162487463, + "loss": 2.5699, + "theoretical_loss": 5.1857727269545, + "tokens_seen": 58327040 + }, + { + "epoch": 0.02, + "learning_rate": 0.000793789368104313, + "loss": 2.5627, + "theoretical_loss": 5.184908680688437, + "tokens_seen": 58392576 + }, + { + "epoch": 0.02, + "learning_rate": 0.0007937733199598797, + "loss": 2.444, + "theoretical_loss": 5.18404587481021, + "tokens_seen": 58458112 + }, + { + "epoch": 0.02, + "learning_rate": 0.0007937572718154463, + "loss": 2.4058, + "theoretical_loss": 5.183184306151595, + "tokens_seen": 58523648 + }, + { + "epoch": 0.02, + "learning_rate": 0.0007937412236710131, + "loss": 2.4112, + "theoretical_loss": 5.182323971555995, + "tokens_seen": 58589184 + }, + { + "epoch": 0.02, + "learning_rate": 0.0007937251755265797, + "loss": 2.615, + "theoretical_loss": 5.181464867878384, + "tokens_seen": 58654720 + }, + { + "epoch": 0.02, + "learning_rate": 0.0007937091273821465, + "loss": 2.7111, + "theoretical_loss": 5.1806069919852575, + "tokens_seen": 58720256 + }, + { + "epoch": 0.02, + "learning_rate": 0.0007936930792377131, + "loss": 2.467, + "theoretical_loss": 5.17975034075457, + "tokens_seen": 58785792 + }, + { + "epoch": 0.02, + "learning_rate": 0.0007936770310932799, + "loss": 2.5219, + "theoretical_loss": 5.1788949110756874, + "tokens_seen": 58851328 + }, + { + "epoch": 0.02, + "learning_rate": 0.0007936609829488465, + "loss": 2.4198, + "theoretical_loss": 5.178040699849327, + "tokens_seen": 58916864 + }, + { + "epoch": 0.02, + "objective/train/docs_used": 29071, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 2.280916213989258, + "objective/train/theoretical_loss": 5.1771877039875065, + "objective/train/tokens_used": 79442400, + "theoretical_loss": 5.1771877039875065, + "tokens_seen": 58982400 + }, + { + "epoch": 0.02, + "learning_rate": 0.0007936449348044133, + "loss": 2.3825, + "theoretical_loss": 5.1771877039875065, + "tokens_seen": 58982400 + }, + { + "epoch": 0.02, + "learning_rate": 0.0007936288866599799, + "loss": 2.4562, + "theoretical_loss": 5.176335920413489, + "tokens_seen": 59047936 + }, + { + "epoch": 0.02, + "learning_rate": 0.0007936128385155467, + "loss": 2.5607, + "theoretical_loss": 5.175485346061732, + "tokens_seen": 59113472 + }, + { + "epoch": 0.02, + "learning_rate": 0.0007935967903711133, + "loss": 2.4495, + "theoretical_loss": 5.174635977877831, + "tokens_seen": 59179008 + }, + { + "epoch": 0.02, + "learning_rate": 0.0007935807422266801, + "loss": 2.5447, + "theoretical_loss": 5.173787812818468, + "tokens_seen": 59244544 + }, + { + "epoch": 0.02, + "learning_rate": 0.0007935646940822467, + "loss": 2.4884, + "theoretical_loss": 5.172940847851359, + "tokens_seen": 59310080 + }, + { + "epoch": 0.02, + "learning_rate": 0.0007935486459378135, + "loss": 2.5016, + "theoretical_loss": 5.172095079955204, + "tokens_seen": 59375616 + }, + { + "epoch": 0.02, + "learning_rate": 0.0007935325977933801, + "loss": 2.315, + "theoretical_loss": 5.171250506119634, + "tokens_seen": 59441152 + }, + { + "epoch": 0.02, + "learning_rate": 0.0007935165496489469, + "loss": 2.613, + "theoretical_loss": 5.170407123345155, + "tokens_seen": 59506688 + }, + { + "epoch": 0.02, + "learning_rate": 0.0007935005015045135, + "loss": 2.444, + "theoretical_loss": 5.169564928643103, + "tokens_seen": 59572224 + }, + { + "epoch": 0.02, + "learning_rate": 0.0007934844533600803, + "loss": 2.4865, + "theoretical_loss": 5.168723919035591, + "tokens_seen": 59637760 + }, + { + "epoch": 0.02, + "learning_rate": 0.0007934684052156469, + "loss": 2.338, + "theoretical_loss": 5.167884091555456, + "tokens_seen": 59703296 + }, + { + "epoch": 0.02, + "learning_rate": 0.0007934523570712137, + "loss": 2.5008, + "theoretical_loss": 5.167045443246213, + "tokens_seen": 59768832 + }, + { + "epoch": 0.02, + "learning_rate": 0.0007934363089267803, + "loss": 2.4123, + "theoretical_loss": 5.166207971162001, + "tokens_seen": 59834368 + }, + { + "epoch": 0.02, + "learning_rate": 0.0007934202607823471, + "loss": 2.4395, + "theoretical_loss": 5.1653716723675345, + "tokens_seen": 59899904 + }, + { + "epoch": 0.02, + "learning_rate": 0.0007934042126379137, + "loss": 2.5566, + "theoretical_loss": 5.164536543938056, + "tokens_seen": 59965440 + }, + { + "epoch": 0.02, + "learning_rate": 0.0007933881644934805, + "loss": 2.3826, + "theoretical_loss": 5.163702582959283, + "tokens_seen": 60030976 + }, + { + "epoch": 0.02, + "learning_rate": 0.0007933721163490471, + "loss": 2.6312, + "theoretical_loss": 5.162869786527363, + "tokens_seen": 60096512 + }, + { + "epoch": 0.02, + "learning_rate": 0.0007933560682046139, + "loss": 2.5485, + "theoretical_loss": 5.1620381517488205, + "tokens_seen": 60162048 + }, + { + "epoch": 0.02, + "learning_rate": 0.0007933400200601805, + "loss": 2.5121, + "theoretical_loss": 5.161207675740515, + "tokens_seen": 60227584 + }, + { + "epoch": 0.02, + "learning_rate": 0.0007933239719157473, + "loss": 2.4624, + "theoretical_loss": 5.160378355629584, + "tokens_seen": 60293120 + }, + { + "epoch": 0.02, + "learning_rate": 0.0007933079237713139, + "loss": 2.4698, + "theoretical_loss": 5.159550188553405, + "tokens_seen": 60358656 + }, + { + "epoch": 0.02, + "learning_rate": 0.0007932918756268807, + "loss": 2.5012, + "theoretical_loss": 5.15872317165954, + "tokens_seen": 60424192 + }, + { + "epoch": 0.02, + "learning_rate": 0.0007932758274824473, + "loss": 2.464, + "theoretical_loss": 5.1578973021056935, + "tokens_seen": 60489728 + }, + { + "epoch": 0.02, + "learning_rate": 0.0007932597793380141, + "loss": 2.4249, + "theoretical_loss": 5.157072577059662, + "tokens_seen": 60555264 + }, + { + "epoch": 0.02, + "learning_rate": 0.0007932437311935807, + "loss": 2.5763, + "theoretical_loss": 5.15624899369929, + "tokens_seen": 60620800 + }, + { + "epoch": 0.02, + "learning_rate": 0.0007932276830491475, + "loss": 2.4868, + "theoretical_loss": 5.155426549212422, + "tokens_seen": 60686336 + }, + { + "epoch": 0.02, + "learning_rate": 0.0007932116349047141, + "loss": 2.4949, + "theoretical_loss": 5.154605240796855, + "tokens_seen": 60751872 + }, + { + "epoch": 0.02, + "learning_rate": 0.0007931955867602809, + "loss": 2.4117, + "theoretical_loss": 5.153785065660296, + "tokens_seen": 60817408 + }, + { + "epoch": 0.02, + "learning_rate": 0.0007931795386158475, + "loss": 2.5438, + "theoretical_loss": 5.152966021020315, + "tokens_seen": 60882944 + }, + { + "epoch": 0.02, + "learning_rate": 0.0007931634904714143, + "loss": 2.4582, + "theoretical_loss": 5.152148104104297, + "tokens_seen": 60948480 + }, + { + "epoch": 0.02, + "learning_rate": 0.0007931474423269809, + "loss": 2.3854, + "theoretical_loss": 5.1513313121494, + "tokens_seen": 61014016 + }, + { + "epoch": 0.02, + "learning_rate": 0.0007931313941825477, + "loss": 2.4946, + "theoretical_loss": 5.15051564240251, + "tokens_seen": 61079552 + }, + { + "epoch": 0.02, + "learning_rate": 0.0007931153460381143, + "loss": 2.5024, + "theoretical_loss": 5.149701092120194, + "tokens_seen": 61145088 + }, + { + "epoch": 0.02, + "learning_rate": 0.0007930992978936811, + "loss": 2.4005, + "theoretical_loss": 5.148887658568658, + "tokens_seen": 61210624 + }, + { + "epoch": 0.02, + "learning_rate": 0.0007930832497492477, + "loss": 2.5115, + "theoretical_loss": 5.148075339023699, + "tokens_seen": 61276160 + }, + { + "epoch": 0.02, + "learning_rate": 0.0007930672016048145, + "loss": 2.4502, + "theoretical_loss": 5.147264130770671, + "tokens_seen": 61341696 + }, + { + "epoch": 0.02, + "learning_rate": 0.0007930511534603811, + "loss": 2.4304, + "theoretical_loss": 5.146454031104426, + "tokens_seen": 61407232 + }, + { + "epoch": 0.02, + "learning_rate": 0.0007930351053159479, + "loss": 2.5771, + "theoretical_loss": 5.1456450373292855, + "tokens_seen": 61472768 + }, + { + "epoch": 0.02, + "learning_rate": 0.0007930190571715145, + "loss": 2.4737, + "theoretical_loss": 5.144837146758987, + "tokens_seen": 61538304 + }, + { + "epoch": 0.02, + "learning_rate": 0.0007930030090270813, + "loss": 2.5004, + "theoretical_loss": 5.144030356716646, + "tokens_seen": 61603840 + }, + { + "epoch": 0.02, + "learning_rate": 0.0007929869608826479, + "loss": 2.4925, + "theoretical_loss": 5.1432246645347135, + "tokens_seen": 61669376 + }, + { + "epoch": 0.02, + "learning_rate": 0.0007929709127382147, + "loss": 2.365, + "theoretical_loss": 5.14242006755493, + "tokens_seen": 61734912 + }, + { + "epoch": 0.02, + "learning_rate": 0.0007929548645937813, + "loss": 2.429, + "theoretical_loss": 5.141616563128288, + "tokens_seen": 61800448 + }, + { + "epoch": 0.02, + "learning_rate": 0.0007929388164493481, + "loss": 2.4815, + "theoretical_loss": 5.140814148614985, + "tokens_seen": 61865984 + }, + { + "epoch": 0.02, + "learning_rate": 0.0007929227683049147, + "loss": 2.448, + "theoretical_loss": 5.1400128213843885, + "tokens_seen": 61931520 + }, + { + "epoch": 0.02, + "learning_rate": 0.0007929067201604815, + "loss": 2.5698, + "theoretical_loss": 5.139212578814986, + "tokens_seen": 61997056 + }, + { + "epoch": 0.02, + "learning_rate": 0.0007928906720160481, + "loss": 2.459, + "theoretical_loss": 5.13841341829435, + "tokens_seen": 62062592 + }, + { + "epoch": 0.02, + "learning_rate": 0.0007928746238716149, + "loss": 2.3611, + "theoretical_loss": 5.137615337219098, + "tokens_seen": 62128128 + }, + { + "epoch": 0.02, + "learning_rate": 0.0007928585757271815, + "loss": 2.4881, + "theoretical_loss": 5.136818332994844, + "tokens_seen": 62193664 + }, + { + "epoch": 0.02, + "objective/train/docs_used": 30276, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 2.6835126876831055, + "objective/train/theoretical_loss": 5.136022403036167, + "objective/train/tokens_used": 82719200, + "theoretical_loss": 5.136022403036167, + "tokens_seen": 62259200 + }, + { + "epoch": 0.02, + "learning_rate": 0.0007928425275827483, + "loss": 2.5467, + "theoretical_loss": 5.136022403036167, + "tokens_seen": 62259200 + }, + { + "epoch": 0.02, + "learning_rate": 0.000792826479438315, + "loss": 2.369, + "theoretical_loss": 5.135227544766565, + "tokens_seen": 62324736 + }, + { + "epoch": 0.02, + "learning_rate": 0.0007928104312938817, + "loss": 2.5265, + "theoretical_loss": 5.134433755618414, + "tokens_seen": 62390272 + }, + { + "epoch": 0.02, + "learning_rate": 0.0007927943831494484, + "loss": 2.2945, + "theoretical_loss": 5.1336410330329345, + "tokens_seen": 62455808 + }, + { + "epoch": 0.02, + "learning_rate": 0.0007927783350050151, + "loss": 2.4148, + "theoretical_loss": 5.132849374460145, + "tokens_seen": 62521344 + }, + { + "epoch": 0.02, + "learning_rate": 0.0007927622868605819, + "loss": 2.4408, + "theoretical_loss": 5.132058777358827, + "tokens_seen": 62586880 + }, + { + "epoch": 0.02, + "learning_rate": 0.0007927462387161485, + "loss": 2.3712, + "theoretical_loss": 5.1312692391964845, + "tokens_seen": 62652416 + }, + { + "epoch": 0.02, + "learning_rate": 0.0007927301905717153, + "loss": 2.3667, + "theoretical_loss": 5.1304807574493, + "tokens_seen": 62717952 + }, + { + "epoch": 0.02, + "learning_rate": 0.0007927141424272819, + "loss": 2.4285, + "theoretical_loss": 5.129693329602108, + "tokens_seen": 62783488 + }, + { + "epoch": 0.02, + "learning_rate": 0.0007926980942828487, + "loss": 2.3586, + "theoretical_loss": 5.128906953148342, + "tokens_seen": 62849024 + }, + { + "epoch": 0.02, + "learning_rate": 0.0007926820461384153, + "loss": 2.4575, + "theoretical_loss": 5.128121625590007, + "tokens_seen": 62914560 + }, + { + "epoch": 0.02, + "learning_rate": 0.0007926659979939821, + "loss": 2.4842, + "theoretical_loss": 5.127337344437633, + "tokens_seen": 62980096 + }, + { + "epoch": 0.02, + "learning_rate": 0.0007926499498495487, + "loss": 2.5172, + "theoretical_loss": 5.126554107210245, + "tokens_seen": 63045632 + }, + { + "epoch": 0.02, + "learning_rate": 0.0007926339017051155, + "loss": 2.4282, + "theoretical_loss": 5.125771911435321, + "tokens_seen": 63111168 + }, + { + "epoch": 0.02, + "learning_rate": 0.0007926178535606821, + "loss": 2.3986, + "theoretical_loss": 5.124990754648754, + "tokens_seen": 63176704 + }, + { + "epoch": 0.02, + "learning_rate": 0.0007926018054162489, + "loss": 2.3867, + "theoretical_loss": 5.124210634394814, + "tokens_seen": 63242240 + }, + { + "epoch": 0.02, + "learning_rate": 0.0007925857572718155, + "loss": 2.3106, + "theoretical_loss": 5.1234315482261135, + "tokens_seen": 63307776 + }, + { + "epoch": 0.02, + "learning_rate": 0.0007925697091273823, + "loss": 2.3384, + "theoretical_loss": 5.122653493703571, + "tokens_seen": 63373312 + }, + { + "epoch": 0.02, + "learning_rate": 0.0007925536609829489, + "loss": 2.4279, + "theoretical_loss": 5.121876468396373, + "tokens_seen": 63438848 + }, + { + "epoch": 0.02, + "learning_rate": 0.0007925376128385157, + "loss": 2.413, + "theoretical_loss": 5.121100469881933, + "tokens_seen": 63504384 + }, + { + "epoch": 0.02, + "learning_rate": 0.0007925215646940823, + "loss": 2.3048, + "theoretical_loss": 5.120325495745863, + "tokens_seen": 63569920 + }, + { + "epoch": 0.02, + "learning_rate": 0.0007925055165496491, + "loss": 2.329, + "theoretical_loss": 5.119551543581933, + "tokens_seen": 63635456 + }, + { + "epoch": 0.02, + "learning_rate": 0.0007924894684052157, + "loss": 2.4087, + "theoretical_loss": 5.118778610992038, + "tokens_seen": 63700992 + }, + { + "epoch": 0.02, + "learning_rate": 0.0007924734202607824, + "loss": 2.4179, + "theoretical_loss": 5.118006695586154, + "tokens_seen": 63766528 + }, + { + "epoch": 0.02, + "learning_rate": 0.0007924573721163491, + "loss": 2.3683, + "theoretical_loss": 5.1172357949823155, + "tokens_seen": 63832064 + }, + { + "epoch": 0.02, + "learning_rate": 0.0007924413239719158, + "loss": 2.4536, + "theoretical_loss": 5.116465906806567, + "tokens_seen": 63897600 + }, + { + "epoch": 0.02, + "learning_rate": 0.0007924252758274825, + "loss": 2.4362, + "theoretical_loss": 5.11569702869294, + "tokens_seen": 63963136 + }, + { + "epoch": 0.02, + "learning_rate": 0.0007924092276830492, + "loss": 2.2716, + "theoretical_loss": 5.114929158283407, + "tokens_seen": 64028672 + }, + { + "epoch": 0.02, + "learning_rate": 0.0007923931795386159, + "loss": 2.3832, + "theoretical_loss": 5.114162293227853, + "tokens_seen": 64094208 + }, + { + "epoch": 0.02, + "learning_rate": 0.0007923771313941826, + "loss": 2.4257, + "theoretical_loss": 5.113396431184043, + "tokens_seen": 64159744 + }, + { + "epoch": 0.02, + "learning_rate": 0.0007923610832497493, + "loss": 2.3293, + "theoretical_loss": 5.112631569817578, + "tokens_seen": 64225280 + }, + { + "epoch": 0.02, + "learning_rate": 0.000792345035105316, + "loss": 2.3443, + "theoretical_loss": 5.111867706801872, + "tokens_seen": 64290816 + }, + { + "epoch": 0.02, + "learning_rate": 0.0007923289869608826, + "loss": 2.3122, + "theoretical_loss": 5.111104839818113, + "tokens_seen": 64356352 + }, + { + "epoch": 0.02, + "learning_rate": 0.0007923129388164494, + "loss": 2.4295, + "theoretical_loss": 5.1103429665552245, + "tokens_seen": 64421888 + }, + { + "epoch": 0.02, + "learning_rate": 0.000792296890672016, + "loss": 2.4573, + "theoretical_loss": 5.109582084709841, + "tokens_seen": 64487424 + }, + { + "epoch": 0.02, + "learning_rate": 0.0007922808425275828, + "loss": 2.4579, + "theoretical_loss": 5.108822191986269, + "tokens_seen": 64552960 + }, + { + "epoch": 0.02, + "learning_rate": 0.0007922647943831494, + "loss": 2.2775, + "theoretical_loss": 5.108063286096453, + "tokens_seen": 64618496 + }, + { + "epoch": 0.02, + "learning_rate": 0.0007922487462387162, + "loss": 2.365, + "theoretical_loss": 5.107305364759943, + "tokens_seen": 64684032 + }, + { + "epoch": 0.02, + "learning_rate": 0.0007922326980942828, + "loss": 2.3217, + "theoretical_loss": 5.106548425703869, + "tokens_seen": 64749568 + }, + { + "epoch": 0.02, + "learning_rate": 0.0007922166499498496, + "loss": 2.4907, + "theoretical_loss": 5.1057924666628915, + "tokens_seen": 64815104 + }, + { + "epoch": 0.02, + "learning_rate": 0.0007922006018054162, + "loss": 2.415, + "theoretical_loss": 5.105037485379189, + "tokens_seen": 64880640 + }, + { + "epoch": 0.02, + "learning_rate": 0.000792184553660983, + "loss": 2.4981, + "theoretical_loss": 5.104283479602407, + "tokens_seen": 64946176 + }, + { + "epoch": 0.02, + "learning_rate": 0.0007921685055165496, + "loss": 2.4132, + "theoretical_loss": 5.103530447089642, + "tokens_seen": 65011712 + }, + { + "epoch": 0.02, + "learning_rate": 0.0007921524573721164, + "loss": 2.3904, + "theoretical_loss": 5.102778385605397, + "tokens_seen": 65077248 + }, + { + "epoch": 0.02, + "learning_rate": 0.000792136409227683, + "loss": 2.4111, + "theoretical_loss": 5.102027292921557, + "tokens_seen": 65142784 + }, + { + "epoch": 0.02, + "learning_rate": 0.0007921203610832498, + "loss": 2.4377, + "theoretical_loss": 5.101277166817351, + "tokens_seen": 65208320 + }, + { + "epoch": 0.02, + "learning_rate": 0.0007921043129388164, + "loss": 2.2709, + "theoretical_loss": 5.100528005079328, + "tokens_seen": 65273856 + }, + { + "epoch": 0.02, + "learning_rate": 0.0007920882647943832, + "loss": 2.3739, + "theoretical_loss": 5.099779805501318, + "tokens_seen": 65339392 + }, + { + "epoch": 0.02, + "learning_rate": 0.0007920722166499498, + "loss": 2.3937, + "theoretical_loss": 5.099032565884409, + "tokens_seen": 65404928 + }, + { + "epoch": 0.02, + "learning_rate": 0.0007920561685055166, + "loss": 2.4467, + "theoretical_loss": 5.098286284036908, + "tokens_seen": 65470464 + }, + { + "epoch": 0.02, + "objective/train/docs_used": 31526, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 2.5525588989257812, + "objective/train/theoretical_loss": 5.097540957774312, + "objective/train/tokens_used": 85996000, + "theoretical_loss": 5.097540957774312, + "tokens_seen": 65536000 + }, + { + "epoch": 0.02, + "learning_rate": 0.0007920401203610832, + "loss": 2.4639, + "theoretical_loss": 5.097540957774312, + "tokens_seen": 65536000 + }, + { + "epoch": 0.02, + "learning_rate": 0.00079202407221665, + "loss": 2.456, + "theoretical_loss": 5.096796584919282, + "tokens_seen": 65601536 + }, + { + "epoch": 0.02, + "learning_rate": 0.0007920080240722166, + "loss": 2.4629, + "theoretical_loss": 5.096053163301608, + "tokens_seen": 65667072 + }, + { + "epoch": 0.02, + "learning_rate": 0.0007919919759277834, + "loss": 2.5165, + "theoretical_loss": 5.095310690758177, + "tokens_seen": 65732608 + }, + { + "epoch": 0.02, + "learning_rate": 0.00079197592778335, + "loss": 2.3805, + "theoretical_loss": 5.094569165132949, + "tokens_seen": 65798144 + }, + { + "epoch": 0.02, + "learning_rate": 0.0007919598796389168, + "loss": 2.3113, + "theoretical_loss": 5.09382858427692, + "tokens_seen": 65863680 + }, + { + "epoch": 0.02, + "learning_rate": 0.0007919438314944834, + "loss": 2.428, + "theoretical_loss": 5.0930889460481, + "tokens_seen": 65929216 + }, + { + "epoch": 0.02, + "learning_rate": 0.0007919277833500502, + "loss": 2.3819, + "theoretical_loss": 5.092350248311471, + "tokens_seen": 65994752 + }, + { + "epoch": 0.02, + "learning_rate": 0.0007919117352056168, + "loss": 2.3552, + "theoretical_loss": 5.091612488938971, + "tokens_seen": 66060288 + }, + { + "epoch": 0.02, + "learning_rate": 0.0007918956870611836, + "loss": 2.3424, + "theoretical_loss": 5.090875665809454, + "tokens_seen": 66125824 + }, + { + "epoch": 0.02, + "learning_rate": 0.0007918796389167502, + "loss": 2.2872, + "theoretical_loss": 5.09013977680867, + "tokens_seen": 66191360 + }, + { + "epoch": 0.02, + "learning_rate": 0.000791863590772317, + "loss": 2.3784, + "theoretical_loss": 5.089404819829225, + "tokens_seen": 66256896 + }, + { + "epoch": 0.02, + "learning_rate": 0.0007918475426278836, + "loss": 2.3865, + "theoretical_loss": 5.088670792770563, + "tokens_seen": 66322432 + }, + { + "epoch": 0.02, + "learning_rate": 0.0007918314944834504, + "loss": 2.3318, + "theoretical_loss": 5.087937693538927, + "tokens_seen": 66387968 + }, + { + "epoch": 0.02, + "learning_rate": 0.000791815446339017, + "loss": 2.337, + "theoretical_loss": 5.0872055200473385, + "tokens_seen": 66453504 + }, + { + "epoch": 0.02, + "learning_rate": 0.0007917993981945838, + "loss": 2.3208, + "theoretical_loss": 5.0864742702155645, + "tokens_seen": 66519040 + }, + { + "epoch": 0.02, + "learning_rate": 0.0007917833500501504, + "loss": 2.4569, + "theoretical_loss": 5.0857439419700885, + "tokens_seen": 66584576 + }, + { + "epoch": 0.02, + "learning_rate": 0.0007917673019057172, + "loss": 2.3318, + "theoretical_loss": 5.085014533244088, + "tokens_seen": 66650112 + }, + { + "epoch": 0.02, + "learning_rate": 0.0007917512537612838, + "loss": 2.3326, + "theoretical_loss": 5.084286041977396, + "tokens_seen": 66715648 + }, + { + "epoch": 0.02, + "learning_rate": 0.0007917352056168506, + "loss": 2.3528, + "theoretical_loss": 5.083558466116487, + "tokens_seen": 66781184 + }, + { + "epoch": 0.02, + "learning_rate": 0.0007917191574724172, + "loss": 2.2131, + "theoretical_loss": 5.082831803614433, + "tokens_seen": 66846720 + }, + { + "epoch": 0.02, + "learning_rate": 0.000791703109327984, + "loss": 2.3724, + "theoretical_loss": 5.082106052430891, + "tokens_seen": 66912256 + }, + { + "epoch": 0.02, + "learning_rate": 0.0007916870611835506, + "loss": 2.3096, + "theoretical_loss": 5.081381210532063, + "tokens_seen": 66977792 + }, + { + "epoch": 0.02, + "learning_rate": 0.0007916710130391174, + "loss": 2.441, + "theoretical_loss": 5.08065727589068, + "tokens_seen": 67043328 + }, + { + "epoch": 0.02, + "learning_rate": 0.000791654964894684, + "loss": 2.3414, + "theoretical_loss": 5.079934246485963, + "tokens_seen": 67108864 + }, + { + "epoch": 0.02, + "learning_rate": 0.0007916389167502508, + "loss": 2.5294, + "theoretical_loss": 5.079212120303606, + "tokens_seen": 67174400 + }, + { + "epoch": 0.02, + "learning_rate": 0.0007916228686058174, + "loss": 2.3024, + "theoretical_loss": 5.078490895335744, + "tokens_seen": 67239936 + }, + { + "epoch": 0.02, + "learning_rate": 0.0007916068204613842, + "loss": 2.3816, + "theoretical_loss": 5.077770569580924, + "tokens_seen": 67305472 + }, + { + "epoch": 0.02, + "learning_rate": 0.000791590772316951, + "loss": 2.2574, + "theoretical_loss": 5.077051141044086, + "tokens_seen": 67371008 + }, + { + "epoch": 0.02, + "learning_rate": 0.0007915747241725176, + "loss": 2.2999, + "theoretical_loss": 5.076332607736527, + "tokens_seen": 67436544 + }, + { + "epoch": 0.02, + "learning_rate": 0.0007915586760280844, + "loss": 2.4052, + "theoretical_loss": 5.0756149676758815, + "tokens_seen": 67502080 + }, + { + "epoch": 0.02, + "learning_rate": 0.000791542627883651, + "loss": 2.5148, + "theoretical_loss": 5.0748982188860925, + "tokens_seen": 67567616 + }, + { + "epoch": 0.02, + "learning_rate": 0.0007915265797392178, + "loss": 2.3924, + "theoretical_loss": 5.074182359397387, + "tokens_seen": 67633152 + }, + { + "epoch": 0.02, + "learning_rate": 0.0007915105315947844, + "loss": 2.2304, + "theoretical_loss": 5.073467387246247, + "tokens_seen": 67698688 + }, + { + "epoch": 0.02, + "learning_rate": 0.0007914944834503512, + "loss": 2.3326, + "theoretical_loss": 5.072753300475385, + "tokens_seen": 67764224 + }, + { + "epoch": 0.02, + "learning_rate": 0.0007914784353059178, + "loss": 2.3414, + "theoretical_loss": 5.072040097133718, + "tokens_seen": 67829760 + }, + { + "epoch": 0.02, + "learning_rate": 0.0007914623871614846, + "loss": 2.2544, + "theoretical_loss": 5.0713277752763455, + "tokens_seen": 67895296 + }, + { + "epoch": 0.02, + "learning_rate": 0.0007914463390170512, + "loss": 2.2938, + "theoretical_loss": 5.070616332964518, + "tokens_seen": 67960832 + }, + { + "epoch": 0.02, + "learning_rate": 0.000791430290872618, + "loss": 2.3502, + "theoretical_loss": 5.069905768265615, + "tokens_seen": 68026368 + }, + { + "epoch": 0.02, + "learning_rate": 0.0007914142427281846, + "loss": 2.2745, + "theoretical_loss": 5.069196079253119, + "tokens_seen": 68091904 + }, + { + "epoch": 0.02, + "learning_rate": 0.0007913981945837514, + "loss": 2.1518, + "theoretical_loss": 5.068487264006592, + "tokens_seen": 68157440 + }, + { + "epoch": 0.02, + "learning_rate": 0.000791382146439318, + "loss": 2.2279, + "theoretical_loss": 5.0677793206116455, + "tokens_seen": 68222976 + }, + { + "epoch": 0.02, + "learning_rate": 0.0007913660982948848, + "loss": 2.3788, + "theoretical_loss": 5.067072247159923, + "tokens_seen": 68288512 + }, + { + "epoch": 0.02, + "learning_rate": 0.0007913500501504514, + "loss": 2.2849, + "theoretical_loss": 5.06636604174907, + "tokens_seen": 68354048 + }, + { + "epoch": 0.02, + "learning_rate": 0.0007913340020060182, + "loss": 2.3645, + "theoretical_loss": 5.065660702482711, + "tokens_seen": 68419584 + }, + { + "epoch": 0.02, + "learning_rate": 0.0007913179538615848, + "loss": 2.2937, + "theoretical_loss": 5.064956227470425, + "tokens_seen": 68485120 + }, + { + "epoch": 0.02, + "learning_rate": 0.0007913019057171516, + "loss": 2.3593, + "theoretical_loss": 5.06425261482772, + "tokens_seen": 68550656 + }, + { + "epoch": 0.02, + "learning_rate": 0.0007912858575727182, + "loss": 2.3938, + "theoretical_loss": 5.063549862676011, + "tokens_seen": 68616192 + }, + { + "epoch": 0.02, + "learning_rate": 0.000791269809428285, + "loss": 2.2554, + "theoretical_loss": 5.062847969142591, + "tokens_seen": 68681728 + }, + { + "epoch": 0.02, + "learning_rate": 0.0007912537612838516, + "loss": 2.3624, + "theoretical_loss": 5.062146932360616, + "tokens_seen": 68747264 + }, + { + "epoch": 0.02, + "objective/train/docs_used": 32629, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 2.248861789703369, + "objective/train/theoretical_loss": 5.061446750469072, + "objective/train/tokens_used": 89272800, + "theoretical_loss": 5.061446750469072, + "tokens_seen": 68812800 + }, + { + "epoch": 0.02, + "learning_rate": 0.0007912377131394184, + "loss": 2.3099, + "theoretical_loss": 5.061446750469072, + "tokens_seen": 68812800 + }, + { + "epoch": 0.02, + "learning_rate": 0.000791221664994985, + "loss": 2.4317, + "theoretical_loss": 5.060747421612758, + "tokens_seen": 68878336 + }, + { + "epoch": 0.02, + "learning_rate": 0.0007912056168505518, + "loss": 2.3972, + "theoretical_loss": 5.060048943942254, + "tokens_seen": 68943872 + }, + { + "epoch": 0.02, + "learning_rate": 0.0007911895687061184, + "loss": 2.2637, + "theoretical_loss": 5.059351315613908, + "tokens_seen": 69009408 + }, + { + "epoch": 0.02, + "learning_rate": 0.0007911735205616852, + "loss": 2.2443, + "theoretical_loss": 5.058654534789804, + "tokens_seen": 69074944 + }, + { + "epoch": 0.02, + "learning_rate": 0.0007911574724172518, + "loss": 2.3448, + "theoretical_loss": 5.057958599637746, + "tokens_seen": 69140480 + }, + { + "epoch": 0.02, + "learning_rate": 0.0007911414242728186, + "loss": 2.3436, + "theoretical_loss": 5.057263508331225, + "tokens_seen": 69206016 + }, + { + "epoch": 0.02, + "learning_rate": 0.0007911253761283852, + "loss": 2.278, + "theoretical_loss": 5.0565692590494065, + "tokens_seen": 69271552 + }, + { + "epoch": 0.02, + "learning_rate": 0.000791109327983952, + "loss": 2.2125, + "theoretical_loss": 5.055875849977101, + "tokens_seen": 69337088 + }, + { + "epoch": 0.02, + "learning_rate": 0.0007910932798395186, + "loss": 2.4201, + "theoretical_loss": 5.055183279304744, + "tokens_seen": 69402624 + }, + { + "epoch": 0.02, + "learning_rate": 0.0007910772316950853, + "loss": 2.331, + "theoretical_loss": 5.0544915452283705, + "tokens_seen": 69468160 + }, + { + "epoch": 0.02, + "learning_rate": 0.000791061183550652, + "loss": 2.2871, + "theoretical_loss": 5.053800645949597, + "tokens_seen": 69533696 + }, + { + "epoch": 0.02, + "learning_rate": 0.0007910451354062187, + "loss": 2.238, + "theoretical_loss": 5.053110579675593, + "tokens_seen": 69599232 + }, + { + "epoch": 0.02, + "learning_rate": 0.0007910290872617854, + "loss": 2.2985, + "theoretical_loss": 5.052421344619064, + "tokens_seen": 69664768 + }, + { + "epoch": 0.02, + "learning_rate": 0.0007910130391173521, + "loss": 2.3492, + "theoretical_loss": 5.051732938998227, + "tokens_seen": 69730304 + }, + { + "epoch": 0.02, + "learning_rate": 0.0007909969909729188, + "loss": 2.2491, + "theoretical_loss": 5.051045361036787, + "tokens_seen": 69795840 + }, + { + "epoch": 0.02, + "learning_rate": 0.0007909809428284855, + "loss": 2.29, + "theoretical_loss": 5.0503586089639185, + "tokens_seen": 69861376 + }, + { + "epoch": 0.02, + "learning_rate": 0.0007909648946840522, + "loss": 2.2723, + "theoretical_loss": 5.04967268101424, + "tokens_seen": 69926912 + }, + { + "epoch": 0.02, + "learning_rate": 0.0007909488465396189, + "loss": 2.2914, + "theoretical_loss": 5.048987575427793, + "tokens_seen": 69992448 + }, + { + "epoch": 0.02, + "learning_rate": 0.0007909327983951855, + "loss": 2.3501, + "theoretical_loss": 5.048303290450023, + "tokens_seen": 70057984 + }, + { + "epoch": 0.02, + "learning_rate": 0.0007909167502507523, + "loss": 2.4285, + "theoretical_loss": 5.047619824331753, + "tokens_seen": 70123520 + }, + { + "epoch": 0.02, + "learning_rate": 0.0007909007021063189, + "loss": 2.2901, + "theoretical_loss": 5.046937175329167, + "tokens_seen": 70189056 + }, + { + "epoch": 0.02, + "learning_rate": 0.0007908846539618857, + "loss": 2.3492, + "theoretical_loss": 5.046255341703786, + "tokens_seen": 70254592 + }, + { + "epoch": 0.02, + "learning_rate": 0.0007908686058174523, + "loss": 2.3145, + "theoretical_loss": 5.045574321722445, + "tokens_seen": 70320128 + }, + { + "epoch": 0.02, + "learning_rate": 0.0007908525576730191, + "loss": 2.2768, + "theoretical_loss": 5.0448941136572785, + "tokens_seen": 70385664 + }, + { + "epoch": 0.02, + "learning_rate": 0.0007908365095285857, + "loss": 2.3515, + "theoretical_loss": 5.044214715785689, + "tokens_seen": 70451200 + }, + { + "epoch": 0.02, + "learning_rate": 0.0007908204613841525, + "loss": 2.2817, + "theoretical_loss": 5.043536126390334, + "tokens_seen": 70516736 + }, + { + "epoch": 0.02, + "learning_rate": 0.0007908044132397191, + "loss": 2.1776, + "theoretical_loss": 5.042858343759107, + "tokens_seen": 70582272 + }, + { + "epoch": 0.02, + "learning_rate": 0.0007907883650952859, + "loss": 2.2317, + "theoretical_loss": 5.0421813661851065, + "tokens_seen": 70647808 + }, + { + "epoch": 0.02, + "learning_rate": 0.0007907723169508525, + "loss": 2.3149, + "theoretical_loss": 5.041505191966625, + "tokens_seen": 70713344 + }, + { + "epoch": 0.02, + "learning_rate": 0.0007907562688064193, + "loss": 2.384, + "theoretical_loss": 5.040829819407123, + "tokens_seen": 70778880 + }, + { + "epoch": 0.02, + "learning_rate": 0.0007907402206619859, + "loss": 2.3183, + "theoretical_loss": 5.040155246815212, + "tokens_seen": 70844416 + }, + { + "epoch": 0.02, + "learning_rate": 0.0007907241725175527, + "loss": 2.4291, + "theoretical_loss": 5.039481472504633, + "tokens_seen": 70909952 + }, + { + "epoch": 0.02, + "learning_rate": 0.0007907081243731193, + "loss": 2.3489, + "theoretical_loss": 5.038808494794232, + "tokens_seen": 70975488 + }, + { + "epoch": 0.02, + "learning_rate": 0.0007906920762286861, + "loss": 2.2079, + "theoretical_loss": 5.038136312007948, + "tokens_seen": 71041024 + }, + { + "epoch": 0.02, + "learning_rate": 0.0007906760280842527, + "loss": 2.2226, + "theoretical_loss": 5.037464922474786, + "tokens_seen": 71106560 + }, + { + "epoch": 0.02, + "learning_rate": 0.0007906599799398195, + "loss": 2.2605, + "theoretical_loss": 5.036794324528802, + "tokens_seen": 71172096 + }, + { + "epoch": 0.02, + "learning_rate": 0.0007906439317953861, + "loss": 2.219, + "theoretical_loss": 5.036124516509076, + "tokens_seen": 71237632 + }, + { + "epoch": 0.02, + "learning_rate": 0.0007906278836509529, + "loss": 2.4072, + "theoretical_loss": 5.035455496759701, + "tokens_seen": 71303168 + }, + { + "epoch": 0.02, + "learning_rate": 0.0007906118355065195, + "loss": 2.3632, + "theoretical_loss": 5.034787263629758, + "tokens_seen": 71368704 + }, + { + "epoch": 0.02, + "learning_rate": 0.0007905957873620863, + "loss": 2.3524, + "theoretical_loss": 5.0341198154732965, + "tokens_seen": 71434240 + }, + { + "epoch": 0.02, + "learning_rate": 0.0007905797392176529, + "loss": 2.3933, + "theoretical_loss": 5.033453150649318, + "tokens_seen": 71499776 + }, + { + "epoch": 0.02, + "learning_rate": 0.0007905636910732197, + "loss": 2.2523, + "theoretical_loss": 5.032787267521752, + "tokens_seen": 71565312 + }, + { + "epoch": 0.02, + "learning_rate": 0.0007905476429287863, + "loss": 2.3117, + "theoretical_loss": 5.032122164459443, + "tokens_seen": 71630848 + }, + { + "epoch": 0.02, + "learning_rate": 0.0007905315947843531, + "loss": 2.3397, + "theoretical_loss": 5.031457839836122, + "tokens_seen": 71696384 + }, + { + "epoch": 0.02, + "learning_rate": 0.0007905155466399197, + "loss": 2.1839, + "theoretical_loss": 5.0307942920304, + "tokens_seen": 71761920 + }, + { + "epoch": 0.02, + "learning_rate": 0.0007904994984954865, + "loss": 2.2821, + "theoretical_loss": 5.030131519425737, + "tokens_seen": 71827456 + }, + { + "epoch": 0.02, + "learning_rate": 0.0007904834503510531, + "loss": 2.36, + "theoretical_loss": 5.02946952041043, + "tokens_seen": 71892992 + }, + { + "epoch": 0.02, + "learning_rate": 0.0007904674022066199, + "loss": 2.3235, + "theoretical_loss": 5.028808293377589, + "tokens_seen": 71958528 + }, + { + "epoch": 0.02, + "learning_rate": 0.0007904513540621865, + "loss": 2.2986, + "theoretical_loss": 5.028147836725127, + "tokens_seen": 72024064 + }, + { + "epoch": 0.02, + "objective/train/docs_used": 33987, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 2.2775955200195312, + "objective/train/theoretical_loss": 5.0274881488557295, + "objective/train/tokens_used": 92549600, + "theoretical_loss": 5.0274881488557295, + "tokens_seen": 72089600 + }, + { + "epoch": 0.02, + "learning_rate": 0.0007904353059177533, + "loss": 2.1512, + "theoretical_loss": 5.0274881488557295, + "tokens_seen": 72089600 + }, + { + "epoch": 0.02, + "learning_rate": 0.0007904192577733199, + "loss": 2.2798, + "theoretical_loss": 5.026829228176848, + "tokens_seen": 72155136 + }, + { + "epoch": 0.02, + "learning_rate": 0.0007904032096288867, + "loss": 2.2989, + "theoretical_loss": 5.026171073100672, + "tokens_seen": 72220672 + }, + { + "epoch": 0.02, + "learning_rate": 0.0007903871614844534, + "loss": 2.2647, + "theoretical_loss": 5.025513682044114, + "tokens_seen": 72286208 + }, + { + "epoch": 0.02, + "learning_rate": 0.0007903711133400201, + "loss": 2.4255, + "theoretical_loss": 5.024857053428793, + "tokens_seen": 72351744 + }, + { + "epoch": 0.02, + "learning_rate": 0.0007903550651955868, + "loss": 2.2659, + "theoretical_loss": 5.024201185681015, + "tokens_seen": 72417280 + }, + { + "epoch": 0.02, + "learning_rate": 0.0007903390170511535, + "loss": 2.3613, + "theoretical_loss": 5.023546077231752, + "tokens_seen": 72482816 + }, + { + "epoch": 0.02, + "learning_rate": 0.0007903229689067202, + "loss": 2.2581, + "theoretical_loss": 5.0228917265166295, + "tokens_seen": 72548352 + }, + { + "epoch": 0.02, + "learning_rate": 0.0007903069207622869, + "loss": 2.0883, + "theoretical_loss": 5.022238131975905, + "tokens_seen": 72613888 + }, + { + "epoch": 0.02, + "learning_rate": 0.0007902908726178537, + "loss": 2.2901, + "theoretical_loss": 5.021585292054446, + "tokens_seen": 72679424 + }, + { + "epoch": 0.02, + "learning_rate": 0.0007902748244734203, + "loss": 2.2835, + "theoretical_loss": 5.020933205201725, + "tokens_seen": 72744960 + }, + { + "epoch": 0.02, + "learning_rate": 0.000790258776328987, + "loss": 2.336, + "theoretical_loss": 5.020281869871787, + "tokens_seen": 72810496 + }, + { + "epoch": 0.02, + "learning_rate": 0.0007902427281845537, + "loss": 2.284, + "theoretical_loss": 5.019631284523241, + "tokens_seen": 72876032 + }, + { + "epoch": 0.02, + "learning_rate": 0.0007902266800401205, + "loss": 2.1953, + "theoretical_loss": 5.01898144761924, + "tokens_seen": 72941568 + }, + { + "epoch": 0.02, + "learning_rate": 0.0007902106318956871, + "loss": 2.1479, + "theoretical_loss": 5.018332357627462, + "tokens_seen": 73007104 + }, + { + "epoch": 0.02, + "learning_rate": 0.0007901945837512539, + "loss": 2.2555, + "theoretical_loss": 5.017684013020096, + "tokens_seen": 73072640 + }, + { + "epoch": 0.02, + "learning_rate": 0.0007901785356068205, + "loss": 2.3896, + "theoretical_loss": 5.017036412273823, + "tokens_seen": 73138176 + }, + { + "epoch": 0.02, + "learning_rate": 0.0007901624874623873, + "loss": 2.298, + "theoretical_loss": 5.016389553869796, + "tokens_seen": 73203712 + }, + { + "epoch": 0.02, + "learning_rate": 0.0007901464393179539, + "loss": 2.2897, + "theoretical_loss": 5.0157434362936275, + "tokens_seen": 73269248 + }, + { + "epoch": 0.02, + "learning_rate": 0.0007901303911735207, + "loss": 2.2989, + "theoretical_loss": 5.01509805803537, + "tokens_seen": 73334784 + }, + { + "epoch": 0.02, + "learning_rate": 0.0007901143430290873, + "loss": 2.2947, + "theoretical_loss": 5.0144534175894995, + "tokens_seen": 73400320 + }, + { + "epoch": 0.02, + "learning_rate": 0.0007900982948846541, + "loss": 2.0987, + "theoretical_loss": 5.013809513454898, + "tokens_seen": 73465856 + }, + { + "epoch": 0.02, + "learning_rate": 0.0007900822467402207, + "loss": 2.1887, + "theoretical_loss": 5.013166344134838, + "tokens_seen": 73531392 + }, + { + "epoch": 0.02, + "learning_rate": 0.0007900661985957875, + "loss": 2.2558, + "theoretical_loss": 5.012523908136963, + "tokens_seen": 73596928 + }, + { + "epoch": 0.02, + "learning_rate": 0.0007900501504513541, + "loss": 2.2905, + "theoretical_loss": 5.011882203973277, + "tokens_seen": 73662464 + }, + { + "epoch": 0.02, + "learning_rate": 0.0007900341023069209, + "loss": 2.1891, + "theoretical_loss": 5.011241230160119, + "tokens_seen": 73728000 + }, + { + "epoch": 0.02, + "learning_rate": 0.0007900180541624875, + "loss": 2.4206, + "theoretical_loss": 5.0106009852181534, + "tokens_seen": 73793536 + }, + { + "epoch": 0.02, + "learning_rate": 0.0007900020060180543, + "loss": 2.3818, + "theoretical_loss": 5.009961467672353, + "tokens_seen": 73859072 + }, + { + "epoch": 0.02, + "learning_rate": 0.0007899859578736209, + "loss": 2.2878, + "theoretical_loss": 5.009322676051976, + "tokens_seen": 73924608 + }, + { + "epoch": 0.02, + "learning_rate": 0.0007899699097291877, + "loss": 2.278, + "theoretical_loss": 5.008684608890562, + "tokens_seen": 73990144 + }, + { + "epoch": 0.02, + "learning_rate": 0.0007899538615847543, + "loss": 2.34, + "theoretical_loss": 5.008047264725901, + "tokens_seen": 74055680 + }, + { + "epoch": 0.02, + "learning_rate": 0.0007899378134403211, + "loss": 2.1904, + "theoretical_loss": 5.007410642100032, + "tokens_seen": 74121216 + }, + { + "epoch": 0.02, + "learning_rate": 0.0007899217652958877, + "loss": 2.464, + "theoretical_loss": 5.006774739559216, + "tokens_seen": 74186752 + }, + { + "epoch": 0.02, + "learning_rate": 0.0007899057171514545, + "loss": 2.3185, + "theoretical_loss": 5.00613955565392, + "tokens_seen": 74252288 + }, + { + "epoch": 0.02, + "learning_rate": 0.0007898896690070211, + "loss": 2.1584, + "theoretical_loss": 5.005505088938815, + "tokens_seen": 74317824 + }, + { + "epoch": 0.02, + "learning_rate": 0.0007898736208625879, + "loss": 2.3252, + "theoretical_loss": 5.004871337972742, + "tokens_seen": 74383360 + }, + { + "epoch": 0.02, + "learning_rate": 0.0007898575727181545, + "loss": 2.2353, + "theoretical_loss": 5.004238301318706, + "tokens_seen": 74448896 + }, + { + "epoch": 0.02, + "learning_rate": 0.0007898415245737213, + "loss": 2.2677, + "theoretical_loss": 5.003605977543861, + "tokens_seen": 74514432 + }, + { + "epoch": 0.02, + "learning_rate": 0.0007898254764292879, + "loss": 2.2295, + "theoretical_loss": 5.00297436521949, + "tokens_seen": 74579968 + }, + { + "epoch": 0.02, + "learning_rate": 0.0007898094282848547, + "loss": 2.1869, + "theoretical_loss": 5.002343462920992, + "tokens_seen": 74645504 + }, + { + "epoch": 0.02, + "learning_rate": 0.0007897933801404213, + "loss": 2.3628, + "theoretical_loss": 5.001713269227867, + "tokens_seen": 74711040 + }, + { + "epoch": 0.02, + "learning_rate": 0.0007897773319959881, + "loss": 2.2555, + "theoretical_loss": 5.0010837827237005, + "tokens_seen": 74776576 + }, + { + "epoch": 0.02, + "learning_rate": 0.0007897612838515547, + "loss": 2.2728, + "theoretical_loss": 5.000455001996143, + "tokens_seen": 74842112 + }, + { + "epoch": 0.02, + "learning_rate": 0.0007897452357071215, + "loss": 2.2935, + "theoretical_loss": 4.999826925636905, + "tokens_seen": 74907648 + }, + { + "epoch": 0.02, + "learning_rate": 0.0007897291875626881, + "loss": 2.3101, + "theoretical_loss": 4.999199552241732, + "tokens_seen": 74973184 + }, + { + "epoch": 0.02, + "learning_rate": 0.0007897131394182549, + "loss": 2.4127, + "theoretical_loss": 4.998572880410396, + "tokens_seen": 75038720 + }, + { + "epoch": 0.02, + "learning_rate": 0.0007896970912738215, + "loss": 2.1707, + "theoretical_loss": 4.997946908746677, + "tokens_seen": 75104256 + }, + { + "epoch": 0.02, + "learning_rate": 0.0007896810431293882, + "loss": 2.2421, + "theoretical_loss": 4.997321635858347, + "tokens_seen": 75169792 + }, + { + "epoch": 0.02, + "learning_rate": 0.0007896649949849549, + "loss": 2.1966, + "theoretical_loss": 4.9966970603571585, + "tokens_seen": 75235328 + }, + { + "epoch": 0.02, + "learning_rate": 0.0007896489468405216, + "loss": 2.3293, + "theoretical_loss": 4.996073180858831, + "tokens_seen": 75300864 + }, + { + "epoch": 0.02, + "objective/train/docs_used": 35047, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 2.2802019119262695, + "objective/train/theoretical_loss": 4.995449995983028, + "objective/train/tokens_used": 95826400, + "theoretical_loss": 4.995449995983028, + "tokens_seen": 75366400 + }, + { + "epoch": 0.02, + "learning_rate": 0.0007896328986960883, + "loss": 2.2961, + "theoretical_loss": 4.995449995983028, + "tokens_seen": 75366400 + }, + { + "epoch": 0.02, + "learning_rate": 0.000789616850551655, + "loss": 2.2703, + "theoretical_loss": 4.994827504353353, + "tokens_seen": 75431936 + }, + { + "epoch": 0.02, + "learning_rate": 0.0007896008024072217, + "loss": 2.2969, + "theoretical_loss": 4.994205704597325, + "tokens_seen": 75497472 + }, + { + "epoch": 0.02, + "learning_rate": 0.0007895847542627884, + "loss": 2.1548, + "theoretical_loss": 4.993584595346373, + "tokens_seen": 75563008 + }, + { + "epoch": 0.02, + "learning_rate": 0.0007895687061183551, + "loss": 2.344, + "theoretical_loss": 4.992964175235814, + "tokens_seen": 75628544 + }, + { + "epoch": 0.02, + "learning_rate": 0.0007895526579739218, + "loss": 2.3171, + "theoretical_loss": 4.992344442904843, + "tokens_seen": 75694080 + }, + { + "epoch": 0.02, + "learning_rate": 0.0007895366098294884, + "loss": 2.1316, + "theoretical_loss": 4.991725396996515, + "tokens_seen": 75759616 + }, + { + "epoch": 0.02, + "learning_rate": 0.0007895205616850552, + "loss": 2.1144, + "theoretical_loss": 4.991107036157738, + "tokens_seen": 75825152 + }, + { + "epoch": 0.02, + "learning_rate": 0.0007895045135406218, + "loss": 2.3832, + "theoretical_loss": 4.990489359039248, + "tokens_seen": 75890688 + }, + { + "epoch": 0.02, + "learning_rate": 0.0007894884653961886, + "loss": 2.3555, + "theoretical_loss": 4.989872364295605, + "tokens_seen": 75956224 + }, + { + "epoch": 0.02, + "learning_rate": 0.0007894724172517552, + "loss": 2.2239, + "theoretical_loss": 4.989256050585174, + "tokens_seen": 76021760 + }, + { + "epoch": 0.02, + "learning_rate": 0.000789456369107322, + "loss": 2.3831, + "theoretical_loss": 4.9886404165701075, + "tokens_seen": 76087296 + }, + { + "epoch": 0.02, + "learning_rate": 0.0007894403209628886, + "loss": 2.3553, + "theoretical_loss": 4.98802546091634, + "tokens_seen": 76152832 + }, + { + "epoch": 0.02, + "learning_rate": 0.0007894242728184554, + "loss": 2.2756, + "theoretical_loss": 4.9874111822935685, + "tokens_seen": 76218368 + }, + { + "epoch": 0.02, + "learning_rate": 0.000789408224674022, + "loss": 2.3223, + "theoretical_loss": 4.986797579375239, + "tokens_seen": 76283904 + }, + { + "epoch": 0.02, + "learning_rate": 0.0007893921765295888, + "loss": 2.3169, + "theoretical_loss": 4.986184650838532, + "tokens_seen": 76349440 + }, + { + "epoch": 0.02, + "learning_rate": 0.0007893761283851554, + "loss": 2.2231, + "theoretical_loss": 4.985572395364354, + "tokens_seen": 76414976 + }, + { + "epoch": 0.02, + "learning_rate": 0.0007893600802407222, + "loss": 2.241, + "theoretical_loss": 4.98496081163732, + "tokens_seen": 76480512 + }, + { + "epoch": 0.02, + "learning_rate": 0.0007893440320962888, + "loss": 2.2073, + "theoretical_loss": 4.984349898345735, + "tokens_seen": 76546048 + }, + { + "epoch": 0.02, + "learning_rate": 0.0007893279839518556, + "loss": 2.1966, + "theoretical_loss": 4.98373965418159, + "tokens_seen": 76611584 + }, + { + "epoch": 0.02, + "learning_rate": 0.0007893119358074222, + "loss": 2.2713, + "theoretical_loss": 4.983130077840543, + "tokens_seen": 76677120 + }, + { + "epoch": 0.02, + "learning_rate": 0.000789295887662989, + "loss": 2.2052, + "theoretical_loss": 4.982521168021906, + "tokens_seen": 76742656 + }, + { + "epoch": 0.02, + "learning_rate": 0.0007892798395185556, + "loss": 2.156, + "theoretical_loss": 4.981912923428634, + "tokens_seen": 76808192 + }, + { + "epoch": 0.02, + "learning_rate": 0.0007892637913741224, + "loss": 2.2599, + "theoretical_loss": 4.981305342767307, + "tokens_seen": 76873728 + }, + { + "epoch": 0.02, + "learning_rate": 0.000789247743229689, + "loss": 2.3021, + "theoretical_loss": 4.980698424748122, + "tokens_seen": 76939264 + }, + { + "epoch": 0.02, + "learning_rate": 0.0007892316950852558, + "loss": 2.2046, + "theoretical_loss": 4.980092168084878, + "tokens_seen": 77004800 + }, + { + "epoch": 0.02, + "learning_rate": 0.0007892156469408224, + "loss": 2.2456, + "theoretical_loss": 4.97948657149496, + "tokens_seen": 77070336 + }, + { + "epoch": 0.02, + "learning_rate": 0.0007891995987963892, + "loss": 2.285, + "theoretical_loss": 4.978881633699331, + "tokens_seen": 77135872 + }, + { + "epoch": 0.02, + "learning_rate": 0.0007891835506519559, + "loss": 2.1765, + "theoretical_loss": 4.9782773534225155, + "tokens_seen": 77201408 + }, + { + "epoch": 0.02, + "learning_rate": 0.0007891675025075226, + "loss": 2.2379, + "theoretical_loss": 4.977673729392586, + "tokens_seen": 77266944 + }, + { + "epoch": 0.02, + "learning_rate": 0.0007891514543630893, + "loss": 2.2035, + "theoretical_loss": 4.977070760341153, + "tokens_seen": 77332480 + }, + { + "epoch": 0.02, + "learning_rate": 0.000789135406218656, + "loss": 2.2417, + "theoretical_loss": 4.976468445003352, + "tokens_seen": 77398016 + }, + { + "epoch": 0.02, + "learning_rate": 0.0007891193580742227, + "loss": 2.2679, + "theoretical_loss": 4.975866782117828, + "tokens_seen": 77463552 + }, + { + "epoch": 0.02, + "learning_rate": 0.0007891033099297894, + "loss": 2.175, + "theoretical_loss": 4.975265770426724, + "tokens_seen": 77529088 + }, + { + "epoch": 0.02, + "learning_rate": 0.0007890872617853561, + "loss": 2.2733, + "theoretical_loss": 4.974665408675667, + "tokens_seen": 77594624 + }, + { + "epoch": 0.02, + "learning_rate": 0.0007890712136409228, + "loss": 2.2511, + "theoretical_loss": 4.974065695613763, + "tokens_seen": 77660160 + }, + { + "epoch": 0.02, + "learning_rate": 0.0007890551654964895, + "loss": 2.2099, + "theoretical_loss": 4.973466629993574, + "tokens_seen": 77725696 + }, + { + "epoch": 0.02, + "learning_rate": 0.0007890391173520562, + "loss": 2.1201, + "theoretical_loss": 4.972868210571109, + "tokens_seen": 77791232 + }, + { + "epoch": 0.02, + "learning_rate": 0.000789023069207623, + "loss": 2.4327, + "theoretical_loss": 4.972270436105815, + "tokens_seen": 77856768 + }, + { + "epoch": 0.02, + "learning_rate": 0.0007890070210631896, + "loss": 2.1807, + "theoretical_loss": 4.971673305360563, + "tokens_seen": 77922304 + }, + { + "epoch": 0.02, + "learning_rate": 0.0007889909729187563, + "loss": 2.1866, + "theoretical_loss": 4.971076817101633, + "tokens_seen": 77987840 + }, + { + "epoch": 0.02, + "learning_rate": 0.000788974924774323, + "loss": 2.1687, + "theoretical_loss": 4.970480970098703, + "tokens_seen": 78053376 + }, + { + "epoch": 0.02, + "learning_rate": 0.0007889588766298897, + "loss": 2.1031, + "theoretical_loss": 4.96988576312484, + "tokens_seen": 78118912 + }, + { + "epoch": 0.02, + "learning_rate": 0.0007889428284854564, + "loss": 2.2877, + "theoretical_loss": 4.969291194956485, + "tokens_seen": 78184448 + }, + { + "epoch": 0.02, + "learning_rate": 0.0007889267803410232, + "loss": 2.2416, + "theoretical_loss": 4.968697264373438, + "tokens_seen": 78249984 + }, + { + "epoch": 0.02, + "learning_rate": 0.0007889107321965898, + "loss": 2.0862, + "theoretical_loss": 4.968103970158854, + "tokens_seen": 78315520 + }, + { + "epoch": 0.02, + "learning_rate": 0.0007888946840521566, + "loss": 2.2375, + "theoretical_loss": 4.967511311099223, + "tokens_seen": 78381056 + }, + { + "epoch": 0.02, + "learning_rate": 0.0007888786359077232, + "loss": 2.0887, + "theoretical_loss": 4.96691928598436, + "tokens_seen": 78446592 + }, + { + "epoch": 0.02, + "learning_rate": 0.00078886258776329, + "loss": 2.2336, + "theoretical_loss": 4.966327893607398, + "tokens_seen": 78512128 + }, + { + "epoch": 0.02, + "learning_rate": 0.0007888465396188566, + "loss": 2.2125, + "theoretical_loss": 4.96573713276477, + "tokens_seen": 78577664 + }, + { + "epoch": 0.02, + "objective/train/docs_used": 36308, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 2.292598009109497, + "objective/train/theoretical_loss": 4.965147002256201, + "objective/train/tokens_used": 99103200, + "theoretical_loss": 4.965147002256201, + "tokens_seen": 78643200 + }, + { + "epoch": 0.02, + "learning_rate": 0.0007888304914744234, + "loss": 2.3008, + "theoretical_loss": 4.965147002256201, + "tokens_seen": 78643200 + }, + { + "epoch": 0.02, + "learning_rate": 0.00078881444332999, + "loss": 2.2264, + "theoretical_loss": 4.964557500884695, + "tokens_seen": 78708736 + }, + { + "epoch": 0.02, + "learning_rate": 0.0007887983951855568, + "loss": 2.3307, + "theoretical_loss": 4.963968627456523, + "tokens_seen": 78774272 + }, + { + "epoch": 0.02, + "learning_rate": 0.0007887823470411234, + "loss": 2.0944, + "theoretical_loss": 4.96338038078121, + "tokens_seen": 78839808 + }, + { + "epoch": 0.02, + "learning_rate": 0.0007887662988966902, + "loss": 2.1684, + "theoretical_loss": 4.962792759671528, + "tokens_seen": 78905344 + }, + { + "epoch": 0.02, + "learning_rate": 0.0007887502507522568, + "loss": 2.1921, + "theoretical_loss": 4.962205762943478, + "tokens_seen": 78970880 + }, + { + "epoch": 0.02, + "learning_rate": 0.0007887342026078236, + "loss": 2.1253, + "theoretical_loss": 4.961619389416288, + "tokens_seen": 79036416 + }, + { + "epoch": 0.02, + "learning_rate": 0.0007887181544633902, + "loss": 2.1325, + "theoretical_loss": 4.96103363791239, + "tokens_seen": 79101952 + }, + { + "epoch": 0.02, + "learning_rate": 0.000788702106318957, + "loss": 2.2259, + "theoretical_loss": 4.960448507257416, + "tokens_seen": 79167488 + }, + { + "epoch": 0.02, + "learning_rate": 0.0007886860581745236, + "loss": 2.2462, + "theoretical_loss": 4.959863996280188, + "tokens_seen": 79233024 + }, + { + "epoch": 0.02, + "learning_rate": 0.0007886700100300904, + "loss": 2.3568, + "theoretical_loss": 4.959280103812697, + "tokens_seen": 79298560 + }, + { + "epoch": 0.02, + "learning_rate": 0.000788653961885657, + "loss": 2.1976, + "theoretical_loss": 4.958696828690105, + "tokens_seen": 79364096 + }, + { + "epoch": 0.02, + "learning_rate": 0.0007886379137412238, + "loss": 2.2042, + "theoretical_loss": 4.958114169750726, + "tokens_seen": 79429632 + }, + { + "epoch": 0.02, + "learning_rate": 0.0007886218655967904, + "loss": 2.2204, + "theoretical_loss": 4.957532125836012, + "tokens_seen": 79495168 + }, + { + "epoch": 0.02, + "learning_rate": 0.0007886058174523572, + "loss": 2.3061, + "theoretical_loss": 4.956950695790551, + "tokens_seen": 79560704 + }, + { + "epoch": 0.02, + "learning_rate": 0.0007885897693079238, + "loss": 2.2156, + "theoretical_loss": 4.95636987846205, + "tokens_seen": 79626240 + }, + { + "epoch": 0.02, + "learning_rate": 0.0007885737211634906, + "loss": 2.2333, + "theoretical_loss": 4.955789672701323, + "tokens_seen": 79691776 + }, + { + "epoch": 0.02, + "learning_rate": 0.0007885576730190572, + "loss": 2.2474, + "theoretical_loss": 4.955210077362285, + "tokens_seen": 79757312 + }, + { + "epoch": 0.02, + "learning_rate": 0.000788541624874624, + "loss": 2.3146, + "theoretical_loss": 4.954631091301934, + "tokens_seen": 79822848 + }, + { + "epoch": 0.02, + "learning_rate": 0.0007885255767301906, + "loss": 2.1544, + "theoretical_loss": 4.954052713380349, + "tokens_seen": 79888384 + }, + { + "epoch": 0.02, + "learning_rate": 0.0007885095285857574, + "loss": 2.1181, + "theoretical_loss": 4.953474942460673, + "tokens_seen": 79953920 + }, + { + "epoch": 0.02, + "learning_rate": 0.000788493480441324, + "loss": 2.171, + "theoretical_loss": 4.952897777409104, + "tokens_seen": 80019456 + }, + { + "epoch": 0.02, + "learning_rate": 0.0007884774322968908, + "loss": 2.2195, + "theoretical_loss": 4.952321217094882, + "tokens_seen": 80084992 + }, + { + "epoch": 0.02, + "learning_rate": 0.0007884613841524574, + "loss": 2.1196, + "theoretical_loss": 4.951745260390284, + "tokens_seen": 80150528 + }, + { + "epoch": 0.02, + "learning_rate": 0.0007884453360080242, + "loss": 2.1398, + "theoretical_loss": 4.951169906170607, + "tokens_seen": 80216064 + }, + { + "epoch": 0.02, + "learning_rate": 0.0007884292878635908, + "loss": 2.2183, + "theoretical_loss": 4.950595153314161, + "tokens_seen": 80281600 + }, + { + "epoch": 0.02, + "learning_rate": 0.0007884132397191576, + "loss": 2.221, + "theoretical_loss": 4.950021000702261, + "tokens_seen": 80347136 + }, + { + "epoch": 0.02, + "learning_rate": 0.0007883971915747242, + "loss": 2.1549, + "theoretical_loss": 4.949447447219208, + "tokens_seen": 80412672 + }, + { + "epoch": 0.02, + "learning_rate": 0.000788381143430291, + "loss": 2.1703, + "theoretical_loss": 4.948874491752287, + "tokens_seen": 80478208 + }, + { + "epoch": 0.02, + "learning_rate": 0.0007883650952858576, + "loss": 2.0932, + "theoretical_loss": 4.9483021331917545, + "tokens_seen": 80543744 + }, + { + "epoch": 0.02, + "learning_rate": 0.0007883490471414244, + "loss": 2.084, + "theoretical_loss": 4.9477303704308255, + "tokens_seen": 80609280 + }, + { + "epoch": 0.02, + "learning_rate": 0.000788332998996991, + "loss": 2.1981, + "theoretical_loss": 4.947159202365664, + "tokens_seen": 80674816 + }, + { + "epoch": 0.02, + "learning_rate": 0.0007883169508525578, + "loss": 2.1613, + "theoretical_loss": 4.946588627895375, + "tokens_seen": 80740352 + }, + { + "epoch": 0.02, + "learning_rate": 0.0007883009027081244, + "loss": 2.1629, + "theoretical_loss": 4.946018645921994, + "tokens_seen": 80805888 + }, + { + "epoch": 0.02, + "learning_rate": 0.0007882848545636911, + "loss": 2.2667, + "theoretical_loss": 4.945449255350473, + "tokens_seen": 80871424 + }, + { + "epoch": 0.02, + "learning_rate": 0.0007882688064192578, + "loss": 2.2449, + "theoretical_loss": 4.944880455088677, + "tokens_seen": 80936960 + }, + { + "epoch": 0.02, + "learning_rate": 0.0007882527582748245, + "loss": 2.2763, + "theoretical_loss": 4.944312244047365, + "tokens_seen": 81002496 + }, + { + "epoch": 0.02, + "learning_rate": 0.0007882367101303912, + "loss": 2.1294, + "theoretical_loss": 4.943744621140189, + "tokens_seen": 81068032 + }, + { + "epoch": 0.02, + "learning_rate": 0.0007882206619859579, + "loss": 2.1047, + "theoretical_loss": 4.94317758528368, + "tokens_seen": 81133568 + }, + { + "epoch": 0.02, + "learning_rate": 0.0007882046138415246, + "loss": 2.2569, + "theoretical_loss": 4.942611135397234, + "tokens_seen": 81199104 + }, + { + "epoch": 0.02, + "learning_rate": 0.0007881885656970913, + "loss": 2.2905, + "theoretical_loss": 4.942045270403112, + "tokens_seen": 81264640 + }, + { + "epoch": 0.02, + "learning_rate": 0.000788172517552658, + "loss": 2.2042, + "theoretical_loss": 4.941479989226421, + "tokens_seen": 81330176 + }, + { + "epoch": 0.02, + "learning_rate": 0.0007881564694082247, + "loss": 2.3198, + "theoretical_loss": 4.940915290795106, + "tokens_seen": 81395712 + }, + { + "epoch": 0.02, + "learning_rate": 0.0007881404212637913, + "loss": 2.2341, + "theoretical_loss": 4.940351174039947, + "tokens_seen": 81461248 + }, + { + "epoch": 0.02, + "learning_rate": 0.0007881243731193581, + "loss": 2.2467, + "theoretical_loss": 4.939787637894537, + "tokens_seen": 81526784 + }, + { + "epoch": 0.02, + "learning_rate": 0.0007881083249749247, + "loss": 2.0796, + "theoretical_loss": 4.939224681295286, + "tokens_seen": 81592320 + }, + { + "epoch": 0.02, + "learning_rate": 0.0007880922768304915, + "loss": 2.2327, + "theoretical_loss": 4.938662303181399, + "tokens_seen": 81657856 + }, + { + "epoch": 0.02, + "learning_rate": 0.0007880762286860581, + "loss": 2.381, + "theoretical_loss": 4.938100502494878, + "tokens_seen": 81723392 + }, + { + "epoch": 0.02, + "learning_rate": 0.0007880601805416249, + "loss": 2.3149, + "theoretical_loss": 4.9375392781805, + "tokens_seen": 81788928 + }, + { + "epoch": 0.02, + "learning_rate": 0.0007880441323971915, + "loss": 2.1916, + "theoretical_loss": 4.936978629185818, + "tokens_seen": 81854464 + }, + { + "epoch": 0.02, + "objective/train/docs_used": 37569, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 2.1436986923217773, + "objective/train/theoretical_loss": 4.936418554461147, + "objective/train/tokens_used": 102380000, + "theoretical_loss": 4.936418554461147, + "tokens_seen": 81920000 + }, + { + "epoch": 0.02, + "learning_rate": 0.0007880280842527583, + "loss": 2.2142, + "theoretical_loss": 4.936418554461147, + "tokens_seen": 81920000 + }, + { + "epoch": 0.02, + "learning_rate": 0.0007880120361083249, + "loss": 2.2398, + "theoretical_loss": 4.935859052959554, + "tokens_seen": 81985536 + }, + { + "epoch": 0.02, + "learning_rate": 0.0007879959879638917, + "loss": 2.0788, + "theoretical_loss": 4.935300123636851, + "tokens_seen": 82051072 + }, + { + "epoch": 0.02, + "learning_rate": 0.0007879799398194584, + "loss": 2.2313, + "theoretical_loss": 4.93474176545158, + "tokens_seen": 82116608 + }, + { + "epoch": 0.02, + "learning_rate": 0.0007879638916750251, + "loss": 2.1998, + "theoretical_loss": 4.934183977365015, + "tokens_seen": 82182144 + }, + { + "epoch": 0.02, + "learning_rate": 0.0007879478435305918, + "loss": 2.1024, + "theoretical_loss": 4.9336267583411395, + "tokens_seen": 82247680 + }, + { + "epoch": 0.02, + "learning_rate": 0.0007879317953861585, + "loss": 2.1041, + "theoretical_loss": 4.933070107346644, + "tokens_seen": 82313216 + }, + { + "epoch": 0.02, + "learning_rate": 0.0007879157472417252, + "loss": 2.1461, + "theoretical_loss": 4.932514023350921, + "tokens_seen": 82378752 + }, + { + "epoch": 0.02, + "learning_rate": 0.0007878996990972919, + "loss": 2.2605, + "theoretical_loss": 4.931958505326042, + "tokens_seen": 82444288 + }, + { + "epoch": 0.03, + "learning_rate": 0.0007878836509528586, + "loss": 2.1596, + "theoretical_loss": 4.931403552246769, + "tokens_seen": 82509824 + }, + { + "epoch": 0.03, + "learning_rate": 0.0007878676028084253, + "loss": 2.1929, + "theoretical_loss": 4.930849163090521, + "tokens_seen": 82575360 + }, + { + "epoch": 0.03, + "learning_rate": 0.000787851554663992, + "loss": 2.1834, + "theoretical_loss": 4.930295336837387, + "tokens_seen": 82640896 + }, + { + "epoch": 0.03, + "learning_rate": 0.0007878355065195587, + "loss": 2.1262, + "theoretical_loss": 4.9297420724701055, + "tokens_seen": 82706432 + }, + { + "epoch": 0.03, + "learning_rate": 0.0007878194583751254, + "loss": 2.1854, + "theoretical_loss": 4.929189368974054, + "tokens_seen": 82771968 + }, + { + "epoch": 0.03, + "learning_rate": 0.0007878034102306921, + "loss": 2.1295, + "theoretical_loss": 4.928637225337246, + "tokens_seen": 82837504 + }, + { + "epoch": 0.03, + "learning_rate": 0.0007877873620862588, + "loss": 2.2654, + "theoretical_loss": 4.9280856405503215, + "tokens_seen": 82903040 + }, + { + "epoch": 0.03, + "learning_rate": 0.0007877713139418255, + "loss": 2.2803, + "theoretical_loss": 4.927534613606534, + "tokens_seen": 82968576 + }, + { + "epoch": 0.03, + "learning_rate": 0.0007877552657973922, + "loss": 2.0939, + "theoretical_loss": 4.926984143501745, + "tokens_seen": 83034112 + }, + { + "epoch": 0.03, + "learning_rate": 0.0007877392176529589, + "loss": 2.1945, + "theoretical_loss": 4.926434229234411, + "tokens_seen": 83099648 + }, + { + "epoch": 0.03, + "learning_rate": 0.0007877231695085256, + "loss": 2.1789, + "theoretical_loss": 4.925884869805586, + "tokens_seen": 83165184 + }, + { + "epoch": 0.03, + "learning_rate": 0.0007877071213640923, + "loss": 2.1121, + "theoretical_loss": 4.925336064218893, + "tokens_seen": 83230720 + }, + { + "epoch": 0.03, + "learning_rate": 0.000787691073219659, + "loss": 2.2444, + "theoretical_loss": 4.924787811480538, + "tokens_seen": 83296256 + }, + { + "epoch": 0.03, + "learning_rate": 0.0007876750250752257, + "loss": 2.2246, + "theoretical_loss": 4.9242401105992855, + "tokens_seen": 83361792 + }, + { + "epoch": 0.03, + "learning_rate": 0.0007876589769307924, + "loss": 2.2329, + "theoretical_loss": 4.923692960586452, + "tokens_seen": 83427328 + }, + { + "epoch": 0.03, + "learning_rate": 0.0007876429287863591, + "loss": 1.9818, + "theoretical_loss": 4.923146360455906, + "tokens_seen": 83492864 + }, + { + "epoch": 0.03, + "learning_rate": 0.0007876268806419258, + "loss": 2.1096, + "theoretical_loss": 4.92260030922405, + "tokens_seen": 83558400 + }, + { + "epoch": 0.03, + "learning_rate": 0.0007876108324974925, + "loss": 2.3329, + "theoretical_loss": 4.922054805909816, + "tokens_seen": 83623936 + }, + { + "epoch": 0.03, + "learning_rate": 0.0007875947843530592, + "loss": 2.153, + "theoretical_loss": 4.921509849534658, + "tokens_seen": 83689472 + }, + { + "epoch": 0.03, + "learning_rate": 0.0007875787362086259, + "loss": 2.1602, + "theoretical_loss": 4.920965439122538, + "tokens_seen": 83755008 + }, + { + "epoch": 0.03, + "learning_rate": 0.0007875626880641927, + "loss": 2.1412, + "theoretical_loss": 4.9204215736999295, + "tokens_seen": 83820544 + }, + { + "epoch": 0.03, + "learning_rate": 0.0007875466399197593, + "loss": 2.1037, + "theoretical_loss": 4.919878252295793, + "tokens_seen": 83886080 + }, + { + "epoch": 0.03, + "learning_rate": 0.000787530591775326, + "loss": 2.2191, + "theoretical_loss": 4.919335473941581, + "tokens_seen": 83951616 + }, + { + "epoch": 0.03, + "learning_rate": 0.0007875145436308927, + "loss": 2.2753, + "theoretical_loss": 4.918793237671224, + "tokens_seen": 84017152 + }, + { + "epoch": 0.03, + "learning_rate": 0.0007874984954864595, + "loss": 2.1998, + "theoretical_loss": 4.918251542521121, + "tokens_seen": 84082688 + }, + { + "epoch": 0.03, + "learning_rate": 0.0007874824473420261, + "loss": 2.2343, + "theoretical_loss": 4.9177103875301365, + "tokens_seen": 84148224 + }, + { + "epoch": 0.03, + "learning_rate": 0.0007874663991975929, + "loss": 2.159, + "theoretical_loss": 4.917169771739586, + "tokens_seen": 84213760 + }, + { + "epoch": 0.03, + "learning_rate": 0.0007874503510531595, + "loss": 2.1038, + "theoretical_loss": 4.916629694193233, + "tokens_seen": 84279296 + }, + { + "epoch": 0.03, + "learning_rate": 0.0007874343029087263, + "loss": 2.1748, + "theoretical_loss": 4.916090153937276, + "tokens_seen": 84344832 + }, + { + "epoch": 0.03, + "learning_rate": 0.0007874182547642929, + "loss": 2.2899, + "theoretical_loss": 4.915551150020346, + "tokens_seen": 84410368 + }, + { + "epoch": 0.03, + "learning_rate": 0.0007874022066198597, + "loss": 2.3333, + "theoretical_loss": 4.915012681493495, + "tokens_seen": 84475904 + }, + { + "epoch": 0.03, + "learning_rate": 0.0007873861584754263, + "loss": 2.2483, + "theoretical_loss": 4.914474747410187, + "tokens_seen": 84541440 + }, + { + "epoch": 0.03, + "learning_rate": 0.0007873701103309931, + "loss": 2.0885, + "theoretical_loss": 4.913937346826293, + "tokens_seen": 84606976 + }, + { + "epoch": 0.03, + "learning_rate": 0.0007873540621865597, + "loss": 2.192, + "theoretical_loss": 4.91340047880008, + "tokens_seen": 84672512 + }, + { + "epoch": 0.03, + "learning_rate": 0.0007873380140421265, + "loss": 2.154, + "theoretical_loss": 4.9128641423922055, + "tokens_seen": 84738048 + }, + { + "epoch": 0.03, + "learning_rate": 0.0007873219658976931, + "loss": 2.1588, + "theoretical_loss": 4.912328336665711, + "tokens_seen": 84803584 + }, + { + "epoch": 0.03, + "learning_rate": 0.0007873059177532599, + "loss": 2.1339, + "theoretical_loss": 4.911793060686007, + "tokens_seen": 84869120 + }, + { + "epoch": 0.03, + "learning_rate": 0.0007872898696088265, + "loss": 2.2585, + "theoretical_loss": 4.9112583135208725, + "tokens_seen": 84934656 + }, + { + "epoch": 0.03, + "learning_rate": 0.0007872738214643933, + "loss": 2.2594, + "theoretical_loss": 4.910724094240447, + "tokens_seen": 85000192 + }, + { + "epoch": 0.03, + "learning_rate": 0.0007872577733199599, + "loss": 2.1494, + "theoretical_loss": 4.910190401917214, + "tokens_seen": 85065728 + }, + { + "epoch": 0.03, + "learning_rate": 0.0007872417251755267, + "loss": 2.2764, + "theoretical_loss": 4.909657235626009, + "tokens_seen": 85131264 + }, + { + "epoch": 0.03, + "objective/train/docs_used": 38678, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 2.1814937591552734, + "objective/train/theoretical_loss": 4.909124594443993, + "objective/train/tokens_used": 105656800, + "theoretical_loss": 4.909124594443993, + "tokens_seen": 85196800 + }, + { + "epoch": 0.03, + "learning_rate": 0.0007872256770310933, + "loss": 2.1822, + "theoretical_loss": 4.909124594443993, + "tokens_seen": 85196800 + }, + { + "epoch": 0.03, + "learning_rate": 0.0007872096288866601, + "loss": 2.135, + "theoretical_loss": 4.908592477450659, + "tokens_seen": 85262336 + }, + { + "epoch": 0.03, + "learning_rate": 0.0007871935807422267, + "loss": 2.2014, + "theoretical_loss": 4.908060883727821, + "tokens_seen": 85327872 + }, + { + "epoch": 0.03, + "learning_rate": 0.0007871775325977935, + "loss": 2.1418, + "theoretical_loss": 4.9075298123596, + "tokens_seen": 85393408 + }, + { + "epoch": 0.03, + "learning_rate": 0.0007871614844533601, + "loss": 2.0806, + "theoretical_loss": 4.906999262432427, + "tokens_seen": 85458944 + }, + { + "epoch": 0.03, + "learning_rate": 0.0007871454363089269, + "loss": 2.1216, + "theoretical_loss": 4.906469233035025, + "tokens_seen": 85524480 + }, + { + "epoch": 0.03, + "learning_rate": 0.0007871293881644935, + "loss": 2.3005, + "theoretical_loss": 4.905939723258408, + "tokens_seen": 85590016 + }, + { + "epoch": 0.03, + "learning_rate": 0.0007871133400200603, + "loss": 2.2055, + "theoretical_loss": 4.905410732195875, + "tokens_seen": 85655552 + }, + { + "epoch": 0.03, + "learning_rate": 0.0007870972918756269, + "loss": 2.2052, + "theoretical_loss": 4.904882258942994, + "tokens_seen": 85721088 + }, + { + "epoch": 0.03, + "learning_rate": 0.0007870812437311937, + "loss": 2.0642, + "theoretical_loss": 4.904354302597603, + "tokens_seen": 85786624 + }, + { + "epoch": 0.03, + "learning_rate": 0.0007870651955867603, + "loss": 2.2495, + "theoretical_loss": 4.903826862259798, + "tokens_seen": 85852160 + }, + { + "epoch": 0.03, + "learning_rate": 0.0007870491474423271, + "loss": 2.1897, + "theoretical_loss": 4.90329993703193, + "tokens_seen": 85917696 + }, + { + "epoch": 0.03, + "learning_rate": 0.0007870330992978937, + "loss": 2.1768, + "theoretical_loss": 4.902773526018591, + "tokens_seen": 85983232 + }, + { + "epoch": 0.03, + "learning_rate": 0.0007870170511534605, + "loss": 2.1515, + "theoretical_loss": 4.902247628326613, + "tokens_seen": 86048768 + }, + { + "epoch": 0.03, + "learning_rate": 0.0007870010030090271, + "loss": 2.2251, + "theoretical_loss": 4.9017222430650556, + "tokens_seen": 86114304 + }, + { + "epoch": 0.03, + "learning_rate": 0.0007869849548645939, + "loss": 2.2248, + "theoretical_loss": 4.901197369345205, + "tokens_seen": 86179840 + }, + { + "epoch": 0.03, + "learning_rate": 0.0007869689067201605, + "loss": 2.0499, + "theoretical_loss": 4.9006730062805595, + "tokens_seen": 86245376 + }, + { + "epoch": 0.03, + "learning_rate": 0.0007869528585757273, + "loss": 2.1936, + "theoretical_loss": 4.9001491529868275, + "tokens_seen": 86310912 + }, + { + "epoch": 0.03, + "learning_rate": 0.0007869368104312939, + "loss": 2.0632, + "theoretical_loss": 4.899625808581918, + "tokens_seen": 86376448 + }, + { + "epoch": 0.03, + "learning_rate": 0.0007869207622868607, + "loss": 2.2845, + "theoretical_loss": 4.899102972185936, + "tokens_seen": 86441984 + }, + { + "epoch": 0.03, + "learning_rate": 0.0007869047141424273, + "loss": 2.0632, + "theoretical_loss": 4.898580642921171, + "tokens_seen": 86507520 + }, + { + "epoch": 0.03, + "learning_rate": 0.0007868886659979941, + "loss": 2.1255, + "theoretical_loss": 4.898058819912093, + "tokens_seen": 86573056 + }, + { + "epoch": 0.03, + "learning_rate": 0.0007868726178535607, + "loss": 2.2628, + "theoretical_loss": 4.897537502285347, + "tokens_seen": 86638592 + }, + { + "epoch": 0.03, + "learning_rate": 0.0007868565697091274, + "loss": 2.2799, + "theoretical_loss": 4.897016689169741, + "tokens_seen": 86704128 + }, + { + "epoch": 0.03, + "learning_rate": 0.0007868405215646941, + "loss": 2.0682, + "theoretical_loss": 4.896496379696243, + "tokens_seen": 86769664 + }, + { + "epoch": 0.03, + "learning_rate": 0.0007868244734202608, + "loss": 2.1795, + "theoretical_loss": 4.895976572997973, + "tokens_seen": 86835200 + }, + { + "epoch": 0.03, + "learning_rate": 0.0007868084252758275, + "loss": 2.1368, + "theoretical_loss": 4.895457268210196, + "tokens_seen": 86900736 + }, + { + "epoch": 0.03, + "learning_rate": 0.0007867923771313942, + "loss": 2.1777, + "theoretical_loss": 4.8949384644703136, + "tokens_seen": 86966272 + }, + { + "epoch": 0.03, + "learning_rate": 0.0007867763289869609, + "loss": 2.1068, + "theoretical_loss": 4.894420160917861, + "tokens_seen": 87031808 + }, + { + "epoch": 0.03, + "learning_rate": 0.0007867602808425276, + "loss": 2.273, + "theoretical_loss": 4.893902356694495, + "tokens_seen": 87097344 + }, + { + "epoch": 0.03, + "learning_rate": 0.0007867442326980943, + "loss": 2.0657, + "theoretical_loss": 4.8933850509439925, + "tokens_seen": 87162880 + }, + { + "epoch": 0.03, + "learning_rate": 0.000786728184553661, + "loss": 2.2019, + "theoretical_loss": 4.89286824281224, + "tokens_seen": 87228416 + }, + { + "epoch": 0.03, + "learning_rate": 0.0007867121364092277, + "loss": 2.0909, + "theoretical_loss": 4.892351931447224, + "tokens_seen": 87293952 + }, + { + "epoch": 0.03, + "learning_rate": 0.0007866960882647944, + "loss": 2.1506, + "theoretical_loss": 4.891836115999034, + "tokens_seen": 87359488 + }, + { + "epoch": 0.03, + "learning_rate": 0.0007866800401203611, + "loss": 2.0881, + "theoretical_loss": 4.891320795619847, + "tokens_seen": 87425024 + }, + { + "epoch": 0.03, + "learning_rate": 0.0007866639919759278, + "loss": 2.1078, + "theoretical_loss": 4.890805969463923, + "tokens_seen": 87490560 + }, + { + "epoch": 0.03, + "learning_rate": 0.0007866479438314945, + "loss": 2.1719, + "theoretical_loss": 4.8902916366876, + "tokens_seen": 87556096 + }, + { + "epoch": 0.03, + "learning_rate": 0.0007866318956870612, + "loss": 2.102, + "theoretical_loss": 4.889777796449285, + "tokens_seen": 87621632 + }, + { + "epoch": 0.03, + "learning_rate": 0.0007866158475426279, + "loss": 2.1014, + "theoretical_loss": 4.889264447909451, + "tokens_seen": 87687168 + }, + { + "epoch": 0.03, + "learning_rate": 0.0007865997993981946, + "loss": 2.1098, + "theoretical_loss": 4.888751590230626, + "tokens_seen": 87752704 + }, + { + "epoch": 0.03, + "learning_rate": 0.0007865837512537613, + "loss": 2.1156, + "theoretical_loss": 4.888239222577389, + "tokens_seen": 87818240 + }, + { + "epoch": 0.03, + "learning_rate": 0.000786567703109328, + "loss": 2.2373, + "theoretical_loss": 4.887727344116364, + "tokens_seen": 87883776 + }, + { + "epoch": 0.03, + "learning_rate": 0.0007865516549648947, + "loss": 2.1571, + "theoretical_loss": 4.8872159540162095, + "tokens_seen": 87949312 + }, + { + "epoch": 0.03, + "learning_rate": 0.0007865356068204614, + "loss": 2.1669, + "theoretical_loss": 4.88670505144762, + "tokens_seen": 88014848 + }, + { + "epoch": 0.03, + "learning_rate": 0.0007865195586760281, + "loss": 2.1493, + "theoretical_loss": 4.8861946355833075, + "tokens_seen": 88080384 + }, + { + "epoch": 0.03, + "learning_rate": 0.0007865035105315948, + "loss": 2.1801, + "theoretical_loss": 4.885684705598009, + "tokens_seen": 88145920 + }, + { + "epoch": 0.03, + "learning_rate": 0.0007864874623871615, + "loss": 2.1357, + "theoretical_loss": 4.885175260668468, + "tokens_seen": 88211456 + }, + { + "epoch": 0.03, + "learning_rate": 0.0007864714142427282, + "loss": 2.1099, + "theoretical_loss": 4.884666299973436, + "tokens_seen": 88276992 + }, + { + "epoch": 0.03, + "learning_rate": 0.0007864553660982949, + "loss": 2.2475, + "theoretical_loss": 4.884157822693663, + "tokens_seen": 88342528 + }, + { + "epoch": 0.03, + "learning_rate": 0.0007864393179538616, + "loss": 2.0579, + "theoretical_loss": 4.88364982801189, + "tokens_seen": 88408064 + }, + { + "epoch": 0.03, + "objective/train/docs_used": 39491, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 2.059610605239868, + "objective/train/theoretical_loss": 4.8831423151128455, + "objective/train/tokens_used": 108933600, + "theoretical_loss": 4.8831423151128455, + "tokens_seen": 88473600 + }, + { + "epoch": 0.03, + "learning_rate": 0.0007864232698094283, + "loss": 2.1604, + "theoretical_loss": 4.8831423151128455, + "tokens_seen": 88473600 + }, + { + "epoch": 0.03, + "learning_rate": 0.000786407221664995, + "loss": 2.1375, + "theoretical_loss": 4.8826352831832365, + "tokens_seen": 88539136 + }, + { + "epoch": 0.03, + "learning_rate": 0.0007863911735205617, + "loss": 2.1913, + "theoretical_loss": 4.882128731411745, + "tokens_seen": 88604672 + }, + { + "epoch": 0.03, + "learning_rate": 0.0007863751253761284, + "loss": 2.2191, + "theoretical_loss": 4.881622658989022, + "tokens_seen": 88670208 + }, + { + "epoch": 0.03, + "learning_rate": 0.0007863590772316951, + "loss": 2.1158, + "theoretical_loss": 4.881117065107675, + "tokens_seen": 88735744 + }, + { + "epoch": 0.03, + "learning_rate": 0.0007863430290872618, + "loss": 2.0575, + "theoretical_loss": 4.880611948962271, + "tokens_seen": 88801280 + }, + { + "epoch": 0.03, + "learning_rate": 0.0007863269809428285, + "loss": 2.2024, + "theoretical_loss": 4.880107309749322, + "tokens_seen": 88866816 + }, + { + "epoch": 0.03, + "learning_rate": 0.0007863109327983952, + "loss": 2.166, + "theoretical_loss": 4.879603146667284, + "tokens_seen": 88932352 + }, + { + "epoch": 0.03, + "learning_rate": 0.000786294884653962, + "loss": 2.1703, + "theoretical_loss": 4.879099458916553, + "tokens_seen": 88997888 + }, + { + "epoch": 0.03, + "learning_rate": 0.0007862788365095286, + "loss": 2.0634, + "theoretical_loss": 4.878596245699449, + "tokens_seen": 89063424 + }, + { + "epoch": 0.03, + "learning_rate": 0.0007862627883650953, + "loss": 2.2927, + "theoretical_loss": 4.878093506220221, + "tokens_seen": 89128960 + }, + { + "epoch": 0.03, + "learning_rate": 0.000786246740220662, + "loss": 2.211, + "theoretical_loss": 4.877591239685035, + "tokens_seen": 89194496 + }, + { + "epoch": 0.03, + "learning_rate": 0.0007862306920762287, + "loss": 2.0606, + "theoretical_loss": 4.877089445301969, + "tokens_seen": 89260032 + }, + { + "epoch": 0.03, + "learning_rate": 0.0007862146439317954, + "loss": 2.0376, + "theoretical_loss": 4.876588122281007, + "tokens_seen": 89325568 + }, + { + "epoch": 0.03, + "learning_rate": 0.0007861985957873622, + "loss": 2.1572, + "theoretical_loss": 4.876087269834034, + "tokens_seen": 89391104 + }, + { + "epoch": 0.03, + "learning_rate": 0.0007861825476429288, + "loss": 2.1462, + "theoretical_loss": 4.875586887174828, + "tokens_seen": 89456640 + }, + { + "epoch": 0.03, + "learning_rate": 0.0007861664994984956, + "loss": 2.1742, + "theoretical_loss": 4.875086973519058, + "tokens_seen": 89522176 + }, + { + "epoch": 0.03, + "learning_rate": 0.0007861504513540622, + "loss": 2.1665, + "theoretical_loss": 4.874587528084271, + "tokens_seen": 89587712 + }, + { + "epoch": 0.03, + "learning_rate": 0.000786134403209629, + "loss": 2.143, + "theoretical_loss": 4.874088550089896, + "tokens_seen": 89653248 + }, + { + "epoch": 0.03, + "learning_rate": 0.0007861183550651956, + "loss": 2.1652, + "theoretical_loss": 4.873590038757229, + "tokens_seen": 89718784 + }, + { + "epoch": 0.03, + "learning_rate": 0.0007861023069207624, + "loss": 1.9886, + "theoretical_loss": 4.87309199330943, + "tokens_seen": 89784320 + }, + { + "epoch": 0.03, + "learning_rate": 0.000786086258776329, + "loss": 1.9777, + "theoretical_loss": 4.872594412971521, + "tokens_seen": 89849856 + }, + { + "epoch": 0.03, + "learning_rate": 0.0007860702106318958, + "loss": 2.1703, + "theoretical_loss": 4.872097296970376, + "tokens_seen": 89915392 + }, + { + "epoch": 0.03, + "learning_rate": 0.0007860541624874624, + "loss": 2.1947, + "theoretical_loss": 4.871600644534717, + "tokens_seen": 89980928 + }, + { + "epoch": 0.03, + "learning_rate": 0.0007860381143430292, + "loss": 2.1076, + "theoretical_loss": 4.871104454895106, + "tokens_seen": 90046464 + }, + { + "epoch": 0.03, + "learning_rate": 0.0007860220661985958, + "loss": 2.1245, + "theoretical_loss": 4.870608727283942, + "tokens_seen": 90112000 + }, + { + "epoch": 0.03, + "learning_rate": 0.0007860060180541626, + "loss": 2.4059, + "theoretical_loss": 4.870113460935455, + "tokens_seen": 90177536 + }, + { + "epoch": 0.03, + "learning_rate": 0.0007859899699097292, + "loss": 2.1042, + "theoretical_loss": 4.869618655085697, + "tokens_seen": 90243072 + }, + { + "epoch": 0.03, + "learning_rate": 0.000785973921765296, + "loss": 2.0954, + "theoretical_loss": 4.869124308972543, + "tokens_seen": 90308608 + }, + { + "epoch": 0.03, + "learning_rate": 0.0007859578736208626, + "loss": 2.2113, + "theoretical_loss": 4.868630421835677, + "tokens_seen": 90374144 + }, + { + "epoch": 0.03, + "learning_rate": 0.0007859418254764294, + "loss": 2.2023, + "theoretical_loss": 4.868136992916593, + "tokens_seen": 90439680 + }, + { + "epoch": 0.03, + "learning_rate": 0.000785925777331996, + "loss": 2.2027, + "theoretical_loss": 4.867644021458585, + "tokens_seen": 90505216 + }, + { + "epoch": 0.03, + "learning_rate": 0.0007859097291875628, + "loss": 2.009, + "theoretical_loss": 4.8671515067067475, + "tokens_seen": 90570752 + }, + { + "epoch": 0.03, + "learning_rate": 0.0007858936810431294, + "loss": 2.2347, + "theoretical_loss": 4.866659447907962, + "tokens_seen": 90636288 + }, + { + "epoch": 0.03, + "learning_rate": 0.0007858776328986962, + "loss": 2.0478, + "theoretical_loss": 4.866167844310896, + "tokens_seen": 90701824 + }, + { + "epoch": 0.03, + "learning_rate": 0.0007858615847542628, + "loss": 2.0795, + "theoretical_loss": 4.865676695165996, + "tokens_seen": 90767360 + }, + { + "epoch": 0.03, + "learning_rate": 0.0007858455366098296, + "loss": 2.074, + "theoretical_loss": 4.865185999725487, + "tokens_seen": 90832896 + }, + { + "epoch": 0.03, + "learning_rate": 0.0007858294884653962, + "loss": 2.1698, + "theoretical_loss": 4.864695757243359, + "tokens_seen": 90898432 + }, + { + "epoch": 0.03, + "learning_rate": 0.000785813440320963, + "loss": 2.2419, + "theoretical_loss": 4.864205966975364, + "tokens_seen": 90963968 + }, + { + "epoch": 0.03, + "learning_rate": 0.0007857973921765296, + "loss": 2.066, + "theoretical_loss": 4.8637166281790165, + "tokens_seen": 91029504 + }, + { + "epoch": 0.03, + "learning_rate": 0.0007857813440320964, + "loss": 2.1828, + "theoretical_loss": 4.86322774011358, + "tokens_seen": 91095040 + }, + { + "epoch": 0.03, + "learning_rate": 0.000785765295887663, + "loss": 2.1555, + "theoretical_loss": 4.862739302040067, + "tokens_seen": 91160576 + }, + { + "epoch": 0.03, + "learning_rate": 0.0007857492477432298, + "loss": 2.0443, + "theoretical_loss": 4.862251313221231, + "tokens_seen": 91226112 + }, + { + "epoch": 0.03, + "learning_rate": 0.0007857331995987964, + "loss": 2.2613, + "theoretical_loss": 4.861763772921561, + "tokens_seen": 91291648 + }, + { + "epoch": 0.03, + "learning_rate": 0.0007857171514543632, + "loss": 2.1571, + "theoretical_loss": 4.861276680407279, + "tokens_seen": 91357184 + }, + { + "epoch": 0.03, + "learning_rate": 0.0007857011033099298, + "loss": 2.2144, + "theoretical_loss": 4.860790034946331, + "tokens_seen": 91422720 + }, + { + "epoch": 0.03, + "learning_rate": 0.0007856850551654966, + "loss": 2.1586, + "theoretical_loss": 4.860303835808384, + "tokens_seen": 91488256 + }, + { + "epoch": 0.03, + "learning_rate": 0.0007856690070210632, + "loss": 2.0643, + "theoretical_loss": 4.859818082264822, + "tokens_seen": 91553792 + }, + { + "epoch": 0.03, + "learning_rate": 0.00078565295887663, + "loss": 2.1204, + "theoretical_loss": 4.859332773588736, + "tokens_seen": 91619328 + }, + { + "epoch": 0.03, + "learning_rate": 0.0007856369107321966, + "loss": 2.0912, + "theoretical_loss": 4.858847909054923, + "tokens_seen": 91684864 + }, + { + "epoch": 0.03, + "objective/train/docs_used": 40748, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.9821250438690186, + "objective/train/theoretical_loss": 4.858363487939879, + "objective/train/tokens_used": 112210400, + "theoretical_loss": 4.858363487939879, + "tokens_seen": 91750400 + }, + { + "epoch": 0.03, + "learning_rate": 0.0007856208625877634, + "loss": 2.096, + "theoretical_loss": 4.858363487939879, + "tokens_seen": 91750400 + }, + { + "epoch": 0.03, + "learning_rate": 0.00078560481444333, + "loss": 2.0943, + "theoretical_loss": 4.857879509521796, + "tokens_seen": 91815936 + }, + { + "epoch": 0.03, + "learning_rate": 0.0007855887662988968, + "loss": 2.1188, + "theoretical_loss": 4.857395973080552, + "tokens_seen": 91881472 + }, + { + "epoch": 0.03, + "learning_rate": 0.0007855727181544634, + "loss": 2.169, + "theoretical_loss": 4.856912877897713, + "tokens_seen": 91947008 + }, + { + "epoch": 0.03, + "learning_rate": 0.0007855566700100302, + "loss": 2.1009, + "theoretical_loss": 4.85643022325652, + "tokens_seen": 92012544 + }, + { + "epoch": 0.03, + "learning_rate": 0.0007855406218655968, + "loss": 2.0681, + "theoretical_loss": 4.855948008441891, + "tokens_seen": 92078080 + }, + { + "epoch": 0.03, + "learning_rate": 0.0007855245737211636, + "loss": 2.1438, + "theoretical_loss": 4.85546623274041, + "tokens_seen": 92143616 + }, + { + "epoch": 0.03, + "learning_rate": 0.0007855085255767302, + "loss": 2.031, + "theoretical_loss": 4.854984895440325, + "tokens_seen": 92209152 + }, + { + "epoch": 0.03, + "learning_rate": 0.000785492477432297, + "loss": 2.1225, + "theoretical_loss": 4.854503995831545, + "tokens_seen": 92274688 + }, + { + "epoch": 0.03, + "learning_rate": 0.0007854764292878636, + "loss": 2.2311, + "theoretical_loss": 4.854023533205627, + "tokens_seen": 92340224 + }, + { + "epoch": 0.03, + "learning_rate": 0.0007854603811434303, + "loss": 2.1629, + "theoretical_loss": 4.853543506855784, + "tokens_seen": 92405760 + }, + { + "epoch": 0.03, + "learning_rate": 0.000785444332998997, + "loss": 2.2363, + "theoretical_loss": 4.853063916076865, + "tokens_seen": 92471296 + }, + { + "epoch": 0.03, + "learning_rate": 0.0007854282848545637, + "loss": 2.0171, + "theoretical_loss": 4.852584760165362, + "tokens_seen": 92536832 + }, + { + "epoch": 0.03, + "learning_rate": 0.0007854122367101304, + "loss": 2.0175, + "theoretical_loss": 4.8521060384194, + "tokens_seen": 92602368 + }, + { + "epoch": 0.03, + "learning_rate": 0.0007853961885656971, + "loss": 2.1106, + "theoretical_loss": 4.851627750138731, + "tokens_seen": 92667904 + }, + { + "epoch": 0.03, + "learning_rate": 0.0007853801404212638, + "loss": 2.1537, + "theoretical_loss": 4.851149894624731, + "tokens_seen": 92733440 + }, + { + "epoch": 0.03, + "learning_rate": 0.0007853640922768305, + "loss": 2.1887, + "theoretical_loss": 4.8506724711803955, + "tokens_seen": 92798976 + }, + { + "epoch": 0.03, + "learning_rate": 0.0007853480441323972, + "loss": 1.9478, + "theoretical_loss": 4.850195479110333, + "tokens_seen": 92864512 + }, + { + "epoch": 0.03, + "learning_rate": 0.0007853319959879639, + "loss": 2.1709, + "theoretical_loss": 4.849718917720762, + "tokens_seen": 92930048 + }, + { + "epoch": 0.03, + "learning_rate": 0.0007853159478435306, + "loss": 2.1829, + "theoretical_loss": 4.849242786319503, + "tokens_seen": 92995584 + }, + { + "epoch": 0.03, + "learning_rate": 0.0007852998996990973, + "loss": 2.1172, + "theoretical_loss": 4.848767084215979, + "tokens_seen": 93061120 + }, + { + "epoch": 0.03, + "learning_rate": 0.000785283851554664, + "loss": 2.0612, + "theoretical_loss": 4.848291810721204, + "tokens_seen": 93126656 + }, + { + "epoch": 0.03, + "learning_rate": 0.0007852678034102307, + "loss": 2.091, + "theoretical_loss": 4.847816965147783, + "tokens_seen": 93192192 + }, + { + "epoch": 0.03, + "learning_rate": 0.0007852517552657974, + "loss": 2.1386, + "theoretical_loss": 4.847342546809907, + "tokens_seen": 93257728 + }, + { + "epoch": 0.03, + "learning_rate": 0.0007852357071213641, + "loss": 2.0505, + "theoretical_loss": 4.846868555023345, + "tokens_seen": 93323264 + }, + { + "epoch": 0.03, + "learning_rate": 0.0007852196589769308, + "loss": 2.1446, + "theoretical_loss": 4.846394989105444, + "tokens_seen": 93388800 + }, + { + "epoch": 0.03, + "learning_rate": 0.0007852036108324975, + "loss": 2.1074, + "theoretical_loss": 4.845921848375118, + "tokens_seen": 93454336 + }, + { + "epoch": 0.03, + "learning_rate": 0.0007851875626880642, + "loss": 2.0925, + "theoretical_loss": 4.845449132152851, + "tokens_seen": 93519872 + }, + { + "epoch": 0.03, + "learning_rate": 0.0007851715145436309, + "loss": 2.1368, + "theoretical_loss": 4.844976839760683, + "tokens_seen": 93585408 + }, + { + "epoch": 0.03, + "learning_rate": 0.0007851554663991976, + "loss": 2.0959, + "theoretical_loss": 4.844504970522217, + "tokens_seen": 93650944 + }, + { + "epoch": 0.03, + "learning_rate": 0.0007851394182547643, + "loss": 2.0829, + "theoretical_loss": 4.844033523762603, + "tokens_seen": 93716480 + }, + { + "epoch": 0.03, + "learning_rate": 0.000785123370110331, + "loss": 2.0814, + "theoretical_loss": 4.843562498808538, + "tokens_seen": 93782016 + }, + { + "epoch": 0.03, + "learning_rate": 0.0007851073219658977, + "loss": 2.2578, + "theoretical_loss": 4.843091894988265, + "tokens_seen": 93847552 + }, + { + "epoch": 0.03, + "learning_rate": 0.0007850912738214644, + "loss": 1.9606, + "theoretical_loss": 4.842621711631564, + "tokens_seen": 93913088 + }, + { + "epoch": 0.03, + "learning_rate": 0.0007850752256770311, + "loss": 2.2158, + "theoretical_loss": 4.842151948069745, + "tokens_seen": 93978624 + }, + { + "epoch": 0.03, + "learning_rate": 0.0007850591775325978, + "loss": 2.0042, + "theoretical_loss": 4.84168260363565, + "tokens_seen": 94044160 + }, + { + "epoch": 0.03, + "learning_rate": 0.0007850431293881645, + "loss": 2.0675, + "theoretical_loss": 4.841213677663648, + "tokens_seen": 94109696 + }, + { + "epoch": 0.03, + "learning_rate": 0.0007850270812437312, + "loss": 2.0728, + "theoretical_loss": 4.840745169489621, + "tokens_seen": 94175232 + }, + { + "epoch": 0.03, + "learning_rate": 0.0007850110330992979, + "loss": 2.0633, + "theoretical_loss": 4.840277078450969, + "tokens_seen": 94240768 + }, + { + "epoch": 0.03, + "learning_rate": 0.0007849949849548646, + "loss": 2.1076, + "theoretical_loss": 4.839809403886606, + "tokens_seen": 94306304 + }, + { + "epoch": 0.03, + "learning_rate": 0.0007849789368104313, + "loss": 1.9625, + "theoretical_loss": 4.839342145136947, + "tokens_seen": 94371840 + }, + { + "epoch": 0.03, + "learning_rate": 0.000784962888665998, + "loss": 2.0539, + "theoretical_loss": 4.838875301543913, + "tokens_seen": 94437376 + }, + { + "epoch": 0.03, + "learning_rate": 0.0007849468405215647, + "loss": 2.0606, + "theoretical_loss": 4.838408872450919, + "tokens_seen": 94502912 + }, + { + "epoch": 0.03, + "learning_rate": 0.0007849307923771314, + "loss": 2.169, + "theoretical_loss": 4.837942857202876, + "tokens_seen": 94568448 + }, + { + "epoch": 0.03, + "learning_rate": 0.0007849147442326981, + "loss": 2.0039, + "theoretical_loss": 4.837477255146178, + "tokens_seen": 94633984 + }, + { + "epoch": 0.03, + "learning_rate": 0.0007848986960882648, + "loss": 2.1074, + "theoretical_loss": 4.837012065628711, + "tokens_seen": 94699520 + }, + { + "epoch": 0.03, + "learning_rate": 0.0007848826479438315, + "loss": 1.968, + "theoretical_loss": 4.836547287999833, + "tokens_seen": 94765056 + }, + { + "epoch": 0.03, + "learning_rate": 0.0007848665997993982, + "loss": 2.0423, + "theoretical_loss": 4.836082921610381, + "tokens_seen": 94830592 + }, + { + "epoch": 0.03, + "learning_rate": 0.0007848505516549649, + "loss": 2.0842, + "theoretical_loss": 4.835618965812664, + "tokens_seen": 94896128 + }, + { + "epoch": 0.03, + "learning_rate": 0.0007848345035105317, + "loss": 2.1007, + "theoretical_loss": 4.835155419960453, + "tokens_seen": 94961664 + }, + { + "epoch": 0.03, + "objective/train/docs_used": 42039, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.9016162157058716, + "objective/train/theoretical_loss": 4.8346922834089865, + "objective/train/tokens_used": 115487200, + "theoretical_loss": 4.8346922834089865, + "tokens_seen": 95027200 + }, + { + "epoch": 0.03, + "learning_rate": 0.0007848184553660983, + "loss": 2.0037, + "theoretical_loss": 4.8346922834089865, + "tokens_seen": 95027200 + }, + { + "epoch": 0.03, + "learning_rate": 0.000784802407221665, + "loss": 2.0444, + "theoretical_loss": 4.834229555514956, + "tokens_seen": 95092736 + }, + { + "epoch": 0.03, + "learning_rate": 0.0007847863590772317, + "loss": 2.0442, + "theoretical_loss": 4.833767235636511, + "tokens_seen": 95158272 + }, + { + "epoch": 0.03, + "learning_rate": 0.0007847703109327985, + "loss": 2.0389, + "theoretical_loss": 4.833305323133246, + "tokens_seen": 95223808 + }, + { + "epoch": 0.03, + "learning_rate": 0.0007847542627883651, + "loss": 2.1117, + "theoretical_loss": 4.832843817366204, + "tokens_seen": 95289344 + }, + { + "epoch": 0.03, + "learning_rate": 0.0007847382146439319, + "loss": 2.0866, + "theoretical_loss": 4.832382717697866, + "tokens_seen": 95354880 + }, + { + "epoch": 0.03, + "learning_rate": 0.0007847221664994985, + "loss": 2.0999, + "theoretical_loss": 4.831922023492153, + "tokens_seen": 95420416 + }, + { + "epoch": 0.03, + "learning_rate": 0.0007847061183550653, + "loss": 2.081, + "theoretical_loss": 4.831461734114415, + "tokens_seen": 95485952 + }, + { + "epoch": 0.03, + "learning_rate": 0.0007846900702106319, + "loss": 2.1345, + "theoretical_loss": 4.831001848931429, + "tokens_seen": 95551488 + }, + { + "epoch": 0.03, + "learning_rate": 0.0007846740220661987, + "loss": 2.109, + "theoretical_loss": 4.830542367311402, + "tokens_seen": 95617024 + }, + { + "epoch": 0.03, + "learning_rate": 0.0007846579739217653, + "loss": 2.1716, + "theoretical_loss": 4.830083288623955, + "tokens_seen": 95682560 + }, + { + "epoch": 0.03, + "learning_rate": 0.0007846419257773321, + "loss": 2.0623, + "theoretical_loss": 4.829624612240126, + "tokens_seen": 95748096 + }, + { + "epoch": 0.03, + "learning_rate": 0.0007846258776328987, + "loss": 2.1428, + "theoretical_loss": 4.829166337532367, + "tokens_seen": 95813632 + }, + { + "epoch": 0.03, + "learning_rate": 0.0007846098294884655, + "loss": 2.0581, + "theoretical_loss": 4.828708463874532, + "tokens_seen": 95879168 + }, + { + "epoch": 0.03, + "learning_rate": 0.0007845937813440321, + "loss": 2.1188, + "theoretical_loss": 4.828250990641886, + "tokens_seen": 95944704 + }, + { + "epoch": 0.03, + "learning_rate": 0.0007845777331995989, + "loss": 2.0968, + "theoretical_loss": 4.827793917211083, + "tokens_seen": 96010240 + }, + { + "epoch": 0.03, + "learning_rate": 0.0007845616850551655, + "loss": 2.1534, + "theoretical_loss": 4.827337242960181, + "tokens_seen": 96075776 + }, + { + "epoch": 0.03, + "learning_rate": 0.0007845456369107323, + "loss": 1.9293, + "theoretical_loss": 4.8268809672686235, + "tokens_seen": 96141312 + }, + { + "epoch": 0.03, + "learning_rate": 0.0007845295887662989, + "loss": 1.9851, + "theoretical_loss": 4.826425089517245, + "tokens_seen": 96206848 + }, + { + "epoch": 0.03, + "learning_rate": 0.0007845135406218657, + "loss": 2.0389, + "theoretical_loss": 4.825969609088258, + "tokens_seen": 96272384 + }, + { + "epoch": 0.03, + "learning_rate": 0.0007844974924774323, + "loss": 2.1921, + "theoretical_loss": 4.825514525365258, + "tokens_seen": 96337920 + }, + { + "epoch": 0.03, + "learning_rate": 0.0007844814443329991, + "loss": 2.0471, + "theoretical_loss": 4.825059837733214, + "tokens_seen": 96403456 + }, + { + "epoch": 0.03, + "learning_rate": 0.0007844653961885657, + "loss": 2.0548, + "theoretical_loss": 4.824605545578463, + "tokens_seen": 96468992 + }, + { + "epoch": 0.03, + "learning_rate": 0.0007844493480441325, + "loss": 2.0559, + "theoretical_loss": 4.824151648288714, + "tokens_seen": 96534528 + }, + { + "epoch": 0.03, + "learning_rate": 0.0007844332998996991, + "loss": 2.0222, + "theoretical_loss": 4.823698145253035, + "tokens_seen": 96600064 + }, + { + "epoch": 0.03, + "learning_rate": 0.0007844172517552659, + "loss": 2.1776, + "theoretical_loss": 4.823245035861852, + "tokens_seen": 96665600 + }, + { + "epoch": 0.03, + "learning_rate": 0.0007844012036108325, + "loss": 2.1387, + "theoretical_loss": 4.822792319506952, + "tokens_seen": 96731136 + }, + { + "epoch": 0.03, + "learning_rate": 0.0007843851554663993, + "loss": 2.0619, + "theoretical_loss": 4.822339995581466, + "tokens_seen": 96796672 + }, + { + "epoch": 0.03, + "learning_rate": 0.000784369107321966, + "loss": 2.0424, + "theoretical_loss": 4.821888063479874, + "tokens_seen": 96862208 + }, + { + "epoch": 0.03, + "learning_rate": 0.0007843530591775327, + "loss": 2.1836, + "theoretical_loss": 4.8214365225980025, + "tokens_seen": 96927744 + }, + { + "epoch": 0.03, + "learning_rate": 0.0007843370110330994, + "loss": 2.0611, + "theoretical_loss": 4.8209853723330145, + "tokens_seen": 96993280 + }, + { + "epoch": 0.03, + "learning_rate": 0.0007843209628886661, + "loss": 1.9802, + "theoretical_loss": 4.820534612083408, + "tokens_seen": 97058816 + }, + { + "epoch": 0.03, + "learning_rate": 0.0007843049147442328, + "loss": 2.1303, + "theoretical_loss": 4.820084241249013, + "tokens_seen": 97124352 + }, + { + "epoch": 0.03, + "learning_rate": 0.0007842888665997995, + "loss": 2.0336, + "theoretical_loss": 4.819634259230991, + "tokens_seen": 97189888 + }, + { + "epoch": 0.03, + "learning_rate": 0.0007842728184553662, + "loss": 1.9678, + "theoretical_loss": 4.819184665431821, + "tokens_seen": 97255424 + }, + { + "epoch": 0.03, + "learning_rate": 0.0007842567703109329, + "loss": 1.9587, + "theoretical_loss": 4.818735459255306, + "tokens_seen": 97320960 + }, + { + "epoch": 0.03, + "learning_rate": 0.0007842407221664996, + "loss": 2.1212, + "theoretical_loss": 4.818286640106564, + "tokens_seen": 97386496 + }, + { + "epoch": 0.03, + "learning_rate": 0.0007842246740220663, + "loss": 2.274, + "theoretical_loss": 4.817838207392027, + "tokens_seen": 97452032 + }, + { + "epoch": 0.03, + "learning_rate": 0.0007842086258776329, + "loss": 2.1901, + "theoretical_loss": 4.817390160519435, + "tokens_seen": 97517568 + }, + { + "epoch": 0.03, + "learning_rate": 0.0007841925777331997, + "loss": 2.2077, + "theoretical_loss": 4.816942498897831, + "tokens_seen": 97583104 + }, + { + "epoch": 0.03, + "learning_rate": 0.0007841765295887663, + "loss": 2.3655, + "theoretical_loss": 4.816495221937563, + "tokens_seen": 97648640 + }, + { + "epoch": 0.03, + "learning_rate": 0.0007841604814443331, + "loss": 2.1031, + "theoretical_loss": 4.816048329050273, + "tokens_seen": 97714176 + }, + { + "epoch": 0.03, + "learning_rate": 0.0007841444332998997, + "loss": 2.163, + "theoretical_loss": 4.815601819648897, + "tokens_seen": 97779712 + }, + { + "epoch": 0.03, + "learning_rate": 0.0007841283851554665, + "loss": 2.001, + "theoretical_loss": 4.815155693147666, + "tokens_seen": 97845248 + }, + { + "epoch": 0.03, + "learning_rate": 0.0007841123370110331, + "loss": 2.0919, + "theoretical_loss": 4.81470994896209, + "tokens_seen": 97910784 + }, + { + "epoch": 0.03, + "learning_rate": 0.0007840962888665999, + "loss": 2.1124, + "theoretical_loss": 4.8142645865089655, + "tokens_seen": 97976320 + }, + { + "epoch": 0.03, + "learning_rate": 0.0007840802407221665, + "loss": 2.0964, + "theoretical_loss": 4.813819605206369, + "tokens_seen": 98041856 + }, + { + "epoch": 0.03, + "learning_rate": 0.0007840641925777332, + "loss": 2.0213, + "theoretical_loss": 4.813375004473651, + "tokens_seen": 98107392 + }, + { + "epoch": 0.03, + "learning_rate": 0.0007840481444332999, + "loss": 1.9749, + "theoretical_loss": 4.812930783731433, + "tokens_seen": 98172928 + }, + { + "epoch": 0.03, + "learning_rate": 0.0007840320962888666, + "loss": 1.9254, + "theoretical_loss": 4.8124869424016055, + "tokens_seen": 98238464 + }, + { + "epoch": 0.03, + "objective/train/docs_used": 43181, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 2.0896377563476562, + "objective/train/theoretical_loss": 4.812043479907323, + "objective/train/tokens_used": 118764000, + "theoretical_loss": 4.812043479907323, + "tokens_seen": 98304000 + }, + { + "epoch": 0.03, + "learning_rate": 0.0007840160481444333, + "loss": 2.0893, + "theoretical_loss": 4.812043479907323, + "tokens_seen": 98304000 + }, + { + "epoch": 0.03, + "learning_rate": 0.000784, + "loss": 2.0275, + "theoretical_loss": 4.8116003956730005, + "tokens_seen": 98369536 + }, + { + "epoch": 0.03, + "learning_rate": 0.0007839839518555667, + "loss": 2.0681, + "theoretical_loss": 4.811157689124311, + "tokens_seen": 98435072 + }, + { + "epoch": 0.03, + "learning_rate": 0.0007839679037111334, + "loss": 1.9257, + "theoretical_loss": 4.8107153596881815, + "tokens_seen": 98500608 + }, + { + "epoch": 0.03, + "learning_rate": 0.0007839518555667001, + "loss": 1.949, + "theoretical_loss": 4.810273406792788, + "tokens_seen": 98566144 + }, + { + "epoch": 0.03, + "learning_rate": 0.0007839358074222668, + "loss": 2.0623, + "theoretical_loss": 4.809831829867555, + "tokens_seen": 98631680 + }, + { + "epoch": 0.03, + "learning_rate": 0.0007839197592778335, + "loss": 2.1768, + "theoretical_loss": 4.809390628343146, + "tokens_seen": 98697216 + }, + { + "epoch": 0.03, + "learning_rate": 0.0007839037111334002, + "loss": 2.0867, + "theoretical_loss": 4.808949801651468, + "tokens_seen": 98762752 + }, + { + "epoch": 0.03, + "learning_rate": 0.0007838876629889669, + "loss": 2.0713, + "theoretical_loss": 4.8085093492256625, + "tokens_seen": 98828288 + }, + { + "epoch": 0.03, + "learning_rate": 0.0007838716148445336, + "loss": 1.9426, + "theoretical_loss": 4.808069270500102, + "tokens_seen": 98893824 + }, + { + "epoch": 0.03, + "learning_rate": 0.0007838555667001003, + "loss": 1.9985, + "theoretical_loss": 4.80762956491039, + "tokens_seen": 98959360 + }, + { + "epoch": 0.03, + "learning_rate": 0.000783839518555667, + "loss": 2.106, + "theoretical_loss": 4.807190231893355, + "tokens_seen": 99024896 + }, + { + "epoch": 0.03, + "learning_rate": 0.0007838234704112337, + "loss": 2.1602, + "theoretical_loss": 4.806751270887048, + "tokens_seen": 99090432 + }, + { + "epoch": 0.03, + "learning_rate": 0.0007838074222668004, + "loss": 2.0577, + "theoretical_loss": 4.806312681330736, + "tokens_seen": 99155968 + }, + { + "epoch": 0.03, + "learning_rate": 0.0007837913741223671, + "loss": 2.0089, + "theoretical_loss": 4.805874462664903, + "tokens_seen": 99221504 + }, + { + "epoch": 0.03, + "learning_rate": 0.0007837753259779338, + "loss": 2.087, + "theoretical_loss": 4.805436614331247, + "tokens_seen": 99287040 + }, + { + "epoch": 0.03, + "learning_rate": 0.0007837592778335005, + "loss": 2.1312, + "theoretical_loss": 4.804999135772668, + "tokens_seen": 99352576 + }, + { + "epoch": 0.03, + "learning_rate": 0.0007837432296890672, + "loss": 2.1462, + "theoretical_loss": 4.804562026433278, + "tokens_seen": 99418112 + }, + { + "epoch": 0.03, + "learning_rate": 0.0007837271815446339, + "loss": 2.07, + "theoretical_loss": 4.804125285758383, + "tokens_seen": 99483648 + }, + { + "epoch": 0.03, + "learning_rate": 0.0007837111334002006, + "loss": 2.0889, + "theoretical_loss": 4.803688913194495, + "tokens_seen": 99549184 + }, + { + "epoch": 0.03, + "learning_rate": 0.0007836950852557673, + "loss": 2.057, + "theoretical_loss": 4.8032529081893145, + "tokens_seen": 99614720 + }, + { + "epoch": 0.03, + "learning_rate": 0.000783679037111334, + "loss": 2.0492, + "theoretical_loss": 4.802817270191736, + "tokens_seen": 99680256 + }, + { + "epoch": 0.03, + "learning_rate": 0.0007836629889669007, + "loss": 2.0618, + "theoretical_loss": 4.802381998651841, + "tokens_seen": 99745792 + }, + { + "epoch": 0.03, + "learning_rate": 0.0007836469408224674, + "loss": 1.9796, + "theoretical_loss": 4.801947093020896, + "tokens_seen": 99811328 + }, + { + "epoch": 0.03, + "learning_rate": 0.0007836308926780341, + "loss": 2.2068, + "theoretical_loss": 4.801512552751349, + "tokens_seen": 99876864 + }, + { + "epoch": 0.03, + "learning_rate": 0.0007836148445336008, + "loss": 1.9506, + "theoretical_loss": 4.801078377296827, + "tokens_seen": 99942400 + }, + { + "epoch": 0.03, + "learning_rate": 0.0007835987963891675, + "loss": 2.0037, + "theoretical_loss": 4.800644566112128, + "tokens_seen": 100007936 + }, + { + "epoch": 0.03, + "learning_rate": 0.0007835827482447342, + "loss": 2.2357, + "theoretical_loss": 4.800211118653225, + "tokens_seen": 100073472 + }, + { + "epoch": 0.03, + "learning_rate": 0.000783566700100301, + "loss": 2.0903, + "theoretical_loss": 4.799778034377258, + "tokens_seen": 100139008 + }, + { + "epoch": 0.03, + "learning_rate": 0.0007835506519558676, + "loss": 2.0861, + "theoretical_loss": 4.799345312742531, + "tokens_seen": 100204544 + }, + { + "epoch": 0.03, + "learning_rate": 0.0007835346038114343, + "loss": 2.0738, + "theoretical_loss": 4.798912953208511, + "tokens_seen": 100270080 + }, + { + "epoch": 0.03, + "learning_rate": 0.000783518555667001, + "loss": 2.0545, + "theoretical_loss": 4.798480955235822, + "tokens_seen": 100335616 + }, + { + "epoch": 0.03, + "learning_rate": 0.0007835025075225677, + "loss": 2.2573, + "theoretical_loss": 4.798049318286243, + "tokens_seen": 100401152 + }, + { + "epoch": 0.03, + "learning_rate": 0.0007834864593781344, + "loss": 2.1606, + "theoretical_loss": 4.797618041822707, + "tokens_seen": 100466688 + }, + { + "epoch": 0.03, + "learning_rate": 0.0007834704112337011, + "loss": 1.9262, + "theoretical_loss": 4.797187125309293, + "tokens_seen": 100532224 + }, + { + "epoch": 0.03, + "learning_rate": 0.0007834543630892678, + "loss": 2.0738, + "theoretical_loss": 4.796756568211227, + "tokens_seen": 100597760 + }, + { + "epoch": 0.03, + "learning_rate": 0.0007834383149448346, + "loss": 1.9877, + "theoretical_loss": 4.796326369994876, + "tokens_seen": 100663296 + }, + { + "epoch": 0.03, + "learning_rate": 0.0007834222668004012, + "loss": 2.034, + "theoretical_loss": 4.795896530127747, + "tokens_seen": 100728832 + }, + { + "epoch": 0.03, + "learning_rate": 0.000783406218655968, + "loss": 2.1454, + "theoretical_loss": 4.795467048078484, + "tokens_seen": 100794368 + }, + { + "epoch": 0.03, + "learning_rate": 0.0007833901705115346, + "loss": 1.9609, + "theoretical_loss": 4.79503792331686, + "tokens_seen": 100859904 + }, + { + "epoch": 0.03, + "learning_rate": 0.0007833741223671014, + "loss": 2.082, + "theoretical_loss": 4.794609155313782, + "tokens_seen": 100925440 + }, + { + "epoch": 0.03, + "learning_rate": 0.000783358074222668, + "loss": 2.0242, + "theoretical_loss": 4.79418074354128, + "tokens_seen": 100990976 + }, + { + "epoch": 0.03, + "learning_rate": 0.0007833420260782348, + "loss": 2.1416, + "theoretical_loss": 4.79375268747251, + "tokens_seen": 101056512 + }, + { + "epoch": 0.03, + "learning_rate": 0.0007833259779338014, + "loss": 2.2108, + "theoretical_loss": 4.793324986581746, + "tokens_seen": 101122048 + }, + { + "epoch": 0.03, + "learning_rate": 0.0007833099297893682, + "loss": 1.9895, + "theoretical_loss": 4.79289764034438, + "tokens_seen": 101187584 + }, + { + "epoch": 0.03, + "learning_rate": 0.0007832938816449348, + "loss": 1.9549, + "theoretical_loss": 4.792470648236918, + "tokens_seen": 101253120 + }, + { + "epoch": 0.03, + "learning_rate": 0.0007832778335005016, + "loss": 2.0135, + "theoretical_loss": 4.792044009736978, + "tokens_seen": 101318656 + }, + { + "epoch": 0.03, + "learning_rate": 0.0007832617853560682, + "loss": 2.0656, + "theoretical_loss": 4.791617724323283, + "tokens_seen": 101384192 + }, + { + "epoch": 0.03, + "learning_rate": 0.000783245737211635, + "loss": 2.0866, + "theoretical_loss": 4.791191791475663, + "tokens_seen": 101449728 + }, + { + "epoch": 0.03, + "learning_rate": 0.0007832296890672016, + "loss": 2.0476, + "theoretical_loss": 4.79076621067505, + "tokens_seen": 101515264 + }, + { + "epoch": 0.03, + "objective/train/docs_used": 44310, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 2.0381083488464355, + "objective/train/theoretical_loss": 4.790340981403474, + "objective/train/tokens_used": 122040800, + "theoretical_loss": 4.790340981403474, + "tokens_seen": 101580800 + }, + { + "epoch": 0.03, + "learning_rate": 0.0007832136409227684, + "loss": 1.9943, + "theoretical_loss": 4.790340981403474, + "tokens_seen": 101580800 + }, + { + "epoch": 0.03, + "learning_rate": 0.000783197592778335, + "loss": 2.0594, + "theoretical_loss": 4.7899161031440585, + "tokens_seen": 101646336 + }, + { + "epoch": 0.03, + "learning_rate": 0.0007831815446339018, + "loss": 2.0189, + "theoretical_loss": 4.7894915753810245, + "tokens_seen": 101711872 + }, + { + "epoch": 0.03, + "learning_rate": 0.0007831654964894685, + "loss": 1.9533, + "theoretical_loss": 4.789067397599677, + "tokens_seen": 101777408 + }, + { + "epoch": 0.03, + "learning_rate": 0.0007831494483450352, + "loss": 2.0695, + "theoretical_loss": 4.788643569286411, + "tokens_seen": 101842944 + }, + { + "epoch": 0.03, + "learning_rate": 0.0007831334002006019, + "loss": 1.9675, + "theoretical_loss": 4.788220089928705, + "tokens_seen": 101908480 + }, + { + "epoch": 0.03, + "learning_rate": 0.0007831173520561686, + "loss": 1.9911, + "theoretical_loss": 4.787796959015118, + "tokens_seen": 101974016 + }, + { + "epoch": 0.03, + "learning_rate": 0.0007831013039117353, + "loss": 2.0552, + "theoretical_loss": 4.787374176035287, + "tokens_seen": 102039552 + }, + { + "epoch": 0.03, + "learning_rate": 0.000783085255767302, + "loss": 2.1802, + "theoretical_loss": 4.78695174047992, + "tokens_seen": 102105088 + }, + { + "epoch": 0.03, + "learning_rate": 0.0007830692076228687, + "loss": 2.0938, + "theoretical_loss": 4.786529651840802, + "tokens_seen": 102170624 + }, + { + "epoch": 0.03, + "learning_rate": 0.0007830531594784354, + "loss": 1.9551, + "theoretical_loss": 4.786107909610784, + "tokens_seen": 102236160 + }, + { + "epoch": 0.03, + "learning_rate": 0.0007830371113340021, + "loss": 1.943, + "theoretical_loss": 4.785686513283783, + "tokens_seen": 102301696 + }, + { + "epoch": 0.03, + "learning_rate": 0.0007830210631895688, + "loss": 1.9479, + "theoretical_loss": 4.785265462354779, + "tokens_seen": 102367232 + }, + { + "epoch": 0.03, + "learning_rate": 0.0007830050150451355, + "loss": 1.9354, + "theoretical_loss": 4.7848447563198135, + "tokens_seen": 102432768 + }, + { + "epoch": 0.03, + "learning_rate": 0.0007829889669007022, + "loss": 2.0428, + "theoretical_loss": 4.784424394675983, + "tokens_seen": 102498304 + }, + { + "epoch": 0.03, + "learning_rate": 0.0007829729187562689, + "loss": 2.0708, + "theoretical_loss": 4.784004376921439, + "tokens_seen": 102563840 + }, + { + "epoch": 0.03, + "learning_rate": 0.0007829568706118356, + "loss": 1.9981, + "theoretical_loss": 4.783584702555387, + "tokens_seen": 102629376 + }, + { + "epoch": 0.03, + "learning_rate": 0.0007829408224674023, + "loss": 2.0329, + "theoretical_loss": 4.783165371078075, + "tokens_seen": 102694912 + }, + { + "epoch": 0.03, + "learning_rate": 0.000782924774322969, + "loss": 2.0085, + "theoretical_loss": 4.782746381990803, + "tokens_seen": 102760448 + }, + { + "epoch": 0.03, + "learning_rate": 0.0007829087261785357, + "loss": 2.1502, + "theoretical_loss": 4.78232773479591, + "tokens_seen": 102825984 + }, + { + "epoch": 0.03, + "learning_rate": 0.0007828926780341024, + "loss": 2.1084, + "theoretical_loss": 4.7819094289967765, + "tokens_seen": 102891520 + }, + { + "epoch": 0.03, + "learning_rate": 0.0007828766298896691, + "loss": 2.037, + "theoretical_loss": 4.781491464097819, + "tokens_seen": 102957056 + }, + { + "epoch": 0.03, + "learning_rate": 0.0007828605817452358, + "loss": 1.9188, + "theoretical_loss": 4.781073839604488, + "tokens_seen": 103022592 + }, + { + "epoch": 0.03, + "learning_rate": 0.0007828445336008025, + "loss": 2.0247, + "theoretical_loss": 4.780656555023268, + "tokens_seen": 103088128 + }, + { + "epoch": 0.03, + "learning_rate": 0.0007828284854563692, + "loss": 2.0653, + "theoretical_loss": 4.780239609861669, + "tokens_seen": 103153664 + }, + { + "epoch": 0.03, + "learning_rate": 0.0007828124373119358, + "loss": 2.0414, + "theoretical_loss": 4.77982300362823, + "tokens_seen": 103219200 + }, + { + "epoch": 0.03, + "learning_rate": 0.0007827963891675026, + "loss": 2.0633, + "theoretical_loss": 4.779406735832509, + "tokens_seen": 103284736 + }, + { + "epoch": 0.03, + "learning_rate": 0.0007827803410230692, + "loss": 2.0903, + "theoretical_loss": 4.778990805985089, + "tokens_seen": 103350272 + }, + { + "epoch": 0.03, + "learning_rate": 0.000782764292878636, + "loss": 1.9622, + "theoretical_loss": 4.778575213597568, + "tokens_seen": 103415808 + }, + { + "epoch": 0.03, + "learning_rate": 0.0007827482447342026, + "loss": 2.0441, + "theoretical_loss": 4.778159958182557, + "tokens_seen": 103481344 + }, + { + "epoch": 0.03, + "learning_rate": 0.0007827321965897694, + "loss": 1.9103, + "theoretical_loss": 4.777745039253682, + "tokens_seen": 103546880 + }, + { + "epoch": 0.03, + "learning_rate": 0.000782716148445336, + "loss": 1.9971, + "theoretical_loss": 4.777330456325579, + "tokens_seen": 103612416 + }, + { + "epoch": 0.03, + "learning_rate": 0.0007827001003009028, + "loss": 2.047, + "theoretical_loss": 4.776916208913887, + "tokens_seen": 103677952 + }, + { + "epoch": 0.03, + "learning_rate": 0.0007826840521564694, + "loss": 2.0709, + "theoretical_loss": 4.776502296535252, + "tokens_seen": 103743488 + }, + { + "epoch": 0.03, + "learning_rate": 0.0007826680040120361, + "loss": 2.0027, + "theoretical_loss": 4.776088718707317, + "tokens_seen": 103809024 + }, + { + "epoch": 0.03, + "learning_rate": 0.0007826519558676028, + "loss": 1.9611, + "theoretical_loss": 4.77567547494873, + "tokens_seen": 103874560 + }, + { + "epoch": 0.03, + "learning_rate": 0.0007826359077231695, + "loss": 2.1213, + "theoretical_loss": 4.775262564779128, + "tokens_seen": 103940096 + }, + { + "epoch": 0.03, + "learning_rate": 0.0007826198595787362, + "loss": 2.1656, + "theoretical_loss": 4.774849987719145, + "tokens_seen": 104005632 + }, + { + "epoch": 0.03, + "learning_rate": 0.0007826038114343029, + "loss": 2.1632, + "theoretical_loss": 4.774437743290404, + "tokens_seen": 104071168 + }, + { + "epoch": 0.03, + "learning_rate": 0.0007825877632898696, + "loss": 2.091, + "theoretical_loss": 4.774025831015517, + "tokens_seen": 104136704 + }, + { + "epoch": 0.03, + "learning_rate": 0.0007825717151454363, + "loss": 1.8857, + "theoretical_loss": 4.7736142504180785, + "tokens_seen": 104202240 + }, + { + "epoch": 0.03, + "learning_rate": 0.000782555667001003, + "loss": 1.8357, + "theoretical_loss": 4.773203001022668, + "tokens_seen": 104267776 + }, + { + "epoch": 0.03, + "learning_rate": 0.0007825396188565697, + "loss": 2.0957, + "theoretical_loss": 4.772792082354843, + "tokens_seen": 104333312 + }, + { + "epoch": 0.03, + "learning_rate": 0.0007825235707121364, + "loss": 2.05, + "theoretical_loss": 4.77238149394114, + "tokens_seen": 104398848 + }, + { + "epoch": 0.03, + "learning_rate": 0.0007825075225677031, + "loss": 2.0812, + "theoretical_loss": 4.771971235309065, + "tokens_seen": 104464384 + }, + { + "epoch": 0.03, + "learning_rate": 0.0007824914744232698, + "loss": 1.9862, + "theoretical_loss": 4.771561305987104, + "tokens_seen": 104529920 + }, + { + "epoch": 0.03, + "learning_rate": 0.0007824754262788365, + "loss": 2.0809, + "theoretical_loss": 4.771151705504707, + "tokens_seen": 104595456 + }, + { + "epoch": 0.03, + "learning_rate": 0.0007824593781344032, + "loss": 2.0564, + "theoretical_loss": 4.770742433392289, + "tokens_seen": 104660992 + }, + { + "epoch": 0.03, + "learning_rate": 0.0007824433299899699, + "loss": 1.9421, + "theoretical_loss": 4.770333489181233, + "tokens_seen": 104726528 + }, + { + "epoch": 0.03, + "learning_rate": 0.0007824272818455366, + "loss": 1.9097, + "theoretical_loss": 4.769924872403883, + "tokens_seen": 104792064 + }, + { + "epoch": 0.03, + "objective/train/docs_used": 45444, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 2.1839513778686523, + "objective/train/theoretical_loss": 4.769516582593539, + "objective/train/tokens_used": 125317600, + "theoretical_loss": 4.769516582593539, + "tokens_seen": 104857600 + }, + { + "epoch": 0.03, + "learning_rate": 0.0007824112337011033, + "loss": 2.0834, + "theoretical_loss": 4.769516582593539, + "tokens_seen": 104857600 + }, + { + "epoch": 0.03, + "learning_rate": 0.00078239518555667, + "loss": 1.9976, + "theoretical_loss": 4.769108619284461, + "tokens_seen": 104923136 + }, + { + "epoch": 0.03, + "learning_rate": 0.0007823791374122367, + "loss": 1.8708, + "theoretical_loss": 4.768700982011859, + "tokens_seen": 104988672 + }, + { + "epoch": 0.03, + "learning_rate": 0.0007823630892678034, + "loss": 2.0233, + "theoretical_loss": 4.768293670311899, + "tokens_seen": 105054208 + }, + { + "epoch": 0.03, + "learning_rate": 0.0007823470411233701, + "loss": 1.9331, + "theoretical_loss": 4.7678866837216916, + "tokens_seen": 105119744 + }, + { + "epoch": 0.03, + "learning_rate": 0.0007823309929789368, + "loss": 2.0091, + "theoretical_loss": 4.767480021779294, + "tokens_seen": 105185280 + }, + { + "epoch": 0.03, + "learning_rate": 0.0007823149448345035, + "loss": 1.9834, + "theoretical_loss": 4.76707368402371, + "tokens_seen": 105250816 + }, + { + "epoch": 0.03, + "learning_rate": 0.0007822988966900702, + "loss": 1.9807, + "theoretical_loss": 4.766667669994881, + "tokens_seen": 105316352 + }, + { + "epoch": 0.03, + "learning_rate": 0.0007822828485456369, + "loss": 1.9751, + "theoretical_loss": 4.76626197923369, + "tokens_seen": 105381888 + }, + { + "epoch": 0.03, + "learning_rate": 0.0007822668004012036, + "loss": 1.9962, + "theoretical_loss": 4.765856611281953, + "tokens_seen": 105447424 + }, + { + "epoch": 0.03, + "learning_rate": 0.0007822507522567703, + "loss": 1.9854, + "theoretical_loss": 4.765451565682423, + "tokens_seen": 105512960 + }, + { + "epoch": 0.03, + "learning_rate": 0.000782234704112337, + "loss": 2.0864, + "theoretical_loss": 4.765046841978783, + "tokens_seen": 105578496 + }, + { + "epoch": 0.03, + "learning_rate": 0.0007822186559679037, + "loss": 1.9781, + "theoretical_loss": 4.764642439715642, + "tokens_seen": 105644032 + }, + { + "epoch": 0.03, + "learning_rate": 0.0007822026078234704, + "loss": 2.2189, + "theoretical_loss": 4.764238358438539, + "tokens_seen": 105709568 + }, + { + "epoch": 0.03, + "learning_rate": 0.0007821865596790371, + "loss": 2.0901, + "theoretical_loss": 4.763834597693935, + "tokens_seen": 105775104 + }, + { + "epoch": 0.03, + "learning_rate": 0.0007821705115346038, + "loss": 1.9416, + "theoretical_loss": 4.763431157029213, + "tokens_seen": 105840640 + }, + { + "epoch": 0.03, + "learning_rate": 0.0007821544633901705, + "loss": 1.9316, + "theoretical_loss": 4.763028035992675, + "tokens_seen": 105906176 + }, + { + "epoch": 0.03, + "learning_rate": 0.0007821384152457372, + "loss": 2.1075, + "theoretical_loss": 4.762625234133537, + "tokens_seen": 105971712 + }, + { + "epoch": 0.03, + "learning_rate": 0.0007821223671013039, + "loss": 1.9779, + "theoretical_loss": 4.762222751001933, + "tokens_seen": 106037248 + }, + { + "epoch": 0.03, + "learning_rate": 0.0007821063189568706, + "loss": 1.9221, + "theoretical_loss": 4.761820586148906, + "tokens_seen": 106102784 + }, + { + "epoch": 0.03, + "learning_rate": 0.0007820902708124373, + "loss": 2.0444, + "theoretical_loss": 4.761418739126407, + "tokens_seen": 106168320 + }, + { + "epoch": 0.03, + "learning_rate": 0.000782074222668004, + "loss": 1.9233, + "theoretical_loss": 4.761017209487298, + "tokens_seen": 106233856 + }, + { + "epoch": 0.03, + "learning_rate": 0.0007820581745235707, + "loss": 2.0154, + "theoretical_loss": 4.760615996785342, + "tokens_seen": 106299392 + }, + { + "epoch": 0.03, + "learning_rate": 0.0007820421263791375, + "loss": 2.0621, + "theoretical_loss": 4.760215100575204, + "tokens_seen": 106364928 + }, + { + "epoch": 0.03, + "learning_rate": 0.0007820260782347041, + "loss": 2.0451, + "theoretical_loss": 4.759814520412451, + "tokens_seen": 106430464 + }, + { + "epoch": 0.03, + "learning_rate": 0.0007820100300902709, + "loss": 1.997, + "theoretical_loss": 4.759414255853547, + "tokens_seen": 106496000 + }, + { + "epoch": 0.03, + "learning_rate": 0.0007819939819458375, + "loss": 2.0659, + "theoretical_loss": 4.759014306455848, + "tokens_seen": 106561536 + }, + { + "epoch": 0.03, + "learning_rate": 0.0007819779338014043, + "loss": 1.9623, + "theoretical_loss": 4.758614671777604, + "tokens_seen": 106627072 + }, + { + "epoch": 0.03, + "learning_rate": 0.000781961885656971, + "loss": 1.9956, + "theoretical_loss": 4.758215351377956, + "tokens_seen": 106692608 + }, + { + "epoch": 0.03, + "learning_rate": 0.0007819458375125377, + "loss": 1.939, + "theoretical_loss": 4.757816344816933, + "tokens_seen": 106758144 + }, + { + "epoch": 0.03, + "learning_rate": 0.0007819297893681044, + "loss": 2.0288, + "theoretical_loss": 4.75741765165545, + "tokens_seen": 106823680 + }, + { + "epoch": 0.03, + "learning_rate": 0.0007819137412236711, + "loss": 2.0478, + "theoretical_loss": 4.7570192714553015, + "tokens_seen": 106889216 + }, + { + "epoch": 0.03, + "learning_rate": 0.0007818976930792378, + "loss": 1.9522, + "theoretical_loss": 4.756621203779169, + "tokens_seen": 106954752 + }, + { + "epoch": 0.03, + "learning_rate": 0.0007818816449348045, + "loss": 2.0456, + "theoretical_loss": 4.756223448190605, + "tokens_seen": 107020288 + }, + { + "epoch": 0.03, + "learning_rate": 0.0007818655967903712, + "loss": 2.1246, + "theoretical_loss": 4.755826004254046, + "tokens_seen": 107085824 + }, + { + "epoch": 0.03, + "learning_rate": 0.0007818495486459379, + "loss": 2.0421, + "theoretical_loss": 4.755428871534798, + "tokens_seen": 107151360 + }, + { + "epoch": 0.03, + "learning_rate": 0.0007818335005015046, + "loss": 1.977, + "theoretical_loss": 4.755032049599039, + "tokens_seen": 107216896 + }, + { + "epoch": 0.03, + "learning_rate": 0.0007818174523570713, + "loss": 2.0269, + "theoretical_loss": 4.754635538013817, + "tokens_seen": 107282432 + }, + { + "epoch": 0.03, + "learning_rate": 0.000781801404212638, + "loss": 2.0247, + "theoretical_loss": 4.754239336347048, + "tokens_seen": 107347968 + }, + { + "epoch": 0.03, + "learning_rate": 0.0007817853560682047, + "loss": 1.9948, + "theoretical_loss": 4.753843444167513, + "tokens_seen": 107413504 + }, + { + "epoch": 0.03, + "learning_rate": 0.0007817693079237714, + "loss": 2.0075, + "theoretical_loss": 4.753447861044853, + "tokens_seen": 107479040 + }, + { + "epoch": 0.03, + "learning_rate": 0.0007817532597793381, + "loss": 2.0549, + "theoretical_loss": 4.753052586549572, + "tokens_seen": 107544576 + }, + { + "epoch": 0.03, + "learning_rate": 0.0007817372116349048, + "loss": 2.0203, + "theoretical_loss": 4.75265762025303, + "tokens_seen": 107610112 + }, + { + "epoch": 0.03, + "learning_rate": 0.0007817211634904715, + "loss": 2.0052, + "theoretical_loss": 4.752262961727446, + "tokens_seen": 107675648 + }, + { + "epoch": 0.03, + "learning_rate": 0.0007817051153460382, + "loss": 2.0736, + "theoretical_loss": 4.75186861054589, + "tokens_seen": 107741184 + }, + { + "epoch": 0.03, + "learning_rate": 0.0007816890672016049, + "loss": 1.9687, + "theoretical_loss": 4.751474566282285, + "tokens_seen": 107806720 + }, + { + "epoch": 0.03, + "learning_rate": 0.0007816730190571716, + "loss": 2.1233, + "theoretical_loss": 4.751080828511402, + "tokens_seen": 107872256 + }, + { + "epoch": 0.03, + "learning_rate": 0.0007816569709127383, + "loss": 2.0741, + "theoretical_loss": 4.750687396808859, + "tokens_seen": 107937792 + }, + { + "epoch": 0.03, + "learning_rate": 0.000781640922768305, + "loss": 1.9122, + "theoretical_loss": 4.7502942707511195, + "tokens_seen": 108003328 + }, + { + "epoch": 0.03, + "learning_rate": 0.0007816248746238717, + "loss": 2.0265, + "theoretical_loss": 4.749901449915489, + "tokens_seen": 108068864 + }, + { + "epoch": 0.03, + "objective/train/docs_used": 46682, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 2.1818137168884277, + "objective/train/theoretical_loss": 4.749508933880117, + "objective/train/tokens_used": 128594400, + "theoretical_loss": 4.749508933880117, + "tokens_seen": 108134400 + }, + { + "epoch": 0.03, + "learning_rate": 0.0007816088264794384, + "loss": 2.0937, + "theoretical_loss": 4.749508933880117, + "tokens_seen": 108134400 + }, + { + "epoch": 0.03, + "learning_rate": 0.0007815927783350051, + "loss": 1.9963, + "theoretical_loss": 4.7491167222239845, + "tokens_seen": 108199936 + }, + { + "epoch": 0.03, + "learning_rate": 0.0007815767301905718, + "loss": 1.9994, + "theoretical_loss": 4.7487248145269145, + "tokens_seen": 108265472 + }, + { + "epoch": 0.03, + "learning_rate": 0.0007815606820461385, + "loss": 2.0375, + "theoretical_loss": 4.7483332103695615, + "tokens_seen": 108331008 + }, + { + "epoch": 0.03, + "learning_rate": 0.0007815446339017052, + "loss": 2.0884, + "theoretical_loss": 4.747941909333411, + "tokens_seen": 108396544 + }, + { + "epoch": 0.03, + "learning_rate": 0.0007815285857572719, + "loss": 1.9831, + "theoretical_loss": 4.747550911000781, + "tokens_seen": 108462080 + }, + { + "epoch": 0.03, + "learning_rate": 0.0007815125376128386, + "loss": 1.949, + "theoretical_loss": 4.747160214954813, + "tokens_seen": 108527616 + }, + { + "epoch": 0.03, + "learning_rate": 0.0007814964894684053, + "loss": 2.0771, + "theoretical_loss": 4.746769820779477, + "tokens_seen": 108593152 + }, + { + "epoch": 0.03, + "learning_rate": 0.000781480441323972, + "loss": 1.9987, + "theoretical_loss": 4.746379728059566, + "tokens_seen": 108658688 + }, + { + "epoch": 0.03, + "learning_rate": 0.0007814643931795387, + "loss": 2.0489, + "theoretical_loss": 4.745989936380692, + "tokens_seen": 108724224 + }, + { + "epoch": 0.03, + "learning_rate": 0.0007814483450351054, + "loss": 2.0695, + "theoretical_loss": 4.745600445329288, + "tokens_seen": 108789760 + }, + { + "epoch": 0.03, + "learning_rate": 0.0007814322968906721, + "loss": 2.0271, + "theoretical_loss": 4.745211254492602, + "tokens_seen": 108855296 + }, + { + "epoch": 0.03, + "learning_rate": 0.0007814162487462388, + "loss": 2.0047, + "theoretical_loss": 4.744822363458699, + "tokens_seen": 108920832 + }, + { + "epoch": 0.03, + "learning_rate": 0.0007814002006018055, + "loss": 2.0183, + "theoretical_loss": 4.744433771816455, + "tokens_seen": 108986368 + }, + { + "epoch": 0.03, + "learning_rate": 0.0007813841524573721, + "loss": 2.0936, + "theoretical_loss": 4.744045479155557, + "tokens_seen": 109051904 + }, + { + "epoch": 0.03, + "learning_rate": 0.0007813681043129389, + "loss": 1.9964, + "theoretical_loss": 4.7436574850665005, + "tokens_seen": 109117440 + }, + { + "epoch": 0.03, + "learning_rate": 0.0007813520561685055, + "loss": 1.9289, + "theoretical_loss": 4.743269789140587, + "tokens_seen": 109182976 + }, + { + "epoch": 0.03, + "learning_rate": 0.0007813360080240723, + "loss": 1.9287, + "theoretical_loss": 4.742882390969925, + "tokens_seen": 109248512 + }, + { + "epoch": 0.03, + "learning_rate": 0.0007813199598796389, + "loss": 2.0137, + "theoretical_loss": 4.742495290147421, + "tokens_seen": 109314048 + }, + { + "epoch": 0.03, + "learning_rate": 0.0007813039117352057, + "loss": 1.8879, + "theoretical_loss": 4.742108486266783, + "tokens_seen": 109379584 + }, + { + "epoch": 0.03, + "learning_rate": 0.0007812878635907723, + "loss": 2.0159, + "theoretical_loss": 4.74172197892252, + "tokens_seen": 109445120 + }, + { + "epoch": 0.03, + "learning_rate": 0.000781271815446339, + "loss": 1.9733, + "theoretical_loss": 4.741335767709934, + "tokens_seen": 109510656 + }, + { + "epoch": 0.03, + "learning_rate": 0.0007812557673019057, + "loss": 2.0156, + "theoretical_loss": 4.740949852225122, + "tokens_seen": 109576192 + }, + { + "epoch": 0.03, + "learning_rate": 0.0007812397191574724, + "loss": 2.0429, + "theoretical_loss": 4.740564232064973, + "tokens_seen": 109641728 + }, + { + "epoch": 0.03, + "learning_rate": 0.0007812236710130391, + "loss": 2.0745, + "theoretical_loss": 4.7401789068271665, + "tokens_seen": 109707264 + }, + { + "epoch": 0.03, + "learning_rate": 0.0007812076228686058, + "loss": 2.0089, + "theoretical_loss": 4.73979387611017, + "tokens_seen": 109772800 + }, + { + "epoch": 0.03, + "learning_rate": 0.0007811915747241725, + "loss": 1.9319, + "theoretical_loss": 4.739409139513235, + "tokens_seen": 109838336 + }, + { + "epoch": 0.03, + "learning_rate": 0.0007811755265797392, + "loss": 2.0464, + "theoretical_loss": 4.739024696636401, + "tokens_seen": 109903872 + }, + { + "epoch": 0.03, + "learning_rate": 0.0007811594784353059, + "loss": 2.0393, + "theoretical_loss": 4.738640547080485, + "tokens_seen": 109969408 + }, + { + "epoch": 0.03, + "learning_rate": 0.0007811434302908726, + "loss": 1.9458, + "theoretical_loss": 4.738256690447088, + "tokens_seen": 110034944 + }, + { + "epoch": 0.03, + "learning_rate": 0.0007811273821464393, + "loss": 2.0637, + "theoretical_loss": 4.7378731263385845, + "tokens_seen": 110100480 + }, + { + "epoch": 0.03, + "learning_rate": 0.000781111334002006, + "loss": 1.988, + "theoretical_loss": 4.737489854358129, + "tokens_seen": 110166016 + }, + { + "epoch": 0.03, + "learning_rate": 0.0007810952858575727, + "loss": 1.9183, + "theoretical_loss": 4.737106874109649, + "tokens_seen": 110231552 + }, + { + "epoch": 0.03, + "learning_rate": 0.0007810792377131394, + "loss": 2.0091, + "theoretical_loss": 4.736724185197844, + "tokens_seen": 110297088 + }, + { + "epoch": 0.03, + "learning_rate": 0.0007810631895687061, + "loss": 2.1092, + "theoretical_loss": 4.7363417872281826, + "tokens_seen": 110362624 + }, + { + "epoch": 0.03, + "learning_rate": 0.0007810471414242728, + "loss": 2.0532, + "theoretical_loss": 4.735959679806903, + "tokens_seen": 110428160 + }, + { + "epoch": 0.03, + "learning_rate": 0.0007810310932798395, + "loss": 2.0346, + "theoretical_loss": 4.735577862541007, + "tokens_seen": 110493696 + }, + { + "epoch": 0.03, + "learning_rate": 0.0007810150451354062, + "loss": 1.9872, + "theoretical_loss": 4.735196335038264, + "tokens_seen": 110559232 + }, + { + "epoch": 0.03, + "learning_rate": 0.0007809989969909729, + "loss": 2.1395, + "theoretical_loss": 4.734815096907203, + "tokens_seen": 110624768 + }, + { + "epoch": 0.03, + "learning_rate": 0.0007809829488465396, + "loss": 1.9127, + "theoretical_loss": 4.734434147757117, + "tokens_seen": 110690304 + }, + { + "epoch": 0.03, + "learning_rate": 0.0007809669007021063, + "loss": 1.9952, + "theoretical_loss": 4.734053487198054, + "tokens_seen": 110755840 + }, + { + "epoch": 0.03, + "learning_rate": 0.000780950852557673, + "loss": 1.9776, + "theoretical_loss": 4.733673114840817, + "tokens_seen": 110821376 + }, + { + "epoch": 0.03, + "learning_rate": 0.0007809348044132397, + "loss": 2.0071, + "theoretical_loss": 4.73329303029697, + "tokens_seen": 110886912 + }, + { + "epoch": 0.03, + "learning_rate": 0.0007809187562688064, + "loss": 2.0838, + "theoretical_loss": 4.732913233178825, + "tokens_seen": 110952448 + }, + { + "epoch": 0.03, + "learning_rate": 0.0007809027081243731, + "loss": 1.9109, + "theoretical_loss": 4.732533723099444, + "tokens_seen": 111017984 + }, + { + "epoch": 0.03, + "learning_rate": 0.0007808866599799398, + "loss": 2.0275, + "theoretical_loss": 4.732154499672642, + "tokens_seen": 111083520 + }, + { + "epoch": 0.03, + "learning_rate": 0.0007808706118355065, + "loss": 1.9429, + "theoretical_loss": 4.731775562512977, + "tokens_seen": 111149056 + }, + { + "epoch": 0.03, + "learning_rate": 0.0007808545636910732, + "loss": 1.9544, + "theoretical_loss": 4.7313969112357555, + "tokens_seen": 111214592 + }, + { + "epoch": 0.03, + "learning_rate": 0.0007808385155466399, + "loss": 1.868, + "theoretical_loss": 4.731018545457024, + "tokens_seen": 111280128 + }, + { + "epoch": 0.03, + "learning_rate": 0.0007808224674022066, + "loss": 2.0314, + "theoretical_loss": 4.730640464793574, + "tokens_seen": 111345664 + }, + { + "epoch": 0.03, + "objective/train/docs_used": 47788, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.9520606994628906, + "objective/train/theoretical_loss": 4.730262668862931, + "objective/train/tokens_used": 131871200, + "theoretical_loss": 4.730262668862931, + "tokens_seen": 111411200 + }, + { + "epoch": 0.03, + "learning_rate": 0.0007808064192577733, + "loss": 1.9923, + "theoretical_loss": 4.730262668862931, + "tokens_seen": 111411200 + }, + { + "epoch": 0.03, + "learning_rate": 0.00078079037111334, + "loss": 2.0842, + "theoretical_loss": 4.729885157283366, + "tokens_seen": 111476736 + }, + { + "epoch": 0.03, + "learning_rate": 0.0007807743229689067, + "loss": 1.9604, + "theoretical_loss": 4.729507929673879, + "tokens_seen": 111542272 + }, + { + "epoch": 0.03, + "learning_rate": 0.0007807582748244735, + "loss": 1.9862, + "theoretical_loss": 4.729130985654208, + "tokens_seen": 111607808 + }, + { + "epoch": 0.03, + "learning_rate": 0.0007807422266800401, + "loss": 2.0065, + "theoretical_loss": 4.72875432484482, + "tokens_seen": 111673344 + }, + { + "epoch": 0.03, + "learning_rate": 0.0007807261785356069, + "loss": 2.0533, + "theoretical_loss": 4.728377946866915, + "tokens_seen": 111738880 + }, + { + "epoch": 0.03, + "learning_rate": 0.0007807101303911736, + "loss": 2.0865, + "theoretical_loss": 4.72800185134242, + "tokens_seen": 111804416 + }, + { + "epoch": 0.03, + "learning_rate": 0.0007806940822467403, + "loss": 2.0986, + "theoretical_loss": 4.727626037893989, + "tokens_seen": 111869952 + }, + { + "epoch": 0.03, + "learning_rate": 0.000780678034102307, + "loss": 1.9703, + "theoretical_loss": 4.727250506145001, + "tokens_seen": 111935488 + }, + { + "epoch": 0.03, + "learning_rate": 0.0007806619859578737, + "loss": 1.9902, + "theoretical_loss": 4.726875255719557, + "tokens_seen": 112001024 + }, + { + "epoch": 0.03, + "learning_rate": 0.0007806459378134404, + "loss": 2.1584, + "theoretical_loss": 4.726500286242482, + "tokens_seen": 112066560 + }, + { + "epoch": 0.03, + "learning_rate": 0.0007806298896690071, + "loss": 2.0306, + "theoretical_loss": 4.7261255973393155, + "tokens_seen": 112132096 + }, + { + "epoch": 0.03, + "learning_rate": 0.0007806138415245738, + "loss": 2.0336, + "theoretical_loss": 4.725751188636319, + "tokens_seen": 112197632 + }, + { + "epoch": 0.03, + "learning_rate": 0.0007805977933801405, + "loss": 1.9923, + "theoretical_loss": 4.725377059760468, + "tokens_seen": 112263168 + }, + { + "epoch": 0.03, + "learning_rate": 0.0007805817452357072, + "loss": 1.9869, + "theoretical_loss": 4.725003210339452, + "tokens_seen": 112328704 + }, + { + "epoch": 0.03, + "learning_rate": 0.0007805656970912739, + "loss": 2.0568, + "theoretical_loss": 4.724629640001672, + "tokens_seen": 112394240 + }, + { + "epoch": 0.03, + "learning_rate": 0.0007805496489468406, + "loss": 2.0175, + "theoretical_loss": 4.724256348376242, + "tokens_seen": 112459776 + }, + { + "epoch": 0.03, + "learning_rate": 0.0007805336008024073, + "loss": 1.8579, + "theoretical_loss": 4.723883335092982, + "tokens_seen": 112525312 + }, + { + "epoch": 0.03, + "learning_rate": 0.000780517552657974, + "loss": 2.0644, + "theoretical_loss": 4.723510599782419, + "tokens_seen": 112590848 + }, + { + "epoch": 0.03, + "learning_rate": 0.0007805015045135407, + "loss": 2.0495, + "theoretical_loss": 4.723138142075788, + "tokens_seen": 112656384 + }, + { + "epoch": 0.03, + "learning_rate": 0.0007804854563691074, + "loss": 1.9843, + "theoretical_loss": 4.722765961605024, + "tokens_seen": 112721920 + }, + { + "epoch": 0.03, + "learning_rate": 0.0007804694082246741, + "loss": 2.075, + "theoretical_loss": 4.722394058002765, + "tokens_seen": 112787456 + }, + { + "epoch": 0.03, + "learning_rate": 0.0007804533600802408, + "loss": 2.0916, + "theoretical_loss": 4.7220224309023475, + "tokens_seen": 112852992 + }, + { + "epoch": 0.03, + "learning_rate": 0.0007804373119358075, + "loss": 1.961, + "theoretical_loss": 4.721651079937811, + "tokens_seen": 112918528 + }, + { + "epoch": 0.03, + "learning_rate": 0.0007804212637913742, + "loss": 2.08, + "theoretical_loss": 4.7212800047438845, + "tokens_seen": 112984064 + }, + { + "epoch": 0.03, + "learning_rate": 0.0007804052156469409, + "loss": 1.9894, + "theoretical_loss": 4.720909204955996, + "tokens_seen": 113049600 + }, + { + "epoch": 0.03, + "learning_rate": 0.0007803891675025076, + "loss": 1.9736, + "theoretical_loss": 4.720538680210268, + "tokens_seen": 113115136 + }, + { + "epoch": 0.03, + "learning_rate": 0.0007803731193580743, + "loss": 2.0159, + "theoretical_loss": 4.720168430143508, + "tokens_seen": 113180672 + }, + { + "epoch": 0.03, + "learning_rate": 0.000780357071213641, + "loss": 1.9955, + "theoretical_loss": 4.719798454393219, + "tokens_seen": 113246208 + }, + { + "epoch": 0.03, + "learning_rate": 0.0007803410230692077, + "loss": 2.0341, + "theoretical_loss": 4.719428752597588, + "tokens_seen": 113311744 + }, + { + "epoch": 0.03, + "learning_rate": 0.0007803249749247744, + "loss": 1.9228, + "theoretical_loss": 4.719059324395493, + "tokens_seen": 113377280 + }, + { + "epoch": 0.03, + "learning_rate": 0.0007803089267803411, + "loss": 1.9722, + "theoretical_loss": 4.71869016942649, + "tokens_seen": 113442816 + }, + { + "epoch": 0.03, + "learning_rate": 0.0007802928786359078, + "loss": 1.9862, + "theoretical_loss": 4.718321287330823, + "tokens_seen": 113508352 + }, + { + "epoch": 0.03, + "learning_rate": 0.0007802768304914745, + "loss": 2.0338, + "theoretical_loss": 4.7179526777494125, + "tokens_seen": 113573888 + }, + { + "epoch": 0.03, + "learning_rate": 0.0007802607823470412, + "loss": 1.9923, + "theoretical_loss": 4.717584340323864, + "tokens_seen": 113639424 + }, + { + "epoch": 0.03, + "learning_rate": 0.0007802447342026079, + "loss": 1.9057, + "theoretical_loss": 4.7172162746964545, + "tokens_seen": 113704960 + }, + { + "epoch": 0.03, + "learning_rate": 0.0007802286860581746, + "loss": 2.0515, + "theoretical_loss": 4.716848480510142, + "tokens_seen": 113770496 + }, + { + "epoch": 0.03, + "learning_rate": 0.0007802126379137413, + "loss": 1.9638, + "theoretical_loss": 4.716480957408556, + "tokens_seen": 113836032 + }, + { + "epoch": 0.03, + "learning_rate": 0.000780196589769308, + "loss": 1.8641, + "theoretical_loss": 4.716113705035999, + "tokens_seen": 113901568 + }, + { + "epoch": 0.03, + "learning_rate": 0.0007801805416248747, + "loss": 1.9445, + "theoretical_loss": 4.715746723037448, + "tokens_seen": 113967104 + }, + { + "epoch": 0.03, + "learning_rate": 0.0007801644934804414, + "loss": 2.0271, + "theoretical_loss": 4.715380011058542, + "tokens_seen": 114032640 + }, + { + "epoch": 0.03, + "learning_rate": 0.0007801484453360081, + "loss": 1.9212, + "theoretical_loss": 4.715013568745597, + "tokens_seen": 114098176 + }, + { + "epoch": 0.03, + "learning_rate": 0.0007801323971915748, + "loss": 1.9785, + "theoretical_loss": 4.714647395745587, + "tokens_seen": 114163712 + }, + { + "epoch": 0.03, + "learning_rate": 0.0007801163490471415, + "loss": 1.946, + "theoretical_loss": 4.714281491706154, + "tokens_seen": 114229248 + }, + { + "epoch": 0.03, + "learning_rate": 0.0007801003009027082, + "loss": 2.025, + "theoretical_loss": 4.713915856275603, + "tokens_seen": 114294784 + }, + { + "epoch": 0.03, + "learning_rate": 0.0007800842527582749, + "loss": 1.9241, + "theoretical_loss": 4.713550489102901, + "tokens_seen": 114360320 + }, + { + "epoch": 0.03, + "learning_rate": 0.0007800682046138416, + "loss": 1.8819, + "theoretical_loss": 4.7131853898376725, + "tokens_seen": 114425856 + }, + { + "epoch": 0.03, + "learning_rate": 0.0007800521564694083, + "loss": 1.9821, + "theoretical_loss": 4.7128205581302005, + "tokens_seen": 114491392 + }, + { + "epoch": 0.03, + "learning_rate": 0.000780036108324975, + "loss": 1.9809, + "theoretical_loss": 4.712455993631425, + "tokens_seen": 114556928 + }, + { + "epoch": 0.03, + "learning_rate": 0.0007800200601805417, + "loss": 2.0504, + "theoretical_loss": 4.712091695992941, + "tokens_seen": 114622464 + }, + { + "epoch": 0.03, + "objective/train/docs_used": 48994, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.9793815612792969, + "objective/train/theoretical_loss": 4.711727664866995, + "objective/train/tokens_used": 135148000, + "theoretical_loss": 4.711727664866995, + "tokens_seen": 114688000 + }, + { + "epoch": 0.03, + "learning_rate": 0.0007800040120361084, + "loss": 1.9351, + "theoretical_loss": 4.711727664866995, + "tokens_seen": 114688000 + }, + { + "epoch": 0.03, + "learning_rate": 0.000779987963891675, + "loss": 1.9642, + "theoretical_loss": 4.7113638999064875, + "tokens_seen": 114753536 + }, + { + "epoch": 0.03, + "learning_rate": 0.0007799719157472418, + "loss": 2.0197, + "theoretical_loss": 4.711000400764967, + "tokens_seen": 114819072 + }, + { + "epoch": 0.03, + "learning_rate": 0.0007799558676028084, + "loss": 1.9212, + "theoretical_loss": 4.71063716709663, + "tokens_seen": 114884608 + }, + { + "epoch": 0.03, + "learning_rate": 0.0007799398194583752, + "loss": 2.1173, + "theoretical_loss": 4.7102741985563235, + "tokens_seen": 114950144 + }, + { + "epoch": 0.03, + "learning_rate": 0.0007799237713139418, + "loss": 2.0131, + "theoretical_loss": 4.709911494799533, + "tokens_seen": 115015680 + }, + { + "epoch": 0.03, + "learning_rate": 0.0007799077231695086, + "loss": 2.1062, + "theoretical_loss": 4.709549055482395, + "tokens_seen": 115081216 + }, + { + "epoch": 0.03, + "learning_rate": 0.0007798916750250752, + "loss": 2.0187, + "theoretical_loss": 4.709186880261683, + "tokens_seen": 115146752 + }, + { + "epoch": 0.03, + "learning_rate": 0.000779875626880642, + "loss": 1.9552, + "theoretical_loss": 4.708824968794813, + "tokens_seen": 115212288 + }, + { + "epoch": 0.03, + "learning_rate": 0.0007798595787362086, + "loss": 1.7727, + "theoretical_loss": 4.70846332073984, + "tokens_seen": 115277824 + }, + { + "epoch": 0.03, + "learning_rate": 0.0007798435305917753, + "loss": 1.9681, + "theoretical_loss": 4.708101935755455, + "tokens_seen": 115343360 + }, + { + "epoch": 0.03, + "learning_rate": 0.000779827482447342, + "loss": 1.9841, + "theoretical_loss": 4.707740813500988, + "tokens_seen": 115408896 + }, + { + "epoch": 0.03, + "learning_rate": 0.0007798114343029087, + "loss": 2.0865, + "theoretical_loss": 4.707379953636399, + "tokens_seen": 115474432 + }, + { + "epoch": 0.04, + "learning_rate": 0.0007797953861584754, + "loss": 1.8866, + "theoretical_loss": 4.7070193558222835, + "tokens_seen": 115539968 + }, + { + "epoch": 0.04, + "learning_rate": 0.0007797793380140421, + "loss": 2.0099, + "theoretical_loss": 4.7066590197198686, + "tokens_seen": 115605504 + }, + { + "epoch": 0.04, + "learning_rate": 0.0007797632898696088, + "loss": 1.9494, + "theoretical_loss": 4.706298944991008, + "tokens_seen": 115671040 + }, + { + "epoch": 0.04, + "learning_rate": 0.0007797472417251755, + "loss": 1.9505, + "theoretical_loss": 4.705939131298189, + "tokens_seen": 115736576 + }, + { + "epoch": 0.04, + "learning_rate": 0.0007797311935807422, + "loss": 2.1335, + "theoretical_loss": 4.705579578304519, + "tokens_seen": 115802112 + }, + { + "epoch": 0.04, + "learning_rate": 0.0007797151454363089, + "loss": 2.0242, + "theoretical_loss": 4.705220285673736, + "tokens_seen": 115867648 + }, + { + "epoch": 0.04, + "learning_rate": 0.0007796990972918756, + "loss": 2.0822, + "theoretical_loss": 4.704861253070198, + "tokens_seen": 115933184 + }, + { + "epoch": 0.04, + "learning_rate": 0.0007796830491474423, + "loss": 2.037, + "theoretical_loss": 4.704502480158887, + "tokens_seen": 115998720 + }, + { + "epoch": 0.04, + "learning_rate": 0.000779667001003009, + "loss": 2.1125, + "theoretical_loss": 4.704143966605406, + "tokens_seen": 116064256 + }, + { + "epoch": 0.04, + "learning_rate": 0.0007796509528585757, + "loss": 1.9198, + "theoretical_loss": 4.703785712075973, + "tokens_seen": 116129792 + }, + { + "epoch": 0.04, + "learning_rate": 0.0007796349047141424, + "loss": 2.2112, + "theoretical_loss": 4.703427716237432, + "tokens_seen": 116195328 + }, + { + "epoch": 0.04, + "learning_rate": 0.0007796188565697091, + "loss": 2.0119, + "theoretical_loss": 4.703069978757233, + "tokens_seen": 116260864 + }, + { + "epoch": 0.04, + "learning_rate": 0.0007796028084252758, + "loss": 1.9567, + "theoretical_loss": 4.702712499303448, + "tokens_seen": 116326400 + }, + { + "epoch": 0.04, + "learning_rate": 0.0007795867602808425, + "loss": 2.0826, + "theoretical_loss": 4.7023552775447595, + "tokens_seen": 116391936 + }, + { + "epoch": 0.04, + "learning_rate": 0.0007795707121364092, + "loss": 2.0972, + "theoretical_loss": 4.701998313150462, + "tokens_seen": 116457472 + }, + { + "epoch": 0.04, + "learning_rate": 0.000779554663991976, + "loss": 1.9584, + "theoretical_loss": 4.701641605790459, + "tokens_seen": 116523008 + }, + { + "epoch": 0.04, + "learning_rate": 0.0007795386158475426, + "loss": 2.0373, + "theoretical_loss": 4.701285155135265, + "tokens_seen": 116588544 + }, + { + "epoch": 0.04, + "learning_rate": 0.0007795225677031094, + "loss": 1.9473, + "theoretical_loss": 4.700928960856, + "tokens_seen": 116654080 + }, + { + "epoch": 0.04, + "learning_rate": 0.000779506519558676, + "loss": 1.9847, + "theoretical_loss": 4.70057302262439, + "tokens_seen": 116719616 + }, + { + "epoch": 0.04, + "learning_rate": 0.0007794904714142428, + "loss": 1.8882, + "theoretical_loss": 4.700217340112767, + "tokens_seen": 116785152 + }, + { + "epoch": 0.04, + "learning_rate": 0.0007794744232698094, + "loss": 1.9837, + "theoretical_loss": 4.699861912994063, + "tokens_seen": 116850688 + }, + { + "epoch": 0.04, + "learning_rate": 0.0007794583751253762, + "loss": 1.9781, + "theoretical_loss": 4.699506740941813, + "tokens_seen": 116916224 + }, + { + "epoch": 0.04, + "learning_rate": 0.0007794423269809428, + "loss": 2.0306, + "theoretical_loss": 4.699151823630155, + "tokens_seen": 116981760 + }, + { + "epoch": 0.04, + "learning_rate": 0.0007794262788365096, + "loss": 1.8564, + "theoretical_loss": 4.698797160733818, + "tokens_seen": 117047296 + }, + { + "epoch": 0.04, + "learning_rate": 0.0007794102306920762, + "loss": 1.9603, + "theoretical_loss": 4.6984427519281375, + "tokens_seen": 117112832 + }, + { + "epoch": 0.04, + "learning_rate": 0.000779394182547643, + "loss": 2.2053, + "theoretical_loss": 4.698088596889037, + "tokens_seen": 117178368 + }, + { + "epoch": 0.04, + "learning_rate": 0.0007793781344032096, + "loss": 1.9264, + "theoretical_loss": 4.697734695293039, + "tokens_seen": 117243904 + }, + { + "epoch": 0.04, + "learning_rate": 0.0007793620862587764, + "loss": 2.0553, + "theoretical_loss": 4.697381046817256, + "tokens_seen": 117309440 + }, + { + "epoch": 0.04, + "learning_rate": 0.000779346038114343, + "loss": 1.9074, + "theoretical_loss": 4.697027651139393, + "tokens_seen": 117374976 + }, + { + "epoch": 0.04, + "learning_rate": 0.0007793299899699098, + "loss": 2.0017, + "theoretical_loss": 4.696674507937746, + "tokens_seen": 117440512 + }, + { + "epoch": 0.04, + "learning_rate": 0.0007793139418254765, + "loss": 1.8457, + "theoretical_loss": 4.696321616891199, + "tokens_seen": 117506048 + }, + { + "epoch": 0.04, + "learning_rate": 0.0007792978936810432, + "loss": 2.0361, + "theoretical_loss": 4.695968977679221, + "tokens_seen": 117571584 + }, + { + "epoch": 0.04, + "learning_rate": 0.0007792818455366099, + "loss": 2.0302, + "theoretical_loss": 4.695616589981871, + "tokens_seen": 117637120 + }, + { + "epoch": 0.04, + "learning_rate": 0.0007792657973921766, + "loss": 1.911, + "theoretical_loss": 4.695264453479789, + "tokens_seen": 117702656 + }, + { + "epoch": 0.04, + "learning_rate": 0.0007792497492477433, + "loss": 1.8988, + "theoretical_loss": 4.6949125678542, + "tokens_seen": 117768192 + }, + { + "epoch": 0.04, + "learning_rate": 0.00077923370110331, + "loss": 2.0897, + "theoretical_loss": 4.69456093278691, + "tokens_seen": 117833728 + }, + { + "epoch": 0.04, + "learning_rate": 0.0007792176529588767, + "loss": 1.963, + "theoretical_loss": 4.694209547960304, + "tokens_seen": 117899264 + }, + { + "epoch": 0.04, + "objective/train/docs_used": 50256, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 2.0614588260650635, + "objective/train/theoretical_loss": 4.693858413057347, + "objective/train/tokens_used": 138424800, + "theoretical_loss": 4.693858413057347, + "tokens_seen": 117964800 + }, + { + "epoch": 0.04, + "learning_rate": 0.0007792016048144434, + "loss": 2.039, + "theoretical_loss": 4.693858413057347, + "tokens_seen": 117964800 + }, + { + "epoch": 0.04, + "learning_rate": 0.0007791855566700101, + "loss": 2.076, + "theoretical_loss": 4.693507527761584, + "tokens_seen": 118030336 + }, + { + "epoch": 0.04, + "learning_rate": 0.0007791695085255768, + "loss": 2.0365, + "theoretical_loss": 4.6931568917571305, + "tokens_seen": 118095872 + }, + { + "epoch": 0.04, + "learning_rate": 0.0007791534603811435, + "loss": 1.9286, + "theoretical_loss": 4.692806504728683, + "tokens_seen": 118161408 + }, + { + "epoch": 0.04, + "learning_rate": 0.0007791374122367102, + "loss": 2.0702, + "theoretical_loss": 4.692456366361506, + "tokens_seen": 118226944 + }, + { + "epoch": 0.04, + "learning_rate": 0.0007791213640922769, + "loss": 2.0729, + "theoretical_loss": 4.692106476341438, + "tokens_seen": 118292480 + }, + { + "epoch": 0.04, + "learning_rate": 0.0007791053159478436, + "loss": 1.8666, + "theoretical_loss": 4.691756834354891, + "tokens_seen": 118358016 + }, + { + "epoch": 0.04, + "learning_rate": 0.0007790892678034103, + "loss": 2.0027, + "theoretical_loss": 4.691407440088843, + "tokens_seen": 118423552 + }, + { + "epoch": 0.04, + "learning_rate": 0.000779073219658977, + "loss": 1.917, + "theoretical_loss": 4.69105829323084, + "tokens_seen": 118489088 + }, + { + "epoch": 0.04, + "learning_rate": 0.0007790571715145437, + "loss": 2.0363, + "theoretical_loss": 4.690709393468995, + "tokens_seen": 118554624 + }, + { + "epoch": 0.04, + "learning_rate": 0.0007790411233701104, + "loss": 1.9703, + "theoretical_loss": 4.690360740491989, + "tokens_seen": 118620160 + }, + { + "epoch": 0.04, + "learning_rate": 0.0007790250752256771, + "loss": 1.899, + "theoretical_loss": 4.690012333989063, + "tokens_seen": 118685696 + }, + { + "epoch": 0.04, + "learning_rate": 0.0007790090270812438, + "loss": 1.9936, + "theoretical_loss": 4.689664173650023, + "tokens_seen": 118751232 + }, + { + "epoch": 0.04, + "learning_rate": 0.0007789929789368105, + "loss": 1.9622, + "theoretical_loss": 4.689316259165235, + "tokens_seen": 118816768 + }, + { + "epoch": 0.04, + "learning_rate": 0.0007789769307923772, + "loss": 1.9465, + "theoretical_loss": 4.6889685902256275, + "tokens_seen": 118882304 + }, + { + "epoch": 0.04, + "learning_rate": 0.0007789608826479439, + "loss": 1.9352, + "theoretical_loss": 4.688621166522684, + "tokens_seen": 118947840 + }, + { + "epoch": 0.04, + "learning_rate": 0.0007789448345035106, + "loss": 2.0093, + "theoretical_loss": 4.688273987748448, + "tokens_seen": 119013376 + }, + { + "epoch": 0.04, + "learning_rate": 0.0007789287863590773, + "loss": 2.0512, + "theoretical_loss": 4.687927053595518, + "tokens_seen": 119078912 + }, + { + "epoch": 0.04, + "learning_rate": 0.000778912738214644, + "loss": 2.0666, + "theoretical_loss": 4.687580363757048, + "tokens_seen": 119144448 + }, + { + "epoch": 0.04, + "learning_rate": 0.0007788966900702107, + "loss": 1.9371, + "theoretical_loss": 4.6872339179267435, + "tokens_seen": 119209984 + }, + { + "epoch": 0.04, + "learning_rate": 0.0007788806419257774, + "loss": 1.995, + "theoretical_loss": 4.686887715798864, + "tokens_seen": 119275520 + }, + { + "epoch": 0.04, + "learning_rate": 0.0007788645937813441, + "loss": 1.8701, + "theoretical_loss": 4.6865417570682215, + "tokens_seen": 119341056 + }, + { + "epoch": 0.04, + "learning_rate": 0.0007788485456369108, + "loss": 1.8799, + "theoretical_loss": 4.686196041430172, + "tokens_seen": 119406592 + }, + { + "epoch": 0.04, + "learning_rate": 0.0007788324974924775, + "loss": 2.0072, + "theoretical_loss": 4.685850568580624, + "tokens_seen": 119472128 + }, + { + "epoch": 0.04, + "learning_rate": 0.0007788164493480442, + "loss": 1.8927, + "theoretical_loss": 4.685505338216032, + "tokens_seen": 119537664 + }, + { + "epoch": 0.04, + "learning_rate": 0.0007788004012036109, + "loss": 2.0609, + "theoretical_loss": 4.685160350033398, + "tokens_seen": 119603200 + }, + { + "epoch": 0.04, + "learning_rate": 0.0007787843530591776, + "loss": 1.8658, + "theoretical_loss": 4.684815603730263, + "tokens_seen": 119668736 + }, + { + "epoch": 0.04, + "learning_rate": 0.0007787683049147443, + "loss": 1.9125, + "theoretical_loss": 4.684471099004718, + "tokens_seen": 119734272 + }, + { + "epoch": 0.04, + "learning_rate": 0.000778752256770311, + "loss": 1.9551, + "theoretical_loss": 4.684126835555389, + "tokens_seen": 119799808 + }, + { + "epoch": 0.04, + "learning_rate": 0.0007787362086258777, + "loss": 2.1576, + "theoretical_loss": 4.683782813081448, + "tokens_seen": 119865344 + }, + { + "epoch": 0.04, + "learning_rate": 0.0007787201604814444, + "loss": 1.9788, + "theoretical_loss": 4.683439031282606, + "tokens_seen": 119930880 + }, + { + "epoch": 0.04, + "learning_rate": 0.0007787041123370111, + "loss": 2.0216, + "theoretical_loss": 4.683095489859107, + "tokens_seen": 119996416 + }, + { + "epoch": 0.04, + "learning_rate": 0.0007786880641925778, + "loss": 1.8387, + "theoretical_loss": 4.682752188511737, + "tokens_seen": 120061952 + }, + { + "epoch": 0.04, + "learning_rate": 0.0007786720160481445, + "loss": 1.9339, + "theoretical_loss": 4.682409126941815, + "tokens_seen": 120127488 + }, + { + "epoch": 0.04, + "learning_rate": 0.0007786559679037112, + "loss": 2.0271, + "theoretical_loss": 4.682066304851195, + "tokens_seen": 120193024 + }, + { + "epoch": 0.04, + "learning_rate": 0.0007786399197592779, + "loss": 1.9572, + "theoretical_loss": 4.681723721942264, + "tokens_seen": 120258560 + }, + { + "epoch": 0.04, + "learning_rate": 0.0007786238716148446, + "loss": 2.0954, + "theoretical_loss": 4.6813813779179405, + "tokens_seen": 120324096 + }, + { + "epoch": 0.04, + "learning_rate": 0.0007786078234704113, + "loss": 1.9603, + "theoretical_loss": 4.681039272481674, + "tokens_seen": 120389632 + }, + { + "epoch": 0.04, + "learning_rate": 0.0007785917753259779, + "loss": 2.0401, + "theoretical_loss": 4.680697405337443, + "tokens_seen": 120455168 + }, + { + "epoch": 0.04, + "learning_rate": 0.0007785757271815447, + "loss": 2.088, + "theoretical_loss": 4.680355776189755, + "tokens_seen": 120520704 + }, + { + "epoch": 0.04, + "learning_rate": 0.0007785596790371113, + "loss": 1.98, + "theoretical_loss": 4.680014384743641, + "tokens_seen": 120586240 + }, + { + "epoch": 0.04, + "learning_rate": 0.0007785436308926781, + "loss": 1.7635, + "theoretical_loss": 4.679673230704664, + "tokens_seen": 120651776 + }, + { + "epoch": 0.04, + "learning_rate": 0.0007785275827482447, + "loss": 2.0184, + "theoretical_loss": 4.679332313778904, + "tokens_seen": 120717312 + }, + { + "epoch": 0.04, + "learning_rate": 0.0007785115346038115, + "loss": 2.0518, + "theoretical_loss": 4.678991633672968, + "tokens_seen": 120782848 + }, + { + "epoch": 0.04, + "learning_rate": 0.0007784954864593781, + "loss": 1.9329, + "theoretical_loss": 4.678651190093988, + "tokens_seen": 120848384 + }, + { + "epoch": 0.04, + "learning_rate": 0.0007784794383149449, + "loss": 1.9051, + "theoretical_loss": 4.67831098274961, + "tokens_seen": 120913920 + }, + { + "epoch": 0.04, + "learning_rate": 0.0007784633901705115, + "loss": 2.1229, + "theoretical_loss": 4.677971011348005, + "tokens_seen": 120979456 + }, + { + "epoch": 0.04, + "learning_rate": 0.0007784473420260782, + "loss": 2.0626, + "theoretical_loss": 4.677631275597859, + "tokens_seen": 121044992 + }, + { + "epoch": 0.04, + "learning_rate": 0.0007784312938816449, + "loss": 1.9069, + "theoretical_loss": 4.677291775208378, + "tokens_seen": 121110528 + }, + { + "epoch": 0.04, + "learning_rate": 0.0007784152457372116, + "loss": 2.0151, + "theoretical_loss": 4.676952509889281, + "tokens_seen": 121176064 + }, + { + "epoch": 0.04, + "objective/train/docs_used": 51395, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 2.0064034461975098, + "objective/train/theoretical_loss": 4.676613479350804, + "objective/train/tokens_used": 141701600, + "theoretical_loss": 4.676613479350804, + "tokens_seen": 121241600 + }, + { + "epoch": 0.04, + "learning_rate": 0.0007783991975927783, + "loss": 2.008, + "theoretical_loss": 4.676613479350804, + "tokens_seen": 121241600 + }, + { + "epoch": 0.04, + "learning_rate": 0.000778383149448345, + "loss": 1.9025, + "theoretical_loss": 4.6762746833036966, + "tokens_seen": 121307136 + }, + { + "epoch": 0.04, + "learning_rate": 0.0007783671013039117, + "loss": 1.9485, + "theoretical_loss": 4.675936121459218, + "tokens_seen": 121372672 + }, + { + "epoch": 0.04, + "learning_rate": 0.0007783510531594785, + "loss": 1.8696, + "theoretical_loss": 4.675597793529143, + "tokens_seen": 121438208 + }, + { + "epoch": 0.04, + "learning_rate": 0.0007783350050150451, + "loss": 1.9407, + "theoretical_loss": 4.675259699225752, + "tokens_seen": 121503744 + }, + { + "epoch": 0.04, + "learning_rate": 0.0007783189568706119, + "loss": 1.9453, + "theoretical_loss": 4.674921838261836, + "tokens_seen": 121569280 + }, + { + "epoch": 0.04, + "learning_rate": 0.0007783029087261785, + "loss": 1.9968, + "theoretical_loss": 4.674584210350695, + "tokens_seen": 121634816 + }, + { + "epoch": 0.04, + "learning_rate": 0.0007782868605817453, + "loss": 1.9482, + "theoretical_loss": 4.674246815206136, + "tokens_seen": 121700352 + }, + { + "epoch": 0.04, + "learning_rate": 0.0007782708124373119, + "loss": 2.02, + "theoretical_loss": 4.673909652542466, + "tokens_seen": 121765888 + }, + { + "epoch": 0.04, + "learning_rate": 0.0007782547642928787, + "loss": 1.8834, + "theoretical_loss": 4.673572722074502, + "tokens_seen": 121831424 + }, + { + "epoch": 0.04, + "learning_rate": 0.0007782387161484453, + "loss": 2.0139, + "theoretical_loss": 4.673236023517562, + "tokens_seen": 121896960 + }, + { + "epoch": 0.04, + "learning_rate": 0.0007782226680040121, + "loss": 1.8887, + "theoretical_loss": 4.672899556587465, + "tokens_seen": 121962496 + }, + { + "epoch": 0.04, + "learning_rate": 0.0007782066198595787, + "loss": 1.954, + "theoretical_loss": 4.672563321000533, + "tokens_seen": 122028032 + }, + { + "epoch": 0.04, + "learning_rate": 0.0007781905717151455, + "loss": 1.9145, + "theoretical_loss": 4.672227316473584, + "tokens_seen": 122093568 + }, + { + "epoch": 0.04, + "learning_rate": 0.0007781745235707121, + "loss": 1.9744, + "theoretical_loss": 4.671891542723937, + "tokens_seen": 122159104 + }, + { + "epoch": 0.04, + "learning_rate": 0.0007781584754262789, + "loss": 1.9245, + "theoretical_loss": 4.671555999469408, + "tokens_seen": 122224640 + }, + { + "epoch": 0.04, + "learning_rate": 0.0007781424272818455, + "loss": 2.1476, + "theoretical_loss": 4.671220686428308, + "tokens_seen": 122290176 + }, + { + "epoch": 0.04, + "learning_rate": 0.0007781263791374123, + "loss": 1.9706, + "theoretical_loss": 4.670885603319445, + "tokens_seen": 122355712 + }, + { + "epoch": 0.04, + "learning_rate": 0.0007781103309929789, + "loss": 1.8682, + "theoretical_loss": 4.670550749862119, + "tokens_seen": 122421248 + }, + { + "epoch": 0.04, + "learning_rate": 0.0007780942828485457, + "loss": 1.8624, + "theoretical_loss": 4.670216125776122, + "tokens_seen": 122486784 + }, + { + "epoch": 0.04, + "learning_rate": 0.0007780782347041123, + "loss": 1.9988, + "theoretical_loss": 4.669881730781741, + "tokens_seen": 122552320 + }, + { + "epoch": 0.04, + "learning_rate": 0.0007780621865596791, + "loss": 1.9027, + "theoretical_loss": 4.66954756459975, + "tokens_seen": 122617856 + }, + { + "epoch": 0.04, + "learning_rate": 0.0007780461384152457, + "loss": 1.9079, + "theoretical_loss": 4.669213626951413, + "tokens_seen": 122683392 + }, + { + "epoch": 0.04, + "learning_rate": 0.0007780300902708125, + "loss": 1.9521, + "theoretical_loss": 4.668879917558484, + "tokens_seen": 122748928 + }, + { + "epoch": 0.04, + "learning_rate": 0.0007780140421263791, + "loss": 1.9457, + "theoretical_loss": 4.6685464361432025, + "tokens_seen": 122814464 + }, + { + "epoch": 0.04, + "learning_rate": 0.0007779979939819459, + "loss": 1.9883, + "theoretical_loss": 4.668213182428295, + "tokens_seen": 122880000 + }, + { + "epoch": 0.04, + "learning_rate": 0.0007779819458375126, + "loss": 2.0671, + "theoretical_loss": 4.667880156136971, + "tokens_seen": 122945536 + }, + { + "epoch": 0.04, + "learning_rate": 0.0007779658976930793, + "loss": 1.97, + "theoretical_loss": 4.667547356992927, + "tokens_seen": 123011072 + }, + { + "epoch": 0.04, + "learning_rate": 0.000777949849548646, + "loss": 1.9568, + "theoretical_loss": 4.667214784720338, + "tokens_seen": 123076608 + }, + { + "epoch": 0.04, + "learning_rate": 0.0007779338014042127, + "loss": 1.8895, + "theoretical_loss": 4.666882439043865, + "tokens_seen": 123142144 + }, + { + "epoch": 0.04, + "learning_rate": 0.0007779177532597794, + "loss": 2.0429, + "theoretical_loss": 4.666550319688646, + "tokens_seen": 123207680 + }, + { + "epoch": 0.04, + "learning_rate": 0.0007779017051153461, + "loss": 1.9892, + "theoretical_loss": 4.6662184263803, + "tokens_seen": 123273216 + }, + { + "epoch": 0.04, + "learning_rate": 0.0007778856569709128, + "loss": 1.9428, + "theoretical_loss": 4.665886758844923, + "tokens_seen": 123338752 + }, + { + "epoch": 0.04, + "learning_rate": 0.0007778696088264795, + "loss": 1.9663, + "theoretical_loss": 4.66555531680909, + "tokens_seen": 123404288 + }, + { + "epoch": 0.04, + "learning_rate": 0.0007778535606820462, + "loss": 2.0296, + "theoretical_loss": 4.665224099999852, + "tokens_seen": 123469824 + }, + { + "epoch": 0.04, + "learning_rate": 0.0007778375125376129, + "loss": 1.8555, + "theoretical_loss": 4.664893108144735, + "tokens_seen": 123535360 + }, + { + "epoch": 0.04, + "learning_rate": 0.0007778214643931796, + "loss": 1.9911, + "theoretical_loss": 4.664562340971735, + "tokens_seen": 123600896 + }, + { + "epoch": 0.04, + "learning_rate": 0.0007778054162487463, + "loss": 1.9142, + "theoretical_loss": 4.664231798209327, + "tokens_seen": 123666432 + }, + { + "epoch": 0.04, + "learning_rate": 0.000777789368104313, + "loss": 2.1513, + "theoretical_loss": 4.663901479586454, + "tokens_seen": 123731968 + }, + { + "epoch": 0.04, + "learning_rate": 0.0007777733199598797, + "loss": 1.911, + "theoretical_loss": 4.663571384832531, + "tokens_seen": 123797504 + }, + { + "epoch": 0.04, + "learning_rate": 0.0007777572718154464, + "loss": 1.9337, + "theoretical_loss": 4.663241513677443, + "tokens_seen": 123863040 + }, + { + "epoch": 0.04, + "learning_rate": 0.0007777412236710131, + "loss": 1.983, + "theoretical_loss": 4.662911865851541, + "tokens_seen": 123928576 + }, + { + "epoch": 0.04, + "learning_rate": 0.0007777251755265798, + "loss": 1.8913, + "theoretical_loss": 4.662582441085648, + "tokens_seen": 123994112 + }, + { + "epoch": 0.04, + "learning_rate": 0.0007777091273821465, + "loss": 1.8664, + "theoretical_loss": 4.662253239111049, + "tokens_seen": 124059648 + }, + { + "epoch": 0.04, + "learning_rate": 0.0007776930792377132, + "loss": 1.9809, + "theoretical_loss": 4.661924259659497, + "tokens_seen": 124125184 + }, + { + "epoch": 0.04, + "learning_rate": 0.0007776770310932799, + "loss": 1.9676, + "theoretical_loss": 4.661595502463211, + "tokens_seen": 124190720 + }, + { + "epoch": 0.04, + "learning_rate": 0.0007776609829488466, + "loss": 1.9237, + "theoretical_loss": 4.661266967254868, + "tokens_seen": 124256256 + }, + { + "epoch": 0.04, + "learning_rate": 0.0007776449348044133, + "loss": 1.8675, + "theoretical_loss": 4.660938653767614, + "tokens_seen": 124321792 + }, + { + "epoch": 0.04, + "learning_rate": 0.00077762888665998, + "loss": 2.0486, + "theoretical_loss": 4.660610561735049, + "tokens_seen": 124387328 + }, + { + "epoch": 0.04, + "learning_rate": 0.0007776128385155467, + "loss": 1.9601, + "theoretical_loss": 4.66028269089124, + "tokens_seen": 124452864 + }, + { + "epoch": 0.04, + "objective/train/docs_used": 52572, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.9033129215240479, + "objective/train/theoretical_loss": 4.659955040970709, + "objective/train/tokens_used": 144978400, + "theoretical_loss": 4.659955040970709, + "tokens_seen": 124518400 + }, + { + "epoch": 0.04, + "learning_rate": 0.0007775967903711134, + "loss": 1.8962, + "theoretical_loss": 4.659955040970709, + "tokens_seen": 124518400 + }, + { + "epoch": 0.04, + "learning_rate": 0.0007775807422266801, + "loss": 1.9094, + "theoretical_loss": 4.659627611708437, + "tokens_seen": 124583936 + }, + { + "epoch": 0.04, + "learning_rate": 0.0007775646940822468, + "loss": 2.0726, + "theoretical_loss": 4.659300402839863, + "tokens_seen": 124649472 + }, + { + "epoch": 0.04, + "learning_rate": 0.0007775486459378135, + "loss": 1.9774, + "theoretical_loss": 4.658973414100881, + "tokens_seen": 124715008 + }, + { + "epoch": 0.04, + "learning_rate": 0.0007775325977933802, + "loss": 2.0897, + "theoretical_loss": 4.658646645227842, + "tokens_seen": 124780544 + }, + { + "epoch": 0.04, + "learning_rate": 0.0007775165496489469, + "loss": 1.9587, + "theoretical_loss": 4.6583200959575475, + "tokens_seen": 124846080 + }, + { + "epoch": 0.04, + "learning_rate": 0.0007775005015045136, + "loss": 1.9478, + "theoretical_loss": 4.657993766027255, + "tokens_seen": 124911616 + }, + { + "epoch": 0.04, + "learning_rate": 0.0007774844533600803, + "loss": 2.0377, + "theoretical_loss": 4.657667655174674, + "tokens_seen": 124977152 + }, + { + "epoch": 0.04, + "learning_rate": 0.000777468405215647, + "loss": 2.0441, + "theoretical_loss": 4.657341763137963, + "tokens_seen": 125042688 + }, + { + "epoch": 0.04, + "learning_rate": 0.0007774523570712137, + "loss": 2.0184, + "theoretical_loss": 4.657016089655732, + "tokens_seen": 125108224 + }, + { + "epoch": 0.04, + "learning_rate": 0.0007774363089267804, + "loss": 2.0196, + "theoretical_loss": 4.65669063446704, + "tokens_seen": 125173760 + }, + { + "epoch": 0.04, + "learning_rate": 0.0007774202607823471, + "loss": 1.9879, + "theoretical_loss": 4.656365397311393, + "tokens_seen": 125239296 + }, + { + "epoch": 0.04, + "learning_rate": 0.0007774042126379138, + "loss": 1.9835, + "theoretical_loss": 4.656040377928743, + "tokens_seen": 125304832 + }, + { + "epoch": 0.04, + "learning_rate": 0.0007773881644934805, + "loss": 1.8133, + "theoretical_loss": 4.655715576059493, + "tokens_seen": 125370368 + }, + { + "epoch": 0.04, + "learning_rate": 0.0007773721163490472, + "loss": 1.9717, + "theoretical_loss": 4.655390991444484, + "tokens_seen": 125435904 + }, + { + "epoch": 0.04, + "learning_rate": 0.0007773560682046139, + "loss": 1.9806, + "theoretical_loss": 4.6550666238250065, + "tokens_seen": 125501440 + }, + { + "epoch": 0.04, + "learning_rate": 0.0007773400200601806, + "loss": 1.9296, + "theoretical_loss": 4.65474247294279, + "tokens_seen": 125566976 + }, + { + "epoch": 0.04, + "learning_rate": 0.0007773239719157473, + "loss": 1.9379, + "theoretical_loss": 4.654418538540009, + "tokens_seen": 125632512 + }, + { + "epoch": 0.04, + "learning_rate": 0.000777307923771314, + "loss": 2.0146, + "theoretical_loss": 4.654094820359278, + "tokens_seen": 125698048 + }, + { + "epoch": 0.04, + "learning_rate": 0.0007772918756268807, + "loss": 2.0256, + "theoretical_loss": 4.653771318143651, + "tokens_seen": 125763584 + }, + { + "epoch": 0.04, + "learning_rate": 0.0007772758274824474, + "loss": 1.9402, + "theoretical_loss": 4.65344803163662, + "tokens_seen": 125829120 + }, + { + "epoch": 0.04, + "learning_rate": 0.0007772597793380141, + "loss": 1.9493, + "theoretical_loss": 4.653124960582118, + "tokens_seen": 125894656 + }, + { + "epoch": 0.04, + "learning_rate": 0.0007772437311935808, + "loss": 2.0109, + "theoretical_loss": 4.652802104724513, + "tokens_seen": 125960192 + }, + { + "epoch": 0.04, + "learning_rate": 0.0007772276830491475, + "loss": 1.9672, + "theoretical_loss": 4.652479463808609, + "tokens_seen": 126025728 + }, + { + "epoch": 0.04, + "learning_rate": 0.0007772116349047142, + "loss": 1.8962, + "theoretical_loss": 4.652157037579647, + "tokens_seen": 126091264 + }, + { + "epoch": 0.04, + "learning_rate": 0.0007771955867602808, + "loss": 1.9394, + "theoretical_loss": 4.6518348257833, + "tokens_seen": 126156800 + }, + { + "epoch": 0.04, + "learning_rate": 0.0007771795386158476, + "loss": 1.858, + "theoretical_loss": 4.651512828165675, + "tokens_seen": 126222336 + }, + { + "epoch": 0.04, + "learning_rate": 0.0007771634904714142, + "loss": 2.0211, + "theoretical_loss": 4.651191044473311, + "tokens_seen": 126287872 + }, + { + "epoch": 0.04, + "learning_rate": 0.000777147442326981, + "loss": 1.9099, + "theoretical_loss": 4.65086947445318, + "tokens_seen": 126353408 + }, + { + "epoch": 0.04, + "learning_rate": 0.0007771313941825476, + "loss": 1.7426, + "theoretical_loss": 4.650548117852681, + "tokens_seen": 126418944 + }, + { + "epoch": 0.04, + "learning_rate": 0.0007771153460381144, + "loss": 1.9774, + "theoretical_loss": 4.650226974419646, + "tokens_seen": 126484480 + }, + { + "epoch": 0.04, + "learning_rate": 0.000777099297893681, + "loss": 2.0101, + "theoretical_loss": 4.649906043902332, + "tokens_seen": 126550016 + }, + { + "epoch": 0.04, + "learning_rate": 0.0007770832497492478, + "loss": 2.0433, + "theoretical_loss": 4.6495853260494275, + "tokens_seen": 126615552 + }, + { + "epoch": 0.04, + "learning_rate": 0.0007770672016048144, + "loss": 1.9632, + "theoretical_loss": 4.649264820610043, + "tokens_seen": 126681088 + }, + { + "epoch": 0.04, + "learning_rate": 0.0007770511534603812, + "loss": 1.8905, + "theoretical_loss": 4.648944527333717, + "tokens_seen": 126746624 + }, + { + "epoch": 0.04, + "learning_rate": 0.0007770351053159478, + "loss": 1.9504, + "theoretical_loss": 4.648624445970414, + "tokens_seen": 126812160 + }, + { + "epoch": 0.04, + "learning_rate": 0.0007770190571715146, + "loss": 1.873, + "theoretical_loss": 4.648304576270519, + "tokens_seen": 126877696 + }, + { + "epoch": 0.04, + "learning_rate": 0.0007770030090270812, + "loss": 2.0246, + "theoretical_loss": 4.647984917984841, + "tokens_seen": 126943232 + }, + { + "epoch": 0.04, + "learning_rate": 0.000776986960882648, + "loss": 2.0287, + "theoretical_loss": 4.647665470864613, + "tokens_seen": 127008768 + }, + { + "epoch": 0.04, + "learning_rate": 0.0007769709127382146, + "loss": 1.9304, + "theoretical_loss": 4.6473462346614856, + "tokens_seen": 127074304 + }, + { + "epoch": 0.04, + "learning_rate": 0.0007769548645937814, + "loss": 2.013, + "theoretical_loss": 4.647027209127531, + "tokens_seen": 127139840 + }, + { + "epoch": 0.04, + "learning_rate": 0.000776938816449348, + "loss": 1.9271, + "theoretical_loss": 4.646708394015242, + "tokens_seen": 127205376 + }, + { + "epoch": 0.04, + "learning_rate": 0.0007769227683049148, + "loss": 1.9322, + "theoretical_loss": 4.646389789077524, + "tokens_seen": 127270912 + }, + { + "epoch": 0.04, + "learning_rate": 0.0007769067201604814, + "loss": 1.9821, + "theoretical_loss": 4.646071394067707, + "tokens_seen": 127336448 + }, + { + "epoch": 0.04, + "learning_rate": 0.0007768906720160482, + "loss": 1.9183, + "theoretical_loss": 4.645753208739533, + "tokens_seen": 127401984 + }, + { + "epoch": 0.04, + "learning_rate": 0.0007768746238716148, + "loss": 1.89, + "theoretical_loss": 4.645435232847159, + "tokens_seen": 127467520 + }, + { + "epoch": 0.04, + "learning_rate": 0.0007768585757271816, + "loss": 1.8929, + "theoretical_loss": 4.645117466145157, + "tokens_seen": 127533056 + }, + { + "epoch": 0.04, + "learning_rate": 0.0007768425275827482, + "loss": 1.8327, + "theoretical_loss": 4.6447999083885145, + "tokens_seen": 127598592 + }, + { + "epoch": 0.04, + "learning_rate": 0.000776826479438315, + "loss": 2.0034, + "theoretical_loss": 4.644482559332629, + "tokens_seen": 127664128 + }, + { + "epoch": 0.04, + "learning_rate": 0.0007768104312938816, + "loss": 1.9974, + "theoretical_loss": 4.644165418733312, + "tokens_seen": 127729664 + }, + { + "epoch": 0.04, + "objective/train/docs_used": 53741, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.999466896057129, + "objective/train/theoretical_loss": 4.6438484863467835, + "objective/train/tokens_used": 148255200, + "theoretical_loss": 4.6438484863467835, + "tokens_seen": 127795200 + }, + { + "epoch": 0.04, + "learning_rate": 0.0007767943831494484, + "loss": 2.0474, + "theoretical_loss": 4.6438484863467835, + "tokens_seen": 127795200 + }, + { + "epoch": 0.04, + "learning_rate": 0.000776778335005015, + "loss": 2.0811, + "theoretical_loss": 4.643531761929674, + "tokens_seen": 127860736 + }, + { + "epoch": 0.04, + "learning_rate": 0.0007767622868605818, + "loss": 1.876, + "theoretical_loss": 4.643215245239025, + "tokens_seen": 127926272 + }, + { + "epoch": 0.04, + "learning_rate": 0.0007767462387161484, + "loss": 1.8602, + "theoretical_loss": 4.642898936032285, + "tokens_seen": 127991808 + }, + { + "epoch": 0.04, + "learning_rate": 0.0007767301905717152, + "loss": 1.886, + "theoretical_loss": 4.6425828340673085, + "tokens_seen": 128057344 + }, + { + "epoch": 0.04, + "learning_rate": 0.0007767141424272818, + "loss": 2.0305, + "theoretical_loss": 4.642266939102357, + "tokens_seen": 128122880 + }, + { + "epoch": 0.04, + "learning_rate": 0.0007766980942828486, + "loss": 1.9045, + "theoretical_loss": 4.641951250896098, + "tokens_seen": 128188416 + }, + { + "epoch": 0.04, + "learning_rate": 0.0007766820461384152, + "loss": 1.9436, + "theoretical_loss": 4.641635769207603, + "tokens_seen": 128253952 + }, + { + "epoch": 0.04, + "learning_rate": 0.000776665997993982, + "loss": 2.0198, + "theoretical_loss": 4.641320493796349, + "tokens_seen": 128319488 + }, + { + "epoch": 0.04, + "learning_rate": 0.0007766499498495486, + "loss": 1.8639, + "theoretical_loss": 4.641005424422213, + "tokens_seen": 128385024 + }, + { + "epoch": 0.04, + "learning_rate": 0.0007766339017051154, + "loss": 1.9184, + "theoretical_loss": 4.640690560845474, + "tokens_seen": 128450560 + }, + { + "epoch": 0.04, + "learning_rate": 0.000776617853560682, + "loss": 1.9848, + "theoretical_loss": 4.640375902826815, + "tokens_seen": 128516096 + }, + { + "epoch": 0.04, + "learning_rate": 0.0007766018054162488, + "loss": 2.0954, + "theoretical_loss": 4.640061450127316, + "tokens_seen": 128581632 + }, + { + "epoch": 0.04, + "learning_rate": 0.0007765857572718155, + "loss": 1.8735, + "theoretical_loss": 4.639747202508459, + "tokens_seen": 128647168 + }, + { + "epoch": 0.04, + "learning_rate": 0.0007765697091273822, + "loss": 1.9605, + "theoretical_loss": 4.639433159732121, + "tokens_seen": 128712704 + }, + { + "epoch": 0.04, + "learning_rate": 0.0007765536609829489, + "loss": 2.0226, + "theoretical_loss": 4.639119321560581, + "tokens_seen": 128778240 + }, + { + "epoch": 0.04, + "learning_rate": 0.0007765376128385156, + "loss": 1.9555, + "theoretical_loss": 4.638805687756511, + "tokens_seen": 128843776 + }, + { + "epoch": 0.04, + "learning_rate": 0.0007765215646940823, + "loss": 1.8859, + "theoretical_loss": 4.63849225808298, + "tokens_seen": 128909312 + }, + { + "epoch": 0.04, + "learning_rate": 0.000776505516549649, + "loss": 1.9896, + "theoretical_loss": 4.6381790323034515, + "tokens_seen": 128974848 + }, + { + "epoch": 0.04, + "learning_rate": 0.0007764894684052157, + "loss": 2.0849, + "theoretical_loss": 4.637866010181784, + "tokens_seen": 129040384 + }, + { + "epoch": 0.04, + "learning_rate": 0.0007764734202607824, + "loss": 1.9602, + "theoretical_loss": 4.63755319148223, + "tokens_seen": 129105920 + }, + { + "epoch": 0.04, + "learning_rate": 0.0007764573721163491, + "loss": 1.9742, + "theoretical_loss": 4.637240575969432, + "tokens_seen": 129171456 + }, + { + "epoch": 0.04, + "learning_rate": 0.0007764413239719158, + "loss": 1.8941, + "theoretical_loss": 4.636928163408426, + "tokens_seen": 129236992 + }, + { + "epoch": 0.04, + "learning_rate": 0.0007764252758274825, + "loss": 2.0185, + "theoretical_loss": 4.636615953564637, + "tokens_seen": 129302528 + }, + { + "epoch": 0.04, + "learning_rate": 0.0007764092276830492, + "loss": 1.7939, + "theoretical_loss": 4.636303946203881, + "tokens_seen": 129368064 + }, + { + "epoch": 0.04, + "learning_rate": 0.0007763931795386159, + "loss": 1.9907, + "theoretical_loss": 4.635992141092364, + "tokens_seen": 129433600 + }, + { + "epoch": 0.04, + "learning_rate": 0.0007763771313941826, + "loss": 1.8782, + "theoretical_loss": 4.635680537996679, + "tokens_seen": 129499136 + }, + { + "epoch": 0.04, + "learning_rate": 0.0007763610832497493, + "loss": 1.9687, + "theoretical_loss": 4.635369136683805, + "tokens_seen": 129564672 + }, + { + "epoch": 0.04, + "learning_rate": 0.000776345035105316, + "loss": 1.895, + "theoretical_loss": 4.6350579369211085, + "tokens_seen": 129630208 + }, + { + "epoch": 0.04, + "learning_rate": 0.0007763289869608827, + "loss": 1.9713, + "theoretical_loss": 4.634746938476344, + "tokens_seen": 129695744 + }, + { + "epoch": 0.04, + "learning_rate": 0.0007763129388164494, + "loss": 1.8656, + "theoretical_loss": 4.634436141117645, + "tokens_seen": 129761280 + }, + { + "epoch": 0.04, + "learning_rate": 0.0007762968906720161, + "loss": 2.0204, + "theoretical_loss": 4.634125544613536, + "tokens_seen": 129826816 + }, + { + "epoch": 0.04, + "learning_rate": 0.0007762808425275828, + "loss": 1.9206, + "theoretical_loss": 4.63381514873292, + "tokens_seen": 129892352 + }, + { + "epoch": 0.04, + "learning_rate": 0.0007762647943831495, + "loss": 2.0952, + "theoretical_loss": 4.633504953245085, + "tokens_seen": 129957888 + }, + { + "epoch": 0.04, + "learning_rate": 0.0007762487462387162, + "loss": 2.0068, + "theoretical_loss": 4.633194957919697, + "tokens_seen": 130023424 + }, + { + "epoch": 0.04, + "learning_rate": 0.0007762326980942829, + "loss": 2.047, + "theoretical_loss": 4.632885162526806, + "tokens_seen": 130088960 + }, + { + "epoch": 0.04, + "learning_rate": 0.0007762166499498496, + "loss": 1.9531, + "theoretical_loss": 4.63257556683684, + "tokens_seen": 130154496 + }, + { + "epoch": 0.04, + "learning_rate": 0.0007762006018054163, + "loss": 1.9598, + "theoretical_loss": 4.632266170620607, + "tokens_seen": 130220032 + }, + { + "epoch": 0.04, + "learning_rate": 0.000776184553660983, + "loss": 2.0026, + "theoretical_loss": 4.631956973649293, + "tokens_seen": 130285568 + }, + { + "epoch": 0.04, + "learning_rate": 0.0007761685055165497, + "loss": 2.068, + "theoretical_loss": 4.631647975694461, + "tokens_seen": 130351104 + }, + { + "epoch": 0.04, + "learning_rate": 0.0007761524573721164, + "loss": 2.041, + "theoretical_loss": 4.631339176528051, + "tokens_seen": 130416640 + }, + { + "epoch": 0.04, + "learning_rate": 0.0007761364092276831, + "loss": 1.919, + "theoretical_loss": 4.631030575922379, + "tokens_seen": 130482176 + }, + { + "epoch": 0.04, + "learning_rate": 0.0007761203610832498, + "loss": 1.8603, + "theoretical_loss": 4.630722173650135, + "tokens_seen": 130547712 + }, + { + "epoch": 0.04, + "learning_rate": 0.0007761043129388165, + "loss": 1.8621, + "theoretical_loss": 4.630413969484383, + "tokens_seen": 130613248 + }, + { + "epoch": 0.04, + "learning_rate": 0.0007760882647943832, + "loss": 1.7854, + "theoretical_loss": 4.630105963198563, + "tokens_seen": 130678784 + }, + { + "epoch": 0.04, + "learning_rate": 0.0007760722166499499, + "loss": 1.9231, + "theoretical_loss": 4.629798154566485, + "tokens_seen": 130744320 + }, + { + "epoch": 0.04, + "learning_rate": 0.0007760561685055166, + "loss": 1.8858, + "theoretical_loss": 4.62949054336233, + "tokens_seen": 130809856 + }, + { + "epoch": 0.04, + "learning_rate": 0.0007760401203610833, + "loss": 2.0357, + "theoretical_loss": 4.629183129360653, + "tokens_seen": 130875392 + }, + { + "epoch": 0.04, + "learning_rate": 0.00077602407221665, + "loss": 1.9207, + "theoretical_loss": 4.628875912336378, + "tokens_seen": 130940928 + }, + { + "epoch": 0.04, + "learning_rate": 0.0007760080240722167, + "loss": 1.9181, + "theoretical_loss": 4.628568892064797, + "tokens_seen": 131006464 + }, + { + "epoch": 0.04, + "objective/train/docs_used": 54897, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.9696239233016968, + "objective/train/theoretical_loss": 4.62826206832157, + "objective/train/tokens_used": 151532000, + "theoretical_loss": 4.62826206832157, + "tokens_seen": 131072000 + }, + { + "epoch": 0.04, + "learning_rate": 0.0007759919759277834, + "loss": 1.8815, + "theoretical_loss": 4.62826206832157, + "tokens_seen": 131072000 + }, + { + "epoch": 0.04, + "learning_rate": 0.0007759759277833502, + "loss": 2.1364, + "theoretical_loss": 4.6279554408827295, + "tokens_seen": 131137536 + }, + { + "epoch": 0.04, + "learning_rate": 0.0007759598796389168, + "loss": 1.8593, + "theoretical_loss": 4.62764900952467, + "tokens_seen": 131203072 + }, + { + "epoch": 0.04, + "learning_rate": 0.0007759438314944836, + "loss": 1.8445, + "theoretical_loss": 4.627342774024155, + "tokens_seen": 131268608 + }, + { + "epoch": 0.04, + "learning_rate": 0.0007759277833500502, + "loss": 1.9955, + "theoretical_loss": 4.6270367341583105, + "tokens_seen": 131334144 + }, + { + "epoch": 0.04, + "learning_rate": 0.0007759117352056169, + "loss": 1.8791, + "theoretical_loss": 4.626730889704632, + "tokens_seen": 131399680 + }, + { + "epoch": 0.04, + "learning_rate": 0.0007758956870611836, + "loss": 1.8115, + "theoretical_loss": 4.626425240440973, + "tokens_seen": 131465216 + }, + { + "epoch": 0.04, + "learning_rate": 0.0007758796389167503, + "loss": 2.0407, + "theoretical_loss": 4.626119786145557, + "tokens_seen": 131530752 + }, + { + "epoch": 0.04, + "learning_rate": 0.000775863590772317, + "loss": 1.9006, + "theoretical_loss": 4.625814526596962, + "tokens_seen": 131596288 + }, + { + "epoch": 0.04, + "learning_rate": 0.0007758475426278837, + "loss": 1.9069, + "theoretical_loss": 4.625509461574135, + "tokens_seen": 131661824 + }, + { + "epoch": 0.04, + "learning_rate": 0.0007758314944834504, + "loss": 2.0992, + "theoretical_loss": 4.625204590856377, + "tokens_seen": 131727360 + }, + { + "epoch": 0.04, + "learning_rate": 0.0007758154463390171, + "loss": 1.9607, + "theoretical_loss": 4.624899914223355, + "tokens_seen": 131792896 + }, + { + "epoch": 0.04, + "learning_rate": 0.0007757993981945837, + "loss": 1.9354, + "theoretical_loss": 4.624595431455091, + "tokens_seen": 131858432 + }, + { + "epoch": 0.04, + "learning_rate": 0.0007757833500501505, + "loss": 1.9455, + "theoretical_loss": 4.62429114233197, + "tokens_seen": 131923968 + }, + { + "epoch": 0.04, + "learning_rate": 0.0007757673019057171, + "loss": 2.0128, + "theoretical_loss": 4.623987046634729, + "tokens_seen": 131989504 + }, + { + "epoch": 0.04, + "learning_rate": 0.0007757512537612839, + "loss": 1.8871, + "theoretical_loss": 4.623683144144465, + "tokens_seen": 132055040 + }, + { + "epoch": 0.04, + "learning_rate": 0.0007757352056168505, + "loss": 1.9176, + "theoretical_loss": 4.623379434642632, + "tokens_seen": 132120576 + }, + { + "epoch": 0.04, + "learning_rate": 0.0007757191574724173, + "loss": 1.9214, + "theoretical_loss": 4.623075917911039, + "tokens_seen": 132186112 + }, + { + "epoch": 0.04, + "learning_rate": 0.0007757031093279839, + "loss": 1.9814, + "theoretical_loss": 4.622772593731848, + "tokens_seen": 132251648 + }, + { + "epoch": 0.04, + "learning_rate": 0.0007756870611835507, + "loss": 1.9933, + "theoretical_loss": 4.622469461887577, + "tokens_seen": 132317184 + }, + { + "epoch": 0.04, + "learning_rate": 0.0007756710130391173, + "loss": 2.0109, + "theoretical_loss": 4.622166522161096, + "tokens_seen": 132382720 + }, + { + "epoch": 0.04, + "learning_rate": 0.0007756549648946841, + "loss": 1.9152, + "theoretical_loss": 4.621863774335628, + "tokens_seen": 132448256 + }, + { + "epoch": 0.04, + "learning_rate": 0.0007756389167502507, + "loss": 2.0563, + "theoretical_loss": 4.621561218194747, + "tokens_seen": 132513792 + }, + { + "epoch": 0.04, + "learning_rate": 0.0007756228686058175, + "loss": 2.0387, + "theoretical_loss": 4.62125885352238, + "tokens_seen": 132579328 + }, + { + "epoch": 0.04, + "learning_rate": 0.0007756068204613841, + "loss": 1.9854, + "theoretical_loss": 4.620956680102801, + "tokens_seen": 132644864 + }, + { + "epoch": 0.04, + "learning_rate": 0.0007755907723169509, + "loss": 2.0074, + "theoretical_loss": 4.620654697720637, + "tokens_seen": 132710400 + }, + { + "epoch": 0.04, + "learning_rate": 0.0007755747241725175, + "loss": 2.1297, + "theoretical_loss": 4.620352906160861, + "tokens_seen": 132775936 + }, + { + "epoch": 0.04, + "learning_rate": 0.0007755586760280843, + "loss": 2.1073, + "theoretical_loss": 4.6200513052087935, + "tokens_seen": 132841472 + }, + { + "epoch": 0.04, + "learning_rate": 0.0007755426278836509, + "loss": 2.0247, + "theoretical_loss": 4.6197498946501065, + "tokens_seen": 132907008 + }, + { + "epoch": 0.04, + "learning_rate": 0.0007755265797392177, + "loss": 2.0022, + "theoretical_loss": 4.6194486742708145, + "tokens_seen": 132972544 + }, + { + "epoch": 0.04, + "learning_rate": 0.0007755105315947843, + "loss": 1.9739, + "theoretical_loss": 4.6191476438572785, + "tokens_seen": 133038080 + }, + { + "epoch": 0.04, + "learning_rate": 0.0007754944834503511, + "loss": 1.9305, + "theoretical_loss": 4.618846803196206, + "tokens_seen": 133103616 + }, + { + "epoch": 0.04, + "learning_rate": 0.0007754784353059177, + "loss": 1.928, + "theoretical_loss": 4.618546152074648, + "tokens_seen": 133169152 + }, + { + "epoch": 0.04, + "learning_rate": 0.0007754623871614845, + "loss": 1.9489, + "theoretical_loss": 4.61824569028, + "tokens_seen": 133234688 + }, + { + "epoch": 0.04, + "learning_rate": 0.0007754463390170511, + "loss": 1.968, + "theoretical_loss": 4.617945417599998, + "tokens_seen": 133300224 + }, + { + "epoch": 0.04, + "learning_rate": 0.0007754302908726179, + "loss": 1.8376, + "theoretical_loss": 4.617645333822722, + "tokens_seen": 133365760 + }, + { + "epoch": 0.04, + "learning_rate": 0.0007754142427281845, + "loss": 1.7964, + "theoretical_loss": 4.617345438736594, + "tokens_seen": 133431296 + }, + { + "epoch": 0.04, + "learning_rate": 0.0007753981945837513, + "loss": 1.9027, + "theoretical_loss": 4.617045732130375, + "tokens_seen": 133496832 + }, + { + "epoch": 0.04, + "learning_rate": 0.0007753821464393179, + "loss": 1.8473, + "theoretical_loss": 4.616746213793167, + "tokens_seen": 133562368 + }, + { + "epoch": 0.04, + "learning_rate": 0.0007753660982948847, + "loss": 2.0037, + "theoretical_loss": 4.616446883514412, + "tokens_seen": 133627904 + }, + { + "epoch": 0.04, + "learning_rate": 0.0007753500501504513, + "loss": 1.889, + "theoretical_loss": 4.616147741083891, + "tokens_seen": 133693440 + }, + { + "epoch": 0.04, + "learning_rate": 0.0007753340020060181, + "loss": 1.924, + "theoretical_loss": 4.6158487862917195, + "tokens_seen": 133758976 + }, + { + "epoch": 0.04, + "learning_rate": 0.0007753179538615847, + "loss": 1.877, + "theoretical_loss": 4.615550018928353, + "tokens_seen": 133824512 + }, + { + "epoch": 0.04, + "learning_rate": 0.0007753019057171515, + "loss": 2.0209, + "theoretical_loss": 4.615251438784584, + "tokens_seen": 133890048 + }, + { + "epoch": 0.04, + "learning_rate": 0.0007752858575727181, + "loss": 2.0147, + "theoretical_loss": 4.61495304565154, + "tokens_seen": 133955584 + }, + { + "epoch": 0.04, + "learning_rate": 0.0007752698094282849, + "loss": 2.0136, + "theoretical_loss": 4.614654839320682, + "tokens_seen": 134021120 + }, + { + "epoch": 0.04, + "learning_rate": 0.0007752537612838515, + "loss": 1.9326, + "theoretical_loss": 4.614356819583808, + "tokens_seen": 134086656 + }, + { + "epoch": 0.04, + "learning_rate": 0.0007752377131394183, + "loss": 2.0197, + "theoretical_loss": 4.614058986233047, + "tokens_seen": 134152192 + }, + { + "epoch": 0.04, + "learning_rate": 0.000775221664994985, + "loss": 2.0111, + "theoretical_loss": 4.613761339060863, + "tokens_seen": 134217728 + }, + { + "epoch": 0.04, + "learning_rate": 0.0007752056168505517, + "loss": 1.887, + "theoretical_loss": 4.613463877860051, + "tokens_seen": 134283264 + }, + { + "epoch": 0.04, + "objective/train/docs_used": 56110, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.9131815433502197, + "objective/train/theoretical_loss": 4.61316660242374, + "objective/train/tokens_used": 154808800, + "theoretical_loss": 4.61316660242374, + "tokens_seen": 134348800 + }, + { + "epoch": 0.04, + "learning_rate": 0.0007751895687061184, + "loss": 2.0582, + "theoretical_loss": 4.61316660242374, + "tokens_seen": 134348800 + }, + { + "epoch": 0.04, + "learning_rate": 0.0007751735205616851, + "loss": 2.027, + "theoretical_loss": 4.6128695125453865, + "tokens_seen": 134414336 + }, + { + "epoch": 0.04, + "learning_rate": 0.0007751574724172518, + "loss": 1.9448, + "theoretical_loss": 4.612572608018779, + "tokens_seen": 134479872 + }, + { + "epoch": 0.04, + "learning_rate": 0.0007751414242728185, + "loss": 1.9141, + "theoretical_loss": 4.612275888638036, + "tokens_seen": 134545408 + }, + { + "epoch": 0.04, + "learning_rate": 0.0007751253761283852, + "loss": 1.8997, + "theoretical_loss": 4.611979354197603, + "tokens_seen": 134610944 + }, + { + "epoch": 0.04, + "learning_rate": 0.0007751093279839519, + "loss": 1.8459, + "theoretical_loss": 4.6116830044922565, + "tokens_seen": 134676480 + }, + { + "epoch": 0.04, + "learning_rate": 0.0007750932798395186, + "loss": 1.9561, + "theoretical_loss": 4.611386839317097, + "tokens_seen": 134742016 + }, + { + "epoch": 0.04, + "learning_rate": 0.0007750772316950853, + "loss": 1.9417, + "theoretical_loss": 4.611090858467554, + "tokens_seen": 134807552 + }, + { + "epoch": 0.04, + "learning_rate": 0.000775061183550652, + "loss": 1.7874, + "theoretical_loss": 4.610795061739384, + "tokens_seen": 134873088 + }, + { + "epoch": 0.04, + "learning_rate": 0.0007750451354062187, + "loss": 1.9422, + "theoretical_loss": 4.610499448928666, + "tokens_seen": 134938624 + }, + { + "epoch": 0.04, + "learning_rate": 0.0007750290872617854, + "loss": 2.0973, + "theoretical_loss": 4.610204019831805, + "tokens_seen": 135004160 + }, + { + "epoch": 0.04, + "learning_rate": 0.0007750130391173521, + "loss": 2.0093, + "theoretical_loss": 4.609908774245532, + "tokens_seen": 135069696 + }, + { + "epoch": 0.04, + "learning_rate": 0.0007749969909729188, + "loss": 2.0617, + "theoretical_loss": 4.609613711966898, + "tokens_seen": 135135232 + }, + { + "epoch": 0.04, + "learning_rate": 0.0007749809428284855, + "loss": 1.8904, + "theoretical_loss": 4.60931883279328, + "tokens_seen": 135200768 + }, + { + "epoch": 0.04, + "learning_rate": 0.0007749648946840522, + "loss": 1.8458, + "theoretical_loss": 4.6090241365223745, + "tokens_seen": 135266304 + }, + { + "epoch": 0.04, + "learning_rate": 0.0007749488465396189, + "loss": 1.8843, + "theoretical_loss": 4.6087296229522, + "tokens_seen": 135331840 + }, + { + "epoch": 0.04, + "learning_rate": 0.0007749327983951856, + "loss": 1.8951, + "theoretical_loss": 4.608435291881099, + "tokens_seen": 135397376 + }, + { + "epoch": 0.04, + "learning_rate": 0.0007749167502507523, + "loss": 1.9429, + "theoretical_loss": 4.608141143107728, + "tokens_seen": 135462912 + }, + { + "epoch": 0.04, + "learning_rate": 0.000774900702106319, + "loss": 1.9606, + "theoretical_loss": 4.607847176431067, + "tokens_seen": 135528448 + }, + { + "epoch": 0.04, + "learning_rate": 0.0007748846539618857, + "loss": 1.8603, + "theoretical_loss": 4.6075533916504146, + "tokens_seen": 135593984 + }, + { + "epoch": 0.04, + "learning_rate": 0.0007748686058174524, + "loss": 1.925, + "theoretical_loss": 4.607259788565385, + "tokens_seen": 135659520 + }, + { + "epoch": 0.04, + "learning_rate": 0.0007748525576730191, + "loss": 1.9986, + "theoretical_loss": 4.606966366975914, + "tokens_seen": 135725056 + }, + { + "epoch": 0.04, + "learning_rate": 0.0007748365095285858, + "loss": 2.1594, + "theoretical_loss": 4.606673126682249, + "tokens_seen": 135790592 + }, + { + "epoch": 0.04, + "learning_rate": 0.0007748204613841525, + "loss": 1.892, + "theoretical_loss": 4.6063800674849595, + "tokens_seen": 135856128 + }, + { + "epoch": 0.04, + "learning_rate": 0.0007748044132397192, + "loss": 1.9469, + "theoretical_loss": 4.606087189184924, + "tokens_seen": 135921664 + }, + { + "epoch": 0.04, + "learning_rate": 0.0007747883650952859, + "loss": 1.9675, + "theoretical_loss": 4.6057944915833415, + "tokens_seen": 135987200 + }, + { + "epoch": 0.04, + "learning_rate": 0.0007747723169508527, + "loss": 1.9691, + "theoretical_loss": 4.6055019744817205, + "tokens_seen": 136052736 + }, + { + "epoch": 0.04, + "learning_rate": 0.0007747562688064193, + "loss": 2.0128, + "theoretical_loss": 4.605209637681886, + "tokens_seen": 136118272 + }, + { + "epoch": 0.04, + "learning_rate": 0.0007747402206619861, + "loss": 1.8155, + "theoretical_loss": 4.604917480985975, + "tokens_seen": 136183808 + }, + { + "epoch": 0.04, + "learning_rate": 0.0007747241725175527, + "loss": 1.8094, + "theoretical_loss": 4.604625504196438, + "tokens_seen": 136249344 + }, + { + "epoch": 0.04, + "learning_rate": 0.0007747081243731195, + "loss": 1.9025, + "theoretical_loss": 4.604333707116033, + "tokens_seen": 136314880 + }, + { + "epoch": 0.04, + "learning_rate": 0.0007746920762286861, + "loss": 2.0645, + "theoretical_loss": 4.604042089547834, + "tokens_seen": 136380416 + }, + { + "epoch": 0.04, + "learning_rate": 0.0007746760280842529, + "loss": 1.9811, + "theoretical_loss": 4.603750651295222, + "tokens_seen": 136445952 + }, + { + "epoch": 0.04, + "learning_rate": 0.0007746599799398195, + "loss": 1.9133, + "theoretical_loss": 4.603459392161888, + "tokens_seen": 136511488 + }, + { + "epoch": 0.04, + "learning_rate": 0.0007746439317953863, + "loss": 1.964, + "theoretical_loss": 4.603168311951835, + "tokens_seen": 136577024 + }, + { + "epoch": 0.04, + "learning_rate": 0.0007746278836509529, + "loss": 2.1206, + "theoretical_loss": 4.60287741046937, + "tokens_seen": 136642560 + }, + { + "epoch": 0.04, + "learning_rate": 0.0007746118355065197, + "loss": 1.896, + "theoretical_loss": 4.602586687519111, + "tokens_seen": 136708096 + }, + { + "epoch": 0.04, + "learning_rate": 0.0007745957873620863, + "loss": 1.9813, + "theoretical_loss": 4.602296142905981, + "tokens_seen": 136773632 + }, + { + "epoch": 0.04, + "learning_rate": 0.0007745797392176531, + "loss": 2.0728, + "theoretical_loss": 4.602005776435211, + "tokens_seen": 136839168 + }, + { + "epoch": 0.04, + "learning_rate": 0.0007745636910732197, + "loss": 1.8462, + "theoretical_loss": 4.601715587912338, + "tokens_seen": 136904704 + }, + { + "epoch": 0.04, + "learning_rate": 0.0007745476429287865, + "loss": 1.9378, + "theoretical_loss": 4.601425577143203, + "tokens_seen": 136970240 + }, + { + "epoch": 0.04, + "learning_rate": 0.0007745315947843531, + "loss": 1.8559, + "theoretical_loss": 4.601135743933952, + "tokens_seen": 137035776 + }, + { + "epoch": 0.04, + "learning_rate": 0.0007745155466399198, + "loss": 1.9412, + "theoretical_loss": 4.600846088091036, + "tokens_seen": 137101312 + }, + { + "epoch": 0.04, + "learning_rate": 0.0007744994984954865, + "loss": 1.8996, + "theoretical_loss": 4.600556609421206, + "tokens_seen": 137166848 + }, + { + "epoch": 0.04, + "learning_rate": 0.0007744834503510532, + "loss": 2.0113, + "theoretical_loss": 4.600267307731521, + "tokens_seen": 137232384 + }, + { + "epoch": 0.04, + "learning_rate": 0.0007744674022066199, + "loss": 1.9956, + "theoretical_loss": 4.5999781828293385, + "tokens_seen": 137297920 + }, + { + "epoch": 0.04, + "learning_rate": 0.0007744513540621866, + "loss": 1.93, + "theoretical_loss": 4.5996892345223195, + "tokens_seen": 137363456 + }, + { + "epoch": 0.04, + "learning_rate": 0.0007744353059177533, + "loss": 1.951, + "theoretical_loss": 4.599400462618423, + "tokens_seen": 137428992 + }, + { + "epoch": 0.04, + "learning_rate": 0.00077441925777332, + "loss": 1.8711, + "theoretical_loss": 4.59911186692591, + "tokens_seen": 137494528 + }, + { + "epoch": 0.04, + "learning_rate": 0.0007744032096288867, + "loss": 1.8381, + "theoretical_loss": 4.598823447253343, + "tokens_seen": 137560064 + }, + { + "epoch": 0.04, + "objective/train/docs_used": 57180, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 2.1245598793029785, + "objective/train/theoretical_loss": 4.598535203409582, + "objective/train/tokens_used": 158085600, + "theoretical_loss": 4.598535203409582, + "tokens_seen": 137625600 + }, + { + "epoch": 0.04, + "learning_rate": 0.0007743871614844534, + "loss": 2.0623, + "theoretical_loss": 4.598535203409582, + "tokens_seen": 137625600 + }, + { + "epoch": 0.04, + "learning_rate": 0.00077437111334002, + "loss": 1.876, + "theoretical_loss": 4.598247135203783, + "tokens_seen": 137691136 + }, + { + "epoch": 0.04, + "learning_rate": 0.0007743550651955868, + "loss": 1.8465, + "theoretical_loss": 4.5979592424454045, + "tokens_seen": 137756672 + }, + { + "epoch": 0.04, + "learning_rate": 0.0007743390170511534, + "loss": 1.9438, + "theoretical_loss": 4.597671524944198, + "tokens_seen": 137822208 + }, + { + "epoch": 0.04, + "learning_rate": 0.0007743229689067202, + "loss": 2.054, + "theoretical_loss": 4.597383982510216, + "tokens_seen": 137887744 + }, + { + "epoch": 0.04, + "learning_rate": 0.0007743069207622868, + "loss": 1.8707, + "theoretical_loss": 4.597096614953802, + "tokens_seen": 137953280 + }, + { + "epoch": 0.04, + "learning_rate": 0.0007742908726178536, + "loss": 1.8806, + "theoretical_loss": 4.596809422085599, + "tokens_seen": 138018816 + }, + { + "epoch": 0.04, + "learning_rate": 0.0007742748244734202, + "loss": 1.966, + "theoretical_loss": 4.596522403716542, + "tokens_seen": 138084352 + }, + { + "epoch": 0.04, + "learning_rate": 0.000774258776328987, + "loss": 2.0858, + "theoretical_loss": 4.596235559657863, + "tokens_seen": 138149888 + }, + { + "epoch": 0.04, + "learning_rate": 0.0007742427281845536, + "loss": 1.7883, + "theoretical_loss": 4.595948889721086, + "tokens_seen": 138215424 + }, + { + "epoch": 0.04, + "learning_rate": 0.0007742266800401204, + "loss": 2.0269, + "theoretical_loss": 4.595662393718026, + "tokens_seen": 138280960 + }, + { + "epoch": 0.04, + "learning_rate": 0.000774210631895687, + "loss": 2.0368, + "theoretical_loss": 4.595376071460796, + "tokens_seen": 138346496 + }, + { + "epoch": 0.04, + "learning_rate": 0.0007741945837512538, + "loss": 1.8723, + "theoretical_loss": 4.595089922761796, + "tokens_seen": 138412032 + }, + { + "epoch": 0.04, + "learning_rate": 0.0007741785356068204, + "loss": 1.9169, + "theoretical_loss": 4.594803947433717, + "tokens_seen": 138477568 + }, + { + "epoch": 0.04, + "learning_rate": 0.0007741624874623872, + "loss": 1.9398, + "theoretical_loss": 4.594518145289545, + "tokens_seen": 138543104 + }, + { + "epoch": 0.04, + "learning_rate": 0.0007741464393179538, + "loss": 2.023, + "theoretical_loss": 4.594232516142553, + "tokens_seen": 138608640 + }, + { + "epoch": 0.04, + "learning_rate": 0.0007741303911735206, + "loss": 2.0685, + "theoretical_loss": 4.593947059806302, + "tokens_seen": 138674176 + }, + { + "epoch": 0.04, + "learning_rate": 0.0007741143430290872, + "loss": 1.8882, + "theoretical_loss": 4.593661776094647, + "tokens_seen": 138739712 + }, + { + "epoch": 0.04, + "learning_rate": 0.000774098294884654, + "loss": 1.8311, + "theoretical_loss": 4.593376664821726, + "tokens_seen": 138805248 + }, + { + "epoch": 0.04, + "learning_rate": 0.0007740822467402206, + "loss": 1.7921, + "theoretical_loss": 4.593091725801967, + "tokens_seen": 138870784 + }, + { + "epoch": 0.04, + "learning_rate": 0.0007740661985957874, + "loss": 1.9446, + "theoretical_loss": 4.592806958850087, + "tokens_seen": 138936320 + }, + { + "epoch": 0.04, + "learning_rate": 0.000774050150451354, + "loss": 1.938, + "theoretical_loss": 4.592522363781086, + "tokens_seen": 139001856 + }, + { + "epoch": 0.04, + "learning_rate": 0.0007740341023069208, + "loss": 1.9288, + "theoretical_loss": 4.592237940410252, + "tokens_seen": 139067392 + }, + { + "epoch": 0.04, + "learning_rate": 0.0007740180541624874, + "loss": 1.8639, + "theoretical_loss": 4.591953688553159, + "tokens_seen": 139132928 + }, + { + "epoch": 0.04, + "learning_rate": 0.0007740020060180542, + "loss": 1.9063, + "theoretical_loss": 4.591669608025663, + "tokens_seen": 139198464 + }, + { + "epoch": 0.04, + "learning_rate": 0.0007739859578736208, + "loss": 2.0279, + "theoretical_loss": 4.591385698643908, + "tokens_seen": 139264000 + }, + { + "epoch": 0.04, + "learning_rate": 0.0007739699097291876, + "loss": 1.9946, + "theoretical_loss": 4.591101960224321, + "tokens_seen": 139329536 + }, + { + "epoch": 0.04, + "learning_rate": 0.0007739538615847542, + "loss": 1.9246, + "theoretical_loss": 4.590818392583608, + "tokens_seen": 139395072 + }, + { + "epoch": 0.04, + "learning_rate": 0.000773937813440321, + "loss": 1.801, + "theoretical_loss": 4.590534995538762, + "tokens_seen": 139460608 + }, + { + "epoch": 0.04, + "learning_rate": 0.0007739217652958876, + "loss": 1.976, + "theoretical_loss": 4.590251768907057, + "tokens_seen": 139526144 + }, + { + "epoch": 0.04, + "learning_rate": 0.0007739057171514544, + "loss": 1.8409, + "theoretical_loss": 4.5899687125060495, + "tokens_seen": 139591680 + }, + { + "epoch": 0.04, + "learning_rate": 0.000773889669007021, + "loss": 2.0215, + "theoretical_loss": 4.589685826153573, + "tokens_seen": 139657216 + }, + { + "epoch": 0.04, + "learning_rate": 0.0007738736208625878, + "loss": 1.8769, + "theoretical_loss": 4.589403109667746, + "tokens_seen": 139722752 + }, + { + "epoch": 0.04, + "learning_rate": 0.0007738575727181545, + "loss": 2.0016, + "theoretical_loss": 4.589120562866961, + "tokens_seen": 139788288 + }, + { + "epoch": 0.04, + "learning_rate": 0.0007738415245737212, + "loss": 1.9915, + "theoretical_loss": 4.588838185569897, + "tokens_seen": 139853824 + }, + { + "epoch": 0.04, + "learning_rate": 0.0007738254764292879, + "loss": 1.8701, + "theoretical_loss": 4.588555977595506, + "tokens_seen": 139919360 + }, + { + "epoch": 0.04, + "learning_rate": 0.0007738094282848546, + "loss": 1.9409, + "theoretical_loss": 4.58827393876302, + "tokens_seen": 139984896 + }, + { + "epoch": 0.04, + "learning_rate": 0.0007737933801404213, + "loss": 1.9258, + "theoretical_loss": 4.587992068891948, + "tokens_seen": 140050432 + }, + { + "epoch": 0.04, + "learning_rate": 0.000773777331995988, + "loss": 1.7907, + "theoretical_loss": 4.587710367802076, + "tokens_seen": 140115968 + }, + { + "epoch": 0.04, + "learning_rate": 0.0007737612838515547, + "loss": 1.911, + "theoretical_loss": 4.587428835313469, + "tokens_seen": 140181504 + }, + { + "epoch": 0.04, + "learning_rate": 0.0007737452357071214, + "loss": 1.9648, + "theoretical_loss": 4.587147471246463, + "tokens_seen": 140247040 + }, + { + "epoch": 0.04, + "learning_rate": 0.0007737291875626881, + "loss": 1.8866, + "theoretical_loss": 4.586866275421673, + "tokens_seen": 140312576 + }, + { + "epoch": 0.04, + "learning_rate": 0.0007737131394182548, + "loss": 1.9725, + "theoretical_loss": 4.586585247659986, + "tokens_seen": 140378112 + }, + { + "epoch": 0.04, + "learning_rate": 0.0007736970912738215, + "loss": 1.8555, + "theoretical_loss": 4.586304387782567, + "tokens_seen": 140443648 + }, + { + "epoch": 0.04, + "learning_rate": 0.0007736810431293882, + "loss": 1.8978, + "theoretical_loss": 4.5860236956108515, + "tokens_seen": 140509184 + }, + { + "epoch": 0.04, + "learning_rate": 0.0007736649949849549, + "loss": 1.9184, + "theoretical_loss": 4.585743170966548, + "tokens_seen": 140574720 + }, + { + "epoch": 0.04, + "learning_rate": 0.0007736489468405216, + "loss": 1.876, + "theoretical_loss": 4.58546281367164, + "tokens_seen": 140640256 + }, + { + "epoch": 0.04, + "learning_rate": 0.0007736328986960883, + "loss": 2.1054, + "theoretical_loss": 4.585182623548381, + "tokens_seen": 140705792 + }, + { + "epoch": 0.04, + "learning_rate": 0.000773616850551655, + "loss": 1.8932, + "theoretical_loss": 4.584902600419295, + "tokens_seen": 140771328 + }, + { + "epoch": 0.04, + "learning_rate": 0.0007736008024072217, + "loss": 1.9339, + "theoretical_loss": 4.584622744107181, + "tokens_seen": 140836864 + }, + { + "epoch": 0.04, + "objective/train/docs_used": 58312, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.7973558902740479, + "objective/train/theoretical_loss": 4.584343054435103, + "objective/train/tokens_used": 161362400, + "theoretical_loss": 4.584343054435103, + "tokens_seen": 140902400 + }, + { + "epoch": 0.04, + "learning_rate": 0.0007735847542627884, + "loss": 1.7845, + "theoretical_loss": 4.584343054435103, + "tokens_seen": 140902400 + }, + { + "epoch": 0.04, + "learning_rate": 0.0007735687061183552, + "loss": 1.925, + "theoretical_loss": 4.5840635312264, + "tokens_seen": 140967936 + }, + { + "epoch": 0.04, + "learning_rate": 0.0007735526579739218, + "loss": 1.8794, + "theoretical_loss": 4.583784174304677, + "tokens_seen": 141033472 + }, + { + "epoch": 0.04, + "learning_rate": 0.0007735366098294886, + "loss": 1.7632, + "theoretical_loss": 4.583504983493809, + "tokens_seen": 141099008 + }, + { + "epoch": 0.04, + "learning_rate": 0.0007735205616850552, + "loss": 2.0098, + "theoretical_loss": 4.583225958617938, + "tokens_seen": 141164544 + }, + { + "epoch": 0.04, + "learning_rate": 0.000773504513540622, + "loss": 1.9389, + "theoretical_loss": 4.582947099501475, + "tokens_seen": 141230080 + }, + { + "epoch": 0.04, + "learning_rate": 0.0007734884653961886, + "loss": 1.929, + "theoretical_loss": 4.5826684059691, + "tokens_seen": 141295616 + }, + { + "epoch": 0.04, + "learning_rate": 0.0007734724172517554, + "loss": 1.9694, + "theoretical_loss": 4.582389877845754, + "tokens_seen": 141361152 + }, + { + "epoch": 0.04, + "learning_rate": 0.000773456369107322, + "loss": 1.9387, + "theoretical_loss": 4.58211151495665, + "tokens_seen": 141426688 + }, + { + "epoch": 0.04, + "learning_rate": 0.0007734403209628888, + "loss": 1.7114, + "theoretical_loss": 4.581833317127263, + "tokens_seen": 141492224 + }, + { + "epoch": 0.04, + "learning_rate": 0.0007734242728184554, + "loss": 1.9004, + "theoretical_loss": 4.581555284183335, + "tokens_seen": 141557760 + }, + { + "epoch": 0.04, + "learning_rate": 0.0007734082246740222, + "loss": 1.9622, + "theoretical_loss": 4.581277415950872, + "tokens_seen": 141623296 + }, + { + "epoch": 0.04, + "learning_rate": 0.0007733921765295888, + "loss": 1.938, + "theoretical_loss": 4.580999712256142, + "tokens_seen": 141688832 + }, + { + "epoch": 0.04, + "learning_rate": 0.0007733761283851556, + "loss": 1.8896, + "theoretical_loss": 4.5807221729256815, + "tokens_seen": 141754368 + }, + { + "epoch": 0.04, + "learning_rate": 0.0007733600802407222, + "loss": 1.9371, + "theoretical_loss": 4.580444797786283, + "tokens_seen": 141819904 + }, + { + "epoch": 0.04, + "learning_rate": 0.000773344032096289, + "loss": 1.9358, + "theoretical_loss": 4.5801675866650085, + "tokens_seen": 141885440 + }, + { + "epoch": 0.04, + "learning_rate": 0.0007733279839518556, + "loss": 1.996, + "theoretical_loss": 4.579890539389177, + "tokens_seen": 141950976 + }, + { + "epoch": 0.04, + "learning_rate": 0.0007733119358074224, + "loss": 1.95, + "theoretical_loss": 4.579613655786369, + "tokens_seen": 142016512 + }, + { + "epoch": 0.04, + "learning_rate": 0.000773295887662989, + "loss": 1.8811, + "theoretical_loss": 4.579336935684431, + "tokens_seen": 142082048 + }, + { + "epoch": 0.04, + "learning_rate": 0.0007732798395185558, + "loss": 1.926, + "theoretical_loss": 4.579060378911464, + "tokens_seen": 142147584 + }, + { + "epoch": 0.04, + "learning_rate": 0.0007732637913741224, + "loss": 1.8637, + "theoretical_loss": 4.578783985295832, + "tokens_seen": 142213120 + }, + { + "epoch": 0.04, + "learning_rate": 0.0007732477432296892, + "loss": 1.9638, + "theoretical_loss": 4.578507754666157, + "tokens_seen": 142278656 + }, + { + "epoch": 0.04, + "learning_rate": 0.0007732316950852558, + "loss": 2.0417, + "theoretical_loss": 4.57823168685132, + "tokens_seen": 142344192 + }, + { + "epoch": 0.04, + "learning_rate": 0.0007732156469408226, + "loss": 1.9843, + "theoretical_loss": 4.577955781680462, + "tokens_seen": 142409728 + }, + { + "epoch": 0.04, + "learning_rate": 0.0007731995987963892, + "loss": 1.8507, + "theoretical_loss": 4.57768003898298, + "tokens_seen": 142475264 + }, + { + "epoch": 0.04, + "learning_rate": 0.000773183550651956, + "loss": 1.8684, + "theoretical_loss": 4.577404458588529, + "tokens_seen": 142540800 + }, + { + "epoch": 0.04, + "learning_rate": 0.0007731675025075226, + "loss": 1.937, + "theoretical_loss": 4.577129040327021, + "tokens_seen": 142606336 + }, + { + "epoch": 0.04, + "learning_rate": 0.0007731514543630894, + "loss": 1.8978, + "theoretical_loss": 4.576853784028622, + "tokens_seen": 142671872 + }, + { + "epoch": 0.04, + "learning_rate": 0.000773135406218656, + "loss": 1.8239, + "theoretical_loss": 4.576578689523759, + "tokens_seen": 142737408 + }, + { + "epoch": 0.04, + "learning_rate": 0.0007731193580742227, + "loss": 1.9249, + "theoretical_loss": 4.576303756643108, + "tokens_seen": 142802944 + }, + { + "epoch": 0.04, + "learning_rate": 0.0007731033099297894, + "loss": 1.8761, + "theoretical_loss": 4.576028985217606, + "tokens_seen": 142868480 + }, + { + "epoch": 0.04, + "learning_rate": 0.0007730872617853561, + "loss": 1.9687, + "theoretical_loss": 4.5757543750784375, + "tokens_seen": 142934016 + }, + { + "epoch": 0.04, + "learning_rate": 0.0007730712136409228, + "loss": 1.9482, + "theoretical_loss": 4.5754799260570485, + "tokens_seen": 142999552 + }, + { + "epoch": 0.04, + "learning_rate": 0.0007730551654964895, + "loss": 2.0182, + "theoretical_loss": 4.575205637985132, + "tokens_seen": 143065088 + }, + { + "epoch": 0.04, + "learning_rate": 0.0007730391173520562, + "loss": 1.9988, + "theoretical_loss": 4.574931510694635, + "tokens_seen": 143130624 + }, + { + "epoch": 0.04, + "learning_rate": 0.0007730230692076229, + "loss": 1.8299, + "theoretical_loss": 4.574657544017761, + "tokens_seen": 143196160 + }, + { + "epoch": 0.04, + "learning_rate": 0.0007730070210631896, + "loss": 1.9824, + "theoretical_loss": 4.57438373778696, + "tokens_seen": 143261696 + }, + { + "epoch": 0.04, + "learning_rate": 0.0007729909729187563, + "loss": 1.9357, + "theoretical_loss": 4.574110091834936, + "tokens_seen": 143327232 + }, + { + "epoch": 0.04, + "learning_rate": 0.0007729749247743229, + "loss": 1.9032, + "theoretical_loss": 4.573836605994645, + "tokens_seen": 143392768 + }, + { + "epoch": 0.04, + "learning_rate": 0.0007729588766298897, + "loss": 1.9374, + "theoretical_loss": 4.573563280099288, + "tokens_seen": 143458304 + }, + { + "epoch": 0.04, + "learning_rate": 0.0007729428284854563, + "loss": 1.9955, + "theoretical_loss": 4.573290113982324, + "tokens_seen": 143523840 + }, + { + "epoch": 0.04, + "learning_rate": 0.0007729267803410231, + "loss": 1.7579, + "theoretical_loss": 4.573017107477455, + "tokens_seen": 143589376 + }, + { + "epoch": 0.04, + "learning_rate": 0.0007729107321965897, + "loss": 1.9261, + "theoretical_loss": 4.572744260418633, + "tokens_seen": 143654912 + }, + { + "epoch": 0.04, + "learning_rate": 0.0007728946840521565, + "loss": 1.9801, + "theoretical_loss": 4.572471572640059, + "tokens_seen": 143720448 + }, + { + "epoch": 0.04, + "learning_rate": 0.0007728786359077231, + "loss": 1.984, + "theoretical_loss": 4.572199043976185, + "tokens_seen": 143785984 + }, + { + "epoch": 0.04, + "learning_rate": 0.0007728625877632899, + "loss": 1.8353, + "theoretical_loss": 4.571926674261705, + "tokens_seen": 143851520 + }, + { + "epoch": 0.04, + "learning_rate": 0.0007728465396188565, + "loss": 1.787, + "theoretical_loss": 4.571654463331563, + "tokens_seen": 143917056 + }, + { + "epoch": 0.04, + "learning_rate": 0.0007728304914744233, + "loss": 1.9136, + "theoretical_loss": 4.571382411020949, + "tokens_seen": 143982592 + }, + { + "epoch": 0.04, + "learning_rate": 0.0007728144433299899, + "loss": 1.9295, + "theoretical_loss": 4.5711105171653, + "tokens_seen": 144048128 + }, + { + "epoch": 0.04, + "learning_rate": 0.0007727983951855567, + "loss": 1.9239, + "theoretical_loss": 4.570838781600296, + "tokens_seen": 144113664 + }, + { + "epoch": 0.04, + "objective/train/docs_used": 59583, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.8416532278060913, + "objective/train/theoretical_loss": 4.570567204161865, + "objective/train/tokens_used": 164639200, + "theoretical_loss": 4.570567204161865, + "tokens_seen": 144179200 + }, + { + "epoch": 0.04, + "learning_rate": 0.0007727823470411233, + "loss": 1.9239, + "theoretical_loss": 4.570567204161865, + "tokens_seen": 144179200 + }, + { + "epoch": 0.04, + "learning_rate": 0.0007727662988966901, + "loss": 1.8244, + "theoretical_loss": 4.570295784686176, + "tokens_seen": 144244736 + }, + { + "epoch": 0.04, + "learning_rate": 0.0007727502507522567, + "loss": 1.9547, + "theoretical_loss": 4.570024523009646, + "tokens_seen": 144310272 + }, + { + "epoch": 0.04, + "learning_rate": 0.0007727342026078235, + "loss": 1.9154, + "theoretical_loss": 4.569753418968933, + "tokens_seen": 144375808 + }, + { + "epoch": 0.04, + "learning_rate": 0.0007727181544633901, + "loss": 1.8646, + "theoretical_loss": 4.569482472400939, + "tokens_seen": 144441344 + }, + { + "epoch": 0.04, + "learning_rate": 0.0007727021063189569, + "loss": 1.9339, + "theoretical_loss": 4.5692116831428065, + "tokens_seen": 144506880 + }, + { + "epoch": 0.04, + "learning_rate": 0.0007726860581745235, + "loss": 1.8719, + "theoretical_loss": 4.568941051031925, + "tokens_seen": 144572416 + }, + { + "epoch": 0.04, + "learning_rate": 0.0007726700100300903, + "loss": 1.9144, + "theoretical_loss": 4.568670575905921, + "tokens_seen": 144637952 + }, + { + "epoch": 0.04, + "learning_rate": 0.0007726539618856569, + "loss": 1.9723, + "theoretical_loss": 4.568400257602666, + "tokens_seen": 144703488 + }, + { + "epoch": 0.04, + "learning_rate": 0.0007726379137412237, + "loss": 1.9682, + "theoretical_loss": 4.568130095960268, + "tokens_seen": 144769024 + }, + { + "epoch": 0.04, + "learning_rate": 0.0007726218655967903, + "loss": 1.9649, + "theoretical_loss": 4.56786009081708, + "tokens_seen": 144834560 + }, + { + "epoch": 0.04, + "learning_rate": 0.0007726058174523571, + "loss": 1.9064, + "theoretical_loss": 4.567590242011692, + "tokens_seen": 144900096 + }, + { + "epoch": 0.04, + "learning_rate": 0.0007725897693079237, + "loss": 1.9126, + "theoretical_loss": 4.567320549382931, + "tokens_seen": 144965632 + }, + { + "epoch": 0.04, + "learning_rate": 0.0007725737211634905, + "loss": 1.9397, + "theoretical_loss": 4.56705101276987, + "tokens_seen": 145031168 + }, + { + "epoch": 0.04, + "learning_rate": 0.0007725576730190571, + "loss": 1.8601, + "theoretical_loss": 4.566781632011814, + "tokens_seen": 145096704 + }, + { + "epoch": 0.04, + "learning_rate": 0.0007725416248746239, + "loss": 1.8038, + "theoretical_loss": 4.566512406948309, + "tokens_seen": 145162240 + }, + { + "epoch": 0.04, + "learning_rate": 0.0007725255767301905, + "loss": 1.8976, + "theoretical_loss": 4.566243337419137, + "tokens_seen": 145227776 + }, + { + "epoch": 0.04, + "learning_rate": 0.0007725095285857573, + "loss": 1.8537, + "theoretical_loss": 4.5659744232643185, + "tokens_seen": 145293312 + }, + { + "epoch": 0.04, + "learning_rate": 0.000772493480441324, + "loss": 2.0941, + "theoretical_loss": 4.565705664324111, + "tokens_seen": 145358848 + }, + { + "epoch": 0.04, + "learning_rate": 0.0007724774322968907, + "loss": 1.784, + "theoretical_loss": 4.565437060439006, + "tokens_seen": 145424384 + }, + { + "epoch": 0.04, + "learning_rate": 0.0007724613841524574, + "loss": 1.9471, + "theoretical_loss": 4.565168611449732, + "tokens_seen": 145489920 + }, + { + "epoch": 0.04, + "learning_rate": 0.0007724453360080241, + "loss": 1.9164, + "theoretical_loss": 4.564900317197251, + "tokens_seen": 145555456 + }, + { + "epoch": 0.04, + "learning_rate": 0.0007724292878635908, + "loss": 2.0431, + "theoretical_loss": 4.564632177522766, + "tokens_seen": 145620992 + }, + { + "epoch": 0.04, + "learning_rate": 0.0007724132397191575, + "loss": 1.8891, + "theoretical_loss": 4.564364192267705, + "tokens_seen": 145686528 + }, + { + "epoch": 0.04, + "learning_rate": 0.0007723971915747242, + "loss": 1.9223, + "theoretical_loss": 4.564096361273736, + "tokens_seen": 145752064 + }, + { + "epoch": 0.04, + "learning_rate": 0.0007723811434302909, + "loss": 1.9461, + "theoretical_loss": 4.563828684382761, + "tokens_seen": 145817600 + }, + { + "epoch": 0.04, + "learning_rate": 0.0007723650952858577, + "loss": 1.9824, + "theoretical_loss": 4.563561161436912, + "tokens_seen": 145883136 + }, + { + "epoch": 0.04, + "learning_rate": 0.0007723490471414243, + "loss": 1.7434, + "theoretical_loss": 4.563293792278554, + "tokens_seen": 145948672 + }, + { + "epoch": 0.04, + "learning_rate": 0.0007723329989969911, + "loss": 1.8492, + "theoretical_loss": 4.563026576750286, + "tokens_seen": 146014208 + }, + { + "epoch": 0.04, + "learning_rate": 0.0007723169508525577, + "loss": 1.8973, + "theoretical_loss": 4.562759514694935, + "tokens_seen": 146079744 + }, + { + "epoch": 0.04, + "learning_rate": 0.0007723009027081245, + "loss": 1.9138, + "theoretical_loss": 4.5624926059555655, + "tokens_seen": 146145280 + }, + { + "epoch": 0.04, + "learning_rate": 0.0007722848545636911, + "loss": 1.8851, + "theoretical_loss": 4.562225850375466, + "tokens_seen": 146210816 + }, + { + "epoch": 0.04, + "learning_rate": 0.0007722688064192579, + "loss": 1.895, + "theoretical_loss": 4.561959247798159, + "tokens_seen": 146276352 + }, + { + "epoch": 0.04, + "learning_rate": 0.0007722527582748245, + "loss": 1.8918, + "theoretical_loss": 4.561692798067396, + "tokens_seen": 146341888 + }, + { + "epoch": 0.04, + "learning_rate": 0.0007722367101303913, + "loss": 1.8126, + "theoretical_loss": 4.561426501027158, + "tokens_seen": 146407424 + }, + { + "epoch": 0.04, + "learning_rate": 0.0007722206619859579, + "loss": 1.92, + "theoretical_loss": 4.561160356521656, + "tokens_seen": 146472960 + }, + { + "epoch": 0.04, + "learning_rate": 0.0007722046138415247, + "loss": 1.8422, + "theoretical_loss": 4.5608943643953275, + "tokens_seen": 146538496 + }, + { + "epoch": 0.04, + "learning_rate": 0.0007721885656970913, + "loss": 1.9869, + "theoretical_loss": 4.56062852449284, + "tokens_seen": 146604032 + }, + { + "epoch": 0.04, + "learning_rate": 0.0007721725175526581, + "loss": 1.8371, + "theoretical_loss": 4.5603628366590865, + "tokens_seen": 146669568 + }, + { + "epoch": 0.04, + "learning_rate": 0.0007721564694082247, + "loss": 2.03, + "theoretical_loss": 4.560097300739191, + "tokens_seen": 146735104 + }, + { + "epoch": 0.04, + "learning_rate": 0.0007721404212637915, + "loss": 1.9058, + "theoretical_loss": 4.5598319165785, + "tokens_seen": 146800640 + }, + { + "epoch": 0.04, + "learning_rate": 0.0007721243731193581, + "loss": 1.8956, + "theoretical_loss": 4.55956668402259, + "tokens_seen": 146866176 + }, + { + "epoch": 0.04, + "learning_rate": 0.0007721083249749249, + "loss": 1.7517, + "theoretical_loss": 4.559301602917261, + "tokens_seen": 146931712 + }, + { + "epoch": 0.04, + "learning_rate": 0.0007720922768304915, + "loss": 1.8711, + "theoretical_loss": 4.55903667310854, + "tokens_seen": 146997248 + }, + { + "epoch": 0.04, + "learning_rate": 0.0007720762286860583, + "loss": 1.8376, + "theoretical_loss": 4.558771894442677, + "tokens_seen": 147062784 + }, + { + "epoch": 0.04, + "learning_rate": 0.0007720601805416249, + "loss": 1.9175, + "theoretical_loss": 4.558507266766151, + "tokens_seen": 147128320 + }, + { + "epoch": 0.04, + "learning_rate": 0.0007720441323971917, + "loss": 1.9104, + "theoretical_loss": 4.558242789925659, + "tokens_seen": 147193856 + }, + { + "epoch": 0.04, + "learning_rate": 0.0007720280842527583, + "loss": 1.8095, + "theoretical_loss": 4.557978463768127, + "tokens_seen": 147259392 + }, + { + "epoch": 0.04, + "learning_rate": 0.0007720120361083251, + "loss": 1.9069, + "theoretical_loss": 4.557714288140703, + "tokens_seen": 147324928 + }, + { + "epoch": 0.04, + "learning_rate": 0.0007719959879638917, + "loss": 1.8426, + "theoretical_loss": 4.557450262890757, + "tokens_seen": 147390464 + }, + { + "epoch": 0.04, + "objective/train/docs_used": 60747, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.758107304573059, + "objective/train/theoretical_loss": 4.557186387865883, + "objective/train/tokens_used": 167916000, + "theoretical_loss": 4.557186387865883, + "tokens_seen": 147456000 + }, + { + "epoch": 0.04, + "learning_rate": 0.0007719799398194585, + "loss": 1.7949, + "theoretical_loss": 4.557186387865883, + "tokens_seen": 147456000 + }, + { + "epoch": 0.04, + "learning_rate": 0.0007719638916750251, + "loss": 1.8605, + "theoretical_loss": 4.556922662913895, + "tokens_seen": 147521536 + }, + { + "epoch": 0.04, + "learning_rate": 0.0007719478435305919, + "loss": 1.883, + "theoretical_loss": 4.55665908788283, + "tokens_seen": 147587072 + }, + { + "epoch": 0.04, + "learning_rate": 0.0007719317953861585, + "loss": 1.9864, + "theoretical_loss": 4.556395662620945, + "tokens_seen": 147652608 + }, + { + "epoch": 0.04, + "learning_rate": 0.0007719157472417253, + "loss": 1.8182, + "theoretical_loss": 4.5561323869767225, + "tokens_seen": 147718144 + }, + { + "epoch": 0.04, + "learning_rate": 0.0007718996990972919, + "loss": 1.8828, + "theoretical_loss": 4.555869260798859, + "tokens_seen": 147783680 + }, + { + "epoch": 0.04, + "learning_rate": 0.0007718836509528587, + "loss": 1.9111, + "theoretical_loss": 4.5556062839362745, + "tokens_seen": 147849216 + }, + { + "epoch": 0.04, + "learning_rate": 0.0007718676028084253, + "loss": 1.9322, + "theoretical_loss": 4.555343456238109, + "tokens_seen": 147914752 + }, + { + "epoch": 0.04, + "learning_rate": 0.0007718515546639921, + "loss": 1.969, + "theoretical_loss": 4.55508077755372, + "tokens_seen": 147980288 + }, + { + "epoch": 0.04, + "learning_rate": 0.0007718355065195587, + "loss": 1.8939, + "theoretical_loss": 4.554818247732685, + "tokens_seen": 148045824 + }, + { + "epoch": 0.04, + "learning_rate": 0.0007718194583751255, + "loss": 1.8989, + "theoretical_loss": 4.554555866624798, + "tokens_seen": 148111360 + }, + { + "epoch": 0.04, + "learning_rate": 0.0007718034102306921, + "loss": 1.9176, + "theoretical_loss": 4.554293634080074, + "tokens_seen": 148176896 + }, + { + "epoch": 0.04, + "learning_rate": 0.0007717873620862589, + "loss": 1.8096, + "theoretical_loss": 4.554031549948743, + "tokens_seen": 148242432 + }, + { + "epoch": 0.04, + "learning_rate": 0.0007717713139418255, + "loss": 1.8436, + "theoretical_loss": 4.5537696140812525, + "tokens_seen": 148307968 + }, + { + "epoch": 0.04, + "learning_rate": 0.0007717552657973923, + "loss": 1.9488, + "theoretical_loss": 4.553507826328268, + "tokens_seen": 148373504 + }, + { + "epoch": 0.04, + "learning_rate": 0.0007717392176529589, + "loss": 1.793, + "theoretical_loss": 4.553246186540671, + "tokens_seen": 148439040 + }, + { + "epoch": 0.05, + "learning_rate": 0.0007717231695085256, + "loss": 1.781, + "theoretical_loss": 4.552984694569555, + "tokens_seen": 148504576 + }, + { + "epoch": 0.05, + "learning_rate": 0.0007717071213640923, + "loss": 1.948, + "theoretical_loss": 4.552723350266236, + "tokens_seen": 148570112 + }, + { + "epoch": 0.05, + "learning_rate": 0.000771691073219659, + "loss": 1.8908, + "theoretical_loss": 4.552462153482239, + "tokens_seen": 148635648 + }, + { + "epoch": 0.05, + "learning_rate": 0.0007716750250752257, + "loss": 1.8752, + "theoretical_loss": 4.552201104069307, + "tokens_seen": 148701184 + }, + { + "epoch": 0.05, + "learning_rate": 0.0007716589769307924, + "loss": 1.9313, + "theoretical_loss": 4.551940201879396, + "tokens_seen": 148766720 + }, + { + "epoch": 0.05, + "learning_rate": 0.0007716429287863591, + "loss": 1.7575, + "theoretical_loss": 4.551679446764677, + "tokens_seen": 148832256 + }, + { + "epoch": 0.05, + "learning_rate": 0.0007716268806419258, + "loss": 1.994, + "theoretical_loss": 4.551418838577531, + "tokens_seen": 148897792 + }, + { + "epoch": 0.05, + "learning_rate": 0.0007716108324974925, + "loss": 1.8271, + "theoretical_loss": 4.551158377170557, + "tokens_seen": 148963328 + }, + { + "epoch": 0.05, + "learning_rate": 0.0007715947843530592, + "loss": 1.6936, + "theoretical_loss": 4.5508980623965645, + "tokens_seen": 149028864 + }, + { + "epoch": 0.05, + "learning_rate": 0.0007715787362086258, + "loss": 1.9869, + "theoretical_loss": 4.550637894108572, + "tokens_seen": 149094400 + }, + { + "epoch": 0.05, + "learning_rate": 0.0007715626880641926, + "loss": 1.7649, + "theoretical_loss": 4.5503778721598165, + "tokens_seen": 149159936 + }, + { + "epoch": 0.05, + "learning_rate": 0.0007715466399197592, + "loss": 1.7466, + "theoretical_loss": 4.55011799640374, + "tokens_seen": 149225472 + }, + { + "epoch": 0.05, + "learning_rate": 0.000771530591775326, + "loss": 1.9025, + "theoretical_loss": 4.549858266694001, + "tokens_seen": 149291008 + }, + { + "epoch": 0.05, + "learning_rate": 0.0007715145436308926, + "loss": 1.8453, + "theoretical_loss": 4.549598682884463, + "tokens_seen": 149356544 + }, + { + "epoch": 0.05, + "learning_rate": 0.0007714984954864594, + "loss": 1.8962, + "theoretical_loss": 4.549339244829204, + "tokens_seen": 149422080 + }, + { + "epoch": 0.05, + "learning_rate": 0.000771482447342026, + "loss": 1.7997, + "theoretical_loss": 4.549079952382511, + "tokens_seen": 149487616 + }, + { + "epoch": 0.05, + "learning_rate": 0.0007714663991975928, + "loss": 1.8836, + "theoretical_loss": 4.548820805398881, + "tokens_seen": 149553152 + }, + { + "epoch": 0.05, + "learning_rate": 0.0007714503510531594, + "loss": 1.7376, + "theoretical_loss": 4.548561803733017, + "tokens_seen": 149618688 + }, + { + "epoch": 0.05, + "learning_rate": 0.0007714343029087262, + "loss": 2.0907, + "theoretical_loss": 4.548302947239834, + "tokens_seen": 149684224 + }, + { + "epoch": 0.05, + "learning_rate": 0.0007714182547642928, + "loss": 1.8553, + "theoretical_loss": 4.548044235774454, + "tokens_seen": 149749760 + }, + { + "epoch": 0.05, + "learning_rate": 0.0007714022066198596, + "loss": 1.9082, + "theoretical_loss": 4.547785669192207, + "tokens_seen": 149815296 + }, + { + "epoch": 0.05, + "learning_rate": 0.0007713861584754262, + "loss": 1.9077, + "theoretical_loss": 4.547527247348631, + "tokens_seen": 149880832 + }, + { + "epoch": 0.05, + "learning_rate": 0.000771370110330993, + "loss": 1.8998, + "theoretical_loss": 4.547268970099471, + "tokens_seen": 149946368 + }, + { + "epoch": 0.05, + "learning_rate": 0.0007713540621865596, + "loss": 1.8647, + "theoretical_loss": 4.547010837300677, + "tokens_seen": 150011904 + }, + { + "epoch": 0.05, + "learning_rate": 0.0007713380140421264, + "loss": 1.8943, + "theoretical_loss": 4.546752848808408, + "tokens_seen": 150077440 + }, + { + "epoch": 0.05, + "learning_rate": 0.000771321965897693, + "loss": 1.8652, + "theoretical_loss": 4.546495004479027, + "tokens_seen": 150142976 + }, + { + "epoch": 0.05, + "learning_rate": 0.0007713059177532598, + "loss": 1.7874, + "theoretical_loss": 4.546237304169104, + "tokens_seen": 150208512 + }, + { + "epoch": 0.05, + "learning_rate": 0.0007712898696088264, + "loss": 1.9278, + "theoretical_loss": 4.545979747735413, + "tokens_seen": 150274048 + }, + { + "epoch": 0.05, + "learning_rate": 0.0007712738214643932, + "loss": 1.7275, + "theoretical_loss": 4.5457223350349345, + "tokens_seen": 150339584 + }, + { + "epoch": 0.05, + "learning_rate": 0.0007712577733199598, + "loss": 1.7957, + "theoretical_loss": 4.545465065924851, + "tokens_seen": 150405120 + }, + { + "epoch": 0.05, + "learning_rate": 0.0007712417251755266, + "loss": 1.7762, + "theoretical_loss": 4.54520794026255, + "tokens_seen": 150470656 + }, + { + "epoch": 0.05, + "learning_rate": 0.0007712256770310932, + "loss": 1.7849, + "theoretical_loss": 4.544950957905625, + "tokens_seen": 150536192 + }, + { + "epoch": 0.05, + "learning_rate": 0.00077120962888666, + "loss": 1.827, + "theoretical_loss": 4.54469411871187, + "tokens_seen": 150601728 + }, + { + "epoch": 0.05, + "learning_rate": 0.0007711935807422266, + "loss": 1.792, + "theoretical_loss": 4.544437422539282, + "tokens_seen": 150667264 + }, + { + "epoch": 0.05, + "objective/train/docs_used": 61752, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.8925809860229492, + "objective/train/theoretical_loss": 4.544180869246063, + "objective/train/tokens_used": 171192800, + "theoretical_loss": 4.544180869246063, + "tokens_seen": 150732800 + }, + { + "epoch": 0.05, + "learning_rate": 0.0007711775325977934, + "loss": 1.8854, + "theoretical_loss": 4.544180869246063, + "tokens_seen": 150732800 + }, + { + "epoch": 0.05, + "learning_rate": 0.0007711614844533602, + "loss": 2.0021, + "theoretical_loss": 4.543924458690612, + "tokens_seen": 150798336 + }, + { + "epoch": 0.05, + "learning_rate": 0.0007711454363089268, + "loss": 1.9505, + "theoretical_loss": 4.5436681907315375, + "tokens_seen": 150863872 + }, + { + "epoch": 0.05, + "learning_rate": 0.0007711293881644936, + "loss": 1.9704, + "theoretical_loss": 4.543412065227642, + "tokens_seen": 150929408 + }, + { + "epoch": 0.05, + "learning_rate": 0.0007711133400200602, + "loss": 1.8212, + "theoretical_loss": 4.5431560820379335, + "tokens_seen": 150994944 + }, + { + "epoch": 0.05, + "learning_rate": 0.000771097291875627, + "loss": 1.8285, + "theoretical_loss": 4.54290024102162, + "tokens_seen": 151060480 + }, + { + "epoch": 0.05, + "learning_rate": 0.0007710812437311936, + "loss": 1.8868, + "theoretical_loss": 4.542644542038108, + "tokens_seen": 151126016 + }, + { + "epoch": 0.05, + "learning_rate": 0.0007710651955867604, + "loss": 1.9823, + "theoretical_loss": 4.542388984947005, + "tokens_seen": 151191552 + }, + { + "epoch": 0.05, + "learning_rate": 0.000771049147442327, + "loss": 1.8297, + "theoretical_loss": 4.5421335696081195, + "tokens_seen": 151257088 + }, + { + "epoch": 0.05, + "learning_rate": 0.0007710330992978938, + "loss": 1.8077, + "theoretical_loss": 4.541878295881456, + "tokens_seen": 151322624 + }, + { + "epoch": 0.05, + "learning_rate": 0.0007710170511534604, + "loss": 1.8002, + "theoretical_loss": 4.541623163627221, + "tokens_seen": 151388160 + }, + { + "epoch": 0.05, + "learning_rate": 0.0007710010030090272, + "loss": 1.7638, + "theoretical_loss": 4.541368172705817, + "tokens_seen": 151453696 + }, + { + "epoch": 0.05, + "learning_rate": 0.0007709849548645938, + "loss": 1.8176, + "theoretical_loss": 4.541113322977846, + "tokens_seen": 151519232 + }, + { + "epoch": 0.05, + "learning_rate": 0.0007709689067201606, + "loss": 1.8712, + "theoretical_loss": 4.540858614304108, + "tokens_seen": 151584768 + }, + { + "epoch": 0.05, + "learning_rate": 0.0007709528585757272, + "loss": 1.8013, + "theoretical_loss": 4.5406040465455995, + "tokens_seen": 151650304 + }, + { + "epoch": 0.05, + "learning_rate": 0.000770936810431294, + "loss": 1.7272, + "theoretical_loss": 4.5403496195635125, + "tokens_seen": 151715840 + }, + { + "epoch": 0.05, + "learning_rate": 0.0007709207622868606, + "loss": 1.9183, + "theoretical_loss": 4.54009533321924, + "tokens_seen": 151781376 + }, + { + "epoch": 0.05, + "learning_rate": 0.0007709047141424274, + "loss": 1.7625, + "theoretical_loss": 4.539841187374366, + "tokens_seen": 151846912 + }, + { + "epoch": 0.05, + "learning_rate": 0.000770888665997994, + "loss": 1.8412, + "theoretical_loss": 4.539587181890674, + "tokens_seen": 151912448 + }, + { + "epoch": 0.05, + "learning_rate": 0.0007708726178535608, + "loss": 1.8532, + "theoretical_loss": 4.539333316630143, + "tokens_seen": 151977984 + }, + { + "epoch": 0.05, + "learning_rate": 0.0007708565697091274, + "loss": 1.7363, + "theoretical_loss": 4.539079591454943, + "tokens_seen": 152043520 + }, + { + "epoch": 0.05, + "learning_rate": 0.0007708405215646942, + "loss": 1.9588, + "theoretical_loss": 4.5388260062274455, + "tokens_seen": 152109056 + }, + { + "epoch": 0.05, + "learning_rate": 0.0007708244734202608, + "loss": 1.8086, + "theoretical_loss": 4.538572560810211, + "tokens_seen": 152174592 + }, + { + "epoch": 0.05, + "learning_rate": 0.0007708084252758276, + "loss": 1.7992, + "theoretical_loss": 4.538319255065995, + "tokens_seen": 152240128 + }, + { + "epoch": 0.05, + "learning_rate": 0.0007707923771313942, + "loss": 1.7912, + "theoretical_loss": 4.53806608885775, + "tokens_seen": 152305664 + }, + { + "epoch": 0.05, + "learning_rate": 0.000770776328986961, + "loss": 2.0232, + "theoretical_loss": 4.537813062048618, + "tokens_seen": 152371200 + }, + { + "epoch": 0.05, + "learning_rate": 0.0007707602808425276, + "loss": 1.7682, + "theoretical_loss": 4.537560174501936, + "tokens_seen": 152436736 + }, + { + "epoch": 0.05, + "learning_rate": 0.0007707442326980944, + "loss": 1.8597, + "theoretical_loss": 4.537307426081232, + "tokens_seen": 152502272 + }, + { + "epoch": 0.05, + "learning_rate": 0.000770728184553661, + "loss": 1.9438, + "theoretical_loss": 4.53705481665023, + "tokens_seen": 152567808 + }, + { + "epoch": 0.05, + "learning_rate": 0.0007707121364092278, + "loss": 1.9179, + "theoretical_loss": 4.536802346072841, + "tokens_seen": 152633344 + }, + { + "epoch": 0.05, + "learning_rate": 0.0007706960882647944, + "loss": 1.8341, + "theoretical_loss": 4.536550014213171, + "tokens_seen": 152698880 + }, + { + "epoch": 0.05, + "learning_rate": 0.0007706800401203612, + "loss": 1.8894, + "theoretical_loss": 4.536297820935517, + "tokens_seen": 152764416 + }, + { + "epoch": 0.05, + "learning_rate": 0.0007706639919759278, + "loss": 1.8117, + "theoretical_loss": 4.536045766104366, + "tokens_seen": 152829952 + }, + { + "epoch": 0.05, + "learning_rate": 0.0007706479438314946, + "loss": 1.9594, + "theoretical_loss": 4.535793849584396, + "tokens_seen": 152895488 + }, + { + "epoch": 0.05, + "learning_rate": 0.0007706318956870612, + "loss": 1.9887, + "theoretical_loss": 4.535542071240473, + "tokens_seen": 152961024 + }, + { + "epoch": 0.05, + "learning_rate": 0.000770615847542628, + "loss": 1.871, + "theoretical_loss": 4.535290430937657, + "tokens_seen": 153026560 + }, + { + "epoch": 0.05, + "learning_rate": 0.0007705997993981946, + "loss": 1.8562, + "theoretical_loss": 4.535038928541194, + "tokens_seen": 153092096 + }, + { + "epoch": 0.05, + "learning_rate": 0.0007705837512537614, + "loss": 1.8401, + "theoretical_loss": 4.534787563916522, + "tokens_seen": 153157632 + }, + { + "epoch": 0.05, + "learning_rate": 0.000770567703109328, + "loss": 1.739, + "theoretical_loss": 4.534536336929264, + "tokens_seen": 153223168 + }, + { + "epoch": 0.05, + "learning_rate": 0.0007705516549648948, + "loss": 1.9402, + "theoretical_loss": 4.534285247445236, + "tokens_seen": 153288704 + }, + { + "epoch": 0.05, + "learning_rate": 0.0007705356068204614, + "loss": 1.8792, + "theoretical_loss": 4.534034295330439, + "tokens_seen": 153354240 + }, + { + "epoch": 0.05, + "learning_rate": 0.0007705195586760282, + "loss": 1.7967, + "theoretical_loss": 4.5337834804510635, + "tokens_seen": 153419776 + }, + { + "epoch": 0.05, + "learning_rate": 0.0007705035105315948, + "loss": 2.0127, + "theoretical_loss": 4.533532802673484, + "tokens_seen": 153485312 + }, + { + "epoch": 0.05, + "learning_rate": 0.0007704874623871616, + "loss": 1.8084, + "theoretical_loss": 4.533282261864267, + "tokens_seen": 153550848 + }, + { + "epoch": 0.05, + "learning_rate": 0.0007704714142427282, + "loss": 1.978, + "theoretical_loss": 4.533031857890164, + "tokens_seen": 153616384 + }, + { + "epoch": 0.05, + "learning_rate": 0.000770455366098295, + "loss": 1.8562, + "theoretical_loss": 4.532781590618109, + "tokens_seen": 153681920 + }, + { + "epoch": 0.05, + "learning_rate": 0.0007704393179538616, + "loss": 1.9746, + "theoretical_loss": 4.532531459915229, + "tokens_seen": 153747456 + }, + { + "epoch": 0.05, + "learning_rate": 0.0007704232698094284, + "loss": 1.9016, + "theoretical_loss": 4.532281465648831, + "tokens_seen": 153812992 + }, + { + "epoch": 0.05, + "learning_rate": 0.000770407221664995, + "loss": 1.8911, + "theoretical_loss": 4.532031607686411, + "tokens_seen": 153878528 + }, + { + "epoch": 0.05, + "learning_rate": 0.0007703911735205618, + "loss": 1.8613, + "theoretical_loss": 4.531781885895647, + "tokens_seen": 153944064 + }, + { + "epoch": 0.05, + "objective/train/docs_used": 63027, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 2.079739809036255, + "objective/train/theoretical_loss": 4.531532300144404, + "objective/train/tokens_used": 174469600, + "theoretical_loss": 4.531532300144404, + "tokens_seen": 154009600 + }, + { + "epoch": 0.05, + "learning_rate": 0.0007703751253761284, + "loss": 1.9679, + "theoretical_loss": 4.531532300144404, + "tokens_seen": 154009600 + }, + { + "epoch": 0.05, + "learning_rate": 0.0007703590772316952, + "loss": 1.9205, + "theoretical_loss": 4.531282850300731, + "tokens_seen": 154075136 + }, + { + "epoch": 0.05, + "learning_rate": 0.0007703430290872618, + "loss": 2.0744, + "theoretical_loss": 4.53103353623286, + "tokens_seen": 154140672 + }, + { + "epoch": 0.05, + "learning_rate": 0.0007703269809428285, + "loss": 1.8919, + "theoretical_loss": 4.530784357809209, + "tokens_seen": 154206208 + }, + { + "epoch": 0.05, + "learning_rate": 0.0007703109327983952, + "loss": 1.7733, + "theoretical_loss": 4.5305353148983745, + "tokens_seen": 154271744 + }, + { + "epoch": 0.05, + "learning_rate": 0.0007702948846539619, + "loss": 1.8827, + "theoretical_loss": 4.530286407369143, + "tokens_seen": 154337280 + }, + { + "epoch": 0.05, + "learning_rate": 0.0007702788365095286, + "loss": 2.0005, + "theoretical_loss": 4.530037635090476, + "tokens_seen": 154402816 + }, + { + "epoch": 0.05, + "learning_rate": 0.0007702627883650953, + "loss": 1.9451, + "theoretical_loss": 4.529788997931526, + "tokens_seen": 154468352 + }, + { + "epoch": 0.05, + "learning_rate": 0.000770246740220662, + "loss": 1.8355, + "theoretical_loss": 4.52954049576162, + "tokens_seen": 154533888 + }, + { + "epoch": 0.05, + "learning_rate": 0.0007702306920762287, + "loss": 1.9377, + "theoretical_loss": 4.529292128450271, + "tokens_seen": 154599424 + }, + { + "epoch": 0.05, + "learning_rate": 0.0007702146439317954, + "loss": 1.8837, + "theoretical_loss": 4.52904389586717, + "tokens_seen": 154664960 + }, + { + "epoch": 0.05, + "learning_rate": 0.0007701985957873621, + "loss": 1.8634, + "theoretical_loss": 4.528795797882193, + "tokens_seen": 154730496 + }, + { + "epoch": 0.05, + "learning_rate": 0.0007701825476429287, + "loss": 1.8578, + "theoretical_loss": 4.528547834365394, + "tokens_seen": 154796032 + }, + { + "epoch": 0.05, + "learning_rate": 0.0007701664994984955, + "loss": 1.7841, + "theoretical_loss": 4.52830000518701, + "tokens_seen": 154861568 + }, + { + "epoch": 0.05, + "learning_rate": 0.0007701504513540621, + "loss": 1.8531, + "theoretical_loss": 4.528052310217453, + "tokens_seen": 154927104 + }, + { + "epoch": 0.05, + "learning_rate": 0.0007701344032096289, + "loss": 1.7901, + "theoretical_loss": 4.527804749327321, + "tokens_seen": 154992640 + }, + { + "epoch": 0.05, + "learning_rate": 0.0007701183550651955, + "loss": 1.9865, + "theoretical_loss": 4.527557322387387, + "tokens_seen": 155058176 + }, + { + "epoch": 0.05, + "learning_rate": 0.0007701023069207623, + "loss": 1.9191, + "theoretical_loss": 4.527310029268604, + "tokens_seen": 155123712 + }, + { + "epoch": 0.05, + "learning_rate": 0.0007700862587763289, + "loss": 1.8633, + "theoretical_loss": 4.527062869842107, + "tokens_seen": 155189248 + }, + { + "epoch": 0.05, + "learning_rate": 0.0007700702106318957, + "loss": 1.9372, + "theoretical_loss": 4.526815843979204, + "tokens_seen": 155254784 + }, + { + "epoch": 0.05, + "learning_rate": 0.0007700541624874623, + "loss": 1.9868, + "theoretical_loss": 4.5265689515513845, + "tokens_seen": 155320320 + }, + { + "epoch": 0.05, + "learning_rate": 0.0007700381143430291, + "loss": 1.9034, + "theoretical_loss": 4.526322192430316, + "tokens_seen": 155385856 + }, + { + "epoch": 0.05, + "learning_rate": 0.0007700220661985957, + "loss": 1.8292, + "theoretical_loss": 4.526075566487842, + "tokens_seen": 155451392 + }, + { + "epoch": 0.05, + "learning_rate": 0.0007700060180541625, + "loss": 1.7711, + "theoretical_loss": 4.525829073595984, + "tokens_seen": 155516928 + }, + { + "epoch": 0.05, + "learning_rate": 0.0007699899699097291, + "loss": 1.7064, + "theoretical_loss": 4.5255827136269415, + "tokens_seen": 155582464 + }, + { + "epoch": 0.05, + "learning_rate": 0.0007699739217652959, + "loss": 1.9416, + "theoretical_loss": 4.525336486453088, + "tokens_seen": 155648000 + }, + { + "epoch": 0.05, + "learning_rate": 0.0007699578736208627, + "loss": 1.9699, + "theoretical_loss": 4.525090391946974, + "tokens_seen": 155713536 + }, + { + "epoch": 0.05, + "learning_rate": 0.0007699418254764293, + "loss": 1.8017, + "theoretical_loss": 4.524844429981327, + "tokens_seen": 155779072 + }, + { + "epoch": 0.05, + "learning_rate": 0.000769925777331996, + "loss": 1.8503, + "theoretical_loss": 4.52459860042905, + "tokens_seen": 155844608 + }, + { + "epoch": 0.05, + "learning_rate": 0.0007699097291875627, + "loss": 2.0436, + "theoretical_loss": 4.524352903163219, + "tokens_seen": 155910144 + }, + { + "epoch": 0.05, + "learning_rate": 0.0007698936810431295, + "loss": 1.7407, + "theoretical_loss": 4.524107338057089, + "tokens_seen": 155975680 + }, + { + "epoch": 0.05, + "learning_rate": 0.0007698776328986961, + "loss": 1.8944, + "theoretical_loss": 4.523861904984086, + "tokens_seen": 156041216 + }, + { + "epoch": 0.05, + "learning_rate": 0.0007698615847542629, + "loss": 1.7379, + "theoretical_loss": 4.5236166038178105, + "tokens_seen": 156106752 + }, + { + "epoch": 0.05, + "learning_rate": 0.0007698455366098295, + "loss": 1.8291, + "theoretical_loss": 4.52337143443204, + "tokens_seen": 156172288 + }, + { + "epoch": 0.05, + "learning_rate": 0.0007698294884653963, + "loss": 1.9129, + "theoretical_loss": 4.523126396700722, + "tokens_seen": 156237824 + }, + { + "epoch": 0.05, + "learning_rate": 0.0007698134403209629, + "loss": 1.9769, + "theoretical_loss": 4.522881490497981, + "tokens_seen": 156303360 + }, + { + "epoch": 0.05, + "learning_rate": 0.0007697973921765297, + "loss": 1.8189, + "theoretical_loss": 4.522636715698111, + "tokens_seen": 156368896 + }, + { + "epoch": 0.05, + "learning_rate": 0.0007697813440320963, + "loss": 1.8737, + "theoretical_loss": 4.522392072175582, + "tokens_seen": 156434432 + }, + { + "epoch": 0.05, + "learning_rate": 0.0007697652958876631, + "loss": 1.8943, + "theoretical_loss": 4.522147559805034, + "tokens_seen": 156499968 + }, + { + "epoch": 0.05, + "learning_rate": 0.0007697492477432297, + "loss": 1.8523, + "theoretical_loss": 4.5219031784612795, + "tokens_seen": 156565504 + }, + { + "epoch": 0.05, + "learning_rate": 0.0007697331995987965, + "loss": 1.7242, + "theoretical_loss": 4.521658928019305, + "tokens_seen": 156631040 + }, + { + "epoch": 0.05, + "learning_rate": 0.0007697171514543631, + "loss": 1.8649, + "theoretical_loss": 4.521414808354266, + "tokens_seen": 156696576 + }, + { + "epoch": 0.05, + "learning_rate": 0.0007697011033099299, + "loss": 1.7674, + "theoretical_loss": 4.52117081934149, + "tokens_seen": 156762112 + }, + { + "epoch": 0.05, + "learning_rate": 0.0007696850551654965, + "loss": 1.7227, + "theoretical_loss": 4.520926960856476, + "tokens_seen": 156827648 + }, + { + "epoch": 0.05, + "learning_rate": 0.0007696690070210633, + "loss": 1.8755, + "theoretical_loss": 4.520683232774893, + "tokens_seen": 156893184 + }, + { + "epoch": 0.05, + "learning_rate": 0.0007696529588766299, + "loss": 1.8622, + "theoretical_loss": 4.520439634972579, + "tokens_seen": 156958720 + }, + { + "epoch": 0.05, + "learning_rate": 0.0007696369107321967, + "loss": 1.7618, + "theoretical_loss": 4.520196167325546, + "tokens_seen": 157024256 + }, + { + "epoch": 0.05, + "learning_rate": 0.0007696208625877633, + "loss": 1.9216, + "theoretical_loss": 4.519952829709972, + "tokens_seen": 157089792 + }, + { + "epoch": 0.05, + "learning_rate": 0.0007696048144433301, + "loss": 1.7893, + "theoretical_loss": 4.519709622002205, + "tokens_seen": 157155328 + }, + { + "epoch": 0.05, + "learning_rate": 0.0007695887662988967, + "loss": 1.8048, + "theoretical_loss": 4.519466544078762, + "tokens_seen": 157220864 + }, + { + "epoch": 0.05, + "objective/train/docs_used": 64097, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 2.0462114810943604, + "objective/train/theoretical_loss": 4.519223595816332, + "objective/train/tokens_used": 177746400, + "theoretical_loss": 4.519223595816332, + "tokens_seen": 157286400 + }, + { + "epoch": 0.05, + "learning_rate": 0.0007695727181544635, + "loss": 2.0524, + "theoretical_loss": 4.519223595816332, + "tokens_seen": 157286400 + }, + { + "epoch": 0.05, + "learning_rate": 0.0007695566700100301, + "loss": 1.7516, + "theoretical_loss": 4.518980777091768, + "tokens_seen": 157351936 + }, + { + "epoch": 0.05, + "learning_rate": 0.0007695406218655969, + "loss": 1.9327, + "theoretical_loss": 4.518738087782093, + "tokens_seen": 157417472 + }, + { + "epoch": 0.05, + "learning_rate": 0.0007695245737211635, + "loss": 1.8548, + "theoretical_loss": 4.5184955277644985, + "tokens_seen": 157483008 + }, + { + "epoch": 0.05, + "learning_rate": 0.0007695085255767303, + "loss": 1.9197, + "theoretical_loss": 4.518253096916341, + "tokens_seen": 157548544 + }, + { + "epoch": 0.05, + "learning_rate": 0.0007694924774322969, + "loss": 1.8639, + "theoretical_loss": 4.518010795115151, + "tokens_seen": 157614080 + }, + { + "epoch": 0.05, + "learning_rate": 0.0007694764292878637, + "loss": 1.8509, + "theoretical_loss": 4.517768622238616, + "tokens_seen": 157679616 + }, + { + "epoch": 0.05, + "learning_rate": 0.0007694603811434303, + "loss": 1.7962, + "theoretical_loss": 4.5175265781646, + "tokens_seen": 157745152 + }, + { + "epoch": 0.05, + "learning_rate": 0.0007694443329989971, + "loss": 1.7359, + "theoretical_loss": 4.517284662771125, + "tokens_seen": 157810688 + }, + { + "epoch": 0.05, + "learning_rate": 0.0007694282848545637, + "loss": 1.8996, + "theoretical_loss": 4.5170428759363865, + "tokens_seen": 157876224 + }, + { + "epoch": 0.05, + "learning_rate": 0.0007694122367101305, + "loss": 1.8833, + "theoretical_loss": 4.516801217538739, + "tokens_seen": 157941760 + }, + { + "epoch": 0.05, + "learning_rate": 0.0007693961885656971, + "loss": 1.8119, + "theoretical_loss": 4.516559687456708, + "tokens_seen": 158007296 + }, + { + "epoch": 0.05, + "learning_rate": 0.0007693801404212639, + "loss": 1.8618, + "theoretical_loss": 4.516318285568982, + "tokens_seen": 158072832 + }, + { + "epoch": 0.05, + "learning_rate": 0.0007693640922768305, + "loss": 1.8556, + "theoretical_loss": 4.516077011754414, + "tokens_seen": 158138368 + }, + { + "epoch": 0.05, + "learning_rate": 0.0007693480441323973, + "loss": 1.8703, + "theoretical_loss": 4.51583586589202, + "tokens_seen": 158203904 + }, + { + "epoch": 0.05, + "learning_rate": 0.0007693319959879639, + "loss": 1.9553, + "theoretical_loss": 4.515594847860985, + "tokens_seen": 158269440 + }, + { + "epoch": 0.05, + "learning_rate": 0.0007693159478435307, + "loss": 1.8805, + "theoretical_loss": 4.515353957540656, + "tokens_seen": 158334976 + }, + { + "epoch": 0.05, + "learning_rate": 0.0007692998996990973, + "loss": 1.9145, + "theoretical_loss": 4.515113194810541, + "tokens_seen": 158400512 + }, + { + "epoch": 0.05, + "learning_rate": 0.0007692838515546641, + "loss": 1.832, + "theoretical_loss": 4.514872559550316, + "tokens_seen": 158466048 + }, + { + "epoch": 0.05, + "learning_rate": 0.0007692678034102307, + "loss": 1.8138, + "theoretical_loss": 4.514632051639815, + "tokens_seen": 158531584 + }, + { + "epoch": 0.05, + "learning_rate": 0.0007692517552657975, + "loss": 1.8415, + "theoretical_loss": 4.51439167095904, + "tokens_seen": 158597120 + }, + { + "epoch": 0.05, + "learning_rate": 0.0007692357071213641, + "loss": 1.8088, + "theoretical_loss": 4.5141514173881525, + "tokens_seen": 158662656 + }, + { + "epoch": 0.05, + "learning_rate": 0.0007692196589769309, + "loss": 1.8236, + "theoretical_loss": 4.513911290807478, + "tokens_seen": 158728192 + }, + { + "epoch": 0.05, + "learning_rate": 0.0007692036108324975, + "loss": 1.8724, + "theoretical_loss": 4.513671291097502, + "tokens_seen": 158793728 + }, + { + "epoch": 0.05, + "learning_rate": 0.0007691875626880643, + "loss": 1.8905, + "theoretical_loss": 4.513431418138874, + "tokens_seen": 158859264 + }, + { + "epoch": 0.05, + "learning_rate": 0.0007691715145436309, + "loss": 1.8179, + "theoretical_loss": 4.513191671812404, + "tokens_seen": 158924800 + }, + { + "epoch": 0.05, + "learning_rate": 0.0007691554663991977, + "loss": 1.8227, + "theoretical_loss": 4.512952051999061, + "tokens_seen": 158990336 + }, + { + "epoch": 0.05, + "learning_rate": 0.0007691394182547643, + "loss": 1.7191, + "theoretical_loss": 4.512712558579979, + "tokens_seen": 159055872 + }, + { + "epoch": 0.05, + "learning_rate": 0.0007691233701103311, + "loss": 1.8339, + "theoretical_loss": 4.512473191436451, + "tokens_seen": 159121408 + }, + { + "epoch": 0.05, + "learning_rate": 0.0007691073219658977, + "loss": 1.8046, + "theoretical_loss": 4.512233950449927, + "tokens_seen": 159186944 + }, + { + "epoch": 0.05, + "learning_rate": 0.0007690912738214645, + "loss": 1.8907, + "theoretical_loss": 4.511994835502023, + "tokens_seen": 159252480 + }, + { + "epoch": 0.05, + "learning_rate": 0.0007690752256770311, + "loss": 1.8199, + "theoretical_loss": 4.511755846474511, + "tokens_seen": 159318016 + }, + { + "epoch": 0.05, + "learning_rate": 0.0007690591775325979, + "loss": 1.8998, + "theoretical_loss": 4.511516983249322, + "tokens_seen": 159383552 + }, + { + "epoch": 0.05, + "learning_rate": 0.0007690431293881645, + "loss": 1.7836, + "theoretical_loss": 4.511278245708548, + "tokens_seen": 159449088 + }, + { + "epoch": 0.05, + "learning_rate": 0.0007690270812437313, + "loss": 1.8252, + "theoretical_loss": 4.511039633734439, + "tokens_seen": 159514624 + }, + { + "epoch": 0.05, + "learning_rate": 0.0007690110330992979, + "loss": 1.9065, + "theoretical_loss": 4.5108011472094045, + "tokens_seen": 159580160 + }, + { + "epoch": 0.05, + "learning_rate": 0.0007689949849548647, + "loss": 1.8691, + "theoretical_loss": 4.510562786016012, + "tokens_seen": 159645696 + }, + { + "epoch": 0.05, + "learning_rate": 0.0007689789368104313, + "loss": 1.8827, + "theoretical_loss": 4.5103245500369855, + "tokens_seen": 159711232 + }, + { + "epoch": 0.05, + "learning_rate": 0.0007689628886659981, + "loss": 1.9134, + "theoretical_loss": 4.510086439155209, + "tokens_seen": 159776768 + }, + { + "epoch": 0.05, + "learning_rate": 0.0007689468405215647, + "loss": 1.7702, + "theoretical_loss": 4.509848453253722, + "tokens_seen": 159842304 + }, + { + "epoch": 0.05, + "learning_rate": 0.0007689307923771315, + "loss": 1.8592, + "theoretical_loss": 4.509610592215726, + "tokens_seen": 159907840 + }, + { + "epoch": 0.05, + "learning_rate": 0.0007689147442326981, + "loss": 1.8146, + "theoretical_loss": 4.509372855924571, + "tokens_seen": 159973376 + }, + { + "epoch": 0.05, + "learning_rate": 0.0007688986960882648, + "loss": 1.6996, + "theoretical_loss": 4.5091352442637715, + "tokens_seen": 160038912 + }, + { + "epoch": 0.05, + "learning_rate": 0.0007688826479438315, + "loss": 1.8617, + "theoretical_loss": 4.508897757116995, + "tokens_seen": 160104448 + }, + { + "epoch": 0.05, + "learning_rate": 0.0007688665997993982, + "loss": 1.836, + "theoretical_loss": 4.508660394368066, + "tokens_seen": 160169984 + }, + { + "epoch": 0.05, + "learning_rate": 0.0007688505516549649, + "loss": 1.806, + "theoretical_loss": 4.508423155900964, + "tokens_seen": 160235520 + }, + { + "epoch": 0.05, + "learning_rate": 0.0007688345035105316, + "loss": 1.9236, + "theoretical_loss": 4.508186041599823, + "tokens_seen": 160301056 + }, + { + "epoch": 0.05, + "learning_rate": 0.0007688184553660983, + "loss": 1.8388, + "theoretical_loss": 4.5079490513489375, + "tokens_seen": 160366592 + }, + { + "epoch": 0.05, + "learning_rate": 0.000768802407221665, + "loss": 2.0459, + "theoretical_loss": 4.50771218503275, + "tokens_seen": 160432128 + }, + { + "epoch": 0.05, + "learning_rate": 0.0007687863590772316, + "loss": 1.9995, + "theoretical_loss": 4.507475442535863, + "tokens_seen": 160497664 + }, + { + "epoch": 0.05, + "objective/train/docs_used": 65316, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.8889963626861572, + "objective/train/theoretical_loss": 4.507238823743033, + "objective/train/tokens_used": 181023200, + "theoretical_loss": 4.507238823743033, + "tokens_seen": 160563200 + }, + { + "epoch": 0.05, + "learning_rate": 0.0007687703109327984, + "loss": 1.8929, + "theoretical_loss": 4.507238823743033, + "tokens_seen": 160563200 + }, + { + "epoch": 0.05, + "learning_rate": 0.0007687542627883651, + "loss": 1.7672, + "theoretical_loss": 4.507002328539167, + "tokens_seen": 160628736 + }, + { + "epoch": 0.05, + "learning_rate": 0.0007687382146439318, + "loss": 1.7564, + "theoretical_loss": 4.506765956809331, + "tokens_seen": 160694272 + }, + { + "epoch": 0.05, + "learning_rate": 0.0007687221664994985, + "loss": 1.991, + "theoretical_loss": 4.50652970843874, + "tokens_seen": 160759808 + }, + { + "epoch": 0.05, + "learning_rate": 0.0007687061183550652, + "loss": 1.8183, + "theoretical_loss": 4.506293583312766, + "tokens_seen": 160825344 + }, + { + "epoch": 0.05, + "learning_rate": 0.000768690070210632, + "loss": 1.8831, + "theoretical_loss": 4.506057581316933, + "tokens_seen": 160890880 + }, + { + "epoch": 0.05, + "learning_rate": 0.0007686740220661986, + "loss": 1.79, + "theoretical_loss": 4.505821702336917, + "tokens_seen": 160956416 + }, + { + "epoch": 0.05, + "learning_rate": 0.0007686579739217653, + "loss": 1.867, + "theoretical_loss": 4.505585946258547, + "tokens_seen": 161021952 + }, + { + "epoch": 0.05, + "learning_rate": 0.000768641925777332, + "loss": 1.8576, + "theoretical_loss": 4.505350312967805, + "tokens_seen": 161087488 + }, + { + "epoch": 0.05, + "learning_rate": 0.0007686258776328987, + "loss": 1.8123, + "theoretical_loss": 4.5051148023508265, + "tokens_seen": 161153024 + }, + { + "epoch": 0.05, + "learning_rate": 0.0007686098294884654, + "loss": 1.7019, + "theoretical_loss": 4.504879414293896, + "tokens_seen": 161218560 + }, + { + "epoch": 0.05, + "learning_rate": 0.0007685937813440322, + "loss": 1.924, + "theoretical_loss": 4.50464414868345, + "tokens_seen": 161284096 + }, + { + "epoch": 0.05, + "learning_rate": 0.0007685777331995988, + "loss": 1.8987, + "theoretical_loss": 4.504409005406078, + "tokens_seen": 161349632 + }, + { + "epoch": 0.05, + "learning_rate": 0.0007685616850551656, + "loss": 1.8477, + "theoretical_loss": 4.504173984348521, + "tokens_seen": 161415168 + }, + { + "epoch": 0.05, + "learning_rate": 0.0007685456369107322, + "loss": 1.887, + "theoretical_loss": 4.503939085397667, + "tokens_seen": 161480704 + }, + { + "epoch": 0.05, + "learning_rate": 0.000768529588766299, + "loss": 1.8581, + "theoretical_loss": 4.5037043084405575, + "tokens_seen": 161546240 + }, + { + "epoch": 0.05, + "learning_rate": 0.0007685135406218656, + "loss": 1.901, + "theoretical_loss": 4.503469653364386, + "tokens_seen": 161611776 + }, + { + "epoch": 0.05, + "learning_rate": 0.0007684974924774324, + "loss": 1.8387, + "theoretical_loss": 4.503235120056491, + "tokens_seen": 161677312 + }, + { + "epoch": 0.05, + "learning_rate": 0.000768481444332999, + "loss": 1.8941, + "theoretical_loss": 4.503000708404365, + "tokens_seen": 161742848 + }, + { + "epoch": 0.05, + "learning_rate": 0.0007684653961885658, + "loss": 1.8091, + "theoretical_loss": 4.502766418295647, + "tokens_seen": 161808384 + }, + { + "epoch": 0.05, + "learning_rate": 0.0007684493480441324, + "loss": 2.0001, + "theoretical_loss": 4.502532249618129, + "tokens_seen": 161873920 + }, + { + "epoch": 0.05, + "learning_rate": 0.0007684332998996992, + "loss": 1.9527, + "theoretical_loss": 4.502298202259748, + "tokens_seen": 161939456 + }, + { + "epoch": 0.05, + "learning_rate": 0.0007684172517552658, + "loss": 1.7787, + "theoretical_loss": 4.502064276108593, + "tokens_seen": 162004992 + }, + { + "epoch": 0.05, + "learning_rate": 0.0007684012036108326, + "loss": 1.8296, + "theoretical_loss": 4.501830471052898, + "tokens_seen": 162070528 + }, + { + "epoch": 0.05, + "learning_rate": 0.0007683851554663992, + "loss": 1.7952, + "theoretical_loss": 4.501596786981047, + "tokens_seen": 162136064 + }, + { + "epoch": 0.05, + "learning_rate": 0.000768369107321966, + "loss": 1.8697, + "theoretical_loss": 4.501363223781574, + "tokens_seen": 162201600 + }, + { + "epoch": 0.05, + "learning_rate": 0.0007683530591775326, + "loss": 1.9012, + "theoretical_loss": 4.501129781343157, + "tokens_seen": 162267136 + }, + { + "epoch": 0.05, + "learning_rate": 0.0007683370110330994, + "loss": 1.8908, + "theoretical_loss": 4.500896459554624, + "tokens_seen": 162332672 + }, + { + "epoch": 0.05, + "learning_rate": 0.000768320962888666, + "loss": 1.8207, + "theoretical_loss": 4.500663258304948, + "tokens_seen": 162398208 + }, + { + "epoch": 0.05, + "learning_rate": 0.0007683049147442328, + "loss": 1.7733, + "theoretical_loss": 4.500430177483253, + "tokens_seen": 162463744 + }, + { + "epoch": 0.05, + "learning_rate": 0.0007682888665997994, + "loss": 1.8951, + "theoretical_loss": 4.500197216978804, + "tokens_seen": 162529280 + }, + { + "epoch": 0.05, + "learning_rate": 0.0007682728184553662, + "loss": 1.9566, + "theoretical_loss": 4.499964376681016, + "tokens_seen": 162594816 + }, + { + "epoch": 0.05, + "learning_rate": 0.0007682567703109328, + "loss": 1.9094, + "theoretical_loss": 4.49973165647945, + "tokens_seen": 162660352 + }, + { + "epoch": 0.05, + "learning_rate": 0.0007682407221664996, + "loss": 1.9899, + "theoretical_loss": 4.4994990562638115, + "tokens_seen": 162725888 + }, + { + "epoch": 0.05, + "learning_rate": 0.0007682246740220662, + "loss": 1.746, + "theoretical_loss": 4.499266575923954, + "tokens_seen": 162791424 + }, + { + "epoch": 0.05, + "learning_rate": 0.000768208625877633, + "loss": 2.0798, + "theoretical_loss": 4.499034215349875, + "tokens_seen": 162856960 + }, + { + "epoch": 0.05, + "learning_rate": 0.0007681925777331996, + "loss": 1.894, + "theoretical_loss": 4.498801974431714, + "tokens_seen": 162922496 + }, + { + "epoch": 0.05, + "learning_rate": 0.0007681765295887664, + "loss": 1.8378, + "theoretical_loss": 4.498569853059762, + "tokens_seen": 162988032 + }, + { + "epoch": 0.05, + "learning_rate": 0.000768160481444333, + "loss": 1.8408, + "theoretical_loss": 4.49833785112445, + "tokens_seen": 163053568 + }, + { + "epoch": 0.05, + "learning_rate": 0.0007681444332998998, + "loss": 1.895, + "theoretical_loss": 4.498105968516354, + "tokens_seen": 163119104 + }, + { + "epoch": 0.05, + "learning_rate": 0.0007681283851554664, + "loss": 1.8152, + "theoretical_loss": 4.497874205126195, + "tokens_seen": 163184640 + }, + { + "epoch": 0.05, + "learning_rate": 0.0007681123370110332, + "loss": 1.8303, + "theoretical_loss": 4.497642560844838, + "tokens_seen": 163250176 + }, + { + "epoch": 0.05, + "learning_rate": 0.0007680962888665998, + "loss": 1.936, + "theoretical_loss": 4.497411035563291, + "tokens_seen": 163315712 + }, + { + "epoch": 0.05, + "learning_rate": 0.0007680802407221666, + "loss": 1.6992, + "theoretical_loss": 4.497179629172704, + "tokens_seen": 163381248 + }, + { + "epoch": 0.05, + "learning_rate": 0.0007680641925777332, + "loss": 1.9578, + "theoretical_loss": 4.4969483415643765, + "tokens_seen": 163446784 + }, + { + "epoch": 0.05, + "learning_rate": 0.0007680481444333, + "loss": 1.781, + "theoretical_loss": 4.496717172629742, + "tokens_seen": 163512320 + }, + { + "epoch": 0.05, + "learning_rate": 0.0007680320962888666, + "loss": 1.9059, + "theoretical_loss": 4.496486122260382, + "tokens_seen": 163577856 + }, + { + "epoch": 0.05, + "learning_rate": 0.0007680160481444334, + "loss": 1.9714, + "theoretical_loss": 4.4962551903480215, + "tokens_seen": 163643392 + }, + { + "epoch": 0.05, + "learning_rate": 0.000768, + "loss": 1.8861, + "theoretical_loss": 4.496024376784522, + "tokens_seen": 163708928 + }, + { + "epoch": 0.05, + "learning_rate": 0.0007679839518555668, + "loss": 1.8099, + "theoretical_loss": 4.495793681461894, + "tokens_seen": 163774464 + }, + { + "epoch": 0.05, + "objective/train/docs_used": 66569, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.765413522720337, + "objective/train/theoretical_loss": 4.495563104272286, + "objective/train/tokens_used": 184300000, + "theoretical_loss": 4.495563104272286, + "tokens_seen": 163840000 + }, + { + "epoch": 0.05, + "learning_rate": 0.0007679679037111334, + "loss": 1.6621, + "theoretical_loss": 4.495563104272286, + "tokens_seen": 163840000 + }, + { + "epoch": 0.05, + "learning_rate": 0.0007679518555667002, + "loss": 1.7919, + "theoretical_loss": 4.495332645107985, + "tokens_seen": 163905536 + }, + { + "epoch": 0.05, + "learning_rate": 0.0007679358074222668, + "loss": 1.9176, + "theoretical_loss": 4.495102303861426, + "tokens_seen": 163971072 + }, + { + "epoch": 0.05, + "learning_rate": 0.0007679197592778336, + "loss": 1.8644, + "theoretical_loss": 4.4948720804251785, + "tokens_seen": 164036608 + }, + { + "epoch": 0.05, + "learning_rate": 0.0007679037111334002, + "loss": 1.9127, + "theoretical_loss": 4.494641974691959, + "tokens_seen": 164102144 + }, + { + "epoch": 0.05, + "learning_rate": 0.000767887662988967, + "loss": 1.9421, + "theoretical_loss": 4.494411986554619, + "tokens_seen": 164167680 + }, + { + "epoch": 0.05, + "learning_rate": 0.0007678716148445336, + "loss": 1.8889, + "theoretical_loss": 4.494182115906153, + "tokens_seen": 164233216 + }, + { + "epoch": 0.05, + "learning_rate": 0.0007678555667001004, + "loss": 1.8291, + "theoretical_loss": 4.493952362639694, + "tokens_seen": 164298752 + }, + { + "epoch": 0.05, + "learning_rate": 0.000767839518555667, + "loss": 1.8916, + "theoretical_loss": 4.493722726648519, + "tokens_seen": 164364288 + }, + { + "epoch": 0.05, + "learning_rate": 0.0007678234704112338, + "loss": 1.806, + "theoretical_loss": 4.493493207826037, + "tokens_seen": 164429824 + }, + { + "epoch": 0.05, + "learning_rate": 0.0007678074222668004, + "loss": 1.8416, + "theoretical_loss": 4.493263806065803, + "tokens_seen": 164495360 + }, + { + "epoch": 0.05, + "learning_rate": 0.0007677913741223672, + "loss": 1.8504, + "theoretical_loss": 4.493034521261509, + "tokens_seen": 164560896 + }, + { + "epoch": 0.05, + "learning_rate": 0.0007677753259779338, + "loss": 1.8243, + "theoretical_loss": 4.492805353306984, + "tokens_seen": 164626432 + }, + { + "epoch": 0.05, + "learning_rate": 0.0007677592778335006, + "loss": 1.823, + "theoretical_loss": 4.492576302096199, + "tokens_seen": 164691968 + }, + { + "epoch": 0.05, + "learning_rate": 0.0007677432296890672, + "loss": 1.7176, + "theoretical_loss": 4.49234736752326, + "tokens_seen": 164757504 + }, + { + "epoch": 0.05, + "learning_rate": 0.000767727181544634, + "loss": 1.952, + "theoretical_loss": 4.492118549482413, + "tokens_seen": 164823040 + }, + { + "epoch": 0.05, + "learning_rate": 0.0007677111334002006, + "loss": 1.8066, + "theoretical_loss": 4.491889847868041, + "tokens_seen": 164888576 + }, + { + "epoch": 0.05, + "learning_rate": 0.0007676950852557674, + "loss": 1.6985, + "theoretical_loss": 4.491661262574665, + "tokens_seen": 164954112 + }, + { + "epoch": 0.05, + "learning_rate": 0.000767679037111334, + "loss": 1.8181, + "theoretical_loss": 4.491432793496945, + "tokens_seen": 165019648 + }, + { + "epoch": 0.05, + "learning_rate": 0.0007676629889669008, + "loss": 1.7364, + "theoretical_loss": 4.4912044405296765, + "tokens_seen": 165085184 + }, + { + "epoch": 0.05, + "learning_rate": 0.0007676469408224674, + "loss": 1.8206, + "theoretical_loss": 4.49097620356779, + "tokens_seen": 165150720 + }, + { + "epoch": 0.05, + "learning_rate": 0.0007676308926780342, + "loss": 1.7919, + "theoretical_loss": 4.490748082506356, + "tokens_seen": 165216256 + }, + { + "epoch": 0.05, + "learning_rate": 0.0007676148445336008, + "loss": 1.7843, + "theoretical_loss": 4.490520077240582, + "tokens_seen": 165281792 + }, + { + "epoch": 0.05, + "learning_rate": 0.0007675987963891676, + "loss": 1.7449, + "theoretical_loss": 4.490292187665808, + "tokens_seen": 165347328 + }, + { + "epoch": 0.05, + "learning_rate": 0.0007675827482447342, + "loss": 1.8429, + "theoretical_loss": 4.490064413677514, + "tokens_seen": 165412864 + }, + { + "epoch": 0.05, + "learning_rate": 0.000767566700100301, + "loss": 1.7919, + "theoretical_loss": 4.4898367551713125, + "tokens_seen": 165478400 + }, + { + "epoch": 0.05, + "learning_rate": 0.0007675506519558676, + "loss": 1.6892, + "theoretical_loss": 4.489609212042954, + "tokens_seen": 165543936 + }, + { + "epoch": 0.05, + "learning_rate": 0.0007675346038114344, + "loss": 1.8806, + "theoretical_loss": 4.489381784188322, + "tokens_seen": 165609472 + }, + { + "epoch": 0.05, + "learning_rate": 0.000767518555667001, + "loss": 1.8386, + "theoretical_loss": 4.489154471503438, + "tokens_seen": 165675008 + }, + { + "epoch": 0.05, + "learning_rate": 0.0007675025075225677, + "loss": 1.6779, + "theoretical_loss": 4.488927273884456, + "tokens_seen": 165740544 + }, + { + "epoch": 0.05, + "learning_rate": 0.0007674864593781344, + "loss": 1.6127, + "theoretical_loss": 4.488700191227665, + "tokens_seen": 165806080 + }, + { + "epoch": 0.05, + "learning_rate": 0.0007674704112337011, + "loss": 1.8683, + "theoretical_loss": 4.488473223429491, + "tokens_seen": 165871616 + }, + { + "epoch": 0.05, + "learning_rate": 0.0007674543630892678, + "loss": 1.7969, + "theoretical_loss": 4.48824637038649, + "tokens_seen": 165937152 + }, + { + "epoch": 0.05, + "learning_rate": 0.0007674383149448345, + "loss": 1.7897, + "theoretical_loss": 4.488019631995354, + "tokens_seen": 166002688 + }, + { + "epoch": 0.05, + "learning_rate": 0.0007674222668004012, + "loss": 1.7916, + "theoretical_loss": 4.48779300815291, + "tokens_seen": 166068224 + }, + { + "epoch": 0.05, + "learning_rate": 0.0007674062186559679, + "loss": 1.7308, + "theoretical_loss": 4.4875664987561175, + "tokens_seen": 166133760 + }, + { + "epoch": 0.05, + "learning_rate": 0.0007673901705115346, + "loss": 1.7794, + "theoretical_loss": 4.487340103702068, + "tokens_seen": 166199296 + }, + { + "epoch": 0.05, + "learning_rate": 0.0007673741223671013, + "loss": 1.8808, + "theoretical_loss": 4.487113822887988, + "tokens_seen": 166264832 + }, + { + "epoch": 0.05, + "learning_rate": 0.000767358074222668, + "loss": 1.8526, + "theoretical_loss": 4.486887656211234, + "tokens_seen": 166330368 + }, + { + "epoch": 0.05, + "learning_rate": 0.0007673420260782347, + "loss": 1.7225, + "theoretical_loss": 4.486661603569301, + "tokens_seen": 166395904 + }, + { + "epoch": 0.05, + "learning_rate": 0.0007673259779338014, + "loss": 1.7826, + "theoretical_loss": 4.486435664859809, + "tokens_seen": 166461440 + }, + { + "epoch": 0.05, + "learning_rate": 0.0007673099297893681, + "loss": 1.888, + "theoretical_loss": 4.486209839980516, + "tokens_seen": 166526976 + }, + { + "epoch": 0.05, + "learning_rate": 0.0007672938816449348, + "loss": 1.825, + "theoretical_loss": 4.485984128829307, + "tokens_seen": 166592512 + }, + { + "epoch": 0.05, + "learning_rate": 0.0007672778335005015, + "loss": 1.7918, + "theoretical_loss": 4.485758531304205, + "tokens_seen": 166658048 + }, + { + "epoch": 0.05, + "learning_rate": 0.0007672617853560682, + "loss": 1.8629, + "theoretical_loss": 4.485533047303358, + "tokens_seen": 166723584 + }, + { + "epoch": 0.05, + "learning_rate": 0.0007672457372116349, + "loss": 1.7826, + "theoretical_loss": 4.485307676725048, + "tokens_seen": 166789120 + }, + { + "epoch": 0.05, + "learning_rate": 0.0007672296890672016, + "loss": 1.7353, + "theoretical_loss": 4.485082419467691, + "tokens_seen": 166854656 + }, + { + "epoch": 0.05, + "learning_rate": 0.0007672136409227683, + "loss": 1.7604, + "theoretical_loss": 4.484857275429828, + "tokens_seen": 166920192 + }, + { + "epoch": 0.05, + "learning_rate": 0.000767197592778335, + "loss": 1.668, + "theoretical_loss": 4.484632244510136, + "tokens_seen": 166985728 + }, + { + "epoch": 0.05, + "learning_rate": 0.0007671815446339017, + "loss": 1.8532, + "theoretical_loss": 4.4844073266074185, + "tokens_seen": 167051264 + }, + { + "epoch": 0.05, + "objective/train/docs_used": 67770, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.9288755655288696, + "objective/train/theoretical_loss": 4.484182521620611, + "objective/train/tokens_used": 187576800, + "theoretical_loss": 4.484182521620611, + "tokens_seen": 167116800 + }, + { + "epoch": 0.05, + "learning_rate": 0.0007671654964894685, + "loss": 1.8776, + "theoretical_loss": 4.484182521620611, + "tokens_seen": 167116800 + }, + { + "epoch": 0.05, + "learning_rate": 0.0007671494483450351, + "loss": 1.8424, + "theoretical_loss": 4.48395782944878, + "tokens_seen": 167182336 + }, + { + "epoch": 0.05, + "learning_rate": 0.0007671334002006019, + "loss": 1.9269, + "theoretical_loss": 4.483733249991118, + "tokens_seen": 167247872 + }, + { + "epoch": 0.05, + "learning_rate": 0.0007671173520561685, + "loss": 1.7486, + "theoretical_loss": 4.483508783146952, + "tokens_seen": 167313408 + }, + { + "epoch": 0.05, + "learning_rate": 0.0007671013039117353, + "loss": 1.7675, + "theoretical_loss": 4.483284428815735, + "tokens_seen": 167378944 + }, + { + "epoch": 0.05, + "learning_rate": 0.0007670852557673019, + "loss": 1.9039, + "theoretical_loss": 4.483060186897049, + "tokens_seen": 167444480 + }, + { + "epoch": 0.05, + "learning_rate": 0.0007670692076228687, + "loss": 1.8137, + "theoretical_loss": 4.482836057290607, + "tokens_seen": 167510016 + }, + { + "epoch": 0.05, + "learning_rate": 0.0007670531594784353, + "loss": 1.8443, + "theoretical_loss": 4.4826120398962495, + "tokens_seen": 167575552 + }, + { + "epoch": 0.05, + "learning_rate": 0.0007670371113340021, + "loss": 1.8892, + "theoretical_loss": 4.482388134613943, + "tokens_seen": 167641088 + }, + { + "epoch": 0.05, + "learning_rate": 0.0007670210631895687, + "loss": 1.7741, + "theoretical_loss": 4.482164341343788, + "tokens_seen": 167706624 + }, + { + "epoch": 0.05, + "learning_rate": 0.0007670050150451355, + "loss": 1.6988, + "theoretical_loss": 4.481940659986008, + "tokens_seen": 167772160 + }, + { + "epoch": 0.05, + "learning_rate": 0.0007669889669007021, + "loss": 1.9236, + "theoretical_loss": 4.481717090440956, + "tokens_seen": 167837696 + }, + { + "epoch": 0.05, + "learning_rate": 0.0007669729187562689, + "loss": 1.812, + "theoretical_loss": 4.481493632609112, + "tokens_seen": 167903232 + }, + { + "epoch": 0.05, + "learning_rate": 0.0007669568706118355, + "loss": 1.6025, + "theoretical_loss": 4.481270286391084, + "tokens_seen": 167968768 + }, + { + "epoch": 0.05, + "learning_rate": 0.0007669408224674023, + "loss": 1.7988, + "theoretical_loss": 4.481047051687609, + "tokens_seen": 168034304 + }, + { + "epoch": 0.05, + "learning_rate": 0.0007669247743229689, + "loss": 1.7987, + "theoretical_loss": 4.480823928399547, + "tokens_seen": 168099840 + }, + { + "epoch": 0.05, + "learning_rate": 0.0007669087261785357, + "loss": 1.7825, + "theoretical_loss": 4.480600916427887, + "tokens_seen": 168165376 + }, + { + "epoch": 0.05, + "learning_rate": 0.0007668926780341023, + "loss": 1.6951, + "theoretical_loss": 4.480378015673745, + "tokens_seen": 168230912 + }, + { + "epoch": 0.05, + "learning_rate": 0.0007668766298896691, + "loss": 1.9209, + "theoretical_loss": 4.480155226038362, + "tokens_seen": 168296448 + }, + { + "epoch": 0.05, + "learning_rate": 0.0007668605817452357, + "loss": 1.8298, + "theoretical_loss": 4.479932547423107, + "tokens_seen": 168361984 + }, + { + "epoch": 0.05, + "learning_rate": 0.0007668445336008025, + "loss": 1.8777, + "theoretical_loss": 4.479709979729472, + "tokens_seen": 168427520 + }, + { + "epoch": 0.05, + "learning_rate": 0.0007668284854563691, + "loss": 1.7754, + "theoretical_loss": 4.479487522859078, + "tokens_seen": 168493056 + }, + { + "epoch": 0.05, + "learning_rate": 0.0007668124373119359, + "loss": 1.8987, + "theoretical_loss": 4.4792651767136675, + "tokens_seen": 168558592 + }, + { + "epoch": 0.05, + "learning_rate": 0.0007667963891675025, + "loss": 1.9154, + "theoretical_loss": 4.479042941195113, + "tokens_seen": 168624128 + }, + { + "epoch": 0.05, + "learning_rate": 0.0007667803410230693, + "loss": 1.7211, + "theoretical_loss": 4.478820816205408, + "tokens_seen": 168689664 + }, + { + "epoch": 0.05, + "learning_rate": 0.0007667642928786359, + "loss": 1.8403, + "theoretical_loss": 4.4785988016466725, + "tokens_seen": 168755200 + }, + { + "epoch": 0.05, + "learning_rate": 0.0007667482447342027, + "loss": 1.8454, + "theoretical_loss": 4.4783768974211515, + "tokens_seen": 168820736 + }, + { + "epoch": 0.05, + "learning_rate": 0.0007667321965897693, + "loss": 1.6564, + "theoretical_loss": 4.478155103431215, + "tokens_seen": 168886272 + }, + { + "epoch": 0.05, + "learning_rate": 0.0007667161484453361, + "loss": 1.8781, + "theoretical_loss": 4.477933419579356, + "tokens_seen": 168951808 + }, + { + "epoch": 0.05, + "learning_rate": 0.0007667001003009027, + "loss": 1.6489, + "theoretical_loss": 4.477711845768191, + "tokens_seen": 169017344 + }, + { + "epoch": 0.05, + "learning_rate": 0.0007666840521564695, + "loss": 1.7112, + "theoretical_loss": 4.477490381900461, + "tokens_seen": 169082880 + }, + { + "epoch": 0.05, + "learning_rate": 0.0007666680040120361, + "loss": 1.9848, + "theoretical_loss": 4.477269027879032, + "tokens_seen": 169148416 + }, + { + "epoch": 0.05, + "learning_rate": 0.0007666519558676029, + "loss": 1.7286, + "theoretical_loss": 4.477047783606892, + "tokens_seen": 169213952 + }, + { + "epoch": 0.05, + "learning_rate": 0.0007666359077231695, + "loss": 1.867, + "theoretical_loss": 4.476826648987151, + "tokens_seen": 169279488 + }, + { + "epoch": 0.05, + "learning_rate": 0.0007666198595787363, + "loss": 1.8574, + "theoretical_loss": 4.476605623923044, + "tokens_seen": 169345024 + }, + { + "epoch": 0.05, + "learning_rate": 0.0007666038114343029, + "loss": 1.8197, + "theoretical_loss": 4.47638470831793, + "tokens_seen": 169410560 + }, + { + "epoch": 0.05, + "learning_rate": 0.0007665877632898697, + "loss": 1.8552, + "theoretical_loss": 4.476163902075286, + "tokens_seen": 169476096 + }, + { + "epoch": 0.05, + "learning_rate": 0.0007665717151454363, + "loss": 1.8546, + "theoretical_loss": 4.475943205098716, + "tokens_seen": 169541632 + }, + { + "epoch": 0.05, + "learning_rate": 0.0007665556670010031, + "loss": 1.9419, + "theoretical_loss": 4.475722617291944, + "tokens_seen": 169607168 + }, + { + "epoch": 0.05, + "learning_rate": 0.0007665396188565697, + "loss": 1.8316, + "theoretical_loss": 4.4755021385588165, + "tokens_seen": 169672704 + }, + { + "epoch": 0.05, + "learning_rate": 0.0007665235707121365, + "loss": 1.7791, + "theoretical_loss": 4.475281768803301, + "tokens_seen": 169738240 + }, + { + "epoch": 0.05, + "learning_rate": 0.0007665075225677031, + "loss": 1.8864, + "theoretical_loss": 4.475061507929487, + "tokens_seen": 169803776 + }, + { + "epoch": 0.05, + "learning_rate": 0.0007664914744232699, + "loss": 1.6121, + "theoretical_loss": 4.474841355841587, + "tokens_seen": 169869312 + }, + { + "epoch": 0.05, + "learning_rate": 0.0007664754262788365, + "loss": 1.9299, + "theoretical_loss": 4.474621312443932, + "tokens_seen": 169934848 + }, + { + "epoch": 0.05, + "learning_rate": 0.0007664593781344033, + "loss": 1.8064, + "theoretical_loss": 4.474401377640977, + "tokens_seen": 170000384 + }, + { + "epoch": 0.05, + "learning_rate": 0.0007664433299899699, + "loss": 1.9013, + "theoretical_loss": 4.474181551337294, + "tokens_seen": 170065920 + }, + { + "epoch": 0.05, + "learning_rate": 0.0007664272818455367, + "loss": 1.7079, + "theoretical_loss": 4.473961833437579, + "tokens_seen": 170131456 + }, + { + "epoch": 0.05, + "learning_rate": 0.0007664112337011033, + "loss": 1.8312, + "theoretical_loss": 4.473742223846646, + "tokens_seen": 170196992 + }, + { + "epoch": 0.05, + "learning_rate": 0.0007663951855566701, + "loss": 1.8998, + "theoretical_loss": 4.473522722469431, + "tokens_seen": 170262528 + }, + { + "epoch": 0.05, + "learning_rate": 0.0007663791374122367, + "loss": 1.9759, + "theoretical_loss": 4.47330332921099, + "tokens_seen": 170328064 + }, + { + "epoch": 0.05, + "objective/train/docs_used": 69022, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.9944404363632202, + "objective/train/theoretical_loss": 4.473084043976495, + "objective/train/tokens_used": 190853600, + "theoretical_loss": 4.473084043976495, + "tokens_seen": 170393600 + }, + { + "epoch": 0.05, + "learning_rate": 0.0007663630892678035, + "loss": 1.8972, + "theoretical_loss": 4.473084043976495, + "tokens_seen": 170393600 + }, + { + "epoch": 0.05, + "learning_rate": 0.0007663470411233702, + "loss": 1.9782, + "theoretical_loss": 4.472864866671245, + "tokens_seen": 170459136 + }, + { + "epoch": 0.05, + "learning_rate": 0.0007663309929789369, + "loss": 1.9052, + "theoretical_loss": 4.472645797200651, + "tokens_seen": 170524672 + }, + { + "epoch": 0.05, + "learning_rate": 0.0007663149448345036, + "loss": 1.7257, + "theoretical_loss": 4.472426835470247, + "tokens_seen": 170590208 + }, + { + "epoch": 0.05, + "learning_rate": 0.0007662988966900703, + "loss": 1.76, + "theoretical_loss": 4.472207981385686, + "tokens_seen": 170655744 + }, + { + "epoch": 0.05, + "learning_rate": 0.000766282848545637, + "loss": 1.8946, + "theoretical_loss": 4.471989234852737, + "tokens_seen": 170721280 + }, + { + "epoch": 0.05, + "learning_rate": 0.0007662668004012037, + "loss": 1.8085, + "theoretical_loss": 4.471770595777292, + "tokens_seen": 170786816 + }, + { + "epoch": 0.05, + "learning_rate": 0.0007662507522567703, + "loss": 1.7893, + "theoretical_loss": 4.4715520640653565, + "tokens_seen": 170852352 + }, + { + "epoch": 0.05, + "learning_rate": 0.0007662347041123371, + "loss": 1.808, + "theoretical_loss": 4.471333639623059, + "tokens_seen": 170917888 + }, + { + "epoch": 0.05, + "learning_rate": 0.0007662186559679037, + "loss": 1.8445, + "theoretical_loss": 4.471115322356643, + "tokens_seen": 170983424 + }, + { + "epoch": 0.05, + "learning_rate": 0.0007662026078234705, + "loss": 1.6524, + "theoretical_loss": 4.470897112172471, + "tokens_seen": 171048960 + }, + { + "epoch": 0.05, + "learning_rate": 0.0007661865596790371, + "loss": 2.0136, + "theoretical_loss": 4.470679008977022, + "tokens_seen": 171114496 + }, + { + "epoch": 0.05, + "learning_rate": 0.0007661705115346039, + "loss": 1.7855, + "theoretical_loss": 4.470461012676894, + "tokens_seen": 171180032 + }, + { + "epoch": 0.05, + "learning_rate": 0.0007661544633901705, + "loss": 1.7765, + "theoretical_loss": 4.470243123178802, + "tokens_seen": 171245568 + }, + { + "epoch": 0.05, + "learning_rate": 0.0007661384152457373, + "loss": 1.83, + "theoretical_loss": 4.4700253403895776, + "tokens_seen": 171311104 + }, + { + "epoch": 0.05, + "learning_rate": 0.0007661223671013039, + "loss": 1.8911, + "theoretical_loss": 4.469807664216168, + "tokens_seen": 171376640 + }, + { + "epoch": 0.05, + "learning_rate": 0.0007661063189568706, + "loss": 1.7666, + "theoretical_loss": 4.469590094565638, + "tokens_seen": 171442176 + }, + { + "epoch": 0.05, + "learning_rate": 0.0007660902708124373, + "loss": 1.8178, + "theoretical_loss": 4.469372631345172, + "tokens_seen": 171507712 + }, + { + "epoch": 0.05, + "learning_rate": 0.000766074222668004, + "loss": 1.7119, + "theoretical_loss": 4.469155274462064, + "tokens_seen": 171573248 + }, + { + "epoch": 0.05, + "learning_rate": 0.0007660581745235707, + "loss": 1.8046, + "theoretical_loss": 4.468938023823732, + "tokens_seen": 171638784 + }, + { + "epoch": 0.05, + "learning_rate": 0.0007660421263791374, + "loss": 1.7993, + "theoretical_loss": 4.468720879337704, + "tokens_seen": 171704320 + }, + { + "epoch": 0.05, + "learning_rate": 0.0007660260782347041, + "loss": 1.7315, + "theoretical_loss": 4.468503840911627, + "tokens_seen": 171769856 + }, + { + "epoch": 0.05, + "learning_rate": 0.0007660100300902708, + "loss": 1.8201, + "theoretical_loss": 4.46828690845326, + "tokens_seen": 171835392 + }, + { + "epoch": 0.05, + "learning_rate": 0.0007659939819458375, + "loss": 1.8, + "theoretical_loss": 4.468070081870483, + "tokens_seen": 171900928 + }, + { + "epoch": 0.05, + "learning_rate": 0.0007659779338014042, + "loss": 1.7194, + "theoretical_loss": 4.467853361071284, + "tokens_seen": 171966464 + }, + { + "epoch": 0.05, + "learning_rate": 0.000765961885656971, + "loss": 1.7586, + "theoretical_loss": 4.467636745963772, + "tokens_seen": 172032000 + }, + { + "epoch": 0.05, + "learning_rate": 0.0007659458375125376, + "loss": 1.9236, + "theoretical_loss": 4.467420236456169, + "tokens_seen": 172097536 + }, + { + "epoch": 0.05, + "learning_rate": 0.0007659297893681043, + "loss": 1.8781, + "theoretical_loss": 4.46720383245681, + "tokens_seen": 172163072 + }, + { + "epoch": 0.05, + "learning_rate": 0.000765913741223671, + "loss": 1.757, + "theoretical_loss": 4.466987533874145, + "tokens_seen": 172228608 + }, + { + "epoch": 0.05, + "learning_rate": 0.0007658976930792377, + "loss": 1.8326, + "theoretical_loss": 4.4667713406167415, + "tokens_seen": 172294144 + }, + { + "epoch": 0.05, + "learning_rate": 0.0007658816449348044, + "loss": 1.7727, + "theoretical_loss": 4.466555252593277, + "tokens_seen": 172359680 + }, + { + "epoch": 0.05, + "learning_rate": 0.0007658655967903711, + "loss": 1.8779, + "theoretical_loss": 4.4663392697125435, + "tokens_seen": 172425216 + }, + { + "epoch": 0.05, + "learning_rate": 0.0007658495486459378, + "loss": 1.7429, + "theoretical_loss": 4.466123391883448, + "tokens_seen": 172490752 + }, + { + "epoch": 0.05, + "learning_rate": 0.0007658335005015046, + "loss": 1.9257, + "theoretical_loss": 4.4659076190150095, + "tokens_seen": 172556288 + }, + { + "epoch": 0.05, + "learning_rate": 0.0007658174523570712, + "loss": 1.7526, + "theoretical_loss": 4.465691951016364, + "tokens_seen": 172621824 + }, + { + "epoch": 0.05, + "learning_rate": 0.000765801404212638, + "loss": 1.9496, + "theoretical_loss": 4.465476387796755, + "tokens_seen": 172687360 + }, + { + "epoch": 0.05, + "learning_rate": 0.0007657853560682046, + "loss": 1.7994, + "theoretical_loss": 4.465260929265543, + "tokens_seen": 172752896 + }, + { + "epoch": 0.05, + "learning_rate": 0.0007657693079237714, + "loss": 1.8289, + "theoretical_loss": 4.4650455753322005, + "tokens_seen": 172818432 + }, + { + "epoch": 0.05, + "learning_rate": 0.000765753259779338, + "loss": 1.7973, + "theoretical_loss": 4.464830325906312, + "tokens_seen": 172883968 + }, + { + "epoch": 0.05, + "learning_rate": 0.0007657372116349048, + "loss": 1.7857, + "theoretical_loss": 4.464615180897574, + "tokens_seen": 172949504 + }, + { + "epoch": 0.05, + "learning_rate": 0.0007657211634904714, + "loss": 1.7464, + "theoretical_loss": 4.464400140215796, + "tokens_seen": 173015040 + }, + { + "epoch": 0.05, + "learning_rate": 0.0007657051153460382, + "loss": 1.8364, + "theoretical_loss": 4.4641852037708984, + "tokens_seen": 173080576 + }, + { + "epoch": 0.05, + "learning_rate": 0.0007656890672016048, + "loss": 1.8291, + "theoretical_loss": 4.463970371472916, + "tokens_seen": 173146112 + }, + { + "epoch": 0.05, + "learning_rate": 0.0007656730190571716, + "loss": 1.9802, + "theoretical_loss": 4.463755643231992, + "tokens_seen": 173211648 + }, + { + "epoch": 0.05, + "learning_rate": 0.0007656569709127382, + "loss": 1.7497, + "theoretical_loss": 4.463541018958384, + "tokens_seen": 173277184 + }, + { + "epoch": 0.05, + "learning_rate": 0.000765640922768305, + "loss": 1.9317, + "theoretical_loss": 4.463326498562459, + "tokens_seen": 173342720 + }, + { + "epoch": 0.05, + "learning_rate": 0.0007656248746238716, + "loss": 1.729, + "theoretical_loss": 4.463112081954696, + "tokens_seen": 173408256 + }, + { + "epoch": 0.05, + "learning_rate": 0.0007656088264794384, + "loss": 1.9382, + "theoretical_loss": 4.462897769045684, + "tokens_seen": 173473792 + }, + { + "epoch": 0.05, + "learning_rate": 0.000765592778335005, + "loss": 1.8451, + "theoretical_loss": 4.462683559746123, + "tokens_seen": 173539328 + }, + { + "epoch": 0.05, + "learning_rate": 0.0007655767301905718, + "loss": 1.8001, + "theoretical_loss": 4.4624694539668255, + "tokens_seen": 173604864 + }, + { + "epoch": 0.05, + "objective/train/docs_used": 70257, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.9520747661590576, + "objective/train/theoretical_loss": 4.462255451618713, + "objective/train/tokens_used": 194130400, + "theoretical_loss": 4.462255451618713, + "tokens_seen": 173670400 + }, + { + "epoch": 0.05, + "learning_rate": 0.0007655606820461384, + "loss": 1.9094, + "theoretical_loss": 4.462255451618713, + "tokens_seen": 173670400 + }, + { + "epoch": 0.05, + "learning_rate": 0.0007655446339017052, + "loss": 1.8846, + "theoretical_loss": 4.462041552612815, + "tokens_seen": 173735936 + }, + { + "epoch": 0.05, + "learning_rate": 0.0007655285857572718, + "loss": 1.8477, + "theoretical_loss": 4.461827756860275, + "tokens_seen": 173801472 + }, + { + "epoch": 0.05, + "learning_rate": 0.0007655125376128386, + "loss": 1.7282, + "theoretical_loss": 4.461614064272345, + "tokens_seen": 173867008 + }, + { + "epoch": 0.05, + "learning_rate": 0.0007654964894684052, + "loss": 1.7527, + "theoretical_loss": 4.461400474760383, + "tokens_seen": 173932544 + }, + { + "epoch": 0.05, + "learning_rate": 0.000765480441323972, + "loss": 1.814, + "theoretical_loss": 4.461186988235864, + "tokens_seen": 173998080 + }, + { + "epoch": 0.05, + "learning_rate": 0.0007654643931795386, + "loss": 1.6243, + "theoretical_loss": 4.460973604610366, + "tokens_seen": 174063616 + }, + { + "epoch": 0.05, + "learning_rate": 0.0007654483450351054, + "loss": 1.7709, + "theoretical_loss": 4.4607603237955775, + "tokens_seen": 174129152 + }, + { + "epoch": 0.05, + "learning_rate": 0.000765432296890672, + "loss": 1.7551, + "theoretical_loss": 4.460547145703298, + "tokens_seen": 174194688 + }, + { + "epoch": 0.05, + "learning_rate": 0.0007654162487462388, + "loss": 1.8164, + "theoretical_loss": 4.460334070245436, + "tokens_seen": 174260224 + }, + { + "epoch": 0.05, + "learning_rate": 0.0007654002006018054, + "loss": 1.8552, + "theoretical_loss": 4.460121097334004, + "tokens_seen": 174325760 + }, + { + "epoch": 0.05, + "learning_rate": 0.0007653841524573722, + "loss": 1.896, + "theoretical_loss": 4.4599082268811285, + "tokens_seen": 174391296 + }, + { + "epoch": 0.05, + "learning_rate": 0.0007653681043129388, + "loss": 1.687, + "theoretical_loss": 4.459695458799041, + "tokens_seen": 174456832 + }, + { + "epoch": 0.05, + "learning_rate": 0.0007653520561685056, + "loss": 1.6572, + "theoretical_loss": 4.459482793000083, + "tokens_seen": 174522368 + }, + { + "epoch": 0.05, + "learning_rate": 0.0007653360080240722, + "loss": 1.7566, + "theoretical_loss": 4.459270229396701, + "tokens_seen": 174587904 + }, + { + "epoch": 0.05, + "learning_rate": 0.000765319959879639, + "loss": 1.8112, + "theoretical_loss": 4.459057767901454, + "tokens_seen": 174653440 + }, + { + "epoch": 0.05, + "learning_rate": 0.0007653039117352056, + "loss": 1.6854, + "theoretical_loss": 4.458845408427004, + "tokens_seen": 174718976 + }, + { + "epoch": 0.05, + "learning_rate": 0.0007652878635907724, + "loss": 1.7561, + "theoretical_loss": 4.458633150886124, + "tokens_seen": 174784512 + }, + { + "epoch": 0.05, + "learning_rate": 0.000765271815446339, + "loss": 1.8463, + "theoretical_loss": 4.458420995191691, + "tokens_seen": 174850048 + }, + { + "epoch": 0.05, + "learning_rate": 0.0007652557673019058, + "loss": 1.9804, + "theoretical_loss": 4.458208941256691, + "tokens_seen": 174915584 + }, + { + "epoch": 0.05, + "learning_rate": 0.0007652397191574724, + "loss": 1.9697, + "theoretical_loss": 4.457996988994218, + "tokens_seen": 174981120 + }, + { + "epoch": 0.05, + "learning_rate": 0.0007652236710130392, + "loss": 1.8001, + "theoretical_loss": 4.457785138317469, + "tokens_seen": 175046656 + }, + { + "epoch": 0.05, + "learning_rate": 0.0007652076228686058, + "loss": 1.8057, + "theoretical_loss": 4.4575733891397515, + "tokens_seen": 175112192 + }, + { + "epoch": 0.05, + "learning_rate": 0.0007651915747241726, + "loss": 1.7149, + "theoretical_loss": 4.457361741374477, + "tokens_seen": 175177728 + }, + { + "epoch": 0.05, + "learning_rate": 0.0007651755265797392, + "loss": 1.8113, + "theoretical_loss": 4.457150194935165, + "tokens_seen": 175243264 + }, + { + "epoch": 0.05, + "learning_rate": 0.000765159478435306, + "loss": 1.8515, + "theoretical_loss": 4.456938749735439, + "tokens_seen": 175308800 + }, + { + "epoch": 0.05, + "learning_rate": 0.0007651434302908727, + "loss": 1.8928, + "theoretical_loss": 4.45672740568903, + "tokens_seen": 175374336 + }, + { + "epoch": 0.05, + "learning_rate": 0.0007651273821464394, + "loss": 1.787, + "theoretical_loss": 4.456516162709774, + "tokens_seen": 175439872 + }, + { + "epoch": 0.05, + "learning_rate": 0.0007651113340020061, + "loss": 1.831, + "theoretical_loss": 4.456305020711613, + "tokens_seen": 175505408 + }, + { + "epoch": 0.05, + "learning_rate": 0.0007650952858575728, + "loss": 1.7875, + "theoretical_loss": 4.456093979608594, + "tokens_seen": 175570944 + }, + { + "epoch": 0.05, + "learning_rate": 0.0007650792377131395, + "loss": 1.7858, + "theoretical_loss": 4.455883039314869, + "tokens_seen": 175636480 + }, + { + "epoch": 0.05, + "learning_rate": 0.0007650631895687062, + "loss": 1.8831, + "theoretical_loss": 4.455672199744695, + "tokens_seen": 175702016 + }, + { + "epoch": 0.05, + "learning_rate": 0.0007650471414242729, + "loss": 1.7559, + "theoretical_loss": 4.455461460812435, + "tokens_seen": 175767552 + }, + { + "epoch": 0.05, + "learning_rate": 0.0007650310932798396, + "loss": 1.8917, + "theoretical_loss": 4.455250822432556, + "tokens_seen": 175833088 + }, + { + "epoch": 0.05, + "learning_rate": 0.0007650150451354063, + "loss": 1.9052, + "theoretical_loss": 4.4550402845196295, + "tokens_seen": 175898624 + }, + { + "epoch": 0.05, + "learning_rate": 0.000764998996990973, + "loss": 1.8579, + "theoretical_loss": 4.45482984698833, + "tokens_seen": 175964160 + }, + { + "epoch": 0.05, + "learning_rate": 0.0007649829488465397, + "loss": 1.7819, + "theoretical_loss": 4.454619509753439, + "tokens_seen": 176029696 + }, + { + "epoch": 0.05, + "learning_rate": 0.0007649669007021064, + "loss": 1.7497, + "theoretical_loss": 4.45440927272984, + "tokens_seen": 176095232 + }, + { + "epoch": 0.05, + "learning_rate": 0.0007649508525576731, + "loss": 1.8104, + "theoretical_loss": 4.454199135832521, + "tokens_seen": 176160768 + }, + { + "epoch": 0.05, + "learning_rate": 0.0007649348044132398, + "loss": 1.8691, + "theoretical_loss": 4.4539890989765745, + "tokens_seen": 176226304 + }, + { + "epoch": 0.05, + "learning_rate": 0.0007649187562688065, + "loss": 1.911, + "theoretical_loss": 4.453779162077193, + "tokens_seen": 176291840 + }, + { + "epoch": 0.05, + "learning_rate": 0.0007649027081243732, + "loss": 1.7735, + "theoretical_loss": 4.453569325049679, + "tokens_seen": 176357376 + }, + { + "epoch": 0.05, + "learning_rate": 0.0007648866599799399, + "loss": 1.8309, + "theoretical_loss": 4.453359587809431, + "tokens_seen": 176422912 + }, + { + "epoch": 0.05, + "learning_rate": 0.0007648706118355066, + "loss": 1.7832, + "theoretical_loss": 4.453149950271956, + "tokens_seen": 176488448 + }, + { + "epoch": 0.05, + "learning_rate": 0.0007648545636910732, + "loss": 1.9784, + "theoretical_loss": 4.452940412352861, + "tokens_seen": 176553984 + }, + { + "epoch": 0.05, + "learning_rate": 0.00076483851554664, + "loss": 1.983, + "theoretical_loss": 4.452730973967856, + "tokens_seen": 176619520 + }, + { + "epoch": 0.05, + "learning_rate": 0.0007648224674022066, + "loss": 1.7126, + "theoretical_loss": 4.452521635032754, + "tokens_seen": 176685056 + }, + { + "epoch": 0.05, + "learning_rate": 0.0007648064192577734, + "loss": 1.9158, + "theoretical_loss": 4.452312395463473, + "tokens_seen": 176750592 + }, + { + "epoch": 0.05, + "learning_rate": 0.00076479037111334, + "loss": 1.8313, + "theoretical_loss": 4.452103255176027, + "tokens_seen": 176816128 + }, + { + "epoch": 0.05, + "learning_rate": 0.0007647743229689068, + "loss": 1.9122, + "theoretical_loss": 4.451894214086537, + "tokens_seen": 176881664 + }, + { + "epoch": 0.05, + "objective/train/docs_used": 71338, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.6509839296340942, + "objective/train/theoretical_loss": 4.4516852721112254, + "objective/train/tokens_used": 197407200, + "theoretical_loss": 4.4516852721112254, + "tokens_seen": 176947200 + }, + { + "epoch": 0.05, + "learning_rate": 0.0007647582748244734, + "loss": 1.7463, + "theoretical_loss": 4.4516852721112254, + "tokens_seen": 176947200 + }, + { + "epoch": 0.05, + "learning_rate": 0.0007647422266800402, + "loss": 1.7638, + "theoretical_loss": 4.451476429166416, + "tokens_seen": 177012736 + }, + { + "epoch": 0.05, + "learning_rate": 0.0007647261785356068, + "loss": 1.8587, + "theoretical_loss": 4.451267685168532, + "tokens_seen": 177078272 + }, + { + "epoch": 0.05, + "learning_rate": 0.0007647101303911735, + "loss": 1.9486, + "theoretical_loss": 4.451059040034102, + "tokens_seen": 177143808 + }, + { + "epoch": 0.05, + "learning_rate": 0.0007646940822467402, + "loss": 1.7953, + "theoretical_loss": 4.450850493679752, + "tokens_seen": 177209344 + }, + { + "epoch": 0.05, + "learning_rate": 0.0007646780341023069, + "loss": 1.9194, + "theoretical_loss": 4.450642046022211, + "tokens_seen": 177274880 + }, + { + "epoch": 0.05, + "learning_rate": 0.0007646619859578736, + "loss": 1.7716, + "theoretical_loss": 4.45043369697831, + "tokens_seen": 177340416 + }, + { + "epoch": 0.05, + "learning_rate": 0.0007646459378134403, + "loss": 1.764, + "theoretical_loss": 4.450225446464978, + "tokens_seen": 177405952 + }, + { + "epoch": 0.05, + "learning_rate": 0.000764629889669007, + "loss": 1.8743, + "theoretical_loss": 4.450017294399247, + "tokens_seen": 177471488 + }, + { + "epoch": 0.05, + "learning_rate": 0.0007646138415245737, + "loss": 1.8148, + "theoretical_loss": 4.449809240698249, + "tokens_seen": 177537024 + }, + { + "epoch": 0.05, + "learning_rate": 0.0007645977933801404, + "loss": 1.7505, + "theoretical_loss": 4.449601285279215, + "tokens_seen": 177602560 + }, + { + "epoch": 0.05, + "learning_rate": 0.0007645817452357071, + "loss": 1.7193, + "theoretical_loss": 4.449393428059478, + "tokens_seen": 177668096 + }, + { + "epoch": 0.05, + "learning_rate": 0.0007645656970912738, + "loss": 1.8737, + "theoretical_loss": 4.449185668956469, + "tokens_seen": 177733632 + }, + { + "epoch": 0.05, + "learning_rate": 0.0007645496489468405, + "loss": 1.829, + "theoretical_loss": 4.448978007887721, + "tokens_seen": 177799168 + }, + { + "epoch": 0.05, + "learning_rate": 0.0007645336008024072, + "loss": 1.667, + "theoretical_loss": 4.448770444770865, + "tokens_seen": 177864704 + }, + { + "epoch": 0.05, + "learning_rate": 0.0007645175526579739, + "loss": 1.7772, + "theoretical_loss": 4.448562979523633, + "tokens_seen": 177930240 + }, + { + "epoch": 0.05, + "learning_rate": 0.0007645015045135406, + "loss": 1.8008, + "theoretical_loss": 4.448355612063855, + "tokens_seen": 177995776 + }, + { + "epoch": 0.05, + "learning_rate": 0.0007644854563691073, + "loss": 1.8303, + "theoretical_loss": 4.4481483423094605, + "tokens_seen": 178061312 + }, + { + "epoch": 0.05, + "learning_rate": 0.000764469408224674, + "loss": 1.8297, + "theoretical_loss": 4.447941170178478, + "tokens_seen": 178126848 + }, + { + "epoch": 0.05, + "learning_rate": 0.0007644533600802407, + "loss": 1.8252, + "theoretical_loss": 4.447734095589036, + "tokens_seen": 178192384 + }, + { + "epoch": 0.05, + "learning_rate": 0.0007644373119358075, + "loss": 1.7918, + "theoretical_loss": 4.4475271184593606, + "tokens_seen": 178257920 + }, + { + "epoch": 0.05, + "learning_rate": 0.0007644212637913741, + "loss": 1.7006, + "theoretical_loss": 4.447320238707777, + "tokens_seen": 178323456 + }, + { + "epoch": 0.05, + "learning_rate": 0.0007644052156469409, + "loss": 1.9111, + "theoretical_loss": 4.447113456252707, + "tokens_seen": 178388992 + }, + { + "epoch": 0.05, + "learning_rate": 0.0007643891675025075, + "loss": 1.7147, + "theoretical_loss": 4.446906771012674, + "tokens_seen": 178454528 + }, + { + "epoch": 0.05, + "learning_rate": 0.0007643731193580743, + "loss": 1.6387, + "theoretical_loss": 4.446700182906298, + "tokens_seen": 178520064 + }, + { + "epoch": 0.05, + "learning_rate": 0.0007643570712136409, + "loss": 1.8803, + "theoretical_loss": 4.446493691852297, + "tokens_seen": 178585600 + }, + { + "epoch": 0.05, + "learning_rate": 0.0007643410230692077, + "loss": 1.7636, + "theoretical_loss": 4.4462872977694845, + "tokens_seen": 178651136 + }, + { + "epoch": 0.05, + "learning_rate": 0.0007643249749247743, + "loss": 1.8743, + "theoretical_loss": 4.4460810005767755, + "tokens_seen": 178716672 + }, + { + "epoch": 0.05, + "learning_rate": 0.0007643089267803411, + "loss": 1.7634, + "theoretical_loss": 4.445874800193181, + "tokens_seen": 178782208 + }, + { + "epoch": 0.05, + "learning_rate": 0.0007642928786359077, + "loss": 1.6941, + "theoretical_loss": 4.445668696537808, + "tokens_seen": 178847744 + }, + { + "epoch": 0.05, + "learning_rate": 0.0007642768304914745, + "loss": 1.7511, + "theoretical_loss": 4.4454626895298635, + "tokens_seen": 178913280 + }, + { + "epoch": 0.05, + "learning_rate": 0.0007642607823470411, + "loss": 1.6063, + "theoretical_loss": 4.4452567790886475, + "tokens_seen": 178978816 + }, + { + "epoch": 0.05, + "learning_rate": 0.0007642447342026079, + "loss": 1.8915, + "theoretical_loss": 4.445050965133561, + "tokens_seen": 179044352 + }, + { + "epoch": 0.05, + "learning_rate": 0.0007642286860581745, + "loss": 1.744, + "theoretical_loss": 4.4448452475841, + "tokens_seen": 179109888 + }, + { + "epoch": 0.05, + "learning_rate": 0.0007642126379137413, + "loss": 1.9056, + "theoretical_loss": 4.444639626359856, + "tokens_seen": 179175424 + }, + { + "epoch": 0.05, + "learning_rate": 0.0007641965897693079, + "loss": 1.8943, + "theoretical_loss": 4.444434101380519, + "tokens_seen": 179240960 + }, + { + "epoch": 0.05, + "learning_rate": 0.0007641805416248747, + "loss": 1.9521, + "theoretical_loss": 4.444228672565874, + "tokens_seen": 179306496 + }, + { + "epoch": 0.05, + "learning_rate": 0.0007641644934804413, + "loss": 1.9364, + "theoretical_loss": 4.444023339835803, + "tokens_seen": 179372032 + }, + { + "epoch": 0.05, + "learning_rate": 0.0007641484453360081, + "loss": 1.989, + "theoretical_loss": 4.443818103110283, + "tokens_seen": 179437568 + }, + { + "epoch": 0.05, + "learning_rate": 0.0007641323971915747, + "loss": 1.7013, + "theoretical_loss": 4.443612962309387, + "tokens_seen": 179503104 + }, + { + "epoch": 0.05, + "learning_rate": 0.0007641163490471415, + "loss": 1.8315, + "theoretical_loss": 4.4434079173532846, + "tokens_seen": 179568640 + }, + { + "epoch": 0.05, + "learning_rate": 0.0007641003009027081, + "loss": 1.8385, + "theoretical_loss": 4.4432029681622405, + "tokens_seen": 179634176 + }, + { + "epoch": 0.05, + "learning_rate": 0.0007640842527582749, + "loss": 1.803, + "theoretical_loss": 4.442998114656614, + "tokens_seen": 179699712 + }, + { + "epoch": 0.05, + "learning_rate": 0.0007640682046138415, + "loss": 1.8953, + "theoretical_loss": 4.442793356756861, + "tokens_seen": 179765248 + }, + { + "epoch": 0.05, + "learning_rate": 0.0007640521564694083, + "loss": 1.7949, + "theoretical_loss": 4.442588694383532, + "tokens_seen": 179830784 + }, + { + "epoch": 0.05, + "learning_rate": 0.0007640361083249749, + "loss": 1.7326, + "theoretical_loss": 4.44238412745727, + "tokens_seen": 179896320 + }, + { + "epoch": 0.05, + "learning_rate": 0.0007640200601805417, + "loss": 1.8672, + "theoretical_loss": 4.442179655898819, + "tokens_seen": 179961856 + }, + { + "epoch": 0.05, + "learning_rate": 0.0007640040120361083, + "loss": 1.8312, + "theoretical_loss": 4.44197527962901, + "tokens_seen": 180027392 + }, + { + "epoch": 0.05, + "learning_rate": 0.0007639879638916751, + "loss": 1.8831, + "theoretical_loss": 4.441770998568773, + "tokens_seen": 180092928 + }, + { + "epoch": 0.05, + "learning_rate": 0.0007639719157472417, + "loss": 1.9057, + "theoretical_loss": 4.441566812639133, + "tokens_seen": 180158464 + }, + { + "epoch": 0.05, + "objective/train/docs_used": 72550, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.7588038444519043, + "objective/train/theoretical_loss": 4.441362721761207, + "objective/train/tokens_used": 200684000, + "theoretical_loss": 4.441362721761207, + "tokens_seen": 180224000 + }, + { + "epoch": 0.05, + "learning_rate": 0.0007639558676028085, + "loss": 1.8291, + "theoretical_loss": 4.441362721761207, + "tokens_seen": 180224000 + }, + { + "epoch": 0.05, + "learning_rate": 0.0007639398194583752, + "loss": 1.9662, + "theoretical_loss": 4.441158725856207, + "tokens_seen": 180289536 + }, + { + "epoch": 0.05, + "learning_rate": 0.0007639237713139419, + "loss": 1.9503, + "theoretical_loss": 4.440954824845438, + "tokens_seen": 180355072 + }, + { + "epoch": 0.05, + "learning_rate": 0.0007639077231695086, + "loss": 1.9281, + "theoretical_loss": 4.4407510186503, + "tokens_seen": 180420608 + }, + { + "epoch": 0.05, + "learning_rate": 0.0007638916750250753, + "loss": 1.7959, + "theoretical_loss": 4.440547307192286, + "tokens_seen": 180486144 + }, + { + "epoch": 0.05, + "learning_rate": 0.000763875626880642, + "loss": 1.6979, + "theoretical_loss": 4.440343690392984, + "tokens_seen": 180551680 + }, + { + "epoch": 0.05, + "learning_rate": 0.0007638595787362087, + "loss": 1.7899, + "theoretical_loss": 4.4401401681740715, + "tokens_seen": 180617216 + }, + { + "epoch": 0.05, + "learning_rate": 0.0007638435305917754, + "loss": 1.8052, + "theoretical_loss": 4.439936740457323, + "tokens_seen": 180682752 + }, + { + "epoch": 0.05, + "learning_rate": 0.0007638274824473421, + "loss": 1.9025, + "theoretical_loss": 4.439733407164605, + "tokens_seen": 180748288 + }, + { + "epoch": 0.05, + "learning_rate": 0.0007638114343029088, + "loss": 1.8485, + "theoretical_loss": 4.439530168217875, + "tokens_seen": 180813824 + }, + { + "epoch": 0.05, + "learning_rate": 0.0007637953861584755, + "loss": 1.6793, + "theoretical_loss": 4.439327023539187, + "tokens_seen": 180879360 + }, + { + "epoch": 0.05, + "learning_rate": 0.0007637793380140422, + "loss": 1.8288, + "theoretical_loss": 4.439123973050684, + "tokens_seen": 180944896 + }, + { + "epoch": 0.05, + "learning_rate": 0.0007637632898696089, + "loss": 1.8819, + "theoretical_loss": 4.438921016674603, + "tokens_seen": 181010432 + }, + { + "epoch": 0.05, + "learning_rate": 0.0007637472417251756, + "loss": 1.7398, + "theoretical_loss": 4.438718154333274, + "tokens_seen": 181075968 + }, + { + "epoch": 0.05, + "learning_rate": 0.0007637311935807423, + "loss": 1.8457, + "theoretical_loss": 4.438515385949118, + "tokens_seen": 181141504 + }, + { + "epoch": 0.05, + "learning_rate": 0.000763715145436309, + "loss": 1.7854, + "theoretical_loss": 4.438312711444647, + "tokens_seen": 181207040 + }, + { + "epoch": 0.05, + "learning_rate": 0.0007636990972918757, + "loss": 1.8993, + "theoretical_loss": 4.438110130742469, + "tokens_seen": 181272576 + }, + { + "epoch": 0.05, + "learning_rate": 0.0007636830491474424, + "loss": 1.8286, + "theoretical_loss": 4.437907643765281, + "tokens_seen": 181338112 + }, + { + "epoch": 0.05, + "learning_rate": 0.0007636670010030091, + "loss": 1.8124, + "theoretical_loss": 4.43770525043587, + "tokens_seen": 181403648 + }, + { + "epoch": 0.05, + "learning_rate": 0.0007636509528585758, + "loss": 1.9934, + "theoretical_loss": 4.437502950677116, + "tokens_seen": 181469184 + }, + { + "epoch": 0.06, + "learning_rate": 0.0007636349047141425, + "loss": 1.8306, + "theoretical_loss": 4.437300744411992, + "tokens_seen": 181534720 + }, + { + "epoch": 0.06, + "learning_rate": 0.0007636188565697092, + "loss": 1.8454, + "theoretical_loss": 4.43709863156356, + "tokens_seen": 181600256 + }, + { + "epoch": 0.06, + "learning_rate": 0.0007636028084252759, + "loss": 1.7705, + "theoretical_loss": 4.436896612054975, + "tokens_seen": 181665792 + }, + { + "epoch": 0.06, + "learning_rate": 0.0007635867602808426, + "loss": 1.7867, + "theoretical_loss": 4.43669468580948, + "tokens_seen": 181731328 + }, + { + "epoch": 0.06, + "learning_rate": 0.0007635707121364093, + "loss": 1.9099, + "theoretical_loss": 4.43649285275041, + "tokens_seen": 181796864 + }, + { + "epoch": 0.06, + "learning_rate": 0.000763554663991976, + "loss": 1.7299, + "theoretical_loss": 4.436291112801193, + "tokens_seen": 181862400 + }, + { + "epoch": 0.06, + "learning_rate": 0.0007635386158475427, + "loss": 1.8906, + "theoretical_loss": 4.436089465885345, + "tokens_seen": 181927936 + }, + { + "epoch": 0.06, + "learning_rate": 0.0007635225677031094, + "loss": 1.7991, + "theoretical_loss": 4.435887911926472, + "tokens_seen": 181993472 + }, + { + "epoch": 0.06, + "learning_rate": 0.0007635065195586761, + "loss": 1.811, + "theoretical_loss": 4.435686450848273, + "tokens_seen": 182059008 + }, + { + "epoch": 0.06, + "learning_rate": 0.0007634904714142428, + "loss": 1.8356, + "theoretical_loss": 4.435485082574533, + "tokens_seen": 182124544 + }, + { + "epoch": 0.06, + "learning_rate": 0.0007634744232698095, + "loss": 1.8561, + "theoretical_loss": 4.43528380702913, + "tokens_seen": 182190080 + }, + { + "epoch": 0.06, + "learning_rate": 0.0007634583751253762, + "loss": 1.7772, + "theoretical_loss": 4.435082624136031, + "tokens_seen": 182255616 + }, + { + "epoch": 0.06, + "learning_rate": 0.0007634423269809429, + "loss": 1.9161, + "theoretical_loss": 4.4348815338192935, + "tokens_seen": 182321152 + }, + { + "epoch": 0.06, + "learning_rate": 0.0007634262788365095, + "loss": 1.7509, + "theoretical_loss": 4.434680536003062, + "tokens_seen": 182386688 + }, + { + "epoch": 0.06, + "learning_rate": 0.0007634102306920763, + "loss": 1.7807, + "theoretical_loss": 4.434479630611573, + "tokens_seen": 182452224 + }, + { + "epoch": 0.06, + "learning_rate": 0.0007633941825476429, + "loss": 1.9051, + "theoretical_loss": 4.43427881756915, + "tokens_seen": 182517760 + }, + { + "epoch": 0.06, + "learning_rate": 0.0007633781344032097, + "loss": 1.8378, + "theoretical_loss": 4.434078096800209, + "tokens_seen": 182583296 + }, + { + "epoch": 0.06, + "learning_rate": 0.0007633620862587763, + "loss": 1.8617, + "theoretical_loss": 4.433877468229251, + "tokens_seen": 182648832 + }, + { + "epoch": 0.06, + "learning_rate": 0.0007633460381143431, + "loss": 1.8269, + "theoretical_loss": 4.433676931780868, + "tokens_seen": 182714368 + }, + { + "epoch": 0.06, + "learning_rate": 0.0007633299899699097, + "loss": 1.9327, + "theoretical_loss": 4.43347648737974, + "tokens_seen": 182779904 + }, + { + "epoch": 0.06, + "learning_rate": 0.0007633139418254764, + "loss": 1.9527, + "theoretical_loss": 4.433276134950637, + "tokens_seen": 182845440 + }, + { + "epoch": 0.06, + "learning_rate": 0.0007632978936810431, + "loss": 1.7459, + "theoretical_loss": 4.4330758744184156, + "tokens_seen": 182910976 + }, + { + "epoch": 0.06, + "learning_rate": 0.0007632818455366098, + "loss": 1.6841, + "theoretical_loss": 4.432875705708021, + "tokens_seen": 182976512 + }, + { + "epoch": 0.06, + "learning_rate": 0.0007632657973921765, + "loss": 1.8327, + "theoretical_loss": 4.432675628744488, + "tokens_seen": 183042048 + }, + { + "epoch": 0.06, + "learning_rate": 0.0007632497492477432, + "loss": 1.8581, + "theoretical_loss": 4.432475643452937, + "tokens_seen": 183107584 + }, + { + "epoch": 0.06, + "learning_rate": 0.0007632337011033099, + "loss": 1.9965, + "theoretical_loss": 4.432275749758579, + "tokens_seen": 183173120 + }, + { + "epoch": 0.06, + "learning_rate": 0.0007632176529588766, + "loss": 1.8614, + "theoretical_loss": 4.432075947586711, + "tokens_seen": 183238656 + }, + { + "epoch": 0.06, + "learning_rate": 0.0007632016048144433, + "loss": 1.7442, + "theoretical_loss": 4.431876236862717, + "tokens_seen": 183304192 + }, + { + "epoch": 0.06, + "learning_rate": 0.00076318555667001, + "loss": 1.7209, + "theoretical_loss": 4.43167661751207, + "tokens_seen": 183369728 + }, + { + "epoch": 0.06, + "learning_rate": 0.0007631695085255767, + "loss": 1.7781, + "theoretical_loss": 4.431477089460331, + "tokens_seen": 183435264 + }, + { + "epoch": 0.06, + "objective/train/docs_used": 73723, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.889539122581482, + "objective/train/theoretical_loss": 4.431277652633145, + "objective/train/tokens_used": 203960800, + "theoretical_loss": 4.431277652633145, + "tokens_seen": 183500800 + }, + { + "epoch": 0.06, + "learning_rate": 0.0007631534603811434, + "loss": 1.8188, + "theoretical_loss": 4.431277652633145, + "tokens_seen": 183500800 + }, + { + "epoch": 0.06, + "learning_rate": 0.0007631374122367101, + "loss": 1.8824, + "theoretical_loss": 4.431078306956248, + "tokens_seen": 183566336 + }, + { + "epoch": 0.06, + "learning_rate": 0.0007631213640922768, + "loss": 1.8385, + "theoretical_loss": 4.43087905235546, + "tokens_seen": 183631872 + }, + { + "epoch": 0.06, + "learning_rate": 0.0007631053159478436, + "loss": 1.7824, + "theoretical_loss": 4.43067988875669, + "tokens_seen": 183697408 + }, + { + "epoch": 0.06, + "learning_rate": 0.0007630892678034102, + "loss": 1.807, + "theoretical_loss": 4.430480816085931, + "tokens_seen": 183762944 + }, + { + "epoch": 0.06, + "learning_rate": 0.000763073219658977, + "loss": 1.8618, + "theoretical_loss": 4.430281834269265, + "tokens_seen": 183828480 + }, + { + "epoch": 0.06, + "learning_rate": 0.0007630571715145436, + "loss": 1.6989, + "theoretical_loss": 4.4300829432328594, + "tokens_seen": 183894016 + }, + { + "epoch": 0.06, + "learning_rate": 0.0007630411233701104, + "loss": 1.8442, + "theoretical_loss": 4.429884142902968, + "tokens_seen": 183959552 + }, + { + "epoch": 0.06, + "learning_rate": 0.000763025075225677, + "loss": 1.8156, + "theoretical_loss": 4.429685433205931, + "tokens_seen": 184025088 + }, + { + "epoch": 0.06, + "learning_rate": 0.0007630090270812438, + "loss": 1.8515, + "theoretical_loss": 4.429486814068174, + "tokens_seen": 184090624 + }, + { + "epoch": 0.06, + "learning_rate": 0.0007629929789368104, + "loss": 1.7891, + "theoretical_loss": 4.4292882854162094, + "tokens_seen": 184156160 + }, + { + "epoch": 0.06, + "learning_rate": 0.0007629769307923772, + "loss": 1.8251, + "theoretical_loss": 4.429089847176634, + "tokens_seen": 184221696 + }, + { + "epoch": 0.06, + "learning_rate": 0.0007629608826479438, + "loss": 1.8668, + "theoretical_loss": 4.428891499276132, + "tokens_seen": 184287232 + }, + { + "epoch": 0.06, + "learning_rate": 0.0007629448345035106, + "loss": 1.9331, + "theoretical_loss": 4.42869324164147, + "tokens_seen": 184352768 + }, + { + "epoch": 0.06, + "learning_rate": 0.0007629287863590772, + "loss": 1.8247, + "theoretical_loss": 4.428495074199505, + "tokens_seen": 184418304 + }, + { + "epoch": 0.06, + "learning_rate": 0.000762912738214644, + "loss": 1.8322, + "theoretical_loss": 4.428296996877176, + "tokens_seen": 184483840 + }, + { + "epoch": 0.06, + "learning_rate": 0.0007628966900702106, + "loss": 1.8399, + "theoretical_loss": 4.428099009601506, + "tokens_seen": 184549376 + }, + { + "epoch": 0.06, + "learning_rate": 0.0007628806419257774, + "loss": 1.801, + "theoretical_loss": 4.4279011122996055, + "tokens_seen": 184614912 + }, + { + "epoch": 0.06, + "learning_rate": 0.000762864593781344, + "loss": 1.9318, + "theoretical_loss": 4.427703304898668, + "tokens_seen": 184680448 + }, + { + "epoch": 0.06, + "learning_rate": 0.0007628485456369108, + "loss": 1.7681, + "theoretical_loss": 4.427505587325974, + "tokens_seen": 184745984 + }, + { + "epoch": 0.06, + "learning_rate": 0.0007628324974924774, + "loss": 1.7728, + "theoretical_loss": 4.427307959508885, + "tokens_seen": 184811520 + }, + { + "epoch": 0.06, + "learning_rate": 0.0007628164493480442, + "loss": 1.7658, + "theoretical_loss": 4.4271104213748504, + "tokens_seen": 184877056 + }, + { + "epoch": 0.06, + "learning_rate": 0.0007628004012036108, + "loss": 1.7785, + "theoretical_loss": 4.426912972851404, + "tokens_seen": 184942592 + }, + { + "epoch": 0.06, + "learning_rate": 0.0007627843530591776, + "loss": 1.7737, + "theoretical_loss": 4.426715613866159, + "tokens_seen": 185008128 + }, + { + "epoch": 0.06, + "learning_rate": 0.0007627683049147442, + "loss": 1.935, + "theoretical_loss": 4.426518344346819, + "tokens_seen": 185073664 + }, + { + "epoch": 0.06, + "learning_rate": 0.000762752256770311, + "loss": 1.962, + "theoretical_loss": 4.426321164221168, + "tokens_seen": 185139200 + }, + { + "epoch": 0.06, + "learning_rate": 0.0007627362086258777, + "loss": 1.8899, + "theoretical_loss": 4.426124073417074, + "tokens_seen": 185204736 + }, + { + "epoch": 0.06, + "learning_rate": 0.0007627201604814444, + "loss": 1.7733, + "theoretical_loss": 4.42592707186249, + "tokens_seen": 185270272 + }, + { + "epoch": 0.06, + "learning_rate": 0.0007627041123370111, + "loss": 1.7836, + "theoretical_loss": 4.42573015948545, + "tokens_seen": 185335808 + }, + { + "epoch": 0.06, + "learning_rate": 0.0007626880641925778, + "loss": 1.953, + "theoretical_loss": 4.425533336214076, + "tokens_seen": 185401344 + }, + { + "epoch": 0.06, + "learning_rate": 0.0007626720160481445, + "loss": 1.8257, + "theoretical_loss": 4.425336601976568, + "tokens_seen": 185466880 + }, + { + "epoch": 0.06, + "learning_rate": 0.0007626559679037112, + "loss": 1.92, + "theoretical_loss": 4.425139956701214, + "tokens_seen": 185532416 + }, + { + "epoch": 0.06, + "learning_rate": 0.0007626399197592779, + "loss": 1.7861, + "theoretical_loss": 4.42494340031638, + "tokens_seen": 185597952 + }, + { + "epoch": 0.06, + "learning_rate": 0.0007626238716148446, + "loss": 1.9178, + "theoretical_loss": 4.424746932750521, + "tokens_seen": 185663488 + }, + { + "epoch": 0.06, + "learning_rate": 0.0007626078234704113, + "loss": 1.8171, + "theoretical_loss": 4.4245505539321694, + "tokens_seen": 185729024 + }, + { + "epoch": 0.06, + "learning_rate": 0.000762591775325978, + "loss": 1.8207, + "theoretical_loss": 4.424354263789943, + "tokens_seen": 185794560 + }, + { + "epoch": 0.06, + "learning_rate": 0.0007625757271815447, + "loss": 1.8171, + "theoretical_loss": 4.424158062252541, + "tokens_seen": 185860096 + }, + { + "epoch": 0.06, + "learning_rate": 0.0007625596790371114, + "loss": 1.8476, + "theoretical_loss": 4.423961949248748, + "tokens_seen": 185925632 + }, + { + "epoch": 0.06, + "learning_rate": 0.0007625436308926781, + "loss": 1.7672, + "theoretical_loss": 4.423765924707427, + "tokens_seen": 185991168 + }, + { + "epoch": 0.06, + "learning_rate": 0.0007625275827482448, + "loss": 1.7954, + "theoretical_loss": 4.4235699885575235, + "tokens_seen": 186056704 + }, + { + "epoch": 0.06, + "learning_rate": 0.0007625115346038115, + "loss": 1.784, + "theoretical_loss": 4.423374140728068, + "tokens_seen": 186122240 + }, + { + "epoch": 0.06, + "learning_rate": 0.0007624954864593782, + "loss": 1.7246, + "theoretical_loss": 4.423178381148172, + "tokens_seen": 186187776 + }, + { + "epoch": 0.06, + "learning_rate": 0.0007624794383149449, + "loss": 1.8013, + "theoretical_loss": 4.422982709747026, + "tokens_seen": 186253312 + }, + { + "epoch": 0.06, + "learning_rate": 0.0007624633901705116, + "loss": 1.7655, + "theoretical_loss": 4.4227871264539065, + "tokens_seen": 186318848 + }, + { + "epoch": 0.06, + "learning_rate": 0.0007624473420260783, + "loss": 1.7561, + "theoretical_loss": 4.422591631198167, + "tokens_seen": 186384384 + }, + { + "epoch": 0.06, + "learning_rate": 0.000762431293881645, + "loss": 1.7239, + "theoretical_loss": 4.422396223909248, + "tokens_seen": 186449920 + }, + { + "epoch": 0.06, + "learning_rate": 0.0007624152457372117, + "loss": 1.7342, + "theoretical_loss": 4.422200904516666, + "tokens_seen": 186515456 + }, + { + "epoch": 0.06, + "learning_rate": 0.0007623991975927784, + "loss": 1.754, + "theoretical_loss": 4.4220056729500214, + "tokens_seen": 186580992 + }, + { + "epoch": 0.06, + "learning_rate": 0.0007623831494483451, + "loss": 1.8461, + "theoretical_loss": 4.4218105291389955, + "tokens_seen": 186646528 + }, + { + "epoch": 0.06, + "learning_rate": 0.0007623671013039118, + "loss": 1.8693, + "theoretical_loss": 4.4216154730133495, + "tokens_seen": 186712064 + }, + { + "epoch": 0.06, + "objective/train/docs_used": 74904, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.7918874025344849, + "objective/train/theoretical_loss": 4.421420504502928, + "objective/train/tokens_used": 207237600, + "theoretical_loss": 4.421420504502928, + "tokens_seen": 186777600 + }, + { + "epoch": 0.06, + "learning_rate": 0.0007623510531594785, + "loss": 1.7741, + "theoretical_loss": 4.421420504502928, + "tokens_seen": 186777600 + }, + { + "epoch": 0.06, + "learning_rate": 0.0007623350050150452, + "loss": 1.9113, + "theoretical_loss": 4.421225623537652, + "tokens_seen": 186843136 + }, + { + "epoch": 0.06, + "learning_rate": 0.0007623189568706119, + "loss": 1.9085, + "theoretical_loss": 4.421030830047528, + "tokens_seen": 186908672 + }, + { + "epoch": 0.06, + "learning_rate": 0.0007623029087261786, + "loss": 1.9484, + "theoretical_loss": 4.420836123962639, + "tokens_seen": 186974208 + }, + { + "epoch": 0.06, + "learning_rate": 0.0007622868605817453, + "loss": 1.7441, + "theoretical_loss": 4.420641505213151, + "tokens_seen": 187039744 + }, + { + "epoch": 0.06, + "learning_rate": 0.000762270812437312, + "loss": 1.7614, + "theoretical_loss": 4.42044697372931, + "tokens_seen": 187105280 + }, + { + "epoch": 0.06, + "learning_rate": 0.0007622547642928787, + "loss": 1.7798, + "theoretical_loss": 4.420252529441439, + "tokens_seen": 187170816 + }, + { + "epoch": 0.06, + "learning_rate": 0.0007622387161484454, + "loss": 1.9087, + "theoretical_loss": 4.420058172279945, + "tokens_seen": 187236352 + }, + { + "epoch": 0.06, + "learning_rate": 0.0007622226680040121, + "loss": 1.8485, + "theoretical_loss": 4.419863902175313, + "tokens_seen": 187301888 + }, + { + "epoch": 0.06, + "learning_rate": 0.0007622066198595788, + "loss": 1.8394, + "theoretical_loss": 4.419669719058108, + "tokens_seen": 187367424 + }, + { + "epoch": 0.06, + "learning_rate": 0.0007621905717151455, + "loss": 1.9122, + "theoretical_loss": 4.419475622858974, + "tokens_seen": 187432960 + }, + { + "epoch": 0.06, + "learning_rate": 0.0007621745235707122, + "loss": 1.8611, + "theoretical_loss": 4.419281613508637, + "tokens_seen": 187498496 + }, + { + "epoch": 0.06, + "learning_rate": 0.0007621584754262789, + "loss": 1.9366, + "theoretical_loss": 4.419087690937898, + "tokens_seen": 187564032 + }, + { + "epoch": 0.06, + "learning_rate": 0.0007621424272818456, + "loss": 1.9038, + "theoretical_loss": 4.418893855077641, + "tokens_seen": 187629568 + }, + { + "epoch": 0.06, + "learning_rate": 0.0007621263791374123, + "loss": 1.8104, + "theoretical_loss": 4.4187001058588296, + "tokens_seen": 187695104 + }, + { + "epoch": 0.06, + "learning_rate": 0.000762110330992979, + "loss": 1.8724, + "theoretical_loss": 4.418506443212503, + "tokens_seen": 187760640 + }, + { + "epoch": 0.06, + "learning_rate": 0.0007620942828485457, + "loss": 1.8593, + "theoretical_loss": 4.4183128670697815, + "tokens_seen": 187826176 + }, + { + "epoch": 0.06, + "learning_rate": 0.0007620782347041124, + "loss": 1.9296, + "theoretical_loss": 4.418119377361863, + "tokens_seen": 187891712 + }, + { + "epoch": 0.06, + "learning_rate": 0.0007620621865596791, + "loss": 1.831, + "theoretical_loss": 4.417925974020028, + "tokens_seen": 187957248 + }, + { + "epoch": 0.06, + "learning_rate": 0.0007620461384152458, + "loss": 1.8337, + "theoretical_loss": 4.4177326569756294, + "tokens_seen": 188022784 + }, + { + "epoch": 0.06, + "learning_rate": 0.0007620300902708124, + "loss": 1.8712, + "theoretical_loss": 4.417539426160104, + "tokens_seen": 188088320 + }, + { + "epoch": 0.06, + "learning_rate": 0.0007620140421263792, + "loss": 1.8065, + "theoretical_loss": 4.417346281504963, + "tokens_seen": 188153856 + }, + { + "epoch": 0.06, + "learning_rate": 0.0007619979939819458, + "loss": 1.8558, + "theoretical_loss": 4.417153222941798, + "tokens_seen": 188219392 + }, + { + "epoch": 0.06, + "learning_rate": 0.0007619819458375126, + "loss": 1.8862, + "theoretical_loss": 4.416960250402279, + "tokens_seen": 188284928 + }, + { + "epoch": 0.06, + "learning_rate": 0.0007619658976930792, + "loss": 1.8649, + "theoretical_loss": 4.416767363818153, + "tokens_seen": 188350464 + }, + { + "epoch": 0.06, + "learning_rate": 0.000761949849548646, + "loss": 1.7228, + "theoretical_loss": 4.4165745631212445, + "tokens_seen": 188416000 + }, + { + "epoch": 0.06, + "learning_rate": 0.0007619338014042126, + "loss": 1.8196, + "theoretical_loss": 4.416381848243456, + "tokens_seen": 188481536 + }, + { + "epoch": 0.06, + "learning_rate": 0.0007619177532597794, + "loss": 1.7512, + "theoretical_loss": 4.41618921911677, + "tokens_seen": 188547072 + }, + { + "epoch": 0.06, + "learning_rate": 0.000761901705115346, + "loss": 1.947, + "theoretical_loss": 4.415996675673242, + "tokens_seen": 188612608 + }, + { + "epoch": 0.06, + "learning_rate": 0.0007618856569709127, + "loss": 1.9449, + "theoretical_loss": 4.415804217845009, + "tokens_seen": 188678144 + }, + { + "epoch": 0.06, + "learning_rate": 0.0007618696088264794, + "loss": 1.9928, + "theoretical_loss": 4.415611845564284, + "tokens_seen": 188743680 + }, + { + "epoch": 0.06, + "learning_rate": 0.0007618535606820461, + "loss": 1.8115, + "theoretical_loss": 4.415419558763356, + "tokens_seen": 188809216 + }, + { + "epoch": 0.06, + "learning_rate": 0.0007618375125376128, + "loss": 1.9148, + "theoretical_loss": 4.415227357374592, + "tokens_seen": 188874752 + }, + { + "epoch": 0.06, + "learning_rate": 0.0007618214643931795, + "loss": 1.8226, + "theoretical_loss": 4.415035241330436, + "tokens_seen": 188940288 + }, + { + "epoch": 0.06, + "learning_rate": 0.0007618054162487462, + "loss": 1.8371, + "theoretical_loss": 4.41484321056341, + "tokens_seen": 189005824 + }, + { + "epoch": 0.06, + "learning_rate": 0.0007617893681043129, + "loss": 1.8593, + "theoretical_loss": 4.4146512650061105, + "tokens_seen": 189071360 + }, + { + "epoch": 0.06, + "learning_rate": 0.0007617733199598796, + "loss": 1.869, + "theoretical_loss": 4.414459404591211, + "tokens_seen": 189136896 + }, + { + "epoch": 0.06, + "learning_rate": 0.0007617572718154463, + "loss": 1.704, + "theoretical_loss": 4.414267629251463, + "tokens_seen": 189202432 + }, + { + "epoch": 0.06, + "learning_rate": 0.000761741223671013, + "loss": 1.7478, + "theoretical_loss": 4.414075938919694, + "tokens_seen": 189267968 + }, + { + "epoch": 0.06, + "learning_rate": 0.0007617251755265797, + "loss": 1.864, + "theoretical_loss": 4.413884333528807, + "tokens_seen": 189333504 + }, + { + "epoch": 0.06, + "learning_rate": 0.0007617091273821465, + "loss": 1.7906, + "theoretical_loss": 4.413692813011781, + "tokens_seen": 189399040 + }, + { + "epoch": 0.06, + "learning_rate": 0.0007616930792377131, + "loss": 1.7711, + "theoretical_loss": 4.413501377301671, + "tokens_seen": 189464576 + }, + { + "epoch": 0.06, + "learning_rate": 0.0007616770310932799, + "loss": 1.7426, + "theoretical_loss": 4.41331002633161, + "tokens_seen": 189530112 + }, + { + "epoch": 0.06, + "learning_rate": 0.0007616609829488465, + "loss": 1.8997, + "theoretical_loss": 4.413118760034804, + "tokens_seen": 189595648 + }, + { + "epoch": 0.06, + "learning_rate": 0.0007616449348044133, + "loss": 1.8338, + "theoretical_loss": 4.4129275783445365, + "tokens_seen": 189661184 + }, + { + "epoch": 0.06, + "learning_rate": 0.0007616288866599799, + "loss": 1.7991, + "theoretical_loss": 4.412736481194166, + "tokens_seen": 189726720 + }, + { + "epoch": 0.06, + "learning_rate": 0.0007616128385155467, + "loss": 1.8253, + "theoretical_loss": 4.412545468517127, + "tokens_seen": 189792256 + }, + { + "epoch": 0.06, + "learning_rate": 0.0007615967903711133, + "loss": 1.8471, + "theoretical_loss": 4.412354540246929, + "tokens_seen": 189857792 + }, + { + "epoch": 0.06, + "learning_rate": 0.0007615807422266801, + "loss": 1.7034, + "theoretical_loss": 4.4121636963171555, + "tokens_seen": 189923328 + }, + { + "epoch": 0.06, + "learning_rate": 0.0007615646940822467, + "loss": 1.913, + "theoretical_loss": 4.411972936661467, + "tokens_seen": 189988864 + }, + { + "epoch": 0.06, + "objective/train/docs_used": 76077, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.8091943264007568, + "objective/train/theoretical_loss": 4.411782261213599, + "objective/train/tokens_used": 210514400, + "theoretical_loss": 4.411782261213599, + "tokens_seen": 190054400 + }, + { + "epoch": 0.06, + "learning_rate": 0.0007615486459378135, + "loss": 1.8699, + "theoretical_loss": 4.411782261213599, + "tokens_seen": 190054400 + }, + { + "epoch": 0.06, + "learning_rate": 0.0007615325977933802, + "loss": 1.817, + "theoretical_loss": 4.41159166990736, + "tokens_seen": 190119936 + }, + { + "epoch": 0.06, + "learning_rate": 0.0007615165496489469, + "loss": 1.8747, + "theoretical_loss": 4.411401162676636, + "tokens_seen": 190185472 + }, + { + "epoch": 0.06, + "learning_rate": 0.0007615005015045136, + "loss": 1.9713, + "theoretical_loss": 4.411210739455386, + "tokens_seen": 190251008 + }, + { + "epoch": 0.06, + "learning_rate": 0.0007614844533600803, + "loss": 1.7723, + "theoretical_loss": 4.4110204001776445, + "tokens_seen": 190316544 + }, + { + "epoch": 0.06, + "learning_rate": 0.000761468405215647, + "loss": 1.6557, + "theoretical_loss": 4.410830144777519, + "tokens_seen": 190382080 + }, + { + "epoch": 0.06, + "learning_rate": 0.0007614523570712137, + "loss": 1.7131, + "theoretical_loss": 4.410639973189192, + "tokens_seen": 190447616 + }, + { + "epoch": 0.06, + "learning_rate": 0.0007614363089267804, + "loss": 1.8117, + "theoretical_loss": 4.410449885346922, + "tokens_seen": 190513152 + }, + { + "epoch": 0.06, + "learning_rate": 0.0007614202607823471, + "loss": 1.9271, + "theoretical_loss": 4.4102598811850395, + "tokens_seen": 190578688 + }, + { + "epoch": 0.06, + "learning_rate": 0.0007614042126379138, + "loss": 1.8775, + "theoretical_loss": 4.410069960637949, + "tokens_seen": 190644224 + }, + { + "epoch": 0.06, + "learning_rate": 0.0007613881644934805, + "loss": 1.878, + "theoretical_loss": 4.409880123640131, + "tokens_seen": 190709760 + }, + { + "epoch": 0.06, + "learning_rate": 0.0007613721163490472, + "loss": 1.8097, + "theoretical_loss": 4.409690370126139, + "tokens_seen": 190775296 + }, + { + "epoch": 0.06, + "learning_rate": 0.0007613560682046139, + "loss": 1.887, + "theoretical_loss": 4.409500700030598, + "tokens_seen": 190840832 + }, + { + "epoch": 0.06, + "learning_rate": 0.0007613400200601806, + "loss": 1.9223, + "theoretical_loss": 4.409311113288211, + "tokens_seen": 190906368 + }, + { + "epoch": 0.06, + "learning_rate": 0.0007613239719157473, + "loss": 1.8445, + "theoretical_loss": 4.409121609833749, + "tokens_seen": 190971904 + }, + { + "epoch": 0.06, + "learning_rate": 0.000761307923771314, + "loss": 1.7254, + "theoretical_loss": 4.408932189602061, + "tokens_seen": 191037440 + }, + { + "epoch": 0.06, + "learning_rate": 0.0007612918756268807, + "loss": 1.8927, + "theoretical_loss": 4.408742852528068, + "tokens_seen": 191102976 + }, + { + "epoch": 0.06, + "learning_rate": 0.0007612758274824474, + "loss": 1.7393, + "theoretical_loss": 4.408553598546764, + "tokens_seen": 191168512 + }, + { + "epoch": 0.06, + "learning_rate": 0.0007612597793380141, + "loss": 1.7956, + "theoretical_loss": 4.408364427593215, + "tokens_seen": 191234048 + }, + { + "epoch": 0.06, + "learning_rate": 0.0007612437311935808, + "loss": 1.8922, + "theoretical_loss": 4.40817533960256, + "tokens_seen": 191299584 + }, + { + "epoch": 0.06, + "learning_rate": 0.0007612276830491475, + "loss": 1.7484, + "theoretical_loss": 4.407986334510014, + "tokens_seen": 191365120 + }, + { + "epoch": 0.06, + "learning_rate": 0.0007612116349047142, + "loss": 1.9984, + "theoretical_loss": 4.407797412250861, + "tokens_seen": 191430656 + }, + { + "epoch": 0.06, + "learning_rate": 0.0007611955867602809, + "loss": 1.8266, + "theoretical_loss": 4.40760857276046, + "tokens_seen": 191496192 + }, + { + "epoch": 0.06, + "learning_rate": 0.0007611795386158476, + "loss": 1.7061, + "theoretical_loss": 4.407419815974242, + "tokens_seen": 191561728 + }, + { + "epoch": 0.06, + "learning_rate": 0.0007611634904714143, + "loss": 1.8339, + "theoretical_loss": 4.407231141827708, + "tokens_seen": 191627264 + }, + { + "epoch": 0.06, + "learning_rate": 0.000761147442326981, + "loss": 1.8733, + "theoretical_loss": 4.407042550256436, + "tokens_seen": 191692800 + }, + { + "epoch": 0.06, + "learning_rate": 0.0007611313941825477, + "loss": 1.8592, + "theoretical_loss": 4.406854041196072, + "tokens_seen": 191758336 + }, + { + "epoch": 0.06, + "learning_rate": 0.0007611153460381144, + "loss": 1.736, + "theoretical_loss": 4.406665614582338, + "tokens_seen": 191823872 + }, + { + "epoch": 0.06, + "learning_rate": 0.0007610992978936811, + "loss": 1.8059, + "theoretical_loss": 4.406477270351024, + "tokens_seen": 191889408 + }, + { + "epoch": 0.06, + "learning_rate": 0.0007610832497492478, + "loss": 1.8014, + "theoretical_loss": 4.4062890084379935, + "tokens_seen": 191954944 + }, + { + "epoch": 0.06, + "learning_rate": 0.0007610672016048145, + "loss": 1.7036, + "theoretical_loss": 4.406100828779183, + "tokens_seen": 192020480 + }, + { + "epoch": 0.06, + "learning_rate": 0.0007610511534603812, + "loss": 1.6871, + "theoretical_loss": 4.4059127313106, + "tokens_seen": 192086016 + }, + { + "epoch": 0.06, + "learning_rate": 0.0007610351053159479, + "loss": 1.8046, + "theoretical_loss": 4.405724715968323, + "tokens_seen": 192151552 + }, + { + "epoch": 0.06, + "learning_rate": 0.0007610190571715146, + "loss": 1.8272, + "theoretical_loss": 4.405536782688502, + "tokens_seen": 192217088 + }, + { + "epoch": 0.06, + "learning_rate": 0.0007610030090270813, + "loss": 1.8762, + "theoretical_loss": 4.405348931407359, + "tokens_seen": 192282624 + }, + { + "epoch": 0.06, + "learning_rate": 0.000760986960882648, + "loss": 1.8256, + "theoretical_loss": 4.405161162061187, + "tokens_seen": 192348160 + }, + { + "epoch": 0.06, + "learning_rate": 0.0007609709127382147, + "loss": 1.7582, + "theoretical_loss": 4.404973474586351, + "tokens_seen": 192413696 + }, + { + "epoch": 0.06, + "learning_rate": 0.0007609548645937814, + "loss": 1.9133, + "theoretical_loss": 4.404785868919285, + "tokens_seen": 192479232 + }, + { + "epoch": 0.06, + "learning_rate": 0.0007609388164493481, + "loss": 1.6787, + "theoretical_loss": 4.404598344996497, + "tokens_seen": 192544768 + }, + { + "epoch": 0.06, + "learning_rate": 0.0007609227683049148, + "loss": 1.7126, + "theoretical_loss": 4.404410902754562, + "tokens_seen": 192610304 + }, + { + "epoch": 0.06, + "learning_rate": 0.0007609067201604815, + "loss": 1.8023, + "theoretical_loss": 4.404223542130129, + "tokens_seen": 192675840 + }, + { + "epoch": 0.06, + "learning_rate": 0.0007608906720160482, + "loss": 1.7743, + "theoretical_loss": 4.404036263059917, + "tokens_seen": 192741376 + }, + { + "epoch": 0.06, + "learning_rate": 0.0007608746238716149, + "loss": 1.7834, + "theoretical_loss": 4.403849065480714, + "tokens_seen": 192806912 + }, + { + "epoch": 0.06, + "learning_rate": 0.0007608585757271816, + "loss": 1.8974, + "theoretical_loss": 4.403661949329381, + "tokens_seen": 192872448 + }, + { + "epoch": 0.06, + "learning_rate": 0.0007608425275827483, + "loss": 1.8979, + "theoretical_loss": 4.403474914542848, + "tokens_seen": 192937984 + }, + { + "epoch": 0.06, + "learning_rate": 0.000760826479438315, + "loss": 1.8118, + "theoretical_loss": 4.403287961058112, + "tokens_seen": 193003520 + }, + { + "epoch": 0.06, + "learning_rate": 0.0007608104312938817, + "loss": 1.6972, + "theoretical_loss": 4.403101088812248, + "tokens_seen": 193069056 + }, + { + "epoch": 0.06, + "learning_rate": 0.0007607943831494484, + "loss": 1.8564, + "theoretical_loss": 4.4029142977423925, + "tokens_seen": 193134592 + }, + { + "epoch": 0.06, + "learning_rate": 0.0007607783350050151, + "loss": 1.7529, + "theoretical_loss": 4.402727587785758, + "tokens_seen": 193200128 + }, + { + "epoch": 0.06, + "learning_rate": 0.0007607622868605818, + "loss": 1.8115, + "theoretical_loss": 4.402540958879625, + "tokens_seen": 193265664 + }, + { + "epoch": 0.06, + "objective/train/docs_used": 77266, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.8879733085632324, + "objective/train/theoretical_loss": 4.402354410961341, + "objective/train/tokens_used": 213791200, + "theoretical_loss": 4.402354410961341, + "tokens_seen": 193331200 + }, + { + "epoch": 0.06, + "learning_rate": 0.0007607462387161485, + "loss": 1.8115, + "theoretical_loss": 4.402354410961341, + "tokens_seen": 193331200 + }, + { + "epoch": 0.06, + "learning_rate": 0.0007607301905717152, + "loss": 1.7442, + "theoretical_loss": 4.402167943968328, + "tokens_seen": 193396736 + }, + { + "epoch": 0.06, + "learning_rate": 0.0007607141424272819, + "loss": 1.7272, + "theoretical_loss": 4.4019815578380745, + "tokens_seen": 193462272 + }, + { + "epoch": 0.06, + "learning_rate": 0.0007606980942828486, + "loss": 1.7488, + "theoretical_loss": 4.401795252508139, + "tokens_seen": 193527808 + }, + { + "epoch": 0.06, + "learning_rate": 0.0007606820461384153, + "loss": 1.7104, + "theoretical_loss": 4.401609027916149, + "tokens_seen": 193593344 + }, + { + "epoch": 0.06, + "learning_rate": 0.000760665997993982, + "loss": 1.6458, + "theoretical_loss": 4.4014228839998015, + "tokens_seen": 193658880 + }, + { + "epoch": 0.06, + "learning_rate": 0.0007606499498495487, + "loss": 1.7987, + "theoretical_loss": 4.401236820696863, + "tokens_seen": 193724416 + }, + { + "epoch": 0.06, + "learning_rate": 0.0007606339017051153, + "loss": 1.7601, + "theoretical_loss": 4.40105083794517, + "tokens_seen": 193789952 + }, + { + "epoch": 0.06, + "learning_rate": 0.0007606178535606821, + "loss": 1.7309, + "theoretical_loss": 4.400864935682625, + "tokens_seen": 193855488 + }, + { + "epoch": 0.06, + "learning_rate": 0.0007606018054162487, + "loss": 1.7613, + "theoretical_loss": 4.400679113847202, + "tokens_seen": 193921024 + }, + { + "epoch": 0.06, + "learning_rate": 0.0007605857572718155, + "loss": 1.9588, + "theoretical_loss": 4.400493372376943, + "tokens_seen": 193986560 + }, + { + "epoch": 0.06, + "learning_rate": 0.0007605697091273821, + "loss": 1.7493, + "theoretical_loss": 4.400307711209957, + "tokens_seen": 194052096 + }, + { + "epoch": 0.06, + "learning_rate": 0.0007605536609829489, + "loss": 1.8228, + "theoretical_loss": 4.400122130284426, + "tokens_seen": 194117632 + }, + { + "epoch": 0.06, + "learning_rate": 0.0007605376128385155, + "loss": 1.7375, + "theoretical_loss": 4.399936629538595, + "tokens_seen": 194183168 + }, + { + "epoch": 0.06, + "learning_rate": 0.0007605215646940823, + "loss": 1.6181, + "theoretical_loss": 4.3997512089107795, + "tokens_seen": 194248704 + }, + { + "epoch": 0.06, + "learning_rate": 0.0007605055165496489, + "loss": 1.7303, + "theoretical_loss": 4.399565868339366, + "tokens_seen": 194314240 + }, + { + "epoch": 0.06, + "learning_rate": 0.0007604894684052156, + "loss": 1.8176, + "theoretical_loss": 4.399380607762804, + "tokens_seen": 194379776 + }, + { + "epoch": 0.06, + "learning_rate": 0.0007604734202607823, + "loss": 1.7715, + "theoretical_loss": 4.399195427119616, + "tokens_seen": 194445312 + }, + { + "epoch": 0.06, + "learning_rate": 0.000760457372116349, + "loss": 1.8505, + "theoretical_loss": 4.399010326348388, + "tokens_seen": 194510848 + }, + { + "epoch": 0.06, + "learning_rate": 0.0007604413239719157, + "loss": 1.8056, + "theoretical_loss": 4.398825305387778, + "tokens_seen": 194576384 + }, + { + "epoch": 0.06, + "learning_rate": 0.0007604252758274824, + "loss": 1.8772, + "theoretical_loss": 4.39864036417651, + "tokens_seen": 194641920 + }, + { + "epoch": 0.06, + "learning_rate": 0.0007604092276830491, + "loss": 1.7723, + "theoretical_loss": 4.398455502653373, + "tokens_seen": 194707456 + }, + { + "epoch": 0.06, + "learning_rate": 0.0007603931795386158, + "loss": 1.8772, + "theoretical_loss": 4.398270720757228, + "tokens_seen": 194772992 + }, + { + "epoch": 0.06, + "learning_rate": 0.0007603771313941825, + "loss": 1.9226, + "theoretical_loss": 4.398086018427001, + "tokens_seen": 194838528 + }, + { + "epoch": 0.06, + "learning_rate": 0.0007603610832497492, + "loss": 1.7992, + "theoretical_loss": 4.397901395601687, + "tokens_seen": 194904064 + }, + { + "epoch": 0.06, + "learning_rate": 0.000760345035105316, + "loss": 1.8819, + "theoretical_loss": 4.397716852220345, + "tokens_seen": 194969600 + }, + { + "epoch": 0.06, + "learning_rate": 0.0007603289869608827, + "loss": 1.8732, + "theoretical_loss": 4.397532388222105, + "tokens_seen": 195035136 + }, + { + "epoch": 0.06, + "learning_rate": 0.0007603129388164494, + "loss": 1.8382, + "theoretical_loss": 4.397348003546161, + "tokens_seen": 195100672 + }, + { + "epoch": 0.06, + "learning_rate": 0.0007602968906720161, + "loss": 1.7448, + "theoretical_loss": 4.397163698131776, + "tokens_seen": 195166208 + }, + { + "epoch": 0.06, + "learning_rate": 0.0007602808425275828, + "loss": 1.6997, + "theoretical_loss": 4.39697947191828, + "tokens_seen": 195231744 + }, + { + "epoch": 0.06, + "learning_rate": 0.0007602647943831495, + "loss": 1.8206, + "theoretical_loss": 4.396795324845069, + "tokens_seen": 195297280 + }, + { + "epoch": 0.06, + "learning_rate": 0.0007602487462387162, + "loss": 1.7654, + "theoretical_loss": 4.396611256851604, + "tokens_seen": 195362816 + }, + { + "epoch": 0.06, + "learning_rate": 0.0007602326980942829, + "loss": 1.72, + "theoretical_loss": 4.3964272678774154, + "tokens_seen": 195428352 + }, + { + "epoch": 0.06, + "learning_rate": 0.0007602166499498496, + "loss": 1.7236, + "theoretical_loss": 4.396243357862099, + "tokens_seen": 195493888 + }, + { + "epoch": 0.06, + "learning_rate": 0.0007602006018054163, + "loss": 1.7991, + "theoretical_loss": 4.396059526745318, + "tokens_seen": 195559424 + }, + { + "epoch": 0.06, + "learning_rate": 0.000760184553660983, + "loss": 1.7774, + "theoretical_loss": 4.395875774466799, + "tokens_seen": 195624960 + }, + { + "epoch": 0.06, + "learning_rate": 0.0007601685055165497, + "loss": 1.7815, + "theoretical_loss": 4.395692100966338, + "tokens_seen": 195690496 + }, + { + "epoch": 0.06, + "learning_rate": 0.0007601524573721164, + "loss": 1.6686, + "theoretical_loss": 4.395508506183796, + "tokens_seen": 195756032 + }, + { + "epoch": 0.06, + "learning_rate": 0.0007601364092276831, + "loss": 1.9592, + "theoretical_loss": 4.395324990059098, + "tokens_seen": 195821568 + }, + { + "epoch": 0.06, + "learning_rate": 0.0007601203610832498, + "loss": 1.7582, + "theoretical_loss": 4.395141552532239, + "tokens_seen": 195887104 + }, + { + "epoch": 0.06, + "learning_rate": 0.0007601043129388165, + "loss": 1.8789, + "theoretical_loss": 4.394958193543277, + "tokens_seen": 195952640 + }, + { + "epoch": 0.06, + "learning_rate": 0.0007600882647943832, + "loss": 1.8123, + "theoretical_loss": 4.394774913032337, + "tokens_seen": 196018176 + }, + { + "epoch": 0.06, + "learning_rate": 0.0007600722166499499, + "loss": 1.7184, + "theoretical_loss": 4.394591710939608, + "tokens_seen": 196083712 + }, + { + "epoch": 0.06, + "learning_rate": 0.0007600561685055166, + "loss": 1.7022, + "theoretical_loss": 4.394408587205344, + "tokens_seen": 196149248 + }, + { + "epoch": 0.06, + "learning_rate": 0.0007600401203610833, + "loss": 1.8023, + "theoretical_loss": 4.39422554176987, + "tokens_seen": 196214784 + }, + { + "epoch": 0.06, + "learning_rate": 0.00076002407221665, + "loss": 1.6827, + "theoretical_loss": 4.39404257457357, + "tokens_seen": 196280320 + }, + { + "epoch": 0.06, + "learning_rate": 0.0007600080240722167, + "loss": 1.6646, + "theoretical_loss": 4.393859685556895, + "tokens_seen": 196345856 + }, + { + "epoch": 0.06, + "learning_rate": 0.0007599919759277834, + "loss": 1.865, + "theoretical_loss": 4.393676874660364, + "tokens_seen": 196411392 + }, + { + "epoch": 0.06, + "learning_rate": 0.0007599759277833501, + "loss": 1.7314, + "theoretical_loss": 4.393494141824557, + "tokens_seen": 196476928 + }, + { + "epoch": 0.06, + "learning_rate": 0.0007599598796389168, + "loss": 1.7691, + "theoretical_loss": 4.393311486990122, + "tokens_seen": 196542464 + }, + { + "epoch": 0.06, + "objective/train/docs_used": 78313, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.5786049365997314, + "objective/train/theoretical_loss": 4.39312891009777, + "objective/train/tokens_used": 217068000, + "theoretical_loss": 4.39312891009777, + "tokens_seen": 196608000 + }, + { + "epoch": 0.06, + "learning_rate": 0.0007599438314944835, + "loss": 1.6781, + "theoretical_loss": 4.39312891009777, + "tokens_seen": 196608000 + }, + { + "epoch": 0.06, + "learning_rate": 0.0007599277833500502, + "loss": 1.7706, + "theoretical_loss": 4.392946411088279, + "tokens_seen": 196673536 + }, + { + "epoch": 0.06, + "learning_rate": 0.0007599117352056169, + "loss": 1.8297, + "theoretical_loss": 4.392763989902489, + "tokens_seen": 196739072 + }, + { + "epoch": 0.06, + "learning_rate": 0.0007598956870611836, + "loss": 1.7464, + "theoretical_loss": 4.3925816464813074, + "tokens_seen": 196804608 + }, + { + "epoch": 0.06, + "learning_rate": 0.0007598796389167503, + "loss": 1.7787, + "theoretical_loss": 4.392399380765704, + "tokens_seen": 196870144 + }, + { + "epoch": 0.06, + "learning_rate": 0.000759863590772317, + "loss": 1.7716, + "theoretical_loss": 4.392217192696712, + "tokens_seen": 196935680 + }, + { + "epoch": 0.06, + "learning_rate": 0.0007598475426278837, + "loss": 1.886, + "theoretical_loss": 4.392035082215434, + "tokens_seen": 197001216 + }, + { + "epoch": 0.06, + "learning_rate": 0.0007598314944834504, + "loss": 1.7435, + "theoretical_loss": 4.391853049263032, + "tokens_seen": 197066752 + }, + { + "epoch": 0.06, + "learning_rate": 0.0007598154463390171, + "loss": 1.7621, + "theoretical_loss": 4.391671093780733, + "tokens_seen": 197132288 + }, + { + "epoch": 0.06, + "learning_rate": 0.0007597993981945838, + "loss": 1.6715, + "theoretical_loss": 4.39148921570983, + "tokens_seen": 197197824 + }, + { + "epoch": 0.06, + "learning_rate": 0.0007597833500501505, + "loss": 1.5949, + "theoretical_loss": 4.391307414991678, + "tokens_seen": 197263360 + }, + { + "epoch": 0.06, + "learning_rate": 0.0007597673019057172, + "loss": 1.8478, + "theoretical_loss": 4.391125691567697, + "tokens_seen": 197328896 + }, + { + "epoch": 0.06, + "learning_rate": 0.0007597512537612839, + "loss": 1.831, + "theoretical_loss": 4.390944045379371, + "tokens_seen": 197394432 + }, + { + "epoch": 0.06, + "learning_rate": 0.0007597352056168506, + "loss": 1.7955, + "theoretical_loss": 4.390762476368245, + "tokens_seen": 197459968 + }, + { + "epoch": 0.06, + "learning_rate": 0.0007597191574724173, + "loss": 1.7715, + "theoretical_loss": 4.390580984475932, + "tokens_seen": 197525504 + }, + { + "epoch": 0.06, + "learning_rate": 0.000759703109327984, + "loss": 1.8446, + "theoretical_loss": 4.3903995696441065, + "tokens_seen": 197591040 + }, + { + "epoch": 0.06, + "learning_rate": 0.0007596870611835507, + "loss": 1.8946, + "theoretical_loss": 4.390218231814504, + "tokens_seen": 197656576 + }, + { + "epoch": 0.06, + "learning_rate": 0.0007596710130391174, + "loss": 1.7017, + "theoretical_loss": 4.390036970928929, + "tokens_seen": 197722112 + }, + { + "epoch": 0.06, + "learning_rate": 0.0007596549648946841, + "loss": 1.6968, + "theoretical_loss": 4.3898557869292425, + "tokens_seen": 197787648 + }, + { + "epoch": 0.06, + "learning_rate": 0.0007596389167502508, + "loss": 1.7624, + "theoretical_loss": 4.389674679757374, + "tokens_seen": 197853184 + }, + { + "epoch": 0.06, + "learning_rate": 0.0007596228686058175, + "loss": 1.7091, + "theoretical_loss": 4.389493649355314, + "tokens_seen": 197918720 + }, + { + "epoch": 0.06, + "learning_rate": 0.0007596068204613842, + "loss": 1.7137, + "theoretical_loss": 4.389312695665115, + "tokens_seen": 197984256 + }, + { + "epoch": 0.06, + "learning_rate": 0.0007595907723169509, + "loss": 1.8276, + "theoretical_loss": 4.389131818628895, + "tokens_seen": 198049792 + }, + { + "epoch": 0.06, + "learning_rate": 0.0007595747241725176, + "loss": 1.8001, + "theoretical_loss": 4.388951018188832, + "tokens_seen": 198115328 + }, + { + "epoch": 0.06, + "learning_rate": 0.0007595586760280843, + "loss": 1.7906, + "theoretical_loss": 4.388770294287169, + "tokens_seen": 198180864 + }, + { + "epoch": 0.06, + "learning_rate": 0.000759542627883651, + "loss": 1.8407, + "theoretical_loss": 4.38858964686621, + "tokens_seen": 198246400 + }, + { + "epoch": 0.06, + "learning_rate": 0.0007595265797392177, + "loss": 1.8326, + "theoretical_loss": 4.388409075868323, + "tokens_seen": 198311936 + }, + { + "epoch": 0.06, + "learning_rate": 0.0007595105315947844, + "loss": 1.8717, + "theoretical_loss": 4.388228581235937, + "tokens_seen": 198377472 + }, + { + "epoch": 0.06, + "learning_rate": 0.0007594944834503511, + "loss": 1.8201, + "theoretical_loss": 4.388048162911545, + "tokens_seen": 198443008 + }, + { + "epoch": 0.06, + "learning_rate": 0.0007594784353059178, + "loss": 1.7396, + "theoretical_loss": 4.3878678208377, + "tokens_seen": 198508544 + }, + { + "epoch": 0.06, + "learning_rate": 0.0007594623871614845, + "loss": 1.8219, + "theoretical_loss": 4.38768755495702, + "tokens_seen": 198574080 + }, + { + "epoch": 0.06, + "learning_rate": 0.0007594463390170512, + "loss": 1.7759, + "theoretical_loss": 4.387507365212182, + "tokens_seen": 198639616 + }, + { + "epoch": 0.06, + "learning_rate": 0.0007594302908726179, + "loss": 1.734, + "theoretical_loss": 4.387327251545927, + "tokens_seen": 198705152 + }, + { + "epoch": 0.06, + "learning_rate": 0.0007594142427281846, + "loss": 1.7317, + "theoretical_loss": 4.387147213901059, + "tokens_seen": 198770688 + }, + { + "epoch": 0.06, + "learning_rate": 0.0007593981945837513, + "loss": 1.7094, + "theoretical_loss": 4.386967252220441, + "tokens_seen": 198836224 + }, + { + "epoch": 0.06, + "learning_rate": 0.000759382146439318, + "loss": 1.8516, + "theoretical_loss": 4.386787366447, + "tokens_seen": 198901760 + }, + { + "epoch": 0.06, + "learning_rate": 0.0007593660982948847, + "loss": 1.8624, + "theoretical_loss": 4.386607556523723, + "tokens_seen": 198967296 + }, + { + "epoch": 0.06, + "learning_rate": 0.0007593500501504514, + "loss": 1.8463, + "theoretical_loss": 4.386427822393661, + "tokens_seen": 199032832 + }, + { + "epoch": 0.06, + "learning_rate": 0.0007593340020060181, + "loss": 1.7913, + "theoretical_loss": 4.386248163999923, + "tokens_seen": 199098368 + }, + { + "epoch": 0.06, + "learning_rate": 0.0007593179538615848, + "loss": 1.8448, + "theoretical_loss": 4.3860685812856826, + "tokens_seen": 199163904 + }, + { + "epoch": 0.06, + "learning_rate": 0.0007593019057171515, + "loss": 1.8689, + "theoretical_loss": 4.385889074194173, + "tokens_seen": 199229440 + }, + { + "epoch": 0.06, + "learning_rate": 0.0007592858575727182, + "loss": 1.7348, + "theoretical_loss": 4.385709642668688, + "tokens_seen": 199294976 + }, + { + "epoch": 0.06, + "learning_rate": 0.000759269809428285, + "loss": 1.6719, + "theoretical_loss": 4.385530286652586, + "tokens_seen": 199360512 + }, + { + "epoch": 0.06, + "learning_rate": 0.0007592537612838516, + "loss": 1.7126, + "theoretical_loss": 4.385351006089283, + "tokens_seen": 199426048 + }, + { + "epoch": 0.06, + "learning_rate": 0.0007592377131394182, + "loss": 1.7359, + "theoretical_loss": 4.385171800922257, + "tokens_seen": 199491584 + }, + { + "epoch": 0.06, + "learning_rate": 0.000759221664994985, + "loss": 1.7748, + "theoretical_loss": 4.384992671095047, + "tokens_seen": 199557120 + }, + { + "epoch": 0.06, + "learning_rate": 0.0007592056168505516, + "loss": 1.7288, + "theoretical_loss": 4.384813616551252, + "tokens_seen": 199622656 + }, + { + "epoch": 0.06, + "learning_rate": 0.0007591895687061184, + "loss": 1.6874, + "theoretical_loss": 4.384634637234534, + "tokens_seen": 199688192 + }, + { + "epoch": 0.06, + "learning_rate": 0.000759173520561685, + "loss": 1.7768, + "theoretical_loss": 4.384455733088612, + "tokens_seen": 199753728 + }, + { + "epoch": 0.06, + "learning_rate": 0.0007591574724172518, + "loss": 1.7854, + "theoretical_loss": 4.38427690405727, + "tokens_seen": 199819264 + }, + { + "epoch": 0.06, + "objective/train/docs_used": 79488, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.6650798320770264, + "objective/train/theoretical_loss": 4.384098150084349, + "objective/train/tokens_used": 220344800, + "theoretical_loss": 4.384098150084349, + "tokens_seen": 199884800 + }, + { + "epoch": 0.06, + "learning_rate": 0.0007591414242728184, + "loss": 1.7311, + "theoretical_loss": 4.384098150084349, + "tokens_seen": 199884800 + }, + { + "epoch": 0.06, + "learning_rate": 0.0007591253761283852, + "loss": 1.8218, + "theoretical_loss": 4.383919471113751, + "tokens_seen": 199950336 + }, + { + "epoch": 0.06, + "learning_rate": 0.0007591093279839518, + "loss": 1.8714, + "theoretical_loss": 4.383740867089441, + "tokens_seen": 200015872 + }, + { + "epoch": 0.06, + "learning_rate": 0.0007590932798395186, + "loss": 1.7704, + "theoretical_loss": 4.383562337955439, + "tokens_seen": 200081408 + }, + { + "epoch": 0.06, + "learning_rate": 0.0007590772316950852, + "loss": 1.6368, + "theoretical_loss": 4.3833838836558305, + "tokens_seen": 200146944 + }, + { + "epoch": 0.06, + "learning_rate": 0.000759061183550652, + "loss": 1.7903, + "theoretical_loss": 4.383205504134758, + "tokens_seen": 200212480 + }, + { + "epoch": 0.06, + "learning_rate": 0.0007590451354062186, + "loss": 1.8471, + "theoretical_loss": 4.383027199336424, + "tokens_seen": 200278016 + }, + { + "epoch": 0.06, + "learning_rate": 0.0007590290872617854, + "loss": 1.7998, + "theoretical_loss": 4.382848969205092, + "tokens_seen": 200343552 + }, + { + "epoch": 0.06, + "learning_rate": 0.000759013039117352, + "loss": 1.885, + "theoretical_loss": 4.3826708136850865, + "tokens_seen": 200409088 + }, + { + "epoch": 0.06, + "learning_rate": 0.0007589969909729188, + "loss": 1.8351, + "theoretical_loss": 4.382492732720786, + "tokens_seen": 200474624 + }, + { + "epoch": 0.06, + "learning_rate": 0.0007589809428284855, + "loss": 1.8994, + "theoretical_loss": 4.382314726256637, + "tokens_seen": 200540160 + }, + { + "epoch": 0.06, + "learning_rate": 0.0007589648946840522, + "loss": 1.7233, + "theoretical_loss": 4.382136794237138, + "tokens_seen": 200605696 + }, + { + "epoch": 0.06, + "learning_rate": 0.0007589488465396189, + "loss": 1.8104, + "theoretical_loss": 4.3819589366068525, + "tokens_seen": 200671232 + }, + { + "epoch": 0.06, + "learning_rate": 0.0007589327983951856, + "loss": 1.7187, + "theoretical_loss": 4.381781153310399, + "tokens_seen": 200736768 + }, + { + "epoch": 0.06, + "learning_rate": 0.0007589167502507523, + "loss": 1.7592, + "theoretical_loss": 4.381603444292459, + "tokens_seen": 200802304 + }, + { + "epoch": 0.06, + "learning_rate": 0.000758900702106319, + "loss": 1.6848, + "theoretical_loss": 4.38142580949777, + "tokens_seen": 200867840 + }, + { + "epoch": 0.06, + "learning_rate": 0.0007588846539618857, + "loss": 1.7401, + "theoretical_loss": 4.381248248871131, + "tokens_seen": 200933376 + }, + { + "epoch": 0.06, + "learning_rate": 0.0007588686058174524, + "loss": 1.7501, + "theoretical_loss": 4.3810707623574, + "tokens_seen": 200998912 + }, + { + "epoch": 0.06, + "learning_rate": 0.0007588525576730191, + "loss": 1.7698, + "theoretical_loss": 4.380893349901491, + "tokens_seen": 201064448 + }, + { + "epoch": 0.06, + "learning_rate": 0.0007588365095285858, + "loss": 1.6793, + "theoretical_loss": 4.380716011448381, + "tokens_seen": 201129984 + }, + { + "epoch": 0.06, + "learning_rate": 0.0007588204613841525, + "loss": 1.8001, + "theoretical_loss": 4.380538746943103, + "tokens_seen": 201195520 + }, + { + "epoch": 0.06, + "learning_rate": 0.0007588044132397192, + "loss": 1.6584, + "theoretical_loss": 4.380361556330749, + "tokens_seen": 201261056 + }, + { + "epoch": 0.06, + "learning_rate": 0.0007587883650952859, + "loss": 1.6249, + "theoretical_loss": 4.380184439556472, + "tokens_seen": 201326592 + }, + { + "epoch": 0.06, + "learning_rate": 0.0007587723169508526, + "loss": 1.7555, + "theoretical_loss": 4.38000739656548, + "tokens_seen": 201392128 + }, + { + "epoch": 0.06, + "learning_rate": 0.0007587562688064193, + "loss": 1.7704, + "theoretical_loss": 4.379830427303041, + "tokens_seen": 201457664 + }, + { + "epoch": 0.06, + "learning_rate": 0.000758740220661986, + "loss": 1.7364, + "theoretical_loss": 4.379653531714483, + "tokens_seen": 201523200 + }, + { + "epoch": 0.06, + "learning_rate": 0.0007587241725175527, + "loss": 1.8765, + "theoretical_loss": 4.37947670974519, + "tokens_seen": 201588736 + }, + { + "epoch": 0.06, + "learning_rate": 0.0007587081243731194, + "loss": 1.6823, + "theoretical_loss": 4.379299961340605, + "tokens_seen": 201654272 + }, + { + "epoch": 0.06, + "learning_rate": 0.0007586920762286861, + "loss": 1.8307, + "theoretical_loss": 4.37912328644623, + "tokens_seen": 201719808 + }, + { + "epoch": 0.06, + "learning_rate": 0.0007586760280842528, + "loss": 1.8128, + "theoretical_loss": 4.378946685007623, + "tokens_seen": 201785344 + }, + { + "epoch": 0.06, + "learning_rate": 0.0007586599799398195, + "loss": 1.8563, + "theoretical_loss": 4.378770156970403, + "tokens_seen": 201850880 + }, + { + "epoch": 0.06, + "learning_rate": 0.0007586439317953862, + "loss": 1.7582, + "theoretical_loss": 4.378593702280243, + "tokens_seen": 201916416 + }, + { + "epoch": 0.06, + "learning_rate": 0.0007586278836509529, + "loss": 1.7233, + "theoretical_loss": 4.378417320882879, + "tokens_seen": 201981952 + }, + { + "epoch": 0.06, + "learning_rate": 0.0007586118355065196, + "loss": 1.6604, + "theoretical_loss": 4.378241012724099, + "tokens_seen": 202047488 + }, + { + "epoch": 0.06, + "learning_rate": 0.0007585957873620863, + "loss": 1.8229, + "theoretical_loss": 4.378064777749753, + "tokens_seen": 202113024 + }, + { + "epoch": 0.06, + "learning_rate": 0.000758579739217653, + "loss": 1.7182, + "theoretical_loss": 4.377888615905746, + "tokens_seen": 202178560 + }, + { + "epoch": 0.06, + "learning_rate": 0.0007585636910732197, + "loss": 1.7892, + "theoretical_loss": 4.377712527138042, + "tokens_seen": 202244096 + }, + { + "epoch": 0.06, + "learning_rate": 0.0007585476429287864, + "loss": 1.8195, + "theoretical_loss": 4.3775365113926625, + "tokens_seen": 202309632 + }, + { + "epoch": 0.06, + "learning_rate": 0.0007585315947843531, + "loss": 1.6676, + "theoretical_loss": 4.377360568615685, + "tokens_seen": 202375168 + }, + { + "epoch": 0.06, + "learning_rate": 0.0007585155466399198, + "loss": 1.7534, + "theoretical_loss": 4.377184698753244, + "tokens_seen": 202440704 + }, + { + "epoch": 0.06, + "learning_rate": 0.0007584994984954865, + "loss": 1.6647, + "theoretical_loss": 4.377008901751536, + "tokens_seen": 202506240 + }, + { + "epoch": 0.06, + "learning_rate": 0.0007584834503510532, + "loss": 1.7792, + "theoretical_loss": 4.376833177556806, + "tokens_seen": 202571776 + }, + { + "epoch": 0.06, + "learning_rate": 0.0007584674022066199, + "loss": 1.8167, + "theoretical_loss": 4.3766575261153635, + "tokens_seen": 202637312 + }, + { + "epoch": 0.06, + "learning_rate": 0.0007584513540621866, + "loss": 1.8286, + "theoretical_loss": 4.3764819473735725, + "tokens_seen": 202702848 + }, + { + "epoch": 0.06, + "learning_rate": 0.0007584353059177533, + "loss": 1.7228, + "theoretical_loss": 4.376306441277852, + "tokens_seen": 202768384 + }, + { + "epoch": 0.06, + "learning_rate": 0.00075841925777332, + "loss": 1.7352, + "theoretical_loss": 4.376131007774681, + "tokens_seen": 202833920 + }, + { + "epoch": 0.06, + "learning_rate": 0.0007584032096288867, + "loss": 1.7809, + "theoretical_loss": 4.375955646810594, + "tokens_seen": 202899456 + }, + { + "epoch": 0.06, + "learning_rate": 0.0007583871614844534, + "loss": 1.681, + "theoretical_loss": 4.37578035833218, + "tokens_seen": 202964992 + }, + { + "epoch": 0.06, + "learning_rate": 0.0007583711133400201, + "loss": 1.7076, + "theoretical_loss": 4.375605142286087, + "tokens_seen": 203030528 + }, + { + "epoch": 0.06, + "learning_rate": 0.0007583550651955868, + "loss": 1.7759, + "theoretical_loss": 4.37542999861902, + "tokens_seen": 203096064 + }, + { + "epoch": 0.06, + "objective/train/docs_used": 80605, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.8541486263275146, + "objective/train/theoretical_loss": 4.375254927277737, + "objective/train/tokens_used": 223621600, + "theoretical_loss": 4.375254927277737, + "tokens_seen": 203161600 + }, + { + "epoch": 0.06, + "learning_rate": 0.0007583390170511535, + "loss": 1.8603, + "theoretical_loss": 4.375254927277737, + "tokens_seen": 203161600 + }, + { + "epoch": 0.06, + "learning_rate": 0.0007583229689067202, + "loss": 1.8581, + "theoretical_loss": 4.375079928209058, + "tokens_seen": 203227136 + }, + { + "epoch": 0.06, + "learning_rate": 0.0007583069207622869, + "loss": 1.6736, + "theoretical_loss": 4.374905001359851, + "tokens_seen": 203292672 + }, + { + "epoch": 0.06, + "learning_rate": 0.0007582908726178536, + "loss": 1.7673, + "theoretical_loss": 4.374730146677049, + "tokens_seen": 203358208 + }, + { + "epoch": 0.06, + "learning_rate": 0.0007582748244734203, + "loss": 1.8092, + "theoretical_loss": 4.374555364107635, + "tokens_seen": 203423744 + }, + { + "epoch": 0.06, + "learning_rate": 0.000758258776328987, + "loss": 1.7492, + "theoretical_loss": 4.37438065359865, + "tokens_seen": 203489280 + }, + { + "epoch": 0.06, + "learning_rate": 0.0007582427281845537, + "loss": 1.8158, + "theoretical_loss": 4.374206015097193, + "tokens_seen": 203554816 + }, + { + "epoch": 0.06, + "learning_rate": 0.0007582266800401204, + "loss": 1.6673, + "theoretical_loss": 4.374031448550414, + "tokens_seen": 203620352 + }, + { + "epoch": 0.06, + "learning_rate": 0.0007582106318956871, + "loss": 1.7598, + "theoretical_loss": 4.373856953905523, + "tokens_seen": 203685888 + }, + { + "epoch": 0.06, + "learning_rate": 0.0007581945837512538, + "loss": 1.8609, + "theoretical_loss": 4.373682531109785, + "tokens_seen": 203751424 + }, + { + "epoch": 0.06, + "learning_rate": 0.0007581785356068205, + "loss": 1.7683, + "theoretical_loss": 4.37350818011052, + "tokens_seen": 203816960 + }, + { + "epoch": 0.06, + "learning_rate": 0.0007581624874623872, + "loss": 1.8358, + "theoretical_loss": 4.3733339008551, + "tokens_seen": 203882496 + }, + { + "epoch": 0.06, + "learning_rate": 0.0007581464393179539, + "loss": 1.7821, + "theoretical_loss": 4.373159693290961, + "tokens_seen": 203948032 + }, + { + "epoch": 0.06, + "learning_rate": 0.0007581303911735206, + "loss": 1.7518, + "theoretical_loss": 4.372985557365586, + "tokens_seen": 204013568 + }, + { + "epoch": 0.06, + "learning_rate": 0.0007581143430290873, + "loss": 1.7735, + "theoretical_loss": 4.372811493026518, + "tokens_seen": 204079104 + }, + { + "epoch": 0.06, + "learning_rate": 0.000758098294884654, + "loss": 1.7306, + "theoretical_loss": 4.3726375002213524, + "tokens_seen": 204144640 + }, + { + "epoch": 0.06, + "learning_rate": 0.0007580822467402207, + "loss": 1.6718, + "theoretical_loss": 4.372463578897743, + "tokens_seen": 204210176 + }, + { + "epoch": 0.06, + "learning_rate": 0.0007580661985957874, + "loss": 1.7891, + "theoretical_loss": 4.372289729003397, + "tokens_seen": 204275712 + }, + { + "epoch": 0.06, + "learning_rate": 0.0007580501504513541, + "loss": 1.6962, + "theoretical_loss": 4.372115950486075, + "tokens_seen": 204341248 + }, + { + "epoch": 0.06, + "learning_rate": 0.0007580341023069208, + "loss": 1.8626, + "theoretical_loss": 4.371942243293595, + "tokens_seen": 204406784 + }, + { + "epoch": 0.06, + "learning_rate": 0.0007580180541624875, + "loss": 1.8004, + "theoretical_loss": 4.371768607373829, + "tokens_seen": 204472320 + }, + { + "epoch": 0.06, + "learning_rate": 0.0007580020060180542, + "loss": 1.8158, + "theoretical_loss": 4.371595042674704, + "tokens_seen": 204537856 + }, + { + "epoch": 0.06, + "learning_rate": 0.0007579859578736209, + "loss": 1.6406, + "theoretical_loss": 4.371421549144202, + "tokens_seen": 204603392 + }, + { + "epoch": 0.06, + "learning_rate": 0.0007579699097291876, + "loss": 1.7805, + "theoretical_loss": 4.371248126730357, + "tokens_seen": 204668928 + }, + { + "epoch": 0.06, + "learning_rate": 0.0007579538615847543, + "loss": 1.9244, + "theoretical_loss": 4.3710747753812615, + "tokens_seen": 204734464 + }, + { + "epoch": 0.06, + "learning_rate": 0.000757937813440321, + "loss": 1.6842, + "theoretical_loss": 4.370901495045061, + "tokens_seen": 204800000 + }, + { + "epoch": 0.06, + "learning_rate": 0.0007579217652958877, + "loss": 1.8736, + "theoretical_loss": 4.370728285669954, + "tokens_seen": 204865536 + }, + { + "epoch": 0.06, + "learning_rate": 0.0007579057171514544, + "loss": 1.8496, + "theoretical_loss": 4.370555147204195, + "tokens_seen": 204931072 + }, + { + "epoch": 0.06, + "learning_rate": 0.0007578896690070211, + "loss": 1.7628, + "theoretical_loss": 4.370382079596092, + "tokens_seen": 204996608 + }, + { + "epoch": 0.06, + "learning_rate": 0.0007578736208625878, + "loss": 1.8865, + "theoretical_loss": 4.370209082794006, + "tokens_seen": 205062144 + }, + { + "epoch": 0.06, + "learning_rate": 0.0007578575727181545, + "loss": 1.7702, + "theoretical_loss": 4.370036156746357, + "tokens_seen": 205127680 + }, + { + "epoch": 0.06, + "learning_rate": 0.0007578415245737211, + "loss": 1.6653, + "theoretical_loss": 4.369863301401612, + "tokens_seen": 205193216 + }, + { + "epoch": 0.06, + "learning_rate": 0.0007578254764292879, + "loss": 1.9384, + "theoretical_loss": 4.369690516708298, + "tokens_seen": 205258752 + }, + { + "epoch": 0.06, + "learning_rate": 0.0007578094282848545, + "loss": 1.7648, + "theoretical_loss": 4.3695178026149915, + "tokens_seen": 205324288 + }, + { + "epoch": 0.06, + "learning_rate": 0.0007577933801404213, + "loss": 1.7348, + "theoretical_loss": 4.369345159070327, + "tokens_seen": 205389824 + }, + { + "epoch": 0.06, + "learning_rate": 0.0007577773319959879, + "loss": 1.6668, + "theoretical_loss": 4.369172586022989, + "tokens_seen": 205455360 + }, + { + "epoch": 0.06, + "learning_rate": 0.0007577612838515547, + "loss": 1.8132, + "theoretical_loss": 4.369000083421716, + "tokens_seen": 205520896 + }, + { + "epoch": 0.06, + "learning_rate": 0.0007577452357071213, + "loss": 1.7406, + "theoretical_loss": 4.368827651215304, + "tokens_seen": 205586432 + }, + { + "epoch": 0.06, + "learning_rate": 0.0007577291875626881, + "loss": 1.7132, + "theoretical_loss": 4.368655289352598, + "tokens_seen": 205651968 + }, + { + "epoch": 0.06, + "learning_rate": 0.0007577131394182547, + "loss": 1.769, + "theoretical_loss": 4.3684829977824995, + "tokens_seen": 205717504 + }, + { + "epoch": 0.06, + "learning_rate": 0.0007576970912738215, + "loss": 1.6695, + "theoretical_loss": 4.368310776453961, + "tokens_seen": 205783040 + }, + { + "epoch": 0.06, + "learning_rate": 0.0007576810431293881, + "loss": 1.7146, + "theoretical_loss": 4.368138625315989, + "tokens_seen": 205848576 + }, + { + "epoch": 0.06, + "learning_rate": 0.0007576649949849549, + "loss": 1.6884, + "theoretical_loss": 4.367966544317646, + "tokens_seen": 205914112 + }, + { + "epoch": 0.06, + "learning_rate": 0.0007576489468405215, + "loss": 1.7859, + "theoretical_loss": 4.367794533408043, + "tokens_seen": 205979648 + }, + { + "epoch": 0.06, + "learning_rate": 0.0007576328986960883, + "loss": 1.6507, + "theoretical_loss": 4.367622592536348, + "tokens_seen": 206045184 + }, + { + "epoch": 0.06, + "learning_rate": 0.000757616850551655, + "loss": 1.7482, + "theoretical_loss": 4.36745072165178, + "tokens_seen": 206110720 + }, + { + "epoch": 0.06, + "learning_rate": 0.0007576008024072217, + "loss": 1.7771, + "theoretical_loss": 4.36727892070361, + "tokens_seen": 206176256 + }, + { + "epoch": 0.06, + "learning_rate": 0.0007575847542627884, + "loss": 1.7471, + "theoretical_loss": 4.367107189641167, + "tokens_seen": 206241792 + }, + { + "epoch": 0.06, + "learning_rate": 0.0007575687061183551, + "loss": 1.7526, + "theoretical_loss": 4.3669355284138245, + "tokens_seen": 206307328 + }, + { + "epoch": 0.06, + "learning_rate": 0.0007575526579739218, + "loss": 1.8058, + "theoretical_loss": 4.366763936971017, + "tokens_seen": 206372864 + }, + { + "epoch": 0.06, + "objective/train/docs_used": 81724, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.7351741790771484, + "objective/train/theoretical_loss": 4.366592415262226, + "objective/train/tokens_used": 226898400, + "theoretical_loss": 4.366592415262226, + "tokens_seen": 206438400 + }, + { + "epoch": 0.06, + "learning_rate": 0.0007575366098294885, + "loss": 1.7357, + "theoretical_loss": 4.366592415262226, + "tokens_seen": 206438400 + }, + { + "epoch": 0.06, + "learning_rate": 0.0007575205616850552, + "loss": 1.8134, + "theoretical_loss": 4.366420963236989, + "tokens_seen": 206503936 + }, + { + "epoch": 0.06, + "learning_rate": 0.0007575045135406219, + "loss": 1.884, + "theoretical_loss": 4.366249580844894, + "tokens_seen": 206569472 + }, + { + "epoch": 0.06, + "learning_rate": 0.0007574884653961886, + "loss": 1.7999, + "theoretical_loss": 4.366078268035581, + "tokens_seen": 206635008 + }, + { + "epoch": 0.06, + "learning_rate": 0.0007574724172517553, + "loss": 1.7456, + "theoretical_loss": 4.365907024758744, + "tokens_seen": 206700544 + }, + { + "epoch": 0.06, + "learning_rate": 0.000757456369107322, + "loss": 1.874, + "theoretical_loss": 4.365735850964131, + "tokens_seen": 206766080 + }, + { + "epoch": 0.06, + "learning_rate": 0.0007574403209628887, + "loss": 1.7, + "theoretical_loss": 4.365564746601537, + "tokens_seen": 206831616 + }, + { + "epoch": 0.06, + "learning_rate": 0.0007574242728184554, + "loss": 1.8957, + "theoretical_loss": 4.365393711620814, + "tokens_seen": 206897152 + }, + { + "epoch": 0.06, + "learning_rate": 0.0007574082246740221, + "loss": 1.7647, + "theoretical_loss": 4.365222745971863, + "tokens_seen": 206962688 + }, + { + "epoch": 0.06, + "learning_rate": 0.0007573921765295888, + "loss": 1.5167, + "theoretical_loss": 4.365051849604638, + "tokens_seen": 207028224 + }, + { + "epoch": 0.06, + "learning_rate": 0.0007573761283851555, + "loss": 1.7862, + "theoretical_loss": 4.364881022469148, + "tokens_seen": 207093760 + }, + { + "epoch": 0.06, + "learning_rate": 0.0007573600802407222, + "loss": 1.84, + "theoretical_loss": 4.364710264515448, + "tokens_seen": 207159296 + }, + { + "epoch": 0.06, + "learning_rate": 0.0007573440320962889, + "loss": 1.6729, + "theoretical_loss": 4.36453957569365, + "tokens_seen": 207224832 + }, + { + "epoch": 0.06, + "learning_rate": 0.0007573279839518556, + "loss": 1.7352, + "theoretical_loss": 4.364368955953913, + "tokens_seen": 207290368 + }, + { + "epoch": 0.06, + "learning_rate": 0.0007573119358074223, + "loss": 1.6955, + "theoretical_loss": 4.364198405246453, + "tokens_seen": 207355904 + }, + { + "epoch": 0.06, + "learning_rate": 0.000757295887662989, + "loss": 1.7661, + "theoretical_loss": 4.364027923521535, + "tokens_seen": 207421440 + }, + { + "epoch": 0.06, + "learning_rate": 0.0007572798395185557, + "loss": 1.6879, + "theoretical_loss": 4.363857510729473, + "tokens_seen": 207486976 + }, + { + "epoch": 0.06, + "learning_rate": 0.0007572637913741224, + "loss": 1.8126, + "theoretical_loss": 4.363687166820637, + "tokens_seen": 207552512 + }, + { + "epoch": 0.06, + "learning_rate": 0.0007572477432296891, + "loss": 1.8704, + "theoretical_loss": 4.363516891745445, + "tokens_seen": 207618048 + }, + { + "epoch": 0.06, + "learning_rate": 0.0007572316950852558, + "loss": 1.9037, + "theoretical_loss": 4.36334668545437, + "tokens_seen": 207683584 + }, + { + "epoch": 0.06, + "learning_rate": 0.0007572156469408225, + "loss": 1.8215, + "theoretical_loss": 4.363176547897931, + "tokens_seen": 207749120 + }, + { + "epoch": 0.06, + "learning_rate": 0.0007571995987963892, + "loss": 1.8017, + "theoretical_loss": 4.363006479026702, + "tokens_seen": 207814656 + }, + { + "epoch": 0.06, + "learning_rate": 0.0007571835506519559, + "loss": 1.7931, + "theoretical_loss": 4.3628364787913085, + "tokens_seen": 207880192 + }, + { + "epoch": 0.06, + "learning_rate": 0.0007571675025075226, + "loss": 1.6724, + "theoretical_loss": 4.362666547142424, + "tokens_seen": 207945728 + }, + { + "epoch": 0.06, + "learning_rate": 0.0007571514543630893, + "loss": 1.582, + "theoretical_loss": 4.362496684030775, + "tokens_seen": 208011264 + }, + { + "epoch": 0.06, + "learning_rate": 0.000757135406218656, + "loss": 1.7175, + "theoretical_loss": 4.36232688940714, + "tokens_seen": 208076800 + }, + { + "epoch": 0.06, + "learning_rate": 0.0007571193580742227, + "loss": 1.8997, + "theoretical_loss": 4.362157163222346, + "tokens_seen": 208142336 + }, + { + "epoch": 0.06, + "learning_rate": 0.0007571033099297894, + "loss": 1.8028, + "theoretical_loss": 4.361987505427272, + "tokens_seen": 208207872 + }, + { + "epoch": 0.06, + "learning_rate": 0.0007570872617853561, + "loss": 1.8416, + "theoretical_loss": 4.361817915972846, + "tokens_seen": 208273408 + }, + { + "epoch": 0.06, + "learning_rate": 0.0007570712136409228, + "loss": 1.7811, + "theoretical_loss": 4.36164839481005, + "tokens_seen": 208338944 + }, + { + "epoch": 0.06, + "learning_rate": 0.0007570551654964895, + "loss": 1.924, + "theoretical_loss": 4.361478941889915, + "tokens_seen": 208404480 + }, + { + "epoch": 0.06, + "learning_rate": 0.0007570391173520562, + "loss": 1.8007, + "theoretical_loss": 4.36130955716352, + "tokens_seen": 208470016 + }, + { + "epoch": 0.06, + "learning_rate": 0.0007570230692076229, + "loss": 1.7394, + "theoretical_loss": 4.361140240581998, + "tokens_seen": 208535552 + }, + { + "epoch": 0.06, + "learning_rate": 0.0007570070210631896, + "loss": 1.7863, + "theoretical_loss": 4.36097099209653, + "tokens_seen": 208601088 + }, + { + "epoch": 0.06, + "learning_rate": 0.0007569909729187563, + "loss": 1.7842, + "theoretical_loss": 4.36080181165835, + "tokens_seen": 208666624 + }, + { + "epoch": 0.06, + "learning_rate": 0.000756974924774323, + "loss": 1.7018, + "theoretical_loss": 4.360632699218739, + "tokens_seen": 208732160 + }, + { + "epoch": 0.06, + "learning_rate": 0.0007569588766298897, + "loss": 1.7091, + "theoretical_loss": 4.360463654729029, + "tokens_seen": 208797696 + }, + { + "epoch": 0.06, + "learning_rate": 0.0007569428284854564, + "loss": 1.7777, + "theoretical_loss": 4.360294678140605, + "tokens_seen": 208863232 + }, + { + "epoch": 0.06, + "learning_rate": 0.0007569267803410231, + "loss": 1.7838, + "theoretical_loss": 4.360125769404897, + "tokens_seen": 208928768 + }, + { + "epoch": 0.06, + "learning_rate": 0.0007569107321965898, + "loss": 1.8651, + "theoretical_loss": 4.35995692847339, + "tokens_seen": 208994304 + }, + { + "epoch": 0.06, + "learning_rate": 0.0007568946840521565, + "loss": 1.684, + "theoretical_loss": 4.359788155297615, + "tokens_seen": 209059840 + }, + { + "epoch": 0.06, + "learning_rate": 0.0007568786359077232, + "loss": 1.7097, + "theoretical_loss": 4.359619449829156, + "tokens_seen": 209125376 + }, + { + "epoch": 0.06, + "learning_rate": 0.0007568625877632899, + "loss": 1.8362, + "theoretical_loss": 4.359450812019643, + "tokens_seen": 209190912 + }, + { + "epoch": 0.06, + "learning_rate": 0.0007568465396188566, + "loss": 1.8159, + "theoretical_loss": 4.359282241820761, + "tokens_seen": 209256448 + }, + { + "epoch": 0.06, + "learning_rate": 0.0007568304914744233, + "loss": 1.6521, + "theoretical_loss": 4.359113739184238, + "tokens_seen": 209321984 + }, + { + "epoch": 0.06, + "learning_rate": 0.00075681444332999, + "loss": 1.8363, + "theoretical_loss": 4.358945304061858, + "tokens_seen": 209387520 + }, + { + "epoch": 0.06, + "learning_rate": 0.0007567983951855567, + "loss": 1.7588, + "theoretical_loss": 4.35877693640545, + "tokens_seen": 209453056 + }, + { + "epoch": 0.06, + "learning_rate": 0.0007567823470411234, + "loss": 1.8708, + "theoretical_loss": 4.358608636166894, + "tokens_seen": 209518592 + }, + { + "epoch": 0.06, + "learning_rate": 0.0007567662988966901, + "loss": 1.8057, + "theoretical_loss": 4.358440403298119, + "tokens_seen": 209584128 + }, + { + "epoch": 0.06, + "learning_rate": 0.0007567502507522569, + "loss": 1.6918, + "theoretical_loss": 4.3582722377511045, + "tokens_seen": 209649664 + }, + { + "epoch": 0.06, + "objective/train/docs_used": 82917, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.6985247135162354, + "objective/train/theoretical_loss": 4.358104139477878, + "objective/train/tokens_used": 230175200, + "theoretical_loss": 4.358104139477878, + "tokens_seen": 209715200 + }, + { + "epoch": 0.06, + "learning_rate": 0.0007567342026078235, + "loss": 1.7597, + "theoretical_loss": 4.358104139477878, + "tokens_seen": 209715200 + }, + { + "epoch": 0.06, + "learning_rate": 0.0007567181544633903, + "loss": 1.7852, + "theoretical_loss": 4.357936108430517, + "tokens_seen": 209780736 + }, + { + "epoch": 0.06, + "learning_rate": 0.0007567021063189569, + "loss": 1.782, + "theoretical_loss": 4.357768144561146, + "tokens_seen": 209846272 + }, + { + "epoch": 0.06, + "learning_rate": 0.0007566860581745237, + "loss": 1.8194, + "theoretical_loss": 4.357600247821942, + "tokens_seen": 209911808 + }, + { + "epoch": 0.06, + "learning_rate": 0.0007566700100300903, + "loss": 1.7356, + "theoretical_loss": 4.3574324181651285, + "tokens_seen": 209977344 + }, + { + "epoch": 0.06, + "learning_rate": 0.0007566539618856571, + "loss": 1.7318, + "theoretical_loss": 4.357264655542978, + "tokens_seen": 210042880 + }, + { + "epoch": 0.06, + "learning_rate": 0.0007566379137412237, + "loss": 1.7166, + "theoretical_loss": 4.357096959907813, + "tokens_seen": 210108416 + }, + { + "epoch": 0.06, + "learning_rate": 0.0007566218655967905, + "loss": 1.6848, + "theoretical_loss": 4.356929331212004, + "tokens_seen": 210173952 + }, + { + "epoch": 0.06, + "learning_rate": 0.0007566058174523571, + "loss": 1.7012, + "theoretical_loss": 4.356761769407969, + "tokens_seen": 210239488 + }, + { + "epoch": 0.06, + "learning_rate": 0.0007565897693079239, + "loss": 1.8898, + "theoretical_loss": 4.356594274448178, + "tokens_seen": 210305024 + }, + { + "epoch": 0.06, + "learning_rate": 0.0007565737211634905, + "loss": 1.9007, + "theoretical_loss": 4.356426846285147, + "tokens_seen": 210370560 + }, + { + "epoch": 0.06, + "learning_rate": 0.0007565576730190572, + "loss": 1.7931, + "theoretical_loss": 4.3562594848714395, + "tokens_seen": 210436096 + }, + { + "epoch": 0.06, + "learning_rate": 0.000756541624874624, + "loss": 1.8175, + "theoretical_loss": 4.3560921901596705, + "tokens_seen": 210501632 + }, + { + "epoch": 0.06, + "learning_rate": 0.0007565255767301906, + "loss": 1.8792, + "theoretical_loss": 4.355924962102502, + "tokens_seen": 210567168 + }, + { + "epoch": 0.06, + "learning_rate": 0.0007565095285857573, + "loss": 1.8446, + "theoretical_loss": 4.3557578006526425, + "tokens_seen": 210632704 + }, + { + "epoch": 0.06, + "learning_rate": 0.000756493480441324, + "loss": 1.7133, + "theoretical_loss": 4.355590705762852, + "tokens_seen": 210698240 + }, + { + "epoch": 0.06, + "learning_rate": 0.0007564774322968907, + "loss": 1.7581, + "theoretical_loss": 4.355423677385937, + "tokens_seen": 210763776 + }, + { + "epoch": 0.06, + "learning_rate": 0.0007564613841524574, + "loss": 1.7343, + "theoretical_loss": 4.355256715474751, + "tokens_seen": 210829312 + }, + { + "epoch": 0.06, + "learning_rate": 0.0007564453360080242, + "loss": 1.8091, + "theoretical_loss": 4.355089819982198, + "tokens_seen": 210894848 + }, + { + "epoch": 0.06, + "learning_rate": 0.0007564292878635908, + "loss": 1.8179, + "theoretical_loss": 4.354922990861229, + "tokens_seen": 210960384 + }, + { + "epoch": 0.06, + "learning_rate": 0.0007564132397191574, + "loss": 1.7894, + "theoretical_loss": 4.354756228064842, + "tokens_seen": 211025920 + }, + { + "epoch": 0.06, + "learning_rate": 0.0007563971915747242, + "loss": 1.8495, + "theoretical_loss": 4.354589531546083, + "tokens_seen": 211091456 + }, + { + "epoch": 0.06, + "learning_rate": 0.0007563811434302908, + "loss": 1.7541, + "theoretical_loss": 4.354422901258048, + "tokens_seen": 211156992 + }, + { + "epoch": 0.06, + "learning_rate": 0.0007563650952858576, + "loss": 1.6377, + "theoretical_loss": 4.354256337153878, + "tokens_seen": 211222528 + }, + { + "epoch": 0.06, + "learning_rate": 0.0007563490471414242, + "loss": 1.73, + "theoretical_loss": 4.354089839186763, + "tokens_seen": 211288064 + }, + { + "epoch": 0.06, + "learning_rate": 0.000756332998996991, + "loss": 1.7677, + "theoretical_loss": 4.35392340730994, + "tokens_seen": 211353600 + }, + { + "epoch": 0.06, + "learning_rate": 0.0007563169508525576, + "loss": 1.7276, + "theoretical_loss": 4.353757041476695, + "tokens_seen": 211419136 + }, + { + "epoch": 0.06, + "learning_rate": 0.0007563009027081244, + "loss": 1.7996, + "theoretical_loss": 4.35359074164036, + "tokens_seen": 211484672 + }, + { + "epoch": 0.06, + "learning_rate": 0.000756284854563691, + "loss": 1.748, + "theoretical_loss": 4.3534245077543146, + "tokens_seen": 211550208 + }, + { + "epoch": 0.06, + "learning_rate": 0.0007562688064192578, + "loss": 1.8376, + "theoretical_loss": 4.353258339771986, + "tokens_seen": 211615744 + }, + { + "epoch": 0.06, + "learning_rate": 0.0007562527582748245, + "loss": 1.7172, + "theoretical_loss": 4.353092237646849, + "tokens_seen": 211681280 + }, + { + "epoch": 0.06, + "learning_rate": 0.0007562367101303912, + "loss": 1.7407, + "theoretical_loss": 4.3529262013324255, + "tokens_seen": 211746816 + }, + { + "epoch": 0.06, + "learning_rate": 0.0007562206619859579, + "loss": 1.7103, + "theoretical_loss": 4.352760230782284, + "tokens_seen": 211812352 + }, + { + "epoch": 0.06, + "learning_rate": 0.0007562046138415246, + "loss": 1.7437, + "theoretical_loss": 4.352594325950041, + "tokens_seen": 211877888 + }, + { + "epoch": 0.06, + "learning_rate": 0.0007561885656970913, + "loss": 1.763, + "theoretical_loss": 4.35242848678936, + "tokens_seen": 211943424 + }, + { + "epoch": 0.06, + "learning_rate": 0.000756172517552658, + "loss": 1.7989, + "theoretical_loss": 4.35226271325395, + "tokens_seen": 212008960 + }, + { + "epoch": 0.06, + "learning_rate": 0.0007561564694082247, + "loss": 1.8709, + "theoretical_loss": 4.352097005297569, + "tokens_seen": 212074496 + }, + { + "epoch": 0.06, + "learning_rate": 0.0007561404212637914, + "loss": 1.7125, + "theoretical_loss": 4.351931362874019, + "tokens_seen": 212140032 + }, + { + "epoch": 0.06, + "learning_rate": 0.0007561243731193581, + "loss": 1.7899, + "theoretical_loss": 4.351765785937154, + "tokens_seen": 212205568 + }, + { + "epoch": 0.06, + "learning_rate": 0.0007561083249749248, + "loss": 1.8153, + "theoretical_loss": 4.3516002744408695, + "tokens_seen": 212271104 + }, + { + "epoch": 0.06, + "learning_rate": 0.0007560922768304915, + "loss": 1.774, + "theoretical_loss": 4.351434828339109, + "tokens_seen": 212336640 + }, + { + "epoch": 0.06, + "learning_rate": 0.0007560762286860582, + "loss": 1.8251, + "theoretical_loss": 4.351269447585864, + "tokens_seen": 212402176 + }, + { + "epoch": 0.06, + "learning_rate": 0.0007560601805416249, + "loss": 1.839, + "theoretical_loss": 4.3511041321351716, + "tokens_seen": 212467712 + }, + { + "epoch": 0.06, + "learning_rate": 0.0007560441323971916, + "loss": 1.7667, + "theoretical_loss": 4.350938881941116, + "tokens_seen": 212533248 + }, + { + "epoch": 0.06, + "learning_rate": 0.0007560280842527583, + "loss": 1.7558, + "theoretical_loss": 4.350773696957827, + "tokens_seen": 212598784 + }, + { + "epoch": 0.06, + "learning_rate": 0.000756012036108325, + "loss": 1.8558, + "theoretical_loss": 4.350608577139481, + "tokens_seen": 212664320 + }, + { + "epoch": 0.06, + "learning_rate": 0.0007559959879638917, + "loss": 1.5722, + "theoretical_loss": 4.350443522440301, + "tokens_seen": 212729856 + }, + { + "epoch": 0.06, + "learning_rate": 0.0007559799398194584, + "loss": 1.8203, + "theoretical_loss": 4.350278532814556, + "tokens_seen": 212795392 + }, + { + "epoch": 0.06, + "learning_rate": 0.0007559638916750251, + "loss": 1.7096, + "theoretical_loss": 4.350113608216563, + "tokens_seen": 212860928 + }, + { + "epoch": 0.06, + "learning_rate": 0.0007559478435305918, + "loss": 1.8002, + "theoretical_loss": 4.3499487486006805, + "tokens_seen": 212926464 + }, + { + "epoch": 0.06, + "objective/train/docs_used": 84140, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.614612102508545, + "objective/train/theoretical_loss": 4.349783953921318, + "objective/train/tokens_used": 233452000, + "theoretical_loss": 4.349783953921318, + "tokens_seen": 212992000 + }, + { + "epoch": 0.06, + "learning_rate": 0.0007559317953861585, + "loss": 1.6402, + "theoretical_loss": 4.349783953921318, + "tokens_seen": 212992000 + }, + { + "epoch": 0.06, + "learning_rate": 0.0007559157472417252, + "loss": 1.6882, + "theoretical_loss": 4.349619224132929, + "tokens_seen": 213057536 + }, + { + "epoch": 0.06, + "learning_rate": 0.0007558996990972919, + "loss": 1.704, + "theoretical_loss": 4.349454559190011, + "tokens_seen": 213123072 + }, + { + "epoch": 0.06, + "learning_rate": 0.0007558836509528586, + "loss": 1.7462, + "theoretical_loss": 4.349289959047112, + "tokens_seen": 213188608 + }, + { + "epoch": 0.06, + "learning_rate": 0.0007558676028084253, + "loss": 1.6969, + "theoretical_loss": 4.349125423658821, + "tokens_seen": 213254144 + }, + { + "epoch": 0.06, + "learning_rate": 0.000755851554663992, + "loss": 1.8129, + "theoretical_loss": 4.348960952979776, + "tokens_seen": 213319680 + }, + { + "epoch": 0.06, + "learning_rate": 0.0007558355065195587, + "loss": 1.753, + "theoretical_loss": 4.3487965469646594, + "tokens_seen": 213385216 + }, + { + "epoch": 0.06, + "learning_rate": 0.0007558194583751254, + "loss": 1.7683, + "theoretical_loss": 4.3486322055682, + "tokens_seen": 213450752 + }, + { + "epoch": 0.06, + "learning_rate": 0.0007558034102306921, + "loss": 1.7801, + "theoretical_loss": 4.348467928745171, + "tokens_seen": 213516288 + }, + { + "epoch": 0.06, + "learning_rate": 0.0007557873620862588, + "loss": 1.6653, + "theoretical_loss": 4.348303716450391, + "tokens_seen": 213581824 + }, + { + "epoch": 0.06, + "learning_rate": 0.0007557713139418255, + "loss": 1.818, + "theoretical_loss": 4.348139568638727, + "tokens_seen": 213647360 + }, + { + "epoch": 0.06, + "learning_rate": 0.0007557552657973922, + "loss": 1.7159, + "theoretical_loss": 4.347975485265087, + "tokens_seen": 213712896 + }, + { + "epoch": 0.06, + "learning_rate": 0.0007557392176529589, + "loss": 1.7684, + "theoretical_loss": 4.347811466284428, + "tokens_seen": 213778432 + }, + { + "epoch": 0.06, + "learning_rate": 0.0007557231695085256, + "loss": 1.7161, + "theoretical_loss": 4.34764751165175, + "tokens_seen": 213843968 + }, + { + "epoch": 0.06, + "learning_rate": 0.0007557071213640923, + "loss": 1.7049, + "theoretical_loss": 4.347483621322099, + "tokens_seen": 213909504 + }, + { + "epoch": 0.06, + "learning_rate": 0.000755691073219659, + "loss": 1.7617, + "theoretical_loss": 4.347319795250567, + "tokens_seen": 213975040 + }, + { + "epoch": 0.06, + "learning_rate": 0.0007556750250752257, + "loss": 1.7365, + "theoretical_loss": 4.347156033392291, + "tokens_seen": 214040576 + }, + { + "epoch": 0.06, + "learning_rate": 0.0007556589769307924, + "loss": 1.8124, + "theoretical_loss": 4.346992335702451, + "tokens_seen": 214106112 + }, + { + "epoch": 0.06, + "learning_rate": 0.0007556429287863591, + "loss": 1.7494, + "theoretical_loss": 4.346828702136273, + "tokens_seen": 214171648 + }, + { + "epoch": 0.06, + "learning_rate": 0.0007556268806419258, + "loss": 1.7668, + "theoretical_loss": 4.34666513264903, + "tokens_seen": 214237184 + }, + { + "epoch": 0.06, + "learning_rate": 0.0007556108324974925, + "loss": 1.7916, + "theoretical_loss": 4.346501627196037, + "tokens_seen": 214302720 + }, + { + "epoch": 0.06, + "learning_rate": 0.0007555947843530592, + "loss": 1.7943, + "theoretical_loss": 4.346338185732655, + "tokens_seen": 214368256 + }, + { + "epoch": 0.06, + "learning_rate": 0.0007555787362086259, + "loss": 1.7979, + "theoretical_loss": 4.3461748082142915, + "tokens_seen": 214433792 + }, + { + "epoch": 0.06, + "learning_rate": 0.0007555626880641926, + "loss": 1.8804, + "theoretical_loss": 4.346011494596395, + "tokens_seen": 214499328 + }, + { + "epoch": 0.07, + "learning_rate": 0.0007555466399197594, + "loss": 1.654, + "theoretical_loss": 4.34584824483446, + "tokens_seen": 214564864 + }, + { + "epoch": 0.07, + "learning_rate": 0.000755530591775326, + "loss": 1.8005, + "theoretical_loss": 4.345685058884029, + "tokens_seen": 214630400 + }, + { + "epoch": 0.07, + "learning_rate": 0.0007555145436308928, + "loss": 1.6943, + "theoretical_loss": 4.345521936700684, + "tokens_seen": 214695936 + }, + { + "epoch": 0.07, + "learning_rate": 0.0007554984954864594, + "loss": 1.8061, + "theoretical_loss": 4.345358878240053, + "tokens_seen": 214761472 + }, + { + "epoch": 0.07, + "learning_rate": 0.0007554824473420262, + "loss": 1.8901, + "theoretical_loss": 4.345195883457811, + "tokens_seen": 214827008 + }, + { + "epoch": 0.07, + "learning_rate": 0.0007554663991975928, + "loss": 1.8234, + "theoretical_loss": 4.3450329523096745, + "tokens_seen": 214892544 + }, + { + "epoch": 0.07, + "learning_rate": 0.0007554503510531596, + "loss": 1.8661, + "theoretical_loss": 4.344870084751406, + "tokens_seen": 214958080 + }, + { + "epoch": 0.07, + "learning_rate": 0.0007554343029087262, + "loss": 1.8713, + "theoretical_loss": 4.34470728073881, + "tokens_seen": 215023616 + }, + { + "epoch": 0.07, + "learning_rate": 0.000755418254764293, + "loss": 1.7635, + "theoretical_loss": 4.344544540227737, + "tokens_seen": 215089152 + }, + { + "epoch": 0.07, + "learning_rate": 0.0007554022066198596, + "loss": 1.7491, + "theoretical_loss": 4.344381863174082, + "tokens_seen": 215154688 + }, + { + "epoch": 0.07, + "learning_rate": 0.0007553861584754264, + "loss": 1.7912, + "theoretical_loss": 4.344219249533783, + "tokens_seen": 215220224 + }, + { + "epoch": 0.07, + "learning_rate": 0.000755370110330993, + "loss": 1.9059, + "theoretical_loss": 4.34405669926282, + "tokens_seen": 215285760 + }, + { + "epoch": 0.07, + "learning_rate": 0.0007553540621865598, + "loss": 1.7646, + "theoretical_loss": 4.343894212317224, + "tokens_seen": 215351296 + }, + { + "epoch": 0.07, + "learning_rate": 0.0007553380140421264, + "loss": 1.7619, + "theoretical_loss": 4.34373178865306, + "tokens_seen": 215416832 + }, + { + "epoch": 0.07, + "learning_rate": 0.0007553219658976932, + "loss": 1.6111, + "theoretical_loss": 4.343569428226445, + "tokens_seen": 215482368 + }, + { + "epoch": 0.07, + "learning_rate": 0.0007553059177532598, + "loss": 1.6482, + "theoretical_loss": 4.343407130993536, + "tokens_seen": 215547904 + }, + { + "epoch": 0.07, + "learning_rate": 0.0007552898696088266, + "loss": 1.8109, + "theoretical_loss": 4.343244896910534, + "tokens_seen": 215613440 + }, + { + "epoch": 0.07, + "learning_rate": 0.0007552738214643932, + "loss": 1.9087, + "theoretical_loss": 4.343082725933684, + "tokens_seen": 215678976 + }, + { + "epoch": 0.07, + "learning_rate": 0.00075525777331996, + "loss": 1.8279, + "theoretical_loss": 4.342920618019275, + "tokens_seen": 215744512 + }, + { + "epoch": 0.07, + "learning_rate": 0.0007552417251755266, + "loss": 1.6458, + "theoretical_loss": 4.342758573123639, + "tokens_seen": 215810048 + }, + { + "epoch": 0.07, + "learning_rate": 0.0007552256770310934, + "loss": 1.8179, + "theoretical_loss": 4.342596591203153, + "tokens_seen": 215875584 + }, + { + "epoch": 0.07, + "learning_rate": 0.00075520962888666, + "loss": 1.6439, + "theoretical_loss": 4.3424346722142335, + "tokens_seen": 215941120 + }, + { + "epoch": 0.07, + "learning_rate": 0.0007551935807422268, + "loss": 1.7023, + "theoretical_loss": 4.342272816113344, + "tokens_seen": 216006656 + }, + { + "epoch": 0.07, + "learning_rate": 0.0007551775325977934, + "loss": 1.8086, + "theoretical_loss": 4.342111022856992, + "tokens_seen": 216072192 + }, + { + "epoch": 0.07, + "learning_rate": 0.0007551614844533601, + "loss": 1.8378, + "theoretical_loss": 4.341949292401725, + "tokens_seen": 216137728 + }, + { + "epoch": 0.07, + "learning_rate": 0.0007551454363089268, + "loss": 1.751, + "theoretical_loss": 4.341787624704135, + "tokens_seen": 216203264 + }, + { + "epoch": 0.07, + "objective/train/docs_used": 85191, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.804856777191162, + "objective/train/theoretical_loss": 4.341626019720858, + "objective/train/tokens_used": 236728800, + "theoretical_loss": 4.341626019720858, + "tokens_seen": 216268800 + }, + { + "epoch": 0.07, + "learning_rate": 0.0007551293881644935, + "loss": 1.728, + "theoretical_loss": 4.341626019720858, + "tokens_seen": 216268800 + }, + { + "epoch": 0.07, + "learning_rate": 0.0007551133400200602, + "loss": 1.7714, + "theoretical_loss": 4.341464477408572, + "tokens_seen": 216334336 + }, + { + "epoch": 0.07, + "learning_rate": 0.0007550972918756269, + "loss": 1.764, + "theoretical_loss": 4.341302997724, + "tokens_seen": 216399872 + }, + { + "epoch": 0.07, + "learning_rate": 0.0007550812437311937, + "loss": 1.8985, + "theoretical_loss": 4.341141580623905, + "tokens_seen": 216465408 + }, + { + "epoch": 0.07, + "learning_rate": 0.0007550651955867603, + "loss": 1.755, + "theoretical_loss": 4.340980226065096, + "tokens_seen": 216530944 + }, + { + "epoch": 0.07, + "learning_rate": 0.000755049147442327, + "loss": 1.8106, + "theoretical_loss": 4.340818934004421, + "tokens_seen": 216596480 + }, + { + "epoch": 0.07, + "learning_rate": 0.0007550330992978937, + "loss": 1.6965, + "theoretical_loss": 4.340657704398774, + "tokens_seen": 216662016 + }, + { + "epoch": 0.07, + "learning_rate": 0.0007550170511534603, + "loss": 1.8092, + "theoretical_loss": 4.340496537205093, + "tokens_seen": 216727552 + }, + { + "epoch": 0.07, + "learning_rate": 0.0007550010030090271, + "loss": 1.7295, + "theoretical_loss": 4.340335432380353, + "tokens_seen": 216793088 + }, + { + "epoch": 0.07, + "learning_rate": 0.0007549849548645937, + "loss": 1.8217, + "theoretical_loss": 4.340174389881579, + "tokens_seen": 216858624 + }, + { + "epoch": 0.07, + "learning_rate": 0.0007549689067201605, + "loss": 1.8343, + "theoretical_loss": 4.340013409665833, + "tokens_seen": 216924160 + }, + { + "epoch": 0.07, + "learning_rate": 0.0007549528585757271, + "loss": 1.6745, + "theoretical_loss": 4.339852491690221, + "tokens_seen": 216989696 + }, + { + "epoch": 0.07, + "learning_rate": 0.0007549368104312939, + "loss": 1.7644, + "theoretical_loss": 4.339691635911892, + "tokens_seen": 217055232 + }, + { + "epoch": 0.07, + "learning_rate": 0.0007549207622868605, + "loss": 1.757, + "theoretical_loss": 4.339530842288037, + "tokens_seen": 217120768 + }, + { + "epoch": 0.07, + "learning_rate": 0.0007549047141424273, + "loss": 1.7117, + "theoretical_loss": 4.339370110775892, + "tokens_seen": 217186304 + }, + { + "epoch": 0.07, + "learning_rate": 0.000754888665997994, + "loss": 1.7265, + "theoretical_loss": 4.33920944133273, + "tokens_seen": 217251840 + }, + { + "epoch": 0.07, + "learning_rate": 0.0007548726178535607, + "loss": 1.8065, + "theoretical_loss": 4.3390488339158715, + "tokens_seen": 217317376 + }, + { + "epoch": 0.07, + "learning_rate": 0.0007548565697091274, + "loss": 1.797, + "theoretical_loss": 4.338888288482675, + "tokens_seen": 217382912 + }, + { + "epoch": 0.07, + "learning_rate": 0.0007548405215646941, + "loss": 1.743, + "theoretical_loss": 4.338727804990544, + "tokens_seen": 217448448 + }, + { + "epoch": 0.07, + "learning_rate": 0.0007548244734202608, + "loss": 1.7099, + "theoretical_loss": 4.338567383396924, + "tokens_seen": 217513984 + }, + { + "epoch": 0.07, + "learning_rate": 0.0007548084252758275, + "loss": 1.7778, + "theoretical_loss": 4.338407023659302, + "tokens_seen": 217579520 + }, + { + "epoch": 0.07, + "learning_rate": 0.0007547923771313942, + "loss": 1.5976, + "theoretical_loss": 4.338246725735205, + "tokens_seen": 217645056 + }, + { + "epoch": 0.07, + "learning_rate": 0.0007547763289869609, + "loss": 1.7478, + "theoretical_loss": 4.338086489582205, + "tokens_seen": 217710592 + }, + { + "epoch": 0.07, + "learning_rate": 0.0007547602808425276, + "loss": 1.7006, + "theoretical_loss": 4.3379263151579135, + "tokens_seen": 217776128 + }, + { + "epoch": 0.07, + "learning_rate": 0.0007547442326980943, + "loss": 1.6422, + "theoretical_loss": 4.337766202419986, + "tokens_seen": 217841664 + }, + { + "epoch": 0.07, + "learning_rate": 0.000754728184553661, + "loss": 1.7968, + "theoretical_loss": 4.337606151326119, + "tokens_seen": 217907200 + }, + { + "epoch": 0.07, + "learning_rate": 0.0007547121364092277, + "loss": 1.7628, + "theoretical_loss": 4.337446161834048, + "tokens_seen": 217972736 + }, + { + "epoch": 0.07, + "learning_rate": 0.0007546960882647944, + "loss": 1.8332, + "theoretical_loss": 4.3372862339015565, + "tokens_seen": 218038272 + }, + { + "epoch": 0.07, + "learning_rate": 0.0007546800401203611, + "loss": 1.678, + "theoretical_loss": 4.337126367486462, + "tokens_seen": 218103808 + }, + { + "epoch": 0.07, + "learning_rate": 0.0007546639919759278, + "loss": 1.7653, + "theoretical_loss": 4.33696656254663, + "tokens_seen": 218169344 + }, + { + "epoch": 0.07, + "learning_rate": 0.0007546479438314945, + "loss": 1.7818, + "theoretical_loss": 4.336806819039962, + "tokens_seen": 218234880 + }, + { + "epoch": 0.07, + "learning_rate": 0.0007546318956870612, + "loss": 1.5455, + "theoretical_loss": 4.336647136924407, + "tokens_seen": 218300416 + }, + { + "epoch": 0.07, + "learning_rate": 0.0007546158475426279, + "loss": 1.6853, + "theoretical_loss": 4.33648751615795, + "tokens_seen": 218365952 + }, + { + "epoch": 0.07, + "learning_rate": 0.0007545997993981946, + "loss": 1.7554, + "theoretical_loss": 4.33632795669862, + "tokens_seen": 218431488 + }, + { + "epoch": 0.07, + "learning_rate": 0.0007545837512537613, + "loss": 1.7851, + "theoretical_loss": 4.336168458504486, + "tokens_seen": 218497024 + }, + { + "epoch": 0.07, + "learning_rate": 0.000754567703109328, + "loss": 1.7715, + "theoretical_loss": 4.336009021533661, + "tokens_seen": 218562560 + }, + { + "epoch": 0.07, + "learning_rate": 0.0007545516549648947, + "loss": 1.8933, + "theoretical_loss": 4.335849645744295, + "tokens_seen": 218628096 + }, + { + "epoch": 0.07, + "learning_rate": 0.0007545356068204614, + "loss": 1.7329, + "theoretical_loss": 4.3356903310945825, + "tokens_seen": 218693632 + }, + { + "epoch": 0.07, + "learning_rate": 0.0007545195586760281, + "loss": 1.737, + "theoretical_loss": 4.335531077542758, + "tokens_seen": 218759168 + }, + { + "epoch": 0.07, + "learning_rate": 0.0007545035105315948, + "loss": 1.7477, + "theoretical_loss": 4.335371885047096, + "tokens_seen": 218824704 + }, + { + "epoch": 0.07, + "learning_rate": 0.0007544874623871615, + "loss": 1.697, + "theoretical_loss": 4.335212753565914, + "tokens_seen": 218890240 + }, + { + "epoch": 0.07, + "learning_rate": 0.0007544714142427282, + "loss": 1.8001, + "theoretical_loss": 4.33505368305757, + "tokens_seen": 218955776 + }, + { + "epoch": 0.07, + "learning_rate": 0.0007544553660982949, + "loss": 1.7792, + "theoretical_loss": 4.33489467348046, + "tokens_seen": 219021312 + }, + { + "epoch": 0.07, + "learning_rate": 0.0007544393179538616, + "loss": 1.6536, + "theoretical_loss": 4.334735724793026, + "tokens_seen": 219086848 + }, + { + "epoch": 0.07, + "learning_rate": 0.0007544232698094283, + "loss": 1.8245, + "theoretical_loss": 4.334576836953744, + "tokens_seen": 219152384 + }, + { + "epoch": 0.07, + "learning_rate": 0.000754407221664995, + "loss": 1.7712, + "theoretical_loss": 4.334418009921139, + "tokens_seen": 219217920 + }, + { + "epoch": 0.07, + "learning_rate": 0.0007543911735205617, + "loss": 1.7243, + "theoretical_loss": 4.334259243653769, + "tokens_seen": 219283456 + }, + { + "epoch": 0.07, + "learning_rate": 0.0007543751253761284, + "loss": 1.8496, + "theoretical_loss": 4.334100538110237, + "tokens_seen": 219348992 + }, + { + "epoch": 0.07, + "learning_rate": 0.0007543590772316951, + "loss": 1.8169, + "theoretical_loss": 4.333941893249184, + "tokens_seen": 219414528 + }, + { + "epoch": 0.07, + "learning_rate": 0.0007543430290872619, + "loss": 1.7389, + "theoretical_loss": 4.333783309029296, + "tokens_seen": 219480064 + }, + { + "epoch": 0.07, + "objective/train/docs_used": 86440, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.684678554534912, + "objective/train/theoretical_loss": 4.3336247854092935, + "objective/train/tokens_used": 240005600, + "theoretical_loss": 4.3336247854092935, + "tokens_seen": 219545600 + }, + { + "epoch": 0.07, + "learning_rate": 0.0007543269809428285, + "loss": 1.6927, + "theoretical_loss": 4.3336247854092935, + "tokens_seen": 219545600 + }, + { + "epoch": 0.07, + "learning_rate": 0.0007543109327983953, + "loss": 1.6785, + "theoretical_loss": 4.333466322347942, + "tokens_seen": 219611136 + }, + { + "epoch": 0.07, + "learning_rate": 0.0007542948846539619, + "loss": 1.8423, + "theoretical_loss": 4.333307919804045, + "tokens_seen": 219676672 + }, + { + "epoch": 0.07, + "learning_rate": 0.0007542788365095287, + "loss": 1.8626, + "theoretical_loss": 4.333149577736446, + "tokens_seen": 219742208 + }, + { + "epoch": 0.07, + "learning_rate": 0.0007542627883650953, + "loss": 1.8314, + "theoretical_loss": 4.332991296104032, + "tokens_seen": 219807744 + }, + { + "epoch": 0.07, + "learning_rate": 0.0007542467402206621, + "loss": 1.7807, + "theoretical_loss": 4.332833074865725, + "tokens_seen": 219873280 + }, + { + "epoch": 0.07, + "learning_rate": 0.0007542306920762287, + "loss": 1.7636, + "theoretical_loss": 4.332674913980492, + "tokens_seen": 219938816 + }, + { + "epoch": 0.07, + "learning_rate": 0.0007542146439317955, + "loss": 1.803, + "theoretical_loss": 4.332516813407337, + "tokens_seen": 220004352 + }, + { + "epoch": 0.07, + "learning_rate": 0.0007541985957873621, + "loss": 1.6885, + "theoretical_loss": 4.332358773105307, + "tokens_seen": 220069888 + }, + { + "epoch": 0.07, + "learning_rate": 0.0007541825476429289, + "loss": 1.7307, + "theoretical_loss": 4.332200793033485, + "tokens_seen": 220135424 + }, + { + "epoch": 0.07, + "learning_rate": 0.0007541664994984955, + "loss": 1.6678, + "theoretical_loss": 4.332042873150997, + "tokens_seen": 220200960 + }, + { + "epoch": 0.07, + "learning_rate": 0.0007541504513540623, + "loss": 1.7824, + "theoretical_loss": 4.3318850134170095, + "tokens_seen": 220266496 + }, + { + "epoch": 0.07, + "learning_rate": 0.0007541344032096289, + "loss": 1.7738, + "theoretical_loss": 4.331727213790725, + "tokens_seen": 220332032 + }, + { + "epoch": 0.07, + "learning_rate": 0.0007541183550651957, + "loss": 1.8076, + "theoretical_loss": 4.331569474231388, + "tokens_seen": 220397568 + }, + { + "epoch": 0.07, + "learning_rate": 0.0007541023069207623, + "loss": 1.8185, + "theoretical_loss": 4.331411794698286, + "tokens_seen": 220463104 + }, + { + "epoch": 0.07, + "learning_rate": 0.0007540862587763291, + "loss": 1.6996, + "theoretical_loss": 4.33125417515074, + "tokens_seen": 220528640 + }, + { + "epoch": 0.07, + "learning_rate": 0.0007540702106318957, + "loss": 1.7716, + "theoretical_loss": 4.331096615548115, + "tokens_seen": 220594176 + }, + { + "epoch": 0.07, + "learning_rate": 0.0007540541624874625, + "loss": 1.807, + "theoretical_loss": 4.330939115849814, + "tokens_seen": 220659712 + }, + { + "epoch": 0.07, + "learning_rate": 0.0007540381143430291, + "loss": 1.7423, + "theoretical_loss": 4.33078167601528, + "tokens_seen": 220725248 + }, + { + "epoch": 0.07, + "learning_rate": 0.0007540220661985959, + "loss": 1.8205, + "theoretical_loss": 4.330624296003996, + "tokens_seen": 220790784 + }, + { + "epoch": 0.07, + "learning_rate": 0.0007540060180541625, + "loss": 1.7504, + "theoretical_loss": 4.330466975775483, + "tokens_seen": 220856320 + }, + { + "epoch": 0.07, + "learning_rate": 0.0007539899699097293, + "loss": 1.8433, + "theoretical_loss": 4.330309715289301, + "tokens_seen": 220921856 + }, + { + "epoch": 0.07, + "learning_rate": 0.0007539739217652959, + "loss": 1.7203, + "theoretical_loss": 4.330152514505053, + "tokens_seen": 220987392 + }, + { + "epoch": 0.07, + "learning_rate": 0.0007539578736208627, + "loss": 1.6328, + "theoretical_loss": 4.329995373382378, + "tokens_seen": 221052928 + }, + { + "epoch": 0.07, + "learning_rate": 0.0007539418254764293, + "loss": 1.7278, + "theoretical_loss": 4.329838291880954, + "tokens_seen": 221118464 + }, + { + "epoch": 0.07, + "learning_rate": 0.0007539257773319961, + "loss": 1.856, + "theoretical_loss": 4.329681269960499, + "tokens_seen": 221184000 + }, + { + "epoch": 0.07, + "learning_rate": 0.0007539097291875627, + "loss": 1.7617, + "theoretical_loss": 4.329524307580773, + "tokens_seen": 221249536 + }, + { + "epoch": 0.07, + "learning_rate": 0.0007538936810431295, + "loss": 1.5823, + "theoretical_loss": 4.3293674047015696, + "tokens_seen": 221315072 + }, + { + "epoch": 0.07, + "learning_rate": 0.0007538776328986961, + "loss": 1.726, + "theoretical_loss": 4.329210561282726, + "tokens_seen": 221380608 + }, + { + "epoch": 0.07, + "learning_rate": 0.0007538615847542629, + "loss": 1.6493, + "theoretical_loss": 4.329053777284115, + "tokens_seen": 221446144 + }, + { + "epoch": 0.07, + "learning_rate": 0.0007538455366098295, + "loss": 1.7107, + "theoretical_loss": 4.328897052665653, + "tokens_seen": 221511680 + }, + { + "epoch": 0.07, + "learning_rate": 0.0007538294884653963, + "loss": 1.8399, + "theoretical_loss": 4.328740387387289, + "tokens_seen": 221577216 + }, + { + "epoch": 0.07, + "learning_rate": 0.000753813440320963, + "loss": 1.6858, + "theoretical_loss": 4.328583781409016, + "tokens_seen": 221642752 + }, + { + "epoch": 0.07, + "learning_rate": 0.0007537973921765297, + "loss": 1.7403, + "theoretical_loss": 4.328427234690864, + "tokens_seen": 221708288 + }, + { + "epoch": 0.07, + "learning_rate": 0.0007537813440320963, + "loss": 1.8017, + "theoretical_loss": 4.328270747192901, + "tokens_seen": 221773824 + }, + { + "epoch": 0.07, + "learning_rate": 0.000753765295887663, + "loss": 1.7683, + "theoretical_loss": 4.328114318875236, + "tokens_seen": 221839360 + }, + { + "epoch": 0.07, + "learning_rate": 0.0007537492477432297, + "loss": 1.8922, + "theoretical_loss": 4.327957949698012, + "tokens_seen": 221904896 + }, + { + "epoch": 0.07, + "learning_rate": 0.0007537331995987964, + "loss": 1.7961, + "theoretical_loss": 4.327801639621417, + "tokens_seen": 221970432 + }, + { + "epoch": 0.07, + "learning_rate": 0.0007537171514543632, + "loss": 1.7909, + "theoretical_loss": 4.3276453886056725, + "tokens_seen": 222035968 + }, + { + "epoch": 0.07, + "learning_rate": 0.0007537011033099298, + "loss": 1.6604, + "theoretical_loss": 4.32748919661104, + "tokens_seen": 222101504 + }, + { + "epoch": 0.07, + "learning_rate": 0.0007536850551654966, + "loss": 1.7646, + "theoretical_loss": 4.327333063597821, + "tokens_seen": 222167040 + }, + { + "epoch": 0.07, + "learning_rate": 0.0007536690070210632, + "loss": 1.7109, + "theoretical_loss": 4.327176989526352, + "tokens_seen": 222232576 + }, + { + "epoch": 0.07, + "learning_rate": 0.00075365295887663, + "loss": 1.7177, + "theoretical_loss": 4.327020974357012, + "tokens_seen": 222298112 + }, + { + "epoch": 0.07, + "learning_rate": 0.0007536369107321966, + "loss": 1.8032, + "theoretical_loss": 4.326865018050215, + "tokens_seen": 222363648 + }, + { + "epoch": 0.07, + "learning_rate": 0.0007536208625877632, + "loss": 1.6964, + "theoretical_loss": 4.326709120566415, + "tokens_seen": 222429184 + }, + { + "epoch": 0.07, + "learning_rate": 0.00075360481444333, + "loss": 1.693, + "theoretical_loss": 4.326553281866103, + "tokens_seen": 222494720 + }, + { + "epoch": 0.07, + "learning_rate": 0.0007535887662988966, + "loss": 1.7181, + "theoretical_loss": 4.326397501909811, + "tokens_seen": 222560256 + }, + { + "epoch": 0.07, + "learning_rate": 0.0007535727181544634, + "loss": 1.6469, + "theoretical_loss": 4.326241780658104, + "tokens_seen": 222625792 + }, + { + "epoch": 0.07, + "learning_rate": 0.00075355667001003, + "loss": 1.7517, + "theoretical_loss": 4.32608611807159, + "tokens_seen": 222691328 + }, + { + "epoch": 0.07, + "learning_rate": 0.0007535406218655968, + "loss": 1.639, + "theoretical_loss": 4.325930514110912, + "tokens_seen": 222756864 + }, + { + "epoch": 0.07, + "objective/train/docs_used": 87601, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.5822908878326416, + "objective/train/theoretical_loss": 4.325774968736753, + "objective/train/tokens_used": 243282400, + "theoretical_loss": 4.325774968736753, + "tokens_seen": 222822400 + }, + { + "epoch": 0.07, + "learning_rate": 0.0007535245737211634, + "loss": 1.6214, + "theoretical_loss": 4.325774968736753, + "tokens_seen": 222822400 + }, + { + "epoch": 0.07, + "learning_rate": 0.0007535085255767302, + "loss": 1.8181, + "theoretical_loss": 4.325619481909833, + "tokens_seen": 222887936 + }, + { + "epoch": 0.07, + "learning_rate": 0.0007534924774322969, + "loss": 1.6323, + "theoretical_loss": 4.325464053590908, + "tokens_seen": 222953472 + }, + { + "epoch": 0.07, + "learning_rate": 0.0007534764292878636, + "loss": 1.7512, + "theoretical_loss": 4.325308683740776, + "tokens_seen": 223019008 + }, + { + "epoch": 0.07, + "learning_rate": 0.0007534603811434303, + "loss": 1.7819, + "theoretical_loss": 4.325153372320269, + "tokens_seen": 223084544 + }, + { + "epoch": 0.07, + "learning_rate": 0.000753444332998997, + "loss": 1.7618, + "theoretical_loss": 4.32499811929026, + "tokens_seen": 223150080 + }, + { + "epoch": 0.07, + "learning_rate": 0.0007534282848545637, + "loss": 1.8032, + "theoretical_loss": 4.324842924611655, + "tokens_seen": 223215616 + }, + { + "epoch": 0.07, + "learning_rate": 0.0007534122367101304, + "loss": 1.7487, + "theoretical_loss": 4.324687788245403, + "tokens_seen": 223281152 + }, + { + "epoch": 0.07, + "learning_rate": 0.0007533961885656971, + "loss": 1.7264, + "theoretical_loss": 4.3245327101524875, + "tokens_seen": 223346688 + }, + { + "epoch": 0.07, + "learning_rate": 0.0007533801404212638, + "loss": 1.7472, + "theoretical_loss": 4.324377690293929, + "tokens_seen": 223412224 + }, + { + "epoch": 0.07, + "learning_rate": 0.0007533640922768305, + "loss": 1.7813, + "theoretical_loss": 4.324222728630787, + "tokens_seen": 223477760 + }, + { + "epoch": 0.07, + "learning_rate": 0.0007533480441323972, + "loss": 1.7512, + "theoretical_loss": 4.324067825124159, + "tokens_seen": 223543296 + }, + { + "epoch": 0.07, + "learning_rate": 0.0007533319959879639, + "loss": 1.7201, + "theoretical_loss": 4.323912979735178, + "tokens_seen": 223608832 + }, + { + "epoch": 0.07, + "learning_rate": 0.0007533159478435306, + "loss": 1.8237, + "theoretical_loss": 4.323758192425016, + "tokens_seen": 223674368 + }, + { + "epoch": 0.07, + "learning_rate": 0.0007532998996990973, + "loss": 1.7116, + "theoretical_loss": 4.323603463154882, + "tokens_seen": 223739904 + }, + { + "epoch": 0.07, + "learning_rate": 0.000753283851554664, + "loss": 1.7584, + "theoretical_loss": 4.323448791886022, + "tokens_seen": 223805440 + }, + { + "epoch": 0.07, + "learning_rate": 0.0007532678034102307, + "loss": 1.8914, + "theoretical_loss": 4.323294178579717, + "tokens_seen": 223870976 + }, + { + "epoch": 0.07, + "learning_rate": 0.0007532517552657974, + "loss": 1.6915, + "theoretical_loss": 4.323139623197289, + "tokens_seen": 223936512 + }, + { + "epoch": 0.07, + "learning_rate": 0.0007532357071213641, + "loss": 1.8091, + "theoretical_loss": 4.322985125700096, + "tokens_seen": 224002048 + }, + { + "epoch": 0.07, + "learning_rate": 0.0007532196589769308, + "loss": 1.6857, + "theoretical_loss": 4.322830686049532, + "tokens_seen": 224067584 + }, + { + "epoch": 0.07, + "learning_rate": 0.0007532036108324975, + "loss": 1.8286, + "theoretical_loss": 4.322676304207027, + "tokens_seen": 224133120 + }, + { + "epoch": 0.07, + "learning_rate": 0.0007531875626880642, + "loss": 1.8001, + "theoretical_loss": 4.32252198013405, + "tokens_seen": 224198656 + }, + { + "epoch": 0.07, + "learning_rate": 0.0007531715145436309, + "loss": 1.7314, + "theoretical_loss": 4.32236771379211, + "tokens_seen": 224264192 + }, + { + "epoch": 0.07, + "learning_rate": 0.0007531554663991976, + "loss": 1.7468, + "theoretical_loss": 4.322213505142744, + "tokens_seen": 224329728 + }, + { + "epoch": 0.07, + "learning_rate": 0.0007531394182547644, + "loss": 1.7816, + "theoretical_loss": 4.322059354147535, + "tokens_seen": 224395264 + }, + { + "epoch": 0.07, + "learning_rate": 0.000753123370110331, + "loss": 1.6007, + "theoretical_loss": 4.321905260768097, + "tokens_seen": 224460800 + }, + { + "epoch": 0.07, + "learning_rate": 0.0007531073219658978, + "loss": 1.7441, + "theoretical_loss": 4.321751224966083, + "tokens_seen": 224526336 + }, + { + "epoch": 0.07, + "learning_rate": 0.0007530912738214644, + "loss": 1.6484, + "theoretical_loss": 4.321597246703183, + "tokens_seen": 224591872 + }, + { + "epoch": 0.07, + "learning_rate": 0.0007530752256770312, + "loss": 1.7817, + "theoretical_loss": 4.321443325941123, + "tokens_seen": 224657408 + }, + { + "epoch": 0.07, + "learning_rate": 0.0007530591775325978, + "loss": 1.733, + "theoretical_loss": 4.3212894626416665, + "tokens_seen": 224722944 + }, + { + "epoch": 0.07, + "learning_rate": 0.0007530431293881646, + "loss": 1.6511, + "theoretical_loss": 4.321135656766611, + "tokens_seen": 224788480 + }, + { + "epoch": 0.07, + "learning_rate": 0.0007530270812437312, + "loss": 1.714, + "theoretical_loss": 4.320981908277793, + "tokens_seen": 224854016 + }, + { + "epoch": 0.07, + "learning_rate": 0.000753011033099298, + "loss": 1.8376, + "theoretical_loss": 4.320828217137086, + "tokens_seen": 224919552 + }, + { + "epoch": 0.07, + "learning_rate": 0.0007529949849548646, + "loss": 1.8269, + "theoretical_loss": 4.320674583306396, + "tokens_seen": 224985088 + }, + { + "epoch": 0.07, + "learning_rate": 0.0007529789368104314, + "loss": 1.8585, + "theoretical_loss": 4.32052100674767, + "tokens_seen": 225050624 + }, + { + "epoch": 0.07, + "learning_rate": 0.000752962888665998, + "loss": 1.9087, + "theoretical_loss": 4.320367487422891, + "tokens_seen": 225116160 + }, + { + "epoch": 0.07, + "learning_rate": 0.0007529468405215648, + "loss": 1.7852, + "theoretical_loss": 4.3202140252940735, + "tokens_seen": 225181696 + }, + { + "epoch": 0.07, + "learning_rate": 0.0007529307923771314, + "loss": 1.7344, + "theoretical_loss": 4.3200606203232725, + "tokens_seen": 225247232 + }, + { + "epoch": 0.07, + "learning_rate": 0.0007529147442326982, + "loss": 1.7692, + "theoretical_loss": 4.319907272472579, + "tokens_seen": 225312768 + }, + { + "epoch": 0.07, + "learning_rate": 0.0007528986960882648, + "loss": 1.6488, + "theoretical_loss": 4.319753981704119, + "tokens_seen": 225378304 + }, + { + "epoch": 0.07, + "learning_rate": 0.0007528826479438316, + "loss": 1.7103, + "theoretical_loss": 4.319600747980056, + "tokens_seen": 225443840 + }, + { + "epoch": 0.07, + "learning_rate": 0.0007528665997993982, + "loss": 1.7511, + "theoretical_loss": 4.319447571262586, + "tokens_seen": 225509376 + }, + { + "epoch": 0.07, + "learning_rate": 0.000752850551654965, + "loss": 1.9236, + "theoretical_loss": 4.3192944515139455, + "tokens_seen": 225574912 + }, + { + "epoch": 0.07, + "learning_rate": 0.0007528345035105316, + "loss": 1.8423, + "theoretical_loss": 4.319141388696404, + "tokens_seen": 225640448 + }, + { + "epoch": 0.07, + "learning_rate": 0.0007528184553660984, + "loss": 1.6845, + "theoretical_loss": 4.318988382772269, + "tokens_seen": 225705984 + }, + { + "epoch": 0.07, + "learning_rate": 0.000752802407221665, + "loss": 1.7915, + "theoretical_loss": 4.318835433703883, + "tokens_seen": 225771520 + }, + { + "epoch": 0.07, + "learning_rate": 0.0007527863590772318, + "loss": 1.8489, + "theoretical_loss": 4.318682541453622, + "tokens_seen": 225837056 + }, + { + "epoch": 0.07, + "learning_rate": 0.0007527703109327984, + "loss": 1.6554, + "theoretical_loss": 4.318529705983903, + "tokens_seen": 225902592 + }, + { + "epoch": 0.07, + "learning_rate": 0.0007527542627883652, + "loss": 1.7774, + "theoretical_loss": 4.318376927257174, + "tokens_seen": 225968128 + }, + { + "epoch": 0.07, + "learning_rate": 0.0007527382146439318, + "loss": 1.7198, + "theoretical_loss": 4.3182242052359205, + "tokens_seen": 226033664 + }, + { + "epoch": 0.07, + "objective/train/docs_used": 88830, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.719179630279541, + "objective/train/theoretical_loss": 4.318071539882665, + "objective/train/tokens_used": 246559200, + "theoretical_loss": 4.318071539882665, + "tokens_seen": 226099200 + }, + { + "epoch": 0.07, + "learning_rate": 0.0007527221664994986, + "loss": 1.7115, + "theoretical_loss": 4.318071539882665, + "tokens_seen": 226099200 + }, + { + "epoch": 0.07, + "learning_rate": 0.0007527061183550652, + "loss": 1.8511, + "theoretical_loss": 4.317918931159963, + "tokens_seen": 226164736 + }, + { + "epoch": 0.07, + "learning_rate": 0.000752690070210632, + "loss": 1.6511, + "theoretical_loss": 4.317766379030407, + "tokens_seen": 226230272 + }, + { + "epoch": 0.07, + "learning_rate": 0.0007526740220661986, + "loss": 1.7335, + "theoretical_loss": 4.317613883456625, + "tokens_seen": 226295808 + }, + { + "epoch": 0.07, + "learning_rate": 0.0007526579739217654, + "loss": 1.6758, + "theoretical_loss": 4.317461444401281, + "tokens_seen": 226361344 + }, + { + "epoch": 0.07, + "learning_rate": 0.000752641925777332, + "loss": 1.7787, + "theoretical_loss": 4.317309061827073, + "tokens_seen": 226426880 + }, + { + "epoch": 0.07, + "learning_rate": 0.0007526258776328988, + "loss": 1.6209, + "theoretical_loss": 4.317156735696736, + "tokens_seen": 226492416 + }, + { + "epoch": 0.07, + "learning_rate": 0.0007526098294884654, + "loss": 1.6159, + "theoretical_loss": 4.3170044659730396, + "tokens_seen": 226557952 + }, + { + "epoch": 0.07, + "learning_rate": 0.0007525937813440322, + "loss": 1.7125, + "theoretical_loss": 4.316852252618788, + "tokens_seen": 226623488 + }, + { + "epoch": 0.07, + "learning_rate": 0.0007525777331995988, + "loss": 1.7883, + "theoretical_loss": 4.316700095596822, + "tokens_seen": 226689024 + }, + { + "epoch": 0.07, + "learning_rate": 0.0007525616850551656, + "loss": 1.7173, + "theoretical_loss": 4.3165479948700165, + "tokens_seen": 226754560 + }, + { + "epoch": 0.07, + "learning_rate": 0.0007525456369107322, + "loss": 1.6463, + "theoretical_loss": 4.316395950401283, + "tokens_seen": 226820096 + }, + { + "epoch": 0.07, + "learning_rate": 0.000752529588766299, + "loss": 1.7169, + "theoretical_loss": 4.316243962153566, + "tokens_seen": 226885632 + }, + { + "epoch": 0.07, + "learning_rate": 0.0007525135406218656, + "loss": 1.7038, + "theoretical_loss": 4.316092030089847, + "tokens_seen": 226951168 + }, + { + "epoch": 0.07, + "learning_rate": 0.0007524974924774324, + "loss": 1.7548, + "theoretical_loss": 4.315940154173143, + "tokens_seen": 227016704 + }, + { + "epoch": 0.07, + "learning_rate": 0.000752481444332999, + "loss": 1.6923, + "theoretical_loss": 4.315788334366504, + "tokens_seen": 227082240 + }, + { + "epoch": 0.07, + "learning_rate": 0.0007524653961885658, + "loss": 1.6762, + "theoretical_loss": 4.315636570633015, + "tokens_seen": 227147776 + }, + { + "epoch": 0.07, + "learning_rate": 0.0007524493480441324, + "loss": 1.6898, + "theoretical_loss": 4.315484862935799, + "tokens_seen": 227213312 + }, + { + "epoch": 0.07, + "learning_rate": 0.0007524332998996992, + "loss": 1.7882, + "theoretical_loss": 4.315333211238009, + "tokens_seen": 227278848 + }, + { + "epoch": 0.07, + "learning_rate": 0.0007524172517552658, + "loss": 1.6381, + "theoretical_loss": 4.315181615502838, + "tokens_seen": 227344384 + }, + { + "epoch": 0.07, + "learning_rate": 0.0007524012036108326, + "loss": 1.6593, + "theoretical_loss": 4.31503007569351, + "tokens_seen": 227409920 + }, + { + "epoch": 0.07, + "learning_rate": 0.0007523851554663992, + "loss": 1.7485, + "theoretical_loss": 4.314878591773287, + "tokens_seen": 227475456 + }, + { + "epoch": 0.07, + "learning_rate": 0.0007523691073219659, + "loss": 1.756, + "theoretical_loss": 4.31472716370546, + "tokens_seen": 227540992 + }, + { + "epoch": 0.07, + "learning_rate": 0.0007523530591775327, + "loss": 1.6357, + "theoretical_loss": 4.314575791453362, + "tokens_seen": 227606528 + }, + { + "epoch": 0.07, + "learning_rate": 0.0007523370110330993, + "loss": 1.7033, + "theoretical_loss": 4.314424474980355, + "tokens_seen": 227672064 + }, + { + "epoch": 0.07, + "learning_rate": 0.000752320962888666, + "loss": 1.7243, + "theoretical_loss": 4.31427321424984, + "tokens_seen": 227737600 + }, + { + "epoch": 0.07, + "learning_rate": 0.0007523049147442327, + "loss": 1.7646, + "theoretical_loss": 4.314122009225248, + "tokens_seen": 227803136 + }, + { + "epoch": 0.07, + "learning_rate": 0.0007522888665997995, + "loss": 1.7521, + "theoretical_loss": 4.313970859870048, + "tokens_seen": 227868672 + }, + { + "epoch": 0.07, + "learning_rate": 0.0007522728184553661, + "loss": 1.6705, + "theoretical_loss": 4.313819766147741, + "tokens_seen": 227934208 + }, + { + "epoch": 0.07, + "learning_rate": 0.0007522567703109329, + "loss": 1.7504, + "theoretical_loss": 4.313668728021866, + "tokens_seen": 227999744 + }, + { + "epoch": 0.07, + "learning_rate": 0.0007522407221664995, + "loss": 1.8794, + "theoretical_loss": 4.313517745455991, + "tokens_seen": 228065280 + }, + { + "epoch": 0.07, + "learning_rate": 0.0007522246740220661, + "loss": 1.6858, + "theoretical_loss": 4.313366818413724, + "tokens_seen": 228130816 + }, + { + "epoch": 0.07, + "learning_rate": 0.0007522086258776329, + "loss": 1.6833, + "theoretical_loss": 4.313215946858702, + "tokens_seen": 228196352 + }, + { + "epoch": 0.07, + "learning_rate": 0.0007521925777331995, + "loss": 1.7844, + "theoretical_loss": 4.313065130754601, + "tokens_seen": 228261888 + }, + { + "epoch": 0.07, + "learning_rate": 0.0007521765295887663, + "loss": 1.7973, + "theoretical_loss": 4.312914370065128, + "tokens_seen": 228327424 + }, + { + "epoch": 0.07, + "learning_rate": 0.000752160481444333, + "loss": 1.7555, + "theoretical_loss": 4.312763664754025, + "tokens_seen": 228392960 + }, + { + "epoch": 0.07, + "learning_rate": 0.0007521444332998997, + "loss": 1.723, + "theoretical_loss": 4.312613014785069, + "tokens_seen": 228458496 + }, + { + "epoch": 0.07, + "learning_rate": 0.0007521283851554664, + "loss": 1.8034, + "theoretical_loss": 4.3124624201220705, + "tokens_seen": 228524032 + }, + { + "epoch": 0.07, + "learning_rate": 0.0007521123370110331, + "loss": 1.7465, + "theoretical_loss": 4.312311880728873, + "tokens_seen": 228589568 + }, + { + "epoch": 0.07, + "learning_rate": 0.0007520962888665998, + "loss": 1.765, + "theoretical_loss": 4.312161396569356, + "tokens_seen": 228655104 + }, + { + "epoch": 0.07, + "learning_rate": 0.0007520802407221665, + "loss": 1.7734, + "theoretical_loss": 4.312010967607431, + "tokens_seen": 228720640 + }, + { + "epoch": 0.07, + "learning_rate": 0.0007520641925777332, + "loss": 1.7811, + "theoretical_loss": 4.311860593807045, + "tokens_seen": 228786176 + }, + { + "epoch": 0.07, + "learning_rate": 0.0007520481444332999, + "loss": 1.7429, + "theoretical_loss": 4.311710275132178, + "tokens_seen": 228851712 + }, + { + "epoch": 0.07, + "learning_rate": 0.0007520320962888666, + "loss": 1.7499, + "theoretical_loss": 4.3115600115468435, + "tokens_seen": 228917248 + }, + { + "epoch": 0.07, + "learning_rate": 0.0007520160481444333, + "loss": 1.7067, + "theoretical_loss": 4.311409803015089, + "tokens_seen": 228982784 + }, + { + "epoch": 0.07, + "learning_rate": 0.000752, + "loss": 1.8546, + "theoretical_loss": 4.3112596495009985, + "tokens_seen": 229048320 + }, + { + "epoch": 0.07, + "learning_rate": 0.0007519839518555667, + "loss": 1.7439, + "theoretical_loss": 4.311109550968684, + "tokens_seen": 229113856 + }, + { + "epoch": 0.07, + "learning_rate": 0.0007519679037111334, + "loss": 1.7964, + "theoretical_loss": 4.310959507382296, + "tokens_seen": 229179392 + }, + { + "epoch": 0.07, + "learning_rate": 0.0007519518555667001, + "loss": 1.726, + "theoretical_loss": 4.310809518706017, + "tokens_seen": 229244928 + }, + { + "epoch": 0.07, + "learning_rate": 0.0007519358074222669, + "loss": 1.7713, + "theoretical_loss": 4.310659584904062, + "tokens_seen": 229310464 + }, + { + "epoch": 0.07, + "objective/train/docs_used": 90061, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.8669874668121338, + "objective/train/theoretical_loss": 4.310509705940682, + "objective/train/tokens_used": 249836000, + "theoretical_loss": 4.310509705940682, + "tokens_seen": 229376000 + }, + { + "epoch": 0.07, + "learning_rate": 0.0007519197592778335, + "loss": 1.8015, + "theoretical_loss": 4.310509705940682, + "tokens_seen": 229376000 + }, + { + "epoch": 0.07, + "learning_rate": 0.0007519037111334003, + "loss": 1.6844, + "theoretical_loss": 4.310359881780159, + "tokens_seen": 229441536 + }, + { + "epoch": 0.07, + "learning_rate": 0.0007518876629889669, + "loss": 1.8372, + "theoretical_loss": 4.310210112386811, + "tokens_seen": 229507072 + }, + { + "epoch": 0.07, + "learning_rate": 0.0007518716148445337, + "loss": 1.8495, + "theoretical_loss": 4.310060397724987, + "tokens_seen": 229572608 + }, + { + "epoch": 0.07, + "learning_rate": 0.0007518555667001003, + "loss": 1.7922, + "theoretical_loss": 4.30991073775907, + "tokens_seen": 229638144 + }, + { + "epoch": 0.07, + "learning_rate": 0.0007518395185556671, + "loss": 1.7052, + "theoretical_loss": 4.3097611324534775, + "tokens_seen": 229703680 + }, + { + "epoch": 0.07, + "learning_rate": 0.0007518234704112337, + "loss": 1.8083, + "theoretical_loss": 4.309611581772659, + "tokens_seen": 229769216 + }, + { + "epoch": 0.07, + "learning_rate": 0.0007518074222668005, + "loss": 1.7807, + "theoretical_loss": 4.309462085681098, + "tokens_seen": 229834752 + }, + { + "epoch": 0.07, + "learning_rate": 0.0007517913741223671, + "loss": 1.853, + "theoretical_loss": 4.3093126441433105, + "tokens_seen": 229900288 + }, + { + "epoch": 0.07, + "learning_rate": 0.0007517753259779339, + "loss": 1.7527, + "theoretical_loss": 4.3091632571238465, + "tokens_seen": 229965824 + }, + { + "epoch": 0.07, + "learning_rate": 0.0007517592778335005, + "loss": 1.7977, + "theoretical_loss": 4.30901392458729, + "tokens_seen": 230031360 + }, + { + "epoch": 0.07, + "learning_rate": 0.0007517432296890673, + "loss": 1.7534, + "theoretical_loss": 4.308864646498254, + "tokens_seen": 230096896 + }, + { + "epoch": 0.07, + "learning_rate": 0.0007517271815446339, + "loss": 1.6965, + "theoretical_loss": 4.30871542282139, + "tokens_seen": 230162432 + }, + { + "epoch": 0.07, + "learning_rate": 0.0007517111334002007, + "loss": 1.5798, + "theoretical_loss": 4.308566253521378, + "tokens_seen": 230227968 + }, + { + "epoch": 0.07, + "learning_rate": 0.0007516950852557673, + "loss": 1.7182, + "theoretical_loss": 4.308417138562934, + "tokens_seen": 230293504 + }, + { + "epoch": 0.07, + "learning_rate": 0.0007516790371113341, + "loss": 1.626, + "theoretical_loss": 4.308268077910806, + "tokens_seen": 230359040 + }, + { + "epoch": 0.07, + "learning_rate": 0.0007516629889669007, + "loss": 1.5869, + "theoretical_loss": 4.308119071529774, + "tokens_seen": 230424576 + }, + { + "epoch": 0.07, + "learning_rate": 0.0007516469408224675, + "loss": 1.666, + "theoretical_loss": 4.3079701193846525, + "tokens_seen": 230490112 + }, + { + "epoch": 0.07, + "learning_rate": 0.0007516308926780341, + "loss": 1.9177, + "theoretical_loss": 4.307821221440286, + "tokens_seen": 230555648 + }, + { + "epoch": 0.07, + "learning_rate": 0.0007516148445336009, + "loss": 1.7043, + "theoretical_loss": 4.3076723776615555, + "tokens_seen": 230621184 + }, + { + "epoch": 0.07, + "learning_rate": 0.0007515987963891675, + "loss": 1.7679, + "theoretical_loss": 4.307523588013371, + "tokens_seen": 230686720 + }, + { + "epoch": 0.07, + "learning_rate": 0.0007515827482447343, + "loss": 1.7001, + "theoretical_loss": 4.307374852460679, + "tokens_seen": 230752256 + }, + { + "epoch": 0.07, + "learning_rate": 0.0007515667001003009, + "loss": 1.6727, + "theoretical_loss": 4.307226170968456, + "tokens_seen": 230817792 + }, + { + "epoch": 0.07, + "learning_rate": 0.0007515506519558677, + "loss": 1.6438, + "theoretical_loss": 4.307077543501713, + "tokens_seen": 230883328 + }, + { + "epoch": 0.07, + "learning_rate": 0.0007515346038114343, + "loss": 1.7717, + "theoretical_loss": 4.3069289700254885, + "tokens_seen": 230948864 + }, + { + "epoch": 0.07, + "learning_rate": 0.0007515185556670011, + "loss": 1.6717, + "theoretical_loss": 4.306780450504861, + "tokens_seen": 231014400 + }, + { + "epoch": 0.07, + "learning_rate": 0.0007515025075225677, + "loss": 1.7269, + "theoretical_loss": 4.3066319849049375, + "tokens_seen": 231079936 + }, + { + "epoch": 0.07, + "learning_rate": 0.0007514864593781345, + "loss": 1.8418, + "theoretical_loss": 4.306483573190857, + "tokens_seen": 231145472 + }, + { + "epoch": 0.07, + "learning_rate": 0.0007514704112337011, + "loss": 1.7501, + "theoretical_loss": 4.306335215327791, + "tokens_seen": 231211008 + }, + { + "epoch": 0.07, + "learning_rate": 0.0007514543630892679, + "loss": 1.7497, + "theoretical_loss": 4.306186911280945, + "tokens_seen": 231276544 + }, + { + "epoch": 0.07, + "learning_rate": 0.0007514383149448345, + "loss": 1.6281, + "theoretical_loss": 4.306038661015557, + "tokens_seen": 231342080 + }, + { + "epoch": 0.07, + "learning_rate": 0.0007514222668004013, + "loss": 1.644, + "theoretical_loss": 4.305890464496895, + "tokens_seen": 231407616 + }, + { + "epoch": 0.07, + "learning_rate": 0.0007514062186559679, + "loss": 1.8915, + "theoretical_loss": 4.305742321690261, + "tokens_seen": 231473152 + }, + { + "epoch": 0.07, + "learning_rate": 0.0007513901705115347, + "loss": 1.7172, + "theoretical_loss": 4.305594232560988, + "tokens_seen": 231538688 + }, + { + "epoch": 0.07, + "learning_rate": 0.0007513741223671013, + "loss": 1.8509, + "theoretical_loss": 4.305446197074443, + "tokens_seen": 231604224 + }, + { + "epoch": 0.07, + "learning_rate": 0.0007513580742226681, + "loss": 1.6805, + "theoretical_loss": 4.305298215196023, + "tokens_seen": 231669760 + }, + { + "epoch": 0.07, + "learning_rate": 0.0007513420260782347, + "loss": 1.754, + "theoretical_loss": 4.305150286891159, + "tokens_seen": 231735296 + }, + { + "epoch": 0.07, + "learning_rate": 0.0007513259779338015, + "loss": 1.881, + "theoretical_loss": 4.3050024121253125, + "tokens_seen": 231800832 + }, + { + "epoch": 0.07, + "learning_rate": 0.0007513099297893681, + "loss": 1.6317, + "theoretical_loss": 4.304854590863979, + "tokens_seen": 231866368 + }, + { + "epoch": 0.07, + "learning_rate": 0.0007512938816449349, + "loss": 1.6523, + "theoretical_loss": 4.304706823072683, + "tokens_seen": 231931904 + }, + { + "epoch": 0.07, + "learning_rate": 0.0007512778335005015, + "loss": 1.8767, + "theoretical_loss": 4.304559108716984, + "tokens_seen": 231997440 + }, + { + "epoch": 0.07, + "learning_rate": 0.0007512617853560683, + "loss": 1.8357, + "theoretical_loss": 4.304411447762472, + "tokens_seen": 232062976 + }, + { + "epoch": 0.07, + "learning_rate": 0.0007512457372116349, + "loss": 1.7207, + "theoretical_loss": 4.304263840174768, + "tokens_seen": 232128512 + }, + { + "epoch": 0.07, + "learning_rate": 0.0007512296890672017, + "loss": 1.5789, + "theoretical_loss": 4.304116285919527, + "tokens_seen": 232194048 + }, + { + "epoch": 0.07, + "learning_rate": 0.0007512136409227683, + "loss": 1.7458, + "theoretical_loss": 4.303968784962433, + "tokens_seen": 232259584 + }, + { + "epoch": 0.07, + "learning_rate": 0.0007511975927783351, + "loss": 1.7072, + "theoretical_loss": 4.303821337269205, + "tokens_seen": 232325120 + }, + { + "epoch": 0.07, + "learning_rate": 0.0007511815446339017, + "loss": 1.7604, + "theoretical_loss": 4.303673942805591, + "tokens_seen": 232390656 + }, + { + "epoch": 0.07, + "learning_rate": 0.0007511654964894685, + "loss": 1.7298, + "theoretical_loss": 4.303526601537371, + "tokens_seen": 232456192 + }, + { + "epoch": 0.07, + "learning_rate": 0.0007511494483450351, + "loss": 1.8322, + "theoretical_loss": 4.303379313430359, + "tokens_seen": 232521728 + }, + { + "epoch": 0.07, + "learning_rate": 0.0007511334002006019, + "loss": 1.6559, + "theoretical_loss": 4.303232078450399, + "tokens_seen": 232587264 + }, + { + "epoch": 0.07, + "objective/train/docs_used": 91258, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.6416524648666382, + "objective/train/theoretical_loss": 4.303084896563364, + "objective/train/tokens_used": 253112800, + "theoretical_loss": 4.303084896563364, + "tokens_seen": 232652800 + }, + { + "epoch": 0.07, + "learning_rate": 0.0007511173520561685, + "loss": 1.7381, + "theoretical_loss": 4.303084896563364, + "tokens_seen": 232652800 + }, + { + "epoch": 0.07, + "learning_rate": 0.0007511013039117353, + "loss": 1.6231, + "theoretical_loss": 4.302937767735164, + "tokens_seen": 232718336 + }, + { + "epoch": 0.07, + "learning_rate": 0.000751085255767302, + "loss": 1.7068, + "theoretical_loss": 4.302790691931736, + "tokens_seen": 232783872 + }, + { + "epoch": 0.07, + "learning_rate": 0.0007510692076228687, + "loss": 1.744, + "theoretical_loss": 4.30264366911905, + "tokens_seen": 232849408 + }, + { + "epoch": 0.07, + "learning_rate": 0.0007510531594784353, + "loss": 1.7707, + "theoretical_loss": 4.3024966992631075, + "tokens_seen": 232914944 + }, + { + "epoch": 0.07, + "learning_rate": 0.0007510371113340021, + "loss": 1.7062, + "theoretical_loss": 4.302349782329942, + "tokens_seen": 232980480 + }, + { + "epoch": 0.07, + "learning_rate": 0.0007510210631895687, + "loss": 1.7569, + "theoretical_loss": 4.302202918285616, + "tokens_seen": 233046016 + }, + { + "epoch": 0.07, + "learning_rate": 0.0007510050150451355, + "loss": 1.7972, + "theoretical_loss": 4.302056107096226, + "tokens_seen": 233111552 + }, + { + "epoch": 0.07, + "learning_rate": 0.0007509889669007021, + "loss": 1.6663, + "theoretical_loss": 4.301909348727899, + "tokens_seen": 233177088 + }, + { + "epoch": 0.07, + "learning_rate": 0.0007509729187562688, + "loss": 1.7164, + "theoretical_loss": 4.301762643146792, + "tokens_seen": 233242624 + }, + { + "epoch": 0.07, + "learning_rate": 0.0007509568706118356, + "loss": 1.8111, + "theoretical_loss": 4.301615990319094, + "tokens_seen": 233308160 + }, + { + "epoch": 0.07, + "learning_rate": 0.0007509408224674022, + "loss": 1.7613, + "theoretical_loss": 4.301469390211025, + "tokens_seen": 233373696 + }, + { + "epoch": 0.07, + "learning_rate": 0.000750924774322969, + "loss": 1.851, + "theoretical_loss": 4.301322842788839, + "tokens_seen": 233439232 + }, + { + "epoch": 0.07, + "learning_rate": 0.0007509087261785356, + "loss": 1.674, + "theoretical_loss": 4.301176348018814, + "tokens_seen": 233504768 + }, + { + "epoch": 0.07, + "learning_rate": 0.0007508926780341024, + "loss": 1.6989, + "theoretical_loss": 4.301029905867265, + "tokens_seen": 233570304 + }, + { + "epoch": 0.07, + "learning_rate": 0.000750876629889669, + "loss": 1.6218, + "theoretical_loss": 4.300883516300539, + "tokens_seen": 233635840 + }, + { + "epoch": 0.07, + "learning_rate": 0.0007508605817452358, + "loss": 1.6716, + "theoretical_loss": 4.300737179285008, + "tokens_seen": 233701376 + }, + { + "epoch": 0.07, + "learning_rate": 0.0007508445336008024, + "loss": 1.7063, + "theoretical_loss": 4.30059089478708, + "tokens_seen": 233766912 + }, + { + "epoch": 0.07, + "learning_rate": 0.000750828485456369, + "loss": 1.753, + "theoretical_loss": 4.300444662773192, + "tokens_seen": 233832448 + }, + { + "epoch": 0.07, + "learning_rate": 0.0007508124373119358, + "loss": 1.6775, + "theoretical_loss": 4.30029848320981, + "tokens_seen": 233897984 + }, + { + "epoch": 0.07, + "learning_rate": 0.0007507963891675024, + "loss": 1.7869, + "theoretical_loss": 4.300152356063435, + "tokens_seen": 233963520 + }, + { + "epoch": 0.07, + "learning_rate": 0.0007507803410230692, + "loss": 1.8083, + "theoretical_loss": 4.300006281300598, + "tokens_seen": 234029056 + }, + { + "epoch": 0.07, + "learning_rate": 0.0007507642928786359, + "loss": 1.7309, + "theoretical_loss": 4.299860258887856, + "tokens_seen": 234094592 + }, + { + "epoch": 0.07, + "learning_rate": 0.0007507482447342026, + "loss": 1.8748, + "theoretical_loss": 4.299714288791802, + "tokens_seen": 234160128 + }, + { + "epoch": 0.07, + "learning_rate": 0.0007507321965897694, + "loss": 1.5748, + "theoretical_loss": 4.299568370979056, + "tokens_seen": 234225664 + }, + { + "epoch": 0.07, + "learning_rate": 0.000750716148445336, + "loss": 1.8127, + "theoretical_loss": 4.299422505416272, + "tokens_seen": 234291200 + }, + { + "epoch": 0.07, + "learning_rate": 0.0007507001003009028, + "loss": 1.7826, + "theoretical_loss": 4.2992766920701335, + "tokens_seen": 234356736 + }, + { + "epoch": 0.07, + "learning_rate": 0.0007506840521564694, + "loss": 1.7522, + "theoretical_loss": 4.299130930907352, + "tokens_seen": 234422272 + }, + { + "epoch": 0.07, + "learning_rate": 0.0007506680040120362, + "loss": 1.7341, + "theoretical_loss": 4.2989852218946725, + "tokens_seen": 234487808 + }, + { + "epoch": 0.07, + "learning_rate": 0.0007506519558676028, + "loss": 1.7595, + "theoretical_loss": 4.298839564998869, + "tokens_seen": 234553344 + }, + { + "epoch": 0.07, + "learning_rate": 0.0007506359077231696, + "loss": 1.8532, + "theoretical_loss": 4.298693960186747, + "tokens_seen": 234618880 + }, + { + "epoch": 0.07, + "learning_rate": 0.0007506198595787362, + "loss": 1.8158, + "theoretical_loss": 4.298548407425141, + "tokens_seen": 234684416 + }, + { + "epoch": 0.07, + "learning_rate": 0.000750603811434303, + "loss": 1.6517, + "theoretical_loss": 4.298402906680918, + "tokens_seen": 234749952 + }, + { + "epoch": 0.07, + "learning_rate": 0.0007505877632898696, + "loss": 1.7427, + "theoretical_loss": 4.298257457920973, + "tokens_seen": 234815488 + }, + { + "epoch": 0.07, + "learning_rate": 0.0007505717151454364, + "loss": 1.7534, + "theoretical_loss": 4.298112061112232, + "tokens_seen": 234881024 + }, + { + "epoch": 0.07, + "learning_rate": 0.000750555667001003, + "loss": 1.8078, + "theoretical_loss": 4.297966716221653, + "tokens_seen": 234946560 + }, + { + "epoch": 0.07, + "learning_rate": 0.0007505396188565698, + "loss": 1.779, + "theoretical_loss": 4.297821423216222, + "tokens_seen": 235012096 + }, + { + "epoch": 0.07, + "learning_rate": 0.0007505235707121364, + "loss": 1.7278, + "theoretical_loss": 4.297676182062955, + "tokens_seen": 235077632 + }, + { + "epoch": 0.07, + "learning_rate": 0.0007505075225677032, + "loss": 1.7346, + "theoretical_loss": 4.297530992728901, + "tokens_seen": 235143168 + }, + { + "epoch": 0.07, + "learning_rate": 0.0007504914744232698, + "loss": 1.7386, + "theoretical_loss": 4.297385855181136, + "tokens_seen": 235208704 + }, + { + "epoch": 0.07, + "learning_rate": 0.0007504754262788366, + "loss": 1.7258, + "theoretical_loss": 4.297240769386768, + "tokens_seen": 235274240 + }, + { + "epoch": 0.07, + "learning_rate": 0.0007504593781344032, + "loss": 1.7867, + "theoretical_loss": 4.297095735312934, + "tokens_seen": 235339776 + }, + { + "epoch": 0.07, + "learning_rate": 0.00075044332998997, + "loss": 1.6718, + "theoretical_loss": 4.296950752926801, + "tokens_seen": 235405312 + }, + { + "epoch": 0.07, + "learning_rate": 0.0007504272818455366, + "loss": 1.6862, + "theoretical_loss": 4.296805822195568, + "tokens_seen": 235470848 + }, + { + "epoch": 0.07, + "learning_rate": 0.0007504112337011034, + "loss": 1.724, + "theoretical_loss": 4.29666094308646, + "tokens_seen": 235536384 + }, + { + "epoch": 0.07, + "learning_rate": 0.00075039518555667, + "loss": 1.7583, + "theoretical_loss": 4.296516115566736, + "tokens_seen": 235601920 + }, + { + "epoch": 0.07, + "learning_rate": 0.0007503791374122368, + "loss": 1.8516, + "theoretical_loss": 4.296371339603684, + "tokens_seen": 235667456 + }, + { + "epoch": 0.07, + "learning_rate": 0.0007503630892678034, + "loss": 1.6938, + "theoretical_loss": 4.296226615164618, + "tokens_seen": 235732992 + }, + { + "epoch": 0.07, + "learning_rate": 0.0007503470411233702, + "loss": 1.7671, + "theoretical_loss": 4.296081942216887, + "tokens_seen": 235798528 + }, + { + "epoch": 0.07, + "learning_rate": 0.0007503309929789368, + "loss": 1.6483, + "theoretical_loss": 4.295937320727866, + "tokens_seen": 235864064 + }, + { + "epoch": 0.07, + "objective/train/docs_used": 92471, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.8403021097183228, + "objective/train/theoretical_loss": 4.295792750664964, + "objective/train/tokens_used": 256389600, + "theoretical_loss": 4.295792750664964, + "tokens_seen": 235929600 + }, + { + "epoch": 0.07, + "learning_rate": 0.0007503149448345036, + "loss": 1.7702, + "theoretical_loss": 4.295792750664964, + "tokens_seen": 235929600 + }, + { + "epoch": 0.07, + "learning_rate": 0.0007502988966900702, + "loss": 1.7537, + "theoretical_loss": 4.295648231995613, + "tokens_seen": 235995136 + }, + { + "epoch": 0.07, + "learning_rate": 0.000750282848545637, + "loss": 1.8606, + "theoretical_loss": 4.295503764687281, + "tokens_seen": 236060672 + }, + { + "epoch": 0.07, + "learning_rate": 0.0007502668004012036, + "loss": 1.672, + "theoretical_loss": 4.295359348707462, + "tokens_seen": 236126208 + }, + { + "epoch": 0.07, + "learning_rate": 0.0007502507522567704, + "loss": 1.677, + "theoretical_loss": 4.2952149840236835, + "tokens_seen": 236191744 + }, + { + "epoch": 0.07, + "learning_rate": 0.000750234704112337, + "loss": 1.7903, + "theoretical_loss": 4.295070670603496, + "tokens_seen": 236257280 + }, + { + "epoch": 0.07, + "learning_rate": 0.0007502186559679038, + "loss": 1.6568, + "theoretical_loss": 4.294926408414486, + "tokens_seen": 236322816 + }, + { + "epoch": 0.07, + "learning_rate": 0.0007502026078234704, + "loss": 1.7027, + "theoretical_loss": 4.294782197424265, + "tokens_seen": 236388352 + }, + { + "epoch": 0.07, + "learning_rate": 0.0007501865596790372, + "loss": 1.7169, + "theoretical_loss": 4.294638037600478, + "tokens_seen": 236453888 + }, + { + "epoch": 0.07, + "learning_rate": 0.0007501705115346038, + "loss": 1.8011, + "theoretical_loss": 4.2944939289107955, + "tokens_seen": 236519424 + }, + { + "epoch": 0.07, + "learning_rate": 0.0007501544633901706, + "loss": 1.8126, + "theoretical_loss": 4.29434987132292, + "tokens_seen": 236584960 + }, + { + "epoch": 0.07, + "learning_rate": 0.0007501384152457372, + "loss": 1.9142, + "theoretical_loss": 4.294205864804581, + "tokens_seen": 236650496 + }, + { + "epoch": 0.07, + "learning_rate": 0.000750122367101304, + "loss": 1.6751, + "theoretical_loss": 4.294061909323539, + "tokens_seen": 236716032 + }, + { + "epoch": 0.07, + "learning_rate": 0.0007501063189568706, + "loss": 1.7729, + "theoretical_loss": 4.293918004847585, + "tokens_seen": 236781568 + }, + { + "epoch": 0.07, + "learning_rate": 0.0007500902708124374, + "loss": 1.7398, + "theoretical_loss": 4.293774151344537, + "tokens_seen": 236847104 + }, + { + "epoch": 0.07, + "learning_rate": 0.000750074222668004, + "loss": 1.8337, + "theoretical_loss": 4.2936303487822425, + "tokens_seen": 236912640 + }, + { + "epoch": 0.07, + "learning_rate": 0.0007500581745235708, + "loss": 1.725, + "theoretical_loss": 4.293486597128578, + "tokens_seen": 236978176 + }, + { + "epoch": 0.07, + "learning_rate": 0.0007500421263791374, + "loss": 1.6685, + "theoretical_loss": 4.293342896351452, + "tokens_seen": 237043712 + }, + { + "epoch": 0.07, + "learning_rate": 0.0007500260782347042, + "loss": 1.6995, + "theoretical_loss": 4.293199246418796, + "tokens_seen": 237109248 + }, + { + "epoch": 0.07, + "learning_rate": 0.0007500100300902708, + "loss": 1.725, + "theoretical_loss": 4.293055647298578, + "tokens_seen": 237174784 + }, + { + "epoch": 0.07, + "learning_rate": 0.0007499939819458376, + "loss": 1.7783, + "theoretical_loss": 4.292912098958791, + "tokens_seen": 237240320 + }, + { + "epoch": 0.07, + "learning_rate": 0.0007499779338014042, + "loss": 1.6958, + "theoretical_loss": 4.292768601367455, + "tokens_seen": 237305856 + }, + { + "epoch": 0.07, + "learning_rate": 0.000749961885656971, + "loss": 1.6506, + "theoretical_loss": 4.292625154492623, + "tokens_seen": 237371392 + }, + { + "epoch": 0.07, + "learning_rate": 0.0007499458375125376, + "loss": 1.658, + "theoretical_loss": 4.292481758302375, + "tokens_seen": 237436928 + }, + { + "epoch": 0.07, + "learning_rate": 0.0007499297893681044, + "loss": 1.6479, + "theoretical_loss": 4.2923384127648205, + "tokens_seen": 237502464 + }, + { + "epoch": 0.07, + "learning_rate": 0.000749913741223671, + "loss": 1.6241, + "theoretical_loss": 4.292195117848098, + "tokens_seen": 237568000 + }, + { + "epoch": 0.07, + "learning_rate": 0.0007498976930792378, + "loss": 1.6924, + "theoretical_loss": 4.292051873520373, + "tokens_seen": 237633536 + }, + { + "epoch": 0.07, + "learning_rate": 0.0007498816449348044, + "loss": 1.6962, + "theoretical_loss": 4.291908679749842, + "tokens_seen": 237699072 + }, + { + "epoch": 0.07, + "learning_rate": 0.0007498655967903712, + "loss": 1.8643, + "theoretical_loss": 4.2917655365047285, + "tokens_seen": 237764608 + }, + { + "epoch": 0.07, + "learning_rate": 0.0007498495486459378, + "loss": 1.633, + "theoretical_loss": 4.291622443753287, + "tokens_seen": 237830144 + }, + { + "epoch": 0.07, + "learning_rate": 0.0007498335005015046, + "loss": 1.7924, + "theoretical_loss": 4.291479401463799, + "tokens_seen": 237895680 + }, + { + "epoch": 0.07, + "learning_rate": 0.0007498174523570712, + "loss": 1.7182, + "theoretical_loss": 4.291336409604574, + "tokens_seen": 237961216 + }, + { + "epoch": 0.07, + "learning_rate": 0.000749801404212638, + "loss": 1.8243, + "theoretical_loss": 4.291193468143952, + "tokens_seen": 238026752 + }, + { + "epoch": 0.07, + "learning_rate": 0.0007497853560682046, + "loss": 1.7804, + "theoretical_loss": 4.2910505770503, + "tokens_seen": 238092288 + }, + { + "epoch": 0.07, + "learning_rate": 0.0007497693079237714, + "loss": 1.6885, + "theoretical_loss": 4.290907736292016, + "tokens_seen": 238157824 + }, + { + "epoch": 0.07, + "learning_rate": 0.000749753259779338, + "loss": 1.667, + "theoretical_loss": 4.290764945837522, + "tokens_seen": 238223360 + }, + { + "epoch": 0.07, + "learning_rate": 0.0007497372116349048, + "loss": 1.7627, + "theoretical_loss": 4.290622205655273, + "tokens_seen": 238288896 + }, + { + "epoch": 0.07, + "learning_rate": 0.0007497211634904714, + "loss": 1.6439, + "theoretical_loss": 4.290479515713751, + "tokens_seen": 238354432 + }, + { + "epoch": 0.07, + "learning_rate": 0.0007497051153460382, + "loss": 1.6521, + "theoretical_loss": 4.290336875981465, + "tokens_seen": 238419968 + }, + { + "epoch": 0.07, + "learning_rate": 0.0007496890672016048, + "loss": 1.6508, + "theoretical_loss": 4.290194286426955, + "tokens_seen": 238485504 + }, + { + "epoch": 0.07, + "learning_rate": 0.0007496730190571716, + "loss": 1.8122, + "theoretical_loss": 4.290051747018785, + "tokens_seen": 238551040 + }, + { + "epoch": 0.07, + "learning_rate": 0.0007496569709127382, + "loss": 1.8236, + "theoretical_loss": 4.289909257725553, + "tokens_seen": 238616576 + }, + { + "epoch": 0.07, + "learning_rate": 0.000749640922768305, + "loss": 1.7384, + "theoretical_loss": 4.289766818515882, + "tokens_seen": 238682112 + }, + { + "epoch": 0.07, + "learning_rate": 0.0007496248746238716, + "loss": 1.8606, + "theoretical_loss": 4.2896244293584225, + "tokens_seen": 238747648 + }, + { + "epoch": 0.07, + "learning_rate": 0.0007496088264794384, + "loss": 1.7366, + "theoretical_loss": 4.2894820902218544, + "tokens_seen": 238813184 + }, + { + "epoch": 0.07, + "learning_rate": 0.000749592778335005, + "loss": 1.6865, + "theoretical_loss": 4.289339801074888, + "tokens_seen": 238878720 + }, + { + "epoch": 0.07, + "learning_rate": 0.0007495767301905718, + "loss": 1.6901, + "theoretical_loss": 4.289197561886258, + "tokens_seen": 238944256 + }, + { + "epoch": 0.07, + "learning_rate": 0.0007495606820461385, + "loss": 1.8293, + "theoretical_loss": 4.289055372624729, + "tokens_seen": 239009792 + }, + { + "epoch": 0.07, + "learning_rate": 0.0007495446339017051, + "loss": 1.7704, + "theoretical_loss": 4.288913233259092, + "tokens_seen": 239075328 + }, + { + "epoch": 0.07, + "learning_rate": 0.0007495285857572719, + "loss": 1.7207, + "theoretical_loss": 4.28877114375817, + "tokens_seen": 239140864 + }, + { + "epoch": 0.07, + "objective/train/docs_used": 93570, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.8132266998291016, + "objective/train/theoretical_loss": 4.288629104090811, + "objective/train/tokens_used": 259666400, + "theoretical_loss": 4.288629104090811, + "tokens_seen": 239206400 + }, + { + "epoch": 0.07, + "learning_rate": 0.0007495125376128385, + "loss": 1.7915, + "theoretical_loss": 4.288629104090811, + "tokens_seen": 239206400 + }, + { + "epoch": 0.07, + "learning_rate": 0.0007494964894684053, + "loss": 1.7168, + "theoretical_loss": 4.28848711422589, + "tokens_seen": 239271936 + }, + { + "epoch": 0.07, + "learning_rate": 0.0007494804413239719, + "loss": 1.7078, + "theoretical_loss": 4.288345174132313, + "tokens_seen": 239337472 + }, + { + "epoch": 0.07, + "learning_rate": 0.0007494643931795387, + "loss": 1.8581, + "theoretical_loss": 4.288203283779012, + "tokens_seen": 239403008 + }, + { + "epoch": 0.07, + "learning_rate": 0.0007494483450351053, + "loss": 1.7838, + "theoretical_loss": 4.288061443134947, + "tokens_seen": 239468544 + }, + { + "epoch": 0.07, + "learning_rate": 0.0007494322968906721, + "loss": 1.7857, + "theoretical_loss": 4.287919652169107, + "tokens_seen": 239534080 + }, + { + "epoch": 0.07, + "learning_rate": 0.0007494162487462387, + "loss": 1.635, + "theoretical_loss": 4.287777910850507, + "tokens_seen": 239599616 + }, + { + "epoch": 0.07, + "learning_rate": 0.0007494002006018055, + "loss": 1.7202, + "theoretical_loss": 4.287636219148191, + "tokens_seen": 239665152 + }, + { + "epoch": 0.07, + "learning_rate": 0.0007493841524573721, + "loss": 1.783, + "theoretical_loss": 4.287494577031231, + "tokens_seen": 239730688 + }, + { + "epoch": 0.07, + "learning_rate": 0.0007493681043129389, + "loss": 1.7387, + "theoretical_loss": 4.287352984468727, + "tokens_seen": 239796224 + }, + { + "epoch": 0.07, + "learning_rate": 0.0007493520561685055, + "loss": 1.8001, + "theoretical_loss": 4.287211441429805, + "tokens_seen": 239861760 + }, + { + "epoch": 0.07, + "learning_rate": 0.0007493360080240723, + "loss": 1.8108, + "theoretical_loss": 4.2870699478836185, + "tokens_seen": 239927296 + }, + { + "epoch": 0.07, + "learning_rate": 0.0007493199598796389, + "loss": 1.7244, + "theoretical_loss": 4.286928503799352, + "tokens_seen": 239992832 + }, + { + "epoch": 0.07, + "learning_rate": 0.0007493039117352057, + "loss": 1.6407, + "theoretical_loss": 4.286787109146213, + "tokens_seen": 240058368 + }, + { + "epoch": 0.07, + "learning_rate": 0.0007492878635907723, + "loss": 1.7601, + "theoretical_loss": 4.286645763893442, + "tokens_seen": 240123904 + }, + { + "epoch": 0.07, + "learning_rate": 0.0007492718154463391, + "loss": 1.7764, + "theoretical_loss": 4.2865044680103015, + "tokens_seen": 240189440 + }, + { + "epoch": 0.07, + "learning_rate": 0.0007492557673019057, + "loss": 1.8299, + "theoretical_loss": 4.286363221466084, + "tokens_seen": 240254976 + }, + { + "epoch": 0.07, + "learning_rate": 0.0007492397191574725, + "loss": 1.752, + "theoretical_loss": 4.286222024230111, + "tokens_seen": 240320512 + }, + { + "epoch": 0.07, + "learning_rate": 0.0007492236710130391, + "loss": 1.6449, + "theoretical_loss": 4.286080876271729, + "tokens_seen": 240386048 + }, + { + "epoch": 0.07, + "learning_rate": 0.0007492076228686059, + "loss": 1.7069, + "theoretical_loss": 4.285939777560312, + "tokens_seen": 240451584 + }, + { + "epoch": 0.07, + "learning_rate": 0.0007491915747241725, + "loss": 1.8281, + "theoretical_loss": 4.285798728065264, + "tokens_seen": 240517120 + }, + { + "epoch": 0.07, + "learning_rate": 0.0007491755265797393, + "loss": 1.8072, + "theoretical_loss": 4.285657727756012, + "tokens_seen": 240582656 + }, + { + "epoch": 0.07, + "learning_rate": 0.0007491594784353059, + "loss": 1.8265, + "theoretical_loss": 4.285516776602016, + "tokens_seen": 240648192 + }, + { + "epoch": 0.07, + "learning_rate": 0.0007491434302908727, + "loss": 1.6805, + "theoretical_loss": 4.2853758745727575, + "tokens_seen": 240713728 + }, + { + "epoch": 0.07, + "learning_rate": 0.0007491273821464393, + "loss": 1.5942, + "theoretical_loss": 4.285235021637748, + "tokens_seen": 240779264 + }, + { + "epoch": 0.07, + "learning_rate": 0.0007491113340020061, + "loss": 1.6507, + "theoretical_loss": 4.285094217766527, + "tokens_seen": 240844800 + }, + { + "epoch": 0.07, + "learning_rate": 0.0007490952858575727, + "loss": 1.6567, + "theoretical_loss": 4.2849534629286605, + "tokens_seen": 240910336 + }, + { + "epoch": 0.07, + "learning_rate": 0.0007490792377131395, + "loss": 1.7083, + "theoretical_loss": 4.284812757093741, + "tokens_seen": 240975872 + }, + { + "epoch": 0.07, + "learning_rate": 0.0007490631895687061, + "loss": 1.6243, + "theoretical_loss": 4.284672100231388, + "tokens_seen": 241041408 + }, + { + "epoch": 0.07, + "learning_rate": 0.0007490471414242729, + "loss": 1.7736, + "theoretical_loss": 4.284531492311248, + "tokens_seen": 241106944 + }, + { + "epoch": 0.07, + "learning_rate": 0.0007490310932798395, + "loss": 1.6471, + "theoretical_loss": 4.284390933302997, + "tokens_seen": 241172480 + }, + { + "epoch": 0.07, + "learning_rate": 0.0007490150451354063, + "loss": 1.6745, + "theoretical_loss": 4.284250423176335, + "tokens_seen": 241238016 + }, + { + "epoch": 0.07, + "learning_rate": 0.0007489989969909729, + "loss": 1.704, + "theoretical_loss": 4.28410996190099, + "tokens_seen": 241303552 + }, + { + "epoch": 0.07, + "learning_rate": 0.0007489829488465397, + "loss": 1.721, + "theoretical_loss": 4.283969549446717, + "tokens_seen": 241369088 + }, + { + "epoch": 0.07, + "learning_rate": 0.0007489669007021063, + "loss": 1.782, + "theoretical_loss": 4.2838291857833, + "tokens_seen": 241434624 + }, + { + "epoch": 0.07, + "learning_rate": 0.0007489508525576731, + "loss": 1.7332, + "theoretical_loss": 4.283688870880545, + "tokens_seen": 241500160 + }, + { + "epoch": 0.07, + "learning_rate": 0.0007489348044132397, + "loss": 1.6264, + "theoretical_loss": 4.28354860470829, + "tokens_seen": 241565696 + }, + { + "epoch": 0.07, + "learning_rate": 0.0007489187562688065, + "loss": 1.719, + "theoretical_loss": 4.283408387236397, + "tokens_seen": 241631232 + }, + { + "epoch": 0.07, + "learning_rate": 0.0007489027081243731, + "loss": 1.7083, + "theoretical_loss": 4.283268218434755, + "tokens_seen": 241696768 + }, + { + "epoch": 0.07, + "learning_rate": 0.0007488866599799399, + "loss": 1.6789, + "theoretical_loss": 4.283128098273281, + "tokens_seen": 241762304 + }, + { + "epoch": 0.07, + "learning_rate": 0.0007488706118355065, + "loss": 1.6976, + "theoretical_loss": 4.282988026721917, + "tokens_seen": 241827840 + }, + { + "epoch": 0.07, + "learning_rate": 0.0007488545636910733, + "loss": 1.5801, + "theoretical_loss": 4.282848003750634, + "tokens_seen": 241893376 + }, + { + "epoch": 0.07, + "learning_rate": 0.0007488385155466399, + "loss": 1.6901, + "theoretical_loss": 4.282708029329427, + "tokens_seen": 241958912 + }, + { + "epoch": 0.07, + "learning_rate": 0.0007488224674022067, + "loss": 1.7822, + "theoretical_loss": 4.2825681034283205, + "tokens_seen": 242024448 + }, + { + "epoch": 0.07, + "learning_rate": 0.0007488064192577733, + "loss": 1.799, + "theoretical_loss": 4.2824282260173625, + "tokens_seen": 242089984 + }, + { + "epoch": 0.07, + "learning_rate": 0.0007487903711133401, + "loss": 1.8296, + "theoretical_loss": 4.282288397066631, + "tokens_seen": 242155520 + }, + { + "epoch": 0.07, + "learning_rate": 0.0007487743229689067, + "loss": 1.7178, + "theoretical_loss": 4.282148616546227, + "tokens_seen": 242221056 + }, + { + "epoch": 0.07, + "learning_rate": 0.0007487582748244735, + "loss": 1.6899, + "theoretical_loss": 4.282008884426281, + "tokens_seen": 242286592 + }, + { + "epoch": 0.07, + "learning_rate": 0.0007487422266800401, + "loss": 1.553, + "theoretical_loss": 4.281869200676948, + "tokens_seen": 242352128 + }, + { + "epoch": 0.07, + "learning_rate": 0.0007487261785356069, + "loss": 1.7794, + "theoretical_loss": 4.281729565268412, + "tokens_seen": 242417664 + }, + { + "epoch": 0.07, + "objective/train/docs_used": 94878, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.7510733604431152, + "objective/train/theoretical_loss": 4.28158997817088, + "objective/train/tokens_used": 262943200, + "theoretical_loss": 4.28158997817088, + "tokens_seen": 242483200 + }, + { + "epoch": 0.07, + "learning_rate": 0.0007487101303911735, + "loss": 1.7041, + "theoretical_loss": 4.28158997817088, + "tokens_seen": 242483200 + }, + { + "epoch": 0.07, + "learning_rate": 0.0007486940822467403, + "loss": 1.743, + "theoretical_loss": 4.281450439354587, + "tokens_seen": 242548736 + }, + { + "epoch": 0.07, + "learning_rate": 0.0007486780341023069, + "loss": 1.7124, + "theoretical_loss": 4.281310948789796, + "tokens_seen": 242614272 + }, + { + "epoch": 0.07, + "learning_rate": 0.0007486619859578737, + "loss": 1.6442, + "theoretical_loss": 4.2811715064467935, + "tokens_seen": 242679808 + }, + { + "epoch": 0.07, + "learning_rate": 0.0007486459378134403, + "loss": 1.756, + "theoretical_loss": 4.281032112295894, + "tokens_seen": 242745344 + }, + { + "epoch": 0.07, + "learning_rate": 0.0007486298896690071, + "loss": 1.723, + "theoretical_loss": 4.2808927663074385, + "tokens_seen": 242810880 + }, + { + "epoch": 0.07, + "learning_rate": 0.0007486138415245737, + "loss": 1.7105, + "theoretical_loss": 4.280753468451793, + "tokens_seen": 242876416 + }, + { + "epoch": 0.07, + "learning_rate": 0.0007485977933801405, + "loss": 1.6751, + "theoretical_loss": 4.28061421869935, + "tokens_seen": 242941952 + }, + { + "epoch": 0.07, + "learning_rate": 0.0007485817452357071, + "loss": 1.7271, + "theoretical_loss": 4.28047501702053, + "tokens_seen": 243007488 + }, + { + "epoch": 0.07, + "learning_rate": 0.0007485656970912739, + "loss": 1.7229, + "theoretical_loss": 4.280335863385778, + "tokens_seen": 243073024 + }, + { + "epoch": 0.07, + "learning_rate": 0.0007485496489468405, + "loss": 1.645, + "theoretical_loss": 4.280196757765564, + "tokens_seen": 243138560 + }, + { + "epoch": 0.07, + "learning_rate": 0.0007485336008024073, + "loss": 1.6585, + "theoretical_loss": 4.280057700130388, + "tokens_seen": 243204096 + }, + { + "epoch": 0.07, + "learning_rate": 0.0007485175526579739, + "loss": 1.7174, + "theoretical_loss": 4.279918690450771, + "tokens_seen": 243269632 + }, + { + "epoch": 0.07, + "learning_rate": 0.0007485015045135407, + "loss": 1.7382, + "theoretical_loss": 4.279779728697263, + "tokens_seen": 243335168 + }, + { + "epoch": 0.07, + "learning_rate": 0.0007484854563691073, + "loss": 1.7215, + "theoretical_loss": 4.279640814840441, + "tokens_seen": 243400704 + }, + { + "epoch": 0.07, + "learning_rate": 0.0007484694082246741, + "loss": 1.7005, + "theoretical_loss": 4.279501948850905, + "tokens_seen": 243466240 + }, + { + "epoch": 0.07, + "learning_rate": 0.0007484533600802407, + "loss": 1.6626, + "theoretical_loss": 4.279363130699284, + "tokens_seen": 243531776 + }, + { + "epoch": 0.07, + "learning_rate": 0.0007484373119358075, + "loss": 1.8033, + "theoretical_loss": 4.279224360356231, + "tokens_seen": 243597312 + }, + { + "epoch": 0.07, + "learning_rate": 0.0007484212637913741, + "loss": 1.7787, + "theoretical_loss": 4.279085637792425, + "tokens_seen": 243662848 + }, + { + "epoch": 0.07, + "learning_rate": 0.0007484052156469409, + "loss": 1.8117, + "theoretical_loss": 4.278946962978573, + "tokens_seen": 243728384 + }, + { + "epoch": 0.07, + "learning_rate": 0.0007483891675025075, + "loss": 1.7271, + "theoretical_loss": 4.278808335885403, + "tokens_seen": 243793920 + }, + { + "epoch": 0.07, + "learning_rate": 0.0007483731193580743, + "loss": 1.7006, + "theoretical_loss": 4.278669756483674, + "tokens_seen": 243859456 + }, + { + "epoch": 0.07, + "learning_rate": 0.0007483570712136409, + "loss": 1.7378, + "theoretical_loss": 4.278531224744167, + "tokens_seen": 243924992 + }, + { + "epoch": 0.07, + "learning_rate": 0.0007483410230692077, + "loss": 1.689, + "theoretical_loss": 4.278392740637694, + "tokens_seen": 243990528 + }, + { + "epoch": 0.07, + "learning_rate": 0.0007483249749247745, + "loss": 1.7157, + "theoretical_loss": 4.278254304135085, + "tokens_seen": 244056064 + }, + { + "epoch": 0.07, + "learning_rate": 0.0007483089267803411, + "loss": 1.8252, + "theoretical_loss": 4.2781159152072, + "tokens_seen": 244121600 + }, + { + "epoch": 0.07, + "learning_rate": 0.0007482928786359077, + "loss": 1.7096, + "theoretical_loss": 4.277977573824928, + "tokens_seen": 244187136 + }, + { + "epoch": 0.07, + "learning_rate": 0.0007482768304914745, + "loss": 1.7419, + "theoretical_loss": 4.277839279959177, + "tokens_seen": 244252672 + }, + { + "epoch": 0.07, + "learning_rate": 0.0007482607823470411, + "loss": 1.6275, + "theoretical_loss": 4.277701033580884, + "tokens_seen": 244318208 + }, + { + "epoch": 0.07, + "learning_rate": 0.0007482447342026079, + "loss": 1.5346, + "theoretical_loss": 4.277562834661013, + "tokens_seen": 244383744 + }, + { + "epoch": 0.07, + "learning_rate": 0.0007482286860581746, + "loss": 1.6855, + "theoretical_loss": 4.277424683170549, + "tokens_seen": 244449280 + }, + { + "epoch": 0.07, + "learning_rate": 0.0007482126379137413, + "loss": 1.6505, + "theoretical_loss": 4.277286579080508, + "tokens_seen": 244514816 + }, + { + "epoch": 0.07, + "learning_rate": 0.000748196589769308, + "loss": 1.7072, + "theoretical_loss": 4.277148522361927, + "tokens_seen": 244580352 + }, + { + "epoch": 0.07, + "learning_rate": 0.0007481805416248747, + "loss": 1.6848, + "theoretical_loss": 4.277010512985871, + "tokens_seen": 244645888 + }, + { + "epoch": 0.07, + "learning_rate": 0.0007481644934804414, + "loss": 1.7946, + "theoretical_loss": 4.2768725509234296, + "tokens_seen": 244711424 + }, + { + "epoch": 0.07, + "learning_rate": 0.000748148445336008, + "loss": 1.6484, + "theoretical_loss": 4.276734636145717, + "tokens_seen": 244776960 + }, + { + "epoch": 0.07, + "learning_rate": 0.0007481323971915748, + "loss": 1.81, + "theoretical_loss": 4.276596768623875, + "tokens_seen": 244842496 + }, + { + "epoch": 0.07, + "learning_rate": 0.0007481163490471414, + "loss": 1.7114, + "theoretical_loss": 4.276458948329068, + "tokens_seen": 244908032 + }, + { + "epoch": 0.07, + "learning_rate": 0.0007481003009027082, + "loss": 1.7432, + "theoretical_loss": 4.276321175232487, + "tokens_seen": 244973568 + }, + { + "epoch": 0.07, + "learning_rate": 0.0007480842527582748, + "loss": 1.6435, + "theoretical_loss": 4.276183449305348, + "tokens_seen": 245039104 + }, + { + "epoch": 0.07, + "learning_rate": 0.0007480682046138416, + "loss": 1.7999, + "theoretical_loss": 4.276045770518895, + "tokens_seen": 245104640 + }, + { + "epoch": 0.07, + "learning_rate": 0.0007480521564694082, + "loss": 1.9077, + "theoretical_loss": 4.275908138844392, + "tokens_seen": 245170176 + }, + { + "epoch": 0.07, + "learning_rate": 0.000748036108324975, + "loss": 1.6817, + "theoretical_loss": 4.275770554253132, + "tokens_seen": 245235712 + }, + { + "epoch": 0.07, + "learning_rate": 0.0007480200601805416, + "loss": 1.6581, + "theoretical_loss": 4.275633016716432, + "tokens_seen": 245301248 + }, + { + "epoch": 0.07, + "learning_rate": 0.0007480040120361084, + "loss": 1.7491, + "theoretical_loss": 4.275495526205635, + "tokens_seen": 245366784 + }, + { + "epoch": 0.07, + "learning_rate": 0.000747987963891675, + "loss": 1.7061, + "theoretical_loss": 4.275358082692106, + "tokens_seen": 245432320 + }, + { + "epoch": 0.07, + "learning_rate": 0.0007479719157472418, + "loss": 1.7529, + "theoretical_loss": 4.275220686147241, + "tokens_seen": 245497856 + }, + { + "epoch": 0.07, + "learning_rate": 0.0007479558676028084, + "loss": 1.5969, + "theoretical_loss": 4.275083336542455, + "tokens_seen": 245563392 + }, + { + "epoch": 0.07, + "learning_rate": 0.0007479398194583752, + "loss": 1.7136, + "theoretical_loss": 4.274946033849189, + "tokens_seen": 245628928 + }, + { + "epoch": 0.07, + "learning_rate": 0.0007479237713139418, + "loss": 1.7013, + "theoretical_loss": 4.2748087780389135, + "tokens_seen": 245694464 + }, + { + "epoch": 0.07, + "objective/train/docs_used": 96014, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.8044483661651611, + "objective/train/theoretical_loss": 4.27467156908312, + "objective/train/tokens_used": 266220000, + "theoretical_loss": 4.27467156908312, + "tokens_seen": 245760000 + }, + { + "epoch": 0.07, + "learning_rate": 0.0007479077231695086, + "loss": 1.6835, + "theoretical_loss": 4.27467156908312, + "tokens_seen": 245760000 + }, + { + "epoch": 0.07, + "learning_rate": 0.0007478916750250752, + "loss": 1.6884, + "theoretical_loss": 4.274534406953325, + "tokens_seen": 245825536 + }, + { + "epoch": 0.07, + "learning_rate": 0.000747875626880642, + "loss": 1.6969, + "theoretical_loss": 4.274397291621073, + "tokens_seen": 245891072 + }, + { + "epoch": 0.07, + "learning_rate": 0.0007478595787362086, + "loss": 1.6677, + "theoretical_loss": 4.274260223057928, + "tokens_seen": 245956608 + }, + { + "epoch": 0.07, + "learning_rate": 0.0007478435305917754, + "loss": 1.7603, + "theoretical_loss": 4.274123201235483, + "tokens_seen": 246022144 + }, + { + "epoch": 0.07, + "learning_rate": 0.000747827482447342, + "loss": 1.7146, + "theoretical_loss": 4.273986226125356, + "tokens_seen": 246087680 + }, + { + "epoch": 0.07, + "learning_rate": 0.0007478114343029088, + "loss": 1.5105, + "theoretical_loss": 4.273849297699186, + "tokens_seen": 246153216 + }, + { + "epoch": 0.07, + "learning_rate": 0.0007477953861584754, + "loss": 1.7729, + "theoretical_loss": 4.273712415928643, + "tokens_seen": 246218752 + }, + { + "epoch": 0.07, + "learning_rate": 0.0007477793380140422, + "loss": 1.7604, + "theoretical_loss": 4.273575580785415, + "tokens_seen": 246284288 + }, + { + "epoch": 0.07, + "learning_rate": 0.0007477632898696088, + "loss": 1.7541, + "theoretical_loss": 4.27343879224122, + "tokens_seen": 246349824 + }, + { + "epoch": 0.07, + "learning_rate": 0.0007477472417251756, + "loss": 1.7456, + "theoretical_loss": 4.273302050267796, + "tokens_seen": 246415360 + }, + { + "epoch": 0.07, + "learning_rate": 0.0007477311935807422, + "loss": 1.7838, + "theoretical_loss": 4.27316535483691, + "tokens_seen": 246480896 + }, + { + "epoch": 0.07, + "learning_rate": 0.000747715145436309, + "loss": 1.7285, + "theoretical_loss": 4.273028705920351, + "tokens_seen": 246546432 + }, + { + "epoch": 0.07, + "learning_rate": 0.0007476990972918756, + "loss": 1.667, + "theoretical_loss": 4.272892103489934, + "tokens_seen": 246611968 + }, + { + "epoch": 0.07, + "learning_rate": 0.0007476830491474424, + "loss": 1.7133, + "theoretical_loss": 4.272755547517495, + "tokens_seen": 246677504 + }, + { + "epoch": 0.07, + "learning_rate": 0.000747667001003009, + "loss": 1.6075, + "theoretical_loss": 4.272619037974902, + "tokens_seen": 246743040 + }, + { + "epoch": 0.07, + "learning_rate": 0.0007476509528585758, + "loss": 1.6131, + "theoretical_loss": 4.272482574834039, + "tokens_seen": 246808576 + }, + { + "epoch": 0.07, + "learning_rate": 0.0007476349047141424, + "loss": 1.6321, + "theoretical_loss": 4.272346158066821, + "tokens_seen": 246874112 + }, + { + "epoch": 0.07, + "learning_rate": 0.0007476188565697092, + "loss": 1.7903, + "theoretical_loss": 4.272209787645183, + "tokens_seen": 246939648 + }, + { + "epoch": 0.07, + "learning_rate": 0.0007476028084252758, + "loss": 1.7218, + "theoretical_loss": 4.272073463541087, + "tokens_seen": 247005184 + }, + { + "epoch": 0.07, + "learning_rate": 0.0007475867602808426, + "loss": 1.6617, + "theoretical_loss": 4.271937185726519, + "tokens_seen": 247070720 + }, + { + "epoch": 0.07, + "learning_rate": 0.0007475707121364092, + "loss": 1.7035, + "theoretical_loss": 4.2718009541734885, + "tokens_seen": 247136256 + }, + { + "epoch": 0.07, + "learning_rate": 0.000747554663991976, + "loss": 1.6743, + "theoretical_loss": 4.27166476885403, + "tokens_seen": 247201792 + }, + { + "epoch": 0.07, + "learning_rate": 0.0007475386158475426, + "loss": 1.8128, + "theoretical_loss": 4.271528629740201, + "tokens_seen": 247267328 + }, + { + "epoch": 0.07, + "learning_rate": 0.0007475225677031094, + "loss": 1.7338, + "theoretical_loss": 4.271392536804088, + "tokens_seen": 247332864 + }, + { + "epoch": 0.07, + "learning_rate": 0.000747506519558676, + "loss": 1.725, + "theoretical_loss": 4.2712564900177945, + "tokens_seen": 247398400 + }, + { + "epoch": 0.07, + "learning_rate": 0.0007474904714142428, + "loss": 1.6911, + "theoretical_loss": 4.271120489353454, + "tokens_seen": 247463936 + }, + { + "epoch": 0.08, + "learning_rate": 0.0007474744232698094, + "loss": 1.6328, + "theoretical_loss": 4.270984534783221, + "tokens_seen": 247529472 + }, + { + "epoch": 0.08, + "learning_rate": 0.0007474583751253762, + "loss": 1.7318, + "theoretical_loss": 4.270848626279275, + "tokens_seen": 247595008 + }, + { + "epoch": 0.08, + "learning_rate": 0.0007474423269809428, + "loss": 1.7288, + "theoretical_loss": 4.270712763813822, + "tokens_seen": 247660544 + }, + { + "epoch": 0.08, + "learning_rate": 0.0007474262788365096, + "loss": 1.7391, + "theoretical_loss": 4.270576947359088, + "tokens_seen": 247726080 + }, + { + "epoch": 0.08, + "learning_rate": 0.0007474102306920762, + "loss": 1.7533, + "theoretical_loss": 4.270441176887327, + "tokens_seen": 247791616 + }, + { + "epoch": 0.08, + "learning_rate": 0.000747394182547643, + "loss": 1.6892, + "theoretical_loss": 4.270305452370814, + "tokens_seen": 247857152 + }, + { + "epoch": 0.08, + "learning_rate": 0.0007473781344032096, + "loss": 1.6365, + "theoretical_loss": 4.27016977378185, + "tokens_seen": 247922688 + }, + { + "epoch": 0.08, + "learning_rate": 0.0007473620862587764, + "loss": 1.7926, + "theoretical_loss": 4.270034141092759, + "tokens_seen": 247988224 + }, + { + "epoch": 0.08, + "learning_rate": 0.000747346038114343, + "loss": 1.717, + "theoretical_loss": 4.269898554275889, + "tokens_seen": 248053760 + }, + { + "epoch": 0.08, + "learning_rate": 0.0007473299899699098, + "loss": 1.6785, + "theoretical_loss": 4.269763013303613, + "tokens_seen": 248119296 + }, + { + "epoch": 0.08, + "learning_rate": 0.0007473139418254764, + "loss": 1.6679, + "theoretical_loss": 4.269627518148326, + "tokens_seen": 248184832 + }, + { + "epoch": 0.08, + "learning_rate": 0.0007472978936810432, + "loss": 1.7058, + "theoretical_loss": 4.26949206878245, + "tokens_seen": 248250368 + }, + { + "epoch": 0.08, + "learning_rate": 0.0007472818455366098, + "loss": 1.8073, + "theoretical_loss": 4.269356665178426, + "tokens_seen": 248315904 + }, + { + "epoch": 0.08, + "learning_rate": 0.0007472657973921766, + "loss": 1.8824, + "theoretical_loss": 4.269221307308725, + "tokens_seen": 248381440 + }, + { + "epoch": 0.08, + "learning_rate": 0.0007472497492477432, + "loss": 1.7697, + "theoretical_loss": 4.269085995145837, + "tokens_seen": 248446976 + }, + { + "epoch": 0.08, + "learning_rate": 0.00074723370110331, + "loss": 1.7275, + "theoretical_loss": 4.268950728662278, + "tokens_seen": 248512512 + }, + { + "epoch": 0.08, + "learning_rate": 0.0007472176529588766, + "loss": 2.0616, + "theoretical_loss": 4.268815507830587, + "tokens_seen": 248578048 + }, + { + "epoch": 0.08, + "learning_rate": 0.0007472016048144434, + "loss": 1.6417, + "theoretical_loss": 4.268680332623328, + "tokens_seen": 248643584 + }, + { + "epoch": 0.08, + "learning_rate": 0.00074718555667001, + "loss": 1.6499, + "theoretical_loss": 4.268545203013085, + "tokens_seen": 248709120 + }, + { + "epoch": 0.08, + "learning_rate": 0.0007471695085255768, + "loss": 1.7156, + "theoretical_loss": 4.26841011897247, + "tokens_seen": 248774656 + }, + { + "epoch": 0.08, + "learning_rate": 0.0007471534603811434, + "loss": 1.8119, + "theoretical_loss": 4.268275080474118, + "tokens_seen": 248840192 + }, + { + "epoch": 0.08, + "learning_rate": 0.0007471374122367102, + "loss": 1.688, + "theoretical_loss": 4.268140087490686, + "tokens_seen": 248905728 + }, + { + "epoch": 0.08, + "learning_rate": 0.000747121364092277, + "loss": 1.7785, + "theoretical_loss": 4.2680051399948535, + "tokens_seen": 248971264 + }, + { + "epoch": 0.08, + "objective/train/docs_used": 97104, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.7050175666809082, + "objective/train/theoretical_loss": 4.267870237959327, + "objective/train/tokens_used": 269496800, + "theoretical_loss": 4.267870237959327, + "tokens_seen": 249036800 + }, + { + "epoch": 0.08, + "learning_rate": 0.0007471053159478436, + "loss": 1.6868, + "theoretical_loss": 4.267870237959327, + "tokens_seen": 249036800 + }, + { + "epoch": 0.08, + "learning_rate": 0.0007470892678034103, + "loss": 1.7011, + "theoretical_loss": 4.267735381356835, + "tokens_seen": 249102336 + }, + { + "epoch": 0.08, + "learning_rate": 0.000747073219658977, + "loss": 1.7533, + "theoretical_loss": 4.26760057016013, + "tokens_seen": 249167872 + }, + { + "epoch": 0.08, + "learning_rate": 0.0007470571715145438, + "loss": 1.8841, + "theoretical_loss": 4.2674658043419855, + "tokens_seen": 249233408 + }, + { + "epoch": 0.08, + "learning_rate": 0.0007470411233701104, + "loss": 1.6654, + "theoretical_loss": 4.267331083875202, + "tokens_seen": 249298944 + }, + { + "epoch": 0.08, + "learning_rate": 0.0007470250752256772, + "loss": 1.6118, + "theoretical_loss": 4.267196408732601, + "tokens_seen": 249364480 + }, + { + "epoch": 0.08, + "learning_rate": 0.0007470090270812438, + "loss": 1.6344, + "theoretical_loss": 4.267061778887029, + "tokens_seen": 249430016 + }, + { + "epoch": 0.08, + "learning_rate": 0.0007469929789368106, + "loss": 1.7346, + "theoretical_loss": 4.2669271943113545, + "tokens_seen": 249495552 + }, + { + "epoch": 0.08, + "learning_rate": 0.0007469769307923772, + "loss": 1.6527, + "theoretical_loss": 4.26679265497847, + "tokens_seen": 249561088 + }, + { + "epoch": 0.08, + "learning_rate": 0.000746960882647944, + "loss": 1.7288, + "theoretical_loss": 4.266658160861293, + "tokens_seen": 249626624 + }, + { + "epoch": 0.08, + "learning_rate": 0.0007469448345035106, + "loss": 1.8412, + "theoretical_loss": 4.266523711932761, + "tokens_seen": 249692160 + }, + { + "epoch": 0.08, + "learning_rate": 0.0007469287863590774, + "loss": 1.8449, + "theoretical_loss": 4.266389308165837, + "tokens_seen": 249757696 + }, + { + "epoch": 0.08, + "learning_rate": 0.000746912738214644, + "loss": 1.7171, + "theoretical_loss": 4.266254949533507, + "tokens_seen": 249823232 + }, + { + "epoch": 0.08, + "learning_rate": 0.0007468966900702106, + "loss": 1.7527, + "theoretical_loss": 4.26612063600878, + "tokens_seen": 249888768 + }, + { + "epoch": 0.08, + "learning_rate": 0.0007468806419257774, + "loss": 1.8751, + "theoretical_loss": 4.265986367564688, + "tokens_seen": 249954304 + }, + { + "epoch": 0.08, + "learning_rate": 0.000746864593781344, + "loss": 1.7722, + "theoretical_loss": 4.265852144174286, + "tokens_seen": 250019840 + }, + { + "epoch": 0.08, + "learning_rate": 0.0007468485456369108, + "loss": 1.6096, + "theoretical_loss": 4.265717965810653, + "tokens_seen": 250085376 + }, + { + "epoch": 0.08, + "learning_rate": 0.0007468324974924775, + "loss": 1.6531, + "theoretical_loss": 4.265583832446892, + "tokens_seen": 250150912 + }, + { + "epoch": 0.08, + "learning_rate": 0.0007468164493480442, + "loss": 1.7547, + "theoretical_loss": 4.265449744056126, + "tokens_seen": 250216448 + }, + { + "epoch": 0.08, + "learning_rate": 0.0007468004012036109, + "loss": 1.7508, + "theoretical_loss": 4.265315700611504, + "tokens_seen": 250281984 + }, + { + "epoch": 0.08, + "learning_rate": 0.0007467843530591776, + "loss": 1.8441, + "theoretical_loss": 4.2651817020861955, + "tokens_seen": 250347520 + }, + { + "epoch": 0.08, + "learning_rate": 0.0007467683049147443, + "loss": 1.7682, + "theoretical_loss": 4.265047748453396, + "tokens_seen": 250413056 + }, + { + "epoch": 0.08, + "learning_rate": 0.0007467522567703109, + "loss": 1.7815, + "theoretical_loss": 4.264913839686323, + "tokens_seen": 250478592 + }, + { + "epoch": 0.08, + "learning_rate": 0.0007467362086258777, + "loss": 1.6966, + "theoretical_loss": 4.264779975758215, + "tokens_seen": 250544128 + }, + { + "epoch": 0.08, + "learning_rate": 0.0007467201604814443, + "loss": 1.6288, + "theoretical_loss": 4.2646461566423355, + "tokens_seen": 250609664 + }, + { + "epoch": 0.08, + "learning_rate": 0.0007467041123370111, + "loss": 1.6935, + "theoretical_loss": 4.26451238231197, + "tokens_seen": 250675200 + }, + { + "epoch": 0.08, + "learning_rate": 0.0007466880641925777, + "loss": 1.8529, + "theoretical_loss": 4.2643786527404295, + "tokens_seen": 250740736 + }, + { + "epoch": 0.08, + "learning_rate": 0.0007466720160481445, + "loss": 1.7577, + "theoretical_loss": 4.264244967901043, + "tokens_seen": 250806272 + }, + { + "epoch": 0.08, + "learning_rate": 0.0007466559679037111, + "loss": 1.6738, + "theoretical_loss": 4.264111327767164, + "tokens_seen": 250871808 + }, + { + "epoch": 0.08, + "learning_rate": 0.0007466399197592779, + "loss": 1.8671, + "theoretical_loss": 4.263977732312174, + "tokens_seen": 250937344 + }, + { + "epoch": 0.08, + "learning_rate": 0.0007466238716148445, + "loss": 1.665, + "theoretical_loss": 4.26384418150947, + "tokens_seen": 251002880 + }, + { + "epoch": 0.08, + "learning_rate": 0.0007466078234704113, + "loss": 1.7992, + "theoretical_loss": 4.263710675332476, + "tokens_seen": 251068416 + }, + { + "epoch": 0.08, + "learning_rate": 0.0007465917753259779, + "loss": 1.8251, + "theoretical_loss": 4.263577213754638, + "tokens_seen": 251133952 + }, + { + "epoch": 0.08, + "learning_rate": 0.0007465757271815447, + "loss": 1.8233, + "theoretical_loss": 4.2634437967494225, + "tokens_seen": 251199488 + }, + { + "epoch": 0.08, + "learning_rate": 0.0007465596790371113, + "loss": 1.7353, + "theoretical_loss": 4.2633104242903235, + "tokens_seen": 251265024 + }, + { + "epoch": 0.08, + "learning_rate": 0.0007465436308926781, + "loss": 1.7695, + "theoretical_loss": 4.263177096350852, + "tokens_seen": 251330560 + }, + { + "epoch": 0.08, + "learning_rate": 0.0007465275827482447, + "loss": 1.6982, + "theoretical_loss": 4.263043812904547, + "tokens_seen": 251396096 + }, + { + "epoch": 0.08, + "learning_rate": 0.0007465115346038115, + "loss": 1.869, + "theoretical_loss": 4.262910573924965, + "tokens_seen": 251461632 + }, + { + "epoch": 0.08, + "learning_rate": 0.0007464954864593781, + "loss": 1.5877, + "theoretical_loss": 4.26277737938569, + "tokens_seen": 251527168 + }, + { + "epoch": 0.08, + "learning_rate": 0.0007464794383149449, + "loss": 1.7694, + "theoretical_loss": 4.262644229260324, + "tokens_seen": 251592704 + }, + { + "epoch": 0.08, + "learning_rate": 0.0007464633901705115, + "loss": 1.816, + "theoretical_loss": 4.262511123522495, + "tokens_seen": 251658240 + }, + { + "epoch": 0.08, + "learning_rate": 0.0007464473420260783, + "loss": 1.6418, + "theoretical_loss": 4.262378062145853, + "tokens_seen": 251723776 + }, + { + "epoch": 0.08, + "learning_rate": 0.0007464312938816449, + "loss": 1.7451, + "theoretical_loss": 4.262245045104068, + "tokens_seen": 251789312 + }, + { + "epoch": 0.08, + "learning_rate": 0.0007464152457372117, + "loss": 1.6533, + "theoretical_loss": 4.262112072370837, + "tokens_seen": 251854848 + }, + { + "epoch": 0.08, + "learning_rate": 0.0007463991975927783, + "loss": 1.7963, + "theoretical_loss": 4.261979143919873, + "tokens_seen": 251920384 + }, + { + "epoch": 0.08, + "learning_rate": 0.0007463831494483451, + "loss": 1.7528, + "theoretical_loss": 4.261846259724917, + "tokens_seen": 251985920 + }, + { + "epoch": 0.08, + "learning_rate": 0.0007463671013039117, + "loss": 1.7288, + "theoretical_loss": 4.261713419759731, + "tokens_seen": 252051456 + }, + { + "epoch": 0.08, + "learning_rate": 0.0007463510531594785, + "loss": 1.6998, + "theoretical_loss": 4.2615806239981, + "tokens_seen": 252116992 + }, + { + "epoch": 0.08, + "learning_rate": 0.0007463350050150451, + "loss": 1.7536, + "theoretical_loss": 4.261447872413827, + "tokens_seen": 252182528 + }, + { + "epoch": 0.08, + "learning_rate": 0.0007463189568706119, + "loss": 1.6487, + "theoretical_loss": 4.261315164980743, + "tokens_seen": 252248064 + }, + { + "epoch": 0.08, + "objective/train/docs_used": 98257, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.7372639179229736, + "objective/train/theoretical_loss": 4.261182501672698, + "objective/train/tokens_used": 272773600, + "theoretical_loss": 4.261182501672698, + "tokens_seen": 252313600 + }, + { + "epoch": 0.08, + "learning_rate": 0.0007463029087261785, + "loss": 1.6653, + "theoretical_loss": 4.261182501672698, + "tokens_seen": 252313600 + }, + { + "epoch": 0.08, + "learning_rate": 0.0007462868605817453, + "loss": 1.7623, + "theoretical_loss": 4.261049882463565, + "tokens_seen": 252379136 + }, + { + "epoch": 0.08, + "learning_rate": 0.0007462708124373119, + "loss": 1.7734, + "theoretical_loss": 4.26091730732724, + "tokens_seen": 252444672 + }, + { + "epoch": 0.08, + "learning_rate": 0.0007462547642928787, + "loss": 1.6221, + "theoretical_loss": 4.2607847762376405, + "tokens_seen": 252510208 + }, + { + "epoch": 0.08, + "learning_rate": 0.0007462387161484453, + "loss": 1.715, + "theoretical_loss": 4.260652289168706, + "tokens_seen": 252575744 + }, + { + "epoch": 0.08, + "learning_rate": 0.0007462226680040121, + "loss": 1.7458, + "theoretical_loss": 4.260519846094399, + "tokens_seen": 252641280 + }, + { + "epoch": 0.08, + "learning_rate": 0.0007462066198595787, + "loss": 1.762, + "theoretical_loss": 4.2603874469887035, + "tokens_seen": 252706816 + }, + { + "epoch": 0.08, + "learning_rate": 0.0007461905717151455, + "loss": 1.7536, + "theoretical_loss": 4.260255091825625, + "tokens_seen": 252772352 + }, + { + "epoch": 0.08, + "learning_rate": 0.0007461745235707121, + "loss": 1.7768, + "theoretical_loss": 4.260122780579193, + "tokens_seen": 252837888 + }, + { + "epoch": 0.08, + "learning_rate": 0.0007461584754262789, + "loss": 1.6594, + "theoretical_loss": 4.259990513223458, + "tokens_seen": 252903424 + }, + { + "epoch": 0.08, + "learning_rate": 0.0007461424272818455, + "loss": 1.807, + "theoretical_loss": 4.259858289732491, + "tokens_seen": 252968960 + }, + { + "epoch": 0.08, + "learning_rate": 0.0007461263791374123, + "loss": 1.6604, + "theoretical_loss": 4.259726110080388, + "tokens_seen": 253034496 + }, + { + "epoch": 0.08, + "learning_rate": 0.0007461103309929789, + "loss": 1.6371, + "theoretical_loss": 4.259593974241264, + "tokens_seen": 253100032 + }, + { + "epoch": 0.08, + "learning_rate": 0.0007460942828485457, + "loss": 1.6921, + "theoretical_loss": 4.25946188218926, + "tokens_seen": 253165568 + }, + { + "epoch": 0.08, + "learning_rate": 0.0007460782347041123, + "loss": 1.6516, + "theoretical_loss": 4.259329833898534, + "tokens_seen": 253231104 + }, + { + "epoch": 0.08, + "learning_rate": 0.0007460621865596791, + "loss": 1.6501, + "theoretical_loss": 4.25919782934327, + "tokens_seen": 253296640 + }, + { + "epoch": 0.08, + "learning_rate": 0.0007460461384152457, + "loss": 1.7773, + "theoretical_loss": 4.259065868497672, + "tokens_seen": 253362176 + }, + { + "epoch": 0.08, + "learning_rate": 0.0007460300902708125, + "loss": 1.7107, + "theoretical_loss": 4.258933951335965, + "tokens_seen": 253427712 + }, + { + "epoch": 0.08, + "learning_rate": 0.0007460140421263791, + "loss": 1.8452, + "theoretical_loss": 4.258802077832399, + "tokens_seen": 253493248 + }, + { + "epoch": 0.08, + "learning_rate": 0.0007459979939819459, + "loss": 1.7053, + "theoretical_loss": 4.258670247961243, + "tokens_seen": 253558784 + }, + { + "epoch": 0.08, + "learning_rate": 0.0007459819458375125, + "loss": 1.7092, + "theoretical_loss": 4.258538461696787, + "tokens_seen": 253624320 + }, + { + "epoch": 0.08, + "learning_rate": 0.0007459658976930793, + "loss": 1.6669, + "theoretical_loss": 4.258406719013347, + "tokens_seen": 253689856 + }, + { + "epoch": 0.08, + "learning_rate": 0.0007459498495486459, + "loss": 1.798, + "theoretical_loss": 4.258275019885257, + "tokens_seen": 253755392 + }, + { + "epoch": 0.08, + "learning_rate": 0.0007459338014042127, + "loss": 1.7625, + "theoretical_loss": 4.258143364286875, + "tokens_seen": 253820928 + }, + { + "epoch": 0.08, + "learning_rate": 0.0007459177532597794, + "loss": 1.6337, + "theoretical_loss": 4.258011752192579, + "tokens_seen": 253886464 + }, + { + "epoch": 0.08, + "learning_rate": 0.0007459017051153461, + "loss": 1.6922, + "theoretical_loss": 4.257880183576768, + "tokens_seen": 253952000 + }, + { + "epoch": 0.08, + "learning_rate": 0.0007458856569709128, + "loss": 1.6999, + "theoretical_loss": 4.257748658413867, + "tokens_seen": 254017536 + }, + { + "epoch": 0.08, + "learning_rate": 0.0007458696088264795, + "loss": 1.6775, + "theoretical_loss": 4.257617176678317, + "tokens_seen": 254083072 + }, + { + "epoch": 0.08, + "learning_rate": 0.0007458535606820462, + "loss": 1.6754, + "theoretical_loss": 4.2574857383445845, + "tokens_seen": 254148608 + }, + { + "epoch": 0.08, + "learning_rate": 0.0007458375125376129, + "loss": 1.7632, + "theoretical_loss": 4.257354343387156, + "tokens_seen": 254214144 + }, + { + "epoch": 0.08, + "learning_rate": 0.0007458214643931796, + "loss": 1.7223, + "theoretical_loss": 4.257222991780541, + "tokens_seen": 254279680 + }, + { + "epoch": 0.08, + "learning_rate": 0.0007458054162487463, + "loss": 1.8358, + "theoretical_loss": 4.257091683499268, + "tokens_seen": 254345216 + }, + { + "epoch": 0.08, + "learning_rate": 0.000745789368104313, + "loss": 1.8973, + "theoretical_loss": 4.25696041851789, + "tokens_seen": 254410752 + }, + { + "epoch": 0.08, + "learning_rate": 0.0007457733199598797, + "loss": 1.7742, + "theoretical_loss": 4.256829196810978, + "tokens_seen": 254476288 + }, + { + "epoch": 0.08, + "learning_rate": 0.0007457572718154464, + "loss": 1.6299, + "theoretical_loss": 4.256698018353129, + "tokens_seen": 254541824 + }, + { + "epoch": 0.08, + "learning_rate": 0.0007457412236710131, + "loss": 1.769, + "theoretical_loss": 4.256566883118958, + "tokens_seen": 254607360 + }, + { + "epoch": 0.08, + "learning_rate": 0.0007457251755265798, + "loss": 1.7065, + "theoretical_loss": 4.256435791083101, + "tokens_seen": 254672896 + }, + { + "epoch": 0.08, + "learning_rate": 0.0007457091273821465, + "loss": 1.7905, + "theoretical_loss": 4.256304742220219, + "tokens_seen": 254738432 + }, + { + "epoch": 0.08, + "learning_rate": 0.0007456930792377133, + "loss": 1.7513, + "theoretical_loss": 4.25617373650499, + "tokens_seen": 254803968 + }, + { + "epoch": 0.08, + "learning_rate": 0.0007456770310932799, + "loss": 1.5869, + "theoretical_loss": 4.256042773912117, + "tokens_seen": 254869504 + }, + { + "epoch": 0.08, + "learning_rate": 0.0007456609829488467, + "loss": 1.7231, + "theoretical_loss": 4.255911854416323, + "tokens_seen": 254935040 + }, + { + "epoch": 0.08, + "learning_rate": 0.0007456449348044133, + "loss": 1.7755, + "theoretical_loss": 4.255780977992352, + "tokens_seen": 255000576 + }, + { + "epoch": 0.08, + "learning_rate": 0.00074562888665998, + "loss": 1.5963, + "theoretical_loss": 4.255650144614968, + "tokens_seen": 255066112 + }, + { + "epoch": 0.08, + "learning_rate": 0.0007456128385155467, + "loss": 1.7501, + "theoretical_loss": 4.255519354258959, + "tokens_seen": 255131648 + }, + { + "epoch": 0.08, + "learning_rate": 0.0007455967903711135, + "loss": 1.7225, + "theoretical_loss": 4.255388606899133, + "tokens_seen": 255197184 + }, + { + "epoch": 0.08, + "learning_rate": 0.0007455807422266801, + "loss": 1.6918, + "theoretical_loss": 4.255257902510319, + "tokens_seen": 255262720 + }, + { + "epoch": 0.08, + "learning_rate": 0.0007455646940822469, + "loss": 1.7493, + "theoretical_loss": 4.255127241067368, + "tokens_seen": 255328256 + }, + { + "epoch": 0.08, + "learning_rate": 0.0007455486459378135, + "loss": 1.7731, + "theoretical_loss": 4.25499662254515, + "tokens_seen": 255393792 + }, + { + "epoch": 0.08, + "learning_rate": 0.0007455325977933803, + "loss": 1.685, + "theoretical_loss": 4.254866046918559, + "tokens_seen": 255459328 + }, + { + "epoch": 0.08, + "learning_rate": 0.0007455165496489469, + "loss": 1.7574, + "theoretical_loss": 4.254735514162508, + "tokens_seen": 255524864 + }, + { + "epoch": 0.08, + "objective/train/docs_used": 99403, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.7591402530670166, + "objective/train/theoretical_loss": 4.254605024251933, + "objective/train/tokens_used": 276050400, + "theoretical_loss": 4.254605024251933, + "tokens_seen": 255590400 + }, + { + "epoch": 0.08, + "learning_rate": 0.0007455005015045136, + "loss": 1.697, + "theoretical_loss": 4.254605024251933, + "tokens_seen": 255590400 + }, + { + "epoch": 0.08, + "learning_rate": 0.0007454844533600803, + "loss": 1.7026, + "theoretical_loss": 4.254474577161789, + "tokens_seen": 255655936 + }, + { + "epoch": 0.08, + "learning_rate": 0.000745468405215647, + "loss": 1.554, + "theoretical_loss": 4.254344172867053, + "tokens_seen": 255721472 + }, + { + "epoch": 0.08, + "learning_rate": 0.0007454523570712137, + "loss": 1.7371, + "theoretical_loss": 4.254213811342723, + "tokens_seen": 255787008 + }, + { + "epoch": 0.08, + "learning_rate": 0.0007454363089267804, + "loss": 1.7308, + "theoretical_loss": 4.254083492563819, + "tokens_seen": 255852544 + }, + { + "epoch": 0.08, + "learning_rate": 0.0007454202607823471, + "loss": 1.6487, + "theoretical_loss": 4.253953216505381, + "tokens_seen": 255918080 + }, + { + "epoch": 0.08, + "learning_rate": 0.0007454042126379138, + "loss": 1.7789, + "theoretical_loss": 4.253822983142468, + "tokens_seen": 255983616 + }, + { + "epoch": 0.08, + "learning_rate": 0.0007453881644934805, + "loss": 1.6448, + "theoretical_loss": 4.253692792450165, + "tokens_seen": 256049152 + }, + { + "epoch": 0.08, + "learning_rate": 0.0007453721163490472, + "loss": 1.7261, + "theoretical_loss": 4.253562644403573, + "tokens_seen": 256114688 + }, + { + "epoch": 0.08, + "learning_rate": 0.0007453560682046138, + "loss": 1.7191, + "theoretical_loss": 4.253432538977816, + "tokens_seen": 256180224 + }, + { + "epoch": 0.08, + "learning_rate": 0.0007453400200601806, + "loss": 1.6398, + "theoretical_loss": 4.253302476148039, + "tokens_seen": 256245760 + }, + { + "epoch": 0.08, + "learning_rate": 0.0007453239719157472, + "loss": 1.7609, + "theoretical_loss": 4.253172455889407, + "tokens_seen": 256311296 + }, + { + "epoch": 0.08, + "learning_rate": 0.000745307923771314, + "loss": 1.7058, + "theoretical_loss": 4.2530424781771075, + "tokens_seen": 256376832 + }, + { + "epoch": 0.08, + "learning_rate": 0.0007452918756268806, + "loss": 1.8598, + "theoretical_loss": 4.252912542986345, + "tokens_seen": 256442368 + }, + { + "epoch": 0.08, + "learning_rate": 0.0007452758274824474, + "loss": 1.592, + "theoretical_loss": 4.2527826502923505, + "tokens_seen": 256507904 + }, + { + "epoch": 0.08, + "learning_rate": 0.000745259779338014, + "loss": 1.7427, + "theoretical_loss": 4.25265280007037, + "tokens_seen": 256573440 + }, + { + "epoch": 0.08, + "learning_rate": 0.0007452437311935808, + "loss": 1.6572, + "theoretical_loss": 4.252522992295676, + "tokens_seen": 256638976 + }, + { + "epoch": 0.08, + "learning_rate": 0.0007452276830491474, + "loss": 1.6466, + "theoretical_loss": 4.252393226943555, + "tokens_seen": 256704512 + }, + { + "epoch": 0.08, + "learning_rate": 0.0007452116349047142, + "loss": 1.6898, + "theoretical_loss": 4.252263503989321, + "tokens_seen": 256770048 + }, + { + "epoch": 0.08, + "learning_rate": 0.0007451955867602808, + "loss": 1.7049, + "theoretical_loss": 4.252133823408303, + "tokens_seen": 256835584 + }, + { + "epoch": 0.08, + "learning_rate": 0.0007451795386158476, + "loss": 1.7661, + "theoretical_loss": 4.2520041851758545, + "tokens_seen": 256901120 + }, + { + "epoch": 0.08, + "learning_rate": 0.0007451634904714142, + "loss": 1.5728, + "theoretical_loss": 4.251874589267347, + "tokens_seen": 256966656 + }, + { + "epoch": 0.08, + "learning_rate": 0.000745147442326981, + "loss": 1.7055, + "theoretical_loss": 4.251745035658176, + "tokens_seen": 257032192 + }, + { + "epoch": 0.08, + "learning_rate": 0.0007451313941825476, + "loss": 1.8156, + "theoretical_loss": 4.251615524323752, + "tokens_seen": 257097728 + }, + { + "epoch": 0.08, + "learning_rate": 0.0007451153460381144, + "loss": 1.8082, + "theoretical_loss": 4.251486055239512, + "tokens_seen": 257163264 + }, + { + "epoch": 0.08, + "learning_rate": 0.000745099297893681, + "loss": 1.8207, + "theoretical_loss": 4.251356628380911, + "tokens_seen": 257228800 + }, + { + "epoch": 0.08, + "learning_rate": 0.0007450832497492478, + "loss": 1.8414, + "theoretical_loss": 4.251227243723424, + "tokens_seen": 257294336 + }, + { + "epoch": 0.08, + "learning_rate": 0.0007450672016048144, + "loss": 1.7233, + "theoretical_loss": 4.251097901242546, + "tokens_seen": 257359872 + }, + { + "epoch": 0.08, + "learning_rate": 0.0007450511534603812, + "loss": 1.8615, + "theoretical_loss": 4.250968600913795, + "tokens_seen": 257425408 + }, + { + "epoch": 0.08, + "learning_rate": 0.0007450351053159478, + "loss": 1.7229, + "theoretical_loss": 4.250839342712707, + "tokens_seen": 257490944 + }, + { + "epoch": 0.08, + "learning_rate": 0.0007450190571715146, + "loss": 1.6302, + "theoretical_loss": 4.250710126614838, + "tokens_seen": 257556480 + }, + { + "epoch": 0.08, + "learning_rate": 0.0007450030090270812, + "loss": 1.8674, + "theoretical_loss": 4.250580952595769, + "tokens_seen": 257622016 + }, + { + "epoch": 0.08, + "learning_rate": 0.000744986960882648, + "loss": 1.7625, + "theoretical_loss": 4.250451820631095, + "tokens_seen": 257687552 + }, + { + "epoch": 0.08, + "learning_rate": 0.0007449709127382146, + "loss": 1.6845, + "theoretical_loss": 4.2503227306964355, + "tokens_seen": 257753088 + }, + { + "epoch": 0.08, + "learning_rate": 0.0007449548645937814, + "loss": 1.743, + "theoretical_loss": 4.25019368276743, + "tokens_seen": 257818624 + }, + { + "epoch": 0.08, + "learning_rate": 0.000744938816449348, + "loss": 1.7327, + "theoretical_loss": 4.250064676819735, + "tokens_seen": 257884160 + }, + { + "epoch": 0.08, + "learning_rate": 0.0007449227683049148, + "loss": 1.6081, + "theoretical_loss": 4.249935712829033, + "tokens_seen": 257949696 + }, + { + "epoch": 0.08, + "learning_rate": 0.0007449067201604814, + "loss": 1.9918, + "theoretical_loss": 4.249806790771022, + "tokens_seen": 258015232 + }, + { + "epoch": 0.08, + "learning_rate": 0.0007448906720160482, + "loss": 1.7628, + "theoretical_loss": 4.249677910621422, + "tokens_seen": 258080768 + }, + { + "epoch": 0.08, + "learning_rate": 0.0007448746238716148, + "loss": 1.8406, + "theoretical_loss": 4.249549072355974, + "tokens_seen": 258146304 + }, + { + "epoch": 0.08, + "learning_rate": 0.0007448585757271816, + "loss": 1.7548, + "theoretical_loss": 4.2494202759504365, + "tokens_seen": 258211840 + }, + { + "epoch": 0.08, + "learning_rate": 0.0007448425275827482, + "loss": 1.7728, + "theoretical_loss": 4.249291521380591, + "tokens_seen": 258277376 + }, + { + "epoch": 0.08, + "learning_rate": 0.000744826479438315, + "loss": 1.8053, + "theoretical_loss": 4.249162808622237, + "tokens_seen": 258342912 + }, + { + "epoch": 0.08, + "learning_rate": 0.0007448104312938816, + "loss": 1.885, + "theoretical_loss": 4.249034137651197, + "tokens_seen": 258408448 + }, + { + "epoch": 0.08, + "learning_rate": 0.0007447943831494484, + "loss": 1.7164, + "theoretical_loss": 4.248905508443311, + "tokens_seen": 258473984 + }, + { + "epoch": 0.08, + "learning_rate": 0.000744778335005015, + "loss": 1.7486, + "theoretical_loss": 4.248776920974439, + "tokens_seen": 258539520 + }, + { + "epoch": 0.08, + "learning_rate": 0.0007447622868605818, + "loss": 1.7495, + "theoretical_loss": 4.248648375220464, + "tokens_seen": 258605056 + }, + { + "epoch": 0.08, + "learning_rate": 0.0007447462387161484, + "loss": 1.715, + "theoretical_loss": 4.248519871157285, + "tokens_seen": 258670592 + }, + { + "epoch": 0.08, + "learning_rate": 0.0007447301905717152, + "loss": 1.8634, + "theoretical_loss": 4.248391408760824, + "tokens_seen": 258736128 + }, + { + "epoch": 0.08, + "learning_rate": 0.0007447141424272819, + "loss": 1.6993, + "theoretical_loss": 4.248262988007022, + "tokens_seen": 258801664 + }, + { + "epoch": 0.08, + "objective/train/docs_used": 100739, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.6957049369812012, + "objective/train/theoretical_loss": 4.248134608871841, + "objective/train/tokens_used": 279327200, + "theoretical_loss": 4.248134608871841, + "tokens_seen": 258867200 + }, + { + "epoch": 0.08, + "learning_rate": 0.0007446980942828486, + "loss": 1.8533, + "theoretical_loss": 4.248134608871841, + "tokens_seen": 258867200 + }, + { + "epoch": 0.08, + "learning_rate": 0.0007446820461384153, + "loss": 1.8847, + "theoretical_loss": 4.24800627133126, + "tokens_seen": 258932736 + }, + { + "epoch": 0.08, + "learning_rate": 0.000744665997993982, + "loss": 1.9464, + "theoretical_loss": 4.247877975361281, + "tokens_seen": 258998272 + }, + { + "epoch": 0.08, + "learning_rate": 0.0007446499498495487, + "loss": 1.7773, + "theoretical_loss": 4.247749720937925, + "tokens_seen": 259063808 + }, + { + "epoch": 0.08, + "learning_rate": 0.0007446339017051154, + "loss": 1.8512, + "theoretical_loss": 4.247621508037232, + "tokens_seen": 259129344 + }, + { + "epoch": 0.08, + "learning_rate": 0.0007446178535606821, + "loss": 1.8379, + "theoretical_loss": 4.247493336635264, + "tokens_seen": 259194880 + }, + { + "epoch": 0.08, + "learning_rate": 0.0007446018054162488, + "loss": 1.7633, + "theoretical_loss": 4.2473652067081, + "tokens_seen": 259260416 + }, + { + "epoch": 0.08, + "learning_rate": 0.0007445857572718155, + "loss": 1.7864, + "theoretical_loss": 4.247237118231841, + "tokens_seen": 259325952 + }, + { + "epoch": 0.08, + "learning_rate": 0.0007445697091273822, + "loss": 1.8378, + "theoretical_loss": 4.247109071182607, + "tokens_seen": 259391488 + }, + { + "epoch": 0.08, + "learning_rate": 0.0007445536609829489, + "loss": 1.8766, + "theoretical_loss": 4.246981065536538, + "tokens_seen": 259457024 + }, + { + "epoch": 0.08, + "learning_rate": 0.0007445376128385156, + "loss": 1.89, + "theoretical_loss": 4.246853101269793, + "tokens_seen": 259522560 + }, + { + "epoch": 0.08, + "learning_rate": 0.0007445215646940823, + "loss": 2.0185, + "theoretical_loss": 4.2467251783585525, + "tokens_seen": 259588096 + }, + { + "epoch": 0.08, + "learning_rate": 0.000744505516549649, + "loss": 1.8617, + "theoretical_loss": 4.246597296779014, + "tokens_seen": 259653632 + }, + { + "epoch": 0.08, + "learning_rate": 0.0007444894684052157, + "loss": 1.896, + "theoretical_loss": 4.246469456507398, + "tokens_seen": 259719168 + }, + { + "epoch": 0.08, + "learning_rate": 0.0007444734202607824, + "loss": 1.847, + "theoretical_loss": 4.246341657519943, + "tokens_seen": 259784704 + }, + { + "epoch": 0.08, + "learning_rate": 0.0007444573721163491, + "loss": 1.8992, + "theoretical_loss": 4.246213899792905, + "tokens_seen": 259850240 + }, + { + "epoch": 0.08, + "learning_rate": 0.0007444413239719158, + "loss": 1.8374, + "theoretical_loss": 4.246086183302564, + "tokens_seen": 259915776 + }, + { + "epoch": 0.08, + "learning_rate": 0.0007444252758274825, + "loss": 1.6729, + "theoretical_loss": 4.2459585080252165, + "tokens_seen": 259981312 + }, + { + "epoch": 0.08, + "learning_rate": 0.0007444092276830492, + "loss": 1.7831, + "theoretical_loss": 4.245830873937179, + "tokens_seen": 260046848 + }, + { + "epoch": 0.08, + "learning_rate": 0.000744393179538616, + "loss": 1.9251, + "theoretical_loss": 4.24570328101479, + "tokens_seen": 260112384 + }, + { + "epoch": 0.08, + "learning_rate": 0.0007443771313941826, + "loss": 1.8155, + "theoretical_loss": 4.245575729234403, + "tokens_seen": 260177920 + }, + { + "epoch": 0.08, + "learning_rate": 0.0007443610832497493, + "loss": 1.9117, + "theoretical_loss": 4.245448218572396, + "tokens_seen": 260243456 + }, + { + "epoch": 0.08, + "learning_rate": 0.000744345035105316, + "loss": 1.8904, + "theoretical_loss": 4.245320749005163, + "tokens_seen": 260308992 + }, + { + "epoch": 0.08, + "learning_rate": 0.0007443289869608828, + "loss": 1.7424, + "theoretical_loss": 4.245193320509118, + "tokens_seen": 260374528 + }, + { + "epoch": 0.08, + "learning_rate": 0.0007443129388164494, + "loss": 1.8501, + "theoretical_loss": 4.245065933060697, + "tokens_seen": 260440064 + }, + { + "epoch": 0.08, + "learning_rate": 0.0007442968906720162, + "loss": 1.8033, + "theoretical_loss": 4.244938586636351, + "tokens_seen": 260505600 + }, + { + "epoch": 0.08, + "learning_rate": 0.0007442808425275828, + "loss": 1.8644, + "theoretical_loss": 4.244811281212556, + "tokens_seen": 260571136 + }, + { + "epoch": 0.08, + "learning_rate": 0.0007442647943831496, + "loss": 1.7738, + "theoretical_loss": 4.244684016765802, + "tokens_seen": 260636672 + }, + { + "epoch": 0.08, + "learning_rate": 0.0007442487462387162, + "loss": 1.8903, + "theoretical_loss": 4.2445567932726025, + "tokens_seen": 260702208 + }, + { + "epoch": 0.08, + "learning_rate": 0.000744232698094283, + "loss": 1.9099, + "theoretical_loss": 4.244429610709488, + "tokens_seen": 260767744 + }, + { + "epoch": 0.08, + "learning_rate": 0.0007442166499498496, + "loss": 1.7934, + "theoretical_loss": 4.244302469053007, + "tokens_seen": 260833280 + }, + { + "epoch": 0.08, + "learning_rate": 0.0007442006018054164, + "loss": 1.7927, + "theoretical_loss": 4.244175368279733, + "tokens_seen": 260898816 + }, + { + "epoch": 0.08, + "learning_rate": 0.000744184553660983, + "loss": 1.6612, + "theoretical_loss": 4.244048308366253, + "tokens_seen": 260964352 + }, + { + "epoch": 0.08, + "learning_rate": 0.0007441685055165498, + "loss": 1.6762, + "theoretical_loss": 4.243921289289175, + "tokens_seen": 261029888 + }, + { + "epoch": 0.08, + "learning_rate": 0.0007441524573721164, + "loss": 1.7418, + "theoretical_loss": 4.243794311025128, + "tokens_seen": 261095424 + }, + { + "epoch": 0.08, + "learning_rate": 0.0007441364092276832, + "loss": 1.8505, + "theoretical_loss": 4.243667373550757, + "tokens_seen": 261160960 + }, + { + "epoch": 0.08, + "learning_rate": 0.0007441203610832498, + "loss": 1.8139, + "theoretical_loss": 4.24354047684273, + "tokens_seen": 261226496 + }, + { + "epoch": 0.08, + "learning_rate": 0.0007441043129388166, + "loss": 1.848, + "theoretical_loss": 4.243413620877732, + "tokens_seen": 261292032 + }, + { + "epoch": 0.08, + "learning_rate": 0.0007440882647943832, + "loss": 1.6818, + "theoretical_loss": 4.243286805632467, + "tokens_seen": 261357568 + }, + { + "epoch": 0.08, + "learning_rate": 0.0007440722166499499, + "loss": 1.7965, + "theoretical_loss": 4.243160031083658, + "tokens_seen": 261423104 + }, + { + "epoch": 0.08, + "learning_rate": 0.0007440561685055166, + "loss": 1.911, + "theoretical_loss": 4.24303329720805, + "tokens_seen": 261488640 + }, + { + "epoch": 0.08, + "learning_rate": 0.0007440401203610833, + "loss": 1.9146, + "theoretical_loss": 4.242906603982402, + "tokens_seen": 261554176 + }, + { + "epoch": 0.08, + "learning_rate": 0.00074402407221665, + "loss": 1.8526, + "theoretical_loss": 4.242779951383497, + "tokens_seen": 261619712 + }, + { + "epoch": 0.08, + "learning_rate": 0.0007440080240722167, + "loss": 1.7957, + "theoretical_loss": 4.242653339388135, + "tokens_seen": 261685248 + }, + { + "epoch": 0.08, + "learning_rate": 0.0007439919759277834, + "loss": 1.8026, + "theoretical_loss": 4.242526767973134, + "tokens_seen": 261750784 + }, + { + "epoch": 0.08, + "learning_rate": 0.0007439759277833501, + "loss": 1.8638, + "theoretical_loss": 4.242400237115333, + "tokens_seen": 261816320 + }, + { + "epoch": 0.08, + "learning_rate": 0.0007439598796389168, + "loss": 1.7631, + "theoretical_loss": 4.242273746791589, + "tokens_seen": 261881856 + }, + { + "epoch": 0.08, + "learning_rate": 0.0007439438314944835, + "loss": 1.7235, + "theoretical_loss": 4.242147296978778, + "tokens_seen": 261947392 + }, + { + "epoch": 0.08, + "learning_rate": 0.0007439277833500501, + "loss": 1.6729, + "theoretical_loss": 4.242020887653795, + "tokens_seen": 262012928 + }, + { + "epoch": 0.08, + "learning_rate": 0.0007439117352056169, + "loss": 1.648, + "theoretical_loss": 4.241894518793555, + "tokens_seen": 262078464 + }, + { + "epoch": 0.08, + "objective/train/docs_used": 101934, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.9144368171691895, + "objective/train/theoretical_loss": 4.24176819037499, + "objective/train/tokens_used": 282604000, + "theoretical_loss": 4.24176819037499, + "tokens_seen": 262144000 + }, + { + "epoch": 0.08, + "learning_rate": 0.0007438956870611835, + "loss": 1.8245, + "theoretical_loss": 4.24176819037499, + "tokens_seen": 262144000 + }, + { + "epoch": 0.08, + "learning_rate": 0.0007438796389167503, + "loss": 1.7725, + "theoretical_loss": 4.2416419023750525, + "tokens_seen": 262209536 + }, + { + "epoch": 0.08, + "learning_rate": 0.0007438635907723169, + "loss": 1.6837, + "theoretical_loss": 4.241515654770713, + "tokens_seen": 262275072 + }, + { + "epoch": 0.08, + "learning_rate": 0.0007438475426278837, + "loss": 1.7756, + "theoretical_loss": 4.241389447538961, + "tokens_seen": 262340608 + }, + { + "epoch": 0.08, + "learning_rate": 0.0007438314944834503, + "loss": 1.7857, + "theoretical_loss": 4.241263280656806, + "tokens_seen": 262406144 + }, + { + "epoch": 0.08, + "learning_rate": 0.0007438154463390171, + "loss": 1.7803, + "theoretical_loss": 4.241137154101275, + "tokens_seen": 262471680 + }, + { + "epoch": 0.08, + "learning_rate": 0.0007437993981945837, + "loss": 1.8726, + "theoretical_loss": 4.2410110678494135, + "tokens_seen": 262537216 + }, + { + "epoch": 0.08, + "learning_rate": 0.0007437833500501505, + "loss": 1.675, + "theoretical_loss": 4.240885021878286, + "tokens_seen": 262602752 + }, + { + "epoch": 0.08, + "learning_rate": 0.0007437673019057171, + "loss": 1.8714, + "theoretical_loss": 4.240759016164978, + "tokens_seen": 262668288 + }, + { + "epoch": 0.08, + "learning_rate": 0.0007437512537612839, + "loss": 1.7747, + "theoretical_loss": 4.240633050686592, + "tokens_seen": 262733824 + }, + { + "epoch": 0.08, + "learning_rate": 0.0007437352056168505, + "loss": 1.7188, + "theoretical_loss": 4.2405071254202475, + "tokens_seen": 262799360 + }, + { + "epoch": 0.08, + "learning_rate": 0.0007437191574724173, + "loss": 1.7428, + "theoretical_loss": 4.240381240343085, + "tokens_seen": 262864896 + }, + { + "epoch": 0.08, + "learning_rate": 0.0007437031093279839, + "loss": 1.6526, + "theoretical_loss": 4.240255395432264, + "tokens_seen": 262930432 + }, + { + "epoch": 0.08, + "learning_rate": 0.0007436870611835507, + "loss": 1.6399, + "theoretical_loss": 4.240129590664961, + "tokens_seen": 262995968 + }, + { + "epoch": 0.08, + "learning_rate": 0.0007436710130391173, + "loss": 1.7486, + "theoretical_loss": 4.240003826018373, + "tokens_seen": 263061504 + }, + { + "epoch": 0.08, + "learning_rate": 0.0007436549648946841, + "loss": 1.6966, + "theoretical_loss": 4.239878101469713, + "tokens_seen": 263127040 + }, + { + "epoch": 0.08, + "learning_rate": 0.0007436389167502507, + "loss": 1.7557, + "theoretical_loss": 4.239752416996215, + "tokens_seen": 263192576 + }, + { + "epoch": 0.08, + "learning_rate": 0.0007436228686058175, + "loss": 1.6558, + "theoretical_loss": 4.239626772575131, + "tokens_seen": 263258112 + }, + { + "epoch": 0.08, + "learning_rate": 0.0007436068204613841, + "loss": 1.861, + "theoretical_loss": 4.239501168183732, + "tokens_seen": 263323648 + }, + { + "epoch": 0.08, + "learning_rate": 0.0007435907723169509, + "loss": 1.7061, + "theoretical_loss": 4.239375603799305, + "tokens_seen": 263389184 + }, + { + "epoch": 0.08, + "learning_rate": 0.0007435747241725175, + "loss": 1.7814, + "theoretical_loss": 4.2392500793991585, + "tokens_seen": 263454720 + }, + { + "epoch": 0.08, + "learning_rate": 0.0007435586760280843, + "loss": 1.8509, + "theoretical_loss": 4.239124594960619, + "tokens_seen": 263520256 + }, + { + "epoch": 0.08, + "learning_rate": 0.000743542627883651, + "loss": 1.8036, + "theoretical_loss": 4.23899915046103, + "tokens_seen": 263585792 + }, + { + "epoch": 0.08, + "learning_rate": 0.0007435265797392177, + "loss": 1.7664, + "theoretical_loss": 4.238873745877756, + "tokens_seen": 263651328 + }, + { + "epoch": 0.08, + "learning_rate": 0.0007435105315947844, + "loss": 1.7624, + "theoretical_loss": 4.238748381188176, + "tokens_seen": 263716864 + }, + { + "epoch": 0.08, + "learning_rate": 0.0007434944834503511, + "loss": 1.772, + "theoretical_loss": 4.238623056369692, + "tokens_seen": 263782400 + }, + { + "epoch": 0.08, + "learning_rate": 0.0007434784353059178, + "loss": 1.803, + "theoretical_loss": 4.238497771399721, + "tokens_seen": 263847936 + }, + { + "epoch": 0.08, + "learning_rate": 0.0007434623871614845, + "loss": 1.7314, + "theoretical_loss": 4.2383725262556995, + "tokens_seen": 263913472 + }, + { + "epoch": 0.08, + "learning_rate": 0.0007434463390170512, + "loss": 1.7749, + "theoretical_loss": 4.238247320915083, + "tokens_seen": 263979008 + }, + { + "epoch": 0.08, + "learning_rate": 0.0007434302908726179, + "loss": 1.7175, + "theoretical_loss": 4.238122155355345, + "tokens_seen": 264044544 + }, + { + "epoch": 0.08, + "learning_rate": 0.0007434142427281846, + "loss": 1.8362, + "theoretical_loss": 4.237997029553977, + "tokens_seen": 264110080 + }, + { + "epoch": 0.08, + "learning_rate": 0.0007433981945837513, + "loss": 1.8471, + "theoretical_loss": 4.237871943488488, + "tokens_seen": 264175616 + }, + { + "epoch": 0.08, + "learning_rate": 0.000743382146439318, + "loss": 1.7321, + "theoretical_loss": 4.237746897136408, + "tokens_seen": 264241152 + }, + { + "epoch": 0.08, + "learning_rate": 0.0007433660982948847, + "loss": 1.7747, + "theoretical_loss": 4.237621890475284, + "tokens_seen": 264306688 + }, + { + "epoch": 0.08, + "learning_rate": 0.0007433500501504514, + "loss": 1.7234, + "theoretical_loss": 4.237496923482678, + "tokens_seen": 264372224 + }, + { + "epoch": 0.08, + "learning_rate": 0.0007433340020060181, + "loss": 1.6787, + "theoretical_loss": 4.237371996136175, + "tokens_seen": 264437760 + }, + { + "epoch": 0.08, + "learning_rate": 0.0007433179538615848, + "loss": 1.7353, + "theoretical_loss": 4.237247108413376, + "tokens_seen": 264503296 + }, + { + "epoch": 0.08, + "learning_rate": 0.0007433019057171515, + "loss": 1.5868, + "theoretical_loss": 4.237122260291901, + "tokens_seen": 264568832 + }, + { + "epoch": 0.08, + "learning_rate": 0.0007432858575727182, + "loss": 1.7438, + "theoretical_loss": 4.236997451749387, + "tokens_seen": 264634368 + }, + { + "epoch": 0.08, + "learning_rate": 0.0007432698094282849, + "loss": 1.6383, + "theoretical_loss": 4.236872682763489, + "tokens_seen": 264699904 + }, + { + "epoch": 0.08, + "learning_rate": 0.0007432537612838516, + "loss": 1.7362, + "theoretical_loss": 4.236747953311884, + "tokens_seen": 264765440 + }, + { + "epoch": 0.08, + "learning_rate": 0.0007432377131394183, + "loss": 1.7981, + "theoretical_loss": 4.236623263372261, + "tokens_seen": 264830976 + }, + { + "epoch": 0.08, + "learning_rate": 0.000743221664994985, + "loss": 1.815, + "theoretical_loss": 4.236498612922331, + "tokens_seen": 264896512 + }, + { + "epoch": 0.08, + "learning_rate": 0.0007432056168505517, + "loss": 1.7597, + "theoretical_loss": 4.236374001939822, + "tokens_seen": 264962048 + }, + { + "epoch": 0.08, + "learning_rate": 0.0007431895687061184, + "loss": 1.8371, + "theoretical_loss": 4.2362494304024825, + "tokens_seen": 265027584 + }, + { + "epoch": 0.08, + "learning_rate": 0.0007431735205616851, + "loss": 1.8106, + "theoretical_loss": 4.236124898288074, + "tokens_seen": 265093120 + }, + { + "epoch": 0.08, + "learning_rate": 0.0007431574724172518, + "loss": 1.7237, + "theoretical_loss": 4.236000405574382, + "tokens_seen": 265158656 + }, + { + "epoch": 0.08, + "learning_rate": 0.0007431414242728185, + "loss": 1.7553, + "theoretical_loss": 4.235875952239204, + "tokens_seen": 265224192 + }, + { + "epoch": 0.08, + "learning_rate": 0.0007431253761283852, + "loss": 1.6739, + "theoretical_loss": 4.235751538260359, + "tokens_seen": 265289728 + }, + { + "epoch": 0.08, + "learning_rate": 0.0007431093279839519, + "loss": 1.7713, + "theoretical_loss": 4.2356271636156855, + "tokens_seen": 265355264 + }, + { + "epoch": 0.08, + "objective/train/docs_used": 103026, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.5847630500793457, + "objective/train/theoretical_loss": 4.235502828283035, + "objective/train/tokens_used": 285880800, + "theoretical_loss": 4.235502828283035, + "tokens_seen": 265420800 + }, + { + "epoch": 0.08, + "learning_rate": 0.0007430932798395186, + "loss": 1.741, + "theoretical_loss": 4.235502828283035, + "tokens_seen": 265420800 + }, + { + "epoch": 0.08, + "learning_rate": 0.0007430772316950853, + "loss": 1.826, + "theoretical_loss": 4.235378532240283, + "tokens_seen": 265486336 + }, + { + "epoch": 0.08, + "learning_rate": 0.000743061183550652, + "loss": 1.8401, + "theoretical_loss": 4.235254275465317, + "tokens_seen": 265551872 + }, + { + "epoch": 0.08, + "learning_rate": 0.0007430451354062187, + "loss": 1.6363, + "theoretical_loss": 4.235130057936045, + "tokens_seen": 265617408 + }, + { + "epoch": 0.08, + "learning_rate": 0.0007430290872617854, + "loss": 1.6438, + "theoretical_loss": 4.235005879630395, + "tokens_seen": 265682944 + }, + { + "epoch": 0.08, + "learning_rate": 0.0007430130391173521, + "loss": 1.7469, + "theoretical_loss": 4.23488174052631, + "tokens_seen": 265748480 + }, + { + "epoch": 0.08, + "learning_rate": 0.0007429969909729188, + "loss": 1.6978, + "theoretical_loss": 4.234757640601751, + "tokens_seen": 265814016 + }, + { + "epoch": 0.08, + "learning_rate": 0.0007429809428284855, + "loss": 1.7184, + "theoretical_loss": 4.234633579834698, + "tokens_seen": 265879552 + }, + { + "epoch": 0.08, + "learning_rate": 0.0007429648946840523, + "loss": 1.6711, + "theoretical_loss": 4.234509558203149, + "tokens_seen": 265945088 + }, + { + "epoch": 0.08, + "learning_rate": 0.0007429488465396189, + "loss": 1.773, + "theoretical_loss": 4.234385575685116, + "tokens_seen": 266010624 + }, + { + "epoch": 0.08, + "learning_rate": 0.0007429327983951857, + "loss": 1.7103, + "theoretical_loss": 4.234261632258636, + "tokens_seen": 266076160 + }, + { + "epoch": 0.08, + "learning_rate": 0.0007429167502507523, + "loss": 1.7284, + "theoretical_loss": 4.234137727901757, + "tokens_seen": 266141696 + }, + { + "epoch": 0.08, + "learning_rate": 0.000742900702106319, + "loss": 1.8367, + "theoretical_loss": 4.234013862592548, + "tokens_seen": 266207232 + }, + { + "epoch": 0.08, + "learning_rate": 0.0007428846539618857, + "loss": 1.7282, + "theoretical_loss": 4.233890036309095, + "tokens_seen": 266272768 + }, + { + "epoch": 0.08, + "learning_rate": 0.0007428686058174525, + "loss": 1.6102, + "theoretical_loss": 4.233766249029501, + "tokens_seen": 266338304 + }, + { + "epoch": 0.08, + "learning_rate": 0.0007428525576730191, + "loss": 1.7502, + "theoretical_loss": 4.233642500731889, + "tokens_seen": 266403840 + }, + { + "epoch": 0.08, + "learning_rate": 0.0007428365095285859, + "loss": 1.8406, + "theoretical_loss": 4.233518791394396, + "tokens_seen": 266469376 + }, + { + "epoch": 0.08, + "learning_rate": 0.0007428204613841525, + "loss": 1.8117, + "theoretical_loss": 4.233395120995181, + "tokens_seen": 266534912 + }, + { + "epoch": 0.08, + "learning_rate": 0.0007428044132397193, + "loss": 1.7561, + "theoretical_loss": 4.233271489512416, + "tokens_seen": 266600448 + }, + { + "epoch": 0.08, + "learning_rate": 0.0007427883650952859, + "loss": 1.7424, + "theoretical_loss": 4.233147896924294, + "tokens_seen": 266665984 + }, + { + "epoch": 0.08, + "learning_rate": 0.0007427723169508527, + "loss": 1.8128, + "theoretical_loss": 4.233024343209023, + "tokens_seen": 266731520 + }, + { + "epoch": 0.08, + "learning_rate": 0.0007427562688064193, + "loss": 1.7449, + "theoretical_loss": 4.232900828344833, + "tokens_seen": 266797056 + }, + { + "epoch": 0.08, + "learning_rate": 0.0007427402206619861, + "loss": 1.6655, + "theoretical_loss": 4.232777352309966, + "tokens_seen": 266862592 + }, + { + "epoch": 0.08, + "learning_rate": 0.0007427241725175527, + "loss": 1.7741, + "theoretical_loss": 4.232653915082684, + "tokens_seen": 266928128 + }, + { + "epoch": 0.08, + "learning_rate": 0.0007427081243731195, + "loss": 1.7288, + "theoretical_loss": 4.232530516641267, + "tokens_seen": 266993664 + }, + { + "epoch": 0.08, + "learning_rate": 0.0007426920762286861, + "loss": 1.7062, + "theoretical_loss": 4.232407156964014, + "tokens_seen": 267059200 + }, + { + "epoch": 0.08, + "learning_rate": 0.0007426760280842528, + "loss": 1.6898, + "theoretical_loss": 4.232283836029237, + "tokens_seen": 267124736 + }, + { + "epoch": 0.08, + "learning_rate": 0.0007426599799398195, + "loss": 1.6657, + "theoretical_loss": 4.232160553815269, + "tokens_seen": 267190272 + }, + { + "epoch": 0.08, + "learning_rate": 0.0007426439317953862, + "loss": 1.7545, + "theoretical_loss": 4.232037310300458, + "tokens_seen": 267255808 + }, + { + "epoch": 0.08, + "learning_rate": 0.0007426278836509529, + "loss": 1.6824, + "theoretical_loss": 4.231914105463173, + "tokens_seen": 267321344 + }, + { + "epoch": 0.08, + "learning_rate": 0.0007426118355065196, + "loss": 1.8148, + "theoretical_loss": 4.231790939281797, + "tokens_seen": 267386880 + }, + { + "epoch": 0.08, + "learning_rate": 0.0007425957873620863, + "loss": 1.7664, + "theoretical_loss": 4.231667811734732, + "tokens_seen": 267452416 + }, + { + "epoch": 0.08, + "learning_rate": 0.000742579739217653, + "loss": 1.8105, + "theoretical_loss": 4.231544722800395, + "tokens_seen": 267517952 + }, + { + "epoch": 0.08, + "learning_rate": 0.0007425636910732197, + "loss": 1.6743, + "theoretical_loss": 4.231421672457225, + "tokens_seen": 267583488 + }, + { + "epoch": 0.08, + "learning_rate": 0.0007425476429287864, + "loss": 1.8326, + "theoretical_loss": 4.231298660683675, + "tokens_seen": 267649024 + }, + { + "epoch": 0.08, + "learning_rate": 0.000742531594784353, + "loss": 1.6611, + "theoretical_loss": 4.231175687458214, + "tokens_seen": 267714560 + }, + { + "epoch": 0.08, + "learning_rate": 0.0007425155466399198, + "loss": 1.7425, + "theoretical_loss": 4.231052752759332, + "tokens_seen": 267780096 + }, + { + "epoch": 0.08, + "learning_rate": 0.0007424994984954864, + "loss": 1.6332, + "theoretical_loss": 4.230929856565535, + "tokens_seen": 267845632 + }, + { + "epoch": 0.08, + "learning_rate": 0.0007424834503510532, + "loss": 1.6871, + "theoretical_loss": 4.230806998855343, + "tokens_seen": 267911168 + }, + { + "epoch": 0.08, + "learning_rate": 0.0007424674022066198, + "loss": 1.7537, + "theoretical_loss": 4.230684179607297, + "tokens_seen": 267976704 + }, + { + "epoch": 0.08, + "learning_rate": 0.0007424513540621866, + "loss": 1.7232, + "theoretical_loss": 4.2305613987999555, + "tokens_seen": 268042240 + }, + { + "epoch": 0.08, + "learning_rate": 0.0007424353059177532, + "loss": 1.6799, + "theoretical_loss": 4.230438656411891, + "tokens_seen": 268107776 + }, + { + "epoch": 0.08, + "learning_rate": 0.00074241925777332, + "loss": 1.6915, + "theoretical_loss": 4.230315952421697, + "tokens_seen": 268173312 + }, + { + "epoch": 0.08, + "learning_rate": 0.0007424032096288866, + "loss": 1.7154, + "theoretical_loss": 4.230193286807981, + "tokens_seen": 268238848 + }, + { + "epoch": 0.08, + "learning_rate": 0.0007423871614844534, + "loss": 1.7564, + "theoretical_loss": 4.230070659549368, + "tokens_seen": 268304384 + }, + { + "epoch": 0.08, + "learning_rate": 0.00074237111334002, + "loss": 1.7275, + "theoretical_loss": 4.229948070624502, + "tokens_seen": 268369920 + }, + { + "epoch": 0.08, + "learning_rate": 0.0007423550651955868, + "loss": 1.7046, + "theoretical_loss": 4.229825520012042, + "tokens_seen": 268435456 + }, + { + "epoch": 0.08, + "learning_rate": 0.0007423390170511535, + "loss": 1.8055, + "theoretical_loss": 4.2297030076906665, + "tokens_seen": 268500992 + }, + { + "epoch": 0.08, + "learning_rate": 0.0007423229689067202, + "loss": 1.7716, + "theoretical_loss": 4.229580533639068, + "tokens_seen": 268566528 + }, + { + "epoch": 0.08, + "learning_rate": 0.0007423069207622869, + "loss": 1.6983, + "theoretical_loss": 4.229458097835959, + "tokens_seen": 268632064 + }, + { + "epoch": 0.08, + "objective/train/docs_used": 104173, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.5929882526397705, + "objective/train/theoretical_loss": 4.2293357002600676, + "objective/train/tokens_used": 289157600, + "theoretical_loss": 4.2293357002600676, + "tokens_seen": 268697600 + }, + { + "epoch": 0.08, + "learning_rate": 0.0007422908726178536, + "loss": 1.6047, + "theoretical_loss": 4.2293357002600676, + "tokens_seen": 268697600 + }, + { + "epoch": 0.08, + "learning_rate": 0.0007422748244734203, + "loss": 1.6948, + "theoretical_loss": 4.229213340890139, + "tokens_seen": 268763136 + }, + { + "epoch": 0.08, + "learning_rate": 0.000742258776328987, + "loss": 1.7341, + "theoretical_loss": 4.229091019704935, + "tokens_seen": 268828672 + }, + { + "epoch": 0.08, + "learning_rate": 0.0007422427281845537, + "loss": 1.8156, + "theoretical_loss": 4.2289687366832345, + "tokens_seen": 268894208 + }, + { + "epoch": 0.08, + "learning_rate": 0.0007422266800401204, + "loss": 1.7336, + "theoretical_loss": 4.228846491803835, + "tokens_seen": 268959744 + }, + { + "epoch": 0.08, + "learning_rate": 0.0007422106318956871, + "loss": 1.7664, + "theoretical_loss": 4.228724285045549, + "tokens_seen": 269025280 + }, + { + "epoch": 0.08, + "learning_rate": 0.0007421945837512538, + "loss": 1.6748, + "theoretical_loss": 4.228602116387206, + "tokens_seen": 269090816 + }, + { + "epoch": 0.08, + "learning_rate": 0.0007421785356068205, + "loss": 1.7264, + "theoretical_loss": 4.228479985807653, + "tokens_seen": 269156352 + }, + { + "epoch": 0.08, + "learning_rate": 0.0007421624874623872, + "loss": 1.5306, + "theoretical_loss": 4.228357893285755, + "tokens_seen": 269221888 + }, + { + "epoch": 0.08, + "learning_rate": 0.0007421464393179539, + "loss": 1.7058, + "theoretical_loss": 4.2282358388003916, + "tokens_seen": 269287424 + }, + { + "epoch": 0.08, + "learning_rate": 0.0007421303911735206, + "loss": 1.641, + "theoretical_loss": 4.228113822330461, + "tokens_seen": 269352960 + }, + { + "epoch": 0.08, + "learning_rate": 0.0007421143430290873, + "loss": 1.5981, + "theoretical_loss": 4.227991843854878, + "tokens_seen": 269418496 + }, + { + "epoch": 0.08, + "learning_rate": 0.000742098294884654, + "loss": 1.6707, + "theoretical_loss": 4.227869903352572, + "tokens_seen": 269484032 + }, + { + "epoch": 0.08, + "learning_rate": 0.0007420822467402207, + "loss": 1.6471, + "theoretical_loss": 4.227748000802493, + "tokens_seen": 269549568 + }, + { + "epoch": 0.08, + "learning_rate": 0.0007420661985957874, + "loss": 1.7609, + "theoretical_loss": 4.227626136183606, + "tokens_seen": 269615104 + }, + { + "epoch": 0.08, + "learning_rate": 0.0007420501504513541, + "loss": 1.7246, + "theoretical_loss": 4.227504309474892, + "tokens_seen": 269680640 + }, + { + "epoch": 0.08, + "learning_rate": 0.0007420341023069208, + "loss": 1.717, + "theoretical_loss": 4.227382520655348, + "tokens_seen": 269746176 + }, + { + "epoch": 0.08, + "learning_rate": 0.0007420180541624875, + "loss": 1.7684, + "theoretical_loss": 4.227260769703991, + "tokens_seen": 269811712 + }, + { + "epoch": 0.08, + "learning_rate": 0.0007420020060180542, + "loss": 1.6063, + "theoretical_loss": 4.227139056599851, + "tokens_seen": 269877248 + }, + { + "epoch": 0.08, + "learning_rate": 0.0007419859578736209, + "loss": 1.6538, + "theoretical_loss": 4.2270173813219785, + "tokens_seen": 269942784 + }, + { + "epoch": 0.08, + "learning_rate": 0.0007419699097291876, + "loss": 1.6595, + "theoretical_loss": 4.226895743849437, + "tokens_seen": 270008320 + }, + { + "epoch": 0.08, + "learning_rate": 0.0007419538615847543, + "loss": 1.7563, + "theoretical_loss": 4.226774144161309, + "tokens_seen": 270073856 + }, + { + "epoch": 0.08, + "learning_rate": 0.000741937813440321, + "loss": 1.7363, + "theoretical_loss": 4.226652582236694, + "tokens_seen": 270139392 + }, + { + "epoch": 0.08, + "learning_rate": 0.0007419217652958877, + "loss": 1.7514, + "theoretical_loss": 4.226531058054706, + "tokens_seen": 270204928 + }, + { + "epoch": 0.08, + "learning_rate": 0.0007419057171514544, + "loss": 1.7576, + "theoretical_loss": 4.226409571594477, + "tokens_seen": 270270464 + }, + { + "epoch": 0.08, + "learning_rate": 0.0007418896690070211, + "loss": 1.7416, + "theoretical_loss": 4.226288122835156, + "tokens_seen": 270336000 + }, + { + "epoch": 0.08, + "learning_rate": 0.0007418736208625878, + "loss": 1.6483, + "theoretical_loss": 4.226166711755906, + "tokens_seen": 270401536 + }, + { + "epoch": 0.08, + "learning_rate": 0.0007418575727181545, + "loss": 1.653, + "theoretical_loss": 4.2260453383359105, + "tokens_seen": 270467072 + }, + { + "epoch": 0.08, + "learning_rate": 0.0007418415245737212, + "loss": 1.8298, + "theoretical_loss": 4.2259240025543665, + "tokens_seen": 270532608 + }, + { + "epoch": 0.08, + "learning_rate": 0.0007418254764292879, + "loss": 1.7826, + "theoretical_loss": 4.225802704390488, + "tokens_seen": 270598144 + }, + { + "epoch": 0.08, + "learning_rate": 0.0007418094282848546, + "loss": 1.6091, + "theoretical_loss": 4.2256814438235075, + "tokens_seen": 270663680 + }, + { + "epoch": 0.08, + "learning_rate": 0.0007417933801404213, + "loss": 1.6292, + "theoretical_loss": 4.225560220832672, + "tokens_seen": 270729216 + }, + { + "epoch": 0.08, + "learning_rate": 0.000741777331995988, + "loss": 1.5817, + "theoretical_loss": 4.225439035397246, + "tokens_seen": 270794752 + }, + { + "epoch": 0.08, + "learning_rate": 0.0007417612838515547, + "loss": 1.6933, + "theoretical_loss": 4.225317887496508, + "tokens_seen": 270860288 + }, + { + "epoch": 0.08, + "learning_rate": 0.0007417452357071214, + "loss": 1.6237, + "theoretical_loss": 4.2251967771097565, + "tokens_seen": 270925824 + }, + { + "epoch": 0.08, + "learning_rate": 0.0007417291875626881, + "loss": 1.7768, + "theoretical_loss": 4.225075704216305, + "tokens_seen": 270991360 + }, + { + "epoch": 0.08, + "learning_rate": 0.0007417131394182548, + "loss": 1.6164, + "theoretical_loss": 4.224954668795483, + "tokens_seen": 271056896 + }, + { + "epoch": 0.08, + "learning_rate": 0.0007416970912738215, + "loss": 1.7091, + "theoretical_loss": 4.224833670826636, + "tokens_seen": 271122432 + }, + { + "epoch": 0.08, + "learning_rate": 0.0007416810431293882, + "loss": 1.7543, + "theoretical_loss": 4.224712710289127, + "tokens_seen": 271187968 + }, + { + "epoch": 0.08, + "learning_rate": 0.000741664994984955, + "loss": 1.79, + "theoretical_loss": 4.2245917871623355, + "tokens_seen": 271253504 + }, + { + "epoch": 0.08, + "learning_rate": 0.0007416489468405216, + "loss": 1.6359, + "theoretical_loss": 4.224470901425655, + "tokens_seen": 271319040 + }, + { + "epoch": 0.08, + "learning_rate": 0.0007416328986960883, + "loss": 1.8391, + "theoretical_loss": 4.224350053058499, + "tokens_seen": 271384576 + }, + { + "epoch": 0.08, + "learning_rate": 0.000741616850551655, + "loss": 1.7596, + "theoretical_loss": 4.224229242040295, + "tokens_seen": 271450112 + }, + { + "epoch": 0.08, + "learning_rate": 0.0007416008024072218, + "loss": 1.691, + "theoretical_loss": 4.224108468350485, + "tokens_seen": 271515648 + }, + { + "epoch": 0.08, + "learning_rate": 0.0007415847542627884, + "loss": 1.8215, + "theoretical_loss": 4.223987731968532, + "tokens_seen": 271581184 + }, + { + "epoch": 0.08, + "learning_rate": 0.0007415687061183552, + "loss": 1.8327, + "theoretical_loss": 4.223867032873912, + "tokens_seen": 271646720 + }, + { + "epoch": 0.08, + "learning_rate": 0.0007415526579739218, + "loss": 1.7407, + "theoretical_loss": 4.223746371046116, + "tokens_seen": 271712256 + }, + { + "epoch": 0.08, + "learning_rate": 0.0007415366098294886, + "loss": 1.6553, + "theoretical_loss": 4.223625746464656, + "tokens_seen": 271777792 + }, + { + "epoch": 0.08, + "learning_rate": 0.0007415205616850552, + "loss": 1.8288, + "theoretical_loss": 4.223505159109054, + "tokens_seen": 271843328 + }, + { + "epoch": 0.08, + "learning_rate": 0.000741504513540622, + "loss": 1.6589, + "theoretical_loss": 4.223384608958855, + "tokens_seen": 271908864 + }, + { + "epoch": 0.08, + "objective/train/docs_used": 105443, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.6452257633209229, + "objective/train/theoretical_loss": 4.223264095993615, + "objective/train/tokens_used": 292434400, + "theoretical_loss": 4.223264095993615, + "tokens_seen": 271974400 + }, + { + "epoch": 0.08, + "learning_rate": 0.0007414884653961886, + "loss": 1.7068, + "theoretical_loss": 4.223264095993615, + "tokens_seen": 271974400 + }, + { + "epoch": 0.08, + "learning_rate": 0.0007414724172517554, + "loss": 1.5908, + "theoretical_loss": 4.223143620192907, + "tokens_seen": 272039936 + }, + { + "epoch": 0.08, + "learning_rate": 0.000741456369107322, + "loss": 1.6794, + "theoretical_loss": 4.223023181536323, + "tokens_seen": 272105472 + }, + { + "epoch": 0.08, + "learning_rate": 0.0007414403209628888, + "loss": 1.6682, + "theoretical_loss": 4.2229027800034675, + "tokens_seen": 272171008 + }, + { + "epoch": 0.08, + "learning_rate": 0.0007414242728184554, + "loss": 1.6452, + "theoretical_loss": 4.222782415573962, + "tokens_seen": 272236544 + }, + { + "epoch": 0.08, + "learning_rate": 0.0007414082246740222, + "loss": 1.6629, + "theoretical_loss": 4.222662088227447, + "tokens_seen": 272302080 + }, + { + "epoch": 0.08, + "learning_rate": 0.0007413921765295888, + "loss": 1.815, + "theoretical_loss": 4.222541797943576, + "tokens_seen": 272367616 + }, + { + "epoch": 0.08, + "learning_rate": 0.0007413761283851556, + "loss": 1.7177, + "theoretical_loss": 4.222421544702018, + "tokens_seen": 272433152 + }, + { + "epoch": 0.08, + "learning_rate": 0.0007413600802407222, + "loss": 1.6889, + "theoretical_loss": 4.2223013284824615, + "tokens_seen": 272498688 + }, + { + "epoch": 0.08, + "learning_rate": 0.000741344032096289, + "loss": 1.6472, + "theoretical_loss": 4.222181149264608, + "tokens_seen": 272564224 + }, + { + "epoch": 0.08, + "learning_rate": 0.0007413279839518556, + "loss": 1.6782, + "theoretical_loss": 4.222061007028175, + "tokens_seen": 272629760 + }, + { + "epoch": 0.08, + "learning_rate": 0.0007413119358074224, + "loss": 1.7058, + "theoretical_loss": 4.221940901752899, + "tokens_seen": 272695296 + }, + { + "epoch": 0.08, + "learning_rate": 0.000741295887662989, + "loss": 1.7659, + "theoretical_loss": 4.221820833418529, + "tokens_seen": 272760832 + }, + { + "epoch": 0.08, + "learning_rate": 0.0007412798395185557, + "loss": 1.7125, + "theoretical_loss": 4.221700802004832, + "tokens_seen": 272826368 + }, + { + "epoch": 0.08, + "learning_rate": 0.0007412637913741224, + "loss": 1.8216, + "theoretical_loss": 4.2215808074915895, + "tokens_seen": 272891904 + }, + { + "epoch": 0.08, + "learning_rate": 0.0007412477432296891, + "loss": 1.7992, + "theoretical_loss": 4.221460849858601, + "tokens_seen": 272957440 + }, + { + "epoch": 0.08, + "learning_rate": 0.0007412316950852558, + "loss": 1.7291, + "theoretical_loss": 4.221340929085679, + "tokens_seen": 273022976 + }, + { + "epoch": 0.08, + "learning_rate": 0.0007412156469408225, + "loss": 1.62, + "theoretical_loss": 4.221221045152655, + "tokens_seen": 273088512 + }, + { + "epoch": 0.08, + "learning_rate": 0.0007411995987963892, + "loss": 1.8229, + "theoretical_loss": 4.221101198039375, + "tokens_seen": 273154048 + }, + { + "epoch": 0.08, + "learning_rate": 0.0007411835506519559, + "loss": 1.8089, + "theoretical_loss": 4.2209813877257, + "tokens_seen": 273219584 + }, + { + "epoch": 0.08, + "learning_rate": 0.0007411675025075226, + "loss": 1.6432, + "theoretical_loss": 4.220861614191508, + "tokens_seen": 273285120 + }, + { + "epoch": 0.08, + "learning_rate": 0.0007411514543630893, + "loss": 1.67, + "theoretical_loss": 4.220741877416691, + "tokens_seen": 273350656 + }, + { + "epoch": 0.08, + "learning_rate": 0.000741135406218656, + "loss": 1.6334, + "theoretical_loss": 4.22062217738116, + "tokens_seen": 273416192 + }, + { + "epoch": 0.08, + "learning_rate": 0.0007411193580742227, + "loss": 1.6947, + "theoretical_loss": 4.22050251406484, + "tokens_seen": 273481728 + }, + { + "epoch": 0.08, + "learning_rate": 0.0007411033099297894, + "loss": 1.804, + "theoretical_loss": 4.220382887447671, + "tokens_seen": 273547264 + }, + { + "epoch": 0.08, + "learning_rate": 0.0007410872617853561, + "loss": 1.7807, + "theoretical_loss": 4.22026329750961, + "tokens_seen": 273612800 + }, + { + "epoch": 0.08, + "learning_rate": 0.0007410712136409228, + "loss": 1.6737, + "theoretical_loss": 4.220143744230629, + "tokens_seen": 273678336 + }, + { + "epoch": 0.08, + "learning_rate": 0.0007410551654964895, + "loss": 1.7156, + "theoretical_loss": 4.220024227590717, + "tokens_seen": 273743872 + }, + { + "epoch": 0.08, + "learning_rate": 0.0007410391173520562, + "loss": 1.681, + "theoretical_loss": 4.219904747569877, + "tokens_seen": 273809408 + }, + { + "epoch": 0.08, + "learning_rate": 0.0007410230692076229, + "loss": 1.7757, + "theoretical_loss": 4.219785304148129, + "tokens_seen": 273874944 + }, + { + "epoch": 0.08, + "learning_rate": 0.0007410070210631896, + "loss": 1.7989, + "theoretical_loss": 4.219665897305506, + "tokens_seen": 273940480 + }, + { + "epoch": 0.08, + "learning_rate": 0.0007409909729187563, + "loss": 1.7158, + "theoretical_loss": 4.219546527022062, + "tokens_seen": 274006016 + }, + { + "epoch": 0.08, + "learning_rate": 0.000740974924774323, + "loss": 1.7044, + "theoretical_loss": 4.219427193277861, + "tokens_seen": 274071552 + }, + { + "epoch": 0.08, + "learning_rate": 0.0007409588766298897, + "loss": 1.6948, + "theoretical_loss": 4.219307896052987, + "tokens_seen": 274137088 + }, + { + "epoch": 0.08, + "learning_rate": 0.0007409428284854564, + "loss": 1.7875, + "theoretical_loss": 4.219188635327535, + "tokens_seen": 274202624 + }, + { + "epoch": 0.08, + "learning_rate": 0.0007409267803410231, + "loss": 1.7205, + "theoretical_loss": 4.219069411081622, + "tokens_seen": 274268160 + }, + { + "epoch": 0.08, + "learning_rate": 0.0007409107321965898, + "loss": 1.5975, + "theoretical_loss": 4.218950223295374, + "tokens_seen": 274333696 + }, + { + "epoch": 0.08, + "learning_rate": 0.0007408946840521565, + "loss": 1.6246, + "theoretical_loss": 4.218831071948936, + "tokens_seen": 274399232 + }, + { + "epoch": 0.08, + "learning_rate": 0.0007408786359077232, + "loss": 1.7599, + "theoretical_loss": 4.218711957022469, + "tokens_seen": 274464768 + }, + { + "epoch": 0.08, + "learning_rate": 0.0007408625877632899, + "loss": 1.5931, + "theoretical_loss": 4.218592878496148, + "tokens_seen": 274530304 + }, + { + "epoch": 0.08, + "learning_rate": 0.0007408465396188566, + "loss": 1.7047, + "theoretical_loss": 4.218473836350164, + "tokens_seen": 274595840 + }, + { + "epoch": 0.08, + "learning_rate": 0.0007408304914744233, + "loss": 1.6342, + "theoretical_loss": 4.2183548305647225, + "tokens_seen": 274661376 + }, + { + "epoch": 0.08, + "learning_rate": 0.00074081444332999, + "loss": 1.7218, + "theoretical_loss": 4.218235861120047, + "tokens_seen": 274726912 + }, + { + "epoch": 0.08, + "learning_rate": 0.0007407983951855567, + "loss": 1.8168, + "theoretical_loss": 4.218116927996375, + "tokens_seen": 274792448 + }, + { + "epoch": 0.08, + "learning_rate": 0.0007407823470411234, + "loss": 1.6461, + "theoretical_loss": 4.217998031173959, + "tokens_seen": 274857984 + }, + { + "epoch": 0.08, + "learning_rate": 0.0007407662988966901, + "loss": 1.7019, + "theoretical_loss": 4.217879170633067, + "tokens_seen": 274923520 + }, + { + "epoch": 0.08, + "learning_rate": 0.0007407502507522568, + "loss": 1.7363, + "theoretical_loss": 4.217760346353984, + "tokens_seen": 274989056 + }, + { + "epoch": 0.08, + "learning_rate": 0.0007407342026078235, + "loss": 1.7649, + "theoretical_loss": 4.217641558317008, + "tokens_seen": 275054592 + }, + { + "epoch": 0.08, + "learning_rate": 0.0007407181544633902, + "loss": 1.5996, + "theoretical_loss": 4.217522806502455, + "tokens_seen": 275120128 + }, + { + "epoch": 0.08, + "learning_rate": 0.0007407021063189569, + "loss": 1.6373, + "theoretical_loss": 4.217404090890653, + "tokens_seen": 275185664 + }, + { + "epoch": 0.08, + "objective/train/docs_used": 106703, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.6383416652679443, + "objective/train/theoretical_loss": 4.217285411461949, + "objective/train/tokens_used": 295711200, + "theoretical_loss": 4.217285411461949, + "tokens_seen": 275251200 + }, + { + "epoch": 0.08, + "learning_rate": 0.0007406860581745236, + "loss": 1.6253, + "theoretical_loss": 4.217285411461949, + "tokens_seen": 275251200 + }, + { + "epoch": 0.08, + "learning_rate": 0.0007406700100300903, + "loss": 1.6686, + "theoretical_loss": 4.217166768196704, + "tokens_seen": 275316736 + }, + { + "epoch": 0.08, + "learning_rate": 0.000740653961885657, + "loss": 1.7709, + "theoretical_loss": 4.217048161075294, + "tokens_seen": 275382272 + }, + { + "epoch": 0.08, + "learning_rate": 0.0007406379137412237, + "loss": 1.6362, + "theoretical_loss": 4.2169295900781085, + "tokens_seen": 275447808 + }, + { + "epoch": 0.08, + "learning_rate": 0.0007406218655967904, + "loss": 1.7757, + "theoretical_loss": 4.216811055185556, + "tokens_seen": 275513344 + }, + { + "epoch": 0.08, + "learning_rate": 0.0007406058174523571, + "loss": 1.7341, + "theoretical_loss": 4.216692556378058, + "tokens_seen": 275578880 + }, + { + "epoch": 0.08, + "learning_rate": 0.0007405897693079238, + "loss": 1.6842, + "theoretical_loss": 4.216574093636053, + "tokens_seen": 275644416 + }, + { + "epoch": 0.08, + "learning_rate": 0.0007405737211634905, + "loss": 1.726, + "theoretical_loss": 4.216455666939991, + "tokens_seen": 275709952 + }, + { + "epoch": 0.08, + "learning_rate": 0.0007405576730190572, + "loss": 1.7236, + "theoretical_loss": 4.216337276270341, + "tokens_seen": 275775488 + }, + { + "epoch": 0.08, + "learning_rate": 0.0007405416248746239, + "loss": 1.5945, + "theoretical_loss": 4.216218921607586, + "tokens_seen": 275841024 + }, + { + "epoch": 0.08, + "learning_rate": 0.0007405255767301906, + "loss": 1.7896, + "theoretical_loss": 4.216100602932224, + "tokens_seen": 275906560 + }, + { + "epoch": 0.08, + "learning_rate": 0.0007405095285857573, + "loss": 1.7486, + "theoretical_loss": 4.21598232022477, + "tokens_seen": 275972096 + }, + { + "epoch": 0.08, + "learning_rate": 0.000740493480441324, + "loss": 1.8183, + "theoretical_loss": 4.215864073465749, + "tokens_seen": 276037632 + }, + { + "epoch": 0.08, + "learning_rate": 0.0007404774322968907, + "loss": 1.7683, + "theoretical_loss": 4.215745862635708, + "tokens_seen": 276103168 + }, + { + "epoch": 0.08, + "learning_rate": 0.0007404613841524574, + "loss": 1.6951, + "theoretical_loss": 4.2156276877152035, + "tokens_seen": 276168704 + }, + { + "epoch": 0.08, + "learning_rate": 0.0007404453360080241, + "loss": 1.8722, + "theoretical_loss": 4.21550954868481, + "tokens_seen": 276234240 + }, + { + "epoch": 0.08, + "learning_rate": 0.0007404292878635908, + "loss": 1.7124, + "theoretical_loss": 4.2153914455251185, + "tokens_seen": 276299776 + }, + { + "epoch": 0.08, + "learning_rate": 0.0007404132397191575, + "loss": 1.7921, + "theoretical_loss": 4.215273378216731, + "tokens_seen": 276365312 + }, + { + "epoch": 0.08, + "learning_rate": 0.0007403971915747242, + "loss": 1.7705, + "theoretical_loss": 4.215155346740268, + "tokens_seen": 276430848 + }, + { + "epoch": 0.08, + "learning_rate": 0.0007403811434302909, + "loss": 1.8177, + "theoretical_loss": 4.215037351076363, + "tokens_seen": 276496384 + }, + { + "epoch": 0.08, + "learning_rate": 0.0007403650952858576, + "loss": 1.5648, + "theoretical_loss": 4.214919391205665, + "tokens_seen": 276561920 + }, + { + "epoch": 0.08, + "learning_rate": 0.0007403490471414243, + "loss": 1.7493, + "theoretical_loss": 4.21480146710884, + "tokens_seen": 276627456 + }, + { + "epoch": 0.08, + "learning_rate": 0.000740332998996991, + "loss": 1.795, + "theoretical_loss": 4.214683578766566, + "tokens_seen": 276692992 + }, + { + "epoch": 0.08, + "learning_rate": 0.0007403169508525577, + "loss": 1.6986, + "theoretical_loss": 4.214565726159539, + "tokens_seen": 276758528 + }, + { + "epoch": 0.08, + "learning_rate": 0.0007403009027081244, + "loss": 1.7076, + "theoretical_loss": 4.214447909268467, + "tokens_seen": 276824064 + }, + { + "epoch": 0.08, + "learning_rate": 0.0007402848545636911, + "loss": 1.7821, + "theoretical_loss": 4.214330128074074, + "tokens_seen": 276889600 + }, + { + "epoch": 0.08, + "learning_rate": 0.0007402688064192578, + "loss": 1.6209, + "theoretical_loss": 4.214212382557101, + "tokens_seen": 276955136 + }, + { + "epoch": 0.08, + "learning_rate": 0.0007402527582748245, + "loss": 1.6704, + "theoretical_loss": 4.214094672698302, + "tokens_seen": 277020672 + }, + { + "epoch": 0.08, + "learning_rate": 0.0007402367101303912, + "loss": 1.6371, + "theoretical_loss": 4.213976998478445, + "tokens_seen": 277086208 + }, + { + "epoch": 0.08, + "learning_rate": 0.0007402206619859579, + "loss": 1.759, + "theoretical_loss": 4.213859359878316, + "tokens_seen": 277151744 + }, + { + "epoch": 0.08, + "learning_rate": 0.0007402046138415247, + "loss": 1.7538, + "theoretical_loss": 4.213741756878713, + "tokens_seen": 277217280 + }, + { + "epoch": 0.08, + "learning_rate": 0.0007401885656970913, + "loss": 1.6653, + "theoretical_loss": 4.213624189460449, + "tokens_seen": 277282816 + }, + { + "epoch": 0.08, + "learning_rate": 0.000740172517552658, + "loss": 1.6814, + "theoretical_loss": 4.213506657604356, + "tokens_seen": 277348352 + }, + { + "epoch": 0.08, + "learning_rate": 0.0007401564694082247, + "loss": 1.6818, + "theoretical_loss": 4.213389161291273, + "tokens_seen": 277413888 + }, + { + "epoch": 0.08, + "learning_rate": 0.0007401404212637915, + "loss": 1.664, + "theoretical_loss": 4.213271700502062, + "tokens_seen": 277479424 + }, + { + "epoch": 0.08, + "learning_rate": 0.0007401243731193581, + "loss": 1.7964, + "theoretical_loss": 4.213154275217597, + "tokens_seen": 277544960 + }, + { + "epoch": 0.08, + "learning_rate": 0.0007401083249749249, + "loss": 1.7626, + "theoretical_loss": 4.213036885418763, + "tokens_seen": 277610496 + }, + { + "epoch": 0.08, + "learning_rate": 0.0007400922768304915, + "loss": 1.6963, + "theoretical_loss": 4.212919531086466, + "tokens_seen": 277676032 + }, + { + "epoch": 0.08, + "learning_rate": 0.0007400762286860583, + "loss": 1.7107, + "theoretical_loss": 4.2128022122016215, + "tokens_seen": 277741568 + }, + { + "epoch": 0.08, + "learning_rate": 0.0007400601805416249, + "loss": 1.7228, + "theoretical_loss": 4.212684928745163, + "tokens_seen": 277807104 + }, + { + "epoch": 0.08, + "learning_rate": 0.0007400441323971917, + "loss": 1.7432, + "theoretical_loss": 4.212567680698038, + "tokens_seen": 277872640 + }, + { + "epoch": 0.08, + "learning_rate": 0.0007400280842527583, + "loss": 1.711, + "theoretical_loss": 4.212450468041208, + "tokens_seen": 277938176 + }, + { + "epoch": 0.08, + "learning_rate": 0.0007400120361083251, + "loss": 1.7981, + "theoretical_loss": 4.21233329075565, + "tokens_seen": 278003712 + }, + { + "epoch": 0.08, + "learning_rate": 0.0007399959879638917, + "loss": 1.6663, + "theoretical_loss": 4.212216148822355, + "tokens_seen": 278069248 + }, + { + "epoch": 0.08, + "learning_rate": 0.0007399799398194585, + "loss": 1.6545, + "theoretical_loss": 4.212099042222331, + "tokens_seen": 278134784 + }, + { + "epoch": 0.08, + "learning_rate": 0.0007399638916750251, + "loss": 1.8317, + "theoretical_loss": 4.211981970936595, + "tokens_seen": 278200320 + }, + { + "epoch": 0.08, + "learning_rate": 0.0007399478435305919, + "loss": 1.6716, + "theoretical_loss": 4.2118649349461865, + "tokens_seen": 278265856 + }, + { + "epoch": 0.08, + "learning_rate": 0.0007399317953861585, + "loss": 1.7654, + "theoretical_loss": 4.211747934232153, + "tokens_seen": 278331392 + }, + { + "epoch": 0.08, + "learning_rate": 0.0007399157472417253, + "loss": 1.7608, + "theoretical_loss": 4.211630968775561, + "tokens_seen": 278396928 + }, + { + "epoch": 0.08, + "learning_rate": 0.0007398996990972919, + "loss": 1.7103, + "theoretical_loss": 4.211514038557488, + "tokens_seen": 278462464 + }, + { + "epoch": 0.08, + "objective/train/docs_used": 107929, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.7213956117630005, + "objective/train/theoretical_loss": 4.211397143559029, + "objective/train/tokens_used": 298988000, + "theoretical_loss": 4.211397143559029, + "tokens_seen": 278528000 + }, + { + "epoch": 0.08, + "learning_rate": 0.0007398836509528586, + "loss": 1.73, + "theoretical_loss": 4.211397143559029, + "tokens_seen": 278528000 + }, + { + "epoch": 0.08, + "learning_rate": 0.0007398676028084253, + "loss": 1.7359, + "theoretical_loss": 4.211280283761294, + "tokens_seen": 278593536 + }, + { + "epoch": 0.08, + "learning_rate": 0.000739851554663992, + "loss": 1.7553, + "theoretical_loss": 4.211163459145403, + "tokens_seen": 278659072 + }, + { + "epoch": 0.08, + "learning_rate": 0.0007398355065195587, + "loss": 1.7627, + "theoretical_loss": 4.211046669692496, + "tokens_seen": 278724608 + }, + { + "epoch": 0.08, + "learning_rate": 0.0007398194583751254, + "loss": 1.7778, + "theoretical_loss": 4.2109299153837245, + "tokens_seen": 278790144 + }, + { + "epoch": 0.08, + "learning_rate": 0.0007398034102306921, + "loss": 1.808, + "theoretical_loss": 4.210813196200255, + "tokens_seen": 278855680 + }, + { + "epoch": 0.08, + "learning_rate": 0.0007397873620862588, + "loss": 1.659, + "theoretical_loss": 4.21069651212327, + "tokens_seen": 278921216 + }, + { + "epoch": 0.08, + "learning_rate": 0.0007397713139418255, + "loss": 1.8267, + "theoretical_loss": 4.2105798631339635, + "tokens_seen": 278986752 + }, + { + "epoch": 0.08, + "learning_rate": 0.0007397552657973922, + "loss": 1.709, + "theoretical_loss": 4.2104632492135465, + "tokens_seen": 279052288 + }, + { + "epoch": 0.08, + "learning_rate": 0.0007397392176529589, + "loss": 1.6104, + "theoretical_loss": 4.210346670343244, + "tokens_seen": 279117824 + }, + { + "epoch": 0.08, + "learning_rate": 0.0007397231695085256, + "loss": 1.7048, + "theoretical_loss": 4.210230126504296, + "tokens_seen": 279183360 + }, + { + "epoch": 0.08, + "learning_rate": 0.0007397071213640923, + "loss": 1.7653, + "theoretical_loss": 4.210113617677955, + "tokens_seen": 279248896 + }, + { + "epoch": 0.08, + "learning_rate": 0.000739691073219659, + "loss": 1.8556, + "theoretical_loss": 4.20999714384549, + "tokens_seen": 279314432 + }, + { + "epoch": 0.08, + "learning_rate": 0.0007396750250752257, + "loss": 1.8738, + "theoretical_loss": 4.209880704988181, + "tokens_seen": 279379968 + }, + { + "epoch": 0.08, + "learning_rate": 0.0007396589769307924, + "loss": 1.7722, + "theoretical_loss": 4.209764301087327, + "tokens_seen": 279445504 + }, + { + "epoch": 0.08, + "learning_rate": 0.0007396429287863591, + "loss": 1.7365, + "theoretical_loss": 4.20964793212424, + "tokens_seen": 279511040 + }, + { + "epoch": 0.08, + "learning_rate": 0.0007396268806419258, + "loss": 1.7316, + "theoretical_loss": 4.209531598080245, + "tokens_seen": 279576576 + }, + { + "epoch": 0.08, + "learning_rate": 0.0007396108324974925, + "loss": 1.8215, + "theoretical_loss": 4.20941529893668, + "tokens_seen": 279642112 + }, + { + "epoch": 0.08, + "learning_rate": 0.0007395947843530592, + "loss": 1.6255, + "theoretical_loss": 4.209299034674902, + "tokens_seen": 279707648 + }, + { + "epoch": 0.08, + "learning_rate": 0.0007395787362086259, + "loss": 1.7352, + "theoretical_loss": 4.209182805276278, + "tokens_seen": 279773184 + }, + { + "epoch": 0.08, + "learning_rate": 0.0007395626880641926, + "loss": 1.7022, + "theoretical_loss": 4.209066610722192, + "tokens_seen": 279838720 + }, + { + "epoch": 0.08, + "learning_rate": 0.0007395466399197593, + "loss": 1.8689, + "theoretical_loss": 4.208950450994041, + "tokens_seen": 279904256 + }, + { + "epoch": 0.08, + "learning_rate": 0.000739530591775326, + "loss": 1.8147, + "theoretical_loss": 4.2088343260732355, + "tokens_seen": 279969792 + }, + { + "epoch": 0.08, + "learning_rate": 0.0007395145436308927, + "loss": 1.6358, + "theoretical_loss": 4.208718235941204, + "tokens_seen": 280035328 + }, + { + "epoch": 0.08, + "learning_rate": 0.0007394984954864594, + "loss": 1.7587, + "theoretical_loss": 4.2086021805793825, + "tokens_seen": 280100864 + }, + { + "epoch": 0.08, + "learning_rate": 0.0007394824473420261, + "loss": 1.726, + "theoretical_loss": 4.208486159969229, + "tokens_seen": 280166400 + }, + { + "epoch": 0.08, + "learning_rate": 0.0007394663991975928, + "loss": 1.6499, + "theoretical_loss": 4.20837017409221, + "tokens_seen": 280231936 + }, + { + "epoch": 0.08, + "learning_rate": 0.0007394503510531595, + "loss": 1.6846, + "theoretical_loss": 4.20825422292981, + "tokens_seen": 280297472 + }, + { + "epoch": 0.08, + "learning_rate": 0.0007394343029087262, + "loss": 1.7777, + "theoretical_loss": 4.208138306463524, + "tokens_seen": 280363008 + }, + { + "epoch": 0.08, + "learning_rate": 0.0007394182547642929, + "loss": 1.7523, + "theoretical_loss": 4.208022424674864, + "tokens_seen": 280428544 + }, + { + "epoch": 0.08, + "learning_rate": 0.0007394022066198596, + "loss": 1.8335, + "theoretical_loss": 4.207906577545356, + "tokens_seen": 280494080 + }, + { + "epoch": 0.09, + "learning_rate": 0.0007393861584754263, + "loss": 1.6543, + "theoretical_loss": 4.207790765056538, + "tokens_seen": 280559616 + }, + { + "epoch": 0.09, + "learning_rate": 0.000739370110330993, + "loss": 1.7447, + "theoretical_loss": 4.207674987189965, + "tokens_seen": 280625152 + }, + { + "epoch": 0.09, + "learning_rate": 0.0007393540621865597, + "loss": 1.6901, + "theoretical_loss": 4.2075592439272045, + "tokens_seen": 280690688 + }, + { + "epoch": 0.09, + "learning_rate": 0.0007393380140421264, + "loss": 1.7692, + "theoretical_loss": 4.207443535249838, + "tokens_seen": 280756224 + }, + { + "epoch": 0.09, + "learning_rate": 0.0007393219658976931, + "loss": 1.6258, + "theoretical_loss": 4.207327861139461, + "tokens_seen": 280821760 + }, + { + "epoch": 0.09, + "learning_rate": 0.0007393059177532598, + "loss": 1.6568, + "theoretical_loss": 4.207212221577684, + "tokens_seen": 280887296 + }, + { + "epoch": 0.09, + "learning_rate": 0.0007392898696088265, + "loss": 1.8187, + "theoretical_loss": 4.207096616546132, + "tokens_seen": 280952832 + }, + { + "epoch": 0.09, + "learning_rate": 0.0007392738214643932, + "loss": 1.6964, + "theoretical_loss": 4.206981046026442, + "tokens_seen": 281018368 + }, + { + "epoch": 0.09, + "learning_rate": 0.0007392577733199599, + "loss": 1.7115, + "theoretical_loss": 4.2068655100002665, + "tokens_seen": 281083904 + }, + { + "epoch": 0.09, + "learning_rate": 0.0007392417251755266, + "loss": 1.6777, + "theoretical_loss": 4.206750008449272, + "tokens_seen": 281149440 + }, + { + "epoch": 0.09, + "learning_rate": 0.0007392256770310933, + "loss": 1.6936, + "theoretical_loss": 4.206634541355138, + "tokens_seen": 281214976 + }, + { + "epoch": 0.09, + "learning_rate": 0.00073920962888666, + "loss": 1.6341, + "theoretical_loss": 4.20651910869956, + "tokens_seen": 281280512 + }, + { + "epoch": 0.09, + "learning_rate": 0.0007391935807422267, + "loss": 1.8313, + "theoretical_loss": 4.2064037104642455, + "tokens_seen": 281346048 + }, + { + "epoch": 0.09, + "learning_rate": 0.0007391775325977934, + "loss": 1.7574, + "theoretical_loss": 4.206288346630917, + "tokens_seen": 281411584 + }, + { + "epoch": 0.09, + "learning_rate": 0.0007391614844533601, + "loss": 1.6461, + "theoretical_loss": 4.206173017181309, + "tokens_seen": 281477120 + }, + { + "epoch": 0.09, + "learning_rate": 0.0007391454363089268, + "loss": 1.6298, + "theoretical_loss": 4.206057722097174, + "tokens_seen": 281542656 + }, + { + "epoch": 0.09, + "learning_rate": 0.0007391293881644935, + "loss": 1.6266, + "theoretical_loss": 4.205942461360276, + "tokens_seen": 281608192 + }, + { + "epoch": 0.09, + "learning_rate": 0.0007391133400200602, + "loss": 1.692, + "theoretical_loss": 4.205827234952391, + "tokens_seen": 281673728 + }, + { + "epoch": 0.09, + "learning_rate": 0.0007390972918756269, + "loss": 1.7718, + "theoretical_loss": 4.205712042855312, + "tokens_seen": 281739264 + }, + { + "epoch": 0.09, + "objective/train/docs_used": 109222, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.6157443523406982, + "objective/train/theoretical_loss": 4.205596885050846, + "objective/train/tokens_used": 302264800, + "theoretical_loss": 4.205596885050846, + "tokens_seen": 281804800 + }, + { + "epoch": 0.09, + "learning_rate": 0.0007390812437311936, + "loss": 1.7052, + "theoretical_loss": 4.205596885050846, + "tokens_seen": 281804800 + }, + { + "epoch": 0.09, + "learning_rate": 0.0007390651955867603, + "loss": 1.7149, + "theoretical_loss": 4.205481761520811, + "tokens_seen": 281870336 + }, + { + "epoch": 0.09, + "learning_rate": 0.000739049147442327, + "loss": 1.6447, + "theoretical_loss": 4.20536667224704, + "tokens_seen": 281935872 + }, + { + "epoch": 0.09, + "learning_rate": 0.0007390330992978937, + "loss": 1.6772, + "theoretical_loss": 4.205251617211382, + "tokens_seen": 282001408 + }, + { + "epoch": 0.09, + "learning_rate": 0.0007390170511534604, + "loss": 1.6222, + "theoretical_loss": 4.205136596395698, + "tokens_seen": 282066944 + }, + { + "epoch": 0.09, + "learning_rate": 0.0007390010030090271, + "loss": 1.6875, + "theoretical_loss": 4.205021609781861, + "tokens_seen": 282132480 + }, + { + "epoch": 0.09, + "learning_rate": 0.0007389849548645938, + "loss": 1.6842, + "theoretical_loss": 4.204906657351763, + "tokens_seen": 282198016 + }, + { + "epoch": 0.09, + "learning_rate": 0.0007389689067201605, + "loss": 1.7007, + "theoretical_loss": 4.2047917390873035, + "tokens_seen": 282263552 + }, + { + "epoch": 0.09, + "learning_rate": 0.0007389528585757272, + "loss": 1.765, + "theoretical_loss": 4.204676854970401, + "tokens_seen": 282329088 + }, + { + "epoch": 0.09, + "learning_rate": 0.000738936810431294, + "loss": 1.697, + "theoretical_loss": 4.204562004982986, + "tokens_seen": 282394624 + }, + { + "epoch": 0.09, + "learning_rate": 0.0007389207622868606, + "loss": 1.6521, + "theoretical_loss": 4.204447189107, + "tokens_seen": 282460160 + }, + { + "epoch": 0.09, + "learning_rate": 0.0007389047141424273, + "loss": 1.7121, + "theoretical_loss": 4.204332407324403, + "tokens_seen": 282525696 + }, + { + "epoch": 0.09, + "learning_rate": 0.000738888665997994, + "loss": 1.7086, + "theoretical_loss": 4.204217659617164, + "tokens_seen": 282591232 + }, + { + "epoch": 0.09, + "learning_rate": 0.0007388726178535607, + "loss": 1.7693, + "theoretical_loss": 4.2041029459672705, + "tokens_seen": 282656768 + }, + { + "epoch": 0.09, + "learning_rate": 0.0007388565697091274, + "loss": 1.6405, + "theoretical_loss": 4.2039882663567205, + "tokens_seen": 282722304 + }, + { + "epoch": 0.09, + "learning_rate": 0.0007388405215646942, + "loss": 1.7454, + "theoretical_loss": 4.203873620767526, + "tokens_seen": 282787840 + }, + { + "epoch": 0.09, + "learning_rate": 0.0007388244734202608, + "loss": 1.7006, + "theoretical_loss": 4.203759009181714, + "tokens_seen": 282853376 + }, + { + "epoch": 0.09, + "learning_rate": 0.0007388084252758276, + "loss": 1.6705, + "theoretical_loss": 4.203644431581323, + "tokens_seen": 282918912 + }, + { + "epoch": 0.09, + "learning_rate": 0.0007387923771313942, + "loss": 1.7079, + "theoretical_loss": 4.203529887948408, + "tokens_seen": 282984448 + }, + { + "epoch": 0.09, + "learning_rate": 0.000738776328986961, + "loss": 1.7163, + "theoretical_loss": 4.203415378265034, + "tokens_seen": 283049984 + }, + { + "epoch": 0.09, + "learning_rate": 0.0007387602808425276, + "loss": 1.6805, + "theoretical_loss": 4.203300902513284, + "tokens_seen": 283115520 + }, + { + "epoch": 0.09, + "learning_rate": 0.0007387442326980944, + "loss": 1.7947, + "theoretical_loss": 4.203186460675251, + "tokens_seen": 283181056 + }, + { + "epoch": 0.09, + "learning_rate": 0.0007387281845536611, + "loss": 1.6748, + "theoretical_loss": 4.203072052733043, + "tokens_seen": 283246592 + }, + { + "epoch": 0.09, + "learning_rate": 0.0007387121364092278, + "loss": 1.685, + "theoretical_loss": 4.202957678668782, + "tokens_seen": 283312128 + }, + { + "epoch": 0.09, + "learning_rate": 0.0007386960882647945, + "loss": 1.6643, + "theoretical_loss": 4.202843338464602, + "tokens_seen": 283377664 + }, + { + "epoch": 0.09, + "learning_rate": 0.0007386800401203612, + "loss": 1.8533, + "theoretical_loss": 4.202729032102653, + "tokens_seen": 283443200 + }, + { + "epoch": 0.09, + "learning_rate": 0.0007386639919759279, + "loss": 1.7096, + "theoretical_loss": 4.202614759565096, + "tokens_seen": 283508736 + }, + { + "epoch": 0.09, + "learning_rate": 0.0007386479438314946, + "loss": 1.7705, + "theoretical_loss": 4.202500520834106, + "tokens_seen": 283574272 + }, + { + "epoch": 0.09, + "learning_rate": 0.0007386318956870613, + "loss": 1.7028, + "theoretical_loss": 4.202386315891874, + "tokens_seen": 283639808 + }, + { + "epoch": 0.09, + "learning_rate": 0.000738615847542628, + "loss": 1.7034, + "theoretical_loss": 4.202272144720602, + "tokens_seen": 283705344 + }, + { + "epoch": 0.09, + "learning_rate": 0.0007385997993981946, + "loss": 1.7116, + "theoretical_loss": 4.202158007302506, + "tokens_seen": 283770880 + }, + { + "epoch": 0.09, + "learning_rate": 0.0007385837512537614, + "loss": 1.7256, + "theoretical_loss": 4.202043903619814, + "tokens_seen": 283836416 + }, + { + "epoch": 0.09, + "learning_rate": 0.000738567703109328, + "loss": 1.6559, + "theoretical_loss": 4.201929833654772, + "tokens_seen": 283901952 + }, + { + "epoch": 0.09, + "learning_rate": 0.0007385516549648948, + "loss": 1.651, + "theoretical_loss": 4.201815797389634, + "tokens_seen": 283967488 + }, + { + "epoch": 0.09, + "learning_rate": 0.0007385356068204614, + "loss": 1.6764, + "theoretical_loss": 4.201701794806672, + "tokens_seen": 284033024 + }, + { + "epoch": 0.09, + "learning_rate": 0.0007385195586760282, + "loss": 1.8475, + "theoretical_loss": 4.201587825888167, + "tokens_seen": 284098560 + }, + { + "epoch": 0.09, + "learning_rate": 0.0007385035105315948, + "loss": 1.5886, + "theoretical_loss": 4.201473890616418, + "tokens_seen": 284164096 + }, + { + "epoch": 0.09, + "learning_rate": 0.0007384874623871615, + "loss": 1.85, + "theoretical_loss": 4.201359988973733, + "tokens_seen": 284229632 + }, + { + "epoch": 0.09, + "learning_rate": 0.0007384714142427282, + "loss": 1.7367, + "theoretical_loss": 4.201246120942438, + "tokens_seen": 284295168 + }, + { + "epoch": 0.09, + "learning_rate": 0.0007384553660982949, + "loss": 1.8059, + "theoretical_loss": 4.201132286504867, + "tokens_seen": 284360704 + }, + { + "epoch": 0.09, + "learning_rate": 0.0007384393179538616, + "loss": 1.789, + "theoretical_loss": 4.201018485643373, + "tokens_seen": 284426240 + }, + { + "epoch": 0.09, + "learning_rate": 0.0007384232698094283, + "loss": 1.6995, + "theoretical_loss": 4.200904718340317, + "tokens_seen": 284491776 + }, + { + "epoch": 0.09, + "learning_rate": 0.000738407221664995, + "loss": 1.7617, + "theoretical_loss": 4.200790984578078, + "tokens_seen": 284557312 + }, + { + "epoch": 0.09, + "learning_rate": 0.0007383911735205617, + "loss": 1.6945, + "theoretical_loss": 4.200677284339045, + "tokens_seen": 284622848 + }, + { + "epoch": 0.09, + "learning_rate": 0.0007383751253761284, + "loss": 1.8084, + "theoretical_loss": 4.2005636176056225, + "tokens_seen": 284688384 + }, + { + "epoch": 0.09, + "learning_rate": 0.0007383590772316951, + "loss": 1.7567, + "theoretical_loss": 4.200449984360226, + "tokens_seen": 284753920 + }, + { + "epoch": 0.09, + "learning_rate": 0.0007383430290872618, + "loss": 1.7587, + "theoretical_loss": 4.200336384585286, + "tokens_seen": 284819456 + }, + { + "epoch": 0.09, + "learning_rate": 0.0007383269809428285, + "loss": 1.7298, + "theoretical_loss": 4.200222818263246, + "tokens_seen": 284884992 + }, + { + "epoch": 0.09, + "learning_rate": 0.0007383109327983952, + "loss": 1.7338, + "theoretical_loss": 4.200109285376563, + "tokens_seen": 284950528 + }, + { + "epoch": 0.09, + "learning_rate": 0.0007382948846539619, + "loss": 1.6662, + "theoretical_loss": 4.1999957859077055, + "tokens_seen": 285016064 + }, + { + "epoch": 0.09, + "objective/train/docs_used": 110428, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.6494081020355225, + "objective/train/theoretical_loss": 4.1998823198391575, + "objective/train/tokens_used": 305541600, + "theoretical_loss": 4.1998823198391575, + "tokens_seen": 285081600 + }, + { + "epoch": 0.09, + "learning_rate": 0.0007382788365095286, + "loss": 1.7091, + "theoretical_loss": 4.1998823198391575, + "tokens_seen": 285081600 + }, + { + "epoch": 0.09, + "learning_rate": 0.0007382627883650953, + "loss": 1.6837, + "theoretical_loss": 4.199768887153415, + "tokens_seen": 285147136 + }, + { + "epoch": 0.09, + "learning_rate": 0.000738246740220662, + "loss": 1.7812, + "theoretical_loss": 4.199655487832987, + "tokens_seen": 285212672 + }, + { + "epoch": 0.09, + "learning_rate": 0.0007382306920762287, + "loss": 1.7756, + "theoretical_loss": 4.199542121860397, + "tokens_seen": 285278208 + }, + { + "epoch": 0.09, + "learning_rate": 0.0007382146439317954, + "loss": 1.7059, + "theoretical_loss": 4.199428789218181, + "tokens_seen": 285343744 + }, + { + "epoch": 0.09, + "learning_rate": 0.0007381985957873621, + "loss": 1.6967, + "theoretical_loss": 4.199315489888885, + "tokens_seen": 285409280 + }, + { + "epoch": 0.09, + "learning_rate": 0.0007381825476429288, + "loss": 1.7051, + "theoretical_loss": 4.199202223855075, + "tokens_seen": 285474816 + }, + { + "epoch": 0.09, + "learning_rate": 0.0007381664994984955, + "loss": 1.7161, + "theoretical_loss": 4.199088991099323, + "tokens_seen": 285540352 + }, + { + "epoch": 0.09, + "learning_rate": 0.0007381504513540622, + "loss": 1.6137, + "theoretical_loss": 4.19897579160422, + "tokens_seen": 285605888 + }, + { + "epoch": 0.09, + "learning_rate": 0.0007381344032096289, + "loss": 1.7261, + "theoretical_loss": 4.1988626253523655, + "tokens_seen": 285671424 + }, + { + "epoch": 0.09, + "learning_rate": 0.0007381183550651956, + "loss": 1.7315, + "theoretical_loss": 4.198749492326374, + "tokens_seen": 285736960 + }, + { + "epoch": 0.09, + "learning_rate": 0.0007381023069207623, + "loss": 1.752, + "theoretical_loss": 4.198636392508874, + "tokens_seen": 285802496 + }, + { + "epoch": 0.09, + "learning_rate": 0.000738086258776329, + "loss": 1.6865, + "theoretical_loss": 4.198523325882505, + "tokens_seen": 285868032 + }, + { + "epoch": 0.09, + "learning_rate": 0.0007380702106318957, + "loss": 1.6543, + "theoretical_loss": 4.198410292429922, + "tokens_seen": 285933568 + }, + { + "epoch": 0.09, + "learning_rate": 0.0007380541624874624, + "loss": 1.6278, + "theoretical_loss": 4.19829729213379, + "tokens_seen": 285999104 + }, + { + "epoch": 0.09, + "learning_rate": 0.0007380381143430291, + "loss": 1.694, + "theoretical_loss": 4.198184324976791, + "tokens_seen": 286064640 + }, + { + "epoch": 0.09, + "learning_rate": 0.0007380220661985958, + "loss": 1.6897, + "theoretical_loss": 4.198071390941616, + "tokens_seen": 286130176 + }, + { + "epoch": 0.09, + "learning_rate": 0.0007380060180541625, + "loss": 1.6426, + "theoretical_loss": 4.197958490010971, + "tokens_seen": 286195712 + }, + { + "epoch": 0.09, + "learning_rate": 0.0007379899699097292, + "loss": 1.7219, + "theoretical_loss": 4.197845622167575, + "tokens_seen": 286261248 + }, + { + "epoch": 0.09, + "learning_rate": 0.0007379739217652959, + "loss": 1.7484, + "theoretical_loss": 4.197732787394161, + "tokens_seen": 286326784 + }, + { + "epoch": 0.09, + "learning_rate": 0.0007379578736208626, + "loss": 1.7159, + "theoretical_loss": 4.197619985673471, + "tokens_seen": 286392320 + }, + { + "epoch": 0.09, + "learning_rate": 0.0007379418254764293, + "loss": 1.6653, + "theoretical_loss": 4.197507216988265, + "tokens_seen": 286457856 + }, + { + "epoch": 0.09, + "learning_rate": 0.000737925777331996, + "loss": 1.6451, + "theoretical_loss": 4.197394481321312, + "tokens_seen": 286523392 + }, + { + "epoch": 0.09, + "learning_rate": 0.0007379097291875627, + "loss": 1.7795, + "theoretical_loss": 4.1972817786553955, + "tokens_seen": 286588928 + }, + { + "epoch": 0.09, + "learning_rate": 0.0007378936810431294, + "loss": 1.6862, + "theoretical_loss": 4.197169108973313, + "tokens_seen": 286654464 + }, + { + "epoch": 0.09, + "learning_rate": 0.0007378776328986961, + "loss": 1.7305, + "theoretical_loss": 4.197056472257874, + "tokens_seen": 286720000 + }, + { + "epoch": 0.09, + "learning_rate": 0.0007378615847542628, + "loss": 1.7809, + "theoretical_loss": 4.196943868491899, + "tokens_seen": 286785536 + }, + { + "epoch": 0.09, + "learning_rate": 0.0007378455366098295, + "loss": 1.7013, + "theoretical_loss": 4.196831297658225, + "tokens_seen": 286851072 + }, + { + "epoch": 0.09, + "learning_rate": 0.0007378294884653962, + "loss": 1.7139, + "theoretical_loss": 4.1967187597396975, + "tokens_seen": 286916608 + }, + { + "epoch": 0.09, + "learning_rate": 0.0007378134403209629, + "loss": 1.6511, + "theoretical_loss": 4.196606254719179, + "tokens_seen": 286982144 + }, + { + "epoch": 0.09, + "learning_rate": 0.0007377973921765296, + "loss": 1.7478, + "theoretical_loss": 4.196493782579544, + "tokens_seen": 287047680 + }, + { + "epoch": 0.09, + "learning_rate": 0.0007377813440320963, + "loss": 1.8014, + "theoretical_loss": 4.196381343303676, + "tokens_seen": 287113216 + }, + { + "epoch": 0.09, + "learning_rate": 0.000737765295887663, + "loss": 1.6889, + "theoretical_loss": 4.1962689368744766, + "tokens_seen": 287178752 + }, + { + "epoch": 0.09, + "learning_rate": 0.0007377492477432297, + "loss": 1.623, + "theoretical_loss": 4.196156563274858, + "tokens_seen": 287244288 + }, + { + "epoch": 0.09, + "learning_rate": 0.0007377331995987964, + "loss": 1.6388, + "theoretical_loss": 4.1960442224877434, + "tokens_seen": 287309824 + }, + { + "epoch": 0.09, + "learning_rate": 0.0007377171514543631, + "loss": 1.6679, + "theoretical_loss": 4.195931914496072, + "tokens_seen": 287375360 + }, + { + "epoch": 0.09, + "learning_rate": 0.0007377011033099298, + "loss": 1.6448, + "theoretical_loss": 4.195819639282792, + "tokens_seen": 287440896 + }, + { + "epoch": 0.09, + "learning_rate": 0.0007376850551654965, + "loss": 1.7951, + "theoretical_loss": 4.195707396830869, + "tokens_seen": 287506432 + }, + { + "epoch": 0.09, + "learning_rate": 0.0007376690070210632, + "loss": 1.6647, + "theoretical_loss": 4.195595187123277, + "tokens_seen": 287571968 + }, + { + "epoch": 0.09, + "learning_rate": 0.0007376529588766299, + "loss": 1.9146, + "theoretical_loss": 4.195483010143006, + "tokens_seen": 287637504 + }, + { + "epoch": 0.09, + "learning_rate": 0.0007376369107321966, + "loss": 1.819, + "theoretical_loss": 4.1953708658730555, + "tokens_seen": 287703040 + }, + { + "epoch": 0.09, + "learning_rate": 0.0007376208625877633, + "loss": 1.7341, + "theoretical_loss": 4.195258754296441, + "tokens_seen": 287768576 + }, + { + "epoch": 0.09, + "learning_rate": 0.00073760481444333, + "loss": 1.7499, + "theoretical_loss": 4.195146675396189, + "tokens_seen": 287834112 + }, + { + "epoch": 0.09, + "learning_rate": 0.0007375887662988967, + "loss": 1.6946, + "theoretical_loss": 4.195034629155338, + "tokens_seen": 287899648 + }, + { + "epoch": 0.09, + "learning_rate": 0.0007375727181544634, + "loss": 1.6304, + "theoretical_loss": 4.194922615556941, + "tokens_seen": 287965184 + }, + { + "epoch": 0.09, + "learning_rate": 0.0007375566700100301, + "loss": 1.7123, + "theoretical_loss": 4.194810634584061, + "tokens_seen": 288030720 + }, + { + "epoch": 0.09, + "learning_rate": 0.0007375406218655968, + "loss": 1.6982, + "theoretical_loss": 4.194698686219778, + "tokens_seen": 288096256 + }, + { + "epoch": 0.09, + "learning_rate": 0.0007375245737211636, + "loss": 1.6847, + "theoretical_loss": 4.194586770447179, + "tokens_seen": 288161792 + }, + { + "epoch": 0.09, + "learning_rate": 0.0007375085255767302, + "loss": 1.6727, + "theoretical_loss": 4.194474887249369, + "tokens_seen": 288227328 + }, + { + "epoch": 0.09, + "learning_rate": 0.000737492477432297, + "loss": 1.6606, + "theoretical_loss": 4.19436303660946, + "tokens_seen": 288292864 + }, + { + "epoch": 0.09, + "objective/train/docs_used": 111591, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.6491329669952393, + "objective/train/theoretical_loss": 4.194251218510583, + "objective/train/tokens_used": 308818400, + "theoretical_loss": 4.194251218510583, + "tokens_seen": 288358400 + }, + { + "epoch": 0.09, + "learning_rate": 0.0007374764292878637, + "loss": 1.6257, + "theoretical_loss": 4.194251218510583, + "tokens_seen": 288358400 + }, + { + "epoch": 0.09, + "learning_rate": 0.0007374603811434304, + "loss": 1.7182, + "theoretical_loss": 4.194139432935877, + "tokens_seen": 288423936 + }, + { + "epoch": 0.09, + "learning_rate": 0.000737444332998997, + "loss": 1.7223, + "theoretical_loss": 4.194027679868495, + "tokens_seen": 288489472 + }, + { + "epoch": 0.09, + "learning_rate": 0.0007374282848545638, + "loss": 1.6705, + "theoretical_loss": 4.193915959291601, + "tokens_seen": 288555008 + }, + { + "epoch": 0.09, + "learning_rate": 0.0007374122367101305, + "loss": 1.7769, + "theoretical_loss": 4.193804271188376, + "tokens_seen": 288620544 + }, + { + "epoch": 0.09, + "learning_rate": 0.0007373961885656972, + "loss": 1.7714, + "theoretical_loss": 4.193692615542007, + "tokens_seen": 288686080 + }, + { + "epoch": 0.09, + "learning_rate": 0.0007373801404212639, + "loss": 1.8298, + "theoretical_loss": 4.193580992335699, + "tokens_seen": 288751616 + }, + { + "epoch": 0.09, + "learning_rate": 0.0007373640922768306, + "loss": 1.7757, + "theoretical_loss": 4.193469401552667, + "tokens_seen": 288817152 + }, + { + "epoch": 0.09, + "learning_rate": 0.0007373480441323973, + "loss": 1.7121, + "theoretical_loss": 4.193357843176139, + "tokens_seen": 288882688 + }, + { + "epoch": 0.09, + "learning_rate": 0.000737331995987964, + "loss": 1.6694, + "theoretical_loss": 4.193246317189356, + "tokens_seen": 288948224 + }, + { + "epoch": 0.09, + "learning_rate": 0.0007373159478435307, + "loss": 1.7629, + "theoretical_loss": 4.19313482357557, + "tokens_seen": 289013760 + }, + { + "epoch": 0.09, + "learning_rate": 0.0007372998996990974, + "loss": 1.8594, + "theoretical_loss": 4.193023362318047, + "tokens_seen": 289079296 + }, + { + "epoch": 0.09, + "learning_rate": 0.0007372838515546641, + "loss": 1.7511, + "theoretical_loss": 4.1929119334000635, + "tokens_seen": 289144832 + }, + { + "epoch": 0.09, + "learning_rate": 0.0007372678034102308, + "loss": 1.6988, + "theoretical_loss": 4.192800536804912, + "tokens_seen": 289210368 + }, + { + "epoch": 0.09, + "learning_rate": 0.0007372517552657975, + "loss": 1.7076, + "theoretical_loss": 4.192689172515893, + "tokens_seen": 289275904 + }, + { + "epoch": 0.09, + "learning_rate": 0.0007372357071213642, + "loss": 1.707, + "theoretical_loss": 4.192577840516323, + "tokens_seen": 289341440 + }, + { + "epoch": 0.09, + "learning_rate": 0.0007372196589769309, + "loss": 1.7032, + "theoretical_loss": 4.192466540789528, + "tokens_seen": 289406976 + }, + { + "epoch": 0.09, + "learning_rate": 0.0007372036108324975, + "loss": 1.7043, + "theoretical_loss": 4.19235527331885, + "tokens_seen": 289472512 + }, + { + "epoch": 0.09, + "learning_rate": 0.0007371875626880643, + "loss": 1.642, + "theoretical_loss": 4.19224403808764, + "tokens_seen": 289538048 + }, + { + "epoch": 0.09, + "learning_rate": 0.0007371715145436309, + "loss": 1.7836, + "theoretical_loss": 4.192132835079262, + "tokens_seen": 289603584 + }, + { + "epoch": 0.09, + "learning_rate": 0.0007371554663991977, + "loss": 1.6842, + "theoretical_loss": 4.192021664277094, + "tokens_seen": 289669120 + }, + { + "epoch": 0.09, + "learning_rate": 0.0007371394182547643, + "loss": 1.8347, + "theoretical_loss": 4.191910525664524, + "tokens_seen": 289734656 + }, + { + "epoch": 0.09, + "learning_rate": 0.0007371233701103311, + "loss": 1.6484, + "theoretical_loss": 4.191799419224955, + "tokens_seen": 289800192 + }, + { + "epoch": 0.09, + "learning_rate": 0.0007371073219658977, + "loss": 1.8959, + "theoretical_loss": 4.1916883449418, + "tokens_seen": 289865728 + }, + { + "epoch": 0.09, + "learning_rate": 0.0007370912738214645, + "loss": 1.8293, + "theoretical_loss": 4.1915773027984855, + "tokens_seen": 289931264 + }, + { + "epoch": 0.09, + "learning_rate": 0.0007370752256770311, + "loss": 1.7501, + "theoretical_loss": 4.19146629277845, + "tokens_seen": 289996800 + }, + { + "epoch": 0.09, + "learning_rate": 0.0007370591775325978, + "loss": 1.7843, + "theoretical_loss": 4.1913553148651435, + "tokens_seen": 290062336 + }, + { + "epoch": 0.09, + "learning_rate": 0.0007370431293881645, + "loss": 1.7068, + "theoretical_loss": 4.191244369042029, + "tokens_seen": 290127872 + }, + { + "epoch": 0.09, + "learning_rate": 0.0007370270812437312, + "loss": 1.7681, + "theoretical_loss": 4.191133455292583, + "tokens_seen": 290193408 + }, + { + "epoch": 0.09, + "learning_rate": 0.0007370110330992979, + "loss": 1.77, + "theoretical_loss": 4.191022573600293, + "tokens_seen": 290258944 + }, + { + "epoch": 0.09, + "learning_rate": 0.0007369949849548646, + "loss": 1.7564, + "theoretical_loss": 4.190911723948657, + "tokens_seen": 290324480 + }, + { + "epoch": 0.09, + "learning_rate": 0.0007369789368104313, + "loss": 1.6675, + "theoretical_loss": 4.190800906321188, + "tokens_seen": 290390016 + }, + { + "epoch": 0.09, + "learning_rate": 0.000736962888665998, + "loss": 1.7481, + "theoretical_loss": 4.190690120701412, + "tokens_seen": 290455552 + }, + { + "epoch": 0.09, + "learning_rate": 0.0007369468405215647, + "loss": 1.6247, + "theoretical_loss": 4.190579367072861, + "tokens_seen": 290521088 + }, + { + "epoch": 0.09, + "learning_rate": 0.0007369307923771314, + "loss": 1.6832, + "theoretical_loss": 4.190468645419086, + "tokens_seen": 290586624 + }, + { + "epoch": 0.09, + "learning_rate": 0.0007369147442326981, + "loss": 1.7791, + "theoretical_loss": 4.190357955723649, + "tokens_seen": 290652160 + }, + { + "epoch": 0.09, + "learning_rate": 0.0007368986960882648, + "loss": 1.609, + "theoretical_loss": 4.190247297970121, + "tokens_seen": 290717696 + }, + { + "epoch": 0.09, + "learning_rate": 0.0007368826479438315, + "loss": 1.7148, + "theoretical_loss": 4.190136672142088, + "tokens_seen": 290783232 + }, + { + "epoch": 0.09, + "learning_rate": 0.0007368665997993982, + "loss": 1.775, + "theoretical_loss": 4.190026078223145, + "tokens_seen": 290848768 + }, + { + "epoch": 0.09, + "learning_rate": 0.0007368505516549649, + "loss": 1.7282, + "theoretical_loss": 4.189915516196904, + "tokens_seen": 290914304 + }, + { + "epoch": 0.09, + "learning_rate": 0.0007368345035105316, + "loss": 1.6803, + "theoretical_loss": 4.189804986046984, + "tokens_seen": 290979840 + }, + { + "epoch": 0.09, + "learning_rate": 0.0007368184553660983, + "loss": 1.7983, + "theoretical_loss": 4.18969448775702, + "tokens_seen": 291045376 + }, + { + "epoch": 0.09, + "learning_rate": 0.000736802407221665, + "loss": 1.8516, + "theoretical_loss": 4.189584021310657, + "tokens_seen": 291110912 + }, + { + "epoch": 0.09, + "learning_rate": 0.0007367863590772317, + "loss": 1.6149, + "theoretical_loss": 4.189473586691554, + "tokens_seen": 291176448 + }, + { + "epoch": 0.09, + "learning_rate": 0.0007367703109327984, + "loss": 1.6599, + "theoretical_loss": 4.189363183883377, + "tokens_seen": 291241984 + }, + { + "epoch": 0.09, + "learning_rate": 0.0007367542627883651, + "loss": 1.6686, + "theoretical_loss": 4.189252812869811, + "tokens_seen": 291307520 + }, + { + "epoch": 0.09, + "learning_rate": 0.0007367382146439318, + "loss": 1.7612, + "theoretical_loss": 4.18914247363455, + "tokens_seen": 291373056 + }, + { + "epoch": 0.09, + "learning_rate": 0.0007367221664994985, + "loss": 1.6751, + "theoretical_loss": 4.189032166161297, + "tokens_seen": 291438592 + }, + { + "epoch": 0.09, + "learning_rate": 0.0007367061183550652, + "loss": 1.6239, + "theoretical_loss": 4.188921890433772, + "tokens_seen": 291504128 + }, + { + "epoch": 0.09, + "learning_rate": 0.0007366900702106319, + "loss": 1.707, + "theoretical_loss": 4.1888116464357035, + "tokens_seen": 291569664 + }, + { + "epoch": 0.09, + "objective/train/docs_used": 112844, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.666099190711975, + "objective/train/theoretical_loss": 4.188701434150834, + "objective/train/tokens_used": 312095200, + "theoretical_loss": 4.188701434150834, + "tokens_seen": 291635200 + }, + { + "epoch": 0.09, + "learning_rate": 0.0007366740220661986, + "loss": 1.7667, + "theoretical_loss": 4.188701434150834, + "tokens_seen": 291635200 + }, + { + "epoch": 0.09, + "learning_rate": 0.0007366579739217653, + "loss": 1.706, + "theoretical_loss": 4.188591253562918, + "tokens_seen": 291700736 + }, + { + "epoch": 0.09, + "learning_rate": 0.000736641925777332, + "loss": 1.6182, + "theoretical_loss": 4.18848110465572, + "tokens_seen": 291766272 + }, + { + "epoch": 0.09, + "learning_rate": 0.0007366258776328987, + "loss": 1.5838, + "theoretical_loss": 4.188370987413018, + "tokens_seen": 291831808 + }, + { + "epoch": 0.09, + "learning_rate": 0.0007366098294884654, + "loss": 1.8023, + "theoretical_loss": 4.188260901818602, + "tokens_seen": 291897344 + }, + { + "epoch": 0.09, + "learning_rate": 0.0007365937813440321, + "loss": 1.7674, + "theoretical_loss": 4.1881508478562735, + "tokens_seen": 291962880 + }, + { + "epoch": 0.09, + "learning_rate": 0.0007365777331995988, + "loss": 1.713, + "theoretical_loss": 4.188040825509845, + "tokens_seen": 292028416 + }, + { + "epoch": 0.09, + "learning_rate": 0.0007365616850551655, + "loss": 1.7364, + "theoretical_loss": 4.1879308347631445, + "tokens_seen": 292093952 + }, + { + "epoch": 0.09, + "learning_rate": 0.0007365456369107322, + "loss": 1.8169, + "theoretical_loss": 4.187820875600006, + "tokens_seen": 292159488 + }, + { + "epoch": 0.09, + "learning_rate": 0.0007365295887662989, + "loss": 1.7631, + "theoretical_loss": 4.187710948004281, + "tokens_seen": 292225024 + }, + { + "epoch": 0.09, + "learning_rate": 0.0007365135406218656, + "loss": 1.7665, + "theoretical_loss": 4.187601051959831, + "tokens_seen": 292290560 + }, + { + "epoch": 0.09, + "learning_rate": 0.0007364974924774323, + "loss": 1.8549, + "theoretical_loss": 4.187491187450527, + "tokens_seen": 292356096 + }, + { + "epoch": 0.09, + "learning_rate": 0.000736481444332999, + "loss": 1.8223, + "theoretical_loss": 4.187381354460255, + "tokens_seen": 292421632 + }, + { + "epoch": 0.09, + "learning_rate": 0.0007364653961885657, + "loss": 1.7657, + "theoretical_loss": 4.187271552972911, + "tokens_seen": 292487168 + }, + { + "epoch": 0.09, + "learning_rate": 0.0007364493480441324, + "loss": 1.7445, + "theoretical_loss": 4.187161782972405, + "tokens_seen": 292552704 + }, + { + "epoch": 0.09, + "learning_rate": 0.0007364332998996991, + "loss": 1.7404, + "theoretical_loss": 4.187052044442655, + "tokens_seen": 292618240 + }, + { + "epoch": 0.09, + "learning_rate": 0.0007364172517552658, + "loss": 1.8939, + "theoretical_loss": 4.1869423373675945, + "tokens_seen": 292683776 + }, + { + "epoch": 0.09, + "learning_rate": 0.0007364012036108325, + "loss": 1.8185, + "theoretical_loss": 4.1868326617311675, + "tokens_seen": 292749312 + }, + { + "epoch": 0.09, + "learning_rate": 0.0007363851554663992, + "loss": 1.7105, + "theoretical_loss": 4.18672301751733, + "tokens_seen": 292814848 + }, + { + "epoch": 0.09, + "learning_rate": 0.0007363691073219659, + "loss": 1.7085, + "theoretical_loss": 4.186613404710048, + "tokens_seen": 292880384 + }, + { + "epoch": 0.09, + "learning_rate": 0.0007363530591775326, + "loss": 1.8454, + "theoretical_loss": 4.186503823293303, + "tokens_seen": 292945920 + }, + { + "epoch": 0.09, + "learning_rate": 0.0007363370110330993, + "loss": 1.6624, + "theoretical_loss": 4.186394273251084, + "tokens_seen": 293011456 + }, + { + "epoch": 0.09, + "learning_rate": 0.0007363209628886661, + "loss": 1.7935, + "theoretical_loss": 4.186284754567395, + "tokens_seen": 293076992 + }, + { + "epoch": 0.09, + "learning_rate": 0.0007363049147442327, + "loss": 1.8297, + "theoretical_loss": 4.1861752672262496, + "tokens_seen": 293142528 + }, + { + "epoch": 0.09, + "learning_rate": 0.0007362888665997995, + "loss": 1.8507, + "theoretical_loss": 4.186065811211675, + "tokens_seen": 293208064 + }, + { + "epoch": 0.09, + "learning_rate": 0.0007362728184553661, + "loss": 1.7137, + "theoretical_loss": 4.185956386507708, + "tokens_seen": 293273600 + }, + { + "epoch": 0.09, + "learning_rate": 0.0007362567703109329, + "loss": 1.7789, + "theoretical_loss": 4.185846993098398, + "tokens_seen": 293339136 + }, + { + "epoch": 0.09, + "learning_rate": 0.0007362407221664995, + "loss": 1.7358, + "theoretical_loss": 4.185737630967808, + "tokens_seen": 293404672 + }, + { + "epoch": 0.09, + "learning_rate": 0.0007362246740220663, + "loss": 1.8608, + "theoretical_loss": 4.18562830010001, + "tokens_seen": 293470208 + }, + { + "epoch": 0.09, + "learning_rate": 0.000736208625877633, + "loss": 1.7296, + "theoretical_loss": 4.185519000479089, + "tokens_seen": 293535744 + }, + { + "epoch": 0.09, + "learning_rate": 0.0007361925777331997, + "loss": 1.7573, + "theoretical_loss": 4.18540973208914, + "tokens_seen": 293601280 + }, + { + "epoch": 0.09, + "learning_rate": 0.0007361765295887663, + "loss": 1.7616, + "theoretical_loss": 4.185300494914273, + "tokens_seen": 293666816 + }, + { + "epoch": 0.09, + "learning_rate": 0.0007361604814443331, + "loss": 1.836, + "theoretical_loss": 4.185191288938606, + "tokens_seen": 293732352 + }, + { + "epoch": 0.09, + "learning_rate": 0.0007361444332998997, + "loss": 1.627, + "theoretical_loss": 4.185082114146271, + "tokens_seen": 293797888 + }, + { + "epoch": 0.09, + "learning_rate": 0.0007361283851554665, + "loss": 1.78, + "theoretical_loss": 4.184972970521409, + "tokens_seen": 293863424 + }, + { + "epoch": 0.09, + "learning_rate": 0.0007361123370110332, + "loss": 1.7657, + "theoretical_loss": 4.184863858048177, + "tokens_seen": 293928960 + }, + { + "epoch": 0.09, + "learning_rate": 0.0007360962888665999, + "loss": 1.7844, + "theoretical_loss": 4.18475477671074, + "tokens_seen": 293994496 + }, + { + "epoch": 0.09, + "learning_rate": 0.0007360802407221666, + "loss": 1.7035, + "theoretical_loss": 4.184645726493275, + "tokens_seen": 294060032 + }, + { + "epoch": 0.09, + "learning_rate": 0.0007360641925777333, + "loss": 1.8892, + "theoretical_loss": 4.184536707379972, + "tokens_seen": 294125568 + }, + { + "epoch": 0.09, + "learning_rate": 0.0007360481444333, + "loss": 1.7247, + "theoretical_loss": 4.184427719355031, + "tokens_seen": 294191104 + }, + { + "epoch": 0.09, + "learning_rate": 0.0007360320962888667, + "loss": 1.8864, + "theoretical_loss": 4.184318762402664, + "tokens_seen": 294256640 + }, + { + "epoch": 0.09, + "learning_rate": 0.0007360160481444334, + "loss": 1.7641, + "theoretical_loss": 4.184209836507096, + "tokens_seen": 294322176 + }, + { + "epoch": 0.09, + "learning_rate": 0.0007360000000000001, + "loss": 1.7185, + "theoretical_loss": 4.18410094165256, + "tokens_seen": 294387712 + }, + { + "epoch": 0.09, + "learning_rate": 0.0007359839518555668, + "loss": 1.8572, + "theoretical_loss": 4.183992077823305, + "tokens_seen": 294453248 + }, + { + "epoch": 0.09, + "learning_rate": 0.0007359679037111335, + "loss": 1.7391, + "theoretical_loss": 4.183883245003588, + "tokens_seen": 294518784 + }, + { + "epoch": 0.09, + "learning_rate": 0.0007359518555667002, + "loss": 1.8038, + "theoretical_loss": 4.183774443177679, + "tokens_seen": 294584320 + }, + { + "epoch": 0.09, + "learning_rate": 0.0007359358074222669, + "loss": 1.947, + "theoretical_loss": 4.183665672329861, + "tokens_seen": 294649856 + }, + { + "epoch": 0.09, + "learning_rate": 0.0007359197592778336, + "loss": 1.7423, + "theoretical_loss": 4.183556932444424, + "tokens_seen": 294715392 + }, + { + "epoch": 0.09, + "learning_rate": 0.0007359037111334003, + "loss": 1.7733, + "theoretical_loss": 4.183448223505673, + "tokens_seen": 294780928 + }, + { + "epoch": 0.09, + "learning_rate": 0.000735887662988967, + "loss": 1.7586, + "theoretical_loss": 4.183339545497924, + "tokens_seen": 294846464 + }, + { + "epoch": 0.09, + "objective/train/docs_used": 114046, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.7973099946975708, + "objective/train/theoretical_loss": 4.1832308984055055, + "objective/train/tokens_used": 315372000, + "theoretical_loss": 4.1832308984055055, + "tokens_seen": 294912000 + }, + { + "epoch": 0.09, + "learning_rate": 0.0007358716148445337, + "loss": 1.7786, + "theoretical_loss": 4.1832308984055055, + "tokens_seen": 294912000 + }, + { + "epoch": 0.09, + "learning_rate": 0.0007358555667001004, + "loss": 1.6798, + "theoretical_loss": 4.183122282212754, + "tokens_seen": 294977536 + }, + { + "epoch": 0.09, + "learning_rate": 0.0007358395185556671, + "loss": 1.9197, + "theoretical_loss": 4.183013696904018, + "tokens_seen": 295043072 + }, + { + "epoch": 0.09, + "learning_rate": 0.0007358234704112338, + "loss": 1.6892, + "theoretical_loss": 4.182905142463662, + "tokens_seen": 295108608 + }, + { + "epoch": 0.09, + "learning_rate": 0.0007358074222668004, + "loss": 1.7392, + "theoretical_loss": 4.182796618876057, + "tokens_seen": 295174144 + }, + { + "epoch": 0.09, + "learning_rate": 0.0007357913741223672, + "loss": 1.825, + "theoretical_loss": 4.182688126125587, + "tokens_seen": 295239680 + }, + { + "epoch": 0.09, + "learning_rate": 0.0007357753259779338, + "loss": 1.8025, + "theoretical_loss": 4.182579664196648, + "tokens_seen": 295305216 + }, + { + "epoch": 0.09, + "learning_rate": 0.0007357592778335006, + "loss": 1.7247, + "theoretical_loss": 4.182471233073647, + "tokens_seen": 295370752 + }, + { + "epoch": 0.09, + "learning_rate": 0.0007357432296890672, + "loss": 1.7748, + "theoretical_loss": 4.182362832741001, + "tokens_seen": 295436288 + }, + { + "epoch": 0.09, + "learning_rate": 0.000735727181544634, + "loss": 1.9126, + "theoretical_loss": 4.182254463183139, + "tokens_seen": 295501824 + }, + { + "epoch": 0.09, + "learning_rate": 0.0007357111334002006, + "loss": 1.6377, + "theoretical_loss": 4.182146124384504, + "tokens_seen": 295567360 + }, + { + "epoch": 0.09, + "learning_rate": 0.0007356950852557674, + "loss": 1.8332, + "theoretical_loss": 4.182037816329547, + "tokens_seen": 295632896 + }, + { + "epoch": 0.09, + "learning_rate": 0.000735679037111334, + "loss": 1.7536, + "theoretical_loss": 4.181929539002731, + "tokens_seen": 295698432 + }, + { + "epoch": 0.09, + "learning_rate": 0.0007356629889669007, + "loss": 1.7708, + "theoretical_loss": 4.1818212923885305, + "tokens_seen": 295763968 + }, + { + "epoch": 0.09, + "learning_rate": 0.0007356469408224674, + "loss": 1.6878, + "theoretical_loss": 4.181713076471432, + "tokens_seen": 295829504 + }, + { + "epoch": 0.09, + "learning_rate": 0.0007356308926780341, + "loss": 1.7055, + "theoretical_loss": 4.181604891235933, + "tokens_seen": 295895040 + }, + { + "epoch": 0.09, + "learning_rate": 0.0007356148445336008, + "loss": 1.6971, + "theoretical_loss": 4.181496736666542, + "tokens_seen": 295960576 + }, + { + "epoch": 0.09, + "learning_rate": 0.0007355987963891675, + "loss": 1.6772, + "theoretical_loss": 4.181388612747778, + "tokens_seen": 296026112 + }, + { + "epoch": 0.09, + "learning_rate": 0.0007355827482447342, + "loss": 1.7997, + "theoretical_loss": 4.181280519464172, + "tokens_seen": 296091648 + }, + { + "epoch": 0.09, + "learning_rate": 0.0007355667001003009, + "loss": 1.71, + "theoretical_loss": 4.181172456800267, + "tokens_seen": 296157184 + }, + { + "epoch": 0.09, + "learning_rate": 0.0007355506519558676, + "loss": 1.7374, + "theoretical_loss": 4.181064424740617, + "tokens_seen": 296222720 + }, + { + "epoch": 0.09, + "learning_rate": 0.0007355346038114343, + "loss": 1.7244, + "theoretical_loss": 4.180956423269784, + "tokens_seen": 296288256 + }, + { + "epoch": 0.09, + "learning_rate": 0.000735518555667001, + "loss": 1.6785, + "theoretical_loss": 4.180848452372347, + "tokens_seen": 296353792 + }, + { + "epoch": 0.09, + "learning_rate": 0.0007355025075225677, + "loss": 1.7608, + "theoretical_loss": 4.180740512032891, + "tokens_seen": 296419328 + }, + { + "epoch": 0.09, + "learning_rate": 0.0007354864593781344, + "loss": 1.7225, + "theoretical_loss": 4.180632602236014, + "tokens_seen": 296484864 + }, + { + "epoch": 0.09, + "learning_rate": 0.0007354704112337011, + "loss": 1.7619, + "theoretical_loss": 4.180524722966327, + "tokens_seen": 296550400 + }, + { + "epoch": 0.09, + "learning_rate": 0.0007354543630892678, + "loss": 1.6837, + "theoretical_loss": 4.180416874208449, + "tokens_seen": 296615936 + }, + { + "epoch": 0.09, + "learning_rate": 0.0007354383149448345, + "loss": 1.8555, + "theoretical_loss": 4.180309055947012, + "tokens_seen": 296681472 + }, + { + "epoch": 0.09, + "learning_rate": 0.0007354222668004012, + "loss": 1.7579, + "theoretical_loss": 4.18020126816666, + "tokens_seen": 296747008 + }, + { + "epoch": 0.09, + "learning_rate": 0.0007354062186559679, + "loss": 1.5818, + "theoretical_loss": 4.180093510852045, + "tokens_seen": 296812544 + }, + { + "epoch": 0.09, + "learning_rate": 0.0007353901705115346, + "loss": 1.7618, + "theoretical_loss": 4.179985783987834, + "tokens_seen": 296878080 + }, + { + "epoch": 0.09, + "learning_rate": 0.0007353741223671013, + "loss": 1.7276, + "theoretical_loss": 4.179878087558702, + "tokens_seen": 296943616 + }, + { + "epoch": 0.09, + "learning_rate": 0.000735358074222668, + "loss": 1.8321, + "theoretical_loss": 4.179770421549336, + "tokens_seen": 297009152 + }, + { + "epoch": 0.09, + "learning_rate": 0.0007353420260782347, + "loss": 1.7049, + "theoretical_loss": 4.179662785944435, + "tokens_seen": 297074688 + }, + { + "epoch": 0.09, + "learning_rate": 0.0007353259779338014, + "loss": 1.7753, + "theoretical_loss": 4.179555180728708, + "tokens_seen": 297140224 + }, + { + "epoch": 0.09, + "learning_rate": 0.0007353099297893681, + "loss": 1.6477, + "theoretical_loss": 4.179447605886876, + "tokens_seen": 297205760 + }, + { + "epoch": 0.09, + "learning_rate": 0.0007352938816449348, + "loss": 1.5375, + "theoretical_loss": 4.1793400614036695, + "tokens_seen": 297271296 + }, + { + "epoch": 0.09, + "learning_rate": 0.0007352778335005015, + "loss": 1.6357, + "theoretical_loss": 4.179232547263831, + "tokens_seen": 297336832 + }, + { + "epoch": 0.09, + "learning_rate": 0.0007352617853560682, + "loss": 1.7096, + "theoretical_loss": 4.1791250634521155, + "tokens_seen": 297402368 + }, + { + "epoch": 0.09, + "learning_rate": 0.0007352457372116349, + "loss": 1.7426, + "theoretical_loss": 4.179017609953287, + "tokens_seen": 297467904 + }, + { + "epoch": 0.09, + "learning_rate": 0.0007352296890672016, + "loss": 1.7282, + "theoretical_loss": 4.17891018675212, + "tokens_seen": 297533440 + }, + { + "epoch": 0.09, + "learning_rate": 0.0007352136409227683, + "loss": 1.7181, + "theoretical_loss": 4.178802793833402, + "tokens_seen": 297598976 + }, + { + "epoch": 0.09, + "learning_rate": 0.000735197592778335, + "loss": 1.7413, + "theoretical_loss": 4.17869543118193, + "tokens_seen": 297664512 + }, + { + "epoch": 0.09, + "learning_rate": 0.0007351815446339017, + "loss": 1.6656, + "theoretical_loss": 4.178588098782513, + "tokens_seen": 297730048 + }, + { + "epoch": 0.09, + "learning_rate": 0.0007351654964894684, + "loss": 1.8001, + "theoretical_loss": 4.178480796619971, + "tokens_seen": 297795584 + }, + { + "epoch": 0.09, + "learning_rate": 0.0007351494483450351, + "loss": 1.6973, + "theoretical_loss": 4.178373524679134, + "tokens_seen": 297861120 + }, + { + "epoch": 0.09, + "learning_rate": 0.0007351334002006018, + "loss": 1.708, + "theoretical_loss": 4.1782662829448425, + "tokens_seen": 297926656 + }, + { + "epoch": 0.09, + "learning_rate": 0.0007351173520561686, + "loss": 1.7886, + "theoretical_loss": 4.17815907140195, + "tokens_seen": 297992192 + }, + { + "epoch": 0.09, + "learning_rate": 0.0007351013039117352, + "loss": 1.7265, + "theoretical_loss": 4.178051890035319, + "tokens_seen": 298057728 + }, + { + "epoch": 0.09, + "learning_rate": 0.000735085255767302, + "loss": 1.6575, + "theoretical_loss": 4.177944738829825, + "tokens_seen": 298123264 + }, + { + "epoch": 0.09, + "objective/train/docs_used": 115247, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.6709059476852417, + "objective/train/theoretical_loss": 4.177837617770352, + "objective/train/tokens_used": 318648800, + "theoretical_loss": 4.177837617770352, + "tokens_seen": 298188800 + }, + { + "epoch": 0.09, + "learning_rate": 0.0007350692076228686, + "loss": 1.7445, + "theoretical_loss": 4.177837617770352, + "tokens_seen": 298188800 + }, + { + "epoch": 0.09, + "learning_rate": 0.0007350531594784354, + "loss": 1.6388, + "theoretical_loss": 4.1777305268417955, + "tokens_seen": 298254336 + }, + { + "epoch": 0.09, + "learning_rate": 0.000735037111334002, + "loss": 1.7788, + "theoretical_loss": 4.177623466029064, + "tokens_seen": 298319872 + }, + { + "epoch": 0.09, + "learning_rate": 0.0007350210631895688, + "loss": 1.8117, + "theoretical_loss": 4.177516435317074, + "tokens_seen": 298385408 + }, + { + "epoch": 0.09, + "learning_rate": 0.0007350050150451354, + "loss": 1.7787, + "theoretical_loss": 4.177409434690754, + "tokens_seen": 298450944 + }, + { + "epoch": 0.09, + "learning_rate": 0.0007349889669007022, + "loss": 1.6712, + "theoretical_loss": 4.177302464135044, + "tokens_seen": 298516480 + }, + { + "epoch": 0.09, + "learning_rate": 0.0007349729187562688, + "loss": 1.7712, + "theoretical_loss": 4.177195523634895, + "tokens_seen": 298582016 + }, + { + "epoch": 0.09, + "learning_rate": 0.0007349568706118356, + "loss": 1.7058, + "theoretical_loss": 4.1770886131752665, + "tokens_seen": 298647552 + }, + { + "epoch": 0.09, + "learning_rate": 0.0007349408224674022, + "loss": 1.7343, + "theoretical_loss": 4.176981732741131, + "tokens_seen": 298713088 + }, + { + "epoch": 0.09, + "learning_rate": 0.000734924774322969, + "loss": 1.7825, + "theoretical_loss": 4.176874882317472, + "tokens_seen": 298778624 + }, + { + "epoch": 0.09, + "learning_rate": 0.0007349087261785356, + "loss": 1.8496, + "theoretical_loss": 4.176768061889283, + "tokens_seen": 298844160 + }, + { + "epoch": 0.09, + "learning_rate": 0.0007348926780341024, + "loss": 1.8905, + "theoretical_loss": 4.176661271441568, + "tokens_seen": 298909696 + }, + { + "epoch": 0.09, + "learning_rate": 0.000734876629889669, + "loss": 1.71, + "theoretical_loss": 4.176554510959342, + "tokens_seen": 298975232 + }, + { + "epoch": 0.09, + "learning_rate": 0.0007348605817452358, + "loss": 1.6873, + "theoretical_loss": 4.1764477804276305, + "tokens_seen": 299040768 + }, + { + "epoch": 0.09, + "learning_rate": 0.0007348445336008024, + "loss": 1.6652, + "theoretical_loss": 4.176341079831471, + "tokens_seen": 299106304 + }, + { + "epoch": 0.09, + "learning_rate": 0.0007348284854563692, + "loss": 1.7692, + "theoretical_loss": 4.176234409155912, + "tokens_seen": 299171840 + }, + { + "epoch": 0.09, + "learning_rate": 0.0007348124373119358, + "loss": 1.6904, + "theoretical_loss": 4.176127768386008, + "tokens_seen": 299237376 + }, + { + "epoch": 0.09, + "learning_rate": 0.0007347963891675026, + "loss": 1.7512, + "theoretical_loss": 4.176021157506832, + "tokens_seen": 299302912 + }, + { + "epoch": 0.09, + "learning_rate": 0.0007347803410230692, + "loss": 1.6858, + "theoretical_loss": 4.1759145765034615, + "tokens_seen": 299368448 + }, + { + "epoch": 0.09, + "learning_rate": 0.000734764292878636, + "loss": 1.6048, + "theoretical_loss": 4.175808025360988, + "tokens_seen": 299433984 + }, + { + "epoch": 0.09, + "learning_rate": 0.0007347482447342027, + "loss": 1.7197, + "theoretical_loss": 4.175701504064511, + "tokens_seen": 299499520 + }, + { + "epoch": 0.09, + "learning_rate": 0.0007347321965897694, + "loss": 1.6316, + "theoretical_loss": 4.175595012599143, + "tokens_seen": 299565056 + }, + { + "epoch": 0.09, + "learning_rate": 0.000734716148445336, + "loss": 1.6773, + "theoretical_loss": 4.175488550950007, + "tokens_seen": 299630592 + }, + { + "epoch": 0.09, + "learning_rate": 0.0007347001003009028, + "loss": 1.7711, + "theoretical_loss": 4.175382119102235, + "tokens_seen": 299696128 + }, + { + "epoch": 0.09, + "learning_rate": 0.0007346840521564695, + "loss": 1.6523, + "theoretical_loss": 4.175275717040972, + "tokens_seen": 299761664 + }, + { + "epoch": 0.09, + "learning_rate": 0.0007346680040120362, + "loss": 1.6845, + "theoretical_loss": 4.175169344751371, + "tokens_seen": 299827200 + }, + { + "epoch": 0.09, + "learning_rate": 0.0007346519558676029, + "loss": 1.7801, + "theoretical_loss": 4.175063002218598, + "tokens_seen": 299892736 + }, + { + "epoch": 0.09, + "learning_rate": 0.0007346359077231696, + "loss": 1.7848, + "theoretical_loss": 4.174956689427829, + "tokens_seen": 299958272 + }, + { + "epoch": 0.09, + "learning_rate": 0.0007346198595787363, + "loss": 1.7924, + "theoretical_loss": 4.17485040636425, + "tokens_seen": 300023808 + }, + { + "epoch": 0.09, + "learning_rate": 0.000734603811434303, + "loss": 1.6965, + "theoretical_loss": 4.174744153013057, + "tokens_seen": 300089344 + }, + { + "epoch": 0.09, + "learning_rate": 0.0007345877632898697, + "loss": 1.7478, + "theoretical_loss": 4.174637929359458, + "tokens_seen": 300154880 + }, + { + "epoch": 0.09, + "learning_rate": 0.0007345717151454364, + "loss": 1.6517, + "theoretical_loss": 4.174531735388673, + "tokens_seen": 300220416 + }, + { + "epoch": 0.09, + "learning_rate": 0.0007345556670010031, + "loss": 1.6726, + "theoretical_loss": 4.174425571085927, + "tokens_seen": 300285952 + }, + { + "epoch": 0.09, + "learning_rate": 0.0007345396188565698, + "loss": 1.7176, + "theoretical_loss": 4.174319436436462, + "tokens_seen": 300351488 + }, + { + "epoch": 0.09, + "learning_rate": 0.0007345235707121365, + "loss": 1.5888, + "theoretical_loss": 4.174213331425527, + "tokens_seen": 300417024 + }, + { + "epoch": 0.09, + "learning_rate": 0.0007345075225677032, + "loss": 1.6816, + "theoretical_loss": 4.174107256038383, + "tokens_seen": 300482560 + }, + { + "epoch": 0.09, + "learning_rate": 0.0007344914744232699, + "loss": 1.7038, + "theoretical_loss": 4.1740012102603, + "tokens_seen": 300548096 + }, + { + "epoch": 0.09, + "learning_rate": 0.0007344754262788366, + "loss": 1.7774, + "theoretical_loss": 4.17389519407656, + "tokens_seen": 300613632 + }, + { + "epoch": 0.09, + "learning_rate": 0.0007344593781344033, + "loss": 1.5742, + "theoretical_loss": 4.173789207472455, + "tokens_seen": 300679168 + }, + { + "epoch": 0.09, + "learning_rate": 0.00073444332998997, + "loss": 1.8255, + "theoretical_loss": 4.173683250433287, + "tokens_seen": 300744704 + }, + { + "epoch": 0.09, + "learning_rate": 0.0007344272818455367, + "loss": 1.7194, + "theoretical_loss": 4.17357732294437, + "tokens_seen": 300810240 + }, + { + "epoch": 0.09, + "learning_rate": 0.0007344112337011033, + "loss": 1.7004, + "theoretical_loss": 4.1734714249910265, + "tokens_seen": 300875776 + }, + { + "epoch": 0.09, + "learning_rate": 0.0007343951855566701, + "loss": 1.873, + "theoretical_loss": 4.173365556558591, + "tokens_seen": 300941312 + }, + { + "epoch": 0.09, + "learning_rate": 0.0007343791374122367, + "loss": 1.7107, + "theoretical_loss": 4.1732597176324076, + "tokens_seen": 301006848 + }, + { + "epoch": 0.09, + "learning_rate": 0.0007343630892678035, + "loss": 1.6005, + "theoretical_loss": 4.173153908197831, + "tokens_seen": 301072384 + }, + { + "epoch": 0.09, + "learning_rate": 0.0007343470411233701, + "loss": 1.8581, + "theoretical_loss": 4.173048128240227, + "tokens_seen": 301137920 + }, + { + "epoch": 0.09, + "learning_rate": 0.0007343309929789369, + "loss": 1.6345, + "theoretical_loss": 4.172942377744972, + "tokens_seen": 301203456 + }, + { + "epoch": 0.09, + "learning_rate": 0.0007343149448345035, + "loss": 1.6218, + "theoretical_loss": 4.17283665669745, + "tokens_seen": 301268992 + }, + { + "epoch": 0.09, + "learning_rate": 0.0007342988966900703, + "loss": 1.611, + "theoretical_loss": 4.172730965083061, + "tokens_seen": 301334528 + }, + { + "epoch": 0.09, + "learning_rate": 0.0007342828485456369, + "loss": 1.6886, + "theoretical_loss": 4.1726253028872105, + "tokens_seen": 301400064 + }, + { + "epoch": 0.09, + "objective/train/docs_used": 116418, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.6034057140350342, + "objective/train/theoretical_loss": 4.172519670095315, + "objective/train/tokens_used": 321925600, + "theoretical_loss": 4.172519670095315, + "tokens_seen": 301465600 + }, + { + "epoch": 0.09, + "learning_rate": 0.0007342668004012036, + "loss": 1.8196, + "theoretical_loss": 4.172519670095315, + "tokens_seen": 301465600 + }, + { + "epoch": 0.09, + "learning_rate": 0.0007342507522567703, + "loss": 1.6165, + "theoretical_loss": 4.172414066692804, + "tokens_seen": 301531136 + }, + { + "epoch": 0.09, + "learning_rate": 0.000734234704112337, + "loss": 1.6753, + "theoretical_loss": 4.172308492665114, + "tokens_seen": 301596672 + }, + { + "epoch": 0.09, + "learning_rate": 0.0007342186559679037, + "loss": 1.7268, + "theoretical_loss": 4.172202947997695, + "tokens_seen": 301662208 + }, + { + "epoch": 0.09, + "learning_rate": 0.0007342026078234704, + "loss": 1.6483, + "theoretical_loss": 4.172097432676006, + "tokens_seen": 301727744 + }, + { + "epoch": 0.09, + "learning_rate": 0.0007341865596790371, + "loss": 1.8326, + "theoretical_loss": 4.171991946685516, + "tokens_seen": 301793280 + }, + { + "epoch": 0.09, + "learning_rate": 0.0007341705115346038, + "loss": 1.6626, + "theoretical_loss": 4.171886490011705, + "tokens_seen": 301858816 + }, + { + "epoch": 0.09, + "learning_rate": 0.0007341544633901705, + "loss": 1.6921, + "theoretical_loss": 4.171781062640063, + "tokens_seen": 301924352 + }, + { + "epoch": 0.09, + "learning_rate": 0.0007341384152457372, + "loss": 1.6975, + "theoretical_loss": 4.17167566455609, + "tokens_seen": 301989888 + }, + { + "epoch": 0.09, + "learning_rate": 0.0007341223671013039, + "loss": 1.7075, + "theoretical_loss": 4.171570295745297, + "tokens_seen": 302055424 + }, + { + "epoch": 0.09, + "learning_rate": 0.0007341063189568706, + "loss": 1.8818, + "theoretical_loss": 4.171464956193205, + "tokens_seen": 302120960 + }, + { + "epoch": 0.09, + "learning_rate": 0.0007340902708124373, + "loss": 1.7207, + "theoretical_loss": 4.171359645885346, + "tokens_seen": 302186496 + }, + { + "epoch": 0.09, + "learning_rate": 0.000734074222668004, + "loss": 1.7166, + "theoretical_loss": 4.17125436480726, + "tokens_seen": 302252032 + }, + { + "epoch": 0.09, + "learning_rate": 0.0007340581745235707, + "loss": 1.7834, + "theoretical_loss": 4.171149112944501, + "tokens_seen": 302317568 + }, + { + "epoch": 0.09, + "learning_rate": 0.0007340421263791374, + "loss": 1.6755, + "theoretical_loss": 4.171043890282629, + "tokens_seen": 302383104 + }, + { + "epoch": 0.09, + "learning_rate": 0.0007340260782347041, + "loss": 1.7353, + "theoretical_loss": 4.170938696807218, + "tokens_seen": 302448640 + }, + { + "epoch": 0.09, + "learning_rate": 0.0007340100300902708, + "loss": 1.6479, + "theoretical_loss": 4.17083353250385, + "tokens_seen": 302514176 + }, + { + "epoch": 0.09, + "learning_rate": 0.0007339939819458375, + "loss": 1.7227, + "theoretical_loss": 4.170728397358118, + "tokens_seen": 302579712 + }, + { + "epoch": 0.09, + "learning_rate": 0.0007339779338014042, + "loss": 1.7237, + "theoretical_loss": 4.170623291355626, + "tokens_seen": 302645248 + }, + { + "epoch": 0.09, + "learning_rate": 0.0007339618856569709, + "loss": 1.7318, + "theoretical_loss": 4.1705182144819855, + "tokens_seen": 302710784 + }, + { + "epoch": 0.09, + "learning_rate": 0.0007339458375125376, + "loss": 1.7141, + "theoretical_loss": 4.170413166722822, + "tokens_seen": 302776320 + }, + { + "epoch": 0.09, + "learning_rate": 0.0007339297893681043, + "loss": 1.6678, + "theoretical_loss": 4.170308148063769, + "tokens_seen": 302841856 + }, + { + "epoch": 0.09, + "learning_rate": 0.0007339137412236711, + "loss": 1.7884, + "theoretical_loss": 4.1702031584904695, + "tokens_seen": 302907392 + }, + { + "epoch": 0.09, + "learning_rate": 0.0007338976930792377, + "loss": 1.6536, + "theoretical_loss": 4.170098197988578, + "tokens_seen": 302972928 + }, + { + "epoch": 0.09, + "learning_rate": 0.0007338816449348045, + "loss": 1.8089, + "theoretical_loss": 4.169993266543759, + "tokens_seen": 303038464 + }, + { + "epoch": 0.09, + "learning_rate": 0.0007338655967903711, + "loss": 1.7048, + "theoretical_loss": 4.169888364141689, + "tokens_seen": 303104000 + }, + { + "epoch": 0.09, + "learning_rate": 0.0007338495486459379, + "loss": 1.6994, + "theoretical_loss": 4.169783490768049, + "tokens_seen": 303169536 + }, + { + "epoch": 0.09, + "learning_rate": 0.0007338335005015045, + "loss": 1.7265, + "theoretical_loss": 4.169678646408536, + "tokens_seen": 303235072 + }, + { + "epoch": 0.09, + "learning_rate": 0.0007338174523570713, + "loss": 1.6332, + "theoretical_loss": 4.169573831048854, + "tokens_seen": 303300608 + }, + { + "epoch": 0.09, + "learning_rate": 0.0007338014042126379, + "loss": 1.8001, + "theoretical_loss": 4.16946904467472, + "tokens_seen": 303366144 + }, + { + "epoch": 0.09, + "learning_rate": 0.0007337853560682047, + "loss": 1.6942, + "theoretical_loss": 4.169364287271858, + "tokens_seen": 303431680 + }, + { + "epoch": 0.09, + "learning_rate": 0.0007337693079237713, + "loss": 1.691, + "theoretical_loss": 4.169259558826003, + "tokens_seen": 303497216 + }, + { + "epoch": 0.09, + "learning_rate": 0.0007337532597793381, + "loss": 1.8677, + "theoretical_loss": 4.169154859322901, + "tokens_seen": 303562752 + }, + { + "epoch": 0.09, + "learning_rate": 0.0007337372116349047, + "loss": 1.8222, + "theoretical_loss": 4.169050188748307, + "tokens_seen": 303628288 + }, + { + "epoch": 0.09, + "learning_rate": 0.0007337211634904715, + "loss": 1.7629, + "theoretical_loss": 4.168945547087987, + "tokens_seen": 303693824 + }, + { + "epoch": 0.09, + "learning_rate": 0.0007337051153460381, + "loss": 1.7293, + "theoretical_loss": 4.168840934327717, + "tokens_seen": 303759360 + }, + { + "epoch": 0.09, + "learning_rate": 0.0007336890672016049, + "loss": 1.6516, + "theoretical_loss": 4.1687363504532815, + "tokens_seen": 303824896 + }, + { + "epoch": 0.09, + "learning_rate": 0.0007336730190571715, + "loss": 1.6091, + "theoretical_loss": 4.168631795450478, + "tokens_seen": 303890432 + }, + { + "epoch": 0.09, + "learning_rate": 0.0007336569709127383, + "loss": 1.6938, + "theoretical_loss": 4.168527269305112, + "tokens_seen": 303955968 + }, + { + "epoch": 0.09, + "learning_rate": 0.0007336409227683049, + "loss": 1.7676, + "theoretical_loss": 4.168422772002998, + "tokens_seen": 304021504 + }, + { + "epoch": 0.09, + "learning_rate": 0.0007336248746238717, + "loss": 1.7218, + "theoretical_loss": 4.168318303529963, + "tokens_seen": 304087040 + }, + { + "epoch": 0.09, + "learning_rate": 0.0007336088264794383, + "loss": 1.7367, + "theoretical_loss": 4.168213863871843, + "tokens_seen": 304152576 + }, + { + "epoch": 0.09, + "learning_rate": 0.0007335927783350051, + "loss": 1.7222, + "theoretical_loss": 4.168109453014484, + "tokens_seen": 304218112 + }, + { + "epoch": 0.09, + "learning_rate": 0.0007335767301905717, + "loss": 1.6648, + "theoretical_loss": 4.1680050709437415, + "tokens_seen": 304283648 + }, + { + "epoch": 0.09, + "learning_rate": 0.0007335606820461385, + "loss": 1.649, + "theoretical_loss": 4.167900717645482, + "tokens_seen": 304349184 + }, + { + "epoch": 0.09, + "learning_rate": 0.0007335446339017051, + "loss": 1.6556, + "theoretical_loss": 4.16779639310558, + "tokens_seen": 304414720 + }, + { + "epoch": 0.09, + "learning_rate": 0.0007335285857572719, + "loss": 1.7109, + "theoretical_loss": 4.167692097309923, + "tokens_seen": 304480256 + }, + { + "epoch": 0.09, + "learning_rate": 0.0007335125376128385, + "loss": 1.5411, + "theoretical_loss": 4.167587830244408, + "tokens_seen": 304545792 + }, + { + "epoch": 0.09, + "learning_rate": 0.0007334964894684053, + "loss": 1.6044, + "theoretical_loss": 4.167483591894937, + "tokens_seen": 304611328 + }, + { + "epoch": 0.09, + "learning_rate": 0.000733480441323972, + "loss": 1.8122, + "theoretical_loss": 4.167379382247429, + "tokens_seen": 304676864 + }, + { + "epoch": 0.09, + "objective/train/docs_used": 117554, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.6125366687774658, + "objective/train/theoretical_loss": 4.167275201287808, + "objective/train/tokens_used": 325202400, + "theoretical_loss": 4.167275201287808, + "tokens_seen": 304742400 + }, + { + "epoch": 0.09, + "learning_rate": 0.0007334643931795387, + "loss": 1.6178, + "theoretical_loss": 4.167275201287808, + "tokens_seen": 304742400 + }, + { + "epoch": 0.09, + "learning_rate": 0.0007334483450351053, + "loss": 1.7959, + "theoretical_loss": 4.16717104900201, + "tokens_seen": 304807936 + }, + { + "epoch": 0.09, + "learning_rate": 0.0007334322968906721, + "loss": 1.8951, + "theoretical_loss": 4.167066925375981, + "tokens_seen": 304873472 + }, + { + "epoch": 0.09, + "learning_rate": 0.0007334162487462387, + "loss": 1.6777, + "theoretical_loss": 4.1669628303956765, + "tokens_seen": 304939008 + }, + { + "epoch": 0.09, + "learning_rate": 0.0007334002006018055, + "loss": 1.7136, + "theoretical_loss": 4.166858764047062, + "tokens_seen": 305004544 + }, + { + "epoch": 0.09, + "learning_rate": 0.0007333841524573721, + "loss": 1.7174, + "theoretical_loss": 4.166754726316111, + "tokens_seen": 305070080 + }, + { + "epoch": 0.09, + "learning_rate": 0.0007333681043129389, + "loss": 1.6579, + "theoretical_loss": 4.16665071718881, + "tokens_seen": 305135616 + }, + { + "epoch": 0.09, + "learning_rate": 0.0007333520561685056, + "loss": 1.8123, + "theoretical_loss": 4.1665467366511555, + "tokens_seen": 305201152 + }, + { + "epoch": 0.09, + "learning_rate": 0.0007333360080240723, + "loss": 1.7296, + "theoretical_loss": 4.166442784689149, + "tokens_seen": 305266688 + }, + { + "epoch": 0.09, + "learning_rate": 0.000733319959879639, + "loss": 1.7432, + "theoretical_loss": 4.166338861288808, + "tokens_seen": 305332224 + }, + { + "epoch": 0.09, + "learning_rate": 0.0007333039117352057, + "loss": 1.7715, + "theoretical_loss": 4.166234966436155, + "tokens_seen": 305397760 + }, + { + "epoch": 0.09, + "learning_rate": 0.0007332878635907724, + "loss": 1.7202, + "theoretical_loss": 4.166131100117226, + "tokens_seen": 305463296 + }, + { + "epoch": 0.09, + "learning_rate": 0.0007332718154463391, + "loss": 1.7248, + "theoretical_loss": 4.166027262318064, + "tokens_seen": 305528832 + }, + { + "epoch": 0.09, + "learning_rate": 0.0007332557673019058, + "loss": 1.7888, + "theoretical_loss": 4.165923453024724, + "tokens_seen": 305594368 + }, + { + "epoch": 0.09, + "learning_rate": 0.0007332397191574725, + "loss": 1.7678, + "theoretical_loss": 4.165819672223269, + "tokens_seen": 305659904 + }, + { + "epoch": 0.09, + "learning_rate": 0.0007332236710130392, + "loss": 1.7789, + "theoretical_loss": 4.165715919899773, + "tokens_seen": 305725440 + }, + { + "epoch": 0.09, + "learning_rate": 0.0007332076228686059, + "loss": 1.6338, + "theoretical_loss": 4.165612196040319, + "tokens_seen": 305790976 + }, + { + "epoch": 0.09, + "learning_rate": 0.0007331915747241726, + "loss": 1.6745, + "theoretical_loss": 4.1655085006310015, + "tokens_seen": 305856512 + }, + { + "epoch": 0.09, + "learning_rate": 0.0007331755265797393, + "loss": 1.7946, + "theoretical_loss": 4.165404833657923, + "tokens_seen": 305922048 + }, + { + "epoch": 0.09, + "learning_rate": 0.000733159478435306, + "loss": 1.7082, + "theoretical_loss": 4.165301195107195, + "tokens_seen": 305987584 + }, + { + "epoch": 0.09, + "learning_rate": 0.0007331434302908727, + "loss": 1.759, + "theoretical_loss": 4.1651975849649405, + "tokens_seen": 306053120 + }, + { + "epoch": 0.09, + "learning_rate": 0.0007331273821464394, + "loss": 1.654, + "theoretical_loss": 4.165094003217293, + "tokens_seen": 306118656 + }, + { + "epoch": 0.09, + "learning_rate": 0.0007331113340020061, + "loss": 1.6675, + "theoretical_loss": 4.164990449850393, + "tokens_seen": 306184192 + }, + { + "epoch": 0.09, + "learning_rate": 0.0007330952858575728, + "loss": 1.8388, + "theoretical_loss": 4.164886924850393, + "tokens_seen": 306249728 + }, + { + "epoch": 0.09, + "learning_rate": 0.0007330792377131395, + "loss": 1.7322, + "theoretical_loss": 4.164783428203453, + "tokens_seen": 306315264 + }, + { + "epoch": 0.09, + "learning_rate": 0.0007330631895687062, + "loss": 1.6639, + "theoretical_loss": 4.1646799598957465, + "tokens_seen": 306380800 + }, + { + "epoch": 0.09, + "learning_rate": 0.0007330471414242729, + "loss": 1.5782, + "theoretical_loss": 4.164576519913453, + "tokens_seen": 306446336 + }, + { + "epoch": 0.09, + "learning_rate": 0.0007330310932798396, + "loss": 1.6565, + "theoretical_loss": 4.164473108242763, + "tokens_seen": 306511872 + }, + { + "epoch": 0.09, + "learning_rate": 0.0007330150451354062, + "loss": 1.9529, + "theoretical_loss": 4.164369724869878, + "tokens_seen": 306577408 + }, + { + "epoch": 0.09, + "learning_rate": 0.000732998996990973, + "loss": 1.4892, + "theoretical_loss": 4.164266369781005, + "tokens_seen": 306642944 + }, + { + "epoch": 0.09, + "learning_rate": 0.0007329829488465396, + "loss": 1.7276, + "theoretical_loss": 4.164163042962366, + "tokens_seen": 306708480 + }, + { + "epoch": 0.09, + "learning_rate": 0.0007329669007021064, + "loss": 1.6383, + "theoretical_loss": 4.1640597444001894, + "tokens_seen": 306774016 + }, + { + "epoch": 0.09, + "learning_rate": 0.000732950852557673, + "loss": 1.7194, + "theoretical_loss": 4.163956474080715, + "tokens_seen": 306839552 + }, + { + "epoch": 0.09, + "learning_rate": 0.0007329348044132398, + "loss": 1.7484, + "theoretical_loss": 4.16385323199019, + "tokens_seen": 306905088 + }, + { + "epoch": 0.09, + "learning_rate": 0.0007329187562688064, + "loss": 1.6868, + "theoretical_loss": 4.1637500181148726, + "tokens_seen": 306970624 + }, + { + "epoch": 0.09, + "learning_rate": 0.0007329027081243732, + "loss": 1.7941, + "theoretical_loss": 4.163646832441031, + "tokens_seen": 307036160 + }, + { + "epoch": 0.09, + "learning_rate": 0.0007328866599799398, + "loss": 1.7326, + "theoretical_loss": 4.163543674954942, + "tokens_seen": 307101696 + }, + { + "epoch": 0.09, + "learning_rate": 0.0007328706118355065, + "loss": 1.8256, + "theoretical_loss": 4.163440545642894, + "tokens_seen": 307167232 + }, + { + "epoch": 0.09, + "learning_rate": 0.0007328545636910732, + "loss": 1.7561, + "theoretical_loss": 4.163337444491183, + "tokens_seen": 307232768 + }, + { + "epoch": 0.09, + "learning_rate": 0.0007328385155466399, + "loss": 1.7677, + "theoretical_loss": 4.163234371486114, + "tokens_seen": 307298304 + }, + { + "epoch": 0.09, + "learning_rate": 0.0007328224674022066, + "loss": 1.6804, + "theoretical_loss": 4.163131326614003, + "tokens_seen": 307363840 + }, + { + "epoch": 0.09, + "learning_rate": 0.0007328064192577733, + "loss": 1.6907, + "theoretical_loss": 4.163028309861176, + "tokens_seen": 307429376 + }, + { + "epoch": 0.09, + "learning_rate": 0.00073279037111334, + "loss": 1.7906, + "theoretical_loss": 4.162925321213968, + "tokens_seen": 307494912 + }, + { + "epoch": 0.09, + "learning_rate": 0.0007327743229689067, + "loss": 1.678, + "theoretical_loss": 4.162822360658723, + "tokens_seen": 307560448 + }, + { + "epoch": 0.09, + "learning_rate": 0.0007327582748244734, + "loss": 1.7791, + "theoretical_loss": 4.162719428181794, + "tokens_seen": 307625984 + }, + { + "epoch": 0.09, + "learning_rate": 0.0007327422266800401, + "loss": 1.5775, + "theoretical_loss": 4.162616523769547, + "tokens_seen": 307691520 + }, + { + "epoch": 0.09, + "learning_rate": 0.0007327261785356068, + "loss": 1.7095, + "theoretical_loss": 4.1625136474083515, + "tokens_seen": 307757056 + }, + { + "epoch": 0.09, + "learning_rate": 0.0007327101303911736, + "loss": 1.6689, + "theoretical_loss": 4.1624107990845935, + "tokens_seen": 307822592 + }, + { + "epoch": 0.09, + "learning_rate": 0.0007326940822467402, + "loss": 1.6153, + "theoretical_loss": 4.162307978784662, + "tokens_seen": 307888128 + }, + { + "epoch": 0.09, + "learning_rate": 0.000732678034102307, + "loss": 1.7628, + "theoretical_loss": 4.162205186494962, + "tokens_seen": 307953664 + }, + { + "epoch": 0.09, + "objective/train/docs_used": 118729, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.7565159797668457, + "objective/train/theoretical_loss": 4.162102422201901, + "objective/train/tokens_used": 328479200, + "theoretical_loss": 4.162102422201901, + "tokens_seen": 308019200 + }, + { + "epoch": 0.09, + "learning_rate": 0.0007326619859578736, + "loss": 1.6893, + "theoretical_loss": 4.162102422201901, + "tokens_seen": 308019200 + }, + { + "epoch": 0.09, + "learning_rate": 0.0007326459378134404, + "loss": 1.7845, + "theoretical_loss": 4.161999685891901, + "tokens_seen": 308084736 + }, + { + "epoch": 0.09, + "learning_rate": 0.000732629889669007, + "loss": 1.7427, + "theoretical_loss": 4.161896977551393, + "tokens_seen": 308150272 + }, + { + "epoch": 0.09, + "learning_rate": 0.0007326138415245738, + "loss": 1.6371, + "theoretical_loss": 4.161794297166815, + "tokens_seen": 308215808 + }, + { + "epoch": 0.09, + "learning_rate": 0.0007325977933801404, + "loss": 1.6365, + "theoretical_loss": 4.161691644724616, + "tokens_seen": 308281344 + }, + { + "epoch": 0.09, + "learning_rate": 0.0007325817452357072, + "loss": 1.8038, + "theoretical_loss": 4.161589020211256, + "tokens_seen": 308346880 + }, + { + "epoch": 0.09, + "learning_rate": 0.0007325656970912738, + "loss": 1.6745, + "theoretical_loss": 4.1614864236132, + "tokens_seen": 308412416 + }, + { + "epoch": 0.09, + "learning_rate": 0.0007325496489468406, + "loss": 1.7819, + "theoretical_loss": 4.161383854916926, + "tokens_seen": 308477952 + }, + { + "epoch": 0.09, + "learning_rate": 0.0007325336008024072, + "loss": 1.7296, + "theoretical_loss": 4.161281314108923, + "tokens_seen": 308543488 + }, + { + "epoch": 0.09, + "learning_rate": 0.000732517552657974, + "loss": 1.7178, + "theoretical_loss": 4.1611788011756845, + "tokens_seen": 308609024 + }, + { + "epoch": 0.09, + "learning_rate": 0.0007325015045135406, + "loss": 1.7708, + "theoretical_loss": 4.161076316103717, + "tokens_seen": 308674560 + }, + { + "epoch": 0.09, + "learning_rate": 0.0007324854563691074, + "loss": 1.7486, + "theoretical_loss": 4.160973858879535, + "tokens_seen": 308740096 + }, + { + "epoch": 0.09, + "learning_rate": 0.000732469408224674, + "loss": 1.6459, + "theoretical_loss": 4.160871429489663, + "tokens_seen": 308805632 + }, + { + "epoch": 0.09, + "learning_rate": 0.0007324533600802408, + "loss": 1.6727, + "theoretical_loss": 4.160769027920635, + "tokens_seen": 308871168 + }, + { + "epoch": 0.09, + "learning_rate": 0.0007324373119358074, + "loss": 1.7087, + "theoretical_loss": 4.160666654158993, + "tokens_seen": 308936704 + }, + { + "epoch": 0.09, + "learning_rate": 0.0007324212637913742, + "loss": 1.7269, + "theoretical_loss": 4.160564308191289, + "tokens_seen": 309002240 + }, + { + "epoch": 0.09, + "learning_rate": 0.0007324052156469408, + "loss": 1.8148, + "theoretical_loss": 4.1604619900040865, + "tokens_seen": 309067776 + }, + { + "epoch": 0.09, + "learning_rate": 0.0007323891675025076, + "loss": 1.7631, + "theoretical_loss": 4.160359699583955, + "tokens_seen": 309133312 + }, + { + "epoch": 0.09, + "learning_rate": 0.0007323731193580742, + "loss": 1.7003, + "theoretical_loss": 4.160257436917476, + "tokens_seen": 309198848 + }, + { + "epoch": 0.09, + "learning_rate": 0.000732357071213641, + "loss": 1.7482, + "theoretical_loss": 4.160155201991238, + "tokens_seen": 309264384 + }, + { + "epoch": 0.09, + "learning_rate": 0.0007323410230692076, + "loss": 1.7761, + "theoretical_loss": 4.160052994791841, + "tokens_seen": 309329920 + }, + { + "epoch": 0.09, + "learning_rate": 0.0007323249749247744, + "loss": 1.6688, + "theoretical_loss": 4.159950815305893, + "tokens_seen": 309395456 + }, + { + "epoch": 0.09, + "learning_rate": 0.000732308926780341, + "loss": 1.6558, + "theoretical_loss": 4.1598486635200125, + "tokens_seen": 309460992 + }, + { + "epoch": 0.09, + "learning_rate": 0.0007322928786359078, + "loss": 1.616, + "theoretical_loss": 4.159746539420825, + "tokens_seen": 309526528 + }, + { + "epoch": 0.09, + "learning_rate": 0.0007322768304914744, + "loss": 1.7797, + "theoretical_loss": 4.159644442994967, + "tokens_seen": 309592064 + }, + { + "epoch": 0.09, + "learning_rate": 0.0007322607823470412, + "loss": 1.7118, + "theoretical_loss": 4.159542374229084, + "tokens_seen": 309657600 + }, + { + "epoch": 0.09, + "learning_rate": 0.0007322447342026078, + "loss": 1.6427, + "theoretical_loss": 4.159440333109832, + "tokens_seen": 309723136 + }, + { + "epoch": 0.09, + "learning_rate": 0.0007322286860581746, + "loss": 1.6694, + "theoretical_loss": 4.159338319623874, + "tokens_seen": 309788672 + }, + { + "epoch": 0.09, + "learning_rate": 0.0007322126379137412, + "loss": 1.6145, + "theoretical_loss": 4.1592363337578835, + "tokens_seen": 309854208 + }, + { + "epoch": 0.09, + "learning_rate": 0.000732196589769308, + "loss": 1.5982, + "theoretical_loss": 4.159134375498542, + "tokens_seen": 309919744 + }, + { + "epoch": 0.09, + "learning_rate": 0.0007321805416248746, + "loss": 1.6677, + "theoretical_loss": 4.159032444832543, + "tokens_seen": 309985280 + }, + { + "epoch": 0.09, + "learning_rate": 0.0007321644934804414, + "loss": 1.679, + "theoretical_loss": 4.158930541746587, + "tokens_seen": 310050816 + }, + { + "epoch": 0.09, + "learning_rate": 0.000732148445336008, + "loss": 1.681, + "theoretical_loss": 4.158828666227383, + "tokens_seen": 310116352 + }, + { + "epoch": 0.09, + "learning_rate": 0.0007321323971915748, + "loss": 1.6715, + "theoretical_loss": 4.158726818261652, + "tokens_seen": 310181888 + }, + { + "epoch": 0.09, + "learning_rate": 0.0007321163490471414, + "loss": 1.6481, + "theoretical_loss": 4.158624997836123, + "tokens_seen": 310247424 + }, + { + "epoch": 0.09, + "learning_rate": 0.0007321003009027082, + "loss": 1.6354, + "theoretical_loss": 4.158523204937531, + "tokens_seen": 310312960 + }, + { + "epoch": 0.09, + "learning_rate": 0.0007320842527582748, + "loss": 1.7324, + "theoretical_loss": 4.158421439552626, + "tokens_seen": 310378496 + }, + { + "epoch": 0.09, + "learning_rate": 0.0007320682046138416, + "loss": 1.656, + "theoretical_loss": 4.158319701668162, + "tokens_seen": 310444032 + }, + { + "epoch": 0.09, + "learning_rate": 0.0007320521564694082, + "loss": 1.5668, + "theoretical_loss": 4.158217991270906, + "tokens_seen": 310509568 + }, + { + "epoch": 0.09, + "learning_rate": 0.000732036108324975, + "loss": 1.7771, + "theoretical_loss": 4.158116308347632, + "tokens_seen": 310575104 + }, + { + "epoch": 0.09, + "learning_rate": 0.0007320200601805416, + "loss": 1.8425, + "theoretical_loss": 4.1580146528851225, + "tokens_seen": 310640640 + }, + { + "epoch": 0.09, + "learning_rate": 0.0007320040120361084, + "loss": 1.6624, + "theoretical_loss": 4.157913024870172, + "tokens_seen": 310706176 + }, + { + "epoch": 0.09, + "learning_rate": 0.000731987963891675, + "loss": 1.682, + "theoretical_loss": 4.157811424289582, + "tokens_seen": 310771712 + }, + { + "epoch": 0.09, + "learning_rate": 0.0007319719157472418, + "loss": 1.6278, + "theoretical_loss": 4.1577098511301624, + "tokens_seen": 310837248 + }, + { + "epoch": 0.09, + "learning_rate": 0.0007319558676028085, + "loss": 1.8102, + "theoretical_loss": 4.157608305378735, + "tokens_seen": 310902784 + }, + { + "epoch": 0.09, + "learning_rate": 0.0007319398194583752, + "loss": 1.6762, + "theoretical_loss": 4.157506787022128, + "tokens_seen": 310968320 + }, + { + "epoch": 0.09, + "learning_rate": 0.0007319237713139419, + "loss": 1.8552, + "theoretical_loss": 4.15740529604718, + "tokens_seen": 311033856 + }, + { + "epoch": 0.09, + "learning_rate": 0.0007319077231695086, + "loss": 1.7637, + "theoretical_loss": 4.15730383244074, + "tokens_seen": 311099392 + }, + { + "epoch": 0.09, + "learning_rate": 0.0007318916750250753, + "loss": 1.7877, + "theoretical_loss": 4.157202396189662, + "tokens_seen": 311164928 + }, + { + "epoch": 0.09, + "learning_rate": 0.000731875626880642, + "loss": 1.4721, + "theoretical_loss": 4.157100987280814, + "tokens_seen": 311230464 + }, + { + "epoch": 0.09, + "objective/train/docs_used": 120076, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.818334937095642, + "objective/train/theoretical_loss": 4.15699960570107, + "objective/train/tokens_used": 331756000, + "theoretical_loss": 4.15699960570107, + "tokens_seen": 311296000 + }, + { + "epoch": 0.09, + "learning_rate": 0.0007318595787362087, + "loss": 1.686, + "theoretical_loss": 4.15699960570107, + "tokens_seen": 311296000 + }, + { + "epoch": 0.09, + "learning_rate": 0.0007318435305917754, + "loss": 1.7126, + "theoretical_loss": 4.156898251437312, + "tokens_seen": 311361536 + }, + { + "epoch": 0.09, + "learning_rate": 0.0007318274824473421, + "loss": 1.6556, + "theoretical_loss": 4.156796924476435, + "tokens_seen": 311427072 + }, + { + "epoch": 0.09, + "learning_rate": 0.0007318114343029088, + "loss": 1.7141, + "theoretical_loss": 4.156695624805341, + "tokens_seen": 311492608 + }, + { + "epoch": 0.09, + "learning_rate": 0.0007317953861584755, + "loss": 1.7688, + "theoretical_loss": 4.15659435241094, + "tokens_seen": 311558144 + }, + { + "epoch": 0.09, + "learning_rate": 0.0007317793380140422, + "loss": 1.7037, + "theoretical_loss": 4.15649310728015, + "tokens_seen": 311623680 + }, + { + "epoch": 0.09, + "learning_rate": 0.0007317632898696089, + "loss": 1.6646, + "theoretical_loss": 4.156391889399904, + "tokens_seen": 311689216 + }, + { + "epoch": 0.09, + "learning_rate": 0.0007317472417251756, + "loss": 1.5443, + "theoretical_loss": 4.156290698757139, + "tokens_seen": 311754752 + }, + { + "epoch": 0.09, + "learning_rate": 0.0007317311935807423, + "loss": 1.7286, + "theoretical_loss": 4.156189535338799, + "tokens_seen": 311820288 + }, + { + "epoch": 0.09, + "learning_rate": 0.000731715145436309, + "loss": 1.6835, + "theoretical_loss": 4.156088399131844, + "tokens_seen": 311885824 + }, + { + "epoch": 0.09, + "learning_rate": 0.0007316990972918757, + "loss": 1.7652, + "theoretical_loss": 4.155987290123236, + "tokens_seen": 311951360 + }, + { + "epoch": 0.09, + "learning_rate": 0.0007316830491474424, + "loss": 1.6989, + "theoretical_loss": 4.15588620829995, + "tokens_seen": 312016896 + }, + { + "epoch": 0.09, + "learning_rate": 0.0007316670010030091, + "loss": 1.745, + "theoretical_loss": 4.15578515364897, + "tokens_seen": 312082432 + }, + { + "epoch": 0.09, + "learning_rate": 0.0007316509528585758, + "loss": 1.7785, + "theoretical_loss": 4.155684126157286, + "tokens_seen": 312147968 + }, + { + "epoch": 0.09, + "learning_rate": 0.0007316349047141425, + "loss": 1.8043, + "theoretical_loss": 4.1555831258119005, + "tokens_seen": 312213504 + }, + { + "epoch": 0.09, + "learning_rate": 0.0007316188565697092, + "loss": 1.8653, + "theoretical_loss": 4.155482152599823, + "tokens_seen": 312279040 + }, + { + "epoch": 0.09, + "learning_rate": 0.0007316028084252759, + "loss": 1.5598, + "theoretical_loss": 4.155381206508071, + "tokens_seen": 312344576 + }, + { + "epoch": 0.09, + "learning_rate": 0.0007315867602808425, + "loss": 1.7186, + "theoretical_loss": 4.155280287523674, + "tokens_seen": 312410112 + }, + { + "epoch": 0.09, + "learning_rate": 0.0007315707121364093, + "loss": 1.6032, + "theoretical_loss": 4.155179395633667, + "tokens_seen": 312475648 + }, + { + "epoch": 0.09, + "learning_rate": 0.0007315546639919759, + "loss": 1.7231, + "theoretical_loss": 4.1550785308250955, + "tokens_seen": 312541184 + }, + { + "epoch": 0.09, + "learning_rate": 0.0007315386158475427, + "loss": 1.5335, + "theoretical_loss": 4.154977693085016, + "tokens_seen": 312606720 + }, + { + "epoch": 0.09, + "learning_rate": 0.0007315225677031093, + "loss": 1.6669, + "theoretical_loss": 4.154876882400489, + "tokens_seen": 312672256 + }, + { + "epoch": 0.09, + "learning_rate": 0.0007315065195586761, + "loss": 1.6878, + "theoretical_loss": 4.154776098758589, + "tokens_seen": 312737792 + }, + { + "epoch": 0.09, + "learning_rate": 0.0007314904714142427, + "loss": 1.6119, + "theoretical_loss": 4.154675342146397, + "tokens_seen": 312803328 + }, + { + "epoch": 0.09, + "learning_rate": 0.0007314744232698095, + "loss": 1.7758, + "theoretical_loss": 4.154574612551001, + "tokens_seen": 312868864 + }, + { + "epoch": 0.09, + "learning_rate": 0.0007314583751253761, + "loss": 1.6961, + "theoretical_loss": 4.154473909959503, + "tokens_seen": 312934400 + }, + { + "epoch": 0.09, + "learning_rate": 0.0007314423269809429, + "loss": 1.796, + "theoretical_loss": 4.154373234359007, + "tokens_seen": 312999936 + }, + { + "epoch": 0.09, + "learning_rate": 0.0007314262788365095, + "loss": 1.835, + "theoretical_loss": 4.154272585736632, + "tokens_seen": 313065472 + }, + { + "epoch": 0.09, + "learning_rate": 0.0007314102306920763, + "loss": 1.7452, + "theoretical_loss": 4.154171964079504, + "tokens_seen": 313131008 + }, + { + "epoch": 0.09, + "learning_rate": 0.0007313941825476429, + "loss": 1.532, + "theoretical_loss": 4.154071369374755, + "tokens_seen": 313196544 + }, + { + "epoch": 0.09, + "learning_rate": 0.0007313781344032097, + "loss": 1.633, + "theoretical_loss": 4.15397080160953, + "tokens_seen": 313262080 + }, + { + "epoch": 0.09, + "learning_rate": 0.0007313620862587763, + "loss": 1.8225, + "theoretical_loss": 4.153870260770979, + "tokens_seen": 313327616 + }, + { + "epoch": 0.09, + "learning_rate": 0.0007313460381143431, + "loss": 1.7727, + "theoretical_loss": 4.153769746846265, + "tokens_seen": 313393152 + }, + { + "epoch": 0.09, + "learning_rate": 0.0007313299899699097, + "loss": 1.6921, + "theoretical_loss": 4.153669259822556, + "tokens_seen": 313458688 + }, + { + "epoch": 0.1, + "learning_rate": 0.0007313139418254765, + "loss": 1.7113, + "theoretical_loss": 4.153568799687031, + "tokens_seen": 313524224 + }, + { + "epoch": 0.1, + "learning_rate": 0.0007312978936810431, + "loss": 1.7319, + "theoretical_loss": 4.153468366426877, + "tokens_seen": 313589760 + }, + { + "epoch": 0.1, + "learning_rate": 0.0007312818455366099, + "loss": 1.684, + "theoretical_loss": 4.153367960029288, + "tokens_seen": 313655296 + }, + { + "epoch": 0.1, + "learning_rate": 0.0007312657973921765, + "loss": 1.691, + "theoretical_loss": 4.153267580481472, + "tokens_seen": 313720832 + }, + { + "epoch": 0.1, + "learning_rate": 0.0007312497492477433, + "loss": 1.9275, + "theoretical_loss": 4.15316722777064, + "tokens_seen": 313786368 + }, + { + "epoch": 0.1, + "learning_rate": 0.0007312337011033099, + "loss": 1.7292, + "theoretical_loss": 4.153066901884016, + "tokens_seen": 313851904 + }, + { + "epoch": 0.1, + "learning_rate": 0.0007312176529588767, + "loss": 1.814, + "theoretical_loss": 4.15296660280883, + "tokens_seen": 313917440 + }, + { + "epoch": 0.1, + "learning_rate": 0.0007312016048144433, + "loss": 1.7392, + "theoretical_loss": 4.152866330532321, + "tokens_seen": 313982976 + }, + { + "epoch": 0.1, + "learning_rate": 0.0007311855566700101, + "loss": 1.7431, + "theoretical_loss": 4.152766085041737, + "tokens_seen": 314048512 + }, + { + "epoch": 0.1, + "learning_rate": 0.0007311695085255767, + "loss": 1.6731, + "theoretical_loss": 4.152665866324337, + "tokens_seen": 314114048 + }, + { + "epoch": 0.1, + "learning_rate": 0.0007311534603811435, + "loss": 1.7467, + "theoretical_loss": 4.152565674367387, + "tokens_seen": 314179584 + }, + { + "epoch": 0.1, + "learning_rate": 0.0007311374122367101, + "loss": 1.723, + "theoretical_loss": 4.15246550915816, + "tokens_seen": 314245120 + }, + { + "epoch": 0.1, + "learning_rate": 0.0007311213640922769, + "loss": 1.6919, + "theoretical_loss": 4.15236537068394, + "tokens_seen": 314310656 + }, + { + "epoch": 0.1, + "learning_rate": 0.0007311053159478435, + "loss": 1.7741, + "theoretical_loss": 4.152265258932019, + "tokens_seen": 314376192 + }, + { + "epoch": 0.1, + "learning_rate": 0.0007310892678034103, + "loss": 1.6972, + "theoretical_loss": 4.152165173889697, + "tokens_seen": 314441728 + }, + { + "epoch": 0.1, + "learning_rate": 0.0007310732196589769, + "loss": 1.607, + "theoretical_loss": 4.152065115544284, + "tokens_seen": 314507264 + }, + { + "epoch": 0.1, + "objective/train/docs_used": 121169, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.7238123416900635, + "objective/train/theoretical_loss": 4.151965083883098, + "objective/train/tokens_used": 335032800, + "theoretical_loss": 4.151965083883098, + "tokens_seen": 314572800 + }, + { + "epoch": 0.1, + "learning_rate": 0.0007310571715145437, + "loss": 1.6484, + "theoretical_loss": 4.151965083883098, + "tokens_seen": 314572800 + }, + { + "epoch": 0.1, + "learning_rate": 0.0007310411233701103, + "loss": 1.6921, + "theoretical_loss": 4.151865078893466, + "tokens_seen": 314638336 + }, + { + "epoch": 0.1, + "learning_rate": 0.0007310250752256771, + "loss": 1.6517, + "theoretical_loss": 4.151765100562723, + "tokens_seen": 314703872 + }, + { + "epoch": 0.1, + "learning_rate": 0.0007310090270812437, + "loss": 1.6811, + "theoretical_loss": 4.151665148878212, + "tokens_seen": 314769408 + }, + { + "epoch": 0.1, + "learning_rate": 0.0007309929789368105, + "loss": 1.7898, + "theoretical_loss": 4.151565223827288, + "tokens_seen": 314834944 + }, + { + "epoch": 0.1, + "learning_rate": 0.0007309769307923771, + "loss": 1.6597, + "theoretical_loss": 4.15146532539731, + "tokens_seen": 314900480 + }, + { + "epoch": 0.1, + "learning_rate": 0.0007309608826479439, + "loss": 1.7177, + "theoretical_loss": 4.151365453575648, + "tokens_seen": 314966016 + }, + { + "epoch": 0.1, + "learning_rate": 0.0007309448345035105, + "loss": 1.6978, + "theoretical_loss": 4.151265608349682, + "tokens_seen": 315031552 + }, + { + "epoch": 0.1, + "learning_rate": 0.0007309287863590773, + "loss": 1.7515, + "theoretical_loss": 4.151165789706798, + "tokens_seen": 315097088 + }, + { + "epoch": 0.1, + "learning_rate": 0.0007309127382146439, + "loss": 1.717, + "theoretical_loss": 4.151065997634393, + "tokens_seen": 315162624 + }, + { + "epoch": 0.1, + "learning_rate": 0.0007308966900702107, + "loss": 1.6669, + "theoretical_loss": 4.1509662321198695, + "tokens_seen": 315228160 + }, + { + "epoch": 0.1, + "learning_rate": 0.0007308806419257773, + "loss": 1.7859, + "theoretical_loss": 4.150866493150642, + "tokens_seen": 315293696 + }, + { + "epoch": 0.1, + "learning_rate": 0.0007308645937813441, + "loss": 1.7265, + "theoretical_loss": 4.150766780714131, + "tokens_seen": 315359232 + }, + { + "epoch": 0.1, + "learning_rate": 0.0007308485456369107, + "loss": 1.6914, + "theoretical_loss": 4.150667094797766, + "tokens_seen": 315424768 + }, + { + "epoch": 0.1, + "learning_rate": 0.0007308324974924775, + "loss": 1.677, + "theoretical_loss": 4.150567435388988, + "tokens_seen": 315490304 + }, + { + "epoch": 0.1, + "learning_rate": 0.0007308164493480441, + "loss": 1.6793, + "theoretical_loss": 4.150467802475241, + "tokens_seen": 315555840 + }, + { + "epoch": 0.1, + "learning_rate": 0.0007308004012036109, + "loss": 1.685, + "theoretical_loss": 4.150368196043983, + "tokens_seen": 315621376 + }, + { + "epoch": 0.1, + "learning_rate": 0.0007307843530591775, + "loss": 1.7068, + "theoretical_loss": 4.150268616082678, + "tokens_seen": 315686912 + }, + { + "epoch": 0.1, + "learning_rate": 0.0007307683049147443, + "loss": 1.5934, + "theoretical_loss": 4.150169062578798, + "tokens_seen": 315752448 + }, + { + "epoch": 0.1, + "learning_rate": 0.0007307522567703109, + "loss": 1.6805, + "theoretical_loss": 4.150069535519825, + "tokens_seen": 315817984 + }, + { + "epoch": 0.1, + "learning_rate": 0.0007307362086258777, + "loss": 1.655, + "theoretical_loss": 4.149970034893249, + "tokens_seen": 315883520 + }, + { + "epoch": 0.1, + "learning_rate": 0.0007307201604814443, + "loss": 1.7066, + "theoretical_loss": 4.149870560686567, + "tokens_seen": 315949056 + }, + { + "epoch": 0.1, + "learning_rate": 0.0007307041123370111, + "loss": 2.0248, + "theoretical_loss": 4.149771112887288, + "tokens_seen": 316014592 + }, + { + "epoch": 0.1, + "learning_rate": 0.0007306880641925777, + "loss": 1.7506, + "theoretical_loss": 4.149671691482926, + "tokens_seen": 316080128 + }, + { + "epoch": 0.1, + "learning_rate": 0.0007306720160481445, + "loss": 1.6686, + "theoretical_loss": 4.149572296461005, + "tokens_seen": 316145664 + }, + { + "epoch": 0.1, + "learning_rate": 0.0007306559679037111, + "loss": 1.6451, + "theoretical_loss": 4.149472927809057, + "tokens_seen": 316211200 + }, + { + "epoch": 0.1, + "learning_rate": 0.0007306399197592779, + "loss": 1.6792, + "theoretical_loss": 4.149373585514625, + "tokens_seen": 316276736 + }, + { + "epoch": 0.1, + "learning_rate": 0.0007306238716148446, + "loss": 1.5694, + "theoretical_loss": 4.149274269565256, + "tokens_seen": 316342272 + }, + { + "epoch": 0.1, + "learning_rate": 0.0007306078234704113, + "loss": 1.6718, + "theoretical_loss": 4.149174979948507, + "tokens_seen": 316407808 + }, + { + "epoch": 0.1, + "learning_rate": 0.000730591775325978, + "loss": 1.5532, + "theoretical_loss": 4.1490757166519465, + "tokens_seen": 316473344 + }, + { + "epoch": 0.1, + "learning_rate": 0.0007305757271815447, + "loss": 1.7812, + "theoretical_loss": 4.148976479663148, + "tokens_seen": 316538880 + }, + { + "epoch": 0.1, + "learning_rate": 0.0007305596790371114, + "loss": 1.6805, + "theoretical_loss": 4.148877268969694, + "tokens_seen": 316604416 + }, + { + "epoch": 0.1, + "learning_rate": 0.0007305436308926781, + "loss": 1.7505, + "theoretical_loss": 4.1487780845591775, + "tokens_seen": 316669952 + }, + { + "epoch": 0.1, + "learning_rate": 0.0007305275827482448, + "loss": 1.6686, + "theoretical_loss": 4.148678926419197, + "tokens_seen": 316735488 + }, + { + "epoch": 0.1, + "learning_rate": 0.0007305115346038115, + "loss": 1.6902, + "theoretical_loss": 4.148579794537362, + "tokens_seen": 316801024 + }, + { + "epoch": 0.1, + "learning_rate": 0.0007304954864593782, + "loss": 1.7654, + "theoretical_loss": 4.148480688901288, + "tokens_seen": 316866560 + }, + { + "epoch": 0.1, + "learning_rate": 0.0007304794383149449, + "loss": 1.6637, + "theoretical_loss": 4.148381609498602, + "tokens_seen": 316932096 + }, + { + "epoch": 0.1, + "learning_rate": 0.0007304633901705116, + "loss": 1.7087, + "theoretical_loss": 4.148282556316936, + "tokens_seen": 316997632 + }, + { + "epoch": 0.1, + "learning_rate": 0.0007304473420260783, + "loss": 1.7401, + "theoretical_loss": 4.148183529343932, + "tokens_seen": 317063168 + }, + { + "epoch": 0.1, + "learning_rate": 0.000730431293881645, + "loss": 1.6517, + "theoretical_loss": 4.148084528567241, + "tokens_seen": 317128704 + }, + { + "epoch": 0.1, + "learning_rate": 0.0007304152457372117, + "loss": 1.6817, + "theoretical_loss": 4.147985553974522, + "tokens_seen": 317194240 + }, + { + "epoch": 0.1, + "learning_rate": 0.0007303991975927784, + "loss": 1.6466, + "theoretical_loss": 4.14788660555344, + "tokens_seen": 317259776 + }, + { + "epoch": 0.1, + "learning_rate": 0.0007303831494483451, + "loss": 1.674, + "theoretical_loss": 4.147787683291672, + "tokens_seen": 317325312 + }, + { + "epoch": 0.1, + "learning_rate": 0.0007303671013039118, + "loss": 1.6404, + "theoretical_loss": 4.147688787176902, + "tokens_seen": 317390848 + }, + { + "epoch": 0.1, + "learning_rate": 0.0007303510531594785, + "loss": 1.8034, + "theoretical_loss": 4.147589917196821, + "tokens_seen": 317456384 + }, + { + "epoch": 0.1, + "learning_rate": 0.0007303350050150452, + "loss": 1.7041, + "theoretical_loss": 4.14749107333913, + "tokens_seen": 317521920 + }, + { + "epoch": 0.1, + "learning_rate": 0.0007303189568706119, + "loss": 1.5829, + "theoretical_loss": 4.147392255591538, + "tokens_seen": 317587456 + }, + { + "epoch": 0.1, + "learning_rate": 0.0007303029087261786, + "loss": 1.6593, + "theoretical_loss": 4.147293463941762, + "tokens_seen": 317652992 + }, + { + "epoch": 0.1, + "learning_rate": 0.0007302868605817453, + "loss": 1.6402, + "theoretical_loss": 4.147194698377526, + "tokens_seen": 317718528 + }, + { + "epoch": 0.1, + "learning_rate": 0.000730270812437312, + "loss": 1.8486, + "theoretical_loss": 4.147095958886566, + "tokens_seen": 317784064 + }, + { + "epoch": 0.1, + "objective/train/docs_used": 122440, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.544681429862976, + "objective/train/theoretical_loss": 4.146997245456623, + "objective/train/tokens_used": 338309600, + "theoretical_loss": 4.146997245456623, + "tokens_seen": 317849600 + }, + { + "epoch": 0.1, + "learning_rate": 0.0007302547642928787, + "loss": 1.6856, + "theoretical_loss": 4.146997245456623, + "tokens_seen": 317849600 + }, + { + "epoch": 0.1, + "learning_rate": 0.0007302387161484454, + "loss": 1.7581, + "theoretical_loss": 4.146898558075446, + "tokens_seen": 317915136 + }, + { + "epoch": 0.1, + "learning_rate": 0.0007302226680040121, + "loss": 1.6719, + "theoretical_loss": 4.146799896730794, + "tokens_seen": 317980672 + }, + { + "epoch": 0.1, + "learning_rate": 0.0007302066198595788, + "loss": 1.6172, + "theoretical_loss": 4.146701261410435, + "tokens_seen": 318046208 + }, + { + "epoch": 0.1, + "learning_rate": 0.0007301905717151454, + "loss": 1.569, + "theoretical_loss": 4.1466026521021435, + "tokens_seen": 318111744 + }, + { + "epoch": 0.1, + "learning_rate": 0.0007301745235707122, + "loss": 1.7414, + "theoretical_loss": 4.146504068793703, + "tokens_seen": 318177280 + }, + { + "epoch": 0.1, + "learning_rate": 0.0007301584754262788, + "loss": 1.5597, + "theoretical_loss": 4.146405511472905, + "tokens_seen": 318242816 + }, + { + "epoch": 0.1, + "learning_rate": 0.0007301424272818456, + "loss": 1.6628, + "theoretical_loss": 4.146306980127548, + "tokens_seen": 318308352 + }, + { + "epoch": 0.1, + "learning_rate": 0.0007301263791374122, + "loss": 1.665, + "theoretical_loss": 4.146208474745442, + "tokens_seen": 318373888 + }, + { + "epoch": 0.1, + "learning_rate": 0.000730110330992979, + "loss": 1.6464, + "theoretical_loss": 4.1461099953144025, + "tokens_seen": 318439424 + }, + { + "epoch": 0.1, + "learning_rate": 0.0007300942828485456, + "loss": 1.7348, + "theoretical_loss": 4.146011541822253, + "tokens_seen": 318504960 + }, + { + "epoch": 0.1, + "learning_rate": 0.0007300782347041124, + "loss": 1.7834, + "theoretical_loss": 4.145913114256829, + "tokens_seen": 318570496 + }, + { + "epoch": 0.1, + "learning_rate": 0.000730062186559679, + "loss": 1.7362, + "theoretical_loss": 4.145814712605969, + "tokens_seen": 318636032 + }, + { + "epoch": 0.1, + "learning_rate": 0.0007300461384152458, + "loss": 1.7104, + "theoretical_loss": 4.1457163368575225, + "tokens_seen": 318701568 + }, + { + "epoch": 0.1, + "learning_rate": 0.0007300300902708124, + "loss": 1.6323, + "theoretical_loss": 4.145617986999348, + "tokens_seen": 318767104 + }, + { + "epoch": 0.1, + "learning_rate": 0.0007300140421263792, + "loss": 1.7372, + "theoretical_loss": 4.145519663019309, + "tokens_seen": 318832640 + }, + { + "epoch": 0.1, + "learning_rate": 0.0007299979939819458, + "loss": 1.7016, + "theoretical_loss": 4.145421364905282, + "tokens_seen": 318898176 + }, + { + "epoch": 0.1, + "learning_rate": 0.0007299819458375126, + "loss": 1.8192, + "theoretical_loss": 4.145323092645146, + "tokens_seen": 318963712 + }, + { + "epoch": 0.1, + "learning_rate": 0.0007299658976930792, + "loss": 1.6481, + "theoretical_loss": 4.145224846226793, + "tokens_seen": 319029248 + }, + { + "epoch": 0.1, + "learning_rate": 0.000729949849548646, + "loss": 1.7163, + "theoretical_loss": 4.14512662563812, + "tokens_seen": 319094784 + }, + { + "epoch": 0.1, + "learning_rate": 0.0007299338014042126, + "loss": 1.6748, + "theoretical_loss": 4.145028430867035, + "tokens_seen": 319160320 + }, + { + "epoch": 0.1, + "learning_rate": 0.0007299177532597794, + "loss": 1.7009, + "theoretical_loss": 4.144930261901451, + "tokens_seen": 319225856 + }, + { + "epoch": 0.1, + "learning_rate": 0.000729901705115346, + "loss": 1.5544, + "theoretical_loss": 4.144832118729291, + "tokens_seen": 319291392 + }, + { + "epoch": 0.1, + "learning_rate": 0.0007298856569709128, + "loss": 1.683, + "theoretical_loss": 4.144734001338486, + "tokens_seen": 319356928 + }, + { + "epoch": 0.1, + "learning_rate": 0.0007298696088264794, + "loss": 1.714, + "theoretical_loss": 4.144635909716976, + "tokens_seen": 319422464 + }, + { + "epoch": 0.1, + "learning_rate": 0.0007298535606820462, + "loss": 1.6802, + "theoretical_loss": 4.144537843852705, + "tokens_seen": 319488000 + }, + { + "epoch": 0.1, + "learning_rate": 0.0007298375125376128, + "loss": 1.8529, + "theoretical_loss": 4.144439803733631, + "tokens_seen": 319553536 + }, + { + "epoch": 0.1, + "learning_rate": 0.0007298214643931796, + "loss": 1.7111, + "theoretical_loss": 4.144341789347715, + "tokens_seen": 319619072 + }, + { + "epoch": 0.1, + "learning_rate": 0.0007298054162487462, + "loss": 1.7965, + "theoretical_loss": 4.144243800682931, + "tokens_seen": 319684608 + }, + { + "epoch": 0.1, + "learning_rate": 0.000729789368104313, + "loss": 1.5613, + "theoretical_loss": 4.1441458377272555, + "tokens_seen": 319750144 + }, + { + "epoch": 0.1, + "learning_rate": 0.0007297733199598796, + "loss": 1.6022, + "theoretical_loss": 4.144047900468678, + "tokens_seen": 319815680 + }, + { + "epoch": 0.1, + "learning_rate": 0.0007297572718154464, + "loss": 1.8574, + "theoretical_loss": 4.143949988895193, + "tokens_seen": 319881216 + }, + { + "epoch": 0.1, + "learning_rate": 0.000729741223671013, + "loss": 1.6895, + "theoretical_loss": 4.1438521029948046, + "tokens_seen": 319946752 + }, + { + "epoch": 0.1, + "learning_rate": 0.0007297251755265798, + "loss": 1.706, + "theoretical_loss": 4.1437542427555245, + "tokens_seen": 320012288 + }, + { + "epoch": 0.1, + "learning_rate": 0.0007297091273821464, + "loss": 1.7095, + "theoretical_loss": 4.143656408165373, + "tokens_seen": 320077824 + }, + { + "epoch": 0.1, + "learning_rate": 0.0007296930792377132, + "loss": 1.8035, + "theoretical_loss": 4.143558599212376, + "tokens_seen": 320143360 + }, + { + "epoch": 0.1, + "learning_rate": 0.0007296770310932798, + "loss": 1.8048, + "theoretical_loss": 4.143460815884571, + "tokens_seen": 320208896 + }, + { + "epoch": 0.1, + "learning_rate": 0.0007296609829488466, + "loss": 1.8103, + "theoretical_loss": 4.143363058170001, + "tokens_seen": 320274432 + }, + { + "epoch": 0.1, + "learning_rate": 0.0007296449348044132, + "loss": 1.6059, + "theoretical_loss": 4.143265326056719, + "tokens_seen": 320339968 + }, + { + "epoch": 0.1, + "learning_rate": 0.00072962888665998, + "loss": 1.6706, + "theoretical_loss": 4.1431676195327825, + "tokens_seen": 320405504 + }, + { + "epoch": 0.1, + "learning_rate": 0.0007296128385155466, + "loss": 1.5998, + "theoretical_loss": 4.143069938586263, + "tokens_seen": 320471040 + }, + { + "epoch": 0.1, + "learning_rate": 0.0007295967903711134, + "loss": 1.6611, + "theoretical_loss": 4.142972283205233, + "tokens_seen": 320536576 + }, + { + "epoch": 0.1, + "learning_rate": 0.00072958074222668, + "loss": 1.75, + "theoretical_loss": 4.142874653377778, + "tokens_seen": 320602112 + }, + { + "epoch": 0.1, + "learning_rate": 0.0007295646940822468, + "loss": 1.7821, + "theoretical_loss": 4.14277704909199, + "tokens_seen": 320667648 + }, + { + "epoch": 0.1, + "learning_rate": 0.0007295486459378134, + "loss": 1.7046, + "theoretical_loss": 4.142679470335969, + "tokens_seen": 320733184 + }, + { + "epoch": 0.1, + "learning_rate": 0.0007295325977933802, + "loss": 1.6689, + "theoretical_loss": 4.142581917097821, + "tokens_seen": 320798720 + }, + { + "epoch": 0.1, + "learning_rate": 0.0007295165496489468, + "loss": 1.6236, + "theoretical_loss": 4.142484389365664, + "tokens_seen": 320864256 + }, + { + "epoch": 0.1, + "learning_rate": 0.0007295005015045136, + "loss": 1.6683, + "theoretical_loss": 4.14238688712762, + "tokens_seen": 320929792 + }, + { + "epoch": 0.1, + "learning_rate": 0.0007294844533600802, + "loss": 1.5962, + "theoretical_loss": 4.142289410371823, + "tokens_seen": 320995328 + }, + { + "epoch": 0.1, + "learning_rate": 0.000729468405215647, + "loss": 1.6061, + "theoretical_loss": 4.14219195908641, + "tokens_seen": 321060864 + }, + { + "epoch": 0.1, + "objective/train/docs_used": 123582, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.6867141723632812, + "objective/train/theoretical_loss": 4.14209453325953, + "objective/train/tokens_used": 341586400, + "theoretical_loss": 4.14209453325953, + "tokens_seen": 321126400 + }, + { + "epoch": 0.1, + "learning_rate": 0.0007294523570712136, + "loss": 1.7088, + "theoretical_loss": 4.14209453325953, + "tokens_seen": 321126400 + }, + { + "epoch": 0.1, + "learning_rate": 0.0007294363089267804, + "loss": 1.6917, + "theoretical_loss": 4.1419971328793395, + "tokens_seen": 321191936 + }, + { + "epoch": 0.1, + "learning_rate": 0.000729420260782347, + "loss": 1.4863, + "theoretical_loss": 4.141899757934, + "tokens_seen": 321257472 + }, + { + "epoch": 0.1, + "learning_rate": 0.0007294042126379138, + "loss": 1.5755, + "theoretical_loss": 4.141802408411683, + "tokens_seen": 321323008 + }, + { + "epoch": 0.1, + "learning_rate": 0.0007293881644934804, + "loss": 1.6586, + "theoretical_loss": 4.14170508430057, + "tokens_seen": 321388544 + }, + { + "epoch": 0.1, + "learning_rate": 0.0007293721163490472, + "loss": 1.7513, + "theoretical_loss": 4.141607785588846, + "tokens_seen": 321454080 + }, + { + "epoch": 0.1, + "learning_rate": 0.0007293560682046138, + "loss": 1.6832, + "theoretical_loss": 4.141510512264707, + "tokens_seen": 321519616 + }, + { + "epoch": 0.1, + "learning_rate": 0.0007293400200601806, + "loss": 1.5708, + "theoretical_loss": 4.141413264316356, + "tokens_seen": 321585152 + }, + { + "epoch": 0.1, + "learning_rate": 0.0007293239719157472, + "loss": 1.7183, + "theoretical_loss": 4.141316041732004, + "tokens_seen": 321650688 + }, + { + "epoch": 0.1, + "learning_rate": 0.000729307923771314, + "loss": 1.7372, + "theoretical_loss": 4.141218844499869, + "tokens_seen": 321716224 + }, + { + "epoch": 0.1, + "learning_rate": 0.0007292918756268806, + "loss": 1.7229, + "theoretical_loss": 4.141121672608178, + "tokens_seen": 321781760 + }, + { + "epoch": 0.1, + "learning_rate": 0.0007292758274824474, + "loss": 1.6512, + "theoretical_loss": 4.141024526045166, + "tokens_seen": 321847296 + }, + { + "epoch": 0.1, + "learning_rate": 0.000729259779338014, + "loss": 1.5841, + "theoretical_loss": 4.140927404799075, + "tokens_seen": 321912832 + }, + { + "epoch": 0.1, + "learning_rate": 0.0007292437311935808, + "loss": 1.6798, + "theoretical_loss": 4.140830308858154, + "tokens_seen": 321978368 + }, + { + "epoch": 0.1, + "learning_rate": 0.0007292276830491475, + "loss": 1.8154, + "theoretical_loss": 4.140733238210664, + "tokens_seen": 322043904 + }, + { + "epoch": 0.1, + "learning_rate": 0.0007292116349047142, + "loss": 1.6837, + "theoretical_loss": 4.140636192844867, + "tokens_seen": 322109440 + }, + { + "epoch": 0.1, + "learning_rate": 0.0007291955867602809, + "loss": 1.6191, + "theoretical_loss": 4.1405391727490395, + "tokens_seen": 322174976 + }, + { + "epoch": 0.1, + "learning_rate": 0.0007291795386158476, + "loss": 1.6922, + "theoretical_loss": 4.140442177911462, + "tokens_seen": 322240512 + }, + { + "epoch": 0.1, + "learning_rate": 0.0007291634904714143, + "loss": 1.7928, + "theoretical_loss": 4.140345208320425, + "tokens_seen": 322306048 + }, + { + "epoch": 0.1, + "learning_rate": 0.000729147442326981, + "loss": 1.5094, + "theoretical_loss": 4.1402482639642235, + "tokens_seen": 322371584 + }, + { + "epoch": 0.1, + "learning_rate": 0.0007291313941825477, + "loss": 1.6999, + "theoretical_loss": 4.140151344831164, + "tokens_seen": 322437120 + }, + { + "epoch": 0.1, + "learning_rate": 0.0007291153460381144, + "loss": 1.6499, + "theoretical_loss": 4.14005445090956, + "tokens_seen": 322502656 + }, + { + "epoch": 0.1, + "learning_rate": 0.0007290992978936812, + "loss": 1.7222, + "theoretical_loss": 4.13995758218773, + "tokens_seen": 322568192 + }, + { + "epoch": 0.1, + "learning_rate": 0.0007290832497492478, + "loss": 1.7248, + "theoretical_loss": 4.1398607386540025, + "tokens_seen": 322633728 + }, + { + "epoch": 0.1, + "learning_rate": 0.0007290672016048146, + "loss": 1.66, + "theoretical_loss": 4.139763920296716, + "tokens_seen": 322699264 + }, + { + "epoch": 0.1, + "learning_rate": 0.0007290511534603812, + "loss": 1.8055, + "theoretical_loss": 4.139667127104211, + "tokens_seen": 322764800 + }, + { + "epoch": 0.1, + "learning_rate": 0.000729035105315948, + "loss": 1.6656, + "theoretical_loss": 4.139570359064842, + "tokens_seen": 322830336 + }, + { + "epoch": 0.1, + "learning_rate": 0.0007290190571715146, + "loss": 1.6534, + "theoretical_loss": 4.139473616166967, + "tokens_seen": 322895872 + }, + { + "epoch": 0.1, + "learning_rate": 0.0007290030090270814, + "loss": 1.5813, + "theoretical_loss": 4.139376898398953, + "tokens_seen": 322961408 + }, + { + "epoch": 0.1, + "learning_rate": 0.000728986960882648, + "loss": 1.6476, + "theoretical_loss": 4.139280205749175, + "tokens_seen": 323026944 + }, + { + "epoch": 0.1, + "learning_rate": 0.0007289709127382148, + "loss": 1.5969, + "theoretical_loss": 4.139183538206017, + "tokens_seen": 323092480 + }, + { + "epoch": 0.1, + "learning_rate": 0.0007289548645937814, + "loss": 1.6109, + "theoretical_loss": 4.139086895757868, + "tokens_seen": 323158016 + }, + { + "epoch": 0.1, + "learning_rate": 0.0007289388164493481, + "loss": 1.7052, + "theoretical_loss": 4.138990278393125, + "tokens_seen": 323223552 + }, + { + "epoch": 0.1, + "learning_rate": 0.0007289227683049148, + "loss": 1.5856, + "theoretical_loss": 4.138893686100196, + "tokens_seen": 323289088 + }, + { + "epoch": 0.1, + "learning_rate": 0.0007289067201604815, + "loss": 1.8015, + "theoretical_loss": 4.138797118867493, + "tokens_seen": 323354624 + }, + { + "epoch": 0.1, + "learning_rate": 0.0007288906720160482, + "loss": 1.5705, + "theoretical_loss": 4.138700576683437, + "tokens_seen": 323420160 + }, + { + "epoch": 0.1, + "learning_rate": 0.0007288746238716149, + "loss": 1.6468, + "theoretical_loss": 4.138604059536458, + "tokens_seen": 323485696 + }, + { + "epoch": 0.1, + "learning_rate": 0.0007288585757271816, + "loss": 1.6534, + "theoretical_loss": 4.138507567414993, + "tokens_seen": 323551232 + }, + { + "epoch": 0.1, + "learning_rate": 0.0007288425275827483, + "loss": 1.7688, + "theoretical_loss": 4.138411100307485, + "tokens_seen": 323616768 + }, + { + "epoch": 0.1, + "learning_rate": 0.000728826479438315, + "loss": 1.5191, + "theoretical_loss": 4.138314658202385, + "tokens_seen": 323682304 + }, + { + "epoch": 0.1, + "learning_rate": 0.0007288104312938817, + "loss": 1.6889, + "theoretical_loss": 4.138218241088156, + "tokens_seen": 323747840 + }, + { + "epoch": 0.1, + "learning_rate": 0.0007287943831494483, + "loss": 1.5635, + "theoretical_loss": 4.138121848953261, + "tokens_seen": 323813376 + }, + { + "epoch": 0.1, + "learning_rate": 0.0007287783350050151, + "loss": 1.6332, + "theoretical_loss": 4.138025481786178, + "tokens_seen": 323878912 + }, + { + "epoch": 0.1, + "learning_rate": 0.0007287622868605817, + "loss": 1.6482, + "theoretical_loss": 4.137929139575387, + "tokens_seen": 323944448 + }, + { + "epoch": 0.1, + "learning_rate": 0.0007287462387161485, + "loss": 1.6532, + "theoretical_loss": 4.137832822309382, + "tokens_seen": 324009984 + }, + { + "epoch": 0.1, + "learning_rate": 0.0007287301905717151, + "loss": 1.7089, + "theoretical_loss": 4.1377365299766575, + "tokens_seen": 324075520 + }, + { + "epoch": 0.1, + "learning_rate": 0.0007287141424272819, + "loss": 1.5545, + "theoretical_loss": 4.13764026256572, + "tokens_seen": 324141056 + }, + { + "epoch": 0.1, + "learning_rate": 0.0007286980942828485, + "loss": 1.6513, + "theoretical_loss": 4.137544020065082, + "tokens_seen": 324206592 + }, + { + "epoch": 0.1, + "learning_rate": 0.0007286820461384153, + "loss": 1.5679, + "theoretical_loss": 4.1374478024632655, + "tokens_seen": 324272128 + }, + { + "epoch": 0.1, + "learning_rate": 0.0007286659979939819, + "loss": 1.6655, + "theoretical_loss": 4.137351609748798, + "tokens_seen": 324337664 + }, + { + "epoch": 0.1, + "objective/train/docs_used": 124782, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.595249891281128, + "objective/train/theoretical_loss": 4.137255441910215, + "objective/train/tokens_used": 344863200, + "theoretical_loss": 4.137255441910215, + "tokens_seen": 324403200 + }, + { + "epoch": 0.1, + "learning_rate": 0.0007286499498495487, + "loss": 1.6349, + "theoretical_loss": 4.137255441910215, + "tokens_seen": 324403200 + }, + { + "epoch": 0.1, + "learning_rate": 0.0007286339017051153, + "loss": 1.6369, + "theoretical_loss": 4.13715929893606, + "tokens_seen": 324468736 + }, + { + "epoch": 0.1, + "learning_rate": 0.0007286178535606821, + "loss": 1.6859, + "theoretical_loss": 4.137063180814886, + "tokens_seen": 324534272 + }, + { + "epoch": 0.1, + "learning_rate": 0.0007286018054162487, + "loss": 1.7276, + "theoretical_loss": 4.136967087535249, + "tokens_seen": 324599808 + }, + { + "epoch": 0.1, + "learning_rate": 0.0007285857572718155, + "loss": 1.6133, + "theoretical_loss": 4.136871019085717, + "tokens_seen": 324665344 + }, + { + "epoch": 0.1, + "learning_rate": 0.0007285697091273821, + "loss": 1.6468, + "theoretical_loss": 4.136774975454863, + "tokens_seen": 324730880 + }, + { + "epoch": 0.1, + "learning_rate": 0.0007285536609829489, + "loss": 1.7783, + "theoretical_loss": 4.136678956631269, + "tokens_seen": 324796416 + }, + { + "epoch": 0.1, + "learning_rate": 0.0007285376128385155, + "loss": 1.656, + "theoretical_loss": 4.136582962603523, + "tokens_seen": 324861952 + }, + { + "epoch": 0.1, + "learning_rate": 0.0007285215646940823, + "loss": 1.6175, + "theoretical_loss": 4.136486993360222, + "tokens_seen": 324927488 + }, + { + "epoch": 0.1, + "learning_rate": 0.0007285055165496489, + "loss": 1.6955, + "theoretical_loss": 4.136391048889971, + "tokens_seen": 324993024 + }, + { + "epoch": 0.1, + "learning_rate": 0.0007284894684052157, + "loss": 1.6054, + "theoretical_loss": 4.136295129181381, + "tokens_seen": 325058560 + }, + { + "epoch": 0.1, + "learning_rate": 0.0007284734202607823, + "loss": 1.5704, + "theoretical_loss": 4.136199234223069, + "tokens_seen": 325124096 + }, + { + "epoch": 0.1, + "learning_rate": 0.0007284573721163491, + "loss": 1.7375, + "theoretical_loss": 4.136103364003665, + "tokens_seen": 325189632 + }, + { + "epoch": 0.1, + "learning_rate": 0.0007284413239719157, + "loss": 1.6142, + "theoretical_loss": 4.1360075185118, + "tokens_seen": 325255168 + }, + { + "epoch": 0.1, + "learning_rate": 0.0007284252758274825, + "loss": 1.695, + "theoretical_loss": 4.135911697736118, + "tokens_seen": 325320704 + }, + { + "epoch": 0.1, + "learning_rate": 0.0007284092276830491, + "loss": 1.6791, + "theoretical_loss": 4.135815901665267, + "tokens_seen": 325386240 + }, + { + "epoch": 0.1, + "learning_rate": 0.0007283931795386159, + "loss": 1.6208, + "theoretical_loss": 4.135720130287904, + "tokens_seen": 325451776 + }, + { + "epoch": 0.1, + "learning_rate": 0.0007283771313941825, + "loss": 1.6768, + "theoretical_loss": 4.135624383592694, + "tokens_seen": 325517312 + }, + { + "epoch": 0.1, + "learning_rate": 0.0007283610832497493, + "loss": 1.6091, + "theoretical_loss": 4.135528661568307, + "tokens_seen": 325582848 + }, + { + "epoch": 0.1, + "learning_rate": 0.0007283450351053159, + "loss": 1.7459, + "theoretical_loss": 4.135432964203422, + "tokens_seen": 325648384 + }, + { + "epoch": 0.1, + "learning_rate": 0.0007283289869608827, + "loss": 1.6344, + "theoretical_loss": 4.135337291486727, + "tokens_seen": 325713920 + }, + { + "epoch": 0.1, + "learning_rate": 0.0007283129388164493, + "loss": 1.657, + "theoretical_loss": 4.135241643406915, + "tokens_seen": 325779456 + }, + { + "epoch": 0.1, + "learning_rate": 0.0007282968906720161, + "loss": 1.7096, + "theoretical_loss": 4.135146019952688, + "tokens_seen": 325844992 + }, + { + "epoch": 0.1, + "learning_rate": 0.0007282808425275827, + "loss": 1.5809, + "theoretical_loss": 4.135050421112755, + "tokens_seen": 325910528 + }, + { + "epoch": 0.1, + "learning_rate": 0.0007282647943831495, + "loss": 1.7253, + "theoretical_loss": 4.134954846875832, + "tokens_seen": 325976064 + }, + { + "epoch": 0.1, + "learning_rate": 0.0007282487462387161, + "loss": 1.7571, + "theoretical_loss": 4.134859297230644, + "tokens_seen": 326041600 + }, + { + "epoch": 0.1, + "learning_rate": 0.0007282326980942829, + "loss": 1.5831, + "theoretical_loss": 4.13476377216592, + "tokens_seen": 326107136 + }, + { + "epoch": 0.1, + "learning_rate": 0.0007282166499498495, + "loss": 1.7223, + "theoretical_loss": 4.134668271670401, + "tokens_seen": 326172672 + }, + { + "epoch": 0.1, + "learning_rate": 0.0007282006018054163, + "loss": 1.6607, + "theoretical_loss": 4.134572795732832, + "tokens_seen": 326238208 + }, + { + "epoch": 0.1, + "learning_rate": 0.0007281845536609829, + "loss": 1.5679, + "theoretical_loss": 4.134477344341967, + "tokens_seen": 326303744 + }, + { + "epoch": 0.1, + "learning_rate": 0.0007281685055165497, + "loss": 1.6929, + "theoretical_loss": 4.134381917486565, + "tokens_seen": 326369280 + }, + { + "epoch": 0.1, + "learning_rate": 0.0007281524573721163, + "loss": 1.6992, + "theoretical_loss": 4.134286515155398, + "tokens_seen": 326434816 + }, + { + "epoch": 0.1, + "learning_rate": 0.0007281364092276831, + "loss": 1.5802, + "theoretical_loss": 4.134191137337238, + "tokens_seen": 326500352 + }, + { + "epoch": 0.1, + "learning_rate": 0.0007281203610832497, + "loss": 1.6796, + "theoretical_loss": 4.134095784020871, + "tokens_seen": 326565888 + }, + { + "epoch": 0.1, + "learning_rate": 0.0007281043129388165, + "loss": 1.7422, + "theoretical_loss": 4.1340004551950855, + "tokens_seen": 326631424 + }, + { + "epoch": 0.1, + "learning_rate": 0.0007280882647943831, + "loss": 1.566, + "theoretical_loss": 4.133905150848681, + "tokens_seen": 326696960 + }, + { + "epoch": 0.1, + "learning_rate": 0.0007280722166499499, + "loss": 1.6093, + "theoretical_loss": 4.133809870970461, + "tokens_seen": 326762496 + }, + { + "epoch": 0.1, + "learning_rate": 0.0007280561685055165, + "loss": 1.6165, + "theoretical_loss": 4.1337146155492395, + "tokens_seen": 326828032 + }, + { + "epoch": 0.1, + "learning_rate": 0.0007280401203610833, + "loss": 1.5444, + "theoretical_loss": 4.133619384573836, + "tokens_seen": 326893568 + }, + { + "epoch": 0.1, + "learning_rate": 0.0007280240722166499, + "loss": 1.5518, + "theoretical_loss": 4.1335241780330785, + "tokens_seen": 326959104 + }, + { + "epoch": 0.1, + "learning_rate": 0.0007280080240722167, + "loss": 1.6327, + "theoretical_loss": 4.1334289959158, + "tokens_seen": 327024640 + }, + { + "epoch": 0.1, + "learning_rate": 0.0007279919759277833, + "loss": 1.6546, + "theoretical_loss": 4.133333838210844, + "tokens_seen": 327090176 + }, + { + "epoch": 0.1, + "learning_rate": 0.0007279759277833501, + "loss": 1.7563, + "theoretical_loss": 4.133238704907059, + "tokens_seen": 327155712 + }, + { + "epoch": 0.1, + "learning_rate": 0.0007279598796389167, + "loss": 1.907, + "theoretical_loss": 4.133143595993303, + "tokens_seen": 327221248 + }, + { + "epoch": 0.1, + "learning_rate": 0.0007279438314944835, + "loss": 1.7002, + "theoretical_loss": 4.133048511458438, + "tokens_seen": 327286784 + }, + { + "epoch": 0.1, + "learning_rate": 0.0007279277833500503, + "loss": 1.6967, + "theoretical_loss": 4.132953451291337, + "tokens_seen": 327352320 + }, + { + "epoch": 0.1, + "learning_rate": 0.0007279117352056169, + "loss": 1.6948, + "theoretical_loss": 4.132858415480877, + "tokens_seen": 327417856 + }, + { + "epoch": 0.1, + "learning_rate": 0.0007278956870611837, + "loss": 1.7828, + "theoretical_loss": 4.132763404015946, + "tokens_seen": 327483392 + }, + { + "epoch": 0.1, + "learning_rate": 0.0007278796389167503, + "loss": 1.6595, + "theoretical_loss": 4.132668416885435, + "tokens_seen": 327548928 + }, + { + "epoch": 0.1, + "learning_rate": 0.0007278635907723171, + "loss": 1.6817, + "theoretical_loss": 4.132573454078246, + "tokens_seen": 327614464 + }, + { + "debugging/Compilability": 1.0, + "debugging/distinct-1-grams": 0.7246169962973871, + "debugging/entropy-1-grams": 5.13190351820591, + "debugging/length": 472.8, + "debugging/num_segments": 10, + "debugging/score": 0.0033104506818385365, + "debugging/score_std": 0.0017105448100298692, + "epoch": 0.1, + "objective/train/docs_used": 126013, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.7461847066879272, + "objective/train/theoretical_loss": 4.132478515583286, + "objective/train/tokens_used": 348140000, + "theoretical_loss": 4.132478515583286, + "tokens_seen": 327680000 + }, + { + "epoch": 0.1, + "learning_rate": 0.0007278475426278837, + "loss": 1.7713, + "theoretical_loss": 4.132478515583286, + "tokens_seen": 327680000 + }, + { + "epoch": 0.1, + "learning_rate": 0.0007278314944834505, + "loss": 1.6751, + "theoretical_loss": 4.132383601389471, + "tokens_seen": 327745536 + }, + { + "epoch": 0.1, + "learning_rate": 0.0007278154463390171, + "loss": 1.7542, + "theoretical_loss": 4.132288711485721, + "tokens_seen": 327811072 + }, + { + "epoch": 0.1, + "learning_rate": 0.0007277993981945839, + "loss": 1.7457, + "theoretical_loss": 4.132193845860969, + "tokens_seen": 327876608 + }, + { + "epoch": 0.1, + "learning_rate": 0.0007277833500501505, + "loss": 1.6538, + "theoretical_loss": 4.132099004504149, + "tokens_seen": 327942144 + }, + { + "epoch": 0.1, + "learning_rate": 0.0007277673019057173, + "loss": 1.7793, + "theoretical_loss": 4.132004187404206, + "tokens_seen": 328007680 + }, + { + "epoch": 0.1, + "learning_rate": 0.0007277512537612839, + "loss": 1.6401, + "theoretical_loss": 4.131909394550092, + "tokens_seen": 328073216 + }, + { + "epoch": 0.1, + "learning_rate": 0.0007277352056168507, + "loss": 1.7683, + "theoretical_loss": 4.131814625930764, + "tokens_seen": 328138752 + }, + { + "epoch": 0.1, + "learning_rate": 0.0007277191574724173, + "loss": 1.6361, + "theoretical_loss": 4.13171988153519, + "tokens_seen": 328204288 + }, + { + "epoch": 0.1, + "learning_rate": 0.0007277031093279841, + "loss": 1.5762, + "theoretical_loss": 4.131625161352342, + "tokens_seen": 328269824 + }, + { + "epoch": 0.1, + "learning_rate": 0.0007276870611835507, + "loss": 1.8849, + "theoretical_loss": 4.131530465371199, + "tokens_seen": 328335360 + }, + { + "epoch": 0.1, + "learning_rate": 0.0007276710130391175, + "loss": 1.6696, + "theoretical_loss": 4.131435793580749, + "tokens_seen": 328400896 + }, + { + "epoch": 0.1, + "learning_rate": 0.0007276549648946841, + "loss": 1.5789, + "theoretical_loss": 4.131341145969989, + "tokens_seen": 328466432 + }, + { + "epoch": 0.1, + "learning_rate": 0.0007276389167502509, + "loss": 1.528, + "theoretical_loss": 4.131246522527918, + "tokens_seen": 328531968 + }, + { + "epoch": 0.1, + "learning_rate": 0.0007276228686058175, + "loss": 1.6979, + "theoretical_loss": 4.1311519232435465, + "tokens_seen": 328597504 + }, + { + "epoch": 0.1, + "learning_rate": 0.0007276068204613843, + "loss": 1.8179, + "theoretical_loss": 4.131057348105889, + "tokens_seen": 328663040 + }, + { + "epoch": 0.1, + "learning_rate": 0.0007275907723169509, + "loss": 1.5262, + "theoretical_loss": 4.1309627971039715, + "tokens_seen": 328728576 + }, + { + "epoch": 0.1, + "learning_rate": 0.0007275747241725177, + "loss": 1.6764, + "theoretical_loss": 4.130868270226824, + "tokens_seen": 328794112 + }, + { + "epoch": 0.1, + "learning_rate": 0.0007275586760280843, + "loss": 1.5785, + "theoretical_loss": 4.130773767463483, + "tokens_seen": 328859648 + }, + { + "epoch": 0.1, + "learning_rate": 0.000727542627883651, + "loss": 1.5802, + "theoretical_loss": 4.130679288802993, + "tokens_seen": 328925184 + }, + { + "epoch": 0.1, + "learning_rate": 0.0007275265797392177, + "loss": 1.6724, + "theoretical_loss": 4.1305848342344085, + "tokens_seen": 328990720 + }, + { + "epoch": 0.1, + "learning_rate": 0.0007275105315947844, + "loss": 1.7448, + "theoretical_loss": 4.130490403746787, + "tokens_seen": 329056256 + }, + { + "epoch": 0.1, + "learning_rate": 0.0007274944834503511, + "loss": 1.6348, + "theoretical_loss": 4.130395997329196, + "tokens_seen": 329121792 + }, + { + "epoch": 0.1, + "learning_rate": 0.0007274784353059178, + "loss": 1.6917, + "theoretical_loss": 4.1303016149707075, + "tokens_seen": 329187328 + }, + { + "epoch": 0.1, + "learning_rate": 0.0007274623871614845, + "loss": 1.6289, + "theoretical_loss": 4.130207256660404, + "tokens_seen": 329252864 + }, + { + "epoch": 0.1, + "learning_rate": 0.0007274463390170512, + "loss": 1.6789, + "theoretical_loss": 4.130112922387372, + "tokens_seen": 329318400 + }, + { + "epoch": 0.1, + "learning_rate": 0.0007274302908726179, + "loss": 1.6537, + "theoretical_loss": 4.130018612140707, + "tokens_seen": 329383936 + }, + { + "epoch": 0.1, + "learning_rate": 0.0007274142427281846, + "loss": 1.6498, + "theoretical_loss": 4.12992432590951, + "tokens_seen": 329449472 + }, + { + "epoch": 0.1, + "learning_rate": 0.0007273981945837512, + "loss": 1.6424, + "theoretical_loss": 4.129830063682892, + "tokens_seen": 329515008 + }, + { + "epoch": 0.1, + "learning_rate": 0.000727382146439318, + "loss": 1.6068, + "theoretical_loss": 4.129735825449968, + "tokens_seen": 329580544 + }, + { + "epoch": 0.1, + "learning_rate": 0.0007273660982948846, + "loss": 1.6234, + "theoretical_loss": 4.129641611199862, + "tokens_seen": 329646080 + }, + { + "epoch": 0.1, + "learning_rate": 0.0007273500501504514, + "loss": 1.7774, + "theoretical_loss": 4.1295474209217025, + "tokens_seen": 329711616 + }, + { + "epoch": 0.1, + "learning_rate": 0.000727334002006018, + "loss": 1.6167, + "theoretical_loss": 4.12945325460463, + "tokens_seen": 329777152 + }, + { + "epoch": 0.1, + "learning_rate": 0.0007273179538615848, + "loss": 1.6614, + "theoretical_loss": 4.129359112237787, + "tokens_seen": 329842688 + }, + { + "epoch": 0.1, + "learning_rate": 0.0007273019057171514, + "loss": 1.6392, + "theoretical_loss": 4.1292649938103265, + "tokens_seen": 329908224 + }, + { + "epoch": 0.1, + "learning_rate": 0.0007272858575727182, + "loss": 1.6729, + "theoretical_loss": 4.129170899311406, + "tokens_seen": 329973760 + }, + { + "epoch": 0.1, + "learning_rate": 0.0007272698094282848, + "loss": 1.7955, + "theoretical_loss": 4.129076828730192, + "tokens_seen": 330039296 + }, + { + "epoch": 0.1, + "learning_rate": 0.0007272537612838516, + "loss": 1.6903, + "theoretical_loss": 4.128982782055858, + "tokens_seen": 330104832 + }, + { + "epoch": 0.1, + "learning_rate": 0.0007272377131394182, + "loss": 1.6232, + "theoretical_loss": 4.128888759277583, + "tokens_seen": 330170368 + }, + { + "epoch": 0.1, + "learning_rate": 0.000727221664994985, + "loss": 1.6741, + "theoretical_loss": 4.1287947603845545, + "tokens_seen": 330235904 + }, + { + "epoch": 0.1, + "learning_rate": 0.0007272056168505516, + "loss": 1.7388, + "theoretical_loss": 4.128700785365965, + "tokens_seen": 330301440 + }, + { + "epoch": 0.1, + "learning_rate": 0.0007271895687061184, + "loss": 1.7391, + "theoretical_loss": 4.128606834211018, + "tokens_seen": 330366976 + }, + { + "epoch": 0.1, + "learning_rate": 0.000727173520561685, + "loss": 1.5852, + "theoretical_loss": 4.128512906908919, + "tokens_seen": 330432512 + }, + { + "epoch": 0.1, + "learning_rate": 0.0007271574724172518, + "loss": 1.6215, + "theoretical_loss": 4.128419003448886, + "tokens_seen": 330498048 + }, + { + "epoch": 0.1, + "learning_rate": 0.0007271414242728184, + "loss": 1.732, + "theoretical_loss": 4.128325123820138, + "tokens_seen": 330563584 + }, + { + "epoch": 0.1, + "learning_rate": 0.0007271253761283852, + "loss": 1.7449, + "theoretical_loss": 4.1282312680119055, + "tokens_seen": 330629120 + }, + { + "epoch": 0.1, + "learning_rate": 0.0007271093279839518, + "loss": 1.7273, + "theoretical_loss": 4.128137436013425, + "tokens_seen": 330694656 + }, + { + "epoch": 0.1, + "learning_rate": 0.0007270932798395186, + "loss": 1.6628, + "theoretical_loss": 4.128043627813939, + "tokens_seen": 330760192 + }, + { + "epoch": 0.1, + "learning_rate": 0.0007270772316950852, + "loss": 1.636, + "theoretical_loss": 4.127949843402699, + "tokens_seen": 330825728 + }, + { + "epoch": 0.1, + "learning_rate": 0.000727061183550652, + "loss": 1.796, + "theoretical_loss": 4.12785608276896, + "tokens_seen": 330891264 + }, + { + "epoch": 0.1, + "objective/train/docs_used": 127130, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.7275724411010742, + "objective/train/theoretical_loss": 4.127762345901987, + "objective/train/tokens_used": 351416800, + "theoretical_loss": 4.127762345901987, + "tokens_seen": 330956800 + }, + { + "epoch": 0.1, + "learning_rate": 0.0007270451354062186, + "loss": 1.7584, + "theoretical_loss": 4.127762345901987, + "tokens_seen": 330956800 + }, + { + "epoch": 0.1, + "learning_rate": 0.0007270290872617854, + "loss": 1.6251, + "theoretical_loss": 4.1276686327910515, + "tokens_seen": 331022336 + }, + { + "epoch": 0.1, + "learning_rate": 0.000727013039117352, + "loss": 1.5752, + "theoretical_loss": 4.127574943425429, + "tokens_seen": 331087872 + }, + { + "epoch": 0.1, + "learning_rate": 0.0007269969909729188, + "loss": 1.8274, + "theoretical_loss": 4.127481277794408, + "tokens_seen": 331153408 + }, + { + "epoch": 0.1, + "learning_rate": 0.0007269809428284854, + "loss": 1.7474, + "theoretical_loss": 4.1273876358872785, + "tokens_seen": 331218944 + }, + { + "epoch": 0.1, + "learning_rate": 0.0007269648946840522, + "loss": 1.6186, + "theoretical_loss": 4.127294017693339, + "tokens_seen": 331284480 + }, + { + "epoch": 0.1, + "learning_rate": 0.0007269488465396188, + "loss": 1.7498, + "theoretical_loss": 4.127200423201897, + "tokens_seen": 331350016 + }, + { + "epoch": 0.1, + "learning_rate": 0.0007269327983951856, + "loss": 1.6575, + "theoretical_loss": 4.127106852402264, + "tokens_seen": 331415552 + }, + { + "epoch": 0.1, + "learning_rate": 0.0007269167502507522, + "loss": 1.6389, + "theoretical_loss": 4.127013305283759, + "tokens_seen": 331481088 + }, + { + "epoch": 0.1, + "learning_rate": 0.000726900702106319, + "loss": 1.6126, + "theoretical_loss": 4.12691978183571, + "tokens_seen": 331546624 + }, + { + "epoch": 0.1, + "learning_rate": 0.0007268846539618856, + "loss": 1.5577, + "theoretical_loss": 4.126826282047449, + "tokens_seen": 331612160 + }, + { + "epoch": 0.1, + "learning_rate": 0.0007268686058174524, + "loss": 1.6788, + "theoretical_loss": 4.1267328059083175, + "tokens_seen": 331677696 + }, + { + "epoch": 0.1, + "learning_rate": 0.000726852557673019, + "loss": 1.6551, + "theoretical_loss": 4.126639353407663, + "tokens_seen": 331743232 + }, + { + "epoch": 0.1, + "learning_rate": 0.0007268365095285858, + "loss": 1.6633, + "theoretical_loss": 4.12654592453484, + "tokens_seen": 331808768 + }, + { + "epoch": 0.1, + "learning_rate": 0.0007268204613841524, + "loss": 1.7105, + "theoretical_loss": 4.126452519279208, + "tokens_seen": 331874304 + }, + { + "epoch": 0.1, + "learning_rate": 0.0007268044132397192, + "loss": 1.6527, + "theoretical_loss": 4.126359137630136, + "tokens_seen": 331939840 + }, + { + "epoch": 0.1, + "learning_rate": 0.0007267883650952858, + "loss": 1.7695, + "theoretical_loss": 4.126265779576999, + "tokens_seen": 332005376 + }, + { + "epoch": 0.1, + "learning_rate": 0.0007267723169508526, + "loss": 1.6924, + "theoretical_loss": 4.126172445109178, + "tokens_seen": 332070912 + }, + { + "epoch": 0.1, + "learning_rate": 0.0007267562688064192, + "loss": 1.7495, + "theoretical_loss": 4.126079134216063, + "tokens_seen": 332136448 + }, + { + "epoch": 0.1, + "learning_rate": 0.000726740220661986, + "loss": 1.5943, + "theoretical_loss": 4.125985846887049, + "tokens_seen": 332201984 + }, + { + "epoch": 0.1, + "learning_rate": 0.0007267241725175528, + "loss": 1.5518, + "theoretical_loss": 4.125892583111538, + "tokens_seen": 332267520 + }, + { + "epoch": 0.1, + "learning_rate": 0.0007267081243731194, + "loss": 1.5459, + "theoretical_loss": 4.125799342878939, + "tokens_seen": 332333056 + }, + { + "epoch": 0.1, + "learning_rate": 0.0007266920762286862, + "loss": 1.6181, + "theoretical_loss": 4.1257061261786685, + "tokens_seen": 332398592 + }, + { + "epoch": 0.1, + "learning_rate": 0.0007266760280842528, + "loss": 1.5781, + "theoretical_loss": 4.12561293300015, + "tokens_seen": 332464128 + }, + { + "epoch": 0.1, + "learning_rate": 0.0007266599799398196, + "loss": 1.5941, + "theoretical_loss": 4.125519763332813, + "tokens_seen": 332529664 + }, + { + "epoch": 0.1, + "learning_rate": 0.0007266439317953862, + "loss": 1.7738, + "theoretical_loss": 4.1254266171660925, + "tokens_seen": 332595200 + }, + { + "epoch": 0.1, + "learning_rate": 0.000726627883650953, + "loss": 1.7125, + "theoretical_loss": 4.125333494489434, + "tokens_seen": 332660736 + }, + { + "epoch": 0.1, + "learning_rate": 0.0007266118355065196, + "loss": 1.7469, + "theoretical_loss": 4.125240395292286, + "tokens_seen": 332726272 + }, + { + "epoch": 0.1, + "learning_rate": 0.0007265957873620864, + "loss": 1.7261, + "theoretical_loss": 4.125147319564107, + "tokens_seen": 332791808 + }, + { + "epoch": 0.1, + "learning_rate": 0.000726579739217653, + "loss": 1.7483, + "theoretical_loss": 4.12505426729436, + "tokens_seen": 332857344 + }, + { + "epoch": 0.1, + "learning_rate": 0.0007265636910732198, + "loss": 1.5552, + "theoretical_loss": 4.124961238472517, + "tokens_seen": 332922880 + }, + { + "epoch": 0.1, + "learning_rate": 0.0007265476429287864, + "loss": 1.6348, + "theoretical_loss": 4.124868233088054, + "tokens_seen": 332988416 + }, + { + "epoch": 0.1, + "learning_rate": 0.0007265315947843532, + "loss": 1.668, + "theoretical_loss": 4.124775251130455, + "tokens_seen": 333053952 + }, + { + "epoch": 0.1, + "learning_rate": 0.0007265155466399198, + "loss": 1.6624, + "theoretical_loss": 4.124682292589212, + "tokens_seen": 333119488 + }, + { + "epoch": 0.1, + "learning_rate": 0.0007264994984954866, + "loss": 1.5358, + "theoretical_loss": 4.124589357453823, + "tokens_seen": 333185024 + }, + { + "epoch": 0.1, + "learning_rate": 0.0007264834503510532, + "loss": 1.6988, + "theoretical_loss": 4.124496445713792, + "tokens_seen": 333250560 + }, + { + "epoch": 0.1, + "learning_rate": 0.00072646740220662, + "loss": 1.6133, + "theoretical_loss": 4.12440355735863, + "tokens_seen": 333316096 + }, + { + "epoch": 0.1, + "learning_rate": 0.0007264513540621866, + "loss": 1.7293, + "theoretical_loss": 4.124310692377857, + "tokens_seen": 333381632 + }, + { + "epoch": 0.1, + "learning_rate": 0.0007264353059177534, + "loss": 1.725, + "theoretical_loss": 4.124217850760996, + "tokens_seen": 333447168 + }, + { + "epoch": 0.1, + "learning_rate": 0.00072641925777332, + "loss": 1.6308, + "theoretical_loss": 4.124125032497578, + "tokens_seen": 333512704 + }, + { + "epoch": 0.1, + "learning_rate": 0.0007264032096288868, + "loss": 1.5885, + "theoretical_loss": 4.124032237577143, + "tokens_seen": 333578240 + }, + { + "epoch": 0.1, + "learning_rate": 0.0007263871614844534, + "loss": 1.5615, + "theoretical_loss": 4.1239394659892366, + "tokens_seen": 333643776 + }, + { + "epoch": 0.1, + "learning_rate": 0.0007263711133400202, + "loss": 1.7113, + "theoretical_loss": 4.12384671772341, + "tokens_seen": 333709312 + }, + { + "epoch": 0.1, + "learning_rate": 0.0007263550651955868, + "loss": 1.649, + "theoretical_loss": 4.12375399276922, + "tokens_seen": 333774848 + }, + { + "epoch": 0.1, + "learning_rate": 0.0007263390170511536, + "loss": 1.7301, + "theoretical_loss": 4.1236612911162345, + "tokens_seen": 333840384 + }, + { + "epoch": 0.1, + "learning_rate": 0.0007263229689067202, + "loss": 1.6086, + "theoretical_loss": 4.123568612754024, + "tokens_seen": 333905920 + }, + { + "epoch": 0.1, + "learning_rate": 0.000726306920762287, + "loss": 1.5694, + "theoretical_loss": 4.123475957672168, + "tokens_seen": 333971456 + }, + { + "epoch": 0.1, + "learning_rate": 0.0007262908726178536, + "loss": 1.7703, + "theoretical_loss": 4.123383325860252, + "tokens_seen": 334036992 + }, + { + "epoch": 0.1, + "learning_rate": 0.0007262748244734204, + "loss": 1.7474, + "theoretical_loss": 4.123290717307867, + "tokens_seen": 334102528 + }, + { + "epoch": 0.1, + "learning_rate": 0.000726258776328987, + "loss": 1.6758, + "theoretical_loss": 4.123198132004614, + "tokens_seen": 334168064 + }, + { + "epoch": 0.1, + "objective/train/docs_used": 128373, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.8325588703155518, + "objective/train/theoretical_loss": 4.123105569940097, + "objective/train/tokens_used": 354693600, + "theoretical_loss": 4.123105569940097, + "tokens_seen": 334233600 + }, + { + "epoch": 0.1, + "learning_rate": 0.0007262427281845538, + "loss": 1.7299, + "theoretical_loss": 4.123105569940097, + "tokens_seen": 334233600 + }, + { + "epoch": 0.1, + "learning_rate": 0.0007262266800401204, + "loss": 1.5525, + "theoretical_loss": 4.123013031103927, + "tokens_seen": 334299136 + }, + { + "epoch": 0.1, + "learning_rate": 0.0007262106318956872, + "loss": 1.7178, + "theoretical_loss": 4.122920515485726, + "tokens_seen": 334364672 + }, + { + "epoch": 0.1, + "learning_rate": 0.0007261945837512538, + "loss": 1.5925, + "theoretical_loss": 4.122828023075117, + "tokens_seen": 334430208 + }, + { + "epoch": 0.1, + "learning_rate": 0.0007261785356068206, + "loss": 1.5875, + "theoretical_loss": 4.122735553861733, + "tokens_seen": 334495744 + }, + { + "epoch": 0.1, + "learning_rate": 0.0007261624874623872, + "loss": 1.7417, + "theoretical_loss": 4.122643107835213, + "tokens_seen": 334561280 + }, + { + "epoch": 0.1, + "learning_rate": 0.000726146439317954, + "loss": 1.5897, + "theoretical_loss": 4.122550684985203, + "tokens_seen": 334626816 + }, + { + "epoch": 0.1, + "learning_rate": 0.0007261303911735206, + "loss": 1.723, + "theoretical_loss": 4.122458285301355, + "tokens_seen": 334692352 + }, + { + "epoch": 0.1, + "learning_rate": 0.0007261143430290873, + "loss": 1.7686, + "theoretical_loss": 4.122365908773327, + "tokens_seen": 334757888 + }, + { + "epoch": 0.1, + "learning_rate": 0.000726098294884654, + "loss": 1.6723, + "theoretical_loss": 4.1222735553907865, + "tokens_seen": 334823424 + }, + { + "epoch": 0.1, + "learning_rate": 0.0007260822467402207, + "loss": 1.6276, + "theoretical_loss": 4.122181225143404, + "tokens_seen": 334888960 + }, + { + "epoch": 0.1, + "learning_rate": 0.0007260661985957874, + "loss": 1.7922, + "theoretical_loss": 4.122088918020857, + "tokens_seen": 334954496 + }, + { + "epoch": 0.1, + "learning_rate": 0.0007260501504513541, + "loss": 1.7212, + "theoretical_loss": 4.121996634012834, + "tokens_seen": 335020032 + }, + { + "epoch": 0.1, + "learning_rate": 0.0007260341023069208, + "loss": 1.6013, + "theoretical_loss": 4.121904373109024, + "tokens_seen": 335085568 + }, + { + "epoch": 0.1, + "learning_rate": 0.0007260180541624875, + "loss": 1.7906, + "theoretical_loss": 4.121812135299129, + "tokens_seen": 335151104 + }, + { + "epoch": 0.1, + "learning_rate": 0.0007260020060180541, + "loss": 1.625, + "theoretical_loss": 4.1217199205728505, + "tokens_seen": 335216640 + }, + { + "epoch": 0.1, + "learning_rate": 0.0007259859578736209, + "loss": 1.6891, + "theoretical_loss": 4.121627728919903, + "tokens_seen": 335282176 + }, + { + "epoch": 0.1, + "learning_rate": 0.0007259699097291875, + "loss": 1.7452, + "theoretical_loss": 4.121535560330004, + "tokens_seen": 335347712 + }, + { + "epoch": 0.1, + "learning_rate": 0.0007259538615847543, + "loss": 1.5871, + "theoretical_loss": 4.1214434147928785, + "tokens_seen": 335413248 + }, + { + "epoch": 0.1, + "learning_rate": 0.0007259378134403209, + "loss": 1.6705, + "theoretical_loss": 4.121351292298257, + "tokens_seen": 335478784 + }, + { + "epoch": 0.1, + "learning_rate": 0.0007259217652958877, + "loss": 1.7067, + "theoretical_loss": 4.12125919283588, + "tokens_seen": 335544320 + }, + { + "epoch": 0.1, + "learning_rate": 0.0007259057171514543, + "loss": 1.6814, + "theoretical_loss": 4.12116711639549, + "tokens_seen": 335609856 + }, + { + "epoch": 0.1, + "learning_rate": 0.0007258896690070211, + "loss": 1.7141, + "theoretical_loss": 4.12107506296684, + "tokens_seen": 335675392 + }, + { + "epoch": 0.1, + "learning_rate": 0.0007258736208625877, + "loss": 1.582, + "theoretical_loss": 4.120983032539687, + "tokens_seen": 335740928 + }, + { + "epoch": 0.1, + "learning_rate": 0.0007258575727181545, + "loss": 1.7005, + "theoretical_loss": 4.120891025103796, + "tokens_seen": 335806464 + }, + { + "epoch": 0.1, + "learning_rate": 0.0007258415245737211, + "loss": 1.654, + "theoretical_loss": 4.1207990406489365, + "tokens_seen": 335872000 + }, + { + "epoch": 0.1, + "learning_rate": 0.0007258254764292879, + "loss": 1.7283, + "theoretical_loss": 4.120707079164887, + "tokens_seen": 335937536 + }, + { + "epoch": 0.1, + "learning_rate": 0.0007258094282848545, + "loss": 1.6701, + "theoretical_loss": 4.120615140641433, + "tokens_seen": 336003072 + }, + { + "epoch": 0.1, + "learning_rate": 0.0007257933801404213, + "loss": 1.631, + "theoretical_loss": 4.1205232250683625, + "tokens_seen": 336068608 + }, + { + "epoch": 0.1, + "learning_rate": 0.0007257773319959879, + "loss": 1.6632, + "theoretical_loss": 4.120431332435475, + "tokens_seen": 336134144 + }, + { + "epoch": 0.1, + "learning_rate": 0.0007257612838515547, + "loss": 1.6172, + "theoretical_loss": 4.120339462732572, + "tokens_seen": 336199680 + }, + { + "epoch": 0.1, + "learning_rate": 0.0007257452357071213, + "loss": 1.6976, + "theoretical_loss": 4.120247615949465, + "tokens_seen": 336265216 + }, + { + "epoch": 0.1, + "learning_rate": 0.0007257291875626881, + "loss": 1.7002, + "theoretical_loss": 4.12015579207597, + "tokens_seen": 336330752 + }, + { + "epoch": 0.1, + "learning_rate": 0.0007257131394182547, + "loss": 1.6965, + "theoretical_loss": 4.12006399110191, + "tokens_seen": 336396288 + }, + { + "epoch": 0.1, + "learning_rate": 0.0007256970912738215, + "loss": 1.7559, + "theoretical_loss": 4.119972213017116, + "tokens_seen": 336461824 + }, + { + "epoch": 0.1, + "learning_rate": 0.0007256810431293881, + "loss": 1.6178, + "theoretical_loss": 4.119880457811424, + "tokens_seen": 336527360 + }, + { + "epoch": 0.1, + "learning_rate": 0.0007256649949849549, + "loss": 1.6174, + "theoretical_loss": 4.119788725474675, + "tokens_seen": 336592896 + }, + { + "epoch": 0.1, + "learning_rate": 0.0007256489468405215, + "loss": 1.7532, + "theoretical_loss": 4.119697015996719, + "tokens_seen": 336658432 + }, + { + "epoch": 0.1, + "learning_rate": 0.0007256328986960883, + "loss": 1.6669, + "theoretical_loss": 4.119605329367413, + "tokens_seen": 336723968 + }, + { + "epoch": 0.1, + "learning_rate": 0.0007256168505516549, + "loss": 1.6543, + "theoretical_loss": 4.119513665576617, + "tokens_seen": 336789504 + }, + { + "epoch": 0.1, + "learning_rate": 0.0007256008024072217, + "loss": 1.7068, + "theoretical_loss": 4.119422024614201, + "tokens_seen": 336855040 + }, + { + "epoch": 0.1, + "learning_rate": 0.0007255847542627883, + "loss": 1.6384, + "theoretical_loss": 4.11933040647004, + "tokens_seen": 336920576 + }, + { + "epoch": 0.1, + "learning_rate": 0.0007255687061183551, + "loss": 1.6557, + "theoretical_loss": 4.119238811134014, + "tokens_seen": 336986112 + }, + { + "epoch": 0.1, + "learning_rate": 0.0007255526579739217, + "loss": 1.6258, + "theoretical_loss": 4.119147238596014, + "tokens_seen": 337051648 + }, + { + "epoch": 0.1, + "learning_rate": 0.0007255366098294885, + "loss": 1.5983, + "theoretical_loss": 4.119055688845931, + "tokens_seen": 337117184 + }, + { + "epoch": 0.1, + "learning_rate": 0.0007255205616850552, + "loss": 1.7568, + "theoretical_loss": 4.118964161873667, + "tokens_seen": 337182720 + }, + { + "epoch": 0.1, + "learning_rate": 0.0007255045135406219, + "loss": 1.6562, + "theoretical_loss": 4.118872657669131, + "tokens_seen": 337248256 + }, + { + "epoch": 0.1, + "learning_rate": 0.0007254884653961886, + "loss": 1.735, + "theoretical_loss": 4.118781176222235, + "tokens_seen": 337313792 + }, + { + "epoch": 0.1, + "learning_rate": 0.0007254724172517553, + "loss": 1.569, + "theoretical_loss": 4.1186897175229, + "tokens_seen": 337379328 + }, + { + "epoch": 0.1, + "learning_rate": 0.000725456369107322, + "loss": 1.7533, + "theoretical_loss": 4.1185982815610505, + "tokens_seen": 337444864 + }, + { + "epoch": 0.1, + "objective/train/docs_used": 129675, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.5856032371520996, + "objective/train/theoretical_loss": 4.118506868326623, + "objective/train/tokens_used": 357970400, + "theoretical_loss": 4.118506868326623, + "tokens_seen": 337510400 + }, + { + "epoch": 0.1, + "learning_rate": 0.0007254403209628887, + "loss": 1.5594, + "theoretical_loss": 4.118506868326623, + "tokens_seen": 337510400 + }, + { + "epoch": 0.1, + "learning_rate": 0.0007254242728184554, + "loss": 1.6836, + "theoretical_loss": 4.118415477809553, + "tokens_seen": 337575936 + }, + { + "epoch": 0.1, + "learning_rate": 0.0007254082246740221, + "loss": 1.5843, + "theoretical_loss": 4.11832410999979, + "tokens_seen": 337641472 + }, + { + "epoch": 0.1, + "learning_rate": 0.0007253921765295888, + "loss": 1.7215, + "theoretical_loss": 4.118232764887285, + "tokens_seen": 337707008 + }, + { + "epoch": 0.1, + "learning_rate": 0.0007253761283851555, + "loss": 1.5607, + "theoretical_loss": 4.118141442461995, + "tokens_seen": 337772544 + }, + { + "epoch": 0.1, + "learning_rate": 0.0007253600802407223, + "loss": 1.602, + "theoretical_loss": 4.118050142713886, + "tokens_seen": 337838080 + }, + { + "epoch": 0.1, + "learning_rate": 0.0007253440320962889, + "loss": 1.6483, + "theoretical_loss": 4.1179588656329305, + "tokens_seen": 337903616 + }, + { + "epoch": 0.1, + "learning_rate": 0.0007253279839518557, + "loss": 1.7324, + "theoretical_loss": 4.117867611209104, + "tokens_seen": 337969152 + }, + { + "epoch": 0.1, + "learning_rate": 0.0007253119358074223, + "loss": 1.7037, + "theoretical_loss": 4.117776379432392, + "tokens_seen": 338034688 + }, + { + "epoch": 0.1, + "learning_rate": 0.000725295887662989, + "loss": 1.5721, + "theoretical_loss": 4.117685170292786, + "tokens_seen": 338100224 + }, + { + "epoch": 0.1, + "learning_rate": 0.0007252798395185557, + "loss": 1.7232, + "theoretical_loss": 4.1175939837802815, + "tokens_seen": 338165760 + }, + { + "epoch": 0.1, + "learning_rate": 0.0007252637913741225, + "loss": 1.5858, + "theoretical_loss": 4.117502819884882, + "tokens_seen": 338231296 + }, + { + "epoch": 0.1, + "learning_rate": 0.0007252477432296891, + "loss": 1.6866, + "theoretical_loss": 4.117411678596596, + "tokens_seen": 338296832 + }, + { + "epoch": 0.1, + "learning_rate": 0.0007252316950852559, + "loss": 1.8084, + "theoretical_loss": 4.117320559905441, + "tokens_seen": 338362368 + }, + { + "epoch": 0.1, + "learning_rate": 0.0007252156469408225, + "loss": 1.758, + "theoretical_loss": 4.117229463801439, + "tokens_seen": 338427904 + }, + { + "epoch": 0.1, + "learning_rate": 0.0007251995987963893, + "loss": 1.7934, + "theoretical_loss": 4.117138390274619, + "tokens_seen": 338493440 + }, + { + "epoch": 0.1, + "learning_rate": 0.0007251835506519559, + "loss": 1.6649, + "theoretical_loss": 4.117047339315014, + "tokens_seen": 338558976 + }, + { + "epoch": 0.1, + "learning_rate": 0.0007251675025075227, + "loss": 1.8013, + "theoretical_loss": 4.116956310912666, + "tokens_seen": 338624512 + }, + { + "epoch": 0.1, + "learning_rate": 0.0007251514543630893, + "loss": 1.8387, + "theoretical_loss": 4.116865305057625, + "tokens_seen": 338690048 + }, + { + "epoch": 0.1, + "learning_rate": 0.0007251354062186561, + "loss": 1.7595, + "theoretical_loss": 4.116774321739942, + "tokens_seen": 338755584 + }, + { + "epoch": 0.1, + "learning_rate": 0.0007251193580742227, + "loss": 1.626, + "theoretical_loss": 4.116683360949678, + "tokens_seen": 338821120 + }, + { + "epoch": 0.1, + "learning_rate": 0.0007251033099297895, + "loss": 1.6181, + "theoretical_loss": 4.1165924226769, + "tokens_seen": 338886656 + }, + { + "epoch": 0.1, + "learning_rate": 0.0007250872617853561, + "loss": 1.7381, + "theoretical_loss": 4.11650150691168, + "tokens_seen": 338952192 + }, + { + "epoch": 0.1, + "learning_rate": 0.0007250712136409229, + "loss": 1.6873, + "theoretical_loss": 4.116410613644098, + "tokens_seen": 339017728 + }, + { + "epoch": 0.1, + "learning_rate": 0.0007250551654964895, + "loss": 1.6373, + "theoretical_loss": 4.116319742864239, + "tokens_seen": 339083264 + }, + { + "epoch": 0.1, + "learning_rate": 0.0007250391173520563, + "loss": 1.6824, + "theoretical_loss": 4.1162288945621945, + "tokens_seen": 339148800 + }, + { + "epoch": 0.1, + "learning_rate": 0.0007250230692076229, + "loss": 1.7403, + "theoretical_loss": 4.116138068728063, + "tokens_seen": 339214336 + }, + { + "epoch": 0.1, + "learning_rate": 0.0007250070210631897, + "loss": 1.5908, + "theoretical_loss": 4.1160472653519475, + "tokens_seen": 339279872 + }, + { + "epoch": 0.1, + "learning_rate": 0.0007249909729187563, + "loss": 1.489, + "theoretical_loss": 4.11595648442396, + "tokens_seen": 339345408 + }, + { + "epoch": 0.1, + "learning_rate": 0.0007249749247743231, + "loss": 1.5505, + "theoretical_loss": 4.115865725934217, + "tokens_seen": 339410944 + }, + { + "epoch": 0.1, + "learning_rate": 0.0007249588766298897, + "loss": 1.684, + "theoretical_loss": 4.11577498987284, + "tokens_seen": 339476480 + }, + { + "epoch": 0.1, + "learning_rate": 0.0007249428284854565, + "loss": 1.7066, + "theoretical_loss": 4.11568427622996, + "tokens_seen": 339542016 + }, + { + "epoch": 0.1, + "learning_rate": 0.0007249267803410231, + "loss": 1.6045, + "theoretical_loss": 4.115593584995712, + "tokens_seen": 339607552 + }, + { + "epoch": 0.1, + "learning_rate": 0.0007249107321965899, + "loss": 1.5944, + "theoretical_loss": 4.115502916160237, + "tokens_seen": 339673088 + }, + { + "epoch": 0.1, + "learning_rate": 0.0007248946840521565, + "loss": 1.6373, + "theoretical_loss": 4.115412269713684, + "tokens_seen": 339738624 + }, + { + "epoch": 0.1, + "learning_rate": 0.0007248786359077233, + "loss": 1.8044, + "theoretical_loss": 4.115321645646207, + "tokens_seen": 339804160 + }, + { + "epoch": 0.1, + "learning_rate": 0.0007248625877632899, + "loss": 1.6099, + "theoretical_loss": 4.115231043947968, + "tokens_seen": 339869696 + }, + { + "epoch": 0.1, + "learning_rate": 0.0007248465396188567, + "loss": 1.5928, + "theoretical_loss": 4.11514046460913, + "tokens_seen": 339935232 + }, + { + "epoch": 0.1, + "learning_rate": 0.0007248304914744233, + "loss": 1.6564, + "theoretical_loss": 4.115049907619868, + "tokens_seen": 340000768 + }, + { + "epoch": 0.1, + "learning_rate": 0.0007248144433299901, + "loss": 1.7186, + "theoretical_loss": 4.114959372970362, + "tokens_seen": 340066304 + }, + { + "epoch": 0.1, + "learning_rate": 0.0007247983951855567, + "loss": 1.6124, + "theoretical_loss": 4.114868860650796, + "tokens_seen": 340131840 + }, + { + "epoch": 0.1, + "learning_rate": 0.0007247823470411235, + "loss": 1.7763, + "theoretical_loss": 4.114778370651361, + "tokens_seen": 340197376 + }, + { + "epoch": 0.1, + "learning_rate": 0.0007247662988966901, + "loss": 1.6565, + "theoretical_loss": 4.114687902962258, + "tokens_seen": 340262912 + }, + { + "epoch": 0.1, + "learning_rate": 0.0007247502507522569, + "loss": 1.667, + "theoretical_loss": 4.114597457573687, + "tokens_seen": 340328448 + }, + { + "epoch": 0.1, + "learning_rate": 0.0007247342026078235, + "loss": 1.7012, + "theoretical_loss": 4.11450703447586, + "tokens_seen": 340393984 + }, + { + "epoch": 0.1, + "learning_rate": 0.0007247181544633902, + "loss": 1.6489, + "theoretical_loss": 4.114416633658992, + "tokens_seen": 340459520 + }, + { + "epoch": 0.1, + "learning_rate": 0.0007247021063189569, + "loss": 1.7053, + "theoretical_loss": 4.114326255113308, + "tokens_seen": 340525056 + }, + { + "epoch": 0.1, + "learning_rate": 0.0007246860581745236, + "loss": 1.5221, + "theoretical_loss": 4.114235898829034, + "tokens_seen": 340590592 + }, + { + "epoch": 0.1, + "learning_rate": 0.0007246700100300903, + "loss": 1.7274, + "theoretical_loss": 4.114145564796406, + "tokens_seen": 340656128 + }, + { + "epoch": 0.1, + "learning_rate": 0.000724653961885657, + "loss": 1.7111, + "theoretical_loss": 4.114055253005665, + "tokens_seen": 340721664 + }, + { + "epoch": 0.1, + "objective/train/docs_used": 130755, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.7197160720825195, + "objective/train/theoretical_loss": 4.113964963447058, + "objective/train/tokens_used": 361247200, + "theoretical_loss": 4.113964963447058, + "tokens_seen": 340787200 + }, + { + "epoch": 0.1, + "learning_rate": 0.0007246379137412237, + "loss": 1.6222, + "theoretical_loss": 4.113964963447058, + "tokens_seen": 340787200 + }, + { + "epoch": 0.1, + "learning_rate": 0.0007246218655967904, + "loss": 1.6192, + "theoretical_loss": 4.113874696110838, + "tokens_seen": 340852736 + }, + { + "epoch": 0.1, + "learning_rate": 0.0007246058174523571, + "loss": 1.7459, + "theoretical_loss": 4.113784450987264, + "tokens_seen": 340918272 + }, + { + "epoch": 0.1, + "learning_rate": 0.0007245897693079238, + "loss": 1.5106, + "theoretical_loss": 4.113694228066604, + "tokens_seen": 340983808 + }, + { + "epoch": 0.1, + "learning_rate": 0.0007245737211634904, + "loss": 1.7371, + "theoretical_loss": 4.113604027339126, + "tokens_seen": 341049344 + }, + { + "epoch": 0.1, + "learning_rate": 0.0007245576730190572, + "loss": 1.6156, + "theoretical_loss": 4.11351384879511, + "tokens_seen": 341114880 + }, + { + "epoch": 0.1, + "learning_rate": 0.0007245416248746238, + "loss": 1.5727, + "theoretical_loss": 4.11342369242484, + "tokens_seen": 341180416 + }, + { + "epoch": 0.1, + "learning_rate": 0.0007245255767301906, + "loss": 1.7787, + "theoretical_loss": 4.1133335582186055, + "tokens_seen": 341245952 + }, + { + "epoch": 0.1, + "learning_rate": 0.0007245095285857572, + "loss": 1.6032, + "theoretical_loss": 4.1132434461667025, + "tokens_seen": 341311488 + }, + { + "epoch": 0.1, + "learning_rate": 0.000724493480441324, + "loss": 1.7881, + "theoretical_loss": 4.113153356259433, + "tokens_seen": 341377024 + }, + { + "epoch": 0.1, + "learning_rate": 0.0007244774322968906, + "loss": 1.7137, + "theoretical_loss": 4.1130632884871074, + "tokens_seen": 341442560 + }, + { + "epoch": 0.1, + "learning_rate": 0.0007244613841524574, + "loss": 1.5736, + "theoretical_loss": 4.112973242840038, + "tokens_seen": 341508096 + }, + { + "epoch": 0.1, + "learning_rate": 0.000724445336008024, + "loss": 1.655, + "theoretical_loss": 4.1128832193085465, + "tokens_seen": 341573632 + }, + { + "epoch": 0.1, + "learning_rate": 0.0007244292878635908, + "loss": 1.6739, + "theoretical_loss": 4.11279321788296, + "tokens_seen": 341639168 + }, + { + "epoch": 0.1, + "learning_rate": 0.0007244132397191574, + "loss": 1.6159, + "theoretical_loss": 4.112703238553609, + "tokens_seen": 341704704 + }, + { + "epoch": 0.1, + "learning_rate": 0.0007243971915747242, + "loss": 1.6707, + "theoretical_loss": 4.112613281310835, + "tokens_seen": 341770240 + }, + { + "epoch": 0.1, + "learning_rate": 0.0007243811434302908, + "loss": 1.5556, + "theoretical_loss": 4.1125233461449815, + "tokens_seen": 341835776 + }, + { + "epoch": 0.1, + "learning_rate": 0.0007243650952858576, + "loss": 1.6903, + "theoretical_loss": 4.1124334330464, + "tokens_seen": 341901312 + }, + { + "epoch": 0.1, + "learning_rate": 0.0007243490471414242, + "loss": 1.7062, + "theoretical_loss": 4.112343542005447, + "tokens_seen": 341966848 + }, + { + "epoch": 0.1, + "learning_rate": 0.000724332998996991, + "loss": 1.4708, + "theoretical_loss": 4.112253673012486, + "tokens_seen": 342032384 + }, + { + "epoch": 0.1, + "learning_rate": 0.0007243169508525577, + "loss": 1.6504, + "theoretical_loss": 4.112163826057886, + "tokens_seen": 342097920 + }, + { + "epoch": 0.1, + "learning_rate": 0.0007243009027081244, + "loss": 1.6184, + "theoretical_loss": 4.112074001132023, + "tokens_seen": 342163456 + }, + { + "epoch": 0.1, + "learning_rate": 0.0007242848545636911, + "loss": 1.6916, + "theoretical_loss": 4.111984198225278, + "tokens_seen": 342228992 + }, + { + "epoch": 0.1, + "learning_rate": 0.0007242688064192578, + "loss": 1.7008, + "theoretical_loss": 4.1118944173280365, + "tokens_seen": 342294528 + }, + { + "epoch": 0.1, + "learning_rate": 0.0007242527582748245, + "loss": 1.6508, + "theoretical_loss": 4.111804658430694, + "tokens_seen": 342360064 + }, + { + "epoch": 0.1, + "learning_rate": 0.0007242367101303912, + "loss": 1.5174, + "theoretical_loss": 4.11171492152365, + "tokens_seen": 342425600 + }, + { + "epoch": 0.1, + "learning_rate": 0.0007242206619859579, + "loss": 1.6476, + "theoretical_loss": 4.111625206597309, + "tokens_seen": 342491136 + }, + { + "epoch": 0.1, + "learning_rate": 0.0007242046138415246, + "loss": 1.7314, + "theoretical_loss": 4.111535513642082, + "tokens_seen": 342556672 + }, + { + "epoch": 0.1, + "learning_rate": 0.0007241885656970913, + "loss": 1.5874, + "theoretical_loss": 4.111445842648387, + "tokens_seen": 342622208 + }, + { + "epoch": 0.1, + "learning_rate": 0.000724172517552658, + "loss": 1.6528, + "theoretical_loss": 4.111356193606648, + "tokens_seen": 342687744 + }, + { + "epoch": 0.1, + "learning_rate": 0.0007241564694082247, + "loss": 1.6563, + "theoretical_loss": 4.111266566507294, + "tokens_seen": 342753280 + }, + { + "epoch": 0.1, + "learning_rate": 0.0007241404212637914, + "loss": 1.5872, + "theoretical_loss": 4.11117696134076, + "tokens_seen": 342818816 + }, + { + "epoch": 0.1, + "learning_rate": 0.0007241243731193581, + "loss": 1.5537, + "theoretical_loss": 4.111087378097488, + "tokens_seen": 342884352 + }, + { + "epoch": 0.1, + "learning_rate": 0.0007241083249749248, + "loss": 1.7304, + "theoretical_loss": 4.1109978167679255, + "tokens_seen": 342949888 + }, + { + "epoch": 0.1, + "learning_rate": 0.0007240922768304915, + "loss": 1.6243, + "theoretical_loss": 4.110908277342526, + "tokens_seen": 343015424 + }, + { + "epoch": 0.1, + "learning_rate": 0.0007240762286860582, + "loss": 1.6869, + "theoretical_loss": 4.110818759811748, + "tokens_seen": 343080960 + }, + { + "epoch": 0.1, + "learning_rate": 0.000724060180541625, + "loss": 1.588, + "theoretical_loss": 4.110729264166058, + "tokens_seen": 343146496 + }, + { + "epoch": 0.1, + "learning_rate": 0.0007240441323971916, + "loss": 1.781, + "theoretical_loss": 4.110639790395926, + "tokens_seen": 343212032 + }, + { + "epoch": 0.1, + "learning_rate": 0.0007240280842527583, + "loss": 1.5875, + "theoretical_loss": 4.110550338491832, + "tokens_seen": 343277568 + }, + { + "epoch": 0.1, + "learning_rate": 0.000724012036108325, + "loss": 1.7854, + "theoretical_loss": 4.1104609084442565, + "tokens_seen": 343343104 + }, + { + "epoch": 0.1, + "learning_rate": 0.0007239959879638917, + "loss": 1.644, + "theoretical_loss": 4.11037150024369, + "tokens_seen": 343408640 + }, + { + "epoch": 0.1, + "learning_rate": 0.0007239799398194584, + "loss": 1.6637, + "theoretical_loss": 4.110282113880627, + "tokens_seen": 343474176 + }, + { + "epoch": 0.1, + "learning_rate": 0.0007239638916750252, + "loss": 1.6211, + "theoretical_loss": 4.11019274934557, + "tokens_seen": 343539712 + }, + { + "epoch": 0.1, + "learning_rate": 0.0007239478435305918, + "loss": 1.6682, + "theoretical_loss": 4.110103406629026, + "tokens_seen": 343605248 + }, + { + "epoch": 0.1, + "learning_rate": 0.0007239317953861586, + "loss": 1.5759, + "theoretical_loss": 4.110014085721506, + "tokens_seen": 343670784 + }, + { + "epoch": 0.1, + "learning_rate": 0.0007239157472417252, + "loss": 1.7124, + "theoretical_loss": 4.109924786613531, + "tokens_seen": 343736320 + }, + { + "epoch": 0.1, + "learning_rate": 0.000723899699097292, + "loss": 1.6343, + "theoretical_loss": 4.109835509295625, + "tokens_seen": 343801856 + }, + { + "epoch": 0.1, + "learning_rate": 0.0007238836509528586, + "loss": 1.6044, + "theoretical_loss": 4.10974625375832, + "tokens_seen": 343867392 + }, + { + "epoch": 0.1, + "learning_rate": 0.0007238676028084254, + "loss": 1.6695, + "theoretical_loss": 4.1096570199921505, + "tokens_seen": 343932928 + }, + { + "epoch": 0.1, + "learning_rate": 0.000723851554663992, + "loss": 1.6381, + "theoretical_loss": 4.109567807987661, + "tokens_seen": 343998464 + }, + { + "epoch": 0.1, + "objective/train/docs_used": 131936, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.7063947916030884, + "objective/train/theoretical_loss": 4.109478617735399, + "objective/train/tokens_used": 364524000, + "theoretical_loss": 4.109478617735399, + "tokens_seen": 344064000 + }, + { + "epoch": 0.1, + "learning_rate": 0.0007238355065195588, + "loss": 1.8138, + "theoretical_loss": 4.109478617735399, + "tokens_seen": 344064000 + }, + { + "epoch": 0.1, + "learning_rate": 0.0007238194583751254, + "loss": 1.5778, + "theoretical_loss": 4.10938944922592, + "tokens_seen": 344129536 + }, + { + "epoch": 0.1, + "learning_rate": 0.0007238034102306922, + "loss": 1.6546, + "theoretical_loss": 4.109300302449784, + "tokens_seen": 344195072 + }, + { + "epoch": 0.1, + "learning_rate": 0.0007237873620862588, + "loss": 1.5391, + "theoretical_loss": 4.109211177397556, + "tokens_seen": 344260608 + }, + { + "epoch": 0.1, + "learning_rate": 0.0007237713139418256, + "loss": 1.5739, + "theoretical_loss": 4.109122074059811, + "tokens_seen": 344326144 + }, + { + "epoch": 0.1, + "learning_rate": 0.0007237552657973922, + "loss": 1.5754, + "theoretical_loss": 4.1090329924271245, + "tokens_seen": 344391680 + }, + { + "epoch": 0.1, + "learning_rate": 0.000723739217652959, + "loss": 1.5847, + "theoretical_loss": 4.108943932490082, + "tokens_seen": 344457216 + }, + { + "epoch": 0.1, + "learning_rate": 0.0007237231695085256, + "loss": 1.7128, + "theoretical_loss": 4.108854894239271, + "tokens_seen": 344522752 + }, + { + "epoch": 0.1, + "learning_rate": 0.0007237071213640924, + "loss": 1.6303, + "theoretical_loss": 4.10876587766529, + "tokens_seen": 344588288 + }, + { + "epoch": 0.1, + "learning_rate": 0.000723691073219659, + "loss": 1.6113, + "theoretical_loss": 4.108676882758738, + "tokens_seen": 344653824 + }, + { + "epoch": 0.1, + "learning_rate": 0.0007236750250752258, + "loss": 1.6593, + "theoretical_loss": 4.1085879095102245, + "tokens_seen": 344719360 + }, + { + "epoch": 0.1, + "learning_rate": 0.0007236589769307924, + "loss": 1.772, + "theoretical_loss": 4.108498957910362, + "tokens_seen": 344784896 + }, + { + "epoch": 0.1, + "learning_rate": 0.0007236429287863592, + "loss": 1.6839, + "theoretical_loss": 4.108410027949769, + "tokens_seen": 344850432 + }, + { + "epoch": 0.1, + "learning_rate": 0.0007236268806419258, + "loss": 1.6658, + "theoretical_loss": 4.10832111961907, + "tokens_seen": 344915968 + }, + { + "epoch": 0.1, + "learning_rate": 0.0007236108324974926, + "loss": 1.6886, + "theoretical_loss": 4.1082322329088985, + "tokens_seen": 344981504 + }, + { + "epoch": 0.1, + "learning_rate": 0.0007235947843530592, + "loss": 1.8071, + "theoretical_loss": 4.108143367809888, + "tokens_seen": 345047040 + }, + { + "epoch": 0.1, + "learning_rate": 0.000723578736208626, + "loss": 1.6042, + "theoretical_loss": 4.108054524312683, + "tokens_seen": 345112576 + }, + { + "epoch": 0.1, + "learning_rate": 0.0007235626880641926, + "loss": 1.6017, + "theoretical_loss": 4.107965702407931, + "tokens_seen": 345178112 + }, + { + "epoch": 0.1, + "learning_rate": 0.0007235466399197594, + "loss": 1.613, + "theoretical_loss": 4.107876902086286, + "tokens_seen": 345243648 + }, + { + "epoch": 0.1, + "learning_rate": 0.000723530591775326, + "loss": 1.655, + "theoretical_loss": 4.107788123338408, + "tokens_seen": 345309184 + }, + { + "epoch": 0.1, + "learning_rate": 0.0007235145436308928, + "loss": 1.6936, + "theoretical_loss": 4.1076993661549634, + "tokens_seen": 345374720 + }, + { + "epoch": 0.1, + "learning_rate": 0.0007234984954864594, + "loss": 1.7146, + "theoretical_loss": 4.107610630526622, + "tokens_seen": 345440256 + }, + { + "epoch": 0.1, + "learning_rate": 0.0007234824473420262, + "loss": 1.8865, + "theoretical_loss": 4.107521916444064, + "tokens_seen": 345505792 + }, + { + "epoch": 0.1, + "learning_rate": 0.0007234663991975928, + "loss": 1.5848, + "theoretical_loss": 4.10743322389797, + "tokens_seen": 345571328 + }, + { + "epoch": 0.1, + "learning_rate": 0.0007234503510531596, + "loss": 1.5706, + "theoretical_loss": 4.107344552879029, + "tokens_seen": 345636864 + }, + { + "epoch": 0.1, + "learning_rate": 0.0007234343029087262, + "loss": 1.6886, + "theoretical_loss": 4.107255903377938, + "tokens_seen": 345702400 + }, + { + "epoch": 0.1, + "learning_rate": 0.000723418254764293, + "loss": 1.7197, + "theoretical_loss": 4.107167275385395, + "tokens_seen": 345767936 + }, + { + "epoch": 0.1, + "learning_rate": 0.0007234022066198596, + "loss": 1.7675, + "theoretical_loss": 4.107078668892108, + "tokens_seen": 345833472 + }, + { + "epoch": 0.1, + "learning_rate": 0.0007233861584754264, + "loss": 1.6309, + "theoretical_loss": 4.106990083888789, + "tokens_seen": 345899008 + }, + { + "epoch": 0.1, + "learning_rate": 0.000723370110330993, + "loss": 1.7005, + "theoretical_loss": 4.106901520366154, + "tokens_seen": 345964544 + }, + { + "epoch": 0.1, + "learning_rate": 0.0007233540621865598, + "loss": 1.6366, + "theoretical_loss": 4.106812978314929, + "tokens_seen": 346030080 + }, + { + "epoch": 0.1, + "learning_rate": 0.0007233380140421264, + "loss": 1.6986, + "theoretical_loss": 4.106724457725842, + "tokens_seen": 346095616 + }, + { + "epoch": 0.1, + "learning_rate": 0.0007233219658976931, + "loss": 1.7318, + "theoretical_loss": 4.106635958589628, + "tokens_seen": 346161152 + }, + { + "epoch": 0.1, + "learning_rate": 0.0007233059177532598, + "loss": 1.6274, + "theoretical_loss": 4.106547480897029, + "tokens_seen": 346226688 + }, + { + "epoch": 0.1, + "learning_rate": 0.0007232898696088265, + "loss": 1.5441, + "theoretical_loss": 4.106459024638791, + "tokens_seen": 346292224 + }, + { + "epoch": 0.1, + "learning_rate": 0.0007232738214643932, + "loss": 1.5893, + "theoretical_loss": 4.106370589805666, + "tokens_seen": 346357760 + }, + { + "epoch": 0.1, + "learning_rate": 0.0007232577733199599, + "loss": 1.7484, + "theoretical_loss": 4.1062821763884125, + "tokens_seen": 346423296 + }, + { + "epoch": 0.1, + "learning_rate": 0.0007232417251755266, + "loss": 1.7867, + "theoretical_loss": 4.106193784377794, + "tokens_seen": 346488832 + }, + { + "epoch": 0.11, + "learning_rate": 0.0007232256770310933, + "loss": 1.6791, + "theoretical_loss": 4.106105413764581, + "tokens_seen": 346554368 + }, + { + "epoch": 0.11, + "learning_rate": 0.00072320962888666, + "loss": 1.559, + "theoretical_loss": 4.106017064539548, + "tokens_seen": 346619904 + }, + { + "epoch": 0.11, + "learning_rate": 0.0007231935807422267, + "loss": 1.7827, + "theoretical_loss": 4.1059287366934765, + "tokens_seen": 346685440 + }, + { + "epoch": 0.11, + "learning_rate": 0.0007231775325977933, + "loss": 1.602, + "theoretical_loss": 4.105840430217153, + "tokens_seen": 346750976 + }, + { + "epoch": 0.11, + "learning_rate": 0.0007231614844533601, + "loss": 1.6888, + "theoretical_loss": 4.10575214510137, + "tokens_seen": 346816512 + }, + { + "epoch": 0.11, + "learning_rate": 0.0007231454363089267, + "loss": 1.5942, + "theoretical_loss": 4.105663881336925, + "tokens_seen": 346882048 + }, + { + "epoch": 0.11, + "learning_rate": 0.0007231293881644935, + "loss": 1.627, + "theoretical_loss": 4.105575638914623, + "tokens_seen": 346947584 + }, + { + "epoch": 0.11, + "learning_rate": 0.0007231133400200602, + "loss": 1.7096, + "theoretical_loss": 4.105487417825273, + "tokens_seen": 347013120 + }, + { + "epoch": 0.11, + "learning_rate": 0.0007230972918756269, + "loss": 1.6719, + "theoretical_loss": 4.10539921805969, + "tokens_seen": 347078656 + }, + { + "epoch": 0.11, + "learning_rate": 0.0007230812437311936, + "loss": 1.5363, + "theoretical_loss": 4.105311039608695, + "tokens_seen": 347144192 + }, + { + "epoch": 0.11, + "learning_rate": 0.0007230651955867603, + "loss": 1.5445, + "theoretical_loss": 4.105222882463115, + "tokens_seen": 347209728 + }, + { + "epoch": 0.11, + "learning_rate": 0.000723049147442327, + "loss": 1.6562, + "theoretical_loss": 4.105134746613783, + "tokens_seen": 347275264 + }, + { + "epoch": 0.11, + "objective/train/docs_used": 133174, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.6303176879882812, + "objective/train/theoretical_loss": 4.105046632051535, + "objective/train/tokens_used": 367800800, + "theoretical_loss": 4.105046632051535, + "tokens_seen": 347340800 + }, + { + "epoch": 0.11, + "learning_rate": 0.0007230330992978937, + "loss": 1.5782, + "theoretical_loss": 4.105046632051535, + "tokens_seen": 347340800 + }, + { + "epoch": 0.11, + "learning_rate": 0.0007230170511534604, + "loss": 1.6261, + "theoretical_loss": 4.1049585387672165, + "tokens_seen": 347406336 + }, + { + "epoch": 0.11, + "learning_rate": 0.0007230010030090271, + "loss": 1.6635, + "theoretical_loss": 4.1048704667516756, + "tokens_seen": 347471872 + }, + { + "epoch": 0.11, + "learning_rate": 0.0007229849548645938, + "loss": 1.642, + "theoretical_loss": 4.104782415995768, + "tokens_seen": 347537408 + }, + { + "epoch": 0.11, + "learning_rate": 0.0007229689067201605, + "loss": 1.6927, + "theoretical_loss": 4.104694386490354, + "tokens_seen": 347602944 + }, + { + "epoch": 0.11, + "learning_rate": 0.0007229528585757272, + "loss": 1.6003, + "theoretical_loss": 4.1046063782263, + "tokens_seen": 347668480 + }, + { + "epoch": 0.11, + "learning_rate": 0.0007229368104312939, + "loss": 1.7073, + "theoretical_loss": 4.104518391194477, + "tokens_seen": 347734016 + }, + { + "epoch": 0.11, + "learning_rate": 0.0007229207622868606, + "loss": 1.5766, + "theoretical_loss": 4.104430425385765, + "tokens_seen": 347799552 + }, + { + "epoch": 0.11, + "learning_rate": 0.0007229047141424273, + "loss": 1.6187, + "theoretical_loss": 4.104342480791044, + "tokens_seen": 347865088 + }, + { + "epoch": 0.11, + "learning_rate": 0.000722888665997994, + "loss": 1.5777, + "theoretical_loss": 4.104254557401205, + "tokens_seen": 347930624 + }, + { + "epoch": 0.11, + "learning_rate": 0.0007228726178535607, + "loss": 1.6162, + "theoretical_loss": 4.104166655207142, + "tokens_seen": 347996160 + }, + { + "epoch": 0.11, + "learning_rate": 0.0007228565697091274, + "loss": 1.6669, + "theoretical_loss": 4.104078774199754, + "tokens_seen": 348061696 + }, + { + "epoch": 0.11, + "learning_rate": 0.0007228405215646941, + "loss": 1.6997, + "theoretical_loss": 4.10399091436995, + "tokens_seen": 348127232 + }, + { + "epoch": 0.11, + "learning_rate": 0.0007228244734202608, + "loss": 1.6456, + "theoretical_loss": 4.103903075708637, + "tokens_seen": 348192768 + }, + { + "epoch": 0.11, + "learning_rate": 0.0007228084252758275, + "loss": 1.5508, + "theoretical_loss": 4.103815258206735, + "tokens_seen": 348258304 + }, + { + "epoch": 0.11, + "learning_rate": 0.0007227923771313942, + "loss": 1.6296, + "theoretical_loss": 4.103727461855165, + "tokens_seen": 348323840 + }, + { + "epoch": 0.11, + "learning_rate": 0.0007227763289869609, + "loss": 1.5141, + "theoretical_loss": 4.103639686644856, + "tokens_seen": 348389376 + }, + { + "epoch": 0.11, + "learning_rate": 0.0007227602808425276, + "loss": 1.7202, + "theoretical_loss": 4.103551932566741, + "tokens_seen": 348454912 + }, + { + "epoch": 0.11, + "learning_rate": 0.0007227442326980943, + "loss": 1.7189, + "theoretical_loss": 4.103464199611759, + "tokens_seen": 348520448 + }, + { + "epoch": 0.11, + "learning_rate": 0.000722728184553661, + "loss": 1.6247, + "theoretical_loss": 4.1033764877708565, + "tokens_seen": 348585984 + }, + { + "epoch": 0.11, + "learning_rate": 0.0007227121364092277, + "loss": 1.8711, + "theoretical_loss": 4.103288797034983, + "tokens_seen": 348651520 + }, + { + "epoch": 0.11, + "learning_rate": 0.0007226960882647944, + "loss": 1.7543, + "theoretical_loss": 4.103201127395095, + "tokens_seen": 348717056 + }, + { + "epoch": 0.11, + "learning_rate": 0.0007226800401203611, + "loss": 1.6477, + "theoretical_loss": 4.103113478842153, + "tokens_seen": 348782592 + }, + { + "epoch": 0.11, + "learning_rate": 0.0007226639919759278, + "loss": 1.5571, + "theoretical_loss": 4.103025851367125, + "tokens_seen": 348848128 + }, + { + "epoch": 0.11, + "learning_rate": 0.0007226479438314945, + "loss": 1.6234, + "theoretical_loss": 4.102938244960985, + "tokens_seen": 348913664 + }, + { + "epoch": 0.11, + "learning_rate": 0.0007226318956870612, + "loss": 1.554, + "theoretical_loss": 4.102850659614709, + "tokens_seen": 348979200 + }, + { + "epoch": 0.11, + "learning_rate": 0.0007226158475426279, + "loss": 1.6322, + "theoretical_loss": 4.102763095319282, + "tokens_seen": 349044736 + }, + { + "epoch": 0.11, + "learning_rate": 0.0007225997993981947, + "loss": 1.6846, + "theoretical_loss": 4.102675552065694, + "tokens_seen": 349110272 + }, + { + "epoch": 0.11, + "learning_rate": 0.0007225837512537613, + "loss": 1.6528, + "theoretical_loss": 4.10258802984494, + "tokens_seen": 349175808 + }, + { + "epoch": 0.11, + "learning_rate": 0.000722567703109328, + "loss": 1.5717, + "theoretical_loss": 4.102500528648019, + "tokens_seen": 349241344 + }, + { + "epoch": 0.11, + "learning_rate": 0.0007225516549648947, + "loss": 1.6957, + "theoretical_loss": 4.102413048465938, + "tokens_seen": 349306880 + }, + { + "epoch": 0.11, + "learning_rate": 0.0007225356068204615, + "loss": 1.5797, + "theoretical_loss": 4.10232558928971, + "tokens_seen": 349372416 + }, + { + "epoch": 0.11, + "learning_rate": 0.0007225195586760281, + "loss": 1.5721, + "theoretical_loss": 4.10223815111035, + "tokens_seen": 349437952 + }, + { + "epoch": 0.11, + "learning_rate": 0.0007225035105315949, + "loss": 1.6423, + "theoretical_loss": 4.102150733918882, + "tokens_seen": 349503488 + }, + { + "epoch": 0.11, + "learning_rate": 0.0007224874623871615, + "loss": 1.8051, + "theoretical_loss": 4.102063337706333, + "tokens_seen": 349569024 + }, + { + "epoch": 0.11, + "learning_rate": 0.0007224714142427283, + "loss": 1.5816, + "theoretical_loss": 4.101975962463738, + "tokens_seen": 349634560 + }, + { + "epoch": 0.11, + "learning_rate": 0.0007224553660982949, + "loss": 1.6508, + "theoretical_loss": 4.101888608182136, + "tokens_seen": 349700096 + }, + { + "epoch": 0.11, + "learning_rate": 0.0007224393179538617, + "loss": 1.6751, + "theoretical_loss": 4.101801274852571, + "tokens_seen": 349765632 + }, + { + "epoch": 0.11, + "learning_rate": 0.0007224232698094283, + "loss": 1.6346, + "theoretical_loss": 4.1017139624660945, + "tokens_seen": 349831168 + }, + { + "epoch": 0.11, + "learning_rate": 0.0007224072216649951, + "loss": 1.4739, + "theoretical_loss": 4.1016266710137606, + "tokens_seen": 349896704 + }, + { + "epoch": 0.11, + "learning_rate": 0.0007223911735205617, + "loss": 1.6231, + "theoretical_loss": 4.101539400486631, + "tokens_seen": 349962240 + }, + { + "epoch": 0.11, + "learning_rate": 0.0007223751253761285, + "loss": 1.5484, + "theoretical_loss": 4.101452150875773, + "tokens_seen": 350027776 + }, + { + "epoch": 0.11, + "learning_rate": 0.0007223590772316951, + "loss": 1.7142, + "theoretical_loss": 4.101364922172259, + "tokens_seen": 350093312 + }, + { + "epoch": 0.11, + "learning_rate": 0.0007223430290872619, + "loss": 1.6179, + "theoretical_loss": 4.101277714367166, + "tokens_seen": 350158848 + }, + { + "epoch": 0.11, + "learning_rate": 0.0007223269809428285, + "loss": 1.5671, + "theoretical_loss": 4.1011905274515765, + "tokens_seen": 350224384 + }, + { + "epoch": 0.11, + "learning_rate": 0.0007223109327983953, + "loss": 1.699, + "theoretical_loss": 4.101103361416581, + "tokens_seen": 350289920 + }, + { + "epoch": 0.11, + "learning_rate": 0.0007222948846539619, + "loss": 1.5336, + "theoretical_loss": 4.101016216253273, + "tokens_seen": 350355456 + }, + { + "epoch": 0.11, + "learning_rate": 0.0007222788365095287, + "loss": 1.6537, + "theoretical_loss": 4.10092909195275, + "tokens_seen": 350420992 + }, + { + "epoch": 0.11, + "learning_rate": 0.0007222627883650953, + "loss": 1.6575, + "theoretical_loss": 4.10084198850612, + "tokens_seen": 350486528 + }, + { + "epoch": 0.11, + "learning_rate": 0.0007222467402206621, + "loss": 1.6608, + "theoretical_loss": 4.100754905904491, + "tokens_seen": 350552064 + }, + { + "epoch": 0.11, + "objective/train/docs_used": 134382, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.6829125881195068, + "objective/train/theoretical_loss": 4.100667844138981, + "objective/train/tokens_used": 371077600, + "theoretical_loss": 4.100667844138981, + "tokens_seen": 350617600 + }, + { + "epoch": 0.11, + "learning_rate": 0.0007222306920762287, + "loss": 1.7003, + "theoretical_loss": 4.100667844138981, + "tokens_seen": 350617600 + }, + { + "epoch": 0.11, + "learning_rate": 0.0007222146439317955, + "loss": 1.6283, + "theoretical_loss": 4.10058080320071, + "tokens_seen": 350683136 + }, + { + "epoch": 0.11, + "learning_rate": 0.0007221985957873621, + "loss": 1.7002, + "theoretical_loss": 4.100493783080806, + "tokens_seen": 350748672 + }, + { + "epoch": 0.11, + "learning_rate": 0.0007221825476429289, + "loss": 1.7905, + "theoretical_loss": 4.1004067837704, + "tokens_seen": 350814208 + }, + { + "epoch": 0.11, + "learning_rate": 0.0007221664994984955, + "loss": 1.6631, + "theoretical_loss": 4.1003198052606304, + "tokens_seen": 350879744 + }, + { + "epoch": 0.11, + "learning_rate": 0.0007221504513540623, + "loss": 1.649, + "theoretical_loss": 4.10023284754264, + "tokens_seen": 350945280 + }, + { + "epoch": 0.11, + "learning_rate": 0.0007221344032096289, + "loss": 1.5811, + "theoretical_loss": 4.100145910607576, + "tokens_seen": 351010816 + }, + { + "epoch": 0.11, + "learning_rate": 0.0007221183550651957, + "loss": 1.5958, + "theoretical_loss": 4.100058994446596, + "tokens_seen": 351076352 + }, + { + "epoch": 0.11, + "learning_rate": 0.0007221023069207623, + "loss": 1.6968, + "theoretical_loss": 4.099972099050856, + "tokens_seen": 351141888 + }, + { + "epoch": 0.11, + "learning_rate": 0.0007220862587763291, + "loss": 1.7475, + "theoretical_loss": 4.099885224411522, + "tokens_seen": 351207424 + }, + { + "epoch": 0.11, + "learning_rate": 0.0007220702106318957, + "loss": 1.5549, + "theoretical_loss": 4.099798370519764, + "tokens_seen": 351272960 + }, + { + "epoch": 0.11, + "learning_rate": 0.0007220541624874625, + "loss": 1.5927, + "theoretical_loss": 4.099711537366758, + "tokens_seen": 351338496 + }, + { + "epoch": 0.11, + "learning_rate": 0.0007220381143430291, + "loss": 1.6744, + "theoretical_loss": 4.099624724943685, + "tokens_seen": 351404032 + }, + { + "epoch": 0.11, + "learning_rate": 0.0007220220661985959, + "loss": 1.6097, + "theoretical_loss": 4.099537933241731, + "tokens_seen": 351469568 + }, + { + "epoch": 0.11, + "learning_rate": 0.0007220060180541625, + "loss": 1.657, + "theoretical_loss": 4.0994511622520875, + "tokens_seen": 351535104 + }, + { + "epoch": 0.11, + "learning_rate": 0.0007219899699097293, + "loss": 1.5677, + "theoretical_loss": 4.099364411965952, + "tokens_seen": 351600640 + }, + { + "epoch": 0.11, + "learning_rate": 0.0007219739217652959, + "loss": 1.7729, + "theoretical_loss": 4.0992776823745265, + "tokens_seen": 351666176 + }, + { + "epoch": 0.11, + "learning_rate": 0.0007219578736208627, + "loss": 1.6116, + "theoretical_loss": 4.09919097346902, + "tokens_seen": 351731712 + }, + { + "epoch": 0.11, + "learning_rate": 0.0007219418254764293, + "loss": 1.6359, + "theoretical_loss": 4.099104285240646, + "tokens_seen": 351797248 + }, + { + "epoch": 0.11, + "learning_rate": 0.000721925777331996, + "loss": 1.7084, + "theoretical_loss": 4.099017617680621, + "tokens_seen": 351862784 + }, + { + "epoch": 0.11, + "learning_rate": 0.0007219097291875627, + "loss": 1.6793, + "theoretical_loss": 4.098930970780172, + "tokens_seen": 351928320 + }, + { + "epoch": 0.11, + "learning_rate": 0.0007218936810431294, + "loss": 1.6407, + "theoretical_loss": 4.098844344530526, + "tokens_seen": 351993856 + }, + { + "epoch": 0.11, + "learning_rate": 0.0007218776328986961, + "loss": 1.7045, + "theoretical_loss": 4.098757738922918, + "tokens_seen": 352059392 + }, + { + "epoch": 0.11, + "learning_rate": 0.0007218615847542628, + "loss": 1.6659, + "theoretical_loss": 4.09867115394859, + "tokens_seen": 352124928 + }, + { + "epoch": 0.11, + "learning_rate": 0.0007218455366098295, + "loss": 1.6179, + "theoretical_loss": 4.098584589598786, + "tokens_seen": 352190464 + }, + { + "epoch": 0.11, + "learning_rate": 0.0007218294884653962, + "loss": 1.5673, + "theoretical_loss": 4.098498045864756, + "tokens_seen": 352256000 + }, + { + "epoch": 0.11, + "learning_rate": 0.0007218134403209629, + "loss": 1.6584, + "theoretical_loss": 4.0984115227377576, + "tokens_seen": 352321536 + }, + { + "epoch": 0.11, + "learning_rate": 0.0007217973921765296, + "loss": 1.6106, + "theoretical_loss": 4.098325020209051, + "tokens_seen": 352387072 + }, + { + "epoch": 0.11, + "learning_rate": 0.0007217813440320963, + "loss": 1.7283, + "theoretical_loss": 4.098238538269904, + "tokens_seen": 352452608 + }, + { + "epoch": 0.11, + "learning_rate": 0.000721765295887663, + "loss": 1.603, + "theoretical_loss": 4.098152076911588, + "tokens_seen": 352518144 + }, + { + "epoch": 0.11, + "learning_rate": 0.0007217492477432297, + "loss": 1.646, + "theoretical_loss": 4.098065636125381, + "tokens_seen": 352583680 + }, + { + "epoch": 0.11, + "learning_rate": 0.0007217331995987964, + "loss": 1.6378, + "theoretical_loss": 4.097979215902566, + "tokens_seen": 352649216 + }, + { + "epoch": 0.11, + "learning_rate": 0.0007217171514543631, + "loss": 1.6731, + "theoretical_loss": 4.097892816234429, + "tokens_seen": 352714752 + }, + { + "epoch": 0.11, + "learning_rate": 0.0007217011033099298, + "loss": 1.5302, + "theoretical_loss": 4.097806437112267, + "tokens_seen": 352780288 + }, + { + "epoch": 0.11, + "learning_rate": 0.0007216850551654965, + "loss": 1.5436, + "theoretical_loss": 4.0977200785273755, + "tokens_seen": 352845824 + }, + { + "epoch": 0.11, + "learning_rate": 0.0007216690070210632, + "loss": 1.5347, + "theoretical_loss": 4.09763374047106, + "tokens_seen": 352911360 + }, + { + "epoch": 0.11, + "learning_rate": 0.0007216529588766299, + "loss": 1.6805, + "theoretical_loss": 4.097547422934629, + "tokens_seen": 352976896 + }, + { + "epoch": 0.11, + "learning_rate": 0.0007216369107321966, + "loss": 1.6137, + "theoretical_loss": 4.097461125909398, + "tokens_seen": 353042432 + }, + { + "epoch": 0.11, + "learning_rate": 0.0007216208625877633, + "loss": 1.7002, + "theoretical_loss": 4.097374849386685, + "tokens_seen": 353107968 + }, + { + "epoch": 0.11, + "learning_rate": 0.00072160481444333, + "loss": 1.7075, + "theoretical_loss": 4.097288593357817, + "tokens_seen": 353173504 + }, + { + "epoch": 0.11, + "learning_rate": 0.0007215887662988967, + "loss": 1.4355, + "theoretical_loss": 4.097202357814124, + "tokens_seen": 353239040 + }, + { + "epoch": 0.11, + "learning_rate": 0.0007215727181544634, + "loss": 1.7483, + "theoretical_loss": 4.097116142746941, + "tokens_seen": 353304576 + }, + { + "epoch": 0.11, + "learning_rate": 0.0007215566700100301, + "loss": 1.6443, + "theoretical_loss": 4.09702994814761, + "tokens_seen": 353370112 + }, + { + "epoch": 0.11, + "learning_rate": 0.0007215406218655968, + "loss": 1.6849, + "theoretical_loss": 4.096943774007476, + "tokens_seen": 353435648 + }, + { + "epoch": 0.11, + "learning_rate": 0.0007215245737211635, + "loss": 1.7195, + "theoretical_loss": 4.096857620317892, + "tokens_seen": 353501184 + }, + { + "epoch": 0.11, + "learning_rate": 0.0007215085255767302, + "loss": 1.6228, + "theoretical_loss": 4.096771487070214, + "tokens_seen": 353566720 + }, + { + "epoch": 0.11, + "learning_rate": 0.0007214924774322969, + "loss": 1.6191, + "theoretical_loss": 4.096685374255803, + "tokens_seen": 353632256 + }, + { + "epoch": 0.11, + "learning_rate": 0.0007214764292878636, + "loss": 1.6328, + "theoretical_loss": 4.096599281866028, + "tokens_seen": 353697792 + }, + { + "epoch": 0.11, + "learning_rate": 0.0007214603811434303, + "loss": 1.7128, + "theoretical_loss": 4.09651320989226, + "tokens_seen": 353763328 + }, + { + "epoch": 0.11, + "learning_rate": 0.000721444332998997, + "loss": 1.5985, + "theoretical_loss": 4.096427158325877, + "tokens_seen": 353828864 + }, + { + "epoch": 0.11, + "objective/train/docs_used": 135614, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.4315263032913208, + "objective/train/theoretical_loss": 4.096341127158263, + "objective/train/tokens_used": 374354400, + "theoretical_loss": 4.096341127158263, + "tokens_seen": 353894400 + }, + { + "epoch": 0.11, + "learning_rate": 0.0007214282848545637, + "loss": 1.5846, + "theoretical_loss": 4.096341127158263, + "tokens_seen": 353894400 + }, + { + "epoch": 0.11, + "learning_rate": 0.0007214122367101304, + "loss": 1.6768, + "theoretical_loss": 4.096255116380806, + "tokens_seen": 353959936 + }, + { + "epoch": 0.11, + "learning_rate": 0.0007213961885656971, + "loss": 1.668, + "theoretical_loss": 4.096169125984899, + "tokens_seen": 354025472 + }, + { + "epoch": 0.11, + "learning_rate": 0.0007213801404212638, + "loss": 1.6471, + "theoretical_loss": 4.0960831559619395, + "tokens_seen": 354091008 + }, + { + "epoch": 0.11, + "learning_rate": 0.0007213640922768305, + "loss": 1.6674, + "theoretical_loss": 4.095997206303333, + "tokens_seen": 354156544 + }, + { + "epoch": 0.11, + "learning_rate": 0.0007213480441323972, + "loss": 1.6317, + "theoretical_loss": 4.095911277000488, + "tokens_seen": 354222080 + }, + { + "epoch": 0.11, + "learning_rate": 0.000721331995987964, + "loss": 1.631, + "theoretical_loss": 4.095825368044819, + "tokens_seen": 354287616 + }, + { + "epoch": 0.11, + "learning_rate": 0.0007213159478435306, + "loss": 1.6847, + "theoretical_loss": 4.0957394794277455, + "tokens_seen": 354353152 + }, + { + "epoch": 0.11, + "learning_rate": 0.0007212998996990973, + "loss": 1.6089, + "theoretical_loss": 4.095653611140692, + "tokens_seen": 354418688 + }, + { + "epoch": 0.11, + "learning_rate": 0.000721283851554664, + "loss": 1.6921, + "theoretical_loss": 4.09556776317509, + "tokens_seen": 354484224 + }, + { + "epoch": 0.11, + "learning_rate": 0.0007212678034102307, + "loss": 1.6862, + "theoretical_loss": 4.095481935522372, + "tokens_seen": 354549760 + }, + { + "epoch": 0.11, + "learning_rate": 0.0007212517552657974, + "loss": 1.6246, + "theoretical_loss": 4.09539612817398, + "tokens_seen": 354615296 + }, + { + "epoch": 0.11, + "learning_rate": 0.0007212357071213642, + "loss": 1.6656, + "theoretical_loss": 4.095310341121358, + "tokens_seen": 354680832 + }, + { + "epoch": 0.11, + "learning_rate": 0.0007212196589769308, + "loss": 1.6927, + "theoretical_loss": 4.095224574355958, + "tokens_seen": 354746368 + }, + { + "epoch": 0.11, + "learning_rate": 0.0007212036108324976, + "loss": 1.714, + "theoretical_loss": 4.095138827869237, + "tokens_seen": 354811904 + }, + { + "epoch": 0.11, + "learning_rate": 0.0007211875626880642, + "loss": 1.6496, + "theoretical_loss": 4.0950531016526535, + "tokens_seen": 354877440 + }, + { + "epoch": 0.11, + "learning_rate": 0.000721171514543631, + "loss": 1.6202, + "theoretical_loss": 4.094967395697675, + "tokens_seen": 354942976 + }, + { + "epoch": 0.11, + "learning_rate": 0.0007211554663991976, + "loss": 1.627, + "theoretical_loss": 4.094881709995773, + "tokens_seen": 355008512 + }, + { + "epoch": 0.11, + "learning_rate": 0.0007211394182547644, + "loss": 1.7245, + "theoretical_loss": 4.094796044538423, + "tokens_seen": 355074048 + }, + { + "epoch": 0.11, + "learning_rate": 0.000721123370110331, + "loss": 1.5969, + "theoretical_loss": 4.094710399317109, + "tokens_seen": 355139584 + }, + { + "epoch": 0.11, + "learning_rate": 0.0007211073219658978, + "loss": 1.6556, + "theoretical_loss": 4.094624774323315, + "tokens_seen": 355205120 + }, + { + "epoch": 0.11, + "learning_rate": 0.0007210912738214644, + "loss": 1.6528, + "theoretical_loss": 4.094539169548535, + "tokens_seen": 355270656 + }, + { + "epoch": 0.11, + "learning_rate": 0.0007210752256770312, + "loss": 1.5882, + "theoretical_loss": 4.094453584984265, + "tokens_seen": 355336192 + }, + { + "epoch": 0.11, + "learning_rate": 0.0007210591775325978, + "loss": 1.619, + "theoretical_loss": 4.094368020622008, + "tokens_seen": 355401728 + }, + { + "epoch": 0.11, + "learning_rate": 0.0007210431293881646, + "loss": 1.722, + "theoretical_loss": 4.094282476453272, + "tokens_seen": 355467264 + }, + { + "epoch": 0.11, + "learning_rate": 0.0007210270812437312, + "loss": 1.6763, + "theoretical_loss": 4.094196952469568, + "tokens_seen": 355532800 + }, + { + "epoch": 0.11, + "learning_rate": 0.000721011033099298, + "loss": 1.7052, + "theoretical_loss": 4.0941114486624155, + "tokens_seen": 355598336 + }, + { + "epoch": 0.11, + "learning_rate": 0.0007209949849548646, + "loss": 1.6454, + "theoretical_loss": 4.094025965023335, + "tokens_seen": 355663872 + }, + { + "epoch": 0.11, + "learning_rate": 0.0007209789368104314, + "loss": 1.5388, + "theoretical_loss": 4.093940501543857, + "tokens_seen": 355729408 + }, + { + "epoch": 0.11, + "learning_rate": 0.000720962888665998, + "loss": 1.5562, + "theoretical_loss": 4.093855058215513, + "tokens_seen": 355794944 + }, + { + "epoch": 0.11, + "learning_rate": 0.0007209468405215648, + "loss": 1.7075, + "theoretical_loss": 4.093769635029841, + "tokens_seen": 355860480 + }, + { + "epoch": 0.11, + "learning_rate": 0.0007209307923771314, + "loss": 1.6596, + "theoretical_loss": 4.0936842319783855, + "tokens_seen": 355926016 + }, + { + "epoch": 0.11, + "learning_rate": 0.0007209147442326982, + "loss": 1.5722, + "theoretical_loss": 4.093598849052694, + "tokens_seen": 355991552 + }, + { + "epoch": 0.11, + "learning_rate": 0.0007208986960882648, + "loss": 1.7093, + "theoretical_loss": 4.093513486244321, + "tokens_seen": 356057088 + }, + { + "epoch": 0.11, + "learning_rate": 0.0007208826479438316, + "loss": 1.5611, + "theoretical_loss": 4.093428143544823, + "tokens_seen": 356122624 + }, + { + "epoch": 0.11, + "learning_rate": 0.0007208665997993982, + "loss": 1.6824, + "theoretical_loss": 4.093342820945766, + "tokens_seen": 356188160 + }, + { + "epoch": 0.11, + "learning_rate": 0.000720850551654965, + "loss": 1.5672, + "theoretical_loss": 4.093257518438718, + "tokens_seen": 356253696 + }, + { + "epoch": 0.11, + "learning_rate": 0.0007208345035105316, + "loss": 1.6248, + "theoretical_loss": 4.093172236015253, + "tokens_seen": 356319232 + }, + { + "epoch": 0.11, + "learning_rate": 0.0007208184553660984, + "loss": 1.6189, + "theoretical_loss": 4.093086973666949, + "tokens_seen": 356384768 + }, + { + "epoch": 0.11, + "learning_rate": 0.000720802407221665, + "loss": 1.6247, + "theoretical_loss": 4.0930017313853915, + "tokens_seen": 356450304 + }, + { + "epoch": 0.11, + "learning_rate": 0.0007207863590772318, + "loss": 1.5719, + "theoretical_loss": 4.092916509162169, + "tokens_seen": 356515840 + }, + { + "epoch": 0.11, + "learning_rate": 0.0007207703109327984, + "loss": 1.6877, + "theoretical_loss": 4.092831306988875, + "tokens_seen": 356581376 + }, + { + "epoch": 0.11, + "learning_rate": 0.0007207542627883652, + "loss": 1.632, + "theoretical_loss": 4.0927461248571095, + "tokens_seen": 356646912 + }, + { + "epoch": 0.11, + "learning_rate": 0.0007207382146439318, + "loss": 1.6197, + "theoretical_loss": 4.092660962758477, + "tokens_seen": 356712448 + }, + { + "epoch": 0.11, + "learning_rate": 0.0007207221664994986, + "loss": 1.6275, + "theoretical_loss": 4.092575820684586, + "tokens_seen": 356777984 + }, + { + "epoch": 0.11, + "learning_rate": 0.0007207061183550653, + "loss": 1.5261, + "theoretical_loss": 4.092490698627052, + "tokens_seen": 356843520 + }, + { + "epoch": 0.11, + "learning_rate": 0.000720690070210632, + "loss": 1.617, + "theoretical_loss": 4.0924055965774935, + "tokens_seen": 356909056 + }, + { + "epoch": 0.11, + "learning_rate": 0.0007206740220661987, + "loss": 1.6815, + "theoretical_loss": 4.092320514527535, + "tokens_seen": 356974592 + }, + { + "epoch": 0.11, + "learning_rate": 0.0007206579739217654, + "loss": 1.6001, + "theoretical_loss": 4.092235452468808, + "tokens_seen": 357040128 + }, + { + "epoch": 0.11, + "learning_rate": 0.000720641925777332, + "loss": 1.6142, + "theoretical_loss": 4.092150410392945, + "tokens_seen": 357105664 + }, + { + "epoch": 0.11, + "objective/train/docs_used": 136331, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.7136940956115723, + "objective/train/theoretical_loss": 4.092065388291585, + "objective/train/tokens_used": 377631200, + "theoretical_loss": 4.092065388291585, + "tokens_seen": 357171200 + }, + { + "epoch": 0.11, + "learning_rate": 0.0007206258776328988, + "loss": 1.7172, + "theoretical_loss": 4.092065388291585, + "tokens_seen": 357171200 + }, + { + "epoch": 0.11, + "learning_rate": 0.0007206098294884654, + "loss": 1.7204, + "theoretical_loss": 4.091980386156376, + "tokens_seen": 357236736 + }, + { + "epoch": 0.11, + "learning_rate": 0.0007205937813440322, + "loss": 1.7092, + "theoretical_loss": 4.091895403978965, + "tokens_seen": 357302272 + }, + { + "epoch": 0.11, + "learning_rate": 0.0007205777331995988, + "loss": 1.6327, + "theoretical_loss": 4.091810441751007, + "tokens_seen": 357367808 + }, + { + "epoch": 0.11, + "learning_rate": 0.0007205616850551656, + "loss": 1.7407, + "theoretical_loss": 4.091725499464163, + "tokens_seen": 357433344 + }, + { + "epoch": 0.11, + "learning_rate": 0.0007205456369107322, + "loss": 1.6873, + "theoretical_loss": 4.091640577110096, + "tokens_seen": 357498880 + }, + { + "epoch": 0.11, + "learning_rate": 0.0007205295887662989, + "loss": 1.5837, + "theoretical_loss": 4.091555674680476, + "tokens_seen": 357564416 + }, + { + "epoch": 0.11, + "learning_rate": 0.0007205135406218656, + "loss": 1.5793, + "theoretical_loss": 4.09147079216698, + "tokens_seen": 357629952 + }, + { + "epoch": 0.11, + "learning_rate": 0.0007204974924774323, + "loss": 1.4291, + "theoretical_loss": 4.0913859295612856, + "tokens_seen": 357695488 + }, + { + "epoch": 0.11, + "learning_rate": 0.000720481444332999, + "loss": 1.6182, + "theoretical_loss": 4.091301086855078, + "tokens_seen": 357761024 + }, + { + "epoch": 0.11, + "learning_rate": 0.0007204653961885657, + "loss": 1.6324, + "theoretical_loss": 4.091216264040048, + "tokens_seen": 357826560 + }, + { + "epoch": 0.11, + "learning_rate": 0.0007204493480441324, + "loss": 1.4679, + "theoretical_loss": 4.091131461107889, + "tokens_seen": 357892096 + }, + { + "epoch": 0.11, + "learning_rate": 0.0007204332998996991, + "loss": 1.5735, + "theoretical_loss": 4.091046678050301, + "tokens_seen": 357957632 + }, + { + "epoch": 0.11, + "learning_rate": 0.0007204172517552658, + "loss": 1.6034, + "theoretical_loss": 4.090961914858989, + "tokens_seen": 358023168 + }, + { + "epoch": 0.11, + "learning_rate": 0.0007204012036108325, + "loss": 1.4949, + "theoretical_loss": 4.090877171525663, + "tokens_seen": 358088704 + }, + { + "epoch": 0.11, + "learning_rate": 0.0007203851554663992, + "loss": 1.7373, + "theoretical_loss": 4.090792448042038, + "tokens_seen": 358154240 + }, + { + "epoch": 0.11, + "learning_rate": 0.0007203691073219659, + "loss": 1.5962, + "theoretical_loss": 4.0907077443998325, + "tokens_seen": 358219776 + }, + { + "epoch": 0.11, + "learning_rate": 0.0007203530591775326, + "loss": 1.7266, + "theoretical_loss": 4.090623060590772, + "tokens_seen": 358285312 + }, + { + "epoch": 0.11, + "learning_rate": 0.0007203370110330993, + "loss": 1.6591, + "theoretical_loss": 4.090538396606585, + "tokens_seen": 358350848 + }, + { + "epoch": 0.11, + "learning_rate": 0.000720320962888666, + "loss": 1.6002, + "theoretical_loss": 4.090453752439008, + "tokens_seen": 358416384 + }, + { + "epoch": 0.11, + "learning_rate": 0.0007203049147442327, + "loss": 1.6469, + "theoretical_loss": 4.09036912807978, + "tokens_seen": 358481920 + }, + { + "epoch": 0.11, + "learning_rate": 0.0007202888665997994, + "loss": 1.6138, + "theoretical_loss": 4.090284523520644, + "tokens_seen": 358547456 + }, + { + "epoch": 0.11, + "learning_rate": 0.0007202728184553661, + "loss": 1.491, + "theoretical_loss": 4.09019993875335, + "tokens_seen": 358612992 + }, + { + "epoch": 0.11, + "learning_rate": 0.0007202567703109328, + "loss": 1.6577, + "theoretical_loss": 4.090115373769653, + "tokens_seen": 358678528 + }, + { + "epoch": 0.11, + "learning_rate": 0.0007202407221664995, + "loss": 1.5577, + "theoretical_loss": 4.0900308285613125, + "tokens_seen": 358744064 + }, + { + "epoch": 0.11, + "learning_rate": 0.0007202246740220662, + "loss": 1.7173, + "theoretical_loss": 4.089946303120092, + "tokens_seen": 358809600 + }, + { + "epoch": 0.11, + "learning_rate": 0.0007202086258776329, + "loss": 1.5857, + "theoretical_loss": 4.089861797437761, + "tokens_seen": 358875136 + }, + { + "epoch": 0.11, + "learning_rate": 0.0007201925777331996, + "loss": 1.6586, + "theoretical_loss": 4.089777311506094, + "tokens_seen": 358940672 + }, + { + "epoch": 0.11, + "learning_rate": 0.0007201765295887663, + "loss": 1.6686, + "theoretical_loss": 4.089692845316868, + "tokens_seen": 359006208 + }, + { + "epoch": 0.11, + "learning_rate": 0.000720160481444333, + "loss": 1.5467, + "theoretical_loss": 4.08960839886187, + "tokens_seen": 359071744 + }, + { + "epoch": 0.11, + "learning_rate": 0.0007201444332998997, + "loss": 1.6093, + "theoretical_loss": 4.089523972132886, + "tokens_seen": 359137280 + }, + { + "epoch": 0.11, + "learning_rate": 0.0007201283851554664, + "loss": 1.6981, + "theoretical_loss": 4.089439565121712, + "tokens_seen": 359202816 + }, + { + "epoch": 0.11, + "learning_rate": 0.0007201123370110331, + "loss": 1.7117, + "theoretical_loss": 4.089355177820146, + "tokens_seen": 359268352 + }, + { + "epoch": 0.11, + "learning_rate": 0.0007200962888665998, + "loss": 1.5819, + "theoretical_loss": 4.0892708102199915, + "tokens_seen": 359333888 + }, + { + "epoch": 0.11, + "learning_rate": 0.0007200802407221665, + "loss": 1.6786, + "theoretical_loss": 4.089186462313056, + "tokens_seen": 359399424 + }, + { + "epoch": 0.11, + "learning_rate": 0.0007200641925777332, + "loss": 1.5782, + "theoretical_loss": 4.089102134091155, + "tokens_seen": 359464960 + }, + { + "epoch": 0.11, + "learning_rate": 0.0007200481444332999, + "loss": 1.6402, + "theoretical_loss": 4.0890178255461045, + "tokens_seen": 359530496 + }, + { + "epoch": 0.11, + "learning_rate": 0.0007200320962888666, + "loss": 1.6548, + "theoretical_loss": 4.08893353666973, + "tokens_seen": 359596032 + }, + { + "epoch": 0.11, + "learning_rate": 0.0007200160481444333, + "loss": 1.5826, + "theoretical_loss": 4.088849267453858, + "tokens_seen": 359661568 + }, + { + "epoch": 0.11, + "learning_rate": 0.00072, + "loss": 1.5694, + "theoretical_loss": 4.088765017890322, + "tokens_seen": 359727104 + }, + { + "epoch": 0.11, + "learning_rate": 0.0007199839518555667, + "loss": 1.5843, + "theoretical_loss": 4.08868078797096, + "tokens_seen": 359792640 + }, + { + "epoch": 0.11, + "learning_rate": 0.0007199679037111334, + "loss": 1.5454, + "theoretical_loss": 4.088596577687614, + "tokens_seen": 359858176 + }, + { + "epoch": 0.11, + "learning_rate": 0.0007199518555667001, + "loss": 1.7449, + "theoretical_loss": 4.088512387032132, + "tokens_seen": 359923712 + }, + { + "epoch": 0.11, + "learning_rate": 0.0007199358074222668, + "loss": 1.5769, + "theoretical_loss": 4.088428215996366, + "tokens_seen": 359989248 + }, + { + "epoch": 0.11, + "learning_rate": 0.0007199197592778335, + "loss": 1.6433, + "theoretical_loss": 4.0883440645721745, + "tokens_seen": 360054784 + }, + { + "epoch": 0.11, + "learning_rate": 0.0007199037111334002, + "loss": 1.6758, + "theoretical_loss": 4.088259932751419, + "tokens_seen": 360120320 + }, + { + "epoch": 0.11, + "learning_rate": 0.0007198876629889669, + "loss": 1.545, + "theoretical_loss": 4.088175820525966, + "tokens_seen": 360185856 + }, + { + "epoch": 0.11, + "learning_rate": 0.0007198716148445337, + "loss": 1.6343, + "theoretical_loss": 4.088091727887688, + "tokens_seen": 360251392 + }, + { + "epoch": 0.11, + "learning_rate": 0.0007198555667001003, + "loss": 1.7332, + "theoretical_loss": 4.088007654828463, + "tokens_seen": 360316928 + }, + { + "epoch": 0.11, + "learning_rate": 0.000719839518555667, + "loss": 1.6791, + "theoretical_loss": 4.087923601340169, + "tokens_seen": 360382464 + }, + { + "epoch": 0.11, + "objective/train/docs_used": 137426, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.7640526294708252, + "objective/train/theoretical_loss": 4.0878395674146955, + "objective/train/tokens_used": 380908000, + "theoretical_loss": 4.0878395674146955, + "tokens_seen": 360448000 + }, + { + "epoch": 0.11, + "learning_rate": 0.0007198234704112337, + "loss": 1.7553, + "theoretical_loss": 4.0878395674146955, + "tokens_seen": 360448000 + }, + { + "epoch": 0.11, + "learning_rate": 0.0007198074222668005, + "loss": 1.6326, + "theoretical_loss": 4.087755553043934, + "tokens_seen": 360513536 + }, + { + "epoch": 0.11, + "learning_rate": 0.0007197913741223671, + "loss": 1.5654, + "theoretical_loss": 4.087671558219778, + "tokens_seen": 360579072 + }, + { + "epoch": 0.11, + "learning_rate": 0.0007197753259779339, + "loss": 1.529, + "theoretical_loss": 4.087587582934131, + "tokens_seen": 360644608 + }, + { + "epoch": 0.11, + "learning_rate": 0.0007197592778335005, + "loss": 1.7114, + "theoretical_loss": 4.087503627178897, + "tokens_seen": 360710144 + }, + { + "epoch": 0.11, + "learning_rate": 0.0007197432296890673, + "loss": 1.7088, + "theoretical_loss": 4.087419690945988, + "tokens_seen": 360775680 + }, + { + "epoch": 0.11, + "learning_rate": 0.0007197271815446339, + "loss": 1.5862, + "theoretical_loss": 4.087335774227318, + "tokens_seen": 360841216 + }, + { + "epoch": 0.11, + "learning_rate": 0.0007197111334002007, + "loss": 1.4933, + "theoretical_loss": 4.0872518770148085, + "tokens_seen": 360906752 + }, + { + "epoch": 0.11, + "learning_rate": 0.0007196950852557673, + "loss": 1.6051, + "theoretical_loss": 4.087167999300384, + "tokens_seen": 360972288 + }, + { + "epoch": 0.11, + "learning_rate": 0.0007196790371113341, + "loss": 1.672, + "theoretical_loss": 4.087084141075974, + "tokens_seen": 361037824 + }, + { + "epoch": 0.11, + "learning_rate": 0.0007196629889669007, + "loss": 1.6922, + "theoretical_loss": 4.087000302333513, + "tokens_seen": 361103360 + }, + { + "epoch": 0.11, + "learning_rate": 0.0007196469408224675, + "loss": 1.7574, + "theoretical_loss": 4.086916483064941, + "tokens_seen": 361168896 + }, + { + "epoch": 0.11, + "learning_rate": 0.0007196308926780341, + "loss": 1.6637, + "theoretical_loss": 4.086832683262203, + "tokens_seen": 361234432 + }, + { + "epoch": 0.11, + "learning_rate": 0.0007196148445336009, + "loss": 1.6472, + "theoretical_loss": 4.086748902917246, + "tokens_seen": 361299968 + }, + { + "epoch": 0.11, + "learning_rate": 0.0007195987963891675, + "loss": 1.6027, + "theoretical_loss": 4.086665142022025, + "tokens_seen": 361365504 + }, + { + "epoch": 0.11, + "learning_rate": 0.0007195827482447343, + "loss": 1.6312, + "theoretical_loss": 4.086581400568499, + "tokens_seen": 361431040 + }, + { + "epoch": 0.11, + "learning_rate": 0.0007195667001003009, + "loss": 1.6123, + "theoretical_loss": 4.086497678548631, + "tokens_seen": 361496576 + }, + { + "epoch": 0.11, + "learning_rate": 0.0007195506519558677, + "loss": 1.7743, + "theoretical_loss": 4.086413975954389, + "tokens_seen": 361562112 + }, + { + "epoch": 0.11, + "learning_rate": 0.0007195346038114343, + "loss": 1.7311, + "theoretical_loss": 4.086330292777745, + "tokens_seen": 361627648 + }, + { + "epoch": 0.11, + "learning_rate": 0.0007195185556670011, + "loss": 1.6226, + "theoretical_loss": 4.086246629010679, + "tokens_seen": 361693184 + }, + { + "epoch": 0.11, + "learning_rate": 0.0007195025075225678, + "loss": 1.6734, + "theoretical_loss": 4.086162984645172, + "tokens_seen": 361758720 + }, + { + "epoch": 0.11, + "learning_rate": 0.0007194864593781345, + "loss": 1.5536, + "theoretical_loss": 4.086079359673212, + "tokens_seen": 361824256 + }, + { + "epoch": 0.11, + "learning_rate": 0.0007194704112337012, + "loss": 1.5532, + "theoretical_loss": 4.085995754086789, + "tokens_seen": 361889792 + }, + { + "epoch": 0.11, + "learning_rate": 0.0007194543630892679, + "loss": 1.4951, + "theoretical_loss": 4.085912167877902, + "tokens_seen": 361955328 + }, + { + "epoch": 0.11, + "learning_rate": 0.0007194383149448346, + "loss": 1.656, + "theoretical_loss": 4.085828601038552, + "tokens_seen": 362020864 + }, + { + "epoch": 0.11, + "learning_rate": 0.0007194222668004013, + "loss": 1.621, + "theoretical_loss": 4.0857450535607445, + "tokens_seen": 362086400 + }, + { + "epoch": 0.11, + "learning_rate": 0.000719406218655968, + "loss": 1.7049, + "theoretical_loss": 4.085661525436492, + "tokens_seen": 362151936 + }, + { + "epoch": 0.11, + "learning_rate": 0.0007193901705115347, + "loss": 1.7409, + "theoretical_loss": 4.0855780166578075, + "tokens_seen": 362217472 + }, + { + "epoch": 0.11, + "learning_rate": 0.0007193741223671014, + "loss": 1.6311, + "theoretical_loss": 4.0854945272167145, + "tokens_seen": 362283008 + }, + { + "epoch": 0.11, + "learning_rate": 0.0007193580742226681, + "loss": 1.5793, + "theoretical_loss": 4.085411057105237, + "tokens_seen": 362348544 + }, + { + "epoch": 0.11, + "learning_rate": 0.0007193420260782348, + "loss": 1.7057, + "theoretical_loss": 4.0853276063154045, + "tokens_seen": 362414080 + }, + { + "epoch": 0.11, + "learning_rate": 0.0007193259779338015, + "loss": 1.6113, + "theoretical_loss": 4.085244174839252, + "tokens_seen": 362479616 + }, + { + "epoch": 0.11, + "learning_rate": 0.0007193099297893682, + "loss": 1.6177, + "theoretical_loss": 4.085160762668819, + "tokens_seen": 362545152 + }, + { + "epoch": 0.11, + "learning_rate": 0.0007192938816449349, + "loss": 1.6461, + "theoretical_loss": 4.085077369796149, + "tokens_seen": 362610688 + }, + { + "epoch": 0.11, + "learning_rate": 0.0007192778335005016, + "loss": 1.6135, + "theoretical_loss": 4.084993996213292, + "tokens_seen": 362676224 + }, + { + "epoch": 0.11, + "learning_rate": 0.0007192617853560683, + "loss": 1.6806, + "theoretical_loss": 4.084910641912301, + "tokens_seen": 362741760 + }, + { + "epoch": 0.11, + "learning_rate": 0.0007192457372116349, + "loss": 1.6797, + "theoretical_loss": 4.084827306885234, + "tokens_seen": 362807296 + }, + { + "epoch": 0.11, + "learning_rate": 0.0007192296890672017, + "loss": 1.7043, + "theoretical_loss": 4.084743991124155, + "tokens_seen": 362872832 + }, + { + "epoch": 0.11, + "learning_rate": 0.0007192136409227683, + "loss": 1.6335, + "theoretical_loss": 4.08466069462113, + "tokens_seen": 362938368 + }, + { + "epoch": 0.11, + "learning_rate": 0.0007191975927783351, + "loss": 1.7987, + "theoretical_loss": 4.084577417368232, + "tokens_seen": 363003904 + }, + { + "epoch": 0.11, + "learning_rate": 0.0007191815446339017, + "loss": 1.6898, + "theoretical_loss": 4.084494159357538, + "tokens_seen": 363069440 + }, + { + "epoch": 0.11, + "learning_rate": 0.0007191654964894685, + "loss": 1.7117, + "theoretical_loss": 4.0844109205811305, + "tokens_seen": 363134976 + }, + { + "epoch": 0.11, + "learning_rate": 0.0007191494483450351, + "loss": 1.6925, + "theoretical_loss": 4.084327701031095, + "tokens_seen": 363200512 + }, + { + "epoch": 0.11, + "learning_rate": 0.0007191334002006019, + "loss": 1.6176, + "theoretical_loss": 4.084244500699524, + "tokens_seen": 363266048 + }, + { + "epoch": 0.11, + "learning_rate": 0.0007191173520561685, + "loss": 1.6554, + "theoretical_loss": 4.084161319578511, + "tokens_seen": 363331584 + }, + { + "epoch": 0.11, + "learning_rate": 0.0007191013039117352, + "loss": 1.6793, + "theoretical_loss": 4.084078157660159, + "tokens_seen": 363397120 + }, + { + "epoch": 0.11, + "learning_rate": 0.0007190852557673019, + "loss": 1.5648, + "theoretical_loss": 4.083995014936571, + "tokens_seen": 363462656 + }, + { + "epoch": 0.11, + "learning_rate": 0.0007190692076228686, + "loss": 1.6968, + "theoretical_loss": 4.0839118913998576, + "tokens_seen": 363528192 + }, + { + "epoch": 0.11, + "learning_rate": 0.0007190531594784353, + "loss": 1.5974, + "theoretical_loss": 4.083828787042134, + "tokens_seen": 363593728 + }, + { + "epoch": 0.11, + "learning_rate": 0.000719037111334002, + "loss": 1.6967, + "theoretical_loss": 4.083745701855518, + "tokens_seen": 363659264 + }, + { + "epoch": 0.11, + "objective/train/docs_used": 138686, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.6887487173080444, + "objective/train/theoretical_loss": 4.083662635832135, + "objective/train/tokens_used": 384184800, + "theoretical_loss": 4.083662635832135, + "tokens_seen": 363724800 + }, + { + "epoch": 0.11, + "learning_rate": 0.0007190210631895687, + "loss": 1.6796, + "theoretical_loss": 4.083662635832135, + "tokens_seen": 363724800 + }, + { + "epoch": 0.11, + "learning_rate": 0.0007190050150451354, + "loss": 1.7068, + "theoretical_loss": 4.0835795889641116, + "tokens_seen": 363790336 + }, + { + "epoch": 0.11, + "learning_rate": 0.0007189889669007021, + "loss": 1.6067, + "theoretical_loss": 4.0834965612435825, + "tokens_seen": 363855872 + }, + { + "epoch": 0.11, + "learning_rate": 0.0007189729187562688, + "loss": 1.7266, + "theoretical_loss": 4.083413552662685, + "tokens_seen": 363921408 + }, + { + "epoch": 0.11, + "learning_rate": 0.0007189568706118355, + "loss": 1.6785, + "theoretical_loss": 4.08333056321356, + "tokens_seen": 363986944 + }, + { + "epoch": 0.11, + "learning_rate": 0.0007189408224674022, + "loss": 1.6006, + "theoretical_loss": 4.083247592888357, + "tokens_seen": 364052480 + }, + { + "epoch": 0.11, + "learning_rate": 0.0007189247743229689, + "loss": 1.6694, + "theoretical_loss": 4.083164641679225, + "tokens_seen": 364118016 + }, + { + "epoch": 0.11, + "learning_rate": 0.0007189087261785356, + "loss": 1.6617, + "theoretical_loss": 4.083081709578323, + "tokens_seen": 364183552 + }, + { + "epoch": 0.11, + "learning_rate": 0.0007188926780341023, + "loss": 1.7299, + "theoretical_loss": 4.082998796577811, + "tokens_seen": 364249088 + }, + { + "epoch": 0.11, + "learning_rate": 0.000718876629889669, + "loss": 1.5965, + "theoretical_loss": 4.082915902669853, + "tokens_seen": 364314624 + }, + { + "epoch": 0.11, + "learning_rate": 0.0007188605817452357, + "loss": 1.6998, + "theoretical_loss": 4.082833027846622, + "tokens_seen": 364380160 + }, + { + "epoch": 0.11, + "learning_rate": 0.0007188445336008024, + "loss": 1.7446, + "theoretical_loss": 4.08275017210029, + "tokens_seen": 364445696 + }, + { + "epoch": 0.11, + "learning_rate": 0.0007188284854563691, + "loss": 1.5997, + "theoretical_loss": 4.082667335423038, + "tokens_seen": 364511232 + }, + { + "epoch": 0.11, + "learning_rate": 0.0007188124373119358, + "loss": 1.672, + "theoretical_loss": 4.0825845178070495, + "tokens_seen": 364576768 + }, + { + "epoch": 0.11, + "learning_rate": 0.0007187963891675025, + "loss": 1.5856, + "theoretical_loss": 4.0825017192445126, + "tokens_seen": 364642304 + }, + { + "epoch": 0.11, + "learning_rate": 0.0007187803410230692, + "loss": 1.6231, + "theoretical_loss": 4.082418939727622, + "tokens_seen": 364707840 + }, + { + "epoch": 0.11, + "learning_rate": 0.0007187642928786359, + "loss": 1.6041, + "theoretical_loss": 4.082336179248574, + "tokens_seen": 364773376 + }, + { + "epoch": 0.11, + "learning_rate": 0.0007187482447342026, + "loss": 1.593, + "theoretical_loss": 4.082253437799572, + "tokens_seen": 364838912 + }, + { + "epoch": 0.11, + "learning_rate": 0.0007187321965897693, + "loss": 1.6836, + "theoretical_loss": 4.082170715372823, + "tokens_seen": 364904448 + }, + { + "epoch": 0.11, + "learning_rate": 0.000718716148445336, + "loss": 1.7691, + "theoretical_loss": 4.082088011960538, + "tokens_seen": 364969984 + }, + { + "epoch": 0.11, + "learning_rate": 0.0007187001003009027, + "loss": 1.6904, + "theoretical_loss": 4.082005327554934, + "tokens_seen": 365035520 + }, + { + "epoch": 0.11, + "learning_rate": 0.0007186840521564694, + "loss": 1.5749, + "theoretical_loss": 4.081922662148231, + "tokens_seen": 365101056 + }, + { + "epoch": 0.11, + "learning_rate": 0.0007186680040120361, + "loss": 1.6869, + "theoretical_loss": 4.081840015732654, + "tokens_seen": 365166592 + }, + { + "epoch": 0.11, + "learning_rate": 0.0007186519558676028, + "loss": 1.626, + "theoretical_loss": 4.081757388300435, + "tokens_seen": 365232128 + }, + { + "epoch": 0.11, + "learning_rate": 0.0007186359077231695, + "loss": 1.5756, + "theoretical_loss": 4.081674779843806, + "tokens_seen": 365297664 + }, + { + "epoch": 0.11, + "learning_rate": 0.0007186198595787362, + "loss": 1.6139, + "theoretical_loss": 4.081592190355008, + "tokens_seen": 365363200 + }, + { + "epoch": 0.11, + "learning_rate": 0.000718603811434303, + "loss": 1.5488, + "theoretical_loss": 4.081509619826283, + "tokens_seen": 365428736 + }, + { + "epoch": 0.11, + "learning_rate": 0.0007185877632898696, + "loss": 1.5643, + "theoretical_loss": 4.081427068249881, + "tokens_seen": 365494272 + }, + { + "epoch": 0.11, + "learning_rate": 0.0007185717151454363, + "loss": 1.7096, + "theoretical_loss": 4.081344535618053, + "tokens_seen": 365559808 + }, + { + "epoch": 0.11, + "learning_rate": 0.000718555667001003, + "loss": 1.6129, + "theoretical_loss": 4.081262021923058, + "tokens_seen": 365625344 + }, + { + "epoch": 0.11, + "learning_rate": 0.0007185396188565697, + "loss": 1.6974, + "theoretical_loss": 4.081179527157156, + "tokens_seen": 365690880 + }, + { + "epoch": 0.11, + "learning_rate": 0.0007185235707121364, + "loss": 1.6126, + "theoretical_loss": 4.081097051312614, + "tokens_seen": 365756416 + }, + { + "epoch": 0.11, + "learning_rate": 0.0007185075225677032, + "loss": 1.6011, + "theoretical_loss": 4.081014594381705, + "tokens_seen": 365821952 + }, + { + "epoch": 0.11, + "learning_rate": 0.0007184914744232698, + "loss": 1.6262, + "theoretical_loss": 4.0809321563567025, + "tokens_seen": 365887488 + }, + { + "epoch": 0.11, + "learning_rate": 0.0007184754262788366, + "loss": 1.5517, + "theoretical_loss": 4.080849737229887, + "tokens_seen": 365953024 + }, + { + "epoch": 0.11, + "learning_rate": 0.0007184593781344032, + "loss": 1.6423, + "theoretical_loss": 4.080767336993542, + "tokens_seen": 366018560 + }, + { + "epoch": 0.11, + "learning_rate": 0.00071844332998997, + "loss": 1.6356, + "theoretical_loss": 4.080684955639958, + "tokens_seen": 366084096 + }, + { + "epoch": 0.11, + "learning_rate": 0.0007184272818455366, + "loss": 1.7032, + "theoretical_loss": 4.080602593161428, + "tokens_seen": 366149632 + }, + { + "epoch": 0.11, + "learning_rate": 0.0007184112337011034, + "loss": 1.62, + "theoretical_loss": 4.0805202495502515, + "tokens_seen": 366215168 + }, + { + "epoch": 0.11, + "learning_rate": 0.00071839518555667, + "loss": 1.6346, + "theoretical_loss": 4.080437924798729, + "tokens_seen": 366280704 + }, + { + "epoch": 0.11, + "learning_rate": 0.0007183791374122368, + "loss": 1.553, + "theoretical_loss": 4.080355618899169, + "tokens_seen": 366346240 + }, + { + "epoch": 0.11, + "learning_rate": 0.0007183630892678034, + "loss": 1.6301, + "theoretical_loss": 4.080273331843884, + "tokens_seen": 366411776 + }, + { + "epoch": 0.11, + "learning_rate": 0.0007183470411233702, + "loss": 1.7083, + "theoretical_loss": 4.080191063625188, + "tokens_seen": 366477312 + }, + { + "epoch": 0.11, + "learning_rate": 0.0007183309929789368, + "loss": 1.5751, + "theoretical_loss": 4.080108814235404, + "tokens_seen": 366542848 + }, + { + "epoch": 0.11, + "learning_rate": 0.0007183149448345036, + "loss": 1.619, + "theoretical_loss": 4.080026583666855, + "tokens_seen": 366608384 + }, + { + "epoch": 0.11, + "learning_rate": 0.0007182988966900703, + "loss": 1.6272, + "theoretical_loss": 4.079944371911871, + "tokens_seen": 366673920 + }, + { + "epoch": 0.11, + "learning_rate": 0.000718282848545637, + "loss": 1.7291, + "theoretical_loss": 4.07986217896279, + "tokens_seen": 366739456 + }, + { + "epoch": 0.11, + "learning_rate": 0.0007182668004012037, + "loss": 1.7519, + "theoretical_loss": 4.079780004811946, + "tokens_seen": 366804992 + }, + { + "epoch": 0.11, + "learning_rate": 0.0007182507522567704, + "loss": 1.6299, + "theoretical_loss": 4.079697849451684, + "tokens_seen": 366870528 + }, + { + "epoch": 0.11, + "learning_rate": 0.0007182347041123371, + "loss": 1.8789, + "theoretical_loss": 4.079615712874352, + "tokens_seen": 366936064 + }, + { + "epoch": 0.11, + "objective/train/docs_used": 139849, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.7035019397735596, + "objective/train/theoretical_loss": 4.079533595072301, + "objective/train/tokens_used": 387461600, + "theoretical_loss": 4.079533595072301, + "tokens_seen": 367001600 + }, + { + "epoch": 0.11, + "learning_rate": 0.0007182186559679038, + "loss": 1.6774, + "theoretical_loss": 4.079533595072301, + "tokens_seen": 367001600 + }, + { + "epoch": 0.11, + "learning_rate": 0.0007182026078234705, + "loss": 1.7, + "theoretical_loss": 4.0794514960378905, + "tokens_seen": 367067136 + }, + { + "epoch": 0.11, + "learning_rate": 0.0007181865596790372, + "loss": 1.6855, + "theoretical_loss": 4.079369415763479, + "tokens_seen": 367132672 + }, + { + "epoch": 0.11, + "learning_rate": 0.0007181705115346039, + "loss": 1.6246, + "theoretical_loss": 4.079287354241432, + "tokens_seen": 367198208 + }, + { + "epoch": 0.11, + "learning_rate": 0.0007181544633901706, + "loss": 1.6673, + "theoretical_loss": 4.07920531146412, + "tokens_seen": 367263744 + }, + { + "epoch": 0.11, + "learning_rate": 0.0007181384152457373, + "loss": 1.6545, + "theoretical_loss": 4.0791232874239185, + "tokens_seen": 367329280 + }, + { + "epoch": 0.11, + "learning_rate": 0.000718122367101304, + "loss": 1.6264, + "theoretical_loss": 4.079041282113206, + "tokens_seen": 367394816 + }, + { + "epoch": 0.11, + "learning_rate": 0.0007181063189568707, + "loss": 1.7621, + "theoretical_loss": 4.078959295524365, + "tokens_seen": 367460352 + }, + { + "epoch": 0.11, + "learning_rate": 0.0007180902708124374, + "loss": 1.6633, + "theoretical_loss": 4.078877327649785, + "tokens_seen": 367525888 + }, + { + "epoch": 0.11, + "learning_rate": 0.0007180742226680041, + "loss": 1.547, + "theoretical_loss": 4.078795378481858, + "tokens_seen": 367591424 + }, + { + "epoch": 0.11, + "learning_rate": 0.0007180581745235708, + "loss": 1.5713, + "theoretical_loss": 4.07871344801298, + "tokens_seen": 367656960 + }, + { + "epoch": 0.11, + "learning_rate": 0.0007180421263791375, + "loss": 1.7252, + "theoretical_loss": 4.0786315362355525, + "tokens_seen": 367722496 + }, + { + "epoch": 0.11, + "learning_rate": 0.0007180260782347042, + "loss": 1.6446, + "theoretical_loss": 4.078549643141981, + "tokens_seen": 367788032 + }, + { + "epoch": 0.11, + "learning_rate": 0.0007180100300902709, + "loss": 1.6173, + "theoretical_loss": 4.078467768724677, + "tokens_seen": 367853568 + }, + { + "epoch": 0.11, + "learning_rate": 0.0007179939819458376, + "loss": 1.5791, + "theoretical_loss": 4.078385912976053, + "tokens_seen": 367919104 + }, + { + "epoch": 0.11, + "learning_rate": 0.0007179779338014043, + "loss": 1.6518, + "theoretical_loss": 4.078304075888529, + "tokens_seen": 367984640 + }, + { + "epoch": 0.11, + "learning_rate": 0.000717961885656971, + "loss": 1.6967, + "theoretical_loss": 4.078222257454529, + "tokens_seen": 368050176 + }, + { + "epoch": 0.11, + "learning_rate": 0.0007179458375125377, + "loss": 1.6739, + "theoretical_loss": 4.07814045766648, + "tokens_seen": 368115712 + }, + { + "epoch": 0.11, + "learning_rate": 0.0007179297893681044, + "loss": 1.6511, + "theoretical_loss": 4.0780586765168145, + "tokens_seen": 368181248 + }, + { + "epoch": 0.11, + "learning_rate": 0.0007179137412236711, + "loss": 1.5136, + "theoretical_loss": 4.07797691399797, + "tokens_seen": 368246784 + }, + { + "epoch": 0.11, + "learning_rate": 0.0007178976930792378, + "loss": 1.7167, + "theoretical_loss": 4.077895170102386, + "tokens_seen": 368312320 + }, + { + "epoch": 0.11, + "learning_rate": 0.0007178816449348045, + "loss": 1.6659, + "theoretical_loss": 4.077813444822509, + "tokens_seen": 368377856 + }, + { + "epoch": 0.11, + "learning_rate": 0.0007178655967903712, + "loss": 1.567, + "theoretical_loss": 4.077731738150789, + "tokens_seen": 368443392 + }, + { + "epoch": 0.11, + "learning_rate": 0.0007178495486459378, + "loss": 1.6116, + "theoretical_loss": 4.077650050079681, + "tokens_seen": 368508928 + }, + { + "epoch": 0.11, + "learning_rate": 0.0007178335005015046, + "loss": 1.5717, + "theoretical_loss": 4.077568380601642, + "tokens_seen": 368574464 + }, + { + "epoch": 0.11, + "learning_rate": 0.0007178174523570712, + "loss": 1.5208, + "theoretical_loss": 4.077486729709136, + "tokens_seen": 368640000 + }, + { + "epoch": 0.11, + "learning_rate": 0.000717801404212638, + "loss": 1.5819, + "theoretical_loss": 4.077405097394631, + "tokens_seen": 368705536 + }, + { + "epoch": 0.11, + "learning_rate": 0.0007177853560682046, + "loss": 1.7143, + "theoretical_loss": 4.077323483650598, + "tokens_seen": 368771072 + }, + { + "epoch": 0.11, + "learning_rate": 0.0007177693079237714, + "loss": 1.7162, + "theoretical_loss": 4.077241888469514, + "tokens_seen": 368836608 + }, + { + "epoch": 0.11, + "learning_rate": 0.000717753259779338, + "loss": 1.6392, + "theoretical_loss": 4.077160311843858, + "tokens_seen": 368902144 + }, + { + "epoch": 0.11, + "learning_rate": 0.0007177372116349048, + "loss": 1.6276, + "theoretical_loss": 4.077078753766118, + "tokens_seen": 368967680 + }, + { + "epoch": 0.11, + "learning_rate": 0.0007177211634904714, + "loss": 1.8095, + "theoretical_loss": 4.076997214228782, + "tokens_seen": 369033216 + }, + { + "epoch": 0.11, + "learning_rate": 0.0007177051153460381, + "loss": 1.6341, + "theoretical_loss": 4.076915693224343, + "tokens_seen": 369098752 + }, + { + "epoch": 0.11, + "learning_rate": 0.0007176890672016048, + "loss": 1.7247, + "theoretical_loss": 4.0768341907453, + "tokens_seen": 369164288 + }, + { + "epoch": 0.11, + "learning_rate": 0.0007176730190571715, + "loss": 1.6215, + "theoretical_loss": 4.0767527067841565, + "tokens_seen": 369229824 + }, + { + "epoch": 0.11, + "learning_rate": 0.0007176569709127382, + "loss": 1.5753, + "theoretical_loss": 4.076671241333417, + "tokens_seen": 369295360 + }, + { + "epoch": 0.11, + "learning_rate": 0.0007176409227683049, + "loss": 1.6145, + "theoretical_loss": 4.076589794385596, + "tokens_seen": 369360896 + }, + { + "epoch": 0.11, + "learning_rate": 0.0007176248746238716, + "loss": 1.576, + "theoretical_loss": 4.076508365933206, + "tokens_seen": 369426432 + }, + { + "epoch": 0.11, + "learning_rate": 0.0007176088264794383, + "loss": 1.6873, + "theoretical_loss": 4.076426955968769, + "tokens_seen": 369491968 + }, + { + "epoch": 0.11, + "learning_rate": 0.000717592778335005, + "loss": 1.6563, + "theoretical_loss": 4.076345564484807, + "tokens_seen": 369557504 + }, + { + "epoch": 0.11, + "learning_rate": 0.0007175767301905717, + "loss": 1.5893, + "theoretical_loss": 4.076264191473852, + "tokens_seen": 369623040 + }, + { + "epoch": 0.11, + "learning_rate": 0.0007175606820461384, + "loss": 1.7037, + "theoretical_loss": 4.076182836928435, + "tokens_seen": 369688576 + }, + { + "epoch": 0.11, + "learning_rate": 0.0007175446339017051, + "loss": 1.7322, + "theoretical_loss": 4.0761015008410935, + "tokens_seen": 369754112 + }, + { + "epoch": 0.11, + "learning_rate": 0.0007175285857572718, + "loss": 1.6815, + "theoretical_loss": 4.07602018320437, + "tokens_seen": 369819648 + }, + { + "epoch": 0.11, + "learning_rate": 0.0007175125376128385, + "loss": 1.5677, + "theoretical_loss": 4.075938884010809, + "tokens_seen": 369885184 + }, + { + "epoch": 0.11, + "learning_rate": 0.0007174964894684052, + "loss": 1.6024, + "theoretical_loss": 4.075857603252963, + "tokens_seen": 369950720 + }, + { + "epoch": 0.11, + "learning_rate": 0.0007174804413239719, + "loss": 1.5981, + "theoretical_loss": 4.075776340923385, + "tokens_seen": 370016256 + }, + { + "epoch": 0.11, + "learning_rate": 0.0007174643931795386, + "loss": 1.6778, + "theoretical_loss": 4.075695097014635, + "tokens_seen": 370081792 + }, + { + "epoch": 0.11, + "learning_rate": 0.0007174483450351053, + "loss": 1.6452, + "theoretical_loss": 4.075613871519276, + "tokens_seen": 370147328 + }, + { + "epoch": 0.11, + "learning_rate": 0.000717432296890672, + "loss": 1.6936, + "theoretical_loss": 4.0755326644298755, + "tokens_seen": 370212864 + }, + { + "epoch": 0.11, + "objective/train/docs_used": 141078, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.61774480342865, + "objective/train/theoretical_loss": 4.075451475739007, + "objective/train/tokens_used": 390738400, + "theoretical_loss": 4.075451475739007, + "tokens_seen": 370278400 + }, + { + "epoch": 0.11, + "learning_rate": 0.0007174162487462387, + "loss": 1.5681, + "theoretical_loss": 4.075451475739007, + "tokens_seen": 370278400 + }, + { + "epoch": 0.11, + "learning_rate": 0.0007174002006018054, + "loss": 1.6328, + "theoretical_loss": 4.075370305439243, + "tokens_seen": 370343936 + }, + { + "epoch": 0.11, + "learning_rate": 0.0007173841524573721, + "loss": 1.5791, + "theoretical_loss": 4.075289153523168, + "tokens_seen": 370409472 + }, + { + "epoch": 0.11, + "learning_rate": 0.0007173681043129388, + "loss": 1.6115, + "theoretical_loss": 4.075208019983366, + "tokens_seen": 370475008 + }, + { + "epoch": 0.11, + "learning_rate": 0.0007173520561685055, + "loss": 1.7275, + "theoretical_loss": 4.075126904812425, + "tokens_seen": 370540544 + }, + { + "epoch": 0.11, + "learning_rate": 0.0007173360080240722, + "loss": 1.5706, + "theoretical_loss": 4.0750458080029395, + "tokens_seen": 370606080 + }, + { + "epoch": 0.11, + "learning_rate": 0.0007173199598796389, + "loss": 1.6614, + "theoretical_loss": 4.074964729547506, + "tokens_seen": 370671616 + }, + { + "epoch": 0.11, + "learning_rate": 0.0007173039117352056, + "loss": 1.5533, + "theoretical_loss": 4.074883669438729, + "tokens_seen": 370737152 + }, + { + "epoch": 0.11, + "learning_rate": 0.0007172878635907723, + "loss": 1.6472, + "theoretical_loss": 4.074802627669213, + "tokens_seen": 370802688 + }, + { + "epoch": 0.11, + "learning_rate": 0.000717271815446339, + "loss": 1.6376, + "theoretical_loss": 4.07472160423157, + "tokens_seen": 370868224 + }, + { + "epoch": 0.11, + "learning_rate": 0.0007172557673019057, + "loss": 1.6108, + "theoretical_loss": 4.074640599118413, + "tokens_seen": 370933760 + }, + { + "epoch": 0.11, + "learning_rate": 0.0007172397191574724, + "loss": 1.6149, + "theoretical_loss": 4.074559612322363, + "tokens_seen": 370999296 + }, + { + "epoch": 0.11, + "learning_rate": 0.0007172236710130391, + "loss": 1.6804, + "theoretical_loss": 4.074478643836043, + "tokens_seen": 371064832 + }, + { + "epoch": 0.11, + "learning_rate": 0.0007172076228686058, + "loss": 1.6601, + "theoretical_loss": 4.074397693652081, + "tokens_seen": 371130368 + }, + { + "epoch": 0.11, + "learning_rate": 0.0007171915747241725, + "loss": 1.664, + "theoretical_loss": 4.074316761763107, + "tokens_seen": 371195904 + }, + { + "epoch": 0.11, + "learning_rate": 0.0007171755265797392, + "loss": 1.6995, + "theoretical_loss": 4.074235848161761, + "tokens_seen": 371261440 + }, + { + "epoch": 0.11, + "learning_rate": 0.0007171594784353059, + "loss": 1.673, + "theoretical_loss": 4.074154952840681, + "tokens_seen": 371326976 + }, + { + "epoch": 0.11, + "learning_rate": 0.0007171434302908726, + "loss": 1.6223, + "theoretical_loss": 4.074074075792513, + "tokens_seen": 371392512 + }, + { + "epoch": 0.11, + "learning_rate": 0.0007171273821464393, + "loss": 1.7006, + "theoretical_loss": 4.073993217009905, + "tokens_seen": 371458048 + }, + { + "epoch": 0.11, + "learning_rate": 0.000717111334002006, + "loss": 1.5466, + "theoretical_loss": 4.073912376485511, + "tokens_seen": 371523584 + }, + { + "epoch": 0.11, + "learning_rate": 0.0007170952858575728, + "loss": 1.614, + "theoretical_loss": 4.073831554211989, + "tokens_seen": 371589120 + }, + { + "epoch": 0.11, + "learning_rate": 0.0007170792377131395, + "loss": 1.5874, + "theoretical_loss": 4.0737507501820005, + "tokens_seen": 371654656 + }, + { + "epoch": 0.11, + "learning_rate": 0.0007170631895687062, + "loss": 1.5996, + "theoretical_loss": 4.073669964388212, + "tokens_seen": 371720192 + }, + { + "epoch": 0.11, + "learning_rate": 0.0007170471414242729, + "loss": 1.779, + "theoretical_loss": 4.073589196823293, + "tokens_seen": 371785728 + }, + { + "epoch": 0.11, + "learning_rate": 0.0007170310932798396, + "loss": 1.7305, + "theoretical_loss": 4.073508447479918, + "tokens_seen": 371851264 + }, + { + "epoch": 0.11, + "learning_rate": 0.0007170150451354063, + "loss": 1.6807, + "theoretical_loss": 4.073427716350767, + "tokens_seen": 371916800 + }, + { + "epoch": 0.11, + "learning_rate": 0.000716998996990973, + "loss": 1.8016, + "theoretical_loss": 4.0733470034285215, + "tokens_seen": 371982336 + }, + { + "epoch": 0.11, + "learning_rate": 0.0007169829488465397, + "loss": 1.5808, + "theoretical_loss": 4.07326630870587, + "tokens_seen": 372047872 + }, + { + "epoch": 0.11, + "learning_rate": 0.0007169669007021064, + "loss": 1.5971, + "theoretical_loss": 4.073185632175504, + "tokens_seen": 372113408 + }, + { + "epoch": 0.11, + "learning_rate": 0.0007169508525576731, + "loss": 1.6919, + "theoretical_loss": 4.073104973830118, + "tokens_seen": 372178944 + }, + { + "epoch": 0.11, + "learning_rate": 0.0007169348044132398, + "loss": 1.7379, + "theoretical_loss": 4.073024333662413, + "tokens_seen": 372244480 + }, + { + "epoch": 0.11, + "learning_rate": 0.0007169187562688065, + "loss": 1.662, + "theoretical_loss": 4.072943711665093, + "tokens_seen": 372310016 + }, + { + "epoch": 0.11, + "learning_rate": 0.0007169027081243732, + "loss": 1.4853, + "theoretical_loss": 4.072863107830866, + "tokens_seen": 372375552 + }, + { + "epoch": 0.11, + "learning_rate": 0.0007168866599799399, + "loss": 1.6556, + "theoretical_loss": 4.072782522152444, + "tokens_seen": 372441088 + }, + { + "epoch": 0.11, + "learning_rate": 0.0007168706118355066, + "loss": 1.6255, + "theoretical_loss": 4.072701954622546, + "tokens_seen": 372506624 + }, + { + "epoch": 0.11, + "learning_rate": 0.0007168545636910733, + "loss": 1.724, + "theoretical_loss": 4.072621405233891, + "tokens_seen": 372572160 + }, + { + "epoch": 0.11, + "learning_rate": 0.00071683851554664, + "loss": 1.6916, + "theoretical_loss": 4.072540873979205, + "tokens_seen": 372637696 + }, + { + "epoch": 0.11, + "learning_rate": 0.0007168224674022067, + "loss": 1.6074, + "theoretical_loss": 4.072460360851217, + "tokens_seen": 372703232 + }, + { + "epoch": 0.11, + "learning_rate": 0.0007168064192577734, + "loss": 1.6332, + "theoretical_loss": 4.07237986584266, + "tokens_seen": 372768768 + }, + { + "epoch": 0.11, + "learning_rate": 0.0007167903711133401, + "loss": 1.593, + "theoretical_loss": 4.0722993889462735, + "tokens_seen": 372834304 + }, + { + "epoch": 0.11, + "learning_rate": 0.0007167743229689068, + "loss": 1.6235, + "theoretical_loss": 4.072218930154797, + "tokens_seen": 372899840 + }, + { + "epoch": 0.11, + "learning_rate": 0.0007167582748244735, + "loss": 1.521, + "theoretical_loss": 4.072138489460979, + "tokens_seen": 372965376 + }, + { + "epoch": 0.11, + "learning_rate": 0.0007167422266800402, + "loss": 1.6076, + "theoretical_loss": 4.0720580668575685, + "tokens_seen": 373030912 + }, + { + "epoch": 0.11, + "learning_rate": 0.0007167261785356069, + "loss": 1.7252, + "theoretical_loss": 4.0719776623373205, + "tokens_seen": 373096448 + }, + { + "epoch": 0.11, + "learning_rate": 0.0007167101303911736, + "loss": 1.6166, + "theoretical_loss": 4.071897275892994, + "tokens_seen": 373161984 + }, + { + "epoch": 0.11, + "learning_rate": 0.0007166940822467403, + "loss": 1.5868, + "theoretical_loss": 4.071816907517349, + "tokens_seen": 373227520 + }, + { + "epoch": 0.11, + "learning_rate": 0.000716678034102307, + "loss": 1.5824, + "theoretical_loss": 4.071736557203157, + "tokens_seen": 373293056 + }, + { + "epoch": 0.11, + "learning_rate": 0.0007166619859578737, + "loss": 1.6028, + "theoretical_loss": 4.071656224943186, + "tokens_seen": 373358592 + }, + { + "epoch": 0.11, + "learning_rate": 0.0007166459378134404, + "loss": 1.6091, + "theoretical_loss": 4.071575910730212, + "tokens_seen": 373424128 + }, + { + "epoch": 0.11, + "learning_rate": 0.0007166298896690071, + "loss": 1.4758, + "theoretical_loss": 4.071495614557016, + "tokens_seen": 373489664 + }, + { + "epoch": 0.11, + "objective/train/docs_used": 142333, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.7061116695404053, + "objective/train/theoretical_loss": 4.071415336416378, + "objective/train/tokens_used": 394015200, + "theoretical_loss": 4.071415336416378, + "tokens_seen": 373555200 + }, + { + "epoch": 0.11, + "learning_rate": 0.0007166138415245738, + "loss": 1.7063, + "theoretical_loss": 4.071415336416378, + "tokens_seen": 373555200 + }, + { + "epoch": 0.11, + "learning_rate": 0.0007165977933801405, + "loss": 1.5161, + "theoretical_loss": 4.071335076301089, + "tokens_seen": 373620736 + }, + { + "epoch": 0.11, + "learning_rate": 0.0007165817452357072, + "loss": 1.6901, + "theoretical_loss": 4.071254834203941, + "tokens_seen": 373686272 + }, + { + "epoch": 0.11, + "learning_rate": 0.0007165656970912739, + "loss": 1.729, + "theoretical_loss": 4.071174610117729, + "tokens_seen": 373751808 + }, + { + "epoch": 0.11, + "learning_rate": 0.0007165496489468406, + "loss": 1.6412, + "theoretical_loss": 4.071094404035253, + "tokens_seen": 373817344 + }, + { + "epoch": 0.11, + "learning_rate": 0.0007165336008024073, + "loss": 1.6348, + "theoretical_loss": 4.071014215949317, + "tokens_seen": 373882880 + }, + { + "epoch": 0.11, + "learning_rate": 0.000716517552657974, + "loss": 1.4928, + "theoretical_loss": 4.070934045852731, + "tokens_seen": 373948416 + }, + { + "epoch": 0.11, + "learning_rate": 0.0007165015045135407, + "loss": 1.6525, + "theoretical_loss": 4.070853893738307, + "tokens_seen": 374013952 + }, + { + "epoch": 0.11, + "learning_rate": 0.0007164854563691074, + "loss": 1.6862, + "theoretical_loss": 4.070773759598861, + "tokens_seen": 374079488 + }, + { + "epoch": 0.11, + "learning_rate": 0.0007164694082246741, + "loss": 1.5912, + "theoretical_loss": 4.070693643427215, + "tokens_seen": 374145024 + }, + { + "epoch": 0.11, + "learning_rate": 0.0007164533600802407, + "loss": 1.5596, + "theoretical_loss": 4.070613545216193, + "tokens_seen": 374210560 + }, + { + "epoch": 0.11, + "learning_rate": 0.0007164373119358075, + "loss": 1.6298, + "theoretical_loss": 4.070533464958626, + "tokens_seen": 374276096 + }, + { + "epoch": 0.11, + "learning_rate": 0.0007164212637913741, + "loss": 1.5965, + "theoretical_loss": 4.070453402647344, + "tokens_seen": 374341632 + }, + { + "epoch": 0.11, + "learning_rate": 0.0007164052156469409, + "loss": 1.6993, + "theoretical_loss": 4.070373358275187, + "tokens_seen": 374407168 + }, + { + "epoch": 0.11, + "learning_rate": 0.0007163891675025075, + "loss": 1.5687, + "theoretical_loss": 4.070293331834996, + "tokens_seen": 374472704 + }, + { + "epoch": 0.11, + "learning_rate": 0.0007163731193580743, + "loss": 1.667, + "theoretical_loss": 4.070213323319616, + "tokens_seen": 374538240 + }, + { + "epoch": 0.11, + "learning_rate": 0.0007163570712136409, + "loss": 1.799, + "theoretical_loss": 4.070133332721897, + "tokens_seen": 374603776 + }, + { + "epoch": 0.11, + "learning_rate": 0.0007163410230692077, + "loss": 1.66, + "theoretical_loss": 4.070053360034693, + "tokens_seen": 374669312 + }, + { + "epoch": 0.11, + "learning_rate": 0.0007163249749247743, + "loss": 1.6204, + "theoretical_loss": 4.069973405250862, + "tokens_seen": 374734848 + }, + { + "epoch": 0.11, + "learning_rate": 0.000716308926780341, + "loss": 1.6565, + "theoretical_loss": 4.069893468363265, + "tokens_seen": 374800384 + }, + { + "epoch": 0.11, + "learning_rate": 0.0007162928786359077, + "loss": 1.6183, + "theoretical_loss": 4.06981354936477, + "tokens_seen": 374865920 + }, + { + "epoch": 0.11, + "learning_rate": 0.0007162768304914744, + "loss": 1.6118, + "theoretical_loss": 4.069733648248246, + "tokens_seen": 374931456 + }, + { + "epoch": 0.11, + "learning_rate": 0.0007162607823470411, + "loss": 1.6054, + "theoretical_loss": 4.069653765006565, + "tokens_seen": 374996992 + }, + { + "epoch": 0.11, + "learning_rate": 0.0007162447342026078, + "loss": 1.5844, + "theoretical_loss": 4.06957389963261, + "tokens_seen": 375062528 + }, + { + "epoch": 0.11, + "learning_rate": 0.0007162286860581745, + "loss": 1.657, + "theoretical_loss": 4.069494052119261, + "tokens_seen": 375128064 + }, + { + "epoch": 0.11, + "learning_rate": 0.0007162126379137412, + "loss": 1.6559, + "theoretical_loss": 4.069414222459403, + "tokens_seen": 375193600 + }, + { + "epoch": 0.11, + "learning_rate": 0.0007161965897693079, + "loss": 1.7173, + "theoretical_loss": 4.069334410645928, + "tokens_seen": 375259136 + }, + { + "epoch": 0.11, + "learning_rate": 0.0007161805416248746, + "loss": 1.5977, + "theoretical_loss": 4.069254616671732, + "tokens_seen": 375324672 + }, + { + "epoch": 0.11, + "learning_rate": 0.0007161644934804413, + "loss": 1.7127, + "theoretical_loss": 4.069174840529712, + "tokens_seen": 375390208 + }, + { + "epoch": 0.11, + "learning_rate": 0.000716148445336008, + "loss": 1.738, + "theoretical_loss": 4.069095082212772, + "tokens_seen": 375455744 + }, + { + "epoch": 0.11, + "learning_rate": 0.0007161323971915747, + "loss": 1.6496, + "theoretical_loss": 4.0690153417138175, + "tokens_seen": 375521280 + }, + { + "epoch": 0.11, + "learning_rate": 0.0007161163490471414, + "loss": 1.7847, + "theoretical_loss": 4.06893561902576, + "tokens_seen": 375586816 + }, + { + "epoch": 0.11, + "learning_rate": 0.0007161003009027081, + "loss": 1.7078, + "theoretical_loss": 4.068855914141515, + "tokens_seen": 375652352 + }, + { + "epoch": 0.11, + "learning_rate": 0.0007160842527582748, + "loss": 1.7301, + "theoretical_loss": 4.068776227054, + "tokens_seen": 375717888 + }, + { + "epoch": 0.11, + "learning_rate": 0.0007160682046138415, + "loss": 1.5806, + "theoretical_loss": 4.06869655775614, + "tokens_seen": 375783424 + }, + { + "epoch": 0.11, + "learning_rate": 0.0007160521564694082, + "loss": 1.6466, + "theoretical_loss": 4.068616906240861, + "tokens_seen": 375848960 + }, + { + "epoch": 0.11, + "learning_rate": 0.0007160361083249749, + "loss": 1.6326, + "theoretical_loss": 4.068537272501095, + "tokens_seen": 375914496 + }, + { + "epoch": 0.11, + "learning_rate": 0.0007160200601805416, + "loss": 1.5778, + "theoretical_loss": 4.068457656529775, + "tokens_seen": 375980032 + }, + { + "epoch": 0.11, + "learning_rate": 0.0007160040120361083, + "loss": 1.7756, + "theoretical_loss": 4.068378058319843, + "tokens_seen": 376045568 + }, + { + "epoch": 0.11, + "learning_rate": 0.000715987963891675, + "loss": 1.7846, + "theoretical_loss": 4.068298477864241, + "tokens_seen": 376111104 + }, + { + "epoch": 0.11, + "learning_rate": 0.0007159719157472417, + "loss": 1.5876, + "theoretical_loss": 4.068218915155917, + "tokens_seen": 376176640 + }, + { + "epoch": 0.11, + "learning_rate": 0.0007159558676028084, + "loss": 1.6276, + "theoretical_loss": 4.068139370187821, + "tokens_seen": 376242176 + }, + { + "epoch": 0.11, + "learning_rate": 0.0007159398194583751, + "loss": 1.6576, + "theoretical_loss": 4.06805984295291, + "tokens_seen": 376307712 + }, + { + "epoch": 0.11, + "learning_rate": 0.0007159237713139418, + "loss": 1.5901, + "theoretical_loss": 4.067980333444142, + "tokens_seen": 376373248 + }, + { + "epoch": 0.11, + "learning_rate": 0.0007159077231695085, + "loss": 1.6792, + "theoretical_loss": 4.067900841654481, + "tokens_seen": 376438784 + }, + { + "epoch": 0.11, + "learning_rate": 0.0007158916750250753, + "loss": 1.707, + "theoretical_loss": 4.067821367576894, + "tokens_seen": 376504320 + }, + { + "epoch": 0.11, + "learning_rate": 0.000715875626880642, + "loss": 1.619, + "theoretical_loss": 4.067741911204354, + "tokens_seen": 376569856 + }, + { + "epoch": 0.11, + "learning_rate": 0.0007158595787362087, + "loss": 1.5852, + "theoretical_loss": 4.067662472529834, + "tokens_seen": 376635392 + }, + { + "epoch": 0.11, + "learning_rate": 0.0007158435305917753, + "loss": 1.8104, + "theoretical_loss": 4.067583051546316, + "tokens_seen": 376700928 + }, + { + "epoch": 0.11, + "learning_rate": 0.0007158274824473421, + "loss": 1.6669, + "theoretical_loss": 4.067503648246781, + "tokens_seen": 376766464 + }, + { + "epoch": 0.11, + "objective/train/docs_used": 143470, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.6984126567840576, + "objective/train/theoretical_loss": 4.067424262624218, + "objective/train/tokens_used": 397292000, + "theoretical_loss": 4.067424262624218, + "tokens_seen": 376832000 + }, + { + "epoch": 0.11, + "learning_rate": 0.0007158114343029087, + "loss": 1.6932, + "theoretical_loss": 4.067424262624218, + "tokens_seen": 376832000 + }, + { + "epoch": 0.11, + "learning_rate": 0.0007157953861584755, + "loss": 1.7424, + "theoretical_loss": 4.06734489467162, + "tokens_seen": 376897536 + }, + { + "epoch": 0.11, + "learning_rate": 0.0007157793380140421, + "loss": 1.6079, + "theoretical_loss": 4.0672655443819785, + "tokens_seen": 376963072 + }, + { + "epoch": 0.11, + "learning_rate": 0.0007157632898696089, + "loss": 1.6034, + "theoretical_loss": 4.067186211748297, + "tokens_seen": 377028608 + }, + { + "epoch": 0.11, + "learning_rate": 0.0007157472417251756, + "loss": 1.6803, + "theoretical_loss": 4.067106896763575, + "tokens_seen": 377094144 + }, + { + "epoch": 0.11, + "learning_rate": 0.0007157311935807423, + "loss": 1.598, + "theoretical_loss": 4.067027599420824, + "tokens_seen": 377159680 + }, + { + "epoch": 0.11, + "learning_rate": 0.000715715145436309, + "loss": 1.5957, + "theoretical_loss": 4.0669483197130525, + "tokens_seen": 377225216 + }, + { + "epoch": 0.11, + "learning_rate": 0.0007156990972918757, + "loss": 1.5787, + "theoretical_loss": 4.066869057633277, + "tokens_seen": 377290752 + }, + { + "epoch": 0.11, + "learning_rate": 0.0007156830491474424, + "loss": 1.5864, + "theoretical_loss": 4.066789813174517, + "tokens_seen": 377356288 + }, + { + "epoch": 0.11, + "learning_rate": 0.0007156670010030091, + "loss": 1.5961, + "theoretical_loss": 4.066710586329796, + "tokens_seen": 377421824 + }, + { + "epoch": 0.11, + "learning_rate": 0.0007156509528585758, + "loss": 1.6396, + "theoretical_loss": 4.066631377092141, + "tokens_seen": 377487360 + }, + { + "epoch": 0.11, + "learning_rate": 0.0007156349047141425, + "loss": 1.6486, + "theoretical_loss": 4.066552185454582, + "tokens_seen": 377552896 + }, + { + "epoch": 0.11, + "learning_rate": 0.0007156188565697092, + "loss": 1.509, + "theoretical_loss": 4.066473011410157, + "tokens_seen": 377618432 + }, + { + "epoch": 0.11, + "learning_rate": 0.0007156028084252759, + "loss": 1.5672, + "theoretical_loss": 4.066393854951903, + "tokens_seen": 377683968 + }, + { + "epoch": 0.11, + "learning_rate": 0.0007155867602808426, + "loss": 1.6641, + "theoretical_loss": 4.066314716072864, + "tokens_seen": 377749504 + }, + { + "epoch": 0.11, + "learning_rate": 0.0007155707121364093, + "loss": 1.5088, + "theoretical_loss": 4.066235594766088, + "tokens_seen": 377815040 + }, + { + "epoch": 0.11, + "learning_rate": 0.000715554663991976, + "loss": 1.5734, + "theoretical_loss": 4.066156491024625, + "tokens_seen": 377880576 + }, + { + "epoch": 0.11, + "learning_rate": 0.0007155386158475427, + "loss": 1.5506, + "theoretical_loss": 4.06607740484153, + "tokens_seen": 377946112 + }, + { + "epoch": 0.11, + "learning_rate": 0.0007155225677031094, + "loss": 1.7745, + "theoretical_loss": 4.065998336209863, + "tokens_seen": 378011648 + }, + { + "epoch": 0.11, + "learning_rate": 0.0007155065195586761, + "loss": 1.6867, + "theoretical_loss": 4.065919285122686, + "tokens_seen": 378077184 + }, + { + "epoch": 0.11, + "learning_rate": 0.0007154904714142428, + "loss": 1.6067, + "theoretical_loss": 4.065840251573066, + "tokens_seen": 378142720 + }, + { + "epoch": 0.11, + "learning_rate": 0.0007154744232698095, + "loss": 1.8035, + "theoretical_loss": 4.065761235554075, + "tokens_seen": 378208256 + }, + { + "epoch": 0.11, + "learning_rate": 0.0007154583751253762, + "loss": 1.5278, + "theoretical_loss": 4.065682237058787, + "tokens_seen": 378273792 + }, + { + "epoch": 0.11, + "learning_rate": 0.0007154423269809429, + "loss": 1.629, + "theoretical_loss": 4.06560325608028, + "tokens_seen": 378339328 + }, + { + "epoch": 0.11, + "learning_rate": 0.0007154262788365096, + "loss": 1.5382, + "theoretical_loss": 4.065524292611637, + "tokens_seen": 378404864 + }, + { + "epoch": 0.11, + "learning_rate": 0.0007154102306920763, + "loss": 1.603, + "theoretical_loss": 4.065445346645945, + "tokens_seen": 378470400 + }, + { + "epoch": 0.11, + "learning_rate": 0.000715394182547643, + "loss": 1.6248, + "theoretical_loss": 4.0653664181762945, + "tokens_seen": 378535936 + }, + { + "epoch": 0.11, + "learning_rate": 0.0007153781344032097, + "loss": 1.7239, + "theoretical_loss": 4.06528750719578, + "tokens_seen": 378601472 + }, + { + "epoch": 0.11, + "learning_rate": 0.0007153620862587764, + "loss": 1.7048, + "theoretical_loss": 4.0652086136975, + "tokens_seen": 378667008 + }, + { + "epoch": 0.11, + "learning_rate": 0.0007153460381143431, + "loss": 1.5878, + "theoretical_loss": 4.065129737674556, + "tokens_seen": 378732544 + }, + { + "epoch": 0.11, + "learning_rate": 0.0007153299899699098, + "loss": 1.503, + "theoretical_loss": 4.065050879120054, + "tokens_seen": 378798080 + }, + { + "epoch": 0.11, + "learning_rate": 0.0007153139418254765, + "loss": 1.5666, + "theoretical_loss": 4.064972038027106, + "tokens_seen": 378863616 + }, + { + "epoch": 0.11, + "learning_rate": 0.0007152978936810432, + "loss": 1.6294, + "theoretical_loss": 4.064893214388825, + "tokens_seen": 378929152 + }, + { + "epoch": 0.11, + "learning_rate": 0.0007152818455366099, + "loss": 1.5769, + "theoretical_loss": 4.064814408198327, + "tokens_seen": 378994688 + }, + { + "epoch": 0.11, + "learning_rate": 0.0007152657973921766, + "loss": 1.6715, + "theoretical_loss": 4.0647356194487365, + "tokens_seen": 379060224 + }, + { + "epoch": 0.11, + "learning_rate": 0.0007152497492477433, + "loss": 1.5349, + "theoretical_loss": 4.064656848133178, + "tokens_seen": 379125760 + }, + { + "epoch": 0.11, + "learning_rate": 0.00071523370110331, + "loss": 1.6167, + "theoretical_loss": 4.064578094244782, + "tokens_seen": 379191296 + }, + { + "epoch": 0.11, + "learning_rate": 0.0007152176529588767, + "loss": 1.5488, + "theoretical_loss": 4.06449935777668, + "tokens_seen": 379256832 + }, + { + "epoch": 0.11, + "learning_rate": 0.0007152016048144434, + "loss": 1.6115, + "theoretical_loss": 4.0644206387220105, + "tokens_seen": 379322368 + }, + { + "epoch": 0.11, + "learning_rate": 0.0007151855566700101, + "loss": 1.619, + "theoretical_loss": 4.064341937073916, + "tokens_seen": 379387904 + }, + { + "epoch": 0.11, + "learning_rate": 0.0007151695085255768, + "loss": 1.706, + "theoretical_loss": 4.0642632528255405, + "tokens_seen": 379453440 + }, + { + "epoch": 0.12, + "learning_rate": 0.0007151534603811435, + "loss": 1.7859, + "theoretical_loss": 4.064184585970033, + "tokens_seen": 379518976 + }, + { + "epoch": 0.12, + "learning_rate": 0.0007151374122367102, + "loss": 1.546, + "theoretical_loss": 4.064105936500547, + "tokens_seen": 379584512 + }, + { + "epoch": 0.12, + "learning_rate": 0.0007151213640922769, + "loss": 1.545, + "theoretical_loss": 4.064027304410239, + "tokens_seen": 379650048 + }, + { + "epoch": 0.12, + "learning_rate": 0.0007151053159478436, + "loss": 1.6316, + "theoretical_loss": 4.063948689692269, + "tokens_seen": 379715584 + }, + { + "epoch": 0.12, + "learning_rate": 0.0007150892678034103, + "loss": 1.7206, + "theoretical_loss": 4.063870092339803, + "tokens_seen": 379781120 + }, + { + "epoch": 0.12, + "learning_rate": 0.000715073219658977, + "loss": 1.6099, + "theoretical_loss": 4.0637915123460076, + "tokens_seen": 379846656 + }, + { + "epoch": 0.12, + "learning_rate": 0.0007150571715145436, + "loss": 1.6265, + "theoretical_loss": 4.0637129497040565, + "tokens_seen": 379912192 + }, + { + "epoch": 0.12, + "learning_rate": 0.0007150411233701104, + "loss": 1.568, + "theoretical_loss": 4.063634404407126, + "tokens_seen": 379977728 + }, + { + "epoch": 0.12, + "learning_rate": 0.000715025075225677, + "loss": 1.6438, + "theoretical_loss": 4.063555876448395, + "tokens_seen": 380043264 + }, + { + "epoch": 0.12, + "objective/train/docs_used": 144698, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.5444930791854858, + "objective/train/theoretical_loss": 4.063477365821048, + "objective/train/tokens_used": 400568800, + "theoretical_loss": 4.063477365821048, + "tokens_seen": 380108800 + }, + { + "epoch": 0.12, + "learning_rate": 0.0007150090270812438, + "loss": 1.5281, + "theoretical_loss": 4.063477365821048, + "tokens_seen": 380108800 + }, + { + "epoch": 0.12, + "learning_rate": 0.0007149929789368104, + "loss": 1.658, + "theoretical_loss": 4.063398872518273, + "tokens_seen": 380174336 + }, + { + "epoch": 0.12, + "learning_rate": 0.0007149769307923772, + "loss": 1.6047, + "theoretical_loss": 4.0633203965332605, + "tokens_seen": 380239872 + }, + { + "epoch": 0.12, + "learning_rate": 0.0007149608826479438, + "loss": 1.6059, + "theoretical_loss": 4.063241937859207, + "tokens_seen": 380305408 + }, + { + "epoch": 0.12, + "learning_rate": 0.0007149448345035106, + "loss": 1.6871, + "theoretical_loss": 4.063163496489312, + "tokens_seen": 380370944 + }, + { + "epoch": 0.12, + "learning_rate": 0.0007149287863590772, + "loss": 1.6641, + "theoretical_loss": 4.0630850724167775, + "tokens_seen": 380436480 + }, + { + "epoch": 0.12, + "learning_rate": 0.0007149127382146439, + "loss": 1.6126, + "theoretical_loss": 4.063006665634811, + "tokens_seen": 380502016 + }, + { + "epoch": 0.12, + "learning_rate": 0.0007148966900702106, + "loss": 1.6812, + "theoretical_loss": 4.062928276136624, + "tokens_seen": 380567552 + }, + { + "epoch": 0.12, + "learning_rate": 0.0007148806419257773, + "loss": 1.5666, + "theoretical_loss": 4.0628499039154296, + "tokens_seen": 380633088 + }, + { + "epoch": 0.12, + "learning_rate": 0.000714864593781344, + "loss": 1.518, + "theoretical_loss": 4.062771548964447, + "tokens_seen": 380698624 + }, + { + "epoch": 0.12, + "learning_rate": 0.0007148485456369107, + "loss": 1.5556, + "theoretical_loss": 4.062693211276899, + "tokens_seen": 380764160 + }, + { + "epoch": 0.12, + "learning_rate": 0.0007148324974924774, + "loss": 1.6571, + "theoretical_loss": 4.062614890846012, + "tokens_seen": 380829696 + }, + { + "epoch": 0.12, + "learning_rate": 0.0007148164493480441, + "loss": 1.651, + "theoretical_loss": 4.062536587665014, + "tokens_seen": 380895232 + }, + { + "epoch": 0.12, + "learning_rate": 0.0007148004012036108, + "loss": 1.565, + "theoretical_loss": 4.0624583017271405, + "tokens_seen": 380960768 + }, + { + "epoch": 0.12, + "learning_rate": 0.0007147843530591775, + "loss": 1.7833, + "theoretical_loss": 4.062380033025629, + "tokens_seen": 381026304 + }, + { + "epoch": 0.12, + "learning_rate": 0.0007147683049147442, + "loss": 1.6693, + "theoretical_loss": 4.062301781553719, + "tokens_seen": 381091840 + }, + { + "epoch": 0.12, + "learning_rate": 0.0007147522567703109, + "loss": 1.5228, + "theoretical_loss": 4.062223547304658, + "tokens_seen": 381157376 + }, + { + "epoch": 0.12, + "learning_rate": 0.0007147362086258776, + "loss": 1.6004, + "theoretical_loss": 4.0621453302716946, + "tokens_seen": 381222912 + }, + { + "epoch": 0.12, + "learning_rate": 0.0007147201604814443, + "loss": 1.5543, + "theoretical_loss": 4.062067130448081, + "tokens_seen": 381288448 + }, + { + "epoch": 0.12, + "learning_rate": 0.000714704112337011, + "loss": 1.5874, + "theoretical_loss": 4.061988947827073, + "tokens_seen": 381353984 + }, + { + "epoch": 0.12, + "learning_rate": 0.0007146880641925778, + "loss": 1.6314, + "theoretical_loss": 4.061910782401933, + "tokens_seen": 381419520 + }, + { + "epoch": 0.12, + "learning_rate": 0.0007146720160481444, + "loss": 1.769, + "theoretical_loss": 4.061832634165923, + "tokens_seen": 381485056 + }, + { + "epoch": 0.12, + "learning_rate": 0.0007146559679037112, + "loss": 1.6047, + "theoretical_loss": 4.061754503112313, + "tokens_seen": 381550592 + }, + { + "epoch": 0.12, + "learning_rate": 0.0007146399197592778, + "loss": 1.51, + "theoretical_loss": 4.061676389234375, + "tokens_seen": 381616128 + }, + { + "epoch": 0.12, + "learning_rate": 0.0007146238716148446, + "loss": 1.5548, + "theoretical_loss": 4.061598292525382, + "tokens_seen": 381681664 + }, + { + "epoch": 0.12, + "learning_rate": 0.0007146078234704112, + "loss": 1.7062, + "theoretical_loss": 4.061520212978616, + "tokens_seen": 381747200 + }, + { + "epoch": 0.12, + "learning_rate": 0.000714591775325978, + "loss": 1.6215, + "theoretical_loss": 4.061442150587358, + "tokens_seen": 381812736 + }, + { + "epoch": 0.12, + "learning_rate": 0.0007145757271815446, + "loss": 1.6426, + "theoretical_loss": 4.0613641053448974, + "tokens_seen": 381878272 + }, + { + "epoch": 0.12, + "learning_rate": 0.0007145596790371114, + "loss": 1.7301, + "theoretical_loss": 4.061286077244523, + "tokens_seen": 381943808 + }, + { + "epoch": 0.12, + "learning_rate": 0.000714543630892678, + "loss": 1.7452, + "theoretical_loss": 4.06120806627953, + "tokens_seen": 382009344 + }, + { + "epoch": 0.12, + "learning_rate": 0.0007145275827482448, + "loss": 1.6606, + "theoretical_loss": 4.061130072443216, + "tokens_seen": 382074880 + }, + { + "epoch": 0.12, + "learning_rate": 0.0007145115346038114, + "loss": 1.5952, + "theoretical_loss": 4.0610520957288845, + "tokens_seen": 382140416 + }, + { + "epoch": 0.12, + "learning_rate": 0.0007144954864593782, + "loss": 1.6226, + "theoretical_loss": 4.06097413612984, + "tokens_seen": 382205952 + }, + { + "epoch": 0.12, + "learning_rate": 0.0007144794383149448, + "loss": 1.5911, + "theoretical_loss": 4.060896193639392, + "tokens_seen": 382271488 + }, + { + "epoch": 0.12, + "learning_rate": 0.0007144633901705116, + "loss": 1.7186, + "theoretical_loss": 4.060818268250855, + "tokens_seen": 382337024 + }, + { + "epoch": 0.12, + "learning_rate": 0.0007144473420260782, + "loss": 1.6811, + "theoretical_loss": 4.060740359957546, + "tokens_seen": 382402560 + }, + { + "epoch": 0.12, + "learning_rate": 0.000714431293881645, + "loss": 1.6081, + "theoretical_loss": 4.060662468752784, + "tokens_seen": 382468096 + }, + { + "epoch": 0.12, + "learning_rate": 0.0007144152457372116, + "loss": 1.7339, + "theoretical_loss": 4.060584594629895, + "tokens_seen": 382533632 + }, + { + "epoch": 0.12, + "learning_rate": 0.0007143991975927784, + "loss": 1.5921, + "theoretical_loss": 4.060506737582208, + "tokens_seen": 382599168 + }, + { + "epoch": 0.12, + "learning_rate": 0.000714383149448345, + "loss": 1.6997, + "theoretical_loss": 4.060428897603054, + "tokens_seen": 382664704 + }, + { + "epoch": 0.12, + "learning_rate": 0.0007143671013039118, + "loss": 1.665, + "theoretical_loss": 4.060351074685769, + "tokens_seen": 382730240 + }, + { + "epoch": 0.12, + "learning_rate": 0.0007143510531594785, + "loss": 1.5317, + "theoretical_loss": 4.060273268823693, + "tokens_seen": 382795776 + }, + { + "epoch": 0.12, + "learning_rate": 0.0007143350050150452, + "loss": 1.6548, + "theoretical_loss": 4.060195480010169, + "tokens_seen": 382861312 + }, + { + "epoch": 0.12, + "learning_rate": 0.0007143189568706119, + "loss": 1.5561, + "theoretical_loss": 4.060117708238544, + "tokens_seen": 382926848 + }, + { + "epoch": 0.12, + "learning_rate": 0.0007143029087261786, + "loss": 1.6994, + "theoretical_loss": 4.060039953502169, + "tokens_seen": 382992384 + }, + { + "epoch": 0.12, + "learning_rate": 0.0007142868605817453, + "loss": 1.5598, + "theoretical_loss": 4.059962215794399, + "tokens_seen": 383057920 + }, + { + "epoch": 0.12, + "learning_rate": 0.000714270812437312, + "loss": 1.5749, + "theoretical_loss": 4.059884495108591, + "tokens_seen": 383123456 + }, + { + "epoch": 0.12, + "learning_rate": 0.0007142547642928787, + "loss": 1.6863, + "theoretical_loss": 4.059806791438108, + "tokens_seen": 383188992 + }, + { + "epoch": 0.12, + "learning_rate": 0.0007142387161484454, + "loss": 1.524, + "theoretical_loss": 4.059729104776316, + "tokens_seen": 383254528 + }, + { + "epoch": 0.12, + "learning_rate": 0.0007142226680040121, + "loss": 1.6369, + "theoretical_loss": 4.059651435116583, + "tokens_seen": 383320064 + }, + { + "epoch": 0.12, + "objective/train/docs_used": 145827, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.628687858581543, + "objective/train/theoretical_loss": 4.059573782452284, + "objective/train/tokens_used": 403845600, + "theoretical_loss": 4.059573782452284, + "tokens_seen": 383385600 + }, + { + "epoch": 0.12, + "learning_rate": 0.0007142066198595788, + "loss": 1.6046, + "theoretical_loss": 4.059573782452284, + "tokens_seen": 383385600 + }, + { + "epoch": 0.12, + "learning_rate": 0.0007141905717151455, + "loss": 1.5903, + "theoretical_loss": 4.0594961467767945, + "tokens_seen": 383451136 + }, + { + "epoch": 0.12, + "learning_rate": 0.0007141745235707122, + "loss": 1.5756, + "theoretical_loss": 4.059418528083495, + "tokens_seen": 383516672 + }, + { + "epoch": 0.12, + "learning_rate": 0.0007141584754262789, + "loss": 1.6542, + "theoretical_loss": 4.0593409263657705, + "tokens_seen": 383582208 + }, + { + "epoch": 0.12, + "learning_rate": 0.0007141424272818456, + "loss": 1.6146, + "theoretical_loss": 4.059263341617009, + "tokens_seen": 383647744 + }, + { + "epoch": 0.12, + "learning_rate": 0.0007141263791374123, + "loss": 1.5073, + "theoretical_loss": 4.059185773830602, + "tokens_seen": 383713280 + }, + { + "epoch": 0.12, + "learning_rate": 0.000714110330992979, + "loss": 1.6973, + "theoretical_loss": 4.059108222999944, + "tokens_seen": 383778816 + }, + { + "epoch": 0.12, + "learning_rate": 0.0007140942828485457, + "loss": 1.6281, + "theoretical_loss": 4.059030689118436, + "tokens_seen": 383844352 + }, + { + "epoch": 0.12, + "learning_rate": 0.0007140782347041124, + "loss": 1.5579, + "theoretical_loss": 4.058953172179479, + "tokens_seen": 383909888 + }, + { + "epoch": 0.12, + "learning_rate": 0.0007140621865596791, + "loss": 1.6811, + "theoretical_loss": 4.05887567217648, + "tokens_seen": 383975424 + }, + { + "epoch": 0.12, + "learning_rate": 0.0007140461384152458, + "loss": 1.5813, + "theoretical_loss": 4.05879818910285, + "tokens_seen": 384040960 + }, + { + "epoch": 0.12, + "learning_rate": 0.0007140300902708125, + "loss": 1.7124, + "theoretical_loss": 4.058720722952002, + "tokens_seen": 384106496 + }, + { + "epoch": 0.12, + "learning_rate": 0.0007140140421263792, + "loss": 1.6188, + "theoretical_loss": 4.058643273717355, + "tokens_seen": 384172032 + }, + { + "epoch": 0.12, + "learning_rate": 0.0007139979939819459, + "loss": 1.5504, + "theoretical_loss": 4.058565841392328, + "tokens_seen": 384237568 + }, + { + "epoch": 0.12, + "learning_rate": 0.0007139819458375126, + "loss": 1.6742, + "theoretical_loss": 4.058488425970347, + "tokens_seen": 384303104 + }, + { + "epoch": 0.12, + "learning_rate": 0.0007139658976930793, + "loss": 1.5306, + "theoretical_loss": 4.058411027444841, + "tokens_seen": 384368640 + }, + { + "epoch": 0.12, + "learning_rate": 0.000713949849548646, + "loss": 1.6376, + "theoretical_loss": 4.05833364580924, + "tokens_seen": 384434176 + }, + { + "epoch": 0.12, + "learning_rate": 0.0007139338014042127, + "loss": 1.622, + "theoretical_loss": 4.058256281056984, + "tokens_seen": 384499712 + }, + { + "epoch": 0.12, + "learning_rate": 0.0007139177532597794, + "loss": 1.6283, + "theoretical_loss": 4.058178933181509, + "tokens_seen": 384565248 + }, + { + "epoch": 0.12, + "learning_rate": 0.0007139017051153461, + "loss": 1.6026, + "theoretical_loss": 4.05810160217626, + "tokens_seen": 384630784 + }, + { + "epoch": 0.12, + "learning_rate": 0.0007138856569709128, + "loss": 1.6336, + "theoretical_loss": 4.058024288034684, + "tokens_seen": 384696320 + }, + { + "epoch": 0.12, + "learning_rate": 0.0007138696088264795, + "loss": 1.5454, + "theoretical_loss": 4.057946990750231, + "tokens_seen": 384761856 + }, + { + "epoch": 0.12, + "learning_rate": 0.0007138535606820462, + "loss": 1.7241, + "theoretical_loss": 4.057869710316355, + "tokens_seen": 384827392 + }, + { + "epoch": 0.12, + "learning_rate": 0.0007138375125376129, + "loss": 1.6373, + "theoretical_loss": 4.0577924467265145, + "tokens_seen": 384892928 + }, + { + "epoch": 0.12, + "learning_rate": 0.0007138214643931796, + "loss": 1.7471, + "theoretical_loss": 4.057715199974171, + "tokens_seen": 384958464 + }, + { + "epoch": 0.12, + "learning_rate": 0.0007138054162487463, + "loss": 1.6167, + "theoretical_loss": 4.05763797005279, + "tokens_seen": 385024000 + }, + { + "epoch": 0.12, + "learning_rate": 0.000713789368104313, + "loss": 1.6074, + "theoretical_loss": 4.05756075695584, + "tokens_seen": 385089536 + }, + { + "epoch": 0.12, + "learning_rate": 0.0007137733199598797, + "loss": 1.6916, + "theoretical_loss": 4.0574835606767925, + "tokens_seen": 385155072 + }, + { + "epoch": 0.12, + "learning_rate": 0.0007137572718154464, + "loss": 1.6738, + "theoretical_loss": 4.057406381209126, + "tokens_seen": 385220608 + }, + { + "epoch": 0.12, + "learning_rate": 0.0007137412236710131, + "loss": 1.6421, + "theoretical_loss": 4.057329218546318, + "tokens_seen": 385286144 + }, + { + "epoch": 0.12, + "learning_rate": 0.0007137251755265798, + "loss": 1.6512, + "theoretical_loss": 4.0572520726818535, + "tokens_seen": 385351680 + }, + { + "epoch": 0.12, + "learning_rate": 0.0007137091273821465, + "loss": 1.6454, + "theoretical_loss": 4.057174943609218, + "tokens_seen": 385417216 + }, + { + "epoch": 0.12, + "learning_rate": 0.0007136930792377132, + "loss": 1.6194, + "theoretical_loss": 4.057097831321904, + "tokens_seen": 385482752 + }, + { + "epoch": 0.12, + "learning_rate": 0.0007136770310932799, + "loss": 1.6177, + "theoretical_loss": 4.057020735813405, + "tokens_seen": 385548288 + }, + { + "epoch": 0.12, + "learning_rate": 0.0007136609829488466, + "loss": 1.6235, + "theoretical_loss": 4.056943657077219, + "tokens_seen": 385613824 + }, + { + "epoch": 0.12, + "learning_rate": 0.0007136449348044133, + "loss": 1.5713, + "theoretical_loss": 4.056866595106847, + "tokens_seen": 385679360 + }, + { + "epoch": 0.12, + "learning_rate": 0.0007136288866599799, + "loss": 1.6924, + "theoretical_loss": 4.056789549895795, + "tokens_seen": 385744896 + }, + { + "epoch": 0.12, + "learning_rate": 0.0007136128385155467, + "loss": 1.6133, + "theoretical_loss": 4.056712521437571, + "tokens_seen": 385810432 + }, + { + "epoch": 0.12, + "learning_rate": 0.0007135967903711133, + "loss": 1.6086, + "theoretical_loss": 4.056635509725689, + "tokens_seen": 385875968 + }, + { + "epoch": 0.12, + "learning_rate": 0.0007135807422266801, + "loss": 1.7305, + "theoretical_loss": 4.056558514753664, + "tokens_seen": 385941504 + }, + { + "epoch": 0.12, + "learning_rate": 0.0007135646940822467, + "loss": 1.5386, + "theoretical_loss": 4.056481536515014, + "tokens_seen": 386007040 + }, + { + "epoch": 0.12, + "learning_rate": 0.0007135486459378135, + "loss": 1.5916, + "theoretical_loss": 4.056404575003266, + "tokens_seen": 386072576 + }, + { + "epoch": 0.12, + "learning_rate": 0.0007135325977933801, + "loss": 1.5872, + "theoretical_loss": 4.0563276302119435, + "tokens_seen": 386138112 + }, + { + "epoch": 0.12, + "learning_rate": 0.0007135165496489468, + "loss": 1.5586, + "theoretical_loss": 4.056250702134578, + "tokens_seen": 386203648 + }, + { + "epoch": 0.12, + "learning_rate": 0.0007135005015045135, + "loss": 1.6418, + "theoretical_loss": 4.0561737907647055, + "tokens_seen": 386269184 + }, + { + "epoch": 0.12, + "learning_rate": 0.0007134844533600803, + "loss": 1.6244, + "theoretical_loss": 4.056096896095861, + "tokens_seen": 386334720 + }, + { + "epoch": 0.12, + "learning_rate": 0.0007134684052156469, + "loss": 1.5695, + "theoretical_loss": 4.056020018121587, + "tokens_seen": 386400256 + }, + { + "epoch": 0.12, + "learning_rate": 0.0007134523570712137, + "loss": 1.5278, + "theoretical_loss": 4.055943156835428, + "tokens_seen": 386465792 + }, + { + "epoch": 0.12, + "learning_rate": 0.0007134363089267803, + "loss": 1.5912, + "theoretical_loss": 4.055866312230933, + "tokens_seen": 386531328 + }, + { + "epoch": 0.12, + "learning_rate": 0.0007134202607823471, + "loss": 1.6199, + "theoretical_loss": 4.055789484301653, + "tokens_seen": 386596864 + }, + { + "epoch": 0.12, + "objective/train/docs_used": 146981, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.7513854503631592, + "objective/train/theoretical_loss": 4.055712673041144, + "objective/train/tokens_used": 407122400, + "theoretical_loss": 4.055712673041144, + "tokens_seen": 386662400 + }, + { + "epoch": 0.12, + "learning_rate": 0.0007134042126379137, + "loss": 1.7525, + "theoretical_loss": 4.055712673041144, + "tokens_seen": 386662400 + }, + { + "epoch": 0.12, + "learning_rate": 0.0007133881644934805, + "loss": 1.6638, + "theoretical_loss": 4.055635878442967, + "tokens_seen": 386727936 + }, + { + "epoch": 0.12, + "learning_rate": 0.0007133721163490471, + "loss": 1.6044, + "theoretical_loss": 4.055559100500682, + "tokens_seen": 386793472 + }, + { + "epoch": 0.12, + "learning_rate": 0.0007133560682046139, + "loss": 1.6154, + "theoretical_loss": 4.055482339207858, + "tokens_seen": 386859008 + }, + { + "epoch": 0.12, + "learning_rate": 0.0007133400200601805, + "loss": 1.4859, + "theoretical_loss": 4.055405594558062, + "tokens_seen": 386924544 + }, + { + "epoch": 0.12, + "learning_rate": 0.0007133239719157473, + "loss": 1.5873, + "theoretical_loss": 4.05532886654487, + "tokens_seen": 386990080 + }, + { + "epoch": 0.12, + "learning_rate": 0.0007133079237713139, + "loss": 1.6966, + "theoretical_loss": 4.055252155161858, + "tokens_seen": 387055616 + }, + { + "epoch": 0.12, + "learning_rate": 0.0007132918756268807, + "loss": 1.7077, + "theoretical_loss": 4.055175460402607, + "tokens_seen": 387121152 + }, + { + "epoch": 0.12, + "learning_rate": 0.0007132758274824473, + "loss": 1.7197, + "theoretical_loss": 4.055098782260701, + "tokens_seen": 387186688 + }, + { + "epoch": 0.12, + "learning_rate": 0.0007132597793380141, + "loss": 1.551, + "theoretical_loss": 4.055022120729728, + "tokens_seen": 387252224 + }, + { + "epoch": 0.12, + "learning_rate": 0.0007132437311935807, + "loss": 1.5555, + "theoretical_loss": 4.054945475803278, + "tokens_seen": 387317760 + }, + { + "epoch": 0.12, + "learning_rate": 0.0007132276830491475, + "loss": 1.7061, + "theoretical_loss": 4.054868847474948, + "tokens_seen": 387383296 + }, + { + "epoch": 0.12, + "learning_rate": 0.0007132116349047141, + "loss": 1.5361, + "theoretical_loss": 4.054792235738335, + "tokens_seen": 387448832 + }, + { + "epoch": 0.12, + "learning_rate": 0.0007131955867602809, + "loss": 1.5471, + "theoretical_loss": 4.054715640587042, + "tokens_seen": 387514368 + }, + { + "epoch": 0.12, + "learning_rate": 0.0007131795386158475, + "loss": 1.6316, + "theoretical_loss": 4.054639062014673, + "tokens_seen": 387579904 + }, + { + "epoch": 0.12, + "learning_rate": 0.0007131634904714143, + "loss": 1.6182, + "theoretical_loss": 4.054562500014839, + "tokens_seen": 387645440 + }, + { + "epoch": 0.12, + "learning_rate": 0.0007131474423269809, + "loss": 1.645, + "theoretical_loss": 4.054485954581152, + "tokens_seen": 387710976 + }, + { + "epoch": 0.12, + "learning_rate": 0.0007131313941825477, + "loss": 1.7088, + "theoretical_loss": 4.054409425707227, + "tokens_seen": 387776512 + }, + { + "epoch": 0.12, + "learning_rate": 0.0007131153460381143, + "loss": 1.6238, + "theoretical_loss": 4.054332913386686, + "tokens_seen": 387842048 + }, + { + "epoch": 0.12, + "learning_rate": 0.0007130992978936811, + "loss": 1.5816, + "theoretical_loss": 4.054256417613149, + "tokens_seen": 387907584 + }, + { + "epoch": 0.12, + "learning_rate": 0.0007130832497492477, + "loss": 1.6087, + "theoretical_loss": 4.054179938380246, + "tokens_seen": 387973120 + }, + { + "epoch": 0.12, + "learning_rate": 0.0007130672016048145, + "loss": 1.5729, + "theoretical_loss": 4.054103475681605, + "tokens_seen": 388038656 + }, + { + "epoch": 0.12, + "learning_rate": 0.0007130511534603811, + "loss": 1.4364, + "theoretical_loss": 4.054027029510861, + "tokens_seen": 388104192 + }, + { + "epoch": 0.12, + "learning_rate": 0.0007130351053159479, + "loss": 1.4969, + "theoretical_loss": 4.053950599861651, + "tokens_seen": 388169728 + }, + { + "epoch": 0.12, + "learning_rate": 0.0007130190571715146, + "loss": 1.5896, + "theoretical_loss": 4.053874186727617, + "tokens_seen": 388235264 + }, + { + "epoch": 0.12, + "learning_rate": 0.0007130030090270813, + "loss": 1.7395, + "theoretical_loss": 4.053797790102402, + "tokens_seen": 388300800 + }, + { + "epoch": 0.12, + "learning_rate": 0.000712986960882648, + "loss": 1.6385, + "theoretical_loss": 4.053721409979654, + "tokens_seen": 388366336 + }, + { + "epoch": 0.12, + "learning_rate": 0.0007129709127382147, + "loss": 1.6698, + "theoretical_loss": 4.053645046353025, + "tokens_seen": 388431872 + }, + { + "epoch": 0.12, + "learning_rate": 0.0007129548645937814, + "loss": 1.4926, + "theoretical_loss": 4.05356869921617, + "tokens_seen": 388497408 + }, + { + "epoch": 0.12, + "learning_rate": 0.0007129388164493481, + "loss": 1.5984, + "theoretical_loss": 4.053492368562747, + "tokens_seen": 388562944 + }, + { + "epoch": 0.12, + "learning_rate": 0.0007129227683049148, + "loss": 1.691, + "theoretical_loss": 4.0534160543864175, + "tokens_seen": 388628480 + }, + { + "epoch": 0.12, + "learning_rate": 0.0007129067201604815, + "loss": 1.6334, + "theoretical_loss": 4.053339756680849, + "tokens_seen": 388694016 + }, + { + "epoch": 0.12, + "learning_rate": 0.0007128906720160482, + "loss": 1.6233, + "theoretical_loss": 4.053263475439707, + "tokens_seen": 388759552 + }, + { + "epoch": 0.12, + "learning_rate": 0.0007128746238716149, + "loss": 1.5284, + "theoretical_loss": 4.053187210656668, + "tokens_seen": 388825088 + }, + { + "epoch": 0.12, + "learning_rate": 0.0007128585757271816, + "loss": 1.6833, + "theoretical_loss": 4.053110962325406, + "tokens_seen": 388890624 + }, + { + "epoch": 0.12, + "learning_rate": 0.0007128425275827483, + "loss": 1.4977, + "theoretical_loss": 4.053034730439601, + "tokens_seen": 388956160 + }, + { + "epoch": 0.12, + "learning_rate": 0.000712826479438315, + "loss": 1.619, + "theoretical_loss": 4.052958514992934, + "tokens_seen": 389021696 + }, + { + "epoch": 0.12, + "learning_rate": 0.0007128104312938817, + "loss": 1.6477, + "theoretical_loss": 4.052882315979094, + "tokens_seen": 389087232 + }, + { + "epoch": 0.12, + "learning_rate": 0.0007127943831494484, + "loss": 1.4743, + "theoretical_loss": 4.052806133391769, + "tokens_seen": 389152768 + }, + { + "epoch": 0.12, + "learning_rate": 0.0007127783350050151, + "loss": 1.6514, + "theoretical_loss": 4.052729967224654, + "tokens_seen": 389218304 + }, + { + "epoch": 0.12, + "learning_rate": 0.0007127622868605818, + "loss": 1.5455, + "theoretical_loss": 4.0526538174714455, + "tokens_seen": 389283840 + }, + { + "epoch": 0.12, + "learning_rate": 0.0007127462387161485, + "loss": 1.6476, + "theoretical_loss": 4.052577684125843, + "tokens_seen": 389349376 + }, + { + "epoch": 0.12, + "learning_rate": 0.0007127301905717152, + "loss": 1.6988, + "theoretical_loss": 4.052501567181552, + "tokens_seen": 389414912 + }, + { + "epoch": 0.12, + "learning_rate": 0.0007127141424272819, + "loss": 1.6693, + "theoretical_loss": 4.052425466632278, + "tokens_seen": 389480448 + }, + { + "epoch": 0.12, + "learning_rate": 0.0007126980942828486, + "loss": 1.6162, + "theoretical_loss": 4.052349382471734, + "tokens_seen": 389545984 + }, + { + "epoch": 0.12, + "learning_rate": 0.0007126820461384153, + "loss": 1.4617, + "theoretical_loss": 4.052273314693632, + "tokens_seen": 389611520 + }, + { + "epoch": 0.12, + "learning_rate": 0.000712665997993982, + "loss": 1.6062, + "theoretical_loss": 4.05219726329169, + "tokens_seen": 389677056 + }, + { + "epoch": 0.12, + "learning_rate": 0.0007126499498495487, + "loss": 1.7045, + "theoretical_loss": 4.052121228259631, + "tokens_seen": 389742592 + }, + { + "epoch": 0.12, + "learning_rate": 0.0007126339017051154, + "loss": 1.6458, + "theoretical_loss": 4.052045209591179, + "tokens_seen": 389808128 + }, + { + "epoch": 0.12, + "learning_rate": 0.0007126178535606821, + "loss": 1.5855, + "theoretical_loss": 4.05196920728006, + "tokens_seen": 389873664 + }, + { + "epoch": 0.12, + "objective/train/docs_used": 148229, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.7241127490997314, + "objective/train/theoretical_loss": 4.051893221320009, + "objective/train/tokens_used": 410399200, + "theoretical_loss": 4.051893221320009, + "tokens_seen": 389939200 + }, + { + "epoch": 0.12, + "learning_rate": 0.0007126018054162488, + "loss": 1.7077, + "theoretical_loss": 4.051893221320009, + "tokens_seen": 389939200 + }, + { + "epoch": 0.12, + "learning_rate": 0.0007125857572718155, + "loss": 1.6347, + "theoretical_loss": 4.05181725170476, + "tokens_seen": 390004736 + }, + { + "epoch": 0.12, + "learning_rate": 0.0007125697091273822, + "loss": 1.6375, + "theoretical_loss": 4.051741298428049, + "tokens_seen": 390070272 + }, + { + "epoch": 0.12, + "learning_rate": 0.0007125536609829489, + "loss": 1.6164, + "theoretical_loss": 4.0516653614836216, + "tokens_seen": 390135808 + }, + { + "epoch": 0.12, + "learning_rate": 0.0007125376128385156, + "loss": 1.5957, + "theoretical_loss": 4.05158944086522, + "tokens_seen": 390201344 + }, + { + "epoch": 0.12, + "learning_rate": 0.0007125215646940823, + "loss": 1.608, + "theoretical_loss": 4.0515135365665955, + "tokens_seen": 390266880 + }, + { + "epoch": 0.12, + "learning_rate": 0.000712505516549649, + "loss": 1.4815, + "theoretical_loss": 4.0514376485815, + "tokens_seen": 390332416 + }, + { + "epoch": 0.12, + "learning_rate": 0.0007124894684052157, + "loss": 1.5895, + "theoretical_loss": 4.051361776903688, + "tokens_seen": 390397952 + }, + { + "epoch": 0.12, + "learning_rate": 0.0007124734202607824, + "loss": 1.6628, + "theoretical_loss": 4.051285921526919, + "tokens_seen": 390463488 + }, + { + "epoch": 0.12, + "learning_rate": 0.0007124573721163491, + "loss": 1.6482, + "theoretical_loss": 4.051210082444956, + "tokens_seen": 390529024 + }, + { + "epoch": 0.12, + "learning_rate": 0.0007124413239719158, + "loss": 1.5451, + "theoretical_loss": 4.051134259651565, + "tokens_seen": 390594560 + }, + { + "epoch": 0.12, + "learning_rate": 0.0007124252758274825, + "loss": 1.716, + "theoretical_loss": 4.0510584531405165, + "tokens_seen": 390660096 + }, + { + "epoch": 0.12, + "learning_rate": 0.0007124092276830492, + "loss": 1.549, + "theoretical_loss": 4.050982662905581, + "tokens_seen": 390725632 + }, + { + "epoch": 0.12, + "learning_rate": 0.0007123931795386159, + "loss": 1.5987, + "theoretical_loss": 4.0509068889405375, + "tokens_seen": 390791168 + }, + { + "epoch": 0.12, + "learning_rate": 0.0007123771313941826, + "loss": 1.7252, + "theoretical_loss": 4.050831131239164, + "tokens_seen": 390856704 + }, + { + "epoch": 0.12, + "learning_rate": 0.0007123610832497493, + "loss": 1.7003, + "theoretical_loss": 4.050755389795244, + "tokens_seen": 390922240 + }, + { + "epoch": 0.12, + "learning_rate": 0.000712345035105316, + "loss": 1.6696, + "theoretical_loss": 4.050679664602564, + "tokens_seen": 390987776 + }, + { + "epoch": 0.12, + "learning_rate": 0.0007123289869608827, + "loss": 1.6625, + "theoretical_loss": 4.050603955654914, + "tokens_seen": 391053312 + }, + { + "epoch": 0.12, + "learning_rate": 0.0007123129388164494, + "loss": 1.5901, + "theoretical_loss": 4.050528262946088, + "tokens_seen": 391118848 + }, + { + "epoch": 0.12, + "learning_rate": 0.0007122968906720161, + "loss": 1.6036, + "theoretical_loss": 4.050452586469882, + "tokens_seen": 391184384 + }, + { + "epoch": 0.12, + "learning_rate": 0.0007122808425275828, + "loss": 1.4128, + "theoretical_loss": 4.050376926220098, + "tokens_seen": 391249920 + }, + { + "epoch": 0.12, + "learning_rate": 0.0007122647943831495, + "loss": 1.529, + "theoretical_loss": 4.050301282190537, + "tokens_seen": 391315456 + }, + { + "epoch": 0.12, + "learning_rate": 0.0007122487462387162, + "loss": 1.6789, + "theoretical_loss": 4.050225654375008, + "tokens_seen": 391380992 + }, + { + "epoch": 0.12, + "learning_rate": 0.0007122326980942828, + "loss": 1.7103, + "theoretical_loss": 4.050150042767321, + "tokens_seen": 391446528 + }, + { + "epoch": 0.12, + "learning_rate": 0.0007122166499498496, + "loss": 1.5652, + "theoretical_loss": 4.050074447361289, + "tokens_seen": 391512064 + }, + { + "epoch": 0.12, + "learning_rate": 0.0007122006018054162, + "loss": 1.5793, + "theoretical_loss": 4.04999886815073, + "tokens_seen": 391577600 + }, + { + "epoch": 0.12, + "learning_rate": 0.000712184553660983, + "loss": 1.5969, + "theoretical_loss": 4.0499233051294645, + "tokens_seen": 391643136 + }, + { + "epoch": 0.12, + "learning_rate": 0.0007121685055165496, + "loss": 1.6479, + "theoretical_loss": 4.049847758291317, + "tokens_seen": 391708672 + }, + { + "epoch": 0.12, + "learning_rate": 0.0007121524573721164, + "loss": 1.6139, + "theoretical_loss": 4.049772227630114, + "tokens_seen": 391774208 + }, + { + "epoch": 0.12, + "learning_rate": 0.000712136409227683, + "loss": 1.6075, + "theoretical_loss": 4.049696713139687, + "tokens_seen": 391839744 + }, + { + "epoch": 0.12, + "learning_rate": 0.0007121203610832498, + "loss": 1.638, + "theoretical_loss": 4.049621214813869, + "tokens_seen": 391905280 + }, + { + "epoch": 0.12, + "learning_rate": 0.0007121043129388164, + "loss": 1.5807, + "theoretical_loss": 4.0495457326465, + "tokens_seen": 391970816 + }, + { + "epoch": 0.12, + "learning_rate": 0.0007120882647943832, + "loss": 1.4724, + "theoretical_loss": 4.049470266631419, + "tokens_seen": 392036352 + }, + { + "epoch": 0.12, + "learning_rate": 0.0007120722166499498, + "loss": 1.6791, + "theoretical_loss": 4.04939481676247, + "tokens_seen": 392101888 + }, + { + "epoch": 0.12, + "learning_rate": 0.0007120561685055166, + "loss": 1.695, + "theoretical_loss": 4.0493193830335015, + "tokens_seen": 392167424 + }, + { + "epoch": 0.12, + "learning_rate": 0.0007120401203610832, + "loss": 1.5693, + "theoretical_loss": 4.0492439654383645, + "tokens_seen": 392232960 + }, + { + "epoch": 0.12, + "learning_rate": 0.00071202407221665, + "loss": 1.6884, + "theoretical_loss": 4.049168563970914, + "tokens_seen": 392298496 + }, + { + "epoch": 0.12, + "learning_rate": 0.0007120080240722166, + "loss": 1.6345, + "theoretical_loss": 4.049093178625007, + "tokens_seen": 392364032 + }, + { + "epoch": 0.12, + "learning_rate": 0.0007119919759277834, + "loss": 1.5625, + "theoretical_loss": 4.049017809394506, + "tokens_seen": 392429568 + }, + { + "epoch": 0.12, + "learning_rate": 0.00071197592778335, + "loss": 1.5021, + "theoretical_loss": 4.048942456273274, + "tokens_seen": 392495104 + }, + { + "epoch": 0.12, + "learning_rate": 0.0007119598796389168, + "loss": 1.5401, + "theoretical_loss": 4.048867119255178, + "tokens_seen": 392560640 + }, + { + "epoch": 0.12, + "learning_rate": 0.0007119438314944834, + "loss": 1.6364, + "theoretical_loss": 4.048791798334092, + "tokens_seen": 392626176 + }, + { + "epoch": 0.12, + "learning_rate": 0.0007119277833500502, + "loss": 1.6589, + "theoretical_loss": 4.048716493503888, + "tokens_seen": 392691712 + }, + { + "epoch": 0.12, + "learning_rate": 0.0007119117352056168, + "loss": 1.7182, + "theoretical_loss": 4.048641204758446, + "tokens_seen": 392757248 + }, + { + "epoch": 0.12, + "learning_rate": 0.0007118956870611836, + "loss": 1.6363, + "theoretical_loss": 4.048565932091646, + "tokens_seen": 392822784 + }, + { + "epoch": 0.12, + "learning_rate": 0.0007118796389167502, + "loss": 1.6736, + "theoretical_loss": 4.048490675497373, + "tokens_seen": 392888320 + }, + { + "epoch": 0.12, + "learning_rate": 0.000711863590772317, + "loss": 1.6343, + "theoretical_loss": 4.048415434969516, + "tokens_seen": 392953856 + }, + { + "epoch": 0.12, + "learning_rate": 0.0007118475426278836, + "loss": 1.5918, + "theoretical_loss": 4.048340210501965, + "tokens_seen": 393019392 + }, + { + "epoch": 0.12, + "learning_rate": 0.0007118314944834504, + "loss": 1.6358, + "theoretical_loss": 4.048265002088614, + "tokens_seen": 393084928 + }, + { + "epoch": 0.12, + "learning_rate": 0.000711815446339017, + "loss": 1.5516, + "theoretical_loss": 4.048189809723364, + "tokens_seen": 393150464 + }, + { + "epoch": 0.12, + "objective/train/docs_used": 149327, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.7545279264450073, + "objective/train/theoretical_loss": 4.048114633400113, + "objective/train/tokens_used": 413676000, + "theoretical_loss": 4.048114633400113, + "tokens_seen": 393216000 + }, + { + "epoch": 0.12, + "learning_rate": 0.0007117993981945838, + "loss": 1.6363, + "theoretical_loss": 4.048114633400113, + "tokens_seen": 393216000 + }, + { + "epoch": 0.12, + "learning_rate": 0.0007117833500501504, + "loss": 1.6494, + "theoretical_loss": 4.048039473112768, + "tokens_seen": 393281536 + }, + { + "epoch": 0.12, + "learning_rate": 0.0007117673019057172, + "loss": 1.5723, + "theoretical_loss": 4.047964328855237, + "tokens_seen": 393347072 + }, + { + "epoch": 0.12, + "learning_rate": 0.0007117512537612838, + "loss": 1.6328, + "theoretical_loss": 4.047889200621428, + "tokens_seen": 393412608 + }, + { + "epoch": 0.12, + "learning_rate": 0.0007117352056168506, + "loss": 1.6961, + "theoretical_loss": 4.047814088405261, + "tokens_seen": 393478144 + }, + { + "epoch": 0.12, + "learning_rate": 0.0007117191574724172, + "loss": 1.7128, + "theoretical_loss": 4.0477389922006495, + "tokens_seen": 393543680 + }, + { + "epoch": 0.12, + "learning_rate": 0.000711703109327984, + "loss": 1.6428, + "theoretical_loss": 4.047663912001518, + "tokens_seen": 393609216 + }, + { + "epoch": 0.12, + "learning_rate": 0.0007116870611835506, + "loss": 1.6497, + "theoretical_loss": 4.047588847801788, + "tokens_seen": 393674752 + }, + { + "epoch": 0.12, + "learning_rate": 0.0007116710130391174, + "loss": 1.6676, + "theoretical_loss": 4.04751379959539, + "tokens_seen": 393740288 + }, + { + "epoch": 0.12, + "learning_rate": 0.000711654964894684, + "loss": 1.6346, + "theoretical_loss": 4.047438767376255, + "tokens_seen": 393805824 + }, + { + "epoch": 0.12, + "learning_rate": 0.0007116389167502508, + "loss": 1.6833, + "theoretical_loss": 4.047363751138316, + "tokens_seen": 393871360 + }, + { + "epoch": 0.12, + "learning_rate": 0.0007116228686058175, + "loss": 1.5737, + "theoretical_loss": 4.047288750875512, + "tokens_seen": 393936896 + }, + { + "epoch": 0.12, + "learning_rate": 0.0007116068204613842, + "loss": 1.651, + "theoretical_loss": 4.047213766581785, + "tokens_seen": 394002432 + }, + { + "epoch": 0.12, + "learning_rate": 0.0007115907723169509, + "loss": 1.47, + "theoretical_loss": 4.047138798251078, + "tokens_seen": 394067968 + }, + { + "epoch": 0.12, + "learning_rate": 0.0007115747241725176, + "loss": 1.6322, + "theoretical_loss": 4.047063845877339, + "tokens_seen": 394133504 + }, + { + "epoch": 0.12, + "learning_rate": 0.0007115586760280843, + "loss": 1.6018, + "theoretical_loss": 4.04698890945452, + "tokens_seen": 394199040 + }, + { + "epoch": 0.12, + "learning_rate": 0.000711542627883651, + "loss": 1.7016, + "theoretical_loss": 4.046913988976575, + "tokens_seen": 394264576 + }, + { + "epoch": 0.12, + "learning_rate": 0.0007115265797392177, + "loss": 1.6217, + "theoretical_loss": 4.046839084437463, + "tokens_seen": 394330112 + }, + { + "epoch": 0.12, + "learning_rate": 0.0007115105315947844, + "loss": 1.657, + "theoretical_loss": 4.046764195831142, + "tokens_seen": 394395648 + }, + { + "epoch": 0.12, + "learning_rate": 0.0007114944834503511, + "loss": 1.6949, + "theoretical_loss": 4.046689323151578, + "tokens_seen": 394461184 + }, + { + "epoch": 0.12, + "learning_rate": 0.0007114784353059178, + "loss": 1.5974, + "theoretical_loss": 4.0466144663927395, + "tokens_seen": 394526720 + }, + { + "epoch": 0.12, + "learning_rate": 0.0007114623871614845, + "loss": 1.7199, + "theoretical_loss": 4.046539625548595, + "tokens_seen": 394592256 + }, + { + "epoch": 0.12, + "learning_rate": 0.0007114463390170512, + "loss": 1.7118, + "theoretical_loss": 4.046464800613121, + "tokens_seen": 394657792 + }, + { + "epoch": 0.12, + "learning_rate": 0.0007114302908726179, + "loss": 1.525, + "theoretical_loss": 4.046389991580294, + "tokens_seen": 394723328 + }, + { + "epoch": 0.12, + "learning_rate": 0.0007114142427281846, + "loss": 1.5365, + "theoretical_loss": 4.046315198444094, + "tokens_seen": 394788864 + }, + { + "epoch": 0.12, + "learning_rate": 0.0007113981945837513, + "loss": 1.5933, + "theoretical_loss": 4.046240421198506, + "tokens_seen": 394854400 + }, + { + "epoch": 0.12, + "learning_rate": 0.000711382146439318, + "loss": 1.6788, + "theoretical_loss": 4.046165659837517, + "tokens_seen": 394919936 + }, + { + "epoch": 0.12, + "learning_rate": 0.0007113660982948847, + "loss": 1.6568, + "theoretical_loss": 4.046090914355117, + "tokens_seen": 394985472 + }, + { + "epoch": 0.12, + "learning_rate": 0.0007113500501504514, + "loss": 1.6441, + "theoretical_loss": 4.046016184745302, + "tokens_seen": 395051008 + }, + { + "epoch": 0.12, + "learning_rate": 0.0007113340020060181, + "loss": 1.6496, + "theoretical_loss": 4.045941471002066, + "tokens_seen": 395116544 + }, + { + "epoch": 0.12, + "learning_rate": 0.0007113179538615848, + "loss": 1.673, + "theoretical_loss": 4.0458667731194105, + "tokens_seen": 395182080 + }, + { + "epoch": 0.12, + "learning_rate": 0.0007113019057171515, + "loss": 1.6358, + "theoretical_loss": 4.045792091091339, + "tokens_seen": 395247616 + }, + { + "epoch": 0.12, + "learning_rate": 0.0007112858575727182, + "loss": 1.6653, + "theoretical_loss": 4.04571742491186, + "tokens_seen": 395313152 + }, + { + "epoch": 0.12, + "learning_rate": 0.0007112698094282849, + "loss": 1.571, + "theoretical_loss": 4.045642774574981, + "tokens_seen": 395378688 + }, + { + "epoch": 0.12, + "learning_rate": 0.0007112537612838516, + "loss": 1.6429, + "theoretical_loss": 4.045568140074717, + "tokens_seen": 395444224 + }, + { + "epoch": 0.12, + "learning_rate": 0.0007112377131394183, + "loss": 1.5375, + "theoretical_loss": 4.045493521405085, + "tokens_seen": 395509760 + }, + { + "epoch": 0.12, + "learning_rate": 0.000711221664994985, + "loss": 1.7246, + "theoretical_loss": 4.045418918560104, + "tokens_seen": 395575296 + }, + { + "epoch": 0.12, + "learning_rate": 0.0007112056168505517, + "loss": 1.55, + "theoretical_loss": 4.045344331533796, + "tokens_seen": 395640832 + }, + { + "epoch": 0.12, + "learning_rate": 0.0007111895687061184, + "loss": 1.7051, + "theoretical_loss": 4.04526976032019, + "tokens_seen": 395706368 + }, + { + "epoch": 0.12, + "learning_rate": 0.0007111735205616851, + "loss": 1.6806, + "theoretical_loss": 4.0451952049133135, + "tokens_seen": 395771904 + }, + { + "epoch": 0.12, + "learning_rate": 0.0007111574724172518, + "loss": 1.6992, + "theoretical_loss": 4.0451206653072, + "tokens_seen": 395837440 + }, + { + "epoch": 0.12, + "learning_rate": 0.0007111414242728185, + "loss": 1.5695, + "theoretical_loss": 4.045046141495887, + "tokens_seen": 395902976 + }, + { + "epoch": 0.12, + "learning_rate": 0.0007111253761283852, + "loss": 1.6266, + "theoretical_loss": 4.044971633473412, + "tokens_seen": 395968512 + }, + { + "epoch": 0.12, + "learning_rate": 0.000711109327983952, + "loss": 1.5558, + "theoretical_loss": 4.044897141233818, + "tokens_seen": 396034048 + }, + { + "epoch": 0.12, + "learning_rate": 0.0007110932798395186, + "loss": 1.6267, + "theoretical_loss": 4.044822664771151, + "tokens_seen": 396099584 + }, + { + "epoch": 0.12, + "learning_rate": 0.0007110772316950854, + "loss": 1.5786, + "theoretical_loss": 4.044748204079459, + "tokens_seen": 396165120 + }, + { + "epoch": 0.12, + "learning_rate": 0.000711061183550652, + "loss": 1.6764, + "theoretical_loss": 4.044673759152797, + "tokens_seen": 396230656 + }, + { + "epoch": 0.12, + "learning_rate": 0.0007110451354062188, + "loss": 1.7233, + "theoretical_loss": 4.0445993299852185, + "tokens_seen": 396296192 + }, + { + "epoch": 0.12, + "learning_rate": 0.0007110290872617854, + "loss": 1.7225, + "theoretical_loss": 4.044524916570782, + "tokens_seen": 396361728 + }, + { + "epoch": 0.12, + "learning_rate": 0.0007110130391173522, + "loss": 1.7308, + "theoretical_loss": 4.04445051890355, + "tokens_seen": 396427264 + }, + { + "epoch": 0.12, + "objective/train/docs_used": 150450, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.601449966430664, + "objective/train/theoretical_loss": 4.044376136977588, + "objective/train/tokens_used": 416952800, + "theoretical_loss": 4.044376136977588, + "tokens_seen": 396492800 + }, + { + "epoch": 0.12, + "learning_rate": 0.0007109969909729188, + "loss": 1.5784, + "theoretical_loss": 4.044376136977588, + "tokens_seen": 396492800 + }, + { + "epoch": 0.12, + "learning_rate": 0.0007109809428284855, + "loss": 1.6137, + "theoretical_loss": 4.044301770786964, + "tokens_seen": 396558336 + }, + { + "epoch": 0.12, + "learning_rate": 0.0007109648946840522, + "loss": 1.6572, + "theoretical_loss": 4.044227420325749, + "tokens_seen": 396623872 + }, + { + "epoch": 0.12, + "learning_rate": 0.0007109488465396189, + "loss": 1.7489, + "theoretical_loss": 4.044153085588018, + "tokens_seen": 396689408 + }, + { + "epoch": 0.12, + "learning_rate": 0.0007109327983951856, + "loss": 1.5549, + "theoretical_loss": 4.044078766567851, + "tokens_seen": 396754944 + }, + { + "epoch": 0.12, + "learning_rate": 0.0007109167502507523, + "loss": 1.676, + "theoretical_loss": 4.044004463259327, + "tokens_seen": 396820480 + }, + { + "epoch": 0.12, + "learning_rate": 0.000710900702106319, + "loss": 1.5399, + "theoretical_loss": 4.043930175656531, + "tokens_seen": 396886016 + }, + { + "epoch": 0.12, + "learning_rate": 0.0007108846539618857, + "loss": 1.4907, + "theoretical_loss": 4.043855903753551, + "tokens_seen": 396951552 + }, + { + "epoch": 0.12, + "learning_rate": 0.0007108686058174524, + "loss": 1.4993, + "theoretical_loss": 4.043781647544477, + "tokens_seen": 397017088 + }, + { + "epoch": 0.12, + "learning_rate": 0.0007108525576730191, + "loss": 1.6906, + "theoretical_loss": 4.043707407023404, + "tokens_seen": 397082624 + }, + { + "epoch": 0.12, + "learning_rate": 0.0007108365095285857, + "loss": 1.6795, + "theoretical_loss": 4.043633182184428, + "tokens_seen": 397148160 + }, + { + "epoch": 0.12, + "learning_rate": 0.0007108204613841525, + "loss": 1.5641, + "theoretical_loss": 4.043558973021652, + "tokens_seen": 397213696 + }, + { + "epoch": 0.12, + "learning_rate": 0.0007108044132397191, + "loss": 1.5855, + "theoretical_loss": 4.043484779529176, + "tokens_seen": 397279232 + }, + { + "epoch": 0.12, + "learning_rate": 0.0007107883650952859, + "loss": 1.723, + "theoretical_loss": 4.04341060170111, + "tokens_seen": 397344768 + }, + { + "epoch": 0.12, + "learning_rate": 0.0007107723169508525, + "loss": 1.7374, + "theoretical_loss": 4.043336439531561, + "tokens_seen": 397410304 + }, + { + "epoch": 0.12, + "learning_rate": 0.0007107562688064193, + "loss": 1.6873, + "theoretical_loss": 4.043262293014646, + "tokens_seen": 397475840 + }, + { + "epoch": 0.12, + "learning_rate": 0.0007107402206619859, + "loss": 1.694, + "theoretical_loss": 4.043188162144478, + "tokens_seen": 397541376 + }, + { + "epoch": 0.12, + "learning_rate": 0.0007107241725175527, + "loss": 1.7009, + "theoretical_loss": 4.0431140469151785, + "tokens_seen": 397606912 + }, + { + "epoch": 0.12, + "learning_rate": 0.0007107081243731193, + "loss": 1.6454, + "theoretical_loss": 4.043039947320869, + "tokens_seen": 397672448 + }, + { + "epoch": 0.12, + "learning_rate": 0.0007106920762286861, + "loss": 1.7645, + "theoretical_loss": 4.0429658633556755, + "tokens_seen": 397737984 + }, + { + "epoch": 0.12, + "learning_rate": 0.0007106760280842527, + "loss": 1.6959, + "theoretical_loss": 4.042891795013729, + "tokens_seen": 397803520 + }, + { + "epoch": 0.12, + "learning_rate": 0.0007106599799398195, + "loss": 1.6798, + "theoretical_loss": 4.042817742289159, + "tokens_seen": 397869056 + }, + { + "epoch": 0.12, + "learning_rate": 0.0007106439317953861, + "loss": 1.6653, + "theoretical_loss": 4.042743705176102, + "tokens_seen": 397934592 + }, + { + "epoch": 0.12, + "learning_rate": 0.0007106278836509529, + "loss": 1.6226, + "theoretical_loss": 4.042669683668698, + "tokens_seen": 398000128 + }, + { + "epoch": 0.12, + "learning_rate": 0.0007106118355065195, + "loss": 1.6647, + "theoretical_loss": 4.042595677761087, + "tokens_seen": 398065664 + }, + { + "epoch": 0.12, + "learning_rate": 0.0007105957873620863, + "loss": 1.5075, + "theoretical_loss": 4.0425216874474135, + "tokens_seen": 398131200 + }, + { + "epoch": 0.12, + "learning_rate": 0.0007105797392176529, + "loss": 1.6271, + "theoretical_loss": 4.0424477127218275, + "tokens_seen": 398196736 + }, + { + "epoch": 0.12, + "learning_rate": 0.0007105636910732197, + "loss": 1.7519, + "theoretical_loss": 4.042373753578479, + "tokens_seen": 398262272 + }, + { + "epoch": 0.12, + "learning_rate": 0.0007105476429287863, + "loss": 1.7084, + "theoretical_loss": 4.042299810011523, + "tokens_seen": 398327808 + }, + { + "epoch": 0.12, + "learning_rate": 0.0007105315947843531, + "loss": 1.7561, + "theoretical_loss": 4.042225882015115, + "tokens_seen": 398393344 + }, + { + "epoch": 0.12, + "learning_rate": 0.0007105155466399197, + "loss": 1.619, + "theoretical_loss": 4.0421519695834185, + "tokens_seen": 398458880 + }, + { + "epoch": 0.12, + "learning_rate": 0.0007104994984954865, + "loss": 1.5026, + "theoretical_loss": 4.042078072710597, + "tokens_seen": 398524416 + }, + { + "epoch": 0.12, + "learning_rate": 0.0007104834503510531, + "loss": 1.542, + "theoretical_loss": 4.042004191390816, + "tokens_seen": 398589952 + }, + { + "epoch": 0.12, + "learning_rate": 0.0007104674022066199, + "loss": 1.6451, + "theoretical_loss": 4.041930325618246, + "tokens_seen": 398655488 + }, + { + "epoch": 0.12, + "learning_rate": 0.0007104513540621865, + "loss": 1.5798, + "theoretical_loss": 4.041856475387061, + "tokens_seen": 398721024 + }, + { + "epoch": 0.12, + "learning_rate": 0.0007104353059177533, + "loss": 1.6246, + "theoretical_loss": 4.041782640691437, + "tokens_seen": 398786560 + }, + { + "epoch": 0.12, + "learning_rate": 0.0007104192577733199, + "loss": 1.6691, + "theoretical_loss": 4.041708821525554, + "tokens_seen": 398852096 + }, + { + "epoch": 0.12, + "learning_rate": 0.0007104032096288867, + "loss": 1.6616, + "theoretical_loss": 4.0416350178835945, + "tokens_seen": 398917632 + }, + { + "epoch": 0.12, + "learning_rate": 0.0007103871614844533, + "loss": 1.543, + "theoretical_loss": 4.041561229759743, + "tokens_seen": 398983168 + }, + { + "epoch": 0.12, + "learning_rate": 0.0007103711133400201, + "loss": 1.6808, + "theoretical_loss": 4.0414874571481905, + "tokens_seen": 399048704 + }, + { + "epoch": 0.12, + "learning_rate": 0.0007103550651955867, + "loss": 1.6037, + "theoretical_loss": 4.041413700043128, + "tokens_seen": 399114240 + }, + { + "epoch": 0.12, + "learning_rate": 0.0007103390170511535, + "loss": 1.6796, + "theoretical_loss": 4.041339958438751, + "tokens_seen": 399179776 + }, + { + "epoch": 0.12, + "learning_rate": 0.0007103229689067201, + "loss": 1.7243, + "theoretical_loss": 4.041266232329257, + "tokens_seen": 399245312 + }, + { + "epoch": 0.12, + "learning_rate": 0.0007103069207622869, + "loss": 1.7109, + "theoretical_loss": 4.041192521708849, + "tokens_seen": 399310848 + }, + { + "epoch": 0.12, + "learning_rate": 0.0007102908726178535, + "loss": 1.6256, + "theoretical_loss": 4.041118826571731, + "tokens_seen": 399376384 + }, + { + "epoch": 0.12, + "learning_rate": 0.0007102748244734203, + "loss": 1.6153, + "theoretical_loss": 4.041045146912111, + "tokens_seen": 399441920 + }, + { + "epoch": 0.12, + "learning_rate": 0.000710258776328987, + "loss": 1.6004, + "theoretical_loss": 4.040971482724199, + "tokens_seen": 399507456 + }, + { + "epoch": 0.12, + "learning_rate": 0.0007102427281845537, + "loss": 1.5374, + "theoretical_loss": 4.040897834002209, + "tokens_seen": 399572992 + }, + { + "epoch": 0.12, + "learning_rate": 0.0007102266800401204, + "loss": 1.6883, + "theoretical_loss": 4.040824200740358, + "tokens_seen": 399638528 + }, + { + "epoch": 0.12, + "learning_rate": 0.0007102106318956871, + "loss": 1.6011, + "theoretical_loss": 4.040750582932868, + "tokens_seen": 399704064 + }, + { + "epoch": 0.12, + "objective/train/docs_used": 151634, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.6212098598480225, + "objective/train/theoretical_loss": 4.04067698057396, + "objective/train/tokens_used": 420229600, + "theoretical_loss": 4.04067698057396, + "tokens_seen": 399769600 + }, + { + "epoch": 0.12, + "learning_rate": 0.0007101945837512538, + "loss": 1.5262, + "theoretical_loss": 4.04067698057396, + "tokens_seen": 399769600 + }, + { + "epoch": 0.12, + "learning_rate": 0.0007101785356068205, + "loss": 1.5166, + "theoretical_loss": 4.040603393657862, + "tokens_seen": 399835136 + }, + { + "epoch": 0.12, + "learning_rate": 0.0007101624874623872, + "loss": 1.5989, + "theoretical_loss": 4.040529822178802, + "tokens_seen": 399900672 + }, + { + "epoch": 0.12, + "learning_rate": 0.0007101464393179539, + "loss": 1.5597, + "theoretical_loss": 4.040456266131013, + "tokens_seen": 399966208 + }, + { + "epoch": 0.12, + "learning_rate": 0.0007101303911735206, + "loss": 1.6047, + "theoretical_loss": 4.0403827255087315, + "tokens_seen": 400031744 + }, + { + "epoch": 0.12, + "learning_rate": 0.0007101143430290873, + "loss": 1.5452, + "theoretical_loss": 4.040309200306195, + "tokens_seen": 400097280 + }, + { + "epoch": 0.12, + "learning_rate": 0.000710098294884654, + "loss": 1.6335, + "theoretical_loss": 4.0402356905176475, + "tokens_seen": 400162816 + }, + { + "epoch": 0.12, + "learning_rate": 0.0007100822467402207, + "loss": 1.6821, + "theoretical_loss": 4.040162196137332, + "tokens_seen": 400228352 + }, + { + "epoch": 0.12, + "learning_rate": 0.0007100661985957874, + "loss": 1.6577, + "theoretical_loss": 4.040088717159497, + "tokens_seen": 400293888 + }, + { + "epoch": 0.12, + "learning_rate": 0.0007100501504513541, + "loss": 1.5389, + "theoretical_loss": 4.040015253578394, + "tokens_seen": 400359424 + }, + { + "epoch": 0.12, + "learning_rate": 0.0007100341023069208, + "loss": 1.5889, + "theoretical_loss": 4.039941805388277, + "tokens_seen": 400424960 + }, + { + "epoch": 0.12, + "learning_rate": 0.0007100180541624875, + "loss": 1.5804, + "theoretical_loss": 4.0398683725834035, + "tokens_seen": 400490496 + }, + { + "epoch": 0.12, + "learning_rate": 0.0007100020060180542, + "loss": 1.5105, + "theoretical_loss": 4.039794955158033, + "tokens_seen": 400556032 + }, + { + "epoch": 0.12, + "learning_rate": 0.0007099859578736209, + "loss": 1.581, + "theoretical_loss": 4.039721553106431, + "tokens_seen": 400621568 + }, + { + "epoch": 0.12, + "learning_rate": 0.0007099699097291876, + "loss": 1.7176, + "theoretical_loss": 4.039648166422863, + "tokens_seen": 400687104 + }, + { + "epoch": 0.12, + "learning_rate": 0.0007099538615847543, + "loss": 1.6494, + "theoretical_loss": 4.039574795101597, + "tokens_seen": 400752640 + }, + { + "epoch": 0.12, + "learning_rate": 0.000709937813440321, + "loss": 1.6576, + "theoretical_loss": 4.039501439136908, + "tokens_seen": 400818176 + }, + { + "epoch": 0.12, + "learning_rate": 0.0007099217652958877, + "loss": 1.6935, + "theoretical_loss": 4.0394280985230715, + "tokens_seen": 400883712 + }, + { + "epoch": 0.12, + "learning_rate": 0.0007099057171514545, + "loss": 1.6463, + "theoretical_loss": 4.039354773254365, + "tokens_seen": 400949248 + }, + { + "epoch": 0.12, + "learning_rate": 0.0007098896690070211, + "loss": 1.5844, + "theoretical_loss": 4.0392814633250715, + "tokens_seen": 401014784 + }, + { + "epoch": 0.12, + "learning_rate": 0.0007098736208625879, + "loss": 1.6337, + "theoretical_loss": 4.039208168729475, + "tokens_seen": 401080320 + }, + { + "epoch": 0.12, + "learning_rate": 0.0007098575727181545, + "loss": 1.6503, + "theoretical_loss": 4.039134889461866, + "tokens_seen": 401145856 + }, + { + "epoch": 0.12, + "learning_rate": 0.0007098415245737213, + "loss": 1.4616, + "theoretical_loss": 4.039061625516532, + "tokens_seen": 401211392 + }, + { + "epoch": 0.12, + "learning_rate": 0.0007098254764292879, + "loss": 1.6464, + "theoretical_loss": 4.038988376887771, + "tokens_seen": 401276928 + }, + { + "epoch": 0.12, + "learning_rate": 0.0007098094282848547, + "loss": 1.6111, + "theoretical_loss": 4.038915143569876, + "tokens_seen": 401342464 + }, + { + "epoch": 0.12, + "learning_rate": 0.0007097933801404213, + "loss": 1.577, + "theoretical_loss": 4.038841925557151, + "tokens_seen": 401408000 + }, + { + "epoch": 0.12, + "learning_rate": 0.0007097773319959881, + "loss": 1.572, + "theoretical_loss": 4.038768722843898, + "tokens_seen": 401473536 + }, + { + "epoch": 0.12, + "learning_rate": 0.0007097612838515547, + "loss": 1.5907, + "theoretical_loss": 4.038695535424422, + "tokens_seen": 401539072 + }, + { + "epoch": 0.12, + "learning_rate": 0.0007097452357071215, + "loss": 1.6304, + "theoretical_loss": 4.038622363293034, + "tokens_seen": 401604608 + }, + { + "epoch": 0.12, + "learning_rate": 0.0007097291875626881, + "loss": 1.554, + "theoretical_loss": 4.0385492064440465, + "tokens_seen": 401670144 + }, + { + "epoch": 0.12, + "learning_rate": 0.0007097131394182549, + "loss": 1.5914, + "theoretical_loss": 4.038476064871775, + "tokens_seen": 401735680 + }, + { + "epoch": 0.12, + "learning_rate": 0.0007096970912738215, + "loss": 1.6395, + "theoretical_loss": 4.0384029385705364, + "tokens_seen": 401801216 + }, + { + "epoch": 0.12, + "learning_rate": 0.0007096810431293883, + "loss": 1.4266, + "theoretical_loss": 4.038329827534654, + "tokens_seen": 401866752 + }, + { + "epoch": 0.12, + "learning_rate": 0.0007096649949849549, + "loss": 1.6691, + "theoretical_loss": 4.038256731758453, + "tokens_seen": 401932288 + }, + { + "epoch": 0.12, + "learning_rate": 0.0007096489468405217, + "loss": 1.5544, + "theoretical_loss": 4.038183651236259, + "tokens_seen": 401997824 + }, + { + "epoch": 0.12, + "learning_rate": 0.0007096328986960883, + "loss": 1.6727, + "theoretical_loss": 4.038110585962404, + "tokens_seen": 402063360 + }, + { + "epoch": 0.12, + "learning_rate": 0.0007096168505516551, + "loss": 1.5538, + "theoretical_loss": 4.038037535931221, + "tokens_seen": 402128896 + }, + { + "epoch": 0.12, + "learning_rate": 0.0007096008024072217, + "loss": 1.6007, + "theoretical_loss": 4.037964501137048, + "tokens_seen": 402194432 + }, + { + "epoch": 0.12, + "learning_rate": 0.0007095847542627884, + "loss": 1.5492, + "theoretical_loss": 4.037891481574224, + "tokens_seen": 402259968 + }, + { + "epoch": 0.12, + "learning_rate": 0.0007095687061183551, + "loss": 1.758, + "theoretical_loss": 4.037818477237091, + "tokens_seen": 402325504 + }, + { + "epoch": 0.12, + "learning_rate": 0.0007095526579739218, + "loss": 1.5858, + "theoretical_loss": 4.037745488119996, + "tokens_seen": 402391040 + }, + { + "epoch": 0.12, + "learning_rate": 0.0007095366098294885, + "loss": 1.5809, + "theoretical_loss": 4.037672514217288, + "tokens_seen": 402456576 + }, + { + "epoch": 0.12, + "learning_rate": 0.0007095205616850552, + "loss": 1.4728, + "theoretical_loss": 4.037599555523318, + "tokens_seen": 402522112 + }, + { + "epoch": 0.12, + "learning_rate": 0.0007095045135406219, + "loss": 1.6019, + "theoretical_loss": 4.037526612032441, + "tokens_seen": 402587648 + }, + { + "epoch": 0.12, + "learning_rate": 0.0007094884653961886, + "loss": 1.5991, + "theoretical_loss": 4.037453683739015, + "tokens_seen": 402653184 + }, + { + "epoch": 0.12, + "learning_rate": 0.0007094724172517553, + "loss": 1.6656, + "theoretical_loss": 4.037380770637402, + "tokens_seen": 402718720 + }, + { + "epoch": 0.12, + "learning_rate": 0.000709456369107322, + "loss": 1.595, + "theoretical_loss": 4.037307872721964, + "tokens_seen": 402784256 + }, + { + "epoch": 0.12, + "learning_rate": 0.0007094403209628886, + "loss": 1.6497, + "theoretical_loss": 4.037234989987069, + "tokens_seen": 402849792 + }, + { + "epoch": 0.12, + "learning_rate": 0.0007094242728184554, + "loss": 1.5669, + "theoretical_loss": 4.0371621224270875, + "tokens_seen": 402915328 + }, + { + "epoch": 0.12, + "learning_rate": 0.000709408224674022, + "loss": 1.6929, + "theoretical_loss": 4.037089270036391, + "tokens_seen": 402980864 + }, + { + "epoch": 0.12, + "objective/train/docs_used": 152845, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.5569102764129639, + "objective/train/theoretical_loss": 4.037016432809357, + "objective/train/tokens_used": 423506400, + "theoretical_loss": 4.037016432809357, + "tokens_seen": 403046400 + }, + { + "epoch": 0.12, + "learning_rate": 0.0007093921765295888, + "loss": 1.6328, + "theoretical_loss": 4.037016432809357, + "tokens_seen": 403046400 + }, + { + "epoch": 0.12, + "learning_rate": 0.0007093761283851554, + "loss": 1.6968, + "theoretical_loss": 4.036943610740362, + "tokens_seen": 403111936 + }, + { + "epoch": 0.12, + "learning_rate": 0.0007093600802407222, + "loss": 1.6074, + "theoretical_loss": 4.036870803823791, + "tokens_seen": 403177472 + }, + { + "epoch": 0.12, + "learning_rate": 0.0007093440320962888, + "loss": 1.6269, + "theoretical_loss": 4.036798012054027, + "tokens_seen": 403243008 + }, + { + "epoch": 0.12, + "learning_rate": 0.0007093279839518556, + "loss": 1.5427, + "theoretical_loss": 4.036725235425458, + "tokens_seen": 403308544 + }, + { + "epoch": 0.12, + "learning_rate": 0.0007093119358074222, + "loss": 1.5832, + "theoretical_loss": 4.0366524739324765, + "tokens_seen": 403374080 + }, + { + "epoch": 0.12, + "learning_rate": 0.000709295887662989, + "loss": 1.6499, + "theoretical_loss": 4.036579727569475, + "tokens_seen": 403439616 + }, + { + "epoch": 0.12, + "learning_rate": 0.0007092798395185556, + "loss": 1.6512, + "theoretical_loss": 4.036506996330849, + "tokens_seen": 403505152 + }, + { + "epoch": 0.12, + "learning_rate": 0.0007092637913741224, + "loss": 1.6159, + "theoretical_loss": 4.036434280211001, + "tokens_seen": 403570688 + }, + { + "epoch": 0.12, + "learning_rate": 0.000709247743229689, + "loss": 1.451, + "theoretical_loss": 4.036361579204334, + "tokens_seen": 403636224 + }, + { + "epoch": 0.12, + "learning_rate": 0.0007092316950852558, + "loss": 1.5882, + "theoretical_loss": 4.036288893305251, + "tokens_seen": 403701760 + }, + { + "epoch": 0.12, + "learning_rate": 0.0007092156469408224, + "loss": 1.6166, + "theoretical_loss": 4.036216222508163, + "tokens_seen": 403767296 + }, + { + "epoch": 0.12, + "learning_rate": 0.0007091995987963892, + "loss": 1.629, + "theoretical_loss": 4.036143566807481, + "tokens_seen": 403832832 + }, + { + "epoch": 0.12, + "learning_rate": 0.0007091835506519558, + "loss": 1.6551, + "theoretical_loss": 4.03607092619762, + "tokens_seen": 403898368 + }, + { + "epoch": 0.12, + "learning_rate": 0.0007091675025075226, + "loss": 1.6806, + "theoretical_loss": 4.035998300672999, + "tokens_seen": 403963904 + }, + { + "epoch": 0.12, + "learning_rate": 0.0007091514543630892, + "loss": 1.5229, + "theoretical_loss": 4.035925690228035, + "tokens_seen": 404029440 + }, + { + "epoch": 0.12, + "learning_rate": 0.000709135406218656, + "loss": 1.609, + "theoretical_loss": 4.035853094857156, + "tokens_seen": 404094976 + }, + { + "epoch": 0.12, + "learning_rate": 0.0007091193580742226, + "loss": 1.5738, + "theoretical_loss": 4.035780514554787, + "tokens_seen": 404160512 + }, + { + "epoch": 0.12, + "learning_rate": 0.0007091033099297894, + "loss": 1.651, + "theoretical_loss": 4.0357079493153565, + "tokens_seen": 404226048 + }, + { + "epoch": 0.12, + "learning_rate": 0.000709087261785356, + "loss": 1.6074, + "theoretical_loss": 4.035635399133298, + "tokens_seen": 404291584 + }, + { + "epoch": 0.12, + "learning_rate": 0.0007090712136409228, + "loss": 1.5971, + "theoretical_loss": 4.035562864003047, + "tokens_seen": 404357120 + }, + { + "epoch": 0.12, + "learning_rate": 0.0007090551654964894, + "loss": 1.629, + "theoretical_loss": 4.035490343919042, + "tokens_seen": 404422656 + }, + { + "epoch": 0.12, + "learning_rate": 0.0007090391173520562, + "loss": 1.5606, + "theoretical_loss": 4.035417838875725, + "tokens_seen": 404488192 + }, + { + "epoch": 0.12, + "learning_rate": 0.0007090230692076228, + "loss": 1.6295, + "theoretical_loss": 4.035345348867539, + "tokens_seen": 404553728 + }, + { + "epoch": 0.12, + "learning_rate": 0.0007090070210631896, + "loss": 1.5969, + "theoretical_loss": 4.035272873888932, + "tokens_seen": 404619264 + }, + { + "epoch": 0.12, + "learning_rate": 0.0007089909729187562, + "loss": 1.7329, + "theoretical_loss": 4.035200413934355, + "tokens_seen": 404684800 + }, + { + "epoch": 0.12, + "learning_rate": 0.000708974924774323, + "loss": 1.6531, + "theoretical_loss": 4.03512796899826, + "tokens_seen": 404750336 + }, + { + "epoch": 0.12, + "learning_rate": 0.0007089588766298896, + "loss": 1.4877, + "theoretical_loss": 4.035055539075104, + "tokens_seen": 404815872 + }, + { + "epoch": 0.12, + "learning_rate": 0.0007089428284854564, + "loss": 1.5553, + "theoretical_loss": 4.034983124159346, + "tokens_seen": 404881408 + }, + { + "epoch": 0.12, + "learning_rate": 0.000708926780341023, + "loss": 1.5935, + "theoretical_loss": 4.0349107242454485, + "tokens_seen": 404946944 + }, + { + "epoch": 0.12, + "learning_rate": 0.0007089107321965898, + "loss": 1.6346, + "theoretical_loss": 4.034838339327876, + "tokens_seen": 405012480 + }, + { + "epoch": 0.12, + "learning_rate": 0.0007088946840521565, + "loss": 1.5903, + "theoretical_loss": 4.034765969401096, + "tokens_seen": 405078016 + }, + { + "epoch": 0.12, + "learning_rate": 0.0007088786359077232, + "loss": 1.6103, + "theoretical_loss": 4.03469361445958, + "tokens_seen": 405143552 + }, + { + "epoch": 0.12, + "learning_rate": 0.0007088625877632899, + "loss": 1.5606, + "theoretical_loss": 4.034621274497802, + "tokens_seen": 405209088 + }, + { + "epoch": 0.12, + "learning_rate": 0.0007088465396188566, + "loss": 1.6416, + "theoretical_loss": 4.0345489495102385, + "tokens_seen": 405274624 + }, + { + "epoch": 0.12, + "learning_rate": 0.0007088304914744233, + "loss": 1.7138, + "theoretical_loss": 4.034476639491368, + "tokens_seen": 405340160 + }, + { + "epoch": 0.12, + "learning_rate": 0.00070881444332999, + "loss": 1.655, + "theoretical_loss": 4.034404344435676, + "tokens_seen": 405405696 + }, + { + "epoch": 0.12, + "learning_rate": 0.0007087983951855567, + "loss": 1.6378, + "theoretical_loss": 4.0343320643376455, + "tokens_seen": 405471232 + }, + { + "epoch": 0.12, + "learning_rate": 0.0007087823470411234, + "loss": 1.8518, + "theoretical_loss": 4.034259799191765, + "tokens_seen": 405536768 + }, + { + "epoch": 0.12, + "learning_rate": 0.0007087662988966901, + "loss": 1.5796, + "theoretical_loss": 4.034187548992527, + "tokens_seen": 405602304 + }, + { + "epoch": 0.12, + "learning_rate": 0.0007087502507522568, + "loss": 1.6143, + "theoretical_loss": 4.034115313734426, + "tokens_seen": 405667840 + }, + { + "epoch": 0.12, + "learning_rate": 0.0007087342026078235, + "loss": 1.6456, + "theoretical_loss": 4.034043093411958, + "tokens_seen": 405733376 + }, + { + "epoch": 0.12, + "learning_rate": 0.0007087181544633902, + "loss": 1.5634, + "theoretical_loss": 4.033970888019624, + "tokens_seen": 405798912 + }, + { + "epoch": 0.12, + "learning_rate": 0.000708702106318957, + "loss": 1.5407, + "theoretical_loss": 4.033898697551926, + "tokens_seen": 405864448 + }, + { + "epoch": 0.12, + "learning_rate": 0.0007086860581745236, + "loss": 1.5393, + "theoretical_loss": 4.0338265220033716, + "tokens_seen": 405929984 + }, + { + "epoch": 0.12, + "learning_rate": 0.0007086700100300904, + "loss": 1.7124, + "theoretical_loss": 4.0337543613684685, + "tokens_seen": 405995520 + }, + { + "epoch": 0.12, + "learning_rate": 0.000708653961885657, + "loss": 1.5351, + "theoretical_loss": 4.033682215641729, + "tokens_seen": 406061056 + }, + { + "epoch": 0.12, + "learning_rate": 0.0007086379137412238, + "loss": 1.499, + "theoretical_loss": 4.033610084817667, + "tokens_seen": 406126592 + }, + { + "epoch": 0.12, + "learning_rate": 0.0007086218655967904, + "loss": 1.6338, + "theoretical_loss": 4.0335379688908, + "tokens_seen": 406192128 + }, + { + "epoch": 0.12, + "learning_rate": 0.0007086058174523572, + "loss": 1.6419, + "theoretical_loss": 4.03346586785565, + "tokens_seen": 406257664 + }, + { + "epoch": 0.12, + "objective/train/docs_used": 154148, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.514775276184082, + "objective/train/theoretical_loss": 4.03339378170674, + "objective/train/tokens_used": 426783200, + "theoretical_loss": 4.03339378170674, + "tokens_seen": 406323200 + }, + { + "epoch": 0.12, + "learning_rate": 0.0007085897693079238, + "loss": 1.5378, + "theoretical_loss": 4.03339378170674, + "tokens_seen": 406323200 + }, + { + "epoch": 0.12, + "learning_rate": 0.0007085737211634906, + "loss": 1.5028, + "theoretical_loss": 4.033321710438594, + "tokens_seen": 406388736 + }, + { + "epoch": 0.12, + "learning_rate": 0.0007085576730190572, + "loss": 1.5153, + "theoretical_loss": 4.033249654045743, + "tokens_seen": 406454272 + }, + { + "epoch": 0.12, + "learning_rate": 0.000708541624874624, + "loss": 1.595, + "theoretical_loss": 4.033177612522721, + "tokens_seen": 406519808 + }, + { + "epoch": 0.12, + "learning_rate": 0.0007085255767301906, + "loss": 1.3989, + "theoretical_loss": 4.0331055858640585, + "tokens_seen": 406585344 + }, + { + "epoch": 0.12, + "learning_rate": 0.0007085095285857574, + "loss": 1.6113, + "theoretical_loss": 4.033033574064297, + "tokens_seen": 406650880 + }, + { + "epoch": 0.12, + "learning_rate": 0.000708493480441324, + "loss": 1.6242, + "theoretical_loss": 4.032961577117976, + "tokens_seen": 406716416 + }, + { + "epoch": 0.12, + "learning_rate": 0.0007084774322968908, + "loss": 1.6357, + "theoretical_loss": 4.032889595019638, + "tokens_seen": 406781952 + }, + { + "epoch": 0.12, + "learning_rate": 0.0007084613841524574, + "loss": 1.7266, + "theoretical_loss": 4.032817627763832, + "tokens_seen": 406847488 + }, + { + "epoch": 0.12, + "learning_rate": 0.0007084453360080242, + "loss": 1.6791, + "theoretical_loss": 4.0327456753451045, + "tokens_seen": 406913024 + }, + { + "epoch": 0.12, + "learning_rate": 0.0007084292878635908, + "loss": 1.6751, + "theoretical_loss": 4.03267373775801, + "tokens_seen": 406978560 + }, + { + "epoch": 0.12, + "learning_rate": 0.0007084132397191576, + "loss": 1.5817, + "theoretical_loss": 4.032601814997102, + "tokens_seen": 407044096 + }, + { + "epoch": 0.12, + "learning_rate": 0.0007083971915747242, + "loss": 1.6017, + "theoretical_loss": 4.03252990705694, + "tokens_seen": 407109632 + }, + { + "epoch": 0.12, + "learning_rate": 0.000708381143430291, + "loss": 1.6537, + "theoretical_loss": 4.032458013932084, + "tokens_seen": 407175168 + }, + { + "epoch": 0.12, + "learning_rate": 0.0007083650952858576, + "loss": 1.5524, + "theoretical_loss": 4.032386135617098, + "tokens_seen": 407240704 + }, + { + "epoch": 0.12, + "learning_rate": 0.0007083490471414244, + "loss": 1.7485, + "theoretical_loss": 4.032314272106548, + "tokens_seen": 407306240 + }, + { + "epoch": 0.12, + "learning_rate": 0.000708332998996991, + "loss": 1.6018, + "theoretical_loss": 4.0322424233950045, + "tokens_seen": 407371776 + }, + { + "epoch": 0.12, + "learning_rate": 0.0007083169508525578, + "loss": 1.6425, + "theoretical_loss": 4.032170589477039, + "tokens_seen": 407437312 + }, + { + "epoch": 0.12, + "learning_rate": 0.0007083009027081244, + "loss": 1.6946, + "theoretical_loss": 4.032098770347228, + "tokens_seen": 407502848 + }, + { + "epoch": 0.12, + "learning_rate": 0.0007082848545636912, + "loss": 1.5392, + "theoretical_loss": 4.0320269660001475, + "tokens_seen": 407568384 + }, + { + "epoch": 0.12, + "learning_rate": 0.0007082688064192578, + "loss": 1.6876, + "theoretical_loss": 4.0319551764303805, + "tokens_seen": 407633920 + }, + { + "epoch": 0.12, + "learning_rate": 0.0007082527582748246, + "loss": 1.6066, + "theoretical_loss": 4.03188340163251, + "tokens_seen": 407699456 + }, + { + "epoch": 0.12, + "learning_rate": 0.0007082367101303912, + "loss": 1.6974, + "theoretical_loss": 4.031811641601122, + "tokens_seen": 407764992 + }, + { + "epoch": 0.12, + "learning_rate": 0.000708220661985958, + "loss": 1.4493, + "theoretical_loss": 4.031739896330807, + "tokens_seen": 407830528 + }, + { + "epoch": 0.12, + "learning_rate": 0.0007082046138415246, + "loss": 1.6327, + "theoretical_loss": 4.031668165816156, + "tokens_seen": 407896064 + }, + { + "epoch": 0.12, + "learning_rate": 0.0007081885656970914, + "loss": 1.6554, + "theoretical_loss": 4.031596450051766, + "tokens_seen": 407961600 + }, + { + "epoch": 0.12, + "learning_rate": 0.000708172517552658, + "loss": 1.5332, + "theoretical_loss": 4.031524749032234, + "tokens_seen": 408027136 + }, + { + "epoch": 0.12, + "learning_rate": 0.0007081564694082247, + "loss": 1.6898, + "theoretical_loss": 4.031453062752161, + "tokens_seen": 408092672 + }, + { + "epoch": 0.12, + "learning_rate": 0.0007081404212637914, + "loss": 1.503, + "theoretical_loss": 4.031381391206151, + "tokens_seen": 408158208 + }, + { + "epoch": 0.12, + "learning_rate": 0.0007081243731193581, + "loss": 1.5192, + "theoretical_loss": 4.03130973438881, + "tokens_seen": 408223744 + }, + { + "epoch": 0.12, + "learning_rate": 0.0007081083249749248, + "loss": 1.6628, + "theoretical_loss": 4.031238092294748, + "tokens_seen": 408289280 + }, + { + "epoch": 0.12, + "learning_rate": 0.0007080922768304915, + "loss": 1.7408, + "theoretical_loss": 4.031166464918576, + "tokens_seen": 408354816 + }, + { + "epoch": 0.12, + "learning_rate": 0.0007080762286860582, + "loss": 1.601, + "theoretical_loss": 4.0310948522549115, + "tokens_seen": 408420352 + }, + { + "epoch": 0.12, + "learning_rate": 0.0007080601805416249, + "loss": 1.6952, + "theoretical_loss": 4.031023254298371, + "tokens_seen": 408485888 + }, + { + "epoch": 0.12, + "learning_rate": 0.0007080441323971915, + "loss": 1.5139, + "theoretical_loss": 4.030951671043574, + "tokens_seen": 408551424 + }, + { + "epoch": 0.12, + "learning_rate": 0.0007080280842527583, + "loss": 1.6573, + "theoretical_loss": 4.030880102485147, + "tokens_seen": 408616960 + }, + { + "epoch": 0.12, + "learning_rate": 0.0007080120361083249, + "loss": 1.5823, + "theoretical_loss": 4.030808548617714, + "tokens_seen": 408682496 + }, + { + "epoch": 0.12, + "learning_rate": 0.0007079959879638917, + "loss": 1.549, + "theoretical_loss": 4.030737009435906, + "tokens_seen": 408748032 + }, + { + "epoch": 0.12, + "learning_rate": 0.0007079799398194583, + "loss": 1.5875, + "theoretical_loss": 4.030665484934353, + "tokens_seen": 408813568 + }, + { + "epoch": 0.12, + "learning_rate": 0.0007079638916750251, + "loss": 1.6287, + "theoretical_loss": 4.030593975107692, + "tokens_seen": 408879104 + }, + { + "epoch": 0.12, + "learning_rate": 0.0007079478435305917, + "loss": 1.5254, + "theoretical_loss": 4.03052247995056, + "tokens_seen": 408944640 + }, + { + "epoch": 0.12, + "learning_rate": 0.0007079317953861585, + "loss": 1.5741, + "theoretical_loss": 4.030450999457596, + "tokens_seen": 409010176 + }, + { + "epoch": 0.12, + "learning_rate": 0.0007079157472417251, + "loss": 1.5952, + "theoretical_loss": 4.030379533623445, + "tokens_seen": 409075712 + }, + { + "epoch": 0.12, + "learning_rate": 0.0007078996990972919, + "loss": 1.594, + "theoretical_loss": 4.030308082442753, + "tokens_seen": 409141248 + }, + { + "epoch": 0.12, + "learning_rate": 0.0007078836509528585, + "loss": 1.5202, + "theoretical_loss": 4.030236645910169, + "tokens_seen": 409206784 + }, + { + "epoch": 0.12, + "learning_rate": 0.0007078676028084253, + "loss": 1.6089, + "theoretical_loss": 4.030165224020345, + "tokens_seen": 409272320 + }, + { + "epoch": 0.12, + "learning_rate": 0.0007078515546639919, + "loss": 1.5635, + "theoretical_loss": 4.030093816767934, + "tokens_seen": 409337856 + }, + { + "epoch": 0.12, + "learning_rate": 0.0007078355065195587, + "loss": 1.7094, + "theoretical_loss": 4.030022424147595, + "tokens_seen": 409403392 + }, + { + "epoch": 0.12, + "learning_rate": 0.0007078194583751253, + "loss": 1.5637, + "theoretical_loss": 4.029951046153988, + "tokens_seen": 409468928 + }, + { + "epoch": 0.12, + "learning_rate": 0.0007078034102306921, + "loss": 1.4649, + "theoretical_loss": 4.029879682781776, + "tokens_seen": 409534464 + }, + { + "epoch": 0.12, + "objective/train/docs_used": 155224, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.7365703582763672, + "objective/train/theoretical_loss": 4.029808334025624, + "objective/train/tokens_used": 430060000, + "theoretical_loss": 4.029808334025624, + "tokens_seen": 409600000 + }, + { + "epoch": 0.12, + "learning_rate": 0.0007077873620862587, + "loss": 1.6135, + "theoretical_loss": 4.029808334025624, + "tokens_seen": 409600000 + }, + { + "epoch": 0.12, + "learning_rate": 0.0007077713139418255, + "loss": 1.4604, + "theoretical_loss": 4.0297369998802015, + "tokens_seen": 409665536 + }, + { + "epoch": 0.12, + "learning_rate": 0.0007077552657973921, + "loss": 1.5184, + "theoretical_loss": 4.0296656803401785, + "tokens_seen": 409731072 + }, + { + "epoch": 0.12, + "learning_rate": 0.0007077392176529589, + "loss": 1.6725, + "theoretical_loss": 4.029594375400231, + "tokens_seen": 409796608 + }, + { + "epoch": 0.12, + "learning_rate": 0.0007077231695085255, + "loss": 1.4153, + "theoretical_loss": 4.029523085055036, + "tokens_seen": 409862144 + }, + { + "epoch": 0.12, + "learning_rate": 0.0007077071213640923, + "loss": 1.5621, + "theoretical_loss": 4.029451809299273, + "tokens_seen": 409927680 + }, + { + "epoch": 0.12, + "learning_rate": 0.0007076910732196589, + "loss": 1.6994, + "theoretical_loss": 4.029380548127622, + "tokens_seen": 409993216 + }, + { + "epoch": 0.12, + "learning_rate": 0.0007076750250752257, + "loss": 1.6012, + "theoretical_loss": 4.029309301534772, + "tokens_seen": 410058752 + }, + { + "epoch": 0.12, + "learning_rate": 0.0007076589769307923, + "loss": 1.5408, + "theoretical_loss": 4.02923806951541, + "tokens_seen": 410124288 + }, + { + "epoch": 0.12, + "learning_rate": 0.0007076429287863591, + "loss": 1.5696, + "theoretical_loss": 4.029166852064225, + "tokens_seen": 410189824 + }, + { + "epoch": 0.12, + "learning_rate": 0.0007076268806419257, + "loss": 1.6951, + "theoretical_loss": 4.029095649175913, + "tokens_seen": 410255360 + }, + { + "epoch": 0.12, + "learning_rate": 0.0007076108324974925, + "loss": 1.6481, + "theoretical_loss": 4.02902446084517, + "tokens_seen": 410320896 + }, + { + "epoch": 0.12, + "learning_rate": 0.0007075947843530591, + "loss": 1.6343, + "theoretical_loss": 4.028953287066695, + "tokens_seen": 410386432 + }, + { + "epoch": 0.12, + "learning_rate": 0.0007075787362086259, + "loss": 1.5999, + "theoretical_loss": 4.02888212783519, + "tokens_seen": 410451968 + }, + { + "epoch": 0.12, + "learning_rate": 0.0007075626880641925, + "loss": 1.6068, + "theoretical_loss": 4.028810983145361, + "tokens_seen": 410517504 + }, + { + "epoch": 0.12, + "learning_rate": 0.0007075466399197593, + "loss": 1.7041, + "theoretical_loss": 4.0287398529919125, + "tokens_seen": 410583040 + }, + { + "epoch": 0.12, + "learning_rate": 0.000707530591775326, + "loss": 1.4612, + "theoretical_loss": 4.028668737369558, + "tokens_seen": 410648576 + }, + { + "epoch": 0.12, + "learning_rate": 0.0007075145436308927, + "loss": 1.6032, + "theoretical_loss": 4.028597636273009, + "tokens_seen": 410714112 + }, + { + "epoch": 0.12, + "learning_rate": 0.0007074984954864595, + "loss": 1.5767, + "theoretical_loss": 4.028526549696982, + "tokens_seen": 410779648 + }, + { + "epoch": 0.12, + "learning_rate": 0.0007074824473420261, + "loss": 1.5887, + "theoretical_loss": 4.0284554776361965, + "tokens_seen": 410845184 + }, + { + "epoch": 0.12, + "learning_rate": 0.0007074663991975929, + "loss": 1.5724, + "theoretical_loss": 4.028384420085372, + "tokens_seen": 410910720 + }, + { + "epoch": 0.12, + "learning_rate": 0.0007074503510531595, + "loss": 1.5949, + "theoretical_loss": 4.028313377039234, + "tokens_seen": 410976256 + }, + { + "epoch": 0.12, + "learning_rate": 0.0007074343029087263, + "loss": 1.6207, + "theoretical_loss": 4.028242348492509, + "tokens_seen": 411041792 + }, + { + "epoch": 0.12, + "learning_rate": 0.0007074182547642929, + "loss": 1.601, + "theoretical_loss": 4.028171334439927, + "tokens_seen": 411107328 + }, + { + "epoch": 0.12, + "learning_rate": 0.0007074022066198597, + "loss": 1.5212, + "theoretical_loss": 4.02810033487622, + "tokens_seen": 411172864 + }, + { + "epoch": 0.12, + "learning_rate": 0.0007073861584754263, + "loss": 1.7169, + "theoretical_loss": 4.028029349796125, + "tokens_seen": 411238400 + }, + { + "epoch": 0.12, + "learning_rate": 0.0007073701103309931, + "loss": 1.5529, + "theoretical_loss": 4.027958379194376, + "tokens_seen": 411303936 + }, + { + "epoch": 0.12, + "learning_rate": 0.0007073540621865597, + "loss": 1.6632, + "theoretical_loss": 4.027887423065717, + "tokens_seen": 411369472 + }, + { + "epoch": 0.12, + "learning_rate": 0.0007073380140421265, + "loss": 1.5867, + "theoretical_loss": 4.027816481404892, + "tokens_seen": 411435008 + }, + { + "epoch": 0.12, + "learning_rate": 0.0007073219658976931, + "loss": 1.5276, + "theoretical_loss": 4.0277455542066445, + "tokens_seen": 411500544 + }, + { + "epoch": 0.12, + "learning_rate": 0.0007073059177532599, + "loss": 1.5809, + "theoretical_loss": 4.027674641465725, + "tokens_seen": 411566080 + }, + { + "epoch": 0.12, + "learning_rate": 0.0007072898696088265, + "loss": 1.7333, + "theoretical_loss": 4.027603743176885, + "tokens_seen": 411631616 + }, + { + "epoch": 0.12, + "learning_rate": 0.0007072738214643933, + "loss": 1.5822, + "theoretical_loss": 4.02753285933488, + "tokens_seen": 411697152 + }, + { + "epoch": 0.12, + "learning_rate": 0.0007072577733199599, + "loss": 1.7061, + "theoretical_loss": 4.027461989934466, + "tokens_seen": 411762688 + }, + { + "epoch": 0.12, + "learning_rate": 0.0007072417251755267, + "loss": 1.5907, + "theoretical_loss": 4.027391134970402, + "tokens_seen": 411828224 + }, + { + "epoch": 0.12, + "learning_rate": 0.0007072256770310933, + "loss": 1.5025, + "theoretical_loss": 4.027320294437452, + "tokens_seen": 411893760 + }, + { + "epoch": 0.12, + "learning_rate": 0.0007072096288866601, + "loss": 1.7353, + "theoretical_loss": 4.027249468330381, + "tokens_seen": 411959296 + }, + { + "epoch": 0.12, + "learning_rate": 0.0007071935807422267, + "loss": 1.7476, + "theoretical_loss": 4.027178656643957, + "tokens_seen": 412024832 + }, + { + "epoch": 0.12, + "learning_rate": 0.0007071775325977935, + "loss": 1.6339, + "theoretical_loss": 4.0271078593729515, + "tokens_seen": 412090368 + }, + { + "epoch": 0.12, + "learning_rate": 0.0007071614844533601, + "loss": 1.6337, + "theoretical_loss": 4.027037076512137, + "tokens_seen": 412155904 + }, + { + "epoch": 0.12, + "learning_rate": 0.0007071454363089269, + "loss": 1.6017, + "theoretical_loss": 4.026966308056291, + "tokens_seen": 412221440 + }, + { + "epoch": 0.12, + "learning_rate": 0.0007071293881644935, + "loss": 1.5405, + "theoretical_loss": 4.026895554000192, + "tokens_seen": 412286976 + }, + { + "epoch": 0.12, + "learning_rate": 0.0007071133400200603, + "loss": 1.6012, + "theoretical_loss": 4.026824814338621, + "tokens_seen": 412352512 + }, + { + "epoch": 0.12, + "learning_rate": 0.0007070972918756269, + "loss": 1.6489, + "theoretical_loss": 4.026754089066364, + "tokens_seen": 412418048 + }, + { + "epoch": 0.12, + "learning_rate": 0.0007070812437311937, + "loss": 1.5208, + "theoretical_loss": 4.026683378178206, + "tokens_seen": 412483584 + }, + { + "epoch": 0.13, + "learning_rate": 0.0007070651955867603, + "loss": 1.6966, + "theoretical_loss": 4.026612681668939, + "tokens_seen": 412549120 + }, + { + "epoch": 0.13, + "learning_rate": 0.0007070491474423271, + "loss": 1.5919, + "theoretical_loss": 4.026541999533355, + "tokens_seen": 412614656 + }, + { + "epoch": 0.13, + "learning_rate": 0.0007070330992978937, + "loss": 1.5842, + "theoretical_loss": 4.026471331766249, + "tokens_seen": 412680192 + }, + { + "epoch": 0.13, + "learning_rate": 0.0007070170511534605, + "loss": 1.7416, + "theoretical_loss": 4.026400678362418, + "tokens_seen": 412745728 + }, + { + "epoch": 0.13, + "learning_rate": 0.0007070010030090271, + "loss": 1.5071, + "theoretical_loss": 4.026330039316665, + "tokens_seen": 412811264 + }, + { + "epoch": 0.13, + "objective/train/docs_used": 156351, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.5006463527679443, + "objective/train/theoretical_loss": 4.026259414623792, + "objective/train/tokens_used": 433336800, + "theoretical_loss": 4.026259414623792, + "tokens_seen": 412876800 + }, + { + "epoch": 0.13, + "learning_rate": 0.0007069849548645939, + "loss": 1.5176, + "theoretical_loss": 4.026259414623792, + "tokens_seen": 412876800 + }, + { + "epoch": 0.13, + "learning_rate": 0.0007069689067201605, + "loss": 1.6805, + "theoretical_loss": 4.026188804278605, + "tokens_seen": 412942336 + }, + { + "epoch": 0.13, + "learning_rate": 0.0007069528585757273, + "loss": 1.5223, + "theoretical_loss": 4.026118208275914, + "tokens_seen": 413007872 + }, + { + "epoch": 0.13, + "learning_rate": 0.0007069368104312939, + "loss": 1.7066, + "theoretical_loss": 4.02604762661053, + "tokens_seen": 413073408 + }, + { + "epoch": 0.13, + "learning_rate": 0.0007069207622868607, + "loss": 1.5287, + "theoretical_loss": 4.025977059277267, + "tokens_seen": 413138944 + }, + { + "epoch": 0.13, + "learning_rate": 0.0007069047141424273, + "loss": 1.5652, + "theoretical_loss": 4.025906506270941, + "tokens_seen": 413204480 + }, + { + "epoch": 0.13, + "learning_rate": 0.0007068886659979941, + "loss": 1.5945, + "theoretical_loss": 4.0258359675863735, + "tokens_seen": 413270016 + }, + { + "epoch": 0.13, + "learning_rate": 0.0007068726178535607, + "loss": 1.565, + "theoretical_loss": 4.025765443218386, + "tokens_seen": 413335552 + }, + { + "epoch": 0.13, + "learning_rate": 0.0007068565697091275, + "loss": 1.6855, + "theoretical_loss": 4.025694933161804, + "tokens_seen": 413401088 + }, + { + "epoch": 0.13, + "learning_rate": 0.0007068405215646941, + "loss": 1.5446, + "theoretical_loss": 4.025624437411455, + "tokens_seen": 413466624 + }, + { + "epoch": 0.13, + "learning_rate": 0.0007068244734202609, + "loss": 1.5223, + "theoretical_loss": 4.0255539559621685, + "tokens_seen": 413532160 + }, + { + "epoch": 0.13, + "learning_rate": 0.0007068084252758275, + "loss": 1.5301, + "theoretical_loss": 4.025483488808779, + "tokens_seen": 413597696 + }, + { + "epoch": 0.13, + "learning_rate": 0.0007067923771313943, + "loss": 1.5843, + "theoretical_loss": 4.025413035946121, + "tokens_seen": 413663232 + }, + { + "epoch": 0.13, + "learning_rate": 0.0007067763289869609, + "loss": 1.6872, + "theoretical_loss": 4.025342597369034, + "tokens_seen": 413728768 + }, + { + "epoch": 0.13, + "learning_rate": 0.0007067602808425276, + "loss": 1.5548, + "theoretical_loss": 4.025272173072359, + "tokens_seen": 413794304 + }, + { + "epoch": 0.13, + "learning_rate": 0.0007067442326980943, + "loss": 1.5162, + "theoretical_loss": 4.025201763050939, + "tokens_seen": 413859840 + }, + { + "epoch": 0.13, + "learning_rate": 0.000706728184553661, + "loss": 1.6179, + "theoretical_loss": 4.025131367299621, + "tokens_seen": 413925376 + }, + { + "epoch": 0.13, + "learning_rate": 0.0007067121364092277, + "loss": 1.4543, + "theoretical_loss": 4.025060985813255, + "tokens_seen": 413990912 + }, + { + "epoch": 0.13, + "learning_rate": 0.0007066960882647944, + "loss": 1.5457, + "theoretical_loss": 4.024990618586692, + "tokens_seen": 414056448 + }, + { + "epoch": 0.13, + "learning_rate": 0.0007066800401203611, + "loss": 1.4967, + "theoretical_loss": 4.024920265614787, + "tokens_seen": 414121984 + }, + { + "epoch": 0.13, + "learning_rate": 0.0007066639919759278, + "loss": 1.6441, + "theoretical_loss": 4.024849926892396, + "tokens_seen": 414187520 + }, + { + "epoch": 0.13, + "learning_rate": 0.0007066479438314945, + "loss": 1.6333, + "theoretical_loss": 4.02477960241438, + "tokens_seen": 414253056 + }, + { + "epoch": 0.13, + "learning_rate": 0.0007066318956870612, + "loss": 1.5529, + "theoretical_loss": 4.024709292175602, + "tokens_seen": 414318592 + }, + { + "epoch": 0.13, + "learning_rate": 0.0007066158475426278, + "loss": 1.5744, + "theoretical_loss": 4.024638996170927, + "tokens_seen": 414384128 + }, + { + "epoch": 0.13, + "learning_rate": 0.0007065997993981946, + "loss": 1.6354, + "theoretical_loss": 4.024568714395221, + "tokens_seen": 414449664 + }, + { + "epoch": 0.13, + "learning_rate": 0.0007065837512537612, + "loss": 1.4886, + "theoretical_loss": 4.0244984468433564, + "tokens_seen": 414515200 + }, + { + "epoch": 0.13, + "learning_rate": 0.000706567703109328, + "loss": 1.5999, + "theoretical_loss": 4.0244281935102055, + "tokens_seen": 414580736 + }, + { + "epoch": 0.13, + "learning_rate": 0.0007065516549648946, + "loss": 1.6236, + "theoretical_loss": 4.024357954390645, + "tokens_seen": 414646272 + }, + { + "epoch": 0.13, + "learning_rate": 0.0007065356068204614, + "loss": 1.6086, + "theoretical_loss": 4.0242877294795525, + "tokens_seen": 414711808 + }, + { + "epoch": 0.13, + "learning_rate": 0.000706519558676028, + "loss": 1.5781, + "theoretical_loss": 4.0242175187718106, + "tokens_seen": 414777344 + }, + { + "epoch": 0.13, + "learning_rate": 0.0007065035105315948, + "loss": 1.574, + "theoretical_loss": 4.024147322262301, + "tokens_seen": 414842880 + }, + { + "epoch": 0.13, + "learning_rate": 0.0007064874623871614, + "loss": 1.7366, + "theoretical_loss": 4.024077139945912, + "tokens_seen": 414908416 + }, + { + "epoch": 0.13, + "learning_rate": 0.0007064714142427282, + "loss": 1.674, + "theoretical_loss": 4.024006971817532, + "tokens_seen": 414973952 + }, + { + "epoch": 0.13, + "learning_rate": 0.0007064553660982948, + "loss": 1.5642, + "theoretical_loss": 4.023936817872053, + "tokens_seen": 415039488 + }, + { + "epoch": 0.13, + "learning_rate": 0.0007064393179538616, + "loss": 1.6421, + "theoretical_loss": 4.023866678104368, + "tokens_seen": 415105024 + }, + { + "epoch": 0.13, + "learning_rate": 0.0007064232698094282, + "loss": 1.5611, + "theoretical_loss": 4.023796552509376, + "tokens_seen": 415170560 + }, + { + "epoch": 0.13, + "learning_rate": 0.000706407221664995, + "loss": 1.3985, + "theoretical_loss": 4.023726441081976, + "tokens_seen": 415236096 + }, + { + "epoch": 0.13, + "learning_rate": 0.0007063911735205616, + "loss": 1.429, + "theoretical_loss": 4.02365634381707, + "tokens_seen": 415301632 + }, + { + "epoch": 0.13, + "learning_rate": 0.0007063751253761284, + "loss": 1.5868, + "theoretical_loss": 4.023586260709564, + "tokens_seen": 415367168 + }, + { + "epoch": 0.13, + "learning_rate": 0.000706359077231695, + "loss": 1.5575, + "theoretical_loss": 4.023516191754363, + "tokens_seen": 415432704 + }, + { + "epoch": 0.13, + "learning_rate": 0.0007063430290872618, + "loss": 1.5799, + "theoretical_loss": 4.02344613694638, + "tokens_seen": 415498240 + }, + { + "epoch": 0.13, + "learning_rate": 0.0007063269809428284, + "loss": 1.5591, + "theoretical_loss": 4.023376096280527, + "tokens_seen": 415563776 + }, + { + "epoch": 0.13, + "learning_rate": 0.0007063109327983952, + "loss": 1.4712, + "theoretical_loss": 4.02330606975172, + "tokens_seen": 415629312 + }, + { + "epoch": 0.13, + "learning_rate": 0.000706294884653962, + "loss": 1.6896, + "theoretical_loss": 4.023236057354875, + "tokens_seen": 415694848 + }, + { + "epoch": 0.13, + "learning_rate": 0.0007062788365095286, + "loss": 1.5865, + "theoretical_loss": 4.023166059084916, + "tokens_seen": 415760384 + }, + { + "epoch": 0.13, + "learning_rate": 0.0007062627883650954, + "loss": 1.6741, + "theoretical_loss": 4.023096074936764, + "tokens_seen": 415825920 + }, + { + "epoch": 0.13, + "learning_rate": 0.000706246740220662, + "loss": 1.4672, + "theoretical_loss": 4.0230261049053455, + "tokens_seen": 415891456 + }, + { + "epoch": 0.13, + "learning_rate": 0.0007062306920762288, + "loss": 1.596, + "theoretical_loss": 4.022956148985589, + "tokens_seen": 415956992 + }, + { + "epoch": 0.13, + "learning_rate": 0.0007062146439317954, + "loss": 1.5769, + "theoretical_loss": 4.0228862071724265, + "tokens_seen": 416022528 + }, + { + "epoch": 0.13, + "learning_rate": 0.0007061985957873622, + "loss": 1.5949, + "theoretical_loss": 4.022816279460791, + "tokens_seen": 416088064 + }, + { + "epoch": 0.13, + "objective/train/docs_used": 157543, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.4583311080932617, + "objective/train/theoretical_loss": 4.02274636584562, + "objective/train/tokens_used": 436613600, + "theoretical_loss": 4.02274636584562, + "tokens_seen": 416153600 + }, + { + "epoch": 0.13, + "learning_rate": 0.0007061825476429288, + "loss": 1.5807, + "theoretical_loss": 4.02274636584562, + "tokens_seen": 416153600 + }, + { + "epoch": 0.13, + "learning_rate": 0.0007061664994984956, + "loss": 1.59, + "theoretical_loss": 4.022676466321853, + "tokens_seen": 416219136 + }, + { + "epoch": 0.13, + "learning_rate": 0.0007061504513540622, + "loss": 1.7119, + "theoretical_loss": 4.022606580884429, + "tokens_seen": 416284672 + }, + { + "epoch": 0.13, + "learning_rate": 0.000706134403209629, + "loss": 1.677, + "theoretical_loss": 4.0225367095282945, + "tokens_seen": 416350208 + }, + { + "epoch": 0.13, + "learning_rate": 0.0007061183550651956, + "loss": 1.6853, + "theoretical_loss": 4.022466852248396, + "tokens_seen": 416415744 + }, + { + "epoch": 0.13, + "learning_rate": 0.0007061023069207624, + "loss": 1.5115, + "theoretical_loss": 4.0223970090396834, + "tokens_seen": 416481280 + }, + { + "epoch": 0.13, + "learning_rate": 0.000706086258776329, + "loss": 1.5972, + "theoretical_loss": 4.0223271798971085, + "tokens_seen": 416546816 + }, + { + "epoch": 0.13, + "learning_rate": 0.0007060702106318958, + "loss": 1.5403, + "theoretical_loss": 4.0222573648156255, + "tokens_seen": 416612352 + }, + { + "epoch": 0.13, + "learning_rate": 0.0007060541624874624, + "loss": 1.6859, + "theoretical_loss": 4.022187563790193, + "tokens_seen": 416677888 + }, + { + "epoch": 0.13, + "learning_rate": 0.0007060381143430292, + "loss": 1.5491, + "theoretical_loss": 4.02211777681577, + "tokens_seen": 416743424 + }, + { + "epoch": 0.13, + "learning_rate": 0.0007060220661985958, + "loss": 1.7026, + "theoretical_loss": 4.022048003887319, + "tokens_seen": 416808960 + }, + { + "epoch": 0.13, + "learning_rate": 0.0007060060180541626, + "loss": 1.5109, + "theoretical_loss": 4.021978244999806, + "tokens_seen": 416874496 + }, + { + "epoch": 0.13, + "learning_rate": 0.0007059899699097292, + "loss": 1.7336, + "theoretical_loss": 4.021908500148198, + "tokens_seen": 416940032 + }, + { + "epoch": 0.13, + "learning_rate": 0.000705973921765296, + "loss": 1.6229, + "theoretical_loss": 4.0218387693274655, + "tokens_seen": 417005568 + }, + { + "epoch": 0.13, + "learning_rate": 0.0007059578736208626, + "loss": 1.4647, + "theoretical_loss": 4.021769052532582, + "tokens_seen": 417071104 + }, + { + "epoch": 0.13, + "learning_rate": 0.0007059418254764294, + "loss": 1.5181, + "theoretical_loss": 4.0216993497585225, + "tokens_seen": 417136640 + }, + { + "epoch": 0.13, + "learning_rate": 0.000705925777331996, + "loss": 1.6888, + "theoretical_loss": 4.021629661000265, + "tokens_seen": 417202176 + }, + { + "epoch": 0.13, + "learning_rate": 0.0007059097291875628, + "loss": 1.6051, + "theoretical_loss": 4.02155998625279, + "tokens_seen": 417267712 + }, + { + "epoch": 0.13, + "learning_rate": 0.0007058936810431294, + "loss": 1.6575, + "theoretical_loss": 4.021490325511082, + "tokens_seen": 417333248 + }, + { + "epoch": 0.13, + "learning_rate": 0.0007058776328986962, + "loss": 1.5971, + "theoretical_loss": 4.021420678770125, + "tokens_seen": 417398784 + }, + { + "epoch": 0.13, + "learning_rate": 0.0007058615847542628, + "loss": 1.6366, + "theoretical_loss": 4.02135104602491, + "tokens_seen": 417464320 + }, + { + "epoch": 0.13, + "learning_rate": 0.0007058455366098296, + "loss": 1.5655, + "theoretical_loss": 4.021281427270425, + "tokens_seen": 417529856 + }, + { + "epoch": 0.13, + "learning_rate": 0.0007058294884653962, + "loss": 1.6452, + "theoretical_loss": 4.021211822501666, + "tokens_seen": 417595392 + }, + { + "epoch": 0.13, + "learning_rate": 0.000705813440320963, + "loss": 1.5535, + "theoretical_loss": 4.021142231713629, + "tokens_seen": 417660928 + }, + { + "epoch": 0.13, + "learning_rate": 0.0007057973921765296, + "loss": 1.5929, + "theoretical_loss": 4.021072654901312, + "tokens_seen": 417726464 + }, + { + "epoch": 0.13, + "learning_rate": 0.0007057813440320964, + "loss": 1.4786, + "theoretical_loss": 4.021003092059716, + "tokens_seen": 417792000 + }, + { + "epoch": 0.13, + "learning_rate": 0.000705765295887663, + "loss": 1.5367, + "theoretical_loss": 4.020933543183845, + "tokens_seen": 417857536 + }, + { + "epoch": 0.13, + "learning_rate": 0.0007057492477432298, + "loss": 1.5904, + "theoretical_loss": 4.020864008268706, + "tokens_seen": 417923072 + }, + { + "epoch": 0.13, + "learning_rate": 0.0007057331995987964, + "loss": 1.4239, + "theoretical_loss": 4.020794487309308, + "tokens_seen": 417988608 + }, + { + "epoch": 0.13, + "learning_rate": 0.0007057171514543632, + "loss": 1.5887, + "theoretical_loss": 4.020724980300663, + "tokens_seen": 418054144 + }, + { + "epoch": 0.13, + "learning_rate": 0.0007057011033099298, + "loss": 1.685, + "theoretical_loss": 4.020655487237785, + "tokens_seen": 418119680 + }, + { + "epoch": 0.13, + "learning_rate": 0.0007056850551654966, + "loss": 1.5687, + "theoretical_loss": 4.02058600811569, + "tokens_seen": 418185216 + }, + { + "epoch": 0.13, + "learning_rate": 0.0007056690070210632, + "loss": 1.6325, + "theoretical_loss": 4.020516542929398, + "tokens_seen": 418250752 + }, + { + "epoch": 0.13, + "learning_rate": 0.00070565295887663, + "loss": 1.4804, + "theoretical_loss": 4.0204470916739306, + "tokens_seen": 418316288 + }, + { + "epoch": 0.13, + "learning_rate": 0.0007056369107321966, + "loss": 1.5728, + "theoretical_loss": 4.020377654344312, + "tokens_seen": 418381824 + }, + { + "epoch": 0.13, + "learning_rate": 0.0007056208625877634, + "loss": 1.684, + "theoretical_loss": 4.020308230935569, + "tokens_seen": 418447360 + }, + { + "epoch": 0.13, + "learning_rate": 0.00070560481444333, + "loss": 1.7115, + "theoretical_loss": 4.020238821442733, + "tokens_seen": 418512896 + }, + { + "epoch": 0.13, + "learning_rate": 0.0007055887662988968, + "loss": 1.6821, + "theoretical_loss": 4.020169425860834, + "tokens_seen": 418578432 + }, + { + "epoch": 0.13, + "learning_rate": 0.0007055727181544634, + "loss": 1.641, + "theoretical_loss": 4.0201000441849075, + "tokens_seen": 418643968 + }, + { + "epoch": 0.13, + "learning_rate": 0.0007055566700100302, + "loss": 1.6932, + "theoretical_loss": 4.0200306764099905, + "tokens_seen": 418709504 + }, + { + "epoch": 0.13, + "learning_rate": 0.0007055406218655968, + "loss": 1.6363, + "theoretical_loss": 4.019961322531122, + "tokens_seen": 418775040 + }, + { + "epoch": 0.13, + "learning_rate": 0.0007055245737211636, + "loss": 1.6081, + "theoretical_loss": 4.019891982543346, + "tokens_seen": 418840576 + }, + { + "epoch": 0.13, + "learning_rate": 0.0007055085255767302, + "loss": 1.5927, + "theoretical_loss": 4.019822656441706, + "tokens_seen": 418906112 + }, + { + "epoch": 0.13, + "learning_rate": 0.000705492477432297, + "loss": 1.5338, + "theoretical_loss": 4.01975334422125, + "tokens_seen": 418971648 + }, + { + "epoch": 0.13, + "learning_rate": 0.0007054764292878636, + "loss": 1.6015, + "theoretical_loss": 4.019684045877027, + "tokens_seen": 419037184 + }, + { + "epoch": 0.13, + "learning_rate": 0.0007054603811434304, + "loss": 1.4542, + "theoretical_loss": 4.019614761404091, + "tokens_seen": 419102720 + }, + { + "epoch": 0.13, + "learning_rate": 0.000705444332998997, + "loss": 1.6329, + "theoretical_loss": 4.019545490797495, + "tokens_seen": 419168256 + }, + { + "epoch": 0.13, + "learning_rate": 0.0007054282848545638, + "loss": 1.5724, + "theoretical_loss": 4.0194762340522985, + "tokens_seen": 419233792 + }, + { + "epoch": 0.13, + "learning_rate": 0.0007054122367101304, + "loss": 1.5817, + "theoretical_loss": 4.01940699116356, + "tokens_seen": 419299328 + }, + { + "epoch": 0.13, + "learning_rate": 0.0007053961885656972, + "loss": 1.6894, + "theoretical_loss": 4.019337762126343, + "tokens_seen": 419364864 + }, + { + "epoch": 0.13, + "objective/train/docs_used": 158751, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.7412222623825073, + "objective/train/theoretical_loss": 4.019268546935713, + "objective/train/tokens_used": 439890400, + "theoretical_loss": 4.019268546935713, + "tokens_seen": 419430400 + }, + { + "epoch": 0.13, + "learning_rate": 0.0007053801404212638, + "loss": 1.6031, + "theoretical_loss": 4.019268546935713, + "tokens_seen": 419430400 + }, + { + "epoch": 0.13, + "learning_rate": 0.0007053640922768305, + "loss": 1.5945, + "theoretical_loss": 4.0191993455867365, + "tokens_seen": 419495936 + }, + { + "epoch": 0.13, + "learning_rate": 0.0007053480441323972, + "loss": 1.7252, + "theoretical_loss": 4.019130158074484, + "tokens_seen": 419561472 + }, + { + "epoch": 0.13, + "learning_rate": 0.0007053319959879639, + "loss": 1.6649, + "theoretical_loss": 4.019060984394028, + "tokens_seen": 419627008 + }, + { + "epoch": 0.13, + "learning_rate": 0.0007053159478435306, + "loss": 1.6739, + "theoretical_loss": 4.018991824540445, + "tokens_seen": 419692544 + }, + { + "epoch": 0.13, + "learning_rate": 0.0007052998996990973, + "loss": 1.6434, + "theoretical_loss": 4.018922678508812, + "tokens_seen": 419758080 + }, + { + "epoch": 0.13, + "learning_rate": 0.000705283851554664, + "loss": 1.6291, + "theoretical_loss": 4.018853546294208, + "tokens_seen": 419823616 + }, + { + "epoch": 0.13, + "learning_rate": 0.0007052678034102307, + "loss": 1.5991, + "theoretical_loss": 4.018784427891718, + "tokens_seen": 419889152 + }, + { + "epoch": 0.13, + "learning_rate": 0.0007052517552657974, + "loss": 1.6365, + "theoretical_loss": 4.018715323296425, + "tokens_seen": 419954688 + }, + { + "epoch": 0.13, + "learning_rate": 0.0007052357071213641, + "loss": 1.488, + "theoretical_loss": 4.018646232503419, + "tokens_seen": 420020224 + }, + { + "epoch": 0.13, + "learning_rate": 0.0007052196589769307, + "loss": 1.4395, + "theoretical_loss": 4.0185771555077885, + "tokens_seen": 420085760 + }, + { + "epoch": 0.13, + "learning_rate": 0.0007052036108324975, + "loss": 1.5856, + "theoretical_loss": 4.018508092304628, + "tokens_seen": 420151296 + }, + { + "epoch": 0.13, + "learning_rate": 0.0007051875626880641, + "loss": 1.6128, + "theoretical_loss": 4.018439042889032, + "tokens_seen": 420216832 + }, + { + "epoch": 0.13, + "learning_rate": 0.0007051715145436309, + "loss": 1.6515, + "theoretical_loss": 4.018370007256098, + "tokens_seen": 420282368 + }, + { + "epoch": 0.13, + "learning_rate": 0.0007051554663991975, + "loss": 1.6632, + "theoretical_loss": 4.018300985400927, + "tokens_seen": 420347904 + }, + { + "epoch": 0.13, + "learning_rate": 0.0007051394182547643, + "loss": 1.6788, + "theoretical_loss": 4.018231977318622, + "tokens_seen": 420413440 + }, + { + "epoch": 0.13, + "learning_rate": 0.0007051233701103309, + "loss": 1.6501, + "theoretical_loss": 4.018162983004287, + "tokens_seen": 420478976 + }, + { + "epoch": 0.13, + "learning_rate": 0.0007051073219658977, + "loss": 1.5045, + "theoretical_loss": 4.018094002453031, + "tokens_seen": 420544512 + }, + { + "epoch": 0.13, + "learning_rate": 0.0007050912738214644, + "loss": 1.6517, + "theoretical_loss": 4.018025035659965, + "tokens_seen": 420610048 + }, + { + "epoch": 0.13, + "learning_rate": 0.0007050752256770311, + "loss": 1.7025, + "theoretical_loss": 4.017956082620201, + "tokens_seen": 420675584 + }, + { + "epoch": 0.13, + "learning_rate": 0.0007050591775325978, + "loss": 1.637, + "theoretical_loss": 4.0178871433288545, + "tokens_seen": 420741120 + }, + { + "epoch": 0.13, + "learning_rate": 0.0007050431293881645, + "loss": 1.6995, + "theoretical_loss": 4.017818217781044, + "tokens_seen": 420806656 + }, + { + "epoch": 0.13, + "learning_rate": 0.0007050270812437312, + "loss": 1.5914, + "theoretical_loss": 4.017749305971889, + "tokens_seen": 420872192 + }, + { + "epoch": 0.13, + "learning_rate": 0.0007050110330992979, + "loss": 1.5281, + "theoretical_loss": 4.0176804078965125, + "tokens_seen": 420937728 + }, + { + "epoch": 0.13, + "learning_rate": 0.0007049949849548647, + "loss": 1.6262, + "theoretical_loss": 4.017611523550039, + "tokens_seen": 421003264 + }, + { + "epoch": 0.13, + "learning_rate": 0.0007049789368104313, + "loss": 1.6659, + "theoretical_loss": 4.017542652927599, + "tokens_seen": 421068800 + }, + { + "epoch": 0.13, + "learning_rate": 0.000704962888665998, + "loss": 1.6325, + "theoretical_loss": 4.01747379602432, + "tokens_seen": 421134336 + }, + { + "epoch": 0.13, + "learning_rate": 0.0007049468405215647, + "loss": 1.6268, + "theoretical_loss": 4.017404952835336, + "tokens_seen": 421199872 + }, + { + "epoch": 0.13, + "learning_rate": 0.0007049307923771315, + "loss": 1.5411, + "theoretical_loss": 4.017336123355783, + "tokens_seen": 421265408 + }, + { + "epoch": 0.13, + "learning_rate": 0.0007049147442326981, + "loss": 1.6313, + "theoretical_loss": 4.017267307580798, + "tokens_seen": 421330944 + }, + { + "epoch": 0.13, + "learning_rate": 0.0007048986960882649, + "loss": 1.6589, + "theoretical_loss": 4.017198505505521, + "tokens_seen": 421396480 + }, + { + "epoch": 0.13, + "learning_rate": 0.0007048826479438315, + "loss": 1.6431, + "theoretical_loss": 4.017129717125094, + "tokens_seen": 421462016 + }, + { + "epoch": 0.13, + "learning_rate": 0.0007048665997993983, + "loss": 1.6706, + "theoretical_loss": 4.017060942434664, + "tokens_seen": 421527552 + }, + { + "epoch": 0.13, + "learning_rate": 0.0007048505516549649, + "loss": 1.5766, + "theoretical_loss": 4.016992181429377, + "tokens_seen": 421593088 + }, + { + "epoch": 0.13, + "learning_rate": 0.0007048345035105317, + "loss": 1.7066, + "theoretical_loss": 4.016923434104386, + "tokens_seen": 421658624 + }, + { + "epoch": 0.13, + "learning_rate": 0.0007048184553660983, + "loss": 1.6173, + "theoretical_loss": 4.01685470045484, + "tokens_seen": 421724160 + }, + { + "epoch": 0.13, + "learning_rate": 0.0007048024072216651, + "loss": 1.6109, + "theoretical_loss": 4.0167859804758965, + "tokens_seen": 421789696 + }, + { + "epoch": 0.13, + "learning_rate": 0.0007047863590772317, + "loss": 1.5642, + "theoretical_loss": 4.016717274162712, + "tokens_seen": 421855232 + }, + { + "epoch": 0.13, + "learning_rate": 0.0007047703109327985, + "loss": 1.5359, + "theoretical_loss": 4.016648581510448, + "tokens_seen": 421920768 + }, + { + "epoch": 0.13, + "learning_rate": 0.0007047542627883651, + "loss": 1.4681, + "theoretical_loss": 4.016579902514264, + "tokens_seen": 421986304 + }, + { + "epoch": 0.13, + "learning_rate": 0.0007047382146439319, + "loss": 1.5693, + "theoretical_loss": 4.016511237169329, + "tokens_seen": 422051840 + }, + { + "epoch": 0.13, + "learning_rate": 0.0007047221664994985, + "loss": 1.6524, + "theoretical_loss": 4.016442585470807, + "tokens_seen": 422117376 + }, + { + "epoch": 0.13, + "learning_rate": 0.0007047061183550653, + "loss": 1.556, + "theoretical_loss": 4.01637394741387, + "tokens_seen": 422182912 + }, + { + "epoch": 0.13, + "learning_rate": 0.0007046900702106319, + "loss": 1.6746, + "theoretical_loss": 4.016305322993689, + "tokens_seen": 422248448 + }, + { + "epoch": 0.13, + "learning_rate": 0.0007046740220661987, + "loss": 1.5218, + "theoretical_loss": 4.016236712205439, + "tokens_seen": 422313984 + }, + { + "epoch": 0.13, + "learning_rate": 0.0007046579739217653, + "loss": 1.6205, + "theoretical_loss": 4.016168115044298, + "tokens_seen": 422379520 + }, + { + "epoch": 0.13, + "learning_rate": 0.0007046419257773321, + "loss": 1.6639, + "theoretical_loss": 4.016099531505445, + "tokens_seen": 422445056 + }, + { + "epoch": 0.13, + "learning_rate": 0.0007046258776328987, + "loss": 1.5517, + "theoretical_loss": 4.016030961584063, + "tokens_seen": 422510592 + }, + { + "epoch": 0.13, + "learning_rate": 0.0007046098294884655, + "loss": 1.5044, + "theoretical_loss": 4.015962405275336, + "tokens_seen": 422576128 + }, + { + "epoch": 0.13, + "learning_rate": 0.0007045937813440321, + "loss": 1.5926, + "theoretical_loss": 4.015893862574451, + "tokens_seen": 422641664 + }, + { + "epoch": 0.13, + "objective/train/docs_used": 159824, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.5569767951965332, + "objective/train/theoretical_loss": 4.015825333476598, + "objective/train/tokens_used": 443167200, + "theoretical_loss": 4.015825333476598, + "tokens_seen": 422707200 + }, + { + "epoch": 0.13, + "learning_rate": 0.0007045777331995989, + "loss": 1.5987, + "theoretical_loss": 4.015825333476598, + "tokens_seen": 422707200 + }, + { + "epoch": 0.13, + "learning_rate": 0.0007045616850551655, + "loss": 1.6804, + "theoretical_loss": 4.0157568179769685, + "tokens_seen": 422772736 + }, + { + "epoch": 0.13, + "learning_rate": 0.0007045456369107323, + "loss": 1.6243, + "theoretical_loss": 4.015688316070757, + "tokens_seen": 422838272 + }, + { + "epoch": 0.13, + "learning_rate": 0.0007045295887662989, + "loss": 1.6777, + "theoretical_loss": 4.01561982775316, + "tokens_seen": 422903808 + }, + { + "epoch": 0.13, + "learning_rate": 0.0007045135406218657, + "loss": 1.5971, + "theoretical_loss": 4.015551353019378, + "tokens_seen": 422969344 + }, + { + "epoch": 0.13, + "learning_rate": 0.0007044974924774323, + "loss": 1.5986, + "theoretical_loss": 4.015482891864612, + "tokens_seen": 423034880 + }, + { + "epoch": 0.13, + "learning_rate": 0.0007044814443329991, + "loss": 1.6381, + "theoretical_loss": 4.015414444284065, + "tokens_seen": 423100416 + }, + { + "epoch": 0.13, + "learning_rate": 0.0007044653961885657, + "loss": 1.5587, + "theoretical_loss": 4.015346010272946, + "tokens_seen": 423165952 + }, + { + "epoch": 0.13, + "learning_rate": 0.0007044493480441325, + "loss": 1.5881, + "theoretical_loss": 4.0152775898264625, + "tokens_seen": 423231488 + }, + { + "epoch": 0.13, + "learning_rate": 0.0007044332998996991, + "loss": 1.7135, + "theoretical_loss": 4.0152091829398255, + "tokens_seen": 423297024 + }, + { + "epoch": 0.13, + "learning_rate": 0.0007044172517552659, + "loss": 1.5737, + "theoretical_loss": 4.0151407896082505, + "tokens_seen": 423362560 + }, + { + "epoch": 0.13, + "learning_rate": 0.0007044012036108325, + "loss": 1.6827, + "theoretical_loss": 4.015072409826953, + "tokens_seen": 423428096 + }, + { + "epoch": 0.13, + "learning_rate": 0.0007043851554663993, + "loss": 1.5063, + "theoretical_loss": 4.015004043591151, + "tokens_seen": 423493632 + }, + { + "epoch": 0.13, + "learning_rate": 0.0007043691073219659, + "loss": 1.6131, + "theoretical_loss": 4.014935690896067, + "tokens_seen": 423559168 + }, + { + "epoch": 0.13, + "learning_rate": 0.0007043530591775327, + "loss": 1.6524, + "theoretical_loss": 4.014867351736924, + "tokens_seen": 423624704 + }, + { + "epoch": 0.13, + "learning_rate": 0.0007043370110330993, + "loss": 1.6441, + "theoretical_loss": 4.014799026108948, + "tokens_seen": 423690240 + }, + { + "epoch": 0.13, + "learning_rate": 0.0007043209628886661, + "loss": 1.5788, + "theoretical_loss": 4.014730714007367, + "tokens_seen": 423755776 + }, + { + "epoch": 0.13, + "learning_rate": 0.0007043049147442327, + "loss": 1.5509, + "theoretical_loss": 4.014662415427413, + "tokens_seen": 423821312 + }, + { + "epoch": 0.13, + "learning_rate": 0.0007042888665997995, + "loss": 1.5335, + "theoretical_loss": 4.014594130364319, + "tokens_seen": 423886848 + }, + { + "epoch": 0.13, + "learning_rate": 0.0007042728184553661, + "loss": 1.4994, + "theoretical_loss": 4.01452585881332, + "tokens_seen": 423952384 + }, + { + "epoch": 0.13, + "learning_rate": 0.0007042567703109329, + "loss": 1.6535, + "theoretical_loss": 4.014457600769655, + "tokens_seen": 424017920 + }, + { + "epoch": 0.13, + "learning_rate": 0.0007042407221664995, + "loss": 1.6975, + "theoretical_loss": 4.0143893562285635, + "tokens_seen": 424083456 + }, + { + "epoch": 0.13, + "learning_rate": 0.0007042246740220663, + "loss": 1.5766, + "theoretical_loss": 4.014321125185289, + "tokens_seen": 424148992 + }, + { + "epoch": 0.13, + "learning_rate": 0.0007042086258776329, + "loss": 1.61, + "theoretical_loss": 4.014252907635077, + "tokens_seen": 424214528 + }, + { + "epoch": 0.13, + "learning_rate": 0.0007041925777331997, + "loss": 1.65, + "theoretical_loss": 4.014184703573176, + "tokens_seen": 424280064 + }, + { + "epoch": 0.13, + "learning_rate": 0.0007041765295887663, + "loss": 1.7025, + "theoretical_loss": 4.014116512994835, + "tokens_seen": 424345600 + }, + { + "epoch": 0.13, + "learning_rate": 0.0007041604814443331, + "loss": 1.7472, + "theoretical_loss": 4.014048335895307, + "tokens_seen": 424411136 + }, + { + "epoch": 0.13, + "learning_rate": 0.0007041444332998997, + "loss": 1.6785, + "theoretical_loss": 4.0139801722698465, + "tokens_seen": 424476672 + }, + { + "epoch": 0.13, + "learning_rate": 0.0007041283851554665, + "loss": 1.6893, + "theoretical_loss": 4.013912022113711, + "tokens_seen": 424542208 + }, + { + "epoch": 0.13, + "learning_rate": 0.0007041123370110331, + "loss": 1.6459, + "theoretical_loss": 4.013843885422162, + "tokens_seen": 424607744 + }, + { + "epoch": 0.13, + "learning_rate": 0.0007040962888665999, + "loss": 1.58, + "theoretical_loss": 4.01377576219046, + "tokens_seen": 424673280 + }, + { + "epoch": 0.13, + "learning_rate": 0.0007040802407221665, + "loss": 1.6192, + "theoretical_loss": 4.01370765241387, + "tokens_seen": 424738816 + }, + { + "epoch": 0.13, + "learning_rate": 0.0007040641925777333, + "loss": 1.5392, + "theoretical_loss": 4.013639556087659, + "tokens_seen": 424804352 + }, + { + "epoch": 0.13, + "learning_rate": 0.0007040481444332999, + "loss": 1.6678, + "theoretical_loss": 4.0135714732070955, + "tokens_seen": 424869888 + }, + { + "epoch": 0.13, + "learning_rate": 0.0007040320962888667, + "loss": 1.6682, + "theoretical_loss": 4.013503403767453, + "tokens_seen": 424935424 + }, + { + "epoch": 0.13, + "learning_rate": 0.0007040160481444333, + "loss": 1.5844, + "theoretical_loss": 4.013435347764005, + "tokens_seen": 425000960 + }, + { + "epoch": 0.13, + "learning_rate": 0.0007040000000000001, + "loss": 1.4933, + "theoretical_loss": 4.013367305192027, + "tokens_seen": 425066496 + }, + { + "epoch": 0.13, + "learning_rate": 0.0007039839518555667, + "loss": 1.5676, + "theoretical_loss": 4.0132992760468, + "tokens_seen": 425132032 + }, + { + "epoch": 0.13, + "learning_rate": 0.0007039679037111334, + "loss": 1.5337, + "theoretical_loss": 4.013231260323603, + "tokens_seen": 425197568 + }, + { + "epoch": 0.13, + "learning_rate": 0.0007039518555667001, + "loss": 1.4475, + "theoretical_loss": 4.0131632580177214, + "tokens_seen": 425263104 + }, + { + "epoch": 0.13, + "learning_rate": 0.0007039358074222668, + "loss": 1.6928, + "theoretical_loss": 4.01309526912444, + "tokens_seen": 425328640 + }, + { + "epoch": 0.13, + "learning_rate": 0.0007039197592778335, + "loss": 1.7201, + "theoretical_loss": 4.013027293639049, + "tokens_seen": 425394176 + }, + { + "epoch": 0.13, + "learning_rate": 0.0007039037111334002, + "loss": 1.588, + "theoretical_loss": 4.0129593315568375, + "tokens_seen": 425459712 + }, + { + "epoch": 0.13, + "learning_rate": 0.0007038876629889669, + "loss": 1.6586, + "theoretical_loss": 4.0128913828731, + "tokens_seen": 425525248 + }, + { + "epoch": 0.13, + "learning_rate": 0.0007038716148445336, + "loss": 1.6235, + "theoretical_loss": 4.012823447583132, + "tokens_seen": 425590784 + }, + { + "epoch": 0.13, + "learning_rate": 0.0007038555667001003, + "loss": 1.7375, + "theoretical_loss": 4.012755525682231, + "tokens_seen": 425656320 + }, + { + "epoch": 0.13, + "learning_rate": 0.000703839518555667, + "loss": 1.6103, + "theoretical_loss": 4.012687617165698, + "tokens_seen": 425721856 + }, + { + "epoch": 0.13, + "learning_rate": 0.0007038234704112337, + "loss": 1.6077, + "theoretical_loss": 4.012619722028835, + "tokens_seen": 425787392 + }, + { + "epoch": 0.13, + "learning_rate": 0.0007038074222668004, + "loss": 1.588, + "theoretical_loss": 4.012551840266948, + "tokens_seen": 425852928 + }, + { + "epoch": 0.13, + "learning_rate": 0.0007037913741223671, + "loss": 1.4664, + "theoretical_loss": 4.0124839718753424, + "tokens_seen": 425918464 + }, + { + "epoch": 0.13, + "objective/train/docs_used": 161088, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.45851469039917, + "objective/train/theoretical_loss": 4.01241611684933, + "objective/train/tokens_used": 446444000, + "theoretical_loss": 4.01241611684933, + "tokens_seen": 425984000 + }, + { + "epoch": 0.13, + "learning_rate": 0.0007037753259779338, + "loss": 1.5567, + "theoretical_loss": 4.01241611684933, + "tokens_seen": 425984000 + }, + { + "epoch": 0.13, + "learning_rate": 0.0007037592778335005, + "loss": 1.6423, + "theoretical_loss": 4.012348275184223, + "tokens_seen": 426049536 + }, + { + "epoch": 0.13, + "learning_rate": 0.0007037432296890672, + "loss": 1.5159, + "theoretical_loss": 4.012280446875336, + "tokens_seen": 426115072 + }, + { + "epoch": 0.13, + "learning_rate": 0.000703727181544634, + "loss": 1.6569, + "theoretical_loss": 4.012212631917986, + "tokens_seen": 426180608 + }, + { + "epoch": 0.13, + "learning_rate": 0.0007037111334002006, + "loss": 1.6286, + "theoretical_loss": 4.012144830307491, + "tokens_seen": 426246144 + }, + { + "epoch": 0.13, + "learning_rate": 0.0007036950852557673, + "loss": 1.5295, + "theoretical_loss": 4.012077042039174, + "tokens_seen": 426311680 + }, + { + "epoch": 0.13, + "learning_rate": 0.000703679037111334, + "loss": 1.6264, + "theoretical_loss": 4.012009267108358, + "tokens_seen": 426377216 + }, + { + "epoch": 0.13, + "learning_rate": 0.0007036629889669007, + "loss": 1.647, + "theoretical_loss": 4.01194150551037, + "tokens_seen": 426442752 + }, + { + "epoch": 0.13, + "learning_rate": 0.0007036469408224674, + "loss": 1.6109, + "theoretical_loss": 4.011873757240539, + "tokens_seen": 426508288 + }, + { + "epoch": 0.13, + "learning_rate": 0.0007036308926780342, + "loss": 1.7098, + "theoretical_loss": 4.011806022294196, + "tokens_seen": 426573824 + }, + { + "epoch": 0.13, + "learning_rate": 0.0007036148445336008, + "loss": 1.6738, + "theoretical_loss": 4.011738300666674, + "tokens_seen": 426639360 + }, + { + "epoch": 0.13, + "learning_rate": 0.0007035987963891676, + "loss": 1.6444, + "theoretical_loss": 4.011670592353309, + "tokens_seen": 426704896 + }, + { + "epoch": 0.13, + "learning_rate": 0.0007035827482447342, + "loss": 1.5332, + "theoretical_loss": 4.011602897349439, + "tokens_seen": 426770432 + }, + { + "epoch": 0.13, + "learning_rate": 0.000703566700100301, + "loss": 1.6083, + "theoretical_loss": 4.011535215650405, + "tokens_seen": 426835968 + }, + { + "epoch": 0.13, + "learning_rate": 0.0007035506519558676, + "loss": 1.5267, + "theoretical_loss": 4.0114675472515495, + "tokens_seen": 426901504 + }, + { + "epoch": 0.13, + "learning_rate": 0.0007035346038114344, + "loss": 1.6154, + "theoretical_loss": 4.011399892148218, + "tokens_seen": 426967040 + }, + { + "epoch": 0.13, + "learning_rate": 0.000703518555667001, + "loss": 1.5702, + "theoretical_loss": 4.011332250335758, + "tokens_seen": 427032576 + }, + { + "epoch": 0.13, + "learning_rate": 0.0007035025075225678, + "loss": 1.6796, + "theoretical_loss": 4.011264621809518, + "tokens_seen": 427098112 + }, + { + "epoch": 0.13, + "learning_rate": 0.0007034864593781344, + "loss": 1.6088, + "theoretical_loss": 4.011197006564853, + "tokens_seen": 427163648 + }, + { + "epoch": 0.13, + "learning_rate": 0.0007034704112337012, + "loss": 1.6088, + "theoretical_loss": 4.011129404597114, + "tokens_seen": 427229184 + }, + { + "epoch": 0.13, + "learning_rate": 0.0007034543630892678, + "loss": 1.7135, + "theoretical_loss": 4.011061815901661, + "tokens_seen": 427294720 + }, + { + "epoch": 0.13, + "learning_rate": 0.0007034383149448346, + "loss": 1.6143, + "theoretical_loss": 4.010994240473853, + "tokens_seen": 427360256 + }, + { + "epoch": 0.13, + "learning_rate": 0.0007034222668004012, + "loss": 1.6278, + "theoretical_loss": 4.0109266783090485, + "tokens_seen": 427425792 + }, + { + "epoch": 0.13, + "learning_rate": 0.000703406218655968, + "loss": 1.628, + "theoretical_loss": 4.010859129402615, + "tokens_seen": 427491328 + }, + { + "epoch": 0.13, + "learning_rate": 0.0007033901705115346, + "loss": 1.5734, + "theoretical_loss": 4.010791593749916, + "tokens_seen": 427556864 + }, + { + "epoch": 0.13, + "learning_rate": 0.0007033741223671014, + "loss": 1.6682, + "theoretical_loss": 4.010724071346322, + "tokens_seen": 427622400 + }, + { + "epoch": 0.13, + "learning_rate": 0.000703358074222668, + "loss": 1.6793, + "theoretical_loss": 4.0106565621872035, + "tokens_seen": 427687936 + }, + { + "epoch": 0.13, + "learning_rate": 0.0007033420260782348, + "loss": 1.5819, + "theoretical_loss": 4.010589066267933, + "tokens_seen": 427753472 + }, + { + "epoch": 0.13, + "learning_rate": 0.0007033259779338014, + "loss": 1.6736, + "theoretical_loss": 4.010521583583886, + "tokens_seen": 427819008 + }, + { + "epoch": 0.13, + "learning_rate": 0.0007033099297893682, + "loss": 1.6315, + "theoretical_loss": 4.010454114130441, + "tokens_seen": 427884544 + }, + { + "epoch": 0.13, + "learning_rate": 0.0007032938816449348, + "loss": 1.6011, + "theoretical_loss": 4.010386657902979, + "tokens_seen": 427950080 + }, + { + "epoch": 0.13, + "learning_rate": 0.0007032778335005016, + "loss": 1.5661, + "theoretical_loss": 4.01031921489688, + "tokens_seen": 428015616 + }, + { + "epoch": 0.13, + "learning_rate": 0.0007032617853560682, + "loss": 1.5961, + "theoretical_loss": 4.01025178510753, + "tokens_seen": 428081152 + }, + { + "epoch": 0.13, + "learning_rate": 0.000703245737211635, + "loss": 1.483, + "theoretical_loss": 4.010184368530318, + "tokens_seen": 428146688 + }, + { + "epoch": 0.13, + "learning_rate": 0.0007032296890672016, + "loss": 1.5512, + "theoretical_loss": 4.0101169651606305, + "tokens_seen": 428212224 + }, + { + "epoch": 0.13, + "learning_rate": 0.0007032136409227684, + "loss": 1.6766, + "theoretical_loss": 4.010049574993861, + "tokens_seen": 428277760 + }, + { + "epoch": 0.13, + "learning_rate": 0.000703197592778335, + "loss": 1.6173, + "theoretical_loss": 4.009982198025404, + "tokens_seen": 428343296 + }, + { + "epoch": 0.13, + "learning_rate": 0.0007031815446339018, + "loss": 1.6195, + "theoretical_loss": 4.0099148342506545, + "tokens_seen": 428408832 + }, + { + "epoch": 0.13, + "learning_rate": 0.0007031654964894684, + "loss": 1.6648, + "theoretical_loss": 4.009847483665011, + "tokens_seen": 428474368 + }, + { + "epoch": 0.13, + "learning_rate": 0.0007031494483450352, + "loss": 1.7611, + "theoretical_loss": 4.009780146263877, + "tokens_seen": 428539904 + }, + { + "epoch": 0.13, + "learning_rate": 0.0007031334002006018, + "loss": 1.5549, + "theoretical_loss": 4.009712822042652, + "tokens_seen": 428605440 + }, + { + "epoch": 0.13, + "learning_rate": 0.0007031173520561686, + "loss": 1.6077, + "theoretical_loss": 4.009645510996744, + "tokens_seen": 428670976 + }, + { + "epoch": 0.13, + "learning_rate": 0.0007031013039117352, + "loss": 1.4947, + "theoretical_loss": 4.0095782131215625, + "tokens_seen": 428736512 + }, + { + "epoch": 0.13, + "learning_rate": 0.000703085255767302, + "loss": 1.5625, + "theoretical_loss": 4.0095109284125146, + "tokens_seen": 428802048 + }, + { + "epoch": 0.13, + "learning_rate": 0.0007030692076228686, + "loss": 1.6729, + "theoretical_loss": 4.009443656865013, + "tokens_seen": 428867584 + }, + { + "epoch": 0.13, + "learning_rate": 0.0007030531594784354, + "loss": 1.5545, + "theoretical_loss": 4.009376398474475, + "tokens_seen": 428933120 + }, + { + "epoch": 0.13, + "learning_rate": 0.000703037111334002, + "loss": 1.5565, + "theoretical_loss": 4.009309153236316, + "tokens_seen": 428998656 + }, + { + "epoch": 0.13, + "learning_rate": 0.0007030210631895688, + "loss": 1.5848, + "theoretical_loss": 4.009241921145956, + "tokens_seen": 429064192 + }, + { + "epoch": 0.13, + "learning_rate": 0.0007030050150451354, + "loss": 1.64, + "theoretical_loss": 4.009174702198816, + "tokens_seen": 429129728 + }, + { + "epoch": 0.13, + "learning_rate": 0.0007029889669007022, + "loss": 1.5232, + "theoretical_loss": 4.009107496390321, + "tokens_seen": 429195264 + }, + { + "epoch": 0.13, + "objective/train/docs_used": 162190, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.6576623916625977, + "objective/train/theoretical_loss": 4.009040303715897, + "objective/train/tokens_used": 449720800, + "theoretical_loss": 4.009040303715897, + "tokens_seen": 429260800 + }, + { + "epoch": 0.13, + "learning_rate": 0.0007029729187562688, + "loss": 1.6672, + "theoretical_loss": 4.009040303715897, + "tokens_seen": 429260800 + }, + { + "epoch": 0.13, + "learning_rate": 0.0007029568706118356, + "loss": 1.5519, + "theoretical_loss": 4.008973124170971, + "tokens_seen": 429326336 + }, + { + "epoch": 0.13, + "learning_rate": 0.0007029408224674022, + "loss": 1.7368, + "theoretical_loss": 4.008905957750976, + "tokens_seen": 429391872 + }, + { + "epoch": 0.13, + "learning_rate": 0.000702924774322969, + "loss": 1.6646, + "theoretical_loss": 4.0088388044513446, + "tokens_seen": 429457408 + }, + { + "epoch": 0.13, + "learning_rate": 0.0007029087261785356, + "loss": 1.514, + "theoretical_loss": 4.0087716642675115, + "tokens_seen": 429522944 + }, + { + "epoch": 0.13, + "learning_rate": 0.0007028926780341024, + "loss": 1.5666, + "theoretical_loss": 4.008704537194916, + "tokens_seen": 429588480 + }, + { + "epoch": 0.13, + "learning_rate": 0.000702876629889669, + "loss": 1.6072, + "theoretical_loss": 4.008637423228996, + "tokens_seen": 429654016 + }, + { + "epoch": 0.13, + "learning_rate": 0.0007028605817452358, + "loss": 1.6855, + "theoretical_loss": 4.008570322365195, + "tokens_seen": 429719552 + }, + { + "epoch": 0.13, + "learning_rate": 0.0007028445336008024, + "loss": 1.6096, + "theoretical_loss": 4.008503234598958, + "tokens_seen": 429785088 + }, + { + "epoch": 0.13, + "learning_rate": 0.0007028284854563692, + "loss": 1.6465, + "theoretical_loss": 4.008436159925731, + "tokens_seen": 429850624 + }, + { + "epoch": 0.13, + "learning_rate": 0.0007028124373119358, + "loss": 1.6578, + "theoretical_loss": 4.008369098340963, + "tokens_seen": 429916160 + }, + { + "epoch": 0.13, + "learning_rate": 0.0007027963891675026, + "loss": 1.5163, + "theoretical_loss": 4.008302049840106, + "tokens_seen": 429981696 + }, + { + "epoch": 0.13, + "learning_rate": 0.0007027803410230692, + "loss": 1.6333, + "theoretical_loss": 4.008235014418614, + "tokens_seen": 430047232 + }, + { + "epoch": 0.13, + "learning_rate": 0.000702764292878636, + "loss": 1.6766, + "theoretical_loss": 4.008167992071941, + "tokens_seen": 430112768 + }, + { + "epoch": 0.13, + "learning_rate": 0.0007027482447342026, + "loss": 1.6243, + "theoretical_loss": 4.008100982795549, + "tokens_seen": 430178304 + }, + { + "epoch": 0.13, + "learning_rate": 0.0007027321965897694, + "loss": 1.4996, + "theoretical_loss": 4.008033986584895, + "tokens_seen": 430243840 + }, + { + "epoch": 0.13, + "learning_rate": 0.000702716148445336, + "loss": 1.68, + "theoretical_loss": 4.007967003435443, + "tokens_seen": 430309376 + }, + { + "epoch": 0.13, + "learning_rate": 0.0007027001003009028, + "loss": 1.4519, + "theoretical_loss": 4.007900033342658, + "tokens_seen": 430374912 + }, + { + "epoch": 0.13, + "learning_rate": 0.0007026840521564694, + "loss": 1.6022, + "theoretical_loss": 4.007833076302006, + "tokens_seen": 430440448 + }, + { + "epoch": 0.13, + "learning_rate": 0.0007026680040120362, + "loss": 1.5665, + "theoretical_loss": 4.00776613230896, + "tokens_seen": 430505984 + }, + { + "epoch": 0.13, + "learning_rate": 0.0007026519558676028, + "loss": 1.5359, + "theoretical_loss": 4.007699201358989, + "tokens_seen": 430571520 + }, + { + "epoch": 0.13, + "learning_rate": 0.0007026359077231696, + "loss": 1.6963, + "theoretical_loss": 4.007632283447567, + "tokens_seen": 430637056 + }, + { + "epoch": 0.13, + "learning_rate": 0.0007026198595787362, + "loss": 1.5029, + "theoretical_loss": 4.007565378570171, + "tokens_seen": 430702592 + }, + { + "epoch": 0.13, + "learning_rate": 0.000702603811434303, + "loss": 1.6302, + "theoretical_loss": 4.00749848672228, + "tokens_seen": 430768128 + }, + { + "epoch": 0.13, + "learning_rate": 0.0007025877632898696, + "loss": 1.539, + "theoretical_loss": 4.007431607899375, + "tokens_seen": 430833664 + }, + { + "epoch": 0.13, + "learning_rate": 0.0007025717151454363, + "loss": 1.4923, + "theoretical_loss": 4.007364742096938, + "tokens_seen": 430899200 + }, + { + "epoch": 0.13, + "learning_rate": 0.000702555667001003, + "loss": 1.5942, + "theoretical_loss": 4.007297889310455, + "tokens_seen": 430964736 + }, + { + "epoch": 0.13, + "learning_rate": 0.0007025396188565697, + "loss": 1.5806, + "theoretical_loss": 4.007231049535413, + "tokens_seen": 431030272 + }, + { + "epoch": 0.13, + "learning_rate": 0.0007025235707121364, + "loss": 1.5919, + "theoretical_loss": 4.007164222767304, + "tokens_seen": 431095808 + }, + { + "epoch": 0.13, + "learning_rate": 0.0007025075225677031, + "loss": 1.6785, + "theoretical_loss": 4.0070974090016165, + "tokens_seen": 431161344 + }, + { + "epoch": 0.13, + "learning_rate": 0.0007024914744232698, + "loss": 1.6331, + "theoretical_loss": 4.007030608233848, + "tokens_seen": 431226880 + }, + { + "epoch": 0.13, + "learning_rate": 0.0007024754262788365, + "loss": 1.7187, + "theoretical_loss": 4.006963820459494, + "tokens_seen": 431292416 + }, + { + "epoch": 0.13, + "learning_rate": 0.0007024593781344032, + "loss": 1.5661, + "theoretical_loss": 4.006897045674053, + "tokens_seen": 431357952 + }, + { + "epoch": 0.13, + "learning_rate": 0.0007024433299899699, + "loss": 1.5943, + "theoretical_loss": 4.006830283873027, + "tokens_seen": 431423488 + }, + { + "epoch": 0.13, + "learning_rate": 0.0007024272818455366, + "loss": 1.5423, + "theoretical_loss": 4.006763535051918, + "tokens_seen": 431489024 + }, + { + "epoch": 0.13, + "learning_rate": 0.0007024112337011033, + "loss": 1.563, + "theoretical_loss": 4.006696799206233, + "tokens_seen": 431554560 + }, + { + "epoch": 0.13, + "learning_rate": 0.00070239518555667, + "loss": 1.5921, + "theoretical_loss": 4.006630076331478, + "tokens_seen": 431620096 + }, + { + "epoch": 0.13, + "learning_rate": 0.0007023791374122367, + "loss": 1.6089, + "theoretical_loss": 4.0065633664231655, + "tokens_seen": 431685632 + }, + { + "epoch": 0.13, + "learning_rate": 0.0007023630892678034, + "loss": 1.5931, + "theoretical_loss": 4.006496669476805, + "tokens_seen": 431751168 + }, + { + "epoch": 0.13, + "learning_rate": 0.0007023470411233701, + "loss": 1.601, + "theoretical_loss": 4.006429985487913, + "tokens_seen": 431816704 + }, + { + "epoch": 0.13, + "learning_rate": 0.0007023309929789368, + "loss": 1.5569, + "theoretical_loss": 4.006363314452007, + "tokens_seen": 431882240 + }, + { + "epoch": 0.13, + "learning_rate": 0.0007023149448345035, + "loss": 1.6183, + "theoretical_loss": 4.006296656364603, + "tokens_seen": 431947776 + }, + { + "epoch": 0.13, + "learning_rate": 0.0007022988966900702, + "loss": 1.6504, + "theoretical_loss": 4.006230011221225, + "tokens_seen": 432013312 + }, + { + "epoch": 0.13, + "learning_rate": 0.0007022828485456369, + "loss": 1.7131, + "theoretical_loss": 4.006163379017394, + "tokens_seen": 432078848 + }, + { + "epoch": 0.13, + "learning_rate": 0.0007022668004012037, + "loss": 1.602, + "theoretical_loss": 4.0060967597486385, + "tokens_seen": 432144384 + }, + { + "epoch": 0.13, + "learning_rate": 0.0007022507522567703, + "loss": 1.6011, + "theoretical_loss": 4.006030153410483, + "tokens_seen": 432209920 + }, + { + "epoch": 0.13, + "learning_rate": 0.000702234704112337, + "loss": 1.6546, + "theoretical_loss": 4.005963559998461, + "tokens_seen": 432275456 + }, + { + "epoch": 0.13, + "learning_rate": 0.0007022186559679037, + "loss": 1.5754, + "theoretical_loss": 4.005896979508103, + "tokens_seen": 432340992 + }, + { + "epoch": 0.13, + "learning_rate": 0.0007022026078234705, + "loss": 1.6521, + "theoretical_loss": 4.005830411934944, + "tokens_seen": 432406528 + }, + { + "epoch": 0.13, + "learning_rate": 0.0007021865596790371, + "loss": 1.6028, + "theoretical_loss": 4.005763857274521, + "tokens_seen": 432472064 + }, + { + "epoch": 0.13, + "objective/train/docs_used": 163281, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.5710229873657227, + "objective/train/theoretical_loss": 4.005697315522371, + "objective/train/tokens_used": 452997600, + "theoretical_loss": 4.005697315522371, + "tokens_seen": 432537600 + }, + { + "epoch": 0.13, + "learning_rate": 0.0007021705115346039, + "loss": 1.6317, + "theoretical_loss": 4.005697315522371, + "tokens_seen": 432537600 + }, + { + "epoch": 0.13, + "learning_rate": 0.0007021544633901705, + "loss": 1.6172, + "theoretical_loss": 4.005630786674039, + "tokens_seen": 432603136 + }, + { + "epoch": 0.13, + "learning_rate": 0.0007021384152457373, + "loss": 1.6338, + "theoretical_loss": 4.005564270725065, + "tokens_seen": 432668672 + }, + { + "epoch": 0.13, + "learning_rate": 0.0007021223671013039, + "loss": 1.6106, + "theoretical_loss": 4.005497767670996, + "tokens_seen": 432734208 + }, + { + "epoch": 0.13, + "learning_rate": 0.0007021063189568707, + "loss": 1.4825, + "theoretical_loss": 4.0054312775073795, + "tokens_seen": 432799744 + }, + { + "epoch": 0.13, + "learning_rate": 0.0007020902708124373, + "loss": 1.5358, + "theoretical_loss": 4.005364800229766, + "tokens_seen": 432865280 + }, + { + "epoch": 0.13, + "learning_rate": 0.0007020742226680041, + "loss": 1.6422, + "theoretical_loss": 4.005298335833708, + "tokens_seen": 432930816 + }, + { + "epoch": 0.13, + "learning_rate": 0.0007020581745235707, + "loss": 1.5811, + "theoretical_loss": 4.005231884314759, + "tokens_seen": 432996352 + }, + { + "epoch": 0.13, + "learning_rate": 0.0007020421263791375, + "loss": 1.5568, + "theoretical_loss": 4.005165445668477, + "tokens_seen": 433061888 + }, + { + "epoch": 0.13, + "learning_rate": 0.0007020260782347041, + "loss": 1.533, + "theoretical_loss": 4.0050990198904195, + "tokens_seen": 433127424 + }, + { + "epoch": 0.13, + "learning_rate": 0.0007020100300902709, + "loss": 1.6386, + "theoretical_loss": 4.005032606976149, + "tokens_seen": 433192960 + }, + { + "epoch": 0.13, + "learning_rate": 0.0007019939819458375, + "loss": 1.5969, + "theoretical_loss": 4.004966206921227, + "tokens_seen": 433258496 + }, + { + "epoch": 0.13, + "learning_rate": 0.0007019779338014043, + "loss": 1.5568, + "theoretical_loss": 4.004899819721221, + "tokens_seen": 433324032 + }, + { + "epoch": 0.13, + "learning_rate": 0.0007019618856569709, + "loss": 1.5963, + "theoretical_loss": 4.004833445371697, + "tokens_seen": 433389568 + }, + { + "epoch": 0.13, + "learning_rate": 0.0007019458375125377, + "loss": 1.6933, + "theoretical_loss": 4.004767083868226, + "tokens_seen": 433455104 + }, + { + "epoch": 0.13, + "learning_rate": 0.0007019297893681043, + "loss": 1.7085, + "theoretical_loss": 4.004700735206381, + "tokens_seen": 433520640 + }, + { + "epoch": 0.13, + "learning_rate": 0.0007019137412236711, + "loss": 1.6475, + "theoretical_loss": 4.004634399381733, + "tokens_seen": 433586176 + }, + { + "epoch": 0.13, + "learning_rate": 0.0007018976930792377, + "loss": 1.7056, + "theoretical_loss": 4.0045680763898615, + "tokens_seen": 433651712 + }, + { + "epoch": 0.13, + "learning_rate": 0.0007018816449348045, + "loss": 1.5372, + "theoretical_loss": 4.004501766226344, + "tokens_seen": 433717248 + }, + { + "epoch": 0.13, + "learning_rate": 0.0007018655967903711, + "loss": 1.5431, + "theoretical_loss": 4.004435468886761, + "tokens_seen": 433782784 + }, + { + "epoch": 0.13, + "learning_rate": 0.0007018495486459379, + "loss": 1.6896, + "theoretical_loss": 4.0043691843666975, + "tokens_seen": 433848320 + }, + { + "epoch": 0.13, + "learning_rate": 0.0007018335005015045, + "loss": 1.4954, + "theoretical_loss": 4.004302912661736, + "tokens_seen": 433913856 + }, + { + "epoch": 0.13, + "learning_rate": 0.0007018174523570713, + "loss": 1.5995, + "theoretical_loss": 4.0042366537674665, + "tokens_seen": 433979392 + }, + { + "epoch": 0.13, + "learning_rate": 0.0007018014042126379, + "loss": 1.5774, + "theoretical_loss": 4.004170407679476, + "tokens_seen": 434044928 + }, + { + "epoch": 0.13, + "learning_rate": 0.0007017853560682047, + "loss": 1.5988, + "theoretical_loss": 4.004104174393359, + "tokens_seen": 434110464 + }, + { + "epoch": 0.13, + "learning_rate": 0.0007017693079237713, + "loss": 1.5903, + "theoretical_loss": 4.004037953904708, + "tokens_seen": 434176000 + }, + { + "epoch": 0.13, + "learning_rate": 0.0007017532597793381, + "loss": 1.5614, + "theoretical_loss": 4.003971746209119, + "tokens_seen": 434241536 + }, + { + "epoch": 0.13, + "learning_rate": 0.0007017372116349047, + "loss": 1.684, + "theoretical_loss": 4.003905551302191, + "tokens_seen": 434307072 + }, + { + "epoch": 0.13, + "learning_rate": 0.0007017211634904715, + "loss": 1.4874, + "theoretical_loss": 4.0038393691795235, + "tokens_seen": 434372608 + }, + { + "epoch": 0.13, + "learning_rate": 0.0007017051153460381, + "loss": 1.652, + "theoretical_loss": 4.003773199836721, + "tokens_seen": 434438144 + }, + { + "epoch": 0.13, + "learning_rate": 0.0007016890672016049, + "loss": 1.4878, + "theoretical_loss": 4.003707043269387, + "tokens_seen": 434503680 + }, + { + "epoch": 0.13, + "learning_rate": 0.0007016730190571715, + "loss": 1.7286, + "theoretical_loss": 4.003640899473128, + "tokens_seen": 434569216 + }, + { + "epoch": 0.13, + "learning_rate": 0.0007016569709127383, + "loss": 1.6185, + "theoretical_loss": 4.0035747684435545, + "tokens_seen": 434634752 + }, + { + "epoch": 0.13, + "learning_rate": 0.0007016409227683049, + "loss": 1.6519, + "theoretical_loss": 4.003508650176278, + "tokens_seen": 434700288 + }, + { + "epoch": 0.13, + "learning_rate": 0.0007016248746238717, + "loss": 1.6645, + "theoretical_loss": 4.003442544666911, + "tokens_seen": 434765824 + }, + { + "epoch": 0.13, + "learning_rate": 0.0007016088264794383, + "loss": 1.6876, + "theoretical_loss": 4.003376451911071, + "tokens_seen": 434831360 + }, + { + "epoch": 0.13, + "learning_rate": 0.0007015927783350051, + "loss": 1.6093, + "theoretical_loss": 4.003310371904373, + "tokens_seen": 434896896 + }, + { + "epoch": 0.13, + "learning_rate": 0.0007015767301905717, + "loss": 1.6206, + "theoretical_loss": 4.00324430464244, + "tokens_seen": 434962432 + }, + { + "epoch": 0.13, + "learning_rate": 0.0007015606820461385, + "loss": 1.6321, + "theoretical_loss": 4.003178250120892, + "tokens_seen": 435027968 + }, + { + "epoch": 0.13, + "learning_rate": 0.0007015446339017051, + "loss": 1.7192, + "theoretical_loss": 4.003112208335355, + "tokens_seen": 435093504 + }, + { + "epoch": 0.13, + "learning_rate": 0.0007015285857572719, + "loss": 1.5402, + "theoretical_loss": 4.003046179281455, + "tokens_seen": 435159040 + }, + { + "epoch": 0.13, + "learning_rate": 0.0007015125376128385, + "loss": 1.679, + "theoretical_loss": 4.00298016295482, + "tokens_seen": 435224576 + }, + { + "epoch": 0.13, + "learning_rate": 0.0007014964894684053, + "loss": 1.4273, + "theoretical_loss": 4.002914159351082, + "tokens_seen": 435290112 + }, + { + "epoch": 0.13, + "learning_rate": 0.000701480441323972, + "loss": 1.5886, + "theoretical_loss": 4.0028481684658725, + "tokens_seen": 435355648 + }, + { + "epoch": 0.13, + "learning_rate": 0.0007014643931795387, + "loss": 1.6108, + "theoretical_loss": 4.002782190294829, + "tokens_seen": 435421184 + }, + { + "epoch": 0.13, + "learning_rate": 0.0007014483450351054, + "loss": 1.6985, + "theoretical_loss": 4.002716224833588, + "tokens_seen": 435486720 + }, + { + "epoch": 0.13, + "learning_rate": 0.0007014322968906721, + "loss": 1.4741, + "theoretical_loss": 4.002650272077787, + "tokens_seen": 435552256 + }, + { + "epoch": 0.13, + "learning_rate": 0.0007014162487462388, + "loss": 1.5535, + "theoretical_loss": 4.00258433202307, + "tokens_seen": 435617792 + }, + { + "epoch": 0.13, + "learning_rate": 0.0007014002006018055, + "loss": 1.6085, + "theoretical_loss": 4.0025184046650795, + "tokens_seen": 435683328 + }, + { + "epoch": 0.13, + "learning_rate": 0.0007013841524573722, + "loss": 1.6967, + "theoretical_loss": 4.002452489999463, + "tokens_seen": 435748864 + }, + { + "epoch": 0.13, + "objective/train/docs_used": 164592, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.5154685974121094, + "objective/train/theoretical_loss": 4.002386588021867, + "objective/train/tokens_used": 456274400, + "theoretical_loss": 4.002386588021867, + "tokens_seen": 435814400 + }, + { + "epoch": 0.13, + "learning_rate": 0.0007013681043129389, + "loss": 1.6358, + "theoretical_loss": 4.002386588021867, + "tokens_seen": 435814400 + }, + { + "epoch": 0.13, + "learning_rate": 0.0007013520561685056, + "loss": 1.5536, + "theoretical_loss": 4.002320698727942, + "tokens_seen": 435879936 + }, + { + "epoch": 0.13, + "learning_rate": 0.0007013360080240723, + "loss": 1.5853, + "theoretical_loss": 4.002254822113341, + "tokens_seen": 435945472 + }, + { + "epoch": 0.13, + "learning_rate": 0.000701319959879639, + "loss": 1.6412, + "theoretical_loss": 4.0021889581737184, + "tokens_seen": 436011008 + }, + { + "epoch": 0.13, + "learning_rate": 0.0007013039117352057, + "loss": 1.6772, + "theoretical_loss": 4.00212310690473, + "tokens_seen": 436076544 + }, + { + "epoch": 0.13, + "learning_rate": 0.0007012878635907723, + "loss": 1.5282, + "theoretical_loss": 4.0020572683020355, + "tokens_seen": 436142080 + }, + { + "epoch": 0.13, + "learning_rate": 0.0007012718154463391, + "loss": 1.6343, + "theoretical_loss": 4.0019914423612954, + "tokens_seen": 436207616 + }, + { + "epoch": 0.13, + "learning_rate": 0.0007012557673019057, + "loss": 1.5542, + "theoretical_loss": 4.0019256290781735, + "tokens_seen": 436273152 + }, + { + "epoch": 0.13, + "learning_rate": 0.0007012397191574725, + "loss": 1.6666, + "theoretical_loss": 4.001859828448334, + "tokens_seen": 436338688 + }, + { + "epoch": 0.13, + "learning_rate": 0.0007012236710130391, + "loss": 1.5928, + "theoretical_loss": 4.001794040467446, + "tokens_seen": 436404224 + }, + { + "epoch": 0.13, + "learning_rate": 0.0007012076228686059, + "loss": 1.5997, + "theoretical_loss": 4.0017282651311765, + "tokens_seen": 436469760 + }, + { + "epoch": 0.13, + "learning_rate": 0.0007011915747241725, + "loss": 1.7025, + "theoretical_loss": 4.001662502435199, + "tokens_seen": 436535296 + }, + { + "epoch": 0.13, + "learning_rate": 0.0007011755265797393, + "loss": 1.4527, + "theoretical_loss": 4.001596752375186, + "tokens_seen": 436600832 + }, + { + "epoch": 0.13, + "learning_rate": 0.0007011594784353059, + "loss": 1.5828, + "theoretical_loss": 4.001531014946815, + "tokens_seen": 436666368 + }, + { + "epoch": 0.13, + "learning_rate": 0.0007011434302908726, + "loss": 1.649, + "theoretical_loss": 4.001465290145763, + "tokens_seen": 436731904 + }, + { + "epoch": 0.13, + "learning_rate": 0.0007011273821464393, + "loss": 1.6637, + "theoretical_loss": 4.00139957796771, + "tokens_seen": 436797440 + }, + { + "epoch": 0.13, + "learning_rate": 0.000701111334002006, + "loss": 1.7111, + "theoretical_loss": 4.001333878408339, + "tokens_seen": 436862976 + }, + { + "epoch": 0.13, + "learning_rate": 0.0007010952858575727, + "loss": 1.5243, + "theoretical_loss": 4.001268191463334, + "tokens_seen": 436928512 + }, + { + "epoch": 0.13, + "learning_rate": 0.0007010792377131394, + "loss": 1.5254, + "theoretical_loss": 4.00120251712838, + "tokens_seen": 436994048 + }, + { + "epoch": 0.13, + "learning_rate": 0.0007010631895687061, + "loss": 1.6567, + "theoretical_loss": 4.001136855399168, + "tokens_seen": 437059584 + }, + { + "epoch": 0.13, + "learning_rate": 0.0007010471414242728, + "loss": 1.6572, + "theoretical_loss": 4.001071206271388, + "tokens_seen": 437125120 + }, + { + "epoch": 0.13, + "learning_rate": 0.0007010310932798395, + "loss": 1.6083, + "theoretical_loss": 4.001005569740732, + "tokens_seen": 437190656 + }, + { + "epoch": 0.13, + "learning_rate": 0.0007010150451354062, + "loss": 1.4764, + "theoretical_loss": 4.000939945802896, + "tokens_seen": 437256192 + }, + { + "epoch": 0.13, + "learning_rate": 0.000700998996990973, + "loss": 1.5354, + "theoretical_loss": 4.000874334453577, + "tokens_seen": 437321728 + }, + { + "epoch": 0.13, + "learning_rate": 0.0007009829488465396, + "loss": 1.563, + "theoretical_loss": 4.000808735688474, + "tokens_seen": 437387264 + }, + { + "epoch": 0.13, + "learning_rate": 0.0007009669007021063, + "loss": 1.6405, + "theoretical_loss": 4.000743149503288, + "tokens_seen": 437452800 + }, + { + "epoch": 0.13, + "learning_rate": 0.000700950852557673, + "loss": 1.6034, + "theoretical_loss": 4.000677575893723, + "tokens_seen": 437518336 + }, + { + "epoch": 0.13, + "learning_rate": 0.0007009348044132397, + "loss": 1.4434, + "theoretical_loss": 4.000612014855484, + "tokens_seen": 437583872 + }, + { + "epoch": 0.13, + "learning_rate": 0.0007009187562688064, + "loss": 1.7228, + "theoretical_loss": 4.00054646638428, + "tokens_seen": 437649408 + }, + { + "epoch": 0.13, + "learning_rate": 0.0007009027081243731, + "loss": 1.5416, + "theoretical_loss": 4.000480930475819, + "tokens_seen": 437714944 + }, + { + "epoch": 0.13, + "learning_rate": 0.0007008866599799398, + "loss": 1.6567, + "theoretical_loss": 4.000415407125814, + "tokens_seen": 437780480 + }, + { + "epoch": 0.13, + "learning_rate": 0.0007008706118355066, + "loss": 1.6696, + "theoretical_loss": 4.000349896329979, + "tokens_seen": 437846016 + }, + { + "epoch": 0.13, + "learning_rate": 0.0007008545636910732, + "loss": 1.6441, + "theoretical_loss": 4.0002843980840295, + "tokens_seen": 437911552 + }, + { + "epoch": 0.13, + "learning_rate": 0.00070083851554664, + "loss": 1.5878, + "theoretical_loss": 4.0002189123836835, + "tokens_seen": 437977088 + }, + { + "epoch": 0.13, + "learning_rate": 0.0007008224674022066, + "loss": 1.594, + "theoretical_loss": 4.000153439224662, + "tokens_seen": 438042624 + }, + { + "epoch": 0.13, + "learning_rate": 0.0007008064192577734, + "loss": 1.4557, + "theoretical_loss": 4.0000879786026875, + "tokens_seen": 438108160 + }, + { + "epoch": 0.13, + "learning_rate": 0.00070079037111334, + "loss": 1.6099, + "theoretical_loss": 4.000022530513484, + "tokens_seen": 438173696 + }, + { + "epoch": 0.13, + "learning_rate": 0.0007007743229689068, + "loss": 1.6393, + "theoretical_loss": 3.999957094952779, + "tokens_seen": 438239232 + }, + { + "epoch": 0.13, + "learning_rate": 0.0007007582748244734, + "loss": 1.6359, + "theoretical_loss": 3.9998916719163002, + "tokens_seen": 438304768 + }, + { + "epoch": 0.13, + "learning_rate": 0.0007007422266800402, + "loss": 1.5887, + "theoretical_loss": 3.999826261399778, + "tokens_seen": 438370304 + }, + { + "epoch": 0.13, + "learning_rate": 0.0007007261785356068, + "loss": 1.4928, + "theoretical_loss": 3.9997608633989463, + "tokens_seen": 438435840 + }, + { + "epoch": 0.13, + "learning_rate": 0.0007007101303911736, + "loss": 1.6151, + "theoretical_loss": 3.9996954779095395, + "tokens_seen": 438501376 + }, + { + "epoch": 0.13, + "learning_rate": 0.0007006940822467402, + "loss": 1.6557, + "theoretical_loss": 3.999630104927295, + "tokens_seen": 438566912 + }, + { + "epoch": 0.13, + "learning_rate": 0.000700678034102307, + "loss": 1.6481, + "theoretical_loss": 3.999564744447952, + "tokens_seen": 438632448 + }, + { + "epoch": 0.13, + "learning_rate": 0.0007006619859578736, + "loss": 1.6973, + "theoretical_loss": 3.9994993964672516, + "tokens_seen": 438697984 + }, + { + "epoch": 0.13, + "learning_rate": 0.0007006459378134404, + "loss": 1.5898, + "theoretical_loss": 3.9994340609809367, + "tokens_seen": 438763520 + }, + { + "epoch": 0.13, + "learning_rate": 0.000700629889669007, + "loss": 1.6288, + "theoretical_loss": 3.999368737984754, + "tokens_seen": 438829056 + }, + { + "epoch": 0.13, + "learning_rate": 0.0007006138415245738, + "loss": 1.5492, + "theoretical_loss": 3.999303427474449, + "tokens_seen": 438894592 + }, + { + "epoch": 0.13, + "learning_rate": 0.0007005977933801404, + "loss": 1.6196, + "theoretical_loss": 3.9992381294457733, + "tokens_seen": 438960128 + }, + { + "epoch": 0.13, + "learning_rate": 0.0007005817452357072, + "loss": 1.5769, + "theoretical_loss": 3.9991728438944767, + "tokens_seen": 439025664 + }, + { + "epoch": 0.13, + "objective/train/docs_used": 165818, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.5964142084121704, + "objective/train/theoretical_loss": 3.9991075708163146, + "objective/train/tokens_used": 459551200, + "theoretical_loss": 3.9991075708163146, + "tokens_seen": 439091200 + }, + { + "epoch": 0.13, + "learning_rate": 0.0007005656970912738, + "loss": 1.6088, + "theoretical_loss": 3.9991075708163146, + "tokens_seen": 439091200 + }, + { + "epoch": 0.13, + "learning_rate": 0.0007005496489468406, + "loss": 1.6869, + "theoretical_loss": 3.9990423102070416, + "tokens_seen": 439156736 + }, + { + "epoch": 0.13, + "learning_rate": 0.0007005336008024072, + "loss": 1.6322, + "theoretical_loss": 3.9989770620624165, + "tokens_seen": 439222272 + }, + { + "epoch": 0.13, + "learning_rate": 0.000700517552657974, + "loss": 1.7324, + "theoretical_loss": 3.9989118263781984, + "tokens_seen": 439287808 + }, + { + "epoch": 0.13, + "learning_rate": 0.0007005015045135406, + "loss": 1.6251, + "theoretical_loss": 3.99884660315015, + "tokens_seen": 439353344 + }, + { + "epoch": 0.13, + "learning_rate": 0.0007004854563691074, + "loss": 1.6993, + "theoretical_loss": 3.9987813923740356, + "tokens_seen": 439418880 + }, + { + "epoch": 0.13, + "learning_rate": 0.000700469408224674, + "loss": 1.5786, + "theoretical_loss": 3.9987161940456204, + "tokens_seen": 439484416 + }, + { + "epoch": 0.13, + "learning_rate": 0.0007004533600802408, + "loss": 1.5472, + "theoretical_loss": 3.9986510081606736, + "tokens_seen": 439549952 + }, + { + "epoch": 0.13, + "learning_rate": 0.0007004373119358074, + "loss": 1.655, + "theoretical_loss": 3.998585834714965, + "tokens_seen": 439615488 + }, + { + "epoch": 0.13, + "learning_rate": 0.0007004212637913742, + "loss": 1.7143, + "theoretical_loss": 3.9985206737042676, + "tokens_seen": 439681024 + }, + { + "epoch": 0.13, + "learning_rate": 0.0007004052156469408, + "loss": 1.6395, + "theoretical_loss": 3.998455525124355, + "tokens_seen": 439746560 + }, + { + "epoch": 0.13, + "learning_rate": 0.0007003891675025076, + "loss": 1.5936, + "theoretical_loss": 3.9983903889710044, + "tokens_seen": 439812096 + }, + { + "epoch": 0.13, + "learning_rate": 0.0007003731193580742, + "loss": 1.624, + "theoretical_loss": 3.998325265239994, + "tokens_seen": 439877632 + }, + { + "epoch": 0.13, + "learning_rate": 0.000700357071213641, + "loss": 1.6107, + "theoretical_loss": 3.998260153927105, + "tokens_seen": 439943168 + }, + { + "epoch": 0.13, + "learning_rate": 0.0007003410230692076, + "loss": 1.5407, + "theoretical_loss": 3.998195055028119, + "tokens_seen": 440008704 + }, + { + "epoch": 0.13, + "learning_rate": 0.0007003249749247744, + "loss": 1.5914, + "theoretical_loss": 3.998129968538822, + "tokens_seen": 440074240 + }, + { + "epoch": 0.13, + "learning_rate": 0.000700308926780341, + "loss": 1.5818, + "theoretical_loss": 3.998064894455, + "tokens_seen": 440139776 + }, + { + "epoch": 0.13, + "learning_rate": 0.0007002928786359078, + "loss": 1.6496, + "theoretical_loss": 3.997999832772443, + "tokens_seen": 440205312 + }, + { + "epoch": 0.13, + "learning_rate": 0.0007002768304914745, + "loss": 1.6154, + "theoretical_loss": 3.997934783486941, + "tokens_seen": 440270848 + }, + { + "epoch": 0.13, + "learning_rate": 0.0007002607823470412, + "loss": 1.5914, + "theoretical_loss": 3.997869746594287, + "tokens_seen": 440336384 + }, + { + "epoch": 0.13, + "learning_rate": 0.0007002447342026079, + "loss": 1.6736, + "theoretical_loss": 3.9978047220902764, + "tokens_seen": 440401920 + }, + { + "epoch": 0.13, + "learning_rate": 0.0007002286860581746, + "loss": 1.6113, + "theoretical_loss": 3.997739709970706, + "tokens_seen": 440467456 + }, + { + "epoch": 0.13, + "learning_rate": 0.0007002126379137413, + "loss": 1.5253, + "theoretical_loss": 3.997674710231375, + "tokens_seen": 440532992 + }, + { + "epoch": 0.13, + "learning_rate": 0.000700196589769308, + "loss": 1.7454, + "theoretical_loss": 3.997609722868085, + "tokens_seen": 440598528 + }, + { + "epoch": 0.13, + "learning_rate": 0.0007001805416248747, + "loss": 1.5116, + "theoretical_loss": 3.9975447478766393, + "tokens_seen": 440664064 + }, + { + "epoch": 0.13, + "learning_rate": 0.0007001644934804414, + "loss": 1.6059, + "theoretical_loss": 3.997479785252842, + "tokens_seen": 440729600 + }, + { + "epoch": 0.13, + "learning_rate": 0.0007001484453360081, + "loss": 1.6564, + "theoretical_loss": 3.997414834992502, + "tokens_seen": 440795136 + }, + { + "epoch": 0.13, + "learning_rate": 0.0007001323971915748, + "loss": 1.5729, + "theoretical_loss": 3.9973498970914285, + "tokens_seen": 440860672 + }, + { + "epoch": 0.13, + "learning_rate": 0.0007001163490471415, + "loss": 1.6291, + "theoretical_loss": 3.9972849715454317, + "tokens_seen": 440926208 + }, + { + "epoch": 0.13, + "learning_rate": 0.0007001003009027082, + "loss": 1.5651, + "theoretical_loss": 3.9972200583503263, + "tokens_seen": 440991744 + }, + { + "epoch": 0.13, + "learning_rate": 0.0007000842527582749, + "loss": 1.6301, + "theoretical_loss": 3.9971551575019273, + "tokens_seen": 441057280 + }, + { + "epoch": 0.13, + "learning_rate": 0.0007000682046138416, + "loss": 1.6889, + "theoretical_loss": 3.997090268996052, + "tokens_seen": 441122816 + }, + { + "epoch": 0.13, + "learning_rate": 0.0007000521564694083, + "loss": 1.5571, + "theoretical_loss": 3.9970253928285215, + "tokens_seen": 441188352 + }, + { + "epoch": 0.13, + "learning_rate": 0.000700036108324975, + "loss": 1.5834, + "theoretical_loss": 3.996960528995155, + "tokens_seen": 441253888 + }, + { + "epoch": 0.13, + "learning_rate": 0.0007000200601805417, + "loss": 1.7046, + "theoretical_loss": 3.996895677491778, + "tokens_seen": 441319424 + }, + { + "epoch": 0.13, + "learning_rate": 0.0007000040120361084, + "loss": 1.6088, + "theoretical_loss": 3.996830838314216, + "tokens_seen": 441384960 + }, + { + "epoch": 0.13, + "learning_rate": 0.0006999879638916751, + "loss": 1.5347, + "theoretical_loss": 3.996766011458296, + "tokens_seen": 441450496 + }, + { + "epoch": 0.13, + "learning_rate": 0.0006999719157472418, + "loss": 1.5956, + "theoretical_loss": 3.9967011969198483, + "tokens_seen": 441516032 + }, + { + "epoch": 0.13, + "learning_rate": 0.0006999558676028085, + "loss": 1.5257, + "theoretical_loss": 3.996636394694705, + "tokens_seen": 441581568 + }, + { + "epoch": 0.13, + "learning_rate": 0.0006999398194583752, + "loss": 1.5839, + "theoretical_loss": 3.9965716047786985, + "tokens_seen": 441647104 + }, + { + "epoch": 0.13, + "learning_rate": 0.0006999237713139419, + "loss": 1.3907, + "theoretical_loss": 3.996506827167667, + "tokens_seen": 441712640 + }, + { + "epoch": 0.13, + "learning_rate": 0.0006999077231695086, + "loss": 1.6416, + "theoretical_loss": 3.9964420618574454, + "tokens_seen": 441778176 + }, + { + "epoch": 0.13, + "learning_rate": 0.0006998916750250752, + "loss": 1.6591, + "theoretical_loss": 3.9963773088438765, + "tokens_seen": 441843712 + }, + { + "epoch": 0.13, + "learning_rate": 0.000699875626880642, + "loss": 1.5938, + "theoretical_loss": 3.996312568122801, + "tokens_seen": 441909248 + }, + { + "epoch": 0.13, + "learning_rate": 0.0006998595787362086, + "loss": 1.5721, + "theoretical_loss": 3.9962478396900627, + "tokens_seen": 441974784 + }, + { + "epoch": 0.13, + "learning_rate": 0.0006998435305917754, + "loss": 1.5602, + "theoretical_loss": 3.996183123541507, + "tokens_seen": 442040320 + }, + { + "epoch": 0.13, + "learning_rate": 0.000699827482447342, + "loss": 1.7715, + "theoretical_loss": 3.9961184196729835, + "tokens_seen": 442105856 + }, + { + "epoch": 0.13, + "learning_rate": 0.0006998114343029088, + "loss": 1.5231, + "theoretical_loss": 3.9960537280803408, + "tokens_seen": 442171392 + }, + { + "epoch": 0.13, + "learning_rate": 0.0006997953861584754, + "loss": 1.5739, + "theoretical_loss": 3.995989048759432, + "tokens_seen": 442236928 + }, + { + "epoch": 0.13, + "learning_rate": 0.0006997793380140422, + "loss": 1.6284, + "theoretical_loss": 3.99592438170611, + "tokens_seen": 442302464 + }, + { + "epoch": 0.13, + "objective/train/docs_used": 166951, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.569860577583313, + "objective/train/theoretical_loss": 3.9958597269162315, + "objective/train/tokens_used": 462828000, + "theoretical_loss": 3.9958597269162315, + "tokens_seen": 442368000 + }, + { + "epoch": 0.13, + "learning_rate": 0.0006997632898696088, + "loss": 1.6351, + "theoretical_loss": 3.9958597269162315, + "tokens_seen": 442368000 + }, + { + "epoch": 0.13, + "learning_rate": 0.0006997472417251755, + "loss": 1.5008, + "theoretical_loss": 3.9957950843856542, + "tokens_seen": 442433536 + }, + { + "epoch": 0.13, + "learning_rate": 0.0006997311935807422, + "loss": 1.6059, + "theoretical_loss": 3.9957304541102396, + "tokens_seen": 442499072 + }, + { + "epoch": 0.13, + "learning_rate": 0.0006997151454363089, + "loss": 1.584, + "theoretical_loss": 3.995665836085848, + "tokens_seen": 442564608 + }, + { + "epoch": 0.13, + "learning_rate": 0.0006996990972918756, + "loss": 1.4876, + "theoretical_loss": 3.9956012303083446, + "tokens_seen": 442630144 + }, + { + "epoch": 0.13, + "learning_rate": 0.0006996830491474423, + "loss": 1.6349, + "theoretical_loss": 3.9955366367735947, + "tokens_seen": 442695680 + }, + { + "epoch": 0.13, + "learning_rate": 0.000699667001003009, + "loss": 1.5163, + "theoretical_loss": 3.995472055477467, + "tokens_seen": 442761216 + }, + { + "epoch": 0.13, + "learning_rate": 0.0006996509528585757, + "loss": 1.7073, + "theoretical_loss": 3.9954074864158313, + "tokens_seen": 442826752 + }, + { + "epoch": 0.13, + "learning_rate": 0.0006996349047141424, + "loss": 1.5442, + "theoretical_loss": 3.9953429295845604, + "tokens_seen": 442892288 + }, + { + "epoch": 0.13, + "learning_rate": 0.0006996188565697091, + "loss": 1.5491, + "theoretical_loss": 3.9952783849795277, + "tokens_seen": 442957824 + }, + { + "epoch": 0.13, + "learning_rate": 0.0006996028084252758, + "loss": 1.4275, + "theoretical_loss": 3.99521385259661, + "tokens_seen": 443023360 + }, + { + "epoch": 0.13, + "learning_rate": 0.0006995867602808425, + "loss": 1.754, + "theoretical_loss": 3.9951493324316845, + "tokens_seen": 443088896 + }, + { + "epoch": 0.13, + "learning_rate": 0.0006995707121364092, + "loss": 1.4668, + "theoretical_loss": 3.9950848244806316, + "tokens_seen": 443154432 + }, + { + "epoch": 0.13, + "learning_rate": 0.0006995546639919759, + "loss": 1.6169, + "theoretical_loss": 3.9950203287393347, + "tokens_seen": 443219968 + }, + { + "epoch": 0.13, + "learning_rate": 0.0006995386158475426, + "loss": 1.5071, + "theoretical_loss": 3.994955845203676, + "tokens_seen": 443285504 + }, + { + "epoch": 0.13, + "learning_rate": 0.0006995225677031093, + "loss": 1.6723, + "theoretical_loss": 3.994891373869543, + "tokens_seen": 443351040 + }, + { + "epoch": 0.13, + "learning_rate": 0.000699506519558676, + "loss": 1.5759, + "theoretical_loss": 3.9948269147328235, + "tokens_seen": 443416576 + }, + { + "epoch": 0.13, + "learning_rate": 0.0006994904714142427, + "loss": 1.6257, + "theoretical_loss": 3.994762467789407, + "tokens_seen": 443482112 + }, + { + "epoch": 0.13, + "learning_rate": 0.0006994744232698095, + "loss": 1.6933, + "theoretical_loss": 3.9946980330351862, + "tokens_seen": 443547648 + }, + { + "epoch": 0.13, + "learning_rate": 0.0006994583751253761, + "loss": 1.6031, + "theoretical_loss": 3.994633610466056, + "tokens_seen": 443613184 + }, + { + "epoch": 0.13, + "learning_rate": 0.0006994423269809429, + "loss": 1.5668, + "theoretical_loss": 3.994569200077911, + "tokens_seen": 443678720 + }, + { + "epoch": 0.13, + "learning_rate": 0.0006994262788365095, + "loss": 1.5945, + "theoretical_loss": 3.9945048018666496, + "tokens_seen": 443744256 + }, + { + "epoch": 0.13, + "learning_rate": 0.0006994102306920763, + "loss": 1.6115, + "theoretical_loss": 3.9944404158281723, + "tokens_seen": 443809792 + }, + { + "epoch": 0.13, + "learning_rate": 0.0006993941825476429, + "loss": 1.559, + "theoretical_loss": 3.994376041958381, + "tokens_seen": 443875328 + }, + { + "epoch": 0.13, + "learning_rate": 0.0006993781344032097, + "loss": 1.5844, + "theoretical_loss": 3.9943116802531797, + "tokens_seen": 443940864 + }, + { + "epoch": 0.13, + "learning_rate": 0.0006993620862587763, + "loss": 1.5402, + "theoretical_loss": 3.9942473307084745, + "tokens_seen": 444006400 + }, + { + "epoch": 0.13, + "learning_rate": 0.0006993460381143431, + "loss": 1.5794, + "theoretical_loss": 3.9941829933201736, + "tokens_seen": 444071936 + }, + { + "epoch": 0.13, + "learning_rate": 0.0006993299899699097, + "loss": 1.6646, + "theoretical_loss": 3.9941186680841865, + "tokens_seen": 444137472 + }, + { + "epoch": 0.13, + "learning_rate": 0.0006993139418254765, + "loss": 1.6818, + "theoretical_loss": 3.9940543549964254, + "tokens_seen": 444203008 + }, + { + "epoch": 0.13, + "learning_rate": 0.0006992978936810431, + "loss": 1.4692, + "theoretical_loss": 3.993990054052804, + "tokens_seen": 444268544 + }, + { + "epoch": 0.13, + "learning_rate": 0.0006992818455366099, + "loss": 1.4914, + "theoretical_loss": 3.993925765249239, + "tokens_seen": 444334080 + }, + { + "epoch": 0.13, + "learning_rate": 0.0006992657973921765, + "loss": 1.6031, + "theoretical_loss": 3.993861488581647, + "tokens_seen": 444399616 + }, + { + "epoch": 0.13, + "learning_rate": 0.0006992497492477433, + "loss": 1.7105, + "theoretical_loss": 3.9937972240459487, + "tokens_seen": 444465152 + }, + { + "epoch": 0.13, + "learning_rate": 0.0006992337011033099, + "loss": 1.5799, + "theoretical_loss": 3.9937329716380665, + "tokens_seen": 444530688 + }, + { + "epoch": 0.13, + "learning_rate": 0.0006992176529588767, + "loss": 1.6276, + "theoretical_loss": 3.9936687313539228, + "tokens_seen": 444596224 + }, + { + "epoch": 0.13, + "learning_rate": 0.0006992016048144433, + "loss": 1.5591, + "theoretical_loss": 3.9936045031894443, + "tokens_seen": 444661760 + }, + { + "epoch": 0.13, + "learning_rate": 0.0006991855566700101, + "loss": 1.5825, + "theoretical_loss": 3.9935402871405588, + "tokens_seen": 444727296 + }, + { + "epoch": 0.13, + "learning_rate": 0.0006991695085255767, + "loss": 1.5729, + "theoretical_loss": 3.9934760832031957, + "tokens_seen": 444792832 + }, + { + "epoch": 0.13, + "learning_rate": 0.0006991534603811435, + "loss": 1.6716, + "theoretical_loss": 3.9934118913732863, + "tokens_seen": 444858368 + }, + { + "epoch": 0.13, + "learning_rate": 0.0006991374122367101, + "loss": 1.6544, + "theoretical_loss": 3.9933477116467646, + "tokens_seen": 444923904 + }, + { + "epoch": 0.13, + "learning_rate": 0.0006991213640922769, + "loss": 1.7195, + "theoretical_loss": 3.993283544019567, + "tokens_seen": 444989440 + }, + { + "epoch": 0.13, + "learning_rate": 0.0006991053159478435, + "loss": 1.6172, + "theoretical_loss": 3.9932193884876304, + "tokens_seen": 445054976 + }, + { + "epoch": 0.13, + "learning_rate": 0.0006990892678034103, + "loss": 1.5853, + "theoretical_loss": 3.9931552450468937, + "tokens_seen": 445120512 + }, + { + "epoch": 0.13, + "learning_rate": 0.000699073219658977, + "loss": 1.5823, + "theoretical_loss": 3.9930911136933, + "tokens_seen": 445186048 + }, + { + "epoch": 0.13, + "learning_rate": 0.0006990571715145437, + "loss": 1.5594, + "theoretical_loss": 3.9930269944227907, + "tokens_seen": 445251584 + }, + { + "epoch": 0.13, + "learning_rate": 0.0006990411233701104, + "loss": 1.6403, + "theoretical_loss": 3.9929628872313128, + "tokens_seen": 445317120 + }, + { + "epoch": 0.13, + "learning_rate": 0.0006990250752256771, + "loss": 1.6085, + "theoretical_loss": 3.9928987921148127, + "tokens_seen": 445382656 + }, + { + "epoch": 0.13, + "learning_rate": 0.0006990090270812438, + "loss": 1.5611, + "theoretical_loss": 3.992834709069241, + "tokens_seen": 445448192 + }, + { + "epoch": 0.14, + "learning_rate": 0.0006989929789368105, + "loss": 1.5787, + "theoretical_loss": 3.9927706380905477, + "tokens_seen": 445513728 + }, + { + "epoch": 0.14, + "learning_rate": 0.0006989769307923772, + "loss": 1.5804, + "theoretical_loss": 3.992706579174686, + "tokens_seen": 445579264 + }, + { + "epoch": 0.14, + "objective/train/docs_used": 168097, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.7664058208465576, + "objective/train/theoretical_loss": 3.9926425323176122, + "objective/train/tokens_used": 466104800, + "theoretical_loss": 3.9926425323176122, + "tokens_seen": 445644800 + }, + { + "epoch": 0.14, + "learning_rate": 0.0006989608826479439, + "loss": 1.6841, + "theoretical_loss": 3.9926425323176122, + "tokens_seen": 445644800 + }, + { + "epoch": 0.14, + "learning_rate": 0.0006989448345035106, + "loss": 1.5435, + "theoretical_loss": 3.992578497515283, + "tokens_seen": 445710336 + }, + { + "epoch": 0.14, + "learning_rate": 0.0006989287863590773, + "loss": 1.6405, + "theoretical_loss": 3.9925144747636576, + "tokens_seen": 445775872 + }, + { + "epoch": 0.14, + "learning_rate": 0.000698912738214644, + "loss": 1.5314, + "theoretical_loss": 3.992450464058696, + "tokens_seen": 445841408 + }, + { + "epoch": 0.14, + "learning_rate": 0.0006988966900702107, + "loss": 1.5025, + "theoretical_loss": 3.9923864653963625, + "tokens_seen": 445906944 + }, + { + "epoch": 0.14, + "learning_rate": 0.0006988806419257774, + "loss": 1.5672, + "theoretical_loss": 3.9923224787726213, + "tokens_seen": 445972480 + }, + { + "epoch": 0.14, + "learning_rate": 0.0006988645937813441, + "loss": 1.5482, + "theoretical_loss": 3.9922585041834395, + "tokens_seen": 446038016 + }, + { + "epoch": 0.14, + "learning_rate": 0.0006988485456369108, + "loss": 1.5476, + "theoretical_loss": 3.9921945416247855, + "tokens_seen": 446103552 + }, + { + "epoch": 0.14, + "learning_rate": 0.0006988324974924775, + "loss": 1.5396, + "theoretical_loss": 3.9921305910926312, + "tokens_seen": 446169088 + }, + { + "epoch": 0.14, + "learning_rate": 0.0006988164493480442, + "loss": 1.5346, + "theoretical_loss": 3.992066652582948, + "tokens_seen": 446234624 + }, + { + "epoch": 0.14, + "learning_rate": 0.0006988004012036109, + "loss": 1.6036, + "theoretical_loss": 3.9920027260917115, + "tokens_seen": 446300160 + }, + { + "epoch": 0.14, + "learning_rate": 0.0006987843530591776, + "loss": 1.6545, + "theoretical_loss": 3.9919388116148977, + "tokens_seen": 446365696 + }, + { + "epoch": 0.14, + "learning_rate": 0.0006987683049147443, + "loss": 1.5502, + "theoretical_loss": 3.991874909148485, + "tokens_seen": 446431232 + }, + { + "epoch": 0.14, + "learning_rate": 0.000698752256770311, + "loss": 1.5739, + "theoretical_loss": 3.9918110186884546, + "tokens_seen": 446496768 + }, + { + "epoch": 0.14, + "learning_rate": 0.0006987362086258777, + "loss": 1.6284, + "theoretical_loss": 3.991747140230788, + "tokens_seen": 446562304 + }, + { + "epoch": 0.14, + "learning_rate": 0.0006987201604814444, + "loss": 1.7083, + "theoretical_loss": 3.9916832737714705, + "tokens_seen": 446627840 + }, + { + "epoch": 0.14, + "learning_rate": 0.0006987041123370111, + "loss": 1.6497, + "theoretical_loss": 3.9916194193064873, + "tokens_seen": 446693376 + }, + { + "epoch": 0.14, + "learning_rate": 0.0006986880641925778, + "loss": 1.6181, + "theoretical_loss": 3.9915555768318276, + "tokens_seen": 446758912 + }, + { + "epoch": 0.14, + "learning_rate": 0.0006986720160481445, + "loss": 1.5582, + "theoretical_loss": 3.9914917463434807, + "tokens_seen": 446824448 + }, + { + "epoch": 0.14, + "learning_rate": 0.0006986559679037112, + "loss": 1.6113, + "theoretical_loss": 3.9914279278374387, + "tokens_seen": 446889984 + }, + { + "epoch": 0.14, + "learning_rate": 0.0006986399197592779, + "loss": 1.5919, + "theoretical_loss": 3.9913641213096964, + "tokens_seen": 446955520 + }, + { + "epoch": 0.14, + "learning_rate": 0.0006986238716148446, + "loss": 1.5142, + "theoretical_loss": 3.9913003267562486, + "tokens_seen": 447021056 + }, + { + "epoch": 0.14, + "learning_rate": 0.0006986078234704113, + "loss": 1.5332, + "theoretical_loss": 3.9912365441730944, + "tokens_seen": 447086592 + }, + { + "epoch": 0.14, + "learning_rate": 0.000698591775325978, + "loss": 1.7142, + "theoretical_loss": 3.9911727735562326, + "tokens_seen": 447152128 + }, + { + "epoch": 0.14, + "learning_rate": 0.0006985757271815447, + "loss": 1.4204, + "theoretical_loss": 3.9911090149016646, + "tokens_seen": 447217664 + }, + { + "epoch": 0.14, + "learning_rate": 0.0006985596790371114, + "loss": 1.6616, + "theoretical_loss": 3.9910452682053954, + "tokens_seen": 447283200 + }, + { + "epoch": 0.14, + "learning_rate": 0.0006985436308926781, + "loss": 1.5401, + "theoretical_loss": 3.9909815334634287, + "tokens_seen": 447348736 + }, + { + "epoch": 0.14, + "learning_rate": 0.0006985275827482448, + "loss": 1.5094, + "theoretical_loss": 3.990917810671773, + "tokens_seen": 447414272 + }, + { + "epoch": 0.14, + "learning_rate": 0.0006985115346038115, + "loss": 1.55, + "theoretical_loss": 3.9908540998264375, + "tokens_seen": 447479808 + }, + { + "epoch": 0.14, + "learning_rate": 0.0006984954864593781, + "loss": 1.6938, + "theoretical_loss": 3.9907904009234336, + "tokens_seen": 447545344 + }, + { + "epoch": 0.14, + "learning_rate": 0.0006984794383149449, + "loss": 1.5668, + "theoretical_loss": 3.9907267139587743, + "tokens_seen": 447610880 + }, + { + "epoch": 0.14, + "learning_rate": 0.0006984633901705115, + "loss": 1.4666, + "theoretical_loss": 3.9906630389284747, + "tokens_seen": 447676416 + }, + { + "epoch": 0.14, + "learning_rate": 0.0006984473420260783, + "loss": 1.5181, + "theoretical_loss": 3.990599375828552, + "tokens_seen": 447741952 + }, + { + "epoch": 0.14, + "learning_rate": 0.0006984312938816449, + "loss": 1.5576, + "theoretical_loss": 3.9905357246550253, + "tokens_seen": 447807488 + }, + { + "epoch": 0.14, + "learning_rate": 0.0006984152457372117, + "loss": 1.5, + "theoretical_loss": 3.9904720854039137, + "tokens_seen": 447873024 + }, + { + "epoch": 0.14, + "learning_rate": 0.0006983991975927783, + "loss": 1.6459, + "theoretical_loss": 3.9904084580712427, + "tokens_seen": 447938560 + }, + { + "epoch": 0.14, + "learning_rate": 0.0006983831494483451, + "loss": 1.5311, + "theoretical_loss": 3.9903448426530352, + "tokens_seen": 448004096 + }, + { + "epoch": 0.14, + "learning_rate": 0.0006983671013039117, + "loss": 1.6949, + "theoretical_loss": 3.990281239145318, + "tokens_seen": 448069632 + }, + { + "epoch": 0.14, + "learning_rate": 0.0006983510531594784, + "loss": 1.5792, + "theoretical_loss": 3.99021764754412, + "tokens_seen": 448135168 + }, + { + "epoch": 0.14, + "learning_rate": 0.0006983350050150451, + "loss": 1.6015, + "theoretical_loss": 3.990154067845472, + "tokens_seen": 448200704 + }, + { + "epoch": 0.14, + "learning_rate": 0.0006983189568706118, + "loss": 1.6055, + "theoretical_loss": 3.990090500045404, + "tokens_seen": 448266240 + }, + { + "epoch": 0.14, + "learning_rate": 0.0006983029087261785, + "loss": 1.6272, + "theoretical_loss": 3.990026944139953, + "tokens_seen": 448331776 + }, + { + "epoch": 0.14, + "learning_rate": 0.0006982868605817452, + "loss": 1.4729, + "theoretical_loss": 3.989963400125153, + "tokens_seen": 448397312 + }, + { + "epoch": 0.14, + "learning_rate": 0.0006982708124373119, + "loss": 1.5439, + "theoretical_loss": 3.9898998679970434, + "tokens_seen": 448462848 + }, + { + "epoch": 0.14, + "learning_rate": 0.0006982547642928786, + "loss": 1.5467, + "theoretical_loss": 3.9898363477516634, + "tokens_seen": 448528384 + }, + { + "epoch": 0.14, + "learning_rate": 0.0006982387161484453, + "loss": 1.6329, + "theoretical_loss": 3.9897728393850547, + "tokens_seen": 448593920 + }, + { + "epoch": 0.14, + "learning_rate": 0.000698222668004012, + "loss": 1.5053, + "theoretical_loss": 3.9897093428932613, + "tokens_seen": 448659456 + }, + { + "epoch": 0.14, + "learning_rate": 0.0006982066198595787, + "loss": 1.4964, + "theoretical_loss": 3.989645858272328, + "tokens_seen": 448724992 + }, + { + "epoch": 0.14, + "learning_rate": 0.0006981905717151454, + "loss": 1.5293, + "theoretical_loss": 3.989582385518304, + "tokens_seen": 448790528 + }, + { + "epoch": 0.14, + "learning_rate": 0.0006981745235707121, + "loss": 1.5651, + "theoretical_loss": 3.989518924627237, + "tokens_seen": 448856064 + }, + { + "epoch": 0.14, + "objective/train/docs_used": 169309, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.6712148189544678, + "objective/train/theoretical_loss": 3.9894554755951788, + "objective/train/tokens_used": 469381600, + "theoretical_loss": 3.9894554755951788, + "tokens_seen": 448921600 + }, + { + "epoch": 0.14, + "learning_rate": 0.0006981584754262788, + "loss": 1.6061, + "theoretical_loss": 3.9894554755951788, + "tokens_seen": 448921600 + }, + { + "epoch": 0.14, + "learning_rate": 0.0006981424272818456, + "loss": 1.5903, + "theoretical_loss": 3.989392038418182, + "tokens_seen": 448987136 + }, + { + "epoch": 0.14, + "learning_rate": 0.0006981263791374122, + "loss": 1.4495, + "theoretical_loss": 3.9893286130923027, + "tokens_seen": 449052672 + }, + { + "epoch": 0.14, + "learning_rate": 0.000698110330992979, + "loss": 1.6695, + "theoretical_loss": 3.989265199613597, + "tokens_seen": 449118208 + }, + { + "epoch": 0.14, + "learning_rate": 0.0006980942828485456, + "loss": 1.5319, + "theoretical_loss": 3.989201797978124, + "tokens_seen": 449183744 + }, + { + "epoch": 0.14, + "learning_rate": 0.0006980782347041124, + "loss": 1.4989, + "theoretical_loss": 3.989138408181944, + "tokens_seen": 449249280 + }, + { + "epoch": 0.14, + "learning_rate": 0.000698062186559679, + "loss": 1.5846, + "theoretical_loss": 3.9890750302211204, + "tokens_seen": 449314816 + }, + { + "epoch": 0.14, + "learning_rate": 0.0006980461384152458, + "loss": 1.5821, + "theoretical_loss": 3.9890116640917173, + "tokens_seen": 449380352 + }, + { + "epoch": 0.14, + "learning_rate": 0.0006980300902708124, + "loss": 1.6644, + "theoretical_loss": 3.9889483097898006, + "tokens_seen": 449445888 + }, + { + "epoch": 0.14, + "learning_rate": 0.0006980140421263792, + "loss": 1.6584, + "theoretical_loss": 3.9888849673114386, + "tokens_seen": 449511424 + }, + { + "epoch": 0.14, + "learning_rate": 0.0006979979939819458, + "loss": 1.6387, + "theoretical_loss": 3.988821636652702, + "tokens_seen": 449576960 + }, + { + "epoch": 0.14, + "learning_rate": 0.0006979819458375126, + "loss": 1.5079, + "theoretical_loss": 3.988758317809662, + "tokens_seen": 449642496 + }, + { + "epoch": 0.14, + "learning_rate": 0.0006979658976930792, + "loss": 1.5727, + "theoretical_loss": 3.988695010778393, + "tokens_seen": 449708032 + }, + { + "epoch": 0.14, + "learning_rate": 0.000697949849548646, + "loss": 1.6666, + "theoretical_loss": 3.9886317155549706, + "tokens_seen": 449773568 + }, + { + "epoch": 0.14, + "learning_rate": 0.0006979338014042126, + "loss": 1.5371, + "theoretical_loss": 3.988568432135472, + "tokens_seen": 449839104 + }, + { + "epoch": 0.14, + "learning_rate": 0.0006979177532597794, + "loss": 1.5821, + "theoretical_loss": 3.988505160515978, + "tokens_seen": 449904640 + }, + { + "epoch": 0.14, + "learning_rate": 0.000697901705115346, + "loss": 1.6404, + "theoretical_loss": 3.9884419006925684, + "tokens_seen": 449970176 + }, + { + "epoch": 0.14, + "learning_rate": 0.0006978856569709128, + "loss": 1.5418, + "theoretical_loss": 3.9883786526613267, + "tokens_seen": 450035712 + }, + { + "epoch": 0.14, + "learning_rate": 0.0006978696088264795, + "loss": 1.5569, + "theoretical_loss": 3.988315416418339, + "tokens_seen": 450101248 + }, + { + "epoch": 0.14, + "learning_rate": 0.0006978535606820462, + "loss": 1.5843, + "theoretical_loss": 3.988252191959692, + "tokens_seen": 450166784 + }, + { + "epoch": 0.14, + "learning_rate": 0.0006978375125376129, + "loss": 1.6824, + "theoretical_loss": 3.9881889792814733, + "tokens_seen": 450232320 + }, + { + "epoch": 0.14, + "learning_rate": 0.0006978214643931796, + "loss": 1.6427, + "theoretical_loss": 3.9881257783797746, + "tokens_seen": 450297856 + }, + { + "epoch": 0.14, + "learning_rate": 0.0006978054162487463, + "loss": 1.5661, + "theoretical_loss": 3.9880625892506885, + "tokens_seen": 450363392 + }, + { + "epoch": 0.14, + "learning_rate": 0.000697789368104313, + "loss": 1.606, + "theoretical_loss": 3.9879994118903093, + "tokens_seen": 450428928 + }, + { + "epoch": 0.14, + "learning_rate": 0.0006977733199598797, + "loss": 1.6257, + "theoretical_loss": 3.9879362462947334, + "tokens_seen": 450494464 + }, + { + "epoch": 0.14, + "learning_rate": 0.0006977572718154464, + "loss": 1.6491, + "theoretical_loss": 3.9878730924600583, + "tokens_seen": 450560000 + }, + { + "epoch": 0.14, + "learning_rate": 0.0006977412236710131, + "loss": 1.6175, + "theoretical_loss": 3.9878099503823847, + "tokens_seen": 450625536 + }, + { + "epoch": 0.14, + "learning_rate": 0.0006977251755265798, + "loss": 1.5898, + "theoretical_loss": 3.9877468200578146, + "tokens_seen": 450691072 + }, + { + "epoch": 0.14, + "learning_rate": 0.0006977091273821465, + "loss": 1.6092, + "theoretical_loss": 3.9876837014824513, + "tokens_seen": 450756608 + }, + { + "epoch": 0.14, + "learning_rate": 0.0006976930792377132, + "loss": 1.6039, + "theoretical_loss": 3.9876205946524, + "tokens_seen": 450822144 + }, + { + "epoch": 0.14, + "learning_rate": 0.0006976770310932799, + "loss": 1.5258, + "theoretical_loss": 3.9875574995637697, + "tokens_seen": 450887680 + }, + { + "epoch": 0.14, + "learning_rate": 0.0006976609829488466, + "loss": 1.5889, + "theoretical_loss": 3.987494416212668, + "tokens_seen": 450953216 + }, + { + "epoch": 0.14, + "learning_rate": 0.0006976449348044133, + "loss": 1.5922, + "theoretical_loss": 3.987431344595207, + "tokens_seen": 451018752 + }, + { + "epoch": 0.14, + "learning_rate": 0.00069762888665998, + "loss": 1.4424, + "theoretical_loss": 3.9873682847074994, + "tokens_seen": 451084288 + }, + { + "epoch": 0.14, + "learning_rate": 0.0006976128385155467, + "loss": 1.5776, + "theoretical_loss": 3.9873052365456605, + "tokens_seen": 451149824 + }, + { + "epoch": 0.14, + "learning_rate": 0.0006975967903711134, + "loss": 1.4422, + "theoretical_loss": 3.9872422001058063, + "tokens_seen": 451215360 + }, + { + "epoch": 0.14, + "learning_rate": 0.0006975807422266801, + "loss": 1.5465, + "theoretical_loss": 3.987179175384056, + "tokens_seen": 451280896 + }, + { + "epoch": 0.14, + "learning_rate": 0.0006975646940822468, + "loss": 1.5387, + "theoretical_loss": 3.9871161623765294, + "tokens_seen": 451346432 + }, + { + "epoch": 0.14, + "learning_rate": 0.0006975486459378135, + "loss": 1.6127, + "theoretical_loss": 3.9870531610793494, + "tokens_seen": 451411968 + }, + { + "epoch": 0.14, + "learning_rate": 0.0006975325977933802, + "loss": 1.5919, + "theoretical_loss": 3.9869901714886398, + "tokens_seen": 451477504 + }, + { + "epoch": 0.14, + "learning_rate": 0.0006975165496489469, + "loss": 1.5275, + "theoretical_loss": 3.9869271936005264, + "tokens_seen": 451543040 + }, + { + "epoch": 0.14, + "learning_rate": 0.0006975005015045136, + "loss": 1.6097, + "theoretical_loss": 3.9868642274111377, + "tokens_seen": 451608576 + }, + { + "epoch": 0.14, + "learning_rate": 0.0006974844533600803, + "loss": 1.5718, + "theoretical_loss": 3.986801272916603, + "tokens_seen": 451674112 + }, + { + "epoch": 0.14, + "learning_rate": 0.000697468405215647, + "loss": 1.5826, + "theoretical_loss": 3.9867383301130532, + "tokens_seen": 451739648 + }, + { + "epoch": 0.14, + "learning_rate": 0.0006974523570712137, + "loss": 1.5749, + "theoretical_loss": 3.9866753989966224, + "tokens_seen": 451805184 + }, + { + "epoch": 0.14, + "learning_rate": 0.0006974363089267804, + "loss": 1.7111, + "theoretical_loss": 3.986612479563445, + "tokens_seen": 451870720 + }, + { + "epoch": 0.14, + "learning_rate": 0.0006974202607823471, + "loss": 1.6483, + "theoretical_loss": 3.9865495718096584, + "tokens_seen": 451936256 + }, + { + "epoch": 0.14, + "learning_rate": 0.0006974042126379138, + "loss": 1.5658, + "theoretical_loss": 3.9864866757314017, + "tokens_seen": 452001792 + }, + { + "epoch": 0.14, + "learning_rate": 0.0006973881644934805, + "loss": 1.5975, + "theoretical_loss": 3.9864237913248157, + "tokens_seen": 452067328 + }, + { + "epoch": 0.14, + "learning_rate": 0.0006973721163490472, + "loss": 1.6068, + "theoretical_loss": 3.9863609185860422, + "tokens_seen": 452132864 + }, + { + "epoch": 0.14, + "objective/train/docs_used": 170598, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.6789288520812988, + "objective/train/theoretical_loss": 3.9862980575112257, + "objective/train/tokens_used": 472658400, + "theoretical_loss": 3.9862980575112257, + "tokens_seen": 452198400 + }, + { + "epoch": 0.14, + "learning_rate": 0.0006973560682046139, + "loss": 1.6308, + "theoretical_loss": 3.9862980575112257, + "tokens_seen": 452198400 + }, + { + "epoch": 0.14, + "learning_rate": 0.0006973400200601806, + "loss": 1.6516, + "theoretical_loss": 3.986235208096513, + "tokens_seen": 452263936 + }, + { + "epoch": 0.14, + "learning_rate": 0.0006973239719157473, + "loss": 1.6448, + "theoretical_loss": 3.9861723703380516, + "tokens_seen": 452329472 + }, + { + "epoch": 0.14, + "learning_rate": 0.000697307923771314, + "loss": 1.6479, + "theoretical_loss": 3.986109544231992, + "tokens_seen": 452395008 + }, + { + "epoch": 0.14, + "learning_rate": 0.0006972918756268807, + "loss": 1.6266, + "theoretical_loss": 3.9860467297744844, + "tokens_seen": 452460544 + }, + { + "epoch": 0.14, + "learning_rate": 0.0006972758274824474, + "loss": 1.5518, + "theoretical_loss": 3.985983926961684, + "tokens_seen": 452526080 + }, + { + "epoch": 0.14, + "learning_rate": 0.0006972597793380141, + "loss": 1.5958, + "theoretical_loss": 3.985921135789745, + "tokens_seen": 452591616 + }, + { + "epoch": 0.14, + "learning_rate": 0.0006972437311935808, + "loss": 1.5059, + "theoretical_loss": 3.9858583562548247, + "tokens_seen": 452657152 + }, + { + "epoch": 0.14, + "learning_rate": 0.0006972276830491475, + "loss": 1.6628, + "theoretical_loss": 3.9857955883530827, + "tokens_seen": 452722688 + }, + { + "epoch": 0.14, + "learning_rate": 0.0006972116349047142, + "loss": 1.6523, + "theoretical_loss": 3.985732832080679, + "tokens_seen": 452788224 + }, + { + "epoch": 0.14, + "learning_rate": 0.0006971955867602809, + "loss": 1.5729, + "theoretical_loss": 3.985670087433777, + "tokens_seen": 452853760 + }, + { + "epoch": 0.14, + "learning_rate": 0.0006971795386158476, + "loss": 1.632, + "theoretical_loss": 3.9856073544085406, + "tokens_seen": 452919296 + }, + { + "epoch": 0.14, + "learning_rate": 0.0006971634904714143, + "loss": 1.7177, + "theoretical_loss": 3.9855446330011364, + "tokens_seen": 452984832 + }, + { + "epoch": 0.14, + "learning_rate": 0.000697147442326981, + "loss": 1.6336, + "theoretical_loss": 3.9854819232077316, + "tokens_seen": 453050368 + }, + { + "epoch": 0.14, + "learning_rate": 0.0006971313941825477, + "loss": 1.5878, + "theoretical_loss": 3.985419225024498, + "tokens_seen": 453115904 + }, + { + "epoch": 0.14, + "learning_rate": 0.0006971153460381144, + "loss": 1.6536, + "theoretical_loss": 3.985356538447605, + "tokens_seen": 453181440 + }, + { + "epoch": 0.14, + "learning_rate": 0.000697099297893681, + "loss": 1.7251, + "theoretical_loss": 3.9852938634732276, + "tokens_seen": 453246976 + }, + { + "epoch": 0.14, + "learning_rate": 0.0006970832497492478, + "loss": 1.6154, + "theoretical_loss": 3.985231200097541, + "tokens_seen": 453312512 + }, + { + "epoch": 0.14, + "learning_rate": 0.0006970672016048144, + "loss": 1.5602, + "theoretical_loss": 3.9851685483167225, + "tokens_seen": 453378048 + }, + { + "epoch": 0.14, + "learning_rate": 0.0006970511534603812, + "loss": 1.559, + "theoretical_loss": 3.9851059081269504, + "tokens_seen": 453443584 + }, + { + "epoch": 0.14, + "learning_rate": 0.0006970351053159478, + "loss": 1.5312, + "theoretical_loss": 3.9850432795244055, + "tokens_seen": 453509120 + }, + { + "epoch": 0.14, + "learning_rate": 0.0006970190571715146, + "loss": 1.5035, + "theoretical_loss": 3.9849806625052713, + "tokens_seen": 453574656 + }, + { + "epoch": 0.14, + "learning_rate": 0.0006970030090270812, + "loss": 1.5081, + "theoretical_loss": 3.984918057065731, + "tokens_seen": 453640192 + }, + { + "epoch": 0.14, + "learning_rate": 0.000696986960882648, + "loss": 1.5428, + "theoretical_loss": 3.984855463201972, + "tokens_seen": 453705728 + }, + { + "epoch": 0.14, + "learning_rate": 0.0006969709127382146, + "loss": 1.5082, + "theoretical_loss": 3.984792880910182, + "tokens_seen": 453771264 + }, + { + "epoch": 0.14, + "learning_rate": 0.0006969548645937813, + "loss": 1.6278, + "theoretical_loss": 3.98473031018655, + "tokens_seen": 453836800 + }, + { + "epoch": 0.14, + "learning_rate": 0.000696938816449348, + "loss": 1.5918, + "theoretical_loss": 3.984667751027268, + "tokens_seen": 453902336 + }, + { + "epoch": 0.14, + "learning_rate": 0.0006969227683049147, + "loss": 1.5387, + "theoretical_loss": 3.98460520342853, + "tokens_seen": 453967872 + }, + { + "epoch": 0.14, + "learning_rate": 0.0006969067201604814, + "loss": 1.5984, + "theoretical_loss": 3.984542667386531, + "tokens_seen": 454033408 + }, + { + "epoch": 0.14, + "learning_rate": 0.0006968906720160481, + "loss": 1.5571, + "theoretical_loss": 3.9844801428974685, + "tokens_seen": 454098944 + }, + { + "epoch": 0.14, + "learning_rate": 0.0006968746238716148, + "loss": 1.5924, + "theoretical_loss": 3.98441762995754, + "tokens_seen": 454164480 + }, + { + "epoch": 0.14, + "learning_rate": 0.0006968585757271815, + "loss": 1.5728, + "theoretical_loss": 3.984355128562947, + "tokens_seen": 454230016 + }, + { + "epoch": 0.14, + "learning_rate": 0.0006968425275827482, + "loss": 1.6776, + "theoretical_loss": 3.984292638709892, + "tokens_seen": 454295552 + }, + { + "epoch": 0.14, + "learning_rate": 0.0006968264794383149, + "loss": 1.5176, + "theoretical_loss": 3.9842301603945787, + "tokens_seen": 454361088 + }, + { + "epoch": 0.14, + "learning_rate": 0.0006968104312938816, + "loss": 1.5786, + "theoretical_loss": 3.9841676936132133, + "tokens_seen": 454426624 + }, + { + "epoch": 0.14, + "learning_rate": 0.0006967943831494483, + "loss": 1.5415, + "theoretical_loss": 3.984105238362005, + "tokens_seen": 454492160 + }, + { + "epoch": 0.14, + "learning_rate": 0.000696778335005015, + "loss": 1.5418, + "theoretical_loss": 3.984042794637161, + "tokens_seen": 454557696 + }, + { + "epoch": 0.14, + "learning_rate": 0.0006967622868605817, + "loss": 1.7188, + "theoretical_loss": 3.9839803624348944, + "tokens_seen": 454623232 + }, + { + "epoch": 0.14, + "learning_rate": 0.0006967462387161485, + "loss": 1.527, + "theoretical_loss": 3.983917941751418, + "tokens_seen": 454688768 + }, + { + "epoch": 0.14, + "learning_rate": 0.0006967301905717151, + "loss": 1.6061, + "theoretical_loss": 3.983855532582947, + "tokens_seen": 454754304 + }, + { + "epoch": 0.14, + "learning_rate": 0.0006967141424272819, + "loss": 1.6029, + "theoretical_loss": 3.983793134925697, + "tokens_seen": 454819840 + }, + { + "epoch": 0.14, + "learning_rate": 0.0006966980942828485, + "loss": 1.5022, + "theoretical_loss": 3.9837307487758884, + "tokens_seen": 454885376 + }, + { + "epoch": 0.14, + "learning_rate": 0.0006966820461384153, + "loss": 1.5934, + "theoretical_loss": 3.9836683741297403, + "tokens_seen": 454950912 + }, + { + "epoch": 0.14, + "learning_rate": 0.000696665997993982, + "loss": 1.5375, + "theoretical_loss": 3.9836060109834746, + "tokens_seen": 455016448 + }, + { + "epoch": 0.14, + "learning_rate": 0.0006966499498495487, + "loss": 1.5445, + "theoretical_loss": 3.983543659333316, + "tokens_seen": 455081984 + }, + { + "epoch": 0.14, + "learning_rate": 0.0006966339017051154, + "loss": 1.6156, + "theoretical_loss": 3.9834813191754908, + "tokens_seen": 455147520 + }, + { + "epoch": 0.14, + "learning_rate": 0.0006966178535606821, + "loss": 1.6466, + "theoretical_loss": 3.983418990506225, + "tokens_seen": 455213056 + }, + { + "epoch": 0.14, + "learning_rate": 0.0006966018054162488, + "loss": 1.5652, + "theoretical_loss": 3.9833566733217483, + "tokens_seen": 455278592 + }, + { + "epoch": 0.14, + "learning_rate": 0.0006965857572718155, + "loss": 1.6271, + "theoretical_loss": 3.983294367618292, + "tokens_seen": 455344128 + }, + { + "epoch": 0.14, + "learning_rate": 0.0006965697091273822, + "loss": 1.5991, + "theoretical_loss": 3.983232073392089, + "tokens_seen": 455409664 + }, + { + "epoch": 0.14, + "objective/train/docs_used": 171735, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.6367623805999756, + "objective/train/theoretical_loss": 3.983169790639374, + "objective/train/tokens_used": 475935200, + "theoretical_loss": 3.983169790639374, + "tokens_seen": 455475200 + }, + { + "epoch": 0.14, + "learning_rate": 0.0006965536609829489, + "loss": 1.6848, + "theoretical_loss": 3.983169790639374, + "tokens_seen": 455475200 + }, + { + "epoch": 0.14, + "learning_rate": 0.0006965376128385156, + "loss": 1.5591, + "theoretical_loss": 3.983107519356383, + "tokens_seen": 455540736 + }, + { + "epoch": 0.14, + "learning_rate": 0.0006965215646940823, + "loss": 1.6615, + "theoretical_loss": 3.9830452595393546, + "tokens_seen": 455606272 + }, + { + "epoch": 0.14, + "learning_rate": 0.000696505516549649, + "loss": 1.5955, + "theoretical_loss": 3.9829830111845284, + "tokens_seen": 455671808 + }, + { + "epoch": 0.14, + "learning_rate": 0.0006964894684052157, + "loss": 1.6063, + "theoretical_loss": 3.9829207742881456, + "tokens_seen": 455737344 + }, + { + "epoch": 0.14, + "learning_rate": 0.0006964734202607824, + "loss": 1.575, + "theoretical_loss": 3.982858548846451, + "tokens_seen": 455802880 + }, + { + "epoch": 0.14, + "learning_rate": 0.0006964573721163491, + "loss": 1.6225, + "theoretical_loss": 3.982796334855689, + "tokens_seen": 455868416 + }, + { + "epoch": 0.14, + "learning_rate": 0.0006964413239719158, + "loss": 1.5625, + "theoretical_loss": 3.9827341323121064, + "tokens_seen": 455933952 + }, + { + "epoch": 0.14, + "learning_rate": 0.0006964252758274825, + "loss": 1.6262, + "theoretical_loss": 3.9826719412119527, + "tokens_seen": 455999488 + }, + { + "epoch": 0.14, + "learning_rate": 0.0006964092276830492, + "loss": 1.5262, + "theoretical_loss": 3.982609761551478, + "tokens_seen": 456065024 + }, + { + "epoch": 0.14, + "learning_rate": 0.0006963931795386159, + "loss": 1.575, + "theoretical_loss": 3.982547593326934, + "tokens_seen": 456130560 + }, + { + "epoch": 0.14, + "learning_rate": 0.0006963771313941826, + "loss": 1.5707, + "theoretical_loss": 3.982485436534576, + "tokens_seen": 456196096 + }, + { + "epoch": 0.14, + "learning_rate": 0.0006963610832497493, + "loss": 1.5514, + "theoretical_loss": 3.98242329117066, + "tokens_seen": 456261632 + }, + { + "epoch": 0.14, + "learning_rate": 0.000696345035105316, + "loss": 1.6324, + "theoretical_loss": 3.982361157231442, + "tokens_seen": 456327168 + }, + { + "epoch": 0.14, + "learning_rate": 0.0006963289869608827, + "loss": 1.547, + "theoretical_loss": 3.982299034713183, + "tokens_seen": 456392704 + }, + { + "epoch": 0.14, + "learning_rate": 0.0006963129388164494, + "loss": 1.5563, + "theoretical_loss": 3.9822369236121427, + "tokens_seen": 456458240 + }, + { + "epoch": 0.14, + "learning_rate": 0.0006962968906720161, + "loss": 1.4718, + "theoretical_loss": 3.982174823924586, + "tokens_seen": 456523776 + }, + { + "epoch": 0.14, + "learning_rate": 0.0006962808425275828, + "loss": 1.6397, + "theoretical_loss": 3.9821127356467754, + "tokens_seen": 456589312 + }, + { + "epoch": 0.14, + "learning_rate": 0.0006962647943831495, + "loss": 1.5661, + "theoretical_loss": 3.9820506587749787, + "tokens_seen": 456654848 + }, + { + "epoch": 0.14, + "learning_rate": 0.0006962487462387162, + "loss": 1.445, + "theoretical_loss": 3.9819885933054637, + "tokens_seen": 456720384 + }, + { + "epoch": 0.14, + "learning_rate": 0.0006962326980942829, + "loss": 1.5401, + "theoretical_loss": 3.9819265392345, + "tokens_seen": 456785920 + }, + { + "epoch": 0.14, + "learning_rate": 0.0006962166499498496, + "loss": 1.5153, + "theoretical_loss": 3.9818644965583596, + "tokens_seen": 456851456 + }, + { + "epoch": 0.14, + "learning_rate": 0.0006962006018054163, + "loss": 1.5113, + "theoretical_loss": 3.981802465273316, + "tokens_seen": 456916992 + }, + { + "epoch": 0.14, + "learning_rate": 0.000696184553660983, + "loss": 1.556, + "theoretical_loss": 3.9817404453756446, + "tokens_seen": 456982528 + }, + { + "epoch": 0.14, + "learning_rate": 0.0006961685055165497, + "loss": 1.634, + "theoretical_loss": 3.981678436861622, + "tokens_seen": 457048064 + }, + { + "epoch": 0.14, + "learning_rate": 0.0006961524573721164, + "loss": 1.5483, + "theoretical_loss": 3.981616439727527, + "tokens_seen": 457113600 + }, + { + "epoch": 0.14, + "learning_rate": 0.0006961364092276831, + "loss": 1.6356, + "theoretical_loss": 3.9815544539696397, + "tokens_seen": 457179136 + }, + { + "epoch": 0.14, + "learning_rate": 0.0006961203610832498, + "loss": 1.6669, + "theoretical_loss": 3.981492479584243, + "tokens_seen": 457244672 + }, + { + "epoch": 0.14, + "learning_rate": 0.0006961043129388165, + "loss": 1.5676, + "theoretical_loss": 3.9814305165676203, + "tokens_seen": 457310208 + }, + { + "epoch": 0.14, + "learning_rate": 0.0006960882647943832, + "loss": 1.5604, + "theoretical_loss": 3.9813685649160577, + "tokens_seen": 457375744 + }, + { + "epoch": 0.14, + "learning_rate": 0.0006960722166499499, + "loss": 1.4125, + "theoretical_loss": 3.981306624625842, + "tokens_seen": 457441280 + }, + { + "epoch": 0.14, + "learning_rate": 0.0006960561685055166, + "loss": 1.6351, + "theoretical_loss": 3.9812446956932632, + "tokens_seen": 457506816 + }, + { + "epoch": 0.14, + "learning_rate": 0.0006960401203610833, + "loss": 1.6956, + "theoretical_loss": 3.9811827781146123, + "tokens_seen": 457572352 + }, + { + "epoch": 0.14, + "learning_rate": 0.00069602407221665, + "loss": 1.5702, + "theoretical_loss": 3.981120871886181, + "tokens_seen": 457637888 + }, + { + "epoch": 0.14, + "learning_rate": 0.0006960080240722167, + "loss": 1.5955, + "theoretical_loss": 3.9810589770042637, + "tokens_seen": 457703424 + }, + { + "epoch": 0.14, + "learning_rate": 0.0006959919759277834, + "loss": 1.5879, + "theoretical_loss": 3.9809970934651577, + "tokens_seen": 457768960 + }, + { + "epoch": 0.14, + "learning_rate": 0.0006959759277833501, + "loss": 1.6368, + "theoretical_loss": 3.9809352212651596, + "tokens_seen": 457834496 + }, + { + "epoch": 0.14, + "learning_rate": 0.0006959598796389168, + "loss": 1.6342, + "theoretical_loss": 3.9808733604005706, + "tokens_seen": 457900032 + }, + { + "epoch": 0.14, + "learning_rate": 0.0006959438314944835, + "loss": 1.5774, + "theoretical_loss": 3.9808115108676905, + "tokens_seen": 457965568 + }, + { + "epoch": 0.14, + "learning_rate": 0.0006959277833500502, + "loss": 1.6235, + "theoretical_loss": 3.9807496726628235, + "tokens_seen": 458031104 + }, + { + "epoch": 0.14, + "learning_rate": 0.0006959117352056169, + "loss": 1.501, + "theoretical_loss": 3.9806878457822736, + "tokens_seen": 458096640 + }, + { + "epoch": 0.14, + "learning_rate": 0.0006958956870611836, + "loss": 1.6025, + "theoretical_loss": 3.9806260302223477, + "tokens_seen": 458162176 + }, + { + "epoch": 0.14, + "learning_rate": 0.0006958796389167503, + "loss": 1.6974, + "theoretical_loss": 3.9805642259793546, + "tokens_seen": 458227712 + }, + { + "epoch": 0.14, + "learning_rate": 0.000695863590772317, + "loss": 1.7237, + "theoretical_loss": 3.9805024330496037, + "tokens_seen": 458293248 + }, + { + "epoch": 0.14, + "learning_rate": 0.0006958475426278837, + "loss": 1.611, + "theoretical_loss": 3.980440651429407, + "tokens_seen": 458358784 + }, + { + "epoch": 0.14, + "learning_rate": 0.0006958314944834504, + "loss": 1.6436, + "theoretical_loss": 3.9803788811150778, + "tokens_seen": 458424320 + }, + { + "epoch": 0.14, + "learning_rate": 0.0006958154463390171, + "loss": 1.5742, + "theoretical_loss": 3.980317122102931, + "tokens_seen": 458489856 + }, + { + "epoch": 0.14, + "learning_rate": 0.0006957993981945838, + "loss": 1.5661, + "theoretical_loss": 3.9802553743892846, + "tokens_seen": 458555392 + }, + { + "epoch": 0.14, + "learning_rate": 0.0006957833500501505, + "loss": 1.5478, + "theoretical_loss": 3.9801936379704568, + "tokens_seen": 458620928 + }, + { + "epoch": 0.14, + "learning_rate": 0.0006957673019057172, + "loss": 1.639, + "theoretical_loss": 3.9801319128427677, + "tokens_seen": 458686464 + }, + { + "epoch": 0.14, + "objective/train/docs_used": 173009, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.5537326335906982, + "objective/train/theoretical_loss": 3.9800701990025393, + "objective/train/tokens_used": 479212000, + "theoretical_loss": 3.9800701990025393, + "tokens_seen": 458752000 + }, + { + "epoch": 0.14, + "learning_rate": 0.0006957512537612839, + "loss": 1.525, + "theoretical_loss": 3.9800701990025393, + "tokens_seen": 458752000 + }, + { + "epoch": 0.14, + "learning_rate": 0.0006957352056168506, + "loss": 1.6425, + "theoretical_loss": 3.9800084964460964, + "tokens_seen": 458817536 + }, + { + "epoch": 0.14, + "learning_rate": 0.0006957191574724173, + "loss": 1.4844, + "theoretical_loss": 3.9799468051697633, + "tokens_seen": 458883072 + }, + { + "epoch": 0.14, + "learning_rate": 0.0006957031093279839, + "loss": 1.5192, + "theoretical_loss": 3.9798851251698686, + "tokens_seen": 458948608 + }, + { + "epoch": 0.14, + "learning_rate": 0.0006956870611835507, + "loss": 1.4638, + "theoretical_loss": 3.97982345644274, + "tokens_seen": 459014144 + }, + { + "epoch": 0.14, + "learning_rate": 0.0006956710130391173, + "loss": 1.6192, + "theoretical_loss": 3.9797617989847094, + "tokens_seen": 459079680 + }, + { + "epoch": 0.14, + "learning_rate": 0.0006956549648946841, + "loss": 1.6405, + "theoretical_loss": 3.979700152792108, + "tokens_seen": 459145216 + }, + { + "epoch": 0.14, + "learning_rate": 0.0006956389167502507, + "loss": 1.6005, + "theoretical_loss": 3.9796385178612717, + "tokens_seen": 459210752 + }, + { + "epoch": 0.14, + "learning_rate": 0.0006956228686058175, + "loss": 1.5473, + "theoretical_loss": 3.9795768941885346, + "tokens_seen": 459276288 + }, + { + "epoch": 0.14, + "learning_rate": 0.0006956068204613841, + "loss": 1.6124, + "theoretical_loss": 3.9795152817702357, + "tokens_seen": 459341824 + }, + { + "epoch": 0.14, + "learning_rate": 0.0006955907723169509, + "loss": 1.5664, + "theoretical_loss": 3.979453680602713, + "tokens_seen": 459407360 + }, + { + "epoch": 0.14, + "learning_rate": 0.0006955747241725175, + "loss": 1.6529, + "theoretical_loss": 3.9793920906823086, + "tokens_seen": 459472896 + }, + { + "epoch": 0.14, + "learning_rate": 0.0006955586760280842, + "loss": 1.4902, + "theoretical_loss": 3.979330512005365, + "tokens_seen": 459538432 + }, + { + "epoch": 0.14, + "learning_rate": 0.0006955426278836509, + "loss": 1.5828, + "theoretical_loss": 3.9792689445682257, + "tokens_seen": 459603968 + }, + { + "epoch": 0.14, + "learning_rate": 0.0006955265797392176, + "loss": 1.5298, + "theoretical_loss": 3.9792073883672385, + "tokens_seen": 459669504 + }, + { + "epoch": 0.14, + "learning_rate": 0.0006955105315947843, + "loss": 1.5899, + "theoretical_loss": 3.9791458433987503, + "tokens_seen": 459735040 + }, + { + "epoch": 0.14, + "learning_rate": 0.0006954944834503511, + "loss": 1.6265, + "theoretical_loss": 3.97908430965911, + "tokens_seen": 459800576 + }, + { + "epoch": 0.14, + "learning_rate": 0.0006954784353059177, + "loss": 1.6322, + "theoretical_loss": 3.979022787144671, + "tokens_seen": 459866112 + }, + { + "epoch": 0.14, + "learning_rate": 0.0006954623871614845, + "loss": 1.6124, + "theoretical_loss": 3.9789612758517836, + "tokens_seen": 459931648 + }, + { + "epoch": 0.14, + "learning_rate": 0.0006954463390170511, + "loss": 1.6273, + "theoretical_loss": 3.9788997757768043, + "tokens_seen": 459997184 + }, + { + "epoch": 0.14, + "learning_rate": 0.0006954302908726179, + "loss": 1.6452, + "theoretical_loss": 3.9788382869160897, + "tokens_seen": 460062720 + }, + { + "epoch": 0.14, + "learning_rate": 0.0006954142427281846, + "loss": 1.6034, + "theoretical_loss": 3.9787768092659963, + "tokens_seen": 460128256 + }, + { + "epoch": 0.14, + "learning_rate": 0.0006953981945837513, + "loss": 1.5914, + "theoretical_loss": 3.978715342822885, + "tokens_seen": 460193792 + }, + { + "epoch": 0.14, + "learning_rate": 0.000695382146439318, + "loss": 1.6569, + "theoretical_loss": 3.9786538875831168, + "tokens_seen": 460259328 + }, + { + "epoch": 0.14, + "learning_rate": 0.0006953660982948847, + "loss": 1.6768, + "theoretical_loss": 3.9785924435430555, + "tokens_seen": 460324864 + }, + { + "epoch": 0.14, + "learning_rate": 0.0006953500501504514, + "loss": 1.546, + "theoretical_loss": 3.978531010699066, + "tokens_seen": 460390400 + }, + { + "epoch": 0.14, + "learning_rate": 0.0006953340020060181, + "loss": 1.5643, + "theoretical_loss": 3.9784695890475144, + "tokens_seen": 460455936 + }, + { + "epoch": 0.14, + "learning_rate": 0.0006953179538615848, + "loss": 1.6097, + "theoretical_loss": 3.9784081785847687, + "tokens_seen": 460521472 + }, + { + "epoch": 0.14, + "learning_rate": 0.0006953019057171515, + "loss": 1.5502, + "theoretical_loss": 3.9783467793071994, + "tokens_seen": 460587008 + }, + { + "epoch": 0.14, + "learning_rate": 0.0006952858575727182, + "loss": 1.5214, + "theoretical_loss": 3.978285391211178, + "tokens_seen": 460652544 + }, + { + "epoch": 0.14, + "learning_rate": 0.0006952698094282849, + "loss": 1.5589, + "theoretical_loss": 3.9782240142930787, + "tokens_seen": 460718080 + }, + { + "epoch": 0.14, + "learning_rate": 0.0006952537612838516, + "loss": 1.5854, + "theoretical_loss": 3.978162648549275, + "tokens_seen": 460783616 + }, + { + "epoch": 0.14, + "learning_rate": 0.0006952377131394183, + "loss": 1.5433, + "theoretical_loss": 3.9781012939761453, + "tokens_seen": 460849152 + }, + { + "epoch": 0.14, + "learning_rate": 0.000695221664994985, + "loss": 1.5474, + "theoretical_loss": 3.978039950570067, + "tokens_seen": 460914688 + }, + { + "epoch": 0.14, + "learning_rate": 0.0006952056168505517, + "loss": 1.6597, + "theoretical_loss": 3.9779786183274197, + "tokens_seen": 460980224 + }, + { + "epoch": 0.14, + "learning_rate": 0.0006951895687061184, + "loss": 1.6007, + "theoretical_loss": 3.977917297244587, + "tokens_seen": 461045760 + }, + { + "epoch": 0.14, + "learning_rate": 0.0006951735205616851, + "loss": 1.6657, + "theoretical_loss": 3.9778559873179504, + "tokens_seen": 461111296 + }, + { + "epoch": 0.14, + "learning_rate": 0.0006951574724172518, + "loss": 1.7203, + "theoretical_loss": 3.977794688543897, + "tokens_seen": 461176832 + }, + { + "epoch": 0.14, + "learning_rate": 0.0006951414242728185, + "loss": 1.5109, + "theoretical_loss": 3.9777334009188117, + "tokens_seen": 461242368 + }, + { + "epoch": 0.14, + "learning_rate": 0.0006951253761283852, + "loss": 1.5696, + "theoretical_loss": 3.9776721244390845, + "tokens_seen": 461307904 + }, + { + "epoch": 0.14, + "learning_rate": 0.0006951093279839519, + "loss": 1.6187, + "theoretical_loss": 3.9776108591011052, + "tokens_seen": 461373440 + }, + { + "epoch": 0.14, + "learning_rate": 0.0006950932798395186, + "loss": 1.5206, + "theoretical_loss": 3.9775496049012657, + "tokens_seen": 461438976 + }, + { + "epoch": 0.14, + "learning_rate": 0.0006950772316950853, + "loss": 1.5196, + "theoretical_loss": 3.9774883618359596, + "tokens_seen": 461504512 + }, + { + "epoch": 0.14, + "learning_rate": 0.000695061183550652, + "loss": 1.5328, + "theoretical_loss": 3.9774271299015824, + "tokens_seen": 461570048 + }, + { + "epoch": 0.14, + "learning_rate": 0.0006950451354062187, + "loss": 1.4379, + "theoretical_loss": 3.977365909094531, + "tokens_seen": 461635584 + }, + { + "epoch": 0.14, + "learning_rate": 0.0006950290872617854, + "loss": 1.7568, + "theoretical_loss": 3.977304699411204, + "tokens_seen": 461701120 + }, + { + "epoch": 0.14, + "learning_rate": 0.0006950130391173521, + "loss": 1.6974, + "theoretical_loss": 3.977243500848001, + "tokens_seen": 461766656 + }, + { + "epoch": 0.14, + "learning_rate": 0.0006949969909729188, + "loss": 1.6132, + "theoretical_loss": 3.977182313401325, + "tokens_seen": 461832192 + }, + { + "epoch": 0.14, + "learning_rate": 0.0006949809428284855, + "loss": 1.6824, + "theoretical_loss": 3.977121137067579, + "tokens_seen": 461897728 + }, + { + "epoch": 0.14, + "learning_rate": 0.0006949648946840522, + "loss": 1.5367, + "theoretical_loss": 3.9770599718431687, + "tokens_seen": 461963264 + }, + { + "epoch": 0.14, + "objective/train/docs_used": 174193, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.7288167476654053, + "objective/train/theoretical_loss": 3.9769988177245015, + "objective/train/tokens_used": 482488800, + "theoretical_loss": 3.9769988177245015, + "tokens_seen": 462028800 + }, + { + "epoch": 0.14, + "learning_rate": 0.0006949488465396189, + "loss": 1.6901, + "theoretical_loss": 3.9769988177245015, + "tokens_seen": 462028800 + }, + { + "epoch": 0.14, + "learning_rate": 0.0006949327983951856, + "loss": 1.6738, + "theoretical_loss": 3.976937674707985, + "tokens_seen": 462094336 + }, + { + "epoch": 0.14, + "learning_rate": 0.0006949167502507523, + "loss": 1.7099, + "theoretical_loss": 3.9768765427900306, + "tokens_seen": 462159872 + }, + { + "epoch": 0.14, + "learning_rate": 0.000694900702106319, + "loss": 1.4486, + "theoretical_loss": 3.97681542196705, + "tokens_seen": 462225408 + }, + { + "epoch": 0.14, + "learning_rate": 0.0006948846539618857, + "loss": 1.5583, + "theoretical_loss": 3.976754312235456, + "tokens_seen": 462290944 + }, + { + "epoch": 0.14, + "learning_rate": 0.0006948686058174524, + "loss": 1.5776, + "theoretical_loss": 3.9766932135916653, + "tokens_seen": 462356480 + }, + { + "epoch": 0.14, + "learning_rate": 0.0006948525576730191, + "loss": 1.6706, + "theoretical_loss": 3.9766321260320945, + "tokens_seen": 462422016 + }, + { + "epoch": 0.14, + "learning_rate": 0.0006948365095285858, + "loss": 1.5813, + "theoretical_loss": 3.976571049553162, + "tokens_seen": 462487552 + }, + { + "epoch": 0.14, + "learning_rate": 0.0006948204613841525, + "loss": 1.5934, + "theoretical_loss": 3.9765099841512885, + "tokens_seen": 462553088 + }, + { + "epoch": 0.14, + "learning_rate": 0.0006948044132397192, + "loss": 1.5769, + "theoretical_loss": 3.976448929822895, + "tokens_seen": 462618624 + }, + { + "epoch": 0.14, + "learning_rate": 0.0006947883650952859, + "loss": 1.474, + "theoretical_loss": 3.9763878865644067, + "tokens_seen": 462684160 + }, + { + "epoch": 0.14, + "learning_rate": 0.0006947723169508526, + "loss": 1.5602, + "theoretical_loss": 3.9763268543722483, + "tokens_seen": 462749696 + }, + { + "epoch": 0.14, + "learning_rate": 0.0006947562688064193, + "loss": 1.5413, + "theoretical_loss": 3.976265833242846, + "tokens_seen": 462815232 + }, + { + "epoch": 0.14, + "learning_rate": 0.000694740220661986, + "loss": 1.6175, + "theoretical_loss": 3.9762048231726297, + "tokens_seen": 462880768 + }, + { + "epoch": 0.14, + "learning_rate": 0.0006947241725175527, + "loss": 1.7297, + "theoretical_loss": 3.9761438241580294, + "tokens_seen": 462946304 + }, + { + "epoch": 0.14, + "learning_rate": 0.0006947081243731194, + "loss": 1.5979, + "theoretical_loss": 3.976082836195477, + "tokens_seen": 463011840 + }, + { + "epoch": 0.14, + "learning_rate": 0.0006946920762286861, + "loss": 1.5814, + "theoretical_loss": 3.976021859281406, + "tokens_seen": 463077376 + }, + { + "epoch": 0.14, + "learning_rate": 0.0006946760280842528, + "loss": 1.5808, + "theoretical_loss": 3.975960893412251, + "tokens_seen": 463142912 + }, + { + "epoch": 0.14, + "learning_rate": 0.0006946599799398195, + "loss": 1.503, + "theoretical_loss": 3.9758999385844502, + "tokens_seen": 463208448 + }, + { + "epoch": 0.14, + "learning_rate": 0.0006946439317953862, + "loss": 1.4918, + "theoretical_loss": 3.975838994794441, + "tokens_seen": 463273984 + }, + { + "epoch": 0.14, + "learning_rate": 0.0006946278836509529, + "loss": 1.5479, + "theoretical_loss": 3.975778062038665, + "tokens_seen": 463339520 + }, + { + "epoch": 0.14, + "learning_rate": 0.0006946118355065196, + "loss": 1.6867, + "theoretical_loss": 3.975717140313564, + "tokens_seen": 463405056 + }, + { + "epoch": 0.14, + "learning_rate": 0.0006945957873620863, + "loss": 1.668, + "theoretical_loss": 3.9756562296155797, + "tokens_seen": 463470592 + }, + { + "epoch": 0.14, + "learning_rate": 0.000694579739217653, + "loss": 1.4495, + "theoretical_loss": 3.975595329941159, + "tokens_seen": 463536128 + }, + { + "epoch": 0.14, + "learning_rate": 0.0006945636910732197, + "loss": 1.5612, + "theoretical_loss": 3.975534441286748, + "tokens_seen": 463601664 + }, + { + "epoch": 0.14, + "learning_rate": 0.0006945476429287864, + "loss": 1.5318, + "theoretical_loss": 3.975473563648796, + "tokens_seen": 463667200 + }, + { + "epoch": 0.14, + "learning_rate": 0.0006945315947843531, + "loss": 1.7074, + "theoretical_loss": 3.975412697023753, + "tokens_seen": 463732736 + }, + { + "epoch": 0.14, + "learning_rate": 0.0006945155466399198, + "loss": 1.5673, + "theoretical_loss": 3.9753518414080693, + "tokens_seen": 463798272 + }, + { + "epoch": 0.14, + "learning_rate": 0.0006944994984954865, + "loss": 1.5443, + "theoretical_loss": 3.9752909967982, + "tokens_seen": 463863808 + }, + { + "epoch": 0.14, + "learning_rate": 0.0006944834503510532, + "loss": 1.6736, + "theoretical_loss": 3.9752301631906, + "tokens_seen": 463929344 + }, + { + "epoch": 0.14, + "learning_rate": 0.0006944674022066199, + "loss": 1.5723, + "theoretical_loss": 3.9751693405817248, + "tokens_seen": 463994880 + }, + { + "epoch": 0.14, + "learning_rate": 0.0006944513540621866, + "loss": 1.7255, + "theoretical_loss": 3.9751085289680343, + "tokens_seen": 464060416 + }, + { + "epoch": 0.14, + "learning_rate": 0.0006944353059177533, + "loss": 1.7928, + "theoretical_loss": 3.975047728345987, + "tokens_seen": 464125952 + }, + { + "epoch": 0.14, + "learning_rate": 0.00069441925777332, + "loss": 1.6863, + "theoretical_loss": 3.9749869387120462, + "tokens_seen": 464191488 + }, + { + "epoch": 0.14, + "learning_rate": 0.0006944032096288867, + "loss": 1.4933, + "theoretical_loss": 3.974926160062674, + "tokens_seen": 464257024 + }, + { + "epoch": 0.14, + "learning_rate": 0.0006943871614844534, + "loss": 1.5732, + "theoretical_loss": 3.9748653923943356, + "tokens_seen": 464322560 + }, + { + "epoch": 0.14, + "learning_rate": 0.0006943711133400201, + "loss": 1.7171, + "theoretical_loss": 3.9748046357034976, + "tokens_seen": 464388096 + }, + { + "epoch": 0.14, + "learning_rate": 0.0006943550651955868, + "loss": 1.5961, + "theoretical_loss": 3.9747438899866276, + "tokens_seen": 464453632 + }, + { + "epoch": 0.14, + "learning_rate": 0.0006943390170511535, + "loss": 1.5412, + "theoretical_loss": 3.9746831552401964, + "tokens_seen": 464519168 + }, + { + "epoch": 0.14, + "learning_rate": 0.0006943229689067202, + "loss": 1.5499, + "theoretical_loss": 3.9746224314606753, + "tokens_seen": 464584704 + }, + { + "epoch": 0.14, + "learning_rate": 0.000694306920762287, + "loss": 1.588, + "theoretical_loss": 3.9745617186445372, + "tokens_seen": 464650240 + }, + { + "epoch": 0.14, + "learning_rate": 0.0006942908726178536, + "loss": 1.7044, + "theoretical_loss": 3.974501016788256, + "tokens_seen": 464715776 + }, + { + "epoch": 0.14, + "learning_rate": 0.0006942748244734202, + "loss": 1.7005, + "theoretical_loss": 3.9744403258883096, + "tokens_seen": 464781312 + }, + { + "epoch": 0.14, + "learning_rate": 0.000694258776328987, + "loss": 1.5065, + "theoretical_loss": 3.974379645941174, + "tokens_seen": 464846848 + }, + { + "epoch": 0.14, + "learning_rate": 0.0006942427281845536, + "loss": 1.5611, + "theoretical_loss": 3.974318976943331, + "tokens_seen": 464912384 + }, + { + "epoch": 0.14, + "learning_rate": 0.0006942266800401204, + "loss": 1.6373, + "theoretical_loss": 3.974258318891261, + "tokens_seen": 464977920 + }, + { + "epoch": 0.14, + "learning_rate": 0.000694210631895687, + "loss": 1.5072, + "theoretical_loss": 3.974197671781446, + "tokens_seen": 465043456 + }, + { + "epoch": 0.14, + "learning_rate": 0.0006941945837512538, + "loss": 1.5089, + "theoretical_loss": 3.9741370356103713, + "tokens_seen": 465108992 + }, + { + "epoch": 0.14, + "learning_rate": 0.0006941785356068204, + "loss": 1.5762, + "theoretical_loss": 3.9740764103745225, + "tokens_seen": 465174528 + }, + { + "epoch": 0.14, + "learning_rate": 0.0006941624874623872, + "loss": 1.6056, + "theoretical_loss": 3.974015796070388, + "tokens_seen": 465240064 + }, + { + "epoch": 0.14, + "objective/train/docs_used": 175471, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.3526719808578491, + "objective/train/theoretical_loss": 3.973955192694457, + "objective/train/tokens_used": 485765600, + "theoretical_loss": 3.973955192694457, + "tokens_seen": 465305600 + }, + { + "epoch": 0.14, + "learning_rate": 0.0006941464393179538, + "loss": 1.4671, + "theoretical_loss": 3.973955192694457, + "tokens_seen": 465305600 + }, + { + "epoch": 0.14, + "learning_rate": 0.0006941303911735206, + "loss": 1.5204, + "theoretical_loss": 3.97389460024322, + "tokens_seen": 465371136 + }, + { + "epoch": 0.14, + "learning_rate": 0.0006941143430290872, + "loss": 1.5704, + "theoretical_loss": 3.9738340187131698, + "tokens_seen": 465436672 + }, + { + "epoch": 0.14, + "learning_rate": 0.000694098294884654, + "loss": 1.619, + "theoretical_loss": 3.973773448100801, + "tokens_seen": 465502208 + }, + { + "epoch": 0.14, + "learning_rate": 0.0006940822467402206, + "loss": 1.5427, + "theoretical_loss": 3.973712888402609, + "tokens_seen": 465567744 + }, + { + "epoch": 0.14, + "learning_rate": 0.0006940661985957874, + "loss": 1.6067, + "theoretical_loss": 3.9736523396150916, + "tokens_seen": 465633280 + }, + { + "epoch": 0.14, + "learning_rate": 0.000694050150451354, + "loss": 1.4913, + "theoretical_loss": 3.973591801734748, + "tokens_seen": 465698816 + }, + { + "epoch": 0.14, + "learning_rate": 0.0006940341023069208, + "loss": 1.617, + "theoretical_loss": 3.9735312747580775, + "tokens_seen": 465764352 + }, + { + "epoch": 0.14, + "learning_rate": 0.0006940180541624875, + "loss": 1.6616, + "theoretical_loss": 3.973470758681584, + "tokens_seen": 465829888 + }, + { + "epoch": 0.14, + "learning_rate": 0.0006940020060180542, + "loss": 1.5875, + "theoretical_loss": 3.9734102535017715, + "tokens_seen": 465895424 + }, + { + "epoch": 0.14, + "learning_rate": 0.0006939859578736209, + "loss": 1.5816, + "theoretical_loss": 3.9733497592151443, + "tokens_seen": 465960960 + }, + { + "epoch": 0.14, + "learning_rate": 0.0006939699097291876, + "loss": 1.465, + "theoretical_loss": 3.9732892758182103, + "tokens_seen": 466026496 + }, + { + "epoch": 0.14, + "learning_rate": 0.0006939538615847543, + "loss": 1.6002, + "theoretical_loss": 3.973228803307478, + "tokens_seen": 466092032 + }, + { + "epoch": 0.14, + "learning_rate": 0.000693937813440321, + "loss": 1.4251, + "theoretical_loss": 3.9731683416794583, + "tokens_seen": 466157568 + }, + { + "epoch": 0.14, + "learning_rate": 0.0006939217652958877, + "loss": 1.5708, + "theoretical_loss": 3.9731078909306623, + "tokens_seen": 466223104 + }, + { + "epoch": 0.14, + "learning_rate": 0.0006939057171514544, + "loss": 1.7334, + "theoretical_loss": 3.9730474510576044, + "tokens_seen": 466288640 + }, + { + "epoch": 0.14, + "learning_rate": 0.0006938896690070211, + "loss": 1.6246, + "theoretical_loss": 3.9729870220567993, + "tokens_seen": 466354176 + }, + { + "epoch": 0.14, + "learning_rate": 0.0006938736208625878, + "loss": 1.651, + "theoretical_loss": 3.9729266039247637, + "tokens_seen": 466419712 + }, + { + "epoch": 0.14, + "learning_rate": 0.0006938575727181545, + "loss": 1.568, + "theoretical_loss": 3.9728661966580163, + "tokens_seen": 466485248 + }, + { + "epoch": 0.14, + "learning_rate": 0.0006938415245737212, + "loss": 1.5295, + "theoretical_loss": 3.9728058002530773, + "tokens_seen": 466550784 + }, + { + "epoch": 0.14, + "learning_rate": 0.0006938254764292879, + "loss": 1.6202, + "theoretical_loss": 3.972745414706468, + "tokens_seen": 466616320 + }, + { + "epoch": 0.14, + "learning_rate": 0.0006938094282848546, + "loss": 1.4942, + "theoretical_loss": 3.972685040014711, + "tokens_seen": 466681856 + }, + { + "epoch": 0.14, + "learning_rate": 0.0006937933801404213, + "loss": 1.4559, + "theoretical_loss": 3.9726246761743322, + "tokens_seen": 466747392 + }, + { + "epoch": 0.14, + "learning_rate": 0.000693777331995988, + "loss": 1.541, + "theoretical_loss": 3.9725643231818575, + "tokens_seen": 466812928 + }, + { + "epoch": 0.14, + "learning_rate": 0.0006937612838515547, + "loss": 1.6105, + "theoretical_loss": 3.9725039810338147, + "tokens_seen": 466878464 + }, + { + "epoch": 0.14, + "learning_rate": 0.0006937452357071214, + "loss": 1.6229, + "theoretical_loss": 3.972443649726734, + "tokens_seen": 466944000 + }, + { + "epoch": 0.14, + "learning_rate": 0.0006937291875626881, + "loss": 1.5586, + "theoretical_loss": 3.9723833292571458, + "tokens_seen": 467009536 + }, + { + "epoch": 0.14, + "learning_rate": 0.0006937131394182548, + "loss": 1.4906, + "theoretical_loss": 3.9723230196215837, + "tokens_seen": 467075072 + }, + { + "epoch": 0.14, + "learning_rate": 0.0006936970912738215, + "loss": 1.6766, + "theoretical_loss": 3.9722627208165817, + "tokens_seen": 467140608 + }, + { + "epoch": 0.14, + "learning_rate": 0.0006936810431293882, + "loss": 1.5681, + "theoretical_loss": 3.9722024328386754, + "tokens_seen": 467206144 + }, + { + "epoch": 0.14, + "learning_rate": 0.0006936649949849549, + "loss": 1.5303, + "theoretical_loss": 3.9721421556844034, + "tokens_seen": 467271680 + }, + { + "epoch": 0.14, + "learning_rate": 0.0006936489468405216, + "loss": 1.5957, + "theoretical_loss": 3.972081889350304, + "tokens_seen": 467337216 + }, + { + "epoch": 0.14, + "learning_rate": 0.0006936328986960883, + "loss": 1.5913, + "theoretical_loss": 3.9720216338329184, + "tokens_seen": 467402752 + }, + { + "epoch": 0.14, + "learning_rate": 0.000693616850551655, + "loss": 1.6153, + "theoretical_loss": 3.971961389128789, + "tokens_seen": 467468288 + }, + { + "epoch": 0.14, + "learning_rate": 0.0006936008024072217, + "loss": 1.5699, + "theoretical_loss": 3.97190115523446, + "tokens_seen": 467533824 + }, + { + "epoch": 0.14, + "learning_rate": 0.0006935847542627884, + "loss": 1.5196, + "theoretical_loss": 3.9718409321464754, + "tokens_seen": 467599360 + }, + { + "epoch": 0.14, + "learning_rate": 0.0006935687061183551, + "loss": 1.646, + "theoretical_loss": 3.9717807198613837, + "tokens_seen": 467664896 + }, + { + "epoch": 0.14, + "learning_rate": 0.0006935526579739218, + "loss": 1.6778, + "theoretical_loss": 3.971720518375734, + "tokens_seen": 467730432 + }, + { + "epoch": 0.14, + "learning_rate": 0.0006935366098294885, + "loss": 1.6097, + "theoretical_loss": 3.971660327686075, + "tokens_seen": 467795968 + }, + { + "epoch": 0.14, + "learning_rate": 0.0006935205616850552, + "loss": 1.5006, + "theoretical_loss": 3.97160014778896, + "tokens_seen": 467861504 + }, + { + "epoch": 0.14, + "learning_rate": 0.0006935045135406219, + "loss": 1.6625, + "theoretical_loss": 3.9715399786809424, + "tokens_seen": 467927040 + }, + { + "epoch": 0.14, + "learning_rate": 0.0006934884653961886, + "loss": 1.5361, + "theoretical_loss": 3.971479820358576, + "tokens_seen": 467992576 + }, + { + "epoch": 0.14, + "learning_rate": 0.0006934724172517553, + "loss": 1.4911, + "theoretical_loss": 3.9714196728184197, + "tokens_seen": 468058112 + }, + { + "epoch": 0.14, + "learning_rate": 0.000693456369107322, + "loss": 1.5114, + "theoretical_loss": 3.971359536057029, + "tokens_seen": 468123648 + }, + { + "epoch": 0.14, + "learning_rate": 0.0006934403209628887, + "loss": 1.5618, + "theoretical_loss": 3.9712994100709653, + "tokens_seen": 468189184 + }, + { + "epoch": 0.14, + "learning_rate": 0.0006934242728184554, + "loss": 1.5737, + "theoretical_loss": 3.97123929485679, + "tokens_seen": 468254720 + }, + { + "epoch": 0.14, + "learning_rate": 0.0006934082246740221, + "loss": 1.5662, + "theoretical_loss": 3.971179190411066, + "tokens_seen": 468320256 + }, + { + "epoch": 0.14, + "learning_rate": 0.0006933921765295888, + "loss": 1.5565, + "theoretical_loss": 3.9711190967303565, + "tokens_seen": 468385792 + }, + { + "epoch": 0.14, + "learning_rate": 0.0006933761283851555, + "loss": 1.5746, + "theoretical_loss": 3.971059013811229, + "tokens_seen": 468451328 + }, + { + "epoch": 0.14, + "learning_rate": 0.0006933600802407222, + "loss": 1.492, + "theoretical_loss": 3.9709989416502514, + "tokens_seen": 468516864 + }, + { + "epoch": 0.14, + "objective/train/docs_used": 176758, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.70401930809021, + "objective/train/theoretical_loss": 3.9709388802439918, + "objective/train/tokens_used": 489042400, + "theoretical_loss": 3.9709388802439918, + "tokens_seen": 468582400 + }, + { + "epoch": 0.14, + "learning_rate": 0.0006933440320962889, + "loss": 1.6639, + "theoretical_loss": 3.9709388802439918, + "tokens_seen": 468582400 + }, + { + "epoch": 0.14, + "learning_rate": 0.0006933279839518556, + "loss": 1.5922, + "theoretical_loss": 3.9708788295890223, + "tokens_seen": 468647936 + }, + { + "epoch": 0.14, + "learning_rate": 0.0006933119358074223, + "loss": 1.6494, + "theoretical_loss": 3.9708187896819136, + "tokens_seen": 468713472 + }, + { + "epoch": 0.14, + "learning_rate": 0.000693295887662989, + "loss": 1.6312, + "theoretical_loss": 3.9707587605192414, + "tokens_seen": 468779008 + }, + { + "epoch": 0.14, + "learning_rate": 0.0006932798395185557, + "loss": 1.5568, + "theoretical_loss": 3.9706987420975803, + "tokens_seen": 468844544 + }, + { + "epoch": 0.14, + "learning_rate": 0.0006932637913741224, + "loss": 1.5788, + "theoretical_loss": 3.970638734413508, + "tokens_seen": 468910080 + }, + { + "epoch": 0.14, + "learning_rate": 0.0006932477432296891, + "loss": 1.5938, + "theoretical_loss": 3.970578737463603, + "tokens_seen": 468975616 + }, + { + "epoch": 0.14, + "learning_rate": 0.0006932316950852558, + "loss": 1.633, + "theoretical_loss": 3.9705187512444446, + "tokens_seen": 469041152 + }, + { + "epoch": 0.14, + "learning_rate": 0.0006932156469408225, + "loss": 1.613, + "theoretical_loss": 3.9704587757526157, + "tokens_seen": 469106688 + }, + { + "epoch": 0.14, + "learning_rate": 0.0006931995987963892, + "loss": 1.6668, + "theoretical_loss": 3.9703988109846997, + "tokens_seen": 469172224 + }, + { + "epoch": 0.14, + "learning_rate": 0.0006931835506519559, + "loss": 1.5262, + "theoretical_loss": 3.970338856937281, + "tokens_seen": 469237760 + }, + { + "epoch": 0.14, + "learning_rate": 0.0006931675025075226, + "loss": 1.5919, + "theoretical_loss": 3.970278913606947, + "tokens_seen": 469303296 + }, + { + "epoch": 0.14, + "learning_rate": 0.0006931514543630893, + "loss": 1.6103, + "theoretical_loss": 3.9702189809902846, + "tokens_seen": 469368832 + }, + { + "epoch": 0.14, + "learning_rate": 0.000693135406218656, + "loss": 1.6034, + "theoretical_loss": 3.9701590590838842, + "tokens_seen": 469434368 + }, + { + "epoch": 0.14, + "learning_rate": 0.0006931193580742227, + "loss": 1.5164, + "theoretical_loss": 3.970099147884336, + "tokens_seen": 469499904 + }, + { + "epoch": 0.14, + "learning_rate": 0.0006931033099297894, + "loss": 1.5574, + "theoretical_loss": 3.9700392473882347, + "tokens_seen": 469565440 + }, + { + "epoch": 0.14, + "learning_rate": 0.0006930872617853562, + "loss": 1.632, + "theoretical_loss": 3.969979357592173, + "tokens_seen": 469630976 + }, + { + "epoch": 0.14, + "learning_rate": 0.0006930712136409228, + "loss": 1.6171, + "theoretical_loss": 3.969919478492747, + "tokens_seen": 469696512 + }, + { + "epoch": 0.14, + "learning_rate": 0.0006930551654964896, + "loss": 1.5846, + "theoretical_loss": 3.9698596100865546, + "tokens_seen": 469762048 + }, + { + "epoch": 0.14, + "learning_rate": 0.0006930391173520562, + "loss": 1.635, + "theoretical_loss": 3.9697997523701947, + "tokens_seen": 469827584 + }, + { + "epoch": 0.14, + "learning_rate": 0.0006930230692076229, + "loss": 1.5778, + "theoretical_loss": 3.969739905340268, + "tokens_seen": 469893120 + }, + { + "epoch": 0.14, + "learning_rate": 0.0006930070210631896, + "loss": 1.6686, + "theoretical_loss": 3.9696800689933758, + "tokens_seen": 469958656 + }, + { + "epoch": 0.14, + "learning_rate": 0.0006929909729187563, + "loss": 1.6406, + "theoretical_loss": 3.9696202433261227, + "tokens_seen": 470024192 + }, + { + "epoch": 0.14, + "learning_rate": 0.000692974924774323, + "loss": 1.5986, + "theoretical_loss": 3.9695604283351136, + "tokens_seen": 470089728 + }, + { + "epoch": 0.14, + "learning_rate": 0.0006929588766298897, + "loss": 1.5189, + "theoretical_loss": 3.9695006240169546, + "tokens_seen": 470155264 + }, + { + "epoch": 0.14, + "learning_rate": 0.0006929428284854564, + "loss": 1.569, + "theoretical_loss": 3.9694408303682556, + "tokens_seen": 470220800 + }, + { + "epoch": 0.14, + "learning_rate": 0.0006929267803410231, + "loss": 1.5353, + "theoretical_loss": 3.9693810473856246, + "tokens_seen": 470286336 + }, + { + "epoch": 0.14, + "learning_rate": 0.0006929107321965898, + "loss": 1.5386, + "theoretical_loss": 3.969321275065674, + "tokens_seen": 470351872 + }, + { + "epoch": 0.14, + "learning_rate": 0.0006928946840521565, + "loss": 1.6356, + "theoretical_loss": 3.969261513405017, + "tokens_seen": 470417408 + }, + { + "epoch": 0.14, + "learning_rate": 0.0006928786359077231, + "loss": 1.6715, + "theoretical_loss": 3.9692017624002682, + "tokens_seen": 470482944 + }, + { + "epoch": 0.14, + "learning_rate": 0.0006928625877632899, + "loss": 1.5722, + "theoretical_loss": 3.9691420220480422, + "tokens_seen": 470548480 + }, + { + "epoch": 0.14, + "learning_rate": 0.0006928465396188565, + "loss": 1.4812, + "theoretical_loss": 3.9690822923449582, + "tokens_seen": 470614016 + }, + { + "epoch": 0.14, + "learning_rate": 0.0006928304914744233, + "loss": 1.6751, + "theoretical_loss": 3.969022573287635, + "tokens_seen": 470679552 + }, + { + "epoch": 0.14, + "learning_rate": 0.0006928144433299899, + "loss": 1.5276, + "theoretical_loss": 3.968962864872693, + "tokens_seen": 470745088 + }, + { + "epoch": 0.14, + "learning_rate": 0.0006927983951855567, + "loss": 1.6029, + "theoretical_loss": 3.9689031670967543, + "tokens_seen": 470810624 + }, + { + "epoch": 0.14, + "learning_rate": 0.0006927823470411233, + "loss": 1.5996, + "theoretical_loss": 3.9688434799564427, + "tokens_seen": 470876160 + }, + { + "epoch": 0.14, + "learning_rate": 0.0006927662988966901, + "loss": 1.569, + "theoretical_loss": 3.9687838034483844, + "tokens_seen": 470941696 + }, + { + "epoch": 0.14, + "learning_rate": 0.0006927502507522567, + "loss": 1.5291, + "theoretical_loss": 3.9687241375692053, + "tokens_seen": 471007232 + }, + { + "epoch": 0.14, + "learning_rate": 0.0006927342026078235, + "loss": 1.5946, + "theoretical_loss": 3.9686644823155337, + "tokens_seen": 471072768 + }, + { + "epoch": 0.14, + "learning_rate": 0.0006927181544633901, + "loss": 1.6797, + "theoretical_loss": 3.9686048376840004, + "tokens_seen": 471138304 + }, + { + "epoch": 0.14, + "learning_rate": 0.0006927021063189569, + "loss": 1.6743, + "theoretical_loss": 3.968545203671236, + "tokens_seen": 471203840 + }, + { + "epoch": 0.14, + "learning_rate": 0.0006926860581745235, + "loss": 1.5878, + "theoretical_loss": 3.9684855802738745, + "tokens_seen": 471269376 + }, + { + "epoch": 0.14, + "learning_rate": 0.0006926700100300903, + "loss": 1.6162, + "theoretical_loss": 3.9684259674885496, + "tokens_seen": 471334912 + }, + { + "epoch": 0.14, + "learning_rate": 0.000692653961885657, + "loss": 1.4805, + "theoretical_loss": 3.9683663653118977, + "tokens_seen": 471400448 + }, + { + "epoch": 0.14, + "learning_rate": 0.0006926379137412237, + "loss": 1.5581, + "theoretical_loss": 3.968306773740556, + "tokens_seen": 471465984 + }, + { + "epoch": 0.14, + "learning_rate": 0.0006926218655967904, + "loss": 1.7074, + "theoretical_loss": 3.9682471927711642, + "tokens_seen": 471531520 + }, + { + "epoch": 0.14, + "learning_rate": 0.0006926058174523571, + "loss": 1.5371, + "theoretical_loss": 3.9681876224003627, + "tokens_seen": 471597056 + }, + { + "epoch": 0.14, + "learning_rate": 0.0006925897693079238, + "loss": 1.6832, + "theoretical_loss": 3.9681280626247943, + "tokens_seen": 471662592 + }, + { + "epoch": 0.14, + "learning_rate": 0.0006925737211634905, + "loss": 1.6076, + "theoretical_loss": 3.9680685134411022, + "tokens_seen": 471728128 + }, + { + "epoch": 0.14, + "learning_rate": 0.0006925576730190572, + "loss": 1.5681, + "theoretical_loss": 3.9680089748459313, + "tokens_seen": 471793664 + }, + { + "epoch": 0.14, + "objective/train/docs_used": 177948, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.6531736850738525, + "objective/train/theoretical_loss": 3.967949446835929, + "objective/train/tokens_used": 492319200, + "theoretical_loss": 3.967949446835929, + "tokens_seen": 471859200 + }, + { + "epoch": 0.14, + "learning_rate": 0.0006925416248746239, + "loss": 1.611, + "theoretical_loss": 3.967949446835929, + "tokens_seen": 471859200 + }, + { + "epoch": 0.14, + "learning_rate": 0.0006925255767301906, + "loss": 1.4401, + "theoretical_loss": 3.967889929407743, + "tokens_seen": 471924736 + }, + { + "epoch": 0.14, + "learning_rate": 0.0006925095285857573, + "loss": 1.5708, + "theoretical_loss": 3.9678304225580248, + "tokens_seen": 471990272 + }, + { + "epoch": 0.14, + "learning_rate": 0.000692493480441324, + "loss": 1.4554, + "theoretical_loss": 3.967770926283424, + "tokens_seen": 472055808 + }, + { + "epoch": 0.14, + "learning_rate": 0.0006924774322968907, + "loss": 1.5342, + "theoretical_loss": 3.9677114405805938, + "tokens_seen": 472121344 + }, + { + "epoch": 0.14, + "learning_rate": 0.0006924613841524574, + "loss": 1.5058, + "theoretical_loss": 3.967651965446189, + "tokens_seen": 472186880 + }, + { + "epoch": 0.14, + "learning_rate": 0.0006924453360080241, + "loss": 1.6298, + "theoretical_loss": 3.9675925008768655, + "tokens_seen": 472252416 + }, + { + "epoch": 0.14, + "learning_rate": 0.0006924292878635908, + "loss": 1.5927, + "theoretical_loss": 3.9675330468692804, + "tokens_seen": 472317952 + }, + { + "epoch": 0.14, + "learning_rate": 0.0006924132397191575, + "loss": 1.5235, + "theoretical_loss": 3.9674736034200935, + "tokens_seen": 472383488 + }, + { + "epoch": 0.14, + "learning_rate": 0.0006923971915747242, + "loss": 1.646, + "theoretical_loss": 3.967414170525965, + "tokens_seen": 472449024 + }, + { + "epoch": 0.14, + "learning_rate": 0.0006923811434302909, + "loss": 1.6031, + "theoretical_loss": 3.9673547481835567, + "tokens_seen": 472514560 + }, + { + "epoch": 0.14, + "learning_rate": 0.0006923650952858576, + "loss": 1.6036, + "theoretical_loss": 3.9672953363895322, + "tokens_seen": 472580096 + }, + { + "epoch": 0.14, + "learning_rate": 0.0006923490471414243, + "loss": 1.568, + "theoretical_loss": 3.9672359351405566, + "tokens_seen": 472645632 + }, + { + "epoch": 0.14, + "learning_rate": 0.000692332998996991, + "loss": 1.661, + "theoretical_loss": 3.967176544433296, + "tokens_seen": 472711168 + }, + { + "epoch": 0.14, + "learning_rate": 0.0006923169508525577, + "loss": 1.6057, + "theoretical_loss": 3.96711716426442, + "tokens_seen": 472776704 + }, + { + "epoch": 0.14, + "learning_rate": 0.0006923009027081244, + "loss": 1.4533, + "theoretical_loss": 3.9670577946305965, + "tokens_seen": 472842240 + }, + { + "epoch": 0.14, + "learning_rate": 0.0006922848545636911, + "loss": 1.723, + "theoretical_loss": 3.9669984355284975, + "tokens_seen": 472907776 + }, + { + "epoch": 0.14, + "learning_rate": 0.0006922688064192578, + "loss": 1.5788, + "theoretical_loss": 3.9669390869547954, + "tokens_seen": 472973312 + }, + { + "epoch": 0.14, + "learning_rate": 0.0006922527582748245, + "loss": 1.5627, + "theoretical_loss": 3.966879748906165, + "tokens_seen": 473038848 + }, + { + "epoch": 0.14, + "learning_rate": 0.0006922367101303912, + "loss": 1.4778, + "theoretical_loss": 3.966820421379281, + "tokens_seen": 473104384 + }, + { + "epoch": 0.14, + "learning_rate": 0.0006922206619859579, + "loss": 1.6626, + "theoretical_loss": 3.966761104370821, + "tokens_seen": 473169920 + }, + { + "epoch": 0.14, + "learning_rate": 0.0006922046138415246, + "loss": 1.5199, + "theoretical_loss": 3.9667017978774632, + "tokens_seen": 473235456 + }, + { + "epoch": 0.14, + "learning_rate": 0.0006921885656970913, + "loss": 1.6412, + "theoretical_loss": 3.966642501895889, + "tokens_seen": 473300992 + }, + { + "epoch": 0.14, + "learning_rate": 0.000692172517552658, + "loss": 1.7117, + "theoretical_loss": 3.9665832164227792, + "tokens_seen": 473366528 + }, + { + "epoch": 0.14, + "learning_rate": 0.0006921564694082247, + "loss": 1.5621, + "theoretical_loss": 3.966523941454817, + "tokens_seen": 473432064 + }, + { + "epoch": 0.14, + "learning_rate": 0.0006921404212637914, + "loss": 1.6063, + "theoretical_loss": 3.9664646769886875, + "tokens_seen": 473497600 + }, + { + "epoch": 0.14, + "learning_rate": 0.0006921243731193581, + "loss": 1.6206, + "theoretical_loss": 3.966405423021077, + "tokens_seen": 473563136 + }, + { + "epoch": 0.14, + "learning_rate": 0.0006921083249749248, + "loss": 1.4934, + "theoretical_loss": 3.9663461795486725, + "tokens_seen": 473628672 + }, + { + "epoch": 0.14, + "learning_rate": 0.0006920922768304915, + "loss": 1.5713, + "theoretical_loss": 3.9662869465681636, + "tokens_seen": 473694208 + }, + { + "epoch": 0.14, + "learning_rate": 0.0006920762286860582, + "loss": 1.5897, + "theoretical_loss": 3.9662277240762416, + "tokens_seen": 473759744 + }, + { + "epoch": 0.14, + "learning_rate": 0.0006920601805416249, + "loss": 1.641, + "theoretical_loss": 3.9661685120695975, + "tokens_seen": 473825280 + }, + { + "epoch": 0.14, + "learning_rate": 0.0006920441323971916, + "loss": 1.524, + "theoretical_loss": 3.9661093105449265, + "tokens_seen": 473890816 + }, + { + "epoch": 0.14, + "learning_rate": 0.0006920280842527583, + "loss": 1.603, + "theoretical_loss": 3.966050119498923, + "tokens_seen": 473956352 + }, + { + "epoch": 0.14, + "learning_rate": 0.000692012036108325, + "loss": 1.4977, + "theoretical_loss": 3.9659909389282832, + "tokens_seen": 474021888 + }, + { + "epoch": 0.14, + "learning_rate": 0.0006919959879638917, + "loss": 1.6196, + "theoretical_loss": 3.9659317688297064, + "tokens_seen": 474087424 + }, + { + "epoch": 0.14, + "learning_rate": 0.0006919799398194584, + "loss": 1.5742, + "theoretical_loss": 3.965872609199892, + "tokens_seen": 474152960 + }, + { + "epoch": 0.14, + "learning_rate": 0.0006919638916750251, + "loss": 1.4963, + "theoretical_loss": 3.965813460035541, + "tokens_seen": 474218496 + }, + { + "epoch": 0.14, + "learning_rate": 0.0006919478435305918, + "loss": 1.6334, + "theoretical_loss": 3.965754321333356, + "tokens_seen": 474284032 + }, + { + "epoch": 0.14, + "learning_rate": 0.0006919317953861585, + "loss": 1.6564, + "theoretical_loss": 3.965695193090042, + "tokens_seen": 474349568 + }, + { + "epoch": 0.14, + "learning_rate": 0.0006919157472417252, + "loss": 1.6023, + "theoretical_loss": 3.9656360753023034, + "tokens_seen": 474415104 + }, + { + "epoch": 0.14, + "learning_rate": 0.0006918996990972919, + "loss": 1.6245, + "theoretical_loss": 3.9655769679668484, + "tokens_seen": 474480640 + }, + { + "epoch": 0.14, + "learning_rate": 0.0006918836509528587, + "loss": 1.6868, + "theoretical_loss": 3.965517871080386, + "tokens_seen": 474546176 + }, + { + "epoch": 0.14, + "learning_rate": 0.0006918676028084253, + "loss": 1.5654, + "theoretical_loss": 3.965458784639625, + "tokens_seen": 474611712 + }, + { + "epoch": 0.14, + "learning_rate": 0.0006918515546639921, + "loss": 1.7474, + "theoretical_loss": 3.965399708641278, + "tokens_seen": 474677248 + }, + { + "epoch": 0.14, + "learning_rate": 0.0006918355065195587, + "loss": 1.6292, + "theoretical_loss": 3.9653406430820586, + "tokens_seen": 474742784 + }, + { + "epoch": 0.14, + "learning_rate": 0.0006918194583751255, + "loss": 1.5518, + "theoretical_loss": 3.9652815879586805, + "tokens_seen": 474808320 + }, + { + "epoch": 0.14, + "learning_rate": 0.0006918034102306921, + "loss": 1.4822, + "theoretical_loss": 3.96522254326786, + "tokens_seen": 474873856 + }, + { + "epoch": 0.14, + "learning_rate": 0.0006917873620862589, + "loss": 1.6641, + "theoretical_loss": 3.9651635090063153, + "tokens_seen": 474939392 + }, + { + "epoch": 0.14, + "learning_rate": 0.0006917713139418255, + "loss": 1.5662, + "theoretical_loss": 3.965104485170765, + "tokens_seen": 475004928 + }, + { + "epoch": 0.14, + "learning_rate": 0.0006917552657973923, + "loss": 1.6297, + "theoretical_loss": 3.9650454717579295, + "tokens_seen": 475070464 + }, + { + "epoch": 0.14, + "objective/train/docs_used": 179087, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.4692811965942383, + "objective/train/theoretical_loss": 3.9649864687645318, + "objective/train/tokens_used": 495596000, + "theoretical_loss": 3.9649864687645318, + "tokens_seen": 475136000 + }, + { + "epoch": 0.14, + "learning_rate": 0.0006917392176529589, + "loss": 1.4789, + "theoretical_loss": 3.9649864687645318, + "tokens_seen": 475136000 + }, + { + "epoch": 0.14, + "learning_rate": 0.0006917231695085257, + "loss": 1.5485, + "theoretical_loss": 3.9649274761872944, + "tokens_seen": 475201536 + }, + { + "epoch": 0.14, + "learning_rate": 0.0006917071213640923, + "loss": 1.5639, + "theoretical_loss": 3.9648684940229426, + "tokens_seen": 475267072 + }, + { + "epoch": 0.14, + "learning_rate": 0.0006916910732196591, + "loss": 1.5657, + "theoretical_loss": 3.9648095222682036, + "tokens_seen": 475332608 + }, + { + "epoch": 0.14, + "learning_rate": 0.0006916750250752257, + "loss": 1.5876, + "theoretical_loss": 3.9647505609198044, + "tokens_seen": 475398144 + }, + { + "epoch": 0.14, + "learning_rate": 0.0006916589769307925, + "loss": 1.7387, + "theoretical_loss": 3.9646916099744756, + "tokens_seen": 475463680 + }, + { + "epoch": 0.14, + "learning_rate": 0.0006916429287863591, + "loss": 1.6174, + "theoretical_loss": 3.9646326694289473, + "tokens_seen": 475529216 + }, + { + "epoch": 0.14, + "learning_rate": 0.0006916268806419258, + "loss": 1.6635, + "theoretical_loss": 3.9645737392799516, + "tokens_seen": 475594752 + }, + { + "epoch": 0.14, + "learning_rate": 0.0006916108324974925, + "loss": 1.5402, + "theoretical_loss": 3.964514819524224, + "tokens_seen": 475660288 + }, + { + "epoch": 0.14, + "learning_rate": 0.0006915947843530592, + "loss": 1.5433, + "theoretical_loss": 3.964455910158498, + "tokens_seen": 475725824 + }, + { + "epoch": 0.14, + "learning_rate": 0.000691578736208626, + "loss": 1.4981, + "theoretical_loss": 3.9643970111795115, + "tokens_seen": 475791360 + }, + { + "epoch": 0.14, + "learning_rate": 0.0006915626880641926, + "loss": 1.6454, + "theoretical_loss": 3.964338122584003, + "tokens_seen": 475856896 + }, + { + "epoch": 0.14, + "learning_rate": 0.0006915466399197593, + "loss": 1.4987, + "theoretical_loss": 3.964279244368712, + "tokens_seen": 475922432 + }, + { + "epoch": 0.14, + "learning_rate": 0.000691530591775326, + "loss": 1.5107, + "theoretical_loss": 3.9642203765303794, + "tokens_seen": 475987968 + }, + { + "epoch": 0.14, + "learning_rate": 0.0006915145436308928, + "loss": 1.4191, + "theoretical_loss": 3.9641615190657484, + "tokens_seen": 476053504 + }, + { + "epoch": 0.14, + "learning_rate": 0.0006914984954864594, + "loss": 1.603, + "theoretical_loss": 3.9641026719715633, + "tokens_seen": 476119040 + }, + { + "epoch": 0.14, + "learning_rate": 0.000691482447342026, + "loss": 1.5629, + "theoretical_loss": 3.9640438352445697, + "tokens_seen": 476184576 + }, + { + "epoch": 0.14, + "learning_rate": 0.0006914663991975928, + "loss": 1.6189, + "theoretical_loss": 3.9639850088815143, + "tokens_seen": 476250112 + }, + { + "epoch": 0.14, + "learning_rate": 0.0006914503510531594, + "loss": 1.5293, + "theoretical_loss": 3.9639261928791463, + "tokens_seen": 476315648 + }, + { + "epoch": 0.14, + "learning_rate": 0.0006914343029087262, + "loss": 1.5161, + "theoretical_loss": 3.963867387234216, + "tokens_seen": 476381184 + }, + { + "epoch": 0.14, + "learning_rate": 0.0006914182547642928, + "loss": 1.5027, + "theoretical_loss": 3.963808591943474, + "tokens_seen": 476446720 + }, + { + "epoch": 0.14, + "learning_rate": 0.0006914022066198596, + "loss": 1.526, + "theoretical_loss": 3.9637498070036745, + "tokens_seen": 476512256 + }, + { + "epoch": 0.14, + "learning_rate": 0.0006913861584754262, + "loss": 1.5956, + "theoretical_loss": 3.963691032411571, + "tokens_seen": 476577792 + }, + { + "epoch": 0.14, + "learning_rate": 0.000691370110330993, + "loss": 1.719, + "theoretical_loss": 3.9636322681639196, + "tokens_seen": 476643328 + }, + { + "epoch": 0.14, + "learning_rate": 0.0006913540621865596, + "loss": 1.476, + "theoretical_loss": 3.9635735142574786, + "tokens_seen": 476708864 + }, + { + "epoch": 0.14, + "learning_rate": 0.0006913380140421264, + "loss": 1.6634, + "theoretical_loss": 3.963514770689006, + "tokens_seen": 476774400 + }, + { + "epoch": 0.14, + "learning_rate": 0.000691321965897693, + "loss": 1.5057, + "theoretical_loss": 3.9634560374552623, + "tokens_seen": 476839936 + }, + { + "epoch": 0.14, + "learning_rate": 0.0006913059177532598, + "loss": 1.5831, + "theoretical_loss": 3.96339731455301, + "tokens_seen": 476905472 + }, + { + "epoch": 0.14, + "learning_rate": 0.0006912898696088265, + "loss": 1.4317, + "theoretical_loss": 3.963338601979011, + "tokens_seen": 476971008 + }, + { + "epoch": 0.14, + "learning_rate": 0.0006912738214643932, + "loss": 1.5479, + "theoretical_loss": 3.9632798997300314, + "tokens_seen": 477036544 + }, + { + "epoch": 0.14, + "learning_rate": 0.0006912577733199599, + "loss": 1.6421, + "theoretical_loss": 3.963221207802836, + "tokens_seen": 477102080 + }, + { + "epoch": 0.14, + "learning_rate": 0.0006912417251755266, + "loss": 1.4193, + "theoretical_loss": 3.9631625261941945, + "tokens_seen": 477167616 + }, + { + "epoch": 0.14, + "learning_rate": 0.0006912256770310933, + "loss": 1.5326, + "theoretical_loss": 3.963103854900874, + "tokens_seen": 477233152 + }, + { + "epoch": 0.14, + "learning_rate": 0.00069120962888666, + "loss": 1.5493, + "theoretical_loss": 3.963045193919646, + "tokens_seen": 477298688 + }, + { + "epoch": 0.14, + "learning_rate": 0.0006911935807422267, + "loss": 1.4845, + "theoretical_loss": 3.9629865432472817, + "tokens_seen": 477364224 + }, + { + "epoch": 0.14, + "learning_rate": 0.0006911775325977934, + "loss": 1.5931, + "theoretical_loss": 3.9629279028805553, + "tokens_seen": 477429760 + }, + { + "epoch": 0.14, + "learning_rate": 0.0006911614844533601, + "loss": 1.6728, + "theoretical_loss": 3.962869272816242, + "tokens_seen": 477495296 + }, + { + "epoch": 0.14, + "learning_rate": 0.0006911454363089268, + "loss": 1.6199, + "theoretical_loss": 3.9628106530511173, + "tokens_seen": 477560832 + }, + { + "epoch": 0.14, + "learning_rate": 0.0006911293881644935, + "loss": 1.6317, + "theoretical_loss": 3.9627520435819594, + "tokens_seen": 477626368 + }, + { + "epoch": 0.14, + "learning_rate": 0.0006911133400200602, + "loss": 1.5913, + "theoretical_loss": 3.9626934444055477, + "tokens_seen": 477691904 + }, + { + "epoch": 0.14, + "learning_rate": 0.0006910972918756269, + "loss": 1.5529, + "theoretical_loss": 3.962634855518663, + "tokens_seen": 477757440 + }, + { + "epoch": 0.14, + "learning_rate": 0.0006910812437311936, + "loss": 1.4957, + "theoretical_loss": 3.962576276918087, + "tokens_seen": 477822976 + }, + { + "epoch": 0.14, + "learning_rate": 0.0006910651955867603, + "loss": 1.5986, + "theoretical_loss": 3.962517708600603, + "tokens_seen": 477888512 + }, + { + "epoch": 0.14, + "learning_rate": 0.000691049147442327, + "loss": 1.622, + "theoretical_loss": 3.9624591505629976, + "tokens_seen": 477954048 + }, + { + "epoch": 0.14, + "learning_rate": 0.0006910330992978937, + "loss": 1.6647, + "theoretical_loss": 3.962400602802056, + "tokens_seen": 478019584 + }, + { + "epoch": 0.14, + "learning_rate": 0.0006910170511534604, + "loss": 1.5956, + "theoretical_loss": 3.962342065314566, + "tokens_seen": 478085120 + }, + { + "epoch": 0.14, + "learning_rate": 0.0006910010030090271, + "loss": 1.5523, + "theoretical_loss": 3.962283538097318, + "tokens_seen": 478150656 + }, + { + "epoch": 0.14, + "learning_rate": 0.0006909849548645938, + "loss": 1.5795, + "theoretical_loss": 3.962225021147102, + "tokens_seen": 478216192 + }, + { + "epoch": 0.14, + "learning_rate": 0.0006909689067201605, + "loss": 1.4795, + "theoretical_loss": 3.9621665144607103, + "tokens_seen": 478281728 + }, + { + "epoch": 0.14, + "learning_rate": 0.0006909528585757272, + "loss": 1.6069, + "theoretical_loss": 3.9621080180349373, + "tokens_seen": 478347264 + }, + { + "epoch": 0.14, + "objective/train/docs_used": 180329, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.6280412673950195, + "objective/train/theoretical_loss": 3.9620495318665774, + "objective/train/tokens_used": 498872800, + "theoretical_loss": 3.9620495318665774, + "tokens_seen": 478412800 + }, + { + "epoch": 0.14, + "learning_rate": 0.0006909368104312939, + "loss": 1.6225, + "theoretical_loss": 3.9620495318665774, + "tokens_seen": 478412800 + }, + { + "epoch": 0.14, + "learning_rate": 0.0006909207622868606, + "loss": 1.6909, + "theoretical_loss": 3.9619910559524274, + "tokens_seen": 478478336 + }, + { + "epoch": 0.15, + "learning_rate": 0.0006909047141424273, + "loss": 1.5174, + "theoretical_loss": 3.9619325902892855, + "tokens_seen": 478543872 + }, + { + "epoch": 0.15, + "learning_rate": 0.000690888665997994, + "loss": 1.568, + "theoretical_loss": 3.961874134873951, + "tokens_seen": 478609408 + }, + { + "epoch": 0.15, + "learning_rate": 0.0006908726178535607, + "loss": 1.5719, + "theoretical_loss": 3.9618156897032257, + "tokens_seen": 478674944 + }, + { + "epoch": 0.15, + "learning_rate": 0.0006908565697091274, + "loss": 1.6408, + "theoretical_loss": 3.961757254773911, + "tokens_seen": 478740480 + }, + { + "epoch": 0.15, + "learning_rate": 0.0006908405215646941, + "loss": 1.5438, + "theoretical_loss": 3.9616988300828098, + "tokens_seen": 478806016 + }, + { + "epoch": 0.15, + "learning_rate": 0.0006908244734202608, + "loss": 1.5945, + "theoretical_loss": 3.9616404156267286, + "tokens_seen": 478871552 + }, + { + "epoch": 0.15, + "learning_rate": 0.0006908084252758275, + "loss": 1.4828, + "theoretical_loss": 3.961582011402475, + "tokens_seen": 478937088 + }, + { + "epoch": 0.15, + "learning_rate": 0.0006907923771313942, + "loss": 1.4829, + "theoretical_loss": 3.9615236174068547, + "tokens_seen": 479002624 + }, + { + "epoch": 0.15, + "learning_rate": 0.0006907763289869609, + "loss": 1.6148, + "theoretical_loss": 3.961465233636679, + "tokens_seen": 479068160 + }, + { + "epoch": 0.15, + "learning_rate": 0.0006907602808425276, + "loss": 1.6042, + "theoretical_loss": 3.9614068600887578, + "tokens_seen": 479133696 + }, + { + "epoch": 0.15, + "learning_rate": 0.0006907442326980943, + "loss": 1.5493, + "theoretical_loss": 3.9613484967599044, + "tokens_seen": 479199232 + }, + { + "epoch": 0.15, + "learning_rate": 0.000690728184553661, + "loss": 1.4159, + "theoretical_loss": 3.961290143646932, + "tokens_seen": 479264768 + }, + { + "epoch": 0.15, + "learning_rate": 0.0006907121364092277, + "loss": 1.654, + "theoretical_loss": 3.961231800746656, + "tokens_seen": 479330304 + }, + { + "epoch": 0.15, + "learning_rate": 0.0006906960882647944, + "loss": 1.518, + "theoretical_loss": 3.961173468055893, + "tokens_seen": 479395840 + }, + { + "epoch": 0.15, + "learning_rate": 0.0006906800401203612, + "loss": 1.4375, + "theoretical_loss": 3.961115145571461, + "tokens_seen": 479461376 + }, + { + "epoch": 0.15, + "learning_rate": 0.0006906639919759278, + "loss": 1.5844, + "theoretical_loss": 3.96105683329018, + "tokens_seen": 479526912 + }, + { + "epoch": 0.15, + "learning_rate": 0.0006906479438314946, + "loss": 1.5608, + "theoretical_loss": 3.9609985312088707, + "tokens_seen": 479592448 + }, + { + "epoch": 0.15, + "learning_rate": 0.0006906318956870612, + "loss": 1.6135, + "theoretical_loss": 3.960940239324355, + "tokens_seen": 479657984 + }, + { + "epoch": 0.15, + "learning_rate": 0.000690615847542628, + "loss": 1.6013, + "theoretical_loss": 3.960881957633457, + "tokens_seen": 479723520 + }, + { + "epoch": 0.15, + "learning_rate": 0.0006905997993981946, + "loss": 1.6409, + "theoretical_loss": 3.9608236861330024, + "tokens_seen": 479789056 + }, + { + "epoch": 0.15, + "learning_rate": 0.0006905837512537614, + "loss": 1.563, + "theoretical_loss": 3.960765424819817, + "tokens_seen": 479854592 + }, + { + "epoch": 0.15, + "learning_rate": 0.000690567703109328, + "loss": 1.5131, + "theoretical_loss": 3.9607071736907296, + "tokens_seen": 479920128 + }, + { + "epoch": 0.15, + "learning_rate": 0.0006905516549648948, + "loss": 1.7211, + "theoretical_loss": 3.9606489327425694, + "tokens_seen": 479985664 + }, + { + "epoch": 0.15, + "learning_rate": 0.0006905356068204614, + "loss": 1.626, + "theoretical_loss": 3.960590701972167, + "tokens_seen": 480051200 + }, + { + "epoch": 0.15, + "learning_rate": 0.0006905195586760282, + "loss": 1.5052, + "theoretical_loss": 3.960532481376355, + "tokens_seen": 480116736 + }, + { + "epoch": 0.15, + "learning_rate": 0.0006905035105315948, + "loss": 1.5383, + "theoretical_loss": 3.960474270951967, + "tokens_seen": 480182272 + }, + { + "epoch": 0.15, + "learning_rate": 0.0006904874623871616, + "loss": 1.5008, + "theoretical_loss": 3.9604160706958385, + "tokens_seen": 480247808 + }, + { + "epoch": 0.15, + "learning_rate": 0.0006904714142427282, + "loss": 1.5922, + "theoretical_loss": 3.960357880604806, + "tokens_seen": 480313344 + }, + { + "epoch": 0.15, + "learning_rate": 0.000690455366098295, + "loss": 1.5382, + "theoretical_loss": 3.9602997006757072, + "tokens_seen": 480378880 + }, + { + "epoch": 0.15, + "learning_rate": 0.0006904393179538616, + "loss": 1.6206, + "theoretical_loss": 3.960241530905382, + "tokens_seen": 480444416 + }, + { + "epoch": 0.15, + "learning_rate": 0.0006904232698094284, + "loss": 1.5673, + "theoretical_loss": 3.96018337129067, + "tokens_seen": 480509952 + }, + { + "epoch": 0.15, + "learning_rate": 0.000690407221664995, + "loss": 1.6695, + "theoretical_loss": 3.960125221828415, + "tokens_seen": 480575488 + }, + { + "epoch": 0.15, + "learning_rate": 0.0006903911735205618, + "loss": 1.6511, + "theoretical_loss": 3.9600670825154594, + "tokens_seen": 480641024 + }, + { + "epoch": 0.15, + "learning_rate": 0.0006903751253761284, + "loss": 1.651, + "theoretical_loss": 3.960008953348649, + "tokens_seen": 480706560 + }, + { + "epoch": 0.15, + "learning_rate": 0.0006903590772316952, + "loss": 1.5801, + "theoretical_loss": 3.9599508343248297, + "tokens_seen": 480772096 + }, + { + "epoch": 0.15, + "learning_rate": 0.0006903430290872618, + "loss": 1.6778, + "theoretical_loss": 3.9598927254408505, + "tokens_seen": 480837632 + }, + { + "epoch": 0.15, + "learning_rate": 0.0006903269809428286, + "loss": 1.622, + "theoretical_loss": 3.9598346266935596, + "tokens_seen": 480903168 + }, + { + "epoch": 0.15, + "learning_rate": 0.0006903109327983952, + "loss": 1.5446, + "theoretical_loss": 3.959776538079808, + "tokens_seen": 480968704 + }, + { + "epoch": 0.15, + "learning_rate": 0.000690294884653962, + "loss": 1.5528, + "theoretical_loss": 3.9597184595964476, + "tokens_seen": 481034240 + }, + { + "epoch": 0.15, + "learning_rate": 0.0006902788365095286, + "loss": 1.4718, + "theoretical_loss": 3.959660391240332, + "tokens_seen": 481099776 + }, + { + "epoch": 0.15, + "learning_rate": 0.0006902627883650954, + "loss": 1.5346, + "theoretical_loss": 3.9596023330083163, + "tokens_seen": 481165312 + }, + { + "epoch": 0.15, + "learning_rate": 0.000690246740220662, + "loss": 1.5394, + "theoretical_loss": 3.9595442848972566, + "tokens_seen": 481230848 + }, + { + "epoch": 0.15, + "learning_rate": 0.0006902306920762287, + "loss": 1.6351, + "theoretical_loss": 3.959486246904011, + "tokens_seen": 481296384 + }, + { + "epoch": 0.15, + "learning_rate": 0.0006902146439317954, + "loss": 1.5559, + "theoretical_loss": 3.959428219025438, + "tokens_seen": 481361920 + }, + { + "epoch": 0.15, + "learning_rate": 0.0006901985957873621, + "loss": 1.5927, + "theoretical_loss": 3.9593702012583987, + "tokens_seen": 481427456 + }, + { + "epoch": 0.15, + "learning_rate": 0.0006901825476429288, + "loss": 1.5695, + "theoretical_loss": 3.9593121935997546, + "tokens_seen": 481492992 + }, + { + "epoch": 0.15, + "learning_rate": 0.0006901664994984955, + "loss": 1.5334, + "theoretical_loss": 3.95925419604637, + "tokens_seen": 481558528 + }, + { + "epoch": 0.15, + "learning_rate": 0.0006901504513540622, + "loss": 1.5211, + "theoretical_loss": 3.959196208595108, + "tokens_seen": 481624064 + }, + { + "epoch": 0.15, + "objective/train/docs_used": 181505, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.5555431842803955, + "objective/train/theoretical_loss": 3.959138231242836, + "objective/train/tokens_used": 502149600, + "theoretical_loss": 3.959138231242836, + "tokens_seen": 481689600 + }, + { + "epoch": 0.15, + "learning_rate": 0.0006901344032096289, + "loss": 1.5169, + "theoretical_loss": 3.959138231242836, + "tokens_seen": 481689600 + }, + { + "epoch": 0.15, + "learning_rate": 0.0006901183550651957, + "loss": 1.6592, + "theoretical_loss": 3.959080263986421, + "tokens_seen": 481755136 + }, + { + "epoch": 0.15, + "learning_rate": 0.0006901023069207623, + "loss": 1.5401, + "theoretical_loss": 3.9590223068227317, + "tokens_seen": 481820672 + }, + { + "epoch": 0.15, + "learning_rate": 0.0006900862587763289, + "loss": 1.5881, + "theoretical_loss": 3.9589643597486393, + "tokens_seen": 481886208 + }, + { + "epoch": 0.15, + "learning_rate": 0.0006900702106318957, + "loss": 1.6444, + "theoretical_loss": 3.9589064227610145, + "tokens_seen": 481951744 + }, + { + "epoch": 0.15, + "learning_rate": 0.0006900541624874623, + "loss": 1.5983, + "theoretical_loss": 3.9588484958567314, + "tokens_seen": 482017280 + }, + { + "epoch": 0.15, + "learning_rate": 0.0006900381143430291, + "loss": 1.6528, + "theoretical_loss": 3.958790579032663, + "tokens_seen": 482082816 + }, + { + "epoch": 0.15, + "learning_rate": 0.0006900220661985957, + "loss": 1.6069, + "theoretical_loss": 3.958732672285687, + "tokens_seen": 482148352 + }, + { + "epoch": 0.15, + "learning_rate": 0.0006900060180541625, + "loss": 1.6706, + "theoretical_loss": 3.95867477561268, + "tokens_seen": 482213888 + }, + { + "epoch": 0.15, + "learning_rate": 0.0006899899699097291, + "loss": 1.5709, + "theoretical_loss": 3.95861688901052, + "tokens_seen": 482279424 + }, + { + "epoch": 0.15, + "learning_rate": 0.0006899739217652959, + "loss": 1.4834, + "theoretical_loss": 3.958559012476088, + "tokens_seen": 482344960 + }, + { + "epoch": 0.15, + "learning_rate": 0.0006899578736208625, + "loss": 1.5225, + "theoretical_loss": 3.958501146006265, + "tokens_seen": 482410496 + }, + { + "epoch": 0.15, + "learning_rate": 0.0006899418254764293, + "loss": 1.5906, + "theoretical_loss": 3.958443289597934, + "tokens_seen": 482476032 + }, + { + "epoch": 0.15, + "learning_rate": 0.000689925777331996, + "loss": 1.5914, + "theoretical_loss": 3.958385443247979, + "tokens_seen": 482541568 + }, + { + "epoch": 0.15, + "learning_rate": 0.0006899097291875627, + "loss": 1.5489, + "theoretical_loss": 3.958327606953286, + "tokens_seen": 482607104 + }, + { + "epoch": 0.15, + "learning_rate": 0.0006898936810431294, + "loss": 1.5948, + "theoretical_loss": 3.958269780710742, + "tokens_seen": 482672640 + }, + { + "epoch": 0.15, + "learning_rate": 0.0006898776328986961, + "loss": 1.5587, + "theoretical_loss": 3.9582119645172344, + "tokens_seen": 482738176 + }, + { + "epoch": 0.15, + "learning_rate": 0.0006898615847542628, + "loss": 1.4786, + "theoretical_loss": 3.9581541583696547, + "tokens_seen": 482803712 + }, + { + "epoch": 0.15, + "learning_rate": 0.0006898455366098295, + "loss": 1.5694, + "theoretical_loss": 3.958096362264893, + "tokens_seen": 482869248 + }, + { + "epoch": 0.15, + "learning_rate": 0.0006898294884653962, + "loss": 1.6169, + "theoretical_loss": 3.958038576199842, + "tokens_seen": 482934784 + }, + { + "epoch": 0.15, + "learning_rate": 0.0006898134403209629, + "loss": 1.6399, + "theoretical_loss": 3.957980800171396, + "tokens_seen": 483000320 + }, + { + "epoch": 0.15, + "learning_rate": 0.0006897973921765296, + "loss": 1.5865, + "theoretical_loss": 3.9579230341764493, + "tokens_seen": 483065856 + }, + { + "epoch": 0.15, + "learning_rate": 0.0006897813440320963, + "loss": 1.6037, + "theoretical_loss": 3.9578652782118997, + "tokens_seen": 483131392 + }, + { + "epoch": 0.15, + "learning_rate": 0.000689765295887663, + "loss": 1.5828, + "theoretical_loss": 3.957807532274645, + "tokens_seen": 483196928 + }, + { + "epoch": 0.15, + "learning_rate": 0.0006897492477432297, + "loss": 1.3385, + "theoretical_loss": 3.957749796361585, + "tokens_seen": 483262464 + }, + { + "epoch": 0.15, + "learning_rate": 0.0006897331995987964, + "loss": 1.548, + "theoretical_loss": 3.9576920704696192, + "tokens_seen": 483328000 + }, + { + "epoch": 0.15, + "learning_rate": 0.0006897171514543631, + "loss": 1.529, + "theoretical_loss": 3.957634354595651, + "tokens_seen": 483393536 + }, + { + "epoch": 0.15, + "learning_rate": 0.0006897011033099298, + "loss": 1.5756, + "theoretical_loss": 3.957576648736584, + "tokens_seen": 483459072 + }, + { + "epoch": 0.15, + "learning_rate": 0.0006896850551654965, + "loss": 1.5035, + "theoretical_loss": 3.9575189528893233, + "tokens_seen": 483524608 + }, + { + "epoch": 0.15, + "learning_rate": 0.0006896690070210632, + "loss": 1.518, + "theoretical_loss": 3.957461267050774, + "tokens_seen": 483590144 + }, + { + "epoch": 0.15, + "learning_rate": 0.0006896529588766299, + "loss": 1.5868, + "theoretical_loss": 3.957403591217845, + "tokens_seen": 483655680 + }, + { + "epoch": 0.15, + "learning_rate": 0.0006896369107321966, + "loss": 1.6081, + "theoretical_loss": 3.9573459253874455, + "tokens_seen": 483721216 + }, + { + "epoch": 0.15, + "learning_rate": 0.0006896208625877633, + "loss": 1.6025, + "theoretical_loss": 3.9572882695564853, + "tokens_seen": 483786752 + }, + { + "epoch": 0.15, + "learning_rate": 0.00068960481444333, + "loss": 1.5226, + "theoretical_loss": 3.9572306237218764, + "tokens_seen": 483852288 + }, + { + "epoch": 0.15, + "learning_rate": 0.0006895887662988967, + "loss": 1.4722, + "theoretical_loss": 3.957172987880532, + "tokens_seen": 483917824 + }, + { + "epoch": 0.15, + "learning_rate": 0.0006895727181544634, + "loss": 1.5638, + "theoretical_loss": 3.9571153620293673, + "tokens_seen": 483983360 + }, + { + "epoch": 0.15, + "learning_rate": 0.0006895566700100301, + "loss": 1.5938, + "theoretical_loss": 3.9570577461652974, + "tokens_seen": 484048896 + }, + { + "epoch": 0.15, + "learning_rate": 0.0006895406218655968, + "loss": 1.6579, + "theoretical_loss": 3.95700014028524, + "tokens_seen": 484114432 + }, + { + "epoch": 0.15, + "learning_rate": 0.0006895245737211635, + "loss": 1.4982, + "theoretical_loss": 3.956942544386114, + "tokens_seen": 484179968 + }, + { + "epoch": 0.15, + "learning_rate": 0.0006895085255767302, + "loss": 1.5697, + "theoretical_loss": 3.9568849584648396, + "tokens_seen": 484245504 + }, + { + "epoch": 0.15, + "learning_rate": 0.0006894924774322969, + "loss": 1.5134, + "theoretical_loss": 3.9568273825183375, + "tokens_seen": 484311040 + }, + { + "epoch": 0.15, + "learning_rate": 0.0006894764292878637, + "loss": 1.6008, + "theoretical_loss": 3.9567698165435305, + "tokens_seen": 484376576 + }, + { + "epoch": 0.15, + "learning_rate": 0.0006894603811434303, + "loss": 1.7031, + "theoretical_loss": 3.9567122605373437, + "tokens_seen": 484442112 + }, + { + "epoch": 0.15, + "learning_rate": 0.0006894443329989971, + "loss": 1.5118, + "theoretical_loss": 3.956654714496702, + "tokens_seen": 484507648 + }, + { + "epoch": 0.15, + "learning_rate": 0.0006894282848545637, + "loss": 1.5885, + "theoretical_loss": 3.9565971784185328, + "tokens_seen": 484573184 + }, + { + "epoch": 0.15, + "learning_rate": 0.0006894122367101305, + "loss": 1.5926, + "theoretical_loss": 3.9565396522997633, + "tokens_seen": 484638720 + }, + { + "epoch": 0.15, + "learning_rate": 0.0006893961885656971, + "loss": 1.5639, + "theoretical_loss": 3.9564821361373244, + "tokens_seen": 484704256 + }, + { + "epoch": 0.15, + "learning_rate": 0.0006893801404212639, + "loss": 1.5981, + "theoretical_loss": 3.956424629928146, + "tokens_seen": 484769792 + }, + { + "epoch": 0.15, + "learning_rate": 0.0006893640922768305, + "loss": 1.5626, + "theoretical_loss": 3.9563671336691613, + "tokens_seen": 484835328 + }, + { + "epoch": 0.15, + "learning_rate": 0.0006893480441323973, + "loss": 1.6562, + "theoretical_loss": 3.956309647357303, + "tokens_seen": 484900864 + }, + { + "epoch": 0.15, + "objective/train/docs_used": 182723, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.3889405727386475, + "objective/train/theoretical_loss": 3.956252170989507, + "objective/train/tokens_used": 505426400, + "theoretical_loss": 3.956252170989507, + "tokens_seen": 484966400 + }, + { + "epoch": 0.15, + "learning_rate": 0.0006893319959879639, + "loss": 1.4917, + "theoretical_loss": 3.956252170989507, + "tokens_seen": 484966400 + }, + { + "epoch": 0.15, + "learning_rate": 0.0006893159478435307, + "loss": 1.5568, + "theoretical_loss": 3.9561947045627095, + "tokens_seen": 485031936 + }, + { + "epoch": 0.15, + "learning_rate": 0.0006892998996990973, + "loss": 1.6379, + "theoretical_loss": 3.9561372480738486, + "tokens_seen": 485097472 + }, + { + "epoch": 0.15, + "learning_rate": 0.0006892838515546641, + "loss": 1.5673, + "theoretical_loss": 3.956079801519863, + "tokens_seen": 485163008 + }, + { + "epoch": 0.15, + "learning_rate": 0.0006892678034102307, + "loss": 1.5759, + "theoretical_loss": 3.9560223648976924, + "tokens_seen": 485228544 + }, + { + "epoch": 0.15, + "learning_rate": 0.0006892517552657975, + "loss": 1.6457, + "theoretical_loss": 3.9559649382042803, + "tokens_seen": 485294080 + }, + { + "epoch": 0.15, + "learning_rate": 0.0006892357071213641, + "loss": 1.5512, + "theoretical_loss": 3.955907521436569, + "tokens_seen": 485359616 + }, + { + "epoch": 0.15, + "learning_rate": 0.0006892196589769309, + "loss": 1.659, + "theoretical_loss": 3.9558501145915033, + "tokens_seen": 485425152 + }, + { + "epoch": 0.15, + "learning_rate": 0.0006892036108324975, + "loss": 1.5669, + "theoretical_loss": 3.9557927176660286, + "tokens_seen": 485490688 + }, + { + "epoch": 0.15, + "learning_rate": 0.0006891875626880643, + "loss": 1.6013, + "theoretical_loss": 3.9557353306570926, + "tokens_seen": 485556224 + }, + { + "epoch": 0.15, + "learning_rate": 0.0006891715145436309, + "loss": 1.5442, + "theoretical_loss": 3.9556779535616444, + "tokens_seen": 485621760 + }, + { + "epoch": 0.15, + "learning_rate": 0.0006891554663991977, + "loss": 1.6068, + "theoretical_loss": 3.955620586376633, + "tokens_seen": 485687296 + }, + { + "epoch": 0.15, + "learning_rate": 0.0006891394182547643, + "loss": 1.5302, + "theoretical_loss": 3.9555632290990097, + "tokens_seen": 485752832 + }, + { + "epoch": 0.15, + "learning_rate": 0.0006891233701103311, + "loss": 1.6401, + "theoretical_loss": 3.9555058817257285, + "tokens_seen": 485818368 + }, + { + "epoch": 0.15, + "learning_rate": 0.0006891073219658977, + "loss": 1.5672, + "theoretical_loss": 3.9554485442537417, + "tokens_seen": 485883904 + }, + { + "epoch": 0.15, + "learning_rate": 0.0006890912738214645, + "loss": 1.4391, + "theoretical_loss": 3.955391216680006, + "tokens_seen": 485949440 + }, + { + "epoch": 0.15, + "learning_rate": 0.0006890752256770311, + "loss": 1.6084, + "theoretical_loss": 3.9553338990014772, + "tokens_seen": 486014976 + }, + { + "epoch": 0.15, + "learning_rate": 0.0006890591775325979, + "loss": 1.5746, + "theoretical_loss": 3.9552765912151138, + "tokens_seen": 486080512 + }, + { + "epoch": 0.15, + "learning_rate": 0.0006890431293881645, + "loss": 1.6001, + "theoretical_loss": 3.9552192933178745, + "tokens_seen": 486146048 + }, + { + "epoch": 0.15, + "learning_rate": 0.0006890270812437313, + "loss": 1.5693, + "theoretical_loss": 3.9551620053067214, + "tokens_seen": 486211584 + }, + { + "epoch": 0.15, + "learning_rate": 0.0006890110330992979, + "loss": 1.4526, + "theoretical_loss": 3.9551047271786155, + "tokens_seen": 486277120 + }, + { + "epoch": 0.15, + "learning_rate": 0.0006889949849548647, + "loss": 1.7152, + "theoretical_loss": 3.9550474589305202, + "tokens_seen": 486342656 + }, + { + "epoch": 0.15, + "learning_rate": 0.0006889789368104313, + "loss": 1.5471, + "theoretical_loss": 3.954990200559401, + "tokens_seen": 486408192 + }, + { + "epoch": 0.15, + "learning_rate": 0.0006889628886659981, + "loss": 1.5489, + "theoretical_loss": 3.9549329520622236, + "tokens_seen": 486473728 + }, + { + "epoch": 0.15, + "learning_rate": 0.0006889468405215647, + "loss": 1.4727, + "theoretical_loss": 3.954875713435955, + "tokens_seen": 486539264 + }, + { + "epoch": 0.15, + "learning_rate": 0.0006889307923771315, + "loss": 1.5463, + "theoretical_loss": 3.9548184846775647, + "tokens_seen": 486604800 + }, + { + "epoch": 0.15, + "learning_rate": 0.0006889147442326981, + "loss": 1.5647, + "theoretical_loss": 3.9547612657840228, + "tokens_seen": 486670336 + }, + { + "epoch": 0.15, + "learning_rate": 0.0006888986960882649, + "loss": 1.5755, + "theoretical_loss": 3.9547040567523, + "tokens_seen": 486735872 + }, + { + "epoch": 0.15, + "learning_rate": 0.0006888826479438315, + "loss": 1.6198, + "theoretical_loss": 3.9546468575793696, + "tokens_seen": 486801408 + }, + { + "epoch": 0.15, + "learning_rate": 0.0006888665997993983, + "loss": 1.5809, + "theoretical_loss": 3.9545896682622064, + "tokens_seen": 486866944 + }, + { + "epoch": 0.15, + "learning_rate": 0.000688850551654965, + "loss": 1.5186, + "theoretical_loss": 3.954532488797785, + "tokens_seen": 486932480 + }, + { + "epoch": 0.15, + "learning_rate": 0.0006888345035105317, + "loss": 1.5148, + "theoretical_loss": 3.954475319183082, + "tokens_seen": 486998016 + }, + { + "epoch": 0.15, + "learning_rate": 0.0006888184553660983, + "loss": 1.5552, + "theoretical_loss": 3.9544181594150762, + "tokens_seen": 487063552 + }, + { + "epoch": 0.15, + "learning_rate": 0.000688802407221665, + "loss": 1.5844, + "theoretical_loss": 3.9543610094907464, + "tokens_seen": 487129088 + }, + { + "epoch": 0.15, + "learning_rate": 0.0006887863590772317, + "loss": 1.4613, + "theoretical_loss": 3.9543038694070747, + "tokens_seen": 487194624 + }, + { + "epoch": 0.15, + "learning_rate": 0.0006887703109327984, + "loss": 1.5161, + "theoretical_loss": 3.9542467391610416, + "tokens_seen": 487260160 + }, + { + "epoch": 0.15, + "learning_rate": 0.0006887542627883652, + "loss": 1.5986, + "theoretical_loss": 3.9541896187496315, + "tokens_seen": 487325696 + }, + { + "epoch": 0.15, + "learning_rate": 0.0006887382146439318, + "loss": 1.5111, + "theoretical_loss": 3.9541325081698293, + "tokens_seen": 487391232 + }, + { + "epoch": 0.15, + "learning_rate": 0.0006887221664994986, + "loss": 1.5661, + "theoretical_loss": 3.954075407418621, + "tokens_seen": 487456768 + }, + { + "epoch": 0.15, + "learning_rate": 0.0006887061183550652, + "loss": 1.4544, + "theoretical_loss": 3.954018316492994, + "tokens_seen": 487522304 + }, + { + "epoch": 0.15, + "learning_rate": 0.0006886900702106318, + "loss": 1.4653, + "theoretical_loss": 3.9539612353899365, + "tokens_seen": 487587840 + }, + { + "epoch": 0.15, + "learning_rate": 0.0006886740220661986, + "loss": 1.5049, + "theoretical_loss": 3.9539041641064396, + "tokens_seen": 487653376 + }, + { + "epoch": 0.15, + "learning_rate": 0.0006886579739217652, + "loss": 1.4901, + "theoretical_loss": 3.953847102639494, + "tokens_seen": 487718912 + }, + { + "epoch": 0.15, + "learning_rate": 0.000688641925777332, + "loss": 1.6048, + "theoretical_loss": 3.9537900509860933, + "tokens_seen": 487784448 + }, + { + "epoch": 0.15, + "learning_rate": 0.0006886258776328986, + "loss": 1.662, + "theoretical_loss": 3.953733009143231, + "tokens_seen": 487849984 + }, + { + "epoch": 0.15, + "learning_rate": 0.0006886098294884654, + "loss": 1.509, + "theoretical_loss": 3.9536759771079018, + "tokens_seen": 487915520 + }, + { + "epoch": 0.15, + "learning_rate": 0.000688593781344032, + "loss": 1.5402, + "theoretical_loss": 3.9536189548771037, + "tokens_seen": 487981056 + }, + { + "epoch": 0.15, + "learning_rate": 0.0006885777331995988, + "loss": 1.5686, + "theoretical_loss": 3.953561942447834, + "tokens_seen": 488046592 + }, + { + "epoch": 0.15, + "learning_rate": 0.0006885616850551655, + "loss": 1.5209, + "theoretical_loss": 3.953504939817093, + "tokens_seen": 488112128 + }, + { + "epoch": 0.15, + "learning_rate": 0.0006885456369107322, + "loss": 1.5404, + "theoretical_loss": 3.953447946981881, + "tokens_seen": 488177664 + }, + { + "epoch": 0.15, + "objective/train/docs_used": 183865, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.6293010711669922, + "objective/train/theoretical_loss": 3.9533909639391993, + "objective/train/tokens_used": 508703200, + "theoretical_loss": 3.9533909639391993, + "tokens_seen": 488243200 + }, + { + "epoch": 0.15, + "learning_rate": 0.0006885295887662989, + "loss": 1.6437, + "theoretical_loss": 3.9533909639391993, + "tokens_seen": 488243200 + }, + { + "epoch": 0.15, + "learning_rate": 0.0006885135406218656, + "loss": 1.515, + "theoretical_loss": 3.953333990686051, + "tokens_seen": 488308736 + }, + { + "epoch": 0.15, + "learning_rate": 0.0006884974924774323, + "loss": 1.5776, + "theoretical_loss": 3.9532770272194426, + "tokens_seen": 488374272 + }, + { + "epoch": 0.15, + "learning_rate": 0.000688481444332999, + "loss": 1.5226, + "theoretical_loss": 3.9532200735363787, + "tokens_seen": 488439808 + }, + { + "epoch": 0.15, + "learning_rate": 0.0006884653961885657, + "loss": 1.6289, + "theoretical_loss": 3.953163129633867, + "tokens_seen": 488505344 + }, + { + "epoch": 0.15, + "learning_rate": 0.0006884493480441324, + "loss": 1.398, + "theoretical_loss": 3.9531061955089157, + "tokens_seen": 488570880 + }, + { + "epoch": 0.15, + "learning_rate": 0.0006884332998996991, + "loss": 1.5191, + "theoretical_loss": 3.9530492711585357, + "tokens_seen": 488636416 + }, + { + "epoch": 0.15, + "learning_rate": 0.0006884172517552658, + "loss": 1.562, + "theoretical_loss": 3.9529923565797374, + "tokens_seen": 488701952 + }, + { + "epoch": 0.15, + "learning_rate": 0.0006884012036108325, + "loss": 1.495, + "theoretical_loss": 3.952935451769533, + "tokens_seen": 488767488 + }, + { + "epoch": 0.15, + "learning_rate": 0.0006883851554663992, + "loss": 1.5231, + "theoretical_loss": 3.9528785567249374, + "tokens_seen": 488833024 + }, + { + "epoch": 0.15, + "learning_rate": 0.0006883691073219659, + "loss": 1.594, + "theoretical_loss": 3.9528216714429654, + "tokens_seen": 488898560 + }, + { + "epoch": 0.15, + "learning_rate": 0.0006883530591775326, + "loss": 1.5806, + "theoretical_loss": 3.9527647959206336, + "tokens_seen": 488964096 + }, + { + "epoch": 0.15, + "learning_rate": 0.0006883370110330993, + "loss": 1.5728, + "theoretical_loss": 3.9527079301549595, + "tokens_seen": 489029632 + }, + { + "epoch": 0.15, + "learning_rate": 0.000688320962888666, + "loss": 1.5814, + "theoretical_loss": 3.9526510741429624, + "tokens_seen": 489095168 + }, + { + "epoch": 0.15, + "learning_rate": 0.0006883049147442327, + "loss": 1.5748, + "theoretical_loss": 3.9525942278816633, + "tokens_seen": 489160704 + }, + { + "epoch": 0.15, + "learning_rate": 0.0006882888665997994, + "loss": 1.6044, + "theoretical_loss": 3.952537391368083, + "tokens_seen": 489226240 + }, + { + "epoch": 0.15, + "learning_rate": 0.0006882728184553662, + "loss": 1.4357, + "theoretical_loss": 3.9524805645992447, + "tokens_seen": 489291776 + }, + { + "epoch": 0.15, + "learning_rate": 0.0006882567703109328, + "loss": 1.5734, + "theoretical_loss": 3.9524237475721726, + "tokens_seen": 489357312 + }, + { + "epoch": 0.15, + "learning_rate": 0.0006882407221664996, + "loss": 1.6073, + "theoretical_loss": 3.952366940283894, + "tokens_seen": 489422848 + }, + { + "epoch": 0.15, + "learning_rate": 0.0006882246740220662, + "loss": 1.5627, + "theoretical_loss": 3.952310142731434, + "tokens_seen": 489488384 + }, + { + "epoch": 0.15, + "learning_rate": 0.000688208625877633, + "loss": 1.5924, + "theoretical_loss": 3.9522533549118215, + "tokens_seen": 489553920 + }, + { + "epoch": 0.15, + "learning_rate": 0.0006881925777331996, + "loss": 1.5072, + "theoretical_loss": 3.952196576822086, + "tokens_seen": 489619456 + }, + { + "epoch": 0.15, + "learning_rate": 0.0006881765295887664, + "loss": 1.679, + "theoretical_loss": 3.9521398084592585, + "tokens_seen": 489684992 + }, + { + "epoch": 0.15, + "learning_rate": 0.000688160481444333, + "loss": 1.6414, + "theoretical_loss": 3.952083049820372, + "tokens_seen": 489750528 + }, + { + "epoch": 0.15, + "learning_rate": 0.0006881444332998998, + "loss": 1.5022, + "theoretical_loss": 3.9520263009024585, + "tokens_seen": 489816064 + }, + { + "epoch": 0.15, + "learning_rate": 0.0006881283851554664, + "loss": 1.611, + "theoretical_loss": 3.951969561702554, + "tokens_seen": 489881600 + }, + { + "epoch": 0.15, + "learning_rate": 0.0006881123370110332, + "loss": 1.6387, + "theoretical_loss": 3.9519128322176935, + "tokens_seen": 489947136 + }, + { + "epoch": 0.15, + "learning_rate": 0.0006880962888665998, + "loss": 1.6859, + "theoretical_loss": 3.951856112444915, + "tokens_seen": 490012672 + }, + { + "epoch": 0.15, + "learning_rate": 0.0006880802407221666, + "loss": 1.568, + "theoretical_loss": 3.951799402381258, + "tokens_seen": 490078208 + }, + { + "epoch": 0.15, + "learning_rate": 0.0006880641925777332, + "loss": 1.4855, + "theoretical_loss": 3.951742702023761, + "tokens_seen": 490143744 + }, + { + "epoch": 0.15, + "learning_rate": 0.0006880481444333, + "loss": 1.489, + "theoretical_loss": 3.951686011369466, + "tokens_seen": 490209280 + }, + { + "epoch": 0.15, + "learning_rate": 0.0006880320962888666, + "loss": 1.6149, + "theoretical_loss": 3.9516293304154155, + "tokens_seen": 490274816 + }, + { + "epoch": 0.15, + "learning_rate": 0.0006880160481444334, + "loss": 1.5513, + "theoretical_loss": 3.9515726591586544, + "tokens_seen": 490340352 + }, + { + "epoch": 0.15, + "learning_rate": 0.000688, + "loss": 1.6129, + "theoretical_loss": 3.9515159975962257, + "tokens_seen": 490405888 + }, + { + "epoch": 0.15, + "learning_rate": 0.0006879839518555668, + "loss": 1.4376, + "theoretical_loss": 3.951459345725178, + "tokens_seen": 490471424 + }, + { + "epoch": 0.15, + "learning_rate": 0.0006879679037111334, + "loss": 1.5118, + "theoretical_loss": 3.9514027035425583, + "tokens_seen": 490536960 + }, + { + "epoch": 0.15, + "learning_rate": 0.0006879518555667002, + "loss": 1.6069, + "theoretical_loss": 3.951346071045415, + "tokens_seen": 490602496 + }, + { + "epoch": 0.15, + "learning_rate": 0.0006879358074222668, + "loss": 1.5679, + "theoretical_loss": 3.951289448230799, + "tokens_seen": 490668032 + }, + { + "epoch": 0.15, + "learning_rate": 0.0006879197592778336, + "loss": 1.5359, + "theoretical_loss": 3.9512328350957624, + "tokens_seen": 490733568 + }, + { + "epoch": 0.15, + "learning_rate": 0.0006879037111334002, + "loss": 1.594, + "theoretical_loss": 3.9511762316373575, + "tokens_seen": 490799104 + }, + { + "epoch": 0.15, + "learning_rate": 0.000687887662988967, + "loss": 1.6459, + "theoretical_loss": 3.951119637852639, + "tokens_seen": 490864640 + }, + { + "epoch": 0.15, + "learning_rate": 0.0006878716148445336, + "loss": 1.5164, + "theoretical_loss": 3.951063053738662, + "tokens_seen": 490930176 + }, + { + "epoch": 0.15, + "learning_rate": 0.0006878555667001004, + "loss": 1.5764, + "theoretical_loss": 3.951006479292484, + "tokens_seen": 490995712 + }, + { + "epoch": 0.15, + "learning_rate": 0.000687839518555667, + "loss": 1.602, + "theoretical_loss": 3.9509499145111624, + "tokens_seen": 491061248 + }, + { + "epoch": 0.15, + "learning_rate": 0.0006878234704112338, + "loss": 1.5709, + "theoretical_loss": 3.950893359391756, + "tokens_seen": 491126784 + }, + { + "epoch": 0.15, + "learning_rate": 0.0006878074222668004, + "loss": 1.4661, + "theoretical_loss": 3.9508368139313275, + "tokens_seen": 491192320 + }, + { + "epoch": 0.15, + "learning_rate": 0.0006877913741223672, + "loss": 1.5775, + "theoretical_loss": 3.950780278126937, + "tokens_seen": 491257856 + }, + { + "epoch": 0.15, + "learning_rate": 0.0006877753259779338, + "loss": 1.5945, + "theoretical_loss": 3.9507237519756484, + "tokens_seen": 491323392 + }, + { + "epoch": 0.15, + "learning_rate": 0.0006877592778335006, + "loss": 1.5727, + "theoretical_loss": 3.9506672354745263, + "tokens_seen": 491388928 + }, + { + "epoch": 0.15, + "learning_rate": 0.0006877432296890672, + "loss": 1.5585, + "theoretical_loss": 3.950610728620636, + "tokens_seen": 491454464 + }, + { + "epoch": 0.15, + "objective/train/docs_used": 185162, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.534618854522705, + "objective/train/theoretical_loss": 3.950554231411046, + "objective/train/tokens_used": 511980000, + "theoretical_loss": 3.950554231411046, + "tokens_seen": 491520000 + }, + { + "epoch": 0.15, + "learning_rate": 0.000687727181544634, + "loss": 1.566, + "theoretical_loss": 3.950554231411046, + "tokens_seen": 491520000 + }, + { + "epoch": 0.15, + "learning_rate": 0.0006877111334002006, + "loss": 1.5093, + "theoretical_loss": 3.9504977438428233, + "tokens_seen": 491585536 + }, + { + "epoch": 0.15, + "learning_rate": 0.0006876950852557674, + "loss": 1.4991, + "theoretical_loss": 3.950441265913038, + "tokens_seen": 491651072 + }, + { + "epoch": 0.15, + "learning_rate": 0.000687679037111334, + "loss": 1.458, + "theoretical_loss": 3.950384797618761, + "tokens_seen": 491716608 + }, + { + "epoch": 0.15, + "learning_rate": 0.0006876629889669008, + "loss": 1.6261, + "theoretical_loss": 3.950328338957064, + "tokens_seen": 491782144 + }, + { + "epoch": 0.15, + "learning_rate": 0.0006876469408224674, + "loss": 1.5267, + "theoretical_loss": 3.950271889925022, + "tokens_seen": 491847680 + }, + { + "epoch": 0.15, + "learning_rate": 0.0006876308926780342, + "loss": 1.6428, + "theoretical_loss": 3.9502154505197087, + "tokens_seen": 491913216 + }, + { + "epoch": 0.15, + "learning_rate": 0.0006876148445336008, + "loss": 1.6596, + "theoretical_loss": 3.9501590207382, + "tokens_seen": 491978752 + }, + { + "epoch": 0.15, + "learning_rate": 0.0006875987963891676, + "loss": 1.5915, + "theoretical_loss": 3.9501026005775732, + "tokens_seen": 492044288 + }, + { + "epoch": 0.15, + "learning_rate": 0.0006875827482447342, + "loss": 1.5647, + "theoretical_loss": 3.9500461900349078, + "tokens_seen": 492109824 + }, + { + "epoch": 0.15, + "learning_rate": 0.000687566700100301, + "loss": 1.5871, + "theoretical_loss": 3.949989789107283, + "tokens_seen": 492175360 + }, + { + "epoch": 0.15, + "learning_rate": 0.0006875506519558676, + "loss": 1.5402, + "theoretical_loss": 3.9499333977917805, + "tokens_seen": 492240896 + }, + { + "epoch": 0.15, + "learning_rate": 0.0006875346038114344, + "loss": 1.6542, + "theoretical_loss": 3.949877016085482, + "tokens_seen": 492306432 + }, + { + "epoch": 0.15, + "learning_rate": 0.000687518555667001, + "loss": 1.5642, + "theoretical_loss": 3.949820643985472, + "tokens_seen": 492371968 + }, + { + "epoch": 0.15, + "learning_rate": 0.0006875025075225678, + "loss": 1.6414, + "theoretical_loss": 3.9497642814888345, + "tokens_seen": 492437504 + }, + { + "epoch": 0.15, + "learning_rate": 0.0006874864593781344, + "loss": 1.5391, + "theoretical_loss": 3.949707928592657, + "tokens_seen": 492503040 + }, + { + "epoch": 0.15, + "learning_rate": 0.0006874704112337012, + "loss": 1.6386, + "theoretical_loss": 3.949651585294026, + "tokens_seen": 492568576 + }, + { + "epoch": 0.15, + "learning_rate": 0.0006874543630892678, + "loss": 1.6293, + "theoretical_loss": 3.9495952515900306, + "tokens_seen": 492634112 + }, + { + "epoch": 0.15, + "learning_rate": 0.0006874383149448346, + "loss": 1.6136, + "theoretical_loss": 3.9495389274777613, + "tokens_seen": 492699648 + }, + { + "epoch": 0.15, + "learning_rate": 0.0006874222668004012, + "loss": 1.4437, + "theoretical_loss": 3.9494826129543084, + "tokens_seen": 492765184 + }, + { + "epoch": 0.15, + "learning_rate": 0.0006874062186559679, + "loss": 1.6245, + "theoretical_loss": 3.9494263080167658, + "tokens_seen": 492830720 + }, + { + "epoch": 0.15, + "learning_rate": 0.0006873901705115347, + "loss": 1.6287, + "theoretical_loss": 3.949370012662227, + "tokens_seen": 492896256 + }, + { + "epoch": 0.15, + "learning_rate": 0.0006873741223671013, + "loss": 1.557, + "theoretical_loss": 3.9493137268877865, + "tokens_seen": 492961792 + }, + { + "epoch": 0.15, + "learning_rate": 0.000687358074222668, + "loss": 1.5787, + "theoretical_loss": 3.9492574506905416, + "tokens_seen": 493027328 + }, + { + "epoch": 0.15, + "learning_rate": 0.0006873420260782347, + "loss": 1.5126, + "theoretical_loss": 3.949201184067589, + "tokens_seen": 493092864 + }, + { + "epoch": 0.15, + "learning_rate": 0.0006873259779338015, + "loss": 1.6109, + "theoretical_loss": 3.949144927016029, + "tokens_seen": 493158400 + }, + { + "epoch": 0.15, + "learning_rate": 0.0006873099297893681, + "loss": 1.5961, + "theoretical_loss": 3.9490886795329603, + "tokens_seen": 493223936 + }, + { + "epoch": 0.15, + "learning_rate": 0.0006872938816449349, + "loss": 1.5471, + "theoretical_loss": 3.9490324416154854, + "tokens_seen": 493289472 + }, + { + "epoch": 0.15, + "learning_rate": 0.0006872778335005015, + "loss": 1.6098, + "theoretical_loss": 3.9489762132607065, + "tokens_seen": 493355008 + }, + { + "epoch": 0.15, + "learning_rate": 0.0006872617853560681, + "loss": 1.5885, + "theoretical_loss": 3.948919994465728, + "tokens_seen": 493420544 + }, + { + "epoch": 0.15, + "learning_rate": 0.0006872457372116349, + "loss": 1.7396, + "theoretical_loss": 3.9488637852276547, + "tokens_seen": 493486080 + }, + { + "epoch": 0.15, + "learning_rate": 0.0006872296890672015, + "loss": 1.6293, + "theoretical_loss": 3.9488075855435936, + "tokens_seen": 493551616 + }, + { + "epoch": 0.15, + "learning_rate": 0.0006872136409227683, + "loss": 1.7213, + "theoretical_loss": 3.9487513954106523, + "tokens_seen": 493617152 + }, + { + "epoch": 0.15, + "learning_rate": 0.000687197592778335, + "loss": 1.5788, + "theoretical_loss": 3.94869521482594, + "tokens_seen": 493682688 + }, + { + "epoch": 0.15, + "learning_rate": 0.0006871815446339017, + "loss": 1.5083, + "theoretical_loss": 3.9486390437865664, + "tokens_seen": 493748224 + }, + { + "epoch": 0.15, + "learning_rate": 0.0006871654964894684, + "loss": 1.5572, + "theoretical_loss": 3.948582882289644, + "tokens_seen": 493813760 + }, + { + "epoch": 0.15, + "learning_rate": 0.0006871494483450351, + "loss": 1.5667, + "theoretical_loss": 3.9485267303322846, + "tokens_seen": 493879296 + }, + { + "epoch": 0.15, + "learning_rate": 0.0006871334002006018, + "loss": 1.4772, + "theoretical_loss": 3.9484705879116033, + "tokens_seen": 493944832 + }, + { + "epoch": 0.15, + "learning_rate": 0.0006871173520561685, + "loss": 1.4599, + "theoretical_loss": 3.9484144550247144, + "tokens_seen": 494010368 + }, + { + "epoch": 0.15, + "learning_rate": 0.0006871013039117352, + "loss": 1.6707, + "theoretical_loss": 3.948358331668735, + "tokens_seen": 494075904 + }, + { + "epoch": 0.15, + "learning_rate": 0.0006870852557673019, + "loss": 1.5428, + "theoretical_loss": 3.9483022178407827, + "tokens_seen": 494141440 + }, + { + "epoch": 0.15, + "learning_rate": 0.0006870692076228687, + "loss": 1.7044, + "theoretical_loss": 3.9482461135379774, + "tokens_seen": 494206976 + }, + { + "epoch": 0.15, + "learning_rate": 0.0006870531594784353, + "loss": 1.5927, + "theoretical_loss": 3.9481900187574386, + "tokens_seen": 494272512 + }, + { + "epoch": 0.15, + "learning_rate": 0.0006870371113340021, + "loss": 1.5825, + "theoretical_loss": 3.948133933496288, + "tokens_seen": 494338048 + }, + { + "epoch": 0.15, + "learning_rate": 0.0006870210631895687, + "loss": 1.5429, + "theoretical_loss": 3.948077857751649, + "tokens_seen": 494403584 + }, + { + "epoch": 0.15, + "learning_rate": 0.0006870050150451355, + "loss": 1.635, + "theoretical_loss": 3.948021791520645, + "tokens_seen": 494469120 + }, + { + "epoch": 0.15, + "learning_rate": 0.0006869889669007021, + "loss": 1.6267, + "theoretical_loss": 3.9479657348004014, + "tokens_seen": 494534656 + }, + { + "epoch": 0.15, + "learning_rate": 0.0006869729187562689, + "loss": 1.6544, + "theoretical_loss": 3.9479096875880453, + "tokens_seen": 494600192 + }, + { + "epoch": 0.15, + "learning_rate": 0.0006869568706118355, + "loss": 1.7407, + "theoretical_loss": 3.947853649880704, + "tokens_seen": 494665728 + }, + { + "epoch": 0.15, + "learning_rate": 0.0006869408224674023, + "loss": 1.5793, + "theoretical_loss": 3.947797621675507, + "tokens_seen": 494731264 + }, + { + "epoch": 0.15, + "objective/train/docs_used": 186433, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.6235318183898926, + "objective/train/theoretical_loss": 3.947741602969585, + "objective/train/tokens_used": 515256800, + "theoretical_loss": 3.947741602969585, + "tokens_seen": 494796800 + }, + { + "epoch": 0.15, + "learning_rate": 0.0006869247743229689, + "loss": 1.5244, + "theoretical_loss": 3.947741602969585, + "tokens_seen": 494796800 + }, + { + "epoch": 0.15, + "learning_rate": 0.0006869087261785357, + "loss": 1.4682, + "theoretical_loss": 3.9476855937600686, + "tokens_seen": 494862336 + }, + { + "epoch": 0.15, + "learning_rate": 0.0006868926780341023, + "loss": 1.4476, + "theoretical_loss": 3.9476295940440913, + "tokens_seen": 494927872 + }, + { + "epoch": 0.15, + "learning_rate": 0.0006868766298896691, + "loss": 1.6045, + "theoretical_loss": 3.9475736038187867, + "tokens_seen": 494993408 + }, + { + "epoch": 0.15, + "learning_rate": 0.0006868605817452357, + "loss": 1.5215, + "theoretical_loss": 3.947517623081291, + "tokens_seen": 495058944 + }, + { + "epoch": 0.15, + "learning_rate": 0.0006868445336008025, + "loss": 1.5711, + "theoretical_loss": 3.9474616518287395, + "tokens_seen": 495124480 + }, + { + "epoch": 0.15, + "learning_rate": 0.0006868284854563691, + "loss": 1.5945, + "theoretical_loss": 3.9474056900582717, + "tokens_seen": 495190016 + }, + { + "epoch": 0.15, + "learning_rate": 0.0006868124373119359, + "loss": 1.6945, + "theoretical_loss": 3.9473497377670244, + "tokens_seen": 495255552 + }, + { + "epoch": 0.15, + "learning_rate": 0.0006867963891675025, + "loss": 1.7119, + "theoretical_loss": 3.94729379495214, + "tokens_seen": 495321088 + }, + { + "epoch": 0.15, + "learning_rate": 0.0006867803410230693, + "loss": 1.5318, + "theoretical_loss": 3.947237861610759, + "tokens_seen": 495386624 + }, + { + "epoch": 0.15, + "learning_rate": 0.0006867642928786359, + "loss": 1.5842, + "theoretical_loss": 3.9471819377400243, + "tokens_seen": 495452160 + }, + { + "epoch": 0.15, + "learning_rate": 0.0006867482447342027, + "loss": 1.4931, + "theoretical_loss": 3.94712602333708, + "tokens_seen": 495517696 + }, + { + "epoch": 0.15, + "learning_rate": 0.0006867321965897693, + "loss": 1.4614, + "theoretical_loss": 3.9470701183990715, + "tokens_seen": 495583232 + }, + { + "epoch": 0.15, + "learning_rate": 0.0006867161484453361, + "loss": 1.657, + "theoretical_loss": 3.947014222923145, + "tokens_seen": 495648768 + }, + { + "epoch": 0.15, + "learning_rate": 0.0006867001003009027, + "loss": 1.5399, + "theoretical_loss": 3.946958336906449, + "tokens_seen": 495714304 + }, + { + "epoch": 0.15, + "learning_rate": 0.0006866840521564695, + "loss": 1.4588, + "theoretical_loss": 3.9469024603461307, + "tokens_seen": 495779840 + }, + { + "epoch": 0.15, + "learning_rate": 0.0006866680040120361, + "loss": 1.7069, + "theoretical_loss": 3.946846593239342, + "tokens_seen": 495845376 + }, + { + "epoch": 0.15, + "learning_rate": 0.0006866519558676029, + "loss": 1.6388, + "theoretical_loss": 3.946790735583234, + "tokens_seen": 495910912 + }, + { + "epoch": 0.15, + "learning_rate": 0.0006866359077231695, + "loss": 1.5266, + "theoretical_loss": 3.9467348873749586, + "tokens_seen": 495976448 + }, + { + "epoch": 0.15, + "learning_rate": 0.0006866198595787363, + "loss": 1.6288, + "theoretical_loss": 3.9466790486116707, + "tokens_seen": 496041984 + }, + { + "epoch": 0.15, + "learning_rate": 0.0006866038114343029, + "loss": 1.6084, + "theoretical_loss": 3.946623219290525, + "tokens_seen": 496107520 + }, + { + "epoch": 0.15, + "learning_rate": 0.0006865877632898697, + "loss": 1.5596, + "theoretical_loss": 3.9465673994086776, + "tokens_seen": 496173056 + }, + { + "epoch": 0.15, + "learning_rate": 0.0006865717151454363, + "loss": 1.6082, + "theoretical_loss": 3.9465115889632862, + "tokens_seen": 496238592 + }, + { + "epoch": 0.15, + "learning_rate": 0.0006865556670010031, + "loss": 1.6171, + "theoretical_loss": 3.9464557879515105, + "tokens_seen": 496304128 + }, + { + "epoch": 0.15, + "learning_rate": 0.0006865396188565697, + "loss": 1.4912, + "theoretical_loss": 3.946399996370509, + "tokens_seen": 496369664 + }, + { + "epoch": 0.15, + "learning_rate": 0.0006865235707121365, + "loss": 1.5728, + "theoretical_loss": 3.9463442142174445, + "tokens_seen": 496435200 + }, + { + "epoch": 0.15, + "learning_rate": 0.0006865075225677031, + "loss": 1.6358, + "theoretical_loss": 3.9462884414894788, + "tokens_seen": 496500736 + }, + { + "epoch": 0.15, + "learning_rate": 0.0006864914744232699, + "loss": 1.5276, + "theoretical_loss": 3.946232678183775, + "tokens_seen": 496566272 + }, + { + "epoch": 0.15, + "learning_rate": 0.0006864754262788365, + "loss": 1.5262, + "theoretical_loss": 3.9461769242974998, + "tokens_seen": 496631808 + }, + { + "epoch": 0.15, + "learning_rate": 0.0006864593781344033, + "loss": 1.6283, + "theoretical_loss": 3.946121179827818, + "tokens_seen": 496697344 + }, + { + "epoch": 0.15, + "learning_rate": 0.0006864433299899699, + "loss": 1.6244, + "theoretical_loss": 3.9460654447718975, + "tokens_seen": 496762880 + }, + { + "epoch": 0.15, + "learning_rate": 0.0006864272818455367, + "loss": 1.6055, + "theoretical_loss": 3.9460097191269066, + "tokens_seen": 496828416 + }, + { + "epoch": 0.15, + "learning_rate": 0.0006864112337011033, + "loss": 1.5854, + "theoretical_loss": 3.9459540028900157, + "tokens_seen": 496893952 + }, + { + "epoch": 0.15, + "learning_rate": 0.0006863951855566701, + "loss": 1.5782, + "theoretical_loss": 3.945898296058396, + "tokens_seen": 496959488 + }, + { + "epoch": 0.15, + "learning_rate": 0.0006863791374122367, + "loss": 1.6092, + "theoretical_loss": 3.9458425986292194, + "tokens_seen": 497025024 + }, + { + "epoch": 0.15, + "learning_rate": 0.0006863630892678035, + "loss": 1.5501, + "theoretical_loss": 3.9457869105996597, + "tokens_seen": 497090560 + }, + { + "epoch": 0.15, + "learning_rate": 0.0006863470411233701, + "loss": 1.5318, + "theoretical_loss": 3.9457312319668914, + "tokens_seen": 497156096 + }, + { + "epoch": 0.15, + "learning_rate": 0.0006863309929789369, + "loss": 1.6296, + "theoretical_loss": 3.9456755627280904, + "tokens_seen": 497221632 + }, + { + "epoch": 0.15, + "learning_rate": 0.0006863149448345035, + "loss": 1.4482, + "theoretical_loss": 3.945619902880434, + "tokens_seen": 497287168 + }, + { + "epoch": 0.15, + "learning_rate": 0.0006862988966900703, + "loss": 1.6612, + "theoretical_loss": 3.945564252421102, + "tokens_seen": 497352704 + }, + { + "epoch": 0.15, + "learning_rate": 0.0006862828485456369, + "loss": 1.6787, + "theoretical_loss": 3.9455086113472717, + "tokens_seen": 497418240 + }, + { + "epoch": 0.15, + "learning_rate": 0.0006862668004012037, + "loss": 1.5548, + "theoretical_loss": 3.9454529796561255, + "tokens_seen": 497483776 + }, + { + "epoch": 0.15, + "learning_rate": 0.0006862507522567703, + "loss": 1.6029, + "theoretical_loss": 3.9453973573448455, + "tokens_seen": 497549312 + }, + { + "epoch": 0.15, + "learning_rate": 0.0006862347041123371, + "loss": 1.5408, + "theoretical_loss": 3.945341744410615, + "tokens_seen": 497614848 + }, + { + "epoch": 0.15, + "learning_rate": 0.0006862186559679037, + "loss": 1.5625, + "theoretical_loss": 3.9452861408506172, + "tokens_seen": 497680384 + }, + { + "epoch": 0.15, + "learning_rate": 0.0006862026078234705, + "loss": 1.5883, + "theoretical_loss": 3.945230546662039, + "tokens_seen": 497745920 + }, + { + "epoch": 0.15, + "learning_rate": 0.0006861865596790371, + "loss": 1.5373, + "theoretical_loss": 3.945174961842068, + "tokens_seen": 497811456 + }, + { + "epoch": 0.15, + "learning_rate": 0.0006861705115346039, + "loss": 1.6018, + "theoretical_loss": 3.9451193863878906, + "tokens_seen": 497876992 + }, + { + "epoch": 0.15, + "learning_rate": 0.0006861544633901705, + "loss": 1.5607, + "theoretical_loss": 3.945063820296698, + "tokens_seen": 497942528 + }, + { + "epoch": 0.15, + "learning_rate": 0.0006861384152457373, + "loss": 1.6158, + "theoretical_loss": 3.9450082635656796, + "tokens_seen": 498008064 + }, + { + "epoch": 0.15, + "objective/train/docs_used": 187487, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.682167887687683, + "objective/train/theoretical_loss": 3.9449527161920273, + "objective/train/tokens_used": 518533600, + "theoretical_loss": 3.9449527161920273, + "tokens_seen": 498073600 + }, + { + "epoch": 0.15, + "learning_rate": 0.000686122367101304, + "loss": 1.6417, + "theoretical_loss": 3.9449527161920273, + "tokens_seen": 498073600 + }, + { + "epoch": 0.15, + "learning_rate": 0.0006861063189568707, + "loss": 1.5373, + "theoretical_loss": 3.944897178172935, + "tokens_seen": 498139136 + }, + { + "epoch": 0.15, + "learning_rate": 0.0006860902708124373, + "loss": 1.6786, + "theoretical_loss": 3.9448416495055962, + "tokens_seen": 498204672 + }, + { + "epoch": 0.15, + "learning_rate": 0.0006860742226680041, + "loss": 1.5967, + "theoretical_loss": 3.9447861301872065, + "tokens_seen": 498270208 + }, + { + "epoch": 0.15, + "learning_rate": 0.0006860581745235707, + "loss": 1.4599, + "theoretical_loss": 3.944730620214962, + "tokens_seen": 498335744 + }, + { + "epoch": 0.15, + "learning_rate": 0.0006860421263791375, + "loss": 1.4873, + "theoretical_loss": 3.944675119586061, + "tokens_seen": 498401280 + }, + { + "epoch": 0.15, + "learning_rate": 0.0006860260782347042, + "loss": 1.5804, + "theoretical_loss": 3.944619628297704, + "tokens_seen": 498466816 + }, + { + "epoch": 0.15, + "learning_rate": 0.0006860100300902708, + "loss": 1.5662, + "theoretical_loss": 3.944564146347089, + "tokens_seen": 498532352 + }, + { + "epoch": 0.15, + "learning_rate": 0.0006859939819458376, + "loss": 1.4675, + "theoretical_loss": 3.9445086737314186, + "tokens_seen": 498597888 + }, + { + "epoch": 0.15, + "learning_rate": 0.0006859779338014042, + "loss": 1.6095, + "theoretical_loss": 3.9444532104478958, + "tokens_seen": 498663424 + }, + { + "epoch": 0.15, + "learning_rate": 0.000685961885656971, + "loss": 1.6014, + "theoretical_loss": 3.9443977564937236, + "tokens_seen": 498728960 + }, + { + "epoch": 0.15, + "learning_rate": 0.0006859458375125376, + "loss": 1.6011, + "theoretical_loss": 3.9443423118661074, + "tokens_seen": 498794496 + }, + { + "epoch": 0.15, + "learning_rate": 0.0006859297893681044, + "loss": 1.5587, + "theoretical_loss": 3.944286876562254, + "tokens_seen": 498860032 + }, + { + "epoch": 0.15, + "learning_rate": 0.000685913741223671, + "loss": 1.4845, + "theoretical_loss": 3.9442314505793705, + "tokens_seen": 498925568 + }, + { + "epoch": 0.15, + "learning_rate": 0.0006858976930792378, + "loss": 1.6731, + "theoretical_loss": 3.944176033914666, + "tokens_seen": 498991104 + }, + { + "epoch": 0.15, + "learning_rate": 0.0006858816449348044, + "loss": 1.6231, + "theoretical_loss": 3.9441206265653497, + "tokens_seen": 499056640 + }, + { + "epoch": 0.15, + "learning_rate": 0.0006858655967903712, + "loss": 1.5512, + "theoretical_loss": 3.9440652285286335, + "tokens_seen": 499122176 + }, + { + "epoch": 0.15, + "learning_rate": 0.0006858495486459378, + "loss": 1.5978, + "theoretical_loss": 3.9440098398017294, + "tokens_seen": 499187712 + }, + { + "epoch": 0.15, + "learning_rate": 0.0006858335005015046, + "loss": 1.5608, + "theoretical_loss": 3.9439544603818506, + "tokens_seen": 499253248 + }, + { + "epoch": 0.15, + "learning_rate": 0.0006858174523570712, + "loss": 1.6568, + "theoretical_loss": 3.9438990902662128, + "tokens_seen": 499318784 + }, + { + "epoch": 0.15, + "learning_rate": 0.000685801404212638, + "loss": 1.5826, + "theoretical_loss": 3.9438437294520305, + "tokens_seen": 499384320 + }, + { + "epoch": 0.15, + "learning_rate": 0.0006857853560682046, + "loss": 1.7045, + "theoretical_loss": 3.943788377936522, + "tokens_seen": 499449856 + }, + { + "epoch": 0.15, + "learning_rate": 0.0006857693079237714, + "loss": 1.5835, + "theoretical_loss": 3.943733035716905, + "tokens_seen": 499515392 + }, + { + "epoch": 0.15, + "learning_rate": 0.000685753259779338, + "loss": 1.5314, + "theoretical_loss": 3.9436777027903993, + "tokens_seen": 499580928 + }, + { + "epoch": 0.15, + "learning_rate": 0.0006857372116349048, + "loss": 1.7151, + "theoretical_loss": 3.9436223791542258, + "tokens_seen": 499646464 + }, + { + "epoch": 0.15, + "learning_rate": 0.0006857211634904714, + "loss": 1.5132, + "theoretical_loss": 3.943567064805606, + "tokens_seen": 499712000 + }, + { + "epoch": 0.15, + "learning_rate": 0.0006857051153460382, + "loss": 1.4704, + "theoretical_loss": 3.943511759741763, + "tokens_seen": 499777536 + }, + { + "epoch": 0.15, + "learning_rate": 0.0006856890672016048, + "loss": 1.485, + "theoretical_loss": 3.9434564639599214, + "tokens_seen": 499843072 + }, + { + "epoch": 0.15, + "learning_rate": 0.0006856730190571716, + "loss": 1.582, + "theoretical_loss": 3.9434011774573063, + "tokens_seen": 499908608 + }, + { + "epoch": 0.15, + "learning_rate": 0.0006856569709127382, + "loss": 1.5337, + "theoretical_loss": 3.9433459002311446, + "tokens_seen": 499974144 + }, + { + "epoch": 0.15, + "learning_rate": 0.000685640922768305, + "loss": 1.5268, + "theoretical_loss": 3.9432906322786643, + "tokens_seen": 500039680 + }, + { + "epoch": 0.15, + "learning_rate": 0.0006856248746238716, + "loss": 1.5824, + "theoretical_loss": 3.943235373597094, + "tokens_seen": 500105216 + }, + { + "epoch": 0.15, + "learning_rate": 0.0006856088264794384, + "loss": 1.6497, + "theoretical_loss": 3.943180124183665, + "tokens_seen": 500170752 + }, + { + "epoch": 0.15, + "learning_rate": 0.000685592778335005, + "loss": 1.5891, + "theoretical_loss": 3.9431248840356075, + "tokens_seen": 500236288 + }, + { + "epoch": 0.15, + "learning_rate": 0.0006855767301905718, + "loss": 1.6055, + "theoretical_loss": 3.943069653150155, + "tokens_seen": 500301824 + }, + { + "epoch": 0.15, + "learning_rate": 0.0006855606820461384, + "loss": 1.584, + "theoretical_loss": 3.9430144315245412, + "tokens_seen": 500367360 + }, + { + "epoch": 0.15, + "learning_rate": 0.0006855446339017052, + "loss": 1.5998, + "theoretical_loss": 3.942959219156, + "tokens_seen": 500432896 + }, + { + "epoch": 0.15, + "learning_rate": 0.0006855285857572718, + "loss": 1.7045, + "theoretical_loss": 3.942904016041769, + "tokens_seen": 500498432 + }, + { + "epoch": 0.15, + "learning_rate": 0.0006855125376128386, + "loss": 1.6169, + "theoretical_loss": 3.9428488221790854, + "tokens_seen": 500563968 + }, + { + "epoch": 0.15, + "learning_rate": 0.0006854964894684052, + "loss": 1.6496, + "theoretical_loss": 3.9427936375651877, + "tokens_seen": 500629504 + }, + { + "epoch": 0.15, + "learning_rate": 0.000685480441323972, + "loss": 1.5686, + "theoretical_loss": 3.9427384621973154, + "tokens_seen": 500695040 + }, + { + "epoch": 0.15, + "learning_rate": 0.0006854643931795386, + "loss": 1.526, + "theoretical_loss": 3.9426832960727087, + "tokens_seen": 500760576 + }, + { + "epoch": 0.15, + "learning_rate": 0.0006854483450351054, + "loss": 1.4728, + "theoretical_loss": 3.9426281391886113, + "tokens_seen": 500826112 + }, + { + "epoch": 0.15, + "learning_rate": 0.000685432296890672, + "loss": 1.5643, + "theoretical_loss": 3.9425729915422654, + "tokens_seen": 500891648 + }, + { + "epoch": 0.15, + "learning_rate": 0.0006854162487462388, + "loss": 1.5243, + "theoretical_loss": 3.942517853130916, + "tokens_seen": 500957184 + }, + { + "epoch": 0.15, + "learning_rate": 0.0006854002006018054, + "loss": 1.5283, + "theoretical_loss": 3.9424627239518086, + "tokens_seen": 501022720 + }, + { + "epoch": 0.15, + "learning_rate": 0.0006853841524573722, + "loss": 1.6602, + "theoretical_loss": 3.9424076040021903, + "tokens_seen": 501088256 + }, + { + "epoch": 0.15, + "learning_rate": 0.0006853681043129388, + "loss": 1.5705, + "theoretical_loss": 3.9423524932793086, + "tokens_seen": 501153792 + }, + { + "epoch": 0.15, + "learning_rate": 0.0006853520561685056, + "loss": 1.6113, + "theoretical_loss": 3.9422973917804134, + "tokens_seen": 501219328 + }, + { + "epoch": 0.15, + "learning_rate": 0.0006853360080240722, + "loss": 1.5656, + "theoretical_loss": 3.9422422995027544, + "tokens_seen": 501284864 + }, + { + "epoch": 0.15, + "objective/train/docs_used": 188727, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.4144394397735596, + "objective/train/theoretical_loss": 3.9421872164435836, + "objective/train/tokens_used": 521810400, + "theoretical_loss": 3.9421872164435836, + "tokens_seen": 501350400 + }, + { + "epoch": 0.15, + "learning_rate": 0.000685319959879639, + "loss": 1.5467, + "theoretical_loss": 3.9421872164435836, + "tokens_seen": 501350400 + }, + { + "epoch": 0.15, + "learning_rate": 0.0006853039117352056, + "loss": 1.5047, + "theoretical_loss": 3.942132142600154, + "tokens_seen": 501415936 + }, + { + "epoch": 0.15, + "learning_rate": 0.0006852878635907724, + "loss": 1.5506, + "theoretical_loss": 3.942077077969719, + "tokens_seen": 501481472 + }, + { + "epoch": 0.15, + "learning_rate": 0.000685271815446339, + "loss": 1.6686, + "theoretical_loss": 3.9420220225495335, + "tokens_seen": 501547008 + }, + { + "epoch": 0.15, + "learning_rate": 0.0006852557673019058, + "loss": 1.4919, + "theoretical_loss": 3.9419669763368548, + "tokens_seen": 501612544 + }, + { + "epoch": 0.15, + "learning_rate": 0.0006852397191574724, + "loss": 1.5975, + "theoretical_loss": 3.9419119393289397, + "tokens_seen": 501678080 + }, + { + "epoch": 0.15, + "learning_rate": 0.0006852236710130392, + "loss": 1.4967, + "theoretical_loss": 3.941856911523047, + "tokens_seen": 501743616 + }, + { + "epoch": 0.15, + "learning_rate": 0.0006852076228686058, + "loss": 1.5238, + "theoretical_loss": 3.9418018929164362, + "tokens_seen": 501809152 + }, + { + "epoch": 0.15, + "learning_rate": 0.0006851915747241726, + "loss": 1.5856, + "theoretical_loss": 3.9417468835063687, + "tokens_seen": 501874688 + }, + { + "epoch": 0.15, + "learning_rate": 0.0006851755265797392, + "loss": 1.6091, + "theoretical_loss": 3.9416918832901064, + "tokens_seen": 501940224 + }, + { + "epoch": 0.15, + "learning_rate": 0.000685159478435306, + "loss": 1.5816, + "theoretical_loss": 3.9416368922649134, + "tokens_seen": 502005760 + }, + { + "epoch": 0.15, + "learning_rate": 0.0006851434302908726, + "loss": 1.5319, + "theoretical_loss": 3.9415819104280527, + "tokens_seen": 502071296 + }, + { + "epoch": 0.15, + "learning_rate": 0.0006851273821464394, + "loss": 1.5851, + "theoretical_loss": 3.941526937776791, + "tokens_seen": 502136832 + }, + { + "epoch": 0.15, + "learning_rate": 0.000685111334002006, + "loss": 1.446, + "theoretical_loss": 3.941471974308395, + "tokens_seen": 502202368 + }, + { + "epoch": 0.15, + "learning_rate": 0.0006850952858575728, + "loss": 1.6732, + "theoretical_loss": 3.9414170200201326, + "tokens_seen": 502267904 + }, + { + "epoch": 0.15, + "learning_rate": 0.0006850792377131394, + "loss": 1.5853, + "theoretical_loss": 3.9413620749092733, + "tokens_seen": 502333440 + }, + { + "epoch": 0.15, + "learning_rate": 0.0006850631895687062, + "loss": 1.4718, + "theoretical_loss": 3.9413071389730874, + "tokens_seen": 502398976 + }, + { + "epoch": 0.15, + "learning_rate": 0.0006850471414242728, + "loss": 1.4982, + "theoretical_loss": 3.941252212208846, + "tokens_seen": 502464512 + }, + { + "epoch": 0.15, + "learning_rate": 0.0006850310932798396, + "loss": 1.6002, + "theoretical_loss": 3.941197294613822, + "tokens_seen": 502530048 + }, + { + "epoch": 0.15, + "learning_rate": 0.0006850150451354062, + "loss": 1.621, + "theoretical_loss": 3.9411423861852892, + "tokens_seen": 502595584 + }, + { + "epoch": 0.15, + "learning_rate": 0.000684998996990973, + "loss": 1.6324, + "theoretical_loss": 3.941087486920523, + "tokens_seen": 502661120 + }, + { + "epoch": 0.15, + "learning_rate": 0.0006849829488465396, + "loss": 1.6196, + "theoretical_loss": 3.9410325968167994, + "tokens_seen": 502726656 + }, + { + "epoch": 0.15, + "learning_rate": 0.0006849669007021064, + "loss": 1.5417, + "theoretical_loss": 3.9409777158713952, + "tokens_seen": 502792192 + }, + { + "epoch": 0.15, + "learning_rate": 0.000684950852557673, + "loss": 1.624, + "theoretical_loss": 3.9409228440815895, + "tokens_seen": 502857728 + }, + { + "epoch": 0.15, + "learning_rate": 0.0006849348044132398, + "loss": 1.6731, + "theoretical_loss": 3.9408679814446623, + "tokens_seen": 502923264 + }, + { + "epoch": 0.15, + "learning_rate": 0.0006849187562688064, + "loss": 1.5957, + "theoretical_loss": 3.9408131279578935, + "tokens_seen": 502988800 + }, + { + "epoch": 0.15, + "learning_rate": 0.0006849027081243732, + "loss": 1.5945, + "theoretical_loss": 3.9407582836185657, + "tokens_seen": 503054336 + }, + { + "epoch": 0.15, + "learning_rate": 0.0006848866599799398, + "loss": 1.676, + "theoretical_loss": 3.940703448423962, + "tokens_seen": 503119872 + }, + { + "epoch": 0.15, + "learning_rate": 0.0006848706118355066, + "loss": 1.6987, + "theoretical_loss": 3.9406486223713664, + "tokens_seen": 503185408 + }, + { + "epoch": 0.15, + "learning_rate": 0.0006848545636910732, + "loss": 1.6542, + "theoretical_loss": 3.940593805458065, + "tokens_seen": 503250944 + }, + { + "epoch": 0.15, + "learning_rate": 0.00068483851554664, + "loss": 1.626, + "theoretical_loss": 3.940538997681344, + "tokens_seen": 503316480 + }, + { + "epoch": 0.15, + "learning_rate": 0.0006848224674022066, + "loss": 1.6217, + "theoretical_loss": 3.9404841990384916, + "tokens_seen": 503382016 + }, + { + "epoch": 0.15, + "learning_rate": 0.0006848064192577734, + "loss": 1.6237, + "theoretical_loss": 3.940429409526796, + "tokens_seen": 503447552 + }, + { + "epoch": 0.15, + "learning_rate": 0.00068479037111334, + "loss": 1.7262, + "theoretical_loss": 3.940374629143548, + "tokens_seen": 503513088 + }, + { + "epoch": 0.15, + "learning_rate": 0.0006847743229689068, + "loss": 1.6648, + "theoretical_loss": 3.940319857886039, + "tokens_seen": 503578624 + }, + { + "epoch": 0.15, + "learning_rate": 0.0006847582748244734, + "loss": 1.6583, + "theoretical_loss": 3.9402650957515606, + "tokens_seen": 503644160 + }, + { + "epoch": 0.15, + "learning_rate": 0.0006847422266800402, + "loss": 1.6812, + "theoretical_loss": 3.940210342737407, + "tokens_seen": 503709696 + }, + { + "epoch": 0.15, + "learning_rate": 0.0006847261785356068, + "loss": 1.6456, + "theoretical_loss": 3.940155598840873, + "tokens_seen": 503775232 + }, + { + "epoch": 0.15, + "learning_rate": 0.0006847101303911736, + "loss": 1.6594, + "theoretical_loss": 3.9401008640592545, + "tokens_seen": 503840768 + }, + { + "epoch": 0.15, + "learning_rate": 0.0006846940822467402, + "loss": 1.7093, + "theoretical_loss": 3.940046138389848, + "tokens_seen": 503906304 + }, + { + "epoch": 0.15, + "learning_rate": 0.000684678034102307, + "loss": 1.7981, + "theoretical_loss": 3.9399914218299523, + "tokens_seen": 503971840 + }, + { + "epoch": 0.15, + "learning_rate": 0.0006846619859578737, + "loss": 1.9029, + "theoretical_loss": 3.9399367143768664, + "tokens_seen": 504037376 + }, + { + "epoch": 0.15, + "learning_rate": 0.0006846459378134404, + "loss": 1.7128, + "theoretical_loss": 3.939882016027891, + "tokens_seen": 504102912 + }, + { + "epoch": 0.15, + "learning_rate": 0.000684629889669007, + "loss": 1.7157, + "theoretical_loss": 3.939827326780328, + "tokens_seen": 504168448 + }, + { + "epoch": 0.15, + "learning_rate": 0.0006846138415245737, + "loss": 1.7476, + "theoretical_loss": 3.9397726466314795, + "tokens_seen": 504233984 + }, + { + "epoch": 0.15, + "learning_rate": 0.0006845977933801405, + "loss": 1.6037, + "theoretical_loss": 3.93971797557865, + "tokens_seen": 504299520 + }, + { + "epoch": 0.15, + "learning_rate": 0.0006845817452357071, + "loss": 1.7578, + "theoretical_loss": 3.939663313619145, + "tokens_seen": 504365056 + }, + { + "epoch": 0.15, + "learning_rate": 0.0006845656970912739, + "loss": 1.6403, + "theoretical_loss": 3.9396086607502694, + "tokens_seen": 504430592 + }, + { + "epoch": 0.15, + "learning_rate": 0.0006845496489468405, + "loss": 1.6514, + "theoretical_loss": 3.939554016969332, + "tokens_seen": 504496128 + }, + { + "epoch": 0.15, + "learning_rate": 0.0006845336008024073, + "loss": 1.6478, + "theoretical_loss": 3.939499382273641, + "tokens_seen": 504561664 + }, + { + "epoch": 0.15, + "objective/train/docs_used": 189911, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.5120371580123901, + "objective/train/theoretical_loss": 3.939444756660505, + "objective/train/tokens_used": 525087200, + "theoretical_loss": 3.939444756660505, + "tokens_seen": 504627200 + }, + { + "epoch": 0.15, + "learning_rate": 0.0006845175526579739, + "loss": 1.6389, + "theoretical_loss": 3.939444756660505, + "tokens_seen": 504627200 + }, + { + "epoch": 0.15, + "learning_rate": 0.0006845015045135407, + "loss": 1.636, + "theoretical_loss": 3.9393901401272364, + "tokens_seen": 504692736 + }, + { + "epoch": 0.15, + "learning_rate": 0.0006844854563691073, + "loss": 1.6353, + "theoretical_loss": 3.939335532671146, + "tokens_seen": 504758272 + }, + { + "epoch": 0.15, + "learning_rate": 0.0006844694082246741, + "loss": 1.6735, + "theoretical_loss": 3.939280934289548, + "tokens_seen": 504823808 + }, + { + "epoch": 0.15, + "learning_rate": 0.0006844533600802407, + "loss": 1.6621, + "theoretical_loss": 3.9392263449797564, + "tokens_seen": 504889344 + }, + { + "epoch": 0.15, + "learning_rate": 0.0006844373119358075, + "loss": 1.6857, + "theoretical_loss": 3.939171764739086, + "tokens_seen": 504954880 + }, + { + "epoch": 0.15, + "learning_rate": 0.0006844212637913741, + "loss": 1.6204, + "theoretical_loss": 3.9391171935648535, + "tokens_seen": 505020416 + }, + { + "epoch": 0.15, + "learning_rate": 0.0006844052156469409, + "loss": 1.7122, + "theoretical_loss": 3.9390626314543766, + "tokens_seen": 505085952 + }, + { + "epoch": 0.15, + "learning_rate": 0.0006843891675025075, + "loss": 1.8205, + "theoretical_loss": 3.939008078404975, + "tokens_seen": 505151488 + }, + { + "epoch": 0.15, + "learning_rate": 0.0006843731193580743, + "loss": 1.6955, + "theoretical_loss": 3.938953534413968, + "tokens_seen": 505217024 + }, + { + "epoch": 0.15, + "learning_rate": 0.0006843570712136409, + "loss": 1.5974, + "theoretical_loss": 3.9388989994786767, + "tokens_seen": 505282560 + }, + { + "epoch": 0.15, + "learning_rate": 0.0006843410230692077, + "loss": 1.6949, + "theoretical_loss": 3.938844473596423, + "tokens_seen": 505348096 + }, + { + "epoch": 0.15, + "learning_rate": 0.0006843249749247743, + "loss": 1.6685, + "theoretical_loss": 3.9387899567645315, + "tokens_seen": 505413632 + }, + { + "epoch": 0.15, + "learning_rate": 0.0006843089267803411, + "loss": 1.6294, + "theoretical_loss": 3.9387354489803252, + "tokens_seen": 505479168 + }, + { + "epoch": 0.15, + "learning_rate": 0.0006842928786359077, + "loss": 1.6699, + "theoretical_loss": 3.938680950241131, + "tokens_seen": 505544704 + }, + { + "epoch": 0.15, + "learning_rate": 0.0006842768304914745, + "loss": 1.6819, + "theoretical_loss": 3.938626460544275, + "tokens_seen": 505610240 + }, + { + "epoch": 0.15, + "learning_rate": 0.0006842607823470411, + "loss": 1.7014, + "theoretical_loss": 3.9385719798870857, + "tokens_seen": 505675776 + }, + { + "epoch": 0.15, + "learning_rate": 0.0006842447342026079, + "loss": 1.7741, + "theoretical_loss": 3.9385175082668926, + "tokens_seen": 505741312 + }, + { + "epoch": 0.15, + "learning_rate": 0.0006842286860581745, + "loss": 1.6518, + "theoretical_loss": 3.9384630456810243, + "tokens_seen": 505806848 + }, + { + "epoch": 0.15, + "learning_rate": 0.0006842126379137413, + "loss": 1.6657, + "theoretical_loss": 3.9384085921268133, + "tokens_seen": 505872384 + }, + { + "epoch": 0.15, + "learning_rate": 0.0006841965897693079, + "loss": 1.6194, + "theoretical_loss": 3.9383541476015926, + "tokens_seen": 505937920 + }, + { + "epoch": 0.15, + "learning_rate": 0.0006841805416248747, + "loss": 1.6528, + "theoretical_loss": 3.9382997121026944, + "tokens_seen": 506003456 + }, + { + "epoch": 0.15, + "learning_rate": 0.0006841644934804413, + "loss": 1.7138, + "theoretical_loss": 3.938245285627455, + "tokens_seen": 506068992 + }, + { + "epoch": 0.15, + "learning_rate": 0.0006841484453360081, + "loss": 1.5723, + "theoretical_loss": 3.9381908681732085, + "tokens_seen": 506134528 + }, + { + "epoch": 0.15, + "learning_rate": 0.0006841323971915747, + "loss": 1.6994, + "theoretical_loss": 3.9381364597372936, + "tokens_seen": 506200064 + }, + { + "epoch": 0.15, + "learning_rate": 0.0006841163490471415, + "loss": 1.7913, + "theoretical_loss": 3.938082060317048, + "tokens_seen": 506265600 + }, + { + "epoch": 0.15, + "learning_rate": 0.0006841003009027081, + "loss": 1.7098, + "theoretical_loss": 3.9380276699098102, + "tokens_seen": 506331136 + }, + { + "epoch": 0.15, + "learning_rate": 0.0006840842527582749, + "loss": 1.5471, + "theoretical_loss": 3.9379732885129215, + "tokens_seen": 506396672 + }, + { + "epoch": 0.15, + "learning_rate": 0.0006840682046138415, + "loss": 1.6097, + "theoretical_loss": 3.9379189161237234, + "tokens_seen": 506462208 + }, + { + "epoch": 0.15, + "learning_rate": 0.0006840521564694083, + "loss": 1.7594, + "theoretical_loss": 3.9378645527395575, + "tokens_seen": 506527744 + }, + { + "epoch": 0.15, + "learning_rate": 0.0006840361083249749, + "loss": 1.5394, + "theoretical_loss": 3.937810198357769, + "tokens_seen": 506593280 + }, + { + "epoch": 0.15, + "learning_rate": 0.0006840200601805417, + "loss": 1.6265, + "theoretical_loss": 3.9377558529757026, + "tokens_seen": 506658816 + }, + { + "epoch": 0.15, + "learning_rate": 0.0006840040120361083, + "loss": 1.6958, + "theoretical_loss": 3.937701516590703, + "tokens_seen": 506724352 + }, + { + "epoch": 0.15, + "learning_rate": 0.0006839879638916751, + "loss": 1.565, + "theoretical_loss": 3.9376471892001197, + "tokens_seen": 506789888 + }, + { + "epoch": 0.15, + "learning_rate": 0.0006839719157472417, + "loss": 1.7363, + "theoretical_loss": 3.9375928708012986, + "tokens_seen": 506855424 + }, + { + "epoch": 0.15, + "learning_rate": 0.0006839558676028085, + "loss": 1.6742, + "theoretical_loss": 3.9375385613915905, + "tokens_seen": 506920960 + }, + { + "epoch": 0.15, + "learning_rate": 0.0006839398194583751, + "loss": 1.6342, + "theoretical_loss": 3.937484260968346, + "tokens_seen": 506986496 + }, + { + "epoch": 0.15, + "learning_rate": 0.0006839237713139419, + "loss": 1.7214, + "theoretical_loss": 3.9374299695289157, + "tokens_seen": 507052032 + }, + { + "epoch": 0.15, + "learning_rate": 0.0006839077231695085, + "loss": 1.6441, + "theoretical_loss": 3.9373756870706536, + "tokens_seen": 507117568 + }, + { + "epoch": 0.15, + "learning_rate": 0.0006838916750250753, + "loss": 1.6718, + "theoretical_loss": 3.937321413590913, + "tokens_seen": 507183104 + }, + { + "epoch": 0.15, + "learning_rate": 0.0006838756268806419, + "loss": 1.7649, + "theoretical_loss": 3.9372671490870492, + "tokens_seen": 507248640 + }, + { + "epoch": 0.15, + "learning_rate": 0.0006838595787362087, + "loss": 1.5116, + "theoretical_loss": 3.937212893556418, + "tokens_seen": 507314176 + }, + { + "epoch": 0.15, + "learning_rate": 0.0006838435305917753, + "loss": 1.604, + "theoretical_loss": 3.9371586469963775, + "tokens_seen": 507379712 + }, + { + "epoch": 0.15, + "learning_rate": 0.0006838274824473421, + "loss": 1.6395, + "theoretical_loss": 3.937104409404285, + "tokens_seen": 507445248 + }, + { + "epoch": 0.15, + "learning_rate": 0.0006838114343029087, + "loss": 1.5427, + "theoretical_loss": 3.9370501807775007, + "tokens_seen": 507510784 + }, + { + "epoch": 0.15, + "learning_rate": 0.0006837953861584755, + "loss": 1.5735, + "theoretical_loss": 3.936995961113385, + "tokens_seen": 507576320 + }, + { + "epoch": 0.15, + "learning_rate": 0.0006837793380140421, + "loss": 1.6086, + "theoretical_loss": 3.9369417504093, + "tokens_seen": 507641856 + }, + { + "epoch": 0.15, + "learning_rate": 0.0006837632898696089, + "loss": 1.6198, + "theoretical_loss": 3.936887548662608, + "tokens_seen": 507707392 + }, + { + "epoch": 0.15, + "learning_rate": 0.0006837472417251755, + "loss": 1.5303, + "theoretical_loss": 3.9368333558706734, + "tokens_seen": 507772928 + }, + { + "epoch": 0.15, + "learning_rate": 0.0006837311935807423, + "loss": 1.6947, + "theoretical_loss": 3.936779172030861, + "tokens_seen": 507838464 + }, + { + "epoch": 0.15, + "objective/train/docs_used": 191028, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.5330253839492798, + "objective/train/theoretical_loss": 3.936724997140538, + "objective/train/tokens_used": 528364000, + "theoretical_loss": 3.936724997140538, + "tokens_seen": 507904000 + }, + { + "epoch": 0.15, + "learning_rate": 0.0006837151454363089, + "loss": 1.53, + "theoretical_loss": 3.936724997140538, + "tokens_seen": 507904000 + }, + { + "epoch": 0.15, + "learning_rate": 0.0006836990972918757, + "loss": 1.6265, + "theoretical_loss": 3.9366708311970706, + "tokens_seen": 507969536 + }, + { + "epoch": 0.15, + "learning_rate": 0.0006836830491474423, + "loss": 1.5108, + "theoretical_loss": 3.9366166741978272, + "tokens_seen": 508035072 + }, + { + "epoch": 0.15, + "learning_rate": 0.0006836670010030091, + "loss": 1.561, + "theoretical_loss": 3.936562526140178, + "tokens_seen": 508100608 + }, + { + "epoch": 0.15, + "learning_rate": 0.0006836509528585757, + "loss": 1.5132, + "theoretical_loss": 3.936508387021494, + "tokens_seen": 508166144 + }, + { + "epoch": 0.15, + "learning_rate": 0.0006836349047141425, + "loss": 1.6205, + "theoretical_loss": 3.9364542568391454, + "tokens_seen": 508231680 + }, + { + "epoch": 0.15, + "learning_rate": 0.0006836188565697091, + "loss": 1.5388, + "theoretical_loss": 3.9364001355905067, + "tokens_seen": 508297216 + }, + { + "epoch": 0.15, + "learning_rate": 0.0006836028084252759, + "loss": 1.6027, + "theoretical_loss": 3.936346023272952, + "tokens_seen": 508362752 + }, + { + "epoch": 0.15, + "learning_rate": 0.0006835867602808425, + "loss": 1.5733, + "theoretical_loss": 3.936291919883855, + "tokens_seen": 508428288 + }, + { + "epoch": 0.15, + "learning_rate": 0.0006835707121364093, + "loss": 1.6851, + "theoretical_loss": 3.9362378254205925, + "tokens_seen": 508493824 + }, + { + "epoch": 0.15, + "learning_rate": 0.0006835546639919759, + "loss": 1.7339, + "theoretical_loss": 3.936183739880543, + "tokens_seen": 508559360 + }, + { + "epoch": 0.15, + "learning_rate": 0.0006835386158475427, + "loss": 1.7074, + "theoretical_loss": 3.9361296632610827, + "tokens_seen": 508624896 + }, + { + "epoch": 0.15, + "learning_rate": 0.0006835225677031093, + "loss": 1.6669, + "theoretical_loss": 3.9360755955595934, + "tokens_seen": 508690432 + }, + { + "epoch": 0.15, + "learning_rate": 0.0006835065195586761, + "loss": 1.7732, + "theoretical_loss": 3.936021536773455, + "tokens_seen": 508755968 + }, + { + "epoch": 0.15, + "learning_rate": 0.0006834904714142427, + "loss": 1.6241, + "theoretical_loss": 3.9359674869000485, + "tokens_seen": 508821504 + }, + { + "epoch": 0.15, + "learning_rate": 0.0006834744232698095, + "loss": 1.618, + "theoretical_loss": 3.9359134459367575, + "tokens_seen": 508887040 + }, + { + "epoch": 0.15, + "learning_rate": 0.0006834583751253763, + "loss": 1.6754, + "theoretical_loss": 3.9358594138809657, + "tokens_seen": 508952576 + }, + { + "epoch": 0.15, + "learning_rate": 0.0006834423269809429, + "loss": 1.6658, + "theoretical_loss": 3.9358053907300588, + "tokens_seen": 509018112 + }, + { + "epoch": 0.15, + "learning_rate": 0.0006834262788365097, + "loss": 1.6223, + "theoretical_loss": 3.935751376481422, + "tokens_seen": 509083648 + }, + { + "epoch": 0.15, + "learning_rate": 0.0006834102306920763, + "loss": 1.5496, + "theoretical_loss": 3.9356973711324432, + "tokens_seen": 509149184 + }, + { + "epoch": 0.15, + "learning_rate": 0.0006833941825476431, + "loss": 1.6484, + "theoretical_loss": 3.9356433746805113, + "tokens_seen": 509214720 + }, + { + "epoch": 0.15, + "learning_rate": 0.0006833781344032097, + "loss": 1.6683, + "theoretical_loss": 3.9355893871230143, + "tokens_seen": 509280256 + }, + { + "epoch": 0.15, + "learning_rate": 0.0006833620862587765, + "loss": 1.642, + "theoretical_loss": 3.935535408457344, + "tokens_seen": 509345792 + }, + { + "epoch": 0.15, + "learning_rate": 0.0006833460381143431, + "loss": 1.5493, + "theoretical_loss": 3.9354814386808923, + "tokens_seen": 509411328 + }, + { + "epoch": 0.15, + "learning_rate": 0.0006833299899699097, + "loss": 1.6046, + "theoretical_loss": 3.935427477791051, + "tokens_seen": 509476864 + }, + { + "epoch": 0.15, + "learning_rate": 0.0006833139418254765, + "loss": 1.6169, + "theoretical_loss": 3.9353735257852147, + "tokens_seen": 509542400 + }, + { + "epoch": 0.15, + "learning_rate": 0.0006832978936810431, + "loss": 1.5592, + "theoretical_loss": 3.935319582660778, + "tokens_seen": 509607936 + }, + { + "epoch": 0.15, + "learning_rate": 0.0006832818455366099, + "loss": 1.6628, + "theoretical_loss": 3.935265648415138, + "tokens_seen": 509673472 + }, + { + "epoch": 0.15, + "learning_rate": 0.0006832657973921766, + "loss": 1.4854, + "theoretical_loss": 3.9352117230456902, + "tokens_seen": 509739008 + }, + { + "epoch": 0.15, + "learning_rate": 0.0006832497492477433, + "loss": 1.6948, + "theoretical_loss": 3.9351578065498343, + "tokens_seen": 509804544 + }, + { + "epoch": 0.15, + "learning_rate": 0.00068323370110331, + "loss": 1.7144, + "theoretical_loss": 3.935103898924969, + "tokens_seen": 509870080 + }, + { + "epoch": 0.15, + "learning_rate": 0.0006832176529588766, + "loss": 1.7164, + "theoretical_loss": 3.935050000168495, + "tokens_seen": 509935616 + }, + { + "epoch": 0.15, + "learning_rate": 0.0006832016048144434, + "loss": 1.5019, + "theoretical_loss": 3.9349961102778135, + "tokens_seen": 510001152 + }, + { + "epoch": 0.15, + "learning_rate": 0.00068318555667001, + "loss": 1.6949, + "theoretical_loss": 3.934942229250328, + "tokens_seen": 510066688 + }, + { + "epoch": 0.15, + "learning_rate": 0.0006831695085255768, + "loss": 1.5639, + "theoretical_loss": 3.934888357083442, + "tokens_seen": 510132224 + }, + { + "epoch": 0.15, + "learning_rate": 0.0006831534603811434, + "loss": 1.5852, + "theoretical_loss": 3.9348344937745594, + "tokens_seen": 510197760 + }, + { + "epoch": 0.15, + "learning_rate": 0.0006831374122367102, + "loss": 1.5632, + "theoretical_loss": 3.9347806393210876, + "tokens_seen": 510263296 + }, + { + "epoch": 0.15, + "learning_rate": 0.0006831213640922768, + "loss": 1.7333, + "theoretical_loss": 3.934726793720433, + "tokens_seen": 510328832 + }, + { + "epoch": 0.15, + "learning_rate": 0.0006831053159478436, + "loss": 1.572, + "theoretical_loss": 3.9346729569700036, + "tokens_seen": 510394368 + }, + { + "epoch": 0.15, + "learning_rate": 0.0006830892678034102, + "loss": 1.5754, + "theoretical_loss": 3.934619129067209, + "tokens_seen": 510459904 + }, + { + "epoch": 0.15, + "learning_rate": 0.000683073219658977, + "loss": 1.5913, + "theoretical_loss": 3.934565310009459, + "tokens_seen": 510525440 + }, + { + "epoch": 0.15, + "learning_rate": 0.0006830571715145436, + "loss": 1.707, + "theoretical_loss": 3.9345114997941657, + "tokens_seen": 510590976 + }, + { + "epoch": 0.15, + "learning_rate": 0.0006830411233701104, + "loss": 1.6294, + "theoretical_loss": 3.934457698418741, + "tokens_seen": 510656512 + }, + { + "epoch": 0.15, + "learning_rate": 0.000683025075225677, + "loss": 1.6298, + "theoretical_loss": 3.9344039058805986, + "tokens_seen": 510722048 + }, + { + "epoch": 0.15, + "learning_rate": 0.0006830090270812438, + "loss": 1.5979, + "theoretical_loss": 3.934350122177154, + "tokens_seen": 510787584 + }, + { + "epoch": 0.15, + "learning_rate": 0.0006829929789368104, + "loss": 1.5589, + "theoretical_loss": 3.9342963473058212, + "tokens_seen": 510853120 + }, + { + "epoch": 0.15, + "learning_rate": 0.0006829769307923772, + "loss": 1.6284, + "theoretical_loss": 3.9342425812640194, + "tokens_seen": 510918656 + }, + { + "epoch": 0.15, + "learning_rate": 0.0006829608826479438, + "loss": 1.4556, + "theoretical_loss": 3.9341888240491647, + "tokens_seen": 510984192 + }, + { + "epoch": 0.15, + "learning_rate": 0.0006829448345035106, + "loss": 1.7204, + "theoretical_loss": 3.9341350756586766, + "tokens_seen": 511049728 + }, + { + "epoch": 0.15, + "learning_rate": 0.0006829287863590772, + "loss": 1.6471, + "theoretical_loss": 3.934081336089976, + "tokens_seen": 511115264 + }, + { + "epoch": 0.15, + "objective/train/docs_used": 192198, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.705810546875, + "objective/train/theoretical_loss": 3.934027605340483, + "objective/train/tokens_used": 531640800, + "theoretical_loss": 3.934027605340483, + "tokens_seen": 511180800 + }, + { + "epoch": 0.15, + "learning_rate": 0.000682912738214644, + "loss": 1.7278, + "theoretical_loss": 3.934027605340483, + "tokens_seen": 511180800 + }, + { + "epoch": 0.15, + "learning_rate": 0.0006828966900702106, + "loss": 1.5156, + "theoretical_loss": 3.93397388340762, + "tokens_seen": 511246336 + }, + { + "epoch": 0.15, + "learning_rate": 0.0006828806419257774, + "loss": 1.6288, + "theoretical_loss": 3.933920170288811, + "tokens_seen": 511311872 + }, + { + "epoch": 0.15, + "learning_rate": 0.000682864593781344, + "loss": 1.5711, + "theoretical_loss": 3.93386646598148, + "tokens_seen": 511377408 + }, + { + "epoch": 0.15, + "learning_rate": 0.0006828485456369108, + "loss": 1.6433, + "theoretical_loss": 3.9338127704830534, + "tokens_seen": 511442944 + }, + { + "epoch": 0.16, + "learning_rate": 0.0006828324974924774, + "loss": 1.612, + "theoretical_loss": 3.933759083790957, + "tokens_seen": 511508480 + }, + { + "epoch": 0.16, + "learning_rate": 0.0006828164493480442, + "loss": 1.5541, + "theoretical_loss": 3.933705405902619, + "tokens_seen": 511574016 + }, + { + "epoch": 0.16, + "learning_rate": 0.0006828004012036108, + "loss": 1.6257, + "theoretical_loss": 3.9336517368154666, + "tokens_seen": 511639552 + }, + { + "epoch": 0.16, + "learning_rate": 0.0006827843530591776, + "loss": 1.7031, + "theoretical_loss": 3.9335980765269323, + "tokens_seen": 511705088 + }, + { + "epoch": 0.16, + "learning_rate": 0.0006827683049147442, + "loss": 1.433, + "theoretical_loss": 3.9335444250344445, + "tokens_seen": 511770624 + }, + { + "epoch": 0.16, + "learning_rate": 0.000682752256770311, + "loss": 1.5689, + "theoretical_loss": 3.933490782335437, + "tokens_seen": 511836160 + }, + { + "epoch": 0.16, + "learning_rate": 0.0006827362086258776, + "loss": 1.6788, + "theoretical_loss": 3.933437148427342, + "tokens_seen": 511901696 + }, + { + "epoch": 0.16, + "learning_rate": 0.0006827201604814444, + "loss": 1.6155, + "theoretical_loss": 3.933383523307594, + "tokens_seen": 511967232 + }, + { + "epoch": 0.16, + "learning_rate": 0.000682704112337011, + "loss": 1.7009, + "theoretical_loss": 3.9333299069736283, + "tokens_seen": 512032768 + }, + { + "epoch": 0.16, + "learning_rate": 0.0006826880641925778, + "loss": 1.476, + "theoretical_loss": 3.9332762994228814, + "tokens_seen": 512098304 + }, + { + "epoch": 0.16, + "learning_rate": 0.0006826720160481444, + "loss": 1.5629, + "theoretical_loss": 3.93322270065279, + "tokens_seen": 512163840 + }, + { + "epoch": 0.16, + "learning_rate": 0.0006826559679037112, + "loss": 1.5482, + "theoretical_loss": 3.933169110660793, + "tokens_seen": 512229376 + }, + { + "epoch": 0.16, + "learning_rate": 0.0006826399197592778, + "loss": 1.5995, + "theoretical_loss": 3.93311552944433, + "tokens_seen": 512294912 + }, + { + "epoch": 0.16, + "learning_rate": 0.0006826238716148446, + "loss": 1.7785, + "theoretical_loss": 3.9330619570008416, + "tokens_seen": 512360448 + }, + { + "epoch": 0.16, + "learning_rate": 0.0006826078234704112, + "loss": 1.5225, + "theoretical_loss": 3.9330083933277695, + "tokens_seen": 512425984 + }, + { + "epoch": 0.16, + "learning_rate": 0.000682591775325978, + "loss": 1.5485, + "theoretical_loss": 3.9329548384225568, + "tokens_seen": 512491520 + }, + { + "epoch": 0.16, + "learning_rate": 0.0006825757271815446, + "loss": 1.5807, + "theoretical_loss": 3.932901292282647, + "tokens_seen": 512557056 + }, + { + "epoch": 0.16, + "learning_rate": 0.0006825596790371114, + "loss": 1.7058, + "theoretical_loss": 3.932847754905485, + "tokens_seen": 512622592 + }, + { + "epoch": 0.16, + "learning_rate": 0.000682543630892678, + "loss": 1.6738, + "theoretical_loss": 3.9327942262885167, + "tokens_seen": 512688128 + }, + { + "epoch": 0.16, + "learning_rate": 0.0006825275827482448, + "loss": 1.6435, + "theoretical_loss": 3.9327407064291897, + "tokens_seen": 512753664 + }, + { + "epoch": 0.16, + "learning_rate": 0.0006825115346038114, + "loss": 1.5101, + "theoretical_loss": 3.9326871953249514, + "tokens_seen": 512819200 + }, + { + "epoch": 0.16, + "learning_rate": 0.0006824954864593782, + "loss": 1.5159, + "theoretical_loss": 3.9326336929732513, + "tokens_seen": 512884736 + }, + { + "epoch": 0.16, + "learning_rate": 0.0006824794383149448, + "loss": 1.575, + "theoretical_loss": 3.9325801993715404, + "tokens_seen": 512950272 + }, + { + "epoch": 0.16, + "learning_rate": 0.0006824633901705116, + "loss": 1.5841, + "theoretical_loss": 3.9325267145172687, + "tokens_seen": 513015808 + }, + { + "epoch": 0.16, + "learning_rate": 0.0006824473420260782, + "loss": 1.6745, + "theoretical_loss": 3.932473238407889, + "tokens_seen": 513081344 + }, + { + "epoch": 0.16, + "learning_rate": 0.000682431293881645, + "loss": 1.6699, + "theoretical_loss": 3.932419771040856, + "tokens_seen": 513146880 + }, + { + "epoch": 0.16, + "learning_rate": 0.0006824152457372116, + "loss": 1.5316, + "theoretical_loss": 3.932366312413623, + "tokens_seen": 513212416 + }, + { + "epoch": 0.16, + "learning_rate": 0.0006823991975927784, + "loss": 1.6223, + "theoretical_loss": 3.932312862523646, + "tokens_seen": 513277952 + }, + { + "epoch": 0.16, + "learning_rate": 0.000682383149448345, + "loss": 1.5865, + "theoretical_loss": 3.9322594213683812, + "tokens_seen": 513343488 + }, + { + "epoch": 0.16, + "learning_rate": 0.0006823671013039118, + "loss": 1.4985, + "theoretical_loss": 3.9322059889452867, + "tokens_seen": 513409024 + }, + { + "epoch": 0.16, + "learning_rate": 0.0006823510531594784, + "loss": 1.6173, + "theoretical_loss": 3.9321525652518217, + "tokens_seen": 513474560 + }, + { + "epoch": 0.16, + "learning_rate": 0.0006823350050150452, + "loss": 1.6063, + "theoretical_loss": 3.9320991502854454, + "tokens_seen": 513540096 + }, + { + "epoch": 0.16, + "learning_rate": 0.0006823189568706118, + "loss": 1.6769, + "theoretical_loss": 3.932045744043619, + "tokens_seen": 513605632 + }, + { + "epoch": 0.16, + "learning_rate": 0.0006823029087261786, + "loss": 1.6557, + "theoretical_loss": 3.9319923465238054, + "tokens_seen": 513671168 + }, + { + "epoch": 0.16, + "learning_rate": 0.0006822868605817452, + "loss": 1.4954, + "theoretical_loss": 3.9319389577234665, + "tokens_seen": 513736704 + }, + { + "epoch": 0.16, + "learning_rate": 0.000682270812437312, + "loss": 1.6475, + "theoretical_loss": 3.931885577640066, + "tokens_seen": 513802240 + }, + { + "epoch": 0.16, + "learning_rate": 0.0006822547642928787, + "loss": 1.5576, + "theoretical_loss": 3.9318322062710704, + "tokens_seen": 513867776 + }, + { + "epoch": 0.16, + "learning_rate": 0.0006822387161484454, + "loss": 1.4589, + "theoretical_loss": 3.931778843613945, + "tokens_seen": 513933312 + }, + { + "epoch": 0.16, + "learning_rate": 0.0006822226680040121, + "loss": 1.5804, + "theoretical_loss": 3.9317254896661575, + "tokens_seen": 513998848 + }, + { + "epoch": 0.16, + "learning_rate": 0.0006822066198595788, + "loss": 1.6546, + "theoretical_loss": 3.9316721444251765, + "tokens_seen": 514064384 + }, + { + "epoch": 0.16, + "learning_rate": 0.0006821905717151455, + "loss": 1.575, + "theoretical_loss": 3.9316188078884706, + "tokens_seen": 514129920 + }, + { + "epoch": 0.16, + "learning_rate": 0.0006821745235707122, + "loss": 1.5117, + "theoretical_loss": 3.931565480053511, + "tokens_seen": 514195456 + }, + { + "epoch": 0.16, + "learning_rate": 0.000682158475426279, + "loss": 1.5987, + "theoretical_loss": 3.9315121609177694, + "tokens_seen": 514260992 + }, + { + "epoch": 0.16, + "learning_rate": 0.0006821424272818456, + "loss": 1.5735, + "theoretical_loss": 3.931458850478718, + "tokens_seen": 514326528 + }, + { + "epoch": 0.16, + "learning_rate": 0.0006821263791374124, + "loss": 1.6057, + "theoretical_loss": 3.9314055487338297, + "tokens_seen": 514392064 + }, + { + "epoch": 0.16, + "objective/train/docs_used": 193343, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.5015358924865723, + "objective/train/theoretical_loss": 3.93135225568058, + "objective/train/tokens_used": 534917600, + "theoretical_loss": 3.93135225568058, + "tokens_seen": 514457600 + }, + { + "epoch": 0.16, + "learning_rate": 0.000682110330992979, + "loss": 1.6002, + "theoretical_loss": 3.93135225568058, + "tokens_seen": 514457600 + }, + { + "epoch": 0.16, + "learning_rate": 0.0006820942828485458, + "loss": 1.6066, + "theoretical_loss": 3.9312989713164455, + "tokens_seen": 514523136 + }, + { + "epoch": 0.16, + "learning_rate": 0.0006820782347041124, + "loss": 1.5792, + "theoretical_loss": 3.9312456956389017, + "tokens_seen": 514588672 + }, + { + "epoch": 0.16, + "learning_rate": 0.0006820621865596792, + "loss": 1.6938, + "theoretical_loss": 3.931192428645427, + "tokens_seen": 514654208 + }, + { + "epoch": 0.16, + "learning_rate": 0.0006820461384152458, + "loss": 1.5246, + "theoretical_loss": 3.9311391703334997, + "tokens_seen": 514719744 + }, + { + "epoch": 0.16, + "learning_rate": 0.0006820300902708126, + "loss": 1.518, + "theoretical_loss": 3.9310859207006006, + "tokens_seen": 514785280 + }, + { + "epoch": 0.16, + "learning_rate": 0.0006820140421263792, + "loss": 1.6257, + "theoretical_loss": 3.9310326797442103, + "tokens_seen": 514850816 + }, + { + "epoch": 0.16, + "learning_rate": 0.000681997993981946, + "loss": 1.5721, + "theoretical_loss": 3.930979447461811, + "tokens_seen": 514916352 + }, + { + "epoch": 0.16, + "learning_rate": 0.0006819819458375126, + "loss": 1.6153, + "theoretical_loss": 3.9309262238508857, + "tokens_seen": 514981888 + }, + { + "epoch": 0.16, + "learning_rate": 0.0006819658976930794, + "loss": 1.6002, + "theoretical_loss": 3.930873008908919, + "tokens_seen": 515047424 + }, + { + "epoch": 0.16, + "learning_rate": 0.000681949849548646, + "loss": 1.6757, + "theoretical_loss": 3.9308198026333954, + "tokens_seen": 515112960 + }, + { + "epoch": 0.16, + "learning_rate": 0.0006819338014042126, + "loss": 1.6174, + "theoretical_loss": 3.9307666050218018, + "tokens_seen": 515178496 + }, + { + "epoch": 0.16, + "learning_rate": 0.0006819177532597794, + "loss": 1.5264, + "theoretical_loss": 3.930713416071625, + "tokens_seen": 515244032 + }, + { + "epoch": 0.16, + "learning_rate": 0.000681901705115346, + "loss": 1.5843, + "theoretical_loss": 3.930660235780354, + "tokens_seen": 515309568 + }, + { + "epoch": 0.16, + "learning_rate": 0.0006818856569709128, + "loss": 1.688, + "theoretical_loss": 3.9306070641454776, + "tokens_seen": 515375104 + }, + { + "epoch": 0.16, + "learning_rate": 0.0006818696088264795, + "loss": 1.5642, + "theoretical_loss": 3.9305539011644863, + "tokens_seen": 515440640 + }, + { + "epoch": 0.16, + "learning_rate": 0.0006818535606820462, + "loss": 1.5336, + "theoretical_loss": 3.9305007468348725, + "tokens_seen": 515506176 + }, + { + "epoch": 0.16, + "learning_rate": 0.0006818375125376129, + "loss": 1.665, + "theoretical_loss": 3.9304476011541274, + "tokens_seen": 515571712 + }, + { + "epoch": 0.16, + "learning_rate": 0.0006818214643931796, + "loss": 1.5775, + "theoretical_loss": 3.930394464119746, + "tokens_seen": 515637248 + }, + { + "epoch": 0.16, + "learning_rate": 0.0006818054162487463, + "loss": 1.6608, + "theoretical_loss": 3.9303413357292216, + "tokens_seen": 515702784 + }, + { + "epoch": 0.16, + "learning_rate": 0.0006817893681043129, + "loss": 1.5623, + "theoretical_loss": 3.930288215980051, + "tokens_seen": 515768320 + }, + { + "epoch": 0.16, + "learning_rate": 0.0006817733199598797, + "loss": 1.6305, + "theoretical_loss": 3.9302351048697304, + "tokens_seen": 515833856 + }, + { + "epoch": 0.16, + "learning_rate": 0.0006817572718154463, + "loss": 1.5846, + "theoretical_loss": 3.9301820023957577, + "tokens_seen": 515899392 + }, + { + "epoch": 0.16, + "learning_rate": 0.0006817412236710131, + "loss": 1.5992, + "theoretical_loss": 3.930128908555632, + "tokens_seen": 515964928 + }, + { + "epoch": 0.16, + "learning_rate": 0.0006817251755265797, + "loss": 1.6838, + "theoretical_loss": 3.9300758233468525, + "tokens_seen": 516030464 + }, + { + "epoch": 0.16, + "learning_rate": 0.0006817091273821465, + "loss": 1.5667, + "theoretical_loss": 3.9300227467669204, + "tokens_seen": 516096000 + }, + { + "epoch": 0.16, + "learning_rate": 0.0006816930792377131, + "loss": 1.5965, + "theoretical_loss": 3.929969678813338, + "tokens_seen": 516161536 + }, + { + "epoch": 0.16, + "learning_rate": 0.0006816770310932799, + "loss": 1.5507, + "theoretical_loss": 3.929916619483608, + "tokens_seen": 516227072 + }, + { + "epoch": 0.16, + "learning_rate": 0.0006816609829488465, + "loss": 1.5044, + "theoretical_loss": 3.9298635687752337, + "tokens_seen": 516292608 + }, + { + "epoch": 0.16, + "learning_rate": 0.0006816449348044133, + "loss": 1.4905, + "theoretical_loss": 3.9298105266857215, + "tokens_seen": 516358144 + }, + { + "epoch": 0.16, + "learning_rate": 0.0006816288866599799, + "loss": 1.568, + "theoretical_loss": 3.929757493212577, + "tokens_seen": 516423680 + }, + { + "epoch": 0.16, + "learning_rate": 0.0006816128385155467, + "loss": 1.5777, + "theoretical_loss": 3.929704468353307, + "tokens_seen": 516489216 + }, + { + "epoch": 0.16, + "learning_rate": 0.0006815967903711133, + "loss": 1.5613, + "theoretical_loss": 3.9296514521054196, + "tokens_seen": 516554752 + }, + { + "epoch": 0.16, + "learning_rate": 0.0006815807422266801, + "loss": 1.5599, + "theoretical_loss": 3.929598444466425, + "tokens_seen": 516620288 + }, + { + "epoch": 0.16, + "learning_rate": 0.0006815646940822467, + "loss": 1.6079, + "theoretical_loss": 3.9295454454338326, + "tokens_seen": 516685824 + }, + { + "epoch": 0.16, + "learning_rate": 0.0006815486459378135, + "loss": 1.527, + "theoretical_loss": 3.9294924550051533, + "tokens_seen": 516751360 + }, + { + "epoch": 0.16, + "learning_rate": 0.0006815325977933801, + "loss": 1.5251, + "theoretical_loss": 3.9294394731779008, + "tokens_seen": 516816896 + }, + { + "epoch": 0.16, + "learning_rate": 0.0006815165496489469, + "loss": 1.6439, + "theoretical_loss": 3.929386499949587, + "tokens_seen": 516882432 + }, + { + "epoch": 0.16, + "learning_rate": 0.0006815005015045135, + "loss": 1.6832, + "theoretical_loss": 3.929333535317727, + "tokens_seen": 516947968 + }, + { + "epoch": 0.16, + "learning_rate": 0.0006814844533600803, + "loss": 1.4488, + "theoretical_loss": 3.929280579279837, + "tokens_seen": 517013504 + }, + { + "epoch": 0.16, + "learning_rate": 0.0006814684052156469, + "loss": 1.4345, + "theoretical_loss": 3.929227631833432, + "tokens_seen": 517079040 + }, + { + "epoch": 0.16, + "learning_rate": 0.0006814523570712137, + "loss": 1.5578, + "theoretical_loss": 3.92917469297603, + "tokens_seen": 517144576 + }, + { + "epoch": 0.16, + "learning_rate": 0.0006814363089267803, + "loss": 1.6003, + "theoretical_loss": 3.9291217627051496, + "tokens_seen": 517210112 + }, + { + "epoch": 0.16, + "learning_rate": 0.0006814202607823471, + "loss": 1.5021, + "theoretical_loss": 3.9290688410183106, + "tokens_seen": 517275648 + }, + { + "epoch": 0.16, + "learning_rate": 0.0006814042126379137, + "loss": 1.61, + "theoretical_loss": 3.9290159279130332, + "tokens_seen": 517341184 + }, + { + "epoch": 0.16, + "learning_rate": 0.0006813881644934805, + "loss": 1.5352, + "theoretical_loss": 3.928963023386839, + "tokens_seen": 517406720 + }, + { + "epoch": 0.16, + "learning_rate": 0.0006813721163490471, + "loss": 1.5614, + "theoretical_loss": 3.9289101274372515, + "tokens_seen": 517472256 + }, + { + "epoch": 0.16, + "learning_rate": 0.0006813560682046139, + "loss": 1.4288, + "theoretical_loss": 3.9288572400617934, + "tokens_seen": 517537792 + }, + { + "epoch": 0.16, + "learning_rate": 0.0006813400200601805, + "loss": 1.5268, + "theoretical_loss": 3.9288043612579893, + "tokens_seen": 517603328 + }, + { + "epoch": 0.16, + "learning_rate": 0.0006813239719157473, + "loss": 1.5898, + "theoretical_loss": 3.928751491023366, + "tokens_seen": 517668864 + }, + { + "epoch": 0.16, + "objective/train/docs_used": 194581, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.6056010723114014, + "objective/train/theoretical_loss": 3.9286986293554493, + "objective/train/tokens_used": 538194400, + "theoretical_loss": 3.9286986293554493, + "tokens_seen": 517734400 + }, + { + "epoch": 0.16, + "learning_rate": 0.0006813079237713139, + "loss": 1.5588, + "theoretical_loss": 3.9286986293554493, + "tokens_seen": 517734400 + }, + { + "epoch": 0.16, + "learning_rate": 0.0006812918756268807, + "loss": 1.5683, + "theoretical_loss": 3.9286457762517673, + "tokens_seen": 517799936 + }, + { + "epoch": 0.16, + "learning_rate": 0.0006812758274824473, + "loss": 1.5707, + "theoretical_loss": 3.928592931709849, + "tokens_seen": 517865472 + }, + { + "epoch": 0.16, + "learning_rate": 0.0006812597793380141, + "loss": 1.6181, + "theoretical_loss": 3.9285400957272234, + "tokens_seen": 517931008 + }, + { + "epoch": 0.16, + "learning_rate": 0.0006812437311935807, + "loss": 1.51, + "theoretical_loss": 3.9284872683014225, + "tokens_seen": 517996544 + }, + { + "epoch": 0.16, + "learning_rate": 0.0006812276830491475, + "loss": 1.5046, + "theoretical_loss": 3.9284344494299774, + "tokens_seen": 518062080 + }, + { + "epoch": 0.16, + "learning_rate": 0.0006812116349047141, + "loss": 1.6166, + "theoretical_loss": 3.9283816391104214, + "tokens_seen": 518127616 + }, + { + "epoch": 0.16, + "learning_rate": 0.0006811955867602809, + "loss": 1.5323, + "theoretical_loss": 3.9283288373402883, + "tokens_seen": 518193152 + }, + { + "epoch": 0.16, + "learning_rate": 0.0006811795386158475, + "loss": 1.5853, + "theoretical_loss": 3.9282760441171134, + "tokens_seen": 518258688 + }, + { + "epoch": 0.16, + "learning_rate": 0.0006811634904714143, + "loss": 1.488, + "theoretical_loss": 3.9282232594384316, + "tokens_seen": 518324224 + }, + { + "epoch": 0.16, + "learning_rate": 0.0006811474423269809, + "loss": 1.442, + "theoretical_loss": 3.9281704833017814, + "tokens_seen": 518389760 + }, + { + "epoch": 0.16, + "learning_rate": 0.0006811313941825477, + "loss": 1.5532, + "theoretical_loss": 3.9281177157047003, + "tokens_seen": 518455296 + }, + { + "epoch": 0.16, + "learning_rate": 0.0006811153460381143, + "loss": 1.5115, + "theoretical_loss": 3.928064956644726, + "tokens_seen": 518520832 + }, + { + "epoch": 0.16, + "learning_rate": 0.0006810992978936811, + "loss": 1.5443, + "theoretical_loss": 3.9280122061194005, + "tokens_seen": 518586368 + }, + { + "epoch": 0.16, + "learning_rate": 0.0006810832497492477, + "loss": 1.5387, + "theoretical_loss": 3.927959464126264, + "tokens_seen": 518651904 + }, + { + "epoch": 0.16, + "learning_rate": 0.0006810672016048145, + "loss": 1.5598, + "theoretical_loss": 3.9279067306628583, + "tokens_seen": 518717440 + }, + { + "epoch": 0.16, + "learning_rate": 0.0006810511534603812, + "loss": 1.5121, + "theoretical_loss": 3.9278540057267266, + "tokens_seen": 518782976 + }, + { + "epoch": 0.16, + "learning_rate": 0.0006810351053159479, + "loss": 1.6723, + "theoretical_loss": 3.927801289315414, + "tokens_seen": 518848512 + }, + { + "epoch": 0.16, + "learning_rate": 0.0006810190571715146, + "loss": 1.5111, + "theoretical_loss": 3.9277485814264645, + "tokens_seen": 518914048 + }, + { + "epoch": 0.16, + "learning_rate": 0.0006810030090270813, + "loss": 1.4921, + "theoretical_loss": 3.9276958820574253, + "tokens_seen": 518979584 + }, + { + "epoch": 0.16, + "learning_rate": 0.000680986960882648, + "loss": 1.5723, + "theoretical_loss": 3.9276431912058425, + "tokens_seen": 519045120 + }, + { + "epoch": 0.16, + "learning_rate": 0.0006809709127382147, + "loss": 1.6162, + "theoretical_loss": 3.9275905088692653, + "tokens_seen": 519110656 + }, + { + "epoch": 0.16, + "learning_rate": 0.0006809548645937814, + "loss": 1.4961, + "theoretical_loss": 3.927537835045242, + "tokens_seen": 519176192 + }, + { + "epoch": 0.16, + "learning_rate": 0.0006809388164493481, + "loss": 1.5161, + "theoretical_loss": 3.927485169731323, + "tokens_seen": 519241728 + }, + { + "epoch": 0.16, + "learning_rate": 0.0006809227683049148, + "loss": 1.5776, + "theoretical_loss": 3.9274325129250602, + "tokens_seen": 519307264 + }, + { + "epoch": 0.16, + "learning_rate": 0.0006809067201604815, + "loss": 1.482, + "theoretical_loss": 3.9273798646240055, + "tokens_seen": 519372800 + }, + { + "epoch": 0.16, + "learning_rate": 0.0006808906720160482, + "loss": 1.563, + "theoretical_loss": 3.9273272248257123, + "tokens_seen": 519438336 + }, + { + "epoch": 0.16, + "learning_rate": 0.0006808746238716149, + "loss": 1.5167, + "theoretical_loss": 3.9272745935277347, + "tokens_seen": 519503872 + }, + { + "epoch": 0.16, + "learning_rate": 0.0006808585757271816, + "loss": 1.4608, + "theoretical_loss": 3.927221970727628, + "tokens_seen": 519569408 + }, + { + "epoch": 0.16, + "learning_rate": 0.0006808425275827483, + "loss": 1.4811, + "theoretical_loss": 3.927169356422948, + "tokens_seen": 519634944 + }, + { + "epoch": 0.16, + "learning_rate": 0.000680826479438315, + "loss": 1.501, + "theoretical_loss": 3.9271167506112534, + "tokens_seen": 519700480 + }, + { + "epoch": 0.16, + "learning_rate": 0.0006808104312938817, + "loss": 1.4824, + "theoretical_loss": 3.927064153290101, + "tokens_seen": 519766016 + }, + { + "epoch": 0.16, + "learning_rate": 0.0006807943831494484, + "loss": 1.5573, + "theoretical_loss": 3.9270115644570516, + "tokens_seen": 519831552 + }, + { + "epoch": 0.16, + "learning_rate": 0.0006807783350050151, + "loss": 1.5932, + "theoretical_loss": 3.9269589841096644, + "tokens_seen": 519897088 + }, + { + "epoch": 0.16, + "learning_rate": 0.0006807622868605819, + "loss": 1.6347, + "theoretical_loss": 3.9269064122455006, + "tokens_seen": 519962624 + }, + { + "epoch": 0.16, + "learning_rate": 0.0006807462387161485, + "loss": 1.4604, + "theoretical_loss": 3.926853848862124, + "tokens_seen": 520028160 + }, + { + "epoch": 0.16, + "learning_rate": 0.0006807301905717153, + "loss": 1.4686, + "theoretical_loss": 3.9268012939570967, + "tokens_seen": 520093696 + }, + { + "epoch": 0.16, + "learning_rate": 0.0006807141424272819, + "loss": 1.477, + "theoretical_loss": 3.9267487475279834, + "tokens_seen": 520159232 + }, + { + "epoch": 0.16, + "learning_rate": 0.0006806980942828487, + "loss": 1.634, + "theoretical_loss": 3.9266962095723494, + "tokens_seen": 520224768 + }, + { + "epoch": 0.16, + "learning_rate": 0.0006806820461384153, + "loss": 1.5457, + "theoretical_loss": 3.9266436800877615, + "tokens_seen": 520290304 + }, + { + "epoch": 0.16, + "learning_rate": 0.000680665997993982, + "loss": 1.5545, + "theoretical_loss": 3.926591159071787, + "tokens_seen": 520355840 + }, + { + "epoch": 0.16, + "learning_rate": 0.0006806499498495487, + "loss": 1.5212, + "theoretical_loss": 3.9265386465219936, + "tokens_seen": 520421376 + }, + { + "epoch": 0.16, + "learning_rate": 0.0006806339017051155, + "loss": 1.498, + "theoretical_loss": 3.9264861424359516, + "tokens_seen": 520486912 + }, + { + "epoch": 0.16, + "learning_rate": 0.0006806178535606821, + "loss": 1.5941, + "theoretical_loss": 3.9264336468112315, + "tokens_seen": 520552448 + }, + { + "epoch": 0.16, + "learning_rate": 0.0006806018054162489, + "loss": 1.456, + "theoretical_loss": 3.9263811596454037, + "tokens_seen": 520617984 + }, + { + "epoch": 0.16, + "learning_rate": 0.0006805857572718155, + "loss": 1.5561, + "theoretical_loss": 3.9263286809360407, + "tokens_seen": 520683520 + }, + { + "epoch": 0.16, + "learning_rate": 0.0006805697091273823, + "loss": 1.5285, + "theoretical_loss": 3.9262762106807174, + "tokens_seen": 520749056 + }, + { + "epoch": 0.16, + "learning_rate": 0.0006805536609829489, + "loss": 1.4868, + "theoretical_loss": 3.926223748877007, + "tokens_seen": 520814592 + }, + { + "epoch": 0.16, + "learning_rate": 0.0006805376128385156, + "loss": 1.4654, + "theoretical_loss": 3.926171295522485, + "tokens_seen": 520880128 + }, + { + "epoch": 0.16, + "learning_rate": 0.0006805215646940823, + "loss": 1.5596, + "theoretical_loss": 3.9261188506147278, + "tokens_seen": 520945664 + }, + { + "epoch": 0.16, + "objective/train/docs_used": 195804, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.6612575054168701, + "objective/train/theoretical_loss": 3.926066414151313, + "objective/train/tokens_used": 541471200, + "theoretical_loss": 3.926066414151313, + "tokens_seen": 521011200 + }, + { + "epoch": 0.16, + "learning_rate": 0.000680505516549649, + "loss": 1.6041, + "theoretical_loss": 3.926066414151313, + "tokens_seen": 521011200 + }, + { + "epoch": 0.16, + "learning_rate": 0.0006804894684052157, + "loss": 1.6128, + "theoretical_loss": 3.926013986129819, + "tokens_seen": 521076736 + }, + { + "epoch": 0.16, + "learning_rate": 0.0006804734202607824, + "loss": 1.5746, + "theoretical_loss": 3.9259615665478256, + "tokens_seen": 521142272 + }, + { + "epoch": 0.16, + "learning_rate": 0.0006804573721163491, + "loss": 1.4821, + "theoretical_loss": 3.9259091554029126, + "tokens_seen": 521207808 + }, + { + "epoch": 0.16, + "learning_rate": 0.0006804413239719158, + "loss": 1.6004, + "theoretical_loss": 3.9258567526926615, + "tokens_seen": 521273344 + }, + { + "epoch": 0.16, + "learning_rate": 0.0006804252758274825, + "loss": 1.4902, + "theoretical_loss": 3.9258043584146556, + "tokens_seen": 521338880 + }, + { + "epoch": 0.16, + "learning_rate": 0.0006804092276830492, + "loss": 1.5623, + "theoretical_loss": 3.925751972566477, + "tokens_seen": 521404416 + }, + { + "epoch": 0.16, + "learning_rate": 0.0006803931795386158, + "loss": 1.6712, + "theoretical_loss": 3.9256995951457103, + "tokens_seen": 521469952 + }, + { + "epoch": 0.16, + "learning_rate": 0.0006803771313941826, + "loss": 1.5212, + "theoretical_loss": 3.925647226149942, + "tokens_seen": 521535488 + }, + { + "epoch": 0.16, + "learning_rate": 0.0006803610832497492, + "loss": 1.7339, + "theoretical_loss": 3.925594865576758, + "tokens_seen": 521601024 + }, + { + "epoch": 0.16, + "learning_rate": 0.000680345035105316, + "loss": 1.4738, + "theoretical_loss": 3.9255425134237445, + "tokens_seen": 521666560 + }, + { + "epoch": 0.16, + "learning_rate": 0.0006803289869608826, + "loss": 1.6473, + "theoretical_loss": 3.9254901696884916, + "tokens_seen": 521732096 + }, + { + "epoch": 0.16, + "learning_rate": 0.0006803129388164494, + "loss": 1.4753, + "theoretical_loss": 3.925437834368588, + "tokens_seen": 521797632 + }, + { + "epoch": 0.16, + "learning_rate": 0.000680296890672016, + "loss": 1.5412, + "theoretical_loss": 3.925385507461624, + "tokens_seen": 521863168 + }, + { + "epoch": 0.16, + "learning_rate": 0.0006802808425275828, + "loss": 1.4833, + "theoretical_loss": 3.9253331889651912, + "tokens_seen": 521928704 + }, + { + "epoch": 0.16, + "learning_rate": 0.0006802647943831494, + "loss": 1.4834, + "theoretical_loss": 3.9252808788768814, + "tokens_seen": 521994240 + }, + { + "epoch": 0.16, + "learning_rate": 0.0006802487462387162, + "loss": 1.513, + "theoretical_loss": 3.9252285771942876, + "tokens_seen": 522059776 + }, + { + "epoch": 0.16, + "learning_rate": 0.0006802326980942828, + "loss": 1.6248, + "theoretical_loss": 3.9251762839150057, + "tokens_seen": 522125312 + }, + { + "epoch": 0.16, + "learning_rate": 0.0006802166499498496, + "loss": 1.6406, + "theoretical_loss": 3.9251239990366305, + "tokens_seen": 522190848 + }, + { + "epoch": 0.16, + "learning_rate": 0.0006802006018054162, + "loss": 1.6336, + "theoretical_loss": 3.925071722556757, + "tokens_seen": 522256384 + }, + { + "epoch": 0.16, + "learning_rate": 0.000680184553660983, + "loss": 1.5256, + "theoretical_loss": 3.9250194544729844, + "tokens_seen": 522321920 + }, + { + "epoch": 0.16, + "learning_rate": 0.0006801685055165496, + "loss": 1.6279, + "theoretical_loss": 3.9249671947829095, + "tokens_seen": 522387456 + }, + { + "epoch": 0.16, + "learning_rate": 0.0006801524573721164, + "loss": 1.5059, + "theoretical_loss": 3.924914943484133, + "tokens_seen": 522452992 + }, + { + "epoch": 0.16, + "learning_rate": 0.000680136409227683, + "loss": 1.4339, + "theoretical_loss": 3.9248627005742533, + "tokens_seen": 522518528 + }, + { + "epoch": 0.16, + "learning_rate": 0.0006801203610832498, + "loss": 1.5058, + "theoretical_loss": 3.9248104660508734, + "tokens_seen": 522584064 + }, + { + "epoch": 0.16, + "learning_rate": 0.0006801043129388164, + "loss": 1.6381, + "theoretical_loss": 3.9247582399115943, + "tokens_seen": 522649600 + }, + { + "epoch": 0.16, + "learning_rate": 0.0006800882647943832, + "loss": 1.5873, + "theoretical_loss": 3.92470602215402, + "tokens_seen": 522715136 + }, + { + "epoch": 0.16, + "learning_rate": 0.0006800722166499498, + "loss": 1.5993, + "theoretical_loss": 3.9246538127757544, + "tokens_seen": 522780672 + }, + { + "epoch": 0.16, + "learning_rate": 0.0006800561685055166, + "loss": 1.6311, + "theoretical_loss": 3.9246016117744036, + "tokens_seen": 522846208 + }, + { + "epoch": 0.16, + "learning_rate": 0.0006800401203610832, + "loss": 1.522, + "theoretical_loss": 3.924549419147572, + "tokens_seen": 522911744 + }, + { + "epoch": 0.16, + "learning_rate": 0.00068002407221665, + "loss": 1.4226, + "theoretical_loss": 3.9244972348928684, + "tokens_seen": 522977280 + }, + { + "epoch": 0.16, + "learning_rate": 0.0006800080240722166, + "loss": 1.6148, + "theoretical_loss": 3.9244450590079003, + "tokens_seen": 523042816 + }, + { + "epoch": 0.16, + "learning_rate": 0.0006799919759277834, + "loss": 1.5619, + "theoretical_loss": 3.9243928914902764, + "tokens_seen": 523108352 + }, + { + "epoch": 0.16, + "learning_rate": 0.00067997592778335, + "loss": 1.5992, + "theoretical_loss": 3.924340732337608, + "tokens_seen": 523173888 + }, + { + "epoch": 0.16, + "learning_rate": 0.0006799598796389168, + "loss": 1.6026, + "theoretical_loss": 3.924288581547505, + "tokens_seen": 523239424 + }, + { + "epoch": 0.16, + "learning_rate": 0.0006799438314944834, + "loss": 1.5208, + "theoretical_loss": 3.9242364391175797, + "tokens_seen": 523304960 + }, + { + "epoch": 0.16, + "learning_rate": 0.0006799277833500502, + "loss": 1.5145, + "theoretical_loss": 3.924184305045446, + "tokens_seen": 523370496 + }, + { + "epoch": 0.16, + "learning_rate": 0.0006799117352056168, + "loss": 1.4722, + "theoretical_loss": 3.9241321793287174, + "tokens_seen": 523436032 + }, + { + "epoch": 0.16, + "learning_rate": 0.0006798956870611836, + "loss": 1.5879, + "theoretical_loss": 3.9240800619650082, + "tokens_seen": 523501568 + }, + { + "epoch": 0.16, + "learning_rate": 0.0006798796389167503, + "loss": 1.5646, + "theoretical_loss": 3.9240279529519357, + "tokens_seen": 523567104 + }, + { + "epoch": 0.16, + "learning_rate": 0.000679863590772317, + "loss": 1.6876, + "theoretical_loss": 3.9239758522871164, + "tokens_seen": 523632640 + }, + { + "epoch": 0.16, + "learning_rate": 0.0006798475426278837, + "loss": 1.4864, + "theoretical_loss": 3.923923759968168, + "tokens_seen": 523698176 + }, + { + "epoch": 0.16, + "learning_rate": 0.0006798314944834504, + "loss": 1.5098, + "theoretical_loss": 3.9238716759927095, + "tokens_seen": 523763712 + }, + { + "epoch": 0.16, + "learning_rate": 0.0006798154463390171, + "loss": 1.5832, + "theoretical_loss": 3.923819600358361, + "tokens_seen": 523829248 + }, + { + "epoch": 0.16, + "learning_rate": 0.0006797993981945838, + "loss": 1.6905, + "theoretical_loss": 3.923767533062743, + "tokens_seen": 523894784 + }, + { + "epoch": 0.16, + "learning_rate": 0.0006797833500501505, + "loss": 1.4721, + "theoretical_loss": 3.9237154741034783, + "tokens_seen": 523960320 + }, + { + "epoch": 0.16, + "learning_rate": 0.0006797673019057172, + "loss": 1.5683, + "theoretical_loss": 3.923663423478189, + "tokens_seen": 524025856 + }, + { + "epoch": 0.16, + "learning_rate": 0.0006797512537612839, + "loss": 1.6094, + "theoretical_loss": 3.9236113811844984, + "tokens_seen": 524091392 + }, + { + "epoch": 0.16, + "learning_rate": 0.0006797352056168506, + "loss": 1.5399, + "theoretical_loss": 3.9235593472200323, + "tokens_seen": 524156928 + }, + { + "epoch": 0.16, + "learning_rate": 0.0006797191574724173, + "loss": 1.5415, + "theoretical_loss": 3.9235073215824166, + "tokens_seen": 524222464 + }, + { + "epoch": 0.16, + "objective/train/docs_used": 196964, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.4494655132293701, + "objective/train/theoretical_loss": 3.9234553042692766, + "objective/train/tokens_used": 544748000, + "theoretical_loss": 3.9234553042692766, + "tokens_seen": 524288000 + }, + { + "epoch": 0.16, + "learning_rate": 0.000679703109327984, + "loss": 1.4826, + "theoretical_loss": 3.9234553042692766, + "tokens_seen": 524288000 + }, + { + "epoch": 0.16, + "learning_rate": 0.0006796870611835507, + "loss": 1.5332, + "theoretical_loss": 3.9234032952782414, + "tokens_seen": 524353536 + }, + { + "epoch": 0.16, + "learning_rate": 0.0006796710130391174, + "loss": 1.6129, + "theoretical_loss": 3.923351294606939, + "tokens_seen": 524419072 + }, + { + "epoch": 0.16, + "learning_rate": 0.0006796549648946841, + "loss": 1.5477, + "theoretical_loss": 3.923299302253, + "tokens_seen": 524484608 + }, + { + "epoch": 0.16, + "learning_rate": 0.0006796389167502508, + "loss": 1.5173, + "theoretical_loss": 3.923247318214054, + "tokens_seen": 524550144 + }, + { + "epoch": 0.16, + "learning_rate": 0.0006796228686058175, + "loss": 1.4478, + "theoretical_loss": 3.923195342487733, + "tokens_seen": 524615680 + }, + { + "epoch": 0.16, + "learning_rate": 0.0006796068204613842, + "loss": 1.5623, + "theoretical_loss": 3.9231433750716693, + "tokens_seen": 524681216 + }, + { + "epoch": 0.16, + "learning_rate": 0.0006795907723169509, + "loss": 1.5695, + "theoretical_loss": 3.9230914159634964, + "tokens_seen": 524746752 + }, + { + "epoch": 0.16, + "learning_rate": 0.0006795747241725176, + "loss": 1.4031, + "theoretical_loss": 3.9230394651608496, + "tokens_seen": 524812288 + }, + { + "epoch": 0.16, + "learning_rate": 0.0006795586760280843, + "loss": 1.7101, + "theoretical_loss": 3.922987522661363, + "tokens_seen": 524877824 + }, + { + "epoch": 0.16, + "learning_rate": 0.000679542627883651, + "loss": 1.5572, + "theoretical_loss": 3.922935588462675, + "tokens_seen": 524943360 + }, + { + "epoch": 0.16, + "learning_rate": 0.0006795265797392177, + "loss": 1.3978, + "theoretical_loss": 3.9228836625624206, + "tokens_seen": 525008896 + }, + { + "epoch": 0.16, + "learning_rate": 0.0006795105315947844, + "loss": 1.5304, + "theoretical_loss": 3.9228317449582404, + "tokens_seen": 525074432 + }, + { + "epoch": 0.16, + "learning_rate": 0.0006794944834503511, + "loss": 1.5129, + "theoretical_loss": 3.9227798356477726, + "tokens_seen": 525139968 + }, + { + "epoch": 0.16, + "learning_rate": 0.0006794784353059178, + "loss": 1.5415, + "theoretical_loss": 3.922727934628658, + "tokens_seen": 525205504 + }, + { + "epoch": 0.16, + "learning_rate": 0.0006794623871614845, + "loss": 1.5755, + "theoretical_loss": 3.922676041898537, + "tokens_seen": 525271040 + }, + { + "epoch": 0.16, + "learning_rate": 0.0006794463390170512, + "loss": 1.6102, + "theoretical_loss": 3.9226241574550533, + "tokens_seen": 525336576 + }, + { + "epoch": 0.16, + "learning_rate": 0.000679430290872618, + "loss": 1.5868, + "theoretical_loss": 3.9225722812958486, + "tokens_seen": 525402112 + }, + { + "epoch": 0.16, + "learning_rate": 0.0006794142427281846, + "loss": 1.6712, + "theoretical_loss": 3.922520413418568, + "tokens_seen": 525467648 + }, + { + "epoch": 0.16, + "learning_rate": 0.0006793981945837513, + "loss": 1.5872, + "theoretical_loss": 3.922468553820856, + "tokens_seen": 525533184 + }, + { + "epoch": 0.16, + "learning_rate": 0.000679382146439318, + "loss": 1.6025, + "theoretical_loss": 3.9224167025003593, + "tokens_seen": 525598720 + }, + { + "epoch": 0.16, + "learning_rate": 0.0006793660982948848, + "loss": 1.591, + "theoretical_loss": 3.922364859454725, + "tokens_seen": 525664256 + }, + { + "epoch": 0.16, + "learning_rate": 0.0006793500501504514, + "loss": 1.4947, + "theoretical_loss": 3.9223130246816007, + "tokens_seen": 525729792 + }, + { + "epoch": 0.16, + "learning_rate": 0.0006793340020060182, + "loss": 1.5136, + "theoretical_loss": 3.9222611981786355, + "tokens_seen": 525795328 + }, + { + "epoch": 0.16, + "learning_rate": 0.0006793179538615848, + "loss": 1.5604, + "theoretical_loss": 3.92220937994348, + "tokens_seen": 525860864 + }, + { + "epoch": 0.16, + "learning_rate": 0.0006793019057171516, + "loss": 1.4916, + "theoretical_loss": 3.9221575699737836, + "tokens_seen": 525926400 + }, + { + "epoch": 0.16, + "learning_rate": 0.0006792858575727182, + "loss": 1.5604, + "theoretical_loss": 3.9221057682672, + "tokens_seen": 525991936 + }, + { + "epoch": 0.16, + "learning_rate": 0.000679269809428285, + "loss": 1.5832, + "theoretical_loss": 3.92205397482138, + "tokens_seen": 526057472 + }, + { + "epoch": 0.16, + "learning_rate": 0.0006792537612838516, + "loss": 1.4387, + "theoretical_loss": 3.9220021896339796, + "tokens_seen": 526123008 + }, + { + "epoch": 0.16, + "learning_rate": 0.0006792377131394184, + "loss": 1.6231, + "theoretical_loss": 3.921950412702652, + "tokens_seen": 526188544 + }, + { + "epoch": 0.16, + "learning_rate": 0.000679221664994985, + "loss": 1.5379, + "theoretical_loss": 3.9218986440250534, + "tokens_seen": 526254080 + }, + { + "epoch": 0.16, + "learning_rate": 0.0006792056168505518, + "loss": 1.6114, + "theoretical_loss": 3.9218468835988403, + "tokens_seen": 526319616 + }, + { + "epoch": 0.16, + "learning_rate": 0.0006791895687061184, + "loss": 1.5921, + "theoretical_loss": 3.9217951314216704, + "tokens_seen": 526385152 + }, + { + "epoch": 0.16, + "learning_rate": 0.0006791735205616852, + "loss": 1.6027, + "theoretical_loss": 3.9217433874912024, + "tokens_seen": 526450688 + }, + { + "epoch": 0.16, + "learning_rate": 0.0006791574724172518, + "loss": 1.4855, + "theoretical_loss": 3.921691651805096, + "tokens_seen": 526516224 + }, + { + "epoch": 0.16, + "learning_rate": 0.0006791414242728185, + "loss": 1.5662, + "theoretical_loss": 3.9216399243610116, + "tokens_seen": 526581760 + }, + { + "epoch": 0.16, + "learning_rate": 0.0006791253761283852, + "loss": 1.55, + "theoretical_loss": 3.9215882051566098, + "tokens_seen": 526647296 + }, + { + "epoch": 0.16, + "learning_rate": 0.0006791093279839519, + "loss": 1.5521, + "theoretical_loss": 3.921536494189554, + "tokens_seen": 526712832 + }, + { + "epoch": 0.16, + "learning_rate": 0.0006790932798395186, + "loss": 1.6197, + "theoretical_loss": 3.921484791457507, + "tokens_seen": 526778368 + }, + { + "epoch": 0.16, + "learning_rate": 0.0006790772316950853, + "loss": 1.4831, + "theoretical_loss": 3.921433096958134, + "tokens_seen": 526843904 + }, + { + "epoch": 0.16, + "learning_rate": 0.000679061183550652, + "loss": 1.5169, + "theoretical_loss": 3.921381410689099, + "tokens_seen": 526909440 + }, + { + "epoch": 0.16, + "learning_rate": 0.0006790451354062187, + "loss": 1.5599, + "theoretical_loss": 3.9213297326480694, + "tokens_seen": 526974976 + }, + { + "epoch": 0.16, + "learning_rate": 0.0006790290872617854, + "loss": 1.4764, + "theoretical_loss": 3.9212780628327115, + "tokens_seen": 527040512 + }, + { + "epoch": 0.16, + "learning_rate": 0.0006790130391173521, + "loss": 1.6027, + "theoretical_loss": 3.921226401240693, + "tokens_seen": 527106048 + }, + { + "epoch": 0.16, + "learning_rate": 0.0006789969909729187, + "loss": 1.4843, + "theoretical_loss": 3.9211747478696846, + "tokens_seen": 527171584 + }, + { + "epoch": 0.16, + "learning_rate": 0.0006789809428284855, + "loss": 1.5065, + "theoretical_loss": 3.9211231027173548, + "tokens_seen": 527237120 + }, + { + "epoch": 0.16, + "learning_rate": 0.0006789648946840521, + "loss": 1.5867, + "theoretical_loss": 3.921071465781375, + "tokens_seen": 527302656 + }, + { + "epoch": 0.16, + "learning_rate": 0.0006789488465396189, + "loss": 1.5904, + "theoretical_loss": 3.9210198370594176, + "tokens_seen": 527368192 + }, + { + "epoch": 0.16, + "learning_rate": 0.0006789327983951855, + "loss": 1.5203, + "theoretical_loss": 3.920968216549155, + "tokens_seen": 527433728 + }, + { + "epoch": 0.16, + "learning_rate": 0.0006789167502507523, + "loss": 1.5543, + "theoretical_loss": 3.9209166042482613, + "tokens_seen": 527499264 + }, + { + "epoch": 0.16, + "objective/train/docs_used": 198271, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.4092450141906738, + "objective/train/theoretical_loss": 3.9208650001544108, + "objective/train/tokens_used": 548024800, + "theoretical_loss": 3.9208650001544108, + "tokens_seen": 527564800 + }, + { + "epoch": 0.16, + "learning_rate": 0.0006789007021063189, + "loss": 1.4891, + "theoretical_loss": 3.9208650001544108, + "tokens_seen": 527564800 + }, + { + "epoch": 0.16, + "learning_rate": 0.0006788846539618857, + "loss": 1.573, + "theoretical_loss": 3.920813404265279, + "tokens_seen": 527630336 + }, + { + "epoch": 0.16, + "learning_rate": 0.0006788686058174523, + "loss": 1.5386, + "theoretical_loss": 3.920761816578543, + "tokens_seen": 527695872 + }, + { + "epoch": 0.16, + "learning_rate": 0.0006788525576730191, + "loss": 1.5888, + "theoretical_loss": 3.920710237091881, + "tokens_seen": 527761408 + }, + { + "epoch": 0.16, + "learning_rate": 0.0006788365095285857, + "loss": 1.4832, + "theoretical_loss": 3.9206586658029714, + "tokens_seen": 527826944 + }, + { + "epoch": 0.16, + "learning_rate": 0.0006788204613841525, + "loss": 1.6367, + "theoretical_loss": 3.920607102709493, + "tokens_seen": 527892480 + }, + { + "epoch": 0.16, + "learning_rate": 0.0006788044132397191, + "loss": 1.5337, + "theoretical_loss": 3.9205555478091254, + "tokens_seen": 527958016 + }, + { + "epoch": 0.16, + "learning_rate": 0.0006787883650952859, + "loss": 1.602, + "theoretical_loss": 3.920504001099552, + "tokens_seen": 528023552 + }, + { + "epoch": 0.16, + "learning_rate": 0.0006787723169508525, + "loss": 1.534, + "theoretical_loss": 3.920452462578454, + "tokens_seen": 528089088 + }, + { + "epoch": 0.16, + "learning_rate": 0.0006787562688064193, + "loss": 1.5802, + "theoretical_loss": 3.920400932243515, + "tokens_seen": 528154624 + }, + { + "epoch": 0.16, + "learning_rate": 0.0006787402206619859, + "loss": 1.4562, + "theoretical_loss": 3.920349410092419, + "tokens_seen": 528220160 + }, + { + "epoch": 0.16, + "learning_rate": 0.0006787241725175527, + "loss": 1.5524, + "theoretical_loss": 3.920297896122851, + "tokens_seen": 528285696 + }, + { + "epoch": 0.16, + "learning_rate": 0.0006787081243731193, + "loss": 1.5781, + "theoretical_loss": 3.9202463903324976, + "tokens_seen": 528351232 + }, + { + "epoch": 0.16, + "learning_rate": 0.0006786920762286861, + "loss": 1.5468, + "theoretical_loss": 3.920194892719045, + "tokens_seen": 528416768 + }, + { + "epoch": 0.16, + "learning_rate": 0.0006786760280842528, + "loss": 1.544, + "theoretical_loss": 3.920143403280182, + "tokens_seen": 528482304 + }, + { + "epoch": 0.16, + "learning_rate": 0.0006786599799398195, + "loss": 1.487, + "theoretical_loss": 3.9200919220135972, + "tokens_seen": 528547840 + }, + { + "epoch": 0.16, + "learning_rate": 0.0006786439317953862, + "loss": 1.583, + "theoretical_loss": 3.920040448916981, + "tokens_seen": 528613376 + }, + { + "epoch": 0.16, + "learning_rate": 0.0006786278836509529, + "loss": 1.5736, + "theoretical_loss": 3.919988983988023, + "tokens_seen": 528678912 + }, + { + "epoch": 0.16, + "learning_rate": 0.0006786118355065196, + "loss": 1.4574, + "theoretical_loss": 3.919937527224416, + "tokens_seen": 528744448 + }, + { + "epoch": 0.16, + "learning_rate": 0.0006785957873620863, + "loss": 1.5227, + "theoretical_loss": 3.9198860786238523, + "tokens_seen": 528809984 + }, + { + "epoch": 0.16, + "learning_rate": 0.000678579739217653, + "loss": 1.6613, + "theoretical_loss": 3.919834638184025, + "tokens_seen": 528875520 + }, + { + "epoch": 0.16, + "learning_rate": 0.0006785636910732197, + "loss": 1.5751, + "theoretical_loss": 3.91978320590263, + "tokens_seen": 528941056 + }, + { + "epoch": 0.16, + "learning_rate": 0.0006785476429287864, + "loss": 1.5807, + "theoretical_loss": 3.919731781777361, + "tokens_seen": 529006592 + }, + { + "epoch": 0.16, + "learning_rate": 0.0006785315947843531, + "loss": 1.5611, + "theoretical_loss": 3.919680365805916, + "tokens_seen": 529072128 + }, + { + "epoch": 0.16, + "learning_rate": 0.0006785155466399198, + "loss": 1.4909, + "theoretical_loss": 3.9196289579859913, + "tokens_seen": 529137664 + }, + { + "epoch": 0.16, + "learning_rate": 0.0006784994984954865, + "loss": 1.5177, + "theoretical_loss": 3.919577558315286, + "tokens_seen": 529203200 + }, + { + "epoch": 0.16, + "learning_rate": 0.0006784834503510532, + "loss": 1.5801, + "theoretical_loss": 3.9195261667914982, + "tokens_seen": 529268736 + }, + { + "epoch": 0.16, + "learning_rate": 0.0006784674022066199, + "loss": 1.4803, + "theoretical_loss": 3.91947478341233, + "tokens_seen": 529334272 + }, + { + "epoch": 0.16, + "learning_rate": 0.0006784513540621866, + "loss": 1.5342, + "theoretical_loss": 3.9194234081754797, + "tokens_seen": 529399808 + }, + { + "epoch": 0.16, + "learning_rate": 0.0006784353059177533, + "loss": 1.5152, + "theoretical_loss": 3.9193720410786517, + "tokens_seen": 529465344 + }, + { + "epoch": 0.16, + "learning_rate": 0.00067841925777332, + "loss": 1.7123, + "theoretical_loss": 3.9193206821195483, + "tokens_seen": 529530880 + }, + { + "epoch": 0.16, + "learning_rate": 0.0006784032096288867, + "loss": 1.4628, + "theoretical_loss": 3.9192693312958733, + "tokens_seen": 529596416 + }, + { + "epoch": 0.16, + "learning_rate": 0.0006783871614844534, + "loss": 1.5661, + "theoretical_loss": 3.919217988605331, + "tokens_seen": 529661952 + }, + { + "epoch": 0.16, + "learning_rate": 0.0006783711133400201, + "loss": 1.5753, + "theoretical_loss": 3.919166654045628, + "tokens_seen": 529727488 + }, + { + "epoch": 0.16, + "learning_rate": 0.0006783550651955868, + "loss": 1.5999, + "theoretical_loss": 3.919115327614471, + "tokens_seen": 529793024 + }, + { + "epoch": 0.16, + "learning_rate": 0.0006783390170511535, + "loss": 1.6098, + "theoretical_loss": 3.9190640093095666, + "tokens_seen": 529858560 + }, + { + "epoch": 0.16, + "learning_rate": 0.0006783229689067202, + "loss": 1.6245, + "theoretical_loss": 3.9190126991286243, + "tokens_seen": 529924096 + }, + { + "epoch": 0.16, + "learning_rate": 0.0006783069207622869, + "loss": 1.5429, + "theoretical_loss": 3.918961397069353, + "tokens_seen": 529989632 + }, + { + "epoch": 0.16, + "learning_rate": 0.0006782908726178536, + "loss": 1.6523, + "theoretical_loss": 3.9189101031294635, + "tokens_seen": 530055168 + }, + { + "epoch": 0.16, + "learning_rate": 0.0006782748244734203, + "loss": 1.4156, + "theoretical_loss": 3.918858817306667, + "tokens_seen": 530120704 + }, + { + "epoch": 0.16, + "learning_rate": 0.000678258776328987, + "loss": 1.4153, + "theoretical_loss": 3.918807539598676, + "tokens_seen": 530186240 + }, + { + "epoch": 0.16, + "learning_rate": 0.0006782427281845537, + "loss": 1.5113, + "theoretical_loss": 3.918756270003203, + "tokens_seen": 530251776 + }, + { + "epoch": 0.16, + "learning_rate": 0.0006782266800401204, + "loss": 1.5588, + "theoretical_loss": 3.9187050085179633, + "tokens_seen": 530317312 + }, + { + "epoch": 0.16, + "learning_rate": 0.0006782106318956871, + "loss": 1.5236, + "theoretical_loss": 3.918653755140671, + "tokens_seen": 530382848 + }, + { + "epoch": 0.16, + "learning_rate": 0.0006781945837512538, + "loss": 1.4759, + "theoretical_loss": 3.918602509869042, + "tokens_seen": 530448384 + }, + { + "epoch": 0.16, + "learning_rate": 0.0006781785356068205, + "loss": 1.5559, + "theoretical_loss": 3.9185512727007934, + "tokens_seen": 530513920 + }, + { + "epoch": 0.16, + "learning_rate": 0.0006781624874623872, + "loss": 1.5392, + "theoretical_loss": 3.9185000436336432, + "tokens_seen": 530579456 + }, + { + "epoch": 0.16, + "learning_rate": 0.0006781464393179539, + "loss": 1.5733, + "theoretical_loss": 3.9184488226653102, + "tokens_seen": 530644992 + }, + { + "epoch": 0.16, + "learning_rate": 0.0006781303911735206, + "loss": 1.5333, + "theoretical_loss": 3.918397609793514, + "tokens_seen": 530710528 + }, + { + "epoch": 0.16, + "learning_rate": 0.0006781143430290873, + "loss": 1.5337, + "theoretical_loss": 3.9183464050159746, + "tokens_seen": 530776064 + }, + { + "epoch": 0.16, + "objective/train/docs_used": 199409, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.5797431468963623, + "objective/train/theoretical_loss": 3.918295208330415, + "objective/train/tokens_used": 551301600, + "theoretical_loss": 3.918295208330415, + "tokens_seen": 530841600 + }, + { + "epoch": 0.16, + "learning_rate": 0.000678098294884654, + "loss": 1.5212, + "theoretical_loss": 3.918295208330415, + "tokens_seen": 530841600 + }, + { + "epoch": 0.16, + "learning_rate": 0.0006780822467402207, + "loss": 1.5535, + "theoretical_loss": 3.918244019734556, + "tokens_seen": 530907136 + }, + { + "epoch": 0.16, + "learning_rate": 0.0006780661985957874, + "loss": 1.5273, + "theoretical_loss": 3.9181928392261214, + "tokens_seen": 530972672 + }, + { + "epoch": 0.16, + "learning_rate": 0.0006780501504513541, + "loss": 1.4338, + "theoretical_loss": 3.9181416668028355, + "tokens_seen": 531038208 + }, + { + "epoch": 0.16, + "learning_rate": 0.0006780341023069208, + "loss": 1.4649, + "theoretical_loss": 3.918090502462425, + "tokens_seen": 531103744 + }, + { + "epoch": 0.16, + "learning_rate": 0.0006780180541624875, + "loss": 1.6405, + "theoretical_loss": 3.9180393462026135, + "tokens_seen": 531169280 + }, + { + "epoch": 0.16, + "learning_rate": 0.0006780020060180543, + "loss": 1.6119, + "theoretical_loss": 3.9179881980211295, + "tokens_seen": 531234816 + }, + { + "epoch": 0.16, + "learning_rate": 0.0006779859578736209, + "loss": 1.5945, + "theoretical_loss": 3.917937057915701, + "tokens_seen": 531300352 + }, + { + "epoch": 0.16, + "learning_rate": 0.0006779699097291877, + "loss": 1.505, + "theoretical_loss": 3.9178859258840566, + "tokens_seen": 531365888 + }, + { + "epoch": 0.16, + "learning_rate": 0.0006779538615847543, + "loss": 1.5208, + "theoretical_loss": 3.9178348019239264, + "tokens_seen": 531431424 + }, + { + "epoch": 0.16, + "learning_rate": 0.000677937813440321, + "loss": 1.5822, + "theoretical_loss": 3.91778368603304, + "tokens_seen": 531496960 + }, + { + "epoch": 0.16, + "learning_rate": 0.0006779217652958877, + "loss": 1.6843, + "theoretical_loss": 3.9177325782091303, + "tokens_seen": 531562496 + }, + { + "epoch": 0.16, + "learning_rate": 0.0006779057171514545, + "loss": 1.5797, + "theoretical_loss": 3.9176814784499303, + "tokens_seen": 531628032 + }, + { + "epoch": 0.16, + "learning_rate": 0.0006778896690070211, + "loss": 1.5549, + "theoretical_loss": 3.9176303867531717, + "tokens_seen": 531693568 + }, + { + "epoch": 0.16, + "learning_rate": 0.0006778736208625879, + "loss": 1.6315, + "theoretical_loss": 3.9175793031165895, + "tokens_seen": 531759104 + }, + { + "epoch": 0.16, + "learning_rate": 0.0006778575727181545, + "loss": 1.6388, + "theoretical_loss": 3.91752822753792, + "tokens_seen": 531824640 + }, + { + "epoch": 0.16, + "learning_rate": 0.0006778415245737213, + "loss": 1.4509, + "theoretical_loss": 3.917477160014898, + "tokens_seen": 531890176 + }, + { + "epoch": 0.16, + "learning_rate": 0.0006778254764292879, + "loss": 1.4993, + "theoretical_loss": 3.9174261005452617, + "tokens_seen": 531955712 + }, + { + "epoch": 0.16, + "learning_rate": 0.0006778094282848547, + "loss": 1.5689, + "theoretical_loss": 3.917375049126749, + "tokens_seen": 532021248 + }, + { + "epoch": 0.16, + "learning_rate": 0.0006777933801404213, + "loss": 1.4604, + "theoretical_loss": 3.9173240057570986, + "tokens_seen": 532086784 + }, + { + "epoch": 0.16, + "learning_rate": 0.0006777773319959881, + "loss": 1.4896, + "theoretical_loss": 3.91727297043405, + "tokens_seen": 532152320 + }, + { + "epoch": 0.16, + "learning_rate": 0.0006777612838515547, + "loss": 1.4285, + "theoretical_loss": 3.917221943155345, + "tokens_seen": 532217856 + }, + { + "epoch": 0.16, + "learning_rate": 0.0006777452357071214, + "loss": 1.4931, + "theoretical_loss": 3.9171709239187247, + "tokens_seen": 532283392 + }, + { + "epoch": 0.16, + "learning_rate": 0.0006777291875626881, + "loss": 1.7391, + "theoretical_loss": 3.9171199127219314, + "tokens_seen": 532348928 + }, + { + "epoch": 0.16, + "learning_rate": 0.0006777131394182548, + "loss": 1.5181, + "theoretical_loss": 3.917068909562709, + "tokens_seen": 532414464 + }, + { + "epoch": 0.16, + "learning_rate": 0.0006776970912738215, + "loss": 1.6097, + "theoretical_loss": 3.917017914438802, + "tokens_seen": 532480000 + }, + { + "epoch": 0.16, + "learning_rate": 0.0006776810431293882, + "loss": 1.6591, + "theoretical_loss": 3.9169669273479553, + "tokens_seen": 532545536 + }, + { + "epoch": 0.16, + "learning_rate": 0.0006776649949849549, + "loss": 1.5948, + "theoretical_loss": 3.916915948287916, + "tokens_seen": 532611072 + }, + { + "epoch": 0.16, + "learning_rate": 0.0006776489468405216, + "loss": 1.5404, + "theoretical_loss": 3.91686497725643, + "tokens_seen": 532676608 + }, + { + "epoch": 0.16, + "learning_rate": 0.0006776328986960883, + "loss": 1.455, + "theoretical_loss": 3.916814014251247, + "tokens_seen": 532742144 + }, + { + "epoch": 0.16, + "learning_rate": 0.000677616850551655, + "loss": 1.56, + "theoretical_loss": 3.9167630592701146, + "tokens_seen": 532807680 + }, + { + "epoch": 0.16, + "learning_rate": 0.0006776008024072216, + "loss": 1.5889, + "theoretical_loss": 3.9167121123107833, + "tokens_seen": 532873216 + }, + { + "epoch": 0.16, + "learning_rate": 0.0006775847542627884, + "loss": 1.5424, + "theoretical_loss": 3.9166611733710037, + "tokens_seen": 532938752 + }, + { + "epoch": 0.16, + "learning_rate": 0.000677568706118355, + "loss": 1.5763, + "theoretical_loss": 3.9166102424485274, + "tokens_seen": 533004288 + }, + { + "epoch": 0.16, + "learning_rate": 0.0006775526579739218, + "loss": 1.5372, + "theoretical_loss": 3.9165593195411077, + "tokens_seen": 533069824 + }, + { + "epoch": 0.16, + "learning_rate": 0.0006775366098294884, + "loss": 1.5788, + "theoretical_loss": 3.916508404646497, + "tokens_seen": 533135360 + }, + { + "epoch": 0.16, + "learning_rate": 0.0006775205616850552, + "loss": 1.5157, + "theoretical_loss": 3.9164574977624507, + "tokens_seen": 533200896 + }, + { + "epoch": 0.16, + "learning_rate": 0.0006775045135406218, + "loss": 1.5592, + "theoretical_loss": 3.916406598886724, + "tokens_seen": 533266432 + }, + { + "epoch": 0.16, + "learning_rate": 0.0006774884653961886, + "loss": 1.523, + "theoretical_loss": 3.916355708017073, + "tokens_seen": 533331968 + }, + { + "epoch": 0.16, + "learning_rate": 0.0006774724172517553, + "loss": 1.5401, + "theoretical_loss": 3.916304825151254, + "tokens_seen": 533397504 + }, + { + "epoch": 0.16, + "learning_rate": 0.000677456369107322, + "loss": 1.5395, + "theoretical_loss": 3.916253950287026, + "tokens_seen": 533463040 + }, + { + "epoch": 0.16, + "learning_rate": 0.0006774403209628887, + "loss": 1.534, + "theoretical_loss": 3.916203083422148, + "tokens_seen": 533528576 + }, + { + "epoch": 0.16, + "learning_rate": 0.0006774242728184554, + "loss": 1.5139, + "theoretical_loss": 3.91615222455438, + "tokens_seen": 533594112 + }, + { + "epoch": 0.16, + "learning_rate": 0.0006774082246740221, + "loss": 1.5919, + "theoretical_loss": 3.9161013736814816, + "tokens_seen": 533659648 + }, + { + "epoch": 0.16, + "learning_rate": 0.0006773921765295888, + "loss": 1.5343, + "theoretical_loss": 3.916050530801215, + "tokens_seen": 533725184 + }, + { + "epoch": 0.16, + "learning_rate": 0.0006773761283851555, + "loss": 1.5456, + "theoretical_loss": 3.915999695911343, + "tokens_seen": 533790720 + }, + { + "epoch": 0.16, + "learning_rate": 0.0006773600802407222, + "loss": 1.4871, + "theoretical_loss": 3.91594886900963, + "tokens_seen": 533856256 + }, + { + "epoch": 0.16, + "learning_rate": 0.0006773440320962889, + "loss": 1.6209, + "theoretical_loss": 3.9158980500938383, + "tokens_seen": 533921792 + }, + { + "epoch": 0.16, + "learning_rate": 0.0006773279839518556, + "loss": 1.5524, + "theoretical_loss": 3.9158472391617343, + "tokens_seen": 533987328 + }, + { + "epoch": 0.16, + "learning_rate": 0.0006773119358074223, + "loss": 1.5791, + "theoretical_loss": 3.9157964362110844, + "tokens_seen": 534052864 + }, + { + "epoch": 0.16, + "objective/train/docs_used": 200516, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.5898058414459229, + "objective/train/theoretical_loss": 3.9157456412396554, + "objective/train/tokens_used": 554578400, + "theoretical_loss": 3.9157456412396554, + "tokens_seen": 534118400 + }, + { + "epoch": 0.16, + "learning_rate": 0.000677295887662989, + "loss": 1.5254, + "theoretical_loss": 3.9157456412396554, + "tokens_seen": 534118400 + }, + { + "epoch": 0.16, + "learning_rate": 0.0006772798395185557, + "loss": 1.6496, + "theoretical_loss": 3.9156948542452144, + "tokens_seen": 534183936 + }, + { + "epoch": 0.16, + "learning_rate": 0.0006772637913741224, + "loss": 1.6357, + "theoretical_loss": 3.915644075225532, + "tokens_seen": 534249472 + }, + { + "epoch": 0.16, + "learning_rate": 0.0006772477432296891, + "loss": 1.5556, + "theoretical_loss": 3.915593304178376, + "tokens_seen": 534315008 + }, + { + "epoch": 0.16, + "learning_rate": 0.0006772316950852558, + "loss": 1.5706, + "theoretical_loss": 3.9155425411015186, + "tokens_seen": 534380544 + }, + { + "epoch": 0.16, + "learning_rate": 0.0006772156469408225, + "loss": 1.5954, + "theoretical_loss": 3.9154917859927307, + "tokens_seen": 534446080 + }, + { + "epoch": 0.16, + "learning_rate": 0.0006771995987963892, + "loss": 1.5782, + "theoretical_loss": 3.9154410388497842, + "tokens_seen": 534511616 + }, + { + "epoch": 0.16, + "learning_rate": 0.0006771835506519559, + "loss": 1.5577, + "theoretical_loss": 3.9153902996704533, + "tokens_seen": 534577152 + }, + { + "epoch": 0.16, + "learning_rate": 0.0006771675025075226, + "loss": 1.4792, + "theoretical_loss": 3.915339568452512, + "tokens_seen": 534642688 + }, + { + "epoch": 0.16, + "learning_rate": 0.0006771514543630893, + "loss": 1.5324, + "theoretical_loss": 3.9152888451937358, + "tokens_seen": 534708224 + }, + { + "epoch": 0.16, + "learning_rate": 0.000677135406218656, + "loss": 1.5354, + "theoretical_loss": 3.9152381298919, + "tokens_seen": 534773760 + }, + { + "epoch": 0.16, + "learning_rate": 0.0006771193580742227, + "loss": 1.6639, + "theoretical_loss": 3.915187422544782, + "tokens_seen": 534839296 + }, + { + "epoch": 0.16, + "learning_rate": 0.0006771033099297894, + "loss": 1.4665, + "theoretical_loss": 3.915136723150159, + "tokens_seen": 534904832 + }, + { + "epoch": 0.16, + "learning_rate": 0.0006770872617853561, + "loss": 1.5456, + "theoretical_loss": 3.91508603170581, + "tokens_seen": 534970368 + }, + { + "epoch": 0.16, + "learning_rate": 0.0006770712136409228, + "loss": 1.5257, + "theoretical_loss": 3.915035348209514, + "tokens_seen": 535035904 + }, + { + "epoch": 0.16, + "learning_rate": 0.0006770551654964895, + "loss": 1.509, + "theoretical_loss": 3.9149846726590534, + "tokens_seen": 535101440 + }, + { + "epoch": 0.16, + "learning_rate": 0.0006770391173520562, + "loss": 1.5801, + "theoretical_loss": 3.914934005052208, + "tokens_seen": 535166976 + }, + { + "epoch": 0.16, + "learning_rate": 0.0006770230692076229, + "loss": 1.6143, + "theoretical_loss": 3.91488334538676, + "tokens_seen": 535232512 + }, + { + "epoch": 0.16, + "learning_rate": 0.0006770070210631896, + "loss": 1.5007, + "theoretical_loss": 3.914832693660493, + "tokens_seen": 535298048 + }, + { + "epoch": 0.16, + "learning_rate": 0.0006769909729187563, + "loss": 1.4722, + "theoretical_loss": 3.9147820498711914, + "tokens_seen": 535363584 + }, + { + "epoch": 0.16, + "learning_rate": 0.000676974924774323, + "loss": 1.4692, + "theoretical_loss": 3.91473141401664, + "tokens_seen": 535429120 + }, + { + "epoch": 0.16, + "learning_rate": 0.0006769588766298897, + "loss": 1.5418, + "theoretical_loss": 3.9146807860946238, + "tokens_seen": 535494656 + }, + { + "epoch": 0.16, + "learning_rate": 0.0006769428284854564, + "loss": 1.5118, + "theoretical_loss": 3.91463016610293, + "tokens_seen": 535560192 + }, + { + "epoch": 0.16, + "learning_rate": 0.0006769267803410231, + "loss": 1.5715, + "theoretical_loss": 3.9145795540393467, + "tokens_seen": 535625728 + }, + { + "epoch": 0.16, + "learning_rate": 0.0006769107321965898, + "loss": 1.5394, + "theoretical_loss": 3.9145289499016616, + "tokens_seen": 535691264 + }, + { + "epoch": 0.16, + "learning_rate": 0.0006768946840521565, + "loss": 1.5566, + "theoretical_loss": 3.9144783536876644, + "tokens_seen": 535756800 + }, + { + "epoch": 0.16, + "learning_rate": 0.0006768786359077232, + "loss": 1.4737, + "theoretical_loss": 3.9144277653951454, + "tokens_seen": 535822336 + }, + { + "epoch": 0.16, + "learning_rate": 0.0006768625877632899, + "loss": 1.6297, + "theoretical_loss": 3.914377185021896, + "tokens_seen": 535887872 + }, + { + "epoch": 0.16, + "learning_rate": 0.0006768465396188566, + "loss": 1.6034, + "theoretical_loss": 3.914326612565708, + "tokens_seen": 535953408 + }, + { + "epoch": 0.16, + "learning_rate": 0.0006768304914744233, + "loss": 1.5764, + "theoretical_loss": 3.914276048024374, + "tokens_seen": 536018944 + }, + { + "epoch": 0.16, + "learning_rate": 0.00067681444332999, + "loss": 1.5338, + "theoretical_loss": 3.9142254913956886, + "tokens_seen": 536084480 + }, + { + "epoch": 0.16, + "learning_rate": 0.0006767983951855567, + "loss": 1.5276, + "theoretical_loss": 3.914174942677446, + "tokens_seen": 536150016 + }, + { + "epoch": 0.16, + "learning_rate": 0.0006767823470411234, + "loss": 1.4913, + "theoretical_loss": 3.9141244018674413, + "tokens_seen": 536215552 + }, + { + "epoch": 0.16, + "learning_rate": 0.0006767662988966901, + "loss": 1.5861, + "theoretical_loss": 3.9140738689634715, + "tokens_seen": 536281088 + }, + { + "epoch": 0.16, + "learning_rate": 0.0006767502507522568, + "loss": 1.5259, + "theoretical_loss": 3.914023343963334, + "tokens_seen": 536346624 + }, + { + "epoch": 0.16, + "learning_rate": 0.0006767342026078235, + "loss": 1.6253, + "theoretical_loss": 3.9139728268648266, + "tokens_seen": 536412160 + }, + { + "epoch": 0.16, + "learning_rate": 0.0006767181544633902, + "loss": 1.5262, + "theoretical_loss": 3.913922317665749, + "tokens_seen": 536477696 + }, + { + "epoch": 0.16, + "learning_rate": 0.000676702106318957, + "loss": 1.4616, + "theoretical_loss": 3.9138718163639012, + "tokens_seen": 536543232 + }, + { + "epoch": 0.16, + "learning_rate": 0.0006766860581745236, + "loss": 1.4446, + "theoretical_loss": 3.9138213229570837, + "tokens_seen": 536608768 + }, + { + "epoch": 0.16, + "learning_rate": 0.0006766700100300903, + "loss": 1.5373, + "theoretical_loss": 3.9137708374430984, + "tokens_seen": 536674304 + }, + { + "epoch": 0.16, + "learning_rate": 0.000676653961885657, + "loss": 1.6161, + "theoretical_loss": 3.913720359819748, + "tokens_seen": 536739840 + }, + { + "epoch": 0.16, + "learning_rate": 0.0006766379137412238, + "loss": 1.5189, + "theoretical_loss": 3.913669890084835, + "tokens_seen": 536805376 + }, + { + "epoch": 0.16, + "learning_rate": 0.0006766218655967904, + "loss": 1.5548, + "theoretical_loss": 3.9136194282361654, + "tokens_seen": 536870912 + }, + { + "epoch": 0.16, + "learning_rate": 0.0006766058174523572, + "loss": 1.4536, + "theoretical_loss": 3.913568974271543, + "tokens_seen": 536936448 + }, + { + "epoch": 0.16, + "learning_rate": 0.0006765897693079238, + "loss": 1.4997, + "theoretical_loss": 3.913518528188775, + "tokens_seen": 537001984 + }, + { + "epoch": 0.16, + "learning_rate": 0.0006765737211634906, + "loss": 1.4853, + "theoretical_loss": 3.9134680899856686, + "tokens_seen": 537067520 + }, + { + "epoch": 0.16, + "learning_rate": 0.0006765576730190572, + "loss": 1.514, + "theoretical_loss": 3.913417659660031, + "tokens_seen": 537133056 + }, + { + "epoch": 0.16, + "learning_rate": 0.000676541624874624, + "loss": 1.4818, + "theoretical_loss": 3.913367237209671, + "tokens_seen": 537198592 + }, + { + "epoch": 0.16, + "learning_rate": 0.0006765255767301906, + "loss": 1.5643, + "theoretical_loss": 3.9133168226323987, + "tokens_seen": 537264128 + }, + { + "epoch": 0.16, + "learning_rate": 0.0006765095285857574, + "loss": 1.5488, + "theoretical_loss": 3.913266415926024, + "tokens_seen": 537329664 + }, + { + "epoch": 0.16, + "objective/train/docs_used": 201641, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.4480133056640625, + "objective/train/theoretical_loss": 3.9132160170883585, + "objective/train/tokens_used": 557855200, + "theoretical_loss": 3.9132160170883585, + "tokens_seen": 537395200 + }, + { + "epoch": 0.16, + "learning_rate": 0.000676493480441324, + "loss": 1.4529, + "theoretical_loss": 3.9132160170883585, + "tokens_seen": 537395200 + }, + { + "epoch": 0.16, + "learning_rate": 0.0006764774322968908, + "loss": 1.5668, + "theoretical_loss": 3.913165626117215, + "tokens_seen": 537460736 + }, + { + "epoch": 0.16, + "learning_rate": 0.0006764613841524574, + "loss": 1.599, + "theoretical_loss": 3.913115243010407, + "tokens_seen": 537526272 + }, + { + "epoch": 0.16, + "learning_rate": 0.0006764453360080242, + "loss": 1.4591, + "theoretical_loss": 3.9130648677657476, + "tokens_seen": 537591808 + }, + { + "epoch": 0.16, + "learning_rate": 0.0006764292878635908, + "loss": 1.4956, + "theoretical_loss": 3.9130145003810517, + "tokens_seen": 537657344 + }, + { + "epoch": 0.16, + "learning_rate": 0.0006764132397191576, + "loss": 1.5913, + "theoretical_loss": 3.912964140854135, + "tokens_seen": 537722880 + }, + { + "epoch": 0.16, + "learning_rate": 0.0006763971915747242, + "loss": 1.5883, + "theoretical_loss": 3.9129137891828147, + "tokens_seen": 537788416 + }, + { + "epoch": 0.16, + "learning_rate": 0.000676381143430291, + "loss": 1.5483, + "theoretical_loss": 3.9128634453649083, + "tokens_seen": 537853952 + }, + { + "epoch": 0.16, + "learning_rate": 0.0006763650952858576, + "loss": 1.5843, + "theoretical_loss": 3.9128131093982343, + "tokens_seen": 537919488 + }, + { + "epoch": 0.16, + "learning_rate": 0.0006763490471414244, + "loss": 1.6088, + "theoretical_loss": 3.9127627812806116, + "tokens_seen": 537985024 + }, + { + "epoch": 0.16, + "learning_rate": 0.000676332998996991, + "loss": 1.5599, + "theoretical_loss": 3.91271246100986, + "tokens_seen": 538050560 + }, + { + "epoch": 0.16, + "learning_rate": 0.0006763169508525577, + "loss": 1.574, + "theoretical_loss": 3.9126621485838013, + "tokens_seen": 538116096 + }, + { + "epoch": 0.16, + "learning_rate": 0.0006763009027081244, + "loss": 1.433, + "theoretical_loss": 3.9126118440002573, + "tokens_seen": 538181632 + }, + { + "epoch": 0.16, + "learning_rate": 0.0006762848545636911, + "loss": 1.5506, + "theoretical_loss": 3.9125615472570505, + "tokens_seen": 538247168 + }, + { + "epoch": 0.16, + "learning_rate": 0.0006762688064192578, + "loss": 1.644, + "theoretical_loss": 3.912511258352004, + "tokens_seen": 538312704 + }, + { + "epoch": 0.16, + "learning_rate": 0.0006762527582748245, + "loss": 1.6297, + "theoretical_loss": 3.9124609772829437, + "tokens_seen": 538378240 + }, + { + "epoch": 0.16, + "learning_rate": 0.0006762367101303912, + "loss": 1.4689, + "theoretical_loss": 3.9124107040476934, + "tokens_seen": 538443776 + }, + { + "epoch": 0.16, + "learning_rate": 0.0006762206619859579, + "loss": 1.5097, + "theoretical_loss": 3.91236043864408, + "tokens_seen": 538509312 + }, + { + "epoch": 0.16, + "learning_rate": 0.0006762046138415246, + "loss": 1.5196, + "theoretical_loss": 3.912310181069931, + "tokens_seen": 538574848 + }, + { + "epoch": 0.16, + "learning_rate": 0.0006761885656970913, + "loss": 1.689, + "theoretical_loss": 3.9122599313230735, + "tokens_seen": 538640384 + }, + { + "epoch": 0.16, + "learning_rate": 0.000676172517552658, + "loss": 1.505, + "theoretical_loss": 3.9122096894013376, + "tokens_seen": 538705920 + }, + { + "epoch": 0.16, + "learning_rate": 0.0006761564694082247, + "loss": 1.6033, + "theoretical_loss": 3.9121594553025516, + "tokens_seen": 538771456 + }, + { + "epoch": 0.16, + "learning_rate": 0.0006761404212637914, + "loss": 1.5209, + "theoretical_loss": 3.912109229024547, + "tokens_seen": 538836992 + }, + { + "epoch": 0.16, + "learning_rate": 0.0006761243731193581, + "loss": 1.5181, + "theoretical_loss": 3.9120590105651547, + "tokens_seen": 538902528 + }, + { + "epoch": 0.16, + "learning_rate": 0.0006761083249749248, + "loss": 1.5821, + "theoretical_loss": 3.9120087999222073, + "tokens_seen": 538968064 + }, + { + "epoch": 0.16, + "learning_rate": 0.0006760922768304915, + "loss": 1.4207, + "theoretical_loss": 3.9119585970935375, + "tokens_seen": 539033600 + }, + { + "epoch": 0.16, + "learning_rate": 0.0006760762286860582, + "loss": 1.4642, + "theoretical_loss": 3.91190840207698, + "tokens_seen": 539099136 + }, + { + "epoch": 0.16, + "learning_rate": 0.0006760601805416249, + "loss": 1.5109, + "theoretical_loss": 3.9118582148703696, + "tokens_seen": 539164672 + }, + { + "epoch": 0.16, + "learning_rate": 0.0006760441323971916, + "loss": 1.5964, + "theoretical_loss": 3.9118080354715414, + "tokens_seen": 539230208 + }, + { + "epoch": 0.16, + "learning_rate": 0.0006760280842527583, + "loss": 1.6208, + "theoretical_loss": 3.911757863878333, + "tokens_seen": 539295744 + }, + { + "epoch": 0.16, + "learning_rate": 0.000676012036108325, + "loss": 1.6713, + "theoretical_loss": 3.9117077000885807, + "tokens_seen": 539361280 + }, + { + "epoch": 0.16, + "learning_rate": 0.0006759959879638917, + "loss": 1.5399, + "theoretical_loss": 3.911657544100124, + "tokens_seen": 539426816 + }, + { + "epoch": 0.16, + "learning_rate": 0.0006759799398194584, + "loss": 1.464, + "theoretical_loss": 3.911607395910801, + "tokens_seen": 539492352 + }, + { + "epoch": 0.16, + "learning_rate": 0.0006759638916750251, + "loss": 1.5493, + "theoretical_loss": 3.911557255518453, + "tokens_seen": 539557888 + }, + { + "epoch": 0.16, + "learning_rate": 0.0006759478435305918, + "loss": 1.5766, + "theoretical_loss": 3.9115071229209195, + "tokens_seen": 539623424 + }, + { + "epoch": 0.16, + "learning_rate": 0.0006759317953861585, + "loss": 1.5732, + "theoretical_loss": 3.9114569981160434, + "tokens_seen": 539688960 + }, + { + "epoch": 0.16, + "learning_rate": 0.0006759157472417252, + "loss": 1.5323, + "theoretical_loss": 3.9114068811016662, + "tokens_seen": 539754496 + }, + { + "epoch": 0.16, + "learning_rate": 0.0006758996990972919, + "loss": 1.6215, + "theoretical_loss": 3.9113567718756332, + "tokens_seen": 539820032 + }, + { + "epoch": 0.16, + "learning_rate": 0.0006758836509528586, + "loss": 1.5503, + "theoretical_loss": 3.9113066704357866, + "tokens_seen": 539885568 + }, + { + "epoch": 0.16, + "learning_rate": 0.0006758676028084253, + "loss": 1.6223, + "theoretical_loss": 3.9112565767799734, + "tokens_seen": 539951104 + }, + { + "epoch": 0.16, + "learning_rate": 0.000675851554663992, + "loss": 1.6077, + "theoretical_loss": 3.9112064909060393, + "tokens_seen": 540016640 + }, + { + "epoch": 0.16, + "learning_rate": 0.0006758355065195587, + "loss": 1.5549, + "theoretical_loss": 3.91115641281183, + "tokens_seen": 540082176 + }, + { + "epoch": 0.16, + "learning_rate": 0.0006758194583751254, + "loss": 1.4746, + "theoretical_loss": 3.911106342495195, + "tokens_seen": 540147712 + }, + { + "epoch": 0.16, + "learning_rate": 0.0006758034102306921, + "loss": 1.7422, + "theoretical_loss": 3.911056279953981, + "tokens_seen": 540213248 + }, + { + "epoch": 0.16, + "learning_rate": 0.0006757873620862588, + "loss": 1.4856, + "theoretical_loss": 3.91100622518604, + "tokens_seen": 540278784 + }, + { + "epoch": 0.16, + "learning_rate": 0.0006757713139418255, + "loss": 1.656, + "theoretical_loss": 3.9109561781892204, + "tokens_seen": 540344320 + }, + { + "epoch": 0.16, + "learning_rate": 0.0006757552657973922, + "loss": 1.5785, + "theoretical_loss": 3.9109061389613737, + "tokens_seen": 540409856 + }, + { + "epoch": 0.16, + "learning_rate": 0.0006757392176529589, + "loss": 1.4879, + "theoretical_loss": 3.9108561075003525, + "tokens_seen": 540475392 + }, + { + "epoch": 0.16, + "learning_rate": 0.0006757231695085256, + "loss": 1.639, + "theoretical_loss": 3.9108060838040095, + "tokens_seen": 540540928 + }, + { + "epoch": 0.16, + "learning_rate": 0.0006757071213640923, + "loss": 1.4798, + "theoretical_loss": 3.9107560678701985, + "tokens_seen": 540606464 + }, + { + "epoch": 0.16, + "objective/train/docs_used": 202818, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.4981238842010498, + "objective/train/theoretical_loss": 3.910706059696774, + "objective/train/tokens_used": 561132000, + "theoretical_loss": 3.910706059696774, + "tokens_seen": 540672000 + }, + { + "epoch": 0.16, + "learning_rate": 0.000675691073219659, + "loss": 1.5368, + "theoretical_loss": 3.910706059696774, + "tokens_seen": 540672000 + }, + { + "epoch": 0.16, + "learning_rate": 0.0006756750250752257, + "loss": 1.6161, + "theoretical_loss": 3.910656059281591, + "tokens_seen": 540737536 + }, + { + "epoch": 0.16, + "learning_rate": 0.0006756589769307924, + "loss": 1.5116, + "theoretical_loss": 3.910606066622507, + "tokens_seen": 540803072 + }, + { + "epoch": 0.16, + "learning_rate": 0.0006756429287863591, + "loss": 1.5183, + "theoretical_loss": 3.9105560817173783, + "tokens_seen": 540868608 + }, + { + "epoch": 0.16, + "learning_rate": 0.0006756268806419258, + "loss": 1.5294, + "theoretical_loss": 3.910506104564063, + "tokens_seen": 540934144 + }, + { + "epoch": 0.16, + "learning_rate": 0.0006756108324974925, + "loss": 1.5071, + "theoretical_loss": 3.9104561351604206, + "tokens_seen": 540999680 + }, + { + "epoch": 0.16, + "learning_rate": 0.0006755947843530592, + "loss": 1.4601, + "theoretical_loss": 3.9104061735043105, + "tokens_seen": 541065216 + }, + { + "epoch": 0.16, + "learning_rate": 0.0006755787362086259, + "loss": 1.5059, + "theoretical_loss": 3.9103562195935933, + "tokens_seen": 541130752 + }, + { + "epoch": 0.16, + "learning_rate": 0.0006755626880641926, + "loss": 1.667, + "theoretical_loss": 3.91030627342613, + "tokens_seen": 541196288 + }, + { + "epoch": 0.16, + "learning_rate": 0.0006755466399197593, + "loss": 1.6614, + "theoretical_loss": 3.910256334999783, + "tokens_seen": 541261824 + }, + { + "epoch": 0.16, + "learning_rate": 0.000675530591775326, + "loss": 1.6021, + "theoretical_loss": 3.910206404312416, + "tokens_seen": 541327360 + }, + { + "epoch": 0.16, + "learning_rate": 0.0006755145436308927, + "loss": 1.5746, + "theoretical_loss": 3.9101564813618923, + "tokens_seen": 541392896 + }, + { + "epoch": 0.16, + "learning_rate": 0.0006754984954864594, + "loss": 1.5643, + "theoretical_loss": 3.9101065661460774, + "tokens_seen": 541458432 + }, + { + "epoch": 0.16, + "learning_rate": 0.0006754824473420261, + "loss": 1.5521, + "theoretical_loss": 3.9100566586628362, + "tokens_seen": 541523968 + }, + { + "epoch": 0.16, + "learning_rate": 0.0006754663991975928, + "loss": 1.6441, + "theoretical_loss": 3.9100067589100362, + "tokens_seen": 541589504 + }, + { + "epoch": 0.16, + "learning_rate": 0.0006754503510531595, + "loss": 1.6462, + "theoretical_loss": 3.909956866885544, + "tokens_seen": 541655040 + }, + { + "epoch": 0.16, + "learning_rate": 0.0006754343029087262, + "loss": 1.5425, + "theoretical_loss": 3.909906982587228, + "tokens_seen": 541720576 + }, + { + "epoch": 0.16, + "learning_rate": 0.0006754182547642929, + "loss": 1.5708, + "theoretical_loss": 3.909857106012957, + "tokens_seen": 541786112 + }, + { + "epoch": 0.16, + "learning_rate": 0.0006754022066198596, + "loss": 1.4538, + "theoretical_loss": 3.9098072371606016, + "tokens_seen": 541851648 + }, + { + "epoch": 0.16, + "learning_rate": 0.0006753861584754263, + "loss": 1.6832, + "theoretical_loss": 3.9097573760280326, + "tokens_seen": 541917184 + }, + { + "epoch": 0.16, + "learning_rate": 0.000675370110330993, + "loss": 1.586, + "theoretical_loss": 3.90970752261312, + "tokens_seen": 541982720 + }, + { + "epoch": 0.16, + "learning_rate": 0.0006753540621865597, + "loss": 1.5737, + "theoretical_loss": 3.9096576769137386, + "tokens_seen": 542048256 + }, + { + "epoch": 0.16, + "learning_rate": 0.0006753380140421264, + "loss": 1.4684, + "theoretical_loss": 3.90960783892776, + "tokens_seen": 542113792 + }, + { + "epoch": 0.16, + "learning_rate": 0.0006753219658976931, + "loss": 1.5201, + "theoretical_loss": 3.9095580086530592, + "tokens_seen": 542179328 + }, + { + "epoch": 0.16, + "learning_rate": 0.0006753059177532598, + "loss": 1.5815, + "theoretical_loss": 3.90950818608751, + "tokens_seen": 542244864 + }, + { + "epoch": 0.16, + "learning_rate": 0.0006752898696088265, + "loss": 1.5549, + "theoretical_loss": 3.9094583712289896, + "tokens_seen": 542310400 + }, + { + "epoch": 0.16, + "learning_rate": 0.0006752738214643933, + "loss": 1.5385, + "theoretical_loss": 3.909408564075374, + "tokens_seen": 542375936 + }, + { + "epoch": 0.16, + "learning_rate": 0.0006752577733199599, + "loss": 1.5081, + "theoretical_loss": 3.9093587646245407, + "tokens_seen": 542441472 + }, + { + "epoch": 0.16, + "learning_rate": 0.0006752417251755267, + "loss": 1.635, + "theoretical_loss": 3.9093089728743684, + "tokens_seen": 542507008 + }, + { + "epoch": 0.16, + "learning_rate": 0.0006752256770310933, + "loss": 1.5879, + "theoretical_loss": 3.9092591888227357, + "tokens_seen": 542572544 + }, + { + "epoch": 0.16, + "learning_rate": 0.00067520962888666, + "loss": 1.5684, + "theoretical_loss": 3.909209412467523, + "tokens_seen": 542638080 + }, + { + "epoch": 0.16, + "learning_rate": 0.0006751935807422267, + "loss": 1.6029, + "theoretical_loss": 3.9091596438066114, + "tokens_seen": 542703616 + }, + { + "epoch": 0.16, + "learning_rate": 0.0006751775325977935, + "loss": 1.5953, + "theoretical_loss": 3.909109882837882, + "tokens_seen": 542769152 + }, + { + "epoch": 0.16, + "learning_rate": 0.0006751614844533601, + "loss": 1.5626, + "theoretical_loss": 3.9090601295592178, + "tokens_seen": 542834688 + }, + { + "epoch": 0.16, + "learning_rate": 0.0006751454363089269, + "loss": 1.4197, + "theoretical_loss": 3.9090103839685013, + "tokens_seen": 542900224 + }, + { + "epoch": 0.16, + "learning_rate": 0.0006751293881644935, + "loss": 1.7414, + "theoretical_loss": 3.908960646063618, + "tokens_seen": 542965760 + }, + { + "epoch": 0.16, + "learning_rate": 0.0006751133400200603, + "loss": 1.5353, + "theoretical_loss": 3.908910915842452, + "tokens_seen": 543031296 + }, + { + "epoch": 0.16, + "learning_rate": 0.0006750972918756269, + "loss": 1.4823, + "theoretical_loss": 3.90886119330289, + "tokens_seen": 543096832 + }, + { + "epoch": 0.16, + "learning_rate": 0.0006750812437311937, + "loss": 1.4661, + "theoretical_loss": 3.908811478442818, + "tokens_seen": 543162368 + }, + { + "epoch": 0.16, + "learning_rate": 0.0006750651955867603, + "loss": 1.5551, + "theoretical_loss": 3.908761771260124, + "tokens_seen": 543227904 + }, + { + "epoch": 0.16, + "learning_rate": 0.0006750491474423271, + "loss": 1.4568, + "theoretical_loss": 3.908712071752696, + "tokens_seen": 543293440 + }, + { + "epoch": 0.16, + "learning_rate": 0.0006750330992978937, + "loss": 1.5933, + "theoretical_loss": 3.908662379918424, + "tokens_seen": 543358976 + }, + { + "epoch": 0.16, + "learning_rate": 0.0006750170511534605, + "loss": 1.5738, + "theoretical_loss": 3.908612695755197, + "tokens_seen": 543424512 + }, + { + "epoch": 0.16, + "learning_rate": 0.0006750010030090271, + "loss": 1.5688, + "theoretical_loss": 3.9085630192609067, + "tokens_seen": 543490048 + }, + { + "epoch": 0.16, + "learning_rate": 0.0006749849548645939, + "loss": 1.5768, + "theoretical_loss": 3.908513350433444, + "tokens_seen": 543555584 + }, + { + "epoch": 0.16, + "learning_rate": 0.0006749689067201605, + "loss": 1.5758, + "theoretical_loss": 3.908463689270703, + "tokens_seen": 543621120 + }, + { + "epoch": 0.16, + "learning_rate": 0.0006749528585757273, + "loss": 1.528, + "theoretical_loss": 3.9084140357705754, + "tokens_seen": 543686656 + }, + { + "epoch": 0.16, + "learning_rate": 0.0006749368104312939, + "loss": 1.5836, + "theoretical_loss": 3.9083643899309566, + "tokens_seen": 543752192 + }, + { + "epoch": 0.16, + "learning_rate": 0.0006749207622868606, + "loss": 1.5989, + "theoretical_loss": 3.9083147517497414, + "tokens_seen": 543817728 + }, + { + "epoch": 0.16, + "learning_rate": 0.0006749047141424273, + "loss": 1.5109, + "theoretical_loss": 3.908265121224825, + "tokens_seen": 543883264 + }, + { + "epoch": 0.16, + "objective/train/docs_used": 204061, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.5775312185287476, + "objective/train/theoretical_loss": 3.908215498354105, + "objective/train/tokens_used": 564408800, + "theoretical_loss": 3.908215498354105, + "tokens_seen": 543948800 + }, + { + "epoch": 0.16, + "learning_rate": 0.000674888665997994, + "loss": 1.5972, + "theoretical_loss": 3.908215498354105, + "tokens_seen": 543948800 + }, + { + "epoch": 0.16, + "learning_rate": 0.0006748726178535607, + "loss": 1.547, + "theoretical_loss": 3.9081658831354784, + "tokens_seen": 544014336 + }, + { + "epoch": 0.16, + "learning_rate": 0.0006748565697091274, + "loss": 1.5506, + "theoretical_loss": 3.908116275566844, + "tokens_seen": 544079872 + }, + { + "epoch": 0.16, + "learning_rate": 0.0006748405215646941, + "loss": 1.5396, + "theoretical_loss": 3.9080666756461007, + "tokens_seen": 544145408 + }, + { + "epoch": 0.16, + "learning_rate": 0.0006748244734202608, + "loss": 1.5912, + "theoretical_loss": 3.908017083371149, + "tokens_seen": 544210944 + }, + { + "epoch": 0.16, + "learning_rate": 0.0006748084252758275, + "loss": 1.5917, + "theoretical_loss": 3.9079674987398896, + "tokens_seen": 544276480 + }, + { + "epoch": 0.16, + "learning_rate": 0.0006747923771313942, + "loss": 1.5907, + "theoretical_loss": 3.907917921750224, + "tokens_seen": 544342016 + }, + { + "epoch": 0.16, + "learning_rate": 0.0006747763289869609, + "loss": 1.3856, + "theoretical_loss": 3.9078683524000546, + "tokens_seen": 544407552 + }, + { + "epoch": 0.16, + "learning_rate": 0.0006747602808425276, + "loss": 1.6137, + "theoretical_loss": 3.907818790687285, + "tokens_seen": 544473088 + }, + { + "epoch": 0.17, + "learning_rate": 0.0006747442326980943, + "loss": 1.5546, + "theoretical_loss": 3.90776923660982, + "tokens_seen": 544538624 + }, + { + "epoch": 0.17, + "learning_rate": 0.000674728184553661, + "loss": 1.4969, + "theoretical_loss": 3.907719690165564, + "tokens_seen": 544604160 + }, + { + "epoch": 0.17, + "learning_rate": 0.0006747121364092277, + "loss": 1.5025, + "theoretical_loss": 3.9076701513524226, + "tokens_seen": 544669696 + }, + { + "epoch": 0.17, + "learning_rate": 0.0006746960882647944, + "loss": 1.5555, + "theoretical_loss": 3.9076206201683026, + "tokens_seen": 544735232 + }, + { + "epoch": 0.17, + "learning_rate": 0.0006746800401203611, + "loss": 1.5128, + "theoretical_loss": 3.907571096611112, + "tokens_seen": 544800768 + }, + { + "epoch": 0.17, + "learning_rate": 0.0006746639919759278, + "loss": 1.5509, + "theoretical_loss": 3.9075215806787584, + "tokens_seen": 544866304 + }, + { + "epoch": 0.17, + "learning_rate": 0.0006746479438314945, + "loss": 1.5621, + "theoretical_loss": 3.907472072369152, + "tokens_seen": 544931840 + }, + { + "epoch": 0.17, + "learning_rate": 0.0006746318956870612, + "loss": 1.7077, + "theoretical_loss": 3.907422571680202, + "tokens_seen": 544997376 + }, + { + "epoch": 0.17, + "learning_rate": 0.0006746158475426279, + "loss": 1.5596, + "theoretical_loss": 3.907373078609819, + "tokens_seen": 545062912 + }, + { + "epoch": 0.17, + "learning_rate": 0.0006745997993981946, + "loss": 1.6374, + "theoretical_loss": 3.9073235931559154, + "tokens_seen": 545128448 + }, + { + "epoch": 0.17, + "learning_rate": 0.0006745837512537613, + "loss": 1.5762, + "theoretical_loss": 3.907274115316403, + "tokens_seen": 545193984 + }, + { + "epoch": 0.17, + "learning_rate": 0.000674567703109328, + "loss": 1.4866, + "theoretical_loss": 3.907224645089195, + "tokens_seen": 545259520 + }, + { + "epoch": 0.17, + "learning_rate": 0.0006745516549648947, + "loss": 1.6781, + "theoretical_loss": 3.907175182472206, + "tokens_seen": 545325056 + }, + { + "epoch": 0.17, + "learning_rate": 0.0006745356068204614, + "loss": 1.5158, + "theoretical_loss": 3.9071257274633506, + "tokens_seen": 545390592 + }, + { + "epoch": 0.17, + "learning_rate": 0.0006745195586760281, + "loss": 1.6639, + "theoretical_loss": 3.9070762800605454, + "tokens_seen": 545456128 + }, + { + "epoch": 0.17, + "learning_rate": 0.0006745035105315948, + "loss": 1.6046, + "theoretical_loss": 3.9070268402617057, + "tokens_seen": 545521664 + }, + { + "epoch": 0.17, + "learning_rate": 0.0006744874623871615, + "loss": 1.5814, + "theoretical_loss": 3.9069774080647495, + "tokens_seen": 545587200 + }, + { + "epoch": 0.17, + "learning_rate": 0.0006744714142427282, + "loss": 1.6504, + "theoretical_loss": 3.9069279834675945, + "tokens_seen": 545652736 + }, + { + "epoch": 0.17, + "learning_rate": 0.0006744553660982949, + "loss": 1.5998, + "theoretical_loss": 3.9068785664681602, + "tokens_seen": 545718272 + }, + { + "epoch": 0.17, + "learning_rate": 0.0006744393179538616, + "loss": 1.5472, + "theoretical_loss": 3.906829157064367, + "tokens_seen": 545783808 + }, + { + "epoch": 0.17, + "learning_rate": 0.0006744232698094283, + "loss": 1.5551, + "theoretical_loss": 3.906779755254134, + "tokens_seen": 545849344 + }, + { + "epoch": 0.17, + "learning_rate": 0.000674407221664995, + "loss": 1.5628, + "theoretical_loss": 3.9067303610353847, + "tokens_seen": 545914880 + }, + { + "epoch": 0.17, + "learning_rate": 0.0006743911735205617, + "loss": 1.6008, + "theoretical_loss": 3.9066809744060396, + "tokens_seen": 545980416 + }, + { + "epoch": 0.17, + "learning_rate": 0.0006743751253761284, + "loss": 1.5355, + "theoretical_loss": 3.9066315953640225, + "tokens_seen": 546045952 + }, + { + "epoch": 0.17, + "learning_rate": 0.0006743590772316951, + "loss": 1.613, + "theoretical_loss": 3.9065822239072574, + "tokens_seen": 546111488 + }, + { + "epoch": 0.17, + "learning_rate": 0.0006743430290872618, + "loss": 1.6047, + "theoretical_loss": 3.9065328600336695, + "tokens_seen": 546177024 + }, + { + "epoch": 0.17, + "learning_rate": 0.0006743269809428285, + "loss": 1.5305, + "theoretical_loss": 3.9064835037411836, + "tokens_seen": 546242560 + }, + { + "epoch": 0.17, + "learning_rate": 0.0006743109327983952, + "loss": 1.5447, + "theoretical_loss": 3.9064341550277266, + "tokens_seen": 546308096 + }, + { + "epoch": 0.17, + "learning_rate": 0.0006742948846539619, + "loss": 1.6237, + "theoretical_loss": 3.9063848138912247, + "tokens_seen": 546373632 + }, + { + "epoch": 0.17, + "learning_rate": 0.0006742788365095286, + "loss": 1.5304, + "theoretical_loss": 3.9063354803296075, + "tokens_seen": 546439168 + }, + { + "epoch": 0.17, + "learning_rate": 0.0006742627883650953, + "loss": 1.5486, + "theoretical_loss": 3.906286154340802, + "tokens_seen": 546504704 + }, + { + "epoch": 0.17, + "learning_rate": 0.000674246740220662, + "loss": 1.589, + "theoretical_loss": 3.9062368359227397, + "tokens_seen": 546570240 + }, + { + "epoch": 0.17, + "learning_rate": 0.0006742306920762287, + "loss": 1.5402, + "theoretical_loss": 3.90618752507335, + "tokens_seen": 546635776 + }, + { + "epoch": 0.17, + "learning_rate": 0.0006742146439317954, + "loss": 1.5989, + "theoretical_loss": 3.906138221790565, + "tokens_seen": 546701312 + }, + { + "epoch": 0.17, + "learning_rate": 0.0006741985957873621, + "loss": 1.5083, + "theoretical_loss": 3.9060889260723153, + "tokens_seen": 546766848 + }, + { + "epoch": 0.17, + "learning_rate": 0.0006741825476429288, + "loss": 1.5911, + "theoretical_loss": 3.906039637916535, + "tokens_seen": 546832384 + }, + { + "epoch": 0.17, + "learning_rate": 0.0006741664994984955, + "loss": 1.5997, + "theoretical_loss": 3.9059903573211576, + "tokens_seen": 546897920 + }, + { + "epoch": 0.17, + "learning_rate": 0.0006741504513540622, + "loss": 1.5676, + "theoretical_loss": 3.905941084284117, + "tokens_seen": 546963456 + }, + { + "epoch": 0.17, + "learning_rate": 0.0006741344032096289, + "loss": 1.3596, + "theoretical_loss": 3.9058918188033496, + "tokens_seen": 547028992 + }, + { + "epoch": 0.17, + "learning_rate": 0.0006741183550651956, + "loss": 1.6146, + "theoretical_loss": 3.905842560876791, + "tokens_seen": 547094528 + }, + { + "epoch": 0.17, + "learning_rate": 0.0006741023069207623, + "loss": 1.5868, + "theoretical_loss": 3.9057933105023777, + "tokens_seen": 547160064 + }, + { + "epoch": 0.17, + "objective/train/docs_used": 204823, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.4403038024902344, + "objective/train/theoretical_loss": 3.9057440676780484, + "objective/train/tokens_used": 567685600, + "theoretical_loss": 3.9057440676780484, + "tokens_seen": 547225600 + }, + { + "epoch": 0.17, + "learning_rate": 0.000674086258776329, + "loss": 1.5045, + "theoretical_loss": 3.9057440676780484, + "tokens_seen": 547225600 + }, + { + "epoch": 0.17, + "learning_rate": 0.0006740702106318957, + "loss": 1.531, + "theoretical_loss": 3.905694832401741, + "tokens_seen": 547291136 + }, + { + "epoch": 0.17, + "learning_rate": 0.0006740541624874624, + "loss": 1.6313, + "theoretical_loss": 3.905645604671395, + "tokens_seen": 547356672 + }, + { + "epoch": 0.17, + "learning_rate": 0.0006740381143430291, + "loss": 1.5193, + "theoretical_loss": 3.9055963844849506, + "tokens_seen": 547422208 + }, + { + "epoch": 0.17, + "learning_rate": 0.0006740220661985958, + "loss": 1.46, + "theoretical_loss": 3.9055471718403494, + "tokens_seen": 547487744 + }, + { + "epoch": 0.17, + "learning_rate": 0.0006740060180541625, + "loss": 1.5352, + "theoretical_loss": 3.905497966735532, + "tokens_seen": 547553280 + }, + { + "epoch": 0.17, + "learning_rate": 0.0006739899699097292, + "loss": 1.5055, + "theoretical_loss": 3.9054487691684425, + "tokens_seen": 547618816 + }, + { + "epoch": 0.17, + "learning_rate": 0.000673973921765296, + "loss": 1.5125, + "theoretical_loss": 3.905399579137023, + "tokens_seen": 547684352 + }, + { + "epoch": 0.17, + "learning_rate": 0.0006739578736208626, + "loss": 1.5323, + "theoretical_loss": 3.9053503966392187, + "tokens_seen": 547749888 + }, + { + "epoch": 0.17, + "learning_rate": 0.0006739418254764293, + "loss": 1.5436, + "theoretical_loss": 3.905301221672974, + "tokens_seen": 547815424 + }, + { + "epoch": 0.17, + "learning_rate": 0.000673925777331996, + "loss": 1.5257, + "theoretical_loss": 3.9052520542362346, + "tokens_seen": 547880960 + }, + { + "epoch": 0.17, + "learning_rate": 0.0006739097291875627, + "loss": 1.6165, + "theoretical_loss": 3.9052028943269477, + "tokens_seen": 547946496 + }, + { + "epoch": 0.17, + "learning_rate": 0.0006738936810431294, + "loss": 1.4825, + "theoretical_loss": 3.905153741943061, + "tokens_seen": 548012032 + }, + { + "epoch": 0.17, + "learning_rate": 0.0006738776328986962, + "loss": 1.5636, + "theoretical_loss": 3.905104597082522, + "tokens_seen": 548077568 + }, + { + "epoch": 0.17, + "learning_rate": 0.0006738615847542629, + "loss": 1.5847, + "theoretical_loss": 3.9050554597432807, + "tokens_seen": 548143104 + }, + { + "epoch": 0.17, + "learning_rate": 0.0006738455366098296, + "loss": 1.5891, + "theoretical_loss": 3.905006329923286, + "tokens_seen": 548208640 + }, + { + "epoch": 0.17, + "learning_rate": 0.0006738294884653963, + "loss": 1.613, + "theoretical_loss": 3.904957207620489, + "tokens_seen": 548274176 + }, + { + "epoch": 0.17, + "learning_rate": 0.000673813440320963, + "loss": 1.5355, + "theoretical_loss": 3.904908092832841, + "tokens_seen": 548339712 + }, + { + "epoch": 0.17, + "learning_rate": 0.0006737973921765297, + "loss": 1.5118, + "theoretical_loss": 3.904858985558295, + "tokens_seen": 548405248 + }, + { + "epoch": 0.17, + "learning_rate": 0.0006737813440320964, + "loss": 1.6144, + "theoretical_loss": 3.904809885794803, + "tokens_seen": 548470784 + }, + { + "epoch": 0.17, + "learning_rate": 0.0006737652958876631, + "loss": 1.4531, + "theoretical_loss": 3.9047607935403197, + "tokens_seen": 548536320 + }, + { + "epoch": 0.17, + "learning_rate": 0.0006737492477432298, + "loss": 1.4702, + "theoretical_loss": 3.9047117087927994, + "tokens_seen": 548601856 + }, + { + "epoch": 0.17, + "learning_rate": 0.0006737331995987965, + "loss": 1.398, + "theoretical_loss": 3.904662631550198, + "tokens_seen": 548667392 + }, + { + "epoch": 0.17, + "learning_rate": 0.0006737171514543632, + "loss": 1.3923, + "theoretical_loss": 3.904613561810471, + "tokens_seen": 548732928 + }, + { + "epoch": 0.17, + "learning_rate": 0.0006737011033099299, + "loss": 1.5505, + "theoretical_loss": 3.9045644995715767, + "tokens_seen": 548798464 + }, + { + "epoch": 0.17, + "learning_rate": 0.0006736850551654966, + "loss": 1.6466, + "theoretical_loss": 3.9045154448314716, + "tokens_seen": 548864000 + }, + { + "epoch": 0.17, + "learning_rate": 0.0006736690070210632, + "loss": 1.5787, + "theoretical_loss": 3.9044663975881155, + "tokens_seen": 548929536 + }, + { + "epoch": 0.17, + "learning_rate": 0.00067365295887663, + "loss": 1.5414, + "theoretical_loss": 3.9044173578394674, + "tokens_seen": 548995072 + }, + { + "epoch": 0.17, + "learning_rate": 0.0006736369107321966, + "loss": 1.5393, + "theoretical_loss": 3.904368325583488, + "tokens_seen": 549060608 + }, + { + "epoch": 0.17, + "learning_rate": 0.0006736208625877634, + "loss": 1.5919, + "theoretical_loss": 3.9043193008181376, + "tokens_seen": 549126144 + }, + { + "epoch": 0.17, + "learning_rate": 0.00067360481444333, + "loss": 1.5234, + "theoretical_loss": 3.9042702835413787, + "tokens_seen": 549191680 + }, + { + "epoch": 0.17, + "learning_rate": 0.0006735887662988968, + "loss": 1.5716, + "theoretical_loss": 3.9042212737511743, + "tokens_seen": 549257216 + }, + { + "epoch": 0.17, + "learning_rate": 0.0006735727181544634, + "loss": 1.5777, + "theoretical_loss": 3.904172271445487, + "tokens_seen": 549322752 + }, + { + "epoch": 0.17, + "learning_rate": 0.0006735566700100302, + "loss": 1.5596, + "theoretical_loss": 3.9041232766222813, + "tokens_seen": 549388288 + }, + { + "epoch": 0.17, + "learning_rate": 0.0006735406218655968, + "loss": 1.5278, + "theoretical_loss": 3.9040742892795226, + "tokens_seen": 549453824 + }, + { + "epoch": 0.17, + "learning_rate": 0.0006735245737211635, + "loss": 1.5968, + "theoretical_loss": 3.9040253094151773, + "tokens_seen": 549519360 + }, + { + "epoch": 0.17, + "learning_rate": 0.0006735085255767302, + "loss": 1.6684, + "theoretical_loss": 3.9039763370272107, + "tokens_seen": 549584896 + }, + { + "epoch": 0.17, + "learning_rate": 0.0006734924774322969, + "loss": 1.4234, + "theoretical_loss": 3.903927372113592, + "tokens_seen": 549650432 + }, + { + "epoch": 0.17, + "learning_rate": 0.0006734764292878636, + "loss": 1.551, + "theoretical_loss": 3.9038784146722874, + "tokens_seen": 549715968 + }, + { + "epoch": 0.17, + "learning_rate": 0.0006734603811434303, + "loss": 1.6019, + "theoretical_loss": 3.9038294647012672, + "tokens_seen": 549781504 + }, + { + "epoch": 0.17, + "learning_rate": 0.000673444332998997, + "loss": 1.5353, + "theoretical_loss": 3.9037805221985016, + "tokens_seen": 549847040 + }, + { + "epoch": 0.17, + "learning_rate": 0.0006734282848545637, + "loss": 1.4861, + "theoretical_loss": 3.90373158716196, + "tokens_seen": 549912576 + }, + { + "epoch": 0.17, + "learning_rate": 0.0006734122367101304, + "loss": 1.5978, + "theoretical_loss": 3.903682659589615, + "tokens_seen": 549978112 + }, + { + "epoch": 0.17, + "learning_rate": 0.0006733961885656971, + "loss": 1.4957, + "theoretical_loss": 3.903633739479438, + "tokens_seen": 550043648 + }, + { + "epoch": 0.17, + "learning_rate": 0.0006733801404212638, + "loss": 1.5126, + "theoretical_loss": 3.9035848268294027, + "tokens_seen": 550109184 + }, + { + "epoch": 0.17, + "learning_rate": 0.0006733640922768305, + "loss": 1.4566, + "theoretical_loss": 3.9035359216374825, + "tokens_seen": 550174720 + }, + { + "epoch": 0.17, + "learning_rate": 0.0006733480441323972, + "loss": 1.6607, + "theoretical_loss": 3.9034870239016524, + "tokens_seen": 550240256 + }, + { + "epoch": 0.17, + "learning_rate": 0.0006733319959879639, + "loss": 1.5756, + "theoretical_loss": 3.903438133619887, + "tokens_seen": 550305792 + }, + { + "epoch": 0.17, + "learning_rate": 0.0006733159478435306, + "loss": 1.3884, + "theoretical_loss": 3.9033892507901635, + "tokens_seen": 550371328 + }, + { + "epoch": 0.17, + "learning_rate": 0.0006732998996990973, + "loss": 1.5891, + "theoretical_loss": 3.9033403754104583, + "tokens_seen": 550436864 + }, + { + "epoch": 0.17, + "objective/train/docs_used": 206347, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.64381742477417, + "objective/train/theoretical_loss": 3.9032915074787486, + "objective/train/tokens_used": 570962400, + "theoretical_loss": 3.9032915074787486, + "tokens_seen": 550502400 + }, + { + "epoch": 0.17, + "learning_rate": 0.000673283851554664, + "loss": 1.5668, + "theoretical_loss": 3.9032915074787486, + "tokens_seen": 550502400 + }, + { + "epoch": 0.17, + "learning_rate": 0.0006732678034102307, + "loss": 1.5784, + "theoretical_loss": 3.903242646993014, + "tokens_seen": 550567936 + }, + { + "epoch": 0.17, + "learning_rate": 0.0006732517552657974, + "loss": 1.5621, + "theoretical_loss": 3.9031937939512336, + "tokens_seen": 550633472 + }, + { + "epoch": 0.17, + "learning_rate": 0.0006732357071213641, + "loss": 1.5139, + "theoretical_loss": 3.903144948351387, + "tokens_seen": 550699008 + }, + { + "epoch": 0.17, + "learning_rate": 0.0006732196589769308, + "loss": 1.4477, + "theoretical_loss": 3.9030961101914556, + "tokens_seen": 550764544 + }, + { + "epoch": 0.17, + "learning_rate": 0.0006732036108324975, + "loss": 1.5992, + "theoretical_loss": 3.903047279469421, + "tokens_seen": 550830080 + }, + { + "epoch": 0.17, + "learning_rate": 0.0006731875626880642, + "loss": 1.5819, + "theoretical_loss": 3.9029984561832656, + "tokens_seen": 550895616 + }, + { + "epoch": 0.17, + "learning_rate": 0.0006731715145436309, + "loss": 1.5205, + "theoretical_loss": 3.9029496403309727, + "tokens_seen": 550961152 + }, + { + "epoch": 0.17, + "learning_rate": 0.0006731554663991976, + "loss": 1.6922, + "theoretical_loss": 3.902900831910527, + "tokens_seen": 551026688 + }, + { + "epoch": 0.17, + "learning_rate": 0.0006731394182547643, + "loss": 1.5918, + "theoretical_loss": 3.902852030919912, + "tokens_seen": 551092224 + }, + { + "epoch": 0.17, + "learning_rate": 0.000673123370110331, + "loss": 1.5323, + "theoretical_loss": 3.902803237357115, + "tokens_seen": 551157760 + }, + { + "epoch": 0.17, + "learning_rate": 0.0006731073219658977, + "loss": 1.5679, + "theoretical_loss": 3.9027544512201207, + "tokens_seen": 551223296 + }, + { + "epoch": 0.17, + "learning_rate": 0.0006730912738214644, + "loss": 1.562, + "theoretical_loss": 3.902705672506918, + "tokens_seen": 551288832 + }, + { + "epoch": 0.17, + "learning_rate": 0.0006730752256770311, + "loss": 1.5465, + "theoretical_loss": 3.9026569012154932, + "tokens_seen": 551354368 + }, + { + "epoch": 0.17, + "learning_rate": 0.0006730591775325978, + "loss": 1.5349, + "theoretical_loss": 3.9026081373438366, + "tokens_seen": 551419904 + }, + { + "epoch": 0.17, + "learning_rate": 0.0006730431293881645, + "loss": 1.5162, + "theoretical_loss": 3.9025593808899366, + "tokens_seen": 551485440 + }, + { + "epoch": 0.17, + "learning_rate": 0.0006730270812437312, + "loss": 1.6557, + "theoretical_loss": 3.9025106318517846, + "tokens_seen": 551550976 + }, + { + "epoch": 0.17, + "learning_rate": 0.0006730110330992979, + "loss": 1.4724, + "theoretical_loss": 3.902461890227371, + "tokens_seen": 551616512 + }, + { + "epoch": 0.17, + "learning_rate": 0.0006729949849548646, + "loss": 1.4511, + "theoretical_loss": 3.902413156014688, + "tokens_seen": 551682048 + }, + { + "epoch": 0.17, + "learning_rate": 0.0006729789368104313, + "loss": 1.5334, + "theoretical_loss": 3.902364429211728, + "tokens_seen": 551747584 + }, + { + "epoch": 0.17, + "learning_rate": 0.000672962888665998, + "loss": 1.5567, + "theoretical_loss": 3.9023157098164853, + "tokens_seen": 551813120 + }, + { + "epoch": 0.17, + "learning_rate": 0.0006729468405215647, + "loss": 1.5868, + "theoretical_loss": 3.902266997826953, + "tokens_seen": 551878656 + }, + { + "epoch": 0.17, + "learning_rate": 0.0006729307923771314, + "loss": 1.4461, + "theoretical_loss": 3.902218293241127, + "tokens_seen": 551944192 + }, + { + "epoch": 0.17, + "learning_rate": 0.0006729147442326981, + "loss": 1.4911, + "theoretical_loss": 3.9021695960570026, + "tokens_seen": 552009728 + }, + { + "epoch": 0.17, + "learning_rate": 0.0006728986960882648, + "loss": 1.5633, + "theoretical_loss": 3.902120906272577, + "tokens_seen": 552075264 + }, + { + "epoch": 0.17, + "learning_rate": 0.0006728826479438315, + "loss": 1.5254, + "theoretical_loss": 3.9020722238858467, + "tokens_seen": 552140800 + }, + { + "epoch": 0.17, + "learning_rate": 0.0006728665997993982, + "loss": 1.51, + "theoretical_loss": 3.902023548894811, + "tokens_seen": 552206336 + }, + { + "epoch": 0.17, + "learning_rate": 0.0006728505516549649, + "loss": 1.6601, + "theoretical_loss": 3.901974881297468, + "tokens_seen": 552271872 + }, + { + "epoch": 0.17, + "learning_rate": 0.0006728345035105316, + "loss": 1.5342, + "theoretical_loss": 3.9019262210918173, + "tokens_seen": 552337408 + }, + { + "epoch": 0.17, + "learning_rate": 0.0006728184553660983, + "loss": 1.411, + "theoretical_loss": 3.90187756827586, + "tokens_seen": 552402944 + }, + { + "epoch": 0.17, + "learning_rate": 0.000672802407221665, + "loss": 1.5242, + "theoretical_loss": 3.9018289228475975, + "tokens_seen": 552468480 + }, + { + "epoch": 0.17, + "learning_rate": 0.0006727863590772317, + "loss": 1.6469, + "theoretical_loss": 3.9017802848050307, + "tokens_seen": 552534016 + }, + { + "epoch": 0.17, + "learning_rate": 0.0006727703109327984, + "loss": 1.4478, + "theoretical_loss": 3.9017316541461637, + "tokens_seen": 552599552 + }, + { + "epoch": 0.17, + "learning_rate": 0.0006727542627883651, + "loss": 1.6469, + "theoretical_loss": 3.9016830308689996, + "tokens_seen": 552665088 + }, + { + "epoch": 0.17, + "learning_rate": 0.0006727382146439318, + "loss": 1.5693, + "theoretical_loss": 3.901634414971543, + "tokens_seen": 552730624 + }, + { + "epoch": 0.17, + "learning_rate": 0.0006727221664994985, + "loss": 1.489, + "theoretical_loss": 3.901585806451799, + "tokens_seen": 552796160 + }, + { + "epoch": 0.17, + "learning_rate": 0.0006727061183550652, + "loss": 1.4429, + "theoretical_loss": 3.901537205307773, + "tokens_seen": 552861696 + }, + { + "epoch": 0.17, + "learning_rate": 0.0006726900702106319, + "loss": 1.391, + "theoretical_loss": 3.9014886115374723, + "tokens_seen": 552927232 + }, + { + "epoch": 0.17, + "learning_rate": 0.0006726740220661986, + "loss": 1.602, + "theoretical_loss": 3.9014400251389043, + "tokens_seen": 552992768 + }, + { + "epoch": 0.17, + "learning_rate": 0.0006726579739217654, + "loss": 1.5272, + "theoretical_loss": 3.901391446110077, + "tokens_seen": 553058304 + }, + { + "epoch": 0.17, + "learning_rate": 0.000672641925777332, + "loss": 1.6125, + "theoretical_loss": 3.901342874449, + "tokens_seen": 553123840 + }, + { + "epoch": 0.17, + "learning_rate": 0.0006726258776328988, + "loss": 1.4653, + "theoretical_loss": 3.9012943101536823, + "tokens_seen": 553189376 + }, + { + "epoch": 0.17, + "learning_rate": 0.0006726098294884654, + "loss": 1.5353, + "theoretical_loss": 3.901245753222135, + "tokens_seen": 553254912 + }, + { + "epoch": 0.17, + "learning_rate": 0.0006725937813440322, + "loss": 1.5686, + "theoretical_loss": 3.90119720365237, + "tokens_seen": 553320448 + }, + { + "epoch": 0.17, + "learning_rate": 0.0006725777331995988, + "loss": 1.5124, + "theoretical_loss": 3.9011486614423982, + "tokens_seen": 553385984 + }, + { + "epoch": 0.17, + "learning_rate": 0.0006725616850551656, + "loss": 1.5627, + "theoretical_loss": 3.901100126590233, + "tokens_seen": 553451520 + }, + { + "epoch": 0.17, + "learning_rate": 0.0006725456369107322, + "loss": 1.4873, + "theoretical_loss": 3.901051599093889, + "tokens_seen": 553517056 + }, + { + "epoch": 0.17, + "learning_rate": 0.000672529588766299, + "loss": 1.6178, + "theoretical_loss": 3.901003078951379, + "tokens_seen": 553582592 + }, + { + "epoch": 0.17, + "learning_rate": 0.0006725135406218657, + "loss": 1.5038, + "theoretical_loss": 3.9009545661607192, + "tokens_seen": 553648128 + }, + { + "epoch": 0.17, + "learning_rate": 0.0006724974924774324, + "loss": 1.6446, + "theoretical_loss": 3.900906060719925, + "tokens_seen": 553713664 + }, + { + "epoch": 0.17, + "objective/train/docs_used": 207589, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.7646839618682861, + "objective/train/theoretical_loss": 3.9008575626270146, + "objective/train/tokens_used": 574239200, + "theoretical_loss": 3.9008575626270146, + "tokens_seen": 553779200 + }, + { + "epoch": 0.17, + "learning_rate": 0.000672481444332999, + "loss": 1.5949, + "theoretical_loss": 3.9008575626270146, + "tokens_seen": 553779200 + }, + { + "epoch": 0.17, + "learning_rate": 0.0006724653961885658, + "loss": 1.4948, + "theoretical_loss": 3.900809071880004, + "tokens_seen": 553844736 + }, + { + "epoch": 0.17, + "learning_rate": 0.0006724493480441325, + "loss": 1.575, + "theoretical_loss": 3.900760588476912, + "tokens_seen": 553910272 + }, + { + "epoch": 0.17, + "learning_rate": 0.0006724332998996992, + "loss": 1.636, + "theoretical_loss": 3.900712112415758, + "tokens_seen": 553975808 + }, + { + "epoch": 0.17, + "learning_rate": 0.0006724172517552659, + "loss": 1.5518, + "theoretical_loss": 3.900663643694561, + "tokens_seen": 554041344 + }, + { + "epoch": 0.17, + "learning_rate": 0.0006724012036108326, + "loss": 1.5257, + "theoretical_loss": 3.900615182311342, + "tokens_seen": 554106880 + }, + { + "epoch": 0.17, + "learning_rate": 0.0006723851554663993, + "loss": 1.5343, + "theoretical_loss": 3.900566728264123, + "tokens_seen": 554172416 + }, + { + "epoch": 0.17, + "learning_rate": 0.000672369107321966, + "loss": 1.5018, + "theoretical_loss": 3.9005182815509256, + "tokens_seen": 554237952 + }, + { + "epoch": 0.17, + "learning_rate": 0.0006723530591775327, + "loss": 1.4893, + "theoretical_loss": 3.9004698421697723, + "tokens_seen": 554303488 + }, + { + "epoch": 0.17, + "learning_rate": 0.0006723370110330994, + "loss": 1.6764, + "theoretical_loss": 3.900421410118688, + "tokens_seen": 554369024 + }, + { + "epoch": 0.17, + "learning_rate": 0.0006723209628886661, + "loss": 1.4528, + "theoretical_loss": 3.9003729853956957, + "tokens_seen": 554434560 + }, + { + "epoch": 0.17, + "learning_rate": 0.0006723049147442328, + "loss": 1.6105, + "theoretical_loss": 3.9003245679988217, + "tokens_seen": 554500096 + }, + { + "epoch": 0.17, + "learning_rate": 0.0006722888665997995, + "loss": 1.5763, + "theoretical_loss": 3.9002761579260907, + "tokens_seen": 554565632 + }, + { + "epoch": 0.17, + "learning_rate": 0.0006722728184553661, + "loss": 1.5002, + "theoretical_loss": 3.9002277551755307, + "tokens_seen": 554631168 + }, + { + "epoch": 0.17, + "learning_rate": 0.0006722567703109329, + "loss": 1.523, + "theoretical_loss": 3.9001793597451684, + "tokens_seen": 554696704 + }, + { + "epoch": 0.17, + "learning_rate": 0.0006722407221664995, + "loss": 1.5555, + "theoretical_loss": 3.900130971633033, + "tokens_seen": 554762240 + }, + { + "epoch": 0.17, + "learning_rate": 0.0006722246740220663, + "loss": 1.5142, + "theoretical_loss": 3.9000825908371524, + "tokens_seen": 554827776 + }, + { + "epoch": 0.17, + "learning_rate": 0.0006722086258776329, + "loss": 1.4769, + "theoretical_loss": 3.9000342173555573, + "tokens_seen": 554893312 + }, + { + "epoch": 0.17, + "learning_rate": 0.0006721925777331997, + "loss": 1.5985, + "theoretical_loss": 3.899985851186277, + "tokens_seen": 554958848 + }, + { + "epoch": 0.17, + "learning_rate": 0.0006721765295887663, + "loss": 1.4712, + "theoretical_loss": 3.8999374923273447, + "tokens_seen": 555024384 + }, + { + "epoch": 0.17, + "learning_rate": 0.0006721604814443331, + "loss": 1.5977, + "theoretical_loss": 3.899889140776791, + "tokens_seen": 555089920 + }, + { + "epoch": 0.17, + "learning_rate": 0.0006721444332998997, + "loss": 1.5134, + "theoretical_loss": 3.8998407965326494, + "tokens_seen": 555155456 + }, + { + "epoch": 0.17, + "learning_rate": 0.0006721283851554664, + "loss": 1.4379, + "theoretical_loss": 3.899792459592953, + "tokens_seen": 555220992 + }, + { + "epoch": 0.17, + "learning_rate": 0.0006721123370110331, + "loss": 1.4483, + "theoretical_loss": 3.899744129955737, + "tokens_seen": 555286528 + }, + { + "epoch": 0.17, + "learning_rate": 0.0006720962888665998, + "loss": 1.4878, + "theoretical_loss": 3.899695807619035, + "tokens_seen": 555352064 + }, + { + "epoch": 0.17, + "learning_rate": 0.0006720802407221665, + "loss": 1.5747, + "theoretical_loss": 3.899647492580885, + "tokens_seen": 555417600 + }, + { + "epoch": 0.17, + "learning_rate": 0.0006720641925777332, + "loss": 1.5975, + "theoretical_loss": 3.899599184839322, + "tokens_seen": 555483136 + }, + { + "epoch": 0.17, + "learning_rate": 0.0006720481444332999, + "loss": 1.4576, + "theoretical_loss": 3.899550884392384, + "tokens_seen": 555548672 + }, + { + "epoch": 0.17, + "learning_rate": 0.0006720320962888666, + "loss": 1.4151, + "theoretical_loss": 3.8995025912381087, + "tokens_seen": 555614208 + }, + { + "epoch": 0.17, + "learning_rate": 0.0006720160481444333, + "loss": 1.5107, + "theoretical_loss": 3.899454305374536, + "tokens_seen": 555679744 + }, + { + "epoch": 0.17, + "learning_rate": 0.000672, + "loss": 1.4955, + "theoretical_loss": 3.899406026799704, + "tokens_seen": 555745280 + }, + { + "epoch": 0.17, + "learning_rate": 0.0006719839518555667, + "loss": 1.4928, + "theoretical_loss": 3.899357755511655, + "tokens_seen": 555810816 + }, + { + "epoch": 0.17, + "learning_rate": 0.0006719679037111334, + "loss": 1.6544, + "theoretical_loss": 3.899309491508429, + "tokens_seen": 555876352 + }, + { + "epoch": 0.17, + "learning_rate": 0.0006719518555667001, + "loss": 1.614, + "theoretical_loss": 3.899261234788068, + "tokens_seen": 555941888 + }, + { + "epoch": 0.17, + "learning_rate": 0.0006719358074222668, + "loss": 1.3133, + "theoretical_loss": 3.8992129853486155, + "tokens_seen": 556007424 + }, + { + "epoch": 0.17, + "learning_rate": 0.0006719197592778335, + "loss": 1.5193, + "theoretical_loss": 3.8991647431881145, + "tokens_seen": 556072960 + }, + { + "epoch": 0.17, + "learning_rate": 0.0006719037111334002, + "loss": 1.6433, + "theoretical_loss": 3.899116508304609, + "tokens_seen": 556138496 + }, + { + "epoch": 0.17, + "learning_rate": 0.0006718876629889669, + "loss": 1.621, + "theoretical_loss": 3.899068280696144, + "tokens_seen": 556204032 + }, + { + "epoch": 0.17, + "learning_rate": 0.0006718716148445336, + "loss": 1.4167, + "theoretical_loss": 3.8990200603607654, + "tokens_seen": 556269568 + }, + { + "epoch": 0.17, + "learning_rate": 0.0006718555667001003, + "loss": 1.5974, + "theoretical_loss": 3.8989718472965196, + "tokens_seen": 556335104 + }, + { + "epoch": 0.17, + "learning_rate": 0.000671839518555667, + "loss": 1.5636, + "theoretical_loss": 3.898923641501454, + "tokens_seen": 556400640 + }, + { + "epoch": 0.17, + "learning_rate": 0.0006718234704112337, + "loss": 1.5134, + "theoretical_loss": 3.898875442973617, + "tokens_seen": 556466176 + }, + { + "epoch": 0.17, + "learning_rate": 0.0006718074222668004, + "loss": 1.4875, + "theoretical_loss": 3.8988272517110563, + "tokens_seen": 556531712 + }, + { + "epoch": 0.17, + "learning_rate": 0.0006717913741223671, + "loss": 1.4776, + "theoretical_loss": 3.898779067711822, + "tokens_seen": 556597248 + }, + { + "epoch": 0.17, + "learning_rate": 0.0006717753259779338, + "loss": 1.4987, + "theoretical_loss": 3.8987308909739644, + "tokens_seen": 556662784 + }, + { + "epoch": 0.17, + "learning_rate": 0.0006717592778335005, + "loss": 1.6468, + "theoretical_loss": 3.898682721495535, + "tokens_seen": 556728320 + }, + { + "epoch": 0.17, + "learning_rate": 0.0006717432296890672, + "loss": 1.5536, + "theoretical_loss": 3.8986345592745844, + "tokens_seen": 556793856 + }, + { + "epoch": 0.17, + "learning_rate": 0.0006717271815446339, + "loss": 1.5359, + "theoretical_loss": 3.898586404309166, + "tokens_seen": 556859392 + }, + { + "epoch": 0.17, + "learning_rate": 0.0006717111334002006, + "loss": 1.4616, + "theoretical_loss": 3.8985382565973326, + "tokens_seen": 556924928 + }, + { + "epoch": 0.17, + "learning_rate": 0.0006716950852557673, + "loss": 1.4828, + "theoretical_loss": 3.898490116137139, + "tokens_seen": 556990464 + }, + { + "epoch": 0.17, + "objective/train/docs_used": 208740, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.582440733909607, + "objective/train/theoretical_loss": 3.898441982926639, + "objective/train/tokens_used": 577516000, + "theoretical_loss": 3.898441982926639, + "tokens_seen": 557056000 + }, + { + "epoch": 0.17, + "learning_rate": 0.000671679037111334, + "loss": 1.4972, + "theoretical_loss": 3.898441982926639, + "tokens_seen": 557056000 + }, + { + "epoch": 0.17, + "learning_rate": 0.0006716629889669007, + "loss": 1.4295, + "theoretical_loss": 3.8983938569638883, + "tokens_seen": 557121536 + }, + { + "epoch": 0.17, + "learning_rate": 0.0006716469408224674, + "loss": 1.5827, + "theoretical_loss": 3.8983457382469435, + "tokens_seen": 557187072 + }, + { + "epoch": 0.17, + "learning_rate": 0.0006716308926780341, + "loss": 1.4769, + "theoretical_loss": 3.8982976267738616, + "tokens_seen": 557252608 + }, + { + "epoch": 0.17, + "learning_rate": 0.0006716148445336008, + "loss": 1.4211, + "theoretical_loss": 3.8982495225427005, + "tokens_seen": 557318144 + }, + { + "epoch": 0.17, + "learning_rate": 0.0006715987963891675, + "loss": 1.477, + "theoretical_loss": 3.8982014255515183, + "tokens_seen": 557383680 + }, + { + "epoch": 0.17, + "learning_rate": 0.0006715827482447342, + "loss": 1.5364, + "theoretical_loss": 3.8981533357983746, + "tokens_seen": 557449216 + }, + { + "epoch": 0.17, + "learning_rate": 0.0006715667001003009, + "loss": 1.5072, + "theoretical_loss": 3.8981052532813294, + "tokens_seen": 557514752 + }, + { + "epoch": 0.17, + "learning_rate": 0.0006715506519558676, + "loss": 1.5025, + "theoretical_loss": 3.898057177998443, + "tokens_seen": 557580288 + }, + { + "epoch": 0.17, + "learning_rate": 0.0006715346038114343, + "loss": 1.4636, + "theoretical_loss": 3.898009109947777, + "tokens_seen": 557645824 + }, + { + "epoch": 0.17, + "learning_rate": 0.000671518555667001, + "loss": 1.4334, + "theoretical_loss": 3.897961049127395, + "tokens_seen": 557711360 + }, + { + "epoch": 0.17, + "learning_rate": 0.0006715025075225677, + "loss": 1.4902, + "theoretical_loss": 3.8979129955353584, + "tokens_seen": 557776896 + }, + { + "epoch": 0.17, + "learning_rate": 0.0006714864593781344, + "loss": 1.502, + "theoretical_loss": 3.8978649491697315, + "tokens_seen": 557842432 + }, + { + "epoch": 0.17, + "learning_rate": 0.0006714704112337011, + "loss": 1.4824, + "theoretical_loss": 3.8978169100285784, + "tokens_seen": 557907968 + }, + { + "epoch": 0.17, + "learning_rate": 0.0006714543630892679, + "loss": 1.4535, + "theoretical_loss": 3.8977688781099653, + "tokens_seen": 557973504 + }, + { + "epoch": 0.17, + "learning_rate": 0.0006714383149448345, + "loss": 1.49, + "theoretical_loss": 3.8977208534119576, + "tokens_seen": 558039040 + }, + { + "epoch": 0.17, + "learning_rate": 0.0006714222668004013, + "loss": 1.6021, + "theoretical_loss": 3.8976728359326223, + "tokens_seen": 558104576 + }, + { + "epoch": 0.17, + "learning_rate": 0.0006714062186559679, + "loss": 1.5004, + "theoretical_loss": 3.897624825670026, + "tokens_seen": 558170112 + }, + { + "epoch": 0.17, + "learning_rate": 0.0006713901705115347, + "loss": 1.5875, + "theoretical_loss": 3.8975768226222383, + "tokens_seen": 558235648 + }, + { + "epoch": 0.17, + "learning_rate": 0.0006713741223671013, + "loss": 1.5321, + "theoretical_loss": 3.897528826787327, + "tokens_seen": 558301184 + }, + { + "epoch": 0.17, + "learning_rate": 0.0006713580742226681, + "loss": 1.5268, + "theoretical_loss": 3.8974808381633625, + "tokens_seen": 558366720 + }, + { + "epoch": 0.17, + "learning_rate": 0.0006713420260782347, + "loss": 1.6611, + "theoretical_loss": 3.8974328567484147, + "tokens_seen": 558432256 + }, + { + "epoch": 0.17, + "learning_rate": 0.0006713259779338015, + "loss": 1.5848, + "theoretical_loss": 3.8973848825405555, + "tokens_seen": 558497792 + }, + { + "epoch": 0.17, + "learning_rate": 0.0006713099297893681, + "loss": 1.5185, + "theoretical_loss": 3.8973369155378563, + "tokens_seen": 558563328 + }, + { + "epoch": 0.17, + "learning_rate": 0.0006712938816449349, + "loss": 1.5352, + "theoretical_loss": 3.89728895573839, + "tokens_seen": 558628864 + }, + { + "epoch": 0.17, + "learning_rate": 0.0006712778335005015, + "loss": 1.5519, + "theoretical_loss": 3.8972410031402296, + "tokens_seen": 558694400 + }, + { + "epoch": 0.17, + "learning_rate": 0.0006712617853560683, + "loss": 1.5777, + "theoretical_loss": 3.8971930577414495, + "tokens_seen": 558759936 + }, + { + "epoch": 0.17, + "learning_rate": 0.000671245737211635, + "loss": 1.5814, + "theoretical_loss": 3.8971451195401254, + "tokens_seen": 558825472 + }, + { + "epoch": 0.17, + "learning_rate": 0.0006712296890672017, + "loss": 1.5516, + "theoretical_loss": 3.897097188534332, + "tokens_seen": 558891008 + }, + { + "epoch": 0.17, + "learning_rate": 0.0006712136409227683, + "loss": 1.5827, + "theoretical_loss": 3.8970492647221464, + "tokens_seen": 558956544 + }, + { + "epoch": 0.17, + "learning_rate": 0.0006711975927783351, + "loss": 1.5942, + "theoretical_loss": 3.897001348101645, + "tokens_seen": 559022080 + }, + { + "epoch": 0.17, + "learning_rate": 0.0006711815446339017, + "loss": 1.4914, + "theoretical_loss": 3.8969534386709053, + "tokens_seen": 559087616 + }, + { + "epoch": 0.17, + "learning_rate": 0.0006711654964894685, + "loss": 1.6098, + "theoretical_loss": 3.896905536428007, + "tokens_seen": 559153152 + }, + { + "epoch": 0.17, + "learning_rate": 0.0006711494483450352, + "loss": 1.5997, + "theoretical_loss": 3.8968576413710294, + "tokens_seen": 559218688 + }, + { + "epoch": 0.17, + "learning_rate": 0.0006711334002006019, + "loss": 1.5389, + "theoretical_loss": 3.8968097534980517, + "tokens_seen": 559284224 + }, + { + "epoch": 0.17, + "learning_rate": 0.0006711173520561686, + "loss": 1.4889, + "theoretical_loss": 3.8967618728071556, + "tokens_seen": 559349760 + }, + { + "epoch": 0.17, + "learning_rate": 0.0006711013039117353, + "loss": 1.5712, + "theoretical_loss": 3.896713999296422, + "tokens_seen": 559415296 + }, + { + "epoch": 0.17, + "learning_rate": 0.000671085255767302, + "loss": 1.5529, + "theoretical_loss": 3.8966661329639334, + "tokens_seen": 559480832 + }, + { + "epoch": 0.17, + "learning_rate": 0.0006710692076228687, + "loss": 1.5749, + "theoretical_loss": 3.896618273807773, + "tokens_seen": 559546368 + }, + { + "epoch": 0.17, + "learning_rate": 0.0006710531594784354, + "loss": 1.5294, + "theoretical_loss": 3.8965704218260244, + "tokens_seen": 559611904 + }, + { + "epoch": 0.17, + "learning_rate": 0.0006710371113340021, + "loss": 1.6145, + "theoretical_loss": 3.896522577016772, + "tokens_seen": 559677440 + }, + { + "epoch": 0.17, + "learning_rate": 0.0006710210631895688, + "loss": 1.5557, + "theoretical_loss": 3.896474739378102, + "tokens_seen": 559742976 + }, + { + "epoch": 0.17, + "learning_rate": 0.0006710050150451355, + "loss": 1.5817, + "theoretical_loss": 3.896426908908099, + "tokens_seen": 559808512 + }, + { + "epoch": 0.17, + "learning_rate": 0.0006709889669007022, + "loss": 1.5264, + "theoretical_loss": 3.89637908560485, + "tokens_seen": 559874048 + }, + { + "epoch": 0.17, + "learning_rate": 0.0006709729187562689, + "loss": 1.5195, + "theoretical_loss": 3.896331269466443, + "tokens_seen": 559939584 + }, + { + "epoch": 0.17, + "learning_rate": 0.0006709568706118356, + "loss": 1.4636, + "theoretical_loss": 3.8962834604909657, + "tokens_seen": 560005120 + }, + { + "epoch": 0.17, + "learning_rate": 0.0006709408224674023, + "loss": 1.4809, + "theoretical_loss": 3.8962356586765075, + "tokens_seen": 560070656 + }, + { + "epoch": 0.17, + "learning_rate": 0.000670924774322969, + "loss": 1.4566, + "theoretical_loss": 3.896187864021157, + "tokens_seen": 560136192 + }, + { + "epoch": 0.17, + "learning_rate": 0.0006709087261785357, + "loss": 1.587, + "theoretical_loss": 3.8961400765230065, + "tokens_seen": 560201728 + }, + { + "epoch": 0.17, + "learning_rate": 0.0006708926780341024, + "loss": 1.5119, + "theoretical_loss": 3.8960922961801447, + "tokens_seen": 560267264 + }, + { + "epoch": 0.17, + "objective/train/docs_used": 209977, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.4891555309295654, + "objective/train/theoretical_loss": 3.8960445229906657, + "objective/train/tokens_used": 580792800, + "theoretical_loss": 3.8960445229906657, + "tokens_seen": 560332800 + }, + { + "epoch": 0.17, + "learning_rate": 0.0006708766298896691, + "loss": 1.4649, + "theoretical_loss": 3.8960445229906657, + "tokens_seen": 560332800 + }, + { + "epoch": 0.17, + "learning_rate": 0.0006708605817452358, + "loss": 1.5073, + "theoretical_loss": 3.895996756952661, + "tokens_seen": 560398336 + }, + { + "epoch": 0.17, + "learning_rate": 0.0006708445336008024, + "loss": 1.4887, + "theoretical_loss": 3.895948998064223, + "tokens_seen": 560463872 + }, + { + "epoch": 0.17, + "learning_rate": 0.0006708284854563692, + "loss": 1.5224, + "theoretical_loss": 3.8959012463234473, + "tokens_seen": 560529408 + }, + { + "epoch": 0.17, + "learning_rate": 0.0006708124373119358, + "loss": 1.3575, + "theoretical_loss": 3.8958535017284284, + "tokens_seen": 560594944 + }, + { + "epoch": 0.17, + "learning_rate": 0.0006707963891675026, + "loss": 1.4763, + "theoretical_loss": 3.895805764277261, + "tokens_seen": 560660480 + }, + { + "epoch": 0.17, + "learning_rate": 0.0006707803410230692, + "loss": 1.5626, + "theoretical_loss": 3.895758033968042, + "tokens_seen": 560726016 + }, + { + "epoch": 0.17, + "learning_rate": 0.000670764292878636, + "loss": 1.4727, + "theoretical_loss": 3.895710310798868, + "tokens_seen": 560791552 + }, + { + "epoch": 0.17, + "learning_rate": 0.0006707482447342026, + "loss": 1.5432, + "theoretical_loss": 3.895662594767837, + "tokens_seen": 560857088 + }, + { + "epoch": 0.17, + "learning_rate": 0.0006707321965897693, + "loss": 1.5923, + "theoretical_loss": 3.895614885873047, + "tokens_seen": 560922624 + }, + { + "epoch": 0.17, + "learning_rate": 0.000670716148445336, + "loss": 1.5416, + "theoretical_loss": 3.895567184112598, + "tokens_seen": 560988160 + }, + { + "epoch": 0.17, + "learning_rate": 0.0006707001003009027, + "loss": 1.6404, + "theoretical_loss": 3.8955194894845886, + "tokens_seen": 561053696 + }, + { + "epoch": 0.17, + "learning_rate": 0.0006706840521564694, + "loss": 1.4986, + "theoretical_loss": 3.89547180198712, + "tokens_seen": 561119232 + }, + { + "epoch": 0.17, + "learning_rate": 0.0006706680040120361, + "loss": 1.4429, + "theoretical_loss": 3.8954241216182943, + "tokens_seen": 561184768 + }, + { + "epoch": 0.17, + "learning_rate": 0.0006706519558676028, + "loss": 1.5499, + "theoretical_loss": 3.8953764483762123, + "tokens_seen": 561250304 + }, + { + "epoch": 0.17, + "learning_rate": 0.0006706359077231695, + "loss": 1.4866, + "theoretical_loss": 3.8953287822589777, + "tokens_seen": 561315840 + }, + { + "epoch": 0.17, + "learning_rate": 0.0006706198595787362, + "loss": 1.4202, + "theoretical_loss": 3.8952811232646933, + "tokens_seen": 561381376 + }, + { + "epoch": 0.17, + "learning_rate": 0.0006706038114343029, + "loss": 1.5772, + "theoretical_loss": 3.895233471391464, + "tokens_seen": 561446912 + }, + { + "epoch": 0.17, + "learning_rate": 0.0006705877632898696, + "loss": 1.4505, + "theoretical_loss": 3.895185826637394, + "tokens_seen": 561512448 + }, + { + "epoch": 0.17, + "learning_rate": 0.0006705717151454363, + "loss": 1.6127, + "theoretical_loss": 3.89513818900059, + "tokens_seen": 561577984 + }, + { + "epoch": 0.17, + "learning_rate": 0.000670555667001003, + "loss": 1.5053, + "theoretical_loss": 3.8950905584791577, + "tokens_seen": 561643520 + }, + { + "epoch": 0.17, + "learning_rate": 0.0006705396188565697, + "loss": 1.5361, + "theoretical_loss": 3.8950429350712037, + "tokens_seen": 561709056 + }, + { + "epoch": 0.17, + "learning_rate": 0.0006705235707121364, + "loss": 1.5016, + "theoretical_loss": 3.8949953187748374, + "tokens_seen": 561774592 + }, + { + "epoch": 0.17, + "learning_rate": 0.0006705075225677031, + "loss": 1.5258, + "theoretical_loss": 3.894947709588166, + "tokens_seen": 561840128 + }, + { + "epoch": 0.17, + "learning_rate": 0.0006704914744232698, + "loss": 1.4065, + "theoretical_loss": 3.8949001075093, + "tokens_seen": 561905664 + }, + { + "epoch": 0.17, + "learning_rate": 0.0006704754262788365, + "loss": 1.5687, + "theoretical_loss": 3.8948525125363487, + "tokens_seen": 561971200 + }, + { + "epoch": 0.17, + "learning_rate": 0.0006704593781344032, + "loss": 1.5608, + "theoretical_loss": 3.8948049246674223, + "tokens_seen": 562036736 + }, + { + "epoch": 0.17, + "learning_rate": 0.0006704433299899699, + "loss": 1.5062, + "theoretical_loss": 3.894757343900633, + "tokens_seen": 562102272 + }, + { + "epoch": 0.17, + "learning_rate": 0.0006704272818455366, + "loss": 1.6068, + "theoretical_loss": 3.894709770234093, + "tokens_seen": 562167808 + }, + { + "epoch": 0.17, + "learning_rate": 0.0006704112337011033, + "loss": 1.5432, + "theoretical_loss": 3.8946622036659155, + "tokens_seen": 562233344 + }, + { + "epoch": 0.17, + "learning_rate": 0.00067039518555667, + "loss": 1.6706, + "theoretical_loss": 3.8946146441942138, + "tokens_seen": 562298880 + }, + { + "epoch": 0.17, + "learning_rate": 0.0006703791374122367, + "loss": 1.5393, + "theoretical_loss": 3.894567091817102, + "tokens_seen": 562364416 + }, + { + "epoch": 0.17, + "learning_rate": 0.0006703630892678034, + "loss": 1.4583, + "theoretical_loss": 3.8945195465326954, + "tokens_seen": 562429952 + }, + { + "epoch": 0.17, + "learning_rate": 0.0006703470411233701, + "loss": 1.5098, + "theoretical_loss": 3.8944720083391093, + "tokens_seen": 562495488 + }, + { + "epoch": 0.17, + "learning_rate": 0.0006703309929789368, + "loss": 1.5555, + "theoretical_loss": 3.8944244772344616, + "tokens_seen": 562561024 + }, + { + "epoch": 0.17, + "learning_rate": 0.0006703149448345035, + "loss": 1.4953, + "theoretical_loss": 3.894376953216868, + "tokens_seen": 562626560 + }, + { + "epoch": 0.17, + "learning_rate": 0.0006702988966900702, + "loss": 1.5626, + "theoretical_loss": 3.8943294362844476, + "tokens_seen": 562692096 + }, + { + "epoch": 0.17, + "learning_rate": 0.0006702828485456369, + "loss": 1.5616, + "theoretical_loss": 3.8942819264353186, + "tokens_seen": 562757632 + }, + { + "epoch": 0.17, + "learning_rate": 0.0006702668004012036, + "loss": 1.5596, + "theoretical_loss": 3.8942344236676005, + "tokens_seen": 562823168 + }, + { + "epoch": 0.17, + "learning_rate": 0.0006702507522567704, + "loss": 1.5557, + "theoretical_loss": 3.894186927979413, + "tokens_seen": 562888704 + }, + { + "epoch": 0.17, + "learning_rate": 0.000670234704112337, + "loss": 1.6581, + "theoretical_loss": 3.8941394393688777, + "tokens_seen": 562954240 + }, + { + "epoch": 0.17, + "learning_rate": 0.0006702186559679038, + "loss": 1.5227, + "theoretical_loss": 3.8940919578341155, + "tokens_seen": 563019776 + }, + { + "epoch": 0.17, + "learning_rate": 0.0006702026078234704, + "loss": 1.5154, + "theoretical_loss": 3.894044483373249, + "tokens_seen": 563085312 + }, + { + "epoch": 0.17, + "learning_rate": 0.0006701865596790372, + "loss": 1.6639, + "theoretical_loss": 3.893997015984401, + "tokens_seen": 563150848 + }, + { + "epoch": 0.17, + "learning_rate": 0.0006701705115346038, + "loss": 1.5539, + "theoretical_loss": 3.893949555665696, + "tokens_seen": 563216384 + }, + { + "epoch": 0.17, + "learning_rate": 0.0006701544633901706, + "loss": 1.4203, + "theoretical_loss": 3.893902102415257, + "tokens_seen": 563281920 + }, + { + "epoch": 0.17, + "learning_rate": 0.0006701384152457372, + "loss": 1.6629, + "theoretical_loss": 3.89385465623121, + "tokens_seen": 563347456 + }, + { + "epoch": 0.17, + "learning_rate": 0.000670122367101304, + "loss": 1.549, + "theoretical_loss": 3.8938072171116804, + "tokens_seen": 563412992 + }, + { + "epoch": 0.17, + "learning_rate": 0.0006701063189568706, + "loss": 1.4734, + "theoretical_loss": 3.893759785054796, + "tokens_seen": 563478528 + }, + { + "epoch": 0.17, + "learning_rate": 0.0006700902708124374, + "loss": 1.4825, + "theoretical_loss": 3.8937123600586823, + "tokens_seen": 563544064 + }, + { + "epoch": 0.17, + "objective/train/docs_used": 211161, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.6450563669204712, + "objective/train/theoretical_loss": 3.8936649421214686, + "objective/train/tokens_used": 584069600, + "theoretical_loss": 3.8936649421214686, + "tokens_seen": 563609600 + }, + { + "epoch": 0.17, + "learning_rate": 0.000670074222668004, + "loss": 1.533, + "theoretical_loss": 3.8936649421214686, + "tokens_seen": 563609600 + }, + { + "epoch": 0.17, + "learning_rate": 0.0006700581745235708, + "loss": 1.5589, + "theoretical_loss": 3.8936175312412833, + "tokens_seen": 563675136 + }, + { + "epoch": 0.17, + "learning_rate": 0.0006700421263791374, + "loss": 1.522, + "theoretical_loss": 3.8935701274162553, + "tokens_seen": 563740672 + }, + { + "epoch": 0.17, + "learning_rate": 0.0006700260782347042, + "loss": 1.4654, + "theoretical_loss": 3.8935227306445155, + "tokens_seen": 563806208 + }, + { + "epoch": 0.17, + "learning_rate": 0.0006700100300902708, + "loss": 1.5974, + "theoretical_loss": 3.8934753409241942, + "tokens_seen": 563871744 + }, + { + "epoch": 0.17, + "learning_rate": 0.0006699939819458376, + "loss": 1.6078, + "theoretical_loss": 3.8934279582534232, + "tokens_seen": 563937280 + }, + { + "epoch": 0.17, + "learning_rate": 0.0006699779338014042, + "loss": 1.5706, + "theoretical_loss": 3.8933805826303347, + "tokens_seen": 564002816 + }, + { + "epoch": 0.17, + "learning_rate": 0.000669961885656971, + "loss": 1.4881, + "theoretical_loss": 3.893333214053062, + "tokens_seen": 564068352 + }, + { + "epoch": 0.17, + "learning_rate": 0.0006699458375125376, + "loss": 1.5906, + "theoretical_loss": 3.8932858525197385, + "tokens_seen": 564133888 + }, + { + "epoch": 0.17, + "learning_rate": 0.0006699297893681044, + "loss": 1.5673, + "theoretical_loss": 3.8932384980284986, + "tokens_seen": 564199424 + }, + { + "epoch": 0.17, + "learning_rate": 0.000669913741223671, + "loss": 1.4499, + "theoretical_loss": 3.8931911505774766, + "tokens_seen": 564264960 + }, + { + "epoch": 0.17, + "learning_rate": 0.0006698976930792378, + "loss": 1.545, + "theoretical_loss": 3.89314381016481, + "tokens_seen": 564330496 + }, + { + "epoch": 0.17, + "learning_rate": 0.0006698816449348044, + "loss": 1.5354, + "theoretical_loss": 3.8930964767886347, + "tokens_seen": 564396032 + }, + { + "epoch": 0.17, + "learning_rate": 0.0006698655967903712, + "loss": 1.457, + "theoretical_loss": 3.893049150447087, + "tokens_seen": 564461568 + }, + { + "epoch": 0.17, + "learning_rate": 0.0006698495486459378, + "loss": 1.508, + "theoretical_loss": 3.8930018311383066, + "tokens_seen": 564527104 + }, + { + "epoch": 0.17, + "learning_rate": 0.0006698335005015046, + "loss": 1.4281, + "theoretical_loss": 3.8929545188604306, + "tokens_seen": 564592640 + }, + { + "epoch": 0.17, + "learning_rate": 0.0006698174523570712, + "loss": 1.5234, + "theoretical_loss": 3.8929072136115987, + "tokens_seen": 564658176 + }, + { + "epoch": 0.17, + "learning_rate": 0.000669801404212638, + "loss": 1.5418, + "theoretical_loss": 3.8928599153899515, + "tokens_seen": 564723712 + }, + { + "epoch": 0.17, + "learning_rate": 0.0006697853560682047, + "loss": 1.4775, + "theoretical_loss": 3.8928126241936294, + "tokens_seen": 564789248 + }, + { + "epoch": 0.17, + "learning_rate": 0.0006697693079237714, + "loss": 1.4704, + "theoretical_loss": 3.8927653400207745, + "tokens_seen": 564854784 + }, + { + "epoch": 0.17, + "learning_rate": 0.000669753259779338, + "loss": 1.4684, + "theoretical_loss": 3.892718062869528, + "tokens_seen": 564920320 + }, + { + "epoch": 0.17, + "learning_rate": 0.0006697372116349048, + "loss": 1.5643, + "theoretical_loss": 3.8926707927380337, + "tokens_seen": 564985856 + }, + { + "epoch": 0.17, + "learning_rate": 0.0006697211634904715, + "loss": 1.5849, + "theoretical_loss": 3.892623529624435, + "tokens_seen": 565051392 + }, + { + "epoch": 0.17, + "learning_rate": 0.0006697051153460382, + "loss": 1.4685, + "theoretical_loss": 3.892576273526876, + "tokens_seen": 565116928 + }, + { + "epoch": 0.17, + "learning_rate": 0.0006696890672016049, + "loss": 1.5196, + "theoretical_loss": 3.892529024443502, + "tokens_seen": 565182464 + }, + { + "epoch": 0.17, + "learning_rate": 0.0006696730190571716, + "loss": 1.6693, + "theoretical_loss": 3.892481782372458, + "tokens_seen": 565248000 + }, + { + "epoch": 0.17, + "learning_rate": 0.0006696569709127383, + "loss": 1.4694, + "theoretical_loss": 3.8924345473118915, + "tokens_seen": 565313536 + }, + { + "epoch": 0.17, + "learning_rate": 0.000669640922768305, + "loss": 1.5888, + "theoretical_loss": 3.892387319259949, + "tokens_seen": 565379072 + }, + { + "epoch": 0.17, + "learning_rate": 0.0006696248746238717, + "loss": 1.5553, + "theoretical_loss": 3.8923400982147784, + "tokens_seen": 565444608 + }, + { + "epoch": 0.17, + "learning_rate": 0.0006696088264794384, + "loss": 1.5463, + "theoretical_loss": 3.8922928841745286, + "tokens_seen": 565510144 + }, + { + "epoch": 0.17, + "learning_rate": 0.0006695927783350051, + "loss": 1.613, + "theoretical_loss": 3.892245677137349, + "tokens_seen": 565575680 + }, + { + "epoch": 0.17, + "learning_rate": 0.0006695767301905718, + "loss": 1.4964, + "theoretical_loss": 3.8921984771013887, + "tokens_seen": 565641216 + }, + { + "epoch": 0.17, + "learning_rate": 0.0006695606820461385, + "loss": 1.5188, + "theoretical_loss": 3.8921512840647985, + "tokens_seen": 565706752 + }, + { + "epoch": 0.17, + "learning_rate": 0.0006695446339017052, + "loss": 1.4971, + "theoretical_loss": 3.892104098025731, + "tokens_seen": 565772288 + }, + { + "epoch": 0.17, + "learning_rate": 0.0006695285857572719, + "loss": 1.5215, + "theoretical_loss": 3.892056918982337, + "tokens_seen": 565837824 + }, + { + "epoch": 0.17, + "learning_rate": 0.0006695125376128386, + "loss": 1.6373, + "theoretical_loss": 3.8920097469327697, + "tokens_seen": 565903360 + }, + { + "epoch": 0.17, + "learning_rate": 0.0006694964894684053, + "loss": 1.4868, + "theoretical_loss": 3.8919625818751826, + "tokens_seen": 565968896 + }, + { + "epoch": 0.17, + "learning_rate": 0.000669480441323972, + "loss": 1.5161, + "theoretical_loss": 3.8919154238077294, + "tokens_seen": 566034432 + }, + { + "epoch": 0.17, + "learning_rate": 0.0006694643931795387, + "loss": 1.4879, + "theoretical_loss": 3.8918682727285665, + "tokens_seen": 566099968 + }, + { + "epoch": 0.17, + "learning_rate": 0.0006694483450351053, + "loss": 1.2927, + "theoretical_loss": 3.8918211286358475, + "tokens_seen": 566165504 + }, + { + "epoch": 0.17, + "learning_rate": 0.0006694322968906721, + "loss": 1.5265, + "theoretical_loss": 3.89177399152773, + "tokens_seen": 566231040 + }, + { + "epoch": 0.17, + "learning_rate": 0.0006694162487462387, + "loss": 1.5608, + "theoretical_loss": 3.89172686140237, + "tokens_seen": 566296576 + }, + { + "epoch": 0.17, + "learning_rate": 0.0006694002006018055, + "loss": 1.5379, + "theoretical_loss": 3.8916797382579262, + "tokens_seen": 566362112 + }, + { + "epoch": 0.17, + "learning_rate": 0.0006693841524573721, + "loss": 1.523, + "theoretical_loss": 3.8916326220925566, + "tokens_seen": 566427648 + }, + { + "epoch": 0.17, + "learning_rate": 0.0006693681043129389, + "loss": 1.4184, + "theoretical_loss": 3.8915855129044195, + "tokens_seen": 566493184 + }, + { + "epoch": 0.17, + "learning_rate": 0.0006693520561685055, + "loss": 1.5095, + "theoretical_loss": 3.891538410691676, + "tokens_seen": 566558720 + }, + { + "epoch": 0.17, + "learning_rate": 0.0006693360080240723, + "loss": 1.4817, + "theoretical_loss": 3.8914913154524857, + "tokens_seen": 566624256 + }, + { + "epoch": 0.17, + "learning_rate": 0.0006693199598796389, + "loss": 1.46, + "theoretical_loss": 3.89144422718501, + "tokens_seen": 566689792 + }, + { + "epoch": 0.17, + "learning_rate": 0.0006693039117352056, + "loss": 1.5227, + "theoretical_loss": 3.891397145887411, + "tokens_seen": 566755328 + }, + { + "epoch": 0.17, + "learning_rate": 0.0006692878635907723, + "loss": 1.5802, + "theoretical_loss": 3.8913500715578504, + "tokens_seen": 566820864 + }, + { + "epoch": 0.17, + "objective/train/docs_used": 212374, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.5325250625610352, + "objective/train/theoretical_loss": 3.8913030041944925, + "objective/train/tokens_used": 587346400, + "theoretical_loss": 3.8913030041944925, + "tokens_seen": 566886400 + }, + { + "epoch": 0.17, + "learning_rate": 0.000669271815446339, + "loss": 1.5873, + "theoretical_loss": 3.8913030041944925, + "tokens_seen": 566886400 + }, + { + "epoch": 0.17, + "learning_rate": 0.0006692557673019057, + "loss": 1.555, + "theoretical_loss": 3.891255943795501, + "tokens_seen": 566951936 + }, + { + "epoch": 0.17, + "learning_rate": 0.0006692397191574724, + "loss": 1.5507, + "theoretical_loss": 3.8912088903590405, + "tokens_seen": 567017472 + }, + { + "epoch": 0.17, + "learning_rate": 0.0006692236710130391, + "loss": 1.4982, + "theoretical_loss": 3.8911618438832756, + "tokens_seen": 567083008 + }, + { + "epoch": 0.17, + "learning_rate": 0.0006692076228686058, + "loss": 1.528, + "theoretical_loss": 3.8911148043663735, + "tokens_seen": 567148544 + }, + { + "epoch": 0.17, + "learning_rate": 0.0006691915747241725, + "loss": 1.4404, + "theoretical_loss": 3.8910677718065, + "tokens_seen": 567214080 + }, + { + "epoch": 0.17, + "learning_rate": 0.0006691755265797392, + "loss": 1.6812, + "theoretical_loss": 3.8910207462018236, + "tokens_seen": 567279616 + }, + { + "epoch": 0.17, + "learning_rate": 0.0006691594784353059, + "loss": 1.5571, + "theoretical_loss": 3.890973727550511, + "tokens_seen": 567345152 + }, + { + "epoch": 0.17, + "learning_rate": 0.0006691434302908726, + "loss": 1.5151, + "theoretical_loss": 3.890926715850733, + "tokens_seen": 567410688 + }, + { + "epoch": 0.17, + "learning_rate": 0.0006691273821464393, + "loss": 1.7016, + "theoretical_loss": 3.890879711100657, + "tokens_seen": 567476224 + }, + { + "epoch": 0.17, + "learning_rate": 0.000669111334002006, + "loss": 1.5362, + "theoretical_loss": 3.8908327132984546, + "tokens_seen": 567541760 + }, + { + "epoch": 0.17, + "learning_rate": 0.0006690952858575727, + "loss": 1.4815, + "theoretical_loss": 3.890785722442296, + "tokens_seen": 567607296 + }, + { + "epoch": 0.17, + "learning_rate": 0.0006690792377131394, + "loss": 1.4937, + "theoretical_loss": 3.890738738530353, + "tokens_seen": 567672832 + }, + { + "epoch": 0.17, + "learning_rate": 0.0006690631895687061, + "loss": 1.5548, + "theoretical_loss": 3.890691761560798, + "tokens_seen": 567738368 + }, + { + "epoch": 0.17, + "learning_rate": 0.0006690471414242729, + "loss": 1.5574, + "theoretical_loss": 3.8906447915318036, + "tokens_seen": 567803904 + }, + { + "epoch": 0.17, + "learning_rate": 0.0006690310932798395, + "loss": 1.3903, + "theoretical_loss": 3.8905978284415443, + "tokens_seen": 567869440 + }, + { + "epoch": 0.17, + "learning_rate": 0.0006690150451354063, + "loss": 1.6089, + "theoretical_loss": 3.890550872288194, + "tokens_seen": 567934976 + }, + { + "epoch": 0.17, + "learning_rate": 0.0006689989969909729, + "loss": 1.5278, + "theoretical_loss": 3.890503923069927, + "tokens_seen": 568000512 + }, + { + "epoch": 0.17, + "learning_rate": 0.0006689829488465397, + "loss": 1.6138, + "theoretical_loss": 3.8904569807849203, + "tokens_seen": 568066048 + }, + { + "epoch": 0.17, + "learning_rate": 0.0006689669007021063, + "loss": 1.5506, + "theoretical_loss": 3.8904100454313495, + "tokens_seen": 568131584 + }, + { + "epoch": 0.17, + "learning_rate": 0.0006689508525576731, + "loss": 1.6011, + "theoretical_loss": 3.890363117007392, + "tokens_seen": 568197120 + }, + { + "epoch": 0.17, + "learning_rate": 0.0006689348044132397, + "loss": 1.517, + "theoretical_loss": 3.890316195511226, + "tokens_seen": 568262656 + }, + { + "epoch": 0.17, + "learning_rate": 0.0006689187562688065, + "loss": 1.5286, + "theoretical_loss": 3.8902692809410295, + "tokens_seen": 568328192 + }, + { + "epoch": 0.17, + "learning_rate": 0.0006689027081243731, + "loss": 1.5462, + "theoretical_loss": 3.8902223732949817, + "tokens_seen": 568393728 + }, + { + "epoch": 0.17, + "learning_rate": 0.0006688866599799399, + "loss": 1.6095, + "theoretical_loss": 3.8901754725712627, + "tokens_seen": 568459264 + }, + { + "epoch": 0.17, + "learning_rate": 0.0006688706118355065, + "loss": 1.5349, + "theoretical_loss": 3.890128578768053, + "tokens_seen": 568524800 + }, + { + "epoch": 0.17, + "learning_rate": 0.0006688545636910733, + "loss": 1.5355, + "theoretical_loss": 3.8900816918835335, + "tokens_seen": 568590336 + }, + { + "epoch": 0.17, + "learning_rate": 0.0006688385155466399, + "loss": 1.5287, + "theoretical_loss": 3.8900348119158865, + "tokens_seen": 568655872 + }, + { + "epoch": 0.17, + "learning_rate": 0.0006688224674022067, + "loss": 1.6321, + "theoretical_loss": 3.8899879388632947, + "tokens_seen": 568721408 + }, + { + "epoch": 0.17, + "learning_rate": 0.0006688064192577733, + "loss": 1.595, + "theoretical_loss": 3.889941072723942, + "tokens_seen": 568786944 + }, + { + "epoch": 0.17, + "learning_rate": 0.0006687903711133401, + "loss": 1.5434, + "theoretical_loss": 3.8898942134960106, + "tokens_seen": 568852480 + }, + { + "epoch": 0.17, + "learning_rate": 0.0006687743229689067, + "loss": 1.5724, + "theoretical_loss": 3.8898473611776865, + "tokens_seen": 568918016 + }, + { + "epoch": 0.17, + "learning_rate": 0.0006687582748244735, + "loss": 1.4651, + "theoretical_loss": 3.8898005157671554, + "tokens_seen": 568983552 + }, + { + "epoch": 0.17, + "learning_rate": 0.0006687422266800401, + "loss": 1.4357, + "theoretical_loss": 3.889753677262603, + "tokens_seen": 569049088 + }, + { + "epoch": 0.17, + "learning_rate": 0.0006687261785356069, + "loss": 1.6196, + "theoretical_loss": 3.8897068456622153, + "tokens_seen": 569114624 + }, + { + "epoch": 0.17, + "learning_rate": 0.0006687101303911735, + "loss": 1.4975, + "theoretical_loss": 3.8896600209641807, + "tokens_seen": 569180160 + }, + { + "epoch": 0.17, + "learning_rate": 0.0006686940822467403, + "loss": 1.5918, + "theoretical_loss": 3.889613203166687, + "tokens_seen": 569245696 + }, + { + "epoch": 0.17, + "learning_rate": 0.0006686780341023069, + "loss": 1.6405, + "theoretical_loss": 3.8895663922679224, + "tokens_seen": 569311232 + }, + { + "epoch": 0.17, + "learning_rate": 0.0006686619859578737, + "loss": 1.4963, + "theoretical_loss": 3.8895195882660776, + "tokens_seen": 569376768 + }, + { + "epoch": 0.17, + "learning_rate": 0.0006686459378134403, + "loss": 1.5356, + "theoretical_loss": 3.8894727911593416, + "tokens_seen": 569442304 + }, + { + "epoch": 0.17, + "learning_rate": 0.0006686298896690071, + "loss": 1.6726, + "theoretical_loss": 3.889426000945906, + "tokens_seen": 569507840 + }, + { + "epoch": 0.17, + "learning_rate": 0.0006686138415245737, + "loss": 1.503, + "theoretical_loss": 3.889379217623962, + "tokens_seen": 569573376 + }, + { + "epoch": 0.17, + "learning_rate": 0.0006685977933801405, + "loss": 1.4394, + "theoretical_loss": 3.8893324411917014, + "tokens_seen": 569638912 + }, + { + "epoch": 0.17, + "learning_rate": 0.0006685817452357071, + "loss": 1.586, + "theoretical_loss": 3.8892856716473183, + "tokens_seen": 569704448 + }, + { + "epoch": 0.17, + "learning_rate": 0.0006685656970912739, + "loss": 1.6604, + "theoretical_loss": 3.8892389089890056, + "tokens_seen": 569769984 + }, + { + "epoch": 0.17, + "learning_rate": 0.0006685496489468405, + "loss": 1.5405, + "theoretical_loss": 3.8891921532149576, + "tokens_seen": 569835520 + }, + { + "epoch": 0.17, + "learning_rate": 0.0006685336008024073, + "loss": 1.5311, + "theoretical_loss": 3.8891454043233686, + "tokens_seen": 569901056 + }, + { + "epoch": 0.17, + "learning_rate": 0.000668517552657974, + "loss": 1.504, + "theoretical_loss": 3.8890986623124353, + "tokens_seen": 569966592 + }, + { + "epoch": 0.17, + "learning_rate": 0.0006685015045135407, + "loss": 1.6563, + "theoretical_loss": 3.889051927180353, + "tokens_seen": 570032128 + }, + { + "epoch": 0.17, + "learning_rate": 0.0006684854563691073, + "loss": 1.4733, + "theoretical_loss": 3.88900519892532, + "tokens_seen": 570097664 + }, + { + "epoch": 0.17, + "objective/train/docs_used": 213119, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.4601891040802002, + "objective/train/theoretical_loss": 3.8889584775455326, + "objective/train/tokens_used": 590623200, + "theoretical_loss": 3.8889584775455326, + "tokens_seen": 570163200 + }, + { + "epoch": 0.17, + "learning_rate": 0.0006684694082246741, + "loss": 1.5535, + "theoretical_loss": 3.8889584775455326, + "tokens_seen": 570163200 + }, + { + "epoch": 0.17, + "learning_rate": 0.0006684533600802407, + "loss": 1.4949, + "theoretical_loss": 3.8889117630391894, + "tokens_seen": 570228736 + }, + { + "epoch": 0.17, + "learning_rate": 0.0006684373119358075, + "loss": 1.4715, + "theoretical_loss": 3.888865055404491, + "tokens_seen": 570294272 + }, + { + "epoch": 0.17, + "learning_rate": 0.0006684212637913742, + "loss": 1.4447, + "theoretical_loss": 3.8888183546396347, + "tokens_seen": 570359808 + }, + { + "epoch": 0.17, + "learning_rate": 0.0006684052156469409, + "loss": 1.5301, + "theoretical_loss": 3.8887716607428224, + "tokens_seen": 570425344 + }, + { + "epoch": 0.17, + "learning_rate": 0.0006683891675025076, + "loss": 1.4499, + "theoretical_loss": 3.8887249737122547, + "tokens_seen": 570490880 + }, + { + "epoch": 0.17, + "learning_rate": 0.0006683731193580743, + "loss": 1.452, + "theoretical_loss": 3.8886782935461333, + "tokens_seen": 570556416 + }, + { + "epoch": 0.17, + "learning_rate": 0.000668357071213641, + "loss": 1.6029, + "theoretical_loss": 3.888631620242661, + "tokens_seen": 570621952 + }, + { + "epoch": 0.17, + "learning_rate": 0.0006683410230692077, + "loss": 1.5653, + "theoretical_loss": 3.8885849538000397, + "tokens_seen": 570687488 + }, + { + "epoch": 0.17, + "learning_rate": 0.0006683249749247744, + "loss": 1.5382, + "theoretical_loss": 3.8885382942164752, + "tokens_seen": 570753024 + }, + { + "epoch": 0.17, + "learning_rate": 0.0006683089267803411, + "loss": 1.4502, + "theoretical_loss": 3.8884916414901705, + "tokens_seen": 570818560 + }, + { + "epoch": 0.17, + "learning_rate": 0.0006682928786359078, + "loss": 1.556, + "theoretical_loss": 3.8884449956193308, + "tokens_seen": 570884096 + }, + { + "epoch": 0.17, + "learning_rate": 0.0006682768304914745, + "loss": 1.6842, + "theoretical_loss": 3.8883983566021625, + "tokens_seen": 570949632 + }, + { + "epoch": 0.17, + "learning_rate": 0.0006682607823470412, + "loss": 1.4297, + "theoretical_loss": 3.8883517244368715, + "tokens_seen": 571015168 + }, + { + "epoch": 0.17, + "learning_rate": 0.0006682447342026079, + "loss": 1.5794, + "theoretical_loss": 3.8883050991216654, + "tokens_seen": 571080704 + }, + { + "epoch": 0.17, + "learning_rate": 0.0006682286860581746, + "loss": 1.5158, + "theoretical_loss": 3.888258480654751, + "tokens_seen": 571146240 + }, + { + "epoch": 0.17, + "learning_rate": 0.0006682126379137413, + "loss": 1.4432, + "theoretical_loss": 3.888211869034339, + "tokens_seen": 571211776 + }, + { + "epoch": 0.17, + "learning_rate": 0.000668196589769308, + "loss": 1.5861, + "theoretical_loss": 3.888165264258636, + "tokens_seen": 571277312 + }, + { + "epoch": 0.17, + "learning_rate": 0.0006681805416248747, + "loss": 1.6212, + "theoretical_loss": 3.888118666325853, + "tokens_seen": 571342848 + }, + { + "epoch": 0.17, + "learning_rate": 0.0006681644934804414, + "loss": 1.5539, + "theoretical_loss": 3.888072075234201, + "tokens_seen": 571408384 + }, + { + "epoch": 0.17, + "learning_rate": 0.0006681484453360081, + "loss": 1.4633, + "theoretical_loss": 3.8880254909818905, + "tokens_seen": 571473920 + }, + { + "epoch": 0.17, + "learning_rate": 0.0006681323971915748, + "loss": 1.6961, + "theoretical_loss": 3.8879789135671334, + "tokens_seen": 571539456 + }, + { + "epoch": 0.17, + "learning_rate": 0.0006681163490471415, + "loss": 1.504, + "theoretical_loss": 3.8879323429881425, + "tokens_seen": 571604992 + }, + { + "epoch": 0.17, + "learning_rate": 0.0006681003009027082, + "loss": 1.5274, + "theoretical_loss": 3.887885779243131, + "tokens_seen": 571670528 + }, + { + "epoch": 0.17, + "learning_rate": 0.0006680842527582749, + "loss": 1.5848, + "theoretical_loss": 3.887839222330313, + "tokens_seen": 571736064 + }, + { + "epoch": 0.17, + "learning_rate": 0.0006680682046138416, + "loss": 1.6348, + "theoretical_loss": 3.887792672247902, + "tokens_seen": 571801600 + }, + { + "epoch": 0.17, + "learning_rate": 0.0006680521564694082, + "loss": 1.4932, + "theoretical_loss": 3.887746128994114, + "tokens_seen": 571867136 + }, + { + "epoch": 0.17, + "learning_rate": 0.000668036108324975, + "loss": 1.6736, + "theoretical_loss": 3.8876995925671656, + "tokens_seen": 571932672 + }, + { + "epoch": 0.17, + "learning_rate": 0.0006680200601805416, + "loss": 1.4478, + "theoretical_loss": 3.8876530629652724, + "tokens_seen": 571998208 + }, + { + "epoch": 0.17, + "learning_rate": 0.0006680040120361084, + "loss": 1.5892, + "theoretical_loss": 3.8876065401866513, + "tokens_seen": 572063744 + }, + { + "epoch": 0.17, + "learning_rate": 0.000667987963891675, + "loss": 1.5512, + "theoretical_loss": 3.887560024229521, + "tokens_seen": 572129280 + }, + { + "epoch": 0.17, + "learning_rate": 0.0006679719157472418, + "loss": 1.5238, + "theoretical_loss": 3.8875135150921, + "tokens_seen": 572194816 + }, + { + "epoch": 0.17, + "learning_rate": 0.0006679558676028084, + "loss": 1.6031, + "theoretical_loss": 3.887467012772607, + "tokens_seen": 572260352 + }, + { + "epoch": 0.17, + "learning_rate": 0.0006679398194583752, + "loss": 1.6568, + "theoretical_loss": 3.8874205172692626, + "tokens_seen": 572325888 + }, + { + "epoch": 0.17, + "learning_rate": 0.0006679237713139418, + "loss": 1.5086, + "theoretical_loss": 3.887374028580287, + "tokens_seen": 572391424 + }, + { + "epoch": 0.17, + "learning_rate": 0.0006679077231695085, + "loss": 1.6675, + "theoretical_loss": 3.8873275467039017, + "tokens_seen": 572456960 + }, + { + "epoch": 0.17, + "learning_rate": 0.0006678916750250752, + "loss": 1.502, + "theoretical_loss": 3.887281071638328, + "tokens_seen": 572522496 + }, + { + "epoch": 0.17, + "learning_rate": 0.0006678756268806419, + "loss": 1.5031, + "theoretical_loss": 3.8872346033817893, + "tokens_seen": 572588032 + }, + { + "epoch": 0.17, + "learning_rate": 0.0006678595787362086, + "loss": 1.6455, + "theoretical_loss": 3.887188141932508, + "tokens_seen": 572653568 + }, + { + "epoch": 0.17, + "learning_rate": 0.0006678435305917754, + "loss": 1.5872, + "theoretical_loss": 3.8871416872887083, + "tokens_seen": 572719104 + }, + { + "epoch": 0.17, + "learning_rate": 0.000667827482447342, + "loss": 1.5449, + "theoretical_loss": 3.8870952394486156, + "tokens_seen": 572784640 + }, + { + "epoch": 0.17, + "learning_rate": 0.0006678114343029088, + "loss": 1.4379, + "theoretical_loss": 3.8870487984104543, + "tokens_seen": 572850176 + }, + { + "epoch": 0.17, + "learning_rate": 0.0006677953861584754, + "loss": 1.3765, + "theoretical_loss": 3.8870023641724503, + "tokens_seen": 572915712 + }, + { + "epoch": 0.17, + "learning_rate": 0.0006677793380140422, + "loss": 1.6085, + "theoretical_loss": 3.886955936732831, + "tokens_seen": 572981248 + }, + { + "epoch": 0.17, + "learning_rate": 0.0006677632898696088, + "loss": 1.4887, + "theoretical_loss": 3.886909516089822, + "tokens_seen": 573046784 + }, + { + "epoch": 0.17, + "learning_rate": 0.0006677472417251756, + "loss": 1.6679, + "theoretical_loss": 3.8868631022416533, + "tokens_seen": 573112320 + }, + { + "epoch": 0.17, + "learning_rate": 0.0006677311935807422, + "loss": 1.6183, + "theoretical_loss": 3.886816695186552, + "tokens_seen": 573177856 + }, + { + "epoch": 0.17, + "learning_rate": 0.000667715145436309, + "loss": 1.5256, + "theoretical_loss": 3.886770294922748, + "tokens_seen": 573243392 + }, + { + "epoch": 0.17, + "learning_rate": 0.0006676990972918756, + "loss": 1.4602, + "theoretical_loss": 3.8867239014484705, + "tokens_seen": 573308928 + }, + { + "epoch": 0.17, + "learning_rate": 0.0006676830491474424, + "loss": 1.6453, + "theoretical_loss": 3.886677514761951, + "tokens_seen": 573374464 + }, + { + "epoch": 0.17, + "objective/train/docs_used": 214228, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.5650763511657715, + "objective/train/theoretical_loss": 3.8866311348614198, + "objective/train/tokens_used": 593900000, + "theoretical_loss": 3.8866311348614198, + "tokens_seen": 573440000 + }, + { + "epoch": 0.17, + "learning_rate": 0.000667667001003009, + "loss": 1.5031, + "theoretical_loss": 3.8866311348614198, + "tokens_seen": 573440000 + }, + { + "epoch": 0.17, + "learning_rate": 0.0006676509528585758, + "loss": 1.4411, + "theoretical_loss": 3.8865847617451097, + "tokens_seen": 573505536 + }, + { + "epoch": 0.17, + "learning_rate": 0.0006676349047141424, + "loss": 1.5939, + "theoretical_loss": 3.8865383954112525, + "tokens_seen": 573571072 + }, + { + "epoch": 0.17, + "learning_rate": 0.0006676188565697092, + "loss": 1.6341, + "theoretical_loss": 3.8864920358580823, + "tokens_seen": 573636608 + }, + { + "epoch": 0.17, + "learning_rate": 0.0006676028084252758, + "loss": 1.5106, + "theoretical_loss": 3.8864456830838314, + "tokens_seen": 573702144 + }, + { + "epoch": 0.17, + "learning_rate": 0.0006675867602808426, + "loss": 1.5776, + "theoretical_loss": 3.8863993370867362, + "tokens_seen": 573767680 + }, + { + "epoch": 0.17, + "learning_rate": 0.0006675707121364092, + "loss": 1.4642, + "theoretical_loss": 3.886352997865031, + "tokens_seen": 573833216 + }, + { + "epoch": 0.17, + "learning_rate": 0.000667554663991976, + "loss": 1.4544, + "theoretical_loss": 3.8863066654169516, + "tokens_seen": 573898752 + }, + { + "epoch": 0.17, + "learning_rate": 0.0006675386158475426, + "loss": 1.5981, + "theoretical_loss": 3.886260339740735, + "tokens_seen": 573964288 + }, + { + "epoch": 0.17, + "learning_rate": 0.0006675225677031094, + "loss": 1.6268, + "theoretical_loss": 3.8862140208346174, + "tokens_seen": 574029824 + }, + { + "epoch": 0.17, + "learning_rate": 0.000667506519558676, + "loss": 1.542, + "theoretical_loss": 3.8861677086968376, + "tokens_seen": 574095360 + }, + { + "epoch": 0.17, + "learning_rate": 0.0006674904714142428, + "loss": 1.4689, + "theoretical_loss": 3.8861214033256344, + "tokens_seen": 574160896 + }, + { + "epoch": 0.17, + "learning_rate": 0.0006674744232698094, + "loss": 1.5502, + "theoretical_loss": 3.8860751047192457, + "tokens_seen": 574226432 + }, + { + "epoch": 0.17, + "learning_rate": 0.0006674583751253762, + "loss": 1.617, + "theoretical_loss": 3.886028812875912, + "tokens_seen": 574291968 + }, + { + "epoch": 0.17, + "learning_rate": 0.0006674423269809428, + "loss": 1.495, + "theoretical_loss": 3.885982527793874, + "tokens_seen": 574357504 + }, + { + "epoch": 0.17, + "learning_rate": 0.0006674262788365096, + "loss": 1.6011, + "theoretical_loss": 3.8859362494713725, + "tokens_seen": 574423040 + }, + { + "epoch": 0.17, + "learning_rate": 0.0006674102306920762, + "loss": 1.6016, + "theoretical_loss": 3.8858899779066496, + "tokens_seen": 574488576 + }, + { + "epoch": 0.17, + "learning_rate": 0.000667394182547643, + "loss": 1.5224, + "theoretical_loss": 3.8858437130979473, + "tokens_seen": 574554112 + }, + { + "epoch": 0.17, + "learning_rate": 0.0006673781344032096, + "loss": 1.6589, + "theoretical_loss": 3.8857974550435097, + "tokens_seen": 574619648 + }, + { + "epoch": 0.17, + "learning_rate": 0.0006673620862587764, + "loss": 1.5766, + "theoretical_loss": 3.885751203741579, + "tokens_seen": 574685184 + }, + { + "epoch": 0.17, + "learning_rate": 0.000667346038114343, + "loss": 1.5258, + "theoretical_loss": 3.8857049591904014, + "tokens_seen": 574750720 + }, + { + "epoch": 0.17, + "learning_rate": 0.0006673299899699098, + "loss": 1.3843, + "theoretical_loss": 3.8856587213882205, + "tokens_seen": 574816256 + }, + { + "epoch": 0.17, + "learning_rate": 0.0006673139418254764, + "loss": 1.4656, + "theoretical_loss": 3.8856124903332825, + "tokens_seen": 574881792 + }, + { + "epoch": 0.17, + "learning_rate": 0.0006672978936810432, + "loss": 1.6385, + "theoretical_loss": 3.885566266023834, + "tokens_seen": 574947328 + }, + { + "epoch": 0.17, + "learning_rate": 0.0006672818455366098, + "loss": 1.5898, + "theoretical_loss": 3.8855200484581225, + "tokens_seen": 575012864 + }, + { + "epoch": 0.17, + "learning_rate": 0.0006672657973921766, + "loss": 1.5239, + "theoretical_loss": 3.885473837634395, + "tokens_seen": 575078400 + }, + { + "epoch": 0.17, + "learning_rate": 0.0006672497492477432, + "loss": 1.4545, + "theoretical_loss": 3.8854276335508997, + "tokens_seen": 575143936 + }, + { + "epoch": 0.17, + "learning_rate": 0.00066723370110331, + "loss": 1.5442, + "theoretical_loss": 3.885381436205886, + "tokens_seen": 575209472 + }, + { + "epoch": 0.17, + "learning_rate": 0.0006672176529588766, + "loss": 1.5742, + "theoretical_loss": 3.885335245597604, + "tokens_seen": 575275008 + }, + { + "epoch": 0.17, + "learning_rate": 0.0006672016048144434, + "loss": 1.5625, + "theoretical_loss": 3.885289061724303, + "tokens_seen": 575340544 + }, + { + "epoch": 0.17, + "learning_rate": 0.00066718555667001, + "loss": 1.5808, + "theoretical_loss": 3.885242884584235, + "tokens_seen": 575406080 + }, + { + "epoch": 0.17, + "learning_rate": 0.0006671695085255768, + "loss": 1.4719, + "theoretical_loss": 3.8851967141756507, + "tokens_seen": 575471616 + }, + { + "epoch": 0.17, + "learning_rate": 0.0006671534603811434, + "loss": 1.4086, + "theoretical_loss": 3.885150550496803, + "tokens_seen": 575537152 + }, + { + "epoch": 0.17, + "learning_rate": 0.0006671374122367102, + "loss": 1.5502, + "theoretical_loss": 3.8851043935459453, + "tokens_seen": 575602688 + }, + { + "epoch": 0.17, + "learning_rate": 0.0006671213640922768, + "loss": 1.6286, + "theoretical_loss": 3.88505824332133, + "tokens_seen": 575668224 + }, + { + "epoch": 0.17, + "learning_rate": 0.0006671053159478436, + "loss": 1.5341, + "theoretical_loss": 3.8850120998212123, + "tokens_seen": 575733760 + }, + { + "epoch": 0.17, + "learning_rate": 0.0006670892678034102, + "loss": 1.7004, + "theoretical_loss": 3.884965963043847, + "tokens_seen": 575799296 + }, + { + "epoch": 0.17, + "learning_rate": 0.000667073219658977, + "loss": 1.4898, + "theoretical_loss": 3.884919832987489, + "tokens_seen": 575864832 + }, + { + "epoch": 0.17, + "learning_rate": 0.0006670571715145436, + "loss": 1.5393, + "theoretical_loss": 3.884873709650395, + "tokens_seen": 575930368 + }, + { + "epoch": 0.17, + "learning_rate": 0.0006670411233701104, + "loss": 1.4929, + "theoretical_loss": 3.8848275930308223, + "tokens_seen": 575995904 + }, + { + "epoch": 0.17, + "learning_rate": 0.000667025075225677, + "loss": 1.3647, + "theoretical_loss": 3.884781483127027, + "tokens_seen": 576061440 + }, + { + "epoch": 0.17, + "learning_rate": 0.0006670090270812438, + "loss": 1.446, + "theoretical_loss": 3.8847353799372684, + "tokens_seen": 576126976 + }, + { + "epoch": 0.17, + "learning_rate": 0.0006669929789368105, + "loss": 1.558, + "theoretical_loss": 3.8846892834598052, + "tokens_seen": 576192512 + }, + { + "epoch": 0.17, + "learning_rate": 0.0006669769307923772, + "loss": 1.5736, + "theoretical_loss": 3.884643193692896, + "tokens_seen": 576258048 + }, + { + "epoch": 0.17, + "learning_rate": 0.0006669608826479439, + "loss": 1.571, + "theoretical_loss": 3.884597110634802, + "tokens_seen": 576323584 + }, + { + "epoch": 0.17, + "learning_rate": 0.0006669448345035106, + "loss": 1.4526, + "theoretical_loss": 3.884551034283784, + "tokens_seen": 576389120 + }, + { + "epoch": 0.17, + "learning_rate": 0.0006669287863590773, + "loss": 1.4393, + "theoretical_loss": 3.8845049646381025, + "tokens_seen": 576454656 + }, + { + "epoch": 0.17, + "learning_rate": 0.000666912738214644, + "loss": 1.5262, + "theoretical_loss": 3.8844589016960196, + "tokens_seen": 576520192 + }, + { + "epoch": 0.17, + "learning_rate": 0.0006668966900702107, + "loss": 1.416, + "theoretical_loss": 3.884412845455799, + "tokens_seen": 576585728 + }, + { + "epoch": 0.17, + "learning_rate": 0.0006668806419257774, + "loss": 1.5032, + "theoretical_loss": 3.8843667959157027, + "tokens_seen": 576651264 + }, + { + "epoch": 0.17, + "objective/train/docs_used": 215374, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.587733507156372, + "objective/train/theoretical_loss": 3.884320753073996, + "objective/train/tokens_used": 597176800, + "theoretical_loss": 3.884320753073996, + "tokens_seen": 576716800 + }, + { + "epoch": 0.17, + "learning_rate": 0.0006668645937813441, + "loss": 1.5311, + "theoretical_loss": 3.884320753073996, + "tokens_seen": 576716800 + }, + { + "epoch": 0.17, + "learning_rate": 0.0006668485456369108, + "loss": 1.5659, + "theoretical_loss": 3.884274716928942, + "tokens_seen": 576782336 + }, + { + "epoch": 0.17, + "learning_rate": 0.0006668324974924775, + "loss": 1.4889, + "theoretical_loss": 3.8842286874788075, + "tokens_seen": 576847872 + }, + { + "epoch": 0.17, + "learning_rate": 0.0006668164493480442, + "loss": 1.5707, + "theoretical_loss": 3.8841826647218576, + "tokens_seen": 576913408 + }, + { + "epoch": 0.17, + "learning_rate": 0.0006668004012036109, + "loss": 1.5512, + "theoretical_loss": 3.8841366486563595, + "tokens_seen": 576978944 + }, + { + "epoch": 0.17, + "learning_rate": 0.0006667843530591776, + "loss": 1.5331, + "theoretical_loss": 3.8840906392805796, + "tokens_seen": 577044480 + }, + { + "epoch": 0.17, + "learning_rate": 0.0006667683049147443, + "loss": 1.5657, + "theoretical_loss": 3.8840446365927863, + "tokens_seen": 577110016 + }, + { + "epoch": 0.17, + "learning_rate": 0.000666752256770311, + "loss": 1.5814, + "theoretical_loss": 3.8839986405912477, + "tokens_seen": 577175552 + }, + { + "epoch": 0.17, + "learning_rate": 0.0006667362086258777, + "loss": 1.5926, + "theoretical_loss": 3.8839526512742335, + "tokens_seen": 577241088 + }, + { + "epoch": 0.17, + "learning_rate": 0.0006667201604814444, + "loss": 1.529, + "theoretical_loss": 3.883906668640013, + "tokens_seen": 577306624 + }, + { + "epoch": 0.17, + "learning_rate": 0.0006667041123370111, + "loss": 1.6871, + "theoretical_loss": 3.8838606926868575, + "tokens_seen": 577372160 + }, + { + "epoch": 0.17, + "learning_rate": 0.0006666880641925778, + "loss": 1.4675, + "theoretical_loss": 3.8838147234130367, + "tokens_seen": 577437696 + }, + { + "epoch": 0.18, + "learning_rate": 0.0006666720160481445, + "loss": 1.3944, + "theoretical_loss": 3.8837687608168237, + "tokens_seen": 577503232 + }, + { + "epoch": 0.18, + "learning_rate": 0.0006666559679037111, + "loss": 1.5022, + "theoretical_loss": 3.8837228048964896, + "tokens_seen": 577568768 + }, + { + "epoch": 0.18, + "learning_rate": 0.0006666399197592779, + "loss": 1.5502, + "theoretical_loss": 3.8836768556503083, + "tokens_seen": 577634304 + }, + { + "epoch": 0.18, + "learning_rate": 0.0006666238716148445, + "loss": 1.5724, + "theoretical_loss": 3.883630913076553, + "tokens_seen": 577699840 + }, + { + "epoch": 0.18, + "learning_rate": 0.0006666078234704113, + "loss": 1.5941, + "theoretical_loss": 3.883584977173499, + "tokens_seen": 577765376 + }, + { + "epoch": 0.18, + "learning_rate": 0.0006665917753259779, + "loss": 1.5679, + "theoretical_loss": 3.8835390479394203, + "tokens_seen": 577830912 + }, + { + "epoch": 0.18, + "learning_rate": 0.0006665757271815447, + "loss": 1.4378, + "theoretical_loss": 3.883493125372592, + "tokens_seen": 577896448 + }, + { + "epoch": 0.18, + "learning_rate": 0.0006665596790371113, + "loss": 1.5684, + "theoretical_loss": 3.8834472094712913, + "tokens_seen": 577961984 + }, + { + "epoch": 0.18, + "learning_rate": 0.0006665436308926781, + "loss": 1.635, + "theoretical_loss": 3.8834013002337953, + "tokens_seen": 578027520 + }, + { + "epoch": 0.18, + "learning_rate": 0.0006665275827482447, + "loss": 1.5058, + "theoretical_loss": 3.883355397658381, + "tokens_seen": 578093056 + }, + { + "epoch": 0.18, + "learning_rate": 0.0006665115346038115, + "loss": 1.5759, + "theoretical_loss": 3.8833095017433257, + "tokens_seen": 578158592 + }, + { + "epoch": 0.18, + "learning_rate": 0.0006664954864593781, + "loss": 1.5276, + "theoretical_loss": 3.8832636124869095, + "tokens_seen": 578224128 + }, + { + "epoch": 0.18, + "learning_rate": 0.0006664794383149449, + "loss": 1.5593, + "theoretical_loss": 3.883217729887411, + "tokens_seen": 578289664 + }, + { + "epoch": 0.18, + "learning_rate": 0.0006664633901705115, + "loss": 1.5631, + "theoretical_loss": 3.883171853943111, + "tokens_seen": 578355200 + }, + { + "epoch": 0.18, + "learning_rate": 0.0006664473420260783, + "loss": 1.6674, + "theoretical_loss": 3.8831259846522896, + "tokens_seen": 578420736 + }, + { + "epoch": 0.18, + "learning_rate": 0.0006664312938816449, + "loss": 1.4532, + "theoretical_loss": 3.8830801220132285, + "tokens_seen": 578486272 + }, + { + "epoch": 0.18, + "learning_rate": 0.0006664152457372117, + "loss": 1.5326, + "theoretical_loss": 3.883034266024209, + "tokens_seen": 578551808 + }, + { + "epoch": 0.18, + "learning_rate": 0.0006663991975927783, + "loss": 1.4378, + "theoretical_loss": 3.8829884166835154, + "tokens_seen": 578617344 + }, + { + "epoch": 0.18, + "learning_rate": 0.0006663831494483451, + "loss": 1.5423, + "theoretical_loss": 3.882942573989429, + "tokens_seen": 578682880 + }, + { + "epoch": 0.18, + "learning_rate": 0.0006663671013039117, + "loss": 1.5549, + "theoretical_loss": 3.882896737940235, + "tokens_seen": 578748416 + }, + { + "epoch": 0.18, + "learning_rate": 0.0006663510531594785, + "loss": 1.4676, + "theoretical_loss": 3.882850908534217, + "tokens_seen": 578813952 + }, + { + "epoch": 0.18, + "learning_rate": 0.0006663350050150451, + "loss": 1.5943, + "theoretical_loss": 3.882805085769661, + "tokens_seen": 578879488 + }, + { + "epoch": 0.18, + "learning_rate": 0.0006663189568706119, + "loss": 1.6222, + "theoretical_loss": 3.8827592696448523, + "tokens_seen": 578945024 + }, + { + "epoch": 0.18, + "learning_rate": 0.0006663029087261785, + "loss": 1.5442, + "theoretical_loss": 3.8827134601580777, + "tokens_seen": 579010560 + }, + { + "epoch": 0.18, + "learning_rate": 0.0006662868605817453, + "loss": 1.4954, + "theoretical_loss": 3.8826676573076235, + "tokens_seen": 579076096 + }, + { + "epoch": 0.18, + "learning_rate": 0.0006662708124373119, + "loss": 1.5486, + "theoretical_loss": 3.8826218610917786, + "tokens_seen": 579141632 + }, + { + "epoch": 0.18, + "learning_rate": 0.0006662547642928787, + "loss": 1.58, + "theoretical_loss": 3.8825760715088307, + "tokens_seen": 579207168 + }, + { + "epoch": 0.18, + "learning_rate": 0.0006662387161484453, + "loss": 1.4693, + "theoretical_loss": 3.882530288557069, + "tokens_seen": 579272704 + }, + { + "epoch": 0.18, + "learning_rate": 0.0006662226680040121, + "loss": 1.5769, + "theoretical_loss": 3.8824845122347824, + "tokens_seen": 579338240 + }, + { + "epoch": 0.18, + "learning_rate": 0.0006662066198595787, + "loss": 1.5206, + "theoretical_loss": 3.8824387425402618, + "tokens_seen": 579403776 + }, + { + "epoch": 0.18, + "learning_rate": 0.0006661905717151455, + "loss": 1.4946, + "theoretical_loss": 3.8823929794717986, + "tokens_seen": 579469312 + }, + { + "epoch": 0.18, + "learning_rate": 0.0006661745235707121, + "loss": 1.6544, + "theoretical_loss": 3.882347223027683, + "tokens_seen": 579534848 + }, + { + "epoch": 0.18, + "learning_rate": 0.0006661584754262789, + "loss": 1.5364, + "theoretical_loss": 3.882301473206209, + "tokens_seen": 579600384 + }, + { + "epoch": 0.18, + "learning_rate": 0.0006661424272818455, + "loss": 1.5567, + "theoretical_loss": 3.8822557300056673, + "tokens_seen": 579665920 + }, + { + "epoch": 0.18, + "learning_rate": 0.0006661263791374123, + "loss": 1.5179, + "theoretical_loss": 3.882209993424353, + "tokens_seen": 579731456 + }, + { + "epoch": 0.18, + "learning_rate": 0.0006661103309929789, + "loss": 1.5202, + "theoretical_loss": 3.8821642634605587, + "tokens_seen": 579796992 + }, + { + "epoch": 0.18, + "learning_rate": 0.0006660942828485457, + "loss": 1.5807, + "theoretical_loss": 3.8821185401125806, + "tokens_seen": 579862528 + }, + { + "epoch": 0.18, + "learning_rate": 0.0006660782347041123, + "loss": 1.5021, + "theoretical_loss": 3.8820728233787136, + "tokens_seen": 579928064 + }, + { + "epoch": 0.18, + "objective/train/docs_used": 216668, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.6637428998947144, + "objective/train/theoretical_loss": 3.8820271132572524, + "objective/train/tokens_used": 600453600, + "theoretical_loss": 3.8820271132572524, + "tokens_seen": 579993600 + }, + { + "epoch": 0.18, + "learning_rate": 0.0006660621865596791, + "loss": 1.604, + "theoretical_loss": 3.8820271132572524, + "tokens_seen": 579993600 + }, + { + "epoch": 0.18, + "learning_rate": 0.0006660461384152457, + "loss": 1.6311, + "theoretical_loss": 3.881981409746495, + "tokens_seen": 580059136 + }, + { + "epoch": 0.18, + "learning_rate": 0.0006660300902708125, + "loss": 1.5264, + "theoretical_loss": 3.8819357128447383, + "tokens_seen": 580124672 + }, + { + "epoch": 0.18, + "learning_rate": 0.0006660140421263791, + "loss": 1.4838, + "theoretical_loss": 3.88189002255028, + "tokens_seen": 580190208 + }, + { + "epoch": 0.18, + "learning_rate": 0.0006659979939819459, + "loss": 1.5524, + "theoretical_loss": 3.8818443388614186, + "tokens_seen": 580255744 + }, + { + "epoch": 0.18, + "learning_rate": 0.0006659819458375125, + "loss": 1.5313, + "theoretical_loss": 3.881798661776453, + "tokens_seen": 580321280 + }, + { + "epoch": 0.18, + "learning_rate": 0.0006659658976930793, + "loss": 1.5947, + "theoretical_loss": 3.8817529912936832, + "tokens_seen": 580386816 + }, + { + "epoch": 0.18, + "learning_rate": 0.0006659498495486459, + "loss": 1.5387, + "theoretical_loss": 3.88170732741141, + "tokens_seen": 580452352 + }, + { + "epoch": 0.18, + "learning_rate": 0.0006659338014042127, + "loss": 1.6039, + "theoretical_loss": 3.8816616701279334, + "tokens_seen": 580517888 + }, + { + "epoch": 0.18, + "learning_rate": 0.0006659177532597793, + "loss": 1.5464, + "theoretical_loss": 3.881616019441556, + "tokens_seen": 580583424 + }, + { + "epoch": 0.18, + "learning_rate": 0.0006659017051153461, + "loss": 1.4593, + "theoretical_loss": 3.881570375350579, + "tokens_seen": 580648960 + }, + { + "epoch": 0.18, + "learning_rate": 0.0006658856569709127, + "loss": 1.5986, + "theoretical_loss": 3.881524737853306, + "tokens_seen": 580714496 + }, + { + "epoch": 0.18, + "learning_rate": 0.0006658696088264795, + "loss": 1.5968, + "theoretical_loss": 3.881479106948041, + "tokens_seen": 580780032 + }, + { + "epoch": 0.18, + "learning_rate": 0.0006658535606820461, + "loss": 1.4646, + "theoretical_loss": 3.8814334826330876, + "tokens_seen": 580845568 + }, + { + "epoch": 0.18, + "learning_rate": 0.0006658375125376129, + "loss": 1.5846, + "theoretical_loss": 3.88138786490675, + "tokens_seen": 580911104 + }, + { + "epoch": 0.18, + "learning_rate": 0.0006658214643931795, + "loss": 1.5256, + "theoretical_loss": 3.8813422537673343, + "tokens_seen": 580976640 + }, + { + "epoch": 0.18, + "learning_rate": 0.0006658054162487463, + "loss": 1.5558, + "theoretical_loss": 3.881296649213146, + "tokens_seen": 581042176 + }, + { + "epoch": 0.18, + "learning_rate": 0.0006657893681043129, + "loss": 1.5059, + "theoretical_loss": 3.8812510512424927, + "tokens_seen": 581107712 + }, + { + "epoch": 0.18, + "learning_rate": 0.0006657733199598797, + "loss": 1.6923, + "theoretical_loss": 3.881205459853681, + "tokens_seen": 581173248 + }, + { + "epoch": 0.18, + "learning_rate": 0.0006657572718154463, + "loss": 1.5314, + "theoretical_loss": 3.8811598750450185, + "tokens_seen": 581238784 + }, + { + "epoch": 0.18, + "learning_rate": 0.0006657412236710131, + "loss": 1.5725, + "theoretical_loss": 3.8811142968148142, + "tokens_seen": 581304320 + }, + { + "epoch": 0.18, + "learning_rate": 0.0006657251755265797, + "loss": 1.4883, + "theoretical_loss": 3.881068725161377, + "tokens_seen": 581369856 + }, + { + "epoch": 0.18, + "learning_rate": 0.0006657091273821465, + "loss": 1.4967, + "theoretical_loss": 3.881023160083018, + "tokens_seen": 581435392 + }, + { + "epoch": 0.18, + "learning_rate": 0.0006656930792377131, + "loss": 1.5737, + "theoretical_loss": 3.8809776015780453, + "tokens_seen": 581500928 + }, + { + "epoch": 0.18, + "learning_rate": 0.0006656770310932799, + "loss": 1.5029, + "theoretical_loss": 3.880932049644771, + "tokens_seen": 581566464 + }, + { + "epoch": 0.18, + "learning_rate": 0.0006656609829488466, + "loss": 1.5279, + "theoretical_loss": 3.8808865042815075, + "tokens_seen": 581632000 + }, + { + "epoch": 0.18, + "learning_rate": 0.0006656449348044133, + "loss": 1.4966, + "theoretical_loss": 3.880840965486566, + "tokens_seen": 581697536 + }, + { + "epoch": 0.18, + "learning_rate": 0.00066562888665998, + "loss": 1.4883, + "theoretical_loss": 3.8807954332582595, + "tokens_seen": 581763072 + }, + { + "epoch": 0.18, + "learning_rate": 0.0006656128385155467, + "loss": 1.608, + "theoretical_loss": 3.880749907594902, + "tokens_seen": 581828608 + }, + { + "epoch": 0.18, + "learning_rate": 0.0006655967903711134, + "loss": 1.5664, + "theoretical_loss": 3.8807043884948085, + "tokens_seen": 581894144 + }, + { + "epoch": 0.18, + "learning_rate": 0.0006655807422266801, + "loss": 1.568, + "theoretical_loss": 3.880658875956292, + "tokens_seen": 581959680 + }, + { + "epoch": 0.18, + "learning_rate": 0.0006655646940822468, + "loss": 1.5118, + "theoretical_loss": 3.8806133699776684, + "tokens_seen": 582025216 + }, + { + "epoch": 0.18, + "learning_rate": 0.0006655486459378135, + "loss": 1.5355, + "theoretical_loss": 3.880567870557254, + "tokens_seen": 582090752 + }, + { + "epoch": 0.18, + "learning_rate": 0.0006655325977933802, + "loss": 1.6621, + "theoretical_loss": 3.8805223776933655, + "tokens_seen": 582156288 + }, + { + "epoch": 0.18, + "learning_rate": 0.0006655165496489469, + "loss": 1.6947, + "theoretical_loss": 3.88047689138432, + "tokens_seen": 582221824 + }, + { + "epoch": 0.18, + "learning_rate": 0.0006655005015045136, + "loss": 1.4968, + "theoretical_loss": 3.880431411628436, + "tokens_seen": 582287360 + }, + { + "epoch": 0.18, + "learning_rate": 0.0006654844533600803, + "loss": 1.593, + "theoretical_loss": 3.8803859384240305, + "tokens_seen": 582352896 + }, + { + "epoch": 0.18, + "learning_rate": 0.000665468405215647, + "loss": 1.4887, + "theoretical_loss": 3.880340471769424, + "tokens_seen": 582418432 + }, + { + "epoch": 0.18, + "learning_rate": 0.0006654523570712137, + "loss": 1.5931, + "theoretical_loss": 3.880295011662936, + "tokens_seen": 582483968 + }, + { + "epoch": 0.18, + "learning_rate": 0.0006654363089267805, + "loss": 1.517, + "theoretical_loss": 3.8802495581028857, + "tokens_seen": 582549504 + }, + { + "epoch": 0.18, + "learning_rate": 0.0006654202607823471, + "loss": 1.7181, + "theoretical_loss": 3.880204111087596, + "tokens_seen": 582615040 + }, + { + "epoch": 0.18, + "learning_rate": 0.0006654042126379139, + "loss": 1.5413, + "theoretical_loss": 3.880158670615387, + "tokens_seen": 582680576 + }, + { + "epoch": 0.18, + "learning_rate": 0.0006653881644934805, + "loss": 1.5136, + "theoretical_loss": 3.880113236684582, + "tokens_seen": 582746112 + }, + { + "epoch": 0.18, + "learning_rate": 0.0006653721163490472, + "loss": 1.6162, + "theoretical_loss": 3.880067809293503, + "tokens_seen": 582811648 + }, + { + "epoch": 0.18, + "learning_rate": 0.0006653560682046139, + "loss": 1.6533, + "theoretical_loss": 3.880022388440474, + "tokens_seen": 582877184 + }, + { + "epoch": 0.18, + "learning_rate": 0.0006653400200601806, + "loss": 1.5603, + "theoretical_loss": 3.879976974123818, + "tokens_seen": 582942720 + }, + { + "epoch": 0.18, + "learning_rate": 0.0006653239719157473, + "loss": 1.586, + "theoretical_loss": 3.8799315663418614, + "tokens_seen": 583008256 + }, + { + "epoch": 0.18, + "learning_rate": 0.000665307923771314, + "loss": 1.4177, + "theoretical_loss": 3.8798861650929277, + "tokens_seen": 583073792 + }, + { + "epoch": 0.18, + "learning_rate": 0.0006652918756268807, + "loss": 1.5825, + "theoretical_loss": 3.8798407703753446, + "tokens_seen": 583139328 + }, + { + "epoch": 0.18, + "learning_rate": 0.0006652758274824474, + "loss": 1.5814, + "theoretical_loss": 3.879795382187438, + "tokens_seen": 583204864 + }, + { + "epoch": 0.18, + "objective/train/docs_used": 217803, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.6120554208755493, + "objective/train/theoretical_loss": 3.879750000527534, + "objective/train/tokens_used": 603730400, + "theoretical_loss": 3.879750000527534, + "tokens_seen": 583270400 + }, + { + "epoch": 0.18, + "learning_rate": 0.000665259779338014, + "loss": 1.5858, + "theoretical_loss": 3.879750000527534, + "tokens_seen": 583270400 + }, + { + "epoch": 0.18, + "learning_rate": 0.0006652437311935808, + "loss": 1.5372, + "theoretical_loss": 3.8797046253939618, + "tokens_seen": 583335936 + }, + { + "epoch": 0.18, + "learning_rate": 0.0006652276830491474, + "loss": 1.5924, + "theoretical_loss": 3.8796592567850485, + "tokens_seen": 583401472 + }, + { + "epoch": 0.18, + "learning_rate": 0.0006652116349047142, + "loss": 1.5496, + "theoretical_loss": 3.8796138946991245, + "tokens_seen": 583467008 + }, + { + "epoch": 0.18, + "learning_rate": 0.0006651955867602808, + "loss": 1.6644, + "theoretical_loss": 3.8795685391345183, + "tokens_seen": 583532544 + }, + { + "epoch": 0.18, + "learning_rate": 0.0006651795386158476, + "loss": 1.5525, + "theoretical_loss": 3.879523190089561, + "tokens_seen": 583598080 + }, + { + "epoch": 0.18, + "learning_rate": 0.0006651634904714142, + "loss": 1.5414, + "theoretical_loss": 3.8794778475625833, + "tokens_seen": 583663616 + }, + { + "epoch": 0.18, + "learning_rate": 0.000665147442326981, + "loss": 1.6299, + "theoretical_loss": 3.879432511551916, + "tokens_seen": 583729152 + }, + { + "epoch": 0.18, + "learning_rate": 0.0006651313941825476, + "loss": 1.4834, + "theoretical_loss": 3.8793871820558916, + "tokens_seen": 583794688 + }, + { + "epoch": 0.18, + "learning_rate": 0.0006651153460381144, + "loss": 1.5941, + "theoretical_loss": 3.8793418590728423, + "tokens_seen": 583860224 + }, + { + "epoch": 0.18, + "learning_rate": 0.000665099297893681, + "loss": 1.529, + "theoretical_loss": 3.879296542601103, + "tokens_seen": 583925760 + }, + { + "epoch": 0.18, + "learning_rate": 0.0006650832497492478, + "loss": 1.4861, + "theoretical_loss": 3.879251232639006, + "tokens_seen": 583991296 + }, + { + "epoch": 0.18, + "learning_rate": 0.0006650672016048144, + "loss": 1.5777, + "theoretical_loss": 3.879205929184886, + "tokens_seen": 584056832 + }, + { + "epoch": 0.18, + "learning_rate": 0.0006650511534603812, + "loss": 1.4783, + "theoretical_loss": 3.8791606322370793, + "tokens_seen": 584122368 + }, + { + "epoch": 0.18, + "learning_rate": 0.0006650351053159478, + "loss": 1.5006, + "theoretical_loss": 3.8791153417939204, + "tokens_seen": 584187904 + }, + { + "epoch": 0.18, + "learning_rate": 0.0006650190571715146, + "loss": 1.4908, + "theoretical_loss": 3.8790700578537463, + "tokens_seen": 584253440 + }, + { + "epoch": 0.18, + "learning_rate": 0.0006650030090270812, + "loss": 1.6616, + "theoretical_loss": 3.879024780414894, + "tokens_seen": 584318976 + }, + { + "epoch": 0.18, + "learning_rate": 0.000664986960882648, + "loss": 1.5618, + "theoretical_loss": 3.878979509475701, + "tokens_seen": 584384512 + }, + { + "epoch": 0.18, + "learning_rate": 0.0006649709127382146, + "loss": 1.5686, + "theoretical_loss": 3.8789342450345052, + "tokens_seen": 584450048 + }, + { + "epoch": 0.18, + "learning_rate": 0.0006649548645937814, + "loss": 1.5357, + "theoretical_loss": 3.8788889870896455, + "tokens_seen": 584515584 + }, + { + "epoch": 0.18, + "learning_rate": 0.000664938816449348, + "loss": 1.5095, + "theoretical_loss": 3.878843735639462, + "tokens_seen": 584581120 + }, + { + "epoch": 0.18, + "learning_rate": 0.0006649227683049148, + "loss": 1.4742, + "theoretical_loss": 3.8787984906822945, + "tokens_seen": 584646656 + }, + { + "epoch": 0.18, + "learning_rate": 0.0006649067201604814, + "loss": 1.5401, + "theoretical_loss": 3.878753252216483, + "tokens_seen": 584712192 + }, + { + "epoch": 0.18, + "learning_rate": 0.0006648906720160482, + "loss": 1.6162, + "theoretical_loss": 3.8787080202403694, + "tokens_seen": 584777728 + }, + { + "epoch": 0.18, + "learning_rate": 0.0006648746238716148, + "loss": 1.5208, + "theoretical_loss": 3.8786627947522954, + "tokens_seen": 584843264 + }, + { + "epoch": 0.18, + "learning_rate": 0.0006648585757271816, + "loss": 1.5928, + "theoretical_loss": 3.878617575750603, + "tokens_seen": 584908800 + }, + { + "epoch": 0.18, + "learning_rate": 0.0006648425275827482, + "loss": 1.5719, + "theoretical_loss": 3.8785723632336366, + "tokens_seen": 584974336 + }, + { + "epoch": 0.18, + "learning_rate": 0.000664826479438315, + "loss": 1.4603, + "theoretical_loss": 3.8785271571997386, + "tokens_seen": 585039872 + }, + { + "epoch": 0.18, + "learning_rate": 0.0006648104312938816, + "loss": 1.6372, + "theoretical_loss": 3.878481957647254, + "tokens_seen": 585105408 + }, + { + "epoch": 0.18, + "learning_rate": 0.0006647943831494484, + "loss": 1.4895, + "theoretical_loss": 3.878436764574527, + "tokens_seen": 585170944 + }, + { + "epoch": 0.18, + "learning_rate": 0.000664778335005015, + "loss": 1.5645, + "theoretical_loss": 3.8783915779799036, + "tokens_seen": 585236480 + }, + { + "epoch": 0.18, + "learning_rate": 0.0006647622868605818, + "loss": 1.6121, + "theoretical_loss": 3.8783463978617303, + "tokens_seen": 585302016 + }, + { + "epoch": 0.18, + "learning_rate": 0.0006647462387161484, + "loss": 1.5353, + "theoretical_loss": 3.878301224218353, + "tokens_seen": 585367552 + }, + { + "epoch": 0.18, + "learning_rate": 0.0006647301905717152, + "loss": 1.5014, + "theoretical_loss": 3.8782560570481204, + "tokens_seen": 585433088 + }, + { + "epoch": 0.18, + "learning_rate": 0.0006647141424272818, + "loss": 1.5148, + "theoretical_loss": 3.8782108963493793, + "tokens_seen": 585498624 + }, + { + "epoch": 0.18, + "learning_rate": 0.0006646980942828486, + "loss": 1.6119, + "theoretical_loss": 3.878165742120478, + "tokens_seen": 585564160 + }, + { + "epoch": 0.18, + "learning_rate": 0.0006646820461384152, + "loss": 1.5141, + "theoretical_loss": 3.8781205943597667, + "tokens_seen": 585629696 + }, + { + "epoch": 0.18, + "learning_rate": 0.000664665997993982, + "loss": 1.5737, + "theoretical_loss": 3.8780754530655948, + "tokens_seen": 585695232 + }, + { + "epoch": 0.18, + "learning_rate": 0.0006646499498495486, + "loss": 1.4742, + "theoretical_loss": 3.8780303182363123, + "tokens_seen": 585760768 + }, + { + "epoch": 0.18, + "learning_rate": 0.0006646339017051154, + "loss": 1.5494, + "theoretical_loss": 3.877985189870271, + "tokens_seen": 585826304 + }, + { + "epoch": 0.18, + "learning_rate": 0.000664617853560682, + "loss": 1.515, + "theoretical_loss": 3.877940067965821, + "tokens_seen": 585891840 + }, + { + "epoch": 0.18, + "learning_rate": 0.0006646018054162488, + "loss": 1.4346, + "theoretical_loss": 3.8778949525213164, + "tokens_seen": 585957376 + }, + { + "epoch": 0.18, + "learning_rate": 0.0006645857572718154, + "loss": 1.5129, + "theoretical_loss": 3.8778498435351088, + "tokens_seen": 586022912 + }, + { + "epoch": 0.18, + "learning_rate": 0.0006645697091273822, + "loss": 1.624, + "theoretical_loss": 3.8778047410055514, + "tokens_seen": 586088448 + }, + { + "epoch": 0.18, + "learning_rate": 0.0006645536609829488, + "loss": 1.5166, + "theoretical_loss": 3.8777596449309994, + "tokens_seen": 586153984 + }, + { + "epoch": 0.18, + "learning_rate": 0.0006645376128385156, + "loss": 1.6359, + "theoretical_loss": 3.8777145553098062, + "tokens_seen": 586219520 + }, + { + "epoch": 0.18, + "learning_rate": 0.0006645215646940822, + "loss": 1.5646, + "theoretical_loss": 3.8776694721403278, + "tokens_seen": 586285056 + }, + { + "epoch": 0.18, + "learning_rate": 0.000664505516549649, + "loss": 1.507, + "theoretical_loss": 3.877624395420919, + "tokens_seen": 586350592 + }, + { + "epoch": 0.18, + "learning_rate": 0.0006644894684052156, + "loss": 1.4929, + "theoretical_loss": 3.8775793251499375, + "tokens_seen": 586416128 + }, + { + "epoch": 0.18, + "learning_rate": 0.0006644734202607824, + "loss": 1.5531, + "theoretical_loss": 3.8775342613257395, + "tokens_seen": 586481664 + }, + { + "epoch": 0.18, + "objective/train/docs_used": 218927, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.7062187194824219, + "objective/train/theoretical_loss": 3.8774892039466833, + "objective/train/tokens_used": 607007200, + "theoretical_loss": 3.8774892039466833, + "tokens_seen": 586547200 + }, + { + "epoch": 0.18, + "learning_rate": 0.000664457372116349, + "loss": 1.6635, + "theoretical_loss": 3.8774892039466833, + "tokens_seen": 586547200 + }, + { + "epoch": 0.18, + "learning_rate": 0.0006644413239719158, + "loss": 1.4991, + "theoretical_loss": 3.877444153011126, + "tokens_seen": 586612736 + }, + { + "epoch": 0.18, + "learning_rate": 0.0006644252758274824, + "loss": 1.502, + "theoretical_loss": 3.8773991085174284, + "tokens_seen": 586678272 + }, + { + "epoch": 0.18, + "learning_rate": 0.0006644092276830492, + "loss": 1.4851, + "theoretical_loss": 3.8773540704639475, + "tokens_seen": 586743808 + }, + { + "epoch": 0.18, + "learning_rate": 0.0006643931795386158, + "loss": 1.5761, + "theoretical_loss": 3.8773090388490448, + "tokens_seen": 586809344 + }, + { + "epoch": 0.18, + "learning_rate": 0.0006643771313941826, + "loss": 1.4779, + "theoretical_loss": 3.8772640136710805, + "tokens_seen": 586874880 + }, + { + "epoch": 0.18, + "learning_rate": 0.0006643610832497492, + "loss": 1.5292, + "theoretical_loss": 3.877218994928416, + "tokens_seen": 586940416 + }, + { + "epoch": 0.18, + "learning_rate": 0.000664345035105316, + "loss": 1.6483, + "theoretical_loss": 3.8771739826194134, + "tokens_seen": 587005952 + }, + { + "epoch": 0.18, + "learning_rate": 0.0006643289869608826, + "loss": 1.5545, + "theoretical_loss": 3.8771289767424344, + "tokens_seen": 587071488 + }, + { + "epoch": 0.18, + "learning_rate": 0.0006643129388164494, + "loss": 1.5969, + "theoretical_loss": 3.8770839772958428, + "tokens_seen": 587137024 + }, + { + "epoch": 0.18, + "learning_rate": 0.000664296890672016, + "loss": 1.5259, + "theoretical_loss": 3.877038984278002, + "tokens_seen": 587202560 + }, + { + "epoch": 0.18, + "learning_rate": 0.0006642808425275828, + "loss": 1.546, + "theoretical_loss": 3.8769939976872756, + "tokens_seen": 587268096 + }, + { + "epoch": 0.18, + "learning_rate": 0.0006642647943831495, + "loss": 1.4947, + "theoretical_loss": 3.8769490175220294, + "tokens_seen": 587333632 + }, + { + "epoch": 0.18, + "learning_rate": 0.0006642487462387162, + "loss": 1.4633, + "theoretical_loss": 3.876904043780628, + "tokens_seen": 587399168 + }, + { + "epoch": 0.18, + "learning_rate": 0.000664232698094283, + "loss": 1.5552, + "theoretical_loss": 3.8768590764614377, + "tokens_seen": 587464704 + }, + { + "epoch": 0.18, + "learning_rate": 0.0006642166499498496, + "loss": 1.5198, + "theoretical_loss": 3.8768141155628255, + "tokens_seen": 587530240 + }, + { + "epoch": 0.18, + "learning_rate": 0.0006642006018054164, + "loss": 1.5352, + "theoretical_loss": 3.8767691610831587, + "tokens_seen": 587595776 + }, + { + "epoch": 0.18, + "learning_rate": 0.000664184553660983, + "loss": 1.5273, + "theoretical_loss": 3.876724213020804, + "tokens_seen": 587661312 + }, + { + "epoch": 0.18, + "learning_rate": 0.0006641685055165498, + "loss": 1.5005, + "theoretical_loss": 3.876679271374131, + "tokens_seen": 587726848 + }, + { + "epoch": 0.18, + "learning_rate": 0.0006641524573721164, + "loss": 1.472, + "theoretical_loss": 3.8766343361415085, + "tokens_seen": 587792384 + }, + { + "epoch": 0.18, + "learning_rate": 0.0006641364092276832, + "loss": 1.5824, + "theoretical_loss": 3.876589407321305, + "tokens_seen": 587857920 + }, + { + "epoch": 0.18, + "learning_rate": 0.0006641203610832498, + "loss": 1.5213, + "theoretical_loss": 3.8765444849118924, + "tokens_seen": 587923456 + }, + { + "epoch": 0.18, + "learning_rate": 0.0006641043129388166, + "loss": 1.4755, + "theoretical_loss": 3.8764995689116404, + "tokens_seen": 587988992 + }, + { + "epoch": 0.18, + "learning_rate": 0.0006640882647943832, + "loss": 1.51, + "theoretical_loss": 3.8764546593189206, + "tokens_seen": 588054528 + }, + { + "epoch": 0.18, + "learning_rate": 0.00066407221664995, + "loss": 1.5701, + "theoretical_loss": 3.876409756132105, + "tokens_seen": 588120064 + }, + { + "epoch": 0.18, + "learning_rate": 0.0006640561685055166, + "loss": 1.5439, + "theoretical_loss": 3.8763648593495663, + "tokens_seen": 588185600 + }, + { + "epoch": 0.18, + "learning_rate": 0.0006640401203610834, + "loss": 1.5976, + "theoretical_loss": 3.8763199689696783, + "tokens_seen": 588251136 + }, + { + "epoch": 0.18, + "learning_rate": 0.00066402407221665, + "loss": 1.4822, + "theoretical_loss": 3.876275084990813, + "tokens_seen": 588316672 + }, + { + "epoch": 0.18, + "learning_rate": 0.0006640080240722168, + "loss": 1.3733, + "theoretical_loss": 3.876230207411347, + "tokens_seen": 588382208 + }, + { + "epoch": 0.18, + "learning_rate": 0.0006639919759277834, + "loss": 1.4232, + "theoretical_loss": 3.8761853362296534, + "tokens_seen": 588447744 + }, + { + "epoch": 0.18, + "learning_rate": 0.0006639759277833501, + "loss": 1.4737, + "theoretical_loss": 3.8761404714441086, + "tokens_seen": 588513280 + }, + { + "epoch": 0.18, + "learning_rate": 0.0006639598796389168, + "loss": 1.4766, + "theoretical_loss": 3.8760956130530886, + "tokens_seen": 588578816 + }, + { + "epoch": 0.18, + "learning_rate": 0.0006639438314944835, + "loss": 1.6052, + "theoretical_loss": 3.8760507610549704, + "tokens_seen": 588644352 + }, + { + "epoch": 0.18, + "learning_rate": 0.0006639277833500502, + "loss": 1.5843, + "theoretical_loss": 3.876005915448131, + "tokens_seen": 588709888 + }, + { + "epoch": 0.18, + "learning_rate": 0.0006639117352056169, + "loss": 1.4921, + "theoretical_loss": 3.8759610762309484, + "tokens_seen": 588775424 + }, + { + "epoch": 0.18, + "learning_rate": 0.0006638956870611836, + "loss": 1.6164, + "theoretical_loss": 3.875916243401801, + "tokens_seen": 588840960 + }, + { + "epoch": 0.18, + "learning_rate": 0.0006638796389167503, + "loss": 1.566, + "theoretical_loss": 3.8758714169590682, + "tokens_seen": 588906496 + }, + { + "epoch": 0.18, + "learning_rate": 0.000663863590772317, + "loss": 1.6208, + "theoretical_loss": 3.87582659690113, + "tokens_seen": 588972032 + }, + { + "epoch": 0.18, + "learning_rate": 0.0006638475426278837, + "loss": 1.6078, + "theoretical_loss": 3.875781783226366, + "tokens_seen": 589037568 + }, + { + "epoch": 0.18, + "learning_rate": 0.0006638314944834503, + "loss": 1.5425, + "theoretical_loss": 3.8757369759331572, + "tokens_seen": 589103104 + }, + { + "epoch": 0.18, + "learning_rate": 0.0006638154463390171, + "loss": 1.4546, + "theoretical_loss": 3.8756921750198856, + "tokens_seen": 589168640 + }, + { + "epoch": 0.18, + "learning_rate": 0.0006637993981945837, + "loss": 1.4605, + "theoretical_loss": 3.8756473804849323, + "tokens_seen": 589234176 + }, + { + "epoch": 0.18, + "learning_rate": 0.0006637833500501505, + "loss": 1.5468, + "theoretical_loss": 3.8756025923266813, + "tokens_seen": 589299712 + }, + { + "epoch": 0.18, + "learning_rate": 0.0006637673019057171, + "loss": 1.5947, + "theoretical_loss": 3.8755578105435147, + "tokens_seen": 589365248 + }, + { + "epoch": 0.18, + "learning_rate": 0.0006637512537612839, + "loss": 1.5204, + "theoretical_loss": 3.8755130351338165, + "tokens_seen": 589430784 + }, + { + "epoch": 0.18, + "learning_rate": 0.0006637352056168505, + "loss": 1.6894, + "theoretical_loss": 3.8754682660959716, + "tokens_seen": 589496320 + }, + { + "epoch": 0.18, + "learning_rate": 0.0006637191574724173, + "loss": 1.4984, + "theoretical_loss": 3.8754235034283653, + "tokens_seen": 589561856 + }, + { + "epoch": 0.18, + "learning_rate": 0.0006637031093279839, + "loss": 1.4946, + "theoretical_loss": 3.875378747129382, + "tokens_seen": 589627392 + }, + { + "epoch": 0.18, + "learning_rate": 0.0006636870611835507, + "loss": 1.5501, + "theoretical_loss": 3.8753339971974086, + "tokens_seen": 589692928 + }, + { + "epoch": 0.18, + "learning_rate": 0.0006636710130391173, + "loss": 1.5254, + "theoretical_loss": 3.8752892536308323, + "tokens_seen": 589758464 + }, + { + "epoch": 0.18, + "objective/train/docs_used": 220165, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.6023643016815186, + "objective/train/theoretical_loss": 3.87524451642804, + "objective/train/tokens_used": 610284000, + "theoretical_loss": 3.87524451642804, + "tokens_seen": 589824000 + }, + { + "epoch": 0.18, + "learning_rate": 0.0006636549648946841, + "loss": 1.6215, + "theoretical_loss": 3.87524451642804, + "tokens_seen": 589824000 + }, + { + "epoch": 0.18, + "learning_rate": 0.0006636389167502507, + "loss": 1.4987, + "theoretical_loss": 3.8751997855874194, + "tokens_seen": 589889536 + }, + { + "epoch": 0.18, + "learning_rate": 0.0006636228686058175, + "loss": 1.5576, + "theoretical_loss": 3.8751550611073595, + "tokens_seen": 589955072 + }, + { + "epoch": 0.18, + "learning_rate": 0.0006636068204613841, + "loss": 1.4903, + "theoretical_loss": 3.875110342986249, + "tokens_seen": 590020608 + }, + { + "epoch": 0.18, + "learning_rate": 0.0006635907723169509, + "loss": 1.5491, + "theoretical_loss": 3.875065631222478, + "tokens_seen": 590086144 + }, + { + "epoch": 0.18, + "learning_rate": 0.0006635747241725175, + "loss": 1.5856, + "theoretical_loss": 3.875020925814437, + "tokens_seen": 590151680 + }, + { + "epoch": 0.18, + "learning_rate": 0.0006635586760280843, + "loss": 1.7113, + "theoretical_loss": 3.8749762267605163, + "tokens_seen": 590217216 + }, + { + "epoch": 0.18, + "learning_rate": 0.0006635426278836509, + "loss": 1.5475, + "theoretical_loss": 3.8749315340591073, + "tokens_seen": 590282752 + }, + { + "epoch": 0.18, + "learning_rate": 0.0006635265797392177, + "loss": 1.5944, + "theoretical_loss": 3.874886847708603, + "tokens_seen": 590348288 + }, + { + "epoch": 0.18, + "learning_rate": 0.0006635105315947843, + "loss": 1.5232, + "theoretical_loss": 3.8748421677073956, + "tokens_seen": 590413824 + }, + { + "epoch": 0.18, + "learning_rate": 0.0006634944834503511, + "loss": 1.5244, + "theoretical_loss": 3.874797494053878, + "tokens_seen": 590479360 + }, + { + "epoch": 0.18, + "learning_rate": 0.0006634784353059177, + "loss": 1.5198, + "theoretical_loss": 3.8747528267464437, + "tokens_seen": 590544896 + }, + { + "epoch": 0.18, + "learning_rate": 0.0006634623871614845, + "loss": 1.3369, + "theoretical_loss": 3.8747081657834874, + "tokens_seen": 590610432 + }, + { + "epoch": 0.18, + "learning_rate": 0.0006634463390170511, + "loss": 1.5546, + "theoretical_loss": 3.874663511163405, + "tokens_seen": 590675968 + }, + { + "epoch": 0.18, + "learning_rate": 0.0006634302908726179, + "loss": 1.3767, + "theoretical_loss": 3.8746188628845912, + "tokens_seen": 590741504 + }, + { + "epoch": 0.18, + "learning_rate": 0.0006634142427281845, + "loss": 1.5319, + "theoretical_loss": 3.8745742209454423, + "tokens_seen": 590807040 + }, + { + "epoch": 0.18, + "learning_rate": 0.0006633981945837513, + "loss": 1.4895, + "theoretical_loss": 3.874529585344355, + "tokens_seen": 590872576 + }, + { + "epoch": 0.18, + "learning_rate": 0.0006633821464393179, + "loss": 1.5188, + "theoretical_loss": 3.874484956079727, + "tokens_seen": 590938112 + }, + { + "epoch": 0.18, + "learning_rate": 0.0006633660982948847, + "loss": 1.5143, + "theoretical_loss": 3.8744403331499555, + "tokens_seen": 591003648 + }, + { + "epoch": 0.18, + "learning_rate": 0.0006633500501504513, + "loss": 1.5582, + "theoretical_loss": 3.8743957165534395, + "tokens_seen": 591069184 + }, + { + "epoch": 0.18, + "learning_rate": 0.0006633340020060181, + "loss": 1.6534, + "theoretical_loss": 3.8743511062885783, + "tokens_seen": 591134720 + }, + { + "epoch": 0.18, + "learning_rate": 0.0006633179538615847, + "loss": 1.4503, + "theoretical_loss": 3.874306502353771, + "tokens_seen": 591200256 + }, + { + "epoch": 0.18, + "learning_rate": 0.0006633019057171515, + "loss": 1.5756, + "theoretical_loss": 3.8742619047474176, + "tokens_seen": 591265792 + }, + { + "epoch": 0.18, + "learning_rate": 0.0006632858575727181, + "loss": 1.5145, + "theoretical_loss": 3.87421731346792, + "tokens_seen": 591331328 + }, + { + "epoch": 0.18, + "learning_rate": 0.0006632698094282849, + "loss": 1.5447, + "theoretical_loss": 3.874172728513678, + "tokens_seen": 591396864 + }, + { + "epoch": 0.18, + "learning_rate": 0.0006632537612838515, + "loss": 1.5166, + "theoretical_loss": 3.874128149883095, + "tokens_seen": 591462400 + }, + { + "epoch": 0.18, + "learning_rate": 0.0006632377131394183, + "loss": 1.5798, + "theoretical_loss": 3.8740835775745732, + "tokens_seen": 591527936 + }, + { + "epoch": 0.18, + "learning_rate": 0.0006632216649949849, + "loss": 1.4855, + "theoretical_loss": 3.8740390115865155, + "tokens_seen": 591593472 + }, + { + "epoch": 0.18, + "learning_rate": 0.0006632056168505517, + "loss": 1.6562, + "theoretical_loss": 3.873994451917326, + "tokens_seen": 591659008 + }, + { + "epoch": 0.18, + "learning_rate": 0.0006631895687061183, + "loss": 1.3794, + "theoretical_loss": 3.8739498985654084, + "tokens_seen": 591724544 + }, + { + "epoch": 0.18, + "learning_rate": 0.0006631735205616851, + "loss": 1.5177, + "theoretical_loss": 3.873905351529168, + "tokens_seen": 591790080 + }, + { + "epoch": 0.18, + "learning_rate": 0.0006631574724172517, + "loss": 1.5469, + "theoretical_loss": 3.8738608108070096, + "tokens_seen": 591855616 + }, + { + "epoch": 0.18, + "learning_rate": 0.0006631414242728185, + "loss": 1.4492, + "theoretical_loss": 3.8738162763973407, + "tokens_seen": 591921152 + }, + { + "epoch": 0.18, + "learning_rate": 0.0006631253761283851, + "loss": 1.534, + "theoretical_loss": 3.8737717482985667, + "tokens_seen": 591986688 + }, + { + "epoch": 0.18, + "learning_rate": 0.0006631093279839519, + "loss": 1.4786, + "theoretical_loss": 3.8737272265090947, + "tokens_seen": 592052224 + }, + { + "epoch": 0.18, + "learning_rate": 0.0006630932798395185, + "loss": 1.4684, + "theoretical_loss": 3.8736827110273335, + "tokens_seen": 592117760 + }, + { + "epoch": 0.18, + "learning_rate": 0.0006630772316950853, + "loss": 1.5448, + "theoretical_loss": 3.8736382018516906, + "tokens_seen": 592183296 + }, + { + "epoch": 0.18, + "learning_rate": 0.000663061183550652, + "loss": 1.5049, + "theoretical_loss": 3.8735936989805744, + "tokens_seen": 592248832 + }, + { + "epoch": 0.18, + "learning_rate": 0.0006630451354062187, + "loss": 1.5921, + "theoretical_loss": 3.873549202412396, + "tokens_seen": 592314368 + }, + { + "epoch": 0.18, + "learning_rate": 0.0006630290872617855, + "loss": 1.5843, + "theoretical_loss": 3.8735047121455644, + "tokens_seen": 592379904 + }, + { + "epoch": 0.18, + "learning_rate": 0.0006630130391173521, + "loss": 1.6103, + "theoretical_loss": 3.873460228178491, + "tokens_seen": 592445440 + }, + { + "epoch": 0.18, + "learning_rate": 0.0006629969909729189, + "loss": 1.4315, + "theoretical_loss": 3.8734157505095856, + "tokens_seen": 592510976 + }, + { + "epoch": 0.18, + "learning_rate": 0.0006629809428284855, + "loss": 1.5224, + "theoretical_loss": 3.8733712791372614, + "tokens_seen": 592576512 + }, + { + "epoch": 0.18, + "learning_rate": 0.0006629648946840523, + "loss": 1.5171, + "theoretical_loss": 3.8733268140599306, + "tokens_seen": 592642048 + }, + { + "epoch": 0.18, + "learning_rate": 0.0006629488465396189, + "loss": 1.4759, + "theoretical_loss": 3.8732823552760056, + "tokens_seen": 592707584 + }, + { + "epoch": 0.18, + "learning_rate": 0.0006629327983951857, + "loss": 1.5308, + "theoretical_loss": 3.8732379027839, + "tokens_seen": 592773120 + }, + { + "epoch": 0.18, + "learning_rate": 0.0006629167502507523, + "loss": 1.5292, + "theoretical_loss": 3.8731934565820287, + "tokens_seen": 592838656 + }, + { + "epoch": 0.18, + "learning_rate": 0.0006629007021063191, + "loss": 1.4885, + "theoretical_loss": 3.8731490166688056, + "tokens_seen": 592904192 + }, + { + "epoch": 0.18, + "learning_rate": 0.0006628846539618857, + "loss": 1.4727, + "theoretical_loss": 3.873104583042646, + "tokens_seen": 592969728 + }, + { + "epoch": 0.18, + "learning_rate": 0.0006628686058174525, + "loss": 1.576, + "theoretical_loss": 3.8730601557019657, + "tokens_seen": 593035264 + }, + { + "epoch": 0.18, + "objective/train/docs_used": 221319, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.5747249126434326, + "objective/train/theoretical_loss": 3.873015734645182, + "objective/train/tokens_used": 613560800, + "theoretical_loss": 3.873015734645182, + "tokens_seen": 593100800 + }, + { + "epoch": 0.18, + "learning_rate": 0.0006628525576730191, + "loss": 1.5386, + "theoretical_loss": 3.873015734645182, + "tokens_seen": 593100800 + }, + { + "epoch": 0.18, + "learning_rate": 0.0006628365095285859, + "loss": 1.528, + "theoretical_loss": 3.872971319870711, + "tokens_seen": 593166336 + }, + { + "epoch": 0.18, + "learning_rate": 0.0006628204613841525, + "loss": 1.3572, + "theoretical_loss": 3.872926911376971, + "tokens_seen": 593231872 + }, + { + "epoch": 0.18, + "learning_rate": 0.0006628044132397193, + "loss": 1.5365, + "theoretical_loss": 3.8728825091623786, + "tokens_seen": 593297408 + }, + { + "epoch": 0.18, + "learning_rate": 0.0006627883650952859, + "loss": 1.5209, + "theoretical_loss": 3.872838113225354, + "tokens_seen": 593362944 + }, + { + "epoch": 0.18, + "learning_rate": 0.0006627723169508527, + "loss": 1.5372, + "theoretical_loss": 3.8727937235643157, + "tokens_seen": 593428480 + }, + { + "epoch": 0.18, + "learning_rate": 0.0006627562688064193, + "loss": 1.4541, + "theoretical_loss": 3.872749340177685, + "tokens_seen": 593494016 + }, + { + "epoch": 0.18, + "learning_rate": 0.0006627402206619861, + "loss": 1.5802, + "theoretical_loss": 3.87270496306388, + "tokens_seen": 593559552 + }, + { + "epoch": 0.18, + "learning_rate": 0.0006627241725175527, + "loss": 1.6086, + "theoretical_loss": 3.8726605922213233, + "tokens_seen": 593625088 + }, + { + "epoch": 0.18, + "learning_rate": 0.0006627081243731195, + "loss": 1.5209, + "theoretical_loss": 3.872616227648436, + "tokens_seen": 593690624 + }, + { + "epoch": 0.18, + "learning_rate": 0.0006626920762286861, + "loss": 1.4681, + "theoretical_loss": 3.8725718693436404, + "tokens_seen": 593756160 + }, + { + "epoch": 0.18, + "learning_rate": 0.0006626760280842529, + "loss": 1.4899, + "theoretical_loss": 3.8725275173053593, + "tokens_seen": 593821696 + }, + { + "epoch": 0.18, + "learning_rate": 0.0006626599799398195, + "loss": 1.4566, + "theoretical_loss": 3.872483171532015, + "tokens_seen": 593887232 + }, + { + "epoch": 0.18, + "learning_rate": 0.0006626439317953863, + "loss": 1.5294, + "theoretical_loss": 3.8724388320220333, + "tokens_seen": 593952768 + }, + { + "epoch": 0.18, + "learning_rate": 0.0006626278836509529, + "loss": 1.5155, + "theoretical_loss": 3.872394498773837, + "tokens_seen": 594018304 + }, + { + "epoch": 0.18, + "learning_rate": 0.0006626118355065197, + "loss": 1.5941, + "theoretical_loss": 3.8723501717858513, + "tokens_seen": 594083840 + }, + { + "epoch": 0.18, + "learning_rate": 0.0006625957873620863, + "loss": 1.4146, + "theoretical_loss": 3.872305851056502, + "tokens_seen": 594149376 + }, + { + "epoch": 0.18, + "learning_rate": 0.000662579739217653, + "loss": 1.5475, + "theoretical_loss": 3.8722615365842157, + "tokens_seen": 594214912 + }, + { + "epoch": 0.18, + "learning_rate": 0.0006625636910732197, + "loss": 1.6088, + "theoretical_loss": 3.8722172283674183, + "tokens_seen": 594280448 + }, + { + "epoch": 0.18, + "learning_rate": 0.0006625476429287864, + "loss": 1.5001, + "theoretical_loss": 3.8721729264045375, + "tokens_seen": 594345984 + }, + { + "epoch": 0.18, + "learning_rate": 0.0006625315947843531, + "loss": 1.4675, + "theoretical_loss": 3.872128630694001, + "tokens_seen": 594411520 + }, + { + "epoch": 0.18, + "learning_rate": 0.0006625155466399198, + "loss": 1.6737, + "theoretical_loss": 3.8720843412342374, + "tokens_seen": 594477056 + }, + { + "epoch": 0.18, + "learning_rate": 0.0006624994984954865, + "loss": 1.5599, + "theoretical_loss": 3.872040058023676, + "tokens_seen": 594542592 + }, + { + "epoch": 0.18, + "learning_rate": 0.0006624834503510532, + "loss": 1.5426, + "theoretical_loss": 3.871995781060745, + "tokens_seen": 594608128 + }, + { + "epoch": 0.18, + "learning_rate": 0.0006624674022066199, + "loss": 1.4421, + "theoretical_loss": 3.8719515103438757, + "tokens_seen": 594673664 + }, + { + "epoch": 0.18, + "learning_rate": 0.0006624513540621866, + "loss": 1.5173, + "theoretical_loss": 3.871907245871499, + "tokens_seen": 594739200 + }, + { + "epoch": 0.18, + "learning_rate": 0.0006624353059177532, + "loss": 1.6406, + "theoretical_loss": 3.871862987642045, + "tokens_seen": 594804736 + }, + { + "epoch": 0.18, + "learning_rate": 0.00066241925777332, + "loss": 1.4026, + "theoretical_loss": 3.871818735653947, + "tokens_seen": 594870272 + }, + { + "epoch": 0.18, + "learning_rate": 0.0006624032096288866, + "loss": 1.4206, + "theoretical_loss": 3.8717744899056354, + "tokens_seen": 594935808 + }, + { + "epoch": 0.18, + "learning_rate": 0.0006623871614844534, + "loss": 1.5414, + "theoretical_loss": 3.8717302503955446, + "tokens_seen": 595001344 + }, + { + "epoch": 0.18, + "learning_rate": 0.00066237111334002, + "loss": 1.518, + "theoretical_loss": 3.8716860171221086, + "tokens_seen": 595066880 + }, + { + "epoch": 0.18, + "learning_rate": 0.0006623550651955868, + "loss": 1.5192, + "theoretical_loss": 3.8716417900837596, + "tokens_seen": 595132416 + }, + { + "epoch": 0.18, + "learning_rate": 0.0006623390170511534, + "loss": 1.5493, + "theoretical_loss": 3.8715975692789337, + "tokens_seen": 595197952 + }, + { + "epoch": 0.18, + "learning_rate": 0.0006623229689067202, + "loss": 1.5061, + "theoretical_loss": 3.871553354706066, + "tokens_seen": 595263488 + }, + { + "epoch": 0.18, + "learning_rate": 0.0006623069207622868, + "loss": 1.6567, + "theoretical_loss": 3.871509146363592, + "tokens_seen": 595329024 + }, + { + "epoch": 0.18, + "learning_rate": 0.0006622908726178536, + "loss": 1.5316, + "theoretical_loss": 3.8714649442499476, + "tokens_seen": 595394560 + }, + { + "epoch": 0.18, + "learning_rate": 0.0006622748244734202, + "loss": 1.4784, + "theoretical_loss": 3.8714207483635703, + "tokens_seen": 595460096 + }, + { + "epoch": 0.18, + "learning_rate": 0.000662258776328987, + "loss": 1.4882, + "theoretical_loss": 3.8713765587028974, + "tokens_seen": 595525632 + }, + { + "epoch": 0.18, + "learning_rate": 0.0006622427281845536, + "loss": 1.5178, + "theoretical_loss": 3.871332375266367, + "tokens_seen": 595591168 + }, + { + "epoch": 0.18, + "learning_rate": 0.0006622266800401204, + "loss": 1.5376, + "theoretical_loss": 3.871288198052418, + "tokens_seen": 595656704 + }, + { + "epoch": 0.18, + "learning_rate": 0.000662210631895687, + "loss": 1.5576, + "theoretical_loss": 3.871244027059489, + "tokens_seen": 595722240 + }, + { + "epoch": 0.18, + "learning_rate": 0.0006621945837512538, + "loss": 1.3972, + "theoretical_loss": 3.8711998622860193, + "tokens_seen": 595787776 + }, + { + "epoch": 0.18, + "learning_rate": 0.0006621785356068204, + "loss": 1.6675, + "theoretical_loss": 3.8711557037304507, + "tokens_seen": 595853312 + }, + { + "epoch": 0.18, + "learning_rate": 0.0006621624874623872, + "loss": 1.5118, + "theoretical_loss": 3.8711115513912224, + "tokens_seen": 595918848 + }, + { + "epoch": 0.18, + "learning_rate": 0.0006621464393179538, + "loss": 1.4343, + "theoretical_loss": 3.8710674052667766, + "tokens_seen": 595984384 + }, + { + "epoch": 0.18, + "learning_rate": 0.0006621303911735206, + "loss": 1.5779, + "theoretical_loss": 3.8710232653555554, + "tokens_seen": 596049920 + }, + { + "epoch": 0.18, + "learning_rate": 0.0006621143430290872, + "loss": 1.4795, + "theoretical_loss": 3.870979131656001, + "tokens_seen": 596115456 + }, + { + "epoch": 0.18, + "learning_rate": 0.000662098294884654, + "loss": 1.6183, + "theoretical_loss": 3.870935004166557, + "tokens_seen": 596180992 + }, + { + "epoch": 0.18, + "learning_rate": 0.0006620822467402206, + "loss": 1.4299, + "theoretical_loss": 3.870890882885667, + "tokens_seen": 596246528 + }, + { + "epoch": 0.18, + "learning_rate": 0.0006620661985957874, + "loss": 1.5008, + "theoretical_loss": 3.8708467678117744, + "tokens_seen": 596312064 + }, + { + "epoch": 0.18, + "objective/train/docs_used": 222574, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.473783254623413, + "objective/train/theoretical_loss": 3.870802658943324, + "objective/train/tokens_used": 616837600, + "theoretical_loss": 3.870802658943324, + "tokens_seen": 596377600 + }, + { + "epoch": 0.18, + "learning_rate": 0.000662050150451354, + "loss": 1.4494, + "theoretical_loss": 3.870802658943324, + "tokens_seen": 596377600 + }, + { + "epoch": 0.18, + "learning_rate": 0.0006620341023069208, + "loss": 1.4641, + "theoretical_loss": 3.8707585562787625, + "tokens_seen": 596443136 + }, + { + "epoch": 0.18, + "learning_rate": 0.0006620180541624874, + "loss": 1.6722, + "theoretical_loss": 3.8707144598165346, + "tokens_seen": 596508672 + }, + { + "epoch": 0.18, + "learning_rate": 0.0006620020060180542, + "loss": 1.64, + "theoretical_loss": 3.8706703695550875, + "tokens_seen": 596574208 + }, + { + "epoch": 0.18, + "learning_rate": 0.0006619859578736208, + "loss": 1.3859, + "theoretical_loss": 3.8706262854928672, + "tokens_seen": 596639744 + }, + { + "epoch": 0.18, + "learning_rate": 0.0006619699097291876, + "loss": 1.5664, + "theoretical_loss": 3.8705822076283223, + "tokens_seen": 596705280 + }, + { + "epoch": 0.18, + "learning_rate": 0.0006619538615847542, + "loss": 1.4505, + "theoretical_loss": 3.870538135959901, + "tokens_seen": 596770816 + }, + { + "epoch": 0.18, + "learning_rate": 0.000661937813440321, + "loss": 1.4318, + "theoretical_loss": 3.8704940704860507, + "tokens_seen": 596836352 + }, + { + "epoch": 0.18, + "learning_rate": 0.0006619217652958876, + "loss": 1.5912, + "theoretical_loss": 3.8704500112052216, + "tokens_seen": 596901888 + }, + { + "epoch": 0.18, + "learning_rate": 0.0006619057171514544, + "loss": 1.539, + "theoretical_loss": 3.870405958115864, + "tokens_seen": 596967424 + }, + { + "epoch": 0.18, + "learning_rate": 0.000661889669007021, + "loss": 1.6504, + "theoretical_loss": 3.870361911216427, + "tokens_seen": 597032960 + }, + { + "epoch": 0.18, + "learning_rate": 0.0006618736208625878, + "loss": 1.603, + "theoretical_loss": 3.8703178705053625, + "tokens_seen": 597098496 + }, + { + "epoch": 0.18, + "learning_rate": 0.0006618575727181545, + "loss": 1.5407, + "theoretical_loss": 3.8702738359811217, + "tokens_seen": 597164032 + }, + { + "epoch": 0.18, + "learning_rate": 0.0006618415245737212, + "loss": 1.5396, + "theoretical_loss": 3.8702298076421564, + "tokens_seen": 597229568 + }, + { + "epoch": 0.18, + "learning_rate": 0.000661825476429288, + "loss": 1.5293, + "theoretical_loss": 3.87018578548692, + "tokens_seen": 597295104 + }, + { + "epoch": 0.18, + "learning_rate": 0.0006618094282848546, + "loss": 1.4045, + "theoretical_loss": 3.8701417695138645, + "tokens_seen": 597360640 + }, + { + "epoch": 0.18, + "learning_rate": 0.0006617933801404213, + "loss": 1.5224, + "theoretical_loss": 3.8700977597214443, + "tokens_seen": 597426176 + }, + { + "epoch": 0.18, + "learning_rate": 0.000661777331995988, + "loss": 1.509, + "theoretical_loss": 3.870053756108114, + "tokens_seen": 597491712 + }, + { + "epoch": 0.18, + "learning_rate": 0.0006617612838515548, + "loss": 1.413, + "theoretical_loss": 3.8700097586723277, + "tokens_seen": 597557248 + }, + { + "epoch": 0.18, + "learning_rate": 0.0006617452357071214, + "loss": 1.5961, + "theoretical_loss": 3.869965767412541, + "tokens_seen": 597622784 + }, + { + "epoch": 0.18, + "learning_rate": 0.0006617291875626882, + "loss": 1.5317, + "theoretical_loss": 3.8699217823272103, + "tokens_seen": 597688320 + }, + { + "epoch": 0.18, + "learning_rate": 0.0006617131394182548, + "loss": 1.662, + "theoretical_loss": 3.8698778034147914, + "tokens_seen": 597753856 + }, + { + "epoch": 0.18, + "learning_rate": 0.0006616970912738216, + "loss": 1.5407, + "theoretical_loss": 3.869833830673742, + "tokens_seen": 597819392 + }, + { + "epoch": 0.18, + "learning_rate": 0.0006616810431293882, + "loss": 1.549, + "theoretical_loss": 3.8697898641025192, + "tokens_seen": 597884928 + }, + { + "epoch": 0.18, + "learning_rate": 0.000661664994984955, + "loss": 1.6671, + "theoretical_loss": 3.8697459036995814, + "tokens_seen": 597950464 + }, + { + "epoch": 0.18, + "learning_rate": 0.0006616489468405216, + "loss": 1.4526, + "theoretical_loss": 3.869701949463387, + "tokens_seen": 598016000 + }, + { + "epoch": 0.18, + "learning_rate": 0.0006616328986960884, + "loss": 1.556, + "theoretical_loss": 3.869658001392396, + "tokens_seen": 598081536 + }, + { + "epoch": 0.18, + "learning_rate": 0.000661616850551655, + "loss": 1.604, + "theoretical_loss": 3.869614059485068, + "tokens_seen": 598147072 + }, + { + "epoch": 0.18, + "learning_rate": 0.0006616008024072218, + "loss": 1.6048, + "theoretical_loss": 3.8695701237398623, + "tokens_seen": 598212608 + }, + { + "epoch": 0.18, + "learning_rate": 0.0006615847542627884, + "loss": 1.6318, + "theoretical_loss": 3.869526194155241, + "tokens_seen": 598278144 + }, + { + "epoch": 0.18, + "learning_rate": 0.0006615687061183552, + "loss": 1.4689, + "theoretical_loss": 3.869482270729665, + "tokens_seen": 598343680 + }, + { + "epoch": 0.18, + "learning_rate": 0.0006615526579739218, + "loss": 1.6061, + "theoretical_loss": 3.8694383534615966, + "tokens_seen": 598409216 + }, + { + "epoch": 0.18, + "learning_rate": 0.0006615366098294886, + "loss": 1.5479, + "theoretical_loss": 3.8693944423494986, + "tokens_seen": 598474752 + }, + { + "epoch": 0.18, + "learning_rate": 0.0006615205616850552, + "loss": 1.5326, + "theoretical_loss": 3.869350537391834, + "tokens_seen": 598540288 + }, + { + "epoch": 0.18, + "learning_rate": 0.000661504513540622, + "loss": 1.4462, + "theoretical_loss": 3.869306638587066, + "tokens_seen": 598605824 + }, + { + "epoch": 0.18, + "learning_rate": 0.0006614884653961886, + "loss": 1.5684, + "theoretical_loss": 3.8692627459336597, + "tokens_seen": 598671360 + }, + { + "epoch": 0.18, + "learning_rate": 0.0006614724172517554, + "loss": 1.5517, + "theoretical_loss": 3.8692188594300787, + "tokens_seen": 598736896 + }, + { + "epoch": 0.18, + "learning_rate": 0.000661456369107322, + "loss": 1.4888, + "theoretical_loss": 3.8691749790747894, + "tokens_seen": 598802432 + }, + { + "epoch": 0.18, + "learning_rate": 0.0006614403209628888, + "loss": 1.4601, + "theoretical_loss": 3.8691311048662573, + "tokens_seen": 598867968 + }, + { + "epoch": 0.18, + "learning_rate": 0.0006614242728184554, + "loss": 1.5157, + "theoretical_loss": 3.869087236802949, + "tokens_seen": 598933504 + }, + { + "epoch": 0.18, + "learning_rate": 0.0006614082246740222, + "loss": 1.5441, + "theoretical_loss": 3.869043374883331, + "tokens_seen": 598999040 + }, + { + "epoch": 0.18, + "learning_rate": 0.0006613921765295888, + "loss": 1.5236, + "theoretical_loss": 3.8689995191058717, + "tokens_seen": 599064576 + }, + { + "epoch": 0.18, + "learning_rate": 0.0006613761283851556, + "loss": 1.5734, + "theoretical_loss": 3.868955669469039, + "tokens_seen": 599130112 + }, + { + "epoch": 0.18, + "learning_rate": 0.0006613600802407222, + "loss": 1.5765, + "theoretical_loss": 3.868911825971301, + "tokens_seen": 599195648 + }, + { + "epoch": 0.18, + "learning_rate": 0.000661344032096289, + "loss": 1.6146, + "theoretical_loss": 3.868867988611127, + "tokens_seen": 599261184 + }, + { + "epoch": 0.18, + "learning_rate": 0.0006613279839518556, + "loss": 1.6667, + "theoretical_loss": 3.868824157386987, + "tokens_seen": 599326720 + }, + { + "epoch": 0.18, + "learning_rate": 0.0006613119358074224, + "loss": 1.4695, + "theoretical_loss": 3.8687803322973506, + "tokens_seen": 599392256 + }, + { + "epoch": 0.18, + "learning_rate": 0.000661295887662989, + "loss": 1.4763, + "theoretical_loss": 3.86873651334069, + "tokens_seen": 599457792 + }, + { + "epoch": 0.18, + "learning_rate": 0.0006612798395185558, + "loss": 1.5218, + "theoretical_loss": 3.8686927005154756, + "tokens_seen": 599523328 + }, + { + "epoch": 0.18, + "learning_rate": 0.0006612637913741224, + "loss": 1.4709, + "theoretical_loss": 3.8686488938201795, + "tokens_seen": 599588864 + }, + { + "epoch": 0.18, + "objective/train/docs_used": 223680, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.604945421218872, + "objective/train/theoretical_loss": 3.868605093253274, + "objective/train/tokens_used": 620114400, + "theoretical_loss": 3.868605093253274, + "tokens_seen": 599654400 + }, + { + "epoch": 0.18, + "learning_rate": 0.0006612477432296892, + "loss": 1.5103, + "theoretical_loss": 3.868605093253274, + "tokens_seen": 599654400 + }, + { + "epoch": 0.18, + "learning_rate": 0.0006612316950852558, + "loss": 1.5628, + "theoretical_loss": 3.8685612988132334, + "tokens_seen": 599719936 + }, + { + "epoch": 0.18, + "learning_rate": 0.0006612156469408226, + "loss": 1.6003, + "theoretical_loss": 3.8685175104985294, + "tokens_seen": 599785472 + }, + { + "epoch": 0.18, + "learning_rate": 0.0006611995987963892, + "loss": 1.4986, + "theoretical_loss": 3.868473728307637, + "tokens_seen": 599851008 + }, + { + "epoch": 0.18, + "learning_rate": 0.0006611835506519559, + "loss": 1.6014, + "theoretical_loss": 3.868429952239031, + "tokens_seen": 599916544 + }, + { + "epoch": 0.18, + "learning_rate": 0.0006611675025075226, + "loss": 1.5846, + "theoretical_loss": 3.8683861822911867, + "tokens_seen": 599982080 + }, + { + "epoch": 0.18, + "learning_rate": 0.0006611514543630893, + "loss": 1.4967, + "theoretical_loss": 3.8683424184625794, + "tokens_seen": 600047616 + }, + { + "epoch": 0.18, + "learning_rate": 0.000661135406218656, + "loss": 1.5056, + "theoretical_loss": 3.8682986607516856, + "tokens_seen": 600113152 + }, + { + "epoch": 0.18, + "learning_rate": 0.0006611193580742227, + "loss": 1.5956, + "theoretical_loss": 3.8682549091569824, + "tokens_seen": 600178688 + }, + { + "epoch": 0.18, + "learning_rate": 0.0006611033099297894, + "loss": 1.509, + "theoretical_loss": 3.868211163676947, + "tokens_seen": 600244224 + }, + { + "epoch": 0.18, + "learning_rate": 0.0006610872617853561, + "loss": 1.6833, + "theoretical_loss": 3.868167424310057, + "tokens_seen": 600309760 + }, + { + "epoch": 0.18, + "learning_rate": 0.0006610712136409228, + "loss": 1.5937, + "theoretical_loss": 3.868123691054792, + "tokens_seen": 600375296 + }, + { + "epoch": 0.18, + "learning_rate": 0.0006610551654964895, + "loss": 1.5046, + "theoretical_loss": 3.8680799639096297, + "tokens_seen": 600440832 + }, + { + "epoch": 0.18, + "learning_rate": 0.0006610391173520561, + "loss": 1.5477, + "theoretical_loss": 3.8680362428730506, + "tokens_seen": 600506368 + }, + { + "epoch": 0.18, + "learning_rate": 0.0006610230692076229, + "loss": 1.5931, + "theoretical_loss": 3.8679925279435343, + "tokens_seen": 600571904 + }, + { + "epoch": 0.18, + "learning_rate": 0.0006610070210631895, + "loss": 1.5224, + "theoretical_loss": 3.8679488191195617, + "tokens_seen": 600637440 + }, + { + "epoch": 0.18, + "learning_rate": 0.0006609909729187563, + "loss": 1.5843, + "theoretical_loss": 3.8679051163996143, + "tokens_seen": 600702976 + }, + { + "epoch": 0.18, + "learning_rate": 0.0006609749247743229, + "loss": 1.4796, + "theoretical_loss": 3.867861419782173, + "tokens_seen": 600768512 + }, + { + "epoch": 0.18, + "learning_rate": 0.0006609588766298897, + "loss": 1.6498, + "theoretical_loss": 3.867817729265721, + "tokens_seen": 600834048 + }, + { + "epoch": 0.18, + "learning_rate": 0.0006609428284854563, + "loss": 1.5907, + "theoretical_loss": 3.8677740448487414, + "tokens_seen": 600899584 + }, + { + "epoch": 0.18, + "learning_rate": 0.0006609267803410231, + "loss": 1.6157, + "theoretical_loss": 3.867730366529716, + "tokens_seen": 600965120 + }, + { + "epoch": 0.18, + "learning_rate": 0.0006609107321965897, + "loss": 1.4929, + "theoretical_loss": 3.86768669430713, + "tokens_seen": 601030656 + }, + { + "epoch": 0.18, + "learning_rate": 0.0006608946840521565, + "loss": 1.5655, + "theoretical_loss": 3.867643028179468, + "tokens_seen": 601096192 + }, + { + "epoch": 0.18, + "learning_rate": 0.0006608786359077231, + "loss": 1.5841, + "theoretical_loss": 3.867599368145214, + "tokens_seen": 601161728 + }, + { + "epoch": 0.18, + "learning_rate": 0.0006608625877632899, + "loss": 1.6179, + "theoretical_loss": 3.8675557142028545, + "tokens_seen": 601227264 + }, + { + "epoch": 0.18, + "learning_rate": 0.0006608465396188565, + "loss": 1.5291, + "theoretical_loss": 3.8675120663508755, + "tokens_seen": 601292800 + }, + { + "epoch": 0.18, + "learning_rate": 0.0006608304914744233, + "loss": 1.61, + "theoretical_loss": 3.867468424587763, + "tokens_seen": 601358336 + }, + { + "epoch": 0.18, + "learning_rate": 0.0006608144433299899, + "loss": 1.6556, + "theoretical_loss": 3.867424788912005, + "tokens_seen": 601423872 + }, + { + "epoch": 0.18, + "learning_rate": 0.0006607983951855567, + "loss": 1.5699, + "theoretical_loss": 3.8673811593220884, + "tokens_seen": 601489408 + }, + { + "epoch": 0.18, + "learning_rate": 0.0006607823470411233, + "loss": 1.4453, + "theoretical_loss": 3.8673375358165014, + "tokens_seen": 601554944 + }, + { + "epoch": 0.18, + "learning_rate": 0.0006607662988966901, + "loss": 1.513, + "theoretical_loss": 3.867293918393734, + "tokens_seen": 601620480 + }, + { + "epoch": 0.18, + "learning_rate": 0.0006607502507522567, + "loss": 1.4583, + "theoretical_loss": 3.8672503070522746, + "tokens_seen": 601686016 + }, + { + "epoch": 0.18, + "learning_rate": 0.0006607342026078235, + "loss": 1.5583, + "theoretical_loss": 3.8672067017906135, + "tokens_seen": 601751552 + }, + { + "epoch": 0.18, + "learning_rate": 0.0006607181544633901, + "loss": 1.4376, + "theoretical_loss": 3.8671631026072406, + "tokens_seen": 601817088 + }, + { + "epoch": 0.18, + "learning_rate": 0.0006607021063189569, + "loss": 1.6076, + "theoretical_loss": 3.8671195095006468, + "tokens_seen": 601882624 + }, + { + "epoch": 0.18, + "learning_rate": 0.0006606860581745235, + "loss": 1.6294, + "theoretical_loss": 3.8670759224693247, + "tokens_seen": 601948160 + }, + { + "epoch": 0.18, + "learning_rate": 0.0006606700100300903, + "loss": 1.4929, + "theoretical_loss": 3.867032341511765, + "tokens_seen": 602013696 + }, + { + "epoch": 0.18, + "learning_rate": 0.000660653961885657, + "loss": 1.5272, + "theoretical_loss": 3.866988766626461, + "tokens_seen": 602079232 + }, + { + "epoch": 0.18, + "learning_rate": 0.0006606379137412237, + "loss": 1.5976, + "theoretical_loss": 3.866945197811906, + "tokens_seen": 602144768 + }, + { + "epoch": 0.18, + "learning_rate": 0.0006606218655967904, + "loss": 1.3975, + "theoretical_loss": 3.866901635066593, + "tokens_seen": 602210304 + }, + { + "epoch": 0.18, + "learning_rate": 0.0006606058174523571, + "loss": 1.5746, + "theoretical_loss": 3.8668580783890167, + "tokens_seen": 602275840 + }, + { + "epoch": 0.18, + "learning_rate": 0.0006605897693079238, + "loss": 1.5118, + "theoretical_loss": 3.866814527777671, + "tokens_seen": 602341376 + }, + { + "epoch": 0.18, + "learning_rate": 0.0006605737211634905, + "loss": 1.5758, + "theoretical_loss": 3.8667709832310524, + "tokens_seen": 602406912 + }, + { + "epoch": 0.18, + "learning_rate": 0.0006605576730190572, + "loss": 1.6292, + "theoretical_loss": 3.866727444747656, + "tokens_seen": 602472448 + }, + { + "epoch": 0.18, + "learning_rate": 0.0006605416248746239, + "loss": 1.4693, + "theoretical_loss": 3.8666839123259784, + "tokens_seen": 602537984 + }, + { + "epoch": 0.18, + "learning_rate": 0.0006605255767301906, + "loss": 1.4931, + "theoretical_loss": 3.8666403859645158, + "tokens_seen": 602603520 + }, + { + "epoch": 0.18, + "learning_rate": 0.0006605095285857573, + "loss": 1.6084, + "theoretical_loss": 3.8665968656617666, + "tokens_seen": 602669056 + }, + { + "epoch": 0.18, + "learning_rate": 0.000660493480441324, + "loss": 1.4922, + "theoretical_loss": 3.8665533514162282, + "tokens_seen": 602734592 + }, + { + "epoch": 0.18, + "learning_rate": 0.0006604774322968907, + "loss": 1.5023, + "theoretical_loss": 3.8665098432263987, + "tokens_seen": 602800128 + }, + { + "epoch": 0.18, + "learning_rate": 0.0006604613841524574, + "loss": 1.5229, + "theoretical_loss": 3.866466341090778, + "tokens_seen": 602865664 + }, + { + "epoch": 0.18, + "objective/train/docs_used": 224736, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.4949688911437988, + "objective/train/theoretical_loss": 3.866422845007865, + "objective/train/tokens_used": 623391200, + "theoretical_loss": 3.866422845007865, + "tokens_seen": 602931200 + }, + { + "epoch": 0.18, + "learning_rate": 0.0006604453360080241, + "loss": 1.4879, + "theoretical_loss": 3.866422845007865, + "tokens_seen": 602931200 + }, + { + "epoch": 0.18, + "learning_rate": 0.0006604292878635908, + "loss": 1.5348, + "theoretical_loss": 3.86637935497616, + "tokens_seen": 602996736 + }, + { + "epoch": 0.18, + "learning_rate": 0.0006604132397191575, + "loss": 1.4358, + "theoretical_loss": 3.8663358709941633, + "tokens_seen": 603062272 + }, + { + "epoch": 0.18, + "learning_rate": 0.0006603971915747243, + "loss": 1.4837, + "theoretical_loss": 3.8662923930603768, + "tokens_seen": 603127808 + }, + { + "epoch": 0.18, + "learning_rate": 0.0006603811434302909, + "loss": 1.3449, + "theoretical_loss": 3.8662489211733018, + "tokens_seen": 603193344 + }, + { + "epoch": 0.18, + "learning_rate": 0.0006603650952858577, + "loss": 1.6166, + "theoretical_loss": 3.8662054553314404, + "tokens_seen": 603258880 + }, + { + "epoch": 0.18, + "learning_rate": 0.0006603490471414243, + "loss": 1.5247, + "theoretical_loss": 3.866161995533295, + "tokens_seen": 603324416 + }, + { + "epoch": 0.18, + "learning_rate": 0.000660332998996991, + "loss": 1.5391, + "theoretical_loss": 3.86611854177737, + "tokens_seen": 603389952 + }, + { + "epoch": 0.18, + "learning_rate": 0.0006603169508525577, + "loss": 1.5261, + "theoretical_loss": 3.866075094062168, + "tokens_seen": 603455488 + }, + { + "epoch": 0.18, + "learning_rate": 0.0006603009027081245, + "loss": 1.4812, + "theoretical_loss": 3.8660316523861944, + "tokens_seen": 603521024 + }, + { + "epoch": 0.18, + "learning_rate": 0.0006602848545636911, + "loss": 1.4541, + "theoretical_loss": 3.8659882167479527, + "tokens_seen": 603586560 + }, + { + "epoch": 0.18, + "learning_rate": 0.0006602688064192579, + "loss": 1.5745, + "theoretical_loss": 3.86594478714595, + "tokens_seen": 603652096 + }, + { + "epoch": 0.18, + "learning_rate": 0.0006602527582748245, + "loss": 1.4469, + "theoretical_loss": 3.8659013635786907, + "tokens_seen": 603717632 + }, + { + "epoch": 0.18, + "learning_rate": 0.0006602367101303913, + "loss": 1.572, + "theoretical_loss": 3.8658579460446827, + "tokens_seen": 603783168 + }, + { + "epoch": 0.18, + "learning_rate": 0.0006602206619859579, + "loss": 1.601, + "theoretical_loss": 3.865814534542432, + "tokens_seen": 603848704 + }, + { + "epoch": 0.18, + "learning_rate": 0.0006602046138415247, + "loss": 1.5419, + "theoretical_loss": 3.8657711290704473, + "tokens_seen": 603914240 + }, + { + "epoch": 0.18, + "learning_rate": 0.0006601885656970913, + "loss": 1.6105, + "theoretical_loss": 3.8657277296272348, + "tokens_seen": 603979776 + }, + { + "epoch": 0.18, + "learning_rate": 0.0006601725175526581, + "loss": 1.4834, + "theoretical_loss": 3.865684336211305, + "tokens_seen": 604045312 + }, + { + "epoch": 0.18, + "learning_rate": 0.0006601564694082247, + "loss": 1.5224, + "theoretical_loss": 3.8656409488211656, + "tokens_seen": 604110848 + }, + { + "epoch": 0.18, + "learning_rate": 0.0006601404212637915, + "loss": 1.5589, + "theoretical_loss": 3.8655975674553273, + "tokens_seen": 604176384 + }, + { + "epoch": 0.18, + "learning_rate": 0.0006601243731193581, + "loss": 1.4989, + "theoretical_loss": 3.8655541921123, + "tokens_seen": 604241920 + }, + { + "epoch": 0.18, + "learning_rate": 0.0006601083249749249, + "loss": 1.4454, + "theoretical_loss": 3.8655108227905943, + "tokens_seen": 604307456 + }, + { + "epoch": 0.18, + "learning_rate": 0.0006600922768304915, + "loss": 1.6272, + "theoretical_loss": 3.8654674594887215, + "tokens_seen": 604372992 + }, + { + "epoch": 0.18, + "learning_rate": 0.0006600762286860583, + "loss": 1.6195, + "theoretical_loss": 3.8654241022051927, + "tokens_seen": 604438528 + }, + { + "epoch": 0.18, + "learning_rate": 0.0006600601805416249, + "loss": 1.5605, + "theoretical_loss": 3.865380750938522, + "tokens_seen": 604504064 + }, + { + "epoch": 0.18, + "learning_rate": 0.0006600441323971917, + "loss": 1.5983, + "theoretical_loss": 3.86533740568722, + "tokens_seen": 604569600 + }, + { + "epoch": 0.18, + "learning_rate": 0.0006600280842527583, + "loss": 1.5592, + "theoretical_loss": 3.8652940664498026, + "tokens_seen": 604635136 + }, + { + "epoch": 0.18, + "learning_rate": 0.0006600120361083251, + "loss": 1.4996, + "theoretical_loss": 3.865250733224781, + "tokens_seen": 604700672 + }, + { + "epoch": 0.18, + "learning_rate": 0.0006599959879638917, + "loss": 1.5046, + "theoretical_loss": 3.865207406010672, + "tokens_seen": 604766208 + }, + { + "epoch": 0.18, + "learning_rate": 0.0006599799398194585, + "loss": 1.5428, + "theoretical_loss": 3.865164084805989, + "tokens_seen": 604831744 + }, + { + "epoch": 0.18, + "learning_rate": 0.0006599638916750251, + "loss": 1.5406, + "theoretical_loss": 3.865120769609248, + "tokens_seen": 604897280 + }, + { + "epoch": 0.18, + "learning_rate": 0.0006599478435305919, + "loss": 1.5098, + "theoretical_loss": 3.8650774604189655, + "tokens_seen": 604962816 + }, + { + "epoch": 0.18, + "learning_rate": 0.0006599317953861585, + "loss": 1.5995, + "theoretical_loss": 3.8650341572336573, + "tokens_seen": 605028352 + }, + { + "epoch": 0.18, + "learning_rate": 0.0006599157472417253, + "loss": 1.4006, + "theoretical_loss": 3.8649908600518406, + "tokens_seen": 605093888 + }, + { + "epoch": 0.18, + "learning_rate": 0.0006598996990972919, + "loss": 1.5283, + "theoretical_loss": 3.8649475688720334, + "tokens_seen": 605159424 + }, + { + "epoch": 0.18, + "learning_rate": 0.0006598836509528587, + "loss": 1.6671, + "theoretical_loss": 3.864904283692754, + "tokens_seen": 605224960 + }, + { + "epoch": 0.18, + "learning_rate": 0.0006598676028084253, + "loss": 1.53, + "theoretical_loss": 3.8648610045125196, + "tokens_seen": 605290496 + }, + { + "epoch": 0.18, + "learning_rate": 0.0006598515546639921, + "loss": 1.5975, + "theoretical_loss": 3.8648177313298513, + "tokens_seen": 605356032 + }, + { + "epoch": 0.18, + "learning_rate": 0.0006598355065195587, + "loss": 1.5842, + "theoretical_loss": 3.8647744641432675, + "tokens_seen": 605421568 + }, + { + "epoch": 0.18, + "learning_rate": 0.0006598194583751255, + "loss": 1.5067, + "theoretical_loss": 3.8647312029512886, + "tokens_seen": 605487104 + }, + { + "epoch": 0.18, + "learning_rate": 0.0006598034102306921, + "loss": 1.4965, + "theoretical_loss": 3.864687947752436, + "tokens_seen": 605552640 + }, + { + "epoch": 0.18, + "learning_rate": 0.0006597873620862588, + "loss": 1.6503, + "theoretical_loss": 3.8646446985452307, + "tokens_seen": 605618176 + }, + { + "epoch": 0.18, + "learning_rate": 0.0006597713139418255, + "loss": 1.5425, + "theoretical_loss": 3.864601455328194, + "tokens_seen": 605683712 + }, + { + "epoch": 0.18, + "learning_rate": 0.0006597552657973922, + "loss": 1.4937, + "theoretical_loss": 3.864558218099849, + "tokens_seen": 605749248 + }, + { + "epoch": 0.18, + "learning_rate": 0.0006597392176529589, + "loss": 1.5229, + "theoretical_loss": 3.864514986858718, + "tokens_seen": 605814784 + }, + { + "epoch": 0.18, + "learning_rate": 0.0006597231695085256, + "loss": 1.5616, + "theoretical_loss": 3.8644717616033244, + "tokens_seen": 605880320 + }, + { + "epoch": 0.18, + "learning_rate": 0.0006597071213640923, + "loss": 1.5166, + "theoretical_loss": 3.8644285423321927, + "tokens_seen": 605945856 + }, + { + "epoch": 0.18, + "learning_rate": 0.000659691073219659, + "loss": 1.4773, + "theoretical_loss": 3.864385329043847, + "tokens_seen": 606011392 + }, + { + "epoch": 0.18, + "learning_rate": 0.0006596750250752257, + "loss": 1.5813, + "theoretical_loss": 3.8643421217368115, + "tokens_seen": 606076928 + }, + { + "epoch": 0.18, + "learning_rate": 0.0006596589769307924, + "loss": 1.4449, + "theoretical_loss": 3.8642989204096123, + "tokens_seen": 606142464 + }, + { + "epoch": 0.18, + "objective/train/docs_used": 225863, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.5240535736083984, + "objective/train/theoretical_loss": 3.8642557250607754, + "objective/train/tokens_used": 626668000, + "theoretical_loss": 3.8642557250607754, + "tokens_seen": 606208000 + }, + { + "epoch": 0.18, + "learning_rate": 0.000659642928786359, + "loss": 1.4898, + "theoretical_loss": 3.8642557250607754, + "tokens_seen": 606208000 + }, + { + "epoch": 0.18, + "learning_rate": 0.0006596268806419258, + "loss": 1.5724, + "theoretical_loss": 3.8642125356888277, + "tokens_seen": 606273536 + }, + { + "epoch": 0.18, + "learning_rate": 0.0006596108324974924, + "loss": 1.5284, + "theoretical_loss": 3.8641693522922953, + "tokens_seen": 606339072 + }, + { + "epoch": 0.18, + "learning_rate": 0.0006595947843530592, + "loss": 1.6036, + "theoretical_loss": 3.8641261748697064, + "tokens_seen": 606404608 + }, + { + "epoch": 0.18, + "learning_rate": 0.0006595787362086258, + "loss": 1.5487, + "theoretical_loss": 3.8640830034195894, + "tokens_seen": 606470144 + }, + { + "epoch": 0.18, + "learning_rate": 0.0006595626880641926, + "loss": 1.621, + "theoretical_loss": 3.864039837940472, + "tokens_seen": 606535680 + }, + { + "epoch": 0.18, + "learning_rate": 0.0006595466399197592, + "loss": 1.524, + "theoretical_loss": 3.863996678430884, + "tokens_seen": 606601216 + }, + { + "epoch": 0.18, + "learning_rate": 0.000659530591775326, + "loss": 1.6004, + "theoretical_loss": 3.8639535248893555, + "tokens_seen": 606666752 + }, + { + "epoch": 0.18, + "learning_rate": 0.0006595145436308926, + "loss": 1.4823, + "theoretical_loss": 3.863910377314415, + "tokens_seen": 606732288 + }, + { + "epoch": 0.18, + "learning_rate": 0.0006594984954864594, + "loss": 1.5635, + "theoretical_loss": 3.863867235704595, + "tokens_seen": 606797824 + }, + { + "epoch": 0.18, + "learning_rate": 0.000659482447342026, + "loss": 1.526, + "theoretical_loss": 3.8638241000584257, + "tokens_seen": 606863360 + }, + { + "epoch": 0.18, + "learning_rate": 0.0006594663991975928, + "loss": 1.4859, + "theoretical_loss": 3.863780970374439, + "tokens_seen": 606928896 + }, + { + "epoch": 0.18, + "learning_rate": 0.0006594503510531595, + "loss": 1.3927, + "theoretical_loss": 3.863737846651167, + "tokens_seen": 606994432 + }, + { + "epoch": 0.18, + "learning_rate": 0.0006594343029087262, + "loss": 1.5374, + "theoretical_loss": 3.8636947288871424, + "tokens_seen": 607059968 + }, + { + "epoch": 0.18, + "learning_rate": 0.0006594182547642929, + "loss": 1.5469, + "theoretical_loss": 3.863651617080899, + "tokens_seen": 607125504 + }, + { + "epoch": 0.18, + "learning_rate": 0.0006594022066198596, + "loss": 1.4459, + "theoretical_loss": 3.8636085112309715, + "tokens_seen": 607191040 + }, + { + "epoch": 0.18, + "learning_rate": 0.0006593861584754263, + "loss": 1.5167, + "theoretical_loss": 3.863565411335892, + "tokens_seen": 607256576 + }, + { + "epoch": 0.18, + "learning_rate": 0.000659370110330993, + "loss": 1.5812, + "theoretical_loss": 3.8635223173941964, + "tokens_seen": 607322112 + }, + { + "epoch": 0.18, + "learning_rate": 0.0006593540621865597, + "loss": 1.4871, + "theoretical_loss": 3.8634792294044202, + "tokens_seen": 607387648 + }, + { + "epoch": 0.18, + "learning_rate": 0.0006593380140421264, + "loss": 1.4625, + "theoretical_loss": 3.8634361473650993, + "tokens_seen": 607453184 + }, + { + "epoch": 0.18, + "learning_rate": 0.0006593219658976931, + "loss": 1.6108, + "theoretical_loss": 3.8633930712747704, + "tokens_seen": 607518720 + }, + { + "epoch": 0.18, + "learning_rate": 0.0006593059177532598, + "loss": 1.5285, + "theoretical_loss": 3.8633500011319692, + "tokens_seen": 607584256 + }, + { + "epoch": 0.18, + "learning_rate": 0.0006592898696088265, + "loss": 1.6288, + "theoretical_loss": 3.8633069369352344, + "tokens_seen": 607649792 + }, + { + "epoch": 0.18, + "learning_rate": 0.0006592738214643932, + "loss": 1.5514, + "theoretical_loss": 3.8632638786831035, + "tokens_seen": 607715328 + }, + { + "epoch": 0.18, + "learning_rate": 0.0006592577733199599, + "loss": 1.6553, + "theoretical_loss": 3.863220826374115, + "tokens_seen": 607780864 + }, + { + "epoch": 0.18, + "learning_rate": 0.0006592417251755266, + "loss": 1.5877, + "theoretical_loss": 3.863177780006808, + "tokens_seen": 607846400 + }, + { + "epoch": 0.18, + "learning_rate": 0.0006592256770310933, + "loss": 1.5772, + "theoretical_loss": 3.8631347395797215, + "tokens_seen": 607911936 + }, + { + "epoch": 0.18, + "learning_rate": 0.00065920962888666, + "loss": 1.499, + "theoretical_loss": 3.8630917050913958, + "tokens_seen": 607977472 + }, + { + "epoch": 0.18, + "learning_rate": 0.0006591935807422267, + "loss": 1.5473, + "theoretical_loss": 3.863048676540372, + "tokens_seen": 608043008 + }, + { + "epoch": 0.18, + "learning_rate": 0.0006591775325977934, + "loss": 1.4075, + "theoretical_loss": 3.8630056539251902, + "tokens_seen": 608108544 + }, + { + "epoch": 0.18, + "learning_rate": 0.0006591614844533601, + "loss": 1.5214, + "theoretical_loss": 3.862962637244393, + "tokens_seen": 608174080 + }, + { + "epoch": 0.18, + "learning_rate": 0.0006591454363089268, + "loss": 1.5158, + "theoretical_loss": 3.862919626496521, + "tokens_seen": 608239616 + }, + { + "epoch": 0.18, + "learning_rate": 0.0006591293881644935, + "loss": 1.7075, + "theoretical_loss": 3.862876621680119, + "tokens_seen": 608305152 + }, + { + "epoch": 0.18, + "learning_rate": 0.0006591133400200602, + "loss": 1.5488, + "theoretical_loss": 3.862833622793728, + "tokens_seen": 608370688 + }, + { + "epoch": 0.18, + "learning_rate": 0.000659097291875627, + "loss": 1.525, + "theoretical_loss": 3.8627906298358923, + "tokens_seen": 608436224 + }, + { + "epoch": 0.18, + "learning_rate": 0.0006590812437311936, + "loss": 1.4264, + "theoretical_loss": 3.862747642805157, + "tokens_seen": 608501760 + }, + { + "epoch": 0.18, + "learning_rate": 0.0006590651955867603, + "loss": 1.624, + "theoretical_loss": 3.8627046617000653, + "tokens_seen": 608567296 + }, + { + "epoch": 0.18, + "learning_rate": 0.000659049147442327, + "loss": 1.4392, + "theoretical_loss": 3.862661686519163, + "tokens_seen": 608632832 + }, + { + "epoch": 0.18, + "learning_rate": 0.0006590330992978938, + "loss": 1.5242, + "theoretical_loss": 3.862618717260996, + "tokens_seen": 608698368 + }, + { + "epoch": 0.18, + "learning_rate": 0.0006590170511534604, + "loss": 1.5037, + "theoretical_loss": 3.8625757539241103, + "tokens_seen": 608763904 + }, + { + "epoch": 0.18, + "learning_rate": 0.0006590010030090272, + "loss": 1.5969, + "theoretical_loss": 3.862532796507053, + "tokens_seen": 608829440 + }, + { + "epoch": 0.18, + "learning_rate": 0.0006589849548645938, + "loss": 1.4881, + "theoretical_loss": 3.8624898450083704, + "tokens_seen": 608894976 + }, + { + "epoch": 0.18, + "learning_rate": 0.0006589689067201606, + "loss": 1.4859, + "theoretical_loss": 3.862446899426611, + "tokens_seen": 608960512 + }, + { + "epoch": 0.18, + "learning_rate": 0.0006589528585757272, + "loss": 1.5648, + "theoretical_loss": 3.8624039597603232, + "tokens_seen": 609026048 + }, + { + "epoch": 0.18, + "learning_rate": 0.000658936810431294, + "loss": 1.556, + "theoretical_loss": 3.862361026008055, + "tokens_seen": 609091584 + }, + { + "epoch": 0.18, + "learning_rate": 0.0006589207622868606, + "loss": 1.4537, + "theoretical_loss": 3.8623180981683563, + "tokens_seen": 609157120 + }, + { + "epoch": 0.18, + "learning_rate": 0.0006589047141424274, + "loss": 1.383, + "theoretical_loss": 3.8622751762397765, + "tokens_seen": 609222656 + }, + { + "epoch": 0.18, + "learning_rate": 0.000658888665997994, + "loss": 1.5581, + "theoretical_loss": 3.862232260220866, + "tokens_seen": 609288192 + }, + { + "epoch": 0.18, + "learning_rate": 0.0006588726178535608, + "loss": 1.5084, + "theoretical_loss": 3.8621893501101763, + "tokens_seen": 609353728 + }, + { + "epoch": 0.18, + "learning_rate": 0.0006588565697091274, + "loss": 1.5499, + "theoretical_loss": 3.8621464459062573, + "tokens_seen": 609419264 + }, + { + "epoch": 0.18, + "objective/train/docs_used": 227108, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.47286057472229, + "objective/train/theoretical_loss": 3.8621035476076617, + "objective/train/tokens_used": 629944800, + "theoretical_loss": 3.8621035476076617, + "tokens_seen": 609484800 + }, + { + "epoch": 0.18, + "learning_rate": 0.0006588405215646942, + "loss": 1.4771, + "theoretical_loss": 3.8621035476076617, + "tokens_seen": 609484800 + }, + { + "epoch": 0.18, + "learning_rate": 0.0006588244734202608, + "loss": 1.5489, + "theoretical_loss": 3.862060655212942, + "tokens_seen": 609550336 + }, + { + "epoch": 0.18, + "learning_rate": 0.0006588084252758276, + "loss": 1.4631, + "theoretical_loss": 3.8620177687206514, + "tokens_seen": 609615872 + }, + { + "epoch": 0.18, + "learning_rate": 0.0006587923771313942, + "loss": 1.4776, + "theoretical_loss": 3.861974888129342, + "tokens_seen": 609681408 + }, + { + "epoch": 0.18, + "learning_rate": 0.000658776328986961, + "loss": 1.5141, + "theoretical_loss": 3.8619320134375688, + "tokens_seen": 609746944 + }, + { + "epoch": 0.18, + "learning_rate": 0.0006587602808425276, + "loss": 1.5504, + "theoretical_loss": 3.861889144643885, + "tokens_seen": 609812480 + }, + { + "epoch": 0.18, + "learning_rate": 0.0006587442326980944, + "loss": 1.3485, + "theoretical_loss": 3.861846281746847, + "tokens_seen": 609878016 + }, + { + "epoch": 0.18, + "learning_rate": 0.000658728184553661, + "loss": 1.5351, + "theoretical_loss": 3.861803424745009, + "tokens_seen": 609943552 + }, + { + "epoch": 0.18, + "learning_rate": 0.0006587121364092278, + "loss": 1.5319, + "theoretical_loss": 3.8617605736369276, + "tokens_seen": 610009088 + }, + { + "epoch": 0.18, + "learning_rate": 0.0006586960882647944, + "loss": 1.6861, + "theoretical_loss": 3.8617177284211595, + "tokens_seen": 610074624 + }, + { + "epoch": 0.18, + "learning_rate": 0.0006586800401203612, + "loss": 1.4675, + "theoretical_loss": 3.8616748890962604, + "tokens_seen": 610140160 + }, + { + "epoch": 0.18, + "learning_rate": 0.0006586639919759278, + "loss": 1.5139, + "theoretical_loss": 3.8616320556607886, + "tokens_seen": 610205696 + }, + { + "epoch": 0.18, + "learning_rate": 0.0006586479438314946, + "loss": 1.6233, + "theoretical_loss": 3.861589228113302, + "tokens_seen": 610271232 + }, + { + "epoch": 0.18, + "learning_rate": 0.0006586318956870612, + "loss": 1.4776, + "theoretical_loss": 3.8615464064523586, + "tokens_seen": 610336768 + }, + { + "epoch": 0.18, + "learning_rate": 0.000658615847542628, + "loss": 1.5545, + "theoretical_loss": 3.861503590676518, + "tokens_seen": 610402304 + }, + { + "epoch": 0.18, + "learning_rate": 0.0006585997993981946, + "loss": 1.5149, + "theoretical_loss": 3.861460780784339, + "tokens_seen": 610467840 + }, + { + "epoch": 0.19, + "learning_rate": 0.0006585837512537614, + "loss": 1.5956, + "theoretical_loss": 3.8614179767743826, + "tokens_seen": 610533376 + }, + { + "epoch": 0.19, + "learning_rate": 0.000658567703109328, + "loss": 1.4049, + "theoretical_loss": 3.861375178645208, + "tokens_seen": 610598912 + }, + { + "epoch": 0.19, + "learning_rate": 0.0006585516549648948, + "loss": 1.6517, + "theoretical_loss": 3.861332386395377, + "tokens_seen": 610664448 + }, + { + "epoch": 0.19, + "learning_rate": 0.0006585356068204614, + "loss": 1.5246, + "theoretical_loss": 3.8612896000234516, + "tokens_seen": 610729984 + }, + { + "epoch": 0.19, + "learning_rate": 0.0006585195586760282, + "loss": 1.5261, + "theoretical_loss": 3.861246819527992, + "tokens_seen": 610795520 + }, + { + "epoch": 0.19, + "learning_rate": 0.0006585035105315948, + "loss": 1.5648, + "theoretical_loss": 3.861204044907563, + "tokens_seen": 610861056 + }, + { + "epoch": 0.19, + "learning_rate": 0.0006584874623871616, + "loss": 1.6007, + "theoretical_loss": 3.8611612761607255, + "tokens_seen": 610926592 + }, + { + "epoch": 0.19, + "learning_rate": 0.0006584714142427282, + "loss": 1.4285, + "theoretical_loss": 3.861118513286044, + "tokens_seen": 610992128 + }, + { + "epoch": 0.19, + "learning_rate": 0.000658455366098295, + "loss": 1.4912, + "theoretical_loss": 3.8610757562820828, + "tokens_seen": 611057664 + }, + { + "epoch": 0.19, + "learning_rate": 0.0006584393179538616, + "loss": 1.4684, + "theoretical_loss": 3.861033005147406, + "tokens_seen": 611123200 + }, + { + "epoch": 0.19, + "learning_rate": 0.0006584232698094284, + "loss": 1.457, + "theoretical_loss": 3.8609902598805785, + "tokens_seen": 611188736 + }, + { + "epoch": 0.19, + "learning_rate": 0.000658407221664995, + "loss": 1.436, + "theoretical_loss": 3.860947520480167, + "tokens_seen": 611254272 + }, + { + "epoch": 0.19, + "learning_rate": 0.0006583911735205618, + "loss": 1.4949, + "theoretical_loss": 3.8609047869447357, + "tokens_seen": 611319808 + }, + { + "epoch": 0.19, + "learning_rate": 0.0006583751253761284, + "loss": 1.529, + "theoretical_loss": 3.8608620592728524, + "tokens_seen": 611385344 + }, + { + "epoch": 0.19, + "learning_rate": 0.0006583590772316951, + "loss": 1.5244, + "theoretical_loss": 3.860819337463084, + "tokens_seen": 611450880 + }, + { + "epoch": 0.19, + "learning_rate": 0.0006583430290872618, + "loss": 1.6167, + "theoretical_loss": 3.860776621513998, + "tokens_seen": 611516416 + }, + { + "epoch": 0.19, + "learning_rate": 0.0006583269809428285, + "loss": 1.5922, + "theoretical_loss": 3.8607339114241617, + "tokens_seen": 611581952 + }, + { + "epoch": 0.19, + "learning_rate": 0.0006583109327983952, + "loss": 1.5525, + "theoretical_loss": 3.860691207192145, + "tokens_seen": 611647488 + }, + { + "epoch": 0.19, + "learning_rate": 0.0006582948846539619, + "loss": 1.4765, + "theoretical_loss": 3.860648508816516, + "tokens_seen": 611713024 + }, + { + "epoch": 0.19, + "learning_rate": 0.0006582788365095286, + "loss": 1.3923, + "theoretical_loss": 3.8606058162958443, + "tokens_seen": 611778560 + }, + { + "epoch": 0.19, + "learning_rate": 0.0006582627883650953, + "loss": 1.4339, + "theoretical_loss": 3.8605631296287006, + "tokens_seen": 611844096 + }, + { + "epoch": 0.19, + "learning_rate": 0.000658246740220662, + "loss": 1.4809, + "theoretical_loss": 3.860520448813655, + "tokens_seen": 611909632 + }, + { + "epoch": 0.19, + "learning_rate": 0.0006582306920762287, + "loss": 1.6139, + "theoretical_loss": 3.8604777738492784, + "tokens_seen": 611975168 + }, + { + "epoch": 0.19, + "learning_rate": 0.0006582146439317954, + "loss": 1.5063, + "theoretical_loss": 3.860435104734143, + "tokens_seen": 612040704 + }, + { + "epoch": 0.19, + "learning_rate": 0.0006581985957873621, + "loss": 1.6165, + "theoretical_loss": 3.86039244146682, + "tokens_seen": 612106240 + }, + { + "epoch": 0.19, + "learning_rate": 0.0006581825476429288, + "loss": 1.3913, + "theoretical_loss": 3.860349784045883, + "tokens_seen": 612171776 + }, + { + "epoch": 0.19, + "learning_rate": 0.0006581664994984955, + "loss": 1.4562, + "theoretical_loss": 3.8603071324699045, + "tokens_seen": 612237312 + }, + { + "epoch": 0.19, + "learning_rate": 0.0006581504513540622, + "loss": 1.4803, + "theoretical_loss": 3.8602644867374583, + "tokens_seen": 612302848 + }, + { + "epoch": 0.19, + "learning_rate": 0.0006581344032096289, + "loss": 1.5496, + "theoretical_loss": 3.860221846847118, + "tokens_seen": 612368384 + }, + { + "epoch": 0.19, + "learning_rate": 0.0006581183550651956, + "loss": 1.6067, + "theoretical_loss": 3.8601792127974583, + "tokens_seen": 612433920 + }, + { + "epoch": 0.19, + "learning_rate": 0.0006581023069207623, + "loss": 1.3196, + "theoretical_loss": 3.860136584587055, + "tokens_seen": 612499456 + }, + { + "epoch": 0.19, + "learning_rate": 0.000658086258776329, + "loss": 1.5047, + "theoretical_loss": 3.860093962214483, + "tokens_seen": 612564992 + }, + { + "epoch": 0.19, + "learning_rate": 0.0006580702106318957, + "loss": 1.4373, + "theoretical_loss": 3.860051345678319, + "tokens_seen": 612630528 + }, + { + "epoch": 0.19, + "learning_rate": 0.0006580541624874624, + "loss": 1.5616, + "theoretical_loss": 3.860008734977139, + "tokens_seen": 612696064 + }, + { + "epoch": 0.19, + "objective/train/docs_used": 228263, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.6882294416427612, + "objective/train/theoretical_loss": 3.8599661301095196, + "objective/train/tokens_used": 633221600, + "theoretical_loss": 3.8599661301095196, + "tokens_seen": 612761600 + }, + { + "epoch": 0.19, + "learning_rate": 0.0006580381143430291, + "loss": 1.6042, + "theoretical_loss": 3.8599661301095196, + "tokens_seen": 612761600 + }, + { + "epoch": 0.19, + "learning_rate": 0.0006580220661985958, + "loss": 1.4185, + "theoretical_loss": 3.85992353107404, + "tokens_seen": 612827136 + }, + { + "epoch": 0.19, + "learning_rate": 0.0006580060180541625, + "loss": 1.4756, + "theoretical_loss": 3.8598809378692764, + "tokens_seen": 612892672 + }, + { + "epoch": 0.19, + "learning_rate": 0.0006579899699097292, + "loss": 1.4205, + "theoretical_loss": 3.859838350493809, + "tokens_seen": 612958208 + }, + { + "epoch": 0.19, + "learning_rate": 0.0006579739217652959, + "loss": 1.518, + "theoretical_loss": 3.8597957689462157, + "tokens_seen": 613023744 + }, + { + "epoch": 0.19, + "learning_rate": 0.0006579578736208626, + "loss": 1.5546, + "theoretical_loss": 3.859753193225077, + "tokens_seen": 613089280 + }, + { + "epoch": 0.19, + "learning_rate": 0.0006579418254764293, + "loss": 1.3394, + "theoretical_loss": 3.859710623328972, + "tokens_seen": 613154816 + }, + { + "epoch": 0.19, + "learning_rate": 0.000657925777331996, + "loss": 1.5273, + "theoretical_loss": 3.8596680592564816, + "tokens_seen": 613220352 + }, + { + "epoch": 0.19, + "learning_rate": 0.0006579097291875627, + "loss": 1.5859, + "theoretical_loss": 3.859625501006188, + "tokens_seen": 613285888 + }, + { + "epoch": 0.19, + "learning_rate": 0.0006578936810431294, + "loss": 1.5428, + "theoretical_loss": 3.859582948576671, + "tokens_seen": 613351424 + }, + { + "epoch": 0.19, + "learning_rate": 0.0006578776328986961, + "loss": 1.5568, + "theoretical_loss": 3.8595404019665143, + "tokens_seen": 613416960 + }, + { + "epoch": 0.19, + "learning_rate": 0.0006578615847542628, + "loss": 1.4762, + "theoretical_loss": 3.859497861174299, + "tokens_seen": 613482496 + }, + { + "epoch": 0.19, + "learning_rate": 0.0006578455366098295, + "loss": 1.5407, + "theoretical_loss": 3.859455326198609, + "tokens_seen": 613548032 + }, + { + "epoch": 0.19, + "learning_rate": 0.0006578294884653962, + "loss": 1.5515, + "theoretical_loss": 3.8594127970380274, + "tokens_seen": 613613568 + }, + { + "epoch": 0.19, + "learning_rate": 0.0006578134403209629, + "loss": 1.4446, + "theoretical_loss": 3.859370273691139, + "tokens_seen": 613679104 + }, + { + "epoch": 0.19, + "learning_rate": 0.0006577973921765296, + "loss": 1.4078, + "theoretical_loss": 3.8593277561565276, + "tokens_seen": 613744640 + }, + { + "epoch": 0.19, + "learning_rate": 0.0006577813440320963, + "loss": 1.5122, + "theoretical_loss": 3.8592852444327788, + "tokens_seen": 613810176 + }, + { + "epoch": 0.19, + "learning_rate": 0.000657765295887663, + "loss": 1.436, + "theoretical_loss": 3.8592427385184775, + "tokens_seen": 613875712 + }, + { + "epoch": 0.19, + "learning_rate": 0.0006577492477432297, + "loss": 1.5571, + "theoretical_loss": 3.85920023841221, + "tokens_seen": 613941248 + }, + { + "epoch": 0.19, + "learning_rate": 0.0006577331995987964, + "loss": 1.4734, + "theoretical_loss": 3.859157744112563, + "tokens_seen": 614006784 + }, + { + "epoch": 0.19, + "learning_rate": 0.0006577171514543631, + "loss": 1.5467, + "theoretical_loss": 3.859115255618123, + "tokens_seen": 614072320 + }, + { + "epoch": 0.19, + "learning_rate": 0.0006577011033099298, + "loss": 1.5334, + "theoretical_loss": 3.8590727729274787, + "tokens_seen": 614137856 + }, + { + "epoch": 0.19, + "learning_rate": 0.0006576850551654965, + "loss": 1.4055, + "theoretical_loss": 3.859030296039217, + "tokens_seen": 614203392 + }, + { + "epoch": 0.19, + "learning_rate": 0.0006576690070210633, + "loss": 1.628, + "theoretical_loss": 3.8589878249519263, + "tokens_seen": 614268928 + }, + { + "epoch": 0.19, + "learning_rate": 0.0006576529588766299, + "loss": 1.5338, + "theoretical_loss": 3.8589453596641965, + "tokens_seen": 614334464 + }, + { + "epoch": 0.19, + "learning_rate": 0.0006576369107321967, + "loss": 1.4922, + "theoretical_loss": 3.858902900174617, + "tokens_seen": 614400000 + }, + { + "epoch": 0.19, + "learning_rate": 0.0006576208625877633, + "loss": 1.398, + "theoretical_loss": 3.858860446481777, + "tokens_seen": 614465536 + }, + { + "epoch": 0.19, + "learning_rate": 0.00065760481444333, + "loss": 1.5402, + "theoretical_loss": 3.8588179985842674, + "tokens_seen": 614531072 + }, + { + "epoch": 0.19, + "learning_rate": 0.0006575887662988967, + "loss": 1.3944, + "theoretical_loss": 3.8587755564806794, + "tokens_seen": 614596608 + }, + { + "epoch": 0.19, + "learning_rate": 0.0006575727181544635, + "loss": 1.5557, + "theoretical_loss": 3.8587331201696045, + "tokens_seen": 614662144 + }, + { + "epoch": 0.19, + "learning_rate": 0.0006575566700100301, + "loss": 1.4874, + "theoretical_loss": 3.8586906896496336, + "tokens_seen": 614727680 + }, + { + "epoch": 0.19, + "learning_rate": 0.0006575406218655969, + "loss": 1.4428, + "theoretical_loss": 3.8586482649193607, + "tokens_seen": 614793216 + }, + { + "epoch": 0.19, + "learning_rate": 0.0006575245737211635, + "loss": 1.4083, + "theoretical_loss": 3.8586058459773778, + "tokens_seen": 614858752 + }, + { + "epoch": 0.19, + "learning_rate": 0.0006575085255767303, + "loss": 1.4143, + "theoretical_loss": 3.8585634328222778, + "tokens_seen": 614924288 + }, + { + "epoch": 0.19, + "learning_rate": 0.0006574924774322969, + "loss": 1.4872, + "theoretical_loss": 3.8585210254526565, + "tokens_seen": 614989824 + }, + { + "epoch": 0.19, + "learning_rate": 0.0006574764292878637, + "loss": 1.4063, + "theoretical_loss": 3.8584786238671063, + "tokens_seen": 615055360 + }, + { + "epoch": 0.19, + "learning_rate": 0.0006574603811434303, + "loss": 1.5262, + "theoretical_loss": 3.8584362280642233, + "tokens_seen": 615120896 + }, + { + "epoch": 0.19, + "learning_rate": 0.0006574443329989971, + "loss": 1.4851, + "theoretical_loss": 3.858393838042603, + "tokens_seen": 615186432 + }, + { + "epoch": 0.19, + "learning_rate": 0.0006574282848545637, + "loss": 1.5471, + "theoretical_loss": 3.8583514538008403, + "tokens_seen": 615251968 + }, + { + "epoch": 0.19, + "learning_rate": 0.0006574122367101305, + "loss": 1.4549, + "theoretical_loss": 3.858309075337532, + "tokens_seen": 615317504 + }, + { + "epoch": 0.19, + "learning_rate": 0.0006573961885656971, + "loss": 1.3736, + "theoretical_loss": 3.8582667026512754, + "tokens_seen": 615383040 + }, + { + "epoch": 0.19, + "learning_rate": 0.0006573801404212639, + "loss": 1.4562, + "theoretical_loss": 3.858224335740668, + "tokens_seen": 615448576 + }, + { + "epoch": 0.19, + "learning_rate": 0.0006573640922768305, + "loss": 1.5244, + "theoretical_loss": 3.8581819746043067, + "tokens_seen": 615514112 + }, + { + "epoch": 0.19, + "learning_rate": 0.0006573480441323973, + "loss": 1.5771, + "theoretical_loss": 3.8581396192407906, + "tokens_seen": 615579648 + }, + { + "epoch": 0.19, + "learning_rate": 0.0006573319959879639, + "loss": 1.4699, + "theoretical_loss": 3.8580972696487184, + "tokens_seen": 615645184 + }, + { + "epoch": 0.19, + "learning_rate": 0.0006573159478435307, + "loss": 1.4499, + "theoretical_loss": 3.858054925826689, + "tokens_seen": 615710720 + }, + { + "epoch": 0.19, + "learning_rate": 0.0006572998996990973, + "loss": 1.5617, + "theoretical_loss": 3.8580125877733025, + "tokens_seen": 615776256 + }, + { + "epoch": 0.19, + "learning_rate": 0.0006572838515546641, + "loss": 1.4423, + "theoretical_loss": 3.8579702554871593, + "tokens_seen": 615841792 + }, + { + "epoch": 0.19, + "learning_rate": 0.0006572678034102307, + "loss": 1.4077, + "theoretical_loss": 3.8579279289668604, + "tokens_seen": 615907328 + }, + { + "epoch": 0.19, + "learning_rate": 0.0006572517552657975, + "loss": 1.4887, + "theoretical_loss": 3.8578856082110065, + "tokens_seen": 615972864 + }, + { + "epoch": 0.19, + "objective/train/docs_used": 229528, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.4605629444122314, + "objective/train/theoretical_loss": 3.8578432932182, + "objective/train/tokens_used": 636498400, + "theoretical_loss": 3.8578432932182, + "tokens_seen": 616038400 + }, + { + "epoch": 0.19, + "learning_rate": 0.0006572357071213641, + "loss": 1.5247, + "theoretical_loss": 3.8578432932182, + "tokens_seen": 616038400 + }, + { + "epoch": 0.19, + "learning_rate": 0.0006572196589769309, + "loss": 1.4835, + "theoretical_loss": 3.8578009839870426, + "tokens_seen": 616103936 + }, + { + "epoch": 0.19, + "learning_rate": 0.0006572036108324975, + "loss": 1.5083, + "theoretical_loss": 3.8577586805161372, + "tokens_seen": 616169472 + }, + { + "epoch": 0.19, + "learning_rate": 0.0006571875626880643, + "loss": 1.473, + "theoretical_loss": 3.857716382804087, + "tokens_seen": 616235008 + }, + { + "epoch": 0.19, + "learning_rate": 0.0006571715145436309, + "loss": 1.5174, + "theoretical_loss": 3.8576740908494958, + "tokens_seen": 616300544 + }, + { + "epoch": 0.19, + "learning_rate": 0.0006571554663991977, + "loss": 1.5804, + "theoretical_loss": 3.8576318046509677, + "tokens_seen": 616366080 + }, + { + "epoch": 0.19, + "learning_rate": 0.0006571394182547643, + "loss": 1.4493, + "theoretical_loss": 3.857589524207108, + "tokens_seen": 616431616 + }, + { + "epoch": 0.19, + "learning_rate": 0.0006571233701103311, + "loss": 1.472, + "theoretical_loss": 3.8575472495165206, + "tokens_seen": 616497152 + }, + { + "epoch": 0.19, + "learning_rate": 0.0006571073219658977, + "loss": 1.5188, + "theoretical_loss": 3.8575049805778123, + "tokens_seen": 616562688 + }, + { + "epoch": 0.19, + "learning_rate": 0.0006570912738214645, + "loss": 1.4718, + "theoretical_loss": 3.8574627173895886, + "tokens_seen": 616628224 + }, + { + "epoch": 0.19, + "learning_rate": 0.0006570752256770311, + "loss": 1.612, + "theoretical_loss": 3.8574204599504567, + "tokens_seen": 616693760 + }, + { + "epoch": 0.19, + "learning_rate": 0.0006570591775325979, + "loss": 1.3371, + "theoretical_loss": 3.857378208259023, + "tokens_seen": 616759296 + }, + { + "epoch": 0.19, + "learning_rate": 0.0006570431293881645, + "loss": 1.4427, + "theoretical_loss": 3.8573359623138956, + "tokens_seen": 616824832 + }, + { + "epoch": 0.19, + "learning_rate": 0.0006570270812437313, + "loss": 1.6086, + "theoretical_loss": 3.857293722113683, + "tokens_seen": 616890368 + }, + { + "epoch": 0.19, + "learning_rate": 0.0006570110330992979, + "loss": 1.6664, + "theoretical_loss": 3.8572514876569923, + "tokens_seen": 616955904 + }, + { + "epoch": 0.19, + "learning_rate": 0.0006569949849548647, + "loss": 1.518, + "theoretical_loss": 3.857209258942434, + "tokens_seen": 617021440 + }, + { + "epoch": 0.19, + "learning_rate": 0.0006569789368104313, + "loss": 1.5218, + "theoretical_loss": 3.857167035968617, + "tokens_seen": 617086976 + }, + { + "epoch": 0.19, + "learning_rate": 0.000656962888665998, + "loss": 1.5973, + "theoretical_loss": 3.8571248187341522, + "tokens_seen": 617152512 + }, + { + "epoch": 0.19, + "learning_rate": 0.0006569468405215647, + "loss": 1.4636, + "theoretical_loss": 3.8570826072376487, + "tokens_seen": 617218048 + }, + { + "epoch": 0.19, + "learning_rate": 0.0006569307923771314, + "loss": 1.5534, + "theoretical_loss": 3.8570404014777178, + "tokens_seen": 617283584 + }, + { + "epoch": 0.19, + "learning_rate": 0.0006569147442326981, + "loss": 1.5186, + "theoretical_loss": 3.8569982014529716, + "tokens_seen": 617349120 + }, + { + "epoch": 0.19, + "learning_rate": 0.0006568986960882648, + "loss": 1.5197, + "theoretical_loss": 3.8569560071620224, + "tokens_seen": 617414656 + }, + { + "epoch": 0.19, + "learning_rate": 0.0006568826479438315, + "loss": 1.4649, + "theoretical_loss": 3.8569138186034815, + "tokens_seen": 617480192 + }, + { + "epoch": 0.19, + "learning_rate": 0.0006568665997993982, + "loss": 1.5218, + "theoretical_loss": 3.8568716357759625, + "tokens_seen": 617545728 + }, + { + "epoch": 0.19, + "learning_rate": 0.0006568505516549649, + "loss": 1.6223, + "theoretical_loss": 3.856829458678079, + "tokens_seen": 617611264 + }, + { + "epoch": 0.19, + "learning_rate": 0.0006568345035105316, + "loss": 1.4927, + "theoretical_loss": 3.856787287308444, + "tokens_seen": 617676800 + }, + { + "epoch": 0.19, + "learning_rate": 0.0006568184553660983, + "loss": 1.5528, + "theoretical_loss": 3.8567451216656723, + "tokens_seen": 617742336 + }, + { + "epoch": 0.19, + "learning_rate": 0.000656802407221665, + "loss": 1.5556, + "theoretical_loss": 3.8567029617483795, + "tokens_seen": 617807872 + }, + { + "epoch": 0.19, + "learning_rate": 0.0006567863590772317, + "loss": 1.4133, + "theoretical_loss": 3.8566608075551803, + "tokens_seen": 617873408 + }, + { + "epoch": 0.19, + "learning_rate": 0.0006567703109327984, + "loss": 1.5485, + "theoretical_loss": 3.8566186590846905, + "tokens_seen": 617938944 + }, + { + "epoch": 0.19, + "learning_rate": 0.0006567542627883651, + "loss": 1.5554, + "theoretical_loss": 3.856576516335526, + "tokens_seen": 618004480 + }, + { + "epoch": 0.19, + "learning_rate": 0.0006567382146439318, + "loss": 1.5224, + "theoretical_loss": 3.8565343793063045, + "tokens_seen": 618070016 + }, + { + "epoch": 0.19, + "learning_rate": 0.0006567221664994985, + "loss": 1.585, + "theoretical_loss": 3.856492247995643, + "tokens_seen": 618135552 + }, + { + "epoch": 0.19, + "learning_rate": 0.0006567061183550652, + "loss": 1.4764, + "theoretical_loss": 3.856450122402159, + "tokens_seen": 618201088 + }, + { + "epoch": 0.19, + "learning_rate": 0.0006566900702106319, + "loss": 1.6547, + "theoretical_loss": 3.8564080025244705, + "tokens_seen": 618266624 + }, + { + "epoch": 0.19, + "learning_rate": 0.0006566740220661986, + "loss": 1.5319, + "theoretical_loss": 3.8563658883611964, + "tokens_seen": 618332160 + }, + { + "epoch": 0.19, + "learning_rate": 0.0006566579739217653, + "loss": 1.4743, + "theoretical_loss": 3.856323779910956, + "tokens_seen": 618397696 + }, + { + "epoch": 0.19, + "learning_rate": 0.000656641925777332, + "loss": 1.6, + "theoretical_loss": 3.856281677172369, + "tokens_seen": 618463232 + }, + { + "epoch": 0.19, + "learning_rate": 0.0006566258776328987, + "loss": 1.4608, + "theoretical_loss": 3.856239580144056, + "tokens_seen": 618528768 + }, + { + "epoch": 0.19, + "learning_rate": 0.0006566098294884654, + "loss": 1.4459, + "theoretical_loss": 3.8561974888246366, + "tokens_seen": 618594304 + }, + { + "epoch": 0.19, + "learning_rate": 0.0006565937813440321, + "loss": 1.5391, + "theoretical_loss": 3.8561554032127323, + "tokens_seen": 618659840 + }, + { + "epoch": 0.19, + "learning_rate": 0.0006565777331995988, + "loss": 1.4865, + "theoretical_loss": 3.856113323306965, + "tokens_seen": 618725376 + }, + { + "epoch": 0.19, + "learning_rate": 0.0006565616850551655, + "loss": 1.5415, + "theoretical_loss": 3.8560712491059563, + "tokens_seen": 618790912 + }, + { + "epoch": 0.19, + "learning_rate": 0.0006565456369107322, + "loss": 1.3627, + "theoretical_loss": 3.8560291806083296, + "tokens_seen": 618856448 + }, + { + "epoch": 0.19, + "learning_rate": 0.0006565295887662989, + "loss": 1.4591, + "theoretical_loss": 3.855987117812707, + "tokens_seen": 618921984 + }, + { + "epoch": 0.19, + "learning_rate": 0.0006565135406218656, + "loss": 1.6096, + "theoretical_loss": 3.8559450607177124, + "tokens_seen": 618987520 + }, + { + "epoch": 0.19, + "learning_rate": 0.0006564974924774323, + "loss": 1.566, + "theoretical_loss": 3.8559030093219695, + "tokens_seen": 619053056 + }, + { + "epoch": 0.19, + "learning_rate": 0.000656481444332999, + "loss": 1.5837, + "theoretical_loss": 3.855860963624103, + "tokens_seen": 619118592 + }, + { + "epoch": 0.19, + "learning_rate": 0.0006564653961885657, + "loss": 1.5106, + "theoretical_loss": 3.855818923622738, + "tokens_seen": 619184128 + }, + { + "epoch": 0.19, + "learning_rate": 0.0006564493480441324, + "loss": 1.4565, + "theoretical_loss": 3.855776889316499, + "tokens_seen": 619249664 + }, + { + "epoch": 0.19, + "objective/train/docs_used": 230719, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.6779125928878784, + "objective/train/theoretical_loss": 3.855734860704014, + "objective/train/tokens_used": 639775200, + "theoretical_loss": 3.855734860704014, + "tokens_seen": 619315200 + }, + { + "epoch": 0.19, + "learning_rate": 0.0006564332998996991, + "loss": 1.6294, + "theoretical_loss": 3.855734860704014, + "tokens_seen": 619315200 + }, + { + "epoch": 0.19, + "learning_rate": 0.0006564172517552658, + "loss": 1.5053, + "theoretical_loss": 3.8556928377839066, + "tokens_seen": 619380736 + }, + { + "epoch": 0.19, + "learning_rate": 0.0006564012036108325, + "loss": 1.6278, + "theoretical_loss": 3.8556508205548057, + "tokens_seen": 619446272 + }, + { + "epoch": 0.19, + "learning_rate": 0.0006563851554663992, + "loss": 1.6319, + "theoretical_loss": 3.8556088090153375, + "tokens_seen": 619511808 + }, + { + "epoch": 0.19, + "learning_rate": 0.000656369107321966, + "loss": 1.5256, + "theoretical_loss": 3.8555668031641304, + "tokens_seen": 619577344 + }, + { + "epoch": 0.19, + "learning_rate": 0.0006563530591775326, + "loss": 1.5941, + "theoretical_loss": 3.8555248029998124, + "tokens_seen": 619642880 + }, + { + "epoch": 0.19, + "learning_rate": 0.0006563370110330993, + "loss": 1.4198, + "theoretical_loss": 3.855482808521012, + "tokens_seen": 619708416 + }, + { + "epoch": 0.19, + "learning_rate": 0.000656320962888666, + "loss": 1.4917, + "theoretical_loss": 3.855440819726359, + "tokens_seen": 619773952 + }, + { + "epoch": 0.19, + "learning_rate": 0.0006563049147442327, + "loss": 1.5485, + "theoretical_loss": 3.8553988366144827, + "tokens_seen": 619839488 + }, + { + "epoch": 0.19, + "learning_rate": 0.0006562888665997994, + "loss": 1.5699, + "theoretical_loss": 3.8553568591840133, + "tokens_seen": 619905024 + }, + { + "epoch": 0.19, + "learning_rate": 0.0006562728184553662, + "loss": 1.5559, + "theoretical_loss": 3.8553148874335816, + "tokens_seen": 619970560 + }, + { + "epoch": 0.19, + "learning_rate": 0.0006562567703109328, + "loss": 1.5364, + "theoretical_loss": 3.855272921361819, + "tokens_seen": 620036096 + }, + { + "epoch": 0.19, + "learning_rate": 0.0006562407221664996, + "loss": 1.5751, + "theoretical_loss": 3.855230960967356, + "tokens_seen": 620101632 + }, + { + "epoch": 0.19, + "learning_rate": 0.0006562246740220662, + "loss": 1.54, + "theoretical_loss": 3.8551890062488265, + "tokens_seen": 620167168 + }, + { + "epoch": 0.19, + "learning_rate": 0.000656208625877633, + "loss": 1.525, + "theoretical_loss": 3.855147057204861, + "tokens_seen": 620232704 + }, + { + "epoch": 0.19, + "learning_rate": 0.0006561925777331996, + "loss": 1.5433, + "theoretical_loss": 3.8551051138340937, + "tokens_seen": 620298240 + }, + { + "epoch": 0.19, + "learning_rate": 0.0006561765295887664, + "loss": 1.5717, + "theoretical_loss": 3.8550631761351575, + "tokens_seen": 620363776 + }, + { + "epoch": 0.19, + "learning_rate": 0.000656160481444333, + "loss": 1.5589, + "theoretical_loss": 3.8550212441066876, + "tokens_seen": 620429312 + }, + { + "epoch": 0.19, + "learning_rate": 0.0006561444332998998, + "loss": 1.4823, + "theoretical_loss": 3.8549793177473166, + "tokens_seen": 620494848 + }, + { + "epoch": 0.19, + "learning_rate": 0.0006561283851554664, + "loss": 1.5216, + "theoretical_loss": 3.8549373970556813, + "tokens_seen": 620560384 + }, + { + "epoch": 0.19, + "learning_rate": 0.0006561123370110332, + "loss": 1.4552, + "theoretical_loss": 3.8548954820304155, + "tokens_seen": 620625920 + }, + { + "epoch": 0.19, + "learning_rate": 0.0006560962888665998, + "loss": 1.5304, + "theoretical_loss": 3.854853572670156, + "tokens_seen": 620691456 + }, + { + "epoch": 0.19, + "learning_rate": 0.0006560802407221666, + "loss": 1.4632, + "theoretical_loss": 3.8548116689735386, + "tokens_seen": 620756992 + }, + { + "epoch": 0.19, + "learning_rate": 0.0006560641925777332, + "loss": 1.4694, + "theoretical_loss": 3.8547697709392006, + "tokens_seen": 620822528 + }, + { + "epoch": 0.19, + "learning_rate": 0.0006560481444333, + "loss": 1.561, + "theoretical_loss": 3.854727878565779, + "tokens_seen": 620888064 + }, + { + "epoch": 0.19, + "learning_rate": 0.0006560320962888666, + "loss": 1.6225, + "theoretical_loss": 3.8546859918519107, + "tokens_seen": 620953600 + }, + { + "epoch": 0.19, + "learning_rate": 0.0006560160481444334, + "loss": 1.4997, + "theoretical_loss": 3.8546441107962357, + "tokens_seen": 621019136 + }, + { + "epoch": 0.19, + "learning_rate": 0.000656, + "loss": 1.6313, + "theoretical_loss": 3.854602235397391, + "tokens_seen": 621084672 + }, + { + "epoch": 0.19, + "learning_rate": 0.0006559839518555668, + "loss": 1.4989, + "theoretical_loss": 3.8545603656540175, + "tokens_seen": 621150208 + }, + { + "epoch": 0.19, + "learning_rate": 0.0006559679037111334, + "loss": 1.4749, + "theoretical_loss": 3.854518501564753, + "tokens_seen": 621215744 + }, + { + "epoch": 0.19, + "learning_rate": 0.0006559518555667002, + "loss": 1.4842, + "theoretical_loss": 3.8544766431282387, + "tokens_seen": 621281280 + }, + { + "epoch": 0.19, + "learning_rate": 0.0006559358074222668, + "loss": 1.5463, + "theoretical_loss": 3.854434790343115, + "tokens_seen": 621346816 + }, + { + "epoch": 0.19, + "learning_rate": 0.0006559197592778336, + "loss": 1.3461, + "theoretical_loss": 3.8543929432080226, + "tokens_seen": 621412352 + }, + { + "epoch": 0.19, + "learning_rate": 0.0006559037111334002, + "loss": 1.5573, + "theoretical_loss": 3.8543511017216034, + "tokens_seen": 621477888 + }, + { + "epoch": 0.19, + "learning_rate": 0.000655887662988967, + "loss": 1.4535, + "theoretical_loss": 3.8543092658825, + "tokens_seen": 621543424 + }, + { + "epoch": 0.19, + "learning_rate": 0.0006558716148445336, + "loss": 1.4982, + "theoretical_loss": 3.8542674356893527, + "tokens_seen": 621608960 + }, + { + "epoch": 0.19, + "learning_rate": 0.0006558555667001004, + "loss": 1.4733, + "theoretical_loss": 3.8542256111408064, + "tokens_seen": 621674496 + }, + { + "epoch": 0.19, + "learning_rate": 0.0006558395185556671, + "loss": 1.5204, + "theoretical_loss": 3.854183792235504, + "tokens_seen": 621740032 + }, + { + "epoch": 0.19, + "learning_rate": 0.0006558234704112338, + "loss": 1.4372, + "theoretical_loss": 3.8541419789720894, + "tokens_seen": 621805568 + }, + { + "epoch": 0.19, + "learning_rate": 0.0006558074222668005, + "loss": 1.5114, + "theoretical_loss": 3.854100171349206, + "tokens_seen": 621871104 + }, + { + "epoch": 0.19, + "learning_rate": 0.0006557913741223672, + "loss": 1.6178, + "theoretical_loss": 3.8540583693655, + "tokens_seen": 621936640 + }, + { + "epoch": 0.19, + "learning_rate": 0.0006557753259779339, + "loss": 1.5169, + "theoretical_loss": 3.854016573019616, + "tokens_seen": 622002176 + }, + { + "epoch": 0.19, + "learning_rate": 0.0006557592778335006, + "loss": 1.5545, + "theoretical_loss": 3.8539747823101997, + "tokens_seen": 622067712 + }, + { + "epoch": 0.19, + "learning_rate": 0.0006557432296890673, + "loss": 1.4973, + "theoretical_loss": 3.853932997235897, + "tokens_seen": 622133248 + }, + { + "epoch": 0.19, + "learning_rate": 0.000655727181544634, + "loss": 1.5601, + "theoretical_loss": 3.8538912177953555, + "tokens_seen": 622198784 + }, + { + "epoch": 0.19, + "learning_rate": 0.0006557111334002006, + "loss": 1.4422, + "theoretical_loss": 3.8538494439872215, + "tokens_seen": 622264320 + }, + { + "epoch": 0.19, + "learning_rate": 0.0006556950852557674, + "loss": 1.3884, + "theoretical_loss": 3.8538076758101427, + "tokens_seen": 622329856 + }, + { + "epoch": 0.19, + "learning_rate": 0.000655679037111334, + "loss": 1.6415, + "theoretical_loss": 3.8537659132627677, + "tokens_seen": 622395392 + }, + { + "epoch": 0.19, + "learning_rate": 0.0006556629889669008, + "loss": 1.6226, + "theoretical_loss": 3.853724156343744, + "tokens_seen": 622460928 + }, + { + "epoch": 0.19, + "learning_rate": 0.0006556469408224674, + "loss": 1.5262, + "theoretical_loss": 3.853682405051722, + "tokens_seen": 622526464 + }, + { + "epoch": 0.19, + "objective/train/docs_used": 231894, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.599900484085083, + "objective/train/theoretical_loss": 3.8536406593853503, + "objective/train/tokens_used": 643052000, + "theoretical_loss": 3.8536406593853503, + "tokens_seen": 622592000 + }, + { + "epoch": 0.19, + "learning_rate": 0.0006556308926780342, + "loss": 1.5395, + "theoretical_loss": 3.8536406593853503, + "tokens_seen": 622592000 + }, + { + "epoch": 0.19, + "learning_rate": 0.0006556148445336008, + "loss": 1.5365, + "theoretical_loss": 3.8535989193432787, + "tokens_seen": 622657536 + }, + { + "epoch": 0.19, + "learning_rate": 0.0006555987963891676, + "loss": 1.5809, + "theoretical_loss": 3.8535571849241577, + "tokens_seen": 622723072 + }, + { + "epoch": 0.19, + "learning_rate": 0.0006555827482447342, + "loss": 1.4026, + "theoretical_loss": 3.8535154561266385, + "tokens_seen": 622788608 + }, + { + "epoch": 0.19, + "learning_rate": 0.0006555667001003009, + "loss": 1.4929, + "theoretical_loss": 3.8534737329493725, + "tokens_seen": 622854144 + }, + { + "epoch": 0.19, + "learning_rate": 0.0006555506519558676, + "loss": 1.5302, + "theoretical_loss": 3.853432015391011, + "tokens_seen": 622919680 + }, + { + "epoch": 0.19, + "learning_rate": 0.0006555346038114343, + "loss": 1.4178, + "theoretical_loss": 3.8533903034502064, + "tokens_seen": 622985216 + }, + { + "epoch": 0.19, + "learning_rate": 0.000655518555667001, + "loss": 1.5148, + "theoretical_loss": 3.8533485971256125, + "tokens_seen": 623050752 + }, + { + "epoch": 0.19, + "learning_rate": 0.0006555025075225677, + "loss": 1.504, + "theoretical_loss": 3.8533068964158805, + "tokens_seen": 623116288 + }, + { + "epoch": 0.19, + "learning_rate": 0.0006554864593781344, + "loss": 1.4778, + "theoretical_loss": 3.8532652013196653, + "tokens_seen": 623181824 + }, + { + "epoch": 0.19, + "learning_rate": 0.0006554704112337011, + "loss": 1.719, + "theoretical_loss": 3.853223511835621, + "tokens_seen": 623247360 + }, + { + "epoch": 0.19, + "learning_rate": 0.0006554543630892678, + "loss": 1.4936, + "theoretical_loss": 3.853181827962402, + "tokens_seen": 623312896 + }, + { + "epoch": 0.19, + "learning_rate": 0.0006554383149448345, + "loss": 1.4941, + "theoretical_loss": 3.853140149698664, + "tokens_seen": 623378432 + }, + { + "epoch": 0.19, + "learning_rate": 0.0006554222668004012, + "loss": 1.4905, + "theoretical_loss": 3.8530984770430607, + "tokens_seen": 623443968 + }, + { + "epoch": 0.19, + "learning_rate": 0.0006554062186559679, + "loss": 1.4627, + "theoretical_loss": 3.85305680999425, + "tokens_seen": 623509504 + }, + { + "epoch": 0.19, + "learning_rate": 0.0006553901705115346, + "loss": 1.5027, + "theoretical_loss": 3.8530151485508872, + "tokens_seen": 623575040 + }, + { + "epoch": 0.19, + "learning_rate": 0.0006553741223671013, + "loss": 1.5715, + "theoretical_loss": 3.85297349271163, + "tokens_seen": 623640576 + }, + { + "epoch": 0.19, + "learning_rate": 0.000655358074222668, + "loss": 1.5138, + "theoretical_loss": 3.852931842475135, + "tokens_seen": 623706112 + }, + { + "epoch": 0.19, + "learning_rate": 0.0006553420260782347, + "loss": 1.4607, + "theoretical_loss": 3.8528901978400603, + "tokens_seen": 623771648 + }, + { + "epoch": 0.19, + "learning_rate": 0.0006553259779338014, + "loss": 1.5037, + "theoretical_loss": 3.852848558805065, + "tokens_seen": 623837184 + }, + { + "epoch": 0.19, + "learning_rate": 0.0006553099297893681, + "loss": 1.4766, + "theoretical_loss": 3.8528069253688066, + "tokens_seen": 623902720 + }, + { + "epoch": 0.19, + "learning_rate": 0.0006552938816449348, + "loss": 1.4428, + "theoretical_loss": 3.852765297529945, + "tokens_seen": 623968256 + }, + { + "epoch": 0.19, + "learning_rate": 0.0006552778335005015, + "loss": 1.601, + "theoretical_loss": 3.8527236752871397, + "tokens_seen": 624033792 + }, + { + "epoch": 0.19, + "learning_rate": 0.0006552617853560682, + "loss": 1.6005, + "theoretical_loss": 3.8526820586390507, + "tokens_seen": 624099328 + }, + { + "epoch": 0.19, + "learning_rate": 0.0006552457372116349, + "loss": 1.5108, + "theoretical_loss": 3.8526404475843385, + "tokens_seen": 624164864 + }, + { + "epoch": 0.19, + "learning_rate": 0.0006552296890672016, + "loss": 1.6024, + "theoretical_loss": 3.8525988421216653, + "tokens_seen": 624230400 + }, + { + "epoch": 0.19, + "learning_rate": 0.0006552136409227683, + "loss": 1.4582, + "theoretical_loss": 3.852557242249691, + "tokens_seen": 624295936 + }, + { + "epoch": 0.19, + "learning_rate": 0.000655197592778335, + "loss": 1.6346, + "theoretical_loss": 3.8525156479670786, + "tokens_seen": 624361472 + }, + { + "epoch": 0.19, + "learning_rate": 0.0006551815446339017, + "loss": 1.5507, + "theoretical_loss": 3.8524740592724904, + "tokens_seen": 624427008 + }, + { + "epoch": 0.19, + "learning_rate": 0.0006551654964894684, + "loss": 1.6687, + "theoretical_loss": 3.8524324761645894, + "tokens_seen": 624492544 + }, + { + "epoch": 0.19, + "learning_rate": 0.0006551494483450351, + "loss": 1.491, + "theoretical_loss": 3.8523908986420383, + "tokens_seen": 624558080 + }, + { + "epoch": 0.19, + "learning_rate": 0.0006551334002006018, + "loss": 1.5161, + "theoretical_loss": 3.8523493267035014, + "tokens_seen": 624623616 + }, + { + "epoch": 0.19, + "learning_rate": 0.0006551173520561685, + "loss": 1.4704, + "theoretical_loss": 3.852307760347643, + "tokens_seen": 624689152 + }, + { + "epoch": 0.19, + "learning_rate": 0.0006551013039117352, + "loss": 1.5244, + "theoretical_loss": 3.852266199573128, + "tokens_seen": 624754688 + }, + { + "epoch": 0.19, + "learning_rate": 0.0006550852557673019, + "loss": 1.5681, + "theoretical_loss": 3.852224644378621, + "tokens_seen": 624820224 + }, + { + "epoch": 0.19, + "learning_rate": 0.0006550692076228686, + "loss": 1.5426, + "theoretical_loss": 3.852183094762789, + "tokens_seen": 624885760 + }, + { + "epoch": 0.19, + "learning_rate": 0.0006550531594784353, + "loss": 1.5544, + "theoretical_loss": 3.852141550724297, + "tokens_seen": 624951296 + }, + { + "epoch": 0.19, + "learning_rate": 0.000655037111334002, + "loss": 1.3611, + "theoretical_loss": 3.852100012261811, + "tokens_seen": 625016832 + }, + { + "epoch": 0.19, + "learning_rate": 0.0006550210631895687, + "loss": 1.4311, + "theoretical_loss": 3.8520584793739996, + "tokens_seen": 625082368 + }, + { + "epoch": 0.19, + "learning_rate": 0.0006550050150451354, + "loss": 1.6771, + "theoretical_loss": 3.852016952059529, + "tokens_seen": 625147904 + }, + { + "epoch": 0.19, + "learning_rate": 0.0006549889669007021, + "loss": 1.4913, + "theoretical_loss": 3.8519754303170677, + "tokens_seen": 625213440 + }, + { + "epoch": 0.19, + "learning_rate": 0.0006549729187562688, + "loss": 1.5556, + "theoretical_loss": 3.8519339141452846, + "tokens_seen": 625278976 + }, + { + "epoch": 0.19, + "learning_rate": 0.0006549568706118355, + "loss": 1.4771, + "theoretical_loss": 3.8518924035428475, + "tokens_seen": 625344512 + }, + { + "epoch": 0.19, + "learning_rate": 0.0006549408224674022, + "loss": 1.4577, + "theoretical_loss": 3.8518508985084265, + "tokens_seen": 625410048 + }, + { + "epoch": 0.19, + "learning_rate": 0.0006549247743229689, + "loss": 1.5267, + "theoretical_loss": 3.8518093990406914, + "tokens_seen": 625475584 + }, + { + "epoch": 0.19, + "learning_rate": 0.0006549087261785357, + "loss": 1.4958, + "theoretical_loss": 3.8517679051383116, + "tokens_seen": 625541120 + }, + { + "epoch": 0.19, + "learning_rate": 0.0006548926780341023, + "loss": 1.5542, + "theoretical_loss": 3.851726416799959, + "tokens_seen": 625606656 + }, + { + "epoch": 0.19, + "learning_rate": 0.000654876629889669, + "loss": 1.4762, + "theoretical_loss": 3.8516849340243042, + "tokens_seen": 625672192 + }, + { + "epoch": 0.19, + "learning_rate": 0.0006548605817452357, + "loss": 1.4964, + "theoretical_loss": 3.8516434568100184, + "tokens_seen": 625737728 + }, + { + "epoch": 0.19, + "learning_rate": 0.0006548445336008025, + "loss": 1.5929, + "theoretical_loss": 3.8516019851557743, + "tokens_seen": 625803264 + }, + { + "epoch": 0.19, + "objective/train/docs_used": 233149, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.390174388885498, + "objective/train/theoretical_loss": 3.8515605190602447, + "objective/train/tokens_used": 646328800, + "theoretical_loss": 3.8515605190602447, + "tokens_seen": 625868800 + }, + { + "epoch": 0.19, + "learning_rate": 0.0006548284854563691, + "loss": 1.4841, + "theoretical_loss": 3.8515605190602447, + "tokens_seen": 625868800 + }, + { + "epoch": 0.19, + "learning_rate": 0.0006548124373119359, + "loss": 1.4283, + "theoretical_loss": 3.8515190585221015, + "tokens_seen": 625934336 + }, + { + "epoch": 0.19, + "learning_rate": 0.0006547963891675025, + "loss": 1.523, + "theoretical_loss": 3.851477603540019, + "tokens_seen": 625999872 + }, + { + "epoch": 0.19, + "learning_rate": 0.0006547803410230693, + "loss": 1.5365, + "theoretical_loss": 3.8514361541126703, + "tokens_seen": 626065408 + }, + { + "epoch": 0.19, + "learning_rate": 0.0006547642928786359, + "loss": 1.5395, + "theoretical_loss": 3.85139471023873, + "tokens_seen": 626130944 + }, + { + "epoch": 0.19, + "learning_rate": 0.0006547482447342027, + "loss": 1.6101, + "theoretical_loss": 3.8513532719168744, + "tokens_seen": 626196480 + }, + { + "epoch": 0.19, + "learning_rate": 0.0006547321965897693, + "loss": 1.5417, + "theoretical_loss": 3.851311839145777, + "tokens_seen": 626262016 + }, + { + "epoch": 0.19, + "learning_rate": 0.0006547161484453361, + "loss": 1.5134, + "theoretical_loss": 3.851270411924114, + "tokens_seen": 626327552 + }, + { + "epoch": 0.19, + "learning_rate": 0.0006547001003009027, + "loss": 1.6123, + "theoretical_loss": 3.8512289902505605, + "tokens_seen": 626393088 + }, + { + "epoch": 0.19, + "learning_rate": 0.0006546840521564695, + "loss": 1.4199, + "theoretical_loss": 3.8511875741237955, + "tokens_seen": 626458624 + }, + { + "epoch": 0.19, + "learning_rate": 0.0006546680040120361, + "loss": 1.6861, + "theoretical_loss": 3.851146163542495, + "tokens_seen": 626524160 + }, + { + "epoch": 0.19, + "learning_rate": 0.0006546519558676029, + "loss": 1.4581, + "theoretical_loss": 3.8511047585053353, + "tokens_seen": 626589696 + }, + { + "epoch": 0.19, + "learning_rate": 0.0006546359077231696, + "loss": 1.6381, + "theoretical_loss": 3.851063359010996, + "tokens_seen": 626655232 + }, + { + "epoch": 0.19, + "learning_rate": 0.0006546198595787363, + "loss": 1.4545, + "theoretical_loss": 3.851021965058155, + "tokens_seen": 626720768 + }, + { + "epoch": 0.19, + "learning_rate": 0.000654603811434303, + "loss": 1.4505, + "theoretical_loss": 3.850980576645491, + "tokens_seen": 626786304 + }, + { + "epoch": 0.19, + "learning_rate": 0.0006545877632898697, + "loss": 1.4464, + "theoretical_loss": 3.8509391937716835, + "tokens_seen": 626851840 + }, + { + "epoch": 0.19, + "learning_rate": 0.0006545717151454364, + "loss": 1.5349, + "theoretical_loss": 3.8508978164354124, + "tokens_seen": 626917376 + }, + { + "epoch": 0.19, + "learning_rate": 0.0006545556670010031, + "loss": 1.5183, + "theoretical_loss": 3.8508564446353573, + "tokens_seen": 626982912 + }, + { + "epoch": 0.19, + "learning_rate": 0.0006545396188565698, + "loss": 1.6613, + "theoretical_loss": 3.8508150783702, + "tokens_seen": 627048448 + }, + { + "epoch": 0.19, + "learning_rate": 0.0006545235707121365, + "loss": 1.4624, + "theoretical_loss": 3.850773717638621, + "tokens_seen": 627113984 + }, + { + "epoch": 0.19, + "learning_rate": 0.0006545075225677032, + "loss": 1.5893, + "theoretical_loss": 3.850732362439302, + "tokens_seen": 627179520 + }, + { + "epoch": 0.19, + "learning_rate": 0.0006544914744232699, + "loss": 1.6196, + "theoretical_loss": 3.850691012770925, + "tokens_seen": 627245056 + }, + { + "epoch": 0.19, + "learning_rate": 0.0006544754262788366, + "loss": 1.5056, + "theoretical_loss": 3.8506496686321725, + "tokens_seen": 627310592 + }, + { + "epoch": 0.19, + "learning_rate": 0.0006544593781344033, + "loss": 1.5613, + "theoretical_loss": 3.8506083300217275, + "tokens_seen": 627376128 + }, + { + "epoch": 0.19, + "learning_rate": 0.00065444332998997, + "loss": 1.5273, + "theoretical_loss": 3.850566996938274, + "tokens_seen": 627441664 + }, + { + "epoch": 0.19, + "learning_rate": 0.0006544272818455367, + "loss": 1.6469, + "theoretical_loss": 3.8505256693804943, + "tokens_seen": 627507200 + }, + { + "epoch": 0.19, + "learning_rate": 0.0006544112337011034, + "loss": 1.5212, + "theoretical_loss": 3.8504843473470745, + "tokens_seen": 627572736 + }, + { + "epoch": 0.19, + "learning_rate": 0.0006543951855566701, + "loss": 1.6191, + "theoretical_loss": 3.850443030836698, + "tokens_seen": 627638272 + }, + { + "epoch": 0.19, + "learning_rate": 0.0006543791374122368, + "loss": 1.4192, + "theoretical_loss": 3.850401719848051, + "tokens_seen": 627703808 + }, + { + "epoch": 0.19, + "learning_rate": 0.0006543630892678035, + "loss": 1.472, + "theoretical_loss": 3.8503604143798187, + "tokens_seen": 627769344 + }, + { + "epoch": 0.19, + "learning_rate": 0.0006543470411233702, + "loss": 1.566, + "theoretical_loss": 3.850319114430687, + "tokens_seen": 627834880 + }, + { + "epoch": 0.19, + "learning_rate": 0.0006543309929789369, + "loss": 1.4551, + "theoretical_loss": 3.8502778199993433, + "tokens_seen": 627900416 + }, + { + "epoch": 0.19, + "learning_rate": 0.0006543149448345035, + "loss": 1.5514, + "theoretical_loss": 3.850236531084473, + "tokens_seen": 627965952 + }, + { + "epoch": 0.19, + "learning_rate": 0.0006542988966900703, + "loss": 1.4248, + "theoretical_loss": 3.8501952476847654, + "tokens_seen": 628031488 + }, + { + "epoch": 0.19, + "learning_rate": 0.0006542828485456369, + "loss": 1.5344, + "theoretical_loss": 3.8501539697989076, + "tokens_seen": 628097024 + }, + { + "epoch": 0.19, + "learning_rate": 0.0006542668004012037, + "loss": 1.6384, + "theoretical_loss": 3.8501126974255877, + "tokens_seen": 628162560 + }, + { + "epoch": 0.19, + "learning_rate": 0.0006542507522567703, + "loss": 1.5695, + "theoretical_loss": 3.850071430563495, + "tokens_seen": 628228096 + }, + { + "epoch": 0.19, + "learning_rate": 0.0006542347041123371, + "loss": 1.517, + "theoretical_loss": 3.850030169211319, + "tokens_seen": 628293632 + }, + { + "epoch": 0.19, + "learning_rate": 0.0006542186559679037, + "loss": 1.5523, + "theoretical_loss": 3.8499889133677483, + "tokens_seen": 628359168 + }, + { + "epoch": 0.19, + "learning_rate": 0.0006542026078234705, + "loss": 1.4371, + "theoretical_loss": 3.8499476630314735, + "tokens_seen": 628424704 + }, + { + "epoch": 0.19, + "learning_rate": 0.0006541865596790371, + "loss": 1.5405, + "theoretical_loss": 3.849906418201186, + "tokens_seen": 628490240 + }, + { + "epoch": 0.19, + "learning_rate": 0.0006541705115346038, + "loss": 1.5333, + "theoretical_loss": 3.849865178875576, + "tokens_seen": 628555776 + }, + { + "epoch": 0.19, + "learning_rate": 0.0006541544633901705, + "loss": 1.5886, + "theoretical_loss": 3.8498239450533354, + "tokens_seen": 628621312 + }, + { + "epoch": 0.19, + "learning_rate": 0.0006541384152457372, + "loss": 1.4478, + "theoretical_loss": 3.8497827167331558, + "tokens_seen": 628686848 + }, + { + "epoch": 0.19, + "learning_rate": 0.0006541223671013039, + "loss": 1.4138, + "theoretical_loss": 3.84974149391373, + "tokens_seen": 628752384 + }, + { + "epoch": 0.19, + "learning_rate": 0.0006541063189568706, + "loss": 1.4298, + "theoretical_loss": 3.849700276593751, + "tokens_seen": 628817920 + }, + { + "epoch": 0.19, + "learning_rate": 0.0006540902708124373, + "loss": 1.4729, + "theoretical_loss": 3.8496590647719113, + "tokens_seen": 628883456 + }, + { + "epoch": 0.19, + "learning_rate": 0.000654074222668004, + "loss": 1.5968, + "theoretical_loss": 3.849617858446905, + "tokens_seen": 628948992 + }, + { + "epoch": 0.19, + "learning_rate": 0.0006540581745235707, + "loss": 1.5647, + "theoretical_loss": 3.8495766576174266, + "tokens_seen": 629014528 + }, + { + "epoch": 0.19, + "learning_rate": 0.0006540421263791374, + "loss": 1.5442, + "theoretical_loss": 3.8495354622821703, + "tokens_seen": 629080064 + }, + { + "epoch": 0.19, + "objective/train/docs_used": 234355, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.415891408920288, + "objective/train/theoretical_loss": 3.8494942724398316, + "objective/train/tokens_used": 649605600, + "theoretical_loss": 3.8494942724398316, + "tokens_seen": 629145600 + }, + { + "epoch": 0.19, + "learning_rate": 0.0006540260782347041, + "loss": 1.4384, + "theoretical_loss": 3.8494942724398316, + "tokens_seen": 629145600 + }, + { + "epoch": 0.19, + "learning_rate": 0.0006540100300902708, + "loss": 1.5006, + "theoretical_loss": 3.8494530880891054, + "tokens_seen": 629211136 + }, + { + "epoch": 0.19, + "learning_rate": 0.0006539939819458375, + "loss": 1.5459, + "theoretical_loss": 3.8494119092286887, + "tokens_seen": 629276672 + }, + { + "epoch": 0.19, + "learning_rate": 0.0006539779338014042, + "loss": 1.5765, + "theoretical_loss": 3.849370735857277, + "tokens_seen": 629342208 + }, + { + "epoch": 0.19, + "learning_rate": 0.0006539618856569709, + "loss": 1.4774, + "theoretical_loss": 3.849329567973567, + "tokens_seen": 629407744 + }, + { + "epoch": 0.19, + "learning_rate": 0.0006539458375125376, + "loss": 1.593, + "theoretical_loss": 3.8492884055762566, + "tokens_seen": 629473280 + }, + { + "epoch": 0.19, + "learning_rate": 0.0006539297893681043, + "loss": 1.4399, + "theoretical_loss": 3.849247248664043, + "tokens_seen": 629538816 + }, + { + "epoch": 0.19, + "learning_rate": 0.000653913741223671, + "loss": 1.4013, + "theoretical_loss": 3.8492060972356255, + "tokens_seen": 629604352 + }, + { + "epoch": 0.19, + "learning_rate": 0.0006538976930792377, + "loss": 1.4167, + "theoretical_loss": 3.8491649512897017, + "tokens_seen": 629669888 + }, + { + "epoch": 0.19, + "learning_rate": 0.0006538816449348044, + "loss": 1.562, + "theoretical_loss": 3.849123810824971, + "tokens_seen": 629735424 + }, + { + "epoch": 0.19, + "learning_rate": 0.0006538655967903711, + "loss": 1.5798, + "theoretical_loss": 3.8490826758401324, + "tokens_seen": 629800960 + }, + { + "epoch": 0.19, + "learning_rate": 0.0006538495486459378, + "loss": 1.6053, + "theoretical_loss": 3.849041546333887, + "tokens_seen": 629866496 + }, + { + "epoch": 0.19, + "learning_rate": 0.0006538335005015045, + "loss": 1.4632, + "theoretical_loss": 3.8490004223049343, + "tokens_seen": 629932032 + }, + { + "epoch": 0.19, + "learning_rate": 0.0006538174523570712, + "loss": 1.5044, + "theoretical_loss": 3.8489593037519754, + "tokens_seen": 629997568 + }, + { + "epoch": 0.19, + "learning_rate": 0.0006538014042126379, + "loss": 1.4733, + "theoretical_loss": 3.8489181906737118, + "tokens_seen": 630063104 + }, + { + "epoch": 0.19, + "learning_rate": 0.0006537853560682046, + "loss": 1.5427, + "theoretical_loss": 3.848877083068845, + "tokens_seen": 630128640 + }, + { + "epoch": 0.19, + "learning_rate": 0.0006537693079237713, + "loss": 1.591, + "theoretical_loss": 3.8488359809360766, + "tokens_seen": 630194176 + }, + { + "epoch": 0.19, + "learning_rate": 0.000653753259779338, + "loss": 1.5607, + "theoretical_loss": 3.8487948842741107, + "tokens_seen": 630259712 + }, + { + "epoch": 0.19, + "learning_rate": 0.0006537372116349047, + "loss": 1.5137, + "theoretical_loss": 3.8487537930816487, + "tokens_seen": 630325248 + }, + { + "epoch": 0.19, + "learning_rate": 0.0006537211634904714, + "loss": 1.5917, + "theoretical_loss": 3.8487127073573957, + "tokens_seen": 630390784 + }, + { + "epoch": 0.19, + "learning_rate": 0.0006537051153460381, + "loss": 1.5998, + "theoretical_loss": 3.8486716271000545, + "tokens_seen": 630456320 + }, + { + "epoch": 0.19, + "learning_rate": 0.0006536890672016048, + "loss": 1.4978, + "theoretical_loss": 3.8486305523083297, + "tokens_seen": 630521856 + }, + { + "epoch": 0.19, + "learning_rate": 0.0006536730190571715, + "loss": 1.5968, + "theoretical_loss": 3.848589482980927, + "tokens_seen": 630587392 + }, + { + "epoch": 0.19, + "learning_rate": 0.0006536569709127382, + "loss": 1.5125, + "theoretical_loss": 3.8485484191165504, + "tokens_seen": 630652928 + }, + { + "epoch": 0.19, + "learning_rate": 0.000653640922768305, + "loss": 1.5427, + "theoretical_loss": 3.8485073607139064, + "tokens_seen": 630718464 + }, + { + "epoch": 0.19, + "learning_rate": 0.0006536248746238716, + "loss": 1.4974, + "theoretical_loss": 3.8484663077717007, + "tokens_seen": 630784000 + }, + { + "epoch": 0.19, + "learning_rate": 0.0006536088264794383, + "loss": 1.4227, + "theoretical_loss": 3.84842526028864, + "tokens_seen": 630849536 + }, + { + "epoch": 0.19, + "learning_rate": 0.000653592778335005, + "loss": 1.6615, + "theoretical_loss": 3.8483842182634325, + "tokens_seen": 630915072 + }, + { + "epoch": 0.19, + "learning_rate": 0.0006535767301905717, + "loss": 1.5116, + "theoretical_loss": 3.848343181694784, + "tokens_seen": 630980608 + }, + { + "epoch": 0.19, + "learning_rate": 0.0006535606820461384, + "loss": 1.4106, + "theoretical_loss": 3.8483021505814032, + "tokens_seen": 631046144 + }, + { + "epoch": 0.19, + "learning_rate": 0.0006535446339017052, + "loss": 1.5236, + "theoretical_loss": 3.8482611249219985, + "tokens_seen": 631111680 + }, + { + "epoch": 0.19, + "learning_rate": 0.0006535285857572718, + "loss": 1.5722, + "theoretical_loss": 3.8482201047152786, + "tokens_seen": 631177216 + }, + { + "epoch": 0.19, + "learning_rate": 0.0006535125376128386, + "loss": 1.5148, + "theoretical_loss": 3.8481790899599524, + "tokens_seen": 631242752 + }, + { + "epoch": 0.19, + "learning_rate": 0.0006534964894684052, + "loss": 1.6038, + "theoretical_loss": 3.84813808065473, + "tokens_seen": 631308288 + }, + { + "epoch": 0.19, + "learning_rate": 0.000653480441323972, + "loss": 1.5325, + "theoretical_loss": 3.8480970767983216, + "tokens_seen": 631373824 + }, + { + "epoch": 0.19, + "learning_rate": 0.0006534643931795386, + "loss": 1.5989, + "theoretical_loss": 3.8480560783894373, + "tokens_seen": 631439360 + }, + { + "epoch": 0.19, + "learning_rate": 0.0006534483450351054, + "loss": 1.5815, + "theoretical_loss": 3.8480150854267885, + "tokens_seen": 631504896 + }, + { + "epoch": 0.19, + "learning_rate": 0.0006534322968906721, + "loss": 1.4854, + "theoretical_loss": 3.847974097909087, + "tokens_seen": 631570432 + }, + { + "epoch": 0.19, + "learning_rate": 0.0006534162487462388, + "loss": 1.4101, + "theoretical_loss": 3.8479331158350436, + "tokens_seen": 631635968 + }, + { + "epoch": 0.19, + "learning_rate": 0.0006534002006018055, + "loss": 1.4517, + "theoretical_loss": 3.8478921392033714, + "tokens_seen": 631701504 + }, + { + "epoch": 0.19, + "learning_rate": 0.0006533841524573722, + "loss": 1.4487, + "theoretical_loss": 3.8478511680127827, + "tokens_seen": 631767040 + }, + { + "epoch": 0.19, + "learning_rate": 0.0006533681043129389, + "loss": 1.5544, + "theoretical_loss": 3.847810202261991, + "tokens_seen": 631832576 + }, + { + "epoch": 0.19, + "learning_rate": 0.0006533520561685056, + "loss": 1.6663, + "theoretical_loss": 3.8477692419497105, + "tokens_seen": 631898112 + }, + { + "epoch": 0.19, + "learning_rate": 0.0006533360080240723, + "loss": 1.503, + "theoretical_loss": 3.847728287074654, + "tokens_seen": 631963648 + }, + { + "epoch": 0.19, + "learning_rate": 0.000653319959879639, + "loss": 1.5514, + "theoretical_loss": 3.847687337635537, + "tokens_seen": 632029184 + }, + { + "epoch": 0.19, + "learning_rate": 0.0006533039117352057, + "loss": 1.5296, + "theoretical_loss": 3.8476463936310736, + "tokens_seen": 632094720 + }, + { + "epoch": 0.19, + "learning_rate": 0.0006532878635907724, + "loss": 1.5755, + "theoretical_loss": 3.8476054550599796, + "tokens_seen": 632160256 + }, + { + "epoch": 0.19, + "learning_rate": 0.0006532718154463391, + "loss": 1.6096, + "theoretical_loss": 3.8475645219209715, + "tokens_seen": 632225792 + }, + { + "epoch": 0.19, + "learning_rate": 0.0006532557673019058, + "loss": 1.5504, + "theoretical_loss": 3.847523594212765, + "tokens_seen": 632291328 + }, + { + "epoch": 0.19, + "learning_rate": 0.0006532397191574725, + "loss": 1.4807, + "theoretical_loss": 3.8474826719340762, + "tokens_seen": 632356864 + }, + { + "epoch": 0.19, + "objective/train/docs_used": 235545, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.8104701042175293, + "objective/train/theoretical_loss": 3.847441755083623, + "objective/train/tokens_used": 652882400, + "theoretical_loss": 3.847441755083623, + "tokens_seen": 632422400 + }, + { + "epoch": 0.19, + "learning_rate": 0.0006532236710130392, + "loss": 1.6805, + "theoretical_loss": 3.847441755083623, + "tokens_seen": 632422400 + }, + { + "epoch": 0.19, + "learning_rate": 0.0006532076228686059, + "loss": 1.5655, + "theoretical_loss": 3.8474008436601226, + "tokens_seen": 632487936 + }, + { + "epoch": 0.19, + "learning_rate": 0.0006531915747241726, + "loss": 1.5566, + "theoretical_loss": 3.8473599376622927, + "tokens_seen": 632553472 + }, + { + "epoch": 0.19, + "learning_rate": 0.0006531755265797393, + "loss": 1.4381, + "theoretical_loss": 3.8473190370888526, + "tokens_seen": 632619008 + }, + { + "epoch": 0.19, + "learning_rate": 0.000653159478435306, + "loss": 1.4915, + "theoretical_loss": 3.8472781419385207, + "tokens_seen": 632684544 + }, + { + "epoch": 0.19, + "learning_rate": 0.0006531434302908727, + "loss": 1.5029, + "theoretical_loss": 3.8472372522100162, + "tokens_seen": 632750080 + }, + { + "epoch": 0.19, + "learning_rate": 0.0006531273821464394, + "loss": 1.431, + "theoretical_loss": 3.8471963679020593, + "tokens_seen": 632815616 + }, + { + "epoch": 0.19, + "learning_rate": 0.0006531113340020061, + "loss": 1.5532, + "theoretical_loss": 3.8471554890133692, + "tokens_seen": 632881152 + }, + { + "epoch": 0.19, + "learning_rate": 0.0006530952858575728, + "loss": 1.4656, + "theoretical_loss": 3.847114615542667, + "tokens_seen": 632946688 + }, + { + "epoch": 0.19, + "learning_rate": 0.0006530792377131395, + "loss": 1.5346, + "theoretical_loss": 3.8470737474886745, + "tokens_seen": 633012224 + }, + { + "epoch": 0.19, + "learning_rate": 0.0006530631895687062, + "loss": 1.4559, + "theoretical_loss": 3.847032884850112, + "tokens_seen": 633077760 + }, + { + "epoch": 0.19, + "learning_rate": 0.0006530471414242729, + "loss": 1.5505, + "theoretical_loss": 3.846992027625702, + "tokens_seen": 633143296 + }, + { + "epoch": 0.19, + "learning_rate": 0.0006530310932798396, + "loss": 1.4911, + "theoretical_loss": 3.8469511758141666, + "tokens_seen": 633208832 + }, + { + "epoch": 0.19, + "learning_rate": 0.0006530150451354063, + "loss": 1.4551, + "theoretical_loss": 3.8469103294142286, + "tokens_seen": 633274368 + }, + { + "epoch": 0.19, + "learning_rate": 0.000652998996990973, + "loss": 1.4952, + "theoretical_loss": 3.8468694884246117, + "tokens_seen": 633339904 + }, + { + "epoch": 0.19, + "learning_rate": 0.0006529829488465397, + "loss": 1.4621, + "theoretical_loss": 3.8468286528440396, + "tokens_seen": 633405440 + }, + { + "epoch": 0.19, + "learning_rate": 0.0006529669007021064, + "loss": 1.4745, + "theoretical_loss": 3.846787822671235, + "tokens_seen": 633470976 + }, + { + "epoch": 0.19, + "learning_rate": 0.0006529508525576731, + "loss": 1.4851, + "theoretical_loss": 3.8467469979049245, + "tokens_seen": 633536512 + }, + { + "epoch": 0.19, + "learning_rate": 0.0006529348044132398, + "loss": 1.493, + "theoretical_loss": 3.8467061785438315, + "tokens_seen": 633602048 + }, + { + "epoch": 0.19, + "learning_rate": 0.0006529187562688064, + "loss": 1.4304, + "theoretical_loss": 3.8466653645866815, + "tokens_seen": 633667584 + }, + { + "epoch": 0.19, + "learning_rate": 0.0006529027081243732, + "loss": 1.5552, + "theoretical_loss": 3.8466245560322005, + "tokens_seen": 633733120 + }, + { + "epoch": 0.19, + "learning_rate": 0.0006528866599799398, + "loss": 1.578, + "theoretical_loss": 3.8465837528791154, + "tokens_seen": 633798656 + }, + { + "epoch": 0.19, + "learning_rate": 0.0006528706118355066, + "loss": 1.5261, + "theoretical_loss": 3.846542955126152, + "tokens_seen": 633864192 + }, + { + "epoch": 0.19, + "learning_rate": 0.0006528545636910732, + "loss": 1.5365, + "theoretical_loss": 3.8465021627720377, + "tokens_seen": 633929728 + }, + { + "epoch": 0.19, + "learning_rate": 0.00065283851554664, + "loss": 1.6227, + "theoretical_loss": 3.8464613758155, + "tokens_seen": 633995264 + }, + { + "epoch": 0.19, + "learning_rate": 0.0006528224674022066, + "loss": 1.5918, + "theoretical_loss": 3.8464205942552665, + "tokens_seen": 634060800 + }, + { + "epoch": 0.19, + "learning_rate": 0.0006528064192577734, + "loss": 1.6174, + "theoretical_loss": 3.8463798180900666, + "tokens_seen": 634126336 + }, + { + "epoch": 0.19, + "learning_rate": 0.00065279037111334, + "loss": 1.5099, + "theoretical_loss": 3.8463390473186285, + "tokens_seen": 634191872 + }, + { + "epoch": 0.19, + "learning_rate": 0.0006527743229689067, + "loss": 1.4995, + "theoretical_loss": 3.8462982819396814, + "tokens_seen": 634257408 + }, + { + "epoch": 0.19, + "learning_rate": 0.0006527582748244734, + "loss": 1.4596, + "theoretical_loss": 3.846257521951955, + "tokens_seen": 634322944 + }, + { + "epoch": 0.19, + "learning_rate": 0.0006527422266800401, + "loss": 1.5005, + "theoretical_loss": 3.84621676735418, + "tokens_seen": 634388480 + }, + { + "epoch": 0.19, + "learning_rate": 0.0006527261785356068, + "loss": 1.5907, + "theoretical_loss": 3.8461760181450857, + "tokens_seen": 634454016 + }, + { + "epoch": 0.19, + "learning_rate": 0.0006527101303911735, + "loss": 1.4642, + "theoretical_loss": 3.8461352743234043, + "tokens_seen": 634519552 + }, + { + "epoch": 0.19, + "learning_rate": 0.0006526940822467402, + "loss": 1.5235, + "theoretical_loss": 3.8460945358878664, + "tokens_seen": 634585088 + }, + { + "epoch": 0.19, + "learning_rate": 0.0006526780341023069, + "loss": 1.5777, + "theoretical_loss": 3.846053802837204, + "tokens_seen": 634650624 + }, + { + "epoch": 0.19, + "learning_rate": 0.0006526619859578736, + "loss": 1.5356, + "theoretical_loss": 3.8460130751701502, + "tokens_seen": 634716160 + }, + { + "epoch": 0.19, + "learning_rate": 0.0006526459378134403, + "loss": 1.4852, + "theoretical_loss": 3.845972352885437, + "tokens_seen": 634781696 + }, + { + "epoch": 0.19, + "learning_rate": 0.000652629889669007, + "loss": 1.5148, + "theoretical_loss": 3.845931635981797, + "tokens_seen": 634847232 + }, + { + "epoch": 0.19, + "learning_rate": 0.0006526138415245737, + "loss": 1.504, + "theoretical_loss": 3.845890924457965, + "tokens_seen": 634912768 + }, + { + "epoch": 0.19, + "learning_rate": 0.0006525977933801404, + "loss": 1.4691, + "theoretical_loss": 3.845850218312674, + "tokens_seen": 634978304 + }, + { + "epoch": 0.19, + "learning_rate": 0.0006525817452357071, + "loss": 1.6182, + "theoretical_loss": 3.8458095175446587, + "tokens_seen": 635043840 + }, + { + "epoch": 0.19, + "learning_rate": 0.0006525656970912738, + "loss": 1.5147, + "theoretical_loss": 3.8457688221526545, + "tokens_seen": 635109376 + }, + { + "epoch": 0.19, + "learning_rate": 0.0006525496489468405, + "loss": 1.4993, + "theoretical_loss": 3.8457281321353953, + "tokens_seen": 635174912 + }, + { + "epoch": 0.19, + "learning_rate": 0.0006525336008024072, + "loss": 1.4924, + "theoretical_loss": 3.8456874474916183, + "tokens_seen": 635240448 + }, + { + "epoch": 0.19, + "learning_rate": 0.0006525175526579739, + "loss": 1.4868, + "theoretical_loss": 3.8456467682200586, + "tokens_seen": 635305984 + }, + { + "epoch": 0.19, + "learning_rate": 0.0006525015045135406, + "loss": 1.5405, + "theoretical_loss": 3.8456060943194537, + "tokens_seen": 635371520 + }, + { + "epoch": 0.19, + "learning_rate": 0.0006524854563691073, + "loss": 1.4769, + "theoretical_loss": 3.845565425788539, + "tokens_seen": 635437056 + }, + { + "epoch": 0.19, + "learning_rate": 0.000652469408224674, + "loss": 1.5327, + "theoretical_loss": 3.8455247626260536, + "tokens_seen": 635502592 + }, + { + "epoch": 0.19, + "learning_rate": 0.0006524533600802407, + "loss": 1.5949, + "theoretical_loss": 3.845484104830734, + "tokens_seen": 635568128 + }, + { + "epoch": 0.19, + "learning_rate": 0.0006524373119358074, + "loss": 1.4983, + "theoretical_loss": 3.84544345240132, + "tokens_seen": 635633664 + }, + { + "epoch": 0.19, + "objective/train/docs_used": 236526, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.5431077480316162, + "objective/train/theoretical_loss": 3.8454028053365494, + "objective/train/tokens_used": 656159200, + "theoretical_loss": 3.8454028053365494, + "tokens_seen": 635699200 + }, + { + "epoch": 0.19, + "learning_rate": 0.0006524212637913741, + "loss": 1.517, + "theoretical_loss": 3.8454028053365494, + "tokens_seen": 635699200 + }, + { + "epoch": 0.19, + "learning_rate": 0.0006524052156469408, + "loss": 1.5461, + "theoretical_loss": 3.8453621636351607, + "tokens_seen": 635764736 + }, + { + "epoch": 0.19, + "learning_rate": 0.0006523891675025075, + "loss": 1.5018, + "theoretical_loss": 3.8453215272958943, + "tokens_seen": 635830272 + }, + { + "epoch": 0.19, + "learning_rate": 0.0006523731193580742, + "loss": 1.5498, + "theoretical_loss": 3.84528089631749, + "tokens_seen": 635895808 + }, + { + "epoch": 0.19, + "learning_rate": 0.0006523570712136409, + "loss": 1.4352, + "theoretical_loss": 3.845240270698688, + "tokens_seen": 635961344 + }, + { + "epoch": 0.19, + "learning_rate": 0.0006523410230692076, + "loss": 1.5061, + "theoretical_loss": 3.845199650438229, + "tokens_seen": 636026880 + }, + { + "epoch": 0.19, + "learning_rate": 0.0006523249749247743, + "loss": 1.5364, + "theoretical_loss": 3.845159035534855, + "tokens_seen": 636092416 + }, + { + "epoch": 0.19, + "learning_rate": 0.000652308926780341, + "loss": 1.5771, + "theoretical_loss": 3.845118425987307, + "tokens_seen": 636157952 + }, + { + "epoch": 0.19, + "learning_rate": 0.0006522928786359077, + "loss": 1.5796, + "theoretical_loss": 3.845077821794327, + "tokens_seen": 636223488 + }, + { + "epoch": 0.19, + "learning_rate": 0.0006522768304914744, + "loss": 1.5889, + "theoretical_loss": 3.845037222954658, + "tokens_seen": 636289024 + }, + { + "epoch": 0.19, + "learning_rate": 0.0006522607823470411, + "loss": 1.5708, + "theoretical_loss": 3.8449966294670426, + "tokens_seen": 636354560 + }, + { + "epoch": 0.19, + "learning_rate": 0.0006522447342026078, + "loss": 1.5372, + "theoretical_loss": 3.8449560413302244, + "tokens_seen": 636420096 + }, + { + "epoch": 0.19, + "learning_rate": 0.0006522286860581746, + "loss": 1.6235, + "theoretical_loss": 3.844915458542947, + "tokens_seen": 636485632 + }, + { + "epoch": 0.19, + "learning_rate": 0.0006522126379137412, + "loss": 1.4937, + "theoretical_loss": 3.844874881103955, + "tokens_seen": 636551168 + }, + { + "epoch": 0.19, + "learning_rate": 0.000652196589769308, + "loss": 1.4658, + "theoretical_loss": 3.844834309011992, + "tokens_seen": 636616704 + }, + { + "epoch": 0.19, + "learning_rate": 0.0006521805416248747, + "loss": 1.4683, + "theoretical_loss": 3.8447937422658045, + "tokens_seen": 636682240 + }, + { + "epoch": 0.19, + "learning_rate": 0.0006521644934804414, + "loss": 1.4945, + "theoretical_loss": 3.8447531808641373, + "tokens_seen": 636747776 + }, + { + "epoch": 0.19, + "learning_rate": 0.000652148445336008, + "loss": 1.4981, + "theoretical_loss": 3.844712624805736, + "tokens_seen": 636813312 + }, + { + "epoch": 0.19, + "learning_rate": 0.0006521323971915748, + "loss": 1.7662, + "theoretical_loss": 3.844672074089347, + "tokens_seen": 636878848 + }, + { + "epoch": 0.19, + "learning_rate": 0.0006521163490471415, + "loss": 1.4707, + "theoretical_loss": 3.844631528713718, + "tokens_seen": 636944384 + }, + { + "epoch": 0.19, + "learning_rate": 0.0006521003009027082, + "loss": 1.5616, + "theoretical_loss": 3.844590988677595, + "tokens_seen": 637009920 + }, + { + "epoch": 0.19, + "learning_rate": 0.0006520842527582749, + "loss": 1.4525, + "theoretical_loss": 3.844550453979726, + "tokens_seen": 637075456 + }, + { + "epoch": 0.19, + "learning_rate": 0.0006520682046138416, + "loss": 1.5429, + "theoretical_loss": 3.8445099246188597, + "tokens_seen": 637140992 + }, + { + "epoch": 0.19, + "learning_rate": 0.0006520521564694083, + "loss": 1.4458, + "theoretical_loss": 3.8444694005937436, + "tokens_seen": 637206528 + }, + { + "epoch": 0.19, + "learning_rate": 0.000652036108324975, + "loss": 1.6, + "theoretical_loss": 3.844428881903127, + "tokens_seen": 637272064 + }, + { + "epoch": 0.19, + "learning_rate": 0.0006520200601805417, + "loss": 1.5108, + "theoretical_loss": 3.844388368545759, + "tokens_seen": 637337600 + }, + { + "epoch": 0.19, + "learning_rate": 0.0006520040120361084, + "loss": 1.4881, + "theoretical_loss": 3.84434786052039, + "tokens_seen": 637403136 + }, + { + "epoch": 0.19, + "learning_rate": 0.0006519879638916751, + "loss": 1.4962, + "theoretical_loss": 3.8443073578257687, + "tokens_seen": 637468672 + }, + { + "epoch": 0.19, + "learning_rate": 0.0006519719157472418, + "loss": 1.4378, + "theoretical_loss": 3.8442668604606474, + "tokens_seen": 637534208 + }, + { + "epoch": 0.19, + "learning_rate": 0.0006519558676028085, + "loss": 1.5242, + "theoretical_loss": 3.844226368423776, + "tokens_seen": 637599744 + }, + { + "epoch": 0.19, + "learning_rate": 0.0006519398194583752, + "loss": 1.3821, + "theoretical_loss": 3.844185881713906, + "tokens_seen": 637665280 + }, + { + "epoch": 0.19, + "learning_rate": 0.0006519237713139419, + "loss": 1.5098, + "theoretical_loss": 3.844145400329789, + "tokens_seen": 637730816 + }, + { + "epoch": 0.19, + "learning_rate": 0.0006519077231695086, + "loss": 1.5252, + "theoretical_loss": 3.8441049242701784, + "tokens_seen": 637796352 + }, + { + "epoch": 0.19, + "learning_rate": 0.0006518916750250753, + "loss": 1.5354, + "theoretical_loss": 3.8440644535338255, + "tokens_seen": 637861888 + }, + { + "epoch": 0.19, + "learning_rate": 0.000651875626880642, + "loss": 1.5085, + "theoretical_loss": 3.844023988119484, + "tokens_seen": 637927424 + }, + { + "epoch": 0.19, + "learning_rate": 0.0006518595787362087, + "loss": 1.4805, + "theoretical_loss": 3.8439835280259076, + "tokens_seen": 637992960 + }, + { + "epoch": 0.19, + "learning_rate": 0.0006518435305917754, + "loss": 1.5224, + "theoretical_loss": 3.84394307325185, + "tokens_seen": 638058496 + }, + { + "epoch": 0.19, + "learning_rate": 0.0006518274824473421, + "loss": 1.6155, + "theoretical_loss": 3.8439026237960654, + "tokens_seen": 638124032 + }, + { + "epoch": 0.19, + "learning_rate": 0.0006518114343029088, + "loss": 1.5042, + "theoretical_loss": 3.8438621796573087, + "tokens_seen": 638189568 + }, + { + "epoch": 0.19, + "learning_rate": 0.0006517953861584755, + "loss": 1.4422, + "theoretical_loss": 3.843821740834335, + "tokens_seen": 638255104 + }, + { + "epoch": 0.19, + "learning_rate": 0.0006517793380140422, + "loss": 1.4155, + "theoretical_loss": 3.8437813073258997, + "tokens_seen": 638320640 + }, + { + "epoch": 0.19, + "learning_rate": 0.0006517632898696089, + "loss": 1.5092, + "theoretical_loss": 3.8437408791307597, + "tokens_seen": 638386176 + }, + { + "epoch": 0.19, + "learning_rate": 0.0006517472417251756, + "loss": 1.551, + "theoretical_loss": 3.843700456247671, + "tokens_seen": 638451712 + }, + { + "epoch": 0.19, + "learning_rate": 0.0006517311935807423, + "loss": 1.518, + "theoretical_loss": 3.84366003867539, + "tokens_seen": 638517248 + }, + { + "epoch": 0.19, + "learning_rate": 0.000651715145436309, + "loss": 1.5945, + "theoretical_loss": 3.8436196264126736, + "tokens_seen": 638582784 + }, + { + "epoch": 0.19, + "learning_rate": 0.0006516990972918757, + "loss": 1.5451, + "theoretical_loss": 3.843579219458281, + "tokens_seen": 638648320 + }, + { + "epoch": 0.19, + "learning_rate": 0.0006516830491474424, + "loss": 1.6382, + "theoretical_loss": 3.843538817810969, + "tokens_seen": 638713856 + }, + { + "epoch": 0.19, + "learning_rate": 0.0006516670010030091, + "loss": 1.5596, + "theoretical_loss": 3.843498421469497, + "tokens_seen": 638779392 + }, + { + "epoch": 0.19, + "learning_rate": 0.0006516509528585758, + "loss": 1.4754, + "theoretical_loss": 3.843458030432623, + "tokens_seen": 638844928 + }, + { + "epoch": 0.19, + "learning_rate": 0.0006516349047141425, + "loss": 1.5149, + "theoretical_loss": 3.8434176446991075, + "tokens_seen": 638910464 + }, + { + "epoch": 0.19, + "objective/train/docs_used": 237858, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.3317413330078125, + "objective/train/theoretical_loss": 3.8433772642677093, + "objective/train/tokens_used": 659436000, + "theoretical_loss": 3.8433772642677093, + "tokens_seen": 638976000 + }, + { + "epoch": 0.19, + "learning_rate": 0.0006516188565697092, + "loss": 1.2983, + "theoretical_loss": 3.8433772642677093, + "tokens_seen": 638976000 + }, + { + "epoch": 0.19, + "learning_rate": 0.0006516028084252759, + "loss": 1.5331, + "theoretical_loss": 3.8433368891371895, + "tokens_seen": 639041536 + }, + { + "epoch": 0.19, + "learning_rate": 0.0006515867602808426, + "loss": 1.5696, + "theoretical_loss": 3.843296519306308, + "tokens_seen": 639107072 + }, + { + "epoch": 0.19, + "learning_rate": 0.0006515707121364093, + "loss": 1.5003, + "theoretical_loss": 3.843256154773826, + "tokens_seen": 639172608 + }, + { + "epoch": 0.19, + "learning_rate": 0.000651554663991976, + "loss": 1.5006, + "theoretical_loss": 3.843215795538505, + "tokens_seen": 639238144 + }, + { + "epoch": 0.19, + "learning_rate": 0.0006515386158475427, + "loss": 1.5699, + "theoretical_loss": 3.843175441599107, + "tokens_seen": 639303680 + }, + { + "epoch": 0.19, + "learning_rate": 0.0006515225677031094, + "loss": 1.6832, + "theoretical_loss": 3.8431350929543946, + "tokens_seen": 639369216 + }, + { + "epoch": 0.19, + "learning_rate": 0.0006515065195586761, + "loss": 1.4959, + "theoretical_loss": 3.8430947496031296, + "tokens_seen": 639434752 + }, + { + "epoch": 0.19, + "learning_rate": 0.0006514904714142427, + "loss": 1.4057, + "theoretical_loss": 3.8430544115440752, + "tokens_seen": 639500288 + }, + { + "epoch": 0.19, + "learning_rate": 0.0006514744232698095, + "loss": 1.5307, + "theoretical_loss": 3.8430140787759957, + "tokens_seen": 639565824 + }, + { + "epoch": 0.19, + "learning_rate": 0.0006514583751253761, + "loss": 1.5433, + "theoretical_loss": 3.842973751297655, + "tokens_seen": 639631360 + }, + { + "epoch": 0.19, + "learning_rate": 0.0006514423269809429, + "loss": 1.5115, + "theoretical_loss": 3.8429334291078168, + "tokens_seen": 639696896 + }, + { + "epoch": 0.19, + "learning_rate": 0.0006514262788365095, + "loss": 1.4969, + "theoretical_loss": 3.842893112205246, + "tokens_seen": 639762432 + }, + { + "epoch": 0.19, + "learning_rate": 0.0006514102306920763, + "loss": 1.5179, + "theoretical_loss": 3.842852800588709, + "tokens_seen": 639827968 + }, + { + "epoch": 0.19, + "learning_rate": 0.0006513941825476429, + "loss": 1.492, + "theoretical_loss": 3.8428124942569695, + "tokens_seen": 639893504 + }, + { + "epoch": 0.19, + "learning_rate": 0.0006513781344032097, + "loss": 1.4784, + "theoretical_loss": 3.842772193208795, + "tokens_seen": 639959040 + }, + { + "epoch": 0.19, + "learning_rate": 0.0006513620862587763, + "loss": 1.4628, + "theoretical_loss": 3.8427318974429507, + "tokens_seen": 640024576 + }, + { + "epoch": 0.19, + "learning_rate": 0.000651346038114343, + "loss": 1.4491, + "theoretical_loss": 3.8426916069582044, + "tokens_seen": 640090112 + }, + { + "epoch": 0.19, + "learning_rate": 0.0006513299899699097, + "loss": 1.536, + "theoretical_loss": 3.842651321753323, + "tokens_seen": 640155648 + }, + { + "epoch": 0.19, + "learning_rate": 0.0006513139418254764, + "loss": 1.5931, + "theoretical_loss": 3.8426110418270745, + "tokens_seen": 640221184 + }, + { + "epoch": 0.19, + "learning_rate": 0.0006512978936810431, + "loss": 1.6238, + "theoretical_loss": 3.842570767178227, + "tokens_seen": 640286720 + }, + { + "epoch": 0.19, + "learning_rate": 0.0006512818455366098, + "loss": 1.4529, + "theoretical_loss": 3.8425304978055483, + "tokens_seen": 640352256 + }, + { + "epoch": 0.19, + "learning_rate": 0.0006512657973921765, + "loss": 1.5431, + "theoretical_loss": 3.8424902337078084, + "tokens_seen": 640417792 + }, + { + "epoch": 0.19, + "learning_rate": 0.0006512497492477432, + "loss": 1.6034, + "theoretical_loss": 3.8424499748837753, + "tokens_seen": 640483328 + }, + { + "epoch": 0.19, + "learning_rate": 0.0006512337011033099, + "loss": 1.4604, + "theoretical_loss": 3.8424097213322197, + "tokens_seen": 640548864 + }, + { + "epoch": 0.19, + "learning_rate": 0.0006512176529588766, + "loss": 1.4291, + "theoretical_loss": 3.8423694730519116, + "tokens_seen": 640614400 + }, + { + "epoch": 0.19, + "learning_rate": 0.0006512016048144433, + "loss": 1.5715, + "theoretical_loss": 3.8423292300416216, + "tokens_seen": 640679936 + }, + { + "epoch": 0.19, + "learning_rate": 0.00065118555667001, + "loss": 1.5065, + "theoretical_loss": 3.8422889923001207, + "tokens_seen": 640745472 + }, + { + "epoch": 0.19, + "learning_rate": 0.0006511695085255767, + "loss": 1.4979, + "theoretical_loss": 3.84224875982618, + "tokens_seen": 640811008 + }, + { + "epoch": 0.19, + "learning_rate": 0.0006511534603811434, + "loss": 1.382, + "theoretical_loss": 3.8422085326185718, + "tokens_seen": 640876544 + }, + { + "epoch": 0.19, + "learning_rate": 0.0006511374122367101, + "loss": 1.5883, + "theoretical_loss": 3.842168310676067, + "tokens_seen": 640942080 + }, + { + "epoch": 0.19, + "learning_rate": 0.0006511213640922768, + "loss": 1.4969, + "theoretical_loss": 3.8421280939974407, + "tokens_seen": 641007616 + }, + { + "epoch": 0.19, + "learning_rate": 0.0006511053159478435, + "loss": 1.5739, + "theoretical_loss": 3.8420878825814637, + "tokens_seen": 641073152 + }, + { + "epoch": 0.19, + "learning_rate": 0.0006510892678034102, + "loss": 1.5612, + "theoretical_loss": 3.84204767642691, + "tokens_seen": 641138688 + }, + { + "epoch": 0.19, + "learning_rate": 0.0006510732196589769, + "loss": 1.51, + "theoretical_loss": 3.8420074755325544, + "tokens_seen": 641204224 + }, + { + "epoch": 0.19, + "learning_rate": 0.0006510571715145436, + "loss": 1.5315, + "theoretical_loss": 3.8419672798971707, + "tokens_seen": 641269760 + }, + { + "epoch": 0.19, + "learning_rate": 0.0006510411233701103, + "loss": 1.5465, + "theoretical_loss": 3.8419270895195323, + "tokens_seen": 641335296 + }, + { + "epoch": 0.19, + "learning_rate": 0.0006510250752256771, + "loss": 1.5257, + "theoretical_loss": 3.8418869043984163, + "tokens_seen": 641400832 + }, + { + "epoch": 0.19, + "learning_rate": 0.0006510090270812437, + "loss": 1.5543, + "theoretical_loss": 3.841846724532597, + "tokens_seen": 641466368 + }, + { + "epoch": 0.19, + "learning_rate": 0.0006509929789368105, + "loss": 1.5121, + "theoretical_loss": 3.8418065499208502, + "tokens_seen": 641531904 + }, + { + "epoch": 0.19, + "learning_rate": 0.0006509769307923771, + "loss": 1.4938, + "theoretical_loss": 3.8417663805619533, + "tokens_seen": 641597440 + }, + { + "epoch": 0.19, + "learning_rate": 0.0006509608826479439, + "loss": 1.4596, + "theoretical_loss": 3.841726216454682, + "tokens_seen": 641662976 + }, + { + "epoch": 0.19, + "learning_rate": 0.0006509448345035105, + "loss": 1.5959, + "theoretical_loss": 3.8416860575978142, + "tokens_seen": 641728512 + }, + { + "epoch": 0.19, + "learning_rate": 0.0006509287863590773, + "loss": 1.4392, + "theoretical_loss": 3.841645903990127, + "tokens_seen": 641794048 + }, + { + "epoch": 0.19, + "learning_rate": 0.000650912738214644, + "loss": 1.4614, + "theoretical_loss": 3.841605755630398, + "tokens_seen": 641859584 + }, + { + "epoch": 0.19, + "learning_rate": 0.0006508966900702107, + "loss": 1.6054, + "theoretical_loss": 3.8415656125174067, + "tokens_seen": 641925120 + }, + { + "epoch": 0.19, + "learning_rate": 0.0006508806419257773, + "loss": 1.464, + "theoretical_loss": 3.8415254746499308, + "tokens_seen": 641990656 + }, + { + "epoch": 0.19, + "learning_rate": 0.0006508645937813441, + "loss": 1.4817, + "theoretical_loss": 3.8414853420267496, + "tokens_seen": 642056192 + }, + { + "epoch": 0.19, + "learning_rate": 0.0006508485456369107, + "loss": 1.3951, + "theoretical_loss": 3.841445214646644, + "tokens_seen": 642121728 + }, + { + "epoch": 0.19, + "learning_rate": 0.0006508324974924775, + "loss": 1.5195, + "theoretical_loss": 3.8414050925083925, + "tokens_seen": 642187264 + }, + { + "epoch": 0.19, + "objective/train/docs_used": 239059, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.5250208377838135, + "objective/train/theoretical_loss": 3.8413649756107757, + "objective/train/tokens_used": 662712800, + "theoretical_loss": 3.8413649756107757, + "tokens_seen": 642252800 + }, + { + "epoch": 0.19, + "learning_rate": 0.0006508164493480441, + "loss": 1.6323, + "theoretical_loss": 3.8413649756107757, + "tokens_seen": 642252800 + }, + { + "epoch": 0.19, + "learning_rate": 0.0006508004012036109, + "loss": 1.5142, + "theoretical_loss": 3.8413248639525754, + "tokens_seen": 642318336 + }, + { + "epoch": 0.19, + "learning_rate": 0.0006507843530591776, + "loss": 1.5262, + "theoretical_loss": 3.8412847575325717, + "tokens_seen": 642383872 + }, + { + "epoch": 0.19, + "learning_rate": 0.0006507683049147443, + "loss": 1.6426, + "theoretical_loss": 3.8412446563495477, + "tokens_seen": 642449408 + }, + { + "epoch": 0.19, + "learning_rate": 0.000650752256770311, + "loss": 1.6462, + "theoretical_loss": 3.8412045604022835, + "tokens_seen": 642514944 + }, + { + "epoch": 0.19, + "learning_rate": 0.0006507362086258777, + "loss": 1.4705, + "theoretical_loss": 3.8411644696895633, + "tokens_seen": 642580480 + }, + { + "epoch": 0.19, + "learning_rate": 0.0006507201604814444, + "loss": 1.4686, + "theoretical_loss": 3.8411243842101688, + "tokens_seen": 642646016 + }, + { + "epoch": 0.19, + "learning_rate": 0.0006507041123370111, + "loss": 1.5527, + "theoretical_loss": 3.841084303962884, + "tokens_seen": 642711552 + }, + { + "epoch": 0.19, + "learning_rate": 0.0006506880641925778, + "loss": 1.5091, + "theoretical_loss": 3.841044228946492, + "tokens_seen": 642777088 + }, + { + "epoch": 0.19, + "learning_rate": 0.0006506720160481445, + "loss": 1.4465, + "theoretical_loss": 3.8410041591597777, + "tokens_seen": 642842624 + }, + { + "epoch": 0.19, + "learning_rate": 0.0006506559679037112, + "loss": 1.5002, + "theoretical_loss": 3.840964094601525, + "tokens_seen": 642908160 + }, + { + "epoch": 0.19, + "learning_rate": 0.0006506399197592779, + "loss": 1.5388, + "theoretical_loss": 3.8409240352705187, + "tokens_seen": 642973696 + }, + { + "epoch": 0.19, + "learning_rate": 0.0006506238716148446, + "loss": 1.5039, + "theoretical_loss": 3.840883981165544, + "tokens_seen": 643039232 + }, + { + "epoch": 0.19, + "learning_rate": 0.0006506078234704113, + "loss": 1.5974, + "theoretical_loss": 3.8408439322853876, + "tokens_seen": 643104768 + }, + { + "epoch": 0.19, + "learning_rate": 0.000650591775325978, + "loss": 1.4358, + "theoretical_loss": 3.8408038886288347, + "tokens_seen": 643170304 + }, + { + "epoch": 0.19, + "learning_rate": 0.0006505757271815447, + "loss": 1.552, + "theoretical_loss": 3.840763850194672, + "tokens_seen": 643235840 + }, + { + "epoch": 0.19, + "learning_rate": 0.0006505596790371114, + "loss": 1.5648, + "theoretical_loss": 3.8407238169816864, + "tokens_seen": 643301376 + }, + { + "epoch": 0.19, + "learning_rate": 0.0006505436308926781, + "loss": 1.5244, + "theoretical_loss": 3.840683788988666, + "tokens_seen": 643366912 + }, + { + "epoch": 0.19, + "learning_rate": 0.0006505275827482448, + "loss": 1.4798, + "theoretical_loss": 3.840643766214397, + "tokens_seen": 643432448 + }, + { + "epoch": 0.19, + "learning_rate": 0.0006505115346038115, + "loss": 1.4849, + "theoretical_loss": 3.840603748657669, + "tokens_seen": 643497984 + }, + { + "epoch": 0.2, + "learning_rate": 0.0006504954864593782, + "loss": 1.4124, + "theoretical_loss": 3.8405637363172693, + "tokens_seen": 643563520 + }, + { + "epoch": 0.2, + "learning_rate": 0.0006504794383149449, + "loss": 1.3958, + "theoretical_loss": 3.8405237291919883, + "tokens_seen": 643629056 + }, + { + "epoch": 0.2, + "learning_rate": 0.0006504633901705116, + "loss": 1.5062, + "theoretical_loss": 3.8404837272806143, + "tokens_seen": 643694592 + }, + { + "epoch": 0.2, + "learning_rate": 0.0006504473420260783, + "loss": 1.5075, + "theoretical_loss": 3.840443730581937, + "tokens_seen": 643760128 + }, + { + "epoch": 0.2, + "learning_rate": 0.000650431293881645, + "loss": 1.5295, + "theoretical_loss": 3.8404037390947474, + "tokens_seen": 643825664 + }, + { + "epoch": 0.2, + "learning_rate": 0.0006504152457372117, + "loss": 1.4657, + "theoretical_loss": 3.8403637528178356, + "tokens_seen": 643891200 + }, + { + "epoch": 0.2, + "learning_rate": 0.0006503991975927784, + "loss": 1.4955, + "theoretical_loss": 3.8403237717499925, + "tokens_seen": 643956736 + }, + { + "epoch": 0.2, + "learning_rate": 0.0006503831494483451, + "loss": 1.5179, + "theoretical_loss": 3.8402837958900093, + "tokens_seen": 644022272 + }, + { + "epoch": 0.2, + "learning_rate": 0.0006503671013039118, + "loss": 1.5006, + "theoretical_loss": 3.840243825236678, + "tokens_seen": 644087808 + }, + { + "epoch": 0.2, + "learning_rate": 0.0006503510531594785, + "loss": 1.4659, + "theoretical_loss": 3.840203859788791, + "tokens_seen": 644153344 + }, + { + "epoch": 0.2, + "learning_rate": 0.0006503350050150452, + "loss": 1.4561, + "theoretical_loss": 3.840163899545141, + "tokens_seen": 644218880 + }, + { + "epoch": 0.2, + "learning_rate": 0.0006503189568706119, + "loss": 1.5276, + "theoretical_loss": 3.8401239445045205, + "tokens_seen": 644284416 + }, + { + "epoch": 0.2, + "learning_rate": 0.0006503029087261786, + "loss": 1.4959, + "theoretical_loss": 3.840083994665723, + "tokens_seen": 644349952 + }, + { + "epoch": 0.2, + "learning_rate": 0.0006502868605817453, + "loss": 1.454, + "theoretical_loss": 3.840044050027543, + "tokens_seen": 644415488 + }, + { + "epoch": 0.2, + "learning_rate": 0.000650270812437312, + "loss": 1.4729, + "theoretical_loss": 3.8400041105887732, + "tokens_seen": 644481024 + }, + { + "epoch": 0.2, + "learning_rate": 0.0006502547642928787, + "loss": 1.5991, + "theoretical_loss": 3.8399641763482095, + "tokens_seen": 644546560 + }, + { + "epoch": 0.2, + "learning_rate": 0.0006502387161484454, + "loss": 1.5533, + "theoretical_loss": 3.8399242473046464, + "tokens_seen": 644612096 + }, + { + "epoch": 0.2, + "learning_rate": 0.0006502226680040121, + "loss": 1.5463, + "theoretical_loss": 3.8398843234568796, + "tokens_seen": 644677632 + }, + { + "epoch": 0.2, + "learning_rate": 0.0006502066198595788, + "loss": 1.5205, + "theoretical_loss": 3.8398444048037046, + "tokens_seen": 644743168 + }, + { + "epoch": 0.2, + "learning_rate": 0.0006501905717151455, + "loss": 1.5124, + "theoretical_loss": 3.8398044913439175, + "tokens_seen": 644808704 + }, + { + "epoch": 0.2, + "learning_rate": 0.0006501745235707122, + "loss": 1.4609, + "theoretical_loss": 3.839764583076316, + "tokens_seen": 644874240 + }, + { + "epoch": 0.2, + "learning_rate": 0.0006501584754262789, + "loss": 1.5895, + "theoretical_loss": 3.839724679999695, + "tokens_seen": 644939776 + }, + { + "epoch": 0.2, + "learning_rate": 0.0006501424272818456, + "loss": 1.5978, + "theoretical_loss": 3.8396847821128537, + "tokens_seen": 645005312 + }, + { + "epoch": 0.2, + "learning_rate": 0.0006501263791374123, + "loss": 1.6308, + "theoretical_loss": 3.83964488941459, + "tokens_seen": 645070848 + }, + { + "epoch": 0.2, + "learning_rate": 0.000650110330992979, + "loss": 1.5215, + "theoretical_loss": 3.8396050019037005, + "tokens_seen": 645136384 + }, + { + "epoch": 0.2, + "learning_rate": 0.0006500942828485456, + "loss": 1.5026, + "theoretical_loss": 3.8395651195789853, + "tokens_seen": 645201920 + }, + { + "epoch": 0.2, + "learning_rate": 0.0006500782347041124, + "loss": 1.5383, + "theoretical_loss": 3.839525242439243, + "tokens_seen": 645267456 + }, + { + "epoch": 0.2, + "learning_rate": 0.000650062186559679, + "loss": 1.5179, + "theoretical_loss": 3.8394853704832723, + "tokens_seen": 645332992 + }, + { + "epoch": 0.2, + "learning_rate": 0.0006500461384152458, + "loss": 1.565, + "theoretical_loss": 3.8394455037098743, + "tokens_seen": 645398528 + }, + { + "epoch": 0.2, + "learning_rate": 0.0006500300902708124, + "loss": 1.5307, + "theoretical_loss": 3.8394056421178484, + "tokens_seen": 645464064 + }, + { + "epoch": 0.2, + "objective/train/docs_used": 240316, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.5348176956176758, + "objective/train/theoretical_loss": 3.8393657857059953, + "objective/train/tokens_used": 665989600, + "theoretical_loss": 3.8393657857059953, + "tokens_seen": 645529600 + }, + { + "epoch": 0.2, + "learning_rate": 0.0006500140421263792, + "loss": 1.4863, + "theoretical_loss": 3.8393657857059953, + "tokens_seen": 645529600 + }, + { + "epoch": 0.2, + "learning_rate": 0.0006499979939819458, + "loss": 1.6117, + "theoretical_loss": 3.839325934473116, + "tokens_seen": 645595136 + }, + { + "epoch": 0.2, + "learning_rate": 0.0006499819458375126, + "loss": 1.7039, + "theoretical_loss": 3.839286088418012, + "tokens_seen": 645660672 + }, + { + "epoch": 0.2, + "learning_rate": 0.0006499658976930792, + "loss": 1.505, + "theoretical_loss": 3.839246247539485, + "tokens_seen": 645726208 + }, + { + "epoch": 0.2, + "learning_rate": 0.0006499498495486459, + "loss": 1.5777, + "theoretical_loss": 3.8392064118363374, + "tokens_seen": 645791744 + }, + { + "epoch": 0.2, + "learning_rate": 0.0006499338014042126, + "loss": 1.5503, + "theoretical_loss": 3.839166581307372, + "tokens_seen": 645857280 + }, + { + "epoch": 0.2, + "learning_rate": 0.0006499177532597793, + "loss": 1.526, + "theoretical_loss": 3.8391267559513915, + "tokens_seen": 645922816 + }, + { + "epoch": 0.2, + "learning_rate": 0.000649901705115346, + "loss": 1.5761, + "theoretical_loss": 3.839086935767199, + "tokens_seen": 645988352 + }, + { + "epoch": 0.2, + "learning_rate": 0.0006498856569709127, + "loss": 1.5771, + "theoretical_loss": 3.839047120753599, + "tokens_seen": 646053888 + }, + { + "epoch": 0.2, + "learning_rate": 0.0006498696088264794, + "loss": 1.4081, + "theoretical_loss": 3.8390073109093947, + "tokens_seen": 646119424 + }, + { + "epoch": 0.2, + "learning_rate": 0.0006498535606820461, + "loss": 1.5169, + "theoretical_loss": 3.8389675062333923, + "tokens_seen": 646184960 + }, + { + "epoch": 0.2, + "learning_rate": 0.0006498375125376128, + "loss": 1.5043, + "theoretical_loss": 3.838927706724395, + "tokens_seen": 646250496 + }, + { + "epoch": 0.2, + "learning_rate": 0.0006498214643931796, + "loss": 1.4406, + "theoretical_loss": 3.8388879123812094, + "tokens_seen": 646316032 + }, + { + "epoch": 0.2, + "learning_rate": 0.0006498054162487462, + "loss": 1.4893, + "theoretical_loss": 3.8388481232026406, + "tokens_seen": 646381568 + }, + { + "epoch": 0.2, + "learning_rate": 0.000649789368104313, + "loss": 1.4568, + "theoretical_loss": 3.838808339187495, + "tokens_seen": 646447104 + }, + { + "epoch": 0.2, + "learning_rate": 0.0006497733199598796, + "loss": 1.5054, + "theoretical_loss": 3.8387685603345805, + "tokens_seen": 646512640 + }, + { + "epoch": 0.2, + "learning_rate": 0.0006497572718154464, + "loss": 1.539, + "theoretical_loss": 3.838728786642702, + "tokens_seen": 646578176 + }, + { + "epoch": 0.2, + "learning_rate": 0.000649741223671013, + "loss": 1.6068, + "theoretical_loss": 3.838689018110667, + "tokens_seen": 646643712 + }, + { + "epoch": 0.2, + "learning_rate": 0.0006497251755265798, + "loss": 1.4617, + "theoretical_loss": 3.838649254737285, + "tokens_seen": 646709248 + }, + { + "epoch": 0.2, + "learning_rate": 0.0006497091273821464, + "loss": 1.6042, + "theoretical_loss": 3.838609496521363, + "tokens_seen": 646774784 + }, + { + "epoch": 0.2, + "learning_rate": 0.0006496930792377132, + "loss": 1.6786, + "theoretical_loss": 3.8385697434617096, + "tokens_seen": 646840320 + }, + { + "epoch": 0.2, + "learning_rate": 0.0006496770310932798, + "loss": 1.526, + "theoretical_loss": 3.838529995557134, + "tokens_seen": 646905856 + }, + { + "epoch": 0.2, + "learning_rate": 0.0006496609829488466, + "loss": 1.404, + "theoretical_loss": 3.8384902528064457, + "tokens_seen": 646971392 + }, + { + "epoch": 0.2, + "learning_rate": 0.0006496449348044132, + "loss": 1.5145, + "theoretical_loss": 3.8384505152084545, + "tokens_seen": 647036928 + }, + { + "epoch": 0.2, + "learning_rate": 0.00064962888665998, + "loss": 1.582, + "theoretical_loss": 3.8384107827619696, + "tokens_seen": 647102464 + }, + { + "epoch": 0.2, + "learning_rate": 0.0006496128385155466, + "loss": 1.5536, + "theoretical_loss": 3.8383710554658026, + "tokens_seen": 647168000 + }, + { + "epoch": 0.2, + "learning_rate": 0.0006495967903711134, + "loss": 1.4962, + "theoretical_loss": 3.838331333318764, + "tokens_seen": 647233536 + }, + { + "epoch": 0.2, + "learning_rate": 0.00064958074222668, + "loss": 1.4811, + "theoretical_loss": 3.8382916163196654, + "tokens_seen": 647299072 + }, + { + "epoch": 0.2, + "learning_rate": 0.0006495646940822468, + "loss": 1.4843, + "theoretical_loss": 3.8382519044673176, + "tokens_seen": 647364608 + }, + { + "epoch": 0.2, + "learning_rate": 0.0006495486459378134, + "loss": 1.516, + "theoretical_loss": 3.8382121977605346, + "tokens_seen": 647430144 + }, + { + "epoch": 0.2, + "learning_rate": 0.0006495325977933802, + "loss": 1.5732, + "theoretical_loss": 3.8381724961981267, + "tokens_seen": 647495680 + }, + { + "epoch": 0.2, + "learning_rate": 0.0006495165496489468, + "loss": 1.5725, + "theoretical_loss": 3.8381327997789088, + "tokens_seen": 647561216 + }, + { + "epoch": 0.2, + "learning_rate": 0.0006495005015045136, + "loss": 1.6248, + "theoretical_loss": 3.8380931085016927, + "tokens_seen": 647626752 + }, + { + "epoch": 0.2, + "learning_rate": 0.0006494844533600802, + "loss": 1.564, + "theoretical_loss": 3.8380534223652933, + "tokens_seen": 647692288 + }, + { + "epoch": 0.2, + "learning_rate": 0.000649468405215647, + "loss": 1.6001, + "theoretical_loss": 3.838013741368523, + "tokens_seen": 647757824 + }, + { + "epoch": 0.2, + "learning_rate": 0.0006494523570712136, + "loss": 1.5464, + "theoretical_loss": 3.837974065510198, + "tokens_seen": 647823360 + }, + { + "epoch": 0.2, + "learning_rate": 0.0006494363089267804, + "loss": 1.4915, + "theoretical_loss": 3.837934394789133, + "tokens_seen": 647888896 + }, + { + "epoch": 0.2, + "learning_rate": 0.000649420260782347, + "loss": 1.484, + "theoretical_loss": 3.8378947292041428, + "tokens_seen": 647954432 + }, + { + "epoch": 0.2, + "learning_rate": 0.0006494042126379138, + "loss": 1.4874, + "theoretical_loss": 3.8378550687540427, + "tokens_seen": 648019968 + }, + { + "epoch": 0.2, + "learning_rate": 0.0006493881644934805, + "loss": 1.5269, + "theoretical_loss": 3.8378154134376494, + "tokens_seen": 648085504 + }, + { + "epoch": 0.2, + "learning_rate": 0.0006493721163490472, + "loss": 1.6225, + "theoretical_loss": 3.837775763253779, + "tokens_seen": 648151040 + }, + { + "epoch": 0.2, + "learning_rate": 0.0006493560682046139, + "loss": 1.6051, + "theoretical_loss": 3.8377361182012484, + "tokens_seen": 648216576 + }, + { + "epoch": 0.2, + "learning_rate": 0.0006493400200601806, + "loss": 1.5941, + "theoretical_loss": 3.8376964782788754, + "tokens_seen": 648282112 + }, + { + "epoch": 0.2, + "learning_rate": 0.0006493239719157473, + "loss": 1.5141, + "theoretical_loss": 3.8376568434854765, + "tokens_seen": 648347648 + }, + { + "epoch": 0.2, + "learning_rate": 0.000649307923771314, + "loss": 1.5469, + "theoretical_loss": 3.837617213819871, + "tokens_seen": 648413184 + }, + { + "epoch": 0.2, + "learning_rate": 0.0006492918756268807, + "loss": 1.4856, + "theoretical_loss": 3.837577589280876, + "tokens_seen": 648478720 + }, + { + "epoch": 0.2, + "learning_rate": 0.0006492758274824474, + "loss": 1.5407, + "theoretical_loss": 3.8375379698673115, + "tokens_seen": 648544256 + }, + { + "epoch": 0.2, + "learning_rate": 0.0006492597793380141, + "loss": 1.4765, + "theoretical_loss": 3.837498355577996, + "tokens_seen": 648609792 + }, + { + "epoch": 0.2, + "learning_rate": 0.0006492437311935808, + "loss": 1.5979, + "theoretical_loss": 3.837458746411749, + "tokens_seen": 648675328 + }, + { + "epoch": 0.2, + "learning_rate": 0.0006492276830491475, + "loss": 1.4858, + "theoretical_loss": 3.8374191423673913, + "tokens_seen": 648740864 + }, + { + "epoch": 0.2, + "objective/train/docs_used": 241544, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.4680938720703125, + "objective/train/theoretical_loss": 3.8373795434437423, + "objective/train/tokens_used": 669266400, + "theoretical_loss": 3.8373795434437423, + "tokens_seen": 648806400 + }, + { + "epoch": 0.2, + "learning_rate": 0.0006492116349047142, + "loss": 1.5375, + "theoretical_loss": 3.8373795434437423, + "tokens_seen": 648806400 + }, + { + "epoch": 0.2, + "learning_rate": 0.0006491955867602809, + "loss": 1.4837, + "theoretical_loss": 3.8373399496396234, + "tokens_seen": 648871936 + }, + { + "epoch": 0.2, + "learning_rate": 0.0006491795386158476, + "loss": 1.5573, + "theoretical_loss": 3.8373003609538556, + "tokens_seen": 648937472 + }, + { + "epoch": 0.2, + "learning_rate": 0.0006491634904714143, + "loss": 1.5696, + "theoretical_loss": 3.8372607773852603, + "tokens_seen": 649003008 + }, + { + "epoch": 0.2, + "learning_rate": 0.000649147442326981, + "loss": 1.6125, + "theoretical_loss": 3.8372211989326592, + "tokens_seen": 649068544 + }, + { + "epoch": 0.2, + "learning_rate": 0.0006491313941825477, + "loss": 1.5622, + "theoretical_loss": 3.8371816255948756, + "tokens_seen": 649134080 + }, + { + "epoch": 0.2, + "learning_rate": 0.0006491153460381144, + "loss": 1.4636, + "theoretical_loss": 3.837142057370731, + "tokens_seen": 649199616 + }, + { + "epoch": 0.2, + "learning_rate": 0.0006490992978936811, + "loss": 1.5665, + "theoretical_loss": 3.837102494259049, + "tokens_seen": 649265152 + }, + { + "epoch": 0.2, + "learning_rate": 0.0006490832497492478, + "loss": 1.5036, + "theoretical_loss": 3.8370629362586532, + "tokens_seen": 649330688 + }, + { + "epoch": 0.2, + "learning_rate": 0.0006490672016048145, + "loss": 1.6063, + "theoretical_loss": 3.8370233833683676, + "tokens_seen": 649396224 + }, + { + "epoch": 0.2, + "learning_rate": 0.0006490511534603812, + "loss": 1.691, + "theoretical_loss": 3.8369838355870165, + "tokens_seen": 649461760 + }, + { + "epoch": 0.2, + "learning_rate": 0.0006490351053159479, + "loss": 1.5198, + "theoretical_loss": 3.836944292913424, + "tokens_seen": 649527296 + }, + { + "epoch": 0.2, + "learning_rate": 0.0006490190571715146, + "loss": 1.4586, + "theoretical_loss": 3.8369047553464153, + "tokens_seen": 649592832 + }, + { + "epoch": 0.2, + "learning_rate": 0.0006490030090270813, + "loss": 1.4843, + "theoretical_loss": 3.836865222884817, + "tokens_seen": 649658368 + }, + { + "epoch": 0.2, + "learning_rate": 0.000648986960882648, + "loss": 1.602, + "theoretical_loss": 3.8368256955274527, + "tokens_seen": 649723904 + }, + { + "epoch": 0.2, + "learning_rate": 0.0006489709127382147, + "loss": 1.5245, + "theoretical_loss": 3.8367861732731505, + "tokens_seen": 649789440 + }, + { + "epoch": 0.2, + "learning_rate": 0.0006489548645937814, + "loss": 1.4378, + "theoretical_loss": 3.836746656120737, + "tokens_seen": 649854976 + }, + { + "epoch": 0.2, + "learning_rate": 0.0006489388164493481, + "loss": 1.5446, + "theoretical_loss": 3.8367071440690372, + "tokens_seen": 649920512 + }, + { + "epoch": 0.2, + "learning_rate": 0.0006489227683049148, + "loss": 1.5242, + "theoretical_loss": 3.836667637116881, + "tokens_seen": 649986048 + }, + { + "epoch": 0.2, + "learning_rate": 0.0006489067201604815, + "loss": 1.4662, + "theoretical_loss": 3.8366281352630947, + "tokens_seen": 650051584 + }, + { + "epoch": 0.2, + "learning_rate": 0.0006488906720160482, + "loss": 1.4631, + "theoretical_loss": 3.8365886385065067, + "tokens_seen": 650117120 + }, + { + "epoch": 0.2, + "learning_rate": 0.0006488746238716149, + "loss": 1.4556, + "theoretical_loss": 3.836549146845946, + "tokens_seen": 650182656 + }, + { + "epoch": 0.2, + "learning_rate": 0.0006488585757271816, + "loss": 1.4752, + "theoretical_loss": 3.8365096602802407, + "tokens_seen": 650248192 + }, + { + "epoch": 0.2, + "learning_rate": 0.0006488425275827483, + "loss": 1.652, + "theoretical_loss": 3.83647017880822, + "tokens_seen": 650313728 + }, + { + "epoch": 0.2, + "learning_rate": 0.000648826479438315, + "loss": 1.4813, + "theoretical_loss": 3.836430702428715, + "tokens_seen": 650379264 + }, + { + "epoch": 0.2, + "learning_rate": 0.0006488104312938817, + "loss": 1.5376, + "theoretical_loss": 3.8363912311405546, + "tokens_seen": 650444800 + }, + { + "epoch": 0.2, + "learning_rate": 0.0006487943831494484, + "loss": 1.5241, + "theoretical_loss": 3.8363517649425702, + "tokens_seen": 650510336 + }, + { + "epoch": 0.2, + "learning_rate": 0.0006487783350050151, + "loss": 1.5454, + "theoretical_loss": 3.8363123038335916, + "tokens_seen": 650575872 + }, + { + "epoch": 0.2, + "learning_rate": 0.0006487622868605818, + "loss": 1.5059, + "theoretical_loss": 3.8362728478124506, + "tokens_seen": 650641408 + }, + { + "epoch": 0.2, + "learning_rate": 0.0006487462387161485, + "loss": 1.563, + "theoretical_loss": 3.8362333968779794, + "tokens_seen": 650706944 + }, + { + "epoch": 0.2, + "learning_rate": 0.0006487301905717152, + "loss": 1.5176, + "theoretical_loss": 3.836193951029008, + "tokens_seen": 650772480 + }, + { + "epoch": 0.2, + "learning_rate": 0.0006487141424272819, + "loss": 1.5936, + "theoretical_loss": 3.8361545102643717, + "tokens_seen": 650838016 + }, + { + "epoch": 0.2, + "learning_rate": 0.0006486980942828485, + "loss": 1.5161, + "theoretical_loss": 3.836115074582901, + "tokens_seen": 650903552 + }, + { + "epoch": 0.2, + "learning_rate": 0.0006486820461384153, + "loss": 1.4945, + "theoretical_loss": 3.8360756439834303, + "tokens_seen": 650969088 + }, + { + "epoch": 0.2, + "learning_rate": 0.0006486659979939819, + "loss": 1.4281, + "theoretical_loss": 3.836036218464792, + "tokens_seen": 651034624 + }, + { + "epoch": 0.2, + "learning_rate": 0.0006486499498495487, + "loss": 1.4151, + "theoretical_loss": 3.8359967980258216, + "tokens_seen": 651100160 + }, + { + "epoch": 0.2, + "learning_rate": 0.0006486339017051153, + "loss": 1.4625, + "theoretical_loss": 3.8359573826653524, + "tokens_seen": 651165696 + }, + { + "epoch": 0.2, + "learning_rate": 0.0006486178535606821, + "loss": 1.5747, + "theoretical_loss": 3.8359179723822194, + "tokens_seen": 651231232 + }, + { + "epoch": 0.2, + "learning_rate": 0.0006486018054162487, + "loss": 1.6319, + "theoretical_loss": 3.835878567175257, + "tokens_seen": 651296768 + }, + { + "epoch": 0.2, + "learning_rate": 0.0006485857572718155, + "loss": 1.5384, + "theoretical_loss": 3.8358391670433023, + "tokens_seen": 651362304 + }, + { + "epoch": 0.2, + "learning_rate": 0.0006485697091273821, + "loss": 1.5515, + "theoretical_loss": 3.8357997719851893, + "tokens_seen": 651427840 + }, + { + "epoch": 0.2, + "learning_rate": 0.0006485536609829489, + "loss": 1.5528, + "theoretical_loss": 3.835760381999756, + "tokens_seen": 651493376 + }, + { + "epoch": 0.2, + "learning_rate": 0.0006485376128385155, + "loss": 1.6193, + "theoretical_loss": 3.835720997085838, + "tokens_seen": 651558912 + }, + { + "epoch": 0.2, + "learning_rate": 0.0006485215646940823, + "loss": 1.5771, + "theoretical_loss": 3.835681617242272, + "tokens_seen": 651624448 + }, + { + "epoch": 0.2, + "learning_rate": 0.0006485055165496489, + "loss": 1.5193, + "theoretical_loss": 3.8356422424678964, + "tokens_seen": 651689984 + }, + { + "epoch": 0.2, + "learning_rate": 0.0006484894684052157, + "loss": 1.5806, + "theoretical_loss": 3.835602872761549, + "tokens_seen": 651755520 + }, + { + "epoch": 0.2, + "learning_rate": 0.0006484734202607823, + "loss": 1.5636, + "theoretical_loss": 3.8355635081220667, + "tokens_seen": 651821056 + }, + { + "epoch": 0.2, + "learning_rate": 0.0006484573721163491, + "loss": 1.5584, + "theoretical_loss": 3.8355241485482887, + "tokens_seen": 651886592 + }, + { + "epoch": 0.2, + "learning_rate": 0.0006484413239719157, + "loss": 1.5104, + "theoretical_loss": 3.835484794039055, + "tokens_seen": 651952128 + }, + { + "epoch": 0.2, + "learning_rate": 0.0006484252758274825, + "loss": 1.5388, + "theoretical_loss": 3.835445444593203, + "tokens_seen": 652017664 + }, + { + "epoch": 0.2, + "objective/train/docs_used": 242832, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.5449106693267822, + "objective/train/theoretical_loss": 3.8354061002095743, + "objective/train/tokens_used": 672543200, + "theoretical_loss": 3.8354061002095743, + "tokens_seen": 652083200 + }, + { + "epoch": 0.2, + "learning_rate": 0.0006484092276830491, + "loss": 1.4909, + "theoretical_loss": 3.8354061002095743, + "tokens_seen": 652083200 + }, + { + "epoch": 0.2, + "learning_rate": 0.0006483931795386159, + "loss": 1.5756, + "theoretical_loss": 3.835366760887008, + "tokens_seen": 652148736 + }, + { + "epoch": 0.2, + "learning_rate": 0.0006483771313941825, + "loss": 1.4767, + "theoretical_loss": 3.8353274266243442, + "tokens_seen": 652214272 + }, + { + "epoch": 0.2, + "learning_rate": 0.0006483610832497493, + "loss": 1.5212, + "theoretical_loss": 3.835288097420425, + "tokens_seen": 652279808 + }, + { + "epoch": 0.2, + "learning_rate": 0.0006483450351053159, + "loss": 1.63, + "theoretical_loss": 3.83524877327409, + "tokens_seen": 652345344 + }, + { + "epoch": 0.2, + "learning_rate": 0.0006483289869608827, + "loss": 1.5818, + "theoretical_loss": 3.8352094541841817, + "tokens_seen": 652410880 + }, + { + "epoch": 0.2, + "learning_rate": 0.0006483129388164493, + "loss": 1.4821, + "theoretical_loss": 3.8351701401495424, + "tokens_seen": 652476416 + }, + { + "epoch": 0.2, + "learning_rate": 0.0006482968906720161, + "loss": 1.476, + "theoretical_loss": 3.8351308311690144, + "tokens_seen": 652541952 + }, + { + "epoch": 0.2, + "learning_rate": 0.0006482808425275827, + "loss": 1.5582, + "theoretical_loss": 3.83509152724144, + "tokens_seen": 652607488 + }, + { + "epoch": 0.2, + "learning_rate": 0.0006482647943831495, + "loss": 1.6408, + "theoretical_loss": 3.8350522283656625, + "tokens_seen": 652673024 + }, + { + "epoch": 0.2, + "learning_rate": 0.0006482487462387161, + "loss": 1.5531, + "theoretical_loss": 3.8350129345405257, + "tokens_seen": 652738560 + }, + { + "epoch": 0.2, + "learning_rate": 0.0006482326980942829, + "loss": 1.5177, + "theoretical_loss": 3.8349736457648733, + "tokens_seen": 652804096 + }, + { + "epoch": 0.2, + "learning_rate": 0.0006482166499498495, + "loss": 1.4834, + "theoretical_loss": 3.83493436203755, + "tokens_seen": 652869632 + }, + { + "epoch": 0.2, + "learning_rate": 0.0006482006018054163, + "loss": 1.5401, + "theoretical_loss": 3.834895083357399, + "tokens_seen": 652935168 + }, + { + "epoch": 0.2, + "learning_rate": 0.0006481845536609829, + "loss": 1.5321, + "theoretical_loss": 3.8348558097232672, + "tokens_seen": 653000704 + }, + { + "epoch": 0.2, + "learning_rate": 0.0006481685055165497, + "loss": 1.5911, + "theoretical_loss": 3.834816541133999, + "tokens_seen": 653066240 + }, + { + "epoch": 0.2, + "learning_rate": 0.0006481524573721163, + "loss": 1.438, + "theoretical_loss": 3.83477727758844, + "tokens_seen": 653131776 + }, + { + "epoch": 0.2, + "learning_rate": 0.0006481364092276831, + "loss": 1.6793, + "theoretical_loss": 3.8347380190854374, + "tokens_seen": 653197312 + }, + { + "epoch": 0.2, + "learning_rate": 0.0006481203610832497, + "loss": 1.5255, + "theoretical_loss": 3.8346987656238367, + "tokens_seen": 653262848 + }, + { + "epoch": 0.2, + "learning_rate": 0.0006481043129388165, + "loss": 1.4954, + "theoretical_loss": 3.8346595172024855, + "tokens_seen": 653328384 + }, + { + "epoch": 0.2, + "learning_rate": 0.0006480882647943831, + "loss": 1.4542, + "theoretical_loss": 3.834620273820231, + "tokens_seen": 653393920 + }, + { + "epoch": 0.2, + "learning_rate": 0.0006480722166499499, + "loss": 1.6051, + "theoretical_loss": 3.834581035475921, + "tokens_seen": 653459456 + }, + { + "epoch": 0.2, + "learning_rate": 0.0006480561685055166, + "loss": 1.5269, + "theoretical_loss": 3.8345418021684035, + "tokens_seen": 653524992 + }, + { + "epoch": 0.2, + "learning_rate": 0.0006480401203610833, + "loss": 1.5427, + "theoretical_loss": 3.8345025738965264, + "tokens_seen": 653590528 + }, + { + "epoch": 0.2, + "learning_rate": 0.00064802407221665, + "loss": 1.5295, + "theoretical_loss": 3.8344633506591395, + "tokens_seen": 653656064 + }, + { + "epoch": 0.2, + "learning_rate": 0.0006480080240722167, + "loss": 1.5792, + "theoretical_loss": 3.8344241324550907, + "tokens_seen": 653721600 + }, + { + "epoch": 0.2, + "learning_rate": 0.0006479919759277834, + "loss": 1.568, + "theoretical_loss": 3.8343849192832313, + "tokens_seen": 653787136 + }, + { + "epoch": 0.2, + "learning_rate": 0.0006479759277833501, + "loss": 1.5781, + "theoretical_loss": 3.83434571114241, + "tokens_seen": 653852672 + }, + { + "epoch": 0.2, + "learning_rate": 0.0006479598796389168, + "loss": 1.4493, + "theoretical_loss": 3.834306508031478, + "tokens_seen": 653918208 + }, + { + "epoch": 0.2, + "learning_rate": 0.0006479438314944835, + "loss": 1.6275, + "theoretical_loss": 3.8342673099492854, + "tokens_seen": 653983744 + }, + { + "epoch": 0.2, + "learning_rate": 0.0006479277833500502, + "loss": 1.5668, + "theoretical_loss": 3.8342281168946832, + "tokens_seen": 654049280 + }, + { + "epoch": 0.2, + "learning_rate": 0.0006479117352056169, + "loss": 1.576, + "theoretical_loss": 3.8341889288665234, + "tokens_seen": 654114816 + }, + { + "epoch": 0.2, + "learning_rate": 0.0006478956870611836, + "loss": 1.5054, + "theoretical_loss": 3.8341497458636575, + "tokens_seen": 654180352 + }, + { + "epoch": 0.2, + "learning_rate": 0.0006478796389167503, + "loss": 1.5108, + "theoretical_loss": 3.834110567884938, + "tokens_seen": 654245888 + }, + { + "epoch": 0.2, + "learning_rate": 0.000647863590772317, + "loss": 1.6155, + "theoretical_loss": 3.8340713949292176, + "tokens_seen": 654311424 + }, + { + "epoch": 0.2, + "learning_rate": 0.0006478475426278837, + "loss": 1.5045, + "theoretical_loss": 3.834032226995349, + "tokens_seen": 654376960 + }, + { + "epoch": 0.2, + "learning_rate": 0.0006478314944834504, + "loss": 1.5233, + "theoretical_loss": 3.8339930640821853, + "tokens_seen": 654442496 + }, + { + "epoch": 0.2, + "learning_rate": 0.0006478154463390171, + "loss": 1.4597, + "theoretical_loss": 3.8339539061885812, + "tokens_seen": 654508032 + }, + { + "epoch": 0.2, + "learning_rate": 0.0006477993981945838, + "loss": 1.4635, + "theoretical_loss": 3.83391475331339, + "tokens_seen": 654573568 + }, + { + "epoch": 0.2, + "learning_rate": 0.0006477833500501505, + "loss": 1.5185, + "theoretical_loss": 3.833875605455466, + "tokens_seen": 654639104 + }, + { + "epoch": 0.2, + "learning_rate": 0.0006477673019057172, + "loss": 1.5293, + "theoretical_loss": 3.8338364626136654, + "tokens_seen": 654704640 + }, + { + "epoch": 0.2, + "learning_rate": 0.0006477512537612839, + "loss": 1.4165, + "theoretical_loss": 3.833797324786842, + "tokens_seen": 654770176 + }, + { + "epoch": 0.2, + "learning_rate": 0.0006477352056168506, + "loss": 1.5568, + "theoretical_loss": 3.8337581919738524, + "tokens_seen": 654835712 + }, + { + "epoch": 0.2, + "learning_rate": 0.0006477191574724173, + "loss": 1.6123, + "theoretical_loss": 3.8337190641735512, + "tokens_seen": 654901248 + }, + { + "epoch": 0.2, + "learning_rate": 0.000647703109327984, + "loss": 1.558, + "theoretical_loss": 3.833679941384797, + "tokens_seen": 654966784 + }, + { + "epoch": 0.2, + "learning_rate": 0.0006476870611835507, + "loss": 1.3939, + "theoretical_loss": 3.833640823606445, + "tokens_seen": 655032320 + }, + { + "epoch": 0.2, + "learning_rate": 0.0006476710130391174, + "loss": 1.5128, + "theoretical_loss": 3.8336017108373523, + "tokens_seen": 655097856 + }, + { + "epoch": 0.2, + "learning_rate": 0.0006476549648946841, + "loss": 1.6233, + "theoretical_loss": 3.833562603076377, + "tokens_seen": 655163392 + }, + { + "epoch": 0.2, + "learning_rate": 0.0006476389167502508, + "loss": 1.5987, + "theoretical_loss": 3.833523500322377, + "tokens_seen": 655228928 + }, + { + "epoch": 0.2, + "learning_rate": 0.0006476228686058175, + "loss": 1.4782, + "theoretical_loss": 3.83348440257421, + "tokens_seen": 655294464 + }, + { + "debugging/Compilability": 0.9230769230769231, + "debugging/distinct-1-grams": 0.7560193640250044, + "debugging/entropy-1-grams": 5.2534808493234895, + "debugging/length": 422.2307692307692, + "debugging/num_segments": 13, + "debugging/score": 0.01638830111902192, + "debugging/score_std": 0.019821935735086043, + "epoch": 0.2, + "objective/train/docs_used": 243976, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.469989538192749, + "objective/train/theoretical_loss": 3.833445309830735, + "objective/train/tokens_used": 675820000, + "theoretical_loss": 3.833445309830735, + "tokens_seen": 655360000 + }, + { + "epoch": 0.2, + "learning_rate": 0.0006476068204613842, + "loss": 1.58, + "theoretical_loss": 3.833445309830735, + "tokens_seen": 655360000 + }, + { + "epoch": 0.2, + "learning_rate": 0.0006475907723169509, + "loss": 1.5522, + "theoretical_loss": 3.833406222090811, + "tokens_seen": 655425536 + }, + { + "epoch": 0.2, + "learning_rate": 0.0006475747241725176, + "loss": 1.4536, + "theoretical_loss": 3.8333671393532973, + "tokens_seen": 655491072 + }, + { + "epoch": 0.2, + "learning_rate": 0.0006475586760280843, + "loss": 1.5522, + "theoretical_loss": 3.8333280616170535, + "tokens_seen": 655556608 + }, + { + "epoch": 0.2, + "learning_rate": 0.000647542627883651, + "loss": 1.5255, + "theoretical_loss": 3.8332889888809403, + "tokens_seen": 655622144 + }, + { + "epoch": 0.2, + "learning_rate": 0.0006475265797392177, + "loss": 1.5044, + "theoretical_loss": 3.8332499211438176, + "tokens_seen": 655687680 + }, + { + "epoch": 0.2, + "learning_rate": 0.0006475105315947844, + "loss": 1.5449, + "theoretical_loss": 3.8332108584045463, + "tokens_seen": 655753216 + }, + { + "epoch": 0.2, + "learning_rate": 0.0006474944834503511, + "loss": 1.5881, + "theoretical_loss": 3.8331718006619884, + "tokens_seen": 655818752 + }, + { + "epoch": 0.2, + "learning_rate": 0.0006474784353059178, + "loss": 1.5166, + "theoretical_loss": 3.833132747915004, + "tokens_seen": 655884288 + }, + { + "epoch": 0.2, + "learning_rate": 0.0006474623871614845, + "loss": 1.4652, + "theoretical_loss": 3.8330937001624568, + "tokens_seen": 655949824 + }, + { + "epoch": 0.2, + "learning_rate": 0.0006474463390170512, + "loss": 1.3869, + "theoretical_loss": 3.833054657403208, + "tokens_seen": 656015360 + }, + { + "epoch": 0.2, + "learning_rate": 0.0006474302908726179, + "loss": 1.5893, + "theoretical_loss": 3.8330156196361207, + "tokens_seen": 656080896 + }, + { + "epoch": 0.2, + "learning_rate": 0.0006474142427281846, + "loss": 1.6088, + "theoretical_loss": 3.8329765868600583, + "tokens_seen": 656146432 + }, + { + "epoch": 0.2, + "learning_rate": 0.0006473981945837513, + "loss": 1.5619, + "theoretical_loss": 3.832937559073884, + "tokens_seen": 656211968 + }, + { + "epoch": 0.2, + "learning_rate": 0.000647382146439318, + "loss": 1.5477, + "theoretical_loss": 3.8328985362764616, + "tokens_seen": 656277504 + }, + { + "epoch": 0.2, + "learning_rate": 0.0006473660982948847, + "loss": 1.7084, + "theoretical_loss": 3.8328595184666554, + "tokens_seen": 656343040 + }, + { + "epoch": 0.2, + "learning_rate": 0.0006473500501504514, + "loss": 1.5477, + "theoretical_loss": 3.8328205056433298, + "tokens_seen": 656408576 + }, + { + "epoch": 0.2, + "learning_rate": 0.0006473340020060181, + "loss": 1.4643, + "theoretical_loss": 3.8327814978053505, + "tokens_seen": 656474112 + }, + { + "epoch": 0.2, + "learning_rate": 0.0006473179538615848, + "loss": 1.5881, + "theoretical_loss": 3.8327424949515816, + "tokens_seen": 656539648 + }, + { + "epoch": 0.2, + "learning_rate": 0.0006473019057171514, + "loss": 1.6402, + "theoretical_loss": 3.83270349708089, + "tokens_seen": 656605184 + }, + { + "epoch": 0.2, + "learning_rate": 0.0006472858575727182, + "loss": 1.5406, + "theoretical_loss": 3.832664504192141, + "tokens_seen": 656670720 + }, + { + "epoch": 0.2, + "learning_rate": 0.0006472698094282848, + "loss": 1.5329, + "theoretical_loss": 3.8326255162842022, + "tokens_seen": 656736256 + }, + { + "epoch": 0.2, + "learning_rate": 0.0006472537612838516, + "loss": 1.4653, + "theoretical_loss": 3.832586533355939, + "tokens_seen": 656801792 + }, + { + "epoch": 0.2, + "learning_rate": 0.0006472377131394182, + "loss": 1.5117, + "theoretical_loss": 3.832547555406219, + "tokens_seen": 656867328 + }, + { + "epoch": 0.2, + "learning_rate": 0.000647221664994985, + "loss": 1.4648, + "theoretical_loss": 3.8325085824339102, + "tokens_seen": 656932864 + }, + { + "epoch": 0.2, + "learning_rate": 0.0006472056168505516, + "loss": 1.51, + "theoretical_loss": 3.8324696144378803, + "tokens_seen": 656998400 + }, + { + "epoch": 0.2, + "learning_rate": 0.0006471895687061184, + "loss": 1.4025, + "theoretical_loss": 3.8324306514169977, + "tokens_seen": 657063936 + }, + { + "epoch": 0.2, + "learning_rate": 0.000647173520561685, + "loss": 1.5189, + "theoretical_loss": 3.832391693370131, + "tokens_seen": 657129472 + }, + { + "epoch": 0.2, + "learning_rate": 0.0006471574724172518, + "loss": 1.6047, + "theoretical_loss": 3.832352740296149, + "tokens_seen": 657195008 + }, + { + "epoch": 0.2, + "learning_rate": 0.0006471414242728184, + "loss": 1.5526, + "theoretical_loss": 3.8323137921939217, + "tokens_seen": 657260544 + }, + { + "epoch": 0.2, + "learning_rate": 0.0006471253761283852, + "loss": 1.4174, + "theoretical_loss": 3.832274849062318, + "tokens_seen": 657326080 + }, + { + "epoch": 0.2, + "learning_rate": 0.0006471093279839518, + "loss": 1.495, + "theoretical_loss": 3.8322359109002093, + "tokens_seen": 657391616 + }, + { + "epoch": 0.2, + "learning_rate": 0.0006470932798395186, + "loss": 1.4776, + "theoretical_loss": 3.8321969777064653, + "tokens_seen": 657457152 + }, + { + "epoch": 0.2, + "learning_rate": 0.0006470772316950852, + "loss": 1.4966, + "theoretical_loss": 3.8321580494799568, + "tokens_seen": 657522688 + }, + { + "epoch": 0.2, + "learning_rate": 0.000647061183550652, + "loss": 1.6411, + "theoretical_loss": 3.8321191262195553, + "tokens_seen": 657588224 + }, + { + "epoch": 0.2, + "learning_rate": 0.0006470451354062186, + "loss": 1.5678, + "theoretical_loss": 3.832080207924133, + "tokens_seen": 657653760 + }, + { + "epoch": 0.2, + "learning_rate": 0.0006470290872617854, + "loss": 1.5019, + "theoretical_loss": 3.832041294592561, + "tokens_seen": 657719296 + }, + { + "epoch": 0.2, + "learning_rate": 0.000647013039117352, + "loss": 1.4981, + "theoretical_loss": 3.832002386223712, + "tokens_seen": 657784832 + }, + { + "epoch": 0.2, + "learning_rate": 0.0006469969909729188, + "loss": 1.4475, + "theoretical_loss": 3.831963482816459, + "tokens_seen": 657850368 + }, + { + "epoch": 0.2, + "learning_rate": 0.0006469809428284854, + "loss": 1.5385, + "theoretical_loss": 3.8319245843696743, + "tokens_seen": 657915904 + }, + { + "epoch": 0.2, + "learning_rate": 0.0006469648946840522, + "loss": 1.578, + "theoretical_loss": 3.8318856908822325, + "tokens_seen": 657981440 + }, + { + "epoch": 0.2, + "learning_rate": 0.0006469488465396188, + "loss": 1.5025, + "theoretical_loss": 3.831846802353007, + "tokens_seen": 658046976 + }, + { + "epoch": 0.2, + "learning_rate": 0.0006469327983951856, + "loss": 1.5892, + "theoretical_loss": 3.8318079187808713, + "tokens_seen": 658112512 + }, + { + "epoch": 0.2, + "learning_rate": 0.0006469167502507522, + "loss": 1.575, + "theoretical_loss": 3.831769040164701, + "tokens_seen": 658178048 + }, + { + "epoch": 0.2, + "learning_rate": 0.000646900702106319, + "loss": 1.4804, + "theoretical_loss": 3.8317301665033705, + "tokens_seen": 658243584 + }, + { + "epoch": 0.2, + "learning_rate": 0.0006468846539618856, + "loss": 1.579, + "theoretical_loss": 3.8316912977957553, + "tokens_seen": 658309120 + }, + { + "epoch": 0.2, + "learning_rate": 0.0006468686058174524, + "loss": 1.509, + "theoretical_loss": 3.8316524340407314, + "tokens_seen": 658374656 + }, + { + "epoch": 0.2, + "learning_rate": 0.000646852557673019, + "loss": 1.5187, + "theoretical_loss": 3.8316135752371743, + "tokens_seen": 658440192 + }, + { + "epoch": 0.2, + "learning_rate": 0.0006468365095285858, + "loss": 1.5195, + "theoretical_loss": 3.8315747213839604, + "tokens_seen": 658505728 + }, + { + "epoch": 0.2, + "learning_rate": 0.0006468204613841524, + "loss": 1.4725, + "theoretical_loss": 3.831535872479967, + "tokens_seen": 658571264 + }, + { + "epoch": 0.2, + "objective/train/docs_used": 245288, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.4209370613098145, + "objective/train/theoretical_loss": 3.831497028524071, + "objective/train/tokens_used": 679096800, + "theoretical_loss": 3.831497028524071, + "tokens_seen": 658636800 + }, + { + "epoch": 0.2, + "learning_rate": 0.0006468044132397192, + "loss": 1.4964, + "theoretical_loss": 3.831497028524071, + "tokens_seen": 658636800 + }, + { + "epoch": 0.2, + "learning_rate": 0.0006467883650952858, + "loss": 1.513, + "theoretical_loss": 3.83145818951515, + "tokens_seen": 658702336 + }, + { + "epoch": 0.2, + "learning_rate": 0.0006467723169508526, + "loss": 1.4786, + "theoretical_loss": 3.8314193554520806, + "tokens_seen": 658767872 + }, + { + "epoch": 0.2, + "learning_rate": 0.0006467562688064192, + "loss": 1.5388, + "theoretical_loss": 3.831380526333743, + "tokens_seen": 658833408 + }, + { + "epoch": 0.2, + "learning_rate": 0.000646740220661986, + "loss": 1.5369, + "theoretical_loss": 3.8313417021590155, + "tokens_seen": 658898944 + }, + { + "epoch": 0.2, + "learning_rate": 0.0006467241725175526, + "loss": 1.5638, + "theoretical_loss": 3.831302882926776, + "tokens_seen": 658964480 + }, + { + "epoch": 0.2, + "learning_rate": 0.0006467081243731194, + "loss": 1.4851, + "theoretical_loss": 3.831264068635905, + "tokens_seen": 659030016 + }, + { + "epoch": 0.2, + "learning_rate": 0.000646692076228686, + "loss": 1.519, + "theoretical_loss": 3.831225259285281, + "tokens_seen": 659095552 + }, + { + "epoch": 0.2, + "learning_rate": 0.0006466760280842528, + "loss": 1.5303, + "theoretical_loss": 3.8311864548737855, + "tokens_seen": 659161088 + }, + { + "epoch": 0.2, + "learning_rate": 0.0006466599799398195, + "loss": 1.5614, + "theoretical_loss": 3.831147655400298, + "tokens_seen": 659226624 + }, + { + "epoch": 0.2, + "learning_rate": 0.0006466439317953862, + "loss": 1.5756, + "theoretical_loss": 3.831108860863699, + "tokens_seen": 659292160 + }, + { + "epoch": 0.2, + "learning_rate": 0.0006466278836509529, + "loss": 1.5758, + "theoretical_loss": 3.8310700712628707, + "tokens_seen": 659357696 + }, + { + "epoch": 0.2, + "learning_rate": 0.0006466118355065196, + "loss": 1.5834, + "theoretical_loss": 3.8310312865966942, + "tokens_seen": 659423232 + }, + { + "epoch": 0.2, + "learning_rate": 0.0006465957873620863, + "loss": 1.5413, + "theoretical_loss": 3.830992506864051, + "tokens_seen": 659488768 + }, + { + "epoch": 0.2, + "learning_rate": 0.000646579739217653, + "loss": 1.5781, + "theoretical_loss": 3.830953732063824, + "tokens_seen": 659554304 + }, + { + "epoch": 0.2, + "learning_rate": 0.0006465636910732197, + "loss": 1.5916, + "theoretical_loss": 3.830914962194895, + "tokens_seen": 659619840 + }, + { + "epoch": 0.2, + "learning_rate": 0.0006465476429287864, + "loss": 1.6245, + "theoretical_loss": 3.830876197256148, + "tokens_seen": 659685376 + }, + { + "epoch": 0.2, + "learning_rate": 0.0006465315947843531, + "loss": 1.524, + "theoretical_loss": 3.8308374372464655, + "tokens_seen": 659750912 + }, + { + "epoch": 0.2, + "learning_rate": 0.0006465155466399198, + "loss": 1.5363, + "theoretical_loss": 3.830798682164732, + "tokens_seen": 659816448 + }, + { + "epoch": 0.2, + "learning_rate": 0.0006464994984954865, + "loss": 1.5012, + "theoretical_loss": 3.8307599320098307, + "tokens_seen": 659881984 + }, + { + "epoch": 0.2, + "learning_rate": 0.0006464834503510532, + "loss": 1.5936, + "theoretical_loss": 3.8307211867806465, + "tokens_seen": 659947520 + }, + { + "epoch": 0.2, + "learning_rate": 0.0006464674022066199, + "loss": 1.4331, + "theoretical_loss": 3.8306824464760645, + "tokens_seen": 660013056 + }, + { + "epoch": 0.2, + "learning_rate": 0.0006464513540621866, + "loss": 1.4985, + "theoretical_loss": 3.83064371109497, + "tokens_seen": 660078592 + }, + { + "epoch": 0.2, + "learning_rate": 0.0006464353059177533, + "loss": 1.5553, + "theoretical_loss": 3.830604980636248, + "tokens_seen": 660144128 + }, + { + "epoch": 0.2, + "learning_rate": 0.00064641925777332, + "loss": 1.5687, + "theoretical_loss": 3.8305662550987845, + "tokens_seen": 660209664 + }, + { + "epoch": 0.2, + "learning_rate": 0.0006464032096288867, + "loss": 1.5518, + "theoretical_loss": 3.8305275344814658, + "tokens_seen": 660275200 + }, + { + "epoch": 0.2, + "learning_rate": 0.0006463871614844534, + "loss": 1.4351, + "theoretical_loss": 3.8304888187831785, + "tokens_seen": 660340736 + }, + { + "epoch": 0.2, + "learning_rate": 0.0006463711133400201, + "loss": 1.519, + "theoretical_loss": 3.8304501080028097, + "tokens_seen": 660406272 + }, + { + "epoch": 0.2, + "learning_rate": 0.0006463550651955868, + "loss": 1.5711, + "theoretical_loss": 3.830411402139247, + "tokens_seen": 660471808 + }, + { + "epoch": 0.2, + "learning_rate": 0.0006463390170511535, + "loss": 1.5307, + "theoretical_loss": 3.830372701191377, + "tokens_seen": 660537344 + }, + { + "epoch": 0.2, + "learning_rate": 0.0006463229689067202, + "loss": 1.5103, + "theoretical_loss": 3.8303340051580896, + "tokens_seen": 660602880 + }, + { + "epoch": 0.2, + "learning_rate": 0.0006463069207622869, + "loss": 1.5606, + "theoretical_loss": 3.8302953140382714, + "tokens_seen": 660668416 + }, + { + "epoch": 0.2, + "learning_rate": 0.0006462908726178536, + "loss": 1.4474, + "theoretical_loss": 3.8302566278308126, + "tokens_seen": 660733952 + }, + { + "epoch": 0.2, + "learning_rate": 0.0006462748244734203, + "loss": 1.4562, + "theoretical_loss": 3.830217946534602, + "tokens_seen": 660799488 + }, + { + "epoch": 0.2, + "learning_rate": 0.000646258776328987, + "loss": 1.4913, + "theoretical_loss": 3.830179270148528, + "tokens_seen": 660865024 + }, + { + "epoch": 0.2, + "learning_rate": 0.0006462427281845538, + "loss": 1.6714, + "theoretical_loss": 3.8301405986714823, + "tokens_seen": 660930560 + }, + { + "epoch": 0.2, + "learning_rate": 0.0006462266800401204, + "loss": 1.5311, + "theoretical_loss": 3.8301019321023544, + "tokens_seen": 660996096 + }, + { + "epoch": 0.2, + "learning_rate": 0.0006462106318956872, + "loss": 1.5247, + "theoretical_loss": 3.830063270440034, + "tokens_seen": 661061632 + }, + { + "epoch": 0.2, + "learning_rate": 0.0006461945837512538, + "loss": 1.4926, + "theoretical_loss": 3.8300246136834133, + "tokens_seen": 661127168 + }, + { + "epoch": 0.2, + "learning_rate": 0.0006461785356068206, + "loss": 1.4104, + "theoretical_loss": 3.829985961831383, + "tokens_seen": 661192704 + }, + { + "epoch": 0.2, + "learning_rate": 0.0006461624874623872, + "loss": 1.5004, + "theoretical_loss": 3.829947314882835, + "tokens_seen": 661258240 + }, + { + "epoch": 0.2, + "learning_rate": 0.000646146439317954, + "loss": 1.4376, + "theoretical_loss": 3.829908672836661, + "tokens_seen": 661323776 + }, + { + "epoch": 0.2, + "learning_rate": 0.0006461303911735206, + "loss": 1.5548, + "theoretical_loss": 3.829870035691754, + "tokens_seen": 661389312 + }, + { + "epoch": 0.2, + "learning_rate": 0.0006461143430290874, + "loss": 1.5807, + "theoretical_loss": 3.829831403447007, + "tokens_seen": 661454848 + }, + { + "epoch": 0.2, + "learning_rate": 0.000646098294884654, + "loss": 1.4549, + "theoretical_loss": 3.8297927761013115, + "tokens_seen": 661520384 + }, + { + "epoch": 0.2, + "learning_rate": 0.0006460822467402208, + "loss": 1.5728, + "theoretical_loss": 3.829754153653563, + "tokens_seen": 661585920 + }, + { + "epoch": 0.2, + "learning_rate": 0.0006460661985957874, + "loss": 1.5434, + "theoretical_loss": 3.829715536102654, + "tokens_seen": 661651456 + }, + { + "epoch": 0.2, + "learning_rate": 0.0006460501504513542, + "loss": 1.5344, + "theoretical_loss": 3.8296769234474795, + "tokens_seen": 661716992 + }, + { + "epoch": 0.2, + "learning_rate": 0.0006460341023069208, + "loss": 1.446, + "theoretical_loss": 3.829638315686933, + "tokens_seen": 661782528 + }, + { + "epoch": 0.2, + "learning_rate": 0.0006460180541624875, + "loss": 1.5951, + "theoretical_loss": 3.8295997128199106, + "tokens_seen": 661848064 + }, + { + "epoch": 0.2, + "objective/train/docs_used": 246530, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.5638822317123413, + "objective/train/theoretical_loss": 3.8295611148453075, + "objective/train/tokens_used": 682373600, + "theoretical_loss": 3.8295611148453075, + "tokens_seen": 661913600 + }, + { + "epoch": 0.2, + "learning_rate": 0.0006460020060180542, + "loss": 1.6339, + "theoretical_loss": 3.8295611148453075, + "tokens_seen": 661913600 + }, + { + "epoch": 0.2, + "learning_rate": 0.0006459859578736209, + "loss": 1.4674, + "theoretical_loss": 3.8295225217620183, + "tokens_seen": 661979136 + }, + { + "epoch": 0.2, + "learning_rate": 0.0006459699097291876, + "loss": 1.5165, + "theoretical_loss": 3.82948393356894, + "tokens_seen": 662044672 + }, + { + "epoch": 0.2, + "learning_rate": 0.0006459538615847543, + "loss": 1.5782, + "theoretical_loss": 3.829445350264969, + "tokens_seen": 662110208 + }, + { + "epoch": 0.2, + "learning_rate": 0.000645937813440321, + "loss": 1.6213, + "theoretical_loss": 3.8294067718490012, + "tokens_seen": 662175744 + }, + { + "epoch": 0.2, + "learning_rate": 0.0006459217652958877, + "loss": 1.628, + "theoretical_loss": 3.829368198319934, + "tokens_seen": 662241280 + }, + { + "epoch": 0.2, + "learning_rate": 0.0006459057171514543, + "loss": 1.5398, + "theoretical_loss": 3.829329629676665, + "tokens_seen": 662306816 + }, + { + "epoch": 0.2, + "learning_rate": 0.0006458896690070211, + "loss": 1.5391, + "theoretical_loss": 3.829291065918092, + "tokens_seen": 662372352 + }, + { + "epoch": 0.2, + "learning_rate": 0.0006458736208625877, + "loss": 1.4443, + "theoretical_loss": 3.8292525070431136, + "tokens_seen": 662437888 + }, + { + "epoch": 0.2, + "learning_rate": 0.0006458575727181545, + "loss": 1.4861, + "theoretical_loss": 3.829213953050627, + "tokens_seen": 662503424 + }, + { + "epoch": 0.2, + "learning_rate": 0.0006458415245737211, + "loss": 1.4522, + "theoretical_loss": 3.8291754039395323, + "tokens_seen": 662568960 + }, + { + "epoch": 0.2, + "learning_rate": 0.0006458254764292879, + "loss": 1.3979, + "theoretical_loss": 3.8291368597087283, + "tokens_seen": 662634496 + }, + { + "epoch": 0.2, + "learning_rate": 0.0006458094282848545, + "loss": 1.4957, + "theoretical_loss": 3.8290983203571147, + "tokens_seen": 662700032 + }, + { + "epoch": 0.2, + "learning_rate": 0.0006457933801404213, + "loss": 1.6095, + "theoretical_loss": 3.829059785883591, + "tokens_seen": 662765568 + }, + { + "epoch": 0.2, + "learning_rate": 0.0006457773319959879, + "loss": 1.4461, + "theoretical_loss": 3.8290212562870583, + "tokens_seen": 662831104 + }, + { + "epoch": 0.2, + "learning_rate": 0.0006457612838515547, + "loss": 1.3787, + "theoretical_loss": 3.8289827315664158, + "tokens_seen": 662896640 + }, + { + "epoch": 0.2, + "learning_rate": 0.0006457452357071213, + "loss": 1.523, + "theoretical_loss": 3.8289442117205663, + "tokens_seen": 662962176 + }, + { + "epoch": 0.2, + "learning_rate": 0.0006457291875626881, + "loss": 1.5657, + "theoretical_loss": 3.82890569674841, + "tokens_seen": 663027712 + }, + { + "epoch": 0.2, + "learning_rate": 0.0006457131394182547, + "loss": 1.5277, + "theoretical_loss": 3.828867186648849, + "tokens_seen": 663093248 + }, + { + "epoch": 0.2, + "learning_rate": 0.0006456970912738215, + "loss": 1.6843, + "theoretical_loss": 3.8288286814207853, + "tokens_seen": 663158784 + }, + { + "epoch": 0.2, + "learning_rate": 0.0006456810431293881, + "loss": 1.5404, + "theoretical_loss": 3.828790181063121, + "tokens_seen": 663224320 + }, + { + "epoch": 0.2, + "learning_rate": 0.0006456649949849549, + "loss": 1.4937, + "theoretical_loss": 3.828751685574759, + "tokens_seen": 663289856 + }, + { + "epoch": 0.2, + "learning_rate": 0.0006456489468405215, + "loss": 1.5065, + "theoretical_loss": 3.828713194954604, + "tokens_seen": 663355392 + }, + { + "epoch": 0.2, + "learning_rate": 0.0006456328986960883, + "loss": 1.4904, + "theoretical_loss": 3.8286747092015574, + "tokens_seen": 663420928 + }, + { + "epoch": 0.2, + "learning_rate": 0.0006456168505516549, + "loss": 1.5612, + "theoretical_loss": 3.828636228314523, + "tokens_seen": 663486464 + }, + { + "epoch": 0.2, + "learning_rate": 0.0006456008024072217, + "loss": 1.5836, + "theoretical_loss": 3.8285977522924073, + "tokens_seen": 663552000 + }, + { + "epoch": 0.2, + "learning_rate": 0.0006455847542627883, + "loss": 1.6205, + "theoretical_loss": 3.8285592811341127, + "tokens_seen": 663617536 + }, + { + "epoch": 0.2, + "learning_rate": 0.0006455687061183551, + "loss": 1.4846, + "theoretical_loss": 3.8285208148385443, + "tokens_seen": 663683072 + }, + { + "epoch": 0.2, + "learning_rate": 0.0006455526579739217, + "loss": 1.4694, + "theoretical_loss": 3.8284823534046084, + "tokens_seen": 663748608 + }, + { + "epoch": 0.2, + "learning_rate": 0.0006455366098294885, + "loss": 1.5035, + "theoretical_loss": 3.8284438968312102, + "tokens_seen": 663814144 + }, + { + "epoch": 0.2, + "learning_rate": 0.0006455205616850551, + "loss": 1.6345, + "theoretical_loss": 3.828405445117255, + "tokens_seen": 663879680 + }, + { + "epoch": 0.2, + "learning_rate": 0.0006455045135406219, + "loss": 1.523, + "theoretical_loss": 3.82836699826165, + "tokens_seen": 663945216 + }, + { + "epoch": 0.2, + "learning_rate": 0.0006454884653961885, + "loss": 1.5203, + "theoretical_loss": 3.828328556263302, + "tokens_seen": 664010752 + }, + { + "epoch": 0.2, + "learning_rate": 0.0006454724172517553, + "loss": 1.4788, + "theoretical_loss": 3.8282901191211174, + "tokens_seen": 664076288 + }, + { + "epoch": 0.2, + "learning_rate": 0.0006454563691073219, + "loss": 1.4997, + "theoretical_loss": 3.8282516868340037, + "tokens_seen": 664141824 + }, + { + "epoch": 0.2, + "learning_rate": 0.0006454403209628887, + "loss": 1.5753, + "theoretical_loss": 3.8282132594008687, + "tokens_seen": 664207360 + }, + { + "epoch": 0.2, + "learning_rate": 0.0006454242728184553, + "loss": 1.4964, + "theoretical_loss": 3.828174836820621, + "tokens_seen": 664272896 + }, + { + "epoch": 0.2, + "learning_rate": 0.0006454082246740221, + "loss": 1.625, + "theoretical_loss": 3.8281364190921683, + "tokens_seen": 664338432 + }, + { + "epoch": 0.2, + "learning_rate": 0.0006453921765295887, + "loss": 1.5219, + "theoretical_loss": 3.82809800621442, + "tokens_seen": 664403968 + }, + { + "epoch": 0.2, + "learning_rate": 0.0006453761283851555, + "loss": 1.383, + "theoretical_loss": 3.8280595981862846, + "tokens_seen": 664469504 + }, + { + "epoch": 0.2, + "learning_rate": 0.0006453600802407221, + "loss": 1.5342, + "theoretical_loss": 3.8280211950066723, + "tokens_seen": 664535040 + }, + { + "epoch": 0.2, + "learning_rate": 0.0006453440320962889, + "loss": 1.5769, + "theoretical_loss": 3.827982796674493, + "tokens_seen": 664600576 + }, + { + "epoch": 0.2, + "learning_rate": 0.0006453279839518556, + "loss": 1.4879, + "theoretical_loss": 3.8279444031886563, + "tokens_seen": 664666112 + }, + { + "epoch": 0.2, + "learning_rate": 0.0006453119358074223, + "loss": 1.557, + "theoretical_loss": 3.827906014548073, + "tokens_seen": 664731648 + }, + { + "epoch": 0.2, + "learning_rate": 0.000645295887662989, + "loss": 1.4425, + "theoretical_loss": 3.827867630751654, + "tokens_seen": 664797184 + }, + { + "epoch": 0.2, + "learning_rate": 0.0006452798395185557, + "loss": 1.5481, + "theoretical_loss": 3.827829251798311, + "tokens_seen": 664862720 + }, + { + "epoch": 0.2, + "learning_rate": 0.0006452637913741224, + "loss": 1.4604, + "theoretical_loss": 3.8277908776869554, + "tokens_seen": 664928256 + }, + { + "epoch": 0.2, + "learning_rate": 0.0006452477432296891, + "loss": 1.5573, + "theoretical_loss": 3.8277525084164994, + "tokens_seen": 664993792 + }, + { + "epoch": 0.2, + "learning_rate": 0.0006452316950852558, + "loss": 1.5382, + "theoretical_loss": 3.8277141439858546, + "tokens_seen": 665059328 + }, + { + "epoch": 0.2, + "learning_rate": 0.0006452156469408225, + "loss": 1.6714, + "theoretical_loss": 3.827675784393934, + "tokens_seen": 665124864 + }, + { + "epoch": 0.2, + "objective/train/docs_used": 247583, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.6360830068588257, + "objective/train/theoretical_loss": 3.827637429639651, + "objective/train/tokens_used": 685650400, + "theoretical_loss": 3.827637429639651, + "tokens_seen": 665190400 + }, + { + "epoch": 0.2, + "learning_rate": 0.0006451995987963892, + "loss": 1.5571, + "theoretical_loss": 3.827637429639651, + "tokens_seen": 665190400 + }, + { + "epoch": 0.2, + "learning_rate": 0.0006451835506519559, + "loss": 1.4367, + "theoretical_loss": 3.827599079721918, + "tokens_seen": 665255936 + }, + { + "epoch": 0.2, + "learning_rate": 0.0006451675025075226, + "loss": 1.5416, + "theoretical_loss": 3.8275607346396505, + "tokens_seen": 665321472 + }, + { + "epoch": 0.2, + "learning_rate": 0.0006451514543630893, + "loss": 1.6172, + "theoretical_loss": 3.827522394391761, + "tokens_seen": 665387008 + }, + { + "epoch": 0.2, + "learning_rate": 0.000645135406218656, + "loss": 1.4874, + "theoretical_loss": 3.8274840589771655, + "tokens_seen": 665452544 + }, + { + "epoch": 0.2, + "learning_rate": 0.0006451193580742227, + "loss": 1.4465, + "theoretical_loss": 3.8274457283947765, + "tokens_seen": 665518080 + }, + { + "epoch": 0.2, + "learning_rate": 0.0006451033099297894, + "loss": 1.4992, + "theoretical_loss": 3.8274074026435114, + "tokens_seen": 665583616 + }, + { + "epoch": 0.2, + "learning_rate": 0.0006450872617853561, + "loss": 1.4914, + "theoretical_loss": 3.8273690817222845, + "tokens_seen": 665649152 + }, + { + "epoch": 0.2, + "learning_rate": 0.0006450712136409228, + "loss": 1.3835, + "theoretical_loss": 3.8273307656300117, + "tokens_seen": 665714688 + }, + { + "epoch": 0.2, + "learning_rate": 0.0006450551654964895, + "loss": 1.4509, + "theoretical_loss": 3.8272924543656095, + "tokens_seen": 665780224 + }, + { + "epoch": 0.2, + "learning_rate": 0.0006450391173520563, + "loss": 1.5266, + "theoretical_loss": 3.8272541479279942, + "tokens_seen": 665845760 + }, + { + "epoch": 0.2, + "learning_rate": 0.0006450230692076229, + "loss": 1.5182, + "theoretical_loss": 3.827215846316083, + "tokens_seen": 665911296 + }, + { + "epoch": 0.2, + "learning_rate": 0.0006450070210631897, + "loss": 1.4986, + "theoretical_loss": 3.827177549528793, + "tokens_seen": 665976832 + }, + { + "epoch": 0.2, + "learning_rate": 0.0006449909729187563, + "loss": 1.5195, + "theoretical_loss": 3.8271392575650416, + "tokens_seen": 666042368 + }, + { + "epoch": 0.2, + "learning_rate": 0.0006449749247743231, + "loss": 1.4888, + "theoretical_loss": 3.827100970423747, + "tokens_seen": 666107904 + }, + { + "epoch": 0.2, + "learning_rate": 0.0006449588766298897, + "loss": 1.5595, + "theoretical_loss": 3.827062688103827, + "tokens_seen": 666173440 + }, + { + "epoch": 0.2, + "learning_rate": 0.0006449428284854565, + "loss": 1.5329, + "theoretical_loss": 3.827024410604201, + "tokens_seen": 666238976 + }, + { + "epoch": 0.2, + "learning_rate": 0.0006449267803410231, + "loss": 1.5353, + "theoretical_loss": 3.8269861379237877, + "tokens_seen": 666304512 + }, + { + "epoch": 0.2, + "learning_rate": 0.0006449107321965899, + "loss": 1.4798, + "theoretical_loss": 3.826947870061506, + "tokens_seen": 666370048 + }, + { + "epoch": 0.2, + "learning_rate": 0.0006448946840521565, + "loss": 1.5086, + "theoretical_loss": 3.826909607016276, + "tokens_seen": 666435584 + }, + { + "epoch": 0.2, + "learning_rate": 0.0006448786359077233, + "loss": 1.5387, + "theoretical_loss": 3.8268713487870176, + "tokens_seen": 666501120 + }, + { + "epoch": 0.2, + "learning_rate": 0.0006448625877632899, + "loss": 1.4249, + "theoretical_loss": 3.8268330953726513, + "tokens_seen": 666566656 + }, + { + "epoch": 0.2, + "learning_rate": 0.0006448465396188567, + "loss": 1.542, + "theoretical_loss": 3.826794846772098, + "tokens_seen": 666632192 + }, + { + "epoch": 0.2, + "learning_rate": 0.0006448304914744233, + "loss": 1.5988, + "theoretical_loss": 3.826756602984278, + "tokens_seen": 666697728 + }, + { + "epoch": 0.2, + "learning_rate": 0.0006448144433299901, + "loss": 1.4711, + "theoretical_loss": 3.826718364008114, + "tokens_seen": 666763264 + }, + { + "epoch": 0.2, + "learning_rate": 0.0006447983951855567, + "loss": 1.5786, + "theoretical_loss": 3.826680129842526, + "tokens_seen": 666828800 + }, + { + "epoch": 0.2, + "learning_rate": 0.0006447823470411235, + "loss": 1.56, + "theoretical_loss": 3.826641900486438, + "tokens_seen": 666894336 + }, + { + "epoch": 0.2, + "learning_rate": 0.0006447662988966901, + "loss": 1.4832, + "theoretical_loss": 3.8266036759387716, + "tokens_seen": 666959872 + }, + { + "epoch": 0.2, + "learning_rate": 0.0006447502507522569, + "loss": 1.4462, + "theoretical_loss": 3.8265654561984492, + "tokens_seen": 667025408 + }, + { + "epoch": 0.2, + "learning_rate": 0.0006447342026078235, + "loss": 1.6325, + "theoretical_loss": 3.8265272412643947, + "tokens_seen": 667090944 + }, + { + "epoch": 0.2, + "learning_rate": 0.0006447181544633903, + "loss": 1.4092, + "theoretical_loss": 3.8264890311355315, + "tokens_seen": 667156480 + }, + { + "epoch": 0.2, + "learning_rate": 0.0006447021063189569, + "loss": 1.4455, + "theoretical_loss": 3.8264508258107828, + "tokens_seen": 667222016 + }, + { + "epoch": 0.2, + "learning_rate": 0.0006446860581745237, + "loss": 1.5424, + "theoretical_loss": 3.8264126252890738, + "tokens_seen": 667287552 + }, + { + "epoch": 0.2, + "learning_rate": 0.0006446700100300903, + "loss": 1.4982, + "theoretical_loss": 3.826374429569328, + "tokens_seen": 667353088 + }, + { + "epoch": 0.2, + "learning_rate": 0.0006446539618856571, + "loss": 1.5104, + "theoretical_loss": 3.8263362386504713, + "tokens_seen": 667418624 + }, + { + "epoch": 0.2, + "learning_rate": 0.0006446379137412237, + "loss": 1.5611, + "theoretical_loss": 3.826298052531428, + "tokens_seen": 667484160 + }, + { + "epoch": 0.2, + "learning_rate": 0.0006446218655967904, + "loss": 1.5581, + "theoretical_loss": 3.8262598712111244, + "tokens_seen": 667549696 + }, + { + "epoch": 0.2, + "learning_rate": 0.0006446058174523571, + "loss": 1.5072, + "theoretical_loss": 3.8262216946884866, + "tokens_seen": 667615232 + }, + { + "epoch": 0.2, + "learning_rate": 0.0006445897693079238, + "loss": 1.5141, + "theoretical_loss": 3.8261835229624395, + "tokens_seen": 667680768 + }, + { + "epoch": 0.2, + "learning_rate": 0.0006445737211634905, + "loss": 1.4641, + "theoretical_loss": 3.8261453560319114, + "tokens_seen": 667746304 + }, + { + "epoch": 0.2, + "learning_rate": 0.0006445576730190572, + "loss": 1.4197, + "theoretical_loss": 3.8261071938958278, + "tokens_seen": 667811840 + }, + { + "epoch": 0.2, + "learning_rate": 0.0006445416248746239, + "loss": 1.5679, + "theoretical_loss": 3.826069036553118, + "tokens_seen": 667877376 + }, + { + "epoch": 0.2, + "learning_rate": 0.0006445255767301906, + "loss": 1.4958, + "theoretical_loss": 3.8260308840027073, + "tokens_seen": 667942912 + }, + { + "epoch": 0.2, + "learning_rate": 0.0006445095285857573, + "loss": 1.5198, + "theoretical_loss": 3.825992736243525, + "tokens_seen": 668008448 + }, + { + "epoch": 0.2, + "learning_rate": 0.000644493480441324, + "loss": 1.4544, + "theoretical_loss": 3.8259545932745, + "tokens_seen": 668073984 + }, + { + "epoch": 0.2, + "learning_rate": 0.0006444774322968906, + "loss": 1.4953, + "theoretical_loss": 3.8259164550945597, + "tokens_seen": 668139520 + }, + { + "epoch": 0.2, + "learning_rate": 0.0006444613841524574, + "loss": 1.5902, + "theoretical_loss": 3.8258783217026338, + "tokens_seen": 668205056 + }, + { + "epoch": 0.2, + "learning_rate": 0.000644445336008024, + "loss": 1.5015, + "theoretical_loss": 3.825840193097652, + "tokens_seen": 668270592 + }, + { + "epoch": 0.2, + "learning_rate": 0.0006444292878635908, + "loss": 1.5294, + "theoretical_loss": 3.8258020692785433, + "tokens_seen": 668336128 + }, + { + "epoch": 0.2, + "learning_rate": 0.0006444132397191574, + "loss": 1.502, + "theoretical_loss": 3.825763950244238, + "tokens_seen": 668401664 + }, + { + "epoch": 0.2, + "objective/train/docs_used": 248744, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.5574264526367188, + "objective/train/theoretical_loss": 3.8257258359936666, + "objective/train/tokens_used": 688927200, + "theoretical_loss": 3.8257258359936666, + "tokens_seen": 668467200 + }, + { + "epoch": 0.2, + "learning_rate": 0.0006443971915747242, + "loss": 1.6049, + "theoretical_loss": 3.8257258359936666, + "tokens_seen": 668467200 + }, + { + "epoch": 0.2, + "learning_rate": 0.0006443811434302908, + "loss": 1.461, + "theoretical_loss": 3.82568772652576, + "tokens_seen": 668532736 + }, + { + "epoch": 0.2, + "learning_rate": 0.0006443650952858576, + "loss": 1.5712, + "theoretical_loss": 3.8256496218394496, + "tokens_seen": 668598272 + }, + { + "epoch": 0.2, + "learning_rate": 0.0006443490471414242, + "loss": 1.4614, + "theoretical_loss": 3.825611521933666, + "tokens_seen": 668663808 + }, + { + "epoch": 0.2, + "learning_rate": 0.000644332998996991, + "loss": 1.4289, + "theoretical_loss": 3.8255734268073414, + "tokens_seen": 668729344 + }, + { + "epoch": 0.2, + "learning_rate": 0.0006443169508525576, + "loss": 1.578, + "theoretical_loss": 3.8255353364594082, + "tokens_seen": 668794880 + }, + { + "epoch": 0.2, + "learning_rate": 0.0006443009027081244, + "loss": 1.4347, + "theoretical_loss": 3.825497250888799, + "tokens_seen": 668860416 + }, + { + "epoch": 0.2, + "learning_rate": 0.000644284854563691, + "loss": 1.5468, + "theoretical_loss": 3.8254591700944456, + "tokens_seen": 668925952 + }, + { + "epoch": 0.2, + "learning_rate": 0.0006442688064192578, + "loss": 1.456, + "theoretical_loss": 3.825421094075282, + "tokens_seen": 668991488 + }, + { + "epoch": 0.2, + "learning_rate": 0.0006442527582748244, + "loss": 1.4528, + "theoretical_loss": 3.8253830228302417, + "tokens_seen": 669057024 + }, + { + "epoch": 0.2, + "learning_rate": 0.0006442367101303912, + "loss": 1.5076, + "theoretical_loss": 3.825344956358258, + "tokens_seen": 669122560 + }, + { + "epoch": 0.2, + "learning_rate": 0.0006442206619859578, + "loss": 1.5567, + "theoretical_loss": 3.825306894658266, + "tokens_seen": 669188096 + }, + { + "epoch": 0.2, + "learning_rate": 0.0006442046138415246, + "loss": 1.4783, + "theoretical_loss": 3.8252688377292, + "tokens_seen": 669253632 + }, + { + "epoch": 0.2, + "learning_rate": 0.0006441885656970912, + "loss": 1.5414, + "theoretical_loss": 3.8252307855699943, + "tokens_seen": 669319168 + }, + { + "epoch": 0.2, + "learning_rate": 0.000644172517552658, + "loss": 1.4999, + "theoretical_loss": 3.8251927381795845, + "tokens_seen": 669384704 + }, + { + "epoch": 0.2, + "learning_rate": 0.0006441564694082246, + "loss": 1.5093, + "theoretical_loss": 3.825154695556906, + "tokens_seen": 669450240 + }, + { + "epoch": 0.2, + "learning_rate": 0.0006441404212637914, + "loss": 1.4994, + "theoretical_loss": 3.825116657700894, + "tokens_seen": 669515776 + }, + { + "epoch": 0.2, + "learning_rate": 0.000644124373119358, + "loss": 1.5943, + "theoretical_loss": 3.8250786246104864, + "tokens_seen": 669581312 + }, + { + "epoch": 0.2, + "learning_rate": 0.0006441083249749248, + "loss": 1.4878, + "theoretical_loss": 3.825040596284619, + "tokens_seen": 669646848 + }, + { + "epoch": 0.2, + "learning_rate": 0.0006440922768304914, + "loss": 1.5009, + "theoretical_loss": 3.8250025727222283, + "tokens_seen": 669712384 + }, + { + "epoch": 0.2, + "learning_rate": 0.0006440762286860582, + "loss": 1.5723, + "theoretical_loss": 3.8249645539222525, + "tokens_seen": 669777920 + }, + { + "epoch": 0.2, + "learning_rate": 0.0006440601805416248, + "loss": 1.4769, + "theoretical_loss": 3.824926539883628, + "tokens_seen": 669843456 + }, + { + "epoch": 0.2, + "learning_rate": 0.0006440441323971916, + "loss": 1.473, + "theoretical_loss": 3.8248885306052935, + "tokens_seen": 669908992 + }, + { + "epoch": 0.2, + "learning_rate": 0.0006440280842527582, + "loss": 1.5879, + "theoretical_loss": 3.8248505260861876, + "tokens_seen": 669974528 + }, + { + "epoch": 0.2, + "learning_rate": 0.000644012036108325, + "loss": 1.4788, + "theoretical_loss": 3.824812526325248, + "tokens_seen": 670040064 + }, + { + "epoch": 0.2, + "learning_rate": 0.0006439959879638916, + "loss": 1.505, + "theoretical_loss": 3.824774531321414, + "tokens_seen": 670105600 + }, + { + "epoch": 0.2, + "learning_rate": 0.0006439799398194584, + "loss": 1.4466, + "theoretical_loss": 3.8247365410736256, + "tokens_seen": 670171136 + }, + { + "epoch": 0.2, + "learning_rate": 0.000643963891675025, + "loss": 1.5941, + "theoretical_loss": 3.824698555580822, + "tokens_seen": 670236672 + }, + { + "epoch": 0.2, + "learning_rate": 0.0006439478435305918, + "loss": 1.5316, + "theoretical_loss": 3.8246605748419427, + "tokens_seen": 670302208 + }, + { + "epoch": 0.2, + "learning_rate": 0.0006439317953861585, + "loss": 1.4676, + "theoretical_loss": 3.824622598855929, + "tokens_seen": 670367744 + }, + { + "epoch": 0.2, + "learning_rate": 0.0006439157472417252, + "loss": 1.3052, + "theoretical_loss": 3.824584627621721, + "tokens_seen": 670433280 + }, + { + "epoch": 0.2, + "learning_rate": 0.0006438996990972919, + "loss": 1.4175, + "theoretical_loss": 3.8245466611382595, + "tokens_seen": 670498816 + }, + { + "epoch": 0.2, + "learning_rate": 0.0006438836509528586, + "loss": 1.4885, + "theoretical_loss": 3.824508699404486, + "tokens_seen": 670564352 + }, + { + "epoch": 0.2, + "learning_rate": 0.0006438676028084253, + "loss": 1.5674, + "theoretical_loss": 3.8244707424193427, + "tokens_seen": 670629888 + }, + { + "epoch": 0.2, + "learning_rate": 0.000643851554663992, + "loss": 1.5093, + "theoretical_loss": 3.8244327901817705, + "tokens_seen": 670695424 + }, + { + "epoch": 0.2, + "learning_rate": 0.0006438355065195588, + "loss": 1.5557, + "theoretical_loss": 3.8243948426907135, + "tokens_seen": 670760960 + }, + { + "epoch": 0.2, + "learning_rate": 0.0006438194583751254, + "loss": 1.5315, + "theoretical_loss": 3.824356899945113, + "tokens_seen": 670826496 + }, + { + "epoch": 0.2, + "learning_rate": 0.0006438034102306922, + "loss": 1.566, + "theoretical_loss": 3.824318961943912, + "tokens_seen": 670892032 + }, + { + "epoch": 0.2, + "learning_rate": 0.0006437873620862588, + "loss": 1.4504, + "theoretical_loss": 3.8242810286860545, + "tokens_seen": 670957568 + }, + { + "epoch": 0.2, + "learning_rate": 0.0006437713139418256, + "loss": 1.4832, + "theoretical_loss": 3.8242431001704844, + "tokens_seen": 671023104 + }, + { + "epoch": 0.2, + "learning_rate": 0.0006437552657973922, + "loss": 1.5898, + "theoretical_loss": 3.8242051763961458, + "tokens_seen": 671088640 + }, + { + "epoch": 0.2, + "learning_rate": 0.000643739217652959, + "loss": 1.5143, + "theoretical_loss": 3.824167257361982, + "tokens_seen": 671154176 + }, + { + "epoch": 0.2, + "learning_rate": 0.0006437231695085256, + "loss": 1.4321, + "theoretical_loss": 3.8241293430669385, + "tokens_seen": 671219712 + }, + { + "epoch": 0.2, + "learning_rate": 0.0006437071213640924, + "loss": 1.4716, + "theoretical_loss": 3.824091433509961, + "tokens_seen": 671285248 + }, + { + "epoch": 0.2, + "learning_rate": 0.000643691073219659, + "loss": 1.4795, + "theoretical_loss": 3.824053528689994, + "tokens_seen": 671350784 + }, + { + "epoch": 0.2, + "learning_rate": 0.0006436750250752258, + "loss": 1.5441, + "theoretical_loss": 3.8240156286059834, + "tokens_seen": 671416320 + }, + { + "epoch": 0.2, + "learning_rate": 0.0006436589769307924, + "loss": 1.5374, + "theoretical_loss": 3.823977733256875, + "tokens_seen": 671481856 + }, + { + "epoch": 0.2, + "learning_rate": 0.0006436429287863592, + "loss": 1.562, + "theoretical_loss": 3.8239398426416162, + "tokens_seen": 671547392 + }, + { + "epoch": 0.2, + "learning_rate": 0.0006436268806419258, + "loss": 1.5878, + "theoretical_loss": 3.823901956759153, + "tokens_seen": 671612928 + }, + { + "epoch": 0.2, + "learning_rate": 0.0006436108324974926, + "loss": 1.5609, + "theoretical_loss": 3.8238640756084328, + "tokens_seen": 671678464 + }, + { + "epoch": 0.2, + "objective/train/docs_used": 249880, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.4048361778259277, + "objective/train/theoretical_loss": 3.823826199188403, + "objective/train/tokens_used": 692204000, + "theoretical_loss": 3.823826199188403, + "tokens_seen": 671744000 + }, + { + "epoch": 0.2, + "learning_rate": 0.0006435947843530592, + "loss": 1.4399, + "theoretical_loss": 3.823826199188403, + "tokens_seen": 671744000 + }, + { + "epoch": 0.2, + "learning_rate": 0.000643578736208626, + "loss": 1.3892, + "theoretical_loss": 3.8237883274980113, + "tokens_seen": 671809536 + }, + { + "epoch": 0.2, + "learning_rate": 0.0006435626880641926, + "loss": 1.4428, + "theoretical_loss": 3.823750460536206, + "tokens_seen": 671875072 + }, + { + "epoch": 0.2, + "learning_rate": 0.0006435466399197594, + "loss": 1.4283, + "theoretical_loss": 3.8237125983019347, + "tokens_seen": 671940608 + }, + { + "epoch": 0.2, + "learning_rate": 0.000643530591775326, + "loss": 1.5617, + "theoretical_loss": 3.8236747407941474, + "tokens_seen": 672006144 + }, + { + "epoch": 0.2, + "learning_rate": 0.0006435145436308928, + "loss": 1.4885, + "theoretical_loss": 3.8236368880117926, + "tokens_seen": 672071680 + }, + { + "epoch": 0.2, + "learning_rate": 0.0006434984954864594, + "loss": 1.4137, + "theoretical_loss": 3.82359903995382, + "tokens_seen": 672137216 + }, + { + "epoch": 0.2, + "learning_rate": 0.0006434824473420262, + "loss": 1.4754, + "theoretical_loss": 3.823561196619179, + "tokens_seen": 672202752 + }, + { + "epoch": 0.2, + "learning_rate": 0.0006434663991975928, + "loss": 1.54, + "theoretical_loss": 3.82352335800682, + "tokens_seen": 672268288 + }, + { + "epoch": 0.2, + "learning_rate": 0.0006434503510531596, + "loss": 1.5267, + "theoretical_loss": 3.8234855241156938, + "tokens_seen": 672333824 + }, + { + "epoch": 0.2, + "learning_rate": 0.0006434343029087262, + "loss": 1.4521, + "theoretical_loss": 3.8234476949447505, + "tokens_seen": 672399360 + }, + { + "epoch": 0.2, + "learning_rate": 0.000643418254764293, + "loss": 1.5515, + "theoretical_loss": 3.8234098704929416, + "tokens_seen": 672464896 + }, + { + "epoch": 0.2, + "learning_rate": 0.0006434022066198596, + "loss": 1.4172, + "theoretical_loss": 3.823372050759218, + "tokens_seen": 672530432 + }, + { + "epoch": 0.2, + "learning_rate": 0.0006433861584754264, + "loss": 1.5512, + "theoretical_loss": 3.823334235742533, + "tokens_seen": 672595968 + }, + { + "epoch": 0.2, + "learning_rate": 0.000643370110330993, + "loss": 1.5247, + "theoretical_loss": 3.8232964254418373, + "tokens_seen": 672661504 + }, + { + "epoch": 0.2, + "learning_rate": 0.0006433540621865598, + "loss": 1.5354, + "theoretical_loss": 3.823258619856084, + "tokens_seen": 672727040 + }, + { + "epoch": 0.2, + "learning_rate": 0.0006433380140421264, + "loss": 1.5758, + "theoretical_loss": 3.8232208189842254, + "tokens_seen": 672792576 + }, + { + "epoch": 0.2, + "learning_rate": 0.0006433219658976932, + "loss": 1.4288, + "theoretical_loss": 3.8231830228252157, + "tokens_seen": 672858112 + }, + { + "epoch": 0.2, + "learning_rate": 0.0006433059177532598, + "loss": 1.5737, + "theoretical_loss": 3.8231452313780068, + "tokens_seen": 672923648 + }, + { + "epoch": 0.2, + "learning_rate": 0.0006432898696088266, + "loss": 1.4694, + "theoretical_loss": 3.823107444641554, + "tokens_seen": 672989184 + }, + { + "epoch": 0.2, + "learning_rate": 0.0006432738214643932, + "loss": 1.55, + "theoretical_loss": 3.8230696626148113, + "tokens_seen": 673054720 + }, + { + "epoch": 0.2, + "learning_rate": 0.00064325777331996, + "loss": 1.4606, + "theoretical_loss": 3.823031885296732, + "tokens_seen": 673120256 + }, + { + "epoch": 0.2, + "learning_rate": 0.0006432417251755266, + "loss": 1.6337, + "theoretical_loss": 3.822994112686272, + "tokens_seen": 673185792 + }, + { + "epoch": 0.2, + "learning_rate": 0.0006432256770310933, + "loss": 1.5487, + "theoretical_loss": 3.8229563447823858, + "tokens_seen": 673251328 + }, + { + "epoch": 0.2, + "learning_rate": 0.00064320962888666, + "loss": 1.4138, + "theoretical_loss": 3.8229185815840294, + "tokens_seen": 673316864 + }, + { + "epoch": 0.2, + "learning_rate": 0.0006431935807422267, + "loss": 1.4982, + "theoretical_loss": 3.8228808230901583, + "tokens_seen": 673382400 + }, + { + "epoch": 0.2, + "learning_rate": 0.0006431775325977934, + "loss": 1.5057, + "theoretical_loss": 3.822843069299729, + "tokens_seen": 673447936 + }, + { + "epoch": 0.2, + "learning_rate": 0.0006431614844533601, + "loss": 1.5416, + "theoretical_loss": 3.822805320211698, + "tokens_seen": 673513472 + }, + { + "epoch": 0.2, + "learning_rate": 0.0006431454363089268, + "loss": 1.5531, + "theoretical_loss": 3.8227675758250212, + "tokens_seen": 673579008 + }, + { + "epoch": 0.2, + "learning_rate": 0.0006431293881644935, + "loss": 1.5202, + "theoretical_loss": 3.822729836138657, + "tokens_seen": 673644544 + }, + { + "epoch": 0.2, + "learning_rate": 0.0006431133400200602, + "loss": 1.4155, + "theoretical_loss": 3.8226921011515618, + "tokens_seen": 673710080 + }, + { + "epoch": 0.2, + "learning_rate": 0.0006430972918756269, + "loss": 1.4428, + "theoretical_loss": 3.822654370862694, + "tokens_seen": 673775616 + }, + { + "epoch": 0.2, + "learning_rate": 0.0006430812437311935, + "loss": 1.4684, + "theoretical_loss": 3.8226166452710117, + "tokens_seen": 673841152 + }, + { + "epoch": 0.2, + "learning_rate": 0.0006430651955867603, + "loss": 1.5644, + "theoretical_loss": 3.822578924375473, + "tokens_seen": 673906688 + }, + { + "epoch": 0.2, + "learning_rate": 0.0006430491474423269, + "loss": 1.534, + "theoretical_loss": 3.8225412081750374, + "tokens_seen": 673972224 + }, + { + "epoch": 0.2, + "learning_rate": 0.0006430330992978937, + "loss": 1.4774, + "theoretical_loss": 3.8225034966686637, + "tokens_seen": 674037760 + }, + { + "epoch": 0.2, + "learning_rate": 0.0006430170511534603, + "loss": 1.5413, + "theoretical_loss": 3.822465789855311, + "tokens_seen": 674103296 + }, + { + "epoch": 0.2, + "learning_rate": 0.0006430010030090271, + "loss": 1.5644, + "theoretical_loss": 3.82242808773394, + "tokens_seen": 674168832 + }, + { + "epoch": 0.2, + "learning_rate": 0.0006429849548645937, + "loss": 1.5112, + "theoretical_loss": 3.82239039030351, + "tokens_seen": 674234368 + }, + { + "epoch": 0.2, + "learning_rate": 0.0006429689067201605, + "loss": 1.5015, + "theoretical_loss": 3.822352697562982, + "tokens_seen": 674299904 + }, + { + "epoch": 0.2, + "learning_rate": 0.0006429528585757271, + "loss": 1.3863, + "theoretical_loss": 3.8223150095113163, + "tokens_seen": 674365440 + }, + { + "epoch": 0.2, + "learning_rate": 0.0006429368104312939, + "loss": 1.5577, + "theoretical_loss": 3.822277326147475, + "tokens_seen": 674430976 + }, + { + "epoch": 0.2, + "learning_rate": 0.0006429207622868605, + "loss": 1.4557, + "theoretical_loss": 3.822239647470418, + "tokens_seen": 674496512 + }, + { + "epoch": 0.2, + "learning_rate": 0.0006429047141424273, + "loss": 1.4642, + "theoretical_loss": 3.822201973479108, + "tokens_seen": 674562048 + }, + { + "epoch": 0.2, + "learning_rate": 0.0006428886659979939, + "loss": 1.5047, + "theoretical_loss": 3.8221643041725075, + "tokens_seen": 674627584 + }, + { + "epoch": 0.2, + "learning_rate": 0.0006428726178535607, + "loss": 1.5183, + "theoretical_loss": 3.8221266395495785, + "tokens_seen": 674693120 + }, + { + "epoch": 0.2, + "learning_rate": 0.0006428565697091273, + "loss": 1.5672, + "theoretical_loss": 3.822088979609284, + "tokens_seen": 674758656 + }, + { + "epoch": 0.2, + "learning_rate": 0.0006428405215646941, + "loss": 1.4209, + "theoretical_loss": 3.8220513243505865, + "tokens_seen": 674824192 + }, + { + "epoch": 0.2, + "learning_rate": 0.0006428244734202607, + "loss": 1.4298, + "theoretical_loss": 3.8220136737724504, + "tokens_seen": 674889728 + }, + { + "epoch": 0.2, + "learning_rate": 0.0006428084252758275, + "loss": 1.4014, + "theoretical_loss": 3.8219760278738386, + "tokens_seen": 674955264 + }, + { + "epoch": 0.2, + "objective/train/docs_used": 251058, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.5011433362960815, + "objective/train/theoretical_loss": 3.821938386653716, + "objective/train/tokens_used": 695480800, + "theoretical_loss": 3.821938386653716, + "tokens_seen": 675020800 + }, + { + "epoch": 0.2, + "learning_rate": 0.0006427923771313941, + "loss": 1.4747, + "theoretical_loss": 3.821938386653716, + "tokens_seen": 675020800 + }, + { + "epoch": 0.2, + "learning_rate": 0.0006427763289869609, + "loss": 1.467, + "theoretical_loss": 3.8219007501110465, + "tokens_seen": 675086336 + }, + { + "epoch": 0.2, + "learning_rate": 0.0006427602808425275, + "loss": 1.5073, + "theoretical_loss": 3.8218631182447944, + "tokens_seen": 675151872 + }, + { + "epoch": 0.2, + "learning_rate": 0.0006427442326980943, + "loss": 1.503, + "theoretical_loss": 3.821825491053926, + "tokens_seen": 675217408 + }, + { + "epoch": 0.2, + "learning_rate": 0.0006427281845536609, + "loss": 1.4456, + "theoretical_loss": 3.821787868537406, + "tokens_seen": 675282944 + }, + { + "epoch": 0.2, + "learning_rate": 0.0006427121364092277, + "loss": 1.498, + "theoretical_loss": 3.8217502506942003, + "tokens_seen": 675348480 + }, + { + "epoch": 0.2, + "learning_rate": 0.0006426960882647943, + "loss": 1.4411, + "theoretical_loss": 3.821712637523275, + "tokens_seen": 675414016 + }, + { + "epoch": 0.2, + "learning_rate": 0.0006426800401203611, + "loss": 1.4438, + "theoretical_loss": 3.8216750290235963, + "tokens_seen": 675479552 + }, + { + "epoch": 0.2, + "learning_rate": 0.0006426639919759277, + "loss": 1.466, + "theoretical_loss": 3.8216374251941314, + "tokens_seen": 675545088 + }, + { + "epoch": 0.2, + "learning_rate": 0.0006426479438314945, + "loss": 1.5131, + "theoretical_loss": 3.8215998260338466, + "tokens_seen": 675610624 + }, + { + "epoch": 0.2, + "learning_rate": 0.0006426318956870613, + "loss": 1.569, + "theoretical_loss": 3.8215622315417104, + "tokens_seen": 675676160 + }, + { + "epoch": 0.2, + "learning_rate": 0.0006426158475426279, + "loss": 1.5115, + "theoretical_loss": 3.8215246417166893, + "tokens_seen": 675741696 + }, + { + "epoch": 0.2, + "learning_rate": 0.0006425997993981947, + "loss": 1.4387, + "theoretical_loss": 3.821487056557752, + "tokens_seen": 675807232 + }, + { + "epoch": 0.2, + "learning_rate": 0.0006425837512537613, + "loss": 1.4596, + "theoretical_loss": 3.8214494760638673, + "tokens_seen": 675872768 + }, + { + "epoch": 0.2, + "learning_rate": 0.0006425677031093281, + "loss": 1.4274, + "theoretical_loss": 3.8214119002340032, + "tokens_seen": 675938304 + }, + { + "epoch": 0.2, + "learning_rate": 0.0006425516549648947, + "loss": 1.5372, + "theoretical_loss": 3.821374329067129, + "tokens_seen": 676003840 + }, + { + "epoch": 0.2, + "learning_rate": 0.0006425356068204615, + "loss": 1.4557, + "theoretical_loss": 3.8213367625622143, + "tokens_seen": 676069376 + }, + { + "epoch": 0.2, + "learning_rate": 0.0006425195586760281, + "loss": 1.432, + "theoretical_loss": 3.8212992007182285, + "tokens_seen": 676134912 + }, + { + "epoch": 0.2, + "learning_rate": 0.0006425035105315949, + "loss": 1.5594, + "theoretical_loss": 3.8212616435341413, + "tokens_seen": 676200448 + }, + { + "epoch": 0.2, + "learning_rate": 0.0006424874623871615, + "loss": 1.4898, + "theoretical_loss": 3.821224091008924, + "tokens_seen": 676265984 + }, + { + "epoch": 0.2, + "learning_rate": 0.0006424714142427283, + "loss": 1.503, + "theoretical_loss": 3.821186543141547, + "tokens_seen": 676331520 + }, + { + "epoch": 0.2, + "learning_rate": 0.0006424553660982949, + "loss": 1.4861, + "theoretical_loss": 3.8211489999309807, + "tokens_seen": 676397056 + }, + { + "epoch": 0.2, + "learning_rate": 0.0006424393179538617, + "loss": 1.5245, + "theoretical_loss": 3.821111461376196, + "tokens_seen": 676462592 + }, + { + "epoch": 0.21, + "learning_rate": 0.0006424232698094283, + "loss": 1.5362, + "theoretical_loss": 3.8210739274761663, + "tokens_seen": 676528128 + }, + { + "epoch": 0.21, + "learning_rate": 0.0006424072216649951, + "loss": 1.4873, + "theoretical_loss": 3.8210363982298623, + "tokens_seen": 676593664 + }, + { + "epoch": 0.21, + "learning_rate": 0.0006423911735205617, + "loss": 1.5383, + "theoretical_loss": 3.820998873636257, + "tokens_seen": 676659200 + }, + { + "epoch": 0.21, + "learning_rate": 0.0006423751253761285, + "loss": 1.5201, + "theoretical_loss": 3.8209613536943223, + "tokens_seen": 676724736 + }, + { + "epoch": 0.21, + "learning_rate": 0.0006423590772316951, + "loss": 1.4221, + "theoretical_loss": 3.8209238384030315, + "tokens_seen": 676790272 + }, + { + "epoch": 0.21, + "learning_rate": 0.0006423430290872619, + "loss": 1.5278, + "theoretical_loss": 3.820886327761358, + "tokens_seen": 676855808 + }, + { + "epoch": 0.21, + "learning_rate": 0.0006423269809428285, + "loss": 1.5339, + "theoretical_loss": 3.820848821768275, + "tokens_seen": 676921344 + }, + { + "epoch": 0.21, + "learning_rate": 0.0006423109327983953, + "loss": 1.3614, + "theoretical_loss": 3.8208113204227576, + "tokens_seen": 676986880 + }, + { + "epoch": 0.21, + "learning_rate": 0.0006422948846539619, + "loss": 1.462, + "theoretical_loss": 3.8207738237237785, + "tokens_seen": 677052416 + }, + { + "epoch": 0.21, + "learning_rate": 0.0006422788365095287, + "loss": 1.502, + "theoretical_loss": 3.8207363316703136, + "tokens_seen": 677117952 + }, + { + "epoch": 0.21, + "learning_rate": 0.0006422627883650953, + "loss": 1.3617, + "theoretical_loss": 3.820698844261337, + "tokens_seen": 677183488 + }, + { + "epoch": 0.21, + "learning_rate": 0.0006422467402206621, + "loss": 1.5573, + "theoretical_loss": 3.8206613614958242, + "tokens_seen": 677249024 + }, + { + "epoch": 0.21, + "learning_rate": 0.0006422306920762287, + "loss": 1.5098, + "theoretical_loss": 3.820623883372751, + "tokens_seen": 677314560 + }, + { + "epoch": 0.21, + "learning_rate": 0.0006422146439317955, + "loss": 1.504, + "theoretical_loss": 3.820586409891093, + "tokens_seen": 677380096 + }, + { + "epoch": 0.21, + "learning_rate": 0.0006421985957873621, + "loss": 1.5154, + "theoretical_loss": 3.820548941049826, + "tokens_seen": 677445632 + }, + { + "epoch": 0.21, + "learning_rate": 0.0006421825476429289, + "loss": 1.4268, + "theoretical_loss": 3.820511476847928, + "tokens_seen": 677511168 + }, + { + "epoch": 0.21, + "learning_rate": 0.0006421664994984955, + "loss": 1.5008, + "theoretical_loss": 3.8204740172843747, + "tokens_seen": 677576704 + }, + { + "epoch": 0.21, + "learning_rate": 0.0006421504513540623, + "loss": 1.5132, + "theoretical_loss": 3.820436562358143, + "tokens_seen": 677642240 + }, + { + "epoch": 0.21, + "learning_rate": 0.0006421344032096289, + "loss": 1.5621, + "theoretical_loss": 3.8203991120682117, + "tokens_seen": 677707776 + }, + { + "epoch": 0.21, + "learning_rate": 0.0006421183550651957, + "loss": 1.4657, + "theoretical_loss": 3.8203616664135573, + "tokens_seen": 677773312 + }, + { + "epoch": 0.21, + "learning_rate": 0.0006421023069207623, + "loss": 1.5049, + "theoretical_loss": 3.8203242253931586, + "tokens_seen": 677838848 + }, + { + "epoch": 0.21, + "learning_rate": 0.0006420862587763291, + "loss": 1.4008, + "theoretical_loss": 3.820286789005994, + "tokens_seen": 677904384 + }, + { + "epoch": 0.21, + "learning_rate": 0.0006420702106318957, + "loss": 1.4735, + "theoretical_loss": 3.820249357251043, + "tokens_seen": 677969920 + }, + { + "epoch": 0.21, + "learning_rate": 0.0006420541624874625, + "loss": 1.505, + "theoretical_loss": 3.8202119301272837, + "tokens_seen": 678035456 + }, + { + "epoch": 0.21, + "learning_rate": 0.0006420381143430291, + "loss": 1.4799, + "theoretical_loss": 3.8201745076336957, + "tokens_seen": 678100992 + }, + { + "epoch": 0.21, + "learning_rate": 0.0006420220661985959, + "loss": 1.5599, + "theoretical_loss": 3.8201370897692595, + "tokens_seen": 678166528 + }, + { + "epoch": 0.21, + "learning_rate": 0.0006420060180541625, + "loss": 1.5501, + "theoretical_loss": 3.820099676532955, + "tokens_seen": 678232064 + }, + { + "epoch": 0.21, + "objective/train/docs_used": 252282, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.4273723363876343, + "objective/train/theoretical_loss": 3.8200622679237615, + "objective/train/tokens_used": 698757600, + "theoretical_loss": 3.8200622679237615, + "tokens_seen": 678297600 + }, + { + "epoch": 0.21, + "learning_rate": 0.0006419899699097293, + "loss": 1.4157, + "theoretical_loss": 3.8200622679237615, + "tokens_seen": 678297600 + }, + { + "epoch": 0.21, + "learning_rate": 0.0006419739217652959, + "loss": 1.4909, + "theoretical_loss": 3.8200248639406613, + "tokens_seen": 678363136 + }, + { + "epoch": 0.21, + "learning_rate": 0.0006419578736208627, + "loss": 1.3912, + "theoretical_loss": 3.819987464582635, + "tokens_seen": 678428672 + }, + { + "epoch": 0.21, + "learning_rate": 0.0006419418254764293, + "loss": 1.566, + "theoretical_loss": 3.8199500698486633, + "tokens_seen": 678494208 + }, + { + "epoch": 0.21, + "learning_rate": 0.0006419257773319961, + "loss": 1.4996, + "theoretical_loss": 3.8199126797377287, + "tokens_seen": 678559744 + }, + { + "epoch": 0.21, + "learning_rate": 0.0006419097291875627, + "loss": 1.6089, + "theoretical_loss": 3.8198752942488134, + "tokens_seen": 678625280 + }, + { + "epoch": 0.21, + "learning_rate": 0.0006418936810431295, + "loss": 1.5567, + "theoretical_loss": 3.8198379133808995, + "tokens_seen": 678690816 + }, + { + "epoch": 0.21, + "learning_rate": 0.0006418776328986961, + "loss": 1.5496, + "theoretical_loss": 3.819800537132969, + "tokens_seen": 678756352 + }, + { + "epoch": 0.21, + "learning_rate": 0.0006418615847542629, + "loss": 1.5901, + "theoretical_loss": 3.8197631655040056, + "tokens_seen": 678821888 + }, + { + "epoch": 0.21, + "learning_rate": 0.0006418455366098295, + "loss": 1.566, + "theoretical_loss": 3.8197257984929927, + "tokens_seen": 678887424 + }, + { + "epoch": 0.21, + "learning_rate": 0.0006418294884653962, + "loss": 1.5155, + "theoretical_loss": 3.8196884360989136, + "tokens_seen": 678952960 + }, + { + "epoch": 0.21, + "learning_rate": 0.0006418134403209629, + "loss": 1.5636, + "theoretical_loss": 3.819651078320753, + "tokens_seen": 679018496 + }, + { + "epoch": 0.21, + "learning_rate": 0.0006417973921765296, + "loss": 1.5756, + "theoretical_loss": 3.819613725157494, + "tokens_seen": 679084032 + }, + { + "epoch": 0.21, + "learning_rate": 0.0006417813440320963, + "loss": 1.559, + "theoretical_loss": 3.819576376608122, + "tokens_seen": 679149568 + }, + { + "epoch": 0.21, + "learning_rate": 0.000641765295887663, + "loss": 1.6511, + "theoretical_loss": 3.8195390326716225, + "tokens_seen": 679215104 + }, + { + "epoch": 0.21, + "learning_rate": 0.0006417492477432297, + "loss": 1.3918, + "theoretical_loss": 3.8195016933469796, + "tokens_seen": 679280640 + }, + { + "epoch": 0.21, + "learning_rate": 0.0006417331995987964, + "loss": 1.4829, + "theoretical_loss": 3.819464358633179, + "tokens_seen": 679346176 + }, + { + "epoch": 0.21, + "learning_rate": 0.0006417171514543631, + "loss": 1.4644, + "theoretical_loss": 3.819427028529208, + "tokens_seen": 679411712 + }, + { + "epoch": 0.21, + "learning_rate": 0.0006417011033099298, + "loss": 1.4965, + "theoretical_loss": 3.8193897030340516, + "tokens_seen": 679477248 + }, + { + "epoch": 0.21, + "learning_rate": 0.0006416850551654964, + "loss": 1.6388, + "theoretical_loss": 3.819352382146696, + "tokens_seen": 679542784 + }, + { + "epoch": 0.21, + "learning_rate": 0.0006416690070210632, + "loss": 1.4806, + "theoretical_loss": 3.819315065866129, + "tokens_seen": 679608320 + }, + { + "epoch": 0.21, + "learning_rate": 0.0006416529588766298, + "loss": 1.5048, + "theoretical_loss": 3.8192777541913374, + "tokens_seen": 679673856 + }, + { + "epoch": 0.21, + "learning_rate": 0.0006416369107321966, + "loss": 1.4525, + "theoretical_loss": 3.819240447121309, + "tokens_seen": 679739392 + }, + { + "epoch": 0.21, + "learning_rate": 0.0006416208625877632, + "loss": 1.5363, + "theoretical_loss": 3.8192031446550314, + "tokens_seen": 679804928 + }, + { + "epoch": 0.21, + "learning_rate": 0.00064160481444333, + "loss": 1.4793, + "theoretical_loss": 3.819165846791493, + "tokens_seen": 679870464 + }, + { + "epoch": 0.21, + "learning_rate": 0.0006415887662988966, + "loss": 1.5418, + "theoretical_loss": 3.8191285535296813, + "tokens_seen": 679936000 + }, + { + "epoch": 0.21, + "learning_rate": 0.0006415727181544634, + "loss": 1.5096, + "theoretical_loss": 3.819091264868586, + "tokens_seen": 680001536 + }, + { + "epoch": 0.21, + "learning_rate": 0.00064155667001003, + "loss": 1.503, + "theoretical_loss": 3.8190539808071966, + "tokens_seen": 680067072 + }, + { + "epoch": 0.21, + "learning_rate": 0.0006415406218655968, + "loss": 1.429, + "theoretical_loss": 3.819016701344501, + "tokens_seen": 680132608 + }, + { + "epoch": 0.21, + "learning_rate": 0.0006415245737211634, + "loss": 1.5004, + "theoretical_loss": 3.8189794264794905, + "tokens_seen": 680198144 + }, + { + "epoch": 0.21, + "learning_rate": 0.0006415085255767302, + "loss": 1.5776, + "theoretical_loss": 3.8189421562111545, + "tokens_seen": 680263680 + }, + { + "epoch": 0.21, + "learning_rate": 0.0006414924774322968, + "loss": 1.4118, + "theoretical_loss": 3.8189048905384837, + "tokens_seen": 680329216 + }, + { + "epoch": 0.21, + "learning_rate": 0.0006414764292878636, + "loss": 1.4782, + "theoretical_loss": 3.8188676294604686, + "tokens_seen": 680394752 + }, + { + "epoch": 0.21, + "learning_rate": 0.0006414603811434302, + "loss": 1.5534, + "theoretical_loss": 3.8188303729761, + "tokens_seen": 680460288 + }, + { + "epoch": 0.21, + "learning_rate": 0.000641444332998997, + "loss": 1.4285, + "theoretical_loss": 3.818793121084369, + "tokens_seen": 680525824 + }, + { + "epoch": 0.21, + "learning_rate": 0.0006414282848545638, + "loss": 1.5255, + "theoretical_loss": 3.818755873784268, + "tokens_seen": 680591360 + }, + { + "epoch": 0.21, + "learning_rate": 0.0006414122367101304, + "loss": 1.5492, + "theoretical_loss": 3.8187186310747885, + "tokens_seen": 680656896 + }, + { + "epoch": 0.21, + "learning_rate": 0.0006413961885656972, + "loss": 1.5099, + "theoretical_loss": 3.8186813929549235, + "tokens_seen": 680722432 + }, + { + "epoch": 0.21, + "learning_rate": 0.0006413801404212638, + "loss": 1.4248, + "theoretical_loss": 3.818644159423665, + "tokens_seen": 680787968 + }, + { + "epoch": 0.21, + "learning_rate": 0.0006413640922768306, + "loss": 1.5297, + "theoretical_loss": 3.818606930480006, + "tokens_seen": 680853504 + }, + { + "epoch": 0.21, + "learning_rate": 0.0006413480441323972, + "loss": 1.5354, + "theoretical_loss": 3.8185697061229398, + "tokens_seen": 680919040 + }, + { + "epoch": 0.21, + "learning_rate": 0.000641331995987964, + "loss": 1.3892, + "theoretical_loss": 3.8185324863514594, + "tokens_seen": 680984576 + }, + { + "epoch": 0.21, + "learning_rate": 0.0006413159478435306, + "loss": 1.5926, + "theoretical_loss": 3.8184952711645597, + "tokens_seen": 681050112 + }, + { + "epoch": 0.21, + "learning_rate": 0.0006412998996990974, + "loss": 1.5228, + "theoretical_loss": 3.8184580605612344, + "tokens_seen": 681115648 + }, + { + "epoch": 0.21, + "learning_rate": 0.000641283851554664, + "loss": 1.4128, + "theoretical_loss": 3.818420854540478, + "tokens_seen": 681181184 + }, + { + "epoch": 0.21, + "learning_rate": 0.0006412678034102308, + "loss": 1.5374, + "theoretical_loss": 3.818383653101286, + "tokens_seen": 681246720 + }, + { + "epoch": 0.21, + "learning_rate": 0.0006412517552657974, + "loss": 1.4842, + "theoretical_loss": 3.8183464562426526, + "tokens_seen": 681312256 + }, + { + "epoch": 0.21, + "learning_rate": 0.0006412357071213642, + "loss": 1.5474, + "theoretical_loss": 3.8183092639635734, + "tokens_seen": 681377792 + }, + { + "epoch": 0.21, + "learning_rate": 0.0006412196589769308, + "loss": 1.5301, + "theoretical_loss": 3.8182720762630447, + "tokens_seen": 681443328 + }, + { + "epoch": 0.21, + "learning_rate": 0.0006412036108324976, + "loss": 1.5347, + "theoretical_loss": 3.8182348931400627, + "tokens_seen": 681508864 + }, + { + "epoch": 0.21, + "objective/train/docs_used": 253498, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.5124435424804688, + "objective/train/theoretical_loss": 3.818197714593623, + "objective/train/tokens_used": 702034400, + "theoretical_loss": 3.818197714593623, + "tokens_seen": 681574400 + }, + { + "epoch": 0.21, + "learning_rate": 0.0006411875626880642, + "loss": 1.537, + "theoretical_loss": 3.818197714593623, + "tokens_seen": 681574400 + }, + { + "epoch": 0.21, + "learning_rate": 0.000641171514543631, + "loss": 1.4781, + "theoretical_loss": 3.818160540622723, + "tokens_seen": 681639936 + }, + { + "epoch": 0.21, + "learning_rate": 0.0006411554663991976, + "loss": 1.613, + "theoretical_loss": 3.8181233712263594, + "tokens_seen": 681705472 + }, + { + "epoch": 0.21, + "learning_rate": 0.0006411394182547644, + "loss": 1.5274, + "theoretical_loss": 3.81808620640353, + "tokens_seen": 681771008 + }, + { + "epoch": 0.21, + "learning_rate": 0.000641123370110331, + "loss": 1.4979, + "theoretical_loss": 3.8180490461532317, + "tokens_seen": 681836544 + }, + { + "epoch": 0.21, + "learning_rate": 0.0006411073219658978, + "loss": 1.5675, + "theoretical_loss": 3.818011890474464, + "tokens_seen": 681902080 + }, + { + "epoch": 0.21, + "learning_rate": 0.0006410912738214644, + "loss": 1.4863, + "theoretical_loss": 3.817974739366224, + "tokens_seen": 681967616 + }, + { + "epoch": 0.21, + "learning_rate": 0.0006410752256770312, + "loss": 1.5349, + "theoretical_loss": 3.8179375928275103, + "tokens_seen": 682033152 + }, + { + "epoch": 0.21, + "learning_rate": 0.0006410591775325978, + "loss": 1.5702, + "theoretical_loss": 3.817900450857322, + "tokens_seen": 682098688 + }, + { + "epoch": 0.21, + "learning_rate": 0.0006410431293881646, + "loss": 1.4639, + "theoretical_loss": 3.817863313454659, + "tokens_seen": 682164224 + }, + { + "epoch": 0.21, + "learning_rate": 0.0006410270812437312, + "loss": 1.5734, + "theoretical_loss": 3.8178261806185207, + "tokens_seen": 682229760 + }, + { + "epoch": 0.21, + "learning_rate": 0.000641011033099298, + "loss": 1.5778, + "theoretical_loss": 3.817789052347906, + "tokens_seen": 682295296 + }, + { + "epoch": 0.21, + "learning_rate": 0.0006409949849548646, + "loss": 1.5714, + "theoretical_loss": 3.8177519286418162, + "tokens_seen": 682360832 + }, + { + "epoch": 0.21, + "learning_rate": 0.0006409789368104314, + "loss": 1.52, + "theoretical_loss": 3.817714809499252, + "tokens_seen": 682426368 + }, + { + "epoch": 0.21, + "learning_rate": 0.000640962888665998, + "loss": 1.5817, + "theoretical_loss": 3.8176776949192135, + "tokens_seen": 682491904 + }, + { + "epoch": 0.21, + "learning_rate": 0.0006409468405215648, + "loss": 1.5272, + "theoretical_loss": 3.8176405849007016, + "tokens_seen": 682557440 + }, + { + "epoch": 0.21, + "learning_rate": 0.0006409307923771314, + "loss": 1.4351, + "theoretical_loss": 3.8176034794427185, + "tokens_seen": 682622976 + }, + { + "epoch": 0.21, + "learning_rate": 0.0006409147442326982, + "loss": 1.541, + "theoretical_loss": 3.8175663785442655, + "tokens_seen": 682688512 + }, + { + "epoch": 0.21, + "learning_rate": 0.0006408986960882648, + "loss": 1.5881, + "theoretical_loss": 3.8175292822043456, + "tokens_seen": 682754048 + }, + { + "epoch": 0.21, + "learning_rate": 0.0006408826479438316, + "loss": 1.587, + "theoretical_loss": 3.81749219042196, + "tokens_seen": 682819584 + }, + { + "epoch": 0.21, + "learning_rate": 0.0006408665997993982, + "loss": 1.3936, + "theoretical_loss": 3.8174551031961124, + "tokens_seen": 682885120 + }, + { + "epoch": 0.21, + "learning_rate": 0.000640850551654965, + "loss": 1.5012, + "theoretical_loss": 3.817418020525805, + "tokens_seen": 682950656 + }, + { + "epoch": 0.21, + "learning_rate": 0.0006408345035105316, + "loss": 1.4664, + "theoretical_loss": 3.817380942410042, + "tokens_seen": 683016192 + }, + { + "epoch": 0.21, + "learning_rate": 0.0006408184553660984, + "loss": 1.4486, + "theoretical_loss": 3.8173438688478267, + "tokens_seen": 683081728 + }, + { + "epoch": 0.21, + "learning_rate": 0.000640802407221665, + "loss": 1.4864, + "theoretical_loss": 3.817306799838163, + "tokens_seen": 683147264 + }, + { + "epoch": 0.21, + "learning_rate": 0.0006407863590772318, + "loss": 1.532, + "theoretical_loss": 3.817269735380055, + "tokens_seen": 683212800 + }, + { + "epoch": 0.21, + "learning_rate": 0.0006407703109327984, + "loss": 1.495, + "theoretical_loss": 3.817232675472508, + "tokens_seen": 683278336 + }, + { + "epoch": 0.21, + "learning_rate": 0.0006407542627883652, + "loss": 1.5444, + "theoretical_loss": 3.817195620114526, + "tokens_seen": 683343872 + }, + { + "epoch": 0.21, + "learning_rate": 0.0006407382146439318, + "loss": 1.5725, + "theoretical_loss": 3.817158569305115, + "tokens_seen": 683409408 + }, + { + "epoch": 0.21, + "learning_rate": 0.0006407221664994986, + "loss": 1.4653, + "theoretical_loss": 3.817121523043281, + "tokens_seen": 683474944 + }, + { + "epoch": 0.21, + "learning_rate": 0.0006407061183550652, + "loss": 1.6843, + "theoretical_loss": 3.8170844813280285, + "tokens_seen": 683540480 + }, + { + "epoch": 0.21, + "learning_rate": 0.0006407061183550652, + "loss": 1.5386, + "theoretical_loss": 3.817047444158364, + "tokens_seen": 683606016 + }, + { + "epoch": 0.21, + "learning_rate": 0.000640690070210632, + "loss": 1.4462, + "theoretical_loss": 3.8170104115332943, + "tokens_seen": 683671552 + }, + { + "epoch": 0.21, + "learning_rate": 0.0006406740220661986, + "loss": 1.5116, + "theoretical_loss": 3.816973383451827, + "tokens_seen": 683737088 + }, + { + "epoch": 0.21, + "learning_rate": 0.0006406579739217654, + "loss": 1.5098, + "theoretical_loss": 3.816936359912968, + "tokens_seen": 683802624 + }, + { + "epoch": 0.21, + "learning_rate": 0.000640641925777332, + "loss": 1.5734, + "theoretical_loss": 3.8168993409157252, + "tokens_seen": 683868160 + }, + { + "epoch": 0.21, + "learning_rate": 0.0006406258776328988, + "loss": 1.4976, + "theoretical_loss": 3.8168623264591064, + "tokens_seen": 683933696 + }, + { + "epoch": 0.21, + "learning_rate": 0.0006406098294884654, + "loss": 1.5115, + "theoretical_loss": 3.8168253165421198, + "tokens_seen": 683999232 + }, + { + "epoch": 0.21, + "learning_rate": 0.0006405937813440322, + "loss": 1.626, + "theoretical_loss": 3.816788311163773, + "tokens_seen": 684064768 + }, + { + "epoch": 0.21, + "learning_rate": 0.0006405777331995988, + "loss": 1.4991, + "theoretical_loss": 3.8167513103230757, + "tokens_seen": 684130304 + }, + { + "epoch": 0.21, + "learning_rate": 0.0006405616850551656, + "loss": 1.4641, + "theoretical_loss": 3.816714314019036, + "tokens_seen": 684195840 + }, + { + "epoch": 0.21, + "learning_rate": 0.0006405456369107322, + "loss": 1.6354, + "theoretical_loss": 3.816677322250664, + "tokens_seen": 684261376 + }, + { + "epoch": 0.21, + "learning_rate": 0.000640529588766299, + "loss": 1.391, + "theoretical_loss": 3.8166403350169684, + "tokens_seen": 684326912 + }, + { + "epoch": 0.21, + "learning_rate": 0.0006405135406218656, + "loss": 1.4421, + "theoretical_loss": 3.81660335231696, + "tokens_seen": 684392448 + }, + { + "epoch": 0.21, + "learning_rate": 0.0006404974924774324, + "loss": 1.5355, + "theoretical_loss": 3.816566374149648, + "tokens_seen": 684457984 + }, + { + "epoch": 0.21, + "learning_rate": 0.000640481444332999, + "loss": 1.4729, + "theoretical_loss": 3.816529400514044, + "tokens_seen": 684523520 + }, + { + "epoch": 0.21, + "learning_rate": 0.0006404653961885658, + "loss": 1.506, + "theoretical_loss": 3.8164924314091584, + "tokens_seen": 684589056 + }, + { + "epoch": 0.21, + "learning_rate": 0.0006404493480441324, + "loss": 1.4402, + "theoretical_loss": 3.816455466834003, + "tokens_seen": 684654592 + }, + { + "epoch": 0.21, + "learning_rate": 0.0006404332998996991, + "loss": 1.5073, + "theoretical_loss": 3.8164185067875875, + "tokens_seen": 684720128 + }, + { + "epoch": 0.21, + "learning_rate": 0.0006404172517552658, + "loss": 1.5192, + "theoretical_loss": 3.8163815512689254, + "tokens_seen": 684785664 + }, + { + "epoch": 0.21, + "objective/train/docs_used": 254674, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.485886812210083, + "objective/train/theoretical_loss": 3.816344600277029, + "objective/train/tokens_used": 705311200, + "theoretical_loss": 3.816344600277029, + "tokens_seen": 684851200 + }, + { + "epoch": 0.21, + "learning_rate": 0.0006404012036108325, + "loss": 1.4925, + "theoretical_loss": 3.816344600277029, + "tokens_seen": 684851200 + }, + { + "epoch": 0.21, + "learning_rate": 0.0006403851554663992, + "loss": 1.5328, + "theoretical_loss": 3.8163076538109086, + "tokens_seen": 684916736 + }, + { + "epoch": 0.21, + "learning_rate": 0.0006403691073219659, + "loss": 1.5459, + "theoretical_loss": 3.816270711869579, + "tokens_seen": 684982272 + }, + { + "epoch": 0.21, + "learning_rate": 0.0006403530591775326, + "loss": 1.4464, + "theoretical_loss": 3.8162337744520523, + "tokens_seen": 685047808 + }, + { + "epoch": 0.21, + "learning_rate": 0.0006403370110330993, + "loss": 1.4581, + "theoretical_loss": 3.816196841557342, + "tokens_seen": 685113344 + }, + { + "epoch": 0.21, + "learning_rate": 0.000640320962888666, + "loss": 1.4609, + "theoretical_loss": 3.816159913184462, + "tokens_seen": 685178880 + }, + { + "epoch": 0.21, + "learning_rate": 0.0006403049147442327, + "loss": 1.5254, + "theoretical_loss": 3.8161229893324258, + "tokens_seen": 685244416 + }, + { + "epoch": 0.21, + "learning_rate": 0.0006402888665997993, + "loss": 1.5624, + "theoretical_loss": 3.8160860700002477, + "tokens_seen": 685309952 + }, + { + "epoch": 0.21, + "learning_rate": 0.0006402728184553661, + "loss": 1.482, + "theoretical_loss": 3.816049155186943, + "tokens_seen": 685375488 + }, + { + "epoch": 0.21, + "learning_rate": 0.0006402567703109327, + "loss": 1.5214, + "theoretical_loss": 3.8160122448915255, + "tokens_seen": 685441024 + }, + { + "epoch": 0.21, + "learning_rate": 0.0006402407221664995, + "loss": 1.483, + "theoretical_loss": 3.815975339113011, + "tokens_seen": 685506560 + }, + { + "epoch": 0.21, + "learning_rate": 0.0006402246740220662, + "loss": 1.6073, + "theoretical_loss": 3.8159384378504146, + "tokens_seen": 685572096 + }, + { + "epoch": 0.21, + "learning_rate": 0.0006402086258776329, + "loss": 1.445, + "theoretical_loss": 3.8159015411027526, + "tokens_seen": 685637632 + }, + { + "epoch": 0.21, + "learning_rate": 0.0006401925777331996, + "loss": 1.5072, + "theoretical_loss": 3.815864648869041, + "tokens_seen": 685703168 + }, + { + "epoch": 0.21, + "learning_rate": 0.0006401765295887663, + "loss": 1.5409, + "theoretical_loss": 3.815827761148296, + "tokens_seen": 685768704 + }, + { + "epoch": 0.21, + "learning_rate": 0.000640160481444333, + "loss": 1.4681, + "theoretical_loss": 3.8157908779395346, + "tokens_seen": 685834240 + }, + { + "epoch": 0.21, + "learning_rate": 0.0006401444332998997, + "loss": 1.5005, + "theoretical_loss": 3.8157539992417737, + "tokens_seen": 685899776 + }, + { + "epoch": 0.21, + "learning_rate": 0.0006401283851554664, + "loss": 1.6516, + "theoretical_loss": 3.8157171250540305, + "tokens_seen": 685965312 + }, + { + "epoch": 0.21, + "learning_rate": 0.0006401123370110331, + "loss": 1.565, + "theoretical_loss": 3.815680255375323, + "tokens_seen": 686030848 + }, + { + "epoch": 0.21, + "learning_rate": 0.0006400962888665998, + "loss": 1.3871, + "theoretical_loss": 3.815643390204669, + "tokens_seen": 686096384 + }, + { + "epoch": 0.21, + "learning_rate": 0.0006400802407221665, + "loss": 1.5103, + "theoretical_loss": 3.8156065295410864, + "tokens_seen": 686161920 + }, + { + "epoch": 0.21, + "learning_rate": 0.0006400641925777332, + "loss": 1.5127, + "theoretical_loss": 3.8155696733835947, + "tokens_seen": 686227456 + }, + { + "epoch": 0.21, + "learning_rate": 0.0006400481444332999, + "loss": 1.5763, + "theoretical_loss": 3.8155328217312117, + "tokens_seen": 686292992 + }, + { + "epoch": 0.21, + "learning_rate": 0.0006400320962888667, + "loss": 1.5527, + "theoretical_loss": 3.815495974582957, + "tokens_seen": 686358528 + }, + { + "epoch": 0.21, + "learning_rate": 0.0006400160481444333, + "loss": 1.4112, + "theoretical_loss": 3.8154591319378506, + "tokens_seen": 686424064 + }, + { + "epoch": 0.21, + "learning_rate": 0.00064, + "loss": 1.5715, + "theoretical_loss": 3.815422293794912, + "tokens_seen": 686489600 + }, + { + "epoch": 0.21, + "learning_rate": 0.0006399839518555667, + "loss": 1.4903, + "theoretical_loss": 3.8153854601531605, + "tokens_seen": 686555136 + }, + { + "epoch": 0.21, + "learning_rate": 0.0006399679037111335, + "loss": 1.5147, + "theoretical_loss": 3.815348631011618, + "tokens_seen": 686620672 + }, + { + "epoch": 0.21, + "learning_rate": 0.0006399518555667001, + "loss": 1.447, + "theoretical_loss": 3.8153118063693037, + "tokens_seen": 686686208 + }, + { + "epoch": 0.21, + "learning_rate": 0.0006399358074222669, + "loss": 1.4544, + "theoretical_loss": 3.8152749862252398, + "tokens_seen": 686751744 + }, + { + "epoch": 0.21, + "learning_rate": 0.0006399197592778335, + "loss": 1.4856, + "theoretical_loss": 3.8152381705784473, + "tokens_seen": 686817280 + }, + { + "epoch": 0.21, + "learning_rate": 0.0006399037111334003, + "loss": 1.4967, + "theoretical_loss": 3.815201359427948, + "tokens_seen": 686882816 + }, + { + "epoch": 0.21, + "learning_rate": 0.0006398876629889669, + "loss": 1.5481, + "theoretical_loss": 3.815164552772763, + "tokens_seen": 686948352 + }, + { + "epoch": 0.21, + "learning_rate": 0.0006398716148445337, + "loss": 1.4544, + "theoretical_loss": 3.8151277506119152, + "tokens_seen": 687013888 + }, + { + "epoch": 0.21, + "learning_rate": 0.0006398555667001003, + "loss": 1.454, + "theoretical_loss": 3.8150909529444275, + "tokens_seen": 687079424 + }, + { + "epoch": 0.21, + "learning_rate": 0.0006398395185556671, + "loss": 1.4805, + "theoretical_loss": 3.815054159769322, + "tokens_seen": 687144960 + }, + { + "epoch": 0.21, + "learning_rate": 0.0006398234704112337, + "loss": 1.5892, + "theoretical_loss": 3.8150173710856228, + "tokens_seen": 687210496 + }, + { + "epoch": 0.21, + "learning_rate": 0.0006398074222668005, + "loss": 1.4452, + "theoretical_loss": 3.8149805868923528, + "tokens_seen": 687276032 + }, + { + "epoch": 0.21, + "learning_rate": 0.0006397913741223671, + "loss": 1.4321, + "theoretical_loss": 3.814943807188536, + "tokens_seen": 687341568 + }, + { + "epoch": 0.21, + "learning_rate": 0.0006397753259779339, + "loss": 1.5247, + "theoretical_loss": 3.8149070319731955, + "tokens_seen": 687407104 + }, + { + "epoch": 0.21, + "learning_rate": 0.0006397592778335005, + "loss": 1.6585, + "theoretical_loss": 3.8148702612453573, + "tokens_seen": 687472640 + }, + { + "epoch": 0.21, + "learning_rate": 0.0006397432296890673, + "loss": 1.5764, + "theoretical_loss": 3.814833495004045, + "tokens_seen": 687538176 + }, + { + "epoch": 0.21, + "learning_rate": 0.0006397271815446339, + "loss": 1.4579, + "theoretical_loss": 3.8147967332482846, + "tokens_seen": 687603712 + }, + { + "epoch": 0.21, + "learning_rate": 0.0006397111334002007, + "loss": 1.6567, + "theoretical_loss": 3.8147599759771, + "tokens_seen": 687669248 + }, + { + "epoch": 0.21, + "learning_rate": 0.0006396950852557673, + "loss": 1.509, + "theoretical_loss": 3.814723223189518, + "tokens_seen": 687734784 + }, + { + "epoch": 0.21, + "learning_rate": 0.0006396790371113341, + "loss": 1.4724, + "theoretical_loss": 3.8146864748845637, + "tokens_seen": 687800320 + }, + { + "epoch": 0.21, + "learning_rate": 0.0006396629889669007, + "loss": 1.3749, + "theoretical_loss": 3.814649731061264, + "tokens_seen": 687865856 + }, + { + "epoch": 0.21, + "learning_rate": 0.0006396469408224675, + "loss": 1.4926, + "theoretical_loss": 3.8146129917186453, + "tokens_seen": 687931392 + }, + { + "epoch": 0.21, + "learning_rate": 0.0006396308926780341, + "loss": 1.4345, + "theoretical_loss": 3.8145762568557338, + "tokens_seen": 687996928 + }, + { + "epoch": 0.21, + "learning_rate": 0.0006396148445336009, + "loss": 1.5574, + "theoretical_loss": 3.814539526471558, + "tokens_seen": 688062464 + }, + { + "epoch": 0.21, + "objective/train/docs_used": 255827, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.4748759269714355, + "objective/train/theoretical_loss": 3.8145028005651436, + "objective/train/tokens_used": 708588000, + "theoretical_loss": 3.8145028005651436, + "tokens_seen": 688128000 + }, + { + "epoch": 0.21, + "learning_rate": 0.0006395987963891675, + "loss": 1.5136, + "theoretical_loss": 3.8145028005651436, + "tokens_seen": 688128000 + }, + { + "epoch": 0.21, + "learning_rate": 0.0006395827482447343, + "loss": 1.4987, + "theoretical_loss": 3.8144660791355194, + "tokens_seen": 688193536 + }, + { + "epoch": 0.21, + "learning_rate": 0.0006395667001003009, + "loss": 1.5396, + "theoretical_loss": 3.8144293621817136, + "tokens_seen": 688259072 + }, + { + "epoch": 0.21, + "learning_rate": 0.0006395506519558677, + "loss": 1.5171, + "theoretical_loss": 3.8143926497027536, + "tokens_seen": 688324608 + }, + { + "epoch": 0.21, + "learning_rate": 0.0006395346038114343, + "loss": 1.5098, + "theoretical_loss": 3.814355941697669, + "tokens_seen": 688390144 + }, + { + "epoch": 0.21, + "learning_rate": 0.0006395185556670011, + "loss": 1.4863, + "theoretical_loss": 3.814319238165489, + "tokens_seen": 688455680 + }, + { + "epoch": 0.21, + "learning_rate": 0.0006395025075225677, + "loss": 1.5702, + "theoretical_loss": 3.8142825391052417, + "tokens_seen": 688521216 + }, + { + "epoch": 0.21, + "learning_rate": 0.0006394864593781345, + "loss": 1.5053, + "theoretical_loss": 3.8142458445159573, + "tokens_seen": 688586752 + }, + { + "epoch": 0.21, + "learning_rate": 0.0006394704112337011, + "loss": 1.5026, + "theoretical_loss": 3.8142091543966656, + "tokens_seen": 688652288 + }, + { + "epoch": 0.21, + "learning_rate": 0.0006394543630892679, + "loss": 1.4132, + "theoretical_loss": 3.814172468746397, + "tokens_seen": 688717824 + }, + { + "epoch": 0.21, + "learning_rate": 0.0006394383149448345, + "loss": 1.5664, + "theoretical_loss": 3.8141357875641813, + "tokens_seen": 688783360 + }, + { + "epoch": 0.21, + "learning_rate": 0.0006394222668004013, + "loss": 1.47, + "theoretical_loss": 3.81409911084905, + "tokens_seen": 688848896 + }, + { + "epoch": 0.21, + "learning_rate": 0.0006394062186559679, + "loss": 1.5319, + "theoretical_loss": 3.8140624386000344, + "tokens_seen": 688914432 + }, + { + "epoch": 0.21, + "learning_rate": 0.0006393901705115347, + "loss": 1.4734, + "theoretical_loss": 3.8140257708161647, + "tokens_seen": 688979968 + }, + { + "epoch": 0.21, + "learning_rate": 0.0006393741223671013, + "loss": 1.5529, + "theoretical_loss": 3.8139891074964734, + "tokens_seen": 689045504 + }, + { + "epoch": 0.21, + "learning_rate": 0.0006393580742226681, + "loss": 1.5087, + "theoretical_loss": 3.8139524486399927, + "tokens_seen": 689111040 + }, + { + "epoch": 0.21, + "learning_rate": 0.0006393420260782347, + "loss": 1.5136, + "theoretical_loss": 3.813915794245754, + "tokens_seen": 689176576 + }, + { + "epoch": 0.21, + "learning_rate": 0.0006393259779338015, + "loss": 1.3995, + "theoretical_loss": 3.813879144312791, + "tokens_seen": 689242112 + }, + { + "epoch": 0.21, + "learning_rate": 0.0006393099297893681, + "loss": 1.5313, + "theoretical_loss": 3.8138424988401365, + "tokens_seen": 689307648 + }, + { + "epoch": 0.21, + "learning_rate": 0.0006392938816449349, + "loss": 1.6425, + "theoretical_loss": 3.8138058578268224, + "tokens_seen": 689373184 + }, + { + "epoch": 0.21, + "learning_rate": 0.0006392778335005015, + "loss": 1.4443, + "theoretical_loss": 3.8137692212718832, + "tokens_seen": 689438720 + }, + { + "epoch": 0.21, + "learning_rate": 0.0006392617853560683, + "loss": 1.5284, + "theoretical_loss": 3.8137325891743528, + "tokens_seen": 689504256 + }, + { + "epoch": 0.21, + "learning_rate": 0.0006392457372116349, + "loss": 1.5181, + "theoretical_loss": 3.8136959615332646, + "tokens_seen": 689569792 + }, + { + "epoch": 0.21, + "learning_rate": 0.0006392296890672017, + "loss": 1.483, + "theoretical_loss": 3.813659338347654, + "tokens_seen": 689635328 + }, + { + "epoch": 0.21, + "learning_rate": 0.0006392136409227683, + "loss": 1.557, + "theoretical_loss": 3.8136227196165544, + "tokens_seen": 689700864 + }, + { + "epoch": 0.21, + "learning_rate": 0.0006391975927783351, + "loss": 1.5252, + "theoretical_loss": 3.813586105339002, + "tokens_seen": 689766400 + }, + { + "epoch": 0.21, + "learning_rate": 0.0006391815446339017, + "loss": 1.5469, + "theoretical_loss": 3.813549495514031, + "tokens_seen": 689831936 + }, + { + "epoch": 0.21, + "learning_rate": 0.0006391654964894685, + "loss": 1.4918, + "theoretical_loss": 3.8135128901406783, + "tokens_seen": 689897472 + }, + { + "epoch": 0.21, + "learning_rate": 0.0006391494483450351, + "loss": 1.5198, + "theoretical_loss": 3.813476289217979, + "tokens_seen": 689963008 + }, + { + "epoch": 0.21, + "learning_rate": 0.0006391334002006019, + "loss": 1.4248, + "theoretical_loss": 3.813439692744969, + "tokens_seen": 690028544 + }, + { + "epoch": 0.21, + "learning_rate": 0.0006391173520561685, + "loss": 1.6104, + "theoretical_loss": 3.813403100720686, + "tokens_seen": 690094080 + }, + { + "epoch": 0.21, + "learning_rate": 0.0006391013039117353, + "loss": 1.4968, + "theoretical_loss": 3.813366513144165, + "tokens_seen": 690159616 + }, + { + "epoch": 0.21, + "learning_rate": 0.0006390852557673019, + "loss": 1.4665, + "theoretical_loss": 3.8133299300144445, + "tokens_seen": 690225152 + }, + { + "epoch": 0.21, + "learning_rate": 0.0006390692076228687, + "loss": 1.5085, + "theoretical_loss": 3.813293351330562, + "tokens_seen": 690290688 + }, + { + "epoch": 0.21, + "learning_rate": 0.0006390531594784353, + "loss": 1.561, + "theoretical_loss": 3.8132567770915546, + "tokens_seen": 690356224 + }, + { + "epoch": 0.21, + "learning_rate": 0.0006390371113340021, + "loss": 1.5879, + "theoretical_loss": 3.8132202072964603, + "tokens_seen": 690421760 + }, + { + "epoch": 0.21, + "learning_rate": 0.0006390210631895687, + "loss": 1.4745, + "theoretical_loss": 3.8131836419443172, + "tokens_seen": 690487296 + }, + { + "epoch": 0.21, + "learning_rate": 0.0006390050150451354, + "loss": 1.4542, + "theoretical_loss": 3.8131470810341646, + "tokens_seen": 690552832 + }, + { + "epoch": 0.21, + "learning_rate": 0.0006389889669007021, + "loss": 1.5354, + "theoretical_loss": 3.8131105245650403, + "tokens_seen": 690618368 + }, + { + "epoch": 0.21, + "learning_rate": 0.0006389729187562688, + "loss": 1.5506, + "theoretical_loss": 3.8130739725359852, + "tokens_seen": 690683904 + }, + { + "epoch": 0.21, + "learning_rate": 0.0006389568706118355, + "loss": 1.5921, + "theoretical_loss": 3.8130374249460375, + "tokens_seen": 690749440 + }, + { + "epoch": 0.21, + "learning_rate": 0.0006389408224674022, + "loss": 1.5515, + "theoretical_loss": 3.8130008817942374, + "tokens_seen": 690814976 + }, + { + "epoch": 0.21, + "learning_rate": 0.0006389247743229689, + "loss": 1.4454, + "theoretical_loss": 3.8129643430796243, + "tokens_seen": 690880512 + }, + { + "epoch": 0.21, + "learning_rate": 0.0006389087261785356, + "loss": 1.3995, + "theoretical_loss": 3.8129278088012395, + "tokens_seen": 690946048 + }, + { + "epoch": 0.21, + "learning_rate": 0.0006388926780341023, + "loss": 1.4533, + "theoretical_loss": 3.8128912789581237, + "tokens_seen": 691011584 + }, + { + "epoch": 0.21, + "learning_rate": 0.000638876629889669, + "loss": 1.6152, + "theoretical_loss": 3.8128547535493174, + "tokens_seen": 691077120 + }, + { + "epoch": 0.21, + "learning_rate": 0.0006388605817452357, + "loss": 1.5916, + "theoretical_loss": 3.8128182325738624, + "tokens_seen": 691142656 + }, + { + "epoch": 0.21, + "learning_rate": 0.0006388445336008024, + "loss": 1.6286, + "theoretical_loss": 3.8127817160307993, + "tokens_seen": 691208192 + }, + { + "epoch": 0.21, + "learning_rate": 0.0006388284854563691, + "loss": 1.4111, + "theoretical_loss": 3.812745203919171, + "tokens_seen": 691273728 + }, + { + "epoch": 0.21, + "learning_rate": 0.0006388124373119358, + "loss": 1.5036, + "theoretical_loss": 3.8127086962380194, + "tokens_seen": 691339264 + }, + { + "epoch": 0.21, + "objective/train/docs_used": 257040, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.4906601905822754, + "objective/train/theoretical_loss": 3.812672192986387, + "objective/train/tokens_used": 711864800, + "theoretical_loss": 3.812672192986387, + "tokens_seen": 691404800 + }, + { + "epoch": 0.21, + "learning_rate": 0.0006387963891675025, + "loss": 1.5135, + "theoretical_loss": 3.812672192986387, + "tokens_seen": 691404800 + }, + { + "epoch": 0.21, + "learning_rate": 0.0006387803410230692, + "loss": 1.3857, + "theoretical_loss": 3.8126356941633164, + "tokens_seen": 691470336 + }, + { + "epoch": 0.21, + "learning_rate": 0.000638764292878636, + "loss": 1.503, + "theoretical_loss": 3.8125991997678503, + "tokens_seen": 691535872 + }, + { + "epoch": 0.21, + "learning_rate": 0.0006387482447342026, + "loss": 1.4217, + "theoretical_loss": 3.8125627097990327, + "tokens_seen": 691601408 + }, + { + "epoch": 0.21, + "learning_rate": 0.0006387321965897693, + "loss": 1.4702, + "theoretical_loss": 3.8125262242559073, + "tokens_seen": 691666944 + }, + { + "epoch": 0.21, + "learning_rate": 0.000638716148445336, + "loss": 1.5154, + "theoretical_loss": 3.812489743137518, + "tokens_seen": 691732480 + }, + { + "epoch": 0.21, + "learning_rate": 0.0006387001003009027, + "loss": 1.4252, + "theoretical_loss": 3.8124532664429087, + "tokens_seen": 691798016 + }, + { + "epoch": 0.21, + "learning_rate": 0.0006386840521564694, + "loss": 1.403, + "theoretical_loss": 3.8124167941711242, + "tokens_seen": 691863552 + }, + { + "epoch": 0.21, + "learning_rate": 0.0006386680040120362, + "loss": 1.5472, + "theoretical_loss": 3.8123803263212097, + "tokens_seen": 691929088 + }, + { + "epoch": 0.21, + "learning_rate": 0.0006386519558676028, + "loss": 1.563, + "theoretical_loss": 3.8123438628922095, + "tokens_seen": 691994624 + }, + { + "epoch": 0.21, + "learning_rate": 0.0006386359077231696, + "loss": 1.4624, + "theoretical_loss": 3.8123074038831697, + "tokens_seen": 692060160 + }, + { + "epoch": 0.21, + "learning_rate": 0.0006386198595787362, + "loss": 1.5551, + "theoretical_loss": 3.8122709492931364, + "tokens_seen": 692125696 + }, + { + "epoch": 0.21, + "learning_rate": 0.000638603811434303, + "loss": 1.454, + "theoretical_loss": 3.8122344991211543, + "tokens_seen": 692191232 + }, + { + "epoch": 0.21, + "learning_rate": 0.0006385877632898696, + "loss": 1.5203, + "theoretical_loss": 3.8121980533662714, + "tokens_seen": 692256768 + }, + { + "epoch": 0.21, + "learning_rate": 0.0006385717151454364, + "loss": 1.3896, + "theoretical_loss": 3.8121616120275332, + "tokens_seen": 692322304 + }, + { + "epoch": 0.21, + "learning_rate": 0.000638555667001003, + "loss": 1.5584, + "theoretical_loss": 3.812125175103987, + "tokens_seen": 692387840 + }, + { + "epoch": 0.21, + "learning_rate": 0.0006385396188565698, + "loss": 1.4771, + "theoretical_loss": 3.81208874259468, + "tokens_seen": 692453376 + }, + { + "epoch": 0.21, + "learning_rate": 0.0006385235707121364, + "loss": 1.5014, + "theoretical_loss": 3.8120523144986596, + "tokens_seen": 692518912 + }, + { + "epoch": 0.21, + "learning_rate": 0.0006385075225677032, + "loss": 1.5202, + "theoretical_loss": 3.8120158908149735, + "tokens_seen": 692584448 + }, + { + "epoch": 0.21, + "learning_rate": 0.0006384914744232698, + "loss": 1.6181, + "theoretical_loss": 3.8119794715426707, + "tokens_seen": 692649984 + }, + { + "epoch": 0.21, + "learning_rate": 0.0006384754262788366, + "loss": 1.5204, + "theoretical_loss": 3.8119430566807986, + "tokens_seen": 692715520 + }, + { + "epoch": 0.21, + "learning_rate": 0.0006384593781344032, + "loss": 1.5415, + "theoretical_loss": 3.8119066462284064, + "tokens_seen": 692781056 + }, + { + "epoch": 0.21, + "learning_rate": 0.00063844332998997, + "loss": 1.4873, + "theoretical_loss": 3.8118702401845423, + "tokens_seen": 692846592 + }, + { + "epoch": 0.21, + "learning_rate": 0.0006384272818455366, + "loss": 1.4905, + "theoretical_loss": 3.8118338385482575, + "tokens_seen": 692912128 + }, + { + "epoch": 0.21, + "learning_rate": 0.0006384112337011034, + "loss": 1.5518, + "theoretical_loss": 3.8117974413185993, + "tokens_seen": 692977664 + }, + { + "epoch": 0.21, + "learning_rate": 0.00063839518555667, + "loss": 1.5023, + "theoretical_loss": 3.8117610484946187, + "tokens_seen": 693043200 + }, + { + "epoch": 0.21, + "learning_rate": 0.0006383791374122368, + "loss": 1.596, + "theoretical_loss": 3.8117246600753667, + "tokens_seen": 693108736 + }, + { + "epoch": 0.21, + "learning_rate": 0.0006383630892678034, + "loss": 1.6358, + "theoretical_loss": 3.8116882760598925, + "tokens_seen": 693174272 + }, + { + "epoch": 0.21, + "learning_rate": 0.0006383470411233702, + "loss": 1.4968, + "theoretical_loss": 3.811651896447247, + "tokens_seen": 693239808 + }, + { + "epoch": 0.21, + "learning_rate": 0.0006383309929789368, + "loss": 1.5207, + "theoretical_loss": 3.8116155212364817, + "tokens_seen": 693305344 + }, + { + "epoch": 0.21, + "learning_rate": 0.0006383149448345036, + "loss": 1.465, + "theoretical_loss": 3.811579150426648, + "tokens_seen": 693370880 + }, + { + "epoch": 0.21, + "learning_rate": 0.0006382988966900702, + "loss": 1.5533, + "theoretical_loss": 3.811542784016797, + "tokens_seen": 693436416 + }, + { + "epoch": 0.21, + "learning_rate": 0.000638282848545637, + "loss": 1.5675, + "theoretical_loss": 3.811506422005981, + "tokens_seen": 693501952 + }, + { + "epoch": 0.21, + "learning_rate": 0.0006382668004012036, + "loss": 1.5152, + "theoretical_loss": 3.8114700643932524, + "tokens_seen": 693567488 + }, + { + "epoch": 0.21, + "learning_rate": 0.0006382507522567704, + "loss": 1.4655, + "theoretical_loss": 3.8114337111776635, + "tokens_seen": 693633024 + }, + { + "epoch": 0.21, + "learning_rate": 0.000638234704112337, + "loss": 1.554, + "theoretical_loss": 3.8113973623582673, + "tokens_seen": 693698560 + }, + { + "epoch": 0.21, + "learning_rate": 0.0006382186559679038, + "loss": 1.5163, + "theoretical_loss": 3.8113610179341166, + "tokens_seen": 693764096 + }, + { + "epoch": 0.21, + "learning_rate": 0.0006382026078234704, + "loss": 1.5632, + "theoretical_loss": 3.811324677904265, + "tokens_seen": 693829632 + }, + { + "epoch": 0.21, + "learning_rate": 0.0006381865596790372, + "loss": 1.5157, + "theoretical_loss": 3.811288342267767, + "tokens_seen": 693895168 + }, + { + "epoch": 0.21, + "learning_rate": 0.0006381705115346038, + "loss": 1.5486, + "theoretical_loss": 3.811252011023675, + "tokens_seen": 693960704 + }, + { + "epoch": 0.21, + "learning_rate": 0.0006381544633901706, + "loss": 1.4224, + "theoretical_loss": 3.811215684171044, + "tokens_seen": 694026240 + }, + { + "epoch": 0.21, + "learning_rate": 0.0006381384152457372, + "loss": 1.5936, + "theoretical_loss": 3.8111793617089296, + "tokens_seen": 694091776 + }, + { + "epoch": 0.21, + "learning_rate": 0.000638122367101304, + "loss": 1.5411, + "theoretical_loss": 3.8111430436363856, + "tokens_seen": 694157312 + }, + { + "epoch": 0.21, + "learning_rate": 0.0006381063189568706, + "loss": 1.6866, + "theoretical_loss": 3.8111067299524666, + "tokens_seen": 694222848 + }, + { + "epoch": 0.21, + "learning_rate": 0.0006380902708124374, + "loss": 1.4932, + "theoretical_loss": 3.81107042065623, + "tokens_seen": 694288384 + }, + { + "epoch": 0.21, + "learning_rate": 0.000638074222668004, + "loss": 1.4335, + "theoretical_loss": 3.811034115746729, + "tokens_seen": 694353920 + }, + { + "epoch": 0.21, + "learning_rate": 0.0006380581745235708, + "loss": 1.4987, + "theoretical_loss": 3.810997815223022, + "tokens_seen": 694419456 + }, + { + "epoch": 0.21, + "learning_rate": 0.0006380421263791374, + "loss": 1.6172, + "theoretical_loss": 3.810961519084164, + "tokens_seen": 694484992 + }, + { + "epoch": 0.21, + "learning_rate": 0.0006380260782347042, + "loss": 1.4348, + "theoretical_loss": 3.8109252273292125, + "tokens_seen": 694550528 + }, + { + "epoch": 0.21, + "learning_rate": 0.0006380100300902708, + "loss": 1.4433, + "theoretical_loss": 3.8108889399572234, + "tokens_seen": 694616064 + }, + { + "epoch": 0.21, + "objective/train/docs_used": 258341, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.5283029079437256, + "objective/train/theoretical_loss": 3.810852656967255, + "objective/train/tokens_used": 715141600, + "theoretical_loss": 3.810852656967255, + "tokens_seen": 694681600 + }, + { + "epoch": 0.21, + "learning_rate": 0.0006379939819458376, + "loss": 1.6086, + "theoretical_loss": 3.810852656967255, + "tokens_seen": 694681600 + }, + { + "epoch": 0.21, + "learning_rate": 0.0006379779338014042, + "loss": 1.5084, + "theoretical_loss": 3.8108163783583637, + "tokens_seen": 694747136 + }, + { + "epoch": 0.21, + "learning_rate": 0.000637961885656971, + "loss": 1.549, + "theoretical_loss": 3.8107801041296083, + "tokens_seen": 694812672 + }, + { + "epoch": 0.21, + "learning_rate": 0.0006379458375125376, + "loss": 1.5883, + "theoretical_loss": 3.8107438342800464, + "tokens_seen": 694878208 + }, + { + "epoch": 0.21, + "learning_rate": 0.0006379297893681044, + "loss": 1.5219, + "theoretical_loss": 3.810707568808736, + "tokens_seen": 694943744 + }, + { + "epoch": 0.21, + "learning_rate": 0.000637913741223671, + "loss": 1.493, + "theoretical_loss": 3.8106713077147365, + "tokens_seen": 695009280 + }, + { + "epoch": 0.21, + "learning_rate": 0.0006378976930792378, + "loss": 1.5648, + "theoretical_loss": 3.810635050997107, + "tokens_seen": 695074816 + }, + { + "epoch": 0.21, + "learning_rate": 0.0006378816449348044, + "loss": 1.4202, + "theoretical_loss": 3.810598798654906, + "tokens_seen": 695140352 + }, + { + "epoch": 0.21, + "learning_rate": 0.0006378655967903712, + "loss": 1.5737, + "theoretical_loss": 3.8105625506871936, + "tokens_seen": 695205888 + }, + { + "epoch": 0.21, + "learning_rate": 0.0006378495486459378, + "loss": 1.5062, + "theoretical_loss": 3.810526307093029, + "tokens_seen": 695271424 + }, + { + "epoch": 0.21, + "learning_rate": 0.0006378335005015046, + "loss": 1.424, + "theoretical_loss": 3.810490067871473, + "tokens_seen": 695336960 + }, + { + "epoch": 0.21, + "learning_rate": 0.0006378174523570713, + "loss": 1.5776, + "theoretical_loss": 3.810453833021586, + "tokens_seen": 695402496 + }, + { + "epoch": 0.21, + "learning_rate": 0.000637801404212638, + "loss": 1.4616, + "theoretical_loss": 3.810417602542428, + "tokens_seen": 695468032 + }, + { + "epoch": 0.21, + "learning_rate": 0.0006377853560682047, + "loss": 1.4823, + "theoretical_loss": 3.8103813764330607, + "tokens_seen": 695533568 + }, + { + "epoch": 0.21, + "learning_rate": 0.0006377693079237714, + "loss": 1.414, + "theoretical_loss": 3.8103451546925453, + "tokens_seen": 695599104 + }, + { + "epoch": 0.21, + "learning_rate": 0.000637753259779338, + "loss": 1.6414, + "theoretical_loss": 3.8103089373199435, + "tokens_seen": 695664640 + }, + { + "epoch": 0.21, + "learning_rate": 0.0006377372116349048, + "loss": 1.4025, + "theoretical_loss": 3.8102727243143164, + "tokens_seen": 695730176 + }, + { + "epoch": 0.21, + "learning_rate": 0.0006377211634904714, + "loss": 1.4996, + "theoretical_loss": 3.8102365156747267, + "tokens_seen": 695795712 + }, + { + "epoch": 0.21, + "learning_rate": 0.0006377051153460382, + "loss": 1.5433, + "theoretical_loss": 3.810200311400237, + "tokens_seen": 695861248 + }, + { + "epoch": 0.21, + "learning_rate": 0.0006376890672016048, + "loss": 1.4733, + "theoretical_loss": 3.8101641114899105, + "tokens_seen": 695926784 + }, + { + "epoch": 0.21, + "learning_rate": 0.0006376730190571716, + "loss": 1.5206, + "theoretical_loss": 3.810127915942809, + "tokens_seen": 695992320 + }, + { + "epoch": 0.21, + "learning_rate": 0.0006376569709127382, + "loss": 1.4093, + "theoretical_loss": 3.8100917247579966, + "tokens_seen": 696057856 + }, + { + "epoch": 0.21, + "learning_rate": 0.000637640922768305, + "loss": 1.4577, + "theoretical_loss": 3.8100555379345367, + "tokens_seen": 696123392 + }, + { + "epoch": 0.21, + "learning_rate": 0.0006376248746238716, + "loss": 1.5577, + "theoretical_loss": 3.810019355471493, + "tokens_seen": 696188928 + }, + { + "epoch": 0.21, + "learning_rate": 0.0006376088264794383, + "loss": 1.5065, + "theoretical_loss": 3.80998317736793, + "tokens_seen": 696254464 + }, + { + "epoch": 0.21, + "learning_rate": 0.000637592778335005, + "loss": 1.5244, + "theoretical_loss": 3.809947003622913, + "tokens_seen": 696320000 + }, + { + "epoch": 0.21, + "learning_rate": 0.0006375767301905717, + "loss": 1.5675, + "theoretical_loss": 3.8099108342355046, + "tokens_seen": 696385536 + }, + { + "epoch": 0.21, + "learning_rate": 0.0006375606820461384, + "loss": 1.4737, + "theoretical_loss": 3.809874669204772, + "tokens_seen": 696451072 + }, + { + "epoch": 0.21, + "learning_rate": 0.0006375446339017051, + "loss": 1.5206, + "theoretical_loss": 3.8098385085297792, + "tokens_seen": 696516608 + }, + { + "epoch": 0.21, + "learning_rate": 0.0006375285857572718, + "loss": 1.4696, + "theoretical_loss": 3.809802352209592, + "tokens_seen": 696582144 + }, + { + "epoch": 0.21, + "learning_rate": 0.0006375125376128385, + "loss": 1.4225, + "theoretical_loss": 3.809766200243277, + "tokens_seen": 696647680 + }, + { + "epoch": 0.21, + "learning_rate": 0.0006374964894684052, + "loss": 1.5759, + "theoretical_loss": 3.8097300526299, + "tokens_seen": 696713216 + }, + { + "epoch": 0.21, + "learning_rate": 0.0006374804413239719, + "loss": 1.6389, + "theoretical_loss": 3.8096939093685274, + "tokens_seen": 696778752 + }, + { + "epoch": 0.21, + "learning_rate": 0.0006374643931795386, + "loss": 1.4198, + "theoretical_loss": 3.8096577704582266, + "tokens_seen": 696844288 + }, + { + "epoch": 0.21, + "learning_rate": 0.0006374483450351053, + "loss": 1.5305, + "theoretical_loss": 3.8096216358980635, + "tokens_seen": 696909824 + }, + { + "epoch": 0.21, + "learning_rate": 0.000637432296890672, + "loss": 1.6055, + "theoretical_loss": 3.809585505687106, + "tokens_seen": 696975360 + }, + { + "epoch": 0.21, + "learning_rate": 0.0006374162487462387, + "loss": 1.5488, + "theoretical_loss": 3.809549379824422, + "tokens_seen": 697040896 + }, + { + "epoch": 0.21, + "learning_rate": 0.0006374002006018054, + "loss": 1.4885, + "theoretical_loss": 3.80951325830908, + "tokens_seen": 697106432 + }, + { + "epoch": 0.21, + "learning_rate": 0.0006373841524573721, + "loss": 1.4303, + "theoretical_loss": 3.809477141140147, + "tokens_seen": 697171968 + }, + { + "epoch": 0.21, + "learning_rate": 0.0006373681043129388, + "loss": 1.4344, + "theoretical_loss": 3.8094410283166917, + "tokens_seen": 697237504 + }, + { + "epoch": 0.21, + "learning_rate": 0.0006373520561685055, + "loss": 1.4836, + "theoretical_loss": 3.809404919837784, + "tokens_seen": 697303040 + }, + { + "epoch": 0.21, + "learning_rate": 0.0006373360080240722, + "loss": 1.6107, + "theoretical_loss": 3.8093688157024914, + "tokens_seen": 697368576 + }, + { + "epoch": 0.21, + "learning_rate": 0.0006373199598796389, + "loss": 1.5229, + "theoretical_loss": 3.8093327159098846, + "tokens_seen": 697434112 + }, + { + "epoch": 0.21, + "learning_rate": 0.0006373039117352057, + "loss": 1.4762, + "theoretical_loss": 3.8092966204590324, + "tokens_seen": 697499648 + }, + { + "epoch": 0.21, + "learning_rate": 0.0006372878635907723, + "loss": 1.6672, + "theoretical_loss": 3.8092605293490056, + "tokens_seen": 697565184 + }, + { + "epoch": 0.21, + "learning_rate": 0.000637271815446339, + "loss": 1.5458, + "theoretical_loss": 3.809224442578873, + "tokens_seen": 697630720 + }, + { + "epoch": 0.21, + "learning_rate": 0.0006372557673019057, + "loss": 1.5555, + "theoretical_loss": 3.809188360147707, + "tokens_seen": 697696256 + }, + { + "epoch": 0.21, + "learning_rate": 0.0006372397191574725, + "loss": 1.4486, + "theoretical_loss": 3.8091522820545762, + "tokens_seen": 697761792 + }, + { + "epoch": 0.21, + "learning_rate": 0.0006372236710130391, + "loss": 1.6238, + "theoretical_loss": 3.809116208298554, + "tokens_seen": 697827328 + }, + { + "epoch": 0.21, + "learning_rate": 0.0006372076228686059, + "loss": 1.5844, + "theoretical_loss": 3.80908013887871, + "tokens_seen": 697892864 + }, + { + "epoch": 0.21, + "objective/train/docs_used": 259086, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.5347952842712402, + "objective/train/theoretical_loss": 3.8090440737941176, + "objective/train/tokens_used": 718418400, + "theoretical_loss": 3.8090440737941176, + "tokens_seen": 697958400 + }, + { + "epoch": 0.21, + "learning_rate": 0.0006371915747241725, + "loss": 1.4945, + "theoretical_loss": 3.8090440737941176, + "tokens_seen": 697958400 + }, + { + "epoch": 0.21, + "learning_rate": 0.0006371755265797393, + "loss": 1.4734, + "theoretical_loss": 3.8090080130438464, + "tokens_seen": 698023936 + }, + { + "epoch": 0.21, + "learning_rate": 0.0006371594784353059, + "loss": 1.5514, + "theoretical_loss": 3.80897195662697, + "tokens_seen": 698089472 + }, + { + "epoch": 0.21, + "learning_rate": 0.0006371434302908727, + "loss": 1.3941, + "theoretical_loss": 3.8089359045425617, + "tokens_seen": 698155008 + }, + { + "epoch": 0.21, + "learning_rate": 0.0006371273821464393, + "loss": 1.4248, + "theoretical_loss": 3.808899856789693, + "tokens_seen": 698220544 + }, + { + "epoch": 0.21, + "learning_rate": 0.0006371113340020061, + "loss": 1.5949, + "theoretical_loss": 3.808863813367437, + "tokens_seen": 698286080 + }, + { + "epoch": 0.21, + "learning_rate": 0.0006370952858575727, + "loss": 1.5318, + "theoretical_loss": 3.808827774274868, + "tokens_seen": 698351616 + }, + { + "epoch": 0.21, + "learning_rate": 0.0006370792377131395, + "loss": 1.477, + "theoretical_loss": 3.808791739511059, + "tokens_seen": 698417152 + }, + { + "epoch": 0.21, + "learning_rate": 0.0006370631895687061, + "loss": 1.5584, + "theoretical_loss": 3.808755709075084, + "tokens_seen": 698482688 + }, + { + "epoch": 0.21, + "learning_rate": 0.0006370471414242729, + "loss": 1.5062, + "theoretical_loss": 3.8087196829660175, + "tokens_seen": 698548224 + }, + { + "epoch": 0.21, + "learning_rate": 0.0006370310932798395, + "loss": 1.5919, + "theoretical_loss": 3.8086836611829336, + "tokens_seen": 698613760 + }, + { + "epoch": 0.21, + "learning_rate": 0.0006370150451354063, + "loss": 1.3739, + "theoretical_loss": 3.8086476437249077, + "tokens_seen": 698679296 + }, + { + "epoch": 0.21, + "learning_rate": 0.0006369989969909729, + "loss": 1.5431, + "theoretical_loss": 3.808611630591014, + "tokens_seen": 698744832 + }, + { + "epoch": 0.21, + "learning_rate": 0.0006369829488465397, + "loss": 1.5225, + "theoretical_loss": 3.808575621780329, + "tokens_seen": 698810368 + }, + { + "epoch": 0.21, + "learning_rate": 0.0006369669007021063, + "loss": 1.5584, + "theoretical_loss": 3.8085396172919275, + "tokens_seen": 698875904 + }, + { + "epoch": 0.21, + "learning_rate": 0.0006369508525576731, + "loss": 1.5733, + "theoretical_loss": 3.8085036171248854, + "tokens_seen": 698941440 + }, + { + "epoch": 0.21, + "learning_rate": 0.0006369348044132397, + "loss": 1.5744, + "theoretical_loss": 3.8084676212782793, + "tokens_seen": 699006976 + }, + { + "epoch": 0.21, + "learning_rate": 0.0006369187562688065, + "loss": 1.6141, + "theoretical_loss": 3.808431629751186, + "tokens_seen": 699072512 + }, + { + "epoch": 0.21, + "learning_rate": 0.0006369027081243731, + "loss": 1.4771, + "theoretical_loss": 3.8083956425426813, + "tokens_seen": 699138048 + }, + { + "epoch": 0.21, + "learning_rate": 0.0006368866599799399, + "loss": 1.5123, + "theoretical_loss": 3.808359659651843, + "tokens_seen": 699203584 + }, + { + "epoch": 0.21, + "learning_rate": 0.0006368706118355065, + "loss": 1.4837, + "theoretical_loss": 3.8083236810777485, + "tokens_seen": 699269120 + }, + { + "epoch": 0.21, + "learning_rate": 0.0006368545636910733, + "loss": 1.4533, + "theoretical_loss": 3.808287706819475, + "tokens_seen": 699334656 + }, + { + "epoch": 0.21, + "learning_rate": 0.0006368385155466399, + "loss": 1.5021, + "theoretical_loss": 3.808251736876101, + "tokens_seen": 699400192 + }, + { + "epoch": 0.21, + "learning_rate": 0.0006368224674022067, + "loss": 1.5484, + "theoretical_loss": 3.8082157712467044, + "tokens_seen": 699465728 + }, + { + "epoch": 0.21, + "learning_rate": 0.0006368064192577733, + "loss": 1.5523, + "theoretical_loss": 3.8081798099303636, + "tokens_seen": 699531264 + }, + { + "epoch": 0.21, + "learning_rate": 0.0006367903711133401, + "loss": 1.5275, + "theoretical_loss": 3.808143852926157, + "tokens_seen": 699596800 + }, + { + "epoch": 0.21, + "learning_rate": 0.0006367743229689067, + "loss": 1.5267, + "theoretical_loss": 3.8081079002331646, + "tokens_seen": 699662336 + }, + { + "epoch": 0.21, + "learning_rate": 0.0006367582748244735, + "loss": 1.4858, + "theoretical_loss": 3.808071951850465, + "tokens_seen": 699727872 + }, + { + "epoch": 0.21, + "learning_rate": 0.0006367422266800401, + "loss": 1.5704, + "theoretical_loss": 3.8080360077771385, + "tokens_seen": 699793408 + }, + { + "epoch": 0.21, + "learning_rate": 0.0006367261785356069, + "loss": 1.4373, + "theoretical_loss": 3.808000068012264, + "tokens_seen": 699858944 + }, + { + "epoch": 0.21, + "learning_rate": 0.0006367101303911735, + "loss": 1.499, + "theoretical_loss": 3.8079641325549223, + "tokens_seen": 699924480 + }, + { + "epoch": 0.21, + "learning_rate": 0.0006366940822467403, + "loss": 1.5815, + "theoretical_loss": 3.807928201404194, + "tokens_seen": 699990016 + }, + { + "epoch": 0.21, + "learning_rate": 0.0006366780341023069, + "loss": 1.3957, + "theoretical_loss": 3.8078922745591597, + "tokens_seen": 700055552 + }, + { + "epoch": 0.21, + "learning_rate": 0.0006366619859578737, + "loss": 1.2423, + "theoretical_loss": 3.8078563520189004, + "tokens_seen": 700121088 + }, + { + "epoch": 0.21, + "learning_rate": 0.0006366459378134403, + "loss": 1.5284, + "theoretical_loss": 3.8078204337824975, + "tokens_seen": 700186624 + }, + { + "epoch": 0.21, + "learning_rate": 0.0006366298896690071, + "loss": 1.5413, + "theoretical_loss": 3.807784519849032, + "tokens_seen": 700252160 + }, + { + "epoch": 0.21, + "learning_rate": 0.0006366138415245738, + "loss": 1.5432, + "theoretical_loss": 3.807748610217587, + "tokens_seen": 700317696 + }, + { + "epoch": 0.21, + "learning_rate": 0.0006365977933801405, + "loss": 1.5434, + "theoretical_loss": 3.8077127048872437, + "tokens_seen": 700383232 + }, + { + "epoch": 0.21, + "learning_rate": 0.0006365817452357072, + "loss": 1.5225, + "theoretical_loss": 3.8076768038570847, + "tokens_seen": 700448768 + }, + { + "epoch": 0.21, + "learning_rate": 0.0006365656970912739, + "loss": 1.5509, + "theoretical_loss": 3.807640907126193, + "tokens_seen": 700514304 + }, + { + "epoch": 0.21, + "learning_rate": 0.0006365496489468406, + "loss": 1.463, + "theoretical_loss": 3.8076050146936513, + "tokens_seen": 700579840 + }, + { + "epoch": 0.21, + "learning_rate": 0.0006365336008024073, + "loss": 1.58, + "theoretical_loss": 3.807569126558543, + "tokens_seen": 700645376 + }, + { + "epoch": 0.21, + "learning_rate": 0.000636517552657974, + "loss": 1.4288, + "theoretical_loss": 3.807533242719952, + "tokens_seen": 700710912 + }, + { + "epoch": 0.21, + "learning_rate": 0.0006365015045135407, + "loss": 1.5504, + "theoretical_loss": 3.807497363176962, + "tokens_seen": 700776448 + }, + { + "epoch": 0.21, + "learning_rate": 0.0006364854563691074, + "loss": 1.4886, + "theoretical_loss": 3.8074614879286566, + "tokens_seen": 700841984 + }, + { + "epoch": 0.21, + "learning_rate": 0.0006364694082246741, + "loss": 1.4972, + "theoretical_loss": 3.8074256169741205, + "tokens_seen": 700907520 + }, + { + "epoch": 0.21, + "learning_rate": 0.0006364533600802408, + "loss": 1.5241, + "theoretical_loss": 3.8073897503124385, + "tokens_seen": 700973056 + }, + { + "epoch": 0.21, + "learning_rate": 0.0006364373119358075, + "loss": 1.4885, + "theoretical_loss": 3.807353887942696, + "tokens_seen": 701038592 + }, + { + "epoch": 0.21, + "learning_rate": 0.0006364212637913742, + "loss": 1.4331, + "theoretical_loss": 3.8073180298639775, + "tokens_seen": 701104128 + }, + { + "epoch": 0.21, + "learning_rate": 0.0006364052156469409, + "loss": 1.4871, + "theoretical_loss": 3.8072821760753683, + "tokens_seen": 701169664 + }, + { + "epoch": 0.21, + "objective/train/docs_used": 260260, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.5225542783737183, + "objective/train/theoretical_loss": 3.8072463265759557, + "objective/train/tokens_used": 721695200, + "theoretical_loss": 3.8072463265759557, + "tokens_seen": 701235200 + }, + { + "epoch": 0.21, + "learning_rate": 0.0006363891675025076, + "loss": 1.5209, + "theoretical_loss": 3.8072463265759557, + "tokens_seen": 701235200 + }, + { + "epoch": 0.21, + "learning_rate": 0.0006363731193580743, + "loss": 1.5221, + "theoretical_loss": 3.8072104813648244, + "tokens_seen": 701300736 + }, + { + "epoch": 0.21, + "learning_rate": 0.0006363570712136409, + "loss": 1.4769, + "theoretical_loss": 3.8071746404410614, + "tokens_seen": 701366272 + }, + { + "epoch": 0.21, + "learning_rate": 0.0006363410230692077, + "loss": 1.4826, + "theoretical_loss": 3.807138803803753, + "tokens_seen": 701431808 + }, + { + "epoch": 0.21, + "learning_rate": 0.0006363249749247743, + "loss": 1.5209, + "theoretical_loss": 3.807102971451986, + "tokens_seen": 701497344 + }, + { + "epoch": 0.21, + "learning_rate": 0.0006363089267803411, + "loss": 1.5249, + "theoretical_loss": 3.807067143384849, + "tokens_seen": 701562880 + }, + { + "epoch": 0.21, + "learning_rate": 0.0006362928786359077, + "loss": 1.5036, + "theoretical_loss": 3.8070313196014274, + "tokens_seen": 701628416 + }, + { + "epoch": 0.21, + "learning_rate": 0.0006362768304914745, + "loss": 1.5476, + "theoretical_loss": 3.8069955001008102, + "tokens_seen": 701693952 + }, + { + "epoch": 0.21, + "learning_rate": 0.0006362607823470411, + "loss": 1.5557, + "theoretical_loss": 3.8069596848820852, + "tokens_seen": 701759488 + }, + { + "epoch": 0.21, + "learning_rate": 0.0006362447342026079, + "loss": 1.5245, + "theoretical_loss": 3.806923873944341, + "tokens_seen": 701825024 + }, + { + "epoch": 0.21, + "learning_rate": 0.0006362286860581745, + "loss": 1.517, + "theoretical_loss": 3.8068880672866667, + "tokens_seen": 701890560 + }, + { + "epoch": 0.21, + "learning_rate": 0.0006362126379137412, + "loss": 1.6098, + "theoretical_loss": 3.8068522649081498, + "tokens_seen": 701956096 + }, + { + "epoch": 0.21, + "learning_rate": 0.0006361965897693079, + "loss": 1.5453, + "theoretical_loss": 3.80681646680788, + "tokens_seen": 702021632 + }, + { + "epoch": 0.21, + "learning_rate": 0.0006361805416248746, + "loss": 1.4728, + "theoretical_loss": 3.8067806729849467, + "tokens_seen": 702087168 + }, + { + "epoch": 0.21, + "learning_rate": 0.0006361644934804413, + "loss": 1.472, + "theoretical_loss": 3.8067448834384408, + "tokens_seen": 702152704 + }, + { + "epoch": 0.21, + "learning_rate": 0.000636148445336008, + "loss": 1.3047, + "theoretical_loss": 3.8067090981674507, + "tokens_seen": 702218240 + }, + { + "epoch": 0.21, + "learning_rate": 0.0006361323971915747, + "loss": 1.5322, + "theoretical_loss": 3.806673317171067, + "tokens_seen": 702283776 + }, + { + "epoch": 0.21, + "learning_rate": 0.0006361163490471414, + "loss": 1.5077, + "theoretical_loss": 3.8066375404483814, + "tokens_seen": 702349312 + }, + { + "epoch": 0.21, + "learning_rate": 0.0006361003009027081, + "loss": 1.434, + "theoretical_loss": 3.806601767998483, + "tokens_seen": 702414848 + }, + { + "epoch": 0.21, + "learning_rate": 0.0006360842527582748, + "loss": 1.51, + "theoretical_loss": 3.806565999820465, + "tokens_seen": 702480384 + }, + { + "epoch": 0.21, + "learning_rate": 0.0006360682046138415, + "loss": 1.4869, + "theoretical_loss": 3.806530235913417, + "tokens_seen": 702545920 + }, + { + "epoch": 0.21, + "learning_rate": 0.0006360521564694082, + "loss": 1.4917, + "theoretical_loss": 3.8064944762764314, + "tokens_seen": 702611456 + }, + { + "epoch": 0.21, + "learning_rate": 0.000636036108324975, + "loss": 1.6427, + "theoretical_loss": 3.8064587209086005, + "tokens_seen": 702676992 + }, + { + "epoch": 0.21, + "learning_rate": 0.0006360200601805416, + "loss": 1.5601, + "theoretical_loss": 3.8064229698090157, + "tokens_seen": 702742528 + }, + { + "epoch": 0.21, + "learning_rate": 0.0006360040120361083, + "loss": 1.5213, + "theoretical_loss": 3.80638722297677, + "tokens_seen": 702808064 + }, + { + "epoch": 0.21, + "learning_rate": 0.000635987963891675, + "loss": 1.5063, + "theoretical_loss": 3.8063514804109557, + "tokens_seen": 702873600 + }, + { + "epoch": 0.21, + "learning_rate": 0.0006359719157472417, + "loss": 1.5767, + "theoretical_loss": 3.806315742110667, + "tokens_seen": 702939136 + }, + { + "epoch": 0.21, + "learning_rate": 0.0006359558676028084, + "loss": 1.537, + "theoretical_loss": 3.806280008074996, + "tokens_seen": 703004672 + }, + { + "epoch": 0.21, + "learning_rate": 0.0006359398194583752, + "loss": 1.5375, + "theoretical_loss": 3.806244278303037, + "tokens_seen": 703070208 + }, + { + "epoch": 0.21, + "learning_rate": 0.0006359237713139418, + "loss": 1.5376, + "theoretical_loss": 3.806208552793884, + "tokens_seen": 703135744 + }, + { + "epoch": 0.21, + "learning_rate": 0.0006359077231695086, + "loss": 1.5653, + "theoretical_loss": 3.8061728315466308, + "tokens_seen": 703201280 + }, + { + "epoch": 0.21, + "learning_rate": 0.0006358916750250752, + "loss": 1.4455, + "theoretical_loss": 3.8061371145603715, + "tokens_seen": 703266816 + }, + { + "epoch": 0.21, + "learning_rate": 0.000635875626880642, + "loss": 1.4939, + "theoretical_loss": 3.806101401834202, + "tokens_seen": 703332352 + }, + { + "epoch": 0.21, + "learning_rate": 0.0006358595787362086, + "loss": 1.4687, + "theoretical_loss": 3.8060656933672163, + "tokens_seen": 703397888 + }, + { + "epoch": 0.21, + "learning_rate": 0.0006358435305917754, + "loss": 1.5546, + "theoretical_loss": 3.8060299891585094, + "tokens_seen": 703463424 + }, + { + "epoch": 0.21, + "learning_rate": 0.000635827482447342, + "loss": 1.4465, + "theoretical_loss": 3.8059942892071783, + "tokens_seen": 703528960 + }, + { + "epoch": 0.21, + "learning_rate": 0.0006358114343029088, + "loss": 1.5727, + "theoretical_loss": 3.8059585935123175, + "tokens_seen": 703594496 + }, + { + "epoch": 0.21, + "learning_rate": 0.0006357953861584754, + "loss": 1.3998, + "theoretical_loss": 3.8059229020730236, + "tokens_seen": 703660032 + }, + { + "epoch": 0.21, + "learning_rate": 0.0006357793380140422, + "loss": 1.4891, + "theoretical_loss": 3.805887214888393, + "tokens_seen": 703725568 + }, + { + "epoch": 0.21, + "learning_rate": 0.0006357632898696088, + "loss": 1.6125, + "theoretical_loss": 3.8058515319575226, + "tokens_seen": 703791104 + }, + { + "epoch": 0.21, + "learning_rate": 0.0006357472417251756, + "loss": 1.5231, + "theoretical_loss": 3.8058158532795088, + "tokens_seen": 703856640 + }, + { + "epoch": 0.21, + "learning_rate": 0.0006357311935807422, + "loss": 1.6348, + "theoretical_loss": 3.805780178853449, + "tokens_seen": 703922176 + }, + { + "epoch": 0.21, + "learning_rate": 0.000635715145436309, + "loss": 1.4602, + "theoretical_loss": 3.8057445086784405, + "tokens_seen": 703987712 + }, + { + "epoch": 0.21, + "learning_rate": 0.0006356990972918756, + "loss": 1.4738, + "theoretical_loss": 3.8057088427535817, + "tokens_seen": 704053248 + }, + { + "epoch": 0.21, + "learning_rate": 0.0006356830491474424, + "loss": 1.5122, + "theoretical_loss": 3.80567318107797, + "tokens_seen": 704118784 + }, + { + "epoch": 0.21, + "learning_rate": 0.000635667001003009, + "loss": 1.5533, + "theoretical_loss": 3.8056375236507045, + "tokens_seen": 704184320 + }, + { + "epoch": 0.21, + "learning_rate": 0.0006356509528585758, + "loss": 1.4977, + "theoretical_loss": 3.805601870470883, + "tokens_seen": 704249856 + }, + { + "epoch": 0.21, + "learning_rate": 0.0006356349047141424, + "loss": 1.5001, + "theoretical_loss": 3.805566221537604, + "tokens_seen": 704315392 + }, + { + "epoch": 0.21, + "learning_rate": 0.0006356188565697092, + "loss": 1.4333, + "theoretical_loss": 3.805530576849968, + "tokens_seen": 704380928 + }, + { + "epoch": 0.21, + "learning_rate": 0.0006356028084252758, + "loss": 1.5364, + "theoretical_loss": 3.8054949364070736, + "tokens_seen": 704446464 + }, + { + "epoch": 0.21, + "objective/train/docs_used": 261464, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.3635611534118652, + "objective/train/theoretical_loss": 3.8054593002080206, + "objective/train/tokens_used": 724972000, + "theoretical_loss": 3.8054593002080206, + "tokens_seen": 704512000 + }, + { + "epoch": 0.21, + "learning_rate": 0.0006355867602808426, + "loss": 1.4365, + "theoretical_loss": 3.8054593002080206, + "tokens_seen": 704512000 + }, + { + "epoch": 0.21, + "learning_rate": 0.0006355707121364092, + "loss": 1.4354, + "theoretical_loss": 3.805423668251909, + "tokens_seen": 704577536 + }, + { + "epoch": 0.21, + "learning_rate": 0.000635554663991976, + "loss": 1.5625, + "theoretical_loss": 3.8053880405378386, + "tokens_seen": 704643072 + }, + { + "epoch": 0.21, + "learning_rate": 0.0006355386158475426, + "loss": 1.5555, + "theoretical_loss": 3.8053524170649107, + "tokens_seen": 704708608 + }, + { + "epoch": 0.21, + "learning_rate": 0.0006355225677031094, + "loss": 1.5279, + "theoretical_loss": 3.805316797832225, + "tokens_seen": 704774144 + }, + { + "epoch": 0.21, + "learning_rate": 0.000635506519558676, + "loss": 1.4662, + "theoretical_loss": 3.805281182838884, + "tokens_seen": 704839680 + }, + { + "epoch": 0.21, + "learning_rate": 0.0006354904714142428, + "loss": 1.6402, + "theoretical_loss": 3.8052455720839875, + "tokens_seen": 704905216 + }, + { + "epoch": 0.21, + "learning_rate": 0.0006354744232698094, + "loss": 1.5208, + "theoretical_loss": 3.8052099655666387, + "tokens_seen": 704970752 + }, + { + "epoch": 0.21, + "learning_rate": 0.0006354583751253762, + "loss": 1.5596, + "theoretical_loss": 3.8051743632859383, + "tokens_seen": 705036288 + }, + { + "epoch": 0.21, + "learning_rate": 0.0006354423269809428, + "loss": 1.5007, + "theoretical_loss": 3.8051387652409887, + "tokens_seen": 705101824 + }, + { + "epoch": 0.21, + "learning_rate": 0.0006354262788365096, + "loss": 1.5474, + "theoretical_loss": 3.8051031714308925, + "tokens_seen": 705167360 + }, + { + "epoch": 0.21, + "learning_rate": 0.0006354102306920763, + "loss": 1.4175, + "theoretical_loss": 3.8050675818547526, + "tokens_seen": 705232896 + }, + { + "epoch": 0.21, + "learning_rate": 0.000635394182547643, + "loss": 1.5003, + "theoretical_loss": 3.8050319965116715, + "tokens_seen": 705298432 + }, + { + "epoch": 0.21, + "learning_rate": 0.0006353781344032097, + "loss": 1.423, + "theoretical_loss": 3.804996415400753, + "tokens_seen": 705363968 + }, + { + "epoch": 0.21, + "learning_rate": 0.0006353620862587764, + "loss": 1.4022, + "theoretical_loss": 3.8049608385211, + "tokens_seen": 705429504 + }, + { + "epoch": 0.21, + "learning_rate": 0.0006353460381143431, + "loss": 1.4297, + "theoretical_loss": 3.804925265871817, + "tokens_seen": 705495040 + }, + { + "epoch": 0.21, + "learning_rate": 0.0006353299899699098, + "loss": 1.56, + "theoretical_loss": 3.804889697452007, + "tokens_seen": 705560576 + }, + { + "epoch": 0.21, + "learning_rate": 0.0006353139418254765, + "loss": 1.5255, + "theoretical_loss": 3.8048541332607755, + "tokens_seen": 705626112 + }, + { + "epoch": 0.21, + "learning_rate": 0.0006352978936810432, + "loss": 1.5218, + "theoretical_loss": 3.8048185732972266, + "tokens_seen": 705691648 + }, + { + "epoch": 0.21, + "learning_rate": 0.0006352818455366099, + "loss": 1.4214, + "theoretical_loss": 3.8047830175604656, + "tokens_seen": 705757184 + }, + { + "epoch": 0.21, + "learning_rate": 0.0006352657973921766, + "loss": 1.5333, + "theoretical_loss": 3.804747466049597, + "tokens_seen": 705822720 + }, + { + "epoch": 0.21, + "learning_rate": 0.0006352497492477433, + "loss": 1.5674, + "theoretical_loss": 3.804711918763726, + "tokens_seen": 705888256 + }, + { + "epoch": 0.21, + "learning_rate": 0.00063523370110331, + "loss": 1.4722, + "theoretical_loss": 3.8046763757019595, + "tokens_seen": 705953792 + }, + { + "epoch": 0.21, + "learning_rate": 0.0006352176529588767, + "loss": 1.4562, + "theoretical_loss": 3.8046408368634026, + "tokens_seen": 706019328 + }, + { + "epoch": 0.21, + "learning_rate": 0.0006352016048144434, + "loss": 1.5119, + "theoretical_loss": 3.804605302247162, + "tokens_seen": 706084864 + }, + { + "epoch": 0.21, + "learning_rate": 0.0006351855566700101, + "loss": 1.5543, + "theoretical_loss": 3.8045697718523437, + "tokens_seen": 706150400 + }, + { + "epoch": 0.21, + "learning_rate": 0.0006351695085255768, + "loss": 1.572, + "theoretical_loss": 3.804534245678055, + "tokens_seen": 706215936 + }, + { + "epoch": 0.21, + "learning_rate": 0.0006351534603811435, + "loss": 1.5528, + "theoretical_loss": 3.8044987237234027, + "tokens_seen": 706281472 + }, + { + "epoch": 0.21, + "learning_rate": 0.0006351374122367102, + "loss": 1.4538, + "theoretical_loss": 3.804463205987494, + "tokens_seen": 706347008 + }, + { + "epoch": 0.21, + "learning_rate": 0.0006351213640922769, + "loss": 1.5013, + "theoretical_loss": 3.804427692469437, + "tokens_seen": 706412544 + }, + { + "epoch": 0.21, + "learning_rate": 0.0006351053159478436, + "loss": 1.4473, + "theoretical_loss": 3.8043921831683396, + "tokens_seen": 706478080 + }, + { + "epoch": 0.21, + "learning_rate": 0.0006350892678034103, + "loss": 1.5322, + "theoretical_loss": 3.804356678083309, + "tokens_seen": 706543616 + }, + { + "epoch": 0.21, + "learning_rate": 0.000635073219658977, + "loss": 1.5636, + "theoretical_loss": 3.8043211772134544, + "tokens_seen": 706609152 + }, + { + "epoch": 0.21, + "learning_rate": 0.0006350571715145437, + "loss": 1.4998, + "theoretical_loss": 3.804285680557885, + "tokens_seen": 706674688 + }, + { + "epoch": 0.21, + "learning_rate": 0.0006350411233701104, + "loss": 1.4749, + "theoretical_loss": 3.8042501881157085, + "tokens_seen": 706740224 + }, + { + "epoch": 0.21, + "learning_rate": 0.0006350250752256771, + "loss": 1.4923, + "theoretical_loss": 3.804214699886035, + "tokens_seen": 706805760 + }, + { + "epoch": 0.21, + "learning_rate": 0.0006350090270812438, + "loss": 1.5992, + "theoretical_loss": 3.8041792158679737, + "tokens_seen": 706871296 + }, + { + "epoch": 0.21, + "learning_rate": 0.0006349929789368105, + "loss": 1.5599, + "theoretical_loss": 3.804143736060635, + "tokens_seen": 706936832 + }, + { + "epoch": 0.21, + "learning_rate": 0.0006349769307923772, + "loss": 1.4936, + "theoretical_loss": 3.8041082604631278, + "tokens_seen": 707002368 + }, + { + "epoch": 0.21, + "learning_rate": 0.0006349608826479438, + "loss": 1.4608, + "theoretical_loss": 3.804072789074563, + "tokens_seen": 707067904 + }, + { + "epoch": 0.21, + "learning_rate": 0.0006349448345035106, + "loss": 1.5746, + "theoretical_loss": 3.8040373218940515, + "tokens_seen": 707133440 + }, + { + "epoch": 0.21, + "learning_rate": 0.0006349287863590772, + "loss": 1.5694, + "theoretical_loss": 3.804001858920704, + "tokens_seen": 707198976 + }, + { + "epoch": 0.21, + "learning_rate": 0.000634912738214644, + "loss": 1.4266, + "theoretical_loss": 3.8039664001536315, + "tokens_seen": 707264512 + }, + { + "epoch": 0.21, + "learning_rate": 0.0006348966900702106, + "loss": 1.4809, + "theoretical_loss": 3.8039309455919454, + "tokens_seen": 707330048 + }, + { + "epoch": 0.21, + "learning_rate": 0.0006348806419257774, + "loss": 1.4147, + "theoretical_loss": 3.803895495234757, + "tokens_seen": 707395584 + }, + { + "epoch": 0.21, + "learning_rate": 0.000634864593781344, + "loss": 1.5234, + "theoretical_loss": 3.8038600490811794, + "tokens_seen": 707461120 + }, + { + "epoch": 0.21, + "learning_rate": 0.0006348485456369108, + "loss": 1.5451, + "theoretical_loss": 3.8038246071303234, + "tokens_seen": 707526656 + }, + { + "epoch": 0.21, + "learning_rate": 0.0006348324974924774, + "loss": 1.4774, + "theoretical_loss": 3.803789169381303, + "tokens_seen": 707592192 + }, + { + "epoch": 0.21, + "learning_rate": 0.0006348164493480441, + "loss": 1.4849, + "theoretical_loss": 3.8037537358332294, + "tokens_seen": 707657728 + }, + { + "epoch": 0.21, + "learning_rate": 0.0006348004012036108, + "loss": 1.5174, + "theoretical_loss": 3.8037183064852162, + "tokens_seen": 707723264 + }, + { + "epoch": 0.21, + "objective/train/docs_used": 262675, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.6891382932662964, + "objective/train/theoretical_loss": 3.8036828813363774, + "objective/train/tokens_used": 728248800, + "theoretical_loss": 3.8036828813363774, + "tokens_seen": 707788800 + }, + { + "epoch": 0.21, + "learning_rate": 0.0006347843530591775, + "loss": 1.6218, + "theoretical_loss": 3.8036828813363774, + "tokens_seen": 707788800 + }, + { + "epoch": 0.21, + "learning_rate": 0.0006347683049147442, + "loss": 1.4273, + "theoretical_loss": 3.8036474603858252, + "tokens_seen": 707854336 + }, + { + "epoch": 0.21, + "learning_rate": 0.0006347522567703109, + "loss": 1.6154, + "theoretical_loss": 3.803612043632675, + "tokens_seen": 707919872 + }, + { + "epoch": 0.21, + "learning_rate": 0.0006347362086258776, + "loss": 1.4958, + "theoretical_loss": 3.80357663107604, + "tokens_seen": 707985408 + }, + { + "epoch": 0.21, + "learning_rate": 0.0006347201604814443, + "loss": 1.3978, + "theoretical_loss": 3.803541222715034, + "tokens_seen": 708050944 + }, + { + "epoch": 0.21, + "learning_rate": 0.000634704112337011, + "loss": 1.5642, + "theoretical_loss": 3.803505818548773, + "tokens_seen": 708116480 + }, + { + "epoch": 0.21, + "learning_rate": 0.0006346880641925777, + "loss": 1.5905, + "theoretical_loss": 3.803470418576371, + "tokens_seen": 708182016 + }, + { + "epoch": 0.21, + "learning_rate": 0.0006346720160481444, + "loss": 1.5192, + "theoretical_loss": 3.8034350227969433, + "tokens_seen": 708247552 + }, + { + "epoch": 0.21, + "learning_rate": 0.0006346559679037111, + "loss": 1.5418, + "theoretical_loss": 3.8033996312096052, + "tokens_seen": 708313088 + }, + { + "epoch": 0.21, + "learning_rate": 0.0006346399197592778, + "loss": 1.5315, + "theoretical_loss": 3.803364243813473, + "tokens_seen": 708378624 + }, + { + "epoch": 0.21, + "learning_rate": 0.0006346238716148445, + "loss": 1.5639, + "theoretical_loss": 3.803328860607662, + "tokens_seen": 708444160 + }, + { + "epoch": 0.21, + "learning_rate": 0.0006346078234704112, + "loss": 1.5525, + "theoretical_loss": 3.8032934815912895, + "tokens_seen": 708509696 + }, + { + "epoch": 0.21, + "learning_rate": 0.0006345917753259779, + "loss": 1.4884, + "theoretical_loss": 3.8032581067634705, + "tokens_seen": 708575232 + }, + { + "epoch": 0.21, + "learning_rate": 0.0006345757271815446, + "loss": 1.527, + "theoretical_loss": 3.803222736123323, + "tokens_seen": 708640768 + }, + { + "epoch": 0.21, + "learning_rate": 0.0006345596790371113, + "loss": 1.4092, + "theoretical_loss": 3.8031873696699634, + "tokens_seen": 708706304 + }, + { + "epoch": 0.21, + "learning_rate": 0.000634543630892678, + "loss": 1.5924, + "theoretical_loss": 3.8031520074025096, + "tokens_seen": 708771840 + }, + { + "epoch": 0.21, + "learning_rate": 0.0006345275827482447, + "loss": 1.3423, + "theoretical_loss": 3.8031166493200788, + "tokens_seen": 708837376 + }, + { + "epoch": 0.21, + "learning_rate": 0.0006345115346038115, + "loss": 1.5165, + "theoretical_loss": 3.803081295421788, + "tokens_seen": 708902912 + }, + { + "epoch": 0.21, + "learning_rate": 0.0006344954864593781, + "loss": 1.4578, + "theoretical_loss": 3.8030459457067574, + "tokens_seen": 708968448 + }, + { + "epoch": 0.21, + "learning_rate": 0.0006344794383149449, + "loss": 1.6347, + "theoretical_loss": 3.803010600174103, + "tokens_seen": 709033984 + }, + { + "epoch": 0.21, + "learning_rate": 0.0006344633901705115, + "loss": 1.5417, + "theoretical_loss": 3.8029752588229453, + "tokens_seen": 709099520 + }, + { + "epoch": 0.21, + "learning_rate": 0.0006344473420260783, + "loss": 1.4819, + "theoretical_loss": 3.8029399216524027, + "tokens_seen": 709165056 + }, + { + "epoch": 0.21, + "learning_rate": 0.0006344312938816449, + "loss": 1.4832, + "theoretical_loss": 3.802904588661594, + "tokens_seen": 709230592 + }, + { + "epoch": 0.21, + "learning_rate": 0.0006344152457372117, + "loss": 1.599, + "theoretical_loss": 3.802869259849639, + "tokens_seen": 709296128 + }, + { + "epoch": 0.21, + "learning_rate": 0.0006343991975927783, + "loss": 1.5458, + "theoretical_loss": 3.8028339352156575, + "tokens_seen": 709361664 + }, + { + "epoch": 0.21, + "learning_rate": 0.0006343831494483451, + "loss": 1.3965, + "theoretical_loss": 3.802798614758769, + "tokens_seen": 709427200 + }, + { + "epoch": 0.21, + "learning_rate": 0.0006343671013039117, + "loss": 1.6047, + "theoretical_loss": 3.8027632984780944, + "tokens_seen": 709492736 + }, + { + "epoch": 0.22, + "learning_rate": 0.0006343510531594785, + "loss": 1.5264, + "theoretical_loss": 3.8027279863727537, + "tokens_seen": 709558272 + }, + { + "epoch": 0.22, + "learning_rate": 0.0006343350050150451, + "loss": 1.5007, + "theoretical_loss": 3.8026926784418675, + "tokens_seen": 709623808 + }, + { + "epoch": 0.22, + "learning_rate": 0.0006343189568706119, + "loss": 1.5915, + "theoretical_loss": 3.802657374684558, + "tokens_seen": 709689344 + }, + { + "epoch": 0.22, + "learning_rate": 0.0006343029087261785, + "loss": 1.5334, + "theoretical_loss": 3.8026220750999453, + "tokens_seen": 709754880 + }, + { + "epoch": 0.22, + "learning_rate": 0.0006342868605817453, + "loss": 1.479, + "theoretical_loss": 3.802586779687152, + "tokens_seen": 709820416 + }, + { + "epoch": 0.22, + "learning_rate": 0.0006342708124373119, + "loss": 1.6034, + "theoretical_loss": 3.8025514884452987, + "tokens_seen": 709885952 + }, + { + "epoch": 0.22, + "learning_rate": 0.0006342547642928787, + "loss": 1.4612, + "theoretical_loss": 3.8025162013735088, + "tokens_seen": 709951488 + }, + { + "epoch": 0.22, + "learning_rate": 0.0006342387161484453, + "loss": 1.5766, + "theoretical_loss": 3.802480918470904, + "tokens_seen": 710017024 + }, + { + "epoch": 0.22, + "learning_rate": 0.0006342226680040121, + "loss": 1.4579, + "theoretical_loss": 3.8024456397366073, + "tokens_seen": 710082560 + }, + { + "epoch": 0.22, + "learning_rate": 0.0006342066198595788, + "loss": 1.4975, + "theoretical_loss": 3.802410365169741, + "tokens_seen": 710148096 + }, + { + "epoch": 0.22, + "learning_rate": 0.0006341905717151455, + "loss": 1.6207, + "theoretical_loss": 3.802375094769429, + "tokens_seen": 710213632 + }, + { + "epoch": 0.22, + "learning_rate": 0.0006341745235707122, + "loss": 1.6306, + "theoretical_loss": 3.8023398285347945, + "tokens_seen": 710279168 + }, + { + "epoch": 0.22, + "learning_rate": 0.0006341584754262789, + "loss": 1.5546, + "theoretical_loss": 3.8023045664649606, + "tokens_seen": 710344704 + }, + { + "epoch": 0.22, + "learning_rate": 0.0006341424272818456, + "loss": 1.5343, + "theoretical_loss": 3.802269308559052, + "tokens_seen": 710410240 + }, + { + "epoch": 0.22, + "learning_rate": 0.0006341263791374123, + "loss": 1.4644, + "theoretical_loss": 3.8022340548161933, + "tokens_seen": 710475776 + }, + { + "epoch": 0.22, + "learning_rate": 0.000634110330992979, + "loss": 1.6484, + "theoretical_loss": 3.8021988052355082, + "tokens_seen": 710541312 + }, + { + "epoch": 0.22, + "learning_rate": 0.0006340942828485457, + "loss": 1.5734, + "theoretical_loss": 3.8021635598161216, + "tokens_seen": 710606848 + }, + { + "epoch": 0.22, + "learning_rate": 0.0006340782347041124, + "loss": 1.5698, + "theoretical_loss": 3.802128318557159, + "tokens_seen": 710672384 + }, + { + "epoch": 0.22, + "learning_rate": 0.0006340621865596791, + "loss": 1.6125, + "theoretical_loss": 3.802093081457745, + "tokens_seen": 710737920 + }, + { + "epoch": 0.22, + "learning_rate": 0.0006340461384152458, + "loss": 1.5731, + "theoretical_loss": 3.8020578485170056, + "tokens_seen": 710803456 + }, + { + "epoch": 0.22, + "learning_rate": 0.0006340300902708125, + "loss": 1.448, + "theoretical_loss": 3.8020226197340667, + "tokens_seen": 710868992 + }, + { + "epoch": 0.22, + "learning_rate": 0.0006340140421263792, + "loss": 1.5919, + "theoretical_loss": 3.801987395108055, + "tokens_seen": 710934528 + }, + { + "epoch": 0.22, + "learning_rate": 0.0006339979939819459, + "loss": 1.525, + "theoretical_loss": 3.8019521746380955, + "tokens_seen": 711000064 + }, + { + "epoch": 0.22, + "objective/train/docs_used": 263840, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.5781018733978271, + "objective/train/theoretical_loss": 3.801916958323315, + "objective/train/tokens_used": 731525600, + "theoretical_loss": 3.801916958323315, + "tokens_seen": 711065600 + }, + { + "epoch": 0.22, + "learning_rate": 0.0006339819458375126, + "loss": 1.5384, + "theoretical_loss": 3.801916958323315, + "tokens_seen": 711065600 + }, + { + "epoch": 0.22, + "learning_rate": 0.0006339658976930793, + "loss": 1.4129, + "theoretical_loss": 3.8018817461628416, + "tokens_seen": 711131136 + }, + { + "epoch": 0.22, + "learning_rate": 0.000633949849548646, + "loss": 1.5245, + "theoretical_loss": 3.801846538155802, + "tokens_seen": 711196672 + }, + { + "epoch": 0.22, + "learning_rate": 0.0006339338014042127, + "loss": 1.5924, + "theoretical_loss": 3.801811334301323, + "tokens_seen": 711262208 + }, + { + "epoch": 0.22, + "learning_rate": 0.0006339177532597794, + "loss": 1.5082, + "theoretical_loss": 3.8017761345985326, + "tokens_seen": 711327744 + }, + { + "epoch": 0.22, + "learning_rate": 0.0006339017051153461, + "loss": 1.5598, + "theoretical_loss": 3.8017409390465584, + "tokens_seen": 711393280 + }, + { + "epoch": 0.22, + "learning_rate": 0.0006338856569709128, + "loss": 1.5745, + "theoretical_loss": 3.8017057476445295, + "tokens_seen": 711458816 + }, + { + "epoch": 0.22, + "learning_rate": 0.0006338696088264795, + "loss": 1.5885, + "theoretical_loss": 3.8016705603915737, + "tokens_seen": 711524352 + }, + { + "epoch": 0.22, + "learning_rate": 0.0006338535606820462, + "loss": 1.5219, + "theoretical_loss": 3.8016353772868197, + "tokens_seen": 711589888 + }, + { + "epoch": 0.22, + "learning_rate": 0.0006338375125376129, + "loss": 1.4178, + "theoretical_loss": 3.801600198329397, + "tokens_seen": 711655424 + }, + { + "epoch": 0.22, + "learning_rate": 0.0006338214643931796, + "loss": 1.4547, + "theoretical_loss": 3.8015650235184344, + "tokens_seen": 711720960 + }, + { + "epoch": 0.22, + "learning_rate": 0.0006338054162487463, + "loss": 1.5014, + "theoretical_loss": 3.8015298528530614, + "tokens_seen": 711786496 + }, + { + "epoch": 0.22, + "learning_rate": 0.000633789368104313, + "loss": 1.5159, + "theoretical_loss": 3.801494686332408, + "tokens_seen": 711852032 + }, + { + "epoch": 0.22, + "learning_rate": 0.0006337733199598797, + "loss": 1.4882, + "theoretical_loss": 3.8014595239556046, + "tokens_seen": 711917568 + }, + { + "epoch": 0.22, + "learning_rate": 0.0006337572718154464, + "loss": 1.6161, + "theoretical_loss": 3.8014243657217808, + "tokens_seen": 711983104 + }, + { + "epoch": 0.22, + "learning_rate": 0.0006337412236710131, + "loss": 1.5037, + "theoretical_loss": 3.8013892116300676, + "tokens_seen": 712048640 + }, + { + "epoch": 0.22, + "learning_rate": 0.0006337251755265798, + "loss": 1.4582, + "theoretical_loss": 3.8013540616795956, + "tokens_seen": 712114176 + }, + { + "epoch": 0.22, + "learning_rate": 0.0006337091273821465, + "loss": 1.5025, + "theoretical_loss": 3.801318915869496, + "tokens_seen": 712179712 + }, + { + "epoch": 0.22, + "learning_rate": 0.0006336930792377132, + "loss": 1.4883, + "theoretical_loss": 3.8012837741989003, + "tokens_seen": 712245248 + }, + { + "epoch": 0.22, + "learning_rate": 0.0006336770310932799, + "loss": 1.5233, + "theoretical_loss": 3.80124863666694, + "tokens_seen": 712310784 + }, + { + "epoch": 0.22, + "learning_rate": 0.0006336609829488466, + "loss": 1.4111, + "theoretical_loss": 3.8012135032727468, + "tokens_seen": 712376320 + }, + { + "epoch": 0.22, + "learning_rate": 0.0006336449348044133, + "loss": 1.4726, + "theoretical_loss": 3.8011783740154534, + "tokens_seen": 712441856 + }, + { + "epoch": 0.22, + "learning_rate": 0.00063362888665998, + "loss": 1.4755, + "theoretical_loss": 3.801143248894192, + "tokens_seen": 712507392 + }, + { + "epoch": 0.22, + "learning_rate": 0.0006336128385155467, + "loss": 1.4963, + "theoretical_loss": 3.801108127908095, + "tokens_seen": 712572928 + }, + { + "epoch": 0.22, + "learning_rate": 0.0006335967903711134, + "loss": 1.4467, + "theoretical_loss": 3.801073011056295, + "tokens_seen": 712638464 + }, + { + "epoch": 0.22, + "learning_rate": 0.0006335807422266801, + "loss": 1.6029, + "theoretical_loss": 3.801037898337926, + "tokens_seen": 712704000 + }, + { + "epoch": 0.22, + "learning_rate": 0.0006335646940822468, + "loss": 1.4027, + "theoretical_loss": 3.801002789752121, + "tokens_seen": 712769536 + }, + { + "epoch": 0.22, + "learning_rate": 0.0006335486459378135, + "loss": 1.6003, + "theoretical_loss": 3.800967685298014, + "tokens_seen": 712835072 + }, + { + "epoch": 0.22, + "learning_rate": 0.0006335325977933801, + "loss": 1.4386, + "theoretical_loss": 3.8009325849747384, + "tokens_seen": 712900608 + }, + { + "epoch": 0.22, + "learning_rate": 0.0006335165496489469, + "loss": 1.5602, + "theoretical_loss": 3.8008974887814295, + "tokens_seen": 712966144 + }, + { + "epoch": 0.22, + "learning_rate": 0.0006335005015045135, + "loss": 1.6244, + "theoretical_loss": 3.8008623967172204, + "tokens_seen": 713031680 + }, + { + "epoch": 0.22, + "learning_rate": 0.0006334844533600803, + "loss": 1.4495, + "theoretical_loss": 3.800827308781247, + "tokens_seen": 713097216 + }, + { + "epoch": 0.22, + "learning_rate": 0.0006334684052156469, + "loss": 1.4226, + "theoretical_loss": 3.8007922249726445, + "tokens_seen": 713162752 + }, + { + "epoch": 0.22, + "learning_rate": 0.0006334523570712137, + "loss": 1.4291, + "theoretical_loss": 3.8007571452905466, + "tokens_seen": 713228288 + }, + { + "epoch": 0.22, + "learning_rate": 0.0006334363089267803, + "loss": 1.5495, + "theoretical_loss": 3.800722069734091, + "tokens_seen": 713293824 + }, + { + "epoch": 0.22, + "learning_rate": 0.000633420260782347, + "loss": 1.5932, + "theoretical_loss": 3.800686998302411, + "tokens_seen": 713359360 + }, + { + "epoch": 0.22, + "learning_rate": 0.0006334042126379137, + "loss": 1.4963, + "theoretical_loss": 3.8006519309946447, + "tokens_seen": 713424896 + }, + { + "epoch": 0.22, + "learning_rate": 0.0006333881644934804, + "loss": 1.5305, + "theoretical_loss": 3.8006168678099277, + "tokens_seen": 713490432 + }, + { + "epoch": 0.22, + "learning_rate": 0.0006333721163490471, + "loss": 1.5242, + "theoretical_loss": 3.800581808747397, + "tokens_seen": 713555968 + }, + { + "epoch": 0.22, + "learning_rate": 0.0006333560682046138, + "loss": 1.3578, + "theoretical_loss": 3.800546753806189, + "tokens_seen": 713621504 + }, + { + "epoch": 0.22, + "learning_rate": 0.0006333400200601805, + "loss": 1.5029, + "theoretical_loss": 3.8005117029854403, + "tokens_seen": 713687040 + }, + { + "epoch": 0.22, + "learning_rate": 0.0006333239719157472, + "loss": 1.5272, + "theoretical_loss": 3.8004766562842893, + "tokens_seen": 713752576 + }, + { + "epoch": 0.22, + "learning_rate": 0.000633307923771314, + "loss": 1.4354, + "theoretical_loss": 3.800441613701873, + "tokens_seen": 713818112 + }, + { + "epoch": 0.22, + "learning_rate": 0.0006332918756268806, + "loss": 1.5719, + "theoretical_loss": 3.8004065752373295, + "tokens_seen": 713883648 + }, + { + "epoch": 0.22, + "learning_rate": 0.0006332758274824473, + "loss": 1.556, + "theoretical_loss": 3.800371540889797, + "tokens_seen": 713949184 + }, + { + "epoch": 0.22, + "learning_rate": 0.000633259779338014, + "loss": 1.4916, + "theoretical_loss": 3.800336510658414, + "tokens_seen": 714014720 + }, + { + "epoch": 0.22, + "learning_rate": 0.0006332437311935807, + "loss": 1.5208, + "theoretical_loss": 3.8003014845423184, + "tokens_seen": 714080256 + }, + { + "epoch": 0.22, + "learning_rate": 0.0006332276830491474, + "loss": 1.5515, + "theoretical_loss": 3.80026646254065, + "tokens_seen": 714145792 + }, + { + "epoch": 0.22, + "learning_rate": 0.0006332116349047141, + "loss": 1.5036, + "theoretical_loss": 3.800231444652548, + "tokens_seen": 714211328 + }, + { + "epoch": 0.22, + "learning_rate": 0.0006331955867602808, + "loss": 1.5628, + "theoretical_loss": 3.800196430877151, + "tokens_seen": 714276864 + }, + { + "epoch": 0.22, + "objective/train/docs_used": 265036, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.4089895486831665, + "objective/train/theoretical_loss": 3.8001614212135997, + "objective/train/tokens_used": 734802400, + "theoretical_loss": 3.8001614212135997, + "tokens_seen": 714342400 + }, + { + "epoch": 0.22, + "learning_rate": 0.0006331795386158476, + "loss": 1.5181, + "theoretical_loss": 3.8001614212135997, + "tokens_seen": 714342400 + }, + { + "epoch": 0.22, + "learning_rate": 0.0006331634904714142, + "loss": 1.5305, + "theoretical_loss": 3.8001264156610333, + "tokens_seen": 714407936 + }, + { + "epoch": 0.22, + "learning_rate": 0.000633147442326981, + "loss": 1.4355, + "theoretical_loss": 3.800091414218592, + "tokens_seen": 714473472 + }, + { + "epoch": 0.22, + "learning_rate": 0.0006331313941825476, + "loss": 1.5374, + "theoretical_loss": 3.800056416885417, + "tokens_seen": 714539008 + }, + { + "epoch": 0.22, + "learning_rate": 0.0006331153460381144, + "loss": 1.4368, + "theoretical_loss": 3.8000214236606484, + "tokens_seen": 714604544 + }, + { + "epoch": 0.22, + "learning_rate": 0.000633099297893681, + "loss": 1.4618, + "theoretical_loss": 3.7999864345434275, + "tokens_seen": 714670080 + }, + { + "epoch": 0.22, + "learning_rate": 0.0006330832497492478, + "loss": 1.486, + "theoretical_loss": 3.7999514495328954, + "tokens_seen": 714735616 + }, + { + "epoch": 0.22, + "learning_rate": 0.0006330672016048144, + "loss": 1.4757, + "theoretical_loss": 3.7999164686281937, + "tokens_seen": 714801152 + }, + { + "epoch": 0.22, + "learning_rate": 0.0006330511534603812, + "loss": 1.5424, + "theoretical_loss": 3.7998814918284642, + "tokens_seen": 714866688 + }, + { + "epoch": 0.22, + "learning_rate": 0.0006330351053159478, + "loss": 1.4907, + "theoretical_loss": 3.7998465191328483, + "tokens_seen": 714932224 + }, + { + "epoch": 0.22, + "learning_rate": 0.0006330190571715146, + "loss": 1.4308, + "theoretical_loss": 3.7998115505404892, + "tokens_seen": 714997760 + }, + { + "epoch": 0.22, + "learning_rate": 0.0006330030090270813, + "loss": 1.6747, + "theoretical_loss": 3.7997765860505295, + "tokens_seen": 715063296 + }, + { + "epoch": 0.22, + "learning_rate": 0.000632986960882648, + "loss": 1.512, + "theoretical_loss": 3.7997416256621106, + "tokens_seen": 715128832 + }, + { + "epoch": 0.22, + "learning_rate": 0.0006329709127382147, + "loss": 1.5533, + "theoretical_loss": 3.7997066693743773, + "tokens_seen": 715194368 + }, + { + "epoch": 0.22, + "learning_rate": 0.0006329548645937814, + "loss": 1.4856, + "theoretical_loss": 3.799671717186472, + "tokens_seen": 715259904 + }, + { + "epoch": 0.22, + "learning_rate": 0.0006329388164493481, + "loss": 1.4911, + "theoretical_loss": 3.799636769097538, + "tokens_seen": 715325440 + }, + { + "epoch": 0.22, + "learning_rate": 0.0006329227683049148, + "loss": 1.5609, + "theoretical_loss": 3.7996018251067203, + "tokens_seen": 715390976 + }, + { + "epoch": 0.22, + "learning_rate": 0.0006329067201604815, + "loss": 1.5392, + "theoretical_loss": 3.799566885213162, + "tokens_seen": 715456512 + }, + { + "epoch": 0.22, + "learning_rate": 0.0006328906720160482, + "loss": 1.5241, + "theoretical_loss": 3.799531949416007, + "tokens_seen": 715522048 + }, + { + "epoch": 0.22, + "learning_rate": 0.0006328746238716149, + "loss": 1.4401, + "theoretical_loss": 3.799497017714401, + "tokens_seen": 715587584 + }, + { + "epoch": 0.22, + "learning_rate": 0.0006328585757271816, + "loss": 1.4284, + "theoretical_loss": 3.7994620901074887, + "tokens_seen": 715653120 + }, + { + "epoch": 0.22, + "learning_rate": 0.0006328425275827483, + "loss": 1.5675, + "theoretical_loss": 3.7994271665944144, + "tokens_seen": 715718656 + }, + { + "epoch": 0.22, + "learning_rate": 0.000632826479438315, + "loss": 1.4545, + "theoretical_loss": 3.799392247174324, + "tokens_seen": 715784192 + }, + { + "epoch": 0.22, + "learning_rate": 0.0006328104312938817, + "loss": 1.4523, + "theoretical_loss": 3.799357331846364, + "tokens_seen": 715849728 + }, + { + "epoch": 0.22, + "learning_rate": 0.0006327943831494484, + "loss": 1.542, + "theoretical_loss": 3.7993224206096787, + "tokens_seen": 715915264 + }, + { + "epoch": 0.22, + "learning_rate": 0.0006327783350050151, + "loss": 1.5578, + "theoretical_loss": 3.7992875134634154, + "tokens_seen": 715980800 + }, + { + "epoch": 0.22, + "learning_rate": 0.0006327622868605818, + "loss": 1.5097, + "theoretical_loss": 3.79925261040672, + "tokens_seen": 716046336 + }, + { + "epoch": 0.22, + "learning_rate": 0.0006327462387161485, + "loss": 1.4906, + "theoretical_loss": 3.7992177114387387, + "tokens_seen": 716111872 + }, + { + "epoch": 0.22, + "learning_rate": 0.0006327301905717152, + "loss": 1.4799, + "theoretical_loss": 3.799182816558619, + "tokens_seen": 716177408 + }, + { + "epoch": 0.22, + "learning_rate": 0.0006327141424272819, + "loss": 1.4705, + "theoretical_loss": 3.7991479257655083, + "tokens_seen": 716242944 + }, + { + "epoch": 0.22, + "learning_rate": 0.0006326980942828486, + "loss": 1.453, + "theoretical_loss": 3.7991130390585535, + "tokens_seen": 716308480 + }, + { + "epoch": 0.22, + "learning_rate": 0.0006326820461384153, + "loss": 1.3782, + "theoretical_loss": 3.7990781564369023, + "tokens_seen": 716374016 + }, + { + "epoch": 0.22, + "learning_rate": 0.000632665997993982, + "loss": 1.4785, + "theoretical_loss": 3.7990432778997034, + "tokens_seen": 716439552 + }, + { + "epoch": 0.22, + "learning_rate": 0.0006326499498495487, + "loss": 1.5326, + "theoretical_loss": 3.799008403446104, + "tokens_seen": 716505088 + }, + { + "epoch": 0.22, + "learning_rate": 0.0006326339017051154, + "loss": 1.4742, + "theoretical_loss": 3.798973533075253, + "tokens_seen": 716570624 + }, + { + "epoch": 0.22, + "learning_rate": 0.0006326178535606821, + "loss": 1.5146, + "theoretical_loss": 3.798938666786299, + "tokens_seen": 716636160 + }, + { + "epoch": 0.22, + "learning_rate": 0.0006326018054162488, + "loss": 1.6421, + "theoretical_loss": 3.798903804578391, + "tokens_seen": 716701696 + }, + { + "epoch": 0.22, + "learning_rate": 0.0006325857572718155, + "loss": 1.5061, + "theoretical_loss": 3.7988689464506784, + "tokens_seen": 716767232 + }, + { + "epoch": 0.22, + "learning_rate": 0.0006325697091273822, + "loss": 1.4326, + "theoretical_loss": 3.798834092402311, + "tokens_seen": 716832768 + }, + { + "epoch": 0.22, + "learning_rate": 0.0006325536609829489, + "loss": 1.4879, + "theoretical_loss": 3.7987992424324366, + "tokens_seen": 716898304 + }, + { + "epoch": 0.22, + "learning_rate": 0.0006325376128385156, + "loss": 1.5658, + "theoretical_loss": 3.7987643965402076, + "tokens_seen": 716963840 + }, + { + "epoch": 0.22, + "learning_rate": 0.0006325215646940823, + "loss": 1.5247, + "theoretical_loss": 3.7987295547247735, + "tokens_seen": 717029376 + }, + { + "epoch": 0.22, + "learning_rate": 0.000632505516549649, + "loss": 1.5713, + "theoretical_loss": 3.7986947169852834, + "tokens_seen": 717094912 + }, + { + "epoch": 0.22, + "learning_rate": 0.0006324894684052157, + "loss": 1.586, + "theoretical_loss": 3.79865988332089, + "tokens_seen": 717160448 + }, + { + "epoch": 0.22, + "learning_rate": 0.0006324734202607824, + "loss": 1.5683, + "theoretical_loss": 3.798625053730743, + "tokens_seen": 717225984 + }, + { + "epoch": 0.22, + "learning_rate": 0.0006324573721163491, + "loss": 1.5381, + "theoretical_loss": 3.7985902282139943, + "tokens_seen": 717291520 + }, + { + "epoch": 0.22, + "learning_rate": 0.0006324413239719158, + "loss": 1.5479, + "theoretical_loss": 3.7985554067697946, + "tokens_seen": 717357056 + }, + { + "epoch": 0.22, + "learning_rate": 0.0006324252758274825, + "loss": 1.5184, + "theoretical_loss": 3.798520589397297, + "tokens_seen": 717422592 + }, + { + "epoch": 0.22, + "learning_rate": 0.0006324092276830492, + "loss": 1.6706, + "theoretical_loss": 3.798485776095652, + "tokens_seen": 717488128 + }, + { + "epoch": 0.22, + "learning_rate": 0.0006323931795386159, + "loss": 1.4568, + "theoretical_loss": 3.7984509668640136, + "tokens_seen": 717553664 + }, + { + "epoch": 0.22, + "objective/train/docs_used": 266305, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.5267771482467651, + "objective/train/theoretical_loss": 3.7984161617015326, + "objective/train/tokens_used": 738079200, + "theoretical_loss": 3.7984161617015326, + "tokens_seen": 717619200 + }, + { + "epoch": 0.22, + "learning_rate": 0.0006323771313941826, + "loss": 1.4645, + "theoretical_loss": 3.7984161617015326, + "tokens_seen": 717619200 + }, + { + "epoch": 0.22, + "learning_rate": 0.0006323610832497493, + "loss": 1.5432, + "theoretical_loss": 3.7983813606073626, + "tokens_seen": 717684736 + }, + { + "epoch": 0.22, + "learning_rate": 0.000632345035105316, + "loss": 1.5235, + "theoretical_loss": 3.7983465635806564, + "tokens_seen": 717750272 + }, + { + "epoch": 0.22, + "learning_rate": 0.0006323289869608827, + "loss": 1.6257, + "theoretical_loss": 3.7983117706205674, + "tokens_seen": 717815808 + }, + { + "epoch": 0.22, + "learning_rate": 0.0006323129388164494, + "loss": 1.6198, + "theoretical_loss": 3.798276981726249, + "tokens_seen": 717881344 + }, + { + "epoch": 0.22, + "learning_rate": 0.0006322968906720161, + "loss": 1.4868, + "theoretical_loss": 3.798242196896856, + "tokens_seen": 717946880 + }, + { + "epoch": 0.22, + "learning_rate": 0.0006322808425275828, + "loss": 1.5499, + "theoretical_loss": 3.798207416131541, + "tokens_seen": 718012416 + }, + { + "epoch": 0.22, + "learning_rate": 0.0006322647943831495, + "loss": 1.5092, + "theoretical_loss": 3.798172639429459, + "tokens_seen": 718077952 + }, + { + "epoch": 0.22, + "learning_rate": 0.0006322487462387162, + "loss": 1.4793, + "theoretical_loss": 3.798137866789764, + "tokens_seen": 718143488 + }, + { + "epoch": 0.22, + "learning_rate": 0.0006322326980942829, + "loss": 1.5161, + "theoretical_loss": 3.798103098211611, + "tokens_seen": 718209024 + }, + { + "epoch": 0.22, + "learning_rate": 0.0006322166499498496, + "loss": 1.4985, + "theoretical_loss": 3.798068333694156, + "tokens_seen": 718274560 + }, + { + "epoch": 0.22, + "learning_rate": 0.0006322006018054163, + "loss": 1.5309, + "theoretical_loss": 3.798033573236553, + "tokens_seen": 718340096 + }, + { + "epoch": 0.22, + "learning_rate": 0.000632184553660983, + "loss": 1.4502, + "theoretical_loss": 3.797998816837959, + "tokens_seen": 718405632 + }, + { + "epoch": 0.22, + "learning_rate": 0.0006321685055165497, + "loss": 1.5254, + "theoretical_loss": 3.7979640644975277, + "tokens_seen": 718471168 + }, + { + "epoch": 0.22, + "learning_rate": 0.0006321524573721164, + "loss": 1.4513, + "theoretical_loss": 3.797929316214417, + "tokens_seen": 718536704 + }, + { + "epoch": 0.22, + "learning_rate": 0.000632136409227683, + "loss": 1.3996, + "theoretical_loss": 3.797894571987783, + "tokens_seen": 718602240 + }, + { + "epoch": 0.22, + "learning_rate": 0.0006321203610832498, + "loss": 1.4706, + "theoretical_loss": 3.7978598318167816, + "tokens_seen": 718667776 + }, + { + "epoch": 0.22, + "learning_rate": 0.0006321043129388164, + "loss": 1.5038, + "theoretical_loss": 3.7978250957005697, + "tokens_seen": 718733312 + }, + { + "epoch": 0.22, + "learning_rate": 0.0006320882647943832, + "loss": 1.5089, + "theoretical_loss": 3.7977903636383044, + "tokens_seen": 718798848 + }, + { + "epoch": 0.22, + "learning_rate": 0.0006320722166499498, + "loss": 1.4352, + "theoretical_loss": 3.7977556356291435, + "tokens_seen": 718864384 + }, + { + "epoch": 0.22, + "learning_rate": 0.0006320561685055166, + "loss": 1.5556, + "theoretical_loss": 3.7977209116722443, + "tokens_seen": 718929920 + }, + { + "epoch": 0.22, + "learning_rate": 0.0006320401203610832, + "loss": 1.4796, + "theoretical_loss": 3.797686191766765, + "tokens_seen": 718995456 + }, + { + "epoch": 0.22, + "learning_rate": 0.00063202407221665, + "loss": 1.4864, + "theoretical_loss": 3.797651475911863, + "tokens_seen": 719060992 + }, + { + "epoch": 0.22, + "learning_rate": 0.0006320080240722166, + "loss": 1.5151, + "theoretical_loss": 3.7976167641066962, + "tokens_seen": 719126528 + }, + { + "epoch": 0.22, + "learning_rate": 0.0006319919759277833, + "loss": 1.4394, + "theoretical_loss": 3.7975820563504246, + "tokens_seen": 719192064 + }, + { + "epoch": 0.22, + "learning_rate": 0.00063197592778335, + "loss": 1.6132, + "theoretical_loss": 3.797547352642206, + "tokens_seen": 719257600 + }, + { + "epoch": 0.22, + "learning_rate": 0.0006319598796389167, + "loss": 1.4798, + "theoretical_loss": 3.7975126529812, + "tokens_seen": 719323136 + }, + { + "epoch": 0.22, + "learning_rate": 0.0006319438314944834, + "loss": 1.4648, + "theoretical_loss": 3.797477957366566, + "tokens_seen": 719388672 + }, + { + "epoch": 0.22, + "learning_rate": 0.0006319277833500501, + "loss": 1.5074, + "theoretical_loss": 3.797443265797463, + "tokens_seen": 719454208 + }, + { + "epoch": 0.22, + "learning_rate": 0.0006319117352056168, + "loss": 1.6058, + "theoretical_loss": 3.7974085782730516, + "tokens_seen": 719519744 + }, + { + "epoch": 0.22, + "learning_rate": 0.0006318956870611835, + "loss": 1.4043, + "theoretical_loss": 3.797373894792491, + "tokens_seen": 719585280 + }, + { + "epoch": 0.22, + "learning_rate": 0.0006318796389167502, + "loss": 1.4919, + "theoretical_loss": 3.797339215354942, + "tokens_seen": 719650816 + }, + { + "epoch": 0.22, + "learning_rate": 0.0006318635907723169, + "loss": 1.566, + "theoretical_loss": 3.7973045399595655, + "tokens_seen": 719716352 + }, + { + "epoch": 0.22, + "learning_rate": 0.0006318475426278836, + "loss": 1.5055, + "theoretical_loss": 3.797269868605522, + "tokens_seen": 719781888 + }, + { + "epoch": 0.22, + "learning_rate": 0.0006318314944834504, + "loss": 1.484, + "theoretical_loss": 3.797235201291973, + "tokens_seen": 719847424 + }, + { + "epoch": 0.22, + "learning_rate": 0.000631815446339017, + "loss": 1.5326, + "theoretical_loss": 3.797200538018079, + "tokens_seen": 719912960 + }, + { + "epoch": 0.22, + "learning_rate": 0.0006317993981945838, + "loss": 1.4973, + "theoretical_loss": 3.7971658787830025, + "tokens_seen": 719978496 + }, + { + "epoch": 0.22, + "learning_rate": 0.0006317833500501505, + "loss": 1.5236, + "theoretical_loss": 3.7971312235859047, + "tokens_seen": 720044032 + }, + { + "epoch": 0.22, + "learning_rate": 0.0006317673019057172, + "loss": 1.588, + "theoretical_loss": 3.797096572425948, + "tokens_seen": 720109568 + }, + { + "epoch": 0.22, + "learning_rate": 0.0006317512537612839, + "loss": 1.5524, + "theoretical_loss": 3.7970619253022946, + "tokens_seen": 720175104 + }, + { + "epoch": 0.22, + "learning_rate": 0.0006317352056168506, + "loss": 1.4277, + "theoretical_loss": 3.797027282214107, + "tokens_seen": 720240640 + }, + { + "epoch": 0.22, + "learning_rate": 0.0006317191574724173, + "loss": 1.4305, + "theoretical_loss": 3.7969926431605483, + "tokens_seen": 720306176 + }, + { + "epoch": 0.22, + "learning_rate": 0.000631703109327984, + "loss": 1.4264, + "theoretical_loss": 3.796958008140782, + "tokens_seen": 720371712 + }, + { + "epoch": 0.22, + "learning_rate": 0.0006316870611835507, + "loss": 1.5, + "theoretical_loss": 3.7969233771539708, + "tokens_seen": 720437248 + }, + { + "epoch": 0.22, + "learning_rate": 0.0006316710130391174, + "loss": 1.5351, + "theoretical_loss": 3.796888750199278, + "tokens_seen": 720502784 + }, + { + "epoch": 0.22, + "learning_rate": 0.0006316549648946841, + "loss": 1.4909, + "theoretical_loss": 3.796854127275868, + "tokens_seen": 720568320 + }, + { + "epoch": 0.22, + "learning_rate": 0.0006316389167502508, + "loss": 1.5492, + "theoretical_loss": 3.796819508382905, + "tokens_seen": 720633856 + }, + { + "epoch": 0.22, + "learning_rate": 0.0006316228686058175, + "loss": 1.4849, + "theoretical_loss": 3.796784893519553, + "tokens_seen": 720699392 + }, + { + "epoch": 0.22, + "learning_rate": 0.0006316068204613842, + "loss": 1.5994, + "theoretical_loss": 3.796750282684977, + "tokens_seen": 720764928 + }, + { + "epoch": 0.22, + "learning_rate": 0.0006315907723169509, + "loss": 1.3832, + "theoretical_loss": 3.7967156758783416, + "tokens_seen": 720830464 + }, + { + "epoch": 0.22, + "objective/train/docs_used": 267624, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.4001801013946533, + "objective/train/theoretical_loss": 3.7966810730988114, + "objective/train/tokens_used": 741356000, + "theoretical_loss": 3.7966810730988114, + "tokens_seen": 720896000 + }, + { + "epoch": 0.22, + "learning_rate": 0.0006315747241725176, + "loss": 1.3963, + "theoretical_loss": 3.7966810730988114, + "tokens_seen": 720896000 + }, + { + "epoch": 0.22, + "learning_rate": 0.0006315586760280843, + "loss": 1.4295, + "theoretical_loss": 3.7966464743455526, + "tokens_seen": 720961536 + }, + { + "epoch": 0.22, + "learning_rate": 0.000631542627883651, + "loss": 1.4973, + "theoretical_loss": 3.79661187961773, + "tokens_seen": 721027072 + }, + { + "epoch": 0.22, + "learning_rate": 0.0006315265797392177, + "loss": 1.4187, + "theoretical_loss": 3.79657728891451, + "tokens_seen": 721092608 + }, + { + "epoch": 0.22, + "learning_rate": 0.0006315105315947844, + "loss": 1.5136, + "theoretical_loss": 3.796542702235059, + "tokens_seen": 721158144 + }, + { + "epoch": 0.22, + "learning_rate": 0.0006314944834503511, + "loss": 1.4651, + "theoretical_loss": 3.7965081195785424, + "tokens_seen": 721223680 + }, + { + "epoch": 0.22, + "learning_rate": 0.0006314784353059178, + "loss": 1.5106, + "theoretical_loss": 3.7964735409441275, + "tokens_seen": 721289216 + }, + { + "epoch": 0.22, + "learning_rate": 0.0006314623871614845, + "loss": 1.4444, + "theoretical_loss": 3.7964389663309808, + "tokens_seen": 721354752 + }, + { + "epoch": 0.22, + "learning_rate": 0.0006314463390170512, + "loss": 1.5566, + "theoretical_loss": 3.796404395738269, + "tokens_seen": 721420288 + }, + { + "epoch": 0.22, + "learning_rate": 0.0006314302908726179, + "loss": 1.4808, + "theoretical_loss": 3.79636982916516, + "tokens_seen": 721485824 + }, + { + "epoch": 0.22, + "learning_rate": 0.0006314142427281846, + "loss": 1.509, + "theoretical_loss": 3.7963352666108214, + "tokens_seen": 721551360 + }, + { + "epoch": 0.22, + "learning_rate": 0.0006313981945837513, + "loss": 1.6075, + "theoretical_loss": 3.796300708074421, + "tokens_seen": 721616896 + }, + { + "epoch": 0.22, + "learning_rate": 0.000631382146439318, + "loss": 1.5845, + "theoretical_loss": 3.796266153555127, + "tokens_seen": 721682432 + }, + { + "epoch": 0.22, + "learning_rate": 0.0006313660982948847, + "loss": 1.5909, + "theoretical_loss": 3.7962316030521066, + "tokens_seen": 721747968 + }, + { + "epoch": 0.22, + "learning_rate": 0.0006313500501504514, + "loss": 1.468, + "theoretical_loss": 3.79619705656453, + "tokens_seen": 721813504 + }, + { + "epoch": 0.22, + "learning_rate": 0.0006313340020060181, + "loss": 1.4872, + "theoretical_loss": 3.7961625140915647, + "tokens_seen": 721879040 + }, + { + "epoch": 0.22, + "learning_rate": 0.0006313179538615848, + "loss": 1.4464, + "theoretical_loss": 3.7961279756323805, + "tokens_seen": 721944576 + }, + { + "epoch": 0.22, + "learning_rate": 0.0006313019057171515, + "loss": 1.5138, + "theoretical_loss": 3.7960934411861462, + "tokens_seen": 722010112 + }, + { + "epoch": 0.22, + "learning_rate": 0.0006312858575727182, + "loss": 1.5686, + "theoretical_loss": 3.7960589107520324, + "tokens_seen": 722075648 + }, + { + "epoch": 0.22, + "learning_rate": 0.0006312698094282849, + "loss": 1.4727, + "theoretical_loss": 3.7960243843292076, + "tokens_seen": 722141184 + }, + { + "epoch": 0.22, + "learning_rate": 0.0006312537612838516, + "loss": 1.5642, + "theoretical_loss": 3.795989861916843, + "tokens_seen": 722206720 + }, + { + "epoch": 0.22, + "learning_rate": 0.0006312377131394183, + "loss": 1.4785, + "theoretical_loss": 3.7959553435141076, + "tokens_seen": 722272256 + }, + { + "epoch": 0.22, + "learning_rate": 0.000631221664994985, + "loss": 1.5282, + "theoretical_loss": 3.795920829120173, + "tokens_seen": 722337792 + }, + { + "epoch": 0.22, + "learning_rate": 0.0006312056168505517, + "loss": 1.4078, + "theoretical_loss": 3.79588631873421, + "tokens_seen": 722403328 + }, + { + "epoch": 0.22, + "learning_rate": 0.0006311895687061184, + "loss": 1.4178, + "theoretical_loss": 3.7958518123553886, + "tokens_seen": 722468864 + }, + { + "epoch": 0.22, + "learning_rate": 0.0006311735205616851, + "loss": 1.5008, + "theoretical_loss": 3.7958173099828816, + "tokens_seen": 722534400 + }, + { + "epoch": 0.22, + "learning_rate": 0.0006311574724172518, + "loss": 1.5421, + "theoretical_loss": 3.795782811615859, + "tokens_seen": 722599936 + }, + { + "epoch": 0.22, + "learning_rate": 0.0006311414242728185, + "loss": 1.5057, + "theoretical_loss": 3.7957483172534934, + "tokens_seen": 722665472 + }, + { + "epoch": 0.22, + "learning_rate": 0.0006311253761283852, + "loss": 1.4763, + "theoretical_loss": 3.795713826894957, + "tokens_seen": 722731008 + }, + { + "epoch": 0.22, + "learning_rate": 0.0006311093279839519, + "loss": 1.6489, + "theoretical_loss": 3.7956793405394214, + "tokens_seen": 722796544 + }, + { + "epoch": 0.22, + "learning_rate": 0.0006310932798395186, + "loss": 1.597, + "theoretical_loss": 3.79564485818606, + "tokens_seen": 722862080 + }, + { + "epoch": 0.22, + "learning_rate": 0.0006310772316950853, + "loss": 1.5431, + "theoretical_loss": 3.7956103798340446, + "tokens_seen": 722927616 + }, + { + "epoch": 0.22, + "learning_rate": 0.000631061183550652, + "loss": 1.5219, + "theoretical_loss": 3.795575905482549, + "tokens_seen": 722993152 + }, + { + "epoch": 0.22, + "learning_rate": 0.0006310451354062187, + "loss": 1.4907, + "theoretical_loss": 3.7955414351307457, + "tokens_seen": 723058688 + }, + { + "epoch": 0.22, + "learning_rate": 0.0006310290872617854, + "loss": 1.5355, + "theoretical_loss": 3.7955069687778087, + "tokens_seen": 723124224 + }, + { + "epoch": 0.22, + "learning_rate": 0.0006310130391173521, + "loss": 1.5965, + "theoretical_loss": 3.795472506422912, + "tokens_seen": 723189760 + }, + { + "epoch": 0.22, + "learning_rate": 0.0006309969909729188, + "loss": 1.5552, + "theoretical_loss": 3.795438048065229, + "tokens_seen": 723255296 + }, + { + "epoch": 0.22, + "learning_rate": 0.0006309809428284855, + "loss": 1.5707, + "theoretical_loss": 3.7954035937039348, + "tokens_seen": 723320832 + }, + { + "epoch": 0.22, + "learning_rate": 0.0006309648946840522, + "loss": 1.5492, + "theoretical_loss": 3.7953691433382026, + "tokens_seen": 723386368 + }, + { + "epoch": 0.22, + "learning_rate": 0.0006309488465396189, + "loss": 1.5626, + "theoretical_loss": 3.795334696967208, + "tokens_seen": 723451904 + }, + { + "epoch": 0.22, + "learning_rate": 0.0006309327983951856, + "loss": 1.5147, + "theoretical_loss": 3.795300254590126, + "tokens_seen": 723517440 + }, + { + "epoch": 0.22, + "learning_rate": 0.0006309167502507523, + "loss": 1.4858, + "theoretical_loss": 3.7952658162061317, + "tokens_seen": 723582976 + }, + { + "epoch": 0.22, + "learning_rate": 0.000630900702106319, + "loss": 1.3867, + "theoretical_loss": 3.7952313818144, + "tokens_seen": 723648512 + }, + { + "epoch": 0.22, + "learning_rate": 0.0006308846539618857, + "loss": 1.5785, + "theoretical_loss": 3.7951969514141073, + "tokens_seen": 723714048 + }, + { + "epoch": 0.22, + "learning_rate": 0.0006308686058174524, + "loss": 1.4874, + "theoretical_loss": 3.795162525004429, + "tokens_seen": 723779584 + }, + { + "epoch": 0.22, + "learning_rate": 0.0006308525576730191, + "loss": 1.4685, + "theoretical_loss": 3.795128102584542, + "tokens_seen": 723845120 + }, + { + "epoch": 0.22, + "learning_rate": 0.0006308365095285858, + "loss": 1.4759, + "theoretical_loss": 3.7950936841536222, + "tokens_seen": 723910656 + }, + { + "epoch": 0.22, + "learning_rate": 0.0006308204613841525, + "loss": 1.5254, + "theoretical_loss": 3.7950592697108467, + "tokens_seen": 723976192 + }, + { + "epoch": 0.22, + "learning_rate": 0.0006308044132397192, + "loss": 1.5338, + "theoretical_loss": 3.795024859255392, + "tokens_seen": 724041728 + }, + { + "epoch": 0.22, + "learning_rate": 0.0006307883650952859, + "loss": 1.4998, + "theoretical_loss": 3.7949904527864353, + "tokens_seen": 724107264 + }, + { + "epoch": 0.22, + "objective/train/docs_used": 268786, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.5310757160186768, + "objective/train/theoretical_loss": 3.7949560503031545, + "objective/train/tokens_used": 744632800, + "theoretical_loss": 3.7949560503031545, + "tokens_seen": 724172800 + }, + { + "epoch": 0.22, + "learning_rate": 0.0006307723169508526, + "loss": 1.5615, + "theoretical_loss": 3.7949560503031545, + "tokens_seen": 724172800 + }, + { + "epoch": 0.22, + "learning_rate": 0.0006307562688064193, + "loss": 1.5639, + "theoretical_loss": 3.794921651804727, + "tokens_seen": 724238336 + }, + { + "epoch": 0.22, + "learning_rate": 0.0006307402206619859, + "loss": 1.7148, + "theoretical_loss": 3.7948872572903305, + "tokens_seen": 724303872 + }, + { + "epoch": 0.22, + "learning_rate": 0.0006307241725175527, + "loss": 1.4425, + "theoretical_loss": 3.7948528667591432, + "tokens_seen": 724369408 + }, + { + "epoch": 0.22, + "learning_rate": 0.0006307081243731193, + "loss": 1.4723, + "theoretical_loss": 3.7948184802103437, + "tokens_seen": 724434944 + }, + { + "epoch": 0.22, + "learning_rate": 0.0006306920762286861, + "loss": 1.5323, + "theoretical_loss": 3.7947840976431104, + "tokens_seen": 724500480 + }, + { + "epoch": 0.22, + "learning_rate": 0.0006306760280842527, + "loss": 1.4397, + "theoretical_loss": 3.7947497190566226, + "tokens_seen": 724566016 + }, + { + "epoch": 0.22, + "learning_rate": 0.0006306599799398195, + "loss": 1.4781, + "theoretical_loss": 3.7947153444500588, + "tokens_seen": 724631552 + }, + { + "epoch": 0.22, + "learning_rate": 0.0006306439317953861, + "loss": 1.5511, + "theoretical_loss": 3.794680973822599, + "tokens_seen": 724697088 + }, + { + "epoch": 0.22, + "learning_rate": 0.0006306278836509529, + "loss": 1.4137, + "theoretical_loss": 3.7946466071734224, + "tokens_seen": 724762624 + }, + { + "epoch": 0.22, + "learning_rate": 0.0006306118355065195, + "loss": 1.541, + "theoretical_loss": 3.794612244501709, + "tokens_seen": 724828160 + }, + { + "epoch": 0.22, + "learning_rate": 0.0006305957873620863, + "loss": 1.4829, + "theoretical_loss": 3.794577885806639, + "tokens_seen": 724893696 + }, + { + "epoch": 0.22, + "learning_rate": 0.0006305797392176529, + "loss": 1.4808, + "theoretical_loss": 3.7945435310873927, + "tokens_seen": 724959232 + }, + { + "epoch": 0.22, + "learning_rate": 0.0006305636910732197, + "loss": 1.5678, + "theoretical_loss": 3.7945091803431503, + "tokens_seen": 725024768 + }, + { + "epoch": 0.22, + "learning_rate": 0.0006305476429287863, + "loss": 1.512, + "theoretical_loss": 3.7944748335730933, + "tokens_seen": 725090304 + }, + { + "epoch": 0.22, + "learning_rate": 0.0006305315947843531, + "loss": 1.5451, + "theoretical_loss": 3.794440490776402, + "tokens_seen": 725155840 + }, + { + "epoch": 0.22, + "learning_rate": 0.0006305155466399197, + "loss": 1.5041, + "theoretical_loss": 3.7944061519522583, + "tokens_seen": 725221376 + }, + { + "epoch": 0.22, + "learning_rate": 0.0006304994984954865, + "loss": 1.4251, + "theoretical_loss": 3.7943718170998437, + "tokens_seen": 725286912 + }, + { + "epoch": 0.22, + "learning_rate": 0.0006304834503510531, + "loss": 1.4346, + "theoretical_loss": 3.79433748621834, + "tokens_seen": 725352448 + }, + { + "epoch": 0.22, + "learning_rate": 0.0006304674022066199, + "loss": 1.5559, + "theoretical_loss": 3.7943031593069287, + "tokens_seen": 725417984 + }, + { + "epoch": 0.22, + "learning_rate": 0.0006304513540621866, + "loss": 1.4972, + "theoretical_loss": 3.794268836364793, + "tokens_seen": 725483520 + }, + { + "epoch": 0.22, + "learning_rate": 0.0006304353059177533, + "loss": 1.4708, + "theoretical_loss": 3.7942345173911147, + "tokens_seen": 725549056 + }, + { + "epoch": 0.22, + "learning_rate": 0.00063041925777332, + "loss": 1.4468, + "theoretical_loss": 3.7942002023850767, + "tokens_seen": 725614592 + }, + { + "epoch": 0.22, + "learning_rate": 0.0006304032096288867, + "loss": 1.4028, + "theoretical_loss": 3.7941658913458625, + "tokens_seen": 725680128 + }, + { + "epoch": 0.22, + "learning_rate": 0.0006303871614844534, + "loss": 1.4926, + "theoretical_loss": 3.794131584272655, + "tokens_seen": 725745664 + }, + { + "epoch": 0.22, + "learning_rate": 0.0006303711133400201, + "loss": 1.4051, + "theoretical_loss": 3.794097281164637, + "tokens_seen": 725811200 + }, + { + "epoch": 0.22, + "learning_rate": 0.0006303550651955868, + "loss": 1.5178, + "theoretical_loss": 3.7940629820209937, + "tokens_seen": 725876736 + }, + { + "epoch": 0.22, + "learning_rate": 0.0006303390170511535, + "loss": 1.506, + "theoretical_loss": 3.794028686840908, + "tokens_seen": 725942272 + }, + { + "epoch": 0.22, + "learning_rate": 0.0006303229689067202, + "loss": 1.492, + "theoretical_loss": 3.7939943956235647, + "tokens_seen": 726007808 + }, + { + "epoch": 0.22, + "learning_rate": 0.0006303069207622869, + "loss": 1.4651, + "theoretical_loss": 3.7939601083681476, + "tokens_seen": 726073344 + }, + { + "epoch": 0.22, + "learning_rate": 0.0006302908726178536, + "loss": 1.5542, + "theoretical_loss": 3.793925825073842, + "tokens_seen": 726138880 + }, + { + "epoch": 0.22, + "learning_rate": 0.0006302748244734203, + "loss": 1.4913, + "theoretical_loss": 3.7938915457398323, + "tokens_seen": 726204416 + }, + { + "epoch": 0.22, + "learning_rate": 0.000630258776328987, + "loss": 1.5205, + "theoretical_loss": 3.793857270365304, + "tokens_seen": 726269952 + }, + { + "epoch": 0.22, + "learning_rate": 0.0006302427281845537, + "loss": 1.5171, + "theoretical_loss": 3.7938229989494427, + "tokens_seen": 726335488 + }, + { + "epoch": 0.22, + "learning_rate": 0.0006302266800401204, + "loss": 1.442, + "theoretical_loss": 3.793788731491434, + "tokens_seen": 726401024 + }, + { + "epoch": 0.22, + "learning_rate": 0.0006302106318956871, + "loss": 1.5358, + "theoretical_loss": 3.7937544679904636, + "tokens_seen": 726466560 + }, + { + "epoch": 0.22, + "learning_rate": 0.0006301945837512538, + "loss": 1.5746, + "theoretical_loss": 3.793720208445718, + "tokens_seen": 726532096 + }, + { + "epoch": 0.22, + "learning_rate": 0.0006301785356068205, + "loss": 1.4222, + "theoretical_loss": 3.793685952856383, + "tokens_seen": 726597632 + }, + { + "epoch": 0.22, + "learning_rate": 0.0006301624874623872, + "loss": 1.3544, + "theoretical_loss": 3.7936517012216457, + "tokens_seen": 726663168 + }, + { + "epoch": 0.22, + "learning_rate": 0.0006301464393179539, + "loss": 1.557, + "theoretical_loss": 3.7936174535406932, + "tokens_seen": 726728704 + }, + { + "epoch": 0.22, + "learning_rate": 0.0006301303911735206, + "loss": 1.4592, + "theoretical_loss": 3.793583209812712, + "tokens_seen": 726794240 + }, + { + "epoch": 0.22, + "learning_rate": 0.0006301143430290873, + "loss": 1.4461, + "theoretical_loss": 3.7935489700368894, + "tokens_seen": 726859776 + }, + { + "epoch": 0.22, + "learning_rate": 0.000630098294884654, + "loss": 1.5072, + "theoretical_loss": 3.7935147342124136, + "tokens_seen": 726925312 + }, + { + "epoch": 0.22, + "learning_rate": 0.0006300822467402207, + "loss": 1.4415, + "theoretical_loss": 3.7934805023384723, + "tokens_seen": 726990848 + }, + { + "epoch": 0.22, + "learning_rate": 0.0006300661985957874, + "loss": 1.6121, + "theoretical_loss": 3.7934462744142534, + "tokens_seen": 727056384 + }, + { + "epoch": 0.22, + "learning_rate": 0.0006300501504513541, + "loss": 1.6367, + "theoretical_loss": 3.7934120504389455, + "tokens_seen": 727121920 + }, + { + "epoch": 0.22, + "learning_rate": 0.0006300341023069208, + "loss": 1.5162, + "theoretical_loss": 3.7933778304117363, + "tokens_seen": 727187456 + }, + { + "epoch": 0.22, + "learning_rate": 0.0006300180541624875, + "loss": 1.5594, + "theoretical_loss": 3.7933436143318153, + "tokens_seen": 727252992 + }, + { + "epoch": 0.22, + "learning_rate": 0.0006300020060180542, + "loss": 1.51, + "theoretical_loss": 3.7933094021983713, + "tokens_seen": 727318528 + }, + { + "epoch": 0.22, + "learning_rate": 0.0006299859578736209, + "loss": 1.4684, + "theoretical_loss": 3.793275194010594, + "tokens_seen": 727384064 + }, + { + "epoch": 0.22, + "objective/train/docs_used": 269904, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.5342673063278198, + "objective/train/theoretical_loss": 3.7932409897676727, + "objective/train/tokens_used": 747909600, + "theoretical_loss": 3.7932409897676727, + "tokens_seen": 727449600 + }, + { + "epoch": 0.22, + "learning_rate": 0.0006299699097291876, + "loss": 1.5411, + "theoretical_loss": 3.7932409897676727, + "tokens_seen": 727449600 + }, + { + "epoch": 0.22, + "learning_rate": 0.0006299538615847543, + "loss": 1.3999, + "theoretical_loss": 3.793206789468797, + "tokens_seen": 727515136 + }, + { + "epoch": 0.22, + "learning_rate": 0.000629937813440321, + "loss": 1.5531, + "theoretical_loss": 3.793172593113157, + "tokens_seen": 727580672 + }, + { + "epoch": 0.22, + "learning_rate": 0.0006299217652958877, + "loss": 1.484, + "theoretical_loss": 3.7931384006999425, + "tokens_seen": 727646208 + }, + { + "epoch": 0.22, + "learning_rate": 0.0006299057171514544, + "loss": 1.4605, + "theoretical_loss": 3.7931042122283447, + "tokens_seen": 727711744 + }, + { + "epoch": 0.22, + "learning_rate": 0.0006298896690070211, + "loss": 1.5566, + "theoretical_loss": 3.7930700276975537, + "tokens_seen": 727777280 + }, + { + "epoch": 0.22, + "learning_rate": 0.0006298736208625878, + "loss": 1.4626, + "theoretical_loss": 3.793035847106761, + "tokens_seen": 727842816 + }, + { + "epoch": 0.22, + "learning_rate": 0.0006298575727181545, + "loss": 1.5535, + "theoretical_loss": 3.7930016704551575, + "tokens_seen": 727908352 + }, + { + "epoch": 0.22, + "learning_rate": 0.0006298415245737212, + "loss": 1.4212, + "theoretical_loss": 3.7929674977419343, + "tokens_seen": 727973888 + }, + { + "epoch": 0.22, + "learning_rate": 0.0006298254764292879, + "loss": 1.4315, + "theoretical_loss": 3.792933328966283, + "tokens_seen": 728039424 + }, + { + "epoch": 0.22, + "learning_rate": 0.0006298094282848546, + "loss": 1.4483, + "theoretical_loss": 3.792899164127397, + "tokens_seen": 728104960 + }, + { + "epoch": 0.22, + "learning_rate": 0.0006297933801404213, + "loss": 1.454, + "theoretical_loss": 3.792865003224467, + "tokens_seen": 728170496 + }, + { + "epoch": 0.22, + "learning_rate": 0.000629777331995988, + "loss": 1.5401, + "theoretical_loss": 3.792830846256685, + "tokens_seen": 728236032 + }, + { + "epoch": 0.22, + "learning_rate": 0.0006297612838515547, + "loss": 1.4686, + "theoretical_loss": 3.792796693223245, + "tokens_seen": 728301568 + }, + { + "epoch": 0.22, + "learning_rate": 0.0006297452357071214, + "loss": 1.4528, + "theoretical_loss": 3.792762544123339, + "tokens_seen": 728367104 + }, + { + "epoch": 0.22, + "learning_rate": 0.0006297291875626881, + "loss": 1.443, + "theoretical_loss": 3.79272839895616, + "tokens_seen": 728432640 + }, + { + "epoch": 0.22, + "learning_rate": 0.0006297131394182548, + "loss": 1.5216, + "theoretical_loss": 3.792694257720902, + "tokens_seen": 728498176 + }, + { + "epoch": 0.22, + "learning_rate": 0.0006296970912738215, + "loss": 1.5649, + "theoretical_loss": 3.792660120416758, + "tokens_seen": 728563712 + }, + { + "epoch": 0.22, + "learning_rate": 0.0006296810431293882, + "loss": 1.493, + "theoretical_loss": 3.792625987042922, + "tokens_seen": 728629248 + }, + { + "epoch": 0.22, + "learning_rate": 0.0006296649949849549, + "loss": 1.4625, + "theoretical_loss": 3.7925918575985884, + "tokens_seen": 728694784 + }, + { + "epoch": 0.22, + "learning_rate": 0.0006296489468405216, + "loss": 1.486, + "theoretical_loss": 3.7925577320829507, + "tokens_seen": 728760320 + }, + { + "epoch": 0.22, + "learning_rate": 0.0006296328986960883, + "loss": 1.4899, + "theoretical_loss": 3.792523610495204, + "tokens_seen": 728825856 + }, + { + "epoch": 0.22, + "learning_rate": 0.000629616850551655, + "loss": 1.454, + "theoretical_loss": 3.7924894928345427, + "tokens_seen": 728891392 + }, + { + "epoch": 0.22, + "learning_rate": 0.0006296008024072217, + "loss": 1.4978, + "theoretical_loss": 3.7924553791001623, + "tokens_seen": 728956928 + }, + { + "epoch": 0.22, + "learning_rate": 0.0006295847542627884, + "loss": 1.4732, + "theoretical_loss": 3.792421269291258, + "tokens_seen": 729022464 + }, + { + "epoch": 0.22, + "learning_rate": 0.0006295687061183551, + "loss": 1.503, + "theoretical_loss": 3.7923871634070245, + "tokens_seen": 729088000 + }, + { + "epoch": 0.22, + "learning_rate": 0.0006295526579739218, + "loss": 1.5633, + "theoretical_loss": 3.792353061446658, + "tokens_seen": 729153536 + }, + { + "epoch": 0.22, + "learning_rate": 0.0006295366098294885, + "loss": 1.4416, + "theoretical_loss": 3.792318963409355, + "tokens_seen": 729219072 + }, + { + "epoch": 0.22, + "learning_rate": 0.0006295205616850552, + "loss": 1.528, + "theoretical_loss": 3.7922848692943103, + "tokens_seen": 729284608 + }, + { + "epoch": 0.22, + "learning_rate": 0.0006295045135406219, + "loss": 1.4399, + "theoretical_loss": 3.792250779100722, + "tokens_seen": 729350144 + }, + { + "epoch": 0.22, + "learning_rate": 0.0006294884653961886, + "loss": 1.4654, + "theoretical_loss": 3.792216692827785, + "tokens_seen": 729415680 + }, + { + "epoch": 0.22, + "learning_rate": 0.0006294724172517553, + "loss": 1.4219, + "theoretical_loss": 3.792182610474698, + "tokens_seen": 729481216 + }, + { + "epoch": 0.22, + "learning_rate": 0.000629456369107322, + "loss": 1.507, + "theoretical_loss": 3.792148532040657, + "tokens_seen": 729546752 + }, + { + "epoch": 0.22, + "learning_rate": 0.0006294403209628887, + "loss": 1.5853, + "theoretical_loss": 3.7921144575248595, + "tokens_seen": 729612288 + }, + { + "epoch": 0.22, + "learning_rate": 0.0006294242728184554, + "loss": 1.3532, + "theoretical_loss": 3.7920803869265027, + "tokens_seen": 729677824 + }, + { + "epoch": 0.22, + "learning_rate": 0.0006294082246740221, + "loss": 1.4816, + "theoretical_loss": 3.792046320244786, + "tokens_seen": 729743360 + }, + { + "epoch": 0.22, + "learning_rate": 0.0006293921765295888, + "loss": 1.4346, + "theoretical_loss": 3.792012257478906, + "tokens_seen": 729808896 + }, + { + "epoch": 0.22, + "learning_rate": 0.0006293761283851555, + "loss": 1.3574, + "theoretical_loss": 3.791978198628061, + "tokens_seen": 729874432 + }, + { + "epoch": 0.22, + "learning_rate": 0.0006293600802407222, + "loss": 1.4514, + "theoretical_loss": 3.7919441436914507, + "tokens_seen": 729939968 + }, + { + "epoch": 0.22, + "learning_rate": 0.0006293440320962888, + "loss": 1.4211, + "theoretical_loss": 3.7919100926682727, + "tokens_seen": 730005504 + }, + { + "epoch": 0.22, + "learning_rate": 0.0006293279839518556, + "loss": 1.508, + "theoretical_loss": 3.7918760455577267, + "tokens_seen": 730071040 + }, + { + "epoch": 0.22, + "learning_rate": 0.0006293119358074222, + "loss": 1.4463, + "theoretical_loss": 3.791842002359012, + "tokens_seen": 730136576 + }, + { + "epoch": 0.22, + "learning_rate": 0.000629295887662989, + "loss": 1.4216, + "theoretical_loss": 3.7918079630713275, + "tokens_seen": 730202112 + }, + { + "epoch": 0.22, + "learning_rate": 0.0006292798395185556, + "loss": 1.5401, + "theoretical_loss": 3.7917739276938733, + "tokens_seen": 730267648 + }, + { + "epoch": 0.22, + "learning_rate": 0.0006292637913741224, + "loss": 1.4354, + "theoretical_loss": 3.7917398962258493, + "tokens_seen": 730333184 + }, + { + "epoch": 0.22, + "learning_rate": 0.000629247743229689, + "loss": 1.525, + "theoretical_loss": 3.791705868666456, + "tokens_seen": 730398720 + }, + { + "epoch": 0.22, + "learning_rate": 0.0006292316950852558, + "loss": 1.3751, + "theoretical_loss": 3.791671845014893, + "tokens_seen": 730464256 + }, + { + "epoch": 0.22, + "learning_rate": 0.0006292156469408224, + "loss": 1.3639, + "theoretical_loss": 3.7916378252703624, + "tokens_seen": 730529792 + }, + { + "epoch": 0.22, + "learning_rate": 0.0006291995987963892, + "loss": 1.6039, + "theoretical_loss": 3.791603809432064, + "tokens_seen": 730595328 + }, + { + "epoch": 0.22, + "learning_rate": 0.0006291835506519558, + "loss": 1.4665, + "theoretical_loss": 3.7915697974991986, + "tokens_seen": 730660864 + }, + { + "epoch": 0.22, + "objective/train/docs_used": 271086, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.4612535238265991, + "objective/train/theoretical_loss": 3.7915357894709687, + "objective/train/tokens_used": 751186400, + "theoretical_loss": 3.7915357894709687, + "tokens_seen": 730726400 + }, + { + "epoch": 0.22, + "learning_rate": 0.0006291675025075226, + "loss": 1.45, + "theoretical_loss": 3.7915357894709687, + "tokens_seen": 730726400 + }, + { + "epoch": 0.22, + "learning_rate": 0.0006291514543630892, + "loss": 1.3879, + "theoretical_loss": 3.791501785346575, + "tokens_seen": 730791936 + }, + { + "epoch": 0.22, + "learning_rate": 0.000629135406218656, + "loss": 1.4545, + "theoretical_loss": 3.79146778512522, + "tokens_seen": 730857472 + }, + { + "epoch": 0.22, + "learning_rate": 0.0006291193580742226, + "loss": 1.5113, + "theoretical_loss": 3.7914337888061054, + "tokens_seen": 730923008 + }, + { + "epoch": 0.22, + "learning_rate": 0.0006291033099297894, + "loss": 1.4387, + "theoretical_loss": 3.7913997963884336, + "tokens_seen": 730988544 + }, + { + "epoch": 0.22, + "learning_rate": 0.000629087261785356, + "loss": 1.4686, + "theoretical_loss": 3.791365807871407, + "tokens_seen": 731054080 + }, + { + "epoch": 0.22, + "learning_rate": 0.0006290712136409228, + "loss": 1.5712, + "theoretical_loss": 3.791331823254229, + "tokens_seen": 731119616 + }, + { + "epoch": 0.22, + "learning_rate": 0.0006290551654964895, + "loss": 1.5098, + "theoretical_loss": 3.791297842536102, + "tokens_seen": 731185152 + }, + { + "epoch": 0.22, + "learning_rate": 0.0006290391173520562, + "loss": 1.4028, + "theoretical_loss": 3.791263865716229, + "tokens_seen": 731250688 + }, + { + "epoch": 0.22, + "learning_rate": 0.0006290230692076229, + "loss": 1.527, + "theoretical_loss": 3.7912298927938144, + "tokens_seen": 731316224 + }, + { + "epoch": 0.22, + "learning_rate": 0.0006290070210631896, + "loss": 1.5058, + "theoretical_loss": 3.7911959237680612, + "tokens_seen": 731381760 + }, + { + "epoch": 0.22, + "learning_rate": 0.0006289909729187563, + "loss": 1.437, + "theoretical_loss": 3.791161958638173, + "tokens_seen": 731447296 + }, + { + "epoch": 0.22, + "learning_rate": 0.000628974924774323, + "loss": 1.5337, + "theoretical_loss": 3.7911279974033554, + "tokens_seen": 731512832 + }, + { + "epoch": 0.22, + "learning_rate": 0.0006289588766298897, + "loss": 1.4193, + "theoretical_loss": 3.7910940400628115, + "tokens_seen": 731578368 + }, + { + "epoch": 0.22, + "learning_rate": 0.0006289428284854564, + "loss": 1.3766, + "theoretical_loss": 3.7910600866157464, + "tokens_seen": 731643904 + }, + { + "epoch": 0.22, + "learning_rate": 0.0006289267803410231, + "loss": 1.4411, + "theoretical_loss": 3.7910261370613654, + "tokens_seen": 731709440 + }, + { + "epoch": 0.22, + "learning_rate": 0.0006289107321965898, + "loss": 1.4839, + "theoretical_loss": 3.790992191398873, + "tokens_seen": 731774976 + }, + { + "epoch": 0.22, + "learning_rate": 0.0006288946840521565, + "loss": 1.5024, + "theoretical_loss": 3.790958249627475, + "tokens_seen": 731840512 + }, + { + "epoch": 0.22, + "learning_rate": 0.0006288786359077232, + "loss": 1.5632, + "theoretical_loss": 3.7909243117463767, + "tokens_seen": 731906048 + }, + { + "epoch": 0.22, + "learning_rate": 0.0006288625877632899, + "loss": 1.4146, + "theoretical_loss": 3.7908903777547835, + "tokens_seen": 731971584 + }, + { + "epoch": 0.22, + "learning_rate": 0.0006288465396188566, + "loss": 1.4201, + "theoretical_loss": 3.7908564476519024, + "tokens_seen": 732037120 + }, + { + "epoch": 0.22, + "learning_rate": 0.0006288304914744233, + "loss": 1.5013, + "theoretical_loss": 3.7908225214369393, + "tokens_seen": 732102656 + }, + { + "epoch": 0.22, + "learning_rate": 0.00062881444332999, + "loss": 1.5193, + "theoretical_loss": 3.7907885991091, + "tokens_seen": 732168192 + }, + { + "epoch": 0.22, + "learning_rate": 0.0006287983951855567, + "loss": 1.4473, + "theoretical_loss": 3.790754680667593, + "tokens_seen": 732233728 + }, + { + "epoch": 0.22, + "learning_rate": 0.0006287823470411234, + "loss": 1.4793, + "theoretical_loss": 3.7907207661116233, + "tokens_seen": 732299264 + }, + { + "epoch": 0.22, + "learning_rate": 0.0006287662988966901, + "loss": 1.5325, + "theoretical_loss": 3.7906868554403994, + "tokens_seen": 732364800 + }, + { + "epoch": 0.22, + "learning_rate": 0.0006287502507522568, + "loss": 1.3895, + "theoretical_loss": 3.7906529486531286, + "tokens_seen": 732430336 + }, + { + "epoch": 0.22, + "learning_rate": 0.0006287342026078235, + "loss": 1.4725, + "theoretical_loss": 3.7906190457490183, + "tokens_seen": 732495872 + }, + { + "epoch": 0.22, + "learning_rate": 0.0006287181544633902, + "loss": 1.4265, + "theoretical_loss": 3.790585146727276, + "tokens_seen": 732561408 + }, + { + "epoch": 0.22, + "learning_rate": 0.0006287021063189569, + "loss": 1.4518, + "theoretical_loss": 3.790551251587111, + "tokens_seen": 732626944 + }, + { + "epoch": 0.22, + "learning_rate": 0.0006286860581745236, + "loss": 1.4956, + "theoretical_loss": 3.7905173603277307, + "tokens_seen": 732692480 + }, + { + "epoch": 0.22, + "learning_rate": 0.0006286700100300903, + "loss": 1.4125, + "theoretical_loss": 3.790483472948344, + "tokens_seen": 732758016 + }, + { + "epoch": 0.22, + "learning_rate": 0.000628653961885657, + "loss": 1.3758, + "theoretical_loss": 3.79044958944816, + "tokens_seen": 732823552 + }, + { + "epoch": 0.22, + "learning_rate": 0.0006286379137412237, + "loss": 1.6214, + "theoretical_loss": 3.7904157098263873, + "tokens_seen": 732889088 + }, + { + "epoch": 0.22, + "learning_rate": 0.0006286218655967904, + "loss": 1.5612, + "theoretical_loss": 3.7903818340822353, + "tokens_seen": 732954624 + }, + { + "epoch": 0.22, + "learning_rate": 0.0006286058174523571, + "loss": 1.5127, + "theoretical_loss": 3.790347962214914, + "tokens_seen": 733020160 + }, + { + "epoch": 0.22, + "learning_rate": 0.0006285897693079238, + "loss": 1.4199, + "theoretical_loss": 3.790314094223633, + "tokens_seen": 733085696 + }, + { + "epoch": 0.22, + "learning_rate": 0.0006285737211634905, + "loss": 1.5116, + "theoretical_loss": 3.7902802301076024, + "tokens_seen": 733151232 + }, + { + "epoch": 0.22, + "learning_rate": 0.0006285576730190572, + "loss": 1.4847, + "theoretical_loss": 3.790246369866032, + "tokens_seen": 733216768 + }, + { + "epoch": 0.22, + "learning_rate": 0.0006285416248746239, + "loss": 1.4566, + "theoretical_loss": 3.790212513498132, + "tokens_seen": 733282304 + }, + { + "epoch": 0.22, + "learning_rate": 0.0006285255767301906, + "loss": 1.5708, + "theoretical_loss": 3.7901786610031145, + "tokens_seen": 733347840 + }, + { + "epoch": 0.22, + "learning_rate": 0.0006285095285857573, + "loss": 1.4476, + "theoretical_loss": 3.7901448123801886, + "tokens_seen": 733413376 + }, + { + "epoch": 0.22, + "learning_rate": 0.000628493480441324, + "loss": 1.5159, + "theoretical_loss": 3.790110967628567, + "tokens_seen": 733478912 + }, + { + "epoch": 0.22, + "learning_rate": 0.0006284774322968907, + "loss": 1.5558, + "theoretical_loss": 3.7900771267474607, + "tokens_seen": 733544448 + }, + { + "epoch": 0.22, + "learning_rate": 0.0006284613841524574, + "loss": 1.4575, + "theoretical_loss": 3.7900432897360803, + "tokens_seen": 733609984 + }, + { + "epoch": 0.22, + "learning_rate": 0.0006284453360080241, + "loss": 1.4502, + "theoretical_loss": 3.790009456593639, + "tokens_seen": 733675520 + }, + { + "epoch": 0.22, + "learning_rate": 0.0006284292878635908, + "loss": 1.4179, + "theoretical_loss": 3.7899756273193486, + "tokens_seen": 733741056 + }, + { + "epoch": 0.22, + "learning_rate": 0.0006284132397191575, + "loss": 1.5117, + "theoretical_loss": 3.7899418019124207, + "tokens_seen": 733806592 + }, + { + "epoch": 0.22, + "learning_rate": 0.0006283971915747242, + "loss": 1.475, + "theoretical_loss": 3.7899079803720688, + "tokens_seen": 733872128 + }, + { + "epoch": 0.22, + "learning_rate": 0.0006283811434302909, + "loss": 1.5424, + "theoretical_loss": 3.7898741626975054, + "tokens_seen": 733937664 + }, + { + "epoch": 0.22, + "objective/train/docs_used": 272255, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.570894718170166, + "objective/train/theoretical_loss": 3.7898403488879424, + "objective/train/tokens_used": 754463200, + "theoretical_loss": 3.7898403488879424, + "tokens_seen": 734003200 + }, + { + "epoch": 0.22, + "learning_rate": 0.0006283650952858576, + "loss": 1.6115, + "theoretical_loss": 3.7898403488879424, + "tokens_seen": 734003200 + }, + { + "epoch": 0.22, + "learning_rate": 0.0006283490471414243, + "loss": 1.5486, + "theoretical_loss": 3.7898065389425946, + "tokens_seen": 734068736 + }, + { + "epoch": 0.22, + "learning_rate": 0.000628332998996991, + "loss": 1.3091, + "theoretical_loss": 3.789772732860675, + "tokens_seen": 734134272 + }, + { + "epoch": 0.22, + "learning_rate": 0.0006283169508525577, + "loss": 1.5038, + "theoretical_loss": 3.7897389306413967, + "tokens_seen": 734199808 + }, + { + "epoch": 0.22, + "learning_rate": 0.0006283009027081244, + "loss": 1.5614, + "theoretical_loss": 3.7897051322839745, + "tokens_seen": 734265344 + }, + { + "epoch": 0.22, + "learning_rate": 0.0006282848545636911, + "loss": 1.4667, + "theoretical_loss": 3.7896713377876217, + "tokens_seen": 734330880 + }, + { + "epoch": 0.22, + "learning_rate": 0.0006282688064192578, + "loss": 1.4945, + "theoretical_loss": 3.7896375471515533, + "tokens_seen": 734396416 + }, + { + "epoch": 0.22, + "learning_rate": 0.0006282527582748245, + "loss": 1.4911, + "theoretical_loss": 3.789603760374984, + "tokens_seen": 734461952 + }, + { + "epoch": 0.22, + "learning_rate": 0.0006282367101303912, + "loss": 1.4767, + "theoretical_loss": 3.789569977457128, + "tokens_seen": 734527488 + }, + { + "epoch": 0.22, + "learning_rate": 0.000628220661985958, + "loss": 1.4969, + "theoretical_loss": 3.789536198397201, + "tokens_seen": 734593024 + }, + { + "epoch": 0.22, + "learning_rate": 0.0006282046138415246, + "loss": 1.6598, + "theoretical_loss": 3.789502423194418, + "tokens_seen": 734658560 + }, + { + "epoch": 0.22, + "learning_rate": 0.0006281885656970914, + "loss": 1.5035, + "theoretical_loss": 3.7894686518479945, + "tokens_seen": 734724096 + }, + { + "epoch": 0.22, + "learning_rate": 0.000628172517552658, + "loss": 1.4539, + "theoretical_loss": 3.789434884357147, + "tokens_seen": 734789632 + }, + { + "epoch": 0.22, + "learning_rate": 0.0006281564694082248, + "loss": 1.5607, + "theoretical_loss": 3.7894011207210907, + "tokens_seen": 734855168 + }, + { + "epoch": 0.22, + "learning_rate": 0.0006281404212637914, + "loss": 1.4228, + "theoretical_loss": 3.7893673609390417, + "tokens_seen": 734920704 + }, + { + "epoch": 0.22, + "learning_rate": 0.0006281243731193582, + "loss": 1.4733, + "theoretical_loss": 3.789333605010217, + "tokens_seen": 734986240 + }, + { + "epoch": 0.22, + "learning_rate": 0.0006281083249749248, + "loss": 1.3917, + "theoretical_loss": 3.7892998529338326, + "tokens_seen": 735051776 + }, + { + "epoch": 0.22, + "learning_rate": 0.0006280922768304916, + "loss": 1.5716, + "theoretical_loss": 3.7892661047091067, + "tokens_seen": 735117312 + }, + { + "epoch": 0.22, + "learning_rate": 0.0006280762286860582, + "loss": 1.4652, + "theoretical_loss": 3.789232360335255, + "tokens_seen": 735182848 + }, + { + "epoch": 0.22, + "learning_rate": 0.0006280601805416249, + "loss": 1.5047, + "theoretical_loss": 3.789198619811496, + "tokens_seen": 735248384 + }, + { + "epoch": 0.22, + "learning_rate": 0.0006280441323971916, + "loss": 1.4298, + "theoretical_loss": 3.7891648831370466, + "tokens_seen": 735313920 + }, + { + "epoch": 0.22, + "learning_rate": 0.0006280280842527583, + "loss": 1.5743, + "theoretical_loss": 3.7891311503111247, + "tokens_seen": 735379456 + }, + { + "epoch": 0.22, + "learning_rate": 0.000628012036108325, + "loss": 1.3357, + "theoretical_loss": 3.789097421332949, + "tokens_seen": 735444992 + }, + { + "epoch": 0.22, + "learning_rate": 0.0006279959879638917, + "loss": 1.5108, + "theoretical_loss": 3.7890636962017368, + "tokens_seen": 735510528 + }, + { + "epoch": 0.22, + "learning_rate": 0.0006279799398194584, + "loss": 1.5051, + "theoretical_loss": 3.7890299749167076, + "tokens_seen": 735576064 + }, + { + "epoch": 0.22, + "learning_rate": 0.0006279638916750251, + "loss": 1.4945, + "theoretical_loss": 3.7889962574770792, + "tokens_seen": 735641600 + }, + { + "epoch": 0.22, + "learning_rate": 0.0006279478435305917, + "loss": 1.4605, + "theoretical_loss": 3.7889625438820715, + "tokens_seen": 735707136 + }, + { + "epoch": 0.22, + "learning_rate": 0.0006279317953861585, + "loss": 1.4139, + "theoretical_loss": 3.788928834130903, + "tokens_seen": 735772672 + }, + { + "epoch": 0.22, + "learning_rate": 0.0006279157472417251, + "loss": 1.4658, + "theoretical_loss": 3.7888951282227934, + "tokens_seen": 735838208 + }, + { + "epoch": 0.22, + "learning_rate": 0.0006278996990972919, + "loss": 1.5194, + "theoretical_loss": 3.788861426156962, + "tokens_seen": 735903744 + }, + { + "epoch": 0.22, + "learning_rate": 0.0006278836509528585, + "loss": 1.4614, + "theoretical_loss": 3.7888277279326292, + "tokens_seen": 735969280 + }, + { + "epoch": 0.22, + "learning_rate": 0.0006278676028084253, + "loss": 1.6013, + "theoretical_loss": 3.7887940335490153, + "tokens_seen": 736034816 + }, + { + "epoch": 0.22, + "learning_rate": 0.0006278515546639919, + "loss": 1.4869, + "theoretical_loss": 3.7887603430053396, + "tokens_seen": 736100352 + }, + { + "epoch": 0.22, + "learning_rate": 0.0006278355065195587, + "loss": 1.4719, + "theoretical_loss": 3.788726656300824, + "tokens_seen": 736165888 + }, + { + "epoch": 0.22, + "learning_rate": 0.0006278194583751253, + "loss": 1.4428, + "theoretical_loss": 3.7886929734346886, + "tokens_seen": 736231424 + }, + { + "epoch": 0.22, + "learning_rate": 0.0006278034102306921, + "loss": 1.5555, + "theoretical_loss": 3.788659294406154, + "tokens_seen": 736296960 + }, + { + "epoch": 0.22, + "learning_rate": 0.0006277873620862587, + "loss": 1.5374, + "theoretical_loss": 3.788625619214442, + "tokens_seen": 736362496 + }, + { + "epoch": 0.22, + "learning_rate": 0.0006277713139418255, + "loss": 1.4136, + "theoretical_loss": 3.7885919478587744, + "tokens_seen": 736428032 + }, + { + "epoch": 0.22, + "learning_rate": 0.0006277552657973921, + "loss": 1.4378, + "theoretical_loss": 3.788558280338372, + "tokens_seen": 736493568 + }, + { + "epoch": 0.22, + "learning_rate": 0.0006277392176529589, + "loss": 1.5028, + "theoretical_loss": 3.7885246166524578, + "tokens_seen": 736559104 + }, + { + "epoch": 0.22, + "learning_rate": 0.0006277231695085255, + "loss": 1.3868, + "theoretical_loss": 3.7884909568002536, + "tokens_seen": 736624640 + }, + { + "epoch": 0.22, + "learning_rate": 0.0006277071213640923, + "loss": 1.4857, + "theoretical_loss": 3.7884573007809808, + "tokens_seen": 736690176 + }, + { + "epoch": 0.22, + "learning_rate": 0.000627691073219659, + "loss": 1.5264, + "theoretical_loss": 3.788423648593863, + "tokens_seen": 736755712 + }, + { + "epoch": 0.22, + "learning_rate": 0.0006276750250752257, + "loss": 1.4381, + "theoretical_loss": 3.788390000238123, + "tokens_seen": 736821248 + }, + { + "epoch": 0.22, + "learning_rate": 0.0006276589769307924, + "loss": 1.5131, + "theoretical_loss": 3.788356355712984, + "tokens_seen": 736886784 + }, + { + "epoch": 0.22, + "learning_rate": 0.0006276429287863591, + "loss": 1.5225, + "theoretical_loss": 3.788322715017668, + "tokens_seen": 736952320 + }, + { + "epoch": 0.22, + "learning_rate": 0.0006276268806419258, + "loss": 1.4882, + "theoretical_loss": 3.7882890781514007, + "tokens_seen": 737017856 + }, + { + "epoch": 0.22, + "learning_rate": 0.0006276108324974925, + "loss": 1.5809, + "theoretical_loss": 3.7882554451134043, + "tokens_seen": 737083392 + }, + { + "epoch": 0.22, + "learning_rate": 0.0006275947843530592, + "loss": 1.3904, + "theoretical_loss": 3.7882218159029026, + "tokens_seen": 737148928 + }, + { + "epoch": 0.22, + "learning_rate": 0.0006275787362086259, + "loss": 1.494, + "theoretical_loss": 3.7881881905191213, + "tokens_seen": 737214464 + }, + { + "epoch": 0.22, + "objective/train/docs_used": 273341, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.5330805778503418, + "objective/train/theoretical_loss": 3.788154568961283, + "objective/train/tokens_used": 757740000, + "theoretical_loss": 3.788154568961283, + "tokens_seen": 737280000 + }, + { + "epoch": 0.22, + "learning_rate": 0.0006275626880641926, + "loss": 1.5181, + "theoretical_loss": 3.788154568961283, + "tokens_seen": 737280000 + }, + { + "epoch": 0.22, + "learning_rate": 0.0006275466399197593, + "loss": 1.5356, + "theoretical_loss": 3.788120951228614, + "tokens_seen": 737345536 + }, + { + "epoch": 0.22, + "learning_rate": 0.000627530591775326, + "loss": 1.5118, + "theoretical_loss": 3.788087337320338, + "tokens_seen": 737411072 + }, + { + "epoch": 0.22, + "learning_rate": 0.0006275145436308927, + "loss": 1.42, + "theoretical_loss": 3.7880537272356802, + "tokens_seen": 737476608 + }, + { + "epoch": 0.22, + "learning_rate": 0.0006274984954864594, + "loss": 1.5392, + "theoretical_loss": 3.788020120973867, + "tokens_seen": 737542144 + }, + { + "epoch": 0.22, + "learning_rate": 0.0006274824473420261, + "loss": 1.5034, + "theoretical_loss": 3.7879865185341224, + "tokens_seen": 737607680 + }, + { + "epoch": 0.22, + "learning_rate": 0.0006274663991975928, + "loss": 1.4547, + "theoretical_loss": 3.787952919915673, + "tokens_seen": 737673216 + }, + { + "epoch": 0.22, + "learning_rate": 0.0006274503510531595, + "loss": 1.4755, + "theoretical_loss": 3.7879193251177457, + "tokens_seen": 737738752 + }, + { + "epoch": 0.22, + "learning_rate": 0.0006274343029087262, + "loss": 1.5271, + "theoretical_loss": 3.7878857341395653, + "tokens_seen": 737804288 + }, + { + "epoch": 0.22, + "learning_rate": 0.0006274182547642929, + "loss": 1.4634, + "theoretical_loss": 3.7878521469803585, + "tokens_seen": 737869824 + }, + { + "epoch": 0.22, + "learning_rate": 0.0006274022066198596, + "loss": 1.5439, + "theoretical_loss": 3.7878185636393527, + "tokens_seen": 737935360 + }, + { + "epoch": 0.22, + "learning_rate": 0.0006273861584754263, + "loss": 1.5129, + "theoretical_loss": 3.787784984115774, + "tokens_seen": 738000896 + }, + { + "epoch": 0.22, + "learning_rate": 0.000627370110330993, + "loss": 1.422, + "theoretical_loss": 3.78775140840885, + "tokens_seen": 738066432 + }, + { + "epoch": 0.22, + "learning_rate": 0.0006273540621865597, + "loss": 1.4924, + "theoretical_loss": 3.7877178365178086, + "tokens_seen": 738131968 + }, + { + "epoch": 0.22, + "learning_rate": 0.0006273380140421264, + "loss": 1.4642, + "theoretical_loss": 3.787684268441876, + "tokens_seen": 738197504 + }, + { + "epoch": 0.22, + "learning_rate": 0.0006273219658976931, + "loss": 1.4747, + "theoretical_loss": 3.787650704180281, + "tokens_seen": 738263040 + }, + { + "epoch": 0.22, + "learning_rate": 0.0006273059177532598, + "loss": 1.5704, + "theoretical_loss": 3.787617143732251, + "tokens_seen": 738328576 + }, + { + "epoch": 0.22, + "learning_rate": 0.0006272898696088265, + "loss": 1.498, + "theoretical_loss": 3.787583587097015, + "tokens_seen": 738394112 + }, + { + "epoch": 0.22, + "learning_rate": 0.0006272738214643932, + "loss": 1.4062, + "theoretical_loss": 3.7875500342738015, + "tokens_seen": 738459648 + }, + { + "epoch": 0.22, + "learning_rate": 0.0006272577733199599, + "loss": 1.4587, + "theoretical_loss": 3.787516485261839, + "tokens_seen": 738525184 + }, + { + "epoch": 0.22, + "learning_rate": 0.0006272417251755266, + "loss": 1.4763, + "theoretical_loss": 3.7874829400603556, + "tokens_seen": 738590720 + }, + { + "epoch": 0.22, + "learning_rate": 0.0006272256770310933, + "loss": 1.3578, + "theoretical_loss": 3.7874493986685813, + "tokens_seen": 738656256 + }, + { + "epoch": 0.22, + "learning_rate": 0.00062720962888666, + "loss": 1.5047, + "theoretical_loss": 3.787415861085745, + "tokens_seen": 738721792 + }, + { + "epoch": 0.22, + "learning_rate": 0.0006271935807422267, + "loss": 1.3874, + "theoretical_loss": 3.787382327311077, + "tokens_seen": 738787328 + }, + { + "epoch": 0.22, + "learning_rate": 0.0006271775325977934, + "loss": 1.4702, + "theoretical_loss": 3.7873487973438067, + "tokens_seen": 738852864 + }, + { + "epoch": 0.22, + "learning_rate": 0.0006271614844533601, + "loss": 1.5505, + "theoretical_loss": 3.7873152711831644, + "tokens_seen": 738918400 + }, + { + "epoch": 0.22, + "learning_rate": 0.0006271454363089268, + "loss": 1.4537, + "theoretical_loss": 3.7872817488283803, + "tokens_seen": 738983936 + }, + { + "epoch": 0.22, + "learning_rate": 0.0006271293881644935, + "loss": 1.4347, + "theoretical_loss": 3.7872482302786845, + "tokens_seen": 739049472 + }, + { + "epoch": 0.22, + "learning_rate": 0.0006271133400200602, + "loss": 1.456, + "theoretical_loss": 3.787214715533308, + "tokens_seen": 739115008 + }, + { + "epoch": 0.22, + "learning_rate": 0.0006270972918756269, + "loss": 1.5559, + "theoretical_loss": 3.787181204591482, + "tokens_seen": 739180544 + }, + { + "epoch": 0.22, + "learning_rate": 0.0006270812437311936, + "loss": 1.3923, + "theoretical_loss": 3.787147697452437, + "tokens_seen": 739246080 + }, + { + "epoch": 0.22, + "learning_rate": 0.0006270651955867603, + "loss": 1.6654, + "theoretical_loss": 3.7871141941154054, + "tokens_seen": 739311616 + }, + { + "epoch": 0.22, + "learning_rate": 0.000627049147442327, + "loss": 1.4738, + "theoretical_loss": 3.7870806945796174, + "tokens_seen": 739377152 + }, + { + "epoch": 0.22, + "learning_rate": 0.0006270330992978937, + "loss": 1.3619, + "theoretical_loss": 3.7870471988443066, + "tokens_seen": 739442688 + }, + { + "epoch": 0.22, + "learning_rate": 0.0006270170511534605, + "loss": 1.4082, + "theoretical_loss": 3.7870137069087035, + "tokens_seen": 739508224 + }, + { + "epoch": 0.22, + "learning_rate": 0.0006270010030090271, + "loss": 1.5008, + "theoretical_loss": 3.786980218772041, + "tokens_seen": 739573760 + }, + { + "epoch": 0.22, + "learning_rate": 0.0006269849548645939, + "loss": 1.6412, + "theoretical_loss": 3.7869467344335526, + "tokens_seen": 739639296 + }, + { + "epoch": 0.22, + "learning_rate": 0.0006269689067201605, + "loss": 1.45, + "theoretical_loss": 3.7869132538924695, + "tokens_seen": 739704832 + }, + { + "epoch": 0.22, + "learning_rate": 0.0006269528585757273, + "loss": 1.5111, + "theoretical_loss": 3.786879777148025, + "tokens_seen": 739770368 + }, + { + "epoch": 0.22, + "learning_rate": 0.0006269368104312939, + "loss": 1.4338, + "theoretical_loss": 3.786846304199453, + "tokens_seen": 739835904 + }, + { + "epoch": 0.22, + "learning_rate": 0.0006269207622868607, + "loss": 1.5112, + "theoretical_loss": 3.7868128350459864, + "tokens_seen": 739901440 + }, + { + "epoch": 0.22, + "learning_rate": 0.0006269047141424273, + "loss": 1.4034, + "theoretical_loss": 3.7867793696868586, + "tokens_seen": 739966976 + }, + { + "epoch": 0.22, + "learning_rate": 0.0006268886659979941, + "loss": 1.5174, + "theoretical_loss": 3.7867459081213037, + "tokens_seen": 740032512 + }, + { + "epoch": 0.22, + "learning_rate": 0.0006268726178535607, + "loss": 1.5405, + "theoretical_loss": 3.7867124503485563, + "tokens_seen": 740098048 + }, + { + "epoch": 0.22, + "learning_rate": 0.0006268565697091275, + "loss": 1.4285, + "theoretical_loss": 3.7866789963678498, + "tokens_seen": 740163584 + }, + { + "epoch": 0.22, + "learning_rate": 0.0006268405215646941, + "loss": 1.5519, + "theoretical_loss": 3.786645546178419, + "tokens_seen": 740229120 + }, + { + "epoch": 0.22, + "learning_rate": 0.0006268244734202609, + "loss": 1.4782, + "theoretical_loss": 3.7866120997794983, + "tokens_seen": 740294656 + }, + { + "epoch": 0.22, + "learning_rate": 0.0006268084252758275, + "loss": 1.6531, + "theoretical_loss": 3.786578657170324, + "tokens_seen": 740360192 + }, + { + "epoch": 0.22, + "learning_rate": 0.0006267923771313943, + "loss": 1.4714, + "theoretical_loss": 3.7865452183501302, + "tokens_seen": 740425728 + }, + { + "epoch": 0.22, + "learning_rate": 0.0006267763289869609, + "loss": 1.5646, + "theoretical_loss": 3.7865117833181525, + "tokens_seen": 740491264 + }, + { + "epoch": 0.22, + "objective/train/docs_used": 274540, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.6244080066680908, + "objective/train/theoretical_loss": 3.786478352073626, + "objective/train/tokens_used": 761016800, + "theoretical_loss": 3.786478352073626, + "tokens_seen": 740556800 + }, + { + "epoch": 0.22, + "learning_rate": 0.0006267602808425277, + "loss": 1.5893, + "theoretical_loss": 3.786478352073626, + "tokens_seen": 740556800 + }, + { + "epoch": 0.22, + "learning_rate": 0.0006267442326980943, + "loss": 1.4461, + "theoretical_loss": 3.786444924615788, + "tokens_seen": 740622336 + }, + { + "epoch": 0.22, + "learning_rate": 0.0006267281845536611, + "loss": 1.5193, + "theoretical_loss": 3.7864115009438724, + "tokens_seen": 740687872 + }, + { + "epoch": 0.22, + "learning_rate": 0.0006267121364092277, + "loss": 1.4332, + "theoretical_loss": 3.7863780810571175, + "tokens_seen": 740753408 + }, + { + "epoch": 0.22, + "learning_rate": 0.0006266960882647945, + "loss": 1.4306, + "theoretical_loss": 3.7863446649547594, + "tokens_seen": 740818944 + }, + { + "epoch": 0.22, + "learning_rate": 0.0006266800401203611, + "loss": 1.5125, + "theoretical_loss": 3.786311252636034, + "tokens_seen": 740884480 + }, + { + "epoch": 0.22, + "learning_rate": 0.0006266639919759278, + "loss": 1.4136, + "theoretical_loss": 3.7862778441001783, + "tokens_seen": 740950016 + }, + { + "epoch": 0.22, + "learning_rate": 0.0006266479438314945, + "loss": 1.3457, + "theoretical_loss": 3.7862444393464307, + "tokens_seen": 741015552 + }, + { + "epoch": 0.22, + "learning_rate": 0.0006266318956870612, + "loss": 1.6687, + "theoretical_loss": 3.7862110383740273, + "tokens_seen": 741081088 + }, + { + "epoch": 0.22, + "learning_rate": 0.000626615847542628, + "loss": 1.3714, + "theoretical_loss": 3.7861776411822063, + "tokens_seen": 741146624 + }, + { + "epoch": 0.22, + "learning_rate": 0.0006265997993981946, + "loss": 1.3885, + "theoretical_loss": 3.7861442477702063, + "tokens_seen": 741212160 + }, + { + "epoch": 0.22, + "learning_rate": 0.0006265837512537613, + "loss": 1.5109, + "theoretical_loss": 3.7861108581372633, + "tokens_seen": 741277696 + }, + { + "epoch": 0.22, + "learning_rate": 0.000626567703109328, + "loss": 1.4391, + "theoretical_loss": 3.786077472282618, + "tokens_seen": 741343232 + }, + { + "epoch": 0.22, + "learning_rate": 0.0006265516549648948, + "loss": 1.514, + "theoretical_loss": 3.786044090205507, + "tokens_seen": 741408768 + }, + { + "epoch": 0.22, + "learning_rate": 0.0006265356068204614, + "loss": 1.5963, + "theoretical_loss": 3.7860107119051696, + "tokens_seen": 741474304 + }, + { + "epoch": 0.22, + "learning_rate": 0.000626519558676028, + "loss": 1.5099, + "theoretical_loss": 3.785977337380846, + "tokens_seen": 741539840 + }, + { + "epoch": 0.22, + "learning_rate": 0.0006265035105315948, + "loss": 1.5001, + "theoretical_loss": 3.7859439666317733, + "tokens_seen": 741605376 + }, + { + "epoch": 0.22, + "learning_rate": 0.0006264874623871614, + "loss": 1.4956, + "theoretical_loss": 3.7859105996571922, + "tokens_seen": 741670912 + }, + { + "epoch": 0.22, + "learning_rate": 0.0006264714142427282, + "loss": 1.4928, + "theoretical_loss": 3.7858772364563418, + "tokens_seen": 741736448 + }, + { + "epoch": 0.22, + "learning_rate": 0.0006264553660982948, + "loss": 1.4322, + "theoretical_loss": 3.785843877028462, + "tokens_seen": 741801984 + }, + { + "epoch": 0.22, + "learning_rate": 0.0006264393179538616, + "loss": 1.5538, + "theoretical_loss": 3.7858105213727935, + "tokens_seen": 741867520 + }, + { + "epoch": 0.22, + "learning_rate": 0.0006264232698094282, + "loss": 1.4734, + "theoretical_loss": 3.7857771694885756, + "tokens_seen": 741933056 + }, + { + "epoch": 0.22, + "learning_rate": 0.000626407221664995, + "loss": 1.5061, + "theoretical_loss": 3.785743821375049, + "tokens_seen": 741998592 + }, + { + "epoch": 0.22, + "learning_rate": 0.0006263911735205616, + "loss": 1.4164, + "theoretical_loss": 3.785710477031455, + "tokens_seen": 742064128 + }, + { + "epoch": 0.22, + "learning_rate": 0.0006263751253761284, + "loss": 1.5097, + "theoretical_loss": 3.785677136457034, + "tokens_seen": 742129664 + }, + { + "epoch": 0.22, + "learning_rate": 0.000626359077231695, + "loss": 1.4381, + "theoretical_loss": 3.7856437996510275, + "tokens_seen": 742195200 + }, + { + "epoch": 0.22, + "learning_rate": 0.0006263430290872618, + "loss": 1.4481, + "theoretical_loss": 3.7856104666126766, + "tokens_seen": 742260736 + }, + { + "epoch": 0.22, + "learning_rate": 0.0006263269809428285, + "loss": 1.5558, + "theoretical_loss": 3.785577137341223, + "tokens_seen": 742326272 + }, + { + "epoch": 0.22, + "learning_rate": 0.0006263109327983952, + "loss": 1.4607, + "theoretical_loss": 3.7855438118359084, + "tokens_seen": 742391808 + }, + { + "epoch": 0.22, + "learning_rate": 0.0006262948846539619, + "loss": 1.4972, + "theoretical_loss": 3.785510490095975, + "tokens_seen": 742457344 + }, + { + "epoch": 0.23, + "learning_rate": 0.0006262788365095286, + "loss": 1.5519, + "theoretical_loss": 3.7854771721206646, + "tokens_seen": 742522880 + }, + { + "epoch": 0.23, + "learning_rate": 0.0006262627883650953, + "loss": 1.4756, + "theoretical_loss": 3.78544385790922, + "tokens_seen": 742588416 + }, + { + "epoch": 0.23, + "learning_rate": 0.000626246740220662, + "loss": 1.5144, + "theoretical_loss": 3.7854105474608843, + "tokens_seen": 742653952 + }, + { + "epoch": 0.23, + "learning_rate": 0.0006262306920762287, + "loss": 1.5878, + "theoretical_loss": 3.785377240774899, + "tokens_seen": 742719488 + }, + { + "epoch": 0.23, + "learning_rate": 0.0006262146439317954, + "loss": 1.2892, + "theoretical_loss": 3.7853439378505094, + "tokens_seen": 742785024 + }, + { + "epoch": 0.23, + "learning_rate": 0.0006261985957873621, + "loss": 1.3347, + "theoretical_loss": 3.7853106386869575, + "tokens_seen": 742850560 + }, + { + "epoch": 0.23, + "learning_rate": 0.0006261825476429288, + "loss": 1.5376, + "theoretical_loss": 3.7852773432834868, + "tokens_seen": 742916096 + }, + { + "epoch": 0.23, + "learning_rate": 0.0006261664994984955, + "loss": 1.5233, + "theoretical_loss": 3.7852440516393413, + "tokens_seen": 742981632 + }, + { + "epoch": 0.23, + "learning_rate": 0.0006261504513540622, + "loss": 1.5287, + "theoretical_loss": 3.785210763753765, + "tokens_seen": 743047168 + }, + { + "epoch": 0.23, + "learning_rate": 0.0006261344032096289, + "loss": 1.5133, + "theoretical_loss": 3.785177479626002, + "tokens_seen": 743112704 + }, + { + "epoch": 0.23, + "learning_rate": 0.0006261183550651956, + "loss": 1.5114, + "theoretical_loss": 3.785144199255297, + "tokens_seen": 743178240 + }, + { + "epoch": 0.23, + "learning_rate": 0.0006261023069207623, + "loss": 1.3628, + "theoretical_loss": 3.7851109226408943, + "tokens_seen": 743243776 + }, + { + "epoch": 0.23, + "learning_rate": 0.000626086258776329, + "loss": 1.5179, + "theoretical_loss": 3.785077649782039, + "tokens_seen": 743309312 + }, + { + "epoch": 0.23, + "learning_rate": 0.0006260702106318957, + "loss": 1.4877, + "theoretical_loss": 3.7850443806779763, + "tokens_seen": 743374848 + }, + { + "epoch": 0.23, + "learning_rate": 0.0006260541624874624, + "loss": 1.5836, + "theoretical_loss": 3.7850111153279515, + "tokens_seen": 743440384 + }, + { + "epoch": 0.23, + "learning_rate": 0.0006260381143430291, + "loss": 1.4574, + "theoretical_loss": 3.78497785373121, + "tokens_seen": 743505920 + }, + { + "epoch": 0.23, + "learning_rate": 0.0006260220661985958, + "loss": 1.4646, + "theoretical_loss": 3.784944595886998, + "tokens_seen": 743571456 + }, + { + "epoch": 0.23, + "learning_rate": 0.0006260060180541625, + "loss": 1.3568, + "theoretical_loss": 3.7849113417945603, + "tokens_seen": 743636992 + }, + { + "epoch": 0.23, + "learning_rate": 0.0006259899699097292, + "loss": 1.5493, + "theoretical_loss": 3.7848780914531437, + "tokens_seen": 743702528 + }, + { + "epoch": 0.23, + "learning_rate": 0.0006259739217652959, + "loss": 1.6477, + "theoretical_loss": 3.784844844861995, + "tokens_seen": 743768064 + }, + { + "epoch": 0.23, + "objective/train/docs_used": 275779, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.596152424812317, + "objective/train/theoretical_loss": 3.784811602020361, + "objective/train/tokens_used": 764293600, + "theoretical_loss": 3.784811602020361, + "tokens_seen": 743833600 + }, + { + "epoch": 0.23, + "learning_rate": 0.0006259578736208626, + "loss": 1.5142, + "theoretical_loss": 3.784811602020361, + "tokens_seen": 743833600 + }, + { + "epoch": 0.23, + "learning_rate": 0.0006259418254764293, + "loss": 1.4673, + "theoretical_loss": 3.7847783629274874, + "tokens_seen": 743899136 + }, + { + "epoch": 0.23, + "learning_rate": 0.000625925777331996, + "loss": 1.5668, + "theoretical_loss": 3.7847451275826227, + "tokens_seen": 743964672 + }, + { + "epoch": 0.23, + "learning_rate": 0.0006259097291875627, + "loss": 1.5386, + "theoretical_loss": 3.784711895985013, + "tokens_seen": 744030208 + }, + { + "epoch": 0.23, + "learning_rate": 0.0006258936810431294, + "loss": 1.5786, + "theoretical_loss": 3.784678668133906, + "tokens_seen": 744095744 + }, + { + "epoch": 0.23, + "learning_rate": 0.0006258776328986961, + "loss": 1.4389, + "theoretical_loss": 3.7846454440285493, + "tokens_seen": 744161280 + }, + { + "epoch": 0.23, + "learning_rate": 0.0006258615847542628, + "loss": 1.4556, + "theoretical_loss": 3.7846122236681916, + "tokens_seen": 744226816 + }, + { + "epoch": 0.23, + "learning_rate": 0.0006258455366098295, + "loss": 1.4356, + "theoretical_loss": 3.7845790070520806, + "tokens_seen": 744292352 + }, + { + "epoch": 0.23, + "learning_rate": 0.0006258294884653962, + "loss": 1.4204, + "theoretical_loss": 3.784545794179464, + "tokens_seen": 744357888 + }, + { + "epoch": 0.23, + "learning_rate": 0.000625813440320963, + "loss": 1.4399, + "theoretical_loss": 3.7845125850495918, + "tokens_seen": 744423424 + }, + { + "epoch": 0.23, + "learning_rate": 0.0006257973921765296, + "loss": 1.4939, + "theoretical_loss": 3.7844793796617116, + "tokens_seen": 744488960 + }, + { + "epoch": 0.23, + "learning_rate": 0.0006257813440320964, + "loss": 1.4315, + "theoretical_loss": 3.784446178015073, + "tokens_seen": 744554496 + }, + { + "epoch": 0.23, + "learning_rate": 0.000625765295887663, + "loss": 1.4832, + "theoretical_loss": 3.7844129801089244, + "tokens_seen": 744620032 + }, + { + "epoch": 0.23, + "learning_rate": 0.0006257492477432298, + "loss": 1.5018, + "theoretical_loss": 3.784379785942517, + "tokens_seen": 744685568 + }, + { + "epoch": 0.23, + "learning_rate": 0.0006257331995987964, + "loss": 1.4822, + "theoretical_loss": 3.784346595515098, + "tokens_seen": 744751104 + }, + { + "epoch": 0.23, + "learning_rate": 0.0006257171514543632, + "loss": 1.5091, + "theoretical_loss": 3.784313408825919, + "tokens_seen": 744816640 + }, + { + "epoch": 0.23, + "learning_rate": 0.0006257011033099298, + "loss": 1.4786, + "theoretical_loss": 3.78428022587423, + "tokens_seen": 744882176 + }, + { + "epoch": 0.23, + "learning_rate": 0.0006256850551654966, + "loss": 1.385, + "theoretical_loss": 3.7842470466592806, + "tokens_seen": 744947712 + }, + { + "epoch": 0.23, + "learning_rate": 0.0006256690070210632, + "loss": 1.4544, + "theoretical_loss": 3.784213871180322, + "tokens_seen": 745013248 + }, + { + "epoch": 0.23, + "learning_rate": 0.00062565295887663, + "loss": 1.4953, + "theoretical_loss": 3.7841806994366047, + "tokens_seen": 745078784 + }, + { + "epoch": 0.23, + "learning_rate": 0.0006256369107321966, + "loss": 1.4452, + "theoretical_loss": 3.7841475314273794, + "tokens_seen": 745144320 + }, + { + "epoch": 0.23, + "learning_rate": 0.0006256208625877634, + "loss": 1.3252, + "theoretical_loss": 3.7841143671518975, + "tokens_seen": 745209856 + }, + { + "epoch": 0.23, + "learning_rate": 0.00062560481444333, + "loss": 1.6687, + "theoretical_loss": 3.7840812066094105, + "tokens_seen": 745275392 + }, + { + "epoch": 0.23, + "learning_rate": 0.0006255887662988968, + "loss": 1.4848, + "theoretical_loss": 3.78404804979917, + "tokens_seen": 745340928 + }, + { + "epoch": 0.23, + "learning_rate": 0.0006255727181544634, + "loss": 1.4772, + "theoretical_loss": 3.7840148967204277, + "tokens_seen": 745406464 + }, + { + "epoch": 0.23, + "learning_rate": 0.0006255566700100302, + "loss": 1.4451, + "theoretical_loss": 3.7839817473724353, + "tokens_seen": 745472000 + }, + { + "epoch": 0.23, + "learning_rate": 0.0006255406218655968, + "loss": 1.5283, + "theoretical_loss": 3.783948601754445, + "tokens_seen": 745537536 + }, + { + "epoch": 0.23, + "learning_rate": 0.0006255245737211636, + "loss": 1.5363, + "theoretical_loss": 3.7839154598657103, + "tokens_seen": 745603072 + }, + { + "epoch": 0.23, + "learning_rate": 0.0006255085255767302, + "loss": 1.4342, + "theoretical_loss": 3.7838823217054833, + "tokens_seen": 745668608 + }, + { + "epoch": 0.23, + "learning_rate": 0.000625492477432297, + "loss": 1.5228, + "theoretical_loss": 3.7838491872730167, + "tokens_seen": 745734144 + }, + { + "epoch": 0.23, + "learning_rate": 0.0006254764292878636, + "loss": 1.5021, + "theoretical_loss": 3.783816056567564, + "tokens_seen": 745799680 + }, + { + "epoch": 0.23, + "learning_rate": 0.0006254603811434304, + "loss": 1.4927, + "theoretical_loss": 3.783782929588378, + "tokens_seen": 745865216 + }, + { + "epoch": 0.23, + "learning_rate": 0.000625444332998997, + "loss": 1.4406, + "theoretical_loss": 3.783749806334712, + "tokens_seen": 745930752 + }, + { + "epoch": 0.23, + "learning_rate": 0.0006254282848545638, + "loss": 1.5366, + "theoretical_loss": 3.783716686805821, + "tokens_seen": 745996288 + }, + { + "epoch": 0.23, + "learning_rate": 0.0006254122367101304, + "loss": 1.4387, + "theoretical_loss": 3.7836835710009584, + "tokens_seen": 746061824 + }, + { + "epoch": 0.23, + "learning_rate": 0.0006253961885656972, + "loss": 1.5229, + "theoretical_loss": 3.783650458919378, + "tokens_seen": 746127360 + }, + { + "epoch": 0.23, + "learning_rate": 0.0006253801404212638, + "loss": 1.4265, + "theoretical_loss": 3.783617350560334, + "tokens_seen": 746192896 + }, + { + "epoch": 0.23, + "learning_rate": 0.0006253640922768306, + "loss": 1.4634, + "theoretical_loss": 3.7835842459230817, + "tokens_seen": 746258432 + }, + { + "epoch": 0.23, + "learning_rate": 0.0006253480441323972, + "loss": 1.3975, + "theoretical_loss": 3.783551145006876, + "tokens_seen": 746323968 + }, + { + "epoch": 0.23, + "learning_rate": 0.000625331995987964, + "loss": 1.4493, + "theoretical_loss": 3.783518047810971, + "tokens_seen": 746389504 + }, + { + "epoch": 0.23, + "learning_rate": 0.0006253159478435306, + "loss": 1.5553, + "theoretical_loss": 3.7834849543346234, + "tokens_seen": 746455040 + }, + { + "epoch": 0.23, + "learning_rate": 0.0006252998996990974, + "loss": 1.4886, + "theoretical_loss": 3.7834518645770876, + "tokens_seen": 746520576 + }, + { + "epoch": 0.23, + "learning_rate": 0.000625283851554664, + "loss": 1.4892, + "theoretical_loss": 3.7834187785376194, + "tokens_seen": 746586112 + }, + { + "epoch": 0.23, + "learning_rate": 0.0006252678034102307, + "loss": 1.5205, + "theoretical_loss": 3.7833856962154755, + "tokens_seen": 746651648 + }, + { + "epoch": 0.23, + "learning_rate": 0.0006252517552657974, + "loss": 1.5059, + "theoretical_loss": 3.7833526176099106, + "tokens_seen": 746717184 + }, + { + "epoch": 0.23, + "learning_rate": 0.0006252357071213641, + "loss": 1.3695, + "theoretical_loss": 3.7833195427201822, + "tokens_seen": 746782720 + }, + { + "epoch": 0.23, + "learning_rate": 0.0006252196589769308, + "loss": 1.4635, + "theoretical_loss": 3.783286471545547, + "tokens_seen": 746848256 + }, + { + "epoch": 0.23, + "learning_rate": 0.0006252036108324975, + "loss": 1.5373, + "theoretical_loss": 3.7832534040852606, + "tokens_seen": 746913792 + }, + { + "epoch": 0.23, + "learning_rate": 0.0006251875626880643, + "loss": 1.4417, + "theoretical_loss": 3.7832203403385805, + "tokens_seen": 746979328 + }, + { + "epoch": 0.23, + "learning_rate": 0.0006251715145436309, + "loss": 1.508, + "theoretical_loss": 3.783187280304764, + "tokens_seen": 747044864 + }, + { + "epoch": 0.23, + "objective/train/docs_used": 277055, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.4873147010803223, + "objective/train/theoretical_loss": 3.783154223983069, + "objective/train/tokens_used": 767570400, + "theoretical_loss": 3.783154223983069, + "tokens_seen": 747110400 + }, + { + "epoch": 0.23, + "learning_rate": 0.0006251554663991977, + "loss": 1.5653, + "theoretical_loss": 3.783154223983069, + "tokens_seen": 747110400 + }, + { + "epoch": 0.23, + "learning_rate": 0.0006251394182547643, + "loss": 1.4056, + "theoretical_loss": 3.783121171372753, + "tokens_seen": 747175936 + }, + { + "epoch": 0.23, + "learning_rate": 0.0006251233701103309, + "loss": 1.4471, + "theoretical_loss": 3.783088122473073, + "tokens_seen": 747241472 + }, + { + "epoch": 0.23, + "learning_rate": 0.0006251073219658977, + "loss": 1.5439, + "theoretical_loss": 3.7830550772832874, + "tokens_seen": 747307008 + }, + { + "epoch": 0.23, + "learning_rate": 0.0006250912738214643, + "loss": 1.531, + "theoretical_loss": 3.7830220358026545, + "tokens_seen": 747372544 + }, + { + "epoch": 0.23, + "learning_rate": 0.0006250752256770311, + "loss": 1.4095, + "theoretical_loss": 3.782988998030433, + "tokens_seen": 747438080 + }, + { + "epoch": 0.23, + "learning_rate": 0.0006250591775325977, + "loss": 1.4788, + "theoretical_loss": 3.7829559639658816, + "tokens_seen": 747503616 + }, + { + "epoch": 0.23, + "learning_rate": 0.0006250431293881645, + "loss": 1.4046, + "theoretical_loss": 3.7829229336082584, + "tokens_seen": 747569152 + }, + { + "epoch": 0.23, + "learning_rate": 0.0006250270812437311, + "loss": 1.4379, + "theoretical_loss": 3.7828899069568243, + "tokens_seen": 747634688 + }, + { + "epoch": 0.23, + "learning_rate": 0.0006250110330992979, + "loss": 1.4858, + "theoretical_loss": 3.7828568840108363, + "tokens_seen": 747700224 + }, + { + "epoch": 0.23, + "learning_rate": 0.0006249949849548645, + "loss": 1.4196, + "theoretical_loss": 3.782823864769556, + "tokens_seen": 747765760 + }, + { + "epoch": 0.23, + "learning_rate": 0.0006249789368104313, + "loss": 1.4175, + "theoretical_loss": 3.7827908492322413, + "tokens_seen": 747831296 + }, + { + "epoch": 0.23, + "learning_rate": 0.000624962888665998, + "loss": 1.4641, + "theoretical_loss": 3.7827578373981536, + "tokens_seen": 747896832 + }, + { + "epoch": 0.23, + "learning_rate": 0.0006249468405215647, + "loss": 1.507, + "theoretical_loss": 3.782724829266553, + "tokens_seen": 747962368 + }, + { + "epoch": 0.23, + "learning_rate": 0.0006249307923771314, + "loss": 1.3685, + "theoretical_loss": 3.782691824836699, + "tokens_seen": 748027904 + }, + { + "epoch": 0.23, + "learning_rate": 0.0006249147442326981, + "loss": 1.5718, + "theoretical_loss": 3.7826588241078527, + "tokens_seen": 748093440 + }, + { + "epoch": 0.23, + "learning_rate": 0.0006248986960882648, + "loss": 1.4112, + "theoretical_loss": 3.7826258270792747, + "tokens_seen": 748158976 + }, + { + "epoch": 0.23, + "learning_rate": 0.0006248826479438315, + "loss": 1.4672, + "theoretical_loss": 3.7825928337502264, + "tokens_seen": 748224512 + }, + { + "epoch": 0.23, + "learning_rate": 0.0006248665997993982, + "loss": 1.6522, + "theoretical_loss": 3.7825598441199695, + "tokens_seen": 748290048 + }, + { + "epoch": 0.23, + "learning_rate": 0.0006248505516549649, + "loss": 1.4679, + "theoretical_loss": 3.7825268581877642, + "tokens_seen": 748355584 + }, + { + "epoch": 0.23, + "learning_rate": 0.0006248345035105316, + "loss": 1.4929, + "theoretical_loss": 3.7824938759528726, + "tokens_seen": 748421120 + }, + { + "epoch": 0.23, + "learning_rate": 0.0006248184553660983, + "loss": 1.4991, + "theoretical_loss": 3.7824608974145573, + "tokens_seen": 748486656 + }, + { + "epoch": 0.23, + "learning_rate": 0.000624802407221665, + "loss": 1.4734, + "theoretical_loss": 3.782427922572079, + "tokens_seen": 748552192 + }, + { + "epoch": 0.23, + "learning_rate": 0.0006247863590772317, + "loss": 1.4985, + "theoretical_loss": 3.7823949514247017, + "tokens_seen": 748617728 + }, + { + "epoch": 0.23, + "learning_rate": 0.0006247703109327984, + "loss": 1.4438, + "theoretical_loss": 3.782361983971686, + "tokens_seen": 748683264 + }, + { + "epoch": 0.23, + "learning_rate": 0.0006247542627883651, + "loss": 1.5335, + "theoretical_loss": 3.7823290202122966, + "tokens_seen": 748748800 + }, + { + "epoch": 0.23, + "learning_rate": 0.0006247382146439318, + "loss": 1.3498, + "theoretical_loss": 3.7822960601457947, + "tokens_seen": 748814336 + }, + { + "epoch": 0.23, + "learning_rate": 0.0006247221664994985, + "loss": 1.4155, + "theoretical_loss": 3.782263103771445, + "tokens_seen": 748879872 + }, + { + "epoch": 0.23, + "learning_rate": 0.0006247061183550652, + "loss": 1.5884, + "theoretical_loss": 3.78223015108851, + "tokens_seen": 748945408 + }, + { + "epoch": 0.23, + "learning_rate": 0.0006246900702106319, + "loss": 1.4592, + "theoretical_loss": 3.7821972020962527, + "tokens_seen": 749010944 + }, + { + "epoch": 0.23, + "learning_rate": 0.0006246740220661986, + "loss": 1.5043, + "theoretical_loss": 3.782164256793938, + "tokens_seen": 749076480 + }, + { + "epoch": 0.23, + "learning_rate": 0.0006246579739217653, + "loss": 1.4703, + "theoretical_loss": 3.782131315180829, + "tokens_seen": 749142016 + }, + { + "epoch": 0.23, + "learning_rate": 0.000624641925777332, + "loss": 1.5092, + "theoretical_loss": 3.782098377256191, + "tokens_seen": 749207552 + }, + { + "epoch": 0.23, + "learning_rate": 0.0006246258776328987, + "loss": 1.3926, + "theoretical_loss": 3.7820654430192873, + "tokens_seen": 749273088 + }, + { + "epoch": 0.23, + "learning_rate": 0.0006246098294884655, + "loss": 1.4441, + "theoretical_loss": 3.782032512469383, + "tokens_seen": 749338624 + }, + { + "epoch": 0.23, + "learning_rate": 0.0006245937813440321, + "loss": 1.528, + "theoretical_loss": 3.781999585605743, + "tokens_seen": 749404160 + }, + { + "epoch": 0.23, + "learning_rate": 0.0006245777331995989, + "loss": 1.6026, + "theoretical_loss": 3.781966662427632, + "tokens_seen": 749469696 + }, + { + "epoch": 0.23, + "learning_rate": 0.0006245616850551655, + "loss": 1.5247, + "theoretical_loss": 3.7819337429343154, + "tokens_seen": 749535232 + }, + { + "epoch": 0.23, + "learning_rate": 0.0006245456369107323, + "loss": 1.479, + "theoretical_loss": 3.7819008271250594, + "tokens_seen": 749600768 + }, + { + "epoch": 0.23, + "learning_rate": 0.0006245295887662989, + "loss": 1.5742, + "theoretical_loss": 3.7818679149991286, + "tokens_seen": 749666304 + }, + { + "epoch": 0.23, + "learning_rate": 0.0006245135406218657, + "loss": 1.5514, + "theoretical_loss": 3.7818350065557893, + "tokens_seen": 749731840 + }, + { + "epoch": 0.23, + "learning_rate": 0.0006244974924774323, + "loss": 1.5924, + "theoretical_loss": 3.7818021017943075, + "tokens_seen": 749797376 + }, + { + "epoch": 0.23, + "learning_rate": 0.0006244814443329991, + "loss": 1.4129, + "theoretical_loss": 3.78176920071395, + "tokens_seen": 749862912 + }, + { + "epoch": 0.23, + "learning_rate": 0.0006244653961885657, + "loss": 1.454, + "theoretical_loss": 3.7817363033139824, + "tokens_seen": 749928448 + }, + { + "epoch": 0.23, + "learning_rate": 0.0006244493480441325, + "loss": 1.4852, + "theoretical_loss": 3.781703409593672, + "tokens_seen": 749993984 + }, + { + "epoch": 0.23, + "learning_rate": 0.0006244332998996991, + "loss": 1.4975, + "theoretical_loss": 3.7816705195522866, + "tokens_seen": 750059520 + }, + { + "epoch": 0.23, + "learning_rate": 0.0006244172517552659, + "loss": 1.5002, + "theoretical_loss": 3.7816376331890913, + "tokens_seen": 750125056 + }, + { + "epoch": 0.23, + "learning_rate": 0.0006244012036108325, + "loss": 1.5638, + "theoretical_loss": 3.7816047505033556, + "tokens_seen": 750190592 + }, + { + "epoch": 0.23, + "learning_rate": 0.0006243851554663993, + "loss": 1.546, + "theoretical_loss": 3.7815718714943456, + "tokens_seen": 750256128 + }, + { + "epoch": 0.23, + "learning_rate": 0.0006243691073219659, + "loss": 1.4295, + "theoretical_loss": 3.7815389961613297, + "tokens_seen": 750321664 + }, + { + "epoch": 0.23, + "objective/train/docs_used": 278273, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.4335839748382568, + "objective/train/theoretical_loss": 3.7815061245035757, + "objective/train/tokens_used": 770847200, + "theoretical_loss": 3.7815061245035757, + "tokens_seen": 750387200 + }, + { + "epoch": 0.23, + "learning_rate": 0.0006243530591775327, + "loss": 1.4593, + "theoretical_loss": 3.7815061245035757, + "tokens_seen": 750387200 + }, + { + "epoch": 0.23, + "learning_rate": 0.0006243370110330993, + "loss": 1.4782, + "theoretical_loss": 3.7814732565203517, + "tokens_seen": 750452736 + }, + { + "epoch": 0.23, + "learning_rate": 0.0006243209628886661, + "loss": 1.5137, + "theoretical_loss": 3.7814403922109263, + "tokens_seen": 750518272 + }, + { + "epoch": 0.23, + "learning_rate": 0.0006243049147442327, + "loss": 1.3339, + "theoretical_loss": 3.781407531574568, + "tokens_seen": 750583808 + }, + { + "epoch": 0.23, + "learning_rate": 0.0006242888665997995, + "loss": 1.4416, + "theoretical_loss": 3.7813746746105465, + "tokens_seen": 750649344 + }, + { + "epoch": 0.23, + "learning_rate": 0.0006242728184553661, + "loss": 1.4155, + "theoretical_loss": 3.7813418213181293, + "tokens_seen": 750714880 + }, + { + "epoch": 0.23, + "learning_rate": 0.0006242567703109329, + "loss": 1.3753, + "theoretical_loss": 3.7813089716965864, + "tokens_seen": 750780416 + }, + { + "epoch": 0.23, + "learning_rate": 0.0006242407221664995, + "loss": 1.4127, + "theoretical_loss": 3.7812761257451872, + "tokens_seen": 750845952 + }, + { + "epoch": 0.23, + "learning_rate": 0.0006242246740220663, + "loss": 1.5196, + "theoretical_loss": 3.7812432834632017, + "tokens_seen": 750911488 + }, + { + "epoch": 0.23, + "learning_rate": 0.0006242086258776329, + "loss": 1.4397, + "theoretical_loss": 3.7812104448498998, + "tokens_seen": 750977024 + }, + { + "epoch": 0.23, + "learning_rate": 0.0006241925777331997, + "loss": 1.3919, + "theoretical_loss": 3.7811776099045504, + "tokens_seen": 751042560 + }, + { + "epoch": 0.23, + "learning_rate": 0.0006241765295887663, + "loss": 1.5204, + "theoretical_loss": 3.781144778626425, + "tokens_seen": 751108096 + }, + { + "epoch": 0.23, + "learning_rate": 0.0006241604814443331, + "loss": 1.507, + "theoretical_loss": 3.781111951014794, + "tokens_seen": 751173632 + }, + { + "epoch": 0.23, + "learning_rate": 0.0006241444332998997, + "loss": 1.5351, + "theoretical_loss": 3.781079127068928, + "tokens_seen": 751239168 + }, + { + "epoch": 0.23, + "learning_rate": 0.0006241283851554665, + "loss": 1.5038, + "theoretical_loss": 3.781046306788097, + "tokens_seen": 751304704 + }, + { + "epoch": 0.23, + "learning_rate": 0.0006241123370110331, + "loss": 1.5508, + "theoretical_loss": 3.7810134901715733, + "tokens_seen": 751370240 + }, + { + "epoch": 0.23, + "learning_rate": 0.0006240962888665999, + "loss": 1.5323, + "theoretical_loss": 3.7809806772186283, + "tokens_seen": 751435776 + }, + { + "epoch": 0.23, + "learning_rate": 0.0006240802407221665, + "loss": 1.5233, + "theoretical_loss": 3.780947867928532, + "tokens_seen": 751501312 + }, + { + "epoch": 0.23, + "learning_rate": 0.0006240641925777333, + "loss": 1.4781, + "theoretical_loss": 3.7809150623005583, + "tokens_seen": 751566848 + }, + { + "epoch": 0.23, + "learning_rate": 0.0006240481444332999, + "loss": 1.5185, + "theoretical_loss": 3.7808822603339776, + "tokens_seen": 751632384 + }, + { + "epoch": 0.23, + "learning_rate": 0.0006240320962888667, + "loss": 1.5327, + "theoretical_loss": 3.780849462028063, + "tokens_seen": 751697920 + }, + { + "epoch": 0.23, + "learning_rate": 0.0006240160481444333, + "loss": 1.4523, + "theoretical_loss": 3.780816667382086, + "tokens_seen": 751763456 + }, + { + "epoch": 0.23, + "learning_rate": 0.0006240000000000001, + "loss": 1.5185, + "theoretical_loss": 3.7807838763953194, + "tokens_seen": 751828992 + }, + { + "epoch": 0.23, + "learning_rate": 0.0006239839518555667, + "loss": 1.39, + "theoretical_loss": 3.780751089067037, + "tokens_seen": 751894528 + }, + { + "epoch": 0.23, + "learning_rate": 0.0006239679037111335, + "loss": 1.4935, + "theoretical_loss": 3.78071830539651, + "tokens_seen": 751960064 + }, + { + "epoch": 0.23, + "learning_rate": 0.0006239518555667001, + "loss": 1.5101, + "theoretical_loss": 3.7806855253830136, + "tokens_seen": 752025600 + }, + { + "epoch": 0.23, + "learning_rate": 0.0006239358074222669, + "loss": 1.5328, + "theoretical_loss": 3.78065274902582, + "tokens_seen": 752091136 + }, + { + "epoch": 0.23, + "learning_rate": 0.0006239197592778335, + "loss": 1.4857, + "theoretical_loss": 3.7806199763242025, + "tokens_seen": 752156672 + }, + { + "epoch": 0.23, + "learning_rate": 0.0006239037111334003, + "loss": 1.4585, + "theoretical_loss": 3.780587207277436, + "tokens_seen": 752222208 + }, + { + "epoch": 0.23, + "learning_rate": 0.000623887662988967, + "loss": 1.4665, + "theoretical_loss": 3.7805544418847936, + "tokens_seen": 752287744 + }, + { + "epoch": 0.23, + "learning_rate": 0.0006238716148445336, + "loss": 1.5297, + "theoretical_loss": 3.7805216801455503, + "tokens_seen": 752353280 + }, + { + "epoch": 0.23, + "learning_rate": 0.0006238555667001003, + "loss": 1.3779, + "theoretical_loss": 3.7804889220589804, + "tokens_seen": 752418816 + }, + { + "epoch": 0.23, + "learning_rate": 0.000623839518555667, + "loss": 1.6051, + "theoretical_loss": 3.780456167624358, + "tokens_seen": 752484352 + }, + { + "epoch": 0.23, + "learning_rate": 0.0006238234704112337, + "loss": 1.498, + "theoretical_loss": 3.780423416840958, + "tokens_seen": 752549888 + }, + { + "epoch": 0.23, + "learning_rate": 0.0006238074222668004, + "loss": 1.6388, + "theoretical_loss": 3.7803906697080567, + "tokens_seen": 752615424 + }, + { + "epoch": 0.23, + "learning_rate": 0.0006237913741223672, + "loss": 1.4457, + "theoretical_loss": 3.780357926224928, + "tokens_seen": 752680960 + }, + { + "epoch": 0.23, + "learning_rate": 0.0006237753259779338, + "loss": 1.5307, + "theoretical_loss": 3.780325186390848, + "tokens_seen": 752746496 + }, + { + "epoch": 0.23, + "learning_rate": 0.0006237592778335006, + "loss": 1.6023, + "theoretical_loss": 3.780292450205092, + "tokens_seen": 752812032 + }, + { + "epoch": 0.23, + "learning_rate": 0.0006237432296890672, + "loss": 1.4785, + "theoretical_loss": 3.7802597176669366, + "tokens_seen": 752877568 + }, + { + "epoch": 0.23, + "learning_rate": 0.0006237271815446338, + "loss": 1.3839, + "theoretical_loss": 3.7802269887756577, + "tokens_seen": 752943104 + }, + { + "epoch": 0.23, + "learning_rate": 0.0006237111334002006, + "loss": 1.3929, + "theoretical_loss": 3.7801942635305306, + "tokens_seen": 753008640 + }, + { + "epoch": 0.23, + "learning_rate": 0.0006236950852557672, + "loss": 1.5331, + "theoretical_loss": 3.780161541930833, + "tokens_seen": 753074176 + }, + { + "epoch": 0.23, + "learning_rate": 0.000623679037111334, + "loss": 1.466, + "theoretical_loss": 3.7801288239758417, + "tokens_seen": 753139712 + }, + { + "epoch": 0.23, + "learning_rate": 0.0006236629889669006, + "loss": 1.5206, + "theoretical_loss": 3.7800961096648327, + "tokens_seen": 753205248 + }, + { + "epoch": 0.23, + "learning_rate": 0.0006236469408224674, + "loss": 1.4832, + "theoretical_loss": 3.7800633989970835, + "tokens_seen": 753270784 + }, + { + "epoch": 0.23, + "learning_rate": 0.000623630892678034, + "loss": 1.4336, + "theoretical_loss": 3.780030691971872, + "tokens_seen": 753336320 + }, + { + "epoch": 0.23, + "learning_rate": 0.0006236148445336008, + "loss": 1.5405, + "theoretical_loss": 3.7799979885884754, + "tokens_seen": 753401856 + }, + { + "epoch": 0.23, + "learning_rate": 0.0006235987963891675, + "loss": 1.5613, + "theoretical_loss": 3.7799652888461712, + "tokens_seen": 753467392 + }, + { + "epoch": 0.23, + "learning_rate": 0.0006235827482447342, + "loss": 1.5436, + "theoretical_loss": 3.7799325927442373, + "tokens_seen": 753532928 + }, + { + "epoch": 0.23, + "learning_rate": 0.0006235667001003009, + "loss": 1.4826, + "theoretical_loss": 3.7798999002819524, + "tokens_seen": 753598464 + }, + { + "epoch": 0.23, + "objective/train/docs_used": 279492, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.5246219635009766, + "objective/train/theoretical_loss": 3.779867211458595, + "objective/train/tokens_used": 774124000, + "theoretical_loss": 3.779867211458595, + "tokens_seen": 753664000 + }, + { + "epoch": 0.23, + "learning_rate": 0.0006235506519558676, + "loss": 1.4749, + "theoretical_loss": 3.779867211458595, + "tokens_seen": 753664000 + }, + { + "epoch": 0.23, + "learning_rate": 0.0006235346038114343, + "loss": 1.4938, + "theoretical_loss": 3.7798345262734427, + "tokens_seen": 753729536 + }, + { + "epoch": 0.23, + "learning_rate": 0.000623518555667001, + "loss": 1.4973, + "theoretical_loss": 3.779801844725775, + "tokens_seen": 753795072 + }, + { + "epoch": 0.23, + "learning_rate": 0.0006235025075225677, + "loss": 1.4651, + "theoretical_loss": 3.779769166814871, + "tokens_seen": 753860608 + }, + { + "epoch": 0.23, + "learning_rate": 0.0006234864593781344, + "loss": 1.4894, + "theoretical_loss": 3.7797364925400094, + "tokens_seen": 753926144 + }, + { + "epoch": 0.23, + "learning_rate": 0.0006234704112337011, + "loss": 1.4931, + "theoretical_loss": 3.77970382190047, + "tokens_seen": 753991680 + }, + { + "epoch": 0.23, + "learning_rate": 0.0006234543630892678, + "loss": 1.4725, + "theoretical_loss": 3.7796711548955324, + "tokens_seen": 754057216 + }, + { + "epoch": 0.23, + "learning_rate": 0.0006234383149448345, + "loss": 1.5268, + "theoretical_loss": 3.7796384915244765, + "tokens_seen": 754122752 + }, + { + "epoch": 0.23, + "learning_rate": 0.0006234222668004012, + "loss": 1.448, + "theoretical_loss": 3.7796058317865815, + "tokens_seen": 754188288 + }, + { + "epoch": 0.23, + "learning_rate": 0.000623406218655968, + "loss": 1.4811, + "theoretical_loss": 3.7795731756811284, + "tokens_seen": 754253824 + }, + { + "epoch": 0.23, + "learning_rate": 0.0006233901705115346, + "loss": 1.5739, + "theoretical_loss": 3.779540523207398, + "tokens_seen": 754319360 + }, + { + "epoch": 0.23, + "learning_rate": 0.0006233741223671014, + "loss": 1.5301, + "theoretical_loss": 3.77950787436467, + "tokens_seen": 754384896 + }, + { + "epoch": 0.23, + "learning_rate": 0.000623358074222668, + "loss": 1.426, + "theoretical_loss": 3.779475229152226, + "tokens_seen": 754450432 + }, + { + "epoch": 0.23, + "learning_rate": 0.0006233420260782348, + "loss": 1.5416, + "theoretical_loss": 3.7794425875693465, + "tokens_seen": 754515968 + }, + { + "epoch": 0.23, + "learning_rate": 0.0006233259779338014, + "loss": 1.4891, + "theoretical_loss": 3.779409949615313, + "tokens_seen": 754581504 + }, + { + "epoch": 0.23, + "learning_rate": 0.0006233099297893682, + "loss": 1.5073, + "theoretical_loss": 3.7793773152894072, + "tokens_seen": 754647040 + }, + { + "epoch": 0.23, + "learning_rate": 0.0006232938816449348, + "loss": 1.5291, + "theoretical_loss": 3.77934468459091, + "tokens_seen": 754712576 + }, + { + "epoch": 0.23, + "learning_rate": 0.0006232778335005016, + "loss": 1.4249, + "theoretical_loss": 3.779312057519104, + "tokens_seen": 754778112 + }, + { + "epoch": 0.23, + "learning_rate": 0.0006232617853560682, + "loss": 1.4964, + "theoretical_loss": 3.7792794340732714, + "tokens_seen": 754843648 + }, + { + "epoch": 0.23, + "learning_rate": 0.000623245737211635, + "loss": 1.5436, + "theoretical_loss": 3.7792468142526934, + "tokens_seen": 754909184 + }, + { + "epoch": 0.23, + "learning_rate": 0.0006232296890672016, + "loss": 1.4279, + "theoretical_loss": 3.7792141980566534, + "tokens_seen": 754974720 + }, + { + "epoch": 0.23, + "learning_rate": 0.0006232136409227684, + "loss": 1.4135, + "theoretical_loss": 3.7791815854844337, + "tokens_seen": 755040256 + }, + { + "epoch": 0.23, + "learning_rate": 0.000623197592778335, + "loss": 1.5046, + "theoretical_loss": 3.7791489765353177, + "tokens_seen": 755105792 + }, + { + "epoch": 0.23, + "learning_rate": 0.0006231815446339018, + "loss": 1.4913, + "theoretical_loss": 3.7791163712085885, + "tokens_seen": 755171328 + }, + { + "epoch": 0.23, + "learning_rate": 0.0006231654964894684, + "loss": 1.4244, + "theoretical_loss": 3.7790837695035284, + "tokens_seen": 755236864 + }, + { + "epoch": 0.23, + "learning_rate": 0.0006231494483450352, + "loss": 1.4424, + "theoretical_loss": 3.7790511714194217, + "tokens_seen": 755302400 + }, + { + "epoch": 0.23, + "learning_rate": 0.0006231334002006018, + "loss": 1.5435, + "theoretical_loss": 3.779018576955552, + "tokens_seen": 755367936 + }, + { + "epoch": 0.23, + "learning_rate": 0.0006231173520561686, + "loss": 1.4671, + "theoretical_loss": 3.778985986111203, + "tokens_seen": 755433472 + }, + { + "epoch": 0.23, + "learning_rate": 0.0006231013039117352, + "loss": 1.599, + "theoretical_loss": 3.778953398885659, + "tokens_seen": 755499008 + }, + { + "epoch": 0.23, + "learning_rate": 0.000623085255767302, + "loss": 1.5345, + "theoretical_loss": 3.7789208152782043, + "tokens_seen": 755564544 + }, + { + "epoch": 0.23, + "learning_rate": 0.0006230692076228686, + "loss": 1.4984, + "theoretical_loss": 3.7788882352881235, + "tokens_seen": 755630080 + }, + { + "epoch": 0.23, + "learning_rate": 0.0006230531594784354, + "loss": 1.4317, + "theoretical_loss": 3.7788556589147015, + "tokens_seen": 755695616 + }, + { + "epoch": 0.23, + "learning_rate": 0.000623037111334002, + "loss": 1.5276, + "theoretical_loss": 3.7788230861572227, + "tokens_seen": 755761152 + }, + { + "epoch": 0.23, + "learning_rate": 0.0006230210631895688, + "loss": 1.4418, + "theoretical_loss": 3.778790517014972, + "tokens_seen": 755826688 + }, + { + "epoch": 0.23, + "learning_rate": 0.0006230050150451354, + "loss": 1.4974, + "theoretical_loss": 3.7787579514872354, + "tokens_seen": 755892224 + }, + { + "epoch": 0.23, + "learning_rate": 0.0006229889669007022, + "loss": 1.4289, + "theoretical_loss": 3.7787253895732986, + "tokens_seen": 755957760 + }, + { + "epoch": 0.23, + "learning_rate": 0.0006229729187562688, + "loss": 1.4363, + "theoretical_loss": 3.778692831272447, + "tokens_seen": 756023296 + }, + { + "epoch": 0.23, + "learning_rate": 0.0006229568706118356, + "loss": 1.5236, + "theoretical_loss": 3.778660276583966, + "tokens_seen": 756088832 + }, + { + "epoch": 0.23, + "learning_rate": 0.0006229408224674022, + "loss": 1.602, + "theoretical_loss": 3.7786277255071425, + "tokens_seen": 756154368 + }, + { + "epoch": 0.23, + "learning_rate": 0.000622924774322969, + "loss": 1.5276, + "theoretical_loss": 3.7785951780412628, + "tokens_seen": 756219904 + }, + { + "epoch": 0.23, + "learning_rate": 0.0006229087261785356, + "loss": 1.4651, + "theoretical_loss": 3.7785626341856133, + "tokens_seen": 756285440 + }, + { + "epoch": 0.23, + "learning_rate": 0.0006228926780341024, + "loss": 1.3926, + "theoretical_loss": 3.77853009393948, + "tokens_seen": 756350976 + }, + { + "epoch": 0.23, + "learning_rate": 0.000622876629889669, + "loss": 1.5644, + "theoretical_loss": 3.7784975573021518, + "tokens_seen": 756416512 + }, + { + "epoch": 0.23, + "learning_rate": 0.0006228605817452358, + "loss": 1.3896, + "theoretical_loss": 3.7784650242729136, + "tokens_seen": 756482048 + }, + { + "epoch": 0.23, + "learning_rate": 0.0006228445336008024, + "loss": 1.4976, + "theoretical_loss": 3.7784324948510544, + "tokens_seen": 756547584 + }, + { + "epoch": 0.23, + "learning_rate": 0.0006228284854563692, + "loss": 1.4507, + "theoretical_loss": 3.7783999690358607, + "tokens_seen": 756613120 + }, + { + "epoch": 0.23, + "learning_rate": 0.0006228124373119358, + "loss": 1.423, + "theoretical_loss": 3.778367446826621, + "tokens_seen": 756678656 + }, + { + "epoch": 0.23, + "learning_rate": 0.0006227963891675026, + "loss": 1.5134, + "theoretical_loss": 3.778334928222623, + "tokens_seen": 756744192 + }, + { + "epoch": 0.23, + "learning_rate": 0.0006227803410230692, + "loss": 1.3877, + "theoretical_loss": 3.7783024132231544, + "tokens_seen": 756809728 + }, + { + "epoch": 0.23, + "learning_rate": 0.000622764292878636, + "loss": 1.5348, + "theoretical_loss": 3.7782699018275046, + "tokens_seen": 756875264 + }, + { + "epoch": 0.23, + "objective/train/docs_used": 280659, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.5990586280822754, + "objective/train/theoretical_loss": 3.778237394034961, + "objective/train/tokens_used": 777400800, + "theoretical_loss": 3.778237394034961, + "tokens_seen": 756940800 + }, + { + "epoch": 0.23, + "learning_rate": 0.0006227482447342026, + "loss": 1.5185, + "theoretical_loss": 3.778237394034961, + "tokens_seen": 756940800 + }, + { + "epoch": 0.23, + "learning_rate": 0.0006227321965897694, + "loss": 1.4749, + "theoretical_loss": 3.7782048898448135, + "tokens_seen": 757006336 + }, + { + "epoch": 0.23, + "learning_rate": 0.000622716148445336, + "loss": 1.549, + "theoretical_loss": 3.7781723892563503, + "tokens_seen": 757071872 + }, + { + "epoch": 0.23, + "learning_rate": 0.0006227001003009028, + "loss": 1.4091, + "theoretical_loss": 3.778139892268861, + "tokens_seen": 757137408 + }, + { + "epoch": 0.23, + "learning_rate": 0.0006226840521564694, + "loss": 1.5613, + "theoretical_loss": 3.7781073988816347, + "tokens_seen": 757202944 + }, + { + "epoch": 0.23, + "learning_rate": 0.0006226680040120362, + "loss": 1.4506, + "theoretical_loss": 3.7780749090939616, + "tokens_seen": 757268480 + }, + { + "epoch": 0.23, + "learning_rate": 0.0006226519558676028, + "loss": 1.556, + "theoretical_loss": 3.7780424229051306, + "tokens_seen": 757334016 + }, + { + "epoch": 0.23, + "learning_rate": 0.0006226359077231696, + "loss": 1.4686, + "theoretical_loss": 3.778009940314432, + "tokens_seen": 757399552 + }, + { + "epoch": 0.23, + "learning_rate": 0.0006226198595787362, + "loss": 1.4578, + "theoretical_loss": 3.7779774613211563, + "tokens_seen": 757465088 + }, + { + "epoch": 0.23, + "learning_rate": 0.000622603811434303, + "loss": 1.5195, + "theoretical_loss": 3.7779449859245933, + "tokens_seen": 757530624 + }, + { + "epoch": 0.23, + "learning_rate": 0.0006225877632898696, + "loss": 1.4805, + "theoretical_loss": 3.777912514124034, + "tokens_seen": 757596160 + }, + { + "epoch": 0.23, + "learning_rate": 0.0006225717151454364, + "loss": 1.4887, + "theoretical_loss": 3.7778800459187694, + "tokens_seen": 757661696 + }, + { + "epoch": 0.23, + "learning_rate": 0.000622555667001003, + "loss": 1.5093, + "theoretical_loss": 3.77784758130809, + "tokens_seen": 757727232 + }, + { + "epoch": 0.23, + "learning_rate": 0.0006225396188565698, + "loss": 1.4582, + "theoretical_loss": 3.7778151202912875, + "tokens_seen": 757792768 + }, + { + "epoch": 0.23, + "learning_rate": 0.0006225235707121364, + "loss": 1.516, + "theoretical_loss": 3.777782662867653, + "tokens_seen": 757858304 + }, + { + "epoch": 0.23, + "learning_rate": 0.0006225075225677032, + "loss": 1.5024, + "theoretical_loss": 3.777750209036478, + "tokens_seen": 757923840 + }, + { + "epoch": 0.23, + "learning_rate": 0.0006224914744232698, + "loss": 1.5109, + "theoretical_loss": 3.7777177587970545, + "tokens_seen": 757989376 + }, + { + "epoch": 0.23, + "learning_rate": 0.0006224754262788365, + "loss": 1.5412, + "theoretical_loss": 3.777685312148674, + "tokens_seen": 758054912 + }, + { + "epoch": 0.23, + "learning_rate": 0.0006224593781344032, + "loss": 1.5235, + "theoretical_loss": 3.77765286909063, + "tokens_seen": 758120448 + }, + { + "epoch": 0.23, + "learning_rate": 0.0006224433299899699, + "loss": 1.3521, + "theoretical_loss": 3.777620429622213, + "tokens_seen": 758185984 + }, + { + "epoch": 0.23, + "learning_rate": 0.0006224272818455367, + "loss": 1.4622, + "theoretical_loss": 3.7775879937427175, + "tokens_seen": 758251520 + }, + { + "epoch": 0.23, + "learning_rate": 0.0006224112337011033, + "loss": 1.4813, + "theoretical_loss": 3.777555561451435, + "tokens_seen": 758317056 + }, + { + "epoch": 0.23, + "learning_rate": 0.00062239518555667, + "loss": 1.5795, + "theoretical_loss": 3.777523132747659, + "tokens_seen": 758382592 + }, + { + "epoch": 0.23, + "learning_rate": 0.0006223791374122367, + "loss": 1.5242, + "theoretical_loss": 3.7774907076306823, + "tokens_seen": 758448128 + }, + { + "epoch": 0.23, + "learning_rate": 0.0006223630892678035, + "loss": 1.4499, + "theoretical_loss": 3.777458286099799, + "tokens_seen": 758513664 + }, + { + "epoch": 0.23, + "learning_rate": 0.0006223470411233701, + "loss": 1.5272, + "theoretical_loss": 3.7774258681543027, + "tokens_seen": 758579200 + }, + { + "epoch": 0.23, + "learning_rate": 0.0006223309929789367, + "loss": 1.6172, + "theoretical_loss": 3.7773934537934863, + "tokens_seen": 758644736 + }, + { + "epoch": 0.23, + "learning_rate": 0.0006223149448345035, + "loss": 1.5927, + "theoretical_loss": 3.7773610430166444, + "tokens_seen": 758710272 + }, + { + "epoch": 0.23, + "learning_rate": 0.0006222988966900701, + "loss": 1.6582, + "theoretical_loss": 3.7773286358230713, + "tokens_seen": 758775808 + }, + { + "epoch": 0.23, + "learning_rate": 0.0006222828485456369, + "loss": 1.4961, + "theoretical_loss": 3.777296232212061, + "tokens_seen": 758841344 + }, + { + "epoch": 0.23, + "learning_rate": 0.0006222668004012035, + "loss": 1.4932, + "theoretical_loss": 3.777263832182909, + "tokens_seen": 758906880 + }, + { + "epoch": 0.23, + "learning_rate": 0.0006222507522567703, + "loss": 1.4056, + "theoretical_loss": 3.7772314357349086, + "tokens_seen": 758972416 + }, + { + "epoch": 0.23, + "learning_rate": 0.000622234704112337, + "loss": 1.5389, + "theoretical_loss": 3.7771990428673563, + "tokens_seen": 759037952 + }, + { + "epoch": 0.23, + "learning_rate": 0.0006222186559679037, + "loss": 1.5327, + "theoretical_loss": 3.7771666535795463, + "tokens_seen": 759103488 + }, + { + "epoch": 0.23, + "learning_rate": 0.0006222026078234705, + "loss": 1.5914, + "theoretical_loss": 3.777134267870774, + "tokens_seen": 759169024 + }, + { + "epoch": 0.23, + "learning_rate": 0.0006221865596790371, + "loss": 1.3931, + "theoretical_loss": 3.7771018857403362, + "tokens_seen": 759234560 + }, + { + "epoch": 0.23, + "learning_rate": 0.0006221705115346039, + "loss": 1.572, + "theoretical_loss": 3.7770695071875275, + "tokens_seen": 759300096 + }, + { + "epoch": 0.23, + "learning_rate": 0.0006221544633901705, + "loss": 1.5227, + "theoretical_loss": 3.7770371322116443, + "tokens_seen": 759365632 + }, + { + "epoch": 0.23, + "learning_rate": 0.0006221384152457373, + "loss": 1.5411, + "theoretical_loss": 3.7770047608119826, + "tokens_seen": 759431168 + }, + { + "epoch": 0.23, + "learning_rate": 0.0006221223671013039, + "loss": 1.5672, + "theoretical_loss": 3.7769723929878394, + "tokens_seen": 759496704 + }, + { + "epoch": 0.23, + "learning_rate": 0.0006221063189568707, + "loss": 1.4435, + "theoretical_loss": 3.77694002873851, + "tokens_seen": 759562240 + }, + { + "epoch": 0.23, + "learning_rate": 0.0006220902708124373, + "loss": 1.4633, + "theoretical_loss": 3.7769076680632927, + "tokens_seen": 759627776 + }, + { + "epoch": 0.23, + "learning_rate": 0.0006220742226680041, + "loss": 1.4844, + "theoretical_loss": 3.7768753109614837, + "tokens_seen": 759693312 + }, + { + "epoch": 0.23, + "learning_rate": 0.0006220581745235707, + "loss": 1.4129, + "theoretical_loss": 3.7768429574323803, + "tokens_seen": 759758848 + }, + { + "epoch": 0.23, + "learning_rate": 0.0006220421263791375, + "loss": 1.3837, + "theoretical_loss": 3.77681060747528, + "tokens_seen": 759824384 + }, + { + "epoch": 0.23, + "learning_rate": 0.0006220260782347041, + "loss": 1.4898, + "theoretical_loss": 3.7767782610894804, + "tokens_seen": 759889920 + }, + { + "epoch": 0.23, + "learning_rate": 0.0006220100300902709, + "loss": 1.5236, + "theoretical_loss": 3.776745918274279, + "tokens_seen": 759955456 + }, + { + "epoch": 0.23, + "learning_rate": 0.0006219939819458375, + "loss": 1.5271, + "theoretical_loss": 3.776713579028974, + "tokens_seen": 760020992 + }, + { + "epoch": 0.23, + "learning_rate": 0.0006219779338014043, + "loss": 1.5302, + "theoretical_loss": 3.776681243352864, + "tokens_seen": 760086528 + }, + { + "epoch": 0.23, + "learning_rate": 0.0006219618856569709, + "loss": 1.646, + "theoretical_loss": 3.7766489112452466, + "tokens_seen": 760152064 + }, + { + "epoch": 0.23, + "objective/train/docs_used": 281753, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.5392972230911255, + "objective/train/theoretical_loss": 3.776616582705421, + "objective/train/tokens_used": 780677600, + "theoretical_loss": 3.776616582705421, + "tokens_seen": 760217600 + }, + { + "epoch": 0.23, + "learning_rate": 0.0006219458375125377, + "loss": 1.4844, + "theoretical_loss": 3.776616582705421, + "tokens_seen": 760217600 + }, + { + "epoch": 0.23, + "learning_rate": 0.0006219297893681043, + "loss": 1.4407, + "theoretical_loss": 3.7765842577326856, + "tokens_seen": 760283136 + }, + { + "epoch": 0.23, + "learning_rate": 0.0006219137412236711, + "loss": 1.6602, + "theoretical_loss": 3.7765519363263396, + "tokens_seen": 760348672 + }, + { + "epoch": 0.23, + "learning_rate": 0.0006218976930792377, + "loss": 1.4514, + "theoretical_loss": 3.7765196184856826, + "tokens_seen": 760414208 + }, + { + "epoch": 0.23, + "learning_rate": 0.0006218816449348045, + "loss": 1.5864, + "theoretical_loss": 3.776487304210012, + "tokens_seen": 760479744 + }, + { + "epoch": 0.23, + "learning_rate": 0.0006218655967903711, + "loss": 1.4294, + "theoretical_loss": 3.7764549934986302, + "tokens_seen": 760545280 + }, + { + "epoch": 0.23, + "learning_rate": 0.0006218495486459379, + "loss": 1.5508, + "theoretical_loss": 3.776422686350836, + "tokens_seen": 760610816 + }, + { + "epoch": 0.23, + "learning_rate": 0.0006218335005015045, + "loss": 1.4619, + "theoretical_loss": 3.7763903827659275, + "tokens_seen": 760676352 + }, + { + "epoch": 0.23, + "learning_rate": 0.0006218174523570713, + "loss": 1.5664, + "theoretical_loss": 3.776358082743208, + "tokens_seen": 760741888 + }, + { + "epoch": 0.23, + "learning_rate": 0.0006218014042126379, + "loss": 1.4168, + "theoretical_loss": 3.776325786281975, + "tokens_seen": 760807424 + }, + { + "epoch": 0.23, + "learning_rate": 0.0006217853560682047, + "loss": 1.4984, + "theoretical_loss": 3.776293493381531, + "tokens_seen": 760872960 + }, + { + "epoch": 0.23, + "learning_rate": 0.0006217693079237713, + "loss": 1.5607, + "theoretical_loss": 3.776261204041176, + "tokens_seen": 760938496 + }, + { + "epoch": 0.23, + "learning_rate": 0.0006217532597793381, + "loss": 1.3921, + "theoretical_loss": 3.7762289182602107, + "tokens_seen": 761004032 + }, + { + "epoch": 0.23, + "learning_rate": 0.0006217372116349047, + "loss": 1.4259, + "theoretical_loss": 3.7761966360379366, + "tokens_seen": 761069568 + }, + { + "epoch": 0.23, + "learning_rate": 0.0006217211634904715, + "loss": 1.4717, + "theoretical_loss": 3.776164357373655, + "tokens_seen": 761135104 + }, + { + "epoch": 0.23, + "learning_rate": 0.0006217051153460381, + "loss": 1.5487, + "theoretical_loss": 3.7761320822666677, + "tokens_seen": 761200640 + }, + { + "epoch": 0.23, + "learning_rate": 0.0006216890672016049, + "loss": 1.5615, + "theoretical_loss": 3.7760998107162767, + "tokens_seen": 761266176 + }, + { + "epoch": 0.23, + "learning_rate": 0.0006216730190571715, + "loss": 1.4359, + "theoretical_loss": 3.7760675427217825, + "tokens_seen": 761331712 + }, + { + "epoch": 0.23, + "learning_rate": 0.0006216569709127383, + "loss": 1.4661, + "theoretical_loss": 3.776035278282489, + "tokens_seen": 761397248 + }, + { + "epoch": 0.23, + "learning_rate": 0.0006216409227683049, + "loss": 1.4307, + "theoretical_loss": 3.7760030173976977, + "tokens_seen": 761462784 + }, + { + "epoch": 0.23, + "learning_rate": 0.0006216248746238717, + "loss": 1.4611, + "theoretical_loss": 3.775970760066711, + "tokens_seen": 761528320 + }, + { + "epoch": 0.23, + "learning_rate": 0.0006216088264794383, + "loss": 1.3507, + "theoretical_loss": 3.7759385062888318, + "tokens_seen": 761593856 + }, + { + "epoch": 0.23, + "learning_rate": 0.0006215927783350051, + "loss": 1.4204, + "theoretical_loss": 3.7759062560633634, + "tokens_seen": 761659392 + }, + { + "epoch": 0.23, + "learning_rate": 0.0006215767301905717, + "loss": 1.5124, + "theoretical_loss": 3.775874009389608, + "tokens_seen": 761724928 + }, + { + "epoch": 0.23, + "learning_rate": 0.0006215606820461385, + "loss": 1.5474, + "theoretical_loss": 3.77584176626687, + "tokens_seen": 761790464 + }, + { + "epoch": 0.23, + "learning_rate": 0.0006215446339017051, + "loss": 1.5057, + "theoretical_loss": 3.7758095266944522, + "tokens_seen": 761856000 + }, + { + "epoch": 0.23, + "learning_rate": 0.0006215285857572719, + "loss": 1.4685, + "theoretical_loss": 3.775777290671659, + "tokens_seen": 761921536 + }, + { + "epoch": 0.23, + "learning_rate": 0.0006215125376128385, + "loss": 1.4753, + "theoretical_loss": 3.775745058197794, + "tokens_seen": 761987072 + }, + { + "epoch": 0.23, + "learning_rate": 0.0006214964894684053, + "loss": 1.574, + "theoretical_loss": 3.7757128292721607, + "tokens_seen": 762052608 + }, + { + "epoch": 0.23, + "learning_rate": 0.0006214804413239719, + "loss": 1.6045, + "theoretical_loss": 3.775680603894064, + "tokens_seen": 762118144 + }, + { + "epoch": 0.23, + "learning_rate": 0.0006214643931795387, + "loss": 1.471, + "theoretical_loss": 3.7756483820628084, + "tokens_seen": 762183680 + }, + { + "epoch": 0.23, + "learning_rate": 0.0006214483450351053, + "loss": 1.4824, + "theoretical_loss": 3.7756161637776984, + "tokens_seen": 762249216 + }, + { + "epoch": 0.23, + "learning_rate": 0.0006214322968906721, + "loss": 1.5127, + "theoretical_loss": 3.775583949038039, + "tokens_seen": 762314752 + }, + { + "epoch": 0.23, + "learning_rate": 0.0006214162487462387, + "loss": 1.4915, + "theoretical_loss": 3.775551737843135, + "tokens_seen": 762380288 + }, + { + "epoch": 0.23, + "learning_rate": 0.0006214002006018055, + "loss": 1.5497, + "theoretical_loss": 3.775519530192293, + "tokens_seen": 762445824 + }, + { + "epoch": 0.23, + "learning_rate": 0.0006213841524573721, + "loss": 1.4942, + "theoretical_loss": 3.775487326084816, + "tokens_seen": 762511360 + }, + { + "epoch": 0.23, + "learning_rate": 0.0006213681043129389, + "loss": 1.4485, + "theoretical_loss": 3.775455125520012, + "tokens_seen": 762576896 + }, + { + "epoch": 0.23, + "learning_rate": 0.0006213520561685055, + "loss": 1.4392, + "theoretical_loss": 3.7754229284971856, + "tokens_seen": 762642432 + }, + { + "epoch": 0.23, + "learning_rate": 0.0006213360080240723, + "loss": 1.6094, + "theoretical_loss": 3.775390735015644, + "tokens_seen": 762707968 + }, + { + "epoch": 0.23, + "learning_rate": 0.0006213199598796389, + "loss": 1.4648, + "theoretical_loss": 3.7753585450746927, + "tokens_seen": 762773504 + }, + { + "epoch": 0.23, + "learning_rate": 0.0006213039117352057, + "loss": 1.4942, + "theoretical_loss": 3.775326358673637, + "tokens_seen": 762839040 + }, + { + "epoch": 0.23, + "learning_rate": 0.0006212878635907723, + "loss": 1.506, + "theoretical_loss": 3.775294175811786, + "tokens_seen": 762904576 + }, + { + "epoch": 0.23, + "learning_rate": 0.0006212718154463391, + "loss": 1.5801, + "theoretical_loss": 3.775261996488445, + "tokens_seen": 762970112 + }, + { + "epoch": 0.23, + "learning_rate": 0.0006212557673019057, + "loss": 1.3736, + "theoretical_loss": 3.775229820702921, + "tokens_seen": 763035648 + }, + { + "epoch": 0.23, + "learning_rate": 0.0006212397191574725, + "loss": 1.4868, + "theoretical_loss": 3.7751976484545224, + "tokens_seen": 763101184 + }, + { + "epoch": 0.23, + "learning_rate": 0.0006212236710130391, + "loss": 1.521, + "theoretical_loss": 3.7751654797425553, + "tokens_seen": 763166720 + }, + { + "epoch": 0.23, + "learning_rate": 0.0006212076228686059, + "loss": 1.5436, + "theoretical_loss": 3.775133314566328, + "tokens_seen": 763232256 + }, + { + "epoch": 0.23, + "learning_rate": 0.0006211915747241725, + "loss": 1.4412, + "theoretical_loss": 3.7751011529251475, + "tokens_seen": 763297792 + }, + { + "epoch": 0.23, + "learning_rate": 0.0006211755265797393, + "loss": 1.5205, + "theoretical_loss": 3.7750689948183234, + "tokens_seen": 763363328 + }, + { + "epoch": 0.23, + "learning_rate": 0.000621159478435306, + "loss": 1.6122, + "theoretical_loss": 3.7750368402451624, + "tokens_seen": 763428864 + }, + { + "epoch": 0.23, + "objective/train/docs_used": 282838, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.400315284729004, + "objective/train/theoretical_loss": 3.775004689204974, + "objective/train/tokens_used": 783954400, + "theoretical_loss": 3.775004689204974, + "tokens_seen": 763494400 + }, + { + "epoch": 0.23, + "learning_rate": 0.0006211434302908727, + "loss": 1.4989, + "theoretical_loss": 3.775004689204974, + "tokens_seen": 763494400 + }, + { + "epoch": 0.23, + "learning_rate": 0.0006211273821464393, + "loss": 1.5924, + "theoretical_loss": 3.7749725416970668, + "tokens_seen": 763559936 + }, + { + "epoch": 0.23, + "learning_rate": 0.0006211113340020061, + "loss": 1.5168, + "theoretical_loss": 3.7749403977207483, + "tokens_seen": 763625472 + }, + { + "epoch": 0.23, + "learning_rate": 0.0006210952858575727, + "loss": 1.4632, + "theoretical_loss": 3.774908257275329, + "tokens_seen": 763691008 + }, + { + "epoch": 0.23, + "learning_rate": 0.0006210792377131395, + "loss": 1.5155, + "theoretical_loss": 3.7748761203601164, + "tokens_seen": 763756544 + }, + { + "epoch": 0.23, + "learning_rate": 0.0006210631895687062, + "loss": 1.5145, + "theoretical_loss": 3.774843986974422, + "tokens_seen": 763822080 + }, + { + "epoch": 0.23, + "learning_rate": 0.0006210471414242728, + "loss": 1.4909, + "theoretical_loss": 3.774811857117554, + "tokens_seen": 763887616 + }, + { + "epoch": 0.23, + "learning_rate": 0.0006210310932798396, + "loss": 1.4643, + "theoretical_loss": 3.7747797307888225, + "tokens_seen": 763953152 + }, + { + "epoch": 0.23, + "learning_rate": 0.0006210150451354062, + "loss": 1.5285, + "theoretical_loss": 3.7747476079875373, + "tokens_seen": 764018688 + }, + { + "epoch": 0.23, + "learning_rate": 0.000620998996990973, + "loss": 1.4797, + "theoretical_loss": 3.7747154887130088, + "tokens_seen": 764084224 + }, + { + "epoch": 0.23, + "learning_rate": 0.0006209829488465396, + "loss": 1.536, + "theoretical_loss": 3.7746833729645473, + "tokens_seen": 764149760 + }, + { + "epoch": 0.23, + "learning_rate": 0.0006209669007021064, + "loss": 1.5232, + "theoretical_loss": 3.774651260741463, + "tokens_seen": 764215296 + }, + { + "epoch": 0.23, + "learning_rate": 0.000620950852557673, + "loss": 1.5945, + "theoretical_loss": 3.7746191520430674, + "tokens_seen": 764280832 + }, + { + "epoch": 0.23, + "learning_rate": 0.0006209348044132398, + "loss": 1.3295, + "theoretical_loss": 3.774587046868671, + "tokens_seen": 764346368 + }, + { + "epoch": 0.23, + "learning_rate": 0.0006209187562688064, + "loss": 1.4802, + "theoretical_loss": 3.7745549452175844, + "tokens_seen": 764411904 + }, + { + "epoch": 0.23, + "learning_rate": 0.0006209027081243732, + "loss": 1.4696, + "theoretical_loss": 3.77452284708912, + "tokens_seen": 764477440 + }, + { + "epoch": 0.23, + "learning_rate": 0.0006208866599799398, + "loss": 1.4557, + "theoretical_loss": 3.7744907524825884, + "tokens_seen": 764542976 + }, + { + "epoch": 0.23, + "learning_rate": 0.0006208706118355066, + "loss": 1.524, + "theoretical_loss": 3.774458661397302, + "tokens_seen": 764608512 + }, + { + "epoch": 0.23, + "learning_rate": 0.0006208545636910732, + "loss": 1.5477, + "theoretical_loss": 3.774426573832572, + "tokens_seen": 764674048 + }, + { + "epoch": 0.23, + "learning_rate": 0.00062083851554664, + "loss": 1.5789, + "theoretical_loss": 3.7743944897877113, + "tokens_seen": 764739584 + }, + { + "epoch": 0.23, + "learning_rate": 0.0006208224674022066, + "loss": 1.5023, + "theoretical_loss": 3.7743624092620314, + "tokens_seen": 764805120 + }, + { + "epoch": 0.23, + "learning_rate": 0.0006208064192577734, + "loss": 1.4683, + "theoretical_loss": 3.774330332254846, + "tokens_seen": 764870656 + }, + { + "epoch": 0.23, + "learning_rate": 0.00062079037111334, + "loss": 1.4335, + "theoretical_loss": 3.774298258765466, + "tokens_seen": 764936192 + }, + { + "epoch": 0.23, + "learning_rate": 0.0006207743229689068, + "loss": 1.515, + "theoretical_loss": 3.774266188793206, + "tokens_seen": 765001728 + }, + { + "epoch": 0.23, + "learning_rate": 0.0006207582748244734, + "loss": 1.554, + "theoretical_loss": 3.7742341223373774, + "tokens_seen": 765067264 + }, + { + "epoch": 0.23, + "learning_rate": 0.0006207422266800402, + "loss": 1.4779, + "theoretical_loss": 3.774202059397295, + "tokens_seen": 765132800 + }, + { + "epoch": 0.23, + "learning_rate": 0.0006207261785356068, + "loss": 1.4733, + "theoretical_loss": 3.774169999972271, + "tokens_seen": 765198336 + }, + { + "epoch": 0.23, + "learning_rate": 0.0006207101303911736, + "loss": 1.4109, + "theoretical_loss": 3.77413794406162, + "tokens_seen": 765263872 + }, + { + "epoch": 0.23, + "learning_rate": 0.0006206940822467402, + "loss": 1.4279, + "theoretical_loss": 3.7741058916646555, + "tokens_seen": 765329408 + }, + { + "epoch": 0.23, + "learning_rate": 0.000620678034102307, + "loss": 1.5131, + "theoretical_loss": 3.7740738427806915, + "tokens_seen": 765394944 + }, + { + "epoch": 0.23, + "learning_rate": 0.0006206619859578736, + "loss": 1.4888, + "theoretical_loss": 3.7740417974090423, + "tokens_seen": 765460480 + }, + { + "epoch": 0.23, + "learning_rate": 0.0006206459378134404, + "loss": 1.5455, + "theoretical_loss": 3.774009755549022, + "tokens_seen": 765526016 + }, + { + "epoch": 0.23, + "learning_rate": 0.000620629889669007, + "loss": 1.5952, + "theoretical_loss": 3.773977717199946, + "tokens_seen": 765591552 + }, + { + "epoch": 0.23, + "learning_rate": 0.0006206138415245738, + "loss": 1.5053, + "theoretical_loss": 3.773945682361128, + "tokens_seen": 765657088 + }, + { + "epoch": 0.23, + "learning_rate": 0.0006205977933801404, + "loss": 1.57, + "theoretical_loss": 3.7739136510318834, + "tokens_seen": 765722624 + }, + { + "epoch": 0.23, + "learning_rate": 0.0006205817452357072, + "loss": 1.3496, + "theoretical_loss": 3.7738816232115275, + "tokens_seen": 765788160 + }, + { + "epoch": 0.23, + "learning_rate": 0.0006205656970912738, + "loss": 1.6333, + "theoretical_loss": 3.773849598899376, + "tokens_seen": 765853696 + }, + { + "epoch": 0.23, + "learning_rate": 0.0006205496489468406, + "loss": 1.566, + "theoretical_loss": 3.773817578094744, + "tokens_seen": 765919232 + }, + { + "epoch": 0.23, + "learning_rate": 0.0006205336008024072, + "loss": 1.5787, + "theoretical_loss": 3.7737855607969477, + "tokens_seen": 765984768 + }, + { + "epoch": 0.23, + "learning_rate": 0.000620517552657974, + "loss": 1.4753, + "theoretical_loss": 3.7737535470053025, + "tokens_seen": 766050304 + }, + { + "epoch": 0.23, + "learning_rate": 0.0006205015045135406, + "loss": 1.4897, + "theoretical_loss": 3.7737215367191252, + "tokens_seen": 766115840 + }, + { + "epoch": 0.23, + "learning_rate": 0.0006204854563691074, + "loss": 1.5154, + "theoretical_loss": 3.773689529937731, + "tokens_seen": 766181376 + }, + { + "epoch": 0.23, + "learning_rate": 0.000620469408224674, + "loss": 1.4372, + "theoretical_loss": 3.773657526660438, + "tokens_seen": 766246912 + }, + { + "epoch": 0.23, + "learning_rate": 0.0006204533600802408, + "loss": 1.4569, + "theoretical_loss": 3.773625526886562, + "tokens_seen": 766312448 + }, + { + "epoch": 0.23, + "learning_rate": 0.0006204373119358074, + "loss": 1.5898, + "theoretical_loss": 3.7735935306154196, + "tokens_seen": 766377984 + }, + { + "epoch": 0.23, + "learning_rate": 0.0006204212637913742, + "loss": 1.4649, + "theoretical_loss": 3.7735615378463283, + "tokens_seen": 766443520 + }, + { + "epoch": 0.23, + "learning_rate": 0.0006204052156469408, + "loss": 1.6425, + "theoretical_loss": 3.7735295485786056, + "tokens_seen": 766509056 + }, + { + "epoch": 0.23, + "learning_rate": 0.0006203891675025076, + "loss": 1.4848, + "theoretical_loss": 3.773497562811569, + "tokens_seen": 766574592 + }, + { + "epoch": 0.23, + "learning_rate": 0.0006203731193580742, + "loss": 1.5131, + "theoretical_loss": 3.773465580544536, + "tokens_seen": 766640128 + }, + { + "epoch": 0.23, + "learning_rate": 0.000620357071213641, + "loss": 1.5651, + "theoretical_loss": 3.7734336017768237, + "tokens_seen": 766705664 + }, + { + "epoch": 0.23, + "objective/train/docs_used": 283988, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.4890892505645752, + "objective/train/theoretical_loss": 3.773401626507752, + "objective/train/tokens_used": 787231200, + "theoretical_loss": 3.773401626507752, + "tokens_seen": 766771200 + }, + { + "epoch": 0.23, + "learning_rate": 0.0006203410230692076, + "loss": 1.4472, + "theoretical_loss": 3.773401626507752, + "tokens_seen": 766771200 + }, + { + "epoch": 0.23, + "learning_rate": 0.0006203249749247744, + "loss": 1.5215, + "theoretical_loss": 3.773369654736637, + "tokens_seen": 766836736 + }, + { + "epoch": 0.23, + "learning_rate": 0.000620308926780341, + "loss": 1.4225, + "theoretical_loss": 3.773337686462799, + "tokens_seen": 766902272 + }, + { + "epoch": 0.23, + "learning_rate": 0.0006202928786359078, + "loss": 1.5257, + "theoretical_loss": 3.7733057216855554, + "tokens_seen": 766967808 + }, + { + "epoch": 0.23, + "learning_rate": 0.0006202768304914744, + "loss": 1.5404, + "theoretical_loss": 3.7732737604042255, + "tokens_seen": 767033344 + }, + { + "epoch": 0.23, + "learning_rate": 0.0006202607823470412, + "loss": 1.4475, + "theoretical_loss": 3.773241802618129, + "tokens_seen": 767098880 + }, + { + "epoch": 0.23, + "learning_rate": 0.0006202447342026078, + "loss": 1.4526, + "theoretical_loss": 3.7732098483265837, + "tokens_seen": 767164416 + }, + { + "epoch": 0.23, + "learning_rate": 0.0006202286860581746, + "loss": 1.5252, + "theoretical_loss": 3.7731778975289094, + "tokens_seen": 767229952 + }, + { + "epoch": 0.23, + "learning_rate": 0.0006202126379137412, + "loss": 1.4904, + "theoretical_loss": 3.7731459502244262, + "tokens_seen": 767295488 + }, + { + "epoch": 0.23, + "learning_rate": 0.000620196589769308, + "loss": 1.5029, + "theoretical_loss": 3.773114006412454, + "tokens_seen": 767361024 + }, + { + "epoch": 0.23, + "learning_rate": 0.0006201805416248746, + "loss": 1.4168, + "theoretical_loss": 3.773082066092312, + "tokens_seen": 767426560 + }, + { + "epoch": 0.23, + "learning_rate": 0.0006201644934804414, + "loss": 1.575, + "theoretical_loss": 3.773050129263321, + "tokens_seen": 767492096 + }, + { + "epoch": 0.23, + "learning_rate": 0.000620148445336008, + "loss": 1.5004, + "theoretical_loss": 3.7730181959248013, + "tokens_seen": 767557632 + }, + { + "epoch": 0.23, + "learning_rate": 0.0006201323971915748, + "loss": 1.4684, + "theoretical_loss": 3.7729862660760727, + "tokens_seen": 767623168 + }, + { + "epoch": 0.23, + "learning_rate": 0.0006201163490471414, + "loss": 1.4925, + "theoretical_loss": 3.7729543397164567, + "tokens_seen": 767688704 + }, + { + "epoch": 0.23, + "learning_rate": 0.0006201003009027082, + "loss": 1.4973, + "theoretical_loss": 3.772922416845274, + "tokens_seen": 767754240 + }, + { + "epoch": 0.23, + "learning_rate": 0.0006200842527582748, + "loss": 1.482, + "theoretical_loss": 3.7728904974618462, + "tokens_seen": 767819776 + }, + { + "epoch": 0.23, + "learning_rate": 0.0006200682046138416, + "loss": 1.4918, + "theoretical_loss": 3.7728585815654934, + "tokens_seen": 767885312 + }, + { + "epoch": 0.23, + "learning_rate": 0.0006200521564694082, + "loss": 1.4707, + "theoretical_loss": 3.7728266691555383, + "tokens_seen": 767950848 + }, + { + "epoch": 0.23, + "learning_rate": 0.000620036108324975, + "loss": 1.4707, + "theoretical_loss": 3.7727947602313017, + "tokens_seen": 768016384 + }, + { + "epoch": 0.23, + "learning_rate": 0.0006200200601805416, + "loss": 1.4499, + "theoretical_loss": 3.772762854792106, + "tokens_seen": 768081920 + }, + { + "epoch": 0.23, + "learning_rate": 0.0006200040120361084, + "loss": 1.4544, + "theoretical_loss": 3.772730952837273, + "tokens_seen": 768147456 + }, + { + "epoch": 0.23, + "learning_rate": 0.000619987963891675, + "loss": 1.4522, + "theoretical_loss": 3.7726990543661256, + "tokens_seen": 768212992 + }, + { + "epoch": 0.23, + "learning_rate": 0.0006199719157472418, + "loss": 1.4878, + "theoretical_loss": 3.7726671593779852, + "tokens_seen": 768278528 + }, + { + "epoch": 0.23, + "learning_rate": 0.0006199558676028084, + "loss": 1.562, + "theoretical_loss": 3.7726352678721753, + "tokens_seen": 768344064 + }, + { + "epoch": 0.23, + "learning_rate": 0.0006199398194583752, + "loss": 1.5527, + "theoretical_loss": 3.772603379848018, + "tokens_seen": 768409600 + }, + { + "epoch": 0.23, + "learning_rate": 0.0006199237713139418, + "loss": 1.4568, + "theoretical_loss": 3.772571495304837, + "tokens_seen": 768475136 + }, + { + "epoch": 0.23, + "learning_rate": 0.0006199077231695086, + "loss": 1.4653, + "theoretical_loss": 3.7725396142419547, + "tokens_seen": 768540672 + }, + { + "epoch": 0.23, + "learning_rate": 0.0006198916750250752, + "loss": 1.4522, + "theoretical_loss": 3.772507736658695, + "tokens_seen": 768606208 + }, + { + "epoch": 0.23, + "learning_rate": 0.000619875626880642, + "loss": 1.506, + "theoretical_loss": 3.7724758625543817, + "tokens_seen": 768671744 + }, + { + "epoch": 0.23, + "learning_rate": 0.0006198595787362086, + "loss": 1.5747, + "theoretical_loss": 3.7724439919283386, + "tokens_seen": 768737280 + }, + { + "epoch": 0.23, + "learning_rate": 0.0006198435305917754, + "loss": 1.6119, + "theoretical_loss": 3.7724121247798887, + "tokens_seen": 768802816 + }, + { + "epoch": 0.23, + "learning_rate": 0.000619827482447342, + "loss": 1.5315, + "theoretical_loss": 3.7723802611083572, + "tokens_seen": 768868352 + }, + { + "epoch": 0.23, + "learning_rate": 0.0006198114343029088, + "loss": 1.411, + "theoretical_loss": 3.7723484009130677, + "tokens_seen": 768933888 + }, + { + "epoch": 0.23, + "learning_rate": 0.0006197953861584754, + "loss": 1.6412, + "theoretical_loss": 3.772316544193345, + "tokens_seen": 768999424 + }, + { + "epoch": 0.23, + "learning_rate": 0.0006197793380140422, + "loss": 1.5247, + "theoretical_loss": 3.772284690948514, + "tokens_seen": 769064960 + }, + { + "epoch": 0.23, + "learning_rate": 0.0006197632898696088, + "loss": 1.4551, + "theoretical_loss": 3.7722528411778997, + "tokens_seen": 769130496 + }, + { + "epoch": 0.23, + "learning_rate": 0.0006197472417251756, + "loss": 1.5164, + "theoretical_loss": 3.7722209948808265, + "tokens_seen": 769196032 + }, + { + "epoch": 0.23, + "learning_rate": 0.0006197311935807422, + "loss": 1.5368, + "theoretical_loss": 3.77218915205662, + "tokens_seen": 769261568 + }, + { + "epoch": 0.23, + "learning_rate": 0.000619715145436309, + "loss": 1.5484, + "theoretical_loss": 3.772157312704606, + "tokens_seen": 769327104 + }, + { + "epoch": 0.23, + "learning_rate": 0.0006196990972918757, + "loss": 1.4811, + "theoretical_loss": 3.7721254768241104, + "tokens_seen": 769392640 + }, + { + "epoch": 0.23, + "learning_rate": 0.0006196830491474424, + "loss": 1.4728, + "theoretical_loss": 3.772093644414458, + "tokens_seen": 769458176 + }, + { + "epoch": 0.23, + "learning_rate": 0.000619667001003009, + "loss": 1.4254, + "theoretical_loss": 3.7720618154749754, + "tokens_seen": 769523712 + }, + { + "epoch": 0.23, + "learning_rate": 0.0006196509528585757, + "loss": 1.481, + "theoretical_loss": 3.772029990004989, + "tokens_seen": 769589248 + }, + { + "epoch": 0.23, + "learning_rate": 0.0006196349047141425, + "loss": 1.5547, + "theoretical_loss": 3.771998168003825, + "tokens_seen": 769654784 + }, + { + "epoch": 0.23, + "learning_rate": 0.0006196188565697091, + "loss": 1.394, + "theoretical_loss": 3.77196634947081, + "tokens_seen": 769720320 + }, + { + "epoch": 0.23, + "learning_rate": 0.0006196028084252759, + "loss": 1.5163, + "theoretical_loss": 3.7719345344052706, + "tokens_seen": 769785856 + }, + { + "epoch": 0.23, + "learning_rate": 0.0006195867602808425, + "loss": 1.5062, + "theoretical_loss": 3.771902722806534, + "tokens_seen": 769851392 + }, + { + "epoch": 0.23, + "learning_rate": 0.0006195707121364093, + "loss": 1.3951, + "theoretical_loss": 3.7718709146739275, + "tokens_seen": 769916928 + }, + { + "epoch": 0.23, + "learning_rate": 0.0006195546639919759, + "loss": 1.5108, + "theoretical_loss": 3.771839110006778, + "tokens_seen": 769982464 + }, + { + "epoch": 0.23, + "objective/train/docs_used": 285055, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.5235095024108887, + "objective/train/theoretical_loss": 3.7718073088044135, + "objective/train/tokens_used": 790508000, + "theoretical_loss": 3.7718073088044135, + "tokens_seen": 770048000 + }, + { + "epoch": 0.23, + "learning_rate": 0.0006195386158475427, + "loss": 1.5167, + "theoretical_loss": 3.7718073088044135, + "tokens_seen": 770048000 + }, + { + "epoch": 0.23, + "learning_rate": 0.0006195225677031093, + "loss": 1.5032, + "theoretical_loss": 3.7717755110661617, + "tokens_seen": 770113536 + }, + { + "epoch": 0.23, + "learning_rate": 0.0006195065195586761, + "loss": 1.4491, + "theoretical_loss": 3.7717437167913497, + "tokens_seen": 770179072 + }, + { + "epoch": 0.23, + "learning_rate": 0.0006194904714142427, + "loss": 1.5029, + "theoretical_loss": 3.771711925979307, + "tokens_seen": 770244608 + }, + { + "epoch": 0.23, + "learning_rate": 0.0006194744232698095, + "loss": 1.5416, + "theoretical_loss": 3.7716801386293604, + "tokens_seen": 770310144 + }, + { + "epoch": 0.23, + "learning_rate": 0.0006194583751253761, + "loss": 1.3911, + "theoretical_loss": 3.7716483547408393, + "tokens_seen": 770375680 + }, + { + "epoch": 0.23, + "learning_rate": 0.0006194423269809429, + "loss": 1.4864, + "theoretical_loss": 3.7716165743130725, + "tokens_seen": 770441216 + }, + { + "epoch": 0.23, + "learning_rate": 0.0006194262788365095, + "loss": 1.3683, + "theoretical_loss": 3.7715847973453878, + "tokens_seen": 770506752 + }, + { + "epoch": 0.23, + "learning_rate": 0.0006194102306920763, + "loss": 1.4575, + "theoretical_loss": 3.7715530238371153, + "tokens_seen": 770572288 + }, + { + "epoch": 0.23, + "learning_rate": 0.0006193941825476429, + "loss": 1.4476, + "theoretical_loss": 3.7715212537875837, + "tokens_seen": 770637824 + }, + { + "epoch": 0.23, + "learning_rate": 0.0006193781344032097, + "loss": 1.4568, + "theoretical_loss": 3.7714894871961233, + "tokens_seen": 770703360 + }, + { + "epoch": 0.23, + "learning_rate": 0.0006193620862587763, + "loss": 1.4837, + "theoretical_loss": 3.7714577240620617, + "tokens_seen": 770768896 + }, + { + "epoch": 0.23, + "learning_rate": 0.0006193460381143431, + "loss": 1.4577, + "theoretical_loss": 3.771425964384731, + "tokens_seen": 770834432 + }, + { + "epoch": 0.23, + "learning_rate": 0.0006193299899699097, + "loss": 1.5225, + "theoretical_loss": 3.7713942081634593, + "tokens_seen": 770899968 + }, + { + "epoch": 0.23, + "learning_rate": 0.0006193139418254765, + "loss": 1.452, + "theoretical_loss": 3.7713624553975778, + "tokens_seen": 770965504 + }, + { + "epoch": 0.23, + "learning_rate": 0.0006192978936810431, + "loss": 1.5251, + "theoretical_loss": 3.7713307060864167, + "tokens_seen": 771031040 + }, + { + "epoch": 0.23, + "learning_rate": 0.0006192818455366099, + "loss": 1.5303, + "theoretical_loss": 3.771298960229306, + "tokens_seen": 771096576 + }, + { + "epoch": 0.23, + "learning_rate": 0.0006192657973921765, + "loss": 1.5315, + "theoretical_loss": 3.7712672178255775, + "tokens_seen": 771162112 + }, + { + "epoch": 0.23, + "learning_rate": 0.0006192497492477433, + "loss": 1.6051, + "theoretical_loss": 3.7712354788745612, + "tokens_seen": 771227648 + }, + { + "epoch": 0.23, + "learning_rate": 0.0006192337011033099, + "loss": 1.3754, + "theoretical_loss": 3.771203743375588, + "tokens_seen": 771293184 + }, + { + "epoch": 0.23, + "learning_rate": 0.0006192176529588767, + "loss": 1.4912, + "theoretical_loss": 3.77117201132799, + "tokens_seen": 771358720 + }, + { + "epoch": 0.23, + "learning_rate": 0.0006192016048144433, + "loss": 1.4925, + "theoretical_loss": 3.771140282731098, + "tokens_seen": 771424256 + }, + { + "epoch": 0.23, + "learning_rate": 0.0006191855566700101, + "loss": 1.4816, + "theoretical_loss": 3.7711085575842436, + "tokens_seen": 771489792 + }, + { + "epoch": 0.23, + "learning_rate": 0.0006191695085255767, + "loss": 1.4708, + "theoretical_loss": 3.7710768358867597, + "tokens_seen": 771555328 + }, + { + "epoch": 0.23, + "learning_rate": 0.0006191534603811435, + "loss": 1.5368, + "theoretical_loss": 3.771045117637977, + "tokens_seen": 771620864 + }, + { + "epoch": 0.23, + "learning_rate": 0.0006191374122367101, + "loss": 1.4799, + "theoretical_loss": 3.7710134028372284, + "tokens_seen": 771686400 + }, + { + "epoch": 0.23, + "learning_rate": 0.0006191213640922769, + "loss": 1.6245, + "theoretical_loss": 3.770981691483846, + "tokens_seen": 771751936 + }, + { + "epoch": 0.23, + "learning_rate": 0.0006191053159478435, + "loss": 1.4853, + "theoretical_loss": 3.7709499835771627, + "tokens_seen": 771817472 + }, + { + "epoch": 0.23, + "learning_rate": 0.0006190892678034103, + "loss": 1.4416, + "theoretical_loss": 3.7709182791165112, + "tokens_seen": 771883008 + }, + { + "epoch": 0.23, + "learning_rate": 0.0006190732196589769, + "loss": 1.4716, + "theoretical_loss": 3.770886578101224, + "tokens_seen": 771948544 + }, + { + "epoch": 0.23, + "learning_rate": 0.0006190571715145437, + "loss": 1.515, + "theoretical_loss": 3.770854880530635, + "tokens_seen": 772014080 + }, + { + "epoch": 0.23, + "learning_rate": 0.0006190411233701103, + "loss": 1.461, + "theoretical_loss": 3.770823186404077, + "tokens_seen": 772079616 + }, + { + "epoch": 0.23, + "learning_rate": 0.0006190250752256771, + "loss": 1.5056, + "theoretical_loss": 3.770791495720883, + "tokens_seen": 772145152 + }, + { + "epoch": 0.23, + "learning_rate": 0.0006190090270812437, + "loss": 1.4574, + "theoretical_loss": 3.770759808480388, + "tokens_seen": 772210688 + }, + { + "epoch": 0.23, + "learning_rate": 0.0006189929789368105, + "loss": 1.3339, + "theoretical_loss": 3.770728124681925, + "tokens_seen": 772276224 + }, + { + "epoch": 0.23, + "learning_rate": 0.0006189769307923771, + "loss": 1.6151, + "theoretical_loss": 3.7706964443248276, + "tokens_seen": 772341760 + }, + { + "epoch": 0.23, + "learning_rate": 0.0006189608826479439, + "loss": 1.5108, + "theoretical_loss": 3.7706647674084315, + "tokens_seen": 772407296 + }, + { + "epoch": 0.23, + "learning_rate": 0.0006189448345035105, + "loss": 1.3958, + "theoretical_loss": 3.77063309393207, + "tokens_seen": 772472832 + }, + { + "epoch": 0.23, + "learning_rate": 0.0006189287863590773, + "loss": 1.5153, + "theoretical_loss": 3.7706014238950782, + "tokens_seen": 772538368 + }, + { + "epoch": 0.23, + "learning_rate": 0.0006189127382146439, + "loss": 1.3771, + "theoretical_loss": 3.7705697572967902, + "tokens_seen": 772603904 + }, + { + "epoch": 0.23, + "learning_rate": 0.0006188966900702107, + "loss": 1.5387, + "theoretical_loss": 3.7705380941365423, + "tokens_seen": 772669440 + }, + { + "epoch": 0.23, + "learning_rate": 0.0006188806419257773, + "loss": 1.3855, + "theoretical_loss": 3.770506434413668, + "tokens_seen": 772734976 + }, + { + "epoch": 0.23, + "learning_rate": 0.0006188645937813441, + "loss": 1.5604, + "theoretical_loss": 3.770474778127504, + "tokens_seen": 772800512 + }, + { + "epoch": 0.23, + "learning_rate": 0.0006188485456369107, + "loss": 1.5024, + "theoretical_loss": 3.7704431252773856, + "tokens_seen": 772866048 + }, + { + "epoch": 0.23, + "learning_rate": 0.0006188324974924775, + "loss": 1.5986, + "theoretical_loss": 3.770411475862648, + "tokens_seen": 772931584 + }, + { + "epoch": 0.23, + "learning_rate": 0.0006188164493480441, + "loss": 1.498, + "theoretical_loss": 3.7703798298826277, + "tokens_seen": 772997120 + }, + { + "epoch": 0.23, + "learning_rate": 0.0006188004012036109, + "loss": 1.6772, + "theoretical_loss": 3.77034818733666, + "tokens_seen": 773062656 + }, + { + "epoch": 0.23, + "learning_rate": 0.0006187843530591775, + "loss": 1.5632, + "theoretical_loss": 3.7703165482240824, + "tokens_seen": 773128192 + }, + { + "epoch": 0.23, + "learning_rate": 0.0006187683049147443, + "loss": 1.4723, + "theoretical_loss": 3.77028491254423, + "tokens_seen": 773193728 + }, + { + "epoch": 0.23, + "learning_rate": 0.0006187522567703109, + "loss": 1.3876, + "theoretical_loss": 3.7702532802964406, + "tokens_seen": 773259264 + }, + { + "epoch": 0.23, + "objective/train/docs_used": 286255, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.3970820903778076, + "objective/train/theoretical_loss": 3.7702216514800506, + "objective/train/tokens_used": 793784800, + "theoretical_loss": 3.7702216514800506, + "tokens_seen": 773324800 + }, + { + "epoch": 0.23, + "learning_rate": 0.0006187362086258777, + "loss": 1.492, + "theoretical_loss": 3.7702216514800506, + "tokens_seen": 773324800 + }, + { + "epoch": 0.23, + "learning_rate": 0.0006187201604814443, + "loss": 1.4865, + "theoretical_loss": 3.770190026094397, + "tokens_seen": 773390336 + }, + { + "epoch": 0.23, + "learning_rate": 0.0006187041123370111, + "loss": 1.4317, + "theoretical_loss": 3.7701584041388165, + "tokens_seen": 773455872 + }, + { + "epoch": 0.23, + "learning_rate": 0.0006186880641925777, + "loss": 1.5308, + "theoretical_loss": 3.7701267856126472, + "tokens_seen": 773521408 + }, + { + "epoch": 0.23, + "learning_rate": 0.0006186720160481445, + "loss": 1.4457, + "theoretical_loss": 3.7700951705152264, + "tokens_seen": 773586944 + }, + { + "epoch": 0.23, + "learning_rate": 0.0006186559679037111, + "loss": 1.5644, + "theoretical_loss": 3.7700635588458926, + "tokens_seen": 773652480 + }, + { + "epoch": 0.23, + "learning_rate": 0.0006186399197592779, + "loss": 1.5216, + "theoretical_loss": 3.7700319506039826, + "tokens_seen": 773718016 + }, + { + "epoch": 0.23, + "learning_rate": 0.0006186238716148445, + "loss": 1.56, + "theoretical_loss": 3.7700003457888345, + "tokens_seen": 773783552 + }, + { + "epoch": 0.23, + "learning_rate": 0.0006186078234704113, + "loss": 1.5235, + "theoretical_loss": 3.7699687443997876, + "tokens_seen": 773849088 + }, + { + "epoch": 0.23, + "learning_rate": 0.000618591775325978, + "loss": 1.5489, + "theoretical_loss": 3.7699371464361797, + "tokens_seen": 773914624 + }, + { + "epoch": 0.23, + "learning_rate": 0.0006185757271815447, + "loss": 1.5091, + "theoretical_loss": 3.7699055518973497, + "tokens_seen": 773980160 + }, + { + "epoch": 0.23, + "learning_rate": 0.0006185596790371114, + "loss": 1.456, + "theoretical_loss": 3.769873960782636, + "tokens_seen": 774045696 + }, + { + "epoch": 0.23, + "learning_rate": 0.0006185436308926781, + "loss": 1.4392, + "theoretical_loss": 3.769842373091379, + "tokens_seen": 774111232 + }, + { + "epoch": 0.23, + "learning_rate": 0.0006185275827482449, + "loss": 1.638, + "theoretical_loss": 3.769810788822916, + "tokens_seen": 774176768 + }, + { + "epoch": 0.23, + "learning_rate": 0.0006185115346038115, + "loss": 1.4776, + "theoretical_loss": 3.769779207976588, + "tokens_seen": 774242304 + }, + { + "epoch": 0.23, + "learning_rate": 0.0006184954864593783, + "loss": 1.6123, + "theoretical_loss": 3.7697476305517337, + "tokens_seen": 774307840 + }, + { + "epoch": 0.23, + "learning_rate": 0.0006184794383149449, + "loss": 1.4747, + "theoretical_loss": 3.7697160565476935, + "tokens_seen": 774373376 + }, + { + "epoch": 0.23, + "learning_rate": 0.0006184633901705117, + "loss": 1.4321, + "theoretical_loss": 3.769684485963807, + "tokens_seen": 774438912 + }, + { + "epoch": 0.23, + "learning_rate": 0.0006184473420260783, + "loss": 1.5738, + "theoretical_loss": 3.7696529187994137, + "tokens_seen": 774504448 + }, + { + "epoch": 0.23, + "learning_rate": 0.0006184312938816451, + "loss": 1.5302, + "theoretical_loss": 3.769621355053855, + "tokens_seen": 774569984 + }, + { + "epoch": 0.23, + "learning_rate": 0.0006184152457372117, + "loss": 1.5904, + "theoretical_loss": 3.769589794726471, + "tokens_seen": 774635520 + }, + { + "epoch": 0.23, + "learning_rate": 0.0006183991975927783, + "loss": 1.485, + "theoretical_loss": 3.7695582378166024, + "tokens_seen": 774701056 + }, + { + "epoch": 0.23, + "learning_rate": 0.0006183831494483451, + "loss": 1.5945, + "theoretical_loss": 3.7695266843235897, + "tokens_seen": 774766592 + }, + { + "epoch": 0.23, + "learning_rate": 0.0006183671013039117, + "loss": 1.537, + "theoretical_loss": 3.7694951342467746, + "tokens_seen": 774832128 + }, + { + "epoch": 0.23, + "learning_rate": 0.0006183510531594785, + "loss": 1.5191, + "theoretical_loss": 3.7694635875854985, + "tokens_seen": 774897664 + }, + { + "epoch": 0.23, + "learning_rate": 0.0006183350050150452, + "loss": 1.5258, + "theoretical_loss": 3.7694320443391023, + "tokens_seen": 774963200 + }, + { + "epoch": 0.23, + "learning_rate": 0.0006183189568706119, + "loss": 1.445, + "theoretical_loss": 3.769400504506927, + "tokens_seen": 775028736 + }, + { + "epoch": 0.23, + "learning_rate": 0.0006183029087261786, + "loss": 1.5931, + "theoretical_loss": 3.7693689680883153, + "tokens_seen": 775094272 + }, + { + "epoch": 0.23, + "learning_rate": 0.0006182868605817453, + "loss": 1.4024, + "theoretical_loss": 3.769337435082609, + "tokens_seen": 775159808 + }, + { + "epoch": 0.23, + "learning_rate": 0.000618270812437312, + "loss": 1.5208, + "theoretical_loss": 3.769305905489151, + "tokens_seen": 775225344 + }, + { + "epoch": 0.23, + "learning_rate": 0.0006182547642928786, + "loss": 1.654, + "theoretical_loss": 3.769274379307282, + "tokens_seen": 775290880 + }, + { + "epoch": 0.23, + "learning_rate": 0.0006182387161484454, + "loss": 1.4575, + "theoretical_loss": 3.7692428565363456, + "tokens_seen": 775356416 + }, + { + "epoch": 0.23, + "learning_rate": 0.000618222668004012, + "loss": 1.5397, + "theoretical_loss": 3.769211337175684, + "tokens_seen": 775421952 + }, + { + "epoch": 0.23, + "learning_rate": 0.0006182066198595788, + "loss": 1.4732, + "theoretical_loss": 3.7691798212246406, + "tokens_seen": 775487488 + }, + { + "epoch": 0.24, + "learning_rate": 0.0006181905717151454, + "loss": 1.4487, + "theoretical_loss": 3.769148308682558, + "tokens_seen": 775553024 + }, + { + "epoch": 0.24, + "learning_rate": 0.0006181745235707122, + "loss": 1.5678, + "theoretical_loss": 3.76911679954878, + "tokens_seen": 775618560 + }, + { + "epoch": 0.24, + "learning_rate": 0.0006181584754262788, + "loss": 1.4856, + "theoretical_loss": 3.7690852938226493, + "tokens_seen": 775684096 + }, + { + "epoch": 0.24, + "learning_rate": 0.0006181424272818456, + "loss": 1.4944, + "theoretical_loss": 3.7690537915035103, + "tokens_seen": 775749632 + }, + { + "epoch": 0.24, + "learning_rate": 0.0006181263791374122, + "loss": 1.6734, + "theoretical_loss": 3.769022292590706, + "tokens_seen": 775815168 + }, + { + "epoch": 0.24, + "learning_rate": 0.000618110330992979, + "loss": 1.5568, + "theoretical_loss": 3.768990797083581, + "tokens_seen": 775880704 + }, + { + "epoch": 0.24, + "learning_rate": 0.0006180942828485456, + "loss": 1.4404, + "theoretical_loss": 3.7689593049814794, + "tokens_seen": 775946240 + }, + { + "epoch": 0.24, + "learning_rate": 0.0006180782347041124, + "loss": 1.4128, + "theoretical_loss": 3.768927816283745, + "tokens_seen": 776011776 + }, + { + "epoch": 0.24, + "learning_rate": 0.000618062186559679, + "loss": 1.5303, + "theoretical_loss": 3.768896330989723, + "tokens_seen": 776077312 + }, + { + "epoch": 0.24, + "learning_rate": 0.0006180461384152458, + "loss": 1.5916, + "theoretical_loss": 3.7688648490987577, + "tokens_seen": 776142848 + }, + { + "epoch": 0.24, + "learning_rate": 0.0006180300902708124, + "loss": 1.5243, + "theoretical_loss": 3.768833370610194, + "tokens_seen": 776208384 + }, + { + "epoch": 0.24, + "learning_rate": 0.0006180140421263792, + "loss": 1.4192, + "theoretical_loss": 3.768801895523377, + "tokens_seen": 776273920 + }, + { + "epoch": 0.24, + "learning_rate": 0.0006179979939819458, + "loss": 1.4645, + "theoretical_loss": 3.768770423837652, + "tokens_seen": 776339456 + }, + { + "epoch": 0.24, + "learning_rate": 0.0006179819458375126, + "loss": 1.6135, + "theoretical_loss": 3.7687389555523643, + "tokens_seen": 776404992 + }, + { + "epoch": 0.24, + "learning_rate": 0.0006179658976930792, + "loss": 1.4693, + "theoretical_loss": 3.7687074906668596, + "tokens_seen": 776470528 + }, + { + "epoch": 0.24, + "learning_rate": 0.000617949849548646, + "loss": 1.492, + "theoretical_loss": 3.768676029180484, + "tokens_seen": 776536064 + }, + { + "epoch": 0.24, + "objective/train/docs_used": 287303, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.5270603895187378, + "objective/train/theoretical_loss": 3.768644571092583, + "objective/train/tokens_used": 797061600, + "theoretical_loss": 3.768644571092583, + "tokens_seen": 776601600 + }, + { + "epoch": 0.24, + "learning_rate": 0.0006179338014042126, + "loss": 1.5374, + "theoretical_loss": 3.768644571092583, + "tokens_seen": 776601600 + }, + { + "epoch": 0.24, + "learning_rate": 0.0006179177532597794, + "loss": 1.7203, + "theoretical_loss": 3.768613116402503, + "tokens_seen": 776667136 + }, + { + "epoch": 0.24, + "learning_rate": 0.000617901705115346, + "loss": 1.4325, + "theoretical_loss": 3.76858166510959, + "tokens_seen": 776732672 + }, + { + "epoch": 0.24, + "learning_rate": 0.0006178856569709128, + "loss": 1.5414, + "theoretical_loss": 3.768550217213191, + "tokens_seen": 776798208 + }, + { + "epoch": 0.24, + "learning_rate": 0.0006178696088264794, + "loss": 1.5648, + "theoretical_loss": 3.7685187727126523, + "tokens_seen": 776863744 + }, + { + "epoch": 0.24, + "learning_rate": 0.0006178535606820462, + "loss": 1.4334, + "theoretical_loss": 3.768487331607321, + "tokens_seen": 776929280 + }, + { + "epoch": 0.24, + "learning_rate": 0.0006178375125376128, + "loss": 1.4573, + "theoretical_loss": 3.7684558938965442, + "tokens_seen": 776994816 + }, + { + "epoch": 0.24, + "learning_rate": 0.0006178214643931796, + "loss": 1.4722, + "theoretical_loss": 3.768424459579669, + "tokens_seen": 777060352 + }, + { + "epoch": 0.24, + "learning_rate": 0.0006178054162487462, + "loss": 1.5928, + "theoretical_loss": 3.7683930286560425, + "tokens_seen": 777125888 + }, + { + "epoch": 0.24, + "learning_rate": 0.000617789368104313, + "loss": 1.5174, + "theoretical_loss": 3.7683616011250134, + "tokens_seen": 777191424 + }, + { + "epoch": 0.24, + "learning_rate": 0.0006177733199598796, + "loss": 1.4136, + "theoretical_loss": 3.768330176985928, + "tokens_seen": 777256960 + }, + { + "epoch": 0.24, + "learning_rate": 0.0006177572718154464, + "loss": 1.5144, + "theoretical_loss": 3.7682987562381354, + "tokens_seen": 777322496 + }, + { + "epoch": 0.24, + "learning_rate": 0.000617741223671013, + "loss": 1.4879, + "theoretical_loss": 3.768267338880983, + "tokens_seen": 777388032 + }, + { + "epoch": 0.24, + "learning_rate": 0.0006177251755265798, + "loss": 1.5288, + "theoretical_loss": 3.7682359249138195, + "tokens_seen": 777453568 + }, + { + "epoch": 0.24, + "learning_rate": 0.0006177091273821464, + "loss": 1.4331, + "theoretical_loss": 3.7682045143359937, + "tokens_seen": 777519104 + }, + { + "epoch": 0.24, + "learning_rate": 0.0006176930792377132, + "loss": 1.5688, + "theoretical_loss": 3.7681731071468536, + "tokens_seen": 777584640 + }, + { + "epoch": 0.24, + "learning_rate": 0.0006176770310932798, + "loss": 1.5619, + "theoretical_loss": 3.7681417033457487, + "tokens_seen": 777650176 + }, + { + "epoch": 0.24, + "learning_rate": 0.0006176609829488466, + "loss": 1.5437, + "theoretical_loss": 3.768110302932027, + "tokens_seen": 777715712 + }, + { + "epoch": 0.24, + "learning_rate": 0.0006176449348044132, + "loss": 1.5017, + "theoretical_loss": 3.768078905905039, + "tokens_seen": 777781248 + }, + { + "epoch": 0.24, + "learning_rate": 0.00061762888665998, + "loss": 1.4758, + "theoretical_loss": 3.7680475122641335, + "tokens_seen": 777846784 + }, + { + "epoch": 0.24, + "learning_rate": 0.0006176128385155466, + "loss": 1.5661, + "theoretical_loss": 3.7680161220086603, + "tokens_seen": 777912320 + }, + { + "epoch": 0.24, + "learning_rate": 0.0006175967903711134, + "loss": 1.4489, + "theoretical_loss": 3.7679847351379685, + "tokens_seen": 777977856 + }, + { + "epoch": 0.24, + "learning_rate": 0.00061758074222668, + "loss": 1.4157, + "theoretical_loss": 3.7679533516514088, + "tokens_seen": 778043392 + }, + { + "epoch": 0.24, + "learning_rate": 0.0006175646940822468, + "loss": 1.569, + "theoretical_loss": 3.767921971548331, + "tokens_seen": 778108928 + }, + { + "epoch": 0.24, + "learning_rate": 0.0006175486459378134, + "loss": 1.4797, + "theoretical_loss": 3.767890594828085, + "tokens_seen": 778174464 + }, + { + "epoch": 0.24, + "learning_rate": 0.0006175325977933802, + "loss": 1.3668, + "theoretical_loss": 3.7678592214900224, + "tokens_seen": 778240000 + }, + { + "epoch": 0.24, + "learning_rate": 0.0006175165496489468, + "loss": 1.4054, + "theoretical_loss": 3.7678278515334926, + "tokens_seen": 778305536 + }, + { + "epoch": 0.24, + "learning_rate": 0.0006175005015045136, + "loss": 1.3845, + "theoretical_loss": 3.7677964849578474, + "tokens_seen": 778371072 + }, + { + "epoch": 0.24, + "learning_rate": 0.0006174844533600802, + "loss": 1.338, + "theoretical_loss": 3.7677651217624373, + "tokens_seen": 778436608 + }, + { + "epoch": 0.24, + "learning_rate": 0.000617468405215647, + "loss": 1.4587, + "theoretical_loss": 3.767733761946613, + "tokens_seen": 778502144 + }, + { + "epoch": 0.24, + "learning_rate": 0.0006174523570712136, + "loss": 1.4959, + "theoretical_loss": 3.767702405509727, + "tokens_seen": 778567680 + }, + { + "epoch": 0.24, + "learning_rate": 0.0006174363089267804, + "loss": 1.4557, + "theoretical_loss": 3.7676710524511305, + "tokens_seen": 778633216 + }, + { + "epoch": 0.24, + "learning_rate": 0.000617420260782347, + "loss": 1.4645, + "theoretical_loss": 3.767639702770175, + "tokens_seen": 778698752 + }, + { + "epoch": 0.24, + "learning_rate": 0.0006174042126379138, + "loss": 1.3261, + "theoretical_loss": 3.767608356466212, + "tokens_seen": 778764288 + }, + { + "epoch": 0.24, + "learning_rate": 0.0006173881644934805, + "loss": 1.3881, + "theoretical_loss": 3.7675770135385944, + "tokens_seen": 778829824 + }, + { + "epoch": 0.24, + "learning_rate": 0.0006173721163490472, + "loss": 1.5285, + "theoretical_loss": 3.7675456739866746, + "tokens_seen": 778895360 + }, + { + "epoch": 0.24, + "learning_rate": 0.0006173560682046139, + "loss": 1.5334, + "theoretical_loss": 3.7675143378098035, + "tokens_seen": 778960896 + }, + { + "epoch": 0.24, + "learning_rate": 0.0006173400200601806, + "loss": 1.4533, + "theoretical_loss": 3.7674830050073354, + "tokens_seen": 779026432 + }, + { + "epoch": 0.24, + "learning_rate": 0.0006173239719157473, + "loss": 1.4847, + "theoretical_loss": 3.7674516755786227, + "tokens_seen": 779091968 + }, + { + "epoch": 0.24, + "learning_rate": 0.000617307923771314, + "loss": 1.4475, + "theoretical_loss": 3.767420349523018, + "tokens_seen": 779157504 + }, + { + "epoch": 0.24, + "learning_rate": 0.0006172918756268807, + "loss": 1.5149, + "theoretical_loss": 3.7673890268398744, + "tokens_seen": 779223040 + }, + { + "epoch": 0.24, + "learning_rate": 0.0006172758274824474, + "loss": 1.5691, + "theoretical_loss": 3.7673577075285456, + "tokens_seen": 779288576 + }, + { + "epoch": 0.24, + "learning_rate": 0.0006172597793380141, + "loss": 1.409, + "theoretical_loss": 3.767326391588385, + "tokens_seen": 779354112 + }, + { + "epoch": 0.24, + "learning_rate": 0.0006172437311935808, + "loss": 1.4976, + "theoretical_loss": 3.767295079018746, + "tokens_seen": 779419648 + }, + { + "epoch": 0.24, + "learning_rate": 0.0006172276830491475, + "loss": 1.474, + "theoretical_loss": 3.7672637698189835, + "tokens_seen": 779485184 + }, + { + "epoch": 0.24, + "learning_rate": 0.0006172116349047142, + "loss": 1.4917, + "theoretical_loss": 3.7672324639884502, + "tokens_seen": 779550720 + }, + { + "epoch": 0.24, + "learning_rate": 0.000617195586760281, + "loss": 1.5616, + "theoretical_loss": 3.7672011615265015, + "tokens_seen": 779616256 + }, + { + "epoch": 0.24, + "learning_rate": 0.0006171795386158476, + "loss": 1.5385, + "theoretical_loss": 3.7671698624324907, + "tokens_seen": 779681792 + }, + { + "epoch": 0.24, + "learning_rate": 0.0006171634904714144, + "loss": 1.5067, + "theoretical_loss": 3.767138566705773, + "tokens_seen": 779747328 + }, + { + "epoch": 0.24, + "learning_rate": 0.000617147442326981, + "loss": 1.5429, + "theoretical_loss": 3.7671072743457032, + "tokens_seen": 779812864 + }, + { + "epoch": 0.24, + "objective/train/docs_used": 288522, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.4652576446533203, + "objective/train/theoretical_loss": 3.767075985351636, + "objective/train/tokens_used": 800338400, + "theoretical_loss": 3.767075985351636, + "tokens_seen": 779878400 + }, + { + "epoch": 0.24, + "learning_rate": 0.0006171313941825478, + "loss": 1.5679, + "theoretical_loss": 3.767075985351636, + "tokens_seen": 779878400 + }, + { + "epoch": 0.24, + "learning_rate": 0.0006171153460381144, + "loss": 1.6335, + "theoretical_loss": 3.767044699722927, + "tokens_seen": 779943936 + }, + { + "epoch": 0.24, + "learning_rate": 0.0006170992978936812, + "loss": 1.5577, + "theoretical_loss": 3.767013417458931, + "tokens_seen": 780009472 + }, + { + "epoch": 0.24, + "learning_rate": 0.0006170832497492478, + "loss": 1.5266, + "theoretical_loss": 3.766982138559004, + "tokens_seen": 780075008 + }, + { + "epoch": 0.24, + "learning_rate": 0.0006170672016048146, + "loss": 1.4663, + "theoretical_loss": 3.7669508630225006, + "tokens_seen": 780140544 + }, + { + "epoch": 0.24, + "learning_rate": 0.0006170511534603812, + "loss": 1.493, + "theoretical_loss": 3.7669195908487776, + "tokens_seen": 780206080 + }, + { + "epoch": 0.24, + "learning_rate": 0.000617035105315948, + "loss": 1.4089, + "theoretical_loss": 3.766888322037191, + "tokens_seen": 780271616 + }, + { + "epoch": 0.24, + "learning_rate": 0.0006170190571715146, + "loss": 1.5316, + "theoretical_loss": 3.7668570565870967, + "tokens_seen": 780337152 + }, + { + "epoch": 0.24, + "learning_rate": 0.0006170030090270812, + "loss": 1.6444, + "theoretical_loss": 3.7668257944978505, + "tokens_seen": 780402688 + }, + { + "epoch": 0.24, + "learning_rate": 0.000616986960882648, + "loss": 1.5268, + "theoretical_loss": 3.7667945357688097, + "tokens_seen": 780468224 + }, + { + "epoch": 0.24, + "learning_rate": 0.0006169709127382146, + "loss": 1.5162, + "theoretical_loss": 3.7667632803993314, + "tokens_seen": 780533760 + }, + { + "epoch": 0.24, + "learning_rate": 0.0006169548645937814, + "loss": 1.5057, + "theoretical_loss": 3.7667320283887715, + "tokens_seen": 780599296 + }, + { + "epoch": 0.24, + "learning_rate": 0.000616938816449348, + "loss": 1.5809, + "theoretical_loss": 3.7667007797364875, + "tokens_seen": 780664832 + }, + { + "epoch": 0.24, + "learning_rate": 0.0006169227683049148, + "loss": 1.4466, + "theoretical_loss": 3.766669534441837, + "tokens_seen": 780730368 + }, + { + "epoch": 0.24, + "learning_rate": 0.0006169067201604815, + "loss": 1.4905, + "theoretical_loss": 3.766638292504177, + "tokens_seen": 780795904 + }, + { + "epoch": 0.24, + "learning_rate": 0.0006168906720160482, + "loss": 1.5678, + "theoretical_loss": 3.766607053922865, + "tokens_seen": 780861440 + }, + { + "epoch": 0.24, + "learning_rate": 0.0006168746238716149, + "loss": 1.4587, + "theoretical_loss": 3.7665758186972584, + "tokens_seen": 780926976 + }, + { + "epoch": 0.24, + "learning_rate": 0.0006168585757271815, + "loss": 1.4968, + "theoretical_loss": 3.7665445868267162, + "tokens_seen": 780992512 + }, + { + "epoch": 0.24, + "learning_rate": 0.0006168425275827483, + "loss": 1.4437, + "theoretical_loss": 3.766513358310596, + "tokens_seen": 781058048 + }, + { + "epoch": 0.24, + "learning_rate": 0.0006168264794383149, + "loss": 1.5512, + "theoretical_loss": 3.7664821331482568, + "tokens_seen": 781123584 + }, + { + "epoch": 0.24, + "learning_rate": 0.0006168104312938817, + "loss": 1.4463, + "theoretical_loss": 3.7664509113390556, + "tokens_seen": 781189120 + }, + { + "epoch": 0.24, + "learning_rate": 0.0006167943831494483, + "loss": 1.4369, + "theoretical_loss": 3.7664196928823523, + "tokens_seen": 781254656 + }, + { + "epoch": 0.24, + "learning_rate": 0.0006167783350050151, + "loss": 1.4617, + "theoretical_loss": 3.7663884777775047, + "tokens_seen": 781320192 + }, + { + "epoch": 0.24, + "learning_rate": 0.0006167622868605817, + "loss": 1.539, + "theoretical_loss": 3.7663572660238733, + "tokens_seen": 781385728 + }, + { + "epoch": 0.24, + "learning_rate": 0.0006167462387161485, + "loss": 1.4818, + "theoretical_loss": 3.766326057620816, + "tokens_seen": 781451264 + }, + { + "epoch": 0.24, + "learning_rate": 0.0006167301905717151, + "loss": 1.4944, + "theoretical_loss": 3.7662948525676927, + "tokens_seen": 781516800 + }, + { + "epoch": 0.24, + "learning_rate": 0.0006167141424272819, + "loss": 1.4151, + "theoretical_loss": 3.7662636508638627, + "tokens_seen": 781582336 + }, + { + "epoch": 0.24, + "learning_rate": 0.0006166980942828485, + "loss": 1.4711, + "theoretical_loss": 3.7662324525086865, + "tokens_seen": 781647872 + }, + { + "epoch": 0.24, + "learning_rate": 0.0006166820461384153, + "loss": 1.6644, + "theoretical_loss": 3.766201257501523, + "tokens_seen": 781713408 + }, + { + "epoch": 0.24, + "learning_rate": 0.0006166659979939819, + "loss": 1.4739, + "theoretical_loss": 3.7661700658417328, + "tokens_seen": 781778944 + }, + { + "epoch": 0.24, + "learning_rate": 0.0006166499498495487, + "loss": 1.3806, + "theoretical_loss": 3.7661388775286753, + "tokens_seen": 781844480 + }, + { + "epoch": 0.24, + "learning_rate": 0.0006166339017051153, + "loss": 1.5181, + "theoretical_loss": 3.7661076925617123, + "tokens_seen": 781910016 + }, + { + "epoch": 0.24, + "learning_rate": 0.0006166178535606821, + "loss": 1.4152, + "theoretical_loss": 3.7660765109402035, + "tokens_seen": 781975552 + }, + { + "epoch": 0.24, + "learning_rate": 0.0006166018054162487, + "loss": 1.5049, + "theoretical_loss": 3.76604533266351, + "tokens_seen": 782041088 + }, + { + "epoch": 0.24, + "learning_rate": 0.0006165857572718155, + "loss": 1.4712, + "theoretical_loss": 3.7660141577309925, + "tokens_seen": 782106624 + }, + { + "epoch": 0.24, + "learning_rate": 0.0006165697091273821, + "loss": 1.498, + "theoretical_loss": 3.7659829861420127, + "tokens_seen": 782172160 + }, + { + "epoch": 0.24, + "learning_rate": 0.0006165536609829489, + "loss": 1.5945, + "theoretical_loss": 3.765951817895931, + "tokens_seen": 782237696 + }, + { + "epoch": 0.24, + "learning_rate": 0.0006165376128385155, + "loss": 1.4743, + "theoretical_loss": 3.765920652992109, + "tokens_seen": 782303232 + }, + { + "epoch": 0.24, + "learning_rate": 0.0006165215646940823, + "loss": 1.4872, + "theoretical_loss": 3.7658894914299097, + "tokens_seen": 782368768 + }, + { + "epoch": 0.24, + "learning_rate": 0.0006165055165496489, + "loss": 1.3867, + "theoretical_loss": 3.7658583332086932, + "tokens_seen": 782434304 + }, + { + "epoch": 0.24, + "learning_rate": 0.0006164894684052157, + "loss": 1.4175, + "theoretical_loss": 3.7658271783278225, + "tokens_seen": 782499840 + }, + { + "epoch": 0.24, + "learning_rate": 0.0006164734202607823, + "loss": 1.5348, + "theoretical_loss": 3.765796026786659, + "tokens_seen": 782565376 + }, + { + "epoch": 0.24, + "learning_rate": 0.0006164573721163491, + "loss": 1.4253, + "theoretical_loss": 3.7657648785845663, + "tokens_seen": 782630912 + }, + { + "epoch": 0.24, + "learning_rate": 0.0006164413239719157, + "loss": 1.4857, + "theoretical_loss": 3.765733733720906, + "tokens_seen": 782696448 + }, + { + "epoch": 0.24, + "learning_rate": 0.0006164252758274825, + "loss": 1.4792, + "theoretical_loss": 3.7657025921950407, + "tokens_seen": 782761984 + }, + { + "epoch": 0.24, + "learning_rate": 0.0006164092276830491, + "loss": 1.3454, + "theoretical_loss": 3.765671454006334, + "tokens_seen": 782827520 + }, + { + "epoch": 0.24, + "learning_rate": 0.0006163931795386159, + "loss": 1.5476, + "theoretical_loss": 3.7656403191541483, + "tokens_seen": 782893056 + }, + { + "epoch": 0.24, + "learning_rate": 0.0006163771313941825, + "loss": 1.5631, + "theoretical_loss": 3.765609187637847, + "tokens_seen": 782958592 + }, + { + "epoch": 0.24, + "learning_rate": 0.0006163610832497493, + "loss": 1.3932, + "theoretical_loss": 3.765578059456794, + "tokens_seen": 783024128 + }, + { + "epoch": 0.24, + "learning_rate": 0.0006163450351053159, + "loss": 1.4497, + "theoretical_loss": 3.765546934610352, + "tokens_seen": 783089664 + }, + { + "epoch": 0.24, + "objective/train/docs_used": 289672, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.6059420108795166, + "objective/train/theoretical_loss": 3.7655158130978856, + "objective/train/tokens_used": 803615200, + "theoretical_loss": 3.7655158130978856, + "tokens_seen": 783155200 + }, + { + "epoch": 0.24, + "learning_rate": 0.0006163289869608827, + "loss": 1.5014, + "theoretical_loss": 3.7655158130978856, + "tokens_seen": 783155200 + }, + { + "epoch": 0.24, + "learning_rate": 0.0006163129388164493, + "loss": 1.5231, + "theoretical_loss": 3.765484694918758, + "tokens_seen": 783220736 + }, + { + "epoch": 0.24, + "learning_rate": 0.0006162968906720161, + "loss": 1.4691, + "theoretical_loss": 3.7654535800723337, + "tokens_seen": 783286272 + }, + { + "epoch": 0.24, + "learning_rate": 0.0006162808425275827, + "loss": 1.3959, + "theoretical_loss": 3.7654224685579765, + "tokens_seen": 783351808 + }, + { + "epoch": 0.24, + "learning_rate": 0.0006162647943831495, + "loss": 1.5218, + "theoretical_loss": 3.765391360375052, + "tokens_seen": 783417344 + }, + { + "epoch": 0.24, + "learning_rate": 0.0006162487462387161, + "loss": 1.5224, + "theoretical_loss": 3.765360255522924, + "tokens_seen": 783482880 + }, + { + "epoch": 0.24, + "learning_rate": 0.0006162326980942829, + "loss": 1.4561, + "theoretical_loss": 3.7653291540009572, + "tokens_seen": 783548416 + }, + { + "epoch": 0.24, + "learning_rate": 0.0006162166499498495, + "loss": 1.5561, + "theoretical_loss": 3.7652980558085165, + "tokens_seen": 783613952 + }, + { + "epoch": 0.24, + "learning_rate": 0.0006162006018054163, + "loss": 1.6035, + "theoretical_loss": 3.7652669609449676, + "tokens_seen": 783679488 + }, + { + "epoch": 0.24, + "learning_rate": 0.000616184553660983, + "loss": 1.6188, + "theoretical_loss": 3.7652358694096755, + "tokens_seen": 783745024 + }, + { + "epoch": 0.24, + "learning_rate": 0.0006161685055165497, + "loss": 1.4563, + "theoretical_loss": 3.765204781202006, + "tokens_seen": 783810560 + }, + { + "epoch": 0.24, + "learning_rate": 0.0006161524573721164, + "loss": 1.4358, + "theoretical_loss": 3.765173696321324, + "tokens_seen": 783876096 + }, + { + "epoch": 0.24, + "learning_rate": 0.0006161364092276831, + "loss": 1.439, + "theoretical_loss": 3.7651426147669964, + "tokens_seen": 783941632 + }, + { + "epoch": 0.24, + "learning_rate": 0.0006161203610832498, + "loss": 1.4867, + "theoretical_loss": 3.765111536538388, + "tokens_seen": 784007168 + }, + { + "epoch": 0.24, + "learning_rate": 0.0006161043129388165, + "loss": 1.4388, + "theoretical_loss": 3.7650804616348665, + "tokens_seen": 784072704 + }, + { + "epoch": 0.24, + "learning_rate": 0.0006160882647943832, + "loss": 1.3329, + "theoretical_loss": 3.765049390055797, + "tokens_seen": 784138240 + }, + { + "epoch": 0.24, + "learning_rate": 0.0006160722166499499, + "loss": 1.5001, + "theoretical_loss": 3.765018321800546, + "tokens_seen": 784203776 + }, + { + "epoch": 0.24, + "learning_rate": 0.0006160561685055166, + "loss": 1.5427, + "theoretical_loss": 3.7649872568684812, + "tokens_seen": 784269312 + }, + { + "epoch": 0.24, + "learning_rate": 0.0006160401203610833, + "loss": 1.5168, + "theoretical_loss": 3.764956195258969, + "tokens_seen": 784334848 + }, + { + "epoch": 0.24, + "learning_rate": 0.00061602407221665, + "loss": 1.347, + "theoretical_loss": 3.7649251369713763, + "tokens_seen": 784400384 + }, + { + "epoch": 0.24, + "learning_rate": 0.0006160080240722167, + "loss": 1.4088, + "theoretical_loss": 3.7648940820050707, + "tokens_seen": 784465920 + }, + { + "epoch": 0.24, + "learning_rate": 0.0006159919759277834, + "loss": 1.3721, + "theoretical_loss": 3.764863030359419, + "tokens_seen": 784531456 + }, + { + "epoch": 0.24, + "learning_rate": 0.0006159759277833501, + "loss": 1.3693, + "theoretical_loss": 3.7648319820337894, + "tokens_seen": 784596992 + }, + { + "epoch": 0.24, + "learning_rate": 0.0006159598796389168, + "loss": 1.479, + "theoretical_loss": 3.7648009370275495, + "tokens_seen": 784662528 + }, + { + "epoch": 0.24, + "learning_rate": 0.0006159438314944835, + "loss": 1.4448, + "theoretical_loss": 3.7647698953400672, + "tokens_seen": 784728064 + }, + { + "epoch": 0.24, + "learning_rate": 0.0006159277833500502, + "loss": 1.4841, + "theoretical_loss": 3.76473885697071, + "tokens_seen": 784793600 + }, + { + "epoch": 0.24, + "learning_rate": 0.0006159117352056169, + "loss": 1.4357, + "theoretical_loss": 3.7647078219188472, + "tokens_seen": 784859136 + }, + { + "epoch": 0.24, + "learning_rate": 0.0006158956870611836, + "loss": 1.5406, + "theoretical_loss": 3.7646767901838465, + "tokens_seen": 784924672 + }, + { + "epoch": 0.24, + "learning_rate": 0.0006158796389167503, + "loss": 1.4916, + "theoretical_loss": 3.764645761765077, + "tokens_seen": 784990208 + }, + { + "epoch": 0.24, + "learning_rate": 0.000615863590772317, + "loss": 1.4297, + "theoretical_loss": 3.7646147366619074, + "tokens_seen": 785055744 + }, + { + "epoch": 0.24, + "learning_rate": 0.0006158475426278837, + "loss": 1.5563, + "theoretical_loss": 3.764583714873706, + "tokens_seen": 785121280 + }, + { + "epoch": 0.24, + "learning_rate": 0.0006158314944834504, + "loss": 1.4305, + "theoretical_loss": 3.764552696399843, + "tokens_seen": 785186816 + }, + { + "epoch": 0.24, + "learning_rate": 0.0006158154463390171, + "loss": 1.4318, + "theoretical_loss": 3.764521681239687, + "tokens_seen": 785252352 + }, + { + "epoch": 0.24, + "learning_rate": 0.0006157993981945839, + "loss": 1.4599, + "theoretical_loss": 3.7644906693926075, + "tokens_seen": 785317888 + }, + { + "epoch": 0.24, + "learning_rate": 0.0006157833500501505, + "loss": 1.638, + "theoretical_loss": 3.7644596608579746, + "tokens_seen": 785383424 + }, + { + "epoch": 0.24, + "learning_rate": 0.0006157673019057173, + "loss": 1.4972, + "theoretical_loss": 3.764428655635158, + "tokens_seen": 785448960 + }, + { + "epoch": 0.24, + "learning_rate": 0.0006157512537612839, + "loss": 1.4118, + "theoretical_loss": 3.764397653723527, + "tokens_seen": 785514496 + }, + { + "epoch": 0.24, + "learning_rate": 0.0006157352056168507, + "loss": 1.4527, + "theoretical_loss": 3.7643666551224526, + "tokens_seen": 785580032 + }, + { + "epoch": 0.24, + "learning_rate": 0.0006157191574724173, + "loss": 1.4505, + "theoretical_loss": 3.764335659831305, + "tokens_seen": 785645568 + }, + { + "epoch": 0.24, + "learning_rate": 0.000615703109327984, + "loss": 1.4126, + "theoretical_loss": 3.7643046678494545, + "tokens_seen": 785711104 + }, + { + "epoch": 0.24, + "learning_rate": 0.0006156870611835507, + "loss": 1.4235, + "theoretical_loss": 3.764273679176272, + "tokens_seen": 785776640 + }, + { + "epoch": 0.24, + "learning_rate": 0.0006156710130391175, + "loss": 1.5925, + "theoretical_loss": 3.764242693811128, + "tokens_seen": 785842176 + }, + { + "epoch": 0.24, + "learning_rate": 0.0006156549648946841, + "loss": 1.5033, + "theoretical_loss": 3.764211711753394, + "tokens_seen": 785907712 + }, + { + "epoch": 0.24, + "learning_rate": 0.0006156389167502509, + "loss": 1.364, + "theoretical_loss": 3.7641807330024406, + "tokens_seen": 785973248 + }, + { + "epoch": 0.24, + "learning_rate": 0.0006156228686058175, + "loss": 1.3923, + "theoretical_loss": 3.7641497575576404, + "tokens_seen": 786038784 + }, + { + "epoch": 0.24, + "learning_rate": 0.0006156068204613843, + "loss": 1.4599, + "theoretical_loss": 3.764118785418363, + "tokens_seen": 786104320 + }, + { + "epoch": 0.24, + "learning_rate": 0.0006155907723169509, + "loss": 1.5791, + "theoretical_loss": 3.7640878165839817, + "tokens_seen": 786169856 + }, + { + "epoch": 0.24, + "learning_rate": 0.0006155747241725176, + "loss": 1.4551, + "theoretical_loss": 3.7640568510538683, + "tokens_seen": 786235392 + }, + { + "epoch": 0.24, + "learning_rate": 0.0006155586760280843, + "loss": 1.4268, + "theoretical_loss": 3.764025888827394, + "tokens_seen": 786300928 + }, + { + "epoch": 0.24, + "learning_rate": 0.000615542627883651, + "loss": 1.4468, + "theoretical_loss": 3.763994929903932, + "tokens_seen": 786366464 + }, + { + "epoch": 0.24, + "objective/train/docs_used": 290810, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.4015960693359375, + "objective/train/theoretical_loss": 3.763963974282854, + "objective/train/tokens_used": 806892000, + "theoretical_loss": 3.763963974282854, + "tokens_seen": 786432000 + }, + { + "epoch": 0.24, + "learning_rate": 0.0006155265797392177, + "loss": 1.4611, + "theoretical_loss": 3.763963974282854, + "tokens_seen": 786432000 + }, + { + "epoch": 0.24, + "learning_rate": 0.0006155105315947844, + "loss": 1.4332, + "theoretical_loss": 3.763933021963533, + "tokens_seen": 786497536 + }, + { + "epoch": 0.24, + "learning_rate": 0.0006154944834503511, + "loss": 1.4619, + "theoretical_loss": 3.763902072945341, + "tokens_seen": 786563072 + }, + { + "epoch": 0.24, + "learning_rate": 0.0006154784353059178, + "loss": 1.4479, + "theoretical_loss": 3.7638711272276524, + "tokens_seen": 786628608 + }, + { + "epoch": 0.24, + "learning_rate": 0.0006154623871614844, + "loss": 1.4787, + "theoretical_loss": 3.7638401848098386, + "tokens_seen": 786694144 + }, + { + "epoch": 0.24, + "learning_rate": 0.0006154463390170512, + "loss": 1.5858, + "theoretical_loss": 3.763809245691274, + "tokens_seen": 786759680 + }, + { + "epoch": 0.24, + "learning_rate": 0.0006154302908726178, + "loss": 1.4254, + "theoretical_loss": 3.7637783098713316, + "tokens_seen": 786825216 + }, + { + "epoch": 0.24, + "learning_rate": 0.0006154142427281846, + "loss": 1.4752, + "theoretical_loss": 3.7637473773493855, + "tokens_seen": 786890752 + }, + { + "epoch": 0.24, + "learning_rate": 0.0006153981945837512, + "loss": 1.4469, + "theoretical_loss": 3.7637164481248084, + "tokens_seen": 786956288 + }, + { + "epoch": 0.24, + "learning_rate": 0.000615382146439318, + "loss": 1.4117, + "theoretical_loss": 3.763685522196975, + "tokens_seen": 787021824 + }, + { + "epoch": 0.24, + "learning_rate": 0.0006153660982948846, + "loss": 1.5249, + "theoretical_loss": 3.7636545995652604, + "tokens_seen": 787087360 + }, + { + "epoch": 0.24, + "learning_rate": 0.0006153500501504514, + "loss": 1.5829, + "theoretical_loss": 3.7636236802290366, + "tokens_seen": 787152896 + }, + { + "epoch": 0.24, + "learning_rate": 0.000615334002006018, + "loss": 1.5226, + "theoretical_loss": 3.7635927641876803, + "tokens_seen": 787218432 + }, + { + "epoch": 0.24, + "learning_rate": 0.0006153179538615848, + "loss": 1.5167, + "theoretical_loss": 3.7635618514405644, + "tokens_seen": 787283968 + }, + { + "epoch": 0.24, + "learning_rate": 0.0006153019057171514, + "loss": 1.5493, + "theoretical_loss": 3.7635309419870646, + "tokens_seen": 787349504 + }, + { + "epoch": 0.24, + "learning_rate": 0.0006152858575727182, + "loss": 1.4139, + "theoretical_loss": 3.763500035826556, + "tokens_seen": 787415040 + }, + { + "epoch": 0.24, + "learning_rate": 0.0006152698094282848, + "loss": 1.4006, + "theoretical_loss": 3.7634691329584133, + "tokens_seen": 787480576 + }, + { + "epoch": 0.24, + "learning_rate": 0.0006152537612838516, + "loss": 1.4621, + "theoretical_loss": 3.763438233382012, + "tokens_seen": 787546112 + }, + { + "epoch": 0.24, + "learning_rate": 0.0006152377131394182, + "loss": 1.5444, + "theoretical_loss": 3.7634073370967274, + "tokens_seen": 787611648 + }, + { + "epoch": 0.24, + "learning_rate": 0.000615221664994985, + "loss": 1.4051, + "theoretical_loss": 3.7633764441019357, + "tokens_seen": 787677184 + }, + { + "epoch": 0.24, + "learning_rate": 0.0006152056168505516, + "loss": 1.3565, + "theoretical_loss": 3.763345554397012, + "tokens_seen": 787742720 + }, + { + "epoch": 0.24, + "learning_rate": 0.0006151895687061184, + "loss": 1.4798, + "theoretical_loss": 3.763314667981333, + "tokens_seen": 787808256 + }, + { + "epoch": 0.24, + "learning_rate": 0.000615173520561685, + "loss": 1.5913, + "theoretical_loss": 3.7632837848542744, + "tokens_seen": 787873792 + }, + { + "epoch": 0.24, + "learning_rate": 0.0006151574724172518, + "loss": 1.3079, + "theoretical_loss": 3.7632529050152126, + "tokens_seen": 787939328 + }, + { + "epoch": 0.24, + "learning_rate": 0.0006151414242728184, + "loss": 1.4553, + "theoretical_loss": 3.7632220284635247, + "tokens_seen": 788004864 + }, + { + "epoch": 0.24, + "learning_rate": 0.0006151253761283852, + "loss": 1.4123, + "theoretical_loss": 3.763191155198586, + "tokens_seen": 788070400 + }, + { + "epoch": 0.24, + "learning_rate": 0.0006151093279839518, + "loss": 1.4663, + "theoretical_loss": 3.7631602852197745, + "tokens_seen": 788135936 + }, + { + "epoch": 0.24, + "learning_rate": 0.0006150932798395186, + "loss": 1.5333, + "theoretical_loss": 3.763129418526467, + "tokens_seen": 788201472 + }, + { + "epoch": 0.24, + "learning_rate": 0.0006150772316950852, + "loss": 1.3906, + "theoretical_loss": 3.7630985551180407, + "tokens_seen": 788267008 + }, + { + "epoch": 0.24, + "learning_rate": 0.000615061183550652, + "loss": 1.416, + "theoretical_loss": 3.7630676949938726, + "tokens_seen": 788332544 + }, + { + "epoch": 0.24, + "learning_rate": 0.0006150451354062186, + "loss": 1.4162, + "theoretical_loss": 3.763036838153341, + "tokens_seen": 788398080 + }, + { + "epoch": 0.24, + "learning_rate": 0.0006150290872617854, + "loss": 1.5505, + "theoretical_loss": 3.763005984595823, + "tokens_seen": 788463616 + }, + { + "epoch": 0.24, + "learning_rate": 0.0006150130391173521, + "loss": 1.5098, + "theoretical_loss": 3.762975134320696, + "tokens_seen": 788529152 + }, + { + "epoch": 0.24, + "learning_rate": 0.0006149969909729188, + "loss": 1.4855, + "theoretical_loss": 3.762944287327339, + "tokens_seen": 788594688 + }, + { + "epoch": 0.24, + "learning_rate": 0.0006149809428284855, + "loss": 1.4069, + "theoretical_loss": 3.7629134436151297, + "tokens_seen": 788660224 + }, + { + "epoch": 0.24, + "learning_rate": 0.0006149648946840522, + "loss": 1.5309, + "theoretical_loss": 3.7628826031834466, + "tokens_seen": 788725760 + }, + { + "epoch": 0.24, + "learning_rate": 0.0006149488465396189, + "loss": 1.5127, + "theoretical_loss": 3.7628517660316687, + "tokens_seen": 788791296 + }, + { + "epoch": 0.24, + "learning_rate": 0.0006149327983951856, + "loss": 1.5097, + "theoretical_loss": 3.7628209321591735, + "tokens_seen": 788856832 + }, + { + "epoch": 0.24, + "learning_rate": 0.0006149167502507523, + "loss": 1.5499, + "theoretical_loss": 3.762790101565341, + "tokens_seen": 788922368 + }, + { + "epoch": 0.24, + "learning_rate": 0.000614900702106319, + "loss": 1.4318, + "theoretical_loss": 3.76275927424955, + "tokens_seen": 788987904 + }, + { + "epoch": 0.24, + "learning_rate": 0.0006148846539618857, + "loss": 1.4221, + "theoretical_loss": 3.7627284502111795, + "tokens_seen": 789053440 + }, + { + "epoch": 0.24, + "learning_rate": 0.0006148686058174524, + "loss": 1.5421, + "theoretical_loss": 3.762697629449609, + "tokens_seen": 789118976 + }, + { + "epoch": 0.24, + "learning_rate": 0.0006148525576730191, + "loss": 1.4305, + "theoretical_loss": 3.762666811964218, + "tokens_seen": 789184512 + }, + { + "epoch": 0.24, + "learning_rate": 0.0006148365095285858, + "loss": 1.5148, + "theoretical_loss": 3.762635997754386, + "tokens_seen": 789250048 + }, + { + "epoch": 0.24, + "learning_rate": 0.0006148204613841525, + "loss": 1.4171, + "theoretical_loss": 3.762605186819494, + "tokens_seen": 789315584 + }, + { + "epoch": 0.24, + "learning_rate": 0.0006148044132397192, + "loss": 1.5243, + "theoretical_loss": 3.76257437915892, + "tokens_seen": 789381120 + }, + { + "epoch": 0.24, + "learning_rate": 0.0006147883650952859, + "loss": 1.4028, + "theoretical_loss": 3.762543574772047, + "tokens_seen": 789446656 + }, + { + "epoch": 0.24, + "learning_rate": 0.0006147723169508526, + "loss": 1.4056, + "theoretical_loss": 3.7625127736582527, + "tokens_seen": 789512192 + }, + { + "epoch": 0.24, + "learning_rate": 0.0006147562688064193, + "loss": 1.4322, + "theoretical_loss": 3.7624819758169195, + "tokens_seen": 789577728 + }, + { + "epoch": 0.24, + "learning_rate": 0.000614740220661986, + "loss": 1.4521, + "theoretical_loss": 3.7624511812474273, + "tokens_seen": 789643264 + }, + { + "epoch": 0.24, + "objective/train/docs_used": 292000, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.663053035736084, + "objective/train/theoretical_loss": 3.762420389949157, + "objective/train/tokens_used": 810168800, + "theoretical_loss": 3.762420389949157, + "tokens_seen": 789708800 + }, + { + "epoch": 0.24, + "learning_rate": 0.0006147241725175527, + "loss": 1.5575, + "theoretical_loss": 3.762420389949157, + "tokens_seen": 789708800 + }, + { + "epoch": 0.24, + "learning_rate": 0.0006147081243731194, + "loss": 1.5106, + "theoretical_loss": 3.76238960192149, + "tokens_seen": 789774336 + }, + { + "epoch": 0.24, + "learning_rate": 0.0006146920762286861, + "loss": 1.4514, + "theoretical_loss": 3.762358817163808, + "tokens_seen": 789839872 + }, + { + "epoch": 0.24, + "learning_rate": 0.0006146760280842528, + "loss": 1.5178, + "theoretical_loss": 3.762328035675491, + "tokens_seen": 789905408 + }, + { + "epoch": 0.24, + "learning_rate": 0.0006146599799398195, + "loss": 1.4531, + "theoretical_loss": 3.7622972574559226, + "tokens_seen": 789970944 + }, + { + "epoch": 0.24, + "learning_rate": 0.0006146439317953862, + "loss": 1.4839, + "theoretical_loss": 3.762266482504483, + "tokens_seen": 790036480 + }, + { + "epoch": 0.24, + "learning_rate": 0.0006146278836509529, + "loss": 1.4999, + "theoretical_loss": 3.762235710820554, + "tokens_seen": 790102016 + }, + { + "epoch": 0.24, + "learning_rate": 0.0006146118355065196, + "loss": 1.5003, + "theoretical_loss": 3.7622049424035184, + "tokens_seen": 790167552 + }, + { + "epoch": 0.24, + "learning_rate": 0.0006145957873620863, + "loss": 1.5529, + "theoretical_loss": 3.762174177252758, + "tokens_seen": 790233088 + }, + { + "epoch": 0.24, + "learning_rate": 0.000614579739217653, + "loss": 1.463, + "theoretical_loss": 3.7621434153676563, + "tokens_seen": 790298624 + }, + { + "epoch": 0.24, + "learning_rate": 0.0006145636910732197, + "loss": 1.6136, + "theoretical_loss": 3.7621126567475947, + "tokens_seen": 790364160 + }, + { + "epoch": 0.24, + "learning_rate": 0.0006145476429287864, + "loss": 1.4896, + "theoretical_loss": 3.762081901391956, + "tokens_seen": 790429696 + }, + { + "epoch": 0.24, + "learning_rate": 0.0006145315947843531, + "loss": 1.4384, + "theoretical_loss": 3.7620511493001243, + "tokens_seen": 790495232 + }, + { + "epoch": 0.24, + "learning_rate": 0.0006145155466399198, + "loss": 1.4925, + "theoretical_loss": 3.7620204004714815, + "tokens_seen": 790560768 + }, + { + "epoch": 0.24, + "learning_rate": 0.0006144994984954865, + "loss": 1.46, + "theoretical_loss": 3.761989654905411, + "tokens_seen": 790626304 + }, + { + "epoch": 0.24, + "learning_rate": 0.0006144834503510532, + "loss": 1.4922, + "theoretical_loss": 3.7619589126012967, + "tokens_seen": 790691840 + }, + { + "epoch": 0.24, + "learning_rate": 0.00061446740220662, + "loss": 1.6275, + "theoretical_loss": 3.761928173558522, + "tokens_seen": 790757376 + }, + { + "epoch": 0.24, + "learning_rate": 0.0006144513540621866, + "loss": 1.519, + "theoretical_loss": 3.7618974377764705, + "tokens_seen": 790822912 + }, + { + "epoch": 0.24, + "learning_rate": 0.0006144353059177534, + "loss": 1.4803, + "theoretical_loss": 3.7618667052545263, + "tokens_seen": 790888448 + }, + { + "epoch": 0.24, + "learning_rate": 0.00061441925777332, + "loss": 1.4849, + "theoretical_loss": 3.7618359759920734, + "tokens_seen": 790953984 + }, + { + "epoch": 0.24, + "learning_rate": 0.0006144032096288868, + "loss": 1.59, + "theoretical_loss": 3.7618052499884964, + "tokens_seen": 791019520 + }, + { + "epoch": 0.24, + "learning_rate": 0.0006143871614844534, + "loss": 1.4441, + "theoretical_loss": 3.7617745272431797, + "tokens_seen": 791085056 + }, + { + "epoch": 0.24, + "learning_rate": 0.0006143711133400202, + "loss": 1.4836, + "theoretical_loss": 3.761743807755507, + "tokens_seen": 791150592 + }, + { + "epoch": 0.24, + "learning_rate": 0.0006143550651955868, + "loss": 1.4861, + "theoretical_loss": 3.7617130915248644, + "tokens_seen": 791216128 + }, + { + "epoch": 0.24, + "learning_rate": 0.0006143390170511536, + "loss": 1.5172, + "theoretical_loss": 3.7616823785506357, + "tokens_seen": 791281664 + }, + { + "epoch": 0.24, + "learning_rate": 0.0006143229689067202, + "loss": 1.5954, + "theoretical_loss": 3.761651668832207, + "tokens_seen": 791347200 + }, + { + "epoch": 0.24, + "learning_rate": 0.000614306920762287, + "loss": 1.4634, + "theoretical_loss": 3.761620962368963, + "tokens_seen": 791412736 + }, + { + "epoch": 0.24, + "learning_rate": 0.0006142908726178536, + "loss": 1.4578, + "theoretical_loss": 3.7615902591602888, + "tokens_seen": 791478272 + }, + { + "epoch": 0.24, + "learning_rate": 0.0006142748244734204, + "loss": 1.4521, + "theoretical_loss": 3.761559559205571, + "tokens_seen": 791543808 + }, + { + "epoch": 0.24, + "learning_rate": 0.000614258776328987, + "loss": 1.482, + "theoretical_loss": 3.7615288625041945, + "tokens_seen": 791609344 + }, + { + "epoch": 0.24, + "learning_rate": 0.0006142427281845538, + "loss": 1.3844, + "theoretical_loss": 3.7614981690555456, + "tokens_seen": 791674880 + }, + { + "epoch": 0.24, + "learning_rate": 0.0006142266800401204, + "loss": 1.4786, + "theoretical_loss": 3.7614674788590103, + "tokens_seen": 791740416 + }, + { + "epoch": 0.24, + "learning_rate": 0.0006142106318956872, + "loss": 1.5675, + "theoretical_loss": 3.761436791913975, + "tokens_seen": 791805952 + }, + { + "epoch": 0.24, + "learning_rate": 0.0006141945837512538, + "loss": 1.481, + "theoretical_loss": 3.761406108219826, + "tokens_seen": 791871488 + }, + { + "epoch": 0.24, + "learning_rate": 0.0006141785356068205, + "loss": 1.3825, + "theoretical_loss": 3.76137542777595, + "tokens_seen": 791937024 + }, + { + "epoch": 0.24, + "learning_rate": 0.0006141624874623872, + "loss": 1.5071, + "theoretical_loss": 3.761344750581734, + "tokens_seen": 792002560 + }, + { + "epoch": 0.24, + "learning_rate": 0.0006141464393179539, + "loss": 1.3916, + "theoretical_loss": 3.7613140766365643, + "tokens_seen": 792068096 + }, + { + "epoch": 0.24, + "learning_rate": 0.0006141303911735206, + "loss": 1.4356, + "theoretical_loss": 3.7612834059398286, + "tokens_seen": 792133632 + }, + { + "epoch": 0.24, + "learning_rate": 0.0006141143430290873, + "loss": 1.5532, + "theoretical_loss": 3.7612527384909136, + "tokens_seen": 792199168 + }, + { + "epoch": 0.24, + "learning_rate": 0.000614098294884654, + "loss": 1.3948, + "theoretical_loss": 3.7612220742892077, + "tokens_seen": 792264704 + }, + { + "epoch": 0.24, + "learning_rate": 0.0006140822467402207, + "loss": 1.4846, + "theoretical_loss": 3.761191413334098, + "tokens_seen": 792330240 + }, + { + "epoch": 0.24, + "learning_rate": 0.0006140661985957874, + "loss": 1.4907, + "theoretical_loss": 3.7611607556249718, + "tokens_seen": 792395776 + }, + { + "epoch": 0.24, + "learning_rate": 0.0006140501504513541, + "loss": 1.3997, + "theoretical_loss": 3.761130101161217, + "tokens_seen": 792461312 + }, + { + "epoch": 0.24, + "learning_rate": 0.0006140341023069207, + "loss": 1.4975, + "theoretical_loss": 3.761099449942223, + "tokens_seen": 792526848 + }, + { + "epoch": 0.24, + "learning_rate": 0.0006140180541624875, + "loss": 1.4128, + "theoretical_loss": 3.7610688019673764, + "tokens_seen": 792592384 + }, + { + "epoch": 0.24, + "learning_rate": 0.0006140020060180541, + "loss": 1.4749, + "theoretical_loss": 3.761038157236067, + "tokens_seen": 792657920 + }, + { + "epoch": 0.24, + "learning_rate": 0.0006139859578736209, + "loss": 1.4049, + "theoretical_loss": 3.7610075157476826, + "tokens_seen": 792723456 + }, + { + "epoch": 0.24, + "learning_rate": 0.0006139699097291875, + "loss": 1.4943, + "theoretical_loss": 3.760976877501612, + "tokens_seen": 792788992 + }, + { + "epoch": 0.24, + "learning_rate": 0.0006139538615847543, + "loss": 1.5094, + "theoretical_loss": 3.7609462424972446, + "tokens_seen": 792854528 + }, + { + "epoch": 0.24, + "learning_rate": 0.0006139378134403209, + "loss": 1.465, + "theoretical_loss": 3.760915610733969, + "tokens_seen": 792920064 + }, + { + "epoch": 0.24, + "objective/train/docs_used": 293128, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.4129773378372192, + "objective/train/theoretical_loss": 3.7608849822111754, + "objective/train/tokens_used": 813445600, + "theoretical_loss": 3.7608849822111754, + "tokens_seen": 792985600 + }, + { + "epoch": 0.24, + "learning_rate": 0.0006139217652958877, + "loss": 1.5249, + "theoretical_loss": 3.7608849822111754, + "tokens_seen": 792985600 + }, + { + "epoch": 0.24, + "learning_rate": 0.0006139057171514543, + "loss": 1.4892, + "theoretical_loss": 3.7608543569282515, + "tokens_seen": 793051136 + }, + { + "epoch": 0.24, + "learning_rate": 0.0006138896690070211, + "loss": 1.443, + "theoretical_loss": 3.7608237348845885, + "tokens_seen": 793116672 + }, + { + "epoch": 0.24, + "learning_rate": 0.0006138736208625877, + "loss": 1.5864, + "theoretical_loss": 3.760793116079576, + "tokens_seen": 793182208 + }, + { + "epoch": 0.24, + "learning_rate": 0.0006138575727181545, + "loss": 1.4592, + "theoretical_loss": 3.7607625005126026, + "tokens_seen": 793247744 + }, + { + "epoch": 0.24, + "learning_rate": 0.0006138415245737211, + "loss": 1.51, + "theoretical_loss": 3.7607318881830594, + "tokens_seen": 793313280 + }, + { + "epoch": 0.24, + "learning_rate": 0.0006138254764292879, + "loss": 1.4989, + "theoretical_loss": 3.7607012790903367, + "tokens_seen": 793378816 + }, + { + "epoch": 0.24, + "learning_rate": 0.0006138094282848546, + "loss": 1.4439, + "theoretical_loss": 3.7606706732338253, + "tokens_seen": 793444352 + }, + { + "epoch": 0.24, + "learning_rate": 0.0006137933801404213, + "loss": 1.595, + "theoretical_loss": 3.7606400706129146, + "tokens_seen": 793509888 + }, + { + "epoch": 0.24, + "learning_rate": 0.000613777331995988, + "loss": 1.5342, + "theoretical_loss": 3.760609471226996, + "tokens_seen": 793575424 + }, + { + "epoch": 0.24, + "learning_rate": 0.0006137612838515547, + "loss": 1.5887, + "theoretical_loss": 3.7605788750754607, + "tokens_seen": 793640960 + }, + { + "epoch": 0.24, + "learning_rate": 0.0006137452357071214, + "loss": 1.512, + "theoretical_loss": 3.7605482821577, + "tokens_seen": 793706496 + }, + { + "epoch": 0.24, + "learning_rate": 0.0006137291875626881, + "loss": 1.4722, + "theoretical_loss": 3.760517692473104, + "tokens_seen": 793772032 + }, + { + "epoch": 0.24, + "learning_rate": 0.0006137131394182548, + "loss": 1.5224, + "theoretical_loss": 3.7604871060210643, + "tokens_seen": 793837568 + }, + { + "epoch": 0.24, + "learning_rate": 0.0006136970912738215, + "loss": 1.5667, + "theoretical_loss": 3.7604565228009736, + "tokens_seen": 793903104 + }, + { + "epoch": 0.24, + "learning_rate": 0.0006136810431293882, + "loss": 1.4578, + "theoretical_loss": 3.760425942812223, + "tokens_seen": 793968640 + }, + { + "epoch": 0.24, + "learning_rate": 0.0006136649949849549, + "loss": 1.4166, + "theoretical_loss": 3.760395366054204, + "tokens_seen": 794034176 + }, + { + "epoch": 0.24, + "learning_rate": 0.0006136489468405216, + "loss": 1.5343, + "theoretical_loss": 3.7603647925263095, + "tokens_seen": 794099712 + }, + { + "epoch": 0.24, + "learning_rate": 0.0006136328986960883, + "loss": 1.3338, + "theoretical_loss": 3.7603342222279315, + "tokens_seen": 794165248 + }, + { + "epoch": 0.24, + "learning_rate": 0.000613616850551655, + "loss": 1.5065, + "theoretical_loss": 3.760303655158461, + "tokens_seen": 794230784 + }, + { + "epoch": 0.24, + "learning_rate": 0.0006136008024072217, + "loss": 1.4336, + "theoretical_loss": 3.760273091317293, + "tokens_seen": 794296320 + }, + { + "epoch": 0.24, + "learning_rate": 0.0006135847542627884, + "loss": 1.527, + "theoretical_loss": 3.7602425307038185, + "tokens_seen": 794361856 + }, + { + "epoch": 0.24, + "learning_rate": 0.0006135687061183551, + "loss": 1.439, + "theoretical_loss": 3.760211973317431, + "tokens_seen": 794427392 + }, + { + "epoch": 0.24, + "learning_rate": 0.0006135526579739218, + "loss": 1.6154, + "theoretical_loss": 3.760181419157523, + "tokens_seen": 794492928 + }, + { + "epoch": 0.24, + "learning_rate": 0.0006135366098294885, + "loss": 1.404, + "theoretical_loss": 3.7601508682234885, + "tokens_seen": 794558464 + }, + { + "epoch": 0.24, + "learning_rate": 0.0006135205616850552, + "loss": 1.5512, + "theoretical_loss": 3.7601203205147202, + "tokens_seen": 794624000 + }, + { + "epoch": 0.24, + "learning_rate": 0.0006135045135406219, + "loss": 1.4535, + "theoretical_loss": 3.760089776030613, + "tokens_seen": 794689536 + }, + { + "epoch": 0.24, + "learning_rate": 0.0006134884653961886, + "loss": 1.4456, + "theoretical_loss": 3.7600592347705586, + "tokens_seen": 794755072 + }, + { + "epoch": 0.24, + "learning_rate": 0.0006134724172517553, + "loss": 1.4992, + "theoretical_loss": 3.7600286967339525, + "tokens_seen": 794820608 + }, + { + "epoch": 0.24, + "learning_rate": 0.000613456369107322, + "loss": 1.5034, + "theoretical_loss": 3.759998161920187, + "tokens_seen": 794886144 + }, + { + "epoch": 0.24, + "learning_rate": 0.0006134403209628887, + "loss": 1.4991, + "theoretical_loss": 3.7599676303286587, + "tokens_seen": 794951680 + }, + { + "epoch": 0.24, + "learning_rate": 0.0006134242728184554, + "loss": 1.5168, + "theoretical_loss": 3.75993710195876, + "tokens_seen": 795017216 + }, + { + "epoch": 0.24, + "learning_rate": 0.0006134082246740221, + "loss": 1.4738, + "theoretical_loss": 3.759906576809886, + "tokens_seen": 795082752 + }, + { + "epoch": 0.24, + "learning_rate": 0.0006133921765295888, + "loss": 1.455, + "theoretical_loss": 3.759876054881432, + "tokens_seen": 795148288 + }, + { + "epoch": 0.24, + "learning_rate": 0.0006133761283851555, + "loss": 1.471, + "theoretical_loss": 3.759845536172792, + "tokens_seen": 795213824 + }, + { + "epoch": 0.24, + "learning_rate": 0.0006133600802407222, + "loss": 1.6298, + "theoretical_loss": 3.7598150206833614, + "tokens_seen": 795279360 + }, + { + "epoch": 0.24, + "learning_rate": 0.0006133440320962889, + "loss": 1.5098, + "theoretical_loss": 3.759784508412536, + "tokens_seen": 795344896 + }, + { + "epoch": 0.24, + "learning_rate": 0.0006133279839518556, + "loss": 1.507, + "theoretical_loss": 3.7597539993597096, + "tokens_seen": 795410432 + }, + { + "epoch": 0.24, + "learning_rate": 0.0006133119358074223, + "loss": 1.4566, + "theoretical_loss": 3.7597234935242794, + "tokens_seen": 795475968 + }, + { + "epoch": 0.24, + "learning_rate": 0.000613295887662989, + "loss": 1.4714, + "theoretical_loss": 3.7596929909056396, + "tokens_seen": 795541504 + }, + { + "epoch": 0.24, + "learning_rate": 0.0006132798395185557, + "loss": 1.4125, + "theoretical_loss": 3.759662491503187, + "tokens_seen": 795607040 + }, + { + "epoch": 0.24, + "learning_rate": 0.0006132637913741224, + "loss": 1.5198, + "theoretical_loss": 3.759631995316318, + "tokens_seen": 795672576 + }, + { + "epoch": 0.24, + "learning_rate": 0.0006132477432296891, + "loss": 1.517, + "theoretical_loss": 3.7596015023444274, + "tokens_seen": 795738112 + }, + { + "epoch": 0.24, + "learning_rate": 0.0006132316950852558, + "loss": 1.4586, + "theoretical_loss": 3.759571012586913, + "tokens_seen": 795803648 + }, + { + "epoch": 0.24, + "learning_rate": 0.0006132156469408225, + "loss": 1.386, + "theoretical_loss": 3.75954052604317, + "tokens_seen": 795869184 + }, + { + "epoch": 0.24, + "learning_rate": 0.0006131995987963892, + "loss": 1.4654, + "theoretical_loss": 3.759510042712596, + "tokens_seen": 795934720 + }, + { + "epoch": 0.24, + "learning_rate": 0.0006131835506519559, + "loss": 1.6529, + "theoretical_loss": 3.7594795625945867, + "tokens_seen": 796000256 + }, + { + "epoch": 0.24, + "learning_rate": 0.0006131675025075226, + "loss": 1.5034, + "theoretical_loss": 3.75944908568854, + "tokens_seen": 796065792 + }, + { + "epoch": 0.24, + "learning_rate": 0.0006131514543630893, + "loss": 1.4501, + "theoretical_loss": 3.759418611993853, + "tokens_seen": 796131328 + }, + { + "epoch": 0.24, + "learning_rate": 0.000613135406218656, + "loss": 1.518, + "theoretical_loss": 3.7593881415099233, + "tokens_seen": 796196864 + }, + { + "epoch": 0.24, + "objective/train/docs_used": 294439, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.5768048763275146, + "objective/train/theoretical_loss": 3.7593576742361474, + "objective/train/tokens_used": 816722400, + "theoretical_loss": 3.7593576742361474, + "tokens_seen": 796262400 + }, + { + "epoch": 0.24, + "learning_rate": 0.0006131193580742227, + "loss": 1.5865, + "theoretical_loss": 3.7593576742361474, + "tokens_seen": 796262400 + }, + { + "epoch": 0.24, + "learning_rate": 0.0006131033099297894, + "loss": 1.457, + "theoretical_loss": 3.759327210171924, + "tokens_seen": 796327936 + }, + { + "epoch": 0.24, + "learning_rate": 0.0006130872617853561, + "loss": 1.5294, + "theoretical_loss": 3.7592967493166496, + "tokens_seen": 796393472 + }, + { + "epoch": 0.24, + "learning_rate": 0.0006130712136409228, + "loss": 1.5606, + "theoretical_loss": 3.7592662916697237, + "tokens_seen": 796459008 + }, + { + "epoch": 0.24, + "learning_rate": 0.0006130551654964895, + "loss": 1.5105, + "theoretical_loss": 3.759235837230543, + "tokens_seen": 796524544 + }, + { + "epoch": 0.24, + "learning_rate": 0.0006130391173520563, + "loss": 1.4948, + "theoretical_loss": 3.7592053859985066, + "tokens_seen": 796590080 + }, + { + "epoch": 0.24, + "learning_rate": 0.0006130230692076229, + "loss": 1.4812, + "theoretical_loss": 3.759174937973013, + "tokens_seen": 796655616 + }, + { + "epoch": 0.24, + "learning_rate": 0.0006130070210631897, + "loss": 1.3602, + "theoretical_loss": 3.7591444931534603, + "tokens_seen": 796721152 + }, + { + "epoch": 0.24, + "learning_rate": 0.0006129909729187563, + "loss": 1.4082, + "theoretical_loss": 3.759114051539248, + "tokens_seen": 796786688 + }, + { + "epoch": 0.24, + "learning_rate": 0.000612974924774323, + "loss": 1.4587, + "theoretical_loss": 3.759083613129774, + "tokens_seen": 796852224 + }, + { + "epoch": 0.24, + "learning_rate": 0.0006129588766298897, + "loss": 1.5056, + "theoretical_loss": 3.7590531779244385, + "tokens_seen": 796917760 + }, + { + "epoch": 0.24, + "learning_rate": 0.0006129428284854565, + "loss": 1.4384, + "theoretical_loss": 3.75902274592264, + "tokens_seen": 796983296 + }, + { + "epoch": 0.24, + "learning_rate": 0.0006129267803410231, + "loss": 1.355, + "theoretical_loss": 3.758992317123778, + "tokens_seen": 797048832 + }, + { + "epoch": 0.24, + "learning_rate": 0.0006129107321965899, + "loss": 1.4363, + "theoretical_loss": 3.7589618915272522, + "tokens_seen": 797114368 + }, + { + "epoch": 0.24, + "learning_rate": 0.0006128946840521565, + "loss": 1.4116, + "theoretical_loss": 3.7589314691324622, + "tokens_seen": 797179904 + }, + { + "epoch": 0.24, + "learning_rate": 0.0006128786359077233, + "loss": 1.457, + "theoretical_loss": 3.7589010499388085, + "tokens_seen": 797245440 + }, + { + "epoch": 0.24, + "learning_rate": 0.0006128625877632899, + "loss": 1.4185, + "theoretical_loss": 3.7588706339456905, + "tokens_seen": 797310976 + }, + { + "epoch": 0.24, + "learning_rate": 0.0006128465396188567, + "loss": 1.48, + "theoretical_loss": 3.7588402211525085, + "tokens_seen": 797376512 + }, + { + "epoch": 0.24, + "learning_rate": 0.0006128304914744233, + "loss": 1.4536, + "theoretical_loss": 3.7588098115586632, + "tokens_seen": 797442048 + }, + { + "epoch": 0.24, + "learning_rate": 0.0006128144433299901, + "loss": 1.5504, + "theoretical_loss": 3.758779405163555, + "tokens_seen": 797507584 + }, + { + "epoch": 0.24, + "learning_rate": 0.0006127983951855567, + "loss": 1.5609, + "theoretical_loss": 3.7587490019665846, + "tokens_seen": 797573120 + }, + { + "epoch": 0.24, + "learning_rate": 0.0006127823470411234, + "loss": 1.5782, + "theoretical_loss": 3.7587186019671526, + "tokens_seen": 797638656 + }, + { + "epoch": 0.24, + "learning_rate": 0.0006127662988966901, + "loss": 1.4973, + "theoretical_loss": 3.7586882051646606, + "tokens_seen": 797704192 + }, + { + "epoch": 0.24, + "learning_rate": 0.0006127502507522568, + "loss": 1.5516, + "theoretical_loss": 3.7586578115585096, + "tokens_seen": 797769728 + }, + { + "epoch": 0.24, + "learning_rate": 0.0006127342026078235, + "loss": 1.4019, + "theoretical_loss": 3.758627421148101, + "tokens_seen": 797835264 + }, + { + "epoch": 0.24, + "learning_rate": 0.0006127181544633902, + "loss": 1.4319, + "theoretical_loss": 3.758597033932836, + "tokens_seen": 797900800 + }, + { + "epoch": 0.24, + "learning_rate": 0.0006127021063189569, + "loss": 1.5336, + "theoretical_loss": 3.758566649912116, + "tokens_seen": 797966336 + }, + { + "epoch": 0.24, + "learning_rate": 0.0006126860581745236, + "loss": 1.4498, + "theoretical_loss": 3.7585362690853445, + "tokens_seen": 798031872 + }, + { + "epoch": 0.24, + "learning_rate": 0.0006126700100300903, + "loss": 1.6203, + "theoretical_loss": 3.758505891451921, + "tokens_seen": 798097408 + }, + { + "epoch": 0.24, + "learning_rate": 0.000612653961885657, + "loss": 1.4872, + "theoretical_loss": 3.7584755170112496, + "tokens_seen": 798162944 + }, + { + "epoch": 0.24, + "learning_rate": 0.0006126379137412236, + "loss": 1.5364, + "theoretical_loss": 3.758445145762732, + "tokens_seen": 798228480 + }, + { + "epoch": 0.24, + "learning_rate": 0.0006126218655967904, + "loss": 1.4863, + "theoretical_loss": 3.758414777705771, + "tokens_seen": 798294016 + }, + { + "epoch": 0.24, + "learning_rate": 0.0006126058174523571, + "loss": 1.4443, + "theoretical_loss": 3.7583844128397685, + "tokens_seen": 798359552 + }, + { + "epoch": 0.24, + "learning_rate": 0.0006125897693079238, + "loss": 1.5105, + "theoretical_loss": 3.758354051164128, + "tokens_seen": 798425088 + }, + { + "epoch": 0.24, + "learning_rate": 0.0006125737211634905, + "loss": 1.4217, + "theoretical_loss": 3.7583236926782515, + "tokens_seen": 798490624 + }, + { + "epoch": 0.24, + "learning_rate": 0.0006125576730190572, + "loss": 1.5392, + "theoretical_loss": 3.7582933373815433, + "tokens_seen": 798556160 + }, + { + "epoch": 0.24, + "learning_rate": 0.0006125416248746239, + "loss": 1.429, + "theoretical_loss": 3.758262985273406, + "tokens_seen": 798621696 + }, + { + "epoch": 0.24, + "learning_rate": 0.0006125255767301906, + "loss": 1.4012, + "theoretical_loss": 3.7582326363532443, + "tokens_seen": 798687232 + }, + { + "epoch": 0.24, + "learning_rate": 0.0006125095285857573, + "loss": 1.5375, + "theoretical_loss": 3.75820229062046, + "tokens_seen": 798752768 + }, + { + "epoch": 0.24, + "learning_rate": 0.000612493480441324, + "loss": 1.5488, + "theoretical_loss": 3.758171948074457, + "tokens_seen": 798818304 + }, + { + "epoch": 0.24, + "learning_rate": 0.0006124774322968907, + "loss": 1.4496, + "theoretical_loss": 3.7581416087146406, + "tokens_seen": 798883840 + }, + { + "epoch": 0.24, + "learning_rate": 0.0006124613841524574, + "loss": 1.4312, + "theoretical_loss": 3.7581112725404147, + "tokens_seen": 798949376 + }, + { + "epoch": 0.24, + "learning_rate": 0.0006124453360080241, + "loss": 1.4683, + "theoretical_loss": 3.7580809395511823, + "tokens_seen": 799014912 + }, + { + "epoch": 0.24, + "learning_rate": 0.0006124292878635908, + "loss": 1.5243, + "theoretical_loss": 3.7580506097463484, + "tokens_seen": 799080448 + }, + { + "epoch": 0.24, + "learning_rate": 0.0006124132397191575, + "loss": 1.4436, + "theoretical_loss": 3.758020283125318, + "tokens_seen": 799145984 + }, + { + "epoch": 0.24, + "learning_rate": 0.0006123971915747242, + "loss": 1.4862, + "theoretical_loss": 3.757989959687496, + "tokens_seen": 799211520 + }, + { + "epoch": 0.24, + "learning_rate": 0.0006123811434302909, + "loss": 1.5644, + "theoretical_loss": 3.7579596394322863, + "tokens_seen": 799277056 + }, + { + "epoch": 0.24, + "learning_rate": 0.0006123650952858576, + "loss": 1.4834, + "theoretical_loss": 3.757929322359095, + "tokens_seen": 799342592 + }, + { + "epoch": 0.24, + "learning_rate": 0.0006123490471414243, + "loss": 1.5104, + "theoretical_loss": 3.7578990084673256, + "tokens_seen": 799408128 + }, + { + "epoch": 0.24, + "learning_rate": 0.000612332998996991, + "loss": 1.5247, + "theoretical_loss": 3.7578686977563858, + "tokens_seen": 799473664 + }, + { + "epoch": 0.24, + "objective/train/docs_used": 295724, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.3935701847076416, + "objective/train/theoretical_loss": 3.75783839022568, + "objective/train/tokens_used": 819999200, + "theoretical_loss": 3.75783839022568, + "tokens_seen": 799539200 + }, + { + "epoch": 0.24, + "learning_rate": 0.0006123169508525577, + "loss": 1.4903, + "theoretical_loss": 3.75783839022568, + "tokens_seen": 799539200 + }, + { + "epoch": 0.24, + "learning_rate": 0.0006123009027081244, + "loss": 1.5325, + "theoretical_loss": 3.757808085874613, + "tokens_seen": 799604736 + }, + { + "epoch": 0.24, + "learning_rate": 0.0006122848545636911, + "loss": 1.4712, + "theoretical_loss": 3.7577777847025926, + "tokens_seen": 799670272 + }, + { + "epoch": 0.24, + "learning_rate": 0.0006122688064192578, + "loss": 1.4685, + "theoretical_loss": 3.757747486709023, + "tokens_seen": 799735808 + }, + { + "epoch": 0.24, + "learning_rate": 0.0006122527582748245, + "loss": 1.3986, + "theoretical_loss": 3.7577171918933114, + "tokens_seen": 799801344 + }, + { + "epoch": 0.24, + "learning_rate": 0.0006122367101303912, + "loss": 1.4318, + "theoretical_loss": 3.757686900254864, + "tokens_seen": 799866880 + }, + { + "epoch": 0.24, + "learning_rate": 0.0006122206619859579, + "loss": 1.5865, + "theoretical_loss": 3.757656611793087, + "tokens_seen": 799932416 + }, + { + "epoch": 0.24, + "learning_rate": 0.0006122046138415246, + "loss": 1.5005, + "theoretical_loss": 3.7576263265073875, + "tokens_seen": 799997952 + }, + { + "epoch": 0.24, + "learning_rate": 0.0006121885656970913, + "loss": 1.3964, + "theoretical_loss": 3.7575960443971717, + "tokens_seen": 800063488 + }, + { + "epoch": 0.24, + "learning_rate": 0.000612172517552658, + "loss": 1.4858, + "theoretical_loss": 3.7575657654618473, + "tokens_seen": 800129024 + }, + { + "epoch": 0.24, + "learning_rate": 0.0006121564694082247, + "loss": 1.4346, + "theoretical_loss": 3.75753548970082, + "tokens_seen": 800194560 + }, + { + "epoch": 0.24, + "learning_rate": 0.0006121404212637914, + "loss": 1.4918, + "theoretical_loss": 3.757505217113499, + "tokens_seen": 800260096 + }, + { + "epoch": 0.24, + "learning_rate": 0.0006121243731193581, + "loss": 1.3974, + "theoretical_loss": 3.757474947699291, + "tokens_seen": 800325632 + }, + { + "epoch": 0.24, + "learning_rate": 0.0006121083249749248, + "loss": 1.4837, + "theoretical_loss": 3.757444681457603, + "tokens_seen": 800391168 + }, + { + "epoch": 0.24, + "learning_rate": 0.0006120922768304915, + "loss": 1.4517, + "theoretical_loss": 3.7574144183878433, + "tokens_seen": 800456704 + }, + { + "epoch": 0.24, + "learning_rate": 0.0006120762286860582, + "loss": 1.4462, + "theoretical_loss": 3.75738415848942, + "tokens_seen": 800522240 + }, + { + "epoch": 0.24, + "learning_rate": 0.0006120601805416249, + "loss": 1.4173, + "theoretical_loss": 3.7573539017617406, + "tokens_seen": 800587776 + }, + { + "epoch": 0.24, + "learning_rate": 0.0006120441323971916, + "loss": 1.4897, + "theoretical_loss": 3.7573236482042134, + "tokens_seen": 800653312 + }, + { + "epoch": 0.24, + "learning_rate": 0.0006120280842527583, + "loss": 1.4676, + "theoretical_loss": 3.7572933978162477, + "tokens_seen": 800718848 + }, + { + "epoch": 0.24, + "learning_rate": 0.000612012036108325, + "loss": 1.4738, + "theoretical_loss": 3.757263150597251, + "tokens_seen": 800784384 + }, + { + "epoch": 0.24, + "learning_rate": 0.0006119959879638917, + "loss": 1.5588, + "theoretical_loss": 3.7572329065466326, + "tokens_seen": 800849920 + }, + { + "epoch": 0.24, + "learning_rate": 0.0006119799398194584, + "loss": 1.4199, + "theoretical_loss": 3.7572026656638013, + "tokens_seen": 800915456 + }, + { + "epoch": 0.24, + "learning_rate": 0.0006119638916750251, + "loss": 1.4108, + "theoretical_loss": 3.757172427948166, + "tokens_seen": 800980992 + }, + { + "epoch": 0.24, + "learning_rate": 0.0006119478435305918, + "loss": 1.5058, + "theoretical_loss": 3.7571421933991362, + "tokens_seen": 801046528 + }, + { + "epoch": 0.24, + "learning_rate": 0.0006119317953861585, + "loss": 1.4757, + "theoretical_loss": 3.7571119620161206, + "tokens_seen": 801112064 + }, + { + "epoch": 0.24, + "learning_rate": 0.0006119157472417252, + "loss": 1.512, + "theoretical_loss": 3.7570817337985294, + "tokens_seen": 801177600 + }, + { + "epoch": 0.24, + "learning_rate": 0.0006118996990972919, + "loss": 1.4938, + "theoretical_loss": 3.757051508745772, + "tokens_seen": 801243136 + }, + { + "epoch": 0.24, + "learning_rate": 0.0006118836509528586, + "loss": 1.5339, + "theoretical_loss": 3.757021286857258, + "tokens_seen": 801308672 + }, + { + "epoch": 0.24, + "learning_rate": 0.0006118676028084253, + "loss": 1.6401, + "theoretical_loss": 3.7569910681323977, + "tokens_seen": 801374208 + }, + { + "epoch": 0.24, + "learning_rate": 0.000611851554663992, + "loss": 1.5131, + "theoretical_loss": 3.756960852570601, + "tokens_seen": 801439744 + }, + { + "epoch": 0.24, + "learning_rate": 0.0006118355065195587, + "loss": 1.5471, + "theoretical_loss": 3.7569306401712783, + "tokens_seen": 801505280 + }, + { + "epoch": 0.24, + "learning_rate": 0.0006118194583751254, + "loss": 1.572, + "theoretical_loss": 3.7569004309338405, + "tokens_seen": 801570816 + }, + { + "epoch": 0.24, + "learning_rate": 0.0006118034102306921, + "loss": 1.5651, + "theoretical_loss": 3.756870224857698, + "tokens_seen": 801636352 + }, + { + "epoch": 0.24, + "learning_rate": 0.0006117873620862588, + "loss": 1.5299, + "theoretical_loss": 3.756840021942261, + "tokens_seen": 801701888 + }, + { + "epoch": 0.24, + "learning_rate": 0.0006117713139418255, + "loss": 1.4716, + "theoretical_loss": 3.756809822186941, + "tokens_seen": 801767424 + }, + { + "epoch": 0.24, + "learning_rate": 0.0006117552657973922, + "loss": 1.5145, + "theoretical_loss": 3.7567796255911485, + "tokens_seen": 801832960 + }, + { + "epoch": 0.24, + "learning_rate": 0.000611739217652959, + "loss": 1.4931, + "theoretical_loss": 3.7567494321542956, + "tokens_seen": 801898496 + }, + { + "epoch": 0.24, + "learning_rate": 0.0006117231695085256, + "loss": 1.5049, + "theoretical_loss": 3.756719241875793, + "tokens_seen": 801964032 + }, + { + "epoch": 0.24, + "learning_rate": 0.0006117071213640923, + "loss": 1.5183, + "theoretical_loss": 3.7566890547550527, + "tokens_seen": 802029568 + }, + { + "epoch": 0.24, + "learning_rate": 0.000611691073219659, + "loss": 1.5476, + "theoretical_loss": 3.7566588707914867, + "tokens_seen": 802095104 + }, + { + "epoch": 0.24, + "learning_rate": 0.0006116750250752258, + "loss": 1.4034, + "theoretical_loss": 3.756628689984506, + "tokens_seen": 802160640 + }, + { + "epoch": 0.24, + "learning_rate": 0.0006116589769307924, + "loss": 1.3599, + "theoretical_loss": 3.756598512333523, + "tokens_seen": 802226176 + }, + { + "epoch": 0.24, + "learning_rate": 0.0006116429287863592, + "loss": 1.4793, + "theoretical_loss": 3.75656833783795, + "tokens_seen": 802291712 + }, + { + "epoch": 0.24, + "learning_rate": 0.0006116268806419258, + "loss": 1.4264, + "theoretical_loss": 3.7565381664972, + "tokens_seen": 802357248 + }, + { + "epoch": 0.24, + "learning_rate": 0.0006116108324974926, + "loss": 1.4665, + "theoretical_loss": 3.7565079983106844, + "tokens_seen": 802422784 + }, + { + "epoch": 0.24, + "learning_rate": 0.0006115947843530592, + "loss": 1.3744, + "theoretical_loss": 3.7564778332778164, + "tokens_seen": 802488320 + }, + { + "epoch": 0.24, + "learning_rate": 0.000611578736208626, + "loss": 1.4266, + "theoretical_loss": 3.756447671398009, + "tokens_seen": 802553856 + }, + { + "epoch": 0.24, + "learning_rate": 0.0006115626880641926, + "loss": 1.4264, + "theoretical_loss": 3.7564175126706743, + "tokens_seen": 802619392 + }, + { + "epoch": 0.24, + "learning_rate": 0.0006115466399197594, + "loss": 1.4526, + "theoretical_loss": 3.7563873570952264, + "tokens_seen": 802684928 + }, + { + "epoch": 0.24, + "learning_rate": 0.000611530591775326, + "loss": 1.3925, + "theoretical_loss": 3.7563572046710787, + "tokens_seen": 802750464 + }, + { + "epoch": 0.24, + "objective/train/docs_used": 296852, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.439468502998352, + "objective/train/theoretical_loss": 3.7563270553976436, + "objective/train/tokens_used": 823276000, + "theoretical_loss": 3.7563270553976436, + "tokens_seen": 802816000 + }, + { + "epoch": 0.24, + "learning_rate": 0.0006115145436308928, + "loss": 1.3846, + "theoretical_loss": 3.7563270553976436, + "tokens_seen": 802816000 + }, + { + "epoch": 0.24, + "learning_rate": 0.0006114984954864594, + "loss": 1.3906, + "theoretical_loss": 3.7562969092743357, + "tokens_seen": 802881536 + }, + { + "epoch": 0.24, + "learning_rate": 0.0006114824473420262, + "loss": 1.4174, + "theoretical_loss": 3.7562667663005684, + "tokens_seen": 802947072 + }, + { + "epoch": 0.24, + "learning_rate": 0.0006114663991975928, + "loss": 1.503, + "theoretical_loss": 3.7562366264757556, + "tokens_seen": 803012608 + }, + { + "epoch": 0.24, + "learning_rate": 0.0006114503510531596, + "loss": 1.5044, + "theoretical_loss": 3.756206489799311, + "tokens_seen": 803078144 + }, + { + "epoch": 0.24, + "learning_rate": 0.0006114343029087262, + "loss": 1.379, + "theoretical_loss": 3.7561763562706503, + "tokens_seen": 803143680 + }, + { + "epoch": 0.24, + "learning_rate": 0.000611418254764293, + "loss": 1.624, + "theoretical_loss": 3.7561462258891862, + "tokens_seen": 803209216 + }, + { + "epoch": 0.24, + "learning_rate": 0.0006114022066198596, + "loss": 1.4981, + "theoretical_loss": 3.756116098654334, + "tokens_seen": 803274752 + }, + { + "epoch": 0.24, + "learning_rate": 0.0006113861584754263, + "loss": 1.5246, + "theoretical_loss": 3.7560859745655084, + "tokens_seen": 803340288 + }, + { + "epoch": 0.24, + "learning_rate": 0.000611370110330993, + "loss": 1.4882, + "theoretical_loss": 3.756055853622124, + "tokens_seen": 803405824 + }, + { + "epoch": 0.24, + "learning_rate": 0.0006113540621865597, + "loss": 1.477, + "theoretical_loss": 3.756025735823596, + "tokens_seen": 803471360 + }, + { + "epoch": 0.24, + "learning_rate": 0.0006113380140421264, + "loss": 1.4178, + "theoretical_loss": 3.7559956211693404, + "tokens_seen": 803536896 + }, + { + "epoch": 0.24, + "learning_rate": 0.0006113219658976931, + "loss": 1.5439, + "theoretical_loss": 3.7559655096587705, + "tokens_seen": 803602432 + }, + { + "epoch": 0.24, + "learning_rate": 0.0006113059177532598, + "loss": 1.466, + "theoretical_loss": 3.7559354012913038, + "tokens_seen": 803667968 + }, + { + "epoch": 0.24, + "learning_rate": 0.0006112898696088265, + "loss": 1.4557, + "theoretical_loss": 3.755905296066355, + "tokens_seen": 803733504 + }, + { + "epoch": 0.24, + "learning_rate": 0.0006112738214643932, + "loss": 1.5077, + "theoretical_loss": 3.75587519398334, + "tokens_seen": 803799040 + }, + { + "epoch": 0.24, + "learning_rate": 0.0006112577733199599, + "loss": 1.5042, + "theoretical_loss": 3.755845095041675, + "tokens_seen": 803864576 + }, + { + "epoch": 0.24, + "learning_rate": 0.0006112417251755266, + "loss": 1.5697, + "theoretical_loss": 3.755814999240776, + "tokens_seen": 803930112 + }, + { + "epoch": 0.24, + "learning_rate": 0.0006112256770310933, + "loss": 1.5976, + "theoretical_loss": 3.7557849065800593, + "tokens_seen": 803995648 + }, + { + "epoch": 0.24, + "learning_rate": 0.00061120962888666, + "loss": 1.5728, + "theoretical_loss": 3.755754817058941, + "tokens_seen": 804061184 + }, + { + "epoch": 0.24, + "learning_rate": 0.0006111935807422267, + "loss": 1.5765, + "theoretical_loss": 3.755724730676838, + "tokens_seen": 804126720 + }, + { + "epoch": 0.24, + "learning_rate": 0.0006111775325977934, + "loss": 1.4441, + "theoretical_loss": 3.7556946474331667, + "tokens_seen": 804192256 + }, + { + "epoch": 0.24, + "learning_rate": 0.0006111614844533601, + "loss": 1.4491, + "theoretical_loss": 3.7556645673273445, + "tokens_seen": 804257792 + }, + { + "epoch": 0.24, + "learning_rate": 0.0006111454363089268, + "loss": 1.5513, + "theoretical_loss": 3.755634490358789, + "tokens_seen": 804323328 + }, + { + "epoch": 0.24, + "learning_rate": 0.0006111293881644935, + "loss": 1.5301, + "theoretical_loss": 3.7556044165269156, + "tokens_seen": 804388864 + }, + { + "epoch": 0.24, + "learning_rate": 0.0006111133400200602, + "loss": 1.4962, + "theoretical_loss": 3.7555743458311435, + "tokens_seen": 804454400 + }, + { + "epoch": 0.24, + "learning_rate": 0.0006110972918756269, + "loss": 1.5929, + "theoretical_loss": 3.7555442782708885, + "tokens_seen": 804519936 + }, + { + "epoch": 0.24, + "learning_rate": 0.0006110812437311936, + "loss": 1.4205, + "theoretical_loss": 3.75551421384557, + "tokens_seen": 804585472 + }, + { + "epoch": 0.24, + "learning_rate": 0.0006110651955867603, + "loss": 1.3433, + "theoretical_loss": 3.755484152554605, + "tokens_seen": 804651008 + }, + { + "epoch": 0.24, + "learning_rate": 0.000611049147442327, + "loss": 1.3927, + "theoretical_loss": 3.755454094397411, + "tokens_seen": 804716544 + }, + { + "epoch": 0.24, + "learning_rate": 0.0006110330992978937, + "loss": 1.4611, + "theoretical_loss": 3.7554240393734073, + "tokens_seen": 804782080 + }, + { + "epoch": 0.24, + "learning_rate": 0.0006110170511534604, + "loss": 1.4954, + "theoretical_loss": 3.755393987482011, + "tokens_seen": 804847616 + }, + { + "epoch": 0.24, + "learning_rate": 0.0006110010030090271, + "loss": 1.4917, + "theoretical_loss": 3.755363938722642, + "tokens_seen": 804913152 + }, + { + "epoch": 0.24, + "learning_rate": 0.0006109849548645938, + "loss": 1.4851, + "theoretical_loss": 3.7553338930947167, + "tokens_seen": 804978688 + }, + { + "epoch": 0.24, + "learning_rate": 0.0006109689067201605, + "loss": 1.4467, + "theoretical_loss": 3.7553038505976564, + "tokens_seen": 805044224 + }, + { + "epoch": 0.24, + "learning_rate": 0.0006109528585757272, + "loss": 1.3892, + "theoretical_loss": 3.7552738112308783, + "tokens_seen": 805109760 + }, + { + "epoch": 0.24, + "learning_rate": 0.0006109368104312939, + "loss": 1.5459, + "theoretical_loss": 3.7552437749938017, + "tokens_seen": 805175296 + }, + { + "epoch": 0.24, + "learning_rate": 0.0006109207622868606, + "loss": 1.5646, + "theoretical_loss": 3.755213741885847, + "tokens_seen": 805240832 + }, + { + "epoch": 0.24, + "learning_rate": 0.0006109047141424273, + "loss": 1.5281, + "theoretical_loss": 3.755183711906432, + "tokens_seen": 805306368 + }, + { + "epoch": 0.24, + "learning_rate": 0.000610888665997994, + "loss": 1.3792, + "theoretical_loss": 3.755153685054977, + "tokens_seen": 805371904 + }, + { + "epoch": 0.24, + "learning_rate": 0.0006108726178535607, + "loss": 1.4938, + "theoretical_loss": 3.755123661330902, + "tokens_seen": 805437440 + }, + { + "epoch": 0.24, + "learning_rate": 0.0006108565697091274, + "loss": 1.5624, + "theoretical_loss": 3.755093640733626, + "tokens_seen": 805502976 + }, + { + "epoch": 0.24, + "learning_rate": 0.0006108405215646941, + "loss": 1.4788, + "theoretical_loss": 3.75506362326257, + "tokens_seen": 805568512 + }, + { + "epoch": 0.24, + "learning_rate": 0.0006108244734202608, + "loss": 1.5741, + "theoretical_loss": 3.755033608917153, + "tokens_seen": 805634048 + }, + { + "epoch": 0.24, + "learning_rate": 0.0006108084252758275, + "loss": 1.4526, + "theoretical_loss": 3.7550035976967964, + "tokens_seen": 805699584 + }, + { + "epoch": 0.24, + "learning_rate": 0.0006107923771313942, + "loss": 1.4364, + "theoretical_loss": 3.7549735896009206, + "tokens_seen": 805765120 + }, + { + "epoch": 0.24, + "learning_rate": 0.0006107763289869609, + "loss": 1.5181, + "theoretical_loss": 3.754943584628945, + "tokens_seen": 805830656 + }, + { + "epoch": 0.24, + "learning_rate": 0.0006107602808425276, + "loss": 1.5623, + "theoretical_loss": 3.7549135827802917, + "tokens_seen": 805896192 + }, + { + "epoch": 0.24, + "learning_rate": 0.0006107442326980943, + "loss": 1.4442, + "theoretical_loss": 3.7548835840543813, + "tokens_seen": 805961728 + }, + { + "epoch": 0.24, + "learning_rate": 0.000610728184553661, + "loss": 1.3875, + "theoretical_loss": 3.7548535884506347, + "tokens_seen": 806027264 + }, + { + "epoch": 0.24, + "objective/train/docs_used": 297892, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.494751214981079, + "objective/train/theoretical_loss": 3.754823595968473, + "objective/train/tokens_used": 826552800, + "theoretical_loss": 3.754823595968473, + "tokens_seen": 806092800 + }, + { + "epoch": 0.24, + "learning_rate": 0.0006107121364092277, + "loss": 1.4098, + "theoretical_loss": 3.754823595968473, + "tokens_seen": 806092800 + }, + { + "epoch": 0.24, + "learning_rate": 0.0006106960882647944, + "loss": 1.516, + "theoretical_loss": 3.7547936066073175, + "tokens_seen": 806158336 + }, + { + "epoch": 0.24, + "learning_rate": 0.0006106800401203611, + "loss": 1.4971, + "theoretical_loss": 3.7547636203665906, + "tokens_seen": 806223872 + }, + { + "epoch": 0.24, + "learning_rate": 0.0006106639919759278, + "loss": 1.4252, + "theoretical_loss": 3.7547336372457134, + "tokens_seen": 806289408 + }, + { + "epoch": 0.24, + "learning_rate": 0.0006106479438314945, + "loss": 1.4208, + "theoretical_loss": 3.7547036572441073, + "tokens_seen": 806354944 + }, + { + "epoch": 0.24, + "learning_rate": 0.0006106318956870612, + "loss": 1.6, + "theoretical_loss": 3.7546736803611953, + "tokens_seen": 806420480 + }, + { + "epoch": 0.24, + "learning_rate": 0.0006106158475426279, + "loss": 1.6105, + "theoretical_loss": 3.754643706596399, + "tokens_seen": 806486016 + }, + { + "epoch": 0.24, + "learning_rate": 0.0006105997993981946, + "loss": 1.5062, + "theoretical_loss": 3.7546137359491407, + "tokens_seen": 806551552 + }, + { + "epoch": 0.24, + "learning_rate": 0.0006105837512537613, + "loss": 1.3542, + "theoretical_loss": 3.7545837684188426, + "tokens_seen": 806617088 + }, + { + "epoch": 0.24, + "learning_rate": 0.000610567703109328, + "loss": 1.3692, + "theoretical_loss": 3.7545538040049284, + "tokens_seen": 806682624 + }, + { + "epoch": 0.24, + "learning_rate": 0.0006105516549648947, + "loss": 1.4566, + "theoretical_loss": 3.7545238427068197, + "tokens_seen": 806748160 + }, + { + "epoch": 0.24, + "learning_rate": 0.0006105356068204614, + "loss": 1.3897, + "theoretical_loss": 3.7544938845239404, + "tokens_seen": 806813696 + }, + { + "epoch": 0.24, + "learning_rate": 0.0006105195586760281, + "loss": 1.4738, + "theoretical_loss": 3.7544639294557127, + "tokens_seen": 806879232 + }, + { + "epoch": 0.24, + "learning_rate": 0.0006105035105315948, + "loss": 1.4881, + "theoretical_loss": 3.7544339775015603, + "tokens_seen": 806944768 + }, + { + "epoch": 0.24, + "learning_rate": 0.0006104874623871615, + "loss": 1.5049, + "theoretical_loss": 3.7544040286609066, + "tokens_seen": 807010304 + }, + { + "epoch": 0.24, + "learning_rate": 0.0006104714142427282, + "loss": 1.3485, + "theoretical_loss": 3.7543740829331744, + "tokens_seen": 807075840 + }, + { + "epoch": 0.24, + "learning_rate": 0.0006104553660982949, + "loss": 1.5087, + "theoretical_loss": 3.754344140317789, + "tokens_seen": 807141376 + }, + { + "epoch": 0.24, + "learning_rate": 0.0006104393179538616, + "loss": 1.4234, + "theoretical_loss": 3.754314200814173, + "tokens_seen": 807206912 + }, + { + "epoch": 0.24, + "learning_rate": 0.0006104232698094283, + "loss": 1.4513, + "theoretical_loss": 3.754284264421751, + "tokens_seen": 807272448 + }, + { + "epoch": 0.24, + "learning_rate": 0.000610407221664995, + "loss": 1.4611, + "theoretical_loss": 3.754254331139946, + "tokens_seen": 807337984 + }, + { + "epoch": 0.24, + "learning_rate": 0.0006103911735205617, + "loss": 1.3825, + "theoretical_loss": 3.7542244009681847, + "tokens_seen": 807403520 + }, + { + "epoch": 0.24, + "learning_rate": 0.0006103751253761284, + "loss": 1.4557, + "theoretical_loss": 3.7541944739058892, + "tokens_seen": 807469056 + }, + { + "epoch": 0.24, + "learning_rate": 0.0006103590772316951, + "loss": 1.4366, + "theoretical_loss": 3.754164549952485, + "tokens_seen": 807534592 + }, + { + "epoch": 0.24, + "learning_rate": 0.0006103430290872618, + "loss": 1.3824, + "theoretical_loss": 3.754134629107397, + "tokens_seen": 807600128 + }, + { + "epoch": 0.24, + "learning_rate": 0.0006103269809428285, + "loss": 1.4126, + "theoretical_loss": 3.7541047113700503, + "tokens_seen": 807665664 + }, + { + "epoch": 0.24, + "learning_rate": 0.0006103109327983953, + "loss": 1.532, + "theoretical_loss": 3.7540747967398698, + "tokens_seen": 807731200 + }, + { + "epoch": 0.24, + "learning_rate": 0.0006102948846539619, + "loss": 1.3846, + "theoretical_loss": 3.75404488521628, + "tokens_seen": 807796736 + }, + { + "epoch": 0.24, + "learning_rate": 0.0006102788365095287, + "loss": 1.4207, + "theoretical_loss": 3.7540149767987074, + "tokens_seen": 807862272 + }, + { + "epoch": 0.24, + "learning_rate": 0.0006102627883650953, + "loss": 1.4627, + "theoretical_loss": 3.753985071486577, + "tokens_seen": 807927808 + }, + { + "epoch": 0.24, + "learning_rate": 0.000610246740220662, + "loss": 1.4582, + "theoretical_loss": 3.7539551692793145, + "tokens_seen": 807993344 + }, + { + "epoch": 0.24, + "learning_rate": 0.0006102306920762287, + "loss": 1.5387, + "theoretical_loss": 3.753925270176346, + "tokens_seen": 808058880 + }, + { + "epoch": 0.24, + "learning_rate": 0.0006102146439317955, + "loss": 1.4948, + "theoretical_loss": 3.7538953741770973, + "tokens_seen": 808124416 + }, + { + "epoch": 0.24, + "learning_rate": 0.0006101985957873622, + "loss": 1.542, + "theoretical_loss": 3.753865481280995, + "tokens_seen": 808189952 + }, + { + "epoch": 0.24, + "learning_rate": 0.0006101825476429289, + "loss": 1.4222, + "theoretical_loss": 3.753835591487464, + "tokens_seen": 808255488 + }, + { + "epoch": 0.24, + "learning_rate": 0.0006101664994984956, + "loss": 1.4572, + "theoretical_loss": 3.7538057047959326, + "tokens_seen": 808321024 + }, + { + "epoch": 0.24, + "learning_rate": 0.0006101504513540623, + "loss": 1.5363, + "theoretical_loss": 3.753775821205826, + "tokens_seen": 808386560 + }, + { + "epoch": 0.24, + "learning_rate": 0.000610134403209629, + "loss": 1.48, + "theoretical_loss": 3.7537459407165716, + "tokens_seen": 808452096 + }, + { + "epoch": 0.25, + "learning_rate": 0.0006101183550651957, + "loss": 1.4046, + "theoretical_loss": 3.7537160633275963, + "tokens_seen": 808517632 + }, + { + "epoch": 0.25, + "learning_rate": 0.0006101023069207623, + "loss": 1.5231, + "theoretical_loss": 3.7536861890383273, + "tokens_seen": 808583168 + }, + { + "epoch": 0.25, + "learning_rate": 0.0006100862587763291, + "loss": 1.5656, + "theoretical_loss": 3.7536563178481916, + "tokens_seen": 808648704 + }, + { + "epoch": 0.25, + "learning_rate": 0.0006100702106318957, + "loss": 1.4317, + "theoretical_loss": 3.753626449756616, + "tokens_seen": 808714240 + }, + { + "epoch": 0.25, + "learning_rate": 0.0006100541624874625, + "loss": 1.4667, + "theoretical_loss": 3.753596584763029, + "tokens_seen": 808779776 + }, + { + "epoch": 0.25, + "learning_rate": 0.0006100381143430291, + "loss": 1.4162, + "theoretical_loss": 3.753566722866858, + "tokens_seen": 808845312 + }, + { + "epoch": 0.25, + "learning_rate": 0.0006100220661985959, + "loss": 1.5584, + "theoretical_loss": 3.753536864067531, + "tokens_seen": 808910848 + }, + { + "epoch": 0.25, + "learning_rate": 0.0006100060180541625, + "loss": 1.6185, + "theoretical_loss": 3.7535070083644753, + "tokens_seen": 808976384 + }, + { + "epoch": 0.25, + "learning_rate": 0.0006099899699097292, + "loss": 1.4584, + "theoretical_loss": 3.7534771557571194, + "tokens_seen": 809041920 + }, + { + "epoch": 0.25, + "learning_rate": 0.0006099739217652959, + "loss": 1.4052, + "theoretical_loss": 3.7534473062448916, + "tokens_seen": 809107456 + }, + { + "epoch": 0.25, + "learning_rate": 0.0006099578736208626, + "loss": 1.4594, + "theoretical_loss": 3.7534174598272205, + "tokens_seen": 809172992 + }, + { + "epoch": 0.25, + "learning_rate": 0.0006099418254764293, + "loss": 1.486, + "theoretical_loss": 3.7533876165035345, + "tokens_seen": 809238528 + }, + { + "epoch": 0.25, + "learning_rate": 0.000609925777331996, + "loss": 1.5254, + "theoretical_loss": 3.7533577762732624, + "tokens_seen": 809304064 + }, + { + "epoch": 0.25, + "objective/train/docs_used": 298985, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.5134739875793457, + "objective/train/theoretical_loss": 3.753327939135833, + "objective/train/tokens_used": 829829600, + "theoretical_loss": 3.753327939135833, + "tokens_seen": 809369600 + }, + { + "epoch": 0.25, + "learning_rate": 0.0006099097291875627, + "loss": 1.5421, + "theoretical_loss": 3.753327939135833, + "tokens_seen": 809369600 + }, + { + "epoch": 0.25, + "learning_rate": 0.0006098936810431294, + "loss": 1.5266, + "theoretical_loss": 3.7532981050906757, + "tokens_seen": 809435136 + }, + { + "epoch": 0.25, + "learning_rate": 0.0006098776328986961, + "loss": 1.5167, + "theoretical_loss": 3.753268274137219, + "tokens_seen": 809500672 + }, + { + "epoch": 0.25, + "learning_rate": 0.0006098615847542628, + "loss": 1.4599, + "theoretical_loss": 3.753238446274893, + "tokens_seen": 809566208 + }, + { + "epoch": 0.25, + "learning_rate": 0.0006098455366098295, + "loss": 1.4234, + "theoretical_loss": 3.7532086215031266, + "tokens_seen": 809631744 + }, + { + "epoch": 0.25, + "learning_rate": 0.0006098294884653962, + "loss": 1.5258, + "theoretical_loss": 3.7531787998213497, + "tokens_seen": 809697280 + }, + { + "epoch": 0.25, + "learning_rate": 0.0006098134403209629, + "loss": 1.4583, + "theoretical_loss": 3.7531489812289927, + "tokens_seen": 809762816 + }, + { + "epoch": 0.25, + "learning_rate": 0.0006097973921765296, + "loss": 1.3904, + "theoretical_loss": 3.753119165725484, + "tokens_seen": 809828352 + }, + { + "epoch": 0.25, + "learning_rate": 0.0006097813440320963, + "loss": 1.4107, + "theoretical_loss": 3.753089353310255, + "tokens_seen": 809893888 + }, + { + "epoch": 0.25, + "learning_rate": 0.000609765295887663, + "loss": 1.4178, + "theoretical_loss": 3.7530595439827357, + "tokens_seen": 809959424 + }, + { + "epoch": 0.25, + "learning_rate": 0.0006097492477432297, + "loss": 1.4332, + "theoretical_loss": 3.7530297377423567, + "tokens_seen": 810024960 + }, + { + "epoch": 0.25, + "learning_rate": 0.0006097331995987964, + "loss": 1.387, + "theoretical_loss": 3.752999934588548, + "tokens_seen": 810090496 + }, + { + "epoch": 0.25, + "learning_rate": 0.0006097171514543631, + "loss": 1.4158, + "theoretical_loss": 3.7529701345207402, + "tokens_seen": 810156032 + }, + { + "epoch": 0.25, + "learning_rate": 0.0006097011033099298, + "loss": 1.5196, + "theoretical_loss": 3.752940337538365, + "tokens_seen": 810221568 + }, + { + "epoch": 0.25, + "learning_rate": 0.0006096850551654965, + "loss": 1.4122, + "theoretical_loss": 3.752910543640853, + "tokens_seen": 810287104 + }, + { + "epoch": 0.25, + "learning_rate": 0.0006096690070210632, + "loss": 1.3854, + "theoretical_loss": 3.752880752827635, + "tokens_seen": 810352640 + }, + { + "epoch": 0.25, + "learning_rate": 0.0006096529588766299, + "loss": 1.5018, + "theoretical_loss": 3.752850965098143, + "tokens_seen": 810418176 + }, + { + "epoch": 0.25, + "learning_rate": 0.0006096369107321966, + "loss": 1.4368, + "theoretical_loss": 3.752821180451808, + "tokens_seen": 810483712 + }, + { + "epoch": 0.25, + "learning_rate": 0.0006096208625877633, + "loss": 1.4684, + "theoretical_loss": 3.7527913988880615, + "tokens_seen": 810549248 + }, + { + "epoch": 0.25, + "learning_rate": 0.00060960481444333, + "loss": 1.5708, + "theoretical_loss": 3.7527616204063357, + "tokens_seen": 810614784 + }, + { + "epoch": 0.25, + "learning_rate": 0.0006095887662988967, + "loss": 1.4424, + "theoretical_loss": 3.7527318450060623, + "tokens_seen": 810680320 + }, + { + "epoch": 0.25, + "learning_rate": 0.0006095727181544634, + "loss": 1.5946, + "theoretical_loss": 3.752702072686673, + "tokens_seen": 810745856 + }, + { + "epoch": 0.25, + "learning_rate": 0.0006095566700100301, + "loss": 1.3912, + "theoretical_loss": 3.7526723034476004, + "tokens_seen": 810811392 + }, + { + "epoch": 0.25, + "learning_rate": 0.0006095406218655968, + "loss": 1.5026, + "theoretical_loss": 3.752642537288277, + "tokens_seen": 810876928 + }, + { + "epoch": 0.25, + "learning_rate": 0.0006095245737211635, + "loss": 1.5035, + "theoretical_loss": 3.752612774208135, + "tokens_seen": 810942464 + }, + { + "epoch": 0.25, + "learning_rate": 0.0006095085255767302, + "loss": 1.4057, + "theoretical_loss": 3.752583014206608, + "tokens_seen": 811008000 + }, + { + "epoch": 0.25, + "learning_rate": 0.0006094924774322969, + "loss": 1.5172, + "theoretical_loss": 3.752553257283127, + "tokens_seen": 811073536 + }, + { + "epoch": 0.25, + "learning_rate": 0.0006094764292878636, + "loss": 1.5249, + "theoretical_loss": 3.7525235034371267, + "tokens_seen": 811139072 + }, + { + "epoch": 0.25, + "learning_rate": 0.0006094603811434303, + "loss": 1.4584, + "theoretical_loss": 3.7524937526680393, + "tokens_seen": 811204608 + }, + { + "epoch": 0.25, + "learning_rate": 0.000609444332998997, + "loss": 1.4282, + "theoretical_loss": 3.752464004975298, + "tokens_seen": 811270144 + }, + { + "epoch": 0.25, + "learning_rate": 0.0006094282848545637, + "loss": 1.4703, + "theoretical_loss": 3.7524342603583367, + "tokens_seen": 811335680 + }, + { + "epoch": 0.25, + "learning_rate": 0.0006094122367101304, + "loss": 1.3646, + "theoretical_loss": 3.752404518816588, + "tokens_seen": 811401216 + }, + { + "epoch": 0.25, + "learning_rate": 0.0006093961885656971, + "loss": 1.4826, + "theoretical_loss": 3.752374780349488, + "tokens_seen": 811466752 + }, + { + "epoch": 0.25, + "learning_rate": 0.0006093801404212638, + "loss": 1.4457, + "theoretical_loss": 3.752345044956468, + "tokens_seen": 811532288 + }, + { + "epoch": 0.25, + "learning_rate": 0.0006093640922768305, + "loss": 1.3723, + "theoretical_loss": 3.7523153126369633, + "tokens_seen": 811597824 + }, + { + "epoch": 0.25, + "learning_rate": 0.0006093480441323972, + "loss": 1.4803, + "theoretical_loss": 3.7522855833904076, + "tokens_seen": 811663360 + }, + { + "epoch": 0.25, + "learning_rate": 0.0006093319959879639, + "loss": 1.511, + "theoretical_loss": 3.7522558572162352, + "tokens_seen": 811728896 + }, + { + "epoch": 0.25, + "learning_rate": 0.0006093159478435306, + "loss": 1.4522, + "theoretical_loss": 3.752226134113881, + "tokens_seen": 811794432 + }, + { + "epoch": 0.25, + "learning_rate": 0.0006092998996990973, + "loss": 1.4795, + "theoretical_loss": 3.7521964140827793, + "tokens_seen": 811859968 + }, + { + "epoch": 0.25, + "learning_rate": 0.000609283851554664, + "loss": 1.4081, + "theoretical_loss": 3.7521666971223646, + "tokens_seen": 811925504 + }, + { + "epoch": 0.25, + "learning_rate": 0.0006092678034102307, + "loss": 1.448, + "theoretical_loss": 3.752136983232073, + "tokens_seen": 811991040 + }, + { + "epoch": 0.25, + "learning_rate": 0.0006092517552657974, + "loss": 1.534, + "theoretical_loss": 3.7521072724113376, + "tokens_seen": 812056576 + }, + { + "epoch": 0.25, + "learning_rate": 0.0006092357071213641, + "loss": 1.4884, + "theoretical_loss": 3.7520775646595954, + "tokens_seen": 812122112 + }, + { + "epoch": 0.25, + "learning_rate": 0.0006092196589769308, + "loss": 1.3997, + "theoretical_loss": 3.752047859976281, + "tokens_seen": 812187648 + }, + { + "epoch": 0.25, + "learning_rate": 0.0006092036108324975, + "loss": 1.4249, + "theoretical_loss": 3.75201815836083, + "tokens_seen": 812253184 + }, + { + "epoch": 0.25, + "learning_rate": 0.0006091875626880642, + "loss": 1.432, + "theoretical_loss": 3.751988459812678, + "tokens_seen": 812318720 + }, + { + "epoch": 0.25, + "learning_rate": 0.0006091715145436309, + "loss": 1.503, + "theoretical_loss": 3.7519587643312606, + "tokens_seen": 812384256 + }, + { + "epoch": 0.25, + "learning_rate": 0.0006091554663991976, + "loss": 1.5315, + "theoretical_loss": 3.751929071916015, + "tokens_seen": 812449792 + }, + { + "epoch": 0.25, + "learning_rate": 0.0006091394182547643, + "loss": 1.4578, + "theoretical_loss": 3.751899382566376, + "tokens_seen": 812515328 + }, + { + "epoch": 0.25, + "learning_rate": 0.000609123370110331, + "loss": 1.3953, + "theoretical_loss": 3.7518696962817795, + "tokens_seen": 812580864 + }, + { + "epoch": 0.25, + "objective/train/docs_used": 300102, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.4162535667419434, + "objective/train/theoretical_loss": 3.7518400130616634, + "objective/train/tokens_used": 833106400, + "theoretical_loss": 3.7518400130616634, + "tokens_seen": 812646400 + }, + { + "epoch": 0.25, + "learning_rate": 0.0006091073219658977, + "loss": 1.4631, + "theoretical_loss": 3.7518400130616634, + "tokens_seen": 812646400 + }, + { + "epoch": 0.25, + "learning_rate": 0.0006090912738214644, + "loss": 1.5735, + "theoretical_loss": 3.751810332905463, + "tokens_seen": 812711936 + }, + { + "epoch": 0.25, + "learning_rate": 0.0006090752256770311, + "loss": 1.4426, + "theoretical_loss": 3.751780655812616, + "tokens_seen": 812777472 + }, + { + "epoch": 0.25, + "learning_rate": 0.0006090591775325978, + "loss": 1.4898, + "theoretical_loss": 3.7517509817825583, + "tokens_seen": 812843008 + }, + { + "epoch": 0.25, + "learning_rate": 0.0006090431293881645, + "loss": 1.4612, + "theoretical_loss": 3.7517213108147276, + "tokens_seen": 812908544 + }, + { + "epoch": 0.25, + "learning_rate": 0.0006090270812437312, + "loss": 1.4768, + "theoretical_loss": 3.7516916429085603, + "tokens_seen": 812974080 + }, + { + "epoch": 0.25, + "learning_rate": 0.000609011033099298, + "loss": 1.5074, + "theoretical_loss": 3.751661978063495, + "tokens_seen": 813039616 + }, + { + "epoch": 0.25, + "learning_rate": 0.0006089949849548647, + "loss": 1.3634, + "theoretical_loss": 3.751632316278968, + "tokens_seen": 813105152 + }, + { + "epoch": 0.25, + "learning_rate": 0.0006089789368104313, + "loss": 1.4869, + "theoretical_loss": 3.751602657554418, + "tokens_seen": 813170688 + }, + { + "epoch": 0.25, + "learning_rate": 0.0006089628886659981, + "loss": 1.3141, + "theoretical_loss": 3.7515730018892812, + "tokens_seen": 813236224 + }, + { + "epoch": 0.25, + "learning_rate": 0.0006089468405215648, + "loss": 1.4058, + "theoretical_loss": 3.7515433492829966, + "tokens_seen": 813301760 + }, + { + "epoch": 0.25, + "learning_rate": 0.0006089307923771315, + "loss": 1.4817, + "theoretical_loss": 3.751513699735002, + "tokens_seen": 813367296 + }, + { + "epoch": 0.25, + "learning_rate": 0.0006089147442326982, + "loss": 1.3572, + "theoretical_loss": 3.7514840532447353, + "tokens_seen": 813432832 + }, + { + "epoch": 0.25, + "learning_rate": 0.0006088986960882649, + "loss": 1.4424, + "theoretical_loss": 3.7514544098116356, + "tokens_seen": 813498368 + }, + { + "epoch": 0.25, + "learning_rate": 0.0006088826479438316, + "loss": 1.5379, + "theoretical_loss": 3.7514247694351406, + "tokens_seen": 813563904 + }, + { + "epoch": 0.25, + "learning_rate": 0.0006088665997993983, + "loss": 1.4825, + "theoretical_loss": 3.7513951321146894, + "tokens_seen": 813629440 + }, + { + "epoch": 0.25, + "learning_rate": 0.000608850551654965, + "loss": 1.4868, + "theoretical_loss": 3.7513654978497213, + "tokens_seen": 813694976 + }, + { + "epoch": 0.25, + "learning_rate": 0.0006088345035105317, + "loss": 1.4068, + "theoretical_loss": 3.751335866639674, + "tokens_seen": 813760512 + }, + { + "epoch": 0.25, + "learning_rate": 0.0006088184553660984, + "loss": 1.4573, + "theoretical_loss": 3.751306238483987, + "tokens_seen": 813826048 + }, + { + "epoch": 0.25, + "learning_rate": 0.0006088024072216651, + "loss": 1.5148, + "theoretical_loss": 3.7512766133820996, + "tokens_seen": 813891584 + }, + { + "epoch": 0.25, + "learning_rate": 0.0006087863590772318, + "loss": 1.4204, + "theoretical_loss": 3.751246991333452, + "tokens_seen": 813957120 + }, + { + "epoch": 0.25, + "learning_rate": 0.0006087703109327985, + "loss": 1.4499, + "theoretical_loss": 3.751217372337482, + "tokens_seen": 814022656 + }, + { + "epoch": 0.25, + "learning_rate": 0.0006087542627883652, + "loss": 1.508, + "theoretical_loss": 3.751187756393631, + "tokens_seen": 814088192 + }, + { + "epoch": 0.25, + "learning_rate": 0.0006087382146439319, + "loss": 1.477, + "theoretical_loss": 3.751158143501338, + "tokens_seen": 814153728 + }, + { + "epoch": 0.25, + "learning_rate": 0.0006087221664994986, + "loss": 1.4558, + "theoretical_loss": 3.7511285336600433, + "tokens_seen": 814219264 + }, + { + "epoch": 0.25, + "learning_rate": 0.0006087061183550652, + "loss": 1.4082, + "theoretical_loss": 3.7510989268691866, + "tokens_seen": 814284800 + }, + { + "epoch": 0.25, + "learning_rate": 0.000608690070210632, + "loss": 1.5923, + "theoretical_loss": 3.751069323128209, + "tokens_seen": 814350336 + }, + { + "epoch": 0.25, + "learning_rate": 0.0006086740220661986, + "loss": 1.4861, + "theoretical_loss": 3.7510397224365493, + "tokens_seen": 814415872 + }, + { + "epoch": 0.25, + "learning_rate": 0.0006086579739217654, + "loss": 1.548, + "theoretical_loss": 3.7510101247936496, + "tokens_seen": 814481408 + }, + { + "epoch": 0.25, + "learning_rate": 0.000608641925777332, + "loss": 1.516, + "theoretical_loss": 3.75098053019895, + "tokens_seen": 814546944 + }, + { + "epoch": 0.25, + "learning_rate": 0.0006086258776328988, + "loss": 1.4526, + "theoretical_loss": 3.7509509386518918, + "tokens_seen": 814612480 + }, + { + "epoch": 0.25, + "learning_rate": 0.0006086098294884654, + "loss": 1.3558, + "theoretical_loss": 3.7509213501519154, + "tokens_seen": 814678016 + }, + { + "epoch": 0.25, + "learning_rate": 0.0006085937813440322, + "loss": 1.4158, + "theoretical_loss": 3.750891764698462, + "tokens_seen": 814743552 + }, + { + "epoch": 0.25, + "learning_rate": 0.0006085777331995988, + "loss": 1.4457, + "theoretical_loss": 3.750862182290973, + "tokens_seen": 814809088 + }, + { + "epoch": 0.25, + "learning_rate": 0.0006085616850551655, + "loss": 1.3432, + "theoretical_loss": 3.7508326029288908, + "tokens_seen": 814874624 + }, + { + "epoch": 0.25, + "learning_rate": 0.0006085456369107322, + "loss": 1.5192, + "theoretical_loss": 3.750803026611656, + "tokens_seen": 814940160 + }, + { + "epoch": 0.25, + "learning_rate": 0.0006085295887662989, + "loss": 1.4488, + "theoretical_loss": 3.7507734533387103, + "tokens_seen": 815005696 + }, + { + "epoch": 0.25, + "learning_rate": 0.0006085135406218656, + "loss": 1.5338, + "theoretical_loss": 3.7507438831094957, + "tokens_seen": 815071232 + }, + { + "epoch": 0.25, + "learning_rate": 0.0006084974924774323, + "loss": 1.5477, + "theoretical_loss": 3.750714315923454, + "tokens_seen": 815136768 + }, + { + "epoch": 0.25, + "learning_rate": 0.000608481444332999, + "loss": 1.5179, + "theoretical_loss": 3.7506847517800286, + "tokens_seen": 815202304 + }, + { + "epoch": 0.25, + "learning_rate": 0.0006084653961885657, + "loss": 1.4374, + "theoretical_loss": 3.7506551906786605, + "tokens_seen": 815267840 + }, + { + "epoch": 0.25, + "learning_rate": 0.0006084493480441324, + "loss": 1.4007, + "theoretical_loss": 3.7506256326187923, + "tokens_seen": 815333376 + }, + { + "epoch": 0.25, + "learning_rate": 0.0006084332998996991, + "loss": 1.578, + "theoretical_loss": 3.7505960775998677, + "tokens_seen": 815398912 + }, + { + "epoch": 0.25, + "learning_rate": 0.0006084172517552658, + "loss": 1.3932, + "theoretical_loss": 3.7505665256213288, + "tokens_seen": 815464448 + }, + { + "epoch": 0.25, + "learning_rate": 0.0006084012036108325, + "loss": 1.5619, + "theoretical_loss": 3.7505369766826178, + "tokens_seen": 815529984 + }, + { + "epoch": 0.25, + "learning_rate": 0.0006083851554663992, + "loss": 1.4515, + "theoretical_loss": 3.7505074307831787, + "tokens_seen": 815595520 + }, + { + "epoch": 0.25, + "learning_rate": 0.0006083691073219659, + "loss": 1.3974, + "theoretical_loss": 3.7504778879224547, + "tokens_seen": 815661056 + }, + { + "epoch": 0.25, + "learning_rate": 0.0006083530591775326, + "loss": 1.3695, + "theoretical_loss": 3.750448348099889, + "tokens_seen": 815726592 + }, + { + "epoch": 0.25, + "learning_rate": 0.0006083370110330993, + "loss": 1.4674, + "theoretical_loss": 3.7504188113149244, + "tokens_seen": 815792128 + }, + { + "epoch": 0.25, + "learning_rate": 0.000608320962888666, + "loss": 1.3207, + "theoretical_loss": 3.750389277567006, + "tokens_seen": 815857664 + }, + { + "epoch": 0.25, + "objective/train/docs_used": 301260, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.4444539546966553, + "objective/train/theoretical_loss": 3.7503597468555765, + "objective/train/tokens_used": 836383200, + "theoretical_loss": 3.7503597468555765, + "tokens_seen": 815923200 + }, + { + "epoch": 0.25, + "learning_rate": 0.0006083049147442327, + "loss": 1.4667, + "theoretical_loss": 3.7503597468555765, + "tokens_seen": 815923200 + }, + { + "epoch": 0.25, + "learning_rate": 0.0006082888665997994, + "loss": 1.3426, + "theoretical_loss": 3.7503302191800802, + "tokens_seen": 815988736 + }, + { + "epoch": 0.25, + "learning_rate": 0.0006082728184553661, + "loss": 1.4832, + "theoretical_loss": 3.750300694539961, + "tokens_seen": 816054272 + }, + { + "epoch": 0.25, + "learning_rate": 0.0006082567703109328, + "loss": 1.3521, + "theoretical_loss": 3.750271172934663, + "tokens_seen": 816119808 + }, + { + "epoch": 0.25, + "learning_rate": 0.0006082407221664995, + "loss": 1.4888, + "theoretical_loss": 3.750241654363631, + "tokens_seen": 816185344 + }, + { + "epoch": 0.25, + "learning_rate": 0.0006082246740220662, + "loss": 1.4517, + "theoretical_loss": 3.75021213882631, + "tokens_seen": 816250880 + }, + { + "epoch": 0.25, + "learning_rate": 0.0006082086258776329, + "loss": 1.5635, + "theoretical_loss": 3.7501826263221436, + "tokens_seen": 816316416 + }, + { + "epoch": 0.25, + "learning_rate": 0.0006081925777331996, + "loss": 1.5211, + "theoretical_loss": 3.7501531168505773, + "tokens_seen": 816381952 + }, + { + "epoch": 0.25, + "learning_rate": 0.0006081765295887663, + "loss": 1.4913, + "theoretical_loss": 3.750123610411056, + "tokens_seen": 816447488 + }, + { + "epoch": 0.25, + "learning_rate": 0.000608160481444333, + "loss": 1.4458, + "theoretical_loss": 3.750094107003024, + "tokens_seen": 816513024 + }, + { + "epoch": 0.25, + "learning_rate": 0.0006081444332998997, + "loss": 1.306, + "theoretical_loss": 3.7500646066259278, + "tokens_seen": 816578560 + }, + { + "epoch": 0.25, + "learning_rate": 0.0006081283851554664, + "loss": 1.5608, + "theoretical_loss": 3.750035109279212, + "tokens_seen": 816644096 + }, + { + "epoch": 0.25, + "learning_rate": 0.0006081123370110331, + "loss": 1.4499, + "theoretical_loss": 3.7500056149623227, + "tokens_seen": 816709632 + }, + { + "epoch": 0.25, + "learning_rate": 0.0006080962888665998, + "loss": 1.3021, + "theoretical_loss": 3.7499761236747053, + "tokens_seen": 816775168 + }, + { + "epoch": 0.25, + "learning_rate": 0.0006080802407221665, + "loss": 1.4204, + "theoretical_loss": 3.7499466354158058, + "tokens_seen": 816840704 + }, + { + "epoch": 0.25, + "learning_rate": 0.0006080641925777332, + "loss": 1.5087, + "theoretical_loss": 3.7499171501850697, + "tokens_seen": 816906240 + }, + { + "epoch": 0.25, + "learning_rate": 0.0006080481444332999, + "loss": 1.4819, + "theoretical_loss": 3.749887667981944, + "tokens_seen": 816971776 + }, + { + "epoch": 0.25, + "learning_rate": 0.0006080320962888666, + "loss": 1.4163, + "theoretical_loss": 3.7498581888058746, + "tokens_seen": 817037312 + }, + { + "epoch": 0.25, + "learning_rate": 0.0006080160481444333, + "loss": 1.5212, + "theoretical_loss": 3.7498287126563072, + "tokens_seen": 817102848 + }, + { + "epoch": 0.25, + "learning_rate": 0.000608, + "loss": 1.4321, + "theoretical_loss": 3.7497992395326896, + "tokens_seen": 817168384 + }, + { + "epoch": 0.25, + "learning_rate": 0.0006079839518555667, + "loss": 1.4751, + "theoretical_loss": 3.749769769434468, + "tokens_seen": 817233920 + }, + { + "epoch": 0.25, + "learning_rate": 0.0006079679037111334, + "loss": 1.5251, + "theoretical_loss": 3.7497403023610882, + "tokens_seen": 817299456 + }, + { + "epoch": 0.25, + "learning_rate": 0.0006079518555667001, + "loss": 1.5127, + "theoretical_loss": 3.7497108383119997, + "tokens_seen": 817364992 + }, + { + "epoch": 0.25, + "learning_rate": 0.0006079358074222668, + "loss": 1.492, + "theoretical_loss": 3.749681377286647, + "tokens_seen": 817430528 + }, + { + "epoch": 0.25, + "learning_rate": 0.0006079197592778335, + "loss": 1.4053, + "theoretical_loss": 3.749651919284479, + "tokens_seen": 817496064 + }, + { + "epoch": 0.25, + "learning_rate": 0.0006079037111334002, + "loss": 1.4495, + "theoretical_loss": 3.7496224643049425, + "tokens_seen": 817561600 + }, + { + "epoch": 0.25, + "learning_rate": 0.0006078876629889669, + "loss": 1.4037, + "theoretical_loss": 3.7495930123474857, + "tokens_seen": 817627136 + }, + { + "epoch": 0.25, + "learning_rate": 0.0006078716148445336, + "loss": 1.5405, + "theoretical_loss": 3.749563563411556, + "tokens_seen": 817692672 + }, + { + "epoch": 0.25, + "learning_rate": 0.0006078555667001003, + "loss": 1.5246, + "theoretical_loss": 3.7495341174966015, + "tokens_seen": 817758208 + }, + { + "epoch": 0.25, + "learning_rate": 0.000607839518555667, + "loss": 1.4727, + "theoretical_loss": 3.7495046746020693, + "tokens_seen": 817823744 + }, + { + "epoch": 0.25, + "learning_rate": 0.0006078234704112337, + "loss": 1.4452, + "theoretical_loss": 3.7494752347274085, + "tokens_seen": 817889280 + }, + { + "epoch": 0.25, + "learning_rate": 0.0006078074222668004, + "loss": 1.4891, + "theoretical_loss": 3.749445797872067, + "tokens_seen": 817954816 + }, + { + "epoch": 0.25, + "learning_rate": 0.0006077913741223672, + "loss": 1.5279, + "theoretical_loss": 3.7494163640354943, + "tokens_seen": 818020352 + }, + { + "epoch": 0.25, + "learning_rate": 0.0006077753259779338, + "loss": 1.4148, + "theoretical_loss": 3.7493869332171377, + "tokens_seen": 818085888 + }, + { + "epoch": 0.25, + "learning_rate": 0.0006077592778335006, + "loss": 1.4494, + "theoretical_loss": 3.7493575054164463, + "tokens_seen": 818151424 + }, + { + "epoch": 0.25, + "learning_rate": 0.0006077432296890672, + "loss": 1.4309, + "theoretical_loss": 3.749328080632869, + "tokens_seen": 818216960 + }, + { + "epoch": 0.25, + "learning_rate": 0.000607727181544634, + "loss": 1.428, + "theoretical_loss": 3.749298658865855, + "tokens_seen": 818282496 + }, + { + "epoch": 0.25, + "learning_rate": 0.0006077111334002006, + "loss": 1.3875, + "theoretical_loss": 3.7492692401148537, + "tokens_seen": 818348032 + }, + { + "epoch": 0.25, + "learning_rate": 0.0006076950852557674, + "loss": 1.4621, + "theoretical_loss": 3.7492398243793144, + "tokens_seen": 818413568 + }, + { + "epoch": 0.25, + "learning_rate": 0.000607679037111334, + "loss": 1.3279, + "theoretical_loss": 3.7492104116586855, + "tokens_seen": 818479104 + }, + { + "epoch": 0.25, + "learning_rate": 0.0006076629889669008, + "loss": 1.555, + "theoretical_loss": 3.749181001952418, + "tokens_seen": 818544640 + }, + { + "epoch": 0.25, + "learning_rate": 0.0006076469408224674, + "loss": 1.4583, + "theoretical_loss": 3.7491515952599617, + "tokens_seen": 818610176 + }, + { + "epoch": 0.25, + "learning_rate": 0.0006076308926780342, + "loss": 1.5035, + "theoretical_loss": 3.7491221915807653, + "tokens_seen": 818675712 + }, + { + "epoch": 0.25, + "learning_rate": 0.0006076148445336008, + "loss": 1.4671, + "theoretical_loss": 3.74909279091428, + "tokens_seen": 818741248 + }, + { + "epoch": 0.25, + "learning_rate": 0.0006075987963891676, + "loss": 1.5044, + "theoretical_loss": 3.749063393259955, + "tokens_seen": 818806784 + }, + { + "epoch": 0.25, + "learning_rate": 0.0006075827482447342, + "loss": 1.5031, + "theoretical_loss": 3.749033998617241, + "tokens_seen": 818872320 + }, + { + "epoch": 0.25, + "learning_rate": 0.000607566700100301, + "loss": 1.3708, + "theoretical_loss": 3.7490046069855896, + "tokens_seen": 818937856 + }, + { + "epoch": 0.25, + "learning_rate": 0.0006075506519558677, + "loss": 1.2928, + "theoretical_loss": 3.74897521836445, + "tokens_seen": 819003392 + }, + { + "epoch": 0.25, + "learning_rate": 0.0006075346038114344, + "loss": 1.3075, + "theoretical_loss": 3.7489458327532734, + "tokens_seen": 819068928 + }, + { + "epoch": 0.25, + "learning_rate": 0.000607518555667001, + "loss": 1.4731, + "theoretical_loss": 3.748916450151511, + "tokens_seen": 819134464 + }, + { + "epoch": 0.25, + "objective/train/docs_used": 302476, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.511124849319458, + "objective/train/theoretical_loss": 3.7488870705586135, + "objective/train/tokens_used": 839660000, + "theoretical_loss": 3.7488870705586135, + "tokens_seen": 819200000 + }, + { + "epoch": 0.25, + "learning_rate": 0.0006075025075225678, + "loss": 1.513, + "theoretical_loss": 3.7488870705586135, + "tokens_seen": 819200000 + }, + { + "epoch": 0.25, + "learning_rate": 0.0006074864593781345, + "loss": 1.3286, + "theoretical_loss": 3.7488576939740326, + "tokens_seen": 819265536 + }, + { + "epoch": 0.25, + "learning_rate": 0.0006074704112337012, + "loss": 1.4241, + "theoretical_loss": 3.7488283203972195, + "tokens_seen": 819331072 + }, + { + "epoch": 0.25, + "learning_rate": 0.0006074543630892679, + "loss": 1.467, + "theoretical_loss": 3.748798949827625, + "tokens_seen": 819396608 + }, + { + "epoch": 0.25, + "learning_rate": 0.0006074383149448346, + "loss": 1.4206, + "theoretical_loss": 3.7487695822647016, + "tokens_seen": 819462144 + }, + { + "epoch": 0.25, + "learning_rate": 0.0006074222668004013, + "loss": 1.5304, + "theoretical_loss": 3.7487402177079003, + "tokens_seen": 819527680 + }, + { + "epoch": 0.25, + "learning_rate": 0.000607406218655968, + "loss": 1.4323, + "theoretical_loss": 3.7487108561566735, + "tokens_seen": 819593216 + }, + { + "epoch": 0.25, + "learning_rate": 0.0006073901705115347, + "loss": 1.4826, + "theoretical_loss": 3.748681497610474, + "tokens_seen": 819658752 + }, + { + "epoch": 0.25, + "learning_rate": 0.0006073741223671014, + "loss": 1.4934, + "theoretical_loss": 3.748652142068753, + "tokens_seen": 819724288 + }, + { + "epoch": 0.25, + "learning_rate": 0.0006073580742226681, + "loss": 1.4308, + "theoretical_loss": 3.748622789530963, + "tokens_seen": 819789824 + }, + { + "epoch": 0.25, + "learning_rate": 0.0006073420260782348, + "loss": 1.5627, + "theoretical_loss": 3.748593439996557, + "tokens_seen": 819855360 + }, + { + "epoch": 0.25, + "learning_rate": 0.0006073259779338015, + "loss": 1.4918, + "theoretical_loss": 3.748564093464987, + "tokens_seen": 819920896 + }, + { + "epoch": 0.25, + "learning_rate": 0.0006073099297893681, + "loss": 1.4258, + "theoretical_loss": 3.748534749935706, + "tokens_seen": 819986432 + }, + { + "epoch": 0.25, + "learning_rate": 0.0006072938816449349, + "loss": 1.378, + "theoretical_loss": 3.748505409408167, + "tokens_seen": 820051968 + }, + { + "epoch": 0.25, + "learning_rate": 0.0006072778335005015, + "loss": 1.353, + "theoretical_loss": 3.7484760718818233, + "tokens_seen": 820117504 + }, + { + "epoch": 0.25, + "learning_rate": 0.0006072617853560683, + "loss": 1.4914, + "theoretical_loss": 3.748446737356128, + "tokens_seen": 820183040 + }, + { + "epoch": 0.25, + "learning_rate": 0.0006072457372116349, + "loss": 1.5001, + "theoretical_loss": 3.7484174058305344, + "tokens_seen": 820248576 + }, + { + "epoch": 0.25, + "learning_rate": 0.0006072296890672017, + "loss": 1.4732, + "theoretical_loss": 3.748388077304496, + "tokens_seen": 820314112 + }, + { + "epoch": 0.25, + "learning_rate": 0.0006072136409227683, + "loss": 1.4212, + "theoretical_loss": 3.7483587517774666, + "tokens_seen": 820379648 + }, + { + "epoch": 0.25, + "learning_rate": 0.0006071975927783351, + "loss": 1.402, + "theoretical_loss": 3.7483294292488996, + "tokens_seen": 820445184 + }, + { + "epoch": 0.25, + "learning_rate": 0.0006071815446339017, + "loss": 1.472, + "theoretical_loss": 3.748300109718249, + "tokens_seen": 820510720 + }, + { + "epoch": 0.25, + "learning_rate": 0.0006071654964894684, + "loss": 1.4055, + "theoretical_loss": 3.74827079318497, + "tokens_seen": 820576256 + }, + { + "epoch": 0.25, + "learning_rate": 0.0006071494483450351, + "loss": 1.6045, + "theoretical_loss": 3.748241479648515, + "tokens_seen": 820641792 + }, + { + "epoch": 0.25, + "learning_rate": 0.0006071334002006018, + "loss": 1.3061, + "theoretical_loss": 3.7482121691083394, + "tokens_seen": 820707328 + }, + { + "epoch": 0.25, + "learning_rate": 0.0006071173520561685, + "loss": 1.4882, + "theoretical_loss": 3.748182861563898, + "tokens_seen": 820772864 + }, + { + "epoch": 0.25, + "learning_rate": 0.0006071013039117352, + "loss": 1.4296, + "theoretical_loss": 3.748153557014645, + "tokens_seen": 820838400 + }, + { + "epoch": 0.25, + "learning_rate": 0.0006070852557673019, + "loss": 1.394, + "theoretical_loss": 3.748124255460035, + "tokens_seen": 820903936 + }, + { + "epoch": 0.25, + "learning_rate": 0.0006070692076228686, + "loss": 1.5774, + "theoretical_loss": 3.748094956899523, + "tokens_seen": 820969472 + }, + { + "epoch": 0.25, + "learning_rate": 0.0006070531594784353, + "loss": 1.5406, + "theoretical_loss": 3.7480656613325642, + "tokens_seen": 821035008 + }, + { + "epoch": 0.25, + "learning_rate": 0.000607037111334002, + "loss": 1.4842, + "theoretical_loss": 3.7480363687586147, + "tokens_seen": 821100544 + }, + { + "epoch": 0.25, + "learning_rate": 0.0006070210631895687, + "loss": 1.413, + "theoretical_loss": 3.7480070791771283, + "tokens_seen": 821166080 + }, + { + "epoch": 0.25, + "learning_rate": 0.0006070050150451354, + "loss": 1.5811, + "theoretical_loss": 3.7479777925875615, + "tokens_seen": 821231616 + }, + { + "epoch": 0.25, + "learning_rate": 0.0006069889669007021, + "loss": 1.5154, + "theoretical_loss": 3.74794850898937, + "tokens_seen": 821297152 + }, + { + "epoch": 0.25, + "learning_rate": 0.0006069729187562688, + "loss": 1.5285, + "theoretical_loss": 3.7479192283820084, + "tokens_seen": 821362688 + }, + { + "epoch": 0.25, + "learning_rate": 0.0006069568706118355, + "loss": 1.4414, + "theoretical_loss": 3.747889950764934, + "tokens_seen": 821428224 + }, + { + "epoch": 0.25, + "learning_rate": 0.0006069408224674022, + "loss": 1.5829, + "theoretical_loss": 3.747860676137603, + "tokens_seen": 821493760 + }, + { + "epoch": 0.25, + "learning_rate": 0.0006069247743229689, + "loss": 1.4655, + "theoretical_loss": 3.74783140449947, + "tokens_seen": 821559296 + }, + { + "epoch": 0.25, + "learning_rate": 0.0006069087261785356, + "loss": 1.448, + "theoretical_loss": 3.747802135849993, + "tokens_seen": 821624832 + }, + { + "epoch": 0.25, + "learning_rate": 0.0006068926780341023, + "loss": 1.5, + "theoretical_loss": 3.747772870188628, + "tokens_seen": 821690368 + }, + { + "epoch": 0.25, + "learning_rate": 0.000606876629889669, + "loss": 1.4689, + "theoretical_loss": 3.747743607514831, + "tokens_seen": 821755904 + }, + { + "epoch": 0.25, + "learning_rate": 0.0006068605817452357, + "loss": 1.4615, + "theoretical_loss": 3.74771434782806, + "tokens_seen": 821821440 + }, + { + "epoch": 0.25, + "learning_rate": 0.0006068445336008024, + "loss": 1.3573, + "theoretical_loss": 3.747685091127771, + "tokens_seen": 821886976 + }, + { + "epoch": 0.25, + "learning_rate": 0.0006068284854563691, + "loss": 1.3704, + "theoretical_loss": 3.7476558374134217, + "tokens_seen": 821952512 + }, + { + "epoch": 0.25, + "learning_rate": 0.0006068124373119358, + "loss": 1.623, + "theoretical_loss": 3.7476265866844685, + "tokens_seen": 822018048 + }, + { + "epoch": 0.25, + "learning_rate": 0.0006067963891675025, + "loss": 1.4187, + "theoretical_loss": 3.7475973389403694, + "tokens_seen": 822083584 + }, + { + "epoch": 0.25, + "learning_rate": 0.0006067803410230692, + "loss": 1.3557, + "theoretical_loss": 3.747568094180582, + "tokens_seen": 822149120 + }, + { + "epoch": 0.25, + "learning_rate": 0.0006067642928786359, + "loss": 1.4579, + "theoretical_loss": 3.747538852404563, + "tokens_seen": 822214656 + }, + { + "epoch": 0.25, + "learning_rate": 0.0006067482447342026, + "loss": 1.3744, + "theoretical_loss": 3.747509613611771, + "tokens_seen": 822280192 + }, + { + "epoch": 0.25, + "learning_rate": 0.0006067321965897693, + "loss": 1.4277, + "theoretical_loss": 3.747480377801664, + "tokens_seen": 822345728 + }, + { + "epoch": 0.25, + "learning_rate": 0.000606716148445336, + "loss": 1.4528, + "theoretical_loss": 3.7474511449737, + "tokens_seen": 822411264 + }, + { + "epoch": 0.25, + "objective/train/docs_used": 303734, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.4952170848846436, + "objective/train/theoretical_loss": 3.7474219151273367, + "objective/train/tokens_used": 842936800, + "theoretical_loss": 3.7474219151273367, + "tokens_seen": 822476800 + }, + { + "epoch": 0.25, + "learning_rate": 0.0006067001003009027, + "loss": 1.5446, + "theoretical_loss": 3.7474219151273367, + "tokens_seen": 822476800 + }, + { + "epoch": 0.25, + "learning_rate": 0.0006066840521564694, + "loss": 1.5289, + "theoretical_loss": 3.7473926882620328, + "tokens_seen": 822542336 + }, + { + "epoch": 0.25, + "learning_rate": 0.0006066680040120361, + "loss": 1.5559, + "theoretical_loss": 3.7473634643772473, + "tokens_seen": 822607872 + }, + { + "epoch": 0.25, + "learning_rate": 0.0006066519558676028, + "loss": 1.5383, + "theoretical_loss": 3.747334243472438, + "tokens_seen": 822673408 + }, + { + "epoch": 0.25, + "learning_rate": 0.0006066359077231695, + "loss": 1.3423, + "theoretical_loss": 3.747305025547064, + "tokens_seen": 822738944 + }, + { + "epoch": 0.25, + "learning_rate": 0.0006066198595787362, + "loss": 1.557, + "theoretical_loss": 3.747275810600584, + "tokens_seen": 822804480 + }, + { + "epoch": 0.25, + "learning_rate": 0.0006066038114343029, + "loss": 1.447, + "theoretical_loss": 3.747246598632458, + "tokens_seen": 822870016 + }, + { + "epoch": 0.25, + "learning_rate": 0.0006065877632898697, + "loss": 1.4366, + "theoretical_loss": 3.7472173896421435, + "tokens_seen": 822935552 + }, + { + "epoch": 0.25, + "learning_rate": 0.0006065717151454363, + "loss": 1.465, + "theoretical_loss": 3.7471881836291017, + "tokens_seen": 823001088 + }, + { + "epoch": 0.25, + "learning_rate": 0.0006065556670010031, + "loss": 1.4303, + "theoretical_loss": 3.7471589805927907, + "tokens_seen": 823066624 + }, + { + "epoch": 0.25, + "learning_rate": 0.0006065396188565697, + "loss": 1.4491, + "theoretical_loss": 3.7471297805326707, + "tokens_seen": 823132160 + }, + { + "epoch": 0.25, + "learning_rate": 0.0006065235707121365, + "loss": 1.5534, + "theoretical_loss": 3.7471005834482014, + "tokens_seen": 823197696 + }, + { + "epoch": 0.25, + "learning_rate": 0.0006065075225677031, + "loss": 1.546, + "theoretical_loss": 3.7470713893388425, + "tokens_seen": 823263232 + }, + { + "epoch": 0.25, + "learning_rate": 0.0006064914744232699, + "loss": 1.3875, + "theoretical_loss": 3.7470421982040545, + "tokens_seen": 823328768 + }, + { + "epoch": 0.25, + "learning_rate": 0.0006064754262788365, + "loss": 1.3907, + "theoretical_loss": 3.7470130100432972, + "tokens_seen": 823394304 + }, + { + "epoch": 0.25, + "learning_rate": 0.0006064593781344033, + "loss": 1.4255, + "theoretical_loss": 3.746983824856031, + "tokens_seen": 823459840 + }, + { + "epoch": 0.25, + "learning_rate": 0.0006064433299899699, + "loss": 1.408, + "theoretical_loss": 3.7469546426417164, + "tokens_seen": 823525376 + }, + { + "epoch": 0.25, + "learning_rate": 0.0006064272818455367, + "loss": 1.5287, + "theoretical_loss": 3.746925463399814, + "tokens_seen": 823590912 + }, + { + "epoch": 0.25, + "learning_rate": 0.0006064112337011033, + "loss": 1.4086, + "theoretical_loss": 3.746896287129785, + "tokens_seen": 823656448 + }, + { + "epoch": 0.25, + "learning_rate": 0.0006063951855566701, + "loss": 1.5115, + "theoretical_loss": 3.7468671138310894, + "tokens_seen": 823721984 + }, + { + "epoch": 0.25, + "learning_rate": 0.0006063791374122367, + "loss": 1.4978, + "theoretical_loss": 3.7468379435031887, + "tokens_seen": 823787520 + }, + { + "epoch": 0.25, + "learning_rate": 0.0006063630892678035, + "loss": 1.4155, + "theoretical_loss": 3.7468087761455444, + "tokens_seen": 823853056 + }, + { + "epoch": 0.25, + "learning_rate": 0.0006063470411233701, + "loss": 1.4246, + "theoretical_loss": 3.746779611757617, + "tokens_seen": 823918592 + }, + { + "epoch": 0.25, + "learning_rate": 0.0006063309929789369, + "loss": 1.308, + "theoretical_loss": 3.746750450338869, + "tokens_seen": 823984128 + }, + { + "epoch": 0.25, + "learning_rate": 0.0006063149448345035, + "loss": 1.479, + "theoretical_loss": 3.746721291888761, + "tokens_seen": 824049664 + }, + { + "epoch": 0.25, + "learning_rate": 0.0006062988966900703, + "loss": 1.4537, + "theoretical_loss": 3.7466921364067556, + "tokens_seen": 824115200 + }, + { + "epoch": 0.25, + "learning_rate": 0.000606282848545637, + "loss": 1.5177, + "theoretical_loss": 3.746662983892314, + "tokens_seen": 824180736 + }, + { + "epoch": 0.25, + "learning_rate": 0.0006062668004012037, + "loss": 1.5167, + "theoretical_loss": 3.746633834344898, + "tokens_seen": 824246272 + }, + { + "epoch": 0.25, + "learning_rate": 0.0006062507522567703, + "loss": 1.4776, + "theoretical_loss": 3.7466046877639707, + "tokens_seen": 824311808 + }, + { + "epoch": 0.25, + "learning_rate": 0.0006062347041123371, + "loss": 1.4356, + "theoretical_loss": 3.7465755441489943, + "tokens_seen": 824377344 + }, + { + "epoch": 0.25, + "learning_rate": 0.0006062186559679037, + "loss": 1.5265, + "theoretical_loss": 3.74654640349943, + "tokens_seen": 824442880 + }, + { + "epoch": 0.25, + "learning_rate": 0.0006062026078234705, + "loss": 1.5666, + "theoretical_loss": 3.7465172658147416, + "tokens_seen": 824508416 + }, + { + "epoch": 0.25, + "learning_rate": 0.0006061865596790372, + "loss": 1.4299, + "theoretical_loss": 3.7464881310943916, + "tokens_seen": 824573952 + }, + { + "epoch": 0.25, + "learning_rate": 0.0006061705115346039, + "loss": 1.3654, + "theoretical_loss": 3.7464589993378423, + "tokens_seen": 824639488 + }, + { + "epoch": 0.25, + "learning_rate": 0.0006061544633901706, + "loss": 1.4463, + "theoretical_loss": 3.7464298705445573, + "tokens_seen": 824705024 + }, + { + "epoch": 0.25, + "learning_rate": 0.0006061384152457373, + "loss": 1.5036, + "theoretical_loss": 3.7464007447139993, + "tokens_seen": 824770560 + }, + { + "epoch": 0.25, + "learning_rate": 0.000606122367101304, + "loss": 1.5233, + "theoretical_loss": 3.746371621845632, + "tokens_seen": 824836096 + }, + { + "epoch": 0.25, + "learning_rate": 0.0006061063189568707, + "loss": 1.3952, + "theoretical_loss": 3.7463425019389187, + "tokens_seen": 824901632 + }, + { + "epoch": 0.25, + "learning_rate": 0.0006060902708124374, + "loss": 1.4414, + "theoretical_loss": 3.746313384993323, + "tokens_seen": 824967168 + }, + { + "epoch": 0.25, + "learning_rate": 0.0006060742226680041, + "loss": 1.4617, + "theoretical_loss": 3.746284271008308, + "tokens_seen": 825032704 + }, + { + "epoch": 0.25, + "learning_rate": 0.0006060581745235708, + "loss": 1.5118, + "theoretical_loss": 3.7462551599833382, + "tokens_seen": 825098240 + }, + { + "epoch": 0.25, + "learning_rate": 0.0006060421263791375, + "loss": 1.4225, + "theoretical_loss": 3.746226051917877, + "tokens_seen": 825163776 + }, + { + "epoch": 0.25, + "learning_rate": 0.0006060260782347042, + "loss": 1.3487, + "theoretical_loss": 3.7461969468113896, + "tokens_seen": 825229312 + }, + { + "epoch": 0.25, + "learning_rate": 0.0006060100300902709, + "loss": 1.5679, + "theoretical_loss": 3.746167844663339, + "tokens_seen": 825294848 + }, + { + "epoch": 0.25, + "learning_rate": 0.0006059939819458376, + "loss": 1.4131, + "theoretical_loss": 3.7461387454731905, + "tokens_seen": 825360384 + }, + { + "epoch": 0.25, + "learning_rate": 0.0006059779338014043, + "loss": 1.3602, + "theoretical_loss": 3.746109649240408, + "tokens_seen": 825425920 + }, + { + "epoch": 0.25, + "learning_rate": 0.000605961885656971, + "loss": 1.5931, + "theoretical_loss": 3.7460805559644568, + "tokens_seen": 825491456 + }, + { + "epoch": 0.25, + "learning_rate": 0.0006059458375125377, + "loss": 1.4332, + "theoretical_loss": 3.7460514656448005, + "tokens_seen": 825556992 + }, + { + "epoch": 0.25, + "learning_rate": 0.0006059297893681044, + "loss": 1.4736, + "theoretical_loss": 3.7460223782809057, + "tokens_seen": 825622528 + }, + { + "epoch": 0.25, + "learning_rate": 0.000605913741223671, + "loss": 1.5233, + "theoretical_loss": 3.7459932938722362, + "tokens_seen": 825688064 + }, + { + "epoch": 0.25, + "objective/train/docs_used": 304950, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.3893260955810547, + "objective/train/theoretical_loss": 3.745964212418258, + "objective/train/tokens_used": 846213600, + "theoretical_loss": 3.745964212418258, + "tokens_seen": 825753600 + }, + { + "epoch": 0.25, + "learning_rate": 0.0006058976930792378, + "loss": 1.4544, + "theoretical_loss": 3.745964212418258, + "tokens_seen": 825753600 + }, + { + "epoch": 0.25, + "learning_rate": 0.0006058816449348044, + "loss": 1.5312, + "theoretical_loss": 3.7459351339184366, + "tokens_seen": 825819136 + }, + { + "epoch": 0.25, + "learning_rate": 0.0006058655967903712, + "loss": 1.5263, + "theoretical_loss": 3.7459060583722366, + "tokens_seen": 825884672 + }, + { + "epoch": 0.25, + "learning_rate": 0.0006058495486459378, + "loss": 1.5134, + "theoretical_loss": 3.745876985779124, + "tokens_seen": 825950208 + }, + { + "epoch": 0.25, + "learning_rate": 0.0006058335005015046, + "loss": 1.4758, + "theoretical_loss": 3.745847916138565, + "tokens_seen": 826015744 + }, + { + "epoch": 0.25, + "learning_rate": 0.0006058174523570712, + "loss": 1.4461, + "theoretical_loss": 3.745818849450025, + "tokens_seen": 826081280 + }, + { + "epoch": 0.25, + "learning_rate": 0.000605801404212638, + "loss": 1.4808, + "theoretical_loss": 3.7457897857129705, + "tokens_seen": 826146816 + }, + { + "epoch": 0.25, + "learning_rate": 0.0006057853560682046, + "loss": 1.476, + "theoretical_loss": 3.7457607249268676, + "tokens_seen": 826212352 + }, + { + "epoch": 0.25, + "learning_rate": 0.0006057693079237713, + "loss": 1.4866, + "theoretical_loss": 3.7457316670911824, + "tokens_seen": 826277888 + }, + { + "epoch": 0.25, + "learning_rate": 0.000605753259779338, + "loss": 1.3824, + "theoretical_loss": 3.7457026122053816, + "tokens_seen": 826343424 + }, + { + "epoch": 0.25, + "learning_rate": 0.0006057372116349047, + "loss": 1.4573, + "theoretical_loss": 3.7456735602689317, + "tokens_seen": 826408960 + }, + { + "epoch": 0.25, + "learning_rate": 0.0006057211634904714, + "loss": 1.4242, + "theoretical_loss": 3.7456445112812995, + "tokens_seen": 826474496 + }, + { + "epoch": 0.25, + "learning_rate": 0.0006057051153460381, + "loss": 1.4791, + "theoretical_loss": 3.745615465241952, + "tokens_seen": 826540032 + }, + { + "epoch": 0.25, + "learning_rate": 0.0006056890672016048, + "loss": 1.3182, + "theoretical_loss": 3.745586422150356, + "tokens_seen": 826605568 + }, + { + "epoch": 0.25, + "learning_rate": 0.0006056730190571715, + "loss": 1.5141, + "theoretical_loss": 3.745557382005979, + "tokens_seen": 826671104 + }, + { + "epoch": 0.25, + "learning_rate": 0.0006056569709127382, + "loss": 1.4878, + "theoretical_loss": 3.7455283448082883, + "tokens_seen": 826736640 + }, + { + "epoch": 0.25, + "learning_rate": 0.0006056409227683049, + "loss": 1.5505, + "theoretical_loss": 3.7454993105567507, + "tokens_seen": 826802176 + }, + { + "epoch": 0.25, + "learning_rate": 0.0006056248746238716, + "loss": 1.4586, + "theoretical_loss": 3.745470279250835, + "tokens_seen": 826867712 + }, + { + "epoch": 0.25, + "learning_rate": 0.0006056088264794383, + "loss": 1.4202, + "theoretical_loss": 3.7454412508900075, + "tokens_seen": 826933248 + }, + { + "epoch": 0.25, + "learning_rate": 0.000605592778335005, + "loss": 1.3944, + "theoretical_loss": 3.7454122254737374, + "tokens_seen": 826998784 + }, + { + "epoch": 0.25, + "learning_rate": 0.0006055767301905717, + "loss": 1.5626, + "theoretical_loss": 3.7453832030014915, + "tokens_seen": 827064320 + }, + { + "epoch": 0.25, + "learning_rate": 0.0006055606820461384, + "loss": 1.5603, + "theoretical_loss": 3.745354183472739, + "tokens_seen": 827129856 + }, + { + "epoch": 0.25, + "learning_rate": 0.0006055446339017051, + "loss": 1.6016, + "theoretical_loss": 3.745325166886947, + "tokens_seen": 827195392 + }, + { + "epoch": 0.25, + "learning_rate": 0.0006055285857572718, + "loss": 1.5273, + "theoretical_loss": 3.7452961532435856, + "tokens_seen": 827260928 + }, + { + "epoch": 0.25, + "learning_rate": 0.0006055125376128385, + "loss": 1.5278, + "theoretical_loss": 3.7452671425421213, + "tokens_seen": 827326464 + }, + { + "epoch": 0.25, + "learning_rate": 0.0006054964894684052, + "loss": 1.5377, + "theoretical_loss": 3.7452381347820247, + "tokens_seen": 827392000 + }, + { + "epoch": 0.25, + "learning_rate": 0.0006054804413239719, + "loss": 1.6345, + "theoretical_loss": 3.7452091299627637, + "tokens_seen": 827457536 + }, + { + "epoch": 0.25, + "learning_rate": 0.0006054643931795386, + "loss": 1.516, + "theoretical_loss": 3.745180128083807, + "tokens_seen": 827523072 + }, + { + "epoch": 0.25, + "learning_rate": 0.0006054483450351053, + "loss": 1.466, + "theoretical_loss": 3.7451511291446242, + "tokens_seen": 827588608 + }, + { + "epoch": 0.25, + "learning_rate": 0.000605432296890672, + "loss": 1.461, + "theoretical_loss": 3.745122133144685, + "tokens_seen": 827654144 + }, + { + "epoch": 0.25, + "learning_rate": 0.0006054162487462387, + "loss": 1.5121, + "theoretical_loss": 3.7450931400834575, + "tokens_seen": 827719680 + }, + { + "epoch": 0.25, + "learning_rate": 0.0006054002006018054, + "loss": 1.5132, + "theoretical_loss": 3.7450641499604123, + "tokens_seen": 827785216 + }, + { + "epoch": 0.25, + "learning_rate": 0.0006053841524573722, + "loss": 1.5292, + "theoretical_loss": 3.7450351627750185, + "tokens_seen": 827850752 + }, + { + "epoch": 0.25, + "learning_rate": 0.0006053681043129388, + "loss": 1.4768, + "theoretical_loss": 3.7450061785267463, + "tokens_seen": 827916288 + }, + { + "epoch": 0.25, + "learning_rate": 0.0006053520561685056, + "loss": 1.5212, + "theoretical_loss": 3.744977197215065, + "tokens_seen": 827981824 + }, + { + "epoch": 0.25, + "learning_rate": 0.0006053360080240722, + "loss": 1.4879, + "theoretical_loss": 3.744948218839446, + "tokens_seen": 828047360 + }, + { + "epoch": 0.25, + "learning_rate": 0.000605319959879639, + "loss": 1.5204, + "theoretical_loss": 3.744919243399358, + "tokens_seen": 828112896 + }, + { + "epoch": 0.25, + "learning_rate": 0.0006053039117352056, + "loss": 1.4464, + "theoretical_loss": 3.744890270894272, + "tokens_seen": 828178432 + }, + { + "epoch": 0.25, + "learning_rate": 0.0006052878635907724, + "loss": 1.553, + "theoretical_loss": 3.7448613013236587, + "tokens_seen": 828243968 + }, + { + "epoch": 0.25, + "learning_rate": 0.000605271815446339, + "loss": 1.5804, + "theoretical_loss": 3.744832334686988, + "tokens_seen": 828309504 + }, + { + "epoch": 0.25, + "learning_rate": 0.0006052557673019058, + "loss": 1.6239, + "theoretical_loss": 3.744803370983732, + "tokens_seen": 828375040 + }, + { + "epoch": 0.25, + "learning_rate": 0.0006052397191574724, + "loss": 1.5604, + "theoretical_loss": 3.7447744102133598, + "tokens_seen": 828440576 + }, + { + "epoch": 0.25, + "learning_rate": 0.0006052236710130392, + "loss": 1.5085, + "theoretical_loss": 3.744745452375344, + "tokens_seen": 828506112 + }, + { + "epoch": 0.25, + "learning_rate": 0.0006052076228686058, + "loss": 1.3829, + "theoretical_loss": 3.7447164974691547, + "tokens_seen": 828571648 + }, + { + "epoch": 0.25, + "learning_rate": 0.0006051915747241726, + "loss": 1.4473, + "theoretical_loss": 3.744687545494264, + "tokens_seen": 828637184 + }, + { + "epoch": 0.25, + "learning_rate": 0.0006051755265797392, + "loss": 1.5112, + "theoretical_loss": 3.744658596450143, + "tokens_seen": 828702720 + }, + { + "epoch": 0.25, + "learning_rate": 0.000605159478435306, + "loss": 1.4817, + "theoretical_loss": 3.744629650336263, + "tokens_seen": 828768256 + }, + { + "epoch": 0.25, + "learning_rate": 0.0006051434302908726, + "loss": 1.4498, + "theoretical_loss": 3.744600707152096, + "tokens_seen": 828833792 + }, + { + "epoch": 0.25, + "learning_rate": 0.0006051273821464394, + "loss": 1.546, + "theoretical_loss": 3.7445717668971144, + "tokens_seen": 828899328 + }, + { + "epoch": 0.25, + "learning_rate": 0.000605111334002006, + "loss": 1.4275, + "theoretical_loss": 3.7445428295707894, + "tokens_seen": 828964864 + }, + { + "epoch": 0.25, + "objective/train/docs_used": 306096, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.362520694732666, + "objective/train/theoretical_loss": 3.7445138951725934, + "objective/train/tokens_used": 849490400, + "theoretical_loss": 3.7445138951725934, + "tokens_seen": 829030400 + }, + { + "epoch": 0.25, + "learning_rate": 0.0006050952858575728, + "loss": 1.4647, + "theoretical_loss": 3.7445138951725934, + "tokens_seen": 829030400 + }, + { + "epoch": 0.25, + "learning_rate": 0.0006050792377131394, + "loss": 1.4125, + "theoretical_loss": 3.7444849637019986, + "tokens_seen": 829095936 + }, + { + "epoch": 0.25, + "learning_rate": 0.0006050631895687062, + "loss": 1.5777, + "theoretical_loss": 3.7444560351584775, + "tokens_seen": 829161472 + }, + { + "epoch": 0.25, + "learning_rate": 0.0006050471414242728, + "loss": 1.4468, + "theoretical_loss": 3.744427109541503, + "tokens_seen": 829227008 + }, + { + "epoch": 0.25, + "learning_rate": 0.0006050310932798396, + "loss": 1.371, + "theoretical_loss": 3.744398186850547, + "tokens_seen": 829292544 + }, + { + "epoch": 0.25, + "learning_rate": 0.0006050150451354062, + "loss": 1.4272, + "theoretical_loss": 3.7443692670850828, + "tokens_seen": 829358080 + }, + { + "epoch": 0.25, + "learning_rate": 0.000604998996990973, + "loss": 1.4941, + "theoretical_loss": 3.744340350244583, + "tokens_seen": 829423616 + }, + { + "epoch": 0.25, + "learning_rate": 0.0006049829488465396, + "loss": 1.507, + "theoretical_loss": 3.7443114363285215, + "tokens_seen": 829489152 + }, + { + "epoch": 0.25, + "learning_rate": 0.0006049669007021064, + "loss": 1.3644, + "theoretical_loss": 3.74428252533637, + "tokens_seen": 829554688 + }, + { + "epoch": 0.25, + "learning_rate": 0.000604950852557673, + "loss": 1.5077, + "theoretical_loss": 3.7442536172676038, + "tokens_seen": 829620224 + }, + { + "epoch": 0.25, + "learning_rate": 0.0006049348044132398, + "loss": 1.4722, + "theoretical_loss": 3.744224712121695, + "tokens_seen": 829685760 + }, + { + "epoch": 0.25, + "learning_rate": 0.0006049187562688064, + "loss": 1.3343, + "theoretical_loss": 3.744195809898118, + "tokens_seen": 829751296 + }, + { + "epoch": 0.25, + "learning_rate": 0.0006049027081243732, + "loss": 1.3803, + "theoretical_loss": 3.7441669105963458, + "tokens_seen": 829816832 + }, + { + "epoch": 0.25, + "learning_rate": 0.0006048866599799398, + "loss": 1.5695, + "theoretical_loss": 3.744138014215853, + "tokens_seen": 829882368 + }, + { + "epoch": 0.25, + "learning_rate": 0.0006048706118355066, + "loss": 1.4436, + "theoretical_loss": 3.7441091207561135, + "tokens_seen": 829947904 + }, + { + "epoch": 0.25, + "learning_rate": 0.0006048545636910732, + "loss": 1.4607, + "theoretical_loss": 3.7440802302166016, + "tokens_seen": 830013440 + }, + { + "epoch": 0.25, + "learning_rate": 0.00060483851554664, + "loss": 1.4737, + "theoretical_loss": 3.7440513425967907, + "tokens_seen": 830078976 + }, + { + "epoch": 0.25, + "learning_rate": 0.0006048224674022067, + "loss": 1.5101, + "theoretical_loss": 3.7440224578961563, + "tokens_seen": 830144512 + }, + { + "epoch": 0.25, + "learning_rate": 0.0006048064192577734, + "loss": 1.4791, + "theoretical_loss": 3.743993576114173, + "tokens_seen": 830210048 + }, + { + "epoch": 0.25, + "learning_rate": 0.00060479037111334, + "loss": 1.3695, + "theoretical_loss": 3.7439646972503144, + "tokens_seen": 830275584 + }, + { + "epoch": 0.25, + "learning_rate": 0.0006047743229689068, + "loss": 1.4977, + "theoretical_loss": 3.7439358213040568, + "tokens_seen": 830341120 + }, + { + "epoch": 0.25, + "learning_rate": 0.0006047582748244735, + "loss": 1.4556, + "theoretical_loss": 3.743906948274874, + "tokens_seen": 830406656 + }, + { + "epoch": 0.25, + "learning_rate": 0.0006047422266800402, + "loss": 1.5391, + "theoretical_loss": 3.7438780781622425, + "tokens_seen": 830472192 + }, + { + "epoch": 0.25, + "learning_rate": 0.0006047261785356069, + "loss": 1.4305, + "theoretical_loss": 3.743849210965636, + "tokens_seen": 830537728 + }, + { + "epoch": 0.25, + "learning_rate": 0.0006047101303911736, + "loss": 1.441, + "theoretical_loss": 3.7438203466845303, + "tokens_seen": 830603264 + }, + { + "epoch": 0.25, + "learning_rate": 0.0006046940822467403, + "loss": 1.4682, + "theoretical_loss": 3.7437914853184022, + "tokens_seen": 830668800 + }, + { + "epoch": 0.25, + "learning_rate": 0.000604678034102307, + "loss": 1.5284, + "theoretical_loss": 3.743762626866726, + "tokens_seen": 830734336 + }, + { + "epoch": 0.25, + "learning_rate": 0.0006046619859578737, + "loss": 1.3843, + "theoretical_loss": 3.7437337713289778, + "tokens_seen": 830799872 + }, + { + "epoch": 0.25, + "learning_rate": 0.0006046459378134404, + "loss": 1.4173, + "theoretical_loss": 3.743704918704634, + "tokens_seen": 830865408 + }, + { + "epoch": 0.25, + "learning_rate": 0.0006046298896690071, + "loss": 1.49, + "theoretical_loss": 3.74367606899317, + "tokens_seen": 830930944 + }, + { + "epoch": 0.25, + "learning_rate": 0.0006046138415245738, + "loss": 1.4521, + "theoretical_loss": 3.743647222194063, + "tokens_seen": 830996480 + }, + { + "epoch": 0.25, + "learning_rate": 0.0006045977933801405, + "loss": 1.5175, + "theoretical_loss": 3.7436183783067882, + "tokens_seen": 831062016 + }, + { + "epoch": 0.25, + "learning_rate": 0.0006045817452357072, + "loss": 1.4029, + "theoretical_loss": 3.7435895373308226, + "tokens_seen": 831127552 + }, + { + "epoch": 0.25, + "learning_rate": 0.0006045656970912739, + "loss": 1.3925, + "theoretical_loss": 3.743560699265643, + "tokens_seen": 831193088 + }, + { + "epoch": 0.25, + "learning_rate": 0.0006045496489468406, + "loss": 1.5228, + "theoretical_loss": 3.743531864110726, + "tokens_seen": 831258624 + }, + { + "epoch": 0.25, + "learning_rate": 0.0006045336008024073, + "loss": 1.4222, + "theoretical_loss": 3.743503031865549, + "tokens_seen": 831324160 + }, + { + "epoch": 0.25, + "learning_rate": 0.0006045175526579739, + "loss": 1.5022, + "theoretical_loss": 3.743474202529588, + "tokens_seen": 831389696 + }, + { + "epoch": 0.25, + "learning_rate": 0.0006045015045135407, + "loss": 1.4158, + "theoretical_loss": 3.743445376102321, + "tokens_seen": 831455232 + }, + { + "epoch": 0.25, + "learning_rate": 0.0006044854563691073, + "loss": 1.4803, + "theoretical_loss": 3.7434165525832253, + "tokens_seen": 831520768 + }, + { + "epoch": 0.25, + "learning_rate": 0.0006044694082246741, + "loss": 1.4232, + "theoretical_loss": 3.7433877319717777, + "tokens_seen": 831586304 + }, + { + "epoch": 0.25, + "learning_rate": 0.0006044533600802407, + "loss": 1.4984, + "theoretical_loss": 3.743358914267456, + "tokens_seen": 831651840 + }, + { + "epoch": 0.25, + "learning_rate": 0.0006044373119358075, + "loss": 1.6068, + "theoretical_loss": 3.743330099469738, + "tokens_seen": 831717376 + }, + { + "epoch": 0.25, + "learning_rate": 0.0006044212637913741, + "loss": 1.3702, + "theoretical_loss": 3.7433012875781015, + "tokens_seen": 831782912 + }, + { + "epoch": 0.25, + "learning_rate": 0.0006044052156469409, + "loss": 1.3638, + "theoretical_loss": 3.743272478592025, + "tokens_seen": 831848448 + }, + { + "epoch": 0.25, + "learning_rate": 0.0006043891675025075, + "loss": 1.602, + "theoretical_loss": 3.743243672510986, + "tokens_seen": 831913984 + }, + { + "epoch": 0.25, + "learning_rate": 0.0006043731193580742, + "loss": 1.5008, + "theoretical_loss": 3.7432148693344627, + "tokens_seen": 831979520 + }, + { + "epoch": 0.25, + "learning_rate": 0.0006043570712136409, + "loss": 1.5571, + "theoretical_loss": 3.7431860690619336, + "tokens_seen": 832045056 + }, + { + "epoch": 0.25, + "learning_rate": 0.0006043410230692076, + "loss": 1.4502, + "theoretical_loss": 3.7431572716928776, + "tokens_seen": 832110592 + }, + { + "epoch": 0.25, + "learning_rate": 0.0006043249749247743, + "loss": 1.4226, + "theoretical_loss": 3.743128477226773, + "tokens_seen": 832176128 + }, + { + "epoch": 0.25, + "learning_rate": 0.000604308926780341, + "loss": 1.4847, + "theoretical_loss": 3.7430996856630987, + "tokens_seen": 832241664 + }, + { + "epoch": 0.25, + "objective/train/docs_used": 307306, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.4665484428405762, + "objective/train/theoretical_loss": 3.743070897001333, + "objective/train/tokens_used": 852767200, + "theoretical_loss": 3.743070897001333, + "tokens_seen": 832307200 + }, + { + "epoch": 0.25, + "learning_rate": 0.0006042928786359077, + "loss": 1.4059, + "theoretical_loss": 3.743070897001333, + "tokens_seen": 832307200 + }, + { + "epoch": 0.25, + "learning_rate": 0.0006042768304914744, + "loss": 1.4848, + "theoretical_loss": 3.743042111240956, + "tokens_seen": 832372736 + }, + { + "epoch": 0.25, + "learning_rate": 0.0006042607823470411, + "loss": 1.4327, + "theoretical_loss": 3.7430133283814464, + "tokens_seen": 832438272 + }, + { + "epoch": 0.25, + "learning_rate": 0.0006042447342026078, + "loss": 1.4026, + "theoretical_loss": 3.7429845484222835, + "tokens_seen": 832503808 + }, + { + "epoch": 0.25, + "learning_rate": 0.0006042286860581745, + "loss": 1.5362, + "theoretical_loss": 3.742955771362947, + "tokens_seen": 832569344 + }, + { + "epoch": 0.25, + "learning_rate": 0.0006042126379137412, + "loss": 1.4573, + "theoretical_loss": 3.7429269972029156, + "tokens_seen": 832634880 + }, + { + "epoch": 0.25, + "learning_rate": 0.0006041965897693079, + "loss": 1.5638, + "theoretical_loss": 3.74289822594167, + "tokens_seen": 832700416 + }, + { + "epoch": 0.25, + "learning_rate": 0.0006041805416248747, + "loss": 1.4352, + "theoretical_loss": 3.7428694575786903, + "tokens_seen": 832765952 + }, + { + "epoch": 0.25, + "learning_rate": 0.0006041644934804413, + "loss": 1.4471, + "theoretical_loss": 3.742840692113455, + "tokens_seen": 832831488 + }, + { + "epoch": 0.25, + "learning_rate": 0.0006041484453360081, + "loss": 1.5771, + "theoretical_loss": 3.7428119295454456, + "tokens_seen": 832897024 + }, + { + "epoch": 0.25, + "learning_rate": 0.0006041323971915747, + "loss": 1.4375, + "theoretical_loss": 3.7427831698741425, + "tokens_seen": 832962560 + }, + { + "epoch": 0.25, + "learning_rate": 0.0006041163490471415, + "loss": 1.3733, + "theoretical_loss": 3.742754413099025, + "tokens_seen": 833028096 + }, + { + "epoch": 0.25, + "learning_rate": 0.0006041003009027081, + "loss": 1.4981, + "theoretical_loss": 3.742725659219574, + "tokens_seen": 833093632 + }, + { + "epoch": 0.25, + "learning_rate": 0.0006040842527582749, + "loss": 1.4564, + "theoretical_loss": 3.7426969082352706, + "tokens_seen": 833159168 + }, + { + "epoch": 0.25, + "learning_rate": 0.0006040682046138415, + "loss": 1.4469, + "theoretical_loss": 3.7426681601455947, + "tokens_seen": 833224704 + }, + { + "epoch": 0.25, + "learning_rate": 0.0006040521564694083, + "loss": 1.4802, + "theoretical_loss": 3.7426394149500286, + "tokens_seen": 833290240 + }, + { + "epoch": 0.25, + "learning_rate": 0.0006040361083249749, + "loss": 1.4982, + "theoretical_loss": 3.742610672648053, + "tokens_seen": 833355776 + }, + { + "epoch": 0.25, + "learning_rate": 0.0006040200601805417, + "loss": 1.5184, + "theoretical_loss": 3.7425819332391477, + "tokens_seen": 833421312 + }, + { + "epoch": 0.25, + "learning_rate": 0.0006040040120361083, + "loss": 1.4629, + "theoretical_loss": 3.7425531967227954, + "tokens_seen": 833486848 + }, + { + "epoch": 0.25, + "learning_rate": 0.0006039879638916751, + "loss": 1.4921, + "theoretical_loss": 3.7425244630984773, + "tokens_seen": 833552384 + }, + { + "epoch": 0.25, + "learning_rate": 0.0006039719157472417, + "loss": 1.5567, + "theoretical_loss": 3.742495732365675, + "tokens_seen": 833617920 + }, + { + "epoch": 0.25, + "learning_rate": 0.0006039558676028085, + "loss": 1.4876, + "theoretical_loss": 3.74246700452387, + "tokens_seen": 833683456 + }, + { + "epoch": 0.25, + "learning_rate": 0.0006039398194583751, + "loss": 1.5225, + "theoretical_loss": 3.7424382795725446, + "tokens_seen": 833748992 + }, + { + "epoch": 0.25, + "learning_rate": 0.0006039237713139419, + "loss": 1.4138, + "theoretical_loss": 3.7424095575111807, + "tokens_seen": 833814528 + }, + { + "epoch": 0.25, + "learning_rate": 0.0006039077231695085, + "loss": 1.3574, + "theoretical_loss": 3.74238083833926, + "tokens_seen": 833880064 + }, + { + "epoch": 0.25, + "learning_rate": 0.0006038916750250753, + "loss": 1.5157, + "theoretical_loss": 3.742352122056265, + "tokens_seen": 833945600 + }, + { + "epoch": 0.25, + "learning_rate": 0.0006038756268806419, + "loss": 1.5098, + "theoretical_loss": 3.7423234086616777, + "tokens_seen": 834011136 + }, + { + "epoch": 0.25, + "learning_rate": 0.0006038595787362087, + "loss": 1.4906, + "theoretical_loss": 3.7422946981549816, + "tokens_seen": 834076672 + }, + { + "epoch": 0.25, + "learning_rate": 0.0006038435305917753, + "loss": 1.555, + "theoretical_loss": 3.7422659905356586, + "tokens_seen": 834142208 + }, + { + "epoch": 0.25, + "learning_rate": 0.0006038274824473421, + "loss": 1.5618, + "theoretical_loss": 3.742237285803192, + "tokens_seen": 834207744 + }, + { + "epoch": 0.25, + "learning_rate": 0.0006038114343029087, + "loss": 1.5246, + "theoretical_loss": 3.7422085839570647, + "tokens_seen": 834273280 + }, + { + "epoch": 0.25, + "learning_rate": 0.0006037953861584755, + "loss": 1.5964, + "theoretical_loss": 3.7421798849967596, + "tokens_seen": 834338816 + }, + { + "epoch": 0.25, + "learning_rate": 0.0006037793380140421, + "loss": 1.5432, + "theoretical_loss": 3.742151188921759, + "tokens_seen": 834404352 + }, + { + "epoch": 0.25, + "learning_rate": 0.0006037632898696089, + "loss": 1.5354, + "theoretical_loss": 3.742122495731548, + "tokens_seen": 834469888 + }, + { + "epoch": 0.25, + "learning_rate": 0.0006037472417251755, + "loss": 1.5554, + "theoretical_loss": 3.742093805425609, + "tokens_seen": 834535424 + }, + { + "epoch": 0.25, + "learning_rate": 0.0006037311935807423, + "loss": 1.5778, + "theoretical_loss": 3.7420651180034255, + "tokens_seen": 834600960 + }, + { + "epoch": 0.25, + "learning_rate": 0.0006037151454363089, + "loss": 1.5505, + "theoretical_loss": 3.742036433464482, + "tokens_seen": 834666496 + }, + { + "epoch": 0.25, + "learning_rate": 0.0006036990972918757, + "loss": 1.497, + "theoretical_loss": 3.742007751808261, + "tokens_seen": 834732032 + }, + { + "epoch": 0.25, + "learning_rate": 0.0006036830491474423, + "loss": 1.434, + "theoretical_loss": 3.741979073034248, + "tokens_seen": 834797568 + }, + { + "epoch": 0.25, + "learning_rate": 0.0006036670010030091, + "loss": 1.5595, + "theoretical_loss": 3.7419503971419266, + "tokens_seen": 834863104 + }, + { + "epoch": 0.25, + "learning_rate": 0.0006036509528585757, + "loss": 1.3771, + "theoretical_loss": 3.7419217241307807, + "tokens_seen": 834928640 + }, + { + "epoch": 0.25, + "learning_rate": 0.0006036349047141425, + "loss": 1.5008, + "theoretical_loss": 3.741893054000295, + "tokens_seen": 834994176 + }, + { + "epoch": 0.25, + "learning_rate": 0.0006036188565697091, + "loss": 1.5595, + "theoretical_loss": 3.7418643867499544, + "tokens_seen": 835059712 + }, + { + "epoch": 0.25, + "learning_rate": 0.0006036028084252759, + "loss": 1.4743, + "theoretical_loss": 3.7418357223792427, + "tokens_seen": 835125248 + }, + { + "epoch": 0.25, + "learning_rate": 0.0006035867602808425, + "loss": 1.479, + "theoretical_loss": 3.7418070608876453, + "tokens_seen": 835190784 + }, + { + "epoch": 0.25, + "learning_rate": 0.0006035707121364093, + "loss": 1.4023, + "theoretical_loss": 3.741778402274647, + "tokens_seen": 835256320 + }, + { + "epoch": 0.25, + "learning_rate": 0.000603554663991976, + "loss": 1.4233, + "theoretical_loss": 3.7417497465397327, + "tokens_seen": 835321856 + }, + { + "epoch": 0.25, + "learning_rate": 0.0006035386158475427, + "loss": 1.4379, + "theoretical_loss": 3.741721093682388, + "tokens_seen": 835387392 + }, + { + "epoch": 0.25, + "learning_rate": 0.0006035225677031093, + "loss": 1.4172, + "theoretical_loss": 3.7416924437020977, + "tokens_seen": 835452928 + }, + { + "epoch": 0.25, + "learning_rate": 0.0006035065195586761, + "loss": 1.5147, + "theoretical_loss": 3.741663796598348, + "tokens_seen": 835518464 + }, + { + "epoch": 0.25, + "objective/train/docs_used": 308483, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.4287724494934082, + "objective/train/theoretical_loss": 3.741635152370624, + "objective/train/tokens_used": 856044000, + "theoretical_loss": 3.741635152370624, + "tokens_seen": 835584000 + }, + { + "epoch": 0.25, + "learning_rate": 0.0006034904714142427, + "loss": 1.4155, + "theoretical_loss": 3.741635152370624, + "tokens_seen": 835584000 + }, + { + "epoch": 0.25, + "learning_rate": 0.0006034744232698095, + "loss": 1.3839, + "theoretical_loss": 3.7416065110184116, + "tokens_seen": 835649536 + }, + { + "epoch": 0.25, + "learning_rate": 0.0006034583751253762, + "loss": 1.5087, + "theoretical_loss": 3.741577872541196, + "tokens_seen": 835715072 + }, + { + "epoch": 0.25, + "learning_rate": 0.0006034423269809429, + "loss": 1.5472, + "theoretical_loss": 3.741549236938464, + "tokens_seen": 835780608 + }, + { + "epoch": 0.25, + "learning_rate": 0.0006034262788365096, + "loss": 1.4604, + "theoretical_loss": 3.7415206042097013, + "tokens_seen": 835846144 + }, + { + "epoch": 0.25, + "learning_rate": 0.0006034102306920763, + "loss": 1.5495, + "theoretical_loss": 3.7414919743543944, + "tokens_seen": 835911680 + }, + { + "epoch": 0.25, + "learning_rate": 0.000603394182547643, + "loss": 1.4227, + "theoretical_loss": 3.7414633473720293, + "tokens_seen": 835977216 + }, + { + "epoch": 0.25, + "learning_rate": 0.0006033781344032097, + "loss": 1.4697, + "theoretical_loss": 3.741434723262093, + "tokens_seen": 836042752 + }, + { + "epoch": 0.25, + "learning_rate": 0.0006033620862587764, + "loss": 1.5033, + "theoretical_loss": 3.7414061020240723, + "tokens_seen": 836108288 + }, + { + "epoch": 0.25, + "learning_rate": 0.0006033460381143431, + "loss": 1.5262, + "theoretical_loss": 3.741377483657453, + "tokens_seen": 836173824 + }, + { + "epoch": 0.25, + "learning_rate": 0.0006033299899699098, + "loss": 1.4137, + "theoretical_loss": 3.741348868161723, + "tokens_seen": 836239360 + }, + { + "epoch": 0.25, + "learning_rate": 0.0006033139418254765, + "loss": 1.4926, + "theoretical_loss": 3.741320255536369, + "tokens_seen": 836304896 + }, + { + "epoch": 0.25, + "learning_rate": 0.0006032978936810432, + "loss": 1.4458, + "theoretical_loss": 3.741291645780878, + "tokens_seen": 836370432 + }, + { + "epoch": 0.25, + "learning_rate": 0.0006032818455366099, + "loss": 1.5264, + "theoretical_loss": 3.741263038894737, + "tokens_seen": 836435968 + }, + { + "epoch": 0.25, + "learning_rate": 0.0006032657973921766, + "loss": 1.4759, + "theoretical_loss": 3.741234434877435, + "tokens_seen": 836501504 + }, + { + "epoch": 0.25, + "learning_rate": 0.0006032497492477433, + "loss": 1.4603, + "theoretical_loss": 3.7412058337284577, + "tokens_seen": 836567040 + }, + { + "epoch": 0.25, + "learning_rate": 0.00060323370110331, + "loss": 1.5169, + "theoretical_loss": 3.741177235447294, + "tokens_seen": 836632576 + }, + { + "epoch": 0.25, + "learning_rate": 0.0006032176529588767, + "loss": 1.5527, + "theoretical_loss": 3.7411486400334315, + "tokens_seen": 836698112 + }, + { + "epoch": 0.25, + "learning_rate": 0.0006032016048144434, + "loss": 1.5, + "theoretical_loss": 3.7411200474863575, + "tokens_seen": 836763648 + }, + { + "epoch": 0.25, + "learning_rate": 0.0006031855566700101, + "loss": 1.4804, + "theoretical_loss": 3.7410914578055605, + "tokens_seen": 836829184 + }, + { + "epoch": 0.25, + "learning_rate": 0.0006031695085255768, + "loss": 1.6297, + "theoretical_loss": 3.741062870990529, + "tokens_seen": 836894720 + }, + { + "epoch": 0.25, + "learning_rate": 0.0006031534603811435, + "loss": 1.397, + "theoretical_loss": 3.741034287040751, + "tokens_seen": 836960256 + }, + { + "epoch": 0.25, + "learning_rate": 0.0006031374122367102, + "loss": 1.5133, + "theoretical_loss": 3.7410057059557156, + "tokens_seen": 837025792 + }, + { + "epoch": 0.25, + "learning_rate": 0.0006031213640922769, + "loss": 1.4738, + "theoretical_loss": 3.7409771277349106, + "tokens_seen": 837091328 + }, + { + "epoch": 0.25, + "learning_rate": 0.0006031053159478436, + "loss": 1.534, + "theoretical_loss": 3.7409485523778256, + "tokens_seen": 837156864 + }, + { + "epoch": 0.25, + "learning_rate": 0.0006030892678034102, + "loss": 1.4223, + "theoretical_loss": 3.740919979883949, + "tokens_seen": 837222400 + }, + { + "epoch": 0.25, + "learning_rate": 0.000603073219658977, + "loss": 1.3975, + "theoretical_loss": 3.740891410252769, + "tokens_seen": 837287936 + }, + { + "epoch": 0.25, + "learning_rate": 0.0006030571715145436, + "loss": 1.4694, + "theoretical_loss": 3.740862843483777, + "tokens_seen": 837353472 + }, + { + "epoch": 0.25, + "learning_rate": 0.0006030411233701104, + "loss": 1.4701, + "theoretical_loss": 3.7408342795764598, + "tokens_seen": 837419008 + }, + { + "epoch": 0.25, + "learning_rate": 0.000603025075225677, + "loss": 1.4622, + "theoretical_loss": 3.740805718530308, + "tokens_seen": 837484544 + }, + { + "epoch": 0.25, + "learning_rate": 0.0006030090270812438, + "loss": 1.5477, + "theoretical_loss": 3.7407771603448117, + "tokens_seen": 837550080 + }, + { + "epoch": 0.25, + "learning_rate": 0.0006029929789368104, + "loss": 1.5859, + "theoretical_loss": 3.7407486050194594, + "tokens_seen": 837615616 + }, + { + "epoch": 0.25, + "learning_rate": 0.0006029769307923772, + "loss": 1.4584, + "theoretical_loss": 3.7407200525537414, + "tokens_seen": 837681152 + }, + { + "epoch": 0.25, + "learning_rate": 0.0006029608826479438, + "loss": 1.4025, + "theoretical_loss": 3.740691502947148, + "tokens_seen": 837746688 + }, + { + "epoch": 0.25, + "learning_rate": 0.0006029448345035106, + "loss": 1.4606, + "theoretical_loss": 3.7406629561991687, + "tokens_seen": 837812224 + }, + { + "epoch": 0.25, + "learning_rate": 0.0006029287863590772, + "loss": 1.543, + "theoretical_loss": 3.740634412309294, + "tokens_seen": 837877760 + }, + { + "epoch": 0.25, + "learning_rate": 0.000602912738214644, + "loss": 1.4281, + "theoretical_loss": 3.740605871277014, + "tokens_seen": 837943296 + }, + { + "epoch": 0.25, + "learning_rate": 0.0006028966900702106, + "loss": 1.5581, + "theoretical_loss": 3.7405773331018195, + "tokens_seen": 838008832 + }, + { + "epoch": 0.25, + "learning_rate": 0.0006028806419257774, + "loss": 1.3723, + "theoretical_loss": 3.740548797783201, + "tokens_seen": 838074368 + }, + { + "epoch": 0.25, + "learning_rate": 0.000602864593781344, + "loss": 1.5309, + "theoretical_loss": 3.740520265320649, + "tokens_seen": 838139904 + }, + { + "epoch": 0.25, + "learning_rate": 0.0006028485456369108, + "loss": 1.4581, + "theoretical_loss": 3.740491735713654, + "tokens_seen": 838205440 + }, + { + "epoch": 0.25, + "learning_rate": 0.0006028324974924774, + "loss": 1.4511, + "theoretical_loss": 3.740463208961708, + "tokens_seen": 838270976 + }, + { + "epoch": 0.25, + "learning_rate": 0.0006028164493480442, + "loss": 1.59, + "theoretical_loss": 3.7404346850643013, + "tokens_seen": 838336512 + }, + { + "epoch": 0.25, + "learning_rate": 0.0006028004012036108, + "loss": 1.4544, + "theoretical_loss": 3.740406164020926, + "tokens_seen": 838402048 + }, + { + "epoch": 0.25, + "learning_rate": 0.0006027843530591776, + "loss": 1.489, + "theoretical_loss": 3.740377645831072, + "tokens_seen": 838467584 + }, + { + "epoch": 0.25, + "learning_rate": 0.0006027683049147442, + "loss": 1.4493, + "theoretical_loss": 3.7403491304942325, + "tokens_seen": 838533120 + }, + { + "epoch": 0.25, + "learning_rate": 0.000602752256770311, + "loss": 1.4577, + "theoretical_loss": 3.740320618009898, + "tokens_seen": 838598656 + }, + { + "epoch": 0.25, + "learning_rate": 0.0006027362086258776, + "loss": 1.4235, + "theoretical_loss": 3.7402921083775604, + "tokens_seen": 838664192 + }, + { + "epoch": 0.25, + "learning_rate": 0.0006027201604814444, + "loss": 1.3869, + "theoretical_loss": 3.7402636015967117, + "tokens_seen": 838729728 + }, + { + "epoch": 0.25, + "learning_rate": 0.000602704112337011, + "loss": 1.477, + "theoretical_loss": 3.7402350976668446, + "tokens_seen": 838795264 + }, + { + "epoch": 0.25, + "objective/train/docs_used": 309674, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.621412992477417, + "objective/train/theoretical_loss": 3.7402065965874502, + "objective/train/tokens_used": 859320800, + "theoretical_loss": 3.7402065965874502, + "tokens_seen": 838860800 + }, + { + "epoch": 0.25, + "learning_rate": 0.0006026880641925778, + "loss": 1.5082, + "theoretical_loss": 3.7402065965874502, + "tokens_seen": 838860800 + }, + { + "epoch": 0.25, + "learning_rate": 0.0006026720160481444, + "loss": 1.4436, + "theoretical_loss": 3.7401780983580215, + "tokens_seen": 838926336 + }, + { + "epoch": 0.25, + "learning_rate": 0.0006026559679037112, + "loss": 1.4374, + "theoretical_loss": 3.7401496029780503, + "tokens_seen": 838991872 + }, + { + "epoch": 0.25, + "learning_rate": 0.0006026399197592778, + "loss": 1.3808, + "theoretical_loss": 3.7401211104470296, + "tokens_seen": 839057408 + }, + { + "epoch": 0.25, + "learning_rate": 0.0006026238716148446, + "loss": 1.4245, + "theoretical_loss": 3.740092620764452, + "tokens_seen": 839122944 + }, + { + "epoch": 0.25, + "learning_rate": 0.0006026078234704112, + "loss": 1.5556, + "theoretical_loss": 3.74006413392981, + "tokens_seen": 839188480 + }, + { + "epoch": 0.25, + "learning_rate": 0.000602591775325978, + "loss": 1.4646, + "theoretical_loss": 3.7400356499425973, + "tokens_seen": 839254016 + }, + { + "epoch": 0.25, + "learning_rate": 0.0006025757271815446, + "loss": 1.5681, + "theoretical_loss": 3.7400071688023067, + "tokens_seen": 839319552 + }, + { + "epoch": 0.25, + "learning_rate": 0.0006025596790371114, + "loss": 1.4128, + "theoretical_loss": 3.7399786905084307, + "tokens_seen": 839385088 + }, + { + "epoch": 0.25, + "learning_rate": 0.000602543630892678, + "loss": 1.4952, + "theoretical_loss": 3.739950215060463, + "tokens_seen": 839450624 + }, + { + "epoch": 0.25, + "learning_rate": 0.0006025275827482448, + "loss": 1.4623, + "theoretical_loss": 3.7399217424578977, + "tokens_seen": 839516160 + }, + { + "epoch": 0.25, + "learning_rate": 0.0006025115346038114, + "loss": 1.4262, + "theoretical_loss": 3.739893272700227, + "tokens_seen": 839581696 + }, + { + "epoch": 0.25, + "learning_rate": 0.0006024954864593782, + "loss": 1.5602, + "theoretical_loss": 3.7398648057869464, + "tokens_seen": 839647232 + }, + { + "epoch": 0.25, + "learning_rate": 0.0006024794383149448, + "loss": 1.4643, + "theoretical_loss": 3.7398363417175484, + "tokens_seen": 839712768 + }, + { + "epoch": 0.25, + "learning_rate": 0.0006024633901705116, + "loss": 1.4633, + "theoretical_loss": 3.7398078804915276, + "tokens_seen": 839778304 + }, + { + "epoch": 0.25, + "learning_rate": 0.0006024473420260782, + "loss": 1.4175, + "theoretical_loss": 3.739779422108377, + "tokens_seen": 839843840 + }, + { + "epoch": 0.25, + "learning_rate": 0.000602431293881645, + "loss": 1.4284, + "theoretical_loss": 3.7397509665675925, + "tokens_seen": 839909376 + }, + { + "epoch": 0.25, + "learning_rate": 0.0006024152457372116, + "loss": 1.4337, + "theoretical_loss": 3.7397225138686676, + "tokens_seen": 839974912 + }, + { + "epoch": 0.25, + "learning_rate": 0.0006023991975927784, + "loss": 1.4282, + "theoretical_loss": 3.739694064011097, + "tokens_seen": 840040448 + }, + { + "epoch": 0.25, + "learning_rate": 0.000602383149448345, + "loss": 1.5244, + "theoretical_loss": 3.7396656169943743, + "tokens_seen": 840105984 + }, + { + "epoch": 0.25, + "learning_rate": 0.0006023671013039118, + "loss": 1.5029, + "theoretical_loss": 3.739637172817996, + "tokens_seen": 840171520 + }, + { + "epoch": 0.25, + "learning_rate": 0.0006023510531594784, + "loss": 1.4262, + "theoretical_loss": 3.739608731481456, + "tokens_seen": 840237056 + }, + { + "epoch": 0.25, + "learning_rate": 0.0006023350050150452, + "loss": 1.4772, + "theoretical_loss": 3.7395802929842494, + "tokens_seen": 840302592 + }, + { + "epoch": 0.25, + "learning_rate": 0.0006023189568706118, + "loss": 1.4959, + "theoretical_loss": 3.739551857325871, + "tokens_seen": 840368128 + }, + { + "epoch": 0.25, + "learning_rate": 0.0006023029087261786, + "loss": 1.4774, + "theoretical_loss": 3.7395234245058164, + "tokens_seen": 840433664 + }, + { + "epoch": 0.25, + "learning_rate": 0.0006022868605817452, + "loss": 1.4616, + "theoretical_loss": 3.7394949945235814, + "tokens_seen": 840499200 + }, + { + "epoch": 0.25, + "learning_rate": 0.000602270812437312, + "loss": 1.552, + "theoretical_loss": 3.7394665673786607, + "tokens_seen": 840564736 + }, + { + "epoch": 0.25, + "learning_rate": 0.0006022547642928786, + "loss": 1.5555, + "theoretical_loss": 3.739438143070551, + "tokens_seen": 840630272 + }, + { + "epoch": 0.25, + "learning_rate": 0.0006022387161484454, + "loss": 1.5259, + "theoretical_loss": 3.739409721598747, + "tokens_seen": 840695808 + }, + { + "epoch": 0.25, + "learning_rate": 0.000602222668004012, + "loss": 1.4496, + "theoretical_loss": 3.739381302962745, + "tokens_seen": 840761344 + }, + { + "epoch": 0.25, + "learning_rate": 0.0006022066198595788, + "loss": 1.3817, + "theoretical_loss": 3.739352887162041, + "tokens_seen": 840826880 + }, + { + "epoch": 0.25, + "learning_rate": 0.0006021905717151454, + "loss": 1.5079, + "theoretical_loss": 3.739324474196132, + "tokens_seen": 840892416 + }, + { + "epoch": 0.25, + "learning_rate": 0.0006021745235707122, + "loss": 1.3379, + "theoretical_loss": 3.739296064064513, + "tokens_seen": 840957952 + }, + { + "epoch": 0.25, + "learning_rate": 0.0006021584754262788, + "loss": 1.4984, + "theoretical_loss": 3.739267656766681, + "tokens_seen": 841023488 + }, + { + "epoch": 0.25, + "learning_rate": 0.0006021424272818456, + "loss": 1.5358, + "theoretical_loss": 3.7392392523021325, + "tokens_seen": 841089024 + }, + { + "epoch": 0.25, + "learning_rate": 0.0006021263791374122, + "loss": 1.5058, + "theoretical_loss": 3.7392108506703647, + "tokens_seen": 841154560 + }, + { + "epoch": 0.25, + "learning_rate": 0.000602110330992979, + "loss": 1.509, + "theoretical_loss": 3.7391824518708736, + "tokens_seen": 841220096 + }, + { + "epoch": 0.25, + "learning_rate": 0.0006020942828485457, + "loss": 1.4729, + "theoretical_loss": 3.739154055903157, + "tokens_seen": 841285632 + }, + { + "epoch": 0.25, + "learning_rate": 0.0006020782347041124, + "loss": 1.4584, + "theoretical_loss": 3.7391256627667104, + "tokens_seen": 841351168 + }, + { + "epoch": 0.25, + "learning_rate": 0.000602062186559679, + "loss": 1.5003, + "theoretical_loss": 3.7390972724610334, + "tokens_seen": 841416704 + }, + { + "epoch": 0.25, + "learning_rate": 0.0006020461384152458, + "loss": 1.3882, + "theoretical_loss": 3.7390688849856213, + "tokens_seen": 841482240 + }, + { + "epoch": 0.26, + "learning_rate": 0.0006020300902708125, + "loss": 1.4232, + "theoretical_loss": 3.7390405003399723, + "tokens_seen": 841547776 + }, + { + "epoch": 0.26, + "learning_rate": 0.0006020140421263792, + "loss": 1.4606, + "theoretical_loss": 3.7390121185235836, + "tokens_seen": 841613312 + }, + { + "epoch": 0.26, + "learning_rate": 0.0006019979939819459, + "loss": 1.4788, + "theoretical_loss": 3.7389837395359535, + "tokens_seen": 841678848 + }, + { + "epoch": 0.26, + "learning_rate": 0.0006019819458375126, + "loss": 1.5072, + "theoretical_loss": 3.7389553633765797, + "tokens_seen": 841744384 + }, + { + "epoch": 0.26, + "learning_rate": 0.0006019658976930793, + "loss": 1.4344, + "theoretical_loss": 3.73892699004496, + "tokens_seen": 841809920 + }, + { + "epoch": 0.26, + "learning_rate": 0.000601949849548646, + "loss": 1.4142, + "theoretical_loss": 3.7388986195405924, + "tokens_seen": 841875456 + }, + { + "epoch": 0.26, + "learning_rate": 0.0006019338014042127, + "loss": 1.3748, + "theoretical_loss": 3.7388702518629753, + "tokens_seen": 841940992 + }, + { + "epoch": 0.26, + "learning_rate": 0.0006019177532597794, + "loss": 1.4906, + "theoretical_loss": 3.738841887011607, + "tokens_seen": 842006528 + }, + { + "epoch": 0.26, + "learning_rate": 0.0006019017051153461, + "loss": 1.4892, + "theoretical_loss": 3.7388135249859866, + "tokens_seen": 842072064 + }, + { + "epoch": 0.26, + "objective/train/docs_used": 310992, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.3815665245056152, + "objective/train/theoretical_loss": 3.738785165785611, + "objective/train/tokens_used": 862597600, + "theoretical_loss": 3.738785165785611, + "tokens_seen": 842137600 + }, + { + "epoch": 0.26, + "learning_rate": 0.0006018856569709128, + "loss": 1.4703, + "theoretical_loss": 3.738785165785611, + "tokens_seen": 842137600 + }, + { + "epoch": 0.26, + "learning_rate": 0.0006018696088264795, + "loss": 1.433, + "theoretical_loss": 3.7387568094099812, + "tokens_seen": 842203136 + }, + { + "epoch": 0.26, + "learning_rate": 0.0006018535606820462, + "loss": 1.5434, + "theoretical_loss": 3.738728455858594, + "tokens_seen": 842268672 + }, + { + "epoch": 0.26, + "learning_rate": 0.0006018375125376129, + "loss": 1.514, + "theoretical_loss": 3.73870010513095, + "tokens_seen": 842334208 + }, + { + "epoch": 0.26, + "learning_rate": 0.0006018214643931796, + "loss": 1.5266, + "theoretical_loss": 3.7386717572265473, + "tokens_seen": 842399744 + }, + { + "epoch": 0.26, + "learning_rate": 0.0006018054162487463, + "loss": 1.5467, + "theoretical_loss": 3.738643412144885, + "tokens_seen": 842465280 + }, + { + "epoch": 0.26, + "learning_rate": 0.000601789368104313, + "loss": 1.5381, + "theoretical_loss": 3.738615069885464, + "tokens_seen": 842530816 + }, + { + "epoch": 0.26, + "learning_rate": 0.0006017733199598797, + "loss": 1.4789, + "theoretical_loss": 3.738586730447782, + "tokens_seen": 842596352 + }, + { + "epoch": 0.26, + "learning_rate": 0.0006017572718154464, + "loss": 1.4396, + "theoretical_loss": 3.7385583938313394, + "tokens_seen": 842661888 + }, + { + "epoch": 0.26, + "learning_rate": 0.0006017412236710131, + "loss": 1.458, + "theoretical_loss": 3.738530060035636, + "tokens_seen": 842727424 + }, + { + "epoch": 0.26, + "learning_rate": 0.0006017251755265798, + "loss": 1.5016, + "theoretical_loss": 3.7385017290601716, + "tokens_seen": 842792960 + }, + { + "epoch": 0.26, + "learning_rate": 0.0006017091273821465, + "loss": 1.4424, + "theoretical_loss": 3.7384734009044465, + "tokens_seen": 842858496 + }, + { + "epoch": 0.26, + "learning_rate": 0.0006016930792377131, + "loss": 1.5213, + "theoretical_loss": 3.7384450755679604, + "tokens_seen": 842924032 + }, + { + "epoch": 0.26, + "learning_rate": 0.0006016770310932799, + "loss": 1.4834, + "theoretical_loss": 3.7384167530502133, + "tokens_seen": 842989568 + }, + { + "epoch": 0.26, + "learning_rate": 0.0006016609829488465, + "loss": 1.4695, + "theoretical_loss": 3.738388433350707, + "tokens_seen": 843055104 + }, + { + "epoch": 0.26, + "learning_rate": 0.0006016449348044133, + "loss": 1.3849, + "theoretical_loss": 3.7383601164689404, + "tokens_seen": 843120640 + }, + { + "epoch": 0.26, + "learning_rate": 0.0006016288866599799, + "loss": 1.4739, + "theoretical_loss": 3.7383318024044145, + "tokens_seen": 843186176 + }, + { + "epoch": 0.26, + "learning_rate": 0.0006016128385155467, + "loss": 1.5251, + "theoretical_loss": 3.738303491156631, + "tokens_seen": 843251712 + }, + { + "epoch": 0.26, + "learning_rate": 0.0006015967903711133, + "loss": 1.4973, + "theoretical_loss": 3.73827518272509, + "tokens_seen": 843317248 + }, + { + "epoch": 0.26, + "learning_rate": 0.0006015807422266801, + "loss": 1.5373, + "theoretical_loss": 3.738246877109293, + "tokens_seen": 843382784 + }, + { + "epoch": 0.26, + "learning_rate": 0.0006015646940822467, + "loss": 1.4399, + "theoretical_loss": 3.7382185743087404, + "tokens_seen": 843448320 + }, + { + "epoch": 0.26, + "learning_rate": 0.0006015486459378135, + "loss": 1.5587, + "theoretical_loss": 3.7381902743229345, + "tokens_seen": 843513856 + }, + { + "epoch": 0.26, + "learning_rate": 0.0006015325977933801, + "loss": 1.5544, + "theoretical_loss": 3.7381619771513757, + "tokens_seen": 843579392 + }, + { + "epoch": 0.26, + "learning_rate": 0.0006015165496489469, + "loss": 1.4431, + "theoretical_loss": 3.738133682793566, + "tokens_seen": 843644928 + }, + { + "epoch": 0.26, + "learning_rate": 0.0006015005015045135, + "loss": 1.5394, + "theoretical_loss": 3.7381053912490074, + "tokens_seen": 843710464 + }, + { + "epoch": 0.26, + "learning_rate": 0.0006014844533600803, + "loss": 1.5008, + "theoretical_loss": 3.7380771025172015, + "tokens_seen": 843776000 + }, + { + "epoch": 0.26, + "learning_rate": 0.0006014684052156469, + "loss": 1.4312, + "theoretical_loss": 3.73804881659765, + "tokens_seen": 843841536 + }, + { + "epoch": 0.26, + "learning_rate": 0.0006014523570712137, + "loss": 1.4754, + "theoretical_loss": 3.738020533489855, + "tokens_seen": 843907072 + }, + { + "epoch": 0.26, + "learning_rate": 0.0006014363089267803, + "loss": 1.4986, + "theoretical_loss": 3.737992253193319, + "tokens_seen": 843972608 + }, + { + "epoch": 0.26, + "learning_rate": 0.0006014202607823471, + "loss": 1.4709, + "theoretical_loss": 3.737963975707544, + "tokens_seen": 844038144 + }, + { + "epoch": 0.26, + "learning_rate": 0.0006014042126379137, + "loss": 1.4735, + "theoretical_loss": 3.7379357010320318, + "tokens_seen": 844103680 + }, + { + "epoch": 0.26, + "learning_rate": 0.0006013881644934805, + "loss": 1.3998, + "theoretical_loss": 3.7379074291662864, + "tokens_seen": 844169216 + }, + { + "epoch": 0.26, + "learning_rate": 0.0006013721163490471, + "loss": 1.4208, + "theoretical_loss": 3.7378791601098094, + "tokens_seen": 844234752 + }, + { + "epoch": 0.26, + "learning_rate": 0.0006013560682046139, + "loss": 1.4355, + "theoretical_loss": 3.737850893862104, + "tokens_seen": 844300288 + }, + { + "epoch": 0.26, + "learning_rate": 0.0006013400200601805, + "loss": 1.5184, + "theoretical_loss": 3.7378226304226727, + "tokens_seen": 844365824 + }, + { + "epoch": 0.26, + "learning_rate": 0.0006013239719157473, + "loss": 1.3873, + "theoretical_loss": 3.737794369791019, + "tokens_seen": 844431360 + }, + { + "epoch": 0.26, + "learning_rate": 0.0006013079237713139, + "loss": 1.5141, + "theoretical_loss": 3.7377661119666463, + "tokens_seen": 844496896 + }, + { + "epoch": 0.26, + "learning_rate": 0.0006012918756268807, + "loss": 1.6177, + "theoretical_loss": 3.7377378569490576, + "tokens_seen": 844562432 + }, + { + "epoch": 0.26, + "learning_rate": 0.0006012758274824473, + "loss": 1.4654, + "theoretical_loss": 3.7377096047377565, + "tokens_seen": 844627968 + }, + { + "epoch": 0.26, + "learning_rate": 0.0006012597793380141, + "loss": 1.4632, + "theoretical_loss": 3.7376813553322465, + "tokens_seen": 844693504 + }, + { + "epoch": 0.26, + "learning_rate": 0.0006012437311935807, + "loss": 1.5655, + "theoretical_loss": 3.737653108732031, + "tokens_seen": 844759040 + }, + { + "epoch": 0.26, + "learning_rate": 0.0006012276830491475, + "loss": 1.558, + "theoretical_loss": 3.7376248649366137, + "tokens_seen": 844824576 + }, + { + "epoch": 0.26, + "learning_rate": 0.0006012116349047141, + "loss": 1.4022, + "theoretical_loss": 3.737596623945499, + "tokens_seen": 844890112 + }, + { + "epoch": 0.26, + "learning_rate": 0.0006011955867602809, + "loss": 1.4346, + "theoretical_loss": 3.737568385758191, + "tokens_seen": 844955648 + }, + { + "epoch": 0.26, + "learning_rate": 0.0006011795386158475, + "loss": 1.4601, + "theoretical_loss": 3.737540150374194, + "tokens_seen": 845021184 + }, + { + "epoch": 0.26, + "learning_rate": 0.0006011634904714143, + "loss": 1.5145, + "theoretical_loss": 3.737511917793012, + "tokens_seen": 845086720 + }, + { + "epoch": 0.26, + "learning_rate": 0.0006011474423269809, + "loss": 1.4415, + "theoretical_loss": 3.7374836880141493, + "tokens_seen": 845152256 + }, + { + "epoch": 0.26, + "learning_rate": 0.0006011313941825477, + "loss": 1.5245, + "theoretical_loss": 3.737455461037111, + "tokens_seen": 845217792 + }, + { + "epoch": 0.26, + "learning_rate": 0.0006011153460381143, + "loss": 1.47, + "theoretical_loss": 3.737427236861401, + "tokens_seen": 845283328 + }, + { + "epoch": 0.26, + "learning_rate": 0.0006010992978936811, + "loss": 1.4509, + "theoretical_loss": 3.7373990154865253, + "tokens_seen": 845348864 + }, + { + "epoch": 0.26, + "objective/train/docs_used": 312167, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.4518039226531982, + "objective/train/theoretical_loss": 3.7373707969119883, + "objective/train/tokens_used": 865874400, + "theoretical_loss": 3.7373707969119883, + "tokens_seen": 845414400 + }, + { + "epoch": 0.26, + "learning_rate": 0.0006010832497492477, + "loss": 1.4453, + "theoretical_loss": 3.7373707969119883, + "tokens_seen": 845414400 + }, + { + "epoch": 0.26, + "learning_rate": 0.0006010672016048145, + "loss": 1.5489, + "theoretical_loss": 3.7373425811372942, + "tokens_seen": 845479936 + }, + { + "epoch": 0.26, + "learning_rate": 0.0006010511534603811, + "loss": 1.5819, + "theoretical_loss": 3.7373143681619494, + "tokens_seen": 845545472 + }, + { + "epoch": 0.26, + "learning_rate": 0.0006010351053159479, + "loss": 1.4124, + "theoretical_loss": 3.737286157985459, + "tokens_seen": 845611008 + }, + { + "epoch": 0.26, + "learning_rate": 0.0006010190571715145, + "loss": 1.508, + "theoretical_loss": 3.737257950607328, + "tokens_seen": 845676544 + }, + { + "epoch": 0.26, + "learning_rate": 0.0006010030090270813, + "loss": 1.4772, + "theoretical_loss": 3.7372297460270625, + "tokens_seen": 845742080 + }, + { + "epoch": 0.26, + "learning_rate": 0.0006009869608826479, + "loss": 1.5203, + "theoretical_loss": 3.7372015442441677, + "tokens_seen": 845807616 + }, + { + "epoch": 0.26, + "learning_rate": 0.0006009709127382147, + "loss": 1.4181, + "theoretical_loss": 3.7371733452581504, + "tokens_seen": 845873152 + }, + { + "epoch": 0.26, + "learning_rate": 0.0006009548645937813, + "loss": 1.367, + "theoretical_loss": 3.737145149068515, + "tokens_seen": 845938688 + }, + { + "epoch": 0.26, + "learning_rate": 0.0006009388164493481, + "loss": 1.4866, + "theoretical_loss": 3.737116955674769, + "tokens_seen": 846004224 + }, + { + "epoch": 0.26, + "learning_rate": 0.0006009227683049147, + "loss": 1.441, + "theoretical_loss": 3.7370887650764177, + "tokens_seen": 846069760 + }, + { + "epoch": 0.26, + "learning_rate": 0.0006009067201604815, + "loss": 1.4657, + "theoretical_loss": 3.7370605772729677, + "tokens_seen": 846135296 + }, + { + "epoch": 0.26, + "learning_rate": 0.0006008906720160481, + "loss": 1.447, + "theoretical_loss": 3.7370323922639264, + "tokens_seen": 846200832 + }, + { + "epoch": 0.26, + "learning_rate": 0.0006008746238716149, + "loss": 1.3775, + "theoretical_loss": 3.7370042100487986, + "tokens_seen": 846266368 + }, + { + "epoch": 0.26, + "learning_rate": 0.0006008585757271815, + "loss": 1.4951, + "theoretical_loss": 3.7369760306270923, + "tokens_seen": 846331904 + }, + { + "epoch": 0.26, + "learning_rate": 0.0006008425275827483, + "loss": 1.3815, + "theoretical_loss": 3.7369478539983145, + "tokens_seen": 846397440 + }, + { + "epoch": 0.26, + "learning_rate": 0.000600826479438315, + "loss": 1.4259, + "theoretical_loss": 3.736919680161971, + "tokens_seen": 846462976 + }, + { + "epoch": 0.26, + "learning_rate": 0.0006008104312938817, + "loss": 1.4782, + "theoretical_loss": 3.7368915091175703, + "tokens_seen": 846528512 + }, + { + "epoch": 0.26, + "learning_rate": 0.0006007943831494483, + "loss": 1.494, + "theoretical_loss": 3.736863340864618, + "tokens_seen": 846594048 + }, + { + "epoch": 0.26, + "learning_rate": 0.0006007783350050151, + "loss": 1.5657, + "theoretical_loss": 3.7368351754026228, + "tokens_seen": 846659584 + }, + { + "epoch": 0.26, + "learning_rate": 0.0006007622868605817, + "loss": 1.4489, + "theoretical_loss": 3.7368070127310915, + "tokens_seen": 846725120 + }, + { + "epoch": 0.26, + "learning_rate": 0.0006007462387161485, + "loss": 1.4465, + "theoretical_loss": 3.736778852849532, + "tokens_seen": 846790656 + }, + { + "epoch": 0.26, + "learning_rate": 0.0006007301905717151, + "loss": 1.5255, + "theoretical_loss": 3.736750695757452, + "tokens_seen": 846856192 + }, + { + "epoch": 0.26, + "learning_rate": 0.0006007141424272819, + "loss": 1.449, + "theoretical_loss": 3.736722541454359, + "tokens_seen": 846921728 + }, + { + "epoch": 0.26, + "learning_rate": 0.0006006980942828486, + "loss": 1.5117, + "theoretical_loss": 3.7366943899397613, + "tokens_seen": 846987264 + }, + { + "epoch": 0.26, + "learning_rate": 0.0006006820461384153, + "loss": 1.3823, + "theoretical_loss": 3.736666241213167, + "tokens_seen": 847052800 + }, + { + "epoch": 0.26, + "learning_rate": 0.000600665997993982, + "loss": 1.5107, + "theoretical_loss": 3.736638095274084, + "tokens_seen": 847118336 + }, + { + "epoch": 0.26, + "learning_rate": 0.0006006499498495487, + "loss": 1.4681, + "theoretical_loss": 3.736609952122021, + "tokens_seen": 847183872 + }, + { + "epoch": 0.26, + "learning_rate": 0.0006006339017051154, + "loss": 1.3739, + "theoretical_loss": 3.736581811756486, + "tokens_seen": 847249408 + }, + { + "epoch": 0.26, + "learning_rate": 0.0006006178535606821, + "loss": 1.367, + "theoretical_loss": 3.736553674176988, + "tokens_seen": 847314944 + }, + { + "epoch": 0.26, + "learning_rate": 0.0006006018054162488, + "loss": 1.4583, + "theoretical_loss": 3.736525539383036, + "tokens_seen": 847380480 + }, + { + "epoch": 0.26, + "learning_rate": 0.0006005857572718155, + "loss": 1.365, + "theoretical_loss": 3.736497407374138, + "tokens_seen": 847446016 + }, + { + "epoch": 0.26, + "learning_rate": 0.0006005697091273823, + "loss": 1.4183, + "theoretical_loss": 3.7364692781498032, + "tokens_seen": 847511552 + }, + { + "epoch": 0.26, + "learning_rate": 0.0006005536609829489, + "loss": 1.425, + "theoretical_loss": 3.7364411517095415, + "tokens_seen": 847577088 + }, + { + "epoch": 0.26, + "learning_rate": 0.0006005376128385157, + "loss": 1.539, + "theoretical_loss": 3.736413028052861, + "tokens_seen": 847642624 + }, + { + "epoch": 0.26, + "learning_rate": 0.0006005215646940823, + "loss": 1.4855, + "theoretical_loss": 3.736384907179272, + "tokens_seen": 847708160 + }, + { + "epoch": 0.26, + "learning_rate": 0.0006005055165496491, + "loss": 1.477, + "theoretical_loss": 3.736356789088283, + "tokens_seen": 847773696 + }, + { + "epoch": 0.26, + "learning_rate": 0.0006004894684052157, + "loss": 1.4104, + "theoretical_loss": 3.7363286737794046, + "tokens_seen": 847839232 + }, + { + "epoch": 0.26, + "learning_rate": 0.0006004734202607825, + "loss": 1.5518, + "theoretical_loss": 3.7363005612521456, + "tokens_seen": 847904768 + }, + { + "epoch": 0.26, + "learning_rate": 0.0006004573721163491, + "loss": 1.42, + "theoretical_loss": 3.736272451506016, + "tokens_seen": 847970304 + }, + { + "epoch": 0.26, + "learning_rate": 0.0006004413239719158, + "loss": 1.486, + "theoretical_loss": 3.7362443445405265, + "tokens_seen": 848035840 + }, + { + "epoch": 0.26, + "learning_rate": 0.0006004252758274825, + "loss": 1.444, + "theoretical_loss": 3.736216240355186, + "tokens_seen": 848101376 + }, + { + "epoch": 0.26, + "learning_rate": 0.0006004092276830492, + "loss": 1.5435, + "theoretical_loss": 3.7361881389495064, + "tokens_seen": 848166912 + }, + { + "epoch": 0.26, + "learning_rate": 0.0006003931795386159, + "loss": 1.5424, + "theoretical_loss": 3.736160040322996, + "tokens_seen": 848232448 + }, + { + "epoch": 0.26, + "learning_rate": 0.0006003771313941826, + "loss": 1.4149, + "theoretical_loss": 3.736131944475167, + "tokens_seen": 848297984 + }, + { + "epoch": 0.26, + "learning_rate": 0.0006003610832497493, + "loss": 1.5394, + "theoretical_loss": 3.7361038514055287, + "tokens_seen": 848363520 + }, + { + "epoch": 0.26, + "learning_rate": 0.000600345035105316, + "loss": 1.4054, + "theoretical_loss": 3.7360757611135926, + "tokens_seen": 848429056 + }, + { + "epoch": 0.26, + "learning_rate": 0.0006003289869608827, + "loss": 1.3625, + "theoretical_loss": 3.7360476735988692, + "tokens_seen": 848494592 + }, + { + "epoch": 0.26, + "learning_rate": 0.0006003129388164494, + "loss": 1.4341, + "theoretical_loss": 3.736019588860869, + "tokens_seen": 848560128 + }, + { + "epoch": 0.26, + "learning_rate": 0.000600296890672016, + "loss": 1.4772, + "theoretical_loss": 3.7359915068991043, + "tokens_seen": 848625664 + }, + { + "epoch": 0.26, + "objective/train/docs_used": 313372, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.5092430114746094, + "objective/train/theoretical_loss": 3.7359634277130853, + "objective/train/tokens_used": 869151200, + "theoretical_loss": 3.7359634277130853, + "tokens_seen": 848691200 + }, + { + "epoch": 0.26, + "learning_rate": 0.0006002808425275828, + "loss": 1.4609, + "theoretical_loss": 3.7359634277130853, + "tokens_seen": 848691200 + }, + { + "epoch": 0.26, + "learning_rate": 0.0006002647943831494, + "loss": 1.5764, + "theoretical_loss": 3.7359353513023237, + "tokens_seen": 848756736 + }, + { + "epoch": 0.26, + "learning_rate": 0.0006002487462387162, + "loss": 1.4245, + "theoretical_loss": 3.7359072776663305, + "tokens_seen": 848822272 + }, + { + "epoch": 0.26, + "learning_rate": 0.0006002326980942828, + "loss": 1.3999, + "theoretical_loss": 3.735879206804618, + "tokens_seen": 848887808 + }, + { + "epoch": 0.26, + "learning_rate": 0.0006002166499498496, + "loss": 1.4999, + "theoretical_loss": 3.735851138716697, + "tokens_seen": 848953344 + }, + { + "epoch": 0.26, + "learning_rate": 0.0006002006018054162, + "loss": 1.4781, + "theoretical_loss": 3.7358230734020803, + "tokens_seen": 849018880 + }, + { + "epoch": 0.26, + "learning_rate": 0.000600184553660983, + "loss": 1.3721, + "theoretical_loss": 3.7357950108602793, + "tokens_seen": 849084416 + }, + { + "epoch": 0.26, + "learning_rate": 0.0006001685055165496, + "loss": 1.4345, + "theoretical_loss": 3.735766951090806, + "tokens_seen": 849149952 + }, + { + "epoch": 0.26, + "learning_rate": 0.0006001524573721164, + "loss": 1.4935, + "theoretical_loss": 3.735738894093173, + "tokens_seen": 849215488 + }, + { + "epoch": 0.26, + "learning_rate": 0.000600136409227683, + "loss": 1.6127, + "theoretical_loss": 3.7357108398668917, + "tokens_seen": 849281024 + }, + { + "epoch": 0.26, + "learning_rate": 0.0006001203610832498, + "loss": 1.5106, + "theoretical_loss": 3.7356827884114754, + "tokens_seen": 849346560 + }, + { + "epoch": 0.26, + "learning_rate": 0.0006001043129388164, + "loss": 1.6439, + "theoretical_loss": 3.735654739726437, + "tokens_seen": 849412096 + }, + { + "epoch": 0.26, + "learning_rate": 0.0006000882647943832, + "loss": 1.4787, + "theoretical_loss": 3.735626693811288, + "tokens_seen": 849477632 + }, + { + "epoch": 0.26, + "learning_rate": 0.0006000722166499498, + "loss": 1.5005, + "theoretical_loss": 3.7355986506655414, + "tokens_seen": 849543168 + }, + { + "epoch": 0.26, + "learning_rate": 0.0006000561685055166, + "loss": 1.4007, + "theoretical_loss": 3.735570610288711, + "tokens_seen": 849608704 + }, + { + "epoch": 0.26, + "learning_rate": 0.0006000401203610832, + "loss": 1.5498, + "theoretical_loss": 3.73554257268031, + "tokens_seen": 849674240 + }, + { + "epoch": 0.26, + "learning_rate": 0.00060002407221665, + "loss": 1.3952, + "theoretical_loss": 3.73551453783985, + "tokens_seen": 849739776 + }, + { + "epoch": 0.26, + "learning_rate": 0.0006000080240722166, + "loss": 1.3701, + "theoretical_loss": 3.7354865057668456, + "tokens_seen": 849805312 + }, + { + "epoch": 0.26, + "learning_rate": 0.0005999919759277834, + "loss": 1.4946, + "theoretical_loss": 3.73545847646081, + "tokens_seen": 849870848 + }, + { + "epoch": 0.26, + "learning_rate": 0.00059997592778335, + "loss": 1.4101, + "theoretical_loss": 3.735430449921256, + "tokens_seen": 849936384 + }, + { + "epoch": 0.26, + "learning_rate": 0.0005999598796389168, + "loss": 1.3588, + "theoretical_loss": 3.735402426147698, + "tokens_seen": 850001920 + }, + { + "epoch": 0.26, + "learning_rate": 0.0005999438314944834, + "loss": 1.3697, + "theoretical_loss": 3.7353744051396505, + "tokens_seen": 850067456 + }, + { + "epoch": 0.26, + "learning_rate": 0.0005999277833500502, + "loss": 1.3754, + "theoretical_loss": 3.7353463868966257, + "tokens_seen": 850132992 + }, + { + "epoch": 0.26, + "learning_rate": 0.0005999117352056168, + "loss": 1.5069, + "theoretical_loss": 3.7353183714181393, + "tokens_seen": 850198528 + }, + { + "epoch": 0.26, + "learning_rate": 0.0005998956870611836, + "loss": 1.5812, + "theoretical_loss": 3.735290358703704, + "tokens_seen": 850264064 + }, + { + "epoch": 0.26, + "learning_rate": 0.0005998796389167502, + "loss": 1.4138, + "theoretical_loss": 3.735262348752835, + "tokens_seen": 850329600 + }, + { + "epoch": 0.26, + "learning_rate": 0.000599863590772317, + "loss": 1.5197, + "theoretical_loss": 3.7352343415650466, + "tokens_seen": 850395136 + }, + { + "epoch": 0.26, + "learning_rate": 0.0005998475426278836, + "loss": 1.5461, + "theoretical_loss": 3.7352063371398527, + "tokens_seen": 850460672 + }, + { + "epoch": 0.26, + "learning_rate": 0.0005998314944834504, + "loss": 1.446, + "theoretical_loss": 3.7351783354767685, + "tokens_seen": 850526208 + }, + { + "epoch": 0.26, + "learning_rate": 0.000599815446339017, + "loss": 1.4493, + "theoretical_loss": 3.735150336575309, + "tokens_seen": 850591744 + }, + { + "epoch": 0.26, + "learning_rate": 0.0005997993981945838, + "loss": 1.4983, + "theoretical_loss": 3.7351223404349887, + "tokens_seen": 850657280 + }, + { + "epoch": 0.26, + "learning_rate": 0.0005997833500501504, + "loss": 1.5546, + "theoretical_loss": 3.7350943470553224, + "tokens_seen": 850722816 + }, + { + "epoch": 0.26, + "learning_rate": 0.0005997673019057172, + "loss": 1.3637, + "theoretical_loss": 3.735066356435826, + "tokens_seen": 850788352 + }, + { + "epoch": 0.26, + "learning_rate": 0.0005997512537612838, + "loss": 1.6033, + "theoretical_loss": 3.735038368576014, + "tokens_seen": 850853888 + }, + { + "epoch": 0.26, + "learning_rate": 0.0005997352056168506, + "loss": 1.5713, + "theoretical_loss": 3.7350103834754025, + "tokens_seen": 850919424 + }, + { + "epoch": 0.26, + "learning_rate": 0.0005997191574724172, + "loss": 1.4979, + "theoretical_loss": 3.734982401133506, + "tokens_seen": 850984960 + }, + { + "epoch": 0.26, + "learning_rate": 0.000599703109327984, + "loss": 1.4813, + "theoretical_loss": 3.734954421549841, + "tokens_seen": 851050496 + }, + { + "epoch": 0.26, + "learning_rate": 0.0005996870611835506, + "loss": 1.3153, + "theoretical_loss": 3.734926444723923, + "tokens_seen": 851116032 + }, + { + "epoch": 0.26, + "learning_rate": 0.0005996710130391174, + "loss": 1.405, + "theoretical_loss": 3.734898470655268, + "tokens_seen": 851181568 + }, + { + "epoch": 0.26, + "learning_rate": 0.000599654964894684, + "loss": 1.3952, + "theoretical_loss": 3.7348704993433914, + "tokens_seen": 851247104 + }, + { + "epoch": 0.26, + "learning_rate": 0.0005996389167502508, + "loss": 1.4892, + "theoretical_loss": 3.7348425307878097, + "tokens_seen": 851312640 + }, + { + "epoch": 0.26, + "learning_rate": 0.0005996228686058174, + "loss": 1.4525, + "theoretical_loss": 3.7348145649880395, + "tokens_seen": 851378176 + }, + { + "epoch": 0.26, + "learning_rate": 0.0005996068204613842, + "loss": 1.4559, + "theoretical_loss": 3.734786601943597, + "tokens_seen": 851443712 + }, + { + "epoch": 0.26, + "learning_rate": 0.0005995907723169508, + "loss": 1.3134, + "theoretical_loss": 3.7347586416539986, + "tokens_seen": 851509248 + }, + { + "epoch": 0.26, + "learning_rate": 0.0005995747241725176, + "loss": 1.4456, + "theoretical_loss": 3.7347306841187606, + "tokens_seen": 851574784 + }, + { + "epoch": 0.26, + "learning_rate": 0.0005995586760280842, + "loss": 1.5568, + "theoretical_loss": 3.7347027293373998, + "tokens_seen": 851640320 + }, + { + "epoch": 0.26, + "learning_rate": 0.000599542627883651, + "loss": 1.4359, + "theoretical_loss": 3.7346747773094338, + "tokens_seen": 851705856 + }, + { + "epoch": 0.26, + "learning_rate": 0.0005995265797392176, + "loss": 1.4763, + "theoretical_loss": 3.7346468280343785, + "tokens_seen": 851771392 + }, + { + "epoch": 0.26, + "learning_rate": 0.0005995105315947844, + "loss": 1.4177, + "theoretical_loss": 3.734618881511752, + "tokens_seen": 851836928 + }, + { + "epoch": 0.26, + "learning_rate": 0.000599494483450351, + "loss": 1.4229, + "theoretical_loss": 3.7345909377410704, + "tokens_seen": 851902464 + }, + { + "epoch": 0.26, + "objective/train/docs_used": 314587, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.4946930408477783, + "objective/train/theoretical_loss": 3.7345629967218517, + "objective/train/tokens_used": 872428000, + "theoretical_loss": 3.7345629967218517, + "tokens_seen": 851968000 + }, + { + "epoch": 0.26, + "learning_rate": 0.0005994784353059178, + "loss": 1.4802, + "theoretical_loss": 3.7345629967218517, + "tokens_seen": 851968000 + }, + { + "epoch": 0.26, + "learning_rate": 0.0005994623871614844, + "loss": 1.4836, + "theoretical_loss": 3.7345350584536137, + "tokens_seen": 852033536 + }, + { + "epoch": 0.26, + "learning_rate": 0.0005994463390170512, + "loss": 1.5834, + "theoretical_loss": 3.7345071229358733, + "tokens_seen": 852099072 + }, + { + "epoch": 0.26, + "learning_rate": 0.0005994302908726178, + "loss": 1.4989, + "theoretical_loss": 3.7344791901681487, + "tokens_seen": 852164608 + }, + { + "epoch": 0.26, + "learning_rate": 0.0005994142427281846, + "loss": 1.443, + "theoretical_loss": 3.734451260149957, + "tokens_seen": 852230144 + }, + { + "epoch": 0.26, + "learning_rate": 0.0005993981945837514, + "loss": 1.4848, + "theoretical_loss": 3.7344233328808176, + "tokens_seen": 852295680 + }, + { + "epoch": 0.26, + "learning_rate": 0.000599382146439318, + "loss": 1.4762, + "theoretical_loss": 3.7343954083602466, + "tokens_seen": 852361216 + }, + { + "epoch": 0.26, + "learning_rate": 0.0005993660982948848, + "loss": 1.5405, + "theoretical_loss": 3.734367486587764, + "tokens_seen": 852426752 + }, + { + "epoch": 0.26, + "learning_rate": 0.0005993500501504514, + "loss": 1.454, + "theoretical_loss": 3.7343395675628868, + "tokens_seen": 852492288 + }, + { + "epoch": 0.26, + "learning_rate": 0.0005993340020060182, + "loss": 1.414, + "theoretical_loss": 3.734311651285134, + "tokens_seen": 852557824 + }, + { + "epoch": 0.26, + "learning_rate": 0.0005993179538615848, + "loss": 1.4294, + "theoretical_loss": 3.7342837377540246, + "tokens_seen": 852623360 + }, + { + "epoch": 0.26, + "learning_rate": 0.0005993019057171516, + "loss": 1.4726, + "theoretical_loss": 3.7342558269690764, + "tokens_seen": 852688896 + }, + { + "epoch": 0.26, + "learning_rate": 0.0005992858575727182, + "loss": 1.4057, + "theoretical_loss": 3.7342279189298093, + "tokens_seen": 852754432 + }, + { + "epoch": 0.26, + "learning_rate": 0.000599269809428285, + "loss": 1.5854, + "theoretical_loss": 3.7342000136357405, + "tokens_seen": 852819968 + }, + { + "epoch": 0.26, + "learning_rate": 0.0005992537612838516, + "loss": 1.4837, + "theoretical_loss": 3.734172111086391, + "tokens_seen": 852885504 + }, + { + "epoch": 0.26, + "learning_rate": 0.0005992377131394184, + "loss": 1.5044, + "theoretical_loss": 3.7341442112812784, + "tokens_seen": 852951040 + }, + { + "epoch": 0.26, + "learning_rate": 0.000599221664994985, + "loss": 1.5012, + "theoretical_loss": 3.7341163142199223, + "tokens_seen": 853016576 + }, + { + "epoch": 0.26, + "learning_rate": 0.0005992056168505518, + "loss": 1.4351, + "theoretical_loss": 3.734088419901843, + "tokens_seen": 853082112 + }, + { + "epoch": 0.26, + "learning_rate": 0.0005991895687061184, + "loss": 1.5842, + "theoretical_loss": 3.734060528326559, + "tokens_seen": 853147648 + }, + { + "epoch": 0.26, + "learning_rate": 0.0005991735205616852, + "loss": 1.3836, + "theoretical_loss": 3.7340326394935905, + "tokens_seen": 853213184 + }, + { + "epoch": 0.26, + "learning_rate": 0.0005991574724172518, + "loss": 1.4637, + "theoretical_loss": 3.734004753402457, + "tokens_seen": 853278720 + }, + { + "epoch": 0.26, + "learning_rate": 0.0005991414242728186, + "loss": 1.4354, + "theoretical_loss": 3.7339768700526785, + "tokens_seen": 853344256 + }, + { + "epoch": 0.26, + "learning_rate": 0.0005991253761283852, + "loss": 1.5037, + "theoretical_loss": 3.7339489894437756, + "tokens_seen": 853409792 + }, + { + "epoch": 0.26, + "learning_rate": 0.000599109327983952, + "loss": 1.484, + "theoretical_loss": 3.733921111575267, + "tokens_seen": 853475328 + }, + { + "epoch": 0.26, + "learning_rate": 0.0005990932798395186, + "loss": 1.5076, + "theoretical_loss": 3.7338932364466744, + "tokens_seen": 853540864 + }, + { + "epoch": 0.26, + "learning_rate": 0.0005990772316950854, + "loss": 1.3949, + "theoretical_loss": 3.7338653640575172, + "tokens_seen": 853606400 + }, + { + "epoch": 0.26, + "learning_rate": 0.000599061183550652, + "loss": 1.4348, + "theoretical_loss": 3.7338374944073163, + "tokens_seen": 853671936 + }, + { + "epoch": 0.26, + "learning_rate": 0.0005990451354062187, + "loss": 1.4091, + "theoretical_loss": 3.7338096274955923, + "tokens_seen": 853737472 + }, + { + "epoch": 0.26, + "learning_rate": 0.0005990290872617854, + "loss": 1.5537, + "theoretical_loss": 3.733781763321866, + "tokens_seen": 853803008 + }, + { + "epoch": 0.26, + "learning_rate": 0.0005990130391173521, + "loss": 1.5403, + "theoretical_loss": 3.7337539018856587, + "tokens_seen": 853868544 + }, + { + "epoch": 0.26, + "learning_rate": 0.0005989969909729188, + "loss": 1.5131, + "theoretical_loss": 3.73372604318649, + "tokens_seen": 853934080 + }, + { + "epoch": 0.26, + "learning_rate": 0.0005989809428284855, + "loss": 1.5583, + "theoretical_loss": 3.733698187223882, + "tokens_seen": 853999616 + }, + { + "epoch": 0.26, + "learning_rate": 0.0005989648946840522, + "loss": 1.4407, + "theoretical_loss": 3.733670333997356, + "tokens_seen": 854065152 + }, + { + "epoch": 0.26, + "learning_rate": 0.0005989488465396189, + "loss": 1.5386, + "theoretical_loss": 3.733642483506433, + "tokens_seen": 854130688 + }, + { + "epoch": 0.26, + "learning_rate": 0.0005989327983951856, + "loss": 1.395, + "theoretical_loss": 3.7336146357506337, + "tokens_seen": 854196224 + }, + { + "epoch": 0.26, + "learning_rate": 0.0005989167502507523, + "loss": 1.517, + "theoretical_loss": 3.7335867907294817, + "tokens_seen": 854261760 + }, + { + "epoch": 0.26, + "learning_rate": 0.0005989007021063189, + "loss": 1.4532, + "theoretical_loss": 3.733558948442497, + "tokens_seen": 854327296 + }, + { + "epoch": 0.26, + "learning_rate": 0.0005988846539618857, + "loss": 1.4764, + "theoretical_loss": 3.733531108889202, + "tokens_seen": 854392832 + }, + { + "epoch": 0.26, + "learning_rate": 0.0005988686058174523, + "loss": 1.3894, + "theoretical_loss": 3.733503272069118, + "tokens_seen": 854458368 + }, + { + "epoch": 0.26, + "learning_rate": 0.0005988525576730191, + "loss": 1.3569, + "theoretical_loss": 3.7334754379817685, + "tokens_seen": 854523904 + }, + { + "epoch": 0.26, + "learning_rate": 0.0005988365095285857, + "loss": 1.4639, + "theoretical_loss": 3.733447606626674, + "tokens_seen": 854589440 + }, + { + "epoch": 0.26, + "learning_rate": 0.0005988204613841525, + "loss": 1.4333, + "theoretical_loss": 3.7334197780033582, + "tokens_seen": 854654976 + }, + { + "epoch": 0.26, + "learning_rate": 0.0005988044132397191, + "loss": 1.3283, + "theoretical_loss": 3.733391952111343, + "tokens_seen": 854720512 + }, + { + "epoch": 0.26, + "learning_rate": 0.0005987883650952859, + "loss": 1.4855, + "theoretical_loss": 3.73336412895015, + "tokens_seen": 854786048 + }, + { + "epoch": 0.26, + "learning_rate": 0.0005987723169508525, + "loss": 1.543, + "theoretical_loss": 3.733336308519303, + "tokens_seen": 854851584 + }, + { + "epoch": 0.26, + "learning_rate": 0.0005987562688064193, + "loss": 1.5026, + "theoretical_loss": 3.7333084908183247, + "tokens_seen": 854917120 + }, + { + "epoch": 0.26, + "learning_rate": 0.0005987402206619859, + "loss": 1.5389, + "theoretical_loss": 3.7332806758467383, + "tokens_seen": 854982656 + }, + { + "epoch": 0.26, + "learning_rate": 0.0005987241725175527, + "loss": 1.4744, + "theoretical_loss": 3.7332528636040654, + "tokens_seen": 855048192 + }, + { + "epoch": 0.26, + "learning_rate": 0.0005987081243731193, + "loss": 1.4522, + "theoretical_loss": 3.7332250540898304, + "tokens_seen": 855113728 + }, + { + "epoch": 0.26, + "learning_rate": 0.0005986920762286861, + "loss": 1.4747, + "theoretical_loss": 3.733197247303556, + "tokens_seen": 855179264 + }, + { + "epoch": 0.26, + "objective/train/docs_used": 315806, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.3717436790466309, + "objective/train/theoretical_loss": 3.733169443244766, + "objective/train/tokens_used": 875704800, + "theoretical_loss": 3.733169443244766, + "tokens_seen": 855244800 + }, + { + "epoch": 0.26, + "learning_rate": 0.0005986760280842527, + "loss": 1.419, + "theoretical_loss": 3.733169443244766, + "tokens_seen": 855244800 + }, + { + "epoch": 0.26, + "learning_rate": 0.0005986599799398195, + "loss": 1.4628, + "theoretical_loss": 3.7331416419129835, + "tokens_seen": 855310336 + }, + { + "epoch": 0.26, + "learning_rate": 0.0005986439317953861, + "loss": 1.4781, + "theoretical_loss": 3.7331138433077324, + "tokens_seen": 855375872 + }, + { + "epoch": 0.26, + "learning_rate": 0.0005986278836509529, + "loss": 1.5201, + "theoretical_loss": 3.733086047428536, + "tokens_seen": 855441408 + }, + { + "epoch": 0.26, + "learning_rate": 0.0005986118355065195, + "loss": 1.3418, + "theoretical_loss": 3.733058254274919, + "tokens_seen": 855506944 + }, + { + "epoch": 0.26, + "learning_rate": 0.0005985957873620863, + "loss": 1.3808, + "theoretical_loss": 3.7330304638464042, + "tokens_seen": 855572480 + }, + { + "epoch": 0.26, + "learning_rate": 0.0005985797392176529, + "loss": 1.5355, + "theoretical_loss": 3.733002676142517, + "tokens_seen": 855638016 + }, + { + "epoch": 0.26, + "learning_rate": 0.0005985636910732197, + "loss": 1.4141, + "theoretical_loss": 3.7329748911627805, + "tokens_seen": 855703552 + }, + { + "epoch": 0.26, + "learning_rate": 0.0005985476429287863, + "loss": 1.291, + "theoretical_loss": 3.7329471089067194, + "tokens_seen": 855769088 + }, + { + "epoch": 0.26, + "learning_rate": 0.0005985315947843531, + "loss": 1.4296, + "theoretical_loss": 3.7329193293738587, + "tokens_seen": 855834624 + }, + { + "epoch": 0.26, + "learning_rate": 0.0005985155466399197, + "loss": 1.4436, + "theoretical_loss": 3.7328915525637227, + "tokens_seen": 855900160 + }, + { + "epoch": 0.26, + "learning_rate": 0.0005984994984954865, + "loss": 1.4624, + "theoretical_loss": 3.7328637784758354, + "tokens_seen": 855965696 + }, + { + "epoch": 0.26, + "learning_rate": 0.0005984834503510531, + "loss": 1.5694, + "theoretical_loss": 3.7328360071097224, + "tokens_seen": 856031232 + }, + { + "epoch": 0.26, + "learning_rate": 0.0005984674022066199, + "loss": 1.5175, + "theoretical_loss": 3.7328082384649086, + "tokens_seen": 856096768 + }, + { + "epoch": 0.26, + "learning_rate": 0.0005984513540621865, + "loss": 1.4571, + "theoretical_loss": 3.7327804725409184, + "tokens_seen": 856162304 + }, + { + "epoch": 0.26, + "learning_rate": 0.0005984353059177533, + "loss": 1.464, + "theoretical_loss": 3.7327527093372774, + "tokens_seen": 856227840 + }, + { + "epoch": 0.26, + "learning_rate": 0.0005984192577733199, + "loss": 1.4449, + "theoretical_loss": 3.7327249488535106, + "tokens_seen": 856293376 + }, + { + "epoch": 0.26, + "learning_rate": 0.0005984032096288867, + "loss": 1.3543, + "theoretical_loss": 3.7326971910891444, + "tokens_seen": 856358912 + }, + { + "epoch": 0.26, + "learning_rate": 0.0005983871614844533, + "loss": 1.4063, + "theoretical_loss": 3.732669436043703, + "tokens_seen": 856424448 + }, + { + "epoch": 0.26, + "learning_rate": 0.0005983711133400201, + "loss": 1.4081, + "theoretical_loss": 3.7326416837167127, + "tokens_seen": 856489984 + }, + { + "epoch": 0.26, + "learning_rate": 0.0005983550651955867, + "loss": 1.4032, + "theoretical_loss": 3.7326139341076994, + "tokens_seen": 856555520 + }, + { + "epoch": 0.26, + "learning_rate": 0.0005983390170511535, + "loss": 1.4474, + "theoretical_loss": 3.7325861872161887, + "tokens_seen": 856621056 + }, + { + "epoch": 0.26, + "learning_rate": 0.0005983229689067201, + "loss": 1.3753, + "theoretical_loss": 3.732558443041707, + "tokens_seen": 856686592 + }, + { + "epoch": 0.26, + "learning_rate": 0.0005983069207622869, + "loss": 1.4686, + "theoretical_loss": 3.73253070158378, + "tokens_seen": 856752128 + }, + { + "epoch": 0.26, + "learning_rate": 0.0005982908726178535, + "loss": 1.4563, + "theoretical_loss": 3.732502962841934, + "tokens_seen": 856817664 + }, + { + "epoch": 0.26, + "learning_rate": 0.0005982748244734203, + "loss": 1.3603, + "theoretical_loss": 3.732475226815695, + "tokens_seen": 856883200 + }, + { + "epoch": 0.26, + "learning_rate": 0.0005982587763289869, + "loss": 1.468, + "theoretical_loss": 3.7324474935045906, + "tokens_seen": 856948736 + }, + { + "epoch": 0.26, + "learning_rate": 0.0005982427281845537, + "loss": 1.4499, + "theoretical_loss": 3.7324197629081466, + "tokens_seen": 857014272 + }, + { + "epoch": 0.26, + "learning_rate": 0.0005982266800401203, + "loss": 1.4539, + "theoretical_loss": 3.7323920350258897, + "tokens_seen": 857079808 + }, + { + "epoch": 0.26, + "learning_rate": 0.0005982106318956871, + "loss": 1.4189, + "theoretical_loss": 3.732364309857347, + "tokens_seen": 857145344 + }, + { + "epoch": 0.26, + "learning_rate": 0.0005981945837512539, + "loss": 1.4365, + "theoretical_loss": 3.7323365874020453, + "tokens_seen": 857210880 + }, + { + "epoch": 0.26, + "learning_rate": 0.0005981785356068205, + "loss": 1.4425, + "theoretical_loss": 3.732308867659511, + "tokens_seen": 857276416 + }, + { + "epoch": 0.26, + "learning_rate": 0.0005981624874623873, + "loss": 1.3876, + "theoretical_loss": 3.732281150629273, + "tokens_seen": 857341952 + }, + { + "epoch": 0.26, + "learning_rate": 0.0005981464393179539, + "loss": 1.5649, + "theoretical_loss": 3.732253436310857, + "tokens_seen": 857407488 + }, + { + "epoch": 0.26, + "learning_rate": 0.0005981303911735207, + "loss": 1.4506, + "theoretical_loss": 3.732225724703791, + "tokens_seen": 857473024 + }, + { + "epoch": 0.26, + "learning_rate": 0.0005981143430290873, + "loss": 1.4543, + "theoretical_loss": 3.7321980158076027, + "tokens_seen": 857538560 + }, + { + "epoch": 0.26, + "learning_rate": 0.000598098294884654, + "loss": 1.4877, + "theoretical_loss": 3.7321703096218197, + "tokens_seen": 857604096 + }, + { + "epoch": 0.26, + "learning_rate": 0.0005980822467402207, + "loss": 1.4127, + "theoretical_loss": 3.7321426061459704, + "tokens_seen": 857669632 + }, + { + "epoch": 0.26, + "learning_rate": 0.0005980661985957875, + "loss": 1.5491, + "theoretical_loss": 3.7321149053795812, + "tokens_seen": 857735168 + }, + { + "epoch": 0.26, + "learning_rate": 0.0005980501504513541, + "loss": 1.4471, + "theoretical_loss": 3.7320872073221807, + "tokens_seen": 857800704 + }, + { + "epoch": 0.26, + "learning_rate": 0.0005980341023069209, + "loss": 1.4722, + "theoretical_loss": 3.732059511973298, + "tokens_seen": 857866240 + }, + { + "epoch": 0.26, + "learning_rate": 0.0005980180541624875, + "loss": 1.4051, + "theoretical_loss": 3.7320318193324598, + "tokens_seen": 857931776 + }, + { + "epoch": 0.26, + "learning_rate": 0.0005980020060180543, + "loss": 1.4451, + "theoretical_loss": 3.732004129399196, + "tokens_seen": 857997312 + }, + { + "epoch": 0.26, + "learning_rate": 0.0005979859578736209, + "loss": 1.4685, + "theoretical_loss": 3.731976442173034, + "tokens_seen": 858062848 + }, + { + "epoch": 0.26, + "learning_rate": 0.0005979699097291877, + "loss": 1.4576, + "theoretical_loss": 3.731948757653503, + "tokens_seen": 858128384 + }, + { + "epoch": 0.26, + "learning_rate": 0.0005979538615847543, + "loss": 1.3537, + "theoretical_loss": 3.7319210758401318, + "tokens_seen": 858193920 + }, + { + "epoch": 0.26, + "learning_rate": 0.0005979378134403211, + "loss": 1.3904, + "theoretical_loss": 3.7318933967324486, + "tokens_seen": 858259456 + }, + { + "epoch": 0.26, + "learning_rate": 0.0005979217652958877, + "loss": 1.5124, + "theoretical_loss": 3.731865720329983, + "tokens_seen": 858324992 + }, + { + "epoch": 0.26, + "learning_rate": 0.0005979057171514545, + "loss": 1.5034, + "theoretical_loss": 3.731838046632264, + "tokens_seen": 858390528 + }, + { + "epoch": 0.26, + "learning_rate": 0.0005978896690070211, + "loss": 1.4474, + "theoretical_loss": 3.73181037563882, + "tokens_seen": 858456064 + }, + { + "epoch": 0.26, + "objective/train/docs_used": 317142, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.4725887775421143, + "objective/train/theoretical_loss": 3.7317827073491814, + "objective/train/tokens_used": 878981600, + "theoretical_loss": 3.7317827073491814, + "tokens_seen": 858521600 + }, + { + "epoch": 0.26, + "learning_rate": 0.0005978736208625879, + "loss": 1.5003, + "theoretical_loss": 3.7317827073491814, + "tokens_seen": 858521600 + }, + { + "epoch": 0.26, + "learning_rate": 0.0005978575727181545, + "loss": 1.4483, + "theoretical_loss": 3.7317550417628773, + "tokens_seen": 858587136 + }, + { + "epoch": 0.26, + "learning_rate": 0.0005978415245737213, + "loss": 1.5687, + "theoretical_loss": 3.731727378879437, + "tokens_seen": 858652672 + }, + { + "epoch": 0.26, + "learning_rate": 0.0005978254764292879, + "loss": 1.3938, + "theoretical_loss": 3.7316997186983905, + "tokens_seen": 858718208 + }, + { + "epoch": 0.26, + "learning_rate": 0.0005978094282848547, + "loss": 1.384, + "theoretical_loss": 3.7316720612192666, + "tokens_seen": 858783744 + }, + { + "epoch": 0.26, + "learning_rate": 0.0005977933801404213, + "loss": 1.3931, + "theoretical_loss": 3.731644406441597, + "tokens_seen": 858849280 + }, + { + "epoch": 0.26, + "learning_rate": 0.0005977773319959881, + "loss": 1.6072, + "theoretical_loss": 3.7316167543649104, + "tokens_seen": 858914816 + }, + { + "epoch": 0.26, + "learning_rate": 0.0005977612838515547, + "loss": 1.4585, + "theoretical_loss": 3.7315891049887373, + "tokens_seen": 858980352 + }, + { + "epoch": 0.26, + "learning_rate": 0.0005977452357071215, + "loss": 1.5977, + "theoretical_loss": 3.7315614583126084, + "tokens_seen": 859045888 + }, + { + "epoch": 0.26, + "learning_rate": 0.0005977291875626881, + "loss": 1.5109, + "theoretical_loss": 3.731533814336053, + "tokens_seen": 859111424 + }, + { + "epoch": 0.26, + "learning_rate": 0.0005977131394182549, + "loss": 1.4364, + "theoretical_loss": 3.7315061730586025, + "tokens_seen": 859176960 + }, + { + "epoch": 0.26, + "learning_rate": 0.0005976970912738215, + "loss": 1.558, + "theoretical_loss": 3.7314785344797867, + "tokens_seen": 859242496 + }, + { + "epoch": 0.26, + "learning_rate": 0.0005976810431293883, + "loss": 1.5175, + "theoretical_loss": 3.731450898599138, + "tokens_seen": 859308032 + }, + { + "epoch": 0.26, + "learning_rate": 0.0005976649949849549, + "loss": 1.4601, + "theoretical_loss": 3.731423265416185, + "tokens_seen": 859373568 + }, + { + "epoch": 0.26, + "learning_rate": 0.0005976489468405216, + "loss": 1.44, + "theoretical_loss": 3.7313956349304602, + "tokens_seen": 859439104 + }, + { + "epoch": 0.26, + "learning_rate": 0.0005976328986960883, + "loss": 1.4314, + "theoretical_loss": 3.7313680071414943, + "tokens_seen": 859504640 + }, + { + "epoch": 0.26, + "learning_rate": 0.000597616850551655, + "loss": 1.4133, + "theoretical_loss": 3.7313403820488187, + "tokens_seen": 859570176 + }, + { + "epoch": 0.26, + "learning_rate": 0.0005976008024072217, + "loss": 1.4083, + "theoretical_loss": 3.7313127596519644, + "tokens_seen": 859635712 + }, + { + "epoch": 0.26, + "learning_rate": 0.0005975847542627884, + "loss": 1.4047, + "theoretical_loss": 3.7312851399504625, + "tokens_seen": 859701248 + }, + { + "epoch": 0.26, + "learning_rate": 0.0005975687061183551, + "loss": 1.5961, + "theoretical_loss": 3.731257522943846, + "tokens_seen": 859766784 + }, + { + "epoch": 0.26, + "learning_rate": 0.0005975526579739218, + "loss": 1.4596, + "theoretical_loss": 3.7312299086316445, + "tokens_seen": 859832320 + }, + { + "epoch": 0.26, + "learning_rate": 0.0005975366098294885, + "loss": 1.5286, + "theoretical_loss": 3.731202297013391, + "tokens_seen": 859897856 + }, + { + "epoch": 0.26, + "learning_rate": 0.0005975205616850552, + "loss": 1.4199, + "theoretical_loss": 3.7311746880886174, + "tokens_seen": 859963392 + }, + { + "epoch": 0.26, + "learning_rate": 0.0005975045135406218, + "loss": 1.4461, + "theoretical_loss": 3.7311470818568555, + "tokens_seen": 860028928 + }, + { + "epoch": 0.26, + "learning_rate": 0.0005974884653961886, + "loss": 1.4575, + "theoretical_loss": 3.7311194783176376, + "tokens_seen": 860094464 + }, + { + "epoch": 0.26, + "learning_rate": 0.0005974724172517552, + "loss": 1.5449, + "theoretical_loss": 3.7310918774704955, + "tokens_seen": 860160000 + }, + { + "epoch": 0.26, + "learning_rate": 0.000597456369107322, + "loss": 1.5251, + "theoretical_loss": 3.731064279314962, + "tokens_seen": 860225536 + }, + { + "epoch": 0.26, + "learning_rate": 0.0005974403209628886, + "loss": 1.3753, + "theoretical_loss": 3.7310366838505704, + "tokens_seen": 860291072 + }, + { + "epoch": 0.26, + "learning_rate": 0.0005974242728184554, + "loss": 1.4578, + "theoretical_loss": 3.7310090910768516, + "tokens_seen": 860356608 + }, + { + "epoch": 0.26, + "learning_rate": 0.000597408224674022, + "loss": 1.3737, + "theoretical_loss": 3.730981500993339, + "tokens_seen": 860422144 + }, + { + "epoch": 0.26, + "learning_rate": 0.0005973921765295888, + "loss": 1.5311, + "theoretical_loss": 3.7309539135995657, + "tokens_seen": 860487680 + }, + { + "epoch": 0.26, + "learning_rate": 0.0005973761283851554, + "loss": 1.3451, + "theoretical_loss": 3.7309263288950647, + "tokens_seen": 860553216 + }, + { + "epoch": 0.26, + "learning_rate": 0.0005973600802407222, + "loss": 1.4296, + "theoretical_loss": 3.730898746879369, + "tokens_seen": 860618752 + }, + { + "epoch": 0.26, + "learning_rate": 0.0005973440320962888, + "loss": 1.6501, + "theoretical_loss": 3.730871167552011, + "tokens_seen": 860684288 + }, + { + "epoch": 0.26, + "learning_rate": 0.0005973279839518556, + "loss": 1.4685, + "theoretical_loss": 3.7308435909125253, + "tokens_seen": 860749824 + }, + { + "epoch": 0.26, + "learning_rate": 0.0005973119358074222, + "loss": 1.5209, + "theoretical_loss": 3.7308160169604445, + "tokens_seen": 860815360 + }, + { + "epoch": 0.26, + "learning_rate": 0.000597295887662989, + "loss": 1.5865, + "theoretical_loss": 3.7307884456953024, + "tokens_seen": 860880896 + }, + { + "epoch": 0.26, + "learning_rate": 0.0005972798395185556, + "loss": 1.388, + "theoretical_loss": 3.7307608771166327, + "tokens_seen": 860946432 + }, + { + "epoch": 0.26, + "learning_rate": 0.0005972637913741224, + "loss": 1.5047, + "theoretical_loss": 3.730733311223969, + "tokens_seen": 861011968 + }, + { + "epoch": 0.26, + "learning_rate": 0.000597247743229689, + "loss": 1.4663, + "theoretical_loss": 3.7307057480168453, + "tokens_seen": 861077504 + }, + { + "epoch": 0.26, + "learning_rate": 0.0005972316950852558, + "loss": 1.4633, + "theoretical_loss": 3.730678187494796, + "tokens_seen": 861143040 + }, + { + "epoch": 0.26, + "learning_rate": 0.0005972156469408224, + "loss": 1.3678, + "theoretical_loss": 3.7306506296573536, + "tokens_seen": 861208576 + }, + { + "epoch": 0.26, + "learning_rate": 0.0005971995987963892, + "loss": 1.4458, + "theoretical_loss": 3.7306230745040545, + "tokens_seen": 861274112 + }, + { + "epoch": 0.26, + "learning_rate": 0.0005971835506519558, + "loss": 1.4941, + "theoretical_loss": 3.730595522034432, + "tokens_seen": 861339648 + }, + { + "epoch": 0.26, + "learning_rate": 0.0005971675025075226, + "loss": 1.354, + "theoretical_loss": 3.73056797224802, + "tokens_seen": 861405184 + }, + { + "epoch": 0.26, + "learning_rate": 0.0005971514543630892, + "loss": 1.4773, + "theoretical_loss": 3.730540425144354, + "tokens_seen": 861470720 + }, + { + "epoch": 0.26, + "learning_rate": 0.000597135406218656, + "loss": 1.4959, + "theoretical_loss": 3.7305128807229684, + "tokens_seen": 861536256 + }, + { + "epoch": 0.26, + "learning_rate": 0.0005971193580742226, + "loss": 1.4743, + "theoretical_loss": 3.730485338983398, + "tokens_seen": 861601792 + }, + { + "epoch": 0.26, + "learning_rate": 0.0005971033099297894, + "loss": 1.4133, + "theoretical_loss": 3.730457799925178, + "tokens_seen": 861667328 + }, + { + "epoch": 0.26, + "learning_rate": 0.000597087261785356, + "loss": 1.4156, + "theoretical_loss": 3.7304302635478432, + "tokens_seen": 861732864 + }, + { + "epoch": 0.26, + "objective/train/docs_used": 318349, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.3925228118896484, + "objective/train/theoretical_loss": 3.730402729850929, + "objective/train/tokens_used": 882258400, + "theoretical_loss": 3.730402729850929, + "tokens_seen": 861798400 + }, + { + "epoch": 0.26, + "learning_rate": 0.0005970712136409228, + "loss": 1.4977, + "theoretical_loss": 3.730402729850929, + "tokens_seen": 861798400 + }, + { + "epoch": 0.26, + "learning_rate": 0.0005970551654964894, + "loss": 1.5968, + "theoretical_loss": 3.7303751988339697, + "tokens_seen": 861863936 + }, + { + "epoch": 0.26, + "learning_rate": 0.0005970391173520562, + "loss": 1.3973, + "theoretical_loss": 3.730347670496502, + "tokens_seen": 861929472 + }, + { + "epoch": 0.26, + "learning_rate": 0.0005970230692076228, + "loss": 1.4693, + "theoretical_loss": 3.7303201448380605, + "tokens_seen": 861995008 + }, + { + "epoch": 0.26, + "learning_rate": 0.0005970070210631896, + "loss": 1.4764, + "theoretical_loss": 3.7302926218581813, + "tokens_seen": 862060544 + }, + { + "epoch": 0.26, + "learning_rate": 0.0005969909729187563, + "loss": 1.3857, + "theoretical_loss": 3.730265101556401, + "tokens_seen": 862126080 + }, + { + "epoch": 0.26, + "learning_rate": 0.000596974924774323, + "loss": 1.3792, + "theoretical_loss": 3.7302375839322535, + "tokens_seen": 862191616 + }, + { + "epoch": 0.26, + "learning_rate": 0.0005969588766298897, + "loss": 1.4686, + "theoretical_loss": 3.7302100689852757, + "tokens_seen": 862257152 + }, + { + "epoch": 0.26, + "learning_rate": 0.0005969428284854564, + "loss": 1.4242, + "theoretical_loss": 3.730182556715004, + "tokens_seen": 862322688 + }, + { + "epoch": 0.26, + "learning_rate": 0.0005969267803410231, + "loss": 1.5389, + "theoretical_loss": 3.7301550471209746, + "tokens_seen": 862388224 + }, + { + "epoch": 0.26, + "learning_rate": 0.0005969107321965898, + "loss": 1.5624, + "theoretical_loss": 3.730127540202723, + "tokens_seen": 862453760 + }, + { + "epoch": 0.26, + "learning_rate": 0.0005968946840521565, + "loss": 1.4471, + "theoretical_loss": 3.730100035959787, + "tokens_seen": 862519296 + }, + { + "epoch": 0.26, + "learning_rate": 0.0005968786359077232, + "loss": 1.5197, + "theoretical_loss": 3.730072534391702, + "tokens_seen": 862584832 + }, + { + "epoch": 0.26, + "learning_rate": 0.00059686258776329, + "loss": 1.4608, + "theoretical_loss": 3.730045035498005, + "tokens_seen": 862650368 + }, + { + "epoch": 0.26, + "learning_rate": 0.0005968465396188566, + "loss": 1.481, + "theoretical_loss": 3.730017539278233, + "tokens_seen": 862715904 + }, + { + "epoch": 0.26, + "learning_rate": 0.0005968304914744233, + "loss": 1.6164, + "theoretical_loss": 3.729990045731922, + "tokens_seen": 862781440 + }, + { + "epoch": 0.26, + "learning_rate": 0.00059681444332999, + "loss": 1.4455, + "theoretical_loss": 3.7299625548586106, + "tokens_seen": 862846976 + }, + { + "epoch": 0.26, + "learning_rate": 0.0005967983951855568, + "loss": 1.4438, + "theoretical_loss": 3.729935066657835, + "tokens_seen": 862912512 + }, + { + "epoch": 0.26, + "learning_rate": 0.0005967823470411234, + "loss": 1.5327, + "theoretical_loss": 3.7299075811291322, + "tokens_seen": 862978048 + }, + { + "epoch": 0.26, + "learning_rate": 0.0005967662988966902, + "loss": 1.3944, + "theoretical_loss": 3.7298800982720404, + "tokens_seen": 863043584 + }, + { + "epoch": 0.26, + "learning_rate": 0.0005967502507522568, + "loss": 1.5556, + "theoretical_loss": 3.7298526180860962, + "tokens_seen": 863109120 + }, + { + "epoch": 0.26, + "learning_rate": 0.0005967342026078236, + "loss": 1.3491, + "theoretical_loss": 3.7298251405708376, + "tokens_seen": 863174656 + }, + { + "epoch": 0.26, + "learning_rate": 0.0005967181544633902, + "loss": 1.4041, + "theoretical_loss": 3.729797665725802, + "tokens_seen": 863240192 + }, + { + "epoch": 0.26, + "learning_rate": 0.000596702106318957, + "loss": 1.4735, + "theoretical_loss": 3.729770193550528, + "tokens_seen": 863305728 + }, + { + "epoch": 0.26, + "learning_rate": 0.0005966860581745236, + "loss": 1.3628, + "theoretical_loss": 3.729742724044553, + "tokens_seen": 863371264 + }, + { + "epoch": 0.26, + "learning_rate": 0.0005966700100300904, + "loss": 1.4389, + "theoretical_loss": 3.7297152572074146, + "tokens_seen": 863436800 + }, + { + "epoch": 0.26, + "learning_rate": 0.000596653961885657, + "loss": 1.4718, + "theoretical_loss": 3.729687793038652, + "tokens_seen": 863502336 + }, + { + "epoch": 0.26, + "learning_rate": 0.0005966379137412238, + "loss": 1.4875, + "theoretical_loss": 3.7296603315378025, + "tokens_seen": 863567872 + }, + { + "epoch": 0.26, + "learning_rate": 0.0005966218655967904, + "loss": 1.4502, + "theoretical_loss": 3.7296328727044052, + "tokens_seen": 863633408 + }, + { + "epoch": 0.26, + "learning_rate": 0.0005966058174523572, + "loss": 1.4565, + "theoretical_loss": 3.729605416537998, + "tokens_seen": 863698944 + }, + { + "epoch": 0.26, + "learning_rate": 0.0005965897693079238, + "loss": 1.4912, + "theoretical_loss": 3.7295779630381203, + "tokens_seen": 863764480 + }, + { + "epoch": 0.26, + "learning_rate": 0.0005965737211634906, + "loss": 1.468, + "theoretical_loss": 3.7295505122043098, + "tokens_seen": 863830016 + }, + { + "epoch": 0.26, + "learning_rate": 0.0005965576730190572, + "loss": 1.4186, + "theoretical_loss": 3.7295230640361066, + "tokens_seen": 863895552 + }, + { + "epoch": 0.26, + "learning_rate": 0.000596541624874624, + "loss": 1.474, + "theoretical_loss": 3.729495618533049, + "tokens_seen": 863961088 + }, + { + "epoch": 0.26, + "learning_rate": 0.0005965255767301906, + "loss": 1.4538, + "theoretical_loss": 3.7294681756946755, + "tokens_seen": 864026624 + }, + { + "epoch": 0.26, + "learning_rate": 0.0005965095285857574, + "loss": 1.4315, + "theoretical_loss": 3.7294407355205257, + "tokens_seen": 864092160 + }, + { + "epoch": 0.26, + "learning_rate": 0.000596493480441324, + "loss": 1.5526, + "theoretical_loss": 3.72941329801014, + "tokens_seen": 864157696 + }, + { + "epoch": 0.26, + "learning_rate": 0.0005964774322968908, + "loss": 1.5283, + "theoretical_loss": 3.7293858631630563, + "tokens_seen": 864223232 + }, + { + "epoch": 0.26, + "learning_rate": 0.0005964613841524574, + "loss": 1.4576, + "theoretical_loss": 3.7293584309788157, + "tokens_seen": 864288768 + }, + { + "epoch": 0.26, + "learning_rate": 0.0005964453360080242, + "loss": 1.4821, + "theoretical_loss": 3.729331001456956, + "tokens_seen": 864354304 + }, + { + "epoch": 0.26, + "learning_rate": 0.0005964292878635908, + "loss": 1.4715, + "theoretical_loss": 3.7293035745970182, + "tokens_seen": 864419840 + }, + { + "epoch": 0.26, + "learning_rate": 0.0005964132397191576, + "loss": 1.42, + "theoretical_loss": 3.729276150398542, + "tokens_seen": 864485376 + }, + { + "epoch": 0.26, + "learning_rate": 0.0005963971915747242, + "loss": 1.4766, + "theoretical_loss": 3.7292487288610676, + "tokens_seen": 864550912 + }, + { + "epoch": 0.26, + "learning_rate": 0.000596381143430291, + "loss": 1.4183, + "theoretical_loss": 3.7292213099841343, + "tokens_seen": 864616448 + }, + { + "epoch": 0.26, + "learning_rate": 0.0005963650952858576, + "loss": 1.4905, + "theoretical_loss": 3.7291938937672833, + "tokens_seen": 864681984 + }, + { + "epoch": 0.26, + "learning_rate": 0.0005963490471414244, + "loss": 1.4185, + "theoretical_loss": 3.729166480210054, + "tokens_seen": 864747520 + }, + { + "epoch": 0.26, + "learning_rate": 0.000596332998996991, + "loss": 1.5781, + "theoretical_loss": 3.729139069311988, + "tokens_seen": 864813056 + }, + { + "epoch": 0.26, + "learning_rate": 0.0005963169508525578, + "loss": 1.3292, + "theoretical_loss": 3.729111661072625, + "tokens_seen": 864878592 + }, + { + "epoch": 0.26, + "learning_rate": 0.0005963009027081244, + "loss": 1.4546, + "theoretical_loss": 3.729084255491506, + "tokens_seen": 864944128 + }, + { + "epoch": 0.26, + "learning_rate": 0.0005962848545636912, + "loss": 1.4215, + "theoretical_loss": 3.7290568525681715, + "tokens_seen": 865009664 + }, + { + "epoch": 0.26, + "objective/train/docs_used": 319421, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.3840081691741943, + "objective/train/theoretical_loss": 3.7290294523021625, + "objective/train/tokens_used": 885535200, + "theoretical_loss": 3.7290294523021625, + "tokens_seen": 865075200 + }, + { + "epoch": 0.26, + "learning_rate": 0.0005962688064192578, + "loss": 1.4255, + "theoretical_loss": 3.7290294523021625, + "tokens_seen": 865075200 + }, + { + "epoch": 0.26, + "learning_rate": 0.0005962527582748246, + "loss": 1.4035, + "theoretical_loss": 3.729002054693021, + "tokens_seen": 865140736 + }, + { + "epoch": 0.26, + "learning_rate": 0.0005962367101303912, + "loss": 1.5581, + "theoretical_loss": 3.7289746597402864, + "tokens_seen": 865206272 + }, + { + "epoch": 0.26, + "learning_rate": 0.0005962206619859579, + "loss": 1.5796, + "theoretical_loss": 3.7289472674435014, + "tokens_seen": 865271808 + }, + { + "epoch": 0.26, + "learning_rate": 0.0005962046138415246, + "loss": 1.4751, + "theoretical_loss": 3.728919877802207, + "tokens_seen": 865337344 + }, + { + "epoch": 0.26, + "learning_rate": 0.0005961885656970913, + "loss": 1.5072, + "theoretical_loss": 3.728892490815944, + "tokens_seen": 865402880 + }, + { + "epoch": 0.26, + "learning_rate": 0.000596172517552658, + "loss": 1.4092, + "theoretical_loss": 3.7288651064842546, + "tokens_seen": 865468416 + }, + { + "epoch": 0.26, + "learning_rate": 0.0005961564694082247, + "loss": 1.5514, + "theoretical_loss": 3.728837724806681, + "tokens_seen": 865533952 + }, + { + "epoch": 0.26, + "learning_rate": 0.0005961404212637914, + "loss": 1.4988, + "theoretical_loss": 3.7288103457827635, + "tokens_seen": 865599488 + }, + { + "epoch": 0.26, + "learning_rate": 0.0005961243731193581, + "loss": 1.5246, + "theoretical_loss": 3.7287829694120456, + "tokens_seen": 865665024 + }, + { + "epoch": 0.26, + "learning_rate": 0.0005961083249749248, + "loss": 1.3343, + "theoretical_loss": 3.728755595694069, + "tokens_seen": 865730560 + }, + { + "epoch": 0.26, + "learning_rate": 0.0005960922768304915, + "loss": 1.4989, + "theoretical_loss": 3.7287282246283757, + "tokens_seen": 865796096 + }, + { + "epoch": 0.26, + "learning_rate": 0.0005960762286860581, + "loss": 1.5443, + "theoretical_loss": 3.728700856214507, + "tokens_seen": 865861632 + }, + { + "epoch": 0.26, + "learning_rate": 0.0005960601805416249, + "loss": 1.473, + "theoretical_loss": 3.7286734904520076, + "tokens_seen": 865927168 + }, + { + "epoch": 0.26, + "learning_rate": 0.0005960441323971915, + "loss": 1.5156, + "theoretical_loss": 3.7286461273404177, + "tokens_seen": 865992704 + }, + { + "epoch": 0.26, + "learning_rate": 0.0005960280842527583, + "loss": 1.4208, + "theoretical_loss": 3.728618766879281, + "tokens_seen": 866058240 + }, + { + "epoch": 0.26, + "learning_rate": 0.0005960120361083249, + "loss": 1.4204, + "theoretical_loss": 3.72859140906814, + "tokens_seen": 866123776 + }, + { + "epoch": 0.26, + "learning_rate": 0.0005959959879638917, + "loss": 1.3464, + "theoretical_loss": 3.7285640539065374, + "tokens_seen": 866189312 + }, + { + "epoch": 0.26, + "learning_rate": 0.0005959799398194583, + "loss": 1.454, + "theoretical_loss": 3.7285367013940167, + "tokens_seen": 866254848 + }, + { + "epoch": 0.26, + "learning_rate": 0.0005959638916750251, + "loss": 1.4832, + "theoretical_loss": 3.7285093515301213, + "tokens_seen": 866320384 + }, + { + "epoch": 0.26, + "learning_rate": 0.0005959478435305917, + "loss": 1.5099, + "theoretical_loss": 3.728482004314393, + "tokens_seen": 866385920 + }, + { + "epoch": 0.26, + "learning_rate": 0.0005959317953861585, + "loss": 1.3155, + "theoretical_loss": 3.728454659746376, + "tokens_seen": 866451456 + }, + { + "epoch": 0.26, + "learning_rate": 0.0005959157472417251, + "loss": 1.45, + "theoretical_loss": 3.7284273178256133, + "tokens_seen": 866516992 + }, + { + "epoch": 0.26, + "learning_rate": 0.0005958996990972919, + "loss": 1.482, + "theoretical_loss": 3.728399978551649, + "tokens_seen": 866582528 + }, + { + "epoch": 0.26, + "learning_rate": 0.0005958836509528585, + "loss": 1.5023, + "theoretical_loss": 3.7283726419240266, + "tokens_seen": 866648064 + }, + { + "epoch": 0.26, + "learning_rate": 0.0005958676028084253, + "loss": 1.5271, + "theoretical_loss": 3.7283453079422895, + "tokens_seen": 866713600 + }, + { + "epoch": 0.26, + "learning_rate": 0.0005958515546639919, + "loss": 1.4242, + "theoretical_loss": 3.7283179766059824, + "tokens_seen": 866779136 + }, + { + "epoch": 0.26, + "learning_rate": 0.0005958355065195587, + "loss": 1.4788, + "theoretical_loss": 3.7282906479146476, + "tokens_seen": 866844672 + }, + { + "epoch": 0.26, + "learning_rate": 0.0005958194583751253, + "loss": 1.4634, + "theoretical_loss": 3.728263321867831, + "tokens_seen": 866910208 + }, + { + "epoch": 0.26, + "learning_rate": 0.0005958034102306921, + "loss": 1.3556, + "theoretical_loss": 3.7282359984650757, + "tokens_seen": 866975744 + }, + { + "epoch": 0.26, + "learning_rate": 0.0005957873620862588, + "loss": 1.5376, + "theoretical_loss": 3.7282086777059265, + "tokens_seen": 867041280 + }, + { + "epoch": 0.26, + "learning_rate": 0.0005957713139418255, + "loss": 1.537, + "theoretical_loss": 3.728181359589928, + "tokens_seen": 867106816 + }, + { + "epoch": 0.26, + "learning_rate": 0.0005957552657973922, + "loss": 1.4452, + "theoretical_loss": 3.728154044116624, + "tokens_seen": 867172352 + }, + { + "epoch": 0.26, + "learning_rate": 0.0005957392176529589, + "loss": 1.3923, + "theoretical_loss": 3.72812673128556, + "tokens_seen": 867237888 + }, + { + "epoch": 0.26, + "learning_rate": 0.0005957231695085256, + "loss": 1.4235, + "theoretical_loss": 3.728099421096281, + "tokens_seen": 867303424 + }, + { + "epoch": 0.26, + "learning_rate": 0.0005957071213640923, + "loss": 1.5163, + "theoretical_loss": 3.72807211354833, + "tokens_seen": 867368960 + }, + { + "epoch": 0.26, + "learning_rate": 0.000595691073219659, + "loss": 1.4161, + "theoretical_loss": 3.728044808641254, + "tokens_seen": 867434496 + }, + { + "epoch": 0.26, + "learning_rate": 0.0005956750250752257, + "loss": 1.6256, + "theoretical_loss": 3.7280175063745973, + "tokens_seen": 867500032 + }, + { + "epoch": 0.26, + "learning_rate": 0.0005956589769307924, + "loss": 1.4438, + "theoretical_loss": 3.727990206747905, + "tokens_seen": 867565568 + }, + { + "epoch": 0.26, + "learning_rate": 0.0005956429287863591, + "loss": 1.634, + "theoretical_loss": 3.7279629097607234, + "tokens_seen": 867631104 + }, + { + "epoch": 0.26, + "learning_rate": 0.0005956268806419258, + "loss": 1.4753, + "theoretical_loss": 3.7279356154125964, + "tokens_seen": 867696640 + }, + { + "epoch": 0.26, + "learning_rate": 0.0005956108324974925, + "loss": 1.4878, + "theoretical_loss": 3.727908323703071, + "tokens_seen": 867762176 + }, + { + "epoch": 0.26, + "learning_rate": 0.0005955947843530592, + "loss": 1.523, + "theoretical_loss": 3.7278810346316913, + "tokens_seen": 867827712 + }, + { + "epoch": 0.26, + "learning_rate": 0.0005955787362086259, + "loss": 1.4139, + "theoretical_loss": 3.7278537481980045, + "tokens_seen": 867893248 + }, + { + "epoch": 0.26, + "learning_rate": 0.0005955626880641926, + "loss": 1.5918, + "theoretical_loss": 3.7278264644015566, + "tokens_seen": 867958784 + }, + { + "epoch": 0.26, + "learning_rate": 0.0005955466399197593, + "loss": 1.5403, + "theoretical_loss": 3.727799183241893, + "tokens_seen": 868024320 + }, + { + "epoch": 0.26, + "learning_rate": 0.000595530591775326, + "loss": 1.4665, + "theoretical_loss": 3.7277719047185593, + "tokens_seen": 868089856 + }, + { + "epoch": 0.26, + "learning_rate": 0.0005955145436308927, + "loss": 1.4062, + "theoretical_loss": 3.7277446288311022, + "tokens_seen": 868155392 + }, + { + "epoch": 0.26, + "learning_rate": 0.0005954984954864594, + "loss": 1.5332, + "theoretical_loss": 3.7277173555790686, + "tokens_seen": 868220928 + }, + { + "epoch": 0.26, + "learning_rate": 0.0005954824473420261, + "loss": 1.5179, + "theoretical_loss": 3.7276900849620045, + "tokens_seen": 868286464 + }, + { + "epoch": 0.26, + "objective/train/docs_used": 320668, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.470707654953003, + "objective/train/theoretical_loss": 3.727662816979456, + "objective/train/tokens_used": 888812000, + "theoretical_loss": 3.727662816979456, + "tokens_seen": 868352000 + }, + { + "epoch": 0.26, + "learning_rate": 0.0005954663991975928, + "loss": 1.4844, + "theoretical_loss": 3.727662816979456, + "tokens_seen": 868352000 + }, + { + "epoch": 0.26, + "learning_rate": 0.0005954503510531595, + "loss": 1.4807, + "theoretical_loss": 3.727635551630971, + "tokens_seen": 868417536 + }, + { + "epoch": 0.26, + "learning_rate": 0.0005954343029087263, + "loss": 1.5203, + "theoretical_loss": 3.7276082889160955, + "tokens_seen": 868483072 + }, + { + "epoch": 0.26, + "learning_rate": 0.0005954182547642929, + "loss": 1.413, + "theoretical_loss": 3.727581028834376, + "tokens_seen": 868548608 + }, + { + "epoch": 0.26, + "learning_rate": 0.0005954022066198597, + "loss": 1.5338, + "theoretical_loss": 3.7275537713853604, + "tokens_seen": 868614144 + }, + { + "epoch": 0.26, + "learning_rate": 0.0005953861584754263, + "loss": 1.4193, + "theoretical_loss": 3.727526516568595, + "tokens_seen": 868679680 + }, + { + "epoch": 0.26, + "learning_rate": 0.000595370110330993, + "loss": 1.5917, + "theoretical_loss": 3.727499264383628, + "tokens_seen": 868745216 + }, + { + "epoch": 0.26, + "learning_rate": 0.0005953540621865597, + "loss": 1.4053, + "theoretical_loss": 3.7274720148300062, + "tokens_seen": 868810752 + }, + { + "epoch": 0.26, + "learning_rate": 0.0005953380140421265, + "loss": 1.4435, + "theoretical_loss": 3.727444767907277, + "tokens_seen": 868876288 + }, + { + "epoch": 0.26, + "learning_rate": 0.0005953219658976931, + "loss": 1.5033, + "theoretical_loss": 3.7274175236149887, + "tokens_seen": 868941824 + }, + { + "epoch": 0.26, + "learning_rate": 0.0005953059177532599, + "loss": 1.4349, + "theoretical_loss": 3.7273902819526876, + "tokens_seen": 869007360 + }, + { + "epoch": 0.26, + "learning_rate": 0.0005952898696088265, + "loss": 1.5077, + "theoretical_loss": 3.7273630429199223, + "tokens_seen": 869072896 + }, + { + "epoch": 0.26, + "learning_rate": 0.0005952738214643933, + "loss": 1.3438, + "theoretical_loss": 3.727335806516241, + "tokens_seen": 869138432 + }, + { + "epoch": 0.26, + "learning_rate": 0.0005952577733199599, + "loss": 1.4561, + "theoretical_loss": 3.727308572741191, + "tokens_seen": 869203968 + }, + { + "epoch": 0.26, + "learning_rate": 0.0005952417251755267, + "loss": 1.4918, + "theoretical_loss": 3.7272813415943222, + "tokens_seen": 869269504 + }, + { + "epoch": 0.26, + "learning_rate": 0.0005952256770310933, + "loss": 1.5112, + "theoretical_loss": 3.7272541130751806, + "tokens_seen": 869335040 + }, + { + "epoch": 0.26, + "learning_rate": 0.0005952096288866601, + "loss": 1.3987, + "theoretical_loss": 3.727226887183315, + "tokens_seen": 869400576 + }, + { + "epoch": 0.26, + "learning_rate": 0.0005951935807422267, + "loss": 1.521, + "theoretical_loss": 3.7271996639182756, + "tokens_seen": 869466112 + }, + { + "epoch": 0.26, + "learning_rate": 0.0005951775325977935, + "loss": 1.4789, + "theoretical_loss": 3.727172443279609, + "tokens_seen": 869531648 + }, + { + "epoch": 0.26, + "learning_rate": 0.0005951614844533601, + "loss": 1.5757, + "theoretical_loss": 3.727145225266865, + "tokens_seen": 869597184 + }, + { + "epoch": 0.26, + "learning_rate": 0.0005951454363089269, + "loss": 1.5781, + "theoretical_loss": 3.727118009879592, + "tokens_seen": 869662720 + }, + { + "epoch": 0.26, + "learning_rate": 0.0005951293881644935, + "loss": 1.4375, + "theoretical_loss": 3.727090797117339, + "tokens_seen": 869728256 + }, + { + "epoch": 0.26, + "learning_rate": 0.0005951133400200603, + "loss": 1.3386, + "theoretical_loss": 3.7270635869796553, + "tokens_seen": 869793792 + }, + { + "epoch": 0.26, + "learning_rate": 0.0005950972918756269, + "loss": 1.485, + "theoretical_loss": 3.7270363794660892, + "tokens_seen": 869859328 + }, + { + "epoch": 0.26, + "learning_rate": 0.0005950812437311937, + "loss": 1.5219, + "theoretical_loss": 3.727009174576191, + "tokens_seen": 869924864 + }, + { + "epoch": 0.26, + "learning_rate": 0.0005950651955867603, + "loss": 1.3366, + "theoretical_loss": 3.7269819723095097, + "tokens_seen": 869990400 + }, + { + "epoch": 0.26, + "learning_rate": 0.0005950491474423271, + "loss": 1.469, + "theoretical_loss": 3.7269547726655947, + "tokens_seen": 870055936 + }, + { + "epoch": 0.26, + "learning_rate": 0.0005950330992978937, + "loss": 1.5591, + "theoretical_loss": 3.7269275756439955, + "tokens_seen": 870121472 + }, + { + "epoch": 0.26, + "learning_rate": 0.0005950170511534605, + "loss": 1.5335, + "theoretical_loss": 3.7269003812442616, + "tokens_seen": 870187008 + }, + { + "epoch": 0.26, + "learning_rate": 0.0005950010030090271, + "loss": 1.5066, + "theoretical_loss": 3.7268731894659437, + "tokens_seen": 870252544 + }, + { + "epoch": 0.26, + "learning_rate": 0.0005949849548645939, + "loss": 1.6351, + "theoretical_loss": 3.726846000308591, + "tokens_seen": 870318080 + }, + { + "epoch": 0.26, + "learning_rate": 0.0005949689067201605, + "loss": 1.4945, + "theoretical_loss": 3.7268188137717537, + "tokens_seen": 870383616 + }, + { + "epoch": 0.26, + "learning_rate": 0.0005949528585757273, + "loss": 1.4388, + "theoretical_loss": 3.7267916298549815, + "tokens_seen": 870449152 + }, + { + "epoch": 0.26, + "learning_rate": 0.0005949368104312939, + "loss": 1.5551, + "theoretical_loss": 3.7267644485578257, + "tokens_seen": 870514688 + }, + { + "epoch": 0.26, + "learning_rate": 0.0005949207622868607, + "loss": 1.4693, + "theoretical_loss": 3.726737269879836, + "tokens_seen": 870580224 + }, + { + "epoch": 0.26, + "learning_rate": 0.0005949047141424273, + "loss": 1.4779, + "theoretical_loss": 3.7267100938205626, + "tokens_seen": 870645760 + }, + { + "epoch": 0.26, + "learning_rate": 0.0005948886659979941, + "loss": 1.6102, + "theoretical_loss": 3.726682920379556, + "tokens_seen": 870711296 + }, + { + "epoch": 0.26, + "learning_rate": 0.0005948726178535607, + "loss": 1.3022, + "theoretical_loss": 3.726655749556368, + "tokens_seen": 870776832 + }, + { + "epoch": 0.26, + "learning_rate": 0.0005948565697091275, + "loss": 1.4151, + "theoretical_loss": 3.726628581350549, + "tokens_seen": 870842368 + }, + { + "epoch": 0.26, + "learning_rate": 0.0005948405215646941, + "loss": 1.513, + "theoretical_loss": 3.7266014157616487, + "tokens_seen": 870907904 + }, + { + "epoch": 0.26, + "learning_rate": 0.0005948244734202608, + "loss": 1.4876, + "theoretical_loss": 3.72657425278922, + "tokens_seen": 870973440 + }, + { + "epoch": 0.26, + "learning_rate": 0.0005948084252758275, + "loss": 1.3606, + "theoretical_loss": 3.7265470924328126, + "tokens_seen": 871038976 + }, + { + "epoch": 0.26, + "learning_rate": 0.0005947923771313942, + "loss": 1.4305, + "theoretical_loss": 3.7265199346919786, + "tokens_seen": 871104512 + }, + { + "epoch": 0.26, + "learning_rate": 0.0005947763289869609, + "loss": 1.5345, + "theoretical_loss": 3.7264927795662692, + "tokens_seen": 871170048 + }, + { + "epoch": 0.26, + "learning_rate": 0.0005947602808425276, + "loss": 1.5738, + "theoretical_loss": 3.7264656270552354, + "tokens_seen": 871235584 + }, + { + "epoch": 0.26, + "learning_rate": 0.0005947442326980943, + "loss": 1.4335, + "theoretical_loss": 3.72643847715843, + "tokens_seen": 871301120 + }, + { + "epoch": 0.26, + "learning_rate": 0.000594728184553661, + "loss": 1.4224, + "theoretical_loss": 3.726411329875403, + "tokens_seen": 871366656 + }, + { + "epoch": 0.26, + "learning_rate": 0.0005947121364092277, + "loss": 1.4414, + "theoretical_loss": 3.726384185205707, + "tokens_seen": 871432192 + }, + { + "epoch": 0.26, + "learning_rate": 0.0005946960882647944, + "loss": 1.4602, + "theoretical_loss": 3.7263570431488944, + "tokens_seen": 871497728 + }, + { + "epoch": 0.26, + "learning_rate": 0.000594680040120361, + "loss": 1.5792, + "theoretical_loss": 3.7263299037045168, + "tokens_seen": 871563264 + }, + { + "epoch": 0.26, + "objective/train/docs_used": 321822, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.6328963041305542, + "objective/train/theoretical_loss": 3.7263027668721262, + "objective/train/tokens_used": 892088800, + "theoretical_loss": 3.7263027668721262, + "tokens_seen": 871628800 + }, + { + "epoch": 0.26, + "learning_rate": 0.0005946639919759278, + "loss": 1.5599, + "theoretical_loss": 3.7263027668721262, + "tokens_seen": 871628800 + }, + { + "epoch": 0.26, + "learning_rate": 0.0005946479438314944, + "loss": 1.5274, + "theoretical_loss": 3.7262756326512747, + "tokens_seen": 871694336 + }, + { + "epoch": 0.26, + "learning_rate": 0.0005946318956870612, + "loss": 1.3778, + "theoretical_loss": 3.7262485010415154, + "tokens_seen": 871759872 + }, + { + "epoch": 0.26, + "learning_rate": 0.0005946158475426278, + "loss": 1.4112, + "theoretical_loss": 3.7262213720424002, + "tokens_seen": 871825408 + }, + { + "epoch": 0.26, + "learning_rate": 0.0005945997993981946, + "loss": 1.4783, + "theoretical_loss": 3.726194245653482, + "tokens_seen": 871890944 + }, + { + "epoch": 0.26, + "learning_rate": 0.0005945837512537613, + "loss": 1.5891, + "theoretical_loss": 3.7261671218743135, + "tokens_seen": 871956480 + }, + { + "epoch": 0.26, + "learning_rate": 0.000594567703109328, + "loss": 1.41, + "theoretical_loss": 3.7261400007044467, + "tokens_seen": 872022016 + }, + { + "epoch": 0.26, + "learning_rate": 0.0005945516549648947, + "loss": 1.448, + "theoretical_loss": 3.7261128821434353, + "tokens_seen": 872087552 + }, + { + "epoch": 0.26, + "learning_rate": 0.0005945356068204614, + "loss": 1.44, + "theoretical_loss": 3.7260857661908324, + "tokens_seen": 872153088 + }, + { + "epoch": 0.26, + "learning_rate": 0.0005945195586760281, + "loss": 1.5418, + "theoretical_loss": 3.726058652846191, + "tokens_seen": 872218624 + }, + { + "epoch": 0.26, + "learning_rate": 0.0005945035105315948, + "loss": 1.4029, + "theoretical_loss": 3.726031542109064, + "tokens_seen": 872284160 + }, + { + "epoch": 0.26, + "learning_rate": 0.0005944874623871615, + "loss": 1.4079, + "theoretical_loss": 3.7260044339790053, + "tokens_seen": 872349696 + }, + { + "epoch": 0.26, + "learning_rate": 0.0005944714142427282, + "loss": 1.5069, + "theoretical_loss": 3.7259773284555675, + "tokens_seen": 872415232 + }, + { + "epoch": 0.26, + "learning_rate": 0.0005944553660982949, + "loss": 1.4375, + "theoretical_loss": 3.7259502255383055, + "tokens_seen": 872480768 + }, + { + "epoch": 0.26, + "learning_rate": 0.0005944393179538616, + "loss": 1.4722, + "theoretical_loss": 3.7259231252267715, + "tokens_seen": 872546304 + }, + { + "epoch": 0.26, + "learning_rate": 0.0005944232698094283, + "loss": 1.48, + "theoretical_loss": 3.72589602752052, + "tokens_seen": 872611840 + }, + { + "epoch": 0.26, + "learning_rate": 0.000594407221664995, + "loss": 1.5191, + "theoretical_loss": 3.7258689324191057, + "tokens_seen": 872677376 + }, + { + "epoch": 0.26, + "learning_rate": 0.0005943911735205617, + "loss": 1.4956, + "theoretical_loss": 3.725841839922081, + "tokens_seen": 872742912 + }, + { + "epoch": 0.26, + "learning_rate": 0.0005943751253761284, + "loss": 1.5118, + "theoretical_loss": 3.7258147500290013, + "tokens_seen": 872808448 + }, + { + "epoch": 0.26, + "learning_rate": 0.0005943590772316951, + "loss": 1.4851, + "theoretical_loss": 3.72578766273942, + "tokens_seen": 872873984 + }, + { + "epoch": 0.26, + "learning_rate": 0.0005943430290872618, + "loss": 1.4423, + "theoretical_loss": 3.725760578052892, + "tokens_seen": 872939520 + }, + { + "epoch": 0.26, + "learning_rate": 0.0005943269809428285, + "loss": 1.4398, + "theoretical_loss": 3.725733495968972, + "tokens_seen": 873005056 + }, + { + "epoch": 0.26, + "learning_rate": 0.0005943109327983952, + "loss": 1.5575, + "theoretical_loss": 3.7257064164872142, + "tokens_seen": 873070592 + }, + { + "epoch": 0.26, + "learning_rate": 0.0005942948846539619, + "loss": 1.4349, + "theoretical_loss": 3.7256793396071726, + "tokens_seen": 873136128 + }, + { + "epoch": 0.26, + "learning_rate": 0.0005942788365095286, + "loss": 1.4126, + "theoretical_loss": 3.725652265328403, + "tokens_seen": 873201664 + }, + { + "epoch": 0.26, + "learning_rate": 0.0005942627883650953, + "loss": 1.5104, + "theoretical_loss": 3.7256251936504596, + "tokens_seen": 873267200 + }, + { + "epoch": 0.26, + "learning_rate": 0.000594246740220662, + "loss": 1.5818, + "theoretical_loss": 3.725598124572898, + "tokens_seen": 873332736 + }, + { + "epoch": 0.26, + "learning_rate": 0.0005942306920762287, + "loss": 1.4446, + "theoretical_loss": 3.725571058095273, + "tokens_seen": 873398272 + }, + { + "epoch": 0.26, + "learning_rate": 0.0005942146439317954, + "loss": 1.4238, + "theoretical_loss": 3.7255439942171398, + "tokens_seen": 873463808 + }, + { + "epoch": 0.26, + "learning_rate": 0.0005941985957873621, + "loss": 1.4295, + "theoretical_loss": 3.7255169329380537, + "tokens_seen": 873529344 + }, + { + "epoch": 0.26, + "learning_rate": 0.0005941825476429288, + "loss": 1.5936, + "theoretical_loss": 3.72548987425757, + "tokens_seen": 873594880 + }, + { + "epoch": 0.26, + "learning_rate": 0.0005941664994984955, + "loss": 1.4599, + "theoretical_loss": 3.725462818175245, + "tokens_seen": 873660416 + }, + { + "epoch": 0.26, + "learning_rate": 0.0005941504513540622, + "loss": 1.3563, + "theoretical_loss": 3.7254357646906335, + "tokens_seen": 873725952 + }, + { + "epoch": 0.26, + "learning_rate": 0.000594134403209629, + "loss": 1.4477, + "theoretical_loss": 3.7254087138032914, + "tokens_seen": 873791488 + }, + { + "epoch": 0.26, + "learning_rate": 0.0005941183550651956, + "loss": 1.4996, + "theoretical_loss": 3.725381665512775, + "tokens_seen": 873857024 + }, + { + "epoch": 0.26, + "learning_rate": 0.0005941023069207623, + "loss": 1.4341, + "theoretical_loss": 3.7253546198186402, + "tokens_seen": 873922560 + }, + { + "epoch": 0.26, + "learning_rate": 0.000594086258776329, + "loss": 1.5216, + "theoretical_loss": 3.7253275767204426, + "tokens_seen": 873988096 + }, + { + "epoch": 0.26, + "learning_rate": 0.0005940702106318958, + "loss": 1.5415, + "theoretical_loss": 3.7253005362177385, + "tokens_seen": 874053632 + }, + { + "epoch": 0.26, + "learning_rate": 0.0005940541624874624, + "loss": 1.4835, + "theoretical_loss": 3.7252734983100853, + "tokens_seen": 874119168 + }, + { + "epoch": 0.26, + "learning_rate": 0.0005940381143430292, + "loss": 1.4782, + "theoretical_loss": 3.7252464629970383, + "tokens_seen": 874184704 + }, + { + "epoch": 0.26, + "learning_rate": 0.0005940220661985958, + "loss": 1.3594, + "theoretical_loss": 3.7252194302781536, + "tokens_seen": 874250240 + }, + { + "epoch": 0.26, + "learning_rate": 0.0005940060180541626, + "loss": 1.5194, + "theoretical_loss": 3.7251924001529892, + "tokens_seen": 874315776 + }, + { + "epoch": 0.26, + "learning_rate": 0.0005939899699097292, + "loss": 1.5723, + "theoretical_loss": 3.7251653726211016, + "tokens_seen": 874381312 + }, + { + "epoch": 0.26, + "learning_rate": 0.000593973921765296, + "loss": 1.4943, + "theoretical_loss": 3.7251383476820465, + "tokens_seen": 874446848 + }, + { + "epoch": 0.27, + "learning_rate": 0.0005939578736208626, + "loss": 1.4858, + "theoretical_loss": 3.725111325335382, + "tokens_seen": 874512384 + }, + { + "epoch": 0.27, + "learning_rate": 0.0005939418254764294, + "loss": 1.4701, + "theoretical_loss": 3.7250843055806646, + "tokens_seen": 874577920 + }, + { + "epoch": 0.27, + "learning_rate": 0.000593925777331996, + "loss": 1.3694, + "theoretical_loss": 3.7250572884174518, + "tokens_seen": 874643456 + }, + { + "epoch": 0.27, + "learning_rate": 0.0005939097291875628, + "loss": 1.5512, + "theoretical_loss": 3.7250302738453005, + "tokens_seen": 874708992 + }, + { + "epoch": 0.27, + "learning_rate": 0.0005938936810431294, + "loss": 1.5969, + "theoretical_loss": 3.725003261863769, + "tokens_seen": 874774528 + }, + { + "epoch": 0.27, + "learning_rate": 0.0005938776328986962, + "loss": 1.4248, + "theoretical_loss": 3.724976252472414, + "tokens_seen": 874840064 + }, + { + "epoch": 0.27, + "objective/train/docs_used": 323008, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.4671552181243896, + "objective/train/theoretical_loss": 3.7249492456707936, + "objective/train/tokens_used": 895365600, + "theoretical_loss": 3.7249492456707936, + "tokens_seen": 874905600 + }, + { + "epoch": 0.27, + "learning_rate": 0.0005938615847542628, + "loss": 1.409, + "theoretical_loss": 3.7249492456707936, + "tokens_seen": 874905600 + }, + { + "epoch": 0.27, + "learning_rate": 0.0005938455366098296, + "loss": 1.5974, + "theoretical_loss": 3.724922241458464, + "tokens_seen": 874971136 + }, + { + "epoch": 0.27, + "learning_rate": 0.0005938294884653962, + "loss": 1.3897, + "theoretical_loss": 3.7248952398349857, + "tokens_seen": 875036672 + }, + { + "epoch": 0.27, + "learning_rate": 0.000593813440320963, + "loss": 1.4495, + "theoretical_loss": 3.7248682407999145, + "tokens_seen": 875102208 + }, + { + "epoch": 0.27, + "learning_rate": 0.0005937973921765296, + "loss": 1.4683, + "theoretical_loss": 3.7248412443528096, + "tokens_seen": 875167744 + }, + { + "epoch": 0.27, + "learning_rate": 0.0005937813440320964, + "loss": 1.4117, + "theoretical_loss": 3.7248142504932282, + "tokens_seen": 875233280 + }, + { + "epoch": 0.27, + "learning_rate": 0.000593765295887663, + "loss": 1.4975, + "theoretical_loss": 3.7247872592207303, + "tokens_seen": 875298816 + }, + { + "epoch": 0.27, + "learning_rate": 0.0005937492477432298, + "loss": 1.4929, + "theoretical_loss": 3.7247602705348717, + "tokens_seen": 875364352 + }, + { + "epoch": 0.27, + "learning_rate": 0.0005937331995987964, + "loss": 1.4378, + "theoretical_loss": 3.7247332844352132, + "tokens_seen": 875429888 + }, + { + "epoch": 0.27, + "learning_rate": 0.0005937171514543632, + "loss": 1.4258, + "theoretical_loss": 3.724706300921312, + "tokens_seen": 875495424 + }, + { + "epoch": 0.27, + "learning_rate": 0.0005937011033099298, + "loss": 1.4859, + "theoretical_loss": 3.7246793199927275, + "tokens_seen": 875560960 + }, + { + "epoch": 0.27, + "learning_rate": 0.0005936850551654966, + "loss": 1.6248, + "theoretical_loss": 3.7246523416490187, + "tokens_seen": 875626496 + }, + { + "epoch": 0.27, + "learning_rate": 0.0005936690070210632, + "loss": 1.4624, + "theoretical_loss": 3.724625365889744, + "tokens_seen": 875692032 + }, + { + "epoch": 0.27, + "learning_rate": 0.00059365295887663, + "loss": 1.3817, + "theoretical_loss": 3.724598392714462, + "tokens_seen": 875757568 + }, + { + "epoch": 0.27, + "learning_rate": 0.0005936369107321966, + "loss": 1.4648, + "theoretical_loss": 3.724571422122733, + "tokens_seen": 875823104 + }, + { + "epoch": 0.27, + "learning_rate": 0.0005936208625877634, + "loss": 1.4361, + "theoretical_loss": 3.724544454114115, + "tokens_seen": 875888640 + }, + { + "epoch": 0.27, + "learning_rate": 0.00059360481444333, + "loss": 1.3884, + "theoretical_loss": 3.7245174886881687, + "tokens_seen": 875954176 + }, + { + "epoch": 0.27, + "learning_rate": 0.0005935887662988968, + "loss": 1.5148, + "theoretical_loss": 3.7244905258444527, + "tokens_seen": 876019712 + }, + { + "epoch": 0.27, + "learning_rate": 0.0005935727181544634, + "loss": 1.5324, + "theoretical_loss": 3.7244635655825267, + "tokens_seen": 876085248 + }, + { + "epoch": 0.27, + "learning_rate": 0.0005935566700100302, + "loss": 1.5575, + "theoretical_loss": 3.7244366079019504, + "tokens_seen": 876150784 + }, + { + "epoch": 0.27, + "learning_rate": 0.0005935406218655968, + "loss": 1.4616, + "theoretical_loss": 3.724409652802284, + "tokens_seen": 876216320 + }, + { + "epoch": 0.27, + "learning_rate": 0.0005935245737211636, + "loss": 1.4909, + "theoretical_loss": 3.7243827002830865, + "tokens_seen": 876281856 + }, + { + "epoch": 0.27, + "learning_rate": 0.0005935085255767302, + "loss": 1.4977, + "theoretical_loss": 3.7243557503439186, + "tokens_seen": 876347392 + }, + { + "epoch": 0.27, + "learning_rate": 0.000593492477432297, + "loss": 1.4433, + "theoretical_loss": 3.72432880298434, + "tokens_seen": 876412928 + }, + { + "epoch": 0.27, + "learning_rate": 0.0005934764292878636, + "loss": 1.5695, + "theoretical_loss": 3.7243018582039116, + "tokens_seen": 876478464 + }, + { + "epoch": 0.27, + "learning_rate": 0.0005934603811434304, + "loss": 1.4454, + "theoretical_loss": 3.724274916002193, + "tokens_seen": 876544000 + }, + { + "epoch": 0.27, + "learning_rate": 0.000593444332998997, + "loss": 1.509, + "theoretical_loss": 3.724247976378745, + "tokens_seen": 876609536 + }, + { + "epoch": 0.27, + "learning_rate": 0.0005934282848545637, + "loss": 1.6049, + "theoretical_loss": 3.724221039333128, + "tokens_seen": 876675072 + }, + { + "epoch": 0.27, + "learning_rate": 0.0005934122367101304, + "loss": 1.4504, + "theoretical_loss": 3.724194104864903, + "tokens_seen": 876740608 + }, + { + "epoch": 0.27, + "learning_rate": 0.0005933961885656971, + "loss": 1.5255, + "theoretical_loss": 3.72416717297363, + "tokens_seen": 876806144 + }, + { + "epoch": 0.27, + "learning_rate": 0.0005933801404212638, + "loss": 1.43, + "theoretical_loss": 3.7241402436588706, + "tokens_seen": 876871680 + }, + { + "epoch": 0.27, + "learning_rate": 0.0005933640922768305, + "loss": 1.3964, + "theoretical_loss": 3.724113316920185, + "tokens_seen": 876937216 + }, + { + "epoch": 0.27, + "learning_rate": 0.0005933480441323972, + "loss": 1.5712, + "theoretical_loss": 3.724086392757136, + "tokens_seen": 877002752 + }, + { + "epoch": 0.27, + "learning_rate": 0.0005933319959879639, + "loss": 1.3884, + "theoretical_loss": 3.7240594711692827, + "tokens_seen": 877068288 + }, + { + "epoch": 0.27, + "learning_rate": 0.0005933159478435306, + "loss": 1.473, + "theoretical_loss": 3.7240325521561872, + "tokens_seen": 877133824 + }, + { + "epoch": 0.27, + "learning_rate": 0.0005932998996990973, + "loss": 1.4382, + "theoretical_loss": 3.724005635717411, + "tokens_seen": 877199360 + }, + { + "epoch": 0.27, + "learning_rate": 0.000593283851554664, + "loss": 1.4613, + "theoretical_loss": 3.7239787218525158, + "tokens_seen": 877264896 + }, + { + "epoch": 0.27, + "learning_rate": 0.0005932678034102307, + "loss": 1.4474, + "theoretical_loss": 3.723951810561063, + "tokens_seen": 877330432 + }, + { + "epoch": 0.27, + "learning_rate": 0.0005932517552657974, + "loss": 1.4545, + "theoretical_loss": 3.723924901842614, + "tokens_seen": 877395968 + }, + { + "epoch": 0.27, + "learning_rate": 0.0005932357071213641, + "loss": 1.482, + "theoretical_loss": 3.7238979956967317, + "tokens_seen": 877461504 + }, + { + "epoch": 0.27, + "learning_rate": 0.0005932196589769308, + "loss": 1.474, + "theoretical_loss": 3.7238710921229767, + "tokens_seen": 877527040 + }, + { + "epoch": 0.27, + "learning_rate": 0.0005932036108324975, + "loss": 1.33, + "theoretical_loss": 3.7238441911209117, + "tokens_seen": 877592576 + }, + { + "epoch": 0.27, + "learning_rate": 0.0005931875626880642, + "loss": 1.4492, + "theoretical_loss": 3.723817292690099, + "tokens_seen": 877658112 + }, + { + "epoch": 0.27, + "learning_rate": 0.0005931715145436309, + "loss": 1.531, + "theoretical_loss": 3.723790396830101, + "tokens_seen": 877723648 + }, + { + "epoch": 0.27, + "learning_rate": 0.0005931554663991976, + "loss": 1.4964, + "theoretical_loss": 3.7237635035404795, + "tokens_seen": 877789184 + }, + { + "epoch": 0.27, + "learning_rate": 0.0005931394182547643, + "loss": 1.5146, + "theoretical_loss": 3.723736612820797, + "tokens_seen": 877854720 + }, + { + "epoch": 0.27, + "learning_rate": 0.000593123370110331, + "loss": 1.5006, + "theoretical_loss": 3.7237097246706163, + "tokens_seen": 877920256 + }, + { + "epoch": 0.27, + "learning_rate": 0.0005931073219658977, + "loss": 1.5039, + "theoretical_loss": 3.7236828390895003, + "tokens_seen": 877985792 + }, + { + "epoch": 0.27, + "learning_rate": 0.0005930912738214644, + "loss": 1.5379, + "theoretical_loss": 3.7236559560770113, + "tokens_seen": 878051328 + }, + { + "epoch": 0.27, + "learning_rate": 0.0005930752256770311, + "loss": 1.4962, + "theoretical_loss": 3.7236290756327133, + "tokens_seen": 878116864 + }, + { + "epoch": 0.27, + "objective/train/docs_used": 324331, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.360034465789795, + "objective/train/theoretical_loss": 3.7236021977561675, + "objective/train/tokens_used": 898642400, + "theoretical_loss": 3.7236021977561675, + "tokens_seen": 878182400 + }, + { + "epoch": 0.27, + "learning_rate": 0.0005930591775325978, + "loss": 1.4782, + "theoretical_loss": 3.7236021977561675, + "tokens_seen": 878182400 + }, + { + "epoch": 0.27, + "learning_rate": 0.0005930431293881645, + "loss": 1.4092, + "theoretical_loss": 3.723575322446939, + "tokens_seen": 878247936 + }, + { + "epoch": 0.27, + "learning_rate": 0.0005930270812437312, + "loss": 1.5074, + "theoretical_loss": 3.723548449704589, + "tokens_seen": 878313472 + }, + { + "epoch": 0.27, + "learning_rate": 0.0005930110330992979, + "loss": 1.3661, + "theoretical_loss": 3.7235215795286827, + "tokens_seen": 878379008 + }, + { + "epoch": 0.27, + "learning_rate": 0.0005929949849548646, + "loss": 1.5437, + "theoretical_loss": 3.7234947119187822, + "tokens_seen": 878444544 + }, + { + "epoch": 0.27, + "learning_rate": 0.0005929789368104313, + "loss": 1.3637, + "theoretical_loss": 3.723467846874452, + "tokens_seen": 878510080 + }, + { + "epoch": 0.27, + "learning_rate": 0.000592962888665998, + "loss": 1.6224, + "theoretical_loss": 3.723440984395255, + "tokens_seen": 878575616 + }, + { + "epoch": 0.27, + "learning_rate": 0.0005929468405215647, + "loss": 1.4517, + "theoretical_loss": 3.7234141244807555, + "tokens_seen": 878641152 + }, + { + "epoch": 0.27, + "learning_rate": 0.0005929307923771314, + "loss": 1.4102, + "theoretical_loss": 3.7233872671305175, + "tokens_seen": 878706688 + }, + { + "epoch": 0.27, + "learning_rate": 0.0005929147442326981, + "loss": 1.4059, + "theoretical_loss": 3.723360412344104, + "tokens_seen": 878772224 + }, + { + "epoch": 0.27, + "learning_rate": 0.0005928986960882648, + "loss": 1.4256, + "theoretical_loss": 3.7233335601210804, + "tokens_seen": 878837760 + }, + { + "epoch": 0.27, + "learning_rate": 0.0005928826479438315, + "loss": 1.4955, + "theoretical_loss": 3.7233067104610096, + "tokens_seen": 878903296 + }, + { + "epoch": 0.27, + "learning_rate": 0.0005928665997993982, + "loss": 1.523, + "theoretical_loss": 3.723279863363457, + "tokens_seen": 878968832 + }, + { + "epoch": 0.27, + "learning_rate": 0.0005928505516549649, + "loss": 1.5411, + "theoretical_loss": 3.723253018827986, + "tokens_seen": 879034368 + }, + { + "epoch": 0.27, + "learning_rate": 0.0005928345035105316, + "loss": 1.3445, + "theoretical_loss": 3.723226176854162, + "tokens_seen": 879099904 + }, + { + "epoch": 0.27, + "learning_rate": 0.0005928184553660983, + "loss": 1.4623, + "theoretical_loss": 3.723199337441549, + "tokens_seen": 879165440 + }, + { + "epoch": 0.27, + "learning_rate": 0.000592802407221665, + "loss": 1.4133, + "theoretical_loss": 3.723172500589712, + "tokens_seen": 879230976 + }, + { + "epoch": 0.27, + "learning_rate": 0.0005927863590772317, + "loss": 1.5171, + "theoretical_loss": 3.7231456662982154, + "tokens_seen": 879296512 + }, + { + "epoch": 0.27, + "learning_rate": 0.0005927703109327984, + "loss": 1.4117, + "theoretical_loss": 3.723118834566625, + "tokens_seen": 879362048 + }, + { + "epoch": 0.27, + "learning_rate": 0.0005927542627883651, + "loss": 1.5425, + "theoretical_loss": 3.7230920053945056, + "tokens_seen": 879427584 + }, + { + "epoch": 0.27, + "learning_rate": 0.0005927382146439318, + "loss": 1.5027, + "theoretical_loss": 3.7230651787814217, + "tokens_seen": 879493120 + }, + { + "epoch": 0.27, + "learning_rate": 0.0005927221664994985, + "loss": 1.3741, + "theoretical_loss": 3.7230383547269392, + "tokens_seen": 879558656 + }, + { + "epoch": 0.27, + "learning_rate": 0.0005927061183550653, + "loss": 1.5134, + "theoretical_loss": 3.723011533230623, + "tokens_seen": 879624192 + }, + { + "epoch": 0.27, + "learning_rate": 0.0005926900702106319, + "loss": 1.5323, + "theoretical_loss": 3.7229847142920383, + "tokens_seen": 879689728 + }, + { + "epoch": 0.27, + "learning_rate": 0.0005926740220661987, + "loss": 1.5258, + "theoretical_loss": 3.7229578979107516, + "tokens_seen": 879755264 + }, + { + "epoch": 0.27, + "learning_rate": 0.0005926579739217653, + "loss": 1.4787, + "theoretical_loss": 3.722931084086328, + "tokens_seen": 879820800 + }, + { + "epoch": 0.27, + "learning_rate": 0.000592641925777332, + "loss": 1.5743, + "theoretical_loss": 3.722904272818334, + "tokens_seen": 879886336 + }, + { + "epoch": 0.27, + "learning_rate": 0.0005926258776328987, + "loss": 1.522, + "theoretical_loss": 3.722877464106334, + "tokens_seen": 879951872 + }, + { + "epoch": 0.27, + "learning_rate": 0.0005926098294884655, + "loss": 1.5177, + "theoretical_loss": 3.722850657949895, + "tokens_seen": 880017408 + }, + { + "epoch": 0.27, + "learning_rate": 0.0005925937813440321, + "loss": 1.5313, + "theoretical_loss": 3.722823854348583, + "tokens_seen": 880082944 + }, + { + "epoch": 0.27, + "learning_rate": 0.0005925777331995989, + "loss": 1.5704, + "theoretical_loss": 3.7227970533019645, + "tokens_seen": 880148480 + }, + { + "epoch": 0.27, + "learning_rate": 0.0005925616850551655, + "loss": 1.4194, + "theoretical_loss": 3.7227702548096056, + "tokens_seen": 880214016 + }, + { + "epoch": 0.27, + "learning_rate": 0.0005925456369107323, + "loss": 1.4466, + "theoretical_loss": 3.722743458871072, + "tokens_seen": 880279552 + }, + { + "epoch": 0.27, + "learning_rate": 0.0005925295887662989, + "loss": 1.4533, + "theoretical_loss": 3.7227166654859314, + "tokens_seen": 880345088 + }, + { + "epoch": 0.27, + "learning_rate": 0.0005925135406218657, + "loss": 1.4836, + "theoretical_loss": 3.7226898746537493, + "tokens_seen": 880410624 + }, + { + "epoch": 0.27, + "learning_rate": 0.0005924974924774323, + "loss": 1.3959, + "theoretical_loss": 3.7226630863740935, + "tokens_seen": 880476160 + }, + { + "epoch": 0.27, + "learning_rate": 0.0005924814443329991, + "loss": 1.4952, + "theoretical_loss": 3.7226363006465304, + "tokens_seen": 880541696 + }, + { + "epoch": 0.27, + "learning_rate": 0.0005924653961885657, + "loss": 1.508, + "theoretical_loss": 3.7226095174706266, + "tokens_seen": 880607232 + }, + { + "epoch": 0.27, + "learning_rate": 0.0005924493480441325, + "loss": 1.3837, + "theoretical_loss": 3.7225827368459496, + "tokens_seen": 880672768 + }, + { + "epoch": 0.27, + "learning_rate": 0.0005924332998996991, + "loss": 1.3341, + "theoretical_loss": 3.722555958772066, + "tokens_seen": 880738304 + }, + { + "epoch": 0.27, + "learning_rate": 0.0005924172517552659, + "loss": 1.4029, + "theoretical_loss": 3.722529183248544, + "tokens_seen": 880803840 + }, + { + "epoch": 0.27, + "learning_rate": 0.0005924012036108325, + "loss": 1.3849, + "theoretical_loss": 3.7225024102749504, + "tokens_seen": 880869376 + }, + { + "epoch": 0.27, + "learning_rate": 0.0005923851554663993, + "loss": 1.4305, + "theoretical_loss": 3.7224756398508525, + "tokens_seen": 880934912 + }, + { + "epoch": 0.27, + "learning_rate": 0.0005923691073219659, + "loss": 1.6031, + "theoretical_loss": 3.7224488719758178, + "tokens_seen": 881000448 + }, + { + "epoch": 0.27, + "learning_rate": 0.0005923530591775327, + "loss": 1.6236, + "theoretical_loss": 3.7224221066494145, + "tokens_seen": 881065984 + }, + { + "epoch": 0.27, + "learning_rate": 0.0005923370110330993, + "loss": 1.5382, + "theoretical_loss": 3.7223953438712103, + "tokens_seen": 881131520 + }, + { + "epoch": 0.27, + "learning_rate": 0.0005923209628886661, + "loss": 1.4969, + "theoretical_loss": 3.7223685836407734, + "tokens_seen": 881197056 + }, + { + "epoch": 0.27, + "learning_rate": 0.0005923049147442327, + "loss": 1.4922, + "theoretical_loss": 3.7223418259576704, + "tokens_seen": 881262592 + }, + { + "epoch": 0.27, + "learning_rate": 0.0005922888665997995, + "loss": 1.4563, + "theoretical_loss": 3.7223150708214714, + "tokens_seen": 881328128 + }, + { + "epoch": 0.27, + "learning_rate": 0.0005922728184553661, + "loss": 1.4401, + "theoretical_loss": 3.722288318231743, + "tokens_seen": 881393664 + }, + { + "epoch": 0.27, + "objective/train/docs_used": 325483, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.4156274795532227, + "objective/train/theoretical_loss": 3.722261568188054, + "objective/train/tokens_used": 901919200, + "theoretical_loss": 3.722261568188054, + "tokens_seen": 881459200 + }, + { + "epoch": 0.27, + "learning_rate": 0.0005922567703109329, + "loss": 1.4077, + "theoretical_loss": 3.722261568188054, + "tokens_seen": 881459200 + }, + { + "epoch": 0.27, + "learning_rate": 0.0005922407221664995, + "loss": 1.544, + "theoretical_loss": 3.722234820689973, + "tokens_seen": 881524736 + }, + { + "epoch": 0.27, + "learning_rate": 0.0005922246740220663, + "loss": 1.4139, + "theoretical_loss": 3.722208075737069, + "tokens_seen": 881590272 + }, + { + "epoch": 0.27, + "learning_rate": 0.0005922086258776329, + "loss": 1.5139, + "theoretical_loss": 3.7221813333289093, + "tokens_seen": 881655808 + }, + { + "epoch": 0.27, + "learning_rate": 0.0005921925777331997, + "loss": 1.4815, + "theoretical_loss": 3.722154593465064, + "tokens_seen": 881721344 + }, + { + "epoch": 0.27, + "learning_rate": 0.0005921765295887663, + "loss": 1.5434, + "theoretical_loss": 3.722127856145101, + "tokens_seen": 881786880 + }, + { + "epoch": 0.27, + "learning_rate": 0.0005921604814443331, + "loss": 1.4538, + "theoretical_loss": 3.72210112136859, + "tokens_seen": 881852416 + }, + { + "epoch": 0.27, + "learning_rate": 0.0005921444332998997, + "loss": 1.4853, + "theoretical_loss": 3.7220743891350994, + "tokens_seen": 881917952 + }, + { + "epoch": 0.27, + "learning_rate": 0.0005921283851554665, + "loss": 1.457, + "theoretical_loss": 3.7220476594441987, + "tokens_seen": 881983488 + }, + { + "epoch": 0.27, + "learning_rate": 0.0005921123370110331, + "loss": 1.5114, + "theoretical_loss": 3.7220209322954574, + "tokens_seen": 882049024 + }, + { + "epoch": 0.27, + "learning_rate": 0.0005920962888665999, + "loss": 1.4714, + "theoretical_loss": 3.7219942076884447, + "tokens_seen": 882114560 + }, + { + "epoch": 0.27, + "learning_rate": 0.0005920802407221665, + "loss": 1.4817, + "theoretical_loss": 3.72196748562273, + "tokens_seen": 882180096 + }, + { + "epoch": 0.27, + "learning_rate": 0.0005920641925777333, + "loss": 1.5039, + "theoretical_loss": 3.721940766097882, + "tokens_seen": 882245632 + }, + { + "epoch": 0.27, + "learning_rate": 0.0005920481444332999, + "loss": 1.5179, + "theoretical_loss": 3.721914049113472, + "tokens_seen": 882311168 + }, + { + "epoch": 0.27, + "learning_rate": 0.0005920320962888666, + "loss": 1.4126, + "theoretical_loss": 3.721887334669069, + "tokens_seen": 882376704 + }, + { + "epoch": 0.27, + "learning_rate": 0.0005920160481444333, + "loss": 1.5742, + "theoretical_loss": 3.7218606227642423, + "tokens_seen": 882442240 + }, + { + "epoch": 0.27, + "learning_rate": 0.000592, + "loss": 1.404, + "theoretical_loss": 3.721833913398563, + "tokens_seen": 882507776 + }, + { + "epoch": 0.27, + "learning_rate": 0.0005919839518555667, + "loss": 1.4765, + "theoretical_loss": 3.7218072065716004, + "tokens_seen": 882573312 + }, + { + "epoch": 0.27, + "learning_rate": 0.0005919679037111334, + "loss": 1.5523, + "theoretical_loss": 3.7217805022829253, + "tokens_seen": 882638848 + }, + { + "epoch": 0.27, + "learning_rate": 0.0005919518555667001, + "loss": 1.5706, + "theoretical_loss": 3.721753800532108, + "tokens_seen": 882704384 + }, + { + "epoch": 0.27, + "learning_rate": 0.0005919358074222668, + "loss": 1.493, + "theoretical_loss": 3.7217271013187183, + "tokens_seen": 882769920 + }, + { + "epoch": 0.27, + "learning_rate": 0.0005919197592778335, + "loss": 1.4295, + "theoretical_loss": 3.721700404642327, + "tokens_seen": 882835456 + }, + { + "epoch": 0.27, + "learning_rate": 0.0005919037111334002, + "loss": 1.433, + "theoretical_loss": 3.721673710502505, + "tokens_seen": 882900992 + }, + { + "epoch": 0.27, + "learning_rate": 0.0005918876629889669, + "loss": 1.4141, + "theoretical_loss": 3.7216470188988224, + "tokens_seen": 882966528 + }, + { + "epoch": 0.27, + "learning_rate": 0.0005918716148445336, + "loss": 1.5134, + "theoretical_loss": 3.721620329830851, + "tokens_seen": 883032064 + }, + { + "epoch": 0.27, + "learning_rate": 0.0005918555667001003, + "loss": 1.5174, + "theoretical_loss": 3.721593643298161, + "tokens_seen": 883097600 + }, + { + "epoch": 0.27, + "learning_rate": 0.000591839518555667, + "loss": 1.4161, + "theoretical_loss": 3.7215669593003238, + "tokens_seen": 883163136 + }, + { + "epoch": 0.27, + "learning_rate": 0.0005918234704112337, + "loss": 1.5807, + "theoretical_loss": 3.72154027783691, + "tokens_seen": 883228672 + }, + { + "epoch": 0.27, + "learning_rate": 0.0005918074222668004, + "loss": 1.4258, + "theoretical_loss": 3.7215135989074914, + "tokens_seen": 883294208 + }, + { + "epoch": 0.27, + "learning_rate": 0.0005917913741223671, + "loss": 1.4339, + "theoretical_loss": 3.7214869225116396, + "tokens_seen": 883359744 + }, + { + "epoch": 0.27, + "learning_rate": 0.0005917753259779338, + "loss": 1.4037, + "theoretical_loss": 3.7214602486489254, + "tokens_seen": 883425280 + }, + { + "epoch": 0.27, + "learning_rate": 0.0005917592778335005, + "loss": 1.4609, + "theoretical_loss": 3.721433577318921, + "tokens_seen": 883490816 + }, + { + "epoch": 0.27, + "learning_rate": 0.0005917432296890672, + "loss": 1.3441, + "theoretical_loss": 3.721406908521197, + "tokens_seen": 883556352 + }, + { + "epoch": 0.27, + "learning_rate": 0.0005917271815446339, + "loss": 1.4252, + "theoretical_loss": 3.7213802422553264, + "tokens_seen": 883621888 + }, + { + "epoch": 0.27, + "learning_rate": 0.0005917111334002006, + "loss": 1.4205, + "theoretical_loss": 3.7213535785208807, + "tokens_seen": 883687424 + }, + { + "epoch": 0.27, + "learning_rate": 0.0005916950852557673, + "loss": 1.5435, + "theoretical_loss": 3.7213269173174313, + "tokens_seen": 883752960 + }, + { + "epoch": 0.27, + "learning_rate": 0.000591679037111334, + "loss": 1.4307, + "theoretical_loss": 3.7213002586445514, + "tokens_seen": 883818496 + }, + { + "epoch": 0.27, + "learning_rate": 0.0005916629889669007, + "loss": 1.4969, + "theoretical_loss": 3.721273602501812, + "tokens_seen": 883884032 + }, + { + "epoch": 0.27, + "learning_rate": 0.0005916469408224674, + "loss": 1.4266, + "theoretical_loss": 3.721246948888786, + "tokens_seen": 883949568 + }, + { + "epoch": 0.27, + "learning_rate": 0.0005916308926780341, + "loss": 1.5853, + "theoretical_loss": 3.7212202978050457, + "tokens_seen": 884015104 + }, + { + "epoch": 0.27, + "learning_rate": 0.0005916148445336008, + "loss": 1.4433, + "theoretical_loss": 3.721193649250164, + "tokens_seen": 884080640 + }, + { + "epoch": 0.27, + "learning_rate": 0.0005915987963891675, + "loss": 1.4928, + "theoretical_loss": 3.7211670032237123, + "tokens_seen": 884146176 + }, + { + "epoch": 0.27, + "learning_rate": 0.0005915827482447342, + "loss": 1.4195, + "theoretical_loss": 3.721140359725265, + "tokens_seen": 884211712 + }, + { + "epoch": 0.27, + "learning_rate": 0.0005915667001003009, + "loss": 1.4651, + "theoretical_loss": 3.7211137187543937, + "tokens_seen": 884277248 + }, + { + "epoch": 0.27, + "learning_rate": 0.0005915506519558676, + "loss": 1.5149, + "theoretical_loss": 3.7210870803106717, + "tokens_seen": 884342784 + }, + { + "epoch": 0.27, + "learning_rate": 0.0005915346038114343, + "loss": 1.2919, + "theoretical_loss": 3.721060444393672, + "tokens_seen": 884408320 + }, + { + "epoch": 0.27, + "learning_rate": 0.000591518555667001, + "loss": 1.5165, + "theoretical_loss": 3.721033811002968, + "tokens_seen": 884473856 + }, + { + "epoch": 0.27, + "learning_rate": 0.0005915025075225677, + "loss": 1.4611, + "theoretical_loss": 3.721007180138132, + "tokens_seen": 884539392 + }, + { + "epoch": 0.27, + "learning_rate": 0.0005914864593781344, + "loss": 1.3328, + "theoretical_loss": 3.720980551798739, + "tokens_seen": 884604928 + }, + { + "epoch": 0.27, + "learning_rate": 0.0005914704112337011, + "loss": 1.5308, + "theoretical_loss": 3.720953925984361, + "tokens_seen": 884670464 + }, + { + "epoch": 0.27, + "objective/train/docs_used": 326745, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.6591248512268066, + "objective/train/theoretical_loss": 3.720927302694572, + "objective/train/tokens_used": 905196000, + "theoretical_loss": 3.720927302694572, + "tokens_seen": 884736000 + }, + { + "epoch": 0.27, + "learning_rate": 0.0005914543630892678, + "loss": 1.5318, + "theoretical_loss": 3.720927302694572, + "tokens_seen": 884736000 + }, + { + "epoch": 0.27, + "learning_rate": 0.0005914383149448345, + "loss": 1.481, + "theoretical_loss": 3.7209006819289456, + "tokens_seen": 884801536 + }, + { + "epoch": 0.27, + "learning_rate": 0.0005914222668004012, + "loss": 1.488, + "theoretical_loss": 3.7208740636870554, + "tokens_seen": 884867072 + }, + { + "epoch": 0.27, + "learning_rate": 0.000591406218655968, + "loss": 1.5023, + "theoretical_loss": 3.7208474479684757, + "tokens_seen": 884932608 + }, + { + "epoch": 0.27, + "learning_rate": 0.0005913901705115346, + "loss": 1.5396, + "theoretical_loss": 3.72082083477278, + "tokens_seen": 884998144 + }, + { + "epoch": 0.27, + "learning_rate": 0.0005913741223671013, + "loss": 1.486, + "theoretical_loss": 3.7207942240995426, + "tokens_seen": 885063680 + }, + { + "epoch": 0.27, + "learning_rate": 0.000591358074222668, + "loss": 1.5031, + "theoretical_loss": 3.7207676159483376, + "tokens_seen": 885129216 + }, + { + "epoch": 0.27, + "learning_rate": 0.0005913420260782348, + "loss": 1.3271, + "theoretical_loss": 3.720741010318739, + "tokens_seen": 885194752 + }, + { + "epoch": 0.27, + "learning_rate": 0.0005913259779338014, + "loss": 1.4082, + "theoretical_loss": 3.720714407210322, + "tokens_seen": 885260288 + }, + { + "epoch": 0.27, + "learning_rate": 0.0005913099297893682, + "loss": 1.3689, + "theoretical_loss": 3.72068780662266, + "tokens_seen": 885325824 + }, + { + "epoch": 0.27, + "learning_rate": 0.0005912938816449348, + "loss": 1.4404, + "theoretical_loss": 3.7206612085553283, + "tokens_seen": 885391360 + }, + { + "epoch": 0.27, + "learning_rate": 0.0005912778335005016, + "loss": 1.4815, + "theoretical_loss": 3.7206346130079018, + "tokens_seen": 885456896 + }, + { + "epoch": 0.27, + "learning_rate": 0.0005912617853560682, + "loss": 1.5034, + "theoretical_loss": 3.720608019979954, + "tokens_seen": 885522432 + }, + { + "epoch": 0.27, + "learning_rate": 0.000591245737211635, + "loss": 1.5218, + "theoretical_loss": 3.7205814294710615, + "tokens_seen": 885587968 + }, + { + "epoch": 0.27, + "learning_rate": 0.0005912296890672016, + "loss": 1.4494, + "theoretical_loss": 3.720554841480798, + "tokens_seen": 885653504 + }, + { + "epoch": 0.27, + "learning_rate": 0.0005912136409227684, + "loss": 1.4774, + "theoretical_loss": 3.720528256008739, + "tokens_seen": 885719040 + }, + { + "epoch": 0.27, + "learning_rate": 0.000591197592778335, + "loss": 1.5399, + "theoretical_loss": 3.72050167305446, + "tokens_seen": 885784576 + }, + { + "epoch": 0.27, + "learning_rate": 0.0005911815446339018, + "loss": 1.4265, + "theoretical_loss": 3.7204750926175354, + "tokens_seen": 885850112 + }, + { + "epoch": 0.27, + "learning_rate": 0.0005911654964894684, + "loss": 1.4649, + "theoretical_loss": 3.7204485146975417, + "tokens_seen": 885915648 + }, + { + "epoch": 0.27, + "learning_rate": 0.0005911494483450352, + "loss": 1.4891, + "theoretical_loss": 3.720421939294054, + "tokens_seen": 885981184 + }, + { + "epoch": 0.27, + "learning_rate": 0.0005911334002006018, + "loss": 1.4369, + "theoretical_loss": 3.7203953664066476, + "tokens_seen": 886046720 + }, + { + "epoch": 0.27, + "learning_rate": 0.0005911173520561686, + "loss": 1.4935, + "theoretical_loss": 3.7203687960348986, + "tokens_seen": 886112256 + }, + { + "epoch": 0.27, + "learning_rate": 0.0005911013039117352, + "loss": 1.3896, + "theoretical_loss": 3.7203422281783824, + "tokens_seen": 886177792 + }, + { + "epoch": 0.27, + "learning_rate": 0.000591085255767302, + "loss": 1.5376, + "theoretical_loss": 3.7203156628366756, + "tokens_seen": 886243328 + }, + { + "epoch": 0.27, + "learning_rate": 0.0005910692076228686, + "loss": 1.555, + "theoretical_loss": 3.720289100009353, + "tokens_seen": 886308864 + }, + { + "epoch": 0.27, + "learning_rate": 0.0005910531594784354, + "loss": 1.3456, + "theoretical_loss": 3.7202625396959923, + "tokens_seen": 886374400 + }, + { + "epoch": 0.27, + "learning_rate": 0.000591037111334002, + "loss": 1.4594, + "theoretical_loss": 3.7202359818961686, + "tokens_seen": 886439936 + }, + { + "epoch": 0.27, + "learning_rate": 0.0005910210631895688, + "loss": 1.5066, + "theoretical_loss": 3.7202094266094585, + "tokens_seen": 886505472 + }, + { + "epoch": 0.27, + "learning_rate": 0.0005910050150451354, + "loss": 1.4671, + "theoretical_loss": 3.720182873835439, + "tokens_seen": 886571008 + }, + { + "epoch": 0.27, + "learning_rate": 0.0005909889669007022, + "loss": 1.3683, + "theoretical_loss": 3.7201563235736854, + "tokens_seen": 886636544 + }, + { + "epoch": 0.27, + "learning_rate": 0.0005909729187562689, + "loss": 1.4242, + "theoretical_loss": 3.7201297758237755, + "tokens_seen": 886702080 + }, + { + "epoch": 0.27, + "learning_rate": 0.0005909568706118356, + "loss": 1.4848, + "theoretical_loss": 3.720103230585286, + "tokens_seen": 886767616 + }, + { + "epoch": 0.27, + "learning_rate": 0.0005909408224674023, + "loss": 1.4524, + "theoretical_loss": 3.7200766878577927, + "tokens_seen": 886833152 + }, + { + "epoch": 0.27, + "learning_rate": 0.000590924774322969, + "loss": 1.4669, + "theoretical_loss": 3.720050147640873, + "tokens_seen": 886898688 + }, + { + "epoch": 0.27, + "learning_rate": 0.0005909087261785357, + "loss": 1.4816, + "theoretical_loss": 3.7200236099341044, + "tokens_seen": 886964224 + }, + { + "epoch": 0.27, + "learning_rate": 0.0005908926780341024, + "loss": 1.4748, + "theoretical_loss": 3.719997074737064, + "tokens_seen": 887029760 + }, + { + "epoch": 0.27, + "learning_rate": 0.0005908766298896691, + "loss": 1.5641, + "theoretical_loss": 3.7199705420493285, + "tokens_seen": 887095296 + }, + { + "epoch": 0.27, + "learning_rate": 0.0005908605817452358, + "loss": 1.5003, + "theoretical_loss": 3.719944011870476, + "tokens_seen": 887160832 + }, + { + "epoch": 0.27, + "learning_rate": 0.0005908445336008025, + "loss": 1.4545, + "theoretical_loss": 3.7199174842000833, + "tokens_seen": 887226368 + }, + { + "epoch": 0.27, + "learning_rate": 0.0005908284854563692, + "loss": 1.5098, + "theoretical_loss": 3.719890959037728, + "tokens_seen": 887291904 + }, + { + "epoch": 0.27, + "learning_rate": 0.0005908124373119359, + "loss": 1.46, + "theoretical_loss": 3.719864436382988, + "tokens_seen": 887357440 + }, + { + "epoch": 0.27, + "learning_rate": 0.0005907963891675026, + "loss": 1.4958, + "theoretical_loss": 3.7198379162354414, + "tokens_seen": 887422976 + }, + { + "epoch": 0.27, + "learning_rate": 0.0005907803410230693, + "loss": 1.4506, + "theoretical_loss": 3.7198113985946657, + "tokens_seen": 887488512 + }, + { + "epoch": 0.27, + "learning_rate": 0.000590764292878636, + "loss": 1.5191, + "theoretical_loss": 3.719784883460238, + "tokens_seen": 887554048 + }, + { + "epoch": 0.27, + "learning_rate": 0.0005907482447342026, + "loss": 1.3856, + "theoretical_loss": 3.719758370831738, + "tokens_seen": 887619584 + }, + { + "epoch": 0.27, + "learning_rate": 0.0005907321965897694, + "loss": 1.4117, + "theoretical_loss": 3.719731860708743, + "tokens_seen": 887685120 + }, + { + "epoch": 0.27, + "learning_rate": 0.000590716148445336, + "loss": 1.4089, + "theoretical_loss": 3.7197053530908315, + "tokens_seen": 887750656 + }, + { + "epoch": 0.27, + "learning_rate": 0.0005907001003009028, + "loss": 1.4987, + "theoretical_loss": 3.7196788479775815, + "tokens_seen": 887816192 + }, + { + "epoch": 0.27, + "learning_rate": 0.0005906840521564694, + "loss": 1.3813, + "theoretical_loss": 3.7196523453685715, + "tokens_seen": 887881728 + }, + { + "epoch": 0.27, + "learning_rate": 0.0005906680040120362, + "loss": 1.5205, + "theoretical_loss": 3.719625845263381, + "tokens_seen": 887947264 + }, + { + "epoch": 0.27, + "objective/train/docs_used": 327973, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.3294477462768555, + "objective/train/theoretical_loss": 3.719599347661587, + "objective/train/tokens_used": 908472800, + "theoretical_loss": 3.719599347661587, + "tokens_seen": 888012800 + }, + { + "epoch": 0.27, + "learning_rate": 0.0005906519558676028, + "loss": 1.4001, + "theoretical_loss": 3.719599347661587, + "tokens_seen": 888012800 + }, + { + "epoch": 0.27, + "learning_rate": 0.0005906359077231695, + "loss": 1.536, + "theoretical_loss": 3.71957285256277, + "tokens_seen": 888078336 + }, + { + "epoch": 0.27, + "learning_rate": 0.0005906198595787362, + "loss": 1.4498, + "theoretical_loss": 3.7195463599665075, + "tokens_seen": 888143872 + }, + { + "epoch": 0.27, + "learning_rate": 0.0005906038114343029, + "loss": 1.48, + "theoretical_loss": 3.7195198698723795, + "tokens_seen": 888209408 + }, + { + "epoch": 0.27, + "learning_rate": 0.0005905877632898696, + "loss": 1.4333, + "theoretical_loss": 3.719493382279965, + "tokens_seen": 888274944 + }, + { + "epoch": 0.27, + "learning_rate": 0.0005905717151454363, + "loss": 1.4408, + "theoretical_loss": 3.719466897188843, + "tokens_seen": 888340480 + }, + { + "epoch": 0.27, + "learning_rate": 0.000590555667001003, + "loss": 1.489, + "theoretical_loss": 3.7194404145985924, + "tokens_seen": 888406016 + }, + { + "epoch": 0.27, + "learning_rate": 0.0005905396188565697, + "loss": 1.3721, + "theoretical_loss": 3.719413934508793, + "tokens_seen": 888471552 + }, + { + "epoch": 0.27, + "learning_rate": 0.0005905235707121364, + "loss": 1.4819, + "theoretical_loss": 3.719387456919024, + "tokens_seen": 888537088 + }, + { + "epoch": 0.27, + "learning_rate": 0.0005905075225677031, + "loss": 1.4704, + "theoretical_loss": 3.7193609818288653, + "tokens_seen": 888602624 + }, + { + "epoch": 0.27, + "learning_rate": 0.0005904914744232698, + "loss": 1.3769, + "theoretical_loss": 3.7193345092378967, + "tokens_seen": 888668160 + }, + { + "epoch": 0.27, + "learning_rate": 0.0005904754262788365, + "loss": 1.4364, + "theoretical_loss": 3.719308039145698, + "tokens_seen": 888733696 + }, + { + "epoch": 0.27, + "learning_rate": 0.0005904593781344032, + "loss": 1.4899, + "theoretical_loss": 3.7192815715518486, + "tokens_seen": 888799232 + }, + { + "epoch": 0.27, + "learning_rate": 0.0005904433299899699, + "loss": 1.4509, + "theoretical_loss": 3.719255106455929, + "tokens_seen": 888864768 + }, + { + "epoch": 0.27, + "learning_rate": 0.0005904272818455366, + "loss": 1.5939, + "theoretical_loss": 3.7192286438575186, + "tokens_seen": 888930304 + }, + { + "epoch": 0.27, + "learning_rate": 0.0005904112337011033, + "loss": 1.4308, + "theoretical_loss": 3.7192021837561984, + "tokens_seen": 888995840 + }, + { + "epoch": 0.27, + "learning_rate": 0.00059039518555667, + "loss": 1.4493, + "theoretical_loss": 3.7191757261515486, + "tokens_seen": 889061376 + }, + { + "epoch": 0.27, + "learning_rate": 0.0005903791374122367, + "loss": 1.4389, + "theoretical_loss": 3.7191492710431495, + "tokens_seen": 889126912 + }, + { + "epoch": 0.27, + "learning_rate": 0.0005903630892678034, + "loss": 1.3965, + "theoretical_loss": 3.7191228184305816, + "tokens_seen": 889192448 + }, + { + "epoch": 0.27, + "learning_rate": 0.0005903470411233701, + "loss": 1.4973, + "theoretical_loss": 3.719096368313425, + "tokens_seen": 889257984 + }, + { + "epoch": 0.27, + "learning_rate": 0.0005903309929789368, + "loss": 1.5081, + "theoretical_loss": 3.719069920691261, + "tokens_seen": 889323520 + }, + { + "epoch": 0.27, + "learning_rate": 0.0005903149448345035, + "loss": 1.5112, + "theoretical_loss": 3.7190434755636708, + "tokens_seen": 889389056 + }, + { + "epoch": 0.27, + "learning_rate": 0.0005902988966900702, + "loss": 1.4631, + "theoretical_loss": 3.7190170329302346, + "tokens_seen": 889454592 + }, + { + "epoch": 0.27, + "learning_rate": 0.0005902828485456369, + "loss": 1.535, + "theoretical_loss": 3.718990592790534, + "tokens_seen": 889520128 + }, + { + "epoch": 0.27, + "learning_rate": 0.0005902668004012036, + "loss": 1.492, + "theoretical_loss": 3.718964155144149, + "tokens_seen": 889585664 + }, + { + "epoch": 0.27, + "learning_rate": 0.0005902507522567703, + "loss": 1.3621, + "theoretical_loss": 3.718937719990662, + "tokens_seen": 889651200 + }, + { + "epoch": 0.27, + "learning_rate": 0.000590234704112337, + "loss": 1.4744, + "theoretical_loss": 3.7189112873296537, + "tokens_seen": 889716736 + }, + { + "epoch": 0.27, + "learning_rate": 0.0005902186559679037, + "loss": 1.4643, + "theoretical_loss": 3.718884857160706, + "tokens_seen": 889782272 + }, + { + "epoch": 0.27, + "learning_rate": 0.0005902026078234704, + "loss": 1.5312, + "theoretical_loss": 3.7188584294834004, + "tokens_seen": 889847808 + }, + { + "epoch": 0.27, + "learning_rate": 0.0005901865596790371, + "loss": 1.4797, + "theoretical_loss": 3.718832004297318, + "tokens_seen": 889913344 + }, + { + "epoch": 0.27, + "learning_rate": 0.0005901705115346038, + "loss": 1.5032, + "theoretical_loss": 3.718805581602041, + "tokens_seen": 889978880 + }, + { + "epoch": 0.27, + "learning_rate": 0.0005901544633901705, + "loss": 1.5191, + "theoretical_loss": 3.7187791613971504, + "tokens_seen": 890044416 + }, + { + "epoch": 0.27, + "learning_rate": 0.0005901384152457372, + "loss": 1.4844, + "theoretical_loss": 3.7187527436822294, + "tokens_seen": 890109952 + }, + { + "epoch": 0.27, + "learning_rate": 0.0005901223671013039, + "loss": 1.4812, + "theoretical_loss": 3.7187263284568592, + "tokens_seen": 890175488 + }, + { + "epoch": 0.27, + "learning_rate": 0.0005901063189568706, + "loss": 1.415, + "theoretical_loss": 3.7186999157206224, + "tokens_seen": 890241024 + }, + { + "epoch": 0.27, + "learning_rate": 0.0005900902708124373, + "loss": 1.3629, + "theoretical_loss": 3.7186735054731006, + "tokens_seen": 890306560 + }, + { + "epoch": 0.27, + "learning_rate": 0.000590074222668004, + "loss": 1.4108, + "theoretical_loss": 3.718647097713877, + "tokens_seen": 890372096 + }, + { + "epoch": 0.27, + "learning_rate": 0.0005900581745235707, + "loss": 1.4453, + "theoretical_loss": 3.718620692442533, + "tokens_seen": 890437632 + }, + { + "epoch": 0.27, + "learning_rate": 0.0005900421263791374, + "loss": 1.4529, + "theoretical_loss": 3.718594289658652, + "tokens_seen": 890503168 + }, + { + "epoch": 0.27, + "learning_rate": 0.0005900260782347041, + "loss": 1.4626, + "theoretical_loss": 3.7185678893618164, + "tokens_seen": 890568704 + }, + { + "epoch": 0.27, + "learning_rate": 0.0005900100300902708, + "loss": 1.4801, + "theoretical_loss": 3.7185414915516084, + "tokens_seen": 890634240 + }, + { + "epoch": 0.27, + "learning_rate": 0.0005899939819458375, + "loss": 1.4559, + "theoretical_loss": 3.7185150962276117, + "tokens_seen": 890699776 + }, + { + "epoch": 0.27, + "learning_rate": 0.0005899779338014042, + "loss": 1.4777, + "theoretical_loss": 3.7184887033894087, + "tokens_seen": 890765312 + }, + { + "epoch": 0.27, + "learning_rate": 0.0005899618856569709, + "loss": 1.3783, + "theoretical_loss": 3.718462313036582, + "tokens_seen": 890830848 + }, + { + "epoch": 0.27, + "learning_rate": 0.0005899458375125377, + "loss": 1.4508, + "theoretical_loss": 3.7184359251687162, + "tokens_seen": 890896384 + }, + { + "epoch": 0.27, + "learning_rate": 0.0005899297893681043, + "loss": 1.4627, + "theoretical_loss": 3.718409539785393, + "tokens_seen": 890961920 + }, + { + "epoch": 0.27, + "learning_rate": 0.000589913741223671, + "loss": 1.5013, + "theoretical_loss": 3.7183831568861967, + "tokens_seen": 891027456 + }, + { + "epoch": 0.27, + "learning_rate": 0.0005898976930792377, + "loss": 1.4259, + "theoretical_loss": 3.718356776470711, + "tokens_seen": 891092992 + }, + { + "epoch": 0.27, + "learning_rate": 0.0005898816449348045, + "loss": 1.5805, + "theoretical_loss": 3.7183303985385177, + "tokens_seen": 891158528 + }, + { + "epoch": 0.27, + "learning_rate": 0.0005898655967903711, + "loss": 1.5878, + "theoretical_loss": 3.7183040230892024, + "tokens_seen": 891224064 + }, + { + "epoch": 0.27, + "objective/train/docs_used": 329220, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.385576605796814, + "objective/train/theoretical_loss": 3.7182776501223476, + "objective/train/tokens_used": 911749600, + "theoretical_loss": 3.7182776501223476, + "tokens_seen": 891289600 + }, + { + "epoch": 0.27, + "learning_rate": 0.0005898495486459379, + "loss": 1.4186, + "theoretical_loss": 3.7182776501223476, + "tokens_seen": 891289600 + }, + { + "epoch": 0.27, + "learning_rate": 0.0005898335005015045, + "loss": 1.3695, + "theoretical_loss": 3.7182512796375375, + "tokens_seen": 891355136 + }, + { + "epoch": 0.27, + "learning_rate": 0.0005898174523570713, + "loss": 1.4487, + "theoretical_loss": 3.7182249116343566, + "tokens_seen": 891420672 + }, + { + "epoch": 0.27, + "learning_rate": 0.0005898014042126379, + "loss": 1.5463, + "theoretical_loss": 3.7181985461123883, + "tokens_seen": 891486208 + }, + { + "epoch": 0.27, + "learning_rate": 0.0005897853560682047, + "loss": 1.5019, + "theoretical_loss": 3.718172183071217, + "tokens_seen": 891551744 + }, + { + "epoch": 0.27, + "learning_rate": 0.0005897693079237714, + "loss": 1.5411, + "theoretical_loss": 3.7181458225104267, + "tokens_seen": 891617280 + }, + { + "epoch": 0.27, + "learning_rate": 0.0005897532597793381, + "loss": 1.3387, + "theoretical_loss": 3.718119464429602, + "tokens_seen": 891682816 + }, + { + "epoch": 0.27, + "learning_rate": 0.0005897372116349048, + "loss": 1.5025, + "theoretical_loss": 3.718093108828327, + "tokens_seen": 891748352 + }, + { + "epoch": 0.27, + "learning_rate": 0.0005897211634904715, + "loss": 1.438, + "theoretical_loss": 3.718066755706187, + "tokens_seen": 891813888 + }, + { + "epoch": 0.27, + "learning_rate": 0.0005897051153460382, + "loss": 1.4184, + "theoretical_loss": 3.718040405062766, + "tokens_seen": 891879424 + }, + { + "epoch": 0.27, + "learning_rate": 0.0005896890672016049, + "loss": 1.5, + "theoretical_loss": 3.7180140568976485, + "tokens_seen": 891944960 + }, + { + "epoch": 0.27, + "learning_rate": 0.0005896730190571716, + "loss": 1.4851, + "theoretical_loss": 3.7179877112104194, + "tokens_seen": 892010496 + }, + { + "epoch": 0.27, + "learning_rate": 0.0005896569709127383, + "loss": 1.4185, + "theoretical_loss": 3.7179613680006645, + "tokens_seen": 892076032 + }, + { + "epoch": 0.27, + "learning_rate": 0.000589640922768305, + "loss": 1.4965, + "theoretical_loss": 3.7179350272679685, + "tokens_seen": 892141568 + }, + { + "epoch": 0.27, + "learning_rate": 0.0005896248746238717, + "loss": 1.5356, + "theoretical_loss": 3.7179086890119164, + "tokens_seen": 892207104 + }, + { + "epoch": 0.27, + "learning_rate": 0.0005896088264794384, + "loss": 1.4728, + "theoretical_loss": 3.717882353232093, + "tokens_seen": 892272640 + }, + { + "epoch": 0.27, + "learning_rate": 0.0005895927783350051, + "loss": 1.5195, + "theoretical_loss": 3.7178560199280843, + "tokens_seen": 892338176 + }, + { + "epoch": 0.27, + "learning_rate": 0.0005895767301905718, + "loss": 1.4678, + "theoretical_loss": 3.7178296890994753, + "tokens_seen": 892403712 + }, + { + "epoch": 0.27, + "learning_rate": 0.0005895606820461385, + "loss": 1.5673, + "theoretical_loss": 3.717803360745852, + "tokens_seen": 892469248 + }, + { + "epoch": 0.27, + "learning_rate": 0.0005895446339017052, + "loss": 1.5381, + "theoretical_loss": 3.717777034866799, + "tokens_seen": 892534784 + }, + { + "epoch": 0.27, + "learning_rate": 0.0005895285857572719, + "loss": 1.4861, + "theoretical_loss": 3.7177507114619037, + "tokens_seen": 892600320 + }, + { + "epoch": 0.27, + "learning_rate": 0.0005895125376128386, + "loss": 1.5857, + "theoretical_loss": 3.7177243905307504, + "tokens_seen": 892665856 + }, + { + "epoch": 0.27, + "learning_rate": 0.0005894964894684053, + "loss": 1.4618, + "theoretical_loss": 3.7176980720729267, + "tokens_seen": 892731392 + }, + { + "epoch": 0.27, + "learning_rate": 0.000589480441323972, + "loss": 1.5968, + "theoretical_loss": 3.717671756088017, + "tokens_seen": 892796928 + }, + { + "epoch": 0.27, + "learning_rate": 0.0005894643931795387, + "loss": 1.4248, + "theoretical_loss": 3.717645442575608, + "tokens_seen": 892862464 + }, + { + "epoch": 0.27, + "learning_rate": 0.0005894483450351054, + "loss": 1.5087, + "theoretical_loss": 3.717619131535286, + "tokens_seen": 892928000 + }, + { + "epoch": 0.27, + "learning_rate": 0.0005894322968906721, + "loss": 1.5724, + "theoretical_loss": 3.7175928229666377, + "tokens_seen": 892993536 + }, + { + "epoch": 0.27, + "learning_rate": 0.0005894162487462388, + "loss": 1.5341, + "theoretical_loss": 3.717566516869249, + "tokens_seen": 893059072 + }, + { + "epoch": 0.27, + "learning_rate": 0.0005894002006018055, + "loss": 1.3194, + "theoretical_loss": 3.7175402132427067, + "tokens_seen": 893124608 + }, + { + "epoch": 0.27, + "learning_rate": 0.0005893841524573722, + "loss": 1.5116, + "theoretical_loss": 3.7175139120865976, + "tokens_seen": 893190144 + }, + { + "epoch": 0.27, + "learning_rate": 0.0005893681043129389, + "loss": 1.4093, + "theoretical_loss": 3.7174876134005075, + "tokens_seen": 893255680 + }, + { + "epoch": 0.27, + "learning_rate": 0.0005893520561685055, + "loss": 1.308, + "theoretical_loss": 3.717461317184025, + "tokens_seen": 893321216 + }, + { + "epoch": 0.27, + "learning_rate": 0.0005893360080240723, + "loss": 1.4838, + "theoretical_loss": 3.7174350234367353, + "tokens_seen": 893386752 + }, + { + "epoch": 0.27, + "learning_rate": 0.0005893199598796389, + "loss": 1.4642, + "theoretical_loss": 3.7174087321582263, + "tokens_seen": 893452288 + }, + { + "epoch": 0.27, + "learning_rate": 0.0005893039117352057, + "loss": 1.4458, + "theoretical_loss": 3.717382443348085, + "tokens_seen": 893517824 + }, + { + "epoch": 0.27, + "learning_rate": 0.0005892878635907723, + "loss": 1.4529, + "theoretical_loss": 3.717356157005899, + "tokens_seen": 893583360 + }, + { + "epoch": 0.27, + "learning_rate": 0.0005892718154463391, + "loss": 1.5431, + "theoretical_loss": 3.717329873131255, + "tokens_seen": 893648896 + }, + { + "epoch": 0.27, + "learning_rate": 0.0005892557673019057, + "loss": 1.5511, + "theoretical_loss": 3.7173035917237405, + "tokens_seen": 893714432 + }, + { + "epoch": 0.27, + "learning_rate": 0.0005892397191574725, + "loss": 1.4935, + "theoretical_loss": 3.7172773127829437, + "tokens_seen": 893779968 + }, + { + "epoch": 0.27, + "learning_rate": 0.0005892236710130391, + "loss": 1.5044, + "theoretical_loss": 3.7172510363084514, + "tokens_seen": 893845504 + }, + { + "epoch": 0.27, + "learning_rate": 0.0005892076228686058, + "loss": 1.376, + "theoretical_loss": 3.717224762299852, + "tokens_seen": 893911040 + }, + { + "epoch": 0.27, + "learning_rate": 0.0005891915747241725, + "loss": 1.459, + "theoretical_loss": 3.717198490756733, + "tokens_seen": 893976576 + }, + { + "epoch": 0.27, + "learning_rate": 0.0005891755265797392, + "loss": 1.4308, + "theoretical_loss": 3.717172221678682, + "tokens_seen": 894042112 + }, + { + "epoch": 0.27, + "learning_rate": 0.0005891594784353059, + "loss": 1.5006, + "theoretical_loss": 3.7171459550652877, + "tokens_seen": 894107648 + }, + { + "epoch": 0.27, + "learning_rate": 0.0005891434302908726, + "loss": 1.5431, + "theoretical_loss": 3.717119690916138, + "tokens_seen": 894173184 + }, + { + "epoch": 0.27, + "learning_rate": 0.0005891273821464393, + "loss": 1.5385, + "theoretical_loss": 3.717093429230821, + "tokens_seen": 894238720 + }, + { + "epoch": 0.27, + "learning_rate": 0.000589111334002006, + "loss": 1.4541, + "theoretical_loss": 3.7170671700089244, + "tokens_seen": 894304256 + }, + { + "epoch": 0.27, + "learning_rate": 0.0005890952858575727, + "loss": 1.4631, + "theoretical_loss": 3.717040913250038, + "tokens_seen": 894369792 + }, + { + "epoch": 0.27, + "learning_rate": 0.0005890792377131394, + "loss": 1.3626, + "theoretical_loss": 3.7170146589537496, + "tokens_seen": 894435328 + }, + { + "epoch": 0.27, + "learning_rate": 0.0005890631895687061, + "loss": 1.6522, + "theoretical_loss": 3.716988407119648, + "tokens_seen": 894500864 + }, + { + "epoch": 0.27, + "objective/train/docs_used": 330336, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.3582024574279785, + "objective/train/theoretical_loss": 3.7169621577473215, + "objective/train/tokens_used": 915026400, + "theoretical_loss": 3.7169621577473215, + "tokens_seen": 894566400 + }, + { + "epoch": 0.27, + "learning_rate": 0.0005890471414242728, + "loss": 1.4285, + "theoretical_loss": 3.7169621577473215, + "tokens_seen": 894566400 + }, + { + "epoch": 0.27, + "learning_rate": 0.0005890310932798395, + "loss": 1.4653, + "theoretical_loss": 3.716935910836359, + "tokens_seen": 894631936 + }, + { + "epoch": 0.27, + "learning_rate": 0.0005890150451354062, + "loss": 1.4596, + "theoretical_loss": 3.71690966638635, + "tokens_seen": 894697472 + }, + { + "epoch": 0.27, + "learning_rate": 0.0005889989969909729, + "loss": 1.4186, + "theoretical_loss": 3.716883424396883, + "tokens_seen": 894763008 + }, + { + "epoch": 0.27, + "learning_rate": 0.0005889829488465396, + "loss": 1.5831, + "theoretical_loss": 3.716857184867547, + "tokens_seen": 894828544 + }, + { + "epoch": 0.27, + "learning_rate": 0.0005889669007021063, + "loss": 1.4135, + "theoretical_loss": 3.716830947797932, + "tokens_seen": 894894080 + }, + { + "epoch": 0.27, + "learning_rate": 0.000588950852557673, + "loss": 1.4315, + "theoretical_loss": 3.7168047131876265, + "tokens_seen": 894959616 + }, + { + "epoch": 0.27, + "learning_rate": 0.0005889348044132397, + "loss": 1.4475, + "theoretical_loss": 3.7167784810362203, + "tokens_seen": 895025152 + }, + { + "epoch": 0.27, + "learning_rate": 0.0005889187562688064, + "loss": 1.5247, + "theoretical_loss": 3.7167522513433022, + "tokens_seen": 895090688 + }, + { + "epoch": 0.27, + "learning_rate": 0.0005889027081243731, + "loss": 1.4697, + "theoretical_loss": 3.716726024108463, + "tokens_seen": 895156224 + }, + { + "epoch": 0.27, + "learning_rate": 0.0005888866599799398, + "loss": 1.4785, + "theoretical_loss": 3.716699799331292, + "tokens_seen": 895221760 + }, + { + "epoch": 0.27, + "learning_rate": 0.0005888706118355065, + "loss": 1.4923, + "theoretical_loss": 3.716673577011379, + "tokens_seen": 895287296 + }, + { + "epoch": 0.27, + "learning_rate": 0.0005888545636910732, + "loss": 1.5336, + "theoretical_loss": 3.716647357148313, + "tokens_seen": 895352832 + }, + { + "epoch": 0.27, + "learning_rate": 0.0005888385155466399, + "loss": 1.5536, + "theoretical_loss": 3.7166211397416853, + "tokens_seen": 895418368 + }, + { + "epoch": 0.27, + "learning_rate": 0.0005888224674022066, + "loss": 1.4379, + "theoretical_loss": 3.716594924791085, + "tokens_seen": 895483904 + }, + { + "epoch": 0.27, + "learning_rate": 0.0005888064192577733, + "loss": 1.5276, + "theoretical_loss": 3.7165687122961035, + "tokens_seen": 895549440 + }, + { + "epoch": 0.27, + "learning_rate": 0.00058879037111334, + "loss": 1.3394, + "theoretical_loss": 3.71654250225633, + "tokens_seen": 895614976 + }, + { + "epoch": 0.27, + "learning_rate": 0.0005887743229689067, + "loss": 1.3696, + "theoretical_loss": 3.7165162946713552, + "tokens_seen": 895680512 + }, + { + "epoch": 0.27, + "learning_rate": 0.0005887582748244734, + "loss": 1.4354, + "theoretical_loss": 3.7164900895407698, + "tokens_seen": 895746048 + }, + { + "epoch": 0.27, + "learning_rate": 0.0005887422266800401, + "loss": 1.5469, + "theoretical_loss": 3.7164638868641635, + "tokens_seen": 895811584 + }, + { + "epoch": 0.27, + "learning_rate": 0.0005887261785356068, + "loss": 1.5306, + "theoretical_loss": 3.716437686641129, + "tokens_seen": 895877120 + }, + { + "epoch": 0.27, + "learning_rate": 0.0005887101303911735, + "loss": 1.4714, + "theoretical_loss": 3.7164114888712545, + "tokens_seen": 895942656 + }, + { + "epoch": 0.27, + "learning_rate": 0.0005886940822467402, + "loss": 1.4254, + "theoretical_loss": 3.7163852935541333, + "tokens_seen": 896008192 + }, + { + "epoch": 0.27, + "learning_rate": 0.000588678034102307, + "loss": 1.4862, + "theoretical_loss": 3.716359100689355, + "tokens_seen": 896073728 + }, + { + "epoch": 0.27, + "learning_rate": 0.0005886619859578736, + "loss": 1.4401, + "theoretical_loss": 3.7163329102765106, + "tokens_seen": 896139264 + }, + { + "epoch": 0.27, + "learning_rate": 0.0005886459378134403, + "loss": 1.4123, + "theoretical_loss": 3.7163067223151924, + "tokens_seen": 896204800 + }, + { + "epoch": 0.27, + "learning_rate": 0.000588629889669007, + "loss": 1.3633, + "theoretical_loss": 3.7162805368049905, + "tokens_seen": 896270336 + }, + { + "epoch": 0.27, + "learning_rate": 0.0005886138415245737, + "loss": 1.5092, + "theoretical_loss": 3.716254353745497, + "tokens_seen": 896335872 + }, + { + "epoch": 0.27, + "learning_rate": 0.0005885977933801404, + "loss": 1.4656, + "theoretical_loss": 3.7162281731363027, + "tokens_seen": 896401408 + }, + { + "epoch": 0.27, + "learning_rate": 0.0005885817452357072, + "loss": 1.421, + "theoretical_loss": 3.7162019949769998, + "tokens_seen": 896466944 + }, + { + "epoch": 0.27, + "learning_rate": 0.0005885656970912739, + "loss": 1.489, + "theoretical_loss": 3.71617581926718, + "tokens_seen": 896532480 + }, + { + "epoch": 0.27, + "learning_rate": 0.0005885496489468406, + "loss": 1.3765, + "theoretical_loss": 3.7161496460064347, + "tokens_seen": 896598016 + }, + { + "epoch": 0.27, + "learning_rate": 0.0005885336008024073, + "loss": 1.4959, + "theoretical_loss": 3.716123475194356, + "tokens_seen": 896663552 + }, + { + "epoch": 0.27, + "learning_rate": 0.000588517552657974, + "loss": 1.359, + "theoretical_loss": 3.7160973068305356, + "tokens_seen": 896729088 + }, + { + "epoch": 0.27, + "learning_rate": 0.0005885015045135407, + "loss": 1.4736, + "theoretical_loss": 3.7160711409145657, + "tokens_seen": 896794624 + }, + { + "epoch": 0.27, + "learning_rate": 0.0005884854563691074, + "loss": 1.4579, + "theoretical_loss": 3.716044977446039, + "tokens_seen": 896860160 + }, + { + "epoch": 0.27, + "learning_rate": 0.0005884694082246741, + "loss": 1.4473, + "theoretical_loss": 3.7160188164245467, + "tokens_seen": 896925696 + }, + { + "epoch": 0.27, + "learning_rate": 0.0005884533600802408, + "loss": 1.4428, + "theoretical_loss": 3.7159926578496822, + "tokens_seen": 896991232 + }, + { + "epoch": 0.27, + "learning_rate": 0.0005884373119358075, + "loss": 1.3778, + "theoretical_loss": 3.715966501721037, + "tokens_seen": 897056768 + }, + { + "epoch": 0.27, + "learning_rate": 0.0005884212637913742, + "loss": 1.4859, + "theoretical_loss": 3.7159403480382043, + "tokens_seen": 897122304 + }, + { + "epoch": 0.27, + "learning_rate": 0.0005884052156469409, + "loss": 1.6178, + "theoretical_loss": 3.715914196800777, + "tokens_seen": 897187840 + }, + { + "epoch": 0.27, + "learning_rate": 0.0005883891675025076, + "loss": 1.4279, + "theoretical_loss": 3.715888048008347, + "tokens_seen": 897253376 + }, + { + "epoch": 0.27, + "learning_rate": 0.0005883731193580743, + "loss": 1.4788, + "theoretical_loss": 3.7158619016605074, + "tokens_seen": 897318912 + }, + { + "epoch": 0.27, + "learning_rate": 0.000588357071213641, + "loss": 1.5224, + "theoretical_loss": 3.715835757756852, + "tokens_seen": 897384448 + }, + { + "epoch": 0.27, + "learning_rate": 0.0005883410230692077, + "loss": 1.3968, + "theoretical_loss": 3.7158096162969727, + "tokens_seen": 897449984 + }, + { + "epoch": 0.27, + "learning_rate": 0.0005883249749247744, + "loss": 1.3922, + "theoretical_loss": 3.715783477280463, + "tokens_seen": 897515520 + }, + { + "epoch": 0.27, + "learning_rate": 0.0005883089267803411, + "loss": 1.4595, + "theoretical_loss": 3.715757340706917, + "tokens_seen": 897581056 + }, + { + "epoch": 0.27, + "learning_rate": 0.0005882928786359078, + "loss": 1.3244, + "theoretical_loss": 3.715731206575927, + "tokens_seen": 897646592 + }, + { + "epoch": 0.27, + "learning_rate": 0.0005882768304914745, + "loss": 1.3475, + "theoretical_loss": 3.7157050748870866, + "tokens_seen": 897712128 + }, + { + "epoch": 0.27, + "learning_rate": 0.0005882607823470412, + "loss": 1.3841, + "theoretical_loss": 3.715678945639989, + "tokens_seen": 897777664 + }, + { + "epoch": 0.27, + "objective/train/docs_used": 331577, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.4565703868865967, + "objective/train/theoretical_loss": 3.7156528188342284, + "objective/train/tokens_used": 918303200, + "theoretical_loss": 3.7156528188342284, + "tokens_seen": 897843200 + }, + { + "epoch": 0.27, + "learning_rate": 0.0005882447342026079, + "loss": 1.5005, + "theoretical_loss": 3.7156528188342284, + "tokens_seen": 897843200 + }, + { + "epoch": 0.27, + "learning_rate": 0.0005882286860581746, + "loss": 1.4618, + "theoretical_loss": 3.7156266944693987, + "tokens_seen": 897908736 + }, + { + "epoch": 0.27, + "learning_rate": 0.0005882126379137413, + "loss": 1.5891, + "theoretical_loss": 3.7156005725450933, + "tokens_seen": 897974272 + }, + { + "epoch": 0.27, + "learning_rate": 0.000588196589769308, + "loss": 1.4048, + "theoretical_loss": 3.7155744530609063, + "tokens_seen": 898039808 + }, + { + "epoch": 0.27, + "learning_rate": 0.0005881805416248747, + "loss": 1.4901, + "theoretical_loss": 3.7155483360164316, + "tokens_seen": 898105344 + }, + { + "epoch": 0.27, + "learning_rate": 0.0005881644934804414, + "loss": 1.3609, + "theoretical_loss": 3.7155222214112635, + "tokens_seen": 898170880 + }, + { + "epoch": 0.27, + "learning_rate": 0.0005881484453360081, + "loss": 1.4204, + "theoretical_loss": 3.715496109244996, + "tokens_seen": 898236416 + }, + { + "epoch": 0.27, + "learning_rate": 0.0005881323971915748, + "loss": 1.5766, + "theoretical_loss": 3.715469999517224, + "tokens_seen": 898301952 + }, + { + "epoch": 0.27, + "learning_rate": 0.0005881163490471415, + "loss": 1.4731, + "theoretical_loss": 3.7154438922275403, + "tokens_seen": 898367488 + }, + { + "epoch": 0.27, + "learning_rate": 0.0005881003009027082, + "loss": 1.4756, + "theoretical_loss": 3.7154177873755416, + "tokens_seen": 898433024 + }, + { + "epoch": 0.27, + "learning_rate": 0.0005880842527582749, + "loss": 1.4361, + "theoretical_loss": 3.715391684960821, + "tokens_seen": 898498560 + }, + { + "epoch": 0.27, + "learning_rate": 0.0005880682046138416, + "loss": 1.4595, + "theoretical_loss": 3.715365584982973, + "tokens_seen": 898564096 + }, + { + "epoch": 0.27, + "learning_rate": 0.0005880521564694083, + "loss": 1.473, + "theoretical_loss": 3.7153394874415935, + "tokens_seen": 898629632 + }, + { + "epoch": 0.27, + "learning_rate": 0.000588036108324975, + "loss": 1.4233, + "theoretical_loss": 3.7153133923362764, + "tokens_seen": 898695168 + }, + { + "epoch": 0.27, + "learning_rate": 0.0005880200601805417, + "loss": 1.4221, + "theoretical_loss": 3.7152872996666177, + "tokens_seen": 898760704 + }, + { + "epoch": 0.27, + "learning_rate": 0.0005880040120361084, + "loss": 1.4991, + "theoretical_loss": 3.715261209432212, + "tokens_seen": 898826240 + }, + { + "epoch": 0.27, + "learning_rate": 0.0005879879638916751, + "loss": 1.4951, + "theoretical_loss": 3.715235121632654, + "tokens_seen": 898891776 + }, + { + "epoch": 0.27, + "learning_rate": 0.0005879719157472418, + "loss": 1.449, + "theoretical_loss": 3.7152090362675394, + "tokens_seen": 898957312 + }, + { + "epoch": 0.27, + "learning_rate": 0.0005879558676028084, + "loss": 1.4402, + "theoretical_loss": 3.7151829533364635, + "tokens_seen": 899022848 + }, + { + "epoch": 0.27, + "learning_rate": 0.0005879398194583752, + "loss": 1.5493, + "theoretical_loss": 3.715156872839022, + "tokens_seen": 899088384 + }, + { + "epoch": 0.27, + "learning_rate": 0.0005879237713139418, + "loss": 1.4108, + "theoretical_loss": 3.7151307947748107, + "tokens_seen": 899153920 + }, + { + "epoch": 0.27, + "learning_rate": 0.0005879077231695086, + "loss": 1.4745, + "theoretical_loss": 3.715104719143424, + "tokens_seen": 899219456 + }, + { + "epoch": 0.27, + "learning_rate": 0.0005878916750250752, + "loss": 1.413, + "theoretical_loss": 3.715078645944459, + "tokens_seen": 899284992 + }, + { + "epoch": 0.27, + "learning_rate": 0.000587875626880642, + "loss": 1.4361, + "theoretical_loss": 3.715052575177511, + "tokens_seen": 899350528 + }, + { + "epoch": 0.27, + "learning_rate": 0.0005878595787362086, + "loss": 1.4347, + "theoretical_loss": 3.715026506842176, + "tokens_seen": 899416064 + }, + { + "epoch": 0.27, + "learning_rate": 0.0005878435305917754, + "loss": 1.3712, + "theoretical_loss": 3.7150004409380504, + "tokens_seen": 899481600 + }, + { + "epoch": 0.27, + "learning_rate": 0.000587827482447342, + "loss": 1.5579, + "theoretical_loss": 3.7149743774647295, + "tokens_seen": 899547136 + }, + { + "epoch": 0.27, + "learning_rate": 0.0005878114343029087, + "loss": 1.4717, + "theoretical_loss": 3.7149483164218102, + "tokens_seen": 899612672 + }, + { + "epoch": 0.27, + "learning_rate": 0.0005877953861584754, + "loss": 1.3964, + "theoretical_loss": 3.714922257808889, + "tokens_seen": 899678208 + }, + { + "epoch": 0.27, + "learning_rate": 0.0005877793380140421, + "loss": 1.4205, + "theoretical_loss": 3.7148962016255616, + "tokens_seen": 899743744 + }, + { + "epoch": 0.27, + "learning_rate": 0.0005877632898696088, + "loss": 1.5342, + "theoretical_loss": 3.714870147871425, + "tokens_seen": 899809280 + }, + { + "epoch": 0.27, + "learning_rate": 0.0005877472417251755, + "loss": 1.4187, + "theoretical_loss": 3.7148440965460763, + "tokens_seen": 899874816 + }, + { + "epoch": 0.27, + "learning_rate": 0.0005877311935807422, + "loss": 1.3726, + "theoretical_loss": 3.7148180476491115, + "tokens_seen": 899940352 + }, + { + "epoch": 0.27, + "learning_rate": 0.0005877151454363089, + "loss": 1.4916, + "theoretical_loss": 3.714792001180127, + "tokens_seen": 900005888 + }, + { + "epoch": 0.27, + "learning_rate": 0.0005876990972918756, + "loss": 1.4589, + "theoretical_loss": 3.714765957138721, + "tokens_seen": 900071424 + }, + { + "epoch": 0.27, + "learning_rate": 0.0005876830491474423, + "loss": 1.4237, + "theoretical_loss": 3.71473991552449, + "tokens_seen": 900136960 + }, + { + "epoch": 0.27, + "learning_rate": 0.000587667001003009, + "loss": 1.3569, + "theoretical_loss": 3.71471387633703, + "tokens_seen": 900202496 + }, + { + "epoch": 0.27, + "learning_rate": 0.0005876509528585757, + "loss": 1.5222, + "theoretical_loss": 3.7146878395759404, + "tokens_seen": 900268032 + }, + { + "epoch": 0.27, + "learning_rate": 0.0005876349047141424, + "loss": 1.5806, + "theoretical_loss": 3.7146618052408167, + "tokens_seen": 900333568 + }, + { + "epoch": 0.27, + "learning_rate": 0.0005876188565697091, + "loss": 1.4763, + "theoretical_loss": 3.714635773331257, + "tokens_seen": 900399104 + }, + { + "epoch": 0.27, + "learning_rate": 0.0005876028084252758, + "loss": 1.5143, + "theoretical_loss": 3.7146097438468586, + "tokens_seen": 900464640 + }, + { + "epoch": 0.27, + "learning_rate": 0.0005875867602808425, + "loss": 1.4836, + "theoretical_loss": 3.714583716787219, + "tokens_seen": 900530176 + }, + { + "epoch": 0.27, + "learning_rate": 0.0005875707121364092, + "loss": 1.3968, + "theoretical_loss": 3.7145576921519363, + "tokens_seen": 900595712 + }, + { + "epoch": 0.27, + "learning_rate": 0.0005875546639919759, + "loss": 1.443, + "theoretical_loss": 3.714531669940608, + "tokens_seen": 900661248 + }, + { + "epoch": 0.27, + "learning_rate": 0.0005875386158475426, + "loss": 1.4936, + "theoretical_loss": 3.714505650152832, + "tokens_seen": 900726784 + }, + { + "epoch": 0.27, + "learning_rate": 0.0005875225677031093, + "loss": 1.4123, + "theoretical_loss": 3.7144796327882066, + "tokens_seen": 900792320 + }, + { + "epoch": 0.27, + "learning_rate": 0.000587506519558676, + "loss": 1.4276, + "theoretical_loss": 3.714453617846329, + "tokens_seen": 900857856 + }, + { + "epoch": 0.27, + "learning_rate": 0.0005874904714142427, + "loss": 1.4589, + "theoretical_loss": 3.714427605326798, + "tokens_seen": 900923392 + }, + { + "epoch": 0.27, + "learning_rate": 0.0005874744232698094, + "loss": 1.556, + "theoretical_loss": 3.714401595229212, + "tokens_seen": 900988928 + }, + { + "epoch": 0.27, + "learning_rate": 0.0005874583751253761, + "loss": 1.4262, + "theoretical_loss": 3.714375587553169, + "tokens_seen": 901054464 + }, + { + "epoch": 0.27, + "objective/train/docs_used": 332819, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.5790141820907593, + "objective/train/theoretical_loss": 3.714349582298268, + "objective/train/tokens_used": 921580000, + "theoretical_loss": 3.714349582298268, + "tokens_seen": 901120000 + }, + { + "epoch": 0.27, + "learning_rate": 0.0005874423269809428, + "loss": 1.5193, + "theoretical_loss": 3.714349582298268, + "tokens_seen": 901120000 + }, + { + "epoch": 0.27, + "learning_rate": 0.0005874262788365095, + "loss": 1.3811, + "theoretical_loss": 3.7143235794641067, + "tokens_seen": 901185536 + }, + { + "epoch": 0.27, + "learning_rate": 0.0005874102306920762, + "loss": 1.4677, + "theoretical_loss": 3.7142975790502843, + "tokens_seen": 901251072 + }, + { + "epoch": 0.27, + "learning_rate": 0.0005873941825476429, + "loss": 1.5225, + "theoretical_loss": 3.7142715810563995, + "tokens_seen": 901316608 + }, + { + "epoch": 0.27, + "learning_rate": 0.0005873781344032096, + "loss": 1.4944, + "theoretical_loss": 3.7142455854820513, + "tokens_seen": 901382144 + }, + { + "epoch": 0.27, + "learning_rate": 0.0005873620862587764, + "loss": 1.4589, + "theoretical_loss": 3.714219592326838, + "tokens_seen": 901447680 + }, + { + "epoch": 0.27, + "learning_rate": 0.000587346038114343, + "loss": 1.4725, + "theoretical_loss": 3.714193601590359, + "tokens_seen": 901513216 + }, + { + "epoch": 0.27, + "learning_rate": 0.0005873299899699098, + "loss": 1.5259, + "theoretical_loss": 3.7141676132722137, + "tokens_seen": 901578752 + }, + { + "epoch": 0.27, + "learning_rate": 0.0005873139418254764, + "loss": 1.4606, + "theoretical_loss": 3.714141627372001, + "tokens_seen": 901644288 + }, + { + "epoch": 0.27, + "learning_rate": 0.0005872978936810432, + "loss": 1.456, + "theoretical_loss": 3.7141156438893206, + "tokens_seen": 901709824 + }, + { + "epoch": 0.27, + "learning_rate": 0.0005872818455366098, + "loss": 1.5024, + "theoretical_loss": 3.7140896628237714, + "tokens_seen": 901775360 + }, + { + "epoch": 0.27, + "learning_rate": 0.0005872657973921766, + "loss": 1.4145, + "theoretical_loss": 3.7140636841749526, + "tokens_seen": 901840896 + }, + { + "epoch": 0.27, + "learning_rate": 0.0005872497492477432, + "loss": 1.3988, + "theoretical_loss": 3.714037707942465, + "tokens_seen": 901906432 + }, + { + "epoch": 0.27, + "learning_rate": 0.00058723370110331, + "loss": 1.5493, + "theoretical_loss": 3.714011734125907, + "tokens_seen": 901971968 + }, + { + "epoch": 0.27, + "learning_rate": 0.0005872176529588767, + "loss": 1.4868, + "theoretical_loss": 3.7139857627248793, + "tokens_seen": 902037504 + }, + { + "epoch": 0.27, + "learning_rate": 0.0005872016048144434, + "loss": 1.5108, + "theoretical_loss": 3.7139597937389817, + "tokens_seen": 902103040 + }, + { + "epoch": 0.27, + "learning_rate": 0.00058718555667001, + "loss": 1.5235, + "theoretical_loss": 3.7139338271678133, + "tokens_seen": 902168576 + }, + { + "epoch": 0.27, + "learning_rate": 0.0005871695085255768, + "loss": 1.4353, + "theoretical_loss": 3.7139078630109754, + "tokens_seen": 902234112 + }, + { + "epoch": 0.27, + "learning_rate": 0.0005871534603811435, + "loss": 1.3629, + "theoretical_loss": 3.713881901268068, + "tokens_seen": 902299648 + }, + { + "epoch": 0.27, + "learning_rate": 0.0005871374122367102, + "loss": 1.4586, + "theoretical_loss": 3.7138559419386903, + "tokens_seen": 902365184 + }, + { + "epoch": 0.27, + "learning_rate": 0.0005871213640922769, + "loss": 1.4235, + "theoretical_loss": 3.713829985022443, + "tokens_seen": 902430720 + }, + { + "epoch": 0.27, + "learning_rate": 0.0005871053159478436, + "loss": 1.4072, + "theoretical_loss": 3.7138040305189275, + "tokens_seen": 902496256 + }, + { + "epoch": 0.27, + "learning_rate": 0.0005870892678034103, + "loss": 1.5726, + "theoretical_loss": 3.713778078427744, + "tokens_seen": 902561792 + }, + { + "epoch": 0.27, + "learning_rate": 0.000587073219658977, + "loss": 1.4525, + "theoretical_loss": 3.7137521287484923, + "tokens_seen": 902627328 + }, + { + "epoch": 0.27, + "learning_rate": 0.0005870571715145437, + "loss": 1.4329, + "theoretical_loss": 3.713726181480774, + "tokens_seen": 902692864 + }, + { + "epoch": 0.27, + "learning_rate": 0.0005870411233701104, + "loss": 1.4304, + "theoretical_loss": 3.71370023662419, + "tokens_seen": 902758400 + }, + { + "epoch": 0.27, + "learning_rate": 0.0005870250752256771, + "loss": 1.5307, + "theoretical_loss": 3.7136742941783405, + "tokens_seen": 902823936 + }, + { + "epoch": 0.27, + "learning_rate": 0.0005870090270812438, + "loss": 1.4571, + "theoretical_loss": 3.713648354142827, + "tokens_seen": 902889472 + }, + { + "epoch": 0.27, + "learning_rate": 0.0005869929789368105, + "loss": 1.5763, + "theoretical_loss": 3.713622416517251, + "tokens_seen": 902955008 + }, + { + "epoch": 0.27, + "learning_rate": 0.0005869769307923772, + "loss": 1.426, + "theoretical_loss": 3.7135964813012126, + "tokens_seen": 903020544 + }, + { + "epoch": 0.27, + "learning_rate": 0.0005869608826479439, + "loss": 1.4038, + "theoretical_loss": 3.7135705484943142, + "tokens_seen": 903086080 + }, + { + "epoch": 0.27, + "learning_rate": 0.0005869448345035106, + "loss": 1.5695, + "theoretical_loss": 3.713544618096157, + "tokens_seen": 903151616 + }, + { + "epoch": 0.27, + "learning_rate": 0.0005869287863590773, + "loss": 1.3057, + "theoretical_loss": 3.713518690106342, + "tokens_seen": 903217152 + }, + { + "epoch": 0.27, + "learning_rate": 0.000586912738214644, + "loss": 1.4786, + "theoretical_loss": 3.7134927645244713, + "tokens_seen": 903282688 + }, + { + "epoch": 0.27, + "learning_rate": 0.0005868966900702107, + "loss": 1.387, + "theoretical_loss": 3.7134668413501464, + "tokens_seen": 903348224 + }, + { + "epoch": 0.27, + "learning_rate": 0.0005868806419257774, + "loss": 1.4518, + "theoretical_loss": 3.713440920582969, + "tokens_seen": 903413760 + }, + { + "epoch": 0.27, + "learning_rate": 0.0005868645937813441, + "loss": 1.3971, + "theoretical_loss": 3.713415002222541, + "tokens_seen": 903479296 + }, + { + "epoch": 0.27, + "learning_rate": 0.0005868485456369108, + "loss": 1.3517, + "theoretical_loss": 3.7133890862684646, + "tokens_seen": 903544832 + }, + { + "epoch": 0.27, + "learning_rate": 0.0005868324974924775, + "loss": 1.4421, + "theoretical_loss": 3.7133631727203418, + "tokens_seen": 903610368 + }, + { + "epoch": 0.27, + "learning_rate": 0.0005868164493480442, + "loss": 1.6011, + "theoretical_loss": 3.7133372615777747, + "tokens_seen": 903675904 + }, + { + "epoch": 0.27, + "learning_rate": 0.0005868004012036109, + "loss": 1.5559, + "theoretical_loss": 3.713311352840366, + "tokens_seen": 903741440 + }, + { + "epoch": 0.27, + "learning_rate": 0.0005867843530591776, + "loss": 1.3704, + "theoretical_loss": 3.713285446507717, + "tokens_seen": 903806976 + }, + { + "epoch": 0.27, + "learning_rate": 0.0005867683049147443, + "loss": 1.4972, + "theoretical_loss": 3.7132595425794306, + "tokens_seen": 903872512 + }, + { + "epoch": 0.27, + "learning_rate": 0.000586752256770311, + "loss": 1.502, + "theoretical_loss": 3.7132336410551097, + "tokens_seen": 903938048 + }, + { + "epoch": 0.27, + "learning_rate": 0.0005867362086258777, + "loss": 1.5336, + "theoretical_loss": 3.713207741934357, + "tokens_seen": 904003584 + }, + { + "epoch": 0.27, + "learning_rate": 0.0005867201604814444, + "loss": 1.4463, + "theoretical_loss": 3.713181845216775, + "tokens_seen": 904069120 + }, + { + "epoch": 0.27, + "learning_rate": 0.0005867041123370111, + "loss": 1.4061, + "theoretical_loss": 3.7131559509019665, + "tokens_seen": 904134656 + }, + { + "epoch": 0.27, + "learning_rate": 0.0005866880641925778, + "loss": 1.3751, + "theoretical_loss": 3.713130058989534, + "tokens_seen": 904200192 + }, + { + "epoch": 0.27, + "learning_rate": 0.0005866720160481445, + "loss": 1.357, + "theoretical_loss": 3.713104169479082, + "tokens_seen": 904265728 + }, + { + "epoch": 0.27, + "learning_rate": 0.0005866559679037112, + "loss": 1.4965, + "theoretical_loss": 3.713078282370211, + "tokens_seen": 904331264 + }, + { + "epoch": 0.27, + "objective/train/docs_used": 333943, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.3974798917770386, + "objective/train/theoretical_loss": 3.713052397662527, + "objective/train/tokens_used": 924856800, + "theoretical_loss": 3.713052397662527, + "tokens_seen": 904396800 + }, + { + "epoch": 0.27, + "learning_rate": 0.0005866399197592779, + "loss": 1.4074, + "theoretical_loss": 3.713052397662527, + "tokens_seen": 904396800 + }, + { + "epoch": 0.27, + "learning_rate": 0.0005866238716148446, + "loss": 1.4516, + "theoretical_loss": 3.713026515355632, + "tokens_seen": 904462336 + }, + { + "epoch": 0.27, + "learning_rate": 0.0005866078234704113, + "loss": 1.4048, + "theoretical_loss": 3.713000635449129, + "tokens_seen": 904527872 + }, + { + "epoch": 0.27, + "learning_rate": 0.000586591775325978, + "loss": 1.3608, + "theoretical_loss": 3.7129747579426224, + "tokens_seen": 904593408 + }, + { + "epoch": 0.27, + "learning_rate": 0.0005865757271815447, + "loss": 1.4363, + "theoretical_loss": 3.712948882835715, + "tokens_seen": 904658944 + }, + { + "epoch": 0.27, + "learning_rate": 0.0005865596790371113, + "loss": 1.4957, + "theoretical_loss": 3.7129230101280113, + "tokens_seen": 904724480 + }, + { + "epoch": 0.27, + "learning_rate": 0.0005865436308926781, + "loss": 1.4183, + "theoretical_loss": 3.7128971398191144, + "tokens_seen": 904790016 + }, + { + "epoch": 0.27, + "learning_rate": 0.0005865275827482447, + "loss": 1.599, + "theoretical_loss": 3.712871271908628, + "tokens_seen": 904855552 + }, + { + "epoch": 0.27, + "learning_rate": 0.0005865115346038115, + "loss": 1.4496, + "theoretical_loss": 3.7128454063961565, + "tokens_seen": 904921088 + }, + { + "epoch": 0.27, + "learning_rate": 0.0005864954864593781, + "loss": 1.4644, + "theoretical_loss": 3.712819543281304, + "tokens_seen": 904986624 + }, + { + "epoch": 0.27, + "learning_rate": 0.0005864794383149449, + "loss": 1.4706, + "theoretical_loss": 3.712793682563675, + "tokens_seen": 905052160 + }, + { + "epoch": 0.27, + "learning_rate": 0.0005864633901705115, + "loss": 1.4495, + "theoretical_loss": 3.712767824242873, + "tokens_seen": 905117696 + }, + { + "epoch": 0.27, + "learning_rate": 0.0005864473420260783, + "loss": 1.4008, + "theoretical_loss": 3.712741968318502, + "tokens_seen": 905183232 + }, + { + "epoch": 0.27, + "learning_rate": 0.0005864312938816449, + "loss": 1.4794, + "theoretical_loss": 3.712716114790168, + "tokens_seen": 905248768 + }, + { + "epoch": 0.27, + "learning_rate": 0.0005864152457372116, + "loss": 1.4236, + "theoretical_loss": 3.712690263657474, + "tokens_seen": 905314304 + }, + { + "epoch": 0.27, + "learning_rate": 0.0005863991975927783, + "loss": 1.4265, + "theoretical_loss": 3.7126644149200247, + "tokens_seen": 905379840 + }, + { + "epoch": 0.27, + "learning_rate": 0.000586383149448345, + "loss": 1.5952, + "theoretical_loss": 3.712638568577426, + "tokens_seen": 905445376 + }, + { + "epoch": 0.27, + "learning_rate": 0.0005863671013039117, + "loss": 1.4216, + "theoretical_loss": 3.712612724629282, + "tokens_seen": 905510912 + }, + { + "epoch": 0.27, + "learning_rate": 0.0005863510531594784, + "loss": 1.3784, + "theoretical_loss": 3.712586883075198, + "tokens_seen": 905576448 + }, + { + "epoch": 0.27, + "learning_rate": 0.0005863350050150451, + "loss": 1.3675, + "theoretical_loss": 3.7125610439147776, + "tokens_seen": 905641984 + }, + { + "epoch": 0.27, + "learning_rate": 0.0005863189568706118, + "loss": 1.4853, + "theoretical_loss": 3.7125352071476274, + "tokens_seen": 905707520 + }, + { + "epoch": 0.27, + "learning_rate": 0.0005863029087261785, + "loss": 1.3523, + "theoretical_loss": 3.7125093727733516, + "tokens_seen": 905773056 + }, + { + "epoch": 0.27, + "learning_rate": 0.0005862868605817452, + "loss": 1.3309, + "theoretical_loss": 3.7124835407915566, + "tokens_seen": 905838592 + }, + { + "epoch": 0.27, + "learning_rate": 0.0005862708124373119, + "loss": 1.5174, + "theoretical_loss": 3.712457711201847, + "tokens_seen": 905904128 + }, + { + "epoch": 0.27, + "learning_rate": 0.0005862547642928786, + "loss": 1.5498, + "theoretical_loss": 3.712431884003828, + "tokens_seen": 905969664 + }, + { + "epoch": 0.27, + "learning_rate": 0.0005862387161484453, + "loss": 1.4217, + "theoretical_loss": 3.7124060591971055, + "tokens_seen": 906035200 + }, + { + "epoch": 0.27, + "learning_rate": 0.000586222668004012, + "loss": 1.4813, + "theoretical_loss": 3.7123802367812857, + "tokens_seen": 906100736 + }, + { + "epoch": 0.27, + "learning_rate": 0.0005862066198595787, + "loss": 1.3553, + "theoretical_loss": 3.712354416755973, + "tokens_seen": 906166272 + }, + { + "epoch": 0.27, + "learning_rate": 0.0005861905717151454, + "loss": 1.5467, + "theoretical_loss": 3.7123285991207746, + "tokens_seen": 906231808 + }, + { + "epoch": 0.27, + "learning_rate": 0.0005861745235707121, + "loss": 1.4615, + "theoretical_loss": 3.712302783875296, + "tokens_seen": 906297344 + }, + { + "epoch": 0.27, + "learning_rate": 0.0005861584754262789, + "loss": 1.4112, + "theoretical_loss": 3.7122769710191426, + "tokens_seen": 906362880 + }, + { + "epoch": 0.27, + "learning_rate": 0.0005861424272818455, + "loss": 1.5127, + "theoretical_loss": 3.7122511605519217, + "tokens_seen": 906428416 + }, + { + "epoch": 0.27, + "learning_rate": 0.0005861263791374123, + "loss": 1.2917, + "theoretical_loss": 3.7122253524732383, + "tokens_seen": 906493952 + }, + { + "epoch": 0.27, + "learning_rate": 0.0005861103309929789, + "loss": 1.364, + "theoretical_loss": 3.712199546782699, + "tokens_seen": 906559488 + }, + { + "epoch": 0.27, + "learning_rate": 0.0005860942828485457, + "loss": 1.4242, + "theoretical_loss": 3.7121737434799114, + "tokens_seen": 906625024 + }, + { + "epoch": 0.27, + "learning_rate": 0.0005860782347041123, + "loss": 1.4463, + "theoretical_loss": 3.71214794256448, + "tokens_seen": 906690560 + }, + { + "epoch": 0.27, + "learning_rate": 0.0005860621865596791, + "loss": 1.4286, + "theoretical_loss": 3.712122144036013, + "tokens_seen": 906756096 + }, + { + "epoch": 0.27, + "learning_rate": 0.0005860461384152457, + "loss": 1.5289, + "theoretical_loss": 3.712096347894116, + "tokens_seen": 906821632 + }, + { + "epoch": 0.27, + "learning_rate": 0.0005860300902708125, + "loss": 1.3849, + "theoretical_loss": 3.7120705541383967, + "tokens_seen": 906887168 + }, + { + "epoch": 0.27, + "learning_rate": 0.0005860140421263791, + "loss": 1.4133, + "theoretical_loss": 3.7120447627684614, + "tokens_seen": 906952704 + }, + { + "epoch": 0.27, + "learning_rate": 0.0005859979939819459, + "loss": 1.5198, + "theoretical_loss": 3.712018973783917, + "tokens_seen": 907018240 + }, + { + "epoch": 0.27, + "learning_rate": 0.0005859819458375125, + "loss": 1.4252, + "theoretical_loss": 3.7119931871843708, + "tokens_seen": 907083776 + }, + { + "epoch": 0.27, + "learning_rate": 0.0005859658976930793, + "loss": 1.3417, + "theoretical_loss": 3.71196740296943, + "tokens_seen": 907149312 + }, + { + "epoch": 0.27, + "learning_rate": 0.000585949849548646, + "loss": 1.499, + "theoretical_loss": 3.7119416211387013, + "tokens_seen": 907214848 + }, + { + "epoch": 0.27, + "learning_rate": 0.0005859338014042127, + "loss": 1.4658, + "theoretical_loss": 3.7119158416917926, + "tokens_seen": 907280384 + }, + { + "epoch": 0.27, + "learning_rate": 0.0005859177532597793, + "loss": 1.3864, + "theoretical_loss": 3.711890064628311, + "tokens_seen": 907345920 + }, + { + "epoch": 0.27, + "learning_rate": 0.0005859017051153461, + "loss": 1.422, + "theoretical_loss": 3.711864289947864, + "tokens_seen": 907411456 + }, + { + "epoch": 0.27, + "learning_rate": 0.0005858856569709127, + "loss": 1.4246, + "theoretical_loss": 3.71183851765006, + "tokens_seen": 907476992 + }, + { + "epoch": 0.28, + "learning_rate": 0.0005858696088264795, + "loss": 1.4442, + "theoretical_loss": 3.7118127477345055, + "tokens_seen": 907542528 + }, + { + "epoch": 0.28, + "learning_rate": 0.0005858535606820462, + "loss": 1.3901, + "theoretical_loss": 3.7117869802008094, + "tokens_seen": 907608064 + }, + { + "epoch": 0.28, + "objective/train/docs_used": 335091, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.3617725372314453, + "objective/train/theoretical_loss": 3.7117612150485786, + "objective/train/tokens_used": 928133600, + "theoretical_loss": 3.7117612150485786, + "tokens_seen": 907673600 + }, + { + "epoch": 0.28, + "learning_rate": 0.0005858375125376129, + "loss": 1.3876, + "theoretical_loss": 3.7117612150485786, + "tokens_seen": 907673600 + }, + { + "epoch": 0.28, + "learning_rate": 0.0005858214643931796, + "loss": 1.4454, + "theoretical_loss": 3.7117354522774217, + "tokens_seen": 907739136 + }, + { + "epoch": 0.28, + "learning_rate": 0.0005858054162487463, + "loss": 1.4292, + "theoretical_loss": 3.7117096918869468, + "tokens_seen": 907804672 + }, + { + "epoch": 0.28, + "learning_rate": 0.000585789368104313, + "loss": 1.5043, + "theoretical_loss": 3.7116839338767615, + "tokens_seen": 907870208 + }, + { + "epoch": 0.28, + "learning_rate": 0.0005857733199598797, + "loss": 1.3853, + "theoretical_loss": 3.7116581782464744, + "tokens_seen": 907935744 + }, + { + "epoch": 0.28, + "learning_rate": 0.0005857572718154464, + "loss": 1.52, + "theoretical_loss": 3.711632424995694, + "tokens_seen": 908001280 + }, + { + "epoch": 0.28, + "learning_rate": 0.0005857412236710131, + "loss": 1.5511, + "theoretical_loss": 3.7116066741240283, + "tokens_seen": 908066816 + }, + { + "epoch": 0.28, + "learning_rate": 0.0005857251755265798, + "loss": 1.377, + "theoretical_loss": 3.7115809256310865, + "tokens_seen": 908132352 + }, + { + "epoch": 0.28, + "learning_rate": 0.0005857091273821465, + "loss": 1.4005, + "theoretical_loss": 3.711555179516477, + "tokens_seen": 908197888 + }, + { + "epoch": 0.28, + "learning_rate": 0.0005856930792377132, + "loss": 1.3615, + "theoretical_loss": 3.7115294357798083, + "tokens_seen": 908263424 + }, + { + "epoch": 0.28, + "learning_rate": 0.0005856770310932799, + "loss": 1.4606, + "theoretical_loss": 3.7115036944206885, + "tokens_seen": 908328960 + }, + { + "epoch": 0.28, + "learning_rate": 0.0005856609829488466, + "loss": 1.5266, + "theoretical_loss": 3.7114779554387285, + "tokens_seen": 908394496 + }, + { + "epoch": 0.28, + "learning_rate": 0.0005856449348044133, + "loss": 1.4902, + "theoretical_loss": 3.7114522188335357, + "tokens_seen": 908460032 + }, + { + "epoch": 0.28, + "learning_rate": 0.00058562888665998, + "loss": 1.4073, + "theoretical_loss": 3.7114264846047194, + "tokens_seen": 908525568 + }, + { + "epoch": 0.28, + "learning_rate": 0.0005856128385155467, + "loss": 1.4486, + "theoretical_loss": 3.7114007527518886, + "tokens_seen": 908591104 + }, + { + "epoch": 0.28, + "learning_rate": 0.0005855967903711134, + "loss": 1.4053, + "theoretical_loss": 3.7113750232746536, + "tokens_seen": 908656640 + }, + { + "epoch": 0.28, + "learning_rate": 0.0005855807422266801, + "loss": 1.4675, + "theoretical_loss": 3.711349296172623, + "tokens_seen": 908722176 + }, + { + "epoch": 0.28, + "learning_rate": 0.0005855646940822468, + "loss": 1.4406, + "theoretical_loss": 3.7113235714454063, + "tokens_seen": 908787712 + }, + { + "epoch": 0.28, + "learning_rate": 0.0005855486459378135, + "loss": 1.4971, + "theoretical_loss": 3.711297849092613, + "tokens_seen": 908853248 + }, + { + "epoch": 0.28, + "learning_rate": 0.0005855325977933802, + "loss": 1.4465, + "theoretical_loss": 3.711272129113853, + "tokens_seen": 908918784 + }, + { + "epoch": 0.28, + "learning_rate": 0.0005855165496489469, + "loss": 1.3887, + "theoretical_loss": 3.7112464115087365, + "tokens_seen": 908984320 + }, + { + "epoch": 0.28, + "learning_rate": 0.0005855005015045136, + "loss": 1.4339, + "theoretical_loss": 3.7112206962768717, + "tokens_seen": 909049856 + }, + { + "epoch": 0.28, + "learning_rate": 0.0005854844533600803, + "loss": 1.3169, + "theoretical_loss": 3.7111949834178706, + "tokens_seen": 909115392 + }, + { + "epoch": 0.28, + "learning_rate": 0.000585468405215647, + "loss": 1.3912, + "theoretical_loss": 3.7111692729313415, + "tokens_seen": 909180928 + }, + { + "epoch": 0.28, + "learning_rate": 0.0005854523570712137, + "loss": 1.5002, + "theoretical_loss": 3.7111435648168953, + "tokens_seen": 909246464 + }, + { + "epoch": 0.28, + "learning_rate": 0.0005854363089267804, + "loss": 1.5056, + "theoretical_loss": 3.7111178590741423, + "tokens_seen": 909312000 + }, + { + "epoch": 0.28, + "learning_rate": 0.0005854202607823471, + "loss": 1.378, + "theoretical_loss": 3.7110921557026924, + "tokens_seen": 909377536 + }, + { + "epoch": 0.28, + "learning_rate": 0.0005854042126379138, + "loss": 1.5096, + "theoretical_loss": 3.7110664547021566, + "tokens_seen": 909443072 + }, + { + "epoch": 0.28, + "learning_rate": 0.0005853881644934805, + "loss": 1.3422, + "theoretical_loss": 3.711040756072144, + "tokens_seen": 909508608 + }, + { + "epoch": 0.28, + "learning_rate": 0.0005853721163490472, + "loss": 1.61, + "theoretical_loss": 3.711015059812267, + "tokens_seen": 909574144 + }, + { + "epoch": 0.28, + "learning_rate": 0.0005853560682046139, + "loss": 1.3829, + "theoretical_loss": 3.710989365922135, + "tokens_seen": 909639680 + }, + { + "epoch": 0.28, + "learning_rate": 0.0005853400200601806, + "loss": 1.4805, + "theoretical_loss": 3.710963674401359, + "tokens_seen": 909705216 + }, + { + "epoch": 0.28, + "learning_rate": 0.0005853239719157473, + "loss": 1.4478, + "theoretical_loss": 3.7109379852495503, + "tokens_seen": 909770752 + }, + { + "epoch": 0.28, + "learning_rate": 0.000585307923771314, + "loss": 1.4858, + "theoretical_loss": 3.7109122984663196, + "tokens_seen": 909836288 + }, + { + "epoch": 0.28, + "learning_rate": 0.0005852918756268807, + "loss": 1.4464, + "theoretical_loss": 3.7108866140512777, + "tokens_seen": 909901824 + }, + { + "epoch": 0.28, + "learning_rate": 0.0005852758274824474, + "loss": 1.4019, + "theoretical_loss": 3.7108609320040356, + "tokens_seen": 909967360 + }, + { + "epoch": 0.28, + "learning_rate": 0.0005852597793380141, + "loss": 1.4753, + "theoretical_loss": 3.7108352523242045, + "tokens_seen": 910032896 + }, + { + "epoch": 0.28, + "learning_rate": 0.0005852437311935808, + "loss": 1.4616, + "theoretical_loss": 3.710809575011397, + "tokens_seen": 910098432 + }, + { + "epoch": 0.28, + "learning_rate": 0.0005852276830491475, + "loss": 1.4526, + "theoretical_loss": 3.7107839000652225, + "tokens_seen": 910163968 + }, + { + "epoch": 0.28, + "learning_rate": 0.0005852116349047142, + "loss": 1.5003, + "theoretical_loss": 3.710758227485294, + "tokens_seen": 910229504 + }, + { + "epoch": 0.28, + "learning_rate": 0.0005851955867602809, + "loss": 1.3859, + "theoretical_loss": 3.710732557271222, + "tokens_seen": 910295040 + }, + { + "epoch": 0.28, + "learning_rate": 0.0005851795386158476, + "loss": 1.39, + "theoretical_loss": 3.710706889422619, + "tokens_seen": 910360576 + }, + { + "epoch": 0.28, + "learning_rate": 0.0005851634904714142, + "loss": 1.4742, + "theoretical_loss": 3.7106812239390967, + "tokens_seen": 910426112 + }, + { + "epoch": 0.28, + "learning_rate": 0.000585147442326981, + "loss": 1.3647, + "theoretical_loss": 3.7106555608202667, + "tokens_seen": 910491648 + }, + { + "epoch": 0.28, + "learning_rate": 0.0005851313941825476, + "loss": 1.4438, + "theoretical_loss": 3.71062990006574, + "tokens_seen": 910557184 + }, + { + "epoch": 0.28, + "learning_rate": 0.0005851153460381144, + "loss": 1.3738, + "theoretical_loss": 3.7106042416751306, + "tokens_seen": 910622720 + }, + { + "epoch": 0.28, + "learning_rate": 0.000585099297893681, + "loss": 1.4344, + "theoretical_loss": 3.7105785856480495, + "tokens_seen": 910688256 + }, + { + "epoch": 0.28, + "learning_rate": 0.0005850832497492478, + "loss": 1.4284, + "theoretical_loss": 3.710552931984109, + "tokens_seen": 910753792 + }, + { + "epoch": 0.28, + "learning_rate": 0.0005850672016048144, + "loss": 1.374, + "theoretical_loss": 3.7105272806829213, + "tokens_seen": 910819328 + }, + { + "epoch": 0.28, + "learning_rate": 0.0005850511534603812, + "loss": 1.3983, + "theoretical_loss": 3.7105016317440986, + "tokens_seen": 910884864 + }, + { + "epoch": 0.28, + "objective/train/docs_used": 336141, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.4937772750854492, + "objective/train/theoretical_loss": 3.710475985167254, + "objective/train/tokens_used": 931410400, + "theoretical_loss": 3.710475985167254, + "tokens_seen": 910950400 + }, + { + "epoch": 0.28, + "learning_rate": 0.0005850351053159478, + "loss": 1.5064, + "theoretical_loss": 3.710475985167254, + "tokens_seen": 910950400 + }, + { + "epoch": 0.28, + "learning_rate": 0.0005850190571715145, + "loss": 1.4069, + "theoretical_loss": 3.710450340952, + "tokens_seen": 911015936 + }, + { + "epoch": 0.28, + "learning_rate": 0.0005850030090270812, + "loss": 1.412, + "theoretical_loss": 3.710424699097949, + "tokens_seen": 911081472 + }, + { + "epoch": 0.28, + "learning_rate": 0.0005849869608826479, + "loss": 1.4269, + "theoretical_loss": 3.710399059604714, + "tokens_seen": 911147008 + }, + { + "epoch": 0.28, + "learning_rate": 0.0005849709127382146, + "loss": 1.458, + "theoretical_loss": 3.710373422471908, + "tokens_seen": 911212544 + }, + { + "epoch": 0.28, + "learning_rate": 0.0005849548645937814, + "loss": 1.3368, + "theoretical_loss": 3.710347787699143, + "tokens_seen": 911278080 + }, + { + "epoch": 0.28, + "learning_rate": 0.000584938816449348, + "loss": 1.3534, + "theoretical_loss": 3.710322155286033, + "tokens_seen": 911343616 + }, + { + "epoch": 0.28, + "learning_rate": 0.0005849227683049148, + "loss": 1.4083, + "theoretical_loss": 3.710296525232191, + "tokens_seen": 911409152 + }, + { + "epoch": 0.28, + "learning_rate": 0.0005849067201604814, + "loss": 1.3937, + "theoretical_loss": 3.71027089753723, + "tokens_seen": 911474688 + }, + { + "epoch": 0.28, + "learning_rate": 0.0005848906720160482, + "loss": 1.381, + "theoretical_loss": 3.7102452722007637, + "tokens_seen": 911540224 + }, + { + "epoch": 0.28, + "learning_rate": 0.0005848746238716148, + "loss": 1.3109, + "theoretical_loss": 3.710219649222404, + "tokens_seen": 911605760 + }, + { + "epoch": 0.28, + "learning_rate": 0.0005848585757271816, + "loss": 1.4239, + "theoretical_loss": 3.710194028601767, + "tokens_seen": 911671296 + }, + { + "epoch": 0.28, + "learning_rate": 0.0005848425275827482, + "loss": 1.4722, + "theoretical_loss": 3.7101684103384645, + "tokens_seen": 911736832 + }, + { + "epoch": 0.28, + "learning_rate": 0.000584826479438315, + "loss": 1.352, + "theoretical_loss": 3.7101427944321106, + "tokens_seen": 911802368 + }, + { + "epoch": 0.28, + "learning_rate": 0.0005848104312938816, + "loss": 1.5759, + "theoretical_loss": 3.7101171808823192, + "tokens_seen": 911867904 + }, + { + "epoch": 0.28, + "learning_rate": 0.0005847943831494484, + "loss": 1.4841, + "theoretical_loss": 3.7100915696887036, + "tokens_seen": 911933440 + }, + { + "epoch": 0.28, + "learning_rate": 0.000584778335005015, + "loss": 1.3546, + "theoretical_loss": 3.710065960850878, + "tokens_seen": 911998976 + }, + { + "epoch": 0.28, + "learning_rate": 0.0005847622868605818, + "loss": 1.4744, + "theoretical_loss": 3.710040354368457, + "tokens_seen": 912064512 + }, + { + "epoch": 0.28, + "learning_rate": 0.0005847462387161484, + "loss": 1.5446, + "theoretical_loss": 3.7100147502410543, + "tokens_seen": 912130048 + }, + { + "epoch": 0.28, + "learning_rate": 0.0005847301905717152, + "loss": 1.4717, + "theoretical_loss": 3.709989148468284, + "tokens_seen": 912195584 + }, + { + "epoch": 0.28, + "learning_rate": 0.0005847141424272818, + "loss": 1.4515, + "theoretical_loss": 3.7099635490497604, + "tokens_seen": 912261120 + }, + { + "epoch": 0.28, + "learning_rate": 0.0005846980942828486, + "loss": 1.4487, + "theoretical_loss": 3.709937951985099, + "tokens_seen": 912326656 + }, + { + "epoch": 0.28, + "learning_rate": 0.0005846820461384152, + "loss": 1.5485, + "theoretical_loss": 3.7099123572739128, + "tokens_seen": 912392192 + }, + { + "epoch": 0.28, + "learning_rate": 0.000584665997993982, + "loss": 1.3747, + "theoretical_loss": 3.709886764915817, + "tokens_seen": 912457728 + }, + { + "epoch": 0.28, + "learning_rate": 0.0005846499498495486, + "loss": 1.4324, + "theoretical_loss": 3.7098611749104258, + "tokens_seen": 912523264 + }, + { + "epoch": 0.28, + "learning_rate": 0.0005846339017051154, + "loss": 1.4485, + "theoretical_loss": 3.709835587257355, + "tokens_seen": 912588800 + }, + { + "epoch": 0.28, + "learning_rate": 0.000584617853560682, + "loss": 1.4236, + "theoretical_loss": 3.7098100019562192, + "tokens_seen": 912654336 + }, + { + "epoch": 0.28, + "learning_rate": 0.0005846018054162488, + "loss": 1.3125, + "theoretical_loss": 3.7097844190066325, + "tokens_seen": 912719872 + }, + { + "epoch": 0.28, + "learning_rate": 0.0005845857572718154, + "loss": 1.4315, + "theoretical_loss": 3.7097588384082107, + "tokens_seen": 912785408 + }, + { + "epoch": 0.28, + "learning_rate": 0.0005845697091273822, + "loss": 1.4686, + "theoretical_loss": 3.7097332601605686, + "tokens_seen": 912850944 + }, + { + "epoch": 0.28, + "learning_rate": 0.0005845536609829488, + "loss": 1.3791, + "theoretical_loss": 3.7097076842633214, + "tokens_seen": 912916480 + }, + { + "epoch": 0.28, + "learning_rate": 0.0005845376128385156, + "loss": 1.371, + "theoretical_loss": 3.709682110716085, + "tokens_seen": 912982016 + }, + { + "epoch": 0.28, + "learning_rate": 0.0005845215646940822, + "loss": 1.4722, + "theoretical_loss": 3.709656539518474, + "tokens_seen": 913047552 + }, + { + "epoch": 0.28, + "learning_rate": 0.000584505516549649, + "loss": 1.4538, + "theoretical_loss": 3.709630970670105, + "tokens_seen": 913113088 + }, + { + "epoch": 0.28, + "learning_rate": 0.0005844894684052156, + "loss": 1.4901, + "theoretical_loss": 3.709605404170592, + "tokens_seen": 913178624 + }, + { + "epoch": 0.28, + "learning_rate": 0.0005844734202607824, + "loss": 1.4337, + "theoretical_loss": 3.7095798400195523, + "tokens_seen": 913244160 + }, + { + "epoch": 0.28, + "learning_rate": 0.000584457372116349, + "loss": 1.4344, + "theoretical_loss": 3.7095542782166, + "tokens_seen": 913309696 + }, + { + "epoch": 0.28, + "learning_rate": 0.0005844413239719158, + "loss": 1.3831, + "theoretical_loss": 3.7095287187613524, + "tokens_seen": 913375232 + }, + { + "epoch": 0.28, + "learning_rate": 0.0005844252758274825, + "loss": 1.3526, + "theoretical_loss": 3.709503161653425, + "tokens_seen": 913440768 + }, + { + "epoch": 0.28, + "learning_rate": 0.0005844092276830492, + "loss": 1.5338, + "theoretical_loss": 3.7094776068924338, + "tokens_seen": 913506304 + }, + { + "epoch": 0.28, + "learning_rate": 0.0005843931795386159, + "loss": 1.4281, + "theoretical_loss": 3.7094520544779948, + "tokens_seen": 913571840 + }, + { + "epoch": 0.28, + "learning_rate": 0.0005843771313941826, + "loss": 1.4735, + "theoretical_loss": 3.7094265044097243, + "tokens_seen": 913637376 + }, + { + "epoch": 0.28, + "learning_rate": 0.0005843610832497493, + "loss": 1.5132, + "theoretical_loss": 3.7094009566872383, + "tokens_seen": 913702912 + }, + { + "epoch": 0.28, + "learning_rate": 0.000584345035105316, + "loss": 1.3973, + "theoretical_loss": 3.7093754113101536, + "tokens_seen": 913768448 + }, + { + "epoch": 0.28, + "learning_rate": 0.0005843289869608827, + "loss": 1.3188, + "theoretical_loss": 3.709349868278087, + "tokens_seen": 913833984 + }, + { + "epoch": 0.28, + "learning_rate": 0.0005843129388164494, + "loss": 1.3827, + "theoretical_loss": 3.7093243275906547, + "tokens_seen": 913899520 + }, + { + "epoch": 0.28, + "learning_rate": 0.0005842968906720161, + "loss": 1.4766, + "theoretical_loss": 3.7092987892474727, + "tokens_seen": 913965056 + }, + { + "epoch": 0.28, + "learning_rate": 0.0005842808425275828, + "loss": 1.4739, + "theoretical_loss": 3.709273253248159, + "tokens_seen": 914030592 + }, + { + "epoch": 0.28, + "learning_rate": 0.0005842647943831495, + "loss": 1.3811, + "theoretical_loss": 3.7092477195923292, + "tokens_seen": 914096128 + }, + { + "epoch": 0.28, + "learning_rate": 0.0005842487462387162, + "loss": 1.3737, + "theoretical_loss": 3.709222188279602, + "tokens_seen": 914161664 + }, + { + "epoch": 0.28, + "objective/train/docs_used": 337261, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.4412689208984375, + "objective/train/theoretical_loss": 3.709196659309592, + "objective/train/tokens_used": 934687200, + "theoretical_loss": 3.709196659309592, + "tokens_seen": 914227200 + }, + { + "epoch": 0.28, + "learning_rate": 0.0005842326980942829, + "loss": 1.5245, + "theoretical_loss": 3.709196659309592, + "tokens_seen": 914227200 + }, + { + "epoch": 0.28, + "learning_rate": 0.0005842166499498496, + "loss": 1.39, + "theoretical_loss": 3.7091711326819192, + "tokens_seen": 914292736 + }, + { + "epoch": 0.28, + "learning_rate": 0.0005842006018054163, + "loss": 1.4462, + "theoretical_loss": 3.7091456083961982, + "tokens_seen": 914358272 + }, + { + "epoch": 0.28, + "learning_rate": 0.000584184553660983, + "loss": 1.3996, + "theoretical_loss": 3.7091200864520477, + "tokens_seen": 914423808 + }, + { + "epoch": 0.28, + "learning_rate": 0.0005841685055165497, + "loss": 1.4515, + "theoretical_loss": 3.709094566849085, + "tokens_seen": 914489344 + }, + { + "epoch": 0.28, + "learning_rate": 0.0005841524573721164, + "loss": 1.4263, + "theoretical_loss": 3.709069049586927, + "tokens_seen": 914554880 + }, + { + "epoch": 0.28, + "learning_rate": 0.0005841364092276831, + "loss": 1.4291, + "theoretical_loss": 3.7090435346651915, + "tokens_seen": 914620416 + }, + { + "epoch": 0.28, + "learning_rate": 0.0005841203610832498, + "loss": 1.4314, + "theoretical_loss": 3.7090180220834967, + "tokens_seen": 914685952 + }, + { + "epoch": 0.28, + "learning_rate": 0.0005841043129388165, + "loss": 1.4081, + "theoretical_loss": 3.7089925118414597, + "tokens_seen": 914751488 + }, + { + "epoch": 0.28, + "learning_rate": 0.0005840882647943832, + "loss": 1.418, + "theoretical_loss": 3.7089670039386986, + "tokens_seen": 914817024 + }, + { + "epoch": 0.28, + "learning_rate": 0.0005840722166499499, + "loss": 1.4088, + "theoretical_loss": 3.7089414983748314, + "tokens_seen": 914882560 + }, + { + "epoch": 0.28, + "learning_rate": 0.0005840561685055166, + "loss": 1.3557, + "theoretical_loss": 3.708915995149476, + "tokens_seen": 914948096 + }, + { + "epoch": 0.28, + "learning_rate": 0.0005840401203610833, + "loss": 1.4272, + "theoretical_loss": 3.70889049426225, + "tokens_seen": 915013632 + }, + { + "epoch": 0.28, + "learning_rate": 0.00058402407221665, + "loss": 1.4702, + "theoretical_loss": 3.708864995712773, + "tokens_seen": 915079168 + }, + { + "epoch": 0.28, + "learning_rate": 0.0005840080240722167, + "loss": 1.4587, + "theoretical_loss": 3.708839499500662, + "tokens_seen": 915144704 + }, + { + "epoch": 0.28, + "learning_rate": 0.0005839919759277834, + "loss": 1.4702, + "theoretical_loss": 3.708814005625536, + "tokens_seen": 915210240 + }, + { + "epoch": 0.28, + "learning_rate": 0.0005839759277833501, + "loss": 1.4659, + "theoretical_loss": 3.708788514087013, + "tokens_seen": 915275776 + }, + { + "epoch": 0.28, + "learning_rate": 0.0005839598796389168, + "loss": 1.4437, + "theoretical_loss": 3.7087630248847123, + "tokens_seen": 915341312 + }, + { + "epoch": 0.28, + "learning_rate": 0.0005839438314944835, + "loss": 1.4212, + "theoretical_loss": 3.7087375380182515, + "tokens_seen": 915406848 + }, + { + "epoch": 0.28, + "learning_rate": 0.0005839277833500502, + "loss": 1.3661, + "theoretical_loss": 3.7087120534872504, + "tokens_seen": 915472384 + }, + { + "epoch": 0.28, + "learning_rate": 0.0005839117352056169, + "loss": 1.4813, + "theoretical_loss": 3.708686571291327, + "tokens_seen": 915537920 + }, + { + "epoch": 0.28, + "learning_rate": 0.0005838956870611836, + "loss": 1.4409, + "theoretical_loss": 3.7086610914301006, + "tokens_seen": 915603456 + }, + { + "epoch": 0.28, + "learning_rate": 0.0005838796389167503, + "loss": 1.3764, + "theoretical_loss": 3.70863561390319, + "tokens_seen": 915668992 + }, + { + "epoch": 0.28, + "learning_rate": 0.000583863590772317, + "loss": 1.4412, + "theoretical_loss": 3.708610138710215, + "tokens_seen": 915734528 + }, + { + "epoch": 0.28, + "learning_rate": 0.0005838475426278837, + "loss": 1.4459, + "theoretical_loss": 3.708584665850794, + "tokens_seen": 915800064 + }, + { + "epoch": 0.28, + "learning_rate": 0.0005838314944834504, + "loss": 1.427, + "theoretical_loss": 3.7085591953245465, + "tokens_seen": 915865600 + }, + { + "epoch": 0.28, + "learning_rate": 0.0005838154463390171, + "loss": 1.4038, + "theoretical_loss": 3.708533727131092, + "tokens_seen": 915931136 + }, + { + "epoch": 0.28, + "learning_rate": 0.0005837993981945838, + "loss": 1.4281, + "theoretical_loss": 3.7085082612700493, + "tokens_seen": 915996672 + }, + { + "epoch": 0.28, + "learning_rate": 0.0005837833500501505, + "loss": 1.4556, + "theoretical_loss": 3.7084827977410395, + "tokens_seen": 916062208 + }, + { + "epoch": 0.28, + "learning_rate": 0.0005837673019057172, + "loss": 1.4509, + "theoretical_loss": 3.70845733654368, + "tokens_seen": 916127744 + }, + { + "epoch": 0.28, + "learning_rate": 0.0005837512537612839, + "loss": 1.5881, + "theoretical_loss": 3.708431877677593, + "tokens_seen": 916193280 + }, + { + "epoch": 0.28, + "learning_rate": 0.0005837352056168505, + "loss": 1.4672, + "theoretical_loss": 3.708406421142396, + "tokens_seen": 916258816 + }, + { + "epoch": 0.28, + "learning_rate": 0.0005837191574724173, + "loss": 1.468, + "theoretical_loss": 3.7083809669377104, + "tokens_seen": 916324352 + }, + { + "epoch": 0.28, + "learning_rate": 0.0005837031093279839, + "loss": 1.4901, + "theoretical_loss": 3.7083555150631557, + "tokens_seen": 916389888 + }, + { + "epoch": 0.28, + "learning_rate": 0.0005836870611835507, + "loss": 1.3536, + "theoretical_loss": 3.708330065518352, + "tokens_seen": 916455424 + }, + { + "epoch": 0.28, + "learning_rate": 0.0005836710130391173, + "loss": 1.5791, + "theoretical_loss": 3.70830461830292, + "tokens_seen": 916520960 + }, + { + "epoch": 0.28, + "learning_rate": 0.0005836549648946841, + "loss": 1.2958, + "theoretical_loss": 3.7082791734164786, + "tokens_seen": 916586496 + }, + { + "epoch": 0.28, + "learning_rate": 0.0005836389167502507, + "loss": 1.3755, + "theoretical_loss": 3.708253730858649, + "tokens_seen": 916652032 + }, + { + "epoch": 0.28, + "learning_rate": 0.0005836228686058175, + "loss": 1.4345, + "theoretical_loss": 3.7082282906290525, + "tokens_seen": 916717568 + }, + { + "epoch": 0.28, + "learning_rate": 0.0005836068204613841, + "loss": 1.4073, + "theoretical_loss": 3.7082028527273083, + "tokens_seen": 916783104 + }, + { + "epoch": 0.28, + "learning_rate": 0.0005835907723169509, + "loss": 1.3746, + "theoretical_loss": 3.708177417153037, + "tokens_seen": 916848640 + }, + { + "epoch": 0.28, + "learning_rate": 0.0005835747241725175, + "loss": 1.4406, + "theoretical_loss": 3.70815198390586, + "tokens_seen": 916914176 + }, + { + "epoch": 0.28, + "learning_rate": 0.0005835586760280843, + "loss": 1.4189, + "theoretical_loss": 3.708126552985398, + "tokens_seen": 916979712 + }, + { + "epoch": 0.28, + "learning_rate": 0.0005835426278836509, + "loss": 1.2605, + "theoretical_loss": 3.7081011243912716, + "tokens_seen": 917045248 + }, + { + "epoch": 0.28, + "learning_rate": 0.0005835265797392177, + "loss": 1.398, + "theoretical_loss": 3.7080756981231016, + "tokens_seen": 917110784 + }, + { + "epoch": 0.28, + "learning_rate": 0.0005835105315947843, + "loss": 1.3471, + "theoretical_loss": 3.7080502741805104, + "tokens_seen": 917176320 + }, + { + "epoch": 0.28, + "learning_rate": 0.0005834944834503511, + "loss": 1.5443, + "theoretical_loss": 3.708024852563117, + "tokens_seen": 917241856 + }, + { + "epoch": 0.28, + "learning_rate": 0.0005834784353059177, + "loss": 1.4372, + "theoretical_loss": 3.707999433270544, + "tokens_seen": 917307392 + }, + { + "epoch": 0.28, + "learning_rate": 0.0005834623871614845, + "loss": 1.411, + "theoretical_loss": 3.7079740163024124, + "tokens_seen": 917372928 + }, + { + "epoch": 0.28, + "learning_rate": 0.0005834463390170511, + "loss": 1.4615, + "theoretical_loss": 3.7079486016583436, + "tokens_seen": 917438464 + }, + { + "epoch": 0.28, + "objective/train/docs_used": 338425, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.5118355751037598, + "objective/train/theoretical_loss": 3.7079231893379596, + "objective/train/tokens_used": 937964000, + "theoretical_loss": 3.7079231893379596, + "tokens_seen": 917504000 + }, + { + "epoch": 0.28, + "learning_rate": 0.0005834302908726179, + "loss": 1.5411, + "theoretical_loss": 3.7079231893379596, + "tokens_seen": 917504000 + }, + { + "epoch": 0.28, + "learning_rate": 0.0005834142427281845, + "loss": 1.3717, + "theoretical_loss": 3.707897779340881, + "tokens_seen": 917569536 + }, + { + "epoch": 0.28, + "learning_rate": 0.0005833981945837513, + "loss": 1.4813, + "theoretical_loss": 3.7078723716667303, + "tokens_seen": 917635072 + }, + { + "epoch": 0.28, + "learning_rate": 0.0005833821464393179, + "loss": 1.3772, + "theoretical_loss": 3.7078469663151292, + "tokens_seen": 917700608 + }, + { + "epoch": 0.28, + "learning_rate": 0.0005833660982948847, + "loss": 1.3964, + "theoretical_loss": 3.707821563285699, + "tokens_seen": 917766144 + }, + { + "epoch": 0.28, + "learning_rate": 0.0005833500501504513, + "loss": 1.503, + "theoretical_loss": 3.707796162578062, + "tokens_seen": 917831680 + }, + { + "epoch": 0.28, + "learning_rate": 0.0005833340020060181, + "loss": 1.4448, + "theoretical_loss": 3.7077707641918405, + "tokens_seen": 917897216 + }, + { + "epoch": 0.28, + "learning_rate": 0.0005833179538615847, + "loss": 1.4938, + "theoretical_loss": 3.7077453681266563, + "tokens_seen": 917962752 + }, + { + "epoch": 0.28, + "learning_rate": 0.0005833019057171515, + "loss": 1.3731, + "theoretical_loss": 3.7077199743821314, + "tokens_seen": 918028288 + }, + { + "epoch": 0.28, + "learning_rate": 0.0005832858575727181, + "loss": 1.4721, + "theoretical_loss": 3.7076945829578882, + "tokens_seen": 918093824 + }, + { + "epoch": 0.28, + "learning_rate": 0.0005832698094282849, + "loss": 1.4762, + "theoretical_loss": 3.7076691938535493, + "tokens_seen": 918159360 + }, + { + "epoch": 0.28, + "learning_rate": 0.0005832537612838515, + "loss": 1.4024, + "theoretical_loss": 3.707643807068737, + "tokens_seen": 918224896 + }, + { + "epoch": 0.28, + "learning_rate": 0.0005832377131394183, + "loss": 1.4462, + "theoretical_loss": 3.7076184226030744, + "tokens_seen": 918290432 + }, + { + "epoch": 0.28, + "learning_rate": 0.000583221664994985, + "loss": 1.4355, + "theoretical_loss": 3.707593040456184, + "tokens_seen": 918355968 + }, + { + "epoch": 0.28, + "learning_rate": 0.0005832056168505517, + "loss": 1.4711, + "theoretical_loss": 3.7075676606276877, + "tokens_seen": 918421504 + }, + { + "epoch": 0.28, + "learning_rate": 0.0005831895687061183, + "loss": 1.4684, + "theoretical_loss": 3.707542283117209, + "tokens_seen": 918487040 + }, + { + "epoch": 0.28, + "learning_rate": 0.0005831735205616851, + "loss": 1.406, + "theoretical_loss": 3.7075169079243704, + "tokens_seen": 918552576 + }, + { + "epoch": 0.28, + "learning_rate": 0.0005831574724172517, + "loss": 1.5336, + "theoretical_loss": 3.7074915350487956, + "tokens_seen": 918618112 + }, + { + "epoch": 0.28, + "learning_rate": 0.0005831414242728185, + "loss": 1.4214, + "theoretical_loss": 3.707466164490107, + "tokens_seen": 918683648 + }, + { + "epoch": 0.28, + "learning_rate": 0.0005831253761283851, + "loss": 1.5675, + "theoretical_loss": 3.707440796247928, + "tokens_seen": 918749184 + }, + { + "epoch": 0.28, + "learning_rate": 0.0005831093279839519, + "loss": 1.4178, + "theoretical_loss": 3.7074154303218823, + "tokens_seen": 918814720 + }, + { + "epoch": 0.28, + "learning_rate": 0.0005830932798395186, + "loss": 1.5745, + "theoretical_loss": 3.707390066711593, + "tokens_seen": 918880256 + }, + { + "epoch": 0.28, + "learning_rate": 0.0005830772316950853, + "loss": 1.4736, + "theoretical_loss": 3.7073647054166834, + "tokens_seen": 918945792 + }, + { + "epoch": 0.28, + "learning_rate": 0.000583061183550652, + "loss": 1.3422, + "theoretical_loss": 3.7073393464367768, + "tokens_seen": 919011328 + }, + { + "epoch": 0.28, + "learning_rate": 0.0005830451354062187, + "loss": 1.5493, + "theoretical_loss": 3.707313989771497, + "tokens_seen": 919076864 + }, + { + "epoch": 0.28, + "learning_rate": 0.0005830290872617854, + "loss": 1.5416, + "theoretical_loss": 3.707288635420468, + "tokens_seen": 919142400 + }, + { + "epoch": 0.28, + "learning_rate": 0.0005830130391173521, + "loss": 1.4012, + "theoretical_loss": 3.7072632833833135, + "tokens_seen": 919207936 + }, + { + "epoch": 0.28, + "learning_rate": 0.0005829969909729188, + "loss": 1.4329, + "theoretical_loss": 3.707237933659657, + "tokens_seen": 919273472 + }, + { + "epoch": 0.28, + "learning_rate": 0.0005829809428284855, + "loss": 1.404, + "theoretical_loss": 3.7072125862491228, + "tokens_seen": 919339008 + }, + { + "epoch": 0.28, + "learning_rate": 0.0005829648946840522, + "loss": 1.4095, + "theoretical_loss": 3.7071872411513347, + "tokens_seen": 919404544 + }, + { + "epoch": 0.28, + "learning_rate": 0.0005829488465396189, + "loss": 1.4964, + "theoretical_loss": 3.707161898365918, + "tokens_seen": 919470080 + }, + { + "epoch": 0.28, + "learning_rate": 0.0005829327983951856, + "loss": 1.466, + "theoretical_loss": 3.707136557892495, + "tokens_seen": 919535616 + }, + { + "epoch": 0.28, + "learning_rate": 0.0005829167502507523, + "loss": 1.4711, + "theoretical_loss": 3.7071112197306917, + "tokens_seen": 919601152 + }, + { + "epoch": 0.28, + "learning_rate": 0.000582900702106319, + "loss": 1.3384, + "theoretical_loss": 3.7070858838801315, + "tokens_seen": 919666688 + }, + { + "epoch": 0.28, + "learning_rate": 0.0005828846539618857, + "loss": 1.4855, + "theoretical_loss": 3.7070605503404392, + "tokens_seen": 919732224 + }, + { + "epoch": 0.28, + "learning_rate": 0.0005828686058174524, + "loss": 1.5092, + "theoretical_loss": 3.7070352191112397, + "tokens_seen": 919797760 + }, + { + "epoch": 0.28, + "learning_rate": 0.0005828525576730191, + "loss": 1.4404, + "theoretical_loss": 3.707009890192157, + "tokens_seen": 919863296 + }, + { + "epoch": 0.28, + "learning_rate": 0.0005828365095285858, + "loss": 1.5537, + "theoretical_loss": 3.7069845635828167, + "tokens_seen": 919928832 + }, + { + "epoch": 0.28, + "learning_rate": 0.0005828204613841525, + "loss": 1.4181, + "theoretical_loss": 3.7069592392828428, + "tokens_seen": 919994368 + }, + { + "epoch": 0.28, + "learning_rate": 0.0005828044132397192, + "loss": 1.3738, + "theoretical_loss": 3.706933917291861, + "tokens_seen": 920059904 + }, + { + "epoch": 0.28, + "learning_rate": 0.0005827883650952859, + "loss": 1.2971, + "theoretical_loss": 3.7069085976094955, + "tokens_seen": 920125440 + }, + { + "epoch": 0.28, + "learning_rate": 0.0005827723169508526, + "loss": 1.3327, + "theoretical_loss": 3.7068832802353726, + "tokens_seen": 920190976 + }, + { + "epoch": 0.28, + "learning_rate": 0.0005827562688064193, + "loss": 1.4677, + "theoretical_loss": 3.7068579651691165, + "tokens_seen": 920256512 + }, + { + "epoch": 0.28, + "learning_rate": 0.000582740220661986, + "loss": 1.3828, + "theoretical_loss": 3.7068326524103528, + "tokens_seen": 920322048 + }, + { + "epoch": 0.28, + "learning_rate": 0.0005827241725175527, + "loss": 1.463, + "theoretical_loss": 3.706807341958706, + "tokens_seen": 920387584 + }, + { + "epoch": 0.28, + "learning_rate": 0.0005827081243731194, + "loss": 1.4996, + "theoretical_loss": 3.706782033813804, + "tokens_seen": 920453120 + }, + { + "epoch": 0.28, + "learning_rate": 0.0005826920762286861, + "loss": 1.4318, + "theoretical_loss": 3.70675672797527, + "tokens_seen": 920518656 + }, + { + "epoch": 0.28, + "learning_rate": 0.0005826760280842528, + "loss": 1.4285, + "theoretical_loss": 3.7067314244427303, + "tokens_seen": 920584192 + }, + { + "epoch": 0.28, + "learning_rate": 0.0005826599799398195, + "loss": 1.4781, + "theoretical_loss": 3.7067061232158105, + "tokens_seen": 920649728 + }, + { + "epoch": 0.28, + "learning_rate": 0.0005826439317953862, + "loss": 1.6334, + "theoretical_loss": 3.7066808242941374, + "tokens_seen": 920715264 + }, + { + "epoch": 0.28, + "objective/train/docs_used": 339576, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.409914493560791, + "objective/train/theoretical_loss": 3.706655527677335, + "objective/train/tokens_used": 941240800, + "theoretical_loss": 3.706655527677335, + "tokens_seen": 920780800 + }, + { + "epoch": 0.28, + "learning_rate": 0.0005826278836509529, + "loss": 1.544, + "theoretical_loss": 3.706655527677335, + "tokens_seen": 920780800 + }, + { + "epoch": 0.28, + "learning_rate": 0.0005826118355065196, + "loss": 1.5836, + "theoretical_loss": 3.706630233365032, + "tokens_seen": 920846336 + }, + { + "epoch": 0.28, + "learning_rate": 0.0005825957873620863, + "loss": 1.4968, + "theoretical_loss": 3.7066049413568516, + "tokens_seen": 920911872 + }, + { + "epoch": 0.28, + "learning_rate": 0.0005825797392176531, + "loss": 1.5674, + "theoretical_loss": 3.7065796516524214, + "tokens_seen": 920977408 + }, + { + "epoch": 0.28, + "learning_rate": 0.0005825636910732197, + "loss": 1.4608, + "theoretical_loss": 3.7065543642513683, + "tokens_seen": 921042944 + }, + { + "epoch": 0.28, + "learning_rate": 0.0005825476429287865, + "loss": 1.4025, + "theoretical_loss": 3.7065290791533174, + "tokens_seen": 921108480 + }, + { + "epoch": 0.28, + "learning_rate": 0.0005825315947843531, + "loss": 1.4632, + "theoretical_loss": 3.7065037963578957, + "tokens_seen": 921174016 + }, + { + "epoch": 0.28, + "learning_rate": 0.0005825155466399199, + "loss": 1.3362, + "theoretical_loss": 3.7064785158647293, + "tokens_seen": 921239552 + }, + { + "epoch": 0.28, + "learning_rate": 0.0005824994984954865, + "loss": 1.3366, + "theoretical_loss": 3.7064532376734456, + "tokens_seen": 921305088 + }, + { + "epoch": 0.28, + "learning_rate": 0.0005824834503510532, + "loss": 1.4683, + "theoretical_loss": 3.70642796178367, + "tokens_seen": 921370624 + }, + { + "epoch": 0.28, + "learning_rate": 0.0005824674022066199, + "loss": 1.4162, + "theoretical_loss": 3.706402688195031, + "tokens_seen": 921436160 + }, + { + "epoch": 0.28, + "learning_rate": 0.0005824513540621866, + "loss": 1.3857, + "theoretical_loss": 3.706377416907154, + "tokens_seen": 921501696 + }, + { + "epoch": 0.28, + "learning_rate": 0.0005824353059177533, + "loss": 1.4886, + "theoretical_loss": 3.7063521479196666, + "tokens_seen": 921567232 + }, + { + "epoch": 0.28, + "learning_rate": 0.00058241925777332, + "loss": 1.501, + "theoretical_loss": 3.7063268812321954, + "tokens_seen": 921632768 + }, + { + "epoch": 0.28, + "learning_rate": 0.0005824032096288867, + "loss": 1.5034, + "theoretical_loss": 3.7063016168443683, + "tokens_seen": 921698304 + }, + { + "epoch": 0.28, + "learning_rate": 0.0005823871614844534, + "loss": 1.4386, + "theoretical_loss": 3.706276354755812, + "tokens_seen": 921763840 + }, + { + "epoch": 0.28, + "learning_rate": 0.0005823711133400201, + "loss": 1.3161, + "theoretical_loss": 3.706251094966153, + "tokens_seen": 921829376 + }, + { + "epoch": 0.28, + "learning_rate": 0.0005823550651955868, + "loss": 1.5195, + "theoretical_loss": 3.7062258374750208, + "tokens_seen": 921894912 + }, + { + "epoch": 0.28, + "learning_rate": 0.0005823390170511534, + "loss": 1.4413, + "theoretical_loss": 3.7062005822820403, + "tokens_seen": 921960448 + }, + { + "epoch": 0.28, + "learning_rate": 0.0005823229689067202, + "loss": 1.4602, + "theoretical_loss": 3.7061753293868414, + "tokens_seen": 922025984 + }, + { + "epoch": 0.28, + "learning_rate": 0.0005823069207622868, + "loss": 1.5883, + "theoretical_loss": 3.70615007878905, + "tokens_seen": 922091520 + }, + { + "epoch": 0.28, + "learning_rate": 0.0005822908726178536, + "loss": 1.3211, + "theoretical_loss": 3.706124830488295, + "tokens_seen": 922157056 + }, + { + "epoch": 0.28, + "learning_rate": 0.0005822748244734202, + "loss": 1.3768, + "theoretical_loss": 3.7060995844842033, + "tokens_seen": 922222592 + }, + { + "epoch": 0.28, + "learning_rate": 0.000582258776328987, + "loss": 1.5958, + "theoretical_loss": 3.706074340776403, + "tokens_seen": 922288128 + }, + { + "epoch": 0.28, + "learning_rate": 0.0005822427281845536, + "loss": 1.421, + "theoretical_loss": 3.7060490993645225, + "tokens_seen": 922353664 + }, + { + "epoch": 0.28, + "learning_rate": 0.0005822266800401204, + "loss": 1.3998, + "theoretical_loss": 3.70602386024819, + "tokens_seen": 922419200 + }, + { + "epoch": 0.28, + "learning_rate": 0.000582210631895687, + "loss": 1.3567, + "theoretical_loss": 3.7059986234270332, + "tokens_seen": 922484736 + }, + { + "epoch": 0.28, + "learning_rate": 0.0005821945837512538, + "loss": 1.4341, + "theoretical_loss": 3.7059733889006807, + "tokens_seen": 922550272 + }, + { + "epoch": 0.28, + "learning_rate": 0.0005821785356068204, + "loss": 1.3597, + "theoretical_loss": 3.70594815666876, + "tokens_seen": 922615808 + }, + { + "epoch": 0.28, + "learning_rate": 0.0005821624874623872, + "loss": 1.4176, + "theoretical_loss": 3.7059229267309006, + "tokens_seen": 922681344 + }, + { + "epoch": 0.28, + "learning_rate": 0.0005821464393179538, + "loss": 1.3957, + "theoretical_loss": 3.7058976990867305, + "tokens_seen": 922746880 + }, + { + "epoch": 0.28, + "learning_rate": 0.0005821303911735206, + "loss": 1.3872, + "theoretical_loss": 3.7058724737358784, + "tokens_seen": 922812416 + }, + { + "epoch": 0.28, + "learning_rate": 0.0005821143430290872, + "loss": 1.4019, + "theoretical_loss": 3.705847250677973, + "tokens_seen": 922877952 + }, + { + "epoch": 0.28, + "learning_rate": 0.000582098294884654, + "loss": 1.4142, + "theoretical_loss": 3.7058220299126425, + "tokens_seen": 922943488 + }, + { + "epoch": 0.28, + "learning_rate": 0.0005820822467402206, + "loss": 1.5328, + "theoretical_loss": 3.705796811439517, + "tokens_seen": 923009024 + }, + { + "epoch": 0.28, + "learning_rate": 0.0005820661985957874, + "loss": 1.4446, + "theoretical_loss": 3.7057715952582244, + "tokens_seen": 923074560 + }, + { + "epoch": 0.28, + "learning_rate": 0.000582050150451354, + "loss": 1.4934, + "theoretical_loss": 3.705746381368394, + "tokens_seen": 923140096 + }, + { + "epoch": 0.28, + "learning_rate": 0.0005820341023069208, + "loss": 1.5427, + "theoretical_loss": 3.705721169769655, + "tokens_seen": 923205632 + }, + { + "epoch": 0.28, + "learning_rate": 0.0005820180541624874, + "loss": 1.4086, + "theoretical_loss": 3.705695960461636, + "tokens_seen": 923271168 + }, + { + "epoch": 0.28, + "learning_rate": 0.0005820020060180542, + "loss": 1.4628, + "theoretical_loss": 3.7056707534439672, + "tokens_seen": 923336704 + }, + { + "epoch": 0.28, + "learning_rate": 0.0005819859578736208, + "loss": 1.4398, + "theoretical_loss": 3.7056455487162783, + "tokens_seen": 923402240 + }, + { + "epoch": 0.28, + "learning_rate": 0.0005819699097291876, + "loss": 1.3568, + "theoretical_loss": 3.705620346278198, + "tokens_seen": 923467776 + }, + { + "epoch": 0.28, + "learning_rate": 0.0005819538615847542, + "loss": 1.4623, + "theoretical_loss": 3.7055951461293546, + "tokens_seen": 923533312 + }, + { + "epoch": 0.28, + "learning_rate": 0.000581937813440321, + "loss": 1.4132, + "theoretical_loss": 3.70556994826938, + "tokens_seen": 923598848 + }, + { + "epoch": 0.28, + "learning_rate": 0.0005819217652958876, + "loss": 1.3043, + "theoretical_loss": 3.705544752697903, + "tokens_seen": 923664384 + }, + { + "epoch": 0.28, + "learning_rate": 0.0005819057171514544, + "loss": 1.3678, + "theoretical_loss": 3.705519559414553, + "tokens_seen": 923729920 + }, + { + "epoch": 0.28, + "learning_rate": 0.000581889669007021, + "loss": 1.3535, + "theoretical_loss": 3.7054943684189605, + "tokens_seen": 923795456 + }, + { + "epoch": 0.28, + "learning_rate": 0.0005818736208625878, + "loss": 1.3793, + "theoretical_loss": 3.705469179710755, + "tokens_seen": 923860992 + }, + { + "epoch": 0.28, + "learning_rate": 0.0005818575727181544, + "loss": 1.484, + "theoretical_loss": 3.705443993289567, + "tokens_seen": 923926528 + }, + { + "epoch": 0.28, + "learning_rate": 0.0005818415245737212, + "loss": 1.3348, + "theoretical_loss": 3.7054188091550264, + "tokens_seen": 923992064 + }, + { + "epoch": 0.28, + "objective/train/docs_used": 340768, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.564995527267456, + "objective/train/theoretical_loss": 3.705393627306764, + "objective/train/tokens_used": 944517600, + "theoretical_loss": 3.705393627306764, + "tokens_seen": 924057600 + }, + { + "epoch": 0.28, + "learning_rate": 0.0005818254764292878, + "loss": 1.4692, + "theoretical_loss": 3.705393627306764, + "tokens_seen": 924057600 + }, + { + "epoch": 0.28, + "learning_rate": 0.0005818094282848546, + "loss": 1.4338, + "theoretical_loss": 3.7053684477444087, + "tokens_seen": 924123136 + }, + { + "epoch": 0.28, + "learning_rate": 0.0005817933801404212, + "loss": 1.3818, + "theoretical_loss": 3.7053432704675924, + "tokens_seen": 924188672 + }, + { + "epoch": 0.28, + "learning_rate": 0.000581777331995988, + "loss": 1.348, + "theoretical_loss": 3.7053180954759446, + "tokens_seen": 924254208 + }, + { + "epoch": 0.28, + "learning_rate": 0.0005817612838515546, + "loss": 1.3669, + "theoretical_loss": 3.7052929227690967, + "tokens_seen": 924319744 + }, + { + "epoch": 0.28, + "learning_rate": 0.0005817452357071214, + "loss": 1.611, + "theoretical_loss": 3.7052677523466784, + "tokens_seen": 924385280 + }, + { + "epoch": 0.28, + "learning_rate": 0.000581729187562688, + "loss": 1.365, + "theoretical_loss": 3.7052425842083214, + "tokens_seen": 924450816 + }, + { + "epoch": 0.28, + "learning_rate": 0.0005817131394182548, + "loss": 1.3955, + "theoretical_loss": 3.705217418353656, + "tokens_seen": 924516352 + }, + { + "epoch": 0.28, + "learning_rate": 0.0005816970912738215, + "loss": 1.4103, + "theoretical_loss": 3.7051922547823137, + "tokens_seen": 924581888 + }, + { + "epoch": 0.28, + "learning_rate": 0.0005816810431293882, + "loss": 1.3324, + "theoretical_loss": 3.7051670934939245, + "tokens_seen": 924647424 + }, + { + "epoch": 0.28, + "learning_rate": 0.0005816649949849549, + "loss": 1.4032, + "theoretical_loss": 3.70514193448812, + "tokens_seen": 924712960 + }, + { + "epoch": 0.28, + "learning_rate": 0.0005816489468405216, + "loss": 1.353, + "theoretical_loss": 3.705116777764532, + "tokens_seen": 924778496 + }, + { + "epoch": 0.28, + "learning_rate": 0.0005816328986960883, + "loss": 1.4108, + "theoretical_loss": 3.7050916233227906, + "tokens_seen": 924844032 + }, + { + "epoch": 0.28, + "learning_rate": 0.000581616850551655, + "loss": 1.5595, + "theoretical_loss": 3.705066471162528, + "tokens_seen": 924909568 + }, + { + "epoch": 0.28, + "learning_rate": 0.0005816008024072217, + "loss": 1.4862, + "theoretical_loss": 3.7050413212833755, + "tokens_seen": 924975104 + }, + { + "epoch": 0.28, + "learning_rate": 0.0005815847542627884, + "loss": 1.3549, + "theoretical_loss": 3.7050161736849647, + "tokens_seen": 925040640 + }, + { + "epoch": 0.28, + "learning_rate": 0.0005815687061183551, + "loss": 1.3516, + "theoretical_loss": 3.7049910283669263, + "tokens_seen": 925106176 + }, + { + "epoch": 0.28, + "learning_rate": 0.0005815526579739218, + "loss": 1.4882, + "theoretical_loss": 3.7049658853288934, + "tokens_seen": 925171712 + }, + { + "epoch": 0.28, + "learning_rate": 0.0005815366098294885, + "loss": 1.4172, + "theoretical_loss": 3.7049407445704965, + "tokens_seen": 925237248 + }, + { + "epoch": 0.28, + "learning_rate": 0.0005815205616850552, + "loss": 1.3439, + "theoretical_loss": 3.704915606091369, + "tokens_seen": 925302784 + }, + { + "epoch": 0.28, + "learning_rate": 0.0005815045135406219, + "loss": 1.4518, + "theoretical_loss": 3.704890469891141, + "tokens_seen": 925368320 + }, + { + "epoch": 0.28, + "learning_rate": 0.0005814884653961886, + "loss": 1.4007, + "theoretical_loss": 3.704865335969446, + "tokens_seen": 925433856 + }, + { + "epoch": 0.28, + "learning_rate": 0.0005814724172517553, + "loss": 1.3852, + "theoretical_loss": 3.7048402043259148, + "tokens_seen": 925499392 + }, + { + "epoch": 0.28, + "learning_rate": 0.000581456369107322, + "loss": 1.4267, + "theoretical_loss": 3.704815074960181, + "tokens_seen": 925564928 + }, + { + "epoch": 0.28, + "learning_rate": 0.0005814403209628887, + "loss": 1.4671, + "theoretical_loss": 3.704789947871876, + "tokens_seen": 925630464 + }, + { + "epoch": 0.28, + "learning_rate": 0.0005814242728184554, + "loss": 1.3792, + "theoretical_loss": 3.7047648230606325, + "tokens_seen": 925696000 + }, + { + "epoch": 0.28, + "learning_rate": 0.0005814082246740221, + "loss": 1.4621, + "theoretical_loss": 3.7047397005260834, + "tokens_seen": 925761536 + }, + { + "epoch": 0.28, + "learning_rate": 0.0005813921765295888, + "loss": 1.3823, + "theoretical_loss": 3.70471458026786, + "tokens_seen": 925827072 + }, + { + "epoch": 0.28, + "learning_rate": 0.0005813761283851556, + "loss": 1.5733, + "theoretical_loss": 3.7046894622855957, + "tokens_seen": 925892608 + }, + { + "epoch": 0.28, + "learning_rate": 0.0005813600802407222, + "loss": 1.4568, + "theoretical_loss": 3.704664346578924, + "tokens_seen": 925958144 + }, + { + "epoch": 0.28, + "learning_rate": 0.000581344032096289, + "loss": 1.4131, + "theoretical_loss": 3.704639233147476, + "tokens_seen": 926023680 + }, + { + "epoch": 0.28, + "learning_rate": 0.0005813279839518556, + "loss": 1.4536, + "theoretical_loss": 3.704614121990886, + "tokens_seen": 926089216 + }, + { + "epoch": 0.28, + "learning_rate": 0.0005813119358074224, + "loss": 1.5176, + "theoretical_loss": 3.7045890131087864, + "tokens_seen": 926154752 + }, + { + "epoch": 0.28, + "learning_rate": 0.000581295887662989, + "loss": 1.3942, + "theoretical_loss": 3.70456390650081, + "tokens_seen": 926220288 + }, + { + "epoch": 0.28, + "learning_rate": 0.0005812798395185558, + "loss": 1.3632, + "theoretical_loss": 3.7045388021665904, + "tokens_seen": 926285824 + }, + { + "epoch": 0.28, + "learning_rate": 0.0005812637913741224, + "loss": 1.3644, + "theoretical_loss": 3.7045137001057613, + "tokens_seen": 926351360 + }, + { + "epoch": 0.28, + "learning_rate": 0.0005812477432296892, + "loss": 1.416, + "theoretical_loss": 3.704488600317955, + "tokens_seen": 926416896 + }, + { + "epoch": 0.28, + "learning_rate": 0.0005812316950852558, + "loss": 1.462, + "theoretical_loss": 3.7044635028028052, + "tokens_seen": 926482432 + }, + { + "epoch": 0.28, + "learning_rate": 0.0005812156469408226, + "loss": 1.4983, + "theoretical_loss": 3.7044384075599455, + "tokens_seen": 926547968 + }, + { + "epoch": 0.28, + "learning_rate": 0.0005811995987963892, + "loss": 1.4284, + "theoretical_loss": 3.7044133145890097, + "tokens_seen": 926613504 + }, + { + "epoch": 0.28, + "learning_rate": 0.000581183550651956, + "loss": 1.3028, + "theoretical_loss": 3.7043882238896306, + "tokens_seen": 926679040 + }, + { + "epoch": 0.28, + "learning_rate": 0.0005811675025075226, + "loss": 1.3359, + "theoretical_loss": 3.704363135461443, + "tokens_seen": 926744576 + }, + { + "epoch": 0.28, + "learning_rate": 0.0005811514543630894, + "loss": 1.388, + "theoretical_loss": 3.7043380493040807, + "tokens_seen": 926810112 + }, + { + "epoch": 0.28, + "learning_rate": 0.000581135406218656, + "loss": 1.4446, + "theoretical_loss": 3.7043129654171767, + "tokens_seen": 926875648 + }, + { + "epoch": 0.28, + "learning_rate": 0.0005811193580742228, + "loss": 1.3571, + "theoretical_loss": 3.704287883800365, + "tokens_seen": 926941184 + }, + { + "epoch": 0.28, + "learning_rate": 0.0005811033099297894, + "loss": 1.4743, + "theoretical_loss": 3.7042628044532813, + "tokens_seen": 927006720 + }, + { + "epoch": 0.28, + "learning_rate": 0.0005810872617853561, + "loss": 1.3616, + "theoretical_loss": 3.704237727375558, + "tokens_seen": 927072256 + }, + { + "epoch": 0.28, + "learning_rate": 0.0005810712136409228, + "loss": 1.3849, + "theoretical_loss": 3.70421265256683, + "tokens_seen": 927137792 + }, + { + "epoch": 0.28, + "learning_rate": 0.0005810551654964895, + "loss": 1.5054, + "theoretical_loss": 3.7041875800267317, + "tokens_seen": 927203328 + }, + { + "epoch": 0.28, + "learning_rate": 0.0005810391173520562, + "loss": 1.3946, + "theoretical_loss": 3.704162509754898, + "tokens_seen": 927268864 + }, + { + "epoch": 0.28, + "objective/train/docs_used": 341815, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.332563877105713, + "objective/train/theoretical_loss": 3.7041374417509623, + "objective/train/tokens_used": 947794400, + "theoretical_loss": 3.7041374417509623, + "tokens_seen": 927334400 + }, + { + "epoch": 0.28, + "learning_rate": 0.0005810230692076229, + "loss": 1.3695, + "theoretical_loss": 3.7041374417509623, + "tokens_seen": 927334400 + }, + { + "epoch": 0.28, + "learning_rate": 0.0005810070210631896, + "loss": 1.3908, + "theoretical_loss": 3.7041123760145593, + "tokens_seen": 927399936 + }, + { + "epoch": 0.28, + "learning_rate": 0.0005809909729187563, + "loss": 1.4642, + "theoretical_loss": 3.704087312545325, + "tokens_seen": 927465472 + }, + { + "epoch": 0.28, + "learning_rate": 0.000580974924774323, + "loss": 1.492, + "theoretical_loss": 3.7040622513428927, + "tokens_seen": 927531008 + }, + { + "epoch": 0.28, + "learning_rate": 0.0005809588766298897, + "loss": 1.2834, + "theoretical_loss": 3.704037192406898, + "tokens_seen": 927596544 + }, + { + "epoch": 0.28, + "learning_rate": 0.0005809428284854563, + "loss": 1.401, + "theoretical_loss": 3.704012135736975, + "tokens_seen": 927662080 + }, + { + "epoch": 0.28, + "learning_rate": 0.0005809267803410231, + "loss": 1.4345, + "theoretical_loss": 3.7039870813327602, + "tokens_seen": 927727616 + }, + { + "epoch": 0.28, + "learning_rate": 0.0005809107321965897, + "loss": 1.3211, + "theoretical_loss": 3.703962029193888, + "tokens_seen": 927793152 + }, + { + "epoch": 0.28, + "learning_rate": 0.0005808946840521565, + "loss": 1.4526, + "theoretical_loss": 3.7039369793199928, + "tokens_seen": 927858688 + }, + { + "epoch": 0.28, + "learning_rate": 0.0005808786359077231, + "loss": 1.3518, + "theoretical_loss": 3.7039119317107105, + "tokens_seen": 927924224 + }, + { + "epoch": 0.28, + "learning_rate": 0.0005808625877632899, + "loss": 1.3275, + "theoretical_loss": 3.7038868863656766, + "tokens_seen": 927989760 + }, + { + "epoch": 0.28, + "learning_rate": 0.0005808465396188565, + "loss": 1.4936, + "theoretical_loss": 3.7038618432845265, + "tokens_seen": 928055296 + }, + { + "epoch": 0.28, + "learning_rate": 0.0005808304914744233, + "loss": 1.3974, + "theoretical_loss": 3.7038368024668955, + "tokens_seen": 928120832 + }, + { + "epoch": 0.28, + "learning_rate": 0.0005808144433299899, + "loss": 1.4998, + "theoretical_loss": 3.703811763912419, + "tokens_seen": 928186368 + }, + { + "epoch": 0.28, + "learning_rate": 0.0005807983951855567, + "loss": 1.3937, + "theoretical_loss": 3.7037867276207335, + "tokens_seen": 928251904 + }, + { + "epoch": 0.28, + "learning_rate": 0.0005807823470411233, + "loss": 1.5611, + "theoretical_loss": 3.703761693591474, + "tokens_seen": 928317440 + }, + { + "epoch": 0.28, + "learning_rate": 0.0005807662988966901, + "loss": 1.4677, + "theoretical_loss": 3.7037366618242764, + "tokens_seen": 928382976 + }, + { + "epoch": 0.28, + "learning_rate": 0.0005807502507522567, + "loss": 1.4314, + "theoretical_loss": 3.703711632318777, + "tokens_seen": 928448512 + }, + { + "epoch": 0.28, + "learning_rate": 0.0005807342026078235, + "loss": 1.4618, + "theoretical_loss": 3.703686605074612, + "tokens_seen": 928514048 + }, + { + "epoch": 0.28, + "learning_rate": 0.0005807181544633901, + "loss": 1.4021, + "theoretical_loss": 3.7036615800914165, + "tokens_seen": 928579584 + }, + { + "epoch": 0.28, + "learning_rate": 0.0005807021063189569, + "loss": 1.4138, + "theoretical_loss": 3.703636557368828, + "tokens_seen": 928645120 + }, + { + "epoch": 0.28, + "learning_rate": 0.0005806860581745235, + "loss": 1.5398, + "theoretical_loss": 3.7036115369064815, + "tokens_seen": 928710656 + }, + { + "epoch": 0.28, + "learning_rate": 0.0005806700100300903, + "loss": 1.5058, + "theoretical_loss": 3.7035865187040145, + "tokens_seen": 928776192 + }, + { + "epoch": 0.28, + "learning_rate": 0.0005806539618856569, + "loss": 1.4333, + "theoretical_loss": 3.703561502761063, + "tokens_seen": 928841728 + }, + { + "epoch": 0.28, + "learning_rate": 0.0005806379137412237, + "loss": 1.4248, + "theoretical_loss": 3.7035364890772637, + "tokens_seen": 928907264 + }, + { + "epoch": 0.28, + "learning_rate": 0.0005806218655967903, + "loss": 1.4908, + "theoretical_loss": 3.7035114776522526, + "tokens_seen": 928972800 + }, + { + "epoch": 0.28, + "learning_rate": 0.0005806058174523571, + "loss": 1.3206, + "theoretical_loss": 3.703486468485667, + "tokens_seen": 929038336 + }, + { + "epoch": 0.28, + "learning_rate": 0.0005805897693079237, + "loss": 1.584, + "theoretical_loss": 3.7034614615771435, + "tokens_seen": 929103872 + }, + { + "epoch": 0.28, + "learning_rate": 0.0005805737211634905, + "loss": 1.5763, + "theoretical_loss": 3.7034364569263185, + "tokens_seen": 929169408 + }, + { + "epoch": 0.28, + "learning_rate": 0.0005805576730190571, + "loss": 1.5195, + "theoretical_loss": 3.70341145453283, + "tokens_seen": 929234944 + }, + { + "epoch": 0.28, + "learning_rate": 0.0005805416248746239, + "loss": 1.5089, + "theoretical_loss": 3.703386454396314, + "tokens_seen": 929300480 + }, + { + "epoch": 0.28, + "learning_rate": 0.0005805255767301905, + "loss": 1.4633, + "theoretical_loss": 3.703361456516408, + "tokens_seen": 929366016 + }, + { + "epoch": 0.28, + "learning_rate": 0.0005805095285857573, + "loss": 1.4291, + "theoretical_loss": 3.7033364608927495, + "tokens_seen": 929431552 + }, + { + "epoch": 0.28, + "learning_rate": 0.0005804934804413239, + "loss": 1.3761, + "theoretical_loss": 3.7033114675249754, + "tokens_seen": 929497088 + }, + { + "epoch": 0.28, + "learning_rate": 0.0005804774322968907, + "loss": 1.4035, + "theoretical_loss": 3.703286476412724, + "tokens_seen": 929562624 + }, + { + "epoch": 0.28, + "learning_rate": 0.0005804613841524573, + "loss": 1.3128, + "theoretical_loss": 3.703261487555631, + "tokens_seen": 929628160 + }, + { + "epoch": 0.28, + "learning_rate": 0.0005804453360080241, + "loss": 1.4005, + "theoretical_loss": 3.703236500953335, + "tokens_seen": 929693696 + }, + { + "epoch": 0.28, + "learning_rate": 0.0005804292878635907, + "loss": 1.4399, + "theoretical_loss": 3.7032115166054735, + "tokens_seen": 929759232 + }, + { + "epoch": 0.28, + "learning_rate": 0.0005804132397191575, + "loss": 1.4861, + "theoretical_loss": 3.7031865345116843, + "tokens_seen": 929824768 + }, + { + "epoch": 0.28, + "learning_rate": 0.0005803971915747241, + "loss": 1.2976, + "theoretical_loss": 3.7031615546716052, + "tokens_seen": 929890304 + }, + { + "epoch": 0.28, + "learning_rate": 0.0005803811434302909, + "loss": 1.4329, + "theoretical_loss": 3.7031365770848743, + "tokens_seen": 929955840 + }, + { + "epoch": 0.28, + "learning_rate": 0.0005803650952858576, + "loss": 1.4094, + "theoretical_loss": 3.7031116017511287, + "tokens_seen": 930021376 + }, + { + "epoch": 0.28, + "learning_rate": 0.0005803490471414243, + "loss": 1.3553, + "theoretical_loss": 3.7030866286700075, + "tokens_seen": 930086912 + }, + { + "epoch": 0.28, + "learning_rate": 0.000580332998996991, + "loss": 1.3463, + "theoretical_loss": 3.703061657841148, + "tokens_seen": 930152448 + }, + { + "epoch": 0.28, + "learning_rate": 0.0005803169508525577, + "loss": 1.4132, + "theoretical_loss": 3.703036689264189, + "tokens_seen": 930217984 + }, + { + "epoch": 0.28, + "learning_rate": 0.0005803009027081244, + "loss": 1.4523, + "theoretical_loss": 3.703011722938768, + "tokens_seen": 930283520 + }, + { + "epoch": 0.28, + "learning_rate": 0.0005802848545636911, + "loss": 1.5696, + "theoretical_loss": 3.702986758864524, + "tokens_seen": 930349056 + }, + { + "epoch": 0.28, + "learning_rate": 0.0005802688064192578, + "loss": 1.4629, + "theoretical_loss": 3.7029617970410955, + "tokens_seen": 930414592 + }, + { + "epoch": 0.28, + "learning_rate": 0.0005802527582748245, + "loss": 1.4263, + "theoretical_loss": 3.702936837468121, + "tokens_seen": 930480128 + }, + { + "epoch": 0.28, + "learning_rate": 0.0005802367101303912, + "loss": 1.3223, + "theoretical_loss": 3.7029118801452388, + "tokens_seen": 930545664 + }, + { + "epoch": 0.28, + "objective/train/docs_used": 343003, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.5371347665786743, + "objective/train/theoretical_loss": 3.7028869250720877, + "objective/train/tokens_used": 951071200, + "theoretical_loss": 3.7028869250720877, + "tokens_seen": 930611200 + }, + { + "epoch": 0.28, + "learning_rate": 0.0005802206619859579, + "loss": 1.5421, + "theoretical_loss": 3.7028869250720877, + "tokens_seen": 930611200 + }, + { + "epoch": 0.28, + "learning_rate": 0.0005802046138415246, + "loss": 1.3555, + "theoretical_loss": 3.702861972248307, + "tokens_seen": 930676736 + }, + { + "epoch": 0.28, + "learning_rate": 0.0005801885656970913, + "loss": 1.4139, + "theoretical_loss": 3.702837021673535, + "tokens_seen": 930742272 + }, + { + "epoch": 0.28, + "learning_rate": 0.0005801725175526581, + "loss": 1.4407, + "theoretical_loss": 3.7028120733474106, + "tokens_seen": 930807808 + }, + { + "epoch": 0.28, + "learning_rate": 0.0005801564694082247, + "loss": 1.4256, + "theoretical_loss": 3.7027871272695734, + "tokens_seen": 930873344 + }, + { + "epoch": 0.28, + "learning_rate": 0.0005801404212637915, + "loss": 1.4597, + "theoretical_loss": 3.7027621834396625, + "tokens_seen": 930938880 + }, + { + "epoch": 0.28, + "learning_rate": 0.0005801243731193581, + "loss": 1.5036, + "theoretical_loss": 3.7027372418573163, + "tokens_seen": 931004416 + }, + { + "epoch": 0.28, + "learning_rate": 0.0005801083249749249, + "loss": 1.5695, + "theoretical_loss": 3.7027123025221744, + "tokens_seen": 931069952 + }, + { + "epoch": 0.28, + "learning_rate": 0.0005800922768304915, + "loss": 1.3546, + "theoretical_loss": 3.702687365433877, + "tokens_seen": 931135488 + }, + { + "epoch": 0.28, + "learning_rate": 0.0005800762286860583, + "loss": 1.452, + "theoretical_loss": 3.7026624305920626, + "tokens_seen": 931201024 + }, + { + "epoch": 0.28, + "learning_rate": 0.0005800601805416249, + "loss": 1.4469, + "theoretical_loss": 3.702637497996371, + "tokens_seen": 931266560 + }, + { + "epoch": 0.28, + "learning_rate": 0.0005800441323971917, + "loss": 1.4336, + "theoretical_loss": 3.702612567646442, + "tokens_seen": 931332096 + }, + { + "epoch": 0.28, + "learning_rate": 0.0005800280842527583, + "loss": 1.5365, + "theoretical_loss": 3.7025876395419153, + "tokens_seen": 931397632 + }, + { + "epoch": 0.28, + "learning_rate": 0.0005800120361083251, + "loss": 1.5947, + "theoretical_loss": 3.702562713682431, + "tokens_seen": 931463168 + }, + { + "epoch": 0.28, + "learning_rate": 0.0005799959879638917, + "loss": 1.3856, + "theoretical_loss": 3.7025377900676277, + "tokens_seen": 931528704 + }, + { + "epoch": 0.28, + "learning_rate": 0.0005799799398194585, + "loss": 1.3885, + "theoretical_loss": 3.7025128686971467, + "tokens_seen": 931594240 + }, + { + "epoch": 0.28, + "learning_rate": 0.0005799638916750251, + "loss": 1.3816, + "theoretical_loss": 3.7024879495706275, + "tokens_seen": 931659776 + }, + { + "epoch": 0.28, + "learning_rate": 0.0005799478435305919, + "loss": 1.4397, + "theoretical_loss": 3.70246303268771, + "tokens_seen": 931725312 + }, + { + "epoch": 0.28, + "learning_rate": 0.0005799317953861585, + "loss": 1.4222, + "theoretical_loss": 3.7024381180480352, + "tokens_seen": 931790848 + }, + { + "epoch": 0.28, + "learning_rate": 0.0005799157472417253, + "loss": 1.4095, + "theoretical_loss": 3.702413205651243, + "tokens_seen": 931856384 + }, + { + "epoch": 0.28, + "learning_rate": 0.0005798996990972919, + "loss": 1.4279, + "theoretical_loss": 3.7023882954969736, + "tokens_seen": 931921920 + }, + { + "epoch": 0.28, + "learning_rate": 0.0005798836509528587, + "loss": 1.3791, + "theoretical_loss": 3.7023633875848674, + "tokens_seen": 931987456 + }, + { + "epoch": 0.28, + "learning_rate": 0.0005798676028084253, + "loss": 1.5307, + "theoretical_loss": 3.7023384819145644, + "tokens_seen": 932052992 + }, + { + "epoch": 0.28, + "learning_rate": 0.0005798515546639921, + "loss": 1.4637, + "theoretical_loss": 3.7023135784857066, + "tokens_seen": 932118528 + }, + { + "epoch": 0.28, + "learning_rate": 0.0005798355065195587, + "loss": 1.4874, + "theoretical_loss": 3.7022886772979344, + "tokens_seen": 932184064 + }, + { + "epoch": 0.28, + "learning_rate": 0.0005798194583751255, + "loss": 1.391, + "theoretical_loss": 3.702263778350887, + "tokens_seen": 932249600 + }, + { + "epoch": 0.28, + "learning_rate": 0.0005798034102306921, + "loss": 1.354, + "theoretical_loss": 3.702238881644208, + "tokens_seen": 932315136 + }, + { + "epoch": 0.28, + "learning_rate": 0.0005797873620862589, + "loss": 1.4366, + "theoretical_loss": 3.7022139871775357, + "tokens_seen": 932380672 + }, + { + "epoch": 0.28, + "learning_rate": 0.0005797713139418255, + "loss": 1.4402, + "theoretical_loss": 3.702189094950512, + "tokens_seen": 932446208 + }, + { + "epoch": 0.28, + "learning_rate": 0.0005797552657973923, + "loss": 1.3873, + "theoretical_loss": 3.702164204962779, + "tokens_seen": 932511744 + }, + { + "epoch": 0.28, + "learning_rate": 0.0005797392176529589, + "loss": 1.4076, + "theoretical_loss": 3.7021393172139767, + "tokens_seen": 932577280 + }, + { + "epoch": 0.28, + "learning_rate": 0.0005797231695085257, + "loss": 1.482, + "theoretical_loss": 3.7021144317037473, + "tokens_seen": 932642816 + }, + { + "epoch": 0.28, + "learning_rate": 0.0005797071213640923, + "loss": 1.4439, + "theoretical_loss": 3.702089548431731, + "tokens_seen": 932708352 + }, + { + "epoch": 0.28, + "learning_rate": 0.000579691073219659, + "loss": 1.4251, + "theoretical_loss": 3.7020646673975706, + "tokens_seen": 932773888 + }, + { + "epoch": 0.28, + "learning_rate": 0.0005796750250752257, + "loss": 1.4195, + "theoretical_loss": 3.7020397886009064, + "tokens_seen": 932839424 + }, + { + "epoch": 0.28, + "learning_rate": 0.0005796589769307924, + "loss": 1.3214, + "theoretical_loss": 3.7020149120413812, + "tokens_seen": 932904960 + }, + { + "epoch": 0.28, + "learning_rate": 0.0005796429287863591, + "loss": 1.3864, + "theoretical_loss": 3.7019900377186357, + "tokens_seen": 932970496 + }, + { + "epoch": 0.28, + "learning_rate": 0.0005796268806419258, + "loss": 1.5516, + "theoretical_loss": 3.7019651656323127, + "tokens_seen": 933036032 + }, + { + "epoch": 0.28, + "learning_rate": 0.0005796108324974925, + "loss": 1.6106, + "theoretical_loss": 3.7019402957820526, + "tokens_seen": 933101568 + }, + { + "epoch": 0.28, + "learning_rate": 0.0005795947843530592, + "loss": 1.4108, + "theoretical_loss": 3.7019154281674984, + "tokens_seen": 933167104 + }, + { + "epoch": 0.28, + "learning_rate": 0.0005795787362086259, + "loss": 1.583, + "theoretical_loss": 3.701890562788292, + "tokens_seen": 933232640 + }, + { + "epoch": 0.28, + "learning_rate": 0.0005795626880641926, + "loss": 1.3006, + "theoretical_loss": 3.7018656996440757, + "tokens_seen": 933298176 + }, + { + "epoch": 0.28, + "learning_rate": 0.0005795466399197592, + "loss": 1.423, + "theoretical_loss": 3.701840838734491, + "tokens_seen": 933363712 + }, + { + "epoch": 0.28, + "learning_rate": 0.000579530591775326, + "loss": 1.4116, + "theoretical_loss": 3.701815980059181, + "tokens_seen": 933429248 + }, + { + "epoch": 0.28, + "learning_rate": 0.0005795145436308926, + "loss": 1.3468, + "theoretical_loss": 3.701791123617787, + "tokens_seen": 933494784 + }, + { + "epoch": 0.28, + "learning_rate": 0.0005794984954864594, + "loss": 1.4517, + "theoretical_loss": 3.7017662694099527, + "tokens_seen": 933560320 + }, + { + "epoch": 0.28, + "learning_rate": 0.000579482447342026, + "loss": 1.4043, + "theoretical_loss": 3.70174141743532, + "tokens_seen": 933625856 + }, + { + "epoch": 0.28, + "learning_rate": 0.0005794663991975928, + "loss": 1.4002, + "theoretical_loss": 3.7017165676935315, + "tokens_seen": 933691392 + }, + { + "epoch": 0.28, + "learning_rate": 0.0005794503510531594, + "loss": 1.356, + "theoretical_loss": 3.701691720184229, + "tokens_seen": 933756928 + }, + { + "epoch": 0.28, + "learning_rate": 0.0005794343029087262, + "loss": 1.4704, + "theoretical_loss": 3.7016668749070574, + "tokens_seen": 933822464 + }, + { + "epoch": 0.28, + "objective/train/docs_used": 344002, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.5559661388397217, + "objective/train/theoretical_loss": 3.7016420318616574, + "objective/train/tokens_used": 954348000, + "theoretical_loss": 3.7016420318616574, + "tokens_seen": 933888000 + }, + { + "epoch": 0.28, + "learning_rate": 0.0005794182547642928, + "loss": 1.5337, + "theoretical_loss": 3.7016420318616574, + "tokens_seen": 933888000 + }, + { + "epoch": 0.28, + "learning_rate": 0.0005794022066198596, + "loss": 1.4555, + "theoretical_loss": 3.7016171910476734, + "tokens_seen": 933953536 + }, + { + "epoch": 0.28, + "learning_rate": 0.0005793861584754262, + "loss": 1.4741, + "theoretical_loss": 3.701592352464748, + "tokens_seen": 934019072 + }, + { + "epoch": 0.28, + "learning_rate": 0.000579370110330993, + "loss": 1.4652, + "theoretical_loss": 3.701567516112523, + "tokens_seen": 934084608 + }, + { + "epoch": 0.28, + "learning_rate": 0.0005793540621865596, + "loss": 1.486, + "theoretical_loss": 3.701542681990644, + "tokens_seen": 934150144 + }, + { + "epoch": 0.28, + "learning_rate": 0.0005793380140421264, + "loss": 1.5221, + "theoretical_loss": 3.7015178500987522, + "tokens_seen": 934215680 + }, + { + "epoch": 0.28, + "learning_rate": 0.000579321965897693, + "loss": 1.4428, + "theoretical_loss": 3.701493020436492, + "tokens_seen": 934281216 + }, + { + "epoch": 0.28, + "learning_rate": 0.0005793059177532598, + "loss": 1.5324, + "theoretical_loss": 3.7014681930035067, + "tokens_seen": 934346752 + }, + { + "epoch": 0.28, + "learning_rate": 0.0005792898696088264, + "loss": 1.4317, + "theoretical_loss": 3.7014433677994396, + "tokens_seen": 934412288 + }, + { + "epoch": 0.28, + "learning_rate": 0.0005792738214643932, + "loss": 1.4219, + "theoretical_loss": 3.7014185448239343, + "tokens_seen": 934477824 + }, + { + "epoch": 0.28, + "learning_rate": 0.0005792577733199598, + "loss": 1.5089, + "theoretical_loss": 3.7013937240766346, + "tokens_seen": 934543360 + }, + { + "epoch": 0.28, + "learning_rate": 0.0005792417251755266, + "loss": 1.4693, + "theoretical_loss": 3.7013689055571843, + "tokens_seen": 934608896 + }, + { + "epoch": 0.28, + "learning_rate": 0.0005792256770310932, + "loss": 1.4741, + "theoretical_loss": 3.701344089265227, + "tokens_seen": 934674432 + }, + { + "epoch": 0.28, + "learning_rate": 0.00057920962888666, + "loss": 1.3411, + "theoretical_loss": 3.7013192752004063, + "tokens_seen": 934739968 + }, + { + "epoch": 0.28, + "learning_rate": 0.0005791935807422266, + "loss": 1.4569, + "theoretical_loss": 3.701294463362367, + "tokens_seen": 934805504 + }, + { + "epoch": 0.28, + "learning_rate": 0.0005791775325977934, + "loss": 1.4265, + "theoretical_loss": 3.701269653750753, + "tokens_seen": 934871040 + }, + { + "epoch": 0.28, + "learning_rate": 0.00057916148445336, + "loss": 1.4804, + "theoretical_loss": 3.701244846365208, + "tokens_seen": 934936576 + }, + { + "epoch": 0.28, + "learning_rate": 0.0005791454363089268, + "loss": 1.4749, + "theoretical_loss": 3.701220041205377, + "tokens_seen": 935002112 + }, + { + "epoch": 0.28, + "learning_rate": 0.0005791293881644934, + "loss": 1.4445, + "theoretical_loss": 3.701195238270903, + "tokens_seen": 935067648 + }, + { + "epoch": 0.28, + "learning_rate": 0.0005791133400200602, + "loss": 1.3483, + "theoretical_loss": 3.701170437561432, + "tokens_seen": 935133184 + }, + { + "epoch": 0.28, + "learning_rate": 0.0005790972918756268, + "loss": 1.4663, + "theoretical_loss": 3.701145639076607, + "tokens_seen": 935198720 + }, + { + "epoch": 0.28, + "learning_rate": 0.0005790812437311936, + "loss": 1.3316, + "theoretical_loss": 3.7011208428160742, + "tokens_seen": 935264256 + }, + { + "epoch": 0.28, + "learning_rate": 0.0005790651955867602, + "loss": 1.402, + "theoretical_loss": 3.701096048779477, + "tokens_seen": 935329792 + }, + { + "epoch": 0.28, + "learning_rate": 0.000579049147442327, + "loss": 1.439, + "theoretical_loss": 3.7010712569664603, + "tokens_seen": 935395328 + }, + { + "epoch": 0.28, + "learning_rate": 0.0005790330992978936, + "loss": 1.3764, + "theoretical_loss": 3.701046467376669, + "tokens_seen": 935460864 + }, + { + "epoch": 0.28, + "learning_rate": 0.0005790170511534604, + "loss": 1.5906, + "theoretical_loss": 3.701021680009749, + "tokens_seen": 935526400 + }, + { + "epoch": 0.28, + "learning_rate": 0.000579001003009027, + "loss": 1.4108, + "theoretical_loss": 3.7009968948653436, + "tokens_seen": 935591936 + }, + { + "epoch": 0.28, + "learning_rate": 0.0005789849548645938, + "loss": 1.3809, + "theoretical_loss": 3.7009721119430985, + "tokens_seen": 935657472 + }, + { + "epoch": 0.28, + "learning_rate": 0.0005789689067201606, + "loss": 1.4093, + "theoretical_loss": 3.7009473312426593, + "tokens_seen": 935723008 + }, + { + "epoch": 0.28, + "learning_rate": 0.0005789528585757272, + "loss": 1.4363, + "theoretical_loss": 3.700922552763671, + "tokens_seen": 935788544 + }, + { + "epoch": 0.28, + "learning_rate": 0.000578936810431294, + "loss": 1.4001, + "theoretical_loss": 3.700897776505779, + "tokens_seen": 935854080 + }, + { + "epoch": 0.28, + "learning_rate": 0.0005789207622868606, + "loss": 1.424, + "theoretical_loss": 3.7008730024686285, + "tokens_seen": 935919616 + }, + { + "epoch": 0.28, + "learning_rate": 0.0005789047141424274, + "loss": 1.4723, + "theoretical_loss": 3.7008482306518644, + "tokens_seen": 935985152 + }, + { + "epoch": 0.28, + "learning_rate": 0.000578888665997994, + "loss": 1.4251, + "theoretical_loss": 3.7008234610551334, + "tokens_seen": 936050688 + }, + { + "epoch": 0.28, + "learning_rate": 0.0005788726178535608, + "loss": 1.5216, + "theoretical_loss": 3.70079869367808, + "tokens_seen": 936116224 + }, + { + "epoch": 0.28, + "learning_rate": 0.0005788565697091274, + "loss": 1.4619, + "theoretical_loss": 3.700773928520351, + "tokens_seen": 936181760 + }, + { + "epoch": 0.28, + "learning_rate": 0.0005788405215646942, + "loss": 1.3768, + "theoretical_loss": 3.7007491655815916, + "tokens_seen": 936247296 + }, + { + "epoch": 0.28, + "learning_rate": 0.0005788244734202608, + "loss": 1.373, + "theoretical_loss": 3.7007244048614476, + "tokens_seen": 936312832 + }, + { + "epoch": 0.28, + "learning_rate": 0.0005788084252758276, + "loss": 1.4625, + "theoretical_loss": 3.700699646359565, + "tokens_seen": 936378368 + }, + { + "epoch": 0.28, + "learning_rate": 0.0005787923771313942, + "loss": 1.4167, + "theoretical_loss": 3.7006748900755904, + "tokens_seen": 936443904 + }, + { + "epoch": 0.28, + "learning_rate": 0.000578776328986961, + "loss": 1.4115, + "theoretical_loss": 3.700650136009169, + "tokens_seen": 936509440 + }, + { + "epoch": 0.28, + "learning_rate": 0.0005787602808425276, + "loss": 1.4464, + "theoretical_loss": 3.700625384159948, + "tokens_seen": 936574976 + }, + { + "epoch": 0.28, + "learning_rate": 0.0005787442326980944, + "loss": 1.4863, + "theoretical_loss": 3.7006006345275724, + "tokens_seen": 936640512 + }, + { + "epoch": 0.28, + "learning_rate": 0.000578728184553661, + "loss": 1.4163, + "theoretical_loss": 3.70057588711169, + "tokens_seen": 936706048 + }, + { + "epoch": 0.28, + "learning_rate": 0.0005787121364092278, + "loss": 1.4628, + "theoretical_loss": 3.700551141911946, + "tokens_seen": 936771584 + }, + { + "epoch": 0.28, + "learning_rate": 0.0005786960882647944, + "loss": 1.4187, + "theoretical_loss": 3.700526398927988, + "tokens_seen": 936837120 + }, + { + "epoch": 0.28, + "learning_rate": 0.0005786800401203612, + "loss": 1.3754, + "theoretical_loss": 3.700501658159462, + "tokens_seen": 936902656 + }, + { + "epoch": 0.28, + "learning_rate": 0.0005786639919759278, + "loss": 1.3711, + "theoretical_loss": 3.7004769196060145, + "tokens_seen": 936968192 + }, + { + "epoch": 0.28, + "learning_rate": 0.0005786479438314946, + "loss": 1.3765, + "theoretical_loss": 3.7004521832672923, + "tokens_seen": 937033728 + }, + { + "epoch": 0.28, + "learning_rate": 0.0005786318956870612, + "loss": 1.3712, + "theoretical_loss": 3.700427449142943, + "tokens_seen": 937099264 + }, + { + "epoch": 0.28, + "objective/train/docs_used": 344763, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.3824193477630615, + "objective/train/theoretical_loss": 3.700402717232613, + "objective/train/tokens_used": 957624800, + "theoretical_loss": 3.700402717232613, + "tokens_seen": 937164800 + }, + { + "epoch": 0.28, + "learning_rate": 0.000578615847542628, + "loss": 1.3969, + "theoretical_loss": 3.700402717232613, + "tokens_seen": 937164800 + }, + { + "epoch": 0.28, + "learning_rate": 0.0005785997993981946, + "loss": 1.2871, + "theoretical_loss": 3.700377987535949, + "tokens_seen": 937230336 + }, + { + "epoch": 0.28, + "learning_rate": 0.0005785837512537614, + "loss": 1.4984, + "theoretical_loss": 3.700353260052599, + "tokens_seen": 937295872 + }, + { + "epoch": 0.28, + "learning_rate": 0.000578567703109328, + "loss": 1.4611, + "theoretical_loss": 3.7003285347822086, + "tokens_seen": 937361408 + }, + { + "epoch": 0.28, + "learning_rate": 0.0005785516549648948, + "loss": 1.3965, + "theoretical_loss": 3.7003038117244267, + "tokens_seen": 937426944 + }, + { + "epoch": 0.28, + "learning_rate": 0.0005785356068204614, + "loss": 1.4704, + "theoretical_loss": 3.7002790908789, + "tokens_seen": 937492480 + }, + { + "epoch": 0.28, + "learning_rate": 0.0005785195586760282, + "loss": 1.354, + "theoretical_loss": 3.7002543722452756, + "tokens_seen": 937558016 + }, + { + "epoch": 0.28, + "learning_rate": 0.0005785035105315948, + "loss": 1.3945, + "theoretical_loss": 3.7002296558232013, + "tokens_seen": 937623552 + }, + { + "epoch": 0.28, + "learning_rate": 0.0005784874623871616, + "loss": 1.4515, + "theoretical_loss": 3.7002049416123253, + "tokens_seen": 937689088 + }, + { + "epoch": 0.28, + "learning_rate": 0.0005784714142427282, + "loss": 1.5025, + "theoretical_loss": 3.700180229612294, + "tokens_seen": 937754624 + }, + { + "epoch": 0.28, + "learning_rate": 0.000578455366098295, + "loss": 1.4593, + "theoretical_loss": 3.700155519822756, + "tokens_seen": 937820160 + }, + { + "epoch": 0.28, + "learning_rate": 0.0005784393179538616, + "loss": 1.4016, + "theoretical_loss": 3.700130812243359, + "tokens_seen": 937885696 + }, + { + "epoch": 0.28, + "learning_rate": 0.0005784232698094284, + "loss": 1.469, + "theoretical_loss": 3.7001061068737506, + "tokens_seen": 937951232 + }, + { + "epoch": 0.28, + "learning_rate": 0.000578407221664995, + "loss": 1.4781, + "theoretical_loss": 3.700081403713579, + "tokens_seen": 938016768 + }, + { + "epoch": 0.28, + "learning_rate": 0.0005783911735205618, + "loss": 1.4021, + "theoretical_loss": 3.700056702762492, + "tokens_seen": 938082304 + }, + { + "epoch": 0.28, + "learning_rate": 0.0005783751253761284, + "loss": 1.4076, + "theoretical_loss": 3.700032004020138, + "tokens_seen": 938147840 + }, + { + "epoch": 0.28, + "learning_rate": 0.0005783590772316952, + "loss": 1.5299, + "theoretical_loss": 3.7000073074861657, + "tokens_seen": 938213376 + }, + { + "epoch": 0.28, + "learning_rate": 0.0005783430290872618, + "loss": 1.4949, + "theoretical_loss": 3.699982613160223, + "tokens_seen": 938278912 + }, + { + "epoch": 0.28, + "learning_rate": 0.0005783269809428286, + "loss": 1.4279, + "theoretical_loss": 3.6999579210419578, + "tokens_seen": 938344448 + }, + { + "epoch": 0.28, + "learning_rate": 0.0005783109327983952, + "loss": 1.4958, + "theoretical_loss": 3.699933231131019, + "tokens_seen": 938409984 + }, + { + "epoch": 0.28, + "learning_rate": 0.000578294884653962, + "loss": 1.5228, + "theoretical_loss": 3.6999085434270547, + "tokens_seen": 938475520 + }, + { + "epoch": 0.28, + "learning_rate": 0.0005782788365095286, + "loss": 1.3308, + "theoretical_loss": 3.6998838579297146, + "tokens_seen": 938541056 + }, + { + "epoch": 0.28, + "learning_rate": 0.0005782627883650953, + "loss": 1.3576, + "theoretical_loss": 3.699859174638646, + "tokens_seen": 938606592 + }, + { + "epoch": 0.28, + "learning_rate": 0.000578246740220662, + "loss": 1.3807, + "theoretical_loss": 3.6998344935534995, + "tokens_seen": 938672128 + }, + { + "epoch": 0.28, + "learning_rate": 0.0005782306920762287, + "loss": 1.4201, + "theoretical_loss": 3.699809814673922, + "tokens_seen": 938737664 + }, + { + "epoch": 0.28, + "learning_rate": 0.0005782146439317954, + "loss": 1.4354, + "theoretical_loss": 3.699785137999563, + "tokens_seen": 938803200 + }, + { + "epoch": 0.28, + "learning_rate": 0.0005781985957873621, + "loss": 1.4407, + "theoretical_loss": 3.6997604635300725, + "tokens_seen": 938868736 + }, + { + "epoch": 0.28, + "learning_rate": 0.0005781825476429288, + "loss": 1.4678, + "theoretical_loss": 3.699735791265099, + "tokens_seen": 938934272 + }, + { + "epoch": 0.28, + "learning_rate": 0.0005781664994984955, + "loss": 1.3551, + "theoretical_loss": 3.6997111212042917, + "tokens_seen": 938999808 + }, + { + "epoch": 0.28, + "learning_rate": 0.0005781504513540621, + "loss": 1.5026, + "theoretical_loss": 3.6996864533473, + "tokens_seen": 939065344 + }, + { + "epoch": 0.28, + "learning_rate": 0.0005781344032096289, + "loss": 1.5173, + "theoretical_loss": 3.6996617876937723, + "tokens_seen": 939130880 + }, + { + "epoch": 0.28, + "learning_rate": 0.0005781183550651955, + "loss": 1.4612, + "theoretical_loss": 3.6996371242433597, + "tokens_seen": 939196416 + }, + { + "epoch": 0.28, + "learning_rate": 0.0005781023069207623, + "loss": 1.4319, + "theoretical_loss": 3.69961246299571, + "tokens_seen": 939261952 + }, + { + "epoch": 0.28, + "learning_rate": 0.0005780862587763289, + "loss": 1.421, + "theoretical_loss": 3.699587803950474, + "tokens_seen": 939327488 + }, + { + "epoch": 0.28, + "learning_rate": 0.0005780702106318957, + "loss": 1.4445, + "theoretical_loss": 3.699563147107301, + "tokens_seen": 939393024 + }, + { + "epoch": 0.28, + "learning_rate": 0.0005780541624874623, + "loss": 1.2821, + "theoretical_loss": 3.6995384924658405, + "tokens_seen": 939458560 + }, + { + "epoch": 0.28, + "learning_rate": 0.0005780381143430291, + "loss": 1.3097, + "theoretical_loss": 3.699513840025743, + "tokens_seen": 939524096 + }, + { + "epoch": 0.28, + "learning_rate": 0.0005780220661985957, + "loss": 1.4863, + "theoretical_loss": 3.6994891897866573, + "tokens_seen": 939589632 + }, + { + "epoch": 0.28, + "learning_rate": 0.0005780060180541625, + "loss": 1.3368, + "theoretical_loss": 3.6994645417482346, + "tokens_seen": 939655168 + }, + { + "epoch": 0.28, + "learning_rate": 0.0005779899699097291, + "loss": 1.459, + "theoretical_loss": 3.6994398959101242, + "tokens_seen": 939720704 + }, + { + "epoch": 0.28, + "learning_rate": 0.0005779739217652959, + "loss": 1.4326, + "theoretical_loss": 3.6994152522719768, + "tokens_seen": 939786240 + }, + { + "epoch": 0.28, + "learning_rate": 0.0005779578736208625, + "loss": 1.4853, + "theoretical_loss": 3.699390610833442, + "tokens_seen": 939851776 + }, + { + "epoch": 0.28, + "learning_rate": 0.0005779418254764293, + "loss": 1.4802, + "theoretical_loss": 3.6993659715941702, + "tokens_seen": 939917312 + }, + { + "epoch": 0.28, + "learning_rate": 0.0005779257773319959, + "loss": 1.5317, + "theoretical_loss": 3.699341334553812, + "tokens_seen": 939982848 + }, + { + "epoch": 0.28, + "learning_rate": 0.0005779097291875627, + "loss": 1.3508, + "theoretical_loss": 3.6993166997120186, + "tokens_seen": 940048384 + }, + { + "epoch": 0.28, + "learning_rate": 0.0005778936810431293, + "loss": 1.4233, + "theoretical_loss": 3.6992920670684395, + "tokens_seen": 940113920 + }, + { + "epoch": 0.28, + "learning_rate": 0.0005778776328986961, + "loss": 1.4711, + "theoretical_loss": 3.6992674366227254, + "tokens_seen": 940179456 + }, + { + "epoch": 0.28, + "learning_rate": 0.0005778615847542627, + "loss": 1.444, + "theoretical_loss": 3.6992428083745272, + "tokens_seen": 940244992 + }, + { + "epoch": 0.28, + "learning_rate": 0.0005778455366098295, + "loss": 1.3447, + "theoretical_loss": 3.699218182323496, + "tokens_seen": 940310528 + }, + { + "epoch": 0.28, + "learning_rate": 0.0005778294884653961, + "loss": 1.4426, + "theoretical_loss": 3.6991935584692826, + "tokens_seen": 940376064 + }, + { + "epoch": 0.28, + "objective/train/docs_used": 345855, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.38521409034729, + "objective/train/theoretical_loss": 3.6991689368115375, + "objective/train/tokens_used": 960901600, + "theoretical_loss": 3.6991689368115375, + "tokens_seen": 940441600 + }, + { + "epoch": 0.28, + "learning_rate": 0.0005778134403209629, + "loss": 1.4239, + "theoretical_loss": 3.6991689368115375, + "tokens_seen": 940441600 + }, + { + "epoch": 0.29, + "learning_rate": 0.0005777973921765295, + "loss": 1.5457, + "theoretical_loss": 3.699144317349912, + "tokens_seen": 940507136 + }, + { + "epoch": 0.29, + "learning_rate": 0.0005777813440320963, + "loss": 1.3495, + "theoretical_loss": 3.6991197000840574, + "tokens_seen": 940572672 + }, + { + "epoch": 0.29, + "learning_rate": 0.000577765295887663, + "loss": 1.449, + "theoretical_loss": 3.6990950850136244, + "tokens_seen": 940638208 + }, + { + "epoch": 0.29, + "learning_rate": 0.0005777492477432297, + "loss": 1.4242, + "theoretical_loss": 3.6990704721382652, + "tokens_seen": 940703744 + }, + { + "epoch": 0.29, + "learning_rate": 0.0005777331995987965, + "loss": 1.2818, + "theoretical_loss": 3.69904586145763, + "tokens_seen": 940769280 + }, + { + "epoch": 0.29, + "learning_rate": 0.0005777171514543631, + "loss": 1.4651, + "theoretical_loss": 3.699021252971371, + "tokens_seen": 940834816 + }, + { + "epoch": 0.29, + "learning_rate": 0.0005777011033099299, + "loss": 1.4055, + "theoretical_loss": 3.698996646679139, + "tokens_seen": 940900352 + }, + { + "epoch": 0.29, + "learning_rate": 0.0005776850551654965, + "loss": 1.4716, + "theoretical_loss": 3.6989720425805865, + "tokens_seen": 940965888 + }, + { + "epoch": 0.29, + "learning_rate": 0.0005776690070210633, + "loss": 1.4901, + "theoretical_loss": 3.6989474406753646, + "tokens_seen": 941031424 + }, + { + "epoch": 0.29, + "learning_rate": 0.0005776529588766299, + "loss": 1.5083, + "theoretical_loss": 3.6989228409631254, + "tokens_seen": 941096960 + }, + { + "epoch": 0.29, + "learning_rate": 0.0005776369107321967, + "loss": 1.5216, + "theoretical_loss": 3.6988982434435203, + "tokens_seen": 941162496 + }, + { + "epoch": 0.29, + "learning_rate": 0.0005776208625877633, + "loss": 1.4654, + "theoretical_loss": 3.6988736481162015, + "tokens_seen": 941228032 + }, + { + "epoch": 0.29, + "learning_rate": 0.0005776048144433301, + "loss": 1.4216, + "theoretical_loss": 3.698849054980821, + "tokens_seen": 941293568 + }, + { + "epoch": 0.29, + "learning_rate": 0.0005775887662988967, + "loss": 1.4201, + "theoretical_loss": 3.6988244640370302, + "tokens_seen": 941359104 + }, + { + "epoch": 0.29, + "learning_rate": 0.0005775727181544635, + "loss": 1.4724, + "theoretical_loss": 3.6987998752844815, + "tokens_seen": 941424640 + }, + { + "epoch": 0.29, + "learning_rate": 0.0005775566700100301, + "loss": 1.468, + "theoretical_loss": 3.6987752887228282, + "tokens_seen": 941490176 + }, + { + "epoch": 0.29, + "learning_rate": 0.0005775406218655969, + "loss": 1.4225, + "theoretical_loss": 3.698750704351722, + "tokens_seen": 941555712 + }, + { + "epoch": 0.29, + "learning_rate": 0.0005775245737211635, + "loss": 1.5109, + "theoretical_loss": 3.6987261221708145, + "tokens_seen": 941621248 + }, + { + "epoch": 0.29, + "learning_rate": 0.0005775085255767303, + "loss": 1.5145, + "theoretical_loss": 3.6987015421797587, + "tokens_seen": 941686784 + }, + { + "epoch": 0.29, + "learning_rate": 0.0005774924774322969, + "loss": 1.5792, + "theoretical_loss": 3.698676964378208, + "tokens_seen": 941752320 + }, + { + "epoch": 0.29, + "learning_rate": 0.0005774764292878637, + "loss": 1.3107, + "theoretical_loss": 3.6986523887658134, + "tokens_seen": 941817856 + }, + { + "epoch": 0.29, + "learning_rate": 0.0005774603811434303, + "loss": 1.3478, + "theoretical_loss": 3.6986278153422285, + "tokens_seen": 941883392 + }, + { + "epoch": 0.29, + "learning_rate": 0.0005774443329989971, + "loss": 1.4625, + "theoretical_loss": 3.6986032441071064, + "tokens_seen": 941948928 + }, + { + "epoch": 0.29, + "learning_rate": 0.0005774282848545637, + "loss": 1.481, + "theoretical_loss": 3.6985786750600993, + "tokens_seen": 942014464 + }, + { + "epoch": 0.29, + "learning_rate": 0.0005774122367101305, + "loss": 1.3907, + "theoretical_loss": 3.6985541082008604, + "tokens_seen": 942080000 + }, + { + "epoch": 0.29, + "learning_rate": 0.0005773961885656971, + "loss": 1.4176, + "theoretical_loss": 3.698529543529043, + "tokens_seen": 942145536 + }, + { + "epoch": 0.29, + "learning_rate": 0.0005773801404212639, + "loss": 1.3884, + "theoretical_loss": 3.6985049810443, + "tokens_seen": 942211072 + }, + { + "epoch": 0.29, + "learning_rate": 0.0005773640922768305, + "loss": 1.4396, + "theoretical_loss": 3.6984804207462836, + "tokens_seen": 942276608 + }, + { + "epoch": 0.29, + "learning_rate": 0.0005773480441323973, + "loss": 1.4705, + "theoretical_loss": 3.698455862634648, + "tokens_seen": 942342144 + }, + { + "epoch": 0.29, + "learning_rate": 0.0005773319959879639, + "loss": 1.4507, + "theoretical_loss": 3.6984313067090477, + "tokens_seen": 942407680 + }, + { + "epoch": 0.29, + "learning_rate": 0.0005773159478435307, + "loss": 1.5281, + "theoretical_loss": 3.698406752969134, + "tokens_seen": 942473216 + }, + { + "epoch": 0.29, + "learning_rate": 0.0005772998996990973, + "loss": 1.4818, + "theoretical_loss": 3.698382201414561, + "tokens_seen": 942538752 + }, + { + "epoch": 0.29, + "learning_rate": 0.0005772838515546641, + "loss": 1.4214, + "theoretical_loss": 3.6983576520449835, + "tokens_seen": 942604288 + }, + { + "epoch": 0.29, + "learning_rate": 0.0005772678034102307, + "loss": 1.3249, + "theoretical_loss": 3.6983331048600534, + "tokens_seen": 942669824 + }, + { + "epoch": 0.29, + "learning_rate": 0.0005772517552657975, + "loss": 1.4297, + "theoretical_loss": 3.6983085598594254, + "tokens_seen": 942735360 + }, + { + "epoch": 0.29, + "learning_rate": 0.0005772357071213641, + "loss": 1.5195, + "theoretical_loss": 3.698284017042753, + "tokens_seen": 942800896 + }, + { + "epoch": 0.29, + "learning_rate": 0.0005772196589769309, + "loss": 1.5612, + "theoretical_loss": 3.6982594764096905, + "tokens_seen": 942866432 + }, + { + "epoch": 0.29, + "learning_rate": 0.0005772036108324975, + "loss": 1.4603, + "theoretical_loss": 3.6982349379598913, + "tokens_seen": 942931968 + }, + { + "epoch": 0.29, + "learning_rate": 0.0005771875626880643, + "loss": 1.3241, + "theoretical_loss": 3.69821040169301, + "tokens_seen": 942997504 + }, + { + "epoch": 0.29, + "learning_rate": 0.0005771715145436309, + "loss": 1.4797, + "theoretical_loss": 3.6981858676087005, + "tokens_seen": 943063040 + }, + { + "epoch": 0.29, + "learning_rate": 0.0005771554663991977, + "loss": 1.4471, + "theoretical_loss": 3.698161335706617, + "tokens_seen": 943128576 + }, + { + "epoch": 0.29, + "learning_rate": 0.0005771394182547643, + "loss": 1.4265, + "theoretical_loss": 3.6981368059864135, + "tokens_seen": 943194112 + }, + { + "epoch": 0.29, + "learning_rate": 0.0005771233701103311, + "loss": 1.5424, + "theoretical_loss": 3.698112278447744, + "tokens_seen": 943259648 + }, + { + "epoch": 0.29, + "learning_rate": 0.0005771073219658977, + "loss": 1.5619, + "theoretical_loss": 3.6980877530902645, + "tokens_seen": 943325184 + }, + { + "epoch": 0.29, + "learning_rate": 0.0005770912738214645, + "loss": 1.4742, + "theoretical_loss": 3.698063229913628, + "tokens_seen": 943390720 + }, + { + "epoch": 0.29, + "learning_rate": 0.0005770752256770311, + "loss": 1.4597, + "theoretical_loss": 3.6980387089174895, + "tokens_seen": 943456256 + }, + { + "epoch": 0.29, + "learning_rate": 0.0005770591775325979, + "loss": 1.4903, + "theoretical_loss": 3.6980141901015045, + "tokens_seen": 943521792 + }, + { + "epoch": 0.29, + "learning_rate": 0.0005770431293881645, + "loss": 1.4835, + "theoretical_loss": 3.697989673465327, + "tokens_seen": 943587328 + }, + { + "epoch": 0.29, + "learning_rate": 0.0005770270812437313, + "loss": 1.5316, + "theoretical_loss": 3.6979651590086107, + "tokens_seen": 943652864 + }, + { + "epoch": 0.29, + "objective/train/docs_used": 346959, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.349886417388916, + "objective/train/theoretical_loss": 3.6979406467310127, + "objective/train/tokens_used": 964178400, + "theoretical_loss": 3.6979406467310127, + "tokens_seen": 943718400 + }, + { + "epoch": 0.29, + "learning_rate": 0.0005770110330992979, + "loss": 1.3424, + "theoretical_loss": 3.6979406467310127, + "tokens_seen": 943718400 + }, + { + "epoch": 0.29, + "learning_rate": 0.0005769949849548647, + "loss": 1.4714, + "theoretical_loss": 3.6979161366321867, + "tokens_seen": 943783936 + }, + { + "epoch": 0.29, + "learning_rate": 0.0005769789368104313, + "loss": 1.431, + "theoretical_loss": 3.6978916287117882, + "tokens_seen": 943849472 + }, + { + "epoch": 0.29, + "learning_rate": 0.0005769628886659981, + "loss": 1.4463, + "theoretical_loss": 3.6978671229694724, + "tokens_seen": 943915008 + }, + { + "epoch": 0.29, + "learning_rate": 0.0005769468405215647, + "loss": 1.5107, + "theoretical_loss": 3.697842619404894, + "tokens_seen": 943980544 + }, + { + "epoch": 0.29, + "learning_rate": 0.0005769307923771315, + "loss": 1.4262, + "theoretical_loss": 3.6978181180177083, + "tokens_seen": 944046080 + }, + { + "epoch": 0.29, + "learning_rate": 0.0005769147442326981, + "loss": 1.513, + "theoretical_loss": 3.6977936188075717, + "tokens_seen": 944111616 + }, + { + "epoch": 0.29, + "learning_rate": 0.0005768986960882649, + "loss": 1.4396, + "theoretical_loss": 3.6977691217741384, + "tokens_seen": 944177152 + }, + { + "epoch": 0.29, + "learning_rate": 0.0005768826479438315, + "loss": 1.4398, + "theoretical_loss": 3.697744626917065, + "tokens_seen": 944242688 + }, + { + "epoch": 0.29, + "learning_rate": 0.0005768665997993982, + "loss": 1.4137, + "theoretical_loss": 3.6977201342360067, + "tokens_seen": 944308224 + }, + { + "epoch": 0.29, + "learning_rate": 0.0005768505516549649, + "loss": 1.4607, + "theoretical_loss": 3.6976956437306185, + "tokens_seen": 944373760 + }, + { + "epoch": 0.29, + "learning_rate": 0.0005768345035105316, + "loss": 1.4277, + "theoretical_loss": 3.697671155400557, + "tokens_seen": 944439296 + }, + { + "epoch": 0.29, + "learning_rate": 0.0005768184553660983, + "loss": 1.4699, + "theoretical_loss": 3.6976466692454784, + "tokens_seen": 944504832 + }, + { + "epoch": 0.29, + "learning_rate": 0.000576802407221665, + "loss": 1.4575, + "theoretical_loss": 3.697622185265037, + "tokens_seen": 944570368 + }, + { + "epoch": 0.29, + "learning_rate": 0.0005767863590772317, + "loss": 1.4659, + "theoretical_loss": 3.6975977034588907, + "tokens_seen": 944635904 + }, + { + "epoch": 0.29, + "learning_rate": 0.0005767703109327984, + "loss": 1.3772, + "theoretical_loss": 3.6975732238266943, + "tokens_seen": 944701440 + }, + { + "epoch": 0.29, + "learning_rate": 0.0005767542627883651, + "loss": 1.4711, + "theoretical_loss": 3.697548746368105, + "tokens_seen": 944766976 + }, + { + "epoch": 0.29, + "learning_rate": 0.0005767382146439318, + "loss": 1.4258, + "theoretical_loss": 3.697524271082778, + "tokens_seen": 944832512 + }, + { + "epoch": 0.29, + "learning_rate": 0.0005767221664994984, + "loss": 1.4455, + "theoretical_loss": 3.6974997979703703, + "tokens_seen": 944898048 + }, + { + "epoch": 0.29, + "learning_rate": 0.0005767061183550652, + "loss": 1.4795, + "theoretical_loss": 3.697475327030538, + "tokens_seen": 944963584 + }, + { + "epoch": 0.29, + "learning_rate": 0.0005766900702106318, + "loss": 1.5239, + "theoretical_loss": 3.697450858262938, + "tokens_seen": 945029120 + }, + { + "epoch": 0.29, + "learning_rate": 0.0005766740220661986, + "loss": 1.3404, + "theoretical_loss": 3.697426391667226, + "tokens_seen": 945094656 + }, + { + "epoch": 0.29, + "learning_rate": 0.0005766579739217652, + "loss": 1.433, + "theoretical_loss": 3.6974019272430594, + "tokens_seen": 945160192 + }, + { + "epoch": 0.29, + "learning_rate": 0.000576641925777332, + "loss": 1.4269, + "theoretical_loss": 3.697377464990095, + "tokens_seen": 945225728 + }, + { + "epoch": 0.29, + "learning_rate": 0.0005766258776328986, + "loss": 1.4304, + "theoretical_loss": 3.697353004907989, + "tokens_seen": 945291264 + }, + { + "epoch": 0.29, + "learning_rate": 0.0005766098294884654, + "loss": 1.4944, + "theoretical_loss": 3.697328546996398, + "tokens_seen": 945356800 + }, + { + "epoch": 0.29, + "learning_rate": 0.000576593781344032, + "loss": 1.3869, + "theoretical_loss": 3.69730409125498, + "tokens_seen": 945422336 + }, + { + "epoch": 0.29, + "learning_rate": 0.0005765777331995988, + "loss": 1.3352, + "theoretical_loss": 3.6972796376833914, + "tokens_seen": 945487872 + }, + { + "epoch": 0.29, + "learning_rate": 0.0005765616850551655, + "loss": 1.4475, + "theoretical_loss": 3.6972551862812892, + "tokens_seen": 945553408 + }, + { + "epoch": 0.29, + "learning_rate": 0.0005765456369107322, + "loss": 1.4059, + "theoretical_loss": 3.697230737048331, + "tokens_seen": 945618944 + }, + { + "epoch": 0.29, + "learning_rate": 0.000576529588766299, + "loss": 1.4013, + "theoretical_loss": 3.697206289984174, + "tokens_seen": 945684480 + }, + { + "epoch": 0.29, + "learning_rate": 0.0005765135406218656, + "loss": 1.3271, + "theoretical_loss": 3.6971818450884752, + "tokens_seen": 945750016 + }, + { + "epoch": 0.29, + "learning_rate": 0.0005764974924774323, + "loss": 1.4728, + "theoretical_loss": 3.6971574023608924, + "tokens_seen": 945815552 + }, + { + "epoch": 0.29, + "learning_rate": 0.000576481444332999, + "loss": 1.4026, + "theoretical_loss": 3.6971329618010826, + "tokens_seen": 945881088 + }, + { + "epoch": 0.29, + "learning_rate": 0.0005764653961885658, + "loss": 1.3792, + "theoretical_loss": 3.697108523408704, + "tokens_seen": 945946624 + }, + { + "epoch": 0.29, + "learning_rate": 0.0005764493480441324, + "loss": 1.4043, + "theoretical_loss": 3.6970840871834136, + "tokens_seen": 946012160 + }, + { + "epoch": 0.29, + "learning_rate": 0.0005764332998996992, + "loss": 1.2978, + "theoretical_loss": 3.697059653124869, + "tokens_seen": 946077696 + }, + { + "epoch": 0.29, + "learning_rate": 0.0005764172517552658, + "loss": 1.5154, + "theoretical_loss": 3.697035221232729, + "tokens_seen": 946143232 + }, + { + "epoch": 0.29, + "learning_rate": 0.0005764012036108326, + "loss": 1.5087, + "theoretical_loss": 3.697010791506651, + "tokens_seen": 946208768 + }, + { + "epoch": 0.29, + "learning_rate": 0.0005763851554663992, + "loss": 1.4231, + "theoretical_loss": 3.6969863639462925, + "tokens_seen": 946274304 + }, + { + "epoch": 0.29, + "learning_rate": 0.000576369107321966, + "loss": 1.5142, + "theoretical_loss": 3.696961938551312, + "tokens_seen": 946339840 + }, + { + "epoch": 0.29, + "learning_rate": 0.0005763530591775326, + "loss": 1.3433, + "theoretical_loss": 3.696937515321368, + "tokens_seen": 946405376 + }, + { + "epoch": 0.29, + "learning_rate": 0.0005763370110330994, + "loss": 1.4807, + "theoretical_loss": 3.696913094256117, + "tokens_seen": 946470912 + }, + { + "epoch": 0.29, + "learning_rate": 0.000576320962888666, + "loss": 1.4616, + "theoretical_loss": 3.6968886753552193, + "tokens_seen": 946536448 + }, + { + "epoch": 0.29, + "learning_rate": 0.0005763049147442328, + "loss": 1.5236, + "theoretical_loss": 3.6968642586183327, + "tokens_seen": 946601984 + }, + { + "epoch": 0.29, + "learning_rate": 0.0005762888665997994, + "loss": 1.3936, + "theoretical_loss": 3.6968398440451145, + "tokens_seen": 946667520 + }, + { + "epoch": 0.29, + "learning_rate": 0.0005762728184553662, + "loss": 1.4202, + "theoretical_loss": 3.696815431635225, + "tokens_seen": 946733056 + }, + { + "epoch": 0.29, + "learning_rate": 0.0005762567703109328, + "loss": 1.5053, + "theoretical_loss": 3.696791021388321, + "tokens_seen": 946798592 + }, + { + "epoch": 0.29, + "learning_rate": 0.0005762407221664996, + "loss": 1.4381, + "theoretical_loss": 3.6967666133040624, + "tokens_seen": 946864128 + }, + { + "epoch": 0.29, + "learning_rate": 0.0005762246740220662, + "loss": 1.5663, + "theoretical_loss": 3.6967422073821075, + "tokens_seen": 946929664 + }, + { + "epoch": 0.29, + "objective/train/docs_used": 348186, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.438401699066162, + "objective/train/theoretical_loss": 3.696717803622115, + "objective/train/tokens_used": 967455200, + "theoretical_loss": 3.696717803622115, + "tokens_seen": 946995200 + }, + { + "epoch": 0.29, + "learning_rate": 0.000576208625877633, + "loss": 1.4347, + "theoretical_loss": 3.696717803622115, + "tokens_seen": 946995200 + }, + { + "epoch": 0.29, + "learning_rate": 0.0005761925777331996, + "loss": 1.3399, + "theoretical_loss": 3.696693402023744, + "tokens_seen": 947060736 + }, + { + "epoch": 0.29, + "learning_rate": 0.0005761765295887664, + "loss": 1.3569, + "theoretical_loss": 3.696669002586653, + "tokens_seen": 947126272 + }, + { + "epoch": 0.29, + "learning_rate": 0.000576160481444333, + "loss": 1.3861, + "theoretical_loss": 3.696644605310502, + "tokens_seen": 947191808 + }, + { + "epoch": 0.29, + "learning_rate": 0.0005761444332998998, + "loss": 1.4475, + "theoretical_loss": 3.696620210194949, + "tokens_seen": 947257344 + }, + { + "epoch": 0.29, + "learning_rate": 0.0005761283851554664, + "loss": 1.5122, + "theoretical_loss": 3.696595817239654, + "tokens_seen": 947322880 + }, + { + "epoch": 0.29, + "learning_rate": 0.0005761123370110332, + "loss": 1.5283, + "theoretical_loss": 3.6965714264442764, + "tokens_seen": 947388416 + }, + { + "epoch": 0.29, + "learning_rate": 0.0005760962888665998, + "loss": 1.47, + "theoretical_loss": 3.6965470378084744, + "tokens_seen": 947453952 + }, + { + "epoch": 0.29, + "learning_rate": 0.0005760802407221666, + "loss": 1.4994, + "theoretical_loss": 3.696522651331909, + "tokens_seen": 947519488 + }, + { + "epoch": 0.29, + "learning_rate": 0.0005760641925777332, + "loss": 1.3772, + "theoretical_loss": 3.6964982670142383, + "tokens_seen": 947585024 + }, + { + "epoch": 0.29, + "learning_rate": 0.0005760481444333, + "loss": 1.559, + "theoretical_loss": 3.6964738848551226, + "tokens_seen": 947650560 + }, + { + "epoch": 0.29, + "learning_rate": 0.0005760320962888666, + "loss": 1.3742, + "theoretical_loss": 3.696449504854222, + "tokens_seen": 947716096 + }, + { + "epoch": 0.29, + "learning_rate": 0.0005760160481444334, + "loss": 1.4783, + "theoretical_loss": 3.696425127011195, + "tokens_seen": 947781632 + }, + { + "epoch": 0.29, + "learning_rate": 0.000576, + "loss": 1.3737, + "theoretical_loss": 3.6964007513257022, + "tokens_seen": 947847168 + }, + { + "epoch": 0.29, + "learning_rate": 0.0005759839518555668, + "loss": 1.5142, + "theoretical_loss": 3.696376377797404, + "tokens_seen": 947912704 + }, + { + "epoch": 0.29, + "learning_rate": 0.0005759679037111334, + "loss": 1.3254, + "theoretical_loss": 3.696352006425959, + "tokens_seen": 947978240 + }, + { + "epoch": 0.29, + "learning_rate": 0.0005759518555667002, + "loss": 1.4671, + "theoretical_loss": 3.696327637211028, + "tokens_seen": 948043776 + }, + { + "epoch": 0.29, + "learning_rate": 0.0005759358074222668, + "loss": 1.291, + "theoretical_loss": 3.6963032701522716, + "tokens_seen": 948109312 + }, + { + "epoch": 0.29, + "learning_rate": 0.0005759197592778336, + "loss": 1.4325, + "theoretical_loss": 3.6962789052493497, + "tokens_seen": 948174848 + }, + { + "epoch": 0.29, + "learning_rate": 0.0005759037111334002, + "loss": 1.3965, + "theoretical_loss": 3.696254542501922, + "tokens_seen": 948240384 + }, + { + "epoch": 0.29, + "learning_rate": 0.000575887662988967, + "loss": 1.4048, + "theoretical_loss": 3.6962301819096495, + "tokens_seen": 948305920 + }, + { + "epoch": 0.29, + "learning_rate": 0.0005758716148445336, + "loss": 1.5135, + "theoretical_loss": 3.696205823472192, + "tokens_seen": 948371456 + }, + { + "epoch": 0.29, + "learning_rate": 0.0005758555667001004, + "loss": 1.4299, + "theoretical_loss": 3.696181467189211, + "tokens_seen": 948436992 + }, + { + "epoch": 0.29, + "learning_rate": 0.000575839518555667, + "loss": 1.4287, + "theoretical_loss": 3.6961571130603663, + "tokens_seen": 948502528 + }, + { + "epoch": 0.29, + "learning_rate": 0.0005758234704112338, + "loss": 1.6366, + "theoretical_loss": 3.6961327610853187, + "tokens_seen": 948568064 + }, + { + "epoch": 0.29, + "learning_rate": 0.0005758074222668004, + "loss": 1.3668, + "theoretical_loss": 3.6961084112637295, + "tokens_seen": 948633600 + }, + { + "epoch": 0.29, + "learning_rate": 0.0005757913741223672, + "loss": 1.4956, + "theoretical_loss": 3.6960840635952583, + "tokens_seen": 948699136 + }, + { + "epoch": 0.29, + "learning_rate": 0.0005757753259779338, + "loss": 1.438, + "theoretical_loss": 3.696059718079567, + "tokens_seen": 948764672 + }, + { + "epoch": 0.29, + "learning_rate": 0.0005757592778335006, + "loss": 1.5274, + "theoretical_loss": 3.696035374716316, + "tokens_seen": 948830208 + }, + { + "epoch": 0.29, + "learning_rate": 0.0005757432296890672, + "loss": 1.4453, + "theoretical_loss": 3.696011033505167, + "tokens_seen": 948895744 + }, + { + "epoch": 0.29, + "learning_rate": 0.000575727181544634, + "loss": 1.5083, + "theoretical_loss": 3.6959866944457813, + "tokens_seen": 948961280 + }, + { + "epoch": 0.29, + "learning_rate": 0.0005757111334002006, + "loss": 1.4412, + "theoretical_loss": 3.6959623575378187, + "tokens_seen": 949026816 + }, + { + "epoch": 0.29, + "learning_rate": 0.0005756950852557674, + "loss": 1.4711, + "theoretical_loss": 3.6959380227809415, + "tokens_seen": 949092352 + }, + { + "epoch": 0.29, + "learning_rate": 0.000575679037111334, + "loss": 1.469, + "theoretical_loss": 3.6959136901748115, + "tokens_seen": 949157888 + }, + { + "epoch": 0.29, + "learning_rate": 0.0005756629889669008, + "loss": 1.4309, + "theoretical_loss": 3.695889359719089, + "tokens_seen": 949223424 + }, + { + "epoch": 0.29, + "learning_rate": 0.0005756469408224674, + "loss": 1.4572, + "theoretical_loss": 3.695865031413436, + "tokens_seen": 949288960 + }, + { + "epoch": 0.29, + "learning_rate": 0.0005756308926780342, + "loss": 1.4702, + "theoretical_loss": 3.695840705257514, + "tokens_seen": 949354496 + }, + { + "epoch": 0.29, + "learning_rate": 0.0005756148445336008, + "loss": 1.4632, + "theoretical_loss": 3.6958163812509857, + "tokens_seen": 949420032 + }, + { + "epoch": 0.29, + "learning_rate": 0.0005755987963891676, + "loss": 1.5321, + "theoretical_loss": 3.695792059393511, + "tokens_seen": 949485568 + }, + { + "epoch": 0.29, + "learning_rate": 0.0005755827482447342, + "loss": 1.5186, + "theoretical_loss": 3.695767739684753, + "tokens_seen": 949551104 + }, + { + "epoch": 0.29, + "learning_rate": 0.000575566700100301, + "loss": 1.4312, + "theoretical_loss": 3.695743422124373, + "tokens_seen": 949616640 + }, + { + "epoch": 0.29, + "learning_rate": 0.0005755506519558676, + "loss": 1.4357, + "theoretical_loss": 3.6957191067120334, + "tokens_seen": 949682176 + }, + { + "epoch": 0.29, + "learning_rate": 0.0005755346038114344, + "loss": 1.4337, + "theoretical_loss": 3.6956947934473963, + "tokens_seen": 949747712 + }, + { + "epoch": 0.29, + "learning_rate": 0.000575518555667001, + "loss": 1.5488, + "theoretical_loss": 3.695670482330123, + "tokens_seen": 949813248 + }, + { + "epoch": 0.29, + "learning_rate": 0.0005755025075225678, + "loss": 1.3974, + "theoretical_loss": 3.6956461733598767, + "tokens_seen": 949878784 + }, + { + "epoch": 0.29, + "learning_rate": 0.0005754864593781344, + "loss": 1.4359, + "theoretical_loss": 3.695621866536319, + "tokens_seen": 949944320 + }, + { + "epoch": 0.29, + "learning_rate": 0.0005754704112337011, + "loss": 1.4014, + "theoretical_loss": 3.6955975618591133, + "tokens_seen": 950009856 + }, + { + "epoch": 0.29, + "learning_rate": 0.0005754543630892678, + "loss": 1.4709, + "theoretical_loss": 3.6955732593279205, + "tokens_seen": 950075392 + }, + { + "epoch": 0.29, + "learning_rate": 0.0005754383149448345, + "loss": 1.4971, + "theoretical_loss": 3.6955489589424033, + "tokens_seen": 950140928 + }, + { + "epoch": 0.29, + "learning_rate": 0.0005754222668004012, + "loss": 1.4205, + "theoretical_loss": 3.6955246607022256, + "tokens_seen": 950206464 + }, + { + "epoch": 0.29, + "objective/train/docs_used": 349336, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.366361379623413, + "objective/train/theoretical_loss": 3.695500364607049, + "objective/train/tokens_used": 970732000, + "theoretical_loss": 3.695500364607049, + "tokens_seen": 950272000 + }, + { + "epoch": 0.29, + "learning_rate": 0.0005754062186559679, + "loss": 1.3866, + "theoretical_loss": 3.695500364607049, + "tokens_seen": 950272000 + }, + { + "epoch": 0.29, + "learning_rate": 0.0005753901705115346, + "loss": 1.3595, + "theoretical_loss": 3.695476070656537, + "tokens_seen": 950337536 + }, + { + "epoch": 0.29, + "learning_rate": 0.0005753741223671013, + "loss": 1.5752, + "theoretical_loss": 3.6954517788503507, + "tokens_seen": 950403072 + }, + { + "epoch": 0.29, + "learning_rate": 0.000575358074222668, + "loss": 1.3023, + "theoretical_loss": 3.695427489188155, + "tokens_seen": 950468608 + }, + { + "epoch": 0.29, + "learning_rate": 0.0005753420260782347, + "loss": 1.376, + "theoretical_loss": 3.6954032016696123, + "tokens_seen": 950534144 + }, + { + "epoch": 0.29, + "learning_rate": 0.0005753259779338014, + "loss": 1.4238, + "theoretical_loss": 3.6953789162943846, + "tokens_seen": 950599680 + }, + { + "epoch": 0.29, + "learning_rate": 0.0005753099297893681, + "loss": 1.5772, + "theoretical_loss": 3.695354633062136, + "tokens_seen": 950665216 + }, + { + "epoch": 0.29, + "learning_rate": 0.0005752938816449348, + "loss": 1.3868, + "theoretical_loss": 3.6953303519725296, + "tokens_seen": 950730752 + }, + { + "epoch": 0.29, + "learning_rate": 0.0005752778335005015, + "loss": 1.3894, + "theoretical_loss": 3.695306073025229, + "tokens_seen": 950796288 + }, + { + "epoch": 0.29, + "learning_rate": 0.0005752617853560682, + "loss": 1.4969, + "theoretical_loss": 3.6952817962198967, + "tokens_seen": 950861824 + }, + { + "epoch": 0.29, + "learning_rate": 0.0005752457372116349, + "loss": 1.48, + "theoretical_loss": 3.6952575215561962, + "tokens_seen": 950927360 + }, + { + "epoch": 0.29, + "learning_rate": 0.0005752296890672016, + "loss": 1.3783, + "theoretical_loss": 3.695233249033792, + "tokens_seen": 950992896 + }, + { + "epoch": 0.29, + "learning_rate": 0.0005752136409227683, + "loss": 1.4152, + "theoretical_loss": 3.6952089786523468, + "tokens_seen": 951058432 + }, + { + "epoch": 0.29, + "learning_rate": 0.000575197592778335, + "loss": 1.3947, + "theoretical_loss": 3.6951847104115236, + "tokens_seen": 951123968 + }, + { + "epoch": 0.29, + "learning_rate": 0.0005751815446339017, + "loss": 1.4804, + "theoretical_loss": 3.6951604443109876, + "tokens_seen": 951189504 + }, + { + "epoch": 0.29, + "learning_rate": 0.0005751654964894684, + "loss": 1.4882, + "theoretical_loss": 3.695136180350402, + "tokens_seen": 951255040 + }, + { + "epoch": 0.29, + "learning_rate": 0.0005751494483450351, + "loss": 1.443, + "theoretical_loss": 3.69511191852943, + "tokens_seen": 951320576 + }, + { + "epoch": 0.29, + "learning_rate": 0.0005751334002006018, + "loss": 1.5086, + "theoretical_loss": 3.695087658847737, + "tokens_seen": 951386112 + }, + { + "epoch": 0.29, + "learning_rate": 0.0005751173520561685, + "loss": 1.4111, + "theoretical_loss": 3.695063401304986, + "tokens_seen": 951451648 + }, + { + "epoch": 0.29, + "learning_rate": 0.0005751013039117353, + "loss": 1.4824, + "theoretical_loss": 3.695039145900841, + "tokens_seen": 951517184 + }, + { + "epoch": 0.29, + "learning_rate": 0.0005750852557673019, + "loss": 1.3389, + "theoretical_loss": 3.6950148926349664, + "tokens_seen": 951582720 + }, + { + "epoch": 0.29, + "learning_rate": 0.0005750692076228687, + "loss": 1.5131, + "theoretical_loss": 3.694990641507027, + "tokens_seen": 951648256 + }, + { + "epoch": 0.29, + "learning_rate": 0.0005750531594784353, + "loss": 1.4587, + "theoretical_loss": 3.694966392516686, + "tokens_seen": 951713792 + }, + { + "epoch": 0.29, + "learning_rate": 0.000575037111334002, + "loss": 1.4166, + "theoretical_loss": 3.6949421456636085, + "tokens_seen": 951779328 + }, + { + "epoch": 0.29, + "learning_rate": 0.0005750210631895687, + "loss": 1.4582, + "theoretical_loss": 3.694917900947459, + "tokens_seen": 951844864 + }, + { + "epoch": 0.29, + "learning_rate": 0.0005750050150451355, + "loss": 1.3991, + "theoretical_loss": 3.6948936583679024, + "tokens_seen": 951910400 + }, + { + "epoch": 0.29, + "learning_rate": 0.0005749889669007021, + "loss": 1.3012, + "theoretical_loss": 3.694869417924603, + "tokens_seen": 951975936 + }, + { + "epoch": 0.29, + "learning_rate": 0.0005749729187562689, + "loss": 1.3613, + "theoretical_loss": 3.694845179617225, + "tokens_seen": 952041472 + }, + { + "epoch": 0.29, + "learning_rate": 0.0005749568706118355, + "loss": 1.5418, + "theoretical_loss": 3.6948209434454338, + "tokens_seen": 952107008 + }, + { + "epoch": 0.29, + "learning_rate": 0.0005749408224674023, + "loss": 1.3715, + "theoretical_loss": 3.6947967094088936, + "tokens_seen": 952172544 + }, + { + "epoch": 0.29, + "learning_rate": 0.0005749247743229689, + "loss": 1.5306, + "theoretical_loss": 3.6947724775072706, + "tokens_seen": 952238080 + }, + { + "epoch": 0.29, + "learning_rate": 0.0005749087261785357, + "loss": 1.406, + "theoretical_loss": 3.6947482477402285, + "tokens_seen": 952303616 + }, + { + "epoch": 0.29, + "learning_rate": 0.0005748926780341023, + "loss": 1.3842, + "theoretical_loss": 3.694724020107433, + "tokens_seen": 952369152 + }, + { + "epoch": 0.29, + "learning_rate": 0.0005748766298896691, + "loss": 1.5147, + "theoretical_loss": 3.6946997946085487, + "tokens_seen": 952434688 + }, + { + "epoch": 0.29, + "learning_rate": 0.0005748605817452357, + "loss": 1.4848, + "theoretical_loss": 3.694675571243242, + "tokens_seen": 952500224 + }, + { + "epoch": 0.29, + "learning_rate": 0.0005748445336008025, + "loss": 1.3889, + "theoretical_loss": 3.6946513500111777, + "tokens_seen": 952565760 + }, + { + "epoch": 0.29, + "learning_rate": 0.0005748284854563691, + "loss": 1.4269, + "theoretical_loss": 3.6946271309120204, + "tokens_seen": 952631296 + }, + { + "epoch": 0.29, + "learning_rate": 0.0005748124373119359, + "loss": 1.4414, + "theoretical_loss": 3.6946029139454364, + "tokens_seen": 952696832 + }, + { + "epoch": 0.29, + "learning_rate": 0.0005747963891675025, + "loss": 1.3439, + "theoretical_loss": 3.6945786991110907, + "tokens_seen": 952762368 + }, + { + "epoch": 0.29, + "learning_rate": 0.0005747803410230693, + "loss": 1.433, + "theoretical_loss": 3.6945544864086495, + "tokens_seen": 952827904 + }, + { + "epoch": 0.29, + "learning_rate": 0.0005747642928786359, + "loss": 1.4428, + "theoretical_loss": 3.6945302758377783, + "tokens_seen": 952893440 + }, + { + "epoch": 0.29, + "learning_rate": 0.0005747482447342027, + "loss": 1.4333, + "theoretical_loss": 3.6945060673981427, + "tokens_seen": 952958976 + }, + { + "epoch": 0.29, + "learning_rate": 0.0005747321965897693, + "loss": 1.3843, + "theoretical_loss": 3.6944818610894083, + "tokens_seen": 953024512 + }, + { + "epoch": 0.29, + "learning_rate": 0.0005747161484453361, + "loss": 1.3783, + "theoretical_loss": 3.6944576569112417, + "tokens_seen": 953090048 + }, + { + "epoch": 0.29, + "learning_rate": 0.0005747001003009027, + "loss": 1.4264, + "theoretical_loss": 3.694433454863309, + "tokens_seen": 953155584 + }, + { + "epoch": 0.29, + "learning_rate": 0.0005746840521564695, + "loss": 1.5075, + "theoretical_loss": 3.694409254945275, + "tokens_seen": 953221120 + }, + { + "epoch": 0.29, + "learning_rate": 0.0005746680040120361, + "loss": 1.5289, + "theoretical_loss": 3.694385057156807, + "tokens_seen": 953286656 + }, + { + "epoch": 0.29, + "learning_rate": 0.0005746519558676029, + "loss": 1.4518, + "theoretical_loss": 3.6943608614975707, + "tokens_seen": 953352192 + }, + { + "epoch": 0.29, + "learning_rate": 0.0005746359077231695, + "loss": 1.5403, + "theoretical_loss": 3.6943366679672325, + "tokens_seen": 953417728 + }, + { + "epoch": 0.29, + "learning_rate": 0.0005746198595787363, + "loss": 1.2568, + "theoretical_loss": 3.6943124765654587, + "tokens_seen": 953483264 + }, + { + "epoch": 0.29, + "objective/train/docs_used": 350619, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.31082022190094, + "objective/train/theoretical_loss": 3.694288287291916, + "objective/train/tokens_used": 974008800, + "theoretical_loss": 3.694288287291916, + "tokens_seen": 953548800 + }, + { + "epoch": 0.29, + "learning_rate": 0.0005746038114343029, + "loss": 1.3714, + "theoretical_loss": 3.694288287291916, + "tokens_seen": 953548800 + }, + { + "epoch": 0.29, + "learning_rate": 0.0005745877632898697, + "loss": 1.3842, + "theoretical_loss": 3.6942641001462713, + "tokens_seen": 953614336 + }, + { + "epoch": 0.29, + "learning_rate": 0.0005745717151454363, + "loss": 1.4531, + "theoretical_loss": 3.6942399151281897, + "tokens_seen": 953679872 + }, + { + "epoch": 0.29, + "learning_rate": 0.0005745556670010031, + "loss": 1.3734, + "theoretical_loss": 3.6942157322373395, + "tokens_seen": 953745408 + }, + { + "epoch": 0.29, + "learning_rate": 0.0005745396188565697, + "loss": 1.5878, + "theoretical_loss": 3.694191551473387, + "tokens_seen": 953810944 + }, + { + "epoch": 0.29, + "learning_rate": 0.0005745235707121365, + "loss": 1.4699, + "theoretical_loss": 3.694167372835998, + "tokens_seen": 953876480 + }, + { + "epoch": 0.29, + "learning_rate": 0.0005745075225677031, + "loss": 1.4537, + "theoretical_loss": 3.6941431963248403, + "tokens_seen": 953942016 + }, + { + "epoch": 0.29, + "learning_rate": 0.0005744914744232699, + "loss": 1.4121, + "theoretical_loss": 3.6941190219395814, + "tokens_seen": 954007552 + }, + { + "epoch": 0.29, + "learning_rate": 0.0005744754262788365, + "loss": 1.4619, + "theoretical_loss": 3.6940948496798875, + "tokens_seen": 954073088 + }, + { + "epoch": 0.29, + "learning_rate": 0.0005744593781344033, + "loss": 1.4103, + "theoretical_loss": 3.694070679545426, + "tokens_seen": 954138624 + }, + { + "epoch": 0.29, + "learning_rate": 0.0005744433299899699, + "loss": 1.3215, + "theoretical_loss": 3.6940465115358636, + "tokens_seen": 954204160 + }, + { + "epoch": 0.29, + "learning_rate": 0.0005744272818455367, + "loss": 1.429, + "theoretical_loss": 3.6940223456508683, + "tokens_seen": 954269696 + }, + { + "epoch": 0.29, + "learning_rate": 0.0005744112337011033, + "loss": 1.4953, + "theoretical_loss": 3.693998181890107, + "tokens_seen": 954335232 + }, + { + "epoch": 0.29, + "learning_rate": 0.0005743951855566701, + "loss": 1.3978, + "theoretical_loss": 3.693974020253248, + "tokens_seen": 954400768 + }, + { + "epoch": 0.29, + "learning_rate": 0.0005743791374122367, + "loss": 1.367, + "theoretical_loss": 3.6939498607399566, + "tokens_seen": 954466304 + }, + { + "epoch": 0.29, + "learning_rate": 0.0005743630892678035, + "loss": 1.3849, + "theoretical_loss": 3.693925703349903, + "tokens_seen": 954531840 + }, + { + "epoch": 0.29, + "learning_rate": 0.0005743470411233701, + "loss": 1.4282, + "theoretical_loss": 3.6939015480827533, + "tokens_seen": 954597376 + }, + { + "epoch": 0.29, + "learning_rate": 0.0005743309929789369, + "loss": 1.4245, + "theoretical_loss": 3.6938773949381756, + "tokens_seen": 954662912 + }, + { + "epoch": 0.29, + "learning_rate": 0.0005743149448345035, + "loss": 1.4652, + "theoretical_loss": 3.6938532439158376, + "tokens_seen": 954728448 + }, + { + "epoch": 0.29, + "learning_rate": 0.0005742988966900703, + "loss": 1.3979, + "theoretical_loss": 3.693829095015407, + "tokens_seen": 954793984 + }, + { + "epoch": 0.29, + "learning_rate": 0.0005742828485456369, + "loss": 1.4649, + "theoretical_loss": 3.6938049482365525, + "tokens_seen": 954859520 + }, + { + "epoch": 0.29, + "learning_rate": 0.0005742668004012037, + "loss": 1.456, + "theoretical_loss": 3.6937808035789415, + "tokens_seen": 954925056 + }, + { + "epoch": 0.29, + "learning_rate": 0.0005742507522567703, + "loss": 1.3679, + "theoretical_loss": 3.6937566610422423, + "tokens_seen": 954990592 + }, + { + "epoch": 0.29, + "learning_rate": 0.0005742347041123371, + "loss": 1.48, + "theoretical_loss": 3.6937325206261225, + "tokens_seen": 955056128 + }, + { + "epoch": 0.29, + "learning_rate": 0.0005742186559679037, + "loss": 1.304, + "theoretical_loss": 3.693708382330251, + "tokens_seen": 955121664 + }, + { + "epoch": 0.29, + "learning_rate": 0.0005742026078234705, + "loss": 1.5103, + "theoretical_loss": 3.6936842461542962, + "tokens_seen": 955187200 + }, + { + "epoch": 0.29, + "learning_rate": 0.0005741865596790371, + "loss": 1.509, + "theoretical_loss": 3.693660112097926, + "tokens_seen": 955252736 + }, + { + "epoch": 0.29, + "learning_rate": 0.0005741705115346039, + "loss": 1.3744, + "theoretical_loss": 3.693635980160809, + "tokens_seen": 955318272 + }, + { + "epoch": 0.29, + "learning_rate": 0.0005741544633901705, + "loss": 1.4294, + "theoretical_loss": 3.693611850342614, + "tokens_seen": 955383808 + }, + { + "epoch": 0.29, + "learning_rate": 0.0005741384152457373, + "loss": 1.486, + "theoretical_loss": 3.6935877226430094, + "tokens_seen": 955449344 + }, + { + "epoch": 0.29, + "learning_rate": 0.0005741223671013039, + "loss": 1.395, + "theoretical_loss": 3.6935635970616643, + "tokens_seen": 955514880 + }, + { + "epoch": 0.29, + "learning_rate": 0.0005741063189568707, + "loss": 1.3905, + "theoretical_loss": 3.6935394735982463, + "tokens_seen": 955580416 + }, + { + "epoch": 0.29, + "learning_rate": 0.0005740902708124373, + "loss": 1.4237, + "theoretical_loss": 3.6935153522524256, + "tokens_seen": 955645952 + }, + { + "epoch": 0.29, + "learning_rate": 0.000574074222668004, + "loss": 1.4386, + "theoretical_loss": 3.6934912330238703, + "tokens_seen": 955711488 + }, + { + "epoch": 0.29, + "learning_rate": 0.0005740581745235707, + "loss": 1.4784, + "theoretical_loss": 3.6934671159122496, + "tokens_seen": 955777024 + }, + { + "epoch": 0.29, + "learning_rate": 0.0005740421263791374, + "loss": 1.4799, + "theoretical_loss": 3.6934430009172328, + "tokens_seen": 955842560 + }, + { + "epoch": 0.29, + "learning_rate": 0.0005740260782347041, + "loss": 1.4647, + "theoretical_loss": 3.693418888038489, + "tokens_seen": 955908096 + }, + { + "epoch": 0.29, + "learning_rate": 0.0005740100300902708, + "loss": 1.3219, + "theoretical_loss": 3.693394777275687, + "tokens_seen": 955973632 + }, + { + "epoch": 0.29, + "learning_rate": 0.0005739939819458375, + "loss": 1.3569, + "theoretical_loss": 3.693370668628496, + "tokens_seen": 956039168 + }, + { + "epoch": 0.29, + "learning_rate": 0.0005739779338014042, + "loss": 1.4924, + "theoretical_loss": 3.693346562096586, + "tokens_seen": 956104704 + }, + { + "epoch": 0.29, + "learning_rate": 0.0005739618856569709, + "loss": 1.4578, + "theoretical_loss": 3.693322457679626, + "tokens_seen": 956170240 + }, + { + "epoch": 0.29, + "learning_rate": 0.0005739458375125376, + "loss": 1.5596, + "theoretical_loss": 3.693298355377286, + "tokens_seen": 956235776 + }, + { + "epoch": 0.29, + "learning_rate": 0.0005739297893681043, + "loss": 1.3819, + "theoretical_loss": 3.693274255189235, + "tokens_seen": 956301312 + }, + { + "epoch": 0.29, + "learning_rate": 0.000573913741223671, + "loss": 1.4245, + "theoretical_loss": 3.693250157115142, + "tokens_seen": 956366848 + }, + { + "epoch": 0.29, + "learning_rate": 0.0005738976930792377, + "loss": 1.486, + "theoretical_loss": 3.6932260611546788, + "tokens_seen": 956432384 + }, + { + "epoch": 0.29, + "learning_rate": 0.0005738816449348044, + "loss": 1.4792, + "theoretical_loss": 3.693201967307514, + "tokens_seen": 956497920 + }, + { + "epoch": 0.29, + "learning_rate": 0.0005738655967903711, + "loss": 1.6111, + "theoretical_loss": 3.6931778755733164, + "tokens_seen": 956563456 + }, + { + "epoch": 0.29, + "learning_rate": 0.0005738495486459378, + "loss": 1.4631, + "theoretical_loss": 3.6931537859517585, + "tokens_seen": 956628992 + }, + { + "epoch": 0.29, + "learning_rate": 0.0005738335005015045, + "loss": 1.4906, + "theoretical_loss": 3.6931296984425077, + "tokens_seen": 956694528 + }, + { + "epoch": 0.29, + "learning_rate": 0.0005738174523570712, + "loss": 1.4016, + "theoretical_loss": 3.693105613045236, + "tokens_seen": 956760064 + }, + { + "epoch": 0.29, + "objective/train/docs_used": 351696, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.454416036605835, + "objective/train/theoretical_loss": 3.693081529759612, + "objective/train/tokens_used": 977285600, + "theoretical_loss": 3.693081529759612, + "tokens_seen": 956825600 + }, + { + "epoch": 0.29, + "learning_rate": 0.000573801404212638, + "loss": 1.4284, + "theoretical_loss": 3.693081529759612, + "tokens_seen": 956825600 + }, + { + "epoch": 0.29, + "learning_rate": 0.0005737853560682046, + "loss": 1.4899, + "theoretical_loss": 3.6930574485853076, + "tokens_seen": 956891136 + }, + { + "epoch": 0.29, + "learning_rate": 0.0005737693079237713, + "loss": 1.5455, + "theoretical_loss": 3.6930333695219915, + "tokens_seen": 956956672 + }, + { + "epoch": 0.29, + "learning_rate": 0.000573753259779338, + "loss": 1.4289, + "theoretical_loss": 3.693009292569336, + "tokens_seen": 957022208 + }, + { + "epoch": 0.29, + "learning_rate": 0.0005737372116349047, + "loss": 1.437, + "theoretical_loss": 3.69298521772701, + "tokens_seen": 957087744 + }, + { + "epoch": 0.29, + "learning_rate": 0.0005737211634904714, + "loss": 1.463, + "theoretical_loss": 3.692961144994684, + "tokens_seen": 957153280 + }, + { + "epoch": 0.29, + "learning_rate": 0.0005737051153460382, + "loss": 1.5524, + "theoretical_loss": 3.6929370743720296, + "tokens_seen": 957218816 + }, + { + "epoch": 0.29, + "learning_rate": 0.0005736890672016048, + "loss": 1.5479, + "theoretical_loss": 3.692913005858717, + "tokens_seen": 957284352 + }, + { + "epoch": 0.29, + "learning_rate": 0.0005736730190571716, + "loss": 1.4213, + "theoretical_loss": 3.6928889394544173, + "tokens_seen": 957349888 + }, + { + "epoch": 0.29, + "learning_rate": 0.0005736569709127382, + "loss": 1.4198, + "theoretical_loss": 3.6928648751588007, + "tokens_seen": 957415424 + }, + { + "epoch": 0.29, + "learning_rate": 0.000573640922768305, + "loss": 1.4886, + "theoretical_loss": 3.6928408129715384, + "tokens_seen": 957480960 + }, + { + "epoch": 0.29, + "learning_rate": 0.0005736248746238716, + "loss": 1.4612, + "theoretical_loss": 3.692816752892302, + "tokens_seen": 957546496 + }, + { + "epoch": 0.29, + "learning_rate": 0.0005736088264794384, + "loss": 1.5139, + "theoretical_loss": 3.6927926949207617, + "tokens_seen": 957612032 + }, + { + "epoch": 0.29, + "learning_rate": 0.000573592778335005, + "loss": 1.4457, + "theoretical_loss": 3.692768639056589, + "tokens_seen": 957677568 + }, + { + "epoch": 0.29, + "learning_rate": 0.0005735767301905718, + "loss": 1.3406, + "theoretical_loss": 3.692744585299455, + "tokens_seen": 957743104 + }, + { + "epoch": 0.29, + "learning_rate": 0.0005735606820461384, + "loss": 1.4264, + "theoretical_loss": 3.692720533649031, + "tokens_seen": 957808640 + }, + { + "epoch": 0.29, + "learning_rate": 0.0005735446339017052, + "loss": 1.3875, + "theoretical_loss": 3.6926964841049887, + "tokens_seen": 957874176 + }, + { + "epoch": 0.29, + "learning_rate": 0.0005735285857572718, + "loss": 1.5288, + "theoretical_loss": 3.692672436666999, + "tokens_seen": 957939712 + }, + { + "epoch": 0.29, + "learning_rate": 0.0005735125376128386, + "loss": 1.3943, + "theoretical_loss": 3.6926483913347337, + "tokens_seen": 958005248 + }, + { + "epoch": 0.29, + "learning_rate": 0.0005734964894684052, + "loss": 1.4473, + "theoretical_loss": 3.692624348107864, + "tokens_seen": 958070784 + }, + { + "epoch": 0.29, + "learning_rate": 0.000573480441323972, + "loss": 1.5039, + "theoretical_loss": 3.692600306986063, + "tokens_seen": 958136320 + }, + { + "epoch": 0.29, + "learning_rate": 0.0005734643931795386, + "loss": 1.4762, + "theoretical_loss": 3.692576267969, + "tokens_seen": 958201856 + }, + { + "epoch": 0.29, + "learning_rate": 0.0005734483450351054, + "loss": 1.4997, + "theoretical_loss": 3.6925522310563488, + "tokens_seen": 958267392 + }, + { + "epoch": 0.29, + "learning_rate": 0.000573432296890672, + "loss": 1.5156, + "theoretical_loss": 3.6925281962477805, + "tokens_seen": 958332928 + }, + { + "epoch": 0.29, + "learning_rate": 0.0005734162487462388, + "loss": 1.4622, + "theoretical_loss": 3.692504163542967, + "tokens_seen": 958398464 + }, + { + "epoch": 0.29, + "learning_rate": 0.0005734002006018054, + "loss": 1.4284, + "theoretical_loss": 3.692480132941581, + "tokens_seen": 958464000 + }, + { + "epoch": 0.29, + "learning_rate": 0.0005733841524573722, + "loss": 1.4914, + "theoretical_loss": 3.6924561044432935, + "tokens_seen": 958529536 + }, + { + "epoch": 0.29, + "learning_rate": 0.0005733681043129388, + "loss": 1.3867, + "theoretical_loss": 3.6924320780477773, + "tokens_seen": 958595072 + }, + { + "epoch": 0.29, + "learning_rate": 0.0005733520561685056, + "loss": 1.63, + "theoretical_loss": 3.692408053754705, + "tokens_seen": 958660608 + }, + { + "epoch": 0.29, + "learning_rate": 0.0005733360080240722, + "loss": 1.4444, + "theoretical_loss": 3.6923840315637477, + "tokens_seen": 958726144 + }, + { + "epoch": 0.29, + "learning_rate": 0.000573319959879639, + "loss": 1.5202, + "theoretical_loss": 3.6923600114745794, + "tokens_seen": 958791680 + }, + { + "epoch": 0.29, + "learning_rate": 0.0005733039117352056, + "loss": 1.4335, + "theoretical_loss": 3.6923359934868714, + "tokens_seen": 958857216 + }, + { + "epoch": 0.29, + "learning_rate": 0.0005732878635907724, + "loss": 1.4262, + "theoretical_loss": 3.692311977600297, + "tokens_seen": 958922752 + }, + { + "epoch": 0.29, + "learning_rate": 0.000573271815446339, + "loss": 1.4679, + "theoretical_loss": 3.6922879638145276, + "tokens_seen": 958988288 + }, + { + "epoch": 0.29, + "learning_rate": 0.0005732557673019058, + "loss": 1.411, + "theoretical_loss": 3.6922639521292373, + "tokens_seen": 959053824 + }, + { + "epoch": 0.29, + "learning_rate": 0.0005732397191574724, + "loss": 1.4017, + "theoretical_loss": 3.6922399425440977, + "tokens_seen": 959119360 + }, + { + "epoch": 0.29, + "learning_rate": 0.0005732397191574724, + "loss": 1.439, + "theoretical_loss": 3.6922159350587824, + "tokens_seen": 959184896 + }, + { + "epoch": 0.29, + "learning_rate": 0.0005732236710130392, + "loss": 1.4434, + "theoretical_loss": 3.692191929672964, + "tokens_seen": 959250432 + }, + { + "epoch": 0.29, + "learning_rate": 0.0005732076228686058, + "loss": 1.4805, + "theoretical_loss": 3.6921679263863156, + "tokens_seen": 959315968 + }, + { + "epoch": 0.29, + "learning_rate": 0.0005731915747241726, + "loss": 1.2954, + "theoretical_loss": 3.6921439251985104, + "tokens_seen": 959381504 + }, + { + "epoch": 0.29, + "learning_rate": 0.0005731755265797392, + "loss": 1.4674, + "theoretical_loss": 3.6921199261092212, + "tokens_seen": 959447040 + }, + { + "epoch": 0.29, + "learning_rate": 0.000573159478435306, + "loss": 1.369, + "theoretical_loss": 3.6920959291181212, + "tokens_seen": 959512576 + }, + { + "epoch": 0.29, + "learning_rate": 0.0005731434302908726, + "loss": 1.5647, + "theoretical_loss": 3.692071934224884, + "tokens_seen": 959578112 + }, + { + "epoch": 0.29, + "learning_rate": 0.0005731273821464394, + "loss": 1.4062, + "theoretical_loss": 3.6920479414291822, + "tokens_seen": 959643648 + }, + { + "epoch": 0.29, + "learning_rate": 0.000573111334002006, + "loss": 1.4974, + "theoretical_loss": 3.6920239507306905, + "tokens_seen": 959709184 + }, + { + "epoch": 0.29, + "learning_rate": 0.0005730952858575728, + "loss": 1.5062, + "theoretical_loss": 3.6919999621290813, + "tokens_seen": 959774720 + }, + { + "epoch": 0.29, + "learning_rate": 0.0005730792377131394, + "loss": 1.5296, + "theoretical_loss": 3.691975975624028, + "tokens_seen": 959840256 + }, + { + "epoch": 0.29, + "learning_rate": 0.0005730631895687062, + "loss": 1.537, + "theoretical_loss": 3.6919519912152055, + "tokens_seen": 959905792 + }, + { + "epoch": 0.29, + "learning_rate": 0.0005730471414242728, + "loss": 1.4099, + "theoretical_loss": 3.691928008902286, + "tokens_seen": 959971328 + }, + { + "epoch": 0.29, + "learning_rate": 0.0005730310932798396, + "loss": 1.4994, + "theoretical_loss": 3.6919040286849443, + "tokens_seen": 960036864 + }, + { + "epoch": 0.29, + "objective/train/docs_used": 352874, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.3856309652328491, + "objective/train/theoretical_loss": 3.6918800505628546, + "objective/train/tokens_used": 980562400, + "theoretical_loss": 3.6918800505628546, + "tokens_seen": 960102400 + }, + { + "epoch": 0.29, + "learning_rate": 0.0005730150451354062, + "loss": 1.4205, + "theoretical_loss": 3.6918800505628546, + "tokens_seen": 960102400 + }, + { + "epoch": 0.29, + "learning_rate": 0.000572998996990973, + "loss": 1.4619, + "theoretical_loss": 3.691856074535689, + "tokens_seen": 960167936 + }, + { + "epoch": 0.29, + "learning_rate": 0.0005729829488465396, + "loss": 1.4306, + "theoretical_loss": 3.691832100603124, + "tokens_seen": 960233472 + }, + { + "epoch": 0.29, + "learning_rate": 0.0005729669007021064, + "loss": 1.5157, + "theoretical_loss": 3.691808128764831, + "tokens_seen": 960299008 + }, + { + "epoch": 0.29, + "learning_rate": 0.0005729508525576731, + "loss": 1.4509, + "theoretical_loss": 3.6917841590204867, + "tokens_seen": 960364544 + }, + { + "epoch": 0.29, + "learning_rate": 0.0005729348044132398, + "loss": 1.4152, + "theoretical_loss": 3.6917601913697635, + "tokens_seen": 960430080 + }, + { + "epoch": 0.29, + "learning_rate": 0.0005729187562688065, + "loss": 1.3683, + "theoretical_loss": 3.6917362258123365, + "tokens_seen": 960495616 + }, + { + "epoch": 0.29, + "learning_rate": 0.0005729027081243732, + "loss": 1.3384, + "theoretical_loss": 3.69171226234788, + "tokens_seen": 960561152 + }, + { + "epoch": 0.29, + "learning_rate": 0.0005728866599799399, + "loss": 1.3997, + "theoretical_loss": 3.691688300976068, + "tokens_seen": 960626688 + }, + { + "epoch": 0.29, + "learning_rate": 0.0005728706118355066, + "loss": 1.4881, + "theoretical_loss": 3.6916643416965753, + "tokens_seen": 960692224 + }, + { + "epoch": 0.29, + "learning_rate": 0.0005728545636910733, + "loss": 1.4364, + "theoretical_loss": 3.6916403845090766, + "tokens_seen": 960757760 + }, + { + "epoch": 0.29, + "learning_rate": 0.00057283851554664, + "loss": 1.3555, + "theoretical_loss": 3.691616429413247, + "tokens_seen": 960823296 + }, + { + "epoch": 0.29, + "learning_rate": 0.0005728224674022067, + "loss": 1.4334, + "theoretical_loss": 3.6915924764087604, + "tokens_seen": 960888832 + }, + { + "epoch": 0.29, + "learning_rate": 0.0005728064192577734, + "loss": 1.4075, + "theoretical_loss": 3.6915685254952915, + "tokens_seen": 960954368 + }, + { + "epoch": 0.29, + "learning_rate": 0.00057279037111334, + "loss": 1.4101, + "theoretical_loss": 3.6915445766725163, + "tokens_seen": 961019904 + }, + { + "epoch": 0.29, + "learning_rate": 0.0005727743229689068, + "loss": 1.4703, + "theoretical_loss": 3.691520629940108, + "tokens_seen": 961085440 + }, + { + "epoch": 0.29, + "learning_rate": 0.0005727582748244734, + "loss": 1.4502, + "theoretical_loss": 3.6914966852977438, + "tokens_seen": 961150976 + }, + { + "epoch": 0.29, + "learning_rate": 0.0005727422266800402, + "loss": 1.3886, + "theoretical_loss": 3.691472742745097, + "tokens_seen": 961216512 + }, + { + "epoch": 0.29, + "learning_rate": 0.0005727261785356068, + "loss": 1.4952, + "theoretical_loss": 3.6914488022818435, + "tokens_seen": 961282048 + }, + { + "epoch": 0.29, + "learning_rate": 0.0005727101303911736, + "loss": 1.4772, + "theoretical_loss": 3.691424863907659, + "tokens_seen": 961347584 + }, + { + "epoch": 0.29, + "learning_rate": 0.0005726940822467402, + "loss": 1.4017, + "theoretical_loss": 3.691400927622217, + "tokens_seen": 961413120 + }, + { + "epoch": 0.29, + "learning_rate": 0.0005726780341023069, + "loss": 1.3931, + "theoretical_loss": 3.6913769934251954, + "tokens_seen": 961478656 + }, + { + "epoch": 0.29, + "learning_rate": 0.0005726619859578736, + "loss": 1.316, + "theoretical_loss": 3.691353061316268, + "tokens_seen": 961544192 + }, + { + "epoch": 0.29, + "learning_rate": 0.0005726459378134403, + "loss": 1.4409, + "theoretical_loss": 3.6913291312951104, + "tokens_seen": 961609728 + }, + { + "epoch": 0.29, + "learning_rate": 0.000572629889669007, + "loss": 1.4704, + "theoretical_loss": 3.6913052033613987, + "tokens_seen": 961675264 + }, + { + "epoch": 0.29, + "learning_rate": 0.0005726138415245737, + "loss": 1.4128, + "theoretical_loss": 3.6912812775148085, + "tokens_seen": 961740800 + }, + { + "epoch": 0.29, + "learning_rate": 0.0005725977933801404, + "loss": 1.4847, + "theoretical_loss": 3.691257353755015, + "tokens_seen": 961806336 + }, + { + "epoch": 0.29, + "learning_rate": 0.0005725817452357071, + "loss": 1.4557, + "theoretical_loss": 3.691233432081695, + "tokens_seen": 961871872 + }, + { + "epoch": 0.29, + "learning_rate": 0.0005725656970912738, + "loss": 1.512, + "theoretical_loss": 3.691209512494524, + "tokens_seen": 961937408 + }, + { + "epoch": 0.29, + "learning_rate": 0.0005725496489468405, + "loss": 1.5647, + "theoretical_loss": 3.6911855949931773, + "tokens_seen": 962002944 + }, + { + "epoch": 0.29, + "learning_rate": 0.0005725336008024072, + "loss": 1.3962, + "theoretical_loss": 3.6911616795773314, + "tokens_seen": 962068480 + }, + { + "epoch": 0.29, + "learning_rate": 0.0005725175526579739, + "loss": 1.448, + "theoretical_loss": 3.691137766246663, + "tokens_seen": 962134016 + }, + { + "epoch": 0.29, + "learning_rate": 0.0005725015045135406, + "loss": 1.5024, + "theoretical_loss": 3.6911138550008475, + "tokens_seen": 962199552 + }, + { + "epoch": 0.29, + "learning_rate": 0.0005724854563691073, + "loss": 1.476, + "theoretical_loss": 3.691089945839561, + "tokens_seen": 962265088 + }, + { + "epoch": 0.29, + "learning_rate": 0.000572469408224674, + "loss": 1.4355, + "theoretical_loss": 3.6910660387624805, + "tokens_seen": 962330624 + }, + { + "epoch": 0.29, + "learning_rate": 0.0005724533600802407, + "loss": 1.3754, + "theoretical_loss": 3.6910421337692823, + "tokens_seen": 962396160 + }, + { + "epoch": 0.29, + "learning_rate": 0.0005724373119358074, + "loss": 1.4292, + "theoretical_loss": 3.6910182308596426, + "tokens_seen": 962461696 + }, + { + "epoch": 0.29, + "learning_rate": 0.0005724212637913741, + "loss": 1.4871, + "theoretical_loss": 3.6909943300332384, + "tokens_seen": 962527232 + }, + { + "epoch": 0.29, + "learning_rate": 0.0005724052156469408, + "loss": 1.4818, + "theoretical_loss": 3.6909704312897453, + "tokens_seen": 962592768 + }, + { + "epoch": 0.29, + "learning_rate": 0.0005723891675025075, + "loss": 1.3522, + "theoretical_loss": 3.690946534628841, + "tokens_seen": 962658304 + }, + { + "epoch": 0.29, + "learning_rate": 0.0005723731193580742, + "loss": 1.3328, + "theoretical_loss": 3.6909226400502018, + "tokens_seen": 962723840 + }, + { + "epoch": 0.29, + "learning_rate": 0.0005723570712136409, + "loss": 1.3993, + "theoretical_loss": 3.6908987475535047, + "tokens_seen": 962789376 + }, + { + "epoch": 0.29, + "learning_rate": 0.0005723410230692077, + "loss": 1.4697, + "theoretical_loss": 3.690874857138426, + "tokens_seen": 962854912 + }, + { + "epoch": 0.29, + "learning_rate": 0.0005723249749247743, + "loss": 1.4691, + "theoretical_loss": 3.690850968804644, + "tokens_seen": 962920448 + }, + { + "epoch": 0.29, + "learning_rate": 0.000572308926780341, + "loss": 1.3377, + "theoretical_loss": 3.6908270825518343, + "tokens_seen": 962985984 + }, + { + "epoch": 0.29, + "learning_rate": 0.0005722928786359077, + "loss": 1.459, + "theoretical_loss": 3.690803198379675, + "tokens_seen": 963051520 + }, + { + "epoch": 0.29, + "learning_rate": 0.0005722768304914745, + "loss": 1.3442, + "theoretical_loss": 3.690779316287843, + "tokens_seen": 963117056 + }, + { + "epoch": 0.29, + "learning_rate": 0.0005722607823470411, + "loss": 1.3622, + "theoretical_loss": 3.6907554362760155, + "tokens_seen": 963182592 + }, + { + "epoch": 0.29, + "learning_rate": 0.0005722447342026079, + "loss": 1.4275, + "theoretical_loss": 3.6907315583438693, + "tokens_seen": 963248128 + }, + { + "epoch": 0.29, + "learning_rate": 0.0005722286860581745, + "loss": 1.5072, + "theoretical_loss": 3.690707682491083, + "tokens_seen": 963313664 + }, + { + "epoch": 0.29, + "objective/train/docs_used": 353969, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.3799344301223755, + "objective/train/theoretical_loss": 3.690683808717333, + "objective/train/tokens_used": 983839200, + "theoretical_loss": 3.690683808717333, + "tokens_seen": 963379200 + }, + { + "epoch": 0.29, + "learning_rate": 0.0005722126379137413, + "loss": 1.3434, + "theoretical_loss": 3.690683808717333, + "tokens_seen": 963379200 + }, + { + "epoch": 0.29, + "learning_rate": 0.0005721965897693079, + "loss": 1.4436, + "theoretical_loss": 3.6906599370222977, + "tokens_seen": 963444736 + }, + { + "epoch": 0.29, + "learning_rate": 0.0005721805416248747, + "loss": 1.5964, + "theoretical_loss": 3.690636067405654, + "tokens_seen": 963510272 + }, + { + "epoch": 0.29, + "learning_rate": 0.0005721644934804413, + "loss": 1.4023, + "theoretical_loss": 3.69061219986708, + "tokens_seen": 963575808 + }, + { + "epoch": 0.29, + "learning_rate": 0.0005721484453360081, + "loss": 1.4017, + "theoretical_loss": 3.6905883344062533, + "tokens_seen": 963641344 + }, + { + "epoch": 0.29, + "learning_rate": 0.0005721323971915747, + "loss": 1.3725, + "theoretical_loss": 3.690564471022852, + "tokens_seen": 963706880 + }, + { + "epoch": 0.29, + "learning_rate": 0.0005721163490471415, + "loss": 1.3628, + "theoretical_loss": 3.690540609716554, + "tokens_seen": 963772416 + }, + { + "epoch": 0.29, + "learning_rate": 0.0005721003009027081, + "loss": 1.3964, + "theoretical_loss": 3.6905167504870366, + "tokens_seen": 963837952 + }, + { + "epoch": 0.29, + "learning_rate": 0.0005720842527582749, + "loss": 1.4062, + "theoretical_loss": 3.6904928933339787, + "tokens_seen": 963903488 + }, + { + "epoch": 0.29, + "learning_rate": 0.0005720682046138415, + "loss": 1.5363, + "theoretical_loss": 3.690469038257058, + "tokens_seen": 963969024 + }, + { + "epoch": 0.29, + "learning_rate": 0.0005720521564694083, + "loss": 1.3797, + "theoretical_loss": 3.6904451852559528, + "tokens_seen": 964034560 + }, + { + "epoch": 0.29, + "learning_rate": 0.0005720361083249749, + "loss": 1.5227, + "theoretical_loss": 3.6904213343303414, + "tokens_seen": 964100096 + }, + { + "epoch": 0.29, + "learning_rate": 0.0005720200601805417, + "loss": 1.3964, + "theoretical_loss": 3.690397485479902, + "tokens_seen": 964165632 + }, + { + "epoch": 0.29, + "learning_rate": 0.0005720040120361083, + "loss": 1.4231, + "theoretical_loss": 3.690373638704313, + "tokens_seen": 964231168 + }, + { + "epoch": 0.29, + "learning_rate": 0.0005719879638916751, + "loss": 1.4538, + "theoretical_loss": 3.6903497940032532, + "tokens_seen": 964296704 + }, + { + "epoch": 0.29, + "learning_rate": 0.0005719719157472417, + "loss": 1.2996, + "theoretical_loss": 3.690325951376401, + "tokens_seen": 964362240 + }, + { + "epoch": 0.29, + "learning_rate": 0.0005719558676028085, + "loss": 1.4673, + "theoretical_loss": 3.690302110823435, + "tokens_seen": 964427776 + }, + { + "epoch": 0.29, + "learning_rate": 0.0005719398194583751, + "loss": 1.5453, + "theoretical_loss": 3.6902782723440337, + "tokens_seen": 964493312 + }, + { + "epoch": 0.29, + "learning_rate": 0.0005719237713139419, + "loss": 1.4513, + "theoretical_loss": 3.6902544359378755, + "tokens_seen": 964558848 + }, + { + "epoch": 0.29, + "learning_rate": 0.0005719077231695085, + "loss": 1.3866, + "theoretical_loss": 3.6902306016046404, + "tokens_seen": 964624384 + }, + { + "epoch": 0.29, + "learning_rate": 0.0005718916750250753, + "loss": 1.5324, + "theoretical_loss": 3.690206769344007, + "tokens_seen": 964689920 + }, + { + "epoch": 0.29, + "learning_rate": 0.0005718756268806419, + "loss": 1.3776, + "theoretical_loss": 3.6901829391556538, + "tokens_seen": 964755456 + }, + { + "epoch": 0.29, + "learning_rate": 0.0005718595787362087, + "loss": 1.5809, + "theoretical_loss": 3.69015911103926, + "tokens_seen": 964820992 + }, + { + "epoch": 0.29, + "learning_rate": 0.0005718435305917753, + "loss": 1.3577, + "theoretical_loss": 3.690135284994504, + "tokens_seen": 964886528 + }, + { + "epoch": 0.29, + "learning_rate": 0.0005718274824473421, + "loss": 1.4772, + "theoretical_loss": 3.6901114610210666, + "tokens_seen": 964952064 + }, + { + "epoch": 0.29, + "learning_rate": 0.0005718114343029087, + "loss": 1.4076, + "theoretical_loss": 3.6900876391186266, + "tokens_seen": 965017600 + }, + { + "epoch": 0.29, + "learning_rate": 0.0005717953861584755, + "loss": 1.3903, + "theoretical_loss": 3.6900638192868627, + "tokens_seen": 965083136 + }, + { + "epoch": 0.29, + "learning_rate": 0.0005717793380140421, + "loss": 1.3292, + "theoretical_loss": 3.690040001525454, + "tokens_seen": 965148672 + }, + { + "epoch": 0.29, + "learning_rate": 0.0005717632898696089, + "loss": 1.4912, + "theoretical_loss": 3.6900161858340814, + "tokens_seen": 965214208 + }, + { + "epoch": 0.29, + "learning_rate": 0.0005717472417251756, + "loss": 1.3065, + "theoretical_loss": 3.689992372212423, + "tokens_seen": 965279744 + }, + { + "epoch": 0.29, + "learning_rate": 0.0005717311935807423, + "loss": 1.5463, + "theoretical_loss": 3.6899685606601595, + "tokens_seen": 965345280 + }, + { + "epoch": 0.29, + "learning_rate": 0.000571715145436309, + "loss": 1.5577, + "theoretical_loss": 3.68994475117697, + "tokens_seen": 965410816 + }, + { + "epoch": 0.29, + "learning_rate": 0.0005716990972918757, + "loss": 1.4583, + "theoretical_loss": 3.6899209437625347, + "tokens_seen": 965476352 + }, + { + "epoch": 0.29, + "learning_rate": 0.0005716830491474424, + "loss": 1.3894, + "theoretical_loss": 3.6898971384165327, + "tokens_seen": 965541888 + }, + { + "epoch": 0.29, + "learning_rate": 0.0005716670010030091, + "loss": 1.3847, + "theoretical_loss": 3.6898733351386444, + "tokens_seen": 965607424 + }, + { + "epoch": 0.29, + "learning_rate": 0.0005716509528585758, + "loss": 1.4011, + "theoretical_loss": 3.68984953392855, + "tokens_seen": 965672960 + }, + { + "epoch": 0.29, + "learning_rate": 0.0005716349047141425, + "loss": 1.4036, + "theoretical_loss": 3.68982573478593, + "tokens_seen": 965738496 + }, + { + "epoch": 0.29, + "learning_rate": 0.0005716188565697092, + "loss": 1.4086, + "theoretical_loss": 3.689801937710463, + "tokens_seen": 965804032 + }, + { + "epoch": 0.29, + "learning_rate": 0.0005716028084252759, + "loss": 1.5143, + "theoretical_loss": 3.6897781427018304, + "tokens_seen": 965869568 + }, + { + "epoch": 0.29, + "learning_rate": 0.0005715867602808426, + "loss": 1.4981, + "theoretical_loss": 3.6897543497597125, + "tokens_seen": 965935104 + }, + { + "epoch": 0.29, + "learning_rate": 0.0005715707121364093, + "loss": 1.4238, + "theoretical_loss": 3.689730558883789, + "tokens_seen": 966000640 + }, + { + "epoch": 0.29, + "learning_rate": 0.000571554663991976, + "loss": 1.4629, + "theoretical_loss": 3.68970677007374, + "tokens_seen": 966066176 + }, + { + "epoch": 0.29, + "learning_rate": 0.0005715386158475427, + "loss": 1.2999, + "theoretical_loss": 3.6896829833292477, + "tokens_seen": 966131712 + }, + { + "epoch": 0.29, + "learning_rate": 0.0005715225677031094, + "loss": 1.4996, + "theoretical_loss": 3.6896591986499905, + "tokens_seen": 966197248 + }, + { + "epoch": 0.29, + "learning_rate": 0.0005715065195586761, + "loss": 1.4081, + "theoretical_loss": 3.689635416035651, + "tokens_seen": 966262784 + }, + { + "epoch": 0.29, + "learning_rate": 0.0005714904714142428, + "loss": 1.4387, + "theoretical_loss": 3.6896116354859085, + "tokens_seen": 966328320 + }, + { + "epoch": 0.29, + "learning_rate": 0.0005714744232698095, + "loss": 1.5398, + "theoretical_loss": 3.6895878570004443, + "tokens_seen": 966393856 + }, + { + "epoch": 0.29, + "learning_rate": 0.0005714583751253762, + "loss": 1.4356, + "theoretical_loss": 3.689564080578939, + "tokens_seen": 966459392 + }, + { + "epoch": 0.29, + "learning_rate": 0.0005714423269809429, + "loss": 1.4611, + "theoretical_loss": 3.689540306221074, + "tokens_seen": 966524928 + }, + { + "epoch": 0.29, + "learning_rate": 0.0005714262788365096, + "loss": 1.4262, + "theoretical_loss": 3.6895165339265295, + "tokens_seen": 966590464 + }, + { + "epoch": 0.29, + "objective/train/docs_used": 355190, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.517873764038086, + "objective/train/theoretical_loss": 3.6894927636949877, + "objective/train/tokens_used": 987116000, + "theoretical_loss": 3.6894927636949877, + "tokens_seen": 966656000 + }, + { + "epoch": 0.29, + "learning_rate": 0.0005714102306920763, + "loss": 1.5305, + "theoretical_loss": 3.6894927636949877, + "tokens_seen": 966656000 + }, + { + "epoch": 0.29, + "learning_rate": 0.0005713941825476429, + "loss": 1.4807, + "theoretical_loss": 3.689468995526128, + "tokens_seen": 966721536 + }, + { + "epoch": 0.29, + "learning_rate": 0.0005713781344032097, + "loss": 1.5416, + "theoretical_loss": 3.6894452294196336, + "tokens_seen": 966787072 + }, + { + "epoch": 0.29, + "learning_rate": 0.0005713620862587763, + "loss": 1.4505, + "theoretical_loss": 3.6894214653751845, + "tokens_seen": 966852608 + }, + { + "epoch": 0.29, + "learning_rate": 0.0005713460381143431, + "loss": 1.4067, + "theoretical_loss": 3.689397703392462, + "tokens_seen": 966918144 + }, + { + "epoch": 0.29, + "learning_rate": 0.0005713299899699097, + "loss": 1.3868, + "theoretical_loss": 3.689373943471148, + "tokens_seen": 966983680 + }, + { + "epoch": 0.29, + "learning_rate": 0.0005713139418254765, + "loss": 1.3536, + "theoretical_loss": 3.689350185610924, + "tokens_seen": 967049216 + }, + { + "epoch": 0.29, + "learning_rate": 0.0005712978936810431, + "loss": 1.3633, + "theoretical_loss": 3.6893264298114716, + "tokens_seen": 967114752 + }, + { + "epoch": 0.29, + "learning_rate": 0.0005712818455366099, + "loss": 1.4205, + "theoretical_loss": 3.689302676072472, + "tokens_seen": 967180288 + }, + { + "epoch": 0.29, + "learning_rate": 0.0005712657973921765, + "loss": 1.3815, + "theoretical_loss": 3.689278924393607, + "tokens_seen": 967245824 + }, + { + "epoch": 0.29, + "learning_rate": 0.0005712497492477432, + "loss": 1.4313, + "theoretical_loss": 3.689255174774558, + "tokens_seen": 967311360 + }, + { + "epoch": 0.29, + "learning_rate": 0.0005712337011033099, + "loss": 1.3955, + "theoretical_loss": 3.6892314272150077, + "tokens_seen": 967376896 + }, + { + "epoch": 0.29, + "learning_rate": 0.0005712176529588766, + "loss": 1.3629, + "theoretical_loss": 3.6892076817146373, + "tokens_seen": 967442432 + }, + { + "epoch": 0.29, + "learning_rate": 0.0005712016048144433, + "loss": 1.3996, + "theoretical_loss": 3.6891839382731297, + "tokens_seen": 967507968 + }, + { + "epoch": 0.29, + "learning_rate": 0.00057118555667001, + "loss": 1.4383, + "theoretical_loss": 3.6891601968901657, + "tokens_seen": 967573504 + }, + { + "epoch": 0.29, + "learning_rate": 0.0005711695085255767, + "loss": 1.351, + "theoretical_loss": 3.689136457565428, + "tokens_seen": 967639040 + }, + { + "epoch": 0.29, + "learning_rate": 0.0005711534603811434, + "loss": 1.4196, + "theoretical_loss": 3.689112720298599, + "tokens_seen": 967704576 + }, + { + "epoch": 0.29, + "learning_rate": 0.0005711374122367101, + "loss": 1.4854, + "theoretical_loss": 3.6890889850893602, + "tokens_seen": 967770112 + }, + { + "epoch": 0.29, + "learning_rate": 0.0005711213640922768, + "loss": 1.4806, + "theoretical_loss": 3.689065251937395, + "tokens_seen": 967835648 + }, + { + "epoch": 0.29, + "learning_rate": 0.0005711053159478435, + "loss": 1.4158, + "theoretical_loss": 3.6890415208423857, + "tokens_seen": 967901184 + }, + { + "epoch": 0.29, + "learning_rate": 0.0005710892678034102, + "loss": 1.3703, + "theoretical_loss": 3.6890177918040132, + "tokens_seen": 967966720 + }, + { + "epoch": 0.29, + "learning_rate": 0.000571073219658977, + "loss": 1.4153, + "theoretical_loss": 3.688994064821962, + "tokens_seen": 968032256 + }, + { + "epoch": 0.29, + "learning_rate": 0.0005710571715145436, + "loss": 1.4977, + "theoretical_loss": 3.6889703398959135, + "tokens_seen": 968097792 + }, + { + "epoch": 0.29, + "learning_rate": 0.0005710411233701103, + "loss": 1.4056, + "theoretical_loss": 3.6889466170255503, + "tokens_seen": 968163328 + }, + { + "epoch": 0.29, + "learning_rate": 0.000571025075225677, + "loss": 1.522, + "theoretical_loss": 3.6889228962105562, + "tokens_seen": 968228864 + }, + { + "epoch": 0.29, + "learning_rate": 0.0005710090270812437, + "loss": 1.4319, + "theoretical_loss": 3.6888991774506135, + "tokens_seen": 968294400 + }, + { + "epoch": 0.29, + "learning_rate": 0.0005709929789368104, + "loss": 1.4186, + "theoretical_loss": 3.688875460745405, + "tokens_seen": 968359936 + }, + { + "epoch": 0.29, + "learning_rate": 0.0005709769307923772, + "loss": 1.4865, + "theoretical_loss": 3.6888517460946133, + "tokens_seen": 968425472 + }, + { + "epoch": 0.29, + "learning_rate": 0.0005709608826479438, + "loss": 1.2791, + "theoretical_loss": 3.6888280334979218, + "tokens_seen": 968491008 + }, + { + "epoch": 0.29, + "learning_rate": 0.0005709448345035106, + "loss": 1.4725, + "theoretical_loss": 3.688804322955014, + "tokens_seen": 968556544 + }, + { + "epoch": 0.29, + "learning_rate": 0.0005709287863590772, + "loss": 1.4177, + "theoretical_loss": 3.688780614465572, + "tokens_seen": 968622080 + }, + { + "epoch": 0.29, + "learning_rate": 0.000570912738214644, + "loss": 1.41, + "theoretical_loss": 3.68875690802928, + "tokens_seen": 968687616 + }, + { + "epoch": 0.29, + "learning_rate": 0.0005708966900702106, + "loss": 1.33, + "theoretical_loss": 3.688733203645821, + "tokens_seen": 968753152 + }, + { + "epoch": 0.29, + "learning_rate": 0.0005708806419257774, + "loss": 1.4959, + "theoretical_loss": 3.6887095013148787, + "tokens_seen": 968818688 + }, + { + "epoch": 0.29, + "learning_rate": 0.000570864593781344, + "loss": 1.4716, + "theoretical_loss": 3.6886858010361356, + "tokens_seen": 968884224 + }, + { + "epoch": 0.29, + "learning_rate": 0.0005708485456369108, + "loss": 1.5178, + "theoretical_loss": 3.6886621028092765, + "tokens_seen": 968949760 + }, + { + "epoch": 0.29, + "learning_rate": 0.0005708324974924774, + "loss": 1.4643, + "theoretical_loss": 3.688638406633984, + "tokens_seen": 969015296 + }, + { + "epoch": 0.29, + "learning_rate": 0.0005708164493480442, + "loss": 1.3101, + "theoretical_loss": 3.688614712509942, + "tokens_seen": 969080832 + }, + { + "epoch": 0.29, + "learning_rate": 0.0005708004012036108, + "loss": 1.3136, + "theoretical_loss": 3.688591020436835, + "tokens_seen": 969146368 + }, + { + "epoch": 0.29, + "learning_rate": 0.0005707843530591776, + "loss": 1.5083, + "theoretical_loss": 3.6885673304143456, + "tokens_seen": 969211904 + }, + { + "epoch": 0.29, + "learning_rate": 0.0005707683049147442, + "loss": 1.4316, + "theoretical_loss": 3.6885436424421583, + "tokens_seen": 969277440 + }, + { + "epoch": 0.29, + "learning_rate": 0.000570752256770311, + "loss": 1.4495, + "theoretical_loss": 3.688519956519957, + "tokens_seen": 969342976 + }, + { + "epoch": 0.29, + "learning_rate": 0.0005707362086258776, + "loss": 1.4487, + "theoretical_loss": 3.6884962726474253, + "tokens_seen": 969408512 + }, + { + "epoch": 0.29, + "learning_rate": 0.0005707201604814444, + "loss": 1.4832, + "theoretical_loss": 3.688472590824248, + "tokens_seen": 969474048 + }, + { + "epoch": 0.29, + "learning_rate": 0.000570704112337011, + "loss": 1.4143, + "theoretical_loss": 3.6884489110501093, + "tokens_seen": 969539584 + }, + { + "epoch": 0.29, + "learning_rate": 0.0005706880641925778, + "loss": 1.3868, + "theoretical_loss": 3.6884252333246925, + "tokens_seen": 969605120 + }, + { + "epoch": 0.29, + "learning_rate": 0.0005706720160481444, + "loss": 1.439, + "theoretical_loss": 3.6884015576476825, + "tokens_seen": 969670656 + }, + { + "epoch": 0.29, + "learning_rate": 0.0005706559679037112, + "loss": 1.4166, + "theoretical_loss": 3.688377884018764, + "tokens_seen": 969736192 + }, + { + "epoch": 0.29, + "learning_rate": 0.0005706399197592778, + "loss": 1.3884, + "theoretical_loss": 3.6883542124376207, + "tokens_seen": 969801728 + }, + { + "epoch": 0.29, + "learning_rate": 0.0005706238716148446, + "loss": 1.5278, + "theoretical_loss": 3.6883305429039375, + "tokens_seen": 969867264 + }, + { + "epoch": 0.29, + "objective/train/docs_used": 356486, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.3945077657699585, + "objective/train/theoretical_loss": 3.688306875417399, + "objective/train/tokens_used": 990392800, + "theoretical_loss": 3.688306875417399, + "tokens_seen": 969932800 + }, + { + "epoch": 0.29, + "learning_rate": 0.0005706078234704112, + "loss": 1.4173, + "theoretical_loss": 3.688306875417399, + "tokens_seen": 969932800 + }, + { + "epoch": 0.29, + "learning_rate": 0.000570591775325978, + "loss": 1.3795, + "theoretical_loss": 3.688283209977689, + "tokens_seen": 969998336 + }, + { + "epoch": 0.29, + "learning_rate": 0.0005705757271815446, + "loss": 1.3757, + "theoretical_loss": 3.6882595465844936, + "tokens_seen": 970063872 + }, + { + "epoch": 0.29, + "learning_rate": 0.0005705596790371114, + "loss": 1.5, + "theoretical_loss": 3.6882358852374972, + "tokens_seen": 970129408 + }, + { + "epoch": 0.29, + "learning_rate": 0.0005705436308926781, + "loss": 1.4857, + "theoretical_loss": 3.688212225936384, + "tokens_seen": 970194944 + }, + { + "epoch": 0.29, + "learning_rate": 0.0005705275827482448, + "loss": 1.3555, + "theoretical_loss": 3.6881885686808396, + "tokens_seen": 970260480 + }, + { + "epoch": 0.29, + "learning_rate": 0.0005705115346038115, + "loss": 1.4629, + "theoretical_loss": 3.688164913470548, + "tokens_seen": 970326016 + }, + { + "epoch": 0.29, + "learning_rate": 0.0005704954864593782, + "loss": 1.5176, + "theoretical_loss": 3.6881412603051955, + "tokens_seen": 970391552 + }, + { + "epoch": 0.29, + "learning_rate": 0.0005704794383149449, + "loss": 1.3539, + "theoretical_loss": 3.6881176091844665, + "tokens_seen": 970457088 + }, + { + "epoch": 0.29, + "learning_rate": 0.0005704633901705116, + "loss": 1.3818, + "theoretical_loss": 3.6880939601080467, + "tokens_seen": 970522624 + }, + { + "epoch": 0.29, + "learning_rate": 0.0005704473420260783, + "loss": 1.455, + "theoretical_loss": 3.688070313075621, + "tokens_seen": 970588160 + }, + { + "epoch": 0.29, + "learning_rate": 0.000570431293881645, + "loss": 1.3621, + "theoretical_loss": 3.6880466680868746, + "tokens_seen": 970653696 + }, + { + "epoch": 0.29, + "learning_rate": 0.0005704152457372117, + "loss": 1.5654, + "theoretical_loss": 3.6880230251414936, + "tokens_seen": 970719232 + }, + { + "epoch": 0.29, + "learning_rate": 0.0005703991975927784, + "loss": 1.4667, + "theoretical_loss": 3.6879993842391627, + "tokens_seen": 970784768 + }, + { + "epoch": 0.29, + "learning_rate": 0.0005703831494483451, + "loss": 1.4116, + "theoretical_loss": 3.6879757453795676, + "tokens_seen": 970850304 + }, + { + "epoch": 0.29, + "learning_rate": 0.0005703671013039118, + "loss": 1.4325, + "theoretical_loss": 3.6879521085623943, + "tokens_seen": 970915840 + }, + { + "epoch": 0.29, + "learning_rate": 0.0005703510531594785, + "loss": 1.4694, + "theoretical_loss": 3.687928473787328, + "tokens_seen": 970981376 + }, + { + "epoch": 0.29, + "learning_rate": 0.0005703350050150452, + "loss": 1.4343, + "theoretical_loss": 3.6879048410540554, + "tokens_seen": 971046912 + }, + { + "epoch": 0.29, + "learning_rate": 0.0005703189568706119, + "loss": 1.3927, + "theoretical_loss": 3.687881210362261, + "tokens_seen": 971112448 + }, + { + "epoch": 0.29, + "learning_rate": 0.0005703029087261786, + "loss": 1.4996, + "theoretical_loss": 3.687857581711631, + "tokens_seen": 971177984 + }, + { + "epoch": 0.29, + "learning_rate": 0.0005702868605817453, + "loss": 1.4125, + "theoretical_loss": 3.687833955101852, + "tokens_seen": 971243520 + }, + { + "epoch": 0.29, + "learning_rate": 0.000570270812437312, + "loss": 1.521, + "theoretical_loss": 3.68781033053261, + "tokens_seen": 971309056 + }, + { + "epoch": 0.29, + "learning_rate": 0.0005702547642928787, + "loss": 1.389, + "theoretical_loss": 3.687786708003591, + "tokens_seen": 971374592 + }, + { + "epoch": 0.29, + "learning_rate": 0.0005702387161484454, + "loss": 1.4887, + "theoretical_loss": 3.6877630875144805, + "tokens_seen": 971440128 + }, + { + "epoch": 0.29, + "learning_rate": 0.0005702226680040121, + "loss": 1.4918, + "theoretical_loss": 3.687739469064966, + "tokens_seen": 971505664 + }, + { + "epoch": 0.29, + "learning_rate": 0.0005702066198595788, + "loss": 1.4217, + "theoretical_loss": 3.6877158526547325, + "tokens_seen": 971571200 + }, + { + "epoch": 0.29, + "learning_rate": 0.0005701905717151455, + "loss": 1.4975, + "theoretical_loss": 3.6876922382834674, + "tokens_seen": 971636736 + }, + { + "epoch": 0.29, + "learning_rate": 0.0005701745235707122, + "loss": 1.4055, + "theoretical_loss": 3.687668625950856, + "tokens_seen": 971702272 + }, + { + "epoch": 0.29, + "learning_rate": 0.0005701584754262789, + "loss": 1.3941, + "theoretical_loss": 3.687645015656586, + "tokens_seen": 971767808 + }, + { + "epoch": 0.29, + "learning_rate": 0.0005701424272818456, + "loss": 1.4662, + "theoretical_loss": 3.6876214074003437, + "tokens_seen": 971833344 + }, + { + "epoch": 0.29, + "learning_rate": 0.0005701263791374123, + "loss": 1.45, + "theoretical_loss": 3.6875978011818153, + "tokens_seen": 971898880 + }, + { + "epoch": 0.29, + "learning_rate": 0.000570110330992979, + "loss": 1.4009, + "theoretical_loss": 3.6875741970006883, + "tokens_seen": 971964416 + }, + { + "epoch": 0.29, + "learning_rate": 0.0005700942828485457, + "loss": 1.4803, + "theoretical_loss": 3.6875505948566487, + "tokens_seen": 972029952 + }, + { + "epoch": 0.29, + "learning_rate": 0.0005700782347041124, + "loss": 1.3868, + "theoretical_loss": 3.6875269947493834, + "tokens_seen": 972095488 + }, + { + "epoch": 0.29, + "learning_rate": 0.0005700621865596791, + "loss": 1.5114, + "theoretical_loss": 3.68750339667858, + "tokens_seen": 972161024 + }, + { + "epoch": 0.29, + "learning_rate": 0.0005700461384152458, + "loss": 1.4132, + "theoretical_loss": 3.687479800643925, + "tokens_seen": 972226560 + }, + { + "epoch": 0.29, + "learning_rate": 0.0005700300902708125, + "loss": 1.3271, + "theoretical_loss": 3.687456206645106, + "tokens_seen": 972292096 + }, + { + "epoch": 0.29, + "learning_rate": 0.0005700140421263792, + "loss": 1.4083, + "theoretical_loss": 3.6874326146818097, + "tokens_seen": 972357632 + }, + { + "epoch": 0.29, + "learning_rate": 0.0005699979939819458, + "loss": 1.4951, + "theoretical_loss": 3.687409024753723, + "tokens_seen": 972423168 + }, + { + "epoch": 0.29, + "learning_rate": 0.0005699819458375126, + "loss": 1.3122, + "theoretical_loss": 3.6873854368605334, + "tokens_seen": 972488704 + }, + { + "epoch": 0.29, + "learning_rate": 0.0005699658976930792, + "loss": 1.5048, + "theoretical_loss": 3.687361851001929, + "tokens_seen": 972554240 + }, + { + "epoch": 0.29, + "learning_rate": 0.000569949849548646, + "loss": 1.378, + "theoretical_loss": 3.687338267177596, + "tokens_seen": 972619776 + }, + { + "epoch": 0.29, + "learning_rate": 0.0005699338014042126, + "loss": 1.4492, + "theoretical_loss": 3.687314685387223, + "tokens_seen": 972685312 + }, + { + "epoch": 0.29, + "learning_rate": 0.0005699177532597794, + "loss": 1.507, + "theoretical_loss": 3.6872911056304973, + "tokens_seen": 972750848 + }, + { + "epoch": 0.29, + "learning_rate": 0.000569901705115346, + "loss": 1.3766, + "theoretical_loss": 3.687267527907106, + "tokens_seen": 972816384 + }, + { + "epoch": 0.29, + "learning_rate": 0.0005698856569709128, + "loss": 1.5332, + "theoretical_loss": 3.687243952216737, + "tokens_seen": 972881920 + }, + { + "epoch": 0.29, + "learning_rate": 0.0005698696088264794, + "loss": 1.3091, + "theoretical_loss": 3.6872203785590782, + "tokens_seen": 972947456 + }, + { + "epoch": 0.29, + "learning_rate": 0.0005698535606820461, + "loss": 1.4112, + "theoretical_loss": 3.6871968069338177, + "tokens_seen": 973012992 + }, + { + "epoch": 0.29, + "learning_rate": 0.0005698375125376128, + "loss": 1.5588, + "theoretical_loss": 3.6871732373406436, + "tokens_seen": 973078528 + }, + { + "epoch": 0.29, + "learning_rate": 0.0005698214643931795, + "loss": 1.5544, + "theoretical_loss": 3.6871496697792425, + "tokens_seen": 973144064 + }, + { + "epoch": 0.29, + "objective/train/docs_used": 357717, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.4963544607162476, + "objective/train/theoretical_loss": 3.6871261042493035, + "objective/train/tokens_used": 993669600, + "theoretical_loss": 3.6871261042493035, + "tokens_seen": 973209600 + }, + { + "epoch": 0.29, + "learning_rate": 0.0005698054162487462, + "loss": 1.5417, + "theoretical_loss": 3.6871261042493035, + "tokens_seen": 973209600 + }, + { + "epoch": 0.29, + "learning_rate": 0.0005697893681043129, + "loss": 1.4662, + "theoretical_loss": 3.687102540750515, + "tokens_seen": 973275136 + }, + { + "epoch": 0.29, + "learning_rate": 0.0005697733199598796, + "loss": 1.3584, + "theoretical_loss": 3.687078979282564, + "tokens_seen": 973340672 + }, + { + "epoch": 0.29, + "learning_rate": 0.0005697572718154463, + "loss": 1.3834, + "theoretical_loss": 3.6870554198451404, + "tokens_seen": 973406208 + }, + { + "epoch": 0.29, + "learning_rate": 0.000569741223671013, + "loss": 1.4771, + "theoretical_loss": 3.687031862437931, + "tokens_seen": 973471744 + }, + { + "epoch": 0.3, + "learning_rate": 0.0005697251755265797, + "loss": 1.3808, + "theoretical_loss": 3.687008307060625, + "tokens_seen": 973537280 + }, + { + "epoch": 0.3, + "learning_rate": 0.0005697091273821464, + "loss": 1.3543, + "theoretical_loss": 3.6869847537129106, + "tokens_seen": 973602816 + }, + { + "epoch": 0.3, + "learning_rate": 0.0005696930792377131, + "loss": 1.4405, + "theoretical_loss": 3.6869612023944764, + "tokens_seen": 973668352 + }, + { + "epoch": 0.3, + "learning_rate": 0.0005696770310932798, + "loss": 1.4665, + "theoretical_loss": 3.6869376531050113, + "tokens_seen": 973733888 + }, + { + "epoch": 0.3, + "learning_rate": 0.0005696609829488465, + "loss": 1.4387, + "theoretical_loss": 3.686914105844203, + "tokens_seen": 973799424 + }, + { + "epoch": 0.3, + "learning_rate": 0.0005696449348044132, + "loss": 1.3368, + "theoretical_loss": 3.686890560611741, + "tokens_seen": 973864960 + }, + { + "epoch": 0.3, + "learning_rate": 0.0005696288866599799, + "loss": 1.4408, + "theoretical_loss": 3.686867017407314, + "tokens_seen": 973930496 + }, + { + "epoch": 0.3, + "learning_rate": 0.0005696128385155467, + "loss": 1.3646, + "theoretical_loss": 3.686843476230611, + "tokens_seen": 973996032 + }, + { + "epoch": 0.3, + "learning_rate": 0.0005695967903711133, + "loss": 1.492, + "theoretical_loss": 3.686819937081321, + "tokens_seen": 974061568 + }, + { + "epoch": 0.3, + "learning_rate": 0.00056958074222668, + "loss": 1.3155, + "theoretical_loss": 3.6867963999591327, + "tokens_seen": 974127104 + }, + { + "epoch": 0.3, + "learning_rate": 0.0005695646940822467, + "loss": 1.3825, + "theoretical_loss": 3.686772864863735, + "tokens_seen": 974192640 + }, + { + "epoch": 0.3, + "learning_rate": 0.0005695486459378135, + "loss": 1.444, + "theoretical_loss": 3.6867493317948172, + "tokens_seen": 974258176 + }, + { + "epoch": 0.3, + "learning_rate": 0.0005695325977933801, + "loss": 1.4767, + "theoretical_loss": 3.686725800752069, + "tokens_seen": 974323712 + }, + { + "epoch": 0.3, + "learning_rate": 0.0005695165496489469, + "loss": 1.3123, + "theoretical_loss": 3.6867022717351787, + "tokens_seen": 974389248 + }, + { + "epoch": 0.3, + "learning_rate": 0.0005695005015045135, + "loss": 1.5351, + "theoretical_loss": 3.686678744743837, + "tokens_seen": 974454784 + }, + { + "epoch": 0.3, + "learning_rate": 0.0005694844533600803, + "loss": 1.4067, + "theoretical_loss": 3.686655219777732, + "tokens_seen": 974520320 + }, + { + "epoch": 0.3, + "learning_rate": 0.0005694684052156469, + "loss": 1.4403, + "theoretical_loss": 3.686631696836554, + "tokens_seen": 974585856 + }, + { + "epoch": 0.3, + "learning_rate": 0.0005694523570712137, + "loss": 1.566, + "theoretical_loss": 3.686608175919992, + "tokens_seen": 974651392 + }, + { + "epoch": 0.3, + "learning_rate": 0.0005694363089267803, + "loss": 1.3492, + "theoretical_loss": 3.6865846570277365, + "tokens_seen": 974716928 + }, + { + "epoch": 0.3, + "learning_rate": 0.0005694202607823471, + "loss": 1.4494, + "theoretical_loss": 3.6865611401594762, + "tokens_seen": 974782464 + }, + { + "epoch": 0.3, + "learning_rate": 0.0005694042126379137, + "loss": 1.4, + "theoretical_loss": 3.6865376253149016, + "tokens_seen": 974848000 + }, + { + "epoch": 0.3, + "learning_rate": 0.0005693881644934805, + "loss": 1.4203, + "theoretical_loss": 3.6865141124937013, + "tokens_seen": 974913536 + }, + { + "epoch": 0.3, + "learning_rate": 0.0005693721163490471, + "loss": 1.4766, + "theoretical_loss": 3.686490601695567, + "tokens_seen": 974979072 + }, + { + "epoch": 0.3, + "learning_rate": 0.0005693560682046139, + "loss": 1.4855, + "theoretical_loss": 3.6864670929201875, + "tokens_seen": 975044608 + }, + { + "epoch": 0.3, + "learning_rate": 0.0005693400200601806, + "loss": 1.4593, + "theoretical_loss": 3.686443586167253, + "tokens_seen": 975110144 + }, + { + "epoch": 0.3, + "learning_rate": 0.0005693239719157473, + "loss": 1.5347, + "theoretical_loss": 3.686420081436454, + "tokens_seen": 975175680 + }, + { + "epoch": 0.3, + "learning_rate": 0.000569307923771314, + "loss": 1.4429, + "theoretical_loss": 3.6863965787274804, + "tokens_seen": 975241216 + }, + { + "epoch": 0.3, + "learning_rate": 0.0005692918756268807, + "loss": 1.474, + "theoretical_loss": 3.686373078040022, + "tokens_seen": 975306752 + }, + { + "epoch": 0.3, + "learning_rate": 0.0005692758274824474, + "loss": 1.4416, + "theoretical_loss": 3.6863495793737697, + "tokens_seen": 975372288 + }, + { + "epoch": 0.3, + "learning_rate": 0.0005692597793380141, + "loss": 1.5242, + "theoretical_loss": 3.6863260827284137, + "tokens_seen": 975437824 + }, + { + "epoch": 0.3, + "learning_rate": 0.0005692437311935808, + "loss": 1.4432, + "theoretical_loss": 3.6863025881036444, + "tokens_seen": 975503360 + }, + { + "epoch": 0.3, + "learning_rate": 0.0005692276830491475, + "loss": 1.3331, + "theoretical_loss": 3.6862790954991524, + "tokens_seen": 975568896 + }, + { + "epoch": 0.3, + "learning_rate": 0.0005692116349047142, + "loss": 1.4022, + "theoretical_loss": 3.686255604914628, + "tokens_seen": 975634432 + }, + { + "epoch": 0.3, + "learning_rate": 0.0005691955867602809, + "loss": 1.5145, + "theoretical_loss": 3.6862321163497622, + "tokens_seen": 975699968 + }, + { + "epoch": 0.3, + "learning_rate": 0.0005691795386158476, + "loss": 1.487, + "theoretical_loss": 3.686208629804246, + "tokens_seen": 975765504 + }, + { + "epoch": 0.3, + "learning_rate": 0.0005691634904714143, + "loss": 1.4724, + "theoretical_loss": 3.686185145277769, + "tokens_seen": 975831040 + }, + { + "epoch": 0.3, + "learning_rate": 0.000569147442326981, + "loss": 1.4475, + "theoretical_loss": 3.686161662770023, + "tokens_seen": 975896576 + }, + { + "epoch": 0.3, + "learning_rate": 0.0005691313941825477, + "loss": 1.3017, + "theoretical_loss": 3.686138182280699, + "tokens_seen": 975962112 + }, + { + "epoch": 0.3, + "learning_rate": 0.0005691153460381144, + "loss": 1.4606, + "theoretical_loss": 3.686114703809488, + "tokens_seen": 976027648 + }, + { + "epoch": 0.3, + "learning_rate": 0.0005690992978936811, + "loss": 1.4396, + "theoretical_loss": 3.68609122735608, + "tokens_seen": 976093184 + }, + { + "epoch": 0.3, + "learning_rate": 0.0005690832497492478, + "loss": 1.3852, + "theoretical_loss": 3.6860677529201675, + "tokens_seen": 976158720 + }, + { + "epoch": 0.3, + "learning_rate": 0.0005690672016048145, + "loss": 1.4347, + "theoretical_loss": 3.686044280501441, + "tokens_seen": 976224256 + }, + { + "epoch": 0.3, + "learning_rate": 0.0005690511534603812, + "loss": 1.4153, + "theoretical_loss": 3.686020810099592, + "tokens_seen": 976289792 + }, + { + "epoch": 0.3, + "learning_rate": 0.0005690351053159479, + "loss": 1.3966, + "theoretical_loss": 3.6859973417143115, + "tokens_seen": 976355328 + }, + { + "epoch": 0.3, + "learning_rate": 0.0005690190571715146, + "loss": 1.4462, + "theoretical_loss": 3.685973875345291, + "tokens_seen": 976420864 + }, + { + "epoch": 0.3, + "objective/train/docs_used": 358882, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.2705471515655518, + "objective/train/theoretical_loss": 3.6859504109922225, + "objective/train/tokens_used": 996946400, + "theoretical_loss": 3.6859504109922225, + "tokens_seen": 976486400 + }, + { + "epoch": 0.3, + "learning_rate": 0.0005690030090270813, + "loss": 1.412, + "theoretical_loss": 3.6859504109922225, + "tokens_seen": 976486400 + }, + { + "epoch": 0.3, + "learning_rate": 0.000568986960882648, + "loss": 1.4773, + "theoretical_loss": 3.6859269486547968, + "tokens_seen": 976551936 + }, + { + "epoch": 0.3, + "learning_rate": 0.0005689709127382147, + "loss": 1.3879, + "theoretical_loss": 3.6859034883327055, + "tokens_seen": 976617472 + }, + { + "epoch": 0.3, + "learning_rate": 0.0005689548645937814, + "loss": 1.4091, + "theoretical_loss": 3.6858800300256407, + "tokens_seen": 976683008 + }, + { + "epoch": 0.3, + "learning_rate": 0.0005689388164493481, + "loss": 1.3966, + "theoretical_loss": 3.685856573733294, + "tokens_seen": 976748544 + }, + { + "epoch": 0.3, + "learning_rate": 0.0005689227683049148, + "loss": 1.4349, + "theoretical_loss": 3.6858331194553573, + "tokens_seen": 976814080 + }, + { + "epoch": 0.3, + "learning_rate": 0.0005689067201604815, + "loss": 1.4456, + "theoretical_loss": 3.6858096671915224, + "tokens_seen": 976879616 + }, + { + "epoch": 0.3, + "learning_rate": 0.0005688906720160482, + "loss": 1.4704, + "theoretical_loss": 3.685786216941481, + "tokens_seen": 976945152 + }, + { + "epoch": 0.3, + "learning_rate": 0.0005688746238716149, + "loss": 1.4031, + "theoretical_loss": 3.6857627687049255, + "tokens_seen": 977010688 + }, + { + "epoch": 0.3, + "learning_rate": 0.0005688585757271816, + "loss": 1.5823, + "theoretical_loss": 3.6857393224815476, + "tokens_seen": 977076224 + }, + { + "epoch": 0.3, + "learning_rate": 0.0005688425275827483, + "loss": 1.5233, + "theoretical_loss": 3.68571587827104, + "tokens_seen": 977141760 + }, + { + "epoch": 0.3, + "learning_rate": 0.000568826479438315, + "loss": 1.3915, + "theoretical_loss": 3.6856924360730936, + "tokens_seen": 977207296 + }, + { + "epoch": 0.3, + "learning_rate": 0.0005688104312938817, + "loss": 1.4692, + "theoretical_loss": 3.6856689958874025, + "tokens_seen": 977272832 + }, + { + "epoch": 0.3, + "learning_rate": 0.0005687943831494484, + "loss": 1.5191, + "theoretical_loss": 3.6856455577136584, + "tokens_seen": 977338368 + }, + { + "epoch": 0.3, + "learning_rate": 0.0005687783350050151, + "loss": 1.5295, + "theoretical_loss": 3.6856221215515528, + "tokens_seen": 977403904 + }, + { + "epoch": 0.3, + "learning_rate": 0.0005687622868605818, + "loss": 1.3574, + "theoretical_loss": 3.685598687400779, + "tokens_seen": 977469440 + }, + { + "epoch": 0.3, + "learning_rate": 0.0005687462387161485, + "loss": 1.5069, + "theoretical_loss": 3.6855752552610292, + "tokens_seen": 977534976 + }, + { + "epoch": 0.3, + "learning_rate": 0.0005687301905717152, + "loss": 1.4665, + "theoretical_loss": 3.6855518251319968, + "tokens_seen": 977600512 + }, + { + "epoch": 0.3, + "learning_rate": 0.0005687141424272819, + "loss": 1.3857, + "theoretical_loss": 3.6855283970133734, + "tokens_seen": 977666048 + }, + { + "epoch": 0.3, + "learning_rate": 0.0005686980942828486, + "loss": 1.37, + "theoretical_loss": 3.6855049709048524, + "tokens_seen": 977731584 + }, + { + "epoch": 0.3, + "learning_rate": 0.0005686820461384153, + "loss": 1.377, + "theoretical_loss": 3.685481546806127, + "tokens_seen": 977797120 + }, + { + "epoch": 0.3, + "learning_rate": 0.000568665997993982, + "loss": 1.3931, + "theoretical_loss": 3.685458124716889, + "tokens_seen": 977862656 + }, + { + "epoch": 0.3, + "learning_rate": 0.0005686499498495487, + "loss": 1.3776, + "theoretical_loss": 3.6854347046368314, + "tokens_seen": 977928192 + }, + { + "epoch": 0.3, + "learning_rate": 0.0005686339017051154, + "loss": 1.5258, + "theoretical_loss": 3.6854112865656488, + "tokens_seen": 977993728 + }, + { + "epoch": 0.3, + "learning_rate": 0.0005686178535606821, + "loss": 1.5903, + "theoretical_loss": 3.6853878705030327, + "tokens_seen": 978059264 + }, + { + "epoch": 0.3, + "learning_rate": 0.0005686018054162487, + "loss": 1.4661, + "theoretical_loss": 3.685364456448677, + "tokens_seen": 978124800 + }, + { + "epoch": 0.3, + "learning_rate": 0.0005685857572718155, + "loss": 1.3584, + "theoretical_loss": 3.685341044402274, + "tokens_seen": 978190336 + }, + { + "epoch": 0.3, + "learning_rate": 0.0005685697091273821, + "loss": 1.4559, + "theoretical_loss": 3.6853176343635186, + "tokens_seen": 978255872 + }, + { + "epoch": 0.3, + "learning_rate": 0.0005685536609829489, + "loss": 1.4704, + "theoretical_loss": 3.6852942263321022, + "tokens_seen": 978321408 + }, + { + "epoch": 0.3, + "learning_rate": 0.0005685376128385155, + "loss": 1.3959, + "theoretical_loss": 3.6852708203077205, + "tokens_seen": 978386944 + }, + { + "epoch": 0.3, + "learning_rate": 0.0005685215646940823, + "loss": 1.3862, + "theoretical_loss": 3.685247416290065, + "tokens_seen": 978452480 + }, + { + "epoch": 0.3, + "learning_rate": 0.0005685055165496489, + "loss": 1.4048, + "theoretical_loss": 3.6852240142788304, + "tokens_seen": 978518016 + }, + { + "epoch": 0.3, + "learning_rate": 0.0005684894684052157, + "loss": 1.4209, + "theoretical_loss": 3.6852006142737093, + "tokens_seen": 978583552 + }, + { + "epoch": 0.3, + "learning_rate": 0.0005684734202607823, + "loss": 1.4685, + "theoretical_loss": 3.685177216274397, + "tokens_seen": 978649088 + }, + { + "epoch": 0.3, + "learning_rate": 0.000568457372116349, + "loss": 1.4233, + "theoretical_loss": 3.6851538202805854, + "tokens_seen": 978714624 + }, + { + "epoch": 0.3, + "learning_rate": 0.0005684413239719157, + "loss": 1.3644, + "theoretical_loss": 3.6851304262919697, + "tokens_seen": 978780160 + }, + { + "epoch": 0.3, + "learning_rate": 0.0005684252758274824, + "loss": 1.5604, + "theoretical_loss": 3.6851070343082433, + "tokens_seen": 978845696 + }, + { + "epoch": 0.3, + "learning_rate": 0.0005684092276830491, + "loss": 1.4281, + "theoretical_loss": 3.6850836443291004, + "tokens_seen": 978911232 + }, + { + "epoch": 0.3, + "learning_rate": 0.0005683931795386158, + "loss": 1.2615, + "theoretical_loss": 3.6850602563542347, + "tokens_seen": 978976768 + }, + { + "epoch": 0.3, + "learning_rate": 0.0005683771313941825, + "loss": 1.4737, + "theoretical_loss": 3.6850368703833403, + "tokens_seen": 979042304 + }, + { + "epoch": 0.3, + "learning_rate": 0.0005683610832497492, + "loss": 1.4043, + "theoretical_loss": 3.6850134864161115, + "tokens_seen": 979107840 + }, + { + "epoch": 0.3, + "learning_rate": 0.000568345035105316, + "loss": 1.3937, + "theoretical_loss": 3.684990104452243, + "tokens_seen": 979173376 + }, + { + "epoch": 0.3, + "learning_rate": 0.0005683289869608826, + "loss": 1.323, + "theoretical_loss": 3.684966724491428, + "tokens_seen": 979238912 + }, + { + "epoch": 0.3, + "learning_rate": 0.0005683129388164493, + "loss": 1.3374, + "theoretical_loss": 3.6849433465333616, + "tokens_seen": 979304448 + }, + { + "epoch": 0.3, + "learning_rate": 0.000568296890672016, + "loss": 1.4163, + "theoretical_loss": 3.6849199705777385, + "tokens_seen": 979369984 + }, + { + "epoch": 0.3, + "learning_rate": 0.0005682808425275827, + "loss": 1.4061, + "theoretical_loss": 3.6848965966242524, + "tokens_seen": 979435520 + }, + { + "epoch": 0.3, + "learning_rate": 0.0005682647943831494, + "loss": 1.4835, + "theoretical_loss": 3.684873224672599, + "tokens_seen": 979501056 + }, + { + "epoch": 0.3, + "learning_rate": 0.0005682487462387162, + "loss": 1.5214, + "theoretical_loss": 3.6848498547224717, + "tokens_seen": 979566592 + }, + { + "epoch": 0.3, + "learning_rate": 0.0005682326980942828, + "loss": 1.506, + "theoretical_loss": 3.6848264867735656, + "tokens_seen": 979632128 + }, + { + "epoch": 0.3, + "learning_rate": 0.0005682166499498496, + "loss": 1.5618, + "theoretical_loss": 3.684803120825576, + "tokens_seen": 979697664 + }, + { + "epoch": 0.3, + "objective/train/docs_used": 360128, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.518088698387146, + "objective/train/theoretical_loss": 3.6847797568781973, + "objective/train/tokens_used": 1000223200, + "theoretical_loss": 3.6847797568781973, + "tokens_seen": 979763200 + }, + { + "epoch": 0.3, + "learning_rate": 0.0005682006018054162, + "loss": 1.4553, + "theoretical_loss": 3.6847797568781973, + "tokens_seen": 979763200 + }, + { + "epoch": 0.3, + "learning_rate": 0.000568184553660983, + "loss": 1.4459, + "theoretical_loss": 3.6847563949311244, + "tokens_seen": 979828736 + }, + { + "epoch": 0.3, + "learning_rate": 0.0005681685055165496, + "loss": 1.3953, + "theoretical_loss": 3.684733034984052, + "tokens_seen": 979894272 + }, + { + "epoch": 0.3, + "learning_rate": 0.0005681524573721164, + "loss": 1.4006, + "theoretical_loss": 3.684709677036676, + "tokens_seen": 979959808 + }, + { + "epoch": 0.3, + "learning_rate": 0.0005681364092276831, + "loss": 1.4744, + "theoretical_loss": 3.6846863210886904, + "tokens_seen": 980025344 + }, + { + "epoch": 0.3, + "learning_rate": 0.0005681203610832498, + "loss": 1.4651, + "theoretical_loss": 3.6846629671397912, + "tokens_seen": 980090880 + }, + { + "epoch": 0.3, + "learning_rate": 0.0005681043129388165, + "loss": 1.4391, + "theoretical_loss": 3.684639615189673, + "tokens_seen": 980156416 + }, + { + "epoch": 0.3, + "learning_rate": 0.0005680882647943832, + "loss": 1.3909, + "theoretical_loss": 3.6846162652380325, + "tokens_seen": 980221952 + }, + { + "epoch": 0.3, + "learning_rate": 0.0005680722166499499, + "loss": 1.4606, + "theoretical_loss": 3.6845929172845633, + "tokens_seen": 980287488 + }, + { + "epoch": 0.3, + "learning_rate": 0.0005680561685055166, + "loss": 1.4023, + "theoretical_loss": 3.6845695713289617, + "tokens_seen": 980353024 + }, + { + "epoch": 0.3, + "learning_rate": 0.0005680401203610833, + "loss": 1.4656, + "theoretical_loss": 3.684546227370923, + "tokens_seen": 980418560 + }, + { + "epoch": 0.3, + "learning_rate": 0.00056802407221665, + "loss": 1.3891, + "theoretical_loss": 3.684522885410143, + "tokens_seen": 980484096 + }, + { + "epoch": 0.3, + "learning_rate": 0.0005680080240722167, + "loss": 1.4037, + "theoretical_loss": 3.684499545446317, + "tokens_seen": 980549632 + }, + { + "epoch": 0.3, + "learning_rate": 0.0005679919759277834, + "loss": 1.3431, + "theoretical_loss": 3.684476207479141, + "tokens_seen": 980615168 + }, + { + "epoch": 0.3, + "learning_rate": 0.0005679759277833501, + "loss": 1.4564, + "theoretical_loss": 3.6844528715083107, + "tokens_seen": 980680704 + }, + { + "epoch": 0.3, + "learning_rate": 0.0005679598796389168, + "loss": 1.449, + "theoretical_loss": 3.6844295375335223, + "tokens_seen": 980746240 + }, + { + "epoch": 0.3, + "learning_rate": 0.0005679438314944835, + "loss": 1.4156, + "theoretical_loss": 3.684406205554471, + "tokens_seen": 980811776 + }, + { + "epoch": 0.3, + "learning_rate": 0.0005679277833500502, + "loss": 1.4098, + "theoretical_loss": 3.6843828755708525, + "tokens_seen": 980877312 + }, + { + "epoch": 0.3, + "learning_rate": 0.0005679117352056169, + "loss": 1.3904, + "theoretical_loss": 3.684359547582364, + "tokens_seen": 980942848 + }, + { + "epoch": 0.3, + "learning_rate": 0.0005678956870611836, + "loss": 1.4635, + "theoretical_loss": 3.684336221588701, + "tokens_seen": 981008384 + }, + { + "epoch": 0.3, + "learning_rate": 0.0005678796389167503, + "loss": 1.4491, + "theoretical_loss": 3.6843128975895603, + "tokens_seen": 981073920 + }, + { + "epoch": 0.3, + "learning_rate": 0.000567863590772317, + "loss": 1.5026, + "theoretical_loss": 3.6842895755846365, + "tokens_seen": 981139456 + }, + { + "epoch": 0.3, + "learning_rate": 0.0005678475426278837, + "loss": 1.4307, + "theoretical_loss": 3.684266255573627, + "tokens_seen": 981204992 + }, + { + "epoch": 0.3, + "learning_rate": 0.0005678314944834504, + "loss": 1.4629, + "theoretical_loss": 3.684242937556229, + "tokens_seen": 981270528 + }, + { + "epoch": 0.3, + "learning_rate": 0.0005678154463390171, + "loss": 1.4605, + "theoretical_loss": 3.6842196215321374, + "tokens_seen": 981336064 + }, + { + "epoch": 0.3, + "learning_rate": 0.0005677993981945838, + "loss": 1.5241, + "theoretical_loss": 3.6841963075010495, + "tokens_seen": 981401600 + }, + { + "epoch": 0.3, + "learning_rate": 0.0005677833500501505, + "loss": 1.3645, + "theoretical_loss": 3.684172995462662, + "tokens_seen": 981467136 + }, + { + "epoch": 0.3, + "learning_rate": 0.0005677673019057172, + "loss": 1.3869, + "theoretical_loss": 3.6841496854166706, + "tokens_seen": 981532672 + }, + { + "epoch": 0.3, + "learning_rate": 0.0005677512537612839, + "loss": 1.5434, + "theoretical_loss": 3.6841263773627735, + "tokens_seen": 981598208 + }, + { + "epoch": 0.3, + "learning_rate": 0.0005677352056168506, + "loss": 1.4677, + "theoretical_loss": 3.684103071300666, + "tokens_seen": 981663744 + }, + { + "epoch": 0.3, + "learning_rate": 0.0005677191574724173, + "loss": 1.429, + "theoretical_loss": 3.6840797672300463, + "tokens_seen": 981729280 + }, + { + "epoch": 0.3, + "learning_rate": 0.000567703109327984, + "loss": 1.4591, + "theoretical_loss": 3.6840564651506096, + "tokens_seen": 981794816 + }, + { + "epoch": 0.3, + "learning_rate": 0.0005676870611835507, + "loss": 1.3739, + "theoretical_loss": 3.684033165062055, + "tokens_seen": 981860352 + }, + { + "epoch": 0.3, + "learning_rate": 0.0005676710130391174, + "loss": 1.4229, + "theoretical_loss": 3.6840098669640775, + "tokens_seen": 981925888 + }, + { + "epoch": 0.3, + "learning_rate": 0.0005676549648946841, + "loss": 1.4484, + "theoretical_loss": 3.6839865708563755, + "tokens_seen": 981991424 + }, + { + "epoch": 0.3, + "learning_rate": 0.0005676389167502508, + "loss": 1.4402, + "theoretical_loss": 3.6839632767386457, + "tokens_seen": 982056960 + }, + { + "epoch": 0.3, + "learning_rate": 0.0005676228686058175, + "loss": 1.5141, + "theoretical_loss": 3.6839399846105847, + "tokens_seen": 982122496 + }, + { + "epoch": 0.3, + "learning_rate": 0.0005676068204613842, + "loss": 1.4475, + "theoretical_loss": 3.6839166944718915, + "tokens_seen": 982188032 + }, + { + "epoch": 0.3, + "learning_rate": 0.0005675907723169509, + "loss": 1.4891, + "theoretical_loss": 3.683893406322262, + "tokens_seen": 982253568 + }, + { + "epoch": 0.3, + "learning_rate": 0.0005675747241725176, + "loss": 1.4159, + "theoretical_loss": 3.6838701201613944, + "tokens_seen": 982319104 + }, + { + "epoch": 0.3, + "learning_rate": 0.0005675586760280843, + "loss": 1.4006, + "theoretical_loss": 3.6838468359889855, + "tokens_seen": 982384640 + }, + { + "epoch": 0.3, + "learning_rate": 0.000567542627883651, + "loss": 1.5004, + "theoretical_loss": 3.683823553804733, + "tokens_seen": 982450176 + }, + { + "epoch": 0.3, + "learning_rate": 0.0005675265797392177, + "loss": 1.4908, + "theoretical_loss": 3.6838002736083353, + "tokens_seen": 982515712 + }, + { + "epoch": 0.3, + "learning_rate": 0.0005675105315947844, + "loss": 1.4923, + "theoretical_loss": 3.6837769953994894, + "tokens_seen": 982581248 + }, + { + "epoch": 0.3, + "learning_rate": 0.0005674944834503511, + "loss": 1.5264, + "theoretical_loss": 3.6837537191778935, + "tokens_seen": 982646784 + }, + { + "epoch": 0.3, + "learning_rate": 0.0005674784353059178, + "loss": 1.4833, + "theoretical_loss": 3.6837304449432446, + "tokens_seen": 982712320 + }, + { + "epoch": 0.3, + "learning_rate": 0.0005674623871614845, + "loss": 1.4812, + "theoretical_loss": 3.683707172695242, + "tokens_seen": 982777856 + }, + { + "epoch": 0.3, + "learning_rate": 0.0005674463390170512, + "loss": 1.4692, + "theoretical_loss": 3.6836839024335823, + "tokens_seen": 982843392 + }, + { + "epoch": 0.3, + "learning_rate": 0.0005674302908726179, + "loss": 1.4125, + "theoretical_loss": 3.683660634157964, + "tokens_seen": 982908928 + }, + { + "epoch": 0.3, + "learning_rate": 0.0005674142427281846, + "loss": 1.3736, + "theoretical_loss": 3.6836373678680854, + "tokens_seen": 982974464 + }, + { + "debugging/Compilability": 1.0, + "debugging/distinct-1-grams": 0.7488293247094114, + "debugging/entropy-1-grams": 5.349564433264487, + "debugging/length": 462.46153846153845, + "debugging/num_segments": 13, + "debugging/score": 0.0073770168052465215, + "debugging/score_std": 0.0054551366535427, + "epoch": 0.3, + "objective/train/docs_used": 361383, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.3451521396636963, + "objective/train/theoretical_loss": 3.6836141035636447, + "objective/train/tokens_used": 1003500000, + "theoretical_loss": 3.6836141035636447, + "tokens_seen": 983040000 + }, + { + "epoch": 0.3, + "learning_rate": 0.0005673981945837513, + "loss": 1.428, + "theoretical_loss": 3.6836141035636447, + "tokens_seen": 983040000 + }, + { + "epoch": 0.3, + "learning_rate": 0.000567382146439318, + "loss": 1.4553, + "theoretical_loss": 3.6835908412443397, + "tokens_seen": 983105536 + }, + { + "epoch": 0.3, + "learning_rate": 0.0005673660982948847, + "loss": 1.5154, + "theoretical_loss": 3.683567580909869, + "tokens_seen": 983171072 + }, + { + "epoch": 0.3, + "learning_rate": 0.0005673500501504514, + "loss": 1.4309, + "theoretical_loss": 3.683544322559931, + "tokens_seen": 983236608 + }, + { + "epoch": 0.3, + "learning_rate": 0.0005673340020060181, + "loss": 1.4022, + "theoretical_loss": 3.683521066194224, + "tokens_seen": 983302144 + }, + { + "epoch": 0.3, + "learning_rate": 0.0005673179538615848, + "loss": 1.4143, + "theoretical_loss": 3.6834978118124466, + "tokens_seen": 983367680 + }, + { + "epoch": 0.3, + "learning_rate": 0.0005673019057171515, + "loss": 1.3943, + "theoretical_loss": 3.683474559414297, + "tokens_seen": 983433216 + }, + { + "epoch": 0.3, + "learning_rate": 0.0005672858575727182, + "loss": 1.5268, + "theoretical_loss": 3.6834513089994743, + "tokens_seen": 983498752 + }, + { + "epoch": 0.3, + "learning_rate": 0.0005672698094282849, + "loss": 1.4596, + "theoretical_loss": 3.683428060567677, + "tokens_seen": 983564288 + }, + { + "epoch": 0.3, + "learning_rate": 0.0005672537612838516, + "loss": 1.4988, + "theoretical_loss": 3.6834048141186035, + "tokens_seen": 983629824 + }, + { + "epoch": 0.3, + "learning_rate": 0.0005672377131394183, + "loss": 1.4797, + "theoretical_loss": 3.683381569651953, + "tokens_seen": 983695360 + }, + { + "epoch": 0.3, + "learning_rate": 0.000567221664994985, + "loss": 1.4314, + "theoretical_loss": 3.683358327167425, + "tokens_seen": 983760896 + }, + { + "epoch": 0.3, + "learning_rate": 0.0005672056168505516, + "loss": 1.4797, + "theoretical_loss": 3.683335086664717, + "tokens_seen": 983826432 + }, + { + "epoch": 0.3, + "learning_rate": 0.0005671895687061184, + "loss": 1.4294, + "theoretical_loss": 3.6833118481435294, + "tokens_seen": 983891968 + }, + { + "epoch": 0.3, + "learning_rate": 0.000567173520561685, + "loss": 1.2717, + "theoretical_loss": 3.6832886116035604, + "tokens_seen": 983957504 + }, + { + "epoch": 0.3, + "learning_rate": 0.0005671574724172518, + "loss": 1.3784, + "theoretical_loss": 3.6832653770445094, + "tokens_seen": 984023040 + }, + { + "epoch": 0.3, + "learning_rate": 0.0005671414242728184, + "loss": 1.386, + "theoretical_loss": 3.6832421444660755, + "tokens_seen": 984088576 + }, + { + "epoch": 0.3, + "learning_rate": 0.0005671253761283852, + "loss": 1.4302, + "theoretical_loss": 3.6832189138679583, + "tokens_seen": 984154112 + }, + { + "epoch": 0.3, + "learning_rate": 0.0005671093279839518, + "loss": 1.4916, + "theoretical_loss": 3.6831956852498573, + "tokens_seen": 984219648 + }, + { + "epoch": 0.3, + "learning_rate": 0.0005670932798395186, + "loss": 1.3778, + "theoretical_loss": 3.6831724586114714, + "tokens_seen": 984285184 + }, + { + "epoch": 0.3, + "learning_rate": 0.0005670772316950852, + "loss": 1.4832, + "theoretical_loss": 3.6831492339525, + "tokens_seen": 984350720 + }, + { + "epoch": 0.3, + "learning_rate": 0.0005670611835506519, + "loss": 1.4023, + "theoretical_loss": 3.683126011272643, + "tokens_seen": 984416256 + }, + { + "epoch": 0.3, + "learning_rate": 0.0005670451354062186, + "loss": 1.4485, + "theoretical_loss": 3.6831027905716, + "tokens_seen": 984481792 + }, + { + "epoch": 0.3, + "learning_rate": 0.0005670290872617853, + "loss": 1.4749, + "theoretical_loss": 3.6830795718490705, + "tokens_seen": 984547328 + }, + { + "epoch": 0.3, + "learning_rate": 0.000567013039117352, + "loss": 1.5717, + "theoretical_loss": 3.683056355104754, + "tokens_seen": 984612864 + }, + { + "epoch": 0.3, + "learning_rate": 0.0005669969909729187, + "loss": 1.3838, + "theoretical_loss": 3.683033140338351, + "tokens_seen": 984678400 + }, + { + "epoch": 0.3, + "learning_rate": 0.0005669809428284854, + "loss": 1.5301, + "theoretical_loss": 3.6830099275495614, + "tokens_seen": 984743936 + }, + { + "epoch": 0.3, + "learning_rate": 0.0005669648946840522, + "loss": 1.473, + "theoretical_loss": 3.6829867167380845, + "tokens_seen": 984809472 + }, + { + "epoch": 0.3, + "learning_rate": 0.0005669488465396188, + "loss": 1.4614, + "theoretical_loss": 3.6829635079036205, + "tokens_seen": 984875008 + }, + { + "epoch": 0.3, + "learning_rate": 0.0005669327983951856, + "loss": 1.3815, + "theoretical_loss": 3.682940301045869, + "tokens_seen": 984940544 + }, + { + "epoch": 0.3, + "learning_rate": 0.0005669167502507522, + "loss": 1.5065, + "theoretical_loss": 3.6829170961645312, + "tokens_seen": 985006080 + }, + { + "epoch": 0.3, + "learning_rate": 0.000566900702106319, + "loss": 1.4833, + "theoretical_loss": 3.6828938932593065, + "tokens_seen": 985071616 + }, + { + "epoch": 0.3, + "learning_rate": 0.0005668846539618856, + "loss": 1.444, + "theoretical_loss": 3.6828706923298955, + "tokens_seen": 985137152 + }, + { + "epoch": 0.3, + "learning_rate": 0.0005668686058174524, + "loss": 1.4882, + "theoretical_loss": 3.6828474933759985, + "tokens_seen": 985202688 + }, + { + "epoch": 0.3, + "learning_rate": 0.000566852557673019, + "loss": 1.4147, + "theoretical_loss": 3.6828242963973157, + "tokens_seen": 985268224 + }, + { + "epoch": 0.3, + "learning_rate": 0.0005668365095285858, + "loss": 1.4508, + "theoretical_loss": 3.6828011013935473, + "tokens_seen": 985333760 + }, + { + "epoch": 0.3, + "learning_rate": 0.0005668204613841525, + "loss": 1.3858, + "theoretical_loss": 3.682777908364395, + "tokens_seen": 985399296 + }, + { + "epoch": 0.3, + "learning_rate": 0.0005668044132397192, + "loss": 1.4933, + "theoretical_loss": 3.682754717309558, + "tokens_seen": 985464832 + }, + { + "epoch": 0.3, + "learning_rate": 0.0005667883650952859, + "loss": 1.4678, + "theoretical_loss": 3.6827315282287376, + "tokens_seen": 985530368 + }, + { + "epoch": 0.3, + "learning_rate": 0.0005667723169508526, + "loss": 1.4153, + "theoretical_loss": 3.682708341121635, + "tokens_seen": 985595904 + }, + { + "epoch": 0.3, + "learning_rate": 0.0005667562688064193, + "loss": 1.5292, + "theoretical_loss": 3.68268515598795, + "tokens_seen": 985661440 + }, + { + "epoch": 0.3, + "learning_rate": 0.000566740220661986, + "loss": 1.4508, + "theoretical_loss": 3.6826619728273835, + "tokens_seen": 985726976 + }, + { + "epoch": 0.3, + "learning_rate": 0.0005667241725175527, + "loss": 1.5097, + "theoretical_loss": 3.6826387916396373, + "tokens_seen": 985792512 + }, + { + "epoch": 0.3, + "learning_rate": 0.0005667081243731194, + "loss": 1.6329, + "theoretical_loss": 3.6826156124244123, + "tokens_seen": 985858048 + }, + { + "epoch": 0.3, + "learning_rate": 0.0005666920762286861, + "loss": 1.3907, + "theoretical_loss": 3.6825924351814088, + "tokens_seen": 985923584 + }, + { + "epoch": 0.3, + "learning_rate": 0.0005666760280842528, + "loss": 1.501, + "theoretical_loss": 3.682569259910328, + "tokens_seen": 985989120 + }, + { + "epoch": 0.3, + "learning_rate": 0.0005666599799398195, + "loss": 1.4063, + "theoretical_loss": 3.6825460866108717, + "tokens_seen": 986054656 + }, + { + "epoch": 0.3, + "learning_rate": 0.0005666439317953862, + "loss": 1.3789, + "theoretical_loss": 3.6825229152827403, + "tokens_seen": 986120192 + }, + { + "epoch": 0.3, + "learning_rate": 0.0005666278836509529, + "loss": 1.4108, + "theoretical_loss": 3.682499745925636, + "tokens_seen": 986185728 + }, + { + "epoch": 0.3, + "learning_rate": 0.0005666118355065196, + "loss": 1.3472, + "theoretical_loss": 3.6824765785392595, + "tokens_seen": 986251264 + }, + { + "epoch": 0.3, + "objective/train/docs_used": 362627, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.4516384601593018, + "objective/train/theoretical_loss": 3.682453413123313, + "objective/train/tokens_used": 1006776800, + "theoretical_loss": 3.682453413123313, + "tokens_seen": 986316800 + }, + { + "epoch": 0.3, + "learning_rate": 0.0005665957873620863, + "loss": 1.4227, + "theoretical_loss": 3.682453413123313, + "tokens_seen": 986316800 + }, + { + "epoch": 0.3, + "learning_rate": 0.000566579739217653, + "loss": 1.4562, + "theoretical_loss": 3.682430249677497, + "tokens_seen": 986382336 + }, + { + "epoch": 0.3, + "learning_rate": 0.0005665636910732197, + "loss": 1.4417, + "theoretical_loss": 3.682407088201514, + "tokens_seen": 986447872 + }, + { + "epoch": 0.3, + "learning_rate": 0.0005665476429287864, + "loss": 1.4347, + "theoretical_loss": 3.6823839286950646, + "tokens_seen": 986513408 + }, + { + "epoch": 0.3, + "learning_rate": 0.0005665315947843531, + "loss": 1.3649, + "theoretical_loss": 3.6823607711578514, + "tokens_seen": 986578944 + }, + { + "epoch": 0.3, + "learning_rate": 0.0005665155466399198, + "loss": 1.3479, + "theoretical_loss": 3.682337615589576, + "tokens_seen": 986644480 + }, + { + "epoch": 0.3, + "learning_rate": 0.0005664994984954865, + "loss": 1.3842, + "theoretical_loss": 3.68231446198994, + "tokens_seen": 986710016 + }, + { + "epoch": 0.3, + "learning_rate": 0.0005664834503510532, + "loss": 1.4475, + "theoretical_loss": 3.682291310358645, + "tokens_seen": 986775552 + }, + { + "epoch": 0.3, + "learning_rate": 0.0005664674022066199, + "loss": 1.5143, + "theoretical_loss": 3.682268160695394, + "tokens_seen": 986841088 + }, + { + "epoch": 0.3, + "learning_rate": 0.0005664513540621866, + "loss": 1.4629, + "theoretical_loss": 3.6822450129998883, + "tokens_seen": 986906624 + }, + { + "epoch": 0.3, + "learning_rate": 0.0005664353059177533, + "loss": 1.4762, + "theoretical_loss": 3.6822218672718297, + "tokens_seen": 986972160 + }, + { + "epoch": 0.3, + "learning_rate": 0.00056641925777332, + "loss": 1.4276, + "theoretical_loss": 3.6821987235109206, + "tokens_seen": 987037696 + }, + { + "epoch": 0.3, + "learning_rate": 0.0005664032096288867, + "loss": 1.5114, + "theoretical_loss": 3.6821755817168635, + "tokens_seen": 987103232 + }, + { + "epoch": 0.3, + "learning_rate": 0.0005663871614844534, + "loss": 1.3515, + "theoretical_loss": 3.6821524418893605, + "tokens_seen": 987168768 + }, + { + "epoch": 0.3, + "learning_rate": 0.0005663711133400201, + "loss": 1.3916, + "theoretical_loss": 3.682129304028114, + "tokens_seen": 987234304 + }, + { + "epoch": 0.3, + "learning_rate": 0.0005663550651955868, + "loss": 1.5455, + "theoretical_loss": 3.682106168132826, + "tokens_seen": 987299840 + }, + { + "epoch": 0.3, + "learning_rate": 0.0005663390170511535, + "loss": 1.4468, + "theoretical_loss": 3.6820830342032, + "tokens_seen": 987365376 + }, + { + "epoch": 0.3, + "learning_rate": 0.0005663229689067202, + "loss": 1.4074, + "theoretical_loss": 3.6820599022389375, + "tokens_seen": 987430912 + }, + { + "epoch": 0.3, + "learning_rate": 0.0005663069207622869, + "loss": 1.4244, + "theoretical_loss": 3.6820367722397416, + "tokens_seen": 987496448 + }, + { + "epoch": 0.3, + "learning_rate": 0.0005662908726178536, + "loss": 1.4978, + "theoretical_loss": 3.6820136442053144, + "tokens_seen": 987561984 + }, + { + "epoch": 0.3, + "learning_rate": 0.0005662748244734203, + "loss": 1.4378, + "theoretical_loss": 3.6819905181353594, + "tokens_seen": 987627520 + }, + { + "epoch": 0.3, + "learning_rate": 0.000566258776328987, + "loss": 1.4376, + "theoretical_loss": 3.6819673940295794, + "tokens_seen": 987693056 + }, + { + "epoch": 0.3, + "learning_rate": 0.0005662427281845537, + "loss": 1.4186, + "theoretical_loss": 3.6819442718876765, + "tokens_seen": 987758592 + }, + { + "epoch": 0.3, + "learning_rate": 0.0005662266800401204, + "loss": 1.5525, + "theoretical_loss": 3.6819211517093544, + "tokens_seen": 987824128 + }, + { + "epoch": 0.3, + "learning_rate": 0.0005662106318956871, + "loss": 1.416, + "theoretical_loss": 3.6818980334943157, + "tokens_seen": 987889664 + }, + { + "epoch": 0.3, + "learning_rate": 0.0005661945837512538, + "loss": 1.4436, + "theoretical_loss": 3.681874917242263, + "tokens_seen": 987955200 + }, + { + "epoch": 0.3, + "learning_rate": 0.0005661785356068205, + "loss": 1.4959, + "theoretical_loss": 3.6818518029529006, + "tokens_seen": 988020736 + }, + { + "epoch": 0.3, + "learning_rate": 0.0005661624874623872, + "loss": 1.4796, + "theoretical_loss": 3.681828690625931, + "tokens_seen": 988086272 + }, + { + "epoch": 0.3, + "learning_rate": 0.0005661464393179539, + "loss": 1.4314, + "theoretical_loss": 3.681805580261057, + "tokens_seen": 988151808 + }, + { + "epoch": 0.3, + "learning_rate": 0.0005661303911735206, + "loss": 1.3516, + "theoretical_loss": 3.681782471857983, + "tokens_seen": 988217344 + }, + { + "epoch": 0.3, + "learning_rate": 0.0005661143430290873, + "loss": 1.3439, + "theoretical_loss": 3.6817593654164114, + "tokens_seen": 988282880 + }, + { + "epoch": 0.3, + "learning_rate": 0.000566098294884654, + "loss": 1.5097, + "theoretical_loss": 3.681736260936046, + "tokens_seen": 988348416 + }, + { + "epoch": 0.3, + "learning_rate": 0.0005660822467402207, + "loss": 1.3102, + "theoretical_loss": 3.6817131584165907, + "tokens_seen": 988413952 + }, + { + "epoch": 0.3, + "learning_rate": 0.0005660661985957874, + "loss": 1.419, + "theoretical_loss": 3.6816900578577485, + "tokens_seen": 988479488 + }, + { + "epoch": 0.3, + "learning_rate": 0.0005660501504513541, + "loss": 1.3673, + "theoretical_loss": 3.681666959259223, + "tokens_seen": 988545024 + }, + { + "epoch": 0.3, + "learning_rate": 0.0005660341023069208, + "loss": 1.4797, + "theoretical_loss": 3.681643862620719, + "tokens_seen": 988610560 + }, + { + "epoch": 0.3, + "learning_rate": 0.0005660180541624875, + "loss": 1.4539, + "theoretical_loss": 3.681620767941939, + "tokens_seen": 988676096 + }, + { + "epoch": 0.3, + "learning_rate": 0.0005660020060180542, + "loss": 1.5154, + "theoretical_loss": 3.6815976752225867, + "tokens_seen": 988741632 + }, + { + "epoch": 0.3, + "learning_rate": 0.0005659859578736209, + "loss": 1.5563, + "theoretical_loss": 3.681574584462367, + "tokens_seen": 988807168 + }, + { + "epoch": 0.3, + "learning_rate": 0.0005659699097291876, + "loss": 1.5715, + "theoretical_loss": 3.6815514956609836, + "tokens_seen": 988872704 + }, + { + "epoch": 0.3, + "learning_rate": 0.0005659538615847543, + "loss": 1.396, + "theoretical_loss": 3.6815284088181404, + "tokens_seen": 988938240 + }, + { + "epoch": 0.3, + "learning_rate": 0.000565937813440321, + "loss": 1.3834, + "theoretical_loss": 3.681505323933541, + "tokens_seen": 989003776 + }, + { + "epoch": 0.3, + "learning_rate": 0.0005659217652958877, + "loss": 1.4727, + "theoretical_loss": 3.68148224100689, + "tokens_seen": 989069312 + }, + { + "epoch": 0.3, + "learning_rate": 0.0005659057171514544, + "loss": 1.3237, + "theoretical_loss": 3.681459160037892, + "tokens_seen": 989134848 + }, + { + "epoch": 0.3, + "learning_rate": 0.0005658896690070211, + "loss": 1.4903, + "theoretical_loss": 3.6814360810262508, + "tokens_seen": 989200384 + }, + { + "epoch": 0.3, + "learning_rate": 0.0005658736208625878, + "loss": 1.4515, + "theoretical_loss": 3.681413003971671, + "tokens_seen": 989265920 + }, + { + "epoch": 0.3, + "learning_rate": 0.0005658575727181545, + "loss": 1.5772, + "theoretical_loss": 3.6813899288738563, + "tokens_seen": 989331456 + }, + { + "epoch": 0.3, + "learning_rate": 0.0005658415245737212, + "loss": 1.5194, + "theoretical_loss": 3.681366855732512, + "tokens_seen": 989396992 + }, + { + "epoch": 0.3, + "learning_rate": 0.0005658254764292879, + "loss": 1.4455, + "theoretical_loss": 3.6813437845473427, + "tokens_seen": 989462528 + }, + { + "epoch": 0.3, + "learning_rate": 0.0005658094282848546, + "loss": 1.4015, + "theoretical_loss": 3.6813207153180523, + "tokens_seen": 989528064 + }, + { + "epoch": 0.3, + "objective/train/docs_used": 363941, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.3814904689788818, + "objective/train/theoretical_loss": 3.681297648044346, + "objective/train/tokens_used": 1010053600, + "theoretical_loss": 3.681297648044346, + "tokens_seen": 989593600 + }, + { + "epoch": 0.3, + "learning_rate": 0.0005657933801404213, + "loss": 1.3702, + "theoretical_loss": 3.681297648044346, + "tokens_seen": 989593600 + }, + { + "epoch": 0.3, + "learning_rate": 0.0005657773319959879, + "loss": 1.3782, + "theoretical_loss": 3.681274582725928, + "tokens_seen": 989659136 + }, + { + "epoch": 0.3, + "learning_rate": 0.0005657612838515547, + "loss": 1.3205, + "theoretical_loss": 3.681251519362504, + "tokens_seen": 989724672 + }, + { + "epoch": 0.3, + "learning_rate": 0.0005657452357071213, + "loss": 1.3381, + "theoretical_loss": 3.681228457953778, + "tokens_seen": 989790208 + }, + { + "epoch": 0.3, + "learning_rate": 0.0005657291875626881, + "loss": 1.529, + "theoretical_loss": 3.681205398499456, + "tokens_seen": 989855744 + }, + { + "epoch": 0.3, + "learning_rate": 0.0005657131394182547, + "loss": 1.3217, + "theoretical_loss": 3.681182340999241, + "tokens_seen": 989921280 + }, + { + "epoch": 0.3, + "learning_rate": 0.0005656970912738215, + "loss": 1.512, + "theoretical_loss": 3.68115928545284, + "tokens_seen": 989986816 + }, + { + "epoch": 0.3, + "learning_rate": 0.0005656810431293881, + "loss": 1.5457, + "theoretical_loss": 3.6811362318599574, + "tokens_seen": 990052352 + }, + { + "epoch": 0.3, + "learning_rate": 0.0005656649949849549, + "loss": 1.4869, + "theoretical_loss": 3.681113180220298, + "tokens_seen": 990117888 + }, + { + "epoch": 0.3, + "learning_rate": 0.0005656489468405215, + "loss": 1.4624, + "theoretical_loss": 3.681090130533568, + "tokens_seen": 990183424 + }, + { + "epoch": 0.3, + "learning_rate": 0.0005656328986960883, + "loss": 1.5049, + "theoretical_loss": 3.6810670827994723, + "tokens_seen": 990248960 + }, + { + "epoch": 0.3, + "learning_rate": 0.0005656168505516549, + "loss": 1.4834, + "theoretical_loss": 3.6810440370177155, + "tokens_seen": 990314496 + }, + { + "epoch": 0.3, + "learning_rate": 0.0005656008024072217, + "loss": 1.4709, + "theoretical_loss": 3.681020993188004, + "tokens_seen": 990380032 + }, + { + "epoch": 0.3, + "learning_rate": 0.0005655847542627883, + "loss": 1.4155, + "theoretical_loss": 3.680997951310043, + "tokens_seen": 990445568 + }, + { + "epoch": 0.3, + "learning_rate": 0.0005655687061183551, + "loss": 1.4381, + "theoretical_loss": 3.6809749113835375, + "tokens_seen": 990511104 + }, + { + "epoch": 0.3, + "learning_rate": 0.0005655526579739217, + "loss": 1.5369, + "theoretical_loss": 3.680951873408194, + "tokens_seen": 990576640 + }, + { + "epoch": 0.3, + "learning_rate": 0.0005655366098294885, + "loss": 1.3982, + "theoretical_loss": 3.680928837383718, + "tokens_seen": 990642176 + }, + { + "epoch": 0.3, + "learning_rate": 0.0005655205616850551, + "loss": 1.3381, + "theoretical_loss": 3.680905803309815, + "tokens_seen": 990707712 + }, + { + "epoch": 0.3, + "learning_rate": 0.0005655045135406219, + "loss": 1.4397, + "theoretical_loss": 3.680882771186191, + "tokens_seen": 990773248 + }, + { + "epoch": 0.3, + "learning_rate": 0.0005654884653961886, + "loss": 1.4036, + "theoretical_loss": 3.6808597410125516, + "tokens_seen": 990838784 + }, + { + "epoch": 0.3, + "learning_rate": 0.0005654724172517553, + "loss": 1.4852, + "theoretical_loss": 3.680836712788603, + "tokens_seen": 990904320 + }, + { + "epoch": 0.3, + "learning_rate": 0.000565456369107322, + "loss": 1.4181, + "theoretical_loss": 3.6808136865140515, + "tokens_seen": 990969856 + }, + { + "epoch": 0.3, + "learning_rate": 0.0005654403209628887, + "loss": 1.3715, + "theoretical_loss": 3.6807906621886026, + "tokens_seen": 991035392 + }, + { + "epoch": 0.3, + "learning_rate": 0.0005654242728184554, + "loss": 1.5271, + "theoretical_loss": 3.6807676398119624, + "tokens_seen": 991100928 + }, + { + "epoch": 0.3, + "learning_rate": 0.0005654082246740221, + "loss": 1.3769, + "theoretical_loss": 3.6807446193838373, + "tokens_seen": 991166464 + }, + { + "epoch": 0.3, + "learning_rate": 0.0005653921765295888, + "loss": 1.3744, + "theoretical_loss": 3.6807216009039343, + "tokens_seen": 991232000 + }, + { + "epoch": 0.3, + "learning_rate": 0.0005653761283851555, + "loss": 1.7799, + "theoretical_loss": 3.680698584371959, + "tokens_seen": 991297536 + }, + { + "epoch": 0.3, + "learning_rate": 0.0005653600802407222, + "loss": 1.3478, + "theoretical_loss": 3.680675569787617, + "tokens_seen": 991363072 + }, + { + "epoch": 0.3, + "learning_rate": 0.0005653440320962889, + "loss": 1.5099, + "theoretical_loss": 3.6806525571506166, + "tokens_seen": 991428608 + }, + { + "epoch": 0.3, + "learning_rate": 0.0005653279839518556, + "loss": 1.4904, + "theoretical_loss": 3.680629546460663, + "tokens_seen": 991494144 + }, + { + "epoch": 0.3, + "learning_rate": 0.0005653119358074223, + "loss": 1.4008, + "theoretical_loss": 3.680606537717463, + "tokens_seen": 991559680 + }, + { + "epoch": 0.3, + "learning_rate": 0.000565295887662989, + "loss": 1.4402, + "theoretical_loss": 3.6805835309207238, + "tokens_seen": 991625216 + }, + { + "epoch": 0.3, + "learning_rate": 0.0005652798395185557, + "loss": 1.4328, + "theoretical_loss": 3.680560526070151, + "tokens_seen": 991690752 + }, + { + "epoch": 0.3, + "learning_rate": 0.0005652637913741224, + "loss": 1.4417, + "theoretical_loss": 3.6805375231654525, + "tokens_seen": 991756288 + }, + { + "epoch": 0.3, + "learning_rate": 0.0005652477432296891, + "loss": 1.5476, + "theoretical_loss": 3.6805145222063347, + "tokens_seen": 991821824 + }, + { + "epoch": 0.3, + "learning_rate": 0.0005652316950852558, + "loss": 1.5067, + "theoretical_loss": 3.6804915231925044, + "tokens_seen": 991887360 + }, + { + "epoch": 0.3, + "learning_rate": 0.0005652156469408225, + "loss": 1.6396, + "theoretical_loss": 3.680468526123669, + "tokens_seen": 991952896 + }, + { + "epoch": 0.3, + "learning_rate": 0.0005651995987963892, + "loss": 1.5291, + "theoretical_loss": 3.6804455309995348, + "tokens_seen": 992018432 + }, + { + "epoch": 0.3, + "learning_rate": 0.0005651835506519559, + "loss": 1.5605, + "theoretical_loss": 3.6804225378198097, + "tokens_seen": 992083968 + }, + { + "epoch": 0.3, + "learning_rate": 0.0005651675025075226, + "loss": 1.3743, + "theoretical_loss": 3.6803995465842, + "tokens_seen": 992149504 + }, + { + "epoch": 0.3, + "learning_rate": 0.0005651514543630893, + "loss": 1.4133, + "theoretical_loss": 3.680376557292414, + "tokens_seen": 992215040 + }, + { + "epoch": 0.3, + "learning_rate": 0.000565135406218656, + "loss": 1.4976, + "theoretical_loss": 3.680353569944158, + "tokens_seen": 992280576 + }, + { + "epoch": 0.3, + "learning_rate": 0.0005651193580742227, + "loss": 1.3846, + "theoretical_loss": 3.6803305845391394, + "tokens_seen": 992346112 + }, + { + "epoch": 0.3, + "learning_rate": 0.0005651033099297894, + "loss": 1.4959, + "theoretical_loss": 3.680307601077066, + "tokens_seen": 992411648 + }, + { + "epoch": 0.3, + "learning_rate": 0.0005650872617853561, + "loss": 1.3815, + "theoretical_loss": 3.6802846195576455, + "tokens_seen": 992477184 + }, + { + "epoch": 0.3, + "learning_rate": 0.0005650712136409228, + "loss": 1.4584, + "theoretical_loss": 3.6802616399805848, + "tokens_seen": 992542720 + }, + { + "epoch": 0.3, + "learning_rate": 0.0005650551654964895, + "loss": 1.4742, + "theoretical_loss": 3.680238662345592, + "tokens_seen": 992608256 + }, + { + "epoch": 0.3, + "learning_rate": 0.0005650391173520562, + "loss": 1.5429, + "theoretical_loss": 3.6802156866523745, + "tokens_seen": 992673792 + }, + { + "epoch": 0.3, + "learning_rate": 0.0005650230692076229, + "loss": 1.4641, + "theoretical_loss": 3.6801927129006407, + "tokens_seen": 992739328 + }, + { + "epoch": 0.3, + "learning_rate": 0.0005650070210631896, + "loss": 1.3931, + "theoretical_loss": 3.680169741090097, + "tokens_seen": 992804864 + }, + { + "epoch": 0.3, + "objective/train/docs_used": 365105, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.4506094455718994, + "objective/train/theoretical_loss": 3.6801467712204525, + "objective/train/tokens_used": 1013330400, + "theoretical_loss": 3.6801467712204525, + "tokens_seen": 992870400 + }, + { + "epoch": 0.3, + "learning_rate": 0.0005649909729187563, + "loss": 1.4011, + "theoretical_loss": 3.6801467712204525, + "tokens_seen": 992870400 + }, + { + "epoch": 0.3, + "learning_rate": 0.000564974924774323, + "loss": 1.5404, + "theoretical_loss": 3.6801238032914148, + "tokens_seen": 992935936 + }, + { + "epoch": 0.3, + "learning_rate": 0.0005649588766298897, + "loss": 1.5384, + "theoretical_loss": 3.6801008373026916, + "tokens_seen": 993001472 + }, + { + "epoch": 0.3, + "learning_rate": 0.0005649428284854564, + "loss": 1.5477, + "theoretical_loss": 3.6800778732539907, + "tokens_seen": 993067008 + }, + { + "epoch": 0.3, + "learning_rate": 0.0005649267803410231, + "loss": 1.4607, + "theoretical_loss": 3.680054911145021, + "tokens_seen": 993132544 + }, + { + "epoch": 0.3, + "learning_rate": 0.0005649107321965898, + "loss": 1.3569, + "theoretical_loss": 3.68003195097549, + "tokens_seen": 993198080 + }, + { + "epoch": 0.3, + "learning_rate": 0.0005648946840521565, + "loss": 1.42, + "theoretical_loss": 3.6800089927451065, + "tokens_seen": 993263616 + }, + { + "epoch": 0.3, + "learning_rate": 0.0005648786359077232, + "loss": 1.4723, + "theoretical_loss": 3.6799860364535784, + "tokens_seen": 993329152 + }, + { + "epoch": 0.3, + "learning_rate": 0.0005648625877632899, + "loss": 1.4848, + "theoretical_loss": 3.679963082100614, + "tokens_seen": 993394688 + }, + { + "epoch": 0.3, + "learning_rate": 0.0005648465396188566, + "loss": 1.4103, + "theoretical_loss": 3.6799401296859227, + "tokens_seen": 993460224 + }, + { + "epoch": 0.3, + "learning_rate": 0.0005648304914744233, + "loss": 1.386, + "theoretical_loss": 3.6799171792092116, + "tokens_seen": 993525760 + }, + { + "epoch": 0.3, + "learning_rate": 0.00056481444332999, + "loss": 1.2779, + "theoretical_loss": 3.6798942306701896, + "tokens_seen": 993591296 + }, + { + "epoch": 0.3, + "learning_rate": 0.0005647983951855567, + "loss": 1.3938, + "theoretical_loss": 3.6798712840685655, + "tokens_seen": 993656832 + }, + { + "epoch": 0.3, + "learning_rate": 0.0005647823470411234, + "loss": 1.4465, + "theoretical_loss": 3.679848339404048, + "tokens_seen": 993722368 + }, + { + "epoch": 0.3, + "learning_rate": 0.0005647662988966901, + "loss": 1.5016, + "theoretical_loss": 3.679825396676346, + "tokens_seen": 993787904 + }, + { + "epoch": 0.3, + "learning_rate": 0.0005647502507522568, + "loss": 1.3619, + "theoretical_loss": 3.6798024558851683, + "tokens_seen": 993853440 + }, + { + "epoch": 0.3, + "learning_rate": 0.0005647342026078235, + "loss": 1.4364, + "theoretical_loss": 3.679779517030224, + "tokens_seen": 993918976 + }, + { + "epoch": 0.3, + "learning_rate": 0.0005647181544633902, + "loss": 1.4054, + "theoretical_loss": 3.6797565801112206, + "tokens_seen": 993984512 + }, + { + "epoch": 0.3, + "learning_rate": 0.0005647021063189569, + "loss": 1.5398, + "theoretical_loss": 3.679733645127869, + "tokens_seen": 994050048 + }, + { + "epoch": 0.3, + "learning_rate": 0.0005646860581745236, + "loss": 1.3947, + "theoretical_loss": 3.679710712079877, + "tokens_seen": 994115584 + }, + { + "epoch": 0.3, + "learning_rate": 0.0005646700100300903, + "loss": 1.4366, + "theoretical_loss": 3.679687780966954, + "tokens_seen": 994181120 + }, + { + "epoch": 0.3, + "learning_rate": 0.000564653961885657, + "loss": 1.3041, + "theoretical_loss": 3.679664851788809, + "tokens_seen": 994246656 + }, + { + "epoch": 0.3, + "learning_rate": 0.0005646379137412237, + "loss": 1.3603, + "theoretical_loss": 3.6796419245451526, + "tokens_seen": 994312192 + }, + { + "epoch": 0.3, + "learning_rate": 0.0005646218655967904, + "loss": 1.5254, + "theoretical_loss": 3.679618999235692, + "tokens_seen": 994377728 + }, + { + "epoch": 0.3, + "learning_rate": 0.0005646058174523571, + "loss": 1.5019, + "theoretical_loss": 3.679596075860138, + "tokens_seen": 994443264 + }, + { + "epoch": 0.3, + "learning_rate": 0.0005645897693079238, + "loss": 1.5339, + "theoretical_loss": 3.6795731544181995, + "tokens_seen": 994508800 + }, + { + "epoch": 0.3, + "learning_rate": 0.0005645737211634905, + "loss": 1.3799, + "theoretical_loss": 3.679550234909586, + "tokens_seen": 994574336 + }, + { + "epoch": 0.3, + "learning_rate": 0.0005645576730190573, + "loss": 1.4762, + "theoretical_loss": 3.6795273173340073, + "tokens_seen": 994639872 + }, + { + "epoch": 0.3, + "learning_rate": 0.0005645416248746239, + "loss": 1.4587, + "theoretical_loss": 3.6795044016911724, + "tokens_seen": 994705408 + }, + { + "epoch": 0.3, + "learning_rate": 0.0005645255767301906, + "loss": 1.4118, + "theoretical_loss": 3.679481487980792, + "tokens_seen": 994770944 + }, + { + "epoch": 0.3, + "learning_rate": 0.0005645095285857573, + "loss": 1.4255, + "theoretical_loss": 3.6794585762025753, + "tokens_seen": 994836480 + }, + { + "epoch": 0.3, + "learning_rate": 0.000564493480441324, + "loss": 1.4682, + "theoretical_loss": 3.679435666356232, + "tokens_seen": 994902016 + }, + { + "epoch": 0.3, + "learning_rate": 0.0005644774322968907, + "loss": 1.5212, + "theoretical_loss": 3.679412758441472, + "tokens_seen": 994967552 + }, + { + "epoch": 0.3, + "learning_rate": 0.0005644613841524574, + "loss": 1.4152, + "theoretical_loss": 3.679389852458005, + "tokens_seen": 995033088 + }, + { + "epoch": 0.3, + "learning_rate": 0.0005644453360080241, + "loss": 1.4356, + "theoretical_loss": 3.6793669484055416, + "tokens_seen": 995098624 + }, + { + "epoch": 0.3, + "learning_rate": 0.0005644292878635908, + "loss": 1.4786, + "theoretical_loss": 3.6793440462837923, + "tokens_seen": 995164160 + }, + { + "epoch": 0.3, + "learning_rate": 0.0005644132397191575, + "loss": 1.4288, + "theoretical_loss": 3.6793211460924655, + "tokens_seen": 995229696 + }, + { + "epoch": 0.3, + "learning_rate": 0.0005643971915747242, + "loss": 1.4354, + "theoretical_loss": 3.679298247831272, + "tokens_seen": 995295232 + }, + { + "epoch": 0.3, + "learning_rate": 0.0005643811434302908, + "loss": 1.4357, + "theoretical_loss": 3.679275351499924, + "tokens_seen": 995360768 + }, + { + "epoch": 0.3, + "learning_rate": 0.0005643650952858576, + "loss": 1.4014, + "theoretical_loss": 3.679252457098129, + "tokens_seen": 995426304 + }, + { + "epoch": 0.3, + "learning_rate": 0.0005643490471414242, + "loss": 1.4796, + "theoretical_loss": 3.6792295646255986, + "tokens_seen": 995491840 + }, + { + "epoch": 0.3, + "learning_rate": 0.000564332998996991, + "loss": 1.5649, + "theoretical_loss": 3.6792066740820433, + "tokens_seen": 995557376 + }, + { + "epoch": 0.3, + "learning_rate": 0.0005643169508525576, + "loss": 1.4996, + "theoretical_loss": 3.679183785467173, + "tokens_seen": 995622912 + }, + { + "epoch": 0.3, + "learning_rate": 0.0005643009027081244, + "loss": 1.47, + "theoretical_loss": 3.6791608987807, + "tokens_seen": 995688448 + }, + { + "epoch": 0.3, + "learning_rate": 0.000564284854563691, + "loss": 1.4709, + "theoretical_loss": 3.6791380140223326, + "tokens_seen": 995753984 + }, + { + "epoch": 0.3, + "learning_rate": 0.0005642688064192578, + "loss": 1.5928, + "theoretical_loss": 3.679115131191783, + "tokens_seen": 995819520 + }, + { + "epoch": 0.3, + "learning_rate": 0.0005642527582748244, + "loss": 1.3916, + "theoretical_loss": 3.6790922502887615, + "tokens_seen": 995885056 + }, + { + "epoch": 0.3, + "learning_rate": 0.0005642367101303912, + "loss": 1.3263, + "theoretical_loss": 3.6790693713129787, + "tokens_seen": 995950592 + }, + { + "epoch": 0.3, + "learning_rate": 0.0005642206619859578, + "loss": 1.4735, + "theoretical_loss": 3.679046494264146, + "tokens_seen": 996016128 + }, + { + "epoch": 0.3, + "learning_rate": 0.0005642046138415246, + "loss": 1.5021, + "theoretical_loss": 3.679023619141974, + "tokens_seen": 996081664 + }, + { + "epoch": 0.3, + "objective/train/docs_used": 366329, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.548703908920288, + "objective/train/theoretical_loss": 3.6790007459461735, + "objective/train/tokens_used": 1016607200, + "theoretical_loss": 3.6790007459461735, + "tokens_seen": 996147200 + }, + { + "epoch": 0.3, + "learning_rate": 0.0005641885656970912, + "loss": 1.4231, + "theoretical_loss": 3.6790007459461735, + "tokens_seen": 996147200 + }, + { + "epoch": 0.3, + "learning_rate": 0.000564172517552658, + "loss": 1.4681, + "theoretical_loss": 3.678977874676456, + "tokens_seen": 996212736 + }, + { + "epoch": 0.3, + "learning_rate": 0.0005641564694082246, + "loss": 1.5496, + "theoretical_loss": 3.678955005332532, + "tokens_seen": 996278272 + }, + { + "epoch": 0.3, + "learning_rate": 0.0005641404212637914, + "loss": 1.2554, + "theoretical_loss": 3.6789321379141136, + "tokens_seen": 996343808 + }, + { + "epoch": 0.3, + "learning_rate": 0.000564124373119358, + "loss": 1.4728, + "theoretical_loss": 3.6789092724209107, + "tokens_seen": 996409344 + }, + { + "epoch": 0.3, + "learning_rate": 0.0005641083249749248, + "loss": 1.3209, + "theoretical_loss": 3.6788864088526356, + "tokens_seen": 996474880 + }, + { + "epoch": 0.3, + "learning_rate": 0.0005640922768304915, + "loss": 1.3755, + "theoretical_loss": 3.678863547209, + "tokens_seen": 996540416 + }, + { + "epoch": 0.3, + "learning_rate": 0.0005640762286860582, + "loss": 1.4449, + "theoretical_loss": 3.6788406874897146, + "tokens_seen": 996605952 + }, + { + "epoch": 0.3, + "learning_rate": 0.0005640601805416249, + "loss": 1.509, + "theoretical_loss": 3.6788178296944913, + "tokens_seen": 996671488 + }, + { + "epoch": 0.3, + "learning_rate": 0.0005640441323971916, + "loss": 1.5483, + "theoretical_loss": 3.6787949738230408, + "tokens_seen": 996737024 + }, + { + "epoch": 0.3, + "learning_rate": 0.0005640280842527583, + "loss": 1.5004, + "theoretical_loss": 3.678772119875076, + "tokens_seen": 996802560 + }, + { + "epoch": 0.3, + "learning_rate": 0.000564012036108325, + "loss": 1.366, + "theoretical_loss": 3.678749267850308, + "tokens_seen": 996868096 + }, + { + "epoch": 0.3, + "learning_rate": 0.0005639959879638917, + "loss": 1.5439, + "theoretical_loss": 3.678726417748448, + "tokens_seen": 996933632 + }, + { + "epoch": 0.3, + "learning_rate": 0.0005639799398194584, + "loss": 1.4163, + "theoretical_loss": 3.6787035695692083, + "tokens_seen": 996999168 + }, + { + "epoch": 0.3, + "learning_rate": 0.0005639638916750251, + "loss": 1.4826, + "theoretical_loss": 3.678680723312301, + "tokens_seen": 997064704 + }, + { + "epoch": 0.3, + "learning_rate": 0.0005639478435305918, + "loss": 1.4224, + "theoretical_loss": 3.678657878977438, + "tokens_seen": 997130240 + }, + { + "epoch": 0.3, + "learning_rate": 0.0005639317953861585, + "loss": 1.3412, + "theoretical_loss": 3.6786350365643306, + "tokens_seen": 997195776 + }, + { + "epoch": 0.3, + "learning_rate": 0.0005639157472417252, + "loss": 1.4453, + "theoretical_loss": 3.6786121960726916, + "tokens_seen": 997261312 + }, + { + "epoch": 0.3, + "learning_rate": 0.0005638996990972919, + "loss": 1.4518, + "theoretical_loss": 3.678589357502233, + "tokens_seen": 997326848 + }, + { + "epoch": 0.3, + "learning_rate": 0.0005638836509528586, + "loss": 1.4207, + "theoretical_loss": 3.6785665208526668, + "tokens_seen": 997392384 + }, + { + "epoch": 0.3, + "learning_rate": 0.0005638676028084253, + "loss": 1.3462, + "theoretical_loss": 3.678543686123705, + "tokens_seen": 997457920 + }, + { + "epoch": 0.3, + "learning_rate": 0.000563851554663992, + "loss": 1.4585, + "theoretical_loss": 3.67852085331506, + "tokens_seen": 997523456 + }, + { + "epoch": 0.3, + "learning_rate": 0.0005638355065195587, + "loss": 1.4763, + "theoretical_loss": 3.6784980224264445, + "tokens_seen": 997588992 + }, + { + "epoch": 0.3, + "learning_rate": 0.0005638194583751254, + "loss": 1.3981, + "theoretical_loss": 3.6784751934575706, + "tokens_seen": 997654528 + }, + { + "epoch": 0.3, + "learning_rate": 0.0005638034102306921, + "loss": 1.5099, + "theoretical_loss": 3.6784523664081514, + "tokens_seen": 997720064 + }, + { + "epoch": 0.3, + "learning_rate": 0.0005637873620862588, + "loss": 1.5406, + "theoretical_loss": 3.6784295412778985, + "tokens_seen": 997785600 + }, + { + "epoch": 0.3, + "learning_rate": 0.0005637713139418255, + "loss": 1.4624, + "theoretical_loss": 3.678406718066525, + "tokens_seen": 997851136 + }, + { + "epoch": 0.3, + "learning_rate": 0.0005637552657973922, + "loss": 1.5529, + "theoretical_loss": 3.6783838967737443, + "tokens_seen": 997916672 + }, + { + "epoch": 0.3, + "learning_rate": 0.0005637392176529589, + "loss": 1.4605, + "theoretical_loss": 3.6783610773992677, + "tokens_seen": 997982208 + }, + { + "epoch": 0.3, + "learning_rate": 0.0005637231695085256, + "loss": 1.5824, + "theoretical_loss": 3.6783382599428087, + "tokens_seen": 998047744 + }, + { + "epoch": 0.3, + "learning_rate": 0.0005637071213640923, + "loss": 1.305, + "theoretical_loss": 3.67831544440408, + "tokens_seen": 998113280 + }, + { + "epoch": 0.3, + "learning_rate": 0.000563691073219659, + "loss": 1.4536, + "theoretical_loss": 3.6782926307827952, + "tokens_seen": 998178816 + }, + { + "epoch": 0.3, + "learning_rate": 0.0005636750250752257, + "loss": 1.5321, + "theoretical_loss": 3.6782698190786665, + "tokens_seen": 998244352 + }, + { + "epoch": 0.3, + "learning_rate": 0.0005636589769307924, + "loss": 1.5366, + "theoretical_loss": 3.6782470092914075, + "tokens_seen": 998309888 + }, + { + "epoch": 0.3, + "learning_rate": 0.0005636429287863591, + "loss": 1.5175, + "theoretical_loss": 3.67822420142073, + "tokens_seen": 998375424 + }, + { + "epoch": 0.3, + "learning_rate": 0.0005636268806419258, + "loss": 1.5736, + "theoretical_loss": 3.678201395466349, + "tokens_seen": 998440960 + }, + { + "epoch": 0.3, + "learning_rate": 0.0005636108324974925, + "loss": 1.4944, + "theoretical_loss": 3.6781785914279768, + "tokens_seen": 998506496 + }, + { + "epoch": 0.3, + "learning_rate": 0.0005635947843530592, + "loss": 1.4694, + "theoretical_loss": 3.6781557893053267, + "tokens_seen": 998572032 + }, + { + "epoch": 0.3, + "learning_rate": 0.0005635787362086259, + "loss": 1.4094, + "theoretical_loss": 3.6781329890981125, + "tokens_seen": 998637568 + }, + { + "epoch": 0.3, + "learning_rate": 0.0005635626880641926, + "loss": 1.3324, + "theoretical_loss": 3.6781101908060467, + "tokens_seen": 998703104 + }, + { + "epoch": 0.3, + "learning_rate": 0.0005635466399197593, + "loss": 1.4649, + "theoretical_loss": 3.678087394428843, + "tokens_seen": 998768640 + }, + { + "epoch": 0.3, + "learning_rate": 0.000563530591775326, + "loss": 1.418, + "theoretical_loss": 3.6780645999662163, + "tokens_seen": 998834176 + }, + { + "epoch": 0.3, + "learning_rate": 0.0005635145436308927, + "loss": 1.6081, + "theoretical_loss": 3.6780418074178782, + "tokens_seen": 998899712 + }, + { + "epoch": 0.3, + "learning_rate": 0.0005634984954864594, + "loss": 1.4762, + "theoretical_loss": 3.678019016783544, + "tokens_seen": 998965248 + }, + { + "epoch": 0.3, + "learning_rate": 0.0005634824473420261, + "loss": 1.5517, + "theoretical_loss": 3.677996228062926, + "tokens_seen": 999030784 + }, + { + "epoch": 0.3, + "learning_rate": 0.0005634663991975928, + "loss": 1.4338, + "theoretical_loss": 3.677973441255739, + "tokens_seen": 999096320 + }, + { + "epoch": 0.3, + "learning_rate": 0.0005634503510531595, + "loss": 1.475, + "theoretical_loss": 3.6779506563616966, + "tokens_seen": 999161856 + }, + { + "epoch": 0.3, + "learning_rate": 0.0005634343029087262, + "loss": 1.4284, + "theoretical_loss": 3.677927873380513, + "tokens_seen": 999227392 + }, + { + "epoch": 0.3, + "learning_rate": 0.0005634182547642929, + "loss": 1.4528, + "theoretical_loss": 3.6779050923119017, + "tokens_seen": 999292928 + }, + { + "epoch": 0.3, + "learning_rate": 0.0005634022066198596, + "loss": 1.4578, + "theoretical_loss": 3.6778823131555765, + "tokens_seen": 999358464 + }, + { + "epoch": 0.3, + "objective/train/docs_used": 367579, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.5975769758224487, + "objective/train/theoretical_loss": 3.677859535911252, + "objective/train/tokens_used": 1019884000, + "theoretical_loss": 3.677859535911252, + "tokens_seen": 999424000 + }, + { + "epoch": 0.3, + "learning_rate": 0.0005633861584754263, + "loss": 1.5073, + "theoretical_loss": 3.677859535911252, + "tokens_seen": 999424000 + }, + { + "epoch": 0.3, + "learning_rate": 0.000563370110330993, + "loss": 1.3907, + "theoretical_loss": 3.677836760578643, + "tokens_seen": 999489536 + }, + { + "epoch": 0.3, + "learning_rate": 0.0005633540621865598, + "loss": 1.4739, + "theoretical_loss": 3.677813987157462, + "tokens_seen": 999555072 + }, + { + "epoch": 0.3, + "learning_rate": 0.0005633380140421264, + "loss": 1.4303, + "theoretical_loss": 3.6777912156474244, + "tokens_seen": 999620608 + }, + { + "epoch": 0.3, + "learning_rate": 0.0005633219658976932, + "loss": 1.4203, + "theoretical_loss": 3.677768446048244, + "tokens_seen": 999686144 + }, + { + "epoch": 0.3, + "learning_rate": 0.0005633059177532598, + "loss": 1.3923, + "theoretical_loss": 3.677745678359636, + "tokens_seen": 999751680 + }, + { + "epoch": 0.3, + "learning_rate": 0.0005632898696088266, + "loss": 1.418, + "theoretical_loss": 3.6777229125813142, + "tokens_seen": 999817216 + }, + { + "epoch": 0.3, + "learning_rate": 0.0005632738214643932, + "loss": 1.4, + "theoretical_loss": 3.6777001487129937, + "tokens_seen": 999882752 + }, + { + "epoch": 0.3, + "learning_rate": 0.00056325777331996, + "loss": 1.5371, + "theoretical_loss": 3.677677386754388, + "tokens_seen": 999948288 + }, + { + "epoch": 0.3, + "learning_rate": 0.0005632417251755266, + "loss": 1.4282, + "theoretical_loss": 3.677654626705213, + "tokens_seen": 1000013824 + }, + { + "epoch": 0.3, + "learning_rate": 0.0005632256770310934, + "loss": 1.4734, + "theoretical_loss": 3.6776318685651823, + "tokens_seen": 1000079360 + }, + { + "epoch": 0.3, + "learning_rate": 0.00056320962888666, + "loss": 1.4518, + "theoretical_loss": 3.6776091123340118, + "tokens_seen": 1000144896 + }, + { + "epoch": 0.3, + "learning_rate": 0.0005631935807422268, + "loss": 1.4437, + "theoretical_loss": 3.6775863580114154, + "tokens_seen": 1000210432 + }, + { + "epoch": 0.3, + "learning_rate": 0.0005631775325977934, + "loss": 1.5461, + "theoretical_loss": 3.6775636055971086, + "tokens_seen": 1000275968 + }, + { + "epoch": 0.3, + "learning_rate": 0.0005631614844533602, + "loss": 1.4207, + "theoretical_loss": 3.6775408550908057, + "tokens_seen": 1000341504 + }, + { + "epoch": 0.3, + "learning_rate": 0.0005631454363089268, + "loss": 1.2845, + "theoretical_loss": 3.677518106492222, + "tokens_seen": 1000407040 + }, + { + "epoch": 0.3, + "learning_rate": 0.0005631293881644935, + "loss": 1.5251, + "theoretical_loss": 3.6774953598010733, + "tokens_seen": 1000472576 + }, + { + "epoch": 0.3, + "learning_rate": 0.0005631133400200602, + "loss": 1.422, + "theoretical_loss": 3.6774726150170736, + "tokens_seen": 1000538112 + }, + { + "epoch": 0.3, + "learning_rate": 0.0005630972918756269, + "loss": 1.499, + "theoretical_loss": 3.6774498721399387, + "tokens_seen": 1000603648 + }, + { + "epoch": 0.3, + "learning_rate": 0.0005630812437311936, + "loss": 1.5219, + "theoretical_loss": 3.6774271311693836, + "tokens_seen": 1000669184 + }, + { + "epoch": 0.3, + "learning_rate": 0.0005630651955867603, + "loss": 1.4791, + "theoretical_loss": 3.6774043921051236, + "tokens_seen": 1000734720 + }, + { + "epoch": 0.3, + "learning_rate": 0.000563049147442327, + "loss": 1.3718, + "theoretical_loss": 3.677381654946875, + "tokens_seen": 1000800256 + }, + { + "epoch": 0.3, + "learning_rate": 0.0005630330992978937, + "loss": 1.4408, + "theoretical_loss": 3.6773589196943517, + "tokens_seen": 1000865792 + }, + { + "epoch": 0.3, + "learning_rate": 0.0005630170511534604, + "loss": 1.385, + "theoretical_loss": 3.6773361863472704, + "tokens_seen": 1000931328 + }, + { + "epoch": 0.3, + "learning_rate": 0.0005630010030090271, + "loss": 1.498, + "theoretical_loss": 3.677313454905346, + "tokens_seen": 1000996864 + }, + { + "epoch": 0.3, + "learning_rate": 0.0005629849548645937, + "loss": 1.4224, + "theoretical_loss": 3.6772907253682945, + "tokens_seen": 1001062400 + }, + { + "epoch": 0.3, + "learning_rate": 0.0005629689067201605, + "loss": 1.3447, + "theoretical_loss": 3.6772679977358314, + "tokens_seen": 1001127936 + }, + { + "epoch": 0.3, + "learning_rate": 0.0005629528585757271, + "loss": 1.3259, + "theoretical_loss": 3.6772452720076725, + "tokens_seen": 1001193472 + }, + { + "epoch": 0.3, + "learning_rate": 0.0005629368104312939, + "loss": 1.2987, + "theoretical_loss": 3.6772225481835337, + "tokens_seen": 1001259008 + }, + { + "epoch": 0.3, + "learning_rate": 0.0005629207622868605, + "loss": 1.4102, + "theoretical_loss": 3.6771998262631307, + "tokens_seen": 1001324544 + }, + { + "epoch": 0.3, + "learning_rate": 0.0005629047141424273, + "loss": 1.3969, + "theoretical_loss": 3.6771771062461793, + "tokens_seen": 1001390080 + }, + { + "epoch": 0.3, + "learning_rate": 0.0005628886659979939, + "loss": 1.4286, + "theoretical_loss": 3.6771543881323963, + "tokens_seen": 1001455616 + }, + { + "epoch": 0.3, + "learning_rate": 0.0005628726178535607, + "loss": 1.5381, + "theoretical_loss": 3.6771316719214964, + "tokens_seen": 1001521152 + }, + { + "epoch": 0.3, + "learning_rate": 0.0005628565697091273, + "loss": 1.4014, + "theoretical_loss": 3.677108957613197, + "tokens_seen": 1001586688 + }, + { + "epoch": 0.3, + "learning_rate": 0.0005628405215646941, + "loss": 1.3655, + "theoretical_loss": 3.6770862452072133, + "tokens_seen": 1001652224 + }, + { + "epoch": 0.3, + "learning_rate": 0.0005628244734202607, + "loss": 1.4327, + "theoretical_loss": 3.6770635347032616, + "tokens_seen": 1001717760 + }, + { + "epoch": 0.3, + "learning_rate": 0.0005628084252758275, + "loss": 1.4574, + "theoretical_loss": 3.6770408261010594, + "tokens_seen": 1001783296 + }, + { + "epoch": 0.3, + "learning_rate": 0.0005627923771313941, + "loss": 1.3668, + "theoretical_loss": 3.677018119400322, + "tokens_seen": 1001848832 + }, + { + "epoch": 0.3, + "learning_rate": 0.0005627763289869609, + "loss": 1.4197, + "theoretical_loss": 3.6769954146007655, + "tokens_seen": 1001914368 + }, + { + "epoch": 0.3, + "learning_rate": 0.0005627602808425276, + "loss": 1.4565, + "theoretical_loss": 3.676972711702107, + "tokens_seen": 1001979904 + }, + { + "epoch": 0.3, + "learning_rate": 0.0005627442326980943, + "loss": 1.3872, + "theoretical_loss": 3.676950010704063, + "tokens_seen": 1002045440 + }, + { + "epoch": 0.3, + "learning_rate": 0.000562728184553661, + "loss": 1.4658, + "theoretical_loss": 3.6769273116063506, + "tokens_seen": 1002110976 + }, + { + "epoch": 0.3, + "learning_rate": 0.0005627121364092277, + "loss": 1.545, + "theoretical_loss": 3.676904614408685, + "tokens_seen": 1002176512 + }, + { + "epoch": 0.3, + "learning_rate": 0.0005626960882647944, + "loss": 1.4263, + "theoretical_loss": 3.6768819191107847, + "tokens_seen": 1002242048 + }, + { + "epoch": 0.3, + "learning_rate": 0.0005626800401203611, + "loss": 1.4636, + "theoretical_loss": 3.6768592257123647, + "tokens_seen": 1002307584 + }, + { + "epoch": 0.3, + "learning_rate": 0.0005626639919759278, + "loss": 1.5706, + "theoretical_loss": 3.6768365342131437, + "tokens_seen": 1002373120 + }, + { + "epoch": 0.3, + "learning_rate": 0.0005626479438314945, + "loss": 1.4353, + "theoretical_loss": 3.6768138446128367, + "tokens_seen": 1002438656 + }, + { + "epoch": 0.3, + "learning_rate": 0.0005626318956870612, + "loss": 1.3643, + "theoretical_loss": 3.676791156911162, + "tokens_seen": 1002504192 + }, + { + "epoch": 0.3, + "learning_rate": 0.0005626158475426279, + "loss": 1.5284, + "theoretical_loss": 3.6767684711078363, + "tokens_seen": 1002569728 + }, + { + "epoch": 0.3, + "learning_rate": 0.0005625997993981946, + "loss": 1.3748, + "theoretical_loss": 3.6767457872025764, + "tokens_seen": 1002635264 + }, + { + "epoch": 0.3, + "objective/train/docs_used": 368716, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.5292733907699585, + "objective/train/theoretical_loss": 3.6767231051950997, + "objective/train/tokens_used": 1023160800, + "theoretical_loss": 3.6767231051950997, + "tokens_seen": 1002700800 + }, + { + "epoch": 0.3, + "learning_rate": 0.0005625837512537613, + "loss": 1.4855, + "theoretical_loss": 3.6767231051950997, + "tokens_seen": 1002700800 + }, + { + "epoch": 0.3, + "learning_rate": 0.000562567703109328, + "loss": 1.444, + "theoretical_loss": 3.676700425085123, + "tokens_seen": 1002766336 + }, + { + "epoch": 0.3, + "learning_rate": 0.0005625516549648947, + "loss": 1.3212, + "theoretical_loss": 3.6766777468723646, + "tokens_seen": 1002831872 + }, + { + "epoch": 0.3, + "learning_rate": 0.0005625356068204614, + "loss": 1.4327, + "theoretical_loss": 3.6766550705565404, + "tokens_seen": 1002897408 + }, + { + "epoch": 0.3, + "learning_rate": 0.0005625195586760281, + "loss": 1.4479, + "theoretical_loss": 3.676632396137369, + "tokens_seen": 1002962944 + }, + { + "epoch": 0.3, + "learning_rate": 0.0005625035105315948, + "loss": 1.4124, + "theoretical_loss": 3.676609723614567, + "tokens_seen": 1003028480 + }, + { + "epoch": 0.3, + "learning_rate": 0.0005624874623871615, + "loss": 1.384, + "theoretical_loss": 3.6765870529878524, + "tokens_seen": 1003094016 + }, + { + "epoch": 0.3, + "learning_rate": 0.0005624714142427282, + "loss": 1.3625, + "theoretical_loss": 3.6765643842569427, + "tokens_seen": 1003159552 + }, + { + "epoch": 0.3, + "learning_rate": 0.0005624553660982949, + "loss": 1.43, + "theoretical_loss": 3.6765417174215553, + "tokens_seen": 1003225088 + }, + { + "epoch": 0.3, + "learning_rate": 0.0005624393179538616, + "loss": 1.5165, + "theoretical_loss": 3.6765190524814084, + "tokens_seen": 1003290624 + }, + { + "epoch": 0.3, + "learning_rate": 0.0005624232698094283, + "loss": 1.5021, + "theoretical_loss": 3.6764963894362195, + "tokens_seen": 1003356160 + }, + { + "epoch": 0.3, + "learning_rate": 0.000562407221664995, + "loss": 1.4778, + "theoretical_loss": 3.676473728285706, + "tokens_seen": 1003421696 + }, + { + "epoch": 0.3, + "learning_rate": 0.0005623911735205617, + "loss": 1.3837, + "theoretical_loss": 3.676451069029586, + "tokens_seen": 1003487232 + }, + { + "epoch": 0.3, + "learning_rate": 0.0005623751253761284, + "loss": 1.5132, + "theoretical_loss": 3.676428411667578, + "tokens_seen": 1003552768 + }, + { + "epoch": 0.3, + "learning_rate": 0.0005623590772316951, + "loss": 1.5029, + "theoretical_loss": 3.676405756199399, + "tokens_seen": 1003618304 + }, + { + "epoch": 0.3, + "learning_rate": 0.0005623430290872618, + "loss": 1.5409, + "theoretical_loss": 3.6763831026247678, + "tokens_seen": 1003683840 + }, + { + "epoch": 0.3, + "learning_rate": 0.0005623269809428285, + "loss": 1.5435, + "theoretical_loss": 3.6763604509434025, + "tokens_seen": 1003749376 + }, + { + "epoch": 0.3, + "learning_rate": 0.0005623109327983952, + "loss": 1.4489, + "theoretical_loss": 3.6763378011550207, + "tokens_seen": 1003814912 + }, + { + "epoch": 0.3, + "learning_rate": 0.0005622948846539619, + "loss": 1.4823, + "theoretical_loss": 3.676315153259341, + "tokens_seen": 1003880448 + }, + { + "epoch": 0.3, + "learning_rate": 0.0005622788365095286, + "loss": 1.5115, + "theoretical_loss": 3.6762925072560817, + "tokens_seen": 1003945984 + }, + { + "epoch": 0.3, + "learning_rate": 0.0005622627883650953, + "loss": 1.4378, + "theoretical_loss": 3.6762698631449613, + "tokens_seen": 1004011520 + }, + { + "epoch": 0.3, + "learning_rate": 0.000562246740220662, + "loss": 1.3513, + "theoretical_loss": 3.676247220925698, + "tokens_seen": 1004077056 + }, + { + "epoch": 0.3, + "learning_rate": 0.0005622306920762287, + "loss": 1.3607, + "theoretical_loss": 3.67622458059801, + "tokens_seen": 1004142592 + }, + { + "epoch": 0.3, + "learning_rate": 0.0005622146439317954, + "loss": 1.5001, + "theoretical_loss": 3.6762019421616166, + "tokens_seen": 1004208128 + }, + { + "epoch": 0.3, + "learning_rate": 0.0005621985957873621, + "loss": 1.5359, + "theoretical_loss": 3.6761793056162357, + "tokens_seen": 1004273664 + }, + { + "epoch": 0.3, + "learning_rate": 0.0005621825476429288, + "loss": 1.3084, + "theoretical_loss": 3.676156670961586, + "tokens_seen": 1004339200 + }, + { + "epoch": 0.3, + "learning_rate": 0.0005621664994984955, + "loss": 1.4927, + "theoretical_loss": 3.676134038197387, + "tokens_seen": 1004404736 + }, + { + "epoch": 0.3, + "learning_rate": 0.0005621504513540623, + "loss": 1.4495, + "theoretical_loss": 3.6761114073233565, + "tokens_seen": 1004470272 + }, + { + "epoch": 0.3, + "learning_rate": 0.0005621344032096289, + "loss": 1.5639, + "theoretical_loss": 3.6760887783392135, + "tokens_seen": 1004535808 + }, + { + "epoch": 0.3, + "learning_rate": 0.0005621183550651957, + "loss": 1.4354, + "theoretical_loss": 3.6760661512446777, + "tokens_seen": 1004601344 + }, + { + "epoch": 0.3, + "learning_rate": 0.0005621023069207623, + "loss": 1.3676, + "theoretical_loss": 3.676043526039467, + "tokens_seen": 1004666880 + }, + { + "epoch": 0.3, + "learning_rate": 0.0005620862587763291, + "loss": 1.3622, + "theoretical_loss": 3.6760209027233013, + "tokens_seen": 1004732416 + }, + { + "epoch": 0.3, + "learning_rate": 0.0005620702106318957, + "loss": 1.3563, + "theoretical_loss": 3.6759982812958993, + "tokens_seen": 1004797952 + }, + { + "epoch": 0.3, + "learning_rate": 0.0005620541624874625, + "loss": 1.3855, + "theoretical_loss": 3.67597566175698, + "tokens_seen": 1004863488 + }, + { + "epoch": 0.3, + "learning_rate": 0.0005620381143430291, + "loss": 1.5797, + "theoretical_loss": 3.6759530441062624, + "tokens_seen": 1004929024 + }, + { + "epoch": 0.3, + "learning_rate": 0.0005620220661985959, + "loss": 1.3871, + "theoretical_loss": 3.675930428343466, + "tokens_seen": 1004994560 + }, + { + "epoch": 0.3, + "learning_rate": 0.0005620060180541625, + "loss": 1.4373, + "theoretical_loss": 3.6759078144683106, + "tokens_seen": 1005060096 + }, + { + "epoch": 0.3, + "learning_rate": 0.0005619899699097293, + "loss": 1.4219, + "theoretical_loss": 3.675885202480515, + "tokens_seen": 1005125632 + }, + { + "epoch": 0.3, + "learning_rate": 0.0005619739217652959, + "loss": 1.447, + "theoretical_loss": 3.675862592379799, + "tokens_seen": 1005191168 + }, + { + "epoch": 0.3, + "learning_rate": 0.0005619578736208627, + "loss": 1.4482, + "theoretical_loss": 3.6758399841658815, + "tokens_seen": 1005256704 + }, + { + "epoch": 0.3, + "learning_rate": 0.0005619418254764293, + "loss": 1.4663, + "theoretical_loss": 3.6758173778384826, + "tokens_seen": 1005322240 + }, + { + "epoch": 0.3, + "learning_rate": 0.0005619257773319961, + "loss": 1.5507, + "theoretical_loss": 3.675794773397322, + "tokens_seen": 1005387776 + }, + { + "epoch": 0.3, + "learning_rate": 0.0005619097291875627, + "loss": 1.4495, + "theoretical_loss": 3.675772170842119, + "tokens_seen": 1005453312 + }, + { + "epoch": 0.3, + "learning_rate": 0.0005618936810431295, + "loss": 1.4334, + "theoretical_loss": 3.6757495701725933, + "tokens_seen": 1005518848 + }, + { + "epoch": 0.3, + "learning_rate": 0.0005618776328986961, + "loss": 1.4474, + "theoretical_loss": 3.675726971388465, + "tokens_seen": 1005584384 + }, + { + "epoch": 0.3, + "learning_rate": 0.0005618615847542629, + "loss": 1.597, + "theoretical_loss": 3.6757043744894538, + "tokens_seen": 1005649920 + }, + { + "epoch": 0.3, + "learning_rate": 0.0005618455366098295, + "loss": 1.3994, + "theoretical_loss": 3.67568177947528, + "tokens_seen": 1005715456 + }, + { + "epoch": 0.3, + "learning_rate": 0.0005618294884653963, + "loss": 1.5157, + "theoretical_loss": 3.675659186345663, + "tokens_seen": 1005780992 + }, + { + "epoch": 0.3, + "learning_rate": 0.0005618134403209629, + "loss": 1.3712, + "theoretical_loss": 3.6756365951003227, + "tokens_seen": 1005846528 + }, + { + "epoch": 0.3, + "learning_rate": 0.0005617973921765297, + "loss": 1.4569, + "theoretical_loss": 3.6756140057389795, + "tokens_seen": 1005912064 + }, + { + "epoch": 0.3, + "objective/train/docs_used": 369909, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.5814447402954102, + "objective/train/theoretical_loss": 3.6755914182613543, + "objective/train/tokens_used": 1026437600, + "theoretical_loss": 3.6755914182613543, + "tokens_seen": 1005977600 + }, + { + "epoch": 0.3, + "learning_rate": 0.0005617813440320963, + "loss": 1.5671, + "theoretical_loss": 3.6755914182613543, + "tokens_seen": 1005977600 + }, + { + "epoch": 0.3, + "learning_rate": 0.0005617652958876631, + "loss": 1.4918, + "theoretical_loss": 3.675568832667166, + "tokens_seen": 1006043136 + }, + { + "epoch": 0.3, + "learning_rate": 0.0005617492477432297, + "loss": 1.4777, + "theoretical_loss": 3.6755462489561355, + "tokens_seen": 1006108672 + }, + { + "epoch": 0.3, + "learning_rate": 0.0005617331995987964, + "loss": 1.4624, + "theoretical_loss": 3.6755236671279836, + "tokens_seen": 1006174208 + }, + { + "epoch": 0.3, + "learning_rate": 0.0005617171514543631, + "loss": 1.3924, + "theoretical_loss": 3.67550108718243, + "tokens_seen": 1006239744 + }, + { + "epoch": 0.3, + "learning_rate": 0.0005617011033099298, + "loss": 1.4139, + "theoretical_loss": 3.6754785091191957, + "tokens_seen": 1006305280 + }, + { + "epoch": 0.3, + "learning_rate": 0.0005616850551654965, + "loss": 1.4958, + "theoretical_loss": 3.675455932938, + "tokens_seen": 1006370816 + }, + { + "epoch": 0.3, + "learning_rate": 0.0005616690070210632, + "loss": 1.4206, + "theoretical_loss": 3.675433358638565, + "tokens_seen": 1006436352 + }, + { + "epoch": 0.31, + "learning_rate": 0.00056165295887663, + "loss": 1.4082, + "theoretical_loss": 3.6754107862206107, + "tokens_seen": 1006501888 + }, + { + "epoch": 0.31, + "learning_rate": 0.0005616369107321966, + "loss": 1.3774, + "theoretical_loss": 3.6753882156838578, + "tokens_seen": 1006567424 + }, + { + "epoch": 0.31, + "learning_rate": 0.0005616208625877633, + "loss": 1.491, + "theoretical_loss": 3.675365647028027, + "tokens_seen": 1006632960 + }, + { + "epoch": 0.31, + "learning_rate": 0.00056160481444333, + "loss": 1.3449, + "theoretical_loss": 3.675343080252839, + "tokens_seen": 1006698496 + }, + { + "epoch": 0.31, + "learning_rate": 0.0005615887662988966, + "loss": 1.499, + "theoretical_loss": 3.675320515358015, + "tokens_seen": 1006764032 + }, + { + "epoch": 0.31, + "learning_rate": 0.0005615727181544634, + "loss": 1.5206, + "theoretical_loss": 3.675297952343276, + "tokens_seen": 1006829568 + }, + { + "epoch": 0.31, + "learning_rate": 0.00056155667001003, + "loss": 1.4418, + "theoretical_loss": 3.6752753912083422, + "tokens_seen": 1006895104 + }, + { + "epoch": 0.31, + "learning_rate": 0.0005615406218655968, + "loss": 1.4164, + "theoretical_loss": 3.6752528319529354, + "tokens_seen": 1006960640 + }, + { + "epoch": 0.31, + "learning_rate": 0.0005615245737211634, + "loss": 1.495, + "theoretical_loss": 3.675230274576777, + "tokens_seen": 1007026176 + }, + { + "epoch": 0.31, + "learning_rate": 0.0005615085255767302, + "loss": 1.4173, + "theoretical_loss": 3.6752077190795873, + "tokens_seen": 1007091712 + }, + { + "epoch": 0.31, + "learning_rate": 0.0005614924774322968, + "loss": 1.4571, + "theoretical_loss": 3.6751851654610874, + "tokens_seen": 1007157248 + }, + { + "epoch": 0.31, + "learning_rate": 0.0005614764292878636, + "loss": 1.3958, + "theoretical_loss": 3.6751626137209996, + "tokens_seen": 1007222784 + }, + { + "epoch": 0.31, + "learning_rate": 0.0005614603811434302, + "loss": 1.5389, + "theoretical_loss": 3.675140063859045, + "tokens_seen": 1007288320 + }, + { + "epoch": 0.31, + "learning_rate": 0.000561444332998997, + "loss": 1.5157, + "theoretical_loss": 3.6751175158749447, + "tokens_seen": 1007353856 + }, + { + "epoch": 0.31, + "learning_rate": 0.0005614282848545636, + "loss": 1.4435, + "theoretical_loss": 3.67509496976842, + "tokens_seen": 1007419392 + }, + { + "epoch": 0.31, + "learning_rate": 0.0005614122367101304, + "loss": 1.3988, + "theoretical_loss": 3.6750724255391924, + "tokens_seen": 1007484928 + }, + { + "epoch": 0.31, + "learning_rate": 0.000561396188565697, + "loss": 1.6103, + "theoretical_loss": 3.6750498831869844, + "tokens_seen": 1007550464 + }, + { + "epoch": 0.31, + "learning_rate": 0.0005613801404212638, + "loss": 1.4985, + "theoretical_loss": 3.6750273427115165, + "tokens_seen": 1007616000 + }, + { + "epoch": 0.31, + "learning_rate": 0.0005613640922768305, + "loss": 1.4913, + "theoretical_loss": 3.675004804112511, + "tokens_seen": 1007681536 + }, + { + "epoch": 0.31, + "learning_rate": 0.0005613480441323972, + "loss": 1.5304, + "theoretical_loss": 3.67498226738969, + "tokens_seen": 1007747072 + }, + { + "epoch": 0.31, + "learning_rate": 0.0005613319959879639, + "loss": 1.4529, + "theoretical_loss": 3.674959732542774, + "tokens_seen": 1007812608 + }, + { + "epoch": 0.31, + "learning_rate": 0.0005613159478435306, + "loss": 1.4555, + "theoretical_loss": 3.6749371995714863, + "tokens_seen": 1007878144 + }, + { + "epoch": 0.31, + "learning_rate": 0.0005612998996990973, + "loss": 1.4216, + "theoretical_loss": 3.674914668475548, + "tokens_seen": 1007943680 + }, + { + "epoch": 0.31, + "learning_rate": 0.000561283851554664, + "loss": 1.3942, + "theoretical_loss": 3.674892139254682, + "tokens_seen": 1008009216 + }, + { + "epoch": 0.31, + "learning_rate": 0.0005612678034102307, + "loss": 1.57, + "theoretical_loss": 3.6748696119086093, + "tokens_seen": 1008074752 + }, + { + "epoch": 0.31, + "learning_rate": 0.0005612517552657974, + "loss": 1.5336, + "theoretical_loss": 3.674847086437052, + "tokens_seen": 1008140288 + }, + { + "epoch": 0.31, + "learning_rate": 0.0005612357071213641, + "loss": 1.3894, + "theoretical_loss": 3.6748245628397336, + "tokens_seen": 1008205824 + }, + { + "epoch": 0.31, + "learning_rate": 0.0005612196589769308, + "loss": 1.4884, + "theoretical_loss": 3.6748020411163753, + "tokens_seen": 1008271360 + }, + { + "epoch": 0.31, + "learning_rate": 0.0005612036108324975, + "loss": 1.5648, + "theoretical_loss": 3.674779521266699, + "tokens_seen": 1008336896 + }, + { + "epoch": 0.31, + "learning_rate": 0.0005611875626880642, + "loss": 1.51, + "theoretical_loss": 3.674757003290429, + "tokens_seen": 1008402432 + }, + { + "epoch": 0.31, + "learning_rate": 0.0005611715145436309, + "loss": 1.5082, + "theoretical_loss": 3.674734487187285, + "tokens_seen": 1008467968 + }, + { + "epoch": 0.31, + "learning_rate": 0.0005611554663991976, + "loss": 1.3886, + "theoretical_loss": 3.6747119729569917, + "tokens_seen": 1008533504 + }, + { + "epoch": 0.31, + "learning_rate": 0.0005611394182547643, + "loss": 1.4161, + "theoretical_loss": 3.6746894605992706, + "tokens_seen": 1008599040 + }, + { + "epoch": 0.31, + "learning_rate": 0.000561123370110331, + "loss": 1.427, + "theoretical_loss": 3.674666950113844, + "tokens_seen": 1008664576 + }, + { + "epoch": 0.31, + "learning_rate": 0.0005611073219658977, + "loss": 1.525, + "theoretical_loss": 3.674644441500435, + "tokens_seen": 1008730112 + }, + { + "epoch": 0.31, + "learning_rate": 0.0005610912738214644, + "loss": 1.408, + "theoretical_loss": 3.674621934758767, + "tokens_seen": 1008795648 + }, + { + "epoch": 0.31, + "learning_rate": 0.0005610752256770311, + "loss": 1.5659, + "theoretical_loss": 3.674599429888562, + "tokens_seen": 1008861184 + }, + { + "epoch": 0.31, + "learning_rate": 0.0005610591775325978, + "loss": 1.4278, + "theoretical_loss": 3.6745769268895425, + "tokens_seen": 1008926720 + }, + { + "epoch": 0.31, + "learning_rate": 0.0005610431293881645, + "loss": 1.5203, + "theoretical_loss": 3.674554425761432, + "tokens_seen": 1008992256 + }, + { + "epoch": 0.31, + "learning_rate": 0.0005610270812437312, + "loss": 1.3744, + "theoretical_loss": 3.6745319265039535, + "tokens_seen": 1009057792 + }, + { + "epoch": 0.31, + "learning_rate": 0.0005610110330992979, + "loss": 1.3615, + "theoretical_loss": 3.6745094291168297, + "tokens_seen": 1009123328 + }, + { + "epoch": 0.31, + "learning_rate": 0.0005609949849548646, + "loss": 1.4629, + "theoretical_loss": 3.6744869335997836, + "tokens_seen": 1009188864 + }, + { + "epoch": 0.31, + "objective/train/docs_used": 370943, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.5502262115478516, + "objective/train/theoretical_loss": 3.6744644399525384, + "objective/train/tokens_used": 1029714400, + "theoretical_loss": 3.6744644399525384, + "tokens_seen": 1009254400 + }, + { + "epoch": 0.31, + "learning_rate": 0.0005609789368104313, + "loss": 1.4676, + "theoretical_loss": 3.6744644399525384, + "tokens_seen": 1009254400 + }, + { + "epoch": 0.31, + "learning_rate": 0.000560962888665998, + "loss": 1.5088, + "theoretical_loss": 3.6744419481748176, + "tokens_seen": 1009319936 + }, + { + "epoch": 0.31, + "learning_rate": 0.0005609468405215648, + "loss": 1.4738, + "theoretical_loss": 3.674419458266344, + "tokens_seen": 1009385472 + }, + { + "epoch": 0.31, + "learning_rate": 0.0005609307923771314, + "loss": 1.4211, + "theoretical_loss": 3.674396970226841, + "tokens_seen": 1009451008 + }, + { + "epoch": 0.31, + "learning_rate": 0.0005609147442326982, + "loss": 1.384, + "theoretical_loss": 3.674374484056032, + "tokens_seen": 1009516544 + }, + { + "epoch": 0.31, + "learning_rate": 0.0005608986960882648, + "loss": 1.4253, + "theoretical_loss": 3.6743519997536414, + "tokens_seen": 1009582080 + }, + { + "epoch": 0.31, + "learning_rate": 0.0005608826479438316, + "loss": 1.5127, + "theoretical_loss": 3.674329517319391, + "tokens_seen": 1009647616 + }, + { + "epoch": 0.31, + "learning_rate": 0.0005608665997993982, + "loss": 1.3449, + "theoretical_loss": 3.674307036753005, + "tokens_seen": 1009713152 + }, + { + "epoch": 0.31, + "learning_rate": 0.000560850551654965, + "loss": 1.4401, + "theoretical_loss": 3.6742845580542074, + "tokens_seen": 1009778688 + }, + { + "epoch": 0.31, + "learning_rate": 0.0005608345035105316, + "loss": 1.455, + "theoretical_loss": 3.6742620812227216, + "tokens_seen": 1009844224 + }, + { + "epoch": 0.31, + "learning_rate": 0.0005608184553660984, + "loss": 1.4055, + "theoretical_loss": 3.6742396062582703, + "tokens_seen": 1009909760 + }, + { + "epoch": 0.31, + "learning_rate": 0.000560802407221665, + "loss": 1.4513, + "theoretical_loss": 3.674217133160579, + "tokens_seen": 1009975296 + }, + { + "epoch": 0.31, + "learning_rate": 0.0005607863590772318, + "loss": 1.477, + "theoretical_loss": 3.6741946619293713, + "tokens_seen": 1010040832 + }, + { + "epoch": 0.31, + "learning_rate": 0.0005607703109327984, + "loss": 1.4871, + "theoretical_loss": 3.6741721925643698, + "tokens_seen": 1010106368 + }, + { + "epoch": 0.31, + "learning_rate": 0.0005607542627883652, + "loss": 1.3975, + "theoretical_loss": 3.6741497250652997, + "tokens_seen": 1010171904 + }, + { + "epoch": 0.31, + "learning_rate": 0.0005607382146439318, + "loss": 1.4379, + "theoretical_loss": 3.6741272594318843, + "tokens_seen": 1010237440 + }, + { + "epoch": 0.31, + "learning_rate": 0.0005607221664994986, + "loss": 1.4163, + "theoretical_loss": 3.6741047956638475, + "tokens_seen": 1010302976 + }, + { + "epoch": 0.31, + "learning_rate": 0.0005607061183550652, + "loss": 1.4812, + "theoretical_loss": 3.6740823337609143, + "tokens_seen": 1010368512 + }, + { + "epoch": 0.31, + "learning_rate": 0.000560690070210632, + "loss": 1.446, + "theoretical_loss": 3.6740598737228085, + "tokens_seen": 1010434048 + }, + { + "epoch": 0.31, + "learning_rate": 0.0005606740220661986, + "loss": 1.554, + "theoretical_loss": 3.674037415549254, + "tokens_seen": 1010499584 + }, + { + "epoch": 0.31, + "learning_rate": 0.0005606579739217654, + "loss": 1.4494, + "theoretical_loss": 3.674014959239975, + "tokens_seen": 1010565120 + }, + { + "epoch": 0.31, + "learning_rate": 0.000560641925777332, + "loss": 1.5057, + "theoretical_loss": 3.6739925047946964, + "tokens_seen": 1010630656 + }, + { + "epoch": 0.31, + "learning_rate": 0.0005606258776328988, + "loss": 1.4686, + "theoretical_loss": 3.6739700522131424, + "tokens_seen": 1010696192 + }, + { + "epoch": 0.31, + "learning_rate": 0.0005606098294884654, + "loss": 1.4087, + "theoretical_loss": 3.673947601495038, + "tokens_seen": 1010761728 + }, + { + "epoch": 0.31, + "learning_rate": 0.0005605937813440322, + "loss": 1.4312, + "theoretical_loss": 3.6739251526401064, + "tokens_seen": 1010827264 + }, + { + "epoch": 0.31, + "learning_rate": 0.0005605777331995988, + "loss": 1.3723, + "theoretical_loss": 3.6739027056480733, + "tokens_seen": 1010892800 + }, + { + "epoch": 0.31, + "learning_rate": 0.0005605616850551656, + "loss": 1.4133, + "theoretical_loss": 3.673880260518663, + "tokens_seen": 1010958336 + }, + { + "epoch": 0.31, + "learning_rate": 0.0005605456369107322, + "loss": 1.4072, + "theoretical_loss": 3.6738578172516005, + "tokens_seen": 1011023872 + }, + { + "epoch": 0.31, + "learning_rate": 0.000560529588766299, + "loss": 1.4483, + "theoretical_loss": 3.6738353758466102, + "tokens_seen": 1011089408 + }, + { + "epoch": 0.31, + "learning_rate": 0.0005605135406218656, + "loss": 1.5811, + "theoretical_loss": 3.6738129363034178, + "tokens_seen": 1011154944 + }, + { + "epoch": 0.31, + "learning_rate": 0.0005604974924774324, + "loss": 1.4935, + "theoretical_loss": 3.6737904986217464, + "tokens_seen": 1011220480 + }, + { + "epoch": 0.31, + "learning_rate": 0.000560481444332999, + "loss": 1.547, + "theoretical_loss": 3.6737680628013223, + "tokens_seen": 1011286016 + }, + { + "epoch": 0.31, + "learning_rate": 0.0005604653961885658, + "loss": 1.5197, + "theoretical_loss": 3.6737456288418704, + "tokens_seen": 1011351552 + }, + { + "epoch": 0.31, + "learning_rate": 0.0005604493480441324, + "loss": 1.5009, + "theoretical_loss": 3.673723196743116, + "tokens_seen": 1011417088 + }, + { + "epoch": 0.31, + "learning_rate": 0.0005604332998996992, + "loss": 1.4118, + "theoretical_loss": 3.673700766504783, + "tokens_seen": 1011482624 + }, + { + "epoch": 0.31, + "learning_rate": 0.0005604172517552658, + "loss": 1.5108, + "theoretical_loss": 3.6736783381265976, + "tokens_seen": 1011548160 + }, + { + "epoch": 0.31, + "learning_rate": 0.0005604012036108326, + "loss": 1.3868, + "theoretical_loss": 3.673655911608285, + "tokens_seen": 1011613696 + }, + { + "epoch": 0.31, + "learning_rate": 0.0005603851554663992, + "loss": 1.5346, + "theoretical_loss": 3.67363348694957, + "tokens_seen": 1011679232 + }, + { + "epoch": 0.31, + "learning_rate": 0.000560369107321966, + "loss": 1.5085, + "theoretical_loss": 3.6736110641501787, + "tokens_seen": 1011744768 + }, + { + "epoch": 0.31, + "learning_rate": 0.0005603530591775326, + "loss": 1.5105, + "theoretical_loss": 3.673588643209836, + "tokens_seen": 1011810304 + }, + { + "epoch": 0.31, + "learning_rate": 0.0005603370110330994, + "loss": 1.5307, + "theoretical_loss": 3.673566224128267, + "tokens_seen": 1011875840 + }, + { + "epoch": 0.31, + "learning_rate": 0.000560320962888666, + "loss": 1.4367, + "theoretical_loss": 3.673543806905198, + "tokens_seen": 1011941376 + }, + { + "epoch": 0.31, + "learning_rate": 0.0005603049147442327, + "loss": 1.4241, + "theoretical_loss": 3.6735213915403544, + "tokens_seen": 1012006912 + }, + { + "epoch": 0.31, + "learning_rate": 0.0005602888665997994, + "loss": 1.4858, + "theoretical_loss": 3.6734989780334617, + "tokens_seen": 1012072448 + }, + { + "epoch": 0.31, + "learning_rate": 0.0005602728184553661, + "loss": 1.4459, + "theoretical_loss": 3.6734765663842452, + "tokens_seen": 1012137984 + }, + { + "epoch": 0.31, + "learning_rate": 0.0005602567703109328, + "loss": 1.4165, + "theoretical_loss": 3.6734541565924315, + "tokens_seen": 1012203520 + }, + { + "epoch": 0.31, + "learning_rate": 0.0005602407221664995, + "loss": 1.4478, + "theoretical_loss": 3.673431748657746, + "tokens_seen": 1012269056 + }, + { + "epoch": 0.31, + "learning_rate": 0.0005602246740220663, + "loss": 1.3836, + "theoretical_loss": 3.673409342579914, + "tokens_seen": 1012334592 + }, + { + "epoch": 0.31, + "learning_rate": 0.0005602086258776329, + "loss": 1.63, + "theoretical_loss": 3.6733869383586626, + "tokens_seen": 1012400128 + }, + { + "epoch": 0.31, + "learning_rate": 0.0005601925777331995, + "loss": 1.4656, + "theoretical_loss": 3.6733645359937173, + "tokens_seen": 1012465664 + }, + { + "epoch": 0.31, + "objective/train/docs_used": 372166, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.473021388053894, + "objective/train/theoretical_loss": 3.6733421354848037, + "objective/train/tokens_used": 1032991200, + "theoretical_loss": 3.6733421354848037, + "tokens_seen": 1012531200 + }, + { + "epoch": 0.31, + "learning_rate": 0.0005601765295887663, + "loss": 1.5082, + "theoretical_loss": 3.6733421354848037, + "tokens_seen": 1012531200 + }, + { + "epoch": 0.31, + "learning_rate": 0.0005601604814443329, + "loss": 1.46, + "theoretical_loss": 3.6733197368316484, + "tokens_seen": 1012596736 + }, + { + "epoch": 0.31, + "learning_rate": 0.0005601444332998997, + "loss": 1.4655, + "theoretical_loss": 3.6732973400339772, + "tokens_seen": 1012662272 + }, + { + "epoch": 0.31, + "learning_rate": 0.0005601283851554663, + "loss": 1.466, + "theoretical_loss": 3.6732749450915168, + "tokens_seen": 1012727808 + }, + { + "epoch": 0.31, + "learning_rate": 0.0005601123370110331, + "loss": 1.5492, + "theoretical_loss": 3.6732525520039934, + "tokens_seen": 1012793344 + }, + { + "epoch": 0.31, + "learning_rate": 0.0005600962888665997, + "loss": 1.444, + "theoretical_loss": 3.6732301607711326, + "tokens_seen": 1012858880 + }, + { + "epoch": 0.31, + "learning_rate": 0.0005600802407221665, + "loss": 1.4014, + "theoretical_loss": 3.6732077713926623, + "tokens_seen": 1012924416 + }, + { + "epoch": 0.31, + "learning_rate": 0.0005600641925777331, + "loss": 1.458, + "theoretical_loss": 3.673185383868307, + "tokens_seen": 1012989952 + }, + { + "epoch": 0.31, + "learning_rate": 0.0005600481444332999, + "loss": 1.4724, + "theoretical_loss": 3.673162998197795, + "tokens_seen": 1013055488 + }, + { + "epoch": 0.31, + "learning_rate": 0.0005600320962888665, + "loss": 1.4334, + "theoretical_loss": 3.673140614380852, + "tokens_seen": 1013121024 + }, + { + "epoch": 0.31, + "learning_rate": 0.0005600160481444333, + "loss": 1.5111, + "theoretical_loss": 3.673118232417205, + "tokens_seen": 1013186560 + }, + { + "epoch": 0.31, + "learning_rate": 0.00056, + "loss": 1.544, + "theoretical_loss": 3.6730958523065804, + "tokens_seen": 1013252096 + }, + { + "epoch": 0.31, + "learning_rate": 0.0005599839518555667, + "loss": 1.4129, + "theoretical_loss": 3.6730734740487048, + "tokens_seen": 1013317632 + }, + { + "epoch": 0.31, + "learning_rate": 0.0005599679037111334, + "loss": 1.5543, + "theoretical_loss": 3.6730510976433055, + "tokens_seen": 1013383168 + }, + { + "epoch": 0.31, + "learning_rate": 0.0005599518555667001, + "loss": 1.5167, + "theoretical_loss": 3.673028723090109, + "tokens_seen": 1013448704 + }, + { + "epoch": 0.31, + "learning_rate": 0.0005599358074222668, + "loss": 1.4286, + "theoretical_loss": 3.6730063503888424, + "tokens_seen": 1013514240 + }, + { + "epoch": 0.31, + "learning_rate": 0.0005599197592778335, + "loss": 1.4941, + "theoretical_loss": 3.6729839795392323, + "tokens_seen": 1013579776 + }, + { + "epoch": 0.31, + "learning_rate": 0.0005599037111334002, + "loss": 1.3525, + "theoretical_loss": 3.672961610541006, + "tokens_seen": 1013645312 + }, + { + "epoch": 0.31, + "learning_rate": 0.0005598876629889669, + "loss": 1.4657, + "theoretical_loss": 3.672939243393891, + "tokens_seen": 1013710848 + }, + { + "epoch": 0.31, + "learning_rate": 0.0005598716148445336, + "loss": 1.3884, + "theoretical_loss": 3.672916878097614, + "tokens_seen": 1013776384 + }, + { + "epoch": 0.31, + "learning_rate": 0.0005598555667001003, + "loss": 1.382, + "theoretical_loss": 3.6728945146519023, + "tokens_seen": 1013841920 + }, + { + "epoch": 0.31, + "learning_rate": 0.000559839518555667, + "loss": 1.4121, + "theoretical_loss": 3.6728721530564834, + "tokens_seen": 1013907456 + }, + { + "epoch": 0.31, + "learning_rate": 0.0005598234704112337, + "loss": 1.523, + "theoretical_loss": 3.672849793311084, + "tokens_seen": 1013972992 + }, + { + "epoch": 0.31, + "learning_rate": 0.0005598074222668004, + "loss": 1.4534, + "theoretical_loss": 3.6728274354154324, + "tokens_seen": 1014038528 + }, + { + "epoch": 0.31, + "learning_rate": 0.0005597913741223671, + "loss": 1.4473, + "theoretical_loss": 3.672805079369255, + "tokens_seen": 1014104064 + }, + { + "epoch": 0.31, + "learning_rate": 0.0005597753259779338, + "loss": 1.482, + "theoretical_loss": 3.67278272517228, + "tokens_seen": 1014169600 + }, + { + "epoch": 0.31, + "learning_rate": 0.0005597592778335005, + "loss": 1.3996, + "theoretical_loss": 3.6727603728242353, + "tokens_seen": 1014235136 + }, + { + "epoch": 0.31, + "learning_rate": 0.0005597432296890673, + "loss": 1.3897, + "theoretical_loss": 3.6727380223248476, + "tokens_seen": 1014300672 + }, + { + "epoch": 0.31, + "learning_rate": 0.0005597271815446339, + "loss": 1.5186, + "theoretical_loss": 3.672715673673845, + "tokens_seen": 1014366208 + }, + { + "epoch": 0.31, + "learning_rate": 0.0005597111334002007, + "loss": 1.534, + "theoretical_loss": 3.672693326870955, + "tokens_seen": 1014431744 + }, + { + "epoch": 0.31, + "learning_rate": 0.0005596950852557673, + "loss": 1.3995, + "theoretical_loss": 3.672670981915906, + "tokens_seen": 1014497280 + }, + { + "epoch": 0.31, + "learning_rate": 0.0005596790371113341, + "loss": 1.4467, + "theoretical_loss": 3.6726486388084254, + "tokens_seen": 1014562816 + }, + { + "epoch": 0.31, + "learning_rate": 0.0005596629889669007, + "loss": 1.3628, + "theoretical_loss": 3.6726262975482413, + "tokens_seen": 1014628352 + }, + { + "epoch": 0.31, + "learning_rate": 0.0005596469408224675, + "loss": 1.3954, + "theoretical_loss": 3.6726039581350816, + "tokens_seen": 1014693888 + }, + { + "epoch": 0.31, + "learning_rate": 0.0005596308926780341, + "loss": 1.5545, + "theoretical_loss": 3.672581620568674, + "tokens_seen": 1014759424 + }, + { + "epoch": 0.31, + "learning_rate": 0.0005596148445336009, + "loss": 1.3842, + "theoretical_loss": 3.672559284848747, + "tokens_seen": 1014824960 + }, + { + "epoch": 0.31, + "learning_rate": 0.0005595987963891675, + "loss": 1.5572, + "theoretical_loss": 3.6725369509750285, + "tokens_seen": 1014890496 + }, + { + "epoch": 0.31, + "learning_rate": 0.0005595827482447343, + "loss": 1.3799, + "theoretical_loss": 3.672514618947247, + "tokens_seen": 1014956032 + }, + { + "epoch": 0.31, + "learning_rate": 0.0005595667001003009, + "loss": 1.4585, + "theoretical_loss": 3.67249228876513, + "tokens_seen": 1015021568 + }, + { + "epoch": 0.31, + "learning_rate": 0.0005595506519558677, + "loss": 1.4233, + "theoretical_loss": 3.672469960428407, + "tokens_seen": 1015087104 + }, + { + "epoch": 0.31, + "learning_rate": 0.0005595346038114343, + "loss": 1.4855, + "theoretical_loss": 3.6724476339368053, + "tokens_seen": 1015152640 + }, + { + "epoch": 0.31, + "learning_rate": 0.0005595185556670011, + "loss": 1.4354, + "theoretical_loss": 3.672425309290054, + "tokens_seen": 1015218176 + }, + { + "epoch": 0.31, + "learning_rate": 0.0005595025075225677, + "loss": 1.467, + "theoretical_loss": 3.6724029864878807, + "tokens_seen": 1015283712 + }, + { + "epoch": 0.31, + "learning_rate": 0.0005594864593781345, + "loss": 1.4411, + "theoretical_loss": 3.6723806655300155, + "tokens_seen": 1015349248 + }, + { + "epoch": 0.31, + "learning_rate": 0.0005594704112337011, + "loss": 1.4591, + "theoretical_loss": 3.6723583464161855, + "tokens_seen": 1015414784 + }, + { + "epoch": 0.31, + "learning_rate": 0.0005594543630892679, + "loss": 1.4944, + "theoretical_loss": 3.67233602914612, + "tokens_seen": 1015480320 + }, + { + "epoch": 0.31, + "learning_rate": 0.0005594383149448345, + "loss": 1.5065, + "theoretical_loss": 3.6723137137195474, + "tokens_seen": 1015545856 + }, + { + "epoch": 0.31, + "learning_rate": 0.0005594222668004013, + "loss": 1.4144, + "theoretical_loss": 3.672291400136197, + "tokens_seen": 1015611392 + }, + { + "epoch": 0.31, + "learning_rate": 0.0005594062186559679, + "loss": 1.4495, + "theoretical_loss": 3.6722690883957974, + "tokens_seen": 1015676928 + }, + { + "epoch": 0.31, + "learning_rate": 0.0005593901705115347, + "loss": 1.5598, + "theoretical_loss": 3.672246778498077, + "tokens_seen": 1015742464 + }, + { + "epoch": 0.31, + "objective/train/docs_used": 372984, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.4150993824005127, + "objective/train/theoretical_loss": 3.6722244704427656, + "objective/train/tokens_used": 1036268000, + "theoretical_loss": 3.6722244704427656, + "tokens_seen": 1015808000 + }, + { + "epoch": 0.31, + "learning_rate": 0.0005593741223671013, + "loss": 1.4409, + "theoretical_loss": 3.6722244704427656, + "tokens_seen": 1015808000 + }, + { + "epoch": 0.31, + "learning_rate": 0.0005593580742226681, + "loss": 1.3899, + "theoretical_loss": 3.6722021642295912, + "tokens_seen": 1015873536 + }, + { + "epoch": 0.31, + "learning_rate": 0.0005593420260782347, + "loss": 1.4307, + "theoretical_loss": 3.6721798598582835, + "tokens_seen": 1015939072 + }, + { + "epoch": 0.31, + "learning_rate": 0.0005593259779338015, + "loss": 1.4854, + "theoretical_loss": 3.6721575573285716, + "tokens_seen": 1016004608 + }, + { + "epoch": 0.31, + "learning_rate": 0.0005593099297893681, + "loss": 1.3283, + "theoretical_loss": 3.6721352566401846, + "tokens_seen": 1016070144 + }, + { + "epoch": 0.31, + "learning_rate": 0.0005592938816449349, + "loss": 1.4686, + "theoretical_loss": 3.6721129577928524, + "tokens_seen": 1016135680 + }, + { + "epoch": 0.31, + "learning_rate": 0.0005592778335005015, + "loss": 1.5465, + "theoretical_loss": 3.672090660786303, + "tokens_seen": 1016201216 + }, + { + "epoch": 0.31, + "learning_rate": 0.0005592617853560683, + "loss": 1.5136, + "theoretical_loss": 3.672068365620266, + "tokens_seen": 1016266752 + }, + { + "epoch": 0.31, + "learning_rate": 0.0005592457372116349, + "loss": 1.4014, + "theoretical_loss": 3.6720460722944717, + "tokens_seen": 1016332288 + }, + { + "epoch": 0.31, + "learning_rate": 0.0005592296890672017, + "loss": 1.5111, + "theoretical_loss": 3.672023780808648, + "tokens_seen": 1016397824 + }, + { + "epoch": 0.31, + "learning_rate": 0.0005592136409227683, + "loss": 1.4457, + "theoretical_loss": 3.6720014911625265, + "tokens_seen": 1016463360 + }, + { + "epoch": 0.31, + "learning_rate": 0.0005591975927783351, + "loss": 1.3982, + "theoretical_loss": 3.671979203355835, + "tokens_seen": 1016528896 + }, + { + "epoch": 0.31, + "learning_rate": 0.0005591815446339017, + "loss": 1.4157, + "theoretical_loss": 3.671956917388304, + "tokens_seen": 1016594432 + }, + { + "epoch": 0.31, + "learning_rate": 0.0005591654964894685, + "loss": 1.4798, + "theoretical_loss": 3.6719346332596636, + "tokens_seen": 1016659968 + }, + { + "epoch": 0.31, + "learning_rate": 0.0005591494483450351, + "loss": 1.5072, + "theoretical_loss": 3.6719123509696425, + "tokens_seen": 1016725504 + }, + { + "epoch": 0.31, + "learning_rate": 0.0005591334002006019, + "loss": 1.5534, + "theoretical_loss": 3.6718900705179705, + "tokens_seen": 1016791040 + }, + { + "epoch": 0.31, + "learning_rate": 0.0005591173520561685, + "loss": 1.3767, + "theoretical_loss": 3.6718677919043783, + "tokens_seen": 1016856576 + }, + { + "epoch": 0.31, + "learning_rate": 0.0005591013039117353, + "loss": 1.4883, + "theoretical_loss": 3.671845515128595, + "tokens_seen": 1016922112 + }, + { + "epoch": 0.31, + "learning_rate": 0.0005590852557673019, + "loss": 1.5247, + "theoretical_loss": 3.671823240190352, + "tokens_seen": 1016987648 + }, + { + "epoch": 0.31, + "learning_rate": 0.0005590692076228687, + "loss": 1.3937, + "theoretical_loss": 3.6718009670893776, + "tokens_seen": 1017053184 + }, + { + "epoch": 0.31, + "learning_rate": 0.0005590531594784353, + "loss": 1.4713, + "theoretical_loss": 3.6717786958254024, + "tokens_seen": 1017118720 + }, + { + "epoch": 0.31, + "learning_rate": 0.0005590371113340021, + "loss": 1.4645, + "theoretical_loss": 3.671756426398157, + "tokens_seen": 1017184256 + }, + { + "epoch": 0.31, + "learning_rate": 0.0005590210631895687, + "loss": 1.4335, + "theoretical_loss": 3.6717341588073715, + "tokens_seen": 1017249792 + }, + { + "epoch": 0.31, + "learning_rate": 0.0005590050150451355, + "loss": 1.496, + "theoretical_loss": 3.671711893052776, + "tokens_seen": 1017315328 + }, + { + "epoch": 0.31, + "learning_rate": 0.0005589889669007021, + "loss": 1.4078, + "theoretical_loss": 3.6716896291341, + "tokens_seen": 1017380864 + }, + { + "epoch": 0.31, + "learning_rate": 0.0005589729187562689, + "loss": 1.4068, + "theoretical_loss": 3.671667367051076, + "tokens_seen": 1017446400 + }, + { + "epoch": 0.31, + "learning_rate": 0.0005589568706118355, + "loss": 1.3584, + "theoretical_loss": 3.671645106803432, + "tokens_seen": 1017511936 + }, + { + "epoch": 0.31, + "learning_rate": 0.0005589408224674023, + "loss": 1.478, + "theoretical_loss": 3.6716228483909, + "tokens_seen": 1017577472 + }, + { + "epoch": 0.31, + "learning_rate": 0.000558924774322969, + "loss": 1.4212, + "theoretical_loss": 3.6716005918132097, + "tokens_seen": 1017643008 + }, + { + "epoch": 0.31, + "learning_rate": 0.0005589087261785356, + "loss": 1.5175, + "theoretical_loss": 3.671578337070093, + "tokens_seen": 1017708544 + }, + { + "epoch": 0.31, + "learning_rate": 0.0005588926780341023, + "loss": 1.3837, + "theoretical_loss": 3.6715560841612795, + "tokens_seen": 1017774080 + }, + { + "epoch": 0.31, + "learning_rate": 0.000558876629889669, + "loss": 1.4569, + "theoretical_loss": 3.6715338330864995, + "tokens_seen": 1017839616 + }, + { + "epoch": 0.31, + "learning_rate": 0.0005588605817452358, + "loss": 1.4217, + "theoretical_loss": 3.6715115838454846, + "tokens_seen": 1017905152 + }, + { + "epoch": 0.31, + "learning_rate": 0.0005588445336008024, + "loss": 1.4813, + "theoretical_loss": 3.6714893364379657, + "tokens_seen": 1017970688 + }, + { + "epoch": 0.31, + "learning_rate": 0.0005588284854563692, + "loss": 1.495, + "theoretical_loss": 3.671467090863673, + "tokens_seen": 1018036224 + }, + { + "epoch": 0.31, + "learning_rate": 0.0005588124373119358, + "loss": 1.4849, + "theoretical_loss": 3.6714448471223378, + "tokens_seen": 1018101760 + }, + { + "epoch": 0.31, + "learning_rate": 0.0005587963891675026, + "loss": 1.437, + "theoretical_loss": 3.6714226052136913, + "tokens_seen": 1018167296 + }, + { + "epoch": 0.31, + "learning_rate": 0.0005587803410230692, + "loss": 1.4647, + "theoretical_loss": 3.6714003651374645, + "tokens_seen": 1018232832 + }, + { + "epoch": 0.31, + "learning_rate": 0.0005587642928786358, + "loss": 1.3759, + "theoretical_loss": 3.671378126893388, + "tokens_seen": 1018298368 + }, + { + "epoch": 0.31, + "learning_rate": 0.0005587482447342026, + "loss": 1.4889, + "theoretical_loss": 3.671355890481194, + "tokens_seen": 1018363904 + }, + { + "epoch": 0.31, + "learning_rate": 0.0005587321965897692, + "loss": 1.3893, + "theoretical_loss": 3.6713336559006127, + "tokens_seen": 1018429440 + }, + { + "epoch": 0.31, + "learning_rate": 0.000558716148445336, + "loss": 1.5309, + "theoretical_loss": 3.6713114231513755, + "tokens_seen": 1018494976 + }, + { + "epoch": 0.31, + "learning_rate": 0.0005587001003009026, + "loss": 1.3979, + "theoretical_loss": 3.671289192233214, + "tokens_seen": 1018560512 + }, + { + "epoch": 0.31, + "learning_rate": 0.0005586840521564694, + "loss": 1.4794, + "theoretical_loss": 3.6712669631458597, + "tokens_seen": 1018626048 + }, + { + "epoch": 0.31, + "learning_rate": 0.000558668004012036, + "loss": 1.4462, + "theoretical_loss": 3.671244735889044, + "tokens_seen": 1018691584 + }, + { + "epoch": 0.31, + "learning_rate": 0.0005586519558676028, + "loss": 1.5217, + "theoretical_loss": 3.6712225104624983, + "tokens_seen": 1018757120 + }, + { + "epoch": 0.31, + "learning_rate": 0.0005586359077231695, + "loss": 1.5762, + "theoretical_loss": 3.6712002868659543, + "tokens_seen": 1018822656 + }, + { + "epoch": 0.31, + "learning_rate": 0.0005586198595787362, + "loss": 1.3723, + "theoretical_loss": 3.6711780650991432, + "tokens_seen": 1018888192 + }, + { + "epoch": 0.31, + "learning_rate": 0.0005586038114343029, + "loss": 1.4457, + "theoretical_loss": 3.671155845161797, + "tokens_seen": 1018953728 + }, + { + "epoch": 0.31, + "learning_rate": 0.0005585877632898696, + "loss": 1.5071, + "theoretical_loss": 3.671133627053648, + "tokens_seen": 1019019264 + }, + { + "epoch": 0.31, + "objective/train/docs_used": 374190, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.3972426652908325, + "objective/train/theoretical_loss": 3.6711114107744267, + "objective/train/tokens_used": 1039544800, + "theoretical_loss": 3.6711114107744267, + "tokens_seen": 1019084800 + }, + { + "epoch": 0.31, + "learning_rate": 0.0005585717151454363, + "loss": 1.4619, + "theoretical_loss": 3.6711114107744267, + "tokens_seen": 1019084800 + }, + { + "epoch": 0.31, + "learning_rate": 0.000558555667001003, + "loss": 1.5282, + "theoretical_loss": 3.671089196323866, + "tokens_seen": 1019150336 + }, + { + "epoch": 0.31, + "learning_rate": 0.0005585396188565698, + "loss": 1.4116, + "theoretical_loss": 3.6710669837016976, + "tokens_seen": 1019215872 + }, + { + "epoch": 0.31, + "learning_rate": 0.0005585235707121364, + "loss": 1.374, + "theoretical_loss": 3.671044772907653, + "tokens_seen": 1019281408 + }, + { + "epoch": 0.31, + "learning_rate": 0.0005585075225677032, + "loss": 1.513, + "theoretical_loss": 3.671022563941465, + "tokens_seen": 1019346944 + }, + { + "epoch": 0.31, + "learning_rate": 0.0005584914744232698, + "loss": 1.4277, + "theoretical_loss": 3.671000356802865, + "tokens_seen": 1019412480 + }, + { + "epoch": 0.31, + "learning_rate": 0.0005584754262788366, + "loss": 1.5657, + "theoretical_loss": 3.6709781514915854, + "tokens_seen": 1019478016 + }, + { + "epoch": 0.31, + "learning_rate": 0.0005584593781344032, + "loss": 1.4709, + "theoretical_loss": 3.670955948007358, + "tokens_seen": 1019543552 + }, + { + "epoch": 0.31, + "learning_rate": 0.00055844332998997, + "loss": 1.391, + "theoretical_loss": 3.6709337463499154, + "tokens_seen": 1019609088 + }, + { + "epoch": 0.31, + "learning_rate": 0.0005584272818455366, + "loss": 1.4967, + "theoretical_loss": 3.6709115465189903, + "tokens_seen": 1019674624 + }, + { + "epoch": 0.31, + "learning_rate": 0.0005584112337011034, + "loss": 1.5388, + "theoretical_loss": 3.670889348514314, + "tokens_seen": 1019740160 + }, + { + "epoch": 0.31, + "learning_rate": 0.00055839518555667, + "loss": 1.4696, + "theoretical_loss": 3.67086715233562, + "tokens_seen": 1019805696 + }, + { + "epoch": 0.31, + "learning_rate": 0.0005583791374122368, + "loss": 1.3265, + "theoretical_loss": 3.6708449579826405, + "tokens_seen": 1019871232 + }, + { + "epoch": 0.31, + "learning_rate": 0.0005583630892678034, + "loss": 1.5355, + "theoretical_loss": 3.6708227654551076, + "tokens_seen": 1019936768 + }, + { + "epoch": 0.31, + "learning_rate": 0.0005583470411233702, + "loss": 1.5146, + "theoretical_loss": 3.6708005747527537, + "tokens_seen": 1020002304 + }, + { + "epoch": 0.31, + "learning_rate": 0.0005583309929789368, + "loss": 1.5251, + "theoretical_loss": 3.6707783858753125, + "tokens_seen": 1020067840 + }, + { + "epoch": 0.31, + "learning_rate": 0.0005583149448345036, + "loss": 1.4778, + "theoretical_loss": 3.6707561988225157, + "tokens_seen": 1020133376 + }, + { + "epoch": 0.31, + "learning_rate": 0.0005582988966900702, + "loss": 1.3035, + "theoretical_loss": 3.670734013594096, + "tokens_seen": 1020198912 + }, + { + "epoch": 0.31, + "learning_rate": 0.000558282848545637, + "loss": 1.5102, + "theoretical_loss": 3.670711830189787, + "tokens_seen": 1020264448 + }, + { + "epoch": 0.31, + "learning_rate": 0.0005582668004012036, + "loss": 1.5051, + "theoretical_loss": 3.6706896486093212, + "tokens_seen": 1020329984 + }, + { + "epoch": 0.31, + "learning_rate": 0.0005582507522567704, + "loss": 1.4925, + "theoretical_loss": 3.6706674688524314, + "tokens_seen": 1020395520 + }, + { + "epoch": 0.31, + "learning_rate": 0.000558234704112337, + "loss": 1.4974, + "theoretical_loss": 3.670645290918851, + "tokens_seen": 1020461056 + }, + { + "epoch": 0.31, + "learning_rate": 0.0005582186559679038, + "loss": 1.4231, + "theoretical_loss": 3.670623114808312, + "tokens_seen": 1020526592 + }, + { + "epoch": 0.31, + "learning_rate": 0.0005582026078234704, + "loss": 1.4131, + "theoretical_loss": 3.670600940520549, + "tokens_seen": 1020592128 + }, + { + "epoch": 0.31, + "learning_rate": 0.0005581865596790372, + "loss": 1.4525, + "theoretical_loss": 3.670578768055294, + "tokens_seen": 1020657664 + }, + { + "epoch": 0.31, + "learning_rate": 0.0005581705115346038, + "loss": 1.4334, + "theoretical_loss": 3.67055659741228, + "tokens_seen": 1020723200 + }, + { + "epoch": 0.31, + "learning_rate": 0.0005581544633901706, + "loss": 1.3968, + "theoretical_loss": 3.670534428591241, + "tokens_seen": 1020788736 + }, + { + "epoch": 0.31, + "learning_rate": 0.0005581384152457372, + "loss": 1.3299, + "theoretical_loss": 3.6705122615919104, + "tokens_seen": 1020854272 + }, + { + "epoch": 0.31, + "learning_rate": 0.000558122367101304, + "loss": 1.5596, + "theoretical_loss": 3.6704900964140212, + "tokens_seen": 1020919808 + }, + { + "epoch": 0.31, + "learning_rate": 0.0005581063189568706, + "loss": 1.5884, + "theoretical_loss": 3.6704679330573065, + "tokens_seen": 1020985344 + }, + { + "epoch": 0.31, + "learning_rate": 0.0005580902708124374, + "loss": 1.43, + "theoretical_loss": 3.6704457715215004, + "tokens_seen": 1021050880 + }, + { + "epoch": 0.31, + "learning_rate": 0.000558074222668004, + "loss": 1.5301, + "theoretical_loss": 3.6704236118063363, + "tokens_seen": 1021116416 + }, + { + "epoch": 0.31, + "learning_rate": 0.0005580581745235708, + "loss": 1.5593, + "theoretical_loss": 3.6704014539115475, + "tokens_seen": 1021181952 + }, + { + "epoch": 0.31, + "learning_rate": 0.0005580421263791374, + "loss": 1.4053, + "theoretical_loss": 3.670379297836868, + "tokens_seen": 1021247488 + }, + { + "epoch": 0.31, + "learning_rate": 0.0005580260782347042, + "loss": 1.3613, + "theoretical_loss": 3.6703571435820312, + "tokens_seen": 1021313024 + }, + { + "epoch": 0.31, + "learning_rate": 0.0005580100300902708, + "loss": 1.4977, + "theoretical_loss": 3.6703349911467713, + "tokens_seen": 1021378560 + }, + { + "epoch": 0.31, + "learning_rate": 0.0005579939819458376, + "loss": 1.5133, + "theoretical_loss": 3.670312840530821, + "tokens_seen": 1021444096 + }, + { + "epoch": 0.31, + "learning_rate": 0.0005579779338014042, + "loss": 1.3643, + "theoretical_loss": 3.670290691733916, + "tokens_seen": 1021509632 + }, + { + "epoch": 0.31, + "learning_rate": 0.000557961885656971, + "loss": 1.3238, + "theoretical_loss": 3.6702685447557886, + "tokens_seen": 1021575168 + }, + { + "epoch": 0.31, + "learning_rate": 0.0005579458375125376, + "loss": 1.4654, + "theoretical_loss": 3.6702463995961736, + "tokens_seen": 1021640704 + }, + { + "epoch": 0.31, + "learning_rate": 0.0005579297893681044, + "loss": 1.4025, + "theoretical_loss": 3.6702242562548046, + "tokens_seen": 1021706240 + }, + { + "epoch": 0.31, + "learning_rate": 0.000557913741223671, + "loss": 1.4499, + "theoretical_loss": 3.670202114731416, + "tokens_seen": 1021771776 + }, + { + "epoch": 0.31, + "learning_rate": 0.0005578976930792378, + "loss": 1.2978, + "theoretical_loss": 3.670179975025742, + "tokens_seen": 1021837312 + }, + { + "epoch": 0.31, + "learning_rate": 0.0005578816449348044, + "loss": 1.3859, + "theoretical_loss": 3.670157837137517, + "tokens_seen": 1021902848 + }, + { + "epoch": 0.31, + "learning_rate": 0.0005578655967903712, + "loss": 1.4953, + "theoretical_loss": 3.6701357010664744, + "tokens_seen": 1021968384 + }, + { + "epoch": 0.31, + "learning_rate": 0.0005578495486459378, + "loss": 1.4342, + "theoretical_loss": 3.6701135668123497, + "tokens_seen": 1022033920 + }, + { + "epoch": 0.31, + "learning_rate": 0.0005578335005015046, + "loss": 1.4226, + "theoretical_loss": 3.670091434374876, + "tokens_seen": 1022099456 + }, + { + "epoch": 0.31, + "learning_rate": 0.0005578174523570712, + "loss": 1.5036, + "theoretical_loss": 3.670069303753789, + "tokens_seen": 1022164992 + }, + { + "epoch": 0.31, + "learning_rate": 0.000557801404212638, + "loss": 1.5188, + "theoretical_loss": 3.670047174948822, + "tokens_seen": 1022230528 + }, + { + "epoch": 0.31, + "learning_rate": 0.0005577853560682046, + "loss": 1.5046, + "theoretical_loss": 3.6700250479597107, + "tokens_seen": 1022296064 + }, + { + "epoch": 0.31, + "objective/train/docs_used": 375283, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.4091981649398804, + "objective/train/theoretical_loss": 3.6700029227861886, + "objective/train/tokens_used": 1042821600, + "theoretical_loss": 3.6700029227861886, + "tokens_seen": 1022361600 + }, + { + "epoch": 0.31, + "learning_rate": 0.0005577693079237714, + "loss": 1.4656, + "theoretical_loss": 3.6700029227861886, + "tokens_seen": 1022361600 + }, + { + "epoch": 0.31, + "learning_rate": 0.000557753259779338, + "loss": 1.4024, + "theoretical_loss": 3.669980799427991, + "tokens_seen": 1022427136 + }, + { + "epoch": 0.31, + "learning_rate": 0.0005577372116349048, + "loss": 1.4399, + "theoretical_loss": 3.6699586778848525, + "tokens_seen": 1022492672 + }, + { + "epoch": 0.31, + "learning_rate": 0.0005577211634904714, + "loss": 1.354, + "theoretical_loss": 3.6699365581565084, + "tokens_seen": 1022558208 + }, + { + "epoch": 0.31, + "learning_rate": 0.0005577051153460382, + "loss": 1.428, + "theoretical_loss": 3.6699144402426924, + "tokens_seen": 1022623744 + }, + { + "epoch": 0.31, + "learning_rate": 0.0005576890672016048, + "loss": 1.4881, + "theoretical_loss": 3.66989232414314, + "tokens_seen": 1022689280 + }, + { + "epoch": 0.31, + "learning_rate": 0.0005576730190571716, + "loss": 1.4178, + "theoretical_loss": 3.6698702098575864, + "tokens_seen": 1022754816 + }, + { + "epoch": 0.31, + "learning_rate": 0.0005576569709127382, + "loss": 1.3437, + "theoretical_loss": 3.669848097385766, + "tokens_seen": 1022820352 + }, + { + "epoch": 0.31, + "learning_rate": 0.000557640922768305, + "loss": 1.4538, + "theoretical_loss": 3.669825986727414, + "tokens_seen": 1022885888 + }, + { + "epoch": 0.31, + "learning_rate": 0.0005576248746238716, + "loss": 1.5493, + "theoretical_loss": 3.669803877882266, + "tokens_seen": 1022951424 + }, + { + "epoch": 0.31, + "learning_rate": 0.0005576088264794384, + "loss": 1.6065, + "theoretical_loss": 3.6697817708500566, + "tokens_seen": 1023016960 + }, + { + "epoch": 0.31, + "learning_rate": 0.000557592778335005, + "loss": 1.4612, + "theoretical_loss": 3.6697596656305214, + "tokens_seen": 1023082496 + }, + { + "epoch": 0.31, + "learning_rate": 0.0005575767301905718, + "loss": 1.3907, + "theoretical_loss": 3.6697375622233954, + "tokens_seen": 1023148032 + }, + { + "epoch": 0.31, + "learning_rate": 0.0005575606820461384, + "loss": 1.4506, + "theoretical_loss": 3.6697154606284137, + "tokens_seen": 1023213568 + }, + { + "epoch": 0.31, + "learning_rate": 0.0005575446339017052, + "loss": 1.4503, + "theoretical_loss": 3.669693360845312, + "tokens_seen": 1023279104 + }, + { + "epoch": 0.31, + "learning_rate": 0.0005575285857572718, + "loss": 1.4444, + "theoretical_loss": 3.669671262873826, + "tokens_seen": 1023344640 + }, + { + "epoch": 0.31, + "learning_rate": 0.0005575125376128385, + "loss": 1.5223, + "theoretical_loss": 3.6696491667136906, + "tokens_seen": 1023410176 + }, + { + "epoch": 0.31, + "learning_rate": 0.0005574964894684052, + "loss": 1.592, + "theoretical_loss": 3.6696270723646416, + "tokens_seen": 1023475712 + }, + { + "epoch": 0.31, + "learning_rate": 0.0005574804413239719, + "loss": 1.5198, + "theoretical_loss": 3.669604979826415, + "tokens_seen": 1023541248 + }, + { + "epoch": 0.31, + "learning_rate": 0.0005574643931795387, + "loss": 1.425, + "theoretical_loss": 3.669582889098746, + "tokens_seen": 1023606784 + }, + { + "epoch": 0.31, + "learning_rate": 0.0005574483450351053, + "loss": 1.4696, + "theoretical_loss": 3.6695608001813707, + "tokens_seen": 1023672320 + }, + { + "epoch": 0.31, + "learning_rate": 0.000557432296890672, + "loss": 1.5605, + "theoretical_loss": 3.669538713074024, + "tokens_seen": 1023737856 + }, + { + "epoch": 0.31, + "learning_rate": 0.0005574162487462387, + "loss": 1.5252, + "theoretical_loss": 3.669516627776442, + "tokens_seen": 1023803392 + }, + { + "epoch": 0.31, + "learning_rate": 0.0005574002006018055, + "loss": 1.3696, + "theoretical_loss": 3.669494544288362, + "tokens_seen": 1023868928 + }, + { + "epoch": 0.31, + "learning_rate": 0.0005573841524573721, + "loss": 1.452, + "theoretical_loss": 3.6694724626095185, + "tokens_seen": 1023934464 + }, + { + "epoch": 0.31, + "learning_rate": 0.0005573681043129387, + "loss": 1.298, + "theoretical_loss": 3.6694503827396474, + "tokens_seen": 1024000000 + }, + { + "epoch": 0.31, + "learning_rate": 0.0005573520561685055, + "loss": 1.3907, + "theoretical_loss": 3.669428304678485, + "tokens_seen": 1024065536 + }, + { + "epoch": 0.31, + "learning_rate": 0.0005573360080240723, + "loss": 1.5223, + "theoretical_loss": 3.669406228425768, + "tokens_seen": 1024131072 + }, + { + "epoch": 0.31, + "learning_rate": 0.0005573199598796389, + "loss": 1.5184, + "theoretical_loss": 3.669384153981232, + "tokens_seen": 1024196608 + }, + { + "epoch": 0.31, + "learning_rate": 0.0005573039117352057, + "loss": 1.4623, + "theoretical_loss": 3.6693620813446133, + "tokens_seen": 1024262144 + }, + { + "epoch": 0.31, + "learning_rate": 0.0005572878635907723, + "loss": 1.4871, + "theoretical_loss": 3.669340010515648, + "tokens_seen": 1024327680 + }, + { + "epoch": 0.31, + "learning_rate": 0.0005572718154463391, + "loss": 1.3766, + "theoretical_loss": 3.6693179414940733, + "tokens_seen": 1024393216 + }, + { + "epoch": 0.31, + "learning_rate": 0.0005572557673019057, + "loss": 1.5222, + "theoretical_loss": 3.6692958742796242, + "tokens_seen": 1024458752 + }, + { + "epoch": 0.31, + "learning_rate": 0.0005572397191574725, + "loss": 1.3607, + "theoretical_loss": 3.6692738088720382, + "tokens_seen": 1024524288 + }, + { + "epoch": 0.31, + "learning_rate": 0.0005572236710130391, + "loss": 1.4997, + "theoretical_loss": 3.669251745271051, + "tokens_seen": 1024589824 + }, + { + "epoch": 0.31, + "learning_rate": 0.0005572076228686059, + "loss": 1.5193, + "theoretical_loss": 3.6692296834764, + "tokens_seen": 1024655360 + }, + { + "epoch": 0.31, + "learning_rate": 0.0005571915747241725, + "loss": 1.4386, + "theoretical_loss": 3.6692076234878215, + "tokens_seen": 1024720896 + }, + { + "epoch": 0.31, + "learning_rate": 0.0005571755265797393, + "loss": 1.4457, + "theoretical_loss": 3.6691855653050514, + "tokens_seen": 1024786432 + }, + { + "epoch": 0.31, + "learning_rate": 0.0005571594784353059, + "loss": 1.4651, + "theoretical_loss": 3.669163508927827, + "tokens_seen": 1024851968 + }, + { + "epoch": 0.31, + "learning_rate": 0.0005571434302908727, + "loss": 1.5148, + "theoretical_loss": 3.6691414543558856, + "tokens_seen": 1024917504 + }, + { + "epoch": 0.31, + "learning_rate": 0.0005571273821464393, + "loss": 1.5215, + "theoretical_loss": 3.669119401588963, + "tokens_seen": 1024983040 + }, + { + "epoch": 0.31, + "learning_rate": 0.0005571113340020061, + "loss": 1.3784, + "theoretical_loss": 3.669097350626797, + "tokens_seen": 1025048576 + }, + { + "epoch": 0.31, + "learning_rate": 0.0005570952858575727, + "loss": 1.4382, + "theoretical_loss": 3.669075301469124, + "tokens_seen": 1025114112 + }, + { + "epoch": 0.31, + "learning_rate": 0.0005570792377131395, + "loss": 1.3591, + "theoretical_loss": 3.6690532541156813, + "tokens_seen": 1025179648 + }, + { + "epoch": 0.31, + "learning_rate": 0.0005570631895687061, + "loss": 1.3791, + "theoretical_loss": 3.6690312085662056, + "tokens_seen": 1025245184 + }, + { + "epoch": 0.31, + "learning_rate": 0.0005570471414242729, + "loss": 1.4935, + "theoretical_loss": 3.6690091648204337, + "tokens_seen": 1025310720 + }, + { + "epoch": 0.31, + "learning_rate": 0.0005570310932798395, + "loss": 1.4058, + "theoretical_loss": 3.6689871228781037, + "tokens_seen": 1025376256 + }, + { + "epoch": 0.31, + "learning_rate": 0.0005570150451354063, + "loss": 1.4479, + "theoretical_loss": 3.6689650827389517, + "tokens_seen": 1025441792 + }, + { + "epoch": 0.31, + "learning_rate": 0.0005569989969909729, + "loss": 1.3513, + "theoretical_loss": 3.6689430444027162, + "tokens_seen": 1025507328 + }, + { + "epoch": 0.31, + "learning_rate": 0.0005569829488465397, + "loss": 1.4891, + "theoretical_loss": 3.668921007869134, + "tokens_seen": 1025572864 + }, + { + "epoch": 0.31, + "objective/train/docs_used": 376516, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.2563257217407227, + "objective/train/theoretical_loss": 3.6688989731379413, + "objective/train/tokens_used": 1046098400, + "theoretical_loss": 3.6688989731379413, + "tokens_seen": 1025638400 + }, + { + "epoch": 0.31, + "learning_rate": 0.0005569669007021063, + "loss": 1.3371, + "theoretical_loss": 3.6688989731379413, + "tokens_seen": 1025638400 + }, + { + "epoch": 0.31, + "learning_rate": 0.0005569508525576731, + "loss": 1.3177, + "theoretical_loss": 3.6688769402088774, + "tokens_seen": 1025703936 + }, + { + "epoch": 0.31, + "learning_rate": 0.0005569348044132397, + "loss": 1.5283, + "theoretical_loss": 3.668854909081679, + "tokens_seen": 1025769472 + }, + { + "epoch": 0.31, + "learning_rate": 0.0005569187562688065, + "loss": 1.3241, + "theoretical_loss": 3.668832879756083, + "tokens_seen": 1025835008 + }, + { + "epoch": 0.31, + "learning_rate": 0.0005569027081243731, + "loss": 1.4777, + "theoretical_loss": 3.668810852231828, + "tokens_seen": 1025900544 + }, + { + "epoch": 0.31, + "learning_rate": 0.0005568866599799399, + "loss": 1.4248, + "theoretical_loss": 3.668788826508651, + "tokens_seen": 1025966080 + }, + { + "epoch": 0.31, + "learning_rate": 0.0005568706118355065, + "loss": 1.4246, + "theoretical_loss": 3.6687668025862905, + "tokens_seen": 1026031616 + }, + { + "epoch": 0.31, + "learning_rate": 0.0005568545636910733, + "loss": 1.469, + "theoretical_loss": 3.668744780464483, + "tokens_seen": 1026097152 + }, + { + "epoch": 0.31, + "learning_rate": 0.0005568385155466399, + "loss": 1.3827, + "theoretical_loss": 3.668722760142967, + "tokens_seen": 1026162688 + }, + { + "epoch": 0.31, + "learning_rate": 0.0005568224674022067, + "loss": 1.3234, + "theoretical_loss": 3.668700741621481, + "tokens_seen": 1026228224 + }, + { + "epoch": 0.31, + "learning_rate": 0.0005568064192577733, + "loss": 1.4878, + "theoretical_loss": 3.6686787248997614, + "tokens_seen": 1026293760 + }, + { + "epoch": 0.31, + "learning_rate": 0.0005567903711133401, + "loss": 1.3607, + "theoretical_loss": 3.6686567099775473, + "tokens_seen": 1026359296 + }, + { + "epoch": 0.31, + "learning_rate": 0.0005567743229689067, + "loss": 1.4724, + "theoretical_loss": 3.6686346968545767, + "tokens_seen": 1026424832 + }, + { + "epoch": 0.31, + "learning_rate": 0.0005567582748244735, + "loss": 1.4722, + "theoretical_loss": 3.6686126855305874, + "tokens_seen": 1026490368 + }, + { + "epoch": 0.31, + "learning_rate": 0.0005567422266800401, + "loss": 1.4466, + "theoretical_loss": 3.6685906760053175, + "tokens_seen": 1026555904 + }, + { + "epoch": 0.31, + "learning_rate": 0.0005567261785356069, + "loss": 1.4488, + "theoretical_loss": 3.6685686682785055, + "tokens_seen": 1026621440 + }, + { + "epoch": 0.31, + "learning_rate": 0.0005567101303911735, + "loss": 1.3769, + "theoretical_loss": 3.6685466623498892, + "tokens_seen": 1026686976 + }, + { + "epoch": 0.31, + "learning_rate": 0.0005566940822467403, + "loss": 1.4292, + "theoretical_loss": 3.6685246582192077, + "tokens_seen": 1026752512 + }, + { + "epoch": 0.31, + "learning_rate": 0.0005566780341023069, + "loss": 1.4762, + "theoretical_loss": 3.668502655886198, + "tokens_seen": 1026818048 + }, + { + "epoch": 0.31, + "learning_rate": 0.0005566619859578737, + "loss": 1.4536, + "theoretical_loss": 3.6684806553505998, + "tokens_seen": 1026883584 + }, + { + "epoch": 0.31, + "learning_rate": 0.0005566459378134403, + "loss": 1.4686, + "theoretical_loss": 3.6684586566121506, + "tokens_seen": 1026949120 + }, + { + "epoch": 0.31, + "learning_rate": 0.0005566298896690071, + "loss": 1.382, + "theoretical_loss": 3.66843665967059, + "tokens_seen": 1027014656 + }, + { + "epoch": 0.31, + "learning_rate": 0.0005566138415245737, + "loss": 1.5434, + "theoretical_loss": 3.668414664525656, + "tokens_seen": 1027080192 + }, + { + "epoch": 0.31, + "learning_rate": 0.0005565977933801405, + "loss": 1.4309, + "theoretical_loss": 3.668392671177087, + "tokens_seen": 1027145728 + }, + { + "epoch": 0.31, + "learning_rate": 0.0005565817452357071, + "loss": 1.5432, + "theoretical_loss": 3.668370679624622, + "tokens_seen": 1027211264 + }, + { + "epoch": 0.31, + "learning_rate": 0.0005565656970912739, + "loss": 1.4804, + "theoretical_loss": 3.6683486898679996, + "tokens_seen": 1027276800 + }, + { + "epoch": 0.31, + "learning_rate": 0.0005565496489468405, + "loss": 1.4234, + "theoretical_loss": 3.6683267019069588, + "tokens_seen": 1027342336 + }, + { + "epoch": 0.31, + "learning_rate": 0.0005565336008024073, + "loss": 1.5163, + "theoretical_loss": 3.668304715741238, + "tokens_seen": 1027407872 + }, + { + "epoch": 0.31, + "learning_rate": 0.0005565175526579739, + "loss": 1.3724, + "theoretical_loss": 3.6682827313705766, + "tokens_seen": 1027473408 + }, + { + "epoch": 0.31, + "learning_rate": 0.0005565015045135407, + "loss": 1.5222, + "theoretical_loss": 3.6682607487947134, + "tokens_seen": 1027538944 + }, + { + "epoch": 0.31, + "learning_rate": 0.0005564854563691073, + "loss": 1.4195, + "theoretical_loss": 3.668238768013387, + "tokens_seen": 1027604480 + }, + { + "epoch": 0.31, + "learning_rate": 0.0005564694082246741, + "loss": 1.5277, + "theoretical_loss": 3.6682167890263377, + "tokens_seen": 1027670016 + }, + { + "epoch": 0.31, + "learning_rate": 0.0005564533600802407, + "loss": 1.4185, + "theoretical_loss": 3.668194811833303, + "tokens_seen": 1027735552 + }, + { + "epoch": 0.31, + "learning_rate": 0.0005564373119358075, + "loss": 1.4171, + "theoretical_loss": 3.6681728364340236, + "tokens_seen": 1027801088 + }, + { + "epoch": 0.31, + "learning_rate": 0.0005564212637913741, + "loss": 1.4832, + "theoretical_loss": 3.668150862828237, + "tokens_seen": 1027866624 + }, + { + "epoch": 0.31, + "learning_rate": 0.0005564052156469409, + "loss": 1.3158, + "theoretical_loss": 3.668128891015684, + "tokens_seen": 1027932160 + }, + { + "epoch": 0.31, + "learning_rate": 0.0005563891675025075, + "loss": 1.5123, + "theoretical_loss": 3.6681069209961037, + "tokens_seen": 1027997696 + }, + { + "epoch": 0.31, + "learning_rate": 0.0005563731193580743, + "loss": 1.5235, + "theoretical_loss": 3.668084952769235, + "tokens_seen": 1028063232 + }, + { + "epoch": 0.31, + "learning_rate": 0.0005563570712136409, + "loss": 1.4617, + "theoretical_loss": 3.6680629863348178, + "tokens_seen": 1028128768 + }, + { + "epoch": 0.31, + "learning_rate": 0.0005563410230692077, + "loss": 1.4031, + "theoretical_loss": 3.6680410216925914, + "tokens_seen": 1028194304 + }, + { + "epoch": 0.31, + "learning_rate": 0.0005563249749247743, + "loss": 1.396, + "theoretical_loss": 3.6680190588422947, + "tokens_seen": 1028259840 + }, + { + "epoch": 0.31, + "learning_rate": 0.0005563089267803411, + "loss": 1.4351, + "theoretical_loss": 3.6679970977836684, + "tokens_seen": 1028325376 + }, + { + "epoch": 0.31, + "learning_rate": 0.0005562928786359077, + "loss": 1.519, + "theoretical_loss": 3.667975138516452, + "tokens_seen": 1028390912 + }, + { + "epoch": 0.31, + "learning_rate": 0.0005562768304914745, + "loss": 1.3919, + "theoretical_loss": 3.6679531810403843, + "tokens_seen": 1028456448 + }, + { + "epoch": 0.31, + "learning_rate": 0.0005562607823470411, + "loss": 1.5321, + "theoretical_loss": 3.6679312253552063, + "tokens_seen": 1028521984 + }, + { + "epoch": 0.31, + "learning_rate": 0.0005562447342026079, + "loss": 1.4168, + "theoretical_loss": 3.6679092714606574, + "tokens_seen": 1028587520 + }, + { + "epoch": 0.31, + "learning_rate": 0.0005562286860581745, + "loss": 1.4011, + "theoretical_loss": 3.6678873193564767, + "tokens_seen": 1028653056 + }, + { + "epoch": 0.31, + "learning_rate": 0.0005562126379137413, + "loss": 1.5095, + "theoretical_loss": 3.6678653690424055, + "tokens_seen": 1028718592 + }, + { + "epoch": 0.31, + "learning_rate": 0.000556196589769308, + "loss": 1.3434, + "theoretical_loss": 3.6678434205181825, + "tokens_seen": 1028784128 + }, + { + "epoch": 0.31, + "learning_rate": 0.0005561805416248747, + "loss": 1.4249, + "theoretical_loss": 3.6678214737835484, + "tokens_seen": 1028849664 + }, + { + "epoch": 0.31, + "objective/train/docs_used": 377682, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.3685638904571533, + "objective/train/theoretical_loss": 3.667799528838244, + "objective/train/tokens_used": 1049375200, + "theoretical_loss": 3.667799528838244, + "tokens_seen": 1028915200 + }, + { + "epoch": 0.31, + "learning_rate": 0.0005561644934804413, + "loss": 1.3986, + "theoretical_loss": 3.667799528838244, + "tokens_seen": 1028915200 + }, + { + "epoch": 0.31, + "learning_rate": 0.0005561484453360081, + "loss": 1.4931, + "theoretical_loss": 3.667777585682008, + "tokens_seen": 1028980736 + }, + { + "epoch": 0.31, + "learning_rate": 0.0005561323971915747, + "loss": 1.553, + "theoretical_loss": 3.667755644314581, + "tokens_seen": 1029046272 + }, + { + "epoch": 0.31, + "learning_rate": 0.0005561163490471414, + "loss": 1.3488, + "theoretical_loss": 3.667733704735704, + "tokens_seen": 1029111808 + }, + { + "epoch": 0.31, + "learning_rate": 0.0005561003009027082, + "loss": 1.4186, + "theoretical_loss": 3.667711766945117, + "tokens_seen": 1029177344 + }, + { + "epoch": 0.31, + "learning_rate": 0.0005560842527582748, + "loss": 1.4893, + "theoretical_loss": 3.6676898309425603, + "tokens_seen": 1029242880 + }, + { + "epoch": 0.31, + "learning_rate": 0.0005560682046138416, + "loss": 1.5085, + "theoretical_loss": 3.6676678967277745, + "tokens_seen": 1029308416 + }, + { + "epoch": 0.31, + "learning_rate": 0.0005560521564694082, + "loss": 1.5175, + "theoretical_loss": 3.6676459643004993, + "tokens_seen": 1029373952 + }, + { + "epoch": 0.31, + "learning_rate": 0.000556036108324975, + "loss": 1.5014, + "theoretical_loss": 3.667624033660476, + "tokens_seen": 1029439488 + }, + { + "epoch": 0.31, + "learning_rate": 0.0005560200601805416, + "loss": 1.4926, + "theoretical_loss": 3.6676021048074454, + "tokens_seen": 1029505024 + }, + { + "epoch": 0.31, + "learning_rate": 0.0005560040120361084, + "loss": 1.3831, + "theoretical_loss": 3.6675801777411476, + "tokens_seen": 1029570560 + }, + { + "epoch": 0.31, + "learning_rate": 0.000555987963891675, + "loss": 1.4332, + "theoretical_loss": 3.6675582524613235, + "tokens_seen": 1029636096 + }, + { + "epoch": 0.31, + "learning_rate": 0.0005559719157472418, + "loss": 1.4924, + "theoretical_loss": 3.667536328967714, + "tokens_seen": 1029701632 + }, + { + "epoch": 0.31, + "learning_rate": 0.0005559558676028084, + "loss": 1.5095, + "theoretical_loss": 3.6675144072600596, + "tokens_seen": 1029767168 + }, + { + "epoch": 0.31, + "learning_rate": 0.0005559398194583752, + "loss": 1.417, + "theoretical_loss": 3.667492487338101, + "tokens_seen": 1029832704 + }, + { + "epoch": 0.31, + "learning_rate": 0.0005559237713139418, + "loss": 1.5489, + "theoretical_loss": 3.66747056920158, + "tokens_seen": 1029898240 + }, + { + "epoch": 0.31, + "learning_rate": 0.0005559077231695086, + "loss": 1.4308, + "theoretical_loss": 3.667448652850237, + "tokens_seen": 1029963776 + }, + { + "epoch": 0.31, + "learning_rate": 0.0005558916750250752, + "loss": 1.5385, + "theoretical_loss": 3.6674267382838126, + "tokens_seen": 1030029312 + }, + { + "epoch": 0.31, + "learning_rate": 0.000555875626880642, + "loss": 1.3806, + "theoretical_loss": 3.6674048255020484, + "tokens_seen": 1030094848 + }, + { + "epoch": 0.31, + "learning_rate": 0.0005558595787362086, + "loss": 1.3626, + "theoretical_loss": 3.6673829145046852, + "tokens_seen": 1030160384 + }, + { + "epoch": 0.31, + "learning_rate": 0.0005558435305917754, + "loss": 1.3931, + "theoretical_loss": 3.667361005291465, + "tokens_seen": 1030225920 + }, + { + "epoch": 0.31, + "learning_rate": 0.000555827482447342, + "loss": 1.4473, + "theoretical_loss": 3.6673390978621283, + "tokens_seen": 1030291456 + }, + { + "epoch": 0.31, + "learning_rate": 0.0005558114343029088, + "loss": 1.4588, + "theoretical_loss": 3.6673171922164163, + "tokens_seen": 1030356992 + }, + { + "epoch": 0.31, + "learning_rate": 0.0005557953861584754, + "loss": 1.3822, + "theoretical_loss": 3.6672952883540706, + "tokens_seen": 1030422528 + }, + { + "epoch": 0.31, + "learning_rate": 0.0005557793380140422, + "loss": 1.4485, + "theoretical_loss": 3.6672733862748332, + "tokens_seen": 1030488064 + }, + { + "epoch": 0.31, + "learning_rate": 0.0005557632898696088, + "loss": 1.3503, + "theoretical_loss": 3.6672514859784444, + "tokens_seen": 1030553600 + }, + { + "epoch": 0.31, + "learning_rate": 0.0005557472417251756, + "loss": 1.4531, + "theoretical_loss": 3.6672295874646466, + "tokens_seen": 1030619136 + }, + { + "epoch": 0.31, + "learning_rate": 0.0005557311935807422, + "loss": 1.4395, + "theoretical_loss": 3.6672076907331808, + "tokens_seen": 1030684672 + }, + { + "epoch": 0.31, + "learning_rate": 0.000555715145436309, + "loss": 1.5836, + "theoretical_loss": 3.6671857957837886, + "tokens_seen": 1030750208 + }, + { + "epoch": 0.31, + "learning_rate": 0.0005556990972918756, + "loss": 1.4327, + "theoretical_loss": 3.6671639026162124, + "tokens_seen": 1030815744 + }, + { + "epoch": 0.31, + "learning_rate": 0.0005556830491474424, + "loss": 1.4306, + "theoretical_loss": 3.6671420112301933, + "tokens_seen": 1030881280 + }, + { + "epoch": 0.31, + "learning_rate": 0.000555667001003009, + "loss": 1.5207, + "theoretical_loss": 3.6671201216254734, + "tokens_seen": 1030946816 + }, + { + "epoch": 0.31, + "learning_rate": 0.0005556509528585758, + "loss": 1.4557, + "theoretical_loss": 3.667098233801794, + "tokens_seen": 1031012352 + }, + { + "epoch": 0.31, + "learning_rate": 0.0005556349047141424, + "loss": 1.3468, + "theoretical_loss": 3.667076347758898, + "tokens_seen": 1031077888 + }, + { + "epoch": 0.31, + "learning_rate": 0.0005556188565697092, + "loss": 1.4672, + "theoretical_loss": 3.6670544634965263, + "tokens_seen": 1031143424 + }, + { + "epoch": 0.31, + "learning_rate": 0.0005556028084252758, + "loss": 1.5658, + "theoretical_loss": 3.6670325810144213, + "tokens_seen": 1031208960 + }, + { + "epoch": 0.31, + "learning_rate": 0.0005555867602808426, + "loss": 1.4319, + "theoretical_loss": 3.6670107003123253, + "tokens_seen": 1031274496 + }, + { + "epoch": 0.31, + "learning_rate": 0.0005555707121364092, + "loss": 1.3767, + "theoretical_loss": 3.66698882138998, + "tokens_seen": 1031340032 + }, + { + "epoch": 0.31, + "learning_rate": 0.000555554663991976, + "loss": 1.5641, + "theoretical_loss": 3.666966944247128, + "tokens_seen": 1031405568 + }, + { + "epoch": 0.31, + "learning_rate": 0.0005555386158475426, + "loss": 1.5351, + "theoretical_loss": 3.666945068883511, + "tokens_seen": 1031471104 + }, + { + "epoch": 0.31, + "learning_rate": 0.0005555225677031094, + "loss": 1.6153, + "theoretical_loss": 3.6669231952988715, + "tokens_seen": 1031536640 + }, + { + "epoch": 0.31, + "learning_rate": 0.000555506519558676, + "loss": 1.4615, + "theoretical_loss": 3.6669013234929517, + "tokens_seen": 1031602176 + }, + { + "epoch": 0.31, + "learning_rate": 0.0005554904714142428, + "loss": 1.4671, + "theoretical_loss": 3.6668794534654943, + "tokens_seen": 1031667712 + }, + { + "epoch": 0.31, + "learning_rate": 0.0005554744232698094, + "loss": 1.386, + "theoretical_loss": 3.6668575852162415, + "tokens_seen": 1031733248 + }, + { + "epoch": 0.31, + "learning_rate": 0.0005554583751253762, + "loss": 1.46, + "theoretical_loss": 3.6668357187449354, + "tokens_seen": 1031798784 + }, + { + "epoch": 0.31, + "learning_rate": 0.0005554423269809428, + "loss": 1.3775, + "theoretical_loss": 3.6668138540513198, + "tokens_seen": 1031864320 + }, + { + "epoch": 0.31, + "learning_rate": 0.0005554262788365096, + "loss": 1.4636, + "theoretical_loss": 3.6667919911351357, + "tokens_seen": 1031929856 + }, + { + "epoch": 0.31, + "learning_rate": 0.0005554102306920762, + "loss": 1.4216, + "theoretical_loss": 3.6667701299961264, + "tokens_seen": 1031995392 + }, + { + "epoch": 0.31, + "learning_rate": 0.000555394182547643, + "loss": 1.3583, + "theoretical_loss": 3.6667482706340353, + "tokens_seen": 1032060928 + }, + { + "epoch": 0.31, + "learning_rate": 0.0005553781344032096, + "loss": 1.3733, + "theoretical_loss": 3.666726413048604, + "tokens_seen": 1032126464 + }, + { + "epoch": 0.31, + "objective/train/docs_used": 378946, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.438767910003662, + "objective/train/theoretical_loss": 3.6667045572395756, + "objective/train/tokens_used": 1052652000, + "theoretical_loss": 3.6667045572395756, + "tokens_seen": 1032192000 + }, + { + "epoch": 0.31, + "learning_rate": 0.0005553620862587764, + "loss": 1.3703, + "theoretical_loss": 3.6667045572395756, + "tokens_seen": 1032192000 + }, + { + "epoch": 0.31, + "learning_rate": 0.000555346038114343, + "loss": 1.3223, + "theoretical_loss": 3.6666827032066935, + "tokens_seen": 1032257536 + }, + { + "epoch": 0.31, + "learning_rate": 0.0005553299899699098, + "loss": 1.4145, + "theoretical_loss": 3.6666608509497003, + "tokens_seen": 1032323072 + }, + { + "epoch": 0.31, + "learning_rate": 0.0005553139418254764, + "loss": 1.3236, + "theoretical_loss": 3.666639000468339, + "tokens_seen": 1032388608 + }, + { + "epoch": 0.31, + "learning_rate": 0.0005552978936810432, + "loss": 1.4085, + "theoretical_loss": 3.6666171517623525, + "tokens_seen": 1032454144 + }, + { + "epoch": 0.31, + "learning_rate": 0.0005552818455366098, + "loss": 1.3613, + "theoretical_loss": 3.6665953048314837, + "tokens_seen": 1032519680 + }, + { + "epoch": 0.31, + "learning_rate": 0.0005552657973921766, + "loss": 1.462, + "theoretical_loss": 3.666573459675476, + "tokens_seen": 1032585216 + }, + { + "epoch": 0.31, + "learning_rate": 0.0005552497492477432, + "loss": 1.4087, + "theoretical_loss": 3.666551616294073, + "tokens_seen": 1032650752 + }, + { + "epoch": 0.31, + "learning_rate": 0.00055523370110331, + "loss": 1.5297, + "theoretical_loss": 3.6665297746870165, + "tokens_seen": 1032716288 + }, + { + "epoch": 0.31, + "learning_rate": 0.0005552176529588766, + "loss": 1.5462, + "theoretical_loss": 3.6665079348540512, + "tokens_seen": 1032781824 + }, + { + "epoch": 0.31, + "learning_rate": 0.0005552016048144434, + "loss": 1.4779, + "theoretical_loss": 3.6664860967949204, + "tokens_seen": 1032847360 + }, + { + "epoch": 0.31, + "learning_rate": 0.00055518555667001, + "loss": 1.5122, + "theoretical_loss": 3.6664642605093665, + "tokens_seen": 1032912896 + }, + { + "epoch": 0.31, + "learning_rate": 0.0005551695085255768, + "loss": 1.3743, + "theoretical_loss": 3.6664424259971335, + "tokens_seen": 1032978432 + }, + { + "epoch": 0.31, + "learning_rate": 0.0005551534603811434, + "loss": 1.2686, + "theoretical_loss": 3.666420593257965, + "tokens_seen": 1033043968 + }, + { + "epoch": 0.31, + "learning_rate": 0.0005551374122367102, + "loss": 1.3163, + "theoretical_loss": 3.6663987622916046, + "tokens_seen": 1033109504 + }, + { + "epoch": 0.31, + "learning_rate": 0.0005551213640922768, + "loss": 1.2639, + "theoretical_loss": 3.6663769330977956, + "tokens_seen": 1033175040 + }, + { + "epoch": 0.31, + "learning_rate": 0.0005551053159478436, + "loss": 1.3755, + "theoretical_loss": 3.6663551056762818, + "tokens_seen": 1033240576 + }, + { + "epoch": 0.31, + "learning_rate": 0.0005550892678034102, + "loss": 1.5097, + "theoretical_loss": 3.666333280026807, + "tokens_seen": 1033306112 + }, + { + "epoch": 0.31, + "learning_rate": 0.000555073219658977, + "loss": 1.4093, + "theoretical_loss": 3.6663114561491144, + "tokens_seen": 1033371648 + }, + { + "epoch": 0.31, + "learning_rate": 0.0005550571715145436, + "loss": 1.4885, + "theoretical_loss": 3.6662896340429487, + "tokens_seen": 1033437184 + }, + { + "epoch": 0.31, + "learning_rate": 0.0005550411233701104, + "loss": 1.457, + "theoretical_loss": 3.6662678137080533, + "tokens_seen": 1033502720 + }, + { + "epoch": 0.31, + "learning_rate": 0.000555025075225677, + "loss": 1.266, + "theoretical_loss": 3.6662459951441724, + "tokens_seen": 1033568256 + }, + { + "epoch": 0.31, + "learning_rate": 0.0005550090270812438, + "loss": 1.3484, + "theoretical_loss": 3.6662241783510496, + "tokens_seen": 1033633792 + }, + { + "epoch": 0.31, + "learning_rate": 0.0005549929789368104, + "loss": 1.3708, + "theoretical_loss": 3.6662023633284293, + "tokens_seen": 1033699328 + }, + { + "epoch": 0.31, + "learning_rate": 0.0005549769307923772, + "loss": 1.3684, + "theoretical_loss": 3.6661805500760547, + "tokens_seen": 1033764864 + }, + { + "epoch": 0.31, + "learning_rate": 0.0005549608826479438, + "loss": 1.4901, + "theoretical_loss": 3.6661587385936714, + "tokens_seen": 1033830400 + }, + { + "epoch": 0.31, + "learning_rate": 0.0005549448345035106, + "loss": 1.5573, + "theoretical_loss": 3.6661369288810226, + "tokens_seen": 1033895936 + }, + { + "epoch": 0.31, + "learning_rate": 0.0005549287863590774, + "loss": 1.3865, + "theoretical_loss": 3.666115120937852, + "tokens_seen": 1033961472 + }, + { + "epoch": 0.31, + "learning_rate": 0.000554912738214644, + "loss": 1.4145, + "theoretical_loss": 3.6660933147639057, + "tokens_seen": 1034027008 + }, + { + "epoch": 0.31, + "learning_rate": 0.0005548966900702108, + "loss": 1.2599, + "theoretical_loss": 3.6660715103589263, + "tokens_seen": 1034092544 + }, + { + "epoch": 0.31, + "learning_rate": 0.0005548806419257774, + "loss": 1.319, + "theoretical_loss": 3.6660497077226593, + "tokens_seen": 1034158080 + }, + { + "epoch": 0.31, + "learning_rate": 0.000554864593781344, + "loss": 1.4456, + "theoretical_loss": 3.6660279068548487, + "tokens_seen": 1034223616 + }, + { + "epoch": 0.31, + "learning_rate": 0.0005548485456369108, + "loss": 1.4651, + "theoretical_loss": 3.666006107755239, + "tokens_seen": 1034289152 + }, + { + "epoch": 0.31, + "learning_rate": 0.0005548324974924774, + "loss": 1.5168, + "theoretical_loss": 3.665984310423574, + "tokens_seen": 1034354688 + }, + { + "epoch": 0.31, + "learning_rate": 0.0005548164493480442, + "loss": 1.4177, + "theoretical_loss": 3.6659625148596002, + "tokens_seen": 1034420224 + }, + { + "epoch": 0.31, + "learning_rate": 0.0005548004012036108, + "loss": 1.4345, + "theoretical_loss": 3.665940721063061, + "tokens_seen": 1034485760 + }, + { + "epoch": 0.31, + "learning_rate": 0.0005547843530591776, + "loss": 1.5441, + "theoretical_loss": 3.6659189290337006, + "tokens_seen": 1034551296 + }, + { + "epoch": 0.31, + "learning_rate": 0.0005547683049147442, + "loss": 1.4639, + "theoretical_loss": 3.665897138771265, + "tokens_seen": 1034616832 + }, + { + "epoch": 0.31, + "learning_rate": 0.000554752256770311, + "loss": 1.3353, + "theoretical_loss": 3.6658753502754986, + "tokens_seen": 1034682368 + }, + { + "epoch": 0.31, + "learning_rate": 0.0005547362086258777, + "loss": 1.3975, + "theoretical_loss": 3.665853563546146, + "tokens_seen": 1034747904 + }, + { + "epoch": 0.31, + "learning_rate": 0.0005547201604814443, + "loss": 1.4699, + "theoretical_loss": 3.6658317785829526, + "tokens_seen": 1034813440 + }, + { + "epoch": 0.31, + "learning_rate": 0.000554704112337011, + "loss": 1.4537, + "theoretical_loss": 3.665809995385663, + "tokens_seen": 1034878976 + }, + { + "epoch": 0.31, + "learning_rate": 0.0005546880641925777, + "loss": 1.4436, + "theoretical_loss": 3.665788213954022, + "tokens_seen": 1034944512 + }, + { + "epoch": 0.31, + "learning_rate": 0.0005546720160481445, + "loss": 1.4772, + "theoretical_loss": 3.6657664342877756, + "tokens_seen": 1035010048 + }, + { + "epoch": 0.31, + "learning_rate": 0.0005546559679037111, + "loss": 1.429, + "theoretical_loss": 3.665744656386668, + "tokens_seen": 1035075584 + }, + { + "epoch": 0.31, + "learning_rate": 0.0005546399197592779, + "loss": 1.4558, + "theoretical_loss": 3.6657228802504447, + "tokens_seen": 1035141120 + }, + { + "epoch": 0.31, + "learning_rate": 0.0005546238716148445, + "loss": 1.4417, + "theoretical_loss": 3.6657011058788513, + "tokens_seen": 1035206656 + }, + { + "epoch": 0.31, + "learning_rate": 0.0005546078234704113, + "loss": 1.3709, + "theoretical_loss": 3.665679333271633, + "tokens_seen": 1035272192 + }, + { + "epoch": 0.31, + "learning_rate": 0.0005545917753259779, + "loss": 1.6127, + "theoretical_loss": 3.6656575624285344, + "tokens_seen": 1035337728 + }, + { + "epoch": 0.31, + "learning_rate": 0.0005545757271815447, + "loss": 1.4771, + "theoretical_loss": 3.665635793349302, + "tokens_seen": 1035403264 + }, + { + "epoch": 0.31, + "objective/train/docs_used": 379908, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.459204912185669, + "objective/train/theoretical_loss": 3.6656140260336807, + "objective/train/tokens_used": 1055928800, + "theoretical_loss": 3.6656140260336807, + "tokens_seen": 1035468800 + }, + { + "epoch": 0.31, + "learning_rate": 0.0005545596790371113, + "loss": 1.4121, + "theoretical_loss": 3.6656140260336807, + "tokens_seen": 1035468800 + }, + { + "epoch": 0.31, + "learning_rate": 0.0005545436308926781, + "loss": 1.4015, + "theoretical_loss": 3.665592260481416, + "tokens_seen": 1035534336 + }, + { + "epoch": 0.31, + "learning_rate": 0.0005545275827482447, + "loss": 1.3204, + "theoretical_loss": 3.6655704966922533, + "tokens_seen": 1035599872 + }, + { + "epoch": 0.31, + "learning_rate": 0.0005545115346038115, + "loss": 1.4409, + "theoretical_loss": 3.665548734665939, + "tokens_seen": 1035665408 + }, + { + "epoch": 0.31, + "learning_rate": 0.0005544954864593781, + "loss": 1.4891, + "theoretical_loss": 3.6655269744022174, + "tokens_seen": 1035730944 + }, + { + "epoch": 0.31, + "learning_rate": 0.0005544794383149449, + "loss": 1.4777, + "theoretical_loss": 3.665505215900836, + "tokens_seen": 1035796480 + }, + { + "epoch": 0.31, + "learning_rate": 0.0005544633901705115, + "loss": 1.4859, + "theoretical_loss": 3.665483459161539, + "tokens_seen": 1035862016 + }, + { + "epoch": 0.31, + "learning_rate": 0.0005544473420260783, + "loss": 1.4183, + "theoretical_loss": 3.6654617041840734, + "tokens_seen": 1035927552 + }, + { + "epoch": 0.31, + "learning_rate": 0.0005544312938816449, + "loss": 1.5053, + "theoretical_loss": 3.6654399509681843, + "tokens_seen": 1035993088 + }, + { + "epoch": 0.31, + "learning_rate": 0.0005544152457372117, + "loss": 1.4582, + "theoretical_loss": 3.665418199513618, + "tokens_seen": 1036058624 + }, + { + "epoch": 0.31, + "learning_rate": 0.0005543991975927783, + "loss": 1.4208, + "theoretical_loss": 3.6653964498201197, + "tokens_seen": 1036124160 + }, + { + "epoch": 0.31, + "learning_rate": 0.0005543831494483451, + "loss": 1.2879, + "theoretical_loss": 3.665374701887437, + "tokens_seen": 1036189696 + }, + { + "epoch": 0.31, + "learning_rate": 0.0005543671013039117, + "loss": 1.523, + "theoretical_loss": 3.6653529557153144, + "tokens_seen": 1036255232 + }, + { + "epoch": 0.31, + "learning_rate": 0.0005543510531594785, + "loss": 1.4073, + "theoretical_loss": 3.6653312113034993, + "tokens_seen": 1036320768 + }, + { + "epoch": 0.31, + "learning_rate": 0.0005543350050150451, + "loss": 1.3677, + "theoretical_loss": 3.665309468651737, + "tokens_seen": 1036386304 + }, + { + "epoch": 0.31, + "learning_rate": 0.0005543189568706119, + "loss": 1.4921, + "theoretical_loss": 3.6652877277597744, + "tokens_seen": 1036451840 + }, + { + "epoch": 0.31, + "learning_rate": 0.0005543029087261785, + "loss": 1.5094, + "theoretical_loss": 3.6652659886273575, + "tokens_seen": 1036517376 + }, + { + "epoch": 0.31, + "learning_rate": 0.0005542868605817453, + "loss": 1.3957, + "theoretical_loss": 3.6652442512542325, + "tokens_seen": 1036582912 + }, + { + "epoch": 0.31, + "learning_rate": 0.0005542708124373119, + "loss": 1.4615, + "theoretical_loss": 3.6652225156401457, + "tokens_seen": 1036648448 + }, + { + "epoch": 0.31, + "learning_rate": 0.0005542547642928787, + "loss": 1.5203, + "theoretical_loss": 3.6652007817848444, + "tokens_seen": 1036713984 + }, + { + "epoch": 0.31, + "learning_rate": 0.0005542387161484453, + "loss": 1.423, + "theoretical_loss": 3.6651790496880747, + "tokens_seen": 1036779520 + }, + { + "epoch": 0.31, + "learning_rate": 0.0005542226680040121, + "loss": 1.5235, + "theoretical_loss": 3.6651573193495826, + "tokens_seen": 1036845056 + }, + { + "epoch": 0.31, + "learning_rate": 0.0005542066198595787, + "loss": 1.4056, + "theoretical_loss": 3.665135590769115, + "tokens_seen": 1036910592 + }, + { + "epoch": 0.31, + "learning_rate": 0.0005541905717151455, + "loss": 1.4877, + "theoretical_loss": 3.665113863946419, + "tokens_seen": 1036976128 + }, + { + "epoch": 0.31, + "learning_rate": 0.0005541745235707121, + "loss": 1.4725, + "theoretical_loss": 3.6650921388812416, + "tokens_seen": 1037041664 + }, + { + "epoch": 0.31, + "learning_rate": 0.0005541584754262789, + "loss": 1.3965, + "theoretical_loss": 3.6650704155733287, + "tokens_seen": 1037107200 + }, + { + "epoch": 0.31, + "learning_rate": 0.0005541424272818455, + "loss": 1.538, + "theoretical_loss": 3.6650486940224267, + "tokens_seen": 1037172736 + }, + { + "epoch": 0.31, + "learning_rate": 0.0005541263791374123, + "loss": 1.5426, + "theoretical_loss": 3.6650269742282844, + "tokens_seen": 1037238272 + }, + { + "epoch": 0.31, + "learning_rate": 0.0005541103309929789, + "loss": 1.3639, + "theoretical_loss": 3.6650052561906468, + "tokens_seen": 1037303808 + }, + { + "epoch": 0.31, + "learning_rate": 0.0005540942828485457, + "loss": 1.5093, + "theoretical_loss": 3.664983539909262, + "tokens_seen": 1037369344 + }, + { + "epoch": 0.31, + "learning_rate": 0.0005540782347041123, + "loss": 1.5, + "theoretical_loss": 3.6649618253838767, + "tokens_seen": 1037434880 + }, + { + "epoch": 0.31, + "learning_rate": 0.0005540621865596791, + "loss": 1.2772, + "theoretical_loss": 3.664940112614238, + "tokens_seen": 1037500416 + }, + { + "epoch": 0.31, + "learning_rate": 0.0005540461384152457, + "loss": 1.4197, + "theoretical_loss": 3.664918401600093, + "tokens_seen": 1037565952 + }, + { + "epoch": 0.31, + "learning_rate": 0.0005540300902708125, + "loss": 1.3627, + "theoretical_loss": 3.664896692341189, + "tokens_seen": 1037631488 + }, + { + "epoch": 0.31, + "learning_rate": 0.0005540140421263791, + "loss": 1.4616, + "theoretical_loss": 3.6648749848372733, + "tokens_seen": 1037697024 + }, + { + "epoch": 0.31, + "learning_rate": 0.0005539979939819459, + "loss": 1.429, + "theoretical_loss": 3.664853279088093, + "tokens_seen": 1037762560 + }, + { + "epoch": 0.31, + "learning_rate": 0.0005539819458375125, + "loss": 1.3865, + "theoretical_loss": 3.6648315750933955, + "tokens_seen": 1037828096 + }, + { + "epoch": 0.31, + "learning_rate": 0.0005539658976930793, + "loss": 1.3835, + "theoretical_loss": 3.664809872852928, + "tokens_seen": 1037893632 + }, + { + "epoch": 0.31, + "learning_rate": 0.0005539498495486459, + "loss": 1.4401, + "theoretical_loss": 3.6647881723664386, + "tokens_seen": 1037959168 + }, + { + "epoch": 0.31, + "learning_rate": 0.0005539338014042127, + "loss": 1.3775, + "theoretical_loss": 3.6647664736336747, + "tokens_seen": 1038024704 + }, + { + "epoch": 0.31, + "learning_rate": 0.0005539177532597793, + "loss": 1.4328, + "theoretical_loss": 3.6647447766543833, + "tokens_seen": 1038090240 + }, + { + "epoch": 0.31, + "learning_rate": 0.0005539017051153461, + "loss": 1.4198, + "theoretical_loss": 3.664723081428312, + "tokens_seen": 1038155776 + }, + { + "epoch": 0.31, + "learning_rate": 0.0005538856569709127, + "loss": 1.3353, + "theoretical_loss": 3.6647013879552093, + "tokens_seen": 1038221312 + }, + { + "epoch": 0.31, + "learning_rate": 0.0005538696088264795, + "loss": 1.486, + "theoretical_loss": 3.664679696234822, + "tokens_seen": 1038286848 + }, + { + "epoch": 0.31, + "learning_rate": 0.0005538535606820461, + "loss": 1.4454, + "theoretical_loss": 3.6646580062668983, + "tokens_seen": 1038352384 + }, + { + "epoch": 0.31, + "learning_rate": 0.0005538375125376129, + "loss": 1.4286, + "theoretical_loss": 3.664636318051186, + "tokens_seen": 1038417920 + }, + { + "epoch": 0.31, + "learning_rate": 0.0005538214643931795, + "loss": 1.3427, + "theoretical_loss": 3.6646146315874333, + "tokens_seen": 1038483456 + }, + { + "epoch": 0.31, + "learning_rate": 0.0005538054162487463, + "loss": 1.3045, + "theoretical_loss": 3.6645929468753877, + "tokens_seen": 1038548992 + }, + { + "epoch": 0.31, + "learning_rate": 0.0005537893681043129, + "loss": 1.5251, + "theoretical_loss": 3.664571263914797, + "tokens_seen": 1038614528 + }, + { + "epoch": 0.31, + "learning_rate": 0.0005537733199598797, + "loss": 1.4474, + "theoretical_loss": 3.66454958270541, + "tokens_seen": 1038680064 + }, + { + "epoch": 0.31, + "objective/train/docs_used": 381080, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.2670519351959229, + "objective/train/theoretical_loss": 3.6645279032469746, + "objective/train/tokens_used": 1059205600, + "theoretical_loss": 3.6645279032469746, + "tokens_seen": 1038745600 + }, + { + "epoch": 0.31, + "learning_rate": 0.0005537733199598797, + "loss": 1.3684, + "theoretical_loss": 3.6645279032469746, + "tokens_seen": 1038745600 + }, + { + "epoch": 0.31, + "learning_rate": 0.0005537572718154463, + "loss": 1.446, + "theoretical_loss": 3.664506225539238, + "tokens_seen": 1038811136 + }, + { + "epoch": 0.31, + "learning_rate": 0.0005537412236710131, + "loss": 1.4462, + "theoretical_loss": 3.664484549581949, + "tokens_seen": 1038876672 + }, + { + "epoch": 0.31, + "learning_rate": 0.0005537251755265798, + "loss": 1.4892, + "theoretical_loss": 3.6644628753748565, + "tokens_seen": 1038942208 + }, + { + "epoch": 0.31, + "learning_rate": 0.0005537091273821465, + "loss": 1.4123, + "theoretical_loss": 3.6644412029177076, + "tokens_seen": 1039007744 + }, + { + "epoch": 0.31, + "learning_rate": 0.0005536930792377132, + "loss": 1.3607, + "theoretical_loss": 3.6644195322102515, + "tokens_seen": 1039073280 + }, + { + "epoch": 0.31, + "learning_rate": 0.0005536770310932799, + "loss": 1.4371, + "theoretical_loss": 3.664397863252236, + "tokens_seen": 1039138816 + }, + { + "epoch": 0.31, + "learning_rate": 0.0005536609829488466, + "loss": 1.3855, + "theoretical_loss": 3.6643761960434107, + "tokens_seen": 1039204352 + }, + { + "epoch": 0.31, + "learning_rate": 0.0005536449348044133, + "loss": 1.4565, + "theoretical_loss": 3.6643545305835223, + "tokens_seen": 1039269888 + }, + { + "epoch": 0.31, + "learning_rate": 0.00055362888665998, + "loss": 1.468, + "theoretical_loss": 3.664332866872321, + "tokens_seen": 1039335424 + }, + { + "epoch": 0.31, + "learning_rate": 0.0005536128385155467, + "loss": 1.3989, + "theoretical_loss": 3.6643112049095543, + "tokens_seen": 1039400960 + }, + { + "epoch": 0.31, + "learning_rate": 0.0005535967903711134, + "loss": 1.3997, + "theoretical_loss": 3.6642895446949715, + "tokens_seen": 1039466496 + }, + { + "epoch": 0.32, + "learning_rate": 0.0005535807422266801, + "loss": 1.4853, + "theoretical_loss": 3.664267886228321, + "tokens_seen": 1039532032 + }, + { + "epoch": 0.32, + "learning_rate": 0.0005535646940822469, + "loss": 1.4623, + "theoretical_loss": 3.6642462295093523, + "tokens_seen": 1039597568 + }, + { + "epoch": 0.32, + "learning_rate": 0.0005535486459378135, + "loss": 1.3441, + "theoretical_loss": 3.664224574537813, + "tokens_seen": 1039663104 + }, + { + "epoch": 0.32, + "learning_rate": 0.0005535325977933803, + "loss": 1.3715, + "theoretical_loss": 3.6642029213134526, + "tokens_seen": 1039728640 + }, + { + "epoch": 0.32, + "learning_rate": 0.0005535165496489469, + "loss": 1.4358, + "theoretical_loss": 3.66418126983602, + "tokens_seen": 1039794176 + }, + { + "epoch": 0.32, + "learning_rate": 0.0005535005015045137, + "loss": 1.4028, + "theoretical_loss": 3.6641596201052637, + "tokens_seen": 1039859712 + }, + { + "epoch": 0.32, + "learning_rate": 0.0005534844533600803, + "loss": 1.3924, + "theoretical_loss": 3.6641379721209337, + "tokens_seen": 1039925248 + }, + { + "epoch": 0.32, + "learning_rate": 0.0005534684052156471, + "loss": 1.3963, + "theoretical_loss": 3.6641163258827785, + "tokens_seen": 1039990784 + }, + { + "epoch": 0.32, + "learning_rate": 0.0005534523570712137, + "loss": 1.446, + "theoretical_loss": 3.664094681390547, + "tokens_seen": 1040056320 + }, + { + "epoch": 0.32, + "learning_rate": 0.0005534363089267803, + "loss": 1.4254, + "theoretical_loss": 3.6640730386439895, + "tokens_seen": 1040121856 + }, + { + "epoch": 0.32, + "learning_rate": 0.0005534202607823471, + "loss": 1.4031, + "theoretical_loss": 3.6640513976428535, + "tokens_seen": 1040187392 + }, + { + "epoch": 0.32, + "learning_rate": 0.0005534042126379137, + "loss": 1.4869, + "theoretical_loss": 3.6640297583868895, + "tokens_seen": 1040252928 + }, + { + "epoch": 0.32, + "learning_rate": 0.0005533881644934805, + "loss": 1.3299, + "theoretical_loss": 3.664008120875846, + "tokens_seen": 1040318464 + }, + { + "epoch": 0.32, + "learning_rate": 0.0005533721163490472, + "loss": 1.5566, + "theoretical_loss": 3.663986485109474, + "tokens_seen": 1040384000 + }, + { + "epoch": 0.32, + "learning_rate": 0.0005533560682046139, + "loss": 1.4828, + "theoretical_loss": 3.6639648510875205, + "tokens_seen": 1040449536 + }, + { + "epoch": 0.32, + "learning_rate": 0.0005533400200601806, + "loss": 1.4131, + "theoretical_loss": 3.663943218809737, + "tokens_seen": 1040515072 + }, + { + "epoch": 0.32, + "learning_rate": 0.0005533239719157473, + "loss": 1.3707, + "theoretical_loss": 3.6639215882758727, + "tokens_seen": 1040580608 + }, + { + "epoch": 0.32, + "learning_rate": 0.000553307923771314, + "loss": 1.4451, + "theoretical_loss": 3.663899959485676, + "tokens_seen": 1040646144 + }, + { + "epoch": 0.32, + "learning_rate": 0.0005532918756268806, + "loss": 1.4832, + "theoretical_loss": 3.6638783324388977, + "tokens_seen": 1040711680 + }, + { + "epoch": 0.32, + "learning_rate": 0.0005532758274824474, + "loss": 1.47, + "theoretical_loss": 3.663856707135287, + "tokens_seen": 1040777216 + }, + { + "epoch": 0.32, + "learning_rate": 0.000553259779338014, + "loss": 1.5086, + "theoretical_loss": 3.6638350835745945, + "tokens_seen": 1040842752 + }, + { + "epoch": 0.32, + "learning_rate": 0.0005532437311935808, + "loss": 1.5501, + "theoretical_loss": 3.663813461756569, + "tokens_seen": 1040908288 + }, + { + "epoch": 0.32, + "learning_rate": 0.0005532276830491474, + "loss": 1.3737, + "theoretical_loss": 3.6637918416809603, + "tokens_seen": 1040973824 + }, + { + "epoch": 0.32, + "learning_rate": 0.0005532116349047142, + "loss": 1.4258, + "theoretical_loss": 3.663770223347519, + "tokens_seen": 1041039360 + }, + { + "epoch": 0.32, + "learning_rate": 0.0005531955867602808, + "loss": 1.4333, + "theoretical_loss": 3.663748606755995, + "tokens_seen": 1041104896 + }, + { + "epoch": 0.32, + "learning_rate": 0.0005531795386158476, + "loss": 1.4556, + "theoretical_loss": 3.6637269919061373, + "tokens_seen": 1041170432 + }, + { + "epoch": 0.32, + "learning_rate": 0.0005531634904714142, + "loss": 1.3916, + "theoretical_loss": 3.6637053787976974, + "tokens_seen": 1041235968 + }, + { + "epoch": 0.32, + "learning_rate": 0.000553147442326981, + "loss": 1.4096, + "theoretical_loss": 3.663683767430424, + "tokens_seen": 1041301504 + }, + { + "epoch": 0.32, + "learning_rate": 0.0005531313941825476, + "loss": 1.3558, + "theoretical_loss": 3.6636621578040685, + "tokens_seen": 1041367040 + }, + { + "epoch": 0.32, + "learning_rate": 0.0005531153460381144, + "loss": 1.3689, + "theoretical_loss": 3.66364054991838, + "tokens_seen": 1041432576 + }, + { + "epoch": 0.32, + "learning_rate": 0.000553099297893681, + "loss": 1.3648, + "theoretical_loss": 3.66361894377311, + "tokens_seen": 1041498112 + }, + { + "epoch": 0.32, + "learning_rate": 0.0005530832497492478, + "loss": 1.3872, + "theoretical_loss": 3.6635973393680077, + "tokens_seen": 1041563648 + }, + { + "epoch": 0.32, + "learning_rate": 0.0005530672016048144, + "loss": 1.3306, + "theoretical_loss": 3.6635757367028248, + "tokens_seen": 1041629184 + }, + { + "epoch": 0.32, + "learning_rate": 0.0005530511534603812, + "loss": 1.3833, + "theoretical_loss": 3.66355413577731, + "tokens_seen": 1041694720 + }, + { + "epoch": 0.32, + "learning_rate": 0.0005530351053159478, + "loss": 1.51, + "theoretical_loss": 3.6635325365912146, + "tokens_seen": 1041760256 + }, + { + "epoch": 0.32, + "learning_rate": 0.0005530190571715146, + "loss": 1.3862, + "theoretical_loss": 3.663510939144289, + "tokens_seen": 1041825792 + }, + { + "epoch": 0.32, + "learning_rate": 0.0005530030090270812, + "loss": 1.4355, + "theoretical_loss": 3.6634893434362845, + "tokens_seen": 1041891328 + }, + { + "epoch": 0.32, + "learning_rate": 0.000552986960882648, + "loss": 1.3736, + "theoretical_loss": 3.6634677494669505, + "tokens_seen": 1041956864 + }, + { + "epoch": 0.32, + "objective/train/docs_used": 382133, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.398244857788086, + "objective/train/theoretical_loss": 3.663446157236039, + "objective/train/tokens_used": 1062482400, + "theoretical_loss": 3.663446157236039, + "tokens_seen": 1042022400 + }, + { + "epoch": 0.32, + "learning_rate": 0.0005529709127382146, + "loss": 1.3829, + "theoretical_loss": 3.663446157236039, + "tokens_seen": 1042022400 + }, + { + "epoch": 0.32, + "learning_rate": 0.0005529548645937814, + "loss": 1.4891, + "theoretical_loss": 3.6634245667432994, + "tokens_seen": 1042087936 + }, + { + "epoch": 0.32, + "learning_rate": 0.000552938816449348, + "loss": 1.4448, + "theoretical_loss": 3.6634029779884836, + "tokens_seen": 1042153472 + }, + { + "epoch": 0.32, + "learning_rate": 0.0005529227683049148, + "loss": 1.5002, + "theoretical_loss": 3.6633813909713417, + "tokens_seen": 1042219008 + }, + { + "epoch": 0.32, + "learning_rate": 0.0005529067201604814, + "loss": 1.3981, + "theoretical_loss": 3.6633598056916243, + "tokens_seen": 1042284544 + }, + { + "epoch": 0.32, + "learning_rate": 0.0005528906720160482, + "loss": 1.4399, + "theoretical_loss": 3.6633382221490836, + "tokens_seen": 1042350080 + }, + { + "epoch": 0.32, + "learning_rate": 0.0005528746238716148, + "loss": 1.4232, + "theoretical_loss": 3.66331664034347, + "tokens_seen": 1042415616 + }, + { + "epoch": 0.32, + "learning_rate": 0.0005528585757271816, + "loss": 1.3422, + "theoretical_loss": 3.6632950602745336, + "tokens_seen": 1042481152 + }, + { + "epoch": 0.32, + "learning_rate": 0.0005528425275827482, + "loss": 1.471, + "theoretical_loss": 3.663273481942026, + "tokens_seen": 1042546688 + }, + { + "epoch": 0.32, + "learning_rate": 0.000552826479438315, + "loss": 1.442, + "theoretical_loss": 3.6632519053456996, + "tokens_seen": 1042612224 + }, + { + "epoch": 0.32, + "learning_rate": 0.0005528104312938816, + "loss": 1.4206, + "theoretical_loss": 3.6632303304853044, + "tokens_seen": 1042677760 + }, + { + "epoch": 0.32, + "learning_rate": 0.0005527943831494484, + "loss": 1.4732, + "theoretical_loss": 3.663208757360591, + "tokens_seen": 1042743296 + }, + { + "epoch": 0.32, + "learning_rate": 0.000552778335005015, + "loss": 1.3961, + "theoretical_loss": 3.6631871859713123, + "tokens_seen": 1042808832 + }, + { + "epoch": 0.32, + "learning_rate": 0.0005527622868605818, + "loss": 1.3601, + "theoretical_loss": 3.663165616317219, + "tokens_seen": 1042874368 + }, + { + "epoch": 0.32, + "learning_rate": 0.0005527462387161484, + "loss": 1.3304, + "theoretical_loss": 3.6631440483980615, + "tokens_seen": 1042939904 + }, + { + "epoch": 0.32, + "learning_rate": 0.0005527301905717152, + "loss": 1.3509, + "theoretical_loss": 3.663122482213593, + "tokens_seen": 1043005440 + }, + { + "epoch": 0.32, + "learning_rate": 0.0005527141424272818, + "loss": 1.4095, + "theoretical_loss": 3.6631009177635634, + "tokens_seen": 1043070976 + }, + { + "epoch": 0.32, + "learning_rate": 0.0005526980942828486, + "loss": 1.3923, + "theoretical_loss": 3.663079355047725, + "tokens_seen": 1043136512 + }, + { + "epoch": 0.32, + "learning_rate": 0.0005526820461384152, + "loss": 1.4392, + "theoretical_loss": 3.6630577940658298, + "tokens_seen": 1043202048 + }, + { + "epoch": 0.32, + "learning_rate": 0.000552665997993982, + "loss": 1.376, + "theoretical_loss": 3.6630362348176284, + "tokens_seen": 1043267584 + }, + { + "epoch": 0.32, + "learning_rate": 0.0005526499498495486, + "loss": 1.5489, + "theoretical_loss": 3.663014677302874, + "tokens_seen": 1043333120 + }, + { + "epoch": 0.32, + "learning_rate": 0.0005526339017051154, + "loss": 1.5206, + "theoretical_loss": 3.6629931215213167, + "tokens_seen": 1043398656 + }, + { + "epoch": 0.32, + "learning_rate": 0.000552617853560682, + "loss": 1.3264, + "theoretical_loss": 3.6629715674727095, + "tokens_seen": 1043464192 + }, + { + "epoch": 0.32, + "learning_rate": 0.0005526018054162488, + "loss": 1.5284, + "theoretical_loss": 3.6629500151568033, + "tokens_seen": 1043529728 + }, + { + "epoch": 0.32, + "learning_rate": 0.0005525857572718154, + "loss": 1.321, + "theoretical_loss": 3.6629284645733504, + "tokens_seen": 1043595264 + }, + { + "epoch": 0.32, + "learning_rate": 0.0005525697091273822, + "loss": 1.4687, + "theoretical_loss": 3.662906915722103, + "tokens_seen": 1043660800 + }, + { + "epoch": 0.32, + "learning_rate": 0.0005525536609829488, + "loss": 1.5563, + "theoretical_loss": 3.6628853686028133, + "tokens_seen": 1043726336 + }, + { + "epoch": 0.32, + "learning_rate": 0.0005525376128385156, + "loss": 1.3988, + "theoretical_loss": 3.662863823215233, + "tokens_seen": 1043791872 + }, + { + "epoch": 0.32, + "learning_rate": 0.0005525215646940823, + "loss": 1.44, + "theoretical_loss": 3.662842279559114, + "tokens_seen": 1043857408 + }, + { + "epoch": 0.32, + "learning_rate": 0.000552505516549649, + "loss": 1.4287, + "theoretical_loss": 3.6628207376342092, + "tokens_seen": 1043922944 + }, + { + "epoch": 0.32, + "learning_rate": 0.0005524894684052157, + "loss": 1.4801, + "theoretical_loss": 3.66279919744027, + "tokens_seen": 1043988480 + }, + { + "epoch": 0.32, + "learning_rate": 0.0005524734202607824, + "loss": 1.4492, + "theoretical_loss": 3.6627776589770487, + "tokens_seen": 1044054016 + }, + { + "epoch": 0.32, + "learning_rate": 0.0005524573721163491, + "loss": 1.3885, + "theoretical_loss": 3.662756122244298, + "tokens_seen": 1044119552 + }, + { + "epoch": 0.32, + "learning_rate": 0.0005524413239719158, + "loss": 1.5263, + "theoretical_loss": 3.6627345872417703, + "tokens_seen": 1044185088 + }, + { + "epoch": 0.32, + "learning_rate": 0.0005524252758274825, + "loss": 1.584, + "theoretical_loss": 3.662713053969218, + "tokens_seen": 1044250624 + }, + { + "epoch": 0.32, + "learning_rate": 0.0005524092276830492, + "loss": 1.3924, + "theoretical_loss": 3.6626915224263934, + "tokens_seen": 1044316160 + }, + { + "epoch": 0.32, + "learning_rate": 0.0005523931795386159, + "loss": 1.4901, + "theoretical_loss": 3.662669992613049, + "tokens_seen": 1044381696 + }, + { + "epoch": 0.32, + "learning_rate": 0.0005523771313941826, + "loss": 1.4792, + "theoretical_loss": 3.6626484645289374, + "tokens_seen": 1044447232 + }, + { + "epoch": 0.32, + "learning_rate": 0.0005523610832497493, + "loss": 1.425, + "theoretical_loss": 3.662626938173811, + "tokens_seen": 1044512768 + }, + { + "epoch": 0.32, + "learning_rate": 0.000552345035105316, + "loss": 1.4083, + "theoretical_loss": 3.662605413547423, + "tokens_seen": 1044578304 + }, + { + "epoch": 0.32, + "learning_rate": 0.0005523289869608827, + "loss": 1.2697, + "theoretical_loss": 3.6625838906495263, + "tokens_seen": 1044643840 + }, + { + "epoch": 0.32, + "learning_rate": 0.0005523129388164494, + "loss": 1.3675, + "theoretical_loss": 3.6625623694798723, + "tokens_seen": 1044709376 + }, + { + "epoch": 0.32, + "learning_rate": 0.0005522968906720161, + "loss": 1.599, + "theoretical_loss": 3.6625408500382153, + "tokens_seen": 1044774912 + }, + { + "epoch": 0.32, + "learning_rate": 0.0005522808425275828, + "loss": 1.495, + "theoretical_loss": 3.6625193323243073, + "tokens_seen": 1044840448 + }, + { + "epoch": 0.32, + "learning_rate": 0.0005522647943831495, + "loss": 1.4576, + "theoretical_loss": 3.6624978163379014, + "tokens_seen": 1044905984 + }, + { + "epoch": 0.32, + "learning_rate": 0.0005522487462387162, + "loss": 1.3434, + "theoretical_loss": 3.662476302078751, + "tokens_seen": 1044971520 + }, + { + "epoch": 0.32, + "learning_rate": 0.000552232698094283, + "loss": 1.3498, + "theoretical_loss": 3.6624547895466093, + "tokens_seen": 1045037056 + }, + { + "epoch": 0.32, + "learning_rate": 0.0005522166499498496, + "loss": 1.4483, + "theoretical_loss": 3.6624332787412284, + "tokens_seen": 1045102592 + }, + { + "epoch": 0.32, + "learning_rate": 0.0005522006018054164, + "loss": 1.4236, + "theoretical_loss": 3.662411769662362, + "tokens_seen": 1045168128 + }, + { + "epoch": 0.32, + "learning_rate": 0.000552184553660983, + "loss": 1.4912, + "theoretical_loss": 3.6623902623097635, + "tokens_seen": 1045233664 + }, + { + "epoch": 0.32, + "objective/train/docs_used": 383342, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.4522275924682617, + "objective/train/theoretical_loss": 3.6623687566831857, + "objective/train/tokens_used": 1065759200, + "theoretical_loss": 3.6623687566831857, + "tokens_seen": 1045299200 + }, + { + "epoch": 0.32, + "learning_rate": 0.0005521685055165498, + "loss": 1.4765, + "theoretical_loss": 3.6623687566831857, + "tokens_seen": 1045299200 + }, + { + "epoch": 0.32, + "learning_rate": 0.0005521524573721164, + "loss": 1.4742, + "theoretical_loss": 3.662347252782382, + "tokens_seen": 1045364736 + }, + { + "epoch": 0.32, + "learning_rate": 0.0005521364092276832, + "loss": 1.3851, + "theoretical_loss": 3.6623257506071054, + "tokens_seen": 1045430272 + }, + { + "epoch": 0.32, + "learning_rate": 0.0005521203610832498, + "loss": 1.4385, + "theoretical_loss": 3.66230425015711, + "tokens_seen": 1045495808 + }, + { + "epoch": 0.32, + "learning_rate": 0.0005521043129388166, + "loss": 1.3581, + "theoretical_loss": 3.662282751432149, + "tokens_seen": 1045561344 + }, + { + "epoch": 0.32, + "learning_rate": 0.0005520882647943832, + "loss": 1.4888, + "theoretical_loss": 3.662261254431976, + "tokens_seen": 1045626880 + }, + { + "epoch": 0.32, + "learning_rate": 0.00055207221664995, + "loss": 1.4683, + "theoretical_loss": 3.6622397591563436, + "tokens_seen": 1045692416 + }, + { + "epoch": 0.32, + "learning_rate": 0.0005520561685055166, + "loss": 1.3988, + "theoretical_loss": 3.6622182656050066, + "tokens_seen": 1045757952 + }, + { + "epoch": 0.32, + "learning_rate": 0.0005520401203610832, + "loss": 1.4559, + "theoretical_loss": 3.662196773777718, + "tokens_seen": 1045823488 + }, + { + "epoch": 0.32, + "learning_rate": 0.00055202407221665, + "loss": 1.4246, + "theoretical_loss": 3.6621752836742316, + "tokens_seen": 1045889024 + }, + { + "epoch": 0.32, + "learning_rate": 0.0005520080240722167, + "loss": 1.4452, + "theoretical_loss": 3.662153795294301, + "tokens_seen": 1045954560 + }, + { + "epoch": 0.32, + "learning_rate": 0.0005519919759277834, + "loss": 1.4059, + "theoretical_loss": 3.6621323086376805, + "tokens_seen": 1046020096 + }, + { + "epoch": 0.32, + "learning_rate": 0.00055197592778335, + "loss": 1.4436, + "theoretical_loss": 3.6621108237041233, + "tokens_seen": 1046085632 + }, + { + "epoch": 0.32, + "learning_rate": 0.0005519598796389168, + "loss": 1.3094, + "theoretical_loss": 3.6620893404933836, + "tokens_seen": 1046151168 + }, + { + "epoch": 0.32, + "learning_rate": 0.0005519438314944835, + "loss": 1.505, + "theoretical_loss": 3.6620678590052154, + "tokens_seen": 1046216704 + }, + { + "epoch": 0.32, + "learning_rate": 0.0005519277833500502, + "loss": 1.3751, + "theoretical_loss": 3.6620463792393725, + "tokens_seen": 1046282240 + }, + { + "epoch": 0.32, + "learning_rate": 0.0005519117352056169, + "loss": 1.4216, + "theoretical_loss": 3.6620249011956094, + "tokens_seen": 1046347776 + }, + { + "epoch": 0.32, + "learning_rate": 0.0005518956870611835, + "loss": 1.4412, + "theoretical_loss": 3.662003424873679, + "tokens_seen": 1046413312 + }, + { + "epoch": 0.32, + "learning_rate": 0.0005518796389167503, + "loss": 1.4846, + "theoretical_loss": 3.6619819502733373, + "tokens_seen": 1046478848 + }, + { + "epoch": 0.32, + "learning_rate": 0.0005518635907723169, + "loss": 1.3339, + "theoretical_loss": 3.6619604773943366, + "tokens_seen": 1046544384 + }, + { + "epoch": 0.32, + "learning_rate": 0.0005518475426278837, + "loss": 1.3449, + "theoretical_loss": 3.661939006236432, + "tokens_seen": 1046609920 + }, + { + "epoch": 0.32, + "learning_rate": 0.0005518314944834503, + "loss": 1.4959, + "theoretical_loss": 3.661917536799378, + "tokens_seen": 1046675456 + }, + { + "epoch": 0.32, + "learning_rate": 0.0005518154463390171, + "loss": 1.4702, + "theoretical_loss": 3.661896069082929, + "tokens_seen": 1046740992 + }, + { + "epoch": 0.32, + "learning_rate": 0.0005517993981945837, + "loss": 1.5364, + "theoretical_loss": 3.661874603086839, + "tokens_seen": 1046806528 + }, + { + "epoch": 0.32, + "learning_rate": 0.0005517833500501505, + "loss": 1.416, + "theoretical_loss": 3.661853138810863, + "tokens_seen": 1046872064 + }, + { + "epoch": 0.32, + "learning_rate": 0.0005517673019057171, + "loss": 1.3791, + "theoretical_loss": 3.6618316762547543, + "tokens_seen": 1046937600 + }, + { + "epoch": 0.32, + "learning_rate": 0.0005517512537612839, + "loss": 1.5024, + "theoretical_loss": 3.6618102154182686, + "tokens_seen": 1047003136 + }, + { + "epoch": 0.32, + "learning_rate": 0.0005517352056168505, + "loss": 1.389, + "theoretical_loss": 3.66178875630116, + "tokens_seen": 1047068672 + }, + { + "epoch": 0.32, + "learning_rate": 0.0005517191574724173, + "loss": 1.4066, + "theoretical_loss": 3.6617672989031833, + "tokens_seen": 1047134208 + }, + { + "epoch": 0.32, + "learning_rate": 0.0005517031093279839, + "loss": 1.5193, + "theoretical_loss": 3.661745843224093, + "tokens_seen": 1047199744 + }, + { + "epoch": 0.32, + "learning_rate": 0.0005516870611835507, + "loss": 1.3258, + "theoretical_loss": 3.6617243892636444, + "tokens_seen": 1047265280 + }, + { + "epoch": 0.32, + "learning_rate": 0.0005516710130391173, + "loss": 1.3676, + "theoretical_loss": 3.6617029370215914, + "tokens_seen": 1047330816 + }, + { + "epoch": 0.32, + "learning_rate": 0.0005516549648946841, + "loss": 1.3679, + "theoretical_loss": 3.66168148649769, + "tokens_seen": 1047396352 + }, + { + "epoch": 0.32, + "learning_rate": 0.0005516389167502507, + "loss": 1.4617, + "theoretical_loss": 3.6616600376916937, + "tokens_seen": 1047461888 + }, + { + "epoch": 0.32, + "learning_rate": 0.0005516228686058175, + "loss": 1.4156, + "theoretical_loss": 3.6616385906033586, + "tokens_seen": 1047527424 + }, + { + "epoch": 0.32, + "learning_rate": 0.0005516068204613841, + "loss": 1.3859, + "theoretical_loss": 3.661617145232439, + "tokens_seen": 1047592960 + }, + { + "epoch": 0.32, + "learning_rate": 0.0005515907723169509, + "loss": 1.563, + "theoretical_loss": 3.6615957015786904, + "tokens_seen": 1047658496 + }, + { + "epoch": 0.32, + "learning_rate": 0.0005515747241725175, + "loss": 1.4455, + "theoretical_loss": 3.661574259641868, + "tokens_seen": 1047724032 + }, + { + "epoch": 0.32, + "learning_rate": 0.0005515586760280843, + "loss": 1.4039, + "theoretical_loss": 3.6615528194217264, + "tokens_seen": 1047789568 + }, + { + "epoch": 0.32, + "learning_rate": 0.0005515426278836509, + "loss": 1.3702, + "theoretical_loss": 3.661531380918021, + "tokens_seen": 1047855104 + }, + { + "epoch": 0.32, + "learning_rate": 0.0005515265797392177, + "loss": 1.3918, + "theoretical_loss": 3.6615099441305077, + "tokens_seen": 1047920640 + }, + { + "epoch": 0.32, + "learning_rate": 0.0005515105315947843, + "loss": 1.2801, + "theoretical_loss": 3.6614885090589406, + "tokens_seen": 1047986176 + }, + { + "epoch": 0.32, + "learning_rate": 0.0005514944834503511, + "loss": 1.5194, + "theoretical_loss": 3.661467075703076, + "tokens_seen": 1048051712 + }, + { + "epoch": 0.32, + "learning_rate": 0.0005514784353059177, + "loss": 1.4109, + "theoretical_loss": 3.6614456440626686, + "tokens_seen": 1048117248 + }, + { + "epoch": 0.32, + "learning_rate": 0.0005514623871614845, + "loss": 1.4176, + "theoretical_loss": 3.6614242141374747, + "tokens_seen": 1048182784 + }, + { + "epoch": 0.32, + "learning_rate": 0.0005514463390170511, + "loss": 1.4935, + "theoretical_loss": 3.6614027859272493, + "tokens_seen": 1048248320 + }, + { + "epoch": 0.32, + "learning_rate": 0.0005514302908726179, + "loss": 1.3456, + "theoretical_loss": 3.6613813594317475, + "tokens_seen": 1048313856 + }, + { + "epoch": 0.32, + "learning_rate": 0.0005514142427281845, + "loss": 1.4065, + "theoretical_loss": 3.6613599346507257, + "tokens_seen": 1048379392 + }, + { + "epoch": 0.32, + "learning_rate": 0.0005513981945837513, + "loss": 1.3773, + "theoretical_loss": 3.6613385115839394, + "tokens_seen": 1048444928 + }, + { + "epoch": 0.32, + "learning_rate": 0.0005513821464393179, + "loss": 1.4213, + "theoretical_loss": 3.661317090231144, + "tokens_seen": 1048510464 + }, + { + "epoch": 0.32, + "objective/train/docs_used": 384505, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.528778314590454, + "objective/train/theoretical_loss": 3.661295670592095, + "objective/train/tokens_used": 1069036000, + "theoretical_loss": 3.661295670592095, + "tokens_seen": 1048576000 + }, + { + "epoch": 0.32, + "learning_rate": 0.0005513660982948847, + "loss": 1.4164, + "theoretical_loss": 3.661295670592095, + "tokens_seen": 1048576000 + }, + { + "epoch": 0.32, + "learning_rate": 0.0005513500501504514, + "loss": 1.3314, + "theoretical_loss": 3.661274252666549, + "tokens_seen": 1048641536 + }, + { + "epoch": 0.32, + "learning_rate": 0.0005513340020060181, + "loss": 1.4074, + "theoretical_loss": 3.6612528364542616, + "tokens_seen": 1048707072 + }, + { + "epoch": 0.32, + "learning_rate": 0.0005513179538615848, + "loss": 1.4961, + "theoretical_loss": 3.661231421954988, + "tokens_seen": 1048772608 + }, + { + "epoch": 0.32, + "learning_rate": 0.0005513019057171515, + "loss": 1.4901, + "theoretical_loss": 3.661210009168485, + "tokens_seen": 1048838144 + }, + { + "epoch": 0.32, + "learning_rate": 0.0005512858575727182, + "loss": 1.3175, + "theoretical_loss": 3.6611885980945083, + "tokens_seen": 1048903680 + }, + { + "epoch": 0.32, + "learning_rate": 0.0005512698094282849, + "loss": 1.4701, + "theoretical_loss": 3.661167188732814, + "tokens_seen": 1048969216 + }, + { + "epoch": 0.32, + "learning_rate": 0.0005512537612838516, + "loss": 1.382, + "theoretical_loss": 3.661145781083158, + "tokens_seen": 1049034752 + }, + { + "epoch": 0.32, + "learning_rate": 0.0005512377131394183, + "loss": 1.414, + "theoretical_loss": 3.6611243751452966, + "tokens_seen": 1049100288 + }, + { + "epoch": 0.32, + "learning_rate": 0.000551221664994985, + "loss": 1.4047, + "theoretical_loss": 3.661102970918986, + "tokens_seen": 1049165824 + }, + { + "epoch": 0.32, + "learning_rate": 0.0005512056168505517, + "loss": 1.4009, + "theoretical_loss": 3.661081568403983, + "tokens_seen": 1049231360 + }, + { + "epoch": 0.32, + "learning_rate": 0.0005511895687061184, + "loss": 1.424, + "theoretical_loss": 3.6610601676000423, + "tokens_seen": 1049296896 + }, + { + "epoch": 0.32, + "learning_rate": 0.0005511735205616851, + "loss": 1.5171, + "theoretical_loss": 3.6610387685069217, + "tokens_seen": 1049362432 + }, + { + "epoch": 0.32, + "learning_rate": 0.0005511574724172518, + "loss": 1.4493, + "theoretical_loss": 3.6610173711243776, + "tokens_seen": 1049427968 + }, + { + "epoch": 0.32, + "learning_rate": 0.0005511414242728185, + "loss": 1.4717, + "theoretical_loss": 3.660995975452166, + "tokens_seen": 1049493504 + }, + { + "epoch": 0.32, + "learning_rate": 0.0005511253761283852, + "loss": 1.5089, + "theoretical_loss": 3.660974581490043, + "tokens_seen": 1049559040 + }, + { + "epoch": 0.32, + "learning_rate": 0.0005511093279839519, + "loss": 1.4251, + "theoretical_loss": 3.660953189237766, + "tokens_seen": 1049624576 + }, + { + "epoch": 0.32, + "learning_rate": 0.0005510932798395186, + "loss": 1.3707, + "theoretical_loss": 3.660931798695091, + "tokens_seen": 1049690112 + }, + { + "epoch": 0.32, + "learning_rate": 0.0005510772316950853, + "loss": 1.4264, + "theoretical_loss": 3.6609104098617746, + "tokens_seen": 1049755648 + }, + { + "epoch": 0.32, + "learning_rate": 0.000551061183550652, + "loss": 1.267, + "theoretical_loss": 3.6608890227375737, + "tokens_seen": 1049821184 + }, + { + "epoch": 0.32, + "learning_rate": 0.0005510451354062187, + "loss": 1.4499, + "theoretical_loss": 3.6608676373222457, + "tokens_seen": 1049886720 + }, + { + "epoch": 0.32, + "learning_rate": 0.0005510290872617854, + "loss": 1.4417, + "theoretical_loss": 3.6608462536155466, + "tokens_seen": 1049952256 + }, + { + "epoch": 0.32, + "learning_rate": 0.0005510130391173521, + "loss": 1.469, + "theoretical_loss": 3.6608248716172334, + "tokens_seen": 1050017792 + }, + { + "epoch": 0.32, + "learning_rate": 0.0005509969909729188, + "loss": 1.4047, + "theoretical_loss": 3.660803491327063, + "tokens_seen": 1050083328 + }, + { + "epoch": 0.32, + "learning_rate": 0.0005509809428284855, + "loss": 1.4357, + "theoretical_loss": 3.6607821127447924, + "tokens_seen": 1050148864 + }, + { + "epoch": 0.32, + "learning_rate": 0.0005509648946840522, + "loss": 1.4187, + "theoretical_loss": 3.6607607358701784, + "tokens_seen": 1050214400 + }, + { + "epoch": 0.32, + "learning_rate": 0.0005509488465396189, + "loss": 1.5671, + "theoretical_loss": 3.660739360702978, + "tokens_seen": 1050279936 + }, + { + "epoch": 0.32, + "learning_rate": 0.0005509327983951856, + "loss": 1.4991, + "theoretical_loss": 3.660717987242949, + "tokens_seen": 1050345472 + }, + { + "epoch": 0.32, + "learning_rate": 0.0005509167502507523, + "loss": 1.3848, + "theoretical_loss": 3.660696615489848, + "tokens_seen": 1050411008 + }, + { + "epoch": 0.32, + "learning_rate": 0.000550900702106319, + "loss": 1.409, + "theoretical_loss": 3.6606752454434326, + "tokens_seen": 1050476544 + }, + { + "epoch": 0.32, + "learning_rate": 0.0005508846539618857, + "loss": 1.4386, + "theoretical_loss": 3.660653877103459, + "tokens_seen": 1050542080 + }, + { + "epoch": 0.32, + "learning_rate": 0.0005508686058174524, + "loss": 1.4043, + "theoretical_loss": 3.660632510469686, + "tokens_seen": 1050607616 + }, + { + "epoch": 0.32, + "learning_rate": 0.0005508525576730191, + "loss": 1.4096, + "theoretical_loss": 3.66061114554187, + "tokens_seen": 1050673152 + }, + { + "epoch": 0.32, + "learning_rate": 0.0005508365095285859, + "loss": 1.3545, + "theoretical_loss": 3.6605897823197684, + "tokens_seen": 1050738688 + }, + { + "epoch": 0.32, + "learning_rate": 0.0005508204613841525, + "loss": 1.5038, + "theoretical_loss": 3.6605684208031386, + "tokens_seen": 1050804224 + }, + { + "epoch": 0.32, + "learning_rate": 0.0005508044132397193, + "loss": 1.5497, + "theoretical_loss": 3.660547060991739, + "tokens_seen": 1050869760 + }, + { + "epoch": 0.32, + "learning_rate": 0.0005507883650952859, + "loss": 1.5254, + "theoretical_loss": 3.6605257028853257, + "tokens_seen": 1050935296 + }, + { + "epoch": 0.32, + "learning_rate": 0.0005507723169508527, + "loss": 1.4665, + "theoretical_loss": 3.6605043464836573, + "tokens_seen": 1051000832 + }, + { + "epoch": 0.32, + "learning_rate": 0.0005507562688064193, + "loss": 1.3854, + "theoretical_loss": 3.660482991786491, + "tokens_seen": 1051066368 + }, + { + "epoch": 0.32, + "learning_rate": 0.0005507402206619861, + "loss": 1.459, + "theoretical_loss": 3.660461638793585, + "tokens_seen": 1051131904 + }, + { + "epoch": 0.32, + "learning_rate": 0.0005507241725175527, + "loss": 1.5168, + "theoretical_loss": 3.660440287504697, + "tokens_seen": 1051197440 + }, + { + "epoch": 0.32, + "learning_rate": 0.0005507081243731195, + "loss": 1.4357, + "theoretical_loss": 3.6604189379195837, + "tokens_seen": 1051262976 + }, + { + "epoch": 0.32, + "learning_rate": 0.0005506920762286861, + "loss": 1.4378, + "theoretical_loss": 3.6603975900380044, + "tokens_seen": 1051328512 + }, + { + "epoch": 0.32, + "learning_rate": 0.0005506760280842529, + "loss": 1.5107, + "theoretical_loss": 3.660376243859716, + "tokens_seen": 1051394048 + }, + { + "epoch": 0.32, + "learning_rate": 0.0005506599799398195, + "loss": 1.5389, + "theoretical_loss": 3.660354899384477, + "tokens_seen": 1051459584 + }, + { + "epoch": 0.32, + "learning_rate": 0.0005506439317953861, + "loss": 1.4711, + "theoretical_loss": 3.6603335566120454, + "tokens_seen": 1051525120 + }, + { + "epoch": 0.32, + "learning_rate": 0.0005506278836509529, + "loss": 1.5236, + "theoretical_loss": 3.660312215542179, + "tokens_seen": 1051590656 + }, + { + "epoch": 0.32, + "learning_rate": 0.0005506118355065196, + "loss": 1.3973, + "theoretical_loss": 3.6602908761746358, + "tokens_seen": 1051656192 + }, + { + "epoch": 0.32, + "learning_rate": 0.0005505957873620863, + "loss": 1.5379, + "theoretical_loss": 3.6602695385091737, + "tokens_seen": 1051721728 + }, + { + "epoch": 0.32, + "learning_rate": 0.000550579739217653, + "loss": 1.4994, + "theoretical_loss": 3.660248202545551, + "tokens_seen": 1051787264 + }, + { + "epoch": 0.32, + "objective/train/docs_used": 385698, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.4052953720092773, + "objective/train/theoretical_loss": 3.660226868283527, + "objective/train/tokens_used": 1072312800, + "theoretical_loss": 3.660226868283527, + "tokens_seen": 1051852800 + }, + { + "epoch": 0.32, + "learning_rate": 0.0005505636910732197, + "loss": 1.3922, + "theoretical_loss": 3.660226868283527, + "tokens_seen": 1051852800 + }, + { + "epoch": 0.32, + "learning_rate": 0.0005505476429287864, + "loss": 1.5008, + "theoretical_loss": 3.660205535722859, + "tokens_seen": 1051918336 + }, + { + "epoch": 0.32, + "learning_rate": 0.0005505315947843531, + "loss": 1.4614, + "theoretical_loss": 3.660184204863305, + "tokens_seen": 1051983872 + }, + { + "epoch": 0.32, + "learning_rate": 0.0005505155466399198, + "loss": 1.337, + "theoretical_loss": 3.6601628757046245, + "tokens_seen": 1052049408 + }, + { + "epoch": 0.32, + "learning_rate": 0.0005504994984954864, + "loss": 1.4096, + "theoretical_loss": 3.660141548246575, + "tokens_seen": 1052114944 + }, + { + "epoch": 0.32, + "learning_rate": 0.0005504834503510532, + "loss": 1.3319, + "theoretical_loss": 3.6601202224889153, + "tokens_seen": 1052180480 + }, + { + "epoch": 0.32, + "learning_rate": 0.0005504674022066198, + "loss": 1.3347, + "theoretical_loss": 3.6600988984314045, + "tokens_seen": 1052246016 + }, + { + "epoch": 0.32, + "learning_rate": 0.0005504513540621866, + "loss": 1.4097, + "theoretical_loss": 3.6600775760738, + "tokens_seen": 1052311552 + }, + { + "epoch": 0.32, + "learning_rate": 0.0005504353059177532, + "loss": 1.2791, + "theoretical_loss": 3.660056255415861, + "tokens_seen": 1052377088 + }, + { + "epoch": 0.32, + "learning_rate": 0.00055041925777332, + "loss": 1.4849, + "theoretical_loss": 3.660034936457347, + "tokens_seen": 1052442624 + }, + { + "epoch": 0.32, + "learning_rate": 0.0005504032096288866, + "loss": 1.3782, + "theoretical_loss": 3.660013619198015, + "tokens_seen": 1052508160 + }, + { + "epoch": 0.32, + "learning_rate": 0.0005503871614844534, + "loss": 1.4054, + "theoretical_loss": 3.659992303637625, + "tokens_seen": 1052573696 + }, + { + "epoch": 0.32, + "learning_rate": 0.00055037111334002, + "loss": 1.3304, + "theoretical_loss": 3.6599709897759363, + "tokens_seen": 1052639232 + }, + { + "epoch": 0.32, + "learning_rate": 0.0005503550651955868, + "loss": 1.3893, + "theoretical_loss": 3.6599496776127065, + "tokens_seen": 1052704768 + }, + { + "epoch": 0.32, + "learning_rate": 0.0005503390170511534, + "loss": 1.4031, + "theoretical_loss": 3.659928367147695, + "tokens_seen": 1052770304 + }, + { + "epoch": 0.32, + "learning_rate": 0.0005503229689067202, + "loss": 1.3567, + "theoretical_loss": 3.6599070583806617, + "tokens_seen": 1052835840 + }, + { + "epoch": 0.32, + "learning_rate": 0.0005503069207622868, + "loss": 1.3783, + "theoretical_loss": 3.6598857513113634, + "tokens_seen": 1052901376 + }, + { + "epoch": 0.32, + "learning_rate": 0.0005502908726178536, + "loss": 1.3129, + "theoretical_loss": 3.6598644459395615, + "tokens_seen": 1052966912 + }, + { + "epoch": 0.32, + "learning_rate": 0.0005502748244734202, + "loss": 1.3712, + "theoretical_loss": 3.6598431422650144, + "tokens_seen": 1053032448 + }, + { + "epoch": 0.32, + "learning_rate": 0.000550258776328987, + "loss": 1.4446, + "theoretical_loss": 3.6598218402874805, + "tokens_seen": 1053097984 + }, + { + "epoch": 0.32, + "learning_rate": 0.0005502427281845536, + "loss": 1.3513, + "theoretical_loss": 3.65980054000672, + "tokens_seen": 1053163520 + }, + { + "epoch": 0.32, + "learning_rate": 0.0005502266800401204, + "loss": 1.3449, + "theoretical_loss": 3.6597792414224912, + "tokens_seen": 1053229056 + }, + { + "epoch": 0.32, + "learning_rate": 0.000550210631895687, + "loss": 1.5059, + "theoretical_loss": 3.6597579445345545, + "tokens_seen": 1053294592 + }, + { + "epoch": 0.32, + "learning_rate": 0.0005501945837512538, + "loss": 1.482, + "theoretical_loss": 3.6597366493426686, + "tokens_seen": 1053360128 + }, + { + "epoch": 0.32, + "learning_rate": 0.0005501785356068204, + "loss": 1.6505, + "theoretical_loss": 3.6597153558465934, + "tokens_seen": 1053425664 + }, + { + "epoch": 0.32, + "learning_rate": 0.0005501624874623872, + "loss": 1.4561, + "theoretical_loss": 3.6596940640460875, + "tokens_seen": 1053491200 + }, + { + "epoch": 0.32, + "learning_rate": 0.0005501464393179539, + "loss": 1.4071, + "theoretical_loss": 3.6596727739409114, + "tokens_seen": 1053556736 + }, + { + "epoch": 0.32, + "learning_rate": 0.0005501303911735206, + "loss": 1.4608, + "theoretical_loss": 3.6596514855308238, + "tokens_seen": 1053622272 + }, + { + "epoch": 0.32, + "learning_rate": 0.0005501143430290873, + "loss": 1.4121, + "theoretical_loss": 3.6596301988155853, + "tokens_seen": 1053687808 + }, + { + "epoch": 0.32, + "learning_rate": 0.000550098294884654, + "loss": 1.4799, + "theoretical_loss": 3.659608913794955, + "tokens_seen": 1053753344 + }, + { + "epoch": 0.32, + "learning_rate": 0.0005500822467402207, + "loss": 1.4979, + "theoretical_loss": 3.659587630468692, + "tokens_seen": 1053818880 + }, + { + "epoch": 0.32, + "learning_rate": 0.0005500661985957874, + "loss": 1.4817, + "theoretical_loss": 3.6595663488365573, + "tokens_seen": 1053884416 + }, + { + "epoch": 0.32, + "learning_rate": 0.0005500501504513541, + "loss": 1.4738, + "theoretical_loss": 3.6595450688983098, + "tokens_seen": 1053949952 + }, + { + "epoch": 0.32, + "learning_rate": 0.0005500341023069208, + "loss": 1.4846, + "theoretical_loss": 3.6595237906537097, + "tokens_seen": 1054015488 + }, + { + "epoch": 0.32, + "learning_rate": 0.0005500180541624875, + "loss": 1.3948, + "theoretical_loss": 3.659502514102517, + "tokens_seen": 1054081024 + }, + { + "epoch": 0.32, + "learning_rate": 0.0005500020060180542, + "loss": 1.3188, + "theoretical_loss": 3.6594812392444913, + "tokens_seen": 1054146560 + }, + { + "epoch": 0.32, + "learning_rate": 0.0005499859578736209, + "loss": 1.4511, + "theoretical_loss": 3.6594599660793934, + "tokens_seen": 1054212096 + }, + { + "epoch": 0.32, + "learning_rate": 0.0005499699097291876, + "loss": 1.4836, + "theoretical_loss": 3.6594386946069823, + "tokens_seen": 1054277632 + }, + { + "epoch": 0.32, + "learning_rate": 0.0005499538615847543, + "loss": 1.4432, + "theoretical_loss": 3.659417424827019, + "tokens_seen": 1054343168 + }, + { + "epoch": 0.32, + "learning_rate": 0.000549937813440321, + "loss": 1.4075, + "theoretical_loss": 3.6593961567392626, + "tokens_seen": 1054408704 + }, + { + "epoch": 0.32, + "learning_rate": 0.0005499217652958877, + "loss": 1.3935, + "theoretical_loss": 3.6593748903434746, + "tokens_seen": 1054474240 + }, + { + "epoch": 0.32, + "learning_rate": 0.0005499057171514544, + "loss": 1.3719, + "theoretical_loss": 3.6593536256394144, + "tokens_seen": 1054539776 + }, + { + "epoch": 0.32, + "learning_rate": 0.0005498896690070211, + "loss": 1.4306, + "theoretical_loss": 3.6593323626268424, + "tokens_seen": 1054605312 + }, + { + "epoch": 0.32, + "learning_rate": 0.0005498736208625878, + "loss": 1.4521, + "theoretical_loss": 3.6593111013055193, + "tokens_seen": 1054670848 + }, + { + "epoch": 0.32, + "learning_rate": 0.0005498575727181545, + "loss": 1.4557, + "theoretical_loss": 3.6592898416752053, + "tokens_seen": 1054736384 + }, + { + "epoch": 0.32, + "learning_rate": 0.0005498415245737212, + "loss": 1.4103, + "theoretical_loss": 3.6592685837356607, + "tokens_seen": 1054801920 + }, + { + "epoch": 0.32, + "learning_rate": 0.0005498254764292879, + "loss": 1.484, + "theoretical_loss": 3.6592473274866464, + "tokens_seen": 1054867456 + }, + { + "epoch": 0.32, + "learning_rate": 0.0005498094282848546, + "loss": 1.2487, + "theoretical_loss": 3.6592260729279222, + "tokens_seen": 1054932992 + }, + { + "epoch": 0.32, + "learning_rate": 0.0005497933801404213, + "loss": 1.4368, + "theoretical_loss": 3.6592048200592497, + "tokens_seen": 1054998528 + }, + { + "epoch": 0.32, + "learning_rate": 0.000549777331995988, + "loss": 1.4327, + "theoretical_loss": 3.6591835688803886, + "tokens_seen": 1055064064 + }, + { + "epoch": 0.32, + "objective/train/docs_used": 386814, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.3752086162567139, + "objective/train/theoretical_loss": 3.6591623193911, + "objective/train/tokens_used": 1075589600, + "theoretical_loss": 3.6591623193911, + "tokens_seen": 1055129600 + }, + { + "epoch": 0.32, + "learning_rate": 0.0005497612838515547, + "loss": 1.2889, + "theoretical_loss": 3.6591623193911, + "tokens_seen": 1055129600 + }, + { + "epoch": 0.32, + "learning_rate": 0.0005497452357071214, + "loss": 1.3512, + "theoretical_loss": 3.6591410715911445, + "tokens_seen": 1055195136 + }, + { + "epoch": 0.32, + "learning_rate": 0.0005497291875626881, + "loss": 1.4458, + "theoretical_loss": 3.6591198254802837, + "tokens_seen": 1055260672 + }, + { + "epoch": 0.32, + "learning_rate": 0.0005497131394182548, + "loss": 1.3585, + "theoretical_loss": 3.6590985810582772, + "tokens_seen": 1055326208 + }, + { + "epoch": 0.32, + "learning_rate": 0.0005496970912738215, + "loss": 1.3957, + "theoretical_loss": 3.6590773383248862, + "tokens_seen": 1055391744 + }, + { + "epoch": 0.32, + "learning_rate": 0.0005496810431293882, + "loss": 1.4428, + "theoretical_loss": 3.6590560972798727, + "tokens_seen": 1055457280 + }, + { + "epoch": 0.32, + "learning_rate": 0.0005496649949849549, + "loss": 1.41, + "theoretical_loss": 3.6590348579229963, + "tokens_seen": 1055522816 + }, + { + "epoch": 0.32, + "learning_rate": 0.0005496489468405216, + "loss": 1.4633, + "theoretical_loss": 3.6590136202540187, + "tokens_seen": 1055588352 + }, + { + "epoch": 0.32, + "learning_rate": 0.0005496328986960883, + "loss": 1.3596, + "theoretical_loss": 3.6589923842727012, + "tokens_seen": 1055653888 + }, + { + "epoch": 0.32, + "learning_rate": 0.000549616850551655, + "loss": 1.3663, + "theoretical_loss": 3.6589711499788042, + "tokens_seen": 1055719424 + }, + { + "epoch": 0.32, + "learning_rate": 0.0005496008024072217, + "loss": 1.4549, + "theoretical_loss": 3.658949917372089, + "tokens_seen": 1055784960 + }, + { + "epoch": 0.32, + "learning_rate": 0.0005495847542627884, + "loss": 1.4122, + "theoretical_loss": 3.6589286864523176, + "tokens_seen": 1055850496 + }, + { + "epoch": 0.32, + "learning_rate": 0.0005495687061183551, + "loss": 1.4112, + "theoretical_loss": 3.658907457219251, + "tokens_seen": 1055916032 + }, + { + "epoch": 0.32, + "learning_rate": 0.0005495526579739218, + "loss": 1.388, + "theoretical_loss": 3.65888622967265, + "tokens_seen": 1055981568 + }, + { + "epoch": 0.32, + "learning_rate": 0.0005495366098294885, + "loss": 1.39, + "theoretical_loss": 3.6588650038122763, + "tokens_seen": 1056047104 + }, + { + "epoch": 0.32, + "learning_rate": 0.0005495205616850552, + "loss": 1.4303, + "theoretical_loss": 3.6588437796378916, + "tokens_seen": 1056112640 + }, + { + "epoch": 0.32, + "learning_rate": 0.000549504513540622, + "loss": 1.5576, + "theoretical_loss": 3.658822557149257, + "tokens_seen": 1056178176 + }, + { + "epoch": 0.32, + "learning_rate": 0.0005494884653961886, + "loss": 1.4611, + "theoretical_loss": 3.658801336346134, + "tokens_seen": 1056243712 + }, + { + "epoch": 0.32, + "learning_rate": 0.0005494724172517554, + "loss": 1.3788, + "theoretical_loss": 3.658780117228284, + "tokens_seen": 1056309248 + }, + { + "epoch": 0.32, + "learning_rate": 0.000549456369107322, + "loss": 1.418, + "theoretical_loss": 3.65875889979547, + "tokens_seen": 1056374784 + }, + { + "epoch": 0.32, + "learning_rate": 0.0005494403209628888, + "loss": 1.5516, + "theoretical_loss": 3.658737684047451, + "tokens_seen": 1056440320 + }, + { + "epoch": 0.32, + "learning_rate": 0.0005494242728184554, + "loss": 1.3729, + "theoretical_loss": 3.6587164699839914, + "tokens_seen": 1056505856 + }, + { + "epoch": 0.32, + "learning_rate": 0.0005494082246740222, + "loss": 1.3876, + "theoretical_loss": 3.6586952576048515, + "tokens_seen": 1056571392 + }, + { + "epoch": 0.32, + "learning_rate": 0.0005493921765295888, + "loss": 1.3752, + "theoretical_loss": 3.658674046909793, + "tokens_seen": 1056636928 + }, + { + "epoch": 0.32, + "learning_rate": 0.0005493761283851556, + "loss": 1.3468, + "theoretical_loss": 3.658652837898579, + "tokens_seen": 1056702464 + }, + { + "epoch": 0.32, + "learning_rate": 0.0005493600802407222, + "loss": 1.3509, + "theoretical_loss": 3.6586316305709703, + "tokens_seen": 1056768000 + }, + { + "epoch": 0.32, + "learning_rate": 0.000549344032096289, + "loss": 1.4839, + "theoretical_loss": 3.658610424926729, + "tokens_seen": 1056833536 + }, + { + "epoch": 0.32, + "learning_rate": 0.0005493279839518556, + "loss": 1.4055, + "theoretical_loss": 3.6585892209656175, + "tokens_seen": 1056899072 + }, + { + "epoch": 0.32, + "learning_rate": 0.0005493119358074224, + "loss": 1.4796, + "theoretical_loss": 3.6585680186873972, + "tokens_seen": 1056964608 + }, + { + "epoch": 0.32, + "learning_rate": 0.000549295887662989, + "loss": 1.3823, + "theoretical_loss": 3.658546818091831, + "tokens_seen": 1057030144 + }, + { + "epoch": 0.32, + "learning_rate": 0.0005492798395185558, + "loss": 1.3653, + "theoretical_loss": 3.658525619178681, + "tokens_seen": 1057095680 + }, + { + "epoch": 0.32, + "learning_rate": 0.0005492637913741224, + "loss": 1.4323, + "theoretical_loss": 3.658504421947708, + "tokens_seen": 1057161216 + }, + { + "epoch": 0.32, + "learning_rate": 0.000549247743229689, + "loss": 1.5168, + "theoretical_loss": 3.658483226398676, + "tokens_seen": 1057226752 + }, + { + "epoch": 0.32, + "learning_rate": 0.0005492316950852558, + "loss": 1.4872, + "theoretical_loss": 3.6584620325313466, + "tokens_seen": 1057292288 + }, + { + "epoch": 0.32, + "learning_rate": 0.0005492156469408225, + "loss": 1.3689, + "theoretical_loss": 3.6584408403454822, + "tokens_seen": 1057357824 + }, + { + "epoch": 0.32, + "learning_rate": 0.0005491995987963892, + "loss": 1.3124, + "theoretical_loss": 3.6584196498408454, + "tokens_seen": 1057423360 + }, + { + "epoch": 0.32, + "learning_rate": 0.0005491835506519559, + "loss": 1.3788, + "theoretical_loss": 3.658398461017198, + "tokens_seen": 1057488896 + }, + { + "epoch": 0.32, + "learning_rate": 0.0005491675025075226, + "loss": 1.459, + "theoretical_loss": 3.6583772738743026, + "tokens_seen": 1057554432 + }, + { + "epoch": 0.32, + "learning_rate": 0.0005491514543630893, + "loss": 1.4737, + "theoretical_loss": 3.6583560884119226, + "tokens_seen": 1057619968 + }, + { + "epoch": 0.32, + "learning_rate": 0.000549135406218656, + "loss": 1.4416, + "theoretical_loss": 3.6583349046298195, + "tokens_seen": 1057685504 + }, + { + "epoch": 0.32, + "learning_rate": 0.0005491193580742227, + "loss": 1.4192, + "theoretical_loss": 3.6583137225277564, + "tokens_seen": 1057751040 + }, + { + "epoch": 0.32, + "learning_rate": 0.0005491033099297893, + "loss": 1.4413, + "theoretical_loss": 3.6582925421054964, + "tokens_seen": 1057816576 + }, + { + "epoch": 0.32, + "learning_rate": 0.0005490872617853561, + "loss": 1.3308, + "theoretical_loss": 3.6582713633628017, + "tokens_seen": 1057882112 + }, + { + "epoch": 0.32, + "learning_rate": 0.0005490712136409227, + "loss": 1.3387, + "theoretical_loss": 3.658250186299435, + "tokens_seen": 1057947648 + }, + { + "epoch": 0.32, + "learning_rate": 0.0005490551654964895, + "loss": 1.3989, + "theoretical_loss": 3.65822901091516, + "tokens_seen": 1058013184 + }, + { + "epoch": 0.32, + "learning_rate": 0.0005490391173520561, + "loss": 1.3133, + "theoretical_loss": 3.658207837209738, + "tokens_seen": 1058078720 + }, + { + "epoch": 0.32, + "learning_rate": 0.0005490230692076229, + "loss": 1.4241, + "theoretical_loss": 3.6581866651829333, + "tokens_seen": 1058144256 + }, + { + "epoch": 0.32, + "learning_rate": 0.0005490070210631895, + "loss": 1.3274, + "theoretical_loss": 3.658165494834508, + "tokens_seen": 1058209792 + }, + { + "epoch": 0.32, + "learning_rate": 0.0005489909729187563, + "loss": 1.3977, + "theoretical_loss": 3.6581443261642264, + "tokens_seen": 1058275328 + }, + { + "epoch": 0.32, + "learning_rate": 0.0005489749247743229, + "loss": 1.3771, + "theoretical_loss": 3.6581231591718497, + "tokens_seen": 1058340864 + }, + { + "epoch": 0.32, + "objective/train/docs_used": 388023, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.3609163761138916, + "objective/train/theoretical_loss": 3.6581019938571426, + "objective/train/tokens_used": 1078866400, + "theoretical_loss": 3.6581019938571426, + "tokens_seen": 1058406400 + }, + { + "epoch": 0.32, + "learning_rate": 0.0005489588766298897, + "loss": 1.3282, + "theoretical_loss": 3.6581019938571426, + "tokens_seen": 1058406400 + }, + { + "epoch": 0.32, + "learning_rate": 0.0005489428284854564, + "loss": 1.3871, + "theoretical_loss": 3.658080830219867, + "tokens_seen": 1058471936 + }, + { + "epoch": 0.32, + "learning_rate": 0.0005489267803410231, + "loss": 1.4874, + "theoretical_loss": 3.6580596682597877, + "tokens_seen": 1058537472 + }, + { + "epoch": 0.32, + "learning_rate": 0.0005489107321965898, + "loss": 1.4302, + "theoretical_loss": 3.6580385079766664, + "tokens_seen": 1058603008 + }, + { + "epoch": 0.32, + "learning_rate": 0.0005488946840521565, + "loss": 1.4896, + "theoretical_loss": 3.658017349370267, + "tokens_seen": 1058668544 + }, + { + "epoch": 0.32, + "learning_rate": 0.0005488786359077232, + "loss": 1.3485, + "theoretical_loss": 3.6579961924403532, + "tokens_seen": 1058734080 + }, + { + "epoch": 0.32, + "learning_rate": 0.0005488625877632899, + "loss": 1.4266, + "theoretical_loss": 3.657975037186688, + "tokens_seen": 1058799616 + }, + { + "epoch": 0.32, + "learning_rate": 0.0005488465396188566, + "loss": 1.4118, + "theoretical_loss": 3.657953883609035, + "tokens_seen": 1058865152 + }, + { + "epoch": 0.32, + "learning_rate": 0.0005488304914744233, + "loss": 1.3641, + "theoretical_loss": 3.6579327317071573, + "tokens_seen": 1058930688 + }, + { + "epoch": 0.32, + "learning_rate": 0.00054881444332999, + "loss": 1.4763, + "theoretical_loss": 3.6579115814808185, + "tokens_seen": 1058996224 + }, + { + "epoch": 0.32, + "learning_rate": 0.0005487983951855567, + "loss": 1.4306, + "theoretical_loss": 3.6578904329297837, + "tokens_seen": 1059061760 + }, + { + "epoch": 0.32, + "learning_rate": 0.0005487823470411234, + "loss": 1.6013, + "theoretical_loss": 3.6578692860538142, + "tokens_seen": 1059127296 + }, + { + "epoch": 0.32, + "learning_rate": 0.0005487662988966901, + "loss": 1.3774, + "theoretical_loss": 3.6578481408526753, + "tokens_seen": 1059192832 + }, + { + "epoch": 0.32, + "learning_rate": 0.0005487502507522568, + "loss": 1.4831, + "theoretical_loss": 3.65782699732613, + "tokens_seen": 1059258368 + }, + { + "epoch": 0.32, + "learning_rate": 0.0005487342026078235, + "loss": 1.5127, + "theoretical_loss": 3.657805855473942, + "tokens_seen": 1059323904 + }, + { + "epoch": 0.32, + "learning_rate": 0.0005487181544633902, + "loss": 1.4185, + "theoretical_loss": 3.6577847152958762, + "tokens_seen": 1059389440 + }, + { + "epoch": 0.32, + "learning_rate": 0.0005487021063189569, + "loss": 1.4169, + "theoretical_loss": 3.6577635767916954, + "tokens_seen": 1059454976 + }, + { + "epoch": 0.32, + "learning_rate": 0.0005486860581745236, + "loss": 1.3388, + "theoretical_loss": 3.657742439961164, + "tokens_seen": 1059520512 + }, + { + "epoch": 0.32, + "learning_rate": 0.0005486700100300903, + "loss": 1.2614, + "theoretical_loss": 3.657721304804046, + "tokens_seen": 1059586048 + }, + { + "epoch": 0.32, + "learning_rate": 0.000548653961885657, + "loss": 1.2882, + "theoretical_loss": 3.657700171320105, + "tokens_seen": 1059651584 + }, + { + "epoch": 0.32, + "learning_rate": 0.0005486379137412237, + "loss": 1.453, + "theoretical_loss": 3.6576790395091052, + "tokens_seen": 1059717120 + }, + { + "epoch": 0.32, + "learning_rate": 0.0005486218655967904, + "loss": 1.3713, + "theoretical_loss": 3.6576579093708115, + "tokens_seen": 1059782656 + }, + { + "epoch": 0.32, + "learning_rate": 0.0005486058174523571, + "loss": 1.3868, + "theoretical_loss": 3.657636780904987, + "tokens_seen": 1059848192 + }, + { + "epoch": 0.32, + "learning_rate": 0.0005485897693079238, + "loss": 1.4748, + "theoretical_loss": 3.6576156541113964, + "tokens_seen": 1059913728 + }, + { + "epoch": 0.32, + "learning_rate": 0.0005485737211634905, + "loss": 1.3924, + "theoretical_loss": 3.657594528989804, + "tokens_seen": 1059979264 + }, + { + "epoch": 0.32, + "learning_rate": 0.0005485576730190572, + "loss": 1.2438, + "theoretical_loss": 3.6575734055399742, + "tokens_seen": 1060044800 + }, + { + "epoch": 0.32, + "learning_rate": 0.0005485416248746239, + "loss": 1.4256, + "theoretical_loss": 3.657552283761671, + "tokens_seen": 1060110336 + }, + { + "epoch": 0.32, + "learning_rate": 0.0005485255767301906, + "loss": 1.333, + "theoretical_loss": 3.6575311636546584, + "tokens_seen": 1060175872 + }, + { + "epoch": 0.32, + "learning_rate": 0.0005485095285857573, + "loss": 1.3502, + "theoretical_loss": 3.6575100452187024, + "tokens_seen": 1060241408 + }, + { + "epoch": 0.32, + "learning_rate": 0.000548493480441324, + "loss": 1.4182, + "theoretical_loss": 3.6574889284535663, + "tokens_seen": 1060306944 + }, + { + "epoch": 0.32, + "learning_rate": 0.0005484774322968907, + "loss": 1.3848, + "theoretical_loss": 3.657467813359015, + "tokens_seen": 1060372480 + }, + { + "epoch": 0.32, + "learning_rate": 0.0005484613841524574, + "loss": 1.4696, + "theoretical_loss": 3.6574466999348125, + "tokens_seen": 1060438016 + }, + { + "epoch": 0.32, + "learning_rate": 0.0005484453360080241, + "loss": 1.3623, + "theoretical_loss": 3.6574255881807245, + "tokens_seen": 1060503552 + }, + { + "epoch": 0.32, + "learning_rate": 0.0005484292878635908, + "loss": 1.4587, + "theoretical_loss": 3.657404478096515, + "tokens_seen": 1060569088 + }, + { + "epoch": 0.32, + "learning_rate": 0.0005484132397191575, + "loss": 1.3166, + "theoretical_loss": 3.657383369681949, + "tokens_seen": 1060634624 + }, + { + "epoch": 0.32, + "learning_rate": 0.0005483971915747242, + "loss": 1.415, + "theoretical_loss": 3.657362262936791, + "tokens_seen": 1060700160 + }, + { + "epoch": 0.32, + "learning_rate": 0.0005483811434302909, + "loss": 1.3816, + "theoretical_loss": 3.6573411578608064, + "tokens_seen": 1060765696 + }, + { + "epoch": 0.32, + "learning_rate": 0.0005483650952858576, + "loss": 1.4534, + "theoretical_loss": 3.657320054453759, + "tokens_seen": 1060831232 + }, + { + "epoch": 0.32, + "learning_rate": 0.0005483490471414243, + "loss": 1.4323, + "theoretical_loss": 3.6572989527154154, + "tokens_seen": 1060896768 + }, + { + "epoch": 0.32, + "learning_rate": 0.000548332998996991, + "loss": 1.4734, + "theoretical_loss": 3.6572778526455387, + "tokens_seen": 1060962304 + }, + { + "epoch": 0.32, + "learning_rate": 0.0005483169508525577, + "loss": 1.5391, + "theoretical_loss": 3.6572567542438956, + "tokens_seen": 1061027840 + }, + { + "epoch": 0.32, + "learning_rate": 0.0005483009027081244, + "loss": 1.3682, + "theoretical_loss": 3.6572356575102503, + "tokens_seen": 1061093376 + }, + { + "epoch": 0.32, + "learning_rate": 0.0005482848545636911, + "loss": 1.5517, + "theoretical_loss": 3.6572145624443677, + "tokens_seen": 1061158912 + }, + { + "epoch": 0.32, + "learning_rate": 0.0005482688064192578, + "loss": 1.47, + "theoretical_loss": 3.6571934690460135, + "tokens_seen": 1061224448 + }, + { + "epoch": 0.32, + "learning_rate": 0.0005482527582748245, + "loss": 1.3547, + "theoretical_loss": 3.6571723773149527, + "tokens_seen": 1061289984 + }, + { + "epoch": 0.32, + "learning_rate": 0.0005482367101303912, + "loss": 1.5132, + "theoretical_loss": 3.657151287250951, + "tokens_seen": 1061355520 + }, + { + "epoch": 0.32, + "learning_rate": 0.0005482206619859579, + "loss": 1.2895, + "theoretical_loss": 3.657130198853773, + "tokens_seen": 1061421056 + }, + { + "epoch": 0.32, + "learning_rate": 0.0005482046138415246, + "loss": 1.391, + "theoretical_loss": 3.6571091121231847, + "tokens_seen": 1061486592 + }, + { + "epoch": 0.32, + "learning_rate": 0.0005481885656970913, + "loss": 1.433, + "theoretical_loss": 3.6570880270589505, + "tokens_seen": 1061552128 + }, + { + "epoch": 0.32, + "learning_rate": 0.000548172517552658, + "loss": 1.3644, + "theoretical_loss": 3.657066943660837, + "tokens_seen": 1061617664 + }, + { + "epoch": 0.32, + "objective/train/docs_used": 389244, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.4860990047454834, + "objective/train/theoretical_loss": 3.6570458619286095, + "objective/train/tokens_used": 1082143200, + "theoretical_loss": 3.6570458619286095, + "tokens_seen": 1061683200 + }, + { + "epoch": 0.32, + "learning_rate": 0.0005481564694082247, + "loss": 1.4726, + "theoretical_loss": 3.6570458619286095, + "tokens_seen": 1061683200 + }, + { + "epoch": 0.32, + "learning_rate": 0.0005481404212637914, + "loss": 1.4597, + "theoretical_loss": 3.6570247818620327, + "tokens_seen": 1061748736 + }, + { + "epoch": 0.32, + "learning_rate": 0.0005481243731193581, + "loss": 1.4466, + "theoretical_loss": 3.657003703460873, + "tokens_seen": 1061814272 + }, + { + "epoch": 0.32, + "learning_rate": 0.0005481083249749249, + "loss": 1.3817, + "theoretical_loss": 3.656982626724896, + "tokens_seen": 1061879808 + }, + { + "epoch": 0.32, + "learning_rate": 0.0005480922768304915, + "loss": 1.4139, + "theoretical_loss": 3.6569615516538674, + "tokens_seen": 1061945344 + }, + { + "epoch": 0.32, + "learning_rate": 0.0005480762286860583, + "loss": 1.5111, + "theoretical_loss": 3.656940478247553, + "tokens_seen": 1062010880 + }, + { + "epoch": 0.32, + "learning_rate": 0.0005480601805416249, + "loss": 1.4429, + "theoretical_loss": 3.6569194065057182, + "tokens_seen": 1062076416 + }, + { + "epoch": 0.32, + "learning_rate": 0.0005480441323971917, + "loss": 1.4899, + "theoretical_loss": 3.656898336428129, + "tokens_seen": 1062141952 + }, + { + "epoch": 0.32, + "learning_rate": 0.0005480280842527583, + "loss": 1.3964, + "theoretical_loss": 3.6568772680145507, + "tokens_seen": 1062207488 + }, + { + "epoch": 0.32, + "learning_rate": 0.000548012036108325, + "loss": 1.4033, + "theoretical_loss": 3.65685620126475, + "tokens_seen": 1062273024 + }, + { + "epoch": 0.32, + "learning_rate": 0.0005479959879638917, + "loss": 1.3299, + "theoretical_loss": 3.6568351361784934, + "tokens_seen": 1062338560 + }, + { + "epoch": 0.32, + "learning_rate": 0.0005479799398194585, + "loss": 1.4594, + "theoretical_loss": 3.656814072755546, + "tokens_seen": 1062404096 + }, + { + "epoch": 0.32, + "learning_rate": 0.0005479638916750251, + "loss": 1.4425, + "theoretical_loss": 3.656793010995674, + "tokens_seen": 1062469632 + }, + { + "epoch": 0.32, + "learning_rate": 0.0005479478435305919, + "loss": 1.3685, + "theoretical_loss": 3.6567719508986434, + "tokens_seen": 1062535168 + }, + { + "epoch": 0.32, + "learning_rate": 0.0005479317953861585, + "loss": 1.3379, + "theoretical_loss": 3.6567508924642205, + "tokens_seen": 1062600704 + }, + { + "epoch": 0.32, + "learning_rate": 0.0005479157472417253, + "loss": 1.4888, + "theoretical_loss": 3.656729835692172, + "tokens_seen": 1062666240 + }, + { + "epoch": 0.32, + "learning_rate": 0.0005478996990972919, + "loss": 1.5168, + "theoretical_loss": 3.656708780582264, + "tokens_seen": 1062731776 + }, + { + "epoch": 0.32, + "learning_rate": 0.0005478836509528587, + "loss": 1.3409, + "theoretical_loss": 3.656687727134262, + "tokens_seen": 1062797312 + }, + { + "epoch": 0.32, + "learning_rate": 0.0005478676028084253, + "loss": 1.358, + "theoretical_loss": 3.656666675347933, + "tokens_seen": 1062862848 + }, + { + "epoch": 0.32, + "learning_rate": 0.000547851554663992, + "loss": 1.5484, + "theoretical_loss": 3.6566456252230437, + "tokens_seen": 1062928384 + }, + { + "epoch": 0.32, + "learning_rate": 0.0005478355065195587, + "loss": 1.4092, + "theoretical_loss": 3.65662457675936, + "tokens_seen": 1062993920 + }, + { + "epoch": 0.32, + "learning_rate": 0.0005478194583751254, + "loss": 1.3579, + "theoretical_loss": 3.6566035299566484, + "tokens_seen": 1063059456 + }, + { + "epoch": 0.32, + "learning_rate": 0.0005478034102306921, + "loss": 1.4064, + "theoretical_loss": 3.6565824848146757, + "tokens_seen": 1063124992 + }, + { + "epoch": 0.32, + "learning_rate": 0.0005477873620862588, + "loss": 1.4338, + "theoretical_loss": 3.656561441333209, + "tokens_seen": 1063190528 + }, + { + "epoch": 0.32, + "learning_rate": 0.0005477713139418255, + "loss": 1.3058, + "theoretical_loss": 3.6565403995120134, + "tokens_seen": 1063256064 + }, + { + "epoch": 0.32, + "learning_rate": 0.0005477552657973922, + "loss": 1.4151, + "theoretical_loss": 3.656519359350857, + "tokens_seen": 1063321600 + }, + { + "epoch": 0.32, + "learning_rate": 0.0005477392176529589, + "loss": 1.2733, + "theoretical_loss": 3.656498320849506, + "tokens_seen": 1063387136 + }, + { + "epoch": 0.32, + "learning_rate": 0.0005477231695085256, + "loss": 1.442, + "theoretical_loss": 3.6564772840077278, + "tokens_seen": 1063452672 + }, + { + "epoch": 0.32, + "learning_rate": 0.0005477071213640923, + "loss": 1.4641, + "theoretical_loss": 3.6564562488252883, + "tokens_seen": 1063518208 + }, + { + "epoch": 0.32, + "learning_rate": 0.000547691073219659, + "loss": 1.4552, + "theoretical_loss": 3.6564352153019546, + "tokens_seen": 1063583744 + }, + { + "epoch": 0.32, + "learning_rate": 0.0005476750250752257, + "loss": 1.4673, + "theoretical_loss": 3.656414183437494, + "tokens_seen": 1063649280 + }, + { + "epoch": 0.32, + "learning_rate": 0.0005476589769307924, + "loss": 1.2559, + "theoretical_loss": 3.656393153231673, + "tokens_seen": 1063714816 + }, + { + "epoch": 0.32, + "learning_rate": 0.0005476429287863591, + "loss": 1.4318, + "theoretical_loss": 3.6563721246842595, + "tokens_seen": 1063780352 + }, + { + "epoch": 0.32, + "learning_rate": 0.0005476268806419258, + "loss": 1.394, + "theoretical_loss": 3.656351097795019, + "tokens_seen": 1063845888 + }, + { + "epoch": 0.32, + "learning_rate": 0.0005476108324974925, + "loss": 1.2844, + "theoretical_loss": 3.6563300725637204, + "tokens_seen": 1063911424 + }, + { + "epoch": 0.32, + "learning_rate": 0.0005475947843530592, + "loss": 1.3472, + "theoretical_loss": 3.6563090489901295, + "tokens_seen": 1063976960 + }, + { + "epoch": 0.32, + "learning_rate": 0.0005475787362086259, + "loss": 1.4701, + "theoretical_loss": 3.656288027074014, + "tokens_seen": 1064042496 + }, + { + "epoch": 0.32, + "learning_rate": 0.0005475626880641926, + "loss": 1.5498, + "theoretical_loss": 3.656267006815142, + "tokens_seen": 1064108032 + }, + { + "epoch": 0.32, + "learning_rate": 0.0005475466399197593, + "loss": 1.245, + "theoretical_loss": 3.6562459882132794, + "tokens_seen": 1064173568 + }, + { + "epoch": 0.32, + "learning_rate": 0.000547530591775326, + "loss": 1.5202, + "theoretical_loss": 3.656224971268194, + "tokens_seen": 1064239104 + }, + { + "epoch": 0.32, + "learning_rate": 0.0005475145436308927, + "loss": 1.3747, + "theoretical_loss": 3.6562039559796533, + "tokens_seen": 1064304640 + }, + { + "epoch": 0.32, + "learning_rate": 0.0005474984954864594, + "loss": 1.4632, + "theoretical_loss": 3.6561829423474252, + "tokens_seen": 1064370176 + }, + { + "epoch": 0.32, + "learning_rate": 0.0005474824473420261, + "loss": 1.5614, + "theoretical_loss": 3.656161930371276, + "tokens_seen": 1064435712 + }, + { + "epoch": 0.32, + "learning_rate": 0.0005474663991975928, + "loss": 1.4436, + "theoretical_loss": 3.6561409200509747, + "tokens_seen": 1064501248 + }, + { + "epoch": 0.32, + "learning_rate": 0.0005474503510531595, + "loss": 1.3906, + "theoretical_loss": 3.6561199113862877, + "tokens_seen": 1064566784 + }, + { + "epoch": 0.32, + "learning_rate": 0.0005474343029087262, + "loss": 1.3611, + "theoretical_loss": 3.656098904376983, + "tokens_seen": 1064632320 + }, + { + "epoch": 0.32, + "learning_rate": 0.0005474182547642929, + "loss": 1.3745, + "theoretical_loss": 3.6560778990228284, + "tokens_seen": 1064697856 + }, + { + "epoch": 0.32, + "learning_rate": 0.0005474022066198596, + "loss": 1.5192, + "theoretical_loss": 3.6560568953235917, + "tokens_seen": 1064763392 + }, + { + "epoch": 0.32, + "learning_rate": 0.0005473861584754263, + "loss": 1.4145, + "theoretical_loss": 3.6560358932790407, + "tokens_seen": 1064828928 + }, + { + "epoch": 0.32, + "learning_rate": 0.000547370110330993, + "loss": 1.4256, + "theoretical_loss": 3.656014892888943, + "tokens_seen": 1064894464 + }, + { + "epoch": 0.32, + "objective/train/docs_used": 390462, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.3413991928100586, + "objective/train/theoretical_loss": 3.6559938941530667, + "objective/train/tokens_used": 1085420000, + "theoretical_loss": 3.6559938941530667, + "tokens_seen": 1064960000 + }, + { + "epoch": 0.32, + "learning_rate": 0.0005473540621865597, + "loss": 1.4184, + "theoretical_loss": 3.6559938941530667, + "tokens_seen": 1064960000 + }, + { + "epoch": 0.32, + "learning_rate": 0.0005473380140421264, + "loss": 1.4244, + "theoretical_loss": 3.655972897071179, + "tokens_seen": 1065025536 + }, + { + "epoch": 0.32, + "learning_rate": 0.0005473219658976931, + "loss": 1.4027, + "theoretical_loss": 3.655951901643049, + "tokens_seen": 1065091072 + }, + { + "epoch": 0.32, + "learning_rate": 0.0005473059177532598, + "loss": 1.4854, + "theoretical_loss": 3.655930907868443, + "tokens_seen": 1065156608 + }, + { + "epoch": 0.32, + "learning_rate": 0.0005472898696088265, + "loss": 1.4947, + "theoretical_loss": 3.6559099157471313, + "tokens_seen": 1065222144 + }, + { + "epoch": 0.32, + "learning_rate": 0.0005472738214643932, + "loss": 1.3701, + "theoretical_loss": 3.65588892527888, + "tokens_seen": 1065287680 + }, + { + "epoch": 0.32, + "learning_rate": 0.0005472577733199599, + "loss": 1.3701, + "theoretical_loss": 3.655867936463459, + "tokens_seen": 1065353216 + }, + { + "epoch": 0.32, + "learning_rate": 0.0005472417251755266, + "loss": 1.4753, + "theoretical_loss": 3.655846949300635, + "tokens_seen": 1065418752 + }, + { + "epoch": 0.32, + "learning_rate": 0.0005472256770310933, + "loss": 1.3935, + "theoretical_loss": 3.6558259637901767, + "tokens_seen": 1065484288 + }, + { + "epoch": 0.32, + "learning_rate": 0.00054720962888666, + "loss": 1.388, + "theoretical_loss": 3.655804979931853, + "tokens_seen": 1065549824 + }, + { + "epoch": 0.32, + "learning_rate": 0.0005471935807422267, + "loss": 1.4013, + "theoretical_loss": 3.655783997725431, + "tokens_seen": 1065615360 + }, + { + "epoch": 0.32, + "learning_rate": 0.0005471775325977934, + "loss": 1.4627, + "theoretical_loss": 3.6557630171706803, + "tokens_seen": 1065680896 + }, + { + "epoch": 0.32, + "learning_rate": 0.0005471614844533601, + "loss": 1.5819, + "theoretical_loss": 3.655742038267369, + "tokens_seen": 1065746432 + }, + { + "epoch": 0.32, + "learning_rate": 0.0005471454363089268, + "loss": 1.4706, + "theoretical_loss": 3.655721061015265, + "tokens_seen": 1065811968 + }, + { + "epoch": 0.32, + "learning_rate": 0.0005471293881644935, + "loss": 1.4663, + "theoretical_loss": 3.6557000854141375, + "tokens_seen": 1065877504 + }, + { + "epoch": 0.32, + "learning_rate": 0.0005471133400200602, + "loss": 1.3858, + "theoretical_loss": 3.6556791114637543, + "tokens_seen": 1065943040 + }, + { + "epoch": 0.32, + "learning_rate": 0.0005470972918756269, + "loss": 1.4785, + "theoretical_loss": 3.6556581391638847, + "tokens_seen": 1066008576 + }, + { + "epoch": 0.32, + "learning_rate": 0.0005470812437311936, + "loss": 1.3829, + "theoretical_loss": 3.655637168514297, + "tokens_seen": 1066074112 + }, + { + "epoch": 0.32, + "learning_rate": 0.0005470651955867603, + "loss": 1.4257, + "theoretical_loss": 3.6556161995147605, + "tokens_seen": 1066139648 + }, + { + "epoch": 0.32, + "learning_rate": 0.000547049147442327, + "loss": 1.4442, + "theoretical_loss": 3.6555952321650436, + "tokens_seen": 1066205184 + }, + { + "epoch": 0.32, + "learning_rate": 0.0005470330992978937, + "loss": 1.269, + "theoretical_loss": 3.6555742664649147, + "tokens_seen": 1066270720 + }, + { + "epoch": 0.32, + "learning_rate": 0.0005470170511534604, + "loss": 1.4605, + "theoretical_loss": 3.6555533024141424, + "tokens_seen": 1066336256 + }, + { + "epoch": 0.32, + "learning_rate": 0.0005470010030090271, + "loss": 1.5285, + "theoretical_loss": 3.655532340012497, + "tokens_seen": 1066401792 + }, + { + "epoch": 0.32, + "learning_rate": 0.0005469849548645938, + "loss": 1.329, + "theoretical_loss": 3.655511379259746, + "tokens_seen": 1066467328 + }, + { + "epoch": 0.32, + "learning_rate": 0.0005469689067201605, + "loss": 1.3973, + "theoretical_loss": 3.655490420155659, + "tokens_seen": 1066532864 + }, + { + "epoch": 0.32, + "learning_rate": 0.0005469528585757272, + "loss": 1.4629, + "theoretical_loss": 3.655469462700005, + "tokens_seen": 1066598400 + }, + { + "epoch": 0.32, + "learning_rate": 0.0005469368104312939, + "loss": 1.3739, + "theoretical_loss": 3.6554485068925535, + "tokens_seen": 1066663936 + }, + { + "epoch": 0.32, + "learning_rate": 0.0005469207622868606, + "loss": 1.4568, + "theoretical_loss": 3.655427552733072, + "tokens_seen": 1066729472 + }, + { + "epoch": 0.32, + "learning_rate": 0.0005469047141424273, + "loss": 1.4085, + "theoretical_loss": 3.6554066002213323, + "tokens_seen": 1066795008 + }, + { + "epoch": 0.32, + "learning_rate": 0.000546888665997994, + "loss": 1.4406, + "theoretical_loss": 3.6553856493571013, + "tokens_seen": 1066860544 + }, + { + "epoch": 0.32, + "learning_rate": 0.0005468726178535607, + "loss": 1.4893, + "theoretical_loss": 3.6553647001401495, + "tokens_seen": 1066926080 + }, + { + "epoch": 0.32, + "learning_rate": 0.0005468565697091274, + "loss": 1.4333, + "theoretical_loss": 3.6553437525702455, + "tokens_seen": 1066991616 + }, + { + "epoch": 0.32, + "learning_rate": 0.0005468405215646941, + "loss": 1.4264, + "theoretical_loss": 3.6553228066471593, + "tokens_seen": 1067057152 + }, + { + "epoch": 0.32, + "learning_rate": 0.0005468244734202608, + "loss": 1.4636, + "theoretical_loss": 3.65530186237066, + "tokens_seen": 1067122688 + }, + { + "epoch": 0.32, + "learning_rate": 0.0005468084252758275, + "loss": 1.4438, + "theoretical_loss": 3.655280919740517, + "tokens_seen": 1067188224 + }, + { + "epoch": 0.32, + "learning_rate": 0.0005467923771313942, + "loss": 1.3493, + "theoretical_loss": 3.6552599787564994, + "tokens_seen": 1067253760 + }, + { + "epoch": 0.32, + "learning_rate": 0.000546776328986961, + "loss": 1.3959, + "theoretical_loss": 3.655239039418378, + "tokens_seen": 1067319296 + }, + { + "epoch": 0.32, + "learning_rate": 0.0005467602808425276, + "loss": 1.2737, + "theoretical_loss": 3.6552181017259207, + "tokens_seen": 1067384832 + }, + { + "epoch": 0.32, + "learning_rate": 0.0005467442326980943, + "loss": 1.506, + "theoretical_loss": 3.6551971656788984, + "tokens_seen": 1067450368 + }, + { + "epoch": 0.32, + "learning_rate": 0.000546728184553661, + "loss": 1.3868, + "theoretical_loss": 3.65517623127708, + "tokens_seen": 1067515904 + }, + { + "epoch": 0.32, + "learning_rate": 0.0005467121364092278, + "loss": 1.3683, + "theoretical_loss": 3.6551552985202362, + "tokens_seen": 1067581440 + }, + { + "epoch": 0.32, + "learning_rate": 0.0005466960882647944, + "loss": 1.451, + "theoretical_loss": 3.6551343674081362, + "tokens_seen": 1067646976 + }, + { + "epoch": 0.32, + "learning_rate": 0.0005466800401203612, + "loss": 1.4378, + "theoretical_loss": 3.655113437940549, + "tokens_seen": 1067712512 + }, + { + "epoch": 0.32, + "learning_rate": 0.0005466639919759278, + "loss": 1.3835, + "theoretical_loss": 3.655092510117246, + "tokens_seen": 1067778048 + }, + { + "epoch": 0.32, + "learning_rate": 0.0005466479438314946, + "loss": 1.4796, + "theoretical_loss": 3.655071583937996, + "tokens_seen": 1067843584 + }, + { + "epoch": 0.32, + "learning_rate": 0.0005466318956870612, + "loss": 1.4239, + "theoretical_loss": 3.65505065940257, + "tokens_seen": 1067909120 + }, + { + "epoch": 0.32, + "learning_rate": 0.000546615847542628, + "loss": 1.4424, + "theoretical_loss": 3.6550297365107367, + "tokens_seen": 1067974656 + }, + { + "epoch": 0.32, + "learning_rate": 0.0005465997993981946, + "loss": 1.3637, + "theoretical_loss": 3.655008815262267, + "tokens_seen": 1068040192 + }, + { + "epoch": 0.32, + "learning_rate": 0.0005465837512537614, + "loss": 1.4793, + "theoretical_loss": 3.65498789565693, + "tokens_seen": 1068105728 + }, + { + "epoch": 0.32, + "learning_rate": 0.000546567703109328, + "loss": 1.4324, + "theoretical_loss": 3.6549669776944977, + "tokens_seen": 1068171264 + }, + { + "epoch": 0.32, + "objective/train/docs_used": 391774, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.496579885482788, + "objective/train/theoretical_loss": 3.654946061374739, + "objective/train/tokens_used": 1088696800, + "theoretical_loss": 3.654946061374739, + "tokens_seen": 1068236800 + }, + { + "epoch": 0.32, + "learning_rate": 0.0005465516549648948, + "loss": 1.4165, + "theoretical_loss": 3.654946061374739, + "tokens_seen": 1068236800 + }, + { + "epoch": 0.32, + "learning_rate": 0.0005465356068204614, + "loss": 1.4305, + "theoretical_loss": 3.6549251466974244, + "tokens_seen": 1068302336 + }, + { + "epoch": 0.32, + "learning_rate": 0.0005465195586760282, + "loss": 1.3263, + "theoretical_loss": 3.6549042336623234, + "tokens_seen": 1068367872 + }, + { + "epoch": 0.32, + "learning_rate": 0.0005465035105315948, + "loss": 1.4678, + "theoretical_loss": 3.6548833222692076, + "tokens_seen": 1068433408 + }, + { + "epoch": 0.32, + "learning_rate": 0.0005464874623871616, + "loss": 1.4394, + "theoretical_loss": 3.654862412517847, + "tokens_seen": 1068498944 + }, + { + "epoch": 0.32, + "learning_rate": 0.0005464714142427282, + "loss": 1.4094, + "theoretical_loss": 3.654841504408012, + "tokens_seen": 1068564480 + }, + { + "epoch": 0.32, + "learning_rate": 0.000546455366098295, + "loss": 1.4045, + "theoretical_loss": 3.6548205979394726, + "tokens_seen": 1068630016 + }, + { + "epoch": 0.32, + "learning_rate": 0.0005464393179538616, + "loss": 1.4812, + "theoretical_loss": 3.6547996931119995, + "tokens_seen": 1068695552 + }, + { + "epoch": 0.32, + "learning_rate": 0.0005464232698094283, + "loss": 1.4546, + "theoretical_loss": 3.6547787899253636, + "tokens_seen": 1068761088 + }, + { + "epoch": 0.32, + "learning_rate": 0.000546407221664995, + "loss": 1.3254, + "theoretical_loss": 3.6547578883793355, + "tokens_seen": 1068826624 + }, + { + "epoch": 0.32, + "learning_rate": 0.0005463911735205617, + "loss": 1.3195, + "theoretical_loss": 3.6547369884736858, + "tokens_seen": 1068892160 + }, + { + "epoch": 0.32, + "learning_rate": 0.0005463751253761284, + "loss": 1.4563, + "theoretical_loss": 3.6547160902081846, + "tokens_seen": 1068957696 + }, + { + "epoch": 0.32, + "learning_rate": 0.0005463590772316951, + "loss": 1.4331, + "theoretical_loss": 3.6546951935826035, + "tokens_seen": 1069023232 + }, + { + "epoch": 0.32, + "learning_rate": 0.0005463430290872618, + "loss": 1.4818, + "theoretical_loss": 3.6546742985967127, + "tokens_seen": 1069088768 + }, + { + "epoch": 0.32, + "learning_rate": 0.0005463269809428285, + "loss": 1.4668, + "theoretical_loss": 3.654653405250283, + "tokens_seen": 1069154304 + }, + { + "epoch": 0.32, + "learning_rate": 0.0005463109327983952, + "loss": 1.3459, + "theoretical_loss": 3.6546325135430857, + "tokens_seen": 1069219840 + }, + { + "epoch": 0.32, + "learning_rate": 0.0005462948846539619, + "loss": 1.4682, + "theoretical_loss": 3.6546116234748913, + "tokens_seen": 1069285376 + }, + { + "epoch": 0.32, + "learning_rate": 0.0005462788365095286, + "loss": 1.5494, + "theoretical_loss": 3.6545907350454714, + "tokens_seen": 1069350912 + }, + { + "epoch": 0.32, + "learning_rate": 0.0005462627883650953, + "loss": 1.3549, + "theoretical_loss": 3.6545698482545963, + "tokens_seen": 1069416448 + }, + { + "epoch": 0.32, + "learning_rate": 0.000546246740220662, + "loss": 1.455, + "theoretical_loss": 3.6545489631020374, + "tokens_seen": 1069481984 + }, + { + "epoch": 0.32, + "learning_rate": 0.0005462306920762287, + "loss": 1.4068, + "theoretical_loss": 3.654528079587566, + "tokens_seen": 1069547520 + }, + { + "epoch": 0.32, + "learning_rate": 0.0005462146439317954, + "loss": 1.3954, + "theoretical_loss": 3.654507197710953, + "tokens_seen": 1069613056 + }, + { + "epoch": 0.32, + "learning_rate": 0.0005461985957873621, + "loss": 1.4665, + "theoretical_loss": 3.65448631747197, + "tokens_seen": 1069678592 + }, + { + "epoch": 0.32, + "learning_rate": 0.0005461825476429288, + "loss": 1.4729, + "theoretical_loss": 3.6544654388703877, + "tokens_seen": 1069744128 + }, + { + "epoch": 0.32, + "learning_rate": 0.0005461664994984955, + "loss": 1.3015, + "theoretical_loss": 3.6544445619059776, + "tokens_seen": 1069809664 + }, + { + "epoch": 0.32, + "learning_rate": 0.0005461504513540622, + "loss": 1.3642, + "theoretical_loss": 3.654423686578511, + "tokens_seen": 1069875200 + }, + { + "epoch": 0.32, + "learning_rate": 0.0005461344032096289, + "loss": 1.3854, + "theoretical_loss": 3.65440281288776, + "tokens_seen": 1069940736 + }, + { + "epoch": 0.32, + "learning_rate": 0.0005461183550651956, + "loss": 1.4192, + "theoretical_loss": 3.6543819408334945, + "tokens_seen": 1070006272 + }, + { + "epoch": 0.32, + "learning_rate": 0.0005461023069207623, + "loss": 1.3925, + "theoretical_loss": 3.6543610704154874, + "tokens_seen": 1070071808 + }, + { + "epoch": 0.32, + "learning_rate": 0.000546086258776329, + "loss": 1.4243, + "theoretical_loss": 3.6543402016335094, + "tokens_seen": 1070137344 + }, + { + "epoch": 0.32, + "learning_rate": 0.0005460702106318957, + "loss": 1.4888, + "theoretical_loss": 3.6543193344873326, + "tokens_seen": 1070202880 + }, + { + "epoch": 0.32, + "learning_rate": 0.0005460541624874624, + "loss": 1.4413, + "theoretical_loss": 3.6542984689767284, + "tokens_seen": 1070268416 + }, + { + "epoch": 0.32, + "learning_rate": 0.0005460381143430291, + "loss": 1.4021, + "theoretical_loss": 3.654277605101468, + "tokens_seen": 1070333952 + }, + { + "epoch": 0.32, + "learning_rate": 0.0005460220661985958, + "loss": 1.4831, + "theoretical_loss": 3.6542567428613237, + "tokens_seen": 1070399488 + }, + { + "epoch": 0.32, + "learning_rate": 0.0005460060180541625, + "loss": 1.4179, + "theoretical_loss": 3.654235882256067, + "tokens_seen": 1070465024 + }, + { + "epoch": 0.32, + "learning_rate": 0.0005459899699097292, + "loss": 1.5227, + "theoretical_loss": 3.6542150232854707, + "tokens_seen": 1070530560 + }, + { + "epoch": 0.32, + "learning_rate": 0.0005459739217652959, + "loss": 1.4982, + "theoretical_loss": 3.6541941659493054, + "tokens_seen": 1070596096 + }, + { + "epoch": 0.32, + "learning_rate": 0.0005459578736208626, + "loss": 1.5187, + "theoretical_loss": 3.6541733102473426, + "tokens_seen": 1070661632 + }, + { + "epoch": 0.32, + "learning_rate": 0.0005459418254764293, + "loss": 1.4915, + "theoretical_loss": 3.654152456179356, + "tokens_seen": 1070727168 + }, + { + "epoch": 0.32, + "learning_rate": 0.000545925777331996, + "loss": 1.327, + "theoretical_loss": 3.6541316037451157, + "tokens_seen": 1070792704 + }, + { + "epoch": 0.32, + "learning_rate": 0.0005459097291875627, + "loss": 1.4112, + "theoretical_loss": 3.654110752944395, + "tokens_seen": 1070858240 + }, + { + "epoch": 0.32, + "learning_rate": 0.0005458936810431294, + "loss": 1.4933, + "theoretical_loss": 3.6540899037769656, + "tokens_seen": 1070923776 + }, + { + "epoch": 0.32, + "learning_rate": 0.0005458776328986961, + "loss": 1.4403, + "theoretical_loss": 3.6540690562425997, + "tokens_seen": 1070989312 + }, + { + "epoch": 0.32, + "learning_rate": 0.0005458615847542628, + "loss": 1.4763, + "theoretical_loss": 3.6540482103410694, + "tokens_seen": 1071054848 + }, + { + "epoch": 0.32, + "learning_rate": 0.0005458455366098295, + "loss": 1.4161, + "theoretical_loss": 3.654027366072147, + "tokens_seen": 1071120384 + }, + { + "epoch": 0.32, + "learning_rate": 0.0005458294884653962, + "loss": 1.4098, + "theoretical_loss": 3.654006523435604, + "tokens_seen": 1071185920 + }, + { + "epoch": 0.32, + "learning_rate": 0.0005458134403209629, + "loss": 1.5442, + "theoretical_loss": 3.6539856824312142, + "tokens_seen": 1071251456 + }, + { + "epoch": 0.32, + "learning_rate": 0.0005457973921765296, + "loss": 1.3663, + "theoretical_loss": 3.6539648430587484, + "tokens_seen": 1071316992 + }, + { + "epoch": 0.32, + "learning_rate": 0.0005457813440320963, + "loss": 1.4353, + "theoretical_loss": 3.65394400531798, + "tokens_seen": 1071382528 + }, + { + "epoch": 0.32, + "learning_rate": 0.000545765295887663, + "loss": 1.3631, + "theoretical_loss": 3.653923169208681, + "tokens_seen": 1071448064 + }, + { + "epoch": 0.32, + "objective/train/docs_used": 393037, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.4863834381103516, + "objective/train/theoretical_loss": 3.6539023347306245, + "objective/train/tokens_used": 1091973600, + "theoretical_loss": 3.6539023347306245, + "tokens_seen": 1071513600 + }, + { + "epoch": 0.32, + "learning_rate": 0.0005457492477432297, + "loss": 1.4324, + "theoretical_loss": 3.6539023347306245, + "tokens_seen": 1071513600 + }, + { + "epoch": 0.32, + "learning_rate": 0.0005457331995987964, + "loss": 1.432, + "theoretical_loss": 3.653881501883582, + "tokens_seen": 1071579136 + }, + { + "epoch": 0.32, + "learning_rate": 0.0005457171514543631, + "loss": 1.4237, + "theoretical_loss": 3.653860670667327, + "tokens_seen": 1071644672 + }, + { + "epoch": 0.32, + "learning_rate": 0.0005457011033099298, + "loss": 1.4789, + "theoretical_loss": 3.6538398410816315, + "tokens_seen": 1071710208 + }, + { + "epoch": 0.32, + "learning_rate": 0.0005456850551654965, + "loss": 1.5034, + "theoretical_loss": 3.6538190131262684, + "tokens_seen": 1071775744 + }, + { + "epoch": 0.32, + "learning_rate": 0.0005456690070210632, + "loss": 1.3583, + "theoretical_loss": 3.653798186801011, + "tokens_seen": 1071841280 + }, + { + "epoch": 0.32, + "learning_rate": 0.0005456529588766299, + "loss": 1.3664, + "theoretical_loss": 3.6537773621056306, + "tokens_seen": 1071906816 + }, + { + "epoch": 0.32, + "learning_rate": 0.0005456369107321966, + "loss": 1.4886, + "theoretical_loss": 3.653756539039902, + "tokens_seen": 1071972352 + }, + { + "epoch": 0.32, + "learning_rate": 0.0005456208625877633, + "loss": 1.3455, + "theoretical_loss": 3.6537357176035963, + "tokens_seen": 1072037888 + }, + { + "epoch": 0.32, + "learning_rate": 0.00054560481444333, + "loss": 1.4807, + "theoretical_loss": 3.653714897796487, + "tokens_seen": 1072103424 + }, + { + "epoch": 0.32, + "learning_rate": 0.0005455887662988967, + "loss": 1.4181, + "theoretical_loss": 3.6536940796183472, + "tokens_seen": 1072168960 + }, + { + "epoch": 0.32, + "learning_rate": 0.0005455727181544634, + "loss": 1.467, + "theoretical_loss": 3.65367326306895, + "tokens_seen": 1072234496 + }, + { + "epoch": 0.32, + "learning_rate": 0.0005455566700100301, + "loss": 1.4646, + "theoretical_loss": 3.653652448148068, + "tokens_seen": 1072300032 + }, + { + "epoch": 0.32, + "learning_rate": 0.0005455406218655968, + "loss": 1.4233, + "theoretical_loss": 3.6536316348554747, + "tokens_seen": 1072365568 + }, + { + "epoch": 0.32, + "learning_rate": 0.0005455245737211635, + "loss": 1.4083, + "theoretical_loss": 3.653610823190943, + "tokens_seen": 1072431104 + }, + { + "epoch": 0.32, + "learning_rate": 0.0005455085255767302, + "loss": 1.3952, + "theoretical_loss": 3.653590013154246, + "tokens_seen": 1072496640 + }, + { + "epoch": 0.33, + "learning_rate": 0.0005454924774322969, + "loss": 1.4554, + "theoretical_loss": 3.653569204745157, + "tokens_seen": 1072562176 + }, + { + "epoch": 0.33, + "learning_rate": 0.0005454764292878636, + "loss": 1.3905, + "theoretical_loss": 3.6535483979634495, + "tokens_seen": 1072627712 + }, + { + "epoch": 0.33, + "learning_rate": 0.0005454603811434303, + "loss": 1.3677, + "theoretical_loss": 3.6535275928088966, + "tokens_seen": 1072693248 + }, + { + "epoch": 0.33, + "learning_rate": 0.000545444332998997, + "loss": 1.4794, + "theoretical_loss": 3.6535067892812716, + "tokens_seen": 1072758784 + }, + { + "epoch": 0.33, + "learning_rate": 0.0005454282848545637, + "loss": 1.3991, + "theoretical_loss": 3.6534859873803476, + "tokens_seen": 1072824320 + }, + { + "epoch": 0.33, + "learning_rate": 0.0005454122367101304, + "loss": 1.4968, + "theoretical_loss": 3.6534651871058985, + "tokens_seen": 1072889856 + }, + { + "epoch": 0.33, + "learning_rate": 0.0005453961885656971, + "loss": 1.4315, + "theoretical_loss": 3.653444388457698, + "tokens_seen": 1072955392 + }, + { + "epoch": 0.33, + "learning_rate": 0.0005453801404212638, + "loss": 1.4792, + "theoretical_loss": 3.6534235914355193, + "tokens_seen": 1073020928 + }, + { + "epoch": 0.33, + "learning_rate": 0.0005453640922768305, + "loss": 1.4419, + "theoretical_loss": 3.6534027960391358, + "tokens_seen": 1073086464 + }, + { + "epoch": 0.33, + "learning_rate": 0.0005453480441323973, + "loss": 1.4709, + "theoretical_loss": 3.653382002268321, + "tokens_seen": 1073152000 + }, + { + "epoch": 0.33, + "learning_rate": 0.000545331995987964, + "loss": 1.3564, + "theoretical_loss": 3.653361210122849, + "tokens_seen": 1073217536 + }, + { + "epoch": 0.33, + "learning_rate": 0.0005453159478435307, + "loss": 1.3976, + "theoretical_loss": 3.653340419602494, + "tokens_seen": 1073283072 + }, + { + "epoch": 0.33, + "learning_rate": 0.0005452998996990974, + "loss": 1.4246, + "theoretical_loss": 3.6533196307070286, + "tokens_seen": 1073348608 + }, + { + "epoch": 0.33, + "learning_rate": 0.000545283851554664, + "loss": 1.4915, + "theoretical_loss": 3.6532988434362275, + "tokens_seen": 1073414144 + }, + { + "epoch": 0.33, + "learning_rate": 0.0005452678034102308, + "loss": 1.3947, + "theoretical_loss": 3.653278057789864, + "tokens_seen": 1073479680 + }, + { + "epoch": 0.33, + "learning_rate": 0.0005452517552657975, + "loss": 1.4317, + "theoretical_loss": 3.6532572737677125, + "tokens_seen": 1073545216 + }, + { + "epoch": 0.33, + "learning_rate": 0.0005452357071213642, + "loss": 1.4077, + "theoretical_loss": 3.6532364913695465, + "tokens_seen": 1073610752 + }, + { + "epoch": 0.33, + "learning_rate": 0.0005452196589769309, + "loss": 1.4884, + "theoretical_loss": 3.6532157105951395, + "tokens_seen": 1073676288 + }, + { + "epoch": 0.33, + "learning_rate": 0.0005452036108324976, + "loss": 1.3778, + "theoretical_loss": 3.6531949314442667, + "tokens_seen": 1073741824 + }, + { + "epoch": 0.33, + "learning_rate": 0.0005451875626880643, + "loss": 1.3447, + "theoretical_loss": 3.6531741539167024, + "tokens_seen": 1073807360 + }, + { + "epoch": 0.33, + "learning_rate": 0.0005451715145436309, + "loss": 1.3718, + "theoretical_loss": 3.653153378012219, + "tokens_seen": 1073872896 + }, + { + "epoch": 0.33, + "learning_rate": 0.0005451554663991977, + "loss": 1.4501, + "theoretical_loss": 3.653132603730592, + "tokens_seen": 1073938432 + }, + { + "epoch": 0.33, + "learning_rate": 0.0005451394182547643, + "loss": 1.4528, + "theoretical_loss": 3.6531118310715947, + "tokens_seen": 1074003968 + }, + { + "epoch": 0.33, + "learning_rate": 0.0005451233701103311, + "loss": 1.4348, + "theoretical_loss": 3.6530910600350026, + "tokens_seen": 1074069504 + }, + { + "epoch": 0.33, + "learning_rate": 0.0005451073219658977, + "loss": 1.4807, + "theoretical_loss": 3.6530702906205894, + "tokens_seen": 1074135040 + }, + { + "epoch": 0.33, + "learning_rate": 0.0005450912738214645, + "loss": 1.3756, + "theoretical_loss": 3.6530495228281286, + "tokens_seen": 1074200576 + }, + { + "epoch": 0.33, + "learning_rate": 0.0005450752256770311, + "loss": 1.5044, + "theoretical_loss": 3.6530287566573962, + "tokens_seen": 1074266112 + }, + { + "epoch": 0.33, + "learning_rate": 0.0005450591775325979, + "loss": 1.5614, + "theoretical_loss": 3.653007992108165, + "tokens_seen": 1074331648 + }, + { + "epoch": 0.33, + "learning_rate": 0.0005450431293881645, + "loss": 1.3091, + "theoretical_loss": 3.6529872291802112, + "tokens_seen": 1074397184 + }, + { + "epoch": 0.33, + "learning_rate": 0.0005450270812437312, + "loss": 1.418, + "theoretical_loss": 3.6529664678733074, + "tokens_seen": 1074462720 + }, + { + "epoch": 0.33, + "learning_rate": 0.0005450110330992979, + "loss": 1.381, + "theoretical_loss": 3.65294570818723, + "tokens_seen": 1074528256 + }, + { + "epoch": 0.33, + "learning_rate": 0.0005449949849548646, + "loss": 1.4054, + "theoretical_loss": 3.6529249501217524, + "tokens_seen": 1074593792 + }, + { + "epoch": 0.33, + "learning_rate": 0.0005449789368104313, + "loss": 1.4222, + "theoretical_loss": 3.65290419367665, + "tokens_seen": 1074659328 + }, + { + "epoch": 0.33, + "learning_rate": 0.000544962888665998, + "loss": 1.4566, + "theoretical_loss": 3.652883438851697, + "tokens_seen": 1074724864 + }, + { + "epoch": 0.33, + "objective/train/docs_used": 394219, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.3403830528259277, + "objective/train/theoretical_loss": 3.652862685646668, + "objective/train/tokens_used": 1095250400, + "theoretical_loss": 3.652862685646668, + "tokens_seen": 1074790400 + }, + { + "epoch": 0.33, + "learning_rate": 0.0005449468405215647, + "loss": 1.4046, + "theoretical_loss": 3.652862685646668, + "tokens_seen": 1074790400 + }, + { + "epoch": 0.33, + "learning_rate": 0.0005449307923771314, + "loss": 1.4203, + "theoretical_loss": 3.652841934061339, + "tokens_seen": 1074855936 + }, + { + "epoch": 0.33, + "learning_rate": 0.0005449147442326981, + "loss": 1.4339, + "theoretical_loss": 3.652821184095483, + "tokens_seen": 1074921472 + }, + { + "epoch": 0.33, + "learning_rate": 0.0005448986960882648, + "loss": 1.3833, + "theoretical_loss": 3.652800435748876, + "tokens_seen": 1074987008 + }, + { + "epoch": 0.33, + "learning_rate": 0.0005448826479438315, + "loss": 1.4195, + "theoretical_loss": 3.6527796890212936, + "tokens_seen": 1075052544 + }, + { + "epoch": 0.33, + "learning_rate": 0.0005448665997993982, + "loss": 1.4058, + "theoretical_loss": 3.6527589439125094, + "tokens_seen": 1075118080 + }, + { + "epoch": 0.33, + "learning_rate": 0.0005448505516549649, + "loss": 1.524, + "theoretical_loss": 3.6527382004222995, + "tokens_seen": 1075183616 + }, + { + "epoch": 0.33, + "learning_rate": 0.0005448345035105316, + "loss": 1.4123, + "theoretical_loss": 3.6527174585504376, + "tokens_seen": 1075249152 + }, + { + "epoch": 0.33, + "learning_rate": 0.0005448184553660983, + "loss": 1.4147, + "theoretical_loss": 3.6526967182967, + "tokens_seen": 1075314688 + }, + { + "epoch": 0.33, + "learning_rate": 0.000544802407221665, + "loss": 1.4741, + "theoretical_loss": 3.652675979660862, + "tokens_seen": 1075380224 + }, + { + "epoch": 0.33, + "learning_rate": 0.0005447863590772317, + "loss": 1.4881, + "theoretical_loss": 3.652655242642698, + "tokens_seen": 1075445760 + }, + { + "epoch": 0.33, + "learning_rate": 0.0005447703109327984, + "loss": 1.3457, + "theoretical_loss": 3.652634507241984, + "tokens_seen": 1075511296 + }, + { + "epoch": 0.33, + "learning_rate": 0.0005447542627883651, + "loss": 1.5101, + "theoretical_loss": 3.6526137734584943, + "tokens_seen": 1075576832 + }, + { + "epoch": 0.33, + "learning_rate": 0.0005447382146439318, + "loss": 1.3483, + "theoretical_loss": 3.6525930412920053, + "tokens_seen": 1075642368 + }, + { + "epoch": 0.33, + "learning_rate": 0.0005447221664994985, + "loss": 1.4057, + "theoretical_loss": 3.6525723107422916, + "tokens_seen": 1075707904 + }, + { + "epoch": 0.33, + "learning_rate": 0.0005447061183550652, + "loss": 1.4625, + "theoretical_loss": 3.6525515818091296, + "tokens_seen": 1075773440 + }, + { + "epoch": 0.33, + "learning_rate": 0.0005446900702106319, + "loss": 1.37, + "theoretical_loss": 3.6525308544922934, + "tokens_seen": 1075838976 + }, + { + "epoch": 0.33, + "learning_rate": 0.0005446740220661986, + "loss": 1.4309, + "theoretical_loss": 3.6525101287915596, + "tokens_seen": 1075904512 + }, + { + "epoch": 0.33, + "learning_rate": 0.0005446579739217653, + "loss": 1.4149, + "theoretical_loss": 3.6524894047067034, + "tokens_seen": 1075970048 + }, + { + "epoch": 0.33, + "learning_rate": 0.000544641925777332, + "loss": 1.528, + "theoretical_loss": 3.6524686822375, + "tokens_seen": 1076035584 + }, + { + "epoch": 0.33, + "learning_rate": 0.0005446258776328987, + "loss": 1.3631, + "theoretical_loss": 3.652447961383726, + "tokens_seen": 1076101120 + }, + { + "epoch": 0.33, + "learning_rate": 0.0005446098294884654, + "loss": 1.4434, + "theoretical_loss": 3.6524272421451567, + "tokens_seen": 1076166656 + }, + { + "epoch": 0.33, + "learning_rate": 0.0005445937813440321, + "loss": 1.4046, + "theoretical_loss": 3.6524065245215676, + "tokens_seen": 1076232192 + }, + { + "epoch": 0.33, + "learning_rate": 0.0005445777331995988, + "loss": 1.336, + "theoretical_loss": 3.652385808512734, + "tokens_seen": 1076297728 + }, + { + "epoch": 0.33, + "learning_rate": 0.0005445616850551655, + "loss": 1.4792, + "theoretical_loss": 3.6523650941184327, + "tokens_seen": 1076363264 + }, + { + "epoch": 0.33, + "learning_rate": 0.0005445456369107322, + "loss": 1.4331, + "theoretical_loss": 3.6523443813384393, + "tokens_seen": 1076428800 + }, + { + "epoch": 0.33, + "learning_rate": 0.0005445295887662989, + "loss": 1.3383, + "theoretical_loss": 3.65232367017253, + "tokens_seen": 1076494336 + }, + { + "epoch": 0.33, + "learning_rate": 0.0005445135406218656, + "loss": 1.45, + "theoretical_loss": 3.6523029606204798, + "tokens_seen": 1076559872 + }, + { + "epoch": 0.33, + "learning_rate": 0.0005444974924774323, + "loss": 1.4958, + "theoretical_loss": 3.6522822526820655, + "tokens_seen": 1076625408 + }, + { + "epoch": 0.33, + "learning_rate": 0.000544481444332999, + "loss": 1.389, + "theoretical_loss": 3.652261546357063, + "tokens_seen": 1076690944 + }, + { + "epoch": 0.33, + "learning_rate": 0.0005444653961885657, + "loss": 1.4063, + "theoretical_loss": 3.6522408416452485, + "tokens_seen": 1076756480 + }, + { + "epoch": 0.33, + "learning_rate": 0.0005444493480441324, + "loss": 1.4261, + "theoretical_loss": 3.6522201385463977, + "tokens_seen": 1076822016 + }, + { + "epoch": 0.33, + "learning_rate": 0.0005444332998996991, + "loss": 1.4232, + "theoretical_loss": 3.6521994370602875, + "tokens_seen": 1076887552 + }, + { + "epoch": 0.33, + "learning_rate": 0.0005444172517552658, + "loss": 1.4192, + "theoretical_loss": 3.6521787371866936, + "tokens_seen": 1076953088 + }, + { + "epoch": 0.33, + "learning_rate": 0.0005444012036108325, + "loss": 1.3551, + "theoretical_loss": 3.6521580389253927, + "tokens_seen": 1077018624 + }, + { + "epoch": 0.33, + "learning_rate": 0.0005443851554663992, + "loss": 1.4582, + "theoretical_loss": 3.65213734227616, + "tokens_seen": 1077084160 + }, + { + "epoch": 0.33, + "learning_rate": 0.0005443691073219659, + "loss": 1.4071, + "theoretical_loss": 3.652116647238773, + "tokens_seen": 1077149696 + }, + { + "epoch": 0.33, + "learning_rate": 0.0005443530591775326, + "loss": 1.5263, + "theoretical_loss": 3.6520959538130087, + "tokens_seen": 1077215232 + }, + { + "epoch": 0.33, + "learning_rate": 0.0005443370110330993, + "loss": 1.5075, + "theoretical_loss": 3.652075261998642, + "tokens_seen": 1077280768 + }, + { + "epoch": 0.33, + "learning_rate": 0.000544320962888666, + "loss": 1.4832, + "theoretical_loss": 3.65205457179545, + "tokens_seen": 1077346304 + }, + { + "epoch": 0.33, + "learning_rate": 0.0005443049147442327, + "loss": 1.3992, + "theoretical_loss": 3.6520338832032095, + "tokens_seen": 1077411840 + }, + { + "epoch": 0.33, + "learning_rate": 0.0005442888665997994, + "loss": 1.3733, + "theoretical_loss": 3.652013196221697, + "tokens_seen": 1077477376 + }, + { + "epoch": 0.33, + "learning_rate": 0.0005442728184553661, + "loss": 1.4433, + "theoretical_loss": 3.6519925108506888, + "tokens_seen": 1077542912 + }, + { + "epoch": 0.33, + "learning_rate": 0.0005442567703109328, + "loss": 1.4379, + "theoretical_loss": 3.6519718270899624, + "tokens_seen": 1077608448 + }, + { + "epoch": 0.33, + "learning_rate": 0.0005442407221664995, + "loss": 1.3355, + "theoretical_loss": 3.651951144939294, + "tokens_seen": 1077673984 + }, + { + "epoch": 0.33, + "learning_rate": 0.0005442246740220662, + "loss": 1.4118, + "theoretical_loss": 3.6519304643984594, + "tokens_seen": 1077739520 + }, + { + "epoch": 0.33, + "learning_rate": 0.0005442086258776329, + "loss": 1.3416, + "theoretical_loss": 3.651909785467237, + "tokens_seen": 1077805056 + }, + { + "epoch": 0.33, + "learning_rate": 0.0005441925777331996, + "loss": 1.4317, + "theoretical_loss": 3.651889108145403, + "tokens_seen": 1077870592 + }, + { + "epoch": 0.33, + "learning_rate": 0.0005441765295887663, + "loss": 1.4858, + "theoretical_loss": 3.6518684324327344, + "tokens_seen": 1077936128 + }, + { + "epoch": 0.33, + "learning_rate": 0.000544160481444333, + "loss": 1.5066, + "theoretical_loss": 3.6518477583290077, + "tokens_seen": 1078001664 + }, + { + "epoch": 0.33, + "objective/train/docs_used": 395298, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.3336665630340576, + "objective/train/theoretical_loss": 3.6518270858340007, + "objective/train/tokens_used": 1098527200, + "theoretical_loss": 3.6518270858340007, + "tokens_seen": 1078067200 + }, + { + "epoch": 0.33, + "learning_rate": 0.0005441444332998997, + "loss": 1.3764, + "theoretical_loss": 3.6518270858340007, + "tokens_seen": 1078067200 + }, + { + "epoch": 0.33, + "learning_rate": 0.0005441283851554665, + "loss": 1.4223, + "theoretical_loss": 3.6518064149474903, + "tokens_seen": 1078132736 + }, + { + "epoch": 0.33, + "learning_rate": 0.0005441123370110331, + "loss": 1.3786, + "theoretical_loss": 3.651785745669253, + "tokens_seen": 1078198272 + }, + { + "epoch": 0.33, + "learning_rate": 0.0005440962888665999, + "loss": 1.3997, + "theoretical_loss": 3.6517650779990656, + "tokens_seen": 1078263808 + }, + { + "epoch": 0.33, + "learning_rate": 0.0005440802407221665, + "loss": 1.4931, + "theoretical_loss": 3.6517444119367064, + "tokens_seen": 1078329344 + }, + { + "epoch": 0.33, + "learning_rate": 0.0005440641925777333, + "loss": 1.4809, + "theoretical_loss": 3.651723747481952, + "tokens_seen": 1078394880 + }, + { + "epoch": 0.33, + "learning_rate": 0.0005440481444333, + "loss": 1.3285, + "theoretical_loss": 3.6517030846345797, + "tokens_seen": 1078460416 + }, + { + "epoch": 0.33, + "learning_rate": 0.0005440320962888667, + "loss": 1.5051, + "theoretical_loss": 3.6516824233943668, + "tokens_seen": 1078525952 + }, + { + "epoch": 0.33, + "learning_rate": 0.0005440160481444333, + "loss": 1.4991, + "theoretical_loss": 3.651661763761091, + "tokens_seen": 1078591488 + }, + { + "epoch": 0.33, + "learning_rate": 0.0005440000000000001, + "loss": 1.4066, + "theoretical_loss": 3.6516411057345293, + "tokens_seen": 1078657024 + }, + { + "epoch": 0.33, + "learning_rate": 0.0005439839518555668, + "loss": 1.2995, + "theoretical_loss": 3.651620449314459, + "tokens_seen": 1078722560 + }, + { + "epoch": 0.33, + "learning_rate": 0.0005439679037111335, + "loss": 1.4597, + "theoretical_loss": 3.651599794500658, + "tokens_seen": 1078788096 + }, + { + "epoch": 0.33, + "learning_rate": 0.0005439518555667002, + "loss": 1.4621, + "theoretical_loss": 3.6515791412929035, + "tokens_seen": 1078853632 + }, + { + "epoch": 0.33, + "learning_rate": 0.0005439358074222669, + "loss": 1.3615, + "theoretical_loss": 3.6515584896909736, + "tokens_seen": 1078919168 + }, + { + "epoch": 0.33, + "learning_rate": 0.0005439197592778336, + "loss": 1.538, + "theoretical_loss": 3.651537839694645, + "tokens_seen": 1078984704 + }, + { + "epoch": 0.33, + "learning_rate": 0.0005439037111334003, + "loss": 1.492, + "theoretical_loss": 3.6515171913036966, + "tokens_seen": 1079050240 + }, + { + "epoch": 0.33, + "learning_rate": 0.000543887662988967, + "loss": 1.4539, + "theoretical_loss": 3.651496544517905, + "tokens_seen": 1079115776 + }, + { + "epoch": 0.33, + "learning_rate": 0.0005438716148445337, + "loss": 1.4061, + "theoretical_loss": 3.651475899337048, + "tokens_seen": 1079181312 + }, + { + "epoch": 0.33, + "learning_rate": 0.0005438555667001004, + "loss": 1.5296, + "theoretical_loss": 3.651455255760904, + "tokens_seen": 1079246848 + }, + { + "epoch": 0.33, + "learning_rate": 0.0005438395185556671, + "loss": 1.4673, + "theoretical_loss": 3.6514346137892506, + "tokens_seen": 1079312384 + }, + { + "epoch": 0.33, + "learning_rate": 0.0005438234704112338, + "loss": 1.432, + "theoretical_loss": 3.6514139734218656, + "tokens_seen": 1079377920 + }, + { + "epoch": 0.33, + "learning_rate": 0.0005438074222668005, + "loss": 1.413, + "theoretical_loss": 3.6513933346585272, + "tokens_seen": 1079443456 + }, + { + "epoch": 0.33, + "learning_rate": 0.0005437913741223672, + "loss": 1.4858, + "theoretical_loss": 3.6513726974990126, + "tokens_seen": 1079508992 + }, + { + "epoch": 0.33, + "learning_rate": 0.0005437753259779338, + "loss": 1.4951, + "theoretical_loss": 3.6513520619431006, + "tokens_seen": 1079574528 + }, + { + "epoch": 0.33, + "learning_rate": 0.0005437592778335006, + "loss": 1.2508, + "theoretical_loss": 3.651331427990569, + "tokens_seen": 1079640064 + }, + { + "epoch": 0.33, + "learning_rate": 0.0005437432296890672, + "loss": 1.4305, + "theoretical_loss": 3.651310795641196, + "tokens_seen": 1079705600 + }, + { + "epoch": 0.33, + "learning_rate": 0.000543727181544634, + "loss": 1.3758, + "theoretical_loss": 3.65129016489476, + "tokens_seen": 1079771136 + }, + { + "epoch": 0.33, + "learning_rate": 0.0005437111334002006, + "loss": 1.3765, + "theoretical_loss": 3.6512695357510383, + "tokens_seen": 1079836672 + }, + { + "epoch": 0.33, + "learning_rate": 0.0005436950852557674, + "loss": 1.3389, + "theoretical_loss": 3.6512489082098103, + "tokens_seen": 1079902208 + }, + { + "epoch": 0.33, + "learning_rate": 0.000543679037111334, + "loss": 1.5629, + "theoretical_loss": 3.6512282822708526, + "tokens_seen": 1079967744 + }, + { + "epoch": 0.33, + "learning_rate": 0.0005436629889669008, + "loss": 1.4678, + "theoretical_loss": 3.6512076579339454, + "tokens_seen": 1080033280 + }, + { + "epoch": 0.33, + "learning_rate": 0.0005436469408224674, + "loss": 1.3994, + "theoretical_loss": 3.6511870351988662, + "tokens_seen": 1080098816 + }, + { + "epoch": 0.33, + "learning_rate": 0.0005436308926780341, + "loss": 1.5181, + "theoretical_loss": 3.6511664140653934, + "tokens_seen": 1080164352 + }, + { + "epoch": 0.33, + "learning_rate": 0.0005436148445336008, + "loss": 1.4161, + "theoretical_loss": 3.6511457945333055, + "tokens_seen": 1080229888 + }, + { + "epoch": 0.33, + "learning_rate": 0.0005435987963891675, + "loss": 1.3703, + "theoretical_loss": 3.6511251766023807, + "tokens_seen": 1080295424 + }, + { + "epoch": 0.33, + "learning_rate": 0.0005435827482447342, + "loss": 1.4807, + "theoretical_loss": 3.651104560272398, + "tokens_seen": 1080360960 + }, + { + "epoch": 0.33, + "learning_rate": 0.0005435667001003009, + "loss": 1.4392, + "theoretical_loss": 3.6510839455431356, + "tokens_seen": 1080426496 + }, + { + "epoch": 0.33, + "learning_rate": 0.0005435506519558676, + "loss": 1.4338, + "theoretical_loss": 3.6510633324143726, + "tokens_seen": 1080492032 + }, + { + "epoch": 0.33, + "learning_rate": 0.0005435346038114343, + "loss": 1.4131, + "theoretical_loss": 3.6510427208858873, + "tokens_seen": 1080557568 + }, + { + "epoch": 0.33, + "learning_rate": 0.000543518555667001, + "loss": 1.411, + "theoretical_loss": 3.651022110957459, + "tokens_seen": 1080623104 + }, + { + "epoch": 0.33, + "learning_rate": 0.0005435025075225677, + "loss": 1.372, + "theoretical_loss": 3.6510015026288656, + "tokens_seen": 1080688640 + }, + { + "epoch": 0.33, + "learning_rate": 0.0005434864593781344, + "loss": 1.5119, + "theoretical_loss": 3.6509808958998864, + "tokens_seen": 1080754176 + }, + { + "epoch": 0.33, + "learning_rate": 0.0005434704112337011, + "loss": 1.3973, + "theoretical_loss": 3.6509602907703003, + "tokens_seen": 1080819712 + }, + { + "epoch": 0.33, + "learning_rate": 0.0005434543630892678, + "loss": 1.4502, + "theoretical_loss": 3.650939687239886, + "tokens_seen": 1080885248 + }, + { + "epoch": 0.33, + "learning_rate": 0.0005434383149448345, + "loss": 1.3261, + "theoretical_loss": 3.6509190853084226, + "tokens_seen": 1080950784 + }, + { + "epoch": 0.33, + "learning_rate": 0.0005434222668004012, + "loss": 1.3832, + "theoretical_loss": 3.6508984849756887, + "tokens_seen": 1081016320 + }, + { + "epoch": 0.33, + "learning_rate": 0.0005434062186559679, + "loss": 1.5211, + "theoretical_loss": 3.6508778862414637, + "tokens_seen": 1081081856 + }, + { + "epoch": 0.33, + "learning_rate": 0.0005433901705115346, + "loss": 1.4911, + "theoretical_loss": 3.650857289105527, + "tokens_seen": 1081147392 + }, + { + "epoch": 0.33, + "learning_rate": 0.0005433741223671013, + "loss": 1.4153, + "theoretical_loss": 3.6508366935676566, + "tokens_seen": 1081212928 + }, + { + "epoch": 0.33, + "learning_rate": 0.000543358074222668, + "loss": 1.3469, + "theoretical_loss": 3.650816099627633, + "tokens_seen": 1081278464 + }, + { + "epoch": 0.33, + "objective/train/docs_used": 396489, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.5110723972320557, + "objective/train/theoretical_loss": 3.6507955072852343, + "objective/train/tokens_used": 1101804000, + "theoretical_loss": 3.6507955072852343, + "tokens_seen": 1081344000 + }, + { + "epoch": 0.33, + "learning_rate": 0.0005433420260782347, + "loss": 1.4754, + "theoretical_loss": 3.6507955072852343, + "tokens_seen": 1081344000 + }, + { + "epoch": 0.33, + "learning_rate": 0.0005433259779338014, + "loss": 1.4394, + "theoretical_loss": 3.6507749165402403, + "tokens_seen": 1081409536 + }, + { + "epoch": 0.33, + "learning_rate": 0.0005433099297893681, + "loss": 1.3606, + "theoretical_loss": 3.6507543273924306, + "tokens_seen": 1081475072 + }, + { + "epoch": 0.33, + "learning_rate": 0.0005432938816449348, + "loss": 1.3654, + "theoretical_loss": 3.6507337398415833, + "tokens_seen": 1081540608 + }, + { + "epoch": 0.33, + "learning_rate": 0.0005432778335005015, + "loss": 1.3726, + "theoretical_loss": 3.6507131538874793, + "tokens_seen": 1081606144 + }, + { + "epoch": 0.33, + "learning_rate": 0.0005432617853560682, + "loss": 1.4856, + "theoretical_loss": 3.6506925695298973, + "tokens_seen": 1081671680 + }, + { + "epoch": 0.33, + "learning_rate": 0.0005432457372116349, + "loss": 1.4031, + "theoretical_loss": 3.650671986768616, + "tokens_seen": 1081737216 + }, + { + "epoch": 0.33, + "learning_rate": 0.0005432296890672016, + "loss": 1.4854, + "theoretical_loss": 3.6506514056034165, + "tokens_seen": 1081802752 + }, + { + "epoch": 0.33, + "learning_rate": 0.0005432136409227683, + "loss": 1.6129, + "theoretical_loss": 3.6506308260340776, + "tokens_seen": 1081868288 + }, + { + "epoch": 0.33, + "learning_rate": 0.000543197592778335, + "loss": 1.449, + "theoretical_loss": 3.6506102480603784, + "tokens_seen": 1081933824 + }, + { + "epoch": 0.33, + "learning_rate": 0.0005431815446339017, + "loss": 1.3594, + "theoretical_loss": 3.6505896716820994, + "tokens_seen": 1081999360 + }, + { + "epoch": 0.33, + "learning_rate": 0.0005431654964894684, + "loss": 1.4001, + "theoretical_loss": 3.6505690968990194, + "tokens_seen": 1082064896 + }, + { + "epoch": 0.33, + "learning_rate": 0.0005431494483450351, + "loss": 1.5257, + "theoretical_loss": 3.650548523710919, + "tokens_seen": 1082130432 + }, + { + "epoch": 0.33, + "learning_rate": 0.0005431334002006018, + "loss": 1.4435, + "theoretical_loss": 3.6505279521175775, + "tokens_seen": 1082195968 + }, + { + "epoch": 0.33, + "learning_rate": 0.0005431173520561685, + "loss": 1.4237, + "theoretical_loss": 3.650507382118775, + "tokens_seen": 1082261504 + }, + { + "epoch": 0.33, + "learning_rate": 0.0005431013039117352, + "loss": 1.4815, + "theoretical_loss": 3.6504868137142905, + "tokens_seen": 1082327040 + }, + { + "epoch": 0.33, + "learning_rate": 0.0005430852557673019, + "loss": 1.3277, + "theoretical_loss": 3.650466246903905, + "tokens_seen": 1082392576 + }, + { + "epoch": 0.33, + "learning_rate": 0.0005430692076228686, + "loss": 1.4748, + "theoretical_loss": 3.6504456816873976, + "tokens_seen": 1082458112 + }, + { + "epoch": 0.33, + "learning_rate": 0.0005430531594784353, + "loss": 1.3934, + "theoretical_loss": 3.6504251180645495, + "tokens_seen": 1082523648 + }, + { + "epoch": 0.33, + "learning_rate": 0.000543037111334002, + "loss": 1.4893, + "theoretical_loss": 3.6504045560351392, + "tokens_seen": 1082589184 + }, + { + "epoch": 0.33, + "learning_rate": 0.0005430210631895687, + "loss": 1.4563, + "theoretical_loss": 3.650383995598948, + "tokens_seen": 1082654720 + }, + { + "epoch": 0.33, + "learning_rate": 0.0005430050150451354, + "loss": 1.5301, + "theoretical_loss": 3.6503634367557547, + "tokens_seen": 1082720256 + }, + { + "epoch": 0.33, + "learning_rate": 0.0005429889669007021, + "loss": 1.4649, + "theoretical_loss": 3.6503428795053408, + "tokens_seen": 1082785792 + }, + { + "epoch": 0.33, + "learning_rate": 0.0005429729187562688, + "loss": 1.3822, + "theoretical_loss": 3.6503223238474862, + "tokens_seen": 1082851328 + }, + { + "epoch": 0.33, + "learning_rate": 0.0005429568706118355, + "loss": 1.5444, + "theoretical_loss": 3.650301769781971, + "tokens_seen": 1082916864 + }, + { + "epoch": 0.33, + "learning_rate": 0.0005429408224674022, + "loss": 1.4382, + "theoretical_loss": 3.6502812173085752, + "tokens_seen": 1082982400 + }, + { + "epoch": 0.33, + "learning_rate": 0.000542924774322969, + "loss": 1.4563, + "theoretical_loss": 3.650260666427079, + "tokens_seen": 1083047936 + }, + { + "epoch": 0.33, + "learning_rate": 0.0005429087261785356, + "loss": 1.4149, + "theoretical_loss": 3.6502401171372636, + "tokens_seen": 1083113472 + }, + { + "epoch": 0.33, + "learning_rate": 0.0005428926780341024, + "loss": 1.3986, + "theoretical_loss": 3.650219569438909, + "tokens_seen": 1083179008 + }, + { + "epoch": 0.33, + "learning_rate": 0.000542876629889669, + "loss": 1.399, + "theoretical_loss": 3.650199023331796, + "tokens_seen": 1083244544 + }, + { + "epoch": 0.33, + "learning_rate": 0.0005428605817452358, + "loss": 1.3432, + "theoretical_loss": 3.650178478815704, + "tokens_seen": 1083310080 + }, + { + "epoch": 0.33, + "learning_rate": 0.0005428445336008024, + "loss": 1.3636, + "theoretical_loss": 3.6501579358904146, + "tokens_seen": 1083375616 + }, + { + "epoch": 0.33, + "learning_rate": 0.0005428284854563692, + "loss": 1.54, + "theoretical_loss": 3.650137394555708, + "tokens_seen": 1083441152 + }, + { + "epoch": 0.33, + "learning_rate": 0.0005428124373119358, + "loss": 1.3674, + "theoretical_loss": 3.6501168548113654, + "tokens_seen": 1083506688 + }, + { + "epoch": 0.33, + "learning_rate": 0.0005427963891675026, + "loss": 1.4491, + "theoretical_loss": 3.6500963166571667, + "tokens_seen": 1083572224 + }, + { + "epoch": 0.33, + "learning_rate": 0.0005427803410230692, + "loss": 1.3647, + "theoretical_loss": 3.650075780092893, + "tokens_seen": 1083637760 + }, + { + "epoch": 0.33, + "learning_rate": 0.000542764292878636, + "loss": 1.5061, + "theoretical_loss": 3.6500552451183252, + "tokens_seen": 1083703296 + }, + { + "epoch": 0.33, + "learning_rate": 0.0005427482447342026, + "loss": 1.4195, + "theoretical_loss": 3.650034711733244, + "tokens_seen": 1083768832 + }, + { + "epoch": 0.33, + "learning_rate": 0.0005427321965897694, + "loss": 1.377, + "theoretical_loss": 3.6500141799374304, + "tokens_seen": 1083834368 + }, + { + "epoch": 0.33, + "learning_rate": 0.000542716148445336, + "loss": 1.4311, + "theoretical_loss": 3.649993649730665, + "tokens_seen": 1083899904 + }, + { + "epoch": 0.33, + "learning_rate": 0.0005427001003009028, + "loss": 1.4852, + "theoretical_loss": 3.649973121112729, + "tokens_seen": 1083965440 + }, + { + "epoch": 0.33, + "learning_rate": 0.0005426840521564694, + "loss": 1.4544, + "theoretical_loss": 3.649952594083403, + "tokens_seen": 1084030976 + }, + { + "epoch": 0.33, + "learning_rate": 0.0005426680040120362, + "loss": 1.4441, + "theoretical_loss": 3.6499320686424683, + "tokens_seen": 1084096512 + }, + { + "epoch": 0.33, + "learning_rate": 0.0005426519558676028, + "loss": 1.4358, + "theoretical_loss": 3.649911544789706, + "tokens_seen": 1084162048 + }, + { + "epoch": 0.33, + "learning_rate": 0.0005426359077231696, + "loss": 1.4088, + "theoretical_loss": 3.6498910225248977, + "tokens_seen": 1084227584 + }, + { + "epoch": 0.33, + "learning_rate": 0.0005426198595787363, + "loss": 1.4252, + "theoretical_loss": 3.649870501847824, + "tokens_seen": 1084293120 + }, + { + "epoch": 0.33, + "learning_rate": 0.000542603811434303, + "loss": 1.4788, + "theoretical_loss": 3.6498499827582664, + "tokens_seen": 1084358656 + }, + { + "epoch": 0.33, + "learning_rate": 0.0005425877632898697, + "loss": 1.3924, + "theoretical_loss": 3.6498294652560053, + "tokens_seen": 1084424192 + }, + { + "epoch": 0.33, + "learning_rate": 0.0005425717151454364, + "loss": 1.3523, + "theoretical_loss": 3.649808949340823, + "tokens_seen": 1084489728 + }, + { + "epoch": 0.33, + "learning_rate": 0.000542555667001003, + "loss": 1.3912, + "theoretical_loss": 3.6497884350125007, + "tokens_seen": 1084555264 + }, + { + "epoch": 0.33, + "objective/train/docs_used": 397685, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.3136082887649536, + "objective/train/theoretical_loss": 3.649767922270819, + "objective/train/tokens_used": 1105080800, + "theoretical_loss": 3.649767922270819, + "tokens_seen": 1084620800 + }, + { + "epoch": 0.33, + "learning_rate": 0.0005425396188565698, + "loss": 1.429, + "theoretical_loss": 3.649767922270819, + "tokens_seen": 1084620800 + }, + { + "epoch": 0.33, + "learning_rate": 0.0005425235707121365, + "loss": 1.496, + "theoretical_loss": 3.64974741111556, + "tokens_seen": 1084686336 + }, + { + "epoch": 0.33, + "learning_rate": 0.0005425075225677032, + "loss": 1.4487, + "theoretical_loss": 3.6497269015465053, + "tokens_seen": 1084751872 + }, + { + "epoch": 0.33, + "learning_rate": 0.0005424914744232699, + "loss": 1.4593, + "theoretical_loss": 3.649706393563436, + "tokens_seen": 1084817408 + }, + { + "epoch": 0.33, + "learning_rate": 0.0005424754262788366, + "loss": 1.5509, + "theoretical_loss": 3.649685887166134, + "tokens_seen": 1084882944 + }, + { + "epoch": 0.33, + "learning_rate": 0.0005424593781344033, + "loss": 1.5098, + "theoretical_loss": 3.6496653823543816, + "tokens_seen": 1084948480 + }, + { + "epoch": 0.33, + "learning_rate": 0.00054244332998997, + "loss": 1.5173, + "theoretical_loss": 3.6496448791279583, + "tokens_seen": 1085014016 + }, + { + "epoch": 0.33, + "learning_rate": 0.0005424272818455367, + "loss": 1.3866, + "theoretical_loss": 3.6496243774866475, + "tokens_seen": 1085079552 + }, + { + "epoch": 0.33, + "learning_rate": 0.0005424112337011034, + "loss": 1.3771, + "theoretical_loss": 3.6496038774302306, + "tokens_seen": 1085145088 + }, + { + "epoch": 0.33, + "learning_rate": 0.0005423951855566701, + "loss": 1.4353, + "theoretical_loss": 3.6495833789584893, + "tokens_seen": 1085210624 + }, + { + "epoch": 0.33, + "learning_rate": 0.0005423791374122367, + "loss": 1.3537, + "theoretical_loss": 3.649562882071205, + "tokens_seen": 1085276160 + }, + { + "epoch": 0.33, + "learning_rate": 0.0005423630892678035, + "loss": 1.4831, + "theoretical_loss": 3.64954238676816, + "tokens_seen": 1085341696 + }, + { + "epoch": 0.33, + "learning_rate": 0.0005423470411233701, + "loss": 1.4349, + "theoretical_loss": 3.649521893049137, + "tokens_seen": 1085407232 + }, + { + "epoch": 0.33, + "learning_rate": 0.0005423309929789369, + "loss": 1.5171, + "theoretical_loss": 3.649501400913916, + "tokens_seen": 1085472768 + }, + { + "epoch": 0.33, + "learning_rate": 0.0005423149448345035, + "loss": 1.4062, + "theoretical_loss": 3.6494809103622803, + "tokens_seen": 1085538304 + }, + { + "epoch": 0.33, + "learning_rate": 0.0005422988966900703, + "loss": 1.4816, + "theoretical_loss": 3.6494604213940125, + "tokens_seen": 1085603840 + }, + { + "epoch": 0.33, + "learning_rate": 0.0005422828485456369, + "loss": 1.4457, + "theoretical_loss": 3.649439934008893, + "tokens_seen": 1085669376 + }, + { + "epoch": 0.33, + "learning_rate": 0.0005422668004012037, + "loss": 1.413, + "theoretical_loss": 3.649419448206705, + "tokens_seen": 1085734912 + }, + { + "epoch": 0.33, + "learning_rate": 0.0005422507522567703, + "loss": 1.4008, + "theoretical_loss": 3.6493989639872306, + "tokens_seen": 1085800448 + }, + { + "epoch": 0.33, + "learning_rate": 0.000542234704112337, + "loss": 1.4375, + "theoretical_loss": 3.6493784813502517, + "tokens_seen": 1085865984 + }, + { + "epoch": 0.33, + "learning_rate": 0.0005422186559679037, + "loss": 1.4335, + "theoretical_loss": 3.649358000295551, + "tokens_seen": 1085931520 + }, + { + "epoch": 0.33, + "learning_rate": 0.0005422026078234704, + "loss": 1.4332, + "theoretical_loss": 3.64933752082291, + "tokens_seen": 1085997056 + }, + { + "epoch": 0.33, + "learning_rate": 0.0005421865596790371, + "loss": 1.3956, + "theoretical_loss": 3.649317042932112, + "tokens_seen": 1086062592 + }, + { + "epoch": 0.33, + "learning_rate": 0.0005421705115346038, + "loss": 1.3882, + "theoretical_loss": 3.649296566622939, + "tokens_seen": 1086128128 + }, + { + "epoch": 0.33, + "learning_rate": 0.0005421544633901705, + "loss": 1.3168, + "theoretical_loss": 3.6492760918951728, + "tokens_seen": 1086193664 + }, + { + "epoch": 0.33, + "learning_rate": 0.0005421384152457372, + "loss": 1.4935, + "theoretical_loss": 3.6492556187485965, + "tokens_seen": 1086259200 + }, + { + "epoch": 0.33, + "learning_rate": 0.0005421223671013039, + "loss": 1.3994, + "theoretical_loss": 3.649235147182993, + "tokens_seen": 1086324736 + }, + { + "epoch": 0.33, + "learning_rate": 0.0005421063189568706, + "loss": 1.4729, + "theoretical_loss": 3.649214677198144, + "tokens_seen": 1086390272 + }, + { + "epoch": 0.33, + "learning_rate": 0.0005420902708124373, + "loss": 1.4197, + "theoretical_loss": 3.6491942087938325, + "tokens_seen": 1086455808 + }, + { + "epoch": 0.33, + "learning_rate": 0.000542074222668004, + "loss": 1.4018, + "theoretical_loss": 3.649173741969841, + "tokens_seen": 1086521344 + }, + { + "epoch": 0.33, + "learning_rate": 0.0005420581745235707, + "loss": 1.4044, + "theoretical_loss": 3.6491532767259525, + "tokens_seen": 1086586880 + }, + { + "epoch": 0.33, + "learning_rate": 0.0005420421263791374, + "loss": 1.3854, + "theoretical_loss": 3.649132813061949, + "tokens_seen": 1086652416 + }, + { + "epoch": 0.33, + "learning_rate": 0.0005420260782347041, + "loss": 1.2734, + "theoretical_loss": 3.6491123509776138, + "tokens_seen": 1086717952 + }, + { + "epoch": 0.33, + "learning_rate": 0.0005420100300902708, + "loss": 1.5219, + "theoretical_loss": 3.64909189047273, + "tokens_seen": 1086783488 + }, + { + "epoch": 0.33, + "learning_rate": 0.0005419939819458375, + "loss": 1.3465, + "theoretical_loss": 3.64907143154708, + "tokens_seen": 1086849024 + }, + { + "epoch": 0.33, + "learning_rate": 0.0005419779338014042, + "loss": 1.3593, + "theoretical_loss": 3.6490509742004464, + "tokens_seen": 1086914560 + }, + { + "epoch": 0.33, + "learning_rate": 0.0005419618856569709, + "loss": 1.4472, + "theoretical_loss": 3.649030518432613, + "tokens_seen": 1086980096 + }, + { + "epoch": 0.33, + "learning_rate": 0.0005419458375125376, + "loss": 1.4102, + "theoretical_loss": 3.6490100642433623, + "tokens_seen": 1087045632 + }, + { + "epoch": 0.33, + "learning_rate": 0.0005419297893681043, + "loss": 1.4508, + "theoretical_loss": 3.648989611632477, + "tokens_seen": 1087111168 + }, + { + "epoch": 0.33, + "learning_rate": 0.000541913741223671, + "loss": 1.3936, + "theoretical_loss": 3.6489691605997403, + "tokens_seen": 1087176704 + }, + { + "epoch": 0.33, + "learning_rate": 0.0005418976930792377, + "loss": 1.3196, + "theoretical_loss": 3.6489487111449357, + "tokens_seen": 1087242240 + }, + { + "epoch": 0.33, + "learning_rate": 0.0005418816449348044, + "loss": 1.3713, + "theoretical_loss": 3.6489282632678464, + "tokens_seen": 1087307776 + }, + { + "epoch": 0.33, + "learning_rate": 0.0005418655967903711, + "loss": 1.4685, + "theoretical_loss": 3.648907816968255, + "tokens_seen": 1087373312 + }, + { + "epoch": 0.33, + "learning_rate": 0.0005418495486459378, + "loss": 1.5409, + "theoretical_loss": 3.6488873722459454, + "tokens_seen": 1087438848 + }, + { + "epoch": 0.33, + "learning_rate": 0.0005418335005015045, + "loss": 1.2892, + "theoretical_loss": 3.6488669291007003, + "tokens_seen": 1087504384 + }, + { + "epoch": 0.33, + "learning_rate": 0.0005418174523570712, + "loss": 1.3818, + "theoretical_loss": 3.6488464875323032, + "tokens_seen": 1087569920 + }, + { + "epoch": 0.33, + "learning_rate": 0.0005418014042126379, + "loss": 1.5105, + "theoretical_loss": 3.6488260475405374, + "tokens_seen": 1087635456 + }, + { + "epoch": 0.33, + "learning_rate": 0.0005417853560682046, + "loss": 1.464, + "theoretical_loss": 3.648805609125187, + "tokens_seen": 1087700992 + }, + { + "epoch": 0.33, + "learning_rate": 0.0005417693079237713, + "loss": 1.3653, + "theoretical_loss": 3.6487851722860345, + "tokens_seen": 1087766528 + }, + { + "epoch": 0.33, + "learning_rate": 0.000541753259779338, + "loss": 1.4156, + "theoretical_loss": 3.648764737022864, + "tokens_seen": 1087832064 + }, + { + "epoch": 0.33, + "objective/train/docs_used": 398862, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.3423185348510742, + "objective/train/theoretical_loss": 3.6487443033354583, + "objective/train/tokens_used": 1108357600, + "theoretical_loss": 3.6487443033354583, + "tokens_seen": 1087897600 + }, + { + "epoch": 0.33, + "learning_rate": 0.0005417372116349047, + "loss": 1.3644, + "theoretical_loss": 3.6487443033354583, + "tokens_seen": 1087897600 + }, + { + "epoch": 0.33, + "learning_rate": 0.0005417211634904715, + "loss": 1.3658, + "theoretical_loss": 3.648723871223602, + "tokens_seen": 1087963136 + }, + { + "epoch": 0.33, + "learning_rate": 0.0005417051153460381, + "loss": 1.3817, + "theoretical_loss": 3.6487034406870777, + "tokens_seen": 1088028672 + }, + { + "epoch": 0.33, + "learning_rate": 0.0005416890672016049, + "loss": 1.4007, + "theoretical_loss": 3.6486830117256703, + "tokens_seen": 1088094208 + }, + { + "epoch": 0.33, + "learning_rate": 0.0005416730190571715, + "loss": 1.3837, + "theoretical_loss": 3.648662584339162, + "tokens_seen": 1088159744 + }, + { + "epoch": 0.33, + "learning_rate": 0.0005416569709127383, + "loss": 1.4027, + "theoretical_loss": 3.6486421585273376, + "tokens_seen": 1088225280 + }, + { + "epoch": 0.33, + "learning_rate": 0.0005416409227683049, + "loss": 1.3223, + "theoretical_loss": 3.648621734289981, + "tokens_seen": 1088290816 + }, + { + "epoch": 0.33, + "learning_rate": 0.0005416248746238717, + "loss": 1.441, + "theoretical_loss": 3.648601311626875, + "tokens_seen": 1088356352 + }, + { + "epoch": 0.33, + "learning_rate": 0.0005416088264794383, + "loss": 1.5485, + "theoretical_loss": 3.6485808905378048, + "tokens_seen": 1088421888 + }, + { + "epoch": 0.33, + "learning_rate": 0.0005415927783350051, + "loss": 1.3888, + "theoretical_loss": 3.648560471022553, + "tokens_seen": 1088487424 + }, + { + "epoch": 0.33, + "learning_rate": 0.0005415767301905717, + "loss": 1.4314, + "theoretical_loss": 3.6485400530809047, + "tokens_seen": 1088552960 + }, + { + "epoch": 0.33, + "learning_rate": 0.0005415606820461385, + "loss": 1.4576, + "theoretical_loss": 3.6485196367126433, + "tokens_seen": 1088618496 + }, + { + "epoch": 0.33, + "learning_rate": 0.0005415446339017051, + "loss": 1.3887, + "theoretical_loss": 3.648499221917553, + "tokens_seen": 1088684032 + }, + { + "epoch": 0.33, + "learning_rate": 0.0005415285857572719, + "loss": 1.3359, + "theoretical_loss": 3.6484788086954176, + "tokens_seen": 1088749568 + }, + { + "epoch": 0.33, + "learning_rate": 0.0005415125376128385, + "loss": 1.4632, + "theoretical_loss": 3.6484583970460216, + "tokens_seen": 1088815104 + }, + { + "epoch": 0.33, + "learning_rate": 0.0005414964894684053, + "loss": 1.4188, + "theoretical_loss": 3.6484379869691494, + "tokens_seen": 1088880640 + }, + { + "epoch": 0.33, + "learning_rate": 0.0005414804413239719, + "loss": 1.4515, + "theoretical_loss": 3.6484175784645845, + "tokens_seen": 1088946176 + }, + { + "epoch": 0.33, + "learning_rate": 0.0005414643931795387, + "loss": 1.5621, + "theoretical_loss": 3.6483971715321117, + "tokens_seen": 1089011712 + }, + { + "epoch": 0.33, + "learning_rate": 0.0005414483450351053, + "loss": 1.5159, + "theoretical_loss": 3.6483767661715154, + "tokens_seen": 1089077248 + }, + { + "epoch": 0.33, + "learning_rate": 0.0005414322968906721, + "loss": 1.3619, + "theoretical_loss": 3.648356362382579, + "tokens_seen": 1089142784 + }, + { + "epoch": 0.33, + "learning_rate": 0.0005414162487462387, + "loss": 1.3624, + "theoretical_loss": 3.648335960165088, + "tokens_seen": 1089208320 + }, + { + "epoch": 0.33, + "learning_rate": 0.0005414002006018055, + "loss": 1.6399, + "theoretical_loss": 3.6483155595188266, + "tokens_seen": 1089273856 + }, + { + "epoch": 0.33, + "learning_rate": 0.0005413841524573721, + "loss": 1.4254, + "theoretical_loss": 3.648295160443579, + "tokens_seen": 1089339392 + }, + { + "epoch": 0.33, + "learning_rate": 0.0005413681043129389, + "loss": 1.4131, + "theoretical_loss": 3.6482747629391294, + "tokens_seen": 1089404928 + }, + { + "epoch": 0.33, + "learning_rate": 0.0005413520561685055, + "loss": 1.4614, + "theoretical_loss": 3.648254367005263, + "tokens_seen": 1089470464 + }, + { + "epoch": 0.33, + "learning_rate": 0.0005413360080240723, + "loss": 1.4888, + "theoretical_loss": 3.6482339726417643, + "tokens_seen": 1089536000 + }, + { + "epoch": 0.33, + "learning_rate": 0.000541319959879639, + "loss": 1.407, + "theoretical_loss": 3.6482135798484174, + "tokens_seen": 1089601536 + }, + { + "epoch": 0.33, + "learning_rate": 0.0005413039117352057, + "loss": 1.5044, + "theoretical_loss": 3.648193188625008, + "tokens_seen": 1089667072 + }, + { + "epoch": 0.33, + "learning_rate": 0.0005412878635907723, + "loss": 1.3556, + "theoretical_loss": 3.64817279897132, + "tokens_seen": 1089732608 + }, + { + "epoch": 0.33, + "learning_rate": 0.0005412718154463391, + "loss": 1.4281, + "theoretical_loss": 3.648152410887138, + "tokens_seen": 1089798144 + }, + { + "epoch": 0.33, + "learning_rate": 0.0005412557673019058, + "loss": 1.5316, + "theoretical_loss": 3.6481320243722477, + "tokens_seen": 1089863680 + }, + { + "epoch": 0.33, + "learning_rate": 0.0005412397191574725, + "loss": 1.4831, + "theoretical_loss": 3.6481116394264337, + "tokens_seen": 1089929216 + }, + { + "epoch": 0.33, + "learning_rate": 0.0005412236710130392, + "loss": 1.4971, + "theoretical_loss": 3.64809125604948, + "tokens_seen": 1089994752 + }, + { + "epoch": 0.33, + "learning_rate": 0.0005412076228686059, + "loss": 1.3796, + "theoretical_loss": 3.6480708742411725, + "tokens_seen": 1090060288 + }, + { + "epoch": 0.33, + "learning_rate": 0.0005411915747241726, + "loss": 1.4259, + "theoretical_loss": 3.6480504940012954, + "tokens_seen": 1090125824 + }, + { + "epoch": 0.33, + "learning_rate": 0.0005411755265797393, + "loss": 1.5508, + "theoretical_loss": 3.648030115329635, + "tokens_seen": 1090191360 + }, + { + "epoch": 0.33, + "learning_rate": 0.000541159478435306, + "loss": 1.4859, + "theoretical_loss": 3.648009738225975, + "tokens_seen": 1090256896 + }, + { + "epoch": 0.33, + "learning_rate": 0.0005411434302908727, + "loss": 1.3996, + "theoretical_loss": 3.647989362690101, + "tokens_seen": 1090322432 + }, + { + "epoch": 0.33, + "learning_rate": 0.0005411273821464394, + "loss": 1.3364, + "theoretical_loss": 3.6479689887217988, + "tokens_seen": 1090387968 + }, + { + "epoch": 0.33, + "learning_rate": 0.0005411113340020061, + "loss": 1.3945, + "theoretical_loss": 3.6479486163208525, + "tokens_seen": 1090453504 + }, + { + "epoch": 0.33, + "learning_rate": 0.0005410952858575728, + "loss": 1.4924, + "theoretical_loss": 3.6479282454870487, + "tokens_seen": 1090519040 + }, + { + "epoch": 0.33, + "learning_rate": 0.0005410792377131395, + "loss": 1.4311, + "theoretical_loss": 3.647907876220171, + "tokens_seen": 1090584576 + }, + { + "epoch": 0.33, + "learning_rate": 0.0005410631895687062, + "loss": 1.3959, + "theoretical_loss": 3.6478875085200055, + "tokens_seen": 1090650112 + }, + { + "epoch": 0.33, + "learning_rate": 0.0005410471414242729, + "loss": 1.4914, + "theoretical_loss": 3.647867142386338, + "tokens_seen": 1090715648 + }, + { + "epoch": 0.33, + "learning_rate": 0.0005410310932798396, + "loss": 1.3325, + "theoretical_loss": 3.6478467778189536, + "tokens_seen": 1090781184 + }, + { + "epoch": 0.33, + "learning_rate": 0.0005410150451354063, + "loss": 1.4775, + "theoretical_loss": 3.647826414817638, + "tokens_seen": 1090846720 + }, + { + "epoch": 0.33, + "learning_rate": 0.000540998996990973, + "loss": 1.4725, + "theoretical_loss": 3.6478060533821757, + "tokens_seen": 1090912256 + }, + { + "epoch": 0.33, + "learning_rate": 0.0005409829488465397, + "loss": 1.5584, + "theoretical_loss": 3.647785693512353, + "tokens_seen": 1090977792 + }, + { + "epoch": 0.33, + "learning_rate": 0.0005409669007021064, + "loss": 1.3514, + "theoretical_loss": 3.6477653352079558, + "tokens_seen": 1091043328 + }, + { + "epoch": 0.33, + "learning_rate": 0.000540950852557673, + "loss": 1.443, + "theoretical_loss": 3.647744978468769, + "tokens_seen": 1091108864 + }, + { + "epoch": 0.33, + "objective/train/docs_used": 400125, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.522972822189331, + "objective/train/theoretical_loss": 3.647724623294579, + "objective/train/tokens_used": 1111634400, + "theoretical_loss": 3.647724623294579, + "tokens_seen": 1091174400 + }, + { + "epoch": 0.33, + "learning_rate": 0.0005409348044132398, + "loss": 1.4202, + "theoretical_loss": 3.647724623294579, + "tokens_seen": 1091174400 + }, + { + "epoch": 0.33, + "learning_rate": 0.0005409187562688064, + "loss": 1.5079, + "theoretical_loss": 3.6477042696851707, + "tokens_seen": 1091239936 + }, + { + "epoch": 0.33, + "learning_rate": 0.0005409027081243732, + "loss": 1.4338, + "theoretical_loss": 3.64768391764033, + "tokens_seen": 1091305472 + }, + { + "epoch": 0.33, + "learning_rate": 0.0005408866599799398, + "loss": 1.3323, + "theoretical_loss": 3.647663567159843, + "tokens_seen": 1091371008 + }, + { + "epoch": 0.33, + "learning_rate": 0.0005408706118355066, + "loss": 1.4291, + "theoretical_loss": 3.6476432182434957, + "tokens_seen": 1091436544 + }, + { + "epoch": 0.33, + "learning_rate": 0.0005408545636910732, + "loss": 1.416, + "theoretical_loss": 3.647622870891074, + "tokens_seen": 1091502080 + }, + { + "epoch": 0.33, + "learning_rate": 0.0005408385155466399, + "loss": 1.5676, + "theoretical_loss": 3.647602525102363, + "tokens_seen": 1091567616 + }, + { + "epoch": 0.33, + "learning_rate": 0.0005408224674022066, + "loss": 1.3228, + "theoretical_loss": 3.6475821808771496, + "tokens_seen": 1091633152 + }, + { + "epoch": 0.33, + "learning_rate": 0.0005408064192577733, + "loss": 1.5457, + "theoretical_loss": 3.6475618382152186, + "tokens_seen": 1091698688 + }, + { + "epoch": 0.33, + "learning_rate": 0.00054079037111334, + "loss": 1.5787, + "theoretical_loss": 3.6475414971163573, + "tokens_seen": 1091764224 + }, + { + "epoch": 0.33, + "learning_rate": 0.0005407743229689067, + "loss": 1.4413, + "theoretical_loss": 3.6475211575803517, + "tokens_seen": 1091829760 + }, + { + "epoch": 0.33, + "learning_rate": 0.0005407582748244734, + "loss": 1.3535, + "theoretical_loss": 3.6475008196069867, + "tokens_seen": 1091895296 + }, + { + "epoch": 0.33, + "learning_rate": 0.0005407422266800401, + "loss": 1.4462, + "theoretical_loss": 3.6474804831960497, + "tokens_seen": 1091960832 + }, + { + "epoch": 0.33, + "learning_rate": 0.0005407261785356068, + "loss": 1.458, + "theoretical_loss": 3.6474601483473266, + "tokens_seen": 1092026368 + }, + { + "epoch": 0.33, + "learning_rate": 0.0005407101303911735, + "loss": 1.4705, + "theoretical_loss": 3.6474398150606033, + "tokens_seen": 1092091904 + }, + { + "epoch": 0.33, + "learning_rate": 0.0005406940822467402, + "loss": 1.4923, + "theoretical_loss": 3.647419483335667, + "tokens_seen": 1092157440 + }, + { + "epoch": 0.33, + "learning_rate": 0.0005406780341023069, + "loss": 1.4676, + "theoretical_loss": 3.6473991531723025, + "tokens_seen": 1092222976 + }, + { + "epoch": 0.33, + "learning_rate": 0.0005406619859578736, + "loss": 1.4494, + "theoretical_loss": 3.647378824570297, + "tokens_seen": 1092288512 + }, + { + "epoch": 0.33, + "learning_rate": 0.0005406459378134403, + "loss": 1.4178, + "theoretical_loss": 3.6473584975294377, + "tokens_seen": 1092354048 + }, + { + "epoch": 0.33, + "learning_rate": 0.000540629889669007, + "loss": 1.4409, + "theoretical_loss": 3.64733817204951, + "tokens_seen": 1092419584 + }, + { + "epoch": 0.33, + "learning_rate": 0.0005406138415245737, + "loss": 1.3551, + "theoretical_loss": 3.6473178481303004, + "tokens_seen": 1092485120 + }, + { + "epoch": 0.33, + "learning_rate": 0.0005405977933801404, + "loss": 1.363, + "theoretical_loss": 3.647297525771596, + "tokens_seen": 1092550656 + }, + { + "epoch": 0.33, + "learning_rate": 0.0005405817452357071, + "loss": 1.4933, + "theoretical_loss": 3.647277204973183, + "tokens_seen": 1092616192 + }, + { + "epoch": 0.33, + "learning_rate": 0.0005405656970912738, + "loss": 1.4107, + "theoretical_loss": 3.6472568857348486, + "tokens_seen": 1092681728 + }, + { + "epoch": 0.33, + "learning_rate": 0.0005405496489468405, + "loss": 1.4446, + "theoretical_loss": 3.647236568056379, + "tokens_seen": 1092747264 + }, + { + "epoch": 0.33, + "learning_rate": 0.0005405336008024072, + "loss": 1.4794, + "theoretical_loss": 3.647216251937561, + "tokens_seen": 1092812800 + }, + { + "epoch": 0.33, + "learning_rate": 0.000540517552657974, + "loss": 1.4307, + "theoretical_loss": 3.647195937378181, + "tokens_seen": 1092878336 + }, + { + "epoch": 0.33, + "learning_rate": 0.0005405015045135406, + "loss": 1.4931, + "theoretical_loss": 3.6471756243780264, + "tokens_seen": 1092943872 + }, + { + "epoch": 0.33, + "learning_rate": 0.0005404854563691074, + "loss": 1.4878, + "theoretical_loss": 3.647155312936883, + "tokens_seen": 1093009408 + }, + { + "epoch": 0.33, + "learning_rate": 0.000540469408224674, + "loss": 1.4206, + "theoretical_loss": 3.6471350030545393, + "tokens_seen": 1093074944 + }, + { + "epoch": 0.33, + "learning_rate": 0.0005404533600802408, + "loss": 1.4967, + "theoretical_loss": 3.6471146947307815, + "tokens_seen": 1093140480 + }, + { + "epoch": 0.33, + "learning_rate": 0.0005404373119358074, + "loss": 1.3282, + "theoretical_loss": 3.647094387965396, + "tokens_seen": 1093206016 + }, + { + "epoch": 0.33, + "learning_rate": 0.0005404212637913742, + "loss": 1.4415, + "theoretical_loss": 3.64707408275817, + "tokens_seen": 1093271552 + }, + { + "epoch": 0.33, + "learning_rate": 0.0005404052156469408, + "loss": 1.4446, + "theoretical_loss": 3.6470537791088913, + "tokens_seen": 1093337088 + }, + { + "epoch": 0.33, + "learning_rate": 0.0005403891675025076, + "loss": 1.3553, + "theoretical_loss": 3.647033477017346, + "tokens_seen": 1093402624 + }, + { + "epoch": 0.33, + "learning_rate": 0.0005403731193580742, + "loss": 1.4416, + "theoretical_loss": 3.6470131764833225, + "tokens_seen": 1093468160 + }, + { + "epoch": 0.33, + "learning_rate": 0.000540357071213641, + "loss": 1.3341, + "theoretical_loss": 3.646992877506606, + "tokens_seen": 1093533696 + }, + { + "epoch": 0.33, + "learning_rate": 0.0005403410230692076, + "loss": 1.4408, + "theoretical_loss": 3.646972580086986, + "tokens_seen": 1093599232 + }, + { + "epoch": 0.33, + "learning_rate": 0.0005403249749247744, + "loss": 1.4617, + "theoretical_loss": 3.646952284224248, + "tokens_seen": 1093664768 + }, + { + "epoch": 0.33, + "learning_rate": 0.000540308926780341, + "loss": 1.3457, + "theoretical_loss": 3.64693198991818, + "tokens_seen": 1093730304 + }, + { + "epoch": 0.33, + "learning_rate": 0.0005402928786359078, + "loss": 1.4191, + "theoretical_loss": 3.6469116971685693, + "tokens_seen": 1093795840 + }, + { + "epoch": 0.33, + "learning_rate": 0.0005402768304914744, + "loss": 1.5409, + "theoretical_loss": 3.6468914059752033, + "tokens_seen": 1093861376 + }, + { + "epoch": 0.33, + "learning_rate": 0.0005402607823470412, + "loss": 1.449, + "theoretical_loss": 3.6468711163378695, + "tokens_seen": 1093926912 + }, + { + "epoch": 0.33, + "learning_rate": 0.0005402447342026078, + "loss": 1.4593, + "theoretical_loss": 3.646850828256355, + "tokens_seen": 1093992448 + }, + { + "epoch": 0.33, + "learning_rate": 0.0005402286860581746, + "loss": 1.5518, + "theoretical_loss": 3.646830541730448, + "tokens_seen": 1094057984 + }, + { + "epoch": 0.33, + "learning_rate": 0.0005402126379137412, + "loss": 1.4347, + "theoretical_loss": 3.6468102567599354, + "tokens_seen": 1094123520 + }, + { + "epoch": 0.33, + "learning_rate": 0.000540196589769308, + "loss": 1.4404, + "theoretical_loss": 3.646789973344605, + "tokens_seen": 1094189056 + }, + { + "epoch": 0.33, + "learning_rate": 0.0005401805416248746, + "loss": 1.3502, + "theoretical_loss": 3.6467696914842445, + "tokens_seen": 1094254592 + }, + { + "epoch": 0.33, + "learning_rate": 0.0005401644934804414, + "loss": 1.4939, + "theoretical_loss": 3.646749411178641, + "tokens_seen": 1094320128 + }, + { + "epoch": 0.33, + "learning_rate": 0.000540148445336008, + "loss": 1.4064, + "theoretical_loss": 3.6467291324275832, + "tokens_seen": 1094385664 + }, + { + "epoch": 0.33, + "objective/train/docs_used": 401268, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.3594894409179688, + "objective/train/theoretical_loss": 3.646708855230859, + "objective/train/tokens_used": 1114911200, + "theoretical_loss": 3.646708855230859, + "tokens_seen": 1094451200 + }, + { + "epoch": 0.33, + "learning_rate": 0.0005401323971915748, + "loss": 1.4591, + "theoretical_loss": 3.646708855230859, + "tokens_seen": 1094451200 + }, + { + "epoch": 0.33, + "learning_rate": 0.0005401163490471414, + "loss": 1.4831, + "theoretical_loss": 3.646688579588255, + "tokens_seen": 1094516736 + }, + { + "epoch": 0.33, + "learning_rate": 0.0005401003009027082, + "loss": 1.4322, + "theoretical_loss": 3.646668305499559, + "tokens_seen": 1094582272 + }, + { + "epoch": 0.33, + "learning_rate": 0.0005400842527582748, + "loss": 1.3084, + "theoretical_loss": 3.6466480329645603, + "tokens_seen": 1094647808 + }, + { + "epoch": 0.33, + "learning_rate": 0.0005400682046138416, + "loss": 1.4258, + "theoretical_loss": 3.646627761983046, + "tokens_seen": 1094713344 + }, + { + "epoch": 0.33, + "learning_rate": 0.0005400521564694082, + "loss": 1.4965, + "theoretical_loss": 3.6466074925548035, + "tokens_seen": 1094778880 + }, + { + "epoch": 0.33, + "learning_rate": 0.000540036108324975, + "loss": 1.4785, + "theoretical_loss": 3.646587224679622, + "tokens_seen": 1094844416 + }, + { + "epoch": 0.33, + "learning_rate": 0.0005400200601805416, + "loss": 1.4049, + "theoretical_loss": 3.646566958357289, + "tokens_seen": 1094909952 + }, + { + "epoch": 0.33, + "learning_rate": 0.0005400040120361084, + "loss": 1.4969, + "theoretical_loss": 3.6465466935875925, + "tokens_seen": 1094975488 + }, + { + "epoch": 0.33, + "learning_rate": 0.000539987963891675, + "loss": 1.3671, + "theoretical_loss": 3.6465264303703204, + "tokens_seen": 1095041024 + }, + { + "epoch": 0.33, + "learning_rate": 0.0005399719157472418, + "loss": 1.4434, + "theoretical_loss": 3.6465061687052613, + "tokens_seen": 1095106560 + }, + { + "epoch": 0.33, + "learning_rate": 0.0005399558676028084, + "loss": 1.3676, + "theoretical_loss": 3.6464859085922035, + "tokens_seen": 1095172096 + }, + { + "epoch": 0.33, + "learning_rate": 0.0005399398194583752, + "loss": 1.3477, + "theoretical_loss": 3.646465650030935, + "tokens_seen": 1095237632 + }, + { + "epoch": 0.33, + "learning_rate": 0.0005399237713139418, + "loss": 1.3786, + "theoretical_loss": 3.6464453930212435, + "tokens_seen": 1095303168 + }, + { + "epoch": 0.33, + "learning_rate": 0.0005399077231695086, + "loss": 1.4429, + "theoretical_loss": 3.6464251375629186, + "tokens_seen": 1095368704 + }, + { + "epoch": 0.33, + "learning_rate": 0.0005398916750250752, + "loss": 1.4925, + "theoretical_loss": 3.646404883655748, + "tokens_seen": 1095434240 + }, + { + "epoch": 0.33, + "learning_rate": 0.000539875626880642, + "loss": 1.4045, + "theoretical_loss": 3.6463846312995196, + "tokens_seen": 1095499776 + }, + { + "epoch": 0.33, + "learning_rate": 0.0005398595787362087, + "loss": 1.3977, + "theoretical_loss": 3.646364380494023, + "tokens_seen": 1095565312 + }, + { + "epoch": 0.33, + "learning_rate": 0.0005398435305917754, + "loss": 1.3995, + "theoretical_loss": 3.6463441312390463, + "tokens_seen": 1095630848 + }, + { + "epoch": 0.33, + "learning_rate": 0.000539827482447342, + "loss": 1.4526, + "theoretical_loss": 3.646323883534378, + "tokens_seen": 1095696384 + }, + { + "epoch": 0.33, + "learning_rate": 0.0005398114343029088, + "loss": 1.5505, + "theoretical_loss": 3.6463036373798063, + "tokens_seen": 1095761920 + }, + { + "epoch": 0.33, + "learning_rate": 0.0005397953861584755, + "loss": 1.308, + "theoretical_loss": 3.64628339277512, + "tokens_seen": 1095827456 + }, + { + "epoch": 0.33, + "learning_rate": 0.0005397793380140422, + "loss": 1.311, + "theoretical_loss": 3.6462631497201077, + "tokens_seen": 1095892992 + }, + { + "epoch": 0.33, + "learning_rate": 0.0005397632898696089, + "loss": 1.4523, + "theoretical_loss": 3.6462429082145587, + "tokens_seen": 1095958528 + }, + { + "epoch": 0.33, + "learning_rate": 0.0005397472417251756, + "loss": 1.4592, + "theoretical_loss": 3.646222668258261, + "tokens_seen": 1096024064 + }, + { + "epoch": 0.33, + "learning_rate": 0.0005397311935807423, + "loss": 1.4152, + "theoretical_loss": 3.6462024298510043, + "tokens_seen": 1096089600 + }, + { + "epoch": 0.33, + "learning_rate": 0.000539715145436309, + "loss": 1.3738, + "theoretical_loss": 3.6461821929925766, + "tokens_seen": 1096155136 + }, + { + "epoch": 0.33, + "learning_rate": 0.0005396990972918757, + "loss": 1.3911, + "theoretical_loss": 3.646161957682767, + "tokens_seen": 1096220672 + }, + { + "epoch": 0.33, + "learning_rate": 0.0005396830491474424, + "loss": 1.4175, + "theoretical_loss": 3.6461417239213647, + "tokens_seen": 1096286208 + }, + { + "epoch": 0.33, + "learning_rate": 0.0005396670010030091, + "loss": 1.4902, + "theoretical_loss": 3.6461214917081586, + "tokens_seen": 1096351744 + }, + { + "epoch": 0.33, + "learning_rate": 0.0005396509528585758, + "loss": 1.3775, + "theoretical_loss": 3.6461012610429373, + "tokens_seen": 1096417280 + }, + { + "epoch": 0.33, + "learning_rate": 0.0005396349047141425, + "loss": 1.4646, + "theoretical_loss": 3.64608103192549, + "tokens_seen": 1096482816 + }, + { + "epoch": 0.33, + "learning_rate": 0.0005396188565697092, + "loss": 1.4007, + "theoretical_loss": 3.6460608043556064, + "tokens_seen": 1096548352 + }, + { + "epoch": 0.33, + "learning_rate": 0.0005396028084252759, + "loss": 1.4807, + "theoretical_loss": 3.6460405783330745, + "tokens_seen": 1096613888 + }, + { + "epoch": 0.33, + "learning_rate": 0.0005395867602808426, + "loss": 1.3088, + "theoretical_loss": 3.6460203538576845, + "tokens_seen": 1096679424 + }, + { + "epoch": 0.33, + "learning_rate": 0.0005395707121364093, + "loss": 1.3826, + "theoretical_loss": 3.6460001309292247, + "tokens_seen": 1096744960 + }, + { + "epoch": 0.33, + "learning_rate": 0.0005395546639919759, + "loss": 1.4683, + "theoretical_loss": 3.6459799095474854, + "tokens_seen": 1096810496 + }, + { + "epoch": 0.33, + "learning_rate": 0.0005395386158475427, + "loss": 1.4506, + "theoretical_loss": 3.6459596897122553, + "tokens_seen": 1096876032 + }, + { + "epoch": 0.33, + "learning_rate": 0.0005395225677031093, + "loss": 1.5101, + "theoretical_loss": 3.6459394714233238, + "tokens_seen": 1096941568 + }, + { + "epoch": 0.33, + "learning_rate": 0.0005395065195586761, + "loss": 1.4909, + "theoretical_loss": 3.64591925468048, + "tokens_seen": 1097007104 + }, + { + "epoch": 0.33, + "learning_rate": 0.0005394904714142427, + "loss": 1.3651, + "theoretical_loss": 3.645899039483514, + "tokens_seen": 1097072640 + }, + { + "epoch": 0.33, + "learning_rate": 0.0005394744232698095, + "loss": 1.4273, + "theoretical_loss": 3.6458788258322143, + "tokens_seen": 1097138176 + }, + { + "epoch": 0.33, + "learning_rate": 0.0005394583751253761, + "loss": 1.4709, + "theoretical_loss": 3.6458586137263715, + "tokens_seen": 1097203712 + }, + { + "epoch": 0.33, + "learning_rate": 0.0005394423269809429, + "loss": 1.3196, + "theoretical_loss": 3.645838403165774, + "tokens_seen": 1097269248 + }, + { + "epoch": 0.33, + "learning_rate": 0.0005394262788365095, + "loss": 1.3947, + "theoretical_loss": 3.6458181941502126, + "tokens_seen": 1097334784 + }, + { + "epoch": 0.33, + "learning_rate": 0.0005394102306920762, + "loss": 1.4025, + "theoretical_loss": 3.645797986679476, + "tokens_seen": 1097400320 + }, + { + "epoch": 0.33, + "learning_rate": 0.0005393941825476429, + "loss": 1.3877, + "theoretical_loss": 3.6457777807533542, + "tokens_seen": 1097465856 + }, + { + "epoch": 0.33, + "learning_rate": 0.0005393781344032096, + "loss": 1.3935, + "theoretical_loss": 3.6457575763716363, + "tokens_seen": 1097531392 + }, + { + "epoch": 0.33, + "learning_rate": 0.0005393620862587763, + "loss": 1.3909, + "theoretical_loss": 3.645737373534113, + "tokens_seen": 1097596928 + }, + { + "epoch": 0.33, + "learning_rate": 0.000539346038114343, + "loss": 1.5326, + "theoretical_loss": 3.645717172240574, + "tokens_seen": 1097662464 + }, + { + "epoch": 0.33, + "objective/train/docs_used": 402429, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.3182363510131836, + "objective/train/theoretical_loss": 3.645696972490809, + "objective/train/tokens_used": 1118188000, + "theoretical_loss": 3.645696972490809, + "tokens_seen": 1097728000 + }, + { + "epoch": 0.33, + "learning_rate": 0.0005393299899699097, + "loss": 1.4487, + "theoretical_loss": 3.645696972490809, + "tokens_seen": 1097728000 + }, + { + "epoch": 0.33, + "learning_rate": 0.0005393139418254765, + "loss": 1.4311, + "theoretical_loss": 3.645676774284607, + "tokens_seen": 1097793536 + }, + { + "epoch": 0.33, + "learning_rate": 0.0005392978936810431, + "loss": 1.4191, + "theoretical_loss": 3.645656577621758, + "tokens_seen": 1097859072 + }, + { + "epoch": 0.33, + "learning_rate": 0.0005392818455366099, + "loss": 1.4671, + "theoretical_loss": 3.6456363825020537, + "tokens_seen": 1097924608 + }, + { + "epoch": 0.33, + "learning_rate": 0.0005392657973921765, + "loss": 1.5127, + "theoretical_loss": 3.645616188925283, + "tokens_seen": 1097990144 + }, + { + "epoch": 0.33, + "learning_rate": 0.0005392497492477433, + "loss": 1.3547, + "theoretical_loss": 3.6455959968912355, + "tokens_seen": 1098055680 + }, + { + "epoch": 0.33, + "learning_rate": 0.0005392337011033099, + "loss": 1.3639, + "theoretical_loss": 3.6455758063997017, + "tokens_seen": 1098121216 + }, + { + "epoch": 0.33, + "learning_rate": 0.0005392176529588767, + "loss": 1.3986, + "theoretical_loss": 3.645555617450472, + "tokens_seen": 1098186752 + }, + { + "epoch": 0.33, + "learning_rate": 0.0005392016048144433, + "loss": 1.4002, + "theoretical_loss": 3.645535430043336, + "tokens_seen": 1098252288 + }, + { + "epoch": 0.33, + "learning_rate": 0.0005391855566700101, + "loss": 1.3531, + "theoretical_loss": 3.6455152441780836, + "tokens_seen": 1098317824 + }, + { + "epoch": 0.33, + "learning_rate": 0.0005391695085255767, + "loss": 1.3753, + "theoretical_loss": 3.645495059854506, + "tokens_seen": 1098383360 + }, + { + "epoch": 0.33, + "learning_rate": 0.0005391534603811435, + "loss": 1.3841, + "theoretical_loss": 3.6454748770723935, + "tokens_seen": 1098448896 + }, + { + "epoch": 0.33, + "learning_rate": 0.0005391374122367101, + "loss": 1.5202, + "theoretical_loss": 3.6454546958315355, + "tokens_seen": 1098514432 + }, + { + "epoch": 0.33, + "learning_rate": 0.0005391213640922769, + "loss": 1.4766, + "theoretical_loss": 3.645434516131723, + "tokens_seen": 1098579968 + }, + { + "epoch": 0.33, + "learning_rate": 0.0005391053159478435, + "loss": 1.4356, + "theoretical_loss": 3.6454143379727464, + "tokens_seen": 1098645504 + }, + { + "epoch": 0.33, + "learning_rate": 0.0005390892678034103, + "loss": 1.4849, + "theoretical_loss": 3.645394161354395, + "tokens_seen": 1098711040 + }, + { + "epoch": 0.33, + "learning_rate": 0.0005390732196589769, + "loss": 1.4344, + "theoretical_loss": 3.6453739862764616, + "tokens_seen": 1098776576 + }, + { + "epoch": 0.33, + "learning_rate": 0.0005390571715145437, + "loss": 1.3928, + "theoretical_loss": 3.6453538127387346, + "tokens_seen": 1098842112 + }, + { + "epoch": 0.33, + "learning_rate": 0.0005390411233701103, + "loss": 1.3994, + "theoretical_loss": 3.645333640741006, + "tokens_seen": 1098907648 + }, + { + "epoch": 0.33, + "learning_rate": 0.0005390250752256771, + "loss": 1.4118, + "theoretical_loss": 3.645313470283065, + "tokens_seen": 1098973184 + }, + { + "epoch": 0.33, + "learning_rate": 0.0005390090270812437, + "loss": 1.4416, + "theoretical_loss": 3.6452933013647035, + "tokens_seen": 1099038720 + }, + { + "epoch": 0.33, + "learning_rate": 0.0005389929789368105, + "loss": 1.4106, + "theoretical_loss": 3.645273133985712, + "tokens_seen": 1099104256 + }, + { + "epoch": 0.33, + "learning_rate": 0.0005389769307923771, + "loss": 1.4051, + "theoretical_loss": 3.6452529681458805, + "tokens_seen": 1099169792 + }, + { + "epoch": 0.33, + "learning_rate": 0.0005389608826479439, + "loss": 1.2654, + "theoretical_loss": 3.6452328038450004, + "tokens_seen": 1099235328 + }, + { + "epoch": 0.33, + "learning_rate": 0.0005389448345035105, + "loss": 1.3581, + "theoretical_loss": 3.6452126410828622, + "tokens_seen": 1099300864 + }, + { + "epoch": 0.33, + "learning_rate": 0.0005389287863590773, + "loss": 1.5046, + "theoretical_loss": 3.645192479859257, + "tokens_seen": 1099366400 + }, + { + "epoch": 0.33, + "learning_rate": 0.0005389127382146439, + "loss": 1.3089, + "theoretical_loss": 3.645172320173976, + "tokens_seen": 1099431936 + }, + { + "epoch": 0.33, + "learning_rate": 0.0005388966900702107, + "loss": 1.5117, + "theoretical_loss": 3.6451521620268093, + "tokens_seen": 1099497472 + }, + { + "epoch": 0.33, + "learning_rate": 0.0005388806419257773, + "loss": 1.5093, + "theoretical_loss": 3.645132005417548, + "tokens_seen": 1099563008 + }, + { + "epoch": 0.33, + "learning_rate": 0.0005388645937813441, + "loss": 1.4049, + "theoretical_loss": 3.645111850345984, + "tokens_seen": 1099628544 + }, + { + "epoch": 0.33, + "learning_rate": 0.0005388485456369107, + "loss": 1.3443, + "theoretical_loss": 3.6450916968119076, + "tokens_seen": 1099694080 + }, + { + "epoch": 0.33, + "learning_rate": 0.0005388324974924775, + "loss": 1.4712, + "theoretical_loss": 3.64507154481511, + "tokens_seen": 1099759616 + }, + { + "epoch": 0.33, + "learning_rate": 0.0005388164493480441, + "loss": 1.3842, + "theoretical_loss": 3.645051394355382, + "tokens_seen": 1099825152 + }, + { + "epoch": 0.33, + "learning_rate": 0.0005388004012036109, + "loss": 1.5236, + "theoretical_loss": 3.6450312454325164, + "tokens_seen": 1099890688 + }, + { + "epoch": 0.33, + "learning_rate": 0.0005387843530591775, + "loss": 1.4283, + "theoretical_loss": 3.645011098046302, + "tokens_seen": 1099956224 + }, + { + "epoch": 0.33, + "learning_rate": 0.0005387683049147443, + "loss": 1.3753, + "theoretical_loss": 3.6449909521965322, + "tokens_seen": 1100021760 + }, + { + "epoch": 0.33, + "learning_rate": 0.0005387522567703109, + "loss": 1.4746, + "theoretical_loss": 3.6449708078829968, + "tokens_seen": 1100087296 + }, + { + "epoch": 0.33, + "learning_rate": 0.0005387362086258777, + "loss": 1.3798, + "theoretical_loss": 3.644950665105488, + "tokens_seen": 1100152832 + }, + { + "epoch": 0.33, + "learning_rate": 0.0005387201604814443, + "loss": 1.4714, + "theoretical_loss": 3.6449305238637972, + "tokens_seen": 1100218368 + }, + { + "epoch": 0.33, + "learning_rate": 0.0005387041123370111, + "loss": 1.4374, + "theoretical_loss": 3.644910384157715, + "tokens_seen": 1100283904 + }, + { + "epoch": 0.33, + "learning_rate": 0.0005386880641925777, + "loss": 1.3789, + "theoretical_loss": 3.6448902459870336, + "tokens_seen": 1100349440 + }, + { + "epoch": 0.33, + "learning_rate": 0.0005386720160481445, + "loss": 1.3233, + "theoretical_loss": 3.6448701093515448, + "tokens_seen": 1100414976 + }, + { + "epoch": 0.33, + "learning_rate": 0.0005386559679037111, + "loss": 1.4653, + "theoretical_loss": 3.644849974251039, + "tokens_seen": 1100480512 + }, + { + "epoch": 0.33, + "learning_rate": 0.0005386399197592779, + "loss": 1.4351, + "theoretical_loss": 3.6448298406853095, + "tokens_seen": 1100546048 + }, + { + "epoch": 0.33, + "learning_rate": 0.0005386238716148445, + "loss": 1.5546, + "theoretical_loss": 3.6448097086541456, + "tokens_seen": 1100611584 + }, + { + "epoch": 0.33, + "learning_rate": 0.0005386078234704113, + "loss": 1.4805, + "theoretical_loss": 3.644789578157341, + "tokens_seen": 1100677120 + }, + { + "epoch": 0.33, + "learning_rate": 0.000538591775325978, + "loss": 1.3516, + "theoretical_loss": 3.6447694491946865, + "tokens_seen": 1100742656 + }, + { + "epoch": 0.33, + "learning_rate": 0.0005385757271815447, + "loss": 1.4952, + "theoretical_loss": 3.6447493217659743, + "tokens_seen": 1100808192 + }, + { + "epoch": 0.33, + "learning_rate": 0.0005385596790371113, + "loss": 1.5107, + "theoretical_loss": 3.6447291958709958, + "tokens_seen": 1100873728 + }, + { + "epoch": 0.33, + "learning_rate": 0.0005385436308926781, + "loss": 1.3558, + "theoretical_loss": 3.644709071509543, + "tokens_seen": 1100939264 + }, + { + "epoch": 0.33, + "objective/train/docs_used": 403530, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.6000986099243164, + "objective/train/theoretical_loss": 3.6446889486814076, + "objective/train/tokens_used": 1121464800, + "theoretical_loss": 3.6446889486814076, + "tokens_seen": 1101004800 + }, + { + "epoch": 0.33, + "learning_rate": 0.0005385275827482447, + "loss": 1.5438, + "theoretical_loss": 3.6446889486814076, + "tokens_seen": 1101004800 + }, + { + "epoch": 0.33, + "learning_rate": 0.0005385115346038115, + "loss": 1.4098, + "theoretical_loss": 3.6446688273863814, + "tokens_seen": 1101070336 + }, + { + "epoch": 0.33, + "learning_rate": 0.0005384954864593782, + "loss": 1.3935, + "theoretical_loss": 3.644648707624257, + "tokens_seen": 1101135872 + }, + { + "epoch": 0.33, + "learning_rate": 0.0005384794383149449, + "loss": 1.4494, + "theoretical_loss": 3.6446285893948263, + "tokens_seen": 1101201408 + }, + { + "epoch": 0.33, + "learning_rate": 0.0005384633901705116, + "loss": 1.3338, + "theoretical_loss": 3.6446084726978807, + "tokens_seen": 1101266944 + }, + { + "epoch": 0.33, + "learning_rate": 0.0005384473420260783, + "loss": 1.3564, + "theoretical_loss": 3.6445883575332125, + "tokens_seen": 1101332480 + }, + { + "epoch": 0.33, + "learning_rate": 0.000538431293881645, + "loss": 1.2872, + "theoretical_loss": 3.644568243900614, + "tokens_seen": 1101398016 + }, + { + "epoch": 0.33, + "learning_rate": 0.0005384152457372117, + "loss": 1.4224, + "theoretical_loss": 3.644548131799877, + "tokens_seen": 1101463552 + }, + { + "epoch": 0.33, + "learning_rate": 0.0005383991975927784, + "loss": 1.4616, + "theoretical_loss": 3.6445280212307942, + "tokens_seen": 1101529088 + }, + { + "epoch": 0.33, + "learning_rate": 0.0005383831494483451, + "loss": 1.4348, + "theoretical_loss": 3.644507912193158, + "tokens_seen": 1101594624 + }, + { + "epoch": 0.33, + "learning_rate": 0.0005383671013039118, + "loss": 1.4807, + "theoretical_loss": 3.64448780468676, + "tokens_seen": 1101660160 + }, + { + "epoch": 0.33, + "learning_rate": 0.0005383510531594785, + "loss": 1.2881, + "theoretical_loss": 3.644467698711393, + "tokens_seen": 1101725696 + }, + { + "epoch": 0.33, + "learning_rate": 0.0005383350050150452, + "loss": 1.6108, + "theoretical_loss": 3.644447594266849, + "tokens_seen": 1101791232 + }, + { + "epoch": 0.33, + "learning_rate": 0.0005383189568706119, + "loss": 1.405, + "theoretical_loss": 3.644427491352921, + "tokens_seen": 1101856768 + }, + { + "epoch": 0.33, + "learning_rate": 0.0005383029087261786, + "loss": 1.3255, + "theoretical_loss": 3.6444073899694005, + "tokens_seen": 1101922304 + }, + { + "epoch": 0.33, + "learning_rate": 0.0005382868605817453, + "loss": 1.2299, + "theoretical_loss": 3.6443872901160805, + "tokens_seen": 1101987840 + }, + { + "epoch": 0.33, + "learning_rate": 0.000538270812437312, + "loss": 1.4561, + "theoretical_loss": 3.644367191792754, + "tokens_seen": 1102053376 + }, + { + "epoch": 0.33, + "learning_rate": 0.0005382547642928787, + "loss": 1.2977, + "theoretical_loss": 3.6443470949992127, + "tokens_seen": 1102118912 + }, + { + "epoch": 0.33, + "learning_rate": 0.0005382387161484454, + "loss": 1.4049, + "theoretical_loss": 3.64432699973525, + "tokens_seen": 1102184448 + }, + { + "epoch": 0.33, + "learning_rate": 0.0005382226680040121, + "loss": 1.3962, + "theoretical_loss": 3.644306906000658, + "tokens_seen": 1102249984 + }, + { + "epoch": 0.33, + "learning_rate": 0.0005382066198595788, + "loss": 1.4792, + "theoretical_loss": 3.6442868137952296, + "tokens_seen": 1102315520 + }, + { + "epoch": 0.33, + "learning_rate": 0.0005381905717151455, + "loss": 1.5806, + "theoretical_loss": 3.6442667231187573, + "tokens_seen": 1102381056 + }, + { + "epoch": 0.33, + "learning_rate": 0.0005381745235707122, + "loss": 1.3642, + "theoretical_loss": 3.644246633971034, + "tokens_seen": 1102446592 + }, + { + "epoch": 0.33, + "learning_rate": 0.0005381584754262788, + "loss": 1.4536, + "theoretical_loss": 3.6442265463518533, + "tokens_seen": 1102512128 + }, + { + "epoch": 0.33, + "learning_rate": 0.0005381424272818456, + "loss": 1.547, + "theoretical_loss": 3.6442064602610063, + "tokens_seen": 1102577664 + }, + { + "epoch": 0.33, + "learning_rate": 0.0005381263791374122, + "loss": 1.4308, + "theoretical_loss": 3.644186375698287, + "tokens_seen": 1102643200 + }, + { + "epoch": 0.33, + "learning_rate": 0.000538110330992979, + "loss": 1.498, + "theoretical_loss": 3.6441662926634883, + "tokens_seen": 1102708736 + }, + { + "epoch": 0.33, + "learning_rate": 0.0005380942828485456, + "loss": 1.478, + "theoretical_loss": 3.6441462111564036, + "tokens_seen": 1102774272 + }, + { + "epoch": 0.33, + "learning_rate": 0.0005380782347041124, + "loss": 1.4878, + "theoretical_loss": 3.644126131176825, + "tokens_seen": 1102839808 + }, + { + "epoch": 0.33, + "learning_rate": 0.000538062186559679, + "loss": 1.3117, + "theoretical_loss": 3.6441060527245455, + "tokens_seen": 1102905344 + }, + { + "epoch": 0.33, + "learning_rate": 0.0005380461384152458, + "loss": 1.5689, + "theoretical_loss": 3.6440859757993596, + "tokens_seen": 1102970880 + }, + { + "epoch": 0.33, + "learning_rate": 0.0005380300902708124, + "loss": 1.5017, + "theoretical_loss": 3.644065900401059, + "tokens_seen": 1103036416 + }, + { + "epoch": 0.33, + "learning_rate": 0.0005380140421263792, + "loss": 1.3972, + "theoretical_loss": 3.644045826529437, + "tokens_seen": 1103101952 + }, + { + "epoch": 0.33, + "learning_rate": 0.0005379979939819458, + "loss": 1.3475, + "theoretical_loss": 3.644025754184287, + "tokens_seen": 1103167488 + }, + { + "epoch": 0.33, + "learning_rate": 0.0005379819458375126, + "loss": 1.4182, + "theoretical_loss": 3.644005683365403, + "tokens_seen": 1103233024 + }, + { + "epoch": 0.33, + "learning_rate": 0.0005379658976930792, + "loss": 1.457, + "theoretical_loss": 3.643985614072577, + "tokens_seen": 1103298560 + }, + { + "epoch": 0.33, + "learning_rate": 0.000537949849548646, + "loss": 1.4122, + "theoretical_loss": 3.643965546305603, + "tokens_seen": 1103364096 + }, + { + "epoch": 0.33, + "learning_rate": 0.0005379338014042126, + "loss": 1.3896, + "theoretical_loss": 3.643945480064275, + "tokens_seen": 1103429632 + }, + { + "epoch": 0.33, + "learning_rate": 0.0005379177532597794, + "loss": 1.4804, + "theoretical_loss": 3.6439254153483853, + "tokens_seen": 1103495168 + }, + { + "epoch": 0.33, + "learning_rate": 0.000537901705115346, + "loss": 1.3585, + "theoretical_loss": 3.643905352157728, + "tokens_seen": 1103560704 + }, + { + "epoch": 0.33, + "learning_rate": 0.0005378856569709128, + "loss": 1.3899, + "theoretical_loss": 3.6438852904920966, + "tokens_seen": 1103626240 + }, + { + "epoch": 0.33, + "learning_rate": 0.0005378696088264794, + "loss": 1.3765, + "theoretical_loss": 3.6438652303512846, + "tokens_seen": 1103691776 + }, + { + "epoch": 0.33, + "learning_rate": 0.0005378535606820462, + "loss": 1.4005, + "theoretical_loss": 3.6438451717350846, + "tokens_seen": 1103757312 + }, + { + "epoch": 0.33, + "learning_rate": 0.0005378375125376128, + "loss": 1.4349, + "theoretical_loss": 3.643825114643292, + "tokens_seen": 1103822848 + }, + { + "epoch": 0.33, + "learning_rate": 0.0005378214643931796, + "loss": 1.4485, + "theoretical_loss": 3.6438050590756985, + "tokens_seen": 1103888384 + }, + { + "epoch": 0.33, + "learning_rate": 0.0005378054162487462, + "loss": 1.4226, + "theoretical_loss": 3.6437850050320995, + "tokens_seen": 1103953920 + }, + { + "epoch": 0.33, + "learning_rate": 0.000537789368104313, + "loss": 1.4249, + "theoretical_loss": 3.6437649525122877, + "tokens_seen": 1104019456 + }, + { + "epoch": 0.33, + "learning_rate": 0.0005377733199598796, + "loss": 1.4493, + "theoretical_loss": 3.6437449015160572, + "tokens_seen": 1104084992 + }, + { + "epoch": 0.33, + "learning_rate": 0.0005377572718154464, + "loss": 1.4866, + "theoretical_loss": 3.643724852043202, + "tokens_seen": 1104150528 + }, + { + "epoch": 0.33, + "learning_rate": 0.000537741223671013, + "loss": 1.4928, + "theoretical_loss": 3.6437048040935154, + "tokens_seen": 1104216064 + }, + { + "epoch": 0.33, + "objective/train/docs_used": 404699, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.5437978506088257, + "objective/train/theoretical_loss": 3.643684757666792, + "objective/train/tokens_used": 1124741600, + "theoretical_loss": 3.643684757666792, + "tokens_seen": 1104281600 + }, + { + "epoch": 0.33, + "learning_rate": 0.0005377251755265798, + "loss": 1.4684, + "theoretical_loss": 3.643684757666792, + "tokens_seen": 1104281600 + }, + { + "epoch": 0.33, + "learning_rate": 0.0005377091273821464, + "loss": 1.515, + "theoretical_loss": 3.643664712762825, + "tokens_seen": 1104347136 + }, + { + "epoch": 0.33, + "learning_rate": 0.0005376930792377132, + "loss": 1.4681, + "theoretical_loss": 3.643644669381409, + "tokens_seen": 1104412672 + }, + { + "epoch": 0.33, + "learning_rate": 0.0005376770310932798, + "loss": 1.26, + "theoretical_loss": 3.6436246275223376, + "tokens_seen": 1104478208 + }, + { + "epoch": 0.33, + "learning_rate": 0.0005376609829488466, + "loss": 1.4817, + "theoretical_loss": 3.643604587185405, + "tokens_seen": 1104543744 + }, + { + "epoch": 0.33, + "learning_rate": 0.0005376449348044132, + "loss": 1.5129, + "theoretical_loss": 3.6435845483704057, + "tokens_seen": 1104609280 + }, + { + "epoch": 0.33, + "learning_rate": 0.00053762888665998, + "loss": 1.471, + "theoretical_loss": 3.643564511077133, + "tokens_seen": 1104674816 + }, + { + "epoch": 0.33, + "learning_rate": 0.0005376128385155466, + "loss": 1.3994, + "theoretical_loss": 3.643544475305382, + "tokens_seen": 1104740352 + }, + { + "epoch": 0.33, + "learning_rate": 0.0005375967903711134, + "loss": 1.4332, + "theoretical_loss": 3.6435244410549457, + "tokens_seen": 1104805888 + }, + { + "epoch": 0.33, + "learning_rate": 0.00053758074222668, + "loss": 1.4185, + "theoretical_loss": 3.6435044083256196, + "tokens_seen": 1104871424 + }, + { + "epoch": 0.33, + "learning_rate": 0.0005375646940822468, + "loss": 1.4483, + "theoretical_loss": 3.6434843771171974, + "tokens_seen": 1104936960 + }, + { + "epoch": 0.33, + "learning_rate": 0.0005375486459378134, + "loss": 1.4383, + "theoretical_loss": 3.6434643474294734, + "tokens_seen": 1105002496 + }, + { + "epoch": 0.33, + "learning_rate": 0.0005375325977933802, + "loss": 1.3317, + "theoretical_loss": 3.643444319262242, + "tokens_seen": 1105068032 + }, + { + "epoch": 0.33, + "learning_rate": 0.0005375165496489468, + "loss": 1.4783, + "theoretical_loss": 3.643424292615298, + "tokens_seen": 1105133568 + }, + { + "epoch": 0.33, + "learning_rate": 0.0005375005015045136, + "loss": 1.3839, + "theoretical_loss": 3.6434042674884357, + "tokens_seen": 1105199104 + }, + { + "epoch": 0.33, + "learning_rate": 0.0005374844533600802, + "loss": 1.461, + "theoretical_loss": 3.6433842438814494, + "tokens_seen": 1105264640 + }, + { + "epoch": 0.33, + "learning_rate": 0.000537468405215647, + "loss": 1.414, + "theoretical_loss": 3.6433642217941333, + "tokens_seen": 1105330176 + }, + { + "epoch": 0.33, + "learning_rate": 0.0005374523570712136, + "loss": 1.413, + "theoretical_loss": 3.6433442012262827, + "tokens_seen": 1105395712 + }, + { + "epoch": 0.33, + "learning_rate": 0.0005374363089267804, + "loss": 1.6406, + "theoretical_loss": 3.6433241821776914, + "tokens_seen": 1105461248 + }, + { + "epoch": 0.34, + "learning_rate": 0.000537420260782347, + "loss": 1.4857, + "theoretical_loss": 3.643304164648155, + "tokens_seen": 1105526784 + }, + { + "epoch": 0.34, + "learning_rate": 0.0005374042126379138, + "loss": 1.4448, + "theoretical_loss": 3.643284148637467, + "tokens_seen": 1105592320 + }, + { + "epoch": 0.34, + "learning_rate": 0.0005373881644934804, + "loss": 1.5379, + "theoretical_loss": 3.643264134145424, + "tokens_seen": 1105657856 + }, + { + "epoch": 0.34, + "learning_rate": 0.0005373721163490472, + "loss": 1.3507, + "theoretical_loss": 3.643244121171819, + "tokens_seen": 1105723392 + }, + { + "epoch": 0.34, + "learning_rate": 0.0005373560682046138, + "loss": 1.446, + "theoretical_loss": 3.6432241097164475, + "tokens_seen": 1105788928 + }, + { + "epoch": 0.34, + "learning_rate": 0.0005373400200601806, + "loss": 1.388, + "theoretical_loss": 3.643204099779104, + "tokens_seen": 1105854464 + }, + { + "epoch": 0.34, + "learning_rate": 0.0005373239719157472, + "loss": 1.4083, + "theoretical_loss": 3.643184091359584, + "tokens_seen": 1105920000 + }, + { + "epoch": 0.34, + "learning_rate": 0.000537307923771314, + "loss": 1.5118, + "theoretical_loss": 3.6431640844576823, + "tokens_seen": 1105985536 + }, + { + "epoch": 0.34, + "learning_rate": 0.0005372918756268806, + "loss": 1.3617, + "theoretical_loss": 3.6431440790731933, + "tokens_seen": 1106051072 + }, + { + "epoch": 0.34, + "learning_rate": 0.0005372758274824474, + "loss": 1.4635, + "theoretical_loss": 3.6431240752059124, + "tokens_seen": 1106116608 + }, + { + "epoch": 0.34, + "learning_rate": 0.000537259779338014, + "loss": 1.2922, + "theoretical_loss": 3.643104072855635, + "tokens_seen": 1106182144 + }, + { + "epoch": 0.34, + "learning_rate": 0.0005372437311935808, + "loss": 1.4008, + "theoretical_loss": 3.6430840720221553, + "tokens_seen": 1106247680 + }, + { + "epoch": 0.34, + "learning_rate": 0.0005372276830491474, + "loss": 1.417, + "theoretical_loss": 3.6430640727052697, + "tokens_seen": 1106313216 + }, + { + "epoch": 0.34, + "learning_rate": 0.0005372116349047142, + "loss": 1.3975, + "theoretical_loss": 3.643044074904772, + "tokens_seen": 1106378752 + }, + { + "epoch": 0.34, + "learning_rate": 0.0005371955867602808, + "loss": 1.3662, + "theoretical_loss": 3.6430240786204586, + "tokens_seen": 1106444288 + }, + { + "epoch": 0.34, + "learning_rate": 0.0005371795386158476, + "loss": 1.4612, + "theoretical_loss": 3.643004083852124, + "tokens_seen": 1106509824 + }, + { + "epoch": 0.34, + "learning_rate": 0.0005371634904714142, + "loss": 1.4065, + "theoretical_loss": 3.6429840905995636, + "tokens_seen": 1106575360 + }, + { + "epoch": 0.34, + "learning_rate": 0.000537147442326981, + "loss": 1.397, + "theoretical_loss": 3.6429640988625724, + "tokens_seen": 1106640896 + }, + { + "epoch": 0.34, + "learning_rate": 0.0005371313941825477, + "loss": 1.409, + "theoretical_loss": 3.642944108640947, + "tokens_seen": 1106706432 + }, + { + "epoch": 0.34, + "learning_rate": 0.0005371153460381144, + "loss": 1.4362, + "theoretical_loss": 3.6429241199344817, + "tokens_seen": 1106771968 + }, + { + "epoch": 0.34, + "learning_rate": 0.000537099297893681, + "loss": 1.5481, + "theoretical_loss": 3.6429041327429723, + "tokens_seen": 1106837504 + }, + { + "epoch": 0.34, + "learning_rate": 0.0005370832497492478, + "loss": 1.4103, + "theoretical_loss": 3.642884147066214, + "tokens_seen": 1106903040 + }, + { + "epoch": 0.34, + "learning_rate": 0.0005370672016048145, + "loss": 1.4014, + "theoretical_loss": 3.642864162904003, + "tokens_seen": 1106968576 + }, + { + "epoch": 0.34, + "learning_rate": 0.0005370511534603812, + "loss": 1.4552, + "theoretical_loss": 3.6428441802561347, + "tokens_seen": 1107034112 + }, + { + "epoch": 0.34, + "learning_rate": 0.0005370351053159479, + "loss": 1.4276, + "theoretical_loss": 3.642824199122404, + "tokens_seen": 1107099648 + }, + { + "epoch": 0.34, + "learning_rate": 0.0005370190571715146, + "loss": 1.415, + "theoretical_loss": 3.642804219502607, + "tokens_seen": 1107165184 + }, + { + "epoch": 0.34, + "learning_rate": 0.0005370030090270813, + "loss": 1.4351, + "theoretical_loss": 3.6427842413965394, + "tokens_seen": 1107230720 + }, + { + "epoch": 0.34, + "learning_rate": 0.000536986960882648, + "loss": 1.475, + "theoretical_loss": 3.642764264803997, + "tokens_seen": 1107296256 + }, + { + "epoch": 0.34, + "learning_rate": 0.0005369709127382147, + "loss": 1.3259, + "theoretical_loss": 3.6427442897247753, + "tokens_seen": 1107361792 + }, + { + "epoch": 0.34, + "learning_rate": 0.0005369548645937814, + "loss": 1.3802, + "theoretical_loss": 3.6427243161586706, + "tokens_seen": 1107427328 + }, + { + "epoch": 0.34, + "learning_rate": 0.0005369388164493481, + "loss": 1.3484, + "theoretical_loss": 3.6427043441054785, + "tokens_seen": 1107492864 + }, + { + "epoch": 0.34, + "objective/train/docs_used": 405946, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.4316960573196411, + "objective/train/theoretical_loss": 3.642684373564995, + "objective/train/tokens_used": 1128018400, + "theoretical_loss": 3.642684373564995, + "tokens_seen": 1107558400 + }, + { + "epoch": 0.34, + "learning_rate": 0.0005369227683049148, + "loss": 1.4975, + "theoretical_loss": 3.642684373564995, + "tokens_seen": 1107558400 + }, + { + "epoch": 0.34, + "learning_rate": 0.0005369067201604815, + "loss": 1.4996, + "theoretical_loss": 3.6426644045370153, + "tokens_seen": 1107623936 + }, + { + "epoch": 0.34, + "learning_rate": 0.0005368906720160482, + "loss": 1.4404, + "theoretical_loss": 3.6426444370213362, + "tokens_seen": 1107689472 + }, + { + "epoch": 0.34, + "learning_rate": 0.0005368746238716149, + "loss": 1.5175, + "theoretical_loss": 3.6426244710177533, + "tokens_seen": 1107755008 + }, + { + "epoch": 0.34, + "learning_rate": 0.0005368585757271816, + "loss": 1.44, + "theoretical_loss": 3.6426045065260633, + "tokens_seen": 1107820544 + }, + { + "epoch": 0.34, + "learning_rate": 0.0005368425275827483, + "loss": 1.4411, + "theoretical_loss": 3.642584543546061, + "tokens_seen": 1107886080 + }, + { + "epoch": 0.34, + "learning_rate": 0.000536826479438315, + "loss": 1.467, + "theoretical_loss": 3.642564582077544, + "tokens_seen": 1107951616 + }, + { + "epoch": 0.34, + "learning_rate": 0.0005368104312938817, + "loss": 1.4, + "theoretical_loss": 3.6425446221203073, + "tokens_seen": 1108017152 + }, + { + "epoch": 0.34, + "learning_rate": 0.0005367943831494484, + "loss": 1.3915, + "theoretical_loss": 3.6425246636741475, + "tokens_seen": 1108082688 + }, + { + "epoch": 0.34, + "learning_rate": 0.0005367783350050151, + "loss": 1.4055, + "theoretical_loss": 3.6425047067388614, + "tokens_seen": 1108148224 + }, + { + "epoch": 0.34, + "learning_rate": 0.0005367622868605817, + "loss": 1.3605, + "theoretical_loss": 3.6424847513142447, + "tokens_seen": 1108213760 + }, + { + "epoch": 0.34, + "learning_rate": 0.0005367462387161485, + "loss": 1.4562, + "theoretical_loss": 3.6424647974000934, + "tokens_seen": 1108279296 + }, + { + "epoch": 0.34, + "learning_rate": 0.0005367301905717151, + "loss": 1.3893, + "theoretical_loss": 3.642444844996205, + "tokens_seen": 1108344832 + }, + { + "epoch": 0.34, + "learning_rate": 0.0005367141424272819, + "loss": 1.4103, + "theoretical_loss": 3.6424248941023745, + "tokens_seen": 1108410368 + }, + { + "epoch": 0.34, + "learning_rate": 0.0005366980942828485, + "loss": 1.4386, + "theoretical_loss": 3.642404944718399, + "tokens_seen": 1108475904 + }, + { + "epoch": 0.34, + "learning_rate": 0.0005366820461384153, + "loss": 1.3653, + "theoretical_loss": 3.6423849968440756, + "tokens_seen": 1108541440 + }, + { + "epoch": 0.34, + "learning_rate": 0.0005366659979939819, + "loss": 1.3691, + "theoretical_loss": 3.6423650504791993, + "tokens_seen": 1108606976 + }, + { + "epoch": 0.34, + "learning_rate": 0.0005366499498495487, + "loss": 1.3612, + "theoretical_loss": 3.6423451056235683, + "tokens_seen": 1108672512 + }, + { + "epoch": 0.34, + "learning_rate": 0.0005366339017051153, + "loss": 1.4071, + "theoretical_loss": 3.642325162276978, + "tokens_seen": 1108738048 + }, + { + "epoch": 0.34, + "learning_rate": 0.0005366178535606821, + "loss": 1.4054, + "theoretical_loss": 3.642305220439226, + "tokens_seen": 1108803584 + }, + { + "epoch": 0.34, + "learning_rate": 0.0005366018054162487, + "loss": 1.4932, + "theoretical_loss": 3.6422852801101078, + "tokens_seen": 1108869120 + }, + { + "epoch": 0.34, + "learning_rate": 0.0005365857572718155, + "loss": 1.4859, + "theoretical_loss": 3.6422653412894217, + "tokens_seen": 1108934656 + }, + { + "epoch": 0.34, + "learning_rate": 0.0005365697091273821, + "loss": 1.4, + "theoretical_loss": 3.642245403976963, + "tokens_seen": 1109000192 + }, + { + "epoch": 0.34, + "learning_rate": 0.0005365536609829489, + "loss": 1.4615, + "theoretical_loss": 3.642225468172529, + "tokens_seen": 1109065728 + }, + { + "epoch": 0.34, + "learning_rate": 0.0005365376128385155, + "loss": 1.3133, + "theoretical_loss": 3.6422055338759165, + "tokens_seen": 1109131264 + }, + { + "epoch": 0.34, + "learning_rate": 0.0005365215646940823, + "loss": 1.4773, + "theoretical_loss": 3.642185601086923, + "tokens_seen": 1109196800 + }, + { + "epoch": 0.34, + "learning_rate": 0.0005365055165496489, + "loss": 1.3535, + "theoretical_loss": 3.642165669805344, + "tokens_seen": 1109262336 + }, + { + "epoch": 0.34, + "learning_rate": 0.0005364894684052157, + "loss": 1.4962, + "theoretical_loss": 3.642145740030978, + "tokens_seen": 1109327872 + }, + { + "epoch": 0.34, + "learning_rate": 0.0005364734202607823, + "loss": 1.4582, + "theoretical_loss": 3.6421258117636217, + "tokens_seen": 1109393408 + }, + { + "epoch": 0.34, + "learning_rate": 0.0005364573721163491, + "loss": 1.5485, + "theoretical_loss": 3.6421058850030708, + "tokens_seen": 1109458944 + }, + { + "epoch": 0.34, + "learning_rate": 0.0005364413239719157, + "loss": 1.4743, + "theoretical_loss": 3.642085959749124, + "tokens_seen": 1109524480 + }, + { + "epoch": 0.34, + "learning_rate": 0.0005364252758274825, + "loss": 1.4686, + "theoretical_loss": 3.6420660360015775, + "tokens_seen": 1109590016 + }, + { + "epoch": 0.34, + "learning_rate": 0.0005364092276830491, + "loss": 1.5229, + "theoretical_loss": 3.6420461137602285, + "tokens_seen": 1109655552 + }, + { + "epoch": 0.34, + "learning_rate": 0.0005363931795386159, + "loss": 1.4321, + "theoretical_loss": 3.6420261930248747, + "tokens_seen": 1109721088 + }, + { + "epoch": 0.34, + "learning_rate": 0.0005363771313941825, + "loss": 1.484, + "theoretical_loss": 3.642006273795313, + "tokens_seen": 1109786624 + }, + { + "epoch": 0.34, + "learning_rate": 0.0005363610832497493, + "loss": 1.3881, + "theoretical_loss": 3.64198635607134, + "tokens_seen": 1109852160 + }, + { + "epoch": 0.34, + "learning_rate": 0.0005363450351053159, + "loss": 1.4238, + "theoretical_loss": 3.6419664398527543, + "tokens_seen": 1109917696 + }, + { + "epoch": 0.34, + "learning_rate": 0.0005363289869608827, + "loss": 1.4085, + "theoretical_loss": 3.6419465251393524, + "tokens_seen": 1109983232 + }, + { + "epoch": 0.34, + "learning_rate": 0.0005363129388164493, + "loss": 1.5228, + "theoretical_loss": 3.641926611930932, + "tokens_seen": 1110048768 + }, + { + "epoch": 0.34, + "learning_rate": 0.0005362968906720161, + "loss": 1.508, + "theoretical_loss": 3.64190670022729, + "tokens_seen": 1110114304 + }, + { + "epoch": 0.34, + "learning_rate": 0.0005362808425275827, + "loss": 1.4646, + "theoretical_loss": 3.641886790028225, + "tokens_seen": 1110179840 + }, + { + "epoch": 0.34, + "learning_rate": 0.0005362647943831495, + "loss": 1.3701, + "theoretical_loss": 3.6418668813335335, + "tokens_seen": 1110245376 + }, + { + "epoch": 0.34, + "learning_rate": 0.0005362487462387161, + "loss": 1.5073, + "theoretical_loss": 3.6418469741430135, + "tokens_seen": 1110310912 + }, + { + "epoch": 0.34, + "learning_rate": 0.0005362326980942829, + "loss": 1.4895, + "theoretical_loss": 3.641827068456462, + "tokens_seen": 1110376448 + }, + { + "epoch": 0.34, + "learning_rate": 0.0005362166499498495, + "loss": 1.4608, + "theoretical_loss": 3.6418071642736773, + "tokens_seen": 1110441984 + }, + { + "epoch": 0.34, + "learning_rate": 0.0005362006018054163, + "loss": 1.4658, + "theoretical_loss": 3.6417872615944566, + "tokens_seen": 1110507520 + }, + { + "epoch": 0.34, + "learning_rate": 0.0005361845536609829, + "loss": 1.2965, + "theoretical_loss": 3.641767360418598, + "tokens_seen": 1110573056 + }, + { + "epoch": 0.34, + "learning_rate": 0.0005361685055165497, + "loss": 1.498, + "theoretical_loss": 3.641747460745899, + "tokens_seen": 1110638592 + }, + { + "epoch": 0.34, + "learning_rate": 0.0005361524573721163, + "loss": 1.4204, + "theoretical_loss": 3.6417275625761576, + "tokens_seen": 1110704128 + }, + { + "epoch": 0.34, + "learning_rate": 0.0005361364092276831, + "loss": 1.461, + "theoretical_loss": 3.6417076659091707, + "tokens_seen": 1110769664 + }, + { + "epoch": 0.34, + "objective/train/docs_used": 406993, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.433119773864746, + "objective/train/theoretical_loss": 3.6416877707447375, + "objective/train/tokens_used": 1131295200, + "theoretical_loss": 3.6416877707447375, + "tokens_seen": 1110835200 + }, + { + "epoch": 0.34, + "learning_rate": 0.0005361203610832497, + "loss": 1.3943, + "theoretical_loss": 3.6416877707447375, + "tokens_seen": 1110835200 + }, + { + "epoch": 0.34, + "learning_rate": 0.0005361043129388165, + "loss": 1.4046, + "theoretical_loss": 3.6416678770826554, + "tokens_seen": 1110900736 + }, + { + "epoch": 0.34, + "learning_rate": 0.0005360882647943831, + "loss": 1.5334, + "theoretical_loss": 3.641647984922722, + "tokens_seen": 1110966272 + }, + { + "epoch": 0.34, + "learning_rate": 0.0005360722166499499, + "loss": 1.5853, + "theoretical_loss": 3.6416280942647354, + "tokens_seen": 1111031808 + }, + { + "epoch": 0.34, + "learning_rate": 0.0005360561685055165, + "loss": 1.3841, + "theoretical_loss": 3.6416082051084935, + "tokens_seen": 1111097344 + }, + { + "epoch": 0.34, + "learning_rate": 0.0005360401203610833, + "loss": 1.4109, + "theoretical_loss": 3.641588317453795, + "tokens_seen": 1111162880 + }, + { + "epoch": 0.34, + "learning_rate": 0.0005360240722166499, + "loss": 1.5253, + "theoretical_loss": 3.641568431300437, + "tokens_seen": 1111228416 + }, + { + "epoch": 0.34, + "learning_rate": 0.0005360080240722167, + "loss": 1.488, + "theoretical_loss": 3.6415485466482185, + "tokens_seen": 1111293952 + }, + { + "epoch": 0.34, + "learning_rate": 0.0005359919759277833, + "loss": 1.6076, + "theoretical_loss": 3.6415286634969375, + "tokens_seen": 1111359488 + }, + { + "epoch": 0.34, + "learning_rate": 0.0005359759277833501, + "loss": 1.4654, + "theoretical_loss": 3.6415087818463916, + "tokens_seen": 1111425024 + }, + { + "epoch": 0.34, + "learning_rate": 0.0005359598796389167, + "loss": 1.4868, + "theoretical_loss": 3.6414889016963796, + "tokens_seen": 1111490560 + }, + { + "epoch": 0.34, + "learning_rate": 0.0005359438314944835, + "loss": 1.4799, + "theoretical_loss": 3.6414690230467, + "tokens_seen": 1111556096 + }, + { + "epoch": 0.34, + "learning_rate": 0.0005359277833500501, + "loss": 1.4864, + "theoretical_loss": 3.6414491458971505, + "tokens_seen": 1111621632 + }, + { + "epoch": 0.34, + "learning_rate": 0.0005359117352056169, + "loss": 1.4484, + "theoretical_loss": 3.6414292702475297, + "tokens_seen": 1111687168 + }, + { + "epoch": 0.34, + "learning_rate": 0.0005358956870611835, + "loss": 1.4292, + "theoretical_loss": 3.641409396097636, + "tokens_seen": 1111752704 + }, + { + "epoch": 0.34, + "learning_rate": 0.0005358796389167503, + "loss": 1.5239, + "theoretical_loss": 3.641389523447268, + "tokens_seen": 1111818240 + }, + { + "epoch": 0.34, + "learning_rate": 0.000535863590772317, + "loss": 1.4991, + "theoretical_loss": 3.6413696522962242, + "tokens_seen": 1111883776 + }, + { + "epoch": 0.34, + "learning_rate": 0.0005358475426278837, + "loss": 1.4215, + "theoretical_loss": 3.641349782644303, + "tokens_seen": 1111949312 + }, + { + "epoch": 0.34, + "learning_rate": 0.0005358314944834503, + "loss": 1.4225, + "theoretical_loss": 3.641329914491303, + "tokens_seen": 1112014848 + }, + { + "epoch": 0.34, + "learning_rate": 0.0005358154463390171, + "loss": 1.4615, + "theoretical_loss": 3.6413100478370226, + "tokens_seen": 1112080384 + }, + { + "epoch": 0.34, + "learning_rate": 0.0005357993981945837, + "loss": 1.3968, + "theoretical_loss": 3.6412901826812605, + "tokens_seen": 1112145920 + }, + { + "epoch": 0.34, + "learning_rate": 0.0005357833500501505, + "loss": 1.5071, + "theoretical_loss": 3.641270319023816, + "tokens_seen": 1112211456 + }, + { + "epoch": 0.34, + "learning_rate": 0.0005357673019057172, + "loss": 1.5182, + "theoretical_loss": 3.6412504568644866, + "tokens_seen": 1112276992 + }, + { + "epoch": 0.34, + "learning_rate": 0.0005357512537612839, + "loss": 1.3989, + "theoretical_loss": 3.641230596203072, + "tokens_seen": 1112342528 + }, + { + "epoch": 0.34, + "learning_rate": 0.0005357352056168507, + "loss": 1.4283, + "theoretical_loss": 3.6412107370393705, + "tokens_seen": 1112408064 + }, + { + "epoch": 0.34, + "learning_rate": 0.0005357191574724173, + "loss": 1.3278, + "theoretical_loss": 3.6411908793731813, + "tokens_seen": 1112473600 + }, + { + "epoch": 0.34, + "learning_rate": 0.0005357031093279841, + "loss": 1.4949, + "theoretical_loss": 3.6411710232043033, + "tokens_seen": 1112539136 + }, + { + "epoch": 0.34, + "learning_rate": 0.0005356870611835507, + "loss": 1.4217, + "theoretical_loss": 3.641151168532535, + "tokens_seen": 1112604672 + }, + { + "epoch": 0.34, + "learning_rate": 0.0005356710130391175, + "loss": 1.4543, + "theoretical_loss": 3.6411313153576756, + "tokens_seen": 1112670208 + }, + { + "epoch": 0.34, + "learning_rate": 0.0005356549648946841, + "loss": 1.4987, + "theoretical_loss": 3.6411114636795237, + "tokens_seen": 1112735744 + }, + { + "epoch": 0.34, + "learning_rate": 0.0005356389167502509, + "loss": 1.4438, + "theoretical_loss": 3.641091613497879, + "tokens_seen": 1112801280 + }, + { + "epoch": 0.34, + "learning_rate": 0.0005356228686058175, + "loss": 1.4764, + "theoretical_loss": 3.6410717648125406, + "tokens_seen": 1112866816 + }, + { + "epoch": 0.34, + "learning_rate": 0.0005356068204613843, + "loss": 1.4551, + "theoretical_loss": 3.6410519176233067, + "tokens_seen": 1112932352 + }, + { + "epoch": 0.34, + "learning_rate": 0.0005355907723169509, + "loss": 1.3366, + "theoretical_loss": 3.641032071929977, + "tokens_seen": 1112997888 + }, + { + "epoch": 0.34, + "learning_rate": 0.0005355747241725177, + "loss": 1.3658, + "theoretical_loss": 3.641012227732351, + "tokens_seen": 1113063424 + }, + { + "epoch": 0.34, + "learning_rate": 0.0005355586760280843, + "loss": 1.4455, + "theoretical_loss": 3.6409923850302275, + "tokens_seen": 1113128960 + }, + { + "epoch": 0.34, + "learning_rate": 0.0005355426278836511, + "loss": 1.4263, + "theoretical_loss": 3.640972543823406, + "tokens_seen": 1113194496 + }, + { + "epoch": 0.34, + "learning_rate": 0.0005355265797392177, + "loss": 1.4092, + "theoretical_loss": 3.6409527041116854, + "tokens_seen": 1113260032 + }, + { + "epoch": 0.34, + "learning_rate": 0.0005355105315947845, + "loss": 1.4567, + "theoretical_loss": 3.640932865894865, + "tokens_seen": 1113325568 + }, + { + "epoch": 0.34, + "learning_rate": 0.0005354944834503511, + "loss": 1.5959, + "theoretical_loss": 3.640913029172745, + "tokens_seen": 1113391104 + }, + { + "epoch": 0.34, + "learning_rate": 0.0005354784353059178, + "loss": 1.3175, + "theoretical_loss": 3.640893193945124, + "tokens_seen": 1113456640 + }, + { + "epoch": 0.34, + "learning_rate": 0.0005354623871614845, + "loss": 1.4707, + "theoretical_loss": 3.6408733602118017, + "tokens_seen": 1113522176 + }, + { + "epoch": 0.34, + "learning_rate": 0.0005354463390170512, + "loss": 1.5648, + "theoretical_loss": 3.6408535279725776, + "tokens_seen": 1113587712 + }, + { + "epoch": 0.34, + "learning_rate": 0.0005354302908726179, + "loss": 1.4719, + "theoretical_loss": 3.640833697227251, + "tokens_seen": 1113653248 + }, + { + "epoch": 0.34, + "learning_rate": 0.0005354142427281846, + "loss": 1.4262, + "theoretical_loss": 3.640813867975622, + "tokens_seen": 1113718784 + }, + { + "epoch": 0.34, + "learning_rate": 0.0005353981945837513, + "loss": 1.5737, + "theoretical_loss": 3.6407940402174894, + "tokens_seen": 1113784320 + }, + { + "epoch": 0.34, + "learning_rate": 0.000535382146439318, + "loss": 1.5789, + "theoretical_loss": 3.640774213952654, + "tokens_seen": 1113849856 + }, + { + "epoch": 0.34, + "learning_rate": 0.0005353660982948846, + "loss": 1.4776, + "theoretical_loss": 3.6407543891809144, + "tokens_seen": 1113915392 + }, + { + "epoch": 0.34, + "learning_rate": 0.0005353500501504514, + "loss": 1.383, + "theoretical_loss": 3.640734565902071, + "tokens_seen": 1113980928 + }, + { + "epoch": 0.34, + "learning_rate": 0.000535334002006018, + "loss": 1.4147, + "theoretical_loss": 3.6407147441159236, + "tokens_seen": 1114046464 + }, + { + "epoch": 0.34, + "objective/train/docs_used": 408181, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.389655590057373, + "objective/train/theoretical_loss": 3.640694923822271, + "objective/train/tokens_used": 1134572000, + "theoretical_loss": 3.640694923822271, + "tokens_seen": 1114112000 + }, + { + "epoch": 0.34, + "learning_rate": 0.0005353179538615848, + "loss": 1.3559, + "theoretical_loss": 3.640694923822271, + "tokens_seen": 1114112000 + }, + { + "epoch": 0.34, + "learning_rate": 0.0005353019057171514, + "loss": 1.4409, + "theoretical_loss": 3.6406751050209145, + "tokens_seen": 1114177536 + }, + { + "epoch": 0.34, + "learning_rate": 0.0005352858575727182, + "loss": 1.4933, + "theoretical_loss": 3.6406552877116525, + "tokens_seen": 1114243072 + }, + { + "epoch": 0.34, + "learning_rate": 0.0005352698094282848, + "loss": 1.4258, + "theoretical_loss": 3.640635471894286, + "tokens_seen": 1114308608 + }, + { + "epoch": 0.34, + "learning_rate": 0.0005352537612838516, + "loss": 1.4654, + "theoretical_loss": 3.640615657568614, + "tokens_seen": 1114374144 + }, + { + "epoch": 0.34, + "learning_rate": 0.0005352377131394182, + "loss": 1.3988, + "theoretical_loss": 3.640595844734438, + "tokens_seen": 1114439680 + }, + { + "epoch": 0.34, + "learning_rate": 0.000535221664994985, + "loss": 1.5301, + "theoretical_loss": 3.6405760333915564, + "tokens_seen": 1114505216 + }, + { + "epoch": 0.34, + "learning_rate": 0.0005352056168505516, + "loss": 1.4127, + "theoretical_loss": 3.640556223539771, + "tokens_seen": 1114570752 + }, + { + "epoch": 0.34, + "learning_rate": 0.0005351895687061184, + "loss": 1.6889, + "theoretical_loss": 3.6405364151788797, + "tokens_seen": 1114636288 + }, + { + "epoch": 0.34, + "learning_rate": 0.000535173520561685, + "loss": 1.543, + "theoretical_loss": 3.640516608308684, + "tokens_seen": 1114701824 + }, + { + "epoch": 0.34, + "learning_rate": 0.0005351574724172518, + "loss": 1.4445, + "theoretical_loss": 3.6404968029289844, + "tokens_seen": 1114767360 + }, + { + "epoch": 0.34, + "learning_rate": 0.0005351414242728184, + "loss": 1.3909, + "theoretical_loss": 3.6404769990395804, + "tokens_seen": 1114832896 + }, + { + "epoch": 0.34, + "learning_rate": 0.0005351253761283852, + "loss": 1.4772, + "theoretical_loss": 3.6404571966402726, + "tokens_seen": 1114898432 + }, + { + "epoch": 0.34, + "learning_rate": 0.0005351093279839518, + "loss": 1.5025, + "theoretical_loss": 3.640437395730861, + "tokens_seen": 1114963968 + }, + { + "epoch": 0.34, + "learning_rate": 0.0005350932798395186, + "loss": 1.4109, + "theoretical_loss": 3.6404175963111465, + "tokens_seen": 1115029504 + }, + { + "epoch": 0.34, + "learning_rate": 0.0005350772316950852, + "loss": 1.3717, + "theoretical_loss": 3.640397798380929, + "tokens_seen": 1115095040 + }, + { + "epoch": 0.34, + "learning_rate": 0.000535061183550652, + "loss": 1.2997, + "theoretical_loss": 3.640378001940009, + "tokens_seen": 1115160576 + }, + { + "epoch": 0.34, + "learning_rate": 0.0005350451354062186, + "loss": 1.4125, + "theoretical_loss": 3.6403582069881866, + "tokens_seen": 1115226112 + }, + { + "epoch": 0.34, + "learning_rate": 0.0005350290872617854, + "loss": 1.4811, + "theoretical_loss": 3.640338413525263, + "tokens_seen": 1115291648 + }, + { + "epoch": 0.34, + "learning_rate": 0.000535013039117352, + "loss": 1.3751, + "theoretical_loss": 3.640318621551039, + "tokens_seen": 1115357184 + }, + { + "epoch": 0.34, + "learning_rate": 0.0005349969909729188, + "loss": 1.3203, + "theoretical_loss": 3.640298831065314, + "tokens_seen": 1115422720 + }, + { + "epoch": 0.34, + "learning_rate": 0.0005349809428284854, + "loss": 1.4386, + "theoretical_loss": 3.6402790420678888, + "tokens_seen": 1115488256 + }, + { + "epoch": 0.34, + "learning_rate": 0.0005349648946840522, + "loss": 1.4785, + "theoretical_loss": 3.640259254558565, + "tokens_seen": 1115553792 + }, + { + "epoch": 0.34, + "learning_rate": 0.0005349488465396188, + "loss": 1.5756, + "theoretical_loss": 3.6402394685371426, + "tokens_seen": 1115619328 + }, + { + "epoch": 0.34, + "learning_rate": 0.0005349327983951856, + "loss": 1.5429, + "theoretical_loss": 3.640219684003422, + "tokens_seen": 1115684864 + }, + { + "epoch": 0.34, + "learning_rate": 0.0005349167502507522, + "loss": 1.4342, + "theoretical_loss": 3.640199900957205, + "tokens_seen": 1115750400 + }, + { + "epoch": 0.34, + "learning_rate": 0.000534900702106319, + "loss": 1.4454, + "theoretical_loss": 3.6401801193982917, + "tokens_seen": 1115815936 + }, + { + "epoch": 0.34, + "learning_rate": 0.0005348846539618856, + "loss": 1.4591, + "theoretical_loss": 3.6401603393264828, + "tokens_seen": 1115881472 + }, + { + "epoch": 0.34, + "learning_rate": 0.0005348686058174524, + "loss": 1.4101, + "theoretical_loss": 3.6401405607415795, + "tokens_seen": 1115947008 + }, + { + "epoch": 0.34, + "learning_rate": 0.000534852557673019, + "loss": 1.4275, + "theoretical_loss": 3.6401207836433827, + "tokens_seen": 1116012544 + }, + { + "epoch": 0.34, + "learning_rate": 0.0005348365095285858, + "loss": 1.3843, + "theoretical_loss": 3.640101008031693, + "tokens_seen": 1116078080 + }, + { + "epoch": 0.34, + "learning_rate": 0.0005348204613841524, + "loss": 1.3182, + "theoretical_loss": 3.6400812339063124, + "tokens_seen": 1116143616 + }, + { + "epoch": 0.34, + "learning_rate": 0.0005348044132397192, + "loss": 1.3619, + "theoretical_loss": 3.6400614612670408, + "tokens_seen": 1116209152 + }, + { + "epoch": 0.34, + "learning_rate": 0.0005347883650952858, + "loss": 1.4824, + "theoretical_loss": 3.6400416901136796, + "tokens_seen": 1116274688 + }, + { + "epoch": 0.34, + "learning_rate": 0.0005347723169508526, + "loss": 1.5199, + "theoretical_loss": 3.6400219204460305, + "tokens_seen": 1116340224 + }, + { + "epoch": 0.34, + "learning_rate": 0.0005347562688064192, + "loss": 1.4376, + "theoretical_loss": 3.6400021522638935, + "tokens_seen": 1116405760 + }, + { + "epoch": 0.34, + "learning_rate": 0.000534740220661986, + "loss": 1.383, + "theoretical_loss": 3.6399823855670705, + "tokens_seen": 1116471296 + }, + { + "epoch": 0.34, + "learning_rate": 0.0005347241725175526, + "loss": 1.4959, + "theoretical_loss": 3.6399626203553628, + "tokens_seen": 1116536832 + }, + { + "epoch": 0.34, + "learning_rate": 0.0005347081243731194, + "loss": 1.4139, + "theoretical_loss": 3.639942856628571, + "tokens_seen": 1116602368 + }, + { + "epoch": 0.34, + "learning_rate": 0.000534692076228686, + "loss": 1.4992, + "theoretical_loss": 3.6399230943864977, + "tokens_seen": 1116667904 + }, + { + "epoch": 0.34, + "learning_rate": 0.0005346760280842528, + "loss": 1.4976, + "theoretical_loss": 3.639903333628943, + "tokens_seen": 1116733440 + }, + { + "epoch": 0.34, + "learning_rate": 0.0005346599799398194, + "loss": 1.3583, + "theoretical_loss": 3.639883574355709, + "tokens_seen": 1116798976 + }, + { + "epoch": 0.34, + "learning_rate": 0.0005346439317953862, + "loss": 1.5132, + "theoretical_loss": 3.6398638165665966, + "tokens_seen": 1116864512 + }, + { + "epoch": 0.34, + "learning_rate": 0.0005346278836509528, + "loss": 1.4629, + "theoretical_loss": 3.6398440602614075, + "tokens_seen": 1116930048 + }, + { + "epoch": 0.34, + "learning_rate": 0.0005346118355065196, + "loss": 1.4207, + "theoretical_loss": 3.6398243054399426, + "tokens_seen": 1116995584 + }, + { + "epoch": 0.34, + "learning_rate": 0.0005345957873620862, + "loss": 1.3779, + "theoretical_loss": 3.639804552102005, + "tokens_seen": 1117061120 + }, + { + "epoch": 0.34, + "learning_rate": 0.000534579739217653, + "loss": 1.4029, + "theoretical_loss": 3.6397848002473943, + "tokens_seen": 1117126656 + }, + { + "epoch": 0.34, + "learning_rate": 0.0005345636910732196, + "loss": 1.424, + "theoretical_loss": 3.6397650498759138, + "tokens_seen": 1117192192 + }, + { + "epoch": 0.34, + "learning_rate": 0.0005345476429287864, + "loss": 1.4314, + "theoretical_loss": 3.639745300987364, + "tokens_seen": 1117257728 + }, + { + "epoch": 0.34, + "learning_rate": 0.0005345315947843532, + "loss": 1.3832, + "theoretical_loss": 3.639725553581547, + "tokens_seen": 1117323264 + }, + { + "epoch": 0.34, + "objective/train/docs_used": 409409, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.500030517578125, + "objective/train/theoretical_loss": 3.639705807658265, + "objective/train/tokens_used": 1137848800, + "theoretical_loss": 3.639705807658265, + "tokens_seen": 1117388800 + }, + { + "epoch": 0.34, + "learning_rate": 0.0005345155466399198, + "loss": 1.3922, + "theoretical_loss": 3.639705807658265, + "tokens_seen": 1117388800 + }, + { + "epoch": 0.34, + "learning_rate": 0.0005344994984954866, + "loss": 1.4935, + "theoretical_loss": 3.6396860632173187, + "tokens_seen": 1117454336 + }, + { + "epoch": 0.34, + "learning_rate": 0.0005344834503510532, + "loss": 1.4067, + "theoretical_loss": 3.6396663202585104, + "tokens_seen": 1117519872 + }, + { + "epoch": 0.34, + "learning_rate": 0.00053446740220662, + "loss": 1.4191, + "theoretical_loss": 3.639646578781642, + "tokens_seen": 1117585408 + }, + { + "epoch": 0.34, + "learning_rate": 0.0005344513540621866, + "loss": 1.4649, + "theoretical_loss": 3.6396268387865156, + "tokens_seen": 1117650944 + }, + { + "epoch": 0.34, + "learning_rate": 0.0005344353059177534, + "loss": 1.5277, + "theoretical_loss": 3.639607100272933, + "tokens_seen": 1117716480 + }, + { + "epoch": 0.34, + "learning_rate": 0.00053441925777332, + "loss": 1.5136, + "theoretical_loss": 3.639587363240696, + "tokens_seen": 1117782016 + }, + { + "epoch": 0.34, + "learning_rate": 0.0005344032096288868, + "loss": 1.4048, + "theoretical_loss": 3.6395676276896065, + "tokens_seen": 1117847552 + }, + { + "epoch": 0.34, + "learning_rate": 0.0005343871614844534, + "loss": 1.4366, + "theoretical_loss": 3.6395478936194663, + "tokens_seen": 1117913088 + }, + { + "epoch": 0.34, + "learning_rate": 0.0005343711133400202, + "loss": 1.5067, + "theoretical_loss": 3.6395281610300776, + "tokens_seen": 1117978624 + }, + { + "epoch": 0.34, + "learning_rate": 0.0005343550651955868, + "loss": 1.4046, + "theoretical_loss": 3.6395084299212432, + "tokens_seen": 1118044160 + }, + { + "epoch": 0.34, + "learning_rate": 0.0005343390170511536, + "loss": 1.4465, + "theoretical_loss": 3.6394887002927647, + "tokens_seen": 1118109696 + }, + { + "epoch": 0.34, + "learning_rate": 0.0005343229689067202, + "loss": 1.4575, + "theoretical_loss": 3.639468972144444, + "tokens_seen": 1118175232 + }, + { + "epoch": 0.34, + "learning_rate": 0.000534306920762287, + "loss": 1.4083, + "theoretical_loss": 3.639449245476084, + "tokens_seen": 1118240768 + }, + { + "epoch": 0.34, + "learning_rate": 0.0005342908726178536, + "loss": 1.4345, + "theoretical_loss": 3.6394295202874862, + "tokens_seen": 1118306304 + }, + { + "epoch": 0.34, + "learning_rate": 0.0005342748244734204, + "loss": 1.483, + "theoretical_loss": 3.6394097965784535, + "tokens_seen": 1118371840 + }, + { + "epoch": 0.34, + "learning_rate": 0.000534258776328987, + "loss": 1.4232, + "theoretical_loss": 3.6393900743487877, + "tokens_seen": 1118437376 + }, + { + "epoch": 0.34, + "learning_rate": 0.0005342427281845538, + "loss": 1.544, + "theoretical_loss": 3.6393703535982915, + "tokens_seen": 1118502912 + }, + { + "epoch": 0.34, + "learning_rate": 0.0005342266800401204, + "loss": 1.4075, + "theoretical_loss": 3.6393506343267674, + "tokens_seen": 1118568448 + }, + { + "epoch": 0.34, + "learning_rate": 0.0005342106318956872, + "loss": 1.4275, + "theoretical_loss": 3.6393309165340177, + "tokens_seen": 1118633984 + }, + { + "epoch": 0.34, + "learning_rate": 0.0005341945837512538, + "loss": 1.3958, + "theoretical_loss": 3.6393112002198444, + "tokens_seen": 1118699520 + }, + { + "epoch": 0.34, + "learning_rate": 0.0005341785356068206, + "loss": 1.501, + "theoretical_loss": 3.6392914853840512, + "tokens_seen": 1118765056 + }, + { + "epoch": 0.34, + "learning_rate": 0.0005341624874623872, + "loss": 1.4244, + "theoretical_loss": 3.639271772026439, + "tokens_seen": 1118830592 + }, + { + "epoch": 0.34, + "learning_rate": 0.000534146439317954, + "loss": 1.5635, + "theoretical_loss": 3.6392520601468124, + "tokens_seen": 1118896128 + }, + { + "epoch": 0.34, + "learning_rate": 0.0005341303911735206, + "loss": 1.3737, + "theoretical_loss": 3.6392323497449723, + "tokens_seen": 1118961664 + }, + { + "epoch": 0.34, + "learning_rate": 0.0005341143430290874, + "loss": 1.4577, + "theoretical_loss": 3.6392126408207224, + "tokens_seen": 1119027200 + }, + { + "epoch": 0.34, + "learning_rate": 0.000534098294884654, + "loss": 1.3806, + "theoretical_loss": 3.6391929333738644, + "tokens_seen": 1119092736 + }, + { + "epoch": 0.34, + "learning_rate": 0.0005340822467402207, + "loss": 1.4772, + "theoretical_loss": 3.639173227404202, + "tokens_seen": 1119158272 + }, + { + "epoch": 0.34, + "learning_rate": 0.0005340661985957874, + "loss": 1.4934, + "theoretical_loss": 3.639153522911538, + "tokens_seen": 1119223808 + }, + { + "epoch": 0.34, + "learning_rate": 0.0005340501504513541, + "loss": 1.3902, + "theoretical_loss": 3.6391338198956746, + "tokens_seen": 1119289344 + }, + { + "epoch": 0.34, + "learning_rate": 0.0005340341023069208, + "loss": 1.5248, + "theoretical_loss": 3.639114118356415, + "tokens_seen": 1119354880 + }, + { + "epoch": 0.34, + "learning_rate": 0.0005340180541624875, + "loss": 1.5289, + "theoretical_loss": 3.6390944182935616, + "tokens_seen": 1119420416 + }, + { + "epoch": 0.34, + "learning_rate": 0.0005340020060180542, + "loss": 1.5543, + "theoretical_loss": 3.639074719706919, + "tokens_seen": 1119485952 + }, + { + "epoch": 0.34, + "learning_rate": 0.0005339859578736209, + "loss": 1.413, + "theoretical_loss": 3.6390550225962874, + "tokens_seen": 1119551488 + }, + { + "epoch": 0.34, + "learning_rate": 0.0005339699097291876, + "loss": 1.4657, + "theoretical_loss": 3.6390353269614724, + "tokens_seen": 1119617024 + }, + { + "epoch": 0.34, + "learning_rate": 0.0005339538615847543, + "loss": 1.5302, + "theoretical_loss": 3.6390156328022756, + "tokens_seen": 1119682560 + }, + { + "epoch": 0.34, + "learning_rate": 0.0005339378134403209, + "loss": 1.4071, + "theoretical_loss": 3.638995940118501, + "tokens_seen": 1119748096 + }, + { + "epoch": 0.34, + "learning_rate": 0.0005339217652958877, + "loss": 1.4431, + "theoretical_loss": 3.6389762489099504, + "tokens_seen": 1119813632 + }, + { + "epoch": 0.34, + "learning_rate": 0.0005339057171514543, + "loss": 1.4862, + "theoretical_loss": 3.638956559176428, + "tokens_seen": 1119879168 + }, + { + "epoch": 0.34, + "learning_rate": 0.0005338896690070211, + "loss": 1.4711, + "theoretical_loss": 3.638936870917737, + "tokens_seen": 1119944704 + }, + { + "epoch": 0.34, + "learning_rate": 0.0005338736208625877, + "loss": 1.5324, + "theoretical_loss": 3.63891718413368, + "tokens_seen": 1120010240 + }, + { + "epoch": 0.34, + "learning_rate": 0.0005338575727181545, + "loss": 1.3883, + "theoretical_loss": 3.638897498824061, + "tokens_seen": 1120075776 + }, + { + "epoch": 0.34, + "learning_rate": 0.0005338415245737211, + "loss": 1.5403, + "theoretical_loss": 3.6388778149886836, + "tokens_seen": 1120141312 + }, + { + "epoch": 0.34, + "learning_rate": 0.0005338254764292879, + "loss": 1.4552, + "theoretical_loss": 3.63885813262735, + "tokens_seen": 1120206848 + }, + { + "epoch": 0.34, + "learning_rate": 0.0005338094282848545, + "loss": 1.3967, + "theoretical_loss": 3.638838451739864, + "tokens_seen": 1120272384 + }, + { + "epoch": 0.34, + "learning_rate": 0.0005337933801404213, + "loss": 1.3773, + "theoretical_loss": 3.6388187723260286, + "tokens_seen": 1120337920 + }, + { + "epoch": 0.34, + "learning_rate": 0.0005337773319959879, + "loss": 1.4688, + "theoretical_loss": 3.6387990943856487, + "tokens_seen": 1120403456 + }, + { + "epoch": 0.34, + "learning_rate": 0.0005337612838515547, + "loss": 1.4657, + "theoretical_loss": 3.638779417918527, + "tokens_seen": 1120468992 + }, + { + "epoch": 0.34, + "learning_rate": 0.0005337452357071213, + "loss": 1.3157, + "theoretical_loss": 3.6387597429244662, + "tokens_seen": 1120534528 + }, + { + "epoch": 0.34, + "learning_rate": 0.0005337291875626881, + "loss": 1.4757, + "theoretical_loss": 3.638740069403271, + "tokens_seen": 1120600064 + }, + { + "epoch": 0.34, + "objective/train/docs_used": 410712, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.4097506999969482, + "objective/train/theoretical_loss": 3.638720397354745, + "objective/train/tokens_used": 1141125600, + "theoretical_loss": 3.638720397354745, + "tokens_seen": 1120665600 + }, + { + "epoch": 0.34, + "learning_rate": 0.0005337131394182547, + "loss": 1.4244, + "theoretical_loss": 3.638720397354745, + "tokens_seen": 1120665600 + }, + { + "epoch": 0.34, + "learning_rate": 0.0005336970912738215, + "loss": 1.5121, + "theoretical_loss": 3.638700726778691, + "tokens_seen": 1120731136 + }, + { + "epoch": 0.34, + "learning_rate": 0.0005336810431293881, + "loss": 1.4875, + "theoretical_loss": 3.6386810576749133, + "tokens_seen": 1120796672 + }, + { + "epoch": 0.34, + "learning_rate": 0.0005336649949849549, + "loss": 1.4136, + "theoretical_loss": 3.638661390043216, + "tokens_seen": 1120862208 + }, + { + "epoch": 0.34, + "learning_rate": 0.0005336489468405215, + "loss": 1.3591, + "theoretical_loss": 3.638641723883402, + "tokens_seen": 1120927744 + }, + { + "epoch": 0.34, + "learning_rate": 0.0005336328986960883, + "loss": 1.5203, + "theoretical_loss": 3.638622059195275, + "tokens_seen": 1120993280 + }, + { + "epoch": 0.34, + "learning_rate": 0.0005336168505516549, + "loss": 1.5095, + "theoretical_loss": 3.6386023959786398, + "tokens_seen": 1121058816 + }, + { + "epoch": 0.34, + "learning_rate": 0.0005336008024072217, + "loss": 1.445, + "theoretical_loss": 3.6385827342333, + "tokens_seen": 1121124352 + }, + { + "epoch": 0.34, + "learning_rate": 0.0005335847542627883, + "loss": 1.5057, + "theoretical_loss": 3.638563073959059, + "tokens_seen": 1121189888 + }, + { + "epoch": 0.34, + "learning_rate": 0.0005335687061183551, + "loss": 1.3359, + "theoretical_loss": 3.638543415155721, + "tokens_seen": 1121255424 + }, + { + "epoch": 0.34, + "learning_rate": 0.0005335526579739217, + "loss": 1.5221, + "theoretical_loss": 3.6385237578230907, + "tokens_seen": 1121320960 + }, + { + "epoch": 0.34, + "learning_rate": 0.0005335366098294885, + "loss": 1.424, + "theoretical_loss": 3.6385041019609705, + "tokens_seen": 1121386496 + }, + { + "epoch": 0.34, + "learning_rate": 0.0005335205616850551, + "loss": 1.4353, + "theoretical_loss": 3.6384844475691662, + "tokens_seen": 1121452032 + }, + { + "epoch": 0.34, + "learning_rate": 0.0005335045135406219, + "loss": 1.4709, + "theoretical_loss": 3.638464794647481, + "tokens_seen": 1121517568 + }, + { + "epoch": 0.34, + "learning_rate": 0.0005334884653961885, + "loss": 1.4107, + "theoretical_loss": 3.638445143195719, + "tokens_seen": 1121583104 + }, + { + "epoch": 0.34, + "learning_rate": 0.0005334724172517553, + "loss": 1.4187, + "theoretical_loss": 3.638425493213685, + "tokens_seen": 1121648640 + }, + { + "epoch": 0.34, + "learning_rate": 0.0005334563691073219, + "loss": 1.3331, + "theoretical_loss": 3.6384058447011824, + "tokens_seen": 1121714176 + }, + { + "epoch": 0.34, + "learning_rate": 0.0005334403209628887, + "loss": 1.6352, + "theoretical_loss": 3.638386197658016, + "tokens_seen": 1121779712 + }, + { + "epoch": 0.34, + "learning_rate": 0.0005334242728184553, + "loss": 1.4031, + "theoretical_loss": 3.6383665520839896, + "tokens_seen": 1121845248 + }, + { + "epoch": 0.34, + "learning_rate": 0.0005334082246740221, + "loss": 1.4264, + "theoretical_loss": 3.638346907978908, + "tokens_seen": 1121910784 + }, + { + "epoch": 0.34, + "learning_rate": 0.0005333921765295887, + "loss": 1.5021, + "theoretical_loss": 3.638327265342576, + "tokens_seen": 1121976320 + }, + { + "epoch": 0.34, + "learning_rate": 0.0005333761283851555, + "loss": 1.2949, + "theoretical_loss": 3.6383076241747965, + "tokens_seen": 1122041856 + }, + { + "epoch": 0.34, + "learning_rate": 0.0005333600802407221, + "loss": 1.5076, + "theoretical_loss": 3.638287984475375, + "tokens_seen": 1122107392 + }, + { + "epoch": 0.34, + "learning_rate": 0.0005333440320962889, + "loss": 1.5444, + "theoretical_loss": 3.6382683462441165, + "tokens_seen": 1122172928 + }, + { + "epoch": 0.34, + "learning_rate": 0.0005333279839518556, + "loss": 1.5139, + "theoretical_loss": 3.6382487094808242, + "tokens_seen": 1122238464 + }, + { + "epoch": 0.34, + "learning_rate": 0.0005333119358074223, + "loss": 1.4452, + "theoretical_loss": 3.638229074185304, + "tokens_seen": 1122304000 + }, + { + "epoch": 0.34, + "learning_rate": 0.000533295887662989, + "loss": 1.5105, + "theoretical_loss": 3.6382094403573593, + "tokens_seen": 1122369536 + }, + { + "epoch": 0.34, + "learning_rate": 0.0005332798395185557, + "loss": 1.3923, + "theoretical_loss": 3.6381898079967954, + "tokens_seen": 1122435072 + }, + { + "epoch": 0.34, + "learning_rate": 0.0005332637913741224, + "loss": 1.4073, + "theoretical_loss": 3.6381701771034165, + "tokens_seen": 1122500608 + }, + { + "epoch": 0.34, + "learning_rate": 0.0005332477432296891, + "loss": 1.4358, + "theoretical_loss": 3.6381505476770277, + "tokens_seen": 1122566144 + }, + { + "epoch": 0.34, + "learning_rate": 0.0005332316950852559, + "loss": 1.4162, + "theoretical_loss": 3.638130919717434, + "tokens_seen": 1122631680 + }, + { + "epoch": 0.34, + "learning_rate": 0.0005332156469408225, + "loss": 1.347, + "theoretical_loss": 3.6381112932244397, + "tokens_seen": 1122697216 + }, + { + "epoch": 0.34, + "learning_rate": 0.0005331995987963893, + "loss": 1.449, + "theoretical_loss": 3.6380916681978492, + "tokens_seen": 1122762752 + }, + { + "epoch": 0.34, + "learning_rate": 0.0005331835506519559, + "loss": 1.4723, + "theoretical_loss": 3.6380720446374686, + "tokens_seen": 1122828288 + }, + { + "epoch": 0.34, + "learning_rate": 0.0005331675025075227, + "loss": 1.462, + "theoretical_loss": 3.638052422543102, + "tokens_seen": 1122893824 + }, + { + "epoch": 0.34, + "learning_rate": 0.0005331514543630893, + "loss": 1.5186, + "theoretical_loss": 3.6380328019145542, + "tokens_seen": 1122959360 + }, + { + "epoch": 0.34, + "learning_rate": 0.0005331354062186561, + "loss": 1.4779, + "theoretical_loss": 3.63801318275163, + "tokens_seen": 1123024896 + }, + { + "epoch": 0.34, + "learning_rate": 0.0005331193580742227, + "loss": 1.3368, + "theoretical_loss": 3.637993565054135, + "tokens_seen": 1123090432 + }, + { + "epoch": 0.34, + "learning_rate": 0.0005331033099297895, + "loss": 1.3329, + "theoretical_loss": 3.637973948821874, + "tokens_seen": 1123155968 + }, + { + "epoch": 0.34, + "learning_rate": 0.0005330872617853561, + "loss": 1.449, + "theoretical_loss": 3.637954334054652, + "tokens_seen": 1123221504 + }, + { + "epoch": 0.34, + "learning_rate": 0.0005330712136409229, + "loss": 1.4063, + "theoretical_loss": 3.6379347207522748, + "tokens_seen": 1123287040 + }, + { + "epoch": 0.34, + "learning_rate": 0.0005330551654964895, + "loss": 1.4134, + "theoretical_loss": 3.6379151089145463, + "tokens_seen": 1123352576 + }, + { + "epoch": 0.34, + "learning_rate": 0.0005330391173520563, + "loss": 1.426, + "theoretical_loss": 3.6378954985412726, + "tokens_seen": 1123418112 + }, + { + "epoch": 0.34, + "learning_rate": 0.0005330230692076229, + "loss": 1.4064, + "theoretical_loss": 3.6378758896322587, + "tokens_seen": 1123483648 + }, + { + "epoch": 0.34, + "learning_rate": 0.0005330070210631897, + "loss": 1.4326, + "theoretical_loss": 3.6378562821873097, + "tokens_seen": 1123549184 + }, + { + "epoch": 0.34, + "learning_rate": 0.0005329909729187563, + "loss": 1.4756, + "theoretical_loss": 3.6378366762062306, + "tokens_seen": 1123614720 + }, + { + "epoch": 0.34, + "learning_rate": 0.0005329749247743231, + "loss": 1.4019, + "theoretical_loss": 3.637817071688828, + "tokens_seen": 1123680256 + }, + { + "epoch": 0.34, + "learning_rate": 0.0005329588766298897, + "loss": 1.5194, + "theoretical_loss": 3.637797468634906, + "tokens_seen": 1123745792 + }, + { + "epoch": 0.34, + "learning_rate": 0.0005329428284854565, + "loss": 1.4057, + "theoretical_loss": 3.63777786704427, + "tokens_seen": 1123811328 + }, + { + "epoch": 0.34, + "learning_rate": 0.0005329267803410231, + "loss": 1.5329, + "theoretical_loss": 3.6377582669167268, + "tokens_seen": 1123876864 + }, + { + "epoch": 0.34, + "objective/train/docs_used": 411904, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.207704782485962, + "objective/train/theoretical_loss": 3.63773866825208, + "objective/train/tokens_used": 1144402400, + "theoretical_loss": 3.63773866825208, + "tokens_seen": 1123942400 + }, + { + "epoch": 0.34, + "learning_rate": 0.0005329107321965899, + "loss": 1.3017, + "theoretical_loss": 3.63773866825208, + "tokens_seen": 1123942400 + }, + { + "epoch": 0.34, + "learning_rate": 0.0005328946840521565, + "loss": 1.4741, + "theoretical_loss": 3.637719071050137, + "tokens_seen": 1124007936 + }, + { + "epoch": 0.34, + "learning_rate": 0.0005328786359077233, + "loss": 1.4297, + "theoretical_loss": 3.6376994753107015, + "tokens_seen": 1124073472 + }, + { + "epoch": 0.34, + "learning_rate": 0.0005328625877632899, + "loss": 1.3755, + "theoretical_loss": 3.6376798810335806, + "tokens_seen": 1124139008 + }, + { + "epoch": 0.34, + "learning_rate": 0.0005328465396188567, + "loss": 1.4442, + "theoretical_loss": 3.6376602882185796, + "tokens_seen": 1124204544 + }, + { + "epoch": 0.34, + "learning_rate": 0.0005328304914744233, + "loss": 1.4625, + "theoretical_loss": 3.6376406968655037, + "tokens_seen": 1124270080 + }, + { + "epoch": 0.34, + "learning_rate": 0.0005328144433299901, + "loss": 1.3843, + "theoretical_loss": 3.6376211069741586, + "tokens_seen": 1124335616 + }, + { + "epoch": 0.34, + "learning_rate": 0.0005327983951855567, + "loss": 1.4502, + "theoretical_loss": 3.6376015185443507, + "tokens_seen": 1124401152 + }, + { + "epoch": 0.34, + "learning_rate": 0.0005327823470411235, + "loss": 1.5376, + "theoretical_loss": 3.6375819315758853, + "tokens_seen": 1124466688 + }, + { + "epoch": 0.34, + "learning_rate": 0.0005327662988966901, + "loss": 1.4665, + "theoretical_loss": 3.6375623460685684, + "tokens_seen": 1124532224 + }, + { + "epoch": 0.34, + "learning_rate": 0.0005327502507522569, + "loss": 1.4116, + "theoretical_loss": 3.6375427620222056, + "tokens_seen": 1124597760 + }, + { + "epoch": 0.34, + "learning_rate": 0.0005327342026078235, + "loss": 1.3738, + "theoretical_loss": 3.637523179436603, + "tokens_seen": 1124663296 + }, + { + "epoch": 0.34, + "learning_rate": 0.0005327181544633903, + "loss": 1.3802, + "theoretical_loss": 3.6375035983115667, + "tokens_seen": 1124728832 + }, + { + "epoch": 0.34, + "learning_rate": 0.0005327021063189569, + "loss": 1.425, + "theoretical_loss": 3.6374840186469024, + "tokens_seen": 1124794368 + }, + { + "epoch": 0.34, + "learning_rate": 0.0005326860581745236, + "loss": 1.3632, + "theoretical_loss": 3.637464440442416, + "tokens_seen": 1124859904 + }, + { + "epoch": 0.34, + "learning_rate": 0.0005326700100300903, + "loss": 1.413, + "theoretical_loss": 3.6374448636979135, + "tokens_seen": 1124925440 + }, + { + "epoch": 0.34, + "learning_rate": 0.000532653961885657, + "loss": 1.4366, + "theoretical_loss": 3.6374252884132017, + "tokens_seen": 1124990976 + }, + { + "epoch": 0.34, + "learning_rate": 0.0005326379137412237, + "loss": 1.4927, + "theoretical_loss": 3.637405714588086, + "tokens_seen": 1125056512 + }, + { + "epoch": 0.34, + "learning_rate": 0.0005326218655967904, + "loss": 1.3853, + "theoretical_loss": 3.6373861422223728, + "tokens_seen": 1125122048 + }, + { + "epoch": 0.34, + "learning_rate": 0.0005326058174523571, + "loss": 1.5699, + "theoretical_loss": 3.637366571315868, + "tokens_seen": 1125187584 + }, + { + "epoch": 0.34, + "learning_rate": 0.0005325897693079238, + "loss": 1.415, + "theoretical_loss": 3.6373470018683784, + "tokens_seen": 1125253120 + }, + { + "epoch": 0.34, + "learning_rate": 0.0005325737211634905, + "loss": 1.5097, + "theoretical_loss": 3.63732743387971, + "tokens_seen": 1125318656 + }, + { + "epoch": 0.34, + "learning_rate": 0.0005325576730190572, + "loss": 1.3569, + "theoretical_loss": 3.637307867349669, + "tokens_seen": 1125384192 + }, + { + "epoch": 0.34, + "learning_rate": 0.0005325416248746238, + "loss": 1.4237, + "theoretical_loss": 3.637288302278061, + "tokens_seen": 1125449728 + }, + { + "epoch": 0.34, + "learning_rate": 0.0005325255767301906, + "loss": 1.4899, + "theoretical_loss": 3.6372687386646945, + "tokens_seen": 1125515264 + }, + { + "epoch": 0.34, + "learning_rate": 0.0005325095285857572, + "loss": 1.3638, + "theoretical_loss": 3.637249176509374, + "tokens_seen": 1125580800 + }, + { + "epoch": 0.34, + "learning_rate": 0.000532493480441324, + "loss": 1.4339, + "theoretical_loss": 3.6372296158119064, + "tokens_seen": 1125646336 + }, + { + "epoch": 0.34, + "learning_rate": 0.0005324774322968906, + "loss": 1.4444, + "theoretical_loss": 3.637210056572098, + "tokens_seen": 1125711872 + }, + { + "epoch": 0.34, + "learning_rate": 0.0005324613841524574, + "loss": 1.4773, + "theoretical_loss": 3.6371904987897565, + "tokens_seen": 1125777408 + }, + { + "epoch": 0.34, + "learning_rate": 0.000532445336008024, + "loss": 1.3494, + "theoretical_loss": 3.637170942464687, + "tokens_seen": 1125842944 + }, + { + "epoch": 0.34, + "learning_rate": 0.0005324292878635908, + "loss": 1.4257, + "theoretical_loss": 3.6371513875966968, + "tokens_seen": 1125908480 + }, + { + "epoch": 0.34, + "learning_rate": 0.0005324132397191574, + "loss": 1.4607, + "theoretical_loss": 3.6371318341855927, + "tokens_seen": 1125974016 + }, + { + "epoch": 0.34, + "learning_rate": 0.0005323971915747242, + "loss": 1.4596, + "theoretical_loss": 3.6371122822311808, + "tokens_seen": 1126039552 + }, + { + "epoch": 0.34, + "learning_rate": 0.0005323811434302908, + "loss": 1.5508, + "theoretical_loss": 3.6370927317332677, + "tokens_seen": 1126105088 + }, + { + "epoch": 0.34, + "learning_rate": 0.0005323650952858576, + "loss": 1.3471, + "theoretical_loss": 3.6370731826916614, + "tokens_seen": 1126170624 + }, + { + "epoch": 0.34, + "learning_rate": 0.0005323490471414242, + "loss": 1.4248, + "theoretical_loss": 3.6370536351061675, + "tokens_seen": 1126236160 + }, + { + "epoch": 0.34, + "learning_rate": 0.000532332998996991, + "loss": 1.3784, + "theoretical_loss": 3.637034088976593, + "tokens_seen": 1126301696 + }, + { + "epoch": 0.34, + "learning_rate": 0.0005323169508525576, + "loss": 1.3839, + "theoretical_loss": 3.6370145443027453, + "tokens_seen": 1126367232 + }, + { + "epoch": 0.34, + "learning_rate": 0.0005323009027081244, + "loss": 1.5358, + "theoretical_loss": 3.6369950010844305, + "tokens_seen": 1126432768 + }, + { + "epoch": 0.34, + "learning_rate": 0.000532284854563691, + "loss": 1.4943, + "theoretical_loss": 3.636975459321456, + "tokens_seen": 1126498304 + }, + { + "epoch": 0.34, + "learning_rate": 0.0005322688064192578, + "loss": 1.544, + "theoretical_loss": 3.636955919013629, + "tokens_seen": 1126563840 + }, + { + "epoch": 0.34, + "learning_rate": 0.0005322527582748244, + "loss": 1.4487, + "theoretical_loss": 3.636936380160756, + "tokens_seen": 1126629376 + }, + { + "epoch": 0.34, + "learning_rate": 0.0005322367101303912, + "loss": 1.5217, + "theoretical_loss": 3.6369168427626444, + "tokens_seen": 1126694912 + }, + { + "epoch": 0.34, + "learning_rate": 0.0005322206619859578, + "loss": 1.4468, + "theoretical_loss": 3.636897306819101, + "tokens_seen": 1126760448 + }, + { + "epoch": 0.34, + "learning_rate": 0.0005322046138415246, + "loss": 1.3025, + "theoretical_loss": 3.6368777723299326, + "tokens_seen": 1126825984 + }, + { + "epoch": 0.34, + "learning_rate": 0.0005321885656970912, + "loss": 1.5622, + "theoretical_loss": 3.6368582392949476, + "tokens_seen": 1126891520 + }, + { + "epoch": 0.34, + "learning_rate": 0.000532172517552658, + "loss": 1.4893, + "theoretical_loss": 3.6368387077139523, + "tokens_seen": 1126957056 + }, + { + "epoch": 0.34, + "learning_rate": 0.0005321564694082246, + "loss": 1.449, + "theoretical_loss": 3.6368191775867533, + "tokens_seen": 1127022592 + }, + { + "epoch": 0.34, + "learning_rate": 0.0005321404212637914, + "loss": 1.4365, + "theoretical_loss": 3.636799648913159, + "tokens_seen": 1127088128 + }, + { + "epoch": 0.34, + "learning_rate": 0.0005321243731193581, + "loss": 1.4406, + "theoretical_loss": 3.6367801216929765, + "tokens_seen": 1127153664 + }, + { + "epoch": 0.34, + "objective/train/docs_used": 413086, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.4999926090240479, + "objective/train/theoretical_loss": 3.6367605959260123, + "objective/train/tokens_used": 1147679200, + "theoretical_loss": 3.6367605959260123, + "tokens_seen": 1127219200 + }, + { + "epoch": 0.34, + "learning_rate": 0.0005321083249749248, + "loss": 1.3675, + "theoretical_loss": 3.6367605959260123, + "tokens_seen": 1127219200 + }, + { + "epoch": 0.34, + "learning_rate": 0.0005320922768304915, + "loss": 1.522, + "theoretical_loss": 3.636741071612075, + "tokens_seen": 1127284736 + }, + { + "epoch": 0.34, + "learning_rate": 0.0005320762286860582, + "loss": 1.4009, + "theoretical_loss": 3.636721548750971, + "tokens_seen": 1127350272 + }, + { + "epoch": 0.34, + "learning_rate": 0.0005320601805416249, + "loss": 1.4933, + "theoretical_loss": 3.6367020273425084, + "tokens_seen": 1127415808 + }, + { + "epoch": 0.34, + "learning_rate": 0.0005320441323971916, + "loss": 1.4727, + "theoretical_loss": 3.636682507386494, + "tokens_seen": 1127481344 + }, + { + "epoch": 0.34, + "learning_rate": 0.0005320280842527583, + "loss": 1.4134, + "theoretical_loss": 3.636662988882736, + "tokens_seen": 1127546880 + }, + { + "epoch": 0.34, + "learning_rate": 0.000532012036108325, + "loss": 1.4246, + "theoretical_loss": 3.6366434718310416, + "tokens_seen": 1127612416 + }, + { + "epoch": 0.34, + "learning_rate": 0.0005319959879638917, + "loss": 1.5262, + "theoretical_loss": 3.636623956231219, + "tokens_seen": 1127677952 + }, + { + "epoch": 0.34, + "learning_rate": 0.0005319799398194584, + "loss": 1.4099, + "theoretical_loss": 3.6366044420830743, + "tokens_seen": 1127743488 + }, + { + "epoch": 0.34, + "learning_rate": 0.0005319638916750251, + "loss": 1.5057, + "theoretical_loss": 3.6365849293864168, + "tokens_seen": 1127809024 + }, + { + "epoch": 0.34, + "learning_rate": 0.0005319478435305918, + "loss": 1.4324, + "theoretical_loss": 3.6365654181410534, + "tokens_seen": 1127874560 + }, + { + "epoch": 0.34, + "learning_rate": 0.0005319317953861585, + "loss": 1.4935, + "theoretical_loss": 3.6365459083467915, + "tokens_seen": 1127940096 + }, + { + "epoch": 0.34, + "learning_rate": 0.0005319157472417252, + "loss": 1.4852, + "theoretical_loss": 3.63652640000344, + "tokens_seen": 1128005632 + }, + { + "epoch": 0.34, + "learning_rate": 0.000531899699097292, + "loss": 1.4254, + "theoretical_loss": 3.636506893110806, + "tokens_seen": 1128071168 + }, + { + "epoch": 0.34, + "learning_rate": 0.0005318836509528586, + "loss": 1.4833, + "theoretical_loss": 3.6364873876686974, + "tokens_seen": 1128136704 + }, + { + "epoch": 0.34, + "learning_rate": 0.0005318676028084254, + "loss": 1.4681, + "theoretical_loss": 3.636467883676922, + "tokens_seen": 1128202240 + }, + { + "epoch": 0.34, + "learning_rate": 0.000531851554663992, + "loss": 1.4142, + "theoretical_loss": 3.636448381135288, + "tokens_seen": 1128267776 + }, + { + "epoch": 0.34, + "learning_rate": 0.0005318355065195588, + "loss": 1.4924, + "theoretical_loss": 3.636428880043603, + "tokens_seen": 1128333312 + }, + { + "epoch": 0.34, + "learning_rate": 0.0005318194583751254, + "loss": 1.4472, + "theoretical_loss": 3.636409380401675, + "tokens_seen": 1128398848 + }, + { + "epoch": 0.34, + "learning_rate": 0.0005318034102306922, + "loss": 1.5282, + "theoretical_loss": 3.636389882209313, + "tokens_seen": 1128464384 + }, + { + "epoch": 0.34, + "learning_rate": 0.0005317873620862588, + "loss": 1.4572, + "theoretical_loss": 3.6363703854663236, + "tokens_seen": 1128529920 + }, + { + "epoch": 0.34, + "learning_rate": 0.0005317713139418256, + "loss": 1.5104, + "theoretical_loss": 3.6363508901725154, + "tokens_seen": 1128595456 + }, + { + "epoch": 0.34, + "learning_rate": 0.0005317552657973922, + "loss": 1.2891, + "theoretical_loss": 3.6363313963276975, + "tokens_seen": 1128660992 + }, + { + "epoch": 0.34, + "learning_rate": 0.000531739217652959, + "loss": 1.4722, + "theoretical_loss": 3.636311903931677, + "tokens_seen": 1128726528 + }, + { + "epoch": 0.34, + "learning_rate": 0.0005317231695085256, + "loss": 1.4166, + "theoretical_loss": 3.6362924129842624, + "tokens_seen": 1128792064 + }, + { + "epoch": 0.34, + "learning_rate": 0.0005317071213640924, + "loss": 1.4648, + "theoretical_loss": 3.6362729234852615, + "tokens_seen": 1128857600 + }, + { + "epoch": 0.34, + "learning_rate": 0.000531691073219659, + "loss": 1.5458, + "theoretical_loss": 3.636253435434484, + "tokens_seen": 1128923136 + }, + { + "epoch": 0.34, + "learning_rate": 0.0005316750250752258, + "loss": 1.4139, + "theoretical_loss": 3.6362339488317366, + "tokens_seen": 1128988672 + }, + { + "epoch": 0.34, + "learning_rate": 0.0005316589769307924, + "loss": 1.3357, + "theoretical_loss": 3.6362144636768283, + "tokens_seen": 1129054208 + }, + { + "epoch": 0.34, + "learning_rate": 0.0005316429287863592, + "loss": 1.3593, + "theoretical_loss": 3.6361949799695683, + "tokens_seen": 1129119744 + }, + { + "epoch": 0.34, + "learning_rate": 0.0005316268806419258, + "loss": 1.4215, + "theoretical_loss": 3.6361754977097633, + "tokens_seen": 1129185280 + }, + { + "epoch": 0.34, + "learning_rate": 0.0005316108324974926, + "loss": 1.4049, + "theoretical_loss": 3.6361560168972233, + "tokens_seen": 1129250816 + }, + { + "epoch": 0.34, + "learning_rate": 0.0005315947843530592, + "loss": 1.663, + "theoretical_loss": 3.6361365375317565, + "tokens_seen": 1129316352 + }, + { + "epoch": 0.34, + "learning_rate": 0.000531578736208626, + "loss": 1.4925, + "theoretical_loss": 3.6361170596131704, + "tokens_seen": 1129381888 + }, + { + "epoch": 0.34, + "learning_rate": 0.0005315626880641926, + "loss": 1.3812, + "theoretical_loss": 3.6360975831412747, + "tokens_seen": 1129447424 + }, + { + "epoch": 0.34, + "learning_rate": 0.0005315466399197594, + "loss": 1.4328, + "theoretical_loss": 3.6360781081158775, + "tokens_seen": 1129512960 + }, + { + "epoch": 0.34, + "learning_rate": 0.000531530591775326, + "loss": 1.437, + "theoretical_loss": 3.6360586345367882, + "tokens_seen": 1129578496 + }, + { + "epoch": 0.34, + "learning_rate": 0.0005315145436308928, + "loss": 1.4308, + "theoretical_loss": 3.6360391624038146, + "tokens_seen": 1129644032 + }, + { + "epoch": 0.34, + "learning_rate": 0.0005314984954864594, + "loss": 1.4325, + "theoretical_loss": 3.6360196917167658, + "tokens_seen": 1129709568 + }, + { + "epoch": 0.34, + "learning_rate": 0.0005314824473420262, + "loss": 1.5111, + "theoretical_loss": 3.6360002224754506, + "tokens_seen": 1129775104 + }, + { + "epoch": 0.34, + "learning_rate": 0.0005314663991975928, + "loss": 1.3686, + "theoretical_loss": 3.6359807546796774, + "tokens_seen": 1129840640 + }, + { + "epoch": 0.34, + "learning_rate": 0.0005314503510531596, + "loss": 1.4643, + "theoretical_loss": 3.6359612883292556, + "tokens_seen": 1129906176 + }, + { + "epoch": 0.34, + "learning_rate": 0.0005314343029087262, + "loss": 1.5359, + "theoretical_loss": 3.6359418234239937, + "tokens_seen": 1129971712 + }, + { + "epoch": 0.34, + "learning_rate": 0.000531418254764293, + "loss": 1.508, + "theoretical_loss": 3.6359223599637005, + "tokens_seen": 1130037248 + }, + { + "epoch": 0.34, + "learning_rate": 0.0005314022066198596, + "loss": 1.5199, + "theoretical_loss": 3.6359028979481853, + "tokens_seen": 1130102784 + }, + { + "epoch": 0.34, + "learning_rate": 0.0005313861584754264, + "loss": 1.3669, + "theoretical_loss": 3.635883437377257, + "tokens_seen": 1130168320 + }, + { + "epoch": 0.34, + "learning_rate": 0.000531370110330993, + "loss": 1.3338, + "theoretical_loss": 3.6358639782507245, + "tokens_seen": 1130233856 + }, + { + "epoch": 0.34, + "learning_rate": 0.0005313540621865598, + "loss": 1.5684, + "theoretical_loss": 3.635844520568397, + "tokens_seen": 1130299392 + }, + { + "epoch": 0.34, + "learning_rate": 0.0005313380140421264, + "loss": 1.5479, + "theoretical_loss": 3.6358250643300836, + "tokens_seen": 1130364928 + }, + { + "epoch": 0.34, + "learning_rate": 0.0005313219658976932, + "loss": 1.4162, + "theoretical_loss": 3.635805609535593, + "tokens_seen": 1130430464 + }, + { + "epoch": 0.34, + "objective/train/docs_used": 414307, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.5912754535675049, + "objective/train/theoretical_loss": 3.635786156184735, + "objective/train/tokens_used": 1150956000, + "theoretical_loss": 3.635786156184735, + "tokens_seen": 1130496000 + }, + { + "epoch": 0.34, + "learning_rate": 0.0005313059177532598, + "loss": 1.4815, + "theoretical_loss": 3.635786156184735, + "tokens_seen": 1130496000 + }, + { + "epoch": 0.34, + "learning_rate": 0.0005312898696088265, + "loss": 1.3307, + "theoretical_loss": 3.635766704277318, + "tokens_seen": 1130561536 + }, + { + "epoch": 0.34, + "learning_rate": 0.0005312738214643932, + "loss": 1.3424, + "theoretical_loss": 3.635747253813152, + "tokens_seen": 1130627072 + }, + { + "epoch": 0.34, + "learning_rate": 0.0005312577733199599, + "loss": 1.369, + "theoretical_loss": 3.635727804792046, + "tokens_seen": 1130692608 + }, + { + "epoch": 0.34, + "learning_rate": 0.0005312417251755266, + "loss": 1.3665, + "theoretical_loss": 3.635708357213809, + "tokens_seen": 1130758144 + }, + { + "epoch": 0.34, + "learning_rate": 0.0005312256770310933, + "loss": 1.3707, + "theoretical_loss": 3.635688911078251, + "tokens_seen": 1130823680 + }, + { + "epoch": 0.34, + "learning_rate": 0.00053120962888666, + "loss": 1.5061, + "theoretical_loss": 3.635669466385181, + "tokens_seen": 1130889216 + }, + { + "epoch": 0.34, + "learning_rate": 0.0005311935807422267, + "loss": 1.4551, + "theoretical_loss": 3.635650023134408, + "tokens_seen": 1130954752 + }, + { + "epoch": 0.34, + "learning_rate": 0.0005311775325977934, + "loss": 1.5157, + "theoretical_loss": 3.635630581325742, + "tokens_seen": 1131020288 + }, + { + "epoch": 0.34, + "learning_rate": 0.0005311614844533601, + "loss": 1.4774, + "theoretical_loss": 3.635611140958993, + "tokens_seen": 1131085824 + }, + { + "epoch": 0.34, + "learning_rate": 0.0005311454363089267, + "loss": 1.3273, + "theoretical_loss": 3.6355917020339685, + "tokens_seen": 1131151360 + }, + { + "epoch": 0.34, + "learning_rate": 0.0005311293881644935, + "loss": 1.3667, + "theoretical_loss": 3.635572264550481, + "tokens_seen": 1131216896 + }, + { + "epoch": 0.34, + "learning_rate": 0.0005311133400200601, + "loss": 1.4719, + "theoretical_loss": 3.635552828508337, + "tokens_seen": 1131282432 + }, + { + "epoch": 0.34, + "learning_rate": 0.0005310972918756269, + "loss": 1.4143, + "theoretical_loss": 3.635533393907349, + "tokens_seen": 1131347968 + }, + { + "epoch": 0.34, + "learning_rate": 0.0005310812437311935, + "loss": 1.4414, + "theoretical_loss": 3.635513960747325, + "tokens_seen": 1131413504 + }, + { + "epoch": 0.34, + "learning_rate": 0.0005310651955867603, + "loss": 1.4826, + "theoretical_loss": 3.6354945290280742, + "tokens_seen": 1131479040 + }, + { + "epoch": 0.34, + "learning_rate": 0.0005310491474423269, + "loss": 1.4575, + "theoretical_loss": 3.6354750987494078, + "tokens_seen": 1131544576 + }, + { + "epoch": 0.34, + "learning_rate": 0.0005310330992978937, + "loss": 1.4463, + "theoretical_loss": 3.635455669911135, + "tokens_seen": 1131610112 + }, + { + "epoch": 0.34, + "learning_rate": 0.0005310170511534603, + "loss": 1.3524, + "theoretical_loss": 3.635436242513065, + "tokens_seen": 1131675648 + }, + { + "epoch": 0.34, + "learning_rate": 0.0005310010030090271, + "loss": 1.5176, + "theoretical_loss": 3.6354168165550087, + "tokens_seen": 1131741184 + }, + { + "epoch": 0.34, + "learning_rate": 0.0005309849548645937, + "loss": 1.464, + "theoretical_loss": 3.6353973920367757, + "tokens_seen": 1131806720 + }, + { + "epoch": 0.34, + "learning_rate": 0.0005309689067201605, + "loss": 1.4514, + "theoretical_loss": 3.635377968958175, + "tokens_seen": 1131872256 + }, + { + "epoch": 0.34, + "learning_rate": 0.0005309528585757271, + "loss": 1.4837, + "theoretical_loss": 3.6353585473190178, + "tokens_seen": 1131937792 + }, + { + "epoch": 0.34, + "learning_rate": 0.0005309368104312939, + "loss": 1.4131, + "theoretical_loss": 3.6353391271191136, + "tokens_seen": 1132003328 + }, + { + "epoch": 0.34, + "learning_rate": 0.0005309207622868606, + "loss": 1.4807, + "theoretical_loss": 3.6353197083582725, + "tokens_seen": 1132068864 + }, + { + "epoch": 0.34, + "learning_rate": 0.0005309047141424273, + "loss": 1.4138, + "theoretical_loss": 3.6353002910363035, + "tokens_seen": 1132134400 + }, + { + "epoch": 0.34, + "learning_rate": 0.000530888665997994, + "loss": 1.4503, + "theoretical_loss": 3.6352808751530183, + "tokens_seen": 1132199936 + }, + { + "epoch": 0.34, + "learning_rate": 0.0005308726178535607, + "loss": 1.452, + "theoretical_loss": 3.6352614607082265, + "tokens_seen": 1132265472 + }, + { + "epoch": 0.34, + "learning_rate": 0.0005308565697091274, + "loss": 1.482, + "theoretical_loss": 3.6352420477017384, + "tokens_seen": 1132331008 + }, + { + "epoch": 0.34, + "learning_rate": 0.0005308405215646941, + "loss": 1.4184, + "theoretical_loss": 3.6352226361333635, + "tokens_seen": 1132396544 + }, + { + "epoch": 0.34, + "learning_rate": 0.0005308244734202608, + "loss": 1.4156, + "theoretical_loss": 3.635203226002912, + "tokens_seen": 1132462080 + }, + { + "epoch": 0.34, + "learning_rate": 0.0005308084252758275, + "loss": 1.4075, + "theoretical_loss": 3.6351838173101956, + "tokens_seen": 1132527616 + }, + { + "epoch": 0.34, + "learning_rate": 0.0005307923771313942, + "loss": 1.5418, + "theoretical_loss": 3.6351644100550233, + "tokens_seen": 1132593152 + }, + { + "epoch": 0.34, + "learning_rate": 0.0005307763289869609, + "loss": 1.3985, + "theoretical_loss": 3.635145004237206, + "tokens_seen": 1132658688 + }, + { + "epoch": 0.34, + "learning_rate": 0.0005307602808425276, + "loss": 1.4044, + "theoretical_loss": 3.635125599856554, + "tokens_seen": 1132724224 + }, + { + "epoch": 0.34, + "learning_rate": 0.0005307442326980943, + "loss": 1.4207, + "theoretical_loss": 3.6351061969128775, + "tokens_seen": 1132789760 + }, + { + "epoch": 0.34, + "learning_rate": 0.000530728184553661, + "loss": 1.5205, + "theoretical_loss": 3.6350867954059876, + "tokens_seen": 1132855296 + }, + { + "epoch": 0.34, + "learning_rate": 0.0005307121364092277, + "loss": 1.5612, + "theoretical_loss": 3.635067395335694, + "tokens_seen": 1132920832 + }, + { + "epoch": 0.34, + "learning_rate": 0.0005306960882647944, + "loss": 1.3475, + "theoretical_loss": 3.6350479967018074, + "tokens_seen": 1132986368 + }, + { + "epoch": 0.34, + "learning_rate": 0.0005306800401203611, + "loss": 1.5443, + "theoretical_loss": 3.6350285995041385, + "tokens_seen": 1133051904 + }, + { + "epoch": 0.34, + "learning_rate": 0.0005306639919759278, + "loss": 1.4504, + "theoretical_loss": 3.635009203742498, + "tokens_seen": 1133117440 + }, + { + "epoch": 0.34, + "learning_rate": 0.0005306479438314945, + "loss": 1.4203, + "theoretical_loss": 3.6349898094166964, + "tokens_seen": 1133182976 + }, + { + "epoch": 0.34, + "learning_rate": 0.0005306318956870612, + "loss": 1.4011, + "theoretical_loss": 3.634970416526545, + "tokens_seen": 1133248512 + }, + { + "epoch": 0.34, + "learning_rate": 0.0005306158475426279, + "loss": 1.5742, + "theoretical_loss": 3.6349510250718535, + "tokens_seen": 1133314048 + }, + { + "epoch": 0.34, + "learning_rate": 0.0005305997993981946, + "loss": 1.5159, + "theoretical_loss": 3.634931635052433, + "tokens_seen": 1133379584 + }, + { + "epoch": 0.34, + "learning_rate": 0.0005305837512537613, + "loss": 1.4966, + "theoretical_loss": 3.6349122464680947, + "tokens_seen": 1133445120 + }, + { + "epoch": 0.34, + "learning_rate": 0.000530567703109328, + "loss": 1.4055, + "theoretical_loss": 3.6348928593186494, + "tokens_seen": 1133510656 + }, + { + "epoch": 0.34, + "learning_rate": 0.0005305516549648947, + "loss": 1.4079, + "theoretical_loss": 3.634873473603907, + "tokens_seen": 1133576192 + }, + { + "epoch": 0.34, + "learning_rate": 0.0005305356068204614, + "loss": 1.3472, + "theoretical_loss": 3.63485408932368, + "tokens_seen": 1133641728 + }, + { + "epoch": 0.34, + "learning_rate": 0.0005305195586760281, + "loss": 1.5083, + "theoretical_loss": 3.6348347064777773, + "tokens_seen": 1133707264 + }, + { + "epoch": 0.34, + "objective/train/docs_used": 415442, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.4380152225494385, + "objective/train/theoretical_loss": 3.6348153250660116, + "objective/train/tokens_used": 1154232800, + "theoretical_loss": 3.6348153250660116, + "tokens_seen": 1133772800 + }, + { + "epoch": 0.34, + "learning_rate": 0.0005305035105315948, + "loss": 1.4546, + "theoretical_loss": 3.6348153250660116, + "tokens_seen": 1133772800 + }, + { + "epoch": 0.34, + "learning_rate": 0.0005304874623871615, + "loss": 1.4908, + "theoretical_loss": 3.634795945088193, + "tokens_seen": 1133838336 + }, + { + "epoch": 0.34, + "learning_rate": 0.0005304714142427283, + "loss": 1.3888, + "theoretical_loss": 3.634776566544133, + "tokens_seen": 1133903872 + }, + { + "epoch": 0.34, + "learning_rate": 0.0005304553660982949, + "loss": 1.5054, + "theoretical_loss": 3.6347571894336426, + "tokens_seen": 1133969408 + }, + { + "epoch": 0.34, + "learning_rate": 0.0005304393179538617, + "loss": 1.4178, + "theoretical_loss": 3.6347378137565323, + "tokens_seen": 1134034944 + }, + { + "epoch": 0.34, + "learning_rate": 0.0005304232698094283, + "loss": 1.4544, + "theoretical_loss": 3.634718439512614, + "tokens_seen": 1134100480 + }, + { + "epoch": 0.34, + "learning_rate": 0.000530407221664995, + "loss": 1.4605, + "theoretical_loss": 3.6346990667016987, + "tokens_seen": 1134166016 + }, + { + "epoch": 0.34, + "learning_rate": 0.0005303911735205617, + "loss": 1.3989, + "theoretical_loss": 3.634679695323597, + "tokens_seen": 1134231552 + }, + { + "epoch": 0.34, + "learning_rate": 0.0005303751253761285, + "loss": 1.4334, + "theoretical_loss": 3.634660325378121, + "tokens_seen": 1134297088 + }, + { + "epoch": 0.34, + "learning_rate": 0.0005303590772316951, + "loss": 1.5004, + "theoretical_loss": 3.634640956865082, + "tokens_seen": 1134362624 + }, + { + "epoch": 0.34, + "learning_rate": 0.0005303430290872619, + "loss": 1.4023, + "theoretical_loss": 3.634621589784291, + "tokens_seen": 1134428160 + }, + { + "epoch": 0.34, + "learning_rate": 0.0005303269809428285, + "loss": 1.4595, + "theoretical_loss": 3.6346022241355587, + "tokens_seen": 1134493696 + }, + { + "epoch": 0.34, + "learning_rate": 0.0005303109327983953, + "loss": 1.4304, + "theoretical_loss": 3.634582859918697, + "tokens_seen": 1134559232 + }, + { + "epoch": 0.34, + "learning_rate": 0.0005302948846539619, + "loss": 1.519, + "theoretical_loss": 3.6345634971335183, + "tokens_seen": 1134624768 + }, + { + "epoch": 0.34, + "learning_rate": 0.0005302788365095287, + "loss": 1.4414, + "theoretical_loss": 3.6345441357798327, + "tokens_seen": 1134690304 + }, + { + "epoch": 0.34, + "learning_rate": 0.0005302627883650953, + "loss": 1.4321, + "theoretical_loss": 3.634524775857452, + "tokens_seen": 1134755840 + }, + { + "epoch": 0.34, + "learning_rate": 0.0005302467402206621, + "loss": 1.4757, + "theoretical_loss": 3.6345054173661886, + "tokens_seen": 1134821376 + }, + { + "epoch": 0.34, + "learning_rate": 0.0005302306920762287, + "loss": 1.5378, + "theoretical_loss": 3.6344860603058526, + "tokens_seen": 1134886912 + }, + { + "epoch": 0.34, + "learning_rate": 0.0005302146439317955, + "loss": 1.4277, + "theoretical_loss": 3.6344667046762567, + "tokens_seen": 1134952448 + }, + { + "epoch": 0.34, + "learning_rate": 0.0005301985957873621, + "loss": 1.5249, + "theoretical_loss": 3.634447350477213, + "tokens_seen": 1135017984 + }, + { + "epoch": 0.34, + "learning_rate": 0.0005301825476429289, + "loss": 1.3984, + "theoretical_loss": 3.6344279977085314, + "tokens_seen": 1135083520 + }, + { + "epoch": 0.34, + "learning_rate": 0.0005301664994984955, + "loss": 1.5346, + "theoretical_loss": 3.6344086463700256, + "tokens_seen": 1135149056 + }, + { + "epoch": 0.34, + "learning_rate": 0.0005301504513540623, + "loss": 1.4662, + "theoretical_loss": 3.6343892964615057, + "tokens_seen": 1135214592 + }, + { + "epoch": 0.34, + "learning_rate": 0.0005301344032096289, + "loss": 1.4647, + "theoretical_loss": 3.6343699479827842, + "tokens_seen": 1135280128 + }, + { + "epoch": 0.34, + "learning_rate": 0.0005301183550651957, + "loss": 1.3829, + "theoretical_loss": 3.6343506009336735, + "tokens_seen": 1135345664 + }, + { + "epoch": 0.34, + "learning_rate": 0.0005301023069207623, + "loss": 1.5531, + "theoretical_loss": 3.6343312553139846, + "tokens_seen": 1135411200 + }, + { + "epoch": 0.34, + "learning_rate": 0.0005300862587763291, + "loss": 1.4037, + "theoretical_loss": 3.63431191112353, + "tokens_seen": 1135476736 + }, + { + "epoch": 0.34, + "learning_rate": 0.0005300702106318957, + "loss": 1.4512, + "theoretical_loss": 3.6342925683621203, + "tokens_seen": 1135542272 + }, + { + "epoch": 0.34, + "learning_rate": 0.0005300541624874625, + "loss": 1.5161, + "theoretical_loss": 3.6342732270295692, + "tokens_seen": 1135607808 + }, + { + "epoch": 0.34, + "learning_rate": 0.0005300381143430291, + "loss": 1.4981, + "theoretical_loss": 3.6342538871256878, + "tokens_seen": 1135673344 + }, + { + "epoch": 0.34, + "learning_rate": 0.0005300220661985959, + "loss": 1.361, + "theoretical_loss": 3.634234548650288, + "tokens_seen": 1135738880 + }, + { + "epoch": 0.34, + "learning_rate": 0.0005300060180541625, + "loss": 1.402, + "theoretical_loss": 3.6342152116031827, + "tokens_seen": 1135804416 + }, + { + "epoch": 0.34, + "learning_rate": 0.0005299899699097293, + "loss": 1.4042, + "theoretical_loss": 3.634195875984183, + "tokens_seen": 1135869952 + }, + { + "epoch": 0.34, + "learning_rate": 0.0005299739217652959, + "loss": 1.4193, + "theoretical_loss": 3.634176541793102, + "tokens_seen": 1135935488 + }, + { + "epoch": 0.34, + "learning_rate": 0.0005299578736208627, + "loss": 1.424, + "theoretical_loss": 3.6341572090297514, + "tokens_seen": 1136001024 + }, + { + "epoch": 0.34, + "learning_rate": 0.0005299418254764293, + "loss": 1.397, + "theoretical_loss": 3.6341378776939433, + "tokens_seen": 1136066560 + }, + { + "epoch": 0.34, + "learning_rate": 0.0005299257773319961, + "loss": 1.487, + "theoretical_loss": 3.63411854778549, + "tokens_seen": 1136132096 + }, + { + "epoch": 0.34, + "learning_rate": 0.0005299097291875627, + "loss": 1.5061, + "theoretical_loss": 3.634099219304203, + "tokens_seen": 1136197632 + }, + { + "epoch": 0.34, + "learning_rate": 0.0005298936810431294, + "loss": 1.4955, + "theoretical_loss": 3.6340798922498965, + "tokens_seen": 1136263168 + }, + { + "epoch": 0.34, + "learning_rate": 0.0005298776328986961, + "loss": 1.5967, + "theoretical_loss": 3.6340605666223817, + "tokens_seen": 1136328704 + }, + { + "epoch": 0.34, + "learning_rate": 0.0005298615847542628, + "loss": 1.3923, + "theoretical_loss": 3.6340412424214708, + "tokens_seen": 1136394240 + }, + { + "epoch": 0.34, + "learning_rate": 0.0005298455366098295, + "loss": 1.4179, + "theoretical_loss": 3.634021919646977, + "tokens_seen": 1136459776 + }, + { + "epoch": 0.34, + "learning_rate": 0.0005298294884653962, + "loss": 1.5615, + "theoretical_loss": 3.6340025982987116, + "tokens_seen": 1136525312 + }, + { + "epoch": 0.34, + "learning_rate": 0.0005298134403209629, + "loss": 1.4776, + "theoretical_loss": 3.633983278376488, + "tokens_seen": 1136590848 + }, + { + "epoch": 0.34, + "learning_rate": 0.0005297973921765296, + "loss": 1.5324, + "theoretical_loss": 3.6339639598801186, + "tokens_seen": 1136656384 + }, + { + "epoch": 0.34, + "learning_rate": 0.0005297813440320963, + "loss": 1.4037, + "theoretical_loss": 3.633944642809416, + "tokens_seen": 1136721920 + }, + { + "epoch": 0.34, + "learning_rate": 0.000529765295887663, + "loss": 1.479, + "theoretical_loss": 3.6339253271641923, + "tokens_seen": 1136787456 + }, + { + "epoch": 0.34, + "learning_rate": 0.0005297492477432296, + "loss": 1.403, + "theoretical_loss": 3.633906012944261, + "tokens_seen": 1136852992 + }, + { + "epoch": 0.34, + "learning_rate": 0.0005297331995987964, + "loss": 1.4577, + "theoretical_loss": 3.633886700149434, + "tokens_seen": 1136918528 + }, + { + "epoch": 0.34, + "learning_rate": 0.0005297171514543631, + "loss": 1.3559, + "theoretical_loss": 3.6338673887795245, + "tokens_seen": 1136984064 + }, + { + "epoch": 0.34, + "objective/train/docs_used": 416597, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.52068030834198, + "objective/train/theoretical_loss": 3.633848078834345, + "objective/train/tokens_used": 1157509600, + "theoretical_loss": 3.633848078834345, + "tokens_seen": 1137049600 + }, + { + "epoch": 0.34, + "learning_rate": 0.0005297011033099298, + "loss": 1.4697, + "theoretical_loss": 3.633848078834345, + "tokens_seen": 1137049600 + }, + { + "epoch": 0.34, + "learning_rate": 0.0005296850551654965, + "loss": 1.4298, + "theoretical_loss": 3.633828770313708, + "tokens_seen": 1137115136 + }, + { + "epoch": 0.34, + "learning_rate": 0.0005296690070210632, + "loss": 1.4581, + "theoretical_loss": 3.6338094632174274, + "tokens_seen": 1137180672 + }, + { + "epoch": 0.34, + "learning_rate": 0.0005296529588766299, + "loss": 1.491, + "theoretical_loss": 3.633790157545315, + "tokens_seen": 1137246208 + }, + { + "epoch": 0.34, + "learning_rate": 0.0005296369107321966, + "loss": 1.4107, + "theoretical_loss": 3.6337708532971833, + "tokens_seen": 1137311744 + }, + { + "epoch": 0.34, + "learning_rate": 0.0005296208625877633, + "loss": 1.4289, + "theoretical_loss": 3.6337515504728466, + "tokens_seen": 1137377280 + }, + { + "epoch": 0.34, + "learning_rate": 0.00052960481444333, + "loss": 1.3004, + "theoretical_loss": 3.633732249072117, + "tokens_seen": 1137442816 + }, + { + "epoch": 0.34, + "learning_rate": 0.0005295887662988967, + "loss": 1.402, + "theoretical_loss": 3.6337129490948072, + "tokens_seen": 1137508352 + }, + { + "epoch": 0.34, + "learning_rate": 0.0005295727181544634, + "loss": 1.4578, + "theoretical_loss": 3.6336936505407316, + "tokens_seen": 1137573888 + }, + { + "epoch": 0.34, + "learning_rate": 0.0005295566700100301, + "loss": 1.4885, + "theoretical_loss": 3.633674353409702, + "tokens_seen": 1137639424 + }, + { + "epoch": 0.34, + "learning_rate": 0.0005295406218655968, + "loss": 1.4755, + "theoretical_loss": 3.6336550577015316, + "tokens_seen": 1137704960 + }, + { + "epoch": 0.34, + "learning_rate": 0.0005295245737211635, + "loss": 1.4357, + "theoretical_loss": 3.633635763416034, + "tokens_seen": 1137770496 + }, + { + "epoch": 0.34, + "learning_rate": 0.0005295085255767302, + "loss": 1.4714, + "theoretical_loss": 3.633616470553022, + "tokens_seen": 1137836032 + }, + { + "epoch": 0.34, + "learning_rate": 0.0005294924774322969, + "loss": 1.5364, + "theoretical_loss": 3.633597179112309, + "tokens_seen": 1137901568 + }, + { + "epoch": 0.34, + "learning_rate": 0.0005294764292878636, + "loss": 1.6333, + "theoretical_loss": 3.6335778890937083, + "tokens_seen": 1137967104 + }, + { + "epoch": 0.34, + "learning_rate": 0.0005294603811434303, + "loss": 1.5054, + "theoretical_loss": 3.633558600497033, + "tokens_seen": 1138032640 + }, + { + "epoch": 0.34, + "learning_rate": 0.000529444332998997, + "loss": 1.4903, + "theoretical_loss": 3.633539313322096, + "tokens_seen": 1138098176 + }, + { + "epoch": 0.34, + "learning_rate": 0.0005294282848545637, + "loss": 1.449, + "theoretical_loss": 3.6335200275687116, + "tokens_seen": 1138163712 + }, + { + "epoch": 0.34, + "learning_rate": 0.0005294122367101304, + "loss": 1.4943, + "theoretical_loss": 3.633500743236693, + "tokens_seen": 1138229248 + }, + { + "epoch": 0.34, + "learning_rate": 0.0005293961885656971, + "loss": 1.3161, + "theoretical_loss": 3.6334814603258527, + "tokens_seen": 1138294784 + }, + { + "epoch": 0.34, + "learning_rate": 0.0005293801404212638, + "loss": 1.5084, + "theoretical_loss": 3.6334621788360053, + "tokens_seen": 1138360320 + }, + { + "epoch": 0.34, + "learning_rate": 0.0005293640922768305, + "loss": 1.4801, + "theoretical_loss": 3.6334428987669636, + "tokens_seen": 1138425856 + }, + { + "epoch": 0.34, + "learning_rate": 0.0005293480441323972, + "loss": 1.4387, + "theoretical_loss": 3.633423620118541, + "tokens_seen": 1138491392 + }, + { + "epoch": 0.35, + "learning_rate": 0.0005293319959879639, + "loss": 1.4266, + "theoretical_loss": 3.633404342890551, + "tokens_seen": 1138556928 + }, + { + "epoch": 0.35, + "learning_rate": 0.0005293159478435306, + "loss": 1.491, + "theoretical_loss": 3.6333850670828083, + "tokens_seen": 1138622464 + }, + { + "epoch": 0.35, + "learning_rate": 0.0005292998996990973, + "loss": 1.4231, + "theoretical_loss": 3.633365792695126, + "tokens_seen": 1138688000 + }, + { + "epoch": 0.35, + "learning_rate": 0.000529283851554664, + "loss": 1.3976, + "theoretical_loss": 3.6333465197273167, + "tokens_seen": 1138753536 + }, + { + "epoch": 0.35, + "learning_rate": 0.0005292678034102307, + "loss": 1.4193, + "theoretical_loss": 3.633327248179195, + "tokens_seen": 1138819072 + }, + { + "epoch": 0.35, + "learning_rate": 0.0005292517552657974, + "loss": 1.492, + "theoretical_loss": 3.633307978050575, + "tokens_seen": 1138884608 + }, + { + "epoch": 0.35, + "learning_rate": 0.0005292357071213641, + "loss": 1.4403, + "theoretical_loss": 3.6332887093412696, + "tokens_seen": 1138950144 + }, + { + "epoch": 0.35, + "learning_rate": 0.0005292196589769308, + "loss": 1.5056, + "theoretical_loss": 3.6332694420510934, + "tokens_seen": 1139015680 + }, + { + "epoch": 0.35, + "learning_rate": 0.0005292036108324975, + "loss": 1.4043, + "theoretical_loss": 3.6332501761798595, + "tokens_seen": 1139081216 + }, + { + "epoch": 0.35, + "learning_rate": 0.0005291875626880642, + "loss": 1.3877, + "theoretical_loss": 3.633230911727382, + "tokens_seen": 1139146752 + }, + { + "epoch": 0.35, + "learning_rate": 0.000529171514543631, + "loss": 1.5277, + "theoretical_loss": 3.633211648693475, + "tokens_seen": 1139212288 + }, + { + "epoch": 0.35, + "learning_rate": 0.0005291554663991976, + "loss": 1.587, + "theoretical_loss": 3.633192387077953, + "tokens_seen": 1139277824 + }, + { + "epoch": 0.35, + "learning_rate": 0.0005291394182547643, + "loss": 1.4724, + "theoretical_loss": 3.633173126880629, + "tokens_seen": 1139343360 + }, + { + "epoch": 0.35, + "learning_rate": 0.000529123370110331, + "loss": 1.4688, + "theoretical_loss": 3.633153868101317, + "tokens_seen": 1139408896 + }, + { + "epoch": 0.35, + "learning_rate": 0.0005291073219658978, + "loss": 1.4249, + "theoretical_loss": 3.633134610739832, + "tokens_seen": 1139474432 + }, + { + "epoch": 0.35, + "learning_rate": 0.0005290912738214644, + "loss": 1.4775, + "theoretical_loss": 3.6331153547959874, + "tokens_seen": 1139539968 + }, + { + "epoch": 0.35, + "learning_rate": 0.0005290752256770312, + "loss": 1.3823, + "theoretical_loss": 3.633096100269597, + "tokens_seen": 1139605504 + }, + { + "epoch": 0.35, + "learning_rate": 0.0005290591775325978, + "loss": 1.5397, + "theoretical_loss": 3.633076847160476, + "tokens_seen": 1139671040 + }, + { + "epoch": 0.35, + "learning_rate": 0.0005290431293881646, + "loss": 1.4521, + "theoretical_loss": 3.633057595468437, + "tokens_seen": 1139736576 + }, + { + "epoch": 0.35, + "learning_rate": 0.0005290270812437312, + "loss": 1.3631, + "theoretical_loss": 3.633038345193296, + "tokens_seen": 1139802112 + }, + { + "epoch": 0.35, + "learning_rate": 0.000529011033099298, + "loss": 1.4231, + "theoretical_loss": 3.6330190963348663, + "tokens_seen": 1139867648 + }, + { + "epoch": 0.35, + "learning_rate": 0.0005289949849548646, + "loss": 1.4692, + "theoretical_loss": 3.632999848892962, + "tokens_seen": 1139933184 + }, + { + "epoch": 0.35, + "learning_rate": 0.0005289789368104314, + "loss": 1.4161, + "theoretical_loss": 3.6329806028673985, + "tokens_seen": 1139998720 + }, + { + "epoch": 0.35, + "learning_rate": 0.000528962888665998, + "loss": 1.4881, + "theoretical_loss": 3.6329613582579885, + "tokens_seen": 1140064256 + }, + { + "epoch": 0.35, + "learning_rate": 0.0005289468405215648, + "loss": 1.4282, + "theoretical_loss": 3.632942115064548, + "tokens_seen": 1140129792 + }, + { + "epoch": 0.35, + "learning_rate": 0.0005289307923771314, + "loss": 1.4654, + "theoretical_loss": 3.6329228732868906, + "tokens_seen": 1140195328 + }, + { + "epoch": 0.35, + "learning_rate": 0.0005289147442326982, + "loss": 1.4271, + "theoretical_loss": 3.6329036329248305, + "tokens_seen": 1140260864 + }, + { + "epoch": 0.35, + "objective/train/docs_used": 417664, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.4124476909637451, + "objective/train/theoretical_loss": 3.632884393978183, + "objective/train/tokens_used": 1160786400, + "theoretical_loss": 3.632884393978183, + "tokens_seen": 1140326400 + }, + { + "epoch": 0.35, + "learning_rate": 0.0005288986960882648, + "loss": 1.4569, + "theoretical_loss": 3.632884393978183, + "tokens_seen": 1140326400 + }, + { + "epoch": 0.35, + "learning_rate": 0.0005288826479438316, + "loss": 1.4937, + "theoretical_loss": 3.632865156446763, + "tokens_seen": 1140391936 + }, + { + "epoch": 0.35, + "learning_rate": 0.0005288665997993982, + "loss": 1.4995, + "theoretical_loss": 3.6328459203303836, + "tokens_seen": 1140457472 + }, + { + "epoch": 0.35, + "learning_rate": 0.000528850551654965, + "loss": 1.4166, + "theoretical_loss": 3.63282668562886, + "tokens_seen": 1140523008 + }, + { + "epoch": 0.35, + "learning_rate": 0.0005288345035105316, + "loss": 1.5096, + "theoretical_loss": 3.632807452342007, + "tokens_seen": 1140588544 + }, + { + "epoch": 0.35, + "learning_rate": 0.0005288184553660984, + "loss": 1.4623, + "theoretical_loss": 3.63278822046964, + "tokens_seen": 1140654080 + }, + { + "epoch": 0.35, + "learning_rate": 0.000528802407221665, + "loss": 1.5011, + "theoretical_loss": 3.6327689900115727, + "tokens_seen": 1140719616 + }, + { + "epoch": 0.35, + "learning_rate": 0.0005287863590772318, + "loss": 1.4289, + "theoretical_loss": 3.63274976096762, + "tokens_seen": 1140785152 + }, + { + "epoch": 0.35, + "learning_rate": 0.0005287703109327984, + "loss": 1.4648, + "theoretical_loss": 3.6327305333375968, + "tokens_seen": 1140850688 + }, + { + "epoch": 0.35, + "learning_rate": 0.0005287542627883652, + "loss": 1.5287, + "theoretical_loss": 3.6327113071213177, + "tokens_seen": 1140916224 + }, + { + "epoch": 0.35, + "learning_rate": 0.0005287382146439318, + "loss": 1.3902, + "theoretical_loss": 3.632692082318598, + "tokens_seen": 1140981760 + }, + { + "epoch": 0.35, + "learning_rate": 0.0005287221664994986, + "loss": 1.44, + "theoretical_loss": 3.6326728589292525, + "tokens_seen": 1141047296 + }, + { + "epoch": 0.35, + "learning_rate": 0.0005287061183550652, + "loss": 1.484, + "theoretical_loss": 3.632653636953096, + "tokens_seen": 1141112832 + }, + { + "epoch": 0.35, + "learning_rate": 0.000528690070210632, + "loss": 1.4386, + "theoretical_loss": 3.6326344163899433, + "tokens_seen": 1141178368 + }, + { + "epoch": 0.35, + "learning_rate": 0.0005286740220661986, + "loss": 1.4044, + "theoretical_loss": 3.63261519723961, + "tokens_seen": 1141243904 + }, + { + "epoch": 0.35, + "learning_rate": 0.0005286579739217654, + "loss": 1.3433, + "theoretical_loss": 3.63259597950191, + "tokens_seen": 1141309440 + }, + { + "epoch": 0.35, + "learning_rate": 0.000528641925777332, + "loss": 1.5269, + "theoretical_loss": 3.63257676317666, + "tokens_seen": 1141374976 + }, + { + "epoch": 0.35, + "learning_rate": 0.0005286258776328988, + "loss": 1.4975, + "theoretical_loss": 3.6325575482636734, + "tokens_seen": 1141440512 + }, + { + "epoch": 0.35, + "learning_rate": 0.0005286098294884654, + "loss": 1.4525, + "theoretical_loss": 3.632538334762766, + "tokens_seen": 1141506048 + }, + { + "epoch": 0.35, + "learning_rate": 0.0005285937813440322, + "loss": 1.4739, + "theoretical_loss": 3.6325191226737537, + "tokens_seen": 1141571584 + }, + { + "epoch": 0.35, + "learning_rate": 0.0005285777331995988, + "loss": 1.5158, + "theoretical_loss": 3.6324999119964505, + "tokens_seen": 1141637120 + }, + { + "epoch": 0.35, + "learning_rate": 0.0005285616850551656, + "loss": 1.4177, + "theoretical_loss": 3.632480702730673, + "tokens_seen": 1141702656 + }, + { + "epoch": 0.35, + "learning_rate": 0.0005285456369107322, + "loss": 1.4878, + "theoretical_loss": 3.6324614948762344, + "tokens_seen": 1141768192 + }, + { + "epoch": 0.35, + "learning_rate": 0.000528529588766299, + "loss": 1.3771, + "theoretical_loss": 3.6324422884329524, + "tokens_seen": 1141833728 + }, + { + "epoch": 0.35, + "learning_rate": 0.0005285135406218656, + "loss": 1.4775, + "theoretical_loss": 3.6324230834006404, + "tokens_seen": 1141899264 + }, + { + "epoch": 0.35, + "learning_rate": 0.0005284974924774324, + "loss": 1.4413, + "theoretical_loss": 3.632403879779115, + "tokens_seen": 1141964800 + }, + { + "epoch": 0.35, + "learning_rate": 0.000528481444332999, + "loss": 1.4868, + "theoretical_loss": 3.632384677568191, + "tokens_seen": 1142030336 + }, + { + "epoch": 0.35, + "learning_rate": 0.0005284653961885657, + "loss": 1.4925, + "theoretical_loss": 3.632365476767684, + "tokens_seen": 1142095872 + }, + { + "epoch": 0.35, + "learning_rate": 0.0005284493480441324, + "loss": 1.4702, + "theoretical_loss": 3.63234627737741, + "tokens_seen": 1142161408 + }, + { + "epoch": 0.35, + "learning_rate": 0.0005284332998996991, + "loss": 1.3011, + "theoretical_loss": 3.6323270793971836, + "tokens_seen": 1142226944 + }, + { + "epoch": 0.35, + "learning_rate": 0.0005284172517552658, + "loss": 1.4061, + "theoretical_loss": 3.632307882826821, + "tokens_seen": 1142292480 + }, + { + "epoch": 0.35, + "learning_rate": 0.0005284012036108325, + "loss": 1.4554, + "theoretical_loss": 3.6322886876661373, + "tokens_seen": 1142358016 + }, + { + "epoch": 0.35, + "learning_rate": 0.0005283851554663992, + "loss": 1.4689, + "theoretical_loss": 3.632269493914949, + "tokens_seen": 1142423552 + }, + { + "epoch": 0.35, + "learning_rate": 0.0005283691073219659, + "loss": 1.5418, + "theoretical_loss": 3.6322503015730705, + "tokens_seen": 1142489088 + }, + { + "epoch": 0.35, + "learning_rate": 0.0005283530591775326, + "loss": 1.3737, + "theoretical_loss": 3.6322311106403182, + "tokens_seen": 1142554624 + }, + { + "epoch": 0.35, + "learning_rate": 0.0005283370110330993, + "loss": 1.3986, + "theoretical_loss": 3.632211921116508, + "tokens_seen": 1142620160 + }, + { + "epoch": 0.35, + "learning_rate": 0.000528320962888666, + "loss": 1.4834, + "theoretical_loss": 3.632192733001455, + "tokens_seen": 1142685696 + }, + { + "epoch": 0.35, + "learning_rate": 0.0005283049147442327, + "loss": 1.4043, + "theoretical_loss": 3.632173546294976, + "tokens_seen": 1142751232 + }, + { + "epoch": 0.35, + "learning_rate": 0.0005282888665997994, + "loss": 1.3334, + "theoretical_loss": 3.6321543609968856, + "tokens_seen": 1142816768 + }, + { + "epoch": 0.35, + "learning_rate": 0.0005282728184553661, + "loss": 1.434, + "theoretical_loss": 3.6321351771070005, + "tokens_seen": 1142882304 + }, + { + "epoch": 0.35, + "learning_rate": 0.0005282567703109328, + "loss": 1.435, + "theoretical_loss": 3.6321159946251367, + "tokens_seen": 1142947840 + }, + { + "epoch": 0.35, + "learning_rate": 0.0005282407221664995, + "loss": 1.4506, + "theoretical_loss": 3.6320968135511094, + "tokens_seen": 1143013376 + }, + { + "epoch": 0.35, + "learning_rate": 0.0005282246740220662, + "loss": 1.3463, + "theoretical_loss": 3.632077633884735, + "tokens_seen": 1143078912 + }, + { + "epoch": 0.35, + "learning_rate": 0.0005282086258776329, + "loss": 1.4769, + "theoretical_loss": 3.6320584556258297, + "tokens_seen": 1143144448 + }, + { + "epoch": 0.35, + "learning_rate": 0.0005281925777331996, + "loss": 1.3257, + "theoretical_loss": 3.6320392787742093, + "tokens_seen": 1143209984 + }, + { + "epoch": 0.35, + "learning_rate": 0.0005281765295887663, + "loss": 1.4205, + "theoretical_loss": 3.6320201033296895, + "tokens_seen": 1143275520 + }, + { + "epoch": 0.35, + "learning_rate": 0.000528160481444333, + "loss": 1.4432, + "theoretical_loss": 3.6320009292920874, + "tokens_seen": 1143341056 + }, + { + "epoch": 0.35, + "learning_rate": 0.0005281444332998997, + "loss": 1.4667, + "theoretical_loss": 3.6319817566612183, + "tokens_seen": 1143406592 + }, + { + "epoch": 0.35, + "learning_rate": 0.0005281283851554664, + "loss": 1.3577, + "theoretical_loss": 3.6319625854368986, + "tokens_seen": 1143472128 + }, + { + "epoch": 0.35, + "learning_rate": 0.0005281123370110331, + "loss": 1.4799, + "theoretical_loss": 3.6319434156189443, + "tokens_seen": 1143537664 + }, + { + "epoch": 0.35, + "objective/train/docs_used": 418760, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.3625807762145996, + "objective/train/theoretical_loss": 3.6319242472071718, + "objective/train/tokens_used": 1164063200, + "theoretical_loss": 3.6319242472071718, + "tokens_seen": 1143603200 + }, + { + "epoch": 0.35, + "learning_rate": 0.0005280962888665998, + "loss": 1.4022, + "theoretical_loss": 3.6319242472071718, + "tokens_seen": 1143603200 + }, + { + "epoch": 0.35, + "learning_rate": 0.0005280802407221665, + "loss": 1.535, + "theoretical_loss": 3.631905080201398, + "tokens_seen": 1143668736 + }, + { + "epoch": 0.35, + "learning_rate": 0.0005280641925777332, + "loss": 1.371, + "theoretical_loss": 3.6318859146014377, + "tokens_seen": 1143734272 + }, + { + "epoch": 0.35, + "learning_rate": 0.0005280481444332999, + "loss": 1.3752, + "theoretical_loss": 3.6318667504071094, + "tokens_seen": 1143799808 + }, + { + "epoch": 0.35, + "learning_rate": 0.0005280320962888666, + "loss": 1.401, + "theoretical_loss": 3.6318475876182275, + "tokens_seen": 1143865344 + }, + { + "epoch": 0.35, + "learning_rate": 0.0005280160481444333, + "loss": 1.4663, + "theoretical_loss": 3.631828426234609, + "tokens_seen": 1143930880 + }, + { + "epoch": 0.35, + "learning_rate": 0.000528, + "loss": 1.3667, + "theoretical_loss": 3.631809266256071, + "tokens_seen": 1143996416 + }, + { + "epoch": 0.35, + "learning_rate": 0.0005279839518555667, + "loss": 1.3792, + "theoretical_loss": 3.631790107682429, + "tokens_seen": 1144061952 + }, + { + "epoch": 0.35, + "learning_rate": 0.0005279679037111334, + "loss": 1.298, + "theoretical_loss": 3.6317709505135003, + "tokens_seen": 1144127488 + }, + { + "epoch": 0.35, + "learning_rate": 0.0005279518555667001, + "loss": 1.2977, + "theoretical_loss": 3.631751794749101, + "tokens_seen": 1144193024 + }, + { + "epoch": 0.35, + "learning_rate": 0.0005279358074222668, + "loss": 1.4809, + "theoretical_loss": 3.6317326403890484, + "tokens_seen": 1144258560 + }, + { + "epoch": 0.35, + "learning_rate": 0.0005279197592778335, + "loss": 1.3161, + "theoretical_loss": 3.6317134874331582, + "tokens_seen": 1144324096 + }, + { + "epoch": 0.35, + "learning_rate": 0.0005279037111334002, + "loss": 1.2813, + "theoretical_loss": 3.6316943358812477, + "tokens_seen": 1144389632 + }, + { + "epoch": 0.35, + "learning_rate": 0.0005278876629889669, + "loss": 1.4679, + "theoretical_loss": 3.631675185733133, + "tokens_seen": 1144455168 + }, + { + "epoch": 0.35, + "learning_rate": 0.0005278716148445336, + "loss": 1.4337, + "theoretical_loss": 3.6316560369886313, + "tokens_seen": 1144520704 + }, + { + "epoch": 0.35, + "learning_rate": 0.0005278555667001003, + "loss": 1.475, + "theoretical_loss": 3.6316368896475586, + "tokens_seen": 1144586240 + }, + { + "epoch": 0.35, + "learning_rate": 0.000527839518555667, + "loss": 1.4521, + "theoretical_loss": 3.6316177437097332, + "tokens_seen": 1144651776 + }, + { + "epoch": 0.35, + "learning_rate": 0.0005278234704112337, + "loss": 1.3007, + "theoretical_loss": 3.6315985991749704, + "tokens_seen": 1144717312 + }, + { + "epoch": 0.35, + "learning_rate": 0.0005278074222668004, + "loss": 1.3379, + "theoretical_loss": 3.631579456043088, + "tokens_seen": 1144782848 + }, + { + "epoch": 0.35, + "learning_rate": 0.0005277913741223671, + "loss": 1.3012, + "theoretical_loss": 3.631560314313903, + "tokens_seen": 1144848384 + }, + { + "epoch": 0.35, + "learning_rate": 0.0005277753259779338, + "loss": 1.412, + "theoretical_loss": 3.631541173987231, + "tokens_seen": 1144913920 + }, + { + "epoch": 0.35, + "learning_rate": 0.0005277592778335005, + "loss": 1.4207, + "theoretical_loss": 3.6315220350628907, + "tokens_seen": 1144979456 + }, + { + "epoch": 0.35, + "learning_rate": 0.0005277432296890673, + "loss": 1.3433, + "theoretical_loss": 3.631502897540698, + "tokens_seen": 1145044992 + }, + { + "epoch": 0.35, + "learning_rate": 0.0005277271815446339, + "loss": 1.4274, + "theoretical_loss": 3.6314837614204705, + "tokens_seen": 1145110528 + }, + { + "epoch": 0.35, + "learning_rate": 0.0005277111334002007, + "loss": 1.3951, + "theoretical_loss": 3.6314646267020247, + "tokens_seen": 1145176064 + }, + { + "epoch": 0.35, + "learning_rate": 0.0005276950852557673, + "loss": 1.4994, + "theoretical_loss": 3.6314454933851783, + "tokens_seen": 1145241600 + }, + { + "epoch": 0.35, + "learning_rate": 0.000527679037111334, + "loss": 1.4095, + "theoretical_loss": 3.6314263614697477, + "tokens_seen": 1145307136 + }, + { + "epoch": 0.35, + "learning_rate": 0.0005276629889669007, + "loss": 1.4563, + "theoretical_loss": 3.6314072309555505, + "tokens_seen": 1145372672 + }, + { + "epoch": 0.35, + "learning_rate": 0.0005276469408224675, + "loss": 1.5277, + "theoretical_loss": 3.631388101842404, + "tokens_seen": 1145438208 + }, + { + "epoch": 0.35, + "learning_rate": 0.0005276308926780341, + "loss": 1.3688, + "theoretical_loss": 3.6313689741301256, + "tokens_seen": 1145503744 + }, + { + "epoch": 0.35, + "learning_rate": 0.0005276148445336009, + "loss": 1.4682, + "theoretical_loss": 3.631349847818532, + "tokens_seen": 1145569280 + }, + { + "epoch": 0.35, + "learning_rate": 0.0005275987963891675, + "loss": 1.4107, + "theoretical_loss": 3.6313307229074407, + "tokens_seen": 1145634816 + }, + { + "epoch": 0.35, + "learning_rate": 0.0005275827482447343, + "loss": 1.3029, + "theoretical_loss": 3.6313115993966694, + "tokens_seen": 1145700352 + }, + { + "epoch": 0.35, + "learning_rate": 0.0005275667001003009, + "loss": 1.481, + "theoretical_loss": 3.6312924772860358, + "tokens_seen": 1145765888 + }, + { + "epoch": 0.35, + "learning_rate": 0.0005275506519558677, + "loss": 1.5234, + "theoretical_loss": 3.631273356575356, + "tokens_seen": 1145831424 + }, + { + "epoch": 0.35, + "learning_rate": 0.0005275346038114343, + "loss": 1.4332, + "theoretical_loss": 3.6312542372644483, + "tokens_seen": 1145896960 + }, + { + "epoch": 0.35, + "learning_rate": 0.0005275185556670011, + "loss": 1.4006, + "theoretical_loss": 3.6312351193531303, + "tokens_seen": 1145962496 + }, + { + "epoch": 0.35, + "learning_rate": 0.0005275025075225677, + "loss": 1.4614, + "theoretical_loss": 3.6312160028412195, + "tokens_seen": 1146028032 + }, + { + "epoch": 0.35, + "learning_rate": 0.0005274864593781345, + "loss": 1.4504, + "theoretical_loss": 3.631196887728533, + "tokens_seen": 1146093568 + }, + { + "epoch": 0.35, + "learning_rate": 0.0005274704112337011, + "loss": 1.4137, + "theoretical_loss": 3.6311777740148887, + "tokens_seen": 1146159104 + }, + { + "epoch": 0.35, + "learning_rate": 0.0005274543630892679, + "loss": 1.4811, + "theoretical_loss": 3.631158661700104, + "tokens_seen": 1146224640 + }, + { + "epoch": 0.35, + "learning_rate": 0.0005274383149448345, + "loss": 1.5481, + "theoretical_loss": 3.631139550783997, + "tokens_seen": 1146290176 + }, + { + "epoch": 0.35, + "learning_rate": 0.0005274222668004013, + "loss": 1.3583, + "theoretical_loss": 3.631120441266385, + "tokens_seen": 1146355712 + }, + { + "epoch": 0.35, + "learning_rate": 0.0005274062186559679, + "loss": 1.4196, + "theoretical_loss": 3.6311013331470856, + "tokens_seen": 1146421248 + }, + { + "epoch": 0.35, + "learning_rate": 0.0005273901705115347, + "loss": 1.3973, + "theoretical_loss": 3.631082226425917, + "tokens_seen": 1146486784 + }, + { + "epoch": 0.35, + "learning_rate": 0.0005273741223671013, + "loss": 1.3799, + "theoretical_loss": 3.6310631211026965, + "tokens_seen": 1146552320 + }, + { + "epoch": 0.35, + "learning_rate": 0.0005273580742226681, + "loss": 1.5066, + "theoretical_loss": 3.6310440171772425, + "tokens_seen": 1146617856 + }, + { + "epoch": 0.35, + "learning_rate": 0.0005273420260782347, + "loss": 1.3921, + "theoretical_loss": 3.6310249146493723, + "tokens_seen": 1146683392 + }, + { + "epoch": 0.35, + "learning_rate": 0.0005273259779338015, + "loss": 1.4365, + "theoretical_loss": 3.631005813518904, + "tokens_seen": 1146748928 + }, + { + "epoch": 0.35, + "learning_rate": 0.0005273099297893682, + "loss": 1.4046, + "theoretical_loss": 3.6309867137856555, + "tokens_seen": 1146814464 + }, + { + "epoch": 0.35, + "objective/train/docs_used": 419954, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.3998653888702393, + "objective/train/theoretical_loss": 3.630967615449445, + "objective/train/tokens_used": 1167340000, + "theoretical_loss": 3.630967615449445, + "tokens_seen": 1146880000 + }, + { + "epoch": 0.35, + "learning_rate": 0.0005272938816449349, + "loss": 1.3229, + "theoretical_loss": 3.630967615449445, + "tokens_seen": 1146880000 + }, + { + "epoch": 0.35, + "learning_rate": 0.0005272778335005016, + "loss": 1.3488, + "theoretical_loss": 3.63094851851009, + "tokens_seen": 1146945536 + }, + { + "epoch": 0.35, + "learning_rate": 0.0005272617853560683, + "loss": 1.5221, + "theoretical_loss": 3.6309294229674087, + "tokens_seen": 1147011072 + }, + { + "epoch": 0.35, + "learning_rate": 0.000527245737211635, + "loss": 1.5227, + "theoretical_loss": 3.63091032882122, + "tokens_seen": 1147076608 + }, + { + "epoch": 0.35, + "learning_rate": 0.0005272296890672017, + "loss": 1.3926, + "theoretical_loss": 3.630891236071341, + "tokens_seen": 1147142144 + }, + { + "epoch": 0.35, + "learning_rate": 0.0005272136409227683, + "loss": 1.4101, + "theoretical_loss": 3.63087214471759, + "tokens_seen": 1147207680 + }, + { + "epoch": 0.35, + "learning_rate": 0.0005271975927783351, + "loss": 1.3317, + "theoretical_loss": 3.6308530547597853, + "tokens_seen": 1147273216 + }, + { + "epoch": 0.35, + "learning_rate": 0.0005271815446339017, + "loss": 1.3852, + "theoretical_loss": 3.6308339661977445, + "tokens_seen": 1147338752 + }, + { + "epoch": 0.35, + "learning_rate": 0.0005271654964894685, + "loss": 1.4321, + "theoretical_loss": 3.630814879031287, + "tokens_seen": 1147404288 + }, + { + "epoch": 0.35, + "learning_rate": 0.0005271494483450351, + "loss": 1.4276, + "theoretical_loss": 3.6307957932602304, + "tokens_seen": 1147469824 + }, + { + "epoch": 0.35, + "learning_rate": 0.0005271334002006019, + "loss": 1.5155, + "theoretical_loss": 3.630776708884393, + "tokens_seen": 1147535360 + }, + { + "epoch": 0.35, + "learning_rate": 0.0005271173520561685, + "loss": 1.4316, + "theoretical_loss": 3.630757625903593, + "tokens_seen": 1147600896 + }, + { + "epoch": 0.35, + "learning_rate": 0.0005271013039117353, + "loss": 1.3617, + "theoretical_loss": 3.630738544317649, + "tokens_seen": 1147666432 + }, + { + "epoch": 0.35, + "learning_rate": 0.0005270852557673019, + "loss": 1.3626, + "theoretical_loss": 3.6307194641263796, + "tokens_seen": 1147731968 + }, + { + "epoch": 0.35, + "learning_rate": 0.0005270692076228686, + "loss": 1.5589, + "theoretical_loss": 3.630700385329603, + "tokens_seen": 1147797504 + }, + { + "epoch": 0.35, + "learning_rate": 0.0005270531594784353, + "loss": 1.4254, + "theoretical_loss": 3.6306813079271376, + "tokens_seen": 1147863040 + }, + { + "epoch": 0.35, + "learning_rate": 0.000527037111334002, + "loss": 1.458, + "theoretical_loss": 3.630662231918802, + "tokens_seen": 1147928576 + }, + { + "epoch": 0.35, + "learning_rate": 0.0005270210631895687, + "loss": 1.5009, + "theoretical_loss": 3.630643157304415, + "tokens_seen": 1147994112 + }, + { + "epoch": 0.35, + "learning_rate": 0.0005270050150451354, + "loss": 1.5188, + "theoretical_loss": 3.6306240840837942, + "tokens_seen": 1148059648 + }, + { + "epoch": 0.35, + "learning_rate": 0.0005269889669007021, + "loss": 1.4484, + "theoretical_loss": 3.6306050122567592, + "tokens_seen": 1148125184 + }, + { + "epoch": 0.35, + "learning_rate": 0.0005269729187562688, + "loss": 1.4087, + "theoretical_loss": 3.6305859418231283, + "tokens_seen": 1148190720 + }, + { + "epoch": 0.35, + "learning_rate": 0.0005269568706118355, + "loss": 1.3622, + "theoretical_loss": 3.6305668727827203, + "tokens_seen": 1148256256 + }, + { + "epoch": 0.35, + "learning_rate": 0.0005269408224674022, + "loss": 1.4302, + "theoretical_loss": 3.630547805135354, + "tokens_seen": 1148321792 + }, + { + "epoch": 0.35, + "learning_rate": 0.0005269247743229689, + "loss": 1.2927, + "theoretical_loss": 3.630528738880847, + "tokens_seen": 1148387328 + }, + { + "epoch": 0.35, + "learning_rate": 0.0005269087261785356, + "loss": 1.4072, + "theoretical_loss": 3.63050967401902, + "tokens_seen": 1148452864 + }, + { + "epoch": 0.35, + "learning_rate": 0.0005268926780341023, + "loss": 1.3853, + "theoretical_loss": 3.630490610549691, + "tokens_seen": 1148518400 + }, + { + "epoch": 0.35, + "learning_rate": 0.000526876629889669, + "loss": 1.4034, + "theoretical_loss": 3.630471548472678, + "tokens_seen": 1148583936 + }, + { + "epoch": 0.35, + "learning_rate": 0.0005268605817452357, + "loss": 1.5044, + "theoretical_loss": 3.6304524877878004, + "tokens_seen": 1148649472 + }, + { + "epoch": 0.35, + "learning_rate": 0.0005268445336008024, + "loss": 1.5153, + "theoretical_loss": 3.630433428494878, + "tokens_seen": 1148715008 + }, + { + "epoch": 0.35, + "learning_rate": 0.0005268284854563691, + "loss": 1.3924, + "theoretical_loss": 3.630414370593728, + "tokens_seen": 1148780544 + }, + { + "epoch": 0.35, + "learning_rate": 0.0005268124373119358, + "loss": 1.4653, + "theoretical_loss": 3.6303953140841716, + "tokens_seen": 1148846080 + }, + { + "epoch": 0.35, + "learning_rate": 0.0005267963891675025, + "loss": 1.4472, + "theoretical_loss": 3.6303762589660256, + "tokens_seen": 1148911616 + }, + { + "epoch": 0.35, + "learning_rate": 0.0005267803410230692, + "loss": 1.4506, + "theoretical_loss": 3.6303572052391107, + "tokens_seen": 1148977152 + }, + { + "epoch": 0.35, + "learning_rate": 0.0005267642928786359, + "loss": 1.3381, + "theoretical_loss": 3.630338152903245, + "tokens_seen": 1149042688 + }, + { + "epoch": 0.35, + "learning_rate": 0.0005267482447342026, + "loss": 1.4771, + "theoretical_loss": 3.630319101958248, + "tokens_seen": 1149108224 + }, + { + "epoch": 0.35, + "learning_rate": 0.0005267321965897693, + "loss": 1.3417, + "theoretical_loss": 3.630300052403939, + "tokens_seen": 1149173760 + }, + { + "epoch": 0.35, + "learning_rate": 0.000526716148445336, + "loss": 1.3962, + "theoretical_loss": 3.6302810042401368, + "tokens_seen": 1149239296 + }, + { + "epoch": 0.35, + "learning_rate": 0.0005267001003009027, + "loss": 1.3537, + "theoretical_loss": 3.6302619574666606, + "tokens_seen": 1149304832 + }, + { + "epoch": 0.35, + "learning_rate": 0.0005266840521564694, + "loss": 1.3809, + "theoretical_loss": 3.6302429120833297, + "tokens_seen": 1149370368 + }, + { + "epoch": 0.35, + "learning_rate": 0.0005266680040120361, + "loss": 1.3346, + "theoretical_loss": 3.630223868089964, + "tokens_seen": 1149435904 + }, + { + "epoch": 0.35, + "learning_rate": 0.0005266519558676028, + "loss": 1.4468, + "theoretical_loss": 3.6302048254863823, + "tokens_seen": 1149501440 + }, + { + "epoch": 0.35, + "learning_rate": 0.0005266359077231695, + "loss": 1.4101, + "theoretical_loss": 3.6301857842724035, + "tokens_seen": 1149566976 + }, + { + "epoch": 0.35, + "learning_rate": 0.0005266198595787362, + "loss": 1.3199, + "theoretical_loss": 3.6301667444478474, + "tokens_seen": 1149632512 + }, + { + "epoch": 0.35, + "learning_rate": 0.0005266038114343029, + "loss": 1.6378, + "theoretical_loss": 3.630147706012534, + "tokens_seen": 1149698048 + }, + { + "epoch": 0.35, + "learning_rate": 0.0005265877632898696, + "loss": 1.4765, + "theoretical_loss": 3.6301286689662824, + "tokens_seen": 1149763584 + }, + { + "epoch": 0.35, + "learning_rate": 0.0005265717151454363, + "loss": 1.3755, + "theoretical_loss": 3.630109633308911, + "tokens_seen": 1149829120 + }, + { + "epoch": 0.35, + "learning_rate": 0.000526555667001003, + "loss": 1.4227, + "theoretical_loss": 3.6300905990402415, + "tokens_seen": 1149894656 + }, + { + "epoch": 0.35, + "learning_rate": 0.0005265396188565697, + "loss": 1.3956, + "theoretical_loss": 3.6300715661600913, + "tokens_seen": 1149960192 + }, + { + "epoch": 0.35, + "learning_rate": 0.0005265235707121364, + "loss": 1.382, + "theoretical_loss": 3.630052534668281, + "tokens_seen": 1150025728 + }, + { + "epoch": 0.35, + "learning_rate": 0.0005265075225677031, + "loss": 1.4341, + "theoretical_loss": 3.63003350456463, + "tokens_seen": 1150091264 + }, + { + "epoch": 0.35, + "objective/train/docs_used": 421268, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.4183952808380127, + "objective/train/theoretical_loss": 3.630014475848958, + "objective/train/tokens_used": 1170616800, + "theoretical_loss": 3.630014475848958, + "tokens_seen": 1150156800 + }, + { + "epoch": 0.35, + "learning_rate": 0.0005264914744232698, + "loss": 1.3937, + "theoretical_loss": 3.630014475848958, + "tokens_seen": 1150156800 + }, + { + "epoch": 0.35, + "learning_rate": 0.0005264754262788365, + "loss": 1.453, + "theoretical_loss": 3.6299954485210852, + "tokens_seen": 1150222336 + }, + { + "epoch": 0.35, + "learning_rate": 0.0005264593781344032, + "loss": 1.3488, + "theoretical_loss": 3.629976422580831, + "tokens_seen": 1150287872 + }, + { + "epoch": 0.35, + "learning_rate": 0.00052644332998997, + "loss": 1.4154, + "theoretical_loss": 3.629957398028014, + "tokens_seen": 1150353408 + }, + { + "epoch": 0.35, + "learning_rate": 0.0005264272818455366, + "loss": 1.3818, + "theoretical_loss": 3.629938374862456, + "tokens_seen": 1150418944 + }, + { + "epoch": 0.35, + "learning_rate": 0.0005264112337011033, + "loss": 1.3684, + "theoretical_loss": 3.6299193530839755, + "tokens_seen": 1150484480 + }, + { + "epoch": 0.35, + "learning_rate": 0.00052639518555667, + "loss": 1.3071, + "theoretical_loss": 3.629900332692392, + "tokens_seen": 1150550016 + }, + { + "epoch": 0.35, + "learning_rate": 0.0005263791374122368, + "loss": 1.4607, + "theoretical_loss": 3.629881313687527, + "tokens_seen": 1150615552 + }, + { + "epoch": 0.35, + "learning_rate": 0.0005263630892678034, + "loss": 1.4058, + "theoretical_loss": 3.6298622960691995, + "tokens_seen": 1150681088 + }, + { + "epoch": 0.35, + "learning_rate": 0.0005263470411233702, + "loss": 1.4073, + "theoretical_loss": 3.629843279837229, + "tokens_seen": 1150746624 + }, + { + "epoch": 0.35, + "learning_rate": 0.0005263309929789368, + "loss": 1.4247, + "theoretical_loss": 3.629824264991436, + "tokens_seen": 1150812160 + }, + { + "epoch": 0.35, + "learning_rate": 0.0005263149448345036, + "loss": 1.3961, + "theoretical_loss": 3.6298052515316406, + "tokens_seen": 1150877696 + }, + { + "epoch": 0.35, + "learning_rate": 0.0005262988966900702, + "loss": 1.504, + "theoretical_loss": 3.6297862394576628, + "tokens_seen": 1150943232 + }, + { + "epoch": 0.35, + "learning_rate": 0.000526282848545637, + "loss": 1.3765, + "theoretical_loss": 3.629767228769323, + "tokens_seen": 1151008768 + }, + { + "epoch": 0.35, + "learning_rate": 0.0005262668004012036, + "loss": 1.3352, + "theoretical_loss": 3.62974821946644, + "tokens_seen": 1151074304 + }, + { + "epoch": 0.35, + "learning_rate": 0.0005262507522567704, + "loss": 1.3761, + "theoretical_loss": 3.6297292115488355, + "tokens_seen": 1151139840 + }, + { + "epoch": 0.35, + "learning_rate": 0.000526234704112337, + "loss": 1.2111, + "theoretical_loss": 3.6297102050163286, + "tokens_seen": 1151205376 + }, + { + "epoch": 0.35, + "learning_rate": 0.0005262186559679038, + "loss": 1.5468, + "theoretical_loss": 3.6296911998687404, + "tokens_seen": 1151270912 + }, + { + "epoch": 0.35, + "learning_rate": 0.0005262026078234704, + "loss": 1.3993, + "theoretical_loss": 3.629672196105891, + "tokens_seen": 1151336448 + }, + { + "epoch": 0.35, + "learning_rate": 0.0005261865596790372, + "loss": 1.4392, + "theoretical_loss": 3.6296531937276004, + "tokens_seen": 1151401984 + }, + { + "epoch": 0.35, + "learning_rate": 0.0005261705115346038, + "loss": 1.4753, + "theoretical_loss": 3.629634192733689, + "tokens_seen": 1151467520 + }, + { + "epoch": 0.35, + "learning_rate": 0.0005261544633901706, + "loss": 1.3184, + "theoretical_loss": 3.6296151931239766, + "tokens_seen": 1151533056 + }, + { + "epoch": 0.35, + "learning_rate": 0.0005261384152457372, + "loss": 1.4944, + "theoretical_loss": 3.629596194898285, + "tokens_seen": 1151598592 + }, + { + "epoch": 0.35, + "learning_rate": 0.000526122367101304, + "loss": 1.4653, + "theoretical_loss": 3.6295771980564333, + "tokens_seen": 1151664128 + }, + { + "epoch": 0.35, + "learning_rate": 0.0005261063189568707, + "loss": 1.341, + "theoretical_loss": 3.6295582025982425, + "tokens_seen": 1151729664 + }, + { + "epoch": 0.35, + "learning_rate": 0.0005260902708124374, + "loss": 1.438, + "theoretical_loss": 3.629539208523533, + "tokens_seen": 1151795200 + }, + { + "epoch": 0.35, + "learning_rate": 0.0005260742226680041, + "loss": 1.4066, + "theoretical_loss": 3.6295202158321254, + "tokens_seen": 1151860736 + }, + { + "epoch": 0.35, + "learning_rate": 0.0005260581745235708, + "loss": 1.386, + "theoretical_loss": 3.6295012245238403, + "tokens_seen": 1151926272 + }, + { + "epoch": 0.35, + "learning_rate": 0.0005260421263791375, + "loss": 1.3387, + "theoretical_loss": 3.6294822345984983, + "tokens_seen": 1151991808 + }, + { + "epoch": 0.35, + "learning_rate": 0.0005260260782347042, + "loss": 1.4038, + "theoretical_loss": 3.62946324605592, + "tokens_seen": 1152057344 + }, + { + "epoch": 0.35, + "learning_rate": 0.0005260100300902709, + "loss": 1.2942, + "theoretical_loss": 3.6294442588959255, + "tokens_seen": 1152122880 + }, + { + "epoch": 0.35, + "learning_rate": 0.0005259939819458376, + "loss": 1.3974, + "theoretical_loss": 3.6294252731183363, + "tokens_seen": 1152188416 + }, + { + "epoch": 0.35, + "learning_rate": 0.0005259779338014043, + "loss": 1.4848, + "theoretical_loss": 3.629406288722973, + "tokens_seen": 1152253952 + }, + { + "epoch": 0.35, + "learning_rate": 0.000525961885656971, + "loss": 1.3558, + "theoretical_loss": 3.629387305709656, + "tokens_seen": 1152319488 + }, + { + "epoch": 0.35, + "learning_rate": 0.0005259458375125377, + "loss": 1.4277, + "theoretical_loss": 3.6293683240782064, + "tokens_seen": 1152385024 + }, + { + "epoch": 0.35, + "learning_rate": 0.0005259297893681044, + "loss": 1.3503, + "theoretical_loss": 3.629349343828445, + "tokens_seen": 1152450560 + }, + { + "epoch": 0.35, + "learning_rate": 0.0005259137412236711, + "loss": 1.4142, + "theoretical_loss": 3.6293303649601922, + "tokens_seen": 1152516096 + }, + { + "epoch": 0.35, + "learning_rate": 0.0005258976930792378, + "loss": 1.4218, + "theoretical_loss": 3.62931138747327, + "tokens_seen": 1152581632 + }, + { + "epoch": 0.35, + "learning_rate": 0.0005258816449348045, + "loss": 1.3915, + "theoretical_loss": 3.6292924113674983, + "tokens_seen": 1152647168 + }, + { + "epoch": 0.35, + "learning_rate": 0.0005258655967903712, + "loss": 1.335, + "theoretical_loss": 3.629273436642698, + "tokens_seen": 1152712704 + }, + { + "epoch": 0.35, + "learning_rate": 0.0005258495486459379, + "loss": 1.4529, + "theoretical_loss": 3.6292544632986905, + "tokens_seen": 1152778240 + }, + { + "epoch": 0.35, + "learning_rate": 0.0005258335005015046, + "loss": 1.4165, + "theoretical_loss": 3.6292354913352973, + "tokens_seen": 1152843776 + }, + { + "epoch": 0.35, + "learning_rate": 0.0005258174523570712, + "loss": 1.5607, + "theoretical_loss": 3.6292165207523386, + "tokens_seen": 1152909312 + }, + { + "epoch": 0.35, + "learning_rate": 0.000525801404212638, + "loss": 1.4979, + "theoretical_loss": 3.629197551549636, + "tokens_seen": 1152974848 + }, + { + "epoch": 0.35, + "learning_rate": 0.0005257853560682046, + "loss": 1.3408, + "theoretical_loss": 3.6291785837270103, + "tokens_seen": 1153040384 + }, + { + "epoch": 0.35, + "learning_rate": 0.0005257693079237714, + "loss": 1.4593, + "theoretical_loss": 3.6291596172842833, + "tokens_seen": 1153105920 + }, + { + "epoch": 0.35, + "learning_rate": 0.000525753259779338, + "loss": 1.5497, + "theoretical_loss": 3.6291406522212757, + "tokens_seen": 1153171456 + }, + { + "epoch": 0.35, + "learning_rate": 0.0005257372116349048, + "loss": 1.3932, + "theoretical_loss": 3.629121688537808, + "tokens_seen": 1153236992 + }, + { + "epoch": 0.35, + "learning_rate": 0.0005257211634904714, + "loss": 1.2646, + "theoretical_loss": 3.629102726233703, + "tokens_seen": 1153302528 + }, + { + "epoch": 0.35, + "learning_rate": 0.0005257051153460382, + "loss": 1.4233, + "theoretical_loss": 3.6290837653087813, + "tokens_seen": 1153368064 + }, + { + "epoch": 0.35, + "objective/train/docs_used": 422447, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.5732917785644531, + "objective/train/theoretical_loss": 3.6290648057628636, + "objective/train/tokens_used": 1173893600, + "theoretical_loss": 3.6290648057628636, + "tokens_seen": 1153433600 + }, + { + "epoch": 0.35, + "learning_rate": 0.0005256890672016048, + "loss": 1.4629, + "theoretical_loss": 3.6290648057628636, + "tokens_seen": 1153433600 + }, + { + "epoch": 0.35, + "learning_rate": 0.0005256730190571715, + "loss": 1.3223, + "theoretical_loss": 3.6290458475957723, + "tokens_seen": 1153499136 + }, + { + "epoch": 0.35, + "learning_rate": 0.0005256569709127382, + "loss": 1.4388, + "theoretical_loss": 3.629026890807328, + "tokens_seen": 1153564672 + }, + { + "epoch": 0.35, + "learning_rate": 0.0005256409227683049, + "loss": 1.3846, + "theoretical_loss": 3.6290079353973526, + "tokens_seen": 1153630208 + }, + { + "epoch": 0.35, + "learning_rate": 0.0005256248746238716, + "loss": 1.4133, + "theoretical_loss": 3.6289889813656675, + "tokens_seen": 1153695744 + }, + { + "epoch": 0.35, + "learning_rate": 0.0005256088264794383, + "loss": 1.3509, + "theoretical_loss": 3.628970028712094, + "tokens_seen": 1153761280 + }, + { + "epoch": 0.35, + "learning_rate": 0.000525592778335005, + "loss": 1.3169, + "theoretical_loss": 3.6289510774364535, + "tokens_seen": 1153826816 + }, + { + "epoch": 0.35, + "learning_rate": 0.0005255767301905717, + "loss": 1.4273, + "theoretical_loss": 3.628932127538568, + "tokens_seen": 1153892352 + }, + { + "epoch": 0.35, + "learning_rate": 0.0005255606820461384, + "loss": 1.3575, + "theoretical_loss": 3.6289131790182587, + "tokens_seen": 1153957888 + }, + { + "epoch": 0.35, + "learning_rate": 0.0005255446339017051, + "loss": 1.367, + "theoretical_loss": 3.628894231875348, + "tokens_seen": 1154023424 + }, + { + "epoch": 0.35, + "learning_rate": 0.0005255285857572718, + "loss": 1.4082, + "theoretical_loss": 3.6288752861096567, + "tokens_seen": 1154088960 + }, + { + "epoch": 0.35, + "learning_rate": 0.0005255125376128385, + "loss": 1.4166, + "theoretical_loss": 3.628856341721007, + "tokens_seen": 1154154496 + }, + { + "epoch": 0.35, + "learning_rate": 0.0005254964894684052, + "loss": 1.4409, + "theoretical_loss": 3.62883739870922, + "tokens_seen": 1154220032 + }, + { + "epoch": 0.35, + "learning_rate": 0.0005254804413239719, + "loss": 1.3154, + "theoretical_loss": 3.628818457074118, + "tokens_seen": 1154285568 + }, + { + "epoch": 0.35, + "learning_rate": 0.0005254643931795386, + "loss": 1.3946, + "theoretical_loss": 3.6287995168155227, + "tokens_seen": 1154351104 + }, + { + "epoch": 0.35, + "learning_rate": 0.0005254483450351053, + "loss": 1.4397, + "theoretical_loss": 3.628780577933256, + "tokens_seen": 1154416640 + }, + { + "epoch": 0.35, + "learning_rate": 0.000525432296890672, + "loss": 1.4733, + "theoretical_loss": 3.6287616404271397, + "tokens_seen": 1154482176 + }, + { + "epoch": 0.35, + "learning_rate": 0.0005254162487462387, + "loss": 1.4352, + "theoretical_loss": 3.6287427042969953, + "tokens_seen": 1154547712 + }, + { + "epoch": 0.35, + "learning_rate": 0.0005254002006018054, + "loss": 1.4983, + "theoretical_loss": 3.6287237695426455, + "tokens_seen": 1154613248 + }, + { + "epoch": 0.35, + "learning_rate": 0.0005253841524573721, + "loss": 1.4205, + "theoretical_loss": 3.6287048361639114, + "tokens_seen": 1154678784 + }, + { + "epoch": 0.35, + "learning_rate": 0.0005253681043129388, + "loss": 1.4468, + "theoretical_loss": 3.6286859041606156, + "tokens_seen": 1154744320 + }, + { + "epoch": 0.35, + "learning_rate": 0.0005253520561685055, + "loss": 1.3663, + "theoretical_loss": 3.62866697353258, + "tokens_seen": 1154809856 + }, + { + "epoch": 0.35, + "learning_rate": 0.0005253360080240722, + "loss": 1.3442, + "theoretical_loss": 3.6286480442796267, + "tokens_seen": 1154875392 + }, + { + "epoch": 0.35, + "learning_rate": 0.0005253199598796389, + "loss": 1.4012, + "theoretical_loss": 3.628629116401578, + "tokens_seen": 1154940928 + }, + { + "epoch": 0.35, + "learning_rate": 0.0005253039117352056, + "loss": 1.4151, + "theoretical_loss": 3.6286101898982555, + "tokens_seen": 1155006464 + }, + { + "epoch": 0.35, + "learning_rate": 0.0005252878635907723, + "loss": 1.3428, + "theoretical_loss": 3.628591264769481, + "tokens_seen": 1155072000 + }, + { + "epoch": 0.35, + "learning_rate": 0.000525271815446339, + "loss": 1.3926, + "theoretical_loss": 3.628572341015078, + "tokens_seen": 1155137536 + }, + { + "epoch": 0.35, + "learning_rate": 0.0005252557673019057, + "loss": 1.3958, + "theoretical_loss": 3.6285534186348682, + "tokens_seen": 1155203072 + }, + { + "epoch": 0.35, + "learning_rate": 0.0005252397191574724, + "loss": 1.3124, + "theoretical_loss": 3.628534497628673, + "tokens_seen": 1155268608 + }, + { + "epoch": 0.35, + "learning_rate": 0.0005252236710130391, + "loss": 1.4117, + "theoretical_loss": 3.628515577996316, + "tokens_seen": 1155334144 + }, + { + "epoch": 0.35, + "learning_rate": 0.0005252076228686058, + "loss": 1.3205, + "theoretical_loss": 3.6284966597376185, + "tokens_seen": 1155399680 + }, + { + "epoch": 0.35, + "learning_rate": 0.0005251915747241725, + "loss": 1.441, + "theoretical_loss": 3.628477742852403, + "tokens_seen": 1155465216 + }, + { + "epoch": 0.35, + "learning_rate": 0.0005251755265797392, + "loss": 1.2762, + "theoretical_loss": 3.6284588273404927, + "tokens_seen": 1155530752 + }, + { + "epoch": 0.35, + "learning_rate": 0.0005251594784353059, + "loss": 1.3841, + "theoretical_loss": 3.628439913201709, + "tokens_seen": 1155596288 + }, + { + "epoch": 0.35, + "learning_rate": 0.0005251434302908726, + "loss": 1.4124, + "theoretical_loss": 3.6284210004358757, + "tokens_seen": 1155661824 + }, + { + "epoch": 0.35, + "learning_rate": 0.0005251273821464393, + "loss": 1.4002, + "theoretical_loss": 3.6284020890428135, + "tokens_seen": 1155727360 + }, + { + "epoch": 0.35, + "learning_rate": 0.000525111334002006, + "loss": 1.373, + "theoretical_loss": 3.6283831790223458, + "tokens_seen": 1155792896 + }, + { + "epoch": 0.35, + "learning_rate": 0.0005250952858575727, + "loss": 1.357, + "theoretical_loss": 3.6283642703742958, + "tokens_seen": 1155858432 + }, + { + "epoch": 0.35, + "learning_rate": 0.0005250792377131394, + "loss": 1.4311, + "theoretical_loss": 3.628345363098485, + "tokens_seen": 1155923968 + }, + { + "epoch": 0.35, + "learning_rate": 0.0005250631895687061, + "loss": 1.3615, + "theoretical_loss": 3.6283264571947367, + "tokens_seen": 1155989504 + }, + { + "epoch": 0.35, + "learning_rate": 0.0005250471414242728, + "loss": 1.3443, + "theoretical_loss": 3.628307552662873, + "tokens_seen": 1156055040 + }, + { + "epoch": 0.35, + "learning_rate": 0.0005250310932798395, + "loss": 1.4604, + "theoretical_loss": 3.628288649502718, + "tokens_seen": 1156120576 + }, + { + "epoch": 0.35, + "learning_rate": 0.0005250150451354063, + "loss": 1.3668, + "theoretical_loss": 3.6282697477140924, + "tokens_seen": 1156186112 + }, + { + "epoch": 0.35, + "learning_rate": 0.0005249989969909729, + "loss": 1.2981, + "theoretical_loss": 3.6282508472968202, + "tokens_seen": 1156251648 + }, + { + "epoch": 0.35, + "learning_rate": 0.0005249829488465397, + "loss": 1.4016, + "theoretical_loss": 3.6282319482507237, + "tokens_seen": 1156317184 + }, + { + "epoch": 0.35, + "learning_rate": 0.0005249669007021063, + "loss": 1.3971, + "theoretical_loss": 3.628213050575626, + "tokens_seen": 1156382720 + }, + { + "epoch": 0.35, + "learning_rate": 0.000524950852557673, + "loss": 1.2904, + "theoretical_loss": 3.62819415427135, + "tokens_seen": 1156448256 + }, + { + "epoch": 0.35, + "learning_rate": 0.0005249348044132397, + "loss": 1.3529, + "theoretical_loss": 3.628175259337719, + "tokens_seen": 1156513792 + }, + { + "epoch": 0.35, + "learning_rate": 0.0005249187562688065, + "loss": 1.3258, + "theoretical_loss": 3.628156365774555, + "tokens_seen": 1156579328 + }, + { + "epoch": 0.35, + "learning_rate": 0.0005249027081243732, + "loss": 1.5033, + "theoretical_loss": 3.628137473581681, + "tokens_seen": 1156644864 + }, + { + "epoch": 0.35, + "objective/train/docs_used": 423626, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.221970558166504, + "objective/train/theoretical_loss": 3.628118582758921, + "objective/train/tokens_used": 1177170400, + "theoretical_loss": 3.628118582758921, + "tokens_seen": 1156710400 + }, + { + "epoch": 0.35, + "learning_rate": 0.0005248866599799399, + "loss": 1.2906, + "theoretical_loss": 3.628118582758921, + "tokens_seen": 1156710400 + }, + { + "epoch": 0.35, + "learning_rate": 0.0005248706118355066, + "loss": 1.3094, + "theoretical_loss": 3.6280996933060967, + "tokens_seen": 1156775936 + }, + { + "epoch": 0.35, + "learning_rate": 0.0005248545636910733, + "loss": 1.3531, + "theoretical_loss": 3.6280808052230316, + "tokens_seen": 1156841472 + }, + { + "epoch": 0.35, + "learning_rate": 0.00052483851554664, + "loss": 1.3183, + "theoretical_loss": 3.6280619185095495, + "tokens_seen": 1156907008 + }, + { + "epoch": 0.35, + "learning_rate": 0.0005248224674022067, + "loss": 1.4309, + "theoretical_loss": 3.628043033165473, + "tokens_seen": 1156972544 + }, + { + "epoch": 0.35, + "learning_rate": 0.0005248064192577734, + "loss": 1.4198, + "theoretical_loss": 3.6280241491906247, + "tokens_seen": 1157038080 + }, + { + "epoch": 0.35, + "learning_rate": 0.0005247903711133401, + "loss": 1.2694, + "theoretical_loss": 3.628005266584829, + "tokens_seen": 1157103616 + }, + { + "epoch": 0.35, + "learning_rate": 0.0005247743229689068, + "loss": 1.4218, + "theoretical_loss": 3.6279863853479077, + "tokens_seen": 1157169152 + }, + { + "epoch": 0.35, + "learning_rate": 0.0005247582748244735, + "loss": 1.4571, + "theoretical_loss": 3.6279675054796856, + "tokens_seen": 1157234688 + }, + { + "epoch": 0.35, + "learning_rate": 0.0005247422266800402, + "loss": 1.4818, + "theoretical_loss": 3.6279486269799843, + "tokens_seen": 1157300224 + }, + { + "epoch": 0.35, + "learning_rate": 0.0005247261785356069, + "loss": 1.3429, + "theoretical_loss": 3.6279297498486285, + "tokens_seen": 1157365760 + }, + { + "epoch": 0.35, + "learning_rate": 0.0005247101303911736, + "loss": 1.3576, + "theoretical_loss": 3.6279108740854413, + "tokens_seen": 1157431296 + }, + { + "epoch": 0.35, + "learning_rate": 0.0005246940822467403, + "loss": 1.2634, + "theoretical_loss": 3.6278919996902452, + "tokens_seen": 1157496832 + }, + { + "epoch": 0.35, + "learning_rate": 0.000524678034102307, + "loss": 1.3928, + "theoretical_loss": 3.6278731266628643, + "tokens_seen": 1157562368 + }, + { + "epoch": 0.35, + "learning_rate": 0.0005246619859578737, + "loss": 1.4643, + "theoretical_loss": 3.6278542550031223, + "tokens_seen": 1157627904 + }, + { + "epoch": 0.35, + "learning_rate": 0.0005246459378134404, + "loss": 1.397, + "theoretical_loss": 3.627835384710842, + "tokens_seen": 1157693440 + }, + { + "epoch": 0.35, + "learning_rate": 0.0005246298896690071, + "loss": 1.4168, + "theoretical_loss": 3.6278165157858475, + "tokens_seen": 1157758976 + }, + { + "epoch": 0.35, + "learning_rate": 0.0005246138415245738, + "loss": 1.3509, + "theoretical_loss": 3.627797648227962, + "tokens_seen": 1157824512 + }, + { + "epoch": 0.35, + "learning_rate": 0.0005245977933801405, + "loss": 1.5083, + "theoretical_loss": 3.6277787820370087, + "tokens_seen": 1157890048 + }, + { + "epoch": 0.35, + "learning_rate": 0.0005245817452357072, + "loss": 1.3802, + "theoretical_loss": 3.6277599172128125, + "tokens_seen": 1157955584 + }, + { + "epoch": 0.35, + "learning_rate": 0.0005245656970912739, + "loss": 1.3715, + "theoretical_loss": 3.6277410537551953, + "tokens_seen": 1158021120 + }, + { + "epoch": 0.35, + "learning_rate": 0.0005245496489468406, + "loss": 1.3452, + "theoretical_loss": 3.627722191663983, + "tokens_seen": 1158086656 + }, + { + "epoch": 0.35, + "learning_rate": 0.0005245336008024073, + "loss": 1.3085, + "theoretical_loss": 3.6277033309389966, + "tokens_seen": 1158152192 + }, + { + "epoch": 0.35, + "learning_rate": 0.000524517552657974, + "loss": 1.381, + "theoretical_loss": 3.627684471580062, + "tokens_seen": 1158217728 + }, + { + "epoch": 0.35, + "learning_rate": 0.0005245015045135407, + "loss": 1.4201, + "theoretical_loss": 3.627665613587002, + "tokens_seen": 1158283264 + }, + { + "epoch": 0.35, + "learning_rate": 0.0005244854563691074, + "loss": 1.4428, + "theoretical_loss": 3.6276467569596402, + "tokens_seen": 1158348800 + }, + { + "epoch": 0.35, + "learning_rate": 0.0005244694082246741, + "loss": 1.2698, + "theoretical_loss": 3.6276279016978012, + "tokens_seen": 1158414336 + }, + { + "epoch": 0.35, + "learning_rate": 0.0005244533600802408, + "loss": 1.361, + "theoretical_loss": 3.6276090478013088, + "tokens_seen": 1158479872 + }, + { + "epoch": 0.35, + "learning_rate": 0.0005244373119358075, + "loss": 1.5199, + "theoretical_loss": 3.6275901952699865, + "tokens_seen": 1158545408 + }, + { + "epoch": 0.35, + "learning_rate": 0.0005244212637913741, + "loss": 1.4718, + "theoretical_loss": 3.627571344103658, + "tokens_seen": 1158610944 + }, + { + "epoch": 0.35, + "learning_rate": 0.0005244052156469409, + "loss": 1.4101, + "theoretical_loss": 3.6275524943021478, + "tokens_seen": 1158676480 + }, + { + "epoch": 0.35, + "learning_rate": 0.0005243891675025075, + "loss": 1.3997, + "theoretical_loss": 3.62753364586528, + "tokens_seen": 1158742016 + }, + { + "epoch": 0.35, + "learning_rate": 0.0005243731193580743, + "loss": 1.4385, + "theoretical_loss": 3.6275147987928777, + "tokens_seen": 1158807552 + }, + { + "epoch": 0.35, + "learning_rate": 0.0005243570712136409, + "loss": 1.3241, + "theoretical_loss": 3.6274959530847664, + "tokens_seen": 1158873088 + }, + { + "epoch": 0.35, + "learning_rate": 0.0005243410230692077, + "loss": 1.4005, + "theoretical_loss": 3.6274771087407687, + "tokens_seen": 1158938624 + }, + { + "epoch": 0.35, + "learning_rate": 0.0005243249749247743, + "loss": 1.4585, + "theoretical_loss": 3.6274582657607097, + "tokens_seen": 1159004160 + }, + { + "epoch": 0.35, + "learning_rate": 0.0005243089267803411, + "loss": 1.3914, + "theoretical_loss": 3.6274394241444132, + "tokens_seen": 1159069696 + }, + { + "epoch": 0.35, + "learning_rate": 0.0005242928786359077, + "loss": 1.4963, + "theoretical_loss": 3.6274205838917037, + "tokens_seen": 1159135232 + }, + { + "epoch": 0.35, + "learning_rate": 0.0005242768304914744, + "loss": 1.4671, + "theoretical_loss": 3.6274017450024054, + "tokens_seen": 1159200768 + }, + { + "epoch": 0.35, + "learning_rate": 0.0005242607823470411, + "loss": 1.3438, + "theoretical_loss": 3.627382907476342, + "tokens_seen": 1159266304 + }, + { + "epoch": 0.35, + "learning_rate": 0.0005242447342026078, + "loss": 1.2351, + "theoretical_loss": 3.6273640713133384, + "tokens_seen": 1159331840 + }, + { + "epoch": 0.35, + "learning_rate": 0.0005242286860581745, + "loss": 1.3169, + "theoretical_loss": 3.6273452365132184, + "tokens_seen": 1159397376 + }, + { + "epoch": 0.35, + "learning_rate": 0.0005242126379137412, + "loss": 1.4161, + "theoretical_loss": 3.627326403075807, + "tokens_seen": 1159462912 + }, + { + "epoch": 0.35, + "learning_rate": 0.0005241965897693079, + "loss": 1.4341, + "theoretical_loss": 3.6273075710009284, + "tokens_seen": 1159528448 + }, + { + "epoch": 0.35, + "learning_rate": 0.0005241805416248746, + "loss": 1.3664, + "theoretical_loss": 3.6272887402884066, + "tokens_seen": 1159593984 + }, + { + "epoch": 0.35, + "learning_rate": 0.0005241644934804413, + "loss": 1.4545, + "theoretical_loss": 3.6272699109380664, + "tokens_seen": 1159659520 + }, + { + "epoch": 0.35, + "learning_rate": 0.000524148445336008, + "loss": 1.3381, + "theoretical_loss": 3.627251082949732, + "tokens_seen": 1159725056 + }, + { + "epoch": 0.35, + "learning_rate": 0.0005241323971915747, + "loss": 1.4313, + "theoretical_loss": 3.6272322563232287, + "tokens_seen": 1159790592 + }, + { + "epoch": 0.35, + "learning_rate": 0.0005241163490471414, + "loss": 1.3048, + "theoretical_loss": 3.6272134310583803, + "tokens_seen": 1159856128 + }, + { + "epoch": 0.35, + "learning_rate": 0.0005241003009027081, + "loss": 1.3836, + "theoretical_loss": 3.627194607155011, + "tokens_seen": 1159921664 + }, + { + "epoch": 0.35, + "objective/train/docs_used": 424758, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.460275411605835, + "objective/train/theoretical_loss": 3.6271757846129464, + "objective/train/tokens_used": 1180447200, + "theoretical_loss": 3.6271757846129464, + "tokens_seen": 1159987200 + }, + { + "epoch": 0.35, + "learning_rate": 0.0005240842527582748, + "loss": 1.5013, + "theoretical_loss": 3.6271757846129464, + "tokens_seen": 1159987200 + }, + { + "epoch": 0.35, + "learning_rate": 0.0005240682046138415, + "loss": 1.3871, + "theoretical_loss": 3.6271569634320113, + "tokens_seen": 1160052736 + }, + { + "epoch": 0.35, + "learning_rate": 0.0005240521564694082, + "loss": 1.4407, + "theoretical_loss": 3.6271381436120294, + "tokens_seen": 1160118272 + }, + { + "epoch": 0.35, + "learning_rate": 0.0005240361083249749, + "loss": 1.4165, + "theoretical_loss": 3.627119325152826, + "tokens_seen": 1160183808 + }, + { + "epoch": 0.35, + "learning_rate": 0.0005240200601805416, + "loss": 1.4158, + "theoretical_loss": 3.627100508054226, + "tokens_seen": 1160249344 + }, + { + "epoch": 0.35, + "learning_rate": 0.0005240040120361083, + "loss": 1.3921, + "theoretical_loss": 3.627081692316053, + "tokens_seen": 1160314880 + }, + { + "epoch": 0.35, + "learning_rate": 0.000523987963891675, + "loss": 1.4014, + "theoretical_loss": 3.6270628779381333, + "tokens_seen": 1160380416 + }, + { + "epoch": 0.35, + "learning_rate": 0.0005239719157472417, + "loss": 1.4359, + "theoretical_loss": 3.627044064920291, + "tokens_seen": 1160445952 + }, + { + "epoch": 0.35, + "learning_rate": 0.0005239558676028084, + "loss": 1.4281, + "theoretical_loss": 3.6270252532623513, + "tokens_seen": 1160511488 + }, + { + "epoch": 0.35, + "learning_rate": 0.0005239398194583751, + "loss": 1.3691, + "theoretical_loss": 3.627006442964139, + "tokens_seen": 1160577024 + }, + { + "epoch": 0.35, + "learning_rate": 0.0005239237713139418, + "loss": 1.4573, + "theoretical_loss": 3.626987634025479, + "tokens_seen": 1160642560 + }, + { + "epoch": 0.35, + "learning_rate": 0.0005239077231695085, + "loss": 1.3764, + "theoretical_loss": 3.626968826446196, + "tokens_seen": 1160708096 + }, + { + "epoch": 0.35, + "learning_rate": 0.0005238916750250752, + "loss": 1.3389, + "theoretical_loss": 3.626950020226116, + "tokens_seen": 1160773632 + }, + { + "epoch": 0.35, + "learning_rate": 0.0005238756268806419, + "loss": 1.3415, + "theoretical_loss": 3.6269312153650626, + "tokens_seen": 1160839168 + }, + { + "epoch": 0.35, + "learning_rate": 0.0005238595787362086, + "loss": 1.3668, + "theoretical_loss": 3.6269124118628615, + "tokens_seen": 1160904704 + }, + { + "epoch": 0.35, + "learning_rate": 0.0005238435305917753, + "loss": 1.4256, + "theoretical_loss": 3.626893609719338, + "tokens_seen": 1160970240 + }, + { + "epoch": 0.35, + "learning_rate": 0.000523827482447342, + "loss": 1.5172, + "theoretical_loss": 3.6268748089343177, + "tokens_seen": 1161035776 + }, + { + "epoch": 0.35, + "learning_rate": 0.0005238114343029087, + "loss": 1.2938, + "theoretical_loss": 3.6268560095076245, + "tokens_seen": 1161101312 + }, + { + "epoch": 0.35, + "learning_rate": 0.0005237953861584754, + "loss": 1.4694, + "theoretical_loss": 3.626837211439085, + "tokens_seen": 1161166848 + }, + { + "epoch": 0.35, + "learning_rate": 0.0005237793380140421, + "loss": 1.3285, + "theoretical_loss": 3.626818414728523, + "tokens_seen": 1161232384 + }, + { + "epoch": 0.35, + "learning_rate": 0.0005237632898696088, + "loss": 1.3497, + "theoretical_loss": 3.626799619375765, + "tokens_seen": 1161297920 + }, + { + "epoch": 0.35, + "learning_rate": 0.0005237472417251755, + "loss": 1.329, + "theoretical_loss": 3.626780825380636, + "tokens_seen": 1161363456 + }, + { + "epoch": 0.35, + "learning_rate": 0.0005237311935807422, + "loss": 1.3943, + "theoretical_loss": 3.6267620327429606, + "tokens_seen": 1161428992 + }, + { + "epoch": 0.35, + "learning_rate": 0.000523715145436309, + "loss": 1.4318, + "theoretical_loss": 3.626743241462565, + "tokens_seen": 1161494528 + }, + { + "epoch": 0.35, + "learning_rate": 0.0005236990972918757, + "loss": 1.5144, + "theoretical_loss": 3.6267244515392743, + "tokens_seen": 1161560064 + }, + { + "epoch": 0.35, + "learning_rate": 0.0005236830491474423, + "loss": 1.3909, + "theoretical_loss": 3.626705662972914, + "tokens_seen": 1161625600 + }, + { + "epoch": 0.35, + "learning_rate": 0.0005236670010030091, + "loss": 1.3489, + "theoretical_loss": 3.6266868757633093, + "tokens_seen": 1161691136 + }, + { + "epoch": 0.35, + "learning_rate": 0.0005236509528585757, + "loss": 1.4154, + "theoretical_loss": 3.626668089910286, + "tokens_seen": 1161756672 + }, + { + "epoch": 0.35, + "learning_rate": 0.0005236349047141425, + "loss": 1.4743, + "theoretical_loss": 3.62664930541367, + "tokens_seen": 1161822208 + }, + { + "epoch": 0.35, + "learning_rate": 0.0005236188565697092, + "loss": 1.4398, + "theoretical_loss": 3.6266305222732855, + "tokens_seen": 1161887744 + }, + { + "epoch": 0.35, + "learning_rate": 0.0005236028084252759, + "loss": 1.4217, + "theoretical_loss": 3.6266117404889595, + "tokens_seen": 1161953280 + }, + { + "epoch": 0.35, + "learning_rate": 0.0005235867602808426, + "loss": 1.4681, + "theoretical_loss": 3.626592960060517, + "tokens_seen": 1162018816 + }, + { + "epoch": 0.35, + "learning_rate": 0.0005235707121364093, + "loss": 1.3495, + "theoretical_loss": 3.6265741809877836, + "tokens_seen": 1162084352 + }, + { + "epoch": 0.35, + "learning_rate": 0.000523554663991976, + "loss": 1.3034, + "theoretical_loss": 3.6265554032705856, + "tokens_seen": 1162149888 + }, + { + "epoch": 0.35, + "learning_rate": 0.0005235386158475427, + "loss": 1.3203, + "theoretical_loss": 3.626536626908748, + "tokens_seen": 1162215424 + }, + { + "epoch": 0.35, + "learning_rate": 0.0005235225677031094, + "loss": 1.4169, + "theoretical_loss": 3.626517851902097, + "tokens_seen": 1162280960 + }, + { + "epoch": 0.35, + "learning_rate": 0.0005235065195586761, + "loss": 1.49, + "theoretical_loss": 3.6264990782504576, + "tokens_seen": 1162346496 + }, + { + "epoch": 0.35, + "learning_rate": 0.0005234904714142428, + "loss": 1.2023, + "theoretical_loss": 3.626480305953657, + "tokens_seen": 1162412032 + }, + { + "epoch": 0.35, + "learning_rate": 0.0005234744232698095, + "loss": 1.5601, + "theoretical_loss": 3.6264615350115195, + "tokens_seen": 1162477568 + }, + { + "epoch": 0.35, + "learning_rate": 0.0005234583751253762, + "loss": 1.3138, + "theoretical_loss": 3.6264427654238727, + "tokens_seen": 1162543104 + }, + { + "epoch": 0.35, + "learning_rate": 0.0005234423269809429, + "loss": 1.3342, + "theoretical_loss": 3.6264239971905408, + "tokens_seen": 1162608640 + }, + { + "epoch": 0.35, + "learning_rate": 0.0005234262788365096, + "loss": 1.3319, + "theoretical_loss": 3.6264052303113505, + "tokens_seen": 1162674176 + }, + { + "epoch": 0.35, + "learning_rate": 0.0005234102306920763, + "loss": 1.294, + "theoretical_loss": 3.626386464786128, + "tokens_seen": 1162739712 + }, + { + "epoch": 0.35, + "learning_rate": 0.000523394182547643, + "loss": 1.2992, + "theoretical_loss": 3.626367700614699, + "tokens_seen": 1162805248 + }, + { + "epoch": 0.35, + "learning_rate": 0.0005233781344032097, + "loss": 1.4773, + "theoretical_loss": 3.6263489377968896, + "tokens_seen": 1162870784 + }, + { + "epoch": 0.35, + "learning_rate": 0.0005233620862587764, + "loss": 1.4802, + "theoretical_loss": 3.6263301763325257, + "tokens_seen": 1162936320 + }, + { + "epoch": 0.35, + "learning_rate": 0.0005233460381143431, + "loss": 1.501, + "theoretical_loss": 3.6263114162214345, + "tokens_seen": 1163001856 + }, + { + "epoch": 0.35, + "learning_rate": 0.0005233299899699098, + "loss": 1.4045, + "theoretical_loss": 3.6262926574634404, + "tokens_seen": 1163067392 + }, + { + "epoch": 0.35, + "learning_rate": 0.0005233139418254765, + "loss": 1.4215, + "theoretical_loss": 3.6262739000583704, + "tokens_seen": 1163132928 + }, + { + "epoch": 0.35, + "learning_rate": 0.0005232978936810432, + "loss": 1.3518, + "theoretical_loss": 3.626255144006051, + "tokens_seen": 1163198464 + }, + { + "epoch": 0.35, + "objective/train/docs_used": 425912, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.403470516204834, + "objective/train/theoretical_loss": 3.6262363893063085, + "objective/train/tokens_used": 1183724000, + "theoretical_loss": 3.6262363893063085, + "tokens_seen": 1163264000 + }, + { + "epoch": 0.35, + "learning_rate": 0.0005232818455366099, + "loss": 1.3511, + "theoretical_loss": 3.6262363893063085, + "tokens_seen": 1163264000 + }, + { + "epoch": 0.35, + "learning_rate": 0.0005232657973921766, + "loss": 1.361, + "theoretical_loss": 3.626217635958968, + "tokens_seen": 1163329536 + }, + { + "epoch": 0.35, + "learning_rate": 0.0005232497492477433, + "loss": 1.4417, + "theoretical_loss": 3.6261988839638573, + "tokens_seen": 1163395072 + }, + { + "epoch": 0.35, + "learning_rate": 0.00052323370110331, + "loss": 1.3678, + "theoretical_loss": 3.626180133320802, + "tokens_seen": 1163460608 + }, + { + "epoch": 0.35, + "learning_rate": 0.0005232176529588767, + "loss": 1.3811, + "theoretical_loss": 3.6261613840296283, + "tokens_seen": 1163526144 + }, + { + "epoch": 0.35, + "learning_rate": 0.0005232016048144434, + "loss": 1.24, + "theoretical_loss": 3.6261426360901625, + "tokens_seen": 1163591680 + }, + { + "epoch": 0.35, + "learning_rate": 0.0005231855566700101, + "loss": 1.4102, + "theoretical_loss": 3.626123889502232, + "tokens_seen": 1163657216 + }, + { + "epoch": 0.35, + "learning_rate": 0.0005231695085255768, + "loss": 1.3608, + "theoretical_loss": 3.6261051442656624, + "tokens_seen": 1163722752 + }, + { + "epoch": 0.35, + "learning_rate": 0.0005231534603811435, + "loss": 1.4581, + "theoretical_loss": 3.62608640038028, + "tokens_seen": 1163788288 + }, + { + "epoch": 0.35, + "learning_rate": 0.0005231374122367102, + "loss": 1.3403, + "theoretical_loss": 3.6260676578459123, + "tokens_seen": 1163853824 + }, + { + "epoch": 0.35, + "learning_rate": 0.0005231213640922769, + "loss": 1.4609, + "theoretical_loss": 3.626048916662385, + "tokens_seen": 1163919360 + }, + { + "epoch": 0.35, + "learning_rate": 0.0005231053159478436, + "loss": 1.3882, + "theoretical_loss": 3.6260301768295253, + "tokens_seen": 1163984896 + }, + { + "epoch": 0.35, + "learning_rate": 0.0005230892678034103, + "loss": 1.3574, + "theoretical_loss": 3.6260114383471587, + "tokens_seen": 1164050432 + }, + { + "epoch": 0.35, + "learning_rate": 0.000523073219658977, + "loss": 1.4274, + "theoretical_loss": 3.625992701215113, + "tokens_seen": 1164115968 + }, + { + "epoch": 0.35, + "learning_rate": 0.0005230571715145437, + "loss": 1.2675, + "theoretical_loss": 3.625973965433215, + "tokens_seen": 1164181504 + }, + { + "epoch": 0.35, + "learning_rate": 0.0005230411233701104, + "loss": 1.3401, + "theoretical_loss": 3.6259552310012904, + "tokens_seen": 1164247040 + }, + { + "epoch": 0.35, + "learning_rate": 0.0005230250752256771, + "loss": 1.3805, + "theoretical_loss": 3.6259364979191666, + "tokens_seen": 1164312576 + }, + { + "epoch": 0.35, + "learning_rate": 0.0005230090270812438, + "loss": 1.3612, + "theoretical_loss": 3.62591776618667, + "tokens_seen": 1164378112 + }, + { + "epoch": 0.35, + "learning_rate": 0.0005229929789368104, + "loss": 1.2412, + "theoretical_loss": 3.625899035803628, + "tokens_seen": 1164443648 + }, + { + "epoch": 0.35, + "learning_rate": 0.0005229769307923772, + "loss": 1.3818, + "theoretical_loss": 3.625880306769867, + "tokens_seen": 1164509184 + }, + { + "epoch": 0.35, + "learning_rate": 0.0005229608826479438, + "loss": 1.3752, + "theoretical_loss": 3.625861579085214, + "tokens_seen": 1164574720 + }, + { + "epoch": 0.35, + "learning_rate": 0.0005229448345035106, + "loss": 1.4681, + "theoretical_loss": 3.625842852749496, + "tokens_seen": 1164640256 + }, + { + "epoch": 0.35, + "learning_rate": 0.0005229287863590772, + "loss": 1.2919, + "theoretical_loss": 3.62582412776254, + "tokens_seen": 1164705792 + }, + { + "epoch": 0.35, + "learning_rate": 0.000522912738214644, + "loss": 1.421, + "theoretical_loss": 3.625805404124172, + "tokens_seen": 1164771328 + }, + { + "epoch": 0.35, + "learning_rate": 0.0005228966900702106, + "loss": 1.466, + "theoretical_loss": 3.6257866818342204, + "tokens_seen": 1164836864 + }, + { + "epoch": 0.35, + "learning_rate": 0.0005228806419257773, + "loss": 1.3712, + "theoretical_loss": 3.6257679608925115, + "tokens_seen": 1164902400 + }, + { + "epoch": 0.35, + "learning_rate": 0.000522864593781344, + "loss": 1.4311, + "theoretical_loss": 3.6257492412988723, + "tokens_seen": 1164967936 + }, + { + "epoch": 0.35, + "learning_rate": 0.0005228485456369107, + "loss": 1.4313, + "theoretical_loss": 3.6257305230531305, + "tokens_seen": 1165033472 + }, + { + "epoch": 0.35, + "learning_rate": 0.0005228324974924774, + "loss": 1.378, + "theoretical_loss": 3.6257118061551132, + "tokens_seen": 1165099008 + }, + { + "epoch": 0.35, + "learning_rate": 0.0005228164493480441, + "loss": 1.3631, + "theoretical_loss": 3.6256930906046465, + "tokens_seen": 1165164544 + }, + { + "epoch": 0.35, + "learning_rate": 0.0005228004012036108, + "loss": 1.4805, + "theoretical_loss": 3.625674376401559, + "tokens_seen": 1165230080 + }, + { + "epoch": 0.35, + "learning_rate": 0.0005227843530591775, + "loss": 1.4494, + "theoretical_loss": 3.6256556635456763, + "tokens_seen": 1165295616 + }, + { + "epoch": 0.35, + "learning_rate": 0.0005227683049147442, + "loss": 1.3524, + "theoretical_loss": 3.6256369520368272, + "tokens_seen": 1165361152 + }, + { + "epoch": 0.35, + "learning_rate": 0.0005227522567703109, + "loss": 1.3558, + "theoretical_loss": 3.6256182418748377, + "tokens_seen": 1165426688 + }, + { + "epoch": 0.35, + "learning_rate": 0.0005227362086258776, + "loss": 1.4064, + "theoretical_loss": 3.6255995330595363, + "tokens_seen": 1165492224 + }, + { + "epoch": 0.35, + "learning_rate": 0.0005227201604814443, + "loss": 1.4715, + "theoretical_loss": 3.6255808255907502, + "tokens_seen": 1165557760 + }, + { + "epoch": 0.35, + "learning_rate": 0.000522704112337011, + "loss": 1.3988, + "theoretical_loss": 3.625562119468306, + "tokens_seen": 1165623296 + }, + { + "epoch": 0.35, + "learning_rate": 0.0005226880641925777, + "loss": 1.5134, + "theoretical_loss": 3.6255434146920313, + "tokens_seen": 1165688832 + }, + { + "epoch": 0.35, + "learning_rate": 0.0005226720160481444, + "loss": 1.4215, + "theoretical_loss": 3.6255247112617544, + "tokens_seen": 1165754368 + }, + { + "epoch": 0.35, + "learning_rate": 0.0005226559679037111, + "loss": 1.2968, + "theoretical_loss": 3.625506009177302, + "tokens_seen": 1165819904 + }, + { + "epoch": 0.35, + "learning_rate": 0.0005226399197592778, + "loss": 1.3486, + "theoretical_loss": 3.625487308438502, + "tokens_seen": 1165885440 + }, + { + "epoch": 0.35, + "learning_rate": 0.0005226238716148445, + "loss": 1.4321, + "theoretical_loss": 3.625468609045181, + "tokens_seen": 1165950976 + }, + { + "epoch": 0.35, + "learning_rate": 0.0005226078234704112, + "loss": 1.4645, + "theoretical_loss": 3.625449910997168, + "tokens_seen": 1166016512 + }, + { + "epoch": 0.35, + "learning_rate": 0.0005225917753259779, + "loss": 1.3431, + "theoretical_loss": 3.6254312142942897, + "tokens_seen": 1166082048 + }, + { + "epoch": 0.35, + "learning_rate": 0.0005225757271815446, + "loss": 1.3245, + "theoretical_loss": 3.6254125189363737, + "tokens_seen": 1166147584 + }, + { + "epoch": 0.35, + "learning_rate": 0.0005225596790371113, + "loss": 1.3724, + "theoretical_loss": 3.6253938249232482, + "tokens_seen": 1166213120 + }, + { + "epoch": 0.35, + "learning_rate": 0.000522543630892678, + "loss": 1.3295, + "theoretical_loss": 3.6253751322547405, + "tokens_seen": 1166278656 + }, + { + "epoch": 0.35, + "learning_rate": 0.0005225275827482447, + "loss": 1.4313, + "theoretical_loss": 3.6253564409306787, + "tokens_seen": 1166344192 + }, + { + "epoch": 0.35, + "learning_rate": 0.0005225115346038114, + "loss": 1.5147, + "theoretical_loss": 3.62533775095089, + "tokens_seen": 1166409728 + }, + { + "epoch": 0.35, + "learning_rate": 0.0005224954864593782, + "loss": 1.4122, + "theoretical_loss": 3.625319062315203, + "tokens_seen": 1166475264 + }, + { + "epoch": 0.35, + "objective/train/docs_used": 427016, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.4632213115692139, + "objective/train/theoretical_loss": 3.625300375023445, + "objective/train/tokens_used": 1187000800, + "theoretical_loss": 3.625300375023445, + "tokens_seen": 1166540800 + }, + { + "epoch": 0.35, + "learning_rate": 0.0005224794383149448, + "loss": 1.4425, + "theoretical_loss": 3.625300375023445, + "tokens_seen": 1166540800 + }, + { + "epoch": 0.35, + "learning_rate": 0.0005224633901705116, + "loss": 1.361, + "theoretical_loss": 3.625281689075444, + "tokens_seen": 1166606336 + }, + { + "epoch": 0.35, + "learning_rate": 0.0005224473420260782, + "loss": 1.4287, + "theoretical_loss": 3.6252630044710275, + "tokens_seen": 1166671872 + }, + { + "epoch": 0.35, + "learning_rate": 0.000522431293881645, + "loss": 1.5014, + "theoretical_loss": 3.625244321210024, + "tokens_seen": 1166737408 + }, + { + "epoch": 0.35, + "learning_rate": 0.0005224152457372116, + "loss": 1.3748, + "theoretical_loss": 3.625225639292261, + "tokens_seen": 1166802944 + }, + { + "epoch": 0.35, + "learning_rate": 0.0005223991975927784, + "loss": 1.3992, + "theoretical_loss": 3.6252069587175666, + "tokens_seen": 1166868480 + }, + { + "epoch": 0.35, + "learning_rate": 0.000522383149448345, + "loss": 1.3784, + "theoretical_loss": 3.6251882794857693, + "tokens_seen": 1166934016 + }, + { + "epoch": 0.35, + "learning_rate": 0.0005223671013039118, + "loss": 1.4483, + "theoretical_loss": 3.6251696015966965, + "tokens_seen": 1166999552 + }, + { + "epoch": 0.35, + "learning_rate": 0.0005223510531594784, + "loss": 1.475, + "theoretical_loss": 3.6251509250501766, + "tokens_seen": 1167065088 + }, + { + "epoch": 0.35, + "learning_rate": 0.0005223350050150452, + "loss": 1.4377, + "theoretical_loss": 3.625132249846038, + "tokens_seen": 1167130624 + }, + { + "epoch": 0.35, + "learning_rate": 0.0005223189568706118, + "loss": 1.4637, + "theoretical_loss": 3.625113575984108, + "tokens_seen": 1167196160 + }, + { + "epoch": 0.35, + "learning_rate": 0.0005223029087261786, + "loss": 1.3349, + "theoretical_loss": 3.6250949034642153, + "tokens_seen": 1167261696 + }, + { + "epoch": 0.35, + "learning_rate": 0.0005222868605817452, + "loss": 1.2963, + "theoretical_loss": 3.6250762322861885, + "tokens_seen": 1167327232 + }, + { + "epoch": 0.35, + "learning_rate": 0.000522270812437312, + "loss": 1.4009, + "theoretical_loss": 3.6250575624498556, + "tokens_seen": 1167392768 + }, + { + "epoch": 0.35, + "learning_rate": 0.0005222547642928787, + "loss": 1.4087, + "theoretical_loss": 3.625038893955044, + "tokens_seen": 1167458304 + }, + { + "epoch": 0.35, + "learning_rate": 0.0005222387161484454, + "loss": 1.433, + "theoretical_loss": 3.625020226801584, + "tokens_seen": 1167523840 + }, + { + "epoch": 0.35, + "learning_rate": 0.000522222668004012, + "loss": 1.4288, + "theoretical_loss": 3.6250015609893014, + "tokens_seen": 1167589376 + }, + { + "epoch": 0.35, + "learning_rate": 0.0005222066198595788, + "loss": 1.4152, + "theoretical_loss": 3.624982896518026, + "tokens_seen": 1167654912 + }, + { + "epoch": 0.35, + "learning_rate": 0.0005221905717151455, + "loss": 1.3413, + "theoretical_loss": 3.624964233387587, + "tokens_seen": 1167720448 + }, + { + "epoch": 0.35, + "learning_rate": 0.0005221745235707122, + "loss": 1.3455, + "theoretical_loss": 3.624945571597811, + "tokens_seen": 1167785984 + }, + { + "epoch": 0.35, + "learning_rate": 0.0005221584754262789, + "loss": 1.4012, + "theoretical_loss": 3.6249269111485276, + "tokens_seen": 1167851520 + }, + { + "epoch": 0.35, + "learning_rate": 0.0005221424272818456, + "loss": 1.3187, + "theoretical_loss": 3.624908252039565, + "tokens_seen": 1167917056 + }, + { + "epoch": 0.35, + "learning_rate": 0.0005221263791374123, + "loss": 1.4028, + "theoretical_loss": 3.6248895942707517, + "tokens_seen": 1167982592 + }, + { + "epoch": 0.35, + "learning_rate": 0.000522110330992979, + "loss": 1.401, + "theoretical_loss": 3.6248709378419166, + "tokens_seen": 1168048128 + }, + { + "epoch": 0.35, + "learning_rate": 0.0005220942828485457, + "loss": 1.4575, + "theoretical_loss": 3.6248522827528875, + "tokens_seen": 1168113664 + }, + { + "epoch": 0.35, + "learning_rate": 0.0005220782347041124, + "loss": 1.44, + "theoretical_loss": 3.624833629003494, + "tokens_seen": 1168179200 + }, + { + "epoch": 0.35, + "learning_rate": 0.0005220621865596791, + "loss": 1.4069, + "theoretical_loss": 3.624814976593564, + "tokens_seen": 1168244736 + }, + { + "epoch": 0.35, + "learning_rate": 0.0005220461384152458, + "loss": 1.3849, + "theoretical_loss": 3.6247963255229267, + "tokens_seen": 1168310272 + }, + { + "epoch": 0.35, + "learning_rate": 0.0005220300902708125, + "loss": 1.4512, + "theoretical_loss": 3.6247776757914103, + "tokens_seen": 1168375808 + }, + { + "epoch": 0.35, + "learning_rate": 0.0005220140421263792, + "loss": 1.3527, + "theoretical_loss": 3.624759027398844, + "tokens_seen": 1168441344 + }, + { + "epoch": 0.35, + "learning_rate": 0.0005219979939819459, + "loss": 1.2836, + "theoretical_loss": 3.624740380345057, + "tokens_seen": 1168506880 + }, + { + "epoch": 0.35, + "learning_rate": 0.0005219819458375126, + "loss": 1.4448, + "theoretical_loss": 3.6247217346298766, + "tokens_seen": 1168572416 + }, + { + "epoch": 0.35, + "learning_rate": 0.0005219658976930793, + "loss": 1.4032, + "theoretical_loss": 3.624703090253133, + "tokens_seen": 1168637952 + }, + { + "epoch": 0.35, + "learning_rate": 0.000521949849548646, + "loss": 1.3572, + "theoretical_loss": 3.6246844472146544, + "tokens_seen": 1168703488 + }, + { + "epoch": 0.35, + "learning_rate": 0.0005219338014042127, + "loss": 1.3433, + "theoretical_loss": 3.62466580551427, + "tokens_seen": 1168769024 + }, + { + "epoch": 0.35, + "learning_rate": 0.0005219177532597794, + "loss": 1.2914, + "theoretical_loss": 3.624647165151809, + "tokens_seen": 1168834560 + }, + { + "epoch": 0.35, + "learning_rate": 0.0005219017051153461, + "loss": 1.4824, + "theoretical_loss": 3.6246285261270996, + "tokens_seen": 1168900096 + }, + { + "epoch": 0.35, + "learning_rate": 0.0005218856569709128, + "loss": 1.3664, + "theoretical_loss": 3.624609888439972, + "tokens_seen": 1168965632 + }, + { + "epoch": 0.35, + "learning_rate": 0.0005218696088264795, + "loss": 1.3427, + "theoretical_loss": 3.6245912520902537, + "tokens_seen": 1169031168 + }, + { + "epoch": 0.35, + "learning_rate": 0.0005218535606820462, + "loss": 1.5281, + "theoretical_loss": 3.624572617077775, + "tokens_seen": 1169096704 + }, + { + "epoch": 0.35, + "learning_rate": 0.0005218375125376129, + "loss": 1.323, + "theoretical_loss": 3.624553983402364, + "tokens_seen": 1169162240 + }, + { + "epoch": 0.35, + "learning_rate": 0.0005218214643931796, + "loss": 1.3573, + "theoretical_loss": 3.6245353510638507, + "tokens_seen": 1169227776 + }, + { + "epoch": 0.35, + "learning_rate": 0.0005218054162487463, + "loss": 1.2819, + "theoretical_loss": 3.624516720062064, + "tokens_seen": 1169293312 + }, + { + "epoch": 0.35, + "learning_rate": 0.000521789368104313, + "loss": 1.3672, + "theoretical_loss": 3.624498090396833, + "tokens_seen": 1169358848 + }, + { + "epoch": 0.35, + "learning_rate": 0.0005217733199598797, + "loss": 1.3831, + "theoretical_loss": 3.6244794620679865, + "tokens_seen": 1169424384 + }, + { + "epoch": 0.35, + "learning_rate": 0.0005217572718154464, + "loss": 1.4745, + "theoretical_loss": 3.6244608350753547, + "tokens_seen": 1169489920 + }, + { + "epoch": 0.35, + "learning_rate": 0.0005217412236710131, + "loss": 1.3221, + "theoretical_loss": 3.6244422094187665, + "tokens_seen": 1169555456 + }, + { + "epoch": 0.35, + "learning_rate": 0.0005217251755265798, + "loss": 1.3672, + "theoretical_loss": 3.6244235850980506, + "tokens_seen": 1169620992 + }, + { + "epoch": 0.35, + "learning_rate": 0.0005217091273821465, + "loss": 1.4632, + "theoretical_loss": 3.624404962113037, + "tokens_seen": 1169686528 + }, + { + "epoch": 0.35, + "learning_rate": 0.0005216930792377132, + "loss": 1.4881, + "theoretical_loss": 3.6243863404635546, + "tokens_seen": 1169752064 + }, + { + "epoch": 0.35, + "objective/train/docs_used": 428177, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.427170753479004, + "objective/train/theoretical_loss": 3.6243677201494338, + "objective/train/tokens_used": 1190277600, + "theoretical_loss": 3.6243677201494338, + "tokens_seen": 1169817600 + }, + { + "epoch": 0.35, + "learning_rate": 0.0005216770310932799, + "loss": 1.4668, + "theoretical_loss": 3.6243677201494338, + "tokens_seen": 1169817600 + }, + { + "epoch": 0.35, + "learning_rate": 0.0005216609829488466, + "loss": 1.486, + "theoretical_loss": 3.6243491011705027, + "tokens_seen": 1169883136 + }, + { + "epoch": 0.35, + "learning_rate": 0.0005216449348044133, + "loss": 1.3811, + "theoretical_loss": 3.6243304835265917, + "tokens_seen": 1169948672 + }, + { + "epoch": 0.35, + "learning_rate": 0.00052162888665998, + "loss": 1.3167, + "theoretical_loss": 3.62431186721753, + "tokens_seen": 1170014208 + }, + { + "epoch": 0.35, + "learning_rate": 0.0005216128385155467, + "loss": 1.4654, + "theoretical_loss": 3.624293252243147, + "tokens_seen": 1170079744 + }, + { + "epoch": 0.35, + "learning_rate": 0.0005215967903711133, + "loss": 1.4948, + "theoretical_loss": 3.6242746386032723, + "tokens_seen": 1170145280 + }, + { + "epoch": 0.35, + "learning_rate": 0.0005215807422266801, + "loss": 1.3472, + "theoretical_loss": 3.624256026297736, + "tokens_seen": 1170210816 + }, + { + "epoch": 0.35, + "learning_rate": 0.0005215646940822467, + "loss": 1.386, + "theoretical_loss": 3.624237415326367, + "tokens_seen": 1170276352 + }, + { + "epoch": 0.35, + "learning_rate": 0.0005215486459378135, + "loss": 1.2783, + "theoretical_loss": 3.6242188056889955, + "tokens_seen": 1170341888 + }, + { + "epoch": 0.35, + "learning_rate": 0.0005215325977933801, + "loss": 1.3454, + "theoretical_loss": 3.6242001973854507, + "tokens_seen": 1170407424 + }, + { + "epoch": 0.35, + "learning_rate": 0.0005215165496489469, + "loss": 1.4129, + "theoretical_loss": 3.6241815904155628, + "tokens_seen": 1170472960 + }, + { + "epoch": 0.35, + "learning_rate": 0.0005215005015045135, + "loss": 1.4225, + "theoretical_loss": 3.6241629847791614, + "tokens_seen": 1170538496 + }, + { + "epoch": 0.35, + "learning_rate": 0.0005214844533600803, + "loss": 1.4839, + "theoretical_loss": 3.6241443804760762, + "tokens_seen": 1170604032 + }, + { + "epoch": 0.35, + "learning_rate": 0.0005214684052156469, + "loss": 1.3999, + "theoretical_loss": 3.624125777506137, + "tokens_seen": 1170669568 + }, + { + "epoch": 0.35, + "learning_rate": 0.0005214523570712136, + "loss": 1.4003, + "theoretical_loss": 3.624107175869174, + "tokens_seen": 1170735104 + }, + { + "epoch": 0.35, + "learning_rate": 0.0005214363089267803, + "loss": 1.3605, + "theoretical_loss": 3.6240885755650165, + "tokens_seen": 1170800640 + }, + { + "epoch": 0.35, + "learning_rate": 0.000521420260782347, + "loss": 1.3943, + "theoretical_loss": 3.6240699765934945, + "tokens_seen": 1170866176 + }, + { + "epoch": 0.35, + "learning_rate": 0.0005214042126379137, + "loss": 1.4669, + "theoretical_loss": 3.624051378954438, + "tokens_seen": 1170931712 + }, + { + "epoch": 0.35, + "learning_rate": 0.0005213881644934804, + "loss": 1.4642, + "theoretical_loss": 3.624032782647678, + "tokens_seen": 1170997248 + }, + { + "epoch": 0.35, + "learning_rate": 0.0005213721163490471, + "loss": 1.423, + "theoretical_loss": 3.6240141876730423, + "tokens_seen": 1171062784 + }, + { + "epoch": 0.35, + "learning_rate": 0.0005213560682046138, + "loss": 1.4605, + "theoretical_loss": 3.6239955940303634, + "tokens_seen": 1171128320 + }, + { + "epoch": 0.35, + "learning_rate": 0.0005213400200601805, + "loss": 1.3072, + "theoretical_loss": 3.6239770017194695, + "tokens_seen": 1171193856 + }, + { + "epoch": 0.35, + "learning_rate": 0.0005213239719157472, + "loss": 1.5084, + "theoretical_loss": 3.6239584107401916, + "tokens_seen": 1171259392 + }, + { + "epoch": 0.35, + "learning_rate": 0.0005213079237713139, + "loss": 1.3902, + "theoretical_loss": 3.6239398210923595, + "tokens_seen": 1171324928 + }, + { + "epoch": 0.35, + "learning_rate": 0.0005212918756268807, + "loss": 1.4391, + "theoretical_loss": 3.6239212327758037, + "tokens_seen": 1171390464 + }, + { + "epoch": 0.35, + "learning_rate": 0.0005212758274824473, + "loss": 1.3281, + "theoretical_loss": 3.623902645790354, + "tokens_seen": 1171456000 + }, + { + "epoch": 0.36, + "learning_rate": 0.0005212597793380141, + "loss": 1.3947, + "theoretical_loss": 3.6238840601358406, + "tokens_seen": 1171521536 + }, + { + "epoch": 0.36, + "learning_rate": 0.0005212437311935807, + "loss": 1.4221, + "theoretical_loss": 3.6238654758120945, + "tokens_seen": 1171587072 + }, + { + "epoch": 0.36, + "learning_rate": 0.0005212276830491475, + "loss": 1.4045, + "theoretical_loss": 3.6238468928189453, + "tokens_seen": 1171652608 + }, + { + "epoch": 0.36, + "learning_rate": 0.0005212116349047141, + "loss": 1.4132, + "theoretical_loss": 3.623828311156223, + "tokens_seen": 1171718144 + }, + { + "epoch": 0.36, + "learning_rate": 0.0005211955867602809, + "loss": 1.4143, + "theoretical_loss": 3.6238097308237585, + "tokens_seen": 1171783680 + }, + { + "epoch": 0.36, + "learning_rate": 0.0005211795386158475, + "loss": 1.3721, + "theoretical_loss": 3.623791151821382, + "tokens_seen": 1171849216 + }, + { + "epoch": 0.36, + "learning_rate": 0.0005211634904714143, + "loss": 1.4751, + "theoretical_loss": 3.623772574148924, + "tokens_seen": 1171914752 + }, + { + "epoch": 0.36, + "learning_rate": 0.0005211474423269809, + "loss": 1.3163, + "theoretical_loss": 3.623753997806215, + "tokens_seen": 1171980288 + }, + { + "epoch": 0.36, + "learning_rate": 0.0005211313941825477, + "loss": 1.3596, + "theoretical_loss": 3.6237354227930854, + "tokens_seen": 1172045824 + }, + { + "epoch": 0.36, + "learning_rate": 0.0005211153460381143, + "loss": 1.4726, + "theoretical_loss": 3.6237168491093654, + "tokens_seen": 1172111360 + }, + { + "epoch": 0.36, + "learning_rate": 0.0005210992978936811, + "loss": 1.253, + "theoretical_loss": 3.6236982767548858, + "tokens_seen": 1172176896 + }, + { + "epoch": 0.36, + "learning_rate": 0.0005210832497492477, + "loss": 1.3588, + "theoretical_loss": 3.6236797057294776, + "tokens_seen": 1172242432 + }, + { + "epoch": 0.36, + "learning_rate": 0.0005210672016048145, + "loss": 1.3112, + "theoretical_loss": 3.623661136032971, + "tokens_seen": 1172307968 + }, + { + "epoch": 0.36, + "learning_rate": 0.0005210511534603811, + "loss": 1.1879, + "theoretical_loss": 3.623642567665196, + "tokens_seen": 1172373504 + }, + { + "epoch": 0.36, + "learning_rate": 0.0005210351053159479, + "loss": 1.3926, + "theoretical_loss": 3.623624000625984, + "tokens_seen": 1172439040 + }, + { + "epoch": 0.36, + "learning_rate": 0.0005210190571715145, + "loss": 1.3517, + "theoretical_loss": 3.6236054349151656, + "tokens_seen": 1172504576 + }, + { + "epoch": 0.36, + "learning_rate": 0.0005210030090270813, + "loss": 1.3881, + "theoretical_loss": 3.623586870532572, + "tokens_seen": 1172570112 + }, + { + "epoch": 0.36, + "learning_rate": 0.000520986960882648, + "loss": 1.3949, + "theoretical_loss": 3.623568307478032, + "tokens_seen": 1172635648 + }, + { + "epoch": 0.36, + "learning_rate": 0.0005209709127382147, + "loss": 1.303, + "theoretical_loss": 3.623549745751379, + "tokens_seen": 1172701184 + }, + { + "epoch": 0.36, + "learning_rate": 0.0005209548645937813, + "loss": 1.489, + "theoretical_loss": 3.623531185352442, + "tokens_seen": 1172766720 + }, + { + "epoch": 0.36, + "learning_rate": 0.0005209388164493481, + "loss": 1.4253, + "theoretical_loss": 3.6235126262810526, + "tokens_seen": 1172832256 + }, + { + "epoch": 0.36, + "learning_rate": 0.0005209227683049147, + "loss": 1.3767, + "theoretical_loss": 3.623494068537042, + "tokens_seen": 1172897792 + }, + { + "epoch": 0.36, + "learning_rate": 0.0005209067201604815, + "loss": 1.5056, + "theoretical_loss": 3.6234755121202396, + "tokens_seen": 1172963328 + }, + { + "epoch": 0.36, + "learning_rate": 0.0005208906720160482, + "loss": 1.4155, + "theoretical_loss": 3.623456957030478, + "tokens_seen": 1173028864 + }, + { + "epoch": 0.36, + "objective/train/docs_used": 429267, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.3980040550231934, + "objective/train/theoretical_loss": 3.623438403267587, + "objective/train/tokens_used": 1193554400, + "theoretical_loss": 3.623438403267587, + "tokens_seen": 1173094400 + }, + { + "epoch": 0.36, + "learning_rate": 0.0005208746238716149, + "loss": 1.3172, + "theoretical_loss": 3.623438403267587, + "tokens_seen": 1173094400 + }, + { + "epoch": 0.36, + "learning_rate": 0.0005208585757271816, + "loss": 1.3654, + "theoretical_loss": 3.6234198508313984, + "tokens_seen": 1173159936 + }, + { + "epoch": 0.36, + "learning_rate": 0.0005208425275827483, + "loss": 1.4291, + "theoretical_loss": 3.623401299721743, + "tokens_seen": 1173225472 + }, + { + "epoch": 0.36, + "learning_rate": 0.000520826479438315, + "loss": 1.4125, + "theoretical_loss": 3.6233827499384517, + "tokens_seen": 1173291008 + }, + { + "epoch": 0.36, + "learning_rate": 0.0005208104312938817, + "loss": 1.4403, + "theoretical_loss": 3.6233642014813556, + "tokens_seen": 1173356544 + }, + { + "epoch": 0.36, + "learning_rate": 0.0005207943831494484, + "loss": 1.4032, + "theoretical_loss": 3.6233456543502855, + "tokens_seen": 1173422080 + }, + { + "epoch": 0.36, + "learning_rate": 0.0005207783350050151, + "loss": 1.3838, + "theoretical_loss": 3.623327108545074, + "tokens_seen": 1173487616 + }, + { + "epoch": 0.36, + "learning_rate": 0.0005207622868605818, + "loss": 1.3173, + "theoretical_loss": 3.6233085640655505, + "tokens_seen": 1173553152 + }, + { + "epoch": 0.36, + "learning_rate": 0.0005207462387161485, + "loss": 1.3436, + "theoretical_loss": 3.623290020911547, + "tokens_seen": 1173618688 + }, + { + "epoch": 0.36, + "learning_rate": 0.0005207301905717152, + "loss": 1.3918, + "theoretical_loss": 3.6232714790828946, + "tokens_seen": 1173684224 + }, + { + "epoch": 0.36, + "learning_rate": 0.0005207141424272819, + "loss": 1.4447, + "theoretical_loss": 3.6232529385794248, + "tokens_seen": 1173749760 + }, + { + "epoch": 0.36, + "learning_rate": 0.0005206980942828486, + "loss": 1.4853, + "theoretical_loss": 3.6232343994009684, + "tokens_seen": 1173815296 + }, + { + "epoch": 0.36, + "learning_rate": 0.0005206820461384153, + "loss": 1.435, + "theoretical_loss": 3.623215861547358, + "tokens_seen": 1173880832 + }, + { + "epoch": 0.36, + "learning_rate": 0.000520665997993982, + "loss": 1.4089, + "theoretical_loss": 3.623197325018423, + "tokens_seen": 1173946368 + }, + { + "epoch": 0.36, + "learning_rate": 0.0005206499498495487, + "loss": 1.3135, + "theoretical_loss": 3.6231787898139967, + "tokens_seen": 1174011904 + }, + { + "epoch": 0.36, + "learning_rate": 0.0005206339017051154, + "loss": 1.4654, + "theoretical_loss": 3.623160255933909, + "tokens_seen": 1174077440 + }, + { + "epoch": 0.36, + "learning_rate": 0.0005206178535606821, + "loss": 1.4097, + "theoretical_loss": 3.6231417233779926, + "tokens_seen": 1174142976 + }, + { + "epoch": 0.36, + "learning_rate": 0.0005206018054162488, + "loss": 1.3317, + "theoretical_loss": 3.6231231921460783, + "tokens_seen": 1174208512 + }, + { + "epoch": 0.36, + "learning_rate": 0.0005205857572718155, + "loss": 1.2995, + "theoretical_loss": 3.6231046622379974, + "tokens_seen": 1174274048 + }, + { + "epoch": 0.36, + "learning_rate": 0.0005205697091273822, + "loss": 1.3613, + "theoretical_loss": 3.6230861336535822, + "tokens_seen": 1174339584 + }, + { + "epoch": 0.36, + "learning_rate": 0.0005205536609829489, + "loss": 1.3449, + "theoretical_loss": 3.623067606392664, + "tokens_seen": 1174405120 + }, + { + "epoch": 0.36, + "learning_rate": 0.0005205376128385156, + "loss": 1.3088, + "theoretical_loss": 3.6230490804550737, + "tokens_seen": 1174470656 + }, + { + "epoch": 0.36, + "learning_rate": 0.0005205215646940823, + "loss": 1.3458, + "theoretical_loss": 3.6230305558406446, + "tokens_seen": 1174536192 + }, + { + "epoch": 0.36, + "learning_rate": 0.000520505516549649, + "loss": 1.4672, + "theoretical_loss": 3.623012032549206, + "tokens_seen": 1174601728 + }, + { + "epoch": 0.36, + "learning_rate": 0.0005204894684052157, + "loss": 1.4542, + "theoretical_loss": 3.622993510580592, + "tokens_seen": 1174667264 + }, + { + "epoch": 0.36, + "learning_rate": 0.0005204734202607824, + "loss": 1.4906, + "theoretical_loss": 3.622974989934633, + "tokens_seen": 1174732800 + }, + { + "epoch": 0.36, + "learning_rate": 0.0005204573721163491, + "loss": 1.4649, + "theoretical_loss": 3.622956470611161, + "tokens_seen": 1174798336 + }, + { + "epoch": 0.36, + "learning_rate": 0.0005204413239719158, + "loss": 1.3595, + "theoretical_loss": 3.622937952610008, + "tokens_seen": 1174863872 + }, + { + "epoch": 0.36, + "learning_rate": 0.0005204252758274825, + "loss": 1.3251, + "theoretical_loss": 3.6229194359310055, + "tokens_seen": 1174929408 + }, + { + "epoch": 0.36, + "learning_rate": 0.0005204092276830492, + "loss": 1.2405, + "theoretical_loss": 3.6229009205739855, + "tokens_seen": 1174994944 + }, + { + "epoch": 0.36, + "learning_rate": 0.0005203931795386159, + "loss": 1.3938, + "theoretical_loss": 3.62288240653878, + "tokens_seen": 1175060480 + }, + { + "epoch": 0.36, + "learning_rate": 0.0005203771313941826, + "loss": 1.4431, + "theoretical_loss": 3.6228638938252216, + "tokens_seen": 1175126016 + }, + { + "epoch": 0.36, + "learning_rate": 0.0005203610832497493, + "loss": 1.407, + "theoretical_loss": 3.6228453824331406, + "tokens_seen": 1175191552 + }, + { + "epoch": 0.36, + "learning_rate": 0.000520345035105316, + "loss": 1.3242, + "theoretical_loss": 3.6228268723623698, + "tokens_seen": 1175257088 + }, + { + "epoch": 0.36, + "learning_rate": 0.0005203289869608827, + "loss": 1.4069, + "theoretical_loss": 3.6228083636127417, + "tokens_seen": 1175322624 + }, + { + "epoch": 0.36, + "learning_rate": 0.0005203129388164494, + "loss": 1.4348, + "theoretical_loss": 3.622789856184088, + "tokens_seen": 1175388160 + }, + { + "epoch": 0.36, + "learning_rate": 0.0005202968906720161, + "loss": 1.4501, + "theoretical_loss": 3.6227713500762406, + "tokens_seen": 1175453696 + }, + { + "epoch": 0.36, + "learning_rate": 0.0005202808425275828, + "loss": 1.4077, + "theoretical_loss": 3.622752845289032, + "tokens_seen": 1175519232 + }, + { + "epoch": 0.36, + "learning_rate": 0.0005202647943831495, + "loss": 1.3081, + "theoretical_loss": 3.6227343418222935, + "tokens_seen": 1175584768 + }, + { + "epoch": 0.36, + "learning_rate": 0.0005202487462387162, + "loss": 1.4638, + "theoretical_loss": 3.622715839675858, + "tokens_seen": 1175650304 + }, + { + "epoch": 0.36, + "learning_rate": 0.0005202326980942829, + "loss": 1.3016, + "theoretical_loss": 3.622697338849558, + "tokens_seen": 1175715840 + }, + { + "epoch": 0.36, + "learning_rate": 0.0005202166499498496, + "loss": 1.5403, + "theoretical_loss": 3.6226788393432248, + "tokens_seen": 1175781376 + }, + { + "epoch": 0.36, + "learning_rate": 0.0005202006018054162, + "loss": 1.2349, + "theoretical_loss": 3.6226603411566916, + "tokens_seen": 1175846912 + }, + { + "epoch": 0.36, + "learning_rate": 0.000520184553660983, + "loss": 1.4315, + "theoretical_loss": 3.62264184428979, + "tokens_seen": 1175912448 + }, + { + "epoch": 0.36, + "learning_rate": 0.0005201685055165496, + "loss": 1.2538, + "theoretical_loss": 3.6226233487423523, + "tokens_seen": 1175977984 + }, + { + "epoch": 0.36, + "learning_rate": 0.0005201524573721164, + "loss": 1.3741, + "theoretical_loss": 3.6226048545142113, + "tokens_seen": 1176043520 + }, + { + "epoch": 0.36, + "learning_rate": 0.000520136409227683, + "loss": 1.458, + "theoretical_loss": 3.622586361605199, + "tokens_seen": 1176109056 + }, + { + "epoch": 0.36, + "learning_rate": 0.0005201203610832498, + "loss": 1.4493, + "theoretical_loss": 3.6225678700151485, + "tokens_seen": 1176174592 + }, + { + "epoch": 0.36, + "learning_rate": 0.0005201043129388164, + "loss": 1.3066, + "theoretical_loss": 3.6225493797438917, + "tokens_seen": 1176240128 + }, + { + "epoch": 0.36, + "learning_rate": 0.0005200882647943832, + "loss": 1.41, + "theoretical_loss": 3.6225308907912606, + "tokens_seen": 1176305664 + }, + { + "epoch": 0.36, + "objective/train/docs_used": 430645, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.5048646926879883, + "objective/train/theoretical_loss": 3.6225124031570886, + "objective/train/tokens_used": 1196831200, + "theoretical_loss": 3.6225124031570886, + "tokens_seen": 1176371200 + }, + { + "epoch": 0.36, + "learning_rate": 0.0005200722166499498, + "loss": 1.4016, + "theoretical_loss": 3.6225124031570886, + "tokens_seen": 1176371200 + }, + { + "epoch": 0.36, + "learning_rate": 0.0005200561685055166, + "loss": 1.29, + "theoretical_loss": 3.622493916841208, + "tokens_seen": 1176436736 + }, + { + "epoch": 0.36, + "learning_rate": 0.0005200401203610832, + "loss": 1.4499, + "theoretical_loss": 3.6224754318434504, + "tokens_seen": 1176502272 + }, + { + "epoch": 0.36, + "learning_rate": 0.00052002407221665, + "loss": 1.4913, + "theoretical_loss": 3.62245694816365, + "tokens_seen": 1176567808 + }, + { + "epoch": 0.36, + "learning_rate": 0.0005200080240722166, + "loss": 1.4736, + "theoretical_loss": 3.622438465801638, + "tokens_seen": 1176633344 + }, + { + "epoch": 0.36, + "learning_rate": 0.0005199919759277834, + "loss": 1.4217, + "theoretical_loss": 3.622419984757248, + "tokens_seen": 1176698880 + }, + { + "epoch": 0.36, + "learning_rate": 0.00051997592778335, + "loss": 1.4877, + "theoretical_loss": 3.6224015050303127, + "tokens_seen": 1176764416 + }, + { + "epoch": 0.36, + "learning_rate": 0.0005199598796389168, + "loss": 1.4204, + "theoretical_loss": 3.6223830266206645, + "tokens_seen": 1176829952 + }, + { + "epoch": 0.36, + "learning_rate": 0.0005199438314944834, + "loss": 1.4139, + "theoretical_loss": 3.622364549528136, + "tokens_seen": 1176895488 + }, + { + "epoch": 0.36, + "learning_rate": 0.0005199277833500502, + "loss": 1.3323, + "theoretical_loss": 3.62234607375256, + "tokens_seen": 1176961024 + }, + { + "epoch": 0.36, + "learning_rate": 0.0005199117352056168, + "loss": 1.3811, + "theoretical_loss": 3.6223275992937696, + "tokens_seen": 1177026560 + }, + { + "epoch": 0.36, + "learning_rate": 0.0005198956870611836, + "loss": 1.4749, + "theoretical_loss": 3.6223091261515976, + "tokens_seen": 1177092096 + }, + { + "epoch": 0.36, + "learning_rate": 0.0005198796389167502, + "loss": 1.5001, + "theoretical_loss": 3.6222906543258766, + "tokens_seen": 1177157632 + }, + { + "epoch": 0.36, + "learning_rate": 0.000519863590772317, + "loss": 1.2719, + "theoretical_loss": 3.6222721838164396, + "tokens_seen": 1177223168 + }, + { + "epoch": 0.36, + "learning_rate": 0.0005198475426278836, + "loss": 1.4551, + "theoretical_loss": 3.6222537146231195, + "tokens_seen": 1177288704 + }, + { + "epoch": 0.36, + "learning_rate": 0.0005198314944834504, + "loss": 1.4136, + "theoretical_loss": 3.62223524674575, + "tokens_seen": 1177354240 + }, + { + "epoch": 0.36, + "learning_rate": 0.000519815446339017, + "loss": 1.291, + "theoretical_loss": 3.622216780184163, + "tokens_seen": 1177419776 + }, + { + "epoch": 0.36, + "learning_rate": 0.0005197993981945838, + "loss": 1.3697, + "theoretical_loss": 3.622198314938192, + "tokens_seen": 1177485312 + }, + { + "epoch": 0.36, + "learning_rate": 0.0005197833500501504, + "loss": 1.4641, + "theoretical_loss": 3.6221798510076706, + "tokens_seen": 1177550848 + }, + { + "epoch": 0.36, + "learning_rate": 0.0005197673019057172, + "loss": 1.4187, + "theoretical_loss": 3.6221613883924304, + "tokens_seen": 1177616384 + }, + { + "epoch": 0.36, + "learning_rate": 0.0005197512537612838, + "loss": 1.4591, + "theoretical_loss": 3.622142927092306, + "tokens_seen": 1177681920 + }, + { + "epoch": 0.36, + "learning_rate": 0.0005197352056168506, + "loss": 1.435, + "theoretical_loss": 3.6221244671071298, + "tokens_seen": 1177747456 + }, + { + "epoch": 0.36, + "learning_rate": 0.0005197191574724172, + "loss": 1.3684, + "theoretical_loss": 3.6221060084367354, + "tokens_seen": 1177812992 + }, + { + "epoch": 0.36, + "learning_rate": 0.000519703109327984, + "loss": 1.5051, + "theoretical_loss": 3.6220875510809556, + "tokens_seen": 1177878528 + }, + { + "epoch": 0.36, + "learning_rate": 0.0005196870611835506, + "loss": 1.3823, + "theoretical_loss": 3.6220690950396235, + "tokens_seen": 1177944064 + }, + { + "epoch": 0.36, + "learning_rate": 0.0005196710130391174, + "loss": 1.4032, + "theoretical_loss": 3.622050640312573, + "tokens_seen": 1178009600 + }, + { + "epoch": 0.36, + "learning_rate": 0.000519654964894684, + "loss": 1.4425, + "theoretical_loss": 3.622032186899637, + "tokens_seen": 1178075136 + }, + { + "epoch": 0.36, + "learning_rate": 0.0005196389167502508, + "loss": 1.3081, + "theoretical_loss": 3.622013734800649, + "tokens_seen": 1178140672 + }, + { + "epoch": 0.36, + "learning_rate": 0.0005196228686058174, + "loss": 1.3482, + "theoretical_loss": 3.6219952840154424, + "tokens_seen": 1178206208 + }, + { + "epoch": 0.36, + "learning_rate": 0.0005196068204613842, + "loss": 1.4714, + "theoretical_loss": 3.62197683454385, + "tokens_seen": 1178271744 + }, + { + "epoch": 0.36, + "learning_rate": 0.0005195907723169508, + "loss": 1.3086, + "theoretical_loss": 3.6219583863857068, + "tokens_seen": 1178337280 + }, + { + "epoch": 0.36, + "learning_rate": 0.0005195747241725176, + "loss": 1.5434, + "theoretical_loss": 3.6219399395408436, + "tokens_seen": 1178402816 + }, + { + "epoch": 0.36, + "learning_rate": 0.0005195586760280842, + "loss": 1.405, + "theoretical_loss": 3.6219214940090962, + "tokens_seen": 1178468352 + }, + { + "epoch": 0.36, + "learning_rate": 0.000519542627883651, + "loss": 1.43, + "theoretical_loss": 3.6219030497902978, + "tokens_seen": 1178533888 + }, + { + "epoch": 0.36, + "learning_rate": 0.0005195265797392177, + "loss": 1.4381, + "theoretical_loss": 3.6218846068842807, + "tokens_seen": 1178599424 + }, + { + "epoch": 0.36, + "learning_rate": 0.0005195105315947844, + "loss": 1.3493, + "theoretical_loss": 3.6218661652908795, + "tokens_seen": 1178664960 + }, + { + "epoch": 0.36, + "learning_rate": 0.000519494483450351, + "loss": 1.5077, + "theoretical_loss": 3.6218477250099275, + "tokens_seen": 1178730496 + }, + { + "epoch": 0.36, + "learning_rate": 0.0005194784353059178, + "loss": 1.3506, + "theoretical_loss": 3.621829286041258, + "tokens_seen": 1178796032 + }, + { + "epoch": 0.36, + "learning_rate": 0.0005194623871614845, + "loss": 1.3155, + "theoretical_loss": 3.6218108483847056, + "tokens_seen": 1178861568 + }, + { + "epoch": 0.36, + "learning_rate": 0.0005194463390170512, + "loss": 1.4634, + "theoretical_loss": 3.6217924120401035, + "tokens_seen": 1178927104 + }, + { + "epoch": 0.36, + "learning_rate": 0.0005194302908726179, + "loss": 1.5196, + "theoretical_loss": 3.621773977007285, + "tokens_seen": 1178992640 + }, + { + "epoch": 0.36, + "learning_rate": 0.0005194142427281846, + "loss": 1.4514, + "theoretical_loss": 3.6217555432860844, + "tokens_seen": 1179058176 + }, + { + "epoch": 0.36, + "learning_rate": 0.0005193981945837513, + "loss": 1.5279, + "theoretical_loss": 3.621737110876335, + "tokens_seen": 1179123712 + }, + { + "epoch": 0.36, + "learning_rate": 0.000519382146439318, + "loss": 1.718, + "theoretical_loss": 3.6217186797778713, + "tokens_seen": 1179189248 + }, + { + "epoch": 0.36, + "learning_rate": 0.0005193660982948847, + "loss": 1.4468, + "theoretical_loss": 3.6217002499905266, + "tokens_seen": 1179254784 + }, + { + "epoch": 0.36, + "learning_rate": 0.0005193500501504514, + "loss": 1.4742, + "theoretical_loss": 3.6216818215141346, + "tokens_seen": 1179320320 + }, + { + "epoch": 0.36, + "learning_rate": 0.0005193340020060181, + "loss": 1.512, + "theoretical_loss": 3.62166339434853, + "tokens_seen": 1179385856 + }, + { + "epoch": 0.36, + "learning_rate": 0.0005193179538615848, + "loss": 1.4644, + "theoretical_loss": 3.6216449684935466, + "tokens_seen": 1179451392 + }, + { + "epoch": 0.36, + "learning_rate": 0.0005193019057171515, + "loss": 1.4211, + "theoretical_loss": 3.6216265439490174, + "tokens_seen": 1179516928 + }, + { + "epoch": 0.36, + "learning_rate": 0.0005192858575727182, + "loss": 1.3627, + "theoretical_loss": 3.621608120714777, + "tokens_seen": 1179582464 + }, + { + "epoch": 0.36, + "objective/train/docs_used": 431781, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.4435052871704102, + "objective/train/theoretical_loss": 3.6215896987906597, + "objective/train/tokens_used": 1200108000, + "theoretical_loss": 3.6215896987906597, + "tokens_seen": 1179648000 + }, + { + "epoch": 0.36, + "learning_rate": 0.0005192698094282849, + "loss": 1.5476, + "theoretical_loss": 3.6215896987906597, + "tokens_seen": 1179648000 + }, + { + "epoch": 0.36, + "learning_rate": 0.0005192537612838516, + "loss": 1.4014, + "theoretical_loss": 3.621571278176499, + "tokens_seen": 1179713536 + }, + { + "epoch": 0.36, + "learning_rate": 0.0005192377131394183, + "loss": 1.4276, + "theoretical_loss": 3.6215528588721297, + "tokens_seen": 1179779072 + }, + { + "epoch": 0.36, + "learning_rate": 0.000519221664994985, + "loss": 1.5051, + "theoretical_loss": 3.6215344408773857, + "tokens_seen": 1179844608 + }, + { + "epoch": 0.36, + "learning_rate": 0.0005192056168505517, + "loss": 1.4458, + "theoretical_loss": 3.6215160241921005, + "tokens_seen": 1179910144 + }, + { + "epoch": 0.36, + "learning_rate": 0.0005191895687061184, + "loss": 1.3718, + "theoretical_loss": 3.621497608816109, + "tokens_seen": 1179975680 + }, + { + "epoch": 0.36, + "learning_rate": 0.0005191735205616851, + "loss": 1.5223, + "theoretical_loss": 3.621479194749245, + "tokens_seen": 1180041216 + }, + { + "epoch": 0.36, + "learning_rate": 0.0005191574724172518, + "loss": 1.4686, + "theoretical_loss": 3.6214607819913427, + "tokens_seen": 1180106752 + }, + { + "epoch": 0.36, + "learning_rate": 0.0005191414242728185, + "loss": 1.4156, + "theoretical_loss": 3.621442370542237, + "tokens_seen": 1180172288 + }, + { + "epoch": 0.36, + "learning_rate": 0.0005191253761283852, + "loss": 1.4081, + "theoretical_loss": 3.6214239604017617, + "tokens_seen": 1180237824 + }, + { + "epoch": 0.36, + "learning_rate": 0.0005191093279839519, + "loss": 1.4357, + "theoretical_loss": 3.621405551569751, + "tokens_seen": 1180303360 + }, + { + "epoch": 0.36, + "learning_rate": 0.0005190932798395186, + "loss": 1.4975, + "theoretical_loss": 3.6213871440460395, + "tokens_seen": 1180368896 + }, + { + "epoch": 0.36, + "learning_rate": 0.0005190772316950853, + "loss": 1.435, + "theoretical_loss": 3.6213687378304615, + "tokens_seen": 1180434432 + }, + { + "epoch": 0.36, + "learning_rate": 0.000519061183550652, + "loss": 1.3004, + "theoretical_loss": 3.6213503329228516, + "tokens_seen": 1180499968 + }, + { + "epoch": 0.36, + "learning_rate": 0.0005190451354062187, + "loss": 1.4416, + "theoretical_loss": 3.621331929323044, + "tokens_seen": 1180565504 + }, + { + "epoch": 0.36, + "learning_rate": 0.0005190290872617854, + "loss": 1.409, + "theoretical_loss": 3.621313527030873, + "tokens_seen": 1180631040 + }, + { + "epoch": 0.36, + "learning_rate": 0.0005190130391173521, + "loss": 1.4913, + "theoretical_loss": 3.621295126046174, + "tokens_seen": 1180696576 + }, + { + "epoch": 0.36, + "learning_rate": 0.0005189969909729188, + "loss": 1.4405, + "theoretical_loss": 3.6212767263687806, + "tokens_seen": 1180762112 + }, + { + "epoch": 0.36, + "learning_rate": 0.0005189809428284855, + "loss": 1.3934, + "theoretical_loss": 3.621258327998528, + "tokens_seen": 1180827648 + }, + { + "epoch": 0.36, + "learning_rate": 0.0005189648946840522, + "loss": 1.4339, + "theoretical_loss": 3.62123993093525, + "tokens_seen": 1180893184 + }, + { + "epoch": 0.36, + "learning_rate": 0.0005189488465396189, + "loss": 1.5014, + "theoretical_loss": 3.621221535178782, + "tokens_seen": 1180958720 + }, + { + "epoch": 0.36, + "learning_rate": 0.0005189327983951856, + "loss": 1.5647, + "theoretical_loss": 3.6212031407289587, + "tokens_seen": 1181024256 + }, + { + "epoch": 0.36, + "learning_rate": 0.0005189167502507523, + "loss": 1.41, + "theoretical_loss": 3.621184747585614, + "tokens_seen": 1181089792 + }, + { + "epoch": 0.36, + "learning_rate": 0.000518900702106319, + "loss": 1.5114, + "theoretical_loss": 3.621166355748583, + "tokens_seen": 1181155328 + }, + { + "epoch": 0.36, + "learning_rate": 0.0005188846539618857, + "loss": 1.3342, + "theoretical_loss": 3.621147965217701, + "tokens_seen": 1181220864 + }, + { + "epoch": 0.36, + "learning_rate": 0.0005188686058174524, + "loss": 1.5116, + "theoretical_loss": 3.621129575992802, + "tokens_seen": 1181286400 + }, + { + "epoch": 0.36, + "learning_rate": 0.0005188525576730191, + "loss": 1.3856, + "theoretical_loss": 3.6211111880737215, + "tokens_seen": 1181351936 + }, + { + "epoch": 0.36, + "learning_rate": 0.0005188365095285858, + "loss": 1.455, + "theoretical_loss": 3.6210928014602937, + "tokens_seen": 1181417472 + }, + { + "epoch": 0.36, + "learning_rate": 0.0005188204613841525, + "loss": 1.4002, + "theoretical_loss": 3.621074416152354, + "tokens_seen": 1181483008 + }, + { + "epoch": 0.36, + "learning_rate": 0.0005188044132397191, + "loss": 1.5215, + "theoretical_loss": 3.621056032149737, + "tokens_seen": 1181548544 + }, + { + "epoch": 0.36, + "learning_rate": 0.0005187883650952859, + "loss": 1.4656, + "theoretical_loss": 3.6210376494522776, + "tokens_seen": 1181614080 + }, + { + "epoch": 0.36, + "learning_rate": 0.0005187723169508525, + "loss": 1.3549, + "theoretical_loss": 3.621019268059811, + "tokens_seen": 1181679616 + }, + { + "epoch": 0.36, + "learning_rate": 0.0005187562688064193, + "loss": 1.3775, + "theoretical_loss": 3.621000887972172, + "tokens_seen": 1181745152 + }, + { + "epoch": 0.36, + "learning_rate": 0.0005187402206619859, + "loss": 1.5235, + "theoretical_loss": 3.6209825091891954, + "tokens_seen": 1181810688 + }, + { + "epoch": 0.36, + "learning_rate": 0.0005187241725175527, + "loss": 1.4178, + "theoretical_loss": 3.620964131710717, + "tokens_seen": 1181876224 + }, + { + "epoch": 0.36, + "learning_rate": 0.0005187081243731193, + "loss": 1.4857, + "theoretical_loss": 3.620945755536571, + "tokens_seen": 1181941760 + }, + { + "epoch": 0.36, + "learning_rate": 0.0005186920762286861, + "loss": 1.446, + "theoretical_loss": 3.620927380666593, + "tokens_seen": 1182007296 + }, + { + "epoch": 0.36, + "learning_rate": 0.0005186760280842527, + "loss": 1.474, + "theoretical_loss": 3.620909007100618, + "tokens_seen": 1182072832 + }, + { + "epoch": 0.36, + "learning_rate": 0.0005186599799398195, + "loss": 1.5286, + "theoretical_loss": 3.6208906348384815, + "tokens_seen": 1182138368 + }, + { + "epoch": 0.36, + "learning_rate": 0.0005186439317953861, + "loss": 1.3582, + "theoretical_loss": 3.6208722638800177, + "tokens_seen": 1182203904 + }, + { + "epoch": 0.36, + "learning_rate": 0.0005186278836509529, + "loss": 1.396, + "theoretical_loss": 3.620853894225063, + "tokens_seen": 1182269440 + }, + { + "epoch": 0.36, + "learning_rate": 0.0005186118355065195, + "loss": 1.4905, + "theoretical_loss": 3.620835525873453, + "tokens_seen": 1182334976 + }, + { + "epoch": 0.36, + "learning_rate": 0.0005185957873620863, + "loss": 1.3487, + "theoretical_loss": 3.6208171588250213, + "tokens_seen": 1182400512 + }, + { + "epoch": 0.36, + "learning_rate": 0.0005185797392176529, + "loss": 1.4135, + "theoretical_loss": 3.6207987930796044, + "tokens_seen": 1182466048 + }, + { + "epoch": 0.36, + "learning_rate": 0.0005185636910732197, + "loss": 1.5123, + "theoretical_loss": 3.620780428637037, + "tokens_seen": 1182531584 + }, + { + "epoch": 0.36, + "learning_rate": 0.0005185476429287863, + "loss": 1.4497, + "theoretical_loss": 3.6207620654971553, + "tokens_seen": 1182597120 + }, + { + "epoch": 0.36, + "learning_rate": 0.0005185315947843531, + "loss": 1.3676, + "theoretical_loss": 3.620743703659794, + "tokens_seen": 1182662656 + }, + { + "epoch": 0.36, + "learning_rate": 0.0005185155466399197, + "loss": 1.3474, + "theoretical_loss": 3.620725343124789, + "tokens_seen": 1182728192 + }, + { + "epoch": 0.36, + "learning_rate": 0.0005184994984954865, + "loss": 1.3575, + "theoretical_loss": 3.620706983891975, + "tokens_seen": 1182793728 + }, + { + "epoch": 0.36, + "learning_rate": 0.0005184834503510531, + "loss": 1.4214, + "theoretical_loss": 3.620688625961189, + "tokens_seen": 1182859264 + }, + { + "epoch": 0.36, + "objective/train/docs_used": 433029, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.4044227600097656, + "objective/train/theoretical_loss": 3.6206702693322654, + "objective/train/tokens_used": 1203384800, + "theoretical_loss": 3.6206702693322654, + "tokens_seen": 1182924800 + }, + { + "epoch": 0.36, + "learning_rate": 0.0005184674022066199, + "loss": 1.3674, + "theoretical_loss": 3.6206702693322654, + "tokens_seen": 1182924800 + }, + { + "epoch": 0.36, + "learning_rate": 0.0005184513540621865, + "loss": 1.426, + "theoretical_loss": 3.6206519140050393, + "tokens_seen": 1182990336 + }, + { + "epoch": 0.36, + "learning_rate": 0.0005184353059177533, + "loss": 1.5036, + "theoretical_loss": 3.620633559979348, + "tokens_seen": 1183055872 + }, + { + "epoch": 0.36, + "learning_rate": 0.0005184192577733199, + "loss": 1.4306, + "theoretical_loss": 3.6206152072550255, + "tokens_seen": 1183121408 + }, + { + "epoch": 0.36, + "learning_rate": 0.0005184032096288867, + "loss": 1.5019, + "theoretical_loss": 3.620596855831908, + "tokens_seen": 1183186944 + }, + { + "epoch": 0.36, + "learning_rate": 0.0005183871614844533, + "loss": 1.3458, + "theoretical_loss": 3.620578505709831, + "tokens_seen": 1183252480 + }, + { + "epoch": 0.36, + "learning_rate": 0.0005183711133400201, + "loss": 1.557, + "theoretical_loss": 3.6205601568886308, + "tokens_seen": 1183318016 + }, + { + "epoch": 0.36, + "learning_rate": 0.0005183550651955867, + "loss": 1.4723, + "theoretical_loss": 3.6205418093681425, + "tokens_seen": 1183383552 + }, + { + "epoch": 0.36, + "learning_rate": 0.0005183390170511535, + "loss": 1.3579, + "theoretical_loss": 3.6205234631482024, + "tokens_seen": 1183449088 + }, + { + "epoch": 0.36, + "learning_rate": 0.0005183229689067201, + "loss": 1.4235, + "theoretical_loss": 3.6205051182286456, + "tokens_seen": 1183514624 + }, + { + "epoch": 0.36, + "learning_rate": 0.0005183069207622869, + "loss": 1.5874, + "theoretical_loss": 3.620486774609309, + "tokens_seen": 1183580160 + }, + { + "epoch": 0.36, + "learning_rate": 0.0005182908726178535, + "loss": 1.3826, + "theoretical_loss": 3.620468432290027, + "tokens_seen": 1183645696 + }, + { + "epoch": 0.36, + "learning_rate": 0.0005182748244734203, + "loss": 1.3915, + "theoretical_loss": 3.620450091270637, + "tokens_seen": 1183711232 + }, + { + "epoch": 0.36, + "learning_rate": 0.000518258776328987, + "loss": 1.3322, + "theoretical_loss": 3.620431751550974, + "tokens_seen": 1183776768 + }, + { + "epoch": 0.36, + "learning_rate": 0.0005182427281845537, + "loss": 1.3545, + "theoretical_loss": 3.6204134131308736, + "tokens_seen": 1183842304 + }, + { + "epoch": 0.36, + "learning_rate": 0.0005182266800401203, + "loss": 1.5746, + "theoretical_loss": 3.620395076010173, + "tokens_seen": 1183907840 + }, + { + "epoch": 0.36, + "learning_rate": 0.0005182106318956871, + "loss": 1.4176, + "theoretical_loss": 3.6203767401887075, + "tokens_seen": 1183973376 + }, + { + "epoch": 0.36, + "learning_rate": 0.0005181945837512537, + "loss": 1.3635, + "theoretical_loss": 3.6203584056663134, + "tokens_seen": 1184038912 + }, + { + "epoch": 0.36, + "learning_rate": 0.0005181785356068205, + "loss": 1.4664, + "theoretical_loss": 3.620340072442826, + "tokens_seen": 1184104448 + }, + { + "epoch": 0.36, + "learning_rate": 0.0005181624874623871, + "loss": 1.479, + "theoretical_loss": 3.620321740518082, + "tokens_seen": 1184169984 + }, + { + "epoch": 0.36, + "learning_rate": 0.0005181464393179539, + "loss": 1.4054, + "theoretical_loss": 3.6203034098919176, + "tokens_seen": 1184235520 + }, + { + "epoch": 0.36, + "learning_rate": 0.0005181303911735206, + "loss": 1.4801, + "theoretical_loss": 3.6202850805641695, + "tokens_seen": 1184301056 + }, + { + "epoch": 0.36, + "learning_rate": 0.0005181143430290873, + "loss": 1.4179, + "theoretical_loss": 3.6202667525346723, + "tokens_seen": 1184366592 + }, + { + "epoch": 0.36, + "learning_rate": 0.000518098294884654, + "loss": 1.4231, + "theoretical_loss": 3.6202484258032634, + "tokens_seen": 1184432128 + }, + { + "epoch": 0.36, + "learning_rate": 0.0005180822467402207, + "loss": 1.4068, + "theoretical_loss": 3.6202301003697785, + "tokens_seen": 1184497664 + }, + { + "epoch": 0.36, + "learning_rate": 0.0005180661985957874, + "loss": 1.3575, + "theoretical_loss": 3.620211776234055, + "tokens_seen": 1184563200 + }, + { + "epoch": 0.36, + "learning_rate": 0.0005180501504513541, + "loss": 1.456, + "theoretical_loss": 3.6201934533959275, + "tokens_seen": 1184628736 + }, + { + "epoch": 0.36, + "learning_rate": 0.0005180341023069208, + "loss": 1.3834, + "theoretical_loss": 3.620175131855234, + "tokens_seen": 1184694272 + }, + { + "epoch": 0.36, + "learning_rate": 0.0005180180541624875, + "loss": 1.4324, + "theoretical_loss": 3.6201568116118095, + "tokens_seen": 1184759808 + }, + { + "epoch": 0.36, + "learning_rate": 0.0005180020060180542, + "loss": 1.4839, + "theoretical_loss": 3.6201384926654914, + "tokens_seen": 1184825344 + }, + { + "epoch": 0.36, + "learning_rate": 0.0005179859578736209, + "loss": 1.3463, + "theoretical_loss": 3.620120175016115, + "tokens_seen": 1184890880 + }, + { + "epoch": 0.36, + "learning_rate": 0.0005179699097291876, + "loss": 1.4352, + "theoretical_loss": 3.620101858663518, + "tokens_seen": 1184956416 + }, + { + "epoch": 0.36, + "learning_rate": 0.0005179538615847543, + "loss": 1.4878, + "theoretical_loss": 3.620083543607536, + "tokens_seen": 1185021952 + }, + { + "epoch": 0.36, + "learning_rate": 0.000517937813440321, + "loss": 1.4013, + "theoretical_loss": 3.620065229848006, + "tokens_seen": 1185087488 + }, + { + "epoch": 0.36, + "learning_rate": 0.0005179217652958877, + "loss": 1.3989, + "theoretical_loss": 3.620046917384764, + "tokens_seen": 1185153024 + }, + { + "epoch": 0.36, + "learning_rate": 0.0005179057171514544, + "loss": 1.3996, + "theoretical_loss": 3.620028606217647, + "tokens_seen": 1185218560 + }, + { + "epoch": 0.36, + "learning_rate": 0.0005178896690070211, + "loss": 1.4359, + "theoretical_loss": 3.620010296346492, + "tokens_seen": 1185284096 + }, + { + "epoch": 0.36, + "learning_rate": 0.0005178736208625878, + "loss": 1.5106, + "theoretical_loss": 3.619991987771135, + "tokens_seen": 1185349632 + }, + { + "epoch": 0.36, + "learning_rate": 0.0005178575727181545, + "loss": 1.438, + "theoretical_loss": 3.619973680491413, + "tokens_seen": 1185415168 + }, + { + "epoch": 0.36, + "learning_rate": 0.0005178415245737212, + "loss": 1.323, + "theoretical_loss": 3.619955374507162, + "tokens_seen": 1185480704 + }, + { + "epoch": 0.36, + "learning_rate": 0.0005178254764292879, + "loss": 1.4581, + "theoretical_loss": 3.61993706981822, + "tokens_seen": 1185546240 + }, + { + "epoch": 0.36, + "learning_rate": 0.0005178094282848546, + "loss": 1.4382, + "theoretical_loss": 3.6199187664244223, + "tokens_seen": 1185611776 + }, + { + "epoch": 0.36, + "learning_rate": 0.0005177933801404213, + "loss": 1.3344, + "theoretical_loss": 3.6199004643256067, + "tokens_seen": 1185677312 + }, + { + "epoch": 0.36, + "learning_rate": 0.000517777331995988, + "loss": 1.4427, + "theoretical_loss": 3.6198821635216096, + "tokens_seen": 1185742848 + }, + { + "epoch": 0.36, + "learning_rate": 0.0005177612838515547, + "loss": 1.5732, + "theoretical_loss": 3.6198638640122676, + "tokens_seen": 1185808384 + }, + { + "epoch": 0.36, + "learning_rate": 0.0005177452357071214, + "loss": 1.433, + "theoretical_loss": 3.6198455657974185, + "tokens_seen": 1185873920 + }, + { + "epoch": 0.36, + "learning_rate": 0.0005177291875626881, + "loss": 1.4538, + "theoretical_loss": 3.6198272688768984, + "tokens_seen": 1185939456 + }, + { + "epoch": 0.36, + "learning_rate": 0.0005177131394182549, + "loss": 1.4348, + "theoretical_loss": 3.619808973250544, + "tokens_seen": 1186004992 + }, + { + "epoch": 0.36, + "learning_rate": 0.0005176970912738215, + "loss": 1.514, + "theoretical_loss": 3.6197906789181933, + "tokens_seen": 1186070528 + }, + { + "epoch": 0.36, + "learning_rate": 0.0005176810431293883, + "loss": 1.464, + "theoretical_loss": 3.619772385879682, + "tokens_seen": 1186136064 + }, + { + "epoch": 0.36, + "objective/train/docs_used": 434243, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.5226637125015259, + "objective/train/theoretical_loss": 3.6197540941348487, + "objective/train/tokens_used": 1206661600, + "theoretical_loss": 3.6197540941348487, + "tokens_seen": 1186201600 + }, + { + "epoch": 0.36, + "learning_rate": 0.0005176649949849549, + "loss": 1.4754, + "theoretical_loss": 3.6197540941348487, + "tokens_seen": 1186201600 + }, + { + "epoch": 0.36, + "learning_rate": 0.0005176489468405217, + "loss": 1.4209, + "theoretical_loss": 3.619735803683529, + "tokens_seen": 1186267136 + }, + { + "epoch": 0.36, + "learning_rate": 0.0005176328986960883, + "loss": 1.3512, + "theoretical_loss": 3.61971751452556, + "tokens_seen": 1186332672 + }, + { + "epoch": 0.36, + "learning_rate": 0.0005176168505516551, + "loss": 1.3755, + "theoretical_loss": 3.6196992266607797, + "tokens_seen": 1186398208 + }, + { + "epoch": 0.36, + "learning_rate": 0.0005176008024072217, + "loss": 1.4995, + "theoretical_loss": 3.619680940089025, + "tokens_seen": 1186463744 + }, + { + "epoch": 0.36, + "learning_rate": 0.0005175847542627885, + "loss": 1.4473, + "theoretical_loss": 3.6196626548101327, + "tokens_seen": 1186529280 + }, + { + "epoch": 0.36, + "learning_rate": 0.0005175687061183551, + "loss": 1.5182, + "theoretical_loss": 3.6196443708239405, + "tokens_seen": 1186594816 + }, + { + "epoch": 0.36, + "learning_rate": 0.0005175526579739219, + "loss": 1.5713, + "theoretical_loss": 3.6196260881302846, + "tokens_seen": 1186660352 + }, + { + "epoch": 0.36, + "learning_rate": 0.0005175366098294885, + "loss": 1.4301, + "theoretical_loss": 3.619607806729004, + "tokens_seen": 1186725888 + }, + { + "epoch": 0.36, + "learning_rate": 0.0005175205616850552, + "loss": 1.3987, + "theoretical_loss": 3.619589526619934, + "tokens_seen": 1186791424 + }, + { + "epoch": 0.36, + "learning_rate": 0.0005175045135406219, + "loss": 1.3889, + "theoretical_loss": 3.6195712478029134, + "tokens_seen": 1186856960 + }, + { + "epoch": 0.36, + "learning_rate": 0.0005174884653961886, + "loss": 1.6038, + "theoretical_loss": 3.6195529702777787, + "tokens_seen": 1186922496 + }, + { + "epoch": 0.36, + "learning_rate": 0.0005174724172517553, + "loss": 1.463, + "theoretical_loss": 3.6195346940443676, + "tokens_seen": 1186988032 + }, + { + "epoch": 0.36, + "learning_rate": 0.000517456369107322, + "loss": 1.4994, + "theoretical_loss": 3.619516419102518, + "tokens_seen": 1187053568 + }, + { + "epoch": 0.36, + "learning_rate": 0.0005174403209628887, + "loss": 1.4332, + "theoretical_loss": 3.619498145452066, + "tokens_seen": 1187119104 + }, + { + "epoch": 0.36, + "learning_rate": 0.0005174242728184554, + "loss": 1.3865, + "theoretical_loss": 3.6194798730928506, + "tokens_seen": 1187184640 + }, + { + "epoch": 0.36, + "learning_rate": 0.000517408224674022, + "loss": 1.4733, + "theoretical_loss": 3.6194616020247077, + "tokens_seen": 1187250176 + }, + { + "epoch": 0.36, + "learning_rate": 0.0005173921765295888, + "loss": 1.4252, + "theoretical_loss": 3.619443332247476, + "tokens_seen": 1187315712 + }, + { + "epoch": 0.36, + "learning_rate": 0.0005173761283851554, + "loss": 1.3317, + "theoretical_loss": 3.6194250637609926, + "tokens_seen": 1187381248 + }, + { + "epoch": 0.36, + "learning_rate": 0.0005173600802407222, + "loss": 1.3823, + "theoretical_loss": 3.6194067965650953, + "tokens_seen": 1187446784 + }, + { + "epoch": 0.36, + "learning_rate": 0.0005173440320962888, + "loss": 1.4888, + "theoretical_loss": 3.619388530659622, + "tokens_seen": 1187512320 + }, + { + "epoch": 0.36, + "learning_rate": 0.0005173279839518556, + "loss": 1.4317, + "theoretical_loss": 3.6193702660444096, + "tokens_seen": 1187577856 + }, + { + "epoch": 0.36, + "learning_rate": 0.0005173119358074222, + "loss": 1.5755, + "theoretical_loss": 3.619352002719296, + "tokens_seen": 1187643392 + }, + { + "epoch": 0.36, + "learning_rate": 0.000517295887662989, + "loss": 1.4779, + "theoretical_loss": 3.619333740684119, + "tokens_seen": 1187708928 + }, + { + "epoch": 0.36, + "learning_rate": 0.0005172798395185556, + "loss": 1.5061, + "theoretical_loss": 3.6193154799387166, + "tokens_seen": 1187774464 + }, + { + "epoch": 0.36, + "learning_rate": 0.0005172637913741224, + "loss": 1.4798, + "theoretical_loss": 3.6192972204829257, + "tokens_seen": 1187840000 + }, + { + "epoch": 0.36, + "learning_rate": 0.000517247743229689, + "loss": 1.3918, + "theoretical_loss": 3.619278962316585, + "tokens_seen": 1187905536 + }, + { + "epoch": 0.36, + "learning_rate": 0.0005172316950852558, + "loss": 1.5365, + "theoretical_loss": 3.619260705439532, + "tokens_seen": 1187971072 + }, + { + "epoch": 0.36, + "learning_rate": 0.0005172156469408224, + "loss": 1.5141, + "theoretical_loss": 3.619242449851604, + "tokens_seen": 1188036608 + }, + { + "epoch": 0.36, + "learning_rate": 0.0005171995987963892, + "loss": 1.3917, + "theoretical_loss": 3.61922419555264, + "tokens_seen": 1188102144 + }, + { + "epoch": 0.36, + "learning_rate": 0.0005171835506519558, + "loss": 1.2963, + "theoretical_loss": 3.619205942542477, + "tokens_seen": 1188167680 + }, + { + "epoch": 0.36, + "learning_rate": 0.0005171675025075226, + "loss": 1.3745, + "theoretical_loss": 3.6191876908209526, + "tokens_seen": 1188233216 + }, + { + "epoch": 0.36, + "learning_rate": 0.0005171514543630892, + "loss": 1.4458, + "theoretical_loss": 3.619169440387906, + "tokens_seen": 1188298752 + }, + { + "epoch": 0.36, + "learning_rate": 0.000517135406218656, + "loss": 1.4334, + "theoretical_loss": 3.6191511912431746, + "tokens_seen": 1188364288 + }, + { + "epoch": 0.36, + "learning_rate": 0.0005171193580742226, + "loss": 1.5016, + "theoretical_loss": 3.619132943386596, + "tokens_seen": 1188429824 + }, + { + "epoch": 0.36, + "learning_rate": 0.0005171033099297894, + "loss": 1.5111, + "theoretical_loss": 3.6191146968180083, + "tokens_seen": 1188495360 + }, + { + "epoch": 0.36, + "learning_rate": 0.000517087261785356, + "loss": 1.5388, + "theoretical_loss": 3.61909645153725, + "tokens_seen": 1188560896 + }, + { + "epoch": 0.36, + "learning_rate": 0.0005170712136409228, + "loss": 1.5599, + "theoretical_loss": 3.6190782075441597, + "tokens_seen": 1188626432 + }, + { + "epoch": 0.36, + "learning_rate": 0.0005170551654964894, + "loss": 1.397, + "theoretical_loss": 3.6190599648385744, + "tokens_seen": 1188691968 + }, + { + "epoch": 0.36, + "learning_rate": 0.0005170391173520562, + "loss": 1.4176, + "theoretical_loss": 3.6190417234203327, + "tokens_seen": 1188757504 + }, + { + "epoch": 0.36, + "learning_rate": 0.0005170230692076228, + "loss": 1.2998, + "theoretical_loss": 3.6190234832892734, + "tokens_seen": 1188823040 + }, + { + "epoch": 0.36, + "learning_rate": 0.0005170070210631896, + "loss": 1.4373, + "theoretical_loss": 3.6190052444452334, + "tokens_seen": 1188888576 + }, + { + "epoch": 0.36, + "learning_rate": 0.0005169909729187562, + "loss": 1.4502, + "theoretical_loss": 3.6189870068880525, + "tokens_seen": 1188954112 + }, + { + "epoch": 0.36, + "learning_rate": 0.000516974924774323, + "loss": 1.6178, + "theoretical_loss": 3.6189687706175677, + "tokens_seen": 1189019648 + }, + { + "epoch": 0.36, + "learning_rate": 0.0005169588766298896, + "loss": 1.3065, + "theoretical_loss": 3.618950535633618, + "tokens_seen": 1189085184 + }, + { + "epoch": 0.36, + "learning_rate": 0.0005169428284854564, + "loss": 1.3635, + "theoretical_loss": 3.6189323019360415, + "tokens_seen": 1189150720 + }, + { + "epoch": 0.36, + "learning_rate": 0.000516926780341023, + "loss": 1.3342, + "theoretical_loss": 3.6189140695246764, + "tokens_seen": 1189216256 + }, + { + "epoch": 0.36, + "learning_rate": 0.0005169107321965898, + "loss": 1.4482, + "theoretical_loss": 3.6188958383993617, + "tokens_seen": 1189281792 + }, + { + "epoch": 0.36, + "learning_rate": 0.0005168946840521564, + "loss": 1.3524, + "theoretical_loss": 3.618877608559935, + "tokens_seen": 1189347328 + }, + { + "epoch": 0.36, + "learning_rate": 0.0005168786359077232, + "loss": 1.4449, + "theoretical_loss": 3.6188593800062354, + "tokens_seen": 1189412864 + }, + { + "epoch": 0.36, + "objective/train/docs_used": 435462, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.4314229488372803, + "objective/train/theoretical_loss": 3.618841152738101, + "objective/train/tokens_used": 1209938400, + "theoretical_loss": 3.618841152738101, + "tokens_seen": 1189478400 + }, + { + "epoch": 0.36, + "learning_rate": 0.0005168625877632898, + "loss": 1.4069, + "theoretical_loss": 3.618841152738101, + "tokens_seen": 1189478400 + }, + { + "epoch": 0.36, + "learning_rate": 0.0005168465396188566, + "loss": 1.5516, + "theoretical_loss": 3.618822926755371, + "tokens_seen": 1189543936 + }, + { + "epoch": 0.36, + "learning_rate": 0.0005168304914744232, + "loss": 1.5688, + "theoretical_loss": 3.618804702057883, + "tokens_seen": 1189609472 + }, + { + "epoch": 0.36, + "learning_rate": 0.00051681444332999, + "loss": 1.4914, + "theoretical_loss": 3.6187864786454758, + "tokens_seen": 1189675008 + }, + { + "epoch": 0.36, + "learning_rate": 0.0005167983951855566, + "loss": 1.5421, + "theoretical_loss": 3.6187682565179884, + "tokens_seen": 1189740544 + }, + { + "epoch": 0.36, + "learning_rate": 0.0005167823470411234, + "loss": 1.43, + "theoretical_loss": 3.6187500356752595, + "tokens_seen": 1189806080 + }, + { + "epoch": 0.36, + "learning_rate": 0.00051676629889669, + "loss": 1.3056, + "theoretical_loss": 3.618731816117127, + "tokens_seen": 1189871616 + }, + { + "epoch": 0.36, + "learning_rate": 0.0005167502507522568, + "loss": 1.4172, + "theoretical_loss": 3.6187135978434304, + "tokens_seen": 1189937152 + }, + { + "epoch": 0.36, + "learning_rate": 0.0005167342026078235, + "loss": 1.4719, + "theoretical_loss": 3.6186953808540085, + "tokens_seen": 1190002688 + }, + { + "epoch": 0.36, + "learning_rate": 0.0005167181544633902, + "loss": 1.4567, + "theoretical_loss": 3.618677165148699, + "tokens_seen": 1190068224 + }, + { + "epoch": 0.36, + "learning_rate": 0.0005167021063189569, + "loss": 1.4578, + "theoretical_loss": 3.6186589507273412, + "tokens_seen": 1190133760 + }, + { + "epoch": 0.36, + "learning_rate": 0.0005166860581745236, + "loss": 1.4879, + "theoretical_loss": 3.618640737589775, + "tokens_seen": 1190199296 + }, + { + "epoch": 0.36, + "learning_rate": 0.0005166700100300903, + "loss": 1.5046, + "theoretical_loss": 3.6186225257358373, + "tokens_seen": 1190264832 + }, + { + "epoch": 0.36, + "learning_rate": 0.000516653961885657, + "loss": 1.309, + "theoretical_loss": 3.618604315165368, + "tokens_seen": 1190330368 + }, + { + "epoch": 0.36, + "learning_rate": 0.0005166379137412237, + "loss": 1.4215, + "theoretical_loss": 3.6185861058782063, + "tokens_seen": 1190395904 + }, + { + "epoch": 0.36, + "learning_rate": 0.0005166218655967904, + "loss": 1.4674, + "theoretical_loss": 3.6185678978741906, + "tokens_seen": 1190461440 + }, + { + "epoch": 0.36, + "learning_rate": 0.0005166058174523571, + "loss": 1.4291, + "theoretical_loss": 3.6185496911531603, + "tokens_seen": 1190526976 + }, + { + "epoch": 0.36, + "learning_rate": 0.0005165897693079238, + "loss": 1.3844, + "theoretical_loss": 3.6185314857149535, + "tokens_seen": 1190592512 + }, + { + "epoch": 0.36, + "learning_rate": 0.0005165737211634905, + "loss": 1.4538, + "theoretical_loss": 3.61851328155941, + "tokens_seen": 1190658048 + }, + { + "epoch": 0.36, + "learning_rate": 0.0005165576730190572, + "loss": 1.5329, + "theoretical_loss": 3.6184950786863688, + "tokens_seen": 1190723584 + }, + { + "epoch": 0.36, + "learning_rate": 0.0005165416248746239, + "loss": 1.3931, + "theoretical_loss": 3.6184768770956683, + "tokens_seen": 1190789120 + }, + { + "epoch": 0.36, + "learning_rate": 0.0005165255767301906, + "loss": 1.3725, + "theoretical_loss": 3.6184586767871485, + "tokens_seen": 1190854656 + }, + { + "epoch": 0.36, + "learning_rate": 0.0005165095285857574, + "loss": 1.3718, + "theoretical_loss": 3.6184404777606485, + "tokens_seen": 1190920192 + }, + { + "epoch": 0.36, + "learning_rate": 0.000516493480441324, + "loss": 1.5363, + "theoretical_loss": 3.6184222800160057, + "tokens_seen": 1190985728 + }, + { + "epoch": 0.36, + "learning_rate": 0.0005164774322968908, + "loss": 1.5486, + "theoretical_loss": 3.6184040835530618, + "tokens_seen": 1191051264 + }, + { + "epoch": 0.36, + "learning_rate": 0.0005164613841524574, + "loss": 1.3783, + "theoretical_loss": 3.6183858883716544, + "tokens_seen": 1191116800 + }, + { + "epoch": 0.36, + "learning_rate": 0.0005164453360080242, + "loss": 1.5799, + "theoretical_loss": 3.6183676944716234, + "tokens_seen": 1191182336 + }, + { + "epoch": 0.36, + "learning_rate": 0.0005164292878635908, + "loss": 1.4117, + "theoretical_loss": 3.6183495018528076, + "tokens_seen": 1191247872 + }, + { + "epoch": 0.36, + "learning_rate": 0.0005164132397191576, + "loss": 1.3951, + "theoretical_loss": 3.618331310515047, + "tokens_seen": 1191313408 + }, + { + "epoch": 0.36, + "learning_rate": 0.0005163971915747242, + "loss": 1.4414, + "theoretical_loss": 3.61831312045818, + "tokens_seen": 1191378944 + }, + { + "epoch": 0.36, + "learning_rate": 0.000516381143430291, + "loss": 1.5287, + "theoretical_loss": 3.6182949316820467, + "tokens_seen": 1191444480 + }, + { + "epoch": 0.36, + "learning_rate": 0.0005163650952858576, + "loss": 1.4015, + "theoretical_loss": 3.6182767441864856, + "tokens_seen": 1191510016 + }, + { + "epoch": 0.36, + "learning_rate": 0.0005163490471414244, + "loss": 1.3682, + "theoretical_loss": 3.6182585579713376, + "tokens_seen": 1191575552 + }, + { + "epoch": 0.36, + "learning_rate": 0.000516332998996991, + "loss": 1.4611, + "theoretical_loss": 3.618240373036441, + "tokens_seen": 1191641088 + }, + { + "epoch": 0.36, + "learning_rate": 0.0005163169508525578, + "loss": 1.3853, + "theoretical_loss": 3.618222189381635, + "tokens_seen": 1191706624 + }, + { + "epoch": 0.36, + "learning_rate": 0.0005163009027081244, + "loss": 1.2856, + "theoretical_loss": 3.6182040070067605, + "tokens_seen": 1191772160 + }, + { + "epoch": 0.36, + "learning_rate": 0.0005162848545636912, + "loss": 1.3726, + "theoretical_loss": 3.618185825911656, + "tokens_seen": 1191837696 + }, + { + "epoch": 0.36, + "learning_rate": 0.0005162688064192578, + "loss": 1.2636, + "theoretical_loss": 3.6181676460961607, + "tokens_seen": 1191903232 + }, + { + "epoch": 0.36, + "learning_rate": 0.0005162527582748246, + "loss": 1.5594, + "theoretical_loss": 3.618149467560115, + "tokens_seen": 1191968768 + }, + { + "epoch": 0.36, + "learning_rate": 0.0005162367101303912, + "loss": 1.4999, + "theoretical_loss": 3.6181312903033587, + "tokens_seen": 1192034304 + }, + { + "epoch": 0.36, + "learning_rate": 0.000516220661985958, + "loss": 1.4069, + "theoretical_loss": 3.6181131143257304, + "tokens_seen": 1192099840 + }, + { + "epoch": 0.36, + "learning_rate": 0.0005162046138415246, + "loss": 1.3989, + "theoretical_loss": 3.6180949396270705, + "tokens_seen": 1192165376 + }, + { + "epoch": 0.36, + "learning_rate": 0.0005161885656970914, + "loss": 1.4455, + "theoretical_loss": 3.618076766207219, + "tokens_seen": 1192230912 + }, + { + "epoch": 0.36, + "learning_rate": 0.000516172517552658, + "loss": 1.3607, + "theoretical_loss": 3.6180585940660146, + "tokens_seen": 1192296448 + }, + { + "epoch": 0.36, + "learning_rate": 0.0005161564694082248, + "loss": 1.5253, + "theoretical_loss": 3.618040423203298, + "tokens_seen": 1192361984 + }, + { + "epoch": 0.36, + "learning_rate": 0.0005161404212637914, + "loss": 1.3702, + "theoretical_loss": 3.618022253618909, + "tokens_seen": 1192427520 + }, + { + "epoch": 0.36, + "learning_rate": 0.0005161243731193581, + "loss": 1.311, + "theoretical_loss": 3.6180040853126862, + "tokens_seen": 1192493056 + }, + { + "epoch": 0.36, + "learning_rate": 0.0005161083249749248, + "loss": 1.3399, + "theoretical_loss": 3.6179859182844707, + "tokens_seen": 1192558592 + }, + { + "epoch": 0.36, + "learning_rate": 0.0005160922768304915, + "loss": 1.6599, + "theoretical_loss": 3.6179677525341027, + "tokens_seen": 1192624128 + }, + { + "epoch": 0.36, + "learning_rate": 0.0005160762286860582, + "loss": 1.442, + "theoretical_loss": 3.6179495880614203, + "tokens_seen": 1192689664 + }, + { + "epoch": 0.36, + "objective/train/docs_used": 436657, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.5040814876556396, + "objective/train/theoretical_loss": 3.6179314248662653, + "objective/train/tokens_used": 1213215200, + "theoretical_loss": 3.6179314248662653, + "tokens_seen": 1192755200 + }, + { + "epoch": 0.36, + "learning_rate": 0.0005160601805416249, + "loss": 1.4182, + "theoretical_loss": 3.6179314248662653, + "tokens_seen": 1192755200 + }, + { + "epoch": 0.36, + "learning_rate": 0.0005160441323971916, + "loss": 1.4382, + "theoretical_loss": 3.617913262948477, + "tokens_seen": 1192820736 + }, + { + "epoch": 0.36, + "learning_rate": 0.0005160280842527583, + "loss": 1.4627, + "theoretical_loss": 3.6178951023078945, + "tokens_seen": 1192886272 + }, + { + "epoch": 0.36, + "learning_rate": 0.000516012036108325, + "loss": 1.3499, + "theoretical_loss": 3.6178769429443594, + "tokens_seen": 1192951808 + }, + { + "epoch": 0.36, + "learning_rate": 0.0005159959879638917, + "loss": 1.481, + "theoretical_loss": 3.6178587848577104, + "tokens_seen": 1193017344 + }, + { + "epoch": 0.36, + "learning_rate": 0.0005159799398194583, + "loss": 1.3906, + "theoretical_loss": 3.6178406280477886, + "tokens_seen": 1193082880 + }, + { + "epoch": 0.36, + "learning_rate": 0.0005159638916750251, + "loss": 1.364, + "theoretical_loss": 3.6178224725144332, + "tokens_seen": 1193148416 + }, + { + "epoch": 0.36, + "learning_rate": 0.0005159478435305917, + "loss": 1.4516, + "theoretical_loss": 3.6178043182574857, + "tokens_seen": 1193213952 + }, + { + "epoch": 0.36, + "learning_rate": 0.0005159317953861585, + "loss": 1.4993, + "theoretical_loss": 3.617786165276785, + "tokens_seen": 1193279488 + }, + { + "epoch": 0.36, + "learning_rate": 0.0005159157472417251, + "loss": 1.5029, + "theoretical_loss": 3.617768013572171, + "tokens_seen": 1193345024 + }, + { + "epoch": 0.36, + "learning_rate": 0.0005158996990972919, + "loss": 1.4003, + "theoretical_loss": 3.6177498631434846, + "tokens_seen": 1193410560 + }, + { + "epoch": 0.36, + "learning_rate": 0.0005158836509528585, + "loss": 1.4071, + "theoretical_loss": 3.6177317139905667, + "tokens_seen": 1193476096 + }, + { + "epoch": 0.36, + "learning_rate": 0.0005158676028084253, + "loss": 1.5012, + "theoretical_loss": 3.617713566113257, + "tokens_seen": 1193541632 + }, + { + "epoch": 0.36, + "learning_rate": 0.0005158515546639919, + "loss": 1.5051, + "theoretical_loss": 3.617695419511395, + "tokens_seen": 1193607168 + }, + { + "epoch": 0.36, + "learning_rate": 0.0005158355065195587, + "loss": 1.4833, + "theoretical_loss": 3.617677274184822, + "tokens_seen": 1193672704 + }, + { + "epoch": 0.36, + "learning_rate": 0.0005158194583751253, + "loss": 1.47, + "theoretical_loss": 3.6176591301333785, + "tokens_seen": 1193738240 + }, + { + "epoch": 0.36, + "learning_rate": 0.0005158034102306921, + "loss": 1.3913, + "theoretical_loss": 3.6176409873569044, + "tokens_seen": 1193803776 + }, + { + "epoch": 0.36, + "learning_rate": 0.0005157873620862587, + "loss": 1.4938, + "theoretical_loss": 3.61762284585524, + "tokens_seen": 1193869312 + }, + { + "epoch": 0.36, + "learning_rate": 0.0005157713139418255, + "loss": 1.5411, + "theoretical_loss": 3.617604705628226, + "tokens_seen": 1193934848 + }, + { + "epoch": 0.36, + "learning_rate": 0.0005157552657973921, + "loss": 1.3878, + "theoretical_loss": 3.617586566675703, + "tokens_seen": 1194000384 + }, + { + "epoch": 0.36, + "learning_rate": 0.0005157392176529589, + "loss": 1.4202, + "theoretical_loss": 3.6175684289975116, + "tokens_seen": 1194065920 + }, + { + "epoch": 0.36, + "learning_rate": 0.0005157231695085255, + "loss": 1.3887, + "theoretical_loss": 3.6175502925934917, + "tokens_seen": 1194131456 + }, + { + "epoch": 0.36, + "learning_rate": 0.0005157071213640923, + "loss": 1.5113, + "theoretical_loss": 3.6175321574634847, + "tokens_seen": 1194196992 + }, + { + "epoch": 0.36, + "learning_rate": 0.0005156910732196589, + "loss": 1.4587, + "theoretical_loss": 3.6175140236073307, + "tokens_seen": 1194262528 + }, + { + "epoch": 0.36, + "learning_rate": 0.0005156750250752257, + "loss": 1.3465, + "theoretical_loss": 3.6174958910248707, + "tokens_seen": 1194328064 + }, + { + "epoch": 0.36, + "learning_rate": 0.0005156589769307923, + "loss": 1.4955, + "theoretical_loss": 3.6174777597159444, + "tokens_seen": 1194393600 + }, + { + "epoch": 0.36, + "learning_rate": 0.0005156429287863591, + "loss": 1.4928, + "theoretical_loss": 3.6174596296803934, + "tokens_seen": 1194459136 + }, + { + "epoch": 0.36, + "learning_rate": 0.0005156268806419257, + "loss": 1.3863, + "theoretical_loss": 3.6174415009180585, + "tokens_seen": 1194524672 + }, + { + "epoch": 0.36, + "learning_rate": 0.0005156108324974925, + "loss": 1.4752, + "theoretical_loss": 3.61742337342878, + "tokens_seen": 1194590208 + }, + { + "epoch": 0.36, + "learning_rate": 0.0005155947843530591, + "loss": 1.5141, + "theoretical_loss": 3.6174052472123983, + "tokens_seen": 1194655744 + }, + { + "epoch": 0.36, + "learning_rate": 0.0005155787362086259, + "loss": 1.4928, + "theoretical_loss": 3.617387122268755, + "tokens_seen": 1194721280 + }, + { + "epoch": 0.36, + "learning_rate": 0.0005155626880641925, + "loss": 1.4512, + "theoretical_loss": 3.6173689985976907, + "tokens_seen": 1194786816 + }, + { + "epoch": 0.36, + "learning_rate": 0.0005155466399197593, + "loss": 1.5859, + "theoretical_loss": 3.617350876199046, + "tokens_seen": 1194852352 + }, + { + "epoch": 0.36, + "learning_rate": 0.0005155305917753259, + "loss": 1.3627, + "theoretical_loss": 3.6173327550726624, + "tokens_seen": 1194917888 + }, + { + "epoch": 0.36, + "learning_rate": 0.0005155145436308927, + "loss": 1.4464, + "theoretical_loss": 3.6173146352183796, + "tokens_seen": 1194983424 + }, + { + "epoch": 0.36, + "learning_rate": 0.0005154984954864593, + "loss": 1.4253, + "theoretical_loss": 3.6172965166360402, + "tokens_seen": 1195048960 + }, + { + "epoch": 0.36, + "learning_rate": 0.0005154824473420261, + "loss": 1.4554, + "theoretical_loss": 3.6172783993254836, + "tokens_seen": 1195114496 + }, + { + "epoch": 0.36, + "learning_rate": 0.0005154663991975927, + "loss": 1.2441, + "theoretical_loss": 3.6172602832865515, + "tokens_seen": 1195180032 + }, + { + "epoch": 0.36, + "learning_rate": 0.0005154503510531595, + "loss": 1.2961, + "theoretical_loss": 3.617242168519085, + "tokens_seen": 1195245568 + }, + { + "epoch": 0.36, + "learning_rate": 0.0005154343029087261, + "loss": 1.56, + "theoretical_loss": 3.617224055022925, + "tokens_seen": 1195311104 + }, + { + "epoch": 0.36, + "learning_rate": 0.0005154182547642929, + "loss": 1.3596, + "theoretical_loss": 3.617205942797913, + "tokens_seen": 1195376640 + }, + { + "epoch": 0.36, + "learning_rate": 0.0005154022066198596, + "loss": 1.4331, + "theoretical_loss": 3.6171878318438893, + "tokens_seen": 1195442176 + }, + { + "epoch": 0.36, + "learning_rate": 0.0005153861584754263, + "loss": 1.4747, + "theoretical_loss": 3.6171697221606953, + "tokens_seen": 1195507712 + }, + { + "epoch": 0.36, + "learning_rate": 0.000515370110330993, + "loss": 1.5458, + "theoretical_loss": 3.617151613748173, + "tokens_seen": 1195573248 + }, + { + "epoch": 0.36, + "learning_rate": 0.0005153540621865597, + "loss": 1.3096, + "theoretical_loss": 3.6171335066061623, + "tokens_seen": 1195638784 + }, + { + "epoch": 0.36, + "learning_rate": 0.0005153380140421264, + "loss": 1.5023, + "theoretical_loss": 3.6171154007345057, + "tokens_seen": 1195704320 + }, + { + "epoch": 0.36, + "learning_rate": 0.0005153219658976931, + "loss": 1.4932, + "theoretical_loss": 3.617097296133043, + "tokens_seen": 1195769856 + }, + { + "epoch": 0.36, + "learning_rate": 0.0005153059177532599, + "loss": 1.375, + "theoretical_loss": 3.617079192801617, + "tokens_seen": 1195835392 + }, + { + "epoch": 0.36, + "learning_rate": 0.0005152898696088265, + "loss": 1.4397, + "theoretical_loss": 3.6170610907400684, + "tokens_seen": 1195900928 + }, + { + "epoch": 0.36, + "learning_rate": 0.0005152738214643933, + "loss": 1.3625, + "theoretical_loss": 3.617042989948238, + "tokens_seen": 1195966464 + }, + { + "epoch": 0.36, + "objective/train/docs_used": 437781, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.49967622756958, + "objective/train/theoretical_loss": 3.617024890425968, + "objective/train/tokens_used": 1216492000, + "theoretical_loss": 3.617024890425968, + "tokens_seen": 1196032000 + }, + { + "epoch": 0.36, + "learning_rate": 0.0005152577733199599, + "loss": 1.5269, + "theoretical_loss": 3.617024890425968, + "tokens_seen": 1196032000 + }, + { + "epoch": 0.36, + "learning_rate": 0.0005152417251755267, + "loss": 1.4228, + "theoretical_loss": 3.617006792173099, + "tokens_seen": 1196097536 + }, + { + "epoch": 0.36, + "learning_rate": 0.0005152256770310933, + "loss": 1.3409, + "theoretical_loss": 3.616988695189473, + "tokens_seen": 1196163072 + }, + { + "epoch": 0.36, + "learning_rate": 0.0005152096288866601, + "loss": 1.4471, + "theoretical_loss": 3.6169705994749317, + "tokens_seen": 1196228608 + }, + { + "epoch": 0.36, + "learning_rate": 0.0005151935807422267, + "loss": 1.4061, + "theoretical_loss": 3.6169525050293156, + "tokens_seen": 1196294144 + }, + { + "epoch": 0.36, + "learning_rate": 0.0005151775325977935, + "loss": 1.5513, + "theoretical_loss": 3.6169344118524673, + "tokens_seen": 1196359680 + }, + { + "epoch": 0.36, + "learning_rate": 0.0005151614844533601, + "loss": 1.4039, + "theoretical_loss": 3.616916319944228, + "tokens_seen": 1196425216 + }, + { + "epoch": 0.36, + "learning_rate": 0.0005151454363089269, + "loss": 1.4364, + "theoretical_loss": 3.6168982293044385, + "tokens_seen": 1196490752 + }, + { + "epoch": 0.36, + "learning_rate": 0.0005151293881644935, + "loss": 1.4523, + "theoretical_loss": 3.6168801399329413, + "tokens_seen": 1196556288 + }, + { + "epoch": 0.36, + "learning_rate": 0.0005151133400200603, + "loss": 1.514, + "theoretical_loss": 3.6168620518295778, + "tokens_seen": 1196621824 + }, + { + "epoch": 0.36, + "learning_rate": 0.0005150972918756269, + "loss": 1.3742, + "theoretical_loss": 3.6168439649941897, + "tokens_seen": 1196687360 + }, + { + "epoch": 0.36, + "learning_rate": 0.0005150812437311937, + "loss": 1.4393, + "theoretical_loss": 3.616825879426618, + "tokens_seen": 1196752896 + }, + { + "epoch": 0.36, + "learning_rate": 0.0005150651955867603, + "loss": 1.3575, + "theoretical_loss": 3.6168077951267055, + "tokens_seen": 1196818432 + }, + { + "epoch": 0.36, + "learning_rate": 0.0005150491474423271, + "loss": 1.4963, + "theoretical_loss": 3.6167897120942936, + "tokens_seen": 1196883968 + }, + { + "epoch": 0.36, + "learning_rate": 0.0005150330992978937, + "loss": 1.3686, + "theoretical_loss": 3.6167716303292234, + "tokens_seen": 1196949504 + }, + { + "epoch": 0.36, + "learning_rate": 0.0005150170511534605, + "loss": 1.3856, + "theoretical_loss": 3.616753549831337, + "tokens_seen": 1197015040 + }, + { + "epoch": 0.36, + "learning_rate": 0.0005150010030090271, + "loss": 1.5711, + "theoretical_loss": 3.616735470600477, + "tokens_seen": 1197080576 + }, + { + "epoch": 0.36, + "learning_rate": 0.0005149849548645939, + "loss": 1.4146, + "theoretical_loss": 3.6167173926364846, + "tokens_seen": 1197146112 + }, + { + "epoch": 0.36, + "learning_rate": 0.0005149689067201605, + "loss": 1.5542, + "theoretical_loss": 3.6166993159392016, + "tokens_seen": 1197211648 + }, + { + "epoch": 0.36, + "learning_rate": 0.0005149528585757273, + "loss": 1.3927, + "theoretical_loss": 3.61668124050847, + "tokens_seen": 1197277184 + }, + { + "epoch": 0.36, + "learning_rate": 0.0005149368104312939, + "loss": 1.4153, + "theoretical_loss": 3.6166631663441318, + "tokens_seen": 1197342720 + }, + { + "epoch": 0.36, + "learning_rate": 0.0005149207622868607, + "loss": 1.4833, + "theoretical_loss": 3.616645093446029, + "tokens_seen": 1197408256 + }, + { + "epoch": 0.36, + "learning_rate": 0.0005149047141424273, + "loss": 1.3835, + "theoretical_loss": 3.6166270218140033, + "tokens_seen": 1197473792 + }, + { + "epoch": 0.36, + "learning_rate": 0.0005148886659979941, + "loss": 1.3764, + "theoretical_loss": 3.6166089514478976, + "tokens_seen": 1197539328 + }, + { + "epoch": 0.36, + "learning_rate": 0.0005148726178535607, + "loss": 1.5036, + "theoretical_loss": 3.6165908823475528, + "tokens_seen": 1197604864 + }, + { + "epoch": 0.36, + "learning_rate": 0.0005148565697091275, + "loss": 1.361, + "theoretical_loss": 3.6165728145128115, + "tokens_seen": 1197670400 + }, + { + "epoch": 0.36, + "learning_rate": 0.0005148405215646941, + "loss": 1.5588, + "theoretical_loss": 3.616554747943516, + "tokens_seen": 1197735936 + }, + { + "epoch": 0.36, + "learning_rate": 0.0005148244734202609, + "loss": 1.4734, + "theoretical_loss": 3.616536682639508, + "tokens_seen": 1197801472 + }, + { + "epoch": 0.36, + "learning_rate": 0.0005148084252758275, + "loss": 1.3862, + "theoretical_loss": 3.6165186186006304, + "tokens_seen": 1197867008 + }, + { + "epoch": 0.36, + "learning_rate": 0.0005147923771313943, + "loss": 1.4878, + "theoretical_loss": 3.6165005558267245, + "tokens_seen": 1197932544 + }, + { + "epoch": 0.36, + "learning_rate": 0.0005147763289869609, + "loss": 1.4628, + "theoretical_loss": 3.616482494317633, + "tokens_seen": 1197998080 + }, + { + "epoch": 0.36, + "learning_rate": 0.0005147602808425277, + "loss": 1.385, + "theoretical_loss": 3.6164644340731984, + "tokens_seen": 1198063616 + }, + { + "epoch": 0.36, + "learning_rate": 0.0005147442326980943, + "loss": 1.4119, + "theoretical_loss": 3.616446375093262, + "tokens_seen": 1198129152 + }, + { + "epoch": 0.36, + "learning_rate": 0.000514728184553661, + "loss": 1.3137, + "theoretical_loss": 3.6164283173776672, + "tokens_seen": 1198194688 + }, + { + "epoch": 0.36, + "learning_rate": 0.0005147121364092277, + "loss": 1.4049, + "theoretical_loss": 3.616410260926256, + "tokens_seen": 1198260224 + }, + { + "epoch": 0.36, + "learning_rate": 0.0005146960882647944, + "loss": 1.3673, + "theoretical_loss": 3.6163922057388707, + "tokens_seen": 1198325760 + }, + { + "epoch": 0.36, + "learning_rate": 0.0005146800401203611, + "loss": 1.3993, + "theoretical_loss": 3.6163741518153536, + "tokens_seen": 1198391296 + }, + { + "epoch": 0.36, + "learning_rate": 0.0005146639919759278, + "loss": 1.3597, + "theoretical_loss": 3.6163560991555466, + "tokens_seen": 1198456832 + }, + { + "epoch": 0.36, + "learning_rate": 0.0005146479438314945, + "loss": 1.3259, + "theoretical_loss": 3.6163380477592932, + "tokens_seen": 1198522368 + }, + { + "epoch": 0.36, + "learning_rate": 0.0005146318956870612, + "loss": 1.4536, + "theoretical_loss": 3.6163199976264355, + "tokens_seen": 1198587904 + }, + { + "epoch": 0.36, + "learning_rate": 0.0005146158475426279, + "loss": 1.3528, + "theoretical_loss": 3.6163019487568158, + "tokens_seen": 1198653440 + }, + { + "epoch": 0.36, + "learning_rate": 0.0005145997993981946, + "loss": 1.4438, + "theoretical_loss": 3.6162839011502763, + "tokens_seen": 1198718976 + }, + { + "epoch": 0.36, + "learning_rate": 0.0005145837512537612, + "loss": 1.2983, + "theoretical_loss": 3.6162658548066604, + "tokens_seen": 1198784512 + }, + { + "epoch": 0.36, + "learning_rate": 0.000514567703109328, + "loss": 1.3671, + "theoretical_loss": 3.6162478097258104, + "tokens_seen": 1198850048 + }, + { + "epoch": 0.36, + "learning_rate": 0.0005145516549648946, + "loss": 1.533, + "theoretical_loss": 3.6162297659075686, + "tokens_seen": 1198915584 + }, + { + "epoch": 0.36, + "learning_rate": 0.0005145356068204614, + "loss": 1.3687, + "theoretical_loss": 3.616211723351778, + "tokens_seen": 1198981120 + }, + { + "epoch": 0.36, + "learning_rate": 0.000514519558676028, + "loss": 1.3659, + "theoretical_loss": 3.616193682058281, + "tokens_seen": 1199046656 + }, + { + "epoch": 0.36, + "learning_rate": 0.0005145035105315948, + "loss": 1.3964, + "theoretical_loss": 3.6161756420269207, + "tokens_seen": 1199112192 + }, + { + "epoch": 0.36, + "learning_rate": 0.0005144874623871614, + "loss": 1.4564, + "theoretical_loss": 3.6161576032575393, + "tokens_seen": 1199177728 + }, + { + "epoch": 0.36, + "learning_rate": 0.0005144714142427282, + "loss": 1.2833, + "theoretical_loss": 3.6161395657499797, + "tokens_seen": 1199243264 + }, + { + "epoch": 0.36, + "objective/train/docs_used": 438488, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.521169900894165, + "objective/train/theoretical_loss": 3.6161215295040847, + "objective/train/tokens_used": 1219768800, + "theoretical_loss": 3.6161215295040847, + "tokens_seen": 1199308800 + }, + { + "epoch": 0.36, + "learning_rate": 0.0005144553660982948, + "loss": 1.4658, + "theoretical_loss": 3.6161215295040847, + "tokens_seen": 1199308800 + }, + { + "epoch": 0.36, + "learning_rate": 0.0005144393179538616, + "loss": 1.3846, + "theoretical_loss": 3.6161034945196975, + "tokens_seen": 1199374336 + }, + { + "epoch": 0.36, + "learning_rate": 0.0005144232698094282, + "loss": 1.3941, + "theoretical_loss": 3.6160854607966604, + "tokens_seen": 1199439872 + }, + { + "epoch": 0.36, + "learning_rate": 0.000514407221664995, + "loss": 1.46, + "theoretical_loss": 3.6160674283348166, + "tokens_seen": 1199505408 + }, + { + "epoch": 0.36, + "learning_rate": 0.0005143911735205616, + "loss": 1.3304, + "theoretical_loss": 3.616049397134009, + "tokens_seen": 1199570944 + }, + { + "epoch": 0.36, + "learning_rate": 0.0005143751253761284, + "loss": 1.4486, + "theoretical_loss": 3.6160313671940805, + "tokens_seen": 1199636480 + }, + { + "epoch": 0.36, + "learning_rate": 0.000514359077231695, + "loss": 1.4713, + "theoretical_loss": 3.6160133385148736, + "tokens_seen": 1199702016 + }, + { + "epoch": 0.36, + "learning_rate": 0.0005143430290872618, + "loss": 1.3671, + "theoretical_loss": 3.615995311096232, + "tokens_seen": 1199767552 + }, + { + "epoch": 0.36, + "learning_rate": 0.0005143269809428284, + "loss": 1.2871, + "theoretical_loss": 3.6159772849379985, + "tokens_seen": 1199833088 + }, + { + "epoch": 0.36, + "learning_rate": 0.0005143109327983952, + "loss": 1.4766, + "theoretical_loss": 3.6159592600400163, + "tokens_seen": 1199898624 + }, + { + "epoch": 0.36, + "learning_rate": 0.0005142948846539618, + "loss": 1.3327, + "theoretical_loss": 3.6159412364021275, + "tokens_seen": 1199964160 + }, + { + "epoch": 0.36, + "learning_rate": 0.0005142788365095286, + "loss": 1.4277, + "theoretical_loss": 3.6159232140241766, + "tokens_seen": 1200029696 + }, + { + "epoch": 0.36, + "learning_rate": 0.0005142627883650952, + "loss": 1.4793, + "theoretical_loss": 3.6159051929060055, + "tokens_seen": 1200095232 + }, + { + "epoch": 0.36, + "learning_rate": 0.000514246740220662, + "loss": 1.5387, + "theoretical_loss": 3.615887173047458, + "tokens_seen": 1200160768 + }, + { + "epoch": 0.36, + "learning_rate": 0.0005142306920762286, + "loss": 1.5112, + "theoretical_loss": 3.6158691544483776, + "tokens_seen": 1200226304 + }, + { + "epoch": 0.36, + "learning_rate": 0.0005142146439317954, + "loss": 1.4166, + "theoretical_loss": 3.6158511371086064, + "tokens_seen": 1200291840 + }, + { + "epoch": 0.36, + "learning_rate": 0.000514198595787362, + "loss": 1.4443, + "theoretical_loss": 3.615833121027989, + "tokens_seen": 1200357376 + }, + { + "epoch": 0.36, + "learning_rate": 0.0005141825476429288, + "loss": 1.3879, + "theoretical_loss": 3.6158151062063673, + "tokens_seen": 1200422912 + }, + { + "epoch": 0.36, + "learning_rate": 0.0005141664994984954, + "loss": 1.43, + "theoretical_loss": 3.6157970926435854, + "tokens_seen": 1200488448 + }, + { + "epoch": 0.36, + "learning_rate": 0.0005141504513540622, + "loss": 1.3094, + "theoretical_loss": 3.615779080339487, + "tokens_seen": 1200553984 + }, + { + "epoch": 0.36, + "learning_rate": 0.0005141344032096288, + "loss": 1.3876, + "theoretical_loss": 3.615761069293914, + "tokens_seen": 1200619520 + }, + { + "epoch": 0.36, + "learning_rate": 0.0005141183550651956, + "loss": 1.3161, + "theoretical_loss": 3.6157430595067113, + "tokens_seen": 1200685056 + }, + { + "epoch": 0.36, + "learning_rate": 0.0005141023069207624, + "loss": 1.431, + "theoretical_loss": 3.615725050977722, + "tokens_seen": 1200750592 + }, + { + "epoch": 0.36, + "learning_rate": 0.000514086258776329, + "loss": 1.4039, + "theoretical_loss": 3.6157070437067884, + "tokens_seen": 1200816128 + }, + { + "epoch": 0.36, + "learning_rate": 0.0005140702106318958, + "loss": 1.4032, + "theoretical_loss": 3.6156890376937554, + "tokens_seen": 1200881664 + }, + { + "epoch": 0.36, + "learning_rate": 0.0005140541624874624, + "loss": 1.5436, + "theoretical_loss": 3.615671032938466, + "tokens_seen": 1200947200 + }, + { + "epoch": 0.36, + "learning_rate": 0.0005140381143430292, + "loss": 1.5235, + "theoretical_loss": 3.615653029440763, + "tokens_seen": 1201012736 + }, + { + "epoch": 0.36, + "learning_rate": 0.0005140220661985958, + "loss": 1.3621, + "theoretical_loss": 3.615635027200491, + "tokens_seen": 1201078272 + }, + { + "epoch": 0.36, + "learning_rate": 0.0005140060180541626, + "loss": 1.3458, + "theoretical_loss": 3.6156170262174934, + "tokens_seen": 1201143808 + }, + { + "epoch": 0.36, + "learning_rate": 0.0005139899699097292, + "loss": 1.4751, + "theoretical_loss": 3.615599026491613, + "tokens_seen": 1201209344 + }, + { + "epoch": 0.36, + "learning_rate": 0.000513973921765296, + "loss": 1.371, + "theoretical_loss": 3.615581028022694, + "tokens_seen": 1201274880 + }, + { + "epoch": 0.36, + "learning_rate": 0.0005139578736208626, + "loss": 1.4517, + "theoretical_loss": 3.61556303081058, + "tokens_seen": 1201340416 + }, + { + "epoch": 0.36, + "learning_rate": 0.0005139418254764294, + "loss": 1.5642, + "theoretical_loss": 3.6155450348551152, + "tokens_seen": 1201405952 + }, + { + "epoch": 0.36, + "learning_rate": 0.000513925777331996, + "loss": 1.4107, + "theoretical_loss": 3.6155270401561426, + "tokens_seen": 1201471488 + }, + { + "epoch": 0.36, + "learning_rate": 0.0005139097291875628, + "loss": 1.462, + "theoretical_loss": 3.615509046713506, + "tokens_seen": 1201537024 + }, + { + "epoch": 0.36, + "learning_rate": 0.0005138936810431294, + "loss": 1.3945, + "theoretical_loss": 3.615491054527049, + "tokens_seen": 1201602560 + }, + { + "epoch": 0.36, + "learning_rate": 0.0005138776328986962, + "loss": 1.3776, + "theoretical_loss": 3.615473063596616, + "tokens_seen": 1201668096 + }, + { + "epoch": 0.36, + "learning_rate": 0.0005138615847542628, + "loss": 1.4394, + "theoretical_loss": 3.6154550739220506, + "tokens_seen": 1201733632 + }, + { + "epoch": 0.36, + "learning_rate": 0.0005138455366098296, + "loss": 1.475, + "theoretical_loss": 3.6154370855031965, + "tokens_seen": 1201799168 + }, + { + "epoch": 0.36, + "learning_rate": 0.0005138294884653962, + "loss": 1.5095, + "theoretical_loss": 3.6154190983398977, + "tokens_seen": 1201864704 + }, + { + "epoch": 0.36, + "learning_rate": 0.000513813440320963, + "loss": 1.3031, + "theoretical_loss": 3.6154011124319982, + "tokens_seen": 1201930240 + }, + { + "epoch": 0.36, + "learning_rate": 0.0005137973921765296, + "loss": 1.3942, + "theoretical_loss": 3.615383127779342, + "tokens_seen": 1201995776 + }, + { + "epoch": 0.36, + "learning_rate": 0.0005137813440320964, + "loss": 1.424, + "theoretical_loss": 3.6153651443817725, + "tokens_seen": 1202061312 + }, + { + "epoch": 0.36, + "learning_rate": 0.000513765295887663, + "loss": 1.3737, + "theoretical_loss": 3.615347162239134, + "tokens_seen": 1202126848 + }, + { + "epoch": 0.36, + "learning_rate": 0.0005137492477432298, + "loss": 1.486, + "theoretical_loss": 3.615329181351271, + "tokens_seen": 1202192384 + }, + { + "epoch": 0.36, + "learning_rate": 0.0005137331995987964, + "loss": 1.3504, + "theoretical_loss": 3.6153112017180264, + "tokens_seen": 1202257920 + }, + { + "epoch": 0.36, + "learning_rate": 0.0005137171514543632, + "loss": 1.3312, + "theoretical_loss": 3.615293223339246, + "tokens_seen": 1202323456 + }, + { + "epoch": 0.36, + "learning_rate": 0.0005137011033099298, + "loss": 1.3996, + "theoretical_loss": 3.6152752462147726, + "tokens_seen": 1202388992 + }, + { + "epoch": 0.36, + "learning_rate": 0.0005136850551654966, + "loss": 1.43, + "theoretical_loss": 3.6152572703444505, + "tokens_seen": 1202454528 + }, + { + "epoch": 0.36, + "learning_rate": 0.0005136690070210632, + "loss": 1.377, + "theoretical_loss": 3.6152392957281236, + "tokens_seen": 1202520064 + }, + { + "epoch": 0.36, + "objective/train/docs_used": 439824, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.4753193855285645, + "objective/train/theoretical_loss": 3.6152213223656373, + "objective/train/tokens_used": 1223045600, + "theoretical_loss": 3.6152213223656373, + "tokens_seen": 1202585600 + }, + { + "epoch": 0.36, + "learning_rate": 0.00051365295887663, + "loss": 1.317, + "theoretical_loss": 3.6152213223656373, + "tokens_seen": 1202585600 + }, + { + "epoch": 0.36, + "learning_rate": 0.0005136369107321966, + "loss": 1.328, + "theoretical_loss": 3.6152033502568344, + "tokens_seen": 1202651136 + }, + { + "epoch": 0.36, + "learning_rate": 0.0005136208625877634, + "loss": 1.3111, + "theoretical_loss": 3.6151853794015603, + "tokens_seen": 1202716672 + }, + { + "epoch": 0.36, + "learning_rate": 0.00051360481444333, + "loss": 1.3888, + "theoretical_loss": 3.615167409799658, + "tokens_seen": 1202782208 + }, + { + "epoch": 0.36, + "learning_rate": 0.0005135887662988968, + "loss": 1.436, + "theoretical_loss": 3.6151494414509733, + "tokens_seen": 1202847744 + }, + { + "epoch": 0.36, + "learning_rate": 0.0005135727181544634, + "loss": 1.5027, + "theoretical_loss": 3.6151314743553495, + "tokens_seen": 1202913280 + }, + { + "epoch": 0.36, + "learning_rate": 0.0005135566700100302, + "loss": 1.408, + "theoretical_loss": 3.615113508512631, + "tokens_seen": 1202978816 + }, + { + "epoch": 0.36, + "learning_rate": 0.0005135406218655968, + "loss": 1.3852, + "theoretical_loss": 3.615095543922662, + "tokens_seen": 1203044352 + }, + { + "epoch": 0.36, + "learning_rate": 0.0005135245737211636, + "loss": 1.3954, + "theoretical_loss": 3.6150775805852877, + "tokens_seen": 1203109888 + }, + { + "epoch": 0.36, + "learning_rate": 0.0005135085255767302, + "loss": 1.3582, + "theoretical_loss": 3.6150596185003527, + "tokens_seen": 1203175424 + }, + { + "epoch": 0.36, + "learning_rate": 0.000513492477432297, + "loss": 1.3958, + "theoretical_loss": 3.6150416576677005, + "tokens_seen": 1203240960 + }, + { + "epoch": 0.36, + "learning_rate": 0.0005134764292878636, + "loss": 1.4242, + "theoretical_loss": 3.6150236980871755, + "tokens_seen": 1203306496 + }, + { + "epoch": 0.36, + "learning_rate": 0.0005134603811434304, + "loss": 1.3551, + "theoretical_loss": 3.6150057397586233, + "tokens_seen": 1203372032 + }, + { + "epoch": 0.36, + "learning_rate": 0.000513444332998997, + "loss": 1.5107, + "theoretical_loss": 3.6149877826818875, + "tokens_seen": 1203437568 + }, + { + "epoch": 0.36, + "learning_rate": 0.0005134282848545638, + "loss": 1.4291, + "theoretical_loss": 3.6149698268568136, + "tokens_seen": 1203503104 + }, + { + "epoch": 0.36, + "learning_rate": 0.0005134122367101304, + "loss": 1.4683, + "theoretical_loss": 3.614951872283245, + "tokens_seen": 1203568640 + }, + { + "epoch": 0.36, + "learning_rate": 0.0005133961885656972, + "loss": 1.3471, + "theoretical_loss": 3.614933918961027, + "tokens_seen": 1203634176 + }, + { + "epoch": 0.36, + "learning_rate": 0.0005133801404212638, + "loss": 1.5165, + "theoretical_loss": 3.614915966890005, + "tokens_seen": 1203699712 + }, + { + "epoch": 0.36, + "learning_rate": 0.0005133640922768306, + "loss": 1.3932, + "theoretical_loss": 3.614898016070022, + "tokens_seen": 1203765248 + }, + { + "epoch": 0.36, + "learning_rate": 0.0005133480441323972, + "loss": 1.3642, + "theoretical_loss": 3.614880066500924, + "tokens_seen": 1203830784 + }, + { + "epoch": 0.36, + "learning_rate": 0.0005133319959879639, + "loss": 1.3798, + "theoretical_loss": 3.6148621181825558, + "tokens_seen": 1203896320 + }, + { + "epoch": 0.36, + "learning_rate": 0.0005133159478435306, + "loss": 1.3303, + "theoretical_loss": 3.614844171114761, + "tokens_seen": 1203961856 + }, + { + "epoch": 0.36, + "learning_rate": 0.0005132998996990973, + "loss": 1.4495, + "theoretical_loss": 3.614826225297386, + "tokens_seen": 1204027392 + }, + { + "epoch": 0.36, + "learning_rate": 0.000513283851554664, + "loss": 1.339, + "theoretical_loss": 3.614808280730274, + "tokens_seen": 1204092928 + }, + { + "epoch": 0.36, + "learning_rate": 0.0005132678034102307, + "loss": 1.4388, + "theoretical_loss": 3.614790337413271, + "tokens_seen": 1204158464 + }, + { + "epoch": 0.36, + "learning_rate": 0.0005132517552657974, + "loss": 1.4505, + "theoretical_loss": 3.6147723953462214, + "tokens_seen": 1204224000 + }, + { + "epoch": 0.36, + "learning_rate": 0.0005132357071213641, + "loss": 1.561, + "theoretical_loss": 3.61475445452897, + "tokens_seen": 1204289536 + }, + { + "epoch": 0.36, + "learning_rate": 0.0005132196589769308, + "loss": 1.4148, + "theoretical_loss": 3.614736514961362, + "tokens_seen": 1204355072 + }, + { + "epoch": 0.36, + "learning_rate": 0.0005132036108324975, + "loss": 1.4749, + "theoretical_loss": 3.6147185766432433, + "tokens_seen": 1204420608 + }, + { + "epoch": 0.36, + "learning_rate": 0.0005131875626880641, + "loss": 1.4745, + "theoretical_loss": 3.6147006395744565, + "tokens_seen": 1204486144 + }, + { + "epoch": 0.37, + "learning_rate": 0.0005131715145436309, + "loss": 1.3314, + "theoretical_loss": 3.6146827037548492, + "tokens_seen": 1204551680 + }, + { + "epoch": 0.37, + "learning_rate": 0.0005131554663991975, + "loss": 1.3986, + "theoretical_loss": 3.6146647691842646, + "tokens_seen": 1204617216 + }, + { + "epoch": 0.37, + "learning_rate": 0.0005131394182547643, + "loss": 1.5669, + "theoretical_loss": 3.6146468358625485, + "tokens_seen": 1204682752 + }, + { + "epoch": 0.37, + "learning_rate": 0.0005131233701103309, + "loss": 1.4165, + "theoretical_loss": 3.614628903789546, + "tokens_seen": 1204748288 + }, + { + "epoch": 0.37, + "learning_rate": 0.0005131073219658977, + "loss": 1.2732, + "theoretical_loss": 3.614610972965102, + "tokens_seen": 1204813824 + }, + { + "epoch": 0.37, + "learning_rate": 0.0005130912738214643, + "loss": 1.5334, + "theoretical_loss": 3.6145930433890623, + "tokens_seen": 1204879360 + }, + { + "epoch": 0.37, + "learning_rate": 0.0005130752256770311, + "loss": 1.5201, + "theoretical_loss": 3.6145751150612715, + "tokens_seen": 1204944896 + }, + { + "epoch": 0.37, + "learning_rate": 0.0005130591775325977, + "loss": 1.3848, + "theoretical_loss": 3.6145571879815748, + "tokens_seen": 1205010432 + }, + { + "epoch": 0.37, + "learning_rate": 0.0005130431293881645, + "loss": 1.4096, + "theoretical_loss": 3.614539262149818, + "tokens_seen": 1205075968 + }, + { + "epoch": 0.37, + "learning_rate": 0.0005130270812437311, + "loss": 1.5151, + "theoretical_loss": 3.6145213375658454, + "tokens_seen": 1205141504 + }, + { + "epoch": 0.37, + "learning_rate": 0.0005130110330992979, + "loss": 1.3458, + "theoretical_loss": 3.614503414229503, + "tokens_seen": 1205207040 + }, + { + "epoch": 0.37, + "learning_rate": 0.0005129949849548645, + "loss": 1.4243, + "theoretical_loss": 3.614485492140636, + "tokens_seen": 1205272576 + }, + { + "epoch": 0.37, + "learning_rate": 0.0005129789368104313, + "loss": 1.469, + "theoretical_loss": 3.61446757129909, + "tokens_seen": 1205338112 + }, + { + "epoch": 0.37, + "learning_rate": 0.0005129628886659979, + "loss": 1.4467, + "theoretical_loss": 3.6144496517047098, + "tokens_seen": 1205403648 + }, + { + "epoch": 0.37, + "learning_rate": 0.0005129468405215647, + "loss": 1.4324, + "theoretical_loss": 3.6144317333573412, + "tokens_seen": 1205469184 + }, + { + "epoch": 0.37, + "learning_rate": 0.0005129307923771313, + "loss": 1.3645, + "theoretical_loss": 3.6144138162568296, + "tokens_seen": 1205534720 + }, + { + "epoch": 0.37, + "learning_rate": 0.0005129147442326981, + "loss": 1.4677, + "theoretical_loss": 3.6143959004030206, + "tokens_seen": 1205600256 + }, + { + "epoch": 0.37, + "learning_rate": 0.0005128986960882648, + "loss": 1.3801, + "theoretical_loss": 3.614377985795759, + "tokens_seen": 1205665792 + }, + { + "epoch": 0.37, + "learning_rate": 0.0005128826479438315, + "loss": 1.3352, + "theoretical_loss": 3.6143600724348914, + "tokens_seen": 1205731328 + }, + { + "epoch": 0.37, + "learning_rate": 0.0005128665997993983, + "loss": 1.4956, + "theoretical_loss": 3.614342160320262, + "tokens_seen": 1205796864 + }, + { + "epoch": 0.37, + "objective/train/docs_used": 441116, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.4246854782104492, + "objective/train/theoretical_loss": 3.614324249451718, + "objective/train/tokens_used": 1226322400, + "theoretical_loss": 3.614324249451718, + "tokens_seen": 1205862400 + }, + { + "epoch": 0.37, + "learning_rate": 0.0005128505516549649, + "loss": 1.4393, + "theoretical_loss": 3.614324249451718, + "tokens_seen": 1205862400 + }, + { + "epoch": 0.37, + "learning_rate": 0.0005128345035105317, + "loss": 1.5369, + "theoretical_loss": 3.6143063398291035, + "tokens_seen": 1205927936 + }, + { + "epoch": 0.37, + "learning_rate": 0.0005128184553660983, + "loss": 1.5202, + "theoretical_loss": 3.6142884314522647, + "tokens_seen": 1205993472 + }, + { + "epoch": 0.37, + "learning_rate": 0.000512802407221665, + "loss": 1.4713, + "theoretical_loss": 3.6142705243210473, + "tokens_seen": 1206059008 + }, + { + "epoch": 0.37, + "learning_rate": 0.0005127863590772317, + "loss": 1.3982, + "theoretical_loss": 3.614252618435297, + "tokens_seen": 1206124544 + }, + { + "epoch": 0.37, + "learning_rate": 0.0005127703109327985, + "loss": 1.4535, + "theoretical_loss": 3.6142347137948603, + "tokens_seen": 1206190080 + }, + { + "epoch": 0.37, + "learning_rate": 0.0005127542627883651, + "loss": 1.4863, + "theoretical_loss": 3.614216810399581, + "tokens_seen": 1206255616 + }, + { + "epoch": 0.37, + "learning_rate": 0.0005127382146439319, + "loss": 1.4365, + "theoretical_loss": 3.614198908249307, + "tokens_seen": 1206321152 + }, + { + "epoch": 0.37, + "learning_rate": 0.0005127221664994985, + "loss": 1.45, + "theoretical_loss": 3.6141810073438823, + "tokens_seen": 1206386688 + }, + { + "epoch": 0.37, + "learning_rate": 0.0005127061183550653, + "loss": 1.4629, + "theoretical_loss": 3.6141631076831544, + "tokens_seen": 1206452224 + }, + { + "epoch": 0.37, + "learning_rate": 0.0005126900702106319, + "loss": 1.3719, + "theoretical_loss": 3.6141452092669675, + "tokens_seen": 1206517760 + }, + { + "epoch": 0.37, + "learning_rate": 0.0005126740220661987, + "loss": 1.2715, + "theoretical_loss": 3.6141273120951682, + "tokens_seen": 1206583296 + }, + { + "epoch": 0.37, + "learning_rate": 0.0005126579739217653, + "loss": 1.4186, + "theoretical_loss": 3.6141094161676026, + "tokens_seen": 1206648832 + }, + { + "epoch": 0.37, + "learning_rate": 0.0005126419257773321, + "loss": 1.6038, + "theoretical_loss": 3.614091521484117, + "tokens_seen": 1206714368 + }, + { + "epoch": 0.37, + "learning_rate": 0.0005126258776328987, + "loss": 1.2878, + "theoretical_loss": 3.6140736280445562, + "tokens_seen": 1206779904 + }, + { + "epoch": 0.37, + "learning_rate": 0.0005126098294884655, + "loss": 1.2947, + "theoretical_loss": 3.6140557358487673, + "tokens_seen": 1206845440 + }, + { + "epoch": 0.37, + "learning_rate": 0.0005125937813440321, + "loss": 1.4041, + "theoretical_loss": 3.614037844896595, + "tokens_seen": 1206910976 + }, + { + "epoch": 0.37, + "learning_rate": 0.0005125777331995989, + "loss": 1.4654, + "theoretical_loss": 3.614019955187887, + "tokens_seen": 1206976512 + }, + { + "epoch": 0.37, + "learning_rate": 0.0005125616850551655, + "loss": 1.3792, + "theoretical_loss": 3.614002066722488, + "tokens_seen": 1207042048 + }, + { + "epoch": 0.37, + "learning_rate": 0.0005125456369107323, + "loss": 1.5332, + "theoretical_loss": 3.6139841795002448, + "tokens_seen": 1207107584 + }, + { + "epoch": 0.37, + "learning_rate": 0.0005125295887662989, + "loss": 1.4633, + "theoretical_loss": 3.6139662935210035, + "tokens_seen": 1207173120 + }, + { + "epoch": 0.37, + "learning_rate": 0.0005125135406218657, + "loss": 1.4871, + "theoretical_loss": 3.6139484087846094, + "tokens_seen": 1207238656 + }, + { + "epoch": 0.37, + "learning_rate": 0.0005124974924774323, + "loss": 1.4589, + "theoretical_loss": 3.61393052529091, + "tokens_seen": 1207304192 + }, + { + "epoch": 0.37, + "learning_rate": 0.0005124814443329991, + "loss": 1.4872, + "theoretical_loss": 3.6139126430397512, + "tokens_seen": 1207369728 + }, + { + "epoch": 0.37, + "learning_rate": 0.0005124653961885657, + "loss": 1.5083, + "theoretical_loss": 3.613894762030978, + "tokens_seen": 1207435264 + }, + { + "epoch": 0.37, + "learning_rate": 0.0005124493480441325, + "loss": 1.3808, + "theoretical_loss": 3.613876882264438, + "tokens_seen": 1207500800 + }, + { + "epoch": 0.37, + "learning_rate": 0.0005124332998996991, + "loss": 1.4098, + "theoretical_loss": 3.613859003739977, + "tokens_seen": 1207566336 + }, + { + "epoch": 0.37, + "learning_rate": 0.0005124172517552659, + "loss": 1.3832, + "theoretical_loss": 3.6138411264574413, + "tokens_seen": 1207631872 + }, + { + "epoch": 0.37, + "learning_rate": 0.0005124012036108325, + "loss": 1.4761, + "theoretical_loss": 3.6138232504166776, + "tokens_seen": 1207697408 + }, + { + "epoch": 0.37, + "learning_rate": 0.0005123851554663993, + "loss": 1.536, + "theoretical_loss": 3.613805375617532, + "tokens_seen": 1207762944 + }, + { + "epoch": 0.37, + "learning_rate": 0.0005123691073219659, + "loss": 1.5428, + "theoretical_loss": 3.61378750205985, + "tokens_seen": 1207828480 + }, + { + "epoch": 0.37, + "learning_rate": 0.0005123530591775327, + "loss": 1.4598, + "theoretical_loss": 3.613769629743479, + "tokens_seen": 1207894016 + }, + { + "epoch": 0.37, + "learning_rate": 0.0005123370110330993, + "loss": 1.4595, + "theoretical_loss": 3.6137517586682666, + "tokens_seen": 1207959552 + }, + { + "epoch": 0.37, + "learning_rate": 0.0005123209628886661, + "loss": 1.4262, + "theoretical_loss": 3.613733888834057, + "tokens_seen": 1208025088 + }, + { + "epoch": 0.37, + "learning_rate": 0.0005123049147442327, + "loss": 1.4659, + "theoretical_loss": 3.6137160202406977, + "tokens_seen": 1208090624 + }, + { + "epoch": 0.37, + "learning_rate": 0.0005122888665997995, + "loss": 1.3759, + "theoretical_loss": 3.6136981528880354, + "tokens_seen": 1208156160 + }, + { + "epoch": 0.37, + "learning_rate": 0.0005122728184553661, + "loss": 1.4967, + "theoretical_loss": 3.6136802867759164, + "tokens_seen": 1208221696 + }, + { + "epoch": 0.37, + "learning_rate": 0.0005122567703109329, + "loss": 1.462, + "theoretical_loss": 3.6136624219041877, + "tokens_seen": 1208287232 + }, + { + "epoch": 0.37, + "learning_rate": 0.0005122407221664995, + "loss": 1.2891, + "theoretical_loss": 3.613644558272695, + "tokens_seen": 1208352768 + }, + { + "epoch": 0.37, + "learning_rate": 0.0005122246740220663, + "loss": 1.4593, + "theoretical_loss": 3.613626695881286, + "tokens_seen": 1208418304 + }, + { + "epoch": 0.37, + "learning_rate": 0.0005122086258776329, + "loss": 1.5449, + "theoretical_loss": 3.6136088347298068, + "tokens_seen": 1208483840 + }, + { + "epoch": 0.37, + "learning_rate": 0.0005121925777331997, + "loss": 1.4402, + "theoretical_loss": 3.613590974818104, + "tokens_seen": 1208549376 + }, + { + "epoch": 0.37, + "learning_rate": 0.0005121765295887663, + "loss": 1.4403, + "theoretical_loss": 3.613573116146025, + "tokens_seen": 1208614912 + }, + { + "epoch": 0.37, + "learning_rate": 0.0005121604814443331, + "loss": 1.4442, + "theoretical_loss": 3.613555258713416, + "tokens_seen": 1208680448 + }, + { + "epoch": 0.37, + "learning_rate": 0.0005121444332998997, + "loss": 1.4811, + "theoretical_loss": 3.6135374025201235, + "tokens_seen": 1208745984 + }, + { + "epoch": 0.37, + "learning_rate": 0.0005121283851554665, + "loss": 1.3794, + "theoretical_loss": 3.6135195475659945, + "tokens_seen": 1208811520 + }, + { + "epoch": 0.37, + "learning_rate": 0.0005121123370110331, + "loss": 1.283, + "theoretical_loss": 3.6135016938508766, + "tokens_seen": 1208877056 + }, + { + "epoch": 0.37, + "learning_rate": 0.0005120962888665999, + "loss": 1.2987, + "theoretical_loss": 3.6134838413746158, + "tokens_seen": 1208942592 + }, + { + "epoch": 0.37, + "learning_rate": 0.0005120802407221665, + "loss": 1.4084, + "theoretical_loss": 3.6134659901370587, + "tokens_seen": 1209008128 + }, + { + "epoch": 0.37, + "learning_rate": 0.0005120641925777333, + "loss": 1.3811, + "theoretical_loss": 3.613448140138053, + "tokens_seen": 1209073664 + }, + { + "epoch": 0.37, + "objective/train/docs_used": 442331, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.2643508911132812, + "objective/train/theoretical_loss": 3.613430291377446, + "objective/train/tokens_used": 1229599200, + "theoretical_loss": 3.613430291377446, + "tokens_seen": 1209139200 + }, + { + "epoch": 0.37, + "learning_rate": 0.0005120481444332999, + "loss": 1.3164, + "theoretical_loss": 3.613430291377446, + "tokens_seen": 1209139200 + }, + { + "epoch": 0.37, + "learning_rate": 0.0005120320962888667, + "loss": 1.3799, + "theoretical_loss": 3.613412443855083, + "tokens_seen": 1209204736 + }, + { + "epoch": 0.37, + "learning_rate": 0.0005120160481444333, + "loss": 1.4369, + "theoretical_loss": 3.613394597570813, + "tokens_seen": 1209270272 + }, + { + "epoch": 0.37, + "learning_rate": 0.0005120000000000001, + "loss": 1.4459, + "theoretical_loss": 3.6133767525244815, + "tokens_seen": 1209335808 + }, + { + "epoch": 0.37, + "learning_rate": 0.0005119839518555667, + "loss": 1.4039, + "theoretical_loss": 3.613358908715936, + "tokens_seen": 1209401344 + }, + { + "epoch": 0.37, + "learning_rate": 0.0005119679037111335, + "loss": 1.4702, + "theoretical_loss": 3.6133410661450243, + "tokens_seen": 1209466880 + }, + { + "epoch": 0.37, + "learning_rate": 0.0005119518555667001, + "loss": 1.461, + "theoretical_loss": 3.6133232248115927, + "tokens_seen": 1209532416 + }, + { + "epoch": 0.37, + "learning_rate": 0.0005119358074222668, + "loss": 1.403, + "theoretical_loss": 3.6133053847154883, + "tokens_seen": 1209597952 + }, + { + "epoch": 0.37, + "learning_rate": 0.0005119197592778335, + "loss": 1.4223, + "theoretical_loss": 3.6132875458565588, + "tokens_seen": 1209663488 + }, + { + "epoch": 0.37, + "learning_rate": 0.0005119037111334002, + "loss": 1.3383, + "theoretical_loss": 3.6132697082346508, + "tokens_seen": 1209729024 + }, + { + "epoch": 0.37, + "learning_rate": 0.0005118876629889669, + "loss": 1.3421, + "theoretical_loss": 3.613251871849612, + "tokens_seen": 1209794560 + }, + { + "epoch": 0.37, + "learning_rate": 0.0005118716148445336, + "loss": 1.3808, + "theoretical_loss": 3.613234036701289, + "tokens_seen": 1209860096 + }, + { + "epoch": 0.37, + "learning_rate": 0.0005118555667001003, + "loss": 1.4648, + "theoretical_loss": 3.6132162027895296, + "tokens_seen": 1209925632 + }, + { + "epoch": 0.37, + "learning_rate": 0.000511839518555667, + "loss": 1.3663, + "theoretical_loss": 3.613198370114181, + "tokens_seen": 1209991168 + }, + { + "epoch": 0.37, + "learning_rate": 0.0005118234704112337, + "loss": 1.4148, + "theoretical_loss": 3.613180538675091, + "tokens_seen": 1210056704 + }, + { + "epoch": 0.37, + "learning_rate": 0.0005118074222668004, + "loss": 1.3735, + "theoretical_loss": 3.613162708472106, + "tokens_seen": 1210122240 + }, + { + "epoch": 0.37, + "learning_rate": 0.000511791374122367, + "loss": 1.4162, + "theoretical_loss": 3.6131448795050733, + "tokens_seen": 1210187776 + }, + { + "epoch": 0.37, + "learning_rate": 0.0005117753259779338, + "loss": 1.4741, + "theoretical_loss": 3.6131270517738416, + "tokens_seen": 1210253312 + }, + { + "epoch": 0.37, + "learning_rate": 0.0005117592778335004, + "loss": 1.3002, + "theoretical_loss": 3.6131092252782575, + "tokens_seen": 1210318848 + }, + { + "epoch": 0.37, + "learning_rate": 0.0005117432296890672, + "loss": 1.3749, + "theoretical_loss": 3.6130914000181686, + "tokens_seen": 1210384384 + }, + { + "epoch": 0.37, + "learning_rate": 0.0005117271815446338, + "loss": 1.3966, + "theoretical_loss": 3.613073575993422, + "tokens_seen": 1210449920 + }, + { + "epoch": 0.37, + "learning_rate": 0.0005117111334002006, + "loss": 1.3759, + "theoretical_loss": 3.6130557532038656, + "tokens_seen": 1210515456 + }, + { + "epoch": 0.37, + "learning_rate": 0.0005116950852557673, + "loss": 1.4072, + "theoretical_loss": 3.6130379316493464, + "tokens_seen": 1210580992 + }, + { + "epoch": 0.37, + "learning_rate": 0.000511679037111334, + "loss": 1.3893, + "theoretical_loss": 3.613020111329713, + "tokens_seen": 1210646528 + }, + { + "epoch": 0.37, + "learning_rate": 0.0005116629889669007, + "loss": 1.5019, + "theoretical_loss": 3.613002292244812, + "tokens_seen": 1210712064 + }, + { + "epoch": 0.37, + "learning_rate": 0.0005116469408224674, + "loss": 1.3969, + "theoretical_loss": 3.6129844743944917, + "tokens_seen": 1210777600 + }, + { + "epoch": 0.37, + "learning_rate": 0.0005116308926780341, + "loss": 1.4647, + "theoretical_loss": 3.6129666577785993, + "tokens_seen": 1210843136 + }, + { + "epoch": 0.37, + "learning_rate": 0.0005116148445336008, + "loss": 1.4473, + "theoretical_loss": 3.6129488423969827, + "tokens_seen": 1210908672 + }, + { + "epoch": 0.37, + "learning_rate": 0.0005115987963891675, + "loss": 1.3881, + "theoretical_loss": 3.6129310282494895, + "tokens_seen": 1210974208 + }, + { + "epoch": 0.37, + "learning_rate": 0.0005115827482447342, + "loss": 1.329, + "theoretical_loss": 3.6129132153359675, + "tokens_seen": 1211039744 + }, + { + "epoch": 0.37, + "learning_rate": 0.000511566700100301, + "loss": 1.4478, + "theoretical_loss": 3.6128954036562644, + "tokens_seen": 1211105280 + }, + { + "epoch": 0.37, + "learning_rate": 0.0005115506519558676, + "loss": 1.4502, + "theoretical_loss": 3.6128775932102277, + "tokens_seen": 1211170816 + }, + { + "epoch": 0.37, + "learning_rate": 0.0005115346038114343, + "loss": 1.4472, + "theoretical_loss": 3.612859783997706, + "tokens_seen": 1211236352 + }, + { + "epoch": 0.37, + "learning_rate": 0.000511518555667001, + "loss": 1.3811, + "theoretical_loss": 3.6128419760185464, + "tokens_seen": 1211301888 + }, + { + "epoch": 0.37, + "learning_rate": 0.0005115025075225678, + "loss": 1.4074, + "theoretical_loss": 3.612824169272597, + "tokens_seen": 1211367424 + }, + { + "epoch": 0.37, + "learning_rate": 0.0005114864593781344, + "loss": 1.5072, + "theoretical_loss": 3.612806363759706, + "tokens_seen": 1211432960 + }, + { + "epoch": 0.37, + "learning_rate": 0.0005114704112337012, + "loss": 1.483, + "theoretical_loss": 3.6127885594797204, + "tokens_seen": 1211498496 + }, + { + "epoch": 0.37, + "learning_rate": 0.0005114543630892678, + "loss": 1.4722, + "theoretical_loss": 3.612770756432489, + "tokens_seen": 1211564032 + }, + { + "epoch": 0.37, + "learning_rate": 0.0005114383149448346, + "loss": 1.4026, + "theoretical_loss": 3.61275295461786, + "tokens_seen": 1211629568 + }, + { + "epoch": 0.37, + "learning_rate": 0.0005114222668004012, + "loss": 1.3985, + "theoretical_loss": 3.61273515403568, + "tokens_seen": 1211695104 + }, + { + "epoch": 0.37, + "learning_rate": 0.000511406218655968, + "loss": 1.4292, + "theoretical_loss": 3.612717354685798, + "tokens_seen": 1211760640 + }, + { + "epoch": 0.37, + "learning_rate": 0.0005113901705115346, + "loss": 1.4263, + "theoretical_loss": 3.612699556568063, + "tokens_seen": 1211826176 + }, + { + "epoch": 0.37, + "learning_rate": 0.0005113741223671014, + "loss": 1.3785, + "theoretical_loss": 3.612681759682321, + "tokens_seen": 1211891712 + }, + { + "epoch": 0.37, + "learning_rate": 0.000511358074222668, + "loss": 1.4052, + "theoretical_loss": 3.612663964028422, + "tokens_seen": 1211957248 + }, + { + "epoch": 0.37, + "learning_rate": 0.0005113420260782348, + "loss": 1.3922, + "theoretical_loss": 3.6126461696062124, + "tokens_seen": 1212022784 + }, + { + "epoch": 0.37, + "learning_rate": 0.0005113259779338014, + "loss": 1.5197, + "theoretical_loss": 3.6126283764155422, + "tokens_seen": 1212088320 + }, + { + "epoch": 0.37, + "learning_rate": 0.0005113099297893682, + "loss": 1.4644, + "theoretical_loss": 3.612610584456258, + "tokens_seen": 1212153856 + }, + { + "epoch": 0.37, + "learning_rate": 0.0005112938816449348, + "loss": 1.3195, + "theoretical_loss": 3.6125927937282087, + "tokens_seen": 1212219392 + }, + { + "epoch": 0.37, + "learning_rate": 0.0005112778335005016, + "loss": 1.5485, + "theoretical_loss": 3.612575004231243, + "tokens_seen": 1212284928 + }, + { + "epoch": 0.37, + "learning_rate": 0.0005112617853560682, + "loss": 1.4191, + "theoretical_loss": 3.612557215965208, + "tokens_seen": 1212350464 + }, + { + "epoch": 0.37, + "objective/train/docs_used": 443456, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.439915657043457, + "objective/train/theoretical_loss": 3.6125394289299533, + "objective/train/tokens_used": 1232876000, + "theoretical_loss": 3.6125394289299533, + "tokens_seen": 1212416000 + }, + { + "epoch": 0.37, + "learning_rate": 0.000511245737211635, + "loss": 1.4615, + "theoretical_loss": 3.6125394289299533, + "tokens_seen": 1212416000 + }, + { + "epoch": 0.37, + "learning_rate": 0.0005112296890672016, + "loss": 1.383, + "theoretical_loss": 3.6125216431253264, + "tokens_seen": 1212481536 + }, + { + "epoch": 0.37, + "learning_rate": 0.0005112136409227684, + "loss": 1.2799, + "theoretical_loss": 3.6125038585511753, + "tokens_seen": 1212547072 + }, + { + "epoch": 0.37, + "learning_rate": 0.000511197592778335, + "loss": 1.3684, + "theoretical_loss": 3.6124860752073493, + "tokens_seen": 1212612608 + }, + { + "epoch": 0.37, + "learning_rate": 0.0005111815446339018, + "loss": 1.2895, + "theoretical_loss": 3.6124682930936967, + "tokens_seen": 1212678144 + }, + { + "epoch": 0.37, + "learning_rate": 0.0005111654964894684, + "loss": 1.359, + "theoretical_loss": 3.6124505122100654, + "tokens_seen": 1212743680 + }, + { + "epoch": 0.37, + "learning_rate": 0.0005111494483450352, + "loss": 1.4663, + "theoretical_loss": 3.6124327325563037, + "tokens_seen": 1212809216 + }, + { + "epoch": 0.37, + "learning_rate": 0.0005111334002006018, + "loss": 1.4284, + "theoretical_loss": 3.6124149541322605, + "tokens_seen": 1212874752 + }, + { + "epoch": 0.37, + "learning_rate": 0.0005111173520561686, + "loss": 1.5028, + "theoretical_loss": 3.612397176937785, + "tokens_seen": 1212940288 + }, + { + "epoch": 0.37, + "learning_rate": 0.0005111013039117352, + "loss": 1.4114, + "theoretical_loss": 3.612379400972724, + "tokens_seen": 1213005824 + }, + { + "epoch": 0.37, + "learning_rate": 0.000511085255767302, + "loss": 1.5051, + "theoretical_loss": 3.612361626236928, + "tokens_seen": 1213071360 + }, + { + "epoch": 0.37, + "learning_rate": 0.0005110692076228686, + "loss": 1.3824, + "theoretical_loss": 3.612343852730244, + "tokens_seen": 1213136896 + }, + { + "epoch": 0.37, + "learning_rate": 0.0005110531594784354, + "loss": 1.4838, + "theoretical_loss": 3.6123260804525215, + "tokens_seen": 1213202432 + }, + { + "epoch": 0.37, + "learning_rate": 0.000511037111334002, + "loss": 1.3232, + "theoretical_loss": 3.612308309403609, + "tokens_seen": 1213267968 + }, + { + "epoch": 0.37, + "learning_rate": 0.0005110210631895688, + "loss": 1.4148, + "theoretical_loss": 3.6122905395833547, + "tokens_seen": 1213333504 + }, + { + "epoch": 0.37, + "learning_rate": 0.0005110050150451354, + "loss": 1.4743, + "theoretical_loss": 3.6122727709916083, + "tokens_seen": 1213399040 + }, + { + "epoch": 0.37, + "learning_rate": 0.0005109889669007022, + "loss": 1.3812, + "theoretical_loss": 3.612255003628217, + "tokens_seen": 1213464576 + }, + { + "epoch": 0.37, + "learning_rate": 0.0005109729187562688, + "loss": 1.4318, + "theoretical_loss": 3.6122372374930314, + "tokens_seen": 1213530112 + }, + { + "epoch": 0.37, + "learning_rate": 0.0005109568706118356, + "loss": 1.3889, + "theoretical_loss": 3.612219472585899, + "tokens_seen": 1213595648 + }, + { + "epoch": 0.37, + "learning_rate": 0.0005109408224674022, + "loss": 1.3422, + "theoretical_loss": 3.6122017089066687, + "tokens_seen": 1213661184 + }, + { + "epoch": 0.37, + "learning_rate": 0.000510924774322969, + "loss": 1.415, + "theoretical_loss": 3.6121839464551897, + "tokens_seen": 1213726720 + }, + { + "epoch": 0.37, + "learning_rate": 0.0005109087261785356, + "loss": 1.4464, + "theoretical_loss": 3.6121661852313105, + "tokens_seen": 1213792256 + }, + { + "epoch": 0.37, + "learning_rate": 0.0005108926780341024, + "loss": 1.414, + "theoretical_loss": 3.6121484252348806, + "tokens_seen": 1213857792 + }, + { + "epoch": 0.37, + "learning_rate": 0.000510876629889669, + "loss": 1.4694, + "theoretical_loss": 3.6121306664657484, + "tokens_seen": 1213923328 + }, + { + "epoch": 0.37, + "learning_rate": 0.0005108605817452358, + "loss": 1.4669, + "theoretical_loss": 3.6121129089237627, + "tokens_seen": 1213988864 + }, + { + "epoch": 0.37, + "learning_rate": 0.0005108445336008024, + "loss": 1.4794, + "theoretical_loss": 3.612095152608773, + "tokens_seen": 1214054400 + }, + { + "epoch": 0.37, + "learning_rate": 0.0005108284854563692, + "loss": 1.5307, + "theoretical_loss": 3.612077397520628, + "tokens_seen": 1214119936 + }, + { + "epoch": 0.37, + "learning_rate": 0.0005108124373119358, + "loss": 1.3985, + "theoretical_loss": 3.6120596436591765, + "tokens_seen": 1214185472 + }, + { + "epoch": 0.37, + "learning_rate": 0.0005107963891675026, + "loss": 1.4279, + "theoretical_loss": 3.6120418910242678, + "tokens_seen": 1214251008 + }, + { + "epoch": 0.37, + "learning_rate": 0.0005107803410230692, + "loss": 1.2987, + "theoretical_loss": 3.612024139615751, + "tokens_seen": 1214316544 + }, + { + "epoch": 0.37, + "learning_rate": 0.000510764292878636, + "loss": 1.4456, + "theoretical_loss": 3.612006389433475, + "tokens_seen": 1214382080 + }, + { + "epoch": 0.37, + "learning_rate": 0.0005107482447342026, + "loss": 1.4442, + "theoretical_loss": 3.611988640477289, + "tokens_seen": 1214447616 + }, + { + "epoch": 0.37, + "learning_rate": 0.0005107321965897694, + "loss": 1.4174, + "theoretical_loss": 3.6119708927470424, + "tokens_seen": 1214513152 + }, + { + "epoch": 0.37, + "learning_rate": 0.000510716148445336, + "loss": 1.4497, + "theoretical_loss": 3.611953146242584, + "tokens_seen": 1214578688 + }, + { + "epoch": 0.37, + "learning_rate": 0.0005107001003009028, + "loss": 1.4699, + "theoretical_loss": 3.6119354009637634, + "tokens_seen": 1214644224 + }, + { + "epoch": 0.37, + "learning_rate": 0.0005106840521564694, + "loss": 1.5013, + "theoretical_loss": 3.6119176569104297, + "tokens_seen": 1214709760 + }, + { + "epoch": 0.37, + "learning_rate": 0.0005106680040120362, + "loss": 1.3406, + "theoretical_loss": 3.611899914082432, + "tokens_seen": 1214775296 + }, + { + "epoch": 0.37, + "learning_rate": 0.0005106519558676028, + "loss": 1.4912, + "theoretical_loss": 3.6118821724796195, + "tokens_seen": 1214840832 + }, + { + "epoch": 0.37, + "learning_rate": 0.0005106359077231696, + "loss": 1.4483, + "theoretical_loss": 3.6118644321018416, + "tokens_seen": 1214906368 + }, + { + "epoch": 0.37, + "learning_rate": 0.0005106198595787362, + "loss": 1.463, + "theoretical_loss": 3.611846692948948, + "tokens_seen": 1214971904 + }, + { + "epoch": 0.37, + "learning_rate": 0.000510603811434303, + "loss": 1.335, + "theoretical_loss": 3.6118289550207874, + "tokens_seen": 1215037440 + }, + { + "epoch": 0.37, + "learning_rate": 0.0005105877632898696, + "loss": 1.2916, + "theoretical_loss": 3.61181121831721, + "tokens_seen": 1215102976 + }, + { + "epoch": 0.37, + "learning_rate": 0.0005105717151454364, + "loss": 1.3108, + "theoretical_loss": 3.6117934828380642, + "tokens_seen": 1215168512 + }, + { + "epoch": 0.37, + "learning_rate": 0.000510555667001003, + "loss": 1.591, + "theoretical_loss": 3.6117757485832005, + "tokens_seen": 1215234048 + }, + { + "epoch": 0.37, + "learning_rate": 0.0005105396188565698, + "loss": 1.4227, + "theoretical_loss": 3.611758015552468, + "tokens_seen": 1215299584 + }, + { + "epoch": 0.37, + "learning_rate": 0.0005105235707121364, + "loss": 1.4812, + "theoretical_loss": 3.6117402837457155, + "tokens_seen": 1215365120 + }, + { + "epoch": 0.37, + "learning_rate": 0.0005105075225677031, + "loss": 1.3515, + "theoretical_loss": 3.6117225531627932, + "tokens_seen": 1215430656 + }, + { + "epoch": 0.37, + "learning_rate": 0.0005104914744232698, + "loss": 1.3961, + "theoretical_loss": 3.6117048238035507, + "tokens_seen": 1215496192 + }, + { + "epoch": 0.37, + "learning_rate": 0.0005104754262788365, + "loss": 1.4444, + "theoretical_loss": 3.611687095667838, + "tokens_seen": 1215561728 + }, + { + "epoch": 0.37, + "learning_rate": 0.0005104593781344032, + "loss": 1.4643, + "theoretical_loss": 3.6116693687555035, + "tokens_seen": 1215627264 + }, + { + "epoch": 0.37, + "objective/train/docs_used": 444688, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.3394641876220703, + "objective/train/theoretical_loss": 3.6116516430663976, + "objective/train/tokens_used": 1236152800, + "theoretical_loss": 3.6116516430663976, + "tokens_seen": 1215692800 + }, + { + "epoch": 0.37, + "learning_rate": 0.0005104433299899699, + "loss": 1.3846, + "theoretical_loss": 3.6116516430663976, + "tokens_seen": 1215692800 + }, + { + "epoch": 0.37, + "learning_rate": 0.0005104272818455366, + "loss": 1.348, + "theoretical_loss": 3.6116339186003694, + "tokens_seen": 1215758336 + }, + { + "epoch": 0.37, + "learning_rate": 0.0005104112337011033, + "loss": 1.4081, + "theoretical_loss": 3.6116161953572696, + "tokens_seen": 1215823872 + }, + { + "epoch": 0.37, + "learning_rate": 0.00051039518555667, + "loss": 1.5108, + "theoretical_loss": 3.6115984733369473, + "tokens_seen": 1215889408 + }, + { + "epoch": 0.37, + "learning_rate": 0.0005103791374122367, + "loss": 1.409, + "theoretical_loss": 3.6115807525392523, + "tokens_seen": 1215954944 + }, + { + "epoch": 0.37, + "learning_rate": 0.0005103630892678034, + "loss": 1.4165, + "theoretical_loss": 3.611563032964034, + "tokens_seen": 1216020480 + }, + { + "epoch": 0.37, + "learning_rate": 0.0005103470411233701, + "loss": 1.3453, + "theoretical_loss": 3.611545314611143, + "tokens_seen": 1216086016 + }, + { + "epoch": 0.37, + "learning_rate": 0.0005103309929789368, + "loss": 1.3986, + "theoretical_loss": 3.6115275974804284, + "tokens_seen": 1216151552 + }, + { + "epoch": 0.37, + "learning_rate": 0.0005103149448345035, + "loss": 1.3784, + "theoretical_loss": 3.61150988157174, + "tokens_seen": 1216217088 + }, + { + "epoch": 0.37, + "learning_rate": 0.0005102988966900702, + "loss": 1.56, + "theoretical_loss": 3.6114921668849282, + "tokens_seen": 1216282624 + }, + { + "epoch": 0.37, + "learning_rate": 0.0005102828485456369, + "loss": 1.3356, + "theoretical_loss": 3.6114744534198424, + "tokens_seen": 1216348160 + }, + { + "epoch": 0.37, + "learning_rate": 0.0005102668004012036, + "loss": 1.3614, + "theoretical_loss": 3.6114567411763328, + "tokens_seen": 1216413696 + }, + { + "epoch": 0.37, + "learning_rate": 0.0005102507522567703, + "loss": 1.3711, + "theoretical_loss": 3.6114390301542496, + "tokens_seen": 1216479232 + }, + { + "epoch": 0.37, + "learning_rate": 0.000510234704112337, + "loss": 1.5108, + "theoretical_loss": 3.6114213203534424, + "tokens_seen": 1216544768 + }, + { + "epoch": 0.37, + "learning_rate": 0.0005102186559679037, + "loss": 1.3963, + "theoretical_loss": 3.611403611773761, + "tokens_seen": 1216610304 + }, + { + "epoch": 0.37, + "learning_rate": 0.0005102026078234704, + "loss": 1.4197, + "theoretical_loss": 3.611385904415056, + "tokens_seen": 1216675840 + }, + { + "epoch": 0.37, + "learning_rate": 0.0005101865596790371, + "loss": 1.4819, + "theoretical_loss": 3.611368198277177, + "tokens_seen": 1216741376 + }, + { + "epoch": 0.37, + "learning_rate": 0.0005101705115346038, + "loss": 1.36, + "theoretical_loss": 3.611350493359974, + "tokens_seen": 1216806912 + }, + { + "epoch": 0.37, + "learning_rate": 0.0005101544633901705, + "loss": 1.3799, + "theoretical_loss": 3.6113327896632983, + "tokens_seen": 1216872448 + }, + { + "epoch": 0.37, + "learning_rate": 0.0005101384152457373, + "loss": 1.337, + "theoretical_loss": 3.6113150871869983, + "tokens_seen": 1216937984 + }, + { + "epoch": 0.37, + "learning_rate": 0.0005101223671013039, + "loss": 1.3791, + "theoretical_loss": 3.611297385930925, + "tokens_seen": 1217003520 + }, + { + "epoch": 0.37, + "learning_rate": 0.0005101063189568707, + "loss": 1.2846, + "theoretical_loss": 3.611279685894929, + "tokens_seen": 1217069056 + }, + { + "epoch": 0.37, + "learning_rate": 0.0005100902708124373, + "loss": 1.3709, + "theoretical_loss": 3.6112619870788594, + "tokens_seen": 1217134592 + }, + { + "epoch": 0.37, + "learning_rate": 0.000510074222668004, + "loss": 1.4016, + "theoretical_loss": 3.611244289482568, + "tokens_seen": 1217200128 + }, + { + "epoch": 0.37, + "learning_rate": 0.0005100581745235707, + "loss": 1.4017, + "theoretical_loss": 3.6112265931059033, + "tokens_seen": 1217265664 + }, + { + "epoch": 0.37, + "learning_rate": 0.0005100421263791375, + "loss": 1.4469, + "theoretical_loss": 3.6112088979487167, + "tokens_seen": 1217331200 + }, + { + "epoch": 0.37, + "learning_rate": 0.0005100260782347041, + "loss": 1.337, + "theoretical_loss": 3.611191204010858, + "tokens_seen": 1217396736 + }, + { + "epoch": 0.37, + "learning_rate": 0.0005100100300902709, + "loss": 1.2708, + "theoretical_loss": 3.6111735112921783, + "tokens_seen": 1217462272 + }, + { + "epoch": 0.37, + "learning_rate": 0.0005099939819458375, + "loss": 1.4219, + "theoretical_loss": 3.6111558197925273, + "tokens_seen": 1217527808 + }, + { + "epoch": 0.37, + "learning_rate": 0.0005099779338014043, + "loss": 1.4312, + "theoretical_loss": 3.6111381295117555, + "tokens_seen": 1217593344 + }, + { + "epoch": 0.37, + "learning_rate": 0.0005099618856569709, + "loss": 1.5222, + "theoretical_loss": 3.6111204404497137, + "tokens_seen": 1217658880 + }, + { + "epoch": 0.37, + "learning_rate": 0.0005099458375125377, + "loss": 1.3974, + "theoretical_loss": 3.611102752606252, + "tokens_seen": 1217724416 + }, + { + "epoch": 0.37, + "learning_rate": 0.0005099297893681043, + "loss": 1.3817, + "theoretical_loss": 3.6110850659812206, + "tokens_seen": 1217789952 + }, + { + "epoch": 0.37, + "learning_rate": 0.0005099137412236711, + "loss": 1.4623, + "theoretical_loss": 3.611067380574471, + "tokens_seen": 1217855488 + }, + { + "epoch": 0.37, + "learning_rate": 0.0005098976930792377, + "loss": 1.5202, + "theoretical_loss": 3.6110496963858525, + "tokens_seen": 1217921024 + }, + { + "epoch": 0.37, + "learning_rate": 0.0005098816449348045, + "loss": 1.3694, + "theoretical_loss": 3.6110320134152163, + "tokens_seen": 1217986560 + }, + { + "epoch": 0.37, + "learning_rate": 0.0005098655967903711, + "loss": 1.3997, + "theoretical_loss": 3.611014331662413, + "tokens_seen": 1218052096 + }, + { + "epoch": 0.37, + "learning_rate": 0.0005098495486459379, + "loss": 1.549, + "theoretical_loss": 3.610996651127293, + "tokens_seen": 1218117632 + }, + { + "epoch": 0.37, + "learning_rate": 0.0005098335005015045, + "loss": 1.3615, + "theoretical_loss": 3.6109789718097076, + "tokens_seen": 1218183168 + }, + { + "epoch": 0.37, + "learning_rate": 0.0005098174523570713, + "loss": 1.4326, + "theoretical_loss": 3.610961293709506, + "tokens_seen": 1218248704 + }, + { + "epoch": 0.37, + "learning_rate": 0.0005098014042126379, + "loss": 1.3729, + "theoretical_loss": 3.6109436168265407, + "tokens_seen": 1218314240 + }, + { + "epoch": 0.37, + "learning_rate": 0.0005097853560682047, + "loss": 1.4642, + "theoretical_loss": 3.6109259411606613, + "tokens_seen": 1218379776 + }, + { + "epoch": 0.37, + "learning_rate": 0.0005097693079237713, + "loss": 1.4436, + "theoretical_loss": 3.6109082667117187, + "tokens_seen": 1218445312 + }, + { + "epoch": 0.37, + "learning_rate": 0.0005097532597793381, + "loss": 1.406, + "theoretical_loss": 3.6108905934795636, + "tokens_seen": 1218510848 + }, + { + "epoch": 0.37, + "learning_rate": 0.0005097372116349047, + "loss": 1.3818, + "theoretical_loss": 3.610872921464047, + "tokens_seen": 1218576384 + }, + { + "epoch": 0.37, + "learning_rate": 0.0005097211634904715, + "loss": 1.3802, + "theoretical_loss": 3.61085525066502, + "tokens_seen": 1218641920 + }, + { + "epoch": 0.37, + "learning_rate": 0.0005097051153460381, + "loss": 1.3973, + "theoretical_loss": 3.610837581082333, + "tokens_seen": 1218707456 + }, + { + "epoch": 0.37, + "learning_rate": 0.0005096890672016049, + "loss": 1.4967, + "theoretical_loss": 3.610819912715837, + "tokens_seen": 1218772992 + }, + { + "epoch": 0.37, + "learning_rate": 0.0005096730190571715, + "loss": 1.4212, + "theoretical_loss": 3.6108022455653828, + "tokens_seen": 1218838528 + }, + { + "epoch": 0.37, + "learning_rate": 0.0005096569709127383, + "loss": 1.3019, + "theoretical_loss": 3.6107845796308213, + "tokens_seen": 1218904064 + }, + { + "epoch": 0.37, + "objective/train/docs_used": 445814, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.5334891080856323, + "objective/train/theoretical_loss": 3.6107669149120034, + "objective/train/tokens_used": 1239429600, + "theoretical_loss": 3.6107669149120034, + "tokens_seen": 1218969600 + }, + { + "epoch": 0.37, + "learning_rate": 0.0005096409227683049, + "loss": 1.4657, + "theoretical_loss": 3.6107669149120034, + "tokens_seen": 1218969600 + }, + { + "epoch": 0.37, + "learning_rate": 0.0005096248746238717, + "loss": 1.3574, + "theoretical_loss": 3.6107492514087802, + "tokens_seen": 1219035136 + }, + { + "epoch": 0.37, + "learning_rate": 0.0005096088264794383, + "loss": 1.3293, + "theoretical_loss": 3.610731589121003, + "tokens_seen": 1219100672 + }, + { + "epoch": 0.37, + "learning_rate": 0.0005095927783350051, + "loss": 1.3404, + "theoretical_loss": 3.610713928048523, + "tokens_seen": 1219166208 + }, + { + "epoch": 0.37, + "learning_rate": 0.0005095767301905717, + "loss": 1.509, + "theoretical_loss": 3.61069626819119, + "tokens_seen": 1219231744 + }, + { + "epoch": 0.37, + "learning_rate": 0.0005095606820461385, + "loss": 1.3291, + "theoretical_loss": 3.6106786095488563, + "tokens_seen": 1219297280 + }, + { + "epoch": 0.37, + "learning_rate": 0.0005095446339017051, + "loss": 1.3555, + "theoretical_loss": 3.610660952121373, + "tokens_seen": 1219362816 + }, + { + "epoch": 0.37, + "learning_rate": 0.0005095285857572719, + "loss": 1.4658, + "theoretical_loss": 3.61064329590859, + "tokens_seen": 1219428352 + }, + { + "epoch": 0.37, + "learning_rate": 0.0005095125376128385, + "loss": 1.4203, + "theoretical_loss": 3.61062564091036, + "tokens_seen": 1219493888 + }, + { + "epoch": 0.37, + "learning_rate": 0.0005094964894684053, + "loss": 1.4416, + "theoretical_loss": 3.6106079871265333, + "tokens_seen": 1219559424 + }, + { + "epoch": 0.37, + "learning_rate": 0.0005094804413239719, + "loss": 1.4124, + "theoretical_loss": 3.6105903345569614, + "tokens_seen": 1219624960 + }, + { + "epoch": 0.37, + "learning_rate": 0.0005094643931795387, + "loss": 1.3796, + "theoretical_loss": 3.6105726832014953, + "tokens_seen": 1219690496 + }, + { + "epoch": 0.37, + "learning_rate": 0.0005094483450351053, + "loss": 1.3437, + "theoretical_loss": 3.6105550330599865, + "tokens_seen": 1219756032 + }, + { + "epoch": 0.37, + "learning_rate": 0.0005094322968906721, + "loss": 1.4187, + "theoretical_loss": 3.6105373841322863, + "tokens_seen": 1219821568 + }, + { + "epoch": 0.37, + "learning_rate": 0.0005094162487462387, + "loss": 1.3904, + "theoretical_loss": 3.610519736418246, + "tokens_seen": 1219887104 + }, + { + "epoch": 0.37, + "learning_rate": 0.0005094002006018055, + "loss": 1.4623, + "theoretical_loss": 3.6105020899177167, + "tokens_seen": 1219952640 + }, + { + "epoch": 0.37, + "learning_rate": 0.0005093841524573721, + "loss": 1.3097, + "theoretical_loss": 3.6104844446305497, + "tokens_seen": 1220018176 + }, + { + "epoch": 0.37, + "learning_rate": 0.0005093681043129389, + "loss": 1.4242, + "theoretical_loss": 3.6104668005565976, + "tokens_seen": 1220083712 + }, + { + "epoch": 0.37, + "learning_rate": 0.0005093520561685055, + "loss": 1.3861, + "theoretical_loss": 3.61044915769571, + "tokens_seen": 1220149248 + }, + { + "epoch": 0.37, + "learning_rate": 0.0005093360080240723, + "loss": 1.4162, + "theoretical_loss": 3.61043151604774, + "tokens_seen": 1220214784 + }, + { + "epoch": 0.37, + "learning_rate": 0.0005093199598796389, + "loss": 1.3433, + "theoretical_loss": 3.6104138756125375, + "tokens_seen": 1220280320 + }, + { + "epoch": 0.37, + "learning_rate": 0.0005093039117352057, + "loss": 1.3912, + "theoretical_loss": 3.6103962363899553, + "tokens_seen": 1220345856 + }, + { + "epoch": 0.37, + "learning_rate": 0.0005092878635907724, + "loss": 1.4765, + "theoretical_loss": 3.610378598379844, + "tokens_seen": 1220411392 + }, + { + "epoch": 0.37, + "learning_rate": 0.0005092718154463391, + "loss": 1.3339, + "theoretical_loss": 3.610360961582056, + "tokens_seen": 1220476928 + }, + { + "epoch": 0.37, + "learning_rate": 0.0005092557673019057, + "loss": 1.4493, + "theoretical_loss": 3.6103433259964426, + "tokens_seen": 1220542464 + }, + { + "epoch": 0.37, + "learning_rate": 0.0005092397191574725, + "loss": 1.3371, + "theoretical_loss": 3.610325691622855, + "tokens_seen": 1220608000 + }, + { + "epoch": 0.37, + "learning_rate": 0.0005092236710130391, + "loss": 1.4072, + "theoretical_loss": 3.610308058461145, + "tokens_seen": 1220673536 + }, + { + "epoch": 0.37, + "learning_rate": 0.0005092076228686059, + "loss": 1.5518, + "theoretical_loss": 3.610290426511165, + "tokens_seen": 1220739072 + }, + { + "epoch": 0.37, + "learning_rate": 0.0005091915747241725, + "loss": 1.4698, + "theoretical_loss": 3.6102727957727656, + "tokens_seen": 1220804608 + }, + { + "epoch": 0.37, + "learning_rate": 0.0005091755265797393, + "loss": 1.3567, + "theoretical_loss": 3.610255166245799, + "tokens_seen": 1220870144 + }, + { + "epoch": 0.37, + "learning_rate": 0.0005091594784353059, + "loss": 1.41, + "theoretical_loss": 3.610237537930117, + "tokens_seen": 1220935680 + }, + { + "epoch": 0.37, + "learning_rate": 0.0005091434302908727, + "loss": 1.3831, + "theoretical_loss": 3.6102199108255713, + "tokens_seen": 1221001216 + }, + { + "epoch": 0.37, + "learning_rate": 0.0005091273821464393, + "loss": 1.3443, + "theoretical_loss": 3.6102022849320132, + "tokens_seen": 1221066752 + }, + { + "epoch": 0.37, + "learning_rate": 0.000509111334002006, + "loss": 1.2844, + "theoretical_loss": 3.6101846602492955, + "tokens_seen": 1221132288 + }, + { + "epoch": 0.37, + "learning_rate": 0.0005090952858575727, + "loss": 1.4265, + "theoretical_loss": 3.61016703677727, + "tokens_seen": 1221197824 + }, + { + "epoch": 0.37, + "learning_rate": 0.0005090792377131394, + "loss": 1.4324, + "theoretical_loss": 3.6101494145157873, + "tokens_seen": 1221263360 + }, + { + "epoch": 0.37, + "learning_rate": 0.0005090631895687061, + "loss": 1.4764, + "theoretical_loss": 3.6101317934647, + "tokens_seen": 1221328896 + }, + { + "epoch": 0.37, + "learning_rate": 0.0005090471414242728, + "loss": 1.5017, + "theoretical_loss": 3.610114173623861, + "tokens_seen": 1221394432 + }, + { + "epoch": 0.37, + "learning_rate": 0.0005090310932798395, + "loss": 1.5552, + "theoretical_loss": 3.6100965549931203, + "tokens_seen": 1221459968 + }, + { + "epoch": 0.37, + "learning_rate": 0.0005090150451354062, + "loss": 1.3411, + "theoretical_loss": 3.610078937572332, + "tokens_seen": 1221525504 + }, + { + "epoch": 0.37, + "learning_rate": 0.0005089989969909729, + "loss": 1.4643, + "theoretical_loss": 3.6100613213613464, + "tokens_seen": 1221591040 + }, + { + "epoch": 0.37, + "learning_rate": 0.0005089829488465396, + "loss": 1.4712, + "theoretical_loss": 3.610043706360016, + "tokens_seen": 1221656576 + }, + { + "epoch": 0.37, + "learning_rate": 0.0005089669007021063, + "loss": 1.3456, + "theoretical_loss": 3.6100260925681935, + "tokens_seen": 1221722112 + }, + { + "epoch": 0.37, + "learning_rate": 0.000508950852557673, + "loss": 1.3862, + "theoretical_loss": 3.61000847998573, + "tokens_seen": 1221787648 + }, + { + "epoch": 0.37, + "learning_rate": 0.0005089348044132397, + "loss": 1.4763, + "theoretical_loss": 3.6099908686124786, + "tokens_seen": 1221853184 + }, + { + "epoch": 0.37, + "learning_rate": 0.0005089187562688064, + "loss": 1.4383, + "theoretical_loss": 3.6099732584482904, + "tokens_seen": 1221918720 + }, + { + "epoch": 0.37, + "learning_rate": 0.0005089027081243731, + "loss": 1.407, + "theoretical_loss": 3.6099556494930187, + "tokens_seen": 1221984256 + }, + { + "epoch": 0.37, + "learning_rate": 0.0005088866599799398, + "loss": 1.3838, + "theoretical_loss": 3.6099380417465143, + "tokens_seen": 1222049792 + }, + { + "epoch": 0.37, + "learning_rate": 0.0005088706118355065, + "loss": 1.5978, + "theoretical_loss": 3.6099204352086307, + "tokens_seen": 1222115328 + }, + { + "epoch": 0.37, + "learning_rate": 0.0005088545636910732, + "loss": 1.4177, + "theoretical_loss": 3.6099028298792195, + "tokens_seen": 1222180864 + }, + { + "epoch": 0.37, + "objective/train/docs_used": 446976, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.3944453001022339, + "objective/train/theoretical_loss": 3.6098852257581333, + "objective/train/tokens_used": 1242706400, + "theoretical_loss": 3.6098852257581333, + "tokens_seen": 1222246400 + }, + { + "epoch": 0.37, + "learning_rate": 0.00050883851554664, + "loss": 1.3802, + "theoretical_loss": 3.6098852257581333, + "tokens_seen": 1222246400 + }, + { + "epoch": 0.37, + "learning_rate": 0.0005088224674022066, + "loss": 1.4168, + "theoretical_loss": 3.609867622845224, + "tokens_seen": 1222311936 + }, + { + "epoch": 0.37, + "learning_rate": 0.0005088064192577733, + "loss": 1.3127, + "theoretical_loss": 3.6098500211403435, + "tokens_seen": 1222377472 + }, + { + "epoch": 0.37, + "learning_rate": 0.00050879037111334, + "loss": 1.362, + "theoretical_loss": 3.609832420643345, + "tokens_seen": 1222443008 + }, + { + "epoch": 0.37, + "learning_rate": 0.0005087743229689068, + "loss": 1.4092, + "theoretical_loss": 3.609814821354081, + "tokens_seen": 1222508544 + }, + { + "epoch": 0.37, + "learning_rate": 0.0005087582748244734, + "loss": 1.4811, + "theoretical_loss": 3.6097972232724027, + "tokens_seen": 1222574080 + }, + { + "epoch": 0.37, + "learning_rate": 0.0005087422266800402, + "loss": 1.4071, + "theoretical_loss": 3.6097796263981636, + "tokens_seen": 1222639616 + }, + { + "epoch": 0.37, + "learning_rate": 0.0005087261785356068, + "loss": 1.394, + "theoretical_loss": 3.6097620307312157, + "tokens_seen": 1222705152 + }, + { + "epoch": 0.37, + "learning_rate": 0.0005087101303911736, + "loss": 1.3069, + "theoretical_loss": 3.609744436271412, + "tokens_seen": 1222770688 + }, + { + "epoch": 0.37, + "learning_rate": 0.0005086940822467402, + "loss": 1.4772, + "theoretical_loss": 3.609726843018604, + "tokens_seen": 1222836224 + }, + { + "epoch": 0.37, + "learning_rate": 0.000508678034102307, + "loss": 1.4062, + "theoretical_loss": 3.609709250972645, + "tokens_seen": 1222901760 + }, + { + "epoch": 0.37, + "learning_rate": 0.0005086619859578736, + "loss": 1.3396, + "theoretical_loss": 3.6096916601333873, + "tokens_seen": 1222967296 + }, + { + "epoch": 0.37, + "learning_rate": 0.0005086459378134404, + "loss": 1.3798, + "theoretical_loss": 3.6096740705006836, + "tokens_seen": 1223032832 + }, + { + "epoch": 0.37, + "learning_rate": 0.000508629889669007, + "loss": 1.3755, + "theoretical_loss": 3.609656482074386, + "tokens_seen": 1223098368 + }, + { + "epoch": 0.37, + "learning_rate": 0.0005086138415245738, + "loss": 1.4573, + "theoretical_loss": 3.609638894854348, + "tokens_seen": 1223163904 + }, + { + "epoch": 0.37, + "learning_rate": 0.0005085977933801404, + "loss": 1.4292, + "theoretical_loss": 3.6096213088404214, + "tokens_seen": 1223229440 + }, + { + "epoch": 0.37, + "learning_rate": 0.0005085817452357072, + "loss": 1.3561, + "theoretical_loss": 3.6096037240324597, + "tokens_seen": 1223294976 + }, + { + "epoch": 0.37, + "learning_rate": 0.0005085656970912738, + "loss": 1.2786, + "theoretical_loss": 3.6095861404303147, + "tokens_seen": 1223360512 + }, + { + "epoch": 0.37, + "learning_rate": 0.0005085496489468406, + "loss": 1.3485, + "theoretical_loss": 3.609568558033839, + "tokens_seen": 1223426048 + }, + { + "epoch": 0.37, + "learning_rate": 0.0005085336008024072, + "loss": 1.4744, + "theoretical_loss": 3.609550976842887, + "tokens_seen": 1223491584 + }, + { + "epoch": 0.37, + "learning_rate": 0.000508517552657974, + "loss": 1.4027, + "theoretical_loss": 3.6095333968573096, + "tokens_seen": 1223557120 + }, + { + "epoch": 0.37, + "learning_rate": 0.0005085015045135406, + "loss": 1.2776, + "theoretical_loss": 3.609515818076961, + "tokens_seen": 1223622656 + }, + { + "epoch": 0.37, + "learning_rate": 0.0005084854563691074, + "loss": 1.3578, + "theoretical_loss": 3.609498240501693, + "tokens_seen": 1223688192 + }, + { + "epoch": 0.37, + "learning_rate": 0.000508469408224674, + "loss": 1.5003, + "theoretical_loss": 3.6094806641313593, + "tokens_seen": 1223753728 + }, + { + "epoch": 0.37, + "learning_rate": 0.0005084533600802408, + "loss": 1.4349, + "theoretical_loss": 3.609463088965812, + "tokens_seen": 1223819264 + }, + { + "epoch": 0.37, + "learning_rate": 0.0005084373119358074, + "loss": 1.4003, + "theoretical_loss": 3.609445515004904, + "tokens_seen": 1223884800 + }, + { + "epoch": 0.37, + "learning_rate": 0.0005084212637913742, + "loss": 1.3658, + "theoretical_loss": 3.609427942248489, + "tokens_seen": 1223950336 + }, + { + "epoch": 0.37, + "learning_rate": 0.0005084052156469408, + "loss": 1.4825, + "theoretical_loss": 3.60941037069642, + "tokens_seen": 1224015872 + }, + { + "epoch": 0.37, + "learning_rate": 0.0005083891675025076, + "loss": 1.4068, + "theoretical_loss": 3.609392800348549, + "tokens_seen": 1224081408 + }, + { + "epoch": 0.37, + "learning_rate": 0.0005083731193580742, + "loss": 1.4085, + "theoretical_loss": 3.609375231204729, + "tokens_seen": 1224146944 + }, + { + "epoch": 0.37, + "learning_rate": 0.000508357071213641, + "loss": 1.4535, + "theoretical_loss": 3.6093576632648148, + "tokens_seen": 1224212480 + }, + { + "epoch": 0.37, + "learning_rate": 0.0005083410230692076, + "loss": 1.424, + "theoretical_loss": 3.6093400965286575, + "tokens_seen": 1224278016 + }, + { + "epoch": 0.37, + "learning_rate": 0.0005083249749247744, + "loss": 1.3402, + "theoretical_loss": 3.609322530996111, + "tokens_seen": 1224343552 + }, + { + "epoch": 0.37, + "learning_rate": 0.000508308926780341, + "loss": 1.3884, + "theoretical_loss": 3.609304966667028, + "tokens_seen": 1224409088 + }, + { + "epoch": 0.37, + "learning_rate": 0.0005082928786359078, + "loss": 1.438, + "theoretical_loss": 3.6092874035412628, + "tokens_seen": 1224474624 + }, + { + "epoch": 0.37, + "learning_rate": 0.0005082768304914744, + "loss": 1.3802, + "theoretical_loss": 3.6092698416186675, + "tokens_seen": 1224540160 + }, + { + "epoch": 0.37, + "learning_rate": 0.0005082607823470412, + "loss": 1.4442, + "theoretical_loss": 3.609252280899095, + "tokens_seen": 1224605696 + }, + { + "epoch": 0.37, + "learning_rate": 0.0005082447342026078, + "loss": 1.4946, + "theoretical_loss": 3.6092347213823994, + "tokens_seen": 1224671232 + }, + { + "epoch": 0.37, + "learning_rate": 0.0005082286860581746, + "loss": 1.5016, + "theoretical_loss": 3.6092171630684335, + "tokens_seen": 1224736768 + }, + { + "epoch": 0.37, + "learning_rate": 0.0005082126379137412, + "loss": 1.4638, + "theoretical_loss": 3.609199605957051, + "tokens_seen": 1224802304 + }, + { + "epoch": 0.37, + "learning_rate": 0.000508196589769308, + "loss": 1.3734, + "theoretical_loss": 3.6091820500481044, + "tokens_seen": 1224867840 + }, + { + "epoch": 0.37, + "learning_rate": 0.0005081805416248746, + "loss": 1.3241, + "theoretical_loss": 3.6091644953414477, + "tokens_seen": 1224933376 + }, + { + "epoch": 0.37, + "learning_rate": 0.0005081644934804414, + "loss": 1.5197, + "theoretical_loss": 3.6091469418369337, + "tokens_seen": 1224998912 + }, + { + "epoch": 0.37, + "learning_rate": 0.000508148445336008, + "loss": 1.335, + "theoretical_loss": 3.6091293895344165, + "tokens_seen": 1225064448 + }, + { + "epoch": 0.37, + "learning_rate": 0.0005081323971915748, + "loss": 1.3991, + "theoretical_loss": 3.609111838433749, + "tokens_seen": 1225129984 + }, + { + "epoch": 0.37, + "learning_rate": 0.0005081163490471414, + "loss": 1.4588, + "theoretical_loss": 3.6090942885347848, + "tokens_seen": 1225195520 + }, + { + "epoch": 0.37, + "learning_rate": 0.0005081003009027082, + "loss": 1.3717, + "theoretical_loss": 3.609076739837377, + "tokens_seen": 1225261056 + }, + { + "epoch": 0.37, + "learning_rate": 0.0005080842527582749, + "loss": 1.4804, + "theoretical_loss": 3.6090591923413795, + "tokens_seen": 1225326592 + }, + { + "epoch": 0.37, + "learning_rate": 0.0005080682046138416, + "loss": 1.4683, + "theoretical_loss": 3.609041646046646, + "tokens_seen": 1225392128 + }, + { + "epoch": 0.37, + "learning_rate": 0.0005080521564694083, + "loss": 1.4997, + "theoretical_loss": 3.609024100953029, + "tokens_seen": 1225457664 + }, + { + "epoch": 0.37, + "objective/train/docs_used": 448161, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.4153162240982056, + "objective/train/theoretical_loss": 3.6090065570603826, + "objective/train/tokens_used": 1245983200, + "theoretical_loss": 3.6090065570603826, + "tokens_seen": 1225523200 + }, + { + "epoch": 0.37, + "learning_rate": 0.000508036108324975, + "loss": 1.3777, + "theoretical_loss": 3.6090065570603826, + "tokens_seen": 1225523200 + }, + { + "epoch": 0.37, + "learning_rate": 0.0005080200601805417, + "loss": 1.3572, + "theoretical_loss": 3.608989014368561, + "tokens_seen": 1225588736 + }, + { + "epoch": 0.37, + "learning_rate": 0.0005080040120361084, + "loss": 1.3894, + "theoretical_loss": 3.608971472877417, + "tokens_seen": 1225654272 + }, + { + "epoch": 0.37, + "learning_rate": 0.0005079879638916751, + "loss": 1.4111, + "theoretical_loss": 3.6089539325868047, + "tokens_seen": 1225719808 + }, + { + "epoch": 0.37, + "learning_rate": 0.0005079719157472418, + "loss": 1.3883, + "theoretical_loss": 3.6089363934965775, + "tokens_seen": 1225785344 + }, + { + "epoch": 0.37, + "learning_rate": 0.0005079558676028085, + "loss": 1.3958, + "theoretical_loss": 3.608918855606589, + "tokens_seen": 1225850880 + }, + { + "epoch": 0.37, + "learning_rate": 0.0005079398194583752, + "loss": 1.4873, + "theoretical_loss": 3.6089013189166934, + "tokens_seen": 1225916416 + }, + { + "epoch": 0.37, + "learning_rate": 0.0005079237713139419, + "loss": 1.3918, + "theoretical_loss": 3.608883783426744, + "tokens_seen": 1225981952 + }, + { + "epoch": 0.37, + "learning_rate": 0.0005079077231695086, + "loss": 1.4735, + "theoretical_loss": 3.6088662491365944, + "tokens_seen": 1226047488 + }, + { + "epoch": 0.37, + "learning_rate": 0.0005078916750250753, + "loss": 1.4201, + "theoretical_loss": 3.608848716046099, + "tokens_seen": 1226113024 + }, + { + "epoch": 0.37, + "learning_rate": 0.000507875626880642, + "loss": 1.457, + "theoretical_loss": 3.6088311841551106, + "tokens_seen": 1226178560 + }, + { + "epoch": 0.37, + "learning_rate": 0.0005078595787362086, + "loss": 1.3655, + "theoretical_loss": 3.608813653463484, + "tokens_seen": 1226244096 + }, + { + "epoch": 0.37, + "learning_rate": 0.0005078435305917754, + "loss": 1.457, + "theoretical_loss": 3.6087961239710724, + "tokens_seen": 1226309632 + }, + { + "epoch": 0.37, + "learning_rate": 0.000507827482447342, + "loss": 1.3986, + "theoretical_loss": 3.6087785956777303, + "tokens_seen": 1226375168 + }, + { + "epoch": 0.37, + "learning_rate": 0.0005078114343029088, + "loss": 1.4352, + "theoretical_loss": 3.608761068583312, + "tokens_seen": 1226440704 + }, + { + "epoch": 0.37, + "learning_rate": 0.0005077953861584754, + "loss": 1.3836, + "theoretical_loss": 3.6087435426876695, + "tokens_seen": 1226506240 + }, + { + "epoch": 0.37, + "learning_rate": 0.0005077793380140422, + "loss": 1.3076, + "theoretical_loss": 3.6087260179906586, + "tokens_seen": 1226571776 + }, + { + "epoch": 0.37, + "learning_rate": 0.0005077632898696088, + "loss": 1.3637, + "theoretical_loss": 3.6087084944921326, + "tokens_seen": 1226637312 + }, + { + "epoch": 0.37, + "learning_rate": 0.0005077472417251756, + "loss": 1.3422, + "theoretical_loss": 3.608690972191946, + "tokens_seen": 1226702848 + }, + { + "epoch": 0.37, + "learning_rate": 0.0005077311935807422, + "loss": 1.4897, + "theoretical_loss": 3.608673451089952, + "tokens_seen": 1226768384 + }, + { + "epoch": 0.37, + "learning_rate": 0.0005077151454363089, + "loss": 1.4226, + "theoretical_loss": 3.6086559311860054, + "tokens_seen": 1226833920 + }, + { + "epoch": 0.37, + "learning_rate": 0.0005076990972918756, + "loss": 1.3131, + "theoretical_loss": 3.6086384124799595, + "tokens_seen": 1226899456 + }, + { + "epoch": 0.37, + "learning_rate": 0.0005076830491474423, + "loss": 1.3702, + "theoretical_loss": 3.6086208949716694, + "tokens_seen": 1226964992 + }, + { + "epoch": 0.37, + "learning_rate": 0.000507667001003009, + "loss": 1.3272, + "theoretical_loss": 3.608603378660989, + "tokens_seen": 1227030528 + }, + { + "epoch": 0.37, + "learning_rate": 0.0005076509528585757, + "loss": 1.3352, + "theoretical_loss": 3.6085858635477717, + "tokens_seen": 1227096064 + }, + { + "epoch": 0.37, + "learning_rate": 0.0005076349047141424, + "loss": 1.5193, + "theoretical_loss": 3.608568349631872, + "tokens_seen": 1227161600 + }, + { + "epoch": 0.37, + "learning_rate": 0.0005076188565697091, + "loss": 1.4579, + "theoretical_loss": 3.6085508369131447, + "tokens_seen": 1227227136 + }, + { + "epoch": 0.37, + "learning_rate": 0.0005076028084252758, + "loss": 1.3747, + "theoretical_loss": 3.608533325391444, + "tokens_seen": 1227292672 + }, + { + "epoch": 0.37, + "learning_rate": 0.0005075867602808425, + "loss": 1.336, + "theoretical_loss": 3.6085158150666232, + "tokens_seen": 1227358208 + }, + { + "epoch": 0.37, + "learning_rate": 0.0005075707121364092, + "loss": 1.3495, + "theoretical_loss": 3.6084983059385376, + "tokens_seen": 1227423744 + }, + { + "epoch": 0.37, + "learning_rate": 0.0005075546639919759, + "loss": 1.4646, + "theoretical_loss": 3.608480798007041, + "tokens_seen": 1227489280 + }, + { + "epoch": 0.37, + "learning_rate": 0.0005075386158475426, + "loss": 1.3535, + "theoretical_loss": 3.6084632912719874, + "tokens_seen": 1227554816 + }, + { + "epoch": 0.37, + "learning_rate": 0.0005075225677031093, + "loss": 1.3866, + "theoretical_loss": 3.6084457857332324, + "tokens_seen": 1227620352 + }, + { + "epoch": 0.37, + "learning_rate": 0.000507506519558676, + "loss": 1.3424, + "theoretical_loss": 3.60842828139063, + "tokens_seen": 1227685888 + }, + { + "epoch": 0.37, + "learning_rate": 0.0005074904714142427, + "loss": 1.4162, + "theoretical_loss": 3.608410778244033, + "tokens_seen": 1227751424 + }, + { + "epoch": 0.37, + "learning_rate": 0.0005074744232698094, + "loss": 1.329, + "theoretical_loss": 3.608393276293298, + "tokens_seen": 1227816960 + }, + { + "epoch": 0.37, + "learning_rate": 0.0005074583751253761, + "loss": 1.3888, + "theoretical_loss": 3.6083757755382786, + "tokens_seen": 1227882496 + }, + { + "epoch": 0.37, + "learning_rate": 0.0005074423269809428, + "loss": 1.4271, + "theoretical_loss": 3.608358275978829, + "tokens_seen": 1227948032 + }, + { + "epoch": 0.37, + "learning_rate": 0.0005074262788365095, + "loss": 1.5431, + "theoretical_loss": 3.608340777614804, + "tokens_seen": 1228013568 + }, + { + "epoch": 0.37, + "learning_rate": 0.0005074102306920762, + "loss": 1.4171, + "theoretical_loss": 3.6083232804460583, + "tokens_seen": 1228079104 + }, + { + "epoch": 0.37, + "learning_rate": 0.0005073941825476429, + "loss": 1.4512, + "theoretical_loss": 3.6083057844724467, + "tokens_seen": 1228144640 + }, + { + "epoch": 0.37, + "learning_rate": 0.0005073781344032097, + "loss": 1.3852, + "theoretical_loss": 3.6082882896938226, + "tokens_seen": 1228210176 + }, + { + "epoch": 0.37, + "learning_rate": 0.0005073620862587763, + "loss": 1.3398, + "theoretical_loss": 3.6082707961100424, + "tokens_seen": 1228275712 + }, + { + "epoch": 0.37, + "learning_rate": 0.000507346038114343, + "loss": 1.3306, + "theoretical_loss": 3.608253303720959, + "tokens_seen": 1228341248 + }, + { + "epoch": 0.37, + "learning_rate": 0.0005073299899699097, + "loss": 1.4923, + "theoretical_loss": 3.6082358125264284, + "tokens_seen": 1228406784 + }, + { + "epoch": 0.37, + "learning_rate": 0.0005073139418254765, + "loss": 1.5426, + "theoretical_loss": 3.6082183225263043, + "tokens_seen": 1228472320 + }, + { + "epoch": 0.37, + "learning_rate": 0.0005072978936810431, + "loss": 1.4425, + "theoretical_loss": 3.608200833720443, + "tokens_seen": 1228537856 + }, + { + "epoch": 0.37, + "learning_rate": 0.0005072818455366099, + "loss": 1.3393, + "theoretical_loss": 3.608183346108697, + "tokens_seen": 1228603392 + }, + { + "epoch": 0.37, + "learning_rate": 0.0005072657973921765, + "loss": 1.3982, + "theoretical_loss": 3.6081658596909225, + "tokens_seen": 1228668928 + }, + { + "epoch": 0.37, + "learning_rate": 0.0005072497492477433, + "loss": 1.485, + "theoretical_loss": 3.6081483744669747, + "tokens_seen": 1228734464 + }, + { + "epoch": 0.37, + "objective/train/docs_used": 449484, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.513756513595581, + "objective/train/theoretical_loss": 3.6081308904367067, + "objective/train/tokens_used": 1249260000, + "theoretical_loss": 3.6081308904367067, + "tokens_seen": 1228800000 + }, + { + "epoch": 0.37, + "learning_rate": 0.0005072497492477433, + "loss": 1.5216, + "theoretical_loss": 3.6081308904367067, + "tokens_seen": 1228800000 + }, + { + "epoch": 0.37, + "learning_rate": 0.0005072337011033099, + "loss": 1.4276, + "theoretical_loss": 3.6081134075999755, + "tokens_seen": 1228865536 + }, + { + "epoch": 0.37, + "learning_rate": 0.0005072176529588767, + "loss": 1.4239, + "theoretical_loss": 3.608095925956634, + "tokens_seen": 1228931072 + }, + { + "epoch": 0.37, + "learning_rate": 0.0005072016048144433, + "loss": 1.3873, + "theoretical_loss": 3.6080784455065387, + "tokens_seen": 1228996608 + }, + { + "epoch": 0.37, + "learning_rate": 0.0005071855566700101, + "loss": 1.4184, + "theoretical_loss": 3.6080609662495435, + "tokens_seen": 1229062144 + }, + { + "epoch": 0.37, + "learning_rate": 0.0005071695085255767, + "loss": 1.3998, + "theoretical_loss": 3.608043488185504, + "tokens_seen": 1229127680 + }, + { + "epoch": 0.37, + "learning_rate": 0.0005071534603811435, + "loss": 1.3798, + "theoretical_loss": 3.6080260113142746, + "tokens_seen": 1229193216 + }, + { + "epoch": 0.37, + "learning_rate": 0.0005071374122367101, + "loss": 1.366, + "theoretical_loss": 3.6080085356357103, + "tokens_seen": 1229258752 + }, + { + "epoch": 0.37, + "learning_rate": 0.0005071213640922769, + "loss": 1.4032, + "theoretical_loss": 3.607991061149667, + "tokens_seen": 1229324288 + }, + { + "epoch": 0.37, + "learning_rate": 0.0005071053159478435, + "loss": 1.3594, + "theoretical_loss": 3.607973587855999, + "tokens_seen": 1229389824 + }, + { + "epoch": 0.37, + "learning_rate": 0.0005070892678034103, + "loss": 1.3291, + "theoretical_loss": 3.607956115754561, + "tokens_seen": 1229455360 + }, + { + "epoch": 0.37, + "learning_rate": 0.0005070732196589769, + "loss": 1.4067, + "theoretical_loss": 3.607938644845209, + "tokens_seen": 1229520896 + }, + { + "epoch": 0.37, + "learning_rate": 0.0005070571715145437, + "loss": 1.4301, + "theoretical_loss": 3.6079211751277978, + "tokens_seen": 1229586432 + }, + { + "epoch": 0.37, + "learning_rate": 0.0005070411233701103, + "loss": 1.4715, + "theoretical_loss": 3.607903706602183, + "tokens_seen": 1229651968 + }, + { + "epoch": 0.37, + "learning_rate": 0.0005070250752256771, + "loss": 1.4697, + "theoretical_loss": 3.6078862392682187, + "tokens_seen": 1229717504 + }, + { + "epoch": 0.37, + "learning_rate": 0.0005070090270812437, + "loss": 1.4458, + "theoretical_loss": 3.607868773125761, + "tokens_seen": 1229783040 + }, + { + "epoch": 0.37, + "learning_rate": 0.0005069929789368105, + "loss": 1.5, + "theoretical_loss": 3.607851308174665, + "tokens_seen": 1229848576 + }, + { + "epoch": 0.37, + "learning_rate": 0.0005069769307923771, + "loss": 1.4838, + "theoretical_loss": 3.6078338444147855, + "tokens_seen": 1229914112 + }, + { + "epoch": 0.37, + "learning_rate": 0.0005069608826479439, + "loss": 1.4753, + "theoretical_loss": 3.6078163818459785, + "tokens_seen": 1229979648 + }, + { + "epoch": 0.37, + "learning_rate": 0.0005069448345035105, + "loss": 1.4088, + "theoretical_loss": 3.607798920468099, + "tokens_seen": 1230045184 + }, + { + "epoch": 0.37, + "learning_rate": 0.0005069287863590773, + "loss": 1.466, + "theoretical_loss": 3.607781460281002, + "tokens_seen": 1230110720 + }, + { + "epoch": 0.37, + "learning_rate": 0.0005069127382146439, + "loss": 1.3765, + "theoretical_loss": 3.6077640012845427, + "tokens_seen": 1230176256 + }, + { + "epoch": 0.37, + "learning_rate": 0.0005068966900702107, + "loss": 1.4971, + "theoretical_loss": 3.6077465434785774, + "tokens_seen": 1230241792 + }, + { + "epoch": 0.37, + "learning_rate": 0.0005068806419257774, + "loss": 1.2928, + "theoretical_loss": 3.607729086862961, + "tokens_seen": 1230307328 + }, + { + "epoch": 0.37, + "learning_rate": 0.0005068645937813441, + "loss": 1.4501, + "theoretical_loss": 3.6077116314375486, + "tokens_seen": 1230372864 + }, + { + "epoch": 0.37, + "learning_rate": 0.0005068485456369108, + "loss": 1.4282, + "theoretical_loss": 3.6076941772021964, + "tokens_seen": 1230438400 + }, + { + "epoch": 0.37, + "learning_rate": 0.0005068324974924775, + "loss": 1.4641, + "theoretical_loss": 3.6076767241567596, + "tokens_seen": 1230503936 + }, + { + "epoch": 0.37, + "learning_rate": 0.0005068164493480442, + "loss": 1.4136, + "theoretical_loss": 3.6076592723010936, + "tokens_seen": 1230569472 + }, + { + "epoch": 0.37, + "learning_rate": 0.0005068004012036109, + "loss": 1.4903, + "theoretical_loss": 3.6076418216350534, + "tokens_seen": 1230635008 + }, + { + "epoch": 0.37, + "learning_rate": 0.0005067843530591776, + "loss": 1.3514, + "theoretical_loss": 3.6076243721584955, + "tokens_seen": 1230700544 + }, + { + "epoch": 0.37, + "learning_rate": 0.0005067683049147443, + "loss": 1.5352, + "theoretical_loss": 3.607606923871275, + "tokens_seen": 1230766080 + }, + { + "epoch": 0.37, + "learning_rate": 0.000506752256770311, + "loss": 1.3902, + "theoretical_loss": 3.6075894767732475, + "tokens_seen": 1230831616 + }, + { + "epoch": 0.37, + "learning_rate": 0.0005067362086258777, + "loss": 1.4242, + "theoretical_loss": 3.607572030864269, + "tokens_seen": 1230897152 + }, + { + "epoch": 0.37, + "learning_rate": 0.0005067201604814444, + "loss": 1.425, + "theoretical_loss": 3.607554586144195, + "tokens_seen": 1230962688 + }, + { + "epoch": 0.37, + "learning_rate": 0.0005067041123370111, + "loss": 1.4467, + "theoretical_loss": 3.607537142612881, + "tokens_seen": 1231028224 + }, + { + "epoch": 0.37, + "learning_rate": 0.0005066880641925778, + "loss": 1.4667, + "theoretical_loss": 3.6075197002701826, + "tokens_seen": 1231093760 + }, + { + "epoch": 0.37, + "learning_rate": 0.0005066720160481445, + "loss": 1.4117, + "theoretical_loss": 3.6075022591159556, + "tokens_seen": 1231159296 + }, + { + "epoch": 0.37, + "learning_rate": 0.0005066559679037112, + "loss": 1.4324, + "theoretical_loss": 3.6074848191500566, + "tokens_seen": 1231224832 + }, + { + "epoch": 0.37, + "learning_rate": 0.0005066399197592779, + "loss": 1.311, + "theoretical_loss": 3.60746738037234, + "tokens_seen": 1231290368 + }, + { + "epoch": 0.37, + "learning_rate": 0.0005066238716148446, + "loss": 1.3355, + "theoretical_loss": 3.6074499427826625, + "tokens_seen": 1231355904 + }, + { + "epoch": 0.37, + "learning_rate": 0.0005066078234704113, + "loss": 1.4619, + "theoretical_loss": 3.6074325063808796, + "tokens_seen": 1231421440 + }, + { + "epoch": 0.37, + "learning_rate": 0.000506591775325978, + "loss": 1.4194, + "theoretical_loss": 3.607415071166848, + "tokens_seen": 1231486976 + }, + { + "epoch": 0.37, + "learning_rate": 0.0005065757271815447, + "loss": 1.5036, + "theoretical_loss": 3.6073976371404224, + "tokens_seen": 1231552512 + }, + { + "epoch": 0.37, + "learning_rate": 0.0005065596790371114, + "loss": 1.3219, + "theoretical_loss": 3.6073802043014593, + "tokens_seen": 1231618048 + }, + { + "epoch": 0.37, + "learning_rate": 0.0005065436308926781, + "loss": 1.47, + "theoretical_loss": 3.6073627726498145, + "tokens_seen": 1231683584 + }, + { + "epoch": 0.37, + "learning_rate": 0.0005065275827482448, + "loss": 1.4171, + "theoretical_loss": 3.6073453421853445, + "tokens_seen": 1231749120 + }, + { + "epoch": 0.37, + "learning_rate": 0.0005065115346038115, + "loss": 1.4594, + "theoretical_loss": 3.607327912907904, + "tokens_seen": 1231814656 + }, + { + "epoch": 0.37, + "learning_rate": 0.0005064954864593782, + "loss": 1.3626, + "theoretical_loss": 3.6073104848173503, + "tokens_seen": 1231880192 + }, + { + "epoch": 0.37, + "learning_rate": 0.0005064794383149449, + "loss": 1.3993, + "theoretical_loss": 3.607293057913539, + "tokens_seen": 1231945728 + }, + { + "epoch": 0.37, + "learning_rate": 0.0005064633901705115, + "loss": 1.3591, + "theoretical_loss": 3.607275632196326, + "tokens_seen": 1232011264 + }, + { + "epoch": 0.37, + "objective/train/docs_used": 450737, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.4912829399108887, + "objective/train/theoretical_loss": 3.6072582076655673, + "objective/train/tokens_used": 1252536800, + "theoretical_loss": 3.6072582076655673, + "tokens_seen": 1232076800 + }, + { + "epoch": 0.37, + "learning_rate": 0.0005064473420260783, + "loss": 1.4557, + "theoretical_loss": 3.6072582076655673, + "tokens_seen": 1232076800 + }, + { + "epoch": 0.37, + "learning_rate": 0.0005064312938816449, + "loss": 1.3463, + "theoretical_loss": 3.60724078432112, + "tokens_seen": 1232142336 + }, + { + "epoch": 0.37, + "learning_rate": 0.0005064152457372117, + "loss": 1.3348, + "theoretical_loss": 3.6072233621628387, + "tokens_seen": 1232207872 + }, + { + "epoch": 0.37, + "learning_rate": 0.0005063991975927783, + "loss": 1.4636, + "theoretical_loss": 3.607205941190581, + "tokens_seen": 1232273408 + }, + { + "epoch": 0.37, + "learning_rate": 0.0005063831494483451, + "loss": 1.3343, + "theoretical_loss": 3.607188521404202, + "tokens_seen": 1232338944 + }, + { + "epoch": 0.37, + "learning_rate": 0.0005063671013039117, + "loss": 1.409, + "theoretical_loss": 3.6071711028035587, + "tokens_seen": 1232404480 + }, + { + "epoch": 0.37, + "learning_rate": 0.0005063510531594785, + "loss": 1.4499, + "theoretical_loss": 3.6071536853885062, + "tokens_seen": 1232470016 + }, + { + "epoch": 0.37, + "learning_rate": 0.0005063350050150451, + "loss": 1.3993, + "theoretical_loss": 3.6071362691589024, + "tokens_seen": 1232535552 + }, + { + "epoch": 0.37, + "learning_rate": 0.0005063189568706118, + "loss": 1.4254, + "theoretical_loss": 3.607118854114603, + "tokens_seen": 1232601088 + }, + { + "epoch": 0.37, + "learning_rate": 0.0005063029087261785, + "loss": 1.3969, + "theoretical_loss": 3.6071014402554633, + "tokens_seen": 1232666624 + }, + { + "epoch": 0.37, + "learning_rate": 0.0005062868605817452, + "loss": 1.4605, + "theoretical_loss": 3.6070840275813403, + "tokens_seen": 1232732160 + }, + { + "epoch": 0.37, + "learning_rate": 0.0005062708124373119, + "loss": 1.4472, + "theoretical_loss": 3.607066616092091, + "tokens_seen": 1232797696 + }, + { + "epoch": 0.37, + "learning_rate": 0.0005062547642928786, + "loss": 1.4903, + "theoretical_loss": 3.607049205787571, + "tokens_seen": 1232863232 + }, + { + "epoch": 0.37, + "learning_rate": 0.0005062387161484453, + "loss": 1.5403, + "theoretical_loss": 3.6070317966676377, + "tokens_seen": 1232928768 + }, + { + "epoch": 0.37, + "learning_rate": 0.000506222668004012, + "loss": 1.3496, + "theoretical_loss": 3.6070143887321455, + "tokens_seen": 1232994304 + }, + { + "epoch": 0.37, + "learning_rate": 0.0005062066198595787, + "loss": 1.3421, + "theoretical_loss": 3.6069969819809535, + "tokens_seen": 1233059840 + }, + { + "epoch": 0.37, + "learning_rate": 0.0005061905717151454, + "loss": 1.3801, + "theoretical_loss": 3.606979576413916, + "tokens_seen": 1233125376 + }, + { + "epoch": 0.37, + "learning_rate": 0.0005061745235707121, + "loss": 1.505, + "theoretical_loss": 3.606962172030891, + "tokens_seen": 1233190912 + }, + { + "epoch": 0.37, + "learning_rate": 0.0005061584754262788, + "loss": 1.4529, + "theoretical_loss": 3.606944768831734, + "tokens_seen": 1233256448 + }, + { + "epoch": 0.37, + "learning_rate": 0.0005061424272818455, + "loss": 1.4133, + "theoretical_loss": 3.6069273668163024, + "tokens_seen": 1233321984 + }, + { + "epoch": 0.37, + "learning_rate": 0.0005061263791374122, + "loss": 1.4463, + "theoretical_loss": 3.6069099659844523, + "tokens_seen": 1233387520 + }, + { + "epoch": 0.37, + "learning_rate": 0.000506110330992979, + "loss": 1.437, + "theoretical_loss": 3.6068925663360405, + "tokens_seen": 1233453056 + }, + { + "epoch": 0.37, + "learning_rate": 0.0005060942828485456, + "loss": 1.375, + "theoretical_loss": 3.606875167870923, + "tokens_seen": 1233518592 + }, + { + "epoch": 0.37, + "learning_rate": 0.0005060782347041123, + "loss": 1.3855, + "theoretical_loss": 3.6068577705889577, + "tokens_seen": 1233584128 + }, + { + "epoch": 0.37, + "learning_rate": 0.000506062186559679, + "loss": 1.5658, + "theoretical_loss": 3.606840374490001, + "tokens_seen": 1233649664 + }, + { + "epoch": 0.37, + "learning_rate": 0.0005060461384152457, + "loss": 1.3684, + "theoretical_loss": 3.606822979573908, + "tokens_seen": 1233715200 + }, + { + "epoch": 0.37, + "learning_rate": 0.0005060300902708124, + "loss": 1.3005, + "theoretical_loss": 3.6068055858405375, + "tokens_seen": 1233780736 + }, + { + "epoch": 0.37, + "learning_rate": 0.0005060140421263792, + "loss": 1.4579, + "theoretical_loss": 3.606788193289745, + "tokens_seen": 1233846272 + }, + { + "epoch": 0.37, + "learning_rate": 0.0005059979939819458, + "loss": 1.392, + "theoretical_loss": 3.606770801921388, + "tokens_seen": 1233911808 + }, + { + "epoch": 0.37, + "learning_rate": 0.0005059819458375126, + "loss": 1.43, + "theoretical_loss": 3.6067534117353235, + "tokens_seen": 1233977344 + }, + { + "epoch": 0.37, + "learning_rate": 0.0005059658976930792, + "loss": 1.4143, + "theoretical_loss": 3.6067360227314067, + "tokens_seen": 1234042880 + }, + { + "epoch": 0.37, + "learning_rate": 0.000505949849548646, + "loss": 1.4604, + "theoretical_loss": 3.606718634909497, + "tokens_seen": 1234108416 + }, + { + "epoch": 0.37, + "learning_rate": 0.0005059338014042126, + "loss": 1.4387, + "theoretical_loss": 3.606701248269449, + "tokens_seen": 1234173952 + }, + { + "epoch": 0.37, + "learning_rate": 0.0005059177532597794, + "loss": 1.3521, + "theoretical_loss": 3.6066838628111215, + "tokens_seen": 1234239488 + }, + { + "epoch": 0.37, + "learning_rate": 0.000505901705115346, + "loss": 1.3784, + "theoretical_loss": 3.6066664785343696, + "tokens_seen": 1234305024 + }, + { + "epoch": 0.37, + "learning_rate": 0.0005058856569709128, + "loss": 1.3974, + "theoretical_loss": 3.6066490954390513, + "tokens_seen": 1234370560 + }, + { + "epoch": 0.37, + "learning_rate": 0.0005058696088264794, + "loss": 1.338, + "theoretical_loss": 3.606631713525024, + "tokens_seen": 1234436096 + }, + { + "epoch": 0.37, + "learning_rate": 0.0005058535606820462, + "loss": 1.2767, + "theoretical_loss": 3.6066143327921436, + "tokens_seen": 1234501632 + }, + { + "epoch": 0.37, + "learning_rate": 0.0005058375125376128, + "loss": 1.424, + "theoretical_loss": 3.606596953240268, + "tokens_seen": 1234567168 + }, + { + "epoch": 0.37, + "learning_rate": 0.0005058214643931796, + "loss": 1.4036, + "theoretical_loss": 3.6065795748692544, + "tokens_seen": 1234632704 + }, + { + "epoch": 0.37, + "learning_rate": 0.0005058054162487462, + "loss": 1.4366, + "theoretical_loss": 3.606562197678959, + "tokens_seen": 1234698240 + }, + { + "epoch": 0.37, + "learning_rate": 0.000505789368104313, + "loss": 1.3684, + "theoretical_loss": 3.6065448216692393, + "tokens_seen": 1234763776 + }, + { + "epoch": 0.37, + "learning_rate": 0.0005057733199598796, + "loss": 1.4509, + "theoretical_loss": 3.606527446839953, + "tokens_seen": 1234829312 + }, + { + "epoch": 0.37, + "learning_rate": 0.0005057572718154464, + "loss": 1.4256, + "theoretical_loss": 3.6065100731909565, + "tokens_seen": 1234894848 + }, + { + "epoch": 0.37, + "learning_rate": 0.000505741223671013, + "loss": 1.3975, + "theoretical_loss": 3.6064927007221073, + "tokens_seen": 1234960384 + }, + { + "epoch": 0.37, + "learning_rate": 0.0005057251755265798, + "loss": 1.483, + "theoretical_loss": 3.6064753294332625, + "tokens_seen": 1235025920 + }, + { + "epoch": 0.37, + "learning_rate": 0.0005057091273821464, + "loss": 1.421, + "theoretical_loss": 3.6064579593242794, + "tokens_seen": 1235091456 + }, + { + "epoch": 0.37, + "learning_rate": 0.0005056930792377132, + "loss": 1.3747, + "theoretical_loss": 3.6064405903950156, + "tokens_seen": 1235156992 + }, + { + "epoch": 0.37, + "learning_rate": 0.0005056770310932799, + "loss": 1.4614, + "theoretical_loss": 3.6064232226453283, + "tokens_seen": 1235222528 + }, + { + "epoch": 0.37, + "learning_rate": 0.0005056609829488466, + "loss": 1.4623, + "theoretical_loss": 3.606405856075074, + "tokens_seen": 1235288064 + }, + { + "epoch": 0.37, + "objective/train/docs_used": 451987, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.4472036361694336, + "objective/train/theoretical_loss": 3.606388490684111, + "objective/train/tokens_used": 1255813600, + "theoretical_loss": 3.606388490684111, + "tokens_seen": 1235353600 + }, + { + "epoch": 0.37, + "learning_rate": 0.0005056449348044133, + "loss": 1.4931, + "theoretical_loss": 3.606388490684111, + "tokens_seen": 1235353600 + }, + { + "epoch": 0.37, + "learning_rate": 0.00050562888665998, + "loss": 1.3984, + "theoretical_loss": 3.606371126472296, + "tokens_seen": 1235419136 + }, + { + "epoch": 0.37, + "learning_rate": 0.0005056128385155467, + "loss": 1.4619, + "theoretical_loss": 3.606353763439487, + "tokens_seen": 1235484672 + }, + { + "epoch": 0.37, + "learning_rate": 0.0005055967903711134, + "loss": 1.3309, + "theoretical_loss": 3.6063364015855415, + "tokens_seen": 1235550208 + }, + { + "epoch": 0.37, + "learning_rate": 0.0005055807422266801, + "loss": 1.5326, + "theoretical_loss": 3.606319040910316, + "tokens_seen": 1235615744 + }, + { + "epoch": 0.37, + "learning_rate": 0.0005055646940822468, + "loss": 1.4123, + "theoretical_loss": 3.606301681413669, + "tokens_seen": 1235681280 + }, + { + "epoch": 0.37, + "learning_rate": 0.0005055486459378135, + "loss": 1.4144, + "theoretical_loss": 3.6062843230954567, + "tokens_seen": 1235746816 + }, + { + "epoch": 0.37, + "learning_rate": 0.0005055325977933802, + "loss": 1.4429, + "theoretical_loss": 3.606266965955538, + "tokens_seen": 1235812352 + }, + { + "epoch": 0.37, + "learning_rate": 0.0005055165496489469, + "loss": 1.5204, + "theoretical_loss": 3.6062496099937698, + "tokens_seen": 1235877888 + }, + { + "epoch": 0.37, + "learning_rate": 0.0005055005015045136, + "loss": 1.364, + "theoretical_loss": 3.60623225521001, + "tokens_seen": 1235943424 + }, + { + "epoch": 0.37, + "learning_rate": 0.0005054844533600803, + "loss": 1.424, + "theoretical_loss": 3.6062149016041154, + "tokens_seen": 1236008960 + }, + { + "epoch": 0.37, + "learning_rate": 0.000505468405215647, + "loss": 1.4121, + "theoretical_loss": 3.606197549175944, + "tokens_seen": 1236074496 + }, + { + "epoch": 0.37, + "learning_rate": 0.0005054523570712137, + "loss": 1.3063, + "theoretical_loss": 3.606180197925354, + "tokens_seen": 1236140032 + }, + { + "epoch": 0.37, + "learning_rate": 0.0005054363089267804, + "loss": 1.4195, + "theoretical_loss": 3.6061628478522025, + "tokens_seen": 1236205568 + }, + { + "epoch": 0.37, + "learning_rate": 0.0005054202607823471, + "loss": 1.4329, + "theoretical_loss": 3.606145498956347, + "tokens_seen": 1236271104 + }, + { + "epoch": 0.37, + "learning_rate": 0.0005054042126379138, + "loss": 1.5209, + "theoretical_loss": 3.6061281512376464, + "tokens_seen": 1236336640 + }, + { + "epoch": 0.37, + "learning_rate": 0.0005053881644934805, + "loss": 1.4928, + "theoretical_loss": 3.6061108046959567, + "tokens_seen": 1236402176 + }, + { + "epoch": 0.37, + "learning_rate": 0.0005053721163490472, + "loss": 1.4158, + "theoretical_loss": 3.6060934593311367, + "tokens_seen": 1236467712 + }, + { + "epoch": 0.37, + "learning_rate": 0.0005053560682046139, + "loss": 1.4123, + "theoretical_loss": 3.6060761151430443, + "tokens_seen": 1236533248 + }, + { + "epoch": 0.37, + "learning_rate": 0.0005053400200601806, + "loss": 1.3646, + "theoretical_loss": 3.6060587721315365, + "tokens_seen": 1236598784 + }, + { + "epoch": 0.37, + "learning_rate": 0.0005053239719157473, + "loss": 1.4049, + "theoretical_loss": 3.6060414302964716, + "tokens_seen": 1236664320 + }, + { + "epoch": 0.37, + "learning_rate": 0.000505307923771314, + "loss": 1.5021, + "theoretical_loss": 3.6060240896377085, + "tokens_seen": 1236729856 + }, + { + "epoch": 0.37, + "learning_rate": 0.0005052918756268807, + "loss": 1.512, + "theoretical_loss": 3.606006750155103, + "tokens_seen": 1236795392 + }, + { + "epoch": 0.37, + "learning_rate": 0.0005052758274824474, + "loss": 1.3964, + "theoretical_loss": 3.6059894118485145, + "tokens_seen": 1236860928 + }, + { + "epoch": 0.37, + "learning_rate": 0.0005052597793380141, + "loss": 1.3833, + "theoretical_loss": 3.6059720747178003, + "tokens_seen": 1236926464 + }, + { + "epoch": 0.37, + "learning_rate": 0.0005052437311935808, + "loss": 1.3431, + "theoretical_loss": 3.605954738762819, + "tokens_seen": 1236992000 + }, + { + "epoch": 0.37, + "learning_rate": 0.0005052276830491475, + "loss": 1.4603, + "theoretical_loss": 3.605937403983428, + "tokens_seen": 1237057536 + }, + { + "epoch": 0.37, + "learning_rate": 0.0005052116349047142, + "loss": 1.3086, + "theoretical_loss": 3.6059200703794856, + "tokens_seen": 1237123072 + }, + { + "epoch": 0.37, + "learning_rate": 0.0005051955867602809, + "loss": 1.4107, + "theoretical_loss": 3.6059027379508493, + "tokens_seen": 1237188608 + }, + { + "epoch": 0.37, + "learning_rate": 0.0005051795386158476, + "loss": 1.3164, + "theoretical_loss": 3.605885406697378, + "tokens_seen": 1237254144 + }, + { + "epoch": 0.37, + "learning_rate": 0.0005051634904714143, + "loss": 1.3863, + "theoretical_loss": 3.605868076618929, + "tokens_seen": 1237319680 + }, + { + "epoch": 0.37, + "learning_rate": 0.000505147442326981, + "loss": 1.4292, + "theoretical_loss": 3.605850747715361, + "tokens_seen": 1237385216 + }, + { + "epoch": 0.37, + "learning_rate": 0.0005051313941825477, + "loss": 1.3633, + "theoretical_loss": 3.6058334199865314, + "tokens_seen": 1237450752 + }, + { + "epoch": 0.38, + "learning_rate": 0.0005051153460381144, + "loss": 1.4203, + "theoretical_loss": 3.605816093432299, + "tokens_seen": 1237516288 + }, + { + "epoch": 0.38, + "learning_rate": 0.0005050992978936811, + "loss": 1.3289, + "theoretical_loss": 3.605798768052522, + "tokens_seen": 1237581824 + }, + { + "epoch": 0.38, + "learning_rate": 0.0005050832497492478, + "loss": 1.3571, + "theoretical_loss": 3.6057814438470586, + "tokens_seen": 1237647360 + }, + { + "epoch": 0.38, + "learning_rate": 0.0005050672016048145, + "loss": 1.3569, + "theoretical_loss": 3.6057641208157665, + "tokens_seen": 1237712896 + }, + { + "epoch": 0.38, + "learning_rate": 0.0005050511534603812, + "loss": 1.4539, + "theoretical_loss": 3.6057467989585046, + "tokens_seen": 1237778432 + }, + { + "epoch": 0.38, + "learning_rate": 0.0005050351053159478, + "loss": 1.4018, + "theoretical_loss": 3.605729478275131, + "tokens_seen": 1237843968 + }, + { + "epoch": 0.38, + "learning_rate": 0.0005050190571715146, + "loss": 1.4259, + "theoretical_loss": 3.6057121587655034, + "tokens_seen": 1237909504 + }, + { + "epoch": 0.38, + "learning_rate": 0.0005050030090270812, + "loss": 1.4011, + "theoretical_loss": 3.6056948404294804, + "tokens_seen": 1237975040 + }, + { + "epoch": 0.38, + "learning_rate": 0.000504986960882648, + "loss": 1.3802, + "theoretical_loss": 3.6056775232669214, + "tokens_seen": 1238040576 + }, + { + "epoch": 0.38, + "learning_rate": 0.0005049709127382146, + "loss": 1.3907, + "theoretical_loss": 3.6056602072776833, + "tokens_seen": 1238106112 + }, + { + "epoch": 0.38, + "learning_rate": 0.0005049548645937814, + "loss": 1.4713, + "theoretical_loss": 3.6056428924616255, + "tokens_seen": 1238171648 + }, + { + "epoch": 0.38, + "learning_rate": 0.000504938816449348, + "loss": 1.296, + "theoretical_loss": 3.605625578818606, + "tokens_seen": 1238237184 + }, + { + "epoch": 0.38, + "learning_rate": 0.0005049227683049147, + "loss": 1.4154, + "theoretical_loss": 3.605608266348483, + "tokens_seen": 1238302720 + }, + { + "epoch": 0.38, + "learning_rate": 0.0005049067201604814, + "loss": 1.3884, + "theoretical_loss": 3.6055909550511154, + "tokens_seen": 1238368256 + }, + { + "epoch": 0.38, + "learning_rate": 0.0005048906720160481, + "loss": 1.3866, + "theoretical_loss": 3.605573644926362, + "tokens_seen": 1238433792 + }, + { + "epoch": 0.38, + "learning_rate": 0.0005048746238716148, + "loss": 1.5238, + "theoretical_loss": 3.6055563359740805, + "tokens_seen": 1238499328 + }, + { + "epoch": 0.38, + "learning_rate": 0.0005048585757271815, + "loss": 1.4288, + "theoretical_loss": 3.60553902819413, + "tokens_seen": 1238564864 + }, + { + "epoch": 0.38, + "objective/train/docs_used": 453261, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.5689985752105713, + "objective/train/theoretical_loss": 3.605521721586369, + "objective/train/tokens_used": 1259090400, + "theoretical_loss": 3.605521721586369, + "tokens_seen": 1238630400 + }, + { + "epoch": 0.38, + "learning_rate": 0.0005048425275827482, + "loss": 1.5097, + "theoretical_loss": 3.605521721586369, + "tokens_seen": 1238630400 + }, + { + "epoch": 0.38, + "learning_rate": 0.0005048264794383149, + "loss": 1.3994, + "theoretical_loss": 3.605504416150656, + "tokens_seen": 1238695936 + }, + { + "epoch": 0.38, + "learning_rate": 0.0005048104312938816, + "loss": 1.471, + "theoretical_loss": 3.6054871118868492, + "tokens_seen": 1238761472 + }, + { + "epoch": 0.38, + "learning_rate": 0.0005047943831494483, + "loss": 1.393, + "theoretical_loss": 3.6054698087948083, + "tokens_seen": 1238827008 + }, + { + "epoch": 0.38, + "learning_rate": 0.000504778335005015, + "loss": 1.3814, + "theoretical_loss": 3.605452506874391, + "tokens_seen": 1238892544 + }, + { + "epoch": 0.38, + "learning_rate": 0.0005047622868605817, + "loss": 1.2828, + "theoretical_loss": 3.6054352061254566, + "tokens_seen": 1238958080 + }, + { + "epoch": 0.38, + "learning_rate": 0.0005047462387161484, + "loss": 1.4774, + "theoretical_loss": 3.6054179065478635, + "tokens_seen": 1239023616 + }, + { + "epoch": 0.38, + "learning_rate": 0.0005047301905717151, + "loss": 1.3564, + "theoretical_loss": 3.605400608141471, + "tokens_seen": 1239089152 + }, + { + "epoch": 0.38, + "learning_rate": 0.0005047141424272818, + "loss": 1.4539, + "theoretical_loss": 3.6053833109061366, + "tokens_seen": 1239154688 + }, + { + "epoch": 0.38, + "learning_rate": 0.0005046980942828485, + "loss": 1.352, + "theoretical_loss": 3.605366014841721, + "tokens_seen": 1239220224 + }, + { + "epoch": 0.38, + "learning_rate": 0.0005046820461384152, + "loss": 1.5702, + "theoretical_loss": 3.605348719948081, + "tokens_seen": 1239285760 + }, + { + "epoch": 0.38, + "learning_rate": 0.0005046659979939819, + "loss": 1.3108, + "theoretical_loss": 3.605331426225076, + "tokens_seen": 1239351296 + }, + { + "epoch": 0.38, + "learning_rate": 0.0005046499498495487, + "loss": 1.5243, + "theoretical_loss": 3.6053141336725663, + "tokens_seen": 1239416832 + }, + { + "epoch": 0.38, + "learning_rate": 0.0005046339017051153, + "loss": 1.464, + "theoretical_loss": 3.605296842290409, + "tokens_seen": 1239482368 + }, + { + "epoch": 0.38, + "learning_rate": 0.000504617853560682, + "loss": 1.3305, + "theoretical_loss": 3.605279552078464, + "tokens_seen": 1239547904 + }, + { + "epoch": 0.38, + "learning_rate": 0.0005046018054162487, + "loss": 1.3707, + "theoretical_loss": 3.6052622630365896, + "tokens_seen": 1239613440 + }, + { + "epoch": 0.38, + "learning_rate": 0.0005045857572718155, + "loss": 1.3567, + "theoretical_loss": 3.605244975164646, + "tokens_seen": 1239678976 + }, + { + "epoch": 0.38, + "learning_rate": 0.0005045697091273821, + "loss": 1.3952, + "theoretical_loss": 3.6052276884624908, + "tokens_seen": 1239744512 + }, + { + "epoch": 0.38, + "learning_rate": 0.0005045536609829489, + "loss": 1.3375, + "theoretical_loss": 3.605210402929983, + "tokens_seen": 1239810048 + }, + { + "epoch": 0.38, + "learning_rate": 0.0005045376128385155, + "loss": 1.3248, + "theoretical_loss": 3.6051931185669828, + "tokens_seen": 1239875584 + }, + { + "epoch": 0.38, + "learning_rate": 0.0005045215646940823, + "loss": 1.4706, + "theoretical_loss": 3.6051758353733487, + "tokens_seen": 1239941120 + }, + { + "epoch": 0.38, + "learning_rate": 0.0005045055165496489, + "loss": 1.4192, + "theoretical_loss": 3.6051585533489394, + "tokens_seen": 1240006656 + }, + { + "epoch": 0.38, + "learning_rate": 0.0005044894684052157, + "loss": 1.45, + "theoretical_loss": 3.605141272493615, + "tokens_seen": 1240072192 + }, + { + "epoch": 0.38, + "learning_rate": 0.0005044734202607824, + "loss": 1.3933, + "theoretical_loss": 3.6051239928072336, + "tokens_seen": 1240137728 + }, + { + "epoch": 0.38, + "learning_rate": 0.0005044573721163491, + "loss": 1.4157, + "theoretical_loss": 3.6051067142896542, + "tokens_seen": 1240203264 + }, + { + "epoch": 0.38, + "learning_rate": 0.0005044413239719158, + "loss": 1.3273, + "theoretical_loss": 3.6050894369407374, + "tokens_seen": 1240268800 + }, + { + "epoch": 0.38, + "learning_rate": 0.0005044252758274825, + "loss": 1.4319, + "theoretical_loss": 3.605072160760341, + "tokens_seen": 1240334336 + }, + { + "epoch": 0.38, + "learning_rate": 0.0005044092276830492, + "loss": 1.416, + "theoretical_loss": 3.605054885748325, + "tokens_seen": 1240399872 + }, + { + "epoch": 0.38, + "learning_rate": 0.0005043931795386159, + "loss": 1.5148, + "theoretical_loss": 3.605037611904548, + "tokens_seen": 1240465408 + }, + { + "epoch": 0.38, + "learning_rate": 0.0005043771313941826, + "loss": 1.4563, + "theoretical_loss": 3.6050203392288704, + "tokens_seen": 1240530944 + }, + { + "epoch": 0.38, + "learning_rate": 0.0005043610832497493, + "loss": 1.5168, + "theoretical_loss": 3.6050030677211504, + "tokens_seen": 1240596480 + }, + { + "epoch": 0.38, + "learning_rate": 0.000504345035105316, + "loss": 1.4058, + "theoretical_loss": 3.604985797381248, + "tokens_seen": 1240662016 + }, + { + "epoch": 0.38, + "learning_rate": 0.0005043289869608827, + "loss": 1.4376, + "theoretical_loss": 3.604968528209022, + "tokens_seen": 1240727552 + }, + { + "epoch": 0.38, + "learning_rate": 0.0005043129388164494, + "loss": 1.3719, + "theoretical_loss": 3.6049512602043325, + "tokens_seen": 1240793088 + }, + { + "epoch": 0.38, + "learning_rate": 0.0005042968906720161, + "loss": 1.3214, + "theoretical_loss": 3.604933993367038, + "tokens_seen": 1240858624 + }, + { + "epoch": 0.38, + "learning_rate": 0.0005042808425275828, + "loss": 1.4107, + "theoretical_loss": 3.6049167276969984, + "tokens_seen": 1240924160 + }, + { + "epoch": 0.38, + "learning_rate": 0.0005042647943831495, + "loss": 1.3251, + "theoretical_loss": 3.604899463194074, + "tokens_seen": 1240989696 + }, + { + "epoch": 0.38, + "learning_rate": 0.0005042487462387162, + "loss": 1.4452, + "theoretical_loss": 3.604882199858123, + "tokens_seen": 1241055232 + }, + { + "epoch": 0.38, + "learning_rate": 0.0005042326980942829, + "loss": 1.39, + "theoretical_loss": 3.6048649376890047, + "tokens_seen": 1241120768 + }, + { + "epoch": 0.38, + "learning_rate": 0.0005042166499498496, + "loss": 1.383, + "theoretical_loss": 3.60484767668658, + "tokens_seen": 1241186304 + }, + { + "epoch": 0.38, + "learning_rate": 0.0005042006018054163, + "loss": 1.4435, + "theoretical_loss": 3.6048304168507075, + "tokens_seen": 1241251840 + }, + { + "epoch": 0.38, + "learning_rate": 0.000504184553660983, + "loss": 1.4734, + "theoretical_loss": 3.604813158181247, + "tokens_seen": 1241317376 + }, + { + "epoch": 0.38, + "learning_rate": 0.0005041685055165497, + "loss": 1.4043, + "theoretical_loss": 3.6047959006780586, + "tokens_seen": 1241382912 + }, + { + "epoch": 0.38, + "learning_rate": 0.0005041524573721164, + "loss": 1.4408, + "theoretical_loss": 3.6047786443410006, + "tokens_seen": 1241448448 + }, + { + "epoch": 0.38, + "learning_rate": 0.0005041364092276831, + "loss": 1.5267, + "theoretical_loss": 3.6047613891699335, + "tokens_seen": 1241513984 + }, + { + "epoch": 0.38, + "learning_rate": 0.0005041203610832498, + "loss": 1.4788, + "theoretical_loss": 3.604744135164718, + "tokens_seen": 1241579520 + }, + { + "epoch": 0.38, + "learning_rate": 0.0005041043129388165, + "loss": 1.4345, + "theoretical_loss": 3.6047268823252114, + "tokens_seen": 1241645056 + }, + { + "epoch": 0.38, + "learning_rate": 0.0005040882647943832, + "loss": 1.48, + "theoretical_loss": 3.604709630651276, + "tokens_seen": 1241710592 + }, + { + "epoch": 0.38, + "learning_rate": 0.0005040722166499499, + "loss": 1.4268, + "theoretical_loss": 3.6046923801427697, + "tokens_seen": 1241776128 + }, + { + "epoch": 0.38, + "learning_rate": 0.0005040561685055166, + "loss": 1.4565, + "theoretical_loss": 3.604675130799553, + "tokens_seen": 1241841664 + }, + { + "epoch": 0.38, + "objective/train/docs_used": 454384, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.4378201961517334, + "objective/train/theoretical_loss": 3.6046578826214852, + "objective/train/tokens_used": 1262367200, + "theoretical_loss": 3.6046578826214852, + "tokens_seen": 1241907200 + }, + { + "epoch": 0.38, + "learning_rate": 0.0005040401203610833, + "loss": 1.3855, + "theoretical_loss": 3.6046578826214852, + "tokens_seen": 1241907200 + }, + { + "epoch": 0.38, + "learning_rate": 0.00050402407221665, + "loss": 1.4651, + "theoretical_loss": 3.604640635608427, + "tokens_seen": 1241972736 + }, + { + "epoch": 0.38, + "learning_rate": 0.0005040080240722167, + "loss": 1.4323, + "theoretical_loss": 3.6046233897602376, + "tokens_seen": 1242038272 + }, + { + "epoch": 0.38, + "learning_rate": 0.0005039919759277834, + "loss": 1.4249, + "theoretical_loss": 3.604606145076777, + "tokens_seen": 1242103808 + }, + { + "epoch": 0.38, + "learning_rate": 0.0005039759277833501, + "loss": 1.435, + "theoretical_loss": 3.6045889015579045, + "tokens_seen": 1242169344 + }, + { + "epoch": 0.38, + "learning_rate": 0.0005039598796389168, + "loss": 1.4186, + "theoretical_loss": 3.6045716592034815, + "tokens_seen": 1242234880 + }, + { + "epoch": 0.38, + "learning_rate": 0.0005039438314944835, + "loss": 1.3669, + "theoretical_loss": 3.6045544180133664, + "tokens_seen": 1242300416 + }, + { + "epoch": 0.38, + "learning_rate": 0.0005039277833500502, + "loss": 1.4279, + "theoretical_loss": 3.6045371779874205, + "tokens_seen": 1242365952 + }, + { + "epoch": 0.38, + "learning_rate": 0.0005039117352056169, + "loss": 1.3385, + "theoretical_loss": 3.6045199391255025, + "tokens_seen": 1242431488 + }, + { + "epoch": 0.38, + "learning_rate": 0.0005038956870611836, + "loss": 1.469, + "theoretical_loss": 3.6045027014274735, + "tokens_seen": 1242497024 + }, + { + "epoch": 0.38, + "learning_rate": 0.0005038796389167503, + "loss": 1.3412, + "theoretical_loss": 3.6044854648931928, + "tokens_seen": 1242562560 + }, + { + "epoch": 0.38, + "learning_rate": 0.000503863590772317, + "loss": 1.5676, + "theoretical_loss": 3.6044682295225203, + "tokens_seen": 1242628096 + }, + { + "epoch": 0.38, + "learning_rate": 0.0005038475426278837, + "loss": 1.4261, + "theoretical_loss": 3.604450995315317, + "tokens_seen": 1242693632 + }, + { + "epoch": 0.38, + "learning_rate": 0.0005038314944834504, + "loss": 1.4585, + "theoretical_loss": 3.604433762271442, + "tokens_seen": 1242759168 + }, + { + "epoch": 0.38, + "learning_rate": 0.0005038154463390171, + "loss": 1.3931, + "theoretical_loss": 3.6044165303907567, + "tokens_seen": 1242824704 + }, + { + "epoch": 0.38, + "learning_rate": 0.0005037993981945838, + "loss": 1.4963, + "theoretical_loss": 3.60439929967312, + "tokens_seen": 1242890240 + }, + { + "epoch": 0.38, + "learning_rate": 0.0005037833500501505, + "loss": 1.5081, + "theoretical_loss": 3.6043820701183926, + "tokens_seen": 1242955776 + }, + { + "epoch": 0.38, + "learning_rate": 0.0005037673019057172, + "loss": 1.3793, + "theoretical_loss": 3.6043648417264347, + "tokens_seen": 1243021312 + }, + { + "epoch": 0.38, + "learning_rate": 0.0005037512537612839, + "loss": 1.491, + "theoretical_loss": 3.6043476144971063, + "tokens_seen": 1243086848 + }, + { + "epoch": 0.38, + "learning_rate": 0.0005037352056168506, + "loss": 1.4657, + "theoretical_loss": 3.6043303884302684, + "tokens_seen": 1243152384 + }, + { + "epoch": 0.38, + "learning_rate": 0.0005037191574724173, + "loss": 1.397, + "theoretical_loss": 3.604313163525781, + "tokens_seen": 1243217920 + }, + { + "epoch": 0.38, + "learning_rate": 0.000503703109327984, + "loss": 1.3682, + "theoretical_loss": 3.604295939783503, + "tokens_seen": 1243283456 + }, + { + "epoch": 0.38, + "learning_rate": 0.0005036870611835507, + "loss": 1.2301, + "theoretical_loss": 3.6042787172032966, + "tokens_seen": 1243348992 + }, + { + "epoch": 0.38, + "learning_rate": 0.0005036710130391174, + "loss": 1.4072, + "theoretical_loss": 3.604261495785021, + "tokens_seen": 1243414528 + }, + { + "epoch": 0.38, + "learning_rate": 0.0005036549648946841, + "loss": 1.3656, + "theoretical_loss": 3.6042442755285373, + "tokens_seen": 1243480064 + }, + { + "epoch": 0.38, + "learning_rate": 0.0005036389167502507, + "loss": 1.4584, + "theoretical_loss": 3.6042270564337056, + "tokens_seen": 1243545600 + }, + { + "epoch": 0.38, + "learning_rate": 0.0005036228686058175, + "loss": 1.4201, + "theoretical_loss": 3.6042098385003865, + "tokens_seen": 1243611136 + }, + { + "epoch": 0.38, + "learning_rate": 0.0005036068204613841, + "loss": 1.4513, + "theoretical_loss": 3.6041926217284397, + "tokens_seen": 1243676672 + }, + { + "epoch": 0.38, + "learning_rate": 0.0005035907723169509, + "loss": 1.4801, + "theoretical_loss": 3.6041754061177267, + "tokens_seen": 1243742208 + }, + { + "epoch": 0.38, + "learning_rate": 0.0005035747241725175, + "loss": 1.4901, + "theoretical_loss": 3.604158191668107, + "tokens_seen": 1243807744 + }, + { + "epoch": 0.38, + "learning_rate": 0.0005035586760280843, + "loss": 1.4132, + "theoretical_loss": 3.6041409783794425, + "tokens_seen": 1243873280 + }, + { + "epoch": 0.38, + "learning_rate": 0.0005035426278836509, + "loss": 1.4372, + "theoretical_loss": 3.6041237662515924, + "tokens_seen": 1243938816 + }, + { + "epoch": 0.38, + "learning_rate": 0.0005035265797392177, + "loss": 1.4481, + "theoretical_loss": 3.6041065552844174, + "tokens_seen": 1244004352 + }, + { + "epoch": 0.38, + "learning_rate": 0.0005035105315947843, + "loss": 1.4117, + "theoretical_loss": 3.604089345477779, + "tokens_seen": 1244069888 + }, + { + "epoch": 0.38, + "learning_rate": 0.000503494483450351, + "loss": 1.5379, + "theoretical_loss": 3.6040721368315367, + "tokens_seen": 1244135424 + }, + { + "epoch": 0.38, + "learning_rate": 0.0005034784353059177, + "loss": 1.3154, + "theoretical_loss": 3.604054929345552, + "tokens_seen": 1244200960 + }, + { + "epoch": 0.38, + "learning_rate": 0.0005034623871614844, + "loss": 1.6081, + "theoretical_loss": 3.604037723019686, + "tokens_seen": 1244266496 + }, + { + "epoch": 0.38, + "learning_rate": 0.0005034463390170511, + "loss": 1.5006, + "theoretical_loss": 3.6040205178537974, + "tokens_seen": 1244332032 + }, + { + "epoch": 0.38, + "learning_rate": 0.0005034302908726178, + "loss": 1.3779, + "theoretical_loss": 3.6040033138477487, + "tokens_seen": 1244397568 + }, + { + "epoch": 0.38, + "learning_rate": 0.0005034142427281845, + "loss": 1.4847, + "theoretical_loss": 3.6039861110014004, + "tokens_seen": 1244463104 + }, + { + "epoch": 0.38, + "learning_rate": 0.0005033981945837512, + "loss": 1.2913, + "theoretical_loss": 3.6039689093146126, + "tokens_seen": 1244528640 + }, + { + "epoch": 0.38, + "learning_rate": 0.000503382146439318, + "loss": 1.4385, + "theoretical_loss": 3.6039517087872466, + "tokens_seen": 1244594176 + }, + { + "epoch": 0.38, + "learning_rate": 0.0005033660982948846, + "loss": 1.4826, + "theoretical_loss": 3.603934509419163, + "tokens_seen": 1244659712 + }, + { + "epoch": 0.38, + "learning_rate": 0.0005033500501504513, + "loss": 1.3782, + "theoretical_loss": 3.603917311210223, + "tokens_seen": 1244725248 + }, + { + "epoch": 0.38, + "learning_rate": 0.000503334002006018, + "loss": 1.3827, + "theoretical_loss": 3.603900114160287, + "tokens_seen": 1244790784 + }, + { + "epoch": 0.38, + "learning_rate": 0.0005033179538615847, + "loss": 1.3928, + "theoretical_loss": 3.6038829182692154, + "tokens_seen": 1244856320 + }, + { + "epoch": 0.38, + "learning_rate": 0.0005033019057171515, + "loss": 1.4423, + "theoretical_loss": 3.6038657235368703, + "tokens_seen": 1244921856 + }, + { + "epoch": 0.38, + "learning_rate": 0.0005032858575727182, + "loss": 1.4258, + "theoretical_loss": 3.603848529963112, + "tokens_seen": 1244987392 + }, + { + "epoch": 0.38, + "learning_rate": 0.0005032698094282849, + "loss": 1.3729, + "theoretical_loss": 3.6038313375478013, + "tokens_seen": 1245052928 + }, + { + "epoch": 0.38, + "learning_rate": 0.0005032537612838516, + "loss": 1.3348, + "theoretical_loss": 3.6038141462907998, + "tokens_seen": 1245118464 + }, + { + "epoch": 0.38, + "objective/train/docs_used": 455531, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.4709217548370361, + "objective/train/theoretical_loss": 3.6037969561919674, + "objective/train/tokens_used": 1265644000, + "theoretical_loss": 3.6037969561919674, + "tokens_seen": 1245184000 + }, + { + "epoch": 0.38, + "learning_rate": 0.0005032377131394183, + "loss": 1.5283, + "theoretical_loss": 3.6037969561919674, + "tokens_seen": 1245184000 + }, + { + "epoch": 0.38, + "learning_rate": 0.000503221664994985, + "loss": 1.3923, + "theoretical_loss": 3.6037797672511664, + "tokens_seen": 1245249536 + }, + { + "epoch": 0.38, + "learning_rate": 0.0005032056168505517, + "loss": 1.5236, + "theoretical_loss": 3.603762579468257, + "tokens_seen": 1245315072 + }, + { + "epoch": 0.38, + "learning_rate": 0.0005031895687061184, + "loss": 1.4223, + "theoretical_loss": 3.6037453928431007, + "tokens_seen": 1245380608 + }, + { + "epoch": 0.38, + "learning_rate": 0.0005031735205616851, + "loss": 1.352, + "theoretical_loss": 3.6037282073755583, + "tokens_seen": 1245446144 + }, + { + "epoch": 0.38, + "learning_rate": 0.0005031574724172518, + "loss": 1.5386, + "theoretical_loss": 3.6037110230654914, + "tokens_seen": 1245511680 + }, + { + "epoch": 0.38, + "learning_rate": 0.0005031414242728185, + "loss": 1.3452, + "theoretical_loss": 3.6036938399127605, + "tokens_seen": 1245577216 + }, + { + "epoch": 0.38, + "learning_rate": 0.0005031253761283852, + "loss": 1.5642, + "theoretical_loss": 3.6036766579172266, + "tokens_seen": 1245642752 + }, + { + "epoch": 0.38, + "learning_rate": 0.0005031093279839519, + "loss": 1.4064, + "theoretical_loss": 3.603659477078752, + "tokens_seen": 1245708288 + }, + { + "epoch": 0.38, + "learning_rate": 0.0005030932798395186, + "loss": 1.3926, + "theoretical_loss": 3.6036422973971973, + "tokens_seen": 1245773824 + }, + { + "epoch": 0.38, + "learning_rate": 0.0005030772316950853, + "loss": 1.4931, + "theoretical_loss": 3.603625118872423, + "tokens_seen": 1245839360 + }, + { + "epoch": 0.38, + "learning_rate": 0.000503061183550652, + "loss": 1.3576, + "theoretical_loss": 3.6036079415042916, + "tokens_seen": 1245904896 + }, + { + "epoch": 0.38, + "learning_rate": 0.0005030451354062187, + "loss": 1.4724, + "theoretical_loss": 3.603590765292664, + "tokens_seen": 1245970432 + }, + { + "epoch": 0.38, + "learning_rate": 0.0005030290872617854, + "loss": 1.3429, + "theoretical_loss": 3.603573590237401, + "tokens_seen": 1246035968 + }, + { + "epoch": 0.38, + "learning_rate": 0.0005030130391173521, + "loss": 1.4392, + "theoretical_loss": 3.603556416338365, + "tokens_seen": 1246101504 + }, + { + "epoch": 0.38, + "learning_rate": 0.0005029969909729188, + "loss": 1.5254, + "theoretical_loss": 3.6035392435954154, + "tokens_seen": 1246167040 + }, + { + "epoch": 0.38, + "learning_rate": 0.0005029809428284855, + "loss": 1.533, + "theoretical_loss": 3.603522072008416, + "tokens_seen": 1246232576 + }, + { + "epoch": 0.38, + "learning_rate": 0.0005029648946840522, + "loss": 1.4766, + "theoretical_loss": 3.6035049015772267, + "tokens_seen": 1246298112 + }, + { + "epoch": 0.38, + "learning_rate": 0.0005029488465396189, + "loss": 1.3618, + "theoretical_loss": 3.6034877323017094, + "tokens_seen": 1246363648 + }, + { + "epoch": 0.38, + "learning_rate": 0.0005029327983951856, + "loss": 1.324, + "theoretical_loss": 3.603470564181725, + "tokens_seen": 1246429184 + }, + { + "epoch": 0.38, + "learning_rate": 0.0005029167502507523, + "loss": 1.4399, + "theoretical_loss": 3.6034533972171356, + "tokens_seen": 1246494720 + }, + { + "epoch": 0.38, + "learning_rate": 0.000502900702106319, + "loss": 1.4597, + "theoretical_loss": 3.6034362314078026, + "tokens_seen": 1246560256 + }, + { + "epoch": 0.38, + "learning_rate": 0.0005028846539618857, + "loss": 1.4499, + "theoretical_loss": 3.6034190667535877, + "tokens_seen": 1246625792 + }, + { + "epoch": 0.38, + "learning_rate": 0.0005028686058174524, + "loss": 1.5599, + "theoretical_loss": 3.603401903254352, + "tokens_seen": 1246691328 + }, + { + "epoch": 0.38, + "learning_rate": 0.0005028525576730191, + "loss": 1.4486, + "theoretical_loss": 3.6033847409099575, + "tokens_seen": 1246756864 + }, + { + "epoch": 0.38, + "learning_rate": 0.0005028365095285858, + "loss": 1.4421, + "theoretical_loss": 3.603367579720265, + "tokens_seen": 1246822400 + }, + { + "epoch": 0.38, + "learning_rate": 0.0005028204613841525, + "loss": 1.365, + "theoretical_loss": 3.603350419685137, + "tokens_seen": 1246887936 + }, + { + "epoch": 0.38, + "learning_rate": 0.0005028044132397192, + "loss": 1.4601, + "theoretical_loss": 3.603333260804435, + "tokens_seen": 1246953472 + }, + { + "epoch": 0.38, + "learning_rate": 0.0005027883650952859, + "loss": 1.5759, + "theoretical_loss": 3.6033161030780203, + "tokens_seen": 1247019008 + }, + { + "epoch": 0.38, + "learning_rate": 0.0005027723169508526, + "loss": 1.4477, + "theoretical_loss": 3.603298946505755, + "tokens_seen": 1247084544 + }, + { + "epoch": 0.38, + "learning_rate": 0.0005027562688064193, + "loss": 1.4598, + "theoretical_loss": 3.6032817910875004, + "tokens_seen": 1247150080 + }, + { + "epoch": 0.38, + "learning_rate": 0.000502740220661986, + "loss": 1.3807, + "theoretical_loss": 3.603264636823119, + "tokens_seen": 1247215616 + }, + { + "epoch": 0.38, + "learning_rate": 0.0005027241725175527, + "loss": 1.3991, + "theoretical_loss": 3.6032474837124715, + "tokens_seen": 1247281152 + }, + { + "epoch": 0.38, + "learning_rate": 0.0005027081243731194, + "loss": 1.3859, + "theoretical_loss": 3.6032303317554204, + "tokens_seen": 1247346688 + }, + { + "epoch": 0.38, + "learning_rate": 0.0005026920762286861, + "loss": 1.3112, + "theoretical_loss": 3.603213180951827, + "tokens_seen": 1247412224 + }, + { + "epoch": 0.38, + "learning_rate": 0.0005026760280842528, + "loss": 1.4421, + "theoretical_loss": 3.603196031301554, + "tokens_seen": 1247477760 + }, + { + "epoch": 0.38, + "learning_rate": 0.0005026599799398195, + "loss": 1.3128, + "theoretical_loss": 3.603178882804462, + "tokens_seen": 1247543296 + }, + { + "epoch": 0.38, + "learning_rate": 0.0005026439317953862, + "loss": 1.4527, + "theoretical_loss": 3.6031617354604144, + "tokens_seen": 1247608832 + }, + { + "epoch": 0.38, + "learning_rate": 0.0005026278836509529, + "loss": 1.4145, + "theoretical_loss": 3.6031445892692724, + "tokens_seen": 1247674368 + }, + { + "epoch": 0.38, + "learning_rate": 0.0005026118355065196, + "loss": 1.3307, + "theoretical_loss": 3.603127444230897, + "tokens_seen": 1247739904 + }, + { + "epoch": 0.38, + "learning_rate": 0.0005025957873620863, + "loss": 1.4527, + "theoretical_loss": 3.603110300345152, + "tokens_seen": 1247805440 + }, + { + "epoch": 0.38, + "learning_rate": 0.000502579739217653, + "loss": 1.3664, + "theoretical_loss": 3.603093157611898, + "tokens_seen": 1247870976 + }, + { + "epoch": 0.38, + "learning_rate": 0.0005025636910732197, + "loss": 1.5829, + "theoretical_loss": 3.603076016030997, + "tokens_seen": 1247936512 + }, + { + "epoch": 0.38, + "learning_rate": 0.0005025476429287864, + "loss": 1.3715, + "theoretical_loss": 3.6030588756023123, + "tokens_seen": 1248002048 + }, + { + "epoch": 0.38, + "learning_rate": 0.0005025315947843531, + "loss": 1.463, + "theoretical_loss": 3.603041736325704, + "tokens_seen": 1248067584 + }, + { + "epoch": 0.38, + "learning_rate": 0.0005025155466399198, + "loss": 1.3661, + "theoretical_loss": 3.603024598201036, + "tokens_seen": 1248133120 + }, + { + "epoch": 0.38, + "learning_rate": 0.0005024994984954865, + "loss": 1.3036, + "theoretical_loss": 3.6030074612281693, + "tokens_seen": 1248198656 + }, + { + "epoch": 0.38, + "learning_rate": 0.0005024834503510532, + "loss": 1.4824, + "theoretical_loss": 3.6029903254069664, + "tokens_seen": 1248264192 + }, + { + "epoch": 0.38, + "learning_rate": 0.0005024674022066199, + "loss": 1.4263, + "theoretical_loss": 3.6029731907372895, + "tokens_seen": 1248329728 + }, + { + "epoch": 0.38, + "learning_rate": 0.0005024513540621866, + "loss": 1.5071, + "theoretical_loss": 3.602956057219001, + "tokens_seen": 1248395264 + }, + { + "epoch": 0.38, + "objective/train/docs_used": 456712, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.373943567276001, + "objective/train/theoretical_loss": 3.602938924851962, + "objective/train/tokens_used": 1268920800, + "theoretical_loss": 3.602938924851962, + "tokens_seen": 1248460800 + }, + { + "epoch": 0.38, + "learning_rate": 0.0005024353059177533, + "loss": 1.3932, + "theoretical_loss": 3.602938924851962, + "tokens_seen": 1248460800 + }, + { + "epoch": 0.38, + "learning_rate": 0.00050241925777332, + "loss": 1.4767, + "theoretical_loss": 3.6029217936360363, + "tokens_seen": 1248526336 + }, + { + "epoch": 0.38, + "learning_rate": 0.0005024032096288867, + "loss": 1.4948, + "theoretical_loss": 3.602904663571085, + "tokens_seen": 1248591872 + }, + { + "epoch": 0.38, + "learning_rate": 0.0005023871614844534, + "loss": 1.3953, + "theoretical_loss": 3.602887534656971, + "tokens_seen": 1248657408 + }, + { + "epoch": 0.38, + "learning_rate": 0.0005023711133400201, + "loss": 1.4565, + "theoretical_loss": 3.6028704068935555, + "tokens_seen": 1248722944 + }, + { + "epoch": 0.38, + "learning_rate": 0.0005023550651955868, + "loss": 1.3383, + "theoretical_loss": 3.6028532802807023, + "tokens_seen": 1248788480 + }, + { + "epoch": 0.38, + "learning_rate": 0.0005023390170511535, + "loss": 1.4049, + "theoretical_loss": 3.602836154818273, + "tokens_seen": 1248854016 + }, + { + "epoch": 0.38, + "learning_rate": 0.0005023229689067202, + "loss": 1.4523, + "theoretical_loss": 3.6028190305061303, + "tokens_seen": 1248919552 + }, + { + "epoch": 0.38, + "learning_rate": 0.0005023069207622869, + "loss": 1.3753, + "theoretical_loss": 3.6028019073441357, + "tokens_seen": 1248985088 + }, + { + "epoch": 0.38, + "learning_rate": 0.0005022908726178536, + "loss": 1.4978, + "theoretical_loss": 3.602784785332152, + "tokens_seen": 1249050624 + }, + { + "epoch": 0.38, + "learning_rate": 0.0005022748244734203, + "loss": 1.3232, + "theoretical_loss": 3.6027676644700426, + "tokens_seen": 1249116160 + }, + { + "epoch": 0.38, + "learning_rate": 0.000502258776328987, + "loss": 1.478, + "theoretical_loss": 3.6027505447576686, + "tokens_seen": 1249181696 + }, + { + "epoch": 0.38, + "learning_rate": 0.0005022427281845536, + "loss": 1.522, + "theoretical_loss": 3.6027334261948933, + "tokens_seen": 1249247232 + }, + { + "epoch": 0.38, + "learning_rate": 0.0005022266800401204, + "loss": 1.443, + "theoretical_loss": 3.6027163087815786, + "tokens_seen": 1249312768 + }, + { + "epoch": 0.38, + "learning_rate": 0.000502210631895687, + "loss": 1.363, + "theoretical_loss": 3.602699192517588, + "tokens_seen": 1249378304 + }, + { + "epoch": 0.38, + "learning_rate": 0.0005021945837512538, + "loss": 1.4911, + "theoretical_loss": 3.6026820774027826, + "tokens_seen": 1249443840 + }, + { + "epoch": 0.38, + "learning_rate": 0.0005021785356068204, + "loss": 1.4553, + "theoretical_loss": 3.6026649634370265, + "tokens_seen": 1249509376 + }, + { + "epoch": 0.38, + "learning_rate": 0.0005021624874623872, + "loss": 1.4128, + "theoretical_loss": 3.6026478506201816, + "tokens_seen": 1249574912 + }, + { + "epoch": 0.38, + "learning_rate": 0.0005021464393179538, + "loss": 1.4622, + "theoretical_loss": 3.6026307389521097, + "tokens_seen": 1249640448 + }, + { + "epoch": 0.38, + "learning_rate": 0.0005021303911735206, + "loss": 1.5161, + "theoretical_loss": 3.602613628432675, + "tokens_seen": 1249705984 + }, + { + "epoch": 0.38, + "learning_rate": 0.0005021143430290872, + "loss": 1.419, + "theoretical_loss": 3.6025965190617395, + "tokens_seen": 1249771520 + }, + { + "epoch": 0.38, + "learning_rate": 0.000502098294884654, + "loss": 1.4299, + "theoretical_loss": 3.6025794108391658, + "tokens_seen": 1249837056 + }, + { + "epoch": 0.38, + "learning_rate": 0.0005020822467402206, + "loss": 1.4064, + "theoretical_loss": 3.602562303764816, + "tokens_seen": 1249902592 + }, + { + "epoch": 0.38, + "learning_rate": 0.0005020661985957874, + "loss": 1.4258, + "theoretical_loss": 3.6025451978385536, + "tokens_seen": 1249968128 + }, + { + "epoch": 0.38, + "learning_rate": 0.000502050150451354, + "loss": 1.4471, + "theoretical_loss": 3.6025280930602417, + "tokens_seen": 1250033664 + }, + { + "epoch": 0.38, + "learning_rate": 0.0005020341023069208, + "loss": 1.4329, + "theoretical_loss": 3.6025109894297427, + "tokens_seen": 1250099200 + }, + { + "epoch": 0.38, + "learning_rate": 0.0005020180541624874, + "loss": 1.3175, + "theoretical_loss": 3.602493886946919, + "tokens_seen": 1250164736 + }, + { + "epoch": 0.38, + "learning_rate": 0.0005020020060180542, + "loss": 1.4827, + "theoretical_loss": 3.6024767856116338, + "tokens_seen": 1250230272 + }, + { + "epoch": 0.38, + "learning_rate": 0.0005019859578736208, + "loss": 1.4869, + "theoretical_loss": 3.60245968542375, + "tokens_seen": 1250295808 + }, + { + "epoch": 0.38, + "learning_rate": 0.0005019699097291876, + "loss": 1.3721, + "theoretical_loss": 3.6024425863831304, + "tokens_seen": 1250361344 + }, + { + "epoch": 0.38, + "learning_rate": 0.0005019538615847542, + "loss": 1.316, + "theoretical_loss": 3.6024254884896374, + "tokens_seen": 1250426880 + }, + { + "epoch": 0.38, + "learning_rate": 0.000501937813440321, + "loss": 1.4055, + "theoretical_loss": 3.6024083917431353, + "tokens_seen": 1250492416 + }, + { + "epoch": 0.38, + "learning_rate": 0.0005019217652958877, + "loss": 1.4195, + "theoretical_loss": 3.602391296143485, + "tokens_seen": 1250557952 + }, + { + "epoch": 0.38, + "learning_rate": 0.0005019057171514544, + "loss": 1.4404, + "theoretical_loss": 3.6023742016905516, + "tokens_seen": 1250623488 + }, + { + "epoch": 0.38, + "learning_rate": 0.000501889669007021, + "loss": 1.4976, + "theoretical_loss": 3.602357108384197, + "tokens_seen": 1250689024 + }, + { + "epoch": 0.38, + "learning_rate": 0.0005018736208625878, + "loss": 1.3596, + "theoretical_loss": 3.602340016224284, + "tokens_seen": 1250754560 + }, + { + "epoch": 0.38, + "learning_rate": 0.0005018575727181545, + "loss": 1.4056, + "theoretical_loss": 3.6023229252106765, + "tokens_seen": 1250820096 + }, + { + "epoch": 0.38, + "learning_rate": 0.0005018415245737212, + "loss": 1.4082, + "theoretical_loss": 3.6023058353432367, + "tokens_seen": 1250885632 + }, + { + "epoch": 0.38, + "learning_rate": 0.0005018254764292879, + "loss": 1.4287, + "theoretical_loss": 3.602288746621828, + "tokens_seen": 1250951168 + }, + { + "epoch": 0.38, + "learning_rate": 0.0005018094282848546, + "loss": 1.5153, + "theoretical_loss": 3.602271659046314, + "tokens_seen": 1251016704 + }, + { + "epoch": 0.38, + "learning_rate": 0.0005017933801404213, + "loss": 1.3089, + "theoretical_loss": 3.602254572616557, + "tokens_seen": 1251082240 + }, + { + "epoch": 0.38, + "learning_rate": 0.000501777331995988, + "loss": 1.3568, + "theoretical_loss": 3.602237487332421, + "tokens_seen": 1251147776 + }, + { + "epoch": 0.38, + "learning_rate": 0.0005017612838515547, + "loss": 1.4398, + "theoretical_loss": 3.6022204031937686, + "tokens_seen": 1251213312 + }, + { + "epoch": 0.38, + "learning_rate": 0.0005017452357071214, + "loss": 1.4879, + "theoretical_loss": 3.602203320200463, + "tokens_seen": 1251278848 + }, + { + "epoch": 0.38, + "learning_rate": 0.0005017291875626881, + "loss": 1.3531, + "theoretical_loss": 3.6021862383523677, + "tokens_seen": 1251344384 + }, + { + "epoch": 0.38, + "learning_rate": 0.0005017131394182548, + "loss": 1.3251, + "theoretical_loss": 3.602169157649346, + "tokens_seen": 1251409920 + }, + { + "epoch": 0.38, + "learning_rate": 0.0005016970912738215, + "loss": 1.4258, + "theoretical_loss": 3.602152078091261, + "tokens_seen": 1251475456 + }, + { + "epoch": 0.38, + "learning_rate": 0.0005016810431293882, + "loss": 1.3277, + "theoretical_loss": 3.6021349996779763, + "tokens_seen": 1251540992 + }, + { + "epoch": 0.38, + "learning_rate": 0.0005016649949849549, + "loss": 1.4205, + "theoretical_loss": 3.6021179224093545, + "tokens_seen": 1251606528 + }, + { + "epoch": 0.38, + "learning_rate": 0.0005016489468405216, + "loss": 1.3417, + "theoretical_loss": 3.6021008462852593, + "tokens_seen": 1251672064 + }, + { + "epoch": 0.38, + "objective/train/docs_used": 458046, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.4109761714935303, + "objective/train/theoretical_loss": 3.6020837713055545, + "objective/train/tokens_used": 1272197600, + "theoretical_loss": 3.6020837713055545, + "tokens_seen": 1251737600 + }, + { + "epoch": 0.38, + "learning_rate": 0.0005016328986960883, + "loss": 1.4322, + "theoretical_loss": 3.6020837713055545, + "tokens_seen": 1251737600 + }, + { + "epoch": 0.38, + "learning_rate": 0.000501616850551655, + "loss": 1.4277, + "theoretical_loss": 3.602066697470103, + "tokens_seen": 1251803136 + }, + { + "epoch": 0.38, + "learning_rate": 0.0005016008024072217, + "loss": 1.4926, + "theoretical_loss": 3.6020496247787688, + "tokens_seen": 1251868672 + }, + { + "epoch": 0.38, + "learning_rate": 0.0005015847542627884, + "loss": 1.3793, + "theoretical_loss": 3.602032553231415, + "tokens_seen": 1251934208 + }, + { + "epoch": 0.38, + "learning_rate": 0.0005015687061183551, + "loss": 1.4853, + "theoretical_loss": 3.6020154828279045, + "tokens_seen": 1251999744 + }, + { + "epoch": 0.38, + "learning_rate": 0.0005015526579739218, + "loss": 1.3413, + "theoretical_loss": 3.6019984135681016, + "tokens_seen": 1252065280 + }, + { + "epoch": 0.38, + "learning_rate": 0.0005015366098294885, + "loss": 1.4173, + "theoretical_loss": 3.601981345451869, + "tokens_seen": 1252130816 + }, + { + "epoch": 0.38, + "learning_rate": 0.0005015205616850552, + "loss": 1.4684, + "theoretical_loss": 3.6019642784790715, + "tokens_seen": 1252196352 + }, + { + "epoch": 0.38, + "learning_rate": 0.0005015045135406219, + "loss": 1.484, + "theoretical_loss": 3.6019472126495717, + "tokens_seen": 1252261888 + }, + { + "epoch": 0.38, + "learning_rate": 0.0005014884653961886, + "loss": 1.3189, + "theoretical_loss": 3.601930147963234, + "tokens_seen": 1252327424 + }, + { + "epoch": 0.38, + "learning_rate": 0.0005014724172517553, + "loss": 1.4264, + "theoretical_loss": 3.6019130844199205, + "tokens_seen": 1252392960 + }, + { + "epoch": 0.38, + "learning_rate": 0.000501456369107322, + "loss": 1.3833, + "theoretical_loss": 3.601896022019496, + "tokens_seen": 1252458496 + }, + { + "epoch": 0.38, + "learning_rate": 0.0005014403209628887, + "loss": 1.4274, + "theoretical_loss": 3.601878960761824, + "tokens_seen": 1252524032 + }, + { + "epoch": 0.38, + "learning_rate": 0.0005014242728184554, + "loss": 1.3941, + "theoretical_loss": 3.6018619006467683, + "tokens_seen": 1252589568 + }, + { + "epoch": 0.38, + "learning_rate": 0.0005014082246740221, + "loss": 1.4746, + "theoretical_loss": 3.601844841674192, + "tokens_seen": 1252655104 + }, + { + "epoch": 0.38, + "learning_rate": 0.0005013921765295888, + "loss": 1.3461, + "theoretical_loss": 3.6018277838439596, + "tokens_seen": 1252720640 + }, + { + "epoch": 0.38, + "learning_rate": 0.0005013761283851555, + "loss": 1.4639, + "theoretical_loss": 3.601810727155934, + "tokens_seen": 1252786176 + }, + { + "epoch": 0.38, + "learning_rate": 0.0005013600802407222, + "loss": 1.3284, + "theoretical_loss": 3.6017936716099794, + "tokens_seen": 1252851712 + }, + { + "epoch": 0.38, + "learning_rate": 0.0005013440320962889, + "loss": 1.4586, + "theoretical_loss": 3.60177661720596, + "tokens_seen": 1252917248 + }, + { + "epoch": 0.38, + "learning_rate": 0.0005013279839518556, + "loss": 1.4529, + "theoretical_loss": 3.601759563943739, + "tokens_seen": 1252982784 + }, + { + "epoch": 0.38, + "learning_rate": 0.0005013119358074223, + "loss": 1.3781, + "theoretical_loss": 3.601742511823181, + "tokens_seen": 1253048320 + }, + { + "epoch": 0.38, + "learning_rate": 0.000501295887662989, + "loss": 1.4678, + "theoretical_loss": 3.6017254608441487, + "tokens_seen": 1253113856 + }, + { + "epoch": 0.38, + "learning_rate": 0.0005012798395185557, + "loss": 1.5021, + "theoretical_loss": 3.6017084110065065, + "tokens_seen": 1253179392 + }, + { + "epoch": 0.38, + "learning_rate": 0.0005012637913741224, + "loss": 1.5204, + "theoretical_loss": 3.601691362310119, + "tokens_seen": 1253244928 + }, + { + "epoch": 0.38, + "learning_rate": 0.0005012477432296891, + "loss": 1.3844, + "theoretical_loss": 3.6016743147548493, + "tokens_seen": 1253310464 + }, + { + "epoch": 0.38, + "learning_rate": 0.0005012316950852558, + "loss": 1.3611, + "theoretical_loss": 3.6016572683405617, + "tokens_seen": 1253376000 + }, + { + "epoch": 0.38, + "learning_rate": 0.0005012156469408225, + "loss": 1.4613, + "theoretical_loss": 3.60164022306712, + "tokens_seen": 1253441536 + }, + { + "epoch": 0.38, + "learning_rate": 0.0005011995987963892, + "loss": 1.4216, + "theoretical_loss": 3.6016231789343887, + "tokens_seen": 1253507072 + }, + { + "epoch": 0.38, + "learning_rate": 0.0005011835506519559, + "loss": 1.4214, + "theoretical_loss": 3.601606135942231, + "tokens_seen": 1253572608 + }, + { + "epoch": 0.38, + "learning_rate": 0.0005011675025075226, + "loss": 1.4297, + "theoretical_loss": 3.601589094090512, + "tokens_seen": 1253638144 + }, + { + "epoch": 0.38, + "learning_rate": 0.0005011514543630893, + "loss": 1.4341, + "theoretical_loss": 3.6015720533790945, + "tokens_seen": 1253703680 + }, + { + "epoch": 0.38, + "learning_rate": 0.000501135406218656, + "loss": 1.4001, + "theoretical_loss": 3.601555013807844, + "tokens_seen": 1253769216 + }, + { + "epoch": 0.38, + "learning_rate": 0.0005011193580742227, + "loss": 1.4257, + "theoretical_loss": 3.6015379753766235, + "tokens_seen": 1253834752 + }, + { + "epoch": 0.38, + "learning_rate": 0.0005011033099297894, + "loss": 1.4307, + "theoretical_loss": 3.6015209380852973, + "tokens_seen": 1253900288 + }, + { + "epoch": 0.38, + "learning_rate": 0.0005010872617853561, + "loss": 1.4514, + "theoretical_loss": 3.60150390193373, + "tokens_seen": 1253965824 + }, + { + "epoch": 0.38, + "learning_rate": 0.0005010712136409228, + "loss": 1.4982, + "theoretical_loss": 3.6014868669217854, + "tokens_seen": 1254031360 + }, + { + "epoch": 0.38, + "learning_rate": 0.0005010551654964895, + "loss": 1.4133, + "theoretical_loss": 3.6014698330493284, + "tokens_seen": 1254096896 + }, + { + "epoch": 0.38, + "learning_rate": 0.0005010391173520562, + "loss": 1.4223, + "theoretical_loss": 3.6014528003162223, + "tokens_seen": 1254162432 + }, + { + "epoch": 0.38, + "learning_rate": 0.0005010230692076229, + "loss": 1.373, + "theoretical_loss": 3.6014357687223324, + "tokens_seen": 1254227968 + }, + { + "epoch": 0.38, + "learning_rate": 0.0005010070210631896, + "loss": 1.3923, + "theoretical_loss": 3.6014187382675216, + "tokens_seen": 1254293504 + }, + { + "epoch": 0.38, + "learning_rate": 0.0005009909729187563, + "loss": 1.475, + "theoretical_loss": 3.6014017089516557, + "tokens_seen": 1254359040 + }, + { + "epoch": 0.38, + "learning_rate": 0.000500974924774323, + "loss": 1.4264, + "theoretical_loss": 3.601384680774598, + "tokens_seen": 1254424576 + }, + { + "epoch": 0.38, + "learning_rate": 0.0005009588766298897, + "loss": 1.3929, + "theoretical_loss": 3.6013676537362134, + "tokens_seen": 1254490112 + }, + { + "epoch": 0.38, + "learning_rate": 0.0005009428284854564, + "loss": 1.4184, + "theoretical_loss": 3.601350627836366, + "tokens_seen": 1254555648 + }, + { + "epoch": 0.38, + "learning_rate": 0.0005009267803410231, + "loss": 1.3962, + "theoretical_loss": 3.60133360307492, + "tokens_seen": 1254621184 + }, + { + "epoch": 0.38, + "learning_rate": 0.0005009107321965898, + "loss": 1.4714, + "theoretical_loss": 3.6013165794517406, + "tokens_seen": 1254686720 + }, + { + "epoch": 0.38, + "learning_rate": 0.0005008946840521565, + "loss": 1.3284, + "theoretical_loss": 3.6012995569666915, + "tokens_seen": 1254752256 + }, + { + "epoch": 0.38, + "learning_rate": 0.0005008786359077232, + "loss": 1.5077, + "theoretical_loss": 3.601282535619638, + "tokens_seen": 1254817792 + }, + { + "epoch": 0.38, + "learning_rate": 0.0005008625877632899, + "loss": 1.3706, + "theoretical_loss": 3.601265515410443, + "tokens_seen": 1254883328 + }, + { + "epoch": 0.38, + "learning_rate": 0.0005008465396188565, + "loss": 1.3701, + "theoretical_loss": 3.6012484963389726, + "tokens_seen": 1254948864 + }, + { + "epoch": 0.38, + "objective/train/docs_used": 459153, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.4797133207321167, + "objective/train/theoretical_loss": 3.60123147840509, + "objective/train/tokens_used": 1275474400, + "theoretical_loss": 3.60123147840509, + "tokens_seen": 1255014400 + }, + { + "epoch": 0.38, + "learning_rate": 0.0005008304914744233, + "loss": 1.4852, + "theoretical_loss": 3.60123147840509, + "tokens_seen": 1255014400 + }, + { + "epoch": 0.38, + "learning_rate": 0.0005008144433299899, + "loss": 1.3658, + "theoretical_loss": 3.601214461608661, + "tokens_seen": 1255079936 + }, + { + "epoch": 0.38, + "learning_rate": 0.0005007983951855567, + "loss": 1.4495, + "theoretical_loss": 3.60119744594955, + "tokens_seen": 1255145472 + }, + { + "epoch": 0.38, + "learning_rate": 0.0005007823470411233, + "loss": 1.3651, + "theoretical_loss": 3.601180431427621, + "tokens_seen": 1255211008 + }, + { + "epoch": 0.38, + "learning_rate": 0.0005007662988966901, + "loss": 1.4155, + "theoretical_loss": 3.6011634180427397, + "tokens_seen": 1255276544 + }, + { + "epoch": 0.38, + "learning_rate": 0.0005007502507522567, + "loss": 1.3919, + "theoretical_loss": 3.6011464057947693, + "tokens_seen": 1255342080 + }, + { + "epoch": 0.38, + "learning_rate": 0.0005007342026078235, + "loss": 1.4212, + "theoretical_loss": 3.601129394683575, + "tokens_seen": 1255407616 + }, + { + "epoch": 0.38, + "learning_rate": 0.0005007181544633901, + "loss": 1.4192, + "theoretical_loss": 3.6011123847090225, + "tokens_seen": 1255473152 + }, + { + "epoch": 0.38, + "learning_rate": 0.0005007021063189569, + "loss": 1.4328, + "theoretical_loss": 3.601095375870975, + "tokens_seen": 1255538688 + }, + { + "epoch": 0.38, + "learning_rate": 0.0005006860581745235, + "loss": 1.4101, + "theoretical_loss": 3.601078368169298, + "tokens_seen": 1255604224 + }, + { + "epoch": 0.38, + "learning_rate": 0.0005006700100300903, + "loss": 1.3934, + "theoretical_loss": 3.6010613616038567, + "tokens_seen": 1255669760 + }, + { + "epoch": 0.38, + "learning_rate": 0.000500653961885657, + "loss": 1.4141, + "theoretical_loss": 3.6010443561745156, + "tokens_seen": 1255735296 + }, + { + "epoch": 0.38, + "learning_rate": 0.0005006379137412237, + "loss": 1.3677, + "theoretical_loss": 3.6010273518811386, + "tokens_seen": 1255800832 + }, + { + "epoch": 0.38, + "learning_rate": 0.0005006218655967903, + "loss": 1.4684, + "theoretical_loss": 3.601010348723592, + "tokens_seen": 1255866368 + }, + { + "epoch": 0.38, + "learning_rate": 0.0005006058174523571, + "loss": 1.3811, + "theoretical_loss": 3.600993346701739, + "tokens_seen": 1255931904 + }, + { + "epoch": 0.38, + "learning_rate": 0.0005005897693079237, + "loss": 1.3433, + "theoretical_loss": 3.6009763458154462, + "tokens_seen": 1255997440 + }, + { + "epoch": 0.38, + "learning_rate": 0.0005005737211634905, + "loss": 1.4103, + "theoretical_loss": 3.6009593460645775, + "tokens_seen": 1256062976 + }, + { + "epoch": 0.38, + "learning_rate": 0.0005005576730190571, + "loss": 1.344, + "theoretical_loss": 3.6009423474489983, + "tokens_seen": 1256128512 + }, + { + "epoch": 0.38, + "learning_rate": 0.0005005416248746239, + "loss": 1.4124, + "theoretical_loss": 3.6009253499685734, + "tokens_seen": 1256194048 + }, + { + "epoch": 0.38, + "learning_rate": 0.0005005255767301906, + "loss": 1.4153, + "theoretical_loss": 3.6009083536231676, + "tokens_seen": 1256259584 + }, + { + "epoch": 0.38, + "learning_rate": 0.0005005095285857573, + "loss": 1.3554, + "theoretical_loss": 3.6008913584126456, + "tokens_seen": 1256325120 + }, + { + "epoch": 0.38, + "learning_rate": 0.000500493480441324, + "loss": 1.3908, + "theoretical_loss": 3.6008743643368737, + "tokens_seen": 1256390656 + }, + { + "epoch": 0.38, + "learning_rate": 0.0005004774322968907, + "loss": 1.4125, + "theoretical_loss": 3.6008573713957155, + "tokens_seen": 1256456192 + }, + { + "epoch": 0.38, + "learning_rate": 0.0005004613841524574, + "loss": 1.4048, + "theoretical_loss": 3.6008403795890365, + "tokens_seen": 1256521728 + }, + { + "epoch": 0.38, + "learning_rate": 0.0005004453360080241, + "loss": 1.5325, + "theoretical_loss": 3.6008233889167025, + "tokens_seen": 1256587264 + }, + { + "epoch": 0.38, + "learning_rate": 0.0005004292878635908, + "loss": 1.4272, + "theoretical_loss": 3.6008063993785777, + "tokens_seen": 1256652800 + }, + { + "epoch": 0.38, + "learning_rate": 0.0005004132397191575, + "loss": 1.3461, + "theoretical_loss": 3.600789410974528, + "tokens_seen": 1256718336 + }, + { + "epoch": 0.38, + "learning_rate": 0.0005003971915747242, + "loss": 1.4017, + "theoretical_loss": 3.600772423704418, + "tokens_seen": 1256783872 + }, + { + "epoch": 0.38, + "learning_rate": 0.0005003811434302909, + "loss": 1.3942, + "theoretical_loss": 3.6007554375681132, + "tokens_seen": 1256849408 + }, + { + "epoch": 0.38, + "learning_rate": 0.0005003650952858576, + "loss": 1.4527, + "theoretical_loss": 3.600738452565478, + "tokens_seen": 1256914944 + }, + { + "epoch": 0.38, + "learning_rate": 0.0005003490471414243, + "loss": 1.3313, + "theoretical_loss": 3.6007214686963795, + "tokens_seen": 1256980480 + }, + { + "epoch": 0.38, + "learning_rate": 0.000500332998996991, + "loss": 1.4087, + "theoretical_loss": 3.600704485960681, + "tokens_seen": 1257046016 + }, + { + "epoch": 0.38, + "learning_rate": 0.0005003169508525577, + "loss": 1.4597, + "theoretical_loss": 3.600687504358249, + "tokens_seen": 1257111552 + }, + { + "epoch": 0.38, + "learning_rate": 0.0005003009027081244, + "loss": 1.4959, + "theoretical_loss": 3.6006705238889483, + "tokens_seen": 1257177088 + }, + { + "epoch": 0.38, + "learning_rate": 0.0005002848545636911, + "loss": 1.3747, + "theoretical_loss": 3.600653544552644, + "tokens_seen": 1257242624 + }, + { + "epoch": 0.38, + "learning_rate": 0.0005002688064192578, + "loss": 1.4046, + "theoretical_loss": 3.600636566349202, + "tokens_seen": 1257308160 + }, + { + "epoch": 0.38, + "learning_rate": 0.0005002527582748245, + "loss": 1.3664, + "theoretical_loss": 3.6006195892784874, + "tokens_seen": 1257373696 + }, + { + "epoch": 0.38, + "learning_rate": 0.0005002367101303912, + "loss": 1.4604, + "theoretical_loss": 3.6006026133403655, + "tokens_seen": 1257439232 + }, + { + "epoch": 0.38, + "learning_rate": 0.0005002206619859579, + "loss": 1.3644, + "theoretical_loss": 3.6005856385347013, + "tokens_seen": 1257504768 + }, + { + "epoch": 0.38, + "learning_rate": 0.0005002046138415246, + "loss": 1.4544, + "theoretical_loss": 3.6005686648613615, + "tokens_seen": 1257570304 + }, + { + "epoch": 0.38, + "learning_rate": 0.0005001885656970913, + "loss": 1.3945, + "theoretical_loss": 3.6005516923202108, + "tokens_seen": 1257635840 + }, + { + "epoch": 0.38, + "learning_rate": 0.000500172517552658, + "loss": 1.321, + "theoretical_loss": 3.6005347209111145, + "tokens_seen": 1257701376 + }, + { + "epoch": 0.38, + "learning_rate": 0.0005001564694082247, + "loss": 1.4114, + "theoretical_loss": 3.600517750633938, + "tokens_seen": 1257766912 + }, + { + "epoch": 0.38, + "learning_rate": 0.0005001404212637914, + "loss": 1.3554, + "theoretical_loss": 3.600500781488548, + "tokens_seen": 1257832448 + }, + { + "epoch": 0.38, + "learning_rate": 0.0005001243731193581, + "loss": 1.3846, + "theoretical_loss": 3.600483813474809, + "tokens_seen": 1257897984 + }, + { + "epoch": 0.38, + "learning_rate": 0.0005001083249749248, + "loss": 1.4603, + "theoretical_loss": 3.600466846592586, + "tokens_seen": 1257963520 + }, + { + "epoch": 0.38, + "learning_rate": 0.0005000922768304915, + "loss": 1.4068, + "theoretical_loss": 3.6004498808417464, + "tokens_seen": 1258029056 + }, + { + "epoch": 0.38, + "learning_rate": 0.0005000762286860582, + "loss": 1.3227, + "theoretical_loss": 3.6004329162221542, + "tokens_seen": 1258094592 + }, + { + "epoch": 0.38, + "learning_rate": 0.0005000601805416249, + "loss": 1.3691, + "theoretical_loss": 3.600415952733676, + "tokens_seen": 1258160128 + }, + { + "epoch": 0.38, + "learning_rate": 0.0005000441323971916, + "loss": 1.3776, + "theoretical_loss": 3.600398990376177, + "tokens_seen": 1258225664 + }, + { + "epoch": 0.38, + "objective/train/docs_used": 460354, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.5418055057525635, + "objective/train/theoretical_loss": 3.600382029149523, + "objective/train/tokens_used": 1278751200, + "theoretical_loss": 3.600382029149523, + "tokens_seen": 1258291200 + }, + { + "epoch": 0.38, + "learning_rate": 0.0005000280842527583, + "loss": 1.4532, + "theoretical_loss": 3.600382029149523, + "tokens_seen": 1258291200 + }, + { + "epoch": 0.38, + "learning_rate": 0.000500012036108325, + "loss": 1.3445, + "theoretical_loss": 3.60036506905358, + "tokens_seen": 1258356736 + }, + { + "epoch": 0.38, + "learning_rate": 0.0004999959879638917, + "loss": 1.3165, + "theoretical_loss": 3.600348110088213, + "tokens_seen": 1258422272 + }, + { + "epoch": 0.38, + "learning_rate": 0.0004999799398194584, + "loss": 1.3655, + "theoretical_loss": 3.600331152253289, + "tokens_seen": 1258487808 + }, + { + "epoch": 0.38, + "learning_rate": 0.0004999638916750251, + "loss": 1.3519, + "theoretical_loss": 3.6003141955486724, + "tokens_seen": 1258553344 + }, + { + "epoch": 0.38, + "learning_rate": 0.0004999478435305918, + "loss": 1.4614, + "theoretical_loss": 3.6002972399742297, + "tokens_seen": 1258618880 + }, + { + "epoch": 0.38, + "learning_rate": 0.0004999317953861585, + "loss": 1.4644, + "theoretical_loss": 3.600280285529827, + "tokens_seen": 1258684416 + }, + { + "epoch": 0.38, + "learning_rate": 0.0004999157472417252, + "loss": 1.4691, + "theoretical_loss": 3.6002633322153295, + "tokens_seen": 1258749952 + }, + { + "epoch": 0.38, + "learning_rate": 0.0004998996990972919, + "loss": 1.4423, + "theoretical_loss": 3.6002463800306037, + "tokens_seen": 1258815488 + }, + { + "epoch": 0.38, + "learning_rate": 0.0004998836509528586, + "loss": 1.4463, + "theoretical_loss": 3.6002294289755152, + "tokens_seen": 1258881024 + }, + { + "epoch": 0.38, + "learning_rate": 0.0004998676028084253, + "loss": 1.3125, + "theoretical_loss": 3.6002124790499295, + "tokens_seen": 1258946560 + }, + { + "epoch": 0.38, + "learning_rate": 0.000499851554663992, + "loss": 1.4634, + "theoretical_loss": 3.600195530253713, + "tokens_seen": 1259012096 + }, + { + "epoch": 0.38, + "learning_rate": 0.0004998355065195587, + "loss": 1.55, + "theoretical_loss": 3.6001785825867323, + "tokens_seen": 1259077632 + }, + { + "epoch": 0.38, + "learning_rate": 0.0004998194583751254, + "loss": 1.4336, + "theoretical_loss": 3.6001616360488526, + "tokens_seen": 1259143168 + }, + { + "epoch": 0.38, + "learning_rate": 0.0004998034102306921, + "loss": 1.3658, + "theoretical_loss": 3.6001446906399397, + "tokens_seen": 1259208704 + }, + { + "epoch": 0.38, + "learning_rate": 0.0004997873620862588, + "loss": 1.4685, + "theoretical_loss": 3.60012774635986, + "tokens_seen": 1259274240 + }, + { + "epoch": 0.38, + "learning_rate": 0.0004997713139418255, + "loss": 1.4553, + "theoretical_loss": 3.60011080320848, + "tokens_seen": 1259339776 + }, + { + "epoch": 0.38, + "learning_rate": 0.0004997552657973922, + "loss": 1.4519, + "theoretical_loss": 3.600093861185665, + "tokens_seen": 1259405312 + }, + { + "epoch": 0.38, + "learning_rate": 0.0004997392176529589, + "loss": 1.4784, + "theoretical_loss": 3.600076920291281, + "tokens_seen": 1259470848 + }, + { + "epoch": 0.38, + "learning_rate": 0.0004997231695085256, + "loss": 1.4514, + "theoretical_loss": 3.600059980525195, + "tokens_seen": 1259536384 + }, + { + "epoch": 0.38, + "learning_rate": 0.0004997071213640923, + "loss": 1.4494, + "theoretical_loss": 3.600043041887273, + "tokens_seen": 1259601920 + }, + { + "epoch": 0.38, + "learning_rate": 0.0004996910732196591, + "loss": 1.3927, + "theoretical_loss": 3.600026104377381, + "tokens_seen": 1259667456 + }, + { + "epoch": 0.38, + "learning_rate": 0.0004996750250752257, + "loss": 1.3876, + "theoretical_loss": 3.6000091679953847, + "tokens_seen": 1259732992 + }, + { + "epoch": 0.38, + "learning_rate": 0.0004996589769307925, + "loss": 1.4396, + "theoretical_loss": 3.5999922327411507, + "tokens_seen": 1259798528 + }, + { + "epoch": 0.38, + "learning_rate": 0.0004996429287863591, + "loss": 1.3441, + "theoretical_loss": 3.5999752986145452, + "tokens_seen": 1259864064 + }, + { + "epoch": 0.38, + "learning_rate": 0.0004996268806419259, + "loss": 1.5159, + "theoretical_loss": 3.599958365615435, + "tokens_seen": 1259929600 + }, + { + "epoch": 0.38, + "learning_rate": 0.0004996108324974925, + "loss": 1.4603, + "theoretical_loss": 3.5999414337436852, + "tokens_seen": 1259995136 + }, + { + "epoch": 0.38, + "learning_rate": 0.0004995947843530592, + "loss": 1.4601, + "theoretical_loss": 3.5999245029991633, + "tokens_seen": 1260060672 + }, + { + "epoch": 0.38, + "learning_rate": 0.0004995787362086259, + "loss": 1.3399, + "theoretical_loss": 3.599907573381735, + "tokens_seen": 1260126208 + }, + { + "epoch": 0.38, + "learning_rate": 0.0004995626880641926, + "loss": 1.4417, + "theoretical_loss": 3.599890644891267, + "tokens_seen": 1260191744 + }, + { + "epoch": 0.38, + "learning_rate": 0.0004995466399197593, + "loss": 1.4701, + "theoretical_loss": 3.5998737175276254, + "tokens_seen": 1260257280 + }, + { + "epoch": 0.38, + "learning_rate": 0.000499530591775326, + "loss": 1.3474, + "theoretical_loss": 3.5998567912906765, + "tokens_seen": 1260322816 + }, + { + "epoch": 0.38, + "learning_rate": 0.0004995145436308927, + "loss": 1.4799, + "theoretical_loss": 3.599839866180287, + "tokens_seen": 1260388352 + }, + { + "epoch": 0.38, + "learning_rate": 0.0004994984954864594, + "loss": 1.3961, + "theoretical_loss": 3.5998229421963233, + "tokens_seen": 1260453888 + }, + { + "epoch": 0.38, + "learning_rate": 0.0004994824473420261, + "loss": 1.5093, + "theoretical_loss": 3.5998060193386516, + "tokens_seen": 1260519424 + }, + { + "epoch": 0.38, + "learning_rate": 0.0004994663991975928, + "loss": 1.4274, + "theoretical_loss": 3.599789097607139, + "tokens_seen": 1260584960 + }, + { + "epoch": 0.38, + "learning_rate": 0.0004994503510531594, + "loss": 1.3052, + "theoretical_loss": 3.5997721770016518, + "tokens_seen": 1260650496 + }, + { + "epoch": 0.38, + "learning_rate": 0.0004994343029087262, + "loss": 1.3842, + "theoretical_loss": 3.599755257522056, + "tokens_seen": 1260716032 + }, + { + "epoch": 0.38, + "learning_rate": 0.0004994182547642928, + "loss": 1.4039, + "theoretical_loss": 3.5997383391682187, + "tokens_seen": 1260781568 + }, + { + "epoch": 0.38, + "learning_rate": 0.0004994022066198596, + "loss": 1.5167, + "theoretical_loss": 3.599721421940006, + "tokens_seen": 1260847104 + }, + { + "epoch": 0.38, + "learning_rate": 0.0004993861584754262, + "loss": 1.4869, + "theoretical_loss": 3.599704505837286, + "tokens_seen": 1260912640 + }, + { + "epoch": 0.38, + "learning_rate": 0.000499370110330993, + "loss": 1.4982, + "theoretical_loss": 3.599687590859923, + "tokens_seen": 1260978176 + }, + { + "epoch": 0.38, + "learning_rate": 0.0004993540621865596, + "loss": 1.326, + "theoretical_loss": 3.599670677007785, + "tokens_seen": 1261043712 + }, + { + "epoch": 0.38, + "learning_rate": 0.0004993380140421264, + "loss": 1.4634, + "theoretical_loss": 3.5996537642807387, + "tokens_seen": 1261109248 + }, + { + "epoch": 0.38, + "learning_rate": 0.000499321965897693, + "loss": 1.398, + "theoretical_loss": 3.5996368526786506, + "tokens_seen": 1261174784 + }, + { + "epoch": 0.38, + "learning_rate": 0.0004993059177532598, + "loss": 1.4662, + "theoretical_loss": 3.5996199422013877, + "tokens_seen": 1261240320 + }, + { + "epoch": 0.38, + "learning_rate": 0.0004992898696088264, + "loss": 1.4073, + "theoretical_loss": 3.599603032848816, + "tokens_seen": 1261305856 + }, + { + "epoch": 0.38, + "learning_rate": 0.0004992738214643932, + "loss": 1.3151, + "theoretical_loss": 3.599586124620803, + "tokens_seen": 1261371392 + }, + { + "epoch": 0.38, + "learning_rate": 0.0004992577733199598, + "loss": 1.3646, + "theoretical_loss": 3.5995692175172147, + "tokens_seen": 1261436928 + }, + { + "epoch": 0.38, + "learning_rate": 0.0004992417251755266, + "loss": 1.3482, + "theoretical_loss": 3.599552311537919, + "tokens_seen": 1261502464 + }, + { + "epoch": 0.38, + "objective/train/docs_used": 461530, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.362824559211731, + "objective/train/theoretical_loss": 3.599535406682782, + "objective/train/tokens_used": 1282028000, + "theoretical_loss": 3.599535406682782, + "tokens_seen": 1261568000 + }, + { + "epoch": 0.38, + "learning_rate": 0.0004992256770310932, + "loss": 1.3355, + "theoretical_loss": 3.599535406682782, + "tokens_seen": 1261568000 + }, + { + "epoch": 0.38, + "learning_rate": 0.00049920962888666, + "loss": 1.4254, + "theoretical_loss": 3.5995185029516708, + "tokens_seen": 1261633536 + }, + { + "epoch": 0.38, + "learning_rate": 0.0004991935807422266, + "loss": 1.4828, + "theoretical_loss": 3.5995016003444524, + "tokens_seen": 1261699072 + }, + { + "epoch": 0.38, + "learning_rate": 0.0004991775325977934, + "loss": 1.4002, + "theoretical_loss": 3.5994846988609934, + "tokens_seen": 1261764608 + }, + { + "epoch": 0.38, + "learning_rate": 0.00049916148445336, + "loss": 1.3853, + "theoretical_loss": 3.59946779850116, + "tokens_seen": 1261830144 + }, + { + "epoch": 0.38, + "learning_rate": 0.0004991454363089268, + "loss": 1.4948, + "theoretical_loss": 3.599450899264821, + "tokens_seen": 1261895680 + }, + { + "epoch": 0.38, + "learning_rate": 0.0004991293881644935, + "loss": 1.2833, + "theoretical_loss": 3.599434001151842, + "tokens_seen": 1261961216 + }, + { + "epoch": 0.38, + "learning_rate": 0.0004991133400200602, + "loss": 1.358, + "theoretical_loss": 3.5994171041620904, + "tokens_seen": 1262026752 + }, + { + "epoch": 0.38, + "learning_rate": 0.0004990972918756269, + "loss": 1.4743, + "theoretical_loss": 3.5994002082954335, + "tokens_seen": 1262092288 + }, + { + "epoch": 0.38, + "learning_rate": 0.0004990812437311936, + "loss": 1.3689, + "theoretical_loss": 3.599383313551737, + "tokens_seen": 1262157824 + }, + { + "epoch": 0.38, + "learning_rate": 0.0004990651955867603, + "loss": 1.3619, + "theoretical_loss": 3.59936641993087, + "tokens_seen": 1262223360 + }, + { + "epoch": 0.38, + "learning_rate": 0.000499049147442327, + "loss": 1.4825, + "theoretical_loss": 3.599349527432698, + "tokens_seen": 1262288896 + }, + { + "epoch": 0.38, + "learning_rate": 0.0004990330992978937, + "loss": 1.3158, + "theoretical_loss": 3.5993326360570888, + "tokens_seen": 1262354432 + }, + { + "epoch": 0.38, + "learning_rate": 0.0004990170511534604, + "loss": 1.4016, + "theoretical_loss": 3.599315745803909, + "tokens_seen": 1262419968 + }, + { + "epoch": 0.38, + "learning_rate": 0.0004990010030090271, + "loss": 1.4519, + "theoretical_loss": 3.599298856673027, + "tokens_seen": 1262485504 + }, + { + "epoch": 0.38, + "learning_rate": 0.0004989849548645938, + "loss": 1.2999, + "theoretical_loss": 3.5992819686643083, + "tokens_seen": 1262551040 + }, + { + "epoch": 0.38, + "learning_rate": 0.0004989689067201605, + "loss": 1.4342, + "theoretical_loss": 3.5992650817776215, + "tokens_seen": 1262616576 + }, + { + "epoch": 0.38, + "learning_rate": 0.0004989528585757272, + "loss": 1.3844, + "theoretical_loss": 3.599248196012833, + "tokens_seen": 1262682112 + }, + { + "epoch": 0.38, + "learning_rate": 0.0004989368104312939, + "loss": 1.4155, + "theoretical_loss": 3.59923131136981, + "tokens_seen": 1262747648 + }, + { + "epoch": 0.38, + "learning_rate": 0.0004989207622868606, + "loss": 1.4102, + "theoretical_loss": 3.5992144278484206, + "tokens_seen": 1262813184 + }, + { + "epoch": 0.38, + "learning_rate": 0.0004989047141424273, + "loss": 1.4183, + "theoretical_loss": 3.599197545448531, + "tokens_seen": 1262878720 + }, + { + "epoch": 0.38, + "learning_rate": 0.000498888665997994, + "loss": 1.431, + "theoretical_loss": 3.5991806641700093, + "tokens_seen": 1262944256 + }, + { + "epoch": 0.38, + "learning_rate": 0.0004988726178535607, + "loss": 1.4014, + "theoretical_loss": 3.599163784012722, + "tokens_seen": 1263009792 + }, + { + "epoch": 0.38, + "learning_rate": 0.0004988565697091274, + "loss": 1.3631, + "theoretical_loss": 3.5991469049765374, + "tokens_seen": 1263075328 + }, + { + "epoch": 0.38, + "learning_rate": 0.0004988405215646941, + "loss": 1.3671, + "theoretical_loss": 3.5991300270613227, + "tokens_seen": 1263140864 + }, + { + "epoch": 0.38, + "learning_rate": 0.0004988244734202608, + "loss": 1.4145, + "theoretical_loss": 3.5991131502669447, + "tokens_seen": 1263206400 + }, + { + "epoch": 0.38, + "learning_rate": 0.0004988084252758275, + "loss": 1.3589, + "theoretical_loss": 3.599096274593272, + "tokens_seen": 1263271936 + }, + { + "epoch": 0.38, + "learning_rate": 0.0004987923771313942, + "loss": 1.3825, + "theoretical_loss": 3.5990794000401705, + "tokens_seen": 1263337472 + }, + { + "epoch": 0.38, + "learning_rate": 0.0004987763289869609, + "loss": 1.4863, + "theoretical_loss": 3.5990625266075087, + "tokens_seen": 1263403008 + }, + { + "epoch": 0.38, + "learning_rate": 0.0004987602808425276, + "loss": 1.4228, + "theoretical_loss": 3.5990456542951534, + "tokens_seen": 1263468544 + }, + { + "epoch": 0.38, + "learning_rate": 0.0004987442326980943, + "loss": 1.4769, + "theoretical_loss": 3.5990287831029733, + "tokens_seen": 1263534080 + }, + { + "epoch": 0.38, + "learning_rate": 0.000498728184553661, + "loss": 1.4834, + "theoretical_loss": 3.5990119130308345, + "tokens_seen": 1263599616 + }, + { + "epoch": 0.38, + "learning_rate": 0.0004987121364092277, + "loss": 1.4721, + "theoretical_loss": 3.5989950440786056, + "tokens_seen": 1263665152 + }, + { + "epoch": 0.38, + "learning_rate": 0.0004986960882647944, + "loss": 1.4429, + "theoretical_loss": 3.5989781762461535, + "tokens_seen": 1263730688 + }, + { + "epoch": 0.38, + "learning_rate": 0.0004986800401203611, + "loss": 1.4219, + "theoretical_loss": 3.5989613095333466, + "tokens_seen": 1263796224 + }, + { + "epoch": 0.38, + "learning_rate": 0.0004986639919759278, + "loss": 1.4338, + "theoretical_loss": 3.5989444439400518, + "tokens_seen": 1263861760 + }, + { + "epoch": 0.38, + "learning_rate": 0.0004986479438314945, + "loss": 1.4245, + "theoretical_loss": 3.598927579466137, + "tokens_seen": 1263927296 + }, + { + "epoch": 0.38, + "learning_rate": 0.0004986318956870612, + "loss": 1.4838, + "theoretical_loss": 3.5989107161114697, + "tokens_seen": 1263992832 + }, + { + "epoch": 0.38, + "learning_rate": 0.0004986158475426279, + "loss": 1.4413, + "theoretical_loss": 3.598893853875918, + "tokens_seen": 1264058368 + }, + { + "epoch": 0.38, + "learning_rate": 0.0004985997993981946, + "loss": 1.2625, + "theoretical_loss": 3.598876992759349, + "tokens_seen": 1264123904 + }, + { + "epoch": 0.38, + "learning_rate": 0.0004985837512537613, + "loss": 1.4476, + "theoretical_loss": 3.598860132761631, + "tokens_seen": 1264189440 + }, + { + "epoch": 0.38, + "learning_rate": 0.000498567703109328, + "loss": 1.4299, + "theoretical_loss": 3.5988432738826317, + "tokens_seen": 1264254976 + }, + { + "epoch": 0.38, + "learning_rate": 0.0004985516549648947, + "loss": 1.4459, + "theoretical_loss": 3.5988264161222183, + "tokens_seen": 1264320512 + }, + { + "epoch": 0.38, + "learning_rate": 0.0004985356068204614, + "loss": 1.4233, + "theoretical_loss": 3.5988095594802596, + "tokens_seen": 1264386048 + }, + { + "epoch": 0.38, + "learning_rate": 0.0004985195586760281, + "loss": 1.4001, + "theoretical_loss": 3.5987927039566228, + "tokens_seen": 1264451584 + }, + { + "epoch": 0.38, + "learning_rate": 0.0004985035105315948, + "loss": 1.5102, + "theoretical_loss": 3.5987758495511755, + "tokens_seen": 1264517120 + }, + { + "epoch": 0.38, + "learning_rate": 0.0004984874623871616, + "loss": 1.3457, + "theoretical_loss": 3.5987589962637863, + "tokens_seen": 1264582656 + }, + { + "epoch": 0.38, + "learning_rate": 0.0004984714142427282, + "loss": 1.4075, + "theoretical_loss": 3.5987421440943224, + "tokens_seen": 1264648192 + }, + { + "epoch": 0.38, + "learning_rate": 0.000498455366098295, + "loss": 1.3576, + "theoretical_loss": 3.5987252930426523, + "tokens_seen": 1264713728 + }, + { + "epoch": 0.38, + "learning_rate": 0.0004984393179538616, + "loss": 1.472, + "theoretical_loss": 3.5987084431086434, + "tokens_seen": 1264779264 + }, + { + "epoch": 0.38, + "objective/train/docs_used": 462829, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.4546477794647217, + "objective/train/theoretical_loss": 3.5986915942921636, + "objective/train/tokens_used": 1285304800, + "theoretical_loss": 3.5986915942921636, + "tokens_seen": 1264844800 + }, + { + "epoch": 0.38, + "learning_rate": 0.0004984232698094284, + "loss": 1.4197, + "theoretical_loss": 3.5986915942921636, + "tokens_seen": 1264844800 + }, + { + "epoch": 0.38, + "learning_rate": 0.000498407221664995, + "loss": 1.3854, + "theoretical_loss": 3.598674746593082, + "tokens_seen": 1264910336 + }, + { + "epoch": 0.38, + "learning_rate": 0.0004983911735205618, + "loss": 1.3011, + "theoretical_loss": 3.598657900011265, + "tokens_seen": 1264975872 + }, + { + "epoch": 0.38, + "learning_rate": 0.0004983751253761284, + "loss": 1.4081, + "theoretical_loss": 3.598641054546582, + "tokens_seen": 1265041408 + }, + { + "epoch": 0.38, + "learning_rate": 0.0004983590772316952, + "loss": 1.4489, + "theoretical_loss": 3.5986242101989006, + "tokens_seen": 1265106944 + }, + { + "epoch": 0.38, + "learning_rate": 0.0004983430290872618, + "loss": 1.3632, + "theoretical_loss": 3.5986073669680883, + "tokens_seen": 1265172480 + }, + { + "epoch": 0.38, + "learning_rate": 0.0004983269809428286, + "loss": 1.4351, + "theoretical_loss": 3.598590524854014, + "tokens_seen": 1265238016 + }, + { + "epoch": 0.38, + "learning_rate": 0.0004983109327983952, + "loss": 1.3587, + "theoretical_loss": 3.5985736838565456, + "tokens_seen": 1265303552 + }, + { + "epoch": 0.38, + "learning_rate": 0.000498294884653962, + "loss": 1.3329, + "theoretical_loss": 3.598556843975551, + "tokens_seen": 1265369088 + }, + { + "epoch": 0.38, + "learning_rate": 0.0004982788365095286, + "loss": 1.3772, + "theoretical_loss": 3.5985400052108982, + "tokens_seen": 1265434624 + }, + { + "epoch": 0.38, + "learning_rate": 0.0004982627883650954, + "loss": 1.426, + "theoretical_loss": 3.598523167562456, + "tokens_seen": 1265500160 + }, + { + "epoch": 0.38, + "learning_rate": 0.000498246740220662, + "loss": 1.4772, + "theoretical_loss": 3.5985063310300918, + "tokens_seen": 1265565696 + }, + { + "epoch": 0.38, + "learning_rate": 0.0004982306920762288, + "loss": 1.4759, + "theoretical_loss": 3.5984894956136753, + "tokens_seen": 1265631232 + }, + { + "epoch": 0.38, + "learning_rate": 0.0004982146439317954, + "loss": 1.3347, + "theoretical_loss": 3.598472661313073, + "tokens_seen": 1265696768 + }, + { + "epoch": 0.38, + "learning_rate": 0.0004981985957873622, + "loss": 1.3363, + "theoretical_loss": 3.598455828128154, + "tokens_seen": 1265762304 + }, + { + "epoch": 0.38, + "learning_rate": 0.0004981825476429288, + "loss": 1.4315, + "theoretical_loss": 3.598438996058787, + "tokens_seen": 1265827840 + }, + { + "epoch": 0.38, + "learning_rate": 0.0004981664994984955, + "loss": 1.3536, + "theoretical_loss": 3.5984221651048394, + "tokens_seen": 1265893376 + }, + { + "epoch": 0.38, + "learning_rate": 0.0004981504513540622, + "loss": 1.4311, + "theoretical_loss": 3.5984053352661807, + "tokens_seen": 1265958912 + }, + { + "epoch": 0.38, + "learning_rate": 0.0004981344032096289, + "loss": 1.4176, + "theoretical_loss": 3.598388506542678, + "tokens_seen": 1266024448 + }, + { + "epoch": 0.38, + "learning_rate": 0.0004981183550651956, + "loss": 1.5093, + "theoretical_loss": 3.5983716789342, + "tokens_seen": 1266089984 + }, + { + "epoch": 0.38, + "learning_rate": 0.0004981023069207623, + "loss": 1.5993, + "theoretical_loss": 3.598354852440616, + "tokens_seen": 1266155520 + }, + { + "epoch": 0.38, + "learning_rate": 0.000498086258776329, + "loss": 1.5055, + "theoretical_loss": 3.5983380270617937, + "tokens_seen": 1266221056 + }, + { + "epoch": 0.38, + "learning_rate": 0.0004980702106318957, + "loss": 1.3227, + "theoretical_loss": 3.5983212027976013, + "tokens_seen": 1266286592 + }, + { + "epoch": 0.38, + "learning_rate": 0.0004980541624874624, + "loss": 1.4677, + "theoretical_loss": 3.5983043796479084, + "tokens_seen": 1266352128 + }, + { + "epoch": 0.38, + "learning_rate": 0.0004980381143430291, + "loss": 1.3515, + "theoretical_loss": 3.598287557612582, + "tokens_seen": 1266417664 + }, + { + "epoch": 0.38, + "learning_rate": 0.0004980220661985957, + "loss": 1.478, + "theoretical_loss": 3.5982707366914917, + "tokens_seen": 1266483200 + }, + { + "epoch": 0.38, + "learning_rate": 0.0004980060180541625, + "loss": 1.3997, + "theoretical_loss": 3.598253916884506, + "tokens_seen": 1266548736 + }, + { + "epoch": 0.38, + "learning_rate": 0.0004979899699097291, + "loss": 1.3167, + "theoretical_loss": 3.5982370981914924, + "tokens_seen": 1266614272 + }, + { + "epoch": 0.38, + "learning_rate": 0.0004979739217652959, + "loss": 1.4274, + "theoretical_loss": 3.5982202806123205, + "tokens_seen": 1266679808 + }, + { + "epoch": 0.38, + "learning_rate": 0.0004979578736208625, + "loss": 1.4087, + "theoretical_loss": 3.5982034641468594, + "tokens_seen": 1266745344 + }, + { + "epoch": 0.38, + "learning_rate": 0.0004979418254764293, + "loss": 1.3572, + "theoretical_loss": 3.598186648794976, + "tokens_seen": 1266810880 + }, + { + "epoch": 0.38, + "learning_rate": 0.0004979257773319959, + "loss": 1.383, + "theoretical_loss": 3.5981698345565407, + "tokens_seen": 1266876416 + }, + { + "epoch": 0.38, + "learning_rate": 0.0004979097291875627, + "loss": 1.4438, + "theoretical_loss": 3.5981530214314215, + "tokens_seen": 1266941952 + }, + { + "epoch": 0.38, + "learning_rate": 0.0004978936810431293, + "loss": 1.3629, + "theoretical_loss": 3.5981362094194864, + "tokens_seen": 1267007488 + }, + { + "epoch": 0.38, + "learning_rate": 0.0004978776328986961, + "loss": 1.4488, + "theoretical_loss": 3.5981193985206055, + "tokens_seen": 1267073024 + }, + { + "epoch": 0.38, + "learning_rate": 0.0004978615847542627, + "loss": 1.4292, + "theoretical_loss": 3.5981025887346467, + "tokens_seen": 1267138560 + }, + { + "epoch": 0.38, + "learning_rate": 0.0004978455366098295, + "loss": 1.4067, + "theoretical_loss": 3.5980857800614787, + "tokens_seen": 1267204096 + }, + { + "epoch": 0.38, + "learning_rate": 0.0004978294884653961, + "loss": 1.4997, + "theoretical_loss": 3.5980689725009705, + "tokens_seen": 1267269632 + }, + { + "epoch": 0.38, + "learning_rate": 0.0004978134403209629, + "loss": 1.5427, + "theoretical_loss": 3.598052166052991, + "tokens_seen": 1267335168 + }, + { + "epoch": 0.38, + "learning_rate": 0.0004977973921765296, + "loss": 1.3572, + "theoretical_loss": 3.5980353607174087, + "tokens_seen": 1267400704 + }, + { + "epoch": 0.38, + "learning_rate": 0.0004977813440320963, + "loss": 1.4399, + "theoretical_loss": 3.5980185564940927, + "tokens_seen": 1267466240 + }, + { + "epoch": 0.38, + "learning_rate": 0.000497765295887663, + "loss": 1.4512, + "theoretical_loss": 3.598001753382912, + "tokens_seen": 1267531776 + }, + { + "epoch": 0.38, + "learning_rate": 0.0004977492477432297, + "loss": 1.5075, + "theoretical_loss": 3.5979849513837356, + "tokens_seen": 1267597312 + }, + { + "epoch": 0.38, + "learning_rate": 0.0004977331995987964, + "loss": 1.3097, + "theoretical_loss": 3.597968150496432, + "tokens_seen": 1267662848 + }, + { + "epoch": 0.38, + "learning_rate": 0.0004977171514543631, + "loss": 1.4977, + "theoretical_loss": 3.5979513507208702, + "tokens_seen": 1267728384 + }, + { + "epoch": 0.38, + "learning_rate": 0.0004977011033099298, + "loss": 1.404, + "theoretical_loss": 3.5979345520569193, + "tokens_seen": 1267793920 + }, + { + "epoch": 0.38, + "learning_rate": 0.0004976850551654965, + "loss": 1.4284, + "theoretical_loss": 3.597917754504448, + "tokens_seen": 1267859456 + }, + { + "epoch": 0.38, + "learning_rate": 0.0004976690070210632, + "loss": 1.4315, + "theoretical_loss": 3.5979009580633265, + "tokens_seen": 1267924992 + }, + { + "epoch": 0.38, + "learning_rate": 0.0004976529588766299, + "loss": 1.3963, + "theoretical_loss": 3.5978841627334224, + "tokens_seen": 1267990528 + }, + { + "epoch": 0.38, + "learning_rate": 0.0004976369107321966, + "loss": 1.4819, + "theoretical_loss": 3.5978673685146054, + "tokens_seen": 1268056064 + }, + { + "epoch": 0.38, + "objective/train/docs_used": 464032, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.4968687295913696, + "objective/train/theoretical_loss": 3.5978505754067442, + "objective/train/tokens_used": 1288581600, + "theoretical_loss": 3.5978505754067442, + "tokens_seen": 1268121600 + }, + { + "epoch": 0.38, + "learning_rate": 0.0004976208625877633, + "loss": 1.4579, + "theoretical_loss": 3.5978505754067442, + "tokens_seen": 1268121600 + }, + { + "epoch": 0.38, + "learning_rate": 0.00049760481444333, + "loss": 1.3963, + "theoretical_loss": 3.5978337834097083, + "tokens_seen": 1268187136 + }, + { + "epoch": 0.38, + "learning_rate": 0.0004975887662988967, + "loss": 1.3787, + "theoretical_loss": 3.5978169925233665, + "tokens_seen": 1268252672 + }, + { + "epoch": 0.38, + "learning_rate": 0.0004975727181544634, + "loss": 1.393, + "theoretical_loss": 3.597800202747589, + "tokens_seen": 1268318208 + }, + { + "epoch": 0.38, + "learning_rate": 0.0004975566700100301, + "loss": 1.4911, + "theoretical_loss": 3.5977834140822433, + "tokens_seen": 1268383744 + }, + { + "epoch": 0.38, + "learning_rate": 0.0004975406218655968, + "loss": 1.3132, + "theoretical_loss": 3.5977666265272, + "tokens_seen": 1268449280 + }, + { + "epoch": 0.38, + "learning_rate": 0.0004975245737211635, + "loss": 1.4501, + "theoretical_loss": 3.5977498400823267, + "tokens_seen": 1268514816 + }, + { + "epoch": 0.38, + "learning_rate": 0.0004975085255767302, + "loss": 1.5083, + "theoretical_loss": 3.5977330547474944, + "tokens_seen": 1268580352 + }, + { + "epoch": 0.38, + "learning_rate": 0.0004974924774322969, + "loss": 1.4265, + "theoretical_loss": 3.5977162705225716, + "tokens_seen": 1268645888 + }, + { + "epoch": 0.38, + "learning_rate": 0.0004974764292878636, + "loss": 1.4118, + "theoretical_loss": 3.5976994874074277, + "tokens_seen": 1268711424 + }, + { + "epoch": 0.38, + "learning_rate": 0.0004974603811434303, + "loss": 1.4168, + "theoretical_loss": 3.5976827054019314, + "tokens_seen": 1268776960 + }, + { + "epoch": 0.38, + "learning_rate": 0.000497444332998997, + "loss": 1.3904, + "theoretical_loss": 3.597665924505953, + "tokens_seen": 1268842496 + }, + { + "epoch": 0.38, + "learning_rate": 0.0004974282848545637, + "loss": 1.359, + "theoretical_loss": 3.5976491447193606, + "tokens_seen": 1268908032 + }, + { + "epoch": 0.38, + "learning_rate": 0.0004974122367101304, + "loss": 1.5459, + "theoretical_loss": 3.597632366042025, + "tokens_seen": 1268973568 + }, + { + "epoch": 0.38, + "learning_rate": 0.0004973961885656971, + "loss": 1.4142, + "theoretical_loss": 3.597615588473815, + "tokens_seen": 1269039104 + }, + { + "epoch": 0.38, + "learning_rate": 0.0004973801404212638, + "loss": 1.4063, + "theoretical_loss": 3.597598812014599, + "tokens_seen": 1269104640 + }, + { + "epoch": 0.38, + "learning_rate": 0.0004973640922768305, + "loss": 1.4036, + "theoretical_loss": 3.5975820366642486, + "tokens_seen": 1269170176 + }, + { + "epoch": 0.38, + "learning_rate": 0.0004973480441323972, + "loss": 1.3609, + "theoretical_loss": 3.5975652624226315, + "tokens_seen": 1269235712 + }, + { + "epoch": 0.38, + "learning_rate": 0.0004973319959879639, + "loss": 1.4294, + "theoretical_loss": 3.5975484892896175, + "tokens_seen": 1269301248 + }, + { + "epoch": 0.38, + "learning_rate": 0.0004973159478435306, + "loss": 1.4104, + "theoretical_loss": 3.597531717265076, + "tokens_seen": 1269366784 + }, + { + "epoch": 0.38, + "learning_rate": 0.0004972998996990973, + "loss": 1.4887, + "theoretical_loss": 3.5975149463488774, + "tokens_seen": 1269432320 + }, + { + "epoch": 0.38, + "learning_rate": 0.0004972838515546641, + "loss": 1.3968, + "theoretical_loss": 3.59749817654089, + "tokens_seen": 1269497856 + }, + { + "epoch": 0.38, + "learning_rate": 0.0004972678034102307, + "loss": 1.4512, + "theoretical_loss": 3.5974814078409847, + "tokens_seen": 1269563392 + }, + { + "epoch": 0.38, + "learning_rate": 0.0004972517552657975, + "loss": 1.3316, + "theoretical_loss": 3.5974646402490302, + "tokens_seen": 1269628928 + }, + { + "epoch": 0.38, + "learning_rate": 0.0004972357071213641, + "loss": 1.4414, + "theoretical_loss": 3.597447873764896, + "tokens_seen": 1269694464 + }, + { + "epoch": 0.38, + "learning_rate": 0.0004972196589769309, + "loss": 1.3611, + "theoretical_loss": 3.5974311083884523, + "tokens_seen": 1269760000 + }, + { + "epoch": 0.38, + "learning_rate": 0.0004972036108324975, + "loss": 1.3842, + "theoretical_loss": 3.597414344119569, + "tokens_seen": 1269825536 + }, + { + "epoch": 0.38, + "learning_rate": 0.0004971875626880643, + "loss": 1.4611, + "theoretical_loss": 3.5973975809581145, + "tokens_seen": 1269891072 + }, + { + "epoch": 0.38, + "learning_rate": 0.0004971715145436309, + "loss": 1.3466, + "theoretical_loss": 3.597380818903959, + "tokens_seen": 1269956608 + }, + { + "epoch": 0.38, + "learning_rate": 0.0004971554663991977, + "loss": 1.3983, + "theoretical_loss": 3.5973640579569732, + "tokens_seen": 1270022144 + }, + { + "epoch": 0.38, + "learning_rate": 0.0004971394182547643, + "loss": 1.4668, + "theoretical_loss": 3.5973472981170262, + "tokens_seen": 1270087680 + }, + { + "epoch": 0.38, + "learning_rate": 0.0004971233701103311, + "loss": 1.4188, + "theoretical_loss": 3.5973305393839876, + "tokens_seen": 1270153216 + }, + { + "epoch": 0.38, + "learning_rate": 0.0004971073219658977, + "loss": 1.3994, + "theoretical_loss": 3.5973137817577268, + "tokens_seen": 1270218752 + }, + { + "epoch": 0.38, + "learning_rate": 0.0004970912738214645, + "loss": 1.3151, + "theoretical_loss": 3.5972970252381145, + "tokens_seen": 1270284288 + }, + { + "epoch": 0.38, + "learning_rate": 0.0004970752256770311, + "loss": 1.3283, + "theoretical_loss": 3.59728026982502, + "tokens_seen": 1270349824 + }, + { + "epoch": 0.38, + "learning_rate": 0.0004970591775325979, + "loss": 1.4097, + "theoretical_loss": 3.5972635155183132, + "tokens_seen": 1270415360 + }, + { + "epoch": 0.38, + "learning_rate": 0.0004970431293881645, + "loss": 1.3857, + "theoretical_loss": 3.597246762317864, + "tokens_seen": 1270480896 + }, + { + "epoch": 0.39, + "learning_rate": 0.0004970270812437313, + "loss": 1.3691, + "theoretical_loss": 3.5972300102235426, + "tokens_seen": 1270546432 + }, + { + "epoch": 0.39, + "learning_rate": 0.0004970110330992979, + "loss": 1.4161, + "theoretical_loss": 3.597213259235219, + "tokens_seen": 1270611968 + }, + { + "epoch": 0.39, + "learning_rate": 0.0004969949849548647, + "loss": 1.364, + "theoretical_loss": 3.5971965093527625, + "tokens_seen": 1270677504 + }, + { + "epoch": 0.39, + "learning_rate": 0.0004969789368104313, + "loss": 1.3924, + "theoretical_loss": 3.597179760576043, + "tokens_seen": 1270743040 + }, + { + "epoch": 0.39, + "learning_rate": 0.0004969628886659981, + "loss": 1.3673, + "theoretical_loss": 3.5971630129049315, + "tokens_seen": 1270808576 + }, + { + "epoch": 0.39, + "learning_rate": 0.0004969468405215647, + "loss": 1.3574, + "theoretical_loss": 3.597146266339297, + "tokens_seen": 1270874112 + }, + { + "epoch": 0.39, + "learning_rate": 0.0004969307923771315, + "loss": 1.3364, + "theoretical_loss": 3.59712952087901, + "tokens_seen": 1270939648 + }, + { + "epoch": 0.39, + "learning_rate": 0.0004969147442326981, + "loss": 1.434, + "theoretical_loss": 3.5971127765239403, + "tokens_seen": 1271005184 + }, + { + "epoch": 0.39, + "learning_rate": 0.0004968986960882649, + "loss": 1.4565, + "theoretical_loss": 3.597096033273958, + "tokens_seen": 1271070720 + }, + { + "epoch": 0.39, + "learning_rate": 0.0004968826479438315, + "loss": 1.5145, + "theoretical_loss": 3.5970792911289333, + "tokens_seen": 1271136256 + }, + { + "epoch": 0.39, + "learning_rate": 0.0004968665997993983, + "loss": 1.3713, + "theoretical_loss": 3.5970625500887365, + "tokens_seen": 1271201792 + }, + { + "epoch": 0.39, + "learning_rate": 0.0004968505516549649, + "loss": 1.3116, + "theoretical_loss": 3.5970458101532374, + "tokens_seen": 1271267328 + }, + { + "epoch": 0.39, + "learning_rate": 0.0004968345035105317, + "loss": 1.3814, + "theoretical_loss": 3.5970290713223063, + "tokens_seen": 1271332864 + }, + { + "epoch": 0.39, + "objective/train/docs_used": 465264, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.4307596683502197, + "objective/train/theoretical_loss": 3.5970123335958135, + "objective/train/tokens_used": 1291858400, + "theoretical_loss": 3.5970123335958135, + "tokens_seen": 1271398400 + }, + { + "epoch": 0.39, + "learning_rate": 0.0004968184553660983, + "loss": 1.4111, + "theoretical_loss": 3.5970123335958135, + "tokens_seen": 1271398400 + }, + { + "epoch": 0.39, + "learning_rate": 0.0004968024072216651, + "loss": 1.4591, + "theoretical_loss": 3.5969955969736294, + "tokens_seen": 1271463936 + }, + { + "epoch": 0.39, + "learning_rate": 0.0004967863590772317, + "loss": 1.4539, + "theoretical_loss": 3.5969788614556237, + "tokens_seen": 1271529472 + }, + { + "epoch": 0.39, + "learning_rate": 0.0004967703109327984, + "loss": 1.3606, + "theoretical_loss": 3.596962127041667, + "tokens_seen": 1271595008 + }, + { + "epoch": 0.39, + "learning_rate": 0.0004967542627883651, + "loss": 1.4035, + "theoretical_loss": 3.5969453937316294, + "tokens_seen": 1271660544 + }, + { + "epoch": 0.39, + "learning_rate": 0.0004967382146439318, + "loss": 1.4078, + "theoretical_loss": 3.5969286615253813, + "tokens_seen": 1271726080 + }, + { + "epoch": 0.39, + "learning_rate": 0.0004967221664994985, + "loss": 1.419, + "theoretical_loss": 3.5969119304227926, + "tokens_seen": 1271791616 + }, + { + "epoch": 0.39, + "learning_rate": 0.0004967061183550652, + "loss": 1.3895, + "theoretical_loss": 3.5968952004237345, + "tokens_seen": 1271857152 + }, + { + "epoch": 0.39, + "learning_rate": 0.000496690070210632, + "loss": 1.3952, + "theoretical_loss": 3.5968784715280764, + "tokens_seen": 1271922688 + }, + { + "epoch": 0.39, + "learning_rate": 0.0004966740220661986, + "loss": 1.413, + "theoretical_loss": 3.5968617437356896, + "tokens_seen": 1271988224 + }, + { + "epoch": 0.39, + "learning_rate": 0.0004966579739217653, + "loss": 1.439, + "theoretical_loss": 3.5968450170464434, + "tokens_seen": 1272053760 + }, + { + "epoch": 0.39, + "learning_rate": 0.000496641925777332, + "loss": 1.3421, + "theoretical_loss": 3.596828291460209, + "tokens_seen": 1272119296 + }, + { + "epoch": 0.39, + "learning_rate": 0.0004966258776328986, + "loss": 1.3514, + "theoretical_loss": 3.596811566976857, + "tokens_seen": 1272184832 + }, + { + "epoch": 0.39, + "learning_rate": 0.0004966098294884654, + "loss": 1.4173, + "theoretical_loss": 3.596794843596258, + "tokens_seen": 1272250368 + }, + { + "epoch": 0.39, + "learning_rate": 0.000496593781344032, + "loss": 1.45, + "theoretical_loss": 3.5967781213182817, + "tokens_seen": 1272315904 + }, + { + "epoch": 0.39, + "learning_rate": 0.0004965777331995988, + "loss": 1.482, + "theoretical_loss": 3.5967614001427988, + "tokens_seen": 1272381440 + }, + { + "epoch": 0.39, + "learning_rate": 0.0004965616850551654, + "loss": 1.3381, + "theoretical_loss": 3.59674468006968, + "tokens_seen": 1272446976 + }, + { + "epoch": 0.39, + "learning_rate": 0.0004965456369107322, + "loss": 1.5095, + "theoretical_loss": 3.5967279610987957, + "tokens_seen": 1272512512 + }, + { + "epoch": 0.39, + "learning_rate": 0.0004965295887662988, + "loss": 1.3097, + "theoretical_loss": 3.596711243230017, + "tokens_seen": 1272578048 + }, + { + "epoch": 0.39, + "learning_rate": 0.0004965135406218656, + "loss": 1.3578, + "theoretical_loss": 3.5966945264632137, + "tokens_seen": 1272643584 + }, + { + "epoch": 0.39, + "learning_rate": 0.0004964974924774322, + "loss": 1.3113, + "theoretical_loss": 3.5966778107982575, + "tokens_seen": 1272709120 + }, + { + "epoch": 0.39, + "learning_rate": 0.000496481444332999, + "loss": 1.4111, + "theoretical_loss": 3.596661096235018, + "tokens_seen": 1272774656 + }, + { + "epoch": 0.39, + "learning_rate": 0.0004964653961885656, + "loss": 1.4648, + "theoretical_loss": 3.5966443827733663, + "tokens_seen": 1272840192 + }, + { + "epoch": 0.39, + "learning_rate": 0.0004964493480441324, + "loss": 1.4445, + "theoretical_loss": 3.596627670413173, + "tokens_seen": 1272905728 + }, + { + "epoch": 0.39, + "learning_rate": 0.000496433299899699, + "loss": 1.3416, + "theoretical_loss": 3.596610959154309, + "tokens_seen": 1272971264 + }, + { + "epoch": 0.39, + "learning_rate": 0.0004964172517552658, + "loss": 1.4718, + "theoretical_loss": 3.5965942489966447, + "tokens_seen": 1273036800 + }, + { + "epoch": 0.39, + "learning_rate": 0.0004964012036108325, + "loss": 1.4763, + "theoretical_loss": 3.5965775399400512, + "tokens_seen": 1273102336 + }, + { + "epoch": 0.39, + "learning_rate": 0.0004963851554663992, + "loss": 1.2934, + "theoretical_loss": 3.596560831984399, + "tokens_seen": 1273167872 + }, + { + "epoch": 0.39, + "learning_rate": 0.0004963691073219659, + "loss": 1.4196, + "theoretical_loss": 3.596544125129559, + "tokens_seen": 1273233408 + }, + { + "epoch": 0.39, + "learning_rate": 0.0004963530591775326, + "loss": 1.4694, + "theoretical_loss": 3.596527419375402, + "tokens_seen": 1273298944 + }, + { + "epoch": 0.39, + "learning_rate": 0.0004963370110330993, + "loss": 1.4229, + "theoretical_loss": 3.5965107147217994, + "tokens_seen": 1273364480 + }, + { + "epoch": 0.39, + "learning_rate": 0.000496320962888666, + "loss": 1.3776, + "theoretical_loss": 3.5964940111686206, + "tokens_seen": 1273430016 + }, + { + "epoch": 0.39, + "learning_rate": 0.0004963049147442327, + "loss": 1.4025, + "theoretical_loss": 3.5964773087157385, + "tokens_seen": 1273495552 + }, + { + "epoch": 0.39, + "learning_rate": 0.0004962888665997994, + "loss": 1.315, + "theoretical_loss": 3.5964606073630216, + "tokens_seen": 1273561088 + }, + { + "epoch": 0.39, + "learning_rate": 0.0004962728184553661, + "loss": 1.4265, + "theoretical_loss": 3.596443907110343, + "tokens_seen": 1273626624 + }, + { + "epoch": 0.39, + "learning_rate": 0.0004962567703109328, + "loss": 1.5534, + "theoretical_loss": 3.5964272079575728, + "tokens_seen": 1273692160 + }, + { + "epoch": 0.39, + "learning_rate": 0.0004962407221664995, + "loss": 1.4588, + "theoretical_loss": 3.5964105099045813, + "tokens_seen": 1273757696 + }, + { + "epoch": 0.39, + "learning_rate": 0.0004962246740220662, + "loss": 1.3674, + "theoretical_loss": 3.596393812951241, + "tokens_seen": 1273823232 + }, + { + "epoch": 0.39, + "learning_rate": 0.0004962086258776329, + "loss": 1.4657, + "theoretical_loss": 3.5963771170974215, + "tokens_seen": 1273888768 + }, + { + "epoch": 0.39, + "learning_rate": 0.0004961925777331996, + "loss": 1.5467, + "theoretical_loss": 3.596360422342995, + "tokens_seen": 1273954304 + }, + { + "epoch": 0.39, + "learning_rate": 0.0004961765295887663, + "loss": 1.4445, + "theoretical_loss": 3.596343728687831, + "tokens_seen": 1274019840 + }, + { + "epoch": 0.39, + "learning_rate": 0.000496160481444333, + "loss": 1.4641, + "theoretical_loss": 3.596327036131802, + "tokens_seen": 1274085376 + }, + { + "epoch": 0.39, + "learning_rate": 0.0004961444332998997, + "loss": 1.4883, + "theoretical_loss": 3.5963103446747784, + "tokens_seen": 1274150912 + }, + { + "epoch": 0.39, + "learning_rate": 0.0004961283851554664, + "loss": 1.3438, + "theoretical_loss": 3.596293654316632, + "tokens_seen": 1274216448 + }, + { + "epoch": 0.39, + "learning_rate": 0.0004961123370110331, + "loss": 1.5657, + "theoretical_loss": 3.5962769650572337, + "tokens_seen": 1274281984 + }, + { + "epoch": 0.39, + "learning_rate": 0.0004960962888665998, + "loss": 1.346, + "theoretical_loss": 3.5962602768964533, + "tokens_seen": 1274347520 + }, + { + "epoch": 0.39, + "learning_rate": 0.0004960802407221666, + "loss": 1.3477, + "theoretical_loss": 3.596243589834164, + "tokens_seen": 1274413056 + }, + { + "epoch": 0.39, + "learning_rate": 0.0004960641925777332, + "loss": 1.3288, + "theoretical_loss": 3.596226903870236, + "tokens_seen": 1274478592 + }, + { + "epoch": 0.39, + "learning_rate": 0.0004960481444333, + "loss": 1.3987, + "theoretical_loss": 3.596210219004541, + "tokens_seen": 1274544128 + }, + { + "epoch": 0.39, + "learning_rate": 0.0004960320962888666, + "loss": 1.4753, + "theoretical_loss": 3.5961935352369494, + "tokens_seen": 1274609664 + }, + { + "epoch": 0.39, + "objective/train/docs_used": 466434, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.383596658706665, + "objective/train/theoretical_loss": 3.5961768525673334, + "objective/train/tokens_used": 1295135200, + "theoretical_loss": 3.5961768525673334, + "tokens_seen": 1274675200 + }, + { + "epoch": 0.39, + "learning_rate": 0.0004960160481444334, + "loss": 1.3258, + "theoretical_loss": 3.5961768525673334, + "tokens_seen": 1274675200 + }, + { + "epoch": 0.39, + "learning_rate": 0.000496, + "loss": 1.405, + "theoretical_loss": 3.5961601709955637, + "tokens_seen": 1274740736 + }, + { + "epoch": 0.39, + "learning_rate": 0.0004959839518555668, + "loss": 1.4668, + "theoretical_loss": 3.596143490521512, + "tokens_seen": 1274806272 + }, + { + "epoch": 0.39, + "learning_rate": 0.0004959679037111334, + "loss": 1.5052, + "theoretical_loss": 3.5961268111450493, + "tokens_seen": 1274871808 + }, + { + "epoch": 0.39, + "learning_rate": 0.0004959518555667002, + "loss": 1.4507, + "theoretical_loss": 3.5961101328660474, + "tokens_seen": 1274937344 + }, + { + "epoch": 0.39, + "learning_rate": 0.0004959358074222668, + "loss": 1.4889, + "theoretical_loss": 3.596093455684377, + "tokens_seen": 1275002880 + }, + { + "epoch": 0.39, + "learning_rate": 0.0004959197592778336, + "loss": 1.3948, + "theoretical_loss": 3.59607677959991, + "tokens_seen": 1275068416 + }, + { + "epoch": 0.39, + "learning_rate": 0.0004959037111334002, + "loss": 1.3867, + "theoretical_loss": 3.596060104612518, + "tokens_seen": 1275133952 + }, + { + "epoch": 0.39, + "learning_rate": 0.000495887662988967, + "loss": 1.4581, + "theoretical_loss": 3.5960434307220717, + "tokens_seen": 1275199488 + }, + { + "epoch": 0.39, + "learning_rate": 0.0004958716148445336, + "loss": 1.385, + "theoretical_loss": 3.5960267579284437, + "tokens_seen": 1275265024 + }, + { + "epoch": 0.39, + "learning_rate": 0.0004958555667001004, + "loss": 1.5224, + "theoretical_loss": 3.5960100862315043, + "tokens_seen": 1275330560 + }, + { + "epoch": 0.39, + "learning_rate": 0.000495839518555667, + "loss": 1.4561, + "theoretical_loss": 3.5959934156311255, + "tokens_seen": 1275396096 + }, + { + "epoch": 0.39, + "learning_rate": 0.0004958234704112338, + "loss": 1.5305, + "theoretical_loss": 3.595976746127179, + "tokens_seen": 1275461632 + }, + { + "epoch": 0.39, + "learning_rate": 0.0004958074222668004, + "loss": 1.3993, + "theoretical_loss": 3.5959600777195364, + "tokens_seen": 1275527168 + }, + { + "epoch": 0.39, + "learning_rate": 0.0004957913741223672, + "loss": 1.4215, + "theoretical_loss": 3.5959434104080694, + "tokens_seen": 1275592704 + }, + { + "epoch": 0.39, + "learning_rate": 0.0004957753259779338, + "loss": 1.47, + "theoretical_loss": 3.5959267441926483, + "tokens_seen": 1275658240 + }, + { + "epoch": 0.39, + "learning_rate": 0.0004957592778335006, + "loss": 1.4732, + "theoretical_loss": 3.5959100790731466, + "tokens_seen": 1275723776 + }, + { + "epoch": 0.39, + "learning_rate": 0.0004957432296890672, + "loss": 1.415, + "theoretical_loss": 3.595893415049435, + "tokens_seen": 1275789312 + }, + { + "epoch": 0.39, + "learning_rate": 0.000495727181544634, + "loss": 1.3692, + "theoretical_loss": 3.5958767521213852, + "tokens_seen": 1275854848 + }, + { + "epoch": 0.39, + "learning_rate": 0.0004957111334002006, + "loss": 1.4431, + "theoretical_loss": 3.595860090288869, + "tokens_seen": 1275920384 + }, + { + "epoch": 0.39, + "learning_rate": 0.0004956950852557674, + "loss": 1.4667, + "theoretical_loss": 3.595843429551758, + "tokens_seen": 1275985920 + }, + { + "epoch": 0.39, + "learning_rate": 0.000495679037111334, + "loss": 1.3971, + "theoretical_loss": 3.595826769909924, + "tokens_seen": 1276051456 + }, + { + "epoch": 0.39, + "learning_rate": 0.0004956629889669008, + "loss": 1.4445, + "theoretical_loss": 3.595810111363238, + "tokens_seen": 1276116992 + }, + { + "epoch": 0.39, + "learning_rate": 0.0004956469408224674, + "loss": 1.4427, + "theoretical_loss": 3.5957934539115732, + "tokens_seen": 1276182528 + }, + { + "epoch": 0.39, + "learning_rate": 0.0004956308926780342, + "loss": 1.5161, + "theoretical_loss": 3.5957767975548007, + "tokens_seen": 1276248064 + }, + { + "epoch": 0.39, + "learning_rate": 0.0004956148445336008, + "loss": 1.3909, + "theoretical_loss": 3.595760142292792, + "tokens_seen": 1276313600 + }, + { + "epoch": 0.39, + "learning_rate": 0.0004955987963891676, + "loss": 1.2951, + "theoretical_loss": 3.5957434881254198, + "tokens_seen": 1276379136 + }, + { + "epoch": 0.39, + "learning_rate": 0.0004955827482447342, + "loss": 1.4569, + "theoretical_loss": 3.595726835052555, + "tokens_seen": 1276444672 + }, + { + "epoch": 0.39, + "learning_rate": 0.000495566700100301, + "loss": 1.4217, + "theoretical_loss": 3.5957101830740696, + "tokens_seen": 1276510208 + }, + { + "epoch": 0.39, + "learning_rate": 0.0004955506519558676, + "loss": 1.4296, + "theoretical_loss": 3.595693532189836, + "tokens_seen": 1276575744 + }, + { + "epoch": 0.39, + "learning_rate": 0.0004955346038114344, + "loss": 1.4328, + "theoretical_loss": 3.595676882399726, + "tokens_seen": 1276641280 + }, + { + "epoch": 0.39, + "learning_rate": 0.000495518555667001, + "loss": 1.5384, + "theoretical_loss": 3.5956602337036117, + "tokens_seen": 1276706816 + }, + { + "epoch": 0.39, + "learning_rate": 0.0004955025075225678, + "loss": 1.5539, + "theoretical_loss": 3.5956435861013643, + "tokens_seen": 1276772352 + }, + { + "epoch": 0.39, + "learning_rate": 0.0004954864593781344, + "loss": 1.3995, + "theoretical_loss": 3.595626939592856, + "tokens_seen": 1276837888 + }, + { + "epoch": 0.39, + "learning_rate": 0.0004954704112337012, + "loss": 1.4694, + "theoretical_loss": 3.59561029417796, + "tokens_seen": 1276903424 + }, + { + "epoch": 0.39, + "learning_rate": 0.0004954543630892678, + "loss": 1.3313, + "theoretical_loss": 3.595593649856547, + "tokens_seen": 1276968960 + }, + { + "epoch": 0.39, + "learning_rate": 0.0004954383149448346, + "loss": 1.413, + "theoretical_loss": 3.5955770066284893, + "tokens_seen": 1277034496 + }, + { + "epoch": 0.39, + "learning_rate": 0.0004954222668004012, + "loss": 1.3809, + "theoretical_loss": 3.595560364493659, + "tokens_seen": 1277100032 + }, + { + "epoch": 0.39, + "learning_rate": 0.000495406218655968, + "loss": 1.5302, + "theoretical_loss": 3.595543723451929, + "tokens_seen": 1277165568 + }, + { + "epoch": 0.39, + "learning_rate": 0.0004953901705115346, + "loss": 1.3781, + "theoretical_loss": 3.5955270835031703, + "tokens_seen": 1277231104 + }, + { + "epoch": 0.39, + "learning_rate": 0.0004953741223671013, + "loss": 1.3012, + "theoretical_loss": 3.595510444647256, + "tokens_seen": 1277296640 + }, + { + "epoch": 0.39, + "learning_rate": 0.000495358074222668, + "loss": 1.3779, + "theoretical_loss": 3.5954938068840576, + "tokens_seen": 1277362176 + }, + { + "epoch": 0.39, + "learning_rate": 0.0004953420260782347, + "loss": 1.4239, + "theoretical_loss": 3.5954771702134476, + "tokens_seen": 1277427712 + }, + { + "epoch": 0.39, + "learning_rate": 0.0004953259779338014, + "loss": 1.433, + "theoretical_loss": 3.5954605346352975, + "tokens_seen": 1277493248 + }, + { + "epoch": 0.39, + "learning_rate": 0.0004953099297893681, + "loss": 1.4405, + "theoretical_loss": 3.5954439001494802, + "tokens_seen": 1277558784 + }, + { + "epoch": 0.39, + "learning_rate": 0.0004952938816449348, + "loss": 1.4369, + "theoretical_loss": 3.595427266755868, + "tokens_seen": 1277624320 + }, + { + "epoch": 0.39, + "learning_rate": 0.0004952778335005015, + "loss": 1.4638, + "theoretical_loss": 3.595410634454333, + "tokens_seen": 1277689856 + }, + { + "epoch": 0.39, + "learning_rate": 0.0004952617853560683, + "loss": 1.4054, + "theoretical_loss": 3.5953940032447473, + "tokens_seen": 1277755392 + }, + { + "epoch": 0.39, + "learning_rate": 0.0004952457372116349, + "loss": 1.3748, + "theoretical_loss": 3.5953773731269836, + "tokens_seen": 1277820928 + }, + { + "epoch": 0.39, + "learning_rate": 0.0004952296890672015, + "loss": 1.496, + "theoretical_loss": 3.595360744100914, + "tokens_seen": 1277886464 + }, + { + "epoch": 0.39, + "objective/train/docs_used": 467639, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.3722994327545166, + "objective/train/theoretical_loss": 3.595344116166411, + "objective/train/tokens_used": 1298412000, + "theoretical_loss": 3.595344116166411, + "tokens_seen": 1277952000 + }, + { + "epoch": 0.39, + "learning_rate": 0.0004952136409227683, + "loss": 1.3982, + "theoretical_loss": 3.595344116166411, + "tokens_seen": 1277952000 + }, + { + "epoch": 0.39, + "learning_rate": 0.0004951975927783349, + "loss": 1.4287, + "theoretical_loss": 3.5953274893233464, + "tokens_seen": 1278017536 + }, + { + "epoch": 0.39, + "learning_rate": 0.0004951815446339017, + "loss": 1.3937, + "theoretical_loss": 3.595310863571594, + "tokens_seen": 1278083072 + }, + { + "epoch": 0.39, + "learning_rate": 0.0004951654964894683, + "loss": 1.448, + "theoretical_loss": 3.5952942389110243, + "tokens_seen": 1278148608 + }, + { + "epoch": 0.39, + "learning_rate": 0.0004951494483450351, + "loss": 1.4291, + "theoretical_loss": 3.5952776153415114, + "tokens_seen": 1278214144 + }, + { + "epoch": 0.39, + "learning_rate": 0.0004951334002006017, + "loss": 1.4523, + "theoretical_loss": 3.5952609928629267, + "tokens_seen": 1278279680 + }, + { + "epoch": 0.39, + "learning_rate": 0.0004951173520561685, + "loss": 1.303, + "theoretical_loss": 3.5952443714751428, + "tokens_seen": 1278345216 + }, + { + "epoch": 0.39, + "learning_rate": 0.0004951013039117351, + "loss": 1.3469, + "theoretical_loss": 3.5952277511780326, + "tokens_seen": 1278410752 + }, + { + "epoch": 0.39, + "learning_rate": 0.0004950852557673019, + "loss": 1.5449, + "theoretical_loss": 3.5952111319714684, + "tokens_seen": 1278476288 + }, + { + "epoch": 0.39, + "learning_rate": 0.0004950692076228685, + "loss": 1.4048, + "theoretical_loss": 3.5951945138553234, + "tokens_seen": 1278541824 + }, + { + "epoch": 0.39, + "learning_rate": 0.0004950531594784353, + "loss": 1.4956, + "theoretical_loss": 3.595177896829469, + "tokens_seen": 1278607360 + }, + { + "epoch": 0.39, + "learning_rate": 0.000495037111334002, + "loss": 1.4513, + "theoretical_loss": 3.5951612808937785, + "tokens_seen": 1278672896 + }, + { + "epoch": 0.39, + "learning_rate": 0.0004950210631895687, + "loss": 1.4612, + "theoretical_loss": 3.5951446660481245, + "tokens_seen": 1278738432 + }, + { + "epoch": 0.39, + "learning_rate": 0.0004950050150451354, + "loss": 1.3874, + "theoretical_loss": 3.5951280522923796, + "tokens_seen": 1278803968 + }, + { + "epoch": 0.39, + "learning_rate": 0.0004949889669007021, + "loss": 1.3474, + "theoretical_loss": 3.5951114396264163, + "tokens_seen": 1278869504 + }, + { + "epoch": 0.39, + "learning_rate": 0.0004949729187562688, + "loss": 1.3607, + "theoretical_loss": 3.595094828050107, + "tokens_seen": 1278935040 + }, + { + "epoch": 0.39, + "learning_rate": 0.0004949568706118355, + "loss": 1.4988, + "theoretical_loss": 3.5950782175633256, + "tokens_seen": 1279000576 + }, + { + "epoch": 0.39, + "learning_rate": 0.0004949408224674022, + "loss": 1.3573, + "theoretical_loss": 3.5950616081659437, + "tokens_seen": 1279066112 + }, + { + "epoch": 0.39, + "learning_rate": 0.0004949247743229689, + "loss": 1.4525, + "theoretical_loss": 3.5950449998578335, + "tokens_seen": 1279131648 + }, + { + "epoch": 0.39, + "learning_rate": 0.0004949087261785356, + "loss": 1.4479, + "theoretical_loss": 3.5950283926388695, + "tokens_seen": 1279197184 + }, + { + "epoch": 0.39, + "learning_rate": 0.0004948926780341023, + "loss": 1.4048, + "theoretical_loss": 3.5950117865089233, + "tokens_seen": 1279262720 + }, + { + "epoch": 0.39, + "learning_rate": 0.0004948766298896691, + "loss": 1.348, + "theoretical_loss": 3.594995181467868, + "tokens_seen": 1279328256 + }, + { + "epoch": 0.39, + "learning_rate": 0.0004948605817452357, + "loss": 1.3352, + "theoretical_loss": 3.594978577515576, + "tokens_seen": 1279393792 + }, + { + "epoch": 0.39, + "learning_rate": 0.0004948445336008025, + "loss": 1.422, + "theoretical_loss": 3.5949619746519206, + "tokens_seen": 1279459328 + }, + { + "epoch": 0.39, + "learning_rate": 0.0004948284854563691, + "loss": 1.3962, + "theoretical_loss": 3.594945372876775, + "tokens_seen": 1279524864 + }, + { + "epoch": 0.39, + "learning_rate": 0.0004948124373119359, + "loss": 1.5224, + "theoretical_loss": 3.594928772190011, + "tokens_seen": 1279590400 + }, + { + "epoch": 0.39, + "learning_rate": 0.0004947963891675025, + "loss": 1.4741, + "theoretical_loss": 3.5949121725915028, + "tokens_seen": 1279655936 + }, + { + "epoch": 0.39, + "learning_rate": 0.0004947803410230693, + "loss": 1.3579, + "theoretical_loss": 3.5948955740811224, + "tokens_seen": 1279721472 + }, + { + "epoch": 0.39, + "learning_rate": 0.0004947642928786359, + "loss": 1.4757, + "theoretical_loss": 3.5948789766587437, + "tokens_seen": 1279787008 + }, + { + "epoch": 0.39, + "learning_rate": 0.0004947482447342027, + "loss": 1.3922, + "theoretical_loss": 3.594862380324238, + "tokens_seen": 1279852544 + }, + { + "epoch": 0.39, + "learning_rate": 0.0004947321965897693, + "loss": 1.3852, + "theoretical_loss": 3.5948457850774798, + "tokens_seen": 1279918080 + }, + { + "epoch": 0.39, + "learning_rate": 0.0004947161484453361, + "loss": 1.4149, + "theoretical_loss": 3.5948291909183423, + "tokens_seen": 1279983616 + }, + { + "epoch": 0.39, + "learning_rate": 0.0004947001003009027, + "loss": 1.4713, + "theoretical_loss": 3.5948125978466967, + "tokens_seen": 1280049152 + }, + { + "epoch": 0.39, + "learning_rate": 0.0004946840521564695, + "loss": 1.4721, + "theoretical_loss": 3.594796005862418, + "tokens_seen": 1280114688 + }, + { + "epoch": 0.39, + "learning_rate": 0.0004946680040120361, + "loss": 1.396, + "theoretical_loss": 3.5947794149653785, + "tokens_seen": 1280180224 + }, + { + "epoch": 0.39, + "learning_rate": 0.0004946519558676029, + "loss": 1.4532, + "theoretical_loss": 3.5947628251554518, + "tokens_seen": 1280245760 + }, + { + "epoch": 0.39, + "learning_rate": 0.0004946359077231695, + "loss": 1.4172, + "theoretical_loss": 3.59474623643251, + "tokens_seen": 1280311296 + }, + { + "epoch": 0.39, + "learning_rate": 0.0004946198595787363, + "loss": 1.4745, + "theoretical_loss": 3.5947296487964264, + "tokens_seen": 1280376832 + }, + { + "epoch": 0.39, + "learning_rate": 0.0004946038114343029, + "loss": 1.3816, + "theoretical_loss": 3.594713062247075, + "tokens_seen": 1280442368 + }, + { + "epoch": 0.39, + "learning_rate": 0.0004945877632898697, + "loss": 1.3534, + "theoretical_loss": 3.594696476784329, + "tokens_seen": 1280507904 + }, + { + "epoch": 0.39, + "learning_rate": 0.0004945717151454363, + "loss": 1.3671, + "theoretical_loss": 3.5946798924080605, + "tokens_seen": 1280573440 + }, + { + "epoch": 0.39, + "learning_rate": 0.0004945556670010031, + "loss": 1.4669, + "theoretical_loss": 3.5946633091181437, + "tokens_seen": 1280638976 + }, + { + "epoch": 0.39, + "learning_rate": 0.0004945396188565697, + "loss": 1.5014, + "theoretical_loss": 3.5946467269144513, + "tokens_seen": 1280704512 + }, + { + "epoch": 0.39, + "learning_rate": 0.0004945235707121365, + "loss": 1.3232, + "theoretical_loss": 3.5946301457968577, + "tokens_seen": 1280770048 + }, + { + "epoch": 0.39, + "learning_rate": 0.0004945075225677031, + "loss": 1.5191, + "theoretical_loss": 3.5946135657652345, + "tokens_seen": 1280835584 + }, + { + "epoch": 0.39, + "learning_rate": 0.0004944914744232699, + "loss": 1.5384, + "theoretical_loss": 3.5945969868194556, + "tokens_seen": 1280901120 + }, + { + "epoch": 0.39, + "learning_rate": 0.0004944754262788365, + "loss": 1.4353, + "theoretical_loss": 3.594580408959395, + "tokens_seen": 1280966656 + }, + { + "epoch": 0.39, + "learning_rate": 0.0004944593781344033, + "loss": 1.5049, + "theoretical_loss": 3.5945638321849254, + "tokens_seen": 1281032192 + }, + { + "epoch": 0.39, + "learning_rate": 0.0004944433299899699, + "loss": 1.4382, + "theoretical_loss": 3.594547256495921, + "tokens_seen": 1281097728 + }, + { + "epoch": 0.39, + "learning_rate": 0.0004944272818455367, + "loss": 1.392, + "theoretical_loss": 3.594530681892254, + "tokens_seen": 1281163264 + }, + { + "epoch": 0.39, + "objective/train/docs_used": 468763, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.5896649360656738, + "objective/train/theoretical_loss": 3.594514108373799, + "objective/train/tokens_used": 1301688800, + "theoretical_loss": 3.594514108373799, + "tokens_seen": 1281228800 + }, + { + "epoch": 0.39, + "learning_rate": 0.0004944112337011033, + "loss": 1.5117, + "theoretical_loss": 3.594514108373799, + "tokens_seen": 1281228800 + }, + { + "epoch": 0.39, + "learning_rate": 0.0004943951855566701, + "loss": 1.4606, + "theoretical_loss": 3.5944975359404285, + "tokens_seen": 1281294336 + }, + { + "epoch": 0.39, + "learning_rate": 0.0004943791374122367, + "loss": 1.3899, + "theoretical_loss": 3.594480964592016, + "tokens_seen": 1281359872 + }, + { + "epoch": 0.39, + "learning_rate": 0.0004943630892678035, + "loss": 1.5473, + "theoretical_loss": 3.594464394328436, + "tokens_seen": 1281425408 + }, + { + "epoch": 0.39, + "learning_rate": 0.0004943470411233701, + "loss": 1.3922, + "theoretical_loss": 3.594447825149561, + "tokens_seen": 1281490944 + }, + { + "epoch": 0.39, + "learning_rate": 0.0004943309929789369, + "loss": 1.4859, + "theoretical_loss": 3.5944312570552643, + "tokens_seen": 1281556480 + }, + { + "epoch": 0.39, + "learning_rate": 0.0004943149448345035, + "loss": 1.4117, + "theoretical_loss": 3.594414690045421, + "tokens_seen": 1281622016 + }, + { + "epoch": 0.39, + "learning_rate": 0.0004942988966900703, + "loss": 1.4201, + "theoretical_loss": 3.594398124119903, + "tokens_seen": 1281687552 + }, + { + "epoch": 0.39, + "learning_rate": 0.0004942828485456369, + "loss": 1.376, + "theoretical_loss": 3.5943815592785846, + "tokens_seen": 1281753088 + }, + { + "epoch": 0.39, + "learning_rate": 0.0004942668004012037, + "loss": 1.4821, + "theoretical_loss": 3.5943649955213397, + "tokens_seen": 1281818624 + }, + { + "epoch": 0.39, + "learning_rate": 0.0004942507522567703, + "loss": 1.3877, + "theoretical_loss": 3.594348432848041, + "tokens_seen": 1281884160 + }, + { + "epoch": 0.39, + "learning_rate": 0.0004942347041123371, + "loss": 1.3904, + "theoretical_loss": 3.5943318712585635, + "tokens_seen": 1281949696 + }, + { + "epoch": 0.39, + "learning_rate": 0.0004942186559679037, + "loss": 1.4088, + "theoretical_loss": 3.59431531075278, + "tokens_seen": 1282015232 + }, + { + "epoch": 0.39, + "learning_rate": 0.0004942026078234705, + "loss": 1.4686, + "theoretical_loss": 3.594298751330564, + "tokens_seen": 1282080768 + }, + { + "epoch": 0.39, + "learning_rate": 0.0004941865596790371, + "loss": 1.4062, + "theoretical_loss": 3.5942821929917894, + "tokens_seen": 1282146304 + }, + { + "epoch": 0.39, + "learning_rate": 0.0004941705115346039, + "loss": 1.4427, + "theoretical_loss": 3.5942656357363307, + "tokens_seen": 1282211840 + }, + { + "epoch": 0.39, + "learning_rate": 0.0004941544633901705, + "loss": 1.4302, + "theoretical_loss": 3.5942490795640607, + "tokens_seen": 1282277376 + }, + { + "epoch": 0.39, + "learning_rate": 0.0004941384152457373, + "loss": 1.371, + "theoretical_loss": 3.5942325244748536, + "tokens_seen": 1282342912 + }, + { + "epoch": 0.39, + "learning_rate": 0.0004941223671013039, + "loss": 1.3861, + "theoretical_loss": 3.594215970468583, + "tokens_seen": 1282408448 + }, + { + "epoch": 0.39, + "learning_rate": 0.0004941063189568707, + "loss": 1.4056, + "theoretical_loss": 3.594199417545123, + "tokens_seen": 1282473984 + }, + { + "epoch": 0.39, + "learning_rate": 0.0004940902708124373, + "loss": 1.2783, + "theoretical_loss": 3.5941828657043473, + "tokens_seen": 1282539520 + }, + { + "epoch": 0.39, + "learning_rate": 0.0004940742226680041, + "loss": 1.5182, + "theoretical_loss": 3.5941663149461296, + "tokens_seen": 1282605056 + }, + { + "epoch": 0.39, + "learning_rate": 0.0004940581745235707, + "loss": 1.264, + "theoretical_loss": 3.594149765270344, + "tokens_seen": 1282670592 + }, + { + "epoch": 0.39, + "learning_rate": 0.0004940421263791375, + "loss": 1.3144, + "theoretical_loss": 3.5941332166768647, + "tokens_seen": 1282736128 + }, + { + "epoch": 0.39, + "learning_rate": 0.0004940260782347041, + "loss": 1.3059, + "theoretical_loss": 3.594116669165565, + "tokens_seen": 1282801664 + }, + { + "epoch": 0.39, + "learning_rate": 0.0004940100300902709, + "loss": 1.2353, + "theoretical_loss": 3.594100122736319, + "tokens_seen": 1282867200 + }, + { + "epoch": 0.39, + "learning_rate": 0.0004939939819458375, + "loss": 1.3743, + "theoretical_loss": 3.594083577389001, + "tokens_seen": 1282932736 + }, + { + "epoch": 0.39, + "learning_rate": 0.0004939779338014042, + "loss": 1.3948, + "theoretical_loss": 3.594067033123485, + "tokens_seen": 1282998272 + }, + { + "epoch": 0.39, + "learning_rate": 0.000493961885656971, + "loss": 1.3336, + "theoretical_loss": 3.5940504899396446, + "tokens_seen": 1283063808 + }, + { + "epoch": 0.39, + "learning_rate": 0.0004939458375125376, + "loss": 1.4487, + "theoretical_loss": 3.594033947837354, + "tokens_seen": 1283129344 + }, + { + "epoch": 0.39, + "learning_rate": 0.0004939297893681043, + "loss": 1.3709, + "theoretical_loss": 3.5940174068164876, + "tokens_seen": 1283194880 + }, + { + "epoch": 0.39, + "learning_rate": 0.000493913741223671, + "loss": 1.4266, + "theoretical_loss": 3.594000866876919, + "tokens_seen": 1283260416 + }, + { + "epoch": 0.39, + "learning_rate": 0.0004938976930792378, + "loss": 1.3509, + "theoretical_loss": 3.5939843280185224, + "tokens_seen": 1283325952 + }, + { + "epoch": 0.39, + "learning_rate": 0.0004938816449348044, + "loss": 1.4004, + "theoretical_loss": 3.593967790241172, + "tokens_seen": 1283391488 + }, + { + "epoch": 0.39, + "learning_rate": 0.0004938655967903712, + "loss": 1.3232, + "theoretical_loss": 3.593951253544742, + "tokens_seen": 1283457024 + }, + { + "epoch": 0.39, + "learning_rate": 0.0004938495486459378, + "loss": 1.3661, + "theoretical_loss": 3.593934717929107, + "tokens_seen": 1283522560 + }, + { + "epoch": 0.39, + "learning_rate": 0.0004938335005015044, + "loss": 1.3544, + "theoretical_loss": 3.5939181833941403, + "tokens_seen": 1283588096 + }, + { + "epoch": 0.39, + "learning_rate": 0.0004938174523570712, + "loss": 1.4475, + "theoretical_loss": 3.593901649939716, + "tokens_seen": 1283653632 + }, + { + "epoch": 0.39, + "learning_rate": 0.0004938014042126378, + "loss": 1.4494, + "theoretical_loss": 3.5938851175657094, + "tokens_seen": 1283719168 + }, + { + "epoch": 0.39, + "learning_rate": 0.0004937853560682046, + "loss": 1.3593, + "theoretical_loss": 3.593868586271994, + "tokens_seen": 1283784704 + }, + { + "epoch": 0.39, + "learning_rate": 0.0004937693079237712, + "loss": 1.5645, + "theoretical_loss": 3.593852056058444, + "tokens_seen": 1283850240 + }, + { + "epoch": 0.39, + "learning_rate": 0.000493753259779338, + "loss": 1.4172, + "theoretical_loss": 3.5938355269249347, + "tokens_seen": 1283915776 + }, + { + "epoch": 0.39, + "learning_rate": 0.0004937372116349046, + "loss": 1.289, + "theoretical_loss": 3.593818998871339, + "tokens_seen": 1283981312 + }, + { + "epoch": 0.39, + "learning_rate": 0.0004937211634904714, + "loss": 1.3845, + "theoretical_loss": 3.593802471897532, + "tokens_seen": 1284046848 + }, + { + "epoch": 0.39, + "learning_rate": 0.000493705115346038, + "loss": 1.4528, + "theoretical_loss": 3.593785946003388, + "tokens_seen": 1284112384 + }, + { + "epoch": 0.39, + "learning_rate": 0.0004936890672016048, + "loss": 1.3849, + "theoretical_loss": 3.593769421188781, + "tokens_seen": 1284177920 + }, + { + "epoch": 0.39, + "learning_rate": 0.0004936730190571716, + "loss": 1.4019, + "theoretical_loss": 3.5937528974535864, + "tokens_seen": 1284243456 + }, + { + "epoch": 0.39, + "learning_rate": 0.0004936569709127382, + "loss": 1.4501, + "theoretical_loss": 3.593736374797677, + "tokens_seen": 1284308992 + }, + { + "epoch": 0.39, + "learning_rate": 0.000493640922768305, + "loss": 1.3089, + "theoretical_loss": 3.5937198532209287, + "tokens_seen": 1284374528 + }, + { + "epoch": 0.39, + "learning_rate": 0.0004936248746238716, + "loss": 1.5018, + "theoretical_loss": 3.593703332723215, + "tokens_seen": 1284440064 + }, + { + "epoch": 0.39, + "objective/train/docs_used": 470107, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.4312293529510498, + "objective/train/theoretical_loss": 3.5936868133044113, + "objective/train/tokens_used": 1304965600, + "theoretical_loss": 3.5936868133044113, + "tokens_seen": 1284505600 + }, + { + "epoch": 0.39, + "learning_rate": 0.0004936088264794384, + "loss": 1.3242, + "theoretical_loss": 3.5936868133044113, + "tokens_seen": 1284505600 + }, + { + "epoch": 0.39, + "learning_rate": 0.000493592778335005, + "loss": 1.4325, + "theoretical_loss": 3.593670294964391, + "tokens_seen": 1284571136 + }, + { + "epoch": 0.39, + "learning_rate": 0.0004935767301905718, + "loss": 1.4704, + "theoretical_loss": 3.5936537777030293, + "tokens_seen": 1284636672 + }, + { + "epoch": 0.39, + "learning_rate": 0.0004935606820461384, + "loss": 1.3448, + "theoretical_loss": 3.593637261520201, + "tokens_seen": 1284702208 + }, + { + "epoch": 0.39, + "learning_rate": 0.0004935446339017052, + "loss": 1.48, + "theoretical_loss": 3.59362074641578, + "tokens_seen": 1284767744 + }, + { + "epoch": 0.39, + "learning_rate": 0.0004935285857572718, + "loss": 1.3919, + "theoretical_loss": 3.593604232389641, + "tokens_seen": 1284833280 + }, + { + "epoch": 0.39, + "learning_rate": 0.0004935125376128386, + "loss": 1.3735, + "theoretical_loss": 3.593587719441659, + "tokens_seen": 1284898816 + }, + { + "epoch": 0.39, + "learning_rate": 0.0004934964894684052, + "loss": 1.3457, + "theoretical_loss": 3.593571207571708, + "tokens_seen": 1284964352 + }, + { + "epoch": 0.39, + "learning_rate": 0.000493480441323972, + "loss": 1.3668, + "theoretical_loss": 3.593554696779664, + "tokens_seen": 1285029888 + }, + { + "epoch": 0.39, + "learning_rate": 0.0004934643931795386, + "loss": 1.3577, + "theoretical_loss": 3.5935381870653997, + "tokens_seen": 1285095424 + }, + { + "epoch": 0.39, + "learning_rate": 0.0004934483450351054, + "loss": 1.4279, + "theoretical_loss": 3.5935216784287913, + "tokens_seen": 1285160960 + }, + { + "epoch": 0.39, + "learning_rate": 0.000493432296890672, + "loss": 1.3773, + "theoretical_loss": 3.5935051708697126, + "tokens_seen": 1285226496 + }, + { + "epoch": 0.39, + "learning_rate": 0.0004934162487462388, + "loss": 1.4168, + "theoretical_loss": 3.5934886643880395, + "tokens_seen": 1285292032 + }, + { + "epoch": 0.39, + "learning_rate": 0.0004934002006018054, + "loss": 1.4189, + "theoretical_loss": 3.593472158983645, + "tokens_seen": 1285357568 + }, + { + "epoch": 0.39, + "learning_rate": 0.0004933841524573722, + "loss": 1.489, + "theoretical_loss": 3.5934556546564056, + "tokens_seen": 1285423104 + }, + { + "epoch": 0.39, + "learning_rate": 0.0004933681043129388, + "loss": 1.4381, + "theoretical_loss": 3.593439151406195, + "tokens_seen": 1285488640 + }, + { + "epoch": 0.39, + "learning_rate": 0.0004933520561685056, + "loss": 1.3727, + "theoretical_loss": 3.5934226492328882, + "tokens_seen": 1285554176 + }, + { + "epoch": 0.39, + "learning_rate": 0.0004933360080240722, + "loss": 1.4527, + "theoretical_loss": 3.593406148136361, + "tokens_seen": 1285619712 + }, + { + "epoch": 0.39, + "learning_rate": 0.000493319959879639, + "loss": 1.4298, + "theoretical_loss": 3.593389648116487, + "tokens_seen": 1285685248 + }, + { + "epoch": 0.39, + "learning_rate": 0.0004933039117352056, + "loss": 1.545, + "theoretical_loss": 3.593373149173141, + "tokens_seen": 1285750784 + }, + { + "epoch": 0.39, + "learning_rate": 0.0004932878635907724, + "loss": 1.4177, + "theoretical_loss": 3.5933566513061987, + "tokens_seen": 1285816320 + }, + { + "epoch": 0.39, + "learning_rate": 0.000493271815446339, + "loss": 1.3228, + "theoretical_loss": 3.593340154515535, + "tokens_seen": 1285881856 + }, + { + "epoch": 0.39, + "learning_rate": 0.0004932557673019058, + "loss": 1.3875, + "theoretical_loss": 3.5933236588010242, + "tokens_seen": 1285947392 + }, + { + "epoch": 0.39, + "learning_rate": 0.0004932397191574724, + "loss": 1.4245, + "theoretical_loss": 3.5933071641625416, + "tokens_seen": 1286012928 + }, + { + "epoch": 0.39, + "learning_rate": 0.0004932236710130392, + "loss": 1.4218, + "theoretical_loss": 3.5932906705999628, + "tokens_seen": 1286078464 + }, + { + "epoch": 0.39, + "learning_rate": 0.0004932076228686058, + "loss": 1.5381, + "theoretical_loss": 3.5932741781131616, + "tokens_seen": 1286144000 + }, + { + "epoch": 0.39, + "learning_rate": 0.0004931915747241726, + "loss": 1.3571, + "theoretical_loss": 3.593257686702014, + "tokens_seen": 1286209536 + }, + { + "epoch": 0.39, + "learning_rate": 0.0004931755265797392, + "loss": 1.3472, + "theoretical_loss": 3.5932411963663946, + "tokens_seen": 1286275072 + }, + { + "epoch": 0.39, + "learning_rate": 0.000493159478435306, + "loss": 1.5079, + "theoretical_loss": 3.5932247071061783, + "tokens_seen": 1286340608 + }, + { + "epoch": 0.39, + "learning_rate": 0.0004931434302908726, + "loss": 1.3752, + "theoretical_loss": 3.593208218921241, + "tokens_seen": 1286406144 + }, + { + "epoch": 0.39, + "learning_rate": 0.0004931273821464394, + "loss": 1.288, + "theoretical_loss": 3.593191731811457, + "tokens_seen": 1286471680 + }, + { + "epoch": 0.39, + "learning_rate": 0.000493111334002006, + "loss": 1.444, + "theoretical_loss": 3.5931752457767008, + "tokens_seen": 1286537216 + }, + { + "epoch": 0.39, + "learning_rate": 0.0004930952858575728, + "loss": 1.4336, + "theoretical_loss": 3.5931587608168494, + "tokens_seen": 1286602752 + }, + { + "epoch": 0.39, + "learning_rate": 0.0004930792377131394, + "loss": 1.3809, + "theoretical_loss": 3.5931422769317765, + "tokens_seen": 1286668288 + }, + { + "epoch": 0.39, + "learning_rate": 0.0004930631895687062, + "loss": 1.3846, + "theoretical_loss": 3.593125794121358, + "tokens_seen": 1286733824 + }, + { + "epoch": 0.39, + "learning_rate": 0.0004930471414242728, + "loss": 1.4274, + "theoretical_loss": 3.5931093123854687, + "tokens_seen": 1286799360 + }, + { + "epoch": 0.39, + "learning_rate": 0.0004930310932798396, + "loss": 1.4299, + "theoretical_loss": 3.5930928317239843, + "tokens_seen": 1286864896 + }, + { + "epoch": 0.39, + "learning_rate": 0.0004930150451354062, + "loss": 1.5308, + "theoretical_loss": 3.5930763521367797, + "tokens_seen": 1286930432 + }, + { + "epoch": 0.39, + "learning_rate": 0.000492998996990973, + "loss": 1.4112, + "theoretical_loss": 3.59305987362373, + "tokens_seen": 1286995968 + }, + { + "epoch": 0.39, + "learning_rate": 0.0004929829488465396, + "loss": 1.3079, + "theoretical_loss": 3.5930433961847106, + "tokens_seen": 1287061504 + }, + { + "epoch": 0.39, + "learning_rate": 0.0004929669007021064, + "loss": 1.4617, + "theoretical_loss": 3.5930269198195974, + "tokens_seen": 1287127040 + }, + { + "epoch": 0.39, + "learning_rate": 0.000492950852557673, + "loss": 1.3035, + "theoretical_loss": 3.593010444528265, + "tokens_seen": 1287192576 + }, + { + "epoch": 0.39, + "learning_rate": 0.0004929348044132398, + "loss": 1.4032, + "theoretical_loss": 3.592993970310589, + "tokens_seen": 1287258112 + }, + { + "epoch": 0.39, + "learning_rate": 0.0004929187562688064, + "loss": 1.4428, + "theoretical_loss": 3.5929774971664448, + "tokens_seen": 1287323648 + }, + { + "epoch": 0.39, + "learning_rate": 0.0004929027081243732, + "loss": 1.4861, + "theoretical_loss": 3.5929610250957076, + "tokens_seen": 1287389184 + }, + { + "epoch": 0.39, + "learning_rate": 0.0004928866599799398, + "loss": 1.4814, + "theoretical_loss": 3.5929445540982536, + "tokens_seen": 1287454720 + }, + { + "epoch": 0.39, + "learning_rate": 0.0004928706118355066, + "loss": 1.4542, + "theoretical_loss": 3.5929280841739573, + "tokens_seen": 1287520256 + }, + { + "epoch": 0.39, + "learning_rate": 0.0004928545636910732, + "loss": 1.2929, + "theoretical_loss": 3.592911615322694, + "tokens_seen": 1287585792 + }, + { + "epoch": 0.39, + "learning_rate": 0.00049283851554664, + "loss": 1.3816, + "theoretical_loss": 3.5928951475443403, + "tokens_seen": 1287651328 + }, + { + "epoch": 0.39, + "learning_rate": 0.0004928224674022066, + "loss": 1.3141, + "theoretical_loss": 3.592878680838771, + "tokens_seen": 1287716864 + }, + { + "epoch": 0.39, + "objective/train/docs_used": 471101, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.3303136825561523, + "objective/train/theoretical_loss": 3.592862215205862, + "objective/train/tokens_used": 1308242400, + "theoretical_loss": 3.592862215205862, + "tokens_seen": 1287782400 + }, + { + "epoch": 0.39, + "learning_rate": 0.0004928064192577734, + "loss": 1.4206, + "theoretical_loss": 3.592862215205862, + "tokens_seen": 1287782400 + }, + { + "epoch": 0.39, + "learning_rate": 0.00049279037111334, + "loss": 1.3392, + "theoretical_loss": 3.592845750645488, + "tokens_seen": 1287847936 + }, + { + "epoch": 0.39, + "learning_rate": 0.0004927743229689068, + "loss": 1.3937, + "theoretical_loss": 3.5928292871575254, + "tokens_seen": 1287913472 + }, + { + "epoch": 0.39, + "learning_rate": 0.0004927582748244734, + "loss": 1.3749, + "theoretical_loss": 3.5928128247418494, + "tokens_seen": 1287979008 + }, + { + "epoch": 0.39, + "learning_rate": 0.0004927422266800402, + "loss": 1.3924, + "theoretical_loss": 3.5927963633983357, + "tokens_seen": 1288044544 + }, + { + "epoch": 0.39, + "learning_rate": 0.0004927261785356068, + "loss": 1.4872, + "theoretical_loss": 3.59277990312686, + "tokens_seen": 1288110080 + }, + { + "epoch": 0.39, + "learning_rate": 0.0004927101303911736, + "loss": 1.2995, + "theoretical_loss": 3.592763443927298, + "tokens_seen": 1288175616 + }, + { + "epoch": 0.39, + "learning_rate": 0.0004926940822467402, + "loss": 1.3215, + "theoretical_loss": 3.592746985799525, + "tokens_seen": 1288241152 + }, + { + "epoch": 0.39, + "learning_rate": 0.000492678034102307, + "loss": 1.4896, + "theoretical_loss": 3.5927305287434175, + "tokens_seen": 1288306688 + }, + { + "epoch": 0.39, + "learning_rate": 0.0004926619859578736, + "loss": 1.4039, + "theoretical_loss": 3.59271407275885, + "tokens_seen": 1288372224 + }, + { + "epoch": 0.39, + "learning_rate": 0.0004926459378134404, + "loss": 1.3759, + "theoretical_loss": 3.5926976178456993, + "tokens_seen": 1288437760 + }, + { + "epoch": 0.39, + "learning_rate": 0.000492629889669007, + "loss": 1.3, + "theoretical_loss": 3.5926811640038405, + "tokens_seen": 1288503296 + }, + { + "epoch": 0.39, + "learning_rate": 0.0004926138415245738, + "loss": 1.3699, + "theoretical_loss": 3.5926647112331493, + "tokens_seen": 1288568832 + }, + { + "epoch": 0.39, + "learning_rate": 0.0004925977933801404, + "loss": 1.3486, + "theoretical_loss": 3.5926482595335023, + "tokens_seen": 1288634368 + }, + { + "epoch": 0.39, + "learning_rate": 0.0004925817452357071, + "loss": 1.381, + "theoretical_loss": 3.5926318089047746, + "tokens_seen": 1288699904 + }, + { + "epoch": 0.39, + "learning_rate": 0.0004925656970912738, + "loss": 1.3563, + "theoretical_loss": 3.5926153593468424, + "tokens_seen": 1288765440 + }, + { + "epoch": 0.39, + "learning_rate": 0.0004925496489468405, + "loss": 1.3916, + "theoretical_loss": 3.5925989108595813, + "tokens_seen": 1288830976 + }, + { + "epoch": 0.39, + "learning_rate": 0.0004925336008024073, + "loss": 1.3588, + "theoretical_loss": 3.5925824634428674, + "tokens_seen": 1288896512 + }, + { + "epoch": 0.39, + "learning_rate": 0.0004925175526579739, + "loss": 1.4231, + "theoretical_loss": 3.5925660170965763, + "tokens_seen": 1288962048 + }, + { + "epoch": 0.39, + "learning_rate": 0.0004925015045135407, + "loss": 1.3304, + "theoretical_loss": 3.5925495718205838, + "tokens_seen": 1289027584 + }, + { + "epoch": 0.39, + "learning_rate": 0.0004924854563691073, + "loss": 1.2517, + "theoretical_loss": 3.5925331276147663, + "tokens_seen": 1289093120 + }, + { + "epoch": 0.39, + "learning_rate": 0.000492469408224674, + "loss": 1.3464, + "theoretical_loss": 3.5925166844789995, + "tokens_seen": 1289158656 + }, + { + "epoch": 0.39, + "learning_rate": 0.0004924533600802407, + "loss": 1.489, + "theoretical_loss": 3.59250024241316, + "tokens_seen": 1289224192 + }, + { + "epoch": 0.39, + "learning_rate": 0.0004924373119358075, + "loss": 1.4126, + "theoretical_loss": 3.5924838014171225, + "tokens_seen": 1289289728 + }, + { + "epoch": 0.39, + "learning_rate": 0.0004924212637913741, + "loss": 1.3618, + "theoretical_loss": 3.592467361490764, + "tokens_seen": 1289355264 + }, + { + "epoch": 0.39, + "learning_rate": 0.0004924052156469409, + "loss": 1.5252, + "theoretical_loss": 3.5924509226339607, + "tokens_seen": 1289420800 + }, + { + "epoch": 0.39, + "learning_rate": 0.0004923891675025075, + "loss": 1.4568, + "theoretical_loss": 3.5924344848465877, + "tokens_seen": 1289486336 + }, + { + "epoch": 0.39, + "learning_rate": 0.0004923731193580743, + "loss": 1.3319, + "theoretical_loss": 3.592418048128522, + "tokens_seen": 1289551872 + }, + { + "epoch": 0.39, + "learning_rate": 0.0004923570712136409, + "loss": 1.3906, + "theoretical_loss": 3.592401612479639, + "tokens_seen": 1289617408 + }, + { + "epoch": 0.39, + "learning_rate": 0.0004923410230692077, + "loss": 1.375, + "theoretical_loss": 3.5923851778998155, + "tokens_seen": 1289682944 + }, + { + "epoch": 0.39, + "learning_rate": 0.0004923249749247743, + "loss": 1.3503, + "theoretical_loss": 3.5923687443889274, + "tokens_seen": 1289748480 + }, + { + "epoch": 0.39, + "learning_rate": 0.0004923089267803411, + "loss": 1.3731, + "theoretical_loss": 3.5923523119468506, + "tokens_seen": 1289814016 + }, + { + "epoch": 0.39, + "learning_rate": 0.0004922928786359077, + "loss": 1.3125, + "theoretical_loss": 3.5923358805734615, + "tokens_seen": 1289879552 + }, + { + "epoch": 0.39, + "learning_rate": 0.0004922768304914745, + "loss": 1.3724, + "theoretical_loss": 3.5923194502686364, + "tokens_seen": 1289945088 + }, + { + "epoch": 0.39, + "learning_rate": 0.0004922607823470411, + "loss": 1.397, + "theoretical_loss": 3.592303021032251, + "tokens_seen": 1290010624 + }, + { + "epoch": 0.39, + "learning_rate": 0.0004922447342026079, + "loss": 1.4305, + "theoretical_loss": 3.5922865928641823, + "tokens_seen": 1290076160 + }, + { + "epoch": 0.39, + "learning_rate": 0.0004922286860581745, + "loss": 1.4939, + "theoretical_loss": 3.592270165764306, + "tokens_seen": 1290141696 + }, + { + "epoch": 0.39, + "learning_rate": 0.0004922126379137413, + "loss": 1.3405, + "theoretical_loss": 3.5922537397324987, + "tokens_seen": 1290207232 + }, + { + "epoch": 0.39, + "learning_rate": 0.0004921965897693079, + "loss": 1.4554, + "theoretical_loss": 3.592237314768637, + "tokens_seen": 1290272768 + }, + { + "epoch": 0.39, + "learning_rate": 0.0004921805416248747, + "loss": 1.2673, + "theoretical_loss": 3.592220890872596, + "tokens_seen": 1290338304 + }, + { + "epoch": 0.39, + "learning_rate": 0.0004921644934804413, + "loss": 1.431, + "theoretical_loss": 3.5922044680442533, + "tokens_seen": 1290403840 + }, + { + "epoch": 0.39, + "learning_rate": 0.0004921484453360081, + "loss": 1.3789, + "theoretical_loss": 3.5921880462834848, + "tokens_seen": 1290469376 + }, + { + "epoch": 0.39, + "learning_rate": 0.0004921323971915747, + "loss": 1.4305, + "theoretical_loss": 3.592171625590167, + "tokens_seen": 1290534912 + }, + { + "epoch": 0.39, + "learning_rate": 0.0004921163490471415, + "loss": 1.3665, + "theoretical_loss": 3.5921552059641764, + "tokens_seen": 1290600448 + }, + { + "epoch": 0.39, + "learning_rate": 0.0004921003009027081, + "loss": 1.2676, + "theoretical_loss": 3.592138787405389, + "tokens_seen": 1290665984 + }, + { + "epoch": 0.39, + "learning_rate": 0.0004920842527582749, + "loss": 1.3622, + "theoretical_loss": 3.5921223699136817, + "tokens_seen": 1290731520 + }, + { + "epoch": 0.39, + "learning_rate": 0.0004920682046138415, + "loss": 1.3389, + "theoretical_loss": 3.592105953488931, + "tokens_seen": 1290797056 + }, + { + "epoch": 0.39, + "learning_rate": 0.0004920521564694083, + "loss": 1.2721, + "theoretical_loss": 3.592089538131013, + "tokens_seen": 1290862592 + }, + { + "epoch": 0.39, + "learning_rate": 0.0004920361083249749, + "loss": 1.5544, + "theoretical_loss": 3.5920731238398043, + "tokens_seen": 1290928128 + }, + { + "epoch": 0.39, + "learning_rate": 0.0004920200601805417, + "loss": 1.387, + "theoretical_loss": 3.5920567106151817, + "tokens_seen": 1290993664 + }, + { + "epoch": 0.39, + "objective/train/docs_used": 472355, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.4911893606185913, + "objective/train/theoretical_loss": 3.5920402984570217, + "objective/train/tokens_used": 1311519200, + "theoretical_loss": 3.5920402984570217, + "tokens_seen": 1291059200 + }, + { + "epoch": 0.39, + "learning_rate": 0.0004920040120361083, + "loss": 1.4208, + "theoretical_loss": 3.5920402984570217, + "tokens_seen": 1291059200 + }, + { + "epoch": 0.39, + "learning_rate": 0.0004919879638916751, + "loss": 1.4314, + "theoretical_loss": 3.5920238873652006, + "tokens_seen": 1291124736 + }, + { + "epoch": 0.39, + "learning_rate": 0.0004919719157472417, + "loss": 1.3107, + "theoretical_loss": 3.592007477339595, + "tokens_seen": 1291190272 + }, + { + "epoch": 0.39, + "learning_rate": 0.0004919558676028085, + "loss": 1.3977, + "theoretical_loss": 3.591991068380082, + "tokens_seen": 1291255808 + }, + { + "epoch": 0.39, + "learning_rate": 0.0004919398194583751, + "loss": 1.4285, + "theoretical_loss": 3.591974660486538, + "tokens_seen": 1291321344 + }, + { + "epoch": 0.39, + "learning_rate": 0.0004919237713139419, + "loss": 1.41, + "theoretical_loss": 3.5919582536588397, + "tokens_seen": 1291386880 + }, + { + "epoch": 0.39, + "learning_rate": 0.0004919077231695085, + "loss": 1.3566, + "theoretical_loss": 3.591941847896863, + "tokens_seen": 1291452416 + }, + { + "epoch": 0.39, + "learning_rate": 0.0004918916750250753, + "loss": 1.3579, + "theoretical_loss": 3.591925443200486, + "tokens_seen": 1291517952 + }, + { + "epoch": 0.39, + "learning_rate": 0.0004918756268806419, + "loss": 1.4464, + "theoretical_loss": 3.5919090395695843, + "tokens_seen": 1291583488 + }, + { + "epoch": 0.39, + "learning_rate": 0.0004918595787362087, + "loss": 1.4434, + "theoretical_loss": 3.5918926370040354, + "tokens_seen": 1291649024 + }, + { + "epoch": 0.39, + "learning_rate": 0.0004918435305917753, + "loss": 1.4189, + "theoretical_loss": 3.5918762355037153, + "tokens_seen": 1291714560 + }, + { + "epoch": 0.39, + "learning_rate": 0.0004918274824473421, + "loss": 1.4186, + "theoretical_loss": 3.5918598350685014, + "tokens_seen": 1291780096 + }, + { + "epoch": 0.39, + "learning_rate": 0.0004918114343029087, + "loss": 1.2937, + "theoretical_loss": 3.5918434356982702, + "tokens_seen": 1291845632 + }, + { + "epoch": 0.39, + "learning_rate": 0.0004917953861584755, + "loss": 1.3894, + "theoretical_loss": 3.591827037392899, + "tokens_seen": 1291911168 + }, + { + "epoch": 0.39, + "learning_rate": 0.0004917793380140421, + "loss": 1.397, + "theoretical_loss": 3.5918106401522634, + "tokens_seen": 1291976704 + }, + { + "epoch": 0.39, + "learning_rate": 0.0004917632898696089, + "loss": 1.2735, + "theoretical_loss": 3.5917942439762416, + "tokens_seen": 1292042240 + }, + { + "epoch": 0.39, + "learning_rate": 0.0004917472417251755, + "loss": 1.3697, + "theoretical_loss": 3.59177784886471, + "tokens_seen": 1292107776 + }, + { + "epoch": 0.39, + "learning_rate": 0.0004917311935807423, + "loss": 1.5049, + "theoretical_loss": 3.5917614548175454, + "tokens_seen": 1292173312 + }, + { + "epoch": 0.39, + "learning_rate": 0.0004917151454363089, + "loss": 1.412, + "theoretical_loss": 3.591745061834625, + "tokens_seen": 1292238848 + }, + { + "epoch": 0.39, + "learning_rate": 0.0004916990972918757, + "loss": 1.4909, + "theoretical_loss": 3.5917286699158253, + "tokens_seen": 1292304384 + }, + { + "epoch": 0.39, + "learning_rate": 0.0004916830491474423, + "loss": 1.4614, + "theoretical_loss": 3.5917122790610234, + "tokens_seen": 1292369920 + }, + { + "epoch": 0.39, + "learning_rate": 0.0004916670010030091, + "loss": 1.4133, + "theoretical_loss": 3.5916958892700968, + "tokens_seen": 1292435456 + }, + { + "epoch": 0.39, + "learning_rate": 0.0004916509528585757, + "loss": 1.4459, + "theoretical_loss": 3.591679500542922, + "tokens_seen": 1292500992 + }, + { + "epoch": 0.39, + "learning_rate": 0.0004916349047141425, + "loss": 1.4166, + "theoretical_loss": 3.5916631128793757, + "tokens_seen": 1292566528 + }, + { + "epoch": 0.39, + "learning_rate": 0.0004916188565697091, + "loss": 1.3335, + "theoretical_loss": 3.591646726279336, + "tokens_seen": 1292632064 + }, + { + "epoch": 0.39, + "learning_rate": 0.0004916028084252759, + "loss": 1.4079, + "theoretical_loss": 3.5916303407426797, + "tokens_seen": 1292697600 + }, + { + "epoch": 0.39, + "learning_rate": 0.0004915867602808425, + "loss": 1.402, + "theoretical_loss": 3.5916139562692826, + "tokens_seen": 1292763136 + }, + { + "epoch": 0.39, + "learning_rate": 0.0004915707121364093, + "loss": 1.2657, + "theoretical_loss": 3.5915975728590235, + "tokens_seen": 1292828672 + }, + { + "epoch": 0.39, + "learning_rate": 0.0004915546639919759, + "loss": 1.4346, + "theoretical_loss": 3.5915811905117785, + "tokens_seen": 1292894208 + }, + { + "epoch": 0.39, + "learning_rate": 0.0004915386158475427, + "loss": 1.4832, + "theoretical_loss": 3.5915648092274246, + "tokens_seen": 1292959744 + }, + { + "epoch": 0.39, + "learning_rate": 0.0004915225677031093, + "loss": 1.4174, + "theoretical_loss": 3.59154842900584, + "tokens_seen": 1293025280 + }, + { + "epoch": 0.39, + "learning_rate": 0.0004915065195586761, + "loss": 1.3527, + "theoretical_loss": 3.5915320498469008, + "tokens_seen": 1293090816 + }, + { + "epoch": 0.39, + "learning_rate": 0.0004914904714142427, + "loss": 1.3689, + "theoretical_loss": 3.5915156717504853, + "tokens_seen": 1293156352 + }, + { + "epoch": 0.39, + "learning_rate": 0.0004914744232698095, + "loss": 1.335, + "theoretical_loss": 3.5914992947164697, + "tokens_seen": 1293221888 + }, + { + "epoch": 0.39, + "learning_rate": 0.0004914583751253761, + "loss": 1.3296, + "theoretical_loss": 3.591482918744732, + "tokens_seen": 1293287424 + }, + { + "epoch": 0.39, + "learning_rate": 0.0004914423269809429, + "loss": 1.3537, + "theoretical_loss": 3.591466543835149, + "tokens_seen": 1293352960 + }, + { + "epoch": 0.39, + "learning_rate": 0.0004914262788365095, + "loss": 1.4365, + "theoretical_loss": 3.5914501699875983, + "tokens_seen": 1293418496 + }, + { + "epoch": 0.39, + "learning_rate": 0.0004914102306920763, + "loss": 1.324, + "theoretical_loss": 3.5914337972019563, + "tokens_seen": 1293484032 + }, + { + "epoch": 0.39, + "learning_rate": 0.0004913941825476429, + "loss": 1.421, + "theoretical_loss": 3.591417425478102, + "tokens_seen": 1293549568 + }, + { + "epoch": 0.39, + "learning_rate": 0.0004913781344032097, + "loss": 1.4529, + "theoretical_loss": 3.591401054815911, + "tokens_seen": 1293615104 + }, + { + "epoch": 0.39, + "learning_rate": 0.0004913620862587763, + "loss": 1.3871, + "theoretical_loss": 3.591384685215263, + "tokens_seen": 1293680640 + }, + { + "epoch": 0.39, + "learning_rate": 0.0004913460381143431, + "loss": 1.4194, + "theoretical_loss": 3.5913683166760326, + "tokens_seen": 1293746176 + }, + { + "epoch": 0.39, + "learning_rate": 0.0004913299899699097, + "loss": 1.4474, + "theoretical_loss": 3.591351949198099, + "tokens_seen": 1293811712 + }, + { + "epoch": 0.39, + "learning_rate": 0.0004913139418254765, + "loss": 1.3741, + "theoretical_loss": 3.591335582781339, + "tokens_seen": 1293877248 + }, + { + "epoch": 0.39, + "learning_rate": 0.0004912978936810431, + "loss": 1.5793, + "theoretical_loss": 3.5913192174256303, + "tokens_seen": 1293942784 + }, + { + "epoch": 0.39, + "learning_rate": 0.0004912818455366099, + "loss": 1.3497, + "theoretical_loss": 3.5913028531308506, + "tokens_seen": 1294008320 + }, + { + "epoch": 0.39, + "learning_rate": 0.0004912657973921765, + "loss": 1.4657, + "theoretical_loss": 3.591286489896877, + "tokens_seen": 1294073856 + }, + { + "epoch": 0.39, + "learning_rate": 0.0004912497492477433, + "loss": 1.4545, + "theoretical_loss": 3.591270127723587, + "tokens_seen": 1294139392 + }, + { + "epoch": 0.39, + "learning_rate": 0.00049123370110331, + "loss": 1.3826, + "theoretical_loss": 3.5912537666108584, + "tokens_seen": 1294204928 + }, + { + "epoch": 0.39, + "learning_rate": 0.0004912176529588767, + "loss": 1.4873, + "theoretical_loss": 3.591237406558568, + "tokens_seen": 1294270464 + }, + { + "epoch": 0.39, + "objective/train/docs_used": 473604, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.4100412130355835, + "objective/train/theoretical_loss": 3.591221047566594, + "objective/train/tokens_used": 1314796000, + "theoretical_loss": 3.591221047566594, + "tokens_seen": 1294336000 + }, + { + "epoch": 0.39, + "learning_rate": 0.0004912016048144433, + "loss": 1.3614, + "theoretical_loss": 3.591221047566594, + "tokens_seen": 1294336000 + }, + { + "epoch": 0.39, + "learning_rate": 0.0004911855566700101, + "loss": 1.3739, + "theoretical_loss": 3.591204689634815, + "tokens_seen": 1294401536 + }, + { + "epoch": 0.39, + "learning_rate": 0.0004911695085255767, + "loss": 1.4978, + "theoretical_loss": 3.5911883327631067, + "tokens_seen": 1294467072 + }, + { + "epoch": 0.39, + "learning_rate": 0.0004911534603811434, + "loss": 1.386, + "theoretical_loss": 3.5911719769513484, + "tokens_seen": 1294532608 + }, + { + "epoch": 0.39, + "learning_rate": 0.0004911374122367102, + "loss": 1.3772, + "theoretical_loss": 3.591155622199416, + "tokens_seen": 1294598144 + }, + { + "epoch": 0.39, + "learning_rate": 0.0004911213640922768, + "loss": 1.356, + "theoretical_loss": 3.591139268507189, + "tokens_seen": 1294663680 + }, + { + "epoch": 0.39, + "learning_rate": 0.0004911053159478436, + "loss": 1.458, + "theoretical_loss": 3.5911229158745437, + "tokens_seen": 1294729216 + }, + { + "epoch": 0.39, + "learning_rate": 0.0004910892678034102, + "loss": 1.3353, + "theoretical_loss": 3.5911065643013593, + "tokens_seen": 1294794752 + }, + { + "epoch": 0.39, + "learning_rate": 0.000491073219658977, + "loss": 1.3975, + "theoretical_loss": 3.5910902137875116, + "tokens_seen": 1294860288 + }, + { + "epoch": 0.39, + "learning_rate": 0.0004910571715145436, + "loss": 1.3632, + "theoretical_loss": 3.59107386433288, + "tokens_seen": 1294925824 + }, + { + "epoch": 0.39, + "learning_rate": 0.0004910411233701104, + "loss": 1.3665, + "theoretical_loss": 3.591057515937342, + "tokens_seen": 1294991360 + }, + { + "epoch": 0.39, + "learning_rate": 0.000491025075225677, + "loss": 1.5592, + "theoretical_loss": 3.591041168600774, + "tokens_seen": 1295056896 + }, + { + "epoch": 0.39, + "learning_rate": 0.0004910090270812438, + "loss": 1.4062, + "theoretical_loss": 3.591024822323056, + "tokens_seen": 1295122432 + }, + { + "epoch": 0.39, + "learning_rate": 0.0004909929789368104, + "loss": 1.4356, + "theoretical_loss": 3.5910084771040642, + "tokens_seen": 1295187968 + }, + { + "epoch": 0.39, + "learning_rate": 0.0004909769307923772, + "loss": 1.3748, + "theoretical_loss": 3.590992132943677, + "tokens_seen": 1295253504 + }, + { + "epoch": 0.39, + "learning_rate": 0.0004909608826479438, + "loss": 1.5161, + "theoretical_loss": 3.5909757898417722, + "tokens_seen": 1295319040 + }, + { + "epoch": 0.39, + "learning_rate": 0.0004909448345035106, + "loss": 1.3846, + "theoretical_loss": 3.590959447798228, + "tokens_seen": 1295384576 + }, + { + "epoch": 0.39, + "learning_rate": 0.0004909287863590772, + "loss": 1.4625, + "theoretical_loss": 3.5909431068129223, + "tokens_seen": 1295450112 + }, + { + "epoch": 0.39, + "learning_rate": 0.000490912738214644, + "loss": 1.4131, + "theoretical_loss": 3.5909267668857323, + "tokens_seen": 1295515648 + }, + { + "epoch": 0.39, + "learning_rate": 0.0004908966900702106, + "loss": 1.4498, + "theoretical_loss": 3.590910428016537, + "tokens_seen": 1295581184 + }, + { + "epoch": 0.39, + "learning_rate": 0.0004908806419257774, + "loss": 1.4355, + "theoretical_loss": 3.5908940902052136, + "tokens_seen": 1295646720 + }, + { + "epoch": 0.39, + "learning_rate": 0.000490864593781344, + "loss": 1.5019, + "theoretical_loss": 3.5908777534516405, + "tokens_seen": 1295712256 + }, + { + "epoch": 0.39, + "learning_rate": 0.0004908485456369108, + "loss": 1.3912, + "theoretical_loss": 3.5908614177556957, + "tokens_seen": 1295777792 + }, + { + "epoch": 0.39, + "learning_rate": 0.0004908324974924774, + "loss": 1.2836, + "theoretical_loss": 3.590845083117257, + "tokens_seen": 1295843328 + }, + { + "epoch": 0.39, + "learning_rate": 0.0004908164493480442, + "loss": 1.4411, + "theoretical_loss": 3.590828749536203, + "tokens_seen": 1295908864 + }, + { + "epoch": 0.39, + "learning_rate": 0.0004908004012036108, + "loss": 1.3833, + "theoretical_loss": 3.590812417012411, + "tokens_seen": 1295974400 + }, + { + "epoch": 0.39, + "learning_rate": 0.0004907843530591776, + "loss": 1.4632, + "theoretical_loss": 3.59079608554576, + "tokens_seen": 1296039936 + }, + { + "epoch": 0.39, + "learning_rate": 0.0004907683049147442, + "loss": 1.3053, + "theoretical_loss": 3.5907797551361273, + "tokens_seen": 1296105472 + }, + { + "epoch": 0.39, + "learning_rate": 0.000490752256770311, + "loss": 1.365, + "theoretical_loss": 3.5907634257833916, + "tokens_seen": 1296171008 + }, + { + "epoch": 0.39, + "learning_rate": 0.0004907362086258776, + "loss": 1.3819, + "theoretical_loss": 3.5907470974874305, + "tokens_seen": 1296236544 + }, + { + "epoch": 0.39, + "learning_rate": 0.0004907201604814444, + "loss": 1.3935, + "theoretical_loss": 3.590730770248123, + "tokens_seen": 1296302080 + }, + { + "epoch": 0.39, + "learning_rate": 0.000490704112337011, + "loss": 1.4097, + "theoretical_loss": 3.5907144440653465, + "tokens_seen": 1296367616 + }, + { + "epoch": 0.39, + "learning_rate": 0.0004906880641925778, + "loss": 1.4375, + "theoretical_loss": 3.5906981189389793, + "tokens_seen": 1296433152 + }, + { + "epoch": 0.39, + "learning_rate": 0.0004906720160481444, + "loss": 1.4295, + "theoretical_loss": 3.5906817948689005, + "tokens_seen": 1296498688 + }, + { + "epoch": 0.39, + "learning_rate": 0.0004906559679037112, + "loss": 1.4917, + "theoretical_loss": 3.5906654718549875, + "tokens_seen": 1296564224 + }, + { + "epoch": 0.39, + "learning_rate": 0.0004906399197592778, + "loss": 1.4303, + "theoretical_loss": 3.590649149897119, + "tokens_seen": 1296629760 + }, + { + "epoch": 0.39, + "learning_rate": 0.0004906238716148446, + "loss": 1.377, + "theoretical_loss": 3.590632828995173, + "tokens_seen": 1296695296 + }, + { + "epoch": 0.39, + "learning_rate": 0.0004906078234704112, + "loss": 1.388, + "theoretical_loss": 3.5906165091490285, + "tokens_seen": 1296760832 + }, + { + "epoch": 0.39, + "learning_rate": 0.000490591775325978, + "loss": 1.3445, + "theoretical_loss": 3.590600190358563, + "tokens_seen": 1296826368 + }, + { + "epoch": 0.39, + "learning_rate": 0.0004905757271815446, + "loss": 1.3919, + "theoretical_loss": 3.5905838726236556, + "tokens_seen": 1296891904 + }, + { + "epoch": 0.39, + "learning_rate": 0.0004905596790371114, + "loss": 1.3841, + "theoretical_loss": 3.590567555944183, + "tokens_seen": 1296957440 + }, + { + "epoch": 0.39, + "learning_rate": 0.000490543630892678, + "loss": 1.3318, + "theoretical_loss": 3.5905512403200266, + "tokens_seen": 1297022976 + }, + { + "epoch": 0.39, + "learning_rate": 0.0004905275827482448, + "loss": 1.3878, + "theoretical_loss": 3.5905349257510624, + "tokens_seen": 1297088512 + }, + { + "epoch": 0.39, + "learning_rate": 0.0004905115346038114, + "loss": 1.4073, + "theoretical_loss": 3.5905186122371697, + "tokens_seen": 1297154048 + }, + { + "epoch": 0.39, + "learning_rate": 0.0004904954864593782, + "loss": 1.4067, + "theoretical_loss": 3.5905022997782265, + "tokens_seen": 1297219584 + }, + { + "epoch": 0.39, + "learning_rate": 0.0004904794383149448, + "loss": 1.3682, + "theoretical_loss": 3.5904859883741125, + "tokens_seen": 1297285120 + }, + { + "epoch": 0.39, + "learning_rate": 0.0004904633901705116, + "loss": 1.3741, + "theoretical_loss": 3.5904696780247045, + "tokens_seen": 1297350656 + }, + { + "epoch": 0.39, + "learning_rate": 0.0004904473420260782, + "loss": 1.362, + "theoretical_loss": 3.5904533687298823, + "tokens_seen": 1297416192 + }, + { + "epoch": 0.39, + "learning_rate": 0.000490431293881645, + "loss": 1.3784, + "theoretical_loss": 3.5904370604895237, + "tokens_seen": 1297481728 + }, + { + "epoch": 0.39, + "learning_rate": 0.0004904152457372116, + "loss": 1.349, + "theoretical_loss": 3.590420753303508, + "tokens_seen": 1297547264 + }, + { + "epoch": 0.39, + "objective/train/docs_used": 474727, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.3544045686721802, + "objective/train/theoretical_loss": 3.5904044471717134, + "objective/train/tokens_used": 1318072800, + "theoretical_loss": 3.5904044471717134, + "tokens_seen": 1297612800 + }, + { + "epoch": 0.39, + "learning_rate": 0.0004903991975927784, + "loss": 1.3991, + "theoretical_loss": 3.5904044471717134, + "tokens_seen": 1297612800 + }, + { + "epoch": 0.39, + "learning_rate": 0.000490383149448345, + "loss": 1.3524, + "theoretical_loss": 3.5903881420940182, + "tokens_seen": 1297678336 + }, + { + "epoch": 0.39, + "learning_rate": 0.0004903671013039118, + "loss": 1.2971, + "theoretical_loss": 3.5903718380703014, + "tokens_seen": 1297743872 + }, + { + "epoch": 0.39, + "learning_rate": 0.0004903510531594784, + "loss": 1.3783, + "theoretical_loss": 3.5903555351004415, + "tokens_seen": 1297809408 + }, + { + "epoch": 0.39, + "learning_rate": 0.0004903350050150452, + "loss": 1.2947, + "theoretical_loss": 3.5903392331843174, + "tokens_seen": 1297874944 + }, + { + "epoch": 0.39, + "learning_rate": 0.0004903189568706118, + "loss": 1.2848, + "theoretical_loss": 3.5903229323218078, + "tokens_seen": 1297940480 + }, + { + "epoch": 0.39, + "learning_rate": 0.0004903029087261786, + "loss": 1.3131, + "theoretical_loss": 3.5903066325127906, + "tokens_seen": 1298006016 + }, + { + "epoch": 0.39, + "learning_rate": 0.0004902868605817452, + "loss": 1.3604, + "theoretical_loss": 3.590290333757146, + "tokens_seen": 1298071552 + }, + { + "epoch": 0.39, + "learning_rate": 0.000490270812437312, + "loss": 1.4098, + "theoretical_loss": 3.590274036054751, + "tokens_seen": 1298137088 + }, + { + "epoch": 0.39, + "learning_rate": 0.0004902547642928786, + "loss": 1.4739, + "theoretical_loss": 3.5902577394054864, + "tokens_seen": 1298202624 + }, + { + "epoch": 0.39, + "learning_rate": 0.0004902387161484454, + "loss": 1.366, + "theoretical_loss": 3.590241443809229, + "tokens_seen": 1298268160 + }, + { + "epoch": 0.39, + "learning_rate": 0.000490222668004012, + "loss": 1.4074, + "theoretical_loss": 3.590225149265859, + "tokens_seen": 1298333696 + }, + { + "epoch": 0.39, + "learning_rate": 0.0004902066198595788, + "loss": 1.2831, + "theoretical_loss": 3.5902088557752547, + "tokens_seen": 1298399232 + }, + { + "epoch": 0.39, + "learning_rate": 0.0004901905717151454, + "loss": 1.4874, + "theoretical_loss": 3.5901925633372946, + "tokens_seen": 1298464768 + }, + { + "epoch": 0.39, + "learning_rate": 0.0004901745235707122, + "loss": 1.2886, + "theoretical_loss": 3.590176271951858, + "tokens_seen": 1298530304 + }, + { + "epoch": 0.39, + "learning_rate": 0.0004901584754262788, + "loss": 1.3825, + "theoretical_loss": 3.5901599816188243, + "tokens_seen": 1298595840 + }, + { + "epoch": 0.39, + "learning_rate": 0.0004901424272818456, + "loss": 1.2997, + "theoretical_loss": 3.5901436923380716, + "tokens_seen": 1298661376 + }, + { + "epoch": 0.39, + "learning_rate": 0.0004901263791374122, + "loss": 1.2521, + "theoretical_loss": 3.5901274041094786, + "tokens_seen": 1298726912 + }, + { + "epoch": 0.39, + "learning_rate": 0.000490110330992979, + "loss": 1.5456, + "theoretical_loss": 3.590111116932925, + "tokens_seen": 1298792448 + }, + { + "epoch": 0.39, + "learning_rate": 0.0004900942828485456, + "loss": 1.3654, + "theoretical_loss": 3.5900948308082894, + "tokens_seen": 1298857984 + }, + { + "epoch": 0.39, + "learning_rate": 0.0004900782347041124, + "loss": 1.3226, + "theoretical_loss": 3.5900785457354516, + "tokens_seen": 1298923520 + }, + { + "epoch": 0.39, + "learning_rate": 0.0004900621865596791, + "loss": 1.4038, + "theoretical_loss": 3.590062261714289, + "tokens_seen": 1298989056 + }, + { + "epoch": 0.39, + "learning_rate": 0.0004900461384152458, + "loss": 1.315, + "theoretical_loss": 3.5900459787446817, + "tokens_seen": 1299054592 + }, + { + "epoch": 0.39, + "learning_rate": 0.0004900300902708125, + "loss": 1.3839, + "theoretical_loss": 3.590029696826509, + "tokens_seen": 1299120128 + }, + { + "epoch": 0.39, + "learning_rate": 0.0004900140421263792, + "loss": 1.3224, + "theoretical_loss": 3.5900134159596493, + "tokens_seen": 1299185664 + }, + { + "epoch": 0.39, + "learning_rate": 0.000489997993981946, + "loss": 1.3357, + "theoretical_loss": 3.5899971361439817, + "tokens_seen": 1299251200 + }, + { + "epoch": 0.39, + "learning_rate": 0.0004899819458375126, + "loss": 1.4794, + "theoretical_loss": 3.589980857379386, + "tokens_seen": 1299316736 + }, + { + "epoch": 0.39, + "learning_rate": 0.0004899658976930794, + "loss": 1.3393, + "theoretical_loss": 3.5899645796657405, + "tokens_seen": 1299382272 + }, + { + "epoch": 0.39, + "learning_rate": 0.000489949849548646, + "loss": 1.4272, + "theoretical_loss": 3.5899483030029247, + "tokens_seen": 1299447808 + }, + { + "epoch": 0.39, + "learning_rate": 0.0004899338014042128, + "loss": 1.4051, + "theoretical_loss": 3.5899320273908177, + "tokens_seen": 1299513344 + }, + { + "epoch": 0.39, + "learning_rate": 0.0004899177532597794, + "loss": 1.4696, + "theoretical_loss": 3.5899157528292993, + "tokens_seen": 1299578880 + }, + { + "epoch": 0.39, + "learning_rate": 0.000489901705115346, + "loss": 1.4087, + "theoretical_loss": 3.5898994793182477, + "tokens_seen": 1299644416 + }, + { + "epoch": 0.39, + "learning_rate": 0.0004898856569709128, + "loss": 1.3927, + "theoretical_loss": 3.589883206857543, + "tokens_seen": 1299709952 + }, + { + "epoch": 0.39, + "learning_rate": 0.0004898696088264794, + "loss": 1.2796, + "theoretical_loss": 3.5898669354470636, + "tokens_seen": 1299775488 + }, + { + "epoch": 0.39, + "learning_rate": 0.0004898535606820462, + "loss": 1.4185, + "theoretical_loss": 3.58985066508669, + "tokens_seen": 1299841024 + }, + { + "epoch": 0.39, + "learning_rate": 0.0004898375125376128, + "loss": 1.4629, + "theoretical_loss": 3.5898343957763004, + "tokens_seen": 1299906560 + }, + { + "epoch": 0.39, + "learning_rate": 0.0004898214643931796, + "loss": 1.4751, + "theoretical_loss": 3.589818127515774, + "tokens_seen": 1299972096 + }, + { + "epoch": 0.39, + "learning_rate": 0.0004898054162487462, + "loss": 1.3886, + "theoretical_loss": 3.5898018603049913, + "tokens_seen": 1300037632 + }, + { + "epoch": 0.39, + "learning_rate": 0.000489789368104313, + "loss": 1.3496, + "theoretical_loss": 3.58978559414383, + "tokens_seen": 1300103168 + }, + { + "epoch": 0.39, + "learning_rate": 0.0004897733199598797, + "loss": 1.3473, + "theoretical_loss": 3.589769329032171, + "tokens_seen": 1300168704 + }, + { + "epoch": 0.39, + "learning_rate": 0.0004897572718154463, + "loss": 1.397, + "theoretical_loss": 3.5897530649698934, + "tokens_seen": 1300234240 + }, + { + "epoch": 0.39, + "learning_rate": 0.000489741223671013, + "loss": 1.3725, + "theoretical_loss": 3.5897368019568754, + "tokens_seen": 1300299776 + }, + { + "epoch": 0.39, + "learning_rate": 0.0004897251755265797, + "loss": 1.4538, + "theoretical_loss": 3.589720539992998, + "tokens_seen": 1300365312 + }, + { + "epoch": 0.39, + "learning_rate": 0.0004897091273821465, + "loss": 1.3424, + "theoretical_loss": 3.58970427907814, + "tokens_seen": 1300430848 + }, + { + "epoch": 0.39, + "learning_rate": 0.0004896930792377131, + "loss": 1.4547, + "theoretical_loss": 3.58968801921218, + "tokens_seen": 1300496384 + }, + { + "epoch": 0.39, + "learning_rate": 0.0004896770310932799, + "loss": 1.3576, + "theoretical_loss": 3.5896717603949995, + "tokens_seen": 1300561920 + }, + { + "epoch": 0.39, + "learning_rate": 0.0004896609829488465, + "loss": 1.3923, + "theoretical_loss": 3.589655502626476, + "tokens_seen": 1300627456 + }, + { + "epoch": 0.39, + "learning_rate": 0.0004896449348044133, + "loss": 1.4866, + "theoretical_loss": 3.5896392459064903, + "tokens_seen": 1300692992 + }, + { + "epoch": 0.39, + "learning_rate": 0.0004896288866599799, + "loss": 1.3626, + "theoretical_loss": 3.5896229902349215, + "tokens_seen": 1300758528 + }, + { + "epoch": 0.39, + "learning_rate": 0.0004896128385155467, + "loss": 1.3798, + "theoretical_loss": 3.5896067356116492, + "tokens_seen": 1300824064 + }, + { + "epoch": 0.39, + "objective/train/docs_used": 475730, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.349228858947754, + "objective/train/theoretical_loss": 3.5895904820365527, + "objective/train/tokens_used": 1321349600, + "theoretical_loss": 3.5895904820365527, + "tokens_seen": 1300889600 + }, + { + "epoch": 0.39, + "learning_rate": 0.0004895967903711133, + "loss": 1.4267, + "theoretical_loss": 3.5895904820365527, + "tokens_seen": 1300889600 + }, + { + "epoch": 0.39, + "learning_rate": 0.0004895807422266801, + "loss": 1.3194, + "theoretical_loss": 3.589574229509512, + "tokens_seen": 1300955136 + }, + { + "epoch": 0.39, + "learning_rate": 0.0004895646940822467, + "loss": 1.4311, + "theoretical_loss": 3.589557978030407, + "tokens_seen": 1301020672 + }, + { + "epoch": 0.39, + "learning_rate": 0.0004895486459378135, + "loss": 1.3892, + "theoretical_loss": 3.5895417275991166, + "tokens_seen": 1301086208 + }, + { + "epoch": 0.39, + "learning_rate": 0.0004895325977933801, + "loss": 1.4767, + "theoretical_loss": 3.5895254782155206, + "tokens_seen": 1301151744 + }, + { + "epoch": 0.39, + "learning_rate": 0.0004895165496489469, + "loss": 1.3306, + "theoretical_loss": 3.589509229879499, + "tokens_seen": 1301217280 + }, + { + "epoch": 0.39, + "learning_rate": 0.0004895005015045135, + "loss": 1.3866, + "theoretical_loss": 3.589492982590932, + "tokens_seen": 1301282816 + }, + { + "epoch": 0.39, + "learning_rate": 0.0004894844533600803, + "loss": 1.3843, + "theoretical_loss": 3.5894767363496984, + "tokens_seen": 1301348352 + }, + { + "epoch": 0.39, + "learning_rate": 0.0004894684052156469, + "loss": 1.466, + "theoretical_loss": 3.5894604911556782, + "tokens_seen": 1301413888 + }, + { + "epoch": 0.39, + "learning_rate": 0.0004894523570712137, + "loss": 1.3866, + "theoretical_loss": 3.5894442470087515, + "tokens_seen": 1301479424 + }, + { + "epoch": 0.39, + "learning_rate": 0.0004894363089267803, + "loss": 1.3453, + "theoretical_loss": 3.5894280039087976, + "tokens_seen": 1301544960 + }, + { + "epoch": 0.39, + "learning_rate": 0.0004894202607823471, + "loss": 1.3136, + "theoretical_loss": 3.589411761855697, + "tokens_seen": 1301610496 + }, + { + "epoch": 0.39, + "learning_rate": 0.0004894042126379137, + "loss": 1.4504, + "theoretical_loss": 3.5893955208493287, + "tokens_seen": 1301676032 + }, + { + "epoch": 0.39, + "learning_rate": 0.0004893881644934805, + "loss": 1.3956, + "theoretical_loss": 3.589379280889573, + "tokens_seen": 1301741568 + }, + { + "epoch": 0.39, + "learning_rate": 0.0004893721163490471, + "loss": 1.2845, + "theoretical_loss": 3.589363041976309, + "tokens_seen": 1301807104 + }, + { + "epoch": 0.39, + "learning_rate": 0.0004893560682046139, + "loss": 1.4998, + "theoretical_loss": 3.5893468041094185, + "tokens_seen": 1301872640 + }, + { + "epoch": 0.39, + "learning_rate": 0.0004893400200601805, + "loss": 1.3528, + "theoretical_loss": 3.58933056728878, + "tokens_seen": 1301938176 + }, + { + "epoch": 0.39, + "learning_rate": 0.0004893239719157473, + "loss": 1.3606, + "theoretical_loss": 3.589314331514273, + "tokens_seen": 1302003712 + }, + { + "epoch": 0.39, + "learning_rate": 0.0004893079237713139, + "loss": 1.3432, + "theoretical_loss": 3.589298096785779, + "tokens_seen": 1302069248 + }, + { + "epoch": 0.39, + "learning_rate": 0.0004892918756268807, + "loss": 1.3767, + "theoretical_loss": 3.589281863103176, + "tokens_seen": 1302134784 + }, + { + "epoch": 0.39, + "learning_rate": 0.0004892758274824473, + "loss": 1.4411, + "theoretical_loss": 3.5892656304663455, + "tokens_seen": 1302200320 + }, + { + "epoch": 0.39, + "learning_rate": 0.0004892597793380141, + "loss": 1.4862, + "theoretical_loss": 3.5892493988751673, + "tokens_seen": 1302265856 + }, + { + "epoch": 0.39, + "learning_rate": 0.0004892437311935807, + "loss": 1.3385, + "theoretical_loss": 3.5892331683295207, + "tokens_seen": 1302331392 + }, + { + "epoch": 0.39, + "learning_rate": 0.0004892276830491475, + "loss": 1.335, + "theoretical_loss": 3.5892169388292867, + "tokens_seen": 1302396928 + }, + { + "epoch": 0.39, + "learning_rate": 0.0004892116349047141, + "loss": 1.3069, + "theoretical_loss": 3.589200710374344, + "tokens_seen": 1302462464 + }, + { + "epoch": 0.39, + "learning_rate": 0.0004891955867602809, + "loss": 1.4086, + "theoretical_loss": 3.5891844829645745, + "tokens_seen": 1302528000 + }, + { + "epoch": 0.39, + "learning_rate": 0.0004891795386158475, + "loss": 1.5171, + "theoretical_loss": 3.589168256599857, + "tokens_seen": 1302593536 + }, + { + "epoch": 0.39, + "learning_rate": 0.0004891634904714143, + "loss": 1.375, + "theoretical_loss": 3.5891520312800718, + "tokens_seen": 1302659072 + }, + { + "epoch": 0.39, + "learning_rate": 0.0004891474423269809, + "loss": 1.3754, + "theoretical_loss": 3.5891358070051, + "tokens_seen": 1302724608 + }, + { + "epoch": 0.39, + "learning_rate": 0.0004891313941825477, + "loss": 1.4104, + "theoretical_loss": 3.5891195837748198, + "tokens_seen": 1302790144 + }, + { + "epoch": 0.39, + "learning_rate": 0.0004891153460381143, + "loss": 1.452, + "theoretical_loss": 3.589103361589113, + "tokens_seen": 1302855680 + }, + { + "epoch": 0.39, + "learning_rate": 0.0004890992978936811, + "loss": 1.3574, + "theoretical_loss": 3.5890871404478597, + "tokens_seen": 1302921216 + }, + { + "epoch": 0.39, + "learning_rate": 0.0004890832497492477, + "loss": 1.4629, + "theoretical_loss": 3.58907092035094, + "tokens_seen": 1302986752 + }, + { + "epoch": 0.39, + "learning_rate": 0.0004890672016048145, + "loss": 1.4321, + "theoretical_loss": 3.5890547012982337, + "tokens_seen": 1303052288 + }, + { + "epoch": 0.39, + "learning_rate": 0.0004890511534603811, + "loss": 1.455, + "theoretical_loss": 3.589038483289621, + "tokens_seen": 1303117824 + }, + { + "epoch": 0.39, + "learning_rate": 0.0004890351053159479, + "loss": 1.3397, + "theoretical_loss": 3.5890222663249824, + "tokens_seen": 1303183360 + }, + { + "epoch": 0.39, + "learning_rate": 0.0004890190571715145, + "loss": 1.4341, + "theoretical_loss": 3.589006050404199, + "tokens_seen": 1303248896 + }, + { + "epoch": 0.39, + "learning_rate": 0.0004890030090270813, + "loss": 1.3795, + "theoretical_loss": 3.58898983552715, + "tokens_seen": 1303314432 + }, + { + "epoch": 0.39, + "learning_rate": 0.0004889869608826479, + "loss": 1.4748, + "theoretical_loss": 3.5889736216937163, + "tokens_seen": 1303379968 + }, + { + "epoch": 0.39, + "learning_rate": 0.0004889709127382147, + "loss": 1.3941, + "theoretical_loss": 3.588957408903778, + "tokens_seen": 1303445504 + }, + { + "epoch": 0.4, + "learning_rate": 0.0004889548645937813, + "loss": 1.3893, + "theoretical_loss": 3.5889411971572156, + "tokens_seen": 1303511040 + }, + { + "epoch": 0.4, + "learning_rate": 0.0004889388164493481, + "loss": 1.4119, + "theoretical_loss": 3.5889249864539097, + "tokens_seen": 1303576576 + }, + { + "epoch": 0.4, + "learning_rate": 0.0004889227683049147, + "loss": 1.2894, + "theoretical_loss": 3.5889087767937404, + "tokens_seen": 1303642112 + }, + { + "epoch": 0.4, + "learning_rate": 0.0004889067201604815, + "loss": 1.4198, + "theoretical_loss": 3.588892568176588, + "tokens_seen": 1303707648 + }, + { + "epoch": 0.4, + "learning_rate": 0.0004888906720160481, + "loss": 1.4464, + "theoretical_loss": 3.5888763606023337, + "tokens_seen": 1303773184 + }, + { + "epoch": 0.4, + "learning_rate": 0.0004888746238716149, + "loss": 1.3648, + "theoretical_loss": 3.588860154070857, + "tokens_seen": 1303838720 + }, + { + "epoch": 0.4, + "learning_rate": 0.0004888585757271816, + "loss": 1.4782, + "theoretical_loss": 3.5888439485820394, + "tokens_seen": 1303904256 + }, + { + "epoch": 0.4, + "learning_rate": 0.0004888425275827483, + "loss": 1.4945, + "theoretical_loss": 3.5888277441357603, + "tokens_seen": 1303969792 + }, + { + "epoch": 0.4, + "learning_rate": 0.000488826479438315, + "loss": 1.3643, + "theoretical_loss": 3.5888115407319017, + "tokens_seen": 1304035328 + }, + { + "epoch": 0.4, + "learning_rate": 0.0004888104312938817, + "loss": 1.3447, + "theoretical_loss": 3.5887953383703426, + "tokens_seen": 1304100864 + }, + { + "epoch": 0.4, + "objective/train/docs_used": 476908, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.3083925247192383, + "objective/train/theoretical_loss": 3.588779137050965, + "objective/train/tokens_used": 1324626400, + "theoretical_loss": 3.588779137050965, + "tokens_seen": 1304166400 + }, + { + "epoch": 0.4, + "learning_rate": 0.0004887943831494484, + "loss": 1.2899, + "theoretical_loss": 3.588779137050965, + "tokens_seen": 1304166400 + }, + { + "epoch": 0.4, + "learning_rate": 0.0004887783350050151, + "loss": 1.527, + "theoretical_loss": 3.588762936773648, + "tokens_seen": 1304231936 + }, + { + "epoch": 0.4, + "learning_rate": 0.0004887622868605818, + "loss": 1.3748, + "theoretical_loss": 3.5887467375382736, + "tokens_seen": 1304297472 + }, + { + "epoch": 0.4, + "learning_rate": 0.0004887462387161485, + "loss": 1.3243, + "theoretical_loss": 3.5887305393447217, + "tokens_seen": 1304363008 + }, + { + "epoch": 0.4, + "learning_rate": 0.0004887301905717152, + "loss": 1.5478, + "theoretical_loss": 3.5887143421928727, + "tokens_seen": 1304428544 + }, + { + "epoch": 0.4, + "learning_rate": 0.0004887141424272819, + "loss": 1.3988, + "theoretical_loss": 3.588698146082608, + "tokens_seen": 1304494080 + }, + { + "epoch": 0.4, + "learning_rate": 0.0004886980942828486, + "loss": 1.3098, + "theoretical_loss": 3.5886819510138084, + "tokens_seen": 1304559616 + }, + { + "epoch": 0.4, + "learning_rate": 0.0004886820461384153, + "loss": 1.4141, + "theoretical_loss": 3.5886657569863543, + "tokens_seen": 1304625152 + }, + { + "epoch": 0.4, + "learning_rate": 0.000488665997993982, + "loss": 1.3464, + "theoretical_loss": 3.588649564000126, + "tokens_seen": 1304690688 + }, + { + "epoch": 0.4, + "learning_rate": 0.0004886499498495487, + "loss": 1.4152, + "theoretical_loss": 3.5886333720550043, + "tokens_seen": 1304756224 + }, + { + "epoch": 0.4, + "learning_rate": 0.0004886339017051155, + "loss": 1.3586, + "theoretical_loss": 3.5886171811508705, + "tokens_seen": 1304821760 + }, + { + "epoch": 0.4, + "learning_rate": 0.0004886178535606821, + "loss": 1.3066, + "theoretical_loss": 3.588600991287605, + "tokens_seen": 1304887296 + }, + { + "epoch": 0.4, + "learning_rate": 0.0004886018054162489, + "loss": 1.5075, + "theoretical_loss": 3.588584802465089, + "tokens_seen": 1304952832 + }, + { + "epoch": 0.4, + "learning_rate": 0.0004885857572718155, + "loss": 1.4701, + "theoretical_loss": 3.5885686146832025, + "tokens_seen": 1305018368 + }, + { + "epoch": 0.4, + "learning_rate": 0.0004885697091273823, + "loss": 1.4599, + "theoretical_loss": 3.5885524279418277, + "tokens_seen": 1305083904 + }, + { + "epoch": 0.4, + "learning_rate": 0.0004885536609829489, + "loss": 1.4668, + "theoretical_loss": 3.5885362422408447, + "tokens_seen": 1305149440 + }, + { + "epoch": 0.4, + "learning_rate": 0.0004885376128385157, + "loss": 1.3868, + "theoretical_loss": 3.5885200575801344, + "tokens_seen": 1305214976 + }, + { + "epoch": 0.4, + "learning_rate": 0.0004885215646940823, + "loss": 1.4081, + "theoretical_loss": 3.588503873959578, + "tokens_seen": 1305280512 + }, + { + "epoch": 0.4, + "learning_rate": 0.000488505516549649, + "loss": 1.3997, + "theoretical_loss": 3.588487691379055, + "tokens_seen": 1305346048 + }, + { + "epoch": 0.4, + "learning_rate": 0.0004884894684052157, + "loss": 1.4279, + "theoretical_loss": 3.588471509838449, + "tokens_seen": 1305411584 + }, + { + "epoch": 0.4, + "learning_rate": 0.0004884734202607823, + "loss": 1.3126, + "theoretical_loss": 3.5884553293376387, + "tokens_seen": 1305477120 + }, + { + "epoch": 0.4, + "learning_rate": 0.0004884573721163491, + "loss": 1.3872, + "theoretical_loss": 3.588439149876506, + "tokens_seen": 1305542656 + }, + { + "epoch": 0.4, + "learning_rate": 0.0004884413239719157, + "loss": 1.326, + "theoretical_loss": 3.5884229714549316, + "tokens_seen": 1305608192 + }, + { + "epoch": 0.4, + "learning_rate": 0.0004884252758274825, + "loss": 1.3075, + "theoretical_loss": 3.588406794072797, + "tokens_seen": 1305673728 + }, + { + "epoch": 0.4, + "learning_rate": 0.0004884092276830492, + "loss": 1.4601, + "theoretical_loss": 3.588390617729983, + "tokens_seen": 1305739264 + }, + { + "epoch": 0.4, + "learning_rate": 0.0004883931795386159, + "loss": 1.4169, + "theoretical_loss": 3.5883744424263706, + "tokens_seen": 1305804800 + }, + { + "epoch": 0.4, + "learning_rate": 0.0004883771313941826, + "loss": 1.2972, + "theoretical_loss": 3.5883582681618407, + "tokens_seen": 1305870336 + }, + { + "epoch": 0.4, + "learning_rate": 0.0004883610832497492, + "loss": 1.3268, + "theoretical_loss": 3.588342094936275, + "tokens_seen": 1305935872 + }, + { + "epoch": 0.4, + "learning_rate": 0.000488345035105316, + "loss": 1.262, + "theoretical_loss": 3.5883259227495543, + "tokens_seen": 1306001408 + }, + { + "epoch": 0.4, + "learning_rate": 0.0004883289869608826, + "loss": 1.3871, + "theoretical_loss": 3.5883097516015594, + "tokens_seen": 1306066944 + }, + { + "epoch": 0.4, + "learning_rate": 0.0004883129388164494, + "loss": 1.409, + "theoretical_loss": 3.588293581492172, + "tokens_seen": 1306132480 + }, + { + "epoch": 0.4, + "learning_rate": 0.000488296890672016, + "loss": 1.2983, + "theoretical_loss": 3.5882774124212733, + "tokens_seen": 1306198016 + }, + { + "epoch": 0.4, + "learning_rate": 0.0004882808425275828, + "loss": 1.3489, + "theoretical_loss": 3.588261244388744, + "tokens_seen": 1306263552 + }, + { + "epoch": 0.4, + "learning_rate": 0.00048826479438314946, + "loss": 1.4798, + "theoretical_loss": 3.5882450773944656, + "tokens_seen": 1306329088 + }, + { + "epoch": 0.4, + "learning_rate": 0.0004882487462387162, + "loss": 1.4264, + "theoretical_loss": 3.5882289114383195, + "tokens_seen": 1306394624 + }, + { + "epoch": 0.4, + "learning_rate": 0.00048823269809428286, + "loss": 1.2764, + "theoretical_loss": 3.5882127465201865, + "tokens_seen": 1306460160 + }, + { + "epoch": 0.4, + "learning_rate": 0.0004882166499498496, + "loss": 1.364, + "theoretical_loss": 3.5881965826399487, + "tokens_seen": 1306525696 + }, + { + "epoch": 0.4, + "learning_rate": 0.00048820060180541626, + "loss": 1.3551, + "theoretical_loss": 3.588180419797487, + "tokens_seen": 1306591232 + }, + { + "epoch": 0.4, + "learning_rate": 0.000488184553660983, + "loss": 1.3897, + "theoretical_loss": 3.5881642579926822, + "tokens_seen": 1306656768 + }, + { + "epoch": 0.4, + "learning_rate": 0.00048816850551654966, + "loss": 1.4954, + "theoretical_loss": 3.588148097225416, + "tokens_seen": 1306722304 + }, + { + "epoch": 0.4, + "learning_rate": 0.0004881524573721164, + "loss": 1.3109, + "theoretical_loss": 3.58813193749557, + "tokens_seen": 1306787840 + }, + { + "epoch": 0.4, + "learning_rate": 0.00048813640922768306, + "loss": 1.4091, + "theoretical_loss": 3.5881157788030253, + "tokens_seen": 1306853376 + }, + { + "epoch": 0.4, + "learning_rate": 0.00048812036108324976, + "loss": 1.4706, + "theoretical_loss": 3.5880996211476637, + "tokens_seen": 1306918912 + }, + { + "epoch": 0.4, + "learning_rate": 0.00048810431293881646, + "loss": 1.3289, + "theoretical_loss": 3.588083464529366, + "tokens_seen": 1306984448 + }, + { + "epoch": 0.4, + "learning_rate": 0.00048808826479438316, + "loss": 1.3888, + "theoretical_loss": 3.5880673089480144, + "tokens_seen": 1307049984 + }, + { + "epoch": 0.4, + "learning_rate": 0.00048807221664994986, + "loss": 1.3682, + "theoretical_loss": 3.5880511544034897, + "tokens_seen": 1307115520 + }, + { + "epoch": 0.4, + "learning_rate": 0.00048805616850551656, + "loss": 1.4466, + "theoretical_loss": 3.5880350008956734, + "tokens_seen": 1307181056 + }, + { + "epoch": 0.4, + "learning_rate": 0.00048804012036108327, + "loss": 1.5329, + "theoretical_loss": 3.588018848424448, + "tokens_seen": 1307246592 + }, + { + "epoch": 0.4, + "learning_rate": 0.00048802407221664997, + "loss": 1.3812, + "theoretical_loss": 3.5880026969896934, + "tokens_seen": 1307312128 + }, + { + "epoch": 0.4, + "learning_rate": 0.00048800802407221667, + "loss": 1.3087, + "theoretical_loss": 3.587986546591292, + "tokens_seen": 1307377664 + }, + { + "epoch": 0.4, + "objective/train/docs_used": 478148, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.4670071601867676, + "objective/train/theoretical_loss": 3.587970397229126, + "objective/train/tokens_used": 1327903200, + "theoretical_loss": 3.587970397229126, + "tokens_seen": 1307443200 + }, + { + "epoch": 0.4, + "learning_rate": 0.00048799197592778337, + "loss": 1.3789, + "theoretical_loss": 3.587970397229126, + "tokens_seen": 1307443200 + }, + { + "epoch": 0.4, + "learning_rate": 0.00048797592778335, + "loss": 1.4573, + "theoretical_loss": 3.587954248903076, + "tokens_seen": 1307508736 + }, + { + "epoch": 0.4, + "learning_rate": 0.00048795987963891677, + "loss": 1.2805, + "theoretical_loss": 3.5879381016130236, + "tokens_seen": 1307574272 + }, + { + "epoch": 0.4, + "learning_rate": 0.0004879438314944834, + "loss": 1.4081, + "theoretical_loss": 3.5879219553588513, + "tokens_seen": 1307639808 + }, + { + "epoch": 0.4, + "learning_rate": 0.00048792778335005017, + "loss": 1.4213, + "theoretical_loss": 3.5879058101404397, + "tokens_seen": 1307705344 + }, + { + "epoch": 0.4, + "learning_rate": 0.0004879117352056168, + "loss": 1.4565, + "theoretical_loss": 3.587889665957671, + "tokens_seen": 1307770880 + }, + { + "epoch": 0.4, + "learning_rate": 0.00048789568706118357, + "loss": 1.3544, + "theoretical_loss": 3.587873522810427, + "tokens_seen": 1307836416 + }, + { + "epoch": 0.4, + "learning_rate": 0.0004878796389167502, + "loss": 1.3579, + "theoretical_loss": 3.5878573806985887, + "tokens_seen": 1307901952 + }, + { + "epoch": 0.4, + "learning_rate": 0.00048786359077231697, + "loss": 1.4572, + "theoretical_loss": 3.5878412396220387, + "tokens_seen": 1307967488 + }, + { + "epoch": 0.4, + "learning_rate": 0.0004878475426278836, + "loss": 1.4433, + "theoretical_loss": 3.587825099580658, + "tokens_seen": 1308033024 + }, + { + "epoch": 0.4, + "learning_rate": 0.0004878314944834504, + "loss": 1.4322, + "theoretical_loss": 3.5878089605743293, + "tokens_seen": 1308098560 + }, + { + "epoch": 0.4, + "learning_rate": 0.000487815446339017, + "loss": 1.4103, + "theoretical_loss": 3.5877928226029336, + "tokens_seen": 1308164096 + }, + { + "epoch": 0.4, + "learning_rate": 0.0004877993981945838, + "loss": 1.3931, + "theoretical_loss": 3.5877766856663524, + "tokens_seen": 1308229632 + }, + { + "epoch": 0.4, + "learning_rate": 0.0004877833500501504, + "loss": 1.3003, + "theoretical_loss": 3.587760549764468, + "tokens_seen": 1308295168 + }, + { + "epoch": 0.4, + "learning_rate": 0.0004877673019057172, + "loss": 1.3372, + "theoretical_loss": 3.5877444148971627, + "tokens_seen": 1308360704 + }, + { + "epoch": 0.4, + "learning_rate": 0.0004877512537612838, + "loss": 1.3626, + "theoretical_loss": 3.5877282810643174, + "tokens_seen": 1308426240 + }, + { + "epoch": 0.4, + "learning_rate": 0.0004877352056168506, + "loss": 1.3603, + "theoretical_loss": 3.5877121482658145, + "tokens_seen": 1308491776 + }, + { + "epoch": 0.4, + "learning_rate": 0.0004877191574724172, + "loss": 1.2942, + "theoretical_loss": 3.587696016501536, + "tokens_seen": 1308557312 + }, + { + "epoch": 0.4, + "learning_rate": 0.000487703109327984, + "loss": 1.3703, + "theoretical_loss": 3.5876798857713634, + "tokens_seen": 1308622848 + }, + { + "epoch": 0.4, + "learning_rate": 0.00048768706118355073, + "loss": 1.5021, + "theoretical_loss": 3.5876637560751785, + "tokens_seen": 1308688384 + }, + { + "epoch": 0.4, + "learning_rate": 0.0004876710130391174, + "loss": 1.3003, + "theoretical_loss": 3.5876476274128644, + "tokens_seen": 1308753920 + }, + { + "epoch": 0.4, + "learning_rate": 0.00048765496489468413, + "loss": 1.3685, + "theoretical_loss": 3.587631499784301, + "tokens_seen": 1308819456 + }, + { + "epoch": 0.4, + "learning_rate": 0.0004876389167502508, + "loss": 1.3863, + "theoretical_loss": 3.5876153731893727, + "tokens_seen": 1308884992 + }, + { + "epoch": 0.4, + "learning_rate": 0.00048762286860581754, + "loss": 1.3725, + "theoretical_loss": 3.58759924762796, + "tokens_seen": 1308950528 + }, + { + "epoch": 0.4, + "learning_rate": 0.0004876068204613842, + "loss": 1.3386, + "theoretical_loss": 3.587583123099945, + "tokens_seen": 1309016064 + }, + { + "epoch": 0.4, + "learning_rate": 0.00048759077231695094, + "loss": 1.4189, + "theoretical_loss": 3.58756699960521, + "tokens_seen": 1309081600 + }, + { + "epoch": 0.4, + "learning_rate": 0.0004875747241725176, + "loss": 1.4566, + "theoretical_loss": 3.587550877143637, + "tokens_seen": 1309147136 + }, + { + "epoch": 0.4, + "learning_rate": 0.00048755867602808434, + "loss": 1.4898, + "theoretical_loss": 3.5875347557151085, + "tokens_seen": 1309212672 + }, + { + "epoch": 0.4, + "learning_rate": 0.000487542627883651, + "loss": 1.4846, + "theoretical_loss": 3.5875186353195065, + "tokens_seen": 1309278208 + }, + { + "epoch": 0.4, + "learning_rate": 0.00048752657973921774, + "loss": 1.4352, + "theoretical_loss": 3.5875025159567118, + "tokens_seen": 1309343744 + }, + { + "epoch": 0.4, + "learning_rate": 0.0004875105315947844, + "loss": 1.4429, + "theoretical_loss": 3.5874863976266083, + "tokens_seen": 1309409280 + }, + { + "epoch": 0.4, + "learning_rate": 0.0004874944834503511, + "loss": 1.4794, + "theoretical_loss": 3.5874702803290774, + "tokens_seen": 1309474816 + }, + { + "epoch": 0.4, + "learning_rate": 0.0004874784353059178, + "loss": 1.3945, + "theoretical_loss": 3.5874541640640016, + "tokens_seen": 1309540352 + }, + { + "epoch": 0.4, + "learning_rate": 0.0004874623871614845, + "loss": 1.3581, + "theoretical_loss": 3.5874380488312623, + "tokens_seen": 1309605888 + }, + { + "epoch": 0.4, + "learning_rate": 0.0004874463390170512, + "loss": 1.3773, + "theoretical_loss": 3.5874219346307425, + "tokens_seen": 1309671424 + }, + { + "epoch": 0.4, + "learning_rate": 0.0004874302908726179, + "loss": 1.381, + "theoretical_loss": 3.5874058214623243, + "tokens_seen": 1309736960 + }, + { + "epoch": 0.4, + "learning_rate": 0.0004874142427281846, + "loss": 1.4116, + "theoretical_loss": 3.5873897093258895, + "tokens_seen": 1309802496 + }, + { + "epoch": 0.4, + "learning_rate": 0.0004873981945837513, + "loss": 1.4323, + "theoretical_loss": 3.5873735982213213, + "tokens_seen": 1309868032 + }, + { + "epoch": 0.4, + "learning_rate": 0.000487382146439318, + "loss": 1.4637, + "theoretical_loss": 3.587357488148501, + "tokens_seen": 1309933568 + }, + { + "epoch": 0.4, + "learning_rate": 0.0004873660982948847, + "loss": 1.4376, + "theoretical_loss": 3.587341379107311, + "tokens_seen": 1309999104 + }, + { + "epoch": 0.4, + "learning_rate": 0.00048735005015045134, + "loss": 1.265, + "theoretical_loss": 3.587325271097635, + "tokens_seen": 1310064640 + }, + { + "epoch": 0.4, + "learning_rate": 0.0004873340020060181, + "loss": 1.4511, + "theoretical_loss": 3.587309164119353, + "tokens_seen": 1310130176 + }, + { + "epoch": 0.4, + "learning_rate": 0.00048731795386158474, + "loss": 1.4866, + "theoretical_loss": 3.5872930581723494, + "tokens_seen": 1310195712 + }, + { + "epoch": 0.4, + "learning_rate": 0.0004873019057171515, + "loss": 1.3189, + "theoretical_loss": 3.587276953256506, + "tokens_seen": 1310261248 + }, + { + "epoch": 0.4, + "learning_rate": 0.00048728585757271814, + "loss": 1.2733, + "theoretical_loss": 3.587260849371705, + "tokens_seen": 1310326784 + }, + { + "epoch": 0.4, + "learning_rate": 0.0004872698094282849, + "loss": 1.3909, + "theoretical_loss": 3.587244746517829, + "tokens_seen": 1310392320 + }, + { + "epoch": 0.4, + "learning_rate": 0.00048725376128385154, + "loss": 1.449, + "theoretical_loss": 3.5872286446947603, + "tokens_seen": 1310457856 + }, + { + "epoch": 0.4, + "learning_rate": 0.0004872377131394183, + "loss": 1.3119, + "theoretical_loss": 3.587212543902381, + "tokens_seen": 1310523392 + }, + { + "epoch": 0.4, + "learning_rate": 0.00048722166499498494, + "loss": 1.3725, + "theoretical_loss": 3.5871964441405746, + "tokens_seen": 1310588928 + }, + { + "epoch": 0.4, + "learning_rate": 0.0004872056168505517, + "loss": 1.3181, + "theoretical_loss": 3.587180345409222, + "tokens_seen": 1310654464 + }, + { + "debugging/Compilability": 0.9, + "debugging/distinct-1-grams": 0.7732995475821178, + "debugging/entropy-1-grams": 5.431960387820077, + "debugging/length": 501.4, + "debugging/num_segments": 10, + "debugging/score": 0.007191963134632398, + "debugging/score_std": 0.006508080924945895, + "epoch": 0.4, + "objective/train/docs_used": 479375, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.434427261352539, + "objective/train/theoretical_loss": 3.587164247708208, + "objective/train/tokens_used": 1331180000, + "theoretical_loss": 3.587164247708208, + "tokens_seen": 1310720000 + }, + { + "epoch": 0.4, + "learning_rate": 0.00048718956870611834, + "loss": 1.412, + "theoretical_loss": 3.587164247708208, + "tokens_seen": 1310720000 + }, + { + "epoch": 0.4, + "learning_rate": 0.0004871735205616851, + "loss": 1.4419, + "theoretical_loss": 3.5871481510374137, + "tokens_seen": 1310785536 + }, + { + "epoch": 0.4, + "learning_rate": 0.00048715747241725175, + "loss": 1.2661, + "theoretical_loss": 3.5871320553967214, + "tokens_seen": 1310851072 + }, + { + "epoch": 0.4, + "learning_rate": 0.0004871414242728185, + "loss": 1.3589, + "theoretical_loss": 3.587115960786014, + "tokens_seen": 1310916608 + }, + { + "epoch": 0.4, + "learning_rate": 0.00048712537612838515, + "loss": 1.3794, + "theoretical_loss": 3.5870998672051746, + "tokens_seen": 1310982144 + }, + { + "epoch": 0.4, + "learning_rate": 0.0004871093279839519, + "loss": 1.3921, + "theoretical_loss": 3.5870837746540856, + "tokens_seen": 1311047680 + }, + { + "epoch": 0.4, + "learning_rate": 0.00048709327983951855, + "loss": 1.3924, + "theoretical_loss": 3.5870676831326294, + "tokens_seen": 1311113216 + }, + { + "epoch": 0.4, + "learning_rate": 0.0004870772316950853, + "loss": 1.3824, + "theoretical_loss": 3.5870515926406887, + "tokens_seen": 1311178752 + }, + { + "epoch": 0.4, + "learning_rate": 0.00048706118355065195, + "loss": 1.4464, + "theoretical_loss": 3.5870355031781465, + "tokens_seen": 1311244288 + }, + { + "epoch": 0.4, + "learning_rate": 0.0004870451354062187, + "loss": 1.4768, + "theoretical_loss": 3.587019414744885, + "tokens_seen": 1311309824 + }, + { + "epoch": 0.4, + "learning_rate": 0.00048702908726178535, + "loss": 1.4463, + "theoretical_loss": 3.587003327340787, + "tokens_seen": 1311375360 + }, + { + "epoch": 0.4, + "learning_rate": 0.0004870130391173521, + "loss": 1.3228, + "theoretical_loss": 3.586987240965736, + "tokens_seen": 1311440896 + }, + { + "epoch": 0.4, + "learning_rate": 0.00048699699097291875, + "loss": 1.409, + "theoretical_loss": 3.5869711556196133, + "tokens_seen": 1311506432 + }, + { + "epoch": 0.4, + "learning_rate": 0.0004869809428284855, + "loss": 1.2821, + "theoretical_loss": 3.5869550713023033, + "tokens_seen": 1311571968 + }, + { + "epoch": 0.4, + "learning_rate": 0.00048696489468405215, + "loss": 1.3954, + "theoretical_loss": 3.5869389880136877, + "tokens_seen": 1311637504 + }, + { + "epoch": 0.4, + "learning_rate": 0.0004869488465396189, + "loss": 1.3894, + "theoretical_loss": 3.5869229057536494, + "tokens_seen": 1311703040 + }, + { + "epoch": 0.4, + "learning_rate": 0.00048693279839518555, + "loss": 1.2357, + "theoretical_loss": 3.586906824522072, + "tokens_seen": 1311768576 + }, + { + "epoch": 0.4, + "learning_rate": 0.0004869167502507523, + "loss": 1.2552, + "theoretical_loss": 3.586890744318838, + "tokens_seen": 1311834112 + }, + { + "epoch": 0.4, + "learning_rate": 0.00048690070210631896, + "loss": 1.256, + "theoretical_loss": 3.5868746651438297, + "tokens_seen": 1311899648 + }, + { + "epoch": 0.4, + "learning_rate": 0.0004868846539618857, + "loss": 1.3537, + "theoretical_loss": 3.5868585869969305, + "tokens_seen": 1311965184 + }, + { + "epoch": 0.4, + "learning_rate": 0.00048686860581745236, + "loss": 1.317, + "theoretical_loss": 3.5868425098780228, + "tokens_seen": 1312030720 + }, + { + "epoch": 0.4, + "learning_rate": 0.0004868525576730191, + "loss": 1.43, + "theoretical_loss": 3.5868264337869906, + "tokens_seen": 1312096256 + }, + { + "epoch": 0.4, + "learning_rate": 0.00048683650952858576, + "loss": 1.3881, + "theoretical_loss": 3.5868103587237155, + "tokens_seen": 1312161792 + }, + { + "epoch": 0.4, + "learning_rate": 0.0004868204613841525, + "loss": 1.4764, + "theoretical_loss": 3.586794284688082, + "tokens_seen": 1312227328 + }, + { + "epoch": 0.4, + "learning_rate": 0.00048680441323971916, + "loss": 1.4236, + "theoretical_loss": 3.586778211679972, + "tokens_seen": 1312292864 + }, + { + "epoch": 0.4, + "learning_rate": 0.0004867883650952859, + "loss": 1.3943, + "theoretical_loss": 3.5867621396992684, + "tokens_seen": 1312358400 + }, + { + "epoch": 0.4, + "learning_rate": 0.00048677231695085256, + "loss": 1.4533, + "theoretical_loss": 3.5867460687458546, + "tokens_seen": 1312423936 + }, + { + "epoch": 0.4, + "learning_rate": 0.0004867562688064193, + "loss": 1.4105, + "theoretical_loss": 3.5867299988196137, + "tokens_seen": 1312489472 + }, + { + "epoch": 0.4, + "learning_rate": 0.00048674022066198596, + "loss": 1.2511, + "theoretical_loss": 3.5867139299204287, + "tokens_seen": 1312555008 + }, + { + "epoch": 0.4, + "learning_rate": 0.00048672417251755266, + "loss": 1.4322, + "theoretical_loss": 3.586697862048183, + "tokens_seen": 1312620544 + }, + { + "epoch": 0.4, + "learning_rate": 0.00048670812437311936, + "loss": 1.3835, + "theoretical_loss": 3.586681795202759, + "tokens_seen": 1312686080 + }, + { + "epoch": 0.4, + "learning_rate": 0.00048669207622868606, + "loss": 1.3953, + "theoretical_loss": 3.5866657293840403, + "tokens_seen": 1312751616 + }, + { + "epoch": 0.4, + "learning_rate": 0.00048667602808425277, + "loss": 1.3429, + "theoretical_loss": 3.58664966459191, + "tokens_seen": 1312817152 + }, + { + "epoch": 0.4, + "learning_rate": 0.00048665997993981947, + "loss": 1.3058, + "theoretical_loss": 3.586633600826251, + "tokens_seen": 1312882688 + }, + { + "epoch": 0.4, + "learning_rate": 0.00048664393179538617, + "loss": 1.5594, + "theoretical_loss": 3.586617538086947, + "tokens_seen": 1312948224 + }, + { + "epoch": 0.4, + "learning_rate": 0.00048662788365095287, + "loss": 1.3753, + "theoretical_loss": 3.5866014763738807, + "tokens_seen": 1313013760 + }, + { + "epoch": 0.4, + "learning_rate": 0.00048661183550651957, + "loss": 1.3055, + "theoretical_loss": 3.5865854156869355, + "tokens_seen": 1313079296 + }, + { + "epoch": 0.4, + "learning_rate": 0.00048659578736208627, + "loss": 1.2826, + "theoretical_loss": 3.5865693560259944, + "tokens_seen": 1313144832 + }, + { + "epoch": 0.4, + "learning_rate": 0.0004865797392176529, + "loss": 1.3454, + "theoretical_loss": 3.586553297390941, + "tokens_seen": 1313210368 + }, + { + "epoch": 0.4, + "learning_rate": 0.00048656369107321967, + "loss": 1.5334, + "theoretical_loss": 3.5865372397816584, + "tokens_seen": 1313275904 + }, + { + "epoch": 0.4, + "learning_rate": 0.0004865476429287863, + "loss": 1.3485, + "theoretical_loss": 3.5865211831980304, + "tokens_seen": 1313341440 + }, + { + "epoch": 0.4, + "learning_rate": 0.00048653159478435307, + "loss": 1.4287, + "theoretical_loss": 3.5865051276399393, + "tokens_seen": 1313406976 + }, + { + "epoch": 0.4, + "learning_rate": 0.0004865155466399197, + "loss": 1.45, + "theoretical_loss": 3.5864890731072694, + "tokens_seen": 1313472512 + }, + { + "epoch": 0.4, + "learning_rate": 0.00048649949849548647, + "loss": 1.3772, + "theoretical_loss": 3.5864730195999037, + "tokens_seen": 1313538048 + }, + { + "epoch": 0.4, + "learning_rate": 0.0004864834503510532, + "loss": 1.4242, + "theoretical_loss": 3.586456967117725, + "tokens_seen": 1313603584 + }, + { + "epoch": 0.4, + "learning_rate": 0.0004864674022066199, + "loss": 1.5166, + "theoretical_loss": 3.5864409156606176, + "tokens_seen": 1313669120 + }, + { + "epoch": 0.4, + "learning_rate": 0.00048645135406218663, + "loss": 1.417, + "theoretical_loss": 3.5864248652284645, + "tokens_seen": 1313734656 + }, + { + "epoch": 0.4, + "learning_rate": 0.0004864353059177533, + "loss": 1.4412, + "theoretical_loss": 3.586408815821149, + "tokens_seen": 1313800192 + }, + { + "epoch": 0.4, + "learning_rate": 0.00048641925777332003, + "loss": 1.3904, + "theoretical_loss": 3.5863927674385545, + "tokens_seen": 1313865728 + }, + { + "epoch": 0.4, + "learning_rate": 0.0004864032096288867, + "loss": 1.4834, + "theoretical_loss": 3.586376720080565, + "tokens_seen": 1313931264 + }, + { + "epoch": 0.4, + "objective/train/docs_used": 480644, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.3999719619750977, + "objective/train/theoretical_loss": 3.5863606737470635, + "objective/train/tokens_used": 1334456800, + "theoretical_loss": 3.5863606737470635, + "tokens_seen": 1313996800 + }, + { + "epoch": 0.4, + "learning_rate": 0.00048638716148445343, + "loss": 1.4023, + "theoretical_loss": 3.5863606737470635, + "tokens_seen": 1313996800 + }, + { + "epoch": 0.4, + "learning_rate": 0.0004863711133400201, + "loss": 1.322, + "theoretical_loss": 3.5863446284379337, + "tokens_seen": 1314062336 + }, + { + "epoch": 0.4, + "learning_rate": 0.00048635506519558683, + "loss": 1.3556, + "theoretical_loss": 3.586328584153059, + "tokens_seen": 1314127872 + }, + { + "epoch": 0.4, + "learning_rate": 0.0004863390170511535, + "loss": 1.3691, + "theoretical_loss": 3.586312540892323, + "tokens_seen": 1314193408 + }, + { + "epoch": 0.4, + "learning_rate": 0.00048632296890672023, + "loss": 1.3281, + "theoretical_loss": 3.586296498655609, + "tokens_seen": 1314258944 + }, + { + "epoch": 0.4, + "learning_rate": 0.0004863069207622869, + "loss": 1.3961, + "theoretical_loss": 3.5862804574428013, + "tokens_seen": 1314324480 + }, + { + "epoch": 0.4, + "learning_rate": 0.00048629087261785363, + "loss": 1.4927, + "theoretical_loss": 3.5862644172537825, + "tokens_seen": 1314390016 + }, + { + "epoch": 0.4, + "learning_rate": 0.0004862748244734203, + "loss": 1.4323, + "theoretical_loss": 3.5862483780884373, + "tokens_seen": 1314455552 + }, + { + "epoch": 0.4, + "learning_rate": 0.00048625877632898704, + "loss": 1.4172, + "theoretical_loss": 3.5862323399466485, + "tokens_seen": 1314521088 + }, + { + "epoch": 0.4, + "learning_rate": 0.0004862427281845537, + "loss": 1.2548, + "theoretical_loss": 3.5862163028283005, + "tokens_seen": 1314586624 + }, + { + "epoch": 0.4, + "learning_rate": 0.00048622668004012044, + "loss": 1.3778, + "theoretical_loss": 3.586200266733276, + "tokens_seen": 1314652160 + }, + { + "epoch": 0.4, + "learning_rate": 0.0004862106318956871, + "loss": 1.4693, + "theoretical_loss": 3.586184231661459, + "tokens_seen": 1314717696 + }, + { + "epoch": 0.4, + "learning_rate": 0.00048619458375125384, + "loss": 1.3654, + "theoretical_loss": 3.586168197612734, + "tokens_seen": 1314783232 + }, + { + "epoch": 0.4, + "learning_rate": 0.0004861785356068205, + "loss": 1.3163, + "theoretical_loss": 3.5861521645869843, + "tokens_seen": 1314848768 + }, + { + "epoch": 0.4, + "learning_rate": 0.00048616248746238724, + "loss": 1.3535, + "theoretical_loss": 3.586136132584093, + "tokens_seen": 1314914304 + }, + { + "epoch": 0.4, + "learning_rate": 0.0004861464393179539, + "loss": 1.3971, + "theoretical_loss": 3.5861201016039446, + "tokens_seen": 1314979840 + }, + { + "epoch": 0.4, + "learning_rate": 0.00048613039117352064, + "loss": 1.3564, + "theoretical_loss": 3.5861040716464228, + "tokens_seen": 1315045376 + }, + { + "epoch": 0.4, + "learning_rate": 0.0004861143430290873, + "loss": 1.4443, + "theoretical_loss": 3.5860880427114106, + "tokens_seen": 1315110912 + }, + { + "epoch": 0.4, + "learning_rate": 0.00048609829488465404, + "loss": 1.4284, + "theoretical_loss": 3.586072014798793, + "tokens_seen": 1315176448 + }, + { + "epoch": 0.4, + "learning_rate": 0.0004860822467402207, + "loss": 1.3007, + "theoretical_loss": 3.5860559879084537, + "tokens_seen": 1315241984 + }, + { + "epoch": 0.4, + "learning_rate": 0.0004860661985957874, + "loss": 1.3516, + "theoretical_loss": 3.5860399620402754, + "tokens_seen": 1315307520 + }, + { + "epoch": 0.4, + "learning_rate": 0.0004860501504513541, + "loss": 1.2493, + "theoretical_loss": 3.5860239371941436, + "tokens_seen": 1315373056 + }, + { + "epoch": 0.4, + "learning_rate": 0.0004860341023069208, + "loss": 1.4884, + "theoretical_loss": 3.586007913369941, + "tokens_seen": 1315438592 + }, + { + "epoch": 0.4, + "learning_rate": 0.0004860180541624875, + "loss": 1.4235, + "theoretical_loss": 3.585991890567552, + "tokens_seen": 1315504128 + }, + { + "epoch": 0.4, + "learning_rate": 0.0004860020060180542, + "loss": 1.3931, + "theoretical_loss": 3.5859758687868606, + "tokens_seen": 1315569664 + }, + { + "epoch": 0.4, + "learning_rate": 0.0004859859578736209, + "loss": 1.4874, + "theoretical_loss": 3.58595984802775, + "tokens_seen": 1315635200 + }, + { + "epoch": 0.4, + "learning_rate": 0.0004859699097291876, + "loss": 1.4593, + "theoretical_loss": 3.5859438282901057, + "tokens_seen": 1315700736 + }, + { + "epoch": 0.4, + "learning_rate": 0.00048595386158475424, + "loss": 1.3205, + "theoretical_loss": 3.5859278095738105, + "tokens_seen": 1315766272 + }, + { + "epoch": 0.4, + "learning_rate": 0.000485937813440321, + "loss": 1.4698, + "theoretical_loss": 3.585911791878748, + "tokens_seen": 1315831808 + }, + { + "epoch": 0.4, + "learning_rate": 0.00048592176529588764, + "loss": 1.4062, + "theoretical_loss": 3.5858957752048037, + "tokens_seen": 1315897344 + }, + { + "epoch": 0.4, + "learning_rate": 0.0004859057171514544, + "loss": 1.2727, + "theoretical_loss": 3.585879759551861, + "tokens_seen": 1315962880 + }, + { + "epoch": 0.4, + "learning_rate": 0.00048588966900702104, + "loss": 1.5415, + "theoretical_loss": 3.5858637449198034, + "tokens_seen": 1316028416 + }, + { + "epoch": 0.4, + "learning_rate": 0.0004858736208625878, + "loss": 1.4623, + "theoretical_loss": 3.5858477313085158, + "tokens_seen": 1316093952 + }, + { + "epoch": 0.4, + "learning_rate": 0.00048585757271815444, + "loss": 1.427, + "theoretical_loss": 3.585831718717882, + "tokens_seen": 1316159488 + }, + { + "epoch": 0.4, + "learning_rate": 0.0004858415245737212, + "loss": 1.4158, + "theoretical_loss": 3.585815707147786, + "tokens_seen": 1316225024 + }, + { + "epoch": 0.4, + "learning_rate": 0.00048582547642928784, + "loss": 1.3106, + "theoretical_loss": 3.5857996965981123, + "tokens_seen": 1316290560 + }, + { + "epoch": 0.4, + "learning_rate": 0.0004858094282848546, + "loss": 1.3466, + "theoretical_loss": 3.5857836870687443, + "tokens_seen": 1316356096 + }, + { + "epoch": 0.4, + "learning_rate": 0.00048579338014042125, + "loss": 1.4961, + "theoretical_loss": 3.585767678559567, + "tokens_seen": 1316421632 + }, + { + "epoch": 0.4, + "learning_rate": 0.000485777331995988, + "loss": 1.3042, + "theoretical_loss": 3.585751671070464, + "tokens_seen": 1316487168 + }, + { + "epoch": 0.4, + "learning_rate": 0.00048576128385155465, + "loss": 1.4836, + "theoretical_loss": 3.5857356646013203, + "tokens_seen": 1316552704 + }, + { + "epoch": 0.4, + "learning_rate": 0.0004857452357071214, + "loss": 1.5224, + "theoretical_loss": 3.5857196591520197, + "tokens_seen": 1316618240 + }, + { + "epoch": 0.4, + "learning_rate": 0.00048572918756268805, + "loss": 1.479, + "theoretical_loss": 3.585703654722446, + "tokens_seen": 1316683776 + }, + { + "epoch": 0.4, + "learning_rate": 0.0004857131394182548, + "loss": 1.3825, + "theoretical_loss": 3.5856876513124845, + "tokens_seen": 1316749312 + }, + { + "epoch": 0.4, + "learning_rate": 0.00048569709127382145, + "loss": 1.4649, + "theoretical_loss": 3.585671648922019, + "tokens_seen": 1316814848 + }, + { + "epoch": 0.4, + "learning_rate": 0.0004856810431293882, + "loss": 1.3351, + "theoretical_loss": 3.585655647550933, + "tokens_seen": 1316880384 + }, + { + "epoch": 0.4, + "learning_rate": 0.00048566499498495485, + "loss": 1.497, + "theoretical_loss": 3.5856396471991117, + "tokens_seen": 1316945920 + }, + { + "epoch": 0.4, + "learning_rate": 0.0004856489468405216, + "loss": 1.365, + "theoretical_loss": 3.5856236478664396, + "tokens_seen": 1317011456 + }, + { + "epoch": 0.4, + "learning_rate": 0.00048563289869608825, + "loss": 1.352, + "theoretical_loss": 3.5856076495528013, + "tokens_seen": 1317076992 + }, + { + "epoch": 0.4, + "learning_rate": 0.000485616850551655, + "loss": 1.4584, + "theoretical_loss": 3.58559165225808, + "tokens_seen": 1317142528 + }, + { + "epoch": 0.4, + "learning_rate": 0.00048560080240722165, + "loss": 1.527, + "theoretical_loss": 3.5855756559821614, + "tokens_seen": 1317208064 + }, + { + "epoch": 0.4, + "objective/train/docs_used": 481887, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.5025660991668701, + "objective/train/theoretical_loss": 3.585559660724929, + "objective/train/tokens_used": 1337733600, + "theoretical_loss": 3.585559660724929, + "tokens_seen": 1317273600 + }, + { + "epoch": 0.4, + "learning_rate": 0.0004855847542627884, + "loss": 1.4469, + "theoretical_loss": 3.585559660724929, + "tokens_seen": 1317273600 + }, + { + "epoch": 0.4, + "learning_rate": 0.00048556870611835505, + "loss": 1.5178, + "theoretical_loss": 3.5855436664862674, + "tokens_seen": 1317339136 + }, + { + "epoch": 0.4, + "learning_rate": 0.0004855526579739218, + "loss": 1.6456, + "theoretical_loss": 3.5855276732660615, + "tokens_seen": 1317404672 + }, + { + "epoch": 0.4, + "learning_rate": 0.00048553660982948846, + "loss": 1.5167, + "theoretical_loss": 3.5855116810641956, + "tokens_seen": 1317470208 + }, + { + "epoch": 0.4, + "learning_rate": 0.0004855205616850552, + "loss": 1.3405, + "theoretical_loss": 3.5854956898805543, + "tokens_seen": 1317535744 + }, + { + "epoch": 0.4, + "learning_rate": 0.00048550451354062186, + "loss": 1.3937, + "theoretical_loss": 3.5854796997150222, + "tokens_seen": 1317601280 + }, + { + "epoch": 0.4, + "learning_rate": 0.0004854884653961886, + "loss": 1.4779, + "theoretical_loss": 3.5854637105674834, + "tokens_seen": 1317666816 + }, + { + "epoch": 0.4, + "learning_rate": 0.00048547241725175526, + "loss": 1.2449, + "theoretical_loss": 3.5854477224378227, + "tokens_seen": 1317732352 + }, + { + "epoch": 0.4, + "learning_rate": 0.000485456369107322, + "loss": 1.514, + "theoretical_loss": 3.5854317353259244, + "tokens_seen": 1317797888 + }, + { + "epoch": 0.4, + "learning_rate": 0.00048544032096288866, + "loss": 1.3386, + "theoretical_loss": 3.585415749231674, + "tokens_seen": 1317863424 + }, + { + "epoch": 0.4, + "learning_rate": 0.0004854242728184554, + "loss": 1.4668, + "theoretical_loss": 3.585399764154955, + "tokens_seen": 1317928960 + }, + { + "epoch": 0.4, + "learning_rate": 0.00048540822467402206, + "loss": 1.4498, + "theoretical_loss": 3.585383780095653, + "tokens_seen": 1317994496 + }, + { + "epoch": 0.4, + "learning_rate": 0.0004853921765295888, + "loss": 1.4184, + "theoretical_loss": 3.585367797053652, + "tokens_seen": 1318060032 + }, + { + "epoch": 0.4, + "learning_rate": 0.00048537612838515546, + "loss": 1.3379, + "theoretical_loss": 3.5853518150288366, + "tokens_seen": 1318125568 + }, + { + "epoch": 0.4, + "learning_rate": 0.0004853600802407222, + "loss": 1.3978, + "theoretical_loss": 3.5853358340210923, + "tokens_seen": 1318191104 + }, + { + "epoch": 0.4, + "learning_rate": 0.00048534403209628886, + "loss": 1.469, + "theoretical_loss": 3.5853198540303035, + "tokens_seen": 1318256640 + }, + { + "epoch": 0.4, + "learning_rate": 0.0004853279839518556, + "loss": 1.3916, + "theoretical_loss": 3.585303875056354, + "tokens_seen": 1318322176 + }, + { + "epoch": 0.4, + "learning_rate": 0.00048531193580742226, + "loss": 1.3827, + "theoretical_loss": 3.5852878970991298, + "tokens_seen": 1318387712 + }, + { + "epoch": 0.4, + "learning_rate": 0.00048529588766298897, + "loss": 1.4186, + "theoretical_loss": 3.5852719201585153, + "tokens_seen": 1318453248 + }, + { + "epoch": 0.4, + "learning_rate": 0.0004852798395185557, + "loss": 1.3692, + "theoretical_loss": 3.5852559442343948, + "tokens_seen": 1318518784 + }, + { + "epoch": 0.4, + "learning_rate": 0.00048526379137412237, + "loss": 1.4734, + "theoretical_loss": 3.5852399693266532, + "tokens_seen": 1318584320 + }, + { + "epoch": 0.4, + "learning_rate": 0.0004852477432296891, + "loss": 1.2579, + "theoretical_loss": 3.585223995435176, + "tokens_seen": 1318649856 + }, + { + "epoch": 0.4, + "learning_rate": 0.00048523169508525577, + "loss": 1.504, + "theoretical_loss": 3.5852080225598475, + "tokens_seen": 1318715392 + }, + { + "epoch": 0.4, + "learning_rate": 0.0004852156469408225, + "loss": 1.3838, + "theoretical_loss": 3.585192050700553, + "tokens_seen": 1318780928 + }, + { + "epoch": 0.4, + "learning_rate": 0.00048519959879638917, + "loss": 1.3484, + "theoretical_loss": 3.5851760798571766, + "tokens_seen": 1318846464 + }, + { + "epoch": 0.4, + "learning_rate": 0.0004851835506519559, + "loss": 1.4092, + "theoretical_loss": 3.585160110029604, + "tokens_seen": 1318912000 + }, + { + "epoch": 0.4, + "learning_rate": 0.00048516750250752257, + "loss": 1.3369, + "theoretical_loss": 3.58514414121772, + "tokens_seen": 1318977536 + }, + { + "epoch": 0.4, + "learning_rate": 0.0004851514543630893, + "loss": 1.3078, + "theoretical_loss": 3.585128173421409, + "tokens_seen": 1319043072 + }, + { + "epoch": 0.4, + "learning_rate": 0.00048513540621865597, + "loss": 1.4983, + "theoretical_loss": 3.585112206640556, + "tokens_seen": 1319108608 + }, + { + "epoch": 0.4, + "learning_rate": 0.0004851193580742227, + "loss": 1.4071, + "theoretical_loss": 3.585096240875047, + "tokens_seen": 1319174144 + }, + { + "epoch": 0.4, + "learning_rate": 0.0004851033099297894, + "loss": 1.3784, + "theoretical_loss": 3.5850802761247658, + "tokens_seen": 1319239680 + }, + { + "epoch": 0.4, + "learning_rate": 0.00048508726178535613, + "loss": 1.3259, + "theoretical_loss": 3.585064312389598, + "tokens_seen": 1319305216 + }, + { + "epoch": 0.4, + "learning_rate": 0.0004850712136409228, + "loss": 1.3456, + "theoretical_loss": 3.585048349669429, + "tokens_seen": 1319370752 + }, + { + "epoch": 0.4, + "learning_rate": 0.00048505516549648953, + "loss": 1.4101, + "theoretical_loss": 3.5850323879641426, + "tokens_seen": 1319436288 + }, + { + "epoch": 0.4, + "learning_rate": 0.0004850391173520562, + "loss": 1.3071, + "theoretical_loss": 3.585016427273625, + "tokens_seen": 1319501824 + }, + { + "epoch": 0.4, + "learning_rate": 0.00048502306920762293, + "loss": 1.4655, + "theoretical_loss": 3.5850004675977614, + "tokens_seen": 1319567360 + }, + { + "epoch": 0.4, + "learning_rate": 0.0004850070210631896, + "loss": 1.4247, + "theoretical_loss": 3.5849845089364356, + "tokens_seen": 1319632896 + }, + { + "epoch": 0.4, + "learning_rate": 0.00048499097291875633, + "loss": 1.3834, + "theoretical_loss": 3.5849685512895344, + "tokens_seen": 1319698432 + }, + { + "epoch": 0.4, + "learning_rate": 0.000484974924774323, + "loss": 1.3804, + "theoretical_loss": 3.584952594656942, + "tokens_seen": 1319763968 + }, + { + "epoch": 0.4, + "learning_rate": 0.00048495887662988973, + "loss": 1.4424, + "theoretical_loss": 3.584936639038543, + "tokens_seen": 1319829504 + }, + { + "epoch": 0.4, + "learning_rate": 0.0004849428284854564, + "loss": 1.3945, + "theoretical_loss": 3.584920684434224, + "tokens_seen": 1319895040 + }, + { + "epoch": 0.4, + "learning_rate": 0.00048492678034102313, + "loss": 1.4659, + "theoretical_loss": 3.584904730843869, + "tokens_seen": 1319960576 + }, + { + "epoch": 0.4, + "learning_rate": 0.0004849107321965898, + "loss": 1.3263, + "theoretical_loss": 3.5848887782673637, + "tokens_seen": 1320026112 + }, + { + "epoch": 0.4, + "learning_rate": 0.00048489468405215654, + "loss": 1.2831, + "theoretical_loss": 3.5848728267045935, + "tokens_seen": 1320091648 + }, + { + "epoch": 0.4, + "learning_rate": 0.0004848786359077232, + "loss": 1.4346, + "theoretical_loss": 3.5848568761554436, + "tokens_seen": 1320157184 + }, + { + "epoch": 0.4, + "learning_rate": 0.00048486258776328994, + "loss": 1.2976, + "theoretical_loss": 3.5848409266197994, + "tokens_seen": 1320222720 + }, + { + "epoch": 0.4, + "learning_rate": 0.0004848465396188566, + "loss": 1.3547, + "theoretical_loss": 3.5848249780975454, + "tokens_seen": 1320288256 + }, + { + "epoch": 0.4, + "learning_rate": 0.00048483049147442334, + "loss": 1.3099, + "theoretical_loss": 3.584809030588568, + "tokens_seen": 1320353792 + }, + { + "epoch": 0.4, + "learning_rate": 0.00048481444332999, + "loss": 1.307, + "theoretical_loss": 3.5847930840927518, + "tokens_seen": 1320419328 + }, + { + "epoch": 0.4, + "learning_rate": 0.00048479839518555674, + "loss": 1.477, + "theoretical_loss": 3.584777138609982, + "tokens_seen": 1320484864 + }, + { + "epoch": 0.4, + "objective/train/docs_used": 483102, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.34702467918396, + "objective/train/theoretical_loss": 3.584761194140145, + "objective/train/tokens_used": 1341010400, + "theoretical_loss": 3.584761194140145, + "tokens_seen": 1320550400 + }, + { + "epoch": 0.4, + "learning_rate": 0.0004847823470411234, + "loss": 1.4091, + "theoretical_loss": 3.584761194140145, + "tokens_seen": 1320550400 + }, + { + "epoch": 0.4, + "learning_rate": 0.00048476629889669014, + "loss": 1.4267, + "theoretical_loss": 3.584745250683125, + "tokens_seen": 1320615936 + }, + { + "epoch": 0.4, + "learning_rate": 0.0004847502507522568, + "loss": 1.2856, + "theoretical_loss": 3.584729308238808, + "tokens_seen": 1320681472 + }, + { + "epoch": 0.4, + "learning_rate": 0.00048473420260782354, + "loss": 1.4072, + "theoretical_loss": 3.5847133668070796, + "tokens_seen": 1320747008 + }, + { + "epoch": 0.4, + "learning_rate": 0.0004847181544633902, + "loss": 1.4747, + "theoretical_loss": 3.584697426387825, + "tokens_seen": 1320812544 + }, + { + "epoch": 0.4, + "learning_rate": 0.00048470210631895694, + "loss": 1.3688, + "theoretical_loss": 3.58468148698093, + "tokens_seen": 1320878080 + }, + { + "epoch": 0.4, + "learning_rate": 0.0004846860581745236, + "loss": 1.4789, + "theoretical_loss": 3.5846655485862793, + "tokens_seen": 1320943616 + }, + { + "epoch": 0.4, + "learning_rate": 0.0004846700100300903, + "loss": 1.4289, + "theoretical_loss": 3.5846496112037585, + "tokens_seen": 1321009152 + }, + { + "epoch": 0.4, + "learning_rate": 0.000484653961885657, + "loss": 1.2817, + "theoretical_loss": 3.5846336748332543, + "tokens_seen": 1321074688 + }, + { + "epoch": 0.4, + "learning_rate": 0.0004846379137412237, + "loss": 1.4191, + "theoretical_loss": 3.584617739474651, + "tokens_seen": 1321140224 + }, + { + "epoch": 0.4, + "learning_rate": 0.0004846218655967904, + "loss": 1.4025, + "theoretical_loss": 3.5846018051278348, + "tokens_seen": 1321205760 + }, + { + "epoch": 0.4, + "learning_rate": 0.0004846058174523571, + "loss": 1.4746, + "theoretical_loss": 3.58458587179269, + "tokens_seen": 1321271296 + }, + { + "epoch": 0.4, + "learning_rate": 0.0004845897693079238, + "loss": 1.4607, + "theoretical_loss": 3.5845699394691044, + "tokens_seen": 1321336832 + }, + { + "epoch": 0.4, + "learning_rate": 0.0004845737211634905, + "loss": 1.5893, + "theoretical_loss": 3.584554008156962, + "tokens_seen": 1321402368 + }, + { + "epoch": 0.4, + "learning_rate": 0.0004845576730190572, + "loss": 1.4456, + "theoretical_loss": 3.5845380778561493, + "tokens_seen": 1321467904 + }, + { + "epoch": 0.4, + "learning_rate": 0.0004845416248746239, + "loss": 1.4511, + "theoretical_loss": 3.584522148566551, + "tokens_seen": 1321533440 + }, + { + "epoch": 0.4, + "learning_rate": 0.00048452557673019054, + "loss": 1.461, + "theoretical_loss": 3.5845062202880538, + "tokens_seen": 1321598976 + }, + { + "epoch": 0.4, + "learning_rate": 0.0004845095285857573, + "loss": 1.407, + "theoretical_loss": 3.5844902930205422, + "tokens_seen": 1321664512 + }, + { + "epoch": 0.4, + "learning_rate": 0.00048449348044132394, + "loss": 1.3547, + "theoretical_loss": 3.584474366763903, + "tokens_seen": 1321730048 + }, + { + "epoch": 0.4, + "learning_rate": 0.0004844774322968907, + "loss": 1.418, + "theoretical_loss": 3.584458441518022, + "tokens_seen": 1321795584 + }, + { + "epoch": 0.4, + "learning_rate": 0.00048446138415245734, + "loss": 1.4869, + "theoretical_loss": 3.584442517282784, + "tokens_seen": 1321861120 + }, + { + "epoch": 0.4, + "learning_rate": 0.0004844453360080241, + "loss": 1.4539, + "theoretical_loss": 3.584426594058075, + "tokens_seen": 1321926656 + }, + { + "epoch": 0.4, + "learning_rate": 0.00048442928786359075, + "loss": 1.3507, + "theoretical_loss": 3.5844106718437807, + "tokens_seen": 1321992192 + }, + { + "epoch": 0.4, + "learning_rate": 0.0004844132397191575, + "loss": 1.3572, + "theoretical_loss": 3.584394750639788, + "tokens_seen": 1322057728 + }, + { + "epoch": 0.4, + "learning_rate": 0.00048439719157472415, + "loss": 1.3656, + "theoretical_loss": 3.5843788304459814, + "tokens_seen": 1322123264 + }, + { + "epoch": 0.4, + "learning_rate": 0.0004843811434302909, + "loss": 1.4713, + "theoretical_loss": 3.5843629112622475, + "tokens_seen": 1322188800 + }, + { + "epoch": 0.4, + "learning_rate": 0.00048436509528585755, + "loss": 1.2142, + "theoretical_loss": 3.584346993088472, + "tokens_seen": 1322254336 + }, + { + "epoch": 0.4, + "learning_rate": 0.0004843490471414243, + "loss": 1.4131, + "theoretical_loss": 3.58433107592454, + "tokens_seen": 1322319872 + }, + { + "epoch": 0.4, + "learning_rate": 0.00048433299899699095, + "loss": 1.2876, + "theoretical_loss": 3.5843151597703393, + "tokens_seen": 1322385408 + }, + { + "epoch": 0.4, + "learning_rate": 0.0004843169508525577, + "loss": 1.4553, + "theoretical_loss": 3.5842992446257536, + "tokens_seen": 1322450944 + }, + { + "epoch": 0.4, + "learning_rate": 0.00048430090270812435, + "loss": 1.3725, + "theoretical_loss": 3.58428333049067, + "tokens_seen": 1322516480 + }, + { + "epoch": 0.4, + "learning_rate": 0.0004842848545636911, + "loss": 1.4067, + "theoretical_loss": 3.5842674173649742, + "tokens_seen": 1322582016 + }, + { + "epoch": 0.4, + "learning_rate": 0.00048426880641925775, + "loss": 1.4078, + "theoretical_loss": 3.5842515052485524, + "tokens_seen": 1322647552 + }, + { + "epoch": 0.4, + "learning_rate": 0.0004842527582748245, + "loss": 1.4506, + "theoretical_loss": 3.5842355941412904, + "tokens_seen": 1322713088 + }, + { + "epoch": 0.4, + "learning_rate": 0.00048423671013039115, + "loss": 1.4194, + "theoretical_loss": 3.5842196840430742, + "tokens_seen": 1322778624 + }, + { + "epoch": 0.4, + "learning_rate": 0.0004842206619859579, + "loss": 1.4053, + "theoretical_loss": 3.5842037749537896, + "tokens_seen": 1322844160 + }, + { + "epoch": 0.4, + "learning_rate": 0.00048420461384152455, + "loss": 1.4194, + "theoretical_loss": 3.5841878668733234, + "tokens_seen": 1322909696 + }, + { + "epoch": 0.4, + "learning_rate": 0.0004841885656970913, + "loss": 1.4502, + "theoretical_loss": 3.5841719598015604, + "tokens_seen": 1322975232 + }, + { + "epoch": 0.4, + "learning_rate": 0.00048417251755265796, + "loss": 1.442, + "theoretical_loss": 3.584156053738388, + "tokens_seen": 1323040768 + }, + { + "epoch": 0.4, + "learning_rate": 0.0004841564694082247, + "loss": 1.363, + "theoretical_loss": 3.5841401486836917, + "tokens_seen": 1323106304 + }, + { + "epoch": 0.4, + "learning_rate": 0.00048414042126379136, + "loss": 1.3791, + "theoretical_loss": 3.5841242446373576, + "tokens_seen": 1323171840 + }, + { + "epoch": 0.4, + "learning_rate": 0.0004841243731193581, + "loss": 1.3898, + "theoretical_loss": 3.584108341599271, + "tokens_seen": 1323237376 + }, + { + "epoch": 0.4, + "learning_rate": 0.00048410832497492476, + "loss": 1.4697, + "theoretical_loss": 3.5840924395693197, + "tokens_seen": 1323302912 + }, + { + "epoch": 0.4, + "learning_rate": 0.0004840922768304915, + "loss": 1.3187, + "theoretical_loss": 3.5840765385473894, + "tokens_seen": 1323368448 + }, + { + "epoch": 0.4, + "learning_rate": 0.00048407622868605827, + "loss": 1.4246, + "theoretical_loss": 3.584060638533365, + "tokens_seen": 1323433984 + }, + { + "epoch": 0.4, + "learning_rate": 0.0004840601805416249, + "loss": 1.3953, + "theoretical_loss": 3.584044739527134, + "tokens_seen": 1323499520 + }, + { + "epoch": 0.4, + "learning_rate": 0.0004840441323971916, + "loss": 1.5277, + "theoretical_loss": 3.584028841528583, + "tokens_seen": 1323565056 + }, + { + "epoch": 0.4, + "learning_rate": 0.0004840280842527583, + "loss": 1.3803, + "theoretical_loss": 3.5840129445375966, + "tokens_seen": 1323630592 + }, + { + "epoch": 0.4, + "learning_rate": 0.000484012036108325, + "loss": 1.4337, + "theoretical_loss": 3.583997048554062, + "tokens_seen": 1323696128 + }, + { + "epoch": 0.4, + "learning_rate": 0.0004839959879638917, + "loss": 1.413, + "theoretical_loss": 3.5839811535778656, + "tokens_seen": 1323761664 + }, + { + "epoch": 0.4, + "objective/train/docs_used": 484332, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.4408965110778809, + "objective/train/theoretical_loss": 3.583965259608894, + "objective/train/tokens_used": 1344287200, + "theoretical_loss": 3.583965259608894, + "tokens_seen": 1323827200 + }, + { + "epoch": 0.4, + "learning_rate": 0.0004839799398194584, + "loss": 1.4126, + "theoretical_loss": 3.583965259608894, + "tokens_seen": 1323827200 + }, + { + "epoch": 0.4, + "learning_rate": 0.0004839638916750251, + "loss": 1.341, + "theoretical_loss": 3.583949366647033, + "tokens_seen": 1323892736 + }, + { + "epoch": 0.4, + "learning_rate": 0.0004839478435305918, + "loss": 1.3866, + "theoretical_loss": 3.5839334746921683, + "tokens_seen": 1323958272 + }, + { + "epoch": 0.4, + "learning_rate": 0.0004839317953861585, + "loss": 1.3495, + "theoretical_loss": 3.5839175837441877, + "tokens_seen": 1324023808 + }, + { + "epoch": 0.4, + "learning_rate": 0.0004839157472417252, + "loss": 1.5057, + "theoretical_loss": 3.583901693802977, + "tokens_seen": 1324089344 + }, + { + "epoch": 0.4, + "learning_rate": 0.00048389969909729187, + "loss": 1.3387, + "theoretical_loss": 3.583885804868422, + "tokens_seen": 1324154880 + }, + { + "epoch": 0.4, + "learning_rate": 0.0004838836509528586, + "loss": 1.2956, + "theoretical_loss": 3.5838699169404094, + "tokens_seen": 1324220416 + }, + { + "epoch": 0.4, + "learning_rate": 0.00048386760280842527, + "loss": 1.5631, + "theoretical_loss": 3.583854030018826, + "tokens_seen": 1324285952 + }, + { + "epoch": 0.4, + "learning_rate": 0.000483851554663992, + "loss": 1.3837, + "theoretical_loss": 3.583838144103558, + "tokens_seen": 1324351488 + }, + { + "epoch": 0.4, + "learning_rate": 0.00048383550651955867, + "loss": 1.3787, + "theoretical_loss": 3.583822259194492, + "tokens_seen": 1324417024 + }, + { + "epoch": 0.4, + "learning_rate": 0.0004838194583751254, + "loss": 1.3492, + "theoretical_loss": 3.5838063752915144, + "tokens_seen": 1324482560 + }, + { + "epoch": 0.4, + "learning_rate": 0.00048380341023069207, + "loss": 1.52, + "theoretical_loss": 3.5837904923945114, + "tokens_seen": 1324548096 + }, + { + "epoch": 0.4, + "learning_rate": 0.0004837873620862588, + "loss": 1.383, + "theoretical_loss": 3.5837746105033697, + "tokens_seen": 1324613632 + }, + { + "epoch": 0.4, + "learning_rate": 0.00048377131394182547, + "loss": 1.4735, + "theoretical_loss": 3.5837587296179763, + "tokens_seen": 1324679168 + }, + { + "epoch": 0.4, + "learning_rate": 0.0004837552657973922, + "loss": 1.517, + "theoretical_loss": 3.5837428497382167, + "tokens_seen": 1324744704 + }, + { + "epoch": 0.4, + "learning_rate": 0.00048373921765295887, + "loss": 1.4635, + "theoretical_loss": 3.583726970863979, + "tokens_seen": 1324810240 + }, + { + "epoch": 0.4, + "learning_rate": 0.00048372316950852563, + "loss": 1.5117, + "theoretical_loss": 3.5837110929951486, + "tokens_seen": 1324875776 + }, + { + "epoch": 0.4, + "learning_rate": 0.0004837071213640923, + "loss": 1.3957, + "theoretical_loss": 3.5836952161316122, + "tokens_seen": 1324941312 + }, + { + "epoch": 0.4, + "learning_rate": 0.00048369107321965903, + "loss": 1.3936, + "theoretical_loss": 3.5836793402732567, + "tokens_seen": 1325006848 + }, + { + "epoch": 0.4, + "learning_rate": 0.0004836750250752257, + "loss": 1.3814, + "theoretical_loss": 3.5836634654199693, + "tokens_seen": 1325072384 + }, + { + "epoch": 0.4, + "learning_rate": 0.00048365897693079243, + "loss": 1.3678, + "theoretical_loss": 3.5836475915716357, + "tokens_seen": 1325137920 + }, + { + "epoch": 0.4, + "learning_rate": 0.0004836429287863591, + "loss": 1.4144, + "theoretical_loss": 3.5836317187281432, + "tokens_seen": 1325203456 + }, + { + "epoch": 0.4, + "learning_rate": 0.00048362688064192583, + "loss": 1.4325, + "theoretical_loss": 3.583615846889378, + "tokens_seen": 1325268992 + }, + { + "epoch": 0.4, + "learning_rate": 0.0004836108324974925, + "loss": 1.3914, + "theoretical_loss": 3.5835999760552273, + "tokens_seen": 1325334528 + }, + { + "epoch": 0.4, + "learning_rate": 0.00048359478435305923, + "loss": 1.4561, + "theoretical_loss": 3.5835841062255773, + "tokens_seen": 1325400064 + }, + { + "epoch": 0.4, + "learning_rate": 0.0004835787362086259, + "loss": 1.3429, + "theoretical_loss": 3.583568237400316, + "tokens_seen": 1325465600 + }, + { + "epoch": 0.4, + "learning_rate": 0.00048356268806419263, + "loss": 1.4083, + "theoretical_loss": 3.583552369579328, + "tokens_seen": 1325531136 + }, + { + "epoch": 0.4, + "learning_rate": 0.0004835466399197593, + "loss": 1.294, + "theoretical_loss": 3.5835365027625024, + "tokens_seen": 1325596672 + }, + { + "epoch": 0.4, + "learning_rate": 0.00048353059177532604, + "loss": 1.5029, + "theoretical_loss": 3.583520636949725, + "tokens_seen": 1325662208 + }, + { + "epoch": 0.4, + "learning_rate": 0.0004835145436308927, + "loss": 1.3485, + "theoretical_loss": 3.5835047721408824, + "tokens_seen": 1325727744 + }, + { + "epoch": 0.4, + "learning_rate": 0.00048349849548645944, + "loss": 1.4786, + "theoretical_loss": 3.583488908335861, + "tokens_seen": 1325793280 + }, + { + "epoch": 0.4, + "learning_rate": 0.0004834824473420261, + "loss": 1.4062, + "theoretical_loss": 3.583473045534549, + "tokens_seen": 1325858816 + }, + { + "epoch": 0.4, + "learning_rate": 0.00048346639919759284, + "loss": 1.2976, + "theoretical_loss": 3.583457183736833, + "tokens_seen": 1325924352 + }, + { + "epoch": 0.4, + "learning_rate": 0.0004834503510531595, + "loss": 1.3693, + "theoretical_loss": 3.583441322942599, + "tokens_seen": 1325989888 + }, + { + "epoch": 0.4, + "learning_rate": 0.00048343430290872624, + "loss": 1.4641, + "theoretical_loss": 3.583425463151735, + "tokens_seen": 1326055424 + }, + { + "epoch": 0.4, + "learning_rate": 0.0004834182547642929, + "loss": 1.3825, + "theoretical_loss": 3.5834096043641264, + "tokens_seen": 1326120960 + }, + { + "epoch": 0.4, + "learning_rate": 0.00048340220661985964, + "loss": 1.4486, + "theoretical_loss": 3.583393746579662, + "tokens_seen": 1326186496 + }, + { + "epoch": 0.4, + "learning_rate": 0.0004833861584754263, + "loss": 1.4534, + "theoretical_loss": 3.5833778897982276, + "tokens_seen": 1326252032 + }, + { + "epoch": 0.4, + "learning_rate": 0.00048337011033099304, + "loss": 1.4905, + "theoretical_loss": 3.5833620340197108, + "tokens_seen": 1326317568 + }, + { + "epoch": 0.4, + "learning_rate": 0.0004833540621865597, + "loss": 1.4792, + "theoretical_loss": 3.5833461792439985, + "tokens_seen": 1326383104 + }, + { + "epoch": 0.4, + "learning_rate": 0.00048333801404212644, + "loss": 1.4249, + "theoretical_loss": 3.5833303254709774, + "tokens_seen": 1326448640 + }, + { + "epoch": 0.4, + "learning_rate": 0.0004833219658976931, + "loss": 1.455, + "theoretical_loss": 3.5833144727005344, + "tokens_seen": 1326514176 + }, + { + "epoch": 0.4, + "learning_rate": 0.00048330591775325984, + "loss": 1.4461, + "theoretical_loss": 3.583298620932557, + "tokens_seen": 1326579712 + }, + { + "epoch": 0.4, + "learning_rate": 0.0004832898696088265, + "loss": 1.4175, + "theoretical_loss": 3.583282770166933, + "tokens_seen": 1326645248 + }, + { + "epoch": 0.4, + "learning_rate": 0.0004832738214643932, + "loss": 1.3509, + "theoretical_loss": 3.583266920403548, + "tokens_seen": 1326710784 + }, + { + "epoch": 0.4, + "learning_rate": 0.0004832577733199599, + "loss": 1.349, + "theoretical_loss": 3.58325107164229, + "tokens_seen": 1326776320 + }, + { + "epoch": 0.4, + "learning_rate": 0.0004832417251755266, + "loss": 1.4276, + "theoretical_loss": 3.583235223883046, + "tokens_seen": 1326841856 + }, + { + "epoch": 0.4, + "learning_rate": 0.0004832256770310933, + "loss": 1.4234, + "theoretical_loss": 3.5832193771257033, + "tokens_seen": 1326907392 + }, + { + "epoch": 0.4, + "learning_rate": 0.00048320962888666, + "loss": 1.4415, + "theoretical_loss": 3.583203531370149, + "tokens_seen": 1326972928 + }, + { + "epoch": 0.4, + "learning_rate": 0.0004831935807422267, + "loss": 1.3972, + "theoretical_loss": 3.5831876866162693, + "tokens_seen": 1327038464 + }, + { + "epoch": 0.4, + "objective/train/docs_used": 485470, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.4843668937683105, + "objective/train/theoretical_loss": 3.5831718428639534, + "objective/train/tokens_used": 1347564000, + "theoretical_loss": 3.5831718428639534, + "tokens_seen": 1327104000 + }, + { + "epoch": 0.4, + "learning_rate": 0.0004831775325977934, + "loss": 1.5184, + "theoretical_loss": 3.5831718428639534, + "tokens_seen": 1327104000 + }, + { + "epoch": 0.4, + "learning_rate": 0.0004831614844533601, + "loss": 1.3178, + "theoretical_loss": 3.583156000113087, + "tokens_seen": 1327169536 + }, + { + "epoch": 0.4, + "learning_rate": 0.0004831454363089268, + "loss": 1.4716, + "theoretical_loss": 3.583140158363558, + "tokens_seen": 1327235072 + }, + { + "epoch": 0.4, + "learning_rate": 0.00048312938816449344, + "loss": 1.3703, + "theoretical_loss": 3.5831243176152534, + "tokens_seen": 1327300608 + }, + { + "epoch": 0.4, + "learning_rate": 0.0004831133400200602, + "loss": 1.536, + "theoretical_loss": 3.583108477868061, + "tokens_seen": 1327366144 + }, + { + "epoch": 0.4, + "learning_rate": 0.00048309729187562684, + "loss": 1.3126, + "theoretical_loss": 3.583092639121867, + "tokens_seen": 1327431680 + }, + { + "epoch": 0.4, + "learning_rate": 0.0004830812437311936, + "loss": 1.3425, + "theoretical_loss": 3.58307680137656, + "tokens_seen": 1327497216 + }, + { + "epoch": 0.4, + "learning_rate": 0.00048306519558676025, + "loss": 1.4186, + "theoretical_loss": 3.5830609646320264, + "tokens_seen": 1327562752 + }, + { + "epoch": 0.4, + "learning_rate": 0.000483049147442327, + "loss": 1.465, + "theoretical_loss": 3.5830451288881537, + "tokens_seen": 1327628288 + }, + { + "epoch": 0.4, + "learning_rate": 0.00048303309929789365, + "loss": 1.4904, + "theoretical_loss": 3.58302929414483, + "tokens_seen": 1327693824 + }, + { + "epoch": 0.4, + "learning_rate": 0.0004830170511534604, + "loss": 1.3313, + "theoretical_loss": 3.583013460401942, + "tokens_seen": 1327759360 + }, + { + "epoch": 0.4, + "learning_rate": 0.00048300100300902705, + "loss": 1.3868, + "theoretical_loss": 3.5829976276593776, + "tokens_seen": 1327824896 + }, + { + "epoch": 0.4, + "learning_rate": 0.0004829849548645938, + "loss": 1.4407, + "theoretical_loss": 3.582981795917024, + "tokens_seen": 1327890432 + }, + { + "epoch": 0.4, + "learning_rate": 0.00048296890672016045, + "loss": 1.4529, + "theoretical_loss": 3.5829659651747683, + "tokens_seen": 1327955968 + }, + { + "epoch": 0.4, + "learning_rate": 0.0004829528585757272, + "loss": 1.4562, + "theoretical_loss": 3.5829501354324984, + "tokens_seen": 1328021504 + }, + { + "epoch": 0.4, + "learning_rate": 0.00048293681043129385, + "loss": 1.4153, + "theoretical_loss": 3.5829343066901016, + "tokens_seen": 1328087040 + }, + { + "epoch": 0.4, + "learning_rate": 0.0004829207622868606, + "loss": 1.3654, + "theoretical_loss": 3.582918478947466, + "tokens_seen": 1328152576 + }, + { + "epoch": 0.4, + "learning_rate": 0.00048290471414242725, + "loss": 1.4424, + "theoretical_loss": 3.582902652204478, + "tokens_seen": 1328218112 + }, + { + "epoch": 0.4, + "learning_rate": 0.000482888665997994, + "loss": 1.3389, + "theoretical_loss": 3.582886826461026, + "tokens_seen": 1328283648 + }, + { + "epoch": 0.4, + "learning_rate": 0.00048287261785356076, + "loss": 1.3758, + "theoretical_loss": 3.582871001716997, + "tokens_seen": 1328349184 + }, + { + "epoch": 0.4, + "learning_rate": 0.0004828565697091274, + "loss": 1.4848, + "theoretical_loss": 3.582855177972279, + "tokens_seen": 1328414720 + }, + { + "epoch": 0.4, + "learning_rate": 0.00048284052156469416, + "loss": 1.3913, + "theoretical_loss": 3.58283935522676, + "tokens_seen": 1328480256 + }, + { + "epoch": 0.4, + "learning_rate": 0.0004828244734202608, + "loss": 1.3954, + "theoretical_loss": 3.582823533480327, + "tokens_seen": 1328545792 + }, + { + "epoch": 0.4, + "learning_rate": 0.00048280842527582756, + "loss": 1.3375, + "theoretical_loss": 3.5828077127328672, + "tokens_seen": 1328611328 + }, + { + "epoch": 0.4, + "learning_rate": 0.0004827923771313942, + "loss": 1.3555, + "theoretical_loss": 3.5827918929842695, + "tokens_seen": 1328676864 + }, + { + "epoch": 0.4, + "learning_rate": 0.00048277632898696097, + "loss": 1.3706, + "theoretical_loss": 3.5827760742344203, + "tokens_seen": 1328742400 + }, + { + "epoch": 0.4, + "learning_rate": 0.0004827602808425276, + "loss": 1.4704, + "theoretical_loss": 3.5827602564832084, + "tokens_seen": 1328807936 + }, + { + "epoch": 0.4, + "learning_rate": 0.00048274423269809437, + "loss": 1.4137, + "theoretical_loss": 3.582744439730521, + "tokens_seen": 1328873472 + }, + { + "epoch": 0.4, + "learning_rate": 0.000482728184553661, + "loss": 1.4162, + "theoretical_loss": 3.5827286239762453, + "tokens_seen": 1328939008 + }, + { + "epoch": 0.4, + "learning_rate": 0.00048271213640922777, + "loss": 1.394, + "theoretical_loss": 3.58271280922027, + "tokens_seen": 1329004544 + }, + { + "epoch": 0.4, + "learning_rate": 0.0004826960882647944, + "loss": 1.4979, + "theoretical_loss": 3.582696995462482, + "tokens_seen": 1329070080 + }, + { + "epoch": 0.4, + "learning_rate": 0.00048268004012036117, + "loss": 1.3735, + "theoretical_loss": 3.5826811827027703, + "tokens_seen": 1329135616 + }, + { + "epoch": 0.4, + "learning_rate": 0.0004826639919759278, + "loss": 1.3606, + "theoretical_loss": 3.582665370941021, + "tokens_seen": 1329201152 + }, + { + "epoch": 0.4, + "learning_rate": 0.0004826479438314945, + "loss": 1.5097, + "theoretical_loss": 3.5826495601771233, + "tokens_seen": 1329266688 + }, + { + "epoch": 0.4, + "learning_rate": 0.0004826318956870612, + "loss": 1.3064, + "theoretical_loss": 3.582633750410964, + "tokens_seen": 1329332224 + }, + { + "epoch": 0.4, + "learning_rate": 0.0004826158475426279, + "loss": 1.3946, + "theoretical_loss": 3.5826179416424324, + "tokens_seen": 1329397760 + }, + { + "epoch": 0.4, + "learning_rate": 0.0004825997993981946, + "loss": 1.2917, + "theoretical_loss": 3.582602133871415, + "tokens_seen": 1329463296 + }, + { + "epoch": 0.4, + "learning_rate": 0.0004825837512537613, + "loss": 1.4648, + "theoretical_loss": 3.5825863270978, + "tokens_seen": 1329528832 + }, + { + "epoch": 0.4, + "learning_rate": 0.000482567703109328, + "loss": 1.3778, + "theoretical_loss": 3.582570521321476, + "tokens_seen": 1329594368 + }, + { + "epoch": 0.4, + "learning_rate": 0.0004825516549648947, + "loss": 1.322, + "theoretical_loss": 3.58255471654233, + "tokens_seen": 1329659904 + }, + { + "epoch": 0.4, + "learning_rate": 0.0004825356068204614, + "loss": 1.3113, + "theoretical_loss": 3.58253891276025, + "tokens_seen": 1329725440 + }, + { + "epoch": 0.4, + "learning_rate": 0.0004825195586760281, + "loss": 1.2974, + "theoretical_loss": 3.582523109975125, + "tokens_seen": 1329790976 + }, + { + "epoch": 0.4, + "learning_rate": 0.00048250351053159477, + "loss": 1.4158, + "theoretical_loss": 3.582507308186842, + "tokens_seen": 1329856512 + }, + { + "epoch": 0.4, + "learning_rate": 0.0004824874623871615, + "loss": 1.3636, + "theoretical_loss": 3.5824915073952894, + "tokens_seen": 1329922048 + }, + { + "epoch": 0.4, + "learning_rate": 0.00048247141424272817, + "loss": 1.4833, + "theoretical_loss": 3.582475707600355, + "tokens_seen": 1329987584 + }, + { + "epoch": 0.4, + "learning_rate": 0.0004824553660982949, + "loss": 1.3254, + "theoretical_loss": 3.582459908801927, + "tokens_seen": 1330053120 + }, + { + "epoch": 0.4, + "learning_rate": 0.00048243931795386157, + "loss": 1.3457, + "theoretical_loss": 3.5824441109998935, + "tokens_seen": 1330118656 + }, + { + "epoch": 0.4, + "learning_rate": 0.0004824232698094283, + "loss": 1.4396, + "theoretical_loss": 3.582428314194142, + "tokens_seen": 1330184192 + }, + { + "epoch": 0.4, + "learning_rate": 0.00048240722166499497, + "loss": 1.344, + "theoretical_loss": 3.5824125183845617, + "tokens_seen": 1330249728 + }, + { + "epoch": 0.4, + "learning_rate": 0.0004823911735205617, + "loss": 1.4541, + "theoretical_loss": 3.5823967235710397, + "tokens_seen": 1330315264 + }, + { + "epoch": 0.4, + "objective/train/docs_used": 486514, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.416332483291626, + "objective/train/theoretical_loss": 3.582380929753465, + "objective/train/tokens_used": 1350840800, + "theoretical_loss": 3.582380929753465, + "tokens_seen": 1330380800 + }, + { + "epoch": 0.4, + "learning_rate": 0.00048237512537612837, + "loss": 1.3599, + "theoretical_loss": 3.582380929753465, + "tokens_seen": 1330380800 + }, + { + "epoch": 0.4, + "learning_rate": 0.00048235907723169513, + "loss": 1.4943, + "theoretical_loss": 3.5823651369317244, + "tokens_seen": 1330446336 + }, + { + "epoch": 0.4, + "learning_rate": 0.0004823430290872618, + "loss": 1.4323, + "theoretical_loss": 3.5823493451057074, + "tokens_seen": 1330511872 + }, + { + "epoch": 0.4, + "learning_rate": 0.00048232698094282853, + "loss": 1.3654, + "theoretical_loss": 3.5823335542753014, + "tokens_seen": 1330577408 + }, + { + "epoch": 0.4, + "learning_rate": 0.0004823109327983952, + "loss": 1.3567, + "theoretical_loss": 3.582317764440395, + "tokens_seen": 1330642944 + }, + { + "epoch": 0.4, + "learning_rate": 0.00048229488465396193, + "loss": 1.4454, + "theoretical_loss": 3.5823019756008767, + "tokens_seen": 1330708480 + }, + { + "epoch": 0.4, + "learning_rate": 0.0004822788365095286, + "loss": 1.4409, + "theoretical_loss": 3.582286187756634, + "tokens_seen": 1330774016 + }, + { + "epoch": 0.4, + "learning_rate": 0.00048226278836509533, + "loss": 1.375, + "theoretical_loss": 3.5822704009075554, + "tokens_seen": 1330839552 + }, + { + "epoch": 0.4, + "learning_rate": 0.000482246740220662, + "loss": 1.3637, + "theoretical_loss": 3.5822546150535293, + "tokens_seen": 1330905088 + }, + { + "epoch": 0.4, + "learning_rate": 0.00048223069207622873, + "loss": 1.4716, + "theoretical_loss": 3.582238830194444, + "tokens_seen": 1330970624 + }, + { + "epoch": 0.4, + "learning_rate": 0.0004822146439317954, + "loss": 1.461, + "theoretical_loss": 3.5822230463301876, + "tokens_seen": 1331036160 + }, + { + "epoch": 0.4, + "learning_rate": 0.00048219859578736213, + "loss": 1.3552, + "theoretical_loss": 3.5822072634606483, + "tokens_seen": 1331101696 + }, + { + "epoch": 0.4, + "learning_rate": 0.0004821825476429288, + "loss": 1.4182, + "theoretical_loss": 3.582191481585715, + "tokens_seen": 1331167232 + }, + { + "epoch": 0.4, + "learning_rate": 0.00048216649949849553, + "loss": 1.4232, + "theoretical_loss": 3.5821757007052764, + "tokens_seen": 1331232768 + }, + { + "epoch": 0.4, + "learning_rate": 0.0004821504513540622, + "loss": 1.3898, + "theoretical_loss": 3.5821599208192194, + "tokens_seen": 1331298304 + }, + { + "epoch": 0.4, + "learning_rate": 0.00048213440320962894, + "loss": 1.3604, + "theoretical_loss": 3.5821441419274334, + "tokens_seen": 1331363840 + }, + { + "epoch": 0.4, + "learning_rate": 0.0004821183550651956, + "loss": 1.4428, + "theoretical_loss": 3.582128364029807, + "tokens_seen": 1331429376 + }, + { + "epoch": 0.4, + "learning_rate": 0.00048210230692076234, + "loss": 1.4803, + "theoretical_loss": 3.582112587126228, + "tokens_seen": 1331494912 + }, + { + "epoch": 0.4, + "learning_rate": 0.000482086258776329, + "loss": 1.4088, + "theoretical_loss": 3.582096811216585, + "tokens_seen": 1331560448 + }, + { + "epoch": 0.4, + "learning_rate": 0.00048207021063189574, + "loss": 1.3666, + "theoretical_loss": 3.5820810363007665, + "tokens_seen": 1331625984 + }, + { + "epoch": 0.4, + "learning_rate": 0.0004820541624874624, + "loss": 1.3595, + "theoretical_loss": 3.5820652623786615, + "tokens_seen": 1331691520 + }, + { + "epoch": 0.4, + "learning_rate": 0.00048203811434302914, + "loss": 1.4374, + "theoretical_loss": 3.582049489450158, + "tokens_seen": 1331757056 + }, + { + "epoch": 0.4, + "learning_rate": 0.0004820220661985958, + "loss": 1.4717, + "theoretical_loss": 3.582033717515144, + "tokens_seen": 1331822592 + }, + { + "epoch": 0.4, + "learning_rate": 0.00048200601805416254, + "loss": 1.4435, + "theoretical_loss": 3.58201794657351, + "tokens_seen": 1331888128 + }, + { + "epoch": 0.4, + "learning_rate": 0.0004819899699097292, + "loss": 1.3469, + "theoretical_loss": 3.582002176625142, + "tokens_seen": 1331953664 + }, + { + "epoch": 0.4, + "learning_rate": 0.00048197392176529594, + "loss": 1.3391, + "theoretical_loss": 3.5819864076699304, + "tokens_seen": 1332019200 + }, + { + "epoch": 0.4, + "learning_rate": 0.0004819578736208626, + "loss": 1.2444, + "theoretical_loss": 3.581970639707763, + "tokens_seen": 1332084736 + }, + { + "epoch": 0.4, + "learning_rate": 0.00048194182547642934, + "loss": 1.3398, + "theoretical_loss": 3.5819548727385286, + "tokens_seen": 1332150272 + }, + { + "epoch": 0.4, + "learning_rate": 0.000481925777331996, + "loss": 1.432, + "theoretical_loss": 3.581939106762116, + "tokens_seen": 1332215808 + }, + { + "epoch": 0.4, + "learning_rate": 0.00048190972918756275, + "loss": 1.29, + "theoretical_loss": 3.5819233417784133, + "tokens_seen": 1332281344 + }, + { + "epoch": 0.4, + "learning_rate": 0.0004818936810431294, + "loss": 1.5632, + "theoretical_loss": 3.5819075777873097, + "tokens_seen": 1332346880 + }, + { + "epoch": 0.4, + "learning_rate": 0.0004818776328986961, + "loss": 1.3898, + "theoretical_loss": 3.581891814788694, + "tokens_seen": 1332412416 + }, + { + "epoch": 0.4, + "learning_rate": 0.0004818615847542628, + "loss": 1.4116, + "theoretical_loss": 3.5818760527824542, + "tokens_seen": 1332477952 + }, + { + "epoch": 0.4, + "learning_rate": 0.0004818455366098295, + "loss": 1.4755, + "theoretical_loss": 3.5818602917684794, + "tokens_seen": 1332543488 + }, + { + "epoch": 0.4, + "learning_rate": 0.0004818294884653962, + "loss": 1.3778, + "theoretical_loss": 3.581844531746659, + "tokens_seen": 1332609024 + }, + { + "epoch": 0.4, + "learning_rate": 0.0004818134403209629, + "loss": 1.3867, + "theoretical_loss": 3.5818287727168805, + "tokens_seen": 1332674560 + }, + { + "epoch": 0.4, + "learning_rate": 0.0004817973921765296, + "loss": 1.4314, + "theoretical_loss": 3.5818130146790335, + "tokens_seen": 1332740096 + }, + { + "epoch": 0.4, + "learning_rate": 0.0004817813440320963, + "loss": 1.3592, + "theoretical_loss": 3.581797257633007, + "tokens_seen": 1332805632 + }, + { + "epoch": 0.4, + "learning_rate": 0.000481765295887663, + "loss": 1.283, + "theoretical_loss": 3.581781501578689, + "tokens_seen": 1332871168 + }, + { + "epoch": 0.4, + "learning_rate": 0.0004817492477432297, + "loss": 1.4233, + "theoretical_loss": 3.581765746515969, + "tokens_seen": 1332936704 + }, + { + "epoch": 0.4, + "learning_rate": 0.00048173319959879634, + "loss": 1.4631, + "theoretical_loss": 3.5817499924447356, + "tokens_seen": 1333002240 + }, + { + "epoch": 0.4, + "learning_rate": 0.0004817171514543631, + "loss": 1.4728, + "theoretical_loss": 3.5817342393648772, + "tokens_seen": 1333067776 + }, + { + "epoch": 0.4, + "learning_rate": 0.00048170110330992974, + "loss": 1.3819, + "theoretical_loss": 3.5817184872762833, + "tokens_seen": 1333133312 + }, + { + "epoch": 0.4, + "learning_rate": 0.0004816850551654965, + "loss": 1.3501, + "theoretical_loss": 3.581702736178843, + "tokens_seen": 1333198848 + }, + { + "epoch": 0.4, + "learning_rate": 0.00048166900702106325, + "loss": 1.4788, + "theoretical_loss": 3.5816869860724445, + "tokens_seen": 1333264384 + }, + { + "epoch": 0.4, + "learning_rate": 0.0004816529588766299, + "loss": 1.2573, + "theoretical_loss": 3.581671236956977, + "tokens_seen": 1333329920 + }, + { + "epoch": 0.4, + "learning_rate": 0.00048163691073219666, + "loss": 1.4568, + "theoretical_loss": 3.5816554888323298, + "tokens_seen": 1333395456 + }, + { + "epoch": 0.4, + "learning_rate": 0.0004816208625877633, + "loss": 1.3891, + "theoretical_loss": 3.5816397416983916, + "tokens_seen": 1333460992 + }, + { + "epoch": 0.4, + "learning_rate": 0.00048160481444333006, + "loss": 1.3996, + "theoretical_loss": 3.5816239955550513, + "tokens_seen": 1333526528 + }, + { + "epoch": 0.4, + "learning_rate": 0.0004815887662988967, + "loss": 1.3731, + "theoretical_loss": 3.581608250402198, + "tokens_seen": 1333592064 + }, + { + "epoch": 0.4, + "objective/train/docs_used": 487665, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.3533046245574951, + "objective/train/theoretical_loss": 3.581592506239721, + "objective/train/tokens_used": 1354117600, + "theoretical_loss": 3.581592506239721, + "tokens_seen": 1333657600 + }, + { + "epoch": 0.4, + "learning_rate": 0.00048157271815446346, + "loss": 1.3581, + "theoretical_loss": 3.581592506239721, + "tokens_seen": 1333657600 + }, + { + "epoch": 0.4, + "learning_rate": 0.0004815566700100301, + "loss": 1.46, + "theoretical_loss": 3.581576763067509, + "tokens_seen": 1333723136 + }, + { + "epoch": 0.4, + "learning_rate": 0.00048154062186559686, + "loss": 1.4343, + "theoretical_loss": 3.581561020885451, + "tokens_seen": 1333788672 + }, + { + "epoch": 0.4, + "learning_rate": 0.0004815245737211635, + "loss": 1.4159, + "theoretical_loss": 3.5815452796934357, + "tokens_seen": 1333854208 + }, + { + "epoch": 0.4, + "learning_rate": 0.00048150852557673026, + "loss": 1.3467, + "theoretical_loss": 3.5815295394913536, + "tokens_seen": 1333919744 + }, + { + "epoch": 0.4, + "learning_rate": 0.0004814924774322969, + "loss": 1.2999, + "theoretical_loss": 3.5815138002790925, + "tokens_seen": 1333985280 + }, + { + "epoch": 0.4, + "learning_rate": 0.00048147642928786366, + "loss": 1.2982, + "theoretical_loss": 3.581498062056542, + "tokens_seen": 1334050816 + }, + { + "epoch": 0.4, + "learning_rate": 0.0004814603811434303, + "loss": 1.3406, + "theoretical_loss": 3.581482324823591, + "tokens_seen": 1334116352 + }, + { + "epoch": 0.4, + "learning_rate": 0.00048144433299899706, + "loss": 1.4182, + "theoretical_loss": 3.5814665885801293, + "tokens_seen": 1334181888 + }, + { + "epoch": 0.4, + "learning_rate": 0.0004814282848545637, + "loss": 1.5007, + "theoretical_loss": 3.581450853326045, + "tokens_seen": 1334247424 + }, + { + "epoch": 0.4, + "learning_rate": 0.00048141223671013046, + "loss": 1.4763, + "theoretical_loss": 3.5814351190612284, + "tokens_seen": 1334312960 + }, + { + "epoch": 0.4, + "learning_rate": 0.0004813961885656971, + "loss": 1.4672, + "theoretical_loss": 3.5814193857855683, + "tokens_seen": 1334378496 + }, + { + "epoch": 0.4, + "learning_rate": 0.00048138014042126387, + "loss": 1.3474, + "theoretical_loss": 3.581403653498954, + "tokens_seen": 1334444032 + }, + { + "epoch": 0.4, + "learning_rate": 0.0004813640922768305, + "loss": 1.3532, + "theoretical_loss": 3.5813879222012748, + "tokens_seen": 1334509568 + }, + { + "epoch": 0.4, + "learning_rate": 0.00048134804413239727, + "loss": 1.4404, + "theoretical_loss": 3.5813721918924193, + "tokens_seen": 1334575104 + }, + { + "epoch": 0.4, + "learning_rate": 0.0004813319959879639, + "loss": 1.4018, + "theoretical_loss": 3.581356462572278, + "tokens_seen": 1334640640 + }, + { + "epoch": 0.4, + "learning_rate": 0.00048131594784353067, + "loss": 1.4571, + "theoretical_loss": 3.5813407342407384, + "tokens_seen": 1334706176 + }, + { + "epoch": 0.4, + "learning_rate": 0.0004812998996990973, + "loss": 1.4301, + "theoretical_loss": 3.581325006897692, + "tokens_seen": 1334771712 + }, + { + "epoch": 0.4, + "learning_rate": 0.00048128385155466407, + "loss": 1.3111, + "theoretical_loss": 3.581309280543027, + "tokens_seen": 1334837248 + }, + { + "epoch": 0.4, + "learning_rate": 0.0004812678034102307, + "loss": 1.3018, + "theoretical_loss": 3.5812935551766327, + "tokens_seen": 1334902784 + }, + { + "epoch": 0.4, + "learning_rate": 0.0004812517552657974, + "loss": 1.3322, + "theoretical_loss": 3.581277830798399, + "tokens_seen": 1334968320 + }, + { + "epoch": 0.4, + "learning_rate": 0.0004812357071213641, + "loss": 1.3364, + "theoretical_loss": 3.5812621074082145, + "tokens_seen": 1335033856 + }, + { + "epoch": 0.4, + "learning_rate": 0.0004812196589769308, + "loss": 1.4394, + "theoretical_loss": 3.5812463850059695, + "tokens_seen": 1335099392 + }, + { + "epoch": 0.4, + "learning_rate": 0.0004812036108324975, + "loss": 1.4209, + "theoretical_loss": 3.5812306635915525, + "tokens_seen": 1335164928 + }, + { + "epoch": 0.4, + "learning_rate": 0.0004811875626880642, + "loss": 1.4892, + "theoretical_loss": 3.581214943164854, + "tokens_seen": 1335230464 + }, + { + "epoch": 0.4, + "learning_rate": 0.0004811715145436309, + "loss": 1.2953, + "theoretical_loss": 3.581199223725762, + "tokens_seen": 1335296000 + }, + { + "epoch": 0.4, + "learning_rate": 0.0004811554663991976, + "loss": 1.3309, + "theoretical_loss": 3.5811835052741685, + "tokens_seen": 1335361536 + }, + { + "epoch": 0.4, + "learning_rate": 0.0004811394182547643, + "loss": 1.3578, + "theoretical_loss": 3.58116778780996, + "tokens_seen": 1335427072 + }, + { + "epoch": 0.4, + "learning_rate": 0.000481123370110331, + "loss": 1.4188, + "theoretical_loss": 3.5811520713330283, + "tokens_seen": 1335492608 + }, + { + "epoch": 0.4, + "learning_rate": 0.00048110732196589767, + "loss": 1.278, + "theoretical_loss": 3.5811363558432623, + "tokens_seen": 1335558144 + }, + { + "epoch": 0.4, + "learning_rate": 0.0004810912738214644, + "loss": 1.4406, + "theoretical_loss": 3.5811206413405507, + "tokens_seen": 1335623680 + }, + { + "epoch": 0.4, + "learning_rate": 0.00048107522567703107, + "loss": 1.333, + "theoretical_loss": 3.581104927824784, + "tokens_seen": 1335689216 + }, + { + "epoch": 0.4, + "learning_rate": 0.0004810591775325978, + "loss": 1.3417, + "theoretical_loss": 3.581089215295852, + "tokens_seen": 1335754752 + }, + { + "epoch": 0.4, + "learning_rate": 0.00048104312938816447, + "loss": 1.2867, + "theoretical_loss": 3.581073503753643, + "tokens_seen": 1335820288 + }, + { + "epoch": 0.4, + "learning_rate": 0.0004810270812437312, + "loss": 1.3487, + "theoretical_loss": 3.5810577931980476, + "tokens_seen": 1335885824 + }, + { + "epoch": 0.4, + "learning_rate": 0.00048101103309929787, + "loss": 1.3302, + "theoretical_loss": 3.5810420836289554, + "tokens_seen": 1335951360 + }, + { + "epoch": 0.4, + "learning_rate": 0.0004809949849548646, + "loss": 1.3096, + "theoretical_loss": 3.5810263750462568, + "tokens_seen": 1336016896 + }, + { + "epoch": 0.4, + "learning_rate": 0.0004809789368104313, + "loss": 1.3895, + "theoretical_loss": 3.58101066744984, + "tokens_seen": 1336082432 + }, + { + "epoch": 0.4, + "learning_rate": 0.00048096288866599803, + "loss": 1.3321, + "theoretical_loss": 3.580994960839595, + "tokens_seen": 1336147968 + }, + { + "epoch": 0.4, + "learning_rate": 0.0004809468405215647, + "loss": 1.2822, + "theoretical_loss": 3.5809792552154125, + "tokens_seen": 1336213504 + }, + { + "epoch": 0.4, + "learning_rate": 0.00048093079237713143, + "loss": 1.2996, + "theoretical_loss": 3.5809635505771817, + "tokens_seen": 1336279040 + }, + { + "epoch": 0.4, + "learning_rate": 0.0004809147442326981, + "loss": 1.4189, + "theoretical_loss": 3.5809478469247917, + "tokens_seen": 1336344576 + }, + { + "epoch": 0.4, + "learning_rate": 0.00048089869608826483, + "loss": 1.2699, + "theoretical_loss": 3.580932144258133, + "tokens_seen": 1336410112 + }, + { + "epoch": 0.4, + "learning_rate": 0.0004808826479438315, + "loss": 1.3561, + "theoretical_loss": 3.580916442577096, + "tokens_seen": 1336475648 + }, + { + "epoch": 0.41, + "learning_rate": 0.00048086659979939823, + "loss": 1.4584, + "theoretical_loss": 3.5809007418815693, + "tokens_seen": 1336541184 + }, + { + "epoch": 0.41, + "learning_rate": 0.0004808505516549649, + "loss": 1.382, + "theoretical_loss": 3.580885042171443, + "tokens_seen": 1336606720 + }, + { + "epoch": 0.41, + "learning_rate": 0.00048083450351053163, + "loss": 1.3973, + "theoretical_loss": 3.5808693434466075, + "tokens_seen": 1336672256 + }, + { + "epoch": 0.41, + "learning_rate": 0.0004808184553660983, + "loss": 1.3068, + "theoretical_loss": 3.580853645706952, + "tokens_seen": 1336737792 + }, + { + "epoch": 0.41, + "learning_rate": 0.00048080240722166503, + "loss": 1.3875, + "theoretical_loss": 3.580837948952367, + "tokens_seen": 1336803328 + }, + { + "epoch": 0.41, + "learning_rate": 0.0004807863590772317, + "loss": 1.5675, + "theoretical_loss": 3.580822253182742, + "tokens_seen": 1336868864 + }, + { + "epoch": 0.41, + "objective/train/docs_used": 488954, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.2208714485168457, + "objective/train/theoretical_loss": 3.5808065583979665, + "objective/train/tokens_used": 1357394400, + "theoretical_loss": 3.5808065583979665, + "tokens_seen": 1336934400 + }, + { + "epoch": 0.41, + "learning_rate": 0.00048077031093279844, + "loss": 1.3412, + "theoretical_loss": 3.5808065583979665, + "tokens_seen": 1336934400 + }, + { + "epoch": 0.41, + "learning_rate": 0.0004807542627883651, + "loss": 1.3724, + "theoretical_loss": 3.5807908645979314, + "tokens_seen": 1336999936 + }, + { + "epoch": 0.41, + "learning_rate": 0.00048073821464393184, + "loss": 1.4052, + "theoretical_loss": 3.580775171782526, + "tokens_seen": 1337065472 + }, + { + "epoch": 0.41, + "learning_rate": 0.0004807221664994985, + "loss": 1.3886, + "theoretical_loss": 3.5807594799516402, + "tokens_seen": 1337131008 + }, + { + "epoch": 0.41, + "learning_rate": 0.00048070611835506524, + "loss": 1.3809, + "theoretical_loss": 3.5807437891051643, + "tokens_seen": 1337196544 + }, + { + "epoch": 0.41, + "learning_rate": 0.0004806900702106319, + "loss": 1.3167, + "theoretical_loss": 3.5807280992429886, + "tokens_seen": 1337262080 + }, + { + "epoch": 0.41, + "learning_rate": 0.00048067402206619864, + "loss": 1.3134, + "theoretical_loss": 3.5807124103650025, + "tokens_seen": 1337327616 + }, + { + "epoch": 0.41, + "learning_rate": 0.0004806579739217653, + "loss": 1.4707, + "theoretical_loss": 3.580696722471096, + "tokens_seen": 1337393152 + }, + { + "epoch": 0.41, + "learning_rate": 0.00048064192577733204, + "loss": 1.4503, + "theoretical_loss": 3.58068103556116, + "tokens_seen": 1337458688 + }, + { + "epoch": 0.41, + "learning_rate": 0.0004806258776328987, + "loss": 1.3133, + "theoretical_loss": 3.5806653496350838, + "tokens_seen": 1337524224 + }, + { + "epoch": 0.41, + "learning_rate": 0.00048060982948846544, + "loss": 1.461, + "theoretical_loss": 3.5806496646927575, + "tokens_seen": 1337589760 + }, + { + "epoch": 0.41, + "learning_rate": 0.0004805937813440321, + "loss": 1.4108, + "theoretical_loss": 3.5806339807340715, + "tokens_seen": 1337655296 + }, + { + "epoch": 0.41, + "learning_rate": 0.00048057773319959884, + "loss": 1.3935, + "theoretical_loss": 3.5806182977589156, + "tokens_seen": 1337720832 + }, + { + "epoch": 0.41, + "learning_rate": 0.0004805616850551655, + "loss": 1.4475, + "theoretical_loss": 3.580602615767181, + "tokens_seen": 1337786368 + }, + { + "epoch": 0.41, + "learning_rate": 0.00048054563691073224, + "loss": 1.3663, + "theoretical_loss": 3.5805869347587564, + "tokens_seen": 1337851904 + }, + { + "epoch": 0.41, + "learning_rate": 0.0004805295887662989, + "loss": 1.3921, + "theoretical_loss": 3.580571254733533, + "tokens_seen": 1337917440 + }, + { + "epoch": 0.41, + "learning_rate": 0.00048051354062186565, + "loss": 1.4047, + "theoretical_loss": 3.5805555756914, + "tokens_seen": 1337982976 + }, + { + "epoch": 0.41, + "learning_rate": 0.0004804974924774323, + "loss": 1.3098, + "theoretical_loss": 3.5805398976322484, + "tokens_seen": 1338048512 + }, + { + "epoch": 0.41, + "learning_rate": 0.000480481444332999, + "loss": 1.4243, + "theoretical_loss": 3.5805242205559686, + "tokens_seen": 1338114048 + }, + { + "epoch": 0.41, + "learning_rate": 0.00048046539618856575, + "loss": 1.3265, + "theoretical_loss": 3.5805085444624503, + "tokens_seen": 1338179584 + }, + { + "epoch": 0.41, + "learning_rate": 0.0004804493480441324, + "loss": 1.3961, + "theoretical_loss": 3.580492869351584, + "tokens_seen": 1338245120 + }, + { + "epoch": 0.41, + "learning_rate": 0.00048043329989969915, + "loss": 1.3032, + "theoretical_loss": 3.58047719522326, + "tokens_seen": 1338310656 + }, + { + "epoch": 0.41, + "learning_rate": 0.0004804172517552658, + "loss": 1.3744, + "theoretical_loss": 3.5804615220773686, + "tokens_seen": 1338376192 + }, + { + "epoch": 0.41, + "learning_rate": 0.00048040120361083255, + "loss": 1.3528, + "theoretical_loss": 3.5804458499138, + "tokens_seen": 1338441728 + }, + { + "epoch": 0.41, + "learning_rate": 0.0004803851554663992, + "loss": 1.3973, + "theoretical_loss": 3.5804301787324446, + "tokens_seen": 1338507264 + }, + { + "epoch": 0.41, + "learning_rate": 0.00048036910732196595, + "loss": 1.3926, + "theoretical_loss": 3.5804145085331927, + "tokens_seen": 1338572800 + }, + { + "epoch": 0.41, + "learning_rate": 0.0004803530591775326, + "loss": 1.2702, + "theoretical_loss": 3.580398839315935, + "tokens_seen": 1338638336 + }, + { + "epoch": 0.41, + "learning_rate": 0.00048033701103309935, + "loss": 1.2978, + "theoretical_loss": 3.5803831710805616, + "tokens_seen": 1338703872 + }, + { + "epoch": 0.41, + "learning_rate": 0.000480320962888666, + "loss": 1.372, + "theoretical_loss": 3.580367503826963, + "tokens_seen": 1338769408 + }, + { + "epoch": 0.41, + "learning_rate": 0.00048030491474423275, + "loss": 1.4327, + "theoretical_loss": 3.5803518375550296, + "tokens_seen": 1338834944 + }, + { + "epoch": 0.41, + "learning_rate": 0.0004802888665997994, + "loss": 1.3833, + "theoretical_loss": 3.580336172264652, + "tokens_seen": 1338900480 + }, + { + "epoch": 0.41, + "learning_rate": 0.00048027281845536616, + "loss": 1.4244, + "theoretical_loss": 3.5803205079557197, + "tokens_seen": 1338966016 + }, + { + "epoch": 0.41, + "learning_rate": 0.0004802567703109328, + "loss": 1.3771, + "theoretical_loss": 3.5803048446281247, + "tokens_seen": 1339031552 + }, + { + "epoch": 0.41, + "learning_rate": 0.00048024072216649956, + "loss": 1.346, + "theoretical_loss": 3.5802891822817564, + "tokens_seen": 1339097088 + }, + { + "epoch": 0.41, + "learning_rate": 0.0004802246740220662, + "loss": 1.3711, + "theoretical_loss": 3.580273520916506, + "tokens_seen": 1339162624 + }, + { + "epoch": 0.41, + "learning_rate": 0.00048020862587763296, + "loss": 1.432, + "theoretical_loss": 3.5802578605322637, + "tokens_seen": 1339228160 + }, + { + "epoch": 0.41, + "learning_rate": 0.0004801925777331996, + "loss": 1.3425, + "theoretical_loss": 3.58024220112892, + "tokens_seen": 1339293696 + }, + { + "epoch": 0.41, + "learning_rate": 0.00048017652958876636, + "loss": 1.4106, + "theoretical_loss": 3.580226542706365, + "tokens_seen": 1339359232 + }, + { + "epoch": 0.41, + "learning_rate": 0.000480160481444333, + "loss": 1.4072, + "theoretical_loss": 3.5802108852644903, + "tokens_seen": 1339424768 + }, + { + "epoch": 0.41, + "learning_rate": 0.00048014443329989976, + "loss": 1.413, + "theoretical_loss": 3.580195228803186, + "tokens_seen": 1339490304 + }, + { + "epoch": 0.41, + "learning_rate": 0.0004801283851554664, + "loss": 1.3501, + "theoretical_loss": 3.5801795733223427, + "tokens_seen": 1339555840 + }, + { + "epoch": 0.41, + "learning_rate": 0.00048011233701103316, + "loss": 1.4462, + "theoretical_loss": 3.5801639188218513, + "tokens_seen": 1339621376 + }, + { + "epoch": 0.41, + "learning_rate": 0.0004800962888665998, + "loss": 1.3508, + "theoretical_loss": 3.5801482653016015, + "tokens_seen": 1339686912 + }, + { + "epoch": 0.41, + "learning_rate": 0.00048008024072216656, + "loss": 1.4352, + "theoretical_loss": 3.5801326127614854, + "tokens_seen": 1339752448 + }, + { + "epoch": 0.41, + "learning_rate": 0.0004800641925777332, + "loss": 1.4134, + "theoretical_loss": 3.5801169612013926, + "tokens_seen": 1339817984 + }, + { + "epoch": 0.41, + "learning_rate": 0.00048004814443329996, + "loss": 1.2417, + "theoretical_loss": 3.580101310621214, + "tokens_seen": 1339883520 + }, + { + "epoch": 0.41, + "learning_rate": 0.0004800320962888666, + "loss": 1.3423, + "theoretical_loss": 3.580085661020841, + "tokens_seen": 1339949056 + }, + { + "epoch": 0.41, + "learning_rate": 0.00048001604814443337, + "loss": 1.4697, + "theoretical_loss": 3.5800700124001636, + "tokens_seen": 1340014592 + }, + { + "epoch": 0.41, + "learning_rate": 0.00048, + "loss": 1.4053, + "theoretical_loss": 3.5800543647590732, + "tokens_seen": 1340080128 + }, + { + "epoch": 0.41, + "learning_rate": 0.00047998395185556677, + "loss": 1.3128, + "theoretical_loss": 3.5800387180974598, + "tokens_seen": 1340145664 + }, + { + "epoch": 0.41, + "objective/train/docs_used": 490192, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.3759901523590088, + "objective/train/theoretical_loss": 3.580023072415215, + "objective/train/tokens_used": 1360671200, + "theoretical_loss": 3.580023072415215, + "tokens_seen": 1340211200 + }, + { + "epoch": 0.41, + "learning_rate": 0.0004799679037111334, + "loss": 1.4005, + "theoretical_loss": 3.580023072415215, + "tokens_seen": 1340211200 + }, + { + "epoch": 0.41, + "learning_rate": 0.00047995185556670017, + "loss": 1.4498, + "theoretical_loss": 3.5800074277122285, + "tokens_seen": 1340276736 + }, + { + "epoch": 0.41, + "learning_rate": 0.0004799358074222668, + "loss": 1.2713, + "theoretical_loss": 3.579991783988392, + "tokens_seen": 1340342272 + }, + { + "epoch": 0.41, + "learning_rate": 0.00047991975927783357, + "loss": 1.3702, + "theoretical_loss": 3.5799761412435966, + "tokens_seen": 1340407808 + }, + { + "epoch": 0.41, + "learning_rate": 0.0004799037111334002, + "loss": 1.39, + "theoretical_loss": 3.5799604994777328, + "tokens_seen": 1340473344 + }, + { + "epoch": 0.41, + "learning_rate": 0.00047988766298896697, + "loss": 1.4604, + "theoretical_loss": 3.5799448586906912, + "tokens_seen": 1340538880 + }, + { + "epoch": 0.41, + "learning_rate": 0.0004798716148445336, + "loss": 1.3948, + "theoretical_loss": 3.579929218882363, + "tokens_seen": 1340604416 + }, + { + "epoch": 0.41, + "learning_rate": 0.00047985556670010037, + "loss": 1.4392, + "theoretical_loss": 3.5799135800526383, + "tokens_seen": 1340669952 + }, + { + "epoch": 0.41, + "learning_rate": 0.000479839518555667, + "loss": 1.5037, + "theoretical_loss": 3.57989794220141, + "tokens_seen": 1340735488 + }, + { + "epoch": 0.41, + "learning_rate": 0.0004798234704112337, + "loss": 1.4474, + "theoretical_loss": 3.579882305328567, + "tokens_seen": 1340801024 + }, + { + "epoch": 0.41, + "learning_rate": 0.0004798074222668004, + "loss": 1.3908, + "theoretical_loss": 3.5798666694340016, + "tokens_seen": 1340866560 + }, + { + "epoch": 0.41, + "learning_rate": 0.0004797913741223671, + "loss": 1.1533, + "theoretical_loss": 3.5798510345176036, + "tokens_seen": 1340932096 + }, + { + "epoch": 0.41, + "learning_rate": 0.0004797753259779338, + "loss": 1.4605, + "theoretical_loss": 3.579835400579265, + "tokens_seen": 1340997632 + }, + { + "epoch": 0.41, + "learning_rate": 0.0004797592778335005, + "loss": 1.511, + "theoretical_loss": 3.5798197676188765, + "tokens_seen": 1341063168 + }, + { + "epoch": 0.41, + "learning_rate": 0.0004797432296890672, + "loss": 1.3631, + "theoretical_loss": 3.57980413563633, + "tokens_seen": 1341128704 + }, + { + "epoch": 0.41, + "learning_rate": 0.0004797271815446339, + "loss": 1.3895, + "theoretical_loss": 3.5797885046315145, + "tokens_seen": 1341194240 + }, + { + "epoch": 0.41, + "learning_rate": 0.00047971113340020057, + "loss": 1.3829, + "theoretical_loss": 3.579772874604323, + "tokens_seen": 1341259776 + }, + { + "epoch": 0.41, + "learning_rate": 0.0004796950852557673, + "loss": 1.4145, + "theoretical_loss": 3.5797572455546454, + "tokens_seen": 1341325312 + }, + { + "epoch": 0.41, + "learning_rate": 0.00047967903711133397, + "loss": 1.4774, + "theoretical_loss": 3.579741617482374, + "tokens_seen": 1341390848 + }, + { + "epoch": 0.41, + "learning_rate": 0.0004796629889669007, + "loss": 1.3882, + "theoretical_loss": 3.5797259903873986, + "tokens_seen": 1341456384 + }, + { + "epoch": 0.41, + "learning_rate": 0.00047964694082246737, + "loss": 1.3047, + "theoretical_loss": 3.579710364269611, + "tokens_seen": 1341521920 + }, + { + "epoch": 0.41, + "learning_rate": 0.0004796308926780341, + "loss": 1.4722, + "theoretical_loss": 3.579694739128902, + "tokens_seen": 1341587456 + }, + { + "epoch": 0.41, + "learning_rate": 0.0004796148445336008, + "loss": 1.4764, + "theoretical_loss": 3.5796791149651637, + "tokens_seen": 1341652992 + }, + { + "epoch": 0.41, + "learning_rate": 0.00047959879638916753, + "loss": 1.3732, + "theoretical_loss": 3.5796634917782866, + "tokens_seen": 1341718528 + }, + { + "epoch": 0.41, + "learning_rate": 0.0004795827482447342, + "loss": 1.3599, + "theoretical_loss": 3.579647869568162, + "tokens_seen": 1341784064 + }, + { + "epoch": 0.41, + "learning_rate": 0.00047956670010030093, + "loss": 1.4027, + "theoretical_loss": 3.579632248334681, + "tokens_seen": 1341849600 + }, + { + "epoch": 0.41, + "learning_rate": 0.0004795506519558676, + "loss": 1.4467, + "theoretical_loss": 3.579616628077735, + "tokens_seen": 1341915136 + }, + { + "epoch": 0.41, + "learning_rate": 0.00047953460381143433, + "loss": 1.4875, + "theoretical_loss": 3.579601008797215, + "tokens_seen": 1341980672 + }, + { + "epoch": 0.41, + "learning_rate": 0.000479518555667001, + "loss": 1.3945, + "theoretical_loss": 3.579585390493013, + "tokens_seen": 1342046208 + }, + { + "epoch": 0.41, + "learning_rate": 0.00047950250752256773, + "loss": 1.4972, + "theoretical_loss": 3.5795697731650193, + "tokens_seen": 1342111744 + }, + { + "epoch": 0.41, + "learning_rate": 0.0004794864593781344, + "loss": 1.5009, + "theoretical_loss": 3.5795541568131264, + "tokens_seen": 1342177280 + }, + { + "epoch": 0.41, + "learning_rate": 0.00047947041123370113, + "loss": 1.3597, + "theoretical_loss": 3.5795385414372247, + "tokens_seen": 1342242816 + }, + { + "epoch": 0.41, + "learning_rate": 0.0004794543630892678, + "loss": 1.3759, + "theoretical_loss": 3.579522927037206, + "tokens_seen": 1342308352 + }, + { + "epoch": 0.41, + "learning_rate": 0.00047943831494483453, + "loss": 1.4219, + "theoretical_loss": 3.5795073136129614, + "tokens_seen": 1342373888 + }, + { + "epoch": 0.41, + "learning_rate": 0.0004794222668004012, + "loss": 1.4791, + "theoretical_loss": 3.5794917011643825, + "tokens_seen": 1342439424 + }, + { + "epoch": 0.41, + "learning_rate": 0.00047940621865596794, + "loss": 1.4111, + "theoretical_loss": 3.57947608969136, + "tokens_seen": 1342504960 + }, + { + "epoch": 0.41, + "learning_rate": 0.0004793901705115346, + "loss": 1.5393, + "theoretical_loss": 3.579460479193787, + "tokens_seen": 1342570496 + }, + { + "epoch": 0.41, + "learning_rate": 0.00047937412236710134, + "loss": 1.351, + "theoretical_loss": 3.579444869671553, + "tokens_seen": 1342636032 + }, + { + "epoch": 0.41, + "learning_rate": 0.000479358074222668, + "loss": 1.4251, + "theoretical_loss": 3.5794292611245506, + "tokens_seen": 1342701568 + }, + { + "epoch": 0.41, + "learning_rate": 0.00047934202607823474, + "loss": 1.3787, + "theoretical_loss": 3.5794136535526713, + "tokens_seen": 1342767104 + }, + { + "epoch": 0.41, + "learning_rate": 0.0004793259779338014, + "loss": 1.368, + "theoretical_loss": 3.5793980469558058, + "tokens_seen": 1342832640 + }, + { + "epoch": 0.41, + "learning_rate": 0.00047930992978936814, + "loss": 1.4054, + "theoretical_loss": 3.5793824413338466, + "tokens_seen": 1342898176 + }, + { + "epoch": 0.41, + "learning_rate": 0.0004792938816449348, + "loss": 1.2546, + "theoretical_loss": 3.579366836686684, + "tokens_seen": 1342963712 + }, + { + "epoch": 0.41, + "learning_rate": 0.00047927783350050154, + "loss": 1.4909, + "theoretical_loss": 3.579351233014211, + "tokens_seen": 1343029248 + }, + { + "epoch": 0.41, + "learning_rate": 0.0004792617853560683, + "loss": 1.4281, + "theoretical_loss": 3.579335630316318, + "tokens_seen": 1343094784 + }, + { + "epoch": 0.41, + "learning_rate": 0.00047924573721163494, + "loss": 1.3666, + "theoretical_loss": 3.5793200285928974, + "tokens_seen": 1343160320 + }, + { + "epoch": 0.41, + "learning_rate": 0.0004792296890672017, + "loss": 1.3516, + "theoretical_loss": 3.57930442784384, + "tokens_seen": 1343225856 + }, + { + "epoch": 0.41, + "learning_rate": 0.00047921364092276834, + "loss": 1.4209, + "theoretical_loss": 3.5792888280690383, + "tokens_seen": 1343291392 + }, + { + "epoch": 0.41, + "learning_rate": 0.00047919759277833504, + "loss": 1.2981, + "theoretical_loss": 3.5792732292683835, + "tokens_seen": 1343356928 + }, + { + "epoch": 0.41, + "learning_rate": 0.00047918154463390174, + "loss": 1.3439, + "theoretical_loss": 3.5792576314417666, + "tokens_seen": 1343422464 + }, + { + "epoch": 0.41, + "objective/train/docs_used": 490931, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.5279899835586548, + "objective/train/theoretical_loss": 3.5792420345890807, + "objective/train/tokens_used": 1363948000, + "theoretical_loss": 3.5792420345890807, + "tokens_seen": 1343488000 + }, + { + "epoch": 0.41, + "learning_rate": 0.00047916549648946845, + "loss": 1.4204, + "theoretical_loss": 3.5792420345890807, + "tokens_seen": 1343488000 + }, + { + "epoch": 0.41, + "learning_rate": 0.00047914944834503515, + "loss": 1.3364, + "theoretical_loss": 3.579226438710216, + "tokens_seen": 1343553536 + }, + { + "epoch": 0.41, + "learning_rate": 0.00047913340020060185, + "loss": 1.3705, + "theoretical_loss": 3.579210843805065, + "tokens_seen": 1343619072 + }, + { + "epoch": 0.41, + "learning_rate": 0.00047911735205616855, + "loss": 1.3217, + "theoretical_loss": 3.5791952498735196, + "tokens_seen": 1343684608 + }, + { + "epoch": 0.41, + "learning_rate": 0.00047910130391173525, + "loss": 1.3814, + "theoretical_loss": 3.579179656915471, + "tokens_seen": 1343750144 + }, + { + "epoch": 0.41, + "learning_rate": 0.00047908525576730195, + "loss": 1.3725, + "theoretical_loss": 3.579164064930811, + "tokens_seen": 1343815680 + }, + { + "epoch": 0.41, + "learning_rate": 0.00047906920762286865, + "loss": 1.3175, + "theoretical_loss": 3.5791484739194317, + "tokens_seen": 1343881216 + }, + { + "epoch": 0.41, + "learning_rate": 0.0004790531594784353, + "loss": 1.394, + "theoretical_loss": 3.5791328838812246, + "tokens_seen": 1343946752 + }, + { + "epoch": 0.41, + "learning_rate": 0.00047903711133400205, + "loss": 1.3153, + "theoretical_loss": 3.579117294816082, + "tokens_seen": 1344012288 + }, + { + "epoch": 0.41, + "learning_rate": 0.0004790210631895687, + "loss": 1.4032, + "theoretical_loss": 3.579101706723895, + "tokens_seen": 1344077824 + }, + { + "epoch": 0.41, + "learning_rate": 0.00047900501504513545, + "loss": 1.3696, + "theoretical_loss": 3.579086119604556, + "tokens_seen": 1344143360 + }, + { + "epoch": 0.41, + "learning_rate": 0.0004789889669007021, + "loss": 1.3514, + "theoretical_loss": 3.5790705334579567, + "tokens_seen": 1344208896 + }, + { + "epoch": 0.41, + "learning_rate": 0.00047897291875626885, + "loss": 1.3505, + "theoretical_loss": 3.5790549482839884, + "tokens_seen": 1344274432 + }, + { + "epoch": 0.41, + "learning_rate": 0.0004789568706118355, + "loss": 1.3984, + "theoretical_loss": 3.579039364082544, + "tokens_seen": 1344339968 + }, + { + "epoch": 0.41, + "learning_rate": 0.00047894082246740225, + "loss": 1.3737, + "theoretical_loss": 3.5790237808535146, + "tokens_seen": 1344405504 + }, + { + "epoch": 0.41, + "learning_rate": 0.0004789247743229689, + "loss": 1.3557, + "theoretical_loss": 3.5790081985967928, + "tokens_seen": 1344471040 + }, + { + "epoch": 0.41, + "learning_rate": 0.00047890872617853566, + "loss": 1.3373, + "theoretical_loss": 3.57899261731227, + "tokens_seen": 1344536576 + }, + { + "epoch": 0.41, + "learning_rate": 0.0004788926780341023, + "loss": 1.366, + "theoretical_loss": 3.5789770369998384, + "tokens_seen": 1344602112 + }, + { + "epoch": 0.41, + "learning_rate": 0.00047887662988966906, + "loss": 1.3584, + "theoretical_loss": 3.57896145765939, + "tokens_seen": 1344667648 + }, + { + "epoch": 0.41, + "learning_rate": 0.0004788605817452357, + "loss": 1.4072, + "theoretical_loss": 3.5789458792908166, + "tokens_seen": 1344733184 + }, + { + "epoch": 0.41, + "learning_rate": 0.00047884453360080246, + "loss": 1.3111, + "theoretical_loss": 3.5789303018940104, + "tokens_seen": 1344798720 + }, + { + "epoch": 0.41, + "learning_rate": 0.0004788284854563691, + "loss": 1.4329, + "theoretical_loss": 3.5789147254688634, + "tokens_seen": 1344864256 + }, + { + "epoch": 0.41, + "learning_rate": 0.00047881243731193586, + "loss": 1.3703, + "theoretical_loss": 3.5788991500152676, + "tokens_seen": 1344929792 + }, + { + "epoch": 0.41, + "learning_rate": 0.0004787963891675025, + "loss": 1.3917, + "theoretical_loss": 3.5788835755331148, + "tokens_seen": 1344995328 + }, + { + "epoch": 0.41, + "learning_rate": 0.00047878034102306926, + "loss": 1.2662, + "theoretical_loss": 3.578868002022298, + "tokens_seen": 1345060864 + }, + { + "epoch": 0.41, + "learning_rate": 0.0004787642928786359, + "loss": 1.3198, + "theoretical_loss": 3.578852429482708, + "tokens_seen": 1345126400 + }, + { + "epoch": 0.41, + "learning_rate": 0.00047874824473420266, + "loss": 1.5104, + "theoretical_loss": 3.578836857914238, + "tokens_seen": 1345191936 + }, + { + "epoch": 0.41, + "learning_rate": 0.0004787321965897693, + "loss": 1.4352, + "theoretical_loss": 3.5788212873167793, + "tokens_seen": 1345257472 + }, + { + "epoch": 0.41, + "learning_rate": 0.00047871614844533606, + "loss": 1.4713, + "theoretical_loss": 3.5788057176902246, + "tokens_seen": 1345323008 + }, + { + "epoch": 0.41, + "learning_rate": 0.0004787001003009027, + "loss": 1.4078, + "theoretical_loss": 3.5787901490344662, + "tokens_seen": 1345388544 + }, + { + "epoch": 0.41, + "learning_rate": 0.00047868405215646946, + "loss": 1.2583, + "theoretical_loss": 3.5787745813493954, + "tokens_seen": 1345454080 + }, + { + "epoch": 0.41, + "learning_rate": 0.0004786680040120361, + "loss": 1.3695, + "theoretical_loss": 3.5787590146349055, + "tokens_seen": 1345519616 + }, + { + "epoch": 0.41, + "learning_rate": 0.00047865195586760287, + "loss": 1.2725, + "theoretical_loss": 3.5787434488908882, + "tokens_seen": 1345585152 + }, + { + "epoch": 0.41, + "learning_rate": 0.0004786359077231695, + "loss": 1.4049, + "theoretical_loss": 3.578727884117236, + "tokens_seen": 1345650688 + }, + { + "epoch": 0.41, + "learning_rate": 0.00047861985957873627, + "loss": 1.4071, + "theoretical_loss": 3.5787123203138402, + "tokens_seen": 1345716224 + }, + { + "epoch": 0.41, + "learning_rate": 0.0004786038114343029, + "loss": 1.4022, + "theoretical_loss": 3.5786967574805946, + "tokens_seen": 1345781760 + }, + { + "epoch": 0.41, + "learning_rate": 0.00047858776328986967, + "loss": 1.3091, + "theoretical_loss": 3.57868119561739, + "tokens_seen": 1345847296 + }, + { + "epoch": 0.41, + "learning_rate": 0.0004785717151454363, + "loss": 1.3204, + "theoretical_loss": 3.5786656347241195, + "tokens_seen": 1345912832 + }, + { + "epoch": 0.41, + "learning_rate": 0.00047855566700100307, + "loss": 1.4548, + "theoretical_loss": 3.5786500748006755, + "tokens_seen": 1345978368 + }, + { + "epoch": 0.41, + "learning_rate": 0.0004785396188565697, + "loss": 1.3912, + "theoretical_loss": 3.57863451584695, + "tokens_seen": 1346043904 + }, + { + "epoch": 0.41, + "learning_rate": 0.00047852357071213647, + "loss": 1.4623, + "theoretical_loss": 3.578618957862836, + "tokens_seen": 1346109440 + }, + { + "epoch": 0.41, + "learning_rate": 0.0004785075225677031, + "loss": 1.2746, + "theoretical_loss": 3.5786034008482246, + "tokens_seen": 1346174976 + }, + { + "epoch": 0.41, + "learning_rate": 0.00047849147442326987, + "loss": 1.4304, + "theoretical_loss": 3.5785878448030095, + "tokens_seen": 1346240512 + }, + { + "epoch": 0.41, + "learning_rate": 0.0004784754262788365, + "loss": 1.3542, + "theoretical_loss": 3.578572289727082, + "tokens_seen": 1346306048 + }, + { + "epoch": 0.41, + "learning_rate": 0.0004784593781344033, + "loss": 1.3801, + "theoretical_loss": 3.5785567356203356, + "tokens_seen": 1346371584 + }, + { + "epoch": 0.41, + "learning_rate": 0.0004784433299899699, + "loss": 1.4226, + "theoretical_loss": 3.5785411824826623, + "tokens_seen": 1346437120 + }, + { + "epoch": 0.41, + "learning_rate": 0.0004784272818455366, + "loss": 1.3627, + "theoretical_loss": 3.578525630313954, + "tokens_seen": 1346502656 + }, + { + "epoch": 0.41, + "learning_rate": 0.0004784112337011033, + "loss": 1.2945, + "theoretical_loss": 3.5785100791141042, + "tokens_seen": 1346568192 + }, + { + "epoch": 0.41, + "learning_rate": 0.00047839518555667, + "loss": 1.3516, + "theoretical_loss": 3.578494528883004, + "tokens_seen": 1346633728 + }, + { + "epoch": 0.41, + "learning_rate": 0.0004783791374122367, + "loss": 1.3403, + "theoretical_loss": 3.578478979620548, + "tokens_seen": 1346699264 + }, + { + "epoch": 0.41, + "objective/train/docs_used": 492127, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.338100552558899, + "objective/train/theoretical_loss": 3.5784634313266266, + "objective/train/tokens_used": 1367224800, + "theoretical_loss": 3.5784634313266266, + "tokens_seen": 1346764800 + }, + { + "epoch": 0.41, + "learning_rate": 0.0004783630892678034, + "loss": 1.3106, + "theoretical_loss": 3.5784634313266266, + "tokens_seen": 1346764800 + }, + { + "epoch": 0.41, + "learning_rate": 0.0004783470411233701, + "loss": 1.3894, + "theoretical_loss": 3.578447884001134, + "tokens_seen": 1346830336 + }, + { + "epoch": 0.41, + "learning_rate": 0.0004783309929789368, + "loss": 1.2141, + "theoretical_loss": 3.578432337643961, + "tokens_seen": 1346895872 + }, + { + "epoch": 0.41, + "learning_rate": 0.0004783149448345035, + "loss": 1.4266, + "theoretical_loss": 3.578416792255002, + "tokens_seen": 1346961408 + }, + { + "epoch": 0.41, + "learning_rate": 0.0004782988966900702, + "loss": 1.3865, + "theoretical_loss": 3.5784012478341483, + "tokens_seen": 1347026944 + }, + { + "epoch": 0.41, + "learning_rate": 0.00047828284854563687, + "loss": 1.3164, + "theoretical_loss": 3.5783857043812937, + "tokens_seen": 1347092480 + }, + { + "epoch": 0.41, + "learning_rate": 0.0004782668004012036, + "loss": 1.403, + "theoretical_loss": 3.5783701618963297, + "tokens_seen": 1347158016 + }, + { + "epoch": 0.41, + "learning_rate": 0.0004782507522567703, + "loss": 1.2599, + "theoretical_loss": 3.578354620379149, + "tokens_seen": 1347223552 + }, + { + "epoch": 0.41, + "learning_rate": 0.00047823470411233703, + "loss": 1.3997, + "theoretical_loss": 3.578339079829645, + "tokens_seen": 1347289088 + }, + { + "epoch": 0.41, + "learning_rate": 0.0004782186559679037, + "loss": 1.4831, + "theoretical_loss": 3.5783235402477103, + "tokens_seen": 1347354624 + }, + { + "epoch": 0.41, + "learning_rate": 0.00047820260782347043, + "loss": 1.2432, + "theoretical_loss": 3.578308001633237, + "tokens_seen": 1347420160 + }, + { + "epoch": 0.41, + "learning_rate": 0.0004781865596790371, + "loss": 1.3732, + "theoretical_loss": 3.578292463986118, + "tokens_seen": 1347485696 + }, + { + "epoch": 0.41, + "learning_rate": 0.00047817051153460383, + "loss": 1.3582, + "theoretical_loss": 3.5782769273062467, + "tokens_seen": 1347551232 + }, + { + "epoch": 0.41, + "learning_rate": 0.0004781544633901705, + "loss": 1.2951, + "theoretical_loss": 3.578261391593515, + "tokens_seen": 1347616768 + }, + { + "epoch": 0.41, + "learning_rate": 0.00047813841524573723, + "loss": 1.3402, + "theoretical_loss": 3.5782458568478157, + "tokens_seen": 1347682304 + }, + { + "epoch": 0.41, + "learning_rate": 0.0004781223671013039, + "loss": 1.387, + "theoretical_loss": 3.578230323069042, + "tokens_seen": 1347747840 + }, + { + "epoch": 0.41, + "learning_rate": 0.00047810631895687063, + "loss": 1.3891, + "theoretical_loss": 3.5782147902570873, + "tokens_seen": 1347813376 + }, + { + "epoch": 0.41, + "learning_rate": 0.0004780902708124373, + "loss": 1.308, + "theoretical_loss": 3.578199258411843, + "tokens_seen": 1347878912 + }, + { + "epoch": 0.41, + "learning_rate": 0.00047807422266800403, + "loss": 1.4037, + "theoretical_loss": 3.5781837275332027, + "tokens_seen": 1347944448 + }, + { + "epoch": 0.41, + "learning_rate": 0.0004780581745235708, + "loss": 1.3413, + "theoretical_loss": 3.5781681976210598, + "tokens_seen": 1348009984 + }, + { + "epoch": 0.41, + "learning_rate": 0.00047804212637913744, + "loss": 1.4662, + "theoretical_loss": 3.578152668675306, + "tokens_seen": 1348075520 + }, + { + "epoch": 0.41, + "learning_rate": 0.0004780260782347042, + "loss": 1.4566, + "theoretical_loss": 3.578137140695835, + "tokens_seen": 1348141056 + }, + { + "epoch": 0.41, + "learning_rate": 0.00047801003009027084, + "loss": 1.344, + "theoretical_loss": 3.578121613682539, + "tokens_seen": 1348206592 + }, + { + "epoch": 0.41, + "learning_rate": 0.0004779939819458376, + "loss": 1.3577, + "theoretical_loss": 3.578106087635312, + "tokens_seen": 1348272128 + }, + { + "epoch": 0.41, + "learning_rate": 0.00047797793380140424, + "loss": 1.3538, + "theoretical_loss": 3.578090562554046, + "tokens_seen": 1348337664 + }, + { + "epoch": 0.41, + "learning_rate": 0.000477961885656971, + "loss": 1.3505, + "theoretical_loss": 3.578075038438634, + "tokens_seen": 1348403200 + }, + { + "epoch": 0.41, + "learning_rate": 0.00047794583751253764, + "loss": 1.425, + "theoretical_loss": 3.5780595152889694, + "tokens_seen": 1348468736 + }, + { + "epoch": 0.41, + "learning_rate": 0.0004779297893681044, + "loss": 1.3511, + "theoretical_loss": 3.5780439931049455, + "tokens_seen": 1348534272 + }, + { + "epoch": 0.41, + "learning_rate": 0.00047791374122367104, + "loss": 1.4593, + "theoretical_loss": 3.5780284718864546, + "tokens_seen": 1348599808 + }, + { + "epoch": 0.41, + "learning_rate": 0.0004778976930792378, + "loss": 1.484, + "theoretical_loss": 3.57801295163339, + "tokens_seen": 1348665344 + }, + { + "epoch": 0.41, + "learning_rate": 0.00047788164493480444, + "loss": 1.3695, + "theoretical_loss": 3.5779974323456445, + "tokens_seen": 1348730880 + }, + { + "epoch": 0.41, + "learning_rate": 0.0004778655967903712, + "loss": 1.4566, + "theoretical_loss": 3.5779819140231117, + "tokens_seen": 1348796416 + }, + { + "epoch": 0.41, + "learning_rate": 0.00047784954864593784, + "loss": 1.3104, + "theoretical_loss": 3.577966396665684, + "tokens_seen": 1348861952 + }, + { + "epoch": 0.41, + "learning_rate": 0.0004778335005015046, + "loss": 1.3636, + "theoretical_loss": 3.5779508802732547, + "tokens_seen": 1348927488 + }, + { + "epoch": 0.41, + "learning_rate": 0.00047781745235707124, + "loss": 1.3927, + "theoretical_loss": 3.5779353648457173, + "tokens_seen": 1348993024 + }, + { + "epoch": 0.41, + "learning_rate": 0.00047780140421263794, + "loss": 1.4316, + "theoretical_loss": 3.577919850382965, + "tokens_seen": 1349058560 + }, + { + "epoch": 0.41, + "learning_rate": 0.00047778535606820465, + "loss": 1.3975, + "theoretical_loss": 3.57790433688489, + "tokens_seen": 1349124096 + }, + { + "epoch": 0.41, + "learning_rate": 0.00047776930792377135, + "loss": 1.3806, + "theoretical_loss": 3.577888824351387, + "tokens_seen": 1349189632 + }, + { + "epoch": 0.41, + "learning_rate": 0.00047775325977933805, + "loss": 1.4002, + "theoretical_loss": 3.577873312782347, + "tokens_seen": 1349255168 + }, + { + "epoch": 0.41, + "learning_rate": 0.00047773721163490475, + "loss": 1.3203, + "theoretical_loss": 3.5778578021776655, + "tokens_seen": 1349320704 + }, + { + "epoch": 0.41, + "learning_rate": 0.00047772116349047145, + "loss": 1.2932, + "theoretical_loss": 3.577842292537234, + "tokens_seen": 1349386240 + }, + { + "epoch": 0.41, + "learning_rate": 0.00047770511534603815, + "loss": 1.3298, + "theoretical_loss": 3.5778267838609468, + "tokens_seen": 1349451776 + }, + { + "epoch": 0.41, + "learning_rate": 0.00047768906720160485, + "loss": 1.3713, + "theoretical_loss": 3.5778112761486964, + "tokens_seen": 1349517312 + }, + { + "epoch": 0.41, + "learning_rate": 0.00047767301905717155, + "loss": 1.4213, + "theoretical_loss": 3.577795769400377, + "tokens_seen": 1349582848 + }, + { + "epoch": 0.41, + "learning_rate": 0.0004776569709127382, + "loss": 1.3785, + "theoretical_loss": 3.577780263615881, + "tokens_seen": 1349648384 + }, + { + "epoch": 0.41, + "learning_rate": 0.00047764092276830495, + "loss": 1.3673, + "theoretical_loss": 3.5777647587951016, + "tokens_seen": 1349713920 + }, + { + "epoch": 0.41, + "learning_rate": 0.0004776248746238716, + "loss": 1.4053, + "theoretical_loss": 3.577749254937933, + "tokens_seen": 1349779456 + }, + { + "epoch": 0.41, + "learning_rate": 0.00047760882647943835, + "loss": 1.4725, + "theoretical_loss": 3.577733752044267, + "tokens_seen": 1349844992 + }, + { + "epoch": 0.41, + "learning_rate": 0.000477592778335005, + "loss": 1.4216, + "theoretical_loss": 3.577718250113999, + "tokens_seen": 1349910528 + }, + { + "epoch": 0.41, + "learning_rate": 0.00047757673019057175, + "loss": 1.3963, + "theoretical_loss": 3.5777027491470212, + "tokens_seen": 1349976064 + }, + { + "epoch": 0.41, + "objective/train/docs_used": 493399, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.4353004693984985, + "objective/train/theoretical_loss": 3.577687249143227, + "objective/train/tokens_used": 1370501600, + "theoretical_loss": 3.577687249143227, + "tokens_seen": 1350041600 + }, + { + "epoch": 0.41, + "learning_rate": 0.0004775606820461384, + "loss": 1.3223, + "theoretical_loss": 3.577687249143227, + "tokens_seen": 1350041600 + }, + { + "epoch": 0.41, + "learning_rate": 0.00047754463390170516, + "loss": 1.3448, + "theoretical_loss": 3.57767175010251, + "tokens_seen": 1350107136 + }, + { + "epoch": 0.41, + "learning_rate": 0.0004775285857572718, + "loss": 1.4046, + "theoretical_loss": 3.577656252024763, + "tokens_seen": 1350172672 + }, + { + "epoch": 0.41, + "learning_rate": 0.00047751253761283856, + "loss": 1.461, + "theoretical_loss": 3.5776407549098805, + "tokens_seen": 1350238208 + }, + { + "epoch": 0.41, + "learning_rate": 0.0004774964894684052, + "loss": 1.3768, + "theoretical_loss": 3.577625258757755, + "tokens_seen": 1350303744 + }, + { + "epoch": 0.41, + "learning_rate": 0.00047748044132397196, + "loss": 1.3674, + "theoretical_loss": 3.5776097635682813, + "tokens_seen": 1350369280 + }, + { + "epoch": 0.41, + "learning_rate": 0.0004774643931795386, + "loss": 1.3892, + "theoretical_loss": 3.577594269341351, + "tokens_seen": 1350434816 + }, + { + "epoch": 0.41, + "learning_rate": 0.00047744834503510536, + "loss": 1.377, + "theoretical_loss": 3.5775787760768587, + "tokens_seen": 1350500352 + }, + { + "epoch": 0.41, + "learning_rate": 0.000477432296890672, + "loss": 1.3779, + "theoretical_loss": 3.577563283774698, + "tokens_seen": 1350565888 + }, + { + "epoch": 0.41, + "learning_rate": 0.00047741624874623876, + "loss": 1.4332, + "theoretical_loss": 3.5775477924347623, + "tokens_seen": 1350631424 + }, + { + "epoch": 0.41, + "learning_rate": 0.0004774002006018054, + "loss": 1.3148, + "theoretical_loss": 3.5775323020569445, + "tokens_seen": 1350696960 + }, + { + "epoch": 0.41, + "learning_rate": 0.00047738415245737216, + "loss": 1.2439, + "theoretical_loss": 3.5775168126411394, + "tokens_seen": 1350762496 + }, + { + "epoch": 0.41, + "learning_rate": 0.0004773681043129388, + "loss": 1.4484, + "theoretical_loss": 3.5775013241872395, + "tokens_seen": 1350828032 + }, + { + "epoch": 0.41, + "learning_rate": 0.00047735205616850556, + "loss": 1.403, + "theoretical_loss": 3.577485836695139, + "tokens_seen": 1350893568 + }, + { + "epoch": 0.41, + "learning_rate": 0.0004773360080240722, + "loss": 1.3772, + "theoretical_loss": 3.5774703501647314, + "tokens_seen": 1350959104 + }, + { + "epoch": 0.41, + "learning_rate": 0.00047731995987963896, + "loss": 1.3983, + "theoretical_loss": 3.5774548645959103, + "tokens_seen": 1351024640 + }, + { + "epoch": 0.41, + "learning_rate": 0.0004773039117352056, + "loss": 1.312, + "theoretical_loss": 3.5774393799885695, + "tokens_seen": 1351090176 + }, + { + "epoch": 0.41, + "learning_rate": 0.00047728786359077237, + "loss": 1.3462, + "theoretical_loss": 3.577423896342602, + "tokens_seen": 1351155712 + }, + { + "epoch": 0.41, + "learning_rate": 0.000477271815446339, + "loss": 1.4506, + "theoretical_loss": 3.577408413657902, + "tokens_seen": 1351221248 + }, + { + "epoch": 0.41, + "learning_rate": 0.00047725576730190577, + "loss": 1.4685, + "theoretical_loss": 3.577392931934363, + "tokens_seen": 1351286784 + }, + { + "epoch": 0.41, + "learning_rate": 0.0004772397191574724, + "loss": 1.4643, + "theoretical_loss": 3.5773774511718797, + "tokens_seen": 1351352320 + }, + { + "epoch": 0.41, + "learning_rate": 0.00047722367101303917, + "loss": 1.5081, + "theoretical_loss": 3.577361971370345, + "tokens_seen": 1351417856 + }, + { + "epoch": 0.41, + "learning_rate": 0.0004772076228686058, + "loss": 1.3242, + "theoretical_loss": 3.577346492529652, + "tokens_seen": 1351483392 + }, + { + "epoch": 0.41, + "learning_rate": 0.00047719157472417257, + "loss": 1.3914, + "theoretical_loss": 3.5773310146496953, + "tokens_seen": 1351548928 + }, + { + "epoch": 0.41, + "learning_rate": 0.0004771755265797392, + "loss": 1.3283, + "theoretical_loss": 3.5773155377303687, + "tokens_seen": 1351614464 + }, + { + "epoch": 0.41, + "learning_rate": 0.00047715947843530597, + "loss": 1.4001, + "theoretical_loss": 3.577300061771566, + "tokens_seen": 1351680000 + }, + { + "epoch": 0.41, + "learning_rate": 0.0004771434302908726, + "loss": 1.3768, + "theoretical_loss": 3.577284586773181, + "tokens_seen": 1351745536 + }, + { + "epoch": 0.41, + "learning_rate": 0.00047712738214643937, + "loss": 1.4617, + "theoretical_loss": 3.577269112735107, + "tokens_seen": 1351811072 + }, + { + "epoch": 0.41, + "learning_rate": 0.000477111334002006, + "loss": 1.3833, + "theoretical_loss": 3.577253639657238, + "tokens_seen": 1351876608 + }, + { + "epoch": 0.41, + "learning_rate": 0.00047709528585757277, + "loss": 1.4255, + "theoretical_loss": 3.577238167539469, + "tokens_seen": 1351942144 + }, + { + "epoch": 0.41, + "learning_rate": 0.0004770792377131394, + "loss": 1.5452, + "theoretical_loss": 3.5772226963816927, + "tokens_seen": 1352007680 + }, + { + "epoch": 0.41, + "learning_rate": 0.0004770631895687062, + "loss": 1.459, + "theoretical_loss": 3.577207226183803, + "tokens_seen": 1352073216 + }, + { + "epoch": 0.41, + "learning_rate": 0.0004770471414242728, + "loss": 1.4734, + "theoretical_loss": 3.5771917569456946, + "tokens_seen": 1352138752 + }, + { + "epoch": 0.41, + "learning_rate": 0.0004770310932798395, + "loss": 1.3598, + "theoretical_loss": 3.5771762886672605, + "tokens_seen": 1352204288 + }, + { + "epoch": 0.41, + "learning_rate": 0.0004770150451354062, + "loss": 1.4005, + "theoretical_loss": 3.5771608213483956, + "tokens_seen": 1352269824 + }, + { + "epoch": 0.41, + "learning_rate": 0.0004769989969909729, + "loss": 1.3141, + "theoretical_loss": 3.5771453549889936, + "tokens_seen": 1352335360 + }, + { + "epoch": 0.41, + "learning_rate": 0.0004769829488465396, + "loss": 1.3469, + "theoretical_loss": 3.577129889588948, + "tokens_seen": 1352400896 + }, + { + "epoch": 0.41, + "learning_rate": 0.0004769669007021063, + "loss": 1.3524, + "theoretical_loss": 3.577114425148153, + "tokens_seen": 1352466432 + }, + { + "epoch": 0.41, + "learning_rate": 0.000476950852557673, + "loss": 1.3473, + "theoretical_loss": 3.577098961666503, + "tokens_seen": 1352531968 + }, + { + "epoch": 0.41, + "learning_rate": 0.0004769348044132397, + "loss": 1.2784, + "theoretical_loss": 3.577083499143892, + "tokens_seen": 1352597504 + }, + { + "epoch": 0.41, + "learning_rate": 0.0004769187562688064, + "loss": 1.3096, + "theoretical_loss": 3.5770680375802137, + "tokens_seen": 1352663040 + }, + { + "epoch": 0.41, + "learning_rate": 0.0004769027081243731, + "loss": 1.3707, + "theoretical_loss": 3.577052576975362, + "tokens_seen": 1352728576 + }, + { + "epoch": 0.41, + "learning_rate": 0.00047688665997993977, + "loss": 1.2977, + "theoretical_loss": 3.5770371173292315, + "tokens_seen": 1352794112 + }, + { + "epoch": 0.41, + "learning_rate": 0.00047687061183550653, + "loss": 1.3263, + "theoretical_loss": 3.5770216586417165, + "tokens_seen": 1352859648 + }, + { + "epoch": 0.41, + "learning_rate": 0.0004768545636910733, + "loss": 1.322, + "theoretical_loss": 3.5770062009127104, + "tokens_seen": 1352925184 + }, + { + "epoch": 0.41, + "learning_rate": 0.00047683851554663993, + "loss": 1.407, + "theoretical_loss": 3.576990744142108, + "tokens_seen": 1352990720 + }, + { + "epoch": 0.41, + "learning_rate": 0.0004768224674022067, + "loss": 1.3736, + "theoretical_loss": 3.5769752883298027, + "tokens_seen": 1353056256 + }, + { + "epoch": 0.41, + "learning_rate": 0.00047680641925777333, + "loss": 1.3883, + "theoretical_loss": 3.576959833475689, + "tokens_seen": 1353121792 + }, + { + "epoch": 0.41, + "learning_rate": 0.0004767903711133401, + "loss": 1.3168, + "theoretical_loss": 3.576944379579662, + "tokens_seen": 1353187328 + }, + { + "epoch": 0.41, + "learning_rate": 0.00047677432296890673, + "loss": 1.344, + "theoretical_loss": 3.5769289266416147, + "tokens_seen": 1353252864 + }, + { + "epoch": 0.41, + "objective/train/docs_used": 494475, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.3597676753997803, + "objective/train/theoretical_loss": 3.5769134746614415, + "objective/train/tokens_used": 1373778400, + "theoretical_loss": 3.5769134746614415, + "tokens_seen": 1353318400 + }, + { + "epoch": 0.41, + "learning_rate": 0.0004767582748244735, + "loss": 1.368, + "theoretical_loss": 3.5769134746614415, + "tokens_seen": 1353318400 + }, + { + "epoch": 0.41, + "learning_rate": 0.00047674222668004013, + "loss": 1.3708, + "theoretical_loss": 3.5768980236390373, + "tokens_seen": 1353383936 + }, + { + "epoch": 0.41, + "learning_rate": 0.0004767261785356069, + "loss": 1.4477, + "theoretical_loss": 3.576882573574296, + "tokens_seen": 1353449472 + }, + { + "epoch": 0.41, + "learning_rate": 0.00047671013039117353, + "loss": 1.3896, + "theoretical_loss": 3.5768671244671113, + "tokens_seen": 1353515008 + }, + { + "epoch": 0.41, + "learning_rate": 0.0004766940822467403, + "loss": 1.3878, + "theoretical_loss": 3.5768516763173785, + "tokens_seen": 1353580544 + }, + { + "epoch": 0.41, + "learning_rate": 0.00047667803410230694, + "loss": 1.4098, + "theoretical_loss": 3.5768362291249916, + "tokens_seen": 1353646080 + }, + { + "epoch": 0.41, + "learning_rate": 0.0004766619859578737, + "loss": 1.3814, + "theoretical_loss": 3.576820782889844, + "tokens_seen": 1353711616 + }, + { + "epoch": 0.41, + "learning_rate": 0.00047664593781344034, + "loss": 1.3895, + "theoretical_loss": 3.5768053376118316, + "tokens_seen": 1353777152 + }, + { + "epoch": 0.41, + "learning_rate": 0.0004766298896690071, + "loss": 1.4067, + "theoretical_loss": 3.576789893290848, + "tokens_seen": 1353842688 + }, + { + "epoch": 0.41, + "learning_rate": 0.00047661384152457374, + "loss": 1.536, + "theoretical_loss": 3.576774449926787, + "tokens_seen": 1353908224 + }, + { + "epoch": 0.41, + "learning_rate": 0.0004765977933801405, + "loss": 1.4694, + "theoretical_loss": 3.576759007519544, + "tokens_seen": 1353973760 + }, + { + "epoch": 0.41, + "learning_rate": 0.00047658174523570714, + "loss": 1.3729, + "theoretical_loss": 3.5767435660690126, + "tokens_seen": 1354039296 + }, + { + "epoch": 0.41, + "learning_rate": 0.0004765656970912739, + "loss": 1.4064, + "theoretical_loss": 3.5767281255750882, + "tokens_seen": 1354104832 + }, + { + "epoch": 0.41, + "learning_rate": 0.00047654964894684054, + "loss": 1.2649, + "theoretical_loss": 3.5767126860376637, + "tokens_seen": 1354170368 + }, + { + "epoch": 0.41, + "learning_rate": 0.0004765336008024073, + "loss": 1.4211, + "theoretical_loss": 3.576697247456635, + "tokens_seen": 1354235904 + }, + { + "epoch": 0.41, + "learning_rate": 0.00047651755265797394, + "loss": 1.4463, + "theoretical_loss": 3.576681809831896, + "tokens_seen": 1354301440 + }, + { + "epoch": 0.41, + "learning_rate": 0.0004765015045135407, + "loss": 1.3639, + "theoretical_loss": 3.5766663731633415, + "tokens_seen": 1354366976 + }, + { + "epoch": 0.41, + "learning_rate": 0.00047648545636910734, + "loss": 1.3047, + "theoretical_loss": 3.5766509374508653, + "tokens_seen": 1354432512 + }, + { + "epoch": 0.41, + "learning_rate": 0.0004764694082246741, + "loss": 1.4548, + "theoretical_loss": 3.5766355026943626, + "tokens_seen": 1354498048 + }, + { + "epoch": 0.41, + "learning_rate": 0.00047645336008024074, + "loss": 1.4717, + "theoretical_loss": 3.576620068893728, + "tokens_seen": 1354563584 + }, + { + "epoch": 0.41, + "learning_rate": 0.0004764373119358075, + "loss": 1.3746, + "theoretical_loss": 3.5766046360488555, + "tokens_seen": 1354629120 + }, + { + "epoch": 0.41, + "learning_rate": 0.00047642126379137415, + "loss": 1.3616, + "theoretical_loss": 3.57658920415964, + "tokens_seen": 1354694656 + }, + { + "epoch": 0.41, + "learning_rate": 0.00047640521564694085, + "loss": 1.3306, + "theoretical_loss": 3.5765737732259764, + "tokens_seen": 1354760192 + }, + { + "epoch": 0.41, + "learning_rate": 0.00047638916750250755, + "loss": 1.3705, + "theoretical_loss": 3.5765583432477586, + "tokens_seen": 1354825728 + }, + { + "epoch": 0.41, + "learning_rate": 0.00047637311935807425, + "loss": 1.4124, + "theoretical_loss": 3.5765429142248815, + "tokens_seen": 1354891264 + }, + { + "epoch": 0.41, + "learning_rate": 0.00047635707121364095, + "loss": 1.4102, + "theoretical_loss": 3.5765274861572403, + "tokens_seen": 1354956800 + }, + { + "epoch": 0.41, + "learning_rate": 0.00047634102306920765, + "loss": 1.3334, + "theoretical_loss": 3.5765120590447292, + "tokens_seen": 1355022336 + }, + { + "epoch": 0.41, + "learning_rate": 0.00047632497492477435, + "loss": 1.3832, + "theoretical_loss": 3.5764966328872427, + "tokens_seen": 1355087872 + }, + { + "epoch": 0.41, + "learning_rate": 0.00047630892678034105, + "loss": 1.3281, + "theoretical_loss": 3.576481207684676, + "tokens_seen": 1355153408 + }, + { + "epoch": 0.41, + "learning_rate": 0.00047629287863590775, + "loss": 1.4673, + "theoretical_loss": 3.576465783436923, + "tokens_seen": 1355218944 + }, + { + "epoch": 0.41, + "learning_rate": 0.00047627683049147445, + "loss": 1.3769, + "theoretical_loss": 3.576450360143879, + "tokens_seen": 1355284480 + }, + { + "epoch": 0.41, + "learning_rate": 0.0004762607823470411, + "loss": 1.4065, + "theoretical_loss": 3.576434937805439, + "tokens_seen": 1355350016 + }, + { + "epoch": 0.41, + "learning_rate": 0.00047624473420260785, + "loss": 1.3956, + "theoretical_loss": 3.576419516421497, + "tokens_seen": 1355415552 + }, + { + "epoch": 0.41, + "learning_rate": 0.0004762286860581745, + "loss": 1.3017, + "theoretical_loss": 3.5764040959919488, + "tokens_seen": 1355481088 + }, + { + "epoch": 0.41, + "learning_rate": 0.00047621263791374125, + "loss": 1.3197, + "theoretical_loss": 3.5763886765166886, + "tokens_seen": 1355546624 + }, + { + "epoch": 0.41, + "learning_rate": 0.0004761965897693079, + "loss": 1.4716, + "theoretical_loss": 3.576373257995611, + "tokens_seen": 1355612160 + }, + { + "epoch": 0.41, + "learning_rate": 0.00047618054162487465, + "loss": 1.4269, + "theoretical_loss": 3.576357840428611, + "tokens_seen": 1355677696 + }, + { + "epoch": 0.41, + "learning_rate": 0.0004761644934804413, + "loss": 1.377, + "theoretical_loss": 3.5763424238155834, + "tokens_seen": 1355743232 + }, + { + "epoch": 0.41, + "learning_rate": 0.00047614844533600806, + "loss": 1.4213, + "theoretical_loss": 3.576327008156423, + "tokens_seen": 1355808768 + }, + { + "epoch": 0.41, + "learning_rate": 0.0004761323971915747, + "loss": 1.4429, + "theoretical_loss": 3.5763115934510257, + "tokens_seen": 1355874304 + }, + { + "epoch": 0.41, + "learning_rate": 0.00047611634904714146, + "loss": 1.3334, + "theoretical_loss": 3.576296179699285, + "tokens_seen": 1355939840 + }, + { + "epoch": 0.41, + "learning_rate": 0.0004761003009027081, + "loss": 1.3956, + "theoretical_loss": 3.576280766901096, + "tokens_seen": 1356005376 + }, + { + "epoch": 0.41, + "learning_rate": 0.00047608425275827486, + "loss": 1.2558, + "theoretical_loss": 3.5762653550563543, + "tokens_seen": 1356070912 + }, + { + "epoch": 0.41, + "learning_rate": 0.0004760682046138415, + "loss": 1.2563, + "theoretical_loss": 3.576249944164955, + "tokens_seen": 1356136448 + }, + { + "epoch": 0.41, + "learning_rate": 0.00047605215646940826, + "loss": 1.3754, + "theoretical_loss": 3.576234534226792, + "tokens_seen": 1356201984 + }, + { + "epoch": 0.41, + "learning_rate": 0.0004760361083249749, + "loss": 1.4055, + "theoretical_loss": 3.576219125241761, + "tokens_seen": 1356267520 + }, + { + "epoch": 0.41, + "learning_rate": 0.00047602006018054166, + "loss": 1.336, + "theoretical_loss": 3.576203717209757, + "tokens_seen": 1356333056 + }, + { + "epoch": 0.41, + "learning_rate": 0.0004760040120361083, + "loss": 1.2518, + "theoretical_loss": 3.576188310130674, + "tokens_seen": 1356398592 + }, + { + "epoch": 0.41, + "learning_rate": 0.00047598796389167506, + "loss": 1.4123, + "theoretical_loss": 3.576172904004409, + "tokens_seen": 1356464128 + }, + { + "epoch": 0.41, + "learning_rate": 0.0004759719157472417, + "loss": 1.3899, + "theoretical_loss": 3.576157498830856, + "tokens_seen": 1356529664 + }, + { + "epoch": 0.41, + "objective/train/docs_used": 495730, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.5682625770568848, + "objective/train/theoretical_loss": 3.576142094609909, + "objective/train/tokens_used": 1377055200, + "theoretical_loss": 3.576142094609909, + "tokens_seen": 1356595200 + }, + { + "epoch": 0.41, + "learning_rate": 0.00047595586760280846, + "loss": 1.4145, + "theoretical_loss": 3.576142094609909, + "tokens_seen": 1356595200 + }, + { + "epoch": 0.41, + "learning_rate": 0.0004759398194583751, + "loss": 1.4107, + "theoretical_loss": 3.5761266913414653, + "tokens_seen": 1356660736 + }, + { + "epoch": 0.41, + "learning_rate": 0.00047592377131394186, + "loss": 1.4486, + "theoretical_loss": 3.576111289025418, + "tokens_seen": 1356726272 + }, + { + "epoch": 0.41, + "learning_rate": 0.0004759077231695085, + "loss": 1.4959, + "theoretical_loss": 3.576095887661663, + "tokens_seen": 1356791808 + }, + { + "epoch": 0.41, + "learning_rate": 0.00047589167502507527, + "loss": 1.3337, + "theoretical_loss": 3.5760804872500955, + "tokens_seen": 1356857344 + }, + { + "epoch": 0.41, + "learning_rate": 0.0004758756268806419, + "loss": 1.3974, + "theoretical_loss": 3.576065087790611, + "tokens_seen": 1356922880 + }, + { + "epoch": 0.41, + "learning_rate": 0.00047585957873620867, + "loss": 1.4998, + "theoretical_loss": 3.5760496892831037, + "tokens_seen": 1356988416 + }, + { + "epoch": 0.41, + "learning_rate": 0.0004758435305917753, + "loss": 1.3986, + "theoretical_loss": 3.5760342917274697, + "tokens_seen": 1357053952 + }, + { + "epoch": 0.41, + "learning_rate": 0.00047582748244734207, + "loss": 1.3959, + "theoretical_loss": 3.576018895123603, + "tokens_seen": 1357119488 + }, + { + "epoch": 0.41, + "learning_rate": 0.0004758114343029087, + "loss": 1.3536, + "theoretical_loss": 3.5760034994714003, + "tokens_seen": 1357185024 + }, + { + "epoch": 0.41, + "learning_rate": 0.00047579538615847547, + "loss": 1.3176, + "theoretical_loss": 3.5759881047707562, + "tokens_seen": 1357250560 + }, + { + "epoch": 0.41, + "learning_rate": 0.0004757793380140421, + "loss": 1.4311, + "theoretical_loss": 3.5759727110215658, + "tokens_seen": 1357316096 + }, + { + "epoch": 0.41, + "learning_rate": 0.00047576328986960887, + "loss": 1.4441, + "theoretical_loss": 3.5759573182237245, + "tokens_seen": 1357381632 + }, + { + "epoch": 0.41, + "learning_rate": 0.0004757472417251755, + "loss": 1.247, + "theoretical_loss": 3.5759419263771273, + "tokens_seen": 1357447168 + }, + { + "epoch": 0.41, + "learning_rate": 0.00047573119358074227, + "loss": 1.4039, + "theoretical_loss": 3.57592653548167, + "tokens_seen": 1357512704 + }, + { + "epoch": 0.41, + "learning_rate": 0.0004757151454363089, + "loss": 1.2391, + "theoretical_loss": 3.575911145537247, + "tokens_seen": 1357578240 + }, + { + "epoch": 0.41, + "learning_rate": 0.0004756990972918757, + "loss": 1.4106, + "theoretical_loss": 3.5758957565437552, + "tokens_seen": 1357643776 + }, + { + "epoch": 0.41, + "learning_rate": 0.0004756830491474423, + "loss": 1.4452, + "theoretical_loss": 3.5758803685010885, + "tokens_seen": 1357709312 + }, + { + "epoch": 0.41, + "learning_rate": 0.0004756670010030091, + "loss": 1.3017, + "theoretical_loss": 3.5758649814091426, + "tokens_seen": 1357774848 + }, + { + "epoch": 0.41, + "learning_rate": 0.0004756509528585758, + "loss": 1.3742, + "theoretical_loss": 3.5758495952678135, + "tokens_seen": 1357840384 + }, + { + "epoch": 0.41, + "learning_rate": 0.0004756349047141424, + "loss": 1.4187, + "theoretical_loss": 3.5758342100769958, + "tokens_seen": 1357905920 + }, + { + "epoch": 0.41, + "learning_rate": 0.0004756188565697092, + "loss": 1.3356, + "theoretical_loss": 3.5758188258365853, + "tokens_seen": 1357971456 + }, + { + "epoch": 0.41, + "learning_rate": 0.0004756028084252758, + "loss": 1.3629, + "theoretical_loss": 3.5758034425464773, + "tokens_seen": 1358036992 + }, + { + "epoch": 0.41, + "learning_rate": 0.0004755867602808426, + "loss": 1.4915, + "theoretical_loss": 3.5757880602065675, + "tokens_seen": 1358102528 + }, + { + "epoch": 0.41, + "learning_rate": 0.0004755707121364092, + "loss": 1.4193, + "theoretical_loss": 3.575772678816751, + "tokens_seen": 1358168064 + }, + { + "epoch": 0.41, + "learning_rate": 0.000475554663991976, + "loss": 1.4575, + "theoretical_loss": 3.5757572983769235, + "tokens_seen": 1358233600 + }, + { + "epoch": 0.41, + "learning_rate": 0.0004755386158475426, + "loss": 1.2962, + "theoretical_loss": 3.5757419188869806, + "tokens_seen": 1358299136 + }, + { + "epoch": 0.41, + "learning_rate": 0.0004755225677031094, + "loss": 1.3433, + "theoretical_loss": 3.5757265403468175, + "tokens_seen": 1358364672 + }, + { + "epoch": 0.41, + "learning_rate": 0.00047550651955867603, + "loss": 1.447, + "theoretical_loss": 3.5757111627563294, + "tokens_seen": 1358430208 + }, + { + "epoch": 0.41, + "learning_rate": 0.0004754904714142428, + "loss": 1.3475, + "theoretical_loss": 3.575695786115413, + "tokens_seen": 1358495744 + }, + { + "epoch": 0.41, + "learning_rate": 0.00047547442326980943, + "loss": 1.3287, + "theoretical_loss": 3.5756804104239626, + "tokens_seen": 1358561280 + }, + { + "epoch": 0.41, + "learning_rate": 0.0004754583751253762, + "loss": 1.4319, + "theoretical_loss": 3.575665035681875, + "tokens_seen": 1358626816 + }, + { + "epoch": 0.41, + "learning_rate": 0.00047544232698094283, + "loss": 1.315, + "theoretical_loss": 3.575649661889045, + "tokens_seen": 1358692352 + }, + { + "epoch": 0.41, + "learning_rate": 0.0004754262788365096, + "loss": 1.4089, + "theoretical_loss": 3.575634289045368, + "tokens_seen": 1358757888 + }, + { + "epoch": 0.41, + "learning_rate": 0.00047541023069207623, + "loss": 1.4714, + "theoretical_loss": 3.5756189171507406, + "tokens_seen": 1358823424 + }, + { + "epoch": 0.41, + "learning_rate": 0.000475394182547643, + "loss": 1.3268, + "theoretical_loss": 3.5756035462050573, + "tokens_seen": 1358888960 + }, + { + "epoch": 0.41, + "learning_rate": 0.00047537813440320963, + "loss": 1.3751, + "theoretical_loss": 3.575588176208214, + "tokens_seen": 1358954496 + }, + { + "epoch": 0.41, + "learning_rate": 0.0004753620862587764, + "loss": 1.4075, + "theoretical_loss": 3.5755728071601074, + "tokens_seen": 1359020032 + }, + { + "epoch": 0.41, + "learning_rate": 0.00047534603811434303, + "loss": 1.4226, + "theoretical_loss": 3.575557439060632, + "tokens_seen": 1359085568 + }, + { + "epoch": 0.41, + "learning_rate": 0.0004753299899699098, + "loss": 1.381, + "theoretical_loss": 3.575542071909684, + "tokens_seen": 1359151104 + }, + { + "epoch": 0.41, + "learning_rate": 0.00047531394182547643, + "loss": 1.3901, + "theoretical_loss": 3.5755267057071594, + "tokens_seen": 1359216640 + }, + { + "epoch": 0.41, + "learning_rate": 0.0004752978936810432, + "loss": 1.3546, + "theoretical_loss": 3.5755113404529535, + "tokens_seen": 1359282176 + }, + { + "epoch": 0.41, + "learning_rate": 0.00047528184553660984, + "loss": 1.2531, + "theoretical_loss": 3.5754959761469616, + "tokens_seen": 1359347712 + }, + { + "epoch": 0.41, + "learning_rate": 0.0004752657973921766, + "loss": 1.4229, + "theoretical_loss": 3.5754806127890806, + "tokens_seen": 1359413248 + }, + { + "epoch": 0.41, + "learning_rate": 0.00047524974924774324, + "loss": 1.3798, + "theoretical_loss": 3.5754652503792057, + "tokens_seen": 1359478784 + }, + { + "epoch": 0.41, + "learning_rate": 0.00047523370110331, + "loss": 1.3901, + "theoretical_loss": 3.575449888917233, + "tokens_seen": 1359544320 + }, + { + "epoch": 0.41, + "learning_rate": 0.00047521765295887664, + "loss": 1.2475, + "theoretical_loss": 3.5754345284030573, + "tokens_seen": 1359609856 + }, + { + "epoch": 0.41, + "learning_rate": 0.0004752016048144434, + "loss": 1.4336, + "theoretical_loss": 3.575419168836576, + "tokens_seen": 1359675392 + }, + { + "epoch": 0.41, + "learning_rate": 0.00047518555667001004, + "loss": 1.3541, + "theoretical_loss": 3.5754038102176837, + "tokens_seen": 1359740928 + }, + { + "epoch": 0.41, + "learning_rate": 0.0004751695085255768, + "loss": 1.4047, + "theoretical_loss": 3.575388452546277, + "tokens_seen": 1359806464 + }, + { + "epoch": 0.41, + "objective/train/docs_used": 496861, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.3347420692443848, + "objective/train/theoretical_loss": 3.5753730958222514, + "objective/train/tokens_used": 1380332000, + "theoretical_loss": 3.5753730958222514, + "tokens_seen": 1359872000 + }, + { + "epoch": 0.41, + "learning_rate": 0.00047515346038114344, + "loss": 1.2374, + "theoretical_loss": 3.5753730958222514, + "tokens_seen": 1359872000 + }, + { + "epoch": 0.41, + "learning_rate": 0.0004751374122367102, + "loss": 1.4482, + "theoretical_loss": 3.5753577400455026, + "tokens_seen": 1359937536 + }, + { + "epoch": 0.41, + "learning_rate": 0.00047512136409227684, + "loss": 1.3495, + "theoretical_loss": 3.575342385215927, + "tokens_seen": 1360003072 + }, + { + "epoch": 0.41, + "learning_rate": 0.0004751053159478436, + "loss": 1.4219, + "theoretical_loss": 3.575327031333421, + "tokens_seen": 1360068608 + }, + { + "epoch": 0.41, + "learning_rate": 0.00047508926780341024, + "loss": 1.3998, + "theoretical_loss": 3.5753116783978793, + "tokens_seen": 1360134144 + }, + { + "epoch": 0.41, + "learning_rate": 0.000475073219658977, + "loss": 1.3309, + "theoretical_loss": 3.5752963264091986, + "tokens_seen": 1360199680 + }, + { + "epoch": 0.41, + "learning_rate": 0.00047505717151454364, + "loss": 1.3924, + "theoretical_loss": 3.575280975367275, + "tokens_seen": 1360265216 + }, + { + "epoch": 0.41, + "learning_rate": 0.0004750411233701104, + "loss": 1.427, + "theoretical_loss": 3.5752656252720048, + "tokens_seen": 1360330752 + }, + { + "epoch": 0.41, + "learning_rate": 0.00047502507522567705, + "loss": 1.5432, + "theoretical_loss": 3.5752502761232825, + "tokens_seen": 1360396288 + }, + { + "epoch": 0.41, + "learning_rate": 0.00047500902708124375, + "loss": 1.3408, + "theoretical_loss": 3.575234927921006, + "tokens_seen": 1360461824 + }, + { + "epoch": 0.41, + "learning_rate": 0.00047499297893681045, + "loss": 1.4535, + "theoretical_loss": 3.57521958066507, + "tokens_seen": 1360527360 + }, + { + "epoch": 0.41, + "learning_rate": 0.00047497693079237715, + "loss": 1.3667, + "theoretical_loss": 3.575204234355371, + "tokens_seen": 1360592896 + }, + { + "epoch": 0.41, + "learning_rate": 0.00047496088264794385, + "loss": 1.3655, + "theoretical_loss": 3.5751888889918053, + "tokens_seen": 1360658432 + }, + { + "epoch": 0.41, + "learning_rate": 0.00047494483450351055, + "loss": 1.382, + "theoretical_loss": 3.575173544574269, + "tokens_seen": 1360723968 + }, + { + "epoch": 0.41, + "learning_rate": 0.00047492878635907725, + "loss": 1.336, + "theoretical_loss": 3.575158201102658, + "tokens_seen": 1360789504 + }, + { + "epoch": 0.41, + "learning_rate": 0.00047491273821464395, + "loss": 1.3108, + "theoretical_loss": 3.5751428585768688, + "tokens_seen": 1360855040 + }, + { + "epoch": 0.41, + "learning_rate": 0.00047489669007021065, + "loss": 1.3717, + "theoretical_loss": 3.5751275169967975, + "tokens_seen": 1360920576 + }, + { + "epoch": 0.41, + "learning_rate": 0.00047488064192577735, + "loss": 1.476, + "theoretical_loss": 3.5751121763623397, + "tokens_seen": 1360986112 + }, + { + "epoch": 0.41, + "learning_rate": 0.000474864593781344, + "loss": 1.305, + "theoretical_loss": 3.5750968366733917, + "tokens_seen": 1361051648 + }, + { + "epoch": 0.41, + "learning_rate": 0.00047484854563691075, + "loss": 1.4279, + "theoretical_loss": 3.57508149792985, + "tokens_seen": 1361117184 + }, + { + "epoch": 0.41, + "learning_rate": 0.0004748324974924774, + "loss": 1.432, + "theoretical_loss": 3.5750661601316116, + "tokens_seen": 1361182720 + }, + { + "epoch": 0.41, + "learning_rate": 0.00047481644934804415, + "loss": 1.3193, + "theoretical_loss": 3.5750508232785707, + "tokens_seen": 1361248256 + }, + { + "epoch": 0.41, + "learning_rate": 0.0004748004012036108, + "loss": 1.3287, + "theoretical_loss": 3.5750354873706254, + "tokens_seen": 1361313792 + }, + { + "epoch": 0.41, + "learning_rate": 0.00047478435305917756, + "loss": 1.2723, + "theoretical_loss": 3.5750201524076712, + "tokens_seen": 1361379328 + }, + { + "epoch": 0.41, + "learning_rate": 0.0004747683049147442, + "loss": 1.3218, + "theoretical_loss": 3.5750048183896044, + "tokens_seen": 1361444864 + }, + { + "epoch": 0.41, + "learning_rate": 0.00047475225677031096, + "loss": 1.4536, + "theoretical_loss": 3.5749894853163218, + "tokens_seen": 1361510400 + }, + { + "epoch": 0.41, + "learning_rate": 0.0004747362086258776, + "loss": 1.2775, + "theoretical_loss": 3.5749741531877186, + "tokens_seen": 1361575936 + }, + { + "epoch": 0.41, + "learning_rate": 0.00047472016048144436, + "loss": 1.5185, + "theoretical_loss": 3.5749588220036923, + "tokens_seen": 1361641472 + }, + { + "epoch": 0.41, + "learning_rate": 0.000474704112337011, + "loss": 1.2921, + "theoretical_loss": 3.574943491764139, + "tokens_seen": 1361707008 + }, + { + "epoch": 0.41, + "learning_rate": 0.00047468806419257776, + "loss": 1.4203, + "theoretical_loss": 3.5749281624689546, + "tokens_seen": 1361772544 + }, + { + "epoch": 0.41, + "learning_rate": 0.0004746720160481444, + "loss": 1.57, + "theoretical_loss": 3.5749128341180354, + "tokens_seen": 1361838080 + }, + { + "epoch": 0.41, + "learning_rate": 0.00047465596790371116, + "loss": 1.3008, + "theoretical_loss": 3.5748975067112783, + "tokens_seen": 1361903616 + }, + { + "epoch": 0.41, + "learning_rate": 0.0004746399197592778, + "loss": 1.4057, + "theoretical_loss": 3.5748821802485797, + "tokens_seen": 1361969152 + }, + { + "epoch": 0.41, + "learning_rate": 0.00047462387161484456, + "loss": 1.3643, + "theoretical_loss": 3.574866854729836, + "tokens_seen": 1362034688 + }, + { + "epoch": 0.41, + "learning_rate": 0.0004746078234704112, + "loss": 1.3478, + "theoretical_loss": 3.5748515301549433, + "tokens_seen": 1362100224 + }, + { + "epoch": 0.41, + "learning_rate": 0.00047459177532597796, + "loss": 1.3719, + "theoretical_loss": 3.5748362065237984, + "tokens_seen": 1362165760 + }, + { + "epoch": 0.41, + "learning_rate": 0.0004745757271815446, + "loss": 1.5077, + "theoretical_loss": 3.5748208838362974, + "tokens_seen": 1362231296 + }, + { + "epoch": 0.41, + "learning_rate": 0.00047455967903711136, + "loss": 1.3383, + "theoretical_loss": 3.574805562092337, + "tokens_seen": 1362296832 + }, + { + "epoch": 0.41, + "learning_rate": 0.000474543630892678, + "loss": 1.4278, + "theoretical_loss": 3.574790241291814, + "tokens_seen": 1362362368 + }, + { + "epoch": 0.41, + "learning_rate": 0.00047452758274824477, + "loss": 1.4584, + "theoretical_loss": 3.574774921434625, + "tokens_seen": 1362427904 + }, + { + "epoch": 0.41, + "learning_rate": 0.0004745115346038114, + "loss": 1.2908, + "theoretical_loss": 3.5747596025206656, + "tokens_seen": 1362493440 + }, + { + "epoch": 0.41, + "learning_rate": 0.00047449548645937817, + "loss": 1.3649, + "theoretical_loss": 3.574744284549834, + "tokens_seen": 1362558976 + }, + { + "epoch": 0.41, + "learning_rate": 0.0004744794383149448, + "loss": 1.3344, + "theoretical_loss": 3.5747289675220246, + "tokens_seen": 1362624512 + }, + { + "epoch": 0.41, + "learning_rate": 0.00047446339017051157, + "loss": 1.4899, + "theoretical_loss": 3.5747136514371363, + "tokens_seen": 1362690048 + }, + { + "epoch": 0.41, + "learning_rate": 0.0004744473420260783, + "loss": 1.5057, + "theoretical_loss": 3.5746983362950635, + "tokens_seen": 1362755584 + }, + { + "epoch": 0.41, + "learning_rate": 0.00047443129388164497, + "loss": 1.3257, + "theoretical_loss": 3.5746830220957047, + "tokens_seen": 1362821120 + }, + { + "epoch": 0.41, + "learning_rate": 0.0004744152457372117, + "loss": 1.2954, + "theoretical_loss": 3.5746677088389553, + "tokens_seen": 1362886656 + }, + { + "epoch": 0.41, + "learning_rate": 0.00047439919759277837, + "loss": 1.472, + "theoretical_loss": 3.574652396524713, + "tokens_seen": 1362952192 + }, + { + "epoch": 0.41, + "learning_rate": 0.0004743831494483451, + "loss": 1.3443, + "theoretical_loss": 3.5746370851528733, + "tokens_seen": 1363017728 + }, + { + "epoch": 0.41, + "learning_rate": 0.00047436710130391177, + "loss": 1.3887, + "theoretical_loss": 3.5746217747233335, + "tokens_seen": 1363083264 + }, + { + "epoch": 0.41, + "objective/train/docs_used": 498147, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.3789830207824707, + "objective/train/theoretical_loss": 3.5746064652359903, + "objective/train/tokens_used": 1383608800, + "theoretical_loss": 3.5746064652359903, + "tokens_seen": 1363148800 + }, + { + "epoch": 0.41, + "learning_rate": 0.0004743510531594785, + "loss": 1.3622, + "theoretical_loss": 3.5746064652359903, + "tokens_seen": 1363148800 + }, + { + "epoch": 0.41, + "learning_rate": 0.0004743350050150452, + "loss": 1.3269, + "theoretical_loss": 3.574591156690741, + "tokens_seen": 1363214336 + }, + { + "epoch": 0.41, + "learning_rate": 0.0004743189568706119, + "loss": 1.3234, + "theoretical_loss": 3.574575849087481, + "tokens_seen": 1363279872 + }, + { + "epoch": 0.41, + "learning_rate": 0.0004743029087261786, + "loss": 1.4083, + "theoretical_loss": 3.5745605424261084, + "tokens_seen": 1363345408 + }, + { + "epoch": 0.41, + "learning_rate": 0.0004742868605817453, + "loss": 1.3617, + "theoretical_loss": 3.574545236706519, + "tokens_seen": 1363410944 + }, + { + "epoch": 0.41, + "learning_rate": 0.000474270812437312, + "loss": 1.3855, + "theoretical_loss": 3.57452993192861, + "tokens_seen": 1363476480 + }, + { + "epoch": 0.41, + "learning_rate": 0.0004742547642928787, + "loss": 1.2776, + "theoretical_loss": 3.5745146280922784, + "tokens_seen": 1363542016 + }, + { + "epoch": 0.41, + "learning_rate": 0.0004742387161484453, + "loss": 1.4862, + "theoretical_loss": 3.5744993251974204, + "tokens_seen": 1363607552 + }, + { + "epoch": 0.41, + "learning_rate": 0.0004742226680040121, + "loss": 1.311, + "theoretical_loss": 3.5744840232439334, + "tokens_seen": 1363673088 + }, + { + "epoch": 0.41, + "learning_rate": 0.0004742066198595787, + "loss": 1.3946, + "theoretical_loss": 3.5744687222317144, + "tokens_seen": 1363738624 + }, + { + "epoch": 0.41, + "learning_rate": 0.0004741905717151455, + "loss": 1.4642, + "theoretical_loss": 3.5744534221606594, + "tokens_seen": 1363804160 + }, + { + "epoch": 0.41, + "learning_rate": 0.0004741745235707121, + "loss": 1.4091, + "theoretical_loss": 3.5744381230306663, + "tokens_seen": 1363869696 + }, + { + "epoch": 0.41, + "learning_rate": 0.0004741584754262789, + "loss": 1.4191, + "theoretical_loss": 3.5744228248416317, + "tokens_seen": 1363935232 + }, + { + "epoch": 0.41, + "learning_rate": 0.0004741424272818455, + "loss": 1.4744, + "theoretical_loss": 3.5744075275934515, + "tokens_seen": 1364000768 + }, + { + "epoch": 0.41, + "learning_rate": 0.0004741263791374123, + "loss": 1.3214, + "theoretical_loss": 3.5743922312860246, + "tokens_seen": 1364066304 + }, + { + "epoch": 0.41, + "learning_rate": 0.00047411033099297893, + "loss": 1.4342, + "theoretical_loss": 3.574376935919246, + "tokens_seen": 1364131840 + }, + { + "epoch": 0.41, + "learning_rate": 0.0004740942828485457, + "loss": 1.4145, + "theoretical_loss": 3.574361641493014, + "tokens_seen": 1364197376 + }, + { + "epoch": 0.41, + "learning_rate": 0.00047407823470411233, + "loss": 1.4494, + "theoretical_loss": 3.574346348007225, + "tokens_seen": 1364262912 + }, + { + "epoch": 0.41, + "learning_rate": 0.0004740621865596791, + "loss": 1.3194, + "theoretical_loss": 3.574331055461776, + "tokens_seen": 1364328448 + }, + { + "epoch": 0.41, + "learning_rate": 0.00047404613841524573, + "loss": 1.5183, + "theoretical_loss": 3.574315763856564, + "tokens_seen": 1364393984 + }, + { + "epoch": 0.41, + "learning_rate": 0.0004740300902708125, + "loss": 1.381, + "theoretical_loss": 3.5743004731914865, + "tokens_seen": 1364459520 + }, + { + "epoch": 0.41, + "learning_rate": 0.00047401404212637913, + "loss": 1.4233, + "theoretical_loss": 3.57428518346644, + "tokens_seen": 1364525056 + }, + { + "epoch": 0.41, + "learning_rate": 0.0004739979939819459, + "loss": 1.31, + "theoretical_loss": 3.574269894681322, + "tokens_seen": 1364590592 + }, + { + "epoch": 0.41, + "learning_rate": 0.00047398194583751253, + "loss": 1.352, + "theoretical_loss": 3.5742546068360292, + "tokens_seen": 1364656128 + }, + { + "epoch": 0.41, + "learning_rate": 0.0004739658976930793, + "loss": 1.2889, + "theoretical_loss": 3.574239319930459, + "tokens_seen": 1364721664 + }, + { + "epoch": 0.41, + "learning_rate": 0.00047394984954864593, + "loss": 1.4063, + "theoretical_loss": 3.5742240339645077, + "tokens_seen": 1364787200 + }, + { + "epoch": 0.41, + "learning_rate": 0.0004739338014042127, + "loss": 1.447, + "theoretical_loss": 3.574208748938074, + "tokens_seen": 1364852736 + }, + { + "epoch": 0.41, + "learning_rate": 0.00047391775325977934, + "loss": 1.3352, + "theoretical_loss": 3.574193464851054, + "tokens_seen": 1364918272 + }, + { + "epoch": 0.41, + "learning_rate": 0.0004739017051153461, + "loss": 1.379, + "theoretical_loss": 3.5741781817033447, + "tokens_seen": 1364983808 + }, + { + "epoch": 0.41, + "learning_rate": 0.00047388565697091274, + "loss": 1.386, + "theoretical_loss": 3.5741628994948433, + "tokens_seen": 1365049344 + }, + { + "epoch": 0.41, + "learning_rate": 0.0004738696088264795, + "loss": 1.3389, + "theoretical_loss": 3.5741476182254477, + "tokens_seen": 1365114880 + }, + { + "epoch": 0.41, + "learning_rate": 0.00047385356068204614, + "loss": 1.5314, + "theoretical_loss": 3.5741323378950542, + "tokens_seen": 1365180416 + }, + { + "epoch": 0.41, + "learning_rate": 0.0004738375125376129, + "loss": 1.3965, + "theoretical_loss": 3.5741170585035613, + "tokens_seen": 1365245952 + }, + { + "epoch": 0.41, + "learning_rate": 0.00047382146439317954, + "loss": 1.5005, + "theoretical_loss": 3.5741017800508645, + "tokens_seen": 1365311488 + }, + { + "epoch": 0.41, + "learning_rate": 0.0004738054162487463, + "loss": 1.4351, + "theoretical_loss": 3.574086502536863, + "tokens_seen": 1365377024 + }, + { + "epoch": 0.41, + "learning_rate": 0.00047378936810431294, + "loss": 1.5232, + "theoretical_loss": 3.574071225961452, + "tokens_seen": 1365442560 + }, + { + "epoch": 0.41, + "learning_rate": 0.0004737733199598797, + "loss": 1.2722, + "theoretical_loss": 3.574055950324531, + "tokens_seen": 1365508096 + }, + { + "epoch": 0.41, + "learning_rate": 0.00047375727181544634, + "loss": 1.3588, + "theoretical_loss": 3.574040675625996, + "tokens_seen": 1365573632 + }, + { + "epoch": 0.41, + "learning_rate": 0.0004737412236710131, + "loss": 1.375, + "theoretical_loss": 3.574025401865744, + "tokens_seen": 1365639168 + }, + { + "epoch": 0.41, + "learning_rate": 0.00047372517552657974, + "loss": 1.3866, + "theoretical_loss": 3.574010129043673, + "tokens_seen": 1365704704 + }, + { + "epoch": 0.41, + "learning_rate": 0.0004737091273821465, + "loss": 1.4504, + "theoretical_loss": 3.5739948571596805, + "tokens_seen": 1365770240 + }, + { + "epoch": 0.41, + "learning_rate": 0.00047369307923771314, + "loss": 1.434, + "theoretical_loss": 3.573979586213663, + "tokens_seen": 1365835776 + }, + { + "epoch": 0.41, + "learning_rate": 0.0004736770310932799, + "loss": 1.36, + "theoretical_loss": 3.573964316205519, + "tokens_seen": 1365901312 + }, + { + "epoch": 0.41, + "learning_rate": 0.00047366098294884655, + "loss": 1.5094, + "theoretical_loss": 3.5739490471351454, + "tokens_seen": 1365966848 + }, + { + "epoch": 0.41, + "learning_rate": 0.0004736449348044133, + "loss": 1.449, + "theoretical_loss": 3.5739337790024397, + "tokens_seen": 1366032384 + }, + { + "epoch": 0.41, + "learning_rate": 0.00047362888665997995, + "loss": 1.5536, + "theoretical_loss": 3.5739185118072987, + "tokens_seen": 1366097920 + }, + { + "epoch": 0.41, + "learning_rate": 0.0004736128385155467, + "loss": 1.2999, + "theoretical_loss": 3.5739032455496207, + "tokens_seen": 1366163456 + }, + { + "epoch": 0.41, + "learning_rate": 0.00047359679037111335, + "loss": 1.3462, + "theoretical_loss": 3.573887980229303, + "tokens_seen": 1366228992 + }, + { + "epoch": 0.41, + "learning_rate": 0.00047358074222668005, + "loss": 1.375, + "theoretical_loss": 3.5738727158462424, + "tokens_seen": 1366294528 + }, + { + "epoch": 0.41, + "learning_rate": 0.00047356469408224675, + "loss": 1.3802, + "theoretical_loss": 3.5738574524003375, + "tokens_seen": 1366360064 + }, + { + "epoch": 0.41, + "objective/train/docs_used": 499217, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.3140311241149902, + "objective/train/theoretical_loss": 3.573842189891485, + "objective/train/tokens_used": 1386885600, + "theoretical_loss": 3.573842189891485, + "tokens_seen": 1366425600 + }, + { + "epoch": 0.41, + "learning_rate": 0.00047354864593781345, + "loss": 1.3124, + "theoretical_loss": 3.573842189891485, + "tokens_seen": 1366425600 + }, + { + "epoch": 0.41, + "learning_rate": 0.00047353259779338015, + "loss": 1.3441, + "theoretical_loss": 3.573826928319583, + "tokens_seen": 1366491136 + }, + { + "epoch": 0.41, + "learning_rate": 0.00047351654964894685, + "loss": 1.3356, + "theoretical_loss": 3.573811667684528, + "tokens_seen": 1366556672 + }, + { + "epoch": 0.41, + "learning_rate": 0.00047350050150451355, + "loss": 1.4117, + "theoretical_loss": 3.5737964079862197, + "tokens_seen": 1366622208 + }, + { + "epoch": 0.41, + "learning_rate": 0.00047348445336008025, + "loss": 1.3142, + "theoretical_loss": 3.573781149224553, + "tokens_seen": 1366687744 + }, + { + "epoch": 0.41, + "learning_rate": 0.0004734684052156469, + "loss": 1.3551, + "theoretical_loss": 3.5737658913994275, + "tokens_seen": 1366753280 + }, + { + "epoch": 0.41, + "learning_rate": 0.00047345235707121365, + "loss": 1.3998, + "theoretical_loss": 3.57375063451074, + "tokens_seen": 1366818816 + }, + { + "epoch": 0.41, + "learning_rate": 0.0004734363089267803, + "loss": 1.3866, + "theoretical_loss": 3.573735378558388, + "tokens_seen": 1366884352 + }, + { + "epoch": 0.41, + "learning_rate": 0.00047342026078234706, + "loss": 1.3834, + "theoretical_loss": 3.5737201235422695, + "tokens_seen": 1366949888 + }, + { + "epoch": 0.41, + "learning_rate": 0.0004734042126379137, + "loss": 1.3944, + "theoretical_loss": 3.573704869462282, + "tokens_seen": 1367015424 + }, + { + "epoch": 0.41, + "learning_rate": 0.00047338816449348046, + "loss": 1.4388, + "theoretical_loss": 3.5736896163183234, + "tokens_seen": 1367080960 + }, + { + "epoch": 0.41, + "learning_rate": 0.0004733721163490471, + "loss": 1.5056, + "theoretical_loss": 3.573674364110291, + "tokens_seen": 1367146496 + }, + { + "epoch": 0.41, + "learning_rate": 0.00047335606820461386, + "loss": 1.3017, + "theoretical_loss": 3.573659112838083, + "tokens_seen": 1367212032 + }, + { + "epoch": 0.41, + "learning_rate": 0.0004733400200601805, + "loss": 1.3583, + "theoretical_loss": 3.573643862501597, + "tokens_seen": 1367277568 + }, + { + "epoch": 0.41, + "learning_rate": 0.00047332397191574726, + "loss": 1.4252, + "theoretical_loss": 3.5736286131007304, + "tokens_seen": 1367343104 + }, + { + "epoch": 0.41, + "learning_rate": 0.0004733079237713139, + "loss": 1.4388, + "theoretical_loss": 3.5736133646353805, + "tokens_seen": 1367408640 + }, + { + "epoch": 0.41, + "learning_rate": 0.00047329187562688066, + "loss": 1.4106, + "theoretical_loss": 3.5735981171054467, + "tokens_seen": 1367474176 + }, + { + "epoch": 0.41, + "learning_rate": 0.0004732758274824473, + "loss": 1.4475, + "theoretical_loss": 3.5735828705108252, + "tokens_seen": 1367539712 + }, + { + "epoch": 0.41, + "learning_rate": 0.00047325977933801406, + "loss": 1.3519, + "theoretical_loss": 3.5735676248514148, + "tokens_seen": 1367605248 + }, + { + "epoch": 0.41, + "learning_rate": 0.0004732437311935808, + "loss": 1.3866, + "theoretical_loss": 3.573552380127113, + "tokens_seen": 1367670784 + }, + { + "epoch": 0.41, + "learning_rate": 0.00047322768304914746, + "loss": 1.3136, + "theoretical_loss": 3.5735371363378174, + "tokens_seen": 1367736320 + }, + { + "epoch": 0.41, + "learning_rate": 0.0004732116349047142, + "loss": 1.3947, + "theoretical_loss": 3.5735218934834263, + "tokens_seen": 1367801856 + }, + { + "epoch": 0.41, + "learning_rate": 0.00047319558676028086, + "loss": 1.3601, + "theoretical_loss": 3.573506651563837, + "tokens_seen": 1367867392 + }, + { + "epoch": 0.41, + "learning_rate": 0.0004731795386158476, + "loss": 1.3736, + "theoretical_loss": 3.5734914105789484, + "tokens_seen": 1367932928 + }, + { + "epoch": 0.41, + "learning_rate": 0.00047316349047141427, + "loss": 1.3177, + "theoretical_loss": 3.5734761705286573, + "tokens_seen": 1367998464 + }, + { + "epoch": 0.41, + "learning_rate": 0.000473147442326981, + "loss": 1.4348, + "theoretical_loss": 3.573460931412862, + "tokens_seen": 1368064000 + }, + { + "epoch": 0.41, + "learning_rate": 0.00047313139418254767, + "loss": 1.3269, + "theoretical_loss": 3.57344569323146, + "tokens_seen": 1368129536 + }, + { + "epoch": 0.41, + "learning_rate": 0.0004731153460381144, + "loss": 1.3537, + "theoretical_loss": 3.5734304559843504, + "tokens_seen": 1368195072 + }, + { + "epoch": 0.41, + "learning_rate": 0.00047309929789368107, + "loss": 1.4808, + "theoretical_loss": 3.57341521967143, + "tokens_seen": 1368260608 + }, + { + "epoch": 0.41, + "learning_rate": 0.0004730832497492478, + "loss": 1.2579, + "theoretical_loss": 3.573399984292598, + "tokens_seen": 1368326144 + }, + { + "epoch": 0.41, + "learning_rate": 0.00047306720160481447, + "loss": 1.3977, + "theoretical_loss": 3.5733847498477513, + "tokens_seen": 1368391680 + }, + { + "epoch": 0.41, + "learning_rate": 0.0004730511534603812, + "loss": 1.4208, + "theoretical_loss": 3.5733695163367885, + "tokens_seen": 1368457216 + }, + { + "epoch": 0.41, + "learning_rate": 0.00047303510531594787, + "loss": 1.4167, + "theoretical_loss": 3.573354283759607, + "tokens_seen": 1368522752 + }, + { + "epoch": 0.41, + "learning_rate": 0.0004730190571715146, + "loss": 1.4394, + "theoretical_loss": 3.5733390521161055, + "tokens_seen": 1368588288 + }, + { + "epoch": 0.41, + "learning_rate": 0.00047300300902708127, + "loss": 1.4076, + "theoretical_loss": 3.573323821406182, + "tokens_seen": 1368653824 + }, + { + "epoch": 0.41, + "learning_rate": 0.000472986960882648, + "loss": 1.4397, + "theoretical_loss": 3.5733085916297345, + "tokens_seen": 1368719360 + }, + { + "epoch": 0.41, + "learning_rate": 0.0004729709127382147, + "loss": 1.4727, + "theoretical_loss": 3.5732933627866608, + "tokens_seen": 1368784896 + }, + { + "epoch": 0.41, + "learning_rate": 0.0004729548645937814, + "loss": 1.2712, + "theoretical_loss": 3.573278134876859, + "tokens_seen": 1368850432 + }, + { + "epoch": 0.41, + "learning_rate": 0.0004729388164493481, + "loss": 1.3762, + "theoretical_loss": 3.5732629079002276, + "tokens_seen": 1368915968 + }, + { + "epoch": 0.41, + "learning_rate": 0.0004729227683049148, + "loss": 1.4178, + "theoretical_loss": 3.5732476818566647, + "tokens_seen": 1368981504 + }, + { + "epoch": 0.41, + "learning_rate": 0.0004729067201604815, + "loss": 1.3016, + "theoretical_loss": 3.5732324567460685, + "tokens_seen": 1369047040 + }, + { + "epoch": 0.41, + "learning_rate": 0.0004728906720160482, + "loss": 1.3816, + "theoretical_loss": 3.573217232568337, + "tokens_seen": 1369112576 + }, + { + "epoch": 0.41, + "learning_rate": 0.0004728746238716149, + "loss": 1.3781, + "theoretical_loss": 3.5732020093233685, + "tokens_seen": 1369178112 + }, + { + "epoch": 0.41, + "learning_rate": 0.0004728585757271816, + "loss": 1.3499, + "theoretical_loss": 3.573186787011061, + "tokens_seen": 1369243648 + }, + { + "epoch": 0.41, + "learning_rate": 0.0004728425275827483, + "loss": 1.3804, + "theoretical_loss": 3.5731715656313128, + "tokens_seen": 1369309184 + }, + { + "epoch": 0.41, + "learning_rate": 0.000472826479438315, + "loss": 1.3485, + "theoretical_loss": 3.5731563451840227, + "tokens_seen": 1369374720 + }, + { + "epoch": 0.41, + "learning_rate": 0.0004728104312938816, + "loss": 1.371, + "theoretical_loss": 3.5731411256690877, + "tokens_seen": 1369440256 + }, + { + "epoch": 0.42, + "learning_rate": 0.0004727943831494484, + "loss": 1.3242, + "theoretical_loss": 3.573125907086407, + "tokens_seen": 1369505792 + }, + { + "epoch": 0.42, + "learning_rate": 0.000472778335005015, + "loss": 1.315, + "theoretical_loss": 3.5731106894358793, + "tokens_seen": 1369571328 + }, + { + "epoch": 0.42, + "learning_rate": 0.0004727622868605818, + "loss": 1.4876, + "theoretical_loss": 3.5730954727174016, + "tokens_seen": 1369636864 + }, + { + "epoch": 0.42, + "objective/train/docs_used": 500422, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.4433863162994385, + "objective/train/theoretical_loss": 3.573080256930874, + "objective/train/tokens_used": 1390162400, + "theoretical_loss": 3.573080256930874, + "tokens_seen": 1369702400 + }, + { + "epoch": 0.42, + "learning_rate": 0.00047274623871614843, + "loss": 1.4362, + "theoretical_loss": 3.573080256930874, + "tokens_seen": 1369702400 + }, + { + "epoch": 0.42, + "learning_rate": 0.0004727301905717152, + "loss": 1.4163, + "theoretical_loss": 3.573065042076193, + "tokens_seen": 1369767936 + }, + { + "epoch": 0.42, + "learning_rate": 0.00047271414242728183, + "loss": 1.4836, + "theoretical_loss": 3.5730498281532577, + "tokens_seen": 1369833472 + }, + { + "epoch": 0.42, + "learning_rate": 0.0004726980942828486, + "loss": 1.3275, + "theoretical_loss": 3.5730346151619665, + "tokens_seen": 1369899008 + }, + { + "epoch": 0.42, + "learning_rate": 0.00047268204613841523, + "loss": 1.3549, + "theoretical_loss": 3.5730194031022178, + "tokens_seen": 1369964544 + }, + { + "epoch": 0.42, + "learning_rate": 0.000472665997993982, + "loss": 1.4364, + "theoretical_loss": 3.57300419197391, + "tokens_seen": 1370030080 + }, + { + "epoch": 0.42, + "learning_rate": 0.00047264994984954863, + "loss": 1.3545, + "theoretical_loss": 3.572988981776942, + "tokens_seen": 1370095616 + }, + { + "epoch": 0.42, + "learning_rate": 0.0004726339017051154, + "loss": 1.3466, + "theoretical_loss": 3.5729737725112116, + "tokens_seen": 1370161152 + }, + { + "epoch": 0.42, + "learning_rate": 0.00047261785356068203, + "loss": 1.3263, + "theoretical_loss": 3.572958564176617, + "tokens_seen": 1370226688 + }, + { + "epoch": 0.42, + "learning_rate": 0.0004726018054162488, + "loss": 1.3818, + "theoretical_loss": 3.5729433567730573, + "tokens_seen": 1370292224 + }, + { + "epoch": 0.42, + "learning_rate": 0.00047258575727181543, + "loss": 1.4271, + "theoretical_loss": 3.5729281503004304, + "tokens_seen": 1370357760 + }, + { + "epoch": 0.42, + "learning_rate": 0.0004725697091273822, + "loss": 1.3888, + "theoretical_loss": 3.5729129447586354, + "tokens_seen": 1370423296 + }, + { + "epoch": 0.42, + "learning_rate": 0.00047255366098294884, + "loss": 1.3411, + "theoretical_loss": 3.5728977401475706, + "tokens_seen": 1370488832 + }, + { + "epoch": 0.42, + "learning_rate": 0.0004725376128385156, + "loss": 1.3203, + "theoretical_loss": 3.572882536467134, + "tokens_seen": 1370554368 + }, + { + "epoch": 0.42, + "learning_rate": 0.00047252156469408224, + "loss": 1.3737, + "theoretical_loss": 3.572867333717225, + "tokens_seen": 1370619904 + }, + { + "epoch": 0.42, + "learning_rate": 0.000472505516549649, + "loss": 1.4129, + "theoretical_loss": 3.5728521318977418, + "tokens_seen": 1370685440 + }, + { + "epoch": 0.42, + "learning_rate": 0.00047248946840521564, + "loss": 1.3695, + "theoretical_loss": 3.572836931008583, + "tokens_seen": 1370750976 + }, + { + "epoch": 0.42, + "learning_rate": 0.0004724734202607824, + "loss": 1.387, + "theoretical_loss": 3.572821731049647, + "tokens_seen": 1370816512 + }, + { + "epoch": 0.42, + "learning_rate": 0.00047245737211634904, + "loss": 1.3933, + "theoretical_loss": 3.5728065320208326, + "tokens_seen": 1370882048 + }, + { + "epoch": 0.42, + "learning_rate": 0.0004724413239719158, + "loss": 1.4359, + "theoretical_loss": 3.572791333922038, + "tokens_seen": 1370947584 + }, + { + "epoch": 0.42, + "learning_rate": 0.00047242527582748244, + "loss": 1.3241, + "theoretical_loss": 3.5727761367531627, + "tokens_seen": 1371013120 + }, + { + "epoch": 0.42, + "learning_rate": 0.0004724092276830492, + "loss": 1.45, + "theoretical_loss": 3.5727609405141045, + "tokens_seen": 1371078656 + }, + { + "epoch": 0.42, + "learning_rate": 0.00047239317953861584, + "loss": 1.5844, + "theoretical_loss": 3.572745745204763, + "tokens_seen": 1371144192 + }, + { + "epoch": 0.42, + "learning_rate": 0.0004723771313941826, + "loss": 1.486, + "theoretical_loss": 3.5727305508250358, + "tokens_seen": 1371209728 + }, + { + "epoch": 0.42, + "learning_rate": 0.00047236108324974924, + "loss": 1.3212, + "theoretical_loss": 3.572715357374822, + "tokens_seen": 1371275264 + }, + { + "epoch": 0.42, + "learning_rate": 0.000472345035105316, + "loss": 1.3715, + "theoretical_loss": 3.5727001648540204, + "tokens_seen": 1371340800 + }, + { + "epoch": 0.42, + "learning_rate": 0.00047232898696088264, + "loss": 1.3749, + "theoretical_loss": 3.57268497326253, + "tokens_seen": 1371406336 + }, + { + "epoch": 0.42, + "learning_rate": 0.0004723129388164494, + "loss": 1.4112, + "theoretical_loss": 3.5726697826002494, + "tokens_seen": 1371471872 + }, + { + "epoch": 0.42, + "learning_rate": 0.00047229689067201605, + "loss": 1.2768, + "theoretical_loss": 3.572654592867077, + "tokens_seen": 1371537408 + }, + { + "epoch": 0.42, + "learning_rate": 0.0004722808425275828, + "loss": 1.3891, + "theoretical_loss": 3.572639404062912, + "tokens_seen": 1371602944 + }, + { + "epoch": 0.42, + "learning_rate": 0.00047226479438314945, + "loss": 1.3141, + "theoretical_loss": 3.572624216187653, + "tokens_seen": 1371668480 + }, + { + "epoch": 0.42, + "learning_rate": 0.0004722487462387162, + "loss": 1.4285, + "theoretical_loss": 3.5726090292411996, + "tokens_seen": 1371734016 + }, + { + "epoch": 0.42, + "learning_rate": 0.00047223269809428285, + "loss": 1.2787, + "theoretical_loss": 3.5725938432234488, + "tokens_seen": 1371799552 + }, + { + "epoch": 0.42, + "learning_rate": 0.0004722166499498496, + "loss": 1.3908, + "theoretical_loss": 3.572578658134301, + "tokens_seen": 1371865088 + }, + { + "epoch": 0.42, + "learning_rate": 0.00047220060180541625, + "loss": 1.2792, + "theoretical_loss": 3.5725634739736547, + "tokens_seen": 1371930624 + }, + { + "epoch": 0.42, + "learning_rate": 0.00047218455366098295, + "loss": 1.3814, + "theoretical_loss": 3.572548290741408, + "tokens_seen": 1371996160 + }, + { + "epoch": 0.42, + "learning_rate": 0.00047216850551654965, + "loss": 1.3126, + "theoretical_loss": 3.5725331084374616, + "tokens_seen": 1372061696 + }, + { + "epoch": 0.42, + "learning_rate": 0.00047215245737211635, + "loss": 1.4099, + "theoretical_loss": 3.5725179270617127, + "tokens_seen": 1372127232 + }, + { + "epoch": 0.42, + "learning_rate": 0.00047213640922768305, + "loss": 1.4842, + "theoretical_loss": 3.572502746614061, + "tokens_seen": 1372192768 + }, + { + "epoch": 0.42, + "learning_rate": 0.00047212036108324975, + "loss": 1.4144, + "theoretical_loss": 3.572487567094405, + "tokens_seen": 1372258304 + }, + { + "epoch": 0.42, + "learning_rate": 0.00047210431293881645, + "loss": 1.3536, + "theoretical_loss": 3.5724723885026437, + "tokens_seen": 1372323840 + }, + { + "epoch": 0.42, + "learning_rate": 0.00047208826479438315, + "loss": 1.3073, + "theoretical_loss": 3.5724572108386767, + "tokens_seen": 1372389376 + }, + { + "epoch": 0.42, + "learning_rate": 0.00047207221664994985, + "loss": 1.2483, + "theoretical_loss": 3.5724420341024024, + "tokens_seen": 1372454912 + }, + { + "epoch": 0.42, + "learning_rate": 0.00047205616850551656, + "loss": 1.3298, + "theoretical_loss": 3.5724268582937198, + "tokens_seen": 1372520448 + }, + { + "epoch": 0.42, + "learning_rate": 0.0004720401203610833, + "loss": 1.323, + "theoretical_loss": 3.572411683412528, + "tokens_seen": 1372585984 + }, + { + "epoch": 0.42, + "learning_rate": 0.00047202407221664996, + "loss": 1.3139, + "theoretical_loss": 3.5723965094587267, + "tokens_seen": 1372651520 + }, + { + "epoch": 0.42, + "learning_rate": 0.0004720080240722167, + "loss": 1.4365, + "theoretical_loss": 3.5723813364322137, + "tokens_seen": 1372717056 + }, + { + "epoch": 0.42, + "learning_rate": 0.00047199197592778336, + "loss": 1.4152, + "theoretical_loss": 3.5723661643328892, + "tokens_seen": 1372782592 + }, + { + "epoch": 0.42, + "learning_rate": 0.0004719759277833501, + "loss": 1.3816, + "theoretical_loss": 3.5723509931606516, + "tokens_seen": 1372848128 + }, + { + "epoch": 0.42, + "learning_rate": 0.00047195987963891676, + "loss": 1.3571, + "theoretical_loss": 3.5723358229154005, + "tokens_seen": 1372913664 + }, + { + "epoch": 0.42, + "objective/train/docs_used": 501474, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.247755527496338, + "objective/train/theoretical_loss": 3.5723206535970347, + "objective/train/tokens_used": 1393439200, + "theoretical_loss": 3.5723206535970347, + "tokens_seen": 1372979200 + }, + { + "epoch": 0.42, + "learning_rate": 0.0004719438314944835, + "loss": 1.3277, + "theoretical_loss": 3.5723206535970347, + "tokens_seen": 1372979200 + }, + { + "epoch": 0.42, + "learning_rate": 0.00047192778335005016, + "loss": 1.4192, + "theoretical_loss": 3.5723054852054528, + "tokens_seen": 1373044736 + }, + { + "epoch": 0.42, + "learning_rate": 0.0004719117352056169, + "loss": 1.385, + "theoretical_loss": 3.572290317740555, + "tokens_seen": 1373110272 + }, + { + "epoch": 0.42, + "learning_rate": 0.00047189568706118356, + "loss": 1.3413, + "theoretical_loss": 3.57227515120224, + "tokens_seen": 1373175808 + }, + { + "epoch": 0.42, + "learning_rate": 0.0004718796389167503, + "loss": 1.5768, + "theoretical_loss": 3.5722599855904065, + "tokens_seen": 1373241344 + }, + { + "epoch": 0.42, + "learning_rate": 0.00047186359077231696, + "loss": 1.5281, + "theoretical_loss": 3.5722448209049547, + "tokens_seen": 1373306880 + }, + { + "epoch": 0.42, + "learning_rate": 0.0004718475426278837, + "loss": 1.4015, + "theoretical_loss": 3.5722296571457828, + "tokens_seen": 1373372416 + }, + { + "epoch": 0.42, + "learning_rate": 0.00047183149448345036, + "loss": 1.3671, + "theoretical_loss": 3.5722144943127905, + "tokens_seen": 1373437952 + }, + { + "epoch": 0.42, + "learning_rate": 0.0004718154463390171, + "loss": 1.4123, + "theoretical_loss": 3.5721993324058774, + "tokens_seen": 1373503488 + }, + { + "epoch": 0.42, + "learning_rate": 0.00047179939819458377, + "loss": 1.431, + "theoretical_loss": 3.572184171424942, + "tokens_seen": 1373569024 + }, + { + "epoch": 0.42, + "learning_rate": 0.0004717833500501505, + "loss": 1.3348, + "theoretical_loss": 3.572169011369884, + "tokens_seen": 1373634560 + }, + { + "epoch": 0.42, + "learning_rate": 0.00047176730190571717, + "loss": 1.5236, + "theoretical_loss": 3.5721538522406027, + "tokens_seen": 1373700096 + }, + { + "epoch": 0.42, + "learning_rate": 0.0004717512537612839, + "loss": 1.4886, + "theoretical_loss": 3.572138694036997, + "tokens_seen": 1373765632 + }, + { + "epoch": 0.42, + "learning_rate": 0.00047173520561685057, + "loss": 1.4002, + "theoretical_loss": 3.572123536758967, + "tokens_seen": 1373831168 + }, + { + "epoch": 0.42, + "learning_rate": 0.0004717191574724173, + "loss": 1.3839, + "theoretical_loss": 3.5721083804064113, + "tokens_seen": 1373896704 + }, + { + "epoch": 0.42, + "learning_rate": 0.00047170310932798397, + "loss": 1.4089, + "theoretical_loss": 3.5720932249792297, + "tokens_seen": 1373962240 + }, + { + "epoch": 0.42, + "learning_rate": 0.0004716870611835507, + "loss": 1.3461, + "theoretical_loss": 3.572078070477321, + "tokens_seen": 1374027776 + }, + { + "epoch": 0.42, + "learning_rate": 0.00047167101303911737, + "loss": 1.3903, + "theoretical_loss": 3.5720629169005855, + "tokens_seen": 1374093312 + }, + { + "epoch": 0.42, + "learning_rate": 0.0004716549648946841, + "loss": 1.348, + "theoretical_loss": 3.572047764248922, + "tokens_seen": 1374158848 + }, + { + "epoch": 0.42, + "learning_rate": 0.00047163891675025077, + "loss": 1.3079, + "theoretical_loss": 3.5720326125222295, + "tokens_seen": 1374224384 + }, + { + "epoch": 0.42, + "learning_rate": 0.0004716228686058175, + "loss": 1.3359, + "theoretical_loss": 3.5720174617204083, + "tokens_seen": 1374289920 + }, + { + "epoch": 0.42, + "learning_rate": 0.0004716068204613842, + "loss": 1.3449, + "theoretical_loss": 3.5720023118433577, + "tokens_seen": 1374355456 + }, + { + "epoch": 0.42, + "learning_rate": 0.00047159077231695093, + "loss": 1.3109, + "theoretical_loss": 3.5719871628909763, + "tokens_seen": 1374420992 + }, + { + "epoch": 0.42, + "learning_rate": 0.0004715747241725176, + "loss": 1.3809, + "theoretical_loss": 3.5719720148631646, + "tokens_seen": 1374486528 + }, + { + "epoch": 0.42, + "learning_rate": 0.0004715586760280843, + "loss": 1.3577, + "theoretical_loss": 3.5719568677598215, + "tokens_seen": 1374552064 + }, + { + "epoch": 0.42, + "learning_rate": 0.000471542627883651, + "loss": 1.3811, + "theoretical_loss": 3.5719417215808473, + "tokens_seen": 1374617600 + }, + { + "epoch": 0.42, + "learning_rate": 0.0004715265797392177, + "loss": 1.2965, + "theoretical_loss": 3.57192657632614, + "tokens_seen": 1374683136 + }, + { + "epoch": 0.42, + "learning_rate": 0.0004715105315947844, + "loss": 1.2526, + "theoretical_loss": 3.571911431995601, + "tokens_seen": 1374748672 + }, + { + "epoch": 0.42, + "learning_rate": 0.0004714944834503511, + "loss": 1.3982, + "theoretical_loss": 3.5718962885891283, + "tokens_seen": 1374814208 + }, + { + "epoch": 0.42, + "learning_rate": 0.0004714784353059178, + "loss": 1.2904, + "theoretical_loss": 3.571881146106622, + "tokens_seen": 1374879744 + }, + { + "epoch": 0.42, + "learning_rate": 0.0004714623871614845, + "loss": 1.4106, + "theoretical_loss": 3.5718660045479824, + "tokens_seen": 1374945280 + }, + { + "epoch": 0.42, + "learning_rate": 0.0004714463390170512, + "loss": 1.3019, + "theoretical_loss": 3.5718508639131077, + "tokens_seen": 1375010816 + }, + { + "epoch": 0.42, + "learning_rate": 0.0004714302908726179, + "loss": 1.3557, + "theoretical_loss": 3.5718357242018985, + "tokens_seen": 1375076352 + }, + { + "epoch": 0.42, + "learning_rate": 0.0004714142427281845, + "loss": 1.3588, + "theoretical_loss": 3.5718205854142546, + "tokens_seen": 1375141888 + }, + { + "epoch": 0.42, + "learning_rate": 0.0004713981945837513, + "loss": 1.3303, + "theoretical_loss": 3.571805447550075, + "tokens_seen": 1375207424 + }, + { + "epoch": 0.42, + "learning_rate": 0.00047138214643931793, + "loss": 1.5074, + "theoretical_loss": 3.57179031060926, + "tokens_seen": 1375272960 + }, + { + "epoch": 0.42, + "learning_rate": 0.0004713660982948847, + "loss": 1.3337, + "theoretical_loss": 3.5717751745917083, + "tokens_seen": 1375338496 + }, + { + "epoch": 0.42, + "learning_rate": 0.00047135005015045133, + "loss": 1.3504, + "theoretical_loss": 3.571760039497321, + "tokens_seen": 1375404032 + }, + { + "epoch": 0.42, + "learning_rate": 0.0004713340020060181, + "loss": 1.4456, + "theoretical_loss": 3.571744905325996, + "tokens_seen": 1375469568 + }, + { + "epoch": 0.42, + "learning_rate": 0.00047131795386158473, + "loss": 1.3626, + "theoretical_loss": 3.571729772077635, + "tokens_seen": 1375535104 + }, + { + "epoch": 0.42, + "learning_rate": 0.0004713019057171515, + "loss": 1.3689, + "theoretical_loss": 3.571714639752136, + "tokens_seen": 1375600640 + }, + { + "epoch": 0.42, + "learning_rate": 0.00047128585757271813, + "loss": 1.534, + "theoretical_loss": 3.5716995083494005, + "tokens_seen": 1375666176 + }, + { + "epoch": 0.42, + "learning_rate": 0.0004712698094282849, + "loss": 1.3518, + "theoretical_loss": 3.5716843778693264, + "tokens_seen": 1375731712 + }, + { + "epoch": 0.42, + "learning_rate": 0.00047125376128385153, + "loss": 1.3154, + "theoretical_loss": 3.571669248311815, + "tokens_seen": 1375797248 + }, + { + "epoch": 0.42, + "learning_rate": 0.0004712377131394183, + "loss": 1.4231, + "theoretical_loss": 3.571654119676765, + "tokens_seen": 1375862784 + }, + { + "epoch": 0.42, + "learning_rate": 0.00047122166499498493, + "loss": 1.5591, + "theoretical_loss": 3.571638991964077, + "tokens_seen": 1375928320 + }, + { + "epoch": 0.42, + "learning_rate": 0.0004712056168505517, + "loss": 1.4238, + "theoretical_loss": 3.571623865173651, + "tokens_seen": 1375993856 + }, + { + "epoch": 0.42, + "learning_rate": 0.00047118956870611834, + "loss": 1.4097, + "theoretical_loss": 3.571608739305386, + "tokens_seen": 1376059392 + }, + { + "epoch": 0.42, + "learning_rate": 0.0004711735205616851, + "loss": 1.3689, + "theoretical_loss": 3.5715936143591827, + "tokens_seen": 1376124928 + }, + { + "epoch": 0.42, + "learning_rate": 0.00047115747241725174, + "loss": 1.3244, + "theoretical_loss": 3.5715784903349403, + "tokens_seen": 1376190464 + }, + { + "epoch": 0.42, + "objective/train/docs_used": 502650, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.2397085428237915, + "objective/train/theoretical_loss": 3.5715633672325593, + "objective/train/tokens_used": 1396716000, + "theoretical_loss": 3.5715633672325593, + "tokens_seen": 1376256000 + }, + { + "epoch": 0.42, + "learning_rate": 0.0004711414242728185, + "loss": 1.304, + "theoretical_loss": 3.5715633672325593, + "tokens_seen": 1376256000 + }, + { + "epoch": 0.42, + "learning_rate": 0.00047112537612838514, + "loss": 1.2649, + "theoretical_loss": 3.5715482450519387, + "tokens_seen": 1376321536 + }, + { + "epoch": 0.42, + "learning_rate": 0.0004711093279839519, + "loss": 1.4224, + "theoretical_loss": 3.5715331237929795, + "tokens_seen": 1376387072 + }, + { + "epoch": 0.42, + "learning_rate": 0.00047109327983951854, + "loss": 1.3374, + "theoretical_loss": 3.571518003455581, + "tokens_seen": 1376452608 + }, + { + "epoch": 0.42, + "learning_rate": 0.0004710772316950853, + "loss": 1.2527, + "theoretical_loss": 3.5715028840396434, + "tokens_seen": 1376518144 + }, + { + "epoch": 0.42, + "learning_rate": 0.00047106118355065194, + "loss": 1.3641, + "theoretical_loss": 3.5714877655450668, + "tokens_seen": 1376583680 + }, + { + "epoch": 0.42, + "learning_rate": 0.0004710451354062187, + "loss": 1.376, + "theoretical_loss": 3.571472647971751, + "tokens_seen": 1376649216 + }, + { + "epoch": 0.42, + "learning_rate": 0.00047102908726178534, + "loss": 1.3367, + "theoretical_loss": 3.571457531319595, + "tokens_seen": 1376714752 + }, + { + "epoch": 0.42, + "learning_rate": 0.0004710130391173521, + "loss": 1.4207, + "theoretical_loss": 3.571442415588501, + "tokens_seen": 1376780288 + }, + { + "epoch": 0.42, + "learning_rate": 0.00047099699097291874, + "loss": 1.4799, + "theoretical_loss": 3.5714273007783675, + "tokens_seen": 1376845824 + }, + { + "epoch": 0.42, + "learning_rate": 0.0004709809428284855, + "loss": 1.3521, + "theoretical_loss": 3.5714121868890953, + "tokens_seen": 1376911360 + }, + { + "epoch": 0.42, + "learning_rate": 0.00047096489468405214, + "loss": 1.3555, + "theoretical_loss": 3.571397073920584, + "tokens_seen": 1376976896 + }, + { + "epoch": 0.42, + "learning_rate": 0.0004709488465396189, + "loss": 1.3716, + "theoretical_loss": 3.5713819618727336, + "tokens_seen": 1377042432 + }, + { + "epoch": 0.42, + "learning_rate": 0.00047093279839518555, + "loss": 1.3221, + "theoretical_loss": 3.571366850745444, + "tokens_seen": 1377107968 + }, + { + "epoch": 0.42, + "learning_rate": 0.0004709167502507523, + "loss": 1.3033, + "theoretical_loss": 3.5713517405386166, + "tokens_seen": 1377173504 + }, + { + "epoch": 0.42, + "learning_rate": 0.00047090070210631895, + "loss": 1.2874, + "theoretical_loss": 3.57133663125215, + "tokens_seen": 1377239040 + }, + { + "epoch": 0.42, + "learning_rate": 0.0004708846539618857, + "loss": 1.364, + "theoretical_loss": 3.5713215228859454, + "tokens_seen": 1377304576 + }, + { + "epoch": 0.42, + "learning_rate": 0.0004708686058174524, + "loss": 1.4049, + "theoretical_loss": 3.571306415439903, + "tokens_seen": 1377370112 + }, + { + "epoch": 0.42, + "learning_rate": 0.0004708525576730191, + "loss": 1.3833, + "theoretical_loss": 3.5712913089139215, + "tokens_seen": 1377435648 + }, + { + "epoch": 0.42, + "learning_rate": 0.0004708365095285858, + "loss": 1.3307, + "theoretical_loss": 3.5712762033079026, + "tokens_seen": 1377501184 + }, + { + "epoch": 0.42, + "learning_rate": 0.0004708204613841525, + "loss": 1.5169, + "theoretical_loss": 3.571261098621746, + "tokens_seen": 1377566720 + }, + { + "epoch": 0.42, + "learning_rate": 0.0004708044132397192, + "loss": 1.308, + "theoretical_loss": 3.571245994855352, + "tokens_seen": 1377632256 + }, + { + "epoch": 0.42, + "learning_rate": 0.00047078836509528585, + "loss": 1.3885, + "theoretical_loss": 3.5712308920086215, + "tokens_seen": 1377697792 + }, + { + "epoch": 0.42, + "learning_rate": 0.0004707723169508526, + "loss": 1.2839, + "theoretical_loss": 3.571215790081453, + "tokens_seen": 1377763328 + }, + { + "epoch": 0.42, + "learning_rate": 0.00047075626880641925, + "loss": 1.3325, + "theoretical_loss": 3.5712006890737484, + "tokens_seen": 1377828864 + }, + { + "epoch": 0.42, + "learning_rate": 0.000470740220661986, + "loss": 1.1846, + "theoretical_loss": 3.571185588985408, + "tokens_seen": 1377894400 + }, + { + "epoch": 0.42, + "learning_rate": 0.00047072417251755265, + "loss": 1.3461, + "theoretical_loss": 3.571170489816331, + "tokens_seen": 1377959936 + }, + { + "epoch": 0.42, + "learning_rate": 0.0004707081243731194, + "loss": 1.3016, + "theoretical_loss": 3.571155391566418, + "tokens_seen": 1378025472 + }, + { + "epoch": 0.42, + "learning_rate": 0.00047069207622868605, + "loss": 1.3344, + "theoretical_loss": 3.57114029423557, + "tokens_seen": 1378091008 + }, + { + "epoch": 0.42, + "learning_rate": 0.0004706760280842528, + "loss": 1.4662, + "theoretical_loss": 3.5711251978236866, + "tokens_seen": 1378156544 + }, + { + "epoch": 0.42, + "learning_rate": 0.00047065997993981946, + "loss": 1.3731, + "theoretical_loss": 3.5711101023306693, + "tokens_seen": 1378222080 + }, + { + "epoch": 0.42, + "learning_rate": 0.0004706439317953862, + "loss": 1.4164, + "theoretical_loss": 3.5710950077564165, + "tokens_seen": 1378287616 + }, + { + "epoch": 0.42, + "learning_rate": 0.00047062788365095286, + "loss": 1.3331, + "theoretical_loss": 3.571079914100831, + "tokens_seen": 1378353152 + }, + { + "epoch": 0.42, + "learning_rate": 0.0004706118355065196, + "loss": 1.4027, + "theoretical_loss": 3.5710648213638114, + "tokens_seen": 1378418688 + }, + { + "epoch": 0.42, + "learning_rate": 0.00047059578736208626, + "loss": 1.4571, + "theoretical_loss": 3.5710497295452592, + "tokens_seen": 1378484224 + }, + { + "epoch": 0.42, + "learning_rate": 0.000470579739217653, + "loss": 1.3816, + "theoretical_loss": 3.5710346386450738, + "tokens_seen": 1378549760 + }, + { + "epoch": 0.42, + "learning_rate": 0.00047056369107321966, + "loss": 1.3851, + "theoretical_loss": 3.571019548663157, + "tokens_seen": 1378615296 + }, + { + "epoch": 0.42, + "learning_rate": 0.0004705476429287864, + "loss": 1.3485, + "theoretical_loss": 3.5710044595994077, + "tokens_seen": 1378680832 + }, + { + "epoch": 0.42, + "learning_rate": 0.00047053159478435306, + "loss": 1.4061, + "theoretical_loss": 3.570989371453728, + "tokens_seen": 1378746368 + }, + { + "epoch": 0.42, + "learning_rate": 0.0004705155466399198, + "loss": 1.435, + "theoretical_loss": 3.5709742842260175, + "tokens_seen": 1378811904 + }, + { + "epoch": 0.42, + "learning_rate": 0.00047049949849548646, + "loss": 1.3876, + "theoretical_loss": 3.570959197916176, + "tokens_seen": 1378877440 + }, + { + "epoch": 0.42, + "learning_rate": 0.0004704834503510532, + "loss": 1.403, + "theoretical_loss": 3.570944112524106, + "tokens_seen": 1378942976 + }, + { + "epoch": 0.42, + "learning_rate": 0.00047046740220661986, + "loss": 1.4037, + "theoretical_loss": 3.5709290280497066, + "tokens_seen": 1379008512 + }, + { + "epoch": 0.42, + "learning_rate": 0.0004704513540621866, + "loss": 1.2435, + "theoretical_loss": 3.5709139444928786, + "tokens_seen": 1379074048 + }, + { + "epoch": 0.42, + "learning_rate": 0.00047043530591775327, + "loss": 1.4715, + "theoretical_loss": 3.5708988618535225, + "tokens_seen": 1379139584 + }, + { + "epoch": 0.42, + "learning_rate": 0.00047041925777332, + "loss": 1.4897, + "theoretical_loss": 3.5708837801315396, + "tokens_seen": 1379205120 + }, + { + "epoch": 0.42, + "learning_rate": 0.00047040320962888667, + "loss": 1.3484, + "theoretical_loss": 3.57086869932683, + "tokens_seen": 1379270656 + }, + { + "epoch": 0.42, + "learning_rate": 0.0004703871614844534, + "loss": 1.4131, + "theoretical_loss": 3.570853619439294, + "tokens_seen": 1379336192 + }, + { + "epoch": 0.42, + "learning_rate": 0.00047037111334002007, + "loss": 1.1737, + "theoretical_loss": 3.5708385404688325, + "tokens_seen": 1379401728 + }, + { + "epoch": 0.42, + "learning_rate": 0.0004703550651955868, + "loss": 1.4149, + "theoretical_loss": 3.5708234624153468, + "tokens_seen": 1379467264 + }, + { + "epoch": 0.42, + "objective/train/docs_used": 503826, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.4086315631866455, + "objective/train/theoretical_loss": 3.5708083852787365, + "objective/train/tokens_used": 1399992800, + "theoretical_loss": 3.5708083852787365, + "tokens_seen": 1379532800 + }, + { + "epoch": 0.42, + "learning_rate": 0.00047033901705115347, + "loss": 1.3975, + "theoretical_loss": 3.5708083852787365, + "tokens_seen": 1379532800 + }, + { + "epoch": 0.42, + "learning_rate": 0.0004703229689067202, + "loss": 1.323, + "theoretical_loss": 3.570793309058903, + "tokens_seen": 1379598336 + }, + { + "epoch": 0.42, + "learning_rate": 0.00047030692076228687, + "loss": 1.3774, + "theoretical_loss": 3.570778233755747, + "tokens_seen": 1379663872 + }, + { + "epoch": 0.42, + "learning_rate": 0.0004702908726178536, + "loss": 1.4726, + "theoretical_loss": 3.5707631593691693, + "tokens_seen": 1379729408 + }, + { + "epoch": 0.42, + "learning_rate": 0.00047027482447342027, + "loss": 1.4175, + "theoretical_loss": 3.57074808589907, + "tokens_seen": 1379794944 + }, + { + "epoch": 0.42, + "learning_rate": 0.000470258776328987, + "loss": 1.4084, + "theoretical_loss": 3.5707330133453503, + "tokens_seen": 1379860480 + }, + { + "epoch": 0.42, + "learning_rate": 0.00047024272818455367, + "loss": 1.2823, + "theoretical_loss": 3.570717941707911, + "tokens_seen": 1379926016 + }, + { + "epoch": 0.42, + "learning_rate": 0.00047022668004012043, + "loss": 1.4118, + "theoretical_loss": 3.570702870986653, + "tokens_seen": 1379991552 + }, + { + "epoch": 0.42, + "learning_rate": 0.0004702106318956871, + "loss": 1.4318, + "theoretical_loss": 3.5706878011814767, + "tokens_seen": 1380057088 + }, + { + "epoch": 0.42, + "learning_rate": 0.00047019458375125383, + "loss": 1.424, + "theoretical_loss": 3.570672732292283, + "tokens_seen": 1380122624 + }, + { + "epoch": 0.42, + "learning_rate": 0.0004701785356068205, + "loss": 1.3615, + "theoretical_loss": 3.570657664318973, + "tokens_seen": 1380188160 + }, + { + "epoch": 0.42, + "learning_rate": 0.0004701624874623872, + "loss": 1.4135, + "theoretical_loss": 3.5706425972614477, + "tokens_seen": 1380253696 + }, + { + "epoch": 0.42, + "learning_rate": 0.0004701464393179539, + "loss": 1.3615, + "theoretical_loss": 3.570627531119607, + "tokens_seen": 1380319232 + }, + { + "epoch": 0.42, + "learning_rate": 0.0004701303911735206, + "loss": 1.3334, + "theoretical_loss": 3.570612465893353, + "tokens_seen": 1380384768 + }, + { + "epoch": 0.42, + "learning_rate": 0.0004701143430290873, + "loss": 1.4433, + "theoretical_loss": 3.570597401582586, + "tokens_seen": 1380450304 + }, + { + "epoch": 0.42, + "learning_rate": 0.000470098294884654, + "loss": 1.3492, + "theoretical_loss": 3.570582338187207, + "tokens_seen": 1380515840 + }, + { + "epoch": 0.42, + "learning_rate": 0.0004700822467402207, + "loss": 1.3621, + "theoretical_loss": 3.5705672757071167, + "tokens_seen": 1380581376 + }, + { + "epoch": 0.42, + "learning_rate": 0.0004700661985957874, + "loss": 1.3832, + "theoretical_loss": 3.570552214142216, + "tokens_seen": 1380646912 + }, + { + "epoch": 0.42, + "learning_rate": 0.0004700501504513541, + "loss": 1.4579, + "theoretical_loss": 3.570537153492406, + "tokens_seen": 1380712448 + }, + { + "epoch": 0.42, + "learning_rate": 0.0004700341023069208, + "loss": 1.3277, + "theoretical_loss": 3.5705220937575888, + "tokens_seen": 1380777984 + }, + { + "epoch": 0.42, + "learning_rate": 0.00047001805416248743, + "loss": 1.3995, + "theoretical_loss": 3.5705070349376635, + "tokens_seen": 1380843520 + }, + { + "epoch": 0.42, + "learning_rate": 0.0004700020060180542, + "loss": 1.2352, + "theoretical_loss": 3.570491977032532, + "tokens_seen": 1380909056 + }, + { + "epoch": 0.42, + "learning_rate": 0.00046998595787362083, + "loss": 1.3653, + "theoretical_loss": 3.570476920042095, + "tokens_seen": 1380974592 + }, + { + "epoch": 0.42, + "learning_rate": 0.0004699699097291876, + "loss": 1.3368, + "theoretical_loss": 3.5704618639662544, + "tokens_seen": 1381040128 + }, + { + "epoch": 0.42, + "learning_rate": 0.00046995386158475423, + "loss": 1.3442, + "theoretical_loss": 3.5704468088049106, + "tokens_seen": 1381105664 + }, + { + "epoch": 0.42, + "learning_rate": 0.000469937813440321, + "loss": 1.4125, + "theoretical_loss": 3.5704317545579647, + "tokens_seen": 1381171200 + }, + { + "epoch": 0.42, + "learning_rate": 0.00046992176529588763, + "loss": 1.4279, + "theoretical_loss": 3.570416701225317, + "tokens_seen": 1381236736 + }, + { + "epoch": 0.42, + "learning_rate": 0.0004699057171514544, + "loss": 1.35, + "theoretical_loss": 3.57040164880687, + "tokens_seen": 1381302272 + }, + { + "epoch": 0.42, + "learning_rate": 0.00046988966900702103, + "loss": 1.3741, + "theoretical_loss": 3.570386597302524, + "tokens_seen": 1381367808 + }, + { + "epoch": 0.42, + "learning_rate": 0.0004698736208625878, + "loss": 1.3103, + "theoretical_loss": 3.57037154671218, + "tokens_seen": 1381433344 + }, + { + "epoch": 0.42, + "learning_rate": 0.00046985757271815443, + "loss": 1.4715, + "theoretical_loss": 3.57035649703574, + "tokens_seen": 1381498880 + }, + { + "epoch": 0.42, + "learning_rate": 0.0004698415245737212, + "loss": 1.3337, + "theoretical_loss": 3.570341448273105, + "tokens_seen": 1381564416 + }, + { + "epoch": 0.42, + "learning_rate": 0.00046982547642928784, + "loss": 1.3697, + "theoretical_loss": 3.570326400424175, + "tokens_seen": 1381629952 + }, + { + "epoch": 0.42, + "learning_rate": 0.0004698094282848546, + "loss": 1.3863, + "theoretical_loss": 3.570311353488852, + "tokens_seen": 1381695488 + }, + { + "epoch": 0.42, + "learning_rate": 0.00046979338014042124, + "loss": 1.3285, + "theoretical_loss": 3.5702963074670375, + "tokens_seen": 1381761024 + }, + { + "epoch": 0.42, + "learning_rate": 0.000469777331995988, + "loss": 1.3964, + "theoretical_loss": 3.570281262358632, + "tokens_seen": 1381826560 + }, + { + "epoch": 0.42, + "learning_rate": 0.00046976128385155464, + "loss": 1.3893, + "theoretical_loss": 3.5702662181635376, + "tokens_seen": 1381892096 + }, + { + "epoch": 0.42, + "learning_rate": 0.0004697452357071214, + "loss": 1.2847, + "theoretical_loss": 3.5702511748816548, + "tokens_seen": 1381957632 + }, + { + "epoch": 0.42, + "learning_rate": 0.00046972918756268804, + "loss": 1.4573, + "theoretical_loss": 3.5702361325128846, + "tokens_seen": 1382023168 + }, + { + "epoch": 0.42, + "learning_rate": 0.0004697131394182548, + "loss": 1.4311, + "theoretical_loss": 3.57022109105713, + "tokens_seen": 1382088704 + }, + { + "epoch": 0.42, + "learning_rate": 0.00046969709127382144, + "loss": 1.3287, + "theoretical_loss": 3.5702060505142903, + "tokens_seen": 1382154240 + }, + { + "epoch": 0.42, + "learning_rate": 0.0004696810431293882, + "loss": 1.4016, + "theoretical_loss": 3.570191010884267, + "tokens_seen": 1382219776 + }, + { + "epoch": 0.42, + "learning_rate": 0.00046966499498495495, + "loss": 1.3592, + "theoretical_loss": 3.5701759721669633, + "tokens_seen": 1382285312 + }, + { + "epoch": 0.42, + "learning_rate": 0.0004696489468405216, + "loss": 1.3568, + "theoretical_loss": 3.5701609343622787, + "tokens_seen": 1382350848 + }, + { + "epoch": 0.42, + "learning_rate": 0.00046963289869608835, + "loss": 1.3181, + "theoretical_loss": 3.5701458974701152, + "tokens_seen": 1382416384 + }, + { + "epoch": 0.42, + "learning_rate": 0.000469616850551655, + "loss": 1.3911, + "theoretical_loss": 3.570130861490374, + "tokens_seen": 1382481920 + }, + { + "epoch": 0.42, + "learning_rate": 0.00046960080240722175, + "loss": 1.2779, + "theoretical_loss": 3.570115826422956, + "tokens_seen": 1382547456 + }, + { + "epoch": 0.42, + "learning_rate": 0.0004695847542627884, + "loss": 1.3568, + "theoretical_loss": 3.5701007922677643, + "tokens_seen": 1382612992 + }, + { + "epoch": 0.42, + "learning_rate": 0.00046956870611835515, + "loss": 1.4852, + "theoretical_loss": 3.570085759024699, + "tokens_seen": 1382678528 + }, + { + "epoch": 0.42, + "learning_rate": 0.0004695526579739218, + "loss": 1.3664, + "theoretical_loss": 3.570070726693661, + "tokens_seen": 1382744064 + }, + { + "epoch": 0.42, + "objective/train/docs_used": 505055, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.3206082582473755, + "objective/train/theoretical_loss": 3.570055695274553, + "objective/train/tokens_used": 1403269600, + "theoretical_loss": 3.570055695274553, + "tokens_seen": 1382809600 + }, + { + "epoch": 0.42, + "learning_rate": 0.0004695366098294885, + "loss": 1.3986, + "theoretical_loss": 3.570055695274553, + "tokens_seen": 1382809600 + }, + { + "epoch": 0.42, + "learning_rate": 0.0004695205616850552, + "loss": 1.3267, + "theoretical_loss": 3.570040664767276, + "tokens_seen": 1382875136 + }, + { + "epoch": 0.42, + "learning_rate": 0.0004695045135406219, + "loss": 1.331, + "theoretical_loss": 3.570025635171731, + "tokens_seen": 1382940672 + }, + { + "epoch": 0.42, + "learning_rate": 0.0004694884653961886, + "loss": 1.3896, + "theoretical_loss": 3.5700106064878203, + "tokens_seen": 1383006208 + }, + { + "epoch": 0.42, + "learning_rate": 0.0004694724172517553, + "loss": 1.4442, + "theoretical_loss": 3.569995578715445, + "tokens_seen": 1383071744 + }, + { + "epoch": 0.42, + "learning_rate": 0.000469456369107322, + "loss": 1.3731, + "theoretical_loss": 3.5699805518545062, + "tokens_seen": 1383137280 + }, + { + "epoch": 0.42, + "learning_rate": 0.0004694403209628887, + "loss": 1.3774, + "theoretical_loss": 3.569965525904906, + "tokens_seen": 1383202816 + }, + { + "epoch": 0.42, + "learning_rate": 0.0004694242728184554, + "loss": 1.3469, + "theoretical_loss": 3.569950500866547, + "tokens_seen": 1383268352 + }, + { + "epoch": 0.42, + "learning_rate": 0.0004694082246740221, + "loss": 1.2761, + "theoretical_loss": 3.569935476739328, + "tokens_seen": 1383333888 + }, + { + "epoch": 0.42, + "learning_rate": 0.00046939217652958875, + "loss": 1.463, + "theoretical_loss": 3.569920453523153, + "tokens_seen": 1383399424 + }, + { + "epoch": 0.42, + "learning_rate": 0.0004693761283851555, + "loss": 1.3973, + "theoretical_loss": 3.569905431217923, + "tokens_seen": 1383464960 + }, + { + "epoch": 0.42, + "learning_rate": 0.00046936008024072215, + "loss": 1.3186, + "theoretical_loss": 3.5698904098235387, + "tokens_seen": 1383530496 + }, + { + "epoch": 0.42, + "learning_rate": 0.0004693440320962889, + "loss": 1.3501, + "theoretical_loss": 3.569875389339903, + "tokens_seen": 1383596032 + }, + { + "epoch": 0.42, + "learning_rate": 0.00046932798395185555, + "loss": 1.3371, + "theoretical_loss": 3.569860369766917, + "tokens_seen": 1383661568 + }, + { + "epoch": 0.42, + "learning_rate": 0.0004693119358074223, + "loss": 1.3632, + "theoretical_loss": 3.5698453511044823, + "tokens_seen": 1383727104 + }, + { + "epoch": 0.42, + "learning_rate": 0.00046929588766298896, + "loss": 1.3894, + "theoretical_loss": 3.5698303333525003, + "tokens_seen": 1383792640 + }, + { + "epoch": 0.42, + "learning_rate": 0.0004692798395185557, + "loss": 1.4389, + "theoretical_loss": 3.5698153165108732, + "tokens_seen": 1383858176 + }, + { + "epoch": 0.42, + "learning_rate": 0.00046926379137412236, + "loss": 1.3507, + "theoretical_loss": 3.5698003005795025, + "tokens_seen": 1383923712 + }, + { + "epoch": 0.42, + "learning_rate": 0.0004692477432296891, + "loss": 1.4224, + "theoretical_loss": 3.5697852855582903, + "tokens_seen": 1383989248 + }, + { + "epoch": 0.42, + "learning_rate": 0.00046923169508525576, + "loss": 1.2564, + "theoretical_loss": 3.5697702714471378, + "tokens_seen": 1384054784 + }, + { + "epoch": 0.42, + "learning_rate": 0.0004692156469408225, + "loss": 1.3241, + "theoretical_loss": 3.5697552582459466, + "tokens_seen": 1384120320 + }, + { + "epoch": 0.42, + "learning_rate": 0.00046919959879638916, + "loss": 1.3331, + "theoretical_loss": 3.569740245954619, + "tokens_seen": 1384185856 + }, + { + "epoch": 0.42, + "learning_rate": 0.0004691835506519559, + "loss": 1.347, + "theoretical_loss": 3.5697252345730566, + "tokens_seen": 1384251392 + }, + { + "epoch": 0.42, + "learning_rate": 0.00046916750250752256, + "loss": 1.3449, + "theoretical_loss": 3.5697102241011613, + "tokens_seen": 1384316928 + }, + { + "epoch": 0.42, + "learning_rate": 0.0004691514543630893, + "loss": 1.4813, + "theoretical_loss": 3.569695214538835, + "tokens_seen": 1384382464 + }, + { + "epoch": 0.42, + "learning_rate": 0.00046913540621865596, + "loss": 1.4301, + "theoretical_loss": 3.569680205885979, + "tokens_seen": 1384448000 + }, + { + "epoch": 0.42, + "learning_rate": 0.0004691193580742227, + "loss": 1.4732, + "theoretical_loss": 3.5696651981424954, + "tokens_seen": 1384513536 + }, + { + "epoch": 0.42, + "learning_rate": 0.00046910330992978936, + "loss": 1.3924, + "theoretical_loss": 3.5696501913082863, + "tokens_seen": 1384579072 + }, + { + "epoch": 0.42, + "learning_rate": 0.0004690872617853561, + "loss": 1.3392, + "theoretical_loss": 3.5696351853832535, + "tokens_seen": 1384644608 + }, + { + "epoch": 0.42, + "learning_rate": 0.00046907121364092276, + "loss": 1.4235, + "theoretical_loss": 3.5696201803672984, + "tokens_seen": 1384710144 + }, + { + "epoch": 0.42, + "learning_rate": 0.0004690551654964895, + "loss": 1.3604, + "theoretical_loss": 3.569605176260324, + "tokens_seen": 1384775680 + }, + { + "epoch": 0.42, + "learning_rate": 0.00046903911735205617, + "loss": 1.3747, + "theoretical_loss": 3.569590173062231, + "tokens_seen": 1384841216 + }, + { + "epoch": 0.42, + "learning_rate": 0.0004690230692076229, + "loss": 1.3145, + "theoretical_loss": 3.569575170772922, + "tokens_seen": 1384906752 + }, + { + "epoch": 0.42, + "learning_rate": 0.00046900702106318957, + "loss": 1.4493, + "theoretical_loss": 3.5695601693922985, + "tokens_seen": 1384972288 + }, + { + "epoch": 0.42, + "learning_rate": 0.0004689909729187563, + "loss": 1.3819, + "theoretical_loss": 3.5695451689202633, + "tokens_seen": 1385037824 + }, + { + "epoch": 0.42, + "learning_rate": 0.00046897492477432297, + "loss": 1.4726, + "theoretical_loss": 3.569530169356718, + "tokens_seen": 1385103360 + }, + { + "epoch": 0.42, + "learning_rate": 0.0004689588766298897, + "loss": 1.322, + "theoretical_loss": 3.569515170701564, + "tokens_seen": 1385168896 + }, + { + "epoch": 0.42, + "learning_rate": 0.00046894282848545637, + "loss": 1.3213, + "theoretical_loss": 3.569500172954704, + "tokens_seen": 1385234432 + }, + { + "epoch": 0.42, + "learning_rate": 0.0004689267803410231, + "loss": 1.496, + "theoretical_loss": 3.5694851761160398, + "tokens_seen": 1385299968 + }, + { + "epoch": 0.42, + "learning_rate": 0.00046891073219658977, + "loss": 1.4619, + "theoretical_loss": 3.5694701801854736, + "tokens_seen": 1385365504 + }, + { + "epoch": 0.42, + "learning_rate": 0.0004688946840521565, + "loss": 1.4427, + "theoretical_loss": 3.569455185162907, + "tokens_seen": 1385431040 + }, + { + "epoch": 0.42, + "learning_rate": 0.00046887863590772317, + "loss": 1.3287, + "theoretical_loss": 3.5694401910482423, + "tokens_seen": 1385496576 + }, + { + "epoch": 0.42, + "learning_rate": 0.00046886258776328993, + "loss": 1.3525, + "theoretical_loss": 3.569425197841382, + "tokens_seen": 1385562112 + }, + { + "epoch": 0.42, + "learning_rate": 0.0004688465396188566, + "loss": 1.3752, + "theoretical_loss": 3.5694102055422277, + "tokens_seen": 1385627648 + }, + { + "epoch": 0.42, + "learning_rate": 0.00046883049147442333, + "loss": 1.333, + "theoretical_loss": 3.5693952141506813, + "tokens_seen": 1385693184 + }, + { + "epoch": 0.42, + "learning_rate": 0.00046881444332999, + "loss": 1.305, + "theoretical_loss": 3.569380223666646, + "tokens_seen": 1385758720 + }, + { + "epoch": 0.42, + "learning_rate": 0.00046879839518555673, + "loss": 1.2978, + "theoretical_loss": 3.5693652340900233, + "tokens_seen": 1385824256 + }, + { + "epoch": 0.42, + "learning_rate": 0.0004687823470411234, + "loss": 1.4116, + "theoretical_loss": 3.569350245420715, + "tokens_seen": 1385889792 + }, + { + "epoch": 0.42, + "learning_rate": 0.0004687662988966901, + "loss": 1.4049, + "theoretical_loss": 3.5693352576586235, + "tokens_seen": 1385955328 + }, + { + "epoch": 0.42, + "learning_rate": 0.0004687502507522568, + "loss": 1.3963, + "theoretical_loss": 3.5693202708036513, + "tokens_seen": 1386020864 + }, + { + "epoch": 0.42, + "objective/train/docs_used": 506209, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.5301488637924194, + "objective/train/theoretical_loss": 3.5693052848557003, + "objective/train/tokens_used": 1406546400, + "theoretical_loss": 3.5693052848557003, + "tokens_seen": 1386086400 + }, + { + "epoch": 0.42, + "learning_rate": 0.0004687342026078235, + "loss": 1.4103, + "theoretical_loss": 3.5693052848557003, + "tokens_seen": 1386086400 + }, + { + "epoch": 0.42, + "learning_rate": 0.0004687181544633902, + "loss": 1.3916, + "theoretical_loss": 3.569290299814673, + "tokens_seen": 1386151936 + }, + { + "epoch": 0.42, + "learning_rate": 0.0004687021063189569, + "loss": 1.3486, + "theoretical_loss": 3.5692753156804713, + "tokens_seen": 1386217472 + }, + { + "epoch": 0.42, + "learning_rate": 0.0004686860581745236, + "loss": 1.2859, + "theoretical_loss": 3.569260332452998, + "tokens_seen": 1386283008 + }, + { + "epoch": 0.42, + "learning_rate": 0.0004686700100300903, + "loss": 1.3501, + "theoretical_loss": 3.569245350132155, + "tokens_seen": 1386348544 + }, + { + "epoch": 0.42, + "learning_rate": 0.000468653961885657, + "loss": 1.2138, + "theoretical_loss": 3.569230368717844, + "tokens_seen": 1386414080 + }, + { + "epoch": 0.42, + "learning_rate": 0.0004686379137412237, + "loss": 1.4345, + "theoretical_loss": 3.569215388209968, + "tokens_seen": 1386479616 + }, + { + "epoch": 0.42, + "learning_rate": 0.00046862186559679033, + "loss": 1.3744, + "theoretical_loss": 3.56920040860843, + "tokens_seen": 1386545152 + }, + { + "epoch": 0.42, + "learning_rate": 0.0004686058174523571, + "loss": 1.3933, + "theoretical_loss": 3.569185429913131, + "tokens_seen": 1386610688 + }, + { + "epoch": 0.42, + "learning_rate": 0.00046858976930792373, + "loss": 1.365, + "theoretical_loss": 3.5691704521239735, + "tokens_seen": 1386676224 + }, + { + "epoch": 0.42, + "learning_rate": 0.0004685737211634905, + "loss": 1.3453, + "theoretical_loss": 3.569155475240861, + "tokens_seen": 1386741760 + }, + { + "epoch": 0.42, + "learning_rate": 0.00046855767301905713, + "loss": 1.3361, + "theoretical_loss": 3.5691404992636944, + "tokens_seen": 1386807296 + }, + { + "epoch": 0.42, + "learning_rate": 0.0004685416248746239, + "loss": 1.2924, + "theoretical_loss": 3.569125524192377, + "tokens_seen": 1386872832 + }, + { + "epoch": 0.42, + "learning_rate": 0.00046852557673019053, + "loss": 1.3273, + "theoretical_loss": 3.5691105500268114, + "tokens_seen": 1386938368 + }, + { + "epoch": 0.42, + "learning_rate": 0.0004685095285857573, + "loss": 1.3208, + "theoretical_loss": 3.5690955767669, + "tokens_seen": 1387003904 + }, + { + "epoch": 0.42, + "learning_rate": 0.00046849348044132393, + "loss": 1.396, + "theoretical_loss": 3.569080604412544, + "tokens_seen": 1387069440 + }, + { + "epoch": 0.42, + "learning_rate": 0.0004684774322968907, + "loss": 1.3885, + "theoretical_loss": 3.569065632963647, + "tokens_seen": 1387134976 + }, + { + "epoch": 0.42, + "learning_rate": 0.00046846138415245744, + "loss": 1.4394, + "theoretical_loss": 3.5690506624201115, + "tokens_seen": 1387200512 + }, + { + "epoch": 0.42, + "learning_rate": 0.0004684453360080241, + "loss": 1.5504, + "theoretical_loss": 3.5690356927818394, + "tokens_seen": 1387266048 + }, + { + "epoch": 0.42, + "learning_rate": 0.00046842928786359084, + "loss": 1.3498, + "theoretical_loss": 3.5690207240487335, + "tokens_seen": 1387331584 + }, + { + "epoch": 0.42, + "learning_rate": 0.0004684132397191575, + "loss": 1.3714, + "theoretical_loss": 3.5690057562206965, + "tokens_seen": 1387397120 + }, + { + "epoch": 0.42, + "learning_rate": 0.00046839719157472425, + "loss": 1.4033, + "theoretical_loss": 3.5689907892976303, + "tokens_seen": 1387462656 + }, + { + "epoch": 0.42, + "learning_rate": 0.0004683811434302909, + "loss": 1.2065, + "theoretical_loss": 3.568975823279438, + "tokens_seen": 1387528192 + }, + { + "epoch": 0.42, + "learning_rate": 0.00046836509528585765, + "loss": 1.3894, + "theoretical_loss": 3.5689608581660224, + "tokens_seen": 1387593728 + }, + { + "epoch": 0.42, + "learning_rate": 0.0004683490471414243, + "loss": 1.6175, + "theoretical_loss": 3.5689458939572853, + "tokens_seen": 1387659264 + }, + { + "epoch": 0.42, + "learning_rate": 0.00046833299899699105, + "loss": 1.3889, + "theoretical_loss": 3.5689309306531296, + "tokens_seen": 1387724800 + }, + { + "epoch": 0.42, + "learning_rate": 0.0004683169508525577, + "loss": 1.2643, + "theoretical_loss": 3.5689159682534584, + "tokens_seen": 1387790336 + }, + { + "epoch": 0.42, + "learning_rate": 0.00046830090270812445, + "loss": 1.3817, + "theoretical_loss": 3.568901006758174, + "tokens_seen": 1387855872 + }, + { + "epoch": 0.42, + "learning_rate": 0.0004682848545636911, + "loss": 1.2853, + "theoretical_loss": 3.5688860461671785, + "tokens_seen": 1387921408 + }, + { + "epoch": 0.42, + "learning_rate": 0.00046826880641925785, + "loss": 1.3702, + "theoretical_loss": 3.5688710864803745, + "tokens_seen": 1387986944 + }, + { + "epoch": 0.42, + "learning_rate": 0.0004682527582748245, + "loss": 1.2846, + "theoretical_loss": 3.568856127697666, + "tokens_seen": 1388052480 + }, + { + "epoch": 0.42, + "learning_rate": 0.00046823671013039125, + "loss": 1.2193, + "theoretical_loss": 3.5688411698189544, + "tokens_seen": 1388118016 + }, + { + "epoch": 0.42, + "learning_rate": 0.0004682206619859579, + "loss": 1.4229, + "theoretical_loss": 3.5688262128441433, + "tokens_seen": 1388183552 + }, + { + "epoch": 0.42, + "learning_rate": 0.00046820461384152465, + "loss": 1.3021, + "theoretical_loss": 3.5688112567731345, + "tokens_seen": 1388249088 + }, + { + "epoch": 0.42, + "learning_rate": 0.0004681885656970913, + "loss": 1.3053, + "theoretical_loss": 3.5687963016058313, + "tokens_seen": 1388314624 + }, + { + "epoch": 0.42, + "learning_rate": 0.00046817251755265805, + "loss": 1.4095, + "theoretical_loss": 3.568781347342136, + "tokens_seen": 1388380160 + }, + { + "epoch": 0.42, + "learning_rate": 0.0004681564694082247, + "loss": 1.2601, + "theoretical_loss": 3.568766393981952, + "tokens_seen": 1388445696 + }, + { + "epoch": 0.42, + "learning_rate": 0.00046814042126379146, + "loss": 1.4581, + "theoretical_loss": 3.5687514415251815, + "tokens_seen": 1388511232 + }, + { + "epoch": 0.42, + "learning_rate": 0.0004681243731193581, + "loss": 1.3565, + "theoretical_loss": 3.5687364899717275, + "tokens_seen": 1388576768 + }, + { + "epoch": 0.42, + "learning_rate": 0.0004681083249749248, + "loss": 1.3937, + "theoretical_loss": 3.5687215393214933, + "tokens_seen": 1388642304 + }, + { + "epoch": 0.42, + "learning_rate": 0.0004680922768304915, + "loss": 1.3057, + "theoretical_loss": 3.5687065895743806, + "tokens_seen": 1388707840 + }, + { + "epoch": 0.42, + "learning_rate": 0.0004680762286860582, + "loss": 1.4183, + "theoretical_loss": 3.568691640730293, + "tokens_seen": 1388773376 + }, + { + "epoch": 0.42, + "learning_rate": 0.0004680601805416249, + "loss": 1.4255, + "theoretical_loss": 3.5686766927891336, + "tokens_seen": 1388838912 + }, + { + "epoch": 0.42, + "learning_rate": 0.0004680441323971916, + "loss": 1.3511, + "theoretical_loss": 3.568661745750804, + "tokens_seen": 1388904448 + }, + { + "epoch": 0.42, + "learning_rate": 0.0004680280842527583, + "loss": 1.3738, + "theoretical_loss": 3.5686467996152085, + "tokens_seen": 1388969984 + }, + { + "epoch": 0.42, + "learning_rate": 0.000468012036108325, + "loss": 1.3946, + "theoretical_loss": 3.5686318543822493, + "tokens_seen": 1389035520 + }, + { + "epoch": 0.42, + "learning_rate": 0.00046799598796389165, + "loss": 1.4958, + "theoretical_loss": 3.5686169100518295, + "tokens_seen": 1389101056 + }, + { + "epoch": 0.42, + "learning_rate": 0.0004679799398194584, + "loss": 1.3822, + "theoretical_loss": 3.5686019666238513, + "tokens_seen": 1389166592 + }, + { + "epoch": 0.42, + "learning_rate": 0.00046796389167502505, + "loss": 1.3239, + "theoretical_loss": 3.568587024098219, + "tokens_seen": 1389232128 + }, + { + "epoch": 0.42, + "learning_rate": 0.0004679478435305918, + "loss": 1.4998, + "theoretical_loss": 3.5685720824748346, + "tokens_seen": 1389297664 + }, + { + "epoch": 0.42, + "objective/train/docs_used": 507487, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.342557668685913, + "objective/train/theoretical_loss": 3.568557141753601, + "objective/train/tokens_used": 1409823200, + "theoretical_loss": 3.568557141753601, + "tokens_seen": 1389363200 + }, + { + "epoch": 0.42, + "learning_rate": 0.00046793179538615846, + "loss": 1.3099, + "theoretical_loss": 3.568557141753601, + "tokens_seen": 1389363200 + }, + { + "epoch": 0.42, + "learning_rate": 0.0004679157472417252, + "loss": 1.3686, + "theoretical_loss": 3.568542201934422, + "tokens_seen": 1389428736 + }, + { + "epoch": 0.42, + "learning_rate": 0.00046789969909729186, + "loss": 1.3559, + "theoretical_loss": 3.5685272630171996, + "tokens_seen": 1389494272 + }, + { + "epoch": 0.42, + "learning_rate": 0.0004678836509528586, + "loss": 1.281, + "theoretical_loss": 3.568512325001837, + "tokens_seen": 1389559808 + }, + { + "epoch": 0.42, + "learning_rate": 0.00046786760280842526, + "loss": 1.3421, + "theoretical_loss": 3.5684973878882382, + "tokens_seen": 1389625344 + }, + { + "epoch": 0.42, + "learning_rate": 0.000467851554663992, + "loss": 1.4788, + "theoretical_loss": 3.568482451676305, + "tokens_seen": 1389690880 + }, + { + "epoch": 0.42, + "learning_rate": 0.00046783550651955866, + "loss": 1.4821, + "theoretical_loss": 3.568467516365941, + "tokens_seen": 1389756416 + }, + { + "epoch": 0.42, + "learning_rate": 0.0004678194583751254, + "loss": 1.3903, + "theoretical_loss": 3.5684525819570494, + "tokens_seen": 1389821952 + }, + { + "epoch": 0.42, + "learning_rate": 0.00046780341023069206, + "loss": 1.358, + "theoretical_loss": 3.5684376484495335, + "tokens_seen": 1389887488 + }, + { + "epoch": 0.42, + "learning_rate": 0.0004677873620862588, + "loss": 1.2589, + "theoretical_loss": 3.5684227158432953, + "tokens_seen": 1389953024 + }, + { + "epoch": 0.42, + "learning_rate": 0.00046777131394182546, + "loss": 1.3725, + "theoretical_loss": 3.5684077841382393, + "tokens_seen": 1390018560 + }, + { + "epoch": 0.42, + "learning_rate": 0.0004677552657973922, + "loss": 1.2762, + "theoretical_loss": 3.5683928533342675, + "tokens_seen": 1390084096 + }, + { + "epoch": 0.42, + "learning_rate": 0.00046773921765295886, + "loss": 1.374, + "theoretical_loss": 3.5683779234312833, + "tokens_seen": 1390149632 + }, + { + "epoch": 0.42, + "learning_rate": 0.0004677231695085256, + "loss": 1.227, + "theoretical_loss": 3.5683629944291906, + "tokens_seen": 1390215168 + }, + { + "epoch": 0.42, + "learning_rate": 0.00046770712136409226, + "loss": 1.3102, + "theoretical_loss": 3.568348066327892, + "tokens_seen": 1390280704 + }, + { + "epoch": 0.42, + "learning_rate": 0.000467691073219659, + "loss": 1.2931, + "theoretical_loss": 3.56833313912729, + "tokens_seen": 1390346240 + }, + { + "epoch": 0.42, + "learning_rate": 0.00046767502507522567, + "loss": 1.4146, + "theoretical_loss": 3.5683182128272897, + "tokens_seen": 1390411776 + }, + { + "epoch": 0.42, + "learning_rate": 0.0004676589769307924, + "loss": 1.4577, + "theoretical_loss": 3.5683032874277925, + "tokens_seen": 1390477312 + }, + { + "epoch": 0.42, + "learning_rate": 0.00046764292878635907, + "loss": 1.3549, + "theoretical_loss": 3.568288362928702, + "tokens_seen": 1390542848 + }, + { + "epoch": 0.42, + "learning_rate": 0.0004676268806419258, + "loss": 1.4304, + "theoretical_loss": 3.5682734393299222, + "tokens_seen": 1390608384 + }, + { + "epoch": 0.42, + "learning_rate": 0.00046761083249749247, + "loss": 1.3958, + "theoretical_loss": 3.568258516631356, + "tokens_seen": 1390673920 + }, + { + "epoch": 0.42, + "learning_rate": 0.0004675947843530592, + "loss": 1.3488, + "theoretical_loss": 3.5682435948329063, + "tokens_seen": 1390739456 + }, + { + "epoch": 0.42, + "learning_rate": 0.00046757873620862587, + "loss": 1.2143, + "theoretical_loss": 3.5682286739344766, + "tokens_seen": 1390804992 + }, + { + "epoch": 0.42, + "learning_rate": 0.0004675626880641926, + "loss": 1.3097, + "theoretical_loss": 3.5682137539359706, + "tokens_seen": 1390870528 + }, + { + "epoch": 0.42, + "learning_rate": 0.00046754663991975927, + "loss": 1.3678, + "theoretical_loss": 3.5681988348372915, + "tokens_seen": 1390936064 + }, + { + "epoch": 0.42, + "learning_rate": 0.000467530591775326, + "loss": 1.3132, + "theoretical_loss": 3.568183916638342, + "tokens_seen": 1391001600 + }, + { + "epoch": 0.42, + "learning_rate": 0.00046751454363089267, + "loss": 1.3868, + "theoretical_loss": 3.5681689993390258, + "tokens_seen": 1391067136 + }, + { + "epoch": 0.42, + "learning_rate": 0.0004674984954864594, + "loss": 1.3344, + "theoretical_loss": 3.568154082939247, + "tokens_seen": 1391132672 + }, + { + "epoch": 0.42, + "learning_rate": 0.0004674824473420261, + "loss": 1.2855, + "theoretical_loss": 3.5681391674389076, + "tokens_seen": 1391198208 + }, + { + "epoch": 0.42, + "learning_rate": 0.00046746639919759283, + "loss": 1.2841, + "theoretical_loss": 3.568124252837912, + "tokens_seen": 1391263744 + }, + { + "epoch": 0.42, + "learning_rate": 0.0004674503510531595, + "loss": 1.3921, + "theoretical_loss": 3.5681093391361633, + "tokens_seen": 1391329280 + }, + { + "epoch": 0.42, + "learning_rate": 0.00046743430290872623, + "loss": 1.3505, + "theoretical_loss": 3.568094426333565, + "tokens_seen": 1391394816 + }, + { + "epoch": 0.42, + "learning_rate": 0.0004674182547642929, + "loss": 1.4409, + "theoretical_loss": 3.5680795144300212, + "tokens_seen": 1391460352 + }, + { + "epoch": 0.42, + "learning_rate": 0.00046740220661985963, + "loss": 1.2673, + "theoretical_loss": 3.5680646034254337, + "tokens_seen": 1391525888 + }, + { + "epoch": 0.42, + "learning_rate": 0.0004673861584754263, + "loss": 1.3517, + "theoretical_loss": 3.5680496933197077, + "tokens_seen": 1391591424 + }, + { + "epoch": 0.42, + "learning_rate": 0.00046737011033099303, + "loss": 1.2196, + "theoretical_loss": 3.5680347841127453, + "tokens_seen": 1391656960 + }, + { + "epoch": 0.42, + "learning_rate": 0.0004673540621865597, + "loss": 1.3646, + "theoretical_loss": 3.5680198758044512, + "tokens_seen": 1391722496 + }, + { + "epoch": 0.42, + "learning_rate": 0.0004673380140421264, + "loss": 1.3802, + "theoretical_loss": 3.568004968394728, + "tokens_seen": 1391788032 + }, + { + "epoch": 0.42, + "learning_rate": 0.0004673219658976931, + "loss": 1.2889, + "theoretical_loss": 3.56799006188348, + "tokens_seen": 1391853568 + }, + { + "epoch": 0.42, + "learning_rate": 0.0004673059177532598, + "loss": 1.3159, + "theoretical_loss": 3.5679751562706103, + "tokens_seen": 1391919104 + }, + { + "epoch": 0.42, + "learning_rate": 0.0004672898696088265, + "loss": 1.3906, + "theoretical_loss": 3.5679602515560225, + "tokens_seen": 1391984640 + }, + { + "epoch": 0.42, + "learning_rate": 0.0004672738214643932, + "loss": 1.3635, + "theoretical_loss": 3.5679453477396197, + "tokens_seen": 1392050176 + }, + { + "epoch": 0.42, + "learning_rate": 0.00046725777331995994, + "loss": 1.3261, + "theoretical_loss": 3.5679304448213065, + "tokens_seen": 1392115712 + }, + { + "epoch": 0.42, + "learning_rate": 0.0004672417251755266, + "loss": 1.2934, + "theoretical_loss": 3.5679155428009857, + "tokens_seen": 1392181248 + }, + { + "epoch": 0.42, + "learning_rate": 0.00046722567703109334, + "loss": 1.3561, + "theoretical_loss": 3.5679006416785617, + "tokens_seen": 1392246784 + }, + { + "epoch": 0.42, + "learning_rate": 0.00046720962888666, + "loss": 1.3615, + "theoretical_loss": 3.567885741453937, + "tokens_seen": 1392312320 + }, + { + "epoch": 0.42, + "learning_rate": 0.00046719358074222674, + "loss": 1.3813, + "theoretical_loss": 3.567870842127016, + "tokens_seen": 1392377856 + }, + { + "epoch": 0.42, + "learning_rate": 0.0004671775325977934, + "loss": 1.2829, + "theoretical_loss": 3.567855943697703, + "tokens_seen": 1392443392 + }, + { + "epoch": 0.42, + "learning_rate": 0.00046716148445336014, + "loss": 1.3957, + "theoretical_loss": 3.5678410461659, + "tokens_seen": 1392508928 + }, + { + "epoch": 0.42, + "learning_rate": 0.0004671454363089268, + "loss": 1.4153, + "theoretical_loss": 3.567826149531512, + "tokens_seen": 1392574464 + }, + { + "epoch": 0.42, + "objective/train/docs_used": 508733, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.24476957321167, + "objective/train/theoretical_loss": 3.567811253794442, + "objective/train/tokens_used": 1413100000, + "theoretical_loss": 3.567811253794442, + "tokens_seen": 1392640000 + }, + { + "epoch": 0.42, + "learning_rate": 0.00046712938816449354, + "loss": 1.2824, + "theoretical_loss": 3.567811253794442, + "tokens_seen": 1392640000 + }, + { + "epoch": 0.42, + "learning_rate": 0.0004671133400200602, + "loss": 1.4179, + "theoretical_loss": 3.5677963589545953, + "tokens_seen": 1392705536 + }, + { + "epoch": 0.42, + "learning_rate": 0.00046709729187562694, + "loss": 1.391, + "theoretical_loss": 3.5677814650118735, + "tokens_seen": 1392771072 + }, + { + "epoch": 0.42, + "learning_rate": 0.0004670812437311936, + "loss": 1.3858, + "theoretical_loss": 3.567766571966181, + "tokens_seen": 1392836608 + }, + { + "epoch": 0.42, + "learning_rate": 0.00046706519558676034, + "loss": 1.379, + "theoretical_loss": 3.5677516798174223, + "tokens_seen": 1392902144 + }, + { + "epoch": 0.42, + "learning_rate": 0.000467049147442327, + "loss": 1.2699, + "theoretical_loss": 3.567736788565501, + "tokens_seen": 1392967680 + }, + { + "epoch": 0.42, + "learning_rate": 0.00046703309929789375, + "loss": 1.3544, + "theoretical_loss": 3.56772189821032, + "tokens_seen": 1393033216 + }, + { + "epoch": 0.42, + "learning_rate": 0.0004670170511534604, + "loss": 1.367, + "theoretical_loss": 3.567707008751784, + "tokens_seen": 1393098752 + }, + { + "epoch": 0.42, + "learning_rate": 0.00046700100300902715, + "loss": 1.3302, + "theoretical_loss": 3.567692120189797, + "tokens_seen": 1393164288 + }, + { + "epoch": 0.42, + "learning_rate": 0.0004669849548645938, + "loss": 1.4529, + "theoretical_loss": 3.567677232524262, + "tokens_seen": 1393229824 + }, + { + "epoch": 0.42, + "learning_rate": 0.00046696890672016055, + "loss": 1.419, + "theoretical_loss": 3.567662345755083, + "tokens_seen": 1393295360 + }, + { + "epoch": 0.42, + "learning_rate": 0.0004669528585757272, + "loss": 1.3062, + "theoretical_loss": 3.5676474598821644, + "tokens_seen": 1393360896 + }, + { + "epoch": 0.42, + "learning_rate": 0.00046693681043129395, + "loss": 1.3307, + "theoretical_loss": 3.56763257490541, + "tokens_seen": 1393426432 + }, + { + "epoch": 0.42, + "learning_rate": 0.0004669207622868606, + "loss": 1.3139, + "theoretical_loss": 3.567617690824724, + "tokens_seen": 1393491968 + }, + { + "epoch": 0.42, + "learning_rate": 0.00046690471414242735, + "loss": 1.4649, + "theoretical_loss": 3.567602807640009, + "tokens_seen": 1393557504 + }, + { + "epoch": 0.42, + "learning_rate": 0.000466888665997994, + "loss": 1.3682, + "theoretical_loss": 3.5675879253511704, + "tokens_seen": 1393623040 + }, + { + "epoch": 0.42, + "learning_rate": 0.00046687261785356075, + "loss": 1.3393, + "theoretical_loss": 3.5675730439581113, + "tokens_seen": 1393688576 + }, + { + "epoch": 0.42, + "learning_rate": 0.0004668565697091274, + "loss": 1.3584, + "theoretical_loss": 3.5675581634607356, + "tokens_seen": 1393754112 + }, + { + "epoch": 0.42, + "learning_rate": 0.00046684052156469415, + "loss": 1.3349, + "theoretical_loss": 3.567543283858948, + "tokens_seen": 1393819648 + }, + { + "epoch": 0.42, + "learning_rate": 0.0004668244734202608, + "loss": 1.4528, + "theoretical_loss": 3.567528405152652, + "tokens_seen": 1393885184 + }, + { + "epoch": 0.42, + "learning_rate": 0.00046680842527582755, + "loss": 1.2771, + "theoretical_loss": 3.5675135273417515, + "tokens_seen": 1393950720 + }, + { + "epoch": 0.42, + "learning_rate": 0.0004667923771313942, + "loss": 1.4342, + "theoretical_loss": 3.5674986504261508, + "tokens_seen": 1394016256 + }, + { + "epoch": 0.42, + "learning_rate": 0.00046677632898696096, + "loss": 1.2479, + "theoretical_loss": 3.5674837744057544, + "tokens_seen": 1394081792 + }, + { + "epoch": 0.42, + "learning_rate": 0.0004667602808425276, + "loss": 1.4336, + "theoretical_loss": 3.567468899280465, + "tokens_seen": 1394147328 + }, + { + "epoch": 0.42, + "learning_rate": 0.00046674423269809436, + "loss": 1.4156, + "theoretical_loss": 3.5674540250501883, + "tokens_seen": 1394212864 + }, + { + "epoch": 0.42, + "learning_rate": 0.000466728184553661, + "loss": 1.3287, + "theoretical_loss": 3.5674391517148267, + "tokens_seen": 1394278400 + }, + { + "epoch": 0.42, + "learning_rate": 0.0004667121364092277, + "loss": 1.4318, + "theoretical_loss": 3.567424279274286, + "tokens_seen": 1394343936 + }, + { + "epoch": 0.42, + "learning_rate": 0.0004666960882647944, + "loss": 1.431, + "theoretical_loss": 3.567409407728469, + "tokens_seen": 1394409472 + }, + { + "epoch": 0.42, + "learning_rate": 0.0004666800401203611, + "loss": 1.3238, + "theoretical_loss": 3.56739453707728, + "tokens_seen": 1394475008 + }, + { + "epoch": 0.42, + "learning_rate": 0.0004666639919759278, + "loss": 1.3848, + "theoretical_loss": 3.567379667320624, + "tokens_seen": 1394540544 + }, + { + "epoch": 0.42, + "learning_rate": 0.0004666479438314945, + "loss": 1.436, + "theoretical_loss": 3.5673647984584043, + "tokens_seen": 1394606080 + }, + { + "epoch": 0.42, + "learning_rate": 0.0004666318956870612, + "loss": 1.3033, + "theoretical_loss": 3.567349930490525, + "tokens_seen": 1394671616 + }, + { + "epoch": 0.42, + "learning_rate": 0.0004666158475426279, + "loss": 1.4536, + "theoretical_loss": 3.567335063416891, + "tokens_seen": 1394737152 + }, + { + "epoch": 0.42, + "learning_rate": 0.0004665997993981946, + "loss": 1.3565, + "theoretical_loss": 3.5673201972374065, + "tokens_seen": 1394802688 + }, + { + "epoch": 0.42, + "learning_rate": 0.0004665837512537613, + "loss": 1.4041, + "theoretical_loss": 3.567305331951975, + "tokens_seen": 1394868224 + }, + { + "epoch": 0.42, + "learning_rate": 0.00046656770310932796, + "loss": 1.4615, + "theoretical_loss": 3.5672904675605013, + "tokens_seen": 1394933760 + }, + { + "epoch": 0.42, + "learning_rate": 0.0004665516549648947, + "loss": 1.324, + "theoretical_loss": 3.567275604062889, + "tokens_seen": 1394999296 + }, + { + "epoch": 0.42, + "learning_rate": 0.00046653560682046136, + "loss": 1.4411, + "theoretical_loss": 3.5672607414590427, + "tokens_seen": 1395064832 + }, + { + "epoch": 0.42, + "learning_rate": 0.0004665195586760281, + "loss": 1.2911, + "theoretical_loss": 3.5672458797488673, + "tokens_seen": 1395130368 + }, + { + "epoch": 0.42, + "learning_rate": 0.00046650351053159476, + "loss": 1.4075, + "theoretical_loss": 3.567231018932266, + "tokens_seen": 1395195904 + }, + { + "epoch": 0.42, + "learning_rate": 0.0004664874623871615, + "loss": 1.3913, + "theoretical_loss": 3.5672161590091442, + "tokens_seen": 1395261440 + }, + { + "epoch": 0.42, + "learning_rate": 0.00046647141424272816, + "loss": 1.3488, + "theoretical_loss": 3.5672012999794056, + "tokens_seen": 1395326976 + }, + { + "epoch": 0.42, + "learning_rate": 0.0004664553660982949, + "loss": 1.5841, + "theoretical_loss": 3.5671864418429546, + "tokens_seen": 1395392512 + }, + { + "epoch": 0.42, + "learning_rate": 0.00046643931795386156, + "loss": 1.3435, + "theoretical_loss": 3.567171584599695, + "tokens_seen": 1395458048 + }, + { + "epoch": 0.42, + "learning_rate": 0.0004664232698094283, + "loss": 1.3911, + "theoretical_loss": 3.5671567282495324, + "tokens_seen": 1395523584 + }, + { + "epoch": 0.42, + "learning_rate": 0.00046640722166499496, + "loss": 1.4088, + "theoretical_loss": 3.5671418727923703, + "tokens_seen": 1395589120 + }, + { + "epoch": 0.42, + "learning_rate": 0.0004663911735205617, + "loss": 1.382, + "theoretical_loss": 3.567127018228113, + "tokens_seen": 1395654656 + }, + { + "epoch": 0.42, + "learning_rate": 0.00046637512537612836, + "loss": 1.382, + "theoretical_loss": 3.5671121645566655, + "tokens_seen": 1395720192 + }, + { + "epoch": 0.42, + "learning_rate": 0.0004663590772316951, + "loss": 1.2725, + "theoretical_loss": 3.567097311777932, + "tokens_seen": 1395785728 + }, + { + "epoch": 0.42, + "learning_rate": 0.00046634302908726176, + "loss": 1.3795, + "theoretical_loss": 3.5670824598918163, + "tokens_seen": 1395851264 + }, + { + "epoch": 0.42, + "objective/train/docs_used": 509897, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.2610862255096436, + "objective/train/theoretical_loss": 3.5670676088982236, + "objective/train/tokens_used": 1416376800, + "theoretical_loss": 3.5670676088982236, + "tokens_seen": 1395916800 + }, + { + "epoch": 0.42, + "learning_rate": 0.0004663269809428285, + "loss": 1.3221, + "theoretical_loss": 3.5670676088982236, + "tokens_seen": 1395916800 + }, + { + "epoch": 0.42, + "learning_rate": 0.00046631093279839517, + "loss": 1.5032, + "theoretical_loss": 3.5670527587970584, + "tokens_seen": 1395982336 + }, + { + "epoch": 0.42, + "learning_rate": 0.0004662948846539619, + "loss": 1.4219, + "theoretical_loss": 3.5670379095882248, + "tokens_seen": 1396047872 + }, + { + "epoch": 0.42, + "learning_rate": 0.00046627883650952857, + "loss": 1.3176, + "theoretical_loss": 3.5670230612716276, + "tokens_seen": 1396113408 + }, + { + "epoch": 0.42, + "learning_rate": 0.0004662627883650953, + "loss": 1.4503, + "theoretical_loss": 3.5670082138471706, + "tokens_seen": 1396178944 + }, + { + "epoch": 0.42, + "learning_rate": 0.00046624674022066197, + "loss": 1.3232, + "theoretical_loss": 3.5669933673147596, + "tokens_seen": 1396244480 + }, + { + "epoch": 0.42, + "learning_rate": 0.0004662306920762287, + "loss": 1.3784, + "theoretical_loss": 3.566978521674298, + "tokens_seen": 1396310016 + }, + { + "epoch": 0.42, + "learning_rate": 0.00046621464393179537, + "loss": 1.4811, + "theoretical_loss": 3.566963676925691, + "tokens_seen": 1396375552 + }, + { + "epoch": 0.42, + "learning_rate": 0.0004661985957873621, + "loss": 1.3238, + "theoretical_loss": 3.5669488330688424, + "tokens_seen": 1396441088 + }, + { + "epoch": 0.42, + "learning_rate": 0.00046618254764292877, + "loss": 1.4105, + "theoretical_loss": 3.566933990103658, + "tokens_seen": 1396506624 + }, + { + "epoch": 0.42, + "learning_rate": 0.0004661664994984955, + "loss": 1.3362, + "theoretical_loss": 3.566919148030041, + "tokens_seen": 1396572160 + }, + { + "epoch": 0.42, + "learning_rate": 0.00046615045135406217, + "loss": 1.4643, + "theoretical_loss": 3.5669043068478974, + "tokens_seen": 1396637696 + }, + { + "epoch": 0.42, + "learning_rate": 0.0004661344032096289, + "loss": 1.3501, + "theoretical_loss": 3.5668894665571305, + "tokens_seen": 1396703232 + }, + { + "epoch": 0.42, + "learning_rate": 0.0004661183550651956, + "loss": 1.419, + "theoretical_loss": 3.566874627157646, + "tokens_seen": 1396768768 + }, + { + "epoch": 0.42, + "learning_rate": 0.00046610230692076233, + "loss": 1.4588, + "theoretical_loss": 3.5668597886493485, + "tokens_seen": 1396834304 + }, + { + "epoch": 0.42, + "learning_rate": 0.000466086258776329, + "loss": 1.4067, + "theoretical_loss": 3.5668449510321416, + "tokens_seen": 1396899840 + }, + { + "epoch": 0.42, + "learning_rate": 0.00046607021063189573, + "loss": 1.4288, + "theoretical_loss": 3.5668301143059313, + "tokens_seen": 1396965376 + }, + { + "epoch": 0.42, + "learning_rate": 0.00046605416248746243, + "loss": 1.274, + "theoretical_loss": 3.5668152784706213, + "tokens_seen": 1397030912 + }, + { + "epoch": 0.42, + "learning_rate": 0.00046603811434302913, + "loss": 1.2993, + "theoretical_loss": 3.566800443526117, + "tokens_seen": 1397096448 + }, + { + "epoch": 0.42, + "learning_rate": 0.00046602206619859583, + "loss": 1.5038, + "theoretical_loss": 3.5667856094723227, + "tokens_seen": 1397161984 + }, + { + "epoch": 0.42, + "learning_rate": 0.00046600601805416253, + "loss": 1.343, + "theoretical_loss": 3.566770776309143, + "tokens_seen": 1397227520 + }, + { + "epoch": 0.42, + "learning_rate": 0.00046598996990972923, + "loss": 1.4145, + "theoretical_loss": 3.5667559440364833, + "tokens_seen": 1397293056 + }, + { + "epoch": 0.42, + "learning_rate": 0.00046597392176529593, + "loss": 1.4389, + "theoretical_loss": 3.566741112654248, + "tokens_seen": 1397358592 + }, + { + "epoch": 0.42, + "learning_rate": 0.00046595787362086263, + "loss": 1.52, + "theoretical_loss": 3.566726282162342, + "tokens_seen": 1397424128 + }, + { + "epoch": 0.42, + "learning_rate": 0.0004659418254764293, + "loss": 1.3571, + "theoretical_loss": 3.5667114525606696, + "tokens_seen": 1397489664 + }, + { + "epoch": 0.42, + "learning_rate": 0.00046592577733199603, + "loss": 1.3214, + "theoretical_loss": 3.566696623849136, + "tokens_seen": 1397555200 + }, + { + "epoch": 0.42, + "learning_rate": 0.0004659097291875627, + "loss": 1.3605, + "theoretical_loss": 3.5666817960276465, + "tokens_seen": 1397620736 + }, + { + "epoch": 0.42, + "learning_rate": 0.00046589368104312944, + "loss": 1.3698, + "theoretical_loss": 3.5666669690961057, + "tokens_seen": 1397686272 + }, + { + "epoch": 0.42, + "learning_rate": 0.0004658776328986961, + "loss": 1.3879, + "theoretical_loss": 3.566652143054418, + "tokens_seen": 1397751808 + }, + { + "epoch": 0.42, + "learning_rate": 0.00046586158475426284, + "loss": 1.3522, + "theoretical_loss": 3.5666373179024884, + "tokens_seen": 1397817344 + }, + { + "epoch": 0.42, + "learning_rate": 0.0004658455366098295, + "loss": 1.4697, + "theoretical_loss": 3.566622493640222, + "tokens_seen": 1397882880 + }, + { + "epoch": 0.42, + "learning_rate": 0.00046582948846539624, + "loss": 1.3498, + "theoretical_loss": 3.5666076702675236, + "tokens_seen": 1397948416 + }, + { + "epoch": 0.42, + "learning_rate": 0.0004658134403209629, + "loss": 1.3037, + "theoretical_loss": 3.5665928477842987, + "tokens_seen": 1398013952 + }, + { + "epoch": 0.42, + "learning_rate": 0.00046579739217652964, + "loss": 1.3631, + "theoretical_loss": 3.566578026190451, + "tokens_seen": 1398079488 + }, + { + "epoch": 0.42, + "learning_rate": 0.0004657813440320963, + "loss": 1.3699, + "theoretical_loss": 3.5665632054858865, + "tokens_seen": 1398145024 + }, + { + "epoch": 0.42, + "learning_rate": 0.00046576529588766304, + "loss": 1.4752, + "theoretical_loss": 3.56654838567051, + "tokens_seen": 1398210560 + }, + { + "epoch": 0.42, + "learning_rate": 0.0004657492477432297, + "loss": 1.4379, + "theoretical_loss": 3.5665335667442255, + "tokens_seen": 1398276096 + }, + { + "epoch": 0.42, + "learning_rate": 0.00046573319959879644, + "loss": 1.4287, + "theoretical_loss": 3.5665187487069394, + "tokens_seen": 1398341632 + }, + { + "epoch": 0.42, + "learning_rate": 0.0004657171514543631, + "loss": 1.352, + "theoretical_loss": 3.5665039315585556, + "tokens_seen": 1398407168 + }, + { + "epoch": 0.42, + "learning_rate": 0.00046570110330992984, + "loss": 1.4224, + "theoretical_loss": 3.56648911529898, + "tokens_seen": 1398472704 + }, + { + "epoch": 0.42, + "learning_rate": 0.0004656850551654965, + "loss": 1.3612, + "theoretical_loss": 3.5664742999281174, + "tokens_seen": 1398538240 + }, + { + "epoch": 0.42, + "learning_rate": 0.00046566900702106325, + "loss": 1.4222, + "theoretical_loss": 3.566459485445873, + "tokens_seen": 1398603776 + }, + { + "epoch": 0.42, + "learning_rate": 0.0004656529588766299, + "loss": 1.3595, + "theoretical_loss": 3.5664446718521505, + "tokens_seen": 1398669312 + }, + { + "epoch": 0.42, + "learning_rate": 0.00046563691073219665, + "loss": 1.3254, + "theoretical_loss": 3.566429859146857, + "tokens_seen": 1398734848 + }, + { + "epoch": 0.42, + "learning_rate": 0.0004656208625877633, + "loss": 1.4152, + "theoretical_loss": 3.566415047329896, + "tokens_seen": 1398800384 + }, + { + "epoch": 0.42, + "learning_rate": 0.00046560481444333005, + "loss": 1.357, + "theoretical_loss": 3.566400236401174, + "tokens_seen": 1398865920 + }, + { + "epoch": 0.42, + "learning_rate": 0.0004655887662988967, + "loss": 1.4352, + "theoretical_loss": 3.5663854263605943, + "tokens_seen": 1398931456 + }, + { + "epoch": 0.42, + "learning_rate": 0.00046557271815446345, + "loss": 1.2978, + "theoretical_loss": 3.5663706172080643, + "tokens_seen": 1398996992 + }, + { + "epoch": 0.42, + "learning_rate": 0.0004655566700100301, + "loss": 1.4041, + "theoretical_loss": 3.566355808943487, + "tokens_seen": 1399062528 + }, + { + "epoch": 0.42, + "learning_rate": 0.00046554062186559685, + "loss": 1.3245, + "theoretical_loss": 3.5663410015667694, + "tokens_seen": 1399128064 + }, + { + "epoch": 0.42, + "objective/train/docs_used": 511124, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.3808932304382324, + "objective/train/theoretical_loss": 3.5663261950778153, + "objective/train/tokens_used": 1419653600, + "theoretical_loss": 3.5663261950778153, + "tokens_seen": 1399193600 + }, + { + "epoch": 0.42, + "learning_rate": 0.0004655245737211635, + "loss": 1.4148, + "theoretical_loss": 3.5663261950778153, + "tokens_seen": 1399193600 + }, + { + "epoch": 0.42, + "learning_rate": 0.00046550852557673025, + "loss": 1.4284, + "theoretical_loss": 3.56631138947653, + "tokens_seen": 1399259136 + }, + { + "epoch": 0.42, + "learning_rate": 0.0004654924774322969, + "loss": 1.4633, + "theoretical_loss": 3.56629658476282, + "tokens_seen": 1399324672 + }, + { + "epoch": 0.42, + "learning_rate": 0.00046547642928786365, + "loss": 1.246, + "theoretical_loss": 3.566281780936589, + "tokens_seen": 1399390208 + }, + { + "epoch": 0.42, + "learning_rate": 0.0004654603811434303, + "loss": 1.4147, + "theoretical_loss": 3.566266977997743, + "tokens_seen": 1399455744 + }, + { + "epoch": 0.42, + "learning_rate": 0.00046544433299899705, + "loss": 1.3208, + "theoretical_loss": 3.5662521759461874, + "tokens_seen": 1399521280 + }, + { + "epoch": 0.42, + "learning_rate": 0.0004654282848545637, + "loss": 1.3496, + "theoretical_loss": 3.566237374781827, + "tokens_seen": 1399586816 + }, + { + "epoch": 0.42, + "learning_rate": 0.00046541223671013046, + "loss": 1.3822, + "theoretical_loss": 3.5662225745045677, + "tokens_seen": 1399652352 + }, + { + "epoch": 0.42, + "learning_rate": 0.0004653961885656971, + "loss": 1.2924, + "theoretical_loss": 3.566207775114314, + "tokens_seen": 1399717888 + }, + { + "epoch": 0.42, + "learning_rate": 0.00046538014042126386, + "loss": 1.3998, + "theoretical_loss": 3.5661929766109717, + "tokens_seen": 1399783424 + }, + { + "epoch": 0.42, + "learning_rate": 0.0004653640922768305, + "loss": 1.3856, + "theoretical_loss": 3.5661781789944462, + "tokens_seen": 1399848960 + }, + { + "epoch": 0.42, + "learning_rate": 0.00046534804413239726, + "loss": 1.3796, + "theoretical_loss": 3.5661633822646426, + "tokens_seen": 1399914496 + }, + { + "epoch": 0.42, + "learning_rate": 0.0004653319959879639, + "loss": 1.4119, + "theoretical_loss": 3.566148586421466, + "tokens_seen": 1399980032 + }, + { + "epoch": 0.42, + "learning_rate": 0.0004653159478435306, + "loss": 1.2732, + "theoretical_loss": 3.5661337914648223, + "tokens_seen": 1400045568 + }, + { + "epoch": 0.42, + "learning_rate": 0.0004652998996990973, + "loss": 1.3486, + "theoretical_loss": 3.566118997394617, + "tokens_seen": 1400111104 + }, + { + "epoch": 0.42, + "learning_rate": 0.000465283851554664, + "loss": 1.4028, + "theoretical_loss": 3.566104204210755, + "tokens_seen": 1400176640 + }, + { + "epoch": 0.42, + "learning_rate": 0.0004652678034102307, + "loss": 1.4143, + "theoretical_loss": 3.566089411913142, + "tokens_seen": 1400242176 + }, + { + "epoch": 0.42, + "learning_rate": 0.0004652517552657974, + "loss": 1.4244, + "theoretical_loss": 3.5660746205016833, + "tokens_seen": 1400307712 + }, + { + "epoch": 0.42, + "learning_rate": 0.0004652357071213641, + "loss": 1.3966, + "theoretical_loss": 3.5660598299762847, + "tokens_seen": 1400373248 + }, + { + "epoch": 0.42, + "learning_rate": 0.0004652196589769308, + "loss": 1.4071, + "theoretical_loss": 3.566045040336851, + "tokens_seen": 1400438784 + }, + { + "epoch": 0.42, + "learning_rate": 0.0004652036108324975, + "loss": 1.3571, + "theoretical_loss": 3.566030251583288, + "tokens_seen": 1400504320 + }, + { + "epoch": 0.42, + "learning_rate": 0.0004651875626880642, + "loss": 1.4013, + "theoretical_loss": 3.5660154637155013, + "tokens_seen": 1400569856 + }, + { + "epoch": 0.42, + "learning_rate": 0.00046517151454363086, + "loss": 1.5395, + "theoretical_loss": 3.5660006767333963, + "tokens_seen": 1400635392 + }, + { + "epoch": 0.42, + "learning_rate": 0.0004651554663991976, + "loss": 1.4284, + "theoretical_loss": 3.5659858906368784, + "tokens_seen": 1400700928 + }, + { + "epoch": 0.42, + "learning_rate": 0.00046513941825476426, + "loss": 1.4806, + "theoretical_loss": 3.5659711054258536, + "tokens_seen": 1400766464 + }, + { + "epoch": 0.42, + "learning_rate": 0.000465123370110331, + "loss": 1.517, + "theoretical_loss": 3.5659563211002268, + "tokens_seen": 1400832000 + }, + { + "epoch": 0.42, + "learning_rate": 0.00046510732196589766, + "loss": 1.3765, + "theoretical_loss": 3.5659415376599037, + "tokens_seen": 1400897536 + }, + { + "epoch": 0.42, + "learning_rate": 0.0004650912738214644, + "loss": 1.4034, + "theoretical_loss": 3.5659267551047904, + "tokens_seen": 1400963072 + }, + { + "epoch": 0.42, + "learning_rate": 0.00046507522567703106, + "loss": 1.2932, + "theoretical_loss": 3.565911973434792, + "tokens_seen": 1401028608 + }, + { + "epoch": 0.42, + "learning_rate": 0.0004650591775325978, + "loss": 1.2714, + "theoretical_loss": 3.565897192649815, + "tokens_seen": 1401094144 + }, + { + "epoch": 0.42, + "learning_rate": 0.00046504312938816446, + "loss": 1.276, + "theoretical_loss": 3.565882412749763, + "tokens_seen": 1401159680 + }, + { + "epoch": 0.42, + "learning_rate": 0.0004650270812437312, + "loss": 1.3713, + "theoretical_loss": 3.5658676337345434, + "tokens_seen": 1401225216 + }, + { + "epoch": 0.42, + "learning_rate": 0.00046501103309929786, + "loss": 1.3992, + "theoretical_loss": 3.5658528556040614, + "tokens_seen": 1401290752 + }, + { + "epoch": 0.42, + "learning_rate": 0.0004649949849548646, + "loss": 1.2832, + "theoretical_loss": 3.5658380783582224, + "tokens_seen": 1401356288 + }, + { + "epoch": 0.42, + "learning_rate": 0.00046497893681043126, + "loss": 1.2984, + "theoretical_loss": 3.5658233019969328, + "tokens_seen": 1401421824 + }, + { + "epoch": 0.42, + "learning_rate": 0.000464962888665998, + "loss": 1.3473, + "theoretical_loss": 3.5658085265200974, + "tokens_seen": 1401487360 + }, + { + "epoch": 0.42, + "learning_rate": 0.00046494684052156467, + "loss": 1.3844, + "theoretical_loss": 3.5657937519276226, + "tokens_seen": 1401552896 + }, + { + "epoch": 0.42, + "learning_rate": 0.0004649307923771314, + "loss": 1.3924, + "theoretical_loss": 3.565778978219413, + "tokens_seen": 1401618432 + }, + { + "epoch": 0.42, + "learning_rate": 0.00046491474423269807, + "loss": 1.3339, + "theoretical_loss": 3.565764205395376, + "tokens_seen": 1401683968 + }, + { + "epoch": 0.42, + "learning_rate": 0.0004648986960882648, + "loss": 1.3623, + "theoretical_loss": 3.565749433455416, + "tokens_seen": 1401749504 + }, + { + "epoch": 0.42, + "learning_rate": 0.00046488264794383147, + "loss": 1.409, + "theoretical_loss": 3.5657346623994393, + "tokens_seen": 1401815040 + }, + { + "epoch": 0.42, + "learning_rate": 0.0004648665997993982, + "loss": 1.4727, + "theoretical_loss": 3.565719892227351, + "tokens_seen": 1401880576 + }, + { + "epoch": 0.42, + "learning_rate": 0.000464850551654965, + "loss": 1.4076, + "theoretical_loss": 3.5657051229390584, + "tokens_seen": 1401946112 + }, + { + "epoch": 0.42, + "learning_rate": 0.0004648345035105316, + "loss": 1.4027, + "theoretical_loss": 3.565690354534466, + "tokens_seen": 1402011648 + }, + { + "epoch": 0.42, + "learning_rate": 0.0004648184553660984, + "loss": 1.3668, + "theoretical_loss": 3.56567558701348, + "tokens_seen": 1402077184 + }, + { + "epoch": 0.42, + "learning_rate": 0.000464802407221665, + "loss": 1.2569, + "theoretical_loss": 3.5656608203760065, + "tokens_seen": 1402142720 + }, + { + "epoch": 0.42, + "learning_rate": 0.0004647863590772318, + "loss": 1.401, + "theoretical_loss": 3.5656460546219515, + "tokens_seen": 1402208256 + }, + { + "epoch": 0.42, + "learning_rate": 0.0004647703109327984, + "loss": 1.3331, + "theoretical_loss": 3.5656312897512197, + "tokens_seen": 1402273792 + }, + { + "epoch": 0.42, + "learning_rate": 0.0004647542627883652, + "loss": 1.4353, + "theoretical_loss": 3.5656165257637182, + "tokens_seen": 1402339328 + }, + { + "epoch": 0.42, + "learning_rate": 0.00046473821464393183, + "loss": 1.5529, + "theoretical_loss": 3.5656017626593526, + "tokens_seen": 1402404864 + }, + { + "epoch": 0.42, + "objective/train/docs_used": 512300, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.5094163417816162, + "objective/train/theoretical_loss": 3.565587000438028, + "objective/train/tokens_used": 1422930400, + "theoretical_loss": 3.565587000438028, + "tokens_seen": 1402470400 + }, + { + "epoch": 0.42, + "learning_rate": 0.0004647221664994986, + "loss": 1.5032, + "theoretical_loss": 3.565587000438028, + "tokens_seen": 1402470400 + }, + { + "epoch": 0.43, + "learning_rate": 0.00046470611835506523, + "loss": 1.3804, + "theoretical_loss": 3.5655722390996516, + "tokens_seen": 1402535936 + }, + { + "epoch": 0.43, + "learning_rate": 0.00046469007021063193, + "loss": 1.291, + "theoretical_loss": 3.5655574786441284, + "tokens_seen": 1402601472 + }, + { + "epoch": 0.43, + "learning_rate": 0.00046467402206619863, + "loss": 1.4625, + "theoretical_loss": 3.565542719071365, + "tokens_seen": 1402667008 + }, + { + "epoch": 0.43, + "learning_rate": 0.00046465797392176533, + "loss": 1.4856, + "theoretical_loss": 3.5655279603812664, + "tokens_seen": 1402732544 + }, + { + "epoch": 0.43, + "learning_rate": 0.00046464192577733203, + "loss": 1.2809, + "theoretical_loss": 3.5655132025737393, + "tokens_seen": 1402798080 + }, + { + "epoch": 0.43, + "learning_rate": 0.00046462587763289873, + "loss": 1.4916, + "theoretical_loss": 3.5654984456486902, + "tokens_seen": 1402863616 + }, + { + "epoch": 0.43, + "learning_rate": 0.00046460982948846543, + "loss": 1.2778, + "theoretical_loss": 3.5654836896060242, + "tokens_seen": 1402929152 + }, + { + "epoch": 0.43, + "learning_rate": 0.00046459378134403213, + "loss": 1.328, + "theoretical_loss": 3.5654689344456476, + "tokens_seen": 1402994688 + }, + { + "epoch": 0.43, + "learning_rate": 0.00046457773319959883, + "loss": 1.3872, + "theoretical_loss": 3.5654541801674666, + "tokens_seen": 1403060224 + }, + { + "epoch": 0.43, + "learning_rate": 0.00046456168505516553, + "loss": 1.3685, + "theoretical_loss": 3.5654394267713867, + "tokens_seen": 1403125760 + }, + { + "epoch": 0.43, + "learning_rate": 0.0004645456369107322, + "loss": 1.3852, + "theoretical_loss": 3.565424674257315, + "tokens_seen": 1403191296 + }, + { + "epoch": 0.43, + "learning_rate": 0.00046452958876629894, + "loss": 1.359, + "theoretical_loss": 3.5654099226251565, + "tokens_seen": 1403256832 + }, + { + "epoch": 0.43, + "learning_rate": 0.0004645135406218656, + "loss": 1.369, + "theoretical_loss": 3.565395171874818, + "tokens_seen": 1403322368 + }, + { + "epoch": 0.43, + "learning_rate": 0.00046449749247743234, + "loss": 1.4197, + "theoretical_loss": 3.5653804220062053, + "tokens_seen": 1403387904 + }, + { + "epoch": 0.43, + "learning_rate": 0.000464481444332999, + "loss": 1.3081, + "theoretical_loss": 3.565365673019225, + "tokens_seen": 1403453440 + }, + { + "epoch": 0.43, + "learning_rate": 0.00046446539618856574, + "loss": 1.4197, + "theoretical_loss": 3.5653509249137825, + "tokens_seen": 1403518976 + }, + { + "epoch": 0.43, + "learning_rate": 0.0004644493480441324, + "loss": 1.3651, + "theoretical_loss": 3.565336177689784, + "tokens_seen": 1403584512 + }, + { + "epoch": 0.43, + "learning_rate": 0.00046443329989969914, + "loss": 1.4622, + "theoretical_loss": 3.5653214313471366, + "tokens_seen": 1403650048 + }, + { + "epoch": 0.43, + "learning_rate": 0.0004644172517552658, + "loss": 1.3139, + "theoretical_loss": 3.565306685885745, + "tokens_seen": 1403715584 + }, + { + "epoch": 0.43, + "learning_rate": 0.00046440120361083254, + "loss": 1.4389, + "theoretical_loss": 3.565291941305517, + "tokens_seen": 1403781120 + }, + { + "epoch": 0.43, + "learning_rate": 0.0004643851554663992, + "loss": 1.408, + "theoretical_loss": 3.5652771976063575, + "tokens_seen": 1403846656 + }, + { + "epoch": 0.43, + "learning_rate": 0.00046436910732196594, + "loss": 1.3832, + "theoretical_loss": 3.5652624547881735, + "tokens_seen": 1403912192 + }, + { + "epoch": 0.43, + "learning_rate": 0.0004643530591775326, + "loss": 1.4089, + "theoretical_loss": 3.565247712850871, + "tokens_seen": 1403977728 + }, + { + "epoch": 0.43, + "learning_rate": 0.00046433701103309934, + "loss": 1.3276, + "theoretical_loss": 3.565232971794356, + "tokens_seen": 1404043264 + }, + { + "epoch": 0.43, + "learning_rate": 0.000464320962888666, + "loss": 1.3229, + "theoretical_loss": 3.5652182316185357, + "tokens_seen": 1404108800 + }, + { + "epoch": 0.43, + "learning_rate": 0.00046430491474423274, + "loss": 1.3411, + "theoretical_loss": 3.565203492323315, + "tokens_seen": 1404174336 + }, + { + "epoch": 0.43, + "learning_rate": 0.0004642888665997994, + "loss": 1.368, + "theoretical_loss": 3.5651887539086013, + "tokens_seen": 1404239872 + }, + { + "epoch": 0.43, + "learning_rate": 0.00046427281845536615, + "loss": 1.3775, + "theoretical_loss": 3.5651740163743004, + "tokens_seen": 1404305408 + }, + { + "epoch": 0.43, + "learning_rate": 0.0004642567703109328, + "loss": 1.3948, + "theoretical_loss": 3.5651592797203184, + "tokens_seen": 1404370944 + }, + { + "epoch": 0.43, + "learning_rate": 0.00046424072216649955, + "loss": 1.4366, + "theoretical_loss": 3.5651445439465625, + "tokens_seen": 1404436480 + }, + { + "epoch": 0.43, + "learning_rate": 0.0004642246740220662, + "loss": 1.3924, + "theoretical_loss": 3.565129809052938, + "tokens_seen": 1404502016 + }, + { + "epoch": 0.43, + "learning_rate": 0.00046420862587763295, + "loss": 1.4341, + "theoretical_loss": 3.565115075039352, + "tokens_seen": 1404567552 + }, + { + "epoch": 0.43, + "learning_rate": 0.0004641925777331996, + "loss": 1.3839, + "theoretical_loss": 3.565100341905711, + "tokens_seen": 1404633088 + }, + { + "epoch": 0.43, + "learning_rate": 0.00046417652958876635, + "loss": 1.3505, + "theoretical_loss": 3.565085609651921, + "tokens_seen": 1404698624 + }, + { + "epoch": 0.43, + "learning_rate": 0.000464160481444333, + "loss": 1.2981, + "theoretical_loss": 3.565070878277888, + "tokens_seen": 1404764160 + }, + { + "epoch": 0.43, + "learning_rate": 0.00046414443329989975, + "loss": 1.4381, + "theoretical_loss": 3.5650561477835194, + "tokens_seen": 1404829696 + }, + { + "epoch": 0.43, + "learning_rate": 0.0004641283851554664, + "loss": 1.3893, + "theoretical_loss": 3.565041418168721, + "tokens_seen": 1404895232 + }, + { + "epoch": 0.43, + "learning_rate": 0.00046411233701103315, + "loss": 1.2936, + "theoretical_loss": 3.5650266894333993, + "tokens_seen": 1404960768 + }, + { + "epoch": 0.43, + "learning_rate": 0.0004640962888665998, + "loss": 1.4636, + "theoretical_loss": 3.565011961577461, + "tokens_seen": 1405026304 + }, + { + "epoch": 0.43, + "learning_rate": 0.00046408024072216655, + "loss": 1.5176, + "theoretical_loss": 3.5649972346008125, + "tokens_seen": 1405091840 + }, + { + "epoch": 0.43, + "learning_rate": 0.0004640641925777332, + "loss": 1.4796, + "theoretical_loss": 3.56498250850336, + "tokens_seen": 1405157376 + }, + { + "epoch": 0.43, + "learning_rate": 0.00046404814443329996, + "loss": 1.3584, + "theoretical_loss": 3.5649677832850104, + "tokens_seen": 1405222912 + }, + { + "epoch": 0.43, + "learning_rate": 0.0004640320962888666, + "loss": 1.4489, + "theoretical_loss": 3.5649530589456697, + "tokens_seen": 1405288448 + }, + { + "epoch": 0.43, + "learning_rate": 0.00046401604814443336, + "loss": 1.3773, + "theoretical_loss": 3.5649383354852455, + "tokens_seen": 1405353984 + }, + { + "epoch": 0.43, + "learning_rate": 0.000464, + "loss": 1.4074, + "theoretical_loss": 3.564923612903643, + "tokens_seen": 1405419520 + }, + { + "epoch": 0.43, + "learning_rate": 0.00046398395185556676, + "loss": 1.4104, + "theoretical_loss": 3.56490889120077, + "tokens_seen": 1405485056 + }, + { + "epoch": 0.43, + "learning_rate": 0.0004639679037111334, + "loss": 1.3281, + "theoretical_loss": 3.5648941703765322, + "tokens_seen": 1405550592 + }, + { + "epoch": 0.43, + "learning_rate": 0.00046395185556670016, + "loss": 1.3214, + "theoretical_loss": 3.5648794504308365, + "tokens_seen": 1405616128 + }, + { + "epoch": 0.43, + "learning_rate": 0.0004639358074222668, + "loss": 1.5505, + "theoretical_loss": 3.5648647313635893, + "tokens_seen": 1405681664 + }, + { + "epoch": 0.43, + "objective/train/docs_used": 513528, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.3503563404083252, + "objective/train/theoretical_loss": 3.5648500131746976, + "objective/train/tokens_used": 1426207200, + "theoretical_loss": 3.5648500131746976, + "tokens_seen": 1405747200 + }, + { + "epoch": 0.43, + "learning_rate": 0.0004639197592778335, + "loss": 1.3928, + "theoretical_loss": 3.5648500131746976, + "tokens_seen": 1405747200 + }, + { + "epoch": 0.43, + "learning_rate": 0.0004639037111334002, + "loss": 1.4664, + "theoretical_loss": 3.5648352958640688, + "tokens_seen": 1405812736 + }, + { + "epoch": 0.43, + "learning_rate": 0.0004638876629889669, + "loss": 1.2814, + "theoretical_loss": 3.5648205794316077, + "tokens_seen": 1405878272 + }, + { + "epoch": 0.43, + "learning_rate": 0.0004638716148445336, + "loss": 1.4233, + "theoretical_loss": 3.564805863877222, + "tokens_seen": 1405943808 + }, + { + "epoch": 0.43, + "learning_rate": 0.0004638555667001003, + "loss": 1.4088, + "theoretical_loss": 3.5647911492008184, + "tokens_seen": 1406009344 + }, + { + "epoch": 0.43, + "learning_rate": 0.000463839518555667, + "loss": 1.4396, + "theoretical_loss": 3.5647764354023037, + "tokens_seen": 1406074880 + }, + { + "epoch": 0.43, + "learning_rate": 0.0004638234704112337, + "loss": 1.4581, + "theoretical_loss": 3.5647617224815837, + "tokens_seen": 1406140416 + }, + { + "epoch": 0.43, + "learning_rate": 0.0004638074222668004, + "loss": 1.2227, + "theoretical_loss": 3.5647470104385666, + "tokens_seen": 1406205952 + }, + { + "epoch": 0.43, + "learning_rate": 0.0004637913741223671, + "loss": 1.4871, + "theoretical_loss": 3.564732299273158, + "tokens_seen": 1406271488 + }, + { + "epoch": 0.43, + "learning_rate": 0.00046377532597793376, + "loss": 1.4584, + "theoretical_loss": 3.564717588985265, + "tokens_seen": 1406337024 + }, + { + "epoch": 0.43, + "learning_rate": 0.0004637592778335005, + "loss": 1.4014, + "theoretical_loss": 3.564702879574795, + "tokens_seen": 1406402560 + }, + { + "epoch": 0.43, + "learning_rate": 0.00046374322968906716, + "loss": 1.3708, + "theoretical_loss": 3.564688171041653, + "tokens_seen": 1406468096 + }, + { + "epoch": 0.43, + "learning_rate": 0.0004637271815446339, + "loss": 1.2796, + "theoretical_loss": 3.5646734633857475, + "tokens_seen": 1406533632 + }, + { + "epoch": 0.43, + "learning_rate": 0.00046371113340020056, + "loss": 1.3243, + "theoretical_loss": 3.5646587566069856, + "tokens_seen": 1406599168 + }, + { + "epoch": 0.43, + "learning_rate": 0.0004636950852557673, + "loss": 1.2513, + "theoretical_loss": 3.5646440507052723, + "tokens_seen": 1406664704 + }, + { + "epoch": 0.43, + "learning_rate": 0.00046367903711133396, + "loss": 1.3659, + "theoretical_loss": 3.5646293456805154, + "tokens_seen": 1406730240 + }, + { + "epoch": 0.43, + "learning_rate": 0.0004636629889669007, + "loss": 1.2957, + "theoretical_loss": 3.5646146415326223, + "tokens_seen": 1406795776 + }, + { + "epoch": 0.43, + "learning_rate": 0.00046364694082246747, + "loss": 1.4563, + "theoretical_loss": 3.564599938261499, + "tokens_seen": 1406861312 + }, + { + "epoch": 0.43, + "learning_rate": 0.0004636308926780341, + "loss": 1.3167, + "theoretical_loss": 3.564585235867053, + "tokens_seen": 1406926848 + }, + { + "epoch": 0.43, + "learning_rate": 0.00046361484453360087, + "loss": 1.3768, + "theoretical_loss": 3.5645705343491914, + "tokens_seen": 1406992384 + }, + { + "epoch": 0.43, + "learning_rate": 0.0004635987963891675, + "loss": 1.426, + "theoretical_loss": 3.5645558337078196, + "tokens_seen": 1407057920 + }, + { + "epoch": 0.43, + "learning_rate": 0.0004635827482447343, + "loss": 1.3943, + "theoretical_loss": 3.5645411339428463, + "tokens_seen": 1407123456 + }, + { + "epoch": 0.43, + "learning_rate": 0.0004635667001003009, + "loss": 1.3386, + "theoretical_loss": 3.5645264350541774, + "tokens_seen": 1407188992 + }, + { + "epoch": 0.43, + "learning_rate": 0.0004635506519558677, + "loss": 1.4001, + "theoretical_loss": 3.56451173704172, + "tokens_seen": 1407254528 + }, + { + "epoch": 0.43, + "learning_rate": 0.0004635346038114343, + "loss": 1.4592, + "theoretical_loss": 3.5644970399053815, + "tokens_seen": 1407320064 + }, + { + "epoch": 0.43, + "learning_rate": 0.0004635185556670011, + "loss": 1.371, + "theoretical_loss": 3.5644823436450688, + "tokens_seen": 1407385600 + }, + { + "epoch": 0.43, + "learning_rate": 0.0004635025075225677, + "loss": 1.3786, + "theoretical_loss": 3.5644676482606883, + "tokens_seen": 1407451136 + }, + { + "epoch": 0.43, + "learning_rate": 0.0004634864593781345, + "loss": 1.5246, + "theoretical_loss": 3.5644529537521477, + "tokens_seen": 1407516672 + }, + { + "epoch": 0.43, + "learning_rate": 0.0004634704112337011, + "loss": 1.2586, + "theoretical_loss": 3.5644382601193536, + "tokens_seen": 1407582208 + }, + { + "epoch": 0.43, + "learning_rate": 0.0004634543630892679, + "loss": 1.4384, + "theoretical_loss": 3.564423567362213, + "tokens_seen": 1407647744 + }, + { + "epoch": 0.43, + "learning_rate": 0.0004634383149448345, + "loss": 1.3515, + "theoretical_loss": 3.564408875480633, + "tokens_seen": 1407713280 + }, + { + "epoch": 0.43, + "learning_rate": 0.0004634222668004013, + "loss": 1.3038, + "theoretical_loss": 3.564394184474521, + "tokens_seen": 1407778816 + }, + { + "epoch": 0.43, + "learning_rate": 0.0004634062186559679, + "loss": 1.3569, + "theoretical_loss": 3.5643794943437843, + "tokens_seen": 1407844352 + }, + { + "epoch": 0.43, + "learning_rate": 0.0004633901705115347, + "loss": 1.3753, + "theoretical_loss": 3.5643648050883288, + "tokens_seen": 1407909888 + }, + { + "epoch": 0.43, + "learning_rate": 0.00046337412236710133, + "loss": 1.3885, + "theoretical_loss": 3.564350116708063, + "tokens_seen": 1407975424 + }, + { + "epoch": 0.43, + "learning_rate": 0.0004633580742226681, + "loss": 1.3887, + "theoretical_loss": 3.5643354292028935, + "tokens_seen": 1408040960 + }, + { + "epoch": 0.43, + "learning_rate": 0.00046334202607823473, + "loss": 1.5194, + "theoretical_loss": 3.564320742572727, + "tokens_seen": 1408106496 + }, + { + "epoch": 0.43, + "learning_rate": 0.0004633259779338015, + "loss": 1.3444, + "theoretical_loss": 3.564306056817471, + "tokens_seen": 1408172032 + }, + { + "epoch": 0.43, + "learning_rate": 0.00046330992978936813, + "loss": 1.4935, + "theoretical_loss": 3.564291371937032, + "tokens_seen": 1408237568 + }, + { + "epoch": 0.43, + "learning_rate": 0.00046329388164493483, + "loss": 1.3509, + "theoretical_loss": 3.5642766879313186, + "tokens_seen": 1408303104 + }, + { + "epoch": 0.43, + "learning_rate": 0.00046327783350050153, + "loss": 1.4725, + "theoretical_loss": 3.5642620048002374, + "tokens_seen": 1408368640 + }, + { + "epoch": 0.43, + "learning_rate": 0.00046326178535606823, + "loss": 1.4504, + "theoretical_loss": 3.5642473225436953, + "tokens_seen": 1408434176 + }, + { + "epoch": 0.43, + "learning_rate": 0.00046324573721163493, + "loss": 1.3473, + "theoretical_loss": 3.5642326411615994, + "tokens_seen": 1408499712 + }, + { + "epoch": 0.43, + "learning_rate": 0.00046322968906720163, + "loss": 1.4875, + "theoretical_loss": 3.564217960653857, + "tokens_seen": 1408565248 + }, + { + "epoch": 0.43, + "learning_rate": 0.00046321364092276833, + "loss": 1.3231, + "theoretical_loss": 3.564203281020376, + "tokens_seen": 1408630784 + }, + { + "epoch": 0.43, + "learning_rate": 0.00046319759277833503, + "loss": 1.41, + "theoretical_loss": 3.564188602261063, + "tokens_seen": 1408696320 + }, + { + "epoch": 0.43, + "learning_rate": 0.00046318154463390174, + "loss": 1.3118, + "theoretical_loss": 3.564173924375826, + "tokens_seen": 1408761856 + }, + { + "epoch": 0.43, + "learning_rate": 0.00046316549648946844, + "loss": 1.4073, + "theoretical_loss": 3.5641592473645707, + "tokens_seen": 1408827392 + }, + { + "epoch": 0.43, + "learning_rate": 0.0004631494483450351, + "loss": 1.2818, + "theoretical_loss": 3.564144571227206, + "tokens_seen": 1408892928 + }, + { + "epoch": 0.43, + "learning_rate": 0.00046313340020060184, + "loss": 1.3609, + "theoretical_loss": 3.5641298959636387, + "tokens_seen": 1408958464 + }, + { + "epoch": 0.43, + "objective/train/docs_used": 514827, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.5618319511413574, + "objective/train/theoretical_loss": 3.5641152215737764, + "objective/train/tokens_used": 1429484000, + "theoretical_loss": 3.5641152215737764, + "tokens_seen": 1409024000 + }, + { + "epoch": 0.43, + "learning_rate": 0.0004631173520561685, + "loss": 1.5014, + "theoretical_loss": 3.5641152215737764, + "tokens_seen": 1409024000 + }, + { + "epoch": 0.43, + "learning_rate": 0.00046310130391173524, + "loss": 1.3576, + "theoretical_loss": 3.5641005480575263, + "tokens_seen": 1409089536 + }, + { + "epoch": 0.43, + "learning_rate": 0.0004630852557673019, + "loss": 1.3868, + "theoretical_loss": 3.5640858754147953, + "tokens_seen": 1409155072 + }, + { + "epoch": 0.43, + "learning_rate": 0.00046306920762286864, + "loss": 1.3257, + "theoretical_loss": 3.564071203645491, + "tokens_seen": 1409220608 + }, + { + "epoch": 0.43, + "learning_rate": 0.0004630531594784353, + "loss": 1.3471, + "theoretical_loss": 3.564056532749521, + "tokens_seen": 1409286144 + }, + { + "epoch": 0.43, + "learning_rate": 0.00046303711133400204, + "loss": 1.4087, + "theoretical_loss": 3.5640418627267927, + "tokens_seen": 1409351680 + }, + { + "epoch": 0.43, + "learning_rate": 0.0004630210631895687, + "loss": 1.4046, + "theoretical_loss": 3.5640271935772136, + "tokens_seen": 1409417216 + }, + { + "epoch": 0.43, + "learning_rate": 0.00046300501504513544, + "loss": 1.3174, + "theoretical_loss": 3.564012525300691, + "tokens_seen": 1409482752 + }, + { + "epoch": 0.43, + "learning_rate": 0.0004629889669007021, + "loss": 1.4293, + "theoretical_loss": 3.563997857897132, + "tokens_seen": 1409548288 + }, + { + "epoch": 0.43, + "learning_rate": 0.00046297291875626884, + "loss": 1.3952, + "theoretical_loss": 3.563983191366445, + "tokens_seen": 1409613824 + }, + { + "epoch": 0.43, + "learning_rate": 0.0004629568706118355, + "loss": 1.4213, + "theoretical_loss": 3.5639685257085363, + "tokens_seen": 1409679360 + }, + { + "epoch": 0.43, + "learning_rate": 0.00046294082246740224, + "loss": 1.3193, + "theoretical_loss": 3.563953860923314, + "tokens_seen": 1409744896 + }, + { + "epoch": 0.43, + "learning_rate": 0.0004629247743229689, + "loss": 1.5631, + "theoretical_loss": 3.5639391970106855, + "tokens_seen": 1409810432 + }, + { + "epoch": 0.43, + "learning_rate": 0.00046290872617853565, + "loss": 1.3614, + "theoretical_loss": 3.5639245339705585, + "tokens_seen": 1409875968 + }, + { + "epoch": 0.43, + "learning_rate": 0.0004628926780341023, + "loss": 1.4304, + "theoretical_loss": 3.5639098718028404, + "tokens_seen": 1409941504 + }, + { + "epoch": 0.43, + "learning_rate": 0.00046287662988966905, + "loss": 1.323, + "theoretical_loss": 3.563895210507439, + "tokens_seen": 1410007040 + }, + { + "epoch": 0.43, + "learning_rate": 0.0004628605817452357, + "loss": 1.4166, + "theoretical_loss": 3.563880550084262, + "tokens_seen": 1410072576 + }, + { + "epoch": 0.43, + "learning_rate": 0.00046284453360080245, + "loss": 1.318, + "theoretical_loss": 3.563865890533216, + "tokens_seen": 1410138112 + }, + { + "epoch": 0.43, + "learning_rate": 0.0004628284854563691, + "loss": 1.3783, + "theoretical_loss": 3.5638512318542093, + "tokens_seen": 1410203648 + }, + { + "epoch": 0.43, + "learning_rate": 0.00046281243731193585, + "loss": 1.4712, + "theoretical_loss": 3.563836574047149, + "tokens_seen": 1410269184 + }, + { + "epoch": 0.43, + "learning_rate": 0.0004627963891675025, + "loss": 1.3617, + "theoretical_loss": 3.5638219171119436, + "tokens_seen": 1410334720 + }, + { + "epoch": 0.43, + "learning_rate": 0.00046278034102306925, + "loss": 1.3723, + "theoretical_loss": 3.5638072610485, + "tokens_seen": 1410400256 + }, + { + "epoch": 0.43, + "learning_rate": 0.0004627642928786359, + "loss": 1.3038, + "theoretical_loss": 3.563792605856726, + "tokens_seen": 1410465792 + }, + { + "epoch": 0.43, + "learning_rate": 0.00046274824473420265, + "loss": 1.3702, + "theoretical_loss": 3.5637779515365295, + "tokens_seen": 1410531328 + }, + { + "epoch": 0.43, + "learning_rate": 0.0004627321965897693, + "loss": 1.395, + "theoretical_loss": 3.563763298087818, + "tokens_seen": 1410596864 + }, + { + "epoch": 0.43, + "learning_rate": 0.00046271614844533605, + "loss": 1.5249, + "theoretical_loss": 3.5637486455104996, + "tokens_seen": 1410662400 + }, + { + "epoch": 0.43, + "learning_rate": 0.0004627001003009027, + "loss": 1.3959, + "theoretical_loss": 3.5637339938044814, + "tokens_seen": 1410727936 + }, + { + "epoch": 0.43, + "learning_rate": 0.00046268405215646945, + "loss": 1.4225, + "theoretical_loss": 3.563719342969671, + "tokens_seen": 1410793472 + }, + { + "epoch": 0.43, + "learning_rate": 0.0004626680040120361, + "loss": 1.3103, + "theoretical_loss": 3.5637046930059766, + "tokens_seen": 1410859008 + }, + { + "epoch": 0.43, + "learning_rate": 0.00046265195586760286, + "loss": 1.4465, + "theoretical_loss": 3.563690043913306, + "tokens_seen": 1410924544 + }, + { + "epoch": 0.43, + "learning_rate": 0.0004626359077231695, + "loss": 1.4591, + "theoretical_loss": 3.5636753956915665, + "tokens_seen": 1410990080 + }, + { + "epoch": 0.43, + "learning_rate": 0.00046261985957873626, + "loss": 1.3176, + "theoretical_loss": 3.5636607483406664, + "tokens_seen": 1411055616 + }, + { + "epoch": 0.43, + "learning_rate": 0.0004626038114343029, + "loss": 1.3775, + "theoretical_loss": 3.5636461018605132, + "tokens_seen": 1411121152 + }, + { + "epoch": 0.43, + "learning_rate": 0.00046258776328986966, + "loss": 1.4453, + "theoretical_loss": 3.5636314562510147, + "tokens_seen": 1411186688 + }, + { + "epoch": 0.43, + "learning_rate": 0.0004625717151454363, + "loss": 1.3831, + "theoretical_loss": 3.5636168115120785, + "tokens_seen": 1411252224 + }, + { + "epoch": 0.43, + "learning_rate": 0.00046255566700100306, + "loss": 1.5045, + "theoretical_loss": 3.5636021676436123, + "tokens_seen": 1411317760 + }, + { + "epoch": 0.43, + "learning_rate": 0.0004625396188565697, + "loss": 1.4697, + "theoretical_loss": 3.5635875246455253, + "tokens_seen": 1411383296 + }, + { + "epoch": 0.43, + "learning_rate": 0.0004625235707121364, + "loss": 1.4235, + "theoretical_loss": 3.5635728825177235, + "tokens_seen": 1411448832 + }, + { + "epoch": 0.43, + "learning_rate": 0.0004625075225677031, + "loss": 1.3939, + "theoretical_loss": 3.5635582412601163, + "tokens_seen": 1411514368 + }, + { + "epoch": 0.43, + "learning_rate": 0.0004624914744232698, + "loss": 1.4497, + "theoretical_loss": 3.5635436008726105, + "tokens_seen": 1411579904 + }, + { + "epoch": 0.43, + "learning_rate": 0.0004624754262788365, + "loss": 1.3846, + "theoretical_loss": 3.5635289613551144, + "tokens_seen": 1411645440 + }, + { + "epoch": 0.43, + "learning_rate": 0.0004624593781344032, + "loss": 1.3899, + "theoretical_loss": 3.563514322707536, + "tokens_seen": 1411710976 + }, + { + "epoch": 0.43, + "learning_rate": 0.00046244332998996996, + "loss": 1.6376, + "theoretical_loss": 3.5634996849297833, + "tokens_seen": 1411776512 + }, + { + "epoch": 0.43, + "learning_rate": 0.0004624272818455366, + "loss": 1.3696, + "theoretical_loss": 3.563485048021764, + "tokens_seen": 1411842048 + }, + { + "epoch": 0.43, + "learning_rate": 0.00046241123370110337, + "loss": 1.3591, + "theoretical_loss": 3.5634704119833858, + "tokens_seen": 1411907584 + }, + { + "epoch": 0.43, + "learning_rate": 0.00046239518555667, + "loss": 1.3883, + "theoretical_loss": 3.563455776814558, + "tokens_seen": 1411973120 + }, + { + "epoch": 0.43, + "learning_rate": 0.00046237913741223677, + "loss": 1.4915, + "theoretical_loss": 3.5634411425151864, + "tokens_seen": 1412038656 + }, + { + "epoch": 0.43, + "learning_rate": 0.0004623630892678034, + "loss": 1.4286, + "theoretical_loss": 3.563426509085181, + "tokens_seen": 1412104192 + }, + { + "epoch": 0.43, + "learning_rate": 0.00046234704112337017, + "loss": 1.3524, + "theoretical_loss": 3.5634118765244485, + "tokens_seen": 1412169728 + }, + { + "epoch": 0.43, + "learning_rate": 0.0004623309929789368, + "loss": 1.2866, + "theoretical_loss": 3.5633972448328977, + "tokens_seen": 1412235264 + }, + { + "epoch": 0.43, + "objective/train/docs_used": 516049, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.4440743923187256, + "objective/train/theoretical_loss": 3.563382614010436, + "objective/train/tokens_used": 1432760800, + "theoretical_loss": 3.563382614010436, + "tokens_seen": 1412300800 + }, + { + "epoch": 0.43, + "learning_rate": 0.00046231494483450357, + "loss": 1.4017, + "theoretical_loss": 3.563382614010436, + "tokens_seen": 1412300800 + }, + { + "epoch": 0.43, + "learning_rate": 0.0004622988966900702, + "loss": 1.4392, + "theoretical_loss": 3.5633679840569723, + "tokens_seen": 1412366336 + }, + { + "epoch": 0.43, + "learning_rate": 0.00046228284854563697, + "loss": 1.4205, + "theoretical_loss": 3.5633533549724143, + "tokens_seen": 1412431872 + }, + { + "epoch": 0.43, + "learning_rate": 0.0004622668004012036, + "loss": 1.3958, + "theoretical_loss": 3.56333872675667, + "tokens_seen": 1412497408 + }, + { + "epoch": 0.43, + "learning_rate": 0.00046225075225677037, + "loss": 1.3694, + "theoretical_loss": 3.563324099409647, + "tokens_seen": 1412562944 + }, + { + "epoch": 0.43, + "learning_rate": 0.000462234704112337, + "loss": 1.3174, + "theoretical_loss": 3.5633094729312536, + "tokens_seen": 1412628480 + }, + { + "epoch": 0.43, + "learning_rate": 0.0004622186559679038, + "loss": 1.4909, + "theoretical_loss": 3.563294847321399, + "tokens_seen": 1412694016 + }, + { + "epoch": 0.43, + "learning_rate": 0.0004622026078234704, + "loss": 1.4024, + "theoretical_loss": 3.5632802225799898, + "tokens_seen": 1412759552 + }, + { + "epoch": 0.43, + "learning_rate": 0.0004621865596790372, + "loss": 1.4605, + "theoretical_loss": 3.563265598706936, + "tokens_seen": 1412825088 + }, + { + "epoch": 0.43, + "learning_rate": 0.0004621705115346038, + "loss": 1.3638, + "theoretical_loss": 3.563250975702143, + "tokens_seen": 1412890624 + }, + { + "epoch": 0.43, + "learning_rate": 0.0004621544633901706, + "loss": 1.4234, + "theoretical_loss": 3.5632363535655216, + "tokens_seen": 1412956160 + }, + { + "epoch": 0.43, + "learning_rate": 0.0004621384152457372, + "loss": 1.3804, + "theoretical_loss": 3.563221732296979, + "tokens_seen": 1413021696 + }, + { + "epoch": 0.43, + "learning_rate": 0.000462122367101304, + "loss": 1.393, + "theoretical_loss": 3.563207111896424, + "tokens_seen": 1413087232 + }, + { + "epoch": 0.43, + "learning_rate": 0.0004621063189568706, + "loss": 1.4014, + "theoretical_loss": 3.5631924923637635, + "tokens_seen": 1413152768 + }, + { + "epoch": 0.43, + "learning_rate": 0.0004620902708124374, + "loss": 1.2813, + "theoretical_loss": 3.5631778736989066, + "tokens_seen": 1413218304 + }, + { + "epoch": 0.43, + "learning_rate": 0.000462074222668004, + "loss": 1.4126, + "theoretical_loss": 3.5631632559017614, + "tokens_seen": 1413283840 + }, + { + "epoch": 0.43, + "learning_rate": 0.0004620581745235708, + "loss": 1.2493, + "theoretical_loss": 3.563148638972236, + "tokens_seen": 1413349376 + }, + { + "epoch": 0.43, + "learning_rate": 0.0004620421263791374, + "loss": 1.4419, + "theoretical_loss": 3.563134022910239, + "tokens_seen": 1413414912 + }, + { + "epoch": 0.43, + "learning_rate": 0.0004620260782347042, + "loss": 1.3359, + "theoretical_loss": 3.5631194077156785, + "tokens_seen": 1413480448 + }, + { + "epoch": 0.43, + "learning_rate": 0.00046201003009027083, + "loss": 1.3874, + "theoretical_loss": 3.5631047933884634, + "tokens_seen": 1413545984 + }, + { + "epoch": 0.43, + "learning_rate": 0.0004619939819458376, + "loss": 1.3536, + "theoretical_loss": 3.563090179928501, + "tokens_seen": 1413611520 + }, + { + "epoch": 0.43, + "learning_rate": 0.00046197793380140423, + "loss": 1.4364, + "theoretical_loss": 3.5630755673357, + "tokens_seen": 1413677056 + }, + { + "epoch": 0.43, + "learning_rate": 0.000461961885656971, + "loss": 1.4814, + "theoretical_loss": 3.563060955609969, + "tokens_seen": 1413742592 + }, + { + "epoch": 0.43, + "learning_rate": 0.00046194583751253763, + "loss": 1.4869, + "theoretical_loss": 3.5630463447512164, + "tokens_seen": 1413808128 + }, + { + "epoch": 0.43, + "learning_rate": 0.0004619297893681044, + "loss": 1.4379, + "theoretical_loss": 3.5630317347593503, + "tokens_seen": 1413873664 + }, + { + "epoch": 0.43, + "learning_rate": 0.00046191374122367103, + "loss": 1.4347, + "theoretical_loss": 3.5630171256342793, + "tokens_seen": 1413939200 + }, + { + "epoch": 0.43, + "learning_rate": 0.0004618976930792378, + "loss": 1.2915, + "theoretical_loss": 3.563002517375911, + "tokens_seen": 1414004736 + }, + { + "epoch": 0.43, + "learning_rate": 0.00046188164493480443, + "loss": 1.4302, + "theoretical_loss": 3.5629879099841553, + "tokens_seen": 1414070272 + }, + { + "epoch": 0.43, + "learning_rate": 0.00046186559679037113, + "loss": 1.3249, + "theoretical_loss": 3.5629733034589197, + "tokens_seen": 1414135808 + }, + { + "epoch": 0.43, + "learning_rate": 0.00046184954864593783, + "loss": 1.4248, + "theoretical_loss": 3.5629586978001124, + "tokens_seen": 1414201344 + }, + { + "epoch": 0.43, + "learning_rate": 0.00046183350050150453, + "loss": 1.4594, + "theoretical_loss": 3.5629440930076424, + "tokens_seen": 1414266880 + }, + { + "epoch": 0.43, + "learning_rate": 0.00046181745235707123, + "loss": 1.492, + "theoretical_loss": 3.562929489081418, + "tokens_seen": 1414332416 + }, + { + "epoch": 0.43, + "learning_rate": 0.00046180140421263794, + "loss": 1.3561, + "theoretical_loss": 3.562914886021348, + "tokens_seen": 1414397952 + }, + { + "epoch": 0.43, + "learning_rate": 0.00046178535606820464, + "loss": 1.3121, + "theoretical_loss": 3.5629002838273403, + "tokens_seen": 1414463488 + }, + { + "epoch": 0.43, + "learning_rate": 0.00046176930792377134, + "loss": 1.3825, + "theoretical_loss": 3.5628856824993034, + "tokens_seen": 1414529024 + }, + { + "epoch": 0.43, + "learning_rate": 0.000461753259779338, + "loss": 1.48, + "theoretical_loss": 3.5628710820371463, + "tokens_seen": 1414594560 + }, + { + "epoch": 0.43, + "learning_rate": 0.00046173721163490474, + "loss": 1.4973, + "theoretical_loss": 3.5628564824407776, + "tokens_seen": 1414660096 + }, + { + "epoch": 0.43, + "learning_rate": 0.0004617211634904714, + "loss": 1.4305, + "theoretical_loss": 3.5628418837101057, + "tokens_seen": 1414725632 + }, + { + "epoch": 0.43, + "learning_rate": 0.00046170511534603814, + "loss": 1.3468, + "theoretical_loss": 3.5628272858450387, + "tokens_seen": 1414791168 + }, + { + "epoch": 0.43, + "learning_rate": 0.0004616890672016048, + "loss": 1.4387, + "theoretical_loss": 3.5628126888454856, + "tokens_seen": 1414856704 + }, + { + "epoch": 0.43, + "learning_rate": 0.00046167301905717154, + "loss": 1.3598, + "theoretical_loss": 3.562798092711355, + "tokens_seen": 1414922240 + }, + { + "epoch": 0.43, + "learning_rate": 0.0004616569709127382, + "loss": 1.3054, + "theoretical_loss": 3.5627834974425556, + "tokens_seen": 1414987776 + }, + { + "epoch": 0.43, + "learning_rate": 0.00046164092276830494, + "loss": 1.5385, + "theoretical_loss": 3.5627689030389957, + "tokens_seen": 1415053312 + }, + { + "epoch": 0.43, + "learning_rate": 0.0004616248746238716, + "loss": 1.3292, + "theoretical_loss": 3.5627543095005842, + "tokens_seen": 1415118848 + }, + { + "epoch": 0.43, + "learning_rate": 0.00046160882647943834, + "loss": 1.3978, + "theoretical_loss": 3.56273971682723, + "tokens_seen": 1415184384 + }, + { + "epoch": 0.43, + "learning_rate": 0.000461592778335005, + "loss": 1.4639, + "theoretical_loss": 3.562725125018841, + "tokens_seen": 1415249920 + }, + { + "epoch": 0.43, + "learning_rate": 0.00046157673019057174, + "loss": 1.4632, + "theoretical_loss": 3.5627105340753267, + "tokens_seen": 1415315456 + }, + { + "epoch": 0.43, + "learning_rate": 0.0004615606820461384, + "loss": 1.3749, + "theoretical_loss": 3.5626959439965953, + "tokens_seen": 1415380992 + }, + { + "epoch": 0.43, + "learning_rate": 0.00046154463390170515, + "loss": 1.3919, + "theoretical_loss": 3.5626813547825558, + "tokens_seen": 1415446528 + }, + { + "epoch": 0.43, + "learning_rate": 0.0004615285857572718, + "loss": 1.4184, + "theoretical_loss": 3.562666766433116, + "tokens_seen": 1415512064 + }, + { + "epoch": 0.43, + "objective/train/docs_used": 517252, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.4471797943115234, + "objective/train/theoretical_loss": 3.5626521789481864, + "objective/train/tokens_used": 1436037600, + "theoretical_loss": 3.5626521789481864, + "tokens_seen": 1415577600 + }, + { + "epoch": 0.43, + "learning_rate": 0.00046151253761283855, + "loss": 1.4333, + "theoretical_loss": 3.5626521789481864, + "tokens_seen": 1415577600 + }, + { + "epoch": 0.43, + "learning_rate": 0.0004614964894684052, + "loss": 1.2362, + "theoretical_loss": 3.5626375923276745, + "tokens_seen": 1415643136 + }, + { + "epoch": 0.43, + "learning_rate": 0.00046148044132397195, + "loss": 1.2935, + "theoretical_loss": 3.562623006571489, + "tokens_seen": 1415708672 + }, + { + "epoch": 0.43, + "learning_rate": 0.0004614643931795386, + "loss": 1.3833, + "theoretical_loss": 3.5626084216795393, + "tokens_seen": 1415774208 + }, + { + "epoch": 0.43, + "learning_rate": 0.00046144834503510535, + "loss": 1.3952, + "theoretical_loss": 3.5625938376517334, + "tokens_seen": 1415839744 + }, + { + "epoch": 0.43, + "learning_rate": 0.000461432296890672, + "loss": 1.3193, + "theoretical_loss": 3.562579254487981, + "tokens_seen": 1415905280 + }, + { + "epoch": 0.43, + "learning_rate": 0.00046141624874623875, + "loss": 1.5274, + "theoretical_loss": 3.5625646721881905, + "tokens_seen": 1415970816 + }, + { + "epoch": 0.43, + "learning_rate": 0.0004614002006018054, + "loss": 1.4406, + "theoretical_loss": 3.56255009075227, + "tokens_seen": 1416036352 + }, + { + "epoch": 0.43, + "learning_rate": 0.00046138415245737215, + "loss": 1.5026, + "theoretical_loss": 3.5625355101801297, + "tokens_seen": 1416101888 + }, + { + "epoch": 0.43, + "learning_rate": 0.0004613681043129388, + "loss": 1.4083, + "theoretical_loss": 3.5625209304716776, + "tokens_seen": 1416167424 + }, + { + "epoch": 0.43, + "learning_rate": 0.00046135205616850555, + "loss": 1.2978, + "theoretical_loss": 3.562506351626823, + "tokens_seen": 1416232960 + }, + { + "epoch": 0.43, + "learning_rate": 0.0004613360080240722, + "loss": 1.4999, + "theoretical_loss": 3.5624917736454744, + "tokens_seen": 1416298496 + }, + { + "epoch": 0.43, + "learning_rate": 0.00046131995987963895, + "loss": 1.4678, + "theoretical_loss": 3.5624771965275412, + "tokens_seen": 1416364032 + }, + { + "epoch": 0.43, + "learning_rate": 0.0004613039117352056, + "loss": 1.3902, + "theoretical_loss": 3.5624626202729317, + "tokens_seen": 1416429568 + }, + { + "epoch": 0.43, + "learning_rate": 0.00046128786359077236, + "loss": 1.5011, + "theoretical_loss": 3.5624480448815556, + "tokens_seen": 1416495104 + }, + { + "epoch": 0.43, + "learning_rate": 0.000461271815446339, + "loss": 1.2992, + "theoretical_loss": 3.562433470353321, + "tokens_seen": 1416560640 + }, + { + "epoch": 0.43, + "learning_rate": 0.00046125576730190576, + "loss": 1.3237, + "theoretical_loss": 3.562418896688137, + "tokens_seen": 1416626176 + }, + { + "epoch": 0.43, + "learning_rate": 0.00046123971915747246, + "loss": 1.4735, + "theoretical_loss": 3.5624043238859135, + "tokens_seen": 1416691712 + }, + { + "epoch": 0.43, + "learning_rate": 0.00046122367101303916, + "loss": 1.4567, + "theoretical_loss": 3.5623897519465584, + "tokens_seen": 1416757248 + }, + { + "epoch": 0.43, + "learning_rate": 0.00046120762286860586, + "loss": 1.3593, + "theoretical_loss": 3.5623751808699806, + "tokens_seen": 1416822784 + }, + { + "epoch": 0.43, + "learning_rate": 0.00046119157472417256, + "loss": 1.398, + "theoretical_loss": 3.5623606106560906, + "tokens_seen": 1416888320 + }, + { + "epoch": 0.43, + "learning_rate": 0.00046117552657973926, + "loss": 1.2338, + "theoretical_loss": 3.562346041304796, + "tokens_seen": 1416953856 + }, + { + "epoch": 0.43, + "learning_rate": 0.00046115947843530596, + "loss": 1.4201, + "theoretical_loss": 3.562331472816006, + "tokens_seen": 1417019392 + }, + { + "epoch": 0.43, + "learning_rate": 0.00046114343029087266, + "loss": 1.3086, + "theoretical_loss": 3.56231690518963, + "tokens_seen": 1417084928 + }, + { + "epoch": 0.43, + "learning_rate": 0.00046112738214643936, + "loss": 1.3862, + "theoretical_loss": 3.562302338425577, + "tokens_seen": 1417150464 + }, + { + "epoch": 0.43, + "learning_rate": 0.00046111133400200606, + "loss": 1.4267, + "theoretical_loss": 3.5622877725237565, + "tokens_seen": 1417216000 + }, + { + "epoch": 0.43, + "learning_rate": 0.0004610952858575727, + "loss": 1.2956, + "theoretical_loss": 3.5622732074840764, + "tokens_seen": 1417281536 + }, + { + "epoch": 0.43, + "learning_rate": 0.00046107923771313946, + "loss": 1.5788, + "theoretical_loss": 3.562258643306447, + "tokens_seen": 1417347072 + }, + { + "epoch": 0.43, + "learning_rate": 0.0004610631895687061, + "loss": 1.4242, + "theoretical_loss": 3.562244079990777, + "tokens_seen": 1417412608 + }, + { + "epoch": 0.43, + "learning_rate": 0.00046104714142427287, + "loss": 1.337, + "theoretical_loss": 3.5622295175369754, + "tokens_seen": 1417478144 + }, + { + "epoch": 0.43, + "learning_rate": 0.0004610310932798395, + "loss": 1.4227, + "theoretical_loss": 3.5622149559449516, + "tokens_seen": 1417543680 + }, + { + "epoch": 0.43, + "learning_rate": 0.00046101504513540627, + "loss": 1.3512, + "theoretical_loss": 3.5622003952146146, + "tokens_seen": 1417609216 + }, + { + "epoch": 0.43, + "learning_rate": 0.0004609989969909729, + "loss": 1.4595, + "theoretical_loss": 3.5621858353458737, + "tokens_seen": 1417674752 + }, + { + "epoch": 0.43, + "learning_rate": 0.00046098294884653967, + "loss": 1.3874, + "theoretical_loss": 3.5621712763386375, + "tokens_seen": 1417740288 + }, + { + "epoch": 0.43, + "learning_rate": 0.0004609669007021063, + "loss": 1.4608, + "theoretical_loss": 3.562156718192816, + "tokens_seen": 1417805824 + }, + { + "epoch": 0.43, + "learning_rate": 0.00046095085255767307, + "loss": 1.337, + "theoretical_loss": 3.5621421609083184, + "tokens_seen": 1417871360 + }, + { + "epoch": 0.43, + "learning_rate": 0.0004609348044132397, + "loss": 1.251, + "theoretical_loss": 3.5621276044850534, + "tokens_seen": 1417936896 + }, + { + "epoch": 0.43, + "learning_rate": 0.00046091875626880647, + "loss": 1.3079, + "theoretical_loss": 3.5621130489229307, + "tokens_seen": 1418002432 + }, + { + "epoch": 0.43, + "learning_rate": 0.0004609027081243731, + "loss": 1.3974, + "theoretical_loss": 3.5620984942218596, + "tokens_seen": 1418067968 + }, + { + "epoch": 0.43, + "learning_rate": 0.00046088665997993987, + "loss": 1.3994, + "theoretical_loss": 3.562083940381749, + "tokens_seen": 1418133504 + }, + { + "epoch": 0.43, + "learning_rate": 0.0004608706118355065, + "loss": 1.3006, + "theoretical_loss": 3.5620693874025076, + "tokens_seen": 1418199040 + }, + { + "epoch": 0.43, + "learning_rate": 0.00046085456369107327, + "loss": 1.3754, + "theoretical_loss": 3.562054835284046, + "tokens_seen": 1418264576 + }, + { + "epoch": 0.43, + "learning_rate": 0.0004608385155466399, + "loss": 1.3005, + "theoretical_loss": 3.5620402840262733, + "tokens_seen": 1418330112 + }, + { + "epoch": 0.43, + "learning_rate": 0.0004608224674022067, + "loss": 1.3557, + "theoretical_loss": 3.5620257336290977, + "tokens_seen": 1418395648 + }, + { + "epoch": 0.43, + "learning_rate": 0.0004608064192577733, + "loss": 1.4284, + "theoretical_loss": 3.5620111840924302, + "tokens_seen": 1418461184 + }, + { + "epoch": 0.43, + "learning_rate": 0.0004607903711133401, + "loss": 1.454, + "theoretical_loss": 3.5619966354161785, + "tokens_seen": 1418526720 + }, + { + "epoch": 0.43, + "learning_rate": 0.0004607743229689067, + "loss": 1.3951, + "theoretical_loss": 3.561982087600253, + "tokens_seen": 1418592256 + }, + { + "epoch": 0.43, + "learning_rate": 0.0004607582748244735, + "loss": 1.3401, + "theoretical_loss": 3.561967540644563, + "tokens_seen": 1418657792 + }, + { + "epoch": 0.43, + "learning_rate": 0.0004607422266800401, + "loss": 1.3977, + "theoretical_loss": 3.561952994549017, + "tokens_seen": 1418723328 + }, + { + "epoch": 0.43, + "learning_rate": 0.0004607261785356069, + "loss": 1.4347, + "theoretical_loss": 3.5619384493135264, + "tokens_seen": 1418788864 + }, + { + "epoch": 0.43, + "objective/train/docs_used": 518371, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.3688263893127441, + "objective/train/theoretical_loss": 3.5619239049379985, + "objective/train/tokens_used": 1439314400, + "theoretical_loss": 3.5619239049379985, + "tokens_seen": 1418854400 + }, + { + "epoch": 0.43, + "learning_rate": 0.0004607101303911735, + "loss": 1.3754, + "theoretical_loss": 3.5619239049379985, + "tokens_seen": 1418854400 + }, + { + "epoch": 0.43, + "learning_rate": 0.0004606940822467403, + "loss": 1.3411, + "theoretical_loss": 3.5619093614223436, + "tokens_seen": 1418919936 + }, + { + "epoch": 0.43, + "learning_rate": 0.0004606780341023069, + "loss": 1.5385, + "theoretical_loss": 3.561894818766471, + "tokens_seen": 1418985472 + }, + { + "epoch": 0.43, + "learning_rate": 0.0004606619859578737, + "loss": 1.3712, + "theoretical_loss": 3.5618802769702906, + "tokens_seen": 1419051008 + }, + { + "epoch": 0.43, + "learning_rate": 0.0004606459378134403, + "loss": 1.4314, + "theoretical_loss": 3.561865736033712, + "tokens_seen": 1419116544 + }, + { + "epoch": 0.43, + "learning_rate": 0.0004606298896690071, + "loss": 1.3985, + "theoretical_loss": 3.5618511959566437, + "tokens_seen": 1419182080 + }, + { + "epoch": 0.43, + "learning_rate": 0.00046061384152457373, + "loss": 1.4343, + "theoretical_loss": 3.561836656738996, + "tokens_seen": 1419247616 + }, + { + "epoch": 0.43, + "learning_rate": 0.0004605977933801405, + "loss": 1.4329, + "theoretical_loss": 3.5618221183806784, + "tokens_seen": 1419313152 + }, + { + "epoch": 0.43, + "learning_rate": 0.00046058174523570713, + "loss": 1.3116, + "theoretical_loss": 3.5618075808815997, + "tokens_seen": 1419378688 + }, + { + "epoch": 0.43, + "learning_rate": 0.0004605656970912739, + "loss": 1.3842, + "theoretical_loss": 3.5617930442416705, + "tokens_seen": 1419444224 + }, + { + "epoch": 0.43, + "learning_rate": 0.00046054964894684053, + "loss": 1.4033, + "theoretical_loss": 3.5617785084608, + "tokens_seen": 1419509760 + }, + { + "epoch": 0.43, + "learning_rate": 0.0004605336008024073, + "loss": 1.4504, + "theoretical_loss": 3.5617639735388975, + "tokens_seen": 1419575296 + }, + { + "epoch": 0.43, + "learning_rate": 0.00046051755265797393, + "loss": 1.3437, + "theoretical_loss": 3.5617494394758724, + "tokens_seen": 1419640832 + }, + { + "epoch": 0.43, + "learning_rate": 0.0004605015045135407, + "loss": 1.5103, + "theoretical_loss": 3.561734906271635, + "tokens_seen": 1419706368 + }, + { + "epoch": 0.43, + "learning_rate": 0.00046048545636910733, + "loss": 1.4143, + "theoretical_loss": 3.561720373926094, + "tokens_seen": 1419771904 + }, + { + "epoch": 0.43, + "learning_rate": 0.00046046940822467403, + "loss": 1.4419, + "theoretical_loss": 3.56170584243916, + "tokens_seen": 1419837440 + }, + { + "epoch": 0.43, + "learning_rate": 0.00046045336008024073, + "loss": 1.3925, + "theoretical_loss": 3.5616913118107423, + "tokens_seen": 1419902976 + }, + { + "epoch": 0.43, + "learning_rate": 0.00046043731193580744, + "loss": 1.3472, + "theoretical_loss": 3.56167678204075, + "tokens_seen": 1419968512 + }, + { + "epoch": 0.43, + "learning_rate": 0.00046042126379137414, + "loss": 1.3331, + "theoretical_loss": 3.561662253129094, + "tokens_seen": 1420034048 + }, + { + "epoch": 0.43, + "learning_rate": 0.00046040521564694084, + "loss": 1.4136, + "theoretical_loss": 3.5616477250756824, + "tokens_seen": 1420099584 + }, + { + "epoch": 0.43, + "learning_rate": 0.00046038916750250754, + "loss": 1.3538, + "theoretical_loss": 3.5616331978804263, + "tokens_seen": 1420165120 + }, + { + "epoch": 0.43, + "learning_rate": 0.00046037311935807424, + "loss": 1.3481, + "theoretical_loss": 3.5616186715432345, + "tokens_seen": 1420230656 + }, + { + "epoch": 0.43, + "learning_rate": 0.00046035707121364094, + "loss": 1.4156, + "theoretical_loss": 3.561604146064017, + "tokens_seen": 1420296192 + }, + { + "epoch": 0.43, + "learning_rate": 0.00046034102306920764, + "loss": 1.3821, + "theoretical_loss": 3.5615896214426837, + "tokens_seen": 1420361728 + }, + { + "epoch": 0.43, + "learning_rate": 0.0004603249749247743, + "loss": 1.3277, + "theoretical_loss": 3.5615750976791447, + "tokens_seen": 1420427264 + }, + { + "epoch": 0.43, + "learning_rate": 0.00046030892678034104, + "loss": 1.3724, + "theoretical_loss": 3.561560574773309, + "tokens_seen": 1420492800 + }, + { + "epoch": 0.43, + "learning_rate": 0.0004602928786359077, + "loss": 1.3719, + "theoretical_loss": 3.5615460527250864, + "tokens_seen": 1420558336 + }, + { + "epoch": 0.43, + "learning_rate": 0.00046027683049147444, + "loss": 1.4102, + "theoretical_loss": 3.5615315315343876, + "tokens_seen": 1420623872 + }, + { + "epoch": 0.43, + "learning_rate": 0.0004602607823470411, + "loss": 1.4473, + "theoretical_loss": 3.561517011201121, + "tokens_seen": 1420689408 + }, + { + "epoch": 0.43, + "learning_rate": 0.00046024473420260784, + "loss": 1.3921, + "theoretical_loss": 3.5615024917251983, + "tokens_seen": 1420754944 + }, + { + "epoch": 0.43, + "learning_rate": 0.00046024473420260784, + "loss": 1.4251, + "theoretical_loss": 3.5614879731065274, + "tokens_seen": 1420820480 + }, + { + "epoch": 0.43, + "learning_rate": 0.0004602286860581745, + "loss": 1.4228, + "theoretical_loss": 3.56147345534502, + "tokens_seen": 1420886016 + }, + { + "epoch": 0.43, + "learning_rate": 0.00046021263791374124, + "loss": 1.3851, + "theoretical_loss": 3.561458938440584, + "tokens_seen": 1420951552 + }, + { + "epoch": 0.43, + "learning_rate": 0.0004601965897693079, + "loss": 1.4223, + "theoretical_loss": 3.5614444223931305, + "tokens_seen": 1421017088 + }, + { + "epoch": 0.43, + "learning_rate": 0.00046018054162487465, + "loss": 1.3996, + "theoretical_loss": 3.561429907202569, + "tokens_seen": 1421082624 + }, + { + "epoch": 0.43, + "learning_rate": 0.0004601644934804413, + "loss": 1.3065, + "theoretical_loss": 3.5614153928688097, + "tokens_seen": 1421148160 + }, + { + "epoch": 0.43, + "learning_rate": 0.00046014844533600805, + "loss": 1.397, + "theoretical_loss": 3.561400879391763, + "tokens_seen": 1421213696 + }, + { + "epoch": 0.43, + "learning_rate": 0.0004601323971915747, + "loss": 1.424, + "theoretical_loss": 3.5613863667713375, + "tokens_seen": 1421279232 + }, + { + "epoch": 0.43, + "learning_rate": 0.00046011634904714145, + "loss": 1.2997, + "theoretical_loss": 3.561371855007444, + "tokens_seen": 1421344768 + }, + { + "epoch": 0.43, + "learning_rate": 0.0004601003009027081, + "loss": 1.4336, + "theoretical_loss": 3.561357344099992, + "tokens_seen": 1421410304 + }, + { + "epoch": 0.43, + "learning_rate": 0.00046008425275827485, + "loss": 1.4794, + "theoretical_loss": 3.561342834048892, + "tokens_seen": 1421475840 + }, + { + "epoch": 0.43, + "learning_rate": 0.0004600682046138415, + "loss": 1.3693, + "theoretical_loss": 3.5613283248540535, + "tokens_seen": 1421541376 + }, + { + "epoch": 0.43, + "learning_rate": 0.00046005215646940825, + "loss": 1.2786, + "theoretical_loss": 3.561313816515387, + "tokens_seen": 1421606912 + }, + { + "epoch": 0.43, + "learning_rate": 0.000460036108324975, + "loss": 1.3165, + "theoretical_loss": 3.5612993090328025, + "tokens_seen": 1421672448 + }, + { + "epoch": 0.43, + "learning_rate": 0.00046002006018054165, + "loss": 1.3515, + "theoretical_loss": 3.5612848024062096, + "tokens_seen": 1421737984 + }, + { + "epoch": 0.43, + "learning_rate": 0.0004600040120361084, + "loss": 1.38, + "theoretical_loss": 3.5612702966355183, + "tokens_seen": 1421803520 + }, + { + "epoch": 0.43, + "learning_rate": 0.00045998796389167505, + "loss": 1.4912, + "theoretical_loss": 3.5612557917206393, + "tokens_seen": 1421869056 + }, + { + "epoch": 0.43, + "learning_rate": 0.0004599719157472418, + "loss": 1.4542, + "theoretical_loss": 3.561241287661482, + "tokens_seen": 1421934592 + }, + { + "epoch": 0.43, + "learning_rate": 0.00045995586760280845, + "loss": 1.3365, + "theoretical_loss": 3.561226784457957, + "tokens_seen": 1422000128 + }, + { + "epoch": 0.43, + "learning_rate": 0.0004599398194583752, + "loss": 1.39, + "theoretical_loss": 3.561212282109974, + "tokens_seen": 1422065664 + }, + { + "epoch": 0.43, + "objective/train/docs_used": 519513, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.5345633029937744, + "objective/train/theoretical_loss": 3.5611977806174426, + "objective/train/tokens_used": 1442591200, + "theoretical_loss": 3.5611977806174426, + "tokens_seen": 1422131200 + }, + { + "epoch": 0.43, + "learning_rate": 0.00045992377131394186, + "loss": 1.4723, + "theoretical_loss": 3.5611977806174426, + "tokens_seen": 1422131200 + }, + { + "epoch": 0.43, + "learning_rate": 0.0004599077231695086, + "loss": 1.3907, + "theoretical_loss": 3.5611832799802743, + "tokens_seen": 1422196736 + }, + { + "epoch": 0.43, + "learning_rate": 0.00045989167502507526, + "loss": 1.3909, + "theoretical_loss": 3.5611687801983782, + "tokens_seen": 1422262272 + }, + { + "epoch": 0.43, + "learning_rate": 0.000459875626880642, + "loss": 1.3199, + "theoretical_loss": 3.561154281271665, + "tokens_seen": 1422327808 + }, + { + "epoch": 0.43, + "learning_rate": 0.00045985957873620866, + "loss": 1.4745, + "theoretical_loss": 3.561139783200044, + "tokens_seen": 1422393344 + }, + { + "epoch": 0.43, + "learning_rate": 0.00045984353059177536, + "loss": 1.3564, + "theoretical_loss": 3.5611252859834264, + "tokens_seen": 1422458880 + }, + { + "epoch": 0.43, + "learning_rate": 0.00045982748244734206, + "loss": 1.3352, + "theoretical_loss": 3.561110789621722, + "tokens_seen": 1422524416 + }, + { + "epoch": 0.43, + "learning_rate": 0.00045981143430290876, + "loss": 1.3318, + "theoretical_loss": 3.561096294114841, + "tokens_seen": 1422589952 + }, + { + "epoch": 0.43, + "learning_rate": 0.00045979538615847546, + "loss": 1.4373, + "theoretical_loss": 3.561081799462693, + "tokens_seen": 1422655488 + }, + { + "epoch": 0.43, + "learning_rate": 0.00045977933801404216, + "loss": 1.4576, + "theoretical_loss": 3.5610673056651896, + "tokens_seen": 1422721024 + }, + { + "epoch": 0.43, + "learning_rate": 0.00045976328986960886, + "loss": 1.3044, + "theoretical_loss": 3.56105281272224, + "tokens_seen": 1422786560 + }, + { + "epoch": 0.43, + "learning_rate": 0.00045974724172517556, + "loss": 1.3074, + "theoretical_loss": 3.561038320633754, + "tokens_seen": 1422852096 + }, + { + "epoch": 0.43, + "learning_rate": 0.00045973119358074226, + "loss": 1.3344, + "theoretical_loss": 3.5610238293996437, + "tokens_seen": 1422917632 + }, + { + "epoch": 0.43, + "learning_rate": 0.00045971514543630896, + "loss": 1.3561, + "theoretical_loss": 3.561009339019818, + "tokens_seen": 1422983168 + }, + { + "epoch": 0.43, + "learning_rate": 0.0004596990972918756, + "loss": 1.3451, + "theoretical_loss": 3.560994849494187, + "tokens_seen": 1423048704 + }, + { + "epoch": 0.43, + "learning_rate": 0.00045968304914744236, + "loss": 1.3848, + "theoretical_loss": 3.5609803608226613, + "tokens_seen": 1423114240 + }, + { + "epoch": 0.43, + "learning_rate": 0.000459667001003009, + "loss": 1.3668, + "theoretical_loss": 3.560965873005152, + "tokens_seen": 1423179776 + }, + { + "epoch": 0.43, + "learning_rate": 0.00045965095285857577, + "loss": 1.4188, + "theoretical_loss": 3.5609513860415687, + "tokens_seen": 1423245312 + }, + { + "epoch": 0.43, + "learning_rate": 0.0004596349047141424, + "loss": 1.2884, + "theoretical_loss": 3.5609368999318214, + "tokens_seen": 1423310848 + }, + { + "epoch": 0.43, + "learning_rate": 0.00045961885656970917, + "loss": 1.4265, + "theoretical_loss": 3.5609224146758214, + "tokens_seen": 1423376384 + }, + { + "epoch": 0.43, + "learning_rate": 0.0004596028084252758, + "loss": 1.327, + "theoretical_loss": 3.560907930273479, + "tokens_seen": 1423441920 + }, + { + "epoch": 0.43, + "learning_rate": 0.00045958676028084257, + "loss": 1.3711, + "theoretical_loss": 3.5608934467247035, + "tokens_seen": 1423507456 + }, + { + "epoch": 0.43, + "learning_rate": 0.0004595707121364092, + "loss": 1.3111, + "theoretical_loss": 3.5608789640294063, + "tokens_seen": 1423572992 + }, + { + "epoch": 0.43, + "learning_rate": 0.00045955466399197597, + "loss": 1.4914, + "theoretical_loss": 3.5608644821874975, + "tokens_seen": 1423638528 + }, + { + "epoch": 0.43, + "learning_rate": 0.0004595386158475426, + "loss": 1.3018, + "theoretical_loss": 3.5608500011988875, + "tokens_seen": 1423704064 + }, + { + "epoch": 0.43, + "learning_rate": 0.00045952256770310937, + "loss": 1.3997, + "theoretical_loss": 3.5608355210634874, + "tokens_seen": 1423769600 + }, + { + "epoch": 0.43, + "learning_rate": 0.000459506519558676, + "loss": 1.2401, + "theoretical_loss": 3.5608210417812067, + "tokens_seen": 1423835136 + }, + { + "epoch": 0.43, + "learning_rate": 0.00045949047141424277, + "loss": 1.3886, + "theoretical_loss": 3.560806563351956, + "tokens_seen": 1423900672 + }, + { + "epoch": 0.43, + "learning_rate": 0.0004594744232698094, + "loss": 1.4315, + "theoretical_loss": 3.5607920857756463, + "tokens_seen": 1423966208 + }, + { + "epoch": 0.43, + "learning_rate": 0.0004594583751253762, + "loss": 1.483, + "theoretical_loss": 3.5607776090521877, + "tokens_seen": 1424031744 + }, + { + "epoch": 0.43, + "learning_rate": 0.0004594423269809428, + "loss": 1.3951, + "theoretical_loss": 3.5607631331814913, + "tokens_seen": 1424097280 + }, + { + "epoch": 0.43, + "learning_rate": 0.0004594262788365096, + "loss": 1.3384, + "theoretical_loss": 3.5607486581634666, + "tokens_seen": 1424162816 + }, + { + "epoch": 0.43, + "learning_rate": 0.0004594102306920762, + "loss": 1.3406, + "theoretical_loss": 3.560734183998025, + "tokens_seen": 1424228352 + }, + { + "epoch": 0.43, + "learning_rate": 0.000459394182547643, + "loss": 1.3339, + "theoretical_loss": 3.5607197106850768, + "tokens_seen": 1424293888 + }, + { + "epoch": 0.43, + "learning_rate": 0.0004593781344032096, + "loss": 1.4594, + "theoretical_loss": 3.560705238224532, + "tokens_seen": 1424359424 + }, + { + "epoch": 0.43, + "learning_rate": 0.0004593620862587764, + "loss": 1.4348, + "theoretical_loss": 3.560690766616303, + "tokens_seen": 1424424960 + }, + { + "epoch": 0.43, + "learning_rate": 0.000459346038114343, + "loss": 1.4856, + "theoretical_loss": 3.560676295860298, + "tokens_seen": 1424490496 + }, + { + "epoch": 0.43, + "learning_rate": 0.0004593299899699098, + "loss": 1.3907, + "theoretical_loss": 3.560661825956429, + "tokens_seen": 1424556032 + }, + { + "epoch": 0.43, + "learning_rate": 0.0004593139418254764, + "loss": 1.3841, + "theoretical_loss": 3.5606473569046067, + "tokens_seen": 1424621568 + }, + { + "epoch": 0.43, + "learning_rate": 0.0004592978936810432, + "loss": 1.3255, + "theoretical_loss": 3.560632888704741, + "tokens_seen": 1424687104 + }, + { + "epoch": 0.43, + "learning_rate": 0.0004592818455366098, + "loss": 1.3435, + "theoretical_loss": 3.560618421356743, + "tokens_seen": 1424752640 + }, + { + "epoch": 0.43, + "learning_rate": 0.0004592657973921766, + "loss": 1.3435, + "theoretical_loss": 3.5606039548605235, + "tokens_seen": 1424818176 + }, + { + "epoch": 0.43, + "learning_rate": 0.00045924974924774323, + "loss": 1.3668, + "theoretical_loss": 3.560589489215993, + "tokens_seen": 1424883712 + }, + { + "epoch": 0.43, + "learning_rate": 0.00045923370110331, + "loss": 1.3355, + "theoretical_loss": 3.560575024423062, + "tokens_seen": 1424949248 + }, + { + "epoch": 0.43, + "learning_rate": 0.00045921765295887663, + "loss": 1.4728, + "theoretical_loss": 3.5605605604816413, + "tokens_seen": 1425014784 + }, + { + "epoch": 0.43, + "learning_rate": 0.0004592016048144434, + "loss": 1.4011, + "theoretical_loss": 3.560546097391642, + "tokens_seen": 1425080320 + }, + { + "epoch": 0.43, + "learning_rate": 0.00045918555667001003, + "loss": 1.3478, + "theoretical_loss": 3.560531635152974, + "tokens_seen": 1425145856 + }, + { + "epoch": 0.43, + "learning_rate": 0.0004591695085255768, + "loss": 1.409, + "theoretical_loss": 3.5605171737655485, + "tokens_seen": 1425211392 + }, + { + "epoch": 0.43, + "learning_rate": 0.00045915346038114343, + "loss": 1.3793, + "theoretical_loss": 3.560502713229277, + "tokens_seen": 1425276928 + }, + { + "epoch": 0.43, + "learning_rate": 0.0004591374122367102, + "loss": 1.4893, + "theoretical_loss": 3.5604882535440696, + "tokens_seen": 1425342464 + }, + { + "epoch": 0.43, + "objective/train/docs_used": 520731, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.2746115922927856, + "objective/train/theoretical_loss": 3.5604737947098366, + "objective/train/tokens_used": 1445868000, + "theoretical_loss": 3.5604737947098366, + "tokens_seen": 1425408000 + }, + { + "epoch": 0.43, + "learning_rate": 0.00045912136409227683, + "loss": 1.2761, + "theoretical_loss": 3.5604737947098366, + "tokens_seen": 1425408000 + }, + { + "epoch": 0.43, + "learning_rate": 0.0004591053159478436, + "loss": 1.3671, + "theoretical_loss": 3.560459336726489, + "tokens_seen": 1425473536 + }, + { + "epoch": 0.43, + "learning_rate": 0.00045908926780341023, + "loss": 1.4526, + "theoretical_loss": 3.5604448795939385, + "tokens_seen": 1425539072 + }, + { + "epoch": 0.43, + "learning_rate": 0.00045907321965897693, + "loss": 1.3714, + "theoretical_loss": 3.560430423312095, + "tokens_seen": 1425604608 + }, + { + "epoch": 0.43, + "learning_rate": 0.00045905717151454364, + "loss": 1.3921, + "theoretical_loss": 3.56041596788087, + "tokens_seen": 1425670144 + }, + { + "epoch": 0.43, + "learning_rate": 0.00045904112337011034, + "loss": 1.3054, + "theoretical_loss": 3.5604015133001736, + "tokens_seen": 1425735680 + }, + { + "epoch": 0.43, + "learning_rate": 0.00045902507522567704, + "loss": 1.4481, + "theoretical_loss": 3.5603870595699174, + "tokens_seen": 1425801216 + }, + { + "epoch": 0.43, + "learning_rate": 0.00045900902708124374, + "loss": 1.353, + "theoretical_loss": 3.560372606690011, + "tokens_seen": 1425866752 + }, + { + "epoch": 0.43, + "learning_rate": 0.00045899297893681044, + "loss": 1.3607, + "theoretical_loss": 3.560358154660367, + "tokens_seen": 1425932288 + }, + { + "epoch": 0.43, + "learning_rate": 0.00045897693079237714, + "loss": 1.4241, + "theoretical_loss": 3.560343703480896, + "tokens_seen": 1425997824 + }, + { + "epoch": 0.43, + "learning_rate": 0.00045896088264794384, + "loss": 1.3007, + "theoretical_loss": 3.560329253151508, + "tokens_seen": 1426063360 + }, + { + "epoch": 0.43, + "learning_rate": 0.00045894483450351054, + "loss": 1.4279, + "theoretical_loss": 3.560314803672114, + "tokens_seen": 1426128896 + }, + { + "epoch": 0.43, + "learning_rate": 0.0004589287863590772, + "loss": 1.3116, + "theoretical_loss": 3.5603003550426258, + "tokens_seen": 1426194432 + }, + { + "epoch": 0.43, + "learning_rate": 0.00045891273821464394, + "loss": 1.3054, + "theoretical_loss": 3.560285907262953, + "tokens_seen": 1426259968 + }, + { + "epoch": 0.43, + "learning_rate": 0.0004588966900702106, + "loss": 1.2959, + "theoretical_loss": 3.5602714603330083, + "tokens_seen": 1426325504 + }, + { + "epoch": 0.43, + "learning_rate": 0.00045888064192577734, + "loss": 1.3835, + "theoretical_loss": 3.560257014252702, + "tokens_seen": 1426391040 + }, + { + "epoch": 0.43, + "learning_rate": 0.000458864593781344, + "loss": 1.3294, + "theoretical_loss": 3.560242569021945, + "tokens_seen": 1426456576 + }, + { + "epoch": 0.43, + "learning_rate": 0.00045884854563691074, + "loss": 1.4087, + "theoretical_loss": 3.5602281246406475, + "tokens_seen": 1426522112 + }, + { + "epoch": 0.43, + "learning_rate": 0.0004588324974924775, + "loss": 1.4082, + "theoretical_loss": 3.5602136811087215, + "tokens_seen": 1426587648 + }, + { + "epoch": 0.43, + "learning_rate": 0.00045881644934804415, + "loss": 1.2761, + "theoretical_loss": 3.560199238426078, + "tokens_seen": 1426653184 + }, + { + "epoch": 0.43, + "learning_rate": 0.0004588004012036109, + "loss": 1.3675, + "theoretical_loss": 3.560184796592628, + "tokens_seen": 1426718720 + }, + { + "epoch": 0.43, + "learning_rate": 0.00045878435305917755, + "loss": 1.4335, + "theoretical_loss": 3.5601703556082818, + "tokens_seen": 1426784256 + }, + { + "epoch": 0.43, + "learning_rate": 0.0004587683049147443, + "loss": 1.428, + "theoretical_loss": 3.5601559154729516, + "tokens_seen": 1426849792 + }, + { + "epoch": 0.43, + "learning_rate": 0.00045875225677031095, + "loss": 1.3793, + "theoretical_loss": 3.5601414761865477, + "tokens_seen": 1426915328 + }, + { + "epoch": 0.43, + "learning_rate": 0.0004587362086258777, + "loss": 1.328, + "theoretical_loss": 3.560127037748982, + "tokens_seen": 1426980864 + }, + { + "epoch": 0.43, + "learning_rate": 0.00045872016048144435, + "loss": 1.4262, + "theoretical_loss": 3.560112600160165, + "tokens_seen": 1427046400 + }, + { + "epoch": 0.43, + "learning_rate": 0.0004587041123370111, + "loss": 1.4289, + "theoretical_loss": 3.5600981634200073, + "tokens_seen": 1427111936 + }, + { + "epoch": 0.43, + "learning_rate": 0.00045868806419257775, + "loss": 1.3997, + "theoretical_loss": 3.5600837275284216, + "tokens_seen": 1427177472 + }, + { + "epoch": 0.43, + "learning_rate": 0.0004586720160481445, + "loss": 1.4435, + "theoretical_loss": 3.5600692924853177, + "tokens_seen": 1427243008 + }, + { + "epoch": 0.43, + "learning_rate": 0.00045865596790371115, + "loss": 1.3318, + "theoretical_loss": 3.5600548582906075, + "tokens_seen": 1427308544 + }, + { + "epoch": 0.43, + "learning_rate": 0.0004586399197592779, + "loss": 1.4579, + "theoretical_loss": 3.5600404249442015, + "tokens_seen": 1427374080 + }, + { + "epoch": 0.43, + "learning_rate": 0.00045862387161484455, + "loss": 1.4986, + "theoretical_loss": 3.5600259924460116, + "tokens_seen": 1427439616 + }, + { + "epoch": 0.43, + "learning_rate": 0.0004586078234704113, + "loss": 1.5426, + "theoretical_loss": 3.5600115607959486, + "tokens_seen": 1427505152 + }, + { + "epoch": 0.43, + "learning_rate": 0.00045859177532597795, + "loss": 1.5202, + "theoretical_loss": 3.559997129993924, + "tokens_seen": 1427570688 + }, + { + "epoch": 0.43, + "learning_rate": 0.0004585757271815447, + "loss": 1.3326, + "theoretical_loss": 3.559982700039849, + "tokens_seen": 1427636224 + }, + { + "epoch": 0.43, + "learning_rate": 0.00045855967903711136, + "loss": 1.3428, + "theoretical_loss": 3.559968270933634, + "tokens_seen": 1427701760 + }, + { + "epoch": 0.43, + "learning_rate": 0.0004585436308926781, + "loss": 1.3411, + "theoretical_loss": 3.559953842675192, + "tokens_seen": 1427767296 + }, + { + "epoch": 0.43, + "learning_rate": 0.00045852758274824476, + "loss": 1.4182, + "theoretical_loss": 3.5599394152644326, + "tokens_seen": 1427832832 + }, + { + "epoch": 0.43, + "learning_rate": 0.0004585115346038115, + "loss": 1.361, + "theoretical_loss": 3.559924988701268, + "tokens_seen": 1427898368 + }, + { + "epoch": 0.43, + "learning_rate": 0.00045849548645937816, + "loss": 1.4019, + "theoretical_loss": 3.5599105629856096, + "tokens_seen": 1427963904 + }, + { + "epoch": 0.43, + "learning_rate": 0.0004584794383149449, + "loss": 1.2918, + "theoretical_loss": 3.5598961381173684, + "tokens_seen": 1428029440 + }, + { + "epoch": 0.43, + "learning_rate": 0.00045846339017051156, + "loss": 1.4503, + "theoretical_loss": 3.5598817140964556, + "tokens_seen": 1428094976 + }, + { + "epoch": 0.43, + "learning_rate": 0.00045844734202607826, + "loss": 1.3512, + "theoretical_loss": 3.5598672909227824, + "tokens_seen": 1428160512 + }, + { + "epoch": 0.43, + "learning_rate": 0.00045843129388164496, + "loss": 1.5468, + "theoretical_loss": 3.559852868596261, + "tokens_seen": 1428226048 + }, + { + "epoch": 0.43, + "learning_rate": 0.00045841524573721166, + "loss": 1.374, + "theoretical_loss": 3.5598384471168014, + "tokens_seen": 1428291584 + }, + { + "epoch": 0.43, + "learning_rate": 0.00045839919759277836, + "loss": 1.3787, + "theoretical_loss": 3.559824026484317, + "tokens_seen": 1428357120 + }, + { + "epoch": 0.43, + "learning_rate": 0.00045838314944834506, + "loss": 1.346, + "theoretical_loss": 3.5598096066987166, + "tokens_seen": 1428422656 + }, + { + "epoch": 0.43, + "learning_rate": 0.00045836710130391176, + "loss": 1.3844, + "theoretical_loss": 3.5597951877599137, + "tokens_seen": 1428488192 + }, + { + "epoch": 0.43, + "learning_rate": 0.00045835105315947846, + "loss": 1.4536, + "theoretical_loss": 3.559780769667819, + "tokens_seen": 1428553728 + }, + { + "epoch": 0.43, + "learning_rate": 0.00045833500501504516, + "loss": 1.4797, + "theoretical_loss": 3.5597663524223444, + "tokens_seen": 1428619264 + }, + { + "epoch": 0.43, + "objective/train/docs_used": 521864, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.391134262084961, + "objective/train/theoretical_loss": 3.559751936023401, + "objective/train/tokens_used": 1449144800, + "theoretical_loss": 3.559751936023401, + "tokens_seen": 1428684800 + }, + { + "epoch": 0.43, + "learning_rate": 0.00045831895687061186, + "loss": 1.4051, + "theoretical_loss": 3.559751936023401, + "tokens_seen": 1428684800 + }, + { + "epoch": 0.43, + "learning_rate": 0.0004583029087261785, + "loss": 1.4337, + "theoretical_loss": 3.5597375204709, + "tokens_seen": 1428750336 + }, + { + "epoch": 0.43, + "learning_rate": 0.00045828686058174527, + "loss": 1.4153, + "theoretical_loss": 3.5597231057647525, + "tokens_seen": 1428815872 + }, + { + "epoch": 0.43, + "learning_rate": 0.0004582708124373119, + "loss": 1.4928, + "theoretical_loss": 3.5597086919048713, + "tokens_seen": 1428881408 + }, + { + "epoch": 0.43, + "learning_rate": 0.00045825476429287867, + "loss": 1.5355, + "theoretical_loss": 3.5596942788911665, + "tokens_seen": 1428946944 + }, + { + "epoch": 0.43, + "learning_rate": 0.0004582387161484453, + "loss": 1.4736, + "theoretical_loss": 3.5596798667235507, + "tokens_seen": 1429012480 + }, + { + "epoch": 0.43, + "learning_rate": 0.00045822266800401207, + "loss": 1.3185, + "theoretical_loss": 3.5596654554019356, + "tokens_seen": 1429078016 + }, + { + "epoch": 0.43, + "learning_rate": 0.0004582066198595787, + "loss": 1.4677, + "theoretical_loss": 3.5596510449262313, + "tokens_seen": 1429143552 + }, + { + "epoch": 0.43, + "learning_rate": 0.00045819057171514547, + "loss": 1.3812, + "theoretical_loss": 3.559636635296351, + "tokens_seen": 1429209088 + }, + { + "epoch": 0.43, + "learning_rate": 0.0004581745235707121, + "loss": 1.3547, + "theoretical_loss": 3.559622226512205, + "tokens_seen": 1429274624 + }, + { + "epoch": 0.43, + "learning_rate": 0.00045815847542627887, + "loss": 1.277, + "theoretical_loss": 3.5596078185737055, + "tokens_seen": 1429340160 + }, + { + "epoch": 0.43, + "learning_rate": 0.0004581424272818455, + "loss": 1.3981, + "theoretical_loss": 3.559593411480764, + "tokens_seen": 1429405696 + }, + { + "epoch": 0.43, + "learning_rate": 0.00045812637913741227, + "loss": 1.442, + "theoretical_loss": 3.5595790052332923, + "tokens_seen": 1429471232 + }, + { + "epoch": 0.43, + "learning_rate": 0.0004581103309929789, + "loss": 1.4299, + "theoretical_loss": 3.5595645998312015, + "tokens_seen": 1429536768 + }, + { + "epoch": 0.43, + "learning_rate": 0.0004580942828485457, + "loss": 1.3998, + "theoretical_loss": 3.559550195274404, + "tokens_seen": 1429602304 + }, + { + "epoch": 0.43, + "learning_rate": 0.0004580782347041123, + "loss": 1.3684, + "theoretical_loss": 3.5595357915628107, + "tokens_seen": 1429667840 + }, + { + "epoch": 0.43, + "learning_rate": 0.0004580621865596791, + "loss": 1.3381, + "theoretical_loss": 3.559521388696334, + "tokens_seen": 1429733376 + }, + { + "epoch": 0.43, + "learning_rate": 0.0004580461384152457, + "loss": 1.432, + "theoretical_loss": 3.5595069866748847, + "tokens_seen": 1429798912 + }, + { + "epoch": 0.43, + "learning_rate": 0.0004580300902708125, + "loss": 1.358, + "theoretical_loss": 3.559492585498375, + "tokens_seen": 1429864448 + }, + { + "epoch": 0.43, + "learning_rate": 0.0004580140421263791, + "loss": 1.4289, + "theoretical_loss": 3.559478185166717, + "tokens_seen": 1429929984 + }, + { + "epoch": 0.43, + "learning_rate": 0.0004579979939819459, + "loss": 1.3515, + "theoretical_loss": 3.5594637856798217, + "tokens_seen": 1429995520 + }, + { + "epoch": 0.43, + "learning_rate": 0.0004579819458375125, + "loss": 1.2848, + "theoretical_loss": 3.5594493870376014, + "tokens_seen": 1430061056 + }, + { + "epoch": 0.43, + "learning_rate": 0.0004579658976930793, + "loss": 1.4186, + "theoretical_loss": 3.5594349892399673, + "tokens_seen": 1430126592 + }, + { + "epoch": 0.43, + "learning_rate": 0.0004579498495486459, + "loss": 1.4314, + "theoretical_loss": 3.5594205922868314, + "tokens_seen": 1430192128 + }, + { + "epoch": 0.43, + "learning_rate": 0.0004579338014042127, + "loss": 1.3573, + "theoretical_loss": 3.559406196178106, + "tokens_seen": 1430257664 + }, + { + "epoch": 0.43, + "learning_rate": 0.0004579177532597793, + "loss": 1.3556, + "theoretical_loss": 3.5593918009137018, + "tokens_seen": 1430323200 + }, + { + "epoch": 0.43, + "learning_rate": 0.0004579017051153461, + "loss": 1.3125, + "theoretical_loss": 3.5593774064935317, + "tokens_seen": 1430388736 + }, + { + "epoch": 0.43, + "learning_rate": 0.00045788565697091273, + "loss": 1.4032, + "theoretical_loss": 3.5593630129175065, + "tokens_seen": 1430454272 + }, + { + "epoch": 0.43, + "learning_rate": 0.0004578696088264795, + "loss": 1.4536, + "theoretical_loss": 3.559348620185539, + "tokens_seen": 1430519808 + }, + { + "epoch": 0.43, + "learning_rate": 0.00045785356068204613, + "loss": 1.3485, + "theoretical_loss": 3.55933422829754, + "tokens_seen": 1430585344 + }, + { + "epoch": 0.43, + "learning_rate": 0.0004578375125376129, + "loss": 1.4971, + "theoretical_loss": 3.5593198372534225, + "tokens_seen": 1430650880 + }, + { + "epoch": 0.43, + "learning_rate": 0.00045782146439317953, + "loss": 1.4467, + "theoretical_loss": 3.5593054470530974, + "tokens_seen": 1430716416 + }, + { + "epoch": 0.43, + "learning_rate": 0.0004578054162487463, + "loss": 1.3173, + "theoretical_loss": 3.559291057696477, + "tokens_seen": 1430781952 + }, + { + "epoch": 0.43, + "learning_rate": 0.00045778936810431293, + "loss": 1.2833, + "theoretical_loss": 3.559276669183473, + "tokens_seen": 1430847488 + }, + { + "epoch": 0.43, + "learning_rate": 0.0004577733199598797, + "loss": 1.4155, + "theoretical_loss": 3.559262281513998, + "tokens_seen": 1430913024 + }, + { + "epoch": 0.43, + "learning_rate": 0.00045775727181544633, + "loss": 1.3942, + "theoretical_loss": 3.559247894687963, + "tokens_seen": 1430978560 + }, + { + "epoch": 0.43, + "learning_rate": 0.0004577412236710131, + "loss": 1.4593, + "theoretical_loss": 3.5592335087052804, + "tokens_seen": 1431044096 + }, + { + "epoch": 0.43, + "learning_rate": 0.00045772517552657973, + "loss": 1.3732, + "theoretical_loss": 3.559219123565862, + "tokens_seen": 1431109632 + }, + { + "epoch": 0.43, + "learning_rate": 0.0004577091273821465, + "loss": 1.4513, + "theoretical_loss": 3.55920473926962, + "tokens_seen": 1431175168 + }, + { + "epoch": 0.43, + "learning_rate": 0.00045769307923771314, + "loss": 1.4059, + "theoretical_loss": 3.559190355816466, + "tokens_seen": 1431240704 + }, + { + "epoch": 0.43, + "learning_rate": 0.00045767703109327984, + "loss": 1.4097, + "theoretical_loss": 3.5591759732063117, + "tokens_seen": 1431306240 + }, + { + "epoch": 0.43, + "learning_rate": 0.00045766098294884654, + "loss": 1.3866, + "theoretical_loss": 3.5591615914390706, + "tokens_seen": 1431371776 + }, + { + "epoch": 0.43, + "learning_rate": 0.00045764493480441324, + "loss": 1.2857, + "theoretical_loss": 3.559147210514653, + "tokens_seen": 1431437312 + }, + { + "epoch": 0.43, + "learning_rate": 0.00045762888665998, + "loss": 1.3933, + "theoretical_loss": 3.5591328304329712, + "tokens_seen": 1431502848 + }, + { + "epoch": 0.43, + "learning_rate": 0.00045761283851554664, + "loss": 1.3632, + "theoretical_loss": 3.559118451193938, + "tokens_seen": 1431568384 + }, + { + "epoch": 0.43, + "learning_rate": 0.0004575967903711134, + "loss": 1.4366, + "theoretical_loss": 3.5591040727974654, + "tokens_seen": 1431633920 + }, + { + "epoch": 0.43, + "learning_rate": 0.00045758074222668004, + "loss": 1.466, + "theoretical_loss": 3.559089695243465, + "tokens_seen": 1431699456 + }, + { + "epoch": 0.43, + "learning_rate": 0.0004575646940822468, + "loss": 1.4223, + "theoretical_loss": 3.559075318531849, + "tokens_seen": 1431764992 + }, + { + "epoch": 0.43, + "learning_rate": 0.00045754864593781344, + "loss": 1.4259, + "theoretical_loss": 3.5590609426625295, + "tokens_seen": 1431830528 + }, + { + "epoch": 0.43, + "learning_rate": 0.0004575325977933802, + "loss": 1.3247, + "theoretical_loss": 3.5590465676354186, + "tokens_seen": 1431896064 + }, + { + "epoch": 0.43, + "objective/train/docs_used": 522998, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.26812744140625, + "objective/train/theoretical_loss": 3.5590321934504283, + "objective/train/tokens_used": 1452421600, + "theoretical_loss": 3.5590321934504283, + "tokens_seen": 1431961600 + }, + { + "epoch": 0.43, + "learning_rate": 0.00045751654964894684, + "loss": 1.3695, + "theoretical_loss": 3.5590321934504283, + "tokens_seen": 1431961600 + }, + { + "epoch": 0.43, + "learning_rate": 0.0004575005015045136, + "loss": 1.3377, + "theoretical_loss": 3.559017820107471, + "tokens_seen": 1432027136 + }, + { + "epoch": 0.43, + "learning_rate": 0.00045748445336008024, + "loss": 1.4503, + "theoretical_loss": 3.5590034476064587, + "tokens_seen": 1432092672 + }, + { + "epoch": 0.43, + "learning_rate": 0.000457468405215647, + "loss": 1.457, + "theoretical_loss": 3.558989075947303, + "tokens_seen": 1432158208 + }, + { + "epoch": 0.43, + "learning_rate": 0.00045745235707121364, + "loss": 1.3325, + "theoretical_loss": 3.558974705129917, + "tokens_seen": 1432223744 + }, + { + "epoch": 0.43, + "learning_rate": 0.0004574363089267804, + "loss": 1.3177, + "theoretical_loss": 3.558960335154213, + "tokens_seen": 1432289280 + }, + { + "epoch": 0.43, + "learning_rate": 0.00045742026078234705, + "loss": 1.3698, + "theoretical_loss": 3.5589459660201026, + "tokens_seen": 1432354816 + }, + { + "epoch": 0.43, + "learning_rate": 0.0004574042126379138, + "loss": 1.3385, + "theoretical_loss": 3.558931597727498, + "tokens_seen": 1432420352 + }, + { + "epoch": 0.43, + "learning_rate": 0.00045738816449348045, + "loss": 1.3767, + "theoretical_loss": 3.558917230276311, + "tokens_seen": 1432485888 + }, + { + "epoch": 0.43, + "learning_rate": 0.0004573721163490472, + "loss": 1.3853, + "theoretical_loss": 3.5589028636664546, + "tokens_seen": 1432551424 + }, + { + "epoch": 0.43, + "learning_rate": 0.00045735606820461385, + "loss": 1.447, + "theoretical_loss": 3.558888497897841, + "tokens_seen": 1432616960 + }, + { + "epoch": 0.43, + "learning_rate": 0.0004573400200601806, + "loss": 1.4281, + "theoretical_loss": 3.5588741329703826, + "tokens_seen": 1432682496 + }, + { + "epoch": 0.43, + "learning_rate": 0.00045732397191574725, + "loss": 1.4732, + "theoretical_loss": 3.558859768883991, + "tokens_seen": 1432748032 + }, + { + "epoch": 0.43, + "learning_rate": 0.000457307923771314, + "loss": 1.4093, + "theoretical_loss": 3.5588454056385785, + "tokens_seen": 1432813568 + }, + { + "epoch": 0.43, + "learning_rate": 0.00045729187562688065, + "loss": 1.3127, + "theoretical_loss": 3.5588310432340577, + "tokens_seen": 1432879104 + }, + { + "epoch": 0.43, + "learning_rate": 0.0004572758274824474, + "loss": 1.2346, + "theoretical_loss": 3.5588166816703417, + "tokens_seen": 1432944640 + }, + { + "epoch": 0.43, + "learning_rate": 0.00045725977933801405, + "loss": 1.3342, + "theoretical_loss": 3.5588023209473416, + "tokens_seen": 1433010176 + }, + { + "epoch": 0.43, + "learning_rate": 0.0004572437311935808, + "loss": 1.3397, + "theoretical_loss": 3.5587879610649704, + "tokens_seen": 1433075712 + }, + { + "epoch": 0.43, + "learning_rate": 0.00045722768304914745, + "loss": 1.3382, + "theoretical_loss": 3.55877360202314, + "tokens_seen": 1433141248 + }, + { + "epoch": 0.43, + "learning_rate": 0.0004572116349047142, + "loss": 1.3108, + "theoretical_loss": 3.558759243821763, + "tokens_seen": 1433206784 + }, + { + "epoch": 0.43, + "learning_rate": 0.00045719558676028085, + "loss": 1.4887, + "theoretical_loss": 3.558744886460752, + "tokens_seen": 1433272320 + }, + { + "epoch": 0.43, + "learning_rate": 0.0004571795386158476, + "loss": 1.301, + "theoretical_loss": 3.558730529940019, + "tokens_seen": 1433337856 + }, + { + "epoch": 0.43, + "learning_rate": 0.00045716349047141426, + "loss": 1.2346, + "theoretical_loss": 3.5587161742594766, + "tokens_seen": 1433403392 + }, + { + "epoch": 0.43, + "learning_rate": 0.000457147442326981, + "loss": 1.4029, + "theoretical_loss": 3.5587018194190376, + "tokens_seen": 1433468928 + }, + { + "epoch": 0.43, + "learning_rate": 0.00045713139418254766, + "loss": 1.2224, + "theoretical_loss": 3.558687465418614, + "tokens_seen": 1433534464 + }, + { + "epoch": 0.43, + "learning_rate": 0.0004571153460381144, + "loss": 1.3375, + "theoretical_loss": 3.558673112258118, + "tokens_seen": 1433600000 + }, + { + "epoch": 0.43, + "learning_rate": 0.00045709929789368106, + "loss": 1.4256, + "theoretical_loss": 3.5586587599374626, + "tokens_seen": 1433665536 + }, + { + "epoch": 0.43, + "learning_rate": 0.0004570832497492478, + "loss": 1.4378, + "theoretical_loss": 3.5586444084565603, + "tokens_seen": 1433731072 + }, + { + "epoch": 0.43, + "learning_rate": 0.00045706720160481446, + "loss": 1.4375, + "theoretical_loss": 3.5586300578153227, + "tokens_seen": 1433796608 + }, + { + "epoch": 0.43, + "learning_rate": 0.00045705115346038116, + "loss": 1.4706, + "theoretical_loss": 3.5586157080136633, + "tokens_seen": 1433862144 + }, + { + "epoch": 0.43, + "learning_rate": 0.00045703510531594786, + "loss": 1.4301, + "theoretical_loss": 3.558601359051494, + "tokens_seen": 1433927680 + }, + { + "epoch": 0.43, + "learning_rate": 0.00045701905717151456, + "loss": 1.3974, + "theoretical_loss": 3.5585870109287274, + "tokens_seen": 1433993216 + }, + { + "epoch": 0.43, + "learning_rate": 0.00045700300902708126, + "loss": 1.4695, + "theoretical_loss": 3.558572663645277, + "tokens_seen": 1434058752 + }, + { + "epoch": 0.43, + "learning_rate": 0.00045698696088264796, + "loss": 1.3902, + "theoretical_loss": 3.5585583172010535, + "tokens_seen": 1434124288 + }, + { + "epoch": 0.43, + "learning_rate": 0.00045697091273821466, + "loss": 1.3725, + "theoretical_loss": 3.5585439715959715, + "tokens_seen": 1434189824 + }, + { + "epoch": 0.43, + "learning_rate": 0.00045695486459378136, + "loss": 1.4331, + "theoretical_loss": 3.558529626829942, + "tokens_seen": 1434255360 + }, + { + "epoch": 0.43, + "learning_rate": 0.00045693881644934807, + "loss": 1.3808, + "theoretical_loss": 3.558515282902878, + "tokens_seen": 1434320896 + }, + { + "epoch": 0.43, + "learning_rate": 0.00045692276830491477, + "loss": 1.4872, + "theoretical_loss": 3.5585009398146927, + "tokens_seen": 1434386432 + }, + { + "epoch": 0.43, + "learning_rate": 0.0004569067201604814, + "loss": 1.3389, + "theoretical_loss": 3.558486597565298, + "tokens_seen": 1434451968 + }, + { + "epoch": 0.43, + "learning_rate": 0.00045689067201604817, + "loss": 1.4259, + "theoretical_loss": 3.558472256154607, + "tokens_seen": 1434517504 + }, + { + "epoch": 0.43, + "learning_rate": 0.0004568746238716148, + "loss": 1.4113, + "theoretical_loss": 3.558457915582532, + "tokens_seen": 1434583040 + }, + { + "epoch": 0.43, + "learning_rate": 0.00045685857572718157, + "loss": 1.3813, + "theoretical_loss": 3.558443575848986, + "tokens_seen": 1434648576 + }, + { + "epoch": 0.43, + "learning_rate": 0.0004568425275827482, + "loss": 1.3416, + "theoretical_loss": 3.558429236953881, + "tokens_seen": 1434714112 + }, + { + "epoch": 0.43, + "learning_rate": 0.00045682647943831497, + "loss": 1.3686, + "theoretical_loss": 3.558414898897131, + "tokens_seen": 1434779648 + }, + { + "epoch": 0.43, + "learning_rate": 0.0004568104312938816, + "loss": 1.3427, + "theoretical_loss": 3.558400561678647, + "tokens_seen": 1434845184 + }, + { + "epoch": 0.43, + "learning_rate": 0.00045679438314944837, + "loss": 1.3486, + "theoretical_loss": 3.558386225298343, + "tokens_seen": 1434910720 + }, + { + "epoch": 0.43, + "learning_rate": 0.000456778335005015, + "loss": 1.236, + "theoretical_loss": 3.558371889756131, + "tokens_seen": 1434976256 + }, + { + "epoch": 0.43, + "learning_rate": 0.00045676228686058177, + "loss": 1.2353, + "theoretical_loss": 3.5583575550519244, + "tokens_seen": 1435041792 + }, + { + "epoch": 0.43, + "learning_rate": 0.0004567462387161484, + "loss": 1.4127, + "theoretical_loss": 3.558343221185635, + "tokens_seen": 1435107328 + }, + { + "epoch": 0.43, + "learning_rate": 0.0004567301905717152, + "loss": 1.3787, + "theoretical_loss": 3.5583288881571766, + "tokens_seen": 1435172864 + }, + { + "epoch": 0.43, + "objective/train/docs_used": 524207, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.445927619934082, + "objective/train/theoretical_loss": 3.5583145559664615, + "objective/train/tokens_used": 1455698400, + "theoretical_loss": 3.5583145559664615, + "tokens_seen": 1435238400 + }, + { + "epoch": 0.43, + "learning_rate": 0.0004567141424272818, + "loss": 1.4349, + "theoretical_loss": 3.5583145559664615, + "tokens_seen": 1435238400 + }, + { + "epoch": 0.43, + "learning_rate": 0.0004566980942828486, + "loss": 1.4353, + "theoretical_loss": 3.5583002246134026, + "tokens_seen": 1435303936 + }, + { + "epoch": 0.43, + "learning_rate": 0.0004566820461384152, + "loss": 1.4039, + "theoretical_loss": 3.558285894097912, + "tokens_seen": 1435369472 + }, + { + "epoch": 0.43, + "learning_rate": 0.000456665997993982, + "loss": 1.3947, + "theoretical_loss": 3.558271564419903, + "tokens_seen": 1435435008 + }, + { + "epoch": 0.44, + "learning_rate": 0.0004566499498495486, + "loss": 1.3365, + "theoretical_loss": 3.558257235579289, + "tokens_seen": 1435500544 + }, + { + "epoch": 0.44, + "learning_rate": 0.0004566339017051154, + "loss": 1.3767, + "theoretical_loss": 3.5582429075759823, + "tokens_seen": 1435566080 + }, + { + "epoch": 0.44, + "learning_rate": 0.000456617853560682, + "loss": 1.3277, + "theoretical_loss": 3.558228580409896, + "tokens_seen": 1435631616 + }, + { + "epoch": 0.44, + "learning_rate": 0.0004566018054162488, + "loss": 1.3118, + "theoretical_loss": 3.5582142540809425, + "tokens_seen": 1435697152 + }, + { + "epoch": 0.44, + "learning_rate": 0.0004565857572718154, + "loss": 1.429, + "theoretical_loss": 3.558199928589035, + "tokens_seen": 1435762688 + }, + { + "epoch": 0.44, + "learning_rate": 0.0004565697091273822, + "loss": 1.4217, + "theoretical_loss": 3.5581856039340862, + "tokens_seen": 1435828224 + }, + { + "epoch": 0.44, + "learning_rate": 0.0004565536609829488, + "loss": 1.3199, + "theoretical_loss": 3.5581712801160092, + "tokens_seen": 1435893760 + }, + { + "epoch": 0.44, + "learning_rate": 0.0004565376128385156, + "loss": 1.3146, + "theoretical_loss": 3.5581569571347167, + "tokens_seen": 1435959296 + }, + { + "epoch": 0.44, + "learning_rate": 0.00045652156469408223, + "loss": 1.3806, + "theoretical_loss": 3.558142634990122, + "tokens_seen": 1436024832 + }, + { + "epoch": 0.44, + "learning_rate": 0.000456505516549649, + "loss": 1.4101, + "theoretical_loss": 3.558128313682138, + "tokens_seen": 1436090368 + }, + { + "epoch": 0.44, + "learning_rate": 0.00045648946840521563, + "loss": 1.3498, + "theoretical_loss": 3.558113993210677, + "tokens_seen": 1436155904 + }, + { + "epoch": 0.44, + "learning_rate": 0.0004564734202607824, + "loss": 1.4349, + "theoretical_loss": 3.5580996735756525, + "tokens_seen": 1436221440 + }, + { + "epoch": 0.44, + "learning_rate": 0.00045645737211634903, + "loss": 1.5016, + "theoretical_loss": 3.5580853547769777, + "tokens_seen": 1436286976 + }, + { + "epoch": 0.44, + "learning_rate": 0.0004564413239719158, + "loss": 1.3312, + "theoretical_loss": 3.558071036814565, + "tokens_seen": 1436352512 + }, + { + "epoch": 0.44, + "learning_rate": 0.0004564252758274825, + "loss": 1.3732, + "theoretical_loss": 3.558056719688328, + "tokens_seen": 1436418048 + }, + { + "epoch": 0.44, + "learning_rate": 0.0004564092276830492, + "loss": 1.3186, + "theoretical_loss": 3.5580424033981792, + "tokens_seen": 1436483584 + }, + { + "epoch": 0.44, + "learning_rate": 0.0004563931795386159, + "loss": 1.4288, + "theoretical_loss": 3.5580280879440322, + "tokens_seen": 1436549120 + }, + { + "epoch": 0.44, + "learning_rate": 0.0004563771313941826, + "loss": 1.4657, + "theoretical_loss": 3.5580137733257997, + "tokens_seen": 1436614656 + }, + { + "epoch": 0.44, + "learning_rate": 0.0004563610832497493, + "loss": 1.2661, + "theoretical_loss": 3.557999459543394, + "tokens_seen": 1436680192 + }, + { + "epoch": 0.44, + "learning_rate": 0.000456345035105316, + "loss": 1.5704, + "theoretical_loss": 3.55798514659673, + "tokens_seen": 1436745728 + }, + { + "epoch": 0.44, + "learning_rate": 0.0004563289869608827, + "loss": 1.4019, + "theoretical_loss": 3.5579708344857197, + "tokens_seen": 1436811264 + }, + { + "epoch": 0.44, + "learning_rate": 0.0004563129388164494, + "loss": 1.3324, + "theoretical_loss": 3.557956523210276, + "tokens_seen": 1436876800 + }, + { + "epoch": 0.44, + "learning_rate": 0.0004562968906720161, + "loss": 1.2408, + "theoretical_loss": 3.5579422127703113, + "tokens_seen": 1436942336 + }, + { + "epoch": 0.44, + "learning_rate": 0.00045628084252758274, + "loss": 1.3815, + "theoretical_loss": 3.5579279031657407, + "tokens_seen": 1437007872 + }, + { + "epoch": 0.44, + "learning_rate": 0.0004562647943831495, + "loss": 1.29, + "theoretical_loss": 3.557913594396476, + "tokens_seen": 1437073408 + }, + { + "epoch": 0.44, + "learning_rate": 0.00045624874623871614, + "loss": 1.2724, + "theoretical_loss": 3.5578992864624306, + "tokens_seen": 1437138944 + }, + { + "epoch": 0.44, + "learning_rate": 0.0004562326980942829, + "loss": 1.3683, + "theoretical_loss": 3.557884979363518, + "tokens_seen": 1437204480 + }, + { + "epoch": 0.44, + "learning_rate": 0.00045621664994984954, + "loss": 1.4435, + "theoretical_loss": 3.557870673099651, + "tokens_seen": 1437270016 + }, + { + "epoch": 0.44, + "learning_rate": 0.0004562006018054163, + "loss": 1.325, + "theoretical_loss": 3.557856367670743, + "tokens_seen": 1437335552 + }, + { + "epoch": 0.44, + "learning_rate": 0.00045618455366098294, + "loss": 1.3487, + "theoretical_loss": 3.557842063076707, + "tokens_seen": 1437401088 + }, + { + "epoch": 0.44, + "learning_rate": 0.0004561685055165497, + "loss": 1.3695, + "theoretical_loss": 3.5578277593174565, + "tokens_seen": 1437466624 + }, + { + "epoch": 0.44, + "learning_rate": 0.00045615245737211634, + "loss": 1.3983, + "theoretical_loss": 3.5578134563929042, + "tokens_seen": 1437532160 + }, + { + "epoch": 0.44, + "learning_rate": 0.0004561364092276831, + "loss": 1.2998, + "theoretical_loss": 3.5577991543029635, + "tokens_seen": 1437597696 + }, + { + "epoch": 0.44, + "learning_rate": 0.00045612036108324974, + "loss": 1.2704, + "theoretical_loss": 3.557784853047548, + "tokens_seen": 1437663232 + }, + { + "epoch": 0.44, + "learning_rate": 0.0004561043129388165, + "loss": 1.3754, + "theoretical_loss": 3.557770552626571, + "tokens_seen": 1437728768 + }, + { + "epoch": 0.44, + "learning_rate": 0.00045608826479438314, + "loss": 1.3635, + "theoretical_loss": 3.557756253039946, + "tokens_seen": 1437794304 + }, + { + "epoch": 0.44, + "learning_rate": 0.0004560722166499499, + "loss": 1.3949, + "theoretical_loss": 3.557741954287585, + "tokens_seen": 1437859840 + }, + { + "epoch": 0.44, + "learning_rate": 0.00045605616850551655, + "loss": 1.4522, + "theoretical_loss": 3.5577276563694022, + "tokens_seen": 1437925376 + }, + { + "epoch": 0.44, + "learning_rate": 0.0004560401203610833, + "loss": 1.4818, + "theoretical_loss": 3.5577133592853114, + "tokens_seen": 1437990912 + }, + { + "epoch": 0.44, + "learning_rate": 0.00045602407221664995, + "loss": 1.4208, + "theoretical_loss": 3.557699063035225, + "tokens_seen": 1438056448 + }, + { + "epoch": 0.44, + "learning_rate": 0.0004560080240722167, + "loss": 1.3292, + "theoretical_loss": 3.5576847676190573, + "tokens_seen": 1438121984 + }, + { + "epoch": 0.44, + "learning_rate": 0.00045599197592778335, + "loss": 1.3466, + "theoretical_loss": 3.557670473036721, + "tokens_seen": 1438187520 + }, + { + "epoch": 0.44, + "learning_rate": 0.0004559759277833501, + "loss": 1.4318, + "theoretical_loss": 3.557656179288129, + "tokens_seen": 1438253056 + }, + { + "epoch": 0.44, + "learning_rate": 0.00045595987963891675, + "loss": 1.3898, + "theoretical_loss": 3.5576418863731956, + "tokens_seen": 1438318592 + }, + { + "epoch": 0.44, + "learning_rate": 0.0004559438314944835, + "loss": 1.4235, + "theoretical_loss": 3.5576275942918345, + "tokens_seen": 1438384128 + }, + { + "epoch": 0.44, + "learning_rate": 0.00045592778335005015, + "loss": 1.4726, + "theoretical_loss": 3.557613303043958, + "tokens_seen": 1438449664 + }, + { + "epoch": 0.44, + "objective/train/docs_used": 525418, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.3809759616851807, + "objective/train/theoretical_loss": 3.5575990126294794, + "objective/train/tokens_used": 1458975200, + "theoretical_loss": 3.5575990126294794, + "tokens_seen": 1438515200 + }, + { + "epoch": 0.44, + "learning_rate": 0.0004559117352056169, + "loss": 1.3201, + "theoretical_loss": 3.5575990126294794, + "tokens_seen": 1438515200 + }, + { + "epoch": 0.44, + "learning_rate": 0.00045589568706118355, + "loss": 1.4758, + "theoretical_loss": 3.5575847230483135, + "tokens_seen": 1438580736 + }, + { + "epoch": 0.44, + "learning_rate": 0.0004558796389167503, + "loss": 1.4357, + "theoretical_loss": 3.5575704343003727, + "tokens_seen": 1438646272 + }, + { + "epoch": 0.44, + "learning_rate": 0.00045586359077231695, + "loss": 1.5008, + "theoretical_loss": 3.557556146385571, + "tokens_seen": 1438711808 + }, + { + "epoch": 0.44, + "learning_rate": 0.0004558475426278837, + "loss": 1.394, + "theoretical_loss": 3.5575418593038215, + "tokens_seen": 1438777344 + }, + { + "epoch": 0.44, + "learning_rate": 0.00045583149448345035, + "loss": 1.3414, + "theoretical_loss": 3.557527573055038, + "tokens_seen": 1438842880 + }, + { + "epoch": 0.44, + "learning_rate": 0.0004558154463390171, + "loss": 1.3788, + "theoretical_loss": 3.5575132876391335, + "tokens_seen": 1438908416 + }, + { + "epoch": 0.44, + "learning_rate": 0.00045579939819458376, + "loss": 1.3189, + "theoretical_loss": 3.557499003056022, + "tokens_seen": 1438973952 + }, + { + "epoch": 0.44, + "learning_rate": 0.0004557833500501505, + "loss": 1.3717, + "theoretical_loss": 3.557484719305617, + "tokens_seen": 1439039488 + }, + { + "epoch": 0.44, + "learning_rate": 0.00045576730190571716, + "loss": 1.4021, + "theoretical_loss": 3.5574704363878316, + "tokens_seen": 1439105024 + }, + { + "epoch": 0.44, + "learning_rate": 0.0004557512537612839, + "loss": 1.4257, + "theoretical_loss": 3.5574561543025798, + "tokens_seen": 1439170560 + }, + { + "epoch": 0.44, + "learning_rate": 0.00045573520561685056, + "loss": 1.3436, + "theoretical_loss": 3.5574418730497745, + "tokens_seen": 1439236096 + }, + { + "epoch": 0.44, + "learning_rate": 0.0004557191574724173, + "loss": 1.3314, + "theoretical_loss": 3.5574275926293306, + "tokens_seen": 1439301632 + }, + { + "epoch": 0.44, + "learning_rate": 0.00045570310932798396, + "loss": 1.2782, + "theoretical_loss": 3.5574133130411605, + "tokens_seen": 1439367168 + }, + { + "epoch": 0.44, + "learning_rate": 0.0004556870611835507, + "loss": 1.4194, + "theoretical_loss": 3.5573990342851785, + "tokens_seen": 1439432704 + }, + { + "epoch": 0.44, + "learning_rate": 0.00045567101303911736, + "loss": 1.4843, + "theoretical_loss": 3.5573847563612975, + "tokens_seen": 1439498240 + }, + { + "epoch": 0.44, + "learning_rate": 0.0004556549648946841, + "loss": 1.449, + "theoretical_loss": 3.557370479269432, + "tokens_seen": 1439563776 + }, + { + "epoch": 0.44, + "learning_rate": 0.00045563891675025076, + "loss": 1.4427, + "theoretical_loss": 3.5573562030094945, + "tokens_seen": 1439629312 + }, + { + "epoch": 0.44, + "learning_rate": 0.00045562286860581746, + "loss": 1.4316, + "theoretical_loss": 3.5573419275813993, + "tokens_seen": 1439694848 + }, + { + "epoch": 0.44, + "learning_rate": 0.00045560682046138416, + "loss": 1.3188, + "theoretical_loss": 3.5573276529850606, + "tokens_seen": 1439760384 + }, + { + "epoch": 0.44, + "learning_rate": 0.00045559077231695086, + "loss": 1.3577, + "theoretical_loss": 3.5573133792203917, + "tokens_seen": 1439825920 + }, + { + "epoch": 0.44, + "learning_rate": 0.00045557472417251756, + "loss": 1.4243, + "theoretical_loss": 3.5572991062873056, + "tokens_seen": 1439891456 + }, + { + "epoch": 0.44, + "learning_rate": 0.00045555867602808427, + "loss": 1.4202, + "theoretical_loss": 3.557284834185717, + "tokens_seen": 1439956992 + }, + { + "epoch": 0.44, + "learning_rate": 0.00045554262788365097, + "loss": 1.4388, + "theoretical_loss": 3.557270562915539, + "tokens_seen": 1440022528 + }, + { + "epoch": 0.44, + "learning_rate": 0.00045552657973921767, + "loss": 1.3653, + "theoretical_loss": 3.5572562924766853, + "tokens_seen": 1440088064 + }, + { + "epoch": 0.44, + "learning_rate": 0.0004555105315947843, + "loss": 1.4149, + "theoretical_loss": 3.5572420228690707, + "tokens_seen": 1440153600 + }, + { + "epoch": 0.44, + "learning_rate": 0.00045549448345035107, + "loss": 1.4547, + "theoretical_loss": 3.557227754092607, + "tokens_seen": 1440219136 + }, + { + "epoch": 0.44, + "learning_rate": 0.0004554784353059177, + "loss": 1.3557, + "theoretical_loss": 3.55721348614721, + "tokens_seen": 1440284672 + }, + { + "epoch": 0.44, + "learning_rate": 0.00045546238716148447, + "loss": 1.5747, + "theoretical_loss": 3.5571992190327926, + "tokens_seen": 1440350208 + }, + { + "epoch": 0.44, + "learning_rate": 0.0004554463390170511, + "loss": 1.4167, + "theoretical_loss": 3.557184952749268, + "tokens_seen": 1440415744 + }, + { + "epoch": 0.44, + "learning_rate": 0.00045543029087261787, + "loss": 1.4023, + "theoretical_loss": 3.5571706872965505, + "tokens_seen": 1440481280 + }, + { + "epoch": 0.44, + "learning_rate": 0.0004554142427281845, + "loss": 1.3405, + "theoretical_loss": 3.5571564226745545, + "tokens_seen": 1440546816 + }, + { + "epoch": 0.44, + "learning_rate": 0.00045539819458375127, + "loss": 1.4005, + "theoretical_loss": 3.5571421588831935, + "tokens_seen": 1440612352 + }, + { + "epoch": 0.44, + "learning_rate": 0.0004553821464393179, + "loss": 1.3287, + "theoretical_loss": 3.5571278959223807, + "tokens_seen": 1440677888 + }, + { + "epoch": 0.44, + "learning_rate": 0.0004553660982948847, + "loss": 1.3459, + "theoretical_loss": 3.5571136337920306, + "tokens_seen": 1440743424 + }, + { + "epoch": 0.44, + "learning_rate": 0.0004553500501504513, + "loss": 1.2289, + "theoretical_loss": 3.5570993724920568, + "tokens_seen": 1440808960 + }, + { + "epoch": 0.44, + "learning_rate": 0.0004553340020060181, + "loss": 1.3468, + "theoretical_loss": 3.5570851120223734, + "tokens_seen": 1440874496 + }, + { + "epoch": 0.44, + "learning_rate": 0.0004553179538615847, + "loss": 1.4695, + "theoretical_loss": 3.5570708523828944, + "tokens_seen": 1440940032 + }, + { + "epoch": 0.44, + "learning_rate": 0.0004553019057171515, + "loss": 1.2905, + "theoretical_loss": 3.5570565935735337, + "tokens_seen": 1441005568 + }, + { + "epoch": 0.44, + "learning_rate": 0.0004552858575727181, + "loss": 1.4714, + "theoretical_loss": 3.5570423355942045, + "tokens_seen": 1441071104 + }, + { + "epoch": 0.44, + "learning_rate": 0.0004552698094282849, + "loss": 1.3861, + "theoretical_loss": 3.5570280784448216, + "tokens_seen": 1441136640 + }, + { + "epoch": 0.44, + "learning_rate": 0.00045525376128385163, + "loss": 1.4611, + "theoretical_loss": 3.557013822125299, + "tokens_seen": 1441202176 + }, + { + "epoch": 0.44, + "learning_rate": 0.0004552377131394183, + "loss": 1.4377, + "theoretical_loss": 3.5569995666355494, + "tokens_seen": 1441267712 + }, + { + "epoch": 0.44, + "learning_rate": 0.00045522166499498503, + "loss": 1.3795, + "theoretical_loss": 3.556985311975488, + "tokens_seen": 1441333248 + }, + { + "epoch": 0.44, + "learning_rate": 0.0004552056168505517, + "loss": 1.2836, + "theoretical_loss": 3.556971058145029, + "tokens_seen": 1441398784 + }, + { + "epoch": 0.44, + "learning_rate": 0.00045518956870611843, + "loss": 1.3831, + "theoretical_loss": 3.556956805144085, + "tokens_seen": 1441464320 + }, + { + "epoch": 0.44, + "learning_rate": 0.0004551735205616851, + "loss": 1.29, + "theoretical_loss": 3.556942552972572, + "tokens_seen": 1441529856 + }, + { + "epoch": 0.44, + "learning_rate": 0.00045515747241725184, + "loss": 1.3924, + "theoretical_loss": 3.556928301630402, + "tokens_seen": 1441595392 + }, + { + "epoch": 0.44, + "learning_rate": 0.0004551414242728185, + "loss": 1.4207, + "theoretical_loss": 3.5569140511174906, + "tokens_seen": 1441660928 + }, + { + "epoch": 0.44, + "learning_rate": 0.00045512537612838524, + "loss": 1.3258, + "theoretical_loss": 3.556899801433751, + "tokens_seen": 1441726464 + }, + { + "epoch": 0.44, + "objective/train/docs_used": 526671, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.2915194034576416, + "objective/train/theoretical_loss": 3.5568855525790974, + "objective/train/tokens_used": 1462252000, + "theoretical_loss": 3.5568855525790974, + "tokens_seen": 1441792000 + }, + { + "epoch": 0.44, + "learning_rate": 0.0004551093279839519, + "loss": 1.35, + "theoretical_loss": 3.5568855525790974, + "tokens_seen": 1441792000 + }, + { + "epoch": 0.44, + "learning_rate": 0.00045509327983951864, + "loss": 1.5635, + "theoretical_loss": 3.556871304553444, + "tokens_seen": 1441857536 + }, + { + "epoch": 0.44, + "learning_rate": 0.0004550772316950853, + "loss": 1.476, + "theoretical_loss": 3.556857057356704, + "tokens_seen": 1441923072 + }, + { + "epoch": 0.44, + "learning_rate": 0.00045506118355065204, + "loss": 1.355, + "theoretical_loss": 3.556842810988794, + "tokens_seen": 1441988608 + }, + { + "epoch": 0.44, + "learning_rate": 0.0004550451354062187, + "loss": 1.375, + "theoretical_loss": 3.5568285654496252, + "tokens_seen": 1442054144 + }, + { + "epoch": 0.44, + "learning_rate": 0.00045502908726178544, + "loss": 1.3222, + "theoretical_loss": 3.5568143207391136, + "tokens_seen": 1442119680 + }, + { + "epoch": 0.44, + "learning_rate": 0.0004550130391173521, + "loss": 1.515, + "theoretical_loss": 3.556800076857172, + "tokens_seen": 1442185216 + }, + { + "epoch": 0.44, + "learning_rate": 0.0004549969909729188, + "loss": 1.3838, + "theoretical_loss": 3.5567858338037164, + "tokens_seen": 1442250752 + }, + { + "epoch": 0.44, + "learning_rate": 0.0004549809428284855, + "loss": 1.3382, + "theoretical_loss": 3.5567715915786593, + "tokens_seen": 1442316288 + }, + { + "epoch": 0.44, + "learning_rate": 0.0004549648946840522, + "loss": 1.3392, + "theoretical_loss": 3.556757350181915, + "tokens_seen": 1442381824 + }, + { + "epoch": 0.44, + "learning_rate": 0.0004549488465396189, + "loss": 1.4484, + "theoretical_loss": 3.556743109613399, + "tokens_seen": 1442447360 + }, + { + "epoch": 0.44, + "learning_rate": 0.0004549327983951856, + "loss": 1.47, + "theoretical_loss": 3.556728869873024, + "tokens_seen": 1442512896 + }, + { + "epoch": 0.44, + "learning_rate": 0.0004549167502507523, + "loss": 1.3995, + "theoretical_loss": 3.5567146309607054, + "tokens_seen": 1442578432 + }, + { + "epoch": 0.44, + "learning_rate": 0.000454900702106319, + "loss": 1.4496, + "theoretical_loss": 3.5567003928763565, + "tokens_seen": 1442643968 + }, + { + "epoch": 0.44, + "learning_rate": 0.0004548846539618857, + "loss": 1.4048, + "theoretical_loss": 3.556686155619892, + "tokens_seen": 1442709504 + }, + { + "epoch": 0.44, + "learning_rate": 0.0004548686058174524, + "loss": 1.3507, + "theoretical_loss": 3.556671919191226, + "tokens_seen": 1442775040 + }, + { + "epoch": 0.44, + "learning_rate": 0.00045485255767301904, + "loss": 1.287, + "theoretical_loss": 3.556657683590273, + "tokens_seen": 1442840576 + }, + { + "epoch": 0.44, + "learning_rate": 0.0004548365095285858, + "loss": 1.3113, + "theoretical_loss": 3.5566434488169474, + "tokens_seen": 1442906112 + }, + { + "epoch": 0.44, + "learning_rate": 0.00045482046138415244, + "loss": 1.3871, + "theoretical_loss": 3.5566292148711627, + "tokens_seen": 1442971648 + }, + { + "epoch": 0.44, + "learning_rate": 0.0004548044132397192, + "loss": 1.332, + "theoretical_loss": 3.5566149817528343, + "tokens_seen": 1443037184 + }, + { + "epoch": 0.44, + "learning_rate": 0.00045478836509528584, + "loss": 1.3324, + "theoretical_loss": 3.556600749461876, + "tokens_seen": 1443102720 + }, + { + "epoch": 0.44, + "learning_rate": 0.0004547723169508526, + "loss": 1.2575, + "theoretical_loss": 3.556586517998202, + "tokens_seen": 1443168256 + }, + { + "epoch": 0.44, + "learning_rate": 0.00045475626880641924, + "loss": 1.4958, + "theoretical_loss": 3.5565722873617265, + "tokens_seen": 1443233792 + }, + { + "epoch": 0.44, + "learning_rate": 0.000454740220661986, + "loss": 1.4103, + "theoretical_loss": 3.5565580575523645, + "tokens_seen": 1443299328 + }, + { + "epoch": 0.44, + "learning_rate": 0.00045472417251755264, + "loss": 1.3755, + "theoretical_loss": 3.55654382857003, + "tokens_seen": 1443364864 + }, + { + "epoch": 0.44, + "learning_rate": 0.0004547081243731194, + "loss": 1.4463, + "theoretical_loss": 3.5565296004146365, + "tokens_seen": 1443430400 + }, + { + "epoch": 0.44, + "learning_rate": 0.00045469207622868605, + "loss": 1.334, + "theoretical_loss": 3.5565153730861, + "tokens_seen": 1443495936 + }, + { + "epoch": 0.44, + "learning_rate": 0.0004546760280842528, + "loss": 1.325, + "theoretical_loss": 3.5565011465843344, + "tokens_seen": 1443561472 + }, + { + "epoch": 0.44, + "learning_rate": 0.00045465997993981945, + "loss": 1.4194, + "theoretical_loss": 3.556486920909254, + "tokens_seen": 1443627008 + }, + { + "epoch": 0.44, + "learning_rate": 0.0004546439317953862, + "loss": 1.3436, + "theoretical_loss": 3.556472696060773, + "tokens_seen": 1443692544 + }, + { + "epoch": 0.44, + "learning_rate": 0.00045462788365095285, + "loss": 1.3176, + "theoretical_loss": 3.5564584720388055, + "tokens_seen": 1443758080 + }, + { + "epoch": 0.44, + "learning_rate": 0.0004546118355065196, + "loss": 1.3828, + "theoretical_loss": 3.556444248843267, + "tokens_seen": 1443823616 + }, + { + "epoch": 0.44, + "learning_rate": 0.00045459578736208625, + "loss": 1.32, + "theoretical_loss": 3.556430026474071, + "tokens_seen": 1443889152 + }, + { + "epoch": 0.44, + "learning_rate": 0.000454579739217653, + "loss": 1.3736, + "theoretical_loss": 3.5564158049311327, + "tokens_seen": 1443954688 + }, + { + "epoch": 0.44, + "learning_rate": 0.00045456369107321965, + "loss": 1.323, + "theoretical_loss": 3.556401584214367, + "tokens_seen": 1444020224 + }, + { + "epoch": 0.44, + "learning_rate": 0.0004545476429287864, + "loss": 1.2444, + "theoretical_loss": 3.5563873643236867, + "tokens_seen": 1444085760 + }, + { + "epoch": 0.44, + "learning_rate": 0.00045453159478435305, + "loss": 1.3851, + "theoretical_loss": 3.5563731452590077, + "tokens_seen": 1444151296 + }, + { + "epoch": 0.44, + "learning_rate": 0.0004545155466399198, + "loss": 1.3752, + "theoretical_loss": 3.5563589270202445, + "tokens_seen": 1444216832 + }, + { + "epoch": 0.44, + "learning_rate": 0.00045449949849548645, + "loss": 1.391, + "theoretical_loss": 3.556344709607311, + "tokens_seen": 1444282368 + }, + { + "epoch": 0.44, + "learning_rate": 0.0004544834503510532, + "loss": 1.3943, + "theoretical_loss": 3.5563304930201225, + "tokens_seen": 1444347904 + }, + { + "epoch": 0.44, + "learning_rate": 0.00045446740220661985, + "loss": 1.4573, + "theoretical_loss": 3.556316277258593, + "tokens_seen": 1444413440 + }, + { + "epoch": 0.44, + "learning_rate": 0.0004544513540621866, + "loss": 1.407, + "theoretical_loss": 3.5563020623226373, + "tokens_seen": 1444478976 + }, + { + "epoch": 0.44, + "learning_rate": 0.00045443530591775326, + "loss": 1.4905, + "theoretical_loss": 3.55628784821217, + "tokens_seen": 1444544512 + }, + { + "epoch": 0.44, + "learning_rate": 0.00045441925777332, + "loss": 1.2869, + "theoretical_loss": 3.5562736349271056, + "tokens_seen": 1444610048 + }, + { + "epoch": 0.44, + "learning_rate": 0.00045440320962888666, + "loss": 1.4627, + "theoretical_loss": 3.5562594224673587, + "tokens_seen": 1444675584 + }, + { + "epoch": 0.44, + "learning_rate": 0.0004543871614844534, + "loss": 1.4122, + "theoretical_loss": 3.556245210832844, + "tokens_seen": 1444741120 + }, + { + "epoch": 0.44, + "learning_rate": 0.00045437111334002006, + "loss": 1.3598, + "theoretical_loss": 3.556231000023477, + "tokens_seen": 1444806656 + }, + { + "epoch": 0.44, + "learning_rate": 0.0004543550651955868, + "loss": 1.4455, + "theoretical_loss": 3.556216790039171, + "tokens_seen": 1444872192 + }, + { + "epoch": 0.44, + "learning_rate": 0.00045433901705115346, + "loss": 1.3945, + "theoretical_loss": 3.5562025808798414, + "tokens_seen": 1444937728 + }, + { + "epoch": 0.44, + "learning_rate": 0.0004543229689067202, + "loss": 1.3845, + "theoretical_loss": 3.556188372545403, + "tokens_seen": 1445003264 + }, + { + "epoch": 0.44, + "objective/train/docs_used": 528002, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.3061717748641968, + "objective/train/theoretical_loss": 3.55617416503577, + "objective/train/tokens_used": 1465528800, + "theoretical_loss": 3.55617416503577, + "tokens_seen": 1445068800 + }, + { + "epoch": 0.44, + "learning_rate": 0.00045430692076228686, + "loss": 1.3009, + "theoretical_loss": 3.55617416503577, + "tokens_seen": 1445068800 + }, + { + "epoch": 0.44, + "learning_rate": 0.0004542908726178536, + "loss": 1.2817, + "theoretical_loss": 3.5561599583508574, + "tokens_seen": 1445134336 + }, + { + "epoch": 0.44, + "learning_rate": 0.00045427482447342026, + "loss": 1.5186, + "theoretical_loss": 3.55614575249058, + "tokens_seen": 1445199872 + }, + { + "epoch": 0.44, + "learning_rate": 0.000454258776328987, + "loss": 1.2582, + "theoretical_loss": 3.5561315474548527, + "tokens_seen": 1445265408 + }, + { + "epoch": 0.44, + "learning_rate": 0.00045424272818455366, + "loss": 1.3534, + "theoretical_loss": 3.5561173432435895, + "tokens_seen": 1445330944 + }, + { + "epoch": 0.44, + "learning_rate": 0.00045422668004012036, + "loss": 1.2143, + "theoretical_loss": 3.556103139856706, + "tokens_seen": 1445396480 + }, + { + "epoch": 0.44, + "learning_rate": 0.00045421063189568706, + "loss": 1.4299, + "theoretical_loss": 3.5560889372941165, + "tokens_seen": 1445462016 + }, + { + "epoch": 0.44, + "learning_rate": 0.00045419458375125377, + "loss": 1.1716, + "theoretical_loss": 3.556074735555736, + "tokens_seen": 1445527552 + }, + { + "epoch": 0.44, + "learning_rate": 0.00045417853560682047, + "loss": 1.4358, + "theoretical_loss": 3.556060534641479, + "tokens_seen": 1445593088 + }, + { + "epoch": 0.44, + "learning_rate": 0.00045416248746238717, + "loss": 1.4629, + "theoretical_loss": 3.5560463345512607, + "tokens_seen": 1445658624 + }, + { + "epoch": 0.44, + "learning_rate": 0.00045414643931795387, + "loss": 1.3549, + "theoretical_loss": 3.556032135284996, + "tokens_seen": 1445724160 + }, + { + "epoch": 0.44, + "learning_rate": 0.00045413039117352057, + "loss": 1.4278, + "theoretical_loss": 3.5560179368425997, + "tokens_seen": 1445789696 + }, + { + "epoch": 0.44, + "learning_rate": 0.00045411434302908727, + "loss": 1.4466, + "theoretical_loss": 3.5560037392239865, + "tokens_seen": 1445855232 + }, + { + "epoch": 0.44, + "learning_rate": 0.00045409829488465397, + "loss": 1.3592, + "theoretical_loss": 3.555989542429071, + "tokens_seen": 1445920768 + }, + { + "epoch": 0.44, + "learning_rate": 0.0004540822467402206, + "loss": 1.3505, + "theoretical_loss": 3.5559753464577684, + "tokens_seen": 1445986304 + }, + { + "epoch": 0.44, + "learning_rate": 0.00045406619859578737, + "loss": 1.3663, + "theoretical_loss": 3.5559611513099934, + "tokens_seen": 1446051840 + }, + { + "epoch": 0.44, + "learning_rate": 0.0004540501504513541, + "loss": 1.4015, + "theoretical_loss": 3.555946956985661, + "tokens_seen": 1446117376 + }, + { + "epoch": 0.44, + "learning_rate": 0.00045403410230692077, + "loss": 1.398, + "theoretical_loss": 3.5559327634846865, + "tokens_seen": 1446182912 + }, + { + "epoch": 0.44, + "learning_rate": 0.0004540180541624875, + "loss": 1.3372, + "theoretical_loss": 3.5559185708069845, + "tokens_seen": 1446248448 + }, + { + "epoch": 0.44, + "learning_rate": 0.00045400200601805417, + "loss": 1.4273, + "theoretical_loss": 3.55590437895247, + "tokens_seen": 1446313984 + }, + { + "epoch": 0.44, + "learning_rate": 0.00045398595787362093, + "loss": 1.3913, + "theoretical_loss": 3.5558901879210576, + "tokens_seen": 1446379520 + }, + { + "epoch": 0.44, + "learning_rate": 0.0004539699097291876, + "loss": 1.3503, + "theoretical_loss": 3.555875997712663, + "tokens_seen": 1446445056 + }, + { + "epoch": 0.44, + "learning_rate": 0.00045395386158475433, + "loss": 1.3108, + "theoretical_loss": 3.5558618083272, + "tokens_seen": 1446510592 + }, + { + "epoch": 0.44, + "learning_rate": 0.000453937813440321, + "loss": 1.3536, + "theoretical_loss": 3.5558476197645854, + "tokens_seen": 1446576128 + }, + { + "epoch": 0.44, + "learning_rate": 0.00045392176529588773, + "loss": 1.4626, + "theoretical_loss": 3.5558334320247322, + "tokens_seen": 1446641664 + }, + { + "epoch": 0.44, + "learning_rate": 0.0004539057171514544, + "loss": 1.3417, + "theoretical_loss": 3.555819245107557, + "tokens_seen": 1446707200 + }, + { + "epoch": 0.44, + "learning_rate": 0.00045388966900702113, + "loss": 1.3551, + "theoretical_loss": 3.555805059012974, + "tokens_seen": 1446772736 + }, + { + "epoch": 0.44, + "learning_rate": 0.0004538736208625878, + "loss": 1.4652, + "theoretical_loss": 3.5557908737408983, + "tokens_seen": 1446838272 + }, + { + "epoch": 0.44, + "learning_rate": 0.00045385757271815453, + "loss": 1.2861, + "theoretical_loss": 3.5557766892912452, + "tokens_seen": 1446903808 + }, + { + "epoch": 0.44, + "learning_rate": 0.0004538415245737212, + "loss": 1.3404, + "theoretical_loss": 3.55576250566393, + "tokens_seen": 1446969344 + }, + { + "epoch": 0.44, + "learning_rate": 0.00045382547642928793, + "loss": 1.3489, + "theoretical_loss": 3.555748322858867, + "tokens_seen": 1447034880 + }, + { + "epoch": 0.44, + "learning_rate": 0.0004538094282848546, + "loss": 1.2783, + "theoretical_loss": 3.555734140875972, + "tokens_seen": 1447100416 + }, + { + "epoch": 0.44, + "learning_rate": 0.00045379338014042134, + "loss": 1.3395, + "theoretical_loss": 3.5557199597151596, + "tokens_seen": 1447165952 + }, + { + "epoch": 0.44, + "learning_rate": 0.000453777331995988, + "loss": 1.2901, + "theoretical_loss": 3.555705779376346, + "tokens_seen": 1447231488 + }, + { + "epoch": 0.44, + "learning_rate": 0.00045376128385155474, + "loss": 1.4161, + "theoretical_loss": 3.5556915998594447, + "tokens_seen": 1447297024 + }, + { + "epoch": 0.44, + "learning_rate": 0.0004537452357071214, + "loss": 1.3611, + "theoretical_loss": 3.555677421164372, + "tokens_seen": 1447362560 + }, + { + "epoch": 0.44, + "learning_rate": 0.00045372918756268814, + "loss": 1.2332, + "theoretical_loss": 3.555663243291043, + "tokens_seen": 1447428096 + }, + { + "epoch": 0.44, + "learning_rate": 0.0004537131394182548, + "loss": 1.3781, + "theoretical_loss": 3.555649066239372, + "tokens_seen": 1447493632 + }, + { + "epoch": 0.44, + "learning_rate": 0.00045369709127382154, + "loss": 1.2967, + "theoretical_loss": 3.555634890009275, + "tokens_seen": 1447559168 + }, + { + "epoch": 0.44, + "learning_rate": 0.0004536810431293882, + "loss": 1.3674, + "theoretical_loss": 3.5556207146006673, + "tokens_seen": 1447624704 + }, + { + "epoch": 0.44, + "learning_rate": 0.00045366499498495494, + "loss": 1.4046, + "theoretical_loss": 3.5556065400134633, + "tokens_seen": 1447690240 + }, + { + "epoch": 0.44, + "learning_rate": 0.0004536489468405216, + "loss": 1.3755, + "theoretical_loss": 3.5555923662475792, + "tokens_seen": 1447755776 + }, + { + "epoch": 0.44, + "learning_rate": 0.00045363289869608834, + "loss": 1.3162, + "theoretical_loss": 3.55557819330293, + "tokens_seen": 1447821312 + }, + { + "epoch": 0.44, + "learning_rate": 0.000453616850551655, + "loss": 1.4196, + "theoretical_loss": 3.55556402117943, + "tokens_seen": 1447886848 + }, + { + "epoch": 0.44, + "learning_rate": 0.0004536008024072217, + "loss": 1.3658, + "theoretical_loss": 3.5555498498769955, + "tokens_seen": 1447952384 + }, + { + "epoch": 0.44, + "learning_rate": 0.0004535847542627884, + "loss": 1.3602, + "theoretical_loss": 3.555535679395541, + "tokens_seen": 1448017920 + }, + { + "epoch": 0.44, + "learning_rate": 0.0004535687061183551, + "loss": 1.3262, + "theoretical_loss": 3.555521509734983, + "tokens_seen": 1448083456 + }, + { + "epoch": 0.44, + "learning_rate": 0.0004535526579739218, + "loss": 1.4064, + "theoretical_loss": 3.555507340895235, + "tokens_seen": 1448148992 + }, + { + "epoch": 0.44, + "learning_rate": 0.0004535366098294885, + "loss": 1.4095, + "theoretical_loss": 3.555493172876214, + "tokens_seen": 1448214528 + }, + { + "epoch": 0.44, + "learning_rate": 0.0004535205616850552, + "loss": 1.3258, + "theoretical_loss": 3.555479005677835, + "tokens_seen": 1448280064 + }, + { + "epoch": 0.44, + "objective/train/docs_used": 529166, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.4011125564575195, + "objective/train/theoretical_loss": 3.5554648393000123, + "objective/train/tokens_used": 1468805600, + "theoretical_loss": 3.5554648393000123, + "tokens_seen": 1448345600 + }, + { + "epoch": 0.44, + "learning_rate": 0.0004535045135406219, + "loss": 1.4097, + "theoretical_loss": 3.5554648393000123, + "tokens_seen": 1448345600 + }, + { + "epoch": 0.44, + "learning_rate": 0.0004534884653961886, + "loss": 1.3359, + "theoretical_loss": 3.555450673742662, + "tokens_seen": 1448411136 + }, + { + "epoch": 0.44, + "learning_rate": 0.0004534724172517553, + "loss": 1.4014, + "theoretical_loss": 3.5554365090057, + "tokens_seen": 1448476672 + }, + { + "epoch": 0.44, + "learning_rate": 0.00045345636910732194, + "loss": 1.4804, + "theoretical_loss": 3.5554223450890405, + "tokens_seen": 1448542208 + }, + { + "epoch": 0.44, + "learning_rate": 0.0004534403209628887, + "loss": 1.3951, + "theoretical_loss": 3.5554081819925996, + "tokens_seen": 1448607744 + }, + { + "epoch": 0.44, + "learning_rate": 0.00045342427281845534, + "loss": 1.3428, + "theoretical_loss": 3.5553940197162923, + "tokens_seen": 1448673280 + }, + { + "epoch": 0.44, + "learning_rate": 0.0004534082246740221, + "loss": 1.4003, + "theoretical_loss": 3.555379858260035, + "tokens_seen": 1448738816 + }, + { + "epoch": 0.44, + "learning_rate": 0.00045339217652958874, + "loss": 1.4176, + "theoretical_loss": 3.5553656976237415, + "tokens_seen": 1448804352 + }, + { + "epoch": 0.44, + "learning_rate": 0.0004533761283851555, + "loss": 1.2691, + "theoretical_loss": 3.5553515378073284, + "tokens_seen": 1448869888 + }, + { + "epoch": 0.44, + "learning_rate": 0.00045336008024072214, + "loss": 1.3536, + "theoretical_loss": 3.5553373788107114, + "tokens_seen": 1448935424 + }, + { + "epoch": 0.44, + "learning_rate": 0.0004533440320962889, + "loss": 1.3713, + "theoretical_loss": 3.555323220633805, + "tokens_seen": 1449000960 + }, + { + "epoch": 0.44, + "learning_rate": 0.00045332798395185555, + "loss": 1.4281, + "theoretical_loss": 3.5553090632765247, + "tokens_seen": 1449066496 + }, + { + "epoch": 0.44, + "learning_rate": 0.0004533119358074223, + "loss": 1.5112, + "theoretical_loss": 3.555294906738787, + "tokens_seen": 1449132032 + }, + { + "epoch": 0.44, + "learning_rate": 0.00045329588766298895, + "loss": 1.29, + "theoretical_loss": 3.5552807510205064, + "tokens_seen": 1449197568 + }, + { + "epoch": 0.44, + "learning_rate": 0.0004532798395185557, + "loss": 1.3799, + "theoretical_loss": 3.555266596121599, + "tokens_seen": 1449263104 + }, + { + "epoch": 0.44, + "learning_rate": 0.00045326379137412235, + "loss": 1.4521, + "theoretical_loss": 3.55525244204198, + "tokens_seen": 1449328640 + }, + { + "epoch": 0.44, + "learning_rate": 0.0004532477432296891, + "loss": 1.5089, + "theoretical_loss": 3.555238288781565, + "tokens_seen": 1449394176 + }, + { + "epoch": 0.44, + "learning_rate": 0.00045323169508525575, + "loss": 1.3509, + "theoretical_loss": 3.5552241363402697, + "tokens_seen": 1449459712 + }, + { + "epoch": 0.44, + "learning_rate": 0.0004532156469408225, + "loss": 1.3586, + "theoretical_loss": 3.55520998471801, + "tokens_seen": 1449525248 + }, + { + "epoch": 0.44, + "learning_rate": 0.00045319959879638915, + "loss": 1.4227, + "theoretical_loss": 3.5551958339147003, + "tokens_seen": 1449590784 + }, + { + "epoch": 0.44, + "learning_rate": 0.0004531835506519559, + "loss": 1.3244, + "theoretical_loss": 3.5551816839302575, + "tokens_seen": 1449656320 + }, + { + "epoch": 0.44, + "learning_rate": 0.00045316750250752255, + "loss": 1.4526, + "theoretical_loss": 3.555167534764596, + "tokens_seen": 1449721856 + }, + { + "epoch": 0.44, + "learning_rate": 0.0004531514543630893, + "loss": 1.222, + "theoretical_loss": 3.555153386417633, + "tokens_seen": 1449787392 + }, + { + "epoch": 0.44, + "learning_rate": 0.00045313540621865595, + "loss": 1.4406, + "theoretical_loss": 3.555139238889282, + "tokens_seen": 1449852928 + }, + { + "epoch": 0.44, + "learning_rate": 0.0004531193580742227, + "loss": 1.3271, + "theoretical_loss": 3.5551250921794604, + "tokens_seen": 1449918464 + }, + { + "epoch": 0.44, + "learning_rate": 0.00045310330992978935, + "loss": 1.3812, + "theoretical_loss": 3.5551109462880834, + "tokens_seen": 1449984000 + }, + { + "epoch": 0.44, + "learning_rate": 0.0004530872617853561, + "loss": 1.4402, + "theoretical_loss": 3.555096801215066, + "tokens_seen": 1450049536 + }, + { + "epoch": 0.44, + "learning_rate": 0.00045307121364092276, + "loss": 1.4754, + "theoretical_loss": 3.555082656960325, + "tokens_seen": 1450115072 + }, + { + "epoch": 0.44, + "learning_rate": 0.0004530551654964895, + "loss": 1.3306, + "theoretical_loss": 3.555068513523775, + "tokens_seen": 1450180608 + }, + { + "epoch": 0.44, + "learning_rate": 0.00045303911735205616, + "loss": 1.4308, + "theoretical_loss": 3.555054370905332, + "tokens_seen": 1450246144 + }, + { + "epoch": 0.44, + "learning_rate": 0.0004530230692076229, + "loss": 1.4407, + "theoretical_loss": 3.5550402291049124, + "tokens_seen": 1450311680 + }, + { + "epoch": 0.44, + "learning_rate": 0.00045300702106318956, + "loss": 1.4822, + "theoretical_loss": 3.555026088122431, + "tokens_seen": 1450377216 + }, + { + "epoch": 0.44, + "learning_rate": 0.0004529909729187563, + "loss": 1.397, + "theoretical_loss": 3.555011947957804, + "tokens_seen": 1450442752 + }, + { + "epoch": 0.44, + "learning_rate": 0.00045297492477432296, + "loss": 1.3008, + "theoretical_loss": 3.5549978086109473, + "tokens_seen": 1450508288 + }, + { + "epoch": 0.44, + "learning_rate": 0.0004529588766298897, + "loss": 1.3255, + "theoretical_loss": 3.554983670081776, + "tokens_seen": 1450573824 + }, + { + "epoch": 0.44, + "learning_rate": 0.00045294282848545636, + "loss": 1.2798, + "theoretical_loss": 3.554969532370207, + "tokens_seen": 1450639360 + }, + { + "epoch": 0.44, + "learning_rate": 0.0004529267803410231, + "loss": 1.4278, + "theoretical_loss": 3.5549553954761546, + "tokens_seen": 1450704896 + }, + { + "epoch": 0.44, + "learning_rate": 0.00045291073219658976, + "loss": 1.4162, + "theoretical_loss": 3.554941259399536, + "tokens_seen": 1450770432 + }, + { + "epoch": 0.44, + "learning_rate": 0.0004528946840521565, + "loss": 1.343, + "theoretical_loss": 3.5549271241402662, + "tokens_seen": 1450835968 + }, + { + "epoch": 0.44, + "learning_rate": 0.00045287863590772316, + "loss": 1.3528, + "theoretical_loss": 3.554912989698261, + "tokens_seen": 1450901504 + }, + { + "epoch": 0.44, + "learning_rate": 0.0004528625877632899, + "loss": 1.4291, + "theoretical_loss": 3.5548988560734367, + "tokens_seen": 1450967040 + }, + { + "epoch": 0.44, + "learning_rate": 0.0004528465396188566, + "loss": 1.3282, + "theoretical_loss": 3.554884723265709, + "tokens_seen": 1451032576 + }, + { + "epoch": 0.44, + "learning_rate": 0.00045283049147442326, + "loss": 1.3567, + "theoretical_loss": 3.5548705912749936, + "tokens_seen": 1451098112 + }, + { + "epoch": 0.44, + "learning_rate": 0.00045281444332999, + "loss": 1.3896, + "theoretical_loss": 3.554856460101206, + "tokens_seen": 1451163648 + }, + { + "epoch": 0.44, + "learning_rate": 0.00045279839518555667, + "loss": 1.3751, + "theoretical_loss": 3.5548423297442633, + "tokens_seen": 1451229184 + }, + { + "epoch": 0.44, + "learning_rate": 0.0004527823470411234, + "loss": 1.3883, + "theoretical_loss": 3.5548282002040805, + "tokens_seen": 1451294720 + }, + { + "epoch": 0.44, + "learning_rate": 0.00045276629889669007, + "loss": 1.4606, + "theoretical_loss": 3.554814071480573, + "tokens_seen": 1451360256 + }, + { + "epoch": 0.44, + "learning_rate": 0.0004527502507522568, + "loss": 1.3341, + "theoretical_loss": 3.554799943573658, + "tokens_seen": 1451425792 + }, + { + "epoch": 0.44, + "learning_rate": 0.00045273420260782347, + "loss": 1.4476, + "theoretical_loss": 3.5547858164832506, + "tokens_seen": 1451491328 + }, + { + "epoch": 0.44, + "learning_rate": 0.0004527181544633902, + "loss": 1.4613, + "theoretical_loss": 3.554771690209267, + "tokens_seen": 1451556864 + }, + { + "epoch": 0.44, + "objective/train/docs_used": 530451, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.4979572296142578, + "objective/train/theoretical_loss": 3.5547575647516227, + "objective/train/tokens_used": 1472082400, + "theoretical_loss": 3.5547575647516227, + "tokens_seen": 1451622400 + }, + { + "epoch": 0.44, + "learning_rate": 0.00045270210631895687, + "loss": 1.5215, + "theoretical_loss": 3.5547575647516227, + "tokens_seen": 1451622400 + }, + { + "epoch": 0.44, + "learning_rate": 0.0004526860581745236, + "loss": 1.3345, + "theoretical_loss": 3.5547434401102347, + "tokens_seen": 1451687936 + }, + { + "epoch": 0.44, + "learning_rate": 0.00045267001003009027, + "loss": 1.3454, + "theoretical_loss": 3.554729316285018, + "tokens_seen": 1451753472 + }, + { + "epoch": 0.44, + "learning_rate": 0.000452653961885657, + "loss": 1.3237, + "theoretical_loss": 3.5547151932758894, + "tokens_seen": 1451819008 + }, + { + "epoch": 0.44, + "learning_rate": 0.00045263791374122367, + "loss": 1.4374, + "theoretical_loss": 3.554701071082764, + "tokens_seen": 1451884544 + }, + { + "epoch": 0.44, + "learning_rate": 0.00045262186559679043, + "loss": 1.322, + "theoretical_loss": 3.5546869497055584, + "tokens_seen": 1451950080 + }, + { + "epoch": 0.44, + "learning_rate": 0.0004526058174523571, + "loss": 1.3869, + "theoretical_loss": 3.554672829144189, + "tokens_seen": 1452015616 + }, + { + "epoch": 0.44, + "learning_rate": 0.00045258976930792383, + "loss": 1.3434, + "theoretical_loss": 3.5546587093985704, + "tokens_seen": 1452081152 + }, + { + "epoch": 0.44, + "learning_rate": 0.0004525737211634905, + "loss": 1.4269, + "theoretical_loss": 3.5546445904686204, + "tokens_seen": 1452146688 + }, + { + "epoch": 0.44, + "learning_rate": 0.00045255767301905723, + "loss": 1.4593, + "theoretical_loss": 3.554630472354254, + "tokens_seen": 1452212224 + }, + { + "epoch": 0.44, + "learning_rate": 0.0004525416248746239, + "loss": 1.312, + "theoretical_loss": 3.554616355055388, + "tokens_seen": 1452277760 + }, + { + "epoch": 0.44, + "learning_rate": 0.00045252557673019063, + "loss": 1.4225, + "theoretical_loss": 3.5546022385719374, + "tokens_seen": 1452343296 + }, + { + "epoch": 0.44, + "learning_rate": 0.0004525095285857573, + "loss": 1.3669, + "theoretical_loss": 3.5545881229038194, + "tokens_seen": 1452408832 + }, + { + "epoch": 0.44, + "learning_rate": 0.00045249348044132403, + "loss": 1.3524, + "theoretical_loss": 3.55457400805095, + "tokens_seen": 1452474368 + }, + { + "epoch": 0.44, + "learning_rate": 0.0004524774322968907, + "loss": 1.46, + "theoretical_loss": 3.554559894013245, + "tokens_seen": 1452539904 + }, + { + "epoch": 0.44, + "learning_rate": 0.00045246138415245743, + "loss": 1.4595, + "theoretical_loss": 3.5545457807906202, + "tokens_seen": 1452605440 + }, + { + "epoch": 0.44, + "learning_rate": 0.0004524453360080241, + "loss": 1.4014, + "theoretical_loss": 3.5545316683829924, + "tokens_seen": 1452670976 + }, + { + "epoch": 0.44, + "learning_rate": 0.00045242928786359083, + "loss": 1.4108, + "theoretical_loss": 3.5545175567902776, + "tokens_seen": 1452736512 + }, + { + "epoch": 0.44, + "learning_rate": 0.0004524132397191575, + "loss": 1.371, + "theoretical_loss": 3.554503446012392, + "tokens_seen": 1452802048 + }, + { + "epoch": 0.44, + "learning_rate": 0.00045239719157472424, + "loss": 1.344, + "theoretical_loss": 3.5544893360492513, + "tokens_seen": 1452867584 + }, + { + "epoch": 0.44, + "learning_rate": 0.0004523811434302909, + "loss": 1.3676, + "theoretical_loss": 3.5544752269007724, + "tokens_seen": 1452933120 + }, + { + "epoch": 0.44, + "learning_rate": 0.00045236509528585764, + "loss": 1.356, + "theoretical_loss": 3.5544611185668713, + "tokens_seen": 1452998656 + }, + { + "epoch": 0.44, + "learning_rate": 0.0004523490471414243, + "loss": 1.4534, + "theoretical_loss": 3.554447011047464, + "tokens_seen": 1453064192 + }, + { + "epoch": 0.44, + "learning_rate": 0.00045233299899699104, + "loss": 1.3569, + "theoretical_loss": 3.554432904342467, + "tokens_seen": 1453129728 + }, + { + "epoch": 0.44, + "learning_rate": 0.0004523169508525577, + "loss": 1.4058, + "theoretical_loss": 3.5544187984517968, + "tokens_seen": 1453195264 + }, + { + "epoch": 0.44, + "learning_rate": 0.00045230090270812444, + "loss": 1.3698, + "theoretical_loss": 3.5544046933753695, + "tokens_seen": 1453260800 + }, + { + "epoch": 0.44, + "learning_rate": 0.0004522848545636911, + "loss": 1.3574, + "theoretical_loss": 3.5543905891131007, + "tokens_seen": 1453326336 + }, + { + "epoch": 0.44, + "learning_rate": 0.00045226880641925784, + "loss": 1.4082, + "theoretical_loss": 3.554376485664907, + "tokens_seen": 1453391872 + }, + { + "epoch": 0.44, + "learning_rate": 0.0004522527582748245, + "loss": 1.3318, + "theoretical_loss": 3.5543623830307056, + "tokens_seen": 1453457408 + }, + { + "epoch": 0.44, + "learning_rate": 0.00045223671013039124, + "loss": 1.4165, + "theoretical_loss": 3.554348281210412, + "tokens_seen": 1453522944 + }, + { + "epoch": 0.44, + "learning_rate": 0.0004522206619859579, + "loss": 1.4581, + "theoretical_loss": 3.5543341802039423, + "tokens_seen": 1453588480 + }, + { + "epoch": 0.44, + "learning_rate": 0.0004522046138415246, + "loss": 1.3781, + "theoretical_loss": 3.5543200800112134, + "tokens_seen": 1453654016 + }, + { + "epoch": 0.44, + "learning_rate": 0.0004521885656970913, + "loss": 1.2696, + "theoretical_loss": 3.554305980632142, + "tokens_seen": 1453719552 + }, + { + "epoch": 0.44, + "learning_rate": 0.000452172517552658, + "loss": 1.3585, + "theoretical_loss": 3.554291882066644, + "tokens_seen": 1453785088 + }, + { + "epoch": 0.44, + "learning_rate": 0.0004521564694082247, + "loss": 1.4281, + "theoretical_loss": 3.554277784314635, + "tokens_seen": 1453850624 + }, + { + "epoch": 0.44, + "learning_rate": 0.0004521404212637914, + "loss": 1.3436, + "theoretical_loss": 3.5542636873760323, + "tokens_seen": 1453916160 + }, + { + "epoch": 0.44, + "learning_rate": 0.0004521243731193581, + "loss": 1.4368, + "theoretical_loss": 3.554249591250753, + "tokens_seen": 1453981696 + }, + { + "epoch": 0.44, + "learning_rate": 0.0004521083249749248, + "loss": 1.2893, + "theoretical_loss": 3.5542354959387117, + "tokens_seen": 1454047232 + }, + { + "epoch": 0.44, + "learning_rate": 0.0004520922768304915, + "loss": 1.5305, + "theoretical_loss": 3.5542214014398263, + "tokens_seen": 1454112768 + }, + { + "epoch": 0.44, + "learning_rate": 0.0004520762286860582, + "loss": 1.2654, + "theoretical_loss": 3.5542073077540124, + "tokens_seen": 1454178304 + }, + { + "epoch": 0.44, + "learning_rate": 0.00045206018054162484, + "loss": 1.3861, + "theoretical_loss": 3.5541932148811872, + "tokens_seen": 1454243840 + }, + { + "epoch": 0.44, + "learning_rate": 0.0004520441323971916, + "loss": 1.5466, + "theoretical_loss": 3.5541791228212665, + "tokens_seen": 1454309376 + }, + { + "epoch": 0.44, + "learning_rate": 0.00045202808425275824, + "loss": 1.4488, + "theoretical_loss": 3.554165031574167, + "tokens_seen": 1454374912 + }, + { + "epoch": 0.44, + "learning_rate": 0.000452012036108325, + "loss": 1.306, + "theoretical_loss": 3.5541509411398056, + "tokens_seen": 1454440448 + }, + { + "epoch": 0.44, + "learning_rate": 0.00045199598796389164, + "loss": 1.3654, + "theoretical_loss": 3.5541368515180984, + "tokens_seen": 1454505984 + }, + { + "epoch": 0.44, + "learning_rate": 0.0004519799398194584, + "loss": 1.4374, + "theoretical_loss": 3.5541227627089613, + "tokens_seen": 1454571520 + }, + { + "epoch": 0.44, + "learning_rate": 0.00045196389167502504, + "loss": 1.3379, + "theoretical_loss": 3.554108674712312, + "tokens_seen": 1454637056 + }, + { + "epoch": 0.44, + "learning_rate": 0.0004519478435305918, + "loss": 1.3888, + "theoretical_loss": 3.554094587528067, + "tokens_seen": 1454702592 + }, + { + "epoch": 0.44, + "learning_rate": 0.00045193179538615845, + "loss": 1.36, + "theoretical_loss": 3.5540805011561423, + "tokens_seen": 1454768128 + }, + { + "epoch": 0.44, + "learning_rate": 0.0004519157472417252, + "loss": 1.3894, + "theoretical_loss": 3.554066415596454, + "tokens_seen": 1454833664 + }, + { + "epoch": 0.44, + "objective/train/docs_used": 531642, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.44892156124115, + "objective/train/theoretical_loss": 3.5540523308489194, + "objective/train/tokens_used": 1475359200, + "theoretical_loss": 3.5540523308489194, + "tokens_seen": 1454899200 + }, + { + "epoch": 0.44, + "learning_rate": 0.00045189969909729185, + "loss": 1.335, + "theoretical_loss": 3.5540523308489194, + "tokens_seen": 1454899200 + }, + { + "epoch": 0.44, + "learning_rate": 0.0004518836509528586, + "loss": 1.3849, + "theoretical_loss": 3.5540382469134553, + "tokens_seen": 1454964736 + }, + { + "epoch": 0.44, + "learning_rate": 0.00045186760280842525, + "loss": 1.294, + "theoretical_loss": 3.5540241637899777, + "tokens_seen": 1455030272 + }, + { + "epoch": 0.44, + "learning_rate": 0.000451851554663992, + "loss": 1.4247, + "theoretical_loss": 3.5540100814784035, + "tokens_seen": 1455095808 + }, + { + "epoch": 0.44, + "learning_rate": 0.00045183550651955865, + "loss": 1.4014, + "theoretical_loss": 3.553995999978649, + "tokens_seen": 1455161344 + }, + { + "epoch": 0.44, + "learning_rate": 0.0004518194583751254, + "loss": 1.33, + "theoretical_loss": 3.5539819192906315, + "tokens_seen": 1455226880 + }, + { + "epoch": 0.44, + "learning_rate": 0.00045180341023069205, + "loss": 1.3032, + "theoretical_loss": 3.553967839414267, + "tokens_seen": 1455292416 + }, + { + "epoch": 0.44, + "learning_rate": 0.0004517873620862588, + "loss": 1.3348, + "theoretical_loss": 3.5539537603494726, + "tokens_seen": 1455357952 + }, + { + "epoch": 0.44, + "learning_rate": 0.00045177131394182545, + "loss": 1.5433, + "theoretical_loss": 3.553939682096165, + "tokens_seen": 1455423488 + }, + { + "epoch": 0.44, + "learning_rate": 0.0004517552657973922, + "loss": 1.2203, + "theoretical_loss": 3.5539256046542604, + "tokens_seen": 1455489024 + }, + { + "epoch": 0.44, + "learning_rate": 0.00045173921765295885, + "loss": 1.4394, + "theoretical_loss": 3.553911528023676, + "tokens_seen": 1455554560 + }, + { + "epoch": 0.44, + "learning_rate": 0.0004517231695085256, + "loss": 1.3444, + "theoretical_loss": 3.5538974522043283, + "tokens_seen": 1455620096 + }, + { + "epoch": 0.44, + "learning_rate": 0.00045170712136409226, + "loss": 1.3655, + "theoretical_loss": 3.553883377196134, + "tokens_seen": 1455685632 + }, + { + "epoch": 0.44, + "learning_rate": 0.000451691073219659, + "loss": 1.3716, + "theoretical_loss": 3.55386930299901, + "tokens_seen": 1455751168 + }, + { + "epoch": 0.44, + "learning_rate": 0.00045167502507522566, + "loss": 1.3258, + "theoretical_loss": 3.553855229612873, + "tokens_seen": 1455816704 + }, + { + "epoch": 0.44, + "learning_rate": 0.0004516589769307924, + "loss": 1.3747, + "theoretical_loss": 3.55384115703764, + "tokens_seen": 1455882240 + }, + { + "epoch": 0.44, + "learning_rate": 0.00045164292878635917, + "loss": 1.358, + "theoretical_loss": 3.5538270852732268, + "tokens_seen": 1455947776 + }, + { + "epoch": 0.44, + "learning_rate": 0.0004516268806419258, + "loss": 1.4858, + "theoretical_loss": 3.553813014319551, + "tokens_seen": 1456013312 + }, + { + "epoch": 0.44, + "learning_rate": 0.00045161083249749257, + "loss": 1.3412, + "theoretical_loss": 3.55379894417653, + "tokens_seen": 1456078848 + }, + { + "epoch": 0.44, + "learning_rate": 0.0004515947843530592, + "loss": 1.3311, + "theoretical_loss": 3.5537848748440792, + "tokens_seen": 1456144384 + }, + { + "epoch": 0.44, + "learning_rate": 0.0004515787362086259, + "loss": 1.3063, + "theoretical_loss": 3.5537708063221167, + "tokens_seen": 1456209920 + }, + { + "epoch": 0.44, + "learning_rate": 0.0004515626880641926, + "loss": 1.3215, + "theoretical_loss": 3.5537567386105584, + "tokens_seen": 1456275456 + }, + { + "epoch": 0.44, + "learning_rate": 0.0004515466399197593, + "loss": 1.3621, + "theoretical_loss": 3.5537426717093217, + "tokens_seen": 1456340992 + }, + { + "epoch": 0.44, + "learning_rate": 0.000451530591775326, + "loss": 1.2836, + "theoretical_loss": 3.553728605618323, + "tokens_seen": 1456406528 + }, + { + "epoch": 0.44, + "learning_rate": 0.0004515145436308927, + "loss": 1.3567, + "theoretical_loss": 3.5537145403374795, + "tokens_seen": 1456472064 + }, + { + "epoch": 0.44, + "learning_rate": 0.0004514984954864594, + "loss": 1.4283, + "theoretical_loss": 3.553700475866708, + "tokens_seen": 1456537600 + }, + { + "epoch": 0.44, + "learning_rate": 0.0004514824473420261, + "loss": 1.3473, + "theoretical_loss": 3.5536864122059257, + "tokens_seen": 1456603136 + }, + { + "epoch": 0.44, + "learning_rate": 0.0004514663991975928, + "loss": 1.3145, + "theoretical_loss": 3.5536723493550495, + "tokens_seen": 1456668672 + }, + { + "epoch": 0.44, + "learning_rate": 0.0004514503510531595, + "loss": 1.3375, + "theoretical_loss": 3.5536582873139952, + "tokens_seen": 1456734208 + }, + { + "epoch": 0.44, + "learning_rate": 0.00045143430290872617, + "loss": 1.4671, + "theoretical_loss": 3.5536442260826813, + "tokens_seen": 1456799744 + }, + { + "epoch": 0.44, + "learning_rate": 0.0004514182547642929, + "loss": 1.544, + "theoretical_loss": 3.5536301656610236, + "tokens_seen": 1456865280 + }, + { + "epoch": 0.44, + "learning_rate": 0.00045140220661985957, + "loss": 1.4047, + "theoretical_loss": 3.55361610604894, + "tokens_seen": 1456930816 + }, + { + "epoch": 0.44, + "learning_rate": 0.0004513861584754263, + "loss": 1.4867, + "theoretical_loss": 3.553602047246347, + "tokens_seen": 1456996352 + }, + { + "epoch": 0.44, + "learning_rate": 0.00045137011033099297, + "loss": 1.341, + "theoretical_loss": 3.5535879892531614, + "tokens_seen": 1457061888 + }, + { + "epoch": 0.44, + "learning_rate": 0.0004513540621865597, + "loss": 1.2838, + "theoretical_loss": 3.5535739320693, + "tokens_seen": 1457127424 + }, + { + "epoch": 0.44, + "learning_rate": 0.00045133801404212637, + "loss": 1.5017, + "theoretical_loss": 3.5535598756946807, + "tokens_seen": 1457192960 + }, + { + "epoch": 0.44, + "learning_rate": 0.0004513219658976931, + "loss": 1.4428, + "theoretical_loss": 3.55354582012922, + "tokens_seen": 1457258496 + }, + { + "epoch": 0.44, + "learning_rate": 0.00045130591775325977, + "loss": 1.4469, + "theoretical_loss": 3.553531765372835, + "tokens_seen": 1457324032 + }, + { + "epoch": 0.44, + "learning_rate": 0.0004512898696088265, + "loss": 1.3217, + "theoretical_loss": 3.5535177114254424, + "tokens_seen": 1457389568 + }, + { + "epoch": 0.44, + "learning_rate": 0.00045127382146439317, + "loss": 1.3078, + "theoretical_loss": 3.5535036582869592, + "tokens_seen": 1457455104 + }, + { + "epoch": 0.44, + "learning_rate": 0.0004512577733199599, + "loss": 1.4349, + "theoretical_loss": 3.5534896059573033, + "tokens_seen": 1457520640 + }, + { + "epoch": 0.44, + "learning_rate": 0.0004512417251755266, + "loss": 1.3537, + "theoretical_loss": 3.5534755544363916, + "tokens_seen": 1457586176 + }, + { + "epoch": 0.44, + "learning_rate": 0.00045122567703109333, + "loss": 1.4282, + "theoretical_loss": 3.55346150372414, + "tokens_seen": 1457651712 + }, + { + "epoch": 0.44, + "learning_rate": 0.00045120962888666, + "loss": 1.425, + "theoretical_loss": 3.5534474538204672, + "tokens_seen": 1457717248 + }, + { + "epoch": 0.44, + "learning_rate": 0.00045119358074222673, + "loss": 1.5104, + "theoretical_loss": 3.55343340472529, + "tokens_seen": 1457782784 + }, + { + "epoch": 0.44, + "learning_rate": 0.0004511775325977934, + "loss": 1.5005, + "theoretical_loss": 3.553419356438524, + "tokens_seen": 1457848320 + }, + { + "epoch": 0.44, + "learning_rate": 0.00045116148445336013, + "loss": 1.4491, + "theoretical_loss": 3.5534053089600883, + "tokens_seen": 1457913856 + }, + { + "epoch": 0.44, + "learning_rate": 0.0004511454363089268, + "loss": 1.4182, + "theoretical_loss": 3.553391262289899, + "tokens_seen": 1457979392 + }, + { + "epoch": 0.44, + "learning_rate": 0.00045112938816449353, + "loss": 1.3173, + "theoretical_loss": 3.5533772164278736, + "tokens_seen": 1458044928 + }, + { + "epoch": 0.44, + "learning_rate": 0.0004511133400200602, + "loss": 1.4196, + "theoretical_loss": 3.5533631713739293, + "tokens_seen": 1458110464 + }, + { + "epoch": 0.44, + "objective/train/docs_used": 532886, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.29374361038208, + "objective/train/theoretical_loss": 3.553349127127983, + "objective/train/tokens_used": 1478636000, + "theoretical_loss": 3.553349127127983, + "tokens_seen": 1458176000 + }, + { + "epoch": 0.44, + "learning_rate": 0.00045109729187562693, + "loss": 1.3756, + "theoretical_loss": 3.553349127127983, + "tokens_seen": 1458176000 + }, + { + "epoch": 0.44, + "learning_rate": 0.0004510812437311936, + "loss": 1.4015, + "theoretical_loss": 3.5533350836899524, + "tokens_seen": 1458241536 + }, + { + "epoch": 0.44, + "learning_rate": 0.00045106519558676033, + "loss": 1.2445, + "theoretical_loss": 3.553321041059754, + "tokens_seen": 1458307072 + }, + { + "epoch": 0.44, + "learning_rate": 0.000451049147442327, + "loss": 1.3986, + "theoretical_loss": 3.553306999237306, + "tokens_seen": 1458372608 + }, + { + "epoch": 0.44, + "learning_rate": 0.00045103309929789374, + "loss": 1.4912, + "theoretical_loss": 3.5532929582225243, + "tokens_seen": 1458438144 + }, + { + "epoch": 0.44, + "learning_rate": 0.0004510170511534604, + "loss": 1.4658, + "theoretical_loss": 3.5532789180153275, + "tokens_seen": 1458503680 + }, + { + "epoch": 0.44, + "learning_rate": 0.00045100100300902714, + "loss": 1.3169, + "theoretical_loss": 3.5532648786156322, + "tokens_seen": 1458569216 + }, + { + "epoch": 0.44, + "learning_rate": 0.0004509849548645938, + "loss": 1.4373, + "theoretical_loss": 3.553250840023356, + "tokens_seen": 1458634752 + }, + { + "epoch": 0.44, + "learning_rate": 0.00045096890672016054, + "loss": 1.3086, + "theoretical_loss": 3.553236802238416, + "tokens_seen": 1458700288 + }, + { + "epoch": 0.44, + "learning_rate": 0.0004509528585757272, + "loss": 1.44, + "theoretical_loss": 3.5532227652607293, + "tokens_seen": 1458765824 + }, + { + "epoch": 0.44, + "learning_rate": 0.00045093681043129394, + "loss": 1.3668, + "theoretical_loss": 3.5532087290902132, + "tokens_seen": 1458831360 + }, + { + "epoch": 0.44, + "learning_rate": 0.0004509207622868606, + "loss": 1.368, + "theoretical_loss": 3.5531946937267858, + "tokens_seen": 1458896896 + }, + { + "epoch": 0.44, + "learning_rate": 0.00045090471414242734, + "loss": 1.3399, + "theoretical_loss": 3.553180659170363, + "tokens_seen": 1458962432 + }, + { + "epoch": 0.44, + "learning_rate": 0.000450888665997994, + "loss": 1.3927, + "theoretical_loss": 3.5531666254208636, + "tokens_seen": 1459027968 + }, + { + "epoch": 0.44, + "learning_rate": 0.00045087261785356074, + "loss": 1.3636, + "theoretical_loss": 3.5531525924782046, + "tokens_seen": 1459093504 + }, + { + "epoch": 0.44, + "learning_rate": 0.0004508565697091274, + "loss": 1.445, + "theoretical_loss": 3.553138560342303, + "tokens_seen": 1459159040 + }, + { + "epoch": 0.44, + "learning_rate": 0.00045084052156469414, + "loss": 1.2998, + "theoretical_loss": 3.553124529013076, + "tokens_seen": 1459224576 + }, + { + "epoch": 0.44, + "learning_rate": 0.0004508244734202608, + "loss": 1.2876, + "theoretical_loss": 3.5531104984904416, + "tokens_seen": 1459290112 + }, + { + "epoch": 0.44, + "learning_rate": 0.0004508084252758275, + "loss": 1.5277, + "theoretical_loss": 3.5530964687743167, + "tokens_seen": 1459355648 + }, + { + "epoch": 0.44, + "learning_rate": 0.0004507923771313942, + "loss": 1.3271, + "theoretical_loss": 3.5530824398646192, + "tokens_seen": 1459421184 + }, + { + "epoch": 0.44, + "learning_rate": 0.0004507763289869609, + "loss": 1.5107, + "theoretical_loss": 3.5530684117612665, + "tokens_seen": 1459486720 + }, + { + "epoch": 0.44, + "learning_rate": 0.0004507602808425276, + "loss": 1.4708, + "theoretical_loss": 3.5530543844641755, + "tokens_seen": 1459552256 + }, + { + "epoch": 0.44, + "learning_rate": 0.0004507442326980943, + "loss": 1.3377, + "theoretical_loss": 3.5530403579732646, + "tokens_seen": 1459617792 + }, + { + "epoch": 0.44, + "learning_rate": 0.000450728184553661, + "loss": 1.3486, + "theoretical_loss": 3.55302633228845, + "tokens_seen": 1459683328 + }, + { + "epoch": 0.44, + "learning_rate": 0.0004507121364092277, + "loss": 1.469, + "theoretical_loss": 3.5530123074096505, + "tokens_seen": 1459748864 + }, + { + "epoch": 0.44, + "learning_rate": 0.0004506960882647944, + "loss": 1.4295, + "theoretical_loss": 3.552998283336782, + "tokens_seen": 1459814400 + }, + { + "epoch": 0.44, + "learning_rate": 0.0004506800401203611, + "loss": 1.3064, + "theoretical_loss": 3.5529842600697643, + "tokens_seen": 1459879936 + }, + { + "epoch": 0.44, + "learning_rate": 0.00045066399197592774, + "loss": 1.4402, + "theoretical_loss": 3.5529702376085126, + "tokens_seen": 1459945472 + }, + { + "epoch": 0.44, + "learning_rate": 0.0004506479438314945, + "loss": 1.2826, + "theoretical_loss": 3.552956215952946, + "tokens_seen": 1460011008 + }, + { + "epoch": 0.44, + "learning_rate": 0.00045063189568706114, + "loss": 1.4397, + "theoretical_loss": 3.5529421951029816, + "tokens_seen": 1460076544 + }, + { + "epoch": 0.44, + "learning_rate": 0.0004506158475426279, + "loss": 1.4209, + "theoretical_loss": 3.5529281750585366, + "tokens_seen": 1460142080 + }, + { + "epoch": 0.44, + "learning_rate": 0.00045059979939819454, + "loss": 1.2923, + "theoretical_loss": 3.5529141558195287, + "tokens_seen": 1460207616 + }, + { + "epoch": 0.44, + "learning_rate": 0.0004505837512537613, + "loss": 1.4391, + "theoretical_loss": 3.552900137385876, + "tokens_seen": 1460273152 + }, + { + "epoch": 0.44, + "learning_rate": 0.00045056770310932795, + "loss": 1.4198, + "theoretical_loss": 3.5528861197574955, + "tokens_seen": 1460338688 + }, + { + "epoch": 0.44, + "learning_rate": 0.0004505516549648947, + "loss": 1.4445, + "theoretical_loss": 3.552872102934305, + "tokens_seen": 1460404224 + }, + { + "epoch": 0.44, + "learning_rate": 0.00045053560682046135, + "loss": 1.3768, + "theoretical_loss": 3.552858086916222, + "tokens_seen": 1460469760 + }, + { + "epoch": 0.44, + "learning_rate": 0.0004505195586760281, + "loss": 1.3215, + "theoretical_loss": 3.552844071703164, + "tokens_seen": 1460535296 + }, + { + "epoch": 0.44, + "learning_rate": 0.00045050351053159475, + "loss": 1.4088, + "theoretical_loss": 3.5528300572950497, + "tokens_seen": 1460600832 + }, + { + "epoch": 0.44, + "learning_rate": 0.0004504874623871615, + "loss": 1.4711, + "theoretical_loss": 3.5528160436917955, + "tokens_seen": 1460666368 + }, + { + "epoch": 0.44, + "learning_rate": 0.00045047141424272815, + "loss": 1.3606, + "theoretical_loss": 3.5528020308933197, + "tokens_seen": 1460731904 + }, + { + "epoch": 0.44, + "learning_rate": 0.0004504553660982949, + "loss": 1.2942, + "theoretical_loss": 3.5527880188995393, + "tokens_seen": 1460797440 + }, + { + "epoch": 0.44, + "learning_rate": 0.00045043931795386166, + "loss": 1.3563, + "theoretical_loss": 3.552774007710373, + "tokens_seen": 1460862976 + }, + { + "epoch": 0.44, + "learning_rate": 0.0004504232698094283, + "loss": 1.5251, + "theoretical_loss": 3.5527599973257376, + "tokens_seen": 1460928512 + }, + { + "epoch": 0.44, + "learning_rate": 0.00045040722166499506, + "loss": 1.4414, + "theoretical_loss": 3.552745987745552, + "tokens_seen": 1460994048 + }, + { + "epoch": 0.44, + "learning_rate": 0.0004503911735205617, + "loss": 1.4615, + "theoretical_loss": 3.5527319789697325, + "tokens_seen": 1461059584 + }, + { + "epoch": 0.44, + "learning_rate": 0.00045037512537612846, + "loss": 1.4363, + "theoretical_loss": 3.5527179709981973, + "tokens_seen": 1461125120 + }, + { + "epoch": 0.44, + "learning_rate": 0.0004503590772316951, + "loss": 1.3551, + "theoretical_loss": 3.5527039638308646, + "tokens_seen": 1461190656 + }, + { + "epoch": 0.44, + "learning_rate": 0.00045034302908726186, + "loss": 1.303, + "theoretical_loss": 3.552689957467652, + "tokens_seen": 1461256192 + }, + { + "epoch": 0.44, + "learning_rate": 0.0004503269809428285, + "loss": 1.455, + "theoretical_loss": 3.552675951908477, + "tokens_seen": 1461321728 + }, + { + "epoch": 0.44, + "learning_rate": 0.00045031093279839526, + "loss": 1.3285, + "theoretical_loss": 3.5526619471532577, + "tokens_seen": 1461387264 + }, + { + "epoch": 0.44, + "objective/train/docs_used": 534190, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.367629051208496, + "objective/train/theoretical_loss": 3.552647943201912, + "objective/train/tokens_used": 1481912800, + "theoretical_loss": 3.552647943201912, + "tokens_seen": 1461452800 + }, + { + "epoch": 0.44, + "learning_rate": 0.0004502948846539619, + "loss": 1.4044, + "theoretical_loss": 3.552647943201912, + "tokens_seen": 1461452800 + }, + { + "epoch": 0.44, + "learning_rate": 0.00045027883650952867, + "loss": 1.4028, + "theoretical_loss": 3.552633940054357, + "tokens_seen": 1461518336 + }, + { + "epoch": 0.44, + "learning_rate": 0.0004502627883650953, + "loss": 1.3917, + "theoretical_loss": 3.5526199377105114, + "tokens_seen": 1461583872 + }, + { + "epoch": 0.44, + "learning_rate": 0.00045024674022066207, + "loss": 1.3401, + "theoretical_loss": 3.552605936170292, + "tokens_seen": 1461649408 + }, + { + "epoch": 0.44, + "learning_rate": 0.0004502306920762287, + "loss": 1.3363, + "theoretical_loss": 3.552591935433618, + "tokens_seen": 1461714944 + }, + { + "epoch": 0.44, + "learning_rate": 0.00045021464393179547, + "loss": 1.3422, + "theoretical_loss": 3.5525779355004063, + "tokens_seen": 1461780480 + }, + { + "epoch": 0.44, + "learning_rate": 0.0004501985957873621, + "loss": 1.3502, + "theoretical_loss": 3.552563936370575, + "tokens_seen": 1461846016 + }, + { + "epoch": 0.44, + "learning_rate": 0.0004501825476429288, + "loss": 1.3773, + "theoretical_loss": 3.5525499380440424, + "tokens_seen": 1461911552 + }, + { + "epoch": 0.44, + "learning_rate": 0.0004501664994984955, + "loss": 1.4123, + "theoretical_loss": 3.5525359405207255, + "tokens_seen": 1461977088 + }, + { + "epoch": 0.44, + "learning_rate": 0.0004501504513540622, + "loss": 1.4218, + "theoretical_loss": 3.552521943800543, + "tokens_seen": 1462042624 + }, + { + "epoch": 0.44, + "learning_rate": 0.0004501344032096289, + "loss": 1.4182, + "theoretical_loss": 3.552507947883412, + "tokens_seen": 1462108160 + }, + { + "epoch": 0.44, + "learning_rate": 0.0004501183550651956, + "loss": 1.4285, + "theoretical_loss": 3.552493952769252, + "tokens_seen": 1462173696 + }, + { + "epoch": 0.44, + "learning_rate": 0.0004501023069207623, + "loss": 1.4321, + "theoretical_loss": 3.552479958457979, + "tokens_seen": 1462239232 + }, + { + "epoch": 0.44, + "learning_rate": 0.000450086258776329, + "loss": 1.3602, + "theoretical_loss": 3.552465964949512, + "tokens_seen": 1462304768 + }, + { + "epoch": 0.44, + "learning_rate": 0.0004500702106318957, + "loss": 1.3515, + "theoretical_loss": 3.5524519722437695, + "tokens_seen": 1462370304 + }, + { + "epoch": 0.44, + "learning_rate": 0.0004500541624874624, + "loss": 1.4833, + "theoretical_loss": 3.5524379803406685, + "tokens_seen": 1462435840 + }, + { + "epoch": 0.44, + "learning_rate": 0.00045003811434302907, + "loss": 1.4451, + "theoretical_loss": 3.552423989240127, + "tokens_seen": 1462501376 + }, + { + "epoch": 0.44, + "learning_rate": 0.0004500220661985958, + "loss": 1.2591, + "theoretical_loss": 3.552409998942063, + "tokens_seen": 1462566912 + }, + { + "epoch": 0.44, + "learning_rate": 0.00045000601805416247, + "loss": 1.3823, + "theoretical_loss": 3.5523960094463956, + "tokens_seen": 1462632448 + }, + { + "epoch": 0.44, + "learning_rate": 0.0004499899699097292, + "loss": 1.4574, + "theoretical_loss": 3.5523820207530417, + "tokens_seen": 1462697984 + }, + { + "epoch": 0.44, + "learning_rate": 0.00044997392176529587, + "loss": 1.382, + "theoretical_loss": 3.55236803286192, + "tokens_seen": 1462763520 + }, + { + "epoch": 0.44, + "learning_rate": 0.0004499578736208626, + "loss": 1.4654, + "theoretical_loss": 3.552354045772948, + "tokens_seen": 1462829056 + }, + { + "epoch": 0.44, + "learning_rate": 0.00044994182547642927, + "loss": 1.3672, + "theoretical_loss": 3.5523400594860437, + "tokens_seen": 1462894592 + }, + { + "epoch": 0.44, + "learning_rate": 0.000449925777331996, + "loss": 1.3712, + "theoretical_loss": 3.552326074001126, + "tokens_seen": 1462960128 + }, + { + "epoch": 0.44, + "learning_rate": 0.00044990972918756267, + "loss": 1.3293, + "theoretical_loss": 3.5523120893181126, + "tokens_seen": 1463025664 + }, + { + "epoch": 0.44, + "learning_rate": 0.0004498936810431294, + "loss": 1.4439, + "theoretical_loss": 3.5522981054369214, + "tokens_seen": 1463091200 + }, + { + "epoch": 0.44, + "learning_rate": 0.0004498776328986961, + "loss": 1.4734, + "theoretical_loss": 3.5522841223574706, + "tokens_seen": 1463156736 + }, + { + "epoch": 0.44, + "learning_rate": 0.00044986158475426283, + "loss": 1.3922, + "theoretical_loss": 3.552270140079678, + "tokens_seen": 1463222272 + }, + { + "epoch": 0.44, + "learning_rate": 0.0004498455366098295, + "loss": 1.3927, + "theoretical_loss": 3.5522561586034627, + "tokens_seen": 1463287808 + }, + { + "epoch": 0.44, + "learning_rate": 0.00044982948846539623, + "loss": 1.3559, + "theoretical_loss": 3.5522421779287416, + "tokens_seen": 1463353344 + }, + { + "epoch": 0.44, + "learning_rate": 0.0004498134403209629, + "loss": 1.3277, + "theoretical_loss": 3.5522281980554338, + "tokens_seen": 1463418880 + }, + { + "epoch": 0.44, + "learning_rate": 0.00044979739217652963, + "loss": 1.4046, + "theoretical_loss": 3.5522142189834574, + "tokens_seen": 1463484416 + }, + { + "epoch": 0.44, + "learning_rate": 0.0004497813440320963, + "loss": 1.4233, + "theoretical_loss": 3.5522002407127298, + "tokens_seen": 1463549952 + }, + { + "epoch": 0.44, + "learning_rate": 0.00044976529588766303, + "loss": 1.3505, + "theoretical_loss": 3.55218626324317, + "tokens_seen": 1463615488 + }, + { + "epoch": 0.44, + "learning_rate": 0.0004497492477432297, + "loss": 1.3417, + "theoretical_loss": 3.552172286574696, + "tokens_seen": 1463681024 + }, + { + "epoch": 0.44, + "learning_rate": 0.00044973319959879643, + "loss": 1.3679, + "theoretical_loss": 3.5521583107072265, + "tokens_seen": 1463746560 + }, + { + "epoch": 0.44, + "learning_rate": 0.0004497171514543631, + "loss": 1.3133, + "theoretical_loss": 3.5521443356406786, + "tokens_seen": 1463812096 + }, + { + "epoch": 0.44, + "learning_rate": 0.00044970110330992983, + "loss": 1.3237, + "theoretical_loss": 3.552130361374971, + "tokens_seen": 1463877632 + }, + { + "epoch": 0.44, + "learning_rate": 0.0004496850551654965, + "loss": 1.3901, + "theoretical_loss": 3.5521163879100226, + "tokens_seen": 1463943168 + }, + { + "epoch": 0.44, + "learning_rate": 0.00044966900702106324, + "loss": 1.3783, + "theoretical_loss": 3.552102415245751, + "tokens_seen": 1464008704 + }, + { + "epoch": 0.44, + "learning_rate": 0.0004496529588766299, + "loss": 1.3664, + "theoretical_loss": 3.5520884433820745, + "tokens_seen": 1464074240 + }, + { + "epoch": 0.44, + "learning_rate": 0.00044963691073219664, + "loss": 1.4415, + "theoretical_loss": 3.552074472318912, + "tokens_seen": 1464139776 + }, + { + "epoch": 0.44, + "learning_rate": 0.0004496208625877633, + "loss": 1.2891, + "theoretical_loss": 3.552060502056181, + "tokens_seen": 1464205312 + }, + { + "epoch": 0.44, + "learning_rate": 0.00044960481444333004, + "loss": 1.2716, + "theoretical_loss": 3.5520465325938004, + "tokens_seen": 1464270848 + }, + { + "epoch": 0.44, + "learning_rate": 0.0004495887662988967, + "loss": 1.3842, + "theoretical_loss": 3.5520325639316885, + "tokens_seen": 1464336384 + }, + { + "epoch": 0.44, + "learning_rate": 0.00044957271815446344, + "loss": 1.3881, + "theoretical_loss": 3.552018596069763, + "tokens_seen": 1464401920 + }, + { + "epoch": 0.44, + "learning_rate": 0.0004495566700100301, + "loss": 1.4341, + "theoretical_loss": 3.552004629007943, + "tokens_seen": 1464467456 + }, + { + "epoch": 0.44, + "learning_rate": 0.00044954062186559684, + "loss": 1.38, + "theoretical_loss": 3.551990662746147, + "tokens_seen": 1464532992 + }, + { + "epoch": 0.44, + "learning_rate": 0.0004495245737211635, + "loss": 1.4465, + "theoretical_loss": 3.5519766972842923, + "tokens_seen": 1464598528 + }, + { + "epoch": 0.44, + "learning_rate": 0.00044950852557673024, + "loss": 1.4357, + "theoretical_loss": 3.5519627326222976, + "tokens_seen": 1464664064 + }, + { + "epoch": 0.44, + "objective/train/docs_used": 534938, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.3610155582427979, + "objective/train/theoretical_loss": 3.5519487687600826, + "objective/train/tokens_used": 1485189600, + "theoretical_loss": 3.5519487687600826, + "tokens_seen": 1464729600 + }, + { + "epoch": 0.44, + "learning_rate": 0.0004494924774322969, + "loss": 1.3553, + "theoretical_loss": 3.5519487687600826, + "tokens_seen": 1464729600 + }, + { + "epoch": 0.44, + "learning_rate": 0.00044947642928786364, + "loss": 1.4139, + "theoretical_loss": 3.5519348056975644, + "tokens_seen": 1464795136 + }, + { + "epoch": 0.44, + "learning_rate": 0.0004494603811434303, + "loss": 1.3832, + "theoretical_loss": 3.5519208434346616, + "tokens_seen": 1464860672 + }, + { + "epoch": 0.44, + "learning_rate": 0.00044944433299899704, + "loss": 1.3163, + "theoretical_loss": 3.551906881971293, + "tokens_seen": 1464926208 + }, + { + "epoch": 0.44, + "learning_rate": 0.0004494282848545637, + "loss": 1.4311, + "theoretical_loss": 3.551892921307377, + "tokens_seen": 1464991744 + }, + { + "epoch": 0.44, + "learning_rate": 0.0004494122367101304, + "loss": 1.3956, + "theoretical_loss": 3.551878961442832, + "tokens_seen": 1465057280 + }, + { + "epoch": 0.44, + "learning_rate": 0.0004493961885656971, + "loss": 1.405, + "theoretical_loss": 3.5518650023775766, + "tokens_seen": 1465122816 + }, + { + "epoch": 0.44, + "learning_rate": 0.0004493801404212638, + "loss": 1.2952, + "theoretical_loss": 3.5518510441115287, + "tokens_seen": 1465188352 + }, + { + "epoch": 0.44, + "learning_rate": 0.0004493640922768305, + "loss": 1.4321, + "theoretical_loss": 3.551837086644607, + "tokens_seen": 1465253888 + }, + { + "epoch": 0.44, + "learning_rate": 0.0004493480441323972, + "loss": 1.4602, + "theoretical_loss": 3.5518231299767304, + "tokens_seen": 1465319424 + }, + { + "epoch": 0.44, + "learning_rate": 0.0004493319959879639, + "loss": 1.3148, + "theoretical_loss": 3.5518091741078175, + "tokens_seen": 1465384960 + }, + { + "epoch": 0.44, + "learning_rate": 0.0004493159478435306, + "loss": 1.4686, + "theoretical_loss": 3.5517952190377864, + "tokens_seen": 1465450496 + }, + { + "epoch": 0.44, + "learning_rate": 0.0004492998996990973, + "loss": 1.4446, + "theoretical_loss": 3.5517812647665563, + "tokens_seen": 1465516032 + }, + { + "epoch": 0.44, + "learning_rate": 0.000449283851554664, + "loss": 1.4925, + "theoretical_loss": 3.5517673112940447, + "tokens_seen": 1465581568 + }, + { + "epoch": 0.44, + "learning_rate": 0.00044926780341023064, + "loss": 1.3367, + "theoretical_loss": 3.5517533586201706, + "tokens_seen": 1465647104 + }, + { + "epoch": 0.44, + "learning_rate": 0.0004492517552657974, + "loss": 1.3415, + "theoretical_loss": 3.5517394067448533, + "tokens_seen": 1465712640 + }, + { + "epoch": 0.44, + "learning_rate": 0.00044923570712136415, + "loss": 1.4254, + "theoretical_loss": 3.5517254556680102, + "tokens_seen": 1465778176 + }, + { + "epoch": 0.44, + "learning_rate": 0.0004492196589769308, + "loss": 1.3077, + "theoretical_loss": 3.5517115053895614, + "tokens_seen": 1465843712 + }, + { + "epoch": 0.44, + "learning_rate": 0.00044920361083249755, + "loss": 1.4796, + "theoretical_loss": 3.5516975559094237, + "tokens_seen": 1465909248 + }, + { + "epoch": 0.44, + "learning_rate": 0.0004491875626880642, + "loss": 1.4817, + "theoretical_loss": 3.551683607227517, + "tokens_seen": 1465974784 + }, + { + "epoch": 0.44, + "learning_rate": 0.00044917151454363096, + "loss": 1.3447, + "theoretical_loss": 3.55166965934376, + "tokens_seen": 1466040320 + }, + { + "epoch": 0.44, + "learning_rate": 0.0004491554663991976, + "loss": 1.4302, + "theoretical_loss": 3.5516557122580705, + "tokens_seen": 1466105856 + }, + { + "epoch": 0.44, + "learning_rate": 0.00044913941825476436, + "loss": 1.3123, + "theoretical_loss": 3.551641765970368, + "tokens_seen": 1466171392 + }, + { + "epoch": 0.44, + "learning_rate": 0.000449123370110331, + "loss": 1.27, + "theoretical_loss": 3.5516278204805705, + "tokens_seen": 1466236928 + }, + { + "epoch": 0.44, + "learning_rate": 0.00044910732196589776, + "loss": 1.3958, + "theoretical_loss": 3.551613875788597, + "tokens_seen": 1466302464 + }, + { + "epoch": 0.44, + "learning_rate": 0.0004490912738214644, + "loss": 1.366, + "theoretical_loss": 3.5515999318943665, + "tokens_seen": 1466368000 + }, + { + "epoch": 0.44, + "learning_rate": 0.00044907522567703116, + "loss": 1.3252, + "theoretical_loss": 3.551585988797797, + "tokens_seen": 1466433536 + }, + { + "epoch": 0.44, + "learning_rate": 0.0004490591775325978, + "loss": 1.4153, + "theoretical_loss": 3.5515720464988076, + "tokens_seen": 1466499072 + }, + { + "epoch": 0.44, + "learning_rate": 0.00044904312938816456, + "loss": 1.4256, + "theoretical_loss": 3.551558104997318, + "tokens_seen": 1466564608 + }, + { + "epoch": 0.44, + "learning_rate": 0.0004490270812437312, + "loss": 1.2996, + "theoretical_loss": 3.551544164293245, + "tokens_seen": 1466630144 + }, + { + "epoch": 0.44, + "learning_rate": 0.00044901103309929796, + "loss": 1.4182, + "theoretical_loss": 3.551530224386509, + "tokens_seen": 1466695680 + }, + { + "epoch": 0.44, + "learning_rate": 0.0004489949849548646, + "loss": 1.3062, + "theoretical_loss": 3.551516285277028, + "tokens_seen": 1466761216 + }, + { + "epoch": 0.44, + "learning_rate": 0.00044897893681043136, + "loss": 1.4926, + "theoretical_loss": 3.5515023469647207, + "tokens_seen": 1466826752 + }, + { + "epoch": 0.44, + "learning_rate": 0.000448962888665998, + "loss": 1.4686, + "theoretical_loss": 3.5514884094495063, + "tokens_seen": 1466892288 + }, + { + "epoch": 0.44, + "learning_rate": 0.00044894684052156476, + "loss": 1.3049, + "theoretical_loss": 3.5514744727313032, + "tokens_seen": 1466957824 + }, + { + "epoch": 0.44, + "learning_rate": 0.0004489307923771314, + "loss": 1.3104, + "theoretical_loss": 3.5514605368100307, + "tokens_seen": 1467023360 + }, + { + "epoch": 0.44, + "learning_rate": 0.00044891474423269817, + "loss": 1.4071, + "theoretical_loss": 3.5514466016856074, + "tokens_seen": 1467088896 + }, + { + "epoch": 0.44, + "learning_rate": 0.0004488986960882648, + "loss": 1.3084, + "theoretical_loss": 3.551432667357952, + "tokens_seen": 1467154432 + }, + { + "epoch": 0.44, + "learning_rate": 0.00044888264794383157, + "loss": 1.3718, + "theoretical_loss": 3.551418733826984, + "tokens_seen": 1467219968 + }, + { + "epoch": 0.44, + "learning_rate": 0.0004488665997993982, + "loss": 1.3572, + "theoretical_loss": 3.5514048010926214, + "tokens_seen": 1467285504 + }, + { + "epoch": 0.44, + "learning_rate": 0.00044885055165496497, + "loss": 1.3648, + "theoretical_loss": 3.5513908691547833, + "tokens_seen": 1467351040 + }, + { + "epoch": 0.44, + "learning_rate": 0.0004488345035105316, + "loss": 1.3971, + "theoretical_loss": 3.551376938013389, + "tokens_seen": 1467416576 + }, + { + "epoch": 0.44, + "learning_rate": 0.00044881845536609837, + "loss": 1.3468, + "theoretical_loss": 3.551363007668357, + "tokens_seen": 1467482112 + }, + { + "epoch": 0.44, + "learning_rate": 0.000448802407221665, + "loss": 1.3484, + "theoretical_loss": 3.551349078119606, + "tokens_seen": 1467547648 + }, + { + "epoch": 0.44, + "learning_rate": 0.00044878635907723177, + "loss": 1.3188, + "theoretical_loss": 3.5513351493670555, + "tokens_seen": 1467613184 + }, + { + "epoch": 0.44, + "learning_rate": 0.0004487703109327984, + "loss": 1.3236, + "theoretical_loss": 3.5513212214106247, + "tokens_seen": 1467678720 + }, + { + "epoch": 0.44, + "learning_rate": 0.0004487542627883651, + "loss": 1.409, + "theoretical_loss": 3.5513072942502317, + "tokens_seen": 1467744256 + }, + { + "epoch": 0.44, + "learning_rate": 0.0004487382146439318, + "loss": 1.4918, + "theoretical_loss": 3.551293367885796, + "tokens_seen": 1467809792 + }, + { + "epoch": 0.44, + "learning_rate": 0.0004487221664994985, + "loss": 1.4372, + "theoretical_loss": 3.5512794423172362, + "tokens_seen": 1467875328 + }, + { + "epoch": 0.44, + "learning_rate": 0.0004487061183550652, + "loss": 1.3706, + "theoretical_loss": 3.551265517544471, + "tokens_seen": 1467940864 + }, + { + "epoch": 0.44, + "objective/train/docs_used": 535989, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.341352939605713, + "objective/train/theoretical_loss": 3.5512515935674207, + "objective/train/tokens_used": 1488466400, + "theoretical_loss": 3.5512515935674207, + "tokens_seen": 1468006400 + }, + { + "epoch": 0.44, + "learning_rate": 0.0004486900702106319, + "loss": 1.3911, + "theoretical_loss": 3.5512515935674207, + "tokens_seen": 1468006400 + }, + { + "epoch": 0.44, + "learning_rate": 0.0004486740220661986, + "loss": 1.3775, + "theoretical_loss": 3.551237670386003, + "tokens_seen": 1468071936 + }, + { + "epoch": 0.44, + "learning_rate": 0.0004486579739217653, + "loss": 1.3437, + "theoretical_loss": 3.5512237480001376, + "tokens_seen": 1468137472 + }, + { + "epoch": 0.44, + "learning_rate": 0.000448641925777332, + "loss": 1.3486, + "theoretical_loss": 3.551209826409744, + "tokens_seen": 1468203008 + }, + { + "epoch": 0.44, + "learning_rate": 0.0004486258776328987, + "loss": 1.388, + "theoretical_loss": 3.5511959056147395, + "tokens_seen": 1468268544 + }, + { + "epoch": 0.44, + "learning_rate": 0.00044860982948846537, + "loss": 1.34, + "theoretical_loss": 3.5511819856150453, + "tokens_seen": 1468334080 + }, + { + "epoch": 0.44, + "learning_rate": 0.0004485937813440321, + "loss": 1.4133, + "theoretical_loss": 3.5511680664105785, + "tokens_seen": 1468399616 + }, + { + "epoch": 0.44, + "learning_rate": 0.00044857773319959877, + "loss": 1.3206, + "theoretical_loss": 3.5511541480012596, + "tokens_seen": 1468465152 + }, + { + "epoch": 0.45, + "learning_rate": 0.0004485616850551655, + "loss": 1.2947, + "theoretical_loss": 3.551140230387007, + "tokens_seen": 1468530688 + }, + { + "epoch": 0.45, + "learning_rate": 0.00044854563691073217, + "loss": 1.3849, + "theoretical_loss": 3.55112631356774, + "tokens_seen": 1468596224 + }, + { + "epoch": 0.45, + "learning_rate": 0.0004485295887662989, + "loss": 1.3629, + "theoretical_loss": 3.551112397543378, + "tokens_seen": 1468661760 + }, + { + "epoch": 0.45, + "learning_rate": 0.0004485135406218656, + "loss": 1.4007, + "theoretical_loss": 3.5510984823138396, + "tokens_seen": 1468727296 + }, + { + "epoch": 0.45, + "learning_rate": 0.00044849749247743233, + "loss": 1.3656, + "theoretical_loss": 3.5510845678790446, + "tokens_seen": 1468792832 + }, + { + "epoch": 0.45, + "learning_rate": 0.000448481444332999, + "loss": 1.3327, + "theoretical_loss": 3.551070654238911, + "tokens_seen": 1468858368 + }, + { + "epoch": 0.45, + "learning_rate": 0.00044846539618856573, + "loss": 1.4771, + "theoretical_loss": 3.551056741393359, + "tokens_seen": 1468923904 + }, + { + "epoch": 0.45, + "learning_rate": 0.0004484493480441324, + "loss": 1.4514, + "theoretical_loss": 3.551042829342308, + "tokens_seen": 1468989440 + }, + { + "epoch": 0.45, + "learning_rate": 0.00044843329989969913, + "loss": 1.2683, + "theoretical_loss": 3.5510289180856764, + "tokens_seen": 1469054976 + }, + { + "epoch": 0.45, + "learning_rate": 0.0004484172517552658, + "loss": 1.446, + "theoretical_loss": 3.5510150076233833, + "tokens_seen": 1469120512 + }, + { + "epoch": 0.45, + "learning_rate": 0.00044840120361083253, + "loss": 1.4305, + "theoretical_loss": 3.551001097955348, + "tokens_seen": 1469186048 + }, + { + "epoch": 0.45, + "learning_rate": 0.0004483851554663992, + "loss": 1.3233, + "theoretical_loss": 3.550987189081491, + "tokens_seen": 1469251584 + }, + { + "epoch": 0.45, + "learning_rate": 0.00044836910732196593, + "loss": 1.2991, + "theoretical_loss": 3.5509732810017303, + "tokens_seen": 1469317120 + }, + { + "epoch": 0.45, + "learning_rate": 0.0004483530591775326, + "loss": 1.3153, + "theoretical_loss": 3.550959373715985, + "tokens_seen": 1469382656 + }, + { + "epoch": 0.45, + "learning_rate": 0.00044833701103309933, + "loss": 1.4623, + "theoretical_loss": 3.5509454672241745, + "tokens_seen": 1469448192 + }, + { + "epoch": 0.45, + "learning_rate": 0.000448320962888666, + "loss": 1.3763, + "theoretical_loss": 3.5509315615262187, + "tokens_seen": 1469513728 + }, + { + "epoch": 0.45, + "learning_rate": 0.00044830491474423274, + "loss": 1.3295, + "theoretical_loss": 3.550917656622037, + "tokens_seen": 1469579264 + }, + { + "epoch": 0.45, + "learning_rate": 0.0004482888665997994, + "loss": 1.2333, + "theoretical_loss": 3.5509037525115477, + "tokens_seen": 1469644800 + }, + { + "epoch": 0.45, + "learning_rate": 0.00044827281845536614, + "loss": 1.4753, + "theoretical_loss": 3.5508898491946703, + "tokens_seen": 1469710336 + }, + { + "epoch": 0.45, + "learning_rate": 0.0004482567703109328, + "loss": 1.2441, + "theoretical_loss": 3.550875946671325, + "tokens_seen": 1469775872 + }, + { + "epoch": 0.45, + "learning_rate": 0.00044824072216649954, + "loss": 1.3697, + "theoretical_loss": 3.55086204494143, + "tokens_seen": 1469841408 + }, + { + "epoch": 0.45, + "learning_rate": 0.0004482246740220662, + "loss": 1.3032, + "theoretical_loss": 3.5508481440049056, + "tokens_seen": 1469906944 + }, + { + "epoch": 0.45, + "learning_rate": 0.00044820862587763294, + "loss": 1.3741, + "theoretical_loss": 3.5508342438616705, + "tokens_seen": 1469972480 + }, + { + "epoch": 0.45, + "learning_rate": 0.0004481925777331996, + "loss": 1.3795, + "theoretical_loss": 3.5508203445116444, + "tokens_seen": 1470038016 + }, + { + "epoch": 0.45, + "learning_rate": 0.00044817652958876634, + "loss": 1.277, + "theoretical_loss": 3.5508064459547466, + "tokens_seen": 1470103552 + }, + { + "epoch": 0.45, + "learning_rate": 0.000448160481444333, + "loss": 1.4985, + "theoretical_loss": 3.5507925481908966, + "tokens_seen": 1470169088 + }, + { + "epoch": 0.45, + "learning_rate": 0.00044814443329989974, + "loss": 1.4508, + "theoretical_loss": 3.5507786512200132, + "tokens_seen": 1470234624 + }, + { + "epoch": 0.45, + "learning_rate": 0.0004481283851554664, + "loss": 1.3799, + "theoretical_loss": 3.5507647550420165, + "tokens_seen": 1470300160 + }, + { + "epoch": 0.45, + "learning_rate": 0.00044811233701103314, + "loss": 1.3234, + "theoretical_loss": 3.550750859656826, + "tokens_seen": 1470365696 + }, + { + "epoch": 0.45, + "learning_rate": 0.0004480962888665998, + "loss": 1.2897, + "theoretical_loss": 3.5507369650643605, + "tokens_seen": 1470431232 + }, + { + "epoch": 0.45, + "learning_rate": 0.00044808024072216654, + "loss": 1.3802, + "theoretical_loss": 3.5507230712645397, + "tokens_seen": 1470496768 + }, + { + "epoch": 0.45, + "learning_rate": 0.0004480641925777332, + "loss": 1.3147, + "theoretical_loss": 3.550709178257284, + "tokens_seen": 1470562304 + }, + { + "epoch": 0.45, + "learning_rate": 0.00044804814443329995, + "loss": 1.3095, + "theoretical_loss": 3.5506952860425107, + "tokens_seen": 1470627840 + }, + { + "epoch": 0.45, + "learning_rate": 0.00044803209628886665, + "loss": 1.3416, + "theoretical_loss": 3.5506813946201414, + "tokens_seen": 1470693376 + }, + { + "epoch": 0.45, + "learning_rate": 0.00044801604814443335, + "loss": 1.3928, + "theoretical_loss": 3.550667503990095, + "tokens_seen": 1470758912 + }, + { + "epoch": 0.45, + "learning_rate": 0.00044800000000000005, + "loss": 1.4087, + "theoretical_loss": 3.5506536141522904, + "tokens_seen": 1470824448 + }, + { + "epoch": 0.45, + "learning_rate": 0.0004479839518555667, + "loss": 1.4198, + "theoretical_loss": 3.5506397251066475, + "tokens_seen": 1470889984 + }, + { + "epoch": 0.45, + "learning_rate": 0.00044796790371113345, + "loss": 1.4418, + "theoretical_loss": 3.5506258368530856, + "tokens_seen": 1470955520 + }, + { + "epoch": 0.45, + "learning_rate": 0.0004479518555667001, + "loss": 1.3833, + "theoretical_loss": 3.5506119493915254, + "tokens_seen": 1471021056 + }, + { + "epoch": 0.45, + "learning_rate": 0.00044793580742226685, + "loss": 1.4232, + "theoretical_loss": 3.5505980627218845, + "tokens_seen": 1471086592 + }, + { + "epoch": 0.45, + "learning_rate": 0.0004479197592778335, + "loss": 1.3162, + "theoretical_loss": 3.550584176844084, + "tokens_seen": 1471152128 + }, + { + "epoch": 0.45, + "learning_rate": 0.00044790371113340025, + "loss": 1.3021, + "theoretical_loss": 3.550570291758043, + "tokens_seen": 1471217664 + }, + { + "epoch": 0.45, + "objective/train/docs_used": 537210, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.3250696659088135, + "objective/train/theoretical_loss": 3.550556407463681, + "objective/train/tokens_used": 1491743200, + "theoretical_loss": 3.550556407463681, + "tokens_seen": 1471283200 + }, + { + "epoch": 0.45, + "learning_rate": 0.0004478876629889669, + "loss": 1.3511, + "theoretical_loss": 3.550556407463681, + "tokens_seen": 1471283200 + }, + { + "epoch": 0.45, + "learning_rate": 0.00044787161484453365, + "loss": 1.3637, + "theoretical_loss": 3.5505425239609174, + "tokens_seen": 1471348736 + }, + { + "epoch": 0.45, + "learning_rate": 0.0004478555667001003, + "loss": 1.3927, + "theoretical_loss": 3.550528641249673, + "tokens_seen": 1471414272 + }, + { + "epoch": 0.45, + "learning_rate": 0.00044783951855566705, + "loss": 1.459, + "theoretical_loss": 3.5505147593298654, + "tokens_seen": 1471479808 + }, + { + "epoch": 0.45, + "learning_rate": 0.0004478234704112337, + "loss": 1.3567, + "theoretical_loss": 3.550500878201416, + "tokens_seen": 1471545344 + }, + { + "epoch": 0.45, + "learning_rate": 0.00044780742226680046, + "loss": 1.4503, + "theoretical_loss": 3.5504869978642435, + "tokens_seen": 1471610880 + }, + { + "epoch": 0.45, + "learning_rate": 0.0004477913741223671, + "loss": 1.3015, + "theoretical_loss": 3.550473118318268, + "tokens_seen": 1471676416 + }, + { + "epoch": 0.45, + "learning_rate": 0.00044777532597793386, + "loss": 1.3867, + "theoretical_loss": 3.550459239563409, + "tokens_seen": 1471741952 + }, + { + "epoch": 0.45, + "learning_rate": 0.0004477592778335005, + "loss": 1.3056, + "theoretical_loss": 3.550445361599586, + "tokens_seen": 1471807488 + }, + { + "epoch": 0.45, + "learning_rate": 0.00044774322968906726, + "loss": 1.4225, + "theoretical_loss": 3.550431484426719, + "tokens_seen": 1471873024 + }, + { + "epoch": 0.45, + "learning_rate": 0.0004477271815446339, + "loss": 1.2918, + "theoretical_loss": 3.550417608044728, + "tokens_seen": 1471938560 + }, + { + "epoch": 0.45, + "learning_rate": 0.00044771113340020066, + "loss": 1.41, + "theoretical_loss": 3.550403732453532, + "tokens_seen": 1472004096 + }, + { + "epoch": 0.45, + "learning_rate": 0.0004476950852557673, + "loss": 1.382, + "theoretical_loss": 3.550389857653051, + "tokens_seen": 1472069632 + }, + { + "epoch": 0.45, + "learning_rate": 0.00044767903711133406, + "loss": 1.3386, + "theoretical_loss": 3.550375983643205, + "tokens_seen": 1472135168 + }, + { + "epoch": 0.45, + "learning_rate": 0.0004476629889669007, + "loss": 1.3151, + "theoretical_loss": 3.5503621104239134, + "tokens_seen": 1472200704 + }, + { + "epoch": 0.45, + "learning_rate": 0.00044764694082246746, + "loss": 1.3926, + "theoretical_loss": 3.5503482379950957, + "tokens_seen": 1472266240 + }, + { + "epoch": 0.45, + "learning_rate": 0.0004476308926780341, + "loss": 1.4329, + "theoretical_loss": 3.5503343663566724, + "tokens_seen": 1472331776 + }, + { + "epoch": 0.45, + "learning_rate": 0.00044761484453360086, + "loss": 1.4297, + "theoretical_loss": 3.550320495508563, + "tokens_seen": 1472397312 + }, + { + "epoch": 0.45, + "learning_rate": 0.0004475987963891675, + "loss": 1.5028, + "theoretical_loss": 3.5503066254506876, + "tokens_seen": 1472462848 + }, + { + "epoch": 0.45, + "learning_rate": 0.00044758274824473426, + "loss": 1.3417, + "theoretical_loss": 3.550292756182965, + "tokens_seen": 1472528384 + }, + { + "epoch": 0.45, + "learning_rate": 0.0004475667001003009, + "loss": 1.4241, + "theoretical_loss": 3.550278887705316, + "tokens_seen": 1472593920 + }, + { + "epoch": 0.45, + "learning_rate": 0.00044755065195586767, + "loss": 1.3322, + "theoretical_loss": 3.5502650200176604, + "tokens_seen": 1472659456 + }, + { + "epoch": 0.45, + "learning_rate": 0.0004475346038114343, + "loss": 1.337, + "theoretical_loss": 3.5502511531199175, + "tokens_seen": 1472724992 + }, + { + "epoch": 0.45, + "learning_rate": 0.00044751855566700107, + "loss": 1.4267, + "theoretical_loss": 3.5502372870120076, + "tokens_seen": 1472790528 + }, + { + "epoch": 0.45, + "learning_rate": 0.0004475025075225677, + "loss": 1.4513, + "theoretical_loss": 3.5502234216938504, + "tokens_seen": 1472856064 + }, + { + "epoch": 0.45, + "learning_rate": 0.00044748645937813447, + "loss": 1.4862, + "theoretical_loss": 3.550209557165366, + "tokens_seen": 1472921600 + }, + { + "epoch": 0.45, + "learning_rate": 0.0004474704112337011, + "loss": 1.4247, + "theoretical_loss": 3.550195693426473, + "tokens_seen": 1472987136 + }, + { + "epoch": 0.45, + "learning_rate": 0.00044745436308926787, + "loss": 1.4074, + "theoretical_loss": 3.5501818304770936, + "tokens_seen": 1473052672 + }, + { + "epoch": 0.45, + "learning_rate": 0.0004474383149448345, + "loss": 1.3478, + "theoretical_loss": 3.550167968317146, + "tokens_seen": 1473118208 + }, + { + "epoch": 0.45, + "learning_rate": 0.00044742226680040127, + "loss": 1.3617, + "theoretical_loss": 3.550154106946551, + "tokens_seen": 1473183744 + }, + { + "epoch": 0.45, + "learning_rate": 0.0004474062186559679, + "loss": 1.3366, + "theoretical_loss": 3.5501402463652276, + "tokens_seen": 1473249280 + }, + { + "epoch": 0.45, + "learning_rate": 0.00044739017051153467, + "loss": 1.4313, + "theoretical_loss": 3.5501263865730968, + "tokens_seen": 1473314816 + }, + { + "epoch": 0.45, + "learning_rate": 0.0004473741223671013, + "loss": 1.3411, + "theoretical_loss": 3.5501125275700782, + "tokens_seen": 1473380352 + }, + { + "epoch": 0.45, + "learning_rate": 0.000447358074222668, + "loss": 1.5343, + "theoretical_loss": 3.5500986693560916, + "tokens_seen": 1473445888 + }, + { + "epoch": 0.45, + "learning_rate": 0.0004473420260782347, + "loss": 1.4046, + "theoretical_loss": 3.5500848119310566, + "tokens_seen": 1473511424 + }, + { + "epoch": 0.45, + "learning_rate": 0.0004473259779338014, + "loss": 1.4396, + "theoretical_loss": 3.5500709552948937, + "tokens_seen": 1473576960 + }, + { + "epoch": 0.45, + "learning_rate": 0.0004473099297893681, + "loss": 1.3081, + "theoretical_loss": 3.550057099447523, + "tokens_seen": 1473642496 + }, + { + "epoch": 0.45, + "learning_rate": 0.0004472938816449348, + "loss": 1.3523, + "theoretical_loss": 3.5500432443888643, + "tokens_seen": 1473708032 + }, + { + "epoch": 0.45, + "learning_rate": 0.0004472778335005015, + "loss": 1.3323, + "theoretical_loss": 3.550029390118838, + "tokens_seen": 1473773568 + }, + { + "epoch": 0.45, + "learning_rate": 0.0004472617853560682, + "loss": 1.4095, + "theoretical_loss": 3.550015536637364, + "tokens_seen": 1473839104 + }, + { + "epoch": 0.45, + "learning_rate": 0.0004472457372116349, + "loss": 1.3672, + "theoretical_loss": 3.550001683944362, + "tokens_seen": 1473904640 + }, + { + "epoch": 0.45, + "learning_rate": 0.0004472296890672016, + "loss": 1.3756, + "theoretical_loss": 3.5499878320397524, + "tokens_seen": 1473970176 + }, + { + "epoch": 0.45, + "learning_rate": 0.00044721364092276827, + "loss": 1.3998, + "theoretical_loss": 3.549973980923455, + "tokens_seen": 1474035712 + }, + { + "epoch": 0.45, + "learning_rate": 0.000447197592778335, + "loss": 1.353, + "theoretical_loss": 3.54996013059539, + "tokens_seen": 1474101248 + }, + { + "epoch": 0.45, + "learning_rate": 0.00044718154463390167, + "loss": 1.3913, + "theoretical_loss": 3.549946281055478, + "tokens_seen": 1474166784 + }, + { + "epoch": 0.45, + "learning_rate": 0.0004471654964894684, + "loss": 1.3502, + "theoretical_loss": 3.5499324323036374, + "tokens_seen": 1474232320 + }, + { + "epoch": 0.45, + "learning_rate": 0.00044714944834503507, + "loss": 1.3604, + "theoretical_loss": 3.5499185843397907, + "tokens_seen": 1474297856 + }, + { + "epoch": 0.45, + "learning_rate": 0.00044713340020060183, + "loss": 1.3584, + "theoretical_loss": 3.549904737163857, + "tokens_seen": 1474363392 + }, + { + "epoch": 0.45, + "learning_rate": 0.0004471173520561685, + "loss": 1.4435, + "theoretical_loss": 3.549890890775756, + "tokens_seen": 1474428928 + }, + { + "epoch": 0.45, + "learning_rate": 0.00044710130391173523, + "loss": 1.4907, + "theoretical_loss": 3.5498770451754083, + "tokens_seen": 1474494464 + }, + { + "epoch": 0.45, + "objective/train/docs_used": 538469, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.3394784927368164, + "objective/train/theoretical_loss": 3.549863200362734, + "objective/train/tokens_used": 1495020000, + "theoretical_loss": 3.549863200362734, + "tokens_seen": 1474560000 + }, + { + "epoch": 0.45, + "learning_rate": 0.0004470852557673019, + "loss": 1.3181, + "theoretical_loss": 3.549863200362734, + "tokens_seen": 1474560000 + }, + { + "epoch": 0.45, + "learning_rate": 0.00044706920762286863, + "loss": 1.2597, + "theoretical_loss": 3.5498493563376536, + "tokens_seen": 1474625536 + }, + { + "epoch": 0.45, + "learning_rate": 0.0004470531594784353, + "loss": 1.3619, + "theoretical_loss": 3.5498355131000867, + "tokens_seen": 1474691072 + }, + { + "epoch": 0.45, + "learning_rate": 0.00044703711133400203, + "loss": 1.2643, + "theoretical_loss": 3.549821670649954, + "tokens_seen": 1474756608 + }, + { + "epoch": 0.45, + "learning_rate": 0.0004470210631895687, + "loss": 1.4378, + "theoretical_loss": 3.5498078289871753, + "tokens_seen": 1474822144 + }, + { + "epoch": 0.45, + "learning_rate": 0.00044700501504513543, + "loss": 1.3948, + "theoretical_loss": 3.549793988111671, + "tokens_seen": 1474887680 + }, + { + "epoch": 0.45, + "learning_rate": 0.0004469889669007021, + "loss": 1.4243, + "theoretical_loss": 3.5497801480233617, + "tokens_seen": 1474953216 + }, + { + "epoch": 0.45, + "learning_rate": 0.00044697291875626883, + "loss": 1.3297, + "theoretical_loss": 3.5497663087221674, + "tokens_seen": 1475018752 + }, + { + "epoch": 0.45, + "learning_rate": 0.0004469568706118355, + "loss": 1.4198, + "theoretical_loss": 3.5497524702080083, + "tokens_seen": 1475084288 + }, + { + "epoch": 0.45, + "learning_rate": 0.00044694082246740224, + "loss": 1.4342, + "theoretical_loss": 3.5497386324808042, + "tokens_seen": 1475149824 + }, + { + "epoch": 0.45, + "learning_rate": 0.0004469247743229689, + "loss": 1.3225, + "theoretical_loss": 3.5497247955404765, + "tokens_seen": 1475215360 + }, + { + "epoch": 0.45, + "learning_rate": 0.00044690872617853564, + "loss": 1.3307, + "theoretical_loss": 3.5497109593869447, + "tokens_seen": 1475280896 + }, + { + "epoch": 0.45, + "learning_rate": 0.0004468926780341023, + "loss": 1.4304, + "theoretical_loss": 3.5496971240201294, + "tokens_seen": 1475346432 + }, + { + "epoch": 0.45, + "learning_rate": 0.00044687662988966904, + "loss": 1.4964, + "theoretical_loss": 3.5496832894399506, + "tokens_seen": 1475411968 + }, + { + "epoch": 0.45, + "learning_rate": 0.0004468605817452357, + "loss": 1.4133, + "theoretical_loss": 3.5496694556463293, + "tokens_seen": 1475477504 + }, + { + "epoch": 0.45, + "learning_rate": 0.00044684453360080244, + "loss": 1.3063, + "theoretical_loss": 3.549655622639185, + "tokens_seen": 1475543040 + }, + { + "epoch": 0.45, + "learning_rate": 0.0004468284854563692, + "loss": 1.3108, + "theoretical_loss": 3.5496417904184385, + "tokens_seen": 1475608576 + }, + { + "epoch": 0.45, + "learning_rate": 0.00044681243731193584, + "loss": 1.4052, + "theoretical_loss": 3.54962795898401, + "tokens_seen": 1475674112 + }, + { + "epoch": 0.45, + "learning_rate": 0.0004467963891675026, + "loss": 1.3858, + "theoretical_loss": 3.5496141283358202, + "tokens_seen": 1475739648 + }, + { + "epoch": 0.45, + "learning_rate": 0.00044678034102306924, + "loss": 1.2327, + "theoretical_loss": 3.5496002984737896, + "tokens_seen": 1475805184 + }, + { + "epoch": 0.45, + "learning_rate": 0.000446764292878636, + "loss": 1.2074, + "theoretical_loss": 3.5495864693978376, + "tokens_seen": 1475870720 + }, + { + "epoch": 0.45, + "learning_rate": 0.00044674824473420264, + "loss": 1.4607, + "theoretical_loss": 3.5495726411078863, + "tokens_seen": 1475936256 + }, + { + "epoch": 0.45, + "learning_rate": 0.00044673219658976934, + "loss": 1.3281, + "theoretical_loss": 3.5495588136038547, + "tokens_seen": 1476001792 + }, + { + "epoch": 0.45, + "learning_rate": 0.00044671614844533604, + "loss": 1.481, + "theoretical_loss": 3.549544986885664, + "tokens_seen": 1476067328 + }, + { + "epoch": 0.45, + "learning_rate": 0.00044670010030090274, + "loss": 1.3748, + "theoretical_loss": 3.549531160953234, + "tokens_seen": 1476132864 + }, + { + "epoch": 0.45, + "learning_rate": 0.00044668405215646945, + "loss": 1.4115, + "theoretical_loss": 3.5495173358064855, + "tokens_seen": 1476198400 + }, + { + "epoch": 0.45, + "learning_rate": 0.00044666800401203615, + "loss": 1.3718, + "theoretical_loss": 3.549503511445339, + "tokens_seen": 1476263936 + }, + { + "epoch": 0.45, + "learning_rate": 0.00044665195586760285, + "loss": 1.3515, + "theoretical_loss": 3.5494896878697153, + "tokens_seen": 1476329472 + }, + { + "epoch": 0.45, + "learning_rate": 0.00044663590772316955, + "loss": 1.3461, + "theoretical_loss": 3.549475865079534, + "tokens_seen": 1476395008 + }, + { + "epoch": 0.45, + "learning_rate": 0.00044661985957873625, + "loss": 1.3406, + "theoretical_loss": 3.5494620430747164, + "tokens_seen": 1476460544 + }, + { + "epoch": 0.45, + "learning_rate": 0.00044660381143430295, + "loss": 1.4682, + "theoretical_loss": 3.549448221855183, + "tokens_seen": 1476526080 + }, + { + "epoch": 0.45, + "learning_rate": 0.0004465877632898696, + "loss": 1.4461, + "theoretical_loss": 3.549434401420854, + "tokens_seen": 1476591616 + }, + { + "epoch": 0.45, + "learning_rate": 0.00044657171514543635, + "loss": 1.5059, + "theoretical_loss": 3.5494205817716504, + "tokens_seen": 1476657152 + }, + { + "epoch": 0.45, + "learning_rate": 0.000446555667001003, + "loss": 1.4618, + "theoretical_loss": 3.549406762907492, + "tokens_seen": 1476722688 + }, + { + "epoch": 0.45, + "learning_rate": 0.00044653961885656975, + "loss": 1.3533, + "theoretical_loss": 3.5493929448283, + "tokens_seen": 1476788224 + }, + { + "epoch": 0.45, + "learning_rate": 0.0004465235707121364, + "loss": 1.4303, + "theoretical_loss": 3.549379127533995, + "tokens_seen": 1476853760 + }, + { + "epoch": 0.45, + "learning_rate": 0.00044650752256770315, + "loss": 1.3562, + "theoretical_loss": 3.5493653110244976, + "tokens_seen": 1476919296 + }, + { + "epoch": 0.45, + "learning_rate": 0.0004464914744232698, + "loss": 1.3542, + "theoretical_loss": 3.549351495299728, + "tokens_seen": 1476984832 + }, + { + "epoch": 0.45, + "learning_rate": 0.00044647542627883655, + "loss": 1.3038, + "theoretical_loss": 3.5493376803596064, + "tokens_seen": 1477050368 + }, + { + "epoch": 0.45, + "learning_rate": 0.0004464593781344032, + "loss": 1.3631, + "theoretical_loss": 3.5493238662040545, + "tokens_seen": 1477115904 + }, + { + "epoch": 0.45, + "learning_rate": 0.00044644332998996995, + "loss": 1.4233, + "theoretical_loss": 3.5493100528329924, + "tokens_seen": 1477181440 + }, + { + "epoch": 0.45, + "learning_rate": 0.0004464272818455366, + "loss": 1.45, + "theoretical_loss": 3.549296240246341, + "tokens_seen": 1477246976 + }, + { + "epoch": 0.45, + "learning_rate": 0.00044641123370110336, + "loss": 1.3836, + "theoretical_loss": 3.5492824284440205, + "tokens_seen": 1477312512 + }, + { + "epoch": 0.45, + "learning_rate": 0.00044639518555667, + "loss": 1.4483, + "theoretical_loss": 3.5492686174259522, + "tokens_seen": 1477378048 + }, + { + "epoch": 0.45, + "learning_rate": 0.00044637913741223676, + "loss": 1.3779, + "theoretical_loss": 3.549254807192056, + "tokens_seen": 1477443584 + }, + { + "epoch": 0.45, + "learning_rate": 0.0004463630892678034, + "loss": 1.3212, + "theoretical_loss": 3.5492409977422534, + "tokens_seen": 1477509120 + }, + { + "epoch": 0.45, + "learning_rate": 0.00044634704112337016, + "loss": 1.3847, + "theoretical_loss": 3.549227189076465, + "tokens_seen": 1477574656 + }, + { + "epoch": 0.45, + "learning_rate": 0.0004463309929789368, + "loss": 1.3512, + "theoretical_loss": 3.54921338119461, + "tokens_seen": 1477640192 + }, + { + "epoch": 0.45, + "learning_rate": 0.00044631494483450356, + "loss": 1.2777, + "theoretical_loss": 3.549199574096612, + "tokens_seen": 1477705728 + }, + { + "epoch": 0.45, + "learning_rate": 0.0004462988966900702, + "loss": 1.3761, + "theoretical_loss": 3.549185767782389, + "tokens_seen": 1477771264 + }, + { + "epoch": 0.45, + "objective/train/docs_used": 539614, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.325514554977417, + "objective/train/theoretical_loss": 3.549171962251863, + "objective/train/tokens_used": 1498296800, + "theoretical_loss": 3.549171962251863, + "tokens_seen": 1477836800 + }, + { + "epoch": 0.45, + "learning_rate": 0.00044628284854563696, + "loss": 1.3018, + "theoretical_loss": 3.549171962251863, + "tokens_seen": 1477836800 + }, + { + "epoch": 0.45, + "learning_rate": 0.0004462668004012036, + "loss": 1.3547, + "theoretical_loss": 3.5491581575049547, + "tokens_seen": 1477902336 + }, + { + "epoch": 0.45, + "learning_rate": 0.00044625075225677036, + "loss": 1.3914, + "theoretical_loss": 3.549144353541585, + "tokens_seen": 1477967872 + }, + { + "epoch": 0.45, + "learning_rate": 0.000446234704112337, + "loss": 1.4294, + "theoretical_loss": 3.5491305503616744, + "tokens_seen": 1478033408 + }, + { + "epoch": 0.45, + "learning_rate": 0.00044621865596790376, + "loss": 1.4189, + "theoretical_loss": 3.5491167479651438, + "tokens_seen": 1478098944 + }, + { + "epoch": 0.45, + "learning_rate": 0.0004462026078234704, + "loss": 1.4117, + "theoretical_loss": 3.5491029463519137, + "tokens_seen": 1478164480 + }, + { + "epoch": 0.45, + "learning_rate": 0.00044618655967903716, + "loss": 1.3923, + "theoretical_loss": 3.5490891455219056, + "tokens_seen": 1478230016 + }, + { + "epoch": 0.45, + "learning_rate": 0.0004461705115346038, + "loss": 1.2946, + "theoretical_loss": 3.54907534547504, + "tokens_seen": 1478295552 + }, + { + "epoch": 0.45, + "learning_rate": 0.00044615446339017057, + "loss": 1.3391, + "theoretical_loss": 3.549061546211237, + "tokens_seen": 1478361088 + }, + { + "epoch": 0.45, + "learning_rate": 0.0004461384152457372, + "loss": 1.3795, + "theoretical_loss": 3.5490477477304188, + "tokens_seen": 1478426624 + }, + { + "epoch": 0.45, + "learning_rate": 0.00044612236710130397, + "loss": 1.3828, + "theoretical_loss": 3.5490339500325057, + "tokens_seen": 1478492160 + }, + { + "epoch": 0.45, + "learning_rate": 0.0004461063189568706, + "loss": 1.3991, + "theoretical_loss": 3.5490201531174184, + "tokens_seen": 1478557696 + }, + { + "epoch": 0.45, + "learning_rate": 0.00044609027081243737, + "loss": 1.3326, + "theoretical_loss": 3.549006356985078, + "tokens_seen": 1478623232 + }, + { + "epoch": 0.45, + "learning_rate": 0.000446074222668004, + "loss": 1.4137, + "theoretical_loss": 3.548992561635405, + "tokens_seen": 1478688768 + }, + { + "epoch": 0.45, + "learning_rate": 0.00044605817452357077, + "loss": 1.37, + "theoretical_loss": 3.548978767068321, + "tokens_seen": 1478754304 + }, + { + "epoch": 0.45, + "learning_rate": 0.0004460421263791374, + "loss": 1.2974, + "theoretical_loss": 3.548964973283746, + "tokens_seen": 1478819840 + }, + { + "epoch": 0.45, + "learning_rate": 0.00044602607823470417, + "loss": 1.4639, + "theoretical_loss": 3.548951180281602, + "tokens_seen": 1478885376 + }, + { + "epoch": 0.45, + "learning_rate": 0.0004460100300902708, + "loss": 1.3941, + "theoretical_loss": 3.548937388061809, + "tokens_seen": 1478950912 + }, + { + "epoch": 0.45, + "learning_rate": 0.00044599398194583757, + "loss": 1.5252, + "theoretical_loss": 3.548923596624288, + "tokens_seen": 1479016448 + }, + { + "epoch": 0.45, + "learning_rate": 0.0004459779338014042, + "loss": 1.3359, + "theoretical_loss": 3.548909805968961, + "tokens_seen": 1479081984 + }, + { + "epoch": 0.45, + "learning_rate": 0.0004459618856569709, + "loss": 1.4013, + "theoretical_loss": 3.548896016095748, + "tokens_seen": 1479147520 + }, + { + "epoch": 0.45, + "learning_rate": 0.0004459458375125376, + "loss": 1.4813, + "theoretical_loss": 3.5488822270045706, + "tokens_seen": 1479213056 + }, + { + "epoch": 0.45, + "learning_rate": 0.0004459297893681043, + "loss": 1.4702, + "theoretical_loss": 3.5488684386953495, + "tokens_seen": 1479278592 + }, + { + "epoch": 0.45, + "learning_rate": 0.000445913741223671, + "loss": 1.348, + "theoretical_loss": 3.5488546511680057, + "tokens_seen": 1479344128 + }, + { + "epoch": 0.45, + "learning_rate": 0.0004458976930792377, + "loss": 1.3924, + "theoretical_loss": 3.54884086442246, + "tokens_seen": 1479409664 + }, + { + "epoch": 0.45, + "learning_rate": 0.0004458816449348044, + "loss": 1.3573, + "theoretical_loss": 3.548827078458634, + "tokens_seen": 1479475200 + }, + { + "epoch": 0.45, + "learning_rate": 0.0004458655967903711, + "loss": 1.4093, + "theoretical_loss": 3.5488132932764485, + "tokens_seen": 1479540736 + }, + { + "epoch": 0.45, + "learning_rate": 0.0004458495486459378, + "loss": 1.2473, + "theoretical_loss": 3.5487995088758244, + "tokens_seen": 1479606272 + }, + { + "epoch": 0.45, + "learning_rate": 0.0004458335005015045, + "loss": 1.413, + "theoretical_loss": 3.548785725256683, + "tokens_seen": 1479671808 + }, + { + "epoch": 0.45, + "learning_rate": 0.00044581745235707117, + "loss": 1.4098, + "theoretical_loss": 3.548771942418945, + "tokens_seen": 1479737344 + }, + { + "epoch": 0.45, + "learning_rate": 0.0004458014042126379, + "loss": 1.4936, + "theoretical_loss": 3.5487581603625324, + "tokens_seen": 1479802880 + }, + { + "epoch": 0.45, + "learning_rate": 0.00044578535606820457, + "loss": 1.4299, + "theoretical_loss": 3.548744379087365, + "tokens_seen": 1479868416 + }, + { + "epoch": 0.45, + "learning_rate": 0.00044576930792377133, + "loss": 1.4118, + "theoretical_loss": 3.548730598593365, + "tokens_seen": 1479933952 + }, + { + "epoch": 0.45, + "learning_rate": 0.000445753259779338, + "loss": 1.3449, + "theoretical_loss": 3.548716818880453, + "tokens_seen": 1479999488 + }, + { + "epoch": 0.45, + "learning_rate": 0.00044573721163490473, + "loss": 1.4224, + "theoretical_loss": 3.5487030399485504, + "tokens_seen": 1480065024 + }, + { + "epoch": 0.45, + "learning_rate": 0.0004457211634904714, + "loss": 1.4985, + "theoretical_loss": 3.5486892617975783, + "tokens_seen": 1480130560 + }, + { + "epoch": 0.45, + "learning_rate": 0.00044570511534603813, + "loss": 1.4024, + "theoretical_loss": 3.5486754844274575, + "tokens_seen": 1480196096 + }, + { + "epoch": 0.45, + "learning_rate": 0.0004456890672016048, + "loss": 1.3495, + "theoretical_loss": 3.54866170783811, + "tokens_seen": 1480261632 + }, + { + "epoch": 0.45, + "learning_rate": 0.00044567301905717153, + "loss": 1.3659, + "theoretical_loss": 3.548647932029456, + "tokens_seen": 1480327168 + }, + { + "epoch": 0.45, + "learning_rate": 0.0004456569709127382, + "loss": 1.4122, + "theoretical_loss": 3.548634157001417, + "tokens_seen": 1480392704 + }, + { + "epoch": 0.45, + "learning_rate": 0.00044564092276830493, + "loss": 1.3913, + "theoretical_loss": 3.548620382753915, + "tokens_seen": 1480458240 + }, + { + "epoch": 0.45, + "learning_rate": 0.0004456248746238717, + "loss": 1.5378, + "theoretical_loss": 3.54860660928687, + "tokens_seen": 1480523776 + }, + { + "epoch": 0.45, + "learning_rate": 0.00044560882647943833, + "loss": 1.3525, + "theoretical_loss": 3.5485928366002044, + "tokens_seen": 1480589312 + }, + { + "epoch": 0.45, + "learning_rate": 0.0004455927783350051, + "loss": 1.4187, + "theoretical_loss": 3.5485790646938384, + "tokens_seen": 1480654848 + }, + { + "epoch": 0.45, + "learning_rate": 0.00044557673019057173, + "loss": 1.4063, + "theoretical_loss": 3.548565293567694, + "tokens_seen": 1480720384 + }, + { + "epoch": 0.45, + "learning_rate": 0.0004455606820461385, + "loss": 1.3919, + "theoretical_loss": 3.5485515232216924, + "tokens_seen": 1480785920 + }, + { + "epoch": 0.45, + "learning_rate": 0.00044554463390170514, + "loss": 1.4411, + "theoretical_loss": 3.5485377536557543, + "tokens_seen": 1480851456 + }, + { + "epoch": 0.45, + "learning_rate": 0.0004455285857572719, + "loss": 1.342, + "theoretical_loss": 3.5485239848698016, + "tokens_seen": 1480916992 + }, + { + "epoch": 0.45, + "learning_rate": 0.00044551253761283854, + "loss": 1.455, + "theoretical_loss": 3.5485102168637557, + "tokens_seen": 1480982528 + }, + { + "epoch": 0.45, + "learning_rate": 0.0004454964894684053, + "loss": 1.4136, + "theoretical_loss": 3.548496449637537, + "tokens_seen": 1481048064 + }, + { + "epoch": 0.45, + "objective/train/docs_used": 540822, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.4331252574920654, + "objective/train/theoretical_loss": 3.548482683191067, + "objective/train/tokens_used": 1501573600, + "theoretical_loss": 3.548482683191067, + "tokens_seen": 1481113600 + }, + { + "epoch": 0.45, + "learning_rate": 0.00044548044132397194, + "loss": 1.4637, + "theoretical_loss": 3.548482683191067, + "tokens_seen": 1481113600 + }, + { + "epoch": 0.45, + "learning_rate": 0.0004454643931795387, + "loss": 1.3335, + "theoretical_loss": 3.5484689175242687, + "tokens_seen": 1481179136 + }, + { + "epoch": 0.45, + "learning_rate": 0.00044544834503510534, + "loss": 1.4985, + "theoretical_loss": 3.548455152637062, + "tokens_seen": 1481244672 + }, + { + "epoch": 0.45, + "learning_rate": 0.0004454322968906721, + "loss": 1.3361, + "theoretical_loss": 3.5484413885293673, + "tokens_seen": 1481310208 + }, + { + "epoch": 0.45, + "learning_rate": 0.00044541624874623874, + "loss": 1.4092, + "theoretical_loss": 3.548427625201108, + "tokens_seen": 1481375744 + }, + { + "epoch": 0.45, + "learning_rate": 0.0004454002006018055, + "loss": 1.4275, + "theoretical_loss": 3.548413862652205, + "tokens_seen": 1481441280 + }, + { + "epoch": 0.45, + "learning_rate": 0.00044538415245737214, + "loss": 1.393, + "theoretical_loss": 3.5484001008825787, + "tokens_seen": 1481506816 + }, + { + "epoch": 0.45, + "learning_rate": 0.0004453681043129389, + "loss": 1.3719, + "theoretical_loss": 3.5483863398921516, + "tokens_seen": 1481572352 + }, + { + "epoch": 0.45, + "learning_rate": 0.00044535205616850554, + "loss": 1.3326, + "theoretical_loss": 3.5483725796808443, + "tokens_seen": 1481637888 + }, + { + "epoch": 0.45, + "learning_rate": 0.00044533600802407224, + "loss": 1.4383, + "theoretical_loss": 3.5483588202485787, + "tokens_seen": 1481703424 + }, + { + "epoch": 0.45, + "learning_rate": 0.00044531995987963895, + "loss": 1.3621, + "theoretical_loss": 3.548345061595276, + "tokens_seen": 1481768960 + }, + { + "epoch": 0.45, + "learning_rate": 0.00044530391173520565, + "loss": 1.418, + "theoretical_loss": 3.5483313037208575, + "tokens_seen": 1481834496 + }, + { + "epoch": 0.45, + "learning_rate": 0.00044528786359077235, + "loss": 1.3076, + "theoretical_loss": 3.5483175466252455, + "tokens_seen": 1481900032 + }, + { + "epoch": 0.45, + "learning_rate": 0.00044527181544633905, + "loss": 1.476, + "theoretical_loss": 3.5483037903083607, + "tokens_seen": 1481965568 + }, + { + "epoch": 0.45, + "learning_rate": 0.00044525576730190575, + "loss": 1.4408, + "theoretical_loss": 3.5482900347701247, + "tokens_seen": 1482031104 + }, + { + "epoch": 0.45, + "learning_rate": 0.00044523971915747245, + "loss": 1.3123, + "theoretical_loss": 3.5482762800104592, + "tokens_seen": 1482096640 + }, + { + "epoch": 0.45, + "learning_rate": 0.00044522367101303915, + "loss": 1.4525, + "theoretical_loss": 3.5482625260292853, + "tokens_seen": 1482162176 + }, + { + "epoch": 0.45, + "learning_rate": 0.00044520762286860585, + "loss": 1.3657, + "theoretical_loss": 3.5482487728265255, + "tokens_seen": 1482227712 + }, + { + "epoch": 0.45, + "learning_rate": 0.0004451915747241725, + "loss": 1.4115, + "theoretical_loss": 3.5482350204020996, + "tokens_seen": 1482293248 + }, + { + "epoch": 0.45, + "learning_rate": 0.00044517552657973925, + "loss": 1.4453, + "theoretical_loss": 3.548221268755931, + "tokens_seen": 1482358784 + }, + { + "epoch": 0.45, + "learning_rate": 0.0004451594784353059, + "loss": 1.3212, + "theoretical_loss": 3.54820751788794, + "tokens_seen": 1482424320 + }, + { + "epoch": 0.45, + "learning_rate": 0.00044514343029087265, + "loss": 1.346, + "theoretical_loss": 3.548193767798049, + "tokens_seen": 1482489856 + }, + { + "epoch": 0.45, + "learning_rate": 0.0004451273821464393, + "loss": 1.4658, + "theoretical_loss": 3.548180018486179, + "tokens_seen": 1482555392 + }, + { + "epoch": 0.45, + "learning_rate": 0.00044511133400200605, + "loss": 1.3222, + "theoretical_loss": 3.548166269952252, + "tokens_seen": 1482620928 + }, + { + "epoch": 0.45, + "learning_rate": 0.0004450952858575727, + "loss": 1.4105, + "theoretical_loss": 3.548152522196189, + "tokens_seen": 1482686464 + }, + { + "epoch": 0.45, + "learning_rate": 0.00044507923771313945, + "loss": 1.3364, + "theoretical_loss": 3.548138775217912, + "tokens_seen": 1482752000 + }, + { + "epoch": 0.45, + "learning_rate": 0.0004450631895687061, + "loss": 1.3685, + "theoretical_loss": 3.548125029017343, + "tokens_seen": 1482817536 + }, + { + "epoch": 0.45, + "learning_rate": 0.00044504714142427286, + "loss": 1.3535, + "theoretical_loss": 3.5481112835944026, + "tokens_seen": 1482883072 + }, + { + "epoch": 0.45, + "learning_rate": 0.0004450310932798395, + "loss": 1.4246, + "theoretical_loss": 3.5480975389490137, + "tokens_seen": 1482948608 + }, + { + "epoch": 0.45, + "learning_rate": 0.00044501504513540626, + "loss": 1.403, + "theoretical_loss": 3.548083795081097, + "tokens_seen": 1483014144 + }, + { + "epoch": 0.45, + "learning_rate": 0.0004449989969909729, + "loss": 1.4027, + "theoretical_loss": 3.5480700519905746, + "tokens_seen": 1483079680 + }, + { + "epoch": 0.45, + "learning_rate": 0.00044498294884653966, + "loss": 1.4235, + "theoretical_loss": 3.548056309677368, + "tokens_seen": 1483145216 + }, + { + "epoch": 0.45, + "learning_rate": 0.0004449669007021063, + "loss": 1.4706, + "theoretical_loss": 3.548042568141399, + "tokens_seen": 1483210752 + }, + { + "epoch": 0.45, + "learning_rate": 0.00044495085255767306, + "loss": 1.3732, + "theoretical_loss": 3.5480288273825895, + "tokens_seen": 1483276288 + }, + { + "epoch": 0.45, + "learning_rate": 0.0004449348044132397, + "loss": 1.5689, + "theoretical_loss": 3.5480150874008602, + "tokens_seen": 1483341824 + }, + { + "epoch": 0.45, + "learning_rate": 0.00044491875626880646, + "loss": 1.4227, + "theoretical_loss": 3.5480013481961343, + "tokens_seen": 1483407360 + }, + { + "epoch": 0.45, + "learning_rate": 0.0004449027081243731, + "loss": 1.4509, + "theoretical_loss": 3.5479876097683327, + "tokens_seen": 1483472896 + }, + { + "epoch": 0.45, + "learning_rate": 0.00044488665997993986, + "loss": 1.4348, + "theoretical_loss": 3.5479738721173772, + "tokens_seen": 1483538432 + }, + { + "epoch": 0.45, + "learning_rate": 0.0004448706118355065, + "loss": 1.3739, + "theoretical_loss": 3.5479601352431898, + "tokens_seen": 1483603968 + }, + { + "epoch": 0.45, + "learning_rate": 0.00044485456369107326, + "loss": 1.465, + "theoretical_loss": 3.547946399145692, + "tokens_seen": 1483669504 + }, + { + "epoch": 0.45, + "learning_rate": 0.0004448385155466399, + "loss": 1.4052, + "theoretical_loss": 3.547932663824806, + "tokens_seen": 1483735040 + }, + { + "epoch": 0.45, + "learning_rate": 0.00044482246740220666, + "loss": 1.4832, + "theoretical_loss": 3.547918929280453, + "tokens_seen": 1483800576 + }, + { + "epoch": 0.45, + "learning_rate": 0.0004448064192577733, + "loss": 1.5936, + "theoretical_loss": 3.5479051955125556, + "tokens_seen": 1483866112 + }, + { + "epoch": 0.45, + "learning_rate": 0.00044479037111334007, + "loss": 1.4267, + "theoretical_loss": 3.5478914625210343, + "tokens_seen": 1483931648 + }, + { + "epoch": 0.45, + "learning_rate": 0.0004447743229689067, + "loss": 1.3538, + "theoretical_loss": 3.547877730305812, + "tokens_seen": 1483997184 + }, + { + "epoch": 0.45, + "learning_rate": 0.00044475827482447347, + "loss": 1.3957, + "theoretical_loss": 3.547863998866811, + "tokens_seen": 1484062720 + }, + { + "epoch": 0.45, + "learning_rate": 0.0004447422266800401, + "loss": 1.2819, + "theoretical_loss": 3.5478502682039514, + "tokens_seen": 1484128256 + }, + { + "epoch": 0.45, + "learning_rate": 0.00044472617853560687, + "loss": 1.3432, + "theoretical_loss": 3.5478365383171564, + "tokens_seen": 1484193792 + }, + { + "epoch": 0.45, + "learning_rate": 0.0004447101303911735, + "loss": 1.3947, + "theoretical_loss": 3.547822809206348, + "tokens_seen": 1484259328 + }, + { + "epoch": 0.45, + "learning_rate": 0.00044469408224674027, + "loss": 1.458, + "theoretical_loss": 3.5478090808714473, + "tokens_seen": 1484324864 + }, + { + "epoch": 0.45, + "objective/train/docs_used": 542078, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.3846638202667236, + "objective/train/theoretical_loss": 3.5477953533123765, + "objective/train/tokens_used": 1504850400, + "theoretical_loss": 3.5477953533123765, + "tokens_seen": 1484390400 + }, + { + "epoch": 0.45, + "learning_rate": 0.0004446780341023069, + "loss": 1.4009, + "theoretical_loss": 3.5477953533123765, + "tokens_seen": 1484390400 + }, + { + "epoch": 0.45, + "learning_rate": 0.00044466198595787367, + "loss": 1.3662, + "theoretical_loss": 3.5477816265290576, + "tokens_seen": 1484455936 + }, + { + "epoch": 0.45, + "learning_rate": 0.0004446459378134403, + "loss": 1.3432, + "theoretical_loss": 3.5477679005214124, + "tokens_seen": 1484521472 + }, + { + "epoch": 0.45, + "learning_rate": 0.00044462988966900707, + "loss": 1.3956, + "theoretical_loss": 3.547754175289363, + "tokens_seen": 1484587008 + }, + { + "epoch": 0.45, + "learning_rate": 0.0004446138415245737, + "loss": 1.3881, + "theoretical_loss": 3.5477404508328307, + "tokens_seen": 1484652544 + }, + { + "epoch": 0.45, + "learning_rate": 0.0004445977933801405, + "loss": 1.39, + "theoretical_loss": 3.5477267271517388, + "tokens_seen": 1484718080 + }, + { + "epoch": 0.45, + "learning_rate": 0.0004445817452357071, + "loss": 1.4521, + "theoretical_loss": 3.547713004246008, + "tokens_seen": 1484783616 + }, + { + "epoch": 0.45, + "learning_rate": 0.0004445656970912738, + "loss": 1.4824, + "theoretical_loss": 3.547699282115561, + "tokens_seen": 1484849152 + }, + { + "epoch": 0.45, + "learning_rate": 0.0004445496489468405, + "loss": 1.3447, + "theoretical_loss": 3.547685560760319, + "tokens_seen": 1484914688 + }, + { + "epoch": 0.45, + "learning_rate": 0.0004445336008024072, + "loss": 1.402, + "theoretical_loss": 3.5476718401802048, + "tokens_seen": 1484980224 + }, + { + "epoch": 0.45, + "learning_rate": 0.0004445175526579739, + "loss": 1.4985, + "theoretical_loss": 3.54765812037514, + "tokens_seen": 1485045760 + }, + { + "epoch": 0.45, + "learning_rate": 0.0004445015045135406, + "loss": 1.4806, + "theoretical_loss": 3.547644401345047, + "tokens_seen": 1485111296 + }, + { + "epoch": 0.45, + "learning_rate": 0.0004444854563691073, + "loss": 1.4458, + "theoretical_loss": 3.5476306830898467, + "tokens_seen": 1485176832 + }, + { + "epoch": 0.45, + "learning_rate": 0.000444469408224674, + "loss": 1.4437, + "theoretical_loss": 3.547616965609463, + "tokens_seen": 1485242368 + }, + { + "epoch": 0.45, + "learning_rate": 0.0004444533600802407, + "loss": 1.3758, + "theoretical_loss": 3.5476032489038163, + "tokens_seen": 1485307904 + }, + { + "epoch": 0.45, + "learning_rate": 0.0004444373119358074, + "loss": 1.2946, + "theoretical_loss": 3.5475895329728298, + "tokens_seen": 1485373440 + }, + { + "epoch": 0.45, + "learning_rate": 0.0004444212637913742, + "loss": 1.3244, + "theoretical_loss": 3.5475758178164245, + "tokens_seen": 1485438976 + }, + { + "epoch": 0.45, + "learning_rate": 0.0004444052156469408, + "loss": 1.3998, + "theoretical_loss": 3.5475621034345233, + "tokens_seen": 1485504512 + }, + { + "epoch": 0.45, + "learning_rate": 0.0004443891675025076, + "loss": 1.423, + "theoretical_loss": 3.547548389827048, + "tokens_seen": 1485570048 + }, + { + "epoch": 0.45, + "learning_rate": 0.00044437311935807423, + "loss": 1.4097, + "theoretical_loss": 3.547534676993921, + "tokens_seen": 1485635584 + }, + { + "epoch": 0.45, + "learning_rate": 0.000444357071213641, + "loss": 1.4473, + "theoretical_loss": 3.547520964935064, + "tokens_seen": 1485701120 + }, + { + "epoch": 0.45, + "learning_rate": 0.00044434102306920763, + "loss": 1.3702, + "theoretical_loss": 3.5475072536503998, + "tokens_seen": 1485766656 + }, + { + "epoch": 0.45, + "learning_rate": 0.0004443249749247744, + "loss": 1.2723, + "theoretical_loss": 3.547493543139849, + "tokens_seen": 1485832192 + }, + { + "epoch": 0.45, + "learning_rate": 0.00044430892678034103, + "loss": 1.4313, + "theoretical_loss": 3.5474798334033357, + "tokens_seen": 1485897728 + }, + { + "epoch": 0.45, + "learning_rate": 0.0004442928786359078, + "loss": 1.4101, + "theoretical_loss": 3.547466124440781, + "tokens_seen": 1485963264 + }, + { + "epoch": 0.45, + "learning_rate": 0.00044427683049147443, + "loss": 1.3843, + "theoretical_loss": 3.5474524162521073, + "tokens_seen": 1486028800 + }, + { + "epoch": 0.45, + "learning_rate": 0.0004442607823470412, + "loss": 1.3772, + "theoretical_loss": 3.5474387088372366, + "tokens_seen": 1486094336 + }, + { + "epoch": 0.45, + "learning_rate": 0.00044424473420260783, + "loss": 1.3996, + "theoretical_loss": 3.547425002196091, + "tokens_seen": 1486159872 + }, + { + "epoch": 0.45, + "learning_rate": 0.0004442286860581746, + "loss": 1.4086, + "theoretical_loss": 3.547411296328593, + "tokens_seen": 1486225408 + }, + { + "epoch": 0.45, + "learning_rate": 0.00044421263791374123, + "loss": 1.3482, + "theoretical_loss": 3.5473975912346654, + "tokens_seen": 1486290944 + }, + { + "epoch": 0.45, + "learning_rate": 0.000444196589769308, + "loss": 1.4637, + "theoretical_loss": 3.547383886914229, + "tokens_seen": 1486356480 + }, + { + "epoch": 0.45, + "learning_rate": 0.00044418054162487464, + "loss": 1.3147, + "theoretical_loss": 3.5473701833672076, + "tokens_seen": 1486422016 + }, + { + "epoch": 0.45, + "learning_rate": 0.0004441644934804414, + "loss": 1.3957, + "theoretical_loss": 3.547356480593522, + "tokens_seen": 1486487552 + }, + { + "epoch": 0.45, + "learning_rate": 0.00044414844533600804, + "loss": 1.3822, + "theoretical_loss": 3.5473427785930958, + "tokens_seen": 1486553088 + }, + { + "epoch": 0.45, + "learning_rate": 0.0004441323971915748, + "loss": 1.371, + "theoretical_loss": 3.5473290773658506, + "tokens_seen": 1486618624 + }, + { + "epoch": 0.45, + "learning_rate": 0.00044411634904714144, + "loss": 1.4093, + "theoretical_loss": 3.5473153769117083, + "tokens_seen": 1486684160 + }, + { + "epoch": 0.45, + "learning_rate": 0.0004441003009027082, + "loss": 1.4429, + "theoretical_loss": 3.5473016772305916, + "tokens_seen": 1486749696 + }, + { + "epoch": 0.45, + "learning_rate": 0.00044408425275827484, + "loss": 1.3013, + "theoretical_loss": 3.5472879783224234, + "tokens_seen": 1486815232 + }, + { + "epoch": 0.45, + "learning_rate": 0.0004440682046138416, + "loss": 1.2975, + "theoretical_loss": 3.547274280187125, + "tokens_seen": 1486880768 + }, + { + "epoch": 0.45, + "learning_rate": 0.00044405215646940824, + "loss": 1.3728, + "theoretical_loss": 3.5472605828246193, + "tokens_seen": 1486946304 + }, + { + "epoch": 0.45, + "learning_rate": 0.000444036108324975, + "loss": 1.4189, + "theoretical_loss": 3.547246886234829, + "tokens_seen": 1487011840 + }, + { + "epoch": 0.45, + "learning_rate": 0.00044402006018054164, + "loss": 1.476, + "theoretical_loss": 3.5472331904176757, + "tokens_seen": 1487077376 + }, + { + "epoch": 0.45, + "learning_rate": 0.0004440040120361084, + "loss": 1.3773, + "theoretical_loss": 3.547219495373082, + "tokens_seen": 1487142912 + }, + { + "epoch": 0.45, + "learning_rate": 0.00044398796389167504, + "loss": 1.3754, + "theoretical_loss": 3.5472058011009704, + "tokens_seen": 1487208448 + }, + { + "epoch": 0.45, + "learning_rate": 0.0004439719157472418, + "loss": 1.3818, + "theoretical_loss": 3.547192107601263, + "tokens_seen": 1487273984 + }, + { + "epoch": 0.45, + "learning_rate": 0.00044395586760280844, + "loss": 1.2292, + "theoretical_loss": 3.5471784148738825, + "tokens_seen": 1487339520 + }, + { + "epoch": 0.45, + "learning_rate": 0.00044393981945837515, + "loss": 1.3516, + "theoretical_loss": 3.5471647229187515, + "tokens_seen": 1487405056 + }, + { + "epoch": 0.45, + "learning_rate": 0.00044392377131394185, + "loss": 1.3547, + "theoretical_loss": 3.547151031735792, + "tokens_seen": 1487470592 + }, + { + "epoch": 0.45, + "learning_rate": 0.00044390772316950855, + "loss": 1.3878, + "theoretical_loss": 3.5471373413249268, + "tokens_seen": 1487536128 + }, + { + "epoch": 0.45, + "learning_rate": 0.00044389167502507525, + "loss": 1.2884, + "theoretical_loss": 3.547123651686078, + "tokens_seen": 1487601664 + }, + { + "epoch": 0.45, + "objective/train/docs_used": 543335, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.334249496459961, + "objective/train/theoretical_loss": 3.5471099628191682, + "objective/train/tokens_used": 1508127200, + "theoretical_loss": 3.5471099628191682, + "tokens_seen": 1487667200 + }, + { + "epoch": 0.45, + "learning_rate": 0.00044387562688064195, + "loss": 1.3611, + "theoretical_loss": 3.5471099628191682, + "tokens_seen": 1487667200 + }, + { + "epoch": 0.45, + "learning_rate": 0.00044385957873620865, + "loss": 1.3784, + "theoretical_loss": 3.5470962747241197, + "tokens_seen": 1487732736 + }, + { + "epoch": 0.45, + "learning_rate": 0.00044384353059177535, + "loss": 1.324, + "theoretical_loss": 3.5470825874008556, + "tokens_seen": 1487798272 + }, + { + "epoch": 0.45, + "learning_rate": 0.00044382748244734205, + "loss": 1.4601, + "theoretical_loss": 3.5470689008492977, + "tokens_seen": 1487863808 + }, + { + "epoch": 0.45, + "learning_rate": 0.00044381143430290875, + "loss": 1.5314, + "theoretical_loss": 3.547055215069369, + "tokens_seen": 1487929344 + }, + { + "epoch": 0.45, + "learning_rate": 0.0004437953861584754, + "loss": 1.3757, + "theoretical_loss": 3.5470415300609917, + "tokens_seen": 1487994880 + }, + { + "epoch": 0.45, + "learning_rate": 0.00044377933801404215, + "loss": 1.495, + "theoretical_loss": 3.547027845824088, + "tokens_seen": 1488060416 + }, + { + "epoch": 0.45, + "learning_rate": 0.0004437632898696088, + "loss": 1.4057, + "theoretical_loss": 3.5470141623585807, + "tokens_seen": 1488125952 + }, + { + "epoch": 0.45, + "learning_rate": 0.00044374724172517555, + "loss": 1.3543, + "theoretical_loss": 3.5470004796643932, + "tokens_seen": 1488191488 + }, + { + "epoch": 0.45, + "learning_rate": 0.0004437311935807422, + "loss": 1.4628, + "theoretical_loss": 3.5469867977414467, + "tokens_seen": 1488257024 + }, + { + "epoch": 0.45, + "learning_rate": 0.00044371514543630895, + "loss": 1.41, + "theoretical_loss": 3.546973116589665, + "tokens_seen": 1488322560 + }, + { + "epoch": 0.45, + "learning_rate": 0.0004436990972918756, + "loss": 1.3866, + "theoretical_loss": 3.5469594362089696, + "tokens_seen": 1488388096 + }, + { + "epoch": 0.45, + "learning_rate": 0.00044368304914744236, + "loss": 1.4169, + "theoretical_loss": 3.5469457565992837, + "tokens_seen": 1488453632 + }, + { + "epoch": 0.45, + "learning_rate": 0.000443667001003009, + "loss": 1.3613, + "theoretical_loss": 3.54693207776053, + "tokens_seen": 1488519168 + }, + { + "epoch": 0.45, + "learning_rate": 0.00044365095285857576, + "loss": 1.3265, + "theoretical_loss": 3.546918399692631, + "tokens_seen": 1488584704 + }, + { + "epoch": 0.45, + "learning_rate": 0.0004436349047141424, + "loss": 1.396, + "theoretical_loss": 3.5469047223955084, + "tokens_seen": 1488650240 + }, + { + "epoch": 0.45, + "learning_rate": 0.00044361885656970916, + "loss": 1.3943, + "theoretical_loss": 3.5468910458690863, + "tokens_seen": 1488715776 + }, + { + "epoch": 0.45, + "learning_rate": 0.0004436028084252758, + "loss": 1.3633, + "theoretical_loss": 3.5468773701132865, + "tokens_seen": 1488781312 + }, + { + "epoch": 0.45, + "learning_rate": 0.00044358676028084256, + "loss": 1.3831, + "theoretical_loss": 3.5468636951280317, + "tokens_seen": 1488846848 + }, + { + "epoch": 0.45, + "learning_rate": 0.0004435707121364092, + "loss": 1.4063, + "theoretical_loss": 3.5468500209132454, + "tokens_seen": 1488912384 + }, + { + "epoch": 0.45, + "learning_rate": 0.00044355466399197596, + "loss": 1.453, + "theoretical_loss": 3.546836347468849, + "tokens_seen": 1488977920 + }, + { + "epoch": 0.45, + "learning_rate": 0.0004435386158475426, + "loss": 1.3785, + "theoretical_loss": 3.546822674794766, + "tokens_seen": 1489043456 + }, + { + "epoch": 0.45, + "learning_rate": 0.00044352256770310936, + "loss": 1.3767, + "theoretical_loss": 3.5468090028909187, + "tokens_seen": 1489108992 + }, + { + "epoch": 0.45, + "learning_rate": 0.000443506519558676, + "loss": 1.3249, + "theoretical_loss": 3.54679533175723, + "tokens_seen": 1489174528 + }, + { + "epoch": 0.45, + "learning_rate": 0.00044349047141424276, + "loss": 1.4061, + "theoretical_loss": 3.5467816613936227, + "tokens_seen": 1489240064 + }, + { + "epoch": 0.45, + "learning_rate": 0.0004434744232698094, + "loss": 1.3129, + "theoretical_loss": 3.54676799180002, + "tokens_seen": 1489305600 + }, + { + "epoch": 0.45, + "learning_rate": 0.00044345837512537616, + "loss": 1.3539, + "theoretical_loss": 3.5467543229763434, + "tokens_seen": 1489371136 + }, + { + "epoch": 0.45, + "learning_rate": 0.0004434423269809428, + "loss": 1.3911, + "theoretical_loss": 3.546740654922517, + "tokens_seen": 1489436672 + }, + { + "epoch": 0.45, + "learning_rate": 0.00044342627883650957, + "loss": 1.3353, + "theoretical_loss": 3.546726987638462, + "tokens_seen": 1489502208 + }, + { + "epoch": 0.45, + "learning_rate": 0.0004434102306920762, + "loss": 1.4245, + "theoretical_loss": 3.5467133211241033, + "tokens_seen": 1489567744 + }, + { + "epoch": 0.45, + "learning_rate": 0.00044339418254764297, + "loss": 1.3708, + "theoretical_loss": 3.5466996553793617, + "tokens_seen": 1489633280 + }, + { + "epoch": 0.45, + "learning_rate": 0.0004433781344032096, + "loss": 1.464, + "theoretical_loss": 3.5466859904041614, + "tokens_seen": 1489698816 + }, + { + "epoch": 0.45, + "learning_rate": 0.00044336208625877637, + "loss": 1.3157, + "theoretical_loss": 3.5466723261984243, + "tokens_seen": 1489764352 + }, + { + "epoch": 0.45, + "learning_rate": 0.000443346038114343, + "loss": 1.4094, + "theoretical_loss": 3.5466586627620735, + "tokens_seen": 1489829888 + }, + { + "epoch": 0.45, + "learning_rate": 0.00044332998996990977, + "loss": 1.4477, + "theoretical_loss": 3.5466450000950323, + "tokens_seen": 1489895424 + }, + { + "epoch": 0.45, + "learning_rate": 0.0004433139418254764, + "loss": 1.4079, + "theoretical_loss": 3.5466313381972228, + "tokens_seen": 1489960960 + }, + { + "epoch": 0.45, + "learning_rate": 0.00044329789368104317, + "loss": 1.3481, + "theoretical_loss": 3.5466176770685682, + "tokens_seen": 1490026496 + }, + { + "epoch": 0.45, + "learning_rate": 0.0004432818455366098, + "loss": 1.4105, + "theoretical_loss": 3.546604016708992, + "tokens_seen": 1490092032 + }, + { + "epoch": 0.45, + "learning_rate": 0.00044326579739217657, + "loss": 1.3587, + "theoretical_loss": 3.546590357118416, + "tokens_seen": 1490157568 + }, + { + "epoch": 0.45, + "learning_rate": 0.0004432497492477432, + "loss": 1.2635, + "theoretical_loss": 3.546576698296763, + "tokens_seen": 1490223104 + }, + { + "epoch": 0.45, + "learning_rate": 0.00044323370110331, + "loss": 1.2896, + "theoretical_loss": 3.546563040243957, + "tokens_seen": 1490288640 + }, + { + "epoch": 0.45, + "learning_rate": 0.0004432176529588767, + "loss": 1.3259, + "theoretical_loss": 3.5465493829599204, + "tokens_seen": 1490354176 + }, + { + "epoch": 0.45, + "learning_rate": 0.0004432016048144434, + "loss": 1.2913, + "theoretical_loss": 3.546535726444576, + "tokens_seen": 1490419712 + }, + { + "epoch": 0.45, + "learning_rate": 0.0004431855566700101, + "loss": 1.3251, + "theoretical_loss": 3.546522070697847, + "tokens_seen": 1490485248 + }, + { + "epoch": 0.45, + "learning_rate": 0.0004431695085255767, + "loss": 1.3479, + "theoretical_loss": 3.546508415719656, + "tokens_seen": 1490550784 + }, + { + "epoch": 0.45, + "learning_rate": 0.0004431534603811435, + "loss": 1.4548, + "theoretical_loss": 3.5464947615099263, + "tokens_seen": 1490616320 + }, + { + "epoch": 0.45, + "learning_rate": 0.0004431374122367101, + "loss": 1.2827, + "theoretical_loss": 3.5464811080685807, + "tokens_seen": 1490681856 + }, + { + "epoch": 0.45, + "learning_rate": 0.0004431213640922769, + "loss": 1.3786, + "theoretical_loss": 3.5464674553955415, + "tokens_seen": 1490747392 + }, + { + "epoch": 0.45, + "learning_rate": 0.0004431053159478435, + "loss": 1.4627, + "theoretical_loss": 3.546453803490733, + "tokens_seen": 1490812928 + }, + { + "epoch": 0.45, + "learning_rate": 0.0004430892678034103, + "loss": 1.368, + "theoretical_loss": 3.5464401523540783, + "tokens_seen": 1490878464 + }, + { + "epoch": 0.45, + "objective/train/docs_used": 544526, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.416668176651001, + "objective/train/theoretical_loss": 3.546426501985499, + "objective/train/tokens_used": 1511404000, + "theoretical_loss": 3.546426501985499, + "tokens_seen": 1490944000 + }, + { + "epoch": 0.45, + "learning_rate": 0.0004430732196589769, + "loss": 1.3835, + "theoretical_loss": 3.546426501985499, + "tokens_seen": 1490944000 + }, + { + "epoch": 0.45, + "learning_rate": 0.0004430571715145437, + "loss": 1.3957, + "theoretical_loss": 3.5464128523849183, + "tokens_seen": 1491009536 + }, + { + "epoch": 0.45, + "learning_rate": 0.0004430411233701103, + "loss": 1.3486, + "theoretical_loss": 3.5463992035522605, + "tokens_seen": 1491075072 + }, + { + "epoch": 0.45, + "learning_rate": 0.0004430250752256771, + "loss": 1.3409, + "theoretical_loss": 3.546385555487448, + "tokens_seen": 1491140608 + }, + { + "epoch": 0.45, + "learning_rate": 0.00044300902708124373, + "loss": 1.3997, + "theoretical_loss": 3.5463719081904035, + "tokens_seen": 1491206144 + }, + { + "epoch": 0.45, + "learning_rate": 0.0004429929789368105, + "loss": 1.3207, + "theoretical_loss": 3.54635826166105, + "tokens_seen": 1491271680 + }, + { + "epoch": 0.45, + "learning_rate": 0.00044297693079237713, + "loss": 1.3862, + "theoretical_loss": 3.5463446158993115, + "tokens_seen": 1491337216 + }, + { + "epoch": 0.45, + "learning_rate": 0.0004429608826479439, + "loss": 1.2495, + "theoretical_loss": 3.5463309709051103, + "tokens_seen": 1491402752 + }, + { + "epoch": 0.45, + "learning_rate": 0.00044294483450351053, + "loss": 1.4073, + "theoretical_loss": 3.54631732667837, + "tokens_seen": 1491468288 + }, + { + "epoch": 0.45, + "learning_rate": 0.0004429287863590773, + "loss": 1.4006, + "theoretical_loss": 3.546303683219013, + "tokens_seen": 1491533824 + }, + { + "epoch": 0.45, + "learning_rate": 0.00044291273821464393, + "loss": 1.3874, + "theoretical_loss": 3.546290040526963, + "tokens_seen": 1491599360 + }, + { + "epoch": 0.45, + "learning_rate": 0.0004428966900702107, + "loss": 1.3059, + "theoretical_loss": 3.546276398602143, + "tokens_seen": 1491664896 + }, + { + "epoch": 0.45, + "learning_rate": 0.00044288064192577733, + "loss": 1.3463, + "theoretical_loss": 3.546262757444476, + "tokens_seen": 1491730432 + }, + { + "epoch": 0.45, + "learning_rate": 0.0004428645937813441, + "loss": 1.4159, + "theoretical_loss": 3.5462491170538852, + "tokens_seen": 1491795968 + }, + { + "epoch": 0.45, + "learning_rate": 0.00044284854563691073, + "loss": 1.4424, + "theoretical_loss": 3.5462354774302947, + "tokens_seen": 1491861504 + }, + { + "epoch": 0.45, + "learning_rate": 0.0004428324974924775, + "loss": 1.347, + "theoretical_loss": 3.546221838573626, + "tokens_seen": 1491927040 + }, + { + "epoch": 0.45, + "learning_rate": 0.00044281644934804414, + "loss": 1.4658, + "theoretical_loss": 3.546208200483804, + "tokens_seen": 1491992576 + }, + { + "epoch": 0.45, + "learning_rate": 0.0004428004012036109, + "loss": 1.4611, + "theoretical_loss": 3.54619456316075, + "tokens_seen": 1492058112 + }, + { + "epoch": 0.45, + "learning_rate": 0.00044278435305917754, + "loss": 1.3795, + "theoretical_loss": 3.546180926604389, + "tokens_seen": 1492123648 + }, + { + "epoch": 0.45, + "learning_rate": 0.0004427683049147443, + "loss": 1.4115, + "theoretical_loss": 3.546167290814643, + "tokens_seen": 1492189184 + }, + { + "epoch": 0.45, + "learning_rate": 0.00044275225677031094, + "loss": 1.3174, + "theoretical_loss": 3.5461536557914357, + "tokens_seen": 1492254720 + }, + { + "epoch": 0.45, + "learning_rate": 0.0004427362086258777, + "loss": 1.3199, + "theoretical_loss": 3.546140021534691, + "tokens_seen": 1492320256 + }, + { + "epoch": 0.45, + "learning_rate": 0.00044272016048144434, + "loss": 1.3605, + "theoretical_loss": 3.546126388044331, + "tokens_seen": 1492385792 + }, + { + "epoch": 0.45, + "learning_rate": 0.0004427041123370111, + "loss": 1.5292, + "theoretical_loss": 3.54611275532028, + "tokens_seen": 1492451328 + }, + { + "epoch": 0.45, + "learning_rate": 0.00044268806419257774, + "loss": 1.3939, + "theoretical_loss": 3.5460991233624606, + "tokens_seen": 1492516864 + }, + { + "epoch": 0.45, + "learning_rate": 0.0004426720160481445, + "loss": 1.2986, + "theoretical_loss": 3.5460854921707963, + "tokens_seen": 1492582400 + }, + { + "epoch": 0.45, + "learning_rate": 0.00044265596790371114, + "loss": 1.4956, + "theoretical_loss": 3.54607186174521, + "tokens_seen": 1492647936 + }, + { + "epoch": 0.45, + "learning_rate": 0.0004426399197592779, + "loss": 1.2471, + "theoretical_loss": 3.5460582320856258, + "tokens_seen": 1492713472 + }, + { + "epoch": 0.45, + "learning_rate": 0.00044262387161484454, + "loss": 1.3619, + "theoretical_loss": 3.5460446031919663, + "tokens_seen": 1492779008 + }, + { + "epoch": 0.45, + "learning_rate": 0.0004426078234704113, + "loss": 1.4147, + "theoretical_loss": 3.546030975064155, + "tokens_seen": 1492844544 + }, + { + "epoch": 0.45, + "learning_rate": 0.00044259177532597794, + "loss": 1.3095, + "theoretical_loss": 3.546017347702116, + "tokens_seen": 1492910080 + }, + { + "epoch": 0.45, + "learning_rate": 0.0004425757271815447, + "loss": 1.4008, + "theoretical_loss": 3.5460037211057718, + "tokens_seen": 1492975616 + }, + { + "epoch": 0.45, + "learning_rate": 0.00044255967903711135, + "loss": 1.3731, + "theoretical_loss": 3.545990095275046, + "tokens_seen": 1493041152 + }, + { + "epoch": 0.45, + "learning_rate": 0.0004425436308926781, + "loss": 1.4145, + "theoretical_loss": 3.5459764702098617, + "tokens_seen": 1493106688 + }, + { + "epoch": 0.45, + "learning_rate": 0.00044252758274824475, + "loss": 1.412, + "theoretical_loss": 3.545962845910143, + "tokens_seen": 1493172224 + }, + { + "epoch": 0.45, + "learning_rate": 0.00044251153460381145, + "loss": 1.2616, + "theoretical_loss": 3.5459492223758127, + "tokens_seen": 1493237760 + }, + { + "epoch": 0.45, + "learning_rate": 0.00044249548645937815, + "loss": 1.3983, + "theoretical_loss": 3.5459355996067945, + "tokens_seen": 1493303296 + }, + { + "epoch": 0.45, + "learning_rate": 0.00044247943831494485, + "loss": 1.3918, + "theoretical_loss": 3.5459219776030118, + "tokens_seen": 1493368832 + }, + { + "epoch": 0.45, + "learning_rate": 0.00044246339017051155, + "loss": 1.4144, + "theoretical_loss": 3.5459083563643876, + "tokens_seen": 1493434368 + }, + { + "epoch": 0.45, + "learning_rate": 0.00044244734202607825, + "loss": 1.3575, + "theoretical_loss": 3.5458947358908466, + "tokens_seen": 1493499904 + }, + { + "epoch": 0.45, + "learning_rate": 0.00044243129388164495, + "loss": 1.3464, + "theoretical_loss": 3.5458811161823105, + "tokens_seen": 1493565440 + }, + { + "epoch": 0.45, + "learning_rate": 0.00044241524573721165, + "loss": 1.2546, + "theoretical_loss": 3.545867497238704, + "tokens_seen": 1493630976 + }, + { + "epoch": 0.45, + "learning_rate": 0.0004423991975927783, + "loss": 1.2859, + "theoretical_loss": 3.5458538790599503, + "tokens_seen": 1493696512 + }, + { + "epoch": 0.45, + "learning_rate": 0.00044238314944834505, + "loss": 1.34, + "theoretical_loss": 3.5458402616459725, + "tokens_seen": 1493762048 + }, + { + "epoch": 0.45, + "learning_rate": 0.0004423671013039117, + "loss": 1.3608, + "theoretical_loss": 3.545826644996694, + "tokens_seen": 1493827584 + }, + { + "epoch": 0.45, + "learning_rate": 0.00044235105315947845, + "loss": 1.3053, + "theoretical_loss": 3.5458130291120398, + "tokens_seen": 1493893120 + }, + { + "epoch": 0.45, + "learning_rate": 0.0004423350050150451, + "loss": 1.4029, + "theoretical_loss": 3.545799413991932, + "tokens_seen": 1493958656 + }, + { + "epoch": 0.45, + "learning_rate": 0.00044231895687061186, + "loss": 1.407, + "theoretical_loss": 3.545785799636294, + "tokens_seen": 1494024192 + }, + { + "epoch": 0.45, + "learning_rate": 0.0004423029087261785, + "loss": 1.2977, + "theoretical_loss": 3.54577218604505, + "tokens_seen": 1494089728 + }, + { + "epoch": 0.45, + "learning_rate": 0.00044228686058174526, + "loss": 1.3675, + "theoretical_loss": 3.545758573218124, + "tokens_seen": 1494155264 + }, + { + "epoch": 0.45, + "objective/train/docs_used": 545793, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.2716798782348633, + "objective/train/theoretical_loss": 3.545744961155438, + "objective/train/tokens_used": 1514680800, + "theoretical_loss": 3.545744961155438, + "tokens_seen": 1494220800 + }, + { + "epoch": 0.45, + "learning_rate": 0.0004422708124373119, + "loss": 1.2636, + "theoretical_loss": 3.545744961155438, + "tokens_seen": 1494220800 + }, + { + "epoch": 0.45, + "learning_rate": 0.00044225476429287866, + "loss": 1.4384, + "theoretical_loss": 3.545731349856917, + "tokens_seen": 1494286336 + }, + { + "epoch": 0.45, + "learning_rate": 0.0004422387161484453, + "loss": 1.3507, + "theoretical_loss": 3.545717739322484, + "tokens_seen": 1494351872 + }, + { + "epoch": 0.45, + "learning_rate": 0.00044222266800401206, + "loss": 1.3256, + "theoretical_loss": 3.545704129552063, + "tokens_seen": 1494417408 + }, + { + "epoch": 0.45, + "learning_rate": 0.0004422066198595787, + "loss": 1.3193, + "theoretical_loss": 3.545690520545577, + "tokens_seen": 1494482944 + }, + { + "epoch": 0.45, + "learning_rate": 0.00044219057171514546, + "loss": 1.3832, + "theoretical_loss": 3.5456769123029503, + "tokens_seen": 1494548480 + }, + { + "epoch": 0.45, + "learning_rate": 0.0004421745235707121, + "loss": 1.4193, + "theoretical_loss": 3.5456633048241057, + "tokens_seen": 1494614016 + }, + { + "epoch": 0.45, + "learning_rate": 0.00044215847542627886, + "loss": 1.4177, + "theoretical_loss": 3.545649698108968, + "tokens_seen": 1494679552 + }, + { + "epoch": 0.45, + "learning_rate": 0.0004421424272818455, + "loss": 1.3104, + "theoretical_loss": 3.5456360921574595, + "tokens_seen": 1494745088 + }, + { + "epoch": 0.45, + "learning_rate": 0.00044212637913741226, + "loss": 1.3605, + "theoretical_loss": 3.5456224869695045, + "tokens_seen": 1494810624 + }, + { + "epoch": 0.45, + "learning_rate": 0.0004421103309929789, + "loss": 1.4068, + "theoretical_loss": 3.5456088825450274, + "tokens_seen": 1494876160 + }, + { + "epoch": 0.45, + "learning_rate": 0.00044209428284854566, + "loss": 1.3866, + "theoretical_loss": 3.5455952788839507, + "tokens_seen": 1494941696 + }, + { + "epoch": 0.45, + "learning_rate": 0.0004420782347041123, + "loss": 1.3349, + "theoretical_loss": 3.545581675986199, + "tokens_seen": 1495007232 + }, + { + "epoch": 0.45, + "learning_rate": 0.00044206218655967907, + "loss": 1.2465, + "theoretical_loss": 3.5455680738516953, + "tokens_seen": 1495072768 + }, + { + "epoch": 0.45, + "learning_rate": 0.0004420461384152457, + "loss": 1.375, + "theoretical_loss": 3.545554472480364, + "tokens_seen": 1495138304 + }, + { + "epoch": 0.45, + "learning_rate": 0.00044203009027081247, + "loss": 1.4002, + "theoretical_loss": 3.545540871872128, + "tokens_seen": 1495203840 + }, + { + "epoch": 0.45, + "learning_rate": 0.0004420140421263792, + "loss": 1.38, + "theoretical_loss": 3.5455272720269115, + "tokens_seen": 1495269376 + }, + { + "epoch": 0.45, + "learning_rate": 0.00044199799398194587, + "loss": 1.338, + "theoretical_loss": 3.5455136729446384, + "tokens_seen": 1495334912 + }, + { + "epoch": 0.45, + "learning_rate": 0.0004419819458375126, + "loss": 1.3897, + "theoretical_loss": 3.5455000746252328, + "tokens_seen": 1495400448 + }, + { + "epoch": 0.45, + "learning_rate": 0.00044196589769307927, + "loss": 1.4099, + "theoretical_loss": 3.5454864770686174, + "tokens_seen": 1495465984 + }, + { + "epoch": 0.45, + "learning_rate": 0.000441949849548646, + "loss": 1.4346, + "theoretical_loss": 3.5454728802747164, + "tokens_seen": 1495531520 + }, + { + "epoch": 0.45, + "learning_rate": 0.00044193380140421267, + "loss": 1.359, + "theoretical_loss": 3.5454592842434547, + "tokens_seen": 1495597056 + }, + { + "epoch": 0.45, + "learning_rate": 0.0004419177532597794, + "loss": 1.4386, + "theoretical_loss": 3.5454456889747545, + "tokens_seen": 1495662592 + }, + { + "epoch": 0.45, + "learning_rate": 0.00044190170511534607, + "loss": 1.4683, + "theoretical_loss": 3.5454320944685405, + "tokens_seen": 1495728128 + }, + { + "epoch": 0.45, + "learning_rate": 0.00044188565697091277, + "loss": 1.3153, + "theoretical_loss": 3.545418500724736, + "tokens_seen": 1495793664 + }, + { + "epoch": 0.45, + "learning_rate": 0.0004418696088264795, + "loss": 1.3932, + "theoretical_loss": 3.545404907743266, + "tokens_seen": 1495859200 + }, + { + "epoch": 0.45, + "learning_rate": 0.0004418535606820462, + "loss": 1.3575, + "theoretical_loss": 3.545391315524053, + "tokens_seen": 1495924736 + }, + { + "epoch": 0.45, + "learning_rate": 0.0004418375125376129, + "loss": 1.4502, + "theoretical_loss": 3.5453777240670217, + "tokens_seen": 1495990272 + }, + { + "epoch": 0.45, + "learning_rate": 0.0004418214643931796, + "loss": 1.3801, + "theoretical_loss": 3.5453641333720953, + "tokens_seen": 1496055808 + }, + { + "epoch": 0.45, + "learning_rate": 0.0004418054162487463, + "loss": 1.2811, + "theoretical_loss": 3.5453505434391985, + "tokens_seen": 1496121344 + }, + { + "epoch": 0.45, + "learning_rate": 0.000441789368104313, + "loss": 1.3795, + "theoretical_loss": 3.5453369542682545, + "tokens_seen": 1496186880 + }, + { + "epoch": 0.45, + "learning_rate": 0.0004417733199598797, + "loss": 1.3899, + "theoretical_loss": 3.545323365859187, + "tokens_seen": 1496252416 + }, + { + "epoch": 0.45, + "learning_rate": 0.0004417572718154464, + "loss": 1.3434, + "theoretical_loss": 3.545309778211921, + "tokens_seen": 1496317952 + }, + { + "epoch": 0.45, + "learning_rate": 0.000441741223671013, + "loss": 1.4278, + "theoretical_loss": 3.54529619132638, + "tokens_seen": 1496383488 + }, + { + "epoch": 0.45, + "learning_rate": 0.0004417251755265798, + "loss": 1.4859, + "theoretical_loss": 3.5452826052024875, + "tokens_seen": 1496449024 + }, + { + "epoch": 0.45, + "learning_rate": 0.0004417091273821464, + "loss": 1.3423, + "theoretical_loss": 3.545269019840167, + "tokens_seen": 1496514560 + }, + { + "epoch": 0.45, + "learning_rate": 0.0004416930792377132, + "loss": 1.5828, + "theoretical_loss": 3.545255435239344, + "tokens_seen": 1496580096 + }, + { + "epoch": 0.45, + "learning_rate": 0.0004416770310932798, + "loss": 1.4488, + "theoretical_loss": 3.5452418513999415, + "tokens_seen": 1496645632 + }, + { + "epoch": 0.45, + "learning_rate": 0.0004416609829488466, + "loss": 1.3003, + "theoretical_loss": 3.5452282683218836, + "tokens_seen": 1496711168 + }, + { + "epoch": 0.45, + "learning_rate": 0.00044164493480441323, + "loss": 1.3355, + "theoretical_loss": 3.5452146860050946, + "tokens_seen": 1496776704 + }, + { + "epoch": 0.45, + "learning_rate": 0.00044162888665998, + "loss": 1.3771, + "theoretical_loss": 3.5452011044494975, + "tokens_seen": 1496842240 + }, + { + "epoch": 0.45, + "learning_rate": 0.00044161283851554663, + "loss": 1.462, + "theoretical_loss": 3.5451875236550174, + "tokens_seen": 1496907776 + }, + { + "epoch": 0.45, + "learning_rate": 0.0004415967903711134, + "loss": 1.3754, + "theoretical_loss": 3.5451739436215783, + "tokens_seen": 1496973312 + }, + { + "epoch": 0.45, + "learning_rate": 0.00044158074222668003, + "loss": 1.3423, + "theoretical_loss": 3.545160364349104, + "tokens_seen": 1497038848 + }, + { + "epoch": 0.45, + "learning_rate": 0.0004415646940822468, + "loss": 1.2108, + "theoretical_loss": 3.5451467858375176, + "tokens_seen": 1497104384 + }, + { + "epoch": 0.45, + "learning_rate": 0.00044154864593781343, + "loss": 1.4017, + "theoretical_loss": 3.5451332080867446, + "tokens_seen": 1497169920 + }, + { + "epoch": 0.45, + "learning_rate": 0.0004415325977933802, + "loss": 1.4547, + "theoretical_loss": 3.5451196310967084, + "tokens_seen": 1497235456 + }, + { + "epoch": 0.45, + "learning_rate": 0.00044151654964894683, + "loss": 1.3796, + "theoretical_loss": 3.5451060548673334, + "tokens_seen": 1497300992 + }, + { + "epoch": 0.45, + "learning_rate": 0.0004415005015045136, + "loss": 1.541, + "theoretical_loss": 3.545092479398543, + "tokens_seen": 1497366528 + }, + { + "epoch": 0.45, + "learning_rate": 0.00044148445336008023, + "loss": 1.4236, + "theoretical_loss": 3.545078904690262, + "tokens_seen": 1497432064 + }, + { + "epoch": 0.45, + "objective/train/docs_used": 547043, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.2956085205078125, + "objective/train/theoretical_loss": 3.545065330742414, + "objective/train/tokens_used": 1517957600, + "theoretical_loss": 3.545065330742414, + "tokens_seen": 1497497600 + }, + { + "epoch": 0.45, + "learning_rate": 0.000441468405215647, + "loss": 1.3373, + "theoretical_loss": 3.545065330742414, + "tokens_seen": 1497497600 + }, + { + "epoch": 0.45, + "learning_rate": 0.00044145235707121364, + "loss": 1.364, + "theoretical_loss": 3.545051757554924, + "tokens_seen": 1497563136 + }, + { + "epoch": 0.45, + "learning_rate": 0.0004414363089267804, + "loss": 1.4403, + "theoretical_loss": 3.545038185127715, + "tokens_seen": 1497628672 + }, + { + "epoch": 0.45, + "learning_rate": 0.00044142026078234704, + "loss": 1.4086, + "theoretical_loss": 3.545024613460712, + "tokens_seen": 1497694208 + }, + { + "epoch": 0.45, + "learning_rate": 0.0004414042126379138, + "loss": 1.4659, + "theoretical_loss": 3.545011042553839, + "tokens_seen": 1497759744 + }, + { + "epoch": 0.45, + "learning_rate": 0.00044138816449348044, + "loss": 1.4128, + "theoretical_loss": 3.5449974724070197, + "tokens_seen": 1497825280 + }, + { + "epoch": 0.45, + "learning_rate": 0.0004413721163490472, + "loss": 1.425, + "theoretical_loss": 3.544983903020178, + "tokens_seen": 1497890816 + }, + { + "epoch": 0.45, + "learning_rate": 0.00044135606820461384, + "loss": 1.2686, + "theoretical_loss": 3.5449703343932395, + "tokens_seen": 1497956352 + }, + { + "epoch": 0.45, + "learning_rate": 0.0004413400200601806, + "loss": 1.3882, + "theoretical_loss": 3.5449567665261275, + "tokens_seen": 1498021888 + }, + { + "epoch": 0.45, + "learning_rate": 0.00044132397191574724, + "loss": 1.3627, + "theoretical_loss": 3.544943199418766, + "tokens_seen": 1498087424 + }, + { + "epoch": 0.45, + "learning_rate": 0.000441307923771314, + "loss": 1.4033, + "theoretical_loss": 3.5449296330710798, + "tokens_seen": 1498152960 + }, + { + "epoch": 0.45, + "learning_rate": 0.00044129187562688064, + "loss": 1.4631, + "theoretical_loss": 3.5449160674829927, + "tokens_seen": 1498218496 + }, + { + "epoch": 0.45, + "learning_rate": 0.0004412758274824474, + "loss": 1.5325, + "theoretical_loss": 3.544902502654429, + "tokens_seen": 1498284032 + }, + { + "epoch": 0.45, + "learning_rate": 0.00044125977933801404, + "loss": 1.3944, + "theoretical_loss": 3.5448889385853137, + "tokens_seen": 1498349568 + }, + { + "epoch": 0.45, + "learning_rate": 0.0004412437311935808, + "loss": 1.3813, + "theoretical_loss": 3.5448753752755695, + "tokens_seen": 1498415104 + }, + { + "epoch": 0.45, + "learning_rate": 0.00044122768304914744, + "loss": 1.3533, + "theoretical_loss": 3.544861812725122, + "tokens_seen": 1498480640 + }, + { + "epoch": 0.45, + "learning_rate": 0.0004412116349047142, + "loss": 1.3837, + "theoretical_loss": 3.544848250933895, + "tokens_seen": 1498546176 + }, + { + "epoch": 0.45, + "learning_rate": 0.00044119558676028085, + "loss": 1.391, + "theoretical_loss": 3.5448346899018133, + "tokens_seen": 1498611712 + }, + { + "epoch": 0.45, + "learning_rate": 0.0004411795386158476, + "loss": 1.3055, + "theoretical_loss": 3.5448211296288004, + "tokens_seen": 1498677248 + }, + { + "epoch": 0.45, + "learning_rate": 0.00044116349047141425, + "loss": 1.2946, + "theoretical_loss": 3.5448075701147808, + "tokens_seen": 1498742784 + }, + { + "epoch": 0.45, + "learning_rate": 0.000441147442326981, + "loss": 1.4547, + "theoretical_loss": 3.54479401135968, + "tokens_seen": 1498808320 + }, + { + "epoch": 0.45, + "learning_rate": 0.00044113139418254765, + "loss": 1.4768, + "theoretical_loss": 3.5447804533634204, + "tokens_seen": 1498873856 + }, + { + "epoch": 0.45, + "learning_rate": 0.00044111534603811435, + "loss": 1.3487, + "theoretical_loss": 3.5447668961259278, + "tokens_seen": 1498939392 + }, + { + "epoch": 0.45, + "learning_rate": 0.00044109929789368105, + "loss": 1.351, + "theoretical_loss": 3.5447533396471256, + "tokens_seen": 1499004928 + }, + { + "epoch": 0.45, + "learning_rate": 0.00044108324974924775, + "loss": 1.3701, + "theoretical_loss": 3.5447397839269392, + "tokens_seen": 1499070464 + }, + { + "epoch": 0.45, + "learning_rate": 0.00044106720160481445, + "loss": 1.4573, + "theoretical_loss": 3.5447262289652928, + "tokens_seen": 1499136000 + }, + { + "epoch": 0.45, + "learning_rate": 0.00044105115346038115, + "loss": 1.3677, + "theoretical_loss": 3.54471267476211, + "tokens_seen": 1499201536 + }, + { + "epoch": 0.45, + "learning_rate": 0.00044103510531594785, + "loss": 1.4672, + "theoretical_loss": 3.5446991213173153, + "tokens_seen": 1499267072 + }, + { + "epoch": 0.45, + "learning_rate": 0.00044101905717151455, + "loss": 1.3376, + "theoretical_loss": 3.544685568630834, + "tokens_seen": 1499332608 + }, + { + "epoch": 0.45, + "learning_rate": 0.00044100300902708125, + "loss": 1.3991, + "theoretical_loss": 3.54467201670259, + "tokens_seen": 1499398144 + }, + { + "epoch": 0.45, + "learning_rate": 0.00044098696088264795, + "loss": 1.2946, + "theoretical_loss": 3.5446584655325073, + "tokens_seen": 1499463680 + }, + { + "epoch": 0.45, + "learning_rate": 0.0004409709127382146, + "loss": 1.4151, + "theoretical_loss": 3.5446449151205113, + "tokens_seen": 1499529216 + }, + { + "epoch": 0.45, + "learning_rate": 0.00044095486459378136, + "loss": 1.4019, + "theoretical_loss": 3.544631365466526, + "tokens_seen": 1499594752 + }, + { + "epoch": 0.45, + "learning_rate": 0.000440938816449348, + "loss": 1.4275, + "theoretical_loss": 3.5446178165704754, + "tokens_seen": 1499660288 + }, + { + "epoch": 0.45, + "learning_rate": 0.00044092276830491476, + "loss": 1.349, + "theoretical_loss": 3.544604268432285, + "tokens_seen": 1499725824 + }, + { + "epoch": 0.45, + "learning_rate": 0.0004409067201604814, + "loss": 1.4176, + "theoretical_loss": 3.544590721051878, + "tokens_seen": 1499791360 + }, + { + "epoch": 0.45, + "learning_rate": 0.00044089067201604816, + "loss": 1.3801, + "theoretical_loss": 3.5445771744291803, + "tokens_seen": 1499856896 + }, + { + "epoch": 0.45, + "learning_rate": 0.0004408746238716148, + "loss": 1.2976, + "theoretical_loss": 3.5445636285641156, + "tokens_seen": 1499922432 + }, + { + "epoch": 0.45, + "learning_rate": 0.00044085857572718156, + "loss": 1.3932, + "theoretical_loss": 3.544550083456608, + "tokens_seen": 1499987968 + }, + { + "epoch": 0.45, + "learning_rate": 0.0004408425275827482, + "loss": 1.5194, + "theoretical_loss": 3.5445365391065833, + "tokens_seen": 1500053504 + }, + { + "epoch": 0.45, + "learning_rate": 0.00044082647943831496, + "loss": 1.3761, + "theoretical_loss": 3.544522995513965, + "tokens_seen": 1500119040 + }, + { + "epoch": 0.45, + "learning_rate": 0.0004408104312938817, + "loss": 1.4658, + "theoretical_loss": 3.5445094526786782, + "tokens_seen": 1500184576 + }, + { + "epoch": 0.45, + "learning_rate": 0.00044079438314944836, + "loss": 1.4226, + "theoretical_loss": 3.5444959106006473, + "tokens_seen": 1500250112 + }, + { + "epoch": 0.45, + "learning_rate": 0.0004407783350050151, + "loss": 1.5108, + "theoretical_loss": 3.5444823692797973, + "tokens_seen": 1500315648 + }, + { + "epoch": 0.45, + "learning_rate": 0.00044076228686058176, + "loss": 1.349, + "theoretical_loss": 3.5444688287160515, + "tokens_seen": 1500381184 + }, + { + "epoch": 0.45, + "learning_rate": 0.0004407462387161485, + "loss": 1.336, + "theoretical_loss": 3.5444552889093357, + "tokens_seen": 1500446720 + }, + { + "epoch": 0.45, + "learning_rate": 0.00044073019057171516, + "loss": 1.4944, + "theoretical_loss": 3.5444417498595744, + "tokens_seen": 1500512256 + }, + { + "epoch": 0.45, + "learning_rate": 0.0004407141424272819, + "loss": 1.4026, + "theoretical_loss": 3.5444282115666916, + "tokens_seen": 1500577792 + }, + { + "epoch": 0.45, + "learning_rate": 0.00044069809428284857, + "loss": 1.3703, + "theoretical_loss": 3.5444146740306124, + "tokens_seen": 1500643328 + }, + { + "epoch": 0.45, + "learning_rate": 0.0004406820461384153, + "loss": 1.2788, + "theoretical_loss": 3.5444011372512616, + "tokens_seen": 1500708864 + }, + { + "epoch": 0.45, + "objective/train/docs_used": 548178, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.3915894031524658, + "objective/train/theoretical_loss": 3.5443876012285638, + "objective/train/tokens_used": 1521234400, + "theoretical_loss": 3.5443876012285638, + "tokens_seen": 1500774400 + }, + { + "epoch": 0.45, + "learning_rate": 0.00044066599799398197, + "loss": 1.2974, + "theoretical_loss": 3.5443876012285638, + "tokens_seen": 1500774400 + }, + { + "epoch": 0.45, + "learning_rate": 0.0004406499498495487, + "loss": 1.3742, + "theoretical_loss": 3.544374065962443, + "tokens_seen": 1500839936 + }, + { + "epoch": 0.45, + "learning_rate": 0.00044063390170511537, + "loss": 1.3524, + "theoretical_loss": 3.544360531452825, + "tokens_seen": 1500905472 + }, + { + "epoch": 0.45, + "learning_rate": 0.0004406178535606821, + "loss": 1.2917, + "theoretical_loss": 3.544346997699633, + "tokens_seen": 1500971008 + }, + { + "epoch": 0.45, + "learning_rate": 0.00044060180541624877, + "loss": 1.4081, + "theoretical_loss": 3.5443334647027935, + "tokens_seen": 1501036544 + }, + { + "epoch": 0.45, + "learning_rate": 0.0004405857572718155, + "loss": 1.4212, + "theoretical_loss": 3.54431993246223, + "tokens_seen": 1501102080 + }, + { + "epoch": 0.45, + "learning_rate": 0.00044056970912738217, + "loss": 1.4804, + "theoretical_loss": 3.5443064009778675, + "tokens_seen": 1501167616 + }, + { + "epoch": 0.45, + "learning_rate": 0.0004405536609829489, + "loss": 1.3466, + "theoretical_loss": 3.54429287024963, + "tokens_seen": 1501233152 + }, + { + "epoch": 0.45, + "learning_rate": 0.00044053761283851557, + "loss": 1.3728, + "theoretical_loss": 3.544279340277444, + "tokens_seen": 1501298688 + }, + { + "epoch": 0.45, + "learning_rate": 0.0004405215646940823, + "loss": 1.3863, + "theoretical_loss": 3.5442658110612326, + "tokens_seen": 1501364224 + }, + { + "epoch": 0.45, + "learning_rate": 0.00044050551654964897, + "loss": 1.3799, + "theoretical_loss": 3.5442522826009215, + "tokens_seen": 1501429760 + }, + { + "epoch": 0.45, + "learning_rate": 0.0004404894684052157, + "loss": 1.3052, + "theoretical_loss": 3.544238754896435, + "tokens_seen": 1501495296 + }, + { + "epoch": 0.46, + "learning_rate": 0.0004404734202607824, + "loss": 1.5138, + "theoretical_loss": 3.544225227947698, + "tokens_seen": 1501560832 + }, + { + "epoch": 0.46, + "learning_rate": 0.0004404573721163491, + "loss": 1.4025, + "theoretical_loss": 3.544211701754636, + "tokens_seen": 1501626368 + }, + { + "epoch": 0.46, + "learning_rate": 0.0004404413239719158, + "loss": 1.49, + "theoretical_loss": 3.5441981763171726, + "tokens_seen": 1501691904 + }, + { + "epoch": 0.46, + "learning_rate": 0.0004404252758274825, + "loss": 1.4451, + "theoretical_loss": 3.5441846516352333, + "tokens_seen": 1501757440 + }, + { + "epoch": 0.46, + "learning_rate": 0.0004404092276830492, + "loss": 1.3246, + "theoretical_loss": 3.5441711277087427, + "tokens_seen": 1501822976 + }, + { + "epoch": 0.46, + "learning_rate": 0.0004403931795386159, + "loss": 1.4879, + "theoretical_loss": 3.5441576045376255, + "tokens_seen": 1501888512 + }, + { + "epoch": 0.46, + "learning_rate": 0.0004403771313941826, + "loss": 1.354, + "theoretical_loss": 3.5441440821218073, + "tokens_seen": 1501954048 + }, + { + "epoch": 0.46, + "learning_rate": 0.0004403610832497493, + "loss": 1.3491, + "theoretical_loss": 3.544130560461212, + "tokens_seen": 1502019584 + }, + { + "epoch": 0.46, + "learning_rate": 0.0004403450351053159, + "loss": 1.392, + "theoretical_loss": 3.5441170395557653, + "tokens_seen": 1502085120 + }, + { + "epoch": 0.46, + "learning_rate": 0.0004403289869608827, + "loss": 1.3739, + "theoretical_loss": 3.5441035194053914, + "tokens_seen": 1502150656 + }, + { + "epoch": 0.46, + "learning_rate": 0.0004403129388164493, + "loss": 1.3658, + "theoretical_loss": 3.5440900000100157, + "tokens_seen": 1502216192 + }, + { + "epoch": 0.46, + "learning_rate": 0.0004402968906720161, + "loss": 1.3472, + "theoretical_loss": 3.5440764813695633, + "tokens_seen": 1502281728 + }, + { + "epoch": 0.46, + "learning_rate": 0.00044028084252758273, + "loss": 1.371, + "theoretical_loss": 3.5440629634839578, + "tokens_seen": 1502347264 + }, + { + "epoch": 0.46, + "learning_rate": 0.0004402647943831495, + "loss": 1.3547, + "theoretical_loss": 3.544049446353126, + "tokens_seen": 1502412800 + }, + { + "epoch": 0.46, + "learning_rate": 0.00044024874623871613, + "loss": 1.3638, + "theoretical_loss": 3.5440359299769915, + "tokens_seen": 1502478336 + }, + { + "epoch": 0.46, + "learning_rate": 0.0004402326980942829, + "loss": 1.5412, + "theoretical_loss": 3.5440224143554793, + "tokens_seen": 1502543872 + }, + { + "epoch": 0.46, + "learning_rate": 0.00044021664994984953, + "loss": 1.3891, + "theoretical_loss": 3.544008899488515, + "tokens_seen": 1502609408 + }, + { + "epoch": 0.46, + "learning_rate": 0.0004402006018054163, + "loss": 1.3014, + "theoretical_loss": 3.5439953853760233, + "tokens_seen": 1502674944 + }, + { + "epoch": 0.46, + "learning_rate": 0.00044018455366098293, + "loss": 1.3239, + "theoretical_loss": 3.543981872017929, + "tokens_seen": 1502740480 + }, + { + "epoch": 0.46, + "learning_rate": 0.0004401685055165497, + "loss": 1.4053, + "theoretical_loss": 3.5439683594141576, + "tokens_seen": 1502806016 + }, + { + "epoch": 0.46, + "learning_rate": 0.00044015245737211633, + "loss": 1.4323, + "theoretical_loss": 3.543954847564633, + "tokens_seen": 1502871552 + }, + { + "epoch": 0.46, + "learning_rate": 0.0004401364092276831, + "loss": 1.3647, + "theoretical_loss": 3.543941336469282, + "tokens_seen": 1502937088 + }, + { + "epoch": 0.46, + "learning_rate": 0.00044012036108324973, + "loss": 1.4172, + "theoretical_loss": 3.5439278261280274, + "tokens_seen": 1503002624 + }, + { + "epoch": 0.46, + "learning_rate": 0.0004401043129388165, + "loss": 1.4357, + "theoretical_loss": 3.5439143165407962, + "tokens_seen": 1503068160 + }, + { + "epoch": 0.46, + "learning_rate": 0.00044008826479438314, + "loss": 1.4034, + "theoretical_loss": 3.5439008077075123, + "tokens_seen": 1503133696 + }, + { + "epoch": 0.46, + "learning_rate": 0.0004400722166499499, + "loss": 1.3855, + "theoretical_loss": 3.5438872996281012, + "tokens_seen": 1503199232 + }, + { + "epoch": 0.46, + "learning_rate": 0.00044005616850551654, + "loss": 1.3991, + "theoretical_loss": 3.543873792302488, + "tokens_seen": 1503264768 + }, + { + "epoch": 0.46, + "learning_rate": 0.0004400401203610833, + "loss": 1.269, + "theoretical_loss": 3.5438602857305974, + "tokens_seen": 1503330304 + }, + { + "epoch": 0.46, + "learning_rate": 0.00044002407221664994, + "loss": 1.3752, + "theoretical_loss": 3.5438467799123545, + "tokens_seen": 1503395840 + }, + { + "epoch": 0.46, + "learning_rate": 0.0004400080240722167, + "loss": 1.3024, + "theoretical_loss": 3.5438332748476853, + "tokens_seen": 1503461376 + }, + { + "epoch": 0.46, + "learning_rate": 0.00043999197592778334, + "loss": 1.4129, + "theoretical_loss": 3.5438197705365138, + "tokens_seen": 1503526912 + }, + { + "epoch": 0.46, + "learning_rate": 0.0004399759277833501, + "loss": 1.374, + "theoretical_loss": 3.5438062669787653, + "tokens_seen": 1503592448 + }, + { + "epoch": 0.46, + "learning_rate": 0.00043995987963891674, + "loss": 1.2712, + "theoretical_loss": 3.5437927641743654, + "tokens_seen": 1503657984 + }, + { + "epoch": 0.46, + "learning_rate": 0.0004399438314944835, + "loss": 1.4153, + "theoretical_loss": 3.5437792621232393, + "tokens_seen": 1503723520 + }, + { + "epoch": 0.46, + "learning_rate": 0.00043992778335005014, + "loss": 1.3882, + "theoretical_loss": 3.5437657608253117, + "tokens_seen": 1503789056 + }, + { + "epoch": 0.46, + "learning_rate": 0.0004399117352056169, + "loss": 1.2796, + "theoretical_loss": 3.543752260280508, + "tokens_seen": 1503854592 + }, + { + "epoch": 0.46, + "learning_rate": 0.00043989568706118354, + "loss": 1.482, + "theoretical_loss": 3.543738760488753, + "tokens_seen": 1503920128 + }, + { + "epoch": 0.46, + "learning_rate": 0.0004398796389167503, + "loss": 1.4573, + "theoretical_loss": 3.543725261449972, + "tokens_seen": 1503985664 + }, + { + "epoch": 0.46, + "objective/train/docs_used": 549351, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.3680205345153809, + "objective/train/theoretical_loss": 3.543711763164091, + "objective/train/tokens_used": 1524511200, + "theoretical_loss": 3.543711763164091, + "tokens_seen": 1504051200 + }, + { + "epoch": 0.46, + "learning_rate": 0.00043986359077231694, + "loss": 1.3766, + "theoretical_loss": 3.543711763164091, + "tokens_seen": 1504051200 + }, + { + "epoch": 0.46, + "learning_rate": 0.0004398475426278837, + "loss": 1.2989, + "theoretical_loss": 3.5436982656310345, + "tokens_seen": 1504116736 + }, + { + "epoch": 0.46, + "learning_rate": 0.00043983149448345035, + "loss": 1.3481, + "theoretical_loss": 3.5436847688507274, + "tokens_seen": 1504182272 + }, + { + "epoch": 0.46, + "learning_rate": 0.0004398154463390171, + "loss": 1.3076, + "theoretical_loss": 3.5436712728230955, + "tokens_seen": 1504247808 + }, + { + "epoch": 0.46, + "learning_rate": 0.00043979939819458375, + "loss": 1.3492, + "theoretical_loss": 3.543657777548064, + "tokens_seen": 1504313344 + }, + { + "epoch": 0.46, + "learning_rate": 0.0004397833500501505, + "loss": 1.2798, + "theoretical_loss": 3.543644283025558, + "tokens_seen": 1504378880 + }, + { + "epoch": 0.46, + "learning_rate": 0.00043976730190571715, + "loss": 1.3928, + "theoretical_loss": 3.5436307892555026, + "tokens_seen": 1504444416 + }, + { + "epoch": 0.46, + "learning_rate": 0.0004397512537612839, + "loss": 1.2852, + "theoretical_loss": 3.543617296237823, + "tokens_seen": 1504509952 + }, + { + "epoch": 0.46, + "learning_rate": 0.00043973520561685055, + "loss": 1.3907, + "theoretical_loss": 3.543603803972445, + "tokens_seen": 1504575488 + }, + { + "epoch": 0.46, + "learning_rate": 0.00043971915747241725, + "loss": 1.34, + "theoretical_loss": 3.543590312459294, + "tokens_seen": 1504641024 + }, + { + "epoch": 0.46, + "learning_rate": 0.00043970310932798395, + "loss": 1.3295, + "theoretical_loss": 3.543576821698294, + "tokens_seen": 1504706560 + }, + { + "epoch": 0.46, + "learning_rate": 0.00043968706118355065, + "loss": 1.3124, + "theoretical_loss": 3.543563331689372, + "tokens_seen": 1504772096 + }, + { + "epoch": 0.46, + "learning_rate": 0.00043967101303911735, + "loss": 1.3897, + "theoretical_loss": 3.5435498424324523, + "tokens_seen": 1504837632 + }, + { + "epoch": 0.46, + "learning_rate": 0.00043965496489468405, + "loss": 1.3868, + "theoretical_loss": 3.5435363539274602, + "tokens_seen": 1504903168 + }, + { + "epoch": 0.46, + "learning_rate": 0.0004396389167502508, + "loss": 1.4162, + "theoretical_loss": 3.543522866174322, + "tokens_seen": 1504968704 + }, + { + "epoch": 0.46, + "learning_rate": 0.00043962286860581745, + "loss": 1.4334, + "theoretical_loss": 3.5435093791729617, + "tokens_seen": 1505034240 + }, + { + "epoch": 0.46, + "learning_rate": 0.0004396068204613842, + "loss": 1.4845, + "theoretical_loss": 3.543495892923305, + "tokens_seen": 1505099776 + }, + { + "epoch": 0.46, + "learning_rate": 0.00043959077231695085, + "loss": 1.3252, + "theoretical_loss": 3.5434824074252784, + "tokens_seen": 1505165312 + }, + { + "epoch": 0.46, + "learning_rate": 0.0004395747241725176, + "loss": 1.5062, + "theoretical_loss": 3.543468922678806, + "tokens_seen": 1505230848 + }, + { + "epoch": 0.46, + "learning_rate": 0.00043955867602808426, + "loss": 1.3066, + "theoretical_loss": 3.543455438683814, + "tokens_seen": 1505296384 + }, + { + "epoch": 0.46, + "learning_rate": 0.000439542627883651, + "loss": 1.2421, + "theoretical_loss": 3.5434419554402274, + "tokens_seen": 1505361920 + }, + { + "epoch": 0.46, + "learning_rate": 0.00043952657973921766, + "loss": 1.3988, + "theoretical_loss": 3.543428472947972, + "tokens_seen": 1505427456 + }, + { + "epoch": 0.46, + "learning_rate": 0.0004395105315947844, + "loss": 1.3741, + "theoretical_loss": 3.5434149912069723, + "tokens_seen": 1505492992 + }, + { + "epoch": 0.46, + "learning_rate": 0.00043949448345035106, + "loss": 1.2674, + "theoretical_loss": 3.543401510217155, + "tokens_seen": 1505558528 + }, + { + "epoch": 0.46, + "learning_rate": 0.0004394784353059178, + "loss": 1.3954, + "theoretical_loss": 3.5433880299784444, + "tokens_seen": 1505624064 + }, + { + "epoch": 0.46, + "learning_rate": 0.00043946238716148446, + "loss": 1.4185, + "theoretical_loss": 3.5433745504907668, + "tokens_seen": 1505689600 + }, + { + "epoch": 0.46, + "learning_rate": 0.0004394463390170512, + "loss": 1.3897, + "theoretical_loss": 3.543361071754047, + "tokens_seen": 1505755136 + }, + { + "epoch": 0.46, + "learning_rate": 0.00043943029087261786, + "loss": 1.3616, + "theoretical_loss": 3.5433475937682113, + "tokens_seen": 1505820672 + }, + { + "epoch": 0.46, + "learning_rate": 0.0004394142427281846, + "loss": 1.3311, + "theoretical_loss": 3.543334116533184, + "tokens_seen": 1505886208 + }, + { + "epoch": 0.46, + "learning_rate": 0.00043939819458375126, + "loss": 1.4253, + "theoretical_loss": 3.543320640048892, + "tokens_seen": 1505951744 + }, + { + "epoch": 0.46, + "learning_rate": 0.000439382146439318, + "loss": 1.3875, + "theoretical_loss": 3.54330716431526, + "tokens_seen": 1506017280 + }, + { + "epoch": 0.46, + "learning_rate": 0.00043936609829488466, + "loss": 1.215, + "theoretical_loss": 3.543293689332214, + "tokens_seen": 1506082816 + }, + { + "epoch": 0.46, + "learning_rate": 0.0004393500501504514, + "loss": 1.4795, + "theoretical_loss": 3.5432802150996787, + "tokens_seen": 1506148352 + }, + { + "epoch": 0.46, + "learning_rate": 0.00043933400200601806, + "loss": 1.282, + "theoretical_loss": 3.5432667416175803, + "tokens_seen": 1506213888 + }, + { + "epoch": 0.46, + "learning_rate": 0.0004393179538615848, + "loss": 1.3529, + "theoretical_loss": 3.5432532688858442, + "tokens_seen": 1506279424 + }, + { + "epoch": 0.46, + "learning_rate": 0.00043930190571715147, + "loss": 1.2572, + "theoretical_loss": 3.543239796904396, + "tokens_seen": 1506344960 + }, + { + "epoch": 0.46, + "learning_rate": 0.0004392858575727182, + "loss": 1.4216, + "theoretical_loss": 3.543226325673161, + "tokens_seen": 1506410496 + }, + { + "epoch": 0.46, + "learning_rate": 0.00043926980942828487, + "loss": 1.1933, + "theoretical_loss": 3.543212855192065, + "tokens_seen": 1506476032 + }, + { + "epoch": 0.46, + "learning_rate": 0.0004392537612838516, + "loss": 1.4052, + "theoretical_loss": 3.543199385461034, + "tokens_seen": 1506541568 + }, + { + "epoch": 0.46, + "learning_rate": 0.00043923771313941827, + "loss": 1.4235, + "theoretical_loss": 3.5431859164799935, + "tokens_seen": 1506607104 + }, + { + "epoch": 0.46, + "learning_rate": 0.000439221664994985, + "loss": 1.3895, + "theoretical_loss": 3.5431724482488685, + "tokens_seen": 1506672640 + }, + { + "epoch": 0.46, + "learning_rate": 0.00043920561685055167, + "loss": 1.3045, + "theoretical_loss": 3.543158980767585, + "tokens_seen": 1506738176 + }, + { + "epoch": 0.46, + "learning_rate": 0.0004391895687061184, + "loss": 1.4652, + "theoretical_loss": 3.543145514036069, + "tokens_seen": 1506803712 + }, + { + "epoch": 0.46, + "learning_rate": 0.00043917352056168507, + "loss": 1.3334, + "theoretical_loss": 3.543132048054245, + "tokens_seen": 1506869248 + }, + { + "epoch": 0.46, + "learning_rate": 0.0004391574724172518, + "loss": 1.4163, + "theoretical_loss": 3.54311858282204, + "tokens_seen": 1506934784 + }, + { + "epoch": 0.46, + "learning_rate": 0.00043914142427281847, + "loss": 1.3547, + "theoretical_loss": 3.543105118339379, + "tokens_seen": 1507000320 + }, + { + "epoch": 0.46, + "learning_rate": 0.00043912537612838523, + "loss": 1.3409, + "theoretical_loss": 3.5430916546061875, + "tokens_seen": 1507065856 + }, + { + "epoch": 0.46, + "learning_rate": 0.0004391093279839519, + "loss": 1.3752, + "theoretical_loss": 3.543078191622392, + "tokens_seen": 1507131392 + }, + { + "epoch": 0.46, + "learning_rate": 0.0004390932798395186, + "loss": 1.4101, + "theoretical_loss": 3.5430647293879174, + "tokens_seen": 1507196928 + }, + { + "epoch": 0.46, + "learning_rate": 0.0004390772316950853, + "loss": 1.4014, + "theoretical_loss": 3.5430512679026895, + "tokens_seen": 1507262464 + }, + { + "epoch": 0.46, + "objective/train/docs_used": 550589, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.4720523357391357, + "objective/train/theoretical_loss": 3.543037807166635, + "objective/train/tokens_used": 1527788000, + "theoretical_loss": 3.543037807166635, + "tokens_seen": 1507328000 + }, + { + "epoch": 0.46, + "learning_rate": 0.000439061183550652, + "loss": 1.3611, + "theoretical_loss": 3.543037807166635, + "tokens_seen": 1507328000 + }, + { + "epoch": 0.46, + "learning_rate": 0.0004390451354062187, + "loss": 1.3472, + "theoretical_loss": 3.5430243471796783, + "tokens_seen": 1507393536 + }, + { + "epoch": 0.46, + "learning_rate": 0.0004390290872617854, + "loss": 1.4484, + "theoretical_loss": 3.543010887941746, + "tokens_seen": 1507459072 + }, + { + "epoch": 0.46, + "learning_rate": 0.0004390130391173521, + "loss": 1.4307, + "theoretical_loss": 3.5429974294527633, + "tokens_seen": 1507524608 + }, + { + "epoch": 0.46, + "learning_rate": 0.0004389969909729188, + "loss": 1.5069, + "theoretical_loss": 3.5429839717126566, + "tokens_seen": 1507590144 + }, + { + "epoch": 0.46, + "learning_rate": 0.0004389809428284855, + "loss": 1.3945, + "theoretical_loss": 3.5429705147213513, + "tokens_seen": 1507655680 + }, + { + "epoch": 0.46, + "learning_rate": 0.0004389648946840522, + "loss": 1.4087, + "theoretical_loss": 3.542957058478773, + "tokens_seen": 1507721216 + }, + { + "epoch": 0.46, + "learning_rate": 0.0004389488465396188, + "loss": 1.3237, + "theoretical_loss": 3.5429436029848476, + "tokens_seen": 1507786752 + }, + { + "epoch": 0.46, + "learning_rate": 0.0004389327983951856, + "loss": 1.3255, + "theoretical_loss": 3.542930148239501, + "tokens_seen": 1507852288 + }, + { + "epoch": 0.46, + "learning_rate": 0.00043891675025075223, + "loss": 1.3367, + "theoretical_loss": 3.5429166942426598, + "tokens_seen": 1507917824 + }, + { + "epoch": 0.46, + "learning_rate": 0.000438900702106319, + "loss": 1.3672, + "theoretical_loss": 3.5429032409942485, + "tokens_seen": 1507983360 + }, + { + "epoch": 0.46, + "learning_rate": 0.00043888465396188563, + "loss": 1.353, + "theoretical_loss": 3.5428897884941932, + "tokens_seen": 1508048896 + }, + { + "epoch": 0.46, + "learning_rate": 0.0004388686058174524, + "loss": 1.3942, + "theoretical_loss": 3.5428763367424203, + "tokens_seen": 1508114432 + }, + { + "epoch": 0.46, + "learning_rate": 0.00043885255767301903, + "loss": 1.4488, + "theoretical_loss": 3.542862885738856, + "tokens_seen": 1508179968 + }, + { + "epoch": 0.46, + "learning_rate": 0.0004388365095285858, + "loss": 1.3883, + "theoretical_loss": 3.542849435483425, + "tokens_seen": 1508245504 + }, + { + "epoch": 0.46, + "learning_rate": 0.00043882046138415243, + "loss": 1.4371, + "theoretical_loss": 3.542835985976054, + "tokens_seen": 1508311040 + }, + { + "epoch": 0.46, + "learning_rate": 0.0004388044132397192, + "loss": 1.3212, + "theoretical_loss": 3.5428225372166686, + "tokens_seen": 1508376576 + }, + { + "epoch": 0.46, + "learning_rate": 0.00043878836509528583, + "loss": 1.4104, + "theoretical_loss": 3.542809089205195, + "tokens_seen": 1508442112 + }, + { + "epoch": 0.46, + "learning_rate": 0.0004387723169508526, + "loss": 1.3465, + "theoretical_loss": 3.5427956419415585, + "tokens_seen": 1508507648 + }, + { + "epoch": 0.46, + "learning_rate": 0.00043875626880641923, + "loss": 1.3367, + "theoretical_loss": 3.542782195425686, + "tokens_seen": 1508573184 + }, + { + "epoch": 0.46, + "learning_rate": 0.000438740220661986, + "loss": 1.3641, + "theoretical_loss": 3.542768749657503, + "tokens_seen": 1508638720 + }, + { + "epoch": 0.46, + "learning_rate": 0.00043872417251755263, + "loss": 1.3875, + "theoretical_loss": 3.5427553046369344, + "tokens_seen": 1508704256 + }, + { + "epoch": 0.46, + "learning_rate": 0.0004387081243731194, + "loss": 1.4204, + "theoretical_loss": 3.5427418603639076, + "tokens_seen": 1508769792 + }, + { + "epoch": 0.46, + "learning_rate": 0.00043869207622868604, + "loss": 1.3742, + "theoretical_loss": 3.542728416838348, + "tokens_seen": 1508835328 + }, + { + "epoch": 0.46, + "learning_rate": 0.0004386760280842528, + "loss": 1.3476, + "theoretical_loss": 3.542714974060182, + "tokens_seen": 1508900864 + }, + { + "epoch": 0.46, + "learning_rate": 0.00043865997993981944, + "loss": 1.5809, + "theoretical_loss": 3.542701532029335, + "tokens_seen": 1508966400 + }, + { + "epoch": 0.46, + "learning_rate": 0.0004386439317953862, + "loss": 1.4189, + "theoretical_loss": 3.5426880907457328, + "tokens_seen": 1509031936 + }, + { + "epoch": 0.46, + "learning_rate": 0.00043862788365095284, + "loss": 1.3514, + "theoretical_loss": 3.5426746502093023, + "tokens_seen": 1509097472 + }, + { + "epoch": 0.46, + "learning_rate": 0.0004386118355065196, + "loss": 1.3168, + "theoretical_loss": 3.542661210419969, + "tokens_seen": 1509163008 + }, + { + "epoch": 0.46, + "learning_rate": 0.00043859578736208624, + "loss": 1.3652, + "theoretical_loss": 3.542647771377659, + "tokens_seen": 1509228544 + }, + { + "epoch": 0.46, + "learning_rate": 0.000438579739217653, + "loss": 1.4191, + "theoretical_loss": 3.5426343330822982, + "tokens_seen": 1509294080 + }, + { + "epoch": 0.46, + "learning_rate": 0.00043856369107321964, + "loss": 1.3739, + "theoretical_loss": 3.5426208955338128, + "tokens_seen": 1509359616 + }, + { + "epoch": 0.46, + "learning_rate": 0.0004385476429287864, + "loss": 1.3673, + "theoretical_loss": 3.542607458732129, + "tokens_seen": 1509425152 + }, + { + "epoch": 0.46, + "learning_rate": 0.00043853159478435304, + "loss": 1.3898, + "theoretical_loss": 3.5425940226771724, + "tokens_seen": 1509490688 + }, + { + "epoch": 0.46, + "learning_rate": 0.0004385155466399198, + "loss": 1.3408, + "theoretical_loss": 3.54258058736887, + "tokens_seen": 1509556224 + }, + { + "epoch": 0.46, + "learning_rate": 0.00043849949849548644, + "loss": 1.3623, + "theoretical_loss": 3.5425671528071465, + "tokens_seen": 1509621760 + }, + { + "epoch": 0.46, + "learning_rate": 0.0004384834503510532, + "loss": 1.3481, + "theoretical_loss": 3.542553718991929, + "tokens_seen": 1509687296 + }, + { + "epoch": 0.46, + "learning_rate": 0.00043846740220661984, + "loss": 1.4156, + "theoretical_loss": 3.5425402859231436, + "tokens_seen": 1509752832 + }, + { + "epoch": 0.46, + "learning_rate": 0.0004384513540621866, + "loss": 1.4372, + "theoretical_loss": 3.542526853600717, + "tokens_seen": 1509818368 + }, + { + "epoch": 0.46, + "learning_rate": 0.0004384353059177533, + "loss": 1.4078, + "theoretical_loss": 3.5425134220245735, + "tokens_seen": 1509883904 + }, + { + "epoch": 0.46, + "learning_rate": 0.00043841925777332, + "loss": 1.3557, + "theoretical_loss": 3.5424999911946404, + "tokens_seen": 1509949440 + }, + { + "epoch": 0.46, + "learning_rate": 0.0004384032096288867, + "loss": 1.3202, + "theoretical_loss": 3.5424865611108443, + "tokens_seen": 1510014976 + }, + { + "epoch": 0.46, + "learning_rate": 0.0004383871614844534, + "loss": 1.5026, + "theoretical_loss": 3.542473131773111, + "tokens_seen": 1510080512 + }, + { + "epoch": 0.46, + "learning_rate": 0.0004383711133400201, + "loss": 1.3868, + "theoretical_loss": 3.542459703181366, + "tokens_seen": 1510146048 + }, + { + "epoch": 0.46, + "learning_rate": 0.0004383550651955868, + "loss": 1.4371, + "theoretical_loss": 3.5424462753355366, + "tokens_seen": 1510211584 + }, + { + "epoch": 0.46, + "learning_rate": 0.0004383390170511535, + "loss": 1.3767, + "theoretical_loss": 3.542432848235548, + "tokens_seen": 1510277120 + }, + { + "epoch": 0.46, + "learning_rate": 0.00043832296890672015, + "loss": 1.41, + "theoretical_loss": 3.5424194218813265, + "tokens_seen": 1510342656 + }, + { + "epoch": 0.46, + "learning_rate": 0.0004383069207622869, + "loss": 1.3518, + "theoretical_loss": 3.542405996272799, + "tokens_seen": 1510408192 + }, + { + "epoch": 0.46, + "learning_rate": 0.00043829087261785355, + "loss": 1.4275, + "theoretical_loss": 3.5423925714098914, + "tokens_seen": 1510473728 + }, + { + "epoch": 0.46, + "learning_rate": 0.0004382748244734203, + "loss": 1.4633, + "theoretical_loss": 3.5423791472925306, + "tokens_seen": 1510539264 + }, + { + "epoch": 0.46, + "objective/train/docs_used": 551665, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.389709234237671, + "objective/train/theoretical_loss": 3.542365723920641, + "objective/train/tokens_used": 1531064800, + "theoretical_loss": 3.542365723920641, + "tokens_seen": 1510604800 + }, + { + "epoch": 0.46, + "learning_rate": 0.00043825877632898695, + "loss": 1.3684, + "theoretical_loss": 3.542365723920641, + "tokens_seen": 1510604800 + }, + { + "epoch": 0.46, + "learning_rate": 0.0004382427281845537, + "loss": 1.3655, + "theoretical_loss": 3.5423523012941507, + "tokens_seen": 1510670336 + }, + { + "epoch": 0.46, + "learning_rate": 0.00043822668004012035, + "loss": 1.3597, + "theoretical_loss": 3.5423388794129855, + "tokens_seen": 1510735872 + }, + { + "epoch": 0.46, + "learning_rate": 0.0004382106318956871, + "loss": 1.3645, + "theoretical_loss": 3.5423254582770713, + "tokens_seen": 1510801408 + }, + { + "epoch": 0.46, + "learning_rate": 0.00043819458375125376, + "loss": 1.2325, + "theoretical_loss": 3.542312037886335, + "tokens_seen": 1510866944 + }, + { + "epoch": 0.46, + "learning_rate": 0.0004381785356068205, + "loss": 1.5039, + "theoretical_loss": 3.542298618240702, + "tokens_seen": 1510932480 + }, + { + "epoch": 0.46, + "learning_rate": 0.00043816248746238716, + "loss": 1.4462, + "theoretical_loss": 3.542285199340099, + "tokens_seen": 1510998016 + }, + { + "epoch": 0.46, + "learning_rate": 0.0004381464393179539, + "loss": 1.4316, + "theoretical_loss": 3.5422717811844526, + "tokens_seen": 1511063552 + }, + { + "epoch": 0.46, + "learning_rate": 0.00043813039117352056, + "loss": 1.3773, + "theoretical_loss": 3.5422583637736893, + "tokens_seen": 1511129088 + }, + { + "epoch": 0.46, + "learning_rate": 0.0004381143430290873, + "loss": 1.438, + "theoretical_loss": 3.5422449471077346, + "tokens_seen": 1511194624 + }, + { + "epoch": 0.46, + "learning_rate": 0.00043809829488465396, + "loss": 1.375, + "theoretical_loss": 3.542231531186516, + "tokens_seen": 1511260160 + }, + { + "epoch": 0.46, + "learning_rate": 0.0004380822467402207, + "loss": 1.3701, + "theoretical_loss": 3.542218116009959, + "tokens_seen": 1511325696 + }, + { + "epoch": 0.46, + "learning_rate": 0.00043806619859578736, + "loss": 1.3742, + "theoretical_loss": 3.5422047015779903, + "tokens_seen": 1511391232 + }, + { + "epoch": 0.46, + "learning_rate": 0.0004380501504513541, + "loss": 1.292, + "theoretical_loss": 3.542191287890536, + "tokens_seen": 1511456768 + }, + { + "epoch": 0.46, + "learning_rate": 0.00043803410230692076, + "loss": 1.342, + "theoretical_loss": 3.542177874947523, + "tokens_seen": 1511522304 + }, + { + "epoch": 0.46, + "learning_rate": 0.0004380180541624875, + "loss": 1.3392, + "theoretical_loss": 3.542164462748877, + "tokens_seen": 1511587840 + }, + { + "epoch": 0.46, + "learning_rate": 0.00043800200601805416, + "loss": 1.4351, + "theoretical_loss": 3.5421510512945256, + "tokens_seen": 1511653376 + }, + { + "epoch": 0.46, + "learning_rate": 0.0004379859578736209, + "loss": 1.3851, + "theoretical_loss": 3.542137640584394, + "tokens_seen": 1511718912 + }, + { + "epoch": 0.46, + "learning_rate": 0.00043796990972918756, + "loss": 1.4628, + "theoretical_loss": 3.542124230618409, + "tokens_seen": 1511784448 + }, + { + "epoch": 0.46, + "learning_rate": 0.0004379538615847543, + "loss": 1.3574, + "theoretical_loss": 3.5421108213964976, + "tokens_seen": 1511849984 + }, + { + "epoch": 0.46, + "learning_rate": 0.00043793781344032097, + "loss": 1.467, + "theoretical_loss": 3.5420974129185856, + "tokens_seen": 1511915520 + }, + { + "epoch": 0.46, + "learning_rate": 0.0004379217652958877, + "loss": 1.3756, + "theoretical_loss": 3.5420840051845994, + "tokens_seen": 1511981056 + }, + { + "epoch": 0.46, + "learning_rate": 0.00043790571715145437, + "loss": 1.4371, + "theoretical_loss": 3.5420705981944662, + "tokens_seen": 1512046592 + }, + { + "epoch": 0.46, + "learning_rate": 0.0004378896690070211, + "loss": 1.2973, + "theoretical_loss": 3.5420571919481123, + "tokens_seen": 1512112128 + }, + { + "epoch": 0.46, + "learning_rate": 0.00043787362086258777, + "loss": 1.3656, + "theoretical_loss": 3.542043786445464, + "tokens_seen": 1512177664 + }, + { + "epoch": 0.46, + "learning_rate": 0.0004378575727181545, + "loss": 1.4086, + "theoretical_loss": 3.5420303816864473, + "tokens_seen": 1512243200 + }, + { + "epoch": 0.46, + "learning_rate": 0.00043784152457372117, + "loss": 1.3667, + "theoretical_loss": 3.5420169776709898, + "tokens_seen": 1512308736 + }, + { + "epoch": 0.46, + "learning_rate": 0.0004378254764292879, + "loss": 1.3284, + "theoretical_loss": 3.542003574399017, + "tokens_seen": 1512374272 + }, + { + "epoch": 0.46, + "learning_rate": 0.00043780942828485457, + "loss": 1.4116, + "theoretical_loss": 3.5419901718704563, + "tokens_seen": 1512439808 + }, + { + "epoch": 0.46, + "learning_rate": 0.0004377933801404213, + "loss": 1.2773, + "theoretical_loss": 3.541976770085234, + "tokens_seen": 1512505344 + }, + { + "epoch": 0.46, + "learning_rate": 0.00043777733199598797, + "loss": 1.3376, + "theoretical_loss": 3.541963369043276, + "tokens_seen": 1512570880 + }, + { + "epoch": 0.46, + "learning_rate": 0.0004377612838515547, + "loss": 1.419, + "theoretical_loss": 3.5419499687445093, + "tokens_seen": 1512636416 + }, + { + "epoch": 0.46, + "learning_rate": 0.0004377452357071214, + "loss": 1.3279, + "theoretical_loss": 3.541936569188861, + "tokens_seen": 1512701952 + }, + { + "epoch": 0.46, + "learning_rate": 0.00043772918756268813, + "loss": 1.4505, + "theoretical_loss": 3.5419231703762577, + "tokens_seen": 1512767488 + }, + { + "epoch": 0.46, + "learning_rate": 0.0004377131394182548, + "loss": 1.2918, + "theoretical_loss": 3.5419097723066253, + "tokens_seen": 1512833024 + }, + { + "epoch": 0.46, + "learning_rate": 0.0004376970912738215, + "loss": 1.4041, + "theoretical_loss": 3.54189637497989, + "tokens_seen": 1512898560 + }, + { + "epoch": 0.46, + "learning_rate": 0.0004376810431293882, + "loss": 1.4042, + "theoretical_loss": 3.5418829783959804, + "tokens_seen": 1512964096 + }, + { + "epoch": 0.46, + "learning_rate": 0.0004376649949849549, + "loss": 1.3377, + "theoretical_loss": 3.5418695825548214, + "tokens_seen": 1513029632 + }, + { + "epoch": 0.46, + "learning_rate": 0.0004376489468405216, + "loss": 1.3652, + "theoretical_loss": 3.54185618745634, + "tokens_seen": 1513095168 + }, + { + "epoch": 0.46, + "learning_rate": 0.0004376328986960883, + "loss": 1.4598, + "theoretical_loss": 3.541842793100463, + "tokens_seen": 1513160704 + }, + { + "epoch": 0.46, + "learning_rate": 0.000437616850551655, + "loss": 1.3805, + "theoretical_loss": 3.541829399487117, + "tokens_seen": 1513226240 + }, + { + "epoch": 0.46, + "learning_rate": 0.0004376008024072217, + "loss": 1.3698, + "theoretical_loss": 3.541816006616229, + "tokens_seen": 1513291776 + }, + { + "epoch": 0.46, + "learning_rate": 0.0004375847542627884, + "loss": 1.3437, + "theoretical_loss": 3.5418026144877257, + "tokens_seen": 1513357312 + }, + { + "epoch": 0.46, + "learning_rate": 0.0004375687061183551, + "loss": 1.3713, + "theoretical_loss": 3.5417892231015333, + "tokens_seen": 1513422848 + }, + { + "epoch": 0.46, + "learning_rate": 0.0004375526579739217, + "loss": 1.4364, + "theoretical_loss": 3.5417758324575788, + "tokens_seen": 1513488384 + }, + { + "epoch": 0.46, + "learning_rate": 0.0004375366098294885, + "loss": 1.3842, + "theoretical_loss": 3.5417624425557888, + "tokens_seen": 1513553920 + }, + { + "epoch": 0.46, + "learning_rate": 0.00043752056168505513, + "loss": 1.4257, + "theoretical_loss": 3.54174905339609, + "tokens_seen": 1513619456 + }, + { + "epoch": 0.46, + "learning_rate": 0.0004375045135406219, + "loss": 1.4103, + "theoretical_loss": 3.5417356649784097, + "tokens_seen": 1513684992 + }, + { + "epoch": 0.46, + "learning_rate": 0.00043748846539618853, + "loss": 1.4401, + "theoretical_loss": 3.541722277302674, + "tokens_seen": 1513750528 + }, + { + "epoch": 0.46, + "learning_rate": 0.0004374724172517553, + "loss": 1.3106, + "theoretical_loss": 3.5417088903688096, + "tokens_seen": 1513816064 + }, + { + "epoch": 0.46, + "objective/train/docs_used": 552897, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.233246088027954, + "objective/train/theoretical_loss": 3.541695504176744, + "objective/train/tokens_used": 1534341600, + "theoretical_loss": 3.541695504176744, + "tokens_seen": 1513881600 + }, + { + "epoch": 0.46, + "learning_rate": 0.00043745636910732193, + "loss": 1.3284, + "theoretical_loss": 3.541695504176744, + "tokens_seen": 1513881600 + }, + { + "epoch": 0.46, + "learning_rate": 0.0004374403209628887, + "loss": 1.3035, + "theoretical_loss": 3.5416821187264036, + "tokens_seen": 1513947136 + }, + { + "epoch": 0.46, + "learning_rate": 0.00043742427281845533, + "loss": 1.3068, + "theoretical_loss": 3.541668734017715, + "tokens_seen": 1514012672 + }, + { + "epoch": 0.46, + "learning_rate": 0.0004374082246740221, + "loss": 1.4314, + "theoretical_loss": 3.541655350050605, + "tokens_seen": 1514078208 + }, + { + "epoch": 0.46, + "learning_rate": 0.00043739217652958873, + "loss": 1.3455, + "theoretical_loss": 3.5416419668250008, + "tokens_seen": 1514143744 + }, + { + "epoch": 0.46, + "learning_rate": 0.0004373761283851555, + "loss": 1.4621, + "theoretical_loss": 3.5416285843408293, + "tokens_seen": 1514209280 + }, + { + "epoch": 0.46, + "learning_rate": 0.00043736008024072213, + "loss": 1.3502, + "theoretical_loss": 3.5416152025980163, + "tokens_seen": 1514274816 + }, + { + "epoch": 0.46, + "learning_rate": 0.0004373440320962889, + "loss": 1.3171, + "theoretical_loss": 3.54160182159649, + "tokens_seen": 1514340352 + }, + { + "epoch": 0.46, + "learning_rate": 0.00043732798395185554, + "loss": 1.3567, + "theoretical_loss": 3.5415884413361765, + "tokens_seen": 1514405888 + }, + { + "epoch": 0.46, + "learning_rate": 0.0004373119358074223, + "loss": 1.4095, + "theoretical_loss": 3.541575061817003, + "tokens_seen": 1514471424 + }, + { + "epoch": 0.46, + "learning_rate": 0.00043729588766298894, + "loss": 1.3544, + "theoretical_loss": 3.5415616830388954, + "tokens_seen": 1514536960 + }, + { + "epoch": 0.46, + "learning_rate": 0.0004372798395185557, + "loss": 1.3692, + "theoretical_loss": 3.541548305001782, + "tokens_seen": 1514602496 + }, + { + "epoch": 0.46, + "learning_rate": 0.00043726379137412234, + "loss": 1.4152, + "theoretical_loss": 3.541534927705589, + "tokens_seen": 1514668032 + }, + { + "epoch": 0.46, + "learning_rate": 0.0004372477432296891, + "loss": 1.4333, + "theoretical_loss": 3.5415215511502436, + "tokens_seen": 1514733568 + }, + { + "epoch": 0.46, + "learning_rate": 0.00043723169508525585, + "loss": 1.2701, + "theoretical_loss": 3.5415081753356725, + "tokens_seen": 1514799104 + }, + { + "epoch": 0.46, + "learning_rate": 0.0004372156469408225, + "loss": 1.4064, + "theoretical_loss": 3.5414948002618023, + "tokens_seen": 1514864640 + }, + { + "epoch": 0.46, + "learning_rate": 0.00043719959879638925, + "loss": 1.3981, + "theoretical_loss": 3.54148142592856, + "tokens_seen": 1514930176 + }, + { + "epoch": 0.46, + "learning_rate": 0.0004371835506519559, + "loss": 1.3909, + "theoretical_loss": 3.5414680523358735, + "tokens_seen": 1514995712 + }, + { + "epoch": 0.46, + "learning_rate": 0.00043716750250752265, + "loss": 1.4628, + "theoretical_loss": 3.541454679483669, + "tokens_seen": 1515061248 + }, + { + "epoch": 0.46, + "learning_rate": 0.0004371514543630893, + "loss": 1.3685, + "theoretical_loss": 3.5414413073718736, + "tokens_seen": 1515126784 + }, + { + "epoch": 0.46, + "learning_rate": 0.00043713540621865605, + "loss": 1.3331, + "theoretical_loss": 3.5414279360004137, + "tokens_seen": 1515192320 + }, + { + "epoch": 0.46, + "learning_rate": 0.0004371193580742227, + "loss": 1.3349, + "theoretical_loss": 3.5414145653692177, + "tokens_seen": 1515257856 + }, + { + "epoch": 0.46, + "learning_rate": 0.00043710330992978945, + "loss": 1.4734, + "theoretical_loss": 3.541401195478211, + "tokens_seen": 1515323392 + }, + { + "epoch": 0.46, + "learning_rate": 0.0004370872617853561, + "loss": 1.2425, + "theoretical_loss": 3.5413878263273215, + "tokens_seen": 1515388928 + }, + { + "epoch": 0.46, + "learning_rate": 0.00043707121364092285, + "loss": 1.4169, + "theoretical_loss": 3.5413744579164765, + "tokens_seen": 1515454464 + }, + { + "epoch": 0.46, + "learning_rate": 0.0004370551654964895, + "loss": 1.4013, + "theoretical_loss": 3.5413610902456023, + "tokens_seen": 1515520000 + }, + { + "epoch": 0.46, + "learning_rate": 0.0004370391173520562, + "loss": 1.3707, + "theoretical_loss": 3.5413477233146264, + "tokens_seen": 1515585536 + }, + { + "epoch": 0.46, + "learning_rate": 0.0004370230692076229, + "loss": 1.4218, + "theoretical_loss": 3.5413343571234757, + "tokens_seen": 1515651072 + }, + { + "epoch": 0.46, + "learning_rate": 0.0004370070210631896, + "loss": 1.4799, + "theoretical_loss": 3.5413209916720776, + "tokens_seen": 1515716608 + }, + { + "epoch": 0.46, + "learning_rate": 0.0004369909729187563, + "loss": 1.3769, + "theoretical_loss": 3.5413076269603585, + "tokens_seen": 1515782144 + }, + { + "epoch": 0.46, + "learning_rate": 0.000436974924774323, + "loss": 1.2839, + "theoretical_loss": 3.541294262988246, + "tokens_seen": 1515847680 + }, + { + "epoch": 0.46, + "learning_rate": 0.0004369588766298897, + "loss": 1.3128, + "theoretical_loss": 3.541280899755667, + "tokens_seen": 1515913216 + }, + { + "epoch": 0.46, + "learning_rate": 0.0004369428284854564, + "loss": 1.3855, + "theoretical_loss": 3.5412675372625486, + "tokens_seen": 1515978752 + }, + { + "epoch": 0.46, + "learning_rate": 0.00043692678034102305, + "loss": 1.466, + "theoretical_loss": 3.5412541755088176, + "tokens_seen": 1516044288 + }, + { + "epoch": 0.46, + "learning_rate": 0.0004369107321965898, + "loss": 1.3568, + "theoretical_loss": 3.541240814494402, + "tokens_seen": 1516109824 + }, + { + "epoch": 0.46, + "learning_rate": 0.00043689468405215645, + "loss": 1.3085, + "theoretical_loss": 3.5412274542192286, + "tokens_seen": 1516175360 + }, + { + "epoch": 0.46, + "learning_rate": 0.0004368786359077232, + "loss": 1.3383, + "theoretical_loss": 3.541214094683224, + "tokens_seen": 1516240896 + }, + { + "epoch": 0.46, + "learning_rate": 0.00043686258776328985, + "loss": 1.3434, + "theoretical_loss": 3.541200735886316, + "tokens_seen": 1516306432 + }, + { + "epoch": 0.46, + "learning_rate": 0.0004368465396188566, + "loss": 1.397, + "theoretical_loss": 3.541187377828431, + "tokens_seen": 1516371968 + }, + { + "epoch": 0.46, + "learning_rate": 0.00043683049147442326, + "loss": 1.3991, + "theoretical_loss": 3.541174020509497, + "tokens_seen": 1516437504 + }, + { + "epoch": 0.46, + "learning_rate": 0.00043681444332999, + "loss": 1.3717, + "theoretical_loss": 3.541160663929441, + "tokens_seen": 1516503040 + }, + { + "epoch": 0.46, + "learning_rate": 0.00043679839518555666, + "loss": 1.3075, + "theoretical_loss": 3.54114730808819, + "tokens_seen": 1516568576 + }, + { + "epoch": 0.46, + "learning_rate": 0.0004367823470411234, + "loss": 1.4708, + "theoretical_loss": 3.5411339529856716, + "tokens_seen": 1516634112 + }, + { + "epoch": 0.46, + "learning_rate": 0.00043676629889669006, + "loss": 1.2059, + "theoretical_loss": 3.5411205986218124, + "tokens_seen": 1516699648 + }, + { + "epoch": 0.46, + "learning_rate": 0.0004367502507522568, + "loss": 1.3199, + "theoretical_loss": 3.5411072449965397, + "tokens_seen": 1516765184 + }, + { + "epoch": 0.46, + "learning_rate": 0.00043673420260782346, + "loss": 1.323, + "theoretical_loss": 3.5410938921097816, + "tokens_seen": 1516830720 + }, + { + "epoch": 0.46, + "learning_rate": 0.0004367181544633902, + "loss": 1.4342, + "theoretical_loss": 3.5410805399614644, + "tokens_seen": 1516896256 + }, + { + "epoch": 0.46, + "learning_rate": 0.00043670210631895686, + "loss": 1.4286, + "theoretical_loss": 3.5410671885515157, + "tokens_seen": 1516961792 + }, + { + "epoch": 0.46, + "learning_rate": 0.0004366860581745236, + "loss": 1.2962, + "theoretical_loss": 3.541053837879863, + "tokens_seen": 1517027328 + }, + { + "epoch": 0.46, + "learning_rate": 0.00043667001003009026, + "loss": 1.349, + "theoretical_loss": 3.5410404879464332, + "tokens_seen": 1517092864 + }, + { + "epoch": 0.46, + "objective/train/docs_used": 554075, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.4171507358551025, + "objective/train/theoretical_loss": 3.5410271387511534, + "objective/train/tokens_used": 1537618400, + "theoretical_loss": 3.5410271387511534, + "tokens_seen": 1517158400 + }, + { + "epoch": 0.46, + "learning_rate": 0.000436653961885657, + "loss": 1.4054, + "theoretical_loss": 3.5410271387511534, + "tokens_seen": 1517158400 + }, + { + "epoch": 0.46, + "learning_rate": 0.00043663791374122366, + "loss": 1.3989, + "theoretical_loss": 3.5410137902939516, + "tokens_seen": 1517223936 + }, + { + "epoch": 0.46, + "learning_rate": 0.0004366218655967904, + "loss": 1.4106, + "theoretical_loss": 3.541000442574755, + "tokens_seen": 1517289472 + }, + { + "epoch": 0.46, + "learning_rate": 0.00043660581745235706, + "loss": 1.4054, + "theoretical_loss": 3.54098709559349, + "tokens_seen": 1517355008 + }, + { + "epoch": 0.46, + "learning_rate": 0.0004365897693079238, + "loss": 1.4093, + "theoretical_loss": 3.5409737493500852, + "tokens_seen": 1517420544 + }, + { + "epoch": 0.46, + "learning_rate": 0.00043657372116349047, + "loss": 1.3643, + "theoretical_loss": 3.5409604038444673, + "tokens_seen": 1517486080 + }, + { + "epoch": 0.46, + "learning_rate": 0.0004365576730190572, + "loss": 1.3937, + "theoretical_loss": 3.5409470590765633, + "tokens_seen": 1517551616 + }, + { + "epoch": 0.46, + "learning_rate": 0.00043654162487462387, + "loss": 1.4419, + "theoretical_loss": 3.540933715046301, + "tokens_seen": 1517617152 + }, + { + "epoch": 0.46, + "learning_rate": 0.0004365255767301906, + "loss": 1.373, + "theoretical_loss": 3.540920371753608, + "tokens_seen": 1517682688 + }, + { + "epoch": 0.46, + "learning_rate": 0.00043650952858575727, + "loss": 1.3152, + "theoretical_loss": 3.540907029198412, + "tokens_seen": 1517748224 + }, + { + "epoch": 0.46, + "learning_rate": 0.000436493480441324, + "loss": 1.3103, + "theoretical_loss": 3.540893687380639, + "tokens_seen": 1517813760 + }, + { + "epoch": 0.46, + "learning_rate": 0.00043647743229689067, + "loss": 1.453, + "theoretical_loss": 3.540880346300218, + "tokens_seen": 1517879296 + }, + { + "epoch": 0.46, + "learning_rate": 0.0004364613841524574, + "loss": 1.3322, + "theoretical_loss": 3.5408670059570753, + "tokens_seen": 1517944832 + }, + { + "epoch": 0.46, + "learning_rate": 0.00043644533600802407, + "loss": 1.3704, + "theoretical_loss": 3.540853666351138, + "tokens_seen": 1518010368 + }, + { + "epoch": 0.46, + "learning_rate": 0.0004364292878635908, + "loss": 1.3648, + "theoretical_loss": 3.5408403274823352, + "tokens_seen": 1518075904 + }, + { + "epoch": 0.46, + "learning_rate": 0.00043641323971915747, + "loss": 1.4167, + "theoretical_loss": 3.540826989350593, + "tokens_seen": 1518141440 + }, + { + "epoch": 0.46, + "learning_rate": 0.0004363971915747242, + "loss": 1.3069, + "theoretical_loss": 3.5408136519558395, + "tokens_seen": 1518206976 + }, + { + "epoch": 0.46, + "learning_rate": 0.0004363811434302909, + "loss": 1.2989, + "theoretical_loss": 3.5408003152980014, + "tokens_seen": 1518272512 + }, + { + "epoch": 0.46, + "learning_rate": 0.00043636509528585763, + "loss": 1.2784, + "theoretical_loss": 3.540786979377007, + "tokens_seen": 1518338048 + }, + { + "epoch": 0.46, + "learning_rate": 0.0004363490471414243, + "loss": 1.4797, + "theoretical_loss": 3.5407736441927833, + "tokens_seen": 1518403584 + }, + { + "epoch": 0.46, + "learning_rate": 0.00043633299899699103, + "loss": 1.3821, + "theoretical_loss": 3.540760309745258, + "tokens_seen": 1518469120 + }, + { + "epoch": 0.46, + "learning_rate": 0.0004363169508525577, + "loss": 1.4042, + "theoretical_loss": 3.5407469760343586, + "tokens_seen": 1518534656 + }, + { + "epoch": 0.46, + "learning_rate": 0.00043630090270812443, + "loss": 1.3725, + "theoretical_loss": 3.5407336430600127, + "tokens_seen": 1518600192 + }, + { + "epoch": 0.46, + "learning_rate": 0.0004362848545636911, + "loss": 1.3104, + "theoretical_loss": 3.5407203108221474, + "tokens_seen": 1518665728 + }, + { + "epoch": 0.46, + "learning_rate": 0.0004362688064192578, + "loss": 1.3931, + "theoretical_loss": 3.540706979320691, + "tokens_seen": 1518731264 + }, + { + "epoch": 0.46, + "learning_rate": 0.0004362527582748245, + "loss": 1.4821, + "theoretical_loss": 3.5406936485555702, + "tokens_seen": 1518796800 + }, + { + "epoch": 0.46, + "learning_rate": 0.0004362367101303912, + "loss": 1.4041, + "theoretical_loss": 3.5406803185267135, + "tokens_seen": 1518862336 + }, + { + "epoch": 0.46, + "learning_rate": 0.0004362206619859579, + "loss": 1.408, + "theoretical_loss": 3.5406669892340474, + "tokens_seen": 1518927872 + }, + { + "epoch": 0.46, + "learning_rate": 0.0004362046138415246, + "loss": 1.3563, + "theoretical_loss": 3.5406536606775, + "tokens_seen": 1518993408 + }, + { + "epoch": 0.46, + "learning_rate": 0.0004361885656970913, + "loss": 1.4587, + "theoretical_loss": 3.540640332856999, + "tokens_seen": 1519058944 + }, + { + "epoch": 0.46, + "learning_rate": 0.000436172517552658, + "loss": 1.3519, + "theoretical_loss": 3.540627005772472, + "tokens_seen": 1519124480 + }, + { + "epoch": 0.46, + "learning_rate": 0.00043615646940822463, + "loss": 1.3331, + "theoretical_loss": 3.5406136794238465, + "tokens_seen": 1519190016 + }, + { + "epoch": 0.46, + "learning_rate": 0.0004361404212637914, + "loss": 1.4252, + "theoretical_loss": 3.5406003538110506, + "tokens_seen": 1519255552 + }, + { + "epoch": 0.46, + "learning_rate": 0.00043612437311935803, + "loss": 1.4014, + "theoretical_loss": 3.540587028934011, + "tokens_seen": 1519321088 + }, + { + "epoch": 0.46, + "learning_rate": 0.0004361083249749248, + "loss": 1.3536, + "theoretical_loss": 3.5405737047926555, + "tokens_seen": 1519386624 + }, + { + "epoch": 0.46, + "learning_rate": 0.00043609227683049143, + "loss": 1.3559, + "theoretical_loss": 3.540560381386913, + "tokens_seen": 1519452160 + }, + { + "epoch": 0.46, + "learning_rate": 0.0004360762286860582, + "loss": 1.2447, + "theoretical_loss": 3.5405470587167094, + "tokens_seen": 1519517696 + }, + { + "epoch": 0.46, + "learning_rate": 0.00043606018054162483, + "loss": 1.3265, + "theoretical_loss": 3.540533736781973, + "tokens_seen": 1519583232 + }, + { + "epoch": 0.46, + "learning_rate": 0.0004360441323971916, + "loss": 1.4125, + "theoretical_loss": 3.5405204155826326, + "tokens_seen": 1519648768 + }, + { + "epoch": 0.46, + "learning_rate": 0.00043602808425275834, + "loss": 1.3782, + "theoretical_loss": 3.5405070951186146, + "tokens_seen": 1519714304 + }, + { + "epoch": 0.46, + "learning_rate": 0.000436012036108325, + "loss": 1.4388, + "theoretical_loss": 3.540493775389847, + "tokens_seen": 1519779840 + }, + { + "epoch": 0.46, + "learning_rate": 0.00043599598796389174, + "loss": 1.2708, + "theoretical_loss": 3.5404804563962573, + "tokens_seen": 1519845376 + }, + { + "epoch": 0.46, + "learning_rate": 0.0004359799398194584, + "loss": 1.3689, + "theoretical_loss": 3.5404671381377737, + "tokens_seen": 1519910912 + }, + { + "epoch": 0.46, + "learning_rate": 0.00043596389167502514, + "loss": 1.4675, + "theoretical_loss": 3.5404538206143243, + "tokens_seen": 1519976448 + }, + { + "epoch": 0.46, + "learning_rate": 0.0004359478435305918, + "loss": 1.3568, + "theoretical_loss": 3.5404405038258355, + "tokens_seen": 1520041984 + }, + { + "epoch": 0.46, + "learning_rate": 0.00043593179538615855, + "loss": 1.3774, + "theoretical_loss": 3.540427187772236, + "tokens_seen": 1520107520 + }, + { + "epoch": 0.46, + "learning_rate": 0.0004359157472417252, + "loss": 1.3897, + "theoretical_loss": 3.540413872453454, + "tokens_seen": 1520173056 + }, + { + "epoch": 0.46, + "learning_rate": 0.00043589969909729195, + "loss": 1.1666, + "theoretical_loss": 3.5404005578694164, + "tokens_seen": 1520238592 + }, + { + "epoch": 0.46, + "learning_rate": 0.0004358836509528586, + "loss": 1.396, + "theoretical_loss": 3.5403872440200512, + "tokens_seen": 1520304128 + }, + { + "epoch": 0.46, + "learning_rate": 0.00043586760280842535, + "loss": 1.311, + "theoretical_loss": 3.540373930905286, + "tokens_seen": 1520369664 + }, + { + "epoch": 0.46, + "objective/train/docs_used": 555119, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.426171064376831, + "objective/train/theoretical_loss": 3.5403606185250496, + "objective/train/tokens_used": 1540895200, + "theoretical_loss": 3.5403606185250496, + "tokens_seen": 1520435200 + }, + { + "epoch": 0.46, + "learning_rate": 0.000435851554663992, + "loss": 1.4629, + "theoretical_loss": 3.5403606185250496, + "tokens_seen": 1520435200 + }, + { + "epoch": 0.46, + "learning_rate": 0.00043583550651955875, + "loss": 1.3607, + "theoretical_loss": 3.5403473068792684, + "tokens_seen": 1520500736 + }, + { + "epoch": 0.46, + "learning_rate": 0.0004358194583751254, + "loss": 1.3659, + "theoretical_loss": 3.5403339959678712, + "tokens_seen": 1520566272 + }, + { + "epoch": 0.46, + "learning_rate": 0.00043580341023069215, + "loss": 1.3135, + "theoretical_loss": 3.5403206857907854, + "tokens_seen": 1520631808 + }, + { + "epoch": 0.46, + "learning_rate": 0.0004357873620862588, + "loss": 1.2993, + "theoretical_loss": 3.540307376347939, + "tokens_seen": 1520697344 + }, + { + "epoch": 0.46, + "learning_rate": 0.00043577131394182555, + "loss": 1.4457, + "theoretical_loss": 3.54029406763926, + "tokens_seen": 1520762880 + }, + { + "epoch": 0.46, + "learning_rate": 0.0004357552657973922, + "loss": 1.5051, + "theoretical_loss": 3.540280759664676, + "tokens_seen": 1520828416 + }, + { + "epoch": 0.46, + "learning_rate": 0.00043573921765295895, + "loss": 1.4239, + "theoretical_loss": 3.540267452424115, + "tokens_seen": 1520893952 + }, + { + "epoch": 0.46, + "learning_rate": 0.0004357231695085256, + "loss": 1.3654, + "theoretical_loss": 3.540254145917505, + "tokens_seen": 1520959488 + }, + { + "epoch": 0.46, + "learning_rate": 0.00043570712136409235, + "loss": 1.3922, + "theoretical_loss": 3.540240840144774, + "tokens_seen": 1521025024 + }, + { + "epoch": 0.46, + "learning_rate": 0.000435691073219659, + "loss": 1.3328, + "theoretical_loss": 3.540227535105849, + "tokens_seen": 1521090560 + }, + { + "epoch": 0.46, + "learning_rate": 0.00043567502507522576, + "loss": 1.2892, + "theoretical_loss": 3.5402142308006592, + "tokens_seen": 1521156096 + }, + { + "epoch": 0.46, + "learning_rate": 0.0004356589769307924, + "loss": 1.367, + "theoretical_loss": 3.540200927229132, + "tokens_seen": 1521221632 + }, + { + "epoch": 0.46, + "learning_rate": 0.0004356429287863591, + "loss": 1.4143, + "theoretical_loss": 3.540187624391195, + "tokens_seen": 1521287168 + }, + { + "epoch": 0.46, + "learning_rate": 0.0004356268806419258, + "loss": 1.3876, + "theoretical_loss": 3.5401743222867763, + "tokens_seen": 1521352704 + }, + { + "epoch": 0.46, + "learning_rate": 0.0004356108324974925, + "loss": 1.4594, + "theoretical_loss": 3.540161020915804, + "tokens_seen": 1521418240 + }, + { + "epoch": 0.46, + "learning_rate": 0.0004355947843530592, + "loss": 1.3021, + "theoretical_loss": 3.540147720278206, + "tokens_seen": 1521483776 + }, + { + "epoch": 0.46, + "learning_rate": 0.0004355787362086259, + "loss": 1.3361, + "theoretical_loss": 3.540134420373911, + "tokens_seen": 1521549312 + }, + { + "epoch": 0.46, + "learning_rate": 0.0004355626880641926, + "loss": 1.4402, + "theoretical_loss": 3.5401211212028456, + "tokens_seen": 1521614848 + }, + { + "epoch": 0.46, + "learning_rate": 0.0004355466399197593, + "loss": 1.302, + "theoretical_loss": 3.5401078227649387, + "tokens_seen": 1521680384 + }, + { + "epoch": 0.46, + "learning_rate": 0.000435530591775326, + "loss": 1.3479, + "theoretical_loss": 3.5400945250601183, + "tokens_seen": 1521745920 + }, + { + "epoch": 0.46, + "learning_rate": 0.0004355145436308927, + "loss": 1.2707, + "theoretical_loss": 3.540081228088312, + "tokens_seen": 1521811456 + }, + { + "epoch": 0.46, + "learning_rate": 0.00043549849548645935, + "loss": 1.3718, + "theoretical_loss": 3.540067931849448, + "tokens_seen": 1521876992 + }, + { + "epoch": 0.46, + "learning_rate": 0.0004354824473420261, + "loss": 1.3753, + "theoretical_loss": 3.540054636343455, + "tokens_seen": 1521942528 + }, + { + "epoch": 0.46, + "learning_rate": 0.00043546639919759276, + "loss": 1.5256, + "theoretical_loss": 3.5400413415702596, + "tokens_seen": 1522008064 + }, + { + "epoch": 0.46, + "learning_rate": 0.0004354503510531595, + "loss": 1.4741, + "theoretical_loss": 3.5400280475297916, + "tokens_seen": 1522073600 + }, + { + "epoch": 0.46, + "learning_rate": 0.00043543430290872616, + "loss": 1.413, + "theoretical_loss": 3.5400147542219775, + "tokens_seen": 1522139136 + }, + { + "epoch": 0.46, + "learning_rate": 0.0004354182547642929, + "loss": 1.3514, + "theoretical_loss": 3.5400014616467463, + "tokens_seen": 1522204672 + }, + { + "epoch": 0.46, + "learning_rate": 0.00043540220661985956, + "loss": 1.4309, + "theoretical_loss": 3.539988169804026, + "tokens_seen": 1522270208 + }, + { + "epoch": 0.46, + "learning_rate": 0.0004353861584754263, + "loss": 1.4821, + "theoretical_loss": 3.5399748786937444, + "tokens_seen": 1522335744 + }, + { + "epoch": 0.46, + "learning_rate": 0.00043537011033099296, + "loss": 1.4246, + "theoretical_loss": 3.5399615883158297, + "tokens_seen": 1522401280 + }, + { + "epoch": 0.46, + "learning_rate": 0.0004353540621865597, + "loss": 1.3986, + "theoretical_loss": 3.53994829867021, + "tokens_seen": 1522466816 + }, + { + "epoch": 0.46, + "learning_rate": 0.00043533801404212636, + "loss": 1.3558, + "theoretical_loss": 3.5399350097568143, + "tokens_seen": 1522532352 + }, + { + "epoch": 0.46, + "learning_rate": 0.0004353219658976931, + "loss": 1.4825, + "theoretical_loss": 3.5399217215755696, + "tokens_seen": 1522597888 + }, + { + "epoch": 0.46, + "learning_rate": 0.00043530591775325976, + "loss": 1.2527, + "theoretical_loss": 3.5399084341264038, + "tokens_seen": 1522663424 + }, + { + "epoch": 0.46, + "learning_rate": 0.0004352898696088265, + "loss": 1.193, + "theoretical_loss": 3.539895147409246, + "tokens_seen": 1522728960 + }, + { + "epoch": 0.46, + "learning_rate": 0.00043527382146439316, + "loss": 1.45, + "theoretical_loss": 3.539881861424025, + "tokens_seen": 1522794496 + }, + { + "epoch": 0.46, + "learning_rate": 0.0004352577733199599, + "loss": 1.3689, + "theoretical_loss": 3.539868576170667, + "tokens_seen": 1522860032 + }, + { + "epoch": 0.46, + "learning_rate": 0.00043524172517552656, + "loss": 1.3324, + "theoretical_loss": 3.5398552916491015, + "tokens_seen": 1522925568 + }, + { + "epoch": 0.46, + "learning_rate": 0.0004352256770310933, + "loss": 1.3866, + "theoretical_loss": 3.539842007859256, + "tokens_seen": 1522991104 + }, + { + "epoch": 0.46, + "learning_rate": 0.00043520962888665997, + "loss": 1.3576, + "theoretical_loss": 3.5398287248010596, + "tokens_seen": 1523056640 + }, + { + "epoch": 0.46, + "learning_rate": 0.0004351935807422267, + "loss": 1.2405, + "theoretical_loss": 3.53981544247444, + "tokens_seen": 1523122176 + }, + { + "epoch": 0.46, + "learning_rate": 0.00043517753259779337, + "loss": 1.3583, + "theoretical_loss": 3.5398021608793258, + "tokens_seen": 1523187712 + }, + { + "epoch": 0.46, + "learning_rate": 0.0004351614844533601, + "loss": 1.3271, + "theoretical_loss": 3.539788880015644, + "tokens_seen": 1523253248 + }, + { + "epoch": 0.46, + "learning_rate": 0.00043514543630892677, + "loss": 1.3703, + "theoretical_loss": 3.539775599883325, + "tokens_seen": 1523318784 + }, + { + "epoch": 0.46, + "learning_rate": 0.0004351293881644935, + "loss": 1.3244, + "theoretical_loss": 3.5397623204822946, + "tokens_seen": 1523384320 + }, + { + "epoch": 0.46, + "learning_rate": 0.00043511334002006017, + "loss": 1.251, + "theoretical_loss": 3.5397490418124833, + "tokens_seen": 1523449856 + }, + { + "epoch": 0.46, + "learning_rate": 0.0004350972918756269, + "loss": 1.4895, + "theoretical_loss": 3.539735763873818, + "tokens_seen": 1523515392 + }, + { + "epoch": 0.46, + "learning_rate": 0.00043508124373119357, + "loss": 1.4782, + "theoretical_loss": 3.5397224866662276, + "tokens_seen": 1523580928 + }, + { + "epoch": 0.46, + "learning_rate": 0.0004350651955867603, + "loss": 1.4219, + "theoretical_loss": 3.53970921018964, + "tokens_seen": 1523646464 + }, + { + "epoch": 0.46, + "objective/train/docs_used": 556367, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.4134796857833862, + "objective/train/theoretical_loss": 3.5396959344439836, + "objective/train/tokens_used": 1544172000, + "theoretical_loss": 3.5396959344439836, + "tokens_seen": 1523712000 + }, + { + "epoch": 0.46, + "learning_rate": 0.00043504914744232697, + "loss": 1.4857, + "theoretical_loss": 3.5396959344439836, + "tokens_seen": 1523712000 + }, + { + "epoch": 0.46, + "learning_rate": 0.0004350330992978937, + "loss": 1.4915, + "theoretical_loss": 3.5396826594291873, + "tokens_seen": 1523777536 + }, + { + "epoch": 0.46, + "learning_rate": 0.0004350170511534604, + "loss": 1.3575, + "theoretical_loss": 3.539669385145179, + "tokens_seen": 1523843072 + }, + { + "epoch": 0.46, + "learning_rate": 0.00043500100300902713, + "loss": 1.3484, + "theoretical_loss": 3.5396561115918863, + "tokens_seen": 1523908608 + }, + { + "epoch": 0.46, + "learning_rate": 0.0004349849548645938, + "loss": 1.4383, + "theoretical_loss": 3.5396428387692387, + "tokens_seen": 1523974144 + }, + { + "epoch": 0.46, + "learning_rate": 0.00043496890672016053, + "loss": 1.373, + "theoretical_loss": 3.5396295666771644, + "tokens_seen": 1524039680 + }, + { + "epoch": 0.46, + "learning_rate": 0.0004349528585757272, + "loss": 1.3505, + "theoretical_loss": 3.539616295315591, + "tokens_seen": 1524105216 + }, + { + "epoch": 0.46, + "learning_rate": 0.00043493681043129393, + "loss": 1.3856, + "theoretical_loss": 3.539603024684448, + "tokens_seen": 1524170752 + }, + { + "epoch": 0.46, + "learning_rate": 0.0004349207622868606, + "loss": 1.439, + "theoretical_loss": 3.5395897547836626, + "tokens_seen": 1524236288 + }, + { + "epoch": 0.46, + "learning_rate": 0.00043490471414242733, + "loss": 1.2746, + "theoretical_loss": 3.539576485613164, + "tokens_seen": 1524301824 + }, + { + "epoch": 0.46, + "learning_rate": 0.000434888665997994, + "loss": 1.3519, + "theoretical_loss": 3.539563217172881, + "tokens_seen": 1524367360 + }, + { + "epoch": 0.46, + "learning_rate": 0.0004348726178535607, + "loss": 1.2859, + "theoretical_loss": 3.5395499494627405, + "tokens_seen": 1524432896 + }, + { + "epoch": 0.46, + "learning_rate": 0.0004348565697091274, + "loss": 1.41, + "theoretical_loss": 3.539536682482672, + "tokens_seen": 1524498432 + }, + { + "epoch": 0.46, + "learning_rate": 0.0004348405215646941, + "loss": 1.4011, + "theoretical_loss": 3.539523416232605, + "tokens_seen": 1524563968 + }, + { + "epoch": 0.46, + "learning_rate": 0.00043482447342026083, + "loss": 1.3643, + "theoretical_loss": 3.5395101507124656, + "tokens_seen": 1524629504 + }, + { + "epoch": 0.46, + "learning_rate": 0.0004348084252758275, + "loss": 1.3712, + "theoretical_loss": 3.5394968859221834, + "tokens_seen": 1524695040 + }, + { + "epoch": 0.46, + "learning_rate": 0.00043479237713139424, + "loss": 1.5015, + "theoretical_loss": 3.539483621861687, + "tokens_seen": 1524760576 + }, + { + "epoch": 0.46, + "learning_rate": 0.0004347763289869609, + "loss": 1.409, + "theoretical_loss": 3.539470358530905, + "tokens_seen": 1524826112 + }, + { + "epoch": 0.46, + "learning_rate": 0.00043476028084252764, + "loss": 1.5051, + "theoretical_loss": 3.539457095929766, + "tokens_seen": 1524891648 + }, + { + "epoch": 0.46, + "learning_rate": 0.0004347442326980943, + "loss": 1.3811, + "theoretical_loss": 3.539443834058198, + "tokens_seen": 1524957184 + }, + { + "epoch": 0.46, + "learning_rate": 0.00043472818455366104, + "loss": 1.3484, + "theoretical_loss": 3.539430572916129, + "tokens_seen": 1525022720 + }, + { + "epoch": 0.46, + "learning_rate": 0.0004347121364092277, + "loss": 1.3264, + "theoretical_loss": 3.539417312503489, + "tokens_seen": 1525088256 + }, + { + "epoch": 0.46, + "learning_rate": 0.00043469608826479444, + "loss": 1.2636, + "theoretical_loss": 3.5394040528202058, + "tokens_seen": 1525153792 + }, + { + "epoch": 0.46, + "learning_rate": 0.0004346800401203611, + "loss": 1.2646, + "theoretical_loss": 3.5393907938662075, + "tokens_seen": 1525219328 + }, + { + "epoch": 0.46, + "learning_rate": 0.00043466399197592784, + "loss": 1.3706, + "theoretical_loss": 3.5393775356414237, + "tokens_seen": 1525284864 + }, + { + "epoch": 0.46, + "learning_rate": 0.0004346479438314945, + "loss": 1.4837, + "theoretical_loss": 3.5393642781457815, + "tokens_seen": 1525350400 + }, + { + "epoch": 0.46, + "learning_rate": 0.00043463189568706124, + "loss": 1.2853, + "theoretical_loss": 3.5393510213792108, + "tokens_seen": 1525415936 + }, + { + "epoch": 0.46, + "learning_rate": 0.0004346158475426279, + "loss": 1.4212, + "theoretical_loss": 3.5393377653416396, + "tokens_seen": 1525481472 + }, + { + "epoch": 0.46, + "learning_rate": 0.00043459979939819464, + "loss": 1.4235, + "theoretical_loss": 3.539324510032997, + "tokens_seen": 1525547008 + }, + { + "epoch": 0.46, + "learning_rate": 0.0004345837512537613, + "loss": 1.3298, + "theoretical_loss": 3.5393112554532102, + "tokens_seen": 1525612544 + }, + { + "epoch": 0.46, + "learning_rate": 0.00043456770310932804, + "loss": 1.3383, + "theoretical_loss": 3.5392980016022095, + "tokens_seen": 1525678080 + }, + { + "epoch": 0.46, + "learning_rate": 0.0004345516549648947, + "loss": 1.2908, + "theoretical_loss": 3.539284748479923, + "tokens_seen": 1525743616 + }, + { + "epoch": 0.46, + "learning_rate": 0.00043453560682046145, + "loss": 1.4338, + "theoretical_loss": 3.539271496086279, + "tokens_seen": 1525809152 + }, + { + "epoch": 0.46, + "learning_rate": 0.0004345195586760281, + "loss": 1.3484, + "theoretical_loss": 3.539258244421206, + "tokens_seen": 1525874688 + }, + { + "epoch": 0.46, + "learning_rate": 0.00043450351053159485, + "loss": 1.2667, + "theoretical_loss": 3.5392449934846333, + "tokens_seen": 1525940224 + }, + { + "epoch": 0.46, + "learning_rate": 0.0004344874623871615, + "loss": 1.4112, + "theoretical_loss": 3.5392317432764893, + "tokens_seen": 1526005760 + }, + { + "epoch": 0.46, + "learning_rate": 0.00043447141424272825, + "loss": 1.4078, + "theoretical_loss": 3.539218493796702, + "tokens_seen": 1526071296 + }, + { + "epoch": 0.46, + "learning_rate": 0.0004344553660982949, + "loss": 1.4409, + "theoretical_loss": 3.539205245045202, + "tokens_seen": 1526136832 + }, + { + "epoch": 0.46, + "learning_rate": 0.00043443931795386165, + "loss": 1.3754, + "theoretical_loss": 3.5391919970219154, + "tokens_seen": 1526202368 + }, + { + "epoch": 0.46, + "learning_rate": 0.0004344232698094283, + "loss": 1.4155, + "theoretical_loss": 3.539178749726773, + "tokens_seen": 1526267904 + }, + { + "epoch": 0.46, + "learning_rate": 0.00043440722166499505, + "loss": 1.3726, + "theoretical_loss": 3.5391655031597025, + "tokens_seen": 1526333440 + }, + { + "epoch": 0.46, + "learning_rate": 0.0004343911735205617, + "loss": 1.4259, + "theoretical_loss": 3.539152257320633, + "tokens_seen": 1526398976 + }, + { + "epoch": 0.46, + "learning_rate": 0.00043437512537612845, + "loss": 1.2302, + "theoretical_loss": 3.5391390122094926, + "tokens_seen": 1526464512 + }, + { + "epoch": 0.46, + "learning_rate": 0.0004343590772316951, + "loss": 1.348, + "theoretical_loss": 3.539125767826211, + "tokens_seen": 1526530048 + }, + { + "epoch": 0.46, + "learning_rate": 0.00043434302908726185, + "loss": 1.4621, + "theoretical_loss": 3.5391125241707164, + "tokens_seen": 1526595584 + }, + { + "epoch": 0.46, + "learning_rate": 0.0004343269809428285, + "loss": 1.3314, + "theoretical_loss": 3.5390992812429376, + "tokens_seen": 1526661120 + }, + { + "epoch": 0.46, + "learning_rate": 0.00043431093279839526, + "loss": 1.2925, + "theoretical_loss": 3.539086039042804, + "tokens_seen": 1526726656 + }, + { + "epoch": 0.46, + "learning_rate": 0.0004342948846539619, + "loss": 1.4236, + "theoretical_loss": 3.5390727975702427, + "tokens_seen": 1526792192 + }, + { + "epoch": 0.46, + "learning_rate": 0.00043427883650952866, + "loss": 1.4234, + "theoretical_loss": 3.539059556825184, + "tokens_seen": 1526857728 + }, + { + "epoch": 0.46, + "learning_rate": 0.0004342627883650953, + "loss": 1.3779, + "theoretical_loss": 3.539046316807557, + "tokens_seen": 1526923264 + }, + { + "epoch": 0.46, + "objective/train/docs_used": 557485, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.3049226999282837, + "objective/train/theoretical_loss": 3.5390330775172893, + "objective/train/tokens_used": 1547448800, + "theoretical_loss": 3.5390330775172893, + "tokens_seen": 1526988800 + }, + { + "epoch": 0.46, + "learning_rate": 0.000434246740220662, + "loss": 1.3618, + "theoretical_loss": 3.5390330775172893, + "tokens_seen": 1526988800 + }, + { + "epoch": 0.46, + "learning_rate": 0.0004342306920762287, + "loss": 1.3691, + "theoretical_loss": 3.53901983895431, + "tokens_seen": 1527054336 + }, + { + "epoch": 0.46, + "learning_rate": 0.0004342146439317954, + "loss": 1.3252, + "theoretical_loss": 3.5390066011185493, + "tokens_seen": 1527119872 + }, + { + "epoch": 0.46, + "learning_rate": 0.0004341985957873621, + "loss": 1.2457, + "theoretical_loss": 3.538993364009934, + "tokens_seen": 1527185408 + }, + { + "epoch": 0.46, + "learning_rate": 0.0004341825476429288, + "loss": 1.3152, + "theoretical_loss": 3.538980127628394, + "tokens_seen": 1527250944 + }, + { + "epoch": 0.46, + "learning_rate": 0.0004341664994984955, + "loss": 1.339, + "theoretical_loss": 3.5389668919738586, + "tokens_seen": 1527316480 + }, + { + "epoch": 0.46, + "learning_rate": 0.0004341504513540622, + "loss": 1.3853, + "theoretical_loss": 3.5389536570462554, + "tokens_seen": 1527382016 + }, + { + "epoch": 0.46, + "learning_rate": 0.0004341344032096289, + "loss": 1.3573, + "theoretical_loss": 3.5389404228455144, + "tokens_seen": 1527447552 + }, + { + "epoch": 0.46, + "learning_rate": 0.0004341183550651956, + "loss": 1.3819, + "theoretical_loss": 3.538927189371564, + "tokens_seen": 1527513088 + }, + { + "epoch": 0.46, + "learning_rate": 0.00043410230692076225, + "loss": 1.3481, + "theoretical_loss": 3.5389139566243335, + "tokens_seen": 1527578624 + }, + { + "epoch": 0.46, + "learning_rate": 0.000434086258776329, + "loss": 1.4113, + "theoretical_loss": 3.538900724603752, + "tokens_seen": 1527644160 + }, + { + "epoch": 0.46, + "learning_rate": 0.00043407021063189566, + "loss": 1.4215, + "theoretical_loss": 3.538887493309747, + "tokens_seen": 1527709696 + }, + { + "epoch": 0.46, + "learning_rate": 0.0004340541624874624, + "loss": 1.3086, + "theoretical_loss": 3.538874262742249, + "tokens_seen": 1527775232 + }, + { + "epoch": 0.46, + "learning_rate": 0.00043403811434302906, + "loss": 1.4892, + "theoretical_loss": 3.5388610329011865, + "tokens_seen": 1527840768 + }, + { + "epoch": 0.46, + "learning_rate": 0.0004340220661985958, + "loss": 1.2896, + "theoretical_loss": 3.5388478037864877, + "tokens_seen": 1527906304 + }, + { + "epoch": 0.46, + "learning_rate": 0.00043400601805416246, + "loss": 1.3305, + "theoretical_loss": 3.5388345753980825, + "tokens_seen": 1527971840 + }, + { + "epoch": 0.46, + "learning_rate": 0.0004339899699097292, + "loss": 1.3556, + "theoretical_loss": 3.5388213477359, + "tokens_seen": 1528037376 + }, + { + "epoch": 0.46, + "learning_rate": 0.00043397392176529586, + "loss": 1.3531, + "theoretical_loss": 3.5388081207998683, + "tokens_seen": 1528102912 + }, + { + "epoch": 0.46, + "learning_rate": 0.0004339578736208626, + "loss": 1.2949, + "theoretical_loss": 3.538794894589917, + "tokens_seen": 1528168448 + }, + { + "epoch": 0.46, + "learning_rate": 0.00043394182547642926, + "loss": 1.4735, + "theoretical_loss": 3.5387816691059752, + "tokens_seen": 1528233984 + }, + { + "epoch": 0.46, + "learning_rate": 0.000433925777331996, + "loss": 1.2665, + "theoretical_loss": 3.5387684443479714, + "tokens_seen": 1528299520 + }, + { + "epoch": 0.46, + "learning_rate": 0.00043390972918756266, + "loss": 1.4204, + "theoretical_loss": 3.538755220315835, + "tokens_seen": 1528365056 + }, + { + "epoch": 0.46, + "learning_rate": 0.0004338936810431294, + "loss": 1.4615, + "theoretical_loss": 3.538741997009495, + "tokens_seen": 1528430592 + }, + { + "epoch": 0.46, + "learning_rate": 0.00043387763289869606, + "loss": 1.3896, + "theoretical_loss": 3.53872877442888, + "tokens_seen": 1528496128 + }, + { + "epoch": 0.46, + "learning_rate": 0.0004338615847542628, + "loss": 1.3703, + "theoretical_loss": 3.53871555257392, + "tokens_seen": 1528561664 + }, + { + "epoch": 0.46, + "learning_rate": 0.00043384553660982947, + "loss": 1.3819, + "theoretical_loss": 3.5387023314445436, + "tokens_seen": 1528627200 + }, + { + "epoch": 0.46, + "learning_rate": 0.0004338294884653962, + "loss": 1.4222, + "theoretical_loss": 3.538689111040679, + "tokens_seen": 1528692736 + }, + { + "epoch": 0.46, + "learning_rate": 0.00043381344032096287, + "loss": 1.38, + "theoretical_loss": 3.5386758913622565, + "tokens_seen": 1528758272 + }, + { + "epoch": 0.46, + "learning_rate": 0.0004337973921765296, + "loss": 1.3142, + "theoretical_loss": 3.538662672409205, + "tokens_seen": 1528823808 + }, + { + "epoch": 0.46, + "learning_rate": 0.00043378134403209627, + "loss": 1.2502, + "theoretical_loss": 3.5386494541814533, + "tokens_seen": 1528889344 + }, + { + "epoch": 0.46, + "learning_rate": 0.000433765295887663, + "loss": 1.3989, + "theoretical_loss": 3.5386362366789306, + "tokens_seen": 1528954880 + }, + { + "epoch": 0.46, + "learning_rate": 0.00043374924774322967, + "loss": 1.5682, + "theoretical_loss": 3.538623019901566, + "tokens_seen": 1529020416 + }, + { + "epoch": 0.46, + "learning_rate": 0.0004337331995987964, + "loss": 1.4212, + "theoretical_loss": 3.5386098038492886, + "tokens_seen": 1529085952 + }, + { + "epoch": 0.46, + "learning_rate": 0.00043371715145436307, + "loss": 1.3688, + "theoretical_loss": 3.5385965885220276, + "tokens_seen": 1529151488 + }, + { + "epoch": 0.46, + "learning_rate": 0.0004337011033099298, + "loss": 1.3203, + "theoretical_loss": 3.538583373919712, + "tokens_seen": 1529217024 + }, + { + "epoch": 0.46, + "learning_rate": 0.00043368505516549647, + "loss": 1.3886, + "theoretical_loss": 3.5385701600422714, + "tokens_seen": 1529282560 + }, + { + "epoch": 0.46, + "learning_rate": 0.0004336690070210632, + "loss": 1.4787, + "theoretical_loss": 3.538556946889635, + "tokens_seen": 1529348096 + }, + { + "epoch": 0.46, + "learning_rate": 0.00043365295887662987, + "loss": 1.4097, + "theoretical_loss": 3.538543734461731, + "tokens_seen": 1529413632 + }, + { + "epoch": 0.46, + "learning_rate": 0.00043363691073219663, + "loss": 1.437, + "theoretical_loss": 3.5385305227584896, + "tokens_seen": 1529479168 + }, + { + "epoch": 0.46, + "learning_rate": 0.00043362086258776333, + "loss": 1.313, + "theoretical_loss": 3.53851731177984, + "tokens_seen": 1529544704 + }, + { + "epoch": 0.46, + "learning_rate": 0.00043360481444333003, + "loss": 1.3381, + "theoretical_loss": 3.538504101525711, + "tokens_seen": 1529610240 + }, + { + "epoch": 0.46, + "learning_rate": 0.00043358876629889673, + "loss": 1.3306, + "theoretical_loss": 3.5384908919960316, + "tokens_seen": 1529675776 + }, + { + "epoch": 0.46, + "learning_rate": 0.00043357271815446343, + "loss": 1.3561, + "theoretical_loss": 3.538477683190732, + "tokens_seen": 1529741312 + }, + { + "epoch": 0.46, + "learning_rate": 0.00043355667001003013, + "loss": 1.4695, + "theoretical_loss": 3.5384644751097403, + "tokens_seen": 1529806848 + }, + { + "epoch": 0.46, + "learning_rate": 0.00043354062186559683, + "loss": 1.2441, + "theoretical_loss": 3.5384512677529862, + "tokens_seen": 1529872384 + }, + { + "epoch": 0.46, + "learning_rate": 0.00043352457372116353, + "loss": 1.4571, + "theoretical_loss": 3.5384380611203987, + "tokens_seen": 1529937920 + }, + { + "epoch": 0.46, + "learning_rate": 0.00043350852557673023, + "loss": 1.4381, + "theoretical_loss": 3.5384248552119084, + "tokens_seen": 1530003456 + }, + { + "epoch": 0.46, + "learning_rate": 0.00043349247743229693, + "loss": 1.4201, + "theoretical_loss": 3.538411650027443, + "tokens_seen": 1530068992 + }, + { + "epoch": 0.46, + "learning_rate": 0.0004334764292878636, + "loss": 1.3579, + "theoretical_loss": 3.5383984455669326, + "tokens_seen": 1530134528 + }, + { + "epoch": 0.46, + "learning_rate": 0.00043346038114343033, + "loss": 1.3623, + "theoretical_loss": 3.538385241830306, + "tokens_seen": 1530200064 + }, + { + "epoch": 0.46, + "objective/train/docs_used": 558616, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.288569688796997, + "objective/train/theoretical_loss": 3.5383720388174935, + "objective/train/tokens_used": 1550725600, + "theoretical_loss": 3.5383720388174935, + "tokens_seen": 1530265600 + }, + { + "epoch": 0.46, + "learning_rate": 0.000433444332998997, + "loss": 1.4002, + "theoretical_loss": 3.5383720388174935, + "tokens_seen": 1530265600 + }, + { + "epoch": 0.46, + "learning_rate": 0.00043342828485456374, + "loss": 1.4226, + "theoretical_loss": 3.538358836528423, + "tokens_seen": 1530331136 + }, + { + "epoch": 0.46, + "learning_rate": 0.0004334122367101304, + "loss": 1.4407, + "theoretical_loss": 3.538345634963025, + "tokens_seen": 1530396672 + }, + { + "epoch": 0.46, + "learning_rate": 0.00043339618856569714, + "loss": 1.3023, + "theoretical_loss": 3.5383324341212283, + "tokens_seen": 1530462208 + }, + { + "epoch": 0.46, + "learning_rate": 0.0004333801404212638, + "loss": 1.4237, + "theoretical_loss": 3.5383192340029623, + "tokens_seen": 1530527744 + }, + { + "epoch": 0.46, + "learning_rate": 0.00043336409227683054, + "loss": 1.4554, + "theoretical_loss": 3.5383060346081567, + "tokens_seen": 1530593280 + }, + { + "epoch": 0.46, + "learning_rate": 0.0004333480441323972, + "loss": 1.4152, + "theoretical_loss": 3.5382928359367405, + "tokens_seen": 1530658816 + }, + { + "epoch": 0.46, + "learning_rate": 0.00043333199598796394, + "loss": 1.4185, + "theoretical_loss": 3.5382796379886434, + "tokens_seen": 1530724352 + }, + { + "epoch": 0.46, + "learning_rate": 0.0004333159478435306, + "loss": 1.2529, + "theoretical_loss": 3.5382664407637945, + "tokens_seen": 1530789888 + }, + { + "epoch": 0.46, + "learning_rate": 0.00043329989969909734, + "loss": 1.4228, + "theoretical_loss": 3.538253244262123, + "tokens_seen": 1530855424 + }, + { + "epoch": 0.46, + "learning_rate": 0.000433283851554664, + "loss": 1.4069, + "theoretical_loss": 3.538240048483559, + "tokens_seen": 1530920960 + }, + { + "epoch": 0.46, + "learning_rate": 0.00043326780341023074, + "loss": 1.3811, + "theoretical_loss": 3.5382268534280317, + "tokens_seen": 1530986496 + }, + { + "epoch": 0.46, + "learning_rate": 0.0004332517552657974, + "loss": 1.3045, + "theoretical_loss": 3.53821365909547, + "tokens_seen": 1531052032 + }, + { + "epoch": 0.46, + "learning_rate": 0.00043323570712136414, + "loss": 1.3327, + "theoretical_loss": 3.5382004654858035, + "tokens_seen": 1531117568 + }, + { + "epoch": 0.46, + "learning_rate": 0.0004332196589769308, + "loss": 1.3752, + "theoretical_loss": 3.5381872725989627, + "tokens_seen": 1531183104 + }, + { + "epoch": 0.46, + "learning_rate": 0.00043320361083249754, + "loss": 1.3028, + "theoretical_loss": 3.5381740804348754, + "tokens_seen": 1531248640 + }, + { + "epoch": 0.46, + "learning_rate": 0.0004331875626880642, + "loss": 1.4535, + "theoretical_loss": 3.538160888993472, + "tokens_seen": 1531314176 + }, + { + "epoch": 0.46, + "learning_rate": 0.00043317151454363095, + "loss": 1.4607, + "theoretical_loss": 3.538147698274683, + "tokens_seen": 1531379712 + }, + { + "epoch": 0.46, + "learning_rate": 0.0004331554663991976, + "loss": 1.411, + "theoretical_loss": 3.5381345082784357, + "tokens_seen": 1531445248 + }, + { + "epoch": 0.46, + "learning_rate": 0.00043313941825476435, + "loss": 1.2874, + "theoretical_loss": 3.538121319004661, + "tokens_seen": 1531510784 + }, + { + "epoch": 0.46, + "learning_rate": 0.000433123370110331, + "loss": 1.4115, + "theoretical_loss": 3.538108130453288, + "tokens_seen": 1531576320 + }, + { + "epoch": 0.46, + "learning_rate": 0.00043310732196589775, + "loss": 1.5365, + "theoretical_loss": 3.5380949426242463, + "tokens_seen": 1531641856 + }, + { + "epoch": 0.46, + "learning_rate": 0.0004330912738214644, + "loss": 1.3283, + "theoretical_loss": 3.538081755517466, + "tokens_seen": 1531707392 + }, + { + "epoch": 0.46, + "learning_rate": 0.00043307522567703115, + "loss": 1.4039, + "theoretical_loss": 3.5380685691328755, + "tokens_seen": 1531772928 + }, + { + "epoch": 0.46, + "learning_rate": 0.0004330591775325978, + "loss": 1.3945, + "theoretical_loss": 3.5380553834704056, + "tokens_seen": 1531838464 + }, + { + "epoch": 0.46, + "learning_rate": 0.00043304312938816455, + "loss": 1.4171, + "theoretical_loss": 3.5380421985299844, + "tokens_seen": 1531904000 + }, + { + "epoch": 0.46, + "learning_rate": 0.0004330270812437312, + "loss": 1.414, + "theoretical_loss": 3.538029014311543, + "tokens_seen": 1531969536 + }, + { + "epoch": 0.46, + "learning_rate": 0.00043301103309929795, + "loss": 1.3307, + "theoretical_loss": 3.5380158308150103, + "tokens_seen": 1532035072 + }, + { + "epoch": 0.46, + "learning_rate": 0.0004329949849548646, + "loss": 1.442, + "theoretical_loss": 3.5380026480403153, + "tokens_seen": 1532100608 + }, + { + "epoch": 0.46, + "learning_rate": 0.00043297893681043135, + "loss": 1.3957, + "theoretical_loss": 3.5379894659873887, + "tokens_seen": 1532166144 + }, + { + "epoch": 0.46, + "learning_rate": 0.000432962888665998, + "loss": 1.2958, + "theoretical_loss": 3.5379762846561595, + "tokens_seen": 1532231680 + }, + { + "epoch": 0.46, + "learning_rate": 0.00043294684052156475, + "loss": 1.2867, + "theoretical_loss": 3.5379631040465576, + "tokens_seen": 1532297216 + }, + { + "epoch": 0.46, + "learning_rate": 0.0004329307923771314, + "loss": 1.3132, + "theoretical_loss": 3.5379499241585117, + "tokens_seen": 1532362752 + }, + { + "epoch": 0.46, + "learning_rate": 0.00043291474423269816, + "loss": 1.3878, + "theoretical_loss": 3.5379367449919528, + "tokens_seen": 1532428288 + }, + { + "epoch": 0.46, + "learning_rate": 0.0004328986960882648, + "loss": 1.2996, + "theoretical_loss": 3.53792356654681, + "tokens_seen": 1532493824 + }, + { + "epoch": 0.46, + "learning_rate": 0.00043288264794383156, + "loss": 1.423, + "theoretical_loss": 3.537910388823013, + "tokens_seen": 1532559360 + }, + { + "epoch": 0.46, + "learning_rate": 0.0004328665997993982, + "loss": 1.3461, + "theoretical_loss": 3.537897211820491, + "tokens_seen": 1532624896 + }, + { + "epoch": 0.46, + "learning_rate": 0.0004328505516549649, + "loss": 1.385, + "theoretical_loss": 3.5378840355391743, + "tokens_seen": 1532690432 + }, + { + "epoch": 0.46, + "learning_rate": 0.0004328345035105316, + "loss": 1.3352, + "theoretical_loss": 3.537870859978992, + "tokens_seen": 1532755968 + }, + { + "epoch": 0.46, + "learning_rate": 0.0004328184553660983, + "loss": 1.4005, + "theoretical_loss": 3.5378576851398744, + "tokens_seen": 1532821504 + }, + { + "epoch": 0.46, + "learning_rate": 0.000432802407221665, + "loss": 1.6197, + "theoretical_loss": 3.537844511021751, + "tokens_seen": 1532887040 + }, + { + "epoch": 0.46, + "learning_rate": 0.0004327863590772317, + "loss": 1.3929, + "theoretical_loss": 3.5378313376245516, + "tokens_seen": 1532952576 + }, + { + "epoch": 0.46, + "learning_rate": 0.0004327703109327984, + "loss": 1.3141, + "theoretical_loss": 3.537818164948206, + "tokens_seen": 1533018112 + }, + { + "epoch": 0.46, + "learning_rate": 0.0004327542627883651, + "loss": 1.3537, + "theoretical_loss": 3.537804992992643, + "tokens_seen": 1533083648 + }, + { + "epoch": 0.46, + "learning_rate": 0.0004327382146439318, + "loss": 1.3591, + "theoretical_loss": 3.5377918217577937, + "tokens_seen": 1533149184 + }, + { + "epoch": 0.46, + "learning_rate": 0.0004327221664994985, + "loss": 1.4265, + "theoretical_loss": 3.5377786512435874, + "tokens_seen": 1533214720 + }, + { + "epoch": 0.46, + "learning_rate": 0.00043270611835506516, + "loss": 1.4605, + "theoretical_loss": 3.537765481449954, + "tokens_seen": 1533280256 + }, + { + "epoch": 0.46, + "learning_rate": 0.0004326900702106319, + "loss": 1.3629, + "theoretical_loss": 3.5377523123768224, + "tokens_seen": 1533345792 + }, + { + "epoch": 0.46, + "learning_rate": 0.00043267402206619856, + "loss": 1.3709, + "theoretical_loss": 3.5377391440241226, + "tokens_seen": 1533411328 + }, + { + "epoch": 0.46, + "learning_rate": 0.0004326579739217653, + "loss": 1.5017, + "theoretical_loss": 3.5377259763917857, + "tokens_seen": 1533476864 + }, + { + "epoch": 0.46, + "objective/train/docs_used": 559697, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.4281322956085205, + "objective/train/theoretical_loss": 3.53771280947974, + "objective/train/tokens_used": 1554002400, + "theoretical_loss": 3.53771280947974, + "tokens_seen": 1533542400 + }, + { + "epoch": 0.46, + "learning_rate": 0.00043264192577733196, + "loss": 1.5322, + "theoretical_loss": 3.53771280947974, + "tokens_seen": 1533542400 + }, + { + "epoch": 0.46, + "learning_rate": 0.0004326258776328987, + "loss": 1.4645, + "theoretical_loss": 3.5376996432879166, + "tokens_seen": 1533607936 + }, + { + "epoch": 0.46, + "learning_rate": 0.00043260982948846536, + "loss": 1.3422, + "theoretical_loss": 3.5376864778162442, + "tokens_seen": 1533673472 + }, + { + "epoch": 0.46, + "learning_rate": 0.0004325937813440321, + "loss": 1.3582, + "theoretical_loss": 3.537673313064653, + "tokens_seen": 1533739008 + }, + { + "epoch": 0.46, + "learning_rate": 0.00043257773319959876, + "loss": 1.385, + "theoretical_loss": 3.5376601490330737, + "tokens_seen": 1533804544 + }, + { + "epoch": 0.46, + "learning_rate": 0.0004325616850551655, + "loss": 1.2877, + "theoretical_loss": 3.5376469857214348, + "tokens_seen": 1533870080 + }, + { + "epoch": 0.46, + "learning_rate": 0.00043254563691073216, + "loss": 1.3925, + "theoretical_loss": 3.537633823129667, + "tokens_seen": 1533935616 + }, + { + "epoch": 0.46, + "learning_rate": 0.0004325295887662989, + "loss": 1.448, + "theoretical_loss": 3.5376206612576997, + "tokens_seen": 1534001152 + }, + { + "epoch": 0.46, + "learning_rate": 0.00043251354062186556, + "loss": 1.4552, + "theoretical_loss": 3.537607500105463, + "tokens_seen": 1534066688 + }, + { + "epoch": 0.46, + "learning_rate": 0.0004324974924774323, + "loss": 1.4057, + "theoretical_loss": 3.5375943396728875, + "tokens_seen": 1534132224 + }, + { + "epoch": 0.46, + "learning_rate": 0.00043248144433299896, + "loss": 1.4214, + "theoretical_loss": 3.5375811799599015, + "tokens_seen": 1534197760 + }, + { + "epoch": 0.46, + "learning_rate": 0.0004324653961885657, + "loss": 1.448, + "theoretical_loss": 3.537568020966437, + "tokens_seen": 1534263296 + }, + { + "epoch": 0.46, + "learning_rate": 0.00043244934804413237, + "loss": 1.3751, + "theoretical_loss": 3.537554862692422, + "tokens_seen": 1534328832 + }, + { + "epoch": 0.46, + "learning_rate": 0.0004324332998996991, + "loss": 1.5788, + "theoretical_loss": 3.5375417051377873, + "tokens_seen": 1534394368 + }, + { + "epoch": 0.46, + "learning_rate": 0.0004324172517552659, + "loss": 1.4306, + "theoretical_loss": 3.537528548302463, + "tokens_seen": 1534459904 + }, + { + "epoch": 0.47, + "learning_rate": 0.0004324012036108325, + "loss": 1.4416, + "theoretical_loss": 3.5375153921863793, + "tokens_seen": 1534525440 + }, + { + "epoch": 0.47, + "learning_rate": 0.0004323851554663993, + "loss": 1.4108, + "theoretical_loss": 3.5375022367894653, + "tokens_seen": 1534590976 + }, + { + "epoch": 0.47, + "learning_rate": 0.0004323691073219659, + "loss": 1.4115, + "theoretical_loss": 3.537489082111651, + "tokens_seen": 1534656512 + }, + { + "epoch": 0.47, + "learning_rate": 0.0004323530591775327, + "loss": 1.4228, + "theoretical_loss": 3.537475928152867, + "tokens_seen": 1534722048 + }, + { + "epoch": 0.47, + "learning_rate": 0.0004323370110330993, + "loss": 1.469, + "theoretical_loss": 3.5374627749130436, + "tokens_seen": 1534787584 + }, + { + "epoch": 0.47, + "learning_rate": 0.0004323209628886661, + "loss": 1.3851, + "theoretical_loss": 3.53744962239211, + "tokens_seen": 1534853120 + }, + { + "epoch": 0.47, + "learning_rate": 0.0004323049147442327, + "loss": 1.4579, + "theoretical_loss": 3.537436470589996, + "tokens_seen": 1534918656 + }, + { + "epoch": 0.47, + "learning_rate": 0.0004322888665997995, + "loss": 1.343, + "theoretical_loss": 3.5374233195066327, + "tokens_seen": 1534984192 + }, + { + "epoch": 0.47, + "learning_rate": 0.00043227281845536613, + "loss": 1.4171, + "theoretical_loss": 3.5374101691419497, + "tokens_seen": 1535049728 + }, + { + "epoch": 0.47, + "learning_rate": 0.0004322567703109329, + "loss": 1.3769, + "theoretical_loss": 3.5373970194958764, + "tokens_seen": 1535115264 + }, + { + "epoch": 0.47, + "learning_rate": 0.00043224072216649953, + "loss": 1.3862, + "theoretical_loss": 3.5373838705683434, + "tokens_seen": 1535180800 + }, + { + "epoch": 0.47, + "learning_rate": 0.00043222467402206623, + "loss": 1.3394, + "theoretical_loss": 3.537370722359281, + "tokens_seen": 1535246336 + }, + { + "epoch": 0.47, + "learning_rate": 0.00043220862587763293, + "loss": 1.3338, + "theoretical_loss": 3.5373575748686186, + "tokens_seen": 1535311872 + }, + { + "epoch": 0.47, + "learning_rate": 0.00043219257773319963, + "loss": 1.3941, + "theoretical_loss": 3.5373444280962874, + "tokens_seen": 1535377408 + }, + { + "epoch": 0.47, + "learning_rate": 0.00043217652958876633, + "loss": 1.5061, + "theoretical_loss": 3.5373312820422163, + "tokens_seen": 1535442944 + }, + { + "epoch": 0.47, + "learning_rate": 0.00043216048144433303, + "loss": 1.3288, + "theoretical_loss": 3.537318136706336, + "tokens_seen": 1535508480 + }, + { + "epoch": 0.47, + "learning_rate": 0.00043214443329989973, + "loss": 1.416, + "theoretical_loss": 3.5373049920885764, + "tokens_seen": 1535574016 + }, + { + "epoch": 0.47, + "learning_rate": 0.00043212838515546643, + "loss": 1.455, + "theoretical_loss": 3.5372918481888673, + "tokens_seen": 1535639552 + }, + { + "epoch": 0.47, + "learning_rate": 0.00043211233701103313, + "loss": 1.5086, + "theoretical_loss": 3.53727870500714, + "tokens_seen": 1535705088 + }, + { + "epoch": 0.47, + "learning_rate": 0.00043209628886659983, + "loss": 1.4109, + "theoretical_loss": 3.5372655625433236, + "tokens_seen": 1535770624 + }, + { + "epoch": 0.47, + "learning_rate": 0.0004320802407221665, + "loss": 1.3683, + "theoretical_loss": 3.5372524207973486, + "tokens_seen": 1535836160 + }, + { + "epoch": 0.47, + "learning_rate": 0.00043206419257773324, + "loss": 1.4688, + "theoretical_loss": 3.537239279769145, + "tokens_seen": 1535901696 + }, + { + "epoch": 0.47, + "learning_rate": 0.0004320481444332999, + "loss": 1.4158, + "theoretical_loss": 3.5372261394586433, + "tokens_seen": 1535967232 + }, + { + "epoch": 0.47, + "learning_rate": 0.00043203209628886664, + "loss": 1.3386, + "theoretical_loss": 3.5372129998657735, + "tokens_seen": 1536032768 + }, + { + "epoch": 0.47, + "learning_rate": 0.0004320160481444333, + "loss": 1.4344, + "theoretical_loss": 3.537199860990466, + "tokens_seen": 1536098304 + }, + { + "epoch": 0.47, + "learning_rate": 0.00043200000000000004, + "loss": 1.4389, + "theoretical_loss": 3.53718672283265, + "tokens_seen": 1536163840 + }, + { + "epoch": 0.47, + "learning_rate": 0.0004319839518555667, + "loss": 1.5465, + "theoretical_loss": 3.5371735853922566, + "tokens_seen": 1536229376 + }, + { + "epoch": 0.47, + "learning_rate": 0.00043196790371113344, + "loss": 1.3933, + "theoretical_loss": 3.5371604486692165, + "tokens_seen": 1536294912 + }, + { + "epoch": 0.47, + "learning_rate": 0.0004319518555667001, + "loss": 1.5395, + "theoretical_loss": 3.5371473126634587, + "tokens_seen": 1536360448 + }, + { + "epoch": 0.47, + "learning_rate": 0.00043193580742226684, + "loss": 1.4514, + "theoretical_loss": 3.5371341773749143, + "tokens_seen": 1536425984 + }, + { + "epoch": 0.47, + "learning_rate": 0.0004319197592778335, + "loss": 1.4032, + "theoretical_loss": 3.5371210428035136, + "tokens_seen": 1536491520 + }, + { + "epoch": 0.47, + "learning_rate": 0.00043190371113340024, + "loss": 1.3456, + "theoretical_loss": 3.537107908949186, + "tokens_seen": 1536557056 + }, + { + "epoch": 0.47, + "learning_rate": 0.0004318876629889669, + "loss": 1.457, + "theoretical_loss": 3.537094775811863, + "tokens_seen": 1536622592 + }, + { + "epoch": 0.47, + "learning_rate": 0.00043187161484453364, + "loss": 1.3015, + "theoretical_loss": 3.537081643391474, + "tokens_seen": 1536688128 + }, + { + "epoch": 0.47, + "learning_rate": 0.0004318555667001003, + "loss": 1.5231, + "theoretical_loss": 3.5370685116879494, + "tokens_seen": 1536753664 + }, + { + "epoch": 0.47, + "objective/train/docs_used": 561026, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.2669241428375244, + "objective/train/theoretical_loss": 3.5370553807012195, + "objective/train/tokens_used": 1557279200, + "theoretical_loss": 3.5370553807012195, + "tokens_seen": 1536819200 + }, + { + "epoch": 0.47, + "learning_rate": 0.00043183951855566704, + "loss": 1.4138, + "theoretical_loss": 3.5370553807012195, + "tokens_seen": 1536819200 + }, + { + "epoch": 0.47, + "learning_rate": 0.0004318234704112337, + "loss": 1.3447, + "theoretical_loss": 3.537042250431215, + "tokens_seen": 1536884736 + }, + { + "epoch": 0.47, + "learning_rate": 0.00043180742226680045, + "loss": 1.4414, + "theoretical_loss": 3.5370291208778655, + "tokens_seen": 1536950272 + }, + { + "epoch": 0.47, + "learning_rate": 0.0004317913741223671, + "loss": 1.3395, + "theoretical_loss": 3.537015992041103, + "tokens_seen": 1537015808 + }, + { + "epoch": 0.47, + "learning_rate": 0.00043177532597793385, + "loss": 1.364, + "theoretical_loss": 3.537002863920855, + "tokens_seen": 1537081344 + }, + { + "epoch": 0.47, + "learning_rate": 0.0004317592778335005, + "loss": 1.437, + "theoretical_loss": 3.5369897365170546, + "tokens_seen": 1537146880 + }, + { + "epoch": 0.47, + "learning_rate": 0.00043174322968906725, + "loss": 1.406, + "theoretical_loss": 3.5369766098296305, + "tokens_seen": 1537212416 + }, + { + "epoch": 0.47, + "learning_rate": 0.0004317271815446339, + "loss": 1.4646, + "theoretical_loss": 3.536963483858514, + "tokens_seen": 1537277952 + }, + { + "epoch": 0.47, + "learning_rate": 0.00043171113340020065, + "loss": 1.383, + "theoretical_loss": 3.536950358603635, + "tokens_seen": 1537343488 + }, + { + "epoch": 0.47, + "learning_rate": 0.0004316950852557673, + "loss": 1.5424, + "theoretical_loss": 3.5369372340649234, + "tokens_seen": 1537409024 + }, + { + "epoch": 0.47, + "learning_rate": 0.00043167903711133405, + "loss": 1.3797, + "theoretical_loss": 3.5369241102423103, + "tokens_seen": 1537474560 + }, + { + "epoch": 0.47, + "learning_rate": 0.0004316629889669007, + "loss": 1.5043, + "theoretical_loss": 3.5369109871357267, + "tokens_seen": 1537540096 + }, + { + "epoch": 0.47, + "learning_rate": 0.00043164694082246745, + "loss": 1.3478, + "theoretical_loss": 3.536897864745101, + "tokens_seen": 1537605632 + }, + { + "epoch": 0.47, + "learning_rate": 0.0004316308926780341, + "loss": 1.4308, + "theoretical_loss": 3.536884743070366, + "tokens_seen": 1537671168 + }, + { + "epoch": 0.47, + "learning_rate": 0.00043161484453360085, + "loss": 1.3181, + "theoretical_loss": 3.536871622111451, + "tokens_seen": 1537736704 + }, + { + "epoch": 0.47, + "learning_rate": 0.0004315987963891675, + "loss": 1.3156, + "theoretical_loss": 3.536858501868286, + "tokens_seen": 1537802240 + }, + { + "epoch": 0.47, + "learning_rate": 0.00043158274824473425, + "loss": 1.3571, + "theoretical_loss": 3.536845382340802, + "tokens_seen": 1537867776 + }, + { + "epoch": 0.47, + "learning_rate": 0.0004315667001003009, + "loss": 1.4695, + "theoretical_loss": 3.5368322635289298, + "tokens_seen": 1537933312 + }, + { + "epoch": 0.47, + "learning_rate": 0.00043155065195586766, + "loss": 1.4382, + "theoretical_loss": 3.536819145432599, + "tokens_seen": 1537998848 + }, + { + "epoch": 0.47, + "learning_rate": 0.0004315346038114343, + "loss": 1.3466, + "theoretical_loss": 3.536806028051741, + "tokens_seen": 1538064384 + }, + { + "epoch": 0.47, + "learning_rate": 0.00043151855566700106, + "loss": 1.4433, + "theoretical_loss": 3.5367929113862853, + "tokens_seen": 1538129920 + }, + { + "epoch": 0.47, + "learning_rate": 0.0004315025075225677, + "loss": 1.3453, + "theoretical_loss": 3.5367797954361633, + "tokens_seen": 1538195456 + }, + { + "epoch": 0.47, + "learning_rate": 0.00043148645937813446, + "loss": 1.4763, + "theoretical_loss": 3.536766680201305, + "tokens_seen": 1538260992 + }, + { + "epoch": 0.47, + "learning_rate": 0.0004314704112337011, + "loss": 1.4256, + "theoretical_loss": 3.536753565681641, + "tokens_seen": 1538326528 + }, + { + "epoch": 0.47, + "learning_rate": 0.0004314543630892678, + "loss": 1.2485, + "theoretical_loss": 3.5367404518771024, + "tokens_seen": 1538392064 + }, + { + "epoch": 0.47, + "learning_rate": 0.0004314383149448345, + "loss": 1.4571, + "theoretical_loss": 3.5367273387876192, + "tokens_seen": 1538457600 + }, + { + "epoch": 0.47, + "learning_rate": 0.0004314222668004012, + "loss": 1.437, + "theoretical_loss": 3.536714226413122, + "tokens_seen": 1538523136 + }, + { + "epoch": 0.47, + "learning_rate": 0.0004314062186559679, + "loss": 1.4113, + "theoretical_loss": 3.5367011147535408, + "tokens_seen": 1538588672 + }, + { + "epoch": 0.47, + "learning_rate": 0.0004313901705115346, + "loss": 1.3086, + "theoretical_loss": 3.5366880038088073, + "tokens_seen": 1538654208 + }, + { + "epoch": 0.47, + "learning_rate": 0.0004313741223671013, + "loss": 1.3988, + "theoretical_loss": 3.5366748935788515, + "tokens_seen": 1538719744 + }, + { + "epoch": 0.47, + "learning_rate": 0.000431358074222668, + "loss": 1.4131, + "theoretical_loss": 3.5366617840636043, + "tokens_seen": 1538785280 + }, + { + "epoch": 0.47, + "learning_rate": 0.0004313420260782347, + "loss": 1.3536, + "theoretical_loss": 3.5366486752629953, + "tokens_seen": 1538850816 + }, + { + "epoch": 0.47, + "learning_rate": 0.0004313259779338014, + "loss": 1.2641, + "theoretical_loss": 3.5366355671769565, + "tokens_seen": 1538916352 + }, + { + "epoch": 0.47, + "learning_rate": 0.00043130992978936806, + "loss": 1.3761, + "theoretical_loss": 3.5366224598054177, + "tokens_seen": 1538981888 + }, + { + "epoch": 0.47, + "learning_rate": 0.0004312938816449348, + "loss": 1.4192, + "theoretical_loss": 3.5366093531483096, + "tokens_seen": 1539047424 + }, + { + "epoch": 0.47, + "learning_rate": 0.00043127783350050146, + "loss": 1.4208, + "theoretical_loss": 3.536596247205563, + "tokens_seen": 1539112960 + }, + { + "epoch": 0.47, + "learning_rate": 0.0004312617853560682, + "loss": 1.4726, + "theoretical_loss": 3.5365831419771085, + "tokens_seen": 1539178496 + }, + { + "epoch": 0.47, + "learning_rate": 0.00043124573721163486, + "loss": 1.4227, + "theoretical_loss": 3.5365700374628766, + "tokens_seen": 1539244032 + }, + { + "epoch": 0.47, + "learning_rate": 0.0004312296890672016, + "loss": 1.2644, + "theoretical_loss": 3.5365569336627987, + "tokens_seen": 1539309568 + }, + { + "epoch": 0.47, + "learning_rate": 0.00043121364092276837, + "loss": 1.5164, + "theoretical_loss": 3.5365438305768038, + "tokens_seen": 1539375104 + }, + { + "epoch": 0.47, + "learning_rate": 0.000431197592778335, + "loss": 1.2681, + "theoretical_loss": 3.5365307282048244, + "tokens_seen": 1539440640 + }, + { + "epoch": 0.47, + "learning_rate": 0.00043118154463390177, + "loss": 1.3807, + "theoretical_loss": 3.5365176265467904, + "tokens_seen": 1539506176 + }, + { + "epoch": 0.47, + "learning_rate": 0.0004311654964894684, + "loss": 1.4141, + "theoretical_loss": 3.5365045256026324, + "tokens_seen": 1539571712 + }, + { + "epoch": 0.47, + "learning_rate": 0.00043114944834503517, + "loss": 1.4426, + "theoretical_loss": 3.5364914253722812, + "tokens_seen": 1539637248 + }, + { + "epoch": 0.47, + "learning_rate": 0.0004311334002006018, + "loss": 1.4684, + "theoretical_loss": 3.5364783258556685, + "tokens_seen": 1539702784 + }, + { + "epoch": 0.47, + "learning_rate": 0.0004311173520561686, + "loss": 1.3548, + "theoretical_loss": 3.536465227052723, + "tokens_seen": 1539768320 + }, + { + "epoch": 0.47, + "learning_rate": 0.0004311013039117352, + "loss": 1.437, + "theoretical_loss": 3.5364521289633775, + "tokens_seen": 1539833856 + }, + { + "epoch": 0.47, + "learning_rate": 0.000431085255767302, + "loss": 1.3866, + "theoretical_loss": 3.5364390315875616, + "tokens_seen": 1539899392 + }, + { + "epoch": 0.47, + "learning_rate": 0.0004310692076228686, + "loss": 1.4933, + "theoretical_loss": 3.536425934925206, + "tokens_seen": 1539964928 + }, + { + "epoch": 0.47, + "learning_rate": 0.0004310531594784354, + "loss": 1.4886, + "theoretical_loss": 3.5364128389762417, + "tokens_seen": 1540030464 + }, + { + "epoch": 0.47, + "objective/train/docs_used": 562258, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.3890984058380127, + "objective/train/theoretical_loss": 3.5363997437406, + "objective/train/tokens_used": 1560556000, + "theoretical_loss": 3.5363997437406, + "tokens_seen": 1540096000 + }, + { + "epoch": 0.47, + "learning_rate": 0.000431037111334002, + "loss": 1.3809, + "theoretical_loss": 3.5363997437406, + "tokens_seen": 1540096000 + }, + { + "epoch": 0.47, + "learning_rate": 0.0004310210631895688, + "loss": 1.4048, + "theoretical_loss": 3.5363866492182114, + "tokens_seen": 1540161536 + }, + { + "epoch": 0.47, + "learning_rate": 0.0004310050150451354, + "loss": 1.5214, + "theoretical_loss": 3.536373555409006, + "tokens_seen": 1540227072 + }, + { + "epoch": 0.47, + "learning_rate": 0.0004309889669007022, + "loss": 1.6403, + "theoretical_loss": 3.536360462312916, + "tokens_seen": 1540292608 + }, + { + "epoch": 0.47, + "learning_rate": 0.0004309729187562688, + "loss": 1.4197, + "theoretical_loss": 3.536347369929871, + "tokens_seen": 1540358144 + }, + { + "epoch": 0.47, + "learning_rate": 0.0004309568706118356, + "loss": 1.5069, + "theoretical_loss": 3.536334278259802, + "tokens_seen": 1540423680 + }, + { + "epoch": 0.47, + "learning_rate": 0.0004309408224674022, + "loss": 1.3772, + "theoretical_loss": 3.536321187302641, + "tokens_seen": 1540489216 + }, + { + "epoch": 0.47, + "learning_rate": 0.000430924774322969, + "loss": 1.4284, + "theoretical_loss": 3.536308097058317, + "tokens_seen": 1540554752 + }, + { + "epoch": 0.47, + "learning_rate": 0.0004309087261785356, + "loss": 1.3767, + "theoretical_loss": 3.5362950075267623, + "tokens_seen": 1540620288 + }, + { + "epoch": 0.47, + "learning_rate": 0.0004308926780341024, + "loss": 1.335, + "theoretical_loss": 3.536281918707907, + "tokens_seen": 1540685824 + }, + { + "epoch": 0.47, + "learning_rate": 0.00043087662988966903, + "loss": 1.4358, + "theoretical_loss": 3.5362688306016823, + "tokens_seen": 1540751360 + }, + { + "epoch": 0.47, + "learning_rate": 0.0004308605817452358, + "loss": 1.382, + "theoretical_loss": 3.5362557432080193, + "tokens_seen": 1540816896 + }, + { + "epoch": 0.47, + "learning_rate": 0.00043084453360080243, + "loss": 1.297, + "theoretical_loss": 3.5362426565268485, + "tokens_seen": 1540882432 + }, + { + "epoch": 0.47, + "learning_rate": 0.0004308284854563692, + "loss": 1.3213, + "theoretical_loss": 3.536229570558101, + "tokens_seen": 1540947968 + }, + { + "epoch": 0.47, + "learning_rate": 0.00043081243731193583, + "loss": 1.4747, + "theoretical_loss": 3.5362164853017077, + "tokens_seen": 1541013504 + }, + { + "epoch": 0.47, + "learning_rate": 0.00043079638916750253, + "loss": 1.3802, + "theoretical_loss": 3.5362034007575995, + "tokens_seen": 1541079040 + }, + { + "epoch": 0.47, + "learning_rate": 0.00043078034102306923, + "loss": 1.4207, + "theoretical_loss": 3.5361903169257074, + "tokens_seen": 1541144576 + }, + { + "epoch": 0.47, + "learning_rate": 0.00043076429287863593, + "loss": 1.4123, + "theoretical_loss": 3.5361772338059625, + "tokens_seen": 1541210112 + }, + { + "epoch": 0.47, + "learning_rate": 0.00043074824473420263, + "loss": 1.4575, + "theoretical_loss": 3.536164151398295, + "tokens_seen": 1541275648 + }, + { + "epoch": 0.47, + "learning_rate": 0.00043073219658976933, + "loss": 1.4016, + "theoretical_loss": 3.536151069702637, + "tokens_seen": 1541341184 + }, + { + "epoch": 0.47, + "learning_rate": 0.00043071614844533603, + "loss": 1.3965, + "theoretical_loss": 3.5361379887189184, + "tokens_seen": 1541406720 + }, + { + "epoch": 0.47, + "learning_rate": 0.00043070010030090274, + "loss": 1.4518, + "theoretical_loss": 3.5361249084470714, + "tokens_seen": 1541472256 + }, + { + "epoch": 0.47, + "learning_rate": 0.0004306840521564694, + "loss": 1.3931, + "theoretical_loss": 3.5361118288870257, + "tokens_seen": 1541537792 + }, + { + "epoch": 0.47, + "learning_rate": 0.00043066800401203614, + "loss": 1.4467, + "theoretical_loss": 3.536098750038713, + "tokens_seen": 1541603328 + }, + { + "epoch": 0.47, + "learning_rate": 0.0004306519558676028, + "loss": 1.4299, + "theoretical_loss": 3.5360856719020646, + "tokens_seen": 1541668864 + }, + { + "epoch": 0.47, + "learning_rate": 0.00043063590772316954, + "loss": 1.3837, + "theoretical_loss": 3.5360725944770106, + "tokens_seen": 1541734400 + }, + { + "epoch": 0.47, + "learning_rate": 0.0004306198595787362, + "loss": 1.4437, + "theoretical_loss": 3.5360595177634826, + "tokens_seen": 1541799936 + }, + { + "epoch": 0.47, + "learning_rate": 0.00043060381143430294, + "loss": 1.288, + "theoretical_loss": 3.536046441761412, + "tokens_seen": 1541865472 + }, + { + "epoch": 0.47, + "learning_rate": 0.0004305877632898696, + "loss": 1.4834, + "theoretical_loss": 3.5360333664707295, + "tokens_seen": 1541931008 + }, + { + "epoch": 0.47, + "learning_rate": 0.00043057171514543634, + "loss": 1.3939, + "theoretical_loss": 3.536020291891366, + "tokens_seen": 1541996544 + }, + { + "epoch": 0.47, + "learning_rate": 0.000430555667001003, + "loss": 1.3753, + "theoretical_loss": 3.5360072180232525, + "tokens_seen": 1542062080 + }, + { + "epoch": 0.47, + "learning_rate": 0.00043053961885656974, + "loss": 1.3974, + "theoretical_loss": 3.5359941448663204, + "tokens_seen": 1542127616 + }, + { + "epoch": 0.47, + "learning_rate": 0.0004305235707121364, + "loss": 1.3483, + "theoretical_loss": 3.535981072420501, + "tokens_seen": 1542193152 + }, + { + "epoch": 0.47, + "learning_rate": 0.00043050752256770314, + "loss": 1.442, + "theoretical_loss": 3.535968000685725, + "tokens_seen": 1542258688 + }, + { + "epoch": 0.47, + "learning_rate": 0.0004304914744232698, + "loss": 1.4057, + "theoretical_loss": 3.5359549296619237, + "tokens_seen": 1542324224 + }, + { + "epoch": 0.47, + "learning_rate": 0.00043047542627883654, + "loss": 1.4376, + "theoretical_loss": 3.5359418593490277, + "tokens_seen": 1542389760 + }, + { + "epoch": 0.47, + "learning_rate": 0.0004304593781344032, + "loss": 1.4017, + "theoretical_loss": 3.535928789746969, + "tokens_seen": 1542455296 + }, + { + "epoch": 0.47, + "learning_rate": 0.00043044332998996995, + "loss": 1.4491, + "theoretical_loss": 3.5359157208556775, + "tokens_seen": 1542520832 + }, + { + "epoch": 0.47, + "learning_rate": 0.0004304272818455366, + "loss": 1.2988, + "theoretical_loss": 3.5359026526750865, + "tokens_seen": 1542586368 + }, + { + "epoch": 0.47, + "learning_rate": 0.00043041123370110335, + "loss": 1.4854, + "theoretical_loss": 3.5358895852051244, + "tokens_seen": 1542651904 + }, + { + "epoch": 0.47, + "learning_rate": 0.00043039518555667, + "loss": 1.4095, + "theoretical_loss": 3.5358765184457246, + "tokens_seen": 1542717440 + }, + { + "epoch": 0.47, + "learning_rate": 0.00043037913741223675, + "loss": 1.3797, + "theoretical_loss": 3.5358634523968178, + "tokens_seen": 1542782976 + }, + { + "epoch": 0.47, + "learning_rate": 0.0004303630892678034, + "loss": 1.3347, + "theoretical_loss": 3.535850387058334, + "tokens_seen": 1542848512 + }, + { + "epoch": 0.47, + "learning_rate": 0.00043034704112337015, + "loss": 1.5013, + "theoretical_loss": 3.535837322430206, + "tokens_seen": 1542914048 + }, + { + "epoch": 0.47, + "learning_rate": 0.0004303309929789368, + "loss": 1.342, + "theoretical_loss": 3.5358242585123634, + "tokens_seen": 1542979584 + }, + { + "epoch": 0.47, + "learning_rate": 0.00043031494483450355, + "loss": 1.4292, + "theoretical_loss": 3.5358111953047393, + "tokens_seen": 1543045120 + }, + { + "epoch": 0.47, + "learning_rate": 0.0004302988966900702, + "loss": 1.4444, + "theoretical_loss": 3.535798132807263, + "tokens_seen": 1543110656 + }, + { + "epoch": 0.47, + "learning_rate": 0.00043028284854563695, + "loss": 1.4365, + "theoretical_loss": 3.535785071019867, + "tokens_seen": 1543176192 + }, + { + "epoch": 0.47, + "learning_rate": 0.0004302668004012036, + "loss": 1.4633, + "theoretical_loss": 3.535772009942482, + "tokens_seen": 1543241728 + }, + { + "epoch": 0.47, + "learning_rate": 0.00043025075225677035, + "loss": 1.3589, + "theoretical_loss": 3.53575894957504, + "tokens_seen": 1543307264 + }, + { + "epoch": 0.47, + "objective/train/docs_used": 563510, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.3954761028289795, + "objective/train/theoretical_loss": 3.5357458899174716, + "objective/train/tokens_used": 1563832800, + "theoretical_loss": 3.5357458899174716, + "tokens_seen": 1543372800 + }, + { + "epoch": 0.47, + "learning_rate": 0.000430234704112337, + "loss": 1.2943, + "theoretical_loss": 3.5357458899174716, + "tokens_seen": 1543372800 + }, + { + "epoch": 0.47, + "learning_rate": 0.00043021865596790375, + "loss": 1.4569, + "theoretical_loss": 3.5357328309697076, + "tokens_seen": 1543438336 + }, + { + "epoch": 0.47, + "learning_rate": 0.0004302026078234704, + "loss": 1.3762, + "theoretical_loss": 3.5357197727316807, + "tokens_seen": 1543503872 + }, + { + "epoch": 0.47, + "learning_rate": 0.00043018655967903716, + "loss": 1.5222, + "theoretical_loss": 3.535706715203321, + "tokens_seen": 1543569408 + }, + { + "epoch": 0.47, + "learning_rate": 0.0004301705115346038, + "loss": 1.3252, + "theoretical_loss": 3.5356936583845595, + "tokens_seen": 1543634944 + }, + { + "epoch": 0.47, + "learning_rate": 0.00043015446339017056, + "loss": 1.4879, + "theoretical_loss": 3.5356806022753284, + "tokens_seen": 1543700480 + }, + { + "epoch": 0.47, + "learning_rate": 0.0004301384152457372, + "loss": 1.4005, + "theoretical_loss": 3.5356675468755596, + "tokens_seen": 1543766016 + }, + { + "epoch": 0.47, + "learning_rate": 0.00043012236710130396, + "loss": 1.397, + "theoretical_loss": 3.5356544921851834, + "tokens_seen": 1543831552 + }, + { + "epoch": 0.47, + "learning_rate": 0.0004301063189568706, + "loss": 1.3656, + "theoretical_loss": 3.535641438204131, + "tokens_seen": 1543897088 + }, + { + "epoch": 0.47, + "learning_rate": 0.00043009027081243736, + "loss": 1.4591, + "theoretical_loss": 3.535628384932334, + "tokens_seen": 1543962624 + }, + { + "epoch": 0.47, + "learning_rate": 0.000430074222668004, + "loss": 1.2977, + "theoretical_loss": 3.5356153323697246, + "tokens_seen": 1544028160 + }, + { + "epoch": 0.47, + "learning_rate": 0.00043005817452357076, + "loss": 1.4083, + "theoretical_loss": 3.5356022805162333, + "tokens_seen": 1544093696 + }, + { + "epoch": 0.47, + "learning_rate": 0.0004300421263791374, + "loss": 1.3681, + "theoretical_loss": 3.5355892293717908, + "tokens_seen": 1544159232 + }, + { + "epoch": 0.47, + "learning_rate": 0.0004300260782347041, + "loss": 1.4696, + "theoretical_loss": 3.5355761789363305, + "tokens_seen": 1544224768 + }, + { + "epoch": 0.47, + "learning_rate": 0.00043001003009027086, + "loss": 1.4377, + "theoretical_loss": 3.5355631292097813, + "tokens_seen": 1544290304 + }, + { + "epoch": 0.47, + "learning_rate": 0.0004299939819458375, + "loss": 1.5179, + "theoretical_loss": 3.535550080192077, + "tokens_seen": 1544355840 + }, + { + "epoch": 0.47, + "learning_rate": 0.00042997793380140426, + "loss": 1.4231, + "theoretical_loss": 3.535537031883148, + "tokens_seen": 1544421376 + }, + { + "epoch": 0.47, + "learning_rate": 0.0004299618856569709, + "loss": 1.5551, + "theoretical_loss": 3.535523984282925, + "tokens_seen": 1544486912 + }, + { + "epoch": 0.47, + "learning_rate": 0.00042994583751253767, + "loss": 1.3306, + "theoretical_loss": 3.535510937391341, + "tokens_seen": 1544552448 + }, + { + "epoch": 0.47, + "learning_rate": 0.0004299297893681043, + "loss": 1.3875, + "theoretical_loss": 3.5354978912083257, + "tokens_seen": 1544617984 + }, + { + "epoch": 0.47, + "learning_rate": 0.00042991374122367107, + "loss": 1.2927, + "theoretical_loss": 3.5354848457338117, + "tokens_seen": 1544683520 + }, + { + "epoch": 0.47, + "learning_rate": 0.0004298976930792377, + "loss": 1.323, + "theoretical_loss": 3.5354718009677306, + "tokens_seen": 1544749056 + }, + { + "epoch": 0.47, + "learning_rate": 0.00042988164493480447, + "loss": 1.424, + "theoretical_loss": 3.535458756910013, + "tokens_seen": 1544814592 + }, + { + "epoch": 0.47, + "learning_rate": 0.0004298655967903711, + "loss": 1.2893, + "theoretical_loss": 3.5354457135605912, + "tokens_seen": 1544880128 + }, + { + "epoch": 0.47, + "learning_rate": 0.00042984954864593787, + "loss": 1.3742, + "theoretical_loss": 3.5354326709193957, + "tokens_seen": 1544945664 + }, + { + "epoch": 0.47, + "learning_rate": 0.0004298335005015045, + "loss": 1.449, + "theoretical_loss": 3.5354196289863595, + "tokens_seen": 1545011200 + }, + { + "epoch": 0.47, + "learning_rate": 0.00042981745235707127, + "loss": 1.4613, + "theoretical_loss": 3.535406587761413, + "tokens_seen": 1545076736 + }, + { + "epoch": 0.47, + "learning_rate": 0.0004298014042126379, + "loss": 1.3494, + "theoretical_loss": 3.535393547244488, + "tokens_seen": 1545142272 + }, + { + "epoch": 0.47, + "learning_rate": 0.00042978535606820467, + "loss": 1.4819, + "theoretical_loss": 3.535380507435516, + "tokens_seen": 1545207808 + }, + { + "epoch": 0.47, + "learning_rate": 0.0004297693079237713, + "loss": 1.3557, + "theoretical_loss": 3.535367468334428, + "tokens_seen": 1545273344 + }, + { + "epoch": 0.47, + "learning_rate": 0.00042975325977933807, + "loss": 1.2885, + "theoretical_loss": 3.5353544299411572, + "tokens_seen": 1545338880 + }, + { + "epoch": 0.47, + "learning_rate": 0.0004297372116349047, + "loss": 1.5131, + "theoretical_loss": 3.535341392255633, + "tokens_seen": 1545404416 + }, + { + "epoch": 0.47, + "learning_rate": 0.0004297211634904715, + "loss": 1.3808, + "theoretical_loss": 3.5353283552777888, + "tokens_seen": 1545469952 + }, + { + "epoch": 0.47, + "learning_rate": 0.0004297051153460381, + "loss": 1.4106, + "theoretical_loss": 3.5353153190075552, + "tokens_seen": 1545535488 + }, + { + "epoch": 0.47, + "learning_rate": 0.0004296890672016049, + "loss": 1.3757, + "theoretical_loss": 3.5353022834448637, + "tokens_seen": 1545601024 + }, + { + "epoch": 0.47, + "learning_rate": 0.0004296730190571715, + "loss": 1.4461, + "theoretical_loss": 3.535289248589647, + "tokens_seen": 1545666560 + }, + { + "epoch": 0.47, + "learning_rate": 0.0004296569709127383, + "loss": 1.4636, + "theoretical_loss": 3.5352762144418355, + "tokens_seen": 1545732096 + }, + { + "epoch": 0.47, + "learning_rate": 0.0004296409227683049, + "loss": 1.5161, + "theoretical_loss": 3.535263181001361, + "tokens_seen": 1545797632 + }, + { + "epoch": 0.47, + "learning_rate": 0.0004296248746238717, + "loss": 1.6076, + "theoretical_loss": 3.5352501482681555, + "tokens_seen": 1545863168 + }, + { + "epoch": 0.47, + "learning_rate": 0.0004296088264794383, + "loss": 1.4224, + "theoretical_loss": 3.535237116242151, + "tokens_seen": 1545928704 + }, + { + "epoch": 0.47, + "learning_rate": 0.0004295927783350051, + "loss": 1.3952, + "theoretical_loss": 3.5352240849232786, + "tokens_seen": 1545994240 + }, + { + "epoch": 0.47, + "learning_rate": 0.0004295767301905717, + "loss": 1.3804, + "theoretical_loss": 3.5352110543114694, + "tokens_seen": 1546059776 + }, + { + "epoch": 0.47, + "learning_rate": 0.0004295606820461385, + "loss": 1.4203, + "theoretical_loss": 3.535198024406656, + "tokens_seen": 1546125312 + }, + { + "epoch": 0.47, + "learning_rate": 0.0004295446339017051, + "loss": 1.3219, + "theoretical_loss": 3.5351849952087697, + "tokens_seen": 1546190848 + }, + { + "epoch": 0.47, + "learning_rate": 0.0004295285857572719, + "loss": 1.4452, + "theoretical_loss": 3.5351719667177424, + "tokens_seen": 1546256384 + }, + { + "epoch": 0.47, + "learning_rate": 0.00042951253761283853, + "loss": 1.5108, + "theoretical_loss": 3.535158938933506, + "tokens_seen": 1546321920 + }, + { + "epoch": 0.47, + "learning_rate": 0.0004294964894684053, + "loss": 1.3798, + "theoretical_loss": 3.5351459118559916, + "tokens_seen": 1546387456 + }, + { + "epoch": 0.47, + "learning_rate": 0.00042948044132397193, + "loss": 1.3078, + "theoretical_loss": 3.5351328854851305, + "tokens_seen": 1546452992 + }, + { + "epoch": 0.47, + "learning_rate": 0.0004294643931795387, + "loss": 1.399, + "theoretical_loss": 3.535119859820856, + "tokens_seen": 1546518528 + }, + { + "epoch": 0.47, + "learning_rate": 0.00042944834503510533, + "loss": 1.2924, + "theoretical_loss": 3.5351068348630985, + "tokens_seen": 1546584064 + }, + { + "epoch": 0.47, + "objective/train/docs_used": 564726, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.3714215755462646, + "objective/train/theoretical_loss": 3.53509381061179, + "objective/train/tokens_used": 1567109600, + "theoretical_loss": 3.53509381061179, + "tokens_seen": 1546649600 + }, + { + "epoch": 0.47, + "learning_rate": 0.0004294322968906721, + "loss": 1.3968, + "theoretical_loss": 3.53509381061179, + "tokens_seen": 1546649600 + }, + { + "epoch": 0.47, + "learning_rate": 0.00042941624874623873, + "loss": 1.4927, + "theoretical_loss": 3.535080787066863, + "tokens_seen": 1546715136 + }, + { + "epoch": 0.47, + "learning_rate": 0.00042940020060180543, + "loss": 1.4585, + "theoretical_loss": 3.5350677642282484, + "tokens_seen": 1546780672 + }, + { + "epoch": 0.47, + "learning_rate": 0.00042938415245737213, + "loss": 1.3979, + "theoretical_loss": 3.5350547420958778, + "tokens_seen": 1546846208 + }, + { + "epoch": 0.47, + "learning_rate": 0.00042936810431293883, + "loss": 1.4248, + "theoretical_loss": 3.5350417206696836, + "tokens_seen": 1546911744 + }, + { + "epoch": 0.47, + "learning_rate": 0.00042935205616850553, + "loss": 1.4865, + "theoretical_loss": 3.535028699949598, + "tokens_seen": 1546977280 + }, + { + "epoch": 0.47, + "learning_rate": 0.00042933600802407223, + "loss": 1.4359, + "theoretical_loss": 3.535015679935552, + "tokens_seen": 1547042816 + }, + { + "epoch": 0.47, + "learning_rate": 0.00042931995987963894, + "loss": 1.4793, + "theoretical_loss": 3.535002660627477, + "tokens_seen": 1547108352 + }, + { + "epoch": 0.47, + "learning_rate": 0.00042930391173520564, + "loss": 1.2992, + "theoretical_loss": 3.534989642025306, + "tokens_seen": 1547173888 + }, + { + "epoch": 0.47, + "learning_rate": 0.00042928786359077234, + "loss": 1.4503, + "theoretical_loss": 3.53497662412897, + "tokens_seen": 1547239424 + }, + { + "epoch": 0.47, + "learning_rate": 0.00042927181544633904, + "loss": 1.4176, + "theoretical_loss": 3.534963606938401, + "tokens_seen": 1547304960 + }, + { + "epoch": 0.47, + "learning_rate": 0.0004292557673019057, + "loss": 1.3628, + "theoretical_loss": 3.5349505904535317, + "tokens_seen": 1547370496 + }, + { + "epoch": 0.47, + "learning_rate": 0.00042923971915747244, + "loss": 1.4023, + "theoretical_loss": 3.5349375746742924, + "tokens_seen": 1547436032 + }, + { + "epoch": 0.47, + "learning_rate": 0.0004292236710130391, + "loss": 1.5049, + "theoretical_loss": 3.5349245596006162, + "tokens_seen": 1547501568 + }, + { + "epoch": 0.47, + "learning_rate": 0.00042920762286860584, + "loss": 1.3245, + "theoretical_loss": 3.5349115452324344, + "tokens_seen": 1547567104 + }, + { + "epoch": 0.47, + "learning_rate": 0.0004291915747241725, + "loss": 1.3778, + "theoretical_loss": 3.534898531569679, + "tokens_seen": 1547632640 + }, + { + "epoch": 0.47, + "learning_rate": 0.00042917552657973924, + "loss": 1.4321, + "theoretical_loss": 3.534885518612282, + "tokens_seen": 1547698176 + }, + { + "epoch": 0.47, + "learning_rate": 0.0004291594784353059, + "loss": 1.5784, + "theoretical_loss": 3.5348725063601756, + "tokens_seen": 1547763712 + }, + { + "epoch": 0.47, + "learning_rate": 0.00042914343029087264, + "loss": 1.4192, + "theoretical_loss": 3.5348594948132908, + "tokens_seen": 1547829248 + }, + { + "epoch": 0.47, + "learning_rate": 0.0004291273821464393, + "loss": 1.5069, + "theoretical_loss": 3.5348464839715605, + "tokens_seen": 1547894784 + }, + { + "epoch": 0.47, + "learning_rate": 0.00042911133400200604, + "loss": 1.4905, + "theoretical_loss": 3.534833473834916, + "tokens_seen": 1547960320 + }, + { + "epoch": 0.47, + "learning_rate": 0.0004290952858575727, + "loss": 1.4986, + "theoretical_loss": 3.534820464403289, + "tokens_seen": 1548025856 + }, + { + "epoch": 0.47, + "learning_rate": 0.00042907923771313945, + "loss": 1.4824, + "theoretical_loss": 3.5348074556766127, + "tokens_seen": 1548091392 + }, + { + "epoch": 0.47, + "learning_rate": 0.0004290631895687061, + "loss": 1.3573, + "theoretical_loss": 3.534794447654818, + "tokens_seen": 1548156928 + }, + { + "epoch": 0.47, + "learning_rate": 0.00042904714142427285, + "loss": 1.3999, + "theoretical_loss": 3.534781440337836, + "tokens_seen": 1548222464 + }, + { + "epoch": 0.47, + "learning_rate": 0.0004290310932798395, + "loss": 1.5471, + "theoretical_loss": 3.5347684337256013, + "tokens_seen": 1548288000 + }, + { + "epoch": 0.47, + "learning_rate": 0.00042901504513540625, + "loss": 1.4291, + "theoretical_loss": 3.5347554278180437, + "tokens_seen": 1548353536 + }, + { + "epoch": 0.47, + "learning_rate": 0.0004289989969909729, + "loss": 1.4791, + "theoretical_loss": 3.534742422615096, + "tokens_seen": 1548419072 + }, + { + "epoch": 0.47, + "learning_rate": 0.00042898294884653965, + "loss": 1.4135, + "theoretical_loss": 3.53472941811669, + "tokens_seen": 1548484608 + }, + { + "epoch": 0.47, + "learning_rate": 0.0004289669007021063, + "loss": 1.4042, + "theoretical_loss": 3.534716414322758, + "tokens_seen": 1548550144 + }, + { + "epoch": 0.47, + "learning_rate": 0.00042895085255767305, + "loss": 1.4942, + "theoretical_loss": 3.534703411233232, + "tokens_seen": 1548615680 + }, + { + "epoch": 0.47, + "learning_rate": 0.0004289348044132397, + "loss": 1.3165, + "theoretical_loss": 3.5346904088480433, + "tokens_seen": 1548681216 + }, + { + "epoch": 0.47, + "learning_rate": 0.00042891875626880645, + "loss": 1.3527, + "theoretical_loss": 3.534677407167125, + "tokens_seen": 1548746752 + }, + { + "epoch": 0.47, + "learning_rate": 0.0004289027081243731, + "loss": 1.3761, + "theoretical_loss": 3.5346644061904082, + "tokens_seen": 1548812288 + }, + { + "epoch": 0.47, + "learning_rate": 0.00042888665997993985, + "loss": 1.4079, + "theoretical_loss": 3.534651405917826, + "tokens_seen": 1548877824 + }, + { + "epoch": 0.47, + "learning_rate": 0.0004288706118355065, + "loss": 1.4775, + "theoretical_loss": 3.5346384063493095, + "tokens_seen": 1548943360 + }, + { + "epoch": 0.47, + "learning_rate": 0.00042885456369107325, + "loss": 1.5102, + "theoretical_loss": 3.534625407484791, + "tokens_seen": 1549008896 + }, + { + "epoch": 0.47, + "learning_rate": 0.0004288385155466399, + "loss": 1.2988, + "theoretical_loss": 3.534612409324203, + "tokens_seen": 1549074432 + }, + { + "epoch": 0.47, + "learning_rate": 0.00042882246740220666, + "loss": 1.325, + "theoretical_loss": 3.534599411867478, + "tokens_seen": 1549139968 + }, + { + "epoch": 0.47, + "learning_rate": 0.0004288064192577734, + "loss": 1.4539, + "theoretical_loss": 3.5345864151145467, + "tokens_seen": 1549205504 + }, + { + "epoch": 0.47, + "learning_rate": 0.00042879037111334006, + "loss": 1.4722, + "theoretical_loss": 3.5345734190653424, + "tokens_seen": 1549271040 + }, + { + "epoch": 0.47, + "learning_rate": 0.00042877432296890676, + "loss": 1.251, + "theoretical_loss": 3.534560423719797, + "tokens_seen": 1549336576 + }, + { + "epoch": 0.47, + "learning_rate": 0.00042875827482447346, + "loss": 1.4811, + "theoretical_loss": 3.5345474290778416, + "tokens_seen": 1549402112 + }, + { + "epoch": 0.47, + "learning_rate": 0.00042874222668004016, + "loss": 1.5344, + "theoretical_loss": 3.5345344351394097, + "tokens_seen": 1549467648 + }, + { + "epoch": 0.47, + "learning_rate": 0.00042872617853560686, + "loss": 1.3778, + "theoretical_loss": 3.5345214419044333, + "tokens_seen": 1549533184 + }, + { + "epoch": 0.47, + "learning_rate": 0.00042871013039117356, + "loss": 1.5768, + "theoretical_loss": 3.534508449372844, + "tokens_seen": 1549598720 + }, + { + "epoch": 0.47, + "learning_rate": 0.00042869408224674026, + "loss": 1.4618, + "theoretical_loss": 3.5344954575445744, + "tokens_seen": 1549664256 + }, + { + "epoch": 0.47, + "learning_rate": 0.00042867803410230696, + "loss": 1.3587, + "theoretical_loss": 3.5344824664195564, + "tokens_seen": 1549729792 + }, + { + "epoch": 0.47, + "learning_rate": 0.00042866198595787366, + "loss": 1.3895, + "theoretical_loss": 3.534469475997722, + "tokens_seen": 1549795328 + }, + { + "epoch": 0.47, + "learning_rate": 0.00042864593781344036, + "loss": 1.4334, + "theoretical_loss": 3.534456486279004, + "tokens_seen": 1549860864 + }, + { + "epoch": 0.47, + "objective/train/docs_used": 565981, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.4784154891967773, + "objective/train/theoretical_loss": 3.534443497263335, + "objective/train/tokens_used": 1570386400, + "theoretical_loss": 3.534443497263335, + "tokens_seen": 1549926400 + }, + { + "epoch": 0.47, + "learning_rate": 0.000428629889669007, + "loss": 1.5011, + "theoretical_loss": 3.534443497263335, + "tokens_seen": 1549926400 + }, + { + "epoch": 0.47, + "learning_rate": 0.00042861384152457376, + "loss": 1.3814, + "theoretical_loss": 3.5344305089506456, + "tokens_seen": 1549991936 + }, + { + "epoch": 0.47, + "learning_rate": 0.0004285977933801404, + "loss": 1.3006, + "theoretical_loss": 3.5344175213408695, + "tokens_seen": 1550057472 + }, + { + "epoch": 0.47, + "learning_rate": 0.00042858174523570716, + "loss": 1.2979, + "theoretical_loss": 3.534404534433938, + "tokens_seen": 1550123008 + }, + { + "epoch": 0.47, + "learning_rate": 0.0004285656970912738, + "loss": 1.4768, + "theoretical_loss": 3.534391548229784, + "tokens_seen": 1550188544 + }, + { + "epoch": 0.47, + "learning_rate": 0.00042854964894684057, + "loss": 1.436, + "theoretical_loss": 3.5343785627283397, + "tokens_seen": 1550254080 + }, + { + "epoch": 0.47, + "learning_rate": 0.0004285336008024072, + "loss": 1.4694, + "theoretical_loss": 3.5343655779295373, + "tokens_seen": 1550319616 + }, + { + "epoch": 0.47, + "learning_rate": 0.00042851755265797397, + "loss": 1.3905, + "theoretical_loss": 3.5343525938333085, + "tokens_seen": 1550385152 + }, + { + "epoch": 0.47, + "learning_rate": 0.0004285015045135406, + "loss": 1.523, + "theoretical_loss": 3.534339610439586, + "tokens_seen": 1550450688 + }, + { + "epoch": 0.47, + "learning_rate": 0.00042848545636910737, + "loss": 1.4558, + "theoretical_loss": 3.534326627748303, + "tokens_seen": 1550516224 + }, + { + "epoch": 0.47, + "learning_rate": 0.000428469408224674, + "loss": 1.3714, + "theoretical_loss": 3.5343136457593904, + "tokens_seen": 1550581760 + }, + { + "epoch": 0.47, + "learning_rate": 0.00042845336008024077, + "loss": 1.4585, + "theoretical_loss": 3.5343006644727817, + "tokens_seen": 1550647296 + }, + { + "epoch": 0.47, + "learning_rate": 0.0004284373119358074, + "loss": 1.3556, + "theoretical_loss": 3.534287683888408, + "tokens_seen": 1550712832 + }, + { + "epoch": 0.47, + "learning_rate": 0.00042842126379137417, + "loss": 1.425, + "theoretical_loss": 3.5342747040062026, + "tokens_seen": 1550778368 + }, + { + "epoch": 0.47, + "learning_rate": 0.0004284052156469408, + "loss": 1.3663, + "theoretical_loss": 3.534261724826097, + "tokens_seen": 1550843904 + }, + { + "epoch": 0.47, + "learning_rate": 0.00042838916750250757, + "loss": 1.4007, + "theoretical_loss": 3.5342487463480246, + "tokens_seen": 1550909440 + }, + { + "epoch": 0.47, + "learning_rate": 0.0004283731193580742, + "loss": 1.4103, + "theoretical_loss": 3.534235768571917, + "tokens_seen": 1550974976 + }, + { + "epoch": 0.47, + "learning_rate": 0.000428357071213641, + "loss": 1.3167, + "theoretical_loss": 3.534222791497707, + "tokens_seen": 1551040512 + }, + { + "epoch": 0.47, + "learning_rate": 0.0004283410230692076, + "loss": 1.3194, + "theoretical_loss": 3.534209815125327, + "tokens_seen": 1551106048 + }, + { + "epoch": 0.47, + "learning_rate": 0.0004283249749247744, + "loss": 1.3772, + "theoretical_loss": 3.5341968394547085, + "tokens_seen": 1551171584 + }, + { + "epoch": 0.47, + "learning_rate": 0.000428308926780341, + "loss": 1.3922, + "theoretical_loss": 3.534183864485785, + "tokens_seen": 1551237120 + }, + { + "epoch": 0.47, + "learning_rate": 0.0004282928786359078, + "loss": 1.4169, + "theoretical_loss": 3.5341708902184887, + "tokens_seen": 1551302656 + }, + { + "epoch": 0.47, + "learning_rate": 0.0004282768304914744, + "loss": 1.5142, + "theoretical_loss": 3.5341579166527515, + "tokens_seen": 1551368192 + }, + { + "epoch": 0.47, + "learning_rate": 0.0004282607823470412, + "loss": 1.3861, + "theoretical_loss": 3.5341449437885064, + "tokens_seen": 1551433728 + }, + { + "epoch": 0.47, + "learning_rate": 0.0004282447342026078, + "loss": 1.3964, + "theoretical_loss": 3.534131971625685, + "tokens_seen": 1551499264 + }, + { + "epoch": 0.47, + "learning_rate": 0.0004282286860581746, + "loss": 1.3239, + "theoretical_loss": 3.534119000164221, + "tokens_seen": 1551564800 + }, + { + "epoch": 0.47, + "learning_rate": 0.0004282126379137412, + "loss": 1.3232, + "theoretical_loss": 3.5341060294040463, + "tokens_seen": 1551630336 + }, + { + "epoch": 0.47, + "learning_rate": 0.000428196589769308, + "loss": 1.4355, + "theoretical_loss": 3.5340930593450928, + "tokens_seen": 1551695872 + }, + { + "epoch": 0.47, + "learning_rate": 0.0004281805416248746, + "loss": 1.3387, + "theoretical_loss": 3.5340800899872935, + "tokens_seen": 1551761408 + }, + { + "epoch": 0.47, + "learning_rate": 0.0004281644934804414, + "loss": 1.3746, + "theoretical_loss": 3.5340671213305814, + "tokens_seen": 1551826944 + }, + { + "epoch": 0.47, + "learning_rate": 0.00042814844533600803, + "loss": 1.3568, + "theoretical_loss": 3.5340541533748877, + "tokens_seen": 1551892480 + }, + { + "epoch": 0.47, + "learning_rate": 0.0004281323971915748, + "loss": 1.3696, + "theoretical_loss": 3.534041186120146, + "tokens_seen": 1551958016 + }, + { + "epoch": 0.47, + "learning_rate": 0.00042811634904714143, + "loss": 1.4259, + "theoretical_loss": 3.534028219566289, + "tokens_seen": 1552023552 + }, + { + "epoch": 0.47, + "learning_rate": 0.0004281003009027082, + "loss": 1.2865, + "theoretical_loss": 3.5340152537132474, + "tokens_seen": 1552089088 + }, + { + "epoch": 0.47, + "learning_rate": 0.00042808425275827483, + "loss": 1.4547, + "theoretical_loss": 3.5340022885609557, + "tokens_seen": 1552154624 + }, + { + "epoch": 0.47, + "learning_rate": 0.0004280682046138416, + "loss": 1.421, + "theoretical_loss": 3.5339893241093456, + "tokens_seen": 1552220160 + }, + { + "epoch": 0.47, + "learning_rate": 0.00042805215646940823, + "loss": 1.2858, + "theoretical_loss": 3.5339763603583503, + "tokens_seen": 1552285696 + }, + { + "epoch": 0.47, + "learning_rate": 0.000428036108324975, + "loss": 1.3425, + "theoretical_loss": 3.533963397307901, + "tokens_seen": 1552351232 + }, + { + "epoch": 0.47, + "learning_rate": 0.00042802006018054163, + "loss": 1.3358, + "theoretical_loss": 3.533950434957932, + "tokens_seen": 1552416768 + }, + { + "epoch": 0.47, + "learning_rate": 0.00042800401203610833, + "loss": 1.3322, + "theoretical_loss": 3.5339374733083746, + "tokens_seen": 1552482304 + }, + { + "epoch": 0.47, + "learning_rate": 0.00042798796389167503, + "loss": 1.3777, + "theoretical_loss": 3.5339245123591616, + "tokens_seen": 1552547840 + }, + { + "epoch": 0.47, + "learning_rate": 0.00042797191574724173, + "loss": 1.4446, + "theoretical_loss": 3.533911552110226, + "tokens_seen": 1552613376 + }, + { + "epoch": 0.47, + "learning_rate": 0.00042795586760280844, + "loss": 1.49, + "theoretical_loss": 3.5338985925615, + "tokens_seen": 1552678912 + }, + { + "epoch": 0.47, + "learning_rate": 0.00042793981945837514, + "loss": 1.4596, + "theoretical_loss": 3.533885633712917, + "tokens_seen": 1552744448 + }, + { + "epoch": 0.47, + "learning_rate": 0.00042792377131394184, + "loss": 1.4276, + "theoretical_loss": 3.5338726755644085, + "tokens_seen": 1552809984 + }, + { + "epoch": 0.47, + "learning_rate": 0.00042790772316950854, + "loss": 1.3965, + "theoretical_loss": 3.5338597181159077, + "tokens_seen": 1552875520 + }, + { + "epoch": 0.47, + "learning_rate": 0.00042789167502507524, + "loss": 1.39, + "theoretical_loss": 3.5338467613673474, + "tokens_seen": 1552941056 + }, + { + "epoch": 0.47, + "learning_rate": 0.00042787562688064194, + "loss": 1.3624, + "theoretical_loss": 3.53383380531866, + "tokens_seen": 1553006592 + }, + { + "epoch": 0.47, + "learning_rate": 0.0004278595787362086, + "loss": 1.3586, + "theoretical_loss": 3.5338208499697785, + "tokens_seen": 1553072128 + }, + { + "epoch": 0.47, + "learning_rate": 0.00042784353059177534, + "loss": 1.3556, + "theoretical_loss": 3.5338078953206358, + "tokens_seen": 1553137664 + }, + { + "epoch": 0.47, + "objective/train/docs_used": 567156, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.4017038345336914, + "objective/train/theoretical_loss": 3.533794941371163, + "objective/train/tokens_used": 1573663200, + "theoretical_loss": 3.533794941371163, + "tokens_seen": 1553203200 + }, + { + "epoch": 0.47, + "learning_rate": 0.000427827482447342, + "loss": 1.3679, + "theoretical_loss": 3.533794941371163, + "tokens_seen": 1553203200 + }, + { + "epoch": 0.47, + "learning_rate": 0.00042781143430290874, + "loss": 1.4031, + "theoretical_loss": 3.5337819881212944, + "tokens_seen": 1553268736 + }, + { + "epoch": 0.47, + "learning_rate": 0.0004277953861584754, + "loss": 1.2753, + "theoretical_loss": 3.533769035570962, + "tokens_seen": 1553334272 + }, + { + "epoch": 0.47, + "learning_rate": 0.00042777933801404214, + "loss": 1.3556, + "theoretical_loss": 3.5337560837200988, + "tokens_seen": 1553399808 + }, + { + "epoch": 0.47, + "learning_rate": 0.0004277632898696088, + "loss": 1.314, + "theoretical_loss": 3.5337431325686377, + "tokens_seen": 1553465344 + }, + { + "epoch": 0.47, + "learning_rate": 0.00042774724172517554, + "loss": 1.4593, + "theoretical_loss": 3.5337301821165106, + "tokens_seen": 1553530880 + }, + { + "epoch": 0.47, + "learning_rate": 0.0004277311935807422, + "loss": 1.4527, + "theoretical_loss": 3.533717232363651, + "tokens_seen": 1553596416 + }, + { + "epoch": 0.47, + "learning_rate": 0.00042771514543630894, + "loss": 1.4584, + "theoretical_loss": 3.533704283309992, + "tokens_seen": 1553661952 + }, + { + "epoch": 0.47, + "learning_rate": 0.0004276990972918756, + "loss": 1.2459, + "theoretical_loss": 3.5336913349554653, + "tokens_seen": 1553727488 + }, + { + "epoch": 0.47, + "learning_rate": 0.00042768304914744235, + "loss": 1.3636, + "theoretical_loss": 3.533678387300004, + "tokens_seen": 1553793024 + }, + { + "epoch": 0.47, + "learning_rate": 0.000427667001003009, + "loss": 1.3766, + "theoretical_loss": 3.5336654403435417, + "tokens_seen": 1553858560 + }, + { + "epoch": 0.47, + "learning_rate": 0.00042765095285857575, + "loss": 1.3535, + "theoretical_loss": 3.5336524940860103, + "tokens_seen": 1553924096 + }, + { + "epoch": 0.47, + "learning_rate": 0.0004276349047141424, + "loss": 1.3787, + "theoretical_loss": 3.533639548527342, + "tokens_seen": 1553989632 + }, + { + "epoch": 0.47, + "learning_rate": 0.00042761885656970915, + "loss": 1.4062, + "theoretical_loss": 3.5336266036674715, + "tokens_seen": 1554055168 + }, + { + "epoch": 0.47, + "learning_rate": 0.0004276028084252759, + "loss": 1.291, + "theoretical_loss": 3.53361365950633, + "tokens_seen": 1554120704 + }, + { + "epoch": 0.47, + "learning_rate": 0.00042758676028084255, + "loss": 1.2487, + "theoretical_loss": 3.5336007160438507, + "tokens_seen": 1554186240 + }, + { + "epoch": 0.47, + "learning_rate": 0.0004275707121364093, + "loss": 1.3157, + "theoretical_loss": 3.533587773279967, + "tokens_seen": 1554251776 + }, + { + "epoch": 0.47, + "learning_rate": 0.00042755466399197595, + "loss": 1.4069, + "theoretical_loss": 3.533574831214611, + "tokens_seen": 1554317312 + }, + { + "epoch": 0.47, + "learning_rate": 0.0004275386158475427, + "loss": 1.3448, + "theoretical_loss": 3.533561889847716, + "tokens_seen": 1554382848 + }, + { + "epoch": 0.47, + "learning_rate": 0.00042752256770310935, + "loss": 1.3437, + "theoretical_loss": 3.533548949179215, + "tokens_seen": 1554448384 + }, + { + "epoch": 0.47, + "learning_rate": 0.0004275065195586761, + "loss": 1.4558, + "theoretical_loss": 3.5335360092090404, + "tokens_seen": 1554513920 + }, + { + "epoch": 0.47, + "learning_rate": 0.00042749047141424275, + "loss": 1.4684, + "theoretical_loss": 3.533523069937125, + "tokens_seen": 1554579456 + }, + { + "epoch": 0.47, + "learning_rate": 0.0004274744232698095, + "loss": 1.4861, + "theoretical_loss": 3.5335101313634025, + "tokens_seen": 1554644992 + }, + { + "epoch": 0.47, + "learning_rate": 0.00042745837512537616, + "loss": 1.3797, + "theoretical_loss": 3.533497193487805, + "tokens_seen": 1554710528 + }, + { + "epoch": 0.47, + "learning_rate": 0.0004274423269809429, + "loss": 1.3864, + "theoretical_loss": 3.5334842563102655, + "tokens_seen": 1554776064 + }, + { + "epoch": 0.47, + "learning_rate": 0.00042742627883650956, + "loss": 1.4861, + "theoretical_loss": 3.533471319830717, + "tokens_seen": 1554841600 + }, + { + "epoch": 0.47, + "learning_rate": 0.0004274102306920763, + "loss": 1.4187, + "theoretical_loss": 3.533458384049093, + "tokens_seen": 1554907136 + }, + { + "epoch": 0.47, + "learning_rate": 0.00042739418254764296, + "loss": 1.2727, + "theoretical_loss": 3.5334454489653258, + "tokens_seen": 1554972672 + }, + { + "epoch": 0.47, + "learning_rate": 0.00042737813440320966, + "loss": 1.4536, + "theoretical_loss": 3.5334325145793484, + "tokens_seen": 1555038208 + }, + { + "epoch": 0.47, + "learning_rate": 0.00042736208625877636, + "loss": 1.4189, + "theoretical_loss": 3.5334195808910938, + "tokens_seen": 1555103744 + }, + { + "epoch": 0.47, + "learning_rate": 0.00042734603811434306, + "loss": 1.2835, + "theoretical_loss": 3.533406647900495, + "tokens_seen": 1555169280 + }, + { + "epoch": 0.47, + "learning_rate": 0.00042732998996990976, + "loss": 1.2706, + "theoretical_loss": 3.5333937156074846, + "tokens_seen": 1555234816 + }, + { + "epoch": 0.47, + "learning_rate": 0.00042731394182547646, + "loss": 1.4826, + "theoretical_loss": 3.533380784011997, + "tokens_seen": 1555300352 + }, + { + "epoch": 0.47, + "learning_rate": 0.00042729789368104316, + "loss": 1.4088, + "theoretical_loss": 3.5333678531139627, + "tokens_seen": 1555365888 + }, + { + "epoch": 0.47, + "learning_rate": 0.00042728184553660986, + "loss": 1.3721, + "theoretical_loss": 3.5333549229133174, + "tokens_seen": 1555431424 + }, + { + "epoch": 0.47, + "learning_rate": 0.00042726579739217656, + "loss": 1.3702, + "theoretical_loss": 3.5333419934099926, + "tokens_seen": 1555496960 + }, + { + "epoch": 0.47, + "learning_rate": 0.00042724974924774326, + "loss": 1.4792, + "theoretical_loss": 3.533329064603921, + "tokens_seen": 1555562496 + }, + { + "epoch": 0.47, + "learning_rate": 0.0004272337011033099, + "loss": 1.4798, + "theoretical_loss": 3.5333161364950367, + "tokens_seen": 1555628032 + }, + { + "epoch": 0.47, + "learning_rate": 0.00042721765295887666, + "loss": 1.4256, + "theoretical_loss": 3.5333032090832717, + "tokens_seen": 1555693568 + }, + { + "epoch": 0.47, + "learning_rate": 0.0004272016048144433, + "loss": 1.3938, + "theoretical_loss": 3.53329028236856, + "tokens_seen": 1555759104 + }, + { + "epoch": 0.47, + "learning_rate": 0.00042718555667001007, + "loss": 1.3216, + "theoretical_loss": 3.533277356350834, + "tokens_seen": 1555824640 + }, + { + "epoch": 0.47, + "learning_rate": 0.0004271695085255767, + "loss": 1.1726, + "theoretical_loss": 3.5332644310300276, + "tokens_seen": 1555890176 + }, + { + "epoch": 0.47, + "learning_rate": 0.00042715346038114347, + "loss": 1.3469, + "theoretical_loss": 3.5332515064060726, + "tokens_seen": 1555955712 + }, + { + "epoch": 0.47, + "learning_rate": 0.0004271374122367101, + "loss": 1.3351, + "theoretical_loss": 3.5332385824789028, + "tokens_seen": 1556021248 + }, + { + "epoch": 0.47, + "learning_rate": 0.00042712136409227687, + "loss": 1.4406, + "theoretical_loss": 3.5332256592484512, + "tokens_seen": 1556086784 + }, + { + "epoch": 0.47, + "learning_rate": 0.0004271053159478435, + "loss": 1.6201, + "theoretical_loss": 3.5332127367146513, + "tokens_seen": 1556152320 + }, + { + "epoch": 0.47, + "learning_rate": 0.00042708926780341027, + "loss": 1.4705, + "theoretical_loss": 3.5331998148774355, + "tokens_seen": 1556217856 + }, + { + "epoch": 0.47, + "learning_rate": 0.0004270732196589769, + "loss": 1.2579, + "theoretical_loss": 3.533186893736737, + "tokens_seen": 1556283392 + }, + { + "epoch": 0.47, + "learning_rate": 0.00042705717151454367, + "loss": 1.3898, + "theoretical_loss": 3.53317397329249, + "tokens_seen": 1556348928 + }, + { + "epoch": 0.47, + "learning_rate": 0.0004270411233701103, + "loss": 1.2926, + "theoretical_loss": 3.5331610535446263, + "tokens_seen": 1556414464 + }, + { + "epoch": 0.47, + "objective/train/docs_used": 568457, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.2760186195373535, + "objective/train/theoretical_loss": 3.5331481344930795, + "objective/train/tokens_used": 1576940000, + "theoretical_loss": 3.5331481344930795, + "tokens_seen": 1556480000 + }, + { + "epoch": 0.47, + "learning_rate": 0.00042702507522567707, + "loss": 1.3219, + "theoretical_loss": 3.5331481344930795, + "tokens_seen": 1556480000 + }, + { + "epoch": 0.47, + "learning_rate": 0.0004270090270812437, + "loss": 1.3528, + "theoretical_loss": 3.533135216137783, + "tokens_seen": 1556545536 + }, + { + "epoch": 0.47, + "learning_rate": 0.0004269929789368105, + "loss": 1.3626, + "theoretical_loss": 3.53312229847867, + "tokens_seen": 1556611072 + }, + { + "epoch": 0.47, + "learning_rate": 0.0004269769307923771, + "loss": 1.3756, + "theoretical_loss": 3.533109381515673, + "tokens_seen": 1556676608 + }, + { + "epoch": 0.47, + "learning_rate": 0.0004269608826479439, + "loss": 1.3388, + "theoretical_loss": 3.5330964652487253, + "tokens_seen": 1556742144 + }, + { + "epoch": 0.47, + "learning_rate": 0.0004269448345035105, + "loss": 1.5567, + "theoretical_loss": 3.5330835496777615, + "tokens_seen": 1556807680 + }, + { + "epoch": 0.47, + "learning_rate": 0.0004269287863590773, + "loss": 1.4444, + "theoretical_loss": 3.533070634802713, + "tokens_seen": 1556873216 + }, + { + "epoch": 0.47, + "learning_rate": 0.0004269127382146439, + "loss": 1.2869, + "theoretical_loss": 3.533057720623514, + "tokens_seen": 1556938752 + }, + { + "epoch": 0.47, + "learning_rate": 0.0004268966900702107, + "loss": 1.428, + "theoretical_loss": 3.5330448071400973, + "tokens_seen": 1557004288 + }, + { + "epoch": 0.47, + "learning_rate": 0.0004268806419257773, + "loss": 1.3222, + "theoretical_loss": 3.5330318943523964, + "tokens_seen": 1557069824 + }, + { + "epoch": 0.47, + "learning_rate": 0.0004268645937813441, + "loss": 1.3397, + "theoretical_loss": 3.5330189822603444, + "tokens_seen": 1557135360 + }, + { + "epoch": 0.47, + "learning_rate": 0.0004268485456369107, + "loss": 1.3914, + "theoretical_loss": 3.5330060708638746, + "tokens_seen": 1557200896 + }, + { + "epoch": 0.47, + "learning_rate": 0.0004268324974924775, + "loss": 1.3781, + "theoretical_loss": 3.5329931601629205, + "tokens_seen": 1557266432 + }, + { + "epoch": 0.47, + "learning_rate": 0.0004268164493480441, + "loss": 1.4712, + "theoretical_loss": 3.5329802501574146, + "tokens_seen": 1557331968 + }, + { + "epoch": 0.47, + "learning_rate": 0.0004268004012036109, + "loss": 1.4007, + "theoretical_loss": 3.532967340847291, + "tokens_seen": 1557397504 + }, + { + "epoch": 0.47, + "learning_rate": 0.00042678435305917753, + "loss": 1.313, + "theoretical_loss": 3.532954432232483, + "tokens_seen": 1557463040 + }, + { + "epoch": 0.47, + "learning_rate": 0.0004267683049147443, + "loss": 1.3962, + "theoretical_loss": 3.532941524312923, + "tokens_seen": 1557528576 + }, + { + "epoch": 0.47, + "learning_rate": 0.00042675225677031093, + "loss": 1.2128, + "theoretical_loss": 3.5329286170885448, + "tokens_seen": 1557594112 + }, + { + "epoch": 0.47, + "learning_rate": 0.0004267362086258777, + "loss": 1.3323, + "theoretical_loss": 3.532915710559282, + "tokens_seen": 1557659648 + }, + { + "epoch": 0.47, + "learning_rate": 0.00042672016048144433, + "loss": 1.3405, + "theoretical_loss": 3.532902804725068, + "tokens_seen": 1557725184 + }, + { + "epoch": 0.47, + "learning_rate": 0.0004267041123370111, + "loss": 1.3531, + "theoretical_loss": 3.532889899585835, + "tokens_seen": 1557790720 + }, + { + "epoch": 0.47, + "learning_rate": 0.00042668806419257773, + "loss": 1.4422, + "theoretical_loss": 3.532876995141518, + "tokens_seen": 1557856256 + }, + { + "epoch": 0.47, + "learning_rate": 0.0004266720160481445, + "loss": 1.3502, + "theoretical_loss": 3.5328640913920495, + "tokens_seen": 1557921792 + }, + { + "epoch": 0.47, + "learning_rate": 0.00042665596790371113, + "loss": 1.3793, + "theoretical_loss": 3.5328511883373626, + "tokens_seen": 1557987328 + }, + { + "epoch": 0.47, + "learning_rate": 0.0004266399197592779, + "loss": 1.4434, + "theoretical_loss": 3.532838285977391, + "tokens_seen": 1558052864 + }, + { + "epoch": 0.47, + "learning_rate": 0.00042662387161484453, + "loss": 1.2669, + "theoretical_loss": 3.5328253843120674, + "tokens_seen": 1558118400 + }, + { + "epoch": 0.47, + "learning_rate": 0.00042660782347041123, + "loss": 1.3773, + "theoretical_loss": 3.532812483341327, + "tokens_seen": 1558183936 + }, + { + "epoch": 0.47, + "learning_rate": 0.00042659177532597794, + "loss": 1.2821, + "theoretical_loss": 3.532799583065101, + "tokens_seen": 1558249472 + }, + { + "epoch": 0.47, + "learning_rate": 0.00042657572718154464, + "loss": 1.4856, + "theoretical_loss": 3.532786683483324, + "tokens_seen": 1558315008 + }, + { + "epoch": 0.47, + "learning_rate": 0.00042655967903711134, + "loss": 1.418, + "theoretical_loss": 3.5327737845959293, + "tokens_seen": 1558380544 + }, + { + "epoch": 0.47, + "learning_rate": 0.00042654363089267804, + "loss": 1.462, + "theoretical_loss": 3.5327608864028504, + "tokens_seen": 1558446080 + }, + { + "epoch": 0.47, + "learning_rate": 0.00042652758274824474, + "loss": 1.3871, + "theoretical_loss": 3.5327479889040205, + "tokens_seen": 1558511616 + }, + { + "epoch": 0.47, + "learning_rate": 0.00042651153460381144, + "loss": 1.3059, + "theoretical_loss": 3.532735092099373, + "tokens_seen": 1558577152 + }, + { + "epoch": 0.47, + "learning_rate": 0.00042649548645937814, + "loss": 1.3851, + "theoretical_loss": 3.5327221959888417, + "tokens_seen": 1558642688 + }, + { + "epoch": 0.47, + "learning_rate": 0.00042647943831494484, + "loss": 1.3387, + "theoretical_loss": 3.5327093005723595, + "tokens_seen": 1558708224 + }, + { + "epoch": 0.47, + "learning_rate": 0.0004264633901705115, + "loss": 1.2474, + "theoretical_loss": 3.53269640584986, + "tokens_seen": 1558773760 + }, + { + "epoch": 0.47, + "learning_rate": 0.00042644734202607824, + "loss": 1.4935, + "theoretical_loss": 3.5326835118212774, + "tokens_seen": 1558839296 + }, + { + "epoch": 0.47, + "learning_rate": 0.0004264312938816449, + "loss": 1.4707, + "theoretical_loss": 3.532670618486544, + "tokens_seen": 1558904832 + }, + { + "epoch": 0.47, + "learning_rate": 0.00042641524573721164, + "loss": 1.3001, + "theoretical_loss": 3.5326577258455947, + "tokens_seen": 1558970368 + }, + { + "epoch": 0.47, + "learning_rate": 0.0004263991975927784, + "loss": 1.3816, + "theoretical_loss": 3.5326448338983623, + "tokens_seen": 1559035904 + }, + { + "epoch": 0.47, + "learning_rate": 0.00042638314944834504, + "loss": 1.3196, + "theoretical_loss": 3.5326319426447794, + "tokens_seen": 1559101440 + }, + { + "epoch": 0.47, + "learning_rate": 0.0004263671013039118, + "loss": 1.4278, + "theoretical_loss": 3.5326190520847813, + "tokens_seen": 1559166976 + }, + { + "epoch": 0.47, + "learning_rate": 0.00042635105315947844, + "loss": 1.3629, + "theoretical_loss": 3.5326061622183, + "tokens_seen": 1559232512 + }, + { + "epoch": 0.47, + "learning_rate": 0.0004263350050150452, + "loss": 1.3462, + "theoretical_loss": 3.53259327304527, + "tokens_seen": 1559298048 + }, + { + "epoch": 0.47, + "learning_rate": 0.00042631895687061185, + "loss": 1.3924, + "theoretical_loss": 3.532580384565624, + "tokens_seen": 1559363584 + }, + { + "epoch": 0.47, + "learning_rate": 0.0004263029087261786, + "loss": 1.341, + "theoretical_loss": 3.5325674967792966, + "tokens_seen": 1559429120 + }, + { + "epoch": 0.47, + "learning_rate": 0.00042628686058174525, + "loss": 1.3181, + "theoretical_loss": 3.532554609686221, + "tokens_seen": 1559494656 + }, + { + "epoch": 0.47, + "learning_rate": 0.000426270812437312, + "loss": 1.4666, + "theoretical_loss": 3.53254172328633, + "tokens_seen": 1559560192 + }, + { + "epoch": 0.47, + "learning_rate": 0.00042625476429287865, + "loss": 1.4944, + "theoretical_loss": 3.5325288375795583, + "tokens_seen": 1559625728 + }, + { + "epoch": 0.47, + "learning_rate": 0.0004262387161484454, + "loss": 1.2748, + "theoretical_loss": 3.5325159525658387, + "tokens_seen": 1559691264 + }, + { + "epoch": 0.47, + "objective/train/docs_used": 569616, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.3957040309906006, + "objective/train/theoretical_loss": 3.5325030682451053, + "objective/train/tokens_used": 1580216800, + "theoretical_loss": 3.5325030682451053, + "tokens_seen": 1559756800 + }, + { + "epoch": 0.47, + "learning_rate": 0.00042622266800401205, + "loss": 1.4107, + "theoretical_loss": 3.5325030682451053, + "tokens_seen": 1559756800 + }, + { + "epoch": 0.47, + "learning_rate": 0.0004262066198595788, + "loss": 1.5, + "theoretical_loss": 3.5324901846172914, + "tokens_seen": 1559822336 + }, + { + "epoch": 0.47, + "learning_rate": 0.00042619057171514545, + "loss": 1.4219, + "theoretical_loss": 3.5324773016823308, + "tokens_seen": 1559887872 + }, + { + "epoch": 0.47, + "learning_rate": 0.0004261745235707122, + "loss": 1.4093, + "theoretical_loss": 3.5324644194401573, + "tokens_seen": 1559953408 + }, + { + "epoch": 0.47, + "learning_rate": 0.00042615847542627885, + "loss": 1.3843, + "theoretical_loss": 3.532451537890704, + "tokens_seen": 1560018944 + }, + { + "epoch": 0.47, + "learning_rate": 0.0004261424272818456, + "loss": 1.3351, + "theoretical_loss": 3.532438657033905, + "tokens_seen": 1560084480 + }, + { + "epoch": 0.47, + "learning_rate": 0.00042612637913741225, + "loss": 1.3254, + "theoretical_loss": 3.532425776869694, + "tokens_seen": 1560150016 + }, + { + "epoch": 0.47, + "learning_rate": 0.000426110330992979, + "loss": 1.4478, + "theoretical_loss": 3.532412897398004, + "tokens_seen": 1560215552 + }, + { + "epoch": 0.47, + "learning_rate": 0.00042609428284854565, + "loss": 1.3299, + "theoretical_loss": 3.5324000186187696, + "tokens_seen": 1560281088 + }, + { + "epoch": 0.47, + "learning_rate": 0.0004260782347041124, + "loss": 1.4593, + "theoretical_loss": 3.532387140531924, + "tokens_seen": 1560346624 + }, + { + "epoch": 0.47, + "learning_rate": 0.00042606218655967906, + "loss": 1.3411, + "theoretical_loss": 3.5323742631374007, + "tokens_seen": 1560412160 + }, + { + "epoch": 0.47, + "learning_rate": 0.0004260461384152458, + "loss": 1.3857, + "theoretical_loss": 3.5323613864351344, + "tokens_seen": 1560477696 + }, + { + "epoch": 0.47, + "learning_rate": 0.00042603009027081246, + "loss": 1.4804, + "theoretical_loss": 3.5323485104250576, + "tokens_seen": 1560543232 + }, + { + "epoch": 0.47, + "learning_rate": 0.0004260140421263792, + "loss": 1.4274, + "theoretical_loss": 3.5323356351071045, + "tokens_seen": 1560608768 + }, + { + "epoch": 0.47, + "learning_rate": 0.00042599799398194586, + "loss": 1.4372, + "theoretical_loss": 3.532322760481209, + "tokens_seen": 1560674304 + }, + { + "epoch": 0.47, + "learning_rate": 0.00042598194583751256, + "loss": 1.3021, + "theoretical_loss": 3.5323098865473046, + "tokens_seen": 1560739840 + }, + { + "epoch": 0.47, + "learning_rate": 0.00042596589769307926, + "loss": 1.4468, + "theoretical_loss": 3.532297013305325, + "tokens_seen": 1560805376 + }, + { + "epoch": 0.47, + "learning_rate": 0.00042594984954864596, + "loss": 1.337, + "theoretical_loss": 3.532284140755204, + "tokens_seen": 1560870912 + }, + { + "epoch": 0.47, + "learning_rate": 0.00042593380140421266, + "loss": 1.3496, + "theoretical_loss": 3.532271268896876, + "tokens_seen": 1560936448 + }, + { + "epoch": 0.47, + "learning_rate": 0.00042591775325977936, + "loss": 1.2724, + "theoretical_loss": 3.532258397730274, + "tokens_seen": 1561001984 + }, + { + "epoch": 0.47, + "learning_rate": 0.00042590170511534606, + "loss": 1.3163, + "theoretical_loss": 3.532245527255332, + "tokens_seen": 1561067520 + }, + { + "epoch": 0.47, + "learning_rate": 0.00042588565697091276, + "loss": 1.4081, + "theoretical_loss": 3.532232657471984, + "tokens_seen": 1561133056 + }, + { + "epoch": 0.47, + "learning_rate": 0.00042586960882647946, + "loss": 1.3375, + "theoretical_loss": 3.5322197883801634, + "tokens_seen": 1561198592 + }, + { + "epoch": 0.47, + "learning_rate": 0.00042585356068204616, + "loss": 1.4099, + "theoretical_loss": 3.5322069199798043, + "tokens_seen": 1561264128 + }, + { + "epoch": 0.47, + "learning_rate": 0.0004258375125376128, + "loss": 1.3569, + "theoretical_loss": 3.5321940522708406, + "tokens_seen": 1561329664 + }, + { + "epoch": 0.47, + "learning_rate": 0.00042582146439317957, + "loss": 1.4573, + "theoretical_loss": 3.532181185253206, + "tokens_seen": 1561395200 + }, + { + "epoch": 0.47, + "learning_rate": 0.0004258054162487462, + "loss": 1.479, + "theoretical_loss": 3.5321683189268342, + "tokens_seen": 1561460736 + }, + { + "epoch": 0.47, + "learning_rate": 0.00042578936810431297, + "loss": 1.4395, + "theoretical_loss": 3.5321554532916597, + "tokens_seen": 1561526272 + }, + { + "epoch": 0.47, + "learning_rate": 0.0004257733199598796, + "loss": 1.4371, + "theoretical_loss": 3.5321425883476154, + "tokens_seen": 1561591808 + }, + { + "epoch": 0.47, + "learning_rate": 0.00042575727181544637, + "loss": 1.463, + "theoretical_loss": 3.5321297240946357, + "tokens_seen": 1561657344 + }, + { + "epoch": 0.47, + "learning_rate": 0.000425741223671013, + "loss": 1.3372, + "theoretical_loss": 3.5321168605326543, + "tokens_seen": 1561722880 + }, + { + "epoch": 0.47, + "learning_rate": 0.00042572517552657977, + "loss": 1.3078, + "theoretical_loss": 3.532103997661606, + "tokens_seen": 1561788416 + }, + { + "epoch": 0.47, + "learning_rate": 0.0004257091273821464, + "loss": 1.3893, + "theoretical_loss": 3.532091135481423, + "tokens_seen": 1561853952 + }, + { + "epoch": 0.47, + "learning_rate": 0.00042569307923771317, + "loss": 1.4772, + "theoretical_loss": 3.5320782739920404, + "tokens_seen": 1561919488 + }, + { + "epoch": 0.47, + "learning_rate": 0.0004256770310932798, + "loss": 1.4811, + "theoretical_loss": 3.5320654131933917, + "tokens_seen": 1561985024 + }, + { + "epoch": 0.47, + "learning_rate": 0.00042566098294884657, + "loss": 1.4917, + "theoretical_loss": 3.5320525530854114, + "tokens_seen": 1562050560 + }, + { + "epoch": 0.47, + "learning_rate": 0.0004256449348044132, + "loss": 1.4523, + "theoretical_loss": 3.532039693668033, + "tokens_seen": 1562116096 + }, + { + "epoch": 0.47, + "learning_rate": 0.00042562888665998, + "loss": 1.3662, + "theoretical_loss": 3.53202683494119, + "tokens_seen": 1562181632 + }, + { + "epoch": 0.47, + "learning_rate": 0.0004256128385155466, + "loss": 1.3622, + "theoretical_loss": 3.532013976904817, + "tokens_seen": 1562247168 + }, + { + "epoch": 0.47, + "learning_rate": 0.0004255967903711134, + "loss": 1.4495, + "theoretical_loss": 3.5320011195588483, + "tokens_seen": 1562312704 + }, + { + "epoch": 0.47, + "learning_rate": 0.00042558074222668, + "loss": 1.4068, + "theoretical_loss": 3.5319882629032167, + "tokens_seen": 1562378240 + }, + { + "epoch": 0.47, + "learning_rate": 0.0004255646940822468, + "loss": 1.3388, + "theoretical_loss": 3.531975406937857, + "tokens_seen": 1562443776 + }, + { + "epoch": 0.47, + "learning_rate": 0.0004255486459378134, + "loss": 1.3524, + "theoretical_loss": 3.5319625516627027, + "tokens_seen": 1562509312 + }, + { + "epoch": 0.47, + "learning_rate": 0.0004255325977933802, + "loss": 1.312, + "theoretical_loss": 3.531949697077688, + "tokens_seen": 1562574848 + }, + { + "epoch": 0.47, + "learning_rate": 0.0004255165496489468, + "loss": 1.3838, + "theoretical_loss": 3.5319368431827476, + "tokens_seen": 1562640384 + }, + { + "epoch": 0.47, + "learning_rate": 0.0004255005015045136, + "loss": 1.3709, + "theoretical_loss": 3.5319239899778143, + "tokens_seen": 1562705920 + }, + { + "epoch": 0.47, + "learning_rate": 0.0004254844533600802, + "loss": 1.4033, + "theoretical_loss": 3.5319111374628234, + "tokens_seen": 1562771456 + }, + { + "epoch": 0.47, + "learning_rate": 0.000425468405215647, + "loss": 1.4357, + "theoretical_loss": 3.5318982856377077, + "tokens_seen": 1562836992 + }, + { + "epoch": 0.47, + "learning_rate": 0.0004254523570712136, + "loss": 1.3278, + "theoretical_loss": 3.531885434502402, + "tokens_seen": 1562902528 + }, + { + "epoch": 0.47, + "learning_rate": 0.0004254363089267804, + "loss": 1.4089, + "theoretical_loss": 3.53187258405684, + "tokens_seen": 1562968064 + }, + { + "epoch": 0.47, + "objective/train/docs_used": 570864, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.1853934526443481, + "objective/train/theoretical_loss": 3.5318597343009563, + "objective/train/tokens_used": 1583493600, + "theoretical_loss": 3.5318597343009563, + "tokens_seen": 1563033600 + }, + { + "epoch": 0.47, + "learning_rate": 0.00042542026078234703, + "loss": 1.4101, + "theoretical_loss": 3.5318597343009563, + "tokens_seen": 1563033600 + }, + { + "epoch": 0.47, + "learning_rate": 0.0004254042126379138, + "loss": 1.369, + "theoretical_loss": 3.5318468852346845, + "tokens_seen": 1563099136 + }, + { + "epoch": 0.47, + "learning_rate": 0.00042538816449348043, + "loss": 1.5258, + "theoretical_loss": 3.5318340368579584, + "tokens_seen": 1563164672 + }, + { + "epoch": 0.47, + "learning_rate": 0.0004253721163490472, + "loss": 1.362, + "theoretical_loss": 3.531821189170713, + "tokens_seen": 1563230208 + }, + { + "epoch": 0.47, + "learning_rate": 0.00042535606820461383, + "loss": 1.2552, + "theoretical_loss": 3.531808342172882, + "tokens_seen": 1563295744 + }, + { + "epoch": 0.47, + "learning_rate": 0.0004253400200601806, + "loss": 1.3384, + "theoretical_loss": 3.5317954958643987, + "tokens_seen": 1563361280 + }, + { + "epoch": 0.47, + "learning_rate": 0.00042532397191574723, + "loss": 1.375, + "theoretical_loss": 3.531782650245198, + "tokens_seen": 1563426816 + }, + { + "epoch": 0.47, + "learning_rate": 0.000425307923771314, + "loss": 1.2078, + "theoretical_loss": 3.5317698053152142, + "tokens_seen": 1563492352 + }, + { + "epoch": 0.47, + "learning_rate": 0.00042529187562688063, + "loss": 1.4517, + "theoretical_loss": 3.5317569610743815, + "tokens_seen": 1563557888 + }, + { + "epoch": 0.47, + "learning_rate": 0.0004252758274824474, + "loss": 1.3667, + "theoretical_loss": 3.531744117522633, + "tokens_seen": 1563623424 + }, + { + "epoch": 0.47, + "learning_rate": 0.00042525977933801403, + "loss": 1.3954, + "theoretical_loss": 3.5317312746599034, + "tokens_seen": 1563688960 + }, + { + "epoch": 0.47, + "learning_rate": 0.00042525977933801403, + "loss": 1.3877, + "theoretical_loss": 3.5317184324861275, + "tokens_seen": 1563754496 + }, + { + "epoch": 0.47, + "learning_rate": 0.0004252437311935808, + "loss": 1.3828, + "theoretical_loss": 3.531705591001239, + "tokens_seen": 1563820032 + }, + { + "epoch": 0.47, + "learning_rate": 0.00042522768304914743, + "loss": 1.346, + "theoretical_loss": 3.531692750205172, + "tokens_seen": 1563885568 + }, + { + "epoch": 0.47, + "learning_rate": 0.00042521163490471414, + "loss": 1.2904, + "theoretical_loss": 3.5316799100978606, + "tokens_seen": 1563951104 + }, + { + "epoch": 0.47, + "learning_rate": 0.0004251955867602809, + "loss": 1.3309, + "theoretical_loss": 3.531667070679239, + "tokens_seen": 1564016640 + }, + { + "epoch": 0.47, + "learning_rate": 0.00042517953861584754, + "loss": 1.3704, + "theoretical_loss": 3.5316542319492417, + "tokens_seen": 1564082176 + }, + { + "epoch": 0.47, + "learning_rate": 0.0004251634904714143, + "loss": 1.35, + "theoretical_loss": 3.531641393907803, + "tokens_seen": 1564147712 + }, + { + "epoch": 0.47, + "learning_rate": 0.00042514744232698094, + "loss": 1.3449, + "theoretical_loss": 3.5316285565548564, + "tokens_seen": 1564213248 + }, + { + "epoch": 0.47, + "learning_rate": 0.0004251313941825477, + "loss": 1.416, + "theoretical_loss": 3.531615719890337, + "tokens_seen": 1564278784 + }, + { + "epoch": 0.47, + "learning_rate": 0.00042511534603811434, + "loss": 1.3495, + "theoretical_loss": 3.5316028839141786, + "tokens_seen": 1564344320 + }, + { + "epoch": 0.47, + "learning_rate": 0.0004250992978936811, + "loss": 1.3196, + "theoretical_loss": 3.5315900486263154, + "tokens_seen": 1564409856 + }, + { + "epoch": 0.47, + "learning_rate": 0.00042508324974924774, + "loss": 1.3652, + "theoretical_loss": 3.5315772140266817, + "tokens_seen": 1564475392 + }, + { + "epoch": 0.47, + "learning_rate": 0.0004250672016048145, + "loss": 1.2901, + "theoretical_loss": 3.5315643801152117, + "tokens_seen": 1564540928 + }, + { + "epoch": 0.47, + "learning_rate": 0.00042505115346038114, + "loss": 1.3805, + "theoretical_loss": 3.5315515468918406, + "tokens_seen": 1564606464 + }, + { + "epoch": 0.47, + "learning_rate": 0.0004250351053159479, + "loss": 1.4387, + "theoretical_loss": 3.5315387143565014, + "tokens_seen": 1564672000 + }, + { + "epoch": 0.47, + "learning_rate": 0.00042501905717151454, + "loss": 1.2799, + "theoretical_loss": 3.5315258825091282, + "tokens_seen": 1564737536 + }, + { + "epoch": 0.47, + "learning_rate": 0.0004250030090270813, + "loss": 1.3594, + "theoretical_loss": 3.531513051349657, + "tokens_seen": 1564803072 + }, + { + "epoch": 0.47, + "learning_rate": 0.00042498696088264794, + "loss": 1.3228, + "theoretical_loss": 3.531500220878021, + "tokens_seen": 1564868608 + }, + { + "epoch": 0.47, + "learning_rate": 0.0004249709127382147, + "loss": 1.3569, + "theoretical_loss": 3.531487391094154, + "tokens_seen": 1564934144 + }, + { + "epoch": 0.47, + "learning_rate": 0.00042495486459378135, + "loss": 1.4329, + "theoretical_loss": 3.5314745619979915, + "tokens_seen": 1564999680 + }, + { + "epoch": 0.47, + "learning_rate": 0.0004249388164493481, + "loss": 1.434, + "theoretical_loss": 3.5314617335894676, + "tokens_seen": 1565065216 + }, + { + "epoch": 0.47, + "learning_rate": 0.00042492276830491475, + "loss": 1.4883, + "theoretical_loss": 3.5314489058685155, + "tokens_seen": 1565130752 + }, + { + "epoch": 0.47, + "learning_rate": 0.0004249067201604815, + "loss": 1.3412, + "theoretical_loss": 3.531436078835071, + "tokens_seen": 1565196288 + }, + { + "epoch": 0.47, + "learning_rate": 0.00042489067201604815, + "loss": 1.4675, + "theoretical_loss": 3.531423252489068, + "tokens_seen": 1565261824 + }, + { + "epoch": 0.47, + "learning_rate": 0.0004248746238716149, + "loss": 1.4679, + "theoretical_loss": 3.5314104268304405, + "tokens_seen": 1565327360 + }, + { + "epoch": 0.47, + "learning_rate": 0.00042485857572718155, + "loss": 1.3707, + "theoretical_loss": 3.531397601859123, + "tokens_seen": 1565392896 + }, + { + "epoch": 0.47, + "learning_rate": 0.0004248425275827483, + "loss": 1.3296, + "theoretical_loss": 3.5313847775750506, + "tokens_seen": 1565458432 + }, + { + "epoch": 0.47, + "learning_rate": 0.00042482647943831495, + "loss": 1.4315, + "theoretical_loss": 3.5313719539781565, + "tokens_seen": 1565523968 + }, + { + "epoch": 0.47, + "learning_rate": 0.0004248104312938817, + "loss": 1.5388, + "theoretical_loss": 3.5313591310683763, + "tokens_seen": 1565589504 + }, + { + "epoch": 0.47, + "learning_rate": 0.00042479438314944835, + "loss": 1.4192, + "theoretical_loss": 3.5313463088456434, + "tokens_seen": 1565655040 + }, + { + "epoch": 0.47, + "learning_rate": 0.0004247783350050151, + "loss": 1.3365, + "theoretical_loss": 3.5313334873098934, + "tokens_seen": 1565720576 + }, + { + "epoch": 0.47, + "learning_rate": 0.00042476228686058175, + "loss": 1.3137, + "theoretical_loss": 3.5313206664610597, + "tokens_seen": 1565786112 + }, + { + "epoch": 0.47, + "learning_rate": 0.0004247462387161485, + "loss": 1.3902, + "theoretical_loss": 3.531307846299077, + "tokens_seen": 1565851648 + }, + { + "epoch": 0.47, + "learning_rate": 0.00042473019057171515, + "loss": 1.4179, + "theoretical_loss": 3.53129502682388, + "tokens_seen": 1565917184 + }, + { + "epoch": 0.47, + "learning_rate": 0.0004247141424272819, + "loss": 1.3282, + "theoretical_loss": 3.531282208035403, + "tokens_seen": 1565982720 + }, + { + "epoch": 0.47, + "learning_rate": 0.00042469809428284856, + "loss": 1.4696, + "theoretical_loss": 3.5312693899335805, + "tokens_seen": 1566048256 + }, + { + "epoch": 0.47, + "learning_rate": 0.0004246820461384153, + "loss": 1.4213, + "theoretical_loss": 3.5312565725183473, + "tokens_seen": 1566113792 + }, + { + "epoch": 0.47, + "learning_rate": 0.00042466599799398196, + "loss": 1.3226, + "theoretical_loss": 3.531243755789637, + "tokens_seen": 1566179328 + }, + { + "epoch": 0.47, + "learning_rate": 0.0004246499498495487, + "loss": 1.288, + "theoretical_loss": 3.5312309397473856, + "tokens_seen": 1566244864 + }, + { + "epoch": 0.47, + "objective/train/docs_used": 572070, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.388120412826538, + "objective/train/theoretical_loss": 3.531218124391526, + "objective/train/tokens_used": 1586770400, + "theoretical_loss": 3.531218124391526, + "tokens_seen": 1566310400 + }, + { + "epoch": 0.47, + "learning_rate": 0.00042463390170511536, + "loss": 1.3338, + "theoretical_loss": 3.531218124391526, + "tokens_seen": 1566310400 + }, + { + "epoch": 0.47, + "learning_rate": 0.0004246178535606821, + "loss": 1.3455, + "theoretical_loss": 3.5312053097219938, + "tokens_seen": 1566375936 + }, + { + "epoch": 0.47, + "learning_rate": 0.00042460180541624876, + "loss": 1.3601, + "theoretical_loss": 3.531192495738723, + "tokens_seen": 1566441472 + }, + { + "epoch": 0.47, + "learning_rate": 0.0004245857572718155, + "loss": 1.378, + "theoretical_loss": 3.531179682441648, + "tokens_seen": 1566507008 + }, + { + "epoch": 0.47, + "learning_rate": 0.00042456970912738216, + "loss": 1.2598, + "theoretical_loss": 3.531166869830704, + "tokens_seen": 1566572544 + }, + { + "epoch": 0.47, + "learning_rate": 0.00042455366098294886, + "loss": 1.3443, + "theoretical_loss": 3.5311540579058254, + "tokens_seen": 1566638080 + }, + { + "epoch": 0.47, + "learning_rate": 0.00042453761283851556, + "loss": 1.3938, + "theoretical_loss": 3.5311412466669463, + "tokens_seen": 1566703616 + }, + { + "epoch": 0.47, + "learning_rate": 0.00042452156469408226, + "loss": 1.4005, + "theoretical_loss": 3.5311284361140016, + "tokens_seen": 1566769152 + }, + { + "epoch": 0.47, + "learning_rate": 0.00042450551654964896, + "loss": 1.4203, + "theoretical_loss": 3.531115626246926, + "tokens_seen": 1566834688 + }, + { + "epoch": 0.47, + "learning_rate": 0.00042448946840521566, + "loss": 1.2791, + "theoretical_loss": 3.531102817065654, + "tokens_seen": 1566900224 + }, + { + "epoch": 0.47, + "learning_rate": 0.00042447342026078236, + "loss": 1.3802, + "theoretical_loss": 3.5310900085701196, + "tokens_seen": 1566965760 + }, + { + "epoch": 0.47, + "learning_rate": 0.00042445737211634907, + "loss": 1.3435, + "theoretical_loss": 3.5310772007602584, + "tokens_seen": 1567031296 + }, + { + "epoch": 0.47, + "learning_rate": 0.0004244413239719157, + "loss": 1.3756, + "theoretical_loss": 3.5310643936360044, + "tokens_seen": 1567096832 + }, + { + "epoch": 0.47, + "learning_rate": 0.00042442527582748247, + "loss": 1.3965, + "theoretical_loss": 3.5310515871972927, + "tokens_seen": 1567162368 + }, + { + "epoch": 0.47, + "learning_rate": 0.0004244092276830491, + "loss": 1.3003, + "theoretical_loss": 3.5310387814440576, + "tokens_seen": 1567227904 + }, + { + "epoch": 0.47, + "learning_rate": 0.00042439317953861587, + "loss": 1.375, + "theoretical_loss": 3.5310259763762333, + "tokens_seen": 1567293440 + }, + { + "epoch": 0.47, + "learning_rate": 0.0004243771313941825, + "loss": 1.335, + "theoretical_loss": 3.5310131719937554, + "tokens_seen": 1567358976 + }, + { + "epoch": 0.47, + "learning_rate": 0.00042436108324974927, + "loss": 1.3645, + "theoretical_loss": 3.531000368296558, + "tokens_seen": 1567424512 + }, + { + "epoch": 0.47, + "learning_rate": 0.0004243450351053159, + "loss": 1.3327, + "theoretical_loss": 3.5309875652845757, + "tokens_seen": 1567490048 + }, + { + "epoch": 0.48, + "learning_rate": 0.00042432898696088267, + "loss": 1.4846, + "theoretical_loss": 3.5309747629577437, + "tokens_seen": 1567555584 + }, + { + "epoch": 0.48, + "learning_rate": 0.0004243129388164493, + "loss": 1.3241, + "theoretical_loss": 3.530961961315996, + "tokens_seen": 1567621120 + }, + { + "epoch": 0.48, + "learning_rate": 0.00042429689067201607, + "loss": 1.3955, + "theoretical_loss": 3.5309491603592678, + "tokens_seen": 1567686656 + }, + { + "epoch": 0.48, + "learning_rate": 0.0004242808425275827, + "loss": 1.3156, + "theoretical_loss": 3.5309363600874937, + "tokens_seen": 1567752192 + }, + { + "epoch": 0.48, + "learning_rate": 0.00042426479438314947, + "loss": 1.3881, + "theoretical_loss": 3.530923560500608, + "tokens_seen": 1567817728 + }, + { + "epoch": 0.48, + "learning_rate": 0.0004242487462387161, + "loss": 1.413, + "theoretical_loss": 3.5309107615985464, + "tokens_seen": 1567883264 + }, + { + "epoch": 0.48, + "learning_rate": 0.0004242326980942829, + "loss": 1.3446, + "theoretical_loss": 3.5308979633812427, + "tokens_seen": 1567948800 + }, + { + "epoch": 0.48, + "learning_rate": 0.0004242166499498495, + "loss": 1.3513, + "theoretical_loss": 3.530885165848632, + "tokens_seen": 1568014336 + }, + { + "epoch": 0.48, + "learning_rate": 0.0004242006018054163, + "loss": 1.3907, + "theoretical_loss": 3.530872369000649, + "tokens_seen": 1568079872 + }, + { + "epoch": 0.48, + "learning_rate": 0.0004241845536609829, + "loss": 1.4347, + "theoretical_loss": 3.5308595728372296, + "tokens_seen": 1568145408 + }, + { + "epoch": 0.48, + "learning_rate": 0.0004241685055165497, + "loss": 1.4276, + "theoretical_loss": 3.530846777358306, + "tokens_seen": 1568210944 + }, + { + "epoch": 0.48, + "learning_rate": 0.0004241524573721163, + "loss": 1.4791, + "theoretical_loss": 3.5308339825638155, + "tokens_seen": 1568276480 + }, + { + "epoch": 0.48, + "learning_rate": 0.0004241364092276831, + "loss": 1.3283, + "theoretical_loss": 3.5308211884536913, + "tokens_seen": 1568342016 + }, + { + "epoch": 0.48, + "learning_rate": 0.0004241203610832497, + "loss": 1.3101, + "theoretical_loss": 3.5308083950278686, + "tokens_seen": 1568407552 + }, + { + "epoch": 0.48, + "learning_rate": 0.0004241043129388165, + "loss": 1.3416, + "theoretical_loss": 3.5307956022862825, + "tokens_seen": 1568473088 + }, + { + "epoch": 0.48, + "learning_rate": 0.0004240882647943831, + "loss": 1.3987, + "theoretical_loss": 3.5307828102288683, + "tokens_seen": 1568538624 + }, + { + "epoch": 0.48, + "learning_rate": 0.0004240722166499499, + "loss": 1.3958, + "theoretical_loss": 3.5307700188555593, + "tokens_seen": 1568604160 + }, + { + "epoch": 0.48, + "learning_rate": 0.0004240561685055165, + "loss": 1.3261, + "theoretical_loss": 3.530757228166292, + "tokens_seen": 1568669696 + }, + { + "epoch": 0.48, + "learning_rate": 0.0004240401203610833, + "loss": 1.3618, + "theoretical_loss": 3.5307444381610003, + "tokens_seen": 1568735232 + }, + { + "epoch": 0.48, + "learning_rate": 0.00042402407221664993, + "loss": 1.3895, + "theoretical_loss": 3.5307316488396188, + "tokens_seen": 1568800768 + }, + { + "epoch": 0.48, + "learning_rate": 0.0004240080240722167, + "loss": 1.383, + "theoretical_loss": 3.530718860202083, + "tokens_seen": 1568866304 + }, + { + "epoch": 0.48, + "learning_rate": 0.00042399197592778344, + "loss": 1.3264, + "theoretical_loss": 3.5307060722483277, + "tokens_seen": 1568931840 + }, + { + "epoch": 0.48, + "learning_rate": 0.0004239759277833501, + "loss": 1.3652, + "theoretical_loss": 3.530693284978288, + "tokens_seen": 1568997376 + }, + { + "epoch": 0.48, + "learning_rate": 0.00042395987963891684, + "loss": 1.293, + "theoretical_loss": 3.530680498391898, + "tokens_seen": 1569062912 + }, + { + "epoch": 0.48, + "learning_rate": 0.0004239438314944835, + "loss": 1.4607, + "theoretical_loss": 3.530667712489093, + "tokens_seen": 1569128448 + }, + { + "epoch": 0.48, + "learning_rate": 0.0004239277833500502, + "loss": 1.3537, + "theoretical_loss": 3.5306549272698082, + "tokens_seen": 1569193984 + }, + { + "epoch": 0.48, + "learning_rate": 0.0004239117352056169, + "loss": 1.2856, + "theoretical_loss": 3.5306421427339783, + "tokens_seen": 1569259520 + }, + { + "epoch": 0.48, + "learning_rate": 0.0004238956870611836, + "loss": 1.4253, + "theoretical_loss": 3.530629358881538, + "tokens_seen": 1569325056 + }, + { + "epoch": 0.48, + "learning_rate": 0.0004238796389167503, + "loss": 1.3142, + "theoretical_loss": 3.5306165757124224, + "tokens_seen": 1569390592 + }, + { + "epoch": 0.48, + "learning_rate": 0.000423863590772317, + "loss": 1.3904, + "theoretical_loss": 3.5306037932265664, + "tokens_seen": 1569456128 + }, + { + "epoch": 0.48, + "learning_rate": 0.0004238475426278837, + "loss": 1.3369, + "theoretical_loss": 3.5305910114239056, + "tokens_seen": 1569521664 + }, + { + "epoch": 0.48, + "objective/train/docs_used": 573246, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.3178856372833252, + "objective/train/theoretical_loss": 3.5305782303043736, + "objective/train/tokens_used": 1590047200, + "theoretical_loss": 3.5305782303043736, + "tokens_seen": 1569587200 + }, + { + "epoch": 0.48, + "learning_rate": 0.0004238314944834504, + "loss": 1.4026, + "theoretical_loss": 3.5305782303043736, + "tokens_seen": 1569587200 + }, + { + "epoch": 0.48, + "learning_rate": 0.0004238154463390171, + "loss": 1.3684, + "theoretical_loss": 3.5305654498679067, + "tokens_seen": 1569652736 + }, + { + "epoch": 0.48, + "learning_rate": 0.0004237993981945838, + "loss": 1.3881, + "theoretical_loss": 3.5305526701144387, + "tokens_seen": 1569718272 + }, + { + "epoch": 0.48, + "learning_rate": 0.00042378335005015044, + "loss": 1.2463, + "theoretical_loss": 3.5305398910439054, + "tokens_seen": 1569783808 + }, + { + "epoch": 0.48, + "learning_rate": 0.0004237673019057172, + "loss": 1.3731, + "theoretical_loss": 3.5305271126562423, + "tokens_seen": 1569849344 + }, + { + "epoch": 0.48, + "learning_rate": 0.00042375125376128384, + "loss": 1.3216, + "theoretical_loss": 3.5305143349513832, + "tokens_seen": 1569914880 + }, + { + "epoch": 0.48, + "learning_rate": 0.0004237352056168506, + "loss": 1.2594, + "theoretical_loss": 3.530501557929264, + "tokens_seen": 1569980416 + }, + { + "epoch": 0.48, + "learning_rate": 0.00042371915747241724, + "loss": 1.2972, + "theoretical_loss": 3.5304887815898187, + "tokens_seen": 1570045952 + }, + { + "epoch": 0.48, + "learning_rate": 0.000423703109327984, + "loss": 1.4625, + "theoretical_loss": 3.530476005932983, + "tokens_seen": 1570111488 + }, + { + "epoch": 0.48, + "learning_rate": 0.00042368706118355064, + "loss": 1.3318, + "theoretical_loss": 3.5304632309586923, + "tokens_seen": 1570177024 + }, + { + "epoch": 0.48, + "learning_rate": 0.0004236710130391174, + "loss": 1.3495, + "theoretical_loss": 3.5304504566668813, + "tokens_seen": 1570242560 + }, + { + "epoch": 0.48, + "learning_rate": 0.00042365496489468404, + "loss": 1.4631, + "theoretical_loss": 3.5304376830574844, + "tokens_seen": 1570308096 + }, + { + "epoch": 0.48, + "learning_rate": 0.0004236389167502508, + "loss": 1.3243, + "theoretical_loss": 3.530424910130438, + "tokens_seen": 1570373632 + }, + { + "epoch": 0.48, + "learning_rate": 0.00042362286860581744, + "loss": 1.3047, + "theoretical_loss": 3.5304121378856763, + "tokens_seen": 1570439168 + }, + { + "epoch": 0.48, + "learning_rate": 0.0004236068204613842, + "loss": 1.3964, + "theoretical_loss": 3.5303993663231346, + "tokens_seen": 1570504704 + }, + { + "epoch": 0.48, + "learning_rate": 0.00042359077231695085, + "loss": 1.2513, + "theoretical_loss": 3.5303865954427476, + "tokens_seen": 1570570240 + }, + { + "epoch": 0.48, + "learning_rate": 0.0004235747241725176, + "loss": 1.2738, + "theoretical_loss": 3.530373825244451, + "tokens_seen": 1570635776 + }, + { + "epoch": 0.48, + "learning_rate": 0.00042355867602808425, + "loss": 1.3053, + "theoretical_loss": 3.5303610557281795, + "tokens_seen": 1570701312 + }, + { + "epoch": 0.48, + "learning_rate": 0.000423542627883651, + "loss": 1.2418, + "theoretical_loss": 3.5303482868938687, + "tokens_seen": 1570766848 + }, + { + "epoch": 0.48, + "learning_rate": 0.00042352657973921765, + "loss": 1.3295, + "theoretical_loss": 3.5303355187414533, + "tokens_seen": 1570832384 + }, + { + "epoch": 0.48, + "learning_rate": 0.0004235105315947844, + "loss": 1.2919, + "theoretical_loss": 3.5303227512708686, + "tokens_seen": 1570897920 + }, + { + "epoch": 0.48, + "learning_rate": 0.00042349448345035105, + "loss": 1.3203, + "theoretical_loss": 3.530309984482049, + "tokens_seen": 1570963456 + }, + { + "epoch": 0.48, + "learning_rate": 0.0004234784353059178, + "loss": 1.3449, + "theoretical_loss": 3.5302972183749306, + "tokens_seen": 1571028992 + }, + { + "epoch": 0.48, + "learning_rate": 0.00042346238716148445, + "loss": 1.3886, + "theoretical_loss": 3.530284452949449, + "tokens_seen": 1571094528 + }, + { + "epoch": 0.48, + "learning_rate": 0.0004234463390170512, + "loss": 1.4321, + "theoretical_loss": 3.5302716882055383, + "tokens_seen": 1571160064 + }, + { + "epoch": 0.48, + "learning_rate": 0.00042343029087261785, + "loss": 1.2794, + "theoretical_loss": 3.530258924143134, + "tokens_seen": 1571225600 + }, + { + "epoch": 0.48, + "learning_rate": 0.0004234142427281846, + "loss": 1.34, + "theoretical_loss": 3.5302461607621716, + "tokens_seen": 1571291136 + }, + { + "epoch": 0.48, + "learning_rate": 0.00042339819458375125, + "loss": 1.4039, + "theoretical_loss": 3.5302333980625855, + "tokens_seen": 1571356672 + }, + { + "epoch": 0.48, + "learning_rate": 0.000423382146439318, + "loss": 1.3639, + "theoretical_loss": 3.5302206360443122, + "tokens_seen": 1571422208 + }, + { + "epoch": 0.48, + "learning_rate": 0.00042336609829488465, + "loss": 1.4, + "theoretical_loss": 3.5302078747072856, + "tokens_seen": 1571487744 + }, + { + "epoch": 0.48, + "learning_rate": 0.0004233500501504514, + "loss": 1.3468, + "theoretical_loss": 3.530195114051442, + "tokens_seen": 1571553280 + }, + { + "epoch": 0.48, + "learning_rate": 0.00042333400200601806, + "loss": 1.2685, + "theoretical_loss": 3.5301823540767154, + "tokens_seen": 1571618816 + }, + { + "epoch": 0.48, + "learning_rate": 0.0004233179538615848, + "loss": 1.3066, + "theoretical_loss": 3.5301695947830423, + "tokens_seen": 1571684352 + }, + { + "epoch": 0.48, + "learning_rate": 0.00042330190571715146, + "loss": 1.3746, + "theoretical_loss": 3.5301568361703577, + "tokens_seen": 1571749888 + }, + { + "epoch": 0.48, + "learning_rate": 0.0004232858575727182, + "loss": 1.4491, + "theoretical_loss": 3.530144078238596, + "tokens_seen": 1571815424 + }, + { + "epoch": 0.48, + "learning_rate": 0.00042326980942828486, + "loss": 1.4165, + "theoretical_loss": 3.5301313209876937, + "tokens_seen": 1571880960 + }, + { + "epoch": 0.48, + "learning_rate": 0.0004232537612838516, + "loss": 1.3658, + "theoretical_loss": 3.530118564417585, + "tokens_seen": 1571946496 + }, + { + "epoch": 0.48, + "learning_rate": 0.00042323771313941826, + "loss": 1.3298, + "theoretical_loss": 3.530105808528206, + "tokens_seen": 1572012032 + }, + { + "epoch": 0.48, + "learning_rate": 0.000423221664994985, + "loss": 1.2907, + "theoretical_loss": 3.5300930533194914, + "tokens_seen": 1572077568 + }, + { + "epoch": 0.48, + "learning_rate": 0.00042320561685055166, + "loss": 1.3463, + "theoretical_loss": 3.5300802987913764, + "tokens_seen": 1572143104 + }, + { + "epoch": 0.48, + "learning_rate": 0.0004231895687061184, + "loss": 1.2428, + "theoretical_loss": 3.5300675449437975, + "tokens_seen": 1572208640 + }, + { + "epoch": 0.48, + "learning_rate": 0.00042317352056168506, + "loss": 1.4324, + "theoretical_loss": 3.5300547917766885, + "tokens_seen": 1572274176 + }, + { + "epoch": 0.48, + "learning_rate": 0.00042315747241725176, + "loss": 1.3547, + "theoretical_loss": 3.530042039289986, + "tokens_seen": 1572339712 + }, + { + "epoch": 0.48, + "learning_rate": 0.00042314142427281846, + "loss": 1.4993, + "theoretical_loss": 3.5300292874836243, + "tokens_seen": 1572405248 + }, + { + "epoch": 0.48, + "learning_rate": 0.00042312537612838516, + "loss": 1.4003, + "theoretical_loss": 3.5300165363575395, + "tokens_seen": 1572470784 + }, + { + "epoch": 0.48, + "learning_rate": 0.00042310932798395186, + "loss": 1.489, + "theoretical_loss": 3.530003785911666, + "tokens_seen": 1572536320 + }, + { + "epoch": 0.48, + "learning_rate": 0.00042309327983951856, + "loss": 1.3781, + "theoretical_loss": 3.5299910361459403, + "tokens_seen": 1572601856 + }, + { + "epoch": 0.48, + "learning_rate": 0.00042307723169508527, + "loss": 1.4046, + "theoretical_loss": 3.529978287060298, + "tokens_seen": 1572667392 + }, + { + "epoch": 0.48, + "learning_rate": 0.00042306118355065197, + "loss": 1.4095, + "theoretical_loss": 3.5299655386546727, + "tokens_seen": 1572732928 + }, + { + "epoch": 0.48, + "learning_rate": 0.00042304513540621867, + "loss": 1.3724, + "theoretical_loss": 3.5299527909290016, + "tokens_seen": 1572798464 + }, + { + "epoch": 0.48, + "objective/train/docs_used": 574423, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.4481154680252075, + "objective/train/theoretical_loss": 3.529940043883219, + "objective/train/tokens_used": 1593324000, + "theoretical_loss": 3.529940043883219, + "tokens_seen": 1572864000 + }, + { + "epoch": 0.48, + "learning_rate": 0.00042302908726178537, + "loss": 1.4685, + "theoretical_loss": 3.529940043883219, + "tokens_seen": 1572864000 + }, + { + "epoch": 0.48, + "learning_rate": 0.000423013039117352, + "loss": 1.2855, + "theoretical_loss": 3.529927297517261, + "tokens_seen": 1572929536 + }, + { + "epoch": 0.48, + "learning_rate": 0.00042299699097291877, + "loss": 1.2846, + "theoretical_loss": 3.5299145518310624, + "tokens_seen": 1572995072 + }, + { + "epoch": 0.48, + "learning_rate": 0.0004229809428284854, + "loss": 1.3977, + "theoretical_loss": 3.5299018068245593, + "tokens_seen": 1573060608 + }, + { + "epoch": 0.48, + "learning_rate": 0.00042296489468405217, + "loss": 1.3181, + "theoretical_loss": 3.529889062497687, + "tokens_seen": 1573126144 + }, + { + "epoch": 0.48, + "learning_rate": 0.0004229488465396188, + "loss": 1.3307, + "theoretical_loss": 3.5298763188503806, + "tokens_seen": 1573191680 + }, + { + "epoch": 0.48, + "learning_rate": 0.00042293279839518557, + "loss": 1.2297, + "theoretical_loss": 3.529863575882575, + "tokens_seen": 1573257216 + }, + { + "epoch": 0.48, + "learning_rate": 0.0004229167502507522, + "loss": 1.2961, + "theoretical_loss": 3.529850833594207, + "tokens_seen": 1573322752 + }, + { + "epoch": 0.48, + "learning_rate": 0.00042290070210631897, + "loss": 1.3648, + "theoretical_loss": 3.5298380919852113, + "tokens_seen": 1573388288 + }, + { + "epoch": 0.48, + "learning_rate": 0.0004228846539618856, + "loss": 1.4163, + "theoretical_loss": 3.529825351055524, + "tokens_seen": 1573453824 + }, + { + "epoch": 0.48, + "learning_rate": 0.0004228686058174524, + "loss": 1.3655, + "theoretical_loss": 3.529812610805079, + "tokens_seen": 1573519360 + }, + { + "epoch": 0.48, + "learning_rate": 0.000422852557673019, + "loss": 1.2144, + "theoretical_loss": 3.529799871233814, + "tokens_seen": 1573584896 + }, + { + "epoch": 0.48, + "learning_rate": 0.0004228365095285858, + "loss": 1.2866, + "theoretical_loss": 3.529787132341663, + "tokens_seen": 1573650432 + }, + { + "epoch": 0.48, + "learning_rate": 0.00042282046138415253, + "loss": 1.3169, + "theoretical_loss": 3.529774394128562, + "tokens_seen": 1573715968 + }, + { + "epoch": 0.48, + "learning_rate": 0.0004228044132397192, + "loss": 1.5521, + "theoretical_loss": 3.529761656594447, + "tokens_seen": 1573781504 + }, + { + "epoch": 0.48, + "learning_rate": 0.00042278836509528593, + "loss": 1.2743, + "theoretical_loss": 3.5297489197392524, + "tokens_seen": 1573847040 + }, + { + "epoch": 0.48, + "learning_rate": 0.0004227723169508526, + "loss": 1.2607, + "theoretical_loss": 3.529736183562914, + "tokens_seen": 1573912576 + }, + { + "epoch": 0.48, + "learning_rate": 0.00042275626880641933, + "loss": 1.3604, + "theoretical_loss": 3.5297234480653685, + "tokens_seen": 1573978112 + }, + { + "epoch": 0.48, + "learning_rate": 0.000422740220661986, + "loss": 1.425, + "theoretical_loss": 3.5297107132465504, + "tokens_seen": 1574043648 + }, + { + "epoch": 0.48, + "learning_rate": 0.00042272417251755273, + "loss": 1.3417, + "theoretical_loss": 3.529697979106395, + "tokens_seen": 1574109184 + }, + { + "epoch": 0.48, + "learning_rate": 0.0004227081243731194, + "loss": 1.2541, + "theoretical_loss": 3.5296852456448393, + "tokens_seen": 1574174720 + }, + { + "epoch": 0.48, + "learning_rate": 0.00042269207622868614, + "loss": 1.383, + "theoretical_loss": 3.5296725128618176, + "tokens_seen": 1574240256 + }, + { + "epoch": 0.48, + "learning_rate": 0.0004226760280842528, + "loss": 1.3719, + "theoretical_loss": 3.5296597807572656, + "tokens_seen": 1574305792 + }, + { + "epoch": 0.48, + "learning_rate": 0.00042265997993981954, + "loss": 1.3345, + "theoretical_loss": 3.5296470493311194, + "tokens_seen": 1574371328 + }, + { + "epoch": 0.48, + "learning_rate": 0.0004226439317953862, + "loss": 1.4064, + "theoretical_loss": 3.529634318583315, + "tokens_seen": 1574436864 + }, + { + "epoch": 0.48, + "learning_rate": 0.00042262788365095294, + "loss": 1.2711, + "theoretical_loss": 3.5296215885137867, + "tokens_seen": 1574502400 + }, + { + "epoch": 0.48, + "learning_rate": 0.0004226118355065196, + "loss": 1.3921, + "theoretical_loss": 3.5296088591224715, + "tokens_seen": 1574567936 + }, + { + "epoch": 0.48, + "learning_rate": 0.00042259578736208634, + "loss": 1.2203, + "theoretical_loss": 3.529596130409304, + "tokens_seen": 1574633472 + }, + { + "epoch": 0.48, + "learning_rate": 0.000422579739217653, + "loss": 1.4181, + "theoretical_loss": 3.5295834023742203, + "tokens_seen": 1574699008 + }, + { + "epoch": 0.48, + "learning_rate": 0.00042256369107321974, + "loss": 1.1378, + "theoretical_loss": 3.529570675017156, + "tokens_seen": 1574764544 + }, + { + "epoch": 0.48, + "learning_rate": 0.0004225476429287864, + "loss": 1.1794, + "theoretical_loss": 3.529557948338047, + "tokens_seen": 1574830080 + }, + { + "epoch": 0.48, + "learning_rate": 0.0004225315947843531, + "loss": 1.3471, + "theoretical_loss": 3.5295452223368287, + "tokens_seen": 1574895616 + }, + { + "epoch": 0.48, + "learning_rate": 0.0004225155466399198, + "loss": 1.3932, + "theoretical_loss": 3.529532497013437, + "tokens_seen": 1574961152 + }, + { + "epoch": 0.48, + "learning_rate": 0.0004224994984954865, + "loss": 1.3936, + "theoretical_loss": 3.5295197723678076, + "tokens_seen": 1575026688 + }, + { + "epoch": 0.48, + "learning_rate": 0.0004224834503510532, + "loss": 1.3062, + "theoretical_loss": 3.5295070483998754, + "tokens_seen": 1575092224 + }, + { + "epoch": 0.48, + "learning_rate": 0.0004224674022066199, + "loss": 1.2478, + "theoretical_loss": 3.529494325109577, + "tokens_seen": 1575157760 + }, + { + "epoch": 0.48, + "learning_rate": 0.0004224513540621866, + "loss": 1.3549, + "theoretical_loss": 3.5294816024968485, + "tokens_seen": 1575223296 + }, + { + "epoch": 0.48, + "learning_rate": 0.0004224353059177533, + "loss": 1.4092, + "theoretical_loss": 3.529468880561624, + "tokens_seen": 1575288832 + }, + { + "epoch": 0.48, + "learning_rate": 0.00042241925777332, + "loss": 1.4595, + "theoretical_loss": 3.529456159303841, + "tokens_seen": 1575354368 + }, + { + "epoch": 0.48, + "learning_rate": 0.0004224032096288867, + "loss": 1.2754, + "theoretical_loss": 3.5294434387234346, + "tokens_seen": 1575419904 + }, + { + "epoch": 0.48, + "learning_rate": 0.00042238716148445334, + "loss": 1.3645, + "theoretical_loss": 3.52943071882034, + "tokens_seen": 1575485440 + }, + { + "epoch": 0.48, + "learning_rate": 0.0004223711133400201, + "loss": 1.2569, + "theoretical_loss": 3.529417999594494, + "tokens_seen": 1575550976 + }, + { + "epoch": 0.48, + "learning_rate": 0.00042235506519558674, + "loss": 1.3463, + "theoretical_loss": 3.529405281045831, + "tokens_seen": 1575616512 + }, + { + "epoch": 0.48, + "learning_rate": 0.0004223390170511535, + "loss": 1.3326, + "theoretical_loss": 3.529392563174288, + "tokens_seen": 1575682048 + }, + { + "epoch": 0.48, + "learning_rate": 0.00042232296890672014, + "loss": 1.3876, + "theoretical_loss": 3.5293798459798005, + "tokens_seen": 1575747584 + }, + { + "epoch": 0.48, + "learning_rate": 0.0004223069207622869, + "loss": 1.1863, + "theoretical_loss": 3.529367129462304, + "tokens_seen": 1575813120 + }, + { + "epoch": 0.48, + "learning_rate": 0.00042229087261785354, + "loss": 1.3266, + "theoretical_loss": 3.5293544136217347, + "tokens_seen": 1575878656 + }, + { + "epoch": 0.48, + "learning_rate": 0.0004222748244734203, + "loss": 1.3659, + "theoretical_loss": 3.529341698458028, + "tokens_seen": 1575944192 + }, + { + "epoch": 0.48, + "learning_rate": 0.00042225877632898694, + "loss": 1.3344, + "theoretical_loss": 3.52932898397112, + "tokens_seen": 1576009728 + }, + { + "epoch": 0.48, + "learning_rate": 0.0004222427281845537, + "loss": 1.3919, + "theoretical_loss": 3.529316270160946, + "tokens_seen": 1576075264 + }, + { + "epoch": 0.48, + "objective/train/docs_used": 575737, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.2119507789611816, + "objective/train/theoretical_loss": 3.529303557027443, + "objective/train/tokens_used": 1596600800, + "theoretical_loss": 3.529303557027443, + "tokens_seen": 1576140800 + }, + { + "epoch": 0.48, + "learning_rate": 0.00042222668004012035, + "loss": 1.2692, + "theoretical_loss": 3.529303557027443, + "tokens_seen": 1576140800 + }, + { + "epoch": 0.48, + "learning_rate": 0.0004222106318956871, + "loss": 1.431, + "theoretical_loss": 3.5292908445705455, + "tokens_seen": 1576206336 + }, + { + "epoch": 0.48, + "learning_rate": 0.00042219458375125375, + "loss": 1.2652, + "theoretical_loss": 3.5292781327901905, + "tokens_seen": 1576271872 + }, + { + "epoch": 0.48, + "learning_rate": 0.0004221785356068205, + "loss": 1.2948, + "theoretical_loss": 3.5292654216863135, + "tokens_seen": 1576337408 + }, + { + "epoch": 0.48, + "learning_rate": 0.00042216248746238715, + "loss": 1.3846, + "theoretical_loss": 3.5292527112588497, + "tokens_seen": 1576402944 + }, + { + "epoch": 0.48, + "learning_rate": 0.0004221464393179539, + "loss": 1.3673, + "theoretical_loss": 3.529240001507736, + "tokens_seen": 1576468480 + }, + { + "epoch": 0.48, + "learning_rate": 0.00042213039117352055, + "loss": 1.3267, + "theoretical_loss": 3.529227292432908, + "tokens_seen": 1576534016 + }, + { + "epoch": 0.48, + "learning_rate": 0.0004221143430290873, + "loss": 1.3249, + "theoretical_loss": 3.5292145840343006, + "tokens_seen": 1576599552 + }, + { + "epoch": 0.48, + "learning_rate": 0.00042209829488465395, + "loss": 1.3286, + "theoretical_loss": 3.529201876311851, + "tokens_seen": 1576665088 + }, + { + "epoch": 0.48, + "learning_rate": 0.0004220822467402207, + "loss": 1.3045, + "theoretical_loss": 3.5291891692654946, + "tokens_seen": 1576730624 + }, + { + "epoch": 0.48, + "learning_rate": 0.00042206619859578735, + "loss": 1.3505, + "theoretical_loss": 3.529176462895168, + "tokens_seen": 1576796160 + }, + { + "epoch": 0.48, + "learning_rate": 0.0004220501504513541, + "loss": 1.408, + "theoretical_loss": 3.529163757200806, + "tokens_seen": 1576861696 + }, + { + "epoch": 0.48, + "learning_rate": 0.00042203410230692075, + "loss": 1.2788, + "theoretical_loss": 3.5291510521823453, + "tokens_seen": 1576927232 + }, + { + "epoch": 0.48, + "learning_rate": 0.0004220180541624875, + "loss": 1.3741, + "theoretical_loss": 3.529138347839721, + "tokens_seen": 1576992768 + }, + { + "epoch": 0.48, + "learning_rate": 0.00042200200601805415, + "loss": 1.3528, + "theoretical_loss": 3.5291256441728702, + "tokens_seen": 1577058304 + }, + { + "epoch": 0.48, + "learning_rate": 0.0004219859578736209, + "loss": 1.3101, + "theoretical_loss": 3.5291129411817286, + "tokens_seen": 1577123840 + }, + { + "epoch": 0.48, + "learning_rate": 0.00042196990972918756, + "loss": 1.3099, + "theoretical_loss": 3.5291002388662323, + "tokens_seen": 1577189376 + }, + { + "epoch": 0.48, + "learning_rate": 0.0004219538615847543, + "loss": 1.3345, + "theoretical_loss": 3.5290875372263164, + "tokens_seen": 1577254912 + }, + { + "epoch": 0.48, + "learning_rate": 0.00042193781344032096, + "loss": 1.2924, + "theoretical_loss": 3.529074836261917, + "tokens_seen": 1577320448 + }, + { + "epoch": 0.48, + "learning_rate": 0.0004219217652958877, + "loss": 1.3079, + "theoretical_loss": 3.5290621359729712, + "tokens_seen": 1577385984 + }, + { + "epoch": 0.48, + "learning_rate": 0.00042190571715145436, + "loss": 1.3108, + "theoretical_loss": 3.5290494363594145, + "tokens_seen": 1577451520 + }, + { + "epoch": 0.48, + "learning_rate": 0.0004218896690070211, + "loss": 1.3173, + "theoretical_loss": 3.5290367374211824, + "tokens_seen": 1577517056 + }, + { + "epoch": 0.48, + "learning_rate": 0.00042187362086258776, + "loss": 1.4055, + "theoretical_loss": 3.5290240391582115, + "tokens_seen": 1577582592 + }, + { + "epoch": 0.48, + "learning_rate": 0.0004218575727181545, + "loss": 1.3688, + "theoretical_loss": 3.5290113415704383, + "tokens_seen": 1577648128 + }, + { + "epoch": 0.48, + "learning_rate": 0.00042184152457372116, + "loss": 1.4318, + "theoretical_loss": 3.5289986446577974, + "tokens_seen": 1577713664 + }, + { + "epoch": 0.48, + "learning_rate": 0.0004218254764292879, + "loss": 1.3814, + "theoretical_loss": 3.528985948420226, + "tokens_seen": 1577779200 + }, + { + "epoch": 0.48, + "learning_rate": 0.00042180942828485456, + "loss": 1.3294, + "theoretical_loss": 3.52897325285766, + "tokens_seen": 1577844736 + }, + { + "epoch": 0.48, + "learning_rate": 0.0004217933801404213, + "loss": 1.3435, + "theoretical_loss": 3.528960557970035, + "tokens_seen": 1577910272 + }, + { + "epoch": 0.48, + "learning_rate": 0.00042177733199598796, + "loss": 1.2253, + "theoretical_loss": 3.5289478637572875, + "tokens_seen": 1577975808 + }, + { + "epoch": 0.48, + "learning_rate": 0.00042176128385155466, + "loss": 1.3524, + "theoretical_loss": 3.528935170219354, + "tokens_seen": 1578041344 + }, + { + "epoch": 0.48, + "learning_rate": 0.00042174523570712136, + "loss": 1.2791, + "theoretical_loss": 3.5289224773561703, + "tokens_seen": 1578106880 + }, + { + "epoch": 0.48, + "learning_rate": 0.00042172918756268806, + "loss": 1.3281, + "theoretical_loss": 3.528909785167672, + "tokens_seen": 1578172416 + }, + { + "epoch": 0.48, + "learning_rate": 0.00042171313941825477, + "loss": 1.3381, + "theoretical_loss": 3.528897093653795, + "tokens_seen": 1578237952 + }, + { + "epoch": 0.48, + "learning_rate": 0.00042169709127382147, + "loss": 1.2598, + "theoretical_loss": 3.5288844028144766, + "tokens_seen": 1578303488 + }, + { + "epoch": 0.48, + "learning_rate": 0.00042168104312938817, + "loss": 1.327, + "theoretical_loss": 3.5288717126496523, + "tokens_seen": 1578369024 + }, + { + "epoch": 0.48, + "learning_rate": 0.00042166499498495487, + "loss": 1.3889, + "theoretical_loss": 3.5288590231592583, + "tokens_seen": 1578434560 + }, + { + "epoch": 0.48, + "learning_rate": 0.00042164894684052157, + "loss": 1.385, + "theoretical_loss": 3.528846334343231, + "tokens_seen": 1578500096 + }, + { + "epoch": 0.48, + "learning_rate": 0.00042163289869608827, + "loss": 1.3166, + "theoretical_loss": 3.528833646201506, + "tokens_seen": 1578565632 + }, + { + "epoch": 0.48, + "learning_rate": 0.000421616850551655, + "loss": 1.411, + "theoretical_loss": 3.52882095873402, + "tokens_seen": 1578631168 + }, + { + "epoch": 0.48, + "learning_rate": 0.00042160080240722167, + "loss": 1.3585, + "theoretical_loss": 3.528808271940709, + "tokens_seen": 1578696704 + }, + { + "epoch": 0.48, + "learning_rate": 0.0004215847542627884, + "loss": 1.405, + "theoretical_loss": 3.5287955858215083, + "tokens_seen": 1578762240 + }, + { + "epoch": 0.48, + "learning_rate": 0.00042156870611835507, + "loss": 1.3641, + "theoretical_loss": 3.5287829003763562, + "tokens_seen": 1578827776 + }, + { + "epoch": 0.48, + "learning_rate": 0.0004215526579739218, + "loss": 1.4267, + "theoretical_loss": 3.5287702156051868, + "tokens_seen": 1578893312 + }, + { + "epoch": 0.48, + "learning_rate": 0.00042153660982948847, + "loss": 1.3535, + "theoretical_loss": 3.528757531507938, + "tokens_seen": 1578958848 + }, + { + "epoch": 0.48, + "learning_rate": 0.00042152056168505523, + "loss": 1.3979, + "theoretical_loss": 3.5287448480845445, + "tokens_seen": 1579024384 + }, + { + "epoch": 0.48, + "learning_rate": 0.0004215045135406219, + "loss": 1.4439, + "theoretical_loss": 3.5287321653349437, + "tokens_seen": 1579089920 + }, + { + "epoch": 0.48, + "learning_rate": 0.00042148846539618863, + "loss": 1.3751, + "theoretical_loss": 3.5287194832590707, + "tokens_seen": 1579155456 + }, + { + "epoch": 0.48, + "learning_rate": 0.0004214724172517553, + "loss": 1.3453, + "theoretical_loss": 3.5287068018568633, + "tokens_seen": 1579220992 + }, + { + "epoch": 0.48, + "learning_rate": 0.00042145636910732203, + "loss": 1.3871, + "theoretical_loss": 3.5286941211282565, + "tokens_seen": 1579286528 + }, + { + "epoch": 0.48, + "learning_rate": 0.0004214403209628887, + "loss": 1.3779, + "theoretical_loss": 3.5286814410731875, + "tokens_seen": 1579352064 + }, + { + "epoch": 0.48, + "objective/train/docs_used": 576827, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.3830909729003906, + "objective/train/theoretical_loss": 3.5286687616915913, + "objective/train/tokens_used": 1599877600, + "theoretical_loss": 3.5286687616915913, + "tokens_seen": 1579417600 + }, + { + "epoch": 0.48, + "learning_rate": 0.00042142427281845543, + "loss": 1.3971, + "theoretical_loss": 3.5286687616915913, + "tokens_seen": 1579417600 + }, + { + "epoch": 0.48, + "learning_rate": 0.0004214082246740221, + "loss": 1.4314, + "theoretical_loss": 3.5286560829834057, + "tokens_seen": 1579483136 + }, + { + "epoch": 0.48, + "learning_rate": 0.00042139217652958883, + "loss": 1.2798, + "theoretical_loss": 3.5286434049485655, + "tokens_seen": 1579548672 + }, + { + "epoch": 0.48, + "learning_rate": 0.0004213761283851555, + "loss": 1.2822, + "theoretical_loss": 3.528630727587008, + "tokens_seen": 1579614208 + }, + { + "epoch": 0.48, + "learning_rate": 0.00042136008024072223, + "loss": 1.3549, + "theoretical_loss": 3.52861805089867, + "tokens_seen": 1579679744 + }, + { + "epoch": 0.48, + "learning_rate": 0.0004213440320962889, + "loss": 1.3583, + "theoretical_loss": 3.5286053748834862, + "tokens_seen": 1579745280 + }, + { + "epoch": 0.48, + "learning_rate": 0.00042132798395185563, + "loss": 1.2285, + "theoretical_loss": 3.5285926995413943, + "tokens_seen": 1579810816 + }, + { + "epoch": 0.48, + "learning_rate": 0.0004213119358074223, + "loss": 1.4937, + "theoretical_loss": 3.52858002487233, + "tokens_seen": 1579876352 + }, + { + "epoch": 0.48, + "learning_rate": 0.00042129588766298904, + "loss": 1.4127, + "theoretical_loss": 3.5285673508762296, + "tokens_seen": 1579941888 + }, + { + "epoch": 0.48, + "learning_rate": 0.0004212798395185557, + "loss": 1.3365, + "theoretical_loss": 3.52855467755303, + "tokens_seen": 1580007424 + }, + { + "epoch": 0.48, + "learning_rate": 0.00042126379137412244, + "loss": 1.3614, + "theoretical_loss": 3.528542004902667, + "tokens_seen": 1580072960 + }, + { + "epoch": 0.48, + "learning_rate": 0.0004212477432296891, + "loss": 1.3862, + "theoretical_loss": 3.528529332925077, + "tokens_seen": 1580138496 + }, + { + "epoch": 0.48, + "learning_rate": 0.00042123169508525584, + "loss": 1.3148, + "theoretical_loss": 3.5285166616201966, + "tokens_seen": 1580204032 + }, + { + "epoch": 0.48, + "learning_rate": 0.0004212156469408225, + "loss": 1.3491, + "theoretical_loss": 3.5285039909879625, + "tokens_seen": 1580269568 + }, + { + "epoch": 0.48, + "learning_rate": 0.00042119959879638924, + "loss": 1.4, + "theoretical_loss": 3.5284913210283113, + "tokens_seen": 1580335104 + }, + { + "epoch": 0.48, + "learning_rate": 0.0004211835506519559, + "loss": 1.2958, + "theoretical_loss": 3.528478651741178, + "tokens_seen": 1580400640 + }, + { + "epoch": 0.48, + "learning_rate": 0.00042116750250752264, + "loss": 1.4162, + "theoretical_loss": 3.5284659831265, + "tokens_seen": 1580466176 + }, + { + "epoch": 0.48, + "learning_rate": 0.0004211514543630893, + "loss": 1.3248, + "theoretical_loss": 3.528453315184214, + "tokens_seen": 1580531712 + }, + { + "epoch": 0.48, + "learning_rate": 0.000421135406218656, + "loss": 1.5469, + "theoretical_loss": 3.528440647914256, + "tokens_seen": 1580597248 + }, + { + "epoch": 0.48, + "learning_rate": 0.0004211193580742227, + "loss": 1.4946, + "theoretical_loss": 3.528427981316562, + "tokens_seen": 1580662784 + }, + { + "epoch": 0.48, + "learning_rate": 0.0004211033099297894, + "loss": 1.3756, + "theoretical_loss": 3.5284153153910696, + "tokens_seen": 1580728320 + }, + { + "epoch": 0.48, + "learning_rate": 0.0004210872617853561, + "loss": 1.3486, + "theoretical_loss": 3.528402650137714, + "tokens_seen": 1580793856 + }, + { + "epoch": 0.48, + "learning_rate": 0.0004210712136409228, + "loss": 1.3905, + "theoretical_loss": 3.528389985556433, + "tokens_seen": 1580859392 + }, + { + "epoch": 0.48, + "learning_rate": 0.0004210551654964895, + "loss": 1.4176, + "theoretical_loss": 3.528377321647162, + "tokens_seen": 1580924928 + }, + { + "epoch": 0.48, + "learning_rate": 0.0004210391173520562, + "loss": 1.3232, + "theoretical_loss": 3.5283646584098376, + "tokens_seen": 1580990464 + }, + { + "epoch": 0.48, + "learning_rate": 0.0004210230692076229, + "loss": 1.3926, + "theoretical_loss": 3.528351995844397, + "tokens_seen": 1581056000 + }, + { + "epoch": 0.48, + "learning_rate": 0.0004210070210631896, + "loss": 1.3258, + "theoretical_loss": 3.5283393339507763, + "tokens_seen": 1581121536 + }, + { + "epoch": 0.48, + "learning_rate": 0.00042099097291875624, + "loss": 1.4894, + "theoretical_loss": 3.5283266727289115, + "tokens_seen": 1581187072 + }, + { + "epoch": 0.48, + "learning_rate": 0.000420974924774323, + "loss": 1.3531, + "theoretical_loss": 3.5283140121787397, + "tokens_seen": 1581252608 + }, + { + "epoch": 0.48, + "learning_rate": 0.00042095887662988964, + "loss": 1.3164, + "theoretical_loss": 3.5283013523001974, + "tokens_seen": 1581318144 + }, + { + "epoch": 0.48, + "learning_rate": 0.0004209428284854564, + "loss": 1.3413, + "theoretical_loss": 3.528288693093221, + "tokens_seen": 1581383680 + }, + { + "epoch": 0.48, + "learning_rate": 0.00042092678034102304, + "loss": 1.3191, + "theoretical_loss": 3.528276034557747, + "tokens_seen": 1581449216 + }, + { + "epoch": 0.48, + "learning_rate": 0.0004209107321965898, + "loss": 1.3976, + "theoretical_loss": 3.5282633766937126, + "tokens_seen": 1581514752 + }, + { + "epoch": 0.48, + "learning_rate": 0.00042089468405215644, + "loss": 1.2871, + "theoretical_loss": 3.5282507195010533, + "tokens_seen": 1581580288 + }, + { + "epoch": 0.48, + "learning_rate": 0.0004208786359077232, + "loss": 1.4089, + "theoretical_loss": 3.528238062979706, + "tokens_seen": 1581645824 + }, + { + "epoch": 0.48, + "learning_rate": 0.00042086258776328984, + "loss": 1.3765, + "theoretical_loss": 3.5282254071296077, + "tokens_seen": 1581711360 + }, + { + "epoch": 0.48, + "learning_rate": 0.0004208465396188566, + "loss": 1.3162, + "theoretical_loss": 3.528212751950695, + "tokens_seen": 1581776896 + }, + { + "epoch": 0.48, + "learning_rate": 0.00042083049147442325, + "loss": 1.3895, + "theoretical_loss": 3.528200097442904, + "tokens_seen": 1581842432 + }, + { + "epoch": 0.48, + "learning_rate": 0.00042081444332999, + "loss": 1.3687, + "theoretical_loss": 3.5281874436061713, + "tokens_seen": 1581907968 + }, + { + "epoch": 0.48, + "learning_rate": 0.00042079839518555665, + "loss": 1.4852, + "theoretical_loss": 3.5281747904404344, + "tokens_seen": 1581973504 + }, + { + "epoch": 0.48, + "learning_rate": 0.0004207823470411234, + "loss": 1.3502, + "theoretical_loss": 3.528162137945629, + "tokens_seen": 1582039040 + }, + { + "epoch": 0.48, + "learning_rate": 0.00042076629889669005, + "loss": 1.4103, + "theoretical_loss": 3.528149486121692, + "tokens_seen": 1582104576 + }, + { + "epoch": 0.48, + "learning_rate": 0.0004207502507522568, + "loss": 1.327, + "theoretical_loss": 3.52813683496856, + "tokens_seen": 1582170112 + }, + { + "epoch": 0.48, + "learning_rate": 0.00042073420260782345, + "loss": 1.34, + "theoretical_loss": 3.52812418448617, + "tokens_seen": 1582235648 + }, + { + "epoch": 0.48, + "learning_rate": 0.0004207181544633902, + "loss": 1.4646, + "theoretical_loss": 3.528111534674458, + "tokens_seen": 1582301184 + }, + { + "epoch": 0.48, + "learning_rate": 0.00042070210631895685, + "loss": 1.449, + "theoretical_loss": 3.5280988855333613, + "tokens_seen": 1582366720 + }, + { + "epoch": 0.48, + "learning_rate": 0.0004206860581745236, + "loss": 1.3422, + "theoretical_loss": 3.5280862370628165, + "tokens_seen": 1582432256 + }, + { + "epoch": 0.48, + "learning_rate": 0.00042067001003009025, + "loss": 1.3744, + "theoretical_loss": 3.5280735892627595, + "tokens_seen": 1582497792 + }, + { + "epoch": 0.48, + "learning_rate": 0.000420653961885657, + "loss": 1.4879, + "theoretical_loss": 3.528060942133128, + "tokens_seen": 1582563328 + }, + { + "epoch": 0.48, + "learning_rate": 0.00042063791374122365, + "loss": 1.4255, + "theoretical_loss": 3.528048295673859, + "tokens_seen": 1582628864 + }, + { + "epoch": 0.48, + "objective/train/docs_used": 578056, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.199568510055542, + "objective/train/theoretical_loss": 3.5280356498848873, + "objective/train/tokens_used": 1603154400, + "theoretical_loss": 3.5280356498848873, + "tokens_seen": 1582694400 + }, + { + "epoch": 0.48, + "learning_rate": 0.0004206218655967904, + "loss": 1.2582, + "theoretical_loss": 3.5280356498848873, + "tokens_seen": 1582694400 + }, + { + "epoch": 0.48, + "learning_rate": 0.00042060581745235705, + "loss": 1.4391, + "theoretical_loss": 3.528023004766151, + "tokens_seen": 1582759936 + }, + { + "epoch": 0.48, + "learning_rate": 0.0004205897693079238, + "loss": 1.3997, + "theoretical_loss": 3.5280103603175874, + "tokens_seen": 1582825472 + }, + { + "epoch": 0.48, + "learning_rate": 0.00042057372116349046, + "loss": 1.3387, + "theoretical_loss": 3.527997716539132, + "tokens_seen": 1582891008 + }, + { + "epoch": 0.48, + "learning_rate": 0.0004205576730190572, + "loss": 1.3968, + "theoretical_loss": 3.527985073430722, + "tokens_seen": 1582956544 + }, + { + "epoch": 0.48, + "learning_rate": 0.00042054162487462386, + "loss": 1.3281, + "theoretical_loss": 3.5279724309922944, + "tokens_seen": 1583022080 + }, + { + "epoch": 0.48, + "learning_rate": 0.0004205255767301906, + "loss": 1.3216, + "theoretical_loss": 3.527959789223786, + "tokens_seen": 1583087616 + }, + { + "epoch": 0.48, + "learning_rate": 0.00042050952858575726, + "loss": 1.3219, + "theoretical_loss": 3.5279471481251328, + "tokens_seen": 1583153152 + }, + { + "epoch": 0.48, + "learning_rate": 0.000420493480441324, + "loss": 1.2996, + "theoretical_loss": 3.5279345076962727, + "tokens_seen": 1583218688 + }, + { + "epoch": 0.48, + "learning_rate": 0.00042047743229689066, + "loss": 1.4005, + "theoretical_loss": 3.527921867937142, + "tokens_seen": 1583284224 + }, + { + "epoch": 0.48, + "learning_rate": 0.0004204613841524574, + "loss": 1.4264, + "theoretical_loss": 3.527909228847676, + "tokens_seen": 1583349760 + }, + { + "epoch": 0.48, + "learning_rate": 0.00042044533600802406, + "loss": 1.3871, + "theoretical_loss": 3.5278965904278143, + "tokens_seen": 1583415296 + }, + { + "epoch": 0.48, + "learning_rate": 0.0004204292878635908, + "loss": 1.419, + "theoretical_loss": 3.5278839526774917, + "tokens_seen": 1583480832 + }, + { + "epoch": 0.48, + "learning_rate": 0.0004204132397191575, + "loss": 1.4121, + "theoretical_loss": 3.527871315596646, + "tokens_seen": 1583546368 + }, + { + "epoch": 0.48, + "learning_rate": 0.0004203971915747242, + "loss": 1.3628, + "theoretical_loss": 3.527858679185213, + "tokens_seen": 1583611904 + }, + { + "epoch": 0.48, + "learning_rate": 0.0004203811434302909, + "loss": 1.3789, + "theoretical_loss": 3.527846043443131, + "tokens_seen": 1583677440 + }, + { + "epoch": 0.48, + "learning_rate": 0.00042036509528585756, + "loss": 1.3621, + "theoretical_loss": 3.527833408370336, + "tokens_seen": 1583742976 + }, + { + "epoch": 0.48, + "learning_rate": 0.0004203490471414243, + "loss": 1.4226, + "theoretical_loss": 3.527820773966764, + "tokens_seen": 1583808512 + }, + { + "epoch": 0.48, + "learning_rate": 0.00042033299899699097, + "loss": 1.3927, + "theoretical_loss": 3.5278081402323536, + "tokens_seen": 1583874048 + }, + { + "epoch": 0.48, + "learning_rate": 0.0004203169508525577, + "loss": 1.3888, + "theoretical_loss": 3.5277955071670406, + "tokens_seen": 1583939584 + }, + { + "epoch": 0.48, + "learning_rate": 0.00042030090270812437, + "loss": 1.4008, + "theoretical_loss": 3.5277828747707622, + "tokens_seen": 1584005120 + }, + { + "epoch": 0.48, + "learning_rate": 0.0004202848545636911, + "loss": 1.3308, + "theoretical_loss": 3.527770243043455, + "tokens_seen": 1584070656 + }, + { + "epoch": 0.48, + "learning_rate": 0.00042026880641925777, + "loss": 1.3724, + "theoretical_loss": 3.5277576119850567, + "tokens_seen": 1584136192 + }, + { + "epoch": 0.48, + "learning_rate": 0.0004202527582748245, + "loss": 1.451, + "theoretical_loss": 3.527744981595503, + "tokens_seen": 1584201728 + }, + { + "epoch": 0.48, + "learning_rate": 0.00042023671013039117, + "loss": 1.2512, + "theoretical_loss": 3.5277323518747314, + "tokens_seen": 1584267264 + }, + { + "epoch": 0.48, + "learning_rate": 0.0004202206619859579, + "loss": 1.3486, + "theoretical_loss": 3.5277197228226793, + "tokens_seen": 1584332800 + }, + { + "epoch": 0.48, + "learning_rate": 0.00042020461384152457, + "loss": 1.4616, + "theoretical_loss": 3.5277070944392825, + "tokens_seen": 1584398336 + }, + { + "epoch": 0.48, + "learning_rate": 0.0004201885656970913, + "loss": 1.3537, + "theoretical_loss": 3.5276944667244794, + "tokens_seen": 1584463872 + }, + { + "epoch": 0.48, + "learning_rate": 0.00042017251755265797, + "loss": 1.3556, + "theoretical_loss": 3.527681839678206, + "tokens_seen": 1584529408 + }, + { + "epoch": 0.48, + "learning_rate": 0.0004201564694082247, + "loss": 1.3942, + "theoretical_loss": 3.527669213300399, + "tokens_seen": 1584594944 + }, + { + "epoch": 0.48, + "learning_rate": 0.0004201404212637914, + "loss": 1.4501, + "theoretical_loss": 3.5276565875909967, + "tokens_seen": 1584660480 + }, + { + "epoch": 0.48, + "learning_rate": 0.00042012437311935813, + "loss": 1.3419, + "theoretical_loss": 3.5276439625499343, + "tokens_seen": 1584726016 + }, + { + "epoch": 0.48, + "learning_rate": 0.0004201083249749248, + "loss": 1.3359, + "theoretical_loss": 3.5276313381771502, + "tokens_seen": 1584791552 + }, + { + "epoch": 0.48, + "learning_rate": 0.00042009227683049153, + "loss": 1.3979, + "theoretical_loss": 3.5276187144725806, + "tokens_seen": 1584857088 + }, + { + "epoch": 0.48, + "learning_rate": 0.0004200762286860582, + "loss": 1.3034, + "theoretical_loss": 3.527606091436163, + "tokens_seen": 1584922624 + }, + { + "epoch": 0.48, + "learning_rate": 0.00042006018054162493, + "loss": 1.3712, + "theoretical_loss": 3.527593469067834, + "tokens_seen": 1584988160 + }, + { + "epoch": 0.48, + "learning_rate": 0.0004200441323971916, + "loss": 1.3616, + "theoretical_loss": 3.5275808473675307, + "tokens_seen": 1585053696 + }, + { + "epoch": 0.48, + "learning_rate": 0.00042002808425275833, + "loss": 1.3263, + "theoretical_loss": 3.52756822633519, + "tokens_seen": 1585119232 + }, + { + "epoch": 0.48, + "learning_rate": 0.000420012036108325, + "loss": 1.3134, + "theoretical_loss": 3.52755560597075, + "tokens_seen": 1585184768 + }, + { + "epoch": 0.48, + "learning_rate": 0.00041999598796389173, + "loss": 1.29, + "theoretical_loss": 3.5275429862741463, + "tokens_seen": 1585250304 + }, + { + "epoch": 0.48, + "learning_rate": 0.0004199799398194584, + "loss": 1.4794, + "theoretical_loss": 3.5275303672453164, + "tokens_seen": 1585315840 + }, + { + "epoch": 0.48, + "learning_rate": 0.00041996389167502513, + "loss": 1.4981, + "theoretical_loss": 3.5275177488841982, + "tokens_seen": 1585381376 + }, + { + "epoch": 0.48, + "learning_rate": 0.0004199478435305918, + "loss": 1.2366, + "theoretical_loss": 3.5275051311907277, + "tokens_seen": 1585446912 + }, + { + "epoch": 0.48, + "learning_rate": 0.00041993179538615854, + "loss": 1.2756, + "theoretical_loss": 3.5274925141648423, + "tokens_seen": 1585512448 + }, + { + "epoch": 0.48, + "learning_rate": 0.0004199157472417252, + "loss": 1.2889, + "theoretical_loss": 3.5274798978064794, + "tokens_seen": 1585577984 + }, + { + "epoch": 0.48, + "learning_rate": 0.00041989969909729194, + "loss": 1.3737, + "theoretical_loss": 3.527467282115576, + "tokens_seen": 1585643520 + }, + { + "epoch": 0.48, + "learning_rate": 0.0004198836509528586, + "loss": 1.3453, + "theoretical_loss": 3.5274546670920683, + "tokens_seen": 1585709056 + }, + { + "epoch": 0.48, + "learning_rate": 0.00041986760280842534, + "loss": 1.3718, + "theoretical_loss": 3.5274420527358945, + "tokens_seen": 1585774592 + }, + { + "epoch": 0.48, + "learning_rate": 0.000419851554663992, + "loss": 1.4105, + "theoretical_loss": 3.5274294390469914, + "tokens_seen": 1585840128 + }, + { + "epoch": 0.48, + "learning_rate": 0.00041983550651955874, + "loss": 1.2686, + "theoretical_loss": 3.5274168260252967, + "tokens_seen": 1585905664 + }, + { + "epoch": 0.48, + "objective/train/docs_used": 579278, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.2089581489562988, + "objective/train/theoretical_loss": 3.527404213670746, + "objective/train/tokens_used": 1606431200, + "theoretical_loss": 3.527404213670746, + "tokens_seen": 1585971200 + }, + { + "epoch": 0.48, + "learning_rate": 0.0004198194583751254, + "loss": 1.2745, + "theoretical_loss": 3.527404213670746, + "tokens_seen": 1585971200 + }, + { + "epoch": 0.48, + "learning_rate": 0.00041980341023069214, + "loss": 1.3726, + "theoretical_loss": 3.527391601983278, + "tokens_seen": 1586036736 + }, + { + "epoch": 0.48, + "learning_rate": 0.0004197873620862588, + "loss": 1.4078, + "theoretical_loss": 3.527378990962829, + "tokens_seen": 1586102272 + }, + { + "epoch": 0.48, + "learning_rate": 0.00041977131394182554, + "loss": 1.4259, + "theoretical_loss": 3.5273663806093367, + "tokens_seen": 1586167808 + }, + { + "epoch": 0.48, + "learning_rate": 0.0004197552657973922, + "loss": 1.2551, + "theoretical_loss": 3.527353770922738, + "tokens_seen": 1586233344 + }, + { + "epoch": 0.48, + "learning_rate": 0.0004197392176529589, + "loss": 1.478, + "theoretical_loss": 3.52734116190297, + "tokens_seen": 1586298880 + }, + { + "epoch": 0.48, + "learning_rate": 0.0004197231695085256, + "loss": 1.3115, + "theoretical_loss": 3.5273285535499697, + "tokens_seen": 1586364416 + }, + { + "epoch": 0.48, + "learning_rate": 0.0004197071213640923, + "loss": 1.3463, + "theoretical_loss": 3.5273159458636747, + "tokens_seen": 1586429952 + }, + { + "epoch": 0.48, + "learning_rate": 0.000419691073219659, + "loss": 1.3217, + "theoretical_loss": 3.527303338844022, + "tokens_seen": 1586495488 + }, + { + "epoch": 0.48, + "learning_rate": 0.0004196750250752257, + "loss": 1.4702, + "theoretical_loss": 3.527290732490949, + "tokens_seen": 1586561024 + }, + { + "epoch": 0.48, + "learning_rate": 0.0004196589769307924, + "loss": 1.284, + "theoretical_loss": 3.5272781268043927, + "tokens_seen": 1586626560 + }, + { + "epoch": 0.48, + "learning_rate": 0.0004196429287863591, + "loss": 1.322, + "theoretical_loss": 3.5272655217842908, + "tokens_seen": 1586692096 + }, + { + "epoch": 0.48, + "learning_rate": 0.0004196268806419258, + "loss": 1.4323, + "theoretical_loss": 3.5272529174305802, + "tokens_seen": 1586757632 + }, + { + "epoch": 0.48, + "learning_rate": 0.0004196108324974925, + "loss": 1.3287, + "theoretical_loss": 3.5272403137431976, + "tokens_seen": 1586823168 + }, + { + "epoch": 0.48, + "learning_rate": 0.00041959478435305914, + "loss": 1.4644, + "theoretical_loss": 3.527227710722081, + "tokens_seen": 1586888704 + }, + { + "epoch": 0.48, + "learning_rate": 0.0004195787362086259, + "loss": 1.4228, + "theoretical_loss": 3.5272151083671677, + "tokens_seen": 1586954240 + }, + { + "epoch": 0.48, + "learning_rate": 0.00041956268806419254, + "loss": 1.4535, + "theoretical_loss": 3.5272025066783943, + "tokens_seen": 1587019776 + }, + { + "epoch": 0.48, + "learning_rate": 0.0004195466399197593, + "loss": 1.4641, + "theoretical_loss": 3.527189905655699, + "tokens_seen": 1587085312 + }, + { + "epoch": 0.48, + "learning_rate": 0.00041953059177532594, + "loss": 1.3237, + "theoretical_loss": 3.527177305299018, + "tokens_seen": 1587150848 + }, + { + "epoch": 0.48, + "learning_rate": 0.0004195145436308927, + "loss": 1.3621, + "theoretical_loss": 3.5271647056082895, + "tokens_seen": 1587216384 + }, + { + "epoch": 0.48, + "learning_rate": 0.00041949849548645934, + "loss": 1.3511, + "theoretical_loss": 3.5271521065834506, + "tokens_seen": 1587281920 + }, + { + "epoch": 0.48, + "learning_rate": 0.0004194824473420261, + "loss": 1.3812, + "theoretical_loss": 3.5271395082244386, + "tokens_seen": 1587347456 + }, + { + "epoch": 0.48, + "learning_rate": 0.00041946639919759275, + "loss": 1.3482, + "theoretical_loss": 3.52712691053119, + "tokens_seen": 1587412992 + }, + { + "epoch": 0.48, + "learning_rate": 0.0004194503510531595, + "loss": 1.3599, + "theoretical_loss": 3.5271143135036436, + "tokens_seen": 1587478528 + }, + { + "epoch": 0.48, + "learning_rate": 0.00041943430290872615, + "loss": 1.3618, + "theoretical_loss": 3.5271017171417363, + "tokens_seen": 1587544064 + }, + { + "epoch": 0.48, + "learning_rate": 0.0004194182547642929, + "loss": 1.3377, + "theoretical_loss": 3.5270891214454045, + "tokens_seen": 1587609600 + }, + { + "epoch": 0.48, + "learning_rate": 0.00041940220661985955, + "loss": 1.403, + "theoretical_loss": 3.5270765264145867, + "tokens_seen": 1587675136 + }, + { + "epoch": 0.48, + "learning_rate": 0.0004193861584754263, + "loss": 1.2876, + "theoretical_loss": 3.527063932049219, + "tokens_seen": 1587740672 + }, + { + "epoch": 0.48, + "learning_rate": 0.00041937011033099295, + "loss": 1.3541, + "theoretical_loss": 3.5270513383492403, + "tokens_seen": 1587806208 + }, + { + "epoch": 0.48, + "learning_rate": 0.0004193540621865597, + "loss": 1.2872, + "theoretical_loss": 3.5270387453145875, + "tokens_seen": 1587871744 + }, + { + "epoch": 0.48, + "learning_rate": 0.00041933801404212635, + "loss": 1.2663, + "theoretical_loss": 3.527026152945197, + "tokens_seen": 1587937280 + }, + { + "epoch": 0.48, + "learning_rate": 0.0004193219658976931, + "loss": 1.2763, + "theoretical_loss": 3.5270135612410076, + "tokens_seen": 1588002816 + }, + { + "epoch": 0.48, + "learning_rate": 0.00041930591775325975, + "loss": 1.3109, + "theoretical_loss": 3.5270009702019562, + "tokens_seen": 1588068352 + }, + { + "epoch": 0.48, + "learning_rate": 0.0004192898696088265, + "loss": 1.4162, + "theoretical_loss": 3.5269883798279795, + "tokens_seen": 1588133888 + }, + { + "epoch": 0.48, + "learning_rate": 0.00041927382146439315, + "loss": 1.3688, + "theoretical_loss": 3.5269757901190157, + "tokens_seen": 1588199424 + }, + { + "epoch": 0.48, + "learning_rate": 0.0004192577733199599, + "loss": 1.2515, + "theoretical_loss": 3.5269632010750023, + "tokens_seen": 1588264960 + }, + { + "epoch": 0.48, + "learning_rate": 0.00041924172517552655, + "loss": 1.2893, + "theoretical_loss": 3.526950612695876, + "tokens_seen": 1588330496 + }, + { + "epoch": 0.48, + "learning_rate": 0.0004192256770310933, + "loss": 1.3537, + "theoretical_loss": 3.5269380249815754, + "tokens_seen": 1588396032 + }, + { + "epoch": 0.48, + "learning_rate": 0.00041920962888666006, + "loss": 1.3358, + "theoretical_loss": 3.5269254379320367, + "tokens_seen": 1588461568 + }, + { + "epoch": 0.48, + "learning_rate": 0.0004191935807422267, + "loss": 1.3257, + "theoretical_loss": 3.5269128515471984, + "tokens_seen": 1588527104 + }, + { + "epoch": 0.48, + "learning_rate": 0.00041917753259779347, + "loss": 1.3864, + "theoretical_loss": 3.5269002658269972, + "tokens_seen": 1588592640 + }, + { + "epoch": 0.48, + "learning_rate": 0.0004191614844533601, + "loss": 1.2809, + "theoretical_loss": 3.526887680771371, + "tokens_seen": 1588658176 + }, + { + "epoch": 0.48, + "learning_rate": 0.00041914543630892687, + "loss": 1.3272, + "theoretical_loss": 3.526875096380257, + "tokens_seen": 1588723712 + }, + { + "epoch": 0.48, + "learning_rate": 0.0004191293881644935, + "loss": 1.3312, + "theoretical_loss": 3.5268625126535937, + "tokens_seen": 1588789248 + }, + { + "epoch": 0.48, + "learning_rate": 0.00041911334002006027, + "loss": 1.3423, + "theoretical_loss": 3.526849929591317, + "tokens_seen": 1588854784 + }, + { + "epoch": 0.48, + "learning_rate": 0.0004190972918756269, + "loss": 1.383, + "theoretical_loss": 3.526837347193366, + "tokens_seen": 1588920320 + }, + { + "epoch": 0.48, + "learning_rate": 0.0004190812437311936, + "loss": 1.3478, + "theoretical_loss": 3.5268247654596774, + "tokens_seen": 1588985856 + }, + { + "epoch": 0.48, + "learning_rate": 0.0004190651955867603, + "loss": 1.3764, + "theoretical_loss": 3.5268121843901885, + "tokens_seen": 1589051392 + }, + { + "epoch": 0.48, + "learning_rate": 0.000419049147442327, + "loss": 1.4361, + "theoretical_loss": 3.526799603984837, + "tokens_seen": 1589116928 + }, + { + "epoch": 0.48, + "learning_rate": 0.0004190330992978937, + "loss": 1.3854, + "theoretical_loss": 3.526787024243561, + "tokens_seen": 1589182464 + }, + { + "epoch": 0.48, + "objective/train/docs_used": 580493, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.3065319061279297, + "objective/train/theoretical_loss": 3.5267744451662977, + "objective/train/tokens_used": 1609708000, + "theoretical_loss": 3.5267744451662977, + "tokens_seen": 1589248000 + }, + { + "epoch": 0.48, + "learning_rate": 0.0004190170511534604, + "loss": 1.285, + "theoretical_loss": 3.5267744451662977, + "tokens_seen": 1589248000 + }, + { + "epoch": 0.48, + "learning_rate": 0.0004190010030090271, + "loss": 1.2808, + "theoretical_loss": 3.5267618667529845, + "tokens_seen": 1589313536 + }, + { + "epoch": 0.48, + "learning_rate": 0.0004189849548645938, + "loss": 1.2988, + "theoretical_loss": 3.5267492890035594, + "tokens_seen": 1589379072 + }, + { + "epoch": 0.48, + "learning_rate": 0.00041896890672016047, + "loss": 1.3843, + "theoretical_loss": 3.52673671191796, + "tokens_seen": 1589444608 + }, + { + "epoch": 0.48, + "learning_rate": 0.0004189528585757272, + "loss": 1.3048, + "theoretical_loss": 3.5267241354961225, + "tokens_seen": 1589510144 + }, + { + "epoch": 0.48, + "learning_rate": 0.00041893681043129387, + "loss": 1.5119, + "theoretical_loss": 3.5267115597379863, + "tokens_seen": 1589575680 + }, + { + "epoch": 0.48, + "learning_rate": 0.0004189207622868606, + "loss": 1.3844, + "theoretical_loss": 3.5266989846434886, + "tokens_seen": 1589641216 + }, + { + "epoch": 0.48, + "learning_rate": 0.00041890471414242727, + "loss": 1.3655, + "theoretical_loss": 3.526686410212567, + "tokens_seen": 1589706752 + }, + { + "epoch": 0.48, + "learning_rate": 0.000418888665997994, + "loss": 1.2811, + "theoretical_loss": 3.526673836445158, + "tokens_seen": 1589772288 + }, + { + "epoch": 0.48, + "learning_rate": 0.00041887261785356067, + "loss": 1.341, + "theoretical_loss": 3.5266612633412007, + "tokens_seen": 1589837824 + }, + { + "epoch": 0.48, + "learning_rate": 0.0004188565697091274, + "loss": 1.2879, + "theoretical_loss": 3.526648690900632, + "tokens_seen": 1589903360 + }, + { + "epoch": 0.48, + "learning_rate": 0.00041884052156469407, + "loss": 1.2647, + "theoretical_loss": 3.52663611912339, + "tokens_seen": 1589968896 + }, + { + "epoch": 0.48, + "learning_rate": 0.0004188244734202608, + "loss": 1.3349, + "theoretical_loss": 3.526623548009412, + "tokens_seen": 1590034432 + }, + { + "epoch": 0.48, + "learning_rate": 0.00041880842527582747, + "loss": 1.2989, + "theoretical_loss": 3.5266109775586356, + "tokens_seen": 1590099968 + }, + { + "epoch": 0.48, + "learning_rate": 0.0004187923771313942, + "loss": 1.345, + "theoretical_loss": 3.5265984077709986, + "tokens_seen": 1590165504 + }, + { + "epoch": 0.48, + "learning_rate": 0.0004187763289869609, + "loss": 1.4203, + "theoretical_loss": 3.526585838646439, + "tokens_seen": 1590231040 + }, + { + "epoch": 0.48, + "learning_rate": 0.00041876028084252763, + "loss": 1.405, + "theoretical_loss": 3.526573270184894, + "tokens_seen": 1590296576 + }, + { + "epoch": 0.48, + "learning_rate": 0.0004187442326980943, + "loss": 1.3782, + "theoretical_loss": 3.526560702386302, + "tokens_seen": 1590362112 + }, + { + "epoch": 0.48, + "learning_rate": 0.00041872818455366103, + "loss": 1.4935, + "theoretical_loss": 3.5265481352505996, + "tokens_seen": 1590427648 + }, + { + "epoch": 0.48, + "learning_rate": 0.0004187121364092277, + "loss": 1.3135, + "theoretical_loss": 3.5265355687777253, + "tokens_seen": 1590493184 + }, + { + "epoch": 0.48, + "learning_rate": 0.00041869608826479443, + "loss": 1.3448, + "theoretical_loss": 3.5265230029676173, + "tokens_seen": 1590558720 + }, + { + "epoch": 0.48, + "learning_rate": 0.0004186800401203611, + "loss": 1.3956, + "theoretical_loss": 3.526510437820212, + "tokens_seen": 1590624256 + }, + { + "epoch": 0.48, + "learning_rate": 0.00041866399197592783, + "loss": 1.2929, + "theoretical_loss": 3.5264978733354484, + "tokens_seen": 1590689792 + }, + { + "epoch": 0.48, + "learning_rate": 0.0004186479438314945, + "loss": 1.3382, + "theoretical_loss": 3.526485309513264, + "tokens_seen": 1590755328 + }, + { + "epoch": 0.48, + "learning_rate": 0.00041863189568706123, + "loss": 1.3831, + "theoretical_loss": 3.526472746353596, + "tokens_seen": 1590820864 + }, + { + "epoch": 0.48, + "learning_rate": 0.0004186158475426279, + "loss": 1.4309, + "theoretical_loss": 3.5264601838563823, + "tokens_seen": 1590886400 + }, + { + "epoch": 0.48, + "learning_rate": 0.00041859979939819463, + "loss": 1.3962, + "theoretical_loss": 3.526447622021561, + "tokens_seen": 1590951936 + }, + { + "epoch": 0.48, + "learning_rate": 0.0004185837512537613, + "loss": 1.3276, + "theoretical_loss": 3.52643506084907, + "tokens_seen": 1591017472 + }, + { + "epoch": 0.48, + "learning_rate": 0.00041856770310932804, + "loss": 1.4589, + "theoretical_loss": 3.526422500338846, + "tokens_seen": 1591083008 + }, + { + "epoch": 0.48, + "learning_rate": 0.0004185516549648947, + "loss": 1.2993, + "theoretical_loss": 3.5264099404908285, + "tokens_seen": 1591148544 + }, + { + "epoch": 0.48, + "learning_rate": 0.00041853560682046144, + "loss": 1.3933, + "theoretical_loss": 3.5263973813049545, + "tokens_seen": 1591214080 + }, + { + "epoch": 0.48, + "learning_rate": 0.0004185195586760281, + "loss": 1.3391, + "theoretical_loss": 3.5263848227811616, + "tokens_seen": 1591279616 + }, + { + "epoch": 0.48, + "learning_rate": 0.00041850351053159484, + "loss": 1.3021, + "theoretical_loss": 3.526372264919388, + "tokens_seen": 1591345152 + }, + { + "epoch": 0.48, + "learning_rate": 0.0004184874623871615, + "loss": 1.2447, + "theoretical_loss": 3.526359707719571, + "tokens_seen": 1591410688 + }, + { + "epoch": 0.48, + "learning_rate": 0.00041847141424272824, + "loss": 1.3616, + "theoretical_loss": 3.5263471511816493, + "tokens_seen": 1591476224 + }, + { + "epoch": 0.48, + "learning_rate": 0.0004184553660982949, + "loss": 1.2173, + "theoretical_loss": 3.52633459530556, + "tokens_seen": 1591541760 + }, + { + "epoch": 0.48, + "learning_rate": 0.00041843931795386164, + "loss": 1.438, + "theoretical_loss": 3.5263220400912414, + "tokens_seen": 1591607296 + }, + { + "epoch": 0.48, + "learning_rate": 0.0004184232698094283, + "loss": 1.2429, + "theoretical_loss": 3.526309485538631, + "tokens_seen": 1591672832 + }, + { + "epoch": 0.48, + "learning_rate": 0.00041840722166499504, + "loss": 1.373, + "theoretical_loss": 3.5262969316476673, + "tokens_seen": 1591738368 + }, + { + "epoch": 0.48, + "learning_rate": 0.0004183911735205617, + "loss": 1.3448, + "theoretical_loss": 3.5262843784182873, + "tokens_seen": 1591803904 + }, + { + "epoch": 0.48, + "learning_rate": 0.00041837512537612844, + "loss": 1.2592, + "theoretical_loss": 3.5262718258504298, + "tokens_seen": 1591869440 + }, + { + "epoch": 0.48, + "learning_rate": 0.0004183590772316951, + "loss": 1.4519, + "theoretical_loss": 3.5262592739440324, + "tokens_seen": 1591934976 + }, + { + "epoch": 0.48, + "learning_rate": 0.00041834302908726184, + "loss": 1.3625, + "theoretical_loss": 3.5262467226990326, + "tokens_seen": 1592000512 + }, + { + "epoch": 0.48, + "learning_rate": 0.0004183269809428285, + "loss": 1.2441, + "theoretical_loss": 3.5262341721153687, + "tokens_seen": 1592066048 + }, + { + "epoch": 0.48, + "learning_rate": 0.0004183109327983952, + "loss": 1.3347, + "theoretical_loss": 3.5262216221929785, + "tokens_seen": 1592131584 + }, + { + "epoch": 0.48, + "learning_rate": 0.0004182948846539619, + "loss": 1.343, + "theoretical_loss": 3.5262090729318, + "tokens_seen": 1592197120 + }, + { + "epoch": 0.48, + "learning_rate": 0.0004182788365095286, + "loss": 1.2826, + "theoretical_loss": 3.5261965243317714, + "tokens_seen": 1592262656 + }, + { + "epoch": 0.48, + "learning_rate": 0.0004182627883650953, + "loss": 1.3071, + "theoretical_loss": 3.52618397639283, + "tokens_seen": 1592328192 + }, + { + "epoch": 0.48, + "learning_rate": 0.000418246740220662, + "loss": 1.3117, + "theoretical_loss": 3.526171429114915, + "tokens_seen": 1592393728 + }, + { + "epoch": 0.48, + "learning_rate": 0.0004182306920762287, + "loss": 1.2916, + "theoretical_loss": 3.526158882497963, + "tokens_seen": 1592459264 + }, + { + "epoch": 0.48, + "objective/train/docs_used": 581545, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.477599859237671, + "objective/train/theoretical_loss": 3.5261463365419123, + "objective/train/tokens_used": 1612984800, + "theoretical_loss": 3.5261463365419123, + "tokens_seen": 1592524800 + }, + { + "epoch": 0.48, + "learning_rate": 0.0004182146439317954, + "loss": 1.4221, + "theoretical_loss": 3.5261463365419123, + "tokens_seen": 1592524800 + }, + { + "epoch": 0.48, + "learning_rate": 0.00041819859578736204, + "loss": 1.3284, + "theoretical_loss": 3.5261337912467017, + "tokens_seen": 1592590336 + }, + { + "epoch": 0.48, + "learning_rate": 0.0004181825476429288, + "loss": 1.2529, + "theoretical_loss": 3.5261212466122682, + "tokens_seen": 1592655872 + }, + { + "epoch": 0.48, + "learning_rate": 0.00041816649949849544, + "loss": 1.3728, + "theoretical_loss": 3.5261087026385507, + "tokens_seen": 1592721408 + }, + { + "epoch": 0.48, + "learning_rate": 0.0004181504513540622, + "loss": 1.4599, + "theoretical_loss": 3.526096159325486, + "tokens_seen": 1592786944 + }, + { + "epoch": 0.48, + "learning_rate": 0.00041813440320962884, + "loss": 1.3698, + "theoretical_loss": 3.526083616673014, + "tokens_seen": 1592852480 + }, + { + "epoch": 0.48, + "learning_rate": 0.0004181183550651956, + "loss": 1.3859, + "theoretical_loss": 3.526071074681071, + "tokens_seen": 1592918016 + }, + { + "epoch": 0.48, + "learning_rate": 0.00041810230692076225, + "loss": 1.4153, + "theoretical_loss": 3.5260585333495955, + "tokens_seen": 1592983552 + }, + { + "epoch": 0.48, + "learning_rate": 0.000418086258776329, + "loss": 1.2893, + "theoretical_loss": 3.526045992678526, + "tokens_seen": 1593049088 + }, + { + "epoch": 0.48, + "learning_rate": 0.00041807021063189565, + "loss": 1.4116, + "theoretical_loss": 3.5260334526678, + "tokens_seen": 1593114624 + }, + { + "epoch": 0.48, + "learning_rate": 0.0004180541624874624, + "loss": 1.3019, + "theoretical_loss": 3.526020913317356, + "tokens_seen": 1593180160 + }, + { + "epoch": 0.48, + "learning_rate": 0.00041803811434302905, + "loss": 1.2887, + "theoretical_loss": 3.526008374627132, + "tokens_seen": 1593245696 + }, + { + "epoch": 0.48, + "learning_rate": 0.0004180220661985958, + "loss": 1.3489, + "theoretical_loss": 3.525995836597066, + "tokens_seen": 1593311232 + }, + { + "epoch": 0.48, + "learning_rate": 0.00041800601805416256, + "loss": 1.3501, + "theoretical_loss": 3.5259832992270956, + "tokens_seen": 1593376768 + }, + { + "epoch": 0.48, + "learning_rate": 0.0004179899699097292, + "loss": 1.3733, + "theoretical_loss": 3.52597076251716, + "tokens_seen": 1593442304 + }, + { + "epoch": 0.48, + "learning_rate": 0.00041797392176529596, + "loss": 1.3151, + "theoretical_loss": 3.5259582264671963, + "tokens_seen": 1593507840 + }, + { + "epoch": 0.48, + "learning_rate": 0.0004179578736208626, + "loss": 1.4227, + "theoretical_loss": 3.525945691077143, + "tokens_seen": 1593573376 + }, + { + "epoch": 0.48, + "learning_rate": 0.00041794182547642936, + "loss": 1.4268, + "theoretical_loss": 3.5259331563469387, + "tokens_seen": 1593638912 + }, + { + "epoch": 0.48, + "learning_rate": 0.000417925777331996, + "loss": 1.4057, + "theoretical_loss": 3.525920622276521, + "tokens_seen": 1593704448 + }, + { + "epoch": 0.48, + "learning_rate": 0.00041790972918756276, + "loss": 1.3987, + "theoretical_loss": 3.525908088865828, + "tokens_seen": 1593769984 + }, + { + "epoch": 0.48, + "learning_rate": 0.0004178936810431294, + "loss": 1.3998, + "theoretical_loss": 3.5258955561147975, + "tokens_seen": 1593835520 + }, + { + "epoch": 0.48, + "learning_rate": 0.00041787763289869616, + "loss": 1.3776, + "theoretical_loss": 3.5258830240233685, + "tokens_seen": 1593901056 + }, + { + "epoch": 0.48, + "learning_rate": 0.0004178615847542628, + "loss": 1.2671, + "theoretical_loss": 3.525870492591479, + "tokens_seen": 1593966592 + }, + { + "epoch": 0.48, + "learning_rate": 0.00041784553660982956, + "loss": 1.3681, + "theoretical_loss": 3.5258579618190664, + "tokens_seen": 1594032128 + }, + { + "epoch": 0.48, + "learning_rate": 0.0004178294884653962, + "loss": 1.4044, + "theoretical_loss": 3.5258454317060695, + "tokens_seen": 1594097664 + }, + { + "epoch": 0.48, + "learning_rate": 0.00041781344032096297, + "loss": 1.3923, + "theoretical_loss": 3.525832902252427, + "tokens_seen": 1594163200 + }, + { + "epoch": 0.48, + "learning_rate": 0.0004177973921765296, + "loss": 1.2781, + "theoretical_loss": 3.5258203734580764, + "tokens_seen": 1594228736 + }, + { + "epoch": 0.48, + "learning_rate": 0.00041778134403209637, + "loss": 1.3584, + "theoretical_loss": 3.5258078453229555, + "tokens_seen": 1594294272 + }, + { + "epoch": 0.48, + "learning_rate": 0.000417765295887663, + "loss": 1.2373, + "theoretical_loss": 3.5257953178470034, + "tokens_seen": 1594359808 + }, + { + "epoch": 0.48, + "learning_rate": 0.00041774924774322977, + "loss": 1.3263, + "theoretical_loss": 3.5257827910301582, + "tokens_seen": 1594425344 + }, + { + "epoch": 0.48, + "learning_rate": 0.0004177331995987964, + "loss": 1.3223, + "theoretical_loss": 3.5257702648723575, + "tokens_seen": 1594490880 + }, + { + "epoch": 0.48, + "learning_rate": 0.00041771715145436317, + "loss": 1.3446, + "theoretical_loss": 3.5257577393735406, + "tokens_seen": 1594556416 + }, + { + "epoch": 0.48, + "learning_rate": 0.0004177011033099298, + "loss": 1.3109, + "theoretical_loss": 3.5257452145336443, + "tokens_seen": 1594621952 + }, + { + "epoch": 0.48, + "learning_rate": 0.0004176850551654965, + "loss": 1.3096, + "theoretical_loss": 3.525732690352608, + "tokens_seen": 1594687488 + }, + { + "epoch": 0.48, + "learning_rate": 0.0004176690070210632, + "loss": 1.4024, + "theoretical_loss": 3.5257201668303697, + "tokens_seen": 1594753024 + }, + { + "epoch": 0.48, + "learning_rate": 0.0004176529588766299, + "loss": 1.3834, + "theoretical_loss": 3.525707643966868, + "tokens_seen": 1594818560 + }, + { + "epoch": 0.48, + "learning_rate": 0.0004176369107321966, + "loss": 1.368, + "theoretical_loss": 3.5256951217620403, + "tokens_seen": 1594884096 + }, + { + "epoch": 0.48, + "learning_rate": 0.0004176208625877633, + "loss": 1.3301, + "theoretical_loss": 3.5256826002158252, + "tokens_seen": 1594949632 + }, + { + "epoch": 0.48, + "learning_rate": 0.00041760481444333, + "loss": 1.2742, + "theoretical_loss": 3.5256700793281617, + "tokens_seen": 1595015168 + }, + { + "epoch": 0.48, + "learning_rate": 0.0004175887662988967, + "loss": 1.1465, + "theoretical_loss": 3.5256575590989874, + "tokens_seen": 1595080704 + }, + { + "epoch": 0.48, + "learning_rate": 0.0004175727181544634, + "loss": 1.3467, + "theoretical_loss": 3.5256450395282406, + "tokens_seen": 1595146240 + }, + { + "epoch": 0.48, + "learning_rate": 0.0004175566700100301, + "loss": 1.3167, + "theoretical_loss": 3.52563252061586, + "tokens_seen": 1595211776 + }, + { + "epoch": 0.48, + "learning_rate": 0.00041754062186559677, + "loss": 1.2422, + "theoretical_loss": 3.5256200023617836, + "tokens_seen": 1595277312 + }, + { + "epoch": 0.48, + "learning_rate": 0.0004175245737211635, + "loss": 1.2639, + "theoretical_loss": 3.52560748476595, + "tokens_seen": 1595342848 + }, + { + "epoch": 0.48, + "learning_rate": 0.00041750852557673017, + "loss": 1.3618, + "theoretical_loss": 3.5255949678282974, + "tokens_seen": 1595408384 + }, + { + "epoch": 0.48, + "learning_rate": 0.0004174924774322969, + "loss": 1.3998, + "theoretical_loss": 3.5255824515487646, + "tokens_seen": 1595473920 + }, + { + "epoch": 0.48, + "learning_rate": 0.00041747642928786357, + "loss": 1.3902, + "theoretical_loss": 3.5255699359272894, + "tokens_seen": 1595539456 + }, + { + "epoch": 0.48, + "learning_rate": 0.0004174603811434303, + "loss": 1.311, + "theoretical_loss": 3.52555742096381, + "tokens_seen": 1595604992 + }, + { + "epoch": 0.48, + "learning_rate": 0.00041744433299899697, + "loss": 1.2348, + "theoretical_loss": 3.5255449066582654, + "tokens_seen": 1595670528 + }, + { + "epoch": 0.48, + "learning_rate": 0.0004174282848545637, + "loss": 1.2535, + "theoretical_loss": 3.525532393010594, + "tokens_seen": 1595736064 + }, + { + "epoch": 0.48, + "objective/train/docs_used": 582715, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.3268332481384277, + "objective/train/theoretical_loss": 3.5255198800207332, + "objective/train/tokens_used": 1616261600, + "theoretical_loss": 3.5255198800207332, + "tokens_seen": 1595801600 + }, + { + "epoch": 0.48, + "learning_rate": 0.00041741223671013037, + "loss": 1.2637, + "theoretical_loss": 3.5255198800207332, + "tokens_seen": 1595801600 + }, + { + "epoch": 0.48, + "learning_rate": 0.00041739618856569713, + "loss": 1.3165, + "theoretical_loss": 3.525507367688623, + "tokens_seen": 1595867136 + }, + { + "epoch": 0.48, + "learning_rate": 0.0004173801404212638, + "loss": 1.2628, + "theoretical_loss": 3.5254948560142, + "tokens_seen": 1595932672 + }, + { + "epoch": 0.48, + "learning_rate": 0.00041736409227683053, + "loss": 1.2986, + "theoretical_loss": 3.525482344997404, + "tokens_seen": 1595998208 + }, + { + "epoch": 0.48, + "learning_rate": 0.0004173480441323972, + "loss": 1.319, + "theoretical_loss": 3.5254698346381734, + "tokens_seen": 1596063744 + }, + { + "epoch": 0.48, + "learning_rate": 0.00041733199598796393, + "loss": 1.4425, + "theoretical_loss": 3.525457324936446, + "tokens_seen": 1596129280 + }, + { + "epoch": 0.48, + "learning_rate": 0.0004173159478435306, + "loss": 1.3722, + "theoretical_loss": 3.5254448158921603, + "tokens_seen": 1596194816 + }, + { + "epoch": 0.48, + "learning_rate": 0.00041729989969909733, + "loss": 1.313, + "theoretical_loss": 3.5254323075052554, + "tokens_seen": 1596260352 + }, + { + "epoch": 0.48, + "learning_rate": 0.000417283851554664, + "loss": 1.3437, + "theoretical_loss": 3.525419799775669, + "tokens_seen": 1596325888 + }, + { + "epoch": 0.48, + "learning_rate": 0.00041726780341023073, + "loss": 1.2587, + "theoretical_loss": 3.52540729270334, + "tokens_seen": 1596391424 + }, + { + "epoch": 0.48, + "learning_rate": 0.0004172517552657974, + "loss": 1.2447, + "theoretical_loss": 3.5253947862882065, + "tokens_seen": 1596456960 + }, + { + "epoch": 0.48, + "learning_rate": 0.00041723570712136413, + "loss": 1.3366, + "theoretical_loss": 3.5253822805302075, + "tokens_seen": 1596522496 + }, + { + "epoch": 0.48, + "learning_rate": 0.0004172196589769308, + "loss": 1.4161, + "theoretical_loss": 3.525369775429281, + "tokens_seen": 1596588032 + }, + { + "epoch": 0.48, + "learning_rate": 0.00041720361083249754, + "loss": 1.3057, + "theoretical_loss": 3.525357270985366, + "tokens_seen": 1596653568 + }, + { + "epoch": 0.48, + "learning_rate": 0.0004171875626880642, + "loss": 1.3608, + "theoretical_loss": 3.5253447671984013, + "tokens_seen": 1596719104 + }, + { + "epoch": 0.48, + "learning_rate": 0.00041717151454363094, + "loss": 1.2972, + "theoretical_loss": 3.5253322640683242, + "tokens_seen": 1596784640 + }, + { + "epoch": 0.48, + "learning_rate": 0.0004171554663991976, + "loss": 1.3823, + "theoretical_loss": 3.5253197615950747, + "tokens_seen": 1596850176 + }, + { + "epoch": 0.48, + "learning_rate": 0.00041713941825476434, + "loss": 1.2895, + "theoretical_loss": 3.5253072597785895, + "tokens_seen": 1596915712 + }, + { + "epoch": 0.48, + "learning_rate": 0.000417123370110331, + "loss": 1.2923, + "theoretical_loss": 3.5252947586188093, + "tokens_seen": 1596981248 + }, + { + "epoch": 0.48, + "learning_rate": 0.00041710732196589774, + "loss": 1.378, + "theoretical_loss": 3.525282258115671, + "tokens_seen": 1597046784 + }, + { + "epoch": 0.48, + "learning_rate": 0.0004170912738214644, + "loss": 1.3364, + "theoretical_loss": 3.525269758269114, + "tokens_seen": 1597112320 + }, + { + "epoch": 0.48, + "learning_rate": 0.00041707522567703114, + "loss": 1.3291, + "theoretical_loss": 3.5252572590790763, + "tokens_seen": 1597177856 + }, + { + "epoch": 0.48, + "learning_rate": 0.0004170591775325978, + "loss": 1.2897, + "theoretical_loss": 3.5252447605454966, + "tokens_seen": 1597243392 + }, + { + "epoch": 0.48, + "learning_rate": 0.00041704312938816454, + "loss": 1.2634, + "theoretical_loss": 3.5252322626683146, + "tokens_seen": 1597308928 + }, + { + "epoch": 0.48, + "learning_rate": 0.0004170270812437312, + "loss": 1.2626, + "theoretical_loss": 3.525219765447467, + "tokens_seen": 1597374464 + }, + { + "epoch": 0.48, + "learning_rate": 0.00041701103309929794, + "loss": 1.3883, + "theoretical_loss": 3.525207268882894, + "tokens_seen": 1597440000 + }, + { + "epoch": 0.48, + "learning_rate": 0.0004169949849548646, + "loss": 1.4219, + "theoretical_loss": 3.5251947729745337, + "tokens_seen": 1597505536 + }, + { + "epoch": 0.48, + "learning_rate": 0.00041697893681043134, + "loss": 1.3013, + "theoretical_loss": 3.5251822777223243, + "tokens_seen": 1597571072 + }, + { + "epoch": 0.48, + "learning_rate": 0.000416962888665998, + "loss": 1.2894, + "theoretical_loss": 3.525169783126205, + "tokens_seen": 1597636608 + }, + { + "epoch": 0.48, + "learning_rate": 0.00041694684052156475, + "loss": 1.2666, + "theoretical_loss": 3.5251572891861143, + "tokens_seen": 1597702144 + }, + { + "epoch": 0.48, + "learning_rate": 0.0004169307923771314, + "loss": 1.4198, + "theoretical_loss": 3.5251447959019906, + "tokens_seen": 1597767680 + }, + { + "epoch": 0.48, + "learning_rate": 0.0004169147442326981, + "loss": 1.3143, + "theoretical_loss": 3.525132303273773, + "tokens_seen": 1597833216 + }, + { + "epoch": 0.48, + "learning_rate": 0.0004168986960882648, + "loss": 1.322, + "theoretical_loss": 3.5251198113013995, + "tokens_seen": 1597898752 + }, + { + "epoch": 0.48, + "learning_rate": 0.0004168826479438315, + "loss": 1.3403, + "theoretical_loss": 3.5251073199848095, + "tokens_seen": 1597964288 + }, + { + "epoch": 0.48, + "learning_rate": 0.0004168665997993982, + "loss": 1.3127, + "theoretical_loss": 3.5250948293239412, + "tokens_seen": 1598029824 + }, + { + "epoch": 0.48, + "learning_rate": 0.0004168505516549649, + "loss": 1.311, + "theoretical_loss": 3.5250823393187334, + "tokens_seen": 1598095360 + }, + { + "epoch": 0.48, + "learning_rate": 0.0004168345035105316, + "loss": 1.2698, + "theoretical_loss": 3.5250698499691246, + "tokens_seen": 1598160896 + }, + { + "epoch": 0.48, + "learning_rate": 0.0004168184553660983, + "loss": 1.2932, + "theoretical_loss": 3.5250573612750538, + "tokens_seen": 1598226432 + }, + { + "epoch": 0.48, + "learning_rate": 0.00041680240722166505, + "loss": 1.2685, + "theoretical_loss": 3.52504487323646, + "tokens_seen": 1598291968 + }, + { + "epoch": 0.48, + "learning_rate": 0.0004167863590772317, + "loss": 1.3893, + "theoretical_loss": 3.5250323858532813, + "tokens_seen": 1598357504 + }, + { + "epoch": 0.48, + "learning_rate": 0.00041677031093279845, + "loss": 1.4203, + "theoretical_loss": 3.5250198991254567, + "tokens_seen": 1598423040 + }, + { + "epoch": 0.48, + "learning_rate": 0.0004167542627883651, + "loss": 1.3266, + "theoretical_loss": 3.5250074130529248, + "tokens_seen": 1598488576 + }, + { + "epoch": 0.48, + "learning_rate": 0.00041673821464393185, + "loss": 1.36, + "theoretical_loss": 3.5249949276356247, + "tokens_seen": 1598554112 + }, + { + "epoch": 0.48, + "learning_rate": 0.0004167221664994985, + "loss": 1.3836, + "theoretical_loss": 3.524982442873495, + "tokens_seen": 1598619648 + }, + { + "epoch": 0.48, + "learning_rate": 0.00041670611835506525, + "loss": 1.3517, + "theoretical_loss": 3.5249699587664742, + "tokens_seen": 1598685184 + }, + { + "epoch": 0.48, + "learning_rate": 0.0004166900702106319, + "loss": 1.3984, + "theoretical_loss": 3.524957475314501, + "tokens_seen": 1598750720 + }, + { + "epoch": 0.48, + "learning_rate": 0.00041667402206619866, + "loss": 1.3443, + "theoretical_loss": 3.524944992517515, + "tokens_seen": 1598816256 + }, + { + "epoch": 0.48, + "learning_rate": 0.0004166579739217653, + "loss": 1.4467, + "theoretical_loss": 3.5249325103754545, + "tokens_seen": 1598881792 + }, + { + "epoch": 0.48, + "learning_rate": 0.00041664192577733206, + "loss": 1.2495, + "theoretical_loss": 3.524920028888258, + "tokens_seen": 1598947328 + }, + { + "epoch": 0.48, + "learning_rate": 0.0004166258776328987, + "loss": 1.339, + "theoretical_loss": 3.5249075480558645, + "tokens_seen": 1599012864 + }, + { + "epoch": 0.48, + "objective/train/docs_used": 583998, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.4502851963043213, + "objective/train/theoretical_loss": 3.5248950678782127, + "objective/train/tokens_used": 1619538400, + "theoretical_loss": 3.5248950678782127, + "tokens_seen": 1599078400 + }, + { + "epoch": 0.48, + "learning_rate": 0.00041660982948846546, + "loss": 1.3556, + "theoretical_loss": 3.5248950678782127, + "tokens_seen": 1599078400 + }, + { + "epoch": 0.48, + "learning_rate": 0.0004165937813440321, + "loss": 1.2842, + "theoretical_loss": 3.5248825883552417, + "tokens_seen": 1599143936 + }, + { + "epoch": 0.48, + "learning_rate": 0.00041657773319959886, + "loss": 1.2919, + "theoretical_loss": 3.5248701094868906, + "tokens_seen": 1599209472 + }, + { + "epoch": 0.48, + "learning_rate": 0.0004165616850551655, + "loss": 1.4566, + "theoretical_loss": 3.5248576312730973, + "tokens_seen": 1599275008 + }, + { + "epoch": 0.48, + "learning_rate": 0.00041654563691073226, + "loss": 1.3051, + "theoretical_loss": 3.5248451537138012, + "tokens_seen": 1599340544 + }, + { + "epoch": 0.48, + "learning_rate": 0.0004165295887662989, + "loss": 1.389, + "theoretical_loss": 3.5248326768089413, + "tokens_seen": 1599406080 + }, + { + "epoch": 0.48, + "learning_rate": 0.00041651354062186566, + "loss": 1.3988, + "theoretical_loss": 3.524820200558456, + "tokens_seen": 1599471616 + }, + { + "epoch": 0.48, + "learning_rate": 0.0004164974924774323, + "loss": 1.3401, + "theoretical_loss": 3.524807724962285, + "tokens_seen": 1599537152 + }, + { + "epoch": 0.48, + "learning_rate": 0.00041648144433299906, + "loss": 1.3993, + "theoretical_loss": 3.5247952500203663, + "tokens_seen": 1599602688 + }, + { + "epoch": 0.48, + "learning_rate": 0.0004164653961885657, + "loss": 1.3557, + "theoretical_loss": 3.5247827757326395, + "tokens_seen": 1599668224 + }, + { + "epoch": 0.48, + "learning_rate": 0.00041644934804413247, + "loss": 1.3711, + "theoretical_loss": 3.524770302099043, + "tokens_seen": 1599733760 + }, + { + "epoch": 0.48, + "learning_rate": 0.0004164332998996991, + "loss": 1.359, + "theoretical_loss": 3.5247578291195154, + "tokens_seen": 1599799296 + }, + { + "epoch": 0.48, + "learning_rate": 0.00041641725175526587, + "loss": 1.4394, + "theoretical_loss": 3.5247453567939964, + "tokens_seen": 1599864832 + }, + { + "epoch": 0.48, + "learning_rate": 0.0004164012036108325, + "loss": 1.4532, + "theoretical_loss": 3.5247328851224244, + "tokens_seen": 1599930368 + }, + { + "epoch": 0.48, + "learning_rate": 0.00041638515546639927, + "loss": 1.3368, + "theoretical_loss": 3.5247204141047392, + "tokens_seen": 1599995904 + }, + { + "epoch": 0.48, + "learning_rate": 0.0004163691073219659, + "loss": 1.3209, + "theoretical_loss": 3.524707943740878, + "tokens_seen": 1600061440 + }, + { + "epoch": 0.48, + "learning_rate": 0.00041635305917753267, + "loss": 1.337, + "theoretical_loss": 3.524695474030781, + "tokens_seen": 1600126976 + }, + { + "epoch": 0.48, + "learning_rate": 0.0004163370110330993, + "loss": 1.4324, + "theoretical_loss": 3.5246830049743876, + "tokens_seen": 1600192512 + }, + { + "epoch": 0.48, + "learning_rate": 0.00041632096288866607, + "loss": 1.3959, + "theoretical_loss": 3.5246705365716355, + "tokens_seen": 1600258048 + }, + { + "epoch": 0.48, + "learning_rate": 0.0004163049147442327, + "loss": 1.3323, + "theoretical_loss": 3.5246580688224647, + "tokens_seen": 1600323584 + }, + { + "epoch": 0.48, + "learning_rate": 0.0004162888665997994, + "loss": 1.2801, + "theoretical_loss": 3.524645601726813, + "tokens_seen": 1600389120 + }, + { + "epoch": 0.48, + "learning_rate": 0.0004162728184553661, + "loss": 1.472, + "theoretical_loss": 3.5246331352846205, + "tokens_seen": 1600454656 + }, + { + "epoch": 0.49, + "learning_rate": 0.0004162567703109328, + "loss": 1.2984, + "theoretical_loss": 3.524620669495826, + "tokens_seen": 1600520192 + }, + { + "epoch": 0.49, + "learning_rate": 0.0004162407221664995, + "loss": 1.4682, + "theoretical_loss": 3.524608204360368, + "tokens_seen": 1600585728 + }, + { + "epoch": 0.49, + "learning_rate": 0.0004162246740220662, + "loss": 1.4593, + "theoretical_loss": 3.524595739878186, + "tokens_seen": 1600651264 + }, + { + "epoch": 0.49, + "learning_rate": 0.0004162086258776329, + "loss": 1.2832, + "theoretical_loss": 3.524583276049219, + "tokens_seen": 1600716800 + }, + { + "epoch": 0.49, + "learning_rate": 0.0004161925777331996, + "loss": 1.3089, + "theoretical_loss": 3.5245708128734057, + "tokens_seen": 1600782336 + }, + { + "epoch": 0.49, + "learning_rate": 0.0004161765295887663, + "loss": 1.3837, + "theoretical_loss": 3.524558350350685, + "tokens_seen": 1600847872 + }, + { + "epoch": 0.49, + "learning_rate": 0.000416160481444333, + "loss": 1.4358, + "theoretical_loss": 3.524545888480996, + "tokens_seen": 1600913408 + }, + { + "epoch": 0.49, + "learning_rate": 0.00041614443329989967, + "loss": 1.3408, + "theoretical_loss": 3.524533427264279, + "tokens_seen": 1600978944 + }, + { + "epoch": 0.49, + "learning_rate": 0.0004161283851554664, + "loss": 1.3652, + "theoretical_loss": 3.5245209667004715, + "tokens_seen": 1601044480 + }, + { + "epoch": 0.49, + "learning_rate": 0.00041611233701103307, + "loss": 1.3824, + "theoretical_loss": 3.524508506789513, + "tokens_seen": 1601110016 + }, + { + "epoch": 0.49, + "learning_rate": 0.0004160962888665998, + "loss": 1.305, + "theoretical_loss": 3.5244960475313425, + "tokens_seen": 1601175552 + }, + { + "epoch": 0.49, + "learning_rate": 0.00041608024072216647, + "loss": 1.3301, + "theoretical_loss": 3.5244835889259, + "tokens_seen": 1601241088 + }, + { + "epoch": 0.49, + "learning_rate": 0.0004160641925777332, + "loss": 1.4558, + "theoretical_loss": 3.524471130973123, + "tokens_seen": 1601306624 + }, + { + "epoch": 0.49, + "learning_rate": 0.00041604814443329987, + "loss": 1.2261, + "theoretical_loss": 3.524458673672952, + "tokens_seen": 1601372160 + }, + { + "epoch": 0.49, + "learning_rate": 0.00041603209628886663, + "loss": 1.4102, + "theoretical_loss": 3.524446217025326, + "tokens_seen": 1601437696 + }, + { + "epoch": 0.49, + "learning_rate": 0.0004160160481444333, + "loss": 1.2302, + "theoretical_loss": 3.524433761030183, + "tokens_seen": 1601503232 + }, + { + "epoch": 0.49, + "learning_rate": 0.00041600000000000003, + "loss": 1.29, + "theoretical_loss": 3.524421305687463, + "tokens_seen": 1601568768 + }, + { + "epoch": 0.49, + "learning_rate": 0.0004159839518555667, + "loss": 1.4312, + "theoretical_loss": 3.524408850997105, + "tokens_seen": 1601634304 + }, + { + "epoch": 0.49, + "learning_rate": 0.00041596790371113343, + "loss": 1.4383, + "theoretical_loss": 3.5243963969590477, + "tokens_seen": 1601699840 + }, + { + "epoch": 0.49, + "learning_rate": 0.0004159518555667001, + "loss": 1.3599, + "theoretical_loss": 3.524383943573231, + "tokens_seen": 1601765376 + }, + { + "epoch": 0.49, + "learning_rate": 0.00041593580742226683, + "loss": 1.3406, + "theoretical_loss": 3.5243714908395933, + "tokens_seen": 1601830912 + }, + { + "epoch": 0.49, + "learning_rate": 0.0004159197592778335, + "loss": 1.3937, + "theoretical_loss": 3.5243590387580745, + "tokens_seen": 1601896448 + }, + { + "epoch": 0.49, + "learning_rate": 0.00041590371113340023, + "loss": 1.4337, + "theoretical_loss": 3.5243465873286137, + "tokens_seen": 1601961984 + }, + { + "epoch": 0.49, + "learning_rate": 0.0004158876629889669, + "loss": 1.3914, + "theoretical_loss": 3.5243341365511496, + "tokens_seen": 1602027520 + }, + { + "epoch": 0.49, + "learning_rate": 0.00041587161484453363, + "loss": 1.3074, + "theoretical_loss": 3.5243216864256213, + "tokens_seen": 1602093056 + }, + { + "epoch": 0.49, + "learning_rate": 0.0004158555667001003, + "loss": 1.5127, + "theoretical_loss": 3.524309236951969, + "tokens_seen": 1602158592 + }, + { + "epoch": 0.49, + "learning_rate": 0.00041583951855566703, + "loss": 1.3321, + "theoretical_loss": 3.5242967881301306, + "tokens_seen": 1602224128 + }, + { + "epoch": 0.49, + "learning_rate": 0.0004158234704112337, + "loss": 1.2964, + "theoretical_loss": 3.5242843399600456, + "tokens_seen": 1602289664 + }, + { + "epoch": 0.49, + "objective/train/docs_used": 585126, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.3763723373413086, + "objective/train/theoretical_loss": 3.524271892441654, + "objective/train/tokens_used": 1622815200, + "theoretical_loss": 3.524271892441654, + "tokens_seen": 1602355200 + }, + { + "epoch": 0.49, + "learning_rate": 0.00041580742226680044, + "loss": 1.3647, + "theoretical_loss": 3.524271892441654, + "tokens_seen": 1602355200 + }, + { + "epoch": 0.49, + "learning_rate": 0.0004157913741223671, + "loss": 1.3639, + "theoretical_loss": 3.5242594455748946, + "tokens_seen": 1602420736 + }, + { + "epoch": 0.49, + "learning_rate": 0.00041577532597793384, + "loss": 1.2745, + "theoretical_loss": 3.5242469993597068, + "tokens_seen": 1602486272 + }, + { + "epoch": 0.49, + "learning_rate": 0.0004157592778335005, + "loss": 1.3573, + "theoretical_loss": 3.5242345537960293, + "tokens_seen": 1602551808 + }, + { + "epoch": 0.49, + "learning_rate": 0.00041574322968906724, + "loss": 1.3598, + "theoretical_loss": 3.5242221088838015, + "tokens_seen": 1602617344 + }, + { + "epoch": 0.49, + "learning_rate": 0.0004157271815446339, + "loss": 1.3259, + "theoretical_loss": 3.5242096646229633, + "tokens_seen": 1602682880 + }, + { + "epoch": 0.49, + "learning_rate": 0.00041571113340020064, + "loss": 1.3801, + "theoretical_loss": 3.5241972210134533, + "tokens_seen": 1602748416 + }, + { + "epoch": 0.49, + "learning_rate": 0.0004156950852557673, + "loss": 1.3318, + "theoretical_loss": 3.5241847780552114, + "tokens_seen": 1602813952 + }, + { + "epoch": 0.49, + "learning_rate": 0.00041567903711133404, + "loss": 1.2904, + "theoretical_loss": 3.524172335748176, + "tokens_seen": 1602879488 + }, + { + "epoch": 0.49, + "learning_rate": 0.0004156629889669007, + "loss": 1.1287, + "theoretical_loss": 3.5241598940922874, + "tokens_seen": 1602945024 + }, + { + "epoch": 0.49, + "learning_rate": 0.00041564694082246744, + "loss": 1.3841, + "theoretical_loss": 3.524147453087484, + "tokens_seen": 1603010560 + }, + { + "epoch": 0.49, + "learning_rate": 0.0004156308926780341, + "loss": 1.4553, + "theoretical_loss": 3.524135012733706, + "tokens_seen": 1603076096 + }, + { + "epoch": 0.49, + "learning_rate": 0.00041561484453360084, + "loss": 1.4079, + "theoretical_loss": 3.5241225730308914, + "tokens_seen": 1603141632 + }, + { + "epoch": 0.49, + "learning_rate": 0.00041559879638916754, + "loss": 1.2818, + "theoretical_loss": 3.524110133978981, + "tokens_seen": 1603207168 + }, + { + "epoch": 0.49, + "learning_rate": 0.00041558274824473425, + "loss": 1.2489, + "theoretical_loss": 3.5240976955779137, + "tokens_seen": 1603272704 + }, + { + "epoch": 0.49, + "learning_rate": 0.00041556670010030095, + "loss": 1.3066, + "theoretical_loss": 3.524085257827628, + "tokens_seen": 1603338240 + }, + { + "epoch": 0.49, + "learning_rate": 0.00041555065195586765, + "loss": 1.3957, + "theoretical_loss": 3.5240728207280645, + "tokens_seen": 1603403776 + }, + { + "epoch": 0.49, + "learning_rate": 0.00041553460381143435, + "loss": 1.336, + "theoretical_loss": 3.5240603842791614, + "tokens_seen": 1603469312 + }, + { + "epoch": 0.49, + "learning_rate": 0.000415518555667001, + "loss": 1.3768, + "theoretical_loss": 3.5240479484808587, + "tokens_seen": 1603534848 + }, + { + "epoch": 0.49, + "learning_rate": 0.00041550250752256775, + "loss": 1.3433, + "theoretical_loss": 3.524035513333096, + "tokens_seen": 1603600384 + }, + { + "epoch": 0.49, + "learning_rate": 0.0004154864593781344, + "loss": 1.3372, + "theoretical_loss": 3.5240230788358122, + "tokens_seen": 1603665920 + }, + { + "epoch": 0.49, + "learning_rate": 0.00041547041123370115, + "loss": 1.4249, + "theoretical_loss": 3.5240106449889472, + "tokens_seen": 1603731456 + }, + { + "epoch": 0.49, + "learning_rate": 0.0004154543630892678, + "loss": 1.3649, + "theoretical_loss": 3.5239982117924398, + "tokens_seen": 1603796992 + }, + { + "epoch": 0.49, + "learning_rate": 0.00041543831494483455, + "loss": 1.2545, + "theoretical_loss": 3.52398577924623, + "tokens_seen": 1603862528 + }, + { + "epoch": 0.49, + "learning_rate": 0.0004154222668004012, + "loss": 1.4424, + "theoretical_loss": 3.5239733473502564, + "tokens_seen": 1603928064 + }, + { + "epoch": 0.49, + "learning_rate": 0.00041540621865596795, + "loss": 1.3564, + "theoretical_loss": 3.5239609161044596, + "tokens_seen": 1603993600 + }, + { + "epoch": 0.49, + "learning_rate": 0.0004153901705115346, + "loss": 1.3867, + "theoretical_loss": 3.5239484855087775, + "tokens_seen": 1604059136 + }, + { + "epoch": 0.49, + "learning_rate": 0.00041537412236710135, + "loss": 1.3395, + "theoretical_loss": 3.523936055563151, + "tokens_seen": 1604124672 + }, + { + "epoch": 0.49, + "learning_rate": 0.000415358074222668, + "loss": 1.4877, + "theoretical_loss": 3.5239236262675186, + "tokens_seen": 1604190208 + }, + { + "epoch": 0.49, + "learning_rate": 0.00041534202607823475, + "loss": 1.3839, + "theoretical_loss": 3.5239111976218203, + "tokens_seen": 1604255744 + }, + { + "epoch": 0.49, + "learning_rate": 0.0004153259779338014, + "loss": 1.3349, + "theoretical_loss": 3.523898769625996, + "tokens_seen": 1604321280 + }, + { + "epoch": 0.49, + "learning_rate": 0.00041530992978936816, + "loss": 1.2833, + "theoretical_loss": 3.523886342279984, + "tokens_seen": 1604386816 + }, + { + "epoch": 0.49, + "learning_rate": 0.0004152938816449348, + "loss": 1.4211, + "theoretical_loss": 3.523873915583724, + "tokens_seen": 1604452352 + }, + { + "epoch": 0.49, + "learning_rate": 0.00041527783350050156, + "loss": 1.2708, + "theoretical_loss": 3.5238614895371563, + "tokens_seen": 1604517888 + }, + { + "epoch": 0.49, + "learning_rate": 0.0004152617853560682, + "loss": 1.4388, + "theoretical_loss": 3.5238490641402196, + "tokens_seen": 1604583424 + }, + { + "epoch": 0.49, + "learning_rate": 0.00041524573721163496, + "loss": 1.2642, + "theoretical_loss": 3.523836639392854, + "tokens_seen": 1604648960 + }, + { + "epoch": 0.49, + "learning_rate": 0.0004152296890672016, + "loss": 1.3001, + "theoretical_loss": 3.5238242152949986, + "tokens_seen": 1604714496 + }, + { + "epoch": 0.49, + "learning_rate": 0.00041521364092276836, + "loss": 1.3548, + "theoretical_loss": 3.5238117918465934, + "tokens_seen": 1604780032 + }, + { + "epoch": 0.49, + "learning_rate": 0.000415197592778335, + "loss": 1.2511, + "theoretical_loss": 3.5237993690475773, + "tokens_seen": 1604845568 + }, + { + "epoch": 0.49, + "learning_rate": 0.00041518154463390176, + "loss": 1.3913, + "theoretical_loss": 3.52378694689789, + "tokens_seen": 1604911104 + }, + { + "epoch": 0.49, + "learning_rate": 0.0004151654964894684, + "loss": 1.4708, + "theoretical_loss": 3.5237745253974717, + "tokens_seen": 1604976640 + }, + { + "epoch": 0.49, + "learning_rate": 0.00041514944834503516, + "loss": 1.323, + "theoretical_loss": 3.523762104546261, + "tokens_seen": 1605042176 + }, + { + "epoch": 0.49, + "learning_rate": 0.0004151334002006018, + "loss": 1.4283, + "theoretical_loss": 3.523749684344198, + "tokens_seen": 1605107712 + }, + { + "epoch": 0.49, + "learning_rate": 0.00041511735205616856, + "loss": 1.436, + "theoretical_loss": 3.5237372647912224, + "tokens_seen": 1605173248 + }, + { + "epoch": 0.49, + "learning_rate": 0.0004151013039117352, + "loss": 1.3461, + "theoretical_loss": 3.5237248458872736, + "tokens_seen": 1605238784 + }, + { + "epoch": 0.49, + "learning_rate": 0.00041508525576730196, + "loss": 1.3307, + "theoretical_loss": 3.5237124276322906, + "tokens_seen": 1605304320 + }, + { + "epoch": 0.49, + "learning_rate": 0.0004150692076228686, + "loss": 1.2546, + "theoretical_loss": 3.523700010026214, + "tokens_seen": 1605369856 + }, + { + "epoch": 0.49, + "learning_rate": 0.00041505315947843537, + "loss": 1.337, + "theoretical_loss": 3.5236875930689826, + "tokens_seen": 1605435392 + }, + { + "epoch": 0.49, + "learning_rate": 0.000415037111334002, + "loss": 1.3714, + "theoretical_loss": 3.5236751767605368, + "tokens_seen": 1605500928 + }, + { + "epoch": 0.49, + "learning_rate": 0.00041502106318956877, + "loss": 1.2455, + "theoretical_loss": 3.5236627611008156, + "tokens_seen": 1605566464 + }, + { + "epoch": 0.49, + "objective/train/docs_used": 586391, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.3409714698791504, + "objective/train/theoretical_loss": 3.5236503460897586, + "objective/train/tokens_used": 1626092000, + "theoretical_loss": 3.5236503460897586, + "tokens_seen": 1605632000 + }, + { + "epoch": 0.49, + "learning_rate": 0.0004150050150451354, + "loss": 1.3446, + "theoretical_loss": 3.5236503460897586, + "tokens_seen": 1605632000 + }, + { + "epoch": 0.49, + "learning_rate": 0.00041498896690070217, + "loss": 1.3375, + "theoretical_loss": 3.523637931727306, + "tokens_seen": 1605697536 + }, + { + "epoch": 0.49, + "learning_rate": 0.0004149729187562688, + "loss": 1.3287, + "theoretical_loss": 3.5236255180133966, + "tokens_seen": 1605763072 + }, + { + "epoch": 0.49, + "learning_rate": 0.00041495687061183557, + "loss": 1.247, + "theoretical_loss": 3.5236131049479713, + "tokens_seen": 1605828608 + }, + { + "epoch": 0.49, + "learning_rate": 0.0004149408224674022, + "loss": 1.3637, + "theoretical_loss": 3.5236006925309686, + "tokens_seen": 1605894144 + }, + { + "epoch": 0.49, + "learning_rate": 0.00041492477432296897, + "loss": 1.378, + "theoretical_loss": 3.523588280762328, + "tokens_seen": 1605959680 + }, + { + "epoch": 0.49, + "learning_rate": 0.0004149087261785356, + "loss": 1.4001, + "theoretical_loss": 3.5235758696419905, + "tokens_seen": 1606025216 + }, + { + "epoch": 0.49, + "learning_rate": 0.0004148926780341023, + "loss": 1.3968, + "theoretical_loss": 3.5235634591698948, + "tokens_seen": 1606090752 + }, + { + "epoch": 0.49, + "learning_rate": 0.000414876629889669, + "loss": 1.2699, + "theoretical_loss": 3.523551049345981, + "tokens_seen": 1606156288 + }, + { + "epoch": 0.49, + "learning_rate": 0.0004148605817452357, + "loss": 1.3029, + "theoretical_loss": 3.5235386401701883, + "tokens_seen": 1606221824 + }, + { + "epoch": 0.49, + "learning_rate": 0.0004148445336008024, + "loss": 1.3683, + "theoretical_loss": 3.5235262316424567, + "tokens_seen": 1606287360 + }, + { + "epoch": 0.49, + "learning_rate": 0.0004148284854563691, + "loss": 1.3277, + "theoretical_loss": 3.523513823762726, + "tokens_seen": 1606352896 + }, + { + "epoch": 0.49, + "learning_rate": 0.0004148124373119358, + "loss": 1.2272, + "theoretical_loss": 3.5235014165309364, + "tokens_seen": 1606418432 + }, + { + "epoch": 0.49, + "learning_rate": 0.0004147963891675025, + "loss": 1.3327, + "theoretical_loss": 3.5234890099470264, + "tokens_seen": 1606483968 + }, + { + "epoch": 0.49, + "learning_rate": 0.0004147803410230692, + "loss": 1.3507, + "theoretical_loss": 3.523476604010937, + "tokens_seen": 1606549504 + }, + { + "epoch": 0.49, + "learning_rate": 0.0004147642928786359, + "loss": 1.3639, + "theoretical_loss": 3.523464198722607, + "tokens_seen": 1606615040 + }, + { + "epoch": 0.49, + "learning_rate": 0.00041474824473420257, + "loss": 1.3361, + "theoretical_loss": 3.5234517940819767, + "tokens_seen": 1606680576 + }, + { + "epoch": 0.49, + "learning_rate": 0.0004147321965897693, + "loss": 1.3703, + "theoretical_loss": 3.5234393900889858, + "tokens_seen": 1606746112 + }, + { + "epoch": 0.49, + "learning_rate": 0.00041471614844533597, + "loss": 1.293, + "theoretical_loss": 3.5234269867435737, + "tokens_seen": 1606811648 + }, + { + "epoch": 0.49, + "learning_rate": 0.0004147001003009027, + "loss": 1.1433, + "theoretical_loss": 3.523414584045681, + "tokens_seen": 1606877184 + }, + { + "epoch": 0.49, + "learning_rate": 0.00041468405215646937, + "loss": 1.3901, + "theoretical_loss": 3.5234021819952464, + "tokens_seen": 1606942720 + }, + { + "epoch": 0.49, + "learning_rate": 0.0004146680040120361, + "loss": 1.2929, + "theoretical_loss": 3.5233897805922103, + "tokens_seen": 1607008256 + }, + { + "epoch": 0.49, + "learning_rate": 0.0004146519558676028, + "loss": 1.4555, + "theoretical_loss": 3.523377379836513, + "tokens_seen": 1607073792 + }, + { + "epoch": 0.49, + "learning_rate": 0.00041463590772316953, + "loss": 1.3824, + "theoretical_loss": 3.5233649797280933, + "tokens_seen": 1607139328 + }, + { + "epoch": 0.49, + "learning_rate": 0.0004146198595787362, + "loss": 1.2969, + "theoretical_loss": 3.5233525802668915, + "tokens_seen": 1607204864 + }, + { + "epoch": 0.49, + "learning_rate": 0.00041460381143430293, + "loss": 1.3359, + "theoretical_loss": 3.5233401814528476, + "tokens_seen": 1607270400 + }, + { + "epoch": 0.49, + "learning_rate": 0.0004145877632898696, + "loss": 1.3164, + "theoretical_loss": 3.523327783285901, + "tokens_seen": 1607335936 + }, + { + "epoch": 0.49, + "learning_rate": 0.00041457171514543633, + "loss": 1.3556, + "theoretical_loss": 3.5233153857659922, + "tokens_seen": 1607401472 + }, + { + "epoch": 0.49, + "learning_rate": 0.000414555667001003, + "loss": 1.2801, + "theoretical_loss": 3.523302988893061, + "tokens_seen": 1607467008 + }, + { + "epoch": 0.49, + "learning_rate": 0.00041453961885656973, + "loss": 1.5151, + "theoretical_loss": 3.5232905926670464, + "tokens_seen": 1607532544 + }, + { + "epoch": 0.49, + "learning_rate": 0.0004145235707121364, + "loss": 1.2542, + "theoretical_loss": 3.5232781970878886, + "tokens_seen": 1607598080 + }, + { + "epoch": 0.49, + "learning_rate": 0.00041450752256770313, + "loss": 1.3035, + "theoretical_loss": 3.523265802155528, + "tokens_seen": 1607663616 + }, + { + "epoch": 0.49, + "learning_rate": 0.0004144914744232698, + "loss": 1.3045, + "theoretical_loss": 3.5232534078699036, + "tokens_seen": 1607729152 + }, + { + "epoch": 0.49, + "learning_rate": 0.00041447542627883653, + "loss": 1.4277, + "theoretical_loss": 3.5232410142309565, + "tokens_seen": 1607794688 + }, + { + "epoch": 0.49, + "learning_rate": 0.0004144593781344032, + "loss": 1.4061, + "theoretical_loss": 3.523228621238626, + "tokens_seen": 1607860224 + }, + { + "epoch": 0.49, + "learning_rate": 0.00041444332998996994, + "loss": 1.396, + "theoretical_loss": 3.523216228892852, + "tokens_seen": 1607925760 + }, + { + "epoch": 0.49, + "learning_rate": 0.0004144272818455366, + "loss": 1.3003, + "theoretical_loss": 3.5232038371935737, + "tokens_seen": 1607991296 + }, + { + "epoch": 0.49, + "learning_rate": 0.00041441123370110334, + "loss": 1.2885, + "theoretical_loss": 3.523191446140732, + "tokens_seen": 1608056832 + }, + { + "epoch": 0.49, + "learning_rate": 0.0004143951855566701, + "loss": 1.2804, + "theoretical_loss": 3.523179055734267, + "tokens_seen": 1608122368 + }, + { + "epoch": 0.49, + "learning_rate": 0.00041437913741223674, + "loss": 1.4153, + "theoretical_loss": 3.523166665974118, + "tokens_seen": 1608187904 + }, + { + "epoch": 0.49, + "learning_rate": 0.0004143630892678035, + "loss": 1.2701, + "theoretical_loss": 3.523154276860225, + "tokens_seen": 1608253440 + }, + { + "epoch": 0.49, + "learning_rate": 0.00041434704112337014, + "loss": 1.3752, + "theoretical_loss": 3.5231418883925283, + "tokens_seen": 1608318976 + }, + { + "epoch": 0.49, + "learning_rate": 0.0004143309929789369, + "loss": 1.3678, + "theoretical_loss": 3.523129500570967, + "tokens_seen": 1608384512 + }, + { + "epoch": 0.49, + "learning_rate": 0.00041431494483450354, + "loss": 1.2812, + "theoretical_loss": 3.5231171133954824, + "tokens_seen": 1608450048 + }, + { + "epoch": 0.49, + "learning_rate": 0.0004142988966900703, + "loss": 1.4323, + "theoretical_loss": 3.5231047268660136, + "tokens_seen": 1608515584 + }, + { + "epoch": 0.49, + "learning_rate": 0.00041428284854563694, + "loss": 1.398, + "theoretical_loss": 3.523092340982501, + "tokens_seen": 1608581120 + }, + { + "epoch": 0.49, + "learning_rate": 0.00041426680040120364, + "loss": 1.4296, + "theoretical_loss": 3.523079955744884, + "tokens_seen": 1608646656 + }, + { + "epoch": 0.49, + "learning_rate": 0.00041425075225677034, + "loss": 1.3552, + "theoretical_loss": 3.5230675711531036, + "tokens_seen": 1608712192 + }, + { + "epoch": 0.49, + "learning_rate": 0.00041423470411233704, + "loss": 1.249, + "theoretical_loss": 3.523055187207099, + "tokens_seen": 1608777728 + }, + { + "epoch": 0.49, + "learning_rate": 0.00041421865596790374, + "loss": 1.4478, + "theoretical_loss": 3.5230428039068102, + "tokens_seen": 1608843264 + }, + { + "epoch": 0.49, + "objective/train/docs_used": 587617, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.5823975801467896, + "objective/train/theoretical_loss": 3.5230304212521775, + "objective/train/tokens_used": 1629368800, + "theoretical_loss": 3.5230304212521775, + "tokens_seen": 1608908800 + }, + { + "epoch": 0.49, + "learning_rate": 0.00041420260782347045, + "loss": 1.4965, + "theoretical_loss": 3.5230304212521775, + "tokens_seen": 1608908800 + }, + { + "epoch": 0.49, + "learning_rate": 0.00041418655967903715, + "loss": 1.393, + "theoretical_loss": 3.523018039243141, + "tokens_seen": 1608974336 + }, + { + "epoch": 0.49, + "learning_rate": 0.00041417051153460385, + "loss": 1.4898, + "theoretical_loss": 3.523005657879641, + "tokens_seen": 1609039872 + }, + { + "epoch": 0.49, + "learning_rate": 0.00041415446339017055, + "loss": 1.4043, + "theoretical_loss": 3.522993277161617, + "tokens_seen": 1609105408 + }, + { + "epoch": 0.49, + "learning_rate": 0.00041413841524573725, + "loss": 1.3389, + "theoretical_loss": 3.522980897089009, + "tokens_seen": 1609170944 + }, + { + "epoch": 0.49, + "learning_rate": 0.0004141223671013039, + "loss": 1.4361, + "theoretical_loss": 3.5229685176617576, + "tokens_seen": 1609236480 + }, + { + "epoch": 0.49, + "learning_rate": 0.00041410631895687065, + "loss": 1.4569, + "theoretical_loss": 3.5229561388798025, + "tokens_seen": 1609302016 + }, + { + "epoch": 0.49, + "learning_rate": 0.0004140902708124373, + "loss": 1.3873, + "theoretical_loss": 3.5229437607430842, + "tokens_seen": 1609367552 + }, + { + "epoch": 0.49, + "learning_rate": 0.00041407422266800405, + "loss": 1.4131, + "theoretical_loss": 3.5229313832515423, + "tokens_seen": 1609433088 + }, + { + "epoch": 0.49, + "learning_rate": 0.0004140581745235707, + "loss": 1.2879, + "theoretical_loss": 3.5229190064051172, + "tokens_seen": 1609498624 + }, + { + "epoch": 0.49, + "learning_rate": 0.00041404212637913745, + "loss": 1.2856, + "theoretical_loss": 3.522906630203749, + "tokens_seen": 1609564160 + }, + { + "epoch": 0.49, + "learning_rate": 0.0004140260782347041, + "loss": 1.3371, + "theoretical_loss": 3.522894254647378, + "tokens_seen": 1609629696 + }, + { + "epoch": 0.49, + "learning_rate": 0.00041401003009027085, + "loss": 1.3531, + "theoretical_loss": 3.522881879735943, + "tokens_seen": 1609695232 + }, + { + "epoch": 0.49, + "learning_rate": 0.0004139939819458375, + "loss": 1.4127, + "theoretical_loss": 3.5228695054693864, + "tokens_seen": 1609760768 + }, + { + "epoch": 0.49, + "learning_rate": 0.00041397793380140425, + "loss": 1.3143, + "theoretical_loss": 3.5228571318476467, + "tokens_seen": 1609826304 + }, + { + "epoch": 0.49, + "learning_rate": 0.0004139618856569709, + "loss": 1.3784, + "theoretical_loss": 3.522844758870664, + "tokens_seen": 1609891840 + }, + { + "epoch": 0.49, + "learning_rate": 0.00041394583751253766, + "loss": 1.4641, + "theoretical_loss": 3.5228323865383793, + "tokens_seen": 1609957376 + }, + { + "epoch": 0.49, + "learning_rate": 0.0004139297893681043, + "loss": 1.394, + "theoretical_loss": 3.5228200148507325, + "tokens_seen": 1610022912 + }, + { + "epoch": 0.49, + "learning_rate": 0.00041391374122367106, + "loss": 1.3034, + "theoretical_loss": 3.5228076438076643, + "tokens_seen": 1610088448 + }, + { + "epoch": 0.49, + "learning_rate": 0.0004138976930792377, + "loss": 1.4104, + "theoretical_loss": 3.5227952734091135, + "tokens_seen": 1610153984 + }, + { + "epoch": 0.49, + "learning_rate": 0.00041388164493480446, + "loss": 1.3387, + "theoretical_loss": 3.5227829036550213, + "tokens_seen": 1610219520 + }, + { + "epoch": 0.49, + "learning_rate": 0.0004138655967903711, + "loss": 1.3894, + "theoretical_loss": 3.5227705345453275, + "tokens_seen": 1610285056 + }, + { + "epoch": 0.49, + "learning_rate": 0.00041384954864593786, + "loss": 1.3819, + "theoretical_loss": 3.5227581660799725, + "tokens_seen": 1610350592 + }, + { + "epoch": 0.49, + "learning_rate": 0.0004138335005015045, + "loss": 1.3777, + "theoretical_loss": 3.5227457982588968, + "tokens_seen": 1610416128 + }, + { + "epoch": 0.49, + "learning_rate": 0.00041381745235707126, + "loss": 1.3611, + "theoretical_loss": 3.52273343108204, + "tokens_seen": 1610481664 + }, + { + "epoch": 0.49, + "learning_rate": 0.0004138014042126379, + "loss": 1.3809, + "theoretical_loss": 3.522721064549343, + "tokens_seen": 1610547200 + }, + { + "epoch": 0.49, + "learning_rate": 0.00041378535606820466, + "loss": 1.3428, + "theoretical_loss": 3.5227086986607454, + "tokens_seen": 1610612736 + }, + { + "epoch": 0.49, + "learning_rate": 0.0004137693079237713, + "loss": 1.3327, + "theoretical_loss": 3.522696333416188, + "tokens_seen": 1610678272 + }, + { + "epoch": 0.49, + "learning_rate": 0.00041375325977933806, + "loss": 1.3794, + "theoretical_loss": 3.5226839688156106, + "tokens_seen": 1610743808 + }, + { + "epoch": 0.49, + "learning_rate": 0.0004137372116349047, + "loss": 1.4669, + "theoretical_loss": 3.522671604858953, + "tokens_seen": 1610809344 + }, + { + "epoch": 0.49, + "learning_rate": 0.00041372116349047146, + "loss": 1.3855, + "theoretical_loss": 3.5226592415461573, + "tokens_seen": 1610874880 + }, + { + "epoch": 0.49, + "learning_rate": 0.0004137051153460381, + "loss": 1.4152, + "theoretical_loss": 3.522646878877162, + "tokens_seen": 1610940416 + }, + { + "epoch": 0.49, + "learning_rate": 0.00041368906720160487, + "loss": 1.2869, + "theoretical_loss": 3.522634516851908, + "tokens_seen": 1611005952 + }, + { + "epoch": 0.49, + "learning_rate": 0.0004136730190571715, + "loss": 1.2536, + "theoretical_loss": 3.5226221554703354, + "tokens_seen": 1611071488 + }, + { + "epoch": 0.49, + "learning_rate": 0.00041365697091273827, + "loss": 1.349, + "theoretical_loss": 3.522609794732385, + "tokens_seen": 1611137024 + }, + { + "epoch": 0.49, + "learning_rate": 0.0004136409227683049, + "loss": 1.2675, + "theoretical_loss": 3.5225974346379965, + "tokens_seen": 1611202560 + }, + { + "epoch": 0.49, + "learning_rate": 0.00041362487462387167, + "loss": 1.4712, + "theoretical_loss": 3.52258507518711, + "tokens_seen": 1611268096 + }, + { + "epoch": 0.49, + "learning_rate": 0.0004136088264794383, + "loss": 1.434, + "theoretical_loss": 3.5225727163796674, + "tokens_seen": 1611333632 + }, + { + "epoch": 0.49, + "learning_rate": 0.00041359277833500507, + "loss": 1.2995, + "theoretical_loss": 3.5225603582156078, + "tokens_seen": 1611399168 + }, + { + "epoch": 0.49, + "learning_rate": 0.0004135767301905717, + "loss": 1.3273, + "theoretical_loss": 3.5225480006948713, + "tokens_seen": 1611464704 + }, + { + "epoch": 0.49, + "learning_rate": 0.00041356068204613847, + "loss": 1.2973, + "theoretical_loss": 3.5225356438173985, + "tokens_seen": 1611530240 + }, + { + "epoch": 0.49, + "learning_rate": 0.0004135446339017051, + "loss": 1.4857, + "theoretical_loss": 3.5225232875831303, + "tokens_seen": 1611595776 + }, + { + "epoch": 0.49, + "learning_rate": 0.00041352858575727187, + "loss": 1.4044, + "theoretical_loss": 3.5225109319920067, + "tokens_seen": 1611661312 + }, + { + "epoch": 0.49, + "learning_rate": 0.0004135125376128385, + "loss": 1.3627, + "theoretical_loss": 3.522498577043968, + "tokens_seen": 1611726848 + }, + { + "epoch": 0.49, + "learning_rate": 0.0004134964894684052, + "loss": 1.449, + "theoretical_loss": 3.522486222738954, + "tokens_seen": 1611792384 + }, + { + "epoch": 0.49, + "learning_rate": 0.0004134804413239719, + "loss": 1.5153, + "theoretical_loss": 3.522473869076906, + "tokens_seen": 1611857920 + }, + { + "epoch": 0.49, + "learning_rate": 0.0004134643931795386, + "loss": 1.3168, + "theoretical_loss": 3.5224615160577644, + "tokens_seen": 1611923456 + }, + { + "epoch": 0.49, + "learning_rate": 0.0004134483450351053, + "loss": 1.4235, + "theoretical_loss": 3.5224491636814697, + "tokens_seen": 1611988992 + }, + { + "epoch": 0.49, + "learning_rate": 0.000413432296890672, + "loss": 1.304, + "theoretical_loss": 3.522436811947961, + "tokens_seen": 1612054528 + }, + { + "epoch": 0.49, + "learning_rate": 0.0004134162487462387, + "loss": 1.3785, + "theoretical_loss": 3.522424460857181, + "tokens_seen": 1612120064 + }, + { + "epoch": 0.49, + "objective/train/docs_used": 588885, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.4180693626403809, + "objective/train/theoretical_loss": 3.5224121104090678, + "objective/train/tokens_used": 1632645600, + "theoretical_loss": 3.5224121104090678, + "tokens_seen": 1612185600 + }, + { + "epoch": 0.49, + "learning_rate": 0.0004134002006018054, + "loss": 1.3421, + "theoretical_loss": 3.5224121104090678, + "tokens_seen": 1612185600 + }, + { + "epoch": 0.49, + "learning_rate": 0.0004133841524573721, + "loss": 1.3584, + "theoretical_loss": 3.5223997606035624, + "tokens_seen": 1612251136 + }, + { + "epoch": 0.49, + "learning_rate": 0.0004133681043129388, + "loss": 1.3179, + "theoretical_loss": 3.5223874114406066, + "tokens_seen": 1612316672 + }, + { + "epoch": 0.49, + "learning_rate": 0.00041335205616850547, + "loss": 1.4749, + "theoretical_loss": 3.5223750629201396, + "tokens_seen": 1612382208 + }, + { + "epoch": 0.49, + "learning_rate": 0.0004133360080240722, + "loss": 1.3987, + "theoretical_loss": 3.522362715042102, + "tokens_seen": 1612447744 + }, + { + "epoch": 0.49, + "learning_rate": 0.00041331995987963887, + "loss": 1.2022, + "theoretical_loss": 3.5223503678064345, + "tokens_seen": 1612513280 + }, + { + "epoch": 0.49, + "learning_rate": 0.0004133039117352056, + "loss": 1.3592, + "theoretical_loss": 3.522338021213078, + "tokens_seen": 1612578816 + }, + { + "epoch": 0.49, + "learning_rate": 0.0004132878635907723, + "loss": 1.2423, + "theoretical_loss": 3.522325675261972, + "tokens_seen": 1612644352 + }, + { + "epoch": 0.49, + "learning_rate": 0.00041327181544633903, + "loss": 1.4434, + "theoretical_loss": 3.5223133299530573, + "tokens_seen": 1612709888 + }, + { + "epoch": 0.49, + "learning_rate": 0.0004132557673019057, + "loss": 1.4043, + "theoretical_loss": 3.522300985286275, + "tokens_seen": 1612775424 + }, + { + "epoch": 0.49, + "learning_rate": 0.00041323971915747243, + "loss": 1.3644, + "theoretical_loss": 3.5222886412615653, + "tokens_seen": 1612840960 + }, + { + "epoch": 0.49, + "learning_rate": 0.0004132236710130391, + "loss": 1.3579, + "theoretical_loss": 3.522276297878869, + "tokens_seen": 1612906496 + }, + { + "epoch": 0.49, + "learning_rate": 0.00041320762286860583, + "loss": 1.4005, + "theoretical_loss": 3.5222639551381256, + "tokens_seen": 1612972032 + }, + { + "epoch": 0.49, + "learning_rate": 0.0004131915747241726, + "loss": 1.3758, + "theoretical_loss": 3.5222516130392765, + "tokens_seen": 1613037568 + }, + { + "epoch": 0.49, + "learning_rate": 0.00041317552657973923, + "loss": 1.3731, + "theoretical_loss": 3.522239271582262, + "tokens_seen": 1613103104 + }, + { + "epoch": 0.49, + "learning_rate": 0.000413159478435306, + "loss": 1.3016, + "theoretical_loss": 3.5222269307670224, + "tokens_seen": 1613168640 + }, + { + "epoch": 0.49, + "learning_rate": 0.00041314343029087263, + "loss": 1.4172, + "theoretical_loss": 3.522214590593499, + "tokens_seen": 1613234176 + }, + { + "epoch": 0.49, + "learning_rate": 0.0004131273821464394, + "loss": 1.3695, + "theoretical_loss": 3.5222022510616315, + "tokens_seen": 1613299712 + }, + { + "epoch": 0.49, + "learning_rate": 0.00041311133400200603, + "loss": 1.3799, + "theoretical_loss": 3.5221899121713616, + "tokens_seen": 1613365248 + }, + { + "epoch": 0.49, + "learning_rate": 0.0004130952858575728, + "loss": 1.4267, + "theoretical_loss": 3.522177573922628, + "tokens_seen": 1613430784 + }, + { + "epoch": 0.49, + "learning_rate": 0.00041307923771313944, + "loss": 1.4301, + "theoretical_loss": 3.5221652363153733, + "tokens_seen": 1613496320 + }, + { + "epoch": 0.49, + "learning_rate": 0.0004130631895687062, + "loss": 1.361, + "theoretical_loss": 3.5221528993495372, + "tokens_seen": 1613561856 + }, + { + "epoch": 0.49, + "learning_rate": 0.00041304714142427284, + "loss": 1.3222, + "theoretical_loss": 3.5221405630250597, + "tokens_seen": 1613627392 + }, + { + "epoch": 0.49, + "learning_rate": 0.0004130310932798396, + "loss": 1.3612, + "theoretical_loss": 3.5221282273418826, + "tokens_seen": 1613692928 + }, + { + "epoch": 0.49, + "learning_rate": 0.00041301504513540624, + "loss": 1.3654, + "theoretical_loss": 3.5221158922999463, + "tokens_seen": 1613758464 + }, + { + "epoch": 0.49, + "learning_rate": 0.000412998996990973, + "loss": 1.4091, + "theoretical_loss": 3.5221035578991904, + "tokens_seen": 1613824000 + }, + { + "epoch": 0.49, + "learning_rate": 0.00041298294884653964, + "loss": 1.3905, + "theoretical_loss": 3.5220912241395563, + "tokens_seen": 1613889536 + }, + { + "epoch": 0.49, + "learning_rate": 0.0004129669007021064, + "loss": 1.4472, + "theoretical_loss": 3.522078891020985, + "tokens_seen": 1613955072 + }, + { + "epoch": 0.49, + "learning_rate": 0.00041295085255767304, + "loss": 1.3568, + "theoretical_loss": 3.5220665585434165, + "tokens_seen": 1614020608 + }, + { + "epoch": 0.49, + "learning_rate": 0.0004129348044132398, + "loss": 1.4302, + "theoretical_loss": 3.522054226706792, + "tokens_seen": 1614086144 + }, + { + "epoch": 0.49, + "learning_rate": 0.00041291875626880644, + "loss": 1.4833, + "theoretical_loss": 3.522041895511051, + "tokens_seen": 1614151680 + }, + { + "epoch": 0.49, + "learning_rate": 0.0004129027081243732, + "loss": 1.4479, + "theoretical_loss": 3.522029564956136, + "tokens_seen": 1614217216 + }, + { + "epoch": 0.49, + "learning_rate": 0.00041288665997993984, + "loss": 1.5043, + "theoretical_loss": 3.522017235041986, + "tokens_seen": 1614282752 + }, + { + "epoch": 0.49, + "learning_rate": 0.0004128706118355066, + "loss": 1.4152, + "theoretical_loss": 3.5220049057685427, + "tokens_seen": 1614348288 + }, + { + "epoch": 0.49, + "learning_rate": 0.00041285456369107324, + "loss": 1.4214, + "theoretical_loss": 3.5219925771357463, + "tokens_seen": 1614413824 + }, + { + "epoch": 0.49, + "learning_rate": 0.00041283851554663995, + "loss": 1.4183, + "theoretical_loss": 3.521980249143538, + "tokens_seen": 1614479360 + }, + { + "epoch": 0.49, + "learning_rate": 0.00041282246740220665, + "loss": 1.4619, + "theoretical_loss": 3.521967921791858, + "tokens_seen": 1614544896 + }, + { + "epoch": 0.49, + "learning_rate": 0.00041280641925777335, + "loss": 1.2914, + "theoretical_loss": 3.521955595080647, + "tokens_seen": 1614610432 + }, + { + "epoch": 0.49, + "learning_rate": 0.00041279037111334005, + "loss": 1.4513, + "theoretical_loss": 3.5219432690098467, + "tokens_seen": 1614675968 + }, + { + "epoch": 0.49, + "learning_rate": 0.00041277432296890675, + "loss": 1.4002, + "theoretical_loss": 3.521930943579397, + "tokens_seen": 1614741504 + }, + { + "epoch": 0.49, + "learning_rate": 0.00041275827482447345, + "loss": 1.4126, + "theoretical_loss": 3.521918618789238, + "tokens_seen": 1614807040 + }, + { + "epoch": 0.49, + "learning_rate": 0.00041274222668004015, + "loss": 1.3416, + "theoretical_loss": 3.521906294639312, + "tokens_seen": 1614872576 + }, + { + "epoch": 0.49, + "learning_rate": 0.0004127261785356068, + "loss": 1.3606, + "theoretical_loss": 3.5218939711295585, + "tokens_seen": 1614938112 + }, + { + "epoch": 0.49, + "learning_rate": 0.00041271013039117355, + "loss": 1.4981, + "theoretical_loss": 3.521881648259919, + "tokens_seen": 1615003648 + }, + { + "epoch": 0.49, + "learning_rate": 0.0004126940822467402, + "loss": 1.3328, + "theoretical_loss": 3.521869326030334, + "tokens_seen": 1615069184 + }, + { + "epoch": 0.49, + "learning_rate": 0.00041267803410230695, + "loss": 1.3854, + "theoretical_loss": 3.521857004440744, + "tokens_seen": 1615134720 + }, + { + "epoch": 0.49, + "learning_rate": 0.0004126619859578736, + "loss": 1.2935, + "theoretical_loss": 3.52184468349109, + "tokens_seen": 1615200256 + }, + { + "epoch": 0.49, + "learning_rate": 0.00041264593781344035, + "loss": 1.337, + "theoretical_loss": 3.5218323631813133, + "tokens_seen": 1615265792 + }, + { + "epoch": 0.49, + "learning_rate": 0.000412629889669007, + "loss": 1.4801, + "theoretical_loss": 3.5218200435113545, + "tokens_seen": 1615331328 + }, + { + "epoch": 0.49, + "learning_rate": 0.00041261384152457375, + "loss": 1.3312, + "theoretical_loss": 3.521807724481154, + "tokens_seen": 1615396864 + }, + { + "epoch": 0.49, + "objective/train/docs_used": 590099, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.317259669303894, + "objective/train/theoretical_loss": 3.5217954060906527, + "objective/train/tokens_used": 1635922400, + "theoretical_loss": 3.5217954060906527, + "tokens_seen": 1615462400 + }, + { + "epoch": 0.49, + "learning_rate": 0.0004125977933801404, + "loss": 1.3476, + "theoretical_loss": 3.5217954060906527, + "tokens_seen": 1615462400 + }, + { + "epoch": 0.49, + "learning_rate": 0.00041258174523570716, + "loss": 1.2869, + "theoretical_loss": 3.5217830883397916, + "tokens_seen": 1615527936 + }, + { + "epoch": 0.49, + "learning_rate": 0.0004125656970912738, + "loss": 1.3812, + "theoretical_loss": 3.521770771228512, + "tokens_seen": 1615593472 + }, + { + "epoch": 0.49, + "learning_rate": 0.00041254964894684056, + "loss": 1.4326, + "theoretical_loss": 3.5217584547567538, + "tokens_seen": 1615659008 + }, + { + "epoch": 0.49, + "learning_rate": 0.0004125336008024072, + "loss": 1.4892, + "theoretical_loss": 3.521746138924458, + "tokens_seen": 1615724544 + }, + { + "epoch": 0.49, + "learning_rate": 0.00041251755265797396, + "loss": 1.3695, + "theoretical_loss": 3.5217338237315667, + "tokens_seen": 1615790080 + }, + { + "epoch": 0.49, + "learning_rate": 0.0004125015045135406, + "loss": 1.2685, + "theoretical_loss": 3.521721509178019, + "tokens_seen": 1615855616 + }, + { + "epoch": 0.49, + "learning_rate": 0.00041248545636910736, + "loss": 1.4092, + "theoretical_loss": 3.5217091952637567, + "tokens_seen": 1615921152 + }, + { + "epoch": 0.49, + "learning_rate": 0.000412469408224674, + "loss": 1.333, + "theoretical_loss": 3.5216968819887216, + "tokens_seen": 1615986688 + }, + { + "epoch": 0.49, + "learning_rate": 0.00041245336008024076, + "loss": 1.281, + "theoretical_loss": 3.5216845693528525, + "tokens_seen": 1616052224 + }, + { + "epoch": 0.49, + "learning_rate": 0.0004124373119358074, + "loss": 1.4499, + "theoretical_loss": 3.5216722573560917, + "tokens_seen": 1616117760 + }, + { + "epoch": 0.49, + "learning_rate": 0.00041242126379137416, + "loss": 1.3063, + "theoretical_loss": 3.5216599459983806, + "tokens_seen": 1616183296 + }, + { + "epoch": 0.49, + "learning_rate": 0.0004124052156469408, + "loss": 1.344, + "theoretical_loss": 3.5216476352796584, + "tokens_seen": 1616248832 + }, + { + "epoch": 0.49, + "learning_rate": 0.00041238916750250756, + "loss": 1.3605, + "theoretical_loss": 3.5216353251998673, + "tokens_seen": 1616314368 + }, + { + "epoch": 0.49, + "learning_rate": 0.0004123731193580742, + "loss": 1.4332, + "theoretical_loss": 3.5216230157589483, + "tokens_seen": 1616379904 + }, + { + "epoch": 0.49, + "learning_rate": 0.00041235707121364096, + "loss": 1.3985, + "theoretical_loss": 3.5216107069568414, + "tokens_seen": 1616445440 + }, + { + "epoch": 0.49, + "learning_rate": 0.0004123410230692076, + "loss": 1.3693, + "theoretical_loss": 3.5215983987934885, + "tokens_seen": 1616510976 + }, + { + "epoch": 0.49, + "learning_rate": 0.00041232497492477437, + "loss": 1.4823, + "theoretical_loss": 3.52158609126883, + "tokens_seen": 1616576512 + }, + { + "epoch": 0.49, + "learning_rate": 0.000412308926780341, + "loss": 1.33, + "theoretical_loss": 3.5215737843828068, + "tokens_seen": 1616642048 + }, + { + "epoch": 0.49, + "learning_rate": 0.00041229287863590777, + "loss": 1.419, + "theoretical_loss": 3.521561478135361, + "tokens_seen": 1616707584 + }, + { + "epoch": 0.49, + "learning_rate": 0.0004122768304914744, + "loss": 1.3508, + "theoretical_loss": 3.5215491725264316, + "tokens_seen": 1616773120 + }, + { + "epoch": 0.49, + "learning_rate": 0.00041226078234704117, + "loss": 1.327, + "theoretical_loss": 3.5215368675559615, + "tokens_seen": 1616838656 + }, + { + "epoch": 0.49, + "learning_rate": 0.0004122447342026078, + "loss": 1.3255, + "theoretical_loss": 3.52152456322389, + "tokens_seen": 1616904192 + }, + { + "epoch": 0.49, + "learning_rate": 0.00041222868605817457, + "loss": 1.3762, + "theoretical_loss": 3.52151225953016, + "tokens_seen": 1616969728 + }, + { + "epoch": 0.49, + "learning_rate": 0.0004122126379137412, + "loss": 1.4405, + "theoretical_loss": 3.521499956474711, + "tokens_seen": 1617035264 + }, + { + "epoch": 0.49, + "learning_rate": 0.00041219658976930797, + "loss": 1.4789, + "theoretical_loss": 3.521487654057484, + "tokens_seen": 1617100800 + }, + { + "epoch": 0.49, + "learning_rate": 0.0004121805416248746, + "loss": 1.4015, + "theoretical_loss": 3.5214753522784212, + "tokens_seen": 1617166336 + }, + { + "epoch": 0.49, + "learning_rate": 0.00041216449348044137, + "loss": 1.4477, + "theoretical_loss": 3.521463051137463, + "tokens_seen": 1617231872 + }, + { + "epoch": 0.49, + "learning_rate": 0.000412148445336008, + "loss": 1.3396, + "theoretical_loss": 3.5214507506345507, + "tokens_seen": 1617297408 + }, + { + "epoch": 0.49, + "learning_rate": 0.0004121323971915748, + "loss": 1.3615, + "theoretical_loss": 3.5214384507696246, + "tokens_seen": 1617362944 + }, + { + "epoch": 0.49, + "learning_rate": 0.0004121163490471414, + "loss": 1.429, + "theoretical_loss": 3.5214261515426264, + "tokens_seen": 1617428480 + }, + { + "epoch": 0.49, + "learning_rate": 0.0004121003009027082, + "loss": 1.36, + "theoretical_loss": 3.521413852953497, + "tokens_seen": 1617494016 + }, + { + "epoch": 0.49, + "learning_rate": 0.0004120842527582748, + "loss": 1.291, + "theoretical_loss": 3.5214015550021776, + "tokens_seen": 1617559552 + }, + { + "epoch": 0.49, + "learning_rate": 0.0004120682046138415, + "loss": 1.2894, + "theoretical_loss": 3.5213892576886088, + "tokens_seen": 1617625088 + }, + { + "epoch": 0.49, + "learning_rate": 0.0004120521564694082, + "loss": 1.2688, + "theoretical_loss": 3.5213769610127326, + "tokens_seen": 1617690624 + }, + { + "epoch": 0.49, + "learning_rate": 0.0004120361083249749, + "loss": 1.3719, + "theoretical_loss": 3.521364664974489, + "tokens_seen": 1617756160 + }, + { + "epoch": 0.49, + "learning_rate": 0.0004120200601805416, + "loss": 1.3331, + "theoretical_loss": 3.52135236957382, + "tokens_seen": 1617821696 + }, + { + "epoch": 0.49, + "learning_rate": 0.0004120040120361083, + "loss": 1.364, + "theoretical_loss": 3.5213400748106665, + "tokens_seen": 1617887232 + }, + { + "epoch": 0.49, + "learning_rate": 0.0004119879638916751, + "loss": 1.4103, + "theoretical_loss": 3.5213277806849694, + "tokens_seen": 1617952768 + }, + { + "epoch": 0.49, + "learning_rate": 0.0004119719157472417, + "loss": 1.3724, + "theoretical_loss": 3.5213154871966696, + "tokens_seen": 1618018304 + }, + { + "epoch": 0.49, + "learning_rate": 0.0004119558676028085, + "loss": 1.4276, + "theoretical_loss": 3.521303194345709, + "tokens_seen": 1618083840 + }, + { + "epoch": 0.49, + "learning_rate": 0.0004119398194583751, + "loss": 1.3772, + "theoretical_loss": 3.5212909021320282, + "tokens_seen": 1618149376 + }, + { + "epoch": 0.49, + "learning_rate": 0.0004119237713139419, + "loss": 1.3356, + "theoretical_loss": 3.5212786105555685, + "tokens_seen": 1618214912 + }, + { + "epoch": 0.49, + "learning_rate": 0.00041190772316950853, + "loss": 1.4679, + "theoretical_loss": 3.521266319616271, + "tokens_seen": 1618280448 + }, + { + "epoch": 0.49, + "learning_rate": 0.0004118916750250753, + "loss": 1.3789, + "theoretical_loss": 3.521254029314077, + "tokens_seen": 1618345984 + }, + { + "epoch": 0.49, + "learning_rate": 0.00041187562688064193, + "loss": 1.3695, + "theoretical_loss": 3.5212417396489273, + "tokens_seen": 1618411520 + }, + { + "epoch": 0.49, + "learning_rate": 0.0004118595787362087, + "loss": 1.3954, + "theoretical_loss": 3.5212294506207638, + "tokens_seen": 1618477056 + }, + { + "epoch": 0.49, + "learning_rate": 0.00041184353059177533, + "loss": 1.3105, + "theoretical_loss": 3.5212171622295267, + "tokens_seen": 1618542592 + }, + { + "epoch": 0.49, + "learning_rate": 0.0004118274824473421, + "loss": 1.3551, + "theoretical_loss": 3.521204874475158, + "tokens_seen": 1618608128 + }, + { + "epoch": 0.49, + "learning_rate": 0.00041181143430290873, + "loss": 1.4268, + "theoretical_loss": 3.521192587357599, + "tokens_seen": 1618673664 + }, + { + "epoch": 0.49, + "objective/train/docs_used": 591377, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.2204684019088745, + "objective/train/theoretical_loss": 3.5211803008767903, + "objective/train/tokens_used": 1639199200, + "theoretical_loss": 3.5211803008767903, + "tokens_seen": 1618739200 + }, + { + "epoch": 0.49, + "learning_rate": 0.0004117953861584755, + "loss": 1.1838, + "theoretical_loss": 3.5211803008767903, + "tokens_seen": 1618739200 + }, + { + "epoch": 0.49, + "learning_rate": 0.00041177933801404213, + "loss": 1.3937, + "theoretical_loss": 3.521168015032674, + "tokens_seen": 1618804736 + }, + { + "epoch": 0.49, + "learning_rate": 0.0004117632898696089, + "loss": 1.3957, + "theoretical_loss": 3.52115572982519, + "tokens_seen": 1618870272 + }, + { + "epoch": 0.49, + "learning_rate": 0.00041174724172517553, + "loss": 1.4319, + "theoretical_loss": 3.521143445254281, + "tokens_seen": 1618935808 + }, + { + "epoch": 0.49, + "learning_rate": 0.0004117311935807423, + "loss": 1.4273, + "theoretical_loss": 3.521131161319887, + "tokens_seen": 1619001344 + }, + { + "epoch": 0.49, + "learning_rate": 0.00041171514543630894, + "loss": 1.3706, + "theoretical_loss": 3.52111887802195, + "tokens_seen": 1619066880 + }, + { + "epoch": 0.49, + "learning_rate": 0.0004116990972918757, + "loss": 1.423, + "theoretical_loss": 3.5211065953604104, + "tokens_seen": 1619132416 + }, + { + "epoch": 0.49, + "learning_rate": 0.00041168304914744234, + "loss": 1.2461, + "theoretical_loss": 3.521094313335211, + "tokens_seen": 1619197952 + }, + { + "epoch": 0.49, + "learning_rate": 0.0004116670010030091, + "loss": 1.3525, + "theoretical_loss": 3.5210820319462917, + "tokens_seen": 1619263488 + }, + { + "epoch": 0.49, + "learning_rate": 0.00041165095285857574, + "loss": 1.396, + "theoretical_loss": 3.521069751193594, + "tokens_seen": 1619329024 + }, + { + "epoch": 0.49, + "learning_rate": 0.0004116349047141425, + "loss": 1.2736, + "theoretical_loss": 3.521057471077061, + "tokens_seen": 1619394560 + }, + { + "epoch": 0.49, + "learning_rate": 0.00041161885656970914, + "loss": 1.3176, + "theoretical_loss": 3.521045191596631, + "tokens_seen": 1619460096 + }, + { + "epoch": 0.49, + "learning_rate": 0.0004116028084252759, + "loss": 1.4802, + "theoretical_loss": 3.521032912752247, + "tokens_seen": 1619525632 + }, + { + "epoch": 0.49, + "learning_rate": 0.00041158676028084254, + "loss": 1.3862, + "theoretical_loss": 3.521020634543851, + "tokens_seen": 1619591168 + }, + { + "epoch": 0.49, + "learning_rate": 0.0004115707121364093, + "loss": 1.329, + "theoretical_loss": 3.5210083569713824, + "tokens_seen": 1619656704 + }, + { + "epoch": 0.49, + "learning_rate": 0.00041155466399197594, + "loss": 1.4524, + "theoretical_loss": 3.5209960800347844, + "tokens_seen": 1619722240 + }, + { + "epoch": 0.49, + "learning_rate": 0.0004115386158475427, + "loss": 1.3959, + "theoretical_loss": 3.520983803733997, + "tokens_seen": 1619787776 + }, + { + "epoch": 0.49, + "learning_rate": 0.00041152256770310934, + "loss": 1.356, + "theoretical_loss": 3.520971528068962, + "tokens_seen": 1619853312 + }, + { + "epoch": 0.49, + "learning_rate": 0.0004115065195586761, + "loss": 1.393, + "theoretical_loss": 3.5209592530396217, + "tokens_seen": 1619918848 + }, + { + "epoch": 0.49, + "learning_rate": 0.00041149047141424274, + "loss": 1.3891, + "theoretical_loss": 3.5209469786459158, + "tokens_seen": 1619984384 + }, + { + "epoch": 0.49, + "learning_rate": 0.0004114744232698095, + "loss": 1.3444, + "theoretical_loss": 3.5209347048877864, + "tokens_seen": 1620049920 + }, + { + "epoch": 0.49, + "learning_rate": 0.00041145837512537615, + "loss": 1.37, + "theoretical_loss": 3.5209224317651753, + "tokens_seen": 1620115456 + }, + { + "epoch": 0.49, + "learning_rate": 0.00041144232698094285, + "loss": 1.4067, + "theoretical_loss": 3.5209101592780234, + "tokens_seen": 1620180992 + }, + { + "epoch": 0.49, + "learning_rate": 0.00041142627883650955, + "loss": 1.5087, + "theoretical_loss": 3.520897887426272, + "tokens_seen": 1620246528 + }, + { + "epoch": 0.49, + "learning_rate": 0.00041141023069207625, + "loss": 1.3194, + "theoretical_loss": 3.5208856162098634, + "tokens_seen": 1620312064 + }, + { + "epoch": 0.49, + "learning_rate": 0.00041139418254764295, + "loss": 1.3859, + "theoretical_loss": 3.520873345628738, + "tokens_seen": 1620377600 + }, + { + "epoch": 0.49, + "learning_rate": 0.00041137813440320965, + "loss": 1.2682, + "theoretical_loss": 3.5208610756828373, + "tokens_seen": 1620443136 + }, + { + "epoch": 0.49, + "learning_rate": 0.00041136208625877635, + "loss": 1.3718, + "theoretical_loss": 3.5208488063721033, + "tokens_seen": 1620508672 + }, + { + "epoch": 0.49, + "learning_rate": 0.00041134603811434305, + "loss": 1.3488, + "theoretical_loss": 3.5208365376964768, + "tokens_seen": 1620574208 + }, + { + "epoch": 0.49, + "learning_rate": 0.00041132998996990975, + "loss": 1.4453, + "theoretical_loss": 3.5208242696558996, + "tokens_seen": 1620639744 + }, + { + "epoch": 0.49, + "learning_rate": 0.00041131394182547645, + "loss": 1.2651, + "theoretical_loss": 3.520812002250314, + "tokens_seen": 1620705280 + }, + { + "epoch": 0.49, + "learning_rate": 0.0004112978936810431, + "loss": 1.3108, + "theoretical_loss": 3.5207997354796596, + "tokens_seen": 1620770816 + }, + { + "epoch": 0.49, + "learning_rate": 0.00041128184553660985, + "loss": 1.325, + "theoretical_loss": 3.5207874693438788, + "tokens_seen": 1620836352 + }, + { + "epoch": 0.49, + "learning_rate": 0.0004112657973921765, + "loss": 1.3133, + "theoretical_loss": 3.5207752038429136, + "tokens_seen": 1620901888 + }, + { + "epoch": 0.49, + "learning_rate": 0.00041124974924774325, + "loss": 1.3938, + "theoretical_loss": 3.520762938976705, + "tokens_seen": 1620967424 + }, + { + "epoch": 0.49, + "learning_rate": 0.0004112337011033099, + "loss": 1.3489, + "theoretical_loss": 3.5207506747451944, + "tokens_seen": 1621032960 + }, + { + "epoch": 0.49, + "learning_rate": 0.00041121765295887666, + "loss": 1.3858, + "theoretical_loss": 3.520738411148323, + "tokens_seen": 1621098496 + }, + { + "epoch": 0.49, + "learning_rate": 0.0004112016048144433, + "loss": 1.3574, + "theoretical_loss": 3.520726148186033, + "tokens_seen": 1621164032 + }, + { + "epoch": 0.49, + "learning_rate": 0.00041118555667001006, + "loss": 1.3752, + "theoretical_loss": 3.520713885858266, + "tokens_seen": 1621229568 + }, + { + "epoch": 0.49, + "learning_rate": 0.0004111695085255767, + "loss": 1.3185, + "theoretical_loss": 3.5207016241649627, + "tokens_seen": 1621295104 + }, + { + "epoch": 0.49, + "learning_rate": 0.00041115346038114346, + "loss": 1.3936, + "theoretical_loss": 3.520689363106065, + "tokens_seen": 1621360640 + }, + { + "epoch": 0.49, + "learning_rate": 0.0004111374122367101, + "loss": 1.4035, + "theoretical_loss": 3.520677102681515, + "tokens_seen": 1621426176 + }, + { + "epoch": 0.49, + "learning_rate": 0.00041112136409227686, + "loss": 1.4739, + "theoretical_loss": 3.520664842891253, + "tokens_seen": 1621491712 + }, + { + "epoch": 0.49, + "learning_rate": 0.0004111053159478435, + "loss": 1.3436, + "theoretical_loss": 3.5206525837352216, + "tokens_seen": 1621557248 + }, + { + "epoch": 0.49, + "learning_rate": 0.00041108926780341026, + "loss": 1.4487, + "theoretical_loss": 3.5206403252133622, + "tokens_seen": 1621622784 + }, + { + "epoch": 0.49, + "learning_rate": 0.0004110732196589769, + "loss": 1.4339, + "theoretical_loss": 3.5206280673256156, + "tokens_seen": 1621688320 + }, + { + "epoch": 0.49, + "learning_rate": 0.00041105717151454366, + "loss": 1.3168, + "theoretical_loss": 3.5206158100719245, + "tokens_seen": 1621753856 + }, + { + "epoch": 0.49, + "learning_rate": 0.0004110411233701103, + "loss": 1.4135, + "theoretical_loss": 3.52060355345223, + "tokens_seen": 1621819392 + }, + { + "epoch": 0.49, + "learning_rate": 0.00041102507522567706, + "loss": 1.3074, + "theoretical_loss": 3.5205912974664737, + "tokens_seen": 1621884928 + }, + { + "epoch": 0.49, + "learning_rate": 0.0004110090270812437, + "loss": 1.3831, + "theoretical_loss": 3.520579042114597, + "tokens_seen": 1621950464 + }, + { + "epoch": 0.49, + "objective/train/docs_used": 592532, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.284725546836853, + "objective/train/theoretical_loss": 3.5205667873965414, + "objective/train/tokens_used": 1642476000, + "theoretical_loss": 3.5205667873965414, + "tokens_seen": 1622016000 + }, + { + "epoch": 0.49, + "learning_rate": 0.00041099297893681046, + "loss": 1.3193, + "theoretical_loss": 3.5205667873965414, + "tokens_seen": 1622016000 + }, + { + "epoch": 0.49, + "learning_rate": 0.0004109769307923771, + "loss": 1.3438, + "theoretical_loss": 3.520554533312249, + "tokens_seen": 1622081536 + }, + { + "epoch": 0.49, + "learning_rate": 0.00041096088264794387, + "loss": 1.333, + "theoretical_loss": 3.5205422798616617, + "tokens_seen": 1622147072 + }, + { + "epoch": 0.49, + "learning_rate": 0.0004109448345035105, + "loss": 1.3634, + "theoretical_loss": 3.52053002704472, + "tokens_seen": 1622212608 + }, + { + "epoch": 0.49, + "learning_rate": 0.00041092878635907727, + "loss": 1.3198, + "theoretical_loss": 3.520517774861366, + "tokens_seen": 1622278144 + }, + { + "epoch": 0.49, + "learning_rate": 0.0004109127382146439, + "loss": 1.3975, + "theoretical_loss": 3.520505523311542, + "tokens_seen": 1622343680 + }, + { + "epoch": 0.49, + "learning_rate": 0.00041089669007021067, + "loss": 1.4252, + "theoretical_loss": 3.520493272395189, + "tokens_seen": 1622409216 + }, + { + "epoch": 0.49, + "learning_rate": 0.0004108806419257773, + "loss": 1.3123, + "theoretical_loss": 3.5204810221122487, + "tokens_seen": 1622474752 + }, + { + "epoch": 0.49, + "learning_rate": 0.00041086459378134407, + "loss": 1.3274, + "theoretical_loss": 3.520468772462663, + "tokens_seen": 1622540288 + }, + { + "epoch": 0.49, + "learning_rate": 0.0004108485456369107, + "loss": 1.4365, + "theoretical_loss": 3.5204565234463736, + "tokens_seen": 1622605824 + }, + { + "epoch": 0.49, + "learning_rate": 0.00041083249749247747, + "loss": 1.4193, + "theoretical_loss": 3.520444275063322, + "tokens_seen": 1622671360 + }, + { + "epoch": 0.49, + "learning_rate": 0.0004108164493480441, + "loss": 1.4161, + "theoretical_loss": 3.5204320273134493, + "tokens_seen": 1622736896 + }, + { + "epoch": 0.49, + "learning_rate": 0.00041080040120361087, + "loss": 1.3612, + "theoretical_loss": 3.5204197801966983, + "tokens_seen": 1622802432 + }, + { + "epoch": 0.49, + "learning_rate": 0.00041078435305917757, + "loss": 1.3492, + "theoretical_loss": 3.5204075337130103, + "tokens_seen": 1622867968 + }, + { + "epoch": 0.49, + "learning_rate": 0.00041076830491474427, + "loss": 1.2255, + "theoretical_loss": 3.5203952878623266, + "tokens_seen": 1622933504 + }, + { + "epoch": 0.49, + "learning_rate": 0.000410752256770311, + "loss": 1.3928, + "theoretical_loss": 3.52038304264459, + "tokens_seen": 1622999040 + }, + { + "epoch": 0.49, + "learning_rate": 0.0004107362086258777, + "loss": 1.2983, + "theoretical_loss": 3.5203707980597407, + "tokens_seen": 1623064576 + }, + { + "epoch": 0.49, + "learning_rate": 0.0004107201604814444, + "loss": 1.3705, + "theoretical_loss": 3.5203585541077214, + "tokens_seen": 1623130112 + }, + { + "epoch": 0.49, + "learning_rate": 0.0004107041123370111, + "loss": 1.3497, + "theoretical_loss": 3.520346310788474, + "tokens_seen": 1623195648 + }, + { + "epoch": 0.49, + "learning_rate": 0.0004106880641925778, + "loss": 1.373, + "theoretical_loss": 3.5203340681019397, + "tokens_seen": 1623261184 + }, + { + "epoch": 0.49, + "learning_rate": 0.0004106720160481444, + "loss": 1.3605, + "theoretical_loss": 3.5203218260480607, + "tokens_seen": 1623326720 + }, + { + "epoch": 0.49, + "learning_rate": 0.0004106559679037112, + "loss": 1.3641, + "theoretical_loss": 3.5203095846267782, + "tokens_seen": 1623392256 + }, + { + "epoch": 0.49, + "learning_rate": 0.0004106399197592778, + "loss": 1.4386, + "theoretical_loss": 3.5202973438380347, + "tokens_seen": 1623457792 + }, + { + "epoch": 0.49, + "learning_rate": 0.0004106238716148446, + "loss": 1.4182, + "theoretical_loss": 3.5202851036817715, + "tokens_seen": 1623523328 + }, + { + "epoch": 0.49, + "learning_rate": 0.0004106078234704112, + "loss": 1.4184, + "theoretical_loss": 3.5202728641579304, + "tokens_seen": 1623588864 + }, + { + "epoch": 0.49, + "learning_rate": 0.000410591775325978, + "loss": 1.4294, + "theoretical_loss": 3.5202606252664532, + "tokens_seen": 1623654400 + }, + { + "epoch": 0.49, + "learning_rate": 0.0004105757271815446, + "loss": 1.4418, + "theoretical_loss": 3.520248387007282, + "tokens_seen": 1623719936 + }, + { + "epoch": 0.49, + "learning_rate": 0.0004105596790371114, + "loss": 1.3967, + "theoretical_loss": 3.5202361493803584, + "tokens_seen": 1623785472 + }, + { + "epoch": 0.49, + "learning_rate": 0.00041054363089267803, + "loss": 1.3052, + "theoretical_loss": 3.520223912385624, + "tokens_seen": 1623851008 + }, + { + "epoch": 0.49, + "learning_rate": 0.0004105275827482448, + "loss": 1.4321, + "theoretical_loss": 3.5202116760230213, + "tokens_seen": 1623916544 + }, + { + "epoch": 0.49, + "learning_rate": 0.00041051153460381143, + "loss": 1.3213, + "theoretical_loss": 3.5201994402924917, + "tokens_seen": 1623982080 + }, + { + "epoch": 0.49, + "learning_rate": 0.0004104954864593782, + "loss": 1.3318, + "theoretical_loss": 3.5201872051939764, + "tokens_seen": 1624047616 + }, + { + "epoch": 0.49, + "learning_rate": 0.00041047943831494483, + "loss": 1.4447, + "theoretical_loss": 3.5201749707274184, + "tokens_seen": 1624113152 + }, + { + "epoch": 0.49, + "learning_rate": 0.0004104633901705116, + "loss": 1.4112, + "theoretical_loss": 3.5201627368927593, + "tokens_seen": 1624178688 + }, + { + "epoch": 0.49, + "learning_rate": 0.00041044734202607823, + "loss": 1.3976, + "theoretical_loss": 3.5201505036899405, + "tokens_seen": 1624244224 + }, + { + "epoch": 0.49, + "learning_rate": 0.000410431293881645, + "loss": 1.3394, + "theoretical_loss": 3.520138271118904, + "tokens_seen": 1624309760 + }, + { + "epoch": 0.49, + "learning_rate": 0.00041041524573721163, + "loss": 1.4263, + "theoretical_loss": 3.520126039179592, + "tokens_seen": 1624375296 + }, + { + "epoch": 0.49, + "learning_rate": 0.0004103991975927784, + "loss": 1.4435, + "theoretical_loss": 3.520113807871946, + "tokens_seen": 1624440832 + }, + { + "epoch": 0.49, + "learning_rate": 0.00041038314944834503, + "loss": 1.2545, + "theoretical_loss": 3.5201015771959083, + "tokens_seen": 1624506368 + }, + { + "epoch": 0.49, + "learning_rate": 0.0004103671013039118, + "loss": 1.3897, + "theoretical_loss": 3.52008934715142, + "tokens_seen": 1624571904 + }, + { + "epoch": 0.49, + "learning_rate": 0.00041035105315947844, + "loss": 1.3198, + "theoretical_loss": 3.520077117738424, + "tokens_seen": 1624637440 + }, + { + "epoch": 0.49, + "learning_rate": 0.0004103350050150452, + "loss": 1.3501, + "theoretical_loss": 3.5200648889568624, + "tokens_seen": 1624702976 + }, + { + "epoch": 0.49, + "learning_rate": 0.00041031895687061184, + "loss": 1.3423, + "theoretical_loss": 3.520052660806676, + "tokens_seen": 1624768512 + }, + { + "epoch": 0.49, + "learning_rate": 0.0004103029087261786, + "loss": 1.3822, + "theoretical_loss": 3.5200404332878072, + "tokens_seen": 1624834048 + }, + { + "epoch": 0.49, + "learning_rate": 0.00041028686058174524, + "loss": 1.4983, + "theoretical_loss": 3.520028206400198, + "tokens_seen": 1624899584 + }, + { + "epoch": 0.49, + "learning_rate": 0.000410270812437312, + "loss": 1.3398, + "theoretical_loss": 3.5200159801437905, + "tokens_seen": 1624965120 + }, + { + "epoch": 0.49, + "learning_rate": 0.00041025476429287864, + "loss": 1.447, + "theoretical_loss": 3.5200037545185268, + "tokens_seen": 1625030656 + }, + { + "epoch": 0.49, + "learning_rate": 0.0004102387161484454, + "loss": 1.3507, + "theoretical_loss": 3.519991529524348, + "tokens_seen": 1625096192 + }, + { + "epoch": 0.49, + "learning_rate": 0.00041022266800401204, + "loss": 1.4374, + "theoretical_loss": 3.5199793051611974, + "tokens_seen": 1625161728 + }, + { + "epoch": 0.49, + "learning_rate": 0.0004102066198595788, + "loss": 1.3404, + "theoretical_loss": 3.519967081429016, + "tokens_seen": 1625227264 + }, + { + "epoch": 0.49, + "objective/train/docs_used": 593743, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.577528476715088, + "objective/train/theoretical_loss": 3.519954858327746, + "objective/train/tokens_used": 1645752800, + "theoretical_loss": 3.519954858327746, + "tokens_seen": 1625292800 + }, + { + "epoch": 0.49, + "learning_rate": 0.00041019057171514544, + "loss": 1.4595, + "theoretical_loss": 3.519954858327746, + "tokens_seen": 1625292800 + }, + { + "epoch": 0.49, + "learning_rate": 0.0004101745235707122, + "loss": 1.3787, + "theoretical_loss": 3.5199426358573294, + "tokens_seen": 1625358336 + }, + { + "epoch": 0.49, + "learning_rate": 0.00041015847542627884, + "loss": 1.4315, + "theoretical_loss": 3.5199304140177086, + "tokens_seen": 1625423872 + }, + { + "epoch": 0.49, + "learning_rate": 0.0004101424272818456, + "loss": 1.5641, + "theoretical_loss": 3.5199181928088246, + "tokens_seen": 1625489408 + }, + { + "epoch": 0.49, + "learning_rate": 0.00041012637913741224, + "loss": 1.345, + "theoretical_loss": 3.519905972230621, + "tokens_seen": 1625554944 + }, + { + "epoch": 0.49, + "learning_rate": 0.000410110330992979, + "loss": 1.4068, + "theoretical_loss": 3.5198937522830382, + "tokens_seen": 1625620480 + }, + { + "epoch": 0.49, + "learning_rate": 0.00041009428284854565, + "loss": 1.3865, + "theoretical_loss": 3.5198815329660196, + "tokens_seen": 1625686016 + }, + { + "epoch": 0.49, + "learning_rate": 0.0004100782347041124, + "loss": 1.3754, + "theoretical_loss": 3.519869314279506, + "tokens_seen": 1625751552 + }, + { + "epoch": 0.49, + "learning_rate": 0.00041006218655967905, + "loss": 1.3604, + "theoretical_loss": 3.51985709622344, + "tokens_seen": 1625817088 + }, + { + "epoch": 0.49, + "learning_rate": 0.00041004613841524575, + "loss": 1.2498, + "theoretical_loss": 3.5198448787977643, + "tokens_seen": 1625882624 + }, + { + "epoch": 0.49, + "learning_rate": 0.00041003009027081245, + "loss": 1.3919, + "theoretical_loss": 3.51983266200242, + "tokens_seen": 1625948160 + }, + { + "epoch": 0.49, + "learning_rate": 0.00041001404212637915, + "loss": 1.3417, + "theoretical_loss": 3.51982044583735, + "tokens_seen": 1626013696 + }, + { + "epoch": 0.49, + "learning_rate": 0.00040999799398194585, + "loss": 1.3898, + "theoretical_loss": 3.5198082303024956, + "tokens_seen": 1626079232 + }, + { + "epoch": 0.49, + "learning_rate": 0.00040998194583751255, + "loss": 1.318, + "theoretical_loss": 3.519796015397799, + "tokens_seen": 1626144768 + }, + { + "epoch": 0.49, + "learning_rate": 0.00040996589769307925, + "loss": 1.4473, + "theoretical_loss": 3.5197838011232028, + "tokens_seen": 1626210304 + }, + { + "epoch": 0.49, + "learning_rate": 0.00040994984954864595, + "loss": 1.466, + "theoretical_loss": 3.5197715874786493, + "tokens_seen": 1626275840 + }, + { + "epoch": 0.49, + "learning_rate": 0.00040993380140421265, + "loss": 1.3411, + "theoretical_loss": 3.51975937446408, + "tokens_seen": 1626341376 + }, + { + "epoch": 0.49, + "learning_rate": 0.00040991775325977935, + "loss": 1.369, + "theoretical_loss": 3.5197471620794367, + "tokens_seen": 1626406912 + }, + { + "epoch": 0.49, + "learning_rate": 0.000409901705115346, + "loss": 1.332, + "theoretical_loss": 3.519734950324662, + "tokens_seen": 1626472448 + }, + { + "epoch": 0.49, + "learning_rate": 0.00040988565697091275, + "loss": 1.3164, + "theoretical_loss": 3.519722739199698, + "tokens_seen": 1626537984 + }, + { + "epoch": 0.49, + "learning_rate": 0.0004098696088264794, + "loss": 1.4192, + "theoretical_loss": 3.519710528704487, + "tokens_seen": 1626603520 + }, + { + "epoch": 0.49, + "learning_rate": 0.00040985356068204615, + "loss": 1.5193, + "theoretical_loss": 3.5196983188389717, + "tokens_seen": 1626669056 + }, + { + "epoch": 0.49, + "learning_rate": 0.0004098375125376128, + "loss": 1.3183, + "theoretical_loss": 3.519686109603093, + "tokens_seen": 1626734592 + }, + { + "epoch": 0.49, + "learning_rate": 0.00040982146439317956, + "loss": 1.3746, + "theoretical_loss": 3.5196739009967937, + "tokens_seen": 1626800128 + }, + { + "epoch": 0.49, + "learning_rate": 0.0004098054162487462, + "loss": 1.4271, + "theoretical_loss": 3.519661693020016, + "tokens_seen": 1626865664 + }, + { + "epoch": 0.49, + "learning_rate": 0.00040978936810431296, + "loss": 1.2803, + "theoretical_loss": 3.5196494856727023, + "tokens_seen": 1626931200 + }, + { + "epoch": 0.49, + "learning_rate": 0.0004097733199598796, + "loss": 1.4683, + "theoretical_loss": 3.519637278954794, + "tokens_seen": 1626996736 + }, + { + "epoch": 0.49, + "learning_rate": 0.00040975727181544636, + "loss": 1.4254, + "theoretical_loss": 3.5196250728662344, + "tokens_seen": 1627062272 + }, + { + "epoch": 0.49, + "learning_rate": 0.000409741223671013, + "loss": 1.3839, + "theoretical_loss": 3.5196128674069644, + "tokens_seen": 1627127808 + }, + { + "epoch": 0.49, + "learning_rate": 0.00040972517552657976, + "loss": 1.3419, + "theoretical_loss": 3.519600662576927, + "tokens_seen": 1627193344 + }, + { + "epoch": 0.49, + "learning_rate": 0.0004097091273821464, + "loss": 1.2998, + "theoretical_loss": 3.519588458376065, + "tokens_seen": 1627258880 + }, + { + "epoch": 0.49, + "learning_rate": 0.00040969307923771316, + "loss": 1.3214, + "theoretical_loss": 3.5195762548043197, + "tokens_seen": 1627324416 + }, + { + "epoch": 0.49, + "learning_rate": 0.0004096770310932798, + "loss": 1.4022, + "theoretical_loss": 3.5195640518616336, + "tokens_seen": 1627389952 + }, + { + "epoch": 0.49, + "learning_rate": 0.00040966098294884656, + "loss": 1.3526, + "theoretical_loss": 3.5195518495479488, + "tokens_seen": 1627455488 + }, + { + "epoch": 0.49, + "learning_rate": 0.0004096449348044132, + "loss": 1.3488, + "theoretical_loss": 3.5195396478632075, + "tokens_seen": 1627521024 + }, + { + "epoch": 0.49, + "learning_rate": 0.00040962888665997996, + "loss": 1.4672, + "theoretical_loss": 3.519527446807353, + "tokens_seen": 1627586560 + }, + { + "epoch": 0.49, + "learning_rate": 0.0004096128385155466, + "loss": 1.4186, + "theoretical_loss": 3.519515246380326, + "tokens_seen": 1627652096 + }, + { + "epoch": 0.49, + "learning_rate": 0.00040959679037111336, + "loss": 1.4747, + "theoretical_loss": 3.519503046582069, + "tokens_seen": 1627717632 + }, + { + "epoch": 0.49, + "learning_rate": 0.0004095807422266801, + "loss": 1.3026, + "theoretical_loss": 3.519490847412526, + "tokens_seen": 1627783168 + }, + { + "epoch": 0.49, + "learning_rate": 0.00040956469408224677, + "loss": 1.4261, + "theoretical_loss": 3.5194786488716376, + "tokens_seen": 1627848704 + }, + { + "epoch": 0.49, + "learning_rate": 0.0004095486459378135, + "loss": 1.3368, + "theoretical_loss": 3.5194664509593463, + "tokens_seen": 1627914240 + }, + { + "epoch": 0.49, + "learning_rate": 0.00040953259779338017, + "loss": 1.377, + "theoretical_loss": 3.519454253675595, + "tokens_seen": 1627979776 + }, + { + "epoch": 0.49, + "learning_rate": 0.0004095165496489469, + "loss": 1.3884, + "theoretical_loss": 3.519442057020325, + "tokens_seen": 1628045312 + }, + { + "epoch": 0.49, + "learning_rate": 0.00040950050150451357, + "loss": 1.4609, + "theoretical_loss": 3.51942986099348, + "tokens_seen": 1628110848 + }, + { + "epoch": 0.49, + "learning_rate": 0.0004094844533600803, + "loss": 1.3251, + "theoretical_loss": 3.519417665595002, + "tokens_seen": 1628176384 + }, + { + "epoch": 0.49, + "learning_rate": 0.00040946840521564697, + "loss": 1.2878, + "theoretical_loss": 3.5194054708248323, + "tokens_seen": 1628241920 + }, + { + "epoch": 0.49, + "learning_rate": 0.0004094523570712137, + "loss": 1.3833, + "theoretical_loss": 3.519393276682914, + "tokens_seen": 1628307456 + }, + { + "epoch": 0.49, + "learning_rate": 0.00040943630892678037, + "loss": 1.2967, + "theoretical_loss": 3.5193810831691894, + "tokens_seen": 1628372992 + }, + { + "epoch": 0.49, + "learning_rate": 0.00040942026078234707, + "loss": 1.4058, + "theoretical_loss": 3.5193688902836007, + "tokens_seen": 1628438528 + }, + { + "epoch": 0.49, + "learning_rate": 0.00040940421263791377, + "loss": 1.3888, + "theoretical_loss": 3.5193566980260904, + "tokens_seen": 1628504064 + }, + { + "epoch": 0.49, + "objective/train/docs_used": 595003, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.3385471105575562, + "objective/train/theoretical_loss": 3.5193445063966013, + "objective/train/tokens_used": 1649029600, + "theoretical_loss": 3.5193445063966013, + "tokens_seen": 1628569600 + }, + { + "epoch": 0.49, + "learning_rate": 0.0004093881644934805, + "loss": 1.2933, + "theoretical_loss": 3.5193445063966013, + "tokens_seen": 1628569600 + }, + { + "epoch": 0.49, + "learning_rate": 0.0004093721163490472, + "loss": 1.3775, + "theoretical_loss": 3.519332315395075, + "tokens_seen": 1628635136 + }, + { + "epoch": 0.49, + "learning_rate": 0.0004093560682046139, + "loss": 1.461, + "theoretical_loss": 3.519320125021454, + "tokens_seen": 1628700672 + }, + { + "epoch": 0.49, + "learning_rate": 0.0004093400200601806, + "loss": 1.3381, + "theoretical_loss": 3.5193079352756813, + "tokens_seen": 1628766208 + }, + { + "epoch": 0.49, + "learning_rate": 0.0004093239719157473, + "loss": 1.3195, + "theoretical_loss": 3.519295746157699, + "tokens_seen": 1628831744 + }, + { + "epoch": 0.49, + "learning_rate": 0.000409307923771314, + "loss": 1.4341, + "theoretical_loss": 3.519283557667449, + "tokens_seen": 1628897280 + }, + { + "epoch": 0.49, + "learning_rate": 0.0004092918756268807, + "loss": 1.3431, + "theoretical_loss": 3.519271369804874, + "tokens_seen": 1628962816 + }, + { + "epoch": 0.49, + "learning_rate": 0.0004092758274824473, + "loss": 1.3975, + "theoretical_loss": 3.5192591825699173, + "tokens_seen": 1629028352 + }, + { + "epoch": 0.49, + "learning_rate": 0.0004092597793380141, + "loss": 1.3825, + "theoretical_loss": 3.5192469959625203, + "tokens_seen": 1629093888 + }, + { + "epoch": 0.49, + "learning_rate": 0.0004092437311935807, + "loss": 1.3589, + "theoretical_loss": 3.5192348099826254, + "tokens_seen": 1629159424 + }, + { + "epoch": 0.49, + "learning_rate": 0.0004092276830491475, + "loss": 1.4627, + "theoretical_loss": 3.5192226246301757, + "tokens_seen": 1629224960 + }, + { + "epoch": 0.49, + "learning_rate": 0.0004092116349047141, + "loss": 1.4588, + "theoretical_loss": 3.5192104399051134, + "tokens_seen": 1629290496 + }, + { + "epoch": 0.49, + "learning_rate": 0.0004091955867602809, + "loss": 1.3085, + "theoretical_loss": 3.519198255807381, + "tokens_seen": 1629356032 + }, + { + "epoch": 0.49, + "learning_rate": 0.00040917953861584753, + "loss": 1.3742, + "theoretical_loss": 3.5191860723369204, + "tokens_seen": 1629421568 + }, + { + "epoch": 0.49, + "learning_rate": 0.0004091634904714143, + "loss": 1.3693, + "theoretical_loss": 3.5191738894936746, + "tokens_seen": 1629487104 + }, + { + "epoch": 0.49, + "learning_rate": 0.00040914744232698093, + "loss": 1.3711, + "theoretical_loss": 3.5191617072775863, + "tokens_seen": 1629552640 + }, + { + "epoch": 0.49, + "learning_rate": 0.0004091313941825477, + "loss": 1.3149, + "theoretical_loss": 3.519149525688598, + "tokens_seen": 1629618176 + }, + { + "epoch": 0.49, + "learning_rate": 0.00040911534603811433, + "loss": 1.2922, + "theoretical_loss": 3.519137344726652, + "tokens_seen": 1629683712 + }, + { + "epoch": 0.49, + "learning_rate": 0.0004090992978936811, + "loss": 1.3911, + "theoretical_loss": 3.5191251643916903, + "tokens_seen": 1629749248 + }, + { + "epoch": 0.49, + "learning_rate": 0.00040908324974924773, + "loss": 1.3313, + "theoretical_loss": 3.519112984683656, + "tokens_seen": 1629814784 + }, + { + "epoch": 0.49, + "learning_rate": 0.0004090672016048145, + "loss": 1.3178, + "theoretical_loss": 3.5191008056024917, + "tokens_seen": 1629880320 + }, + { + "epoch": 0.49, + "learning_rate": 0.00040905115346038113, + "loss": 1.3626, + "theoretical_loss": 3.5190886271481396, + "tokens_seen": 1629945856 + }, + { + "epoch": 0.49, + "learning_rate": 0.0004090351053159479, + "loss": 1.3166, + "theoretical_loss": 3.5190764493205426, + "tokens_seen": 1630011392 + }, + { + "epoch": 0.49, + "learning_rate": 0.00040901905717151453, + "loss": 1.3978, + "theoretical_loss": 3.5190642721196426, + "tokens_seen": 1630076928 + }, + { + "epoch": 0.49, + "learning_rate": 0.0004090030090270813, + "loss": 1.4222, + "theoretical_loss": 3.5190520955453826, + "tokens_seen": 1630142464 + }, + { + "epoch": 0.49, + "learning_rate": 0.00040898696088264793, + "loss": 1.315, + "theoretical_loss": 3.519039919597706, + "tokens_seen": 1630208000 + }, + { + "epoch": 0.49, + "learning_rate": 0.0004089709127382147, + "loss": 1.4046, + "theoretical_loss": 3.5190277442765536, + "tokens_seen": 1630273536 + }, + { + "epoch": 0.49, + "learning_rate": 0.00040895486459378134, + "loss": 1.3689, + "theoretical_loss": 3.519015569581869, + "tokens_seen": 1630339072 + }, + { + "epoch": 0.49, + "learning_rate": 0.0004089388164493481, + "loss": 1.258, + "theoretical_loss": 3.5190033955135944, + "tokens_seen": 1630404608 + }, + { + "epoch": 0.49, + "learning_rate": 0.00040892276830491474, + "loss": 1.409, + "theoretical_loss": 3.5189912220716733, + "tokens_seen": 1630470144 + }, + { + "epoch": 0.49, + "learning_rate": 0.0004089067201604815, + "loss": 1.3323, + "theoretical_loss": 3.518979049256047, + "tokens_seen": 1630535680 + }, + { + "epoch": 0.49, + "learning_rate": 0.00040889067201604814, + "loss": 1.3603, + "theoretical_loss": 3.5189668770666596, + "tokens_seen": 1630601216 + }, + { + "epoch": 0.49, + "learning_rate": 0.0004088746238716149, + "loss": 1.3581, + "theoretical_loss": 3.518954705503452, + "tokens_seen": 1630666752 + }, + { + "epoch": 0.49, + "learning_rate": 0.00040885857572718154, + "loss": 1.4272, + "theoretical_loss": 3.5189425345663685, + "tokens_seen": 1630732288 + }, + { + "epoch": 0.49, + "learning_rate": 0.0004088425275827483, + "loss": 1.4101, + "theoretical_loss": 3.5189303642553504, + "tokens_seen": 1630797824 + }, + { + "epoch": 0.49, + "learning_rate": 0.00040882647943831494, + "loss": 1.3993, + "theoretical_loss": 3.5189181945703414, + "tokens_seen": 1630863360 + }, + { + "epoch": 0.49, + "learning_rate": 0.0004088104312938817, + "loss": 1.3614, + "theoretical_loss": 3.518906025511283, + "tokens_seen": 1630928896 + }, + { + "epoch": 0.49, + "learning_rate": 0.00040879438314944834, + "loss": 1.3138, + "theoretical_loss": 3.518893857078119, + "tokens_seen": 1630994432 + }, + { + "epoch": 0.49, + "learning_rate": 0.0004087783350050151, + "loss": 1.4083, + "theoretical_loss": 3.518881689270791, + "tokens_seen": 1631059968 + }, + { + "epoch": 0.49, + "learning_rate": 0.00040876228686058174, + "loss": 1.3111, + "theoretical_loss": 3.518869522089243, + "tokens_seen": 1631125504 + }, + { + "epoch": 0.49, + "learning_rate": 0.0004087462387161485, + "loss": 1.354, + "theoretical_loss": 3.5188573555334166, + "tokens_seen": 1631191040 + }, + { + "epoch": 0.49, + "learning_rate": 0.00040873019057171515, + "loss": 1.2242, + "theoretical_loss": 3.518845189603254, + "tokens_seen": 1631256576 + }, + { + "epoch": 0.49, + "learning_rate": 0.0004087141424272819, + "loss": 1.3006, + "theoretical_loss": 3.5188330242987, + "tokens_seen": 1631322112 + }, + { + "epoch": 0.49, + "learning_rate": 0.00040869809428284855, + "loss": 1.4296, + "theoretical_loss": 3.5188208596196944, + "tokens_seen": 1631387648 + }, + { + "epoch": 0.49, + "learning_rate": 0.0004086820461384153, + "loss": 1.2739, + "theoretical_loss": 3.5188086955661824, + "tokens_seen": 1631453184 + }, + { + "epoch": 0.49, + "learning_rate": 0.00040866599799398195, + "loss": 1.4431, + "theoretical_loss": 3.518796532138106, + "tokens_seen": 1631518720 + }, + { + "epoch": 0.49, + "learning_rate": 0.00040864994984954865, + "loss": 1.3903, + "theoretical_loss": 3.518784369335407, + "tokens_seen": 1631584256 + }, + { + "epoch": 0.49, + "learning_rate": 0.00040863390170511535, + "loss": 1.4481, + "theoretical_loss": 3.5187722071580296, + "tokens_seen": 1631649792 + }, + { + "epoch": 0.49, + "learning_rate": 0.00040861785356068205, + "loss": 1.3367, + "theoretical_loss": 3.518760045605915, + "tokens_seen": 1631715328 + }, + { + "epoch": 0.49, + "learning_rate": 0.00040860180541624875, + "loss": 1.344, + "theoretical_loss": 3.518747884679007, + "tokens_seen": 1631780864 + }, + { + "epoch": 0.49, + "objective/train/docs_used": 596193, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.4816763401031494, + "objective/train/theoretical_loss": 3.5187357243772484, + "objective/train/tokens_used": 1652306400, + "theoretical_loss": 3.5187357243772484, + "tokens_seen": 1631846400 + }, + { + "epoch": 0.49, + "learning_rate": 0.00040858575727181545, + "loss": 1.4181, + "theoretical_loss": 3.5187357243772484, + "tokens_seen": 1631846400 + }, + { + "epoch": 0.49, + "learning_rate": 0.00040856970912738215, + "loss": 1.408, + "theoretical_loss": 3.518723564700581, + "tokens_seen": 1631911936 + }, + { + "epoch": 0.49, + "learning_rate": 0.00040855366098294885, + "loss": 1.4103, + "theoretical_loss": 3.518711405648949, + "tokens_seen": 1631977472 + }, + { + "epoch": 0.49, + "learning_rate": 0.00040853761283851555, + "loss": 1.356, + "theoretical_loss": 3.518699247222293, + "tokens_seen": 1632043008 + }, + { + "epoch": 0.49, + "learning_rate": 0.00040852156469408225, + "loss": 1.3966, + "theoretical_loss": 3.5186870894205584, + "tokens_seen": 1632108544 + }, + { + "epoch": 0.49, + "learning_rate": 0.0004085055165496489, + "loss": 1.4121, + "theoretical_loss": 3.518674932243686, + "tokens_seen": 1632174080 + }, + { + "epoch": 0.49, + "learning_rate": 0.00040848946840521565, + "loss": 1.3383, + "theoretical_loss": 3.51866277569162, + "tokens_seen": 1632239616 + }, + { + "epoch": 0.49, + "learning_rate": 0.0004084734202607823, + "loss": 1.3917, + "theoretical_loss": 3.518650619764302, + "tokens_seen": 1632305152 + }, + { + "epoch": 0.49, + "learning_rate": 0.00040845737211634906, + "loss": 1.3841, + "theoretical_loss": 3.5186384644616755, + "tokens_seen": 1632370688 + }, + { + "epoch": 0.49, + "learning_rate": 0.0004084413239719157, + "loss": 1.4505, + "theoretical_loss": 3.518626309783683, + "tokens_seen": 1632436224 + }, + { + "epoch": 0.49, + "learning_rate": 0.00040842527582748246, + "loss": 1.4276, + "theoretical_loss": 3.5186141557302677, + "tokens_seen": 1632501760 + }, + { + "epoch": 0.49, + "learning_rate": 0.0004084092276830491, + "loss": 1.443, + "theoretical_loss": 3.518602002301372, + "tokens_seen": 1632567296 + }, + { + "epoch": 0.49, + "learning_rate": 0.00040839317953861586, + "loss": 1.409, + "theoretical_loss": 3.518589849496939, + "tokens_seen": 1632632832 + }, + { + "epoch": 0.49, + "learning_rate": 0.0004083771313941826, + "loss": 1.39, + "theoretical_loss": 3.518577697316912, + "tokens_seen": 1632698368 + }, + { + "epoch": 0.49, + "learning_rate": 0.00040836108324974926, + "loss": 1.4028, + "theoretical_loss": 3.5185655457612324, + "tokens_seen": 1632763904 + }, + { + "epoch": 0.49, + "learning_rate": 0.000408345035105316, + "loss": 1.3815, + "theoretical_loss": 3.518553394829845, + "tokens_seen": 1632829440 + }, + { + "epoch": 0.49, + "learning_rate": 0.00040832898696088266, + "loss": 1.2718, + "theoretical_loss": 3.518541244522691, + "tokens_seen": 1632894976 + }, + { + "epoch": 0.49, + "learning_rate": 0.0004083129388164494, + "loss": 1.2841, + "theoretical_loss": 3.5185290948397148, + "tokens_seen": 1632960512 + }, + { + "epoch": 0.49, + "learning_rate": 0.00040829689067201606, + "loss": 1.4098, + "theoretical_loss": 3.518516945780858, + "tokens_seen": 1633026048 + }, + { + "epoch": 0.49, + "learning_rate": 0.0004082808425275828, + "loss": 1.3758, + "theoretical_loss": 3.518504797346064, + "tokens_seen": 1633091584 + }, + { + "epoch": 0.49, + "learning_rate": 0.00040826479438314946, + "loss": 1.3362, + "theoretical_loss": 3.5184926495352755, + "tokens_seen": 1633157120 + }, + { + "epoch": 0.49, + "learning_rate": 0.0004082487462387162, + "loss": 1.2977, + "theoretical_loss": 3.518480502348435, + "tokens_seen": 1633222656 + }, + { + "epoch": 0.49, + "learning_rate": 0.00040823269809428286, + "loss": 1.3484, + "theoretical_loss": 3.518468355785487, + "tokens_seen": 1633288192 + }, + { + "epoch": 0.49, + "learning_rate": 0.0004082166499498496, + "loss": 1.3615, + "theoretical_loss": 3.518456209846373, + "tokens_seen": 1633353728 + }, + { + "epoch": 0.49, + "learning_rate": 0.00040820060180541627, + "loss": 1.4587, + "theoretical_loss": 3.5184440645310366, + "tokens_seen": 1633419264 + }, + { + "epoch": 0.49, + "learning_rate": 0.000408184553660983, + "loss": 1.3424, + "theoretical_loss": 3.5184319198394203, + "tokens_seen": 1633484800 + }, + { + "epoch": 0.5, + "learning_rate": 0.00040816850551654967, + "loss": 1.3783, + "theoretical_loss": 3.5184197757714673, + "tokens_seen": 1633550336 + }, + { + "epoch": 0.5, + "learning_rate": 0.0004081524573721164, + "loss": 1.4326, + "theoretical_loss": 3.5184076323271203, + "tokens_seen": 1633615872 + }, + { + "epoch": 0.5, + "learning_rate": 0.00040813640922768307, + "loss": 1.2903, + "theoretical_loss": 3.5183954895063225, + "tokens_seen": 1633681408 + }, + { + "epoch": 0.5, + "learning_rate": 0.0004081203610832498, + "loss": 1.3579, + "theoretical_loss": 3.518383347309017, + "tokens_seen": 1633746944 + }, + { + "epoch": 0.5, + "learning_rate": 0.00040810431293881647, + "loss": 1.3633, + "theoretical_loss": 3.5183712057351464, + "tokens_seen": 1633812480 + }, + { + "epoch": 0.5, + "learning_rate": 0.0004080882647943832, + "loss": 1.282, + "theoretical_loss": 3.5183590647846543, + "tokens_seen": 1633878016 + }, + { + "epoch": 0.5, + "learning_rate": 0.00040807221664994987, + "loss": 1.3347, + "theoretical_loss": 3.5183469244574828, + "tokens_seen": 1633943552 + }, + { + "epoch": 0.5, + "learning_rate": 0.0004080561685055166, + "loss": 1.3449, + "theoretical_loss": 3.5183347847535753, + "tokens_seen": 1634009088 + }, + { + "epoch": 0.5, + "learning_rate": 0.00040804012036108327, + "loss": 1.4026, + "theoretical_loss": 3.5183226456728756, + "tokens_seen": 1634074624 + }, + { + "epoch": 0.5, + "learning_rate": 0.00040802407221665, + "loss": 1.3503, + "theoretical_loss": 3.5183105072153253, + "tokens_seen": 1634140160 + }, + { + "epoch": 0.5, + "learning_rate": 0.0004080080240722167, + "loss": 1.3511, + "theoretical_loss": 3.5182983693808683, + "tokens_seen": 1634205696 + }, + { + "epoch": 0.5, + "learning_rate": 0.0004079919759277834, + "loss": 1.3852, + "theoretical_loss": 3.518286232169447, + "tokens_seen": 1634271232 + }, + { + "epoch": 0.5, + "learning_rate": 0.0004079759277833501, + "loss": 1.3101, + "theoretical_loss": 3.5182740955810057, + "tokens_seen": 1634336768 + }, + { + "epoch": 0.5, + "learning_rate": 0.0004079598796389168, + "loss": 1.3579, + "theoretical_loss": 3.5182619596154865, + "tokens_seen": 1634402304 + }, + { + "epoch": 0.5, + "learning_rate": 0.0004079438314944835, + "loss": 1.4088, + "theoretical_loss": 3.518249824272832, + "tokens_seen": 1634467840 + }, + { + "epoch": 0.5, + "learning_rate": 0.0004079277833500502, + "loss": 1.3451, + "theoretical_loss": 3.5182376895529863, + "tokens_seen": 1634533376 + }, + { + "epoch": 0.5, + "learning_rate": 0.0004079117352056169, + "loss": 1.3375, + "theoretical_loss": 3.518225555455892, + "tokens_seen": 1634598912 + }, + { + "epoch": 0.5, + "learning_rate": 0.0004078956870611836, + "loss": 1.2944, + "theoretical_loss": 3.5182134219814922, + "tokens_seen": 1634664448 + }, + { + "epoch": 0.5, + "learning_rate": 0.0004078796389167502, + "loss": 1.341, + "theoretical_loss": 3.51820128912973, + "tokens_seen": 1634729984 + }, + { + "epoch": 0.5, + "learning_rate": 0.000407863590772317, + "loss": 1.4136, + "theoretical_loss": 3.518189156900548, + "tokens_seen": 1634795520 + }, + { + "epoch": 0.5, + "learning_rate": 0.0004078475426278836, + "loss": 1.3565, + "theoretical_loss": 3.51817702529389, + "tokens_seen": 1634861056 + }, + { + "epoch": 0.5, + "learning_rate": 0.0004078314944834504, + "loss": 1.3798, + "theoretical_loss": 3.5181648943096993, + "tokens_seen": 1634926592 + }, + { + "epoch": 0.5, + "learning_rate": 0.000407815446339017, + "loss": 1.3914, + "theoretical_loss": 3.518152763947918, + "tokens_seen": 1634992128 + }, + { + "epoch": 0.5, + "learning_rate": 0.0004077993981945838, + "loss": 1.3792, + "theoretical_loss": 3.51814063420849, + "tokens_seen": 1635057664 + }, + { + "epoch": 0.5, + "objective/train/docs_used": 597373, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.3973095417022705, + "objective/train/theoretical_loss": 3.518128505091358, + "objective/train/tokens_used": 1655583200, + "theoretical_loss": 3.518128505091358, + "tokens_seen": 1635123200 + }, + { + "epoch": 0.5, + "learning_rate": 0.00040778335005015043, + "loss": 1.3459, + "theoretical_loss": 3.518128505091358, + "tokens_seen": 1635123200 + }, + { + "epoch": 0.5, + "learning_rate": 0.0004077673019057172, + "loss": 1.3643, + "theoretical_loss": 3.5181163765964656, + "tokens_seen": 1635188736 + }, + { + "epoch": 0.5, + "learning_rate": 0.00040775125376128383, + "loss": 1.3907, + "theoretical_loss": 3.5181042487237555, + "tokens_seen": 1635254272 + }, + { + "epoch": 0.5, + "learning_rate": 0.0004077352056168506, + "loss": 1.4255, + "theoretical_loss": 3.5180921214731713, + "tokens_seen": 1635319808 + }, + { + "epoch": 0.5, + "learning_rate": 0.00040771915747241723, + "loss": 1.2416, + "theoretical_loss": 3.5180799948446557, + "tokens_seen": 1635385344 + }, + { + "epoch": 0.5, + "learning_rate": 0.000407703109327984, + "loss": 1.4707, + "theoretical_loss": 3.5180678688381524, + "tokens_seen": 1635450880 + }, + { + "epoch": 0.5, + "learning_rate": 0.00040768706118355063, + "loss": 1.2987, + "theoretical_loss": 3.518055743453604, + "tokens_seen": 1635516416 + }, + { + "epoch": 0.5, + "learning_rate": 0.0004076710130391174, + "loss": 1.3014, + "theoretical_loss": 3.5180436186909536, + "tokens_seen": 1635581952 + }, + { + "epoch": 0.5, + "learning_rate": 0.00040765496489468403, + "loss": 1.2819, + "theoretical_loss": 3.518031494550145, + "tokens_seen": 1635647488 + }, + { + "epoch": 0.5, + "learning_rate": 0.0004076389167502508, + "loss": 1.3092, + "theoretical_loss": 3.518019371031121, + "tokens_seen": 1635713024 + }, + { + "epoch": 0.5, + "learning_rate": 0.00040762286860581743, + "loss": 1.2996, + "theoretical_loss": 3.518007248133825, + "tokens_seen": 1635778560 + }, + { + "epoch": 0.5, + "learning_rate": 0.0004076068204613842, + "loss": 1.3351, + "theoretical_loss": 3.5179951258582, + "tokens_seen": 1635844096 + }, + { + "epoch": 0.5, + "learning_rate": 0.00040759077231695084, + "loss": 1.3358, + "theoretical_loss": 3.517983004204189, + "tokens_seen": 1635909632 + }, + { + "epoch": 0.5, + "learning_rate": 0.0004075747241725176, + "loss": 1.2047, + "theoretical_loss": 3.5179708831717362, + "tokens_seen": 1635975168 + }, + { + "epoch": 0.5, + "learning_rate": 0.00040755867602808424, + "loss": 1.2874, + "theoretical_loss": 3.5179587627607836, + "tokens_seen": 1636040704 + }, + { + "epoch": 0.5, + "learning_rate": 0.000407542627883651, + "loss": 1.3987, + "theoretical_loss": 3.5179466429712756, + "tokens_seen": 1636106240 + }, + { + "epoch": 0.5, + "learning_rate": 0.00040752657973921764, + "loss": 1.4215, + "theoretical_loss": 3.5179345238031545, + "tokens_seen": 1636171776 + }, + { + "epoch": 0.5, + "learning_rate": 0.0004075105315947844, + "loss": 1.4744, + "theoretical_loss": 3.517922405256364, + "tokens_seen": 1636237312 + }, + { + "epoch": 0.5, + "learning_rate": 0.00040749448345035104, + "loss": 1.4673, + "theoretical_loss": 3.5179102873308468, + "tokens_seen": 1636302848 + }, + { + "epoch": 0.5, + "learning_rate": 0.0004074784353059178, + "loss": 1.1657, + "theoretical_loss": 3.5178981700265473, + "tokens_seen": 1636368384 + }, + { + "epoch": 0.5, + "learning_rate": 0.00040746238716148444, + "loss": 1.2885, + "theoretical_loss": 3.5178860533434078, + "tokens_seen": 1636433920 + }, + { + "epoch": 0.5, + "learning_rate": 0.0004074463390170512, + "loss": 1.3538, + "theoretical_loss": 3.5178739372813723, + "tokens_seen": 1636499456 + }, + { + "epoch": 0.5, + "learning_rate": 0.00040743029087261784, + "loss": 1.3094, + "theoretical_loss": 3.517861821840383, + "tokens_seen": 1636564992 + }, + { + "epoch": 0.5, + "learning_rate": 0.0004074142427281846, + "loss": 1.3268, + "theoretical_loss": 3.5178497070203845, + "tokens_seen": 1636630528 + }, + { + "epoch": 0.5, + "learning_rate": 0.00040739819458375124, + "loss": 1.2972, + "theoretical_loss": 3.5178375928213192, + "tokens_seen": 1636696064 + }, + { + "epoch": 0.5, + "learning_rate": 0.000407382146439318, + "loss": 1.3732, + "theoretical_loss": 3.5178254792431307, + "tokens_seen": 1636761600 + }, + { + "epoch": 0.5, + "learning_rate": 0.00040736609829488464, + "loss": 1.4748, + "theoretical_loss": 3.517813366285762, + "tokens_seen": 1636827136 + }, + { + "epoch": 0.5, + "learning_rate": 0.0004073500501504514, + "loss": 1.2603, + "theoretical_loss": 3.517801253949157, + "tokens_seen": 1636892672 + }, + { + "epoch": 0.5, + "learning_rate": 0.00040733400200601805, + "loss": 1.4487, + "theoretical_loss": 3.5177891422332594, + "tokens_seen": 1636958208 + }, + { + "epoch": 0.5, + "learning_rate": 0.0004073179538615848, + "loss": 1.4376, + "theoretical_loss": 3.5177770311380114, + "tokens_seen": 1637023744 + }, + { + "epoch": 0.5, + "learning_rate": 0.00040730190571715145, + "loss": 1.3964, + "theoretical_loss": 3.5177649206633568, + "tokens_seen": 1637089280 + }, + { + "epoch": 0.5, + "learning_rate": 0.0004072858575727182, + "loss": 1.3186, + "theoretical_loss": 3.5177528108092395, + "tokens_seen": 1637154816 + }, + { + "epoch": 0.5, + "learning_rate": 0.00040726980942828485, + "loss": 1.2986, + "theoretical_loss": 3.517740701575602, + "tokens_seen": 1637220352 + }, + { + "epoch": 0.5, + "learning_rate": 0.00040725376128385155, + "loss": 1.3303, + "theoretical_loss": 3.5177285929623885, + "tokens_seen": 1637285888 + }, + { + "epoch": 0.5, + "learning_rate": 0.00040723771313941825, + "loss": 1.2718, + "theoretical_loss": 3.5177164849695415, + "tokens_seen": 1637351424 + }, + { + "epoch": 0.5, + "learning_rate": 0.00040722166499498495, + "loss": 1.3411, + "theoretical_loss": 3.5177043775970054, + "tokens_seen": 1637416960 + }, + { + "epoch": 0.5, + "learning_rate": 0.0004072056168505517, + "loss": 1.3447, + "theoretical_loss": 3.517692270844722, + "tokens_seen": 1637482496 + }, + { + "epoch": 0.5, + "learning_rate": 0.00040718956870611835, + "loss": 1.2912, + "theoretical_loss": 3.517680164712637, + "tokens_seen": 1637548032 + }, + { + "epoch": 0.5, + "learning_rate": 0.0004071735205616851, + "loss": 1.333, + "theoretical_loss": 3.517668059200692, + "tokens_seen": 1637613568 + }, + { + "epoch": 0.5, + "learning_rate": 0.00040715747241725175, + "loss": 1.3442, + "theoretical_loss": 3.5176559543088306, + "tokens_seen": 1637679104 + }, + { + "epoch": 0.5, + "learning_rate": 0.0004071414242728185, + "loss": 1.2109, + "theoretical_loss": 3.517643850036997, + "tokens_seen": 1637744640 + }, + { + "epoch": 0.5, + "learning_rate": 0.00040712537612838515, + "loss": 1.3554, + "theoretical_loss": 3.5176317463851343, + "tokens_seen": 1637810176 + }, + { + "epoch": 0.5, + "learning_rate": 0.0004071093279839519, + "loss": 1.5034, + "theoretical_loss": 3.5176196433531857, + "tokens_seen": 1637875712 + }, + { + "epoch": 0.5, + "learning_rate": 0.00040709327983951856, + "loss": 1.4368, + "theoretical_loss": 3.517607540941095, + "tokens_seen": 1637941248 + }, + { + "epoch": 0.5, + "learning_rate": 0.0004070772316950853, + "loss": 1.3307, + "theoretical_loss": 3.5175954391488053, + "tokens_seen": 1638006784 + }, + { + "epoch": 0.5, + "learning_rate": 0.00040706118355065196, + "loss": 1.3354, + "theoretical_loss": 3.5175833379762604, + "tokens_seen": 1638072320 + }, + { + "epoch": 0.5, + "learning_rate": 0.0004070451354062187, + "loss": 1.4778, + "theoretical_loss": 3.5175712374234034, + "tokens_seen": 1638137856 + }, + { + "epoch": 0.5, + "learning_rate": 0.00040702908726178536, + "loss": 1.4495, + "theoretical_loss": 3.5175591374901782, + "tokens_seen": 1638203392 + }, + { + "epoch": 0.5, + "learning_rate": 0.0004070130391173521, + "loss": 1.3751, + "theoretical_loss": 3.517547038176528, + "tokens_seen": 1638268928 + }, + { + "epoch": 0.5, + "learning_rate": 0.00040699699097291876, + "loss": 1.4071, + "theoretical_loss": 3.5175349394823963, + "tokens_seen": 1638334464 + }, + { + "debugging/Compilability": 1.0, + "debugging/distinct-1-grams": 0.7474626180525646, + "debugging/entropy-1-grams": 5.599013278347353, + "debugging/length": 499.42857142857144, + "debugging/num_segments": 14, + "debugging/score": 0.0110880360129877, + "debugging/score_std": 0.01573414919444883, + "epoch": 0.5, + "objective/train/docs_used": 598618, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.3616701364517212, + "objective/train/theoretical_loss": 3.5175228414077266, + "objective/train/tokens_used": 1658860000, + "theoretical_loss": 3.5175228414077266, + "tokens_seen": 1638400000 + }, + { + "epoch": 0.5, + "learning_rate": 0.0004069809428284855, + "loss": 1.2968, + "theoretical_loss": 3.5175228414077266, + "tokens_seen": 1638400000 + }, + { + "epoch": 0.5, + "learning_rate": 0.00040696489468405216, + "loss": 1.4071, + "theoretical_loss": 3.517510743952463, + "tokens_seen": 1638465536 + }, + { + "epoch": 0.5, + "learning_rate": 0.0004069488465396189, + "loss": 1.4272, + "theoretical_loss": 3.5174986471165477, + "tokens_seen": 1638531072 + }, + { + "epoch": 0.5, + "learning_rate": 0.00040693279839518556, + "loss": 1.4646, + "theoretical_loss": 3.5174865508999256, + "tokens_seen": 1638596608 + }, + { + "epoch": 0.5, + "learning_rate": 0.0004069167502507523, + "loss": 1.4774, + "theoretical_loss": 3.5174744553025388, + "tokens_seen": 1638662144 + }, + { + "epoch": 0.5, + "learning_rate": 0.00040690070210631896, + "loss": 1.4333, + "theoretical_loss": 3.5174623603243322, + "tokens_seen": 1638727680 + }, + { + "epoch": 0.5, + "learning_rate": 0.0004068846539618857, + "loss": 1.3905, + "theoretical_loss": 3.517450265965249, + "tokens_seen": 1638793216 + }, + { + "epoch": 0.5, + "learning_rate": 0.00040686860581745236, + "loss": 1.1611, + "theoretical_loss": 3.5174381722252326, + "tokens_seen": 1638858752 + }, + { + "epoch": 0.5, + "learning_rate": 0.0004068525576730191, + "loss": 1.3715, + "theoretical_loss": 3.517426079104226, + "tokens_seen": 1638924288 + }, + { + "epoch": 0.5, + "learning_rate": 0.00040683650952858577, + "loss": 1.3706, + "theoretical_loss": 3.5174139866021736, + "tokens_seen": 1638989824 + }, + { + "epoch": 0.5, + "learning_rate": 0.0004068204613841525, + "loss": 1.3862, + "theoretical_loss": 3.5174018947190193, + "tokens_seen": 1639055360 + }, + { + "epoch": 0.5, + "learning_rate": 0.00040680441323971917, + "loss": 1.3904, + "theoretical_loss": 3.517389803454705, + "tokens_seen": 1639120896 + }, + { + "epoch": 0.5, + "learning_rate": 0.0004067883650952859, + "loss": 1.4178, + "theoretical_loss": 3.517377712809176, + "tokens_seen": 1639186432 + }, + { + "epoch": 0.5, + "learning_rate": 0.00040677231695085257, + "loss": 1.3383, + "theoretical_loss": 3.517365622782375, + "tokens_seen": 1639251968 + }, + { + "epoch": 0.5, + "learning_rate": 0.0004067562688064193, + "loss": 1.2378, + "theoretical_loss": 3.5173535333742456, + "tokens_seen": 1639317504 + }, + { + "epoch": 0.5, + "learning_rate": 0.00040674022066198597, + "loss": 1.4389, + "theoretical_loss": 3.5173414445847317, + "tokens_seen": 1639383040 + }, + { + "epoch": 0.5, + "learning_rate": 0.0004067241725175527, + "loss": 1.4005, + "theoretical_loss": 3.517329356413777, + "tokens_seen": 1639448576 + }, + { + "epoch": 0.5, + "learning_rate": 0.00040670812437311937, + "loss": 1.2903, + "theoretical_loss": 3.5173172688613255, + "tokens_seen": 1639514112 + }, + { + "epoch": 0.5, + "learning_rate": 0.0004066920762286861, + "loss": 1.3442, + "theoretical_loss": 3.5173051819273193, + "tokens_seen": 1639579648 + }, + { + "epoch": 0.5, + "learning_rate": 0.00040667602808425277, + "loss": 1.3047, + "theoretical_loss": 3.517293095611704, + "tokens_seen": 1639645184 + }, + { + "epoch": 0.5, + "learning_rate": 0.0004066599799398195, + "loss": 1.4456, + "theoretical_loss": 3.5172810099144214, + "tokens_seen": 1639710720 + }, + { + "epoch": 0.5, + "learning_rate": 0.0004066439317953862, + "loss": 1.3235, + "theoretical_loss": 3.5172689248354168, + "tokens_seen": 1639776256 + }, + { + "epoch": 0.5, + "learning_rate": 0.00040662788365095293, + "loss": 1.3992, + "theoretical_loss": 3.5172568403746327, + "tokens_seen": 1639841792 + }, + { + "epoch": 0.5, + "learning_rate": 0.0004066118355065196, + "loss": 1.371, + "theoretical_loss": 3.517244756532013, + "tokens_seen": 1639907328 + }, + { + "epoch": 0.5, + "learning_rate": 0.0004065957873620863, + "loss": 1.5356, + "theoretical_loss": 3.517232673307502, + "tokens_seen": 1639972864 + }, + { + "epoch": 0.5, + "learning_rate": 0.000406579739217653, + "loss": 1.3137, + "theoretical_loss": 3.5172205907010428, + "tokens_seen": 1640038400 + }, + { + "epoch": 0.5, + "learning_rate": 0.0004065636910732197, + "loss": 1.4395, + "theoretical_loss": 3.517208508712579, + "tokens_seen": 1640103936 + }, + { + "epoch": 0.5, + "learning_rate": 0.0004065476429287864, + "loss": 1.3393, + "theoretical_loss": 3.5171964273420544, + "tokens_seen": 1640169472 + }, + { + "epoch": 0.5, + "learning_rate": 0.0004065315947843531, + "loss": 1.3958, + "theoretical_loss": 3.5171843465894135, + "tokens_seen": 1640235008 + }, + { + "epoch": 0.5, + "learning_rate": 0.0004065155466399198, + "loss": 1.3419, + "theoretical_loss": 3.5171722664545992, + "tokens_seen": 1640300544 + }, + { + "epoch": 0.5, + "learning_rate": 0.0004064994984954865, + "loss": 1.2402, + "theoretical_loss": 3.517160186937555, + "tokens_seen": 1640366080 + }, + { + "epoch": 0.5, + "learning_rate": 0.0004064834503510531, + "loss": 1.4021, + "theoretical_loss": 3.517148108038225, + "tokens_seen": 1640431616 + }, + { + "epoch": 0.5, + "learning_rate": 0.0004064674022066199, + "loss": 1.2954, + "theoretical_loss": 3.5171360297565535, + "tokens_seen": 1640497152 + }, + { + "epoch": 0.5, + "learning_rate": 0.0004064513540621865, + "loss": 1.3657, + "theoretical_loss": 3.517123952092483, + "tokens_seen": 1640562688 + }, + { + "epoch": 0.5, + "learning_rate": 0.0004064353059177533, + "loss": 1.1681, + "theoretical_loss": 3.5171118750459582, + "tokens_seen": 1640628224 + }, + { + "epoch": 0.5, + "learning_rate": 0.00040641925777331993, + "loss": 1.3061, + "theoretical_loss": 3.517099798616923, + "tokens_seen": 1640693760 + }, + { + "epoch": 0.5, + "learning_rate": 0.0004064032096288867, + "loss": 1.4476, + "theoretical_loss": 3.51708772280532, + "tokens_seen": 1640759296 + }, + { + "epoch": 0.5, + "learning_rate": 0.00040638716148445333, + "loss": 1.2807, + "theoretical_loss": 3.5170756476110943, + "tokens_seen": 1640824832 + }, + { + "epoch": 0.5, + "learning_rate": 0.0004063711133400201, + "loss": 1.3272, + "theoretical_loss": 3.517063573034189, + "tokens_seen": 1640890368 + }, + { + "epoch": 0.5, + "learning_rate": 0.00040635506519558673, + "loss": 1.4066, + "theoretical_loss": 3.517051499074548, + "tokens_seen": 1640955904 + }, + { + "epoch": 0.5, + "learning_rate": 0.0004063390170511535, + "loss": 1.4381, + "theoretical_loss": 3.517039425732115, + "tokens_seen": 1641021440 + }, + { + "epoch": 0.5, + "learning_rate": 0.00040632296890672013, + "loss": 1.3193, + "theoretical_loss": 3.5170273530068337, + "tokens_seen": 1641086976 + }, + { + "epoch": 0.5, + "learning_rate": 0.0004063069207622869, + "loss": 1.3729, + "theoretical_loss": 3.517015280898648, + "tokens_seen": 1641152512 + }, + { + "epoch": 0.5, + "learning_rate": 0.00040629087261785353, + "loss": 1.3502, + "theoretical_loss": 3.5170032094075023, + "tokens_seen": 1641218048 + }, + { + "epoch": 0.5, + "learning_rate": 0.0004062748244734203, + "loss": 1.3584, + "theoretical_loss": 3.51699113853334, + "tokens_seen": 1641283584 + }, + { + "epoch": 0.5, + "learning_rate": 0.00040625877632898693, + "loss": 1.3634, + "theoretical_loss": 3.516979068276105, + "tokens_seen": 1641349120 + }, + { + "epoch": 0.5, + "learning_rate": 0.0004062427281845537, + "loss": 1.2633, + "theoretical_loss": 3.5169669986357404, + "tokens_seen": 1641414656 + }, + { + "epoch": 0.5, + "learning_rate": 0.00040622668004012034, + "loss": 1.5521, + "theoretical_loss": 3.516954929612191, + "tokens_seen": 1641480192 + }, + { + "epoch": 0.5, + "learning_rate": 0.0004062106318956871, + "loss": 1.3729, + "theoretical_loss": 3.5169428612053997, + "tokens_seen": 1641545728 + }, + { + "epoch": 0.5, + "learning_rate": 0.00040619458375125374, + "loss": 1.4389, + "theoretical_loss": 3.5169307934153116, + "tokens_seen": 1641611264 + }, + { + "epoch": 0.5, + "objective/train/docs_used": 599774, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.3523476123809814, + "objective/train/theoretical_loss": 3.5169187262418697, + "objective/train/tokens_used": 1662136800, + "theoretical_loss": 3.5169187262418697, + "tokens_seen": 1641676800 + }, + { + "epoch": 0.5, + "learning_rate": 0.0004061785356068205, + "loss": 1.3417, + "theoretical_loss": 3.5169187262418697, + "tokens_seen": 1641676800 + }, + { + "epoch": 0.5, + "learning_rate": 0.00040616248746238714, + "loss": 1.396, + "theoretical_loss": 3.5169066596850183, + "tokens_seen": 1641742336 + }, + { + "epoch": 0.5, + "learning_rate": 0.0004061464393179539, + "loss": 1.4148, + "theoretical_loss": 3.516894593744701, + "tokens_seen": 1641807872 + }, + { + "epoch": 0.5, + "learning_rate": 0.00040613039117352054, + "loss": 1.4064, + "theoretical_loss": 3.516882528420862, + "tokens_seen": 1641873408 + }, + { + "epoch": 0.5, + "learning_rate": 0.0004061143430290873, + "loss": 1.4089, + "theoretical_loss": 3.5168704637134445, + "tokens_seen": 1641938944 + }, + { + "epoch": 0.5, + "learning_rate": 0.00040609829488465394, + "loss": 1.459, + "theoretical_loss": 3.5168583996223934, + "tokens_seen": 1642004480 + }, + { + "epoch": 0.5, + "learning_rate": 0.0004060822467402207, + "loss": 1.4126, + "theoretical_loss": 3.516846336147652, + "tokens_seen": 1642070016 + }, + { + "epoch": 0.5, + "learning_rate": 0.00040606619859578734, + "loss": 1.2095, + "theoretical_loss": 3.516834273289164, + "tokens_seen": 1642135552 + }, + { + "epoch": 0.5, + "learning_rate": 0.0004060501504513541, + "loss": 1.3537, + "theoretical_loss": 3.516822211046874, + "tokens_seen": 1642201088 + }, + { + "epoch": 0.5, + "learning_rate": 0.00040603410230692074, + "loss": 1.3651, + "theoretical_loss": 3.516810149420725, + "tokens_seen": 1642266624 + }, + { + "epoch": 0.5, + "learning_rate": 0.0004060180541624875, + "loss": 1.4339, + "theoretical_loss": 3.5167980884106624, + "tokens_seen": 1642332160 + }, + { + "epoch": 0.5, + "learning_rate": 0.00040600200601805425, + "loss": 1.327, + "theoretical_loss": 3.5167860280166288, + "tokens_seen": 1642397696 + }, + { + "epoch": 0.5, + "learning_rate": 0.0004059859578736209, + "loss": 1.4029, + "theoretical_loss": 3.516773968238569, + "tokens_seen": 1642463232 + }, + { + "epoch": 0.5, + "learning_rate": 0.0004059699097291876, + "loss": 1.3817, + "theoretical_loss": 3.5167619090764264, + "tokens_seen": 1642528768 + }, + { + "epoch": 0.5, + "learning_rate": 0.0004059538615847543, + "loss": 1.385, + "theoretical_loss": 3.516749850530145, + "tokens_seen": 1642594304 + }, + { + "epoch": 0.5, + "learning_rate": 0.000405937813440321, + "loss": 1.3634, + "theoretical_loss": 3.5167377925996686, + "tokens_seen": 1642659840 + }, + { + "epoch": 0.5, + "learning_rate": 0.0004059217652958877, + "loss": 1.4677, + "theoretical_loss": 3.516725735284942, + "tokens_seen": 1642725376 + }, + { + "epoch": 0.5, + "learning_rate": 0.0004059057171514544, + "loss": 1.3659, + "theoretical_loss": 3.5167136785859086, + "tokens_seen": 1642790912 + }, + { + "epoch": 0.5, + "learning_rate": 0.0004058896690070211, + "loss": 1.4394, + "theoretical_loss": 3.5167016225025125, + "tokens_seen": 1642856448 + }, + { + "epoch": 0.5, + "learning_rate": 0.0004058736208625878, + "loss": 1.3608, + "theoretical_loss": 3.516689567034698, + "tokens_seen": 1642921984 + }, + { + "epoch": 0.5, + "learning_rate": 0.0004058575727181545, + "loss": 1.4089, + "theoretical_loss": 3.516677512182408, + "tokens_seen": 1642987520 + }, + { + "epoch": 0.5, + "learning_rate": 0.0004058415245737212, + "loss": 1.4271, + "theoretical_loss": 3.5166654579455887, + "tokens_seen": 1643053056 + }, + { + "epoch": 0.5, + "learning_rate": 0.00040582547642928785, + "loss": 1.4669, + "theoretical_loss": 3.516653404324182, + "tokens_seen": 1643118592 + }, + { + "epoch": 0.5, + "learning_rate": 0.0004058094282848546, + "loss": 1.315, + "theoretical_loss": 3.5166413513181327, + "tokens_seen": 1643184128 + }, + { + "epoch": 0.5, + "learning_rate": 0.00040579338014042125, + "loss": 1.3453, + "theoretical_loss": 3.516629298927385, + "tokens_seen": 1643249664 + }, + { + "epoch": 0.5, + "learning_rate": 0.000405777331995988, + "loss": 1.4066, + "theoretical_loss": 3.516617247151883, + "tokens_seen": 1643315200 + }, + { + "epoch": 0.5, + "learning_rate": 0.00040576128385155465, + "loss": 1.4506, + "theoretical_loss": 3.5166051959915703, + "tokens_seen": 1643380736 + }, + { + "epoch": 0.5, + "learning_rate": 0.0004057452357071214, + "loss": 1.4346, + "theoretical_loss": 3.5165931454463912, + "tokens_seen": 1643446272 + }, + { + "epoch": 0.5, + "learning_rate": 0.00040572918756268806, + "loss": 1.2767, + "theoretical_loss": 3.5165810955162904, + "tokens_seen": 1643511808 + }, + { + "epoch": 0.5, + "learning_rate": 0.0004057131394182548, + "loss": 1.3743, + "theoretical_loss": 3.516569046201211, + "tokens_seen": 1643577344 + }, + { + "epoch": 0.5, + "learning_rate": 0.00040569709127382146, + "loss": 1.418, + "theoretical_loss": 3.5165569975010973, + "tokens_seen": 1643642880 + }, + { + "epoch": 0.5, + "learning_rate": 0.0004056810431293882, + "loss": 1.3792, + "theoretical_loss": 3.516544949415894, + "tokens_seen": 1643708416 + }, + { + "epoch": 0.5, + "learning_rate": 0.00040566499498495486, + "loss": 1.2836, + "theoretical_loss": 3.5165329019455447, + "tokens_seen": 1643773952 + }, + { + "epoch": 0.5, + "learning_rate": 0.0004056489468405216, + "loss": 1.3773, + "theoretical_loss": 3.5165208550899933, + "tokens_seen": 1643839488 + }, + { + "epoch": 0.5, + "learning_rate": 0.00040563289869608826, + "loss": 1.3061, + "theoretical_loss": 3.516508808849184, + "tokens_seen": 1643905024 + }, + { + "epoch": 0.5, + "learning_rate": 0.000405616850551655, + "loss": 1.3837, + "theoretical_loss": 3.5164967632230617, + "tokens_seen": 1643970560 + }, + { + "epoch": 0.5, + "learning_rate": 0.00040560080240722166, + "loss": 1.3155, + "theoretical_loss": 3.51648471821157, + "tokens_seen": 1644036096 + }, + { + "epoch": 0.5, + "learning_rate": 0.0004055847542627884, + "loss": 1.3137, + "theoretical_loss": 3.5164726738146532, + "tokens_seen": 1644101632 + }, + { + "epoch": 0.5, + "learning_rate": 0.00040556870611835506, + "loss": 1.2552, + "theoretical_loss": 3.5164606300322547, + "tokens_seen": 1644167168 + }, + { + "epoch": 0.5, + "learning_rate": 0.0004055526579739218, + "loss": 1.3535, + "theoretical_loss": 3.5164485868643194, + "tokens_seen": 1644232704 + }, + { + "epoch": 0.5, + "learning_rate": 0.00040553660982948846, + "loss": 1.3663, + "theoretical_loss": 3.5164365443107917, + "tokens_seen": 1644298240 + }, + { + "epoch": 0.5, + "learning_rate": 0.0004055205616850552, + "loss": 1.4031, + "theoretical_loss": 3.5164245023716147, + "tokens_seen": 1644363776 + }, + { + "epoch": 0.5, + "learning_rate": 0.00040550451354062186, + "loss": 1.3725, + "theoretical_loss": 3.5164124610467336, + "tokens_seen": 1644429312 + }, + { + "epoch": 0.5, + "learning_rate": 0.0004054884653961886, + "loss": 1.4007, + "theoretical_loss": 3.516400420336092, + "tokens_seen": 1644494848 + }, + { + "epoch": 0.5, + "learning_rate": 0.00040547241725175527, + "loss": 1.3941, + "theoretical_loss": 3.5163883802396345, + "tokens_seen": 1644560384 + }, + { + "epoch": 0.5, + "learning_rate": 0.000405456369107322, + "loss": 1.4312, + "theoretical_loss": 3.516376340757305, + "tokens_seen": 1644625920 + }, + { + "epoch": 0.5, + "learning_rate": 0.00040544032096288867, + "loss": 1.3233, + "theoretical_loss": 3.516364301889048, + "tokens_seen": 1644691456 + }, + { + "epoch": 0.5, + "learning_rate": 0.0004054242728184554, + "loss": 1.5058, + "theoretical_loss": 3.516352263634807, + "tokens_seen": 1644756992 + }, + { + "epoch": 0.5, + "learning_rate": 0.00040540822467402207, + "loss": 1.4193, + "theoretical_loss": 3.516340225994527, + "tokens_seen": 1644822528 + }, + { + "epoch": 0.5, + "learning_rate": 0.0004053921765295888, + "loss": 1.4202, + "theoretical_loss": 3.516328188968152, + "tokens_seen": 1644888064 + }, + { + "epoch": 0.5, + "objective/train/docs_used": 601010, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.3907251358032227, + "objective/train/theoretical_loss": 3.5163161525556266, + "objective/train/tokens_used": 1665413600, + "theoretical_loss": 3.5163161525556266, + "tokens_seen": 1644953600 + }, + { + "epoch": 0.5, + "learning_rate": 0.00040537612838515547, + "loss": 1.3325, + "theoretical_loss": 3.5163161525556266, + "tokens_seen": 1644953600 + }, + { + "epoch": 0.5, + "learning_rate": 0.0004053600802407222, + "loss": 1.3617, + "theoretical_loss": 3.516304116756894, + "tokens_seen": 1645019136 + }, + { + "epoch": 0.5, + "learning_rate": 0.00040534403209628887, + "loss": 1.316, + "theoretical_loss": 3.5162920815718994, + "tokens_seen": 1645084672 + }, + { + "epoch": 0.5, + "learning_rate": 0.0004053279839518556, + "loss": 1.4367, + "theoretical_loss": 3.5162800470005866, + "tokens_seen": 1645150208 + }, + { + "epoch": 0.5, + "learning_rate": 0.00040531193580742227, + "loss": 1.5021, + "theoretical_loss": 3.5162680130429003, + "tokens_seen": 1645215744 + }, + { + "epoch": 0.5, + "learning_rate": 0.000405295887662989, + "loss": 1.4484, + "theoretical_loss": 3.516255979698784, + "tokens_seen": 1645281280 + }, + { + "epoch": 0.5, + "learning_rate": 0.0004052798395185557, + "loss": 1.3583, + "theoretical_loss": 3.5162439469681828, + "tokens_seen": 1645346816 + }, + { + "epoch": 0.5, + "learning_rate": 0.00040526379137412243, + "loss": 1.4576, + "theoretical_loss": 3.5162319148510406, + "tokens_seen": 1645412352 + }, + { + "epoch": 0.5, + "learning_rate": 0.0004052477432296891, + "loss": 1.4079, + "theoretical_loss": 3.5162198833473015, + "tokens_seen": 1645477888 + }, + { + "epoch": 0.5, + "learning_rate": 0.00040523169508525583, + "loss": 1.4447, + "theoretical_loss": 3.51620785245691, + "tokens_seen": 1645543424 + }, + { + "epoch": 0.5, + "learning_rate": 0.0004052156469408225, + "loss": 1.4404, + "theoretical_loss": 3.5161958221798106, + "tokens_seen": 1645608960 + }, + { + "epoch": 0.5, + "learning_rate": 0.0004051995987963892, + "loss": 1.303, + "theoretical_loss": 3.516183792515948, + "tokens_seen": 1645674496 + }, + { + "epoch": 0.5, + "learning_rate": 0.0004051835506519559, + "loss": 1.338, + "theoretical_loss": 3.516171763465265, + "tokens_seen": 1645740032 + }, + { + "epoch": 0.5, + "learning_rate": 0.0004051675025075226, + "loss": 1.4123, + "theoretical_loss": 3.516159735027707, + "tokens_seen": 1645805568 + }, + { + "epoch": 0.5, + "learning_rate": 0.0004051514543630893, + "loss": 1.3932, + "theoretical_loss": 3.516147707203219, + "tokens_seen": 1645871104 + }, + { + "epoch": 0.5, + "learning_rate": 0.000405135406218656, + "loss": 1.3673, + "theoretical_loss": 3.516135679991744, + "tokens_seen": 1645936640 + }, + { + "epoch": 0.5, + "learning_rate": 0.0004051193580742227, + "loss": 1.2996, + "theoretical_loss": 3.516123653393227, + "tokens_seen": 1646002176 + }, + { + "epoch": 0.5, + "learning_rate": 0.0004051033099297894, + "loss": 1.4353, + "theoretical_loss": 3.5161116274076125, + "tokens_seen": 1646067712 + }, + { + "epoch": 0.5, + "learning_rate": 0.0004050872617853561, + "loss": 1.4842, + "theoretical_loss": 3.5160996020348447, + "tokens_seen": 1646133248 + }, + { + "epoch": 0.5, + "learning_rate": 0.0004050712136409228, + "loss": 1.3783, + "theoretical_loss": 3.5160875772748676, + "tokens_seen": 1646198784 + }, + { + "epoch": 0.5, + "learning_rate": 0.00040505516549648943, + "loss": 1.4475, + "theoretical_loss": 3.516075553127626, + "tokens_seen": 1646264320 + }, + { + "epoch": 0.5, + "learning_rate": 0.0004050391173520562, + "loss": 1.2784, + "theoretical_loss": 3.5160635295930645, + "tokens_seen": 1646329856 + }, + { + "epoch": 0.5, + "learning_rate": 0.00040502306920762283, + "loss": 1.3233, + "theoretical_loss": 3.5160515066711264, + "tokens_seen": 1646395392 + }, + { + "epoch": 0.5, + "learning_rate": 0.0004050070210631896, + "loss": 1.4143, + "theoretical_loss": 3.516039484361758, + "tokens_seen": 1646460928 + }, + { + "epoch": 0.5, + "learning_rate": 0.00040499097291875623, + "loss": 1.4094, + "theoretical_loss": 3.5160274626649017, + "tokens_seen": 1646526464 + }, + { + "epoch": 0.5, + "learning_rate": 0.000404974924774323, + "loss": 1.3733, + "theoretical_loss": 3.516015441580503, + "tokens_seen": 1646592000 + }, + { + "epoch": 0.5, + "learning_rate": 0.00040495887662988963, + "loss": 1.3235, + "theoretical_loss": 3.5160034211085067, + "tokens_seen": 1646657536 + }, + { + "epoch": 0.5, + "learning_rate": 0.0004049428284854564, + "loss": 1.5011, + "theoretical_loss": 3.515991401248856, + "tokens_seen": 1646723072 + }, + { + "epoch": 0.5, + "learning_rate": 0.00040492678034102303, + "loss": 1.2682, + "theoretical_loss": 3.515979382001496, + "tokens_seen": 1646788608 + }, + { + "epoch": 0.5, + "learning_rate": 0.0004049107321965898, + "loss": 1.3793, + "theoretical_loss": 3.5159673633663715, + "tokens_seen": 1646854144 + }, + { + "epoch": 0.5, + "learning_rate": 0.00040489468405215643, + "loss": 1.2973, + "theoretical_loss": 3.5159553453434267, + "tokens_seen": 1646919680 + }, + { + "epoch": 0.5, + "learning_rate": 0.0004048786359077232, + "loss": 1.3553, + "theoretical_loss": 3.5159433279326056, + "tokens_seen": 1646985216 + }, + { + "epoch": 0.5, + "learning_rate": 0.00040486258776328984, + "loss": 1.3316, + "theoretical_loss": 3.515931311133853, + "tokens_seen": 1647050752 + }, + { + "epoch": 0.5, + "learning_rate": 0.0004048465396188566, + "loss": 1.3626, + "theoretical_loss": 3.515919294947114, + "tokens_seen": 1647116288 + }, + { + "epoch": 0.5, + "learning_rate": 0.00040483049147442324, + "loss": 1.2657, + "theoretical_loss": 3.515907279372332, + "tokens_seen": 1647181824 + }, + { + "epoch": 0.5, + "learning_rate": 0.00040481444332999, + "loss": 1.3584, + "theoretical_loss": 3.5158952644094517, + "tokens_seen": 1647247360 + }, + { + "epoch": 0.5, + "learning_rate": 0.00040479839518555675, + "loss": 1.3484, + "theoretical_loss": 3.515883250058418, + "tokens_seen": 1647312896 + }, + { + "epoch": 0.5, + "learning_rate": 0.0004047823470411234, + "loss": 1.3297, + "theoretical_loss": 3.5158712363191755, + "tokens_seen": 1647378432 + }, + { + "epoch": 0.5, + "learning_rate": 0.00040476629889669015, + "loss": 1.387, + "theoretical_loss": 3.5158592231916685, + "tokens_seen": 1647443968 + }, + { + "epoch": 0.5, + "learning_rate": 0.0004047502507522568, + "loss": 1.4229, + "theoretical_loss": 3.5158472106758416, + "tokens_seen": 1647509504 + }, + { + "epoch": 0.5, + "learning_rate": 0.00040473420260782355, + "loss": 1.3827, + "theoretical_loss": 3.515835198771639, + "tokens_seen": 1647575040 + }, + { + "epoch": 0.5, + "learning_rate": 0.0004047181544633902, + "loss": 1.4182, + "theoretical_loss": 3.5158231874790054, + "tokens_seen": 1647640576 + }, + { + "epoch": 0.5, + "learning_rate": 0.00040470210631895695, + "loss": 1.2683, + "theoretical_loss": 3.5158111767978855, + "tokens_seen": 1647706112 + }, + { + "epoch": 0.5, + "learning_rate": 0.0004046860581745236, + "loss": 1.4465, + "theoretical_loss": 3.5157991667282236, + "tokens_seen": 1647771648 + }, + { + "epoch": 0.5, + "learning_rate": 0.00040467001003009035, + "loss": 1.2568, + "theoretical_loss": 3.5157871572699646, + "tokens_seen": 1647837184 + }, + { + "epoch": 0.5, + "learning_rate": 0.000404653961885657, + "loss": 1.3368, + "theoretical_loss": 3.515775148423053, + "tokens_seen": 1647902720 + }, + { + "epoch": 0.5, + "learning_rate": 0.00040463791374122375, + "loss": 1.374, + "theoretical_loss": 3.515763140187433, + "tokens_seen": 1647968256 + }, + { + "epoch": 0.5, + "learning_rate": 0.0004046218655967904, + "loss": 1.3562, + "theoretical_loss": 3.5157511325630493, + "tokens_seen": 1648033792 + }, + { + "epoch": 0.5, + "learning_rate": 0.00040460581745235715, + "loss": 1.3898, + "theoretical_loss": 3.5157391255498465, + "tokens_seen": 1648099328 + }, + { + "epoch": 0.5, + "learning_rate": 0.0004045897693079238, + "loss": 1.3594, + "theoretical_loss": 3.5157271191477695, + "tokens_seen": 1648164864 + }, + { + "epoch": 0.5, + "objective/train/docs_used": 602144, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.4319968223571777, + "objective/train/theoretical_loss": 3.5157151133567623, + "objective/train/tokens_used": 1668690400, + "theoretical_loss": 3.5157151133567623, + "tokens_seen": 1648230400 + }, + { + "epoch": 0.5, + "learning_rate": 0.0004045737211634905, + "loss": 1.3946, + "theoretical_loss": 3.5157151133567623, + "tokens_seen": 1648230400 + }, + { + "epoch": 0.5, + "learning_rate": 0.0004045576730190572, + "loss": 1.4956, + "theoretical_loss": 3.5157031081767705, + "tokens_seen": 1648295936 + }, + { + "epoch": 0.5, + "learning_rate": 0.0004045416248746239, + "loss": 1.3327, + "theoretical_loss": 3.515691103607738, + "tokens_seen": 1648361472 + }, + { + "epoch": 0.5, + "learning_rate": 0.0004045255767301906, + "loss": 1.4095, + "theoretical_loss": 3.515679099649609, + "tokens_seen": 1648427008 + }, + { + "epoch": 0.5, + "learning_rate": 0.0004045095285857573, + "loss": 1.4023, + "theoretical_loss": 3.515667096302329, + "tokens_seen": 1648492544 + }, + { + "epoch": 0.5, + "learning_rate": 0.000404493480441324, + "loss": 1.3333, + "theoretical_loss": 3.515655093565842, + "tokens_seen": 1648558080 + }, + { + "epoch": 0.5, + "learning_rate": 0.0004044774322968907, + "loss": 1.4407, + "theoretical_loss": 3.5156430914400927, + "tokens_seen": 1648623616 + }, + { + "epoch": 0.5, + "learning_rate": 0.0004044613841524574, + "loss": 1.4593, + "theoretical_loss": 3.5156310899250265, + "tokens_seen": 1648689152 + }, + { + "epoch": 0.5, + "learning_rate": 0.0004044453360080241, + "loss": 1.3826, + "theoretical_loss": 3.5156190890205874, + "tokens_seen": 1648754688 + }, + { + "epoch": 0.5, + "learning_rate": 0.00040442928786359075, + "loss": 1.2911, + "theoretical_loss": 3.5156070887267195, + "tokens_seen": 1648820224 + }, + { + "epoch": 0.5, + "learning_rate": 0.0004044132397191575, + "loss": 1.3511, + "theoretical_loss": 3.515595089043369, + "tokens_seen": 1648885760 + }, + { + "epoch": 0.5, + "learning_rate": 0.00040439719157472415, + "loss": 1.2578, + "theoretical_loss": 3.5155830899704794, + "tokens_seen": 1648951296 + }, + { + "epoch": 0.5, + "learning_rate": 0.0004043811434302909, + "loss": 1.372, + "theoretical_loss": 3.5155710915079954, + "tokens_seen": 1649016832 + }, + { + "epoch": 0.5, + "learning_rate": 0.00040436509528585755, + "loss": 1.2742, + "theoretical_loss": 3.5155590936558623, + "tokens_seen": 1649082368 + }, + { + "epoch": 0.5, + "learning_rate": 0.0004043490471414243, + "loss": 1.4508, + "theoretical_loss": 3.5155470964140245, + "tokens_seen": 1649147904 + }, + { + "epoch": 0.5, + "learning_rate": 0.00040433299899699096, + "loss": 1.3227, + "theoretical_loss": 3.5155350997824266, + "tokens_seen": 1649213440 + }, + { + "epoch": 0.5, + "learning_rate": 0.0004043169508525577, + "loss": 1.3268, + "theoretical_loss": 3.5155231037610135, + "tokens_seen": 1649278976 + }, + { + "epoch": 0.5, + "learning_rate": 0.00040430090270812436, + "loss": 1.4108, + "theoretical_loss": 3.5155111083497297, + "tokens_seen": 1649344512 + }, + { + "epoch": 0.5, + "learning_rate": 0.0004042848545636911, + "loss": 1.2858, + "theoretical_loss": 3.51549911354852, + "tokens_seen": 1649410048 + }, + { + "epoch": 0.5, + "learning_rate": 0.00040426880641925776, + "loss": 1.4626, + "theoretical_loss": 3.5154871193573296, + "tokens_seen": 1649475584 + }, + { + "epoch": 0.5, + "learning_rate": 0.0004042527582748245, + "loss": 1.2899, + "theoretical_loss": 3.5154751257761023, + "tokens_seen": 1649541120 + }, + { + "epoch": 0.5, + "learning_rate": 0.00040423671013039116, + "loss": 1.4315, + "theoretical_loss": 3.5154631328047836, + "tokens_seen": 1649606656 + }, + { + "epoch": 0.5, + "learning_rate": 0.0004042206619859579, + "loss": 1.4191, + "theoretical_loss": 3.515451140443318, + "tokens_seen": 1649672192 + }, + { + "epoch": 0.5, + "learning_rate": 0.00040420461384152456, + "loss": 1.4338, + "theoretical_loss": 3.51543914869165, + "tokens_seen": 1649737728 + }, + { + "epoch": 0.5, + "learning_rate": 0.0004041885656970913, + "loss": 1.3737, + "theoretical_loss": 3.515427157549725, + "tokens_seen": 1649803264 + }, + { + "epoch": 0.5, + "learning_rate": 0.00040417251755265796, + "loss": 1.3187, + "theoretical_loss": 3.5154151670174874, + "tokens_seen": 1649868800 + }, + { + "epoch": 0.5, + "learning_rate": 0.0004041564694082247, + "loss": 1.3705, + "theoretical_loss": 3.515403177094882, + "tokens_seen": 1649934336 + }, + { + "epoch": 0.5, + "learning_rate": 0.00040414042126379136, + "loss": 1.3402, + "theoretical_loss": 3.5153911877818533, + "tokens_seen": 1649999872 + }, + { + "epoch": 0.5, + "learning_rate": 0.0004041243731193581, + "loss": 1.4029, + "theoretical_loss": 3.515379199078347, + "tokens_seen": 1650065408 + }, + { + "epoch": 0.5, + "learning_rate": 0.00040410832497492477, + "loss": 1.3728, + "theoretical_loss": 3.515367210984307, + "tokens_seen": 1650130944 + }, + { + "epoch": 0.5, + "learning_rate": 0.0004040922768304915, + "loss": 1.2393, + "theoretical_loss": 3.515355223499678, + "tokens_seen": 1650196480 + }, + { + "epoch": 0.5, + "learning_rate": 0.00040407622868605817, + "loss": 1.3306, + "theoretical_loss": 3.515343236624406, + "tokens_seen": 1650262016 + }, + { + "epoch": 0.5, + "learning_rate": 0.0004040601805416249, + "loss": 1.45, + "theoretical_loss": 3.5153312503584346, + "tokens_seen": 1650327552 + }, + { + "epoch": 0.5, + "learning_rate": 0.00040404413239719157, + "loss": 1.4805, + "theoretical_loss": 3.5153192647017093, + "tokens_seen": 1650393088 + }, + { + "epoch": 0.5, + "learning_rate": 0.0004040280842527583, + "loss": 1.3955, + "theoretical_loss": 3.5153072796541744, + "tokens_seen": 1650458624 + }, + { + "epoch": 0.5, + "learning_rate": 0.00040401203610832497, + "loss": 1.342, + "theoretical_loss": 3.5152952952157754, + "tokens_seen": 1650524160 + }, + { + "epoch": 0.5, + "learning_rate": 0.0004039959879638917, + "loss": 1.3984, + "theoretical_loss": 3.515283311386457, + "tokens_seen": 1650589696 + }, + { + "epoch": 0.5, + "learning_rate": 0.00040397993981945837, + "loss": 1.4205, + "theoretical_loss": 3.515271328166164, + "tokens_seen": 1650655232 + }, + { + "epoch": 0.5, + "learning_rate": 0.0004039638916750251, + "loss": 1.3664, + "theoretical_loss": 3.515259345554841, + "tokens_seen": 1650720768 + }, + { + "epoch": 0.5, + "learning_rate": 0.00040394784353059177, + "loss": 1.3681, + "theoretical_loss": 3.515247363552433, + "tokens_seen": 1650786304 + }, + { + "epoch": 0.5, + "learning_rate": 0.0004039317953861585, + "loss": 1.3963, + "theoretical_loss": 3.515235382158885, + "tokens_seen": 1650851840 + }, + { + "epoch": 0.5, + "learning_rate": 0.00040391574724172517, + "loss": 1.3441, + "theoretical_loss": 3.5152234013741417, + "tokens_seen": 1650917376 + }, + { + "epoch": 0.5, + "learning_rate": 0.00040389969909729193, + "loss": 1.3076, + "theoretical_loss": 3.5152114211981487, + "tokens_seen": 1650982912 + }, + { + "epoch": 0.5, + "learning_rate": 0.0004038836509528586, + "loss": 1.4369, + "theoretical_loss": 3.51519944163085, + "tokens_seen": 1651048448 + }, + { + "epoch": 0.5, + "learning_rate": 0.00040386760280842533, + "loss": 1.279, + "theoretical_loss": 3.5151874626721904, + "tokens_seen": 1651113984 + }, + { + "epoch": 0.5, + "learning_rate": 0.000403851554663992, + "loss": 1.3935, + "theoretical_loss": 3.515175484322116, + "tokens_seen": 1651179520 + }, + { + "epoch": 0.5, + "learning_rate": 0.00040383550651955873, + "loss": 1.4106, + "theoretical_loss": 3.5151635065805706, + "tokens_seen": 1651245056 + }, + { + "epoch": 0.5, + "learning_rate": 0.0004038194583751254, + "loss": 1.2792, + "theoretical_loss": 3.5151515294474995, + "tokens_seen": 1651310592 + }, + { + "epoch": 0.5, + "learning_rate": 0.0004038034102306921, + "loss": 1.2486, + "theoretical_loss": 3.515139552922848, + "tokens_seen": 1651376128 + }, + { + "epoch": 0.5, + "learning_rate": 0.0004037873620862588, + "loss": 1.3569, + "theoretical_loss": 3.5151275770065604, + "tokens_seen": 1651441664 + }, + { + "epoch": 0.5, + "objective/train/docs_used": 603335, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.2888708114624023, + "objective/train/theoretical_loss": 3.5151156016985823, + "objective/train/tokens_used": 1671967200, + "theoretical_loss": 3.5151156016985823, + "tokens_seen": 1651507200 + }, + { + "epoch": 0.5, + "learning_rate": 0.0004037713139418255, + "loss": 1.3245, + "theoretical_loss": 3.5151156016985823, + "tokens_seen": 1651507200 + }, + { + "epoch": 0.5, + "learning_rate": 0.0004037552657973922, + "loss": 1.4221, + "theoretical_loss": 3.5151036269988585, + "tokens_seen": 1651572736 + }, + { + "epoch": 0.5, + "learning_rate": 0.0004037392176529589, + "loss": 1.4993, + "theoretical_loss": 3.5150916529073335, + "tokens_seen": 1651638272 + }, + { + "epoch": 0.5, + "learning_rate": 0.0004037231695085256, + "loss": 1.3515, + "theoretical_loss": 3.5150796794239527, + "tokens_seen": 1651703808 + }, + { + "epoch": 0.5, + "learning_rate": 0.0004037071213640923, + "loss": 1.3487, + "theoretical_loss": 3.5150677065486615, + "tokens_seen": 1651769344 + }, + { + "epoch": 0.5, + "learning_rate": 0.000403691073219659, + "loss": 1.4758, + "theoretical_loss": 3.515055734281404, + "tokens_seen": 1651834880 + }, + { + "epoch": 0.5, + "learning_rate": 0.0004036750250752257, + "loss": 1.353, + "theoretical_loss": 3.515043762622125, + "tokens_seen": 1651900416 + }, + { + "epoch": 0.5, + "learning_rate": 0.00040365897693079233, + "loss": 1.3077, + "theoretical_loss": 3.5150317915707707, + "tokens_seen": 1651965952 + }, + { + "epoch": 0.5, + "learning_rate": 0.0004036429287863591, + "loss": 1.3001, + "theoretical_loss": 3.5150198211272854, + "tokens_seen": 1652031488 + }, + { + "epoch": 0.5, + "learning_rate": 0.00040362688064192573, + "loss": 1.4162, + "theoretical_loss": 3.5150078512916143, + "tokens_seen": 1652097024 + }, + { + "epoch": 0.5, + "learning_rate": 0.0004036108324974925, + "loss": 1.3532, + "theoretical_loss": 3.514995882063702, + "tokens_seen": 1652162560 + }, + { + "epoch": 0.5, + "learning_rate": 0.00040359478435305924, + "loss": 1.3522, + "theoretical_loss": 3.5149839134434946, + "tokens_seen": 1652228096 + }, + { + "epoch": 0.5, + "learning_rate": 0.0004035787362086259, + "loss": 1.2825, + "theoretical_loss": 3.5149719454309363, + "tokens_seen": 1652293632 + }, + { + "epoch": 0.5, + "learning_rate": 0.00040356268806419264, + "loss": 1.4085, + "theoretical_loss": 3.5149599780259715, + "tokens_seen": 1652359168 + }, + { + "epoch": 0.5, + "learning_rate": 0.0004035466399197593, + "loss": 1.5619, + "theoretical_loss": 3.514948011228547, + "tokens_seen": 1652424704 + }, + { + "epoch": 0.5, + "learning_rate": 0.00040353059177532604, + "loss": 1.3745, + "theoretical_loss": 3.5149360450386062, + "tokens_seen": 1652490240 + }, + { + "epoch": 0.5, + "learning_rate": 0.0004035145436308927, + "loss": 1.4631, + "theoretical_loss": 3.514924079456095, + "tokens_seen": 1652555776 + }, + { + "epoch": 0.5, + "learning_rate": 0.00040349849548645944, + "loss": 1.5133, + "theoretical_loss": 3.5149121144809587, + "tokens_seen": 1652621312 + }, + { + "epoch": 0.5, + "learning_rate": 0.0004034824473420261, + "loss": 1.3449, + "theoretical_loss": 3.5149001501131423, + "tokens_seen": 1652686848 + }, + { + "epoch": 0.5, + "learning_rate": 0.00040346639919759284, + "loss": 1.3376, + "theoretical_loss": 3.5148881863525903, + "tokens_seen": 1652752384 + }, + { + "epoch": 0.5, + "learning_rate": 0.0004034503510531595, + "loss": 1.3819, + "theoretical_loss": 3.5148762231992476, + "tokens_seen": 1652817920 + }, + { + "epoch": 0.5, + "learning_rate": 0.00040343430290872625, + "loss": 1.4095, + "theoretical_loss": 3.5148642606530602, + "tokens_seen": 1652883456 + }, + { + "epoch": 0.5, + "learning_rate": 0.0004034182547642929, + "loss": 1.3877, + "theoretical_loss": 3.514852298713973, + "tokens_seen": 1652948992 + }, + { + "epoch": 0.5, + "learning_rate": 0.00040340220661985965, + "loss": 1.3629, + "theoretical_loss": 3.514840337381931, + "tokens_seen": 1653014528 + }, + { + "epoch": 0.5, + "learning_rate": 0.0004033861584754263, + "loss": 1.4115, + "theoretical_loss": 3.5148283766568795, + "tokens_seen": 1653080064 + }, + { + "epoch": 0.5, + "learning_rate": 0.00040337011033099305, + "loss": 1.4195, + "theoretical_loss": 3.514816416538763, + "tokens_seen": 1653145600 + }, + { + "epoch": 0.5, + "learning_rate": 0.0004033540621865597, + "loss": 1.3182, + "theoretical_loss": 3.5148044570275276, + "tokens_seen": 1653211136 + }, + { + "epoch": 0.5, + "learning_rate": 0.00040333801404212645, + "loss": 1.374, + "theoretical_loss": 3.5147924981231173, + "tokens_seen": 1653276672 + }, + { + "epoch": 0.5, + "learning_rate": 0.0004033219658976931, + "loss": 1.2938, + "theoretical_loss": 3.5147805398254786, + "tokens_seen": 1653342208 + }, + { + "epoch": 0.5, + "learning_rate": 0.00040330591775325985, + "loss": 1.2952, + "theoretical_loss": 3.514768582134555, + "tokens_seen": 1653407744 + }, + { + "epoch": 0.5, + "learning_rate": 0.0004032898696088265, + "loss": 1.3548, + "theoretical_loss": 3.5147566250502935, + "tokens_seen": 1653473280 + }, + { + "epoch": 0.5, + "learning_rate": 0.00040327382146439325, + "loss": 1.367, + "theoretical_loss": 3.514744668572638, + "tokens_seen": 1653538816 + }, + { + "epoch": 0.5, + "learning_rate": 0.0004032577733199599, + "loss": 1.4489, + "theoretical_loss": 3.514732712701534, + "tokens_seen": 1653604352 + }, + { + "epoch": 0.5, + "learning_rate": 0.00040324172517552665, + "loss": 1.2677, + "theoretical_loss": 3.514720757436927, + "tokens_seen": 1653669888 + }, + { + "epoch": 0.5, + "learning_rate": 0.0004032256770310933, + "loss": 1.2882, + "theoretical_loss": 3.5147088027787623, + "tokens_seen": 1653735424 + }, + { + "epoch": 0.5, + "learning_rate": 0.00040320962888666005, + "loss": 1.4898, + "theoretical_loss": 3.5146968487269845, + "tokens_seen": 1653800960 + }, + { + "epoch": 0.5, + "learning_rate": 0.0004031935807422267, + "loss": 1.3978, + "theoretical_loss": 3.5146848952815386, + "tokens_seen": 1653866496 + }, + { + "epoch": 0.5, + "learning_rate": 0.0004031775325977934, + "loss": 1.3654, + "theoretical_loss": 3.5146729424423713, + "tokens_seen": 1653932032 + }, + { + "epoch": 0.5, + "learning_rate": 0.0004031614844533601, + "loss": 1.4753, + "theoretical_loss": 3.514660990209426, + "tokens_seen": 1653997568 + }, + { + "epoch": 0.5, + "learning_rate": 0.0004031454363089268, + "loss": 1.3783, + "theoretical_loss": 3.5146490385826494, + "tokens_seen": 1654063104 + }, + { + "epoch": 0.5, + "learning_rate": 0.0004031293881644935, + "loss": 1.3196, + "theoretical_loss": 3.5146370875619857, + "tokens_seen": 1654128640 + }, + { + "epoch": 0.5, + "learning_rate": 0.0004031133400200602, + "loss": 1.1628, + "theoretical_loss": 3.5146251371473807, + "tokens_seen": 1654194176 + }, + { + "epoch": 0.5, + "learning_rate": 0.0004030972918756269, + "loss": 1.4076, + "theoretical_loss": 3.51461318733878, + "tokens_seen": 1654259712 + }, + { + "epoch": 0.5, + "learning_rate": 0.0004030812437311936, + "loss": 1.4034, + "theoretical_loss": 3.5146012381361276, + "tokens_seen": 1654325248 + }, + { + "epoch": 0.5, + "learning_rate": 0.0004030651955867603, + "loss": 1.4099, + "theoretical_loss": 3.51458928953937, + "tokens_seen": 1654390784 + }, + { + "epoch": 0.5, + "learning_rate": 0.000403049147442327, + "loss": 1.3085, + "theoretical_loss": 3.514577341548452, + "tokens_seen": 1654456320 + }, + { + "epoch": 0.5, + "learning_rate": 0.00040303309929789365, + "loss": 1.2695, + "theoretical_loss": 3.5145653941633186, + "tokens_seen": 1654521856 + }, + { + "epoch": 0.5, + "learning_rate": 0.0004030170511534604, + "loss": 1.379, + "theoretical_loss": 3.514553447383916, + "tokens_seen": 1654587392 + }, + { + "epoch": 0.5, + "learning_rate": 0.00040300100300902705, + "loss": 1.3892, + "theoretical_loss": 3.5145415012101884, + "tokens_seen": 1654652928 + }, + { + "epoch": 0.5, + "learning_rate": 0.0004029849548645938, + "loss": 1.4003, + "theoretical_loss": 3.514529555642082, + "tokens_seen": 1654718464 + }, + { + "epoch": 0.5, + "objective/train/docs_used": 604403, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.4673120975494385, + "objective/train/theoretical_loss": 3.5145176106795417, + "objective/train/tokens_used": 1675244000, + "theoretical_loss": 3.5145176106795417, + "tokens_seen": 1654784000 + }, + { + "epoch": 0.5, + "learning_rate": 0.00040296890672016046, + "loss": 1.4661, + "theoretical_loss": 3.5145176106795417, + "tokens_seen": 1654784000 + }, + { + "epoch": 0.5, + "learning_rate": 0.0004029528585757272, + "loss": 1.4666, + "theoretical_loss": 3.514505666322513, + "tokens_seen": 1654849536 + }, + { + "epoch": 0.5, + "learning_rate": 0.00040293681043129386, + "loss": 1.3375, + "theoretical_loss": 3.5144937225709407, + "tokens_seen": 1654915072 + }, + { + "epoch": 0.5, + "learning_rate": 0.0004029207622868606, + "loss": 1.3582, + "theoretical_loss": 3.514481779424771, + "tokens_seen": 1654980608 + }, + { + "epoch": 0.5, + "learning_rate": 0.00040290471414242726, + "loss": 1.2788, + "theoretical_loss": 3.5144698368839484, + "tokens_seen": 1655046144 + }, + { + "epoch": 0.5, + "learning_rate": 0.000402888665997994, + "loss": 1.32, + "theoretical_loss": 3.5144578949484186, + "tokens_seen": 1655111680 + }, + { + "epoch": 0.5, + "learning_rate": 0.00040287261785356066, + "loss": 1.3711, + "theoretical_loss": 3.514445953618127, + "tokens_seen": 1655177216 + }, + { + "epoch": 0.5, + "learning_rate": 0.0004028565697091274, + "loss": 1.3338, + "theoretical_loss": 3.5144340128930196, + "tokens_seen": 1655242752 + }, + { + "epoch": 0.5, + "learning_rate": 0.00040284052156469406, + "loss": 1.3805, + "theoretical_loss": 3.51442207277304, + "tokens_seen": 1655308288 + }, + { + "epoch": 0.5, + "learning_rate": 0.0004028244734202608, + "loss": 1.3227, + "theoretical_loss": 3.5144101332581354, + "tokens_seen": 1655373824 + }, + { + "epoch": 0.5, + "learning_rate": 0.00040280842527582746, + "loss": 1.3021, + "theoretical_loss": 3.5143981943482503, + "tokens_seen": 1655439360 + }, + { + "epoch": 0.5, + "learning_rate": 0.0004027923771313942, + "loss": 1.3829, + "theoretical_loss": 3.5143862560433305, + "tokens_seen": 1655504896 + }, + { + "epoch": 0.5, + "learning_rate": 0.00040277632898696086, + "loss": 1.4407, + "theoretical_loss": 3.5143743183433207, + "tokens_seen": 1655570432 + }, + { + "epoch": 0.5, + "learning_rate": 0.0004027602808425276, + "loss": 1.3111, + "theoretical_loss": 3.514362381248167, + "tokens_seen": 1655635968 + }, + { + "epoch": 0.5, + "learning_rate": 0.00040274423269809426, + "loss": 1.4982, + "theoretical_loss": 3.514350444757815, + "tokens_seen": 1655701504 + }, + { + "epoch": 0.5, + "learning_rate": 0.000402728184553661, + "loss": 1.3327, + "theoretical_loss": 3.5143385088722088, + "tokens_seen": 1655767040 + }, + { + "epoch": 0.5, + "learning_rate": 0.00040271213640922767, + "loss": 1.3258, + "theoretical_loss": 3.5143265735912954, + "tokens_seen": 1655832576 + }, + { + "epoch": 0.5, + "learning_rate": 0.0004026960882647944, + "loss": 1.3301, + "theoretical_loss": 3.5143146389150193, + "tokens_seen": 1655898112 + }, + { + "epoch": 0.5, + "learning_rate": 0.00040268004012036107, + "loss": 1.3604, + "theoretical_loss": 3.5143027048433266, + "tokens_seen": 1655963648 + }, + { + "epoch": 0.5, + "learning_rate": 0.0004026639919759278, + "loss": 1.3402, + "theoretical_loss": 3.514290771376162, + "tokens_seen": 1656029184 + }, + { + "epoch": 0.5, + "learning_rate": 0.00040264794383149447, + "loss": 1.421, + "theoretical_loss": 3.514278838513471, + "tokens_seen": 1656094720 + }, + { + "epoch": 0.5, + "learning_rate": 0.0004026318956870612, + "loss": 1.3221, + "theoretical_loss": 3.5142669062552, + "tokens_seen": 1656160256 + }, + { + "epoch": 0.5, + "learning_rate": 0.00040261584754262787, + "loss": 1.2721, + "theoretical_loss": 3.514254974601293, + "tokens_seen": 1656225792 + }, + { + "epoch": 0.5, + "learning_rate": 0.0004025997993981946, + "loss": 1.4566, + "theoretical_loss": 3.5142430435516974, + "tokens_seen": 1656291328 + }, + { + "epoch": 0.5, + "learning_rate": 0.00040258375125376127, + "loss": 1.3959, + "theoretical_loss": 3.514231113106357, + "tokens_seen": 1656356864 + }, + { + "epoch": 0.5, + "learning_rate": 0.000402567703109328, + "loss": 1.3951, + "theoretical_loss": 3.514219183265218, + "tokens_seen": 1656422400 + }, + { + "epoch": 0.5, + "learning_rate": 0.00040255165496489467, + "loss": 1.3433, + "theoretical_loss": 3.514207254028226, + "tokens_seen": 1656487936 + }, + { + "epoch": 0.5, + "learning_rate": 0.00040253560682046143, + "loss": 1.3998, + "theoretical_loss": 3.5141953253953258, + "tokens_seen": 1656553472 + }, + { + "epoch": 0.5, + "learning_rate": 0.0004025195586760281, + "loss": 1.335, + "theoretical_loss": 3.5141833973664633, + "tokens_seen": 1656619008 + }, + { + "epoch": 0.5, + "learning_rate": 0.00040250351053159483, + "loss": 1.3451, + "theoretical_loss": 3.5141714699415845, + "tokens_seen": 1656684544 + }, + { + "epoch": 0.5, + "learning_rate": 0.0004024874623871615, + "loss": 1.3483, + "theoretical_loss": 3.5141595431206345, + "tokens_seen": 1656750080 + }, + { + "epoch": 0.5, + "learning_rate": 0.00040247141424272823, + "loss": 1.3367, + "theoretical_loss": 3.514147616903559, + "tokens_seen": 1656815616 + }, + { + "epoch": 0.5, + "learning_rate": 0.0004024553660982949, + "loss": 1.3624, + "theoretical_loss": 3.5141356912903032, + "tokens_seen": 1656881152 + }, + { + "epoch": 0.5, + "learning_rate": 0.00040243931795386163, + "loss": 1.3876, + "theoretical_loss": 3.5141237662808127, + "tokens_seen": 1656946688 + }, + { + "epoch": 0.5, + "learning_rate": 0.0004024232698094283, + "loss": 1.2823, + "theoretical_loss": 3.5141118418750334, + "tokens_seen": 1657012224 + }, + { + "epoch": 0.5, + "learning_rate": 0.000402407221664995, + "loss": 1.4388, + "theoretical_loss": 3.514099918072911, + "tokens_seen": 1657077760 + }, + { + "epoch": 0.5, + "learning_rate": 0.00040239117352056173, + "loss": 1.272, + "theoretical_loss": 3.514087994874391, + "tokens_seen": 1657143296 + }, + { + "epoch": 0.5, + "learning_rate": 0.0004023751253761284, + "loss": 1.2446, + "theoretical_loss": 3.5140760722794178, + "tokens_seen": 1657208832 + }, + { + "epoch": 0.5, + "learning_rate": 0.00040235907723169513, + "loss": 1.332, + "theoretical_loss": 3.5140641502879384, + "tokens_seen": 1657274368 + }, + { + "epoch": 0.5, + "learning_rate": 0.0004023430290872618, + "loss": 1.411, + "theoretical_loss": 3.5140522288998977, + "tokens_seen": 1657339904 + }, + { + "epoch": 0.5, + "learning_rate": 0.00040232698094282854, + "loss": 1.3413, + "theoretical_loss": 3.514040308115242, + "tokens_seen": 1657405440 + }, + { + "epoch": 0.5, + "learning_rate": 0.0004023109327983952, + "loss": 1.2424, + "theoretical_loss": 3.514028387933916, + "tokens_seen": 1657470976 + }, + { + "epoch": 0.5, + "learning_rate": 0.00040229488465396194, + "loss": 1.3175, + "theoretical_loss": 3.514016468355866, + "tokens_seen": 1657536512 + }, + { + "epoch": 0.5, + "learning_rate": 0.0004022788365095286, + "loss": 1.3221, + "theoretical_loss": 3.5140045493810366, + "tokens_seen": 1657602048 + }, + { + "epoch": 0.5, + "learning_rate": 0.00040226278836509534, + "loss": 1.3868, + "theoretical_loss": 3.5139926310093745, + "tokens_seen": 1657667584 + }, + { + "epoch": 0.5, + "learning_rate": 0.000402246740220662, + "loss": 1.3092, + "theoretical_loss": 3.5139807132408256, + "tokens_seen": 1657733120 + }, + { + "epoch": 0.5, + "learning_rate": 0.00040223069207622874, + "loss": 1.2746, + "theoretical_loss": 3.513968796075334, + "tokens_seen": 1657798656 + }, + { + "epoch": 0.5, + "learning_rate": 0.0004022146439317954, + "loss": 1.4613, + "theoretical_loss": 3.513956879512847, + "tokens_seen": 1657864192 + }, + { + "epoch": 0.5, + "learning_rate": 0.00040219859578736214, + "loss": 1.3116, + "theoretical_loss": 3.513944963553309, + "tokens_seen": 1657929728 + }, + { + "epoch": 0.5, + "learning_rate": 0.0004021825476429288, + "loss": 1.3511, + "theoretical_loss": 3.513933048196667, + "tokens_seen": 1657995264 + }, + { + "epoch": 0.5, + "objective/train/docs_used": 605625, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.3391187191009521, + "objective/train/theoretical_loss": 3.5139211334428646, + "objective/train/tokens_used": 1678520800, + "theoretical_loss": 3.5139211334428646, + "tokens_seen": 1658060800 + }, + { + "epoch": 0.5, + "learning_rate": 0.00040216649949849554, + "loss": 1.3019, + "theoretical_loss": 3.5139211334428646, + "tokens_seen": 1658060800 + }, + { + "epoch": 0.5, + "learning_rate": 0.0004021504513540622, + "loss": 1.2861, + "theoretical_loss": 3.5139092192918495, + "tokens_seen": 1658126336 + }, + { + "epoch": 0.5, + "learning_rate": 0.00040213440320962894, + "loss": 1.4493, + "theoretical_loss": 3.5138973057435665, + "tokens_seen": 1658191872 + }, + { + "epoch": 0.5, + "learning_rate": 0.0004021183550651956, + "loss": 1.3197, + "theoretical_loss": 3.5138853927979614, + "tokens_seen": 1658257408 + }, + { + "epoch": 0.5, + "learning_rate": 0.00040210230692076234, + "loss": 1.2267, + "theoretical_loss": 3.51387348045498, + "tokens_seen": 1658322944 + }, + { + "epoch": 0.5, + "learning_rate": 0.000402086258776329, + "loss": 1.3766, + "theoretical_loss": 3.513861568714568, + "tokens_seen": 1658388480 + }, + { + "epoch": 0.5, + "learning_rate": 0.00040207021063189575, + "loss": 1.3995, + "theoretical_loss": 3.513849657576671, + "tokens_seen": 1658454016 + }, + { + "epoch": 0.5, + "learning_rate": 0.0004020541624874624, + "loss": 1.3804, + "theoretical_loss": 3.513837747041234, + "tokens_seen": 1658519552 + }, + { + "epoch": 0.5, + "learning_rate": 0.00040203811434302915, + "loss": 1.3325, + "theoretical_loss": 3.513825837108204, + "tokens_seen": 1658585088 + }, + { + "epoch": 0.5, + "learning_rate": 0.0004020220661985958, + "loss": 1.4385, + "theoretical_loss": 3.513813927777526, + "tokens_seen": 1658650624 + }, + { + "epoch": 0.5, + "learning_rate": 0.00040200601805416255, + "loss": 1.445, + "theoretical_loss": 3.5138020190491464, + "tokens_seen": 1658716160 + }, + { + "epoch": 0.5, + "learning_rate": 0.0004019899699097292, + "loss": 1.2413, + "theoretical_loss": 3.5137901109230096, + "tokens_seen": 1658781696 + }, + { + "epoch": 0.5, + "learning_rate": 0.00040197392176529595, + "loss": 1.3409, + "theoretical_loss": 3.513778203399063, + "tokens_seen": 1658847232 + }, + { + "epoch": 0.5, + "learning_rate": 0.0004019578736208626, + "loss": 1.3124, + "theoretical_loss": 3.513766296477251, + "tokens_seen": 1658912768 + }, + { + "epoch": 0.5, + "learning_rate": 0.00040194182547642935, + "loss": 1.2505, + "theoretical_loss": 3.51375439015752, + "tokens_seen": 1658978304 + }, + { + "epoch": 0.5, + "learning_rate": 0.000401925777331996, + "loss": 1.3578, + "theoretical_loss": 3.5137424844398164, + "tokens_seen": 1659043840 + }, + { + "epoch": 0.5, + "learning_rate": 0.00040190972918756275, + "loss": 1.3294, + "theoretical_loss": 3.5137305793240845, + "tokens_seen": 1659109376 + }, + { + "epoch": 0.5, + "learning_rate": 0.0004018936810431294, + "loss": 1.3791, + "theoretical_loss": 3.5137186748102716, + "tokens_seen": 1659174912 + }, + { + "epoch": 0.5, + "learning_rate": 0.00040187763289869615, + "loss": 1.3705, + "theoretical_loss": 3.513706770898322, + "tokens_seen": 1659240448 + }, + { + "epoch": 0.5, + "learning_rate": 0.0004018615847542628, + "loss": 1.3508, + "theoretical_loss": 3.5136948675881827, + "tokens_seen": 1659305984 + }, + { + "epoch": 0.5, + "learning_rate": 0.00040184553660982955, + "loss": 1.3632, + "theoretical_loss": 3.5136829648797985, + "tokens_seen": 1659371520 + }, + { + "epoch": 0.5, + "learning_rate": 0.0004018294884653962, + "loss": 1.257, + "theoretical_loss": 3.5136710627731165, + "tokens_seen": 1659437056 + }, + { + "epoch": 0.5, + "learning_rate": 0.00040181344032096296, + "loss": 1.3533, + "theoretical_loss": 3.5136591612680816, + "tokens_seen": 1659502592 + }, + { + "epoch": 0.5, + "learning_rate": 0.0004017973921765296, + "loss": 1.3297, + "theoretical_loss": 3.51364726036464, + "tokens_seen": 1659568128 + }, + { + "epoch": 0.5, + "learning_rate": 0.0004017813440320963, + "loss": 1.4662, + "theoretical_loss": 3.513635360062737, + "tokens_seen": 1659633664 + }, + { + "epoch": 0.5, + "learning_rate": 0.000401765295887663, + "loss": 1.3187, + "theoretical_loss": 3.513623460362319, + "tokens_seen": 1659699200 + }, + { + "epoch": 0.5, + "learning_rate": 0.0004017492477432297, + "loss": 1.3721, + "theoretical_loss": 3.5136115612633314, + "tokens_seen": 1659764736 + }, + { + "epoch": 0.5, + "learning_rate": 0.0004017331995987964, + "loss": 1.4124, + "theoretical_loss": 3.5135996627657207, + "tokens_seen": 1659830272 + }, + { + "epoch": 0.5, + "learning_rate": 0.0004017171514543631, + "loss": 1.285, + "theoretical_loss": 3.513587764869432, + "tokens_seen": 1659895808 + }, + { + "epoch": 0.5, + "learning_rate": 0.0004017011033099298, + "loss": 1.3707, + "theoretical_loss": 3.5135758675744118, + "tokens_seen": 1659961344 + }, + { + "epoch": 0.5, + "learning_rate": 0.0004016850551654965, + "loss": 1.3783, + "theoretical_loss": 3.5135639708806057, + "tokens_seen": 1660026880 + }, + { + "epoch": 0.5, + "learning_rate": 0.0004016690070210632, + "loss": 1.3953, + "theoretical_loss": 3.5135520747879596, + "tokens_seen": 1660092416 + }, + { + "epoch": 0.5, + "learning_rate": 0.0004016529588766299, + "loss": 1.271, + "theoretical_loss": 3.51354017929642, + "tokens_seen": 1660157952 + }, + { + "epoch": 0.5, + "learning_rate": 0.00040163691073219655, + "loss": 1.32, + "theoretical_loss": 3.5135282844059317, + "tokens_seen": 1660223488 + }, + { + "epoch": 0.5, + "learning_rate": 0.0004016208625877633, + "loss": 1.3156, + "theoretical_loss": 3.5135163901164406, + "tokens_seen": 1660289024 + }, + { + "epoch": 0.5, + "learning_rate": 0.00040160481444332996, + "loss": 1.3893, + "theoretical_loss": 3.5135044964278936, + "tokens_seen": 1660354560 + }, + { + "epoch": 0.5, + "learning_rate": 0.0004015887662988967, + "loss": 1.4374, + "theoretical_loss": 3.5134926033402367, + "tokens_seen": 1660420096 + }, + { + "epoch": 0.5, + "learning_rate": 0.00040157271815446336, + "loss": 1.3201, + "theoretical_loss": 3.5134807108534147, + "tokens_seen": 1660485632 + }, + { + "epoch": 0.5, + "learning_rate": 0.0004015566700100301, + "loss": 1.3204, + "theoretical_loss": 3.513468818967374, + "tokens_seen": 1660551168 + }, + { + "epoch": 0.5, + "learning_rate": 0.00040154062186559676, + "loss": 1.4262, + "theoretical_loss": 3.5134569276820606, + "tokens_seen": 1660616704 + }, + { + "epoch": 0.5, + "learning_rate": 0.0004015245737211635, + "loss": 1.2822, + "theoretical_loss": 3.513445036997421, + "tokens_seen": 1660682240 + }, + { + "epoch": 0.5, + "learning_rate": 0.00040150852557673016, + "loss": 1.4617, + "theoretical_loss": 3.5134331469134006, + "tokens_seen": 1660747776 + }, + { + "epoch": 0.5, + "learning_rate": 0.0004014924774322969, + "loss": 1.2269, + "theoretical_loss": 3.513421257429945, + "tokens_seen": 1660813312 + }, + { + "epoch": 0.5, + "learning_rate": 0.00040147642928786356, + "loss": 1.4433, + "theoretical_loss": 3.5134093685470007, + "tokens_seen": 1660878848 + }, + { + "epoch": 0.5, + "learning_rate": 0.0004014603811434303, + "loss": 1.3425, + "theoretical_loss": 3.513397480264514, + "tokens_seen": 1660944384 + }, + { + "epoch": 0.5, + "learning_rate": 0.00040144433299899696, + "loss": 1.3597, + "theoretical_loss": 3.5133855925824298, + "tokens_seen": 1661009920 + }, + { + "epoch": 0.5, + "learning_rate": 0.0004014282848545637, + "loss": 1.421, + "theoretical_loss": 3.513373705500695, + "tokens_seen": 1661075456 + }, + { + "epoch": 0.5, + "learning_rate": 0.00040141223671013036, + "loss": 1.437, + "theoretical_loss": 3.5133618190192553, + "tokens_seen": 1661140992 + }, + { + "epoch": 0.5, + "learning_rate": 0.0004013961885656971, + "loss": 1.3247, + "theoretical_loss": 3.513349933138057, + "tokens_seen": 1661206528 + }, + { + "epoch": 0.5, + "learning_rate": 0.00040138014042126376, + "loss": 1.3129, + "theoretical_loss": 3.5133380478570455, + "tokens_seen": 1661272064 + }, + { + "epoch": 0.5, + "objective/train/docs_used": 606795, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.4029767513275146, + "objective/train/theoretical_loss": 3.513326163176167, + "objective/train/tokens_used": 1681797600, + "theoretical_loss": 3.513326163176167, + "tokens_seen": 1661337600 + }, + { + "epoch": 0.5, + "learning_rate": 0.0004013640922768305, + "loss": 1.3508, + "theoretical_loss": 3.513326163176167, + "tokens_seen": 1661337600 + }, + { + "epoch": 0.5, + "learning_rate": 0.00040134804413239717, + "loss": 1.2974, + "theoretical_loss": 3.5133142790953684, + "tokens_seen": 1661403136 + }, + { + "epoch": 0.5, + "learning_rate": 0.0004013319959879639, + "loss": 1.2035, + "theoretical_loss": 3.513302395614595, + "tokens_seen": 1661468672 + }, + { + "epoch": 0.5, + "learning_rate": 0.00040131594784353057, + "loss": 1.3578, + "theoretical_loss": 3.513290512733792, + "tokens_seen": 1661534208 + }, + { + "epoch": 0.5, + "learning_rate": 0.0004012998996990973, + "loss": 1.3821, + "theoretical_loss": 3.513278630452907, + "tokens_seen": 1661599744 + }, + { + "epoch": 0.5, + "learning_rate": 0.00040128385155466397, + "loss": 1.3352, + "theoretical_loss": 3.513266748771885, + "tokens_seen": 1661665280 + }, + { + "epoch": 0.5, + "learning_rate": 0.0004012678034102307, + "loss": 1.3052, + "theoretical_loss": 3.5132548676906725, + "tokens_seen": 1661730816 + }, + { + "epoch": 0.5, + "learning_rate": 0.00040125175526579737, + "loss": 1.2113, + "theoretical_loss": 3.5132429872092157, + "tokens_seen": 1661796352 + }, + { + "epoch": 0.5, + "learning_rate": 0.0004012357071213641, + "loss": 1.3179, + "theoretical_loss": 3.51323110732746, + "tokens_seen": 1661861888 + }, + { + "epoch": 0.5, + "learning_rate": 0.00040121965897693077, + "loss": 1.248, + "theoretical_loss": 3.5132192280453527, + "tokens_seen": 1661927424 + }, + { + "epoch": 0.5, + "learning_rate": 0.0004012036108324975, + "loss": 1.3589, + "theoretical_loss": 3.5132073493628386, + "tokens_seen": 1661992960 + }, + { + "epoch": 0.5, + "learning_rate": 0.0004011875626880643, + "loss": 1.362, + "theoretical_loss": 3.5131954712798645, + "tokens_seen": 1662058496 + }, + { + "epoch": 0.5, + "learning_rate": 0.0004011715145436309, + "loss": 1.3458, + "theoretical_loss": 3.513183593796376, + "tokens_seen": 1662124032 + }, + { + "epoch": 0.5, + "learning_rate": 0.00040115546639919763, + "loss": 1.3606, + "theoretical_loss": 3.5131717169123196, + "tokens_seen": 1662189568 + }, + { + "epoch": 0.5, + "learning_rate": 0.00040113941825476433, + "loss": 1.3875, + "theoretical_loss": 3.5131598406276416, + "tokens_seen": 1662255104 + }, + { + "epoch": 0.5, + "learning_rate": 0.00040112337011033103, + "loss": 1.3602, + "theoretical_loss": 3.513147964942288, + "tokens_seen": 1662320640 + }, + { + "epoch": 0.5, + "learning_rate": 0.00040110732196589773, + "loss": 1.3952, + "theoretical_loss": 3.5131360898562045, + "tokens_seen": 1662386176 + }, + { + "epoch": 0.5, + "learning_rate": 0.00040109127382146443, + "loss": 1.4331, + "theoretical_loss": 3.513124215369338, + "tokens_seen": 1662451712 + }, + { + "epoch": 0.5, + "learning_rate": 0.00040107522567703113, + "loss": 1.3464, + "theoretical_loss": 3.5131123414816336, + "tokens_seen": 1662517248 + }, + { + "epoch": 0.5, + "learning_rate": 0.00040105917753259783, + "loss": 1.3931, + "theoretical_loss": 3.513100468193038, + "tokens_seen": 1662582784 + }, + { + "epoch": 0.5, + "learning_rate": 0.00040104312938816453, + "loss": 1.396, + "theoretical_loss": 3.513088595503498, + "tokens_seen": 1662648320 + }, + { + "epoch": 0.5, + "learning_rate": 0.00040102708124373123, + "loss": 1.3155, + "theoretical_loss": 3.5130767234129587, + "tokens_seen": 1662713856 + }, + { + "epoch": 0.5, + "learning_rate": 0.0004010110330992979, + "loss": 1.3713, + "theoretical_loss": 3.5130648519213663, + "tokens_seen": 1662779392 + }, + { + "epoch": 0.5, + "learning_rate": 0.00040099498495486463, + "loss": 1.5362, + "theoretical_loss": 3.513052981028668, + "tokens_seen": 1662844928 + }, + { + "epoch": 0.5, + "learning_rate": 0.0004009789368104313, + "loss": 1.2979, + "theoretical_loss": 3.513041110734809, + "tokens_seen": 1662910464 + }, + { + "epoch": 0.5, + "learning_rate": 0.00040096288866599804, + "loss": 1.2572, + "theoretical_loss": 3.513029241039737, + "tokens_seen": 1662976000 + }, + { + "epoch": 0.5, + "learning_rate": 0.0004009468405215647, + "loss": 1.3221, + "theoretical_loss": 3.5130173719433957, + "tokens_seen": 1663041536 + }, + { + "epoch": 0.5, + "learning_rate": 0.00040093079237713144, + "loss": 1.3502, + "theoretical_loss": 3.5130055034457337, + "tokens_seen": 1663107072 + }, + { + "epoch": 0.5, + "learning_rate": 0.0004009147442326981, + "loss": 1.3863, + "theoretical_loss": 3.5129936355466955, + "tokens_seen": 1663172608 + }, + { + "epoch": 0.5, + "learning_rate": 0.00040089869608826484, + "loss": 1.3329, + "theoretical_loss": 3.5129817682462283, + "tokens_seen": 1663238144 + }, + { + "epoch": 0.5, + "learning_rate": 0.0004008826479438315, + "loss": 1.4639, + "theoretical_loss": 3.512969901544278, + "tokens_seen": 1663303680 + }, + { + "epoch": 0.5, + "learning_rate": 0.00040086659979939824, + "loss": 1.3712, + "theoretical_loss": 3.512958035440791, + "tokens_seen": 1663369216 + }, + { + "epoch": 0.5, + "learning_rate": 0.0004008505516549649, + "loss": 1.4065, + "theoretical_loss": 3.512946169935713, + "tokens_seen": 1663434752 + }, + { + "epoch": 0.5, + "learning_rate": 0.00040083450351053164, + "loss": 1.3697, + "theoretical_loss": 3.512934305028991, + "tokens_seen": 1663500288 + }, + { + "epoch": 0.5, + "learning_rate": 0.0004008184553660983, + "loss": 1.2795, + "theoretical_loss": 3.512922440720571, + "tokens_seen": 1663565824 + }, + { + "epoch": 0.5, + "learning_rate": 0.00040080240722166504, + "loss": 1.3637, + "theoretical_loss": 3.512910577010399, + "tokens_seen": 1663631360 + }, + { + "epoch": 0.5, + "learning_rate": 0.0004007863590772317, + "loss": 1.3242, + "theoretical_loss": 3.5128987138984216, + "tokens_seen": 1663696896 + }, + { + "epoch": 0.5, + "learning_rate": 0.00040077031093279844, + "loss": 1.2533, + "theoretical_loss": 3.512886851384585, + "tokens_seen": 1663762432 + }, + { + "epoch": 0.5, + "learning_rate": 0.0004007542627883651, + "loss": 1.3989, + "theoretical_loss": 3.512874989468835, + "tokens_seen": 1663827968 + }, + { + "epoch": 0.5, + "learning_rate": 0.00040073821464393184, + "loss": 1.3958, + "theoretical_loss": 3.5128631281511185, + "tokens_seen": 1663893504 + }, + { + "epoch": 0.5, + "learning_rate": 0.0004007221664994985, + "loss": 1.3215, + "theoretical_loss": 3.512851267431382, + "tokens_seen": 1663959040 + }, + { + "epoch": 0.5, + "learning_rate": 0.00040070611835506525, + "loss": 1.3234, + "theoretical_loss": 3.5128394073095706, + "tokens_seen": 1664024576 + }, + { + "epoch": 0.5, + "learning_rate": 0.0004006900702106319, + "loss": 1.3125, + "theoretical_loss": 3.512827547785632, + "tokens_seen": 1664090112 + }, + { + "epoch": 0.5, + "learning_rate": 0.00040067402206619865, + "loss": 1.3709, + "theoretical_loss": 3.512815688859512, + "tokens_seen": 1664155648 + }, + { + "epoch": 0.5, + "learning_rate": 0.0004006579739217653, + "loss": 1.1552, + "theoretical_loss": 3.5128038305311566, + "tokens_seen": 1664221184 + }, + { + "epoch": 0.5, + "learning_rate": 0.00040064192577733205, + "loss": 1.3972, + "theoretical_loss": 3.512791972800512, + "tokens_seen": 1664286720 + }, + { + "epoch": 0.5, + "learning_rate": 0.0004006258776328987, + "loss": 1.4504, + "theoretical_loss": 3.5127801156675256, + "tokens_seen": 1664352256 + }, + { + "epoch": 0.5, + "learning_rate": 0.00040060982948846545, + "loss": 1.4006, + "theoretical_loss": 3.5127682591321427, + "tokens_seen": 1664417792 + }, + { + "epoch": 0.5, + "learning_rate": 0.0004005937813440321, + "loss": 1.392, + "theoretical_loss": 3.5127564031943104, + "tokens_seen": 1664483328 + }, + { + "epoch": 0.5, + "learning_rate": 0.00040057773319959885, + "loss": 1.4502, + "theoretical_loss": 3.5127445478539743, + "tokens_seen": 1664548864 + }, + { + "epoch": 0.5, + "objective/train/docs_used": 608106, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.2445247173309326, + "objective/train/theoretical_loss": 3.512732693111081, + "objective/train/tokens_used": 1685074400, + "theoretical_loss": 3.512732693111081, + "tokens_seen": 1664614400 + }, + { + "epoch": 0.5, + "learning_rate": 0.0004005616850551655, + "loss": 1.3224, + "theoretical_loss": 3.512732693111081, + "tokens_seen": 1664614400 + }, + { + "epoch": 0.5, + "learning_rate": 0.00040054563691073225, + "loss": 1.407, + "theoretical_loss": 3.5127208389655777, + "tokens_seen": 1664679936 + }, + { + "epoch": 0.5, + "learning_rate": 0.0004005295887662989, + "loss": 1.3835, + "theoretical_loss": 3.51270898541741, + "tokens_seen": 1664745472 + }, + { + "epoch": 0.5, + "learning_rate": 0.00040051354062186565, + "loss": 1.308, + "theoretical_loss": 3.5126971324665237, + "tokens_seen": 1664811008 + }, + { + "epoch": 0.5, + "learning_rate": 0.0004004974924774323, + "loss": 1.2591, + "theoretical_loss": 3.512685280112866, + "tokens_seen": 1664876544 + }, + { + "epoch": 0.5, + "learning_rate": 0.00040048144433299905, + "loss": 1.3752, + "theoretical_loss": 3.5126734283563836, + "tokens_seen": 1664942080 + }, + { + "epoch": 0.5, + "learning_rate": 0.0004004653961885657, + "loss": 1.4218, + "theoretical_loss": 3.512661577197022, + "tokens_seen": 1665007616 + }, + { + "epoch": 0.5, + "learning_rate": 0.00040044934804413246, + "loss": 1.3297, + "theoretical_loss": 3.5126497266347285, + "tokens_seen": 1665073152 + }, + { + "epoch": 0.5, + "learning_rate": 0.0004004332998996991, + "loss": 1.234, + "theoretical_loss": 3.512637876669449, + "tokens_seen": 1665138688 + }, + { + "epoch": 0.5, + "learning_rate": 0.00040041725175526586, + "loss": 1.4524, + "theoretical_loss": 3.5126260273011303, + "tokens_seen": 1665204224 + }, + { + "epoch": 0.5, + "learning_rate": 0.0004004012036108325, + "loss": 1.336, + "theoretical_loss": 3.5126141785297182, + "tokens_seen": 1665269760 + }, + { + "epoch": 0.5, + "learning_rate": 0.0004003851554663992, + "loss": 1.3473, + "theoretical_loss": 3.5126023303551595, + "tokens_seen": 1665335296 + }, + { + "epoch": 0.5, + "learning_rate": 0.0004003691073219659, + "loss": 1.2685, + "theoretical_loss": 3.5125904827774006, + "tokens_seen": 1665400832 + }, + { + "epoch": 0.5, + "learning_rate": 0.0004003530591775326, + "loss": 1.4755, + "theoretical_loss": 3.5125786357963884, + "tokens_seen": 1665466368 + }, + { + "epoch": 0.5, + "learning_rate": 0.0004003370110330993, + "loss": 1.344, + "theoretical_loss": 3.5125667894120687, + "tokens_seen": 1665531904 + }, + { + "epoch": 0.5, + "learning_rate": 0.000400320962888666, + "loss": 1.3553, + "theoretical_loss": 3.5125549436243886, + "tokens_seen": 1665597440 + }, + { + "epoch": 0.5, + "learning_rate": 0.0004003049147442327, + "loss": 1.3987, + "theoretical_loss": 3.512543098433294, + "tokens_seen": 1665662976 + }, + { + "epoch": 0.5, + "learning_rate": 0.0004002888665997994, + "loss": 1.4299, + "theoretical_loss": 3.5125312538387314, + "tokens_seen": 1665728512 + }, + { + "epoch": 0.5, + "learning_rate": 0.0004002728184553661, + "loss": 1.414, + "theoretical_loss": 3.512519409840648, + "tokens_seen": 1665794048 + }, + { + "epoch": 0.5, + "learning_rate": 0.0004002567703109328, + "loss": 1.2915, + "theoretical_loss": 3.5125075664389893, + "tokens_seen": 1665859584 + }, + { + "epoch": 0.5, + "learning_rate": 0.00040024072216649946, + "loss": 1.3196, + "theoretical_loss": 3.5124957236337027, + "tokens_seen": 1665925120 + }, + { + "epoch": 0.5, + "learning_rate": 0.0004002246740220662, + "loss": 1.3511, + "theoretical_loss": 3.512483881424734, + "tokens_seen": 1665990656 + }, + { + "epoch": 0.5, + "learning_rate": 0.00040020862587763286, + "loss": 1.2807, + "theoretical_loss": 3.5124720398120304, + "tokens_seen": 1666056192 + }, + { + "epoch": 0.5, + "learning_rate": 0.0004001925777331996, + "loss": 1.3023, + "theoretical_loss": 3.5124601987955377, + "tokens_seen": 1666121728 + }, + { + "epoch": 0.5, + "learning_rate": 0.00040017652958876626, + "loss": 1.3843, + "theoretical_loss": 3.5124483583752033, + "tokens_seen": 1666187264 + }, + { + "epoch": 0.5, + "learning_rate": 0.000400160481444333, + "loss": 1.3572, + "theoretical_loss": 3.5124365185509725, + "tokens_seen": 1666252800 + }, + { + "epoch": 0.5, + "learning_rate": 0.00040014443329989966, + "loss": 1.3862, + "theoretical_loss": 3.512424679322793, + "tokens_seen": 1666318336 + }, + { + "epoch": 0.5, + "learning_rate": 0.0004001283851554664, + "loss": 1.3896, + "theoretical_loss": 3.5124128406906108, + "tokens_seen": 1666383872 + }, + { + "epoch": 0.5, + "learning_rate": 0.00040011233701103306, + "loss": 1.3067, + "theoretical_loss": 3.512401002654373, + "tokens_seen": 1666449408 + }, + { + "epoch": 0.51, + "learning_rate": 0.0004000962888665998, + "loss": 1.3861, + "theoretical_loss": 3.5123891652140253, + "tokens_seen": 1666514944 + }, + { + "epoch": 0.51, + "learning_rate": 0.00040008024072216646, + "loss": 1.4142, + "theoretical_loss": 3.5123773283695146, + "tokens_seen": 1666580480 + }, + { + "epoch": 0.51, + "learning_rate": 0.0004000641925777332, + "loss": 1.4625, + "theoretical_loss": 3.512365492120788, + "tokens_seen": 1666646016 + }, + { + "epoch": 0.51, + "learning_rate": 0.00040004814443329986, + "loss": 1.3305, + "theoretical_loss": 3.5123536564677913, + "tokens_seen": 1666711552 + }, + { + "epoch": 0.51, + "learning_rate": 0.0004000320962888666, + "loss": 1.4369, + "theoretical_loss": 3.5123418214104722, + "tokens_seen": 1666777088 + }, + { + "epoch": 0.51, + "learning_rate": 0.00040001604814443326, + "loss": 1.4377, + "theoretical_loss": 3.5123299869487763, + "tokens_seen": 1666842624 + }, + { + "epoch": 0.51, + "learning_rate": 0.0004, + "loss": 1.4966, + "theoretical_loss": 3.51231815308265, + "tokens_seen": 1666908160 + }, + { + "epoch": 0.51, + "learning_rate": 0.0003999839518555667, + "loss": 1.3535, + "theoretical_loss": 3.5123063198120406, + "tokens_seen": 1666973696 + }, + { + "epoch": 0.51, + "learning_rate": 0.0003999679037111334, + "loss": 1.3553, + "theoretical_loss": 3.5122944871368946, + "tokens_seen": 1667039232 + }, + { + "epoch": 0.51, + "learning_rate": 0.0003999518555667001, + "loss": 1.2364, + "theoretical_loss": 3.5122826550571586, + "tokens_seen": 1667104768 + }, + { + "epoch": 0.51, + "learning_rate": 0.0003999358074222668, + "loss": 1.4211, + "theoretical_loss": 3.5122708235727798, + "tokens_seen": 1667170304 + }, + { + "epoch": 0.51, + "learning_rate": 0.0003999197592778335, + "loss": 1.3604, + "theoretical_loss": 3.5122589926837033, + "tokens_seen": 1667235840 + }, + { + "epoch": 0.51, + "learning_rate": 0.0003999037111334002, + "loss": 1.3881, + "theoretical_loss": 3.512247162389877, + "tokens_seen": 1667301376 + }, + { + "epoch": 0.51, + "learning_rate": 0.0003998876629889669, + "loss": 1.2633, + "theoretical_loss": 3.5122353326912474, + "tokens_seen": 1667366912 + }, + { + "epoch": 0.51, + "learning_rate": 0.0003998716148445336, + "loss": 1.2432, + "theoretical_loss": 3.512223503587761, + "tokens_seen": 1667432448 + }, + { + "epoch": 0.51, + "learning_rate": 0.0003998555667001003, + "loss": 1.4743, + "theoretical_loss": 3.512211675079364, + "tokens_seen": 1667497984 + }, + { + "epoch": 0.51, + "learning_rate": 0.000399839518555667, + "loss": 1.2675, + "theoretical_loss": 3.512199847166004, + "tokens_seen": 1667563520 + }, + { + "epoch": 0.51, + "learning_rate": 0.0003998234704112337, + "loss": 1.364, + "theoretical_loss": 3.512188019847627, + "tokens_seen": 1667629056 + }, + { + "epoch": 0.51, + "learning_rate": 0.0003998074222668004, + "loss": 1.4467, + "theoretical_loss": 3.5121761931241804, + "tokens_seen": 1667694592 + }, + { + "epoch": 0.51, + "learning_rate": 0.00039979137412236713, + "loss": 1.397, + "theoretical_loss": 3.5121643669956097, + "tokens_seen": 1667760128 + }, + { + "epoch": 0.51, + "learning_rate": 0.00039977532597793383, + "loss": 1.3833, + "theoretical_loss": 3.512152541461863, + "tokens_seen": 1667825664 + }, + { + "epoch": 0.51, + "objective/train/docs_used": 609331, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.47962486743927, + "objective/train/theoretical_loss": 3.512140716522886, + "objective/train/tokens_used": 1688351200, + "theoretical_loss": 3.512140716522886, + "tokens_seen": 1667891200 + }, + { + "epoch": 0.51, + "learning_rate": 0.00039975927783350053, + "loss": 1.421, + "theoretical_loss": 3.512140716522886, + "tokens_seen": 1667891200 + }, + { + "epoch": 0.51, + "learning_rate": 0.00039974322968906723, + "loss": 1.3135, + "theoretical_loss": 3.512128892178626, + "tokens_seen": 1667956736 + }, + { + "epoch": 0.51, + "learning_rate": 0.00039972718154463393, + "loss": 1.3404, + "theoretical_loss": 3.5121170684290286, + "tokens_seen": 1668022272 + }, + { + "epoch": 0.51, + "learning_rate": 0.00039971113340020063, + "loss": 1.2828, + "theoretical_loss": 3.5121052452740424, + "tokens_seen": 1668087808 + }, + { + "epoch": 0.51, + "learning_rate": 0.00039969508525576733, + "loss": 1.3391, + "theoretical_loss": 3.5120934227136127, + "tokens_seen": 1668153344 + }, + { + "epoch": 0.51, + "learning_rate": 0.00039967903711133403, + "loss": 1.1786, + "theoretical_loss": 3.512081600747687, + "tokens_seen": 1668218880 + }, + { + "epoch": 0.51, + "learning_rate": 0.00039966298896690073, + "loss": 1.3325, + "theoretical_loss": 3.5120697793762115, + "tokens_seen": 1668284416 + }, + { + "epoch": 0.51, + "learning_rate": 0.00039964694082246743, + "loss": 1.2982, + "theoretical_loss": 3.512057958599133, + "tokens_seen": 1668349952 + }, + { + "epoch": 0.51, + "learning_rate": 0.00039963089267803413, + "loss": 1.3245, + "theoretical_loss": 3.512046138416399, + "tokens_seen": 1668415488 + }, + { + "epoch": 0.51, + "learning_rate": 0.00039961484453360083, + "loss": 1.3488, + "theoretical_loss": 3.512034318827955, + "tokens_seen": 1668481024 + }, + { + "epoch": 0.51, + "learning_rate": 0.0003995987963891675, + "loss": 1.4199, + "theoretical_loss": 3.512022499833749, + "tokens_seen": 1668546560 + }, + { + "epoch": 0.51, + "learning_rate": 0.0003995827482447342, + "loss": 1.2581, + "theoretical_loss": 3.5120106814337273, + "tokens_seen": 1668612096 + }, + { + "epoch": 0.51, + "learning_rate": 0.0003995667001003009, + "loss": 1.3859, + "theoretical_loss": 3.511998863627837, + "tokens_seen": 1668677632 + }, + { + "epoch": 0.51, + "learning_rate": 0.0003995506519558676, + "loss": 1.3777, + "theoretical_loss": 3.511987046416024, + "tokens_seen": 1668743168 + }, + { + "epoch": 0.51, + "learning_rate": 0.0003995346038114343, + "loss": 1.2616, + "theoretical_loss": 3.511975229798236, + "tokens_seen": 1668808704 + }, + { + "epoch": 0.51, + "learning_rate": 0.000399518555667001, + "loss": 1.3382, + "theoretical_loss": 3.5119634137744193, + "tokens_seen": 1668874240 + }, + { + "epoch": 0.51, + "learning_rate": 0.0003995025075225677, + "loss": 1.2223, + "theoretical_loss": 3.5119515983445213, + "tokens_seen": 1668939776 + }, + { + "epoch": 0.51, + "learning_rate": 0.0003994864593781344, + "loss": 1.3987, + "theoretical_loss": 3.5119397835084882, + "tokens_seen": 1669005312 + }, + { + "epoch": 0.51, + "learning_rate": 0.0003994704112337011, + "loss": 1.457, + "theoretical_loss": 3.5119279692662673, + "tokens_seen": 1669070848 + }, + { + "epoch": 0.51, + "learning_rate": 0.0003994543630892678, + "loss": 1.3445, + "theoretical_loss": 3.5119161556178047, + "tokens_seen": 1669136384 + }, + { + "epoch": 0.51, + "learning_rate": 0.0003994383149448345, + "loss": 1.3128, + "theoretical_loss": 3.511904342563048, + "tokens_seen": 1669201920 + }, + { + "epoch": 0.51, + "learning_rate": 0.0003994222668004012, + "loss": 1.4001, + "theoretical_loss": 3.511892530101944, + "tokens_seen": 1669267456 + }, + { + "epoch": 0.51, + "learning_rate": 0.0003994062186559679, + "loss": 1.2802, + "theoretical_loss": 3.5118807182344396, + "tokens_seen": 1669332992 + }, + { + "epoch": 0.51, + "learning_rate": 0.00039939017051153464, + "loss": 1.5129, + "theoretical_loss": 3.511868906960481, + "tokens_seen": 1669398528 + }, + { + "epoch": 0.51, + "learning_rate": 0.00039937412236710134, + "loss": 1.4606, + "theoretical_loss": 3.511857096280016, + "tokens_seen": 1669464064 + }, + { + "epoch": 0.51, + "learning_rate": 0.00039935807422266804, + "loss": 1.4201, + "theoretical_loss": 3.5118452861929907, + "tokens_seen": 1669529600 + }, + { + "epoch": 0.51, + "learning_rate": 0.00039934202607823475, + "loss": 1.3632, + "theoretical_loss": 3.5118334766993526, + "tokens_seen": 1669595136 + }, + { + "epoch": 0.51, + "learning_rate": 0.00039932597793380145, + "loss": 1.327, + "theoretical_loss": 3.5118216677990484, + "tokens_seen": 1669660672 + }, + { + "epoch": 0.51, + "learning_rate": 0.00039930992978936815, + "loss": 1.3169, + "theoretical_loss": 3.5118098594920246, + "tokens_seen": 1669726208 + }, + { + "epoch": 0.51, + "learning_rate": 0.00039929388164493485, + "loss": 1.4296, + "theoretical_loss": 3.5117980517782286, + "tokens_seen": 1669791744 + }, + { + "epoch": 0.51, + "learning_rate": 0.00039927783350050155, + "loss": 1.4207, + "theoretical_loss": 3.511786244657607, + "tokens_seen": 1669857280 + }, + { + "epoch": 0.51, + "learning_rate": 0.00039926178535606825, + "loss": 1.3883, + "theoretical_loss": 3.5117744381301064, + "tokens_seen": 1669922816 + }, + { + "epoch": 0.51, + "learning_rate": 0.00039924573721163495, + "loss": 1.1972, + "theoretical_loss": 3.5117626321956745, + "tokens_seen": 1669988352 + }, + { + "epoch": 0.51, + "learning_rate": 0.00039922968906720165, + "loss": 1.3017, + "theoretical_loss": 3.5117508268542585, + "tokens_seen": 1670053888 + }, + { + "epoch": 0.51, + "learning_rate": 0.00039921364092276835, + "loss": 1.4104, + "theoretical_loss": 3.5117390221058042, + "tokens_seen": 1670119424 + }, + { + "epoch": 0.51, + "learning_rate": 0.00039919759277833505, + "loss": 1.4479, + "theoretical_loss": 3.511727217950259, + "tokens_seen": 1670184960 + }, + { + "epoch": 0.51, + "learning_rate": 0.00039918154463390175, + "loss": 1.3736, + "theoretical_loss": 3.5117154143875706, + "tokens_seen": 1670250496 + }, + { + "epoch": 0.51, + "learning_rate": 0.00039916549648946845, + "loss": 1.321, + "theoretical_loss": 3.511703611417685, + "tokens_seen": 1670316032 + }, + { + "epoch": 0.51, + "learning_rate": 0.00039914944834503515, + "loss": 1.434, + "theoretical_loss": 3.5116918090405496, + "tokens_seen": 1670381568 + }, + { + "epoch": 0.51, + "learning_rate": 0.00039913340020060185, + "loss": 1.3511, + "theoretical_loss": 3.5116800072561114, + "tokens_seen": 1670447104 + }, + { + "epoch": 0.51, + "learning_rate": 0.00039911735205616855, + "loss": 1.4066, + "theoretical_loss": 3.511668206064317, + "tokens_seen": 1670512640 + }, + { + "epoch": 0.51, + "learning_rate": 0.00039910130391173525, + "loss": 1.3081, + "theoretical_loss": 3.5116564054651134, + "tokens_seen": 1670578176 + }, + { + "epoch": 0.51, + "learning_rate": 0.00039908525576730196, + "loss": 1.353, + "theoretical_loss": 3.5116446054584483, + "tokens_seen": 1670643712 + }, + { + "epoch": 0.51, + "learning_rate": 0.00039906920762286866, + "loss": 1.4046, + "theoretical_loss": 3.5116328060442683, + "tokens_seen": 1670709248 + }, + { + "epoch": 0.51, + "learning_rate": 0.00039905315947843536, + "loss": 1.3013, + "theoretical_loss": 3.5116210072225202, + "tokens_seen": 1670774784 + }, + { + "epoch": 0.51, + "learning_rate": 0.00039903711133400206, + "loss": 1.3526, + "theoretical_loss": 3.511609208993151, + "tokens_seen": 1670840320 + }, + { + "epoch": 0.51, + "learning_rate": 0.00039902106318956876, + "loss": 1.4837, + "theoretical_loss": 3.5115974113561084, + "tokens_seen": 1670905856 + }, + { + "epoch": 0.51, + "learning_rate": 0.00039900501504513546, + "loss": 1.3886, + "theoretical_loss": 3.511585614311339, + "tokens_seen": 1670971392 + }, + { + "epoch": 0.51, + "learning_rate": 0.00039898896690070216, + "loss": 1.5495, + "theoretical_loss": 3.5115738178587894, + "tokens_seen": 1671036928 + }, + { + "epoch": 0.51, + "learning_rate": 0.0003989729187562688, + "loss": 1.5464, + "theoretical_loss": 3.511562021998407, + "tokens_seen": 1671102464 + }, + { + "epoch": 0.51, + "objective/train/docs_used": 610761, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.3039546012878418, + "objective/train/theoretical_loss": 3.5115502267301393, + "objective/train/tokens_used": 1691628000, + "theoretical_loss": 3.5115502267301393, + "tokens_seen": 1671168000 + }, + { + "epoch": 0.51, + "learning_rate": 0.0003989568706118355, + "loss": 1.3194, + "theoretical_loss": 3.5115502267301393, + "tokens_seen": 1671168000 + }, + { + "epoch": 0.51, + "learning_rate": 0.0003989408224674022, + "loss": 1.3813, + "theoretical_loss": 3.5115384320539325, + "tokens_seen": 1671233536 + }, + { + "epoch": 0.51, + "learning_rate": 0.0003989247743229689, + "loss": 1.2763, + "theoretical_loss": 3.5115266379697343, + "tokens_seen": 1671299072 + }, + { + "epoch": 0.51, + "learning_rate": 0.0003989087261785356, + "loss": 1.2935, + "theoretical_loss": 3.5115148444774915, + "tokens_seen": 1671364608 + }, + { + "epoch": 0.51, + "learning_rate": 0.0003988926780341023, + "loss": 1.1678, + "theoretical_loss": 3.5115030515771517, + "tokens_seen": 1671430144 + }, + { + "epoch": 0.51, + "learning_rate": 0.000398876629889669, + "loss": 1.4685, + "theoretical_loss": 3.511491259268661, + "tokens_seen": 1671495680 + }, + { + "epoch": 0.51, + "learning_rate": 0.0003988605817452357, + "loss": 1.2667, + "theoretical_loss": 3.5114794675519674, + "tokens_seen": 1671561216 + }, + { + "epoch": 0.51, + "learning_rate": 0.0003988445336008024, + "loss": 1.2083, + "theoretical_loss": 3.5114676764270176, + "tokens_seen": 1671626752 + }, + { + "epoch": 0.51, + "learning_rate": 0.0003988284854563691, + "loss": 1.4056, + "theoretical_loss": 3.5114558858937586, + "tokens_seen": 1671692288 + }, + { + "epoch": 0.51, + "learning_rate": 0.0003988124373119358, + "loss": 1.3861, + "theoretical_loss": 3.511444095952138, + "tokens_seen": 1671757824 + }, + { + "epoch": 0.51, + "learning_rate": 0.0003987963891675025, + "loss": 1.3136, + "theoretical_loss": 3.511432306602102, + "tokens_seen": 1671823360 + }, + { + "epoch": 0.51, + "learning_rate": 0.0003987803410230692, + "loss": 1.3399, + "theoretical_loss": 3.5114205178435984, + "tokens_seen": 1671888896 + }, + { + "epoch": 0.51, + "learning_rate": 0.0003987642928786359, + "loss": 1.4025, + "theoretical_loss": 3.511408729676575, + "tokens_seen": 1671954432 + }, + { + "epoch": 0.51, + "learning_rate": 0.0003987482447342026, + "loss": 1.3774, + "theoretical_loss": 3.5113969421009776, + "tokens_seen": 1672019968 + }, + { + "epoch": 0.51, + "learning_rate": 0.0003987321965897693, + "loss": 1.3319, + "theoretical_loss": 3.511385155116754, + "tokens_seen": 1672085504 + }, + { + "epoch": 0.51, + "learning_rate": 0.000398716148445336, + "loss": 1.2904, + "theoretical_loss": 3.5113733687238513, + "tokens_seen": 1672151040 + }, + { + "epoch": 0.51, + "learning_rate": 0.0003987001003009027, + "loss": 1.2689, + "theoretical_loss": 3.511361582922217, + "tokens_seen": 1672216576 + }, + { + "epoch": 0.51, + "learning_rate": 0.0003986840521564694, + "loss": 1.2686, + "theoretical_loss": 3.5113497977117976, + "tokens_seen": 1672282112 + }, + { + "epoch": 0.51, + "learning_rate": 0.0003986680040120361, + "loss": 1.3835, + "theoretical_loss": 3.5113380130925407, + "tokens_seen": 1672347648 + }, + { + "epoch": 0.51, + "learning_rate": 0.0003986519558676028, + "loss": 1.3103, + "theoretical_loss": 3.5113262290643936, + "tokens_seen": 1672413184 + }, + { + "epoch": 0.51, + "learning_rate": 0.0003986359077231695, + "loss": 1.2667, + "theoretical_loss": 3.5113144456273027, + "tokens_seen": 1672478720 + }, + { + "epoch": 0.51, + "learning_rate": 0.0003986198595787362, + "loss": 1.3851, + "theoretical_loss": 3.5113026627812163, + "tokens_seen": 1672544256 + }, + { + "epoch": 0.51, + "learning_rate": 0.0003986038114343029, + "loss": 1.266, + "theoretical_loss": 3.5112908805260803, + "tokens_seen": 1672609792 + }, + { + "epoch": 0.51, + "learning_rate": 0.0003985877632898696, + "loss": 1.3568, + "theoretical_loss": 3.5112790988618436, + "tokens_seen": 1672675328 + }, + { + "epoch": 0.51, + "learning_rate": 0.0003985717151454363, + "loss": 1.3049, + "theoretical_loss": 3.511267317788452, + "tokens_seen": 1672740864 + }, + { + "epoch": 0.51, + "learning_rate": 0.000398555667001003, + "loss": 1.3848, + "theoretical_loss": 3.5112555373058534, + "tokens_seen": 1672806400 + }, + { + "epoch": 0.51, + "learning_rate": 0.0003985396188565697, + "loss": 1.3379, + "theoretical_loss": 3.511243757413995, + "tokens_seen": 1672871936 + }, + { + "epoch": 0.51, + "learning_rate": 0.0003985235707121364, + "loss": 1.4074, + "theoretical_loss": 3.511231978112823, + "tokens_seen": 1672937472 + }, + { + "epoch": 0.51, + "learning_rate": 0.0003985075225677031, + "loss": 1.2809, + "theoretical_loss": 3.5112201994022865, + "tokens_seen": 1673003008 + }, + { + "epoch": 0.51, + "learning_rate": 0.0003984914744232698, + "loss": 1.3609, + "theoretical_loss": 3.5112084212823316, + "tokens_seen": 1673068544 + }, + { + "epoch": 0.51, + "learning_rate": 0.0003984754262788365, + "loss": 1.1703, + "theoretical_loss": 3.5111966437529056, + "tokens_seen": 1673134080 + }, + { + "epoch": 0.51, + "learning_rate": 0.0003984593781344032, + "loss": 1.3976, + "theoretical_loss": 3.511184866813956, + "tokens_seen": 1673199616 + }, + { + "epoch": 0.51, + "learning_rate": 0.0003984433299899699, + "loss": 1.4227, + "theoretical_loss": 3.5111730904654297, + "tokens_seen": 1673265152 + }, + { + "epoch": 0.51, + "learning_rate": 0.00039842728184553663, + "loss": 1.3142, + "theoretical_loss": 3.5111613147072744, + "tokens_seen": 1673330688 + }, + { + "epoch": 0.51, + "learning_rate": 0.00039841123370110333, + "loss": 1.383, + "theoretical_loss": 3.511149539539437, + "tokens_seen": 1673396224 + }, + { + "epoch": 0.51, + "learning_rate": 0.00039839518555667003, + "loss": 1.4503, + "theoretical_loss": 3.5111377649618656, + "tokens_seen": 1673461760 + }, + { + "epoch": 0.51, + "learning_rate": 0.00039837913741223673, + "loss": 1.3652, + "theoretical_loss": 3.5111259909745067, + "tokens_seen": 1673527296 + }, + { + "epoch": 0.51, + "learning_rate": 0.00039836308926780343, + "loss": 1.3751, + "theoretical_loss": 3.5111142175773074, + "tokens_seen": 1673592832 + }, + { + "epoch": 0.51, + "learning_rate": 0.00039834704112337013, + "loss": 1.356, + "theoretical_loss": 3.5111024447702155, + "tokens_seen": 1673658368 + }, + { + "epoch": 0.51, + "learning_rate": 0.00039833099297893683, + "loss": 1.3351, + "theoretical_loss": 3.5110906725531787, + "tokens_seen": 1673723904 + }, + { + "epoch": 0.51, + "learning_rate": 0.00039831494483450353, + "loss": 1.3427, + "theoretical_loss": 3.5110789009261434, + "tokens_seen": 1673789440 + }, + { + "epoch": 0.51, + "learning_rate": 0.00039829889669007023, + "loss": 1.403, + "theoretical_loss": 3.5110671298890574, + "tokens_seen": 1673854976 + }, + { + "epoch": 0.51, + "learning_rate": 0.00039828284854563693, + "loss": 1.2522, + "theoretical_loss": 3.511055359441868, + "tokens_seen": 1673920512 + }, + { + "epoch": 0.51, + "learning_rate": 0.00039826680040120363, + "loss": 1.3446, + "theoretical_loss": 3.5110435895845233, + "tokens_seen": 1673986048 + }, + { + "epoch": 0.51, + "learning_rate": 0.00039825075225677033, + "loss": 1.393, + "theoretical_loss": 3.511031820316969, + "tokens_seen": 1674051584 + }, + { + "epoch": 0.51, + "learning_rate": 0.00039823470411233703, + "loss": 1.4181, + "theoretical_loss": 3.511020051639154, + "tokens_seen": 1674117120 + }, + { + "epoch": 0.51, + "learning_rate": 0.00039821865596790374, + "loss": 1.3111, + "theoretical_loss": 3.511008283551025, + "tokens_seen": 1674182656 + }, + { + "epoch": 0.51, + "learning_rate": 0.0003982026078234704, + "loss": 1.3344, + "theoretical_loss": 3.510996516052529, + "tokens_seen": 1674248192 + }, + { + "epoch": 0.51, + "learning_rate": 0.00039818655967903714, + "loss": 1.3617, + "theoretical_loss": 3.5109847491436144, + "tokens_seen": 1674313728 + }, + { + "epoch": 0.51, + "learning_rate": 0.00039817051153460384, + "loss": 1.3481, + "theoretical_loss": 3.5109729828242275, + "tokens_seen": 1674379264 + }, + { + "epoch": 0.51, + "objective/train/docs_used": 611357, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.331943154335022, + "objective/train/theoretical_loss": 3.5109612170943163, + "objective/train/tokens_used": 1694904800, + "theoretical_loss": 3.5109612170943163, + "tokens_seen": 1674444800 + }, + { + "epoch": 0.51, + "learning_rate": 0.00039815446339017054, + "loss": 1.277, + "theoretical_loss": 3.5109612170943163, + "tokens_seen": 1674444800 + }, + { + "epoch": 0.51, + "learning_rate": 0.00039813841524573724, + "loss": 1.389, + "theoretical_loss": 3.5109494519538282, + "tokens_seen": 1674510336 + }, + { + "epoch": 0.51, + "learning_rate": 0.00039812236710130394, + "loss": 1.1816, + "theoretical_loss": 3.5109376874027105, + "tokens_seen": 1674575872 + }, + { + "epoch": 0.51, + "learning_rate": 0.00039810631895687064, + "loss": 1.3668, + "theoretical_loss": 3.5109259234409107, + "tokens_seen": 1674641408 + }, + { + "epoch": 0.51, + "learning_rate": 0.00039809027081243734, + "loss": 1.3274, + "theoretical_loss": 3.510914160068376, + "tokens_seen": 1674706944 + }, + { + "epoch": 0.51, + "learning_rate": 0.00039807422266800404, + "loss": 1.2967, + "theoretical_loss": 3.5109023972850535, + "tokens_seen": 1674772480 + }, + { + "epoch": 0.51, + "learning_rate": 0.00039805817452357074, + "loss": 1.3277, + "theoretical_loss": 3.5108906350908917, + "tokens_seen": 1674838016 + }, + { + "epoch": 0.51, + "learning_rate": 0.00039804212637913744, + "loss": 1.3905, + "theoretical_loss": 3.510878873485837, + "tokens_seen": 1674903552 + }, + { + "epoch": 0.51, + "learning_rate": 0.00039802607823470414, + "loss": 1.2845, + "theoretical_loss": 3.510867112469837, + "tokens_seen": 1674969088 + }, + { + "epoch": 0.51, + "learning_rate": 0.00039801003009027084, + "loss": 1.307, + "theoretical_loss": 3.51085535204284, + "tokens_seen": 1675034624 + }, + { + "epoch": 0.51, + "learning_rate": 0.00039799398194583754, + "loss": 1.3703, + "theoretical_loss": 3.5108435922047927, + "tokens_seen": 1675100160 + }, + { + "epoch": 0.51, + "learning_rate": 0.00039797793380140424, + "loss": 1.2779, + "theoretical_loss": 3.5108318329556427, + "tokens_seen": 1675165696 + }, + { + "epoch": 0.51, + "learning_rate": 0.00039796188565697095, + "loss": 1.3883, + "theoretical_loss": 3.5108200742953377, + "tokens_seen": 1675231232 + }, + { + "epoch": 0.51, + "learning_rate": 0.00039794583751253765, + "loss": 1.342, + "theoretical_loss": 3.510808316223825, + "tokens_seen": 1675296768 + }, + { + "epoch": 0.51, + "learning_rate": 0.00039792978936810435, + "loss": 1.3234, + "theoretical_loss": 3.5107965587410517, + "tokens_seen": 1675362304 + }, + { + "epoch": 0.51, + "learning_rate": 0.00039791374122367105, + "loss": 1.3089, + "theoretical_loss": 3.510784801846966, + "tokens_seen": 1675427840 + }, + { + "epoch": 0.51, + "learning_rate": 0.00039789769307923775, + "loss": 1.3406, + "theoretical_loss": 3.5107730455415145, + "tokens_seen": 1675493376 + }, + { + "epoch": 0.51, + "learning_rate": 0.00039788164493480445, + "loss": 1.3611, + "theoretical_loss": 3.5107612898246456, + "tokens_seen": 1675558912 + }, + { + "epoch": 0.51, + "learning_rate": 0.00039786559679037115, + "loss": 1.2828, + "theoretical_loss": 3.5107495346963065, + "tokens_seen": 1675624448 + }, + { + "epoch": 0.51, + "learning_rate": 0.00039784954864593785, + "loss": 1.3931, + "theoretical_loss": 3.510737780156445, + "tokens_seen": 1675689984 + }, + { + "epoch": 0.51, + "learning_rate": 0.00039783350050150455, + "loss": 1.2252, + "theoretical_loss": 3.510726026205008, + "tokens_seen": 1675755520 + }, + { + "epoch": 0.51, + "learning_rate": 0.00039781745235707125, + "loss": 1.302, + "theoretical_loss": 3.5107142728419434, + "tokens_seen": 1675821056 + }, + { + "epoch": 0.51, + "learning_rate": 0.00039780140421263795, + "loss": 1.2943, + "theoretical_loss": 3.5107025200671984, + "tokens_seen": 1675886592 + }, + { + "epoch": 0.51, + "learning_rate": 0.00039778535606820465, + "loss": 1.2366, + "theoretical_loss": 3.510690767880721, + "tokens_seen": 1675952128 + }, + { + "epoch": 0.51, + "learning_rate": 0.00039776930792377135, + "loss": 1.3751, + "theoretical_loss": 3.510679016282459, + "tokens_seen": 1676017664 + }, + { + "epoch": 0.51, + "learning_rate": 0.00039775325977933805, + "loss": 1.4062, + "theoretical_loss": 3.5106672652723594, + "tokens_seen": 1676083200 + }, + { + "epoch": 0.51, + "learning_rate": 0.00039773721163490475, + "loss": 1.343, + "theoretical_loss": 3.51065551485037, + "tokens_seen": 1676148736 + }, + { + "epoch": 0.51, + "learning_rate": 0.00039772116349047146, + "loss": 1.3757, + "theoretical_loss": 3.5106437650164377, + "tokens_seen": 1676214272 + }, + { + "epoch": 0.51, + "learning_rate": 0.00039770511534603816, + "loss": 1.3268, + "theoretical_loss": 3.5106320157705113, + "tokens_seen": 1676279808 + }, + { + "epoch": 0.51, + "learning_rate": 0.00039768906720160486, + "loss": 1.3566, + "theoretical_loss": 3.5106202671125377, + "tokens_seen": 1676345344 + }, + { + "epoch": 0.51, + "learning_rate": 0.00039767301905717156, + "loss": 1.2462, + "theoretical_loss": 3.5106085190424645, + "tokens_seen": 1676410880 + }, + { + "epoch": 0.51, + "learning_rate": 0.00039765697091273826, + "loss": 1.3135, + "theoretical_loss": 3.5105967715602393, + "tokens_seen": 1676476416 + }, + { + "epoch": 0.51, + "learning_rate": 0.00039764092276830496, + "loss": 1.4127, + "theoretical_loss": 3.5105850246658097, + "tokens_seen": 1676541952 + }, + { + "epoch": 0.51, + "learning_rate": 0.00039762487462387166, + "loss": 1.4574, + "theoretical_loss": 3.5105732783591237, + "tokens_seen": 1676607488 + }, + { + "epoch": 0.51, + "learning_rate": 0.00039760882647943836, + "loss": 1.415, + "theoretical_loss": 3.5105615326401285, + "tokens_seen": 1676673024 + }, + { + "epoch": 0.51, + "learning_rate": 0.00039759277833500506, + "loss": 1.301, + "theoretical_loss": 3.5105497875087717, + "tokens_seen": 1676738560 + }, + { + "epoch": 0.51, + "learning_rate": 0.0003975767301905717, + "loss": 1.2019, + "theoretical_loss": 3.510538042965001, + "tokens_seen": 1676804096 + }, + { + "epoch": 0.51, + "learning_rate": 0.0003975606820461384, + "loss": 1.3662, + "theoretical_loss": 3.510526299008764, + "tokens_seen": 1676869632 + }, + { + "epoch": 0.51, + "learning_rate": 0.0003975446339017051, + "loss": 1.294, + "theoretical_loss": 3.5105145556400092, + "tokens_seen": 1676935168 + }, + { + "epoch": 0.51, + "learning_rate": 0.0003975285857572718, + "loss": 1.3424, + "theoretical_loss": 3.5105028128586824, + "tokens_seen": 1677000704 + }, + { + "epoch": 0.51, + "learning_rate": 0.0003975125376128385, + "loss": 1.2919, + "theoretical_loss": 3.5104910706647336, + "tokens_seen": 1677066240 + }, + { + "epoch": 0.51, + "learning_rate": 0.0003974964894684052, + "loss": 1.3234, + "theoretical_loss": 3.5104793290581084, + "tokens_seen": 1677131776 + }, + { + "epoch": 0.51, + "learning_rate": 0.0003974804413239719, + "loss": 1.3757, + "theoretical_loss": 3.5104675880387552, + "tokens_seen": 1677197312 + }, + { + "epoch": 0.51, + "learning_rate": 0.0003974643931795386, + "loss": 1.3296, + "theoretical_loss": 3.510455847606622, + "tokens_seen": 1677262848 + }, + { + "epoch": 0.51, + "learning_rate": 0.0003974483450351053, + "loss": 1.2333, + "theoretical_loss": 3.5104441077616566, + "tokens_seen": 1677328384 + }, + { + "epoch": 0.51, + "learning_rate": 0.000397432296890672, + "loss": 1.325, + "theoretical_loss": 3.510432368503806, + "tokens_seen": 1677393920 + }, + { + "epoch": 0.51, + "learning_rate": 0.0003974162487462387, + "loss": 1.2721, + "theoretical_loss": 3.5104206298330185, + "tokens_seen": 1677459456 + }, + { + "epoch": 0.51, + "learning_rate": 0.0003974002006018054, + "loss": 1.3667, + "theoretical_loss": 3.5104088917492415, + "tokens_seen": 1677524992 + }, + { + "epoch": 0.51, + "learning_rate": 0.0003973841524573721, + "loss": 1.359, + "theoretical_loss": 3.510397154252423, + "tokens_seen": 1677590528 + }, + { + "epoch": 0.51, + "learning_rate": 0.0003973681043129388, + "loss": 1.2723, + "theoretical_loss": 3.51038541734251, + "tokens_seen": 1677656064 + }, + { + "epoch": 0.51, + "objective/train/docs_used": 612648, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.26918363571167, + "objective/train/theoretical_loss": 3.5103736810194515, + "objective/train/tokens_used": 1698181600, + "theoretical_loss": 3.5103736810194515, + "tokens_seen": 1677721600 + }, + { + "epoch": 0.51, + "learning_rate": 0.0003973520561685055, + "loss": 1.3025, + "theoretical_loss": 3.5103736810194515, + "tokens_seen": 1677721600 + }, + { + "epoch": 0.51, + "learning_rate": 0.0003973360080240722, + "loss": 1.316, + "theoretical_loss": 3.5103619452831936, + "tokens_seen": 1677787136 + }, + { + "epoch": 0.51, + "learning_rate": 0.0003973199598796389, + "loss": 1.3475, + "theoretical_loss": 3.5103502101336854, + "tokens_seen": 1677852672 + }, + { + "epoch": 0.51, + "learning_rate": 0.0003973039117352056, + "loss": 1.41, + "theoretical_loss": 3.510338475570874, + "tokens_seen": 1677918208 + }, + { + "epoch": 0.51, + "learning_rate": 0.0003972878635907723, + "loss": 1.3475, + "theoretical_loss": 3.510326741594708, + "tokens_seen": 1677983744 + }, + { + "epoch": 0.51, + "learning_rate": 0.000397271815446339, + "loss": 1.3789, + "theoretical_loss": 3.5103150082051338, + "tokens_seen": 1678049280 + }, + { + "epoch": 0.51, + "learning_rate": 0.0003972557673019057, + "loss": 1.3718, + "theoretical_loss": 3.5103032754021, + "tokens_seen": 1678114816 + }, + { + "epoch": 0.51, + "learning_rate": 0.0003972397191574724, + "loss": 1.4161, + "theoretical_loss": 3.5102915431855544, + "tokens_seen": 1678180352 + }, + { + "epoch": 0.51, + "learning_rate": 0.0003972236710130391, + "loss": 1.1968, + "theoretical_loss": 3.5102798115554448, + "tokens_seen": 1678245888 + }, + { + "epoch": 0.51, + "learning_rate": 0.0003972076228686058, + "loss": 1.3167, + "theoretical_loss": 3.510268080511718, + "tokens_seen": 1678311424 + }, + { + "epoch": 0.51, + "learning_rate": 0.0003971915747241725, + "loss": 1.4049, + "theoretical_loss": 3.5102563500543233, + "tokens_seen": 1678376960 + }, + { + "epoch": 0.51, + "learning_rate": 0.0003971755265797392, + "loss": 1.447, + "theoretical_loss": 3.510244620183207, + "tokens_seen": 1678442496 + }, + { + "epoch": 0.51, + "learning_rate": 0.0003971594784353059, + "loss": 1.3031, + "theoretical_loss": 3.510232890898319, + "tokens_seen": 1678508032 + }, + { + "epoch": 0.51, + "learning_rate": 0.0003971434302908726, + "loss": 1.3378, + "theoretical_loss": 3.510221162199605, + "tokens_seen": 1678573568 + }, + { + "epoch": 0.51, + "learning_rate": 0.0003971273821464393, + "loss": 1.3104, + "theoretical_loss": 3.5102094340870136, + "tokens_seen": 1678639104 + }, + { + "epoch": 0.51, + "learning_rate": 0.000397111334002006, + "loss": 1.2752, + "theoretical_loss": 3.5101977065604926, + "tokens_seen": 1678704640 + }, + { + "epoch": 0.51, + "learning_rate": 0.0003970952858575727, + "loss": 1.3917, + "theoretical_loss": 3.5101859796199903, + "tokens_seen": 1678770176 + }, + { + "epoch": 0.51, + "learning_rate": 0.0003970792377131394, + "loss": 1.3371, + "theoretical_loss": 3.510174253265454, + "tokens_seen": 1678835712 + }, + { + "epoch": 0.51, + "learning_rate": 0.0003970631895687061, + "loss": 1.4064, + "theoretical_loss": 3.5101625274968318, + "tokens_seen": 1678901248 + }, + { + "epoch": 0.51, + "learning_rate": 0.00039704714142427283, + "loss": 1.2424, + "theoretical_loss": 3.510150802314071, + "tokens_seen": 1678966784 + }, + { + "epoch": 0.51, + "learning_rate": 0.00039703109327983953, + "loss": 1.3151, + "theoretical_loss": 3.5101390777171204, + "tokens_seen": 1679032320 + }, + { + "epoch": 0.51, + "learning_rate": 0.00039701504513540623, + "loss": 1.3459, + "theoretical_loss": 3.5101273537059274, + "tokens_seen": 1679097856 + }, + { + "epoch": 0.51, + "learning_rate": 0.00039699899699097293, + "loss": 1.2194, + "theoretical_loss": 3.5101156302804393, + "tokens_seen": 1679163392 + }, + { + "epoch": 0.51, + "learning_rate": 0.0003969829488465397, + "loss": 1.3913, + "theoretical_loss": 3.5101039074406053, + "tokens_seen": 1679228928 + }, + { + "epoch": 0.51, + "learning_rate": 0.0003969669007021064, + "loss": 1.3081, + "theoretical_loss": 3.5100921851863722, + "tokens_seen": 1679294464 + }, + { + "epoch": 0.51, + "learning_rate": 0.0003969508525576731, + "loss": 1.3721, + "theoretical_loss": 3.5100804635176877, + "tokens_seen": 1679360000 + }, + { + "epoch": 0.51, + "learning_rate": 0.00039693480441323973, + "loss": 1.3841, + "theoretical_loss": 3.510068742434501, + "tokens_seen": 1679425536 + }, + { + "epoch": 0.51, + "learning_rate": 0.00039691875626880643, + "loss": 1.3635, + "theoretical_loss": 3.510057021936759, + "tokens_seen": 1679491072 + }, + { + "epoch": 0.51, + "learning_rate": 0.00039690270812437313, + "loss": 1.303, + "theoretical_loss": 3.5100453020244093, + "tokens_seen": 1679556608 + }, + { + "epoch": 0.51, + "learning_rate": 0.00039688665997993983, + "loss": 1.4001, + "theoretical_loss": 3.510033582697401, + "tokens_seen": 1679622144 + }, + { + "epoch": 0.51, + "learning_rate": 0.00039687061183550653, + "loss": 1.3434, + "theoretical_loss": 3.5100218639556813, + "tokens_seen": 1679687680 + }, + { + "epoch": 0.51, + "learning_rate": 0.00039685456369107324, + "loss": 1.3888, + "theoretical_loss": 3.510010145799198, + "tokens_seen": 1679753216 + }, + { + "epoch": 0.51, + "learning_rate": 0.00039683851554663994, + "loss": 1.2498, + "theoretical_loss": 3.5099984282278993, + "tokens_seen": 1679818752 + }, + { + "epoch": 0.51, + "learning_rate": 0.00039682246740220664, + "loss": 1.2209, + "theoretical_loss": 3.5099867112417336, + "tokens_seen": 1679884288 + }, + { + "epoch": 0.51, + "learning_rate": 0.00039680641925777334, + "loss": 1.5, + "theoretical_loss": 3.509974994840648, + "tokens_seen": 1679949824 + }, + { + "epoch": 0.51, + "learning_rate": 0.00039679037111334004, + "loss": 1.4598, + "theoretical_loss": 3.5099632790245905, + "tokens_seen": 1680015360 + }, + { + "epoch": 0.51, + "learning_rate": 0.00039677432296890674, + "loss": 1.3908, + "theoretical_loss": 3.5099515637935097, + "tokens_seen": 1680080896 + }, + { + "epoch": 0.51, + "learning_rate": 0.00039675827482447344, + "loss": 1.3651, + "theoretical_loss": 3.5099398491473535, + "tokens_seen": 1680146432 + }, + { + "epoch": 0.51, + "learning_rate": 0.00039674222668004014, + "loss": 1.2894, + "theoretical_loss": 3.5099281350860694, + "tokens_seen": 1680211968 + }, + { + "epoch": 0.51, + "learning_rate": 0.00039672617853560684, + "loss": 1.3234, + "theoretical_loss": 3.5099164216096055, + "tokens_seen": 1680277504 + }, + { + "epoch": 0.51, + "learning_rate": 0.00039671013039117354, + "loss": 1.3715, + "theoretical_loss": 3.5099047087179107, + "tokens_seen": 1680343040 + }, + { + "epoch": 0.51, + "learning_rate": 0.00039669408224674024, + "loss": 1.3957, + "theoretical_loss": 3.5098929964109313, + "tokens_seen": 1680408576 + }, + { + "epoch": 0.51, + "learning_rate": 0.00039667803410230694, + "loss": 1.3009, + "theoretical_loss": 3.5098812846886167, + "tokens_seen": 1680474112 + }, + { + "epoch": 0.51, + "learning_rate": 0.00039666198595787364, + "loss": 1.4099, + "theoretical_loss": 3.5098695735509144, + "tokens_seen": 1680539648 + }, + { + "epoch": 0.51, + "learning_rate": 0.00039664593781344034, + "loss": 1.4064, + "theoretical_loss": 3.509857862997773, + "tokens_seen": 1680605184 + }, + { + "epoch": 0.51, + "learning_rate": 0.00039662988966900704, + "loss": 1.3901, + "theoretical_loss": 3.509846153029139, + "tokens_seen": 1680670720 + }, + { + "epoch": 0.51, + "learning_rate": 0.00039661384152457374, + "loss": 1.3864, + "theoretical_loss": 3.509834443644962, + "tokens_seen": 1680736256 + }, + { + "epoch": 0.51, + "learning_rate": 0.00039659779338014045, + "loss": 1.4959, + "theoretical_loss": 3.50982273484519, + "tokens_seen": 1680801792 + }, + { + "epoch": 0.51, + "learning_rate": 0.00039658174523570715, + "loss": 1.4368, + "theoretical_loss": 3.50981102662977, + "tokens_seen": 1680867328 + }, + { + "epoch": 0.51, + "learning_rate": 0.00039656569709127385, + "loss": 1.3701, + "theoretical_loss": 3.5097993189986507, + "tokens_seen": 1680932864 + }, + { + "epoch": 0.51, + "objective/train/docs_used": 613922, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.3106536865234375, + "objective/train/theoretical_loss": 3.50978761195178, + "objective/train/tokens_used": 1701458400, + "theoretical_loss": 3.50978761195178, + "tokens_seen": 1680998400 + }, + { + "epoch": 0.51, + "learning_rate": 0.00039654964894684055, + "loss": 1.354, + "theoretical_loss": 3.50978761195178, + "tokens_seen": 1680998400 + }, + { + "epoch": 0.51, + "learning_rate": 0.00039653360080240725, + "loss": 1.3781, + "theoretical_loss": 3.5097759054891062, + "tokens_seen": 1681063936 + }, + { + "epoch": 0.51, + "learning_rate": 0.00039651755265797395, + "loss": 1.4772, + "theoretical_loss": 3.5097641996105766, + "tokens_seen": 1681129472 + }, + { + "epoch": 0.51, + "learning_rate": 0.00039650150451354065, + "loss": 1.4124, + "theoretical_loss": 3.5097524943161407, + "tokens_seen": 1681195008 + }, + { + "epoch": 0.51, + "learning_rate": 0.00039648545636910735, + "loss": 1.2925, + "theoretical_loss": 3.5097407896057455, + "tokens_seen": 1681260544 + }, + { + "epoch": 0.51, + "learning_rate": 0.00039646940822467405, + "loss": 1.4311, + "theoretical_loss": 3.5097290854793393, + "tokens_seen": 1681326080 + }, + { + "epoch": 0.51, + "learning_rate": 0.00039645336008024075, + "loss": 1.3804, + "theoretical_loss": 3.50971738193687, + "tokens_seen": 1681391616 + }, + { + "epoch": 0.51, + "learning_rate": 0.00039643731193580745, + "loss": 1.3691, + "theoretical_loss": 3.5097056789782863, + "tokens_seen": 1681457152 + }, + { + "epoch": 0.51, + "learning_rate": 0.00039642126379137415, + "loss": 1.4251, + "theoretical_loss": 3.509693976603536, + "tokens_seen": 1681522688 + }, + { + "epoch": 0.51, + "learning_rate": 0.00039640521564694085, + "loss": 1.4086, + "theoretical_loss": 3.509682274812567, + "tokens_seen": 1681588224 + }, + { + "epoch": 0.51, + "learning_rate": 0.00039638916750250755, + "loss": 1.3996, + "theoretical_loss": 3.509670573605328, + "tokens_seen": 1681653760 + }, + { + "epoch": 0.51, + "learning_rate": 0.00039637311935807425, + "loss": 1.367, + "theoretical_loss": 3.509658872981767, + "tokens_seen": 1681719296 + }, + { + "epoch": 0.51, + "learning_rate": 0.00039635707121364095, + "loss": 1.3818, + "theoretical_loss": 3.509647172941831, + "tokens_seen": 1681784832 + }, + { + "epoch": 0.51, + "learning_rate": 0.00039634102306920766, + "loss": 1.3362, + "theoretical_loss": 3.50963547348547, + "tokens_seen": 1681850368 + }, + { + "epoch": 0.51, + "learning_rate": 0.00039632497492477436, + "loss": 1.3672, + "theoretical_loss": 3.5096237746126304, + "tokens_seen": 1681915904 + }, + { + "epoch": 0.51, + "learning_rate": 0.00039630892678034106, + "loss": 1.3584, + "theoretical_loss": 3.509612076323261, + "tokens_seen": 1681981440 + }, + { + "epoch": 0.51, + "learning_rate": 0.00039629287863590776, + "loss": 1.307, + "theoretical_loss": 3.5096003786173116, + "tokens_seen": 1682046976 + }, + { + "epoch": 0.51, + "learning_rate": 0.00039627683049147446, + "loss": 1.3472, + "theoretical_loss": 3.5095886814947277, + "tokens_seen": 1682112512 + }, + { + "epoch": 0.51, + "learning_rate": 0.00039626078234704116, + "loss": 1.3525, + "theoretical_loss": 3.509576984955459, + "tokens_seen": 1682178048 + }, + { + "epoch": 0.51, + "learning_rate": 0.00039624473420260786, + "loss": 1.4714, + "theoretical_loss": 3.509565288999453, + "tokens_seen": 1682243584 + }, + { + "epoch": 0.51, + "learning_rate": 0.00039622868605817456, + "loss": 1.2846, + "theoretical_loss": 3.5095535936266593, + "tokens_seen": 1682309120 + }, + { + "epoch": 0.51, + "learning_rate": 0.00039621263791374126, + "loss": 1.3377, + "theoretical_loss": 3.509541898837024, + "tokens_seen": 1682374656 + }, + { + "epoch": 0.51, + "learning_rate": 0.00039619658976930796, + "loss": 1.3328, + "theoretical_loss": 3.5095302046304973, + "tokens_seen": 1682440192 + }, + { + "epoch": 0.51, + "learning_rate": 0.00039618054162487466, + "loss": 1.3012, + "theoretical_loss": 3.509518511007026, + "tokens_seen": 1682505728 + }, + { + "epoch": 0.51, + "learning_rate": 0.0003961644934804413, + "loss": 1.3751, + "theoretical_loss": 3.5095068179665585, + "tokens_seen": 1682571264 + }, + { + "epoch": 0.51, + "learning_rate": 0.000396148445336008, + "loss": 1.317, + "theoretical_loss": 3.5094951255090443, + "tokens_seen": 1682636800 + }, + { + "epoch": 0.51, + "learning_rate": 0.0003961323971915747, + "loss": 1.362, + "theoretical_loss": 3.50948343363443, + "tokens_seen": 1682702336 + }, + { + "epoch": 0.51, + "learning_rate": 0.0003961163490471414, + "loss": 1.3348, + "theoretical_loss": 3.5094717423426642, + "tokens_seen": 1682767872 + }, + { + "epoch": 0.51, + "learning_rate": 0.0003961003009027081, + "loss": 1.3454, + "theoretical_loss": 3.5094600516336962, + "tokens_seen": 1682833408 + }, + { + "epoch": 0.51, + "learning_rate": 0.0003960842527582748, + "loss": 1.2785, + "theoretical_loss": 3.5094483615074736, + "tokens_seen": 1682898944 + }, + { + "epoch": 0.51, + "learning_rate": 0.0003960682046138415, + "loss": 1.4296, + "theoretical_loss": 3.509436671963944, + "tokens_seen": 1682964480 + }, + { + "epoch": 0.51, + "learning_rate": 0.0003960521564694082, + "loss": 1.2433, + "theoretical_loss": 3.5094249830030564, + "tokens_seen": 1683030016 + }, + { + "epoch": 0.51, + "learning_rate": 0.0003960361083249749, + "loss": 1.4153, + "theoretical_loss": 3.509413294624759, + "tokens_seen": 1683095552 + }, + { + "epoch": 0.51, + "learning_rate": 0.0003960200601805416, + "loss": 1.3775, + "theoretical_loss": 3.509401606829, + "tokens_seen": 1683161088 + }, + { + "epoch": 0.51, + "learning_rate": 0.0003960040120361083, + "loss": 1.4106, + "theoretical_loss": 3.509389919615728, + "tokens_seen": 1683226624 + }, + { + "epoch": 0.51, + "learning_rate": 0.000395987963891675, + "loss": 1.373, + "theoretical_loss": 3.509378232984891, + "tokens_seen": 1683292160 + }, + { + "epoch": 0.51, + "learning_rate": 0.0003959719157472417, + "loss": 1.3543, + "theoretical_loss": 3.5093665469364366, + "tokens_seen": 1683357696 + }, + { + "epoch": 0.51, + "learning_rate": 0.0003959558676028084, + "loss": 1.3668, + "theoretical_loss": 3.509354861470314, + "tokens_seen": 1683423232 + }, + { + "epoch": 0.51, + "learning_rate": 0.0003959398194583751, + "loss": 1.3887, + "theoretical_loss": 3.509343176586472, + "tokens_seen": 1683488768 + }, + { + "epoch": 0.51, + "learning_rate": 0.0003959237713139418, + "loss": 1.2975, + "theoretical_loss": 3.5093314922848577, + "tokens_seen": 1683554304 + }, + { + "epoch": 0.51, + "learning_rate": 0.0003959077231695085, + "loss": 1.4319, + "theoretical_loss": 3.5093198085654196, + "tokens_seen": 1683619840 + }, + { + "epoch": 0.51, + "learning_rate": 0.0003958916750250752, + "loss": 1.3464, + "theoretical_loss": 3.509308125428107, + "tokens_seen": 1683685376 + }, + { + "epoch": 0.51, + "learning_rate": 0.0003958756268806419, + "loss": 1.3821, + "theoretical_loss": 3.5092964428728672, + "tokens_seen": 1683750912 + }, + { + "epoch": 0.51, + "learning_rate": 0.0003958595787362086, + "loss": 1.2789, + "theoretical_loss": 3.509284760899649, + "tokens_seen": 1683816448 + }, + { + "epoch": 0.51, + "learning_rate": 0.0003958435305917753, + "loss": 1.3796, + "theoretical_loss": 3.509273079508401, + "tokens_seen": 1683881984 + }, + { + "epoch": 0.51, + "learning_rate": 0.000395827482447342, + "loss": 1.5066, + "theoretical_loss": 3.5092613986990715, + "tokens_seen": 1683947520 + }, + { + "epoch": 0.51, + "learning_rate": 0.0003958114343029087, + "loss": 1.4284, + "theoretical_loss": 3.509249718471608, + "tokens_seen": 1684013056 + }, + { + "epoch": 0.51, + "learning_rate": 0.0003957953861584755, + "loss": 1.4347, + "theoretical_loss": 3.50923803882596, + "tokens_seen": 1684078592 + }, + { + "epoch": 0.51, + "learning_rate": 0.0003957793380140422, + "loss": 1.2365, + "theoretical_loss": 3.509226359762075, + "tokens_seen": 1684144128 + }, + { + "epoch": 0.51, + "learning_rate": 0.0003957632898696089, + "loss": 1.3996, + "theoretical_loss": 3.5092146812799023, + "tokens_seen": 1684209664 + }, + { + "epoch": 0.51, + "objective/train/docs_used": 615159, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.2541990280151367, + "objective/train/theoretical_loss": 3.5092030033793895, + "objective/train/tokens_used": 1704735200, + "theoretical_loss": 3.5092030033793895, + "tokens_seen": 1684275200 + }, + { + "epoch": 0.51, + "learning_rate": 0.0003957472417251756, + "loss": 1.304, + "theoretical_loss": 3.5092030033793895, + "tokens_seen": 1684275200 + }, + { + "epoch": 0.51, + "learning_rate": 0.0003957311935807423, + "loss": 1.3511, + "theoretical_loss": 3.5091913260604857, + "tokens_seen": 1684340736 + }, + { + "epoch": 0.51, + "learning_rate": 0.000395715145436309, + "loss": 1.2975, + "theoretical_loss": 3.509179649323138, + "tokens_seen": 1684406272 + }, + { + "epoch": 0.51, + "learning_rate": 0.0003956990972918757, + "loss": 1.363, + "theoretical_loss": 3.5091679731672967, + "tokens_seen": 1684471808 + }, + { + "epoch": 0.51, + "learning_rate": 0.0003956830491474424, + "loss": 1.3599, + "theoretical_loss": 3.5091562975929085, + "tokens_seen": 1684537344 + }, + { + "epoch": 0.51, + "learning_rate": 0.0003956670010030091, + "loss": 1.3829, + "theoretical_loss": 3.5091446225999228, + "tokens_seen": 1684602880 + }, + { + "epoch": 0.51, + "learning_rate": 0.0003956509528585758, + "loss": 1.3565, + "theoretical_loss": 3.509132948188287, + "tokens_seen": 1684668416 + }, + { + "epoch": 0.51, + "learning_rate": 0.0003956349047141425, + "loss": 1.2675, + "theoretical_loss": 3.5091212743579514, + "tokens_seen": 1684733952 + }, + { + "epoch": 0.51, + "learning_rate": 0.0003956188565697092, + "loss": 1.3014, + "theoretical_loss": 3.5091096011088627, + "tokens_seen": 1684799488 + }, + { + "epoch": 0.51, + "learning_rate": 0.0003956028084252759, + "loss": 1.3917, + "theoretical_loss": 3.5090979284409705, + "tokens_seen": 1684865024 + }, + { + "epoch": 0.51, + "learning_rate": 0.0003955867602808426, + "loss": 1.371, + "theoretical_loss": 3.5090862563542227, + "tokens_seen": 1684930560 + }, + { + "epoch": 0.51, + "learning_rate": 0.0003955707121364093, + "loss": 1.3103, + "theoretical_loss": 3.509074584848568, + "tokens_seen": 1684996096 + }, + { + "epoch": 0.51, + "learning_rate": 0.000395554663991976, + "loss": 1.2232, + "theoretical_loss": 3.5090629139239535, + "tokens_seen": 1685061632 + }, + { + "epoch": 0.51, + "learning_rate": 0.00039553861584754263, + "loss": 1.3388, + "theoretical_loss": 3.50905124358033, + "tokens_seen": 1685127168 + }, + { + "epoch": 0.51, + "learning_rate": 0.00039552256770310933, + "loss": 1.3506, + "theoretical_loss": 3.5090395738176445, + "tokens_seen": 1685192704 + }, + { + "epoch": 0.51, + "learning_rate": 0.00039550651955867603, + "loss": 1.2373, + "theoretical_loss": 3.509027904635846, + "tokens_seen": 1685258240 + }, + { + "epoch": 0.51, + "learning_rate": 0.00039549047141424273, + "loss": 1.4538, + "theoretical_loss": 3.509016236034883, + "tokens_seen": 1685323776 + }, + { + "epoch": 0.51, + "learning_rate": 0.00039547442326980944, + "loss": 1.4068, + "theoretical_loss": 3.5090045680147037, + "tokens_seen": 1685389312 + }, + { + "epoch": 0.51, + "learning_rate": 0.00039545837512537614, + "loss": 1.3176, + "theoretical_loss": 3.5089929005752563, + "tokens_seen": 1685454848 + }, + { + "epoch": 0.51, + "learning_rate": 0.00039544232698094284, + "loss": 1.4519, + "theoretical_loss": 3.50898123371649, + "tokens_seen": 1685520384 + }, + { + "epoch": 0.51, + "learning_rate": 0.00039542627883650954, + "loss": 1.3943, + "theoretical_loss": 3.5089695674383536, + "tokens_seen": 1685585920 + }, + { + "epoch": 0.51, + "learning_rate": 0.00039541023069207624, + "loss": 1.4132, + "theoretical_loss": 3.5089579017407946, + "tokens_seen": 1685651456 + }, + { + "epoch": 0.51, + "learning_rate": 0.00039539418254764294, + "loss": 1.3973, + "theoretical_loss": 3.5089462366237623, + "tokens_seen": 1685716992 + }, + { + "epoch": 0.51, + "learning_rate": 0.00039537813440320964, + "loss": 1.3521, + "theoretical_loss": 3.508934572087205, + "tokens_seen": 1685782528 + }, + { + "epoch": 0.51, + "learning_rate": 0.00039536208625877634, + "loss": 1.2811, + "theoretical_loss": 3.5089229081310713, + "tokens_seen": 1685848064 + }, + { + "epoch": 0.51, + "learning_rate": 0.00039534603811434304, + "loss": 1.3256, + "theoretical_loss": 3.5089112447553097, + "tokens_seen": 1685913600 + }, + { + "epoch": 0.51, + "learning_rate": 0.00039532998996990974, + "loss": 1.3818, + "theoretical_loss": 3.508899581959869, + "tokens_seen": 1685979136 + }, + { + "epoch": 0.51, + "learning_rate": 0.00039531394182547644, + "loss": 1.3328, + "theoretical_loss": 3.5088879197446974, + "tokens_seen": 1686044672 + }, + { + "epoch": 0.51, + "learning_rate": 0.00039529789368104314, + "loss": 1.3383, + "theoretical_loss": 3.5088762581097432, + "tokens_seen": 1686110208 + }, + { + "epoch": 0.51, + "learning_rate": 0.00039528184553660984, + "loss": 1.4369, + "theoretical_loss": 3.5088645970549557, + "tokens_seen": 1686175744 + }, + { + "epoch": 0.51, + "learning_rate": 0.00039526579739217654, + "loss": 1.3911, + "theoretical_loss": 3.5088529365802836, + "tokens_seen": 1686241280 + }, + { + "epoch": 0.51, + "learning_rate": 0.00039524974924774324, + "loss": 1.3357, + "theoretical_loss": 3.5088412766856747, + "tokens_seen": 1686306816 + }, + { + "epoch": 0.51, + "learning_rate": 0.00039523370110330995, + "loss": 1.3163, + "theoretical_loss": 3.5088296173710782, + "tokens_seen": 1686372352 + }, + { + "epoch": 0.51, + "learning_rate": 0.00039521765295887665, + "loss": 1.3289, + "theoretical_loss": 3.508817958636442, + "tokens_seen": 1686437888 + }, + { + "epoch": 0.51, + "learning_rate": 0.00039520160481444335, + "loss": 1.2524, + "theoretical_loss": 3.508806300481716, + "tokens_seen": 1686503424 + }, + { + "epoch": 0.51, + "learning_rate": 0.00039518555667001005, + "loss": 1.3116, + "theoretical_loss": 3.5087946429068477, + "tokens_seen": 1686568960 + }, + { + "epoch": 0.51, + "learning_rate": 0.00039516950852557675, + "loss": 1.4079, + "theoretical_loss": 3.5087829859117856, + "tokens_seen": 1686634496 + }, + { + "epoch": 0.51, + "learning_rate": 0.00039515346038114345, + "loss": 1.2775, + "theoretical_loss": 3.5087713294964793, + "tokens_seen": 1686700032 + }, + { + "epoch": 0.51, + "learning_rate": 0.00039513741223671015, + "loss": 1.3611, + "theoretical_loss": 3.508759673660877, + "tokens_seen": 1686765568 + }, + { + "epoch": 0.51, + "learning_rate": 0.00039512136409227685, + "loss": 1.3969, + "theoretical_loss": 3.508748018404927, + "tokens_seen": 1686831104 + }, + { + "epoch": 0.51, + "learning_rate": 0.00039510531594784355, + "loss": 1.4001, + "theoretical_loss": 3.5087363637285787, + "tokens_seen": 1686896640 + }, + { + "epoch": 0.51, + "learning_rate": 0.00039508926780341025, + "loss": 1.4674, + "theoretical_loss": 3.50872470963178, + "tokens_seen": 1686962176 + }, + { + "epoch": 0.51, + "learning_rate": 0.00039507321965897695, + "loss": 1.5632, + "theoretical_loss": 3.5087130561144795, + "tokens_seen": 1687027712 + }, + { + "epoch": 0.51, + "learning_rate": 0.00039505717151454365, + "loss": 1.4705, + "theoretical_loss": 3.5087014031766266, + "tokens_seen": 1687093248 + }, + { + "epoch": 0.51, + "learning_rate": 0.00039504112337011035, + "loss": 1.2844, + "theoretical_loss": 3.50868975081817, + "tokens_seen": 1687158784 + }, + { + "epoch": 0.51, + "learning_rate": 0.00039502507522567705, + "loss": 1.4534, + "theoretical_loss": 3.5086780990390576, + "tokens_seen": 1687224320 + }, + { + "epoch": 0.51, + "learning_rate": 0.00039500902708124375, + "loss": 1.3723, + "theoretical_loss": 3.5086664478392384, + "tokens_seen": 1687289856 + }, + { + "epoch": 0.51, + "learning_rate": 0.00039499297893681045, + "loss": 1.5201, + "theoretical_loss": 3.5086547972186617, + "tokens_seen": 1687355392 + }, + { + "epoch": 0.51, + "learning_rate": 0.00039497693079237716, + "loss": 1.389, + "theoretical_loss": 3.5086431471772754, + "tokens_seen": 1687420928 + }, + { + "epoch": 0.51, + "learning_rate": 0.00039496088264794386, + "loss": 1.3659, + "theoretical_loss": 3.5086314977150286, + "tokens_seen": 1687486464 + }, + { + "epoch": 0.51, + "objective/train/docs_used": 616363, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.4115370512008667, + "objective/train/theoretical_loss": 3.50861984883187, + "objective/train/tokens_used": 1708012000, + "theoretical_loss": 3.50861984883187, + "tokens_seen": 1687552000 + }, + { + "epoch": 0.51, + "learning_rate": 0.00039494483450351056, + "loss": 1.3935, + "theoretical_loss": 3.50861984883187, + "tokens_seen": 1687552000 + }, + { + "epoch": 0.51, + "learning_rate": 0.00039492878635907726, + "loss": 1.3967, + "theoretical_loss": 3.508608200527749, + "tokens_seen": 1687617536 + }, + { + "epoch": 0.51, + "learning_rate": 0.00039491273821464396, + "loss": 1.468, + "theoretical_loss": 3.5085965528026124, + "tokens_seen": 1687683072 + }, + { + "epoch": 0.51, + "learning_rate": 0.00039489669007021066, + "loss": 1.2826, + "theoretical_loss": 3.5085849056564107, + "tokens_seen": 1687748608 + }, + { + "epoch": 0.51, + "learning_rate": 0.00039488064192577736, + "loss": 1.3647, + "theoretical_loss": 3.508573259089092, + "tokens_seen": 1687814144 + }, + { + "epoch": 0.51, + "learning_rate": 0.00039486459378134406, + "loss": 1.4406, + "theoretical_loss": 3.508561613100605, + "tokens_seen": 1687879680 + }, + { + "epoch": 0.51, + "learning_rate": 0.00039484854563691076, + "loss": 1.3281, + "theoretical_loss": 3.5085499676908993, + "tokens_seen": 1687945216 + }, + { + "epoch": 0.51, + "learning_rate": 0.00039483249749247746, + "loss": 1.3396, + "theoretical_loss": 3.5085383228599225, + "tokens_seen": 1688010752 + }, + { + "epoch": 0.51, + "learning_rate": 0.00039481644934804416, + "loss": 1.3697, + "theoretical_loss": 3.5085266786076237, + "tokens_seen": 1688076288 + }, + { + "epoch": 0.51, + "learning_rate": 0.00039480040120361086, + "loss": 1.3844, + "theoretical_loss": 3.5085150349339522, + "tokens_seen": 1688141824 + }, + { + "epoch": 0.51, + "learning_rate": 0.00039478435305917756, + "loss": 1.3484, + "theoretical_loss": 3.508503391838856, + "tokens_seen": 1688207360 + }, + { + "epoch": 0.51, + "learning_rate": 0.0003947683049147442, + "loss": 1.2591, + "theoretical_loss": 3.5084917493222845, + "tokens_seen": 1688272896 + }, + { + "epoch": 0.51, + "learning_rate": 0.0003947522567703109, + "loss": 1.444, + "theoretical_loss": 3.5084801073841865, + "tokens_seen": 1688338432 + }, + { + "epoch": 0.51, + "learning_rate": 0.0003947362086258776, + "loss": 1.3218, + "theoretical_loss": 3.5084684660245107, + "tokens_seen": 1688403968 + }, + { + "epoch": 0.51, + "learning_rate": 0.0003947201604814443, + "loss": 1.2592, + "theoretical_loss": 3.5084568252432056, + "tokens_seen": 1688469504 + }, + { + "epoch": 0.51, + "learning_rate": 0.000394704112337011, + "loss": 1.2407, + "theoretical_loss": 3.5084451850402205, + "tokens_seen": 1688535040 + }, + { + "epoch": 0.51, + "learning_rate": 0.0003946880641925777, + "loss": 1.3427, + "theoretical_loss": 3.5084335454155036, + "tokens_seen": 1688600576 + }, + { + "epoch": 0.51, + "learning_rate": 0.0003946720160481444, + "loss": 1.4764, + "theoretical_loss": 3.508421906369004, + "tokens_seen": 1688666112 + }, + { + "epoch": 0.51, + "learning_rate": 0.0003946559679037111, + "loss": 1.3978, + "theoretical_loss": 3.508410267900671, + "tokens_seen": 1688731648 + }, + { + "epoch": 0.51, + "learning_rate": 0.0003946399197592778, + "loss": 1.2337, + "theoretical_loss": 3.5083986300104533, + "tokens_seen": 1688797184 + }, + { + "epoch": 0.51, + "learning_rate": 0.0003946238716148445, + "loss": 1.4034, + "theoretical_loss": 3.508386992698299, + "tokens_seen": 1688862720 + }, + { + "epoch": 0.51, + "learning_rate": 0.0003946078234704112, + "loss": 1.1864, + "theoretical_loss": 3.5083753559641577, + "tokens_seen": 1688928256 + }, + { + "epoch": 0.51, + "learning_rate": 0.00039459177532597797, + "loss": 1.323, + "theoretical_loss": 3.508363719807978, + "tokens_seen": 1688993792 + }, + { + "epoch": 0.51, + "learning_rate": 0.00039457572718154467, + "loss": 1.3561, + "theoretical_loss": 3.508352084229709, + "tokens_seen": 1689059328 + }, + { + "epoch": 0.51, + "learning_rate": 0.00039455967903711137, + "loss": 1.3525, + "theoretical_loss": 3.5083404492292996, + "tokens_seen": 1689124864 + }, + { + "epoch": 0.51, + "learning_rate": 0.00039454363089267807, + "loss": 1.1924, + "theoretical_loss": 3.508328814806698, + "tokens_seen": 1689190400 + }, + { + "epoch": 0.51, + "learning_rate": 0.0003945275827482448, + "loss": 1.318, + "theoretical_loss": 3.5083171809618534, + "tokens_seen": 1689255936 + }, + { + "epoch": 0.51, + "learning_rate": 0.0003945115346038115, + "loss": 1.45, + "theoretical_loss": 3.5083055476947154, + "tokens_seen": 1689321472 + }, + { + "epoch": 0.51, + "learning_rate": 0.0003944954864593782, + "loss": 1.3726, + "theoretical_loss": 3.508293915005232, + "tokens_seen": 1689387008 + }, + { + "epoch": 0.51, + "learning_rate": 0.0003944794383149449, + "loss": 1.5332, + "theoretical_loss": 3.5082822828933526, + "tokens_seen": 1689452544 + }, + { + "epoch": 0.51, + "learning_rate": 0.0003944633901705116, + "loss": 1.4015, + "theoretical_loss": 3.5082706513590263, + "tokens_seen": 1689518080 + }, + { + "epoch": 0.51, + "learning_rate": 0.0003944473420260783, + "loss": 1.2157, + "theoretical_loss": 3.5082590204022015, + "tokens_seen": 1689583616 + }, + { + "epoch": 0.51, + "learning_rate": 0.000394431293881645, + "loss": 1.4767, + "theoretical_loss": 3.5082473900228273, + "tokens_seen": 1689649152 + }, + { + "epoch": 0.51, + "learning_rate": 0.0003944152457372117, + "loss": 1.4474, + "theoretical_loss": 3.5082357602208525, + "tokens_seen": 1689714688 + }, + { + "epoch": 0.51, + "learning_rate": 0.0003943991975927784, + "loss": 1.3795, + "theoretical_loss": 3.5082241309962265, + "tokens_seen": 1689780224 + }, + { + "epoch": 0.51, + "learning_rate": 0.0003943831494483451, + "loss": 1.2316, + "theoretical_loss": 3.508212502348898, + "tokens_seen": 1689845760 + }, + { + "epoch": 0.51, + "learning_rate": 0.0003943671013039118, + "loss": 1.4114, + "theoretical_loss": 3.508200874278816, + "tokens_seen": 1689911296 + }, + { + "epoch": 0.51, + "learning_rate": 0.0003943510531594785, + "loss": 1.39, + "theoretical_loss": 3.5081892467859292, + "tokens_seen": 1689976832 + }, + { + "epoch": 0.51, + "learning_rate": 0.0003943350050150452, + "loss": 1.4279, + "theoretical_loss": 3.5081776198701866, + "tokens_seen": 1690042368 + }, + { + "epoch": 0.51, + "learning_rate": 0.0003943189568706119, + "loss": 1.4058, + "theoretical_loss": 3.5081659935315375, + "tokens_seen": 1690107904 + }, + { + "epoch": 0.51, + "learning_rate": 0.0003943029087261786, + "loss": 1.2944, + "theoretical_loss": 3.5081543677699307, + "tokens_seen": 1690173440 + }, + { + "epoch": 0.51, + "learning_rate": 0.0003942868605817453, + "loss": 1.3122, + "theoretical_loss": 3.5081427425853153, + "tokens_seen": 1690238976 + }, + { + "epoch": 0.51, + "learning_rate": 0.000394270812437312, + "loss": 1.3539, + "theoretical_loss": 3.5081311179776398, + "tokens_seen": 1690304512 + }, + { + "epoch": 0.51, + "learning_rate": 0.0003942547642928787, + "loss": 1.3785, + "theoretical_loss": 3.508119493946854, + "tokens_seen": 1690370048 + }, + { + "epoch": 0.51, + "learning_rate": 0.0003942387161484454, + "loss": 1.3371, + "theoretical_loss": 3.5081078704929065, + "tokens_seen": 1690435584 + }, + { + "epoch": 0.51, + "learning_rate": 0.0003942226680040121, + "loss": 1.4102, + "theoretical_loss": 3.5080962476157462, + "tokens_seen": 1690501120 + }, + { + "epoch": 0.51, + "learning_rate": 0.0003942066198595788, + "loss": 1.335, + "theoretical_loss": 3.5080846253153224, + "tokens_seen": 1690566656 + }, + { + "epoch": 0.51, + "learning_rate": 0.0003941905717151455, + "loss": 1.3755, + "theoretical_loss": 3.508073003591584, + "tokens_seen": 1690632192 + }, + { + "epoch": 0.51, + "learning_rate": 0.0003941745235707122, + "loss": 1.4151, + "theoretical_loss": 3.5080613824444793, + "tokens_seen": 1690697728 + }, + { + "epoch": 0.51, + "learning_rate": 0.0003941584754262789, + "loss": 1.2561, + "theoretical_loss": 3.508049761873959, + "tokens_seen": 1690763264 + }, + { + "epoch": 0.51, + "objective/train/docs_used": 617433, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.372788429260254, + "objective/train/theoretical_loss": 3.5080381418799704, + "objective/train/tokens_used": 1711288800, + "theoretical_loss": 3.5080381418799704, + "tokens_seen": 1690828800 + }, + { + "epoch": 0.51, + "learning_rate": 0.00039414242728184553, + "loss": 1.3585, + "theoretical_loss": 3.5080381418799704, + "tokens_seen": 1690828800 + }, + { + "epoch": 0.51, + "learning_rate": 0.00039412637913741223, + "loss": 1.3861, + "theoretical_loss": 3.5080265224624636, + "tokens_seen": 1690894336 + }, + { + "epoch": 0.51, + "learning_rate": 0.00039411033099297894, + "loss": 1.3248, + "theoretical_loss": 3.5080149036213877, + "tokens_seen": 1690959872 + }, + { + "epoch": 0.51, + "learning_rate": 0.00039409428284854564, + "loss": 1.298, + "theoretical_loss": 3.508003285356691, + "tokens_seen": 1691025408 + }, + { + "epoch": 0.51, + "learning_rate": 0.00039407823470411234, + "loss": 1.4039, + "theoretical_loss": 3.5079916676683234, + "tokens_seen": 1691090944 + }, + { + "epoch": 0.51, + "learning_rate": 0.00039406218655967904, + "loss": 1.4108, + "theoretical_loss": 3.507980050556233, + "tokens_seen": 1691156480 + }, + { + "epoch": 0.51, + "learning_rate": 0.00039404613841524574, + "loss": 1.3714, + "theoretical_loss": 3.50796843402037, + "tokens_seen": 1691222016 + }, + { + "epoch": 0.51, + "learning_rate": 0.00039403009027081244, + "loss": 1.3501, + "theoretical_loss": 3.507956818060683, + "tokens_seen": 1691287552 + }, + { + "epoch": 0.51, + "learning_rate": 0.00039401404212637914, + "loss": 1.3362, + "theoretical_loss": 3.5079452026771207, + "tokens_seen": 1691353088 + }, + { + "epoch": 0.51, + "learning_rate": 0.00039399799398194584, + "loss": 1.3963, + "theoretical_loss": 3.507933587869633, + "tokens_seen": 1691418624 + }, + { + "epoch": 0.51, + "learning_rate": 0.00039398194583751254, + "loss": 1.3643, + "theoretical_loss": 3.5079219736381684, + "tokens_seen": 1691484160 + }, + { + "epoch": 0.51, + "learning_rate": 0.00039396589769307924, + "loss": 1.3684, + "theoretical_loss": 3.507910359982676, + "tokens_seen": 1691549696 + }, + { + "epoch": 0.51, + "learning_rate": 0.00039394984954864594, + "loss": 1.376, + "theoretical_loss": 3.5078987469031055, + "tokens_seen": 1691615232 + }, + { + "epoch": 0.51, + "learning_rate": 0.00039393380140421264, + "loss": 1.3368, + "theoretical_loss": 3.507887134399405, + "tokens_seen": 1691680768 + }, + { + "epoch": 0.51, + "learning_rate": 0.00039391775325977934, + "loss": 1.467, + "theoretical_loss": 3.507875522471525, + "tokens_seen": 1691746304 + }, + { + "epoch": 0.51, + "learning_rate": 0.00039390170511534604, + "loss": 1.2516, + "theoretical_loss": 3.5078639111194136, + "tokens_seen": 1691811840 + }, + { + "epoch": 0.51, + "learning_rate": 0.00039388565697091274, + "loss": 1.3363, + "theoretical_loss": 3.5078523003430204, + "tokens_seen": 1691877376 + }, + { + "epoch": 0.51, + "learning_rate": 0.00039386960882647944, + "loss": 1.4022, + "theoretical_loss": 3.5078406901422943, + "tokens_seen": 1691942912 + }, + { + "epoch": 0.51, + "learning_rate": 0.00039385356068204615, + "loss": 1.3757, + "theoretical_loss": 3.5078290805171846, + "tokens_seen": 1692008448 + }, + { + "epoch": 0.51, + "learning_rate": 0.00039383751253761285, + "loss": 1.3589, + "theoretical_loss": 3.5078174714676407, + "tokens_seen": 1692073984 + }, + { + "epoch": 0.51, + "learning_rate": 0.00039382146439317955, + "loss": 1.4245, + "theoretical_loss": 3.507805862993611, + "tokens_seen": 1692139520 + }, + { + "epoch": 0.51, + "learning_rate": 0.00039380541624874625, + "loss": 1.4529, + "theoretical_loss": 3.507794255095046, + "tokens_seen": 1692205056 + }, + { + "epoch": 0.51, + "learning_rate": 0.00039378936810431295, + "loss": 1.3651, + "theoretical_loss": 3.5077826477718936, + "tokens_seen": 1692270592 + }, + { + "epoch": 0.51, + "learning_rate": 0.00039377331995987965, + "loss": 1.3544, + "theoretical_loss": 3.5077710410241036, + "tokens_seen": 1692336128 + }, + { + "epoch": 0.51, + "learning_rate": 0.00039375727181544635, + "loss": 1.3788, + "theoretical_loss": 3.507759434851625, + "tokens_seen": 1692401664 + }, + { + "epoch": 0.51, + "learning_rate": 0.00039374122367101305, + "loss": 1.3481, + "theoretical_loss": 3.5077478292544075, + "tokens_seen": 1692467200 + }, + { + "epoch": 0.51, + "learning_rate": 0.00039372517552657975, + "loss": 1.3439, + "theoretical_loss": 3.5077362242323993, + "tokens_seen": 1692532736 + }, + { + "epoch": 0.51, + "learning_rate": 0.00039370912738214645, + "loss": 1.3142, + "theoretical_loss": 3.5077246197855505, + "tokens_seen": 1692598272 + }, + { + "epoch": 0.51, + "learning_rate": 0.00039369307923771315, + "loss": 1.2587, + "theoretical_loss": 3.5077130159138106, + "tokens_seen": 1692663808 + }, + { + "epoch": 0.51, + "learning_rate": 0.00039367703109327985, + "loss": 1.3999, + "theoretical_loss": 3.507701412617128, + "tokens_seen": 1692729344 + }, + { + "epoch": 0.51, + "learning_rate": 0.00039366098294884655, + "loss": 1.3595, + "theoretical_loss": 3.507689809895452, + "tokens_seen": 1692794880 + }, + { + "epoch": 0.51, + "learning_rate": 0.00039364493480441325, + "loss": 1.3454, + "theoretical_loss": 3.5076782077487323, + "tokens_seen": 1692860416 + }, + { + "epoch": 0.51, + "learning_rate": 0.00039362888665997995, + "loss": 1.3401, + "theoretical_loss": 3.507666606176918, + "tokens_seen": 1692925952 + }, + { + "epoch": 0.51, + "learning_rate": 0.00039361283851554665, + "loss": 1.3776, + "theoretical_loss": 3.5076550051799575, + "tokens_seen": 1692991488 + }, + { + "epoch": 0.51, + "learning_rate": 0.00039359679037111336, + "loss": 1.4412, + "theoretical_loss": 3.507643404757802, + "tokens_seen": 1693057024 + }, + { + "epoch": 0.51, + "learning_rate": 0.00039358074222668006, + "loss": 1.4467, + "theoretical_loss": 3.5076318049103987, + "tokens_seen": 1693122560 + }, + { + "epoch": 0.51, + "learning_rate": 0.00039356469408224676, + "loss": 1.3524, + "theoretical_loss": 3.507620205637698, + "tokens_seen": 1693188096 + }, + { + "epoch": 0.51, + "learning_rate": 0.00039354864593781346, + "loss": 1.471, + "theoretical_loss": 3.50760860693965, + "tokens_seen": 1693253632 + }, + { + "epoch": 0.51, + "learning_rate": 0.00039353259779338016, + "loss": 1.3309, + "theoretical_loss": 3.507597008816202, + "tokens_seen": 1693319168 + }, + { + "epoch": 0.51, + "learning_rate": 0.00039351654964894686, + "loss": 1.3834, + "theoretical_loss": 3.5075854112673044, + "tokens_seen": 1693384704 + }, + { + "epoch": 0.51, + "learning_rate": 0.00039350050150451356, + "loss": 1.2737, + "theoretical_loss": 3.5075738142929067, + "tokens_seen": 1693450240 + }, + { + "epoch": 0.51, + "learning_rate": 0.00039348445336008026, + "loss": 1.2866, + "theoretical_loss": 3.507562217892958, + "tokens_seen": 1693515776 + }, + { + "epoch": 0.51, + "learning_rate": 0.00039346840521564696, + "loss": 1.2397, + "theoretical_loss": 3.5075506220674066, + "tokens_seen": 1693581312 + }, + { + "epoch": 0.51, + "learning_rate": 0.00039345235707121366, + "loss": 1.3428, + "theoretical_loss": 3.507539026816204, + "tokens_seen": 1693646848 + }, + { + "epoch": 0.51, + "learning_rate": 0.00039343630892678036, + "loss": 1.3625, + "theoretical_loss": 3.5075274321392973, + "tokens_seen": 1693712384 + }, + { + "epoch": 0.51, + "learning_rate": 0.00039342026078234706, + "loss": 1.3876, + "theoretical_loss": 3.507515838036637, + "tokens_seen": 1693777920 + }, + { + "epoch": 0.51, + "learning_rate": 0.00039340421263791376, + "loss": 1.2507, + "theoretical_loss": 3.5075042445081728, + "tokens_seen": 1693843456 + }, + { + "epoch": 0.51, + "learning_rate": 0.00039338816449348046, + "loss": 1.348, + "theoretical_loss": 3.5074926515538527, + "tokens_seen": 1693908992 + }, + { + "epoch": 0.51, + "learning_rate": 0.00039337211634904716, + "loss": 1.2645, + "theoretical_loss": 3.507481059173627, + "tokens_seen": 1693974528 + }, + { + "epoch": 0.51, + "learning_rate": 0.00039335606820461387, + "loss": 1.3794, + "theoretical_loss": 3.5074694673674456, + "tokens_seen": 1694040064 + }, + { + "epoch": 0.51, + "objective/train/docs_used": 618718, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.4453774690628052, + "objective/train/theoretical_loss": 3.507457876135257, + "objective/train/tokens_used": 1714565600, + "theoretical_loss": 3.507457876135257, + "tokens_seen": 1694105600 + }, + { + "epoch": 0.51, + "learning_rate": 0.00039334002006018057, + "loss": 1.4491, + "theoretical_loss": 3.507457876135257, + "tokens_seen": 1694105600 + }, + { + "epoch": 0.51, + "learning_rate": 0.00039332397191574727, + "loss": 1.4042, + "theoretical_loss": 3.5074462854770103, + "tokens_seen": 1694171136 + }, + { + "epoch": 0.51, + "learning_rate": 0.00039330792377131397, + "loss": 1.3292, + "theoretical_loss": 3.5074346953926554, + "tokens_seen": 1694236672 + }, + { + "epoch": 0.51, + "learning_rate": 0.00039329187562688067, + "loss": 1.3731, + "theoretical_loss": 3.507423105882142, + "tokens_seen": 1694302208 + }, + { + "epoch": 0.51, + "learning_rate": 0.00039327582748244737, + "loss": 1.2666, + "theoretical_loss": 3.5074115169454188, + "tokens_seen": 1694367744 + }, + { + "epoch": 0.51, + "learning_rate": 0.00039325977933801407, + "loss": 1.3245, + "theoretical_loss": 3.5073999285824353, + "tokens_seen": 1694433280 + }, + { + "epoch": 0.51, + "learning_rate": 0.00039324373119358077, + "loss": 1.2348, + "theoretical_loss": 3.507388340793142, + "tokens_seen": 1694498816 + }, + { + "epoch": 0.51, + "learning_rate": 0.00039322768304914747, + "loss": 1.3138, + "theoretical_loss": 3.5073767535774865, + "tokens_seen": 1694564352 + }, + { + "epoch": 0.51, + "learning_rate": 0.00039321163490471417, + "loss": 1.413, + "theoretical_loss": 3.5073651669354193, + "tokens_seen": 1694629888 + }, + { + "epoch": 0.51, + "learning_rate": 0.00039319558676028087, + "loss": 1.3664, + "theoretical_loss": 3.50735358086689, + "tokens_seen": 1694695424 + }, + { + "epoch": 0.51, + "learning_rate": 0.00039317953861584757, + "loss": 1.2627, + "theoretical_loss": 3.5073419953718474, + "tokens_seen": 1694760960 + }, + { + "epoch": 0.51, + "learning_rate": 0.00039316349047141427, + "loss": 1.3186, + "theoretical_loss": 3.507330410450242, + "tokens_seen": 1694826496 + }, + { + "epoch": 0.51, + "learning_rate": 0.000393147442326981, + "loss": 1.3686, + "theoretical_loss": 3.507318826102022, + "tokens_seen": 1694892032 + }, + { + "epoch": 0.51, + "learning_rate": 0.0003931313941825477, + "loss": 1.3617, + "theoretical_loss": 3.507307242327137, + "tokens_seen": 1694957568 + }, + { + "epoch": 0.51, + "learning_rate": 0.0003931153460381144, + "loss": 1.3934, + "theoretical_loss": 3.5072956591255373, + "tokens_seen": 1695023104 + }, + { + "epoch": 0.51, + "learning_rate": 0.0003930992978936811, + "loss": 1.3066, + "theoretical_loss": 3.5072840764971716, + "tokens_seen": 1695088640 + }, + { + "epoch": 0.51, + "learning_rate": 0.0003930832497492478, + "loss": 1.3659, + "theoretical_loss": 3.50727249444199, + "tokens_seen": 1695154176 + }, + { + "epoch": 0.51, + "learning_rate": 0.0003930672016048145, + "loss": 1.2988, + "theoretical_loss": 3.507260912959941, + "tokens_seen": 1695219712 + }, + { + "epoch": 0.51, + "learning_rate": 0.0003930511534603812, + "loss": 1.3191, + "theoretical_loss": 3.507249332050975, + "tokens_seen": 1695285248 + }, + { + "epoch": 0.51, + "learning_rate": 0.0003930351053159479, + "loss": 1.3355, + "theoretical_loss": 3.5072377517150413, + "tokens_seen": 1695350784 + }, + { + "epoch": 0.51, + "learning_rate": 0.0003930190571715146, + "loss": 1.2478, + "theoretical_loss": 3.507226171952089, + "tokens_seen": 1695416320 + }, + { + "epoch": 0.51, + "learning_rate": 0.0003930030090270813, + "loss": 1.3148, + "theoretical_loss": 3.507214592762068, + "tokens_seen": 1695481856 + }, + { + "epoch": 0.51, + "learning_rate": 0.000392986960882648, + "loss": 1.3251, + "theoretical_loss": 3.5072030141449284, + "tokens_seen": 1695547392 + }, + { + "epoch": 0.51, + "learning_rate": 0.0003929709127382147, + "loss": 1.1578, + "theoretical_loss": 3.507191436100618, + "tokens_seen": 1695612928 + }, + { + "epoch": 0.51, + "learning_rate": 0.0003929548645937814, + "loss": 1.4125, + "theoretical_loss": 3.5071798586290877, + "tokens_seen": 1695678464 + }, + { + "epoch": 0.51, + "learning_rate": 0.0003929388164493481, + "loss": 1.424, + "theoretical_loss": 3.507168281730287, + "tokens_seen": 1695744000 + }, + { + "epoch": 0.51, + "learning_rate": 0.0003929227683049148, + "loss": 1.3181, + "theoretical_loss": 3.5071567054041646, + "tokens_seen": 1695809536 + }, + { + "epoch": 0.51, + "learning_rate": 0.0003929067201604815, + "loss": 1.2315, + "theoretical_loss": 3.507145129650671, + "tokens_seen": 1695875072 + }, + { + "epoch": 0.51, + "learning_rate": 0.0003928906720160482, + "loss": 1.4866, + "theoretical_loss": 3.507133554469755, + "tokens_seen": 1695940608 + }, + { + "epoch": 0.51, + "learning_rate": 0.0003928746238716149, + "loss": 1.3131, + "theoretical_loss": 3.507121979861367, + "tokens_seen": 1696006144 + }, + { + "epoch": 0.51, + "learning_rate": 0.0003928585757271816, + "loss": 1.4564, + "theoretical_loss": 3.507110405825455, + "tokens_seen": 1696071680 + }, + { + "epoch": 0.51, + "learning_rate": 0.0003928425275827483, + "loss": 1.3247, + "theoretical_loss": 3.50709883236197, + "tokens_seen": 1696137216 + }, + { + "epoch": 0.51, + "learning_rate": 0.000392826479438315, + "loss": 1.2652, + "theoretical_loss": 3.507087259470861, + "tokens_seen": 1696202752 + }, + { + "epoch": 0.51, + "learning_rate": 0.0003928104312938817, + "loss": 1.3467, + "theoretical_loss": 3.5070756871520783, + "tokens_seen": 1696268288 + }, + { + "epoch": 0.51, + "learning_rate": 0.0003927943831494484, + "loss": 1.3449, + "theoretical_loss": 3.50706411540557, + "tokens_seen": 1696333824 + }, + { + "epoch": 0.51, + "learning_rate": 0.0003927783350050151, + "loss": 1.3583, + "theoretical_loss": 3.5070525442312874, + "tokens_seen": 1696399360 + }, + { + "epoch": 0.51, + "learning_rate": 0.0003927622868605818, + "loss": 1.3309, + "theoretical_loss": 3.507040973629179, + "tokens_seen": 1696464896 + }, + { + "epoch": 0.51, + "learning_rate": 0.0003927462387161485, + "loss": 1.2913, + "theoretical_loss": 3.507029403599195, + "tokens_seen": 1696530432 + }, + { + "epoch": 0.51, + "learning_rate": 0.00039273019057171514, + "loss": 1.3745, + "theoretical_loss": 3.5070178341412843, + "tokens_seen": 1696595968 + }, + { + "epoch": 0.51, + "learning_rate": 0.00039271414242728184, + "loss": 1.3951, + "theoretical_loss": 3.5070062652553977, + "tokens_seen": 1696661504 + }, + { + "epoch": 0.51, + "learning_rate": 0.00039269809428284854, + "loss": 1.3246, + "theoretical_loss": 3.506994696941483, + "tokens_seen": 1696727040 + }, + { + "epoch": 0.51, + "learning_rate": 0.00039268204613841524, + "loss": 1.3865, + "theoretical_loss": 3.5069831291994915, + "tokens_seen": 1696792576 + }, + { + "epoch": 0.51, + "learning_rate": 0.00039266599799398194, + "loss": 1.2595, + "theoretical_loss": 3.506971562029372, + "tokens_seen": 1696858112 + }, + { + "epoch": 0.51, + "learning_rate": 0.00039264994984954864, + "loss": 1.2745, + "theoretical_loss": 3.5069599954310746, + "tokens_seen": 1696923648 + }, + { + "epoch": 0.51, + "learning_rate": 0.00039263390170511534, + "loss": 1.4481, + "theoretical_loss": 3.5069484294045483, + "tokens_seen": 1696989184 + }, + { + "epoch": 0.51, + "learning_rate": 0.00039261785356068204, + "loss": 1.3985, + "theoretical_loss": 3.5069368639497434, + "tokens_seen": 1697054720 + }, + { + "epoch": 0.51, + "learning_rate": 0.00039260180541624874, + "loss": 1.4339, + "theoretical_loss": 3.5069252990666095, + "tokens_seen": 1697120256 + }, + { + "epoch": 0.51, + "learning_rate": 0.00039258575727181544, + "loss": 1.3155, + "theoretical_loss": 3.506913734755096, + "tokens_seen": 1697185792 + }, + { + "epoch": 0.51, + "learning_rate": 0.00039256970912738214, + "loss": 1.3023, + "theoretical_loss": 3.5069021710151524, + "tokens_seen": 1697251328 + }, + { + "epoch": 0.51, + "learning_rate": 0.00039255366098294884, + "loss": 1.4397, + "theoretical_loss": 3.506890607846729, + "tokens_seen": 1697316864 + }, + { + "epoch": 0.51, + "objective/train/docs_used": 619912, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.3029978275299072, + "objective/train/theoretical_loss": 3.506879045249775, + "objective/train/tokens_used": 1717842400, + "theoretical_loss": 3.506879045249775, + "tokens_seen": 1697382400 + }, + { + "epoch": 0.51, + "learning_rate": 0.00039253761283851554, + "loss": 1.386, + "theoretical_loss": 3.506879045249775, + "tokens_seen": 1697382400 + }, + { + "epoch": 0.51, + "learning_rate": 0.00039252156469408224, + "loss": 1.3801, + "theoretical_loss": 3.5068674832242404, + "tokens_seen": 1697447936 + }, + { + "epoch": 0.51, + "learning_rate": 0.00039250551654964894, + "loss": 1.245, + "theoretical_loss": 3.5068559217700748, + "tokens_seen": 1697513472 + }, + { + "epoch": 0.51, + "learning_rate": 0.00039248946840521565, + "loss": 1.4169, + "theoretical_loss": 3.5068443608872277, + "tokens_seen": 1697579008 + }, + { + "epoch": 0.51, + "learning_rate": 0.00039247342026078235, + "loss": 1.2263, + "theoretical_loss": 3.5068328005756486, + "tokens_seen": 1697644544 + }, + { + "epoch": 0.51, + "learning_rate": 0.00039245737211634905, + "loss": 1.3916, + "theoretical_loss": 3.506821240835288, + "tokens_seen": 1697710080 + }, + { + "epoch": 0.51, + "learning_rate": 0.00039244132397191575, + "loss": 1.4194, + "theoretical_loss": 3.5068096816660956, + "tokens_seen": 1697775616 + }, + { + "epoch": 0.51, + "learning_rate": 0.00039242527582748245, + "loss": 1.4042, + "theoretical_loss": 3.5067981230680196, + "tokens_seen": 1697841152 + }, + { + "epoch": 0.51, + "learning_rate": 0.00039240922768304915, + "loss": 1.275, + "theoretical_loss": 3.506786565041012, + "tokens_seen": 1697906688 + }, + { + "epoch": 0.51, + "learning_rate": 0.00039239317953861585, + "loss": 1.3822, + "theoretical_loss": 3.506775007585021, + "tokens_seen": 1697972224 + }, + { + "epoch": 0.51, + "learning_rate": 0.00039237713139418255, + "loss": 1.3277, + "theoretical_loss": 3.5067634506999967, + "tokens_seen": 1698037760 + }, + { + "epoch": 0.51, + "learning_rate": 0.00039236108324974925, + "loss": 1.4222, + "theoretical_loss": 3.5067518943858893, + "tokens_seen": 1698103296 + }, + { + "epoch": 0.51, + "learning_rate": 0.00039234503510531595, + "loss": 1.401, + "theoretical_loss": 3.506740338642648, + "tokens_seen": 1698168832 + }, + { + "epoch": 0.51, + "learning_rate": 0.00039232898696088265, + "loss": 1.1663, + "theoretical_loss": 3.5067287834702223, + "tokens_seen": 1698234368 + }, + { + "epoch": 0.51, + "learning_rate": 0.00039231293881644935, + "loss": 1.3267, + "theoretical_loss": 3.506717228868563, + "tokens_seen": 1698299904 + }, + { + "epoch": 0.51, + "learning_rate": 0.00039229689067201605, + "loss": 1.3644, + "theoretical_loss": 3.506705674837619, + "tokens_seen": 1698365440 + }, + { + "epoch": 0.51, + "learning_rate": 0.00039228084252758275, + "loss": 1.4165, + "theoretical_loss": 3.5066941213773406, + "tokens_seen": 1698430976 + }, + { + "epoch": 0.51, + "learning_rate": 0.00039226479438314945, + "loss": 1.4106, + "theoretical_loss": 3.5066825684876775, + "tokens_seen": 1698496512 + }, + { + "epoch": 0.51, + "learning_rate": 0.00039224874623871615, + "loss": 1.3333, + "theoretical_loss": 3.5066710161685792, + "tokens_seen": 1698562048 + }, + { + "epoch": 0.51, + "learning_rate": 0.00039223269809428286, + "loss": 1.2895, + "theoretical_loss": 3.5066594644199958, + "tokens_seen": 1698627584 + }, + { + "epoch": 0.51, + "learning_rate": 0.00039221664994984956, + "loss": 1.3257, + "theoretical_loss": 3.506647913241877, + "tokens_seen": 1698693120 + }, + { + "epoch": 0.51, + "learning_rate": 0.00039220060180541626, + "loss": 1.3374, + "theoretical_loss": 3.5066363626341728, + "tokens_seen": 1698758656 + }, + { + "epoch": 0.51, + "learning_rate": 0.000392184553660983, + "loss": 1.3681, + "theoretical_loss": 3.506624812596833, + "tokens_seen": 1698824192 + }, + { + "epoch": 0.51, + "learning_rate": 0.0003921685055165497, + "loss": 1.3069, + "theoretical_loss": 3.5066132631298075, + "tokens_seen": 1698889728 + }, + { + "epoch": 0.51, + "learning_rate": 0.0003921524573721164, + "loss": 1.3811, + "theoretical_loss": 3.506601714233045, + "tokens_seen": 1698955264 + }, + { + "epoch": 0.51, + "learning_rate": 0.0003921364092276831, + "loss": 1.2748, + "theoretical_loss": 3.5065901659064975, + "tokens_seen": 1699020800 + }, + { + "epoch": 0.51, + "learning_rate": 0.0003921203610832498, + "loss": 1.3808, + "theoretical_loss": 3.5065786181501126, + "tokens_seen": 1699086336 + }, + { + "epoch": 0.51, + "learning_rate": 0.00039210431293881646, + "loss": 1.311, + "theoretical_loss": 3.5065670709638423, + "tokens_seen": 1699151872 + }, + { + "epoch": 0.51, + "learning_rate": 0.00039208826479438316, + "loss": 1.4506, + "theoretical_loss": 3.5065555243476347, + "tokens_seen": 1699217408 + }, + { + "epoch": 0.51, + "learning_rate": 0.00039207221664994986, + "loss": 1.3221, + "theoretical_loss": 3.5065439783014405, + "tokens_seen": 1699282944 + }, + { + "epoch": 0.51, + "learning_rate": 0.00039205616850551656, + "loss": 1.2901, + "theoretical_loss": 3.5065324328252094, + "tokens_seen": 1699348480 + }, + { + "epoch": 0.51, + "learning_rate": 0.00039204012036108326, + "loss": 1.3545, + "theoretical_loss": 3.5065208879188914, + "tokens_seen": 1699414016 + }, + { + "epoch": 0.51, + "learning_rate": 0.00039202407221664996, + "loss": 1.3607, + "theoretical_loss": 3.506509343582436, + "tokens_seen": 1699479552 + }, + { + "epoch": 0.52, + "learning_rate": 0.00039200802407221666, + "loss": 1.313, + "theoretical_loss": 3.506497799815794, + "tokens_seen": 1699545088 + }, + { + "epoch": 0.52, + "learning_rate": 0.00039199197592778336, + "loss": 1.3192, + "theoretical_loss": 3.5064862566189143, + "tokens_seen": 1699610624 + }, + { + "epoch": 0.52, + "learning_rate": 0.00039197592778335007, + "loss": 1.2951, + "theoretical_loss": 3.5064747139917474, + "tokens_seen": 1699676160 + }, + { + "epoch": 0.52, + "learning_rate": 0.00039195987963891677, + "loss": 1.4248, + "theoretical_loss": 3.506463171934243, + "tokens_seen": 1699741696 + }, + { + "epoch": 0.52, + "learning_rate": 0.00039194383149448347, + "loss": 1.3552, + "theoretical_loss": 3.5064516304463513, + "tokens_seen": 1699807232 + }, + { + "epoch": 0.52, + "learning_rate": 0.00039192778335005017, + "loss": 1.2343, + "theoretical_loss": 3.506440089528022, + "tokens_seen": 1699872768 + }, + { + "epoch": 0.52, + "learning_rate": 0.00039191173520561687, + "loss": 1.2681, + "theoretical_loss": 3.5064285491792044, + "tokens_seen": 1699938304 + }, + { + "epoch": 0.52, + "learning_rate": 0.00039189568706118357, + "loss": 1.2642, + "theoretical_loss": 3.5064170093998497, + "tokens_seen": 1700003840 + }, + { + "epoch": 0.52, + "learning_rate": 0.00039187963891675027, + "loss": 1.4146, + "theoretical_loss": 3.506405470189907, + "tokens_seen": 1700069376 + }, + { + "epoch": 0.52, + "learning_rate": 0.00039186359077231697, + "loss": 1.3741, + "theoretical_loss": 3.5063939315493267, + "tokens_seen": 1700134912 + }, + { + "epoch": 0.52, + "learning_rate": 0.00039184754262788367, + "loss": 1.3307, + "theoretical_loss": 3.5063823934780585, + "tokens_seen": 1700200448 + }, + { + "epoch": 0.52, + "learning_rate": 0.00039183149448345037, + "loss": 1.3878, + "theoretical_loss": 3.506370855976052, + "tokens_seen": 1700265984 + }, + { + "epoch": 0.52, + "learning_rate": 0.00039181544633901707, + "loss": 1.4009, + "theoretical_loss": 3.5063593190432574, + "tokens_seen": 1700331520 + }, + { + "epoch": 0.52, + "learning_rate": 0.00039179939819458377, + "loss": 1.3155, + "theoretical_loss": 3.506347782679626, + "tokens_seen": 1700397056 + }, + { + "epoch": 0.52, + "learning_rate": 0.0003917833500501505, + "loss": 1.2579, + "theoretical_loss": 3.5063362468851054, + "tokens_seen": 1700462592 + }, + { + "epoch": 0.52, + "learning_rate": 0.0003917673019057172, + "loss": 1.3633, + "theoretical_loss": 3.5063247116596474, + "tokens_seen": 1700528128 + }, + { + "epoch": 0.52, + "learning_rate": 0.0003917512537612839, + "loss": 1.3579, + "theoretical_loss": 3.506313177003202, + "tokens_seen": 1700593664 + }, + { + "epoch": 0.52, + "objective/train/docs_used": 621118, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.4528954029083252, + "objective/train/theoretical_loss": 3.5063016429157177, + "objective/train/tokens_used": 1721119200, + "theoretical_loss": 3.5063016429157177, + "tokens_seen": 1700659200 + }, + { + "epoch": 0.52, + "learning_rate": 0.0003917352056168506, + "loss": 1.32, + "theoretical_loss": 3.5063016429157177, + "tokens_seen": 1700659200 + }, + { + "epoch": 0.52, + "learning_rate": 0.0003917191574724173, + "loss": 1.3527, + "theoretical_loss": 3.506290109397146, + "tokens_seen": 1700724736 + }, + { + "epoch": 0.52, + "learning_rate": 0.000391703109327984, + "loss": 1.3717, + "theoretical_loss": 3.5062785764474365, + "tokens_seen": 1700790272 + }, + { + "epoch": 0.52, + "learning_rate": 0.0003916870611835507, + "loss": 1.3455, + "theoretical_loss": 3.506267044066538, + "tokens_seen": 1700855808 + }, + { + "epoch": 0.52, + "learning_rate": 0.0003916710130391174, + "loss": 1.3408, + "theoretical_loss": 3.506255512254403, + "tokens_seen": 1700921344 + }, + { + "epoch": 0.52, + "learning_rate": 0.0003916549648946841, + "loss": 1.291, + "theoretical_loss": 3.5062439810109796, + "tokens_seen": 1700986880 + }, + { + "epoch": 0.52, + "learning_rate": 0.0003916389167502508, + "loss": 1.2959, + "theoretical_loss": 3.5062324503362183, + "tokens_seen": 1701052416 + }, + { + "epoch": 0.52, + "learning_rate": 0.0003916228686058175, + "loss": 1.2965, + "theoretical_loss": 3.5062209202300694, + "tokens_seen": 1701117952 + }, + { + "epoch": 0.52, + "learning_rate": 0.0003916068204613842, + "loss": 1.4617, + "theoretical_loss": 3.506209390692483, + "tokens_seen": 1701183488 + }, + { + "epoch": 0.52, + "learning_rate": 0.0003915907723169509, + "loss": 1.4254, + "theoretical_loss": 3.5061978617234084, + "tokens_seen": 1701249024 + }, + { + "epoch": 0.52, + "learning_rate": 0.0003915747241725176, + "loss": 1.3076, + "theoretical_loss": 3.5061863333227965, + "tokens_seen": 1701314560 + }, + { + "epoch": 0.52, + "learning_rate": 0.0003915586760280843, + "loss": 1.3312, + "theoretical_loss": 3.5061748054905975, + "tokens_seen": 1701380096 + }, + { + "epoch": 0.52, + "learning_rate": 0.000391542627883651, + "loss": 1.3716, + "theoretical_loss": 3.5061632782267607, + "tokens_seen": 1701445632 + }, + { + "epoch": 0.52, + "learning_rate": 0.0003915265797392177, + "loss": 1.3188, + "theoretical_loss": 3.506151751531237, + "tokens_seen": 1701511168 + }, + { + "epoch": 0.52, + "learning_rate": 0.0003915105315947844, + "loss": 1.2638, + "theoretical_loss": 3.5061402254039757, + "tokens_seen": 1701576704 + }, + { + "epoch": 0.52, + "learning_rate": 0.0003914944834503511, + "loss": 1.2915, + "theoretical_loss": 3.5061286998449273, + "tokens_seen": 1701642240 + }, + { + "epoch": 0.52, + "learning_rate": 0.0003914784353059178, + "loss": 1.2714, + "theoretical_loss": 3.506117174854042, + "tokens_seen": 1701707776 + }, + { + "epoch": 0.52, + "learning_rate": 0.0003914623871614845, + "loss": 1.2884, + "theoretical_loss": 3.50610565043127, + "tokens_seen": 1701773312 + }, + { + "epoch": 0.52, + "learning_rate": 0.0003914463390170512, + "loss": 1.3628, + "theoretical_loss": 3.506094126576561, + "tokens_seen": 1701838848 + }, + { + "epoch": 0.52, + "learning_rate": 0.0003914302908726179, + "loss": 1.4306, + "theoretical_loss": 3.5060826032898653, + "tokens_seen": 1701904384 + }, + { + "epoch": 0.52, + "learning_rate": 0.0003914142427281846, + "loss": 1.348, + "theoretical_loss": 3.5060710805711333, + "tokens_seen": 1701969920 + }, + { + "epoch": 0.52, + "learning_rate": 0.0003913981945837513, + "loss": 1.3934, + "theoretical_loss": 3.506059558420314, + "tokens_seen": 1702035456 + }, + { + "epoch": 0.52, + "learning_rate": 0.000391382146439318, + "loss": 1.4358, + "theoretical_loss": 3.5060480368373597, + "tokens_seen": 1702100992 + }, + { + "epoch": 0.52, + "learning_rate": 0.0003913660982948847, + "loss": 1.3657, + "theoretical_loss": 3.5060365158222186, + "tokens_seen": 1702166528 + }, + { + "epoch": 0.52, + "learning_rate": 0.0003913500501504514, + "loss": 1.3684, + "theoretical_loss": 3.5060249953748417, + "tokens_seen": 1702232064 + }, + { + "epoch": 0.52, + "learning_rate": 0.00039133400200601804, + "loss": 1.2929, + "theoretical_loss": 3.506013475495179, + "tokens_seen": 1702297600 + }, + { + "epoch": 0.52, + "learning_rate": 0.00039131795386158474, + "loss": 1.3515, + "theoretical_loss": 3.506001956183181, + "tokens_seen": 1702363136 + }, + { + "epoch": 0.52, + "learning_rate": 0.00039130190571715144, + "loss": 1.3062, + "theoretical_loss": 3.5059904374387973, + "tokens_seen": 1702428672 + }, + { + "epoch": 0.52, + "learning_rate": 0.00039128585757271814, + "loss": 1.2889, + "theoretical_loss": 3.5059789192619784, + "tokens_seen": 1702494208 + }, + { + "epoch": 0.52, + "learning_rate": 0.00039126980942828484, + "loss": 1.3734, + "theoretical_loss": 3.5059674016526747, + "tokens_seen": 1702559744 + }, + { + "epoch": 0.52, + "learning_rate": 0.00039125376128385154, + "loss": 1.3554, + "theoretical_loss": 3.5059558846108354, + "tokens_seen": 1702625280 + }, + { + "epoch": 0.52, + "learning_rate": 0.00039123771313941824, + "loss": 1.3267, + "theoretical_loss": 3.5059443681364124, + "tokens_seen": 1702690816 + }, + { + "epoch": 0.52, + "learning_rate": 0.00039122166499498494, + "loss": 1.375, + "theoretical_loss": 3.5059328522293542, + "tokens_seen": 1702756352 + }, + { + "epoch": 0.52, + "learning_rate": 0.00039120561685055164, + "loss": 1.3548, + "theoretical_loss": 3.505921336889612, + "tokens_seen": 1702821888 + }, + { + "epoch": 0.52, + "learning_rate": 0.00039118956870611834, + "loss": 1.358, + "theoretical_loss": 3.505909822117136, + "tokens_seen": 1702887424 + }, + { + "epoch": 0.52, + "learning_rate": 0.00039117352056168504, + "loss": 1.3251, + "theoretical_loss": 3.5058983079118757, + "tokens_seen": 1702952960 + }, + { + "epoch": 0.52, + "learning_rate": 0.00039115747241725174, + "loss": 1.2334, + "theoretical_loss": 3.505886794273782, + "tokens_seen": 1703018496 + }, + { + "epoch": 0.52, + "learning_rate": 0.00039114142427281844, + "loss": 1.3138, + "theoretical_loss": 3.505875281202805, + "tokens_seen": 1703084032 + }, + { + "epoch": 0.52, + "learning_rate": 0.00039112537612838514, + "loss": 1.284, + "theoretical_loss": 3.5058637686988954, + "tokens_seen": 1703149568 + }, + { + "epoch": 0.52, + "learning_rate": 0.00039110932798395185, + "loss": 1.3556, + "theoretical_loss": 3.5058522567620023, + "tokens_seen": 1703215104 + }, + { + "epoch": 0.52, + "learning_rate": 0.00039109327983951855, + "loss": 1.2637, + "theoretical_loss": 3.505840745392077, + "tokens_seen": 1703280640 + }, + { + "epoch": 0.52, + "learning_rate": 0.00039107723169508525, + "loss": 1.4439, + "theoretical_loss": 3.5058292345890694, + "tokens_seen": 1703346176 + }, + { + "epoch": 0.52, + "learning_rate": 0.00039106118355065195, + "loss": 1.27, + "theoretical_loss": 3.505817724352929, + "tokens_seen": 1703411712 + }, + { + "epoch": 0.52, + "learning_rate": 0.00039104513540621865, + "loss": 1.3364, + "theoretical_loss": 3.505806214683608, + "tokens_seen": 1703477248 + }, + { + "epoch": 0.52, + "learning_rate": 0.00039102908726178535, + "loss": 1.4265, + "theoretical_loss": 3.5057947055810548, + "tokens_seen": 1703542784 + }, + { + "epoch": 0.52, + "learning_rate": 0.00039101303911735205, + "loss": 1.2817, + "theoretical_loss": 3.5057831970452207, + "tokens_seen": 1703608320 + }, + { + "epoch": 0.52, + "learning_rate": 0.00039099699097291875, + "loss": 1.4572, + "theoretical_loss": 3.5057716890760555, + "tokens_seen": 1703673856 + }, + { + "epoch": 0.52, + "learning_rate": 0.0003909809428284855, + "loss": 1.3606, + "theoretical_loss": 3.5057601816735096, + "tokens_seen": 1703739392 + }, + { + "epoch": 0.52, + "learning_rate": 0.0003909648946840522, + "loss": 1.2142, + "theoretical_loss": 3.505748674837534, + "tokens_seen": 1703804928 + }, + { + "epoch": 0.52, + "learning_rate": 0.0003909488465396189, + "loss": 1.4811, + "theoretical_loss": 3.5057371685680776, + "tokens_seen": 1703870464 + }, + { + "epoch": 0.52, + "objective/train/docs_used": 622235, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.3871040344238281, + "objective/train/theoretical_loss": 3.505725662865092, + "objective/train/tokens_used": 1724396000, + "theoretical_loss": 3.505725662865092, + "tokens_seen": 1703936000 + }, + { + "epoch": 0.52, + "learning_rate": 0.0003909327983951856, + "loss": 1.3621, + "theoretical_loss": 3.505725662865092, + "tokens_seen": 1703936000 + }, + { + "epoch": 0.52, + "learning_rate": 0.0003909167502507523, + "loss": 1.2636, + "theoretical_loss": 3.5057141577285273, + "tokens_seen": 1704001536 + }, + { + "epoch": 0.52, + "learning_rate": 0.000390900702106319, + "loss": 1.3439, + "theoretical_loss": 3.505702653158333, + "tokens_seen": 1704067072 + }, + { + "epoch": 0.52, + "learning_rate": 0.0003908846539618857, + "loss": 1.269, + "theoretical_loss": 3.5056911491544605, + "tokens_seen": 1704132608 + }, + { + "epoch": 0.52, + "learning_rate": 0.0003908686058174524, + "loss": 1.3738, + "theoretical_loss": 3.5056796457168593, + "tokens_seen": 1704198144 + }, + { + "epoch": 0.52, + "learning_rate": 0.0003908525576730191, + "loss": 1.338, + "theoretical_loss": 3.5056681428454803, + "tokens_seen": 1704263680 + }, + { + "epoch": 0.52, + "learning_rate": 0.0003908365095285858, + "loss": 1.3396, + "theoretical_loss": 3.5056566405402743, + "tokens_seen": 1704329216 + }, + { + "epoch": 0.52, + "learning_rate": 0.0003908204613841525, + "loss": 1.3277, + "theoretical_loss": 3.5056451388011904, + "tokens_seen": 1704394752 + }, + { + "epoch": 0.52, + "learning_rate": 0.0003908044132397192, + "loss": 1.3222, + "theoretical_loss": 3.5056336376281796, + "tokens_seen": 1704460288 + }, + { + "epoch": 0.52, + "learning_rate": 0.0003907883650952859, + "loss": 1.3987, + "theoretical_loss": 3.5056221370211924, + "tokens_seen": 1704525824 + }, + { + "epoch": 0.52, + "learning_rate": 0.0003907723169508526, + "loss": 1.342, + "theoretical_loss": 3.505610636980179, + "tokens_seen": 1704591360 + }, + { + "epoch": 0.52, + "learning_rate": 0.0003907562688064193, + "loss": 1.3302, + "theoretical_loss": 3.50559913750509, + "tokens_seen": 1704656896 + }, + { + "epoch": 0.52, + "learning_rate": 0.000390740220661986, + "loss": 1.2966, + "theoretical_loss": 3.505587638595876, + "tokens_seen": 1704722432 + }, + { + "epoch": 0.52, + "learning_rate": 0.0003907241725175527, + "loss": 1.3088, + "theoretical_loss": 3.5055761402524865, + "tokens_seen": 1704787968 + }, + { + "epoch": 0.52, + "learning_rate": 0.0003907081243731194, + "loss": 1.3458, + "theoretical_loss": 3.5055646424748725, + "tokens_seen": 1704853504 + }, + { + "epoch": 0.52, + "learning_rate": 0.00039069207622868606, + "loss": 1.3916, + "theoretical_loss": 3.5055531452629847, + "tokens_seen": 1704919040 + }, + { + "epoch": 0.52, + "learning_rate": 0.00039067602808425276, + "loss": 1.3159, + "theoretical_loss": 3.505541648616773, + "tokens_seen": 1704984576 + }, + { + "epoch": 0.52, + "learning_rate": 0.00039065997993981946, + "loss": 1.3357, + "theoretical_loss": 3.5055301525361884, + "tokens_seen": 1705050112 + }, + { + "epoch": 0.52, + "learning_rate": 0.00039064393179538616, + "loss": 1.2855, + "theoretical_loss": 3.5055186570211805, + "tokens_seen": 1705115648 + }, + { + "epoch": 0.52, + "learning_rate": 0.00039062788365095286, + "loss": 1.3262, + "theoretical_loss": 3.5055071620717007, + "tokens_seen": 1705181184 + }, + { + "epoch": 0.52, + "learning_rate": 0.00039061183550651957, + "loss": 1.4282, + "theoretical_loss": 3.5054956676876987, + "tokens_seen": 1705246720 + }, + { + "epoch": 0.52, + "learning_rate": 0.00039059578736208627, + "loss": 1.2702, + "theoretical_loss": 3.505484173869125, + "tokens_seen": 1705312256 + }, + { + "epoch": 0.52, + "learning_rate": 0.00039057973921765297, + "loss": 1.4513, + "theoretical_loss": 3.5054726806159304, + "tokens_seen": 1705377792 + }, + { + "epoch": 0.52, + "learning_rate": 0.00039056369107321967, + "loss": 1.3506, + "theoretical_loss": 3.505461187928065, + "tokens_seen": 1705443328 + }, + { + "epoch": 0.52, + "learning_rate": 0.00039054764292878637, + "loss": 1.3822, + "theoretical_loss": 3.50544969580548, + "tokens_seen": 1705508864 + }, + { + "epoch": 0.52, + "learning_rate": 0.00039053159478435307, + "loss": 1.441, + "theoretical_loss": 3.505438204248125, + "tokens_seen": 1705574400 + }, + { + "epoch": 0.52, + "learning_rate": 0.00039051554663991977, + "loss": 1.4191, + "theoretical_loss": 3.505426713255951, + "tokens_seen": 1705639936 + }, + { + "epoch": 0.52, + "learning_rate": 0.00039049949849548647, + "loss": 1.3126, + "theoretical_loss": 3.5054152228289084, + "tokens_seen": 1705705472 + }, + { + "epoch": 0.52, + "learning_rate": 0.00039048345035105317, + "loss": 1.2186, + "theoretical_loss": 3.5054037329669474, + "tokens_seen": 1705771008 + }, + { + "epoch": 0.52, + "learning_rate": 0.00039046740220661987, + "loss": 1.3507, + "theoretical_loss": 3.505392243670019, + "tokens_seen": 1705836544 + }, + { + "epoch": 0.52, + "learning_rate": 0.00039045135406218657, + "loss": 1.5354, + "theoretical_loss": 3.5053807549380736, + "tokens_seen": 1705902080 + }, + { + "epoch": 0.52, + "learning_rate": 0.00039043530591775327, + "loss": 1.4243, + "theoretical_loss": 3.5053692667710608, + "tokens_seen": 1705967616 + }, + { + "epoch": 0.52, + "learning_rate": 0.00039041925777331997, + "loss": 1.2748, + "theoretical_loss": 3.5053577791689325, + "tokens_seen": 1706033152 + }, + { + "epoch": 0.52, + "learning_rate": 0.0003904032096288867, + "loss": 1.3404, + "theoretical_loss": 3.5053462921316383, + "tokens_seen": 1706098688 + }, + { + "epoch": 0.52, + "learning_rate": 0.0003903871614844534, + "loss": 1.4242, + "theoretical_loss": 3.505334805659129, + "tokens_seen": 1706164224 + }, + { + "epoch": 0.52, + "learning_rate": 0.0003903711133400201, + "loss": 1.3525, + "theoretical_loss": 3.505323319751356, + "tokens_seen": 1706229760 + }, + { + "epoch": 0.52, + "learning_rate": 0.0003903550651955868, + "loss": 1.2513, + "theoretical_loss": 3.505311834408268, + "tokens_seen": 1706295296 + }, + { + "epoch": 0.52, + "learning_rate": 0.0003903390170511535, + "loss": 1.3084, + "theoretical_loss": 3.505300349629817, + "tokens_seen": 1706360832 + }, + { + "epoch": 0.52, + "learning_rate": 0.0003903229689067202, + "loss": 1.3733, + "theoretical_loss": 3.5052888654159533, + "tokens_seen": 1706426368 + }, + { + "epoch": 0.52, + "learning_rate": 0.0003903069207622869, + "loss": 1.3151, + "theoretical_loss": 3.5052773817666267, + "tokens_seen": 1706491904 + }, + { + "epoch": 0.52, + "learning_rate": 0.0003902908726178536, + "loss": 1.2903, + "theoretical_loss": 3.505265898681789, + "tokens_seen": 1706557440 + }, + { + "epoch": 0.52, + "learning_rate": 0.0003902748244734203, + "loss": 1.4208, + "theoretical_loss": 3.5052544161613897, + "tokens_seen": 1706622976 + }, + { + "epoch": 0.52, + "learning_rate": 0.000390258776328987, + "loss": 1.4214, + "theoretical_loss": 3.50524293420538, + "tokens_seen": 1706688512 + }, + { + "epoch": 0.52, + "learning_rate": 0.0003902427281845537, + "loss": 1.2919, + "theoretical_loss": 3.5052314528137103, + "tokens_seen": 1706754048 + }, + { + "epoch": 0.52, + "learning_rate": 0.0003902266800401204, + "loss": 1.337, + "theoretical_loss": 3.5052199719863317, + "tokens_seen": 1706819584 + }, + { + "epoch": 0.52, + "learning_rate": 0.0003902106318956871, + "loss": 1.3624, + "theoretical_loss": 3.5052084917231934, + "tokens_seen": 1706885120 + }, + { + "epoch": 0.52, + "learning_rate": 0.0003901945837512538, + "loss": 1.4042, + "theoretical_loss": 3.5051970120242473, + "tokens_seen": 1706950656 + }, + { + "epoch": 0.52, + "learning_rate": 0.0003901785356068205, + "loss": 1.3317, + "theoretical_loss": 3.5051855328894437, + "tokens_seen": 1707016192 + }, + { + "epoch": 0.52, + "learning_rate": 0.0003901624874623872, + "loss": 1.3859, + "theoretical_loss": 3.5051740543187333, + "tokens_seen": 1707081728 + }, + { + "epoch": 0.52, + "learning_rate": 0.0003901464393179539, + "loss": 1.2829, + "theoretical_loss": 3.5051625763120664, + "tokens_seen": 1707147264 + }, + { + "epoch": 0.52, + "objective/train/docs_used": 623332, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.4871866703033447, + "objective/train/theoretical_loss": 3.5051510988693937, + "objective/train/tokens_used": 1727672800, + "theoretical_loss": 3.5051510988693937, + "tokens_seen": 1707212800 + }, + { + "epoch": 0.52, + "learning_rate": 0.0003901303911735206, + "loss": 1.4743, + "theoretical_loss": 3.5051510988693937, + "tokens_seen": 1707212800 + }, + { + "epoch": 0.52, + "learning_rate": 0.0003901143430290873, + "loss": 1.248, + "theoretical_loss": 3.505139621990666, + "tokens_seen": 1707278336 + }, + { + "epoch": 0.52, + "learning_rate": 0.000390098294884654, + "loss": 1.2691, + "theoretical_loss": 3.505128145675834, + "tokens_seen": 1707343872 + }, + { + "epoch": 0.52, + "learning_rate": 0.0003900822467402207, + "loss": 1.2674, + "theoretical_loss": 3.505116669924848, + "tokens_seen": 1707409408 + }, + { + "epoch": 0.52, + "learning_rate": 0.0003900661985957874, + "loss": 1.2854, + "theoretical_loss": 3.5051051947376592, + "tokens_seen": 1707474944 + }, + { + "epoch": 0.52, + "learning_rate": 0.0003900501504513541, + "loss": 1.277, + "theoretical_loss": 3.5050937201142176, + "tokens_seen": 1707540480 + }, + { + "epoch": 0.52, + "learning_rate": 0.0003900341023069208, + "loss": 1.3205, + "theoretical_loss": 3.5050822460544744, + "tokens_seen": 1707606016 + }, + { + "epoch": 0.52, + "learning_rate": 0.0003900180541624875, + "loss": 1.2811, + "theoretical_loss": 3.5050707725583803, + "tokens_seen": 1707671552 + }, + { + "epoch": 0.52, + "learning_rate": 0.0003900020060180542, + "loss": 1.3902, + "theoretical_loss": 3.5050592996258856, + "tokens_seen": 1707737088 + }, + { + "epoch": 0.52, + "learning_rate": 0.0003899859578736209, + "loss": 1.2551, + "theoretical_loss": 3.5050478272569414, + "tokens_seen": 1707802624 + }, + { + "epoch": 0.52, + "learning_rate": 0.0003899699097291876, + "loss": 1.3978, + "theoretical_loss": 3.505036355451498, + "tokens_seen": 1707868160 + }, + { + "epoch": 0.52, + "learning_rate": 0.0003899538615847543, + "loss": 1.3589, + "theoretical_loss": 3.505024884209506, + "tokens_seen": 1707933696 + }, + { + "epoch": 0.52, + "learning_rate": 0.000389937813440321, + "loss": 1.4254, + "theoretical_loss": 3.5050134135309166, + "tokens_seen": 1707999232 + }, + { + "epoch": 0.52, + "learning_rate": 0.00038992176529588764, + "loss": 1.3262, + "theoretical_loss": 3.50500194341568, + "tokens_seen": 1708064768 + }, + { + "epoch": 0.52, + "learning_rate": 0.00038990571715145434, + "loss": 1.3055, + "theoretical_loss": 3.5049904738637476, + "tokens_seen": 1708130304 + }, + { + "epoch": 0.52, + "learning_rate": 0.00038988966900702104, + "loss": 1.187, + "theoretical_loss": 3.5049790048750697, + "tokens_seen": 1708195840 + }, + { + "epoch": 0.52, + "learning_rate": 0.00038987362086258774, + "loss": 1.3506, + "theoretical_loss": 3.504967536449597, + "tokens_seen": 1708261376 + }, + { + "epoch": 0.52, + "learning_rate": 0.00038985757271815444, + "loss": 1.3042, + "theoretical_loss": 3.50495606858728, + "tokens_seen": 1708326912 + }, + { + "epoch": 0.52, + "learning_rate": 0.00038984152457372114, + "loss": 1.326, + "theoretical_loss": 3.50494460128807, + "tokens_seen": 1708392448 + }, + { + "epoch": 0.52, + "learning_rate": 0.00038982547642928784, + "loss": 1.3283, + "theoretical_loss": 3.5049331345519175, + "tokens_seen": 1708457984 + }, + { + "epoch": 0.52, + "learning_rate": 0.00038980942828485454, + "loss": 1.3451, + "theoretical_loss": 3.504921668378773, + "tokens_seen": 1708523520 + }, + { + "epoch": 0.52, + "learning_rate": 0.00038979338014042124, + "loss": 1.3675, + "theoretical_loss": 3.504910202768588, + "tokens_seen": 1708589056 + }, + { + "epoch": 0.52, + "learning_rate": 0.000389777331995988, + "loss": 1.3275, + "theoretical_loss": 3.5048987377213123, + "tokens_seen": 1708654592 + }, + { + "epoch": 0.52, + "learning_rate": 0.0003897612838515547, + "loss": 1.3069, + "theoretical_loss": 3.5048872732368976, + "tokens_seen": 1708720128 + }, + { + "epoch": 0.52, + "learning_rate": 0.0003897452357071214, + "loss": 1.3311, + "theoretical_loss": 3.504875809315294, + "tokens_seen": 1708785664 + }, + { + "epoch": 0.52, + "learning_rate": 0.0003897291875626881, + "loss": 1.5028, + "theoretical_loss": 3.5048643459564524, + "tokens_seen": 1708851200 + }, + { + "epoch": 0.52, + "learning_rate": 0.0003897131394182548, + "loss": 1.3897, + "theoretical_loss": 3.5048528831603236, + "tokens_seen": 1708916736 + }, + { + "epoch": 0.52, + "learning_rate": 0.0003896970912738215, + "loss": 1.3628, + "theoretical_loss": 3.5048414209268586, + "tokens_seen": 1708982272 + }, + { + "epoch": 0.52, + "learning_rate": 0.0003896810431293882, + "loss": 1.2562, + "theoretical_loss": 3.504829959256009, + "tokens_seen": 1709047808 + }, + { + "epoch": 0.52, + "learning_rate": 0.0003896649949849549, + "loss": 1.3433, + "theoretical_loss": 3.5048184981477233, + "tokens_seen": 1709113344 + }, + { + "epoch": 0.52, + "learning_rate": 0.0003896489468405216, + "loss": 1.3819, + "theoretical_loss": 3.504807037601955, + "tokens_seen": 1709178880 + }, + { + "epoch": 0.52, + "learning_rate": 0.0003896328986960883, + "loss": 1.4073, + "theoretical_loss": 3.504795577618653, + "tokens_seen": 1709244416 + }, + { + "epoch": 0.52, + "learning_rate": 0.000389616850551655, + "loss": 1.307, + "theoretical_loss": 3.504784118197769, + "tokens_seen": 1709309952 + }, + { + "epoch": 0.52, + "learning_rate": 0.0003896008024072217, + "loss": 1.2988, + "theoretical_loss": 3.5047726593392534, + "tokens_seen": 1709375488 + }, + { + "epoch": 0.52, + "learning_rate": 0.0003895847542627884, + "loss": 1.5243, + "theoretical_loss": 3.504761201043057, + "tokens_seen": 1709441024 + }, + { + "epoch": 0.52, + "learning_rate": 0.0003895687061183551, + "loss": 1.3884, + "theoretical_loss": 3.5047497433091315, + "tokens_seen": 1709506560 + }, + { + "epoch": 0.52, + "learning_rate": 0.0003895526579739218, + "loss": 1.3431, + "theoretical_loss": 3.504738286137427, + "tokens_seen": 1709572096 + }, + { + "epoch": 0.52, + "learning_rate": 0.0003895366098294885, + "loss": 1.3548, + "theoretical_loss": 3.5047268295278946, + "tokens_seen": 1709637632 + }, + { + "epoch": 0.52, + "learning_rate": 0.0003895205616850552, + "loss": 1.2631, + "theoretical_loss": 3.504715373480485, + "tokens_seen": 1709703168 + }, + { + "epoch": 0.52, + "learning_rate": 0.0003895045135406219, + "loss": 1.4419, + "theoretical_loss": 3.5047039179951494, + "tokens_seen": 1709768704 + }, + { + "epoch": 0.52, + "learning_rate": 0.0003894884653961886, + "loss": 1.4074, + "theoretical_loss": 3.504692463071838, + "tokens_seen": 1709834240 + }, + { + "epoch": 0.52, + "learning_rate": 0.0003894724172517553, + "loss": 1.3328, + "theoretical_loss": 3.5046810087105027, + "tokens_seen": 1709899776 + }, + { + "epoch": 0.52, + "learning_rate": 0.000389456369107322, + "loss": 1.2852, + "theoretical_loss": 3.5046695549110938, + "tokens_seen": 1709965312 + }, + { + "epoch": 0.52, + "learning_rate": 0.0003894403209628887, + "loss": 1.427, + "theoretical_loss": 3.504658101673562, + "tokens_seen": 1710030848 + }, + { + "epoch": 0.52, + "learning_rate": 0.0003894242728184554, + "loss": 1.2852, + "theoretical_loss": 3.5046466489978583, + "tokens_seen": 1710096384 + }, + { + "epoch": 0.52, + "learning_rate": 0.0003894082246740221, + "loss": 1.35, + "theoretical_loss": 3.504635196883934, + "tokens_seen": 1710161920 + }, + { + "epoch": 0.52, + "learning_rate": 0.0003893921765295888, + "loss": 1.2597, + "theoretical_loss": 3.50462374533174, + "tokens_seen": 1710227456 + }, + { + "epoch": 0.52, + "learning_rate": 0.0003893761283851555, + "loss": 1.4008, + "theoretical_loss": 3.504612294341227, + "tokens_seen": 1710292992 + }, + { + "epoch": 0.52, + "learning_rate": 0.0003893600802407222, + "loss": 1.3536, + "theoretical_loss": 3.5046008439123453, + "tokens_seen": 1710358528 + }, + { + "epoch": 0.52, + "learning_rate": 0.0003893440320962889, + "loss": 1.3399, + "theoretical_loss": 3.5045893940450465, + "tokens_seen": 1710424064 + }, + { + "epoch": 0.52, + "objective/train/docs_used": 624643, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.4083890914916992, + "objective/train/theoretical_loss": 3.504577944739282, + "objective/train/tokens_used": 1730949600, + "theoretical_loss": 3.504577944739282, + "tokens_seen": 1710489600 + }, + { + "epoch": 0.52, + "learning_rate": 0.0003893279839518556, + "loss": 1.353, + "theoretical_loss": 3.504577944739282, + "tokens_seen": 1710489600 + }, + { + "epoch": 0.52, + "learning_rate": 0.0003893119358074223, + "loss": 1.3353, + "theoretical_loss": 3.504566495995002, + "tokens_seen": 1710555136 + }, + { + "epoch": 0.52, + "learning_rate": 0.00038929588766298896, + "loss": 1.41, + "theoretical_loss": 3.504555047812158, + "tokens_seen": 1710620672 + }, + { + "epoch": 0.52, + "learning_rate": 0.00038927983951855566, + "loss": 1.2827, + "theoretical_loss": 3.5045436001907007, + "tokens_seen": 1710686208 + }, + { + "epoch": 0.52, + "learning_rate": 0.00038926379137412236, + "loss": 1.4545, + "theoretical_loss": 3.5045321531305804, + "tokens_seen": 1710751744 + }, + { + "epoch": 0.52, + "learning_rate": 0.00038924774322968906, + "loss": 1.284, + "theoretical_loss": 3.5045207066317494, + "tokens_seen": 1710817280 + }, + { + "epoch": 0.52, + "learning_rate": 0.00038923169508525577, + "loss": 1.4543, + "theoretical_loss": 3.5045092606941575, + "tokens_seen": 1710882816 + }, + { + "epoch": 0.52, + "learning_rate": 0.00038921564694082247, + "loss": 1.1828, + "theoretical_loss": 3.5044978153177566, + "tokens_seen": 1710948352 + }, + { + "epoch": 0.52, + "learning_rate": 0.00038919959879638917, + "loss": 1.4264, + "theoretical_loss": 3.504486370502497, + "tokens_seen": 1711013888 + }, + { + "epoch": 0.52, + "learning_rate": 0.00038918355065195587, + "loss": 1.4001, + "theoretical_loss": 3.5044749262483297, + "tokens_seen": 1711079424 + }, + { + "epoch": 0.52, + "learning_rate": 0.00038916750250752257, + "loss": 1.3734, + "theoretical_loss": 3.504463482555207, + "tokens_seen": 1711144960 + }, + { + "epoch": 0.52, + "learning_rate": 0.00038915145436308927, + "loss": 1.338, + "theoretical_loss": 3.504452039423078, + "tokens_seen": 1711210496 + }, + { + "epoch": 0.52, + "learning_rate": 0.00038913540621865597, + "loss": 1.387, + "theoretical_loss": 3.5044405968518944, + "tokens_seen": 1711276032 + }, + { + "epoch": 0.52, + "learning_rate": 0.00038911935807422267, + "loss": 1.4008, + "theoretical_loss": 3.5044291548416076, + "tokens_seen": 1711341568 + }, + { + "epoch": 0.52, + "learning_rate": 0.00038910330992978937, + "loss": 1.2721, + "theoretical_loss": 3.5044177133921686, + "tokens_seen": 1711407104 + }, + { + "epoch": 0.52, + "learning_rate": 0.00038908726178535607, + "loss": 1.2871, + "theoretical_loss": 3.504406272503528, + "tokens_seen": 1711472640 + }, + { + "epoch": 0.52, + "learning_rate": 0.00038907121364092277, + "loss": 1.1646, + "theoretical_loss": 3.5043948321756373, + "tokens_seen": 1711538176 + }, + { + "epoch": 0.52, + "learning_rate": 0.00038905516549648947, + "loss": 1.3117, + "theoretical_loss": 3.5043833924084473, + "tokens_seen": 1711603712 + }, + { + "epoch": 0.52, + "learning_rate": 0.0003890391173520562, + "loss": 1.3371, + "theoretical_loss": 3.5043719532019093, + "tokens_seen": 1711669248 + }, + { + "epoch": 0.52, + "learning_rate": 0.0003890230692076229, + "loss": 1.4644, + "theoretical_loss": 3.504360514555974, + "tokens_seen": 1711734784 + }, + { + "epoch": 0.52, + "learning_rate": 0.0003890070210631896, + "loss": 1.3371, + "theoretical_loss": 3.504349076470593, + "tokens_seen": 1711800320 + }, + { + "epoch": 0.52, + "learning_rate": 0.0003889909729187563, + "loss": 1.4698, + "theoretical_loss": 3.5043376389457164, + "tokens_seen": 1711865856 + }, + { + "epoch": 0.52, + "learning_rate": 0.000388974924774323, + "loss": 1.2971, + "theoretical_loss": 3.504326201981296, + "tokens_seen": 1711931392 + }, + { + "epoch": 0.52, + "learning_rate": 0.0003889588766298897, + "loss": 1.4026, + "theoretical_loss": 3.504314765577283, + "tokens_seen": 1711996928 + }, + { + "epoch": 0.52, + "learning_rate": 0.0003889428284854564, + "loss": 1.402, + "theoretical_loss": 3.504303329733628, + "tokens_seen": 1712062464 + }, + { + "epoch": 0.52, + "learning_rate": 0.0003889267803410231, + "loss": 1.4784, + "theoretical_loss": 3.5042918944502826, + "tokens_seen": 1712128000 + }, + { + "epoch": 0.52, + "learning_rate": 0.0003889107321965898, + "loss": 1.3211, + "theoretical_loss": 3.5042804597271973, + "tokens_seen": 1712193536 + }, + { + "epoch": 0.52, + "learning_rate": 0.0003888946840521565, + "loss": 1.3187, + "theoretical_loss": 3.504269025564324, + "tokens_seen": 1712259072 + }, + { + "epoch": 0.52, + "learning_rate": 0.0003888786359077232, + "loss": 1.3258, + "theoretical_loss": 3.504257591961613, + "tokens_seen": 1712324608 + }, + { + "epoch": 0.52, + "learning_rate": 0.0003888625877632899, + "loss": 1.2362, + "theoretical_loss": 3.504246158919016, + "tokens_seen": 1712390144 + }, + { + "epoch": 0.52, + "learning_rate": 0.0003888465396188566, + "loss": 1.301, + "theoretical_loss": 3.5042347264364837, + "tokens_seen": 1712455680 + }, + { + "epoch": 0.52, + "learning_rate": 0.0003888304914744233, + "loss": 1.3799, + "theoretical_loss": 3.5042232945139675, + "tokens_seen": 1712521216 + }, + { + "epoch": 0.52, + "learning_rate": 0.00038881444332999, + "loss": 1.4491, + "theoretical_loss": 3.5042118631514185, + "tokens_seen": 1712586752 + }, + { + "epoch": 0.52, + "learning_rate": 0.0003887983951855567, + "loss": 1.4369, + "theoretical_loss": 3.504200432348788, + "tokens_seen": 1712652288 + }, + { + "epoch": 0.52, + "learning_rate": 0.0003887823470411234, + "loss": 1.3924, + "theoretical_loss": 3.5041890021060267, + "tokens_seen": 1712717824 + }, + { + "epoch": 0.52, + "learning_rate": 0.0003887662988966901, + "loss": 1.4086, + "theoretical_loss": 3.504177572423086, + "tokens_seen": 1712783360 + }, + { + "epoch": 0.52, + "learning_rate": 0.0003887502507522568, + "loss": 1.3237, + "theoretical_loss": 3.5041661432999174, + "tokens_seen": 1712848896 + }, + { + "epoch": 0.52, + "learning_rate": 0.0003887342026078235, + "loss": 1.4023, + "theoretical_loss": 3.504154714736472, + "tokens_seen": 1712914432 + }, + { + "epoch": 0.52, + "learning_rate": 0.0003887181544633902, + "loss": 1.4098, + "theoretical_loss": 3.5041432867327007, + "tokens_seen": 1712979968 + }, + { + "epoch": 0.52, + "learning_rate": 0.0003887021063189569, + "loss": 1.3002, + "theoretical_loss": 3.504131859288554, + "tokens_seen": 1713045504 + }, + { + "epoch": 0.52, + "learning_rate": 0.0003886860581745236, + "loss": 1.3496, + "theoretical_loss": 3.504120432403984, + "tokens_seen": 1713111040 + }, + { + "epoch": 0.52, + "learning_rate": 0.0003886700100300903, + "loss": 1.366, + "theoretical_loss": 3.5041090060789424, + "tokens_seen": 1713176576 + }, + { + "epoch": 0.52, + "learning_rate": 0.000388653961885657, + "loss": 1.3495, + "theoretical_loss": 3.504097580313379, + "tokens_seen": 1713242112 + }, + { + "epoch": 0.52, + "learning_rate": 0.0003886379137412237, + "loss": 1.4581, + "theoretical_loss": 3.504086155107246, + "tokens_seen": 1713307648 + }, + { + "epoch": 0.52, + "learning_rate": 0.0003886218655967904, + "loss": 1.3672, + "theoretical_loss": 3.504074730460494, + "tokens_seen": 1713373184 + }, + { + "epoch": 0.52, + "learning_rate": 0.0003886058174523571, + "loss": 1.3847, + "theoretical_loss": 3.504063306373075, + "tokens_seen": 1713438720 + }, + { + "epoch": 0.52, + "learning_rate": 0.0003885897693079238, + "loss": 1.3311, + "theoretical_loss": 3.50405188284494, + "tokens_seen": 1713504256 + }, + { + "epoch": 0.52, + "learning_rate": 0.0003885737211634905, + "loss": 1.2846, + "theoretical_loss": 3.5040404598760393, + "tokens_seen": 1713569792 + }, + { + "epoch": 0.52, + "learning_rate": 0.0003885576730190572, + "loss": 1.2712, + "theoretical_loss": 3.504029037466325, + "tokens_seen": 1713635328 + }, + { + "epoch": 0.52, + "learning_rate": 0.0003885416248746239, + "loss": 1.4955, + "theoretical_loss": 3.504017615615748, + "tokens_seen": 1713700864 + }, + { + "epoch": 0.52, + "objective/train/docs_used": 625897, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.2722280025482178, + "objective/train/theoretical_loss": 3.50400619432426, + "objective/train/tokens_used": 1734226400, + "theoretical_loss": 3.50400619432426, + "tokens_seen": 1713766400 + }, + { + "epoch": 0.52, + "learning_rate": 0.0003885255767301906, + "loss": 1.2759, + "theoretical_loss": 3.50400619432426, + "tokens_seen": 1713766400 + }, + { + "epoch": 0.52, + "learning_rate": 0.0003885095285857573, + "loss": 1.3165, + "theoretical_loss": 3.5039947735918124, + "tokens_seen": 1713831936 + }, + { + "epoch": 0.52, + "learning_rate": 0.000388493480441324, + "loss": 1.3419, + "theoretical_loss": 3.5039833534183553, + "tokens_seen": 1713897472 + }, + { + "epoch": 0.52, + "learning_rate": 0.0003884774322968907, + "loss": 1.4077, + "theoretical_loss": 3.5039719338038413, + "tokens_seen": 1713963008 + }, + { + "epoch": 0.52, + "learning_rate": 0.0003884613841524574, + "loss": 1.2323, + "theoretical_loss": 3.5039605147482202, + "tokens_seen": 1714028544 + }, + { + "epoch": 0.52, + "learning_rate": 0.0003884453360080241, + "loss": 1.282, + "theoretical_loss": 3.503949096251445, + "tokens_seen": 1714094080 + }, + { + "epoch": 0.52, + "learning_rate": 0.0003884292878635908, + "loss": 1.4219, + "theoretical_loss": 3.5039376783134655, + "tokens_seen": 1714159616 + }, + { + "epoch": 0.52, + "learning_rate": 0.0003884132397191575, + "loss": 1.319, + "theoretical_loss": 3.503926260934234, + "tokens_seen": 1714225152 + }, + { + "epoch": 0.52, + "learning_rate": 0.0003883971915747242, + "loss": 1.2773, + "theoretical_loss": 3.5039148441137016, + "tokens_seen": 1714290688 + }, + { + "epoch": 0.52, + "learning_rate": 0.0003883811434302909, + "loss": 1.3964, + "theoretical_loss": 3.503903427851819, + "tokens_seen": 1714356224 + }, + { + "epoch": 0.52, + "learning_rate": 0.0003883650952858576, + "loss": 1.3376, + "theoretical_loss": 3.5038920121485386, + "tokens_seen": 1714421760 + }, + { + "epoch": 0.52, + "learning_rate": 0.0003883490471414243, + "loss": 1.3767, + "theoretical_loss": 3.5038805970038105, + "tokens_seen": 1714487296 + }, + { + "epoch": 0.52, + "learning_rate": 0.000388332998996991, + "loss": 1.4112, + "theoretical_loss": 3.5038691824175867, + "tokens_seen": 1714552832 + }, + { + "epoch": 0.52, + "learning_rate": 0.0003883169508525577, + "loss": 1.3371, + "theoretical_loss": 3.5038577683898184, + "tokens_seen": 1714618368 + }, + { + "epoch": 0.52, + "learning_rate": 0.0003883009027081244, + "loss": 1.2672, + "theoretical_loss": 3.503846354920457, + "tokens_seen": 1714683904 + }, + { + "epoch": 0.52, + "learning_rate": 0.0003882848545636911, + "loss": 1.3771, + "theoretical_loss": 3.503834942009454, + "tokens_seen": 1714749440 + }, + { + "epoch": 0.52, + "learning_rate": 0.0003882688064192578, + "loss": 1.4666, + "theoretical_loss": 3.50382352965676, + "tokens_seen": 1714814976 + }, + { + "epoch": 0.52, + "learning_rate": 0.0003882527582748245, + "loss": 1.4306, + "theoretical_loss": 3.5038121178623274, + "tokens_seen": 1714880512 + }, + { + "epoch": 0.52, + "learning_rate": 0.0003882367101303912, + "loss": 1.2987, + "theoretical_loss": 3.503800706626107, + "tokens_seen": 1714946048 + }, + { + "epoch": 0.52, + "learning_rate": 0.0003882206619859579, + "loss": 1.3505, + "theoretical_loss": 3.5037892959480494, + "tokens_seen": 1715011584 + }, + { + "epoch": 0.52, + "learning_rate": 0.0003882046138415246, + "loss": 1.3899, + "theoretical_loss": 3.5037778858281077, + "tokens_seen": 1715077120 + }, + { + "epoch": 0.52, + "learning_rate": 0.0003881885656970913, + "loss": 1.3926, + "theoretical_loss": 3.503766476266232, + "tokens_seen": 1715142656 + }, + { + "epoch": 0.52, + "learning_rate": 0.000388172517552658, + "loss": 1.4764, + "theoretical_loss": 3.503755067262374, + "tokens_seen": 1715208192 + }, + { + "epoch": 0.52, + "learning_rate": 0.0003881564694082247, + "loss": 1.3765, + "theoretical_loss": 3.503743658816485, + "tokens_seen": 1715273728 + }, + { + "epoch": 0.52, + "learning_rate": 0.0003881404212637914, + "loss": 1.5192, + "theoretical_loss": 3.503732250928517, + "tokens_seen": 1715339264 + }, + { + "epoch": 0.52, + "learning_rate": 0.0003881243731193581, + "loss": 1.363, + "theoretical_loss": 3.5037208435984204, + "tokens_seen": 1715404800 + }, + { + "epoch": 0.52, + "learning_rate": 0.0003881083249749248, + "loss": 1.4003, + "theoretical_loss": 3.5037094368261474, + "tokens_seen": 1715470336 + }, + { + "epoch": 0.52, + "learning_rate": 0.0003880922768304915, + "loss": 1.3742, + "theoretical_loss": 3.503698030611649, + "tokens_seen": 1715535872 + }, + { + "epoch": 0.52, + "learning_rate": 0.0003880762286860582, + "loss": 1.3135, + "theoretical_loss": 3.503686624954877, + "tokens_seen": 1715601408 + }, + { + "epoch": 0.52, + "learning_rate": 0.0003880601805416249, + "loss": 1.3514, + "theoretical_loss": 3.503675219855782, + "tokens_seen": 1715666944 + }, + { + "epoch": 0.52, + "learning_rate": 0.0003880441323971916, + "loss": 1.3697, + "theoretical_loss": 3.503663815314317, + "tokens_seen": 1715732480 + }, + { + "epoch": 0.52, + "learning_rate": 0.0003880280842527583, + "loss": 1.3577, + "theoretical_loss": 3.5036524113304317, + "tokens_seen": 1715798016 + }, + { + "epoch": 0.52, + "learning_rate": 0.000388012036108325, + "loss": 1.3521, + "theoretical_loss": 3.5036410079040783, + "tokens_seen": 1715863552 + }, + { + "epoch": 0.52, + "learning_rate": 0.0003879959879638917, + "loss": 1.3397, + "theoretical_loss": 3.5036296050352087, + "tokens_seen": 1715929088 + }, + { + "epoch": 0.52, + "learning_rate": 0.0003879799398194584, + "loss": 1.362, + "theoretical_loss": 3.5036182027237732, + "tokens_seen": 1715994624 + }, + { + "epoch": 0.52, + "learning_rate": 0.0003879638916750251, + "loss": 1.378, + "theoretical_loss": 3.5036068009697248, + "tokens_seen": 1716060160 + }, + { + "epoch": 0.52, + "learning_rate": 0.0003879478435305918, + "loss": 1.2866, + "theoretical_loss": 3.5035953997730136, + "tokens_seen": 1716125696 + }, + { + "epoch": 0.52, + "learning_rate": 0.0003879317953861585, + "loss": 1.3291, + "theoretical_loss": 3.5035839991335918, + "tokens_seen": 1716191232 + }, + { + "epoch": 0.52, + "learning_rate": 0.0003879157472417252, + "loss": 1.2956, + "theoretical_loss": 3.5035725990514104, + "tokens_seen": 1716256768 + }, + { + "epoch": 0.52, + "learning_rate": 0.00038789969909729186, + "loss": 1.3369, + "theoretical_loss": 3.5035611995264215, + "tokens_seen": 1716322304 + }, + { + "epoch": 0.52, + "learning_rate": 0.00038788365095285856, + "loss": 1.3643, + "theoretical_loss": 3.503549800558576, + "tokens_seen": 1716387840 + }, + { + "epoch": 0.52, + "learning_rate": 0.00038786760280842527, + "loss": 1.3223, + "theoretical_loss": 3.503538402147826, + "tokens_seen": 1716453376 + }, + { + "epoch": 0.52, + "learning_rate": 0.00038785155466399197, + "loss": 1.3871, + "theoretical_loss": 3.5035270042941224, + "tokens_seen": 1716518912 + }, + { + "epoch": 0.52, + "learning_rate": 0.00038783550651955867, + "loss": 1.3672, + "theoretical_loss": 3.503515606997417, + "tokens_seen": 1716584448 + }, + { + "epoch": 0.52, + "learning_rate": 0.00038781945837512537, + "loss": 1.3777, + "theoretical_loss": 3.5035042102576615, + "tokens_seen": 1716649984 + }, + { + "epoch": 0.52, + "learning_rate": 0.00038780341023069207, + "loss": 1.2758, + "theoretical_loss": 3.5034928140748067, + "tokens_seen": 1716715520 + }, + { + "epoch": 0.52, + "learning_rate": 0.00038778736208625877, + "loss": 1.3635, + "theoretical_loss": 3.5034814184488052, + "tokens_seen": 1716781056 + }, + { + "epoch": 0.52, + "learning_rate": 0.00038777131394182547, + "loss": 1.3746, + "theoretical_loss": 3.5034700233796077, + "tokens_seen": 1716846592 + }, + { + "epoch": 0.52, + "learning_rate": 0.00038775526579739217, + "loss": 1.2492, + "theoretical_loss": 3.5034586288671656, + "tokens_seen": 1716912128 + }, + { + "epoch": 0.52, + "learning_rate": 0.00038773921765295887, + "loss": 1.2921, + "theoretical_loss": 3.5034472349114316, + "tokens_seen": 1716977664 + }, + { + "epoch": 0.52, + "objective/train/docs_used": 627098, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.3281793594360352, + "objective/train/theoretical_loss": 3.5034358415123563, + "objective/train/tokens_used": 1737503200, + "theoretical_loss": 3.5034358415123563, + "tokens_seen": 1717043200 + }, + { + "epoch": 0.52, + "learning_rate": 0.00038772316950852557, + "loss": 1.2713, + "theoretical_loss": 3.5034358415123563, + "tokens_seen": 1717043200 + }, + { + "epoch": 0.52, + "learning_rate": 0.00038770712136409227, + "loss": 1.3871, + "theoretical_loss": 3.5034244486698913, + "tokens_seen": 1717108736 + }, + { + "epoch": 0.52, + "learning_rate": 0.00038769107321965897, + "loss": 1.3581, + "theoretical_loss": 3.5034130563839887, + "tokens_seen": 1717174272 + }, + { + "epoch": 0.52, + "learning_rate": 0.00038767502507522567, + "loss": 1.2693, + "theoretical_loss": 3.5034016646545996, + "tokens_seen": 1717239808 + }, + { + "epoch": 0.52, + "learning_rate": 0.0003876589769307924, + "loss": 1.3681, + "theoretical_loss": 3.503390273481676, + "tokens_seen": 1717305344 + }, + { + "epoch": 0.52, + "learning_rate": 0.0003876429287863591, + "loss": 1.3671, + "theoretical_loss": 3.503378882865168, + "tokens_seen": 1717370880 + }, + { + "epoch": 0.52, + "learning_rate": 0.0003876268806419258, + "loss": 1.3337, + "theoretical_loss": 3.5033674928050296, + "tokens_seen": 1717436416 + }, + { + "epoch": 0.52, + "learning_rate": 0.0003876108324974925, + "loss": 1.3146, + "theoretical_loss": 3.5033561033012104, + "tokens_seen": 1717501952 + }, + { + "epoch": 0.52, + "learning_rate": 0.0003875947843530592, + "loss": 1.3429, + "theoretical_loss": 3.5033447143536636, + "tokens_seen": 1717567488 + }, + { + "epoch": 0.52, + "learning_rate": 0.0003875787362086259, + "loss": 1.3888, + "theoretical_loss": 3.5033333259623394, + "tokens_seen": 1717633024 + }, + { + "epoch": 0.52, + "learning_rate": 0.0003875626880641926, + "loss": 1.3355, + "theoretical_loss": 3.50332193812719, + "tokens_seen": 1717698560 + }, + { + "epoch": 0.52, + "learning_rate": 0.0003875466399197593, + "loss": 1.4156, + "theoretical_loss": 3.503310550848167, + "tokens_seen": 1717764096 + }, + { + "epoch": 0.52, + "learning_rate": 0.000387530591775326, + "loss": 1.4277, + "theoretical_loss": 3.503299164125222, + "tokens_seen": 1717829632 + }, + { + "epoch": 0.52, + "learning_rate": 0.0003875145436308927, + "loss": 1.3927, + "theoretical_loss": 3.503287777958307, + "tokens_seen": 1717895168 + }, + { + "epoch": 0.52, + "learning_rate": 0.0003874984954864594, + "loss": 1.365, + "theoretical_loss": 3.503276392347373, + "tokens_seen": 1717960704 + }, + { + "epoch": 0.52, + "learning_rate": 0.0003874824473420261, + "loss": 1.2693, + "theoretical_loss": 3.503265007292372, + "tokens_seen": 1718026240 + }, + { + "epoch": 0.52, + "learning_rate": 0.0003874663991975928, + "loss": 1.4043, + "theoretical_loss": 3.5032536227932556, + "tokens_seen": 1718091776 + }, + { + "epoch": 0.52, + "learning_rate": 0.0003874503510531595, + "loss": 1.3242, + "theoretical_loss": 3.5032422388499755, + "tokens_seen": 1718157312 + }, + { + "epoch": 0.52, + "learning_rate": 0.0003874343029087262, + "loss": 1.302, + "theoretical_loss": 3.503230855462483, + "tokens_seen": 1718222848 + }, + { + "epoch": 0.52, + "learning_rate": 0.0003874182547642929, + "loss": 1.2713, + "theoretical_loss": 3.5032194726307306, + "tokens_seen": 1718288384 + }, + { + "epoch": 0.52, + "learning_rate": 0.0003874022066198596, + "loss": 1.3416, + "theoretical_loss": 3.503208090354669, + "tokens_seen": 1718353920 + }, + { + "epoch": 0.52, + "learning_rate": 0.00038738615847542634, + "loss": 1.3174, + "theoretical_loss": 3.5031967086342504, + "tokens_seen": 1718419456 + }, + { + "epoch": 0.52, + "learning_rate": 0.00038737011033099304, + "loss": 1.2718, + "theoretical_loss": 3.5031853274694265, + "tokens_seen": 1718484992 + }, + { + "epoch": 0.52, + "learning_rate": 0.00038735406218655974, + "loss": 1.3731, + "theoretical_loss": 3.503173946860149, + "tokens_seen": 1718550528 + }, + { + "epoch": 0.52, + "learning_rate": 0.00038733801404212644, + "loss": 1.3919, + "theoretical_loss": 3.5031625668063695, + "tokens_seen": 1718616064 + }, + { + "epoch": 0.52, + "learning_rate": 0.00038732196589769314, + "loss": 1.3566, + "theoretical_loss": 3.5031511873080396, + "tokens_seen": 1718681600 + }, + { + "epoch": 0.52, + "learning_rate": 0.00038730591775325984, + "loss": 1.3384, + "theoretical_loss": 3.5031398083651113, + "tokens_seen": 1718747136 + }, + { + "epoch": 0.52, + "learning_rate": 0.00038728986960882654, + "loss": 1.2565, + "theoretical_loss": 3.503128429977536, + "tokens_seen": 1718812672 + }, + { + "epoch": 0.52, + "learning_rate": 0.00038727382146439324, + "loss": 1.3439, + "theoretical_loss": 3.5031170521452655, + "tokens_seen": 1718878208 + }, + { + "epoch": 0.52, + "learning_rate": 0.0003872577733199599, + "loss": 1.3157, + "theoretical_loss": 3.5031056748682516, + "tokens_seen": 1718943744 + }, + { + "epoch": 0.52, + "learning_rate": 0.0003872417251755266, + "loss": 1.2796, + "theoretical_loss": 3.503094298146446, + "tokens_seen": 1719009280 + }, + { + "epoch": 0.52, + "learning_rate": 0.0003872256770310933, + "loss": 1.2782, + "theoretical_loss": 3.503082921979801, + "tokens_seen": 1719074816 + }, + { + "epoch": 0.52, + "learning_rate": 0.00038720962888666, + "loss": 1.4287, + "theoretical_loss": 3.5030715463682673, + "tokens_seen": 1719140352 + }, + { + "epoch": 0.52, + "learning_rate": 0.0003871935807422267, + "loss": 1.4297, + "theoretical_loss": 3.5030601713117973, + "tokens_seen": 1719205888 + }, + { + "epoch": 0.52, + "learning_rate": 0.0003871775325977934, + "loss": 1.3766, + "theoretical_loss": 3.5030487968103428, + "tokens_seen": 1719271424 + }, + { + "epoch": 0.52, + "learning_rate": 0.0003871614844533601, + "loss": 1.3085, + "theoretical_loss": 3.503037422863855, + "tokens_seen": 1719336960 + }, + { + "epoch": 0.52, + "learning_rate": 0.0003871454363089268, + "loss": 1.3599, + "theoretical_loss": 3.5030260494722865, + "tokens_seen": 1719402496 + }, + { + "epoch": 0.52, + "learning_rate": 0.0003871293881644935, + "loss": 1.4769, + "theoretical_loss": 3.5030146766355887, + "tokens_seen": 1719468032 + }, + { + "epoch": 0.52, + "learning_rate": 0.0003871133400200602, + "loss": 1.4131, + "theoretical_loss": 3.5030033043537125, + "tokens_seen": 1719533568 + }, + { + "epoch": 0.52, + "learning_rate": 0.0003870972918756269, + "loss": 1.279, + "theoretical_loss": 3.5029919326266112, + "tokens_seen": 1719599104 + }, + { + "epoch": 0.52, + "learning_rate": 0.0003870812437311936, + "loss": 1.4034, + "theoretical_loss": 3.502980561454236, + "tokens_seen": 1719664640 + }, + { + "epoch": 0.52, + "learning_rate": 0.0003870651955867603, + "loss": 1.4519, + "theoretical_loss": 3.502969190836538, + "tokens_seen": 1719730176 + }, + { + "epoch": 0.52, + "learning_rate": 0.000387049147442327, + "loss": 1.4022, + "theoretical_loss": 3.50295782077347, + "tokens_seen": 1719795712 + }, + { + "epoch": 0.52, + "learning_rate": 0.0003870330992978937, + "loss": 1.3662, + "theoretical_loss": 3.5029464512649833, + "tokens_seen": 1719861248 + }, + { + "epoch": 0.52, + "learning_rate": 0.0003870170511534604, + "loss": 1.3126, + "theoretical_loss": 3.50293508231103, + "tokens_seen": 1719926784 + }, + { + "epoch": 0.52, + "learning_rate": 0.0003870010030090271, + "loss": 1.4656, + "theoretical_loss": 3.502923713911562, + "tokens_seen": 1719992320 + }, + { + "epoch": 0.52, + "learning_rate": 0.0003869849548645938, + "loss": 1.3409, + "theoretical_loss": 3.5029123460665303, + "tokens_seen": 1720057856 + }, + { + "epoch": 0.52, + "learning_rate": 0.0003869689067201605, + "loss": 1.3819, + "theoretical_loss": 3.502900978775888, + "tokens_seen": 1720123392 + }, + { + "epoch": 0.52, + "learning_rate": 0.0003869528585757272, + "loss": 1.3792, + "theoretical_loss": 3.502889612039586, + "tokens_seen": 1720188928 + }, + { + "epoch": 0.52, + "learning_rate": 0.0003869368104312939, + "loss": 1.3358, + "theoretical_loss": 3.502878245857576, + "tokens_seen": 1720254464 + }, + { + "epoch": 0.52, + "objective/train/docs_used": 628251, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.5476046800613403, + "objective/train/theoretical_loss": 3.502866880229811, + "objective/train/tokens_used": 1740780000, + "theoretical_loss": 3.502866880229811, + "tokens_seen": 1720320000 + }, + { + "epoch": 0.52, + "learning_rate": 0.0003869207622868606, + "loss": 1.4272, + "theoretical_loss": 3.502866880229811, + "tokens_seen": 1720320000 + }, + { + "epoch": 0.52, + "learning_rate": 0.0003869047141424273, + "loss": 1.3228, + "theoretical_loss": 3.5028555151562415, + "tokens_seen": 1720385536 + }, + { + "epoch": 0.52, + "learning_rate": 0.000386888665997994, + "loss": 1.3382, + "theoretical_loss": 3.5028441506368204, + "tokens_seen": 1720451072 + }, + { + "epoch": 0.52, + "learning_rate": 0.0003868726178535607, + "loss": 1.3845, + "theoretical_loss": 3.502832786671499, + "tokens_seen": 1720516608 + }, + { + "epoch": 0.52, + "learning_rate": 0.0003868565697091274, + "loss": 1.3758, + "theoretical_loss": 3.5028214232602295, + "tokens_seen": 1720582144 + }, + { + "epoch": 0.52, + "learning_rate": 0.0003868405215646941, + "loss": 1.3745, + "theoretical_loss": 3.5028100604029637, + "tokens_seen": 1720647680 + }, + { + "epoch": 0.52, + "learning_rate": 0.0003868244734202608, + "loss": 1.2741, + "theoretical_loss": 3.5027986980996535, + "tokens_seen": 1720713216 + }, + { + "epoch": 0.52, + "learning_rate": 0.0003868084252758275, + "loss": 1.3772, + "theoretical_loss": 3.5027873363502504, + "tokens_seen": 1720778752 + }, + { + "epoch": 0.52, + "learning_rate": 0.0003867923771313942, + "loss": 1.3392, + "theoretical_loss": 3.502775975154707, + "tokens_seen": 1720844288 + }, + { + "epoch": 0.52, + "learning_rate": 0.0003867763289869609, + "loss": 1.3766, + "theoretical_loss": 3.5027646145129747, + "tokens_seen": 1720909824 + }, + { + "epoch": 0.52, + "learning_rate": 0.0003867602808425276, + "loss": 1.3168, + "theoretical_loss": 3.5027532544250057, + "tokens_seen": 1720975360 + }, + { + "epoch": 0.52, + "learning_rate": 0.0003867442326980943, + "loss": 1.422, + "theoretical_loss": 3.502741894890751, + "tokens_seen": 1721040896 + }, + { + "epoch": 0.52, + "learning_rate": 0.000386728184553661, + "loss": 1.4068, + "theoretical_loss": 3.5027305359101644, + "tokens_seen": 1721106432 + }, + { + "epoch": 0.52, + "learning_rate": 0.0003867121364092277, + "loss": 1.3658, + "theoretical_loss": 3.502719177483196, + "tokens_seen": 1721171968 + }, + { + "epoch": 0.52, + "learning_rate": 0.0003866960882647944, + "loss": 1.309, + "theoretical_loss": 3.502707819609799, + "tokens_seen": 1721237504 + }, + { + "epoch": 0.52, + "learning_rate": 0.0003866800401203611, + "loss": 1.3225, + "theoretical_loss": 3.502696462289925, + "tokens_seen": 1721303040 + }, + { + "epoch": 0.52, + "learning_rate": 0.0003866639919759278, + "loss": 1.3198, + "theoretical_loss": 3.502685105523525, + "tokens_seen": 1721368576 + }, + { + "epoch": 0.52, + "learning_rate": 0.0003866479438314945, + "loss": 1.3823, + "theoretical_loss": 3.5026737493105524, + "tokens_seen": 1721434112 + }, + { + "epoch": 0.52, + "learning_rate": 0.0003866318956870612, + "loss": 1.3654, + "theoretical_loss": 3.502662393650958, + "tokens_seen": 1721499648 + }, + { + "epoch": 0.52, + "learning_rate": 0.0003866158475426279, + "loss": 1.3975, + "theoretical_loss": 3.5026510385446947, + "tokens_seen": 1721565184 + }, + { + "epoch": 0.52, + "learning_rate": 0.0003865997993981946, + "loss": 1.3436, + "theoretical_loss": 3.502639683991714, + "tokens_seen": 1721630720 + }, + { + "epoch": 0.52, + "learning_rate": 0.0003865837512537613, + "loss": 1.421, + "theoretical_loss": 3.5026283299919676, + "tokens_seen": 1721696256 + }, + { + "epoch": 0.52, + "learning_rate": 0.000386567703109328, + "loss": 1.3322, + "theoretical_loss": 3.5026169765454083, + "tokens_seen": 1721761792 + }, + { + "epoch": 0.52, + "learning_rate": 0.0003865516549648947, + "loss": 1.3283, + "theoretical_loss": 3.5026056236519874, + "tokens_seen": 1721827328 + }, + { + "epoch": 0.52, + "learning_rate": 0.0003865356068204614, + "loss": 1.3034, + "theoretical_loss": 3.502594271311657, + "tokens_seen": 1721892864 + }, + { + "epoch": 0.52, + "learning_rate": 0.0003865195586760281, + "loss": 1.4254, + "theoretical_loss": 3.502582919524369, + "tokens_seen": 1721958400 + }, + { + "epoch": 0.52, + "learning_rate": 0.0003865035105315948, + "loss": 1.3365, + "theoretical_loss": 3.502571568290076, + "tokens_seen": 1722023936 + }, + { + "epoch": 0.52, + "learning_rate": 0.00038648746238716147, + "loss": 1.3963, + "theoretical_loss": 3.50256021760873, + "tokens_seen": 1722089472 + }, + { + "epoch": 0.52, + "learning_rate": 0.00038647141424272817, + "loss": 1.3111, + "theoretical_loss": 3.502548867480282, + "tokens_seen": 1722155008 + }, + { + "epoch": 0.52, + "learning_rate": 0.00038645536609829487, + "loss": 1.3787, + "theoretical_loss": 3.5025375179046847, + "tokens_seen": 1722220544 + }, + { + "epoch": 0.52, + "learning_rate": 0.00038643931795386157, + "loss": 1.2803, + "theoretical_loss": 3.5025261688818903, + "tokens_seen": 1722286080 + }, + { + "epoch": 0.52, + "learning_rate": 0.00038642326980942827, + "loss": 1.2837, + "theoretical_loss": 3.502514820411851, + "tokens_seen": 1722351616 + }, + { + "epoch": 0.52, + "learning_rate": 0.00038640722166499497, + "loss": 1.3289, + "theoretical_loss": 3.502503472494518, + "tokens_seen": 1722417152 + }, + { + "epoch": 0.52, + "learning_rate": 0.00038639117352056167, + "loss": 1.4198, + "theoretical_loss": 3.5024921251298444, + "tokens_seen": 1722482688 + }, + { + "epoch": 0.52, + "learning_rate": 0.00038637512537612837, + "loss": 1.3144, + "theoretical_loss": 3.502480778317781, + "tokens_seen": 1722548224 + }, + { + "epoch": 0.52, + "learning_rate": 0.00038635907723169507, + "loss": 1.4154, + "theoretical_loss": 3.502469432058281, + "tokens_seen": 1722613760 + }, + { + "epoch": 0.52, + "learning_rate": 0.00038634302908726177, + "loss": 1.3834, + "theoretical_loss": 3.5024580863512966, + "tokens_seen": 1722679296 + }, + { + "epoch": 0.52, + "learning_rate": 0.00038632698094282847, + "loss": 1.3622, + "theoretical_loss": 3.5024467411967786, + "tokens_seen": 1722744832 + }, + { + "epoch": 0.52, + "learning_rate": 0.00038631093279839517, + "loss": 1.3768, + "theoretical_loss": 3.50243539659468, + "tokens_seen": 1722810368 + }, + { + "epoch": 0.52, + "learning_rate": 0.0003862948846539619, + "loss": 1.3138, + "theoretical_loss": 3.502424052544953, + "tokens_seen": 1722875904 + }, + { + "epoch": 0.52, + "learning_rate": 0.0003862788365095286, + "loss": 1.3798, + "theoretical_loss": 3.502412709047549, + "tokens_seen": 1722941440 + }, + { + "epoch": 0.52, + "learning_rate": 0.0003862627883650953, + "loss": 1.4073, + "theoretical_loss": 3.502401366102421, + "tokens_seen": 1723006976 + }, + { + "epoch": 0.52, + "learning_rate": 0.000386246740220662, + "loss": 1.3465, + "theoretical_loss": 3.5023900237095202, + "tokens_seen": 1723072512 + }, + { + "epoch": 0.52, + "learning_rate": 0.0003862306920762287, + "loss": 1.3293, + "theoretical_loss": 3.502378681868799, + "tokens_seen": 1723138048 + }, + { + "epoch": 0.52, + "learning_rate": 0.0003862146439317954, + "loss": 1.3512, + "theoretical_loss": 3.50236734058021, + "tokens_seen": 1723203584 + }, + { + "epoch": 0.52, + "learning_rate": 0.0003861985957873621, + "loss": 1.3646, + "theoretical_loss": 3.5023559998437053, + "tokens_seen": 1723269120 + }, + { + "epoch": 0.52, + "learning_rate": 0.00038618254764292883, + "loss": 1.4299, + "theoretical_loss": 3.5023446596592365, + "tokens_seen": 1723334656 + }, + { + "epoch": 0.52, + "learning_rate": 0.00038616649949849553, + "loss": 1.3379, + "theoretical_loss": 3.5023333200267555, + "tokens_seen": 1723400192 + }, + { + "epoch": 0.52, + "learning_rate": 0.00038615045135406223, + "loss": 1.3744, + "theoretical_loss": 3.502321980946215, + "tokens_seen": 1723465728 + }, + { + "epoch": 0.52, + "learning_rate": 0.00038613440320962893, + "loss": 1.4293, + "theoretical_loss": 3.502310642417567, + "tokens_seen": 1723531264 + }, + { + "epoch": 0.52, + "objective/train/docs_used": 629447, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.4761786460876465, + "objective/train/theoretical_loss": 3.502299304440764, + "objective/train/tokens_used": 1744056800, + "theoretical_loss": 3.502299304440764, + "tokens_seen": 1723596800 + }, + { + "epoch": 0.52, + "learning_rate": 0.00038611835506519563, + "loss": 1.3923, + "theoretical_loss": 3.502299304440764, + "tokens_seen": 1723596800 + }, + { + "epoch": 0.52, + "learning_rate": 0.00038610230692076234, + "loss": 1.4182, + "theoretical_loss": 3.5022879670157576, + "tokens_seen": 1723662336 + }, + { + "epoch": 0.52, + "learning_rate": 0.00038608625877632904, + "loss": 1.379, + "theoretical_loss": 3.5022766301425, + "tokens_seen": 1723727872 + }, + { + "epoch": 0.52, + "learning_rate": 0.00038607021063189574, + "loss": 1.247, + "theoretical_loss": 3.502265293820944, + "tokens_seen": 1723793408 + }, + { + "epoch": 0.52, + "learning_rate": 0.00038605416248746244, + "loss": 1.3246, + "theoretical_loss": 3.5022539580510417, + "tokens_seen": 1723858944 + }, + { + "epoch": 0.52, + "learning_rate": 0.00038603811434302914, + "loss": 1.3712, + "theoretical_loss": 3.5022426228327443, + "tokens_seen": 1723924480 + }, + { + "epoch": 0.52, + "learning_rate": 0.00038602206619859584, + "loss": 1.3493, + "theoretical_loss": 3.502231288166005, + "tokens_seen": 1723990016 + }, + { + "epoch": 0.52, + "learning_rate": 0.00038600601805416254, + "loss": 1.4056, + "theoretical_loss": 3.5022199540507755, + "tokens_seen": 1724055552 + }, + { + "epoch": 0.52, + "learning_rate": 0.00038598996990972924, + "loss": 1.3914, + "theoretical_loss": 3.502208620487008, + "tokens_seen": 1724121088 + }, + { + "epoch": 0.52, + "learning_rate": 0.00038597392176529594, + "loss": 1.2937, + "theoretical_loss": 3.5021972874746554, + "tokens_seen": 1724186624 + }, + { + "epoch": 0.52, + "learning_rate": 0.00038595787362086264, + "loss": 1.3708, + "theoretical_loss": 3.502185955013669, + "tokens_seen": 1724252160 + }, + { + "epoch": 0.52, + "learning_rate": 0.00038594182547642934, + "loss": 1.3013, + "theoretical_loss": 3.5021746231040014, + "tokens_seen": 1724317696 + }, + { + "epoch": 0.52, + "learning_rate": 0.00038592577733199604, + "loss": 1.4092, + "theoretical_loss": 3.5021632917456045, + "tokens_seen": 1724383232 + }, + { + "epoch": 0.52, + "learning_rate": 0.00038590972918756274, + "loss": 1.3071, + "theoretical_loss": 3.502151960938431, + "tokens_seen": 1724448768 + }, + { + "epoch": 0.52, + "learning_rate": 0.00038589368104312944, + "loss": 1.3938, + "theoretical_loss": 3.502140630682433, + "tokens_seen": 1724514304 + }, + { + "epoch": 0.52, + "learning_rate": 0.00038587763289869614, + "loss": 1.3068, + "theoretical_loss": 3.502129300977563, + "tokens_seen": 1724579840 + }, + { + "epoch": 0.52, + "learning_rate": 0.0003858615847542628, + "loss": 1.3847, + "theoretical_loss": 3.5021179718237727, + "tokens_seen": 1724645376 + }, + { + "epoch": 0.52, + "learning_rate": 0.0003858455366098295, + "loss": 1.338, + "theoretical_loss": 3.5021066432210146, + "tokens_seen": 1724710912 + }, + { + "epoch": 0.52, + "learning_rate": 0.0003858294884653962, + "loss": 1.5247, + "theoretical_loss": 3.5020953151692416, + "tokens_seen": 1724776448 + }, + { + "epoch": 0.52, + "learning_rate": 0.0003858134403209629, + "loss": 1.2457, + "theoretical_loss": 3.5020839876684047, + "tokens_seen": 1724841984 + }, + { + "epoch": 0.52, + "learning_rate": 0.0003857973921765296, + "loss": 1.336, + "theoretical_loss": 3.502072660718457, + "tokens_seen": 1724907520 + }, + { + "epoch": 0.52, + "learning_rate": 0.0003857813440320963, + "loss": 1.4822, + "theoretical_loss": 3.5020613343193507, + "tokens_seen": 1724973056 + }, + { + "epoch": 0.52, + "learning_rate": 0.000385765295887663, + "loss": 1.3657, + "theoretical_loss": 3.5020500084710378, + "tokens_seen": 1725038592 + }, + { + "epoch": 0.52, + "learning_rate": 0.0003857492477432297, + "loss": 1.3754, + "theoretical_loss": 3.502038683173471, + "tokens_seen": 1725104128 + }, + { + "epoch": 0.52, + "learning_rate": 0.0003857331995987964, + "loss": 1.3914, + "theoretical_loss": 3.502027358426602, + "tokens_seen": 1725169664 + }, + { + "epoch": 0.52, + "learning_rate": 0.0003857171514543631, + "loss": 1.304, + "theoretical_loss": 3.502016034230384, + "tokens_seen": 1725235200 + }, + { + "epoch": 0.52, + "learning_rate": 0.0003857011033099298, + "loss": 1.3761, + "theoretical_loss": 3.5020047105847687, + "tokens_seen": 1725300736 + }, + { + "epoch": 0.52, + "learning_rate": 0.0003856850551654965, + "loss": 1.4726, + "theoretical_loss": 3.5019933874897085, + "tokens_seen": 1725366272 + }, + { + "epoch": 0.52, + "learning_rate": 0.0003856690070210632, + "loss": 1.4771, + "theoretical_loss": 3.5019820649451554, + "tokens_seen": 1725431808 + }, + { + "epoch": 0.52, + "learning_rate": 0.0003856529588766299, + "loss": 1.3115, + "theoretical_loss": 3.5019707429510625, + "tokens_seen": 1725497344 + }, + { + "epoch": 0.52, + "learning_rate": 0.0003856369107321966, + "loss": 1.3847, + "theoretical_loss": 3.5019594215073813, + "tokens_seen": 1725562880 + }, + { + "epoch": 0.52, + "learning_rate": 0.0003856208625877633, + "loss": 1.3556, + "theoretical_loss": 3.501948100614065, + "tokens_seen": 1725628416 + }, + { + "epoch": 0.52, + "learning_rate": 0.00038560481444333, + "loss": 1.3935, + "theoretical_loss": 3.5019367802710653, + "tokens_seen": 1725693952 + }, + { + "epoch": 0.52, + "learning_rate": 0.0003855887662988967, + "loss": 1.2759, + "theoretical_loss": 3.5019254604783345, + "tokens_seen": 1725759488 + }, + { + "epoch": 0.52, + "learning_rate": 0.0003855727181544634, + "loss": 1.3347, + "theoretical_loss": 3.5019141412358255, + "tokens_seen": 1725825024 + }, + { + "epoch": 0.52, + "learning_rate": 0.0003855566700100301, + "loss": 1.5678, + "theoretical_loss": 3.5019028225434905, + "tokens_seen": 1725890560 + }, + { + "epoch": 0.52, + "learning_rate": 0.0003855406218655968, + "loss": 1.3451, + "theoretical_loss": 3.5018915044012813, + "tokens_seen": 1725956096 + }, + { + "epoch": 0.52, + "learning_rate": 0.0003855245737211635, + "loss": 1.3486, + "theoretical_loss": 3.5018801868091507, + "tokens_seen": 1726021632 + }, + { + "epoch": 0.52, + "learning_rate": 0.0003855085255767302, + "loss": 1.3674, + "theoretical_loss": 3.5018688697670513, + "tokens_seen": 1726087168 + }, + { + "epoch": 0.52, + "learning_rate": 0.0003854924774322969, + "loss": 1.479, + "theoretical_loss": 3.501857553274935, + "tokens_seen": 1726152704 + }, + { + "epoch": 0.52, + "learning_rate": 0.0003854764292878636, + "loss": 1.4288, + "theoretical_loss": 3.501846237332755, + "tokens_seen": 1726218240 + }, + { + "epoch": 0.52, + "learning_rate": 0.0003854603811434303, + "loss": 1.3799, + "theoretical_loss": 3.5018349219404628, + "tokens_seen": 1726283776 + }, + { + "epoch": 0.52, + "learning_rate": 0.000385444332998997, + "loss": 1.3693, + "theoretical_loss": 3.501823607098011, + "tokens_seen": 1726349312 + }, + { + "epoch": 0.52, + "learning_rate": 0.0003854282848545637, + "loss": 1.3732, + "theoretical_loss": 3.5018122928053526, + "tokens_seen": 1726414848 + }, + { + "epoch": 0.52, + "learning_rate": 0.0003854122367101304, + "loss": 1.3055, + "theoretical_loss": 3.5018009790624394, + "tokens_seen": 1726480384 + }, + { + "epoch": 0.52, + "learning_rate": 0.0003853961885656971, + "loss": 1.4539, + "theoretical_loss": 3.5017896658692234, + "tokens_seen": 1726545920 + }, + { + "epoch": 0.52, + "learning_rate": 0.0003853801404212638, + "loss": 1.3657, + "theoretical_loss": 3.5017783532256583, + "tokens_seen": 1726611456 + }, + { + "epoch": 0.52, + "learning_rate": 0.0003853640922768305, + "loss": 1.405, + "theoretical_loss": 3.501767041131696, + "tokens_seen": 1726676992 + }, + { + "epoch": 0.52, + "learning_rate": 0.0003853480441323972, + "loss": 1.4024, + "theoretical_loss": 3.5017557295872885, + "tokens_seen": 1726742528 + }, + { + "epoch": 0.52, + "learning_rate": 0.0003853319959879639, + "loss": 1.3415, + "theoretical_loss": 3.5017444185923887, + "tokens_seen": 1726808064 + }, + { + "epoch": 0.52, + "objective/train/docs_used": 630619, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.4286601543426514, + "objective/train/theoretical_loss": 3.501733108146949, + "objective/train/tokens_used": 1747333600, + "theoretical_loss": 3.501733108146949, + "tokens_seen": 1726873600 + }, + { + "epoch": 0.52, + "learning_rate": 0.0003853159478435306, + "loss": 1.4481, + "theoretical_loss": 3.501733108146949, + "tokens_seen": 1726873600 + }, + { + "epoch": 0.52, + "learning_rate": 0.0003852998996990973, + "loss": 1.3799, + "theoretical_loss": 3.5017217982509212, + "tokens_seen": 1726939136 + }, + { + "epoch": 0.52, + "learning_rate": 0.000385283851554664, + "loss": 1.3021, + "theoretical_loss": 3.5017104889042585, + "tokens_seen": 1727004672 + }, + { + "epoch": 0.52, + "learning_rate": 0.0003852678034102307, + "loss": 1.2355, + "theoretical_loss": 3.5016991801069137, + "tokens_seen": 1727070208 + }, + { + "epoch": 0.52, + "learning_rate": 0.0003852517552657974, + "loss": 1.4704, + "theoretical_loss": 3.501687871858839, + "tokens_seen": 1727135744 + }, + { + "epoch": 0.52, + "learning_rate": 0.0003852357071213641, + "loss": 1.4379, + "theoretical_loss": 3.501676564159986, + "tokens_seen": 1727201280 + }, + { + "epoch": 0.52, + "learning_rate": 0.0003852196589769308, + "loss": 1.4146, + "theoretical_loss": 3.501665257010308, + "tokens_seen": 1727266816 + }, + { + "epoch": 0.52, + "learning_rate": 0.0003852036108324975, + "loss": 1.3157, + "theoretical_loss": 3.5016539504097572, + "tokens_seen": 1727332352 + }, + { + "epoch": 0.52, + "learning_rate": 0.0003851875626880642, + "loss": 1.2711, + "theoretical_loss": 3.501642644358286, + "tokens_seen": 1727397888 + }, + { + "epoch": 0.52, + "learning_rate": 0.0003851715145436309, + "loss": 1.393, + "theoretical_loss": 3.501631338855848, + "tokens_seen": 1727463424 + }, + { + "epoch": 0.52, + "learning_rate": 0.0003851554663991976, + "loss": 1.3619, + "theoretical_loss": 3.5016200339023946, + "tokens_seen": 1727528960 + }, + { + "epoch": 0.52, + "learning_rate": 0.0003851394182547643, + "loss": 1.4234, + "theoretical_loss": 3.501608729497878, + "tokens_seen": 1727594496 + }, + { + "epoch": 0.52, + "learning_rate": 0.000385123370110331, + "loss": 1.3271, + "theoretical_loss": 3.5015974256422515, + "tokens_seen": 1727660032 + }, + { + "epoch": 0.52, + "learning_rate": 0.0003851073219658977, + "loss": 1.3189, + "theoretical_loss": 3.501586122335468, + "tokens_seen": 1727725568 + }, + { + "epoch": 0.52, + "learning_rate": 0.00038509127382146437, + "loss": 1.3326, + "theoretical_loss": 3.501574819577479, + "tokens_seen": 1727791104 + }, + { + "epoch": 0.52, + "learning_rate": 0.00038507522567703107, + "loss": 1.3309, + "theoretical_loss": 3.5015635173682376, + "tokens_seen": 1727856640 + }, + { + "epoch": 0.52, + "learning_rate": 0.00038505917753259777, + "loss": 1.3468, + "theoretical_loss": 3.5015522157076964, + "tokens_seen": 1727922176 + }, + { + "epoch": 0.52, + "learning_rate": 0.00038504312938816447, + "loss": 1.3023, + "theoretical_loss": 3.5015409145958074, + "tokens_seen": 1727987712 + }, + { + "epoch": 0.52, + "learning_rate": 0.00038502708124373117, + "loss": 1.3515, + "theoretical_loss": 3.5015296140325236, + "tokens_seen": 1728053248 + }, + { + "epoch": 0.52, + "learning_rate": 0.00038501103309929787, + "loss": 1.4757, + "theoretical_loss": 3.5015183140177975, + "tokens_seen": 1728118784 + }, + { + "epoch": 0.52, + "learning_rate": 0.00038499498495486457, + "loss": 1.3472, + "theoretical_loss": 3.5015070145515823, + "tokens_seen": 1728184320 + }, + { + "epoch": 0.52, + "learning_rate": 0.0003849789368104313, + "loss": 1.306, + "theoretical_loss": 3.5014957156338293, + "tokens_seen": 1728249856 + }, + { + "epoch": 0.52, + "learning_rate": 0.000384962888665998, + "loss": 1.2929, + "theoretical_loss": 3.501484417264492, + "tokens_seen": 1728315392 + }, + { + "epoch": 0.52, + "learning_rate": 0.0003849468405215647, + "loss": 1.4059, + "theoretical_loss": 3.5014731194435225, + "tokens_seen": 1728380928 + }, + { + "epoch": 0.52, + "learning_rate": 0.00038493079237713143, + "loss": 1.4389, + "theoretical_loss": 3.501461822170874, + "tokens_seen": 1728446464 + }, + { + "epoch": 0.52, + "learning_rate": 0.00038491474423269813, + "loss": 1.3653, + "theoretical_loss": 3.5014505254464985, + "tokens_seen": 1728512000 + }, + { + "epoch": 0.52, + "learning_rate": 0.00038489869608826483, + "loss": 1.277, + "theoretical_loss": 3.501439229270349, + "tokens_seen": 1728577536 + }, + { + "epoch": 0.52, + "learning_rate": 0.00038488264794383153, + "loss": 1.2207, + "theoretical_loss": 3.5014279336423777, + "tokens_seen": 1728643072 + }, + { + "epoch": 0.52, + "learning_rate": 0.00038486659979939823, + "loss": 1.3852, + "theoretical_loss": 3.501416638562538, + "tokens_seen": 1728708608 + }, + { + "epoch": 0.52, + "learning_rate": 0.00038485055165496493, + "loss": 1.4154, + "theoretical_loss": 3.501405344030781, + "tokens_seen": 1728774144 + }, + { + "epoch": 0.52, + "learning_rate": 0.00038483450351053163, + "loss": 1.2273, + "theoretical_loss": 3.501394050047061, + "tokens_seen": 1728839680 + }, + { + "epoch": 0.52, + "learning_rate": 0.00038481845536609833, + "loss": 1.3621, + "theoretical_loss": 3.5013827566113296, + "tokens_seen": 1728905216 + }, + { + "epoch": 0.52, + "learning_rate": 0.00038480240722166503, + "loss": 1.2805, + "theoretical_loss": 3.5013714637235402, + "tokens_seen": 1728970752 + }, + { + "epoch": 0.52, + "learning_rate": 0.00038478635907723173, + "loss": 1.3744, + "theoretical_loss": 3.501360171383645, + "tokens_seen": 1729036288 + }, + { + "epoch": 0.52, + "learning_rate": 0.00038477031093279843, + "loss": 1.3795, + "theoretical_loss": 3.5013488795915966, + "tokens_seen": 1729101824 + }, + { + "epoch": 0.52, + "learning_rate": 0.00038475426278836513, + "loss": 1.3647, + "theoretical_loss": 3.5013375883473477, + "tokens_seen": 1729167360 + }, + { + "epoch": 0.52, + "learning_rate": 0.00038473821464393183, + "loss": 1.2376, + "theoretical_loss": 3.501326297650851, + "tokens_seen": 1729232896 + }, + { + "epoch": 0.52, + "learning_rate": 0.00038472216649949854, + "loss": 1.3142, + "theoretical_loss": 3.5013150075020594, + "tokens_seen": 1729298432 + }, + { + "epoch": 0.52, + "learning_rate": 0.00038470611835506524, + "loss": 1.3176, + "theoretical_loss": 3.5013037179009245, + "tokens_seen": 1729363968 + }, + { + "epoch": 0.52, + "learning_rate": 0.00038469007021063194, + "loss": 1.3543, + "theoretical_loss": 3.5012924288474006, + "tokens_seen": 1729429504 + }, + { + "epoch": 0.52, + "learning_rate": 0.00038467402206619864, + "loss": 1.4095, + "theoretical_loss": 3.5012811403414394, + "tokens_seen": 1729495040 + }, + { + "epoch": 0.52, + "learning_rate": 0.00038465797392176534, + "loss": 1.3219, + "theoretical_loss": 3.501269852382994, + "tokens_seen": 1729560576 + }, + { + "epoch": 0.52, + "learning_rate": 0.00038464192577733204, + "loss": 1.3958, + "theoretical_loss": 3.501258564972017, + "tokens_seen": 1729626112 + }, + { + "epoch": 0.52, + "learning_rate": 0.00038462587763289874, + "loss": 1.3721, + "theoretical_loss": 3.501247278108461, + "tokens_seen": 1729691648 + }, + { + "epoch": 0.52, + "learning_rate": 0.00038460982948846544, + "loss": 1.3768, + "theoretical_loss": 3.501235991792279, + "tokens_seen": 1729757184 + }, + { + "epoch": 0.52, + "learning_rate": 0.00038459378134403214, + "loss": 1.3136, + "theoretical_loss": 3.5012247060234225, + "tokens_seen": 1729822720 + }, + { + "epoch": 0.52, + "learning_rate": 0.00038457773319959884, + "loss": 1.4332, + "theoretical_loss": 3.5012134208018457, + "tokens_seen": 1729888256 + }, + { + "epoch": 0.52, + "learning_rate": 0.00038456168505516554, + "loss": 1.3011, + "theoretical_loss": 3.501202136127501, + "tokens_seen": 1729953792 + }, + { + "epoch": 0.52, + "learning_rate": 0.00038454563691073224, + "loss": 1.1915, + "theoretical_loss": 3.5011908520003407, + "tokens_seen": 1730019328 + }, + { + "epoch": 0.52, + "learning_rate": 0.00038452958876629894, + "loss": 1.3433, + "theoretical_loss": 3.501179568420318, + "tokens_seen": 1730084864 + }, + { + "epoch": 0.52, + "objective/train/docs_used": 631803, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.4266982078552246, + "objective/train/theoretical_loss": 3.5011682853873856, + "objective/train/tokens_used": 1750610400, + "theoretical_loss": 3.5011682853873856, + "tokens_seen": 1730150400 + }, + { + "epoch": 0.52, + "learning_rate": 0.00038451354062186564, + "loss": 1.3831, + "theoretical_loss": 3.5011682853873856, + "tokens_seen": 1730150400 + }, + { + "epoch": 0.52, + "learning_rate": 0.00038449749247743234, + "loss": 1.4318, + "theoretical_loss": 3.501157002901495, + "tokens_seen": 1730215936 + }, + { + "epoch": 0.52, + "learning_rate": 0.00038448144433299904, + "loss": 1.431, + "theoretical_loss": 3.501145720962601, + "tokens_seen": 1730281472 + }, + { + "epoch": 0.52, + "learning_rate": 0.00038446539618856575, + "loss": 1.3888, + "theoretical_loss": 3.5011344395706554, + "tokens_seen": 1730347008 + }, + { + "epoch": 0.52, + "learning_rate": 0.0003844493480441324, + "loss": 1.3968, + "theoretical_loss": 3.501123158725611, + "tokens_seen": 1730412544 + }, + { + "epoch": 0.52, + "learning_rate": 0.0003844332998996991, + "loss": 1.3833, + "theoretical_loss": 3.5011118784274204, + "tokens_seen": 1730478080 + }, + { + "epoch": 0.52, + "learning_rate": 0.0003844172517552658, + "loss": 1.3537, + "theoretical_loss": 3.501100598676036, + "tokens_seen": 1730543616 + }, + { + "epoch": 0.52, + "learning_rate": 0.0003844012036108325, + "loss": 1.2539, + "theoretical_loss": 3.501089319471412, + "tokens_seen": 1730609152 + }, + { + "epoch": 0.52, + "learning_rate": 0.0003843851554663992, + "loss": 1.4527, + "theoretical_loss": 3.5010780408135, + "tokens_seen": 1730674688 + }, + { + "epoch": 0.52, + "learning_rate": 0.0003843691073219659, + "loss": 1.4136, + "theoretical_loss": 3.501066762702253, + "tokens_seen": 1730740224 + }, + { + "epoch": 0.52, + "learning_rate": 0.0003843530591775326, + "loss": 1.2109, + "theoretical_loss": 3.501055485137624, + "tokens_seen": 1730805760 + }, + { + "epoch": 0.52, + "learning_rate": 0.0003843370110330993, + "loss": 1.4199, + "theoretical_loss": 3.5010442081195654, + "tokens_seen": 1730871296 + }, + { + "epoch": 0.52, + "learning_rate": 0.000384320962888666, + "loss": 1.3995, + "theoretical_loss": 3.5010329316480306, + "tokens_seen": 1730936832 + }, + { + "epoch": 0.52, + "learning_rate": 0.0003843049147442327, + "loss": 1.3408, + "theoretical_loss": 3.501021655722972, + "tokens_seen": 1731002368 + }, + { + "epoch": 0.52, + "learning_rate": 0.0003842888665997994, + "loss": 1.3388, + "theoretical_loss": 3.501010380344343, + "tokens_seen": 1731067904 + }, + { + "epoch": 0.52, + "learning_rate": 0.0003842728184553661, + "loss": 1.4085, + "theoretical_loss": 3.5009991055120957, + "tokens_seen": 1731133440 + }, + { + "epoch": 0.52, + "learning_rate": 0.0003842567703109328, + "loss": 1.4987, + "theoretical_loss": 3.5009878312261833, + "tokens_seen": 1731198976 + }, + { + "epoch": 0.52, + "learning_rate": 0.0003842407221664995, + "loss": 1.3572, + "theoretical_loss": 3.500976557486559, + "tokens_seen": 1731264512 + }, + { + "epoch": 0.52, + "learning_rate": 0.0003842246740220662, + "loss": 1.3614, + "theoretical_loss": 3.5009652842931747, + "tokens_seen": 1731330048 + }, + { + "epoch": 0.52, + "learning_rate": 0.0003842086258776329, + "loss": 1.3892, + "theoretical_loss": 3.5009540116459843, + "tokens_seen": 1731395584 + }, + { + "epoch": 0.52, + "learning_rate": 0.0003841925777331996, + "loss": 1.4357, + "theoretical_loss": 3.5009427395449397, + "tokens_seen": 1731461120 + }, + { + "epoch": 0.52, + "learning_rate": 0.0003841765295887663, + "loss": 1.3902, + "theoretical_loss": 3.5009314679899948, + "tokens_seen": 1731526656 + }, + { + "epoch": 0.52, + "learning_rate": 0.000384160481444333, + "loss": 1.3483, + "theoretical_loss": 3.5009201969811015, + "tokens_seen": 1731592192 + }, + { + "epoch": 0.52, + "learning_rate": 0.0003841444332998997, + "loss": 1.2701, + "theoretical_loss": 3.500908926518213, + "tokens_seen": 1731657728 + }, + { + "epoch": 0.52, + "learning_rate": 0.0003841283851554664, + "loss": 1.3082, + "theoretical_loss": 3.500897656601283, + "tokens_seen": 1731723264 + }, + { + "epoch": 0.52, + "learning_rate": 0.0003841123370110331, + "loss": 1.3044, + "theoretical_loss": 3.5008863872302634, + "tokens_seen": 1731788800 + }, + { + "epoch": 0.52, + "learning_rate": 0.0003840962888665998, + "loss": 1.3886, + "theoretical_loss": 3.5008751184051072, + "tokens_seen": 1731854336 + }, + { + "epoch": 0.52, + "learning_rate": 0.0003840802407221665, + "loss": 1.4578, + "theoretical_loss": 3.500863850125768, + "tokens_seen": 1731919872 + }, + { + "epoch": 0.52, + "learning_rate": 0.0003840641925777332, + "loss": 1.414, + "theoretical_loss": 3.5008525823921977, + "tokens_seen": 1731985408 + }, + { + "epoch": 0.52, + "learning_rate": 0.0003840481444332999, + "loss": 1.3157, + "theoretical_loss": 3.5008413152043496, + "tokens_seen": 1732050944 + }, + { + "epoch": 0.52, + "learning_rate": 0.0003840320962888666, + "loss": 1.397, + "theoretical_loss": 3.500830048562177, + "tokens_seen": 1732116480 + }, + { + "epoch": 0.52, + "learning_rate": 0.0003840160481444333, + "loss": 1.3448, + "theoretical_loss": 3.5008187824656325, + "tokens_seen": 1732182016 + }, + { + "epoch": 0.52, + "learning_rate": 0.000384, + "loss": 1.3534, + "theoretical_loss": 3.5008075169146693, + "tokens_seen": 1732247552 + }, + { + "epoch": 0.52, + "learning_rate": 0.0003839839518555667, + "loss": 1.347, + "theoretical_loss": 3.50079625190924, + "tokens_seen": 1732313088 + }, + { + "epoch": 0.52, + "learning_rate": 0.0003839679037111334, + "loss": 1.3161, + "theoretical_loss": 3.500784987449298, + "tokens_seen": 1732378624 + }, + { + "epoch": 0.52, + "learning_rate": 0.0003839518555667001, + "loss": 1.2872, + "theoretical_loss": 3.5007737235347953, + "tokens_seen": 1732444160 + }, + { + "epoch": 0.53, + "learning_rate": 0.0003839358074222668, + "loss": 1.3337, + "theoretical_loss": 3.500762460165686, + "tokens_seen": 1732509696 + }, + { + "epoch": 0.53, + "learning_rate": 0.0003839197592778335, + "loss": 1.3148, + "theoretical_loss": 3.5007511973419225, + "tokens_seen": 1732575232 + }, + { + "epoch": 0.53, + "learning_rate": 0.0003839037111334002, + "loss": 1.3155, + "theoretical_loss": 3.5007399350634576, + "tokens_seen": 1732640768 + }, + { + "epoch": 0.53, + "learning_rate": 0.0003838876629889669, + "loss": 1.2668, + "theoretical_loss": 3.5007286733302445, + "tokens_seen": 1732706304 + }, + { + "epoch": 0.53, + "learning_rate": 0.0003838716148445336, + "loss": 1.2474, + "theoretical_loss": 3.5007174121422366, + "tokens_seen": 1732771840 + }, + { + "epoch": 0.53, + "learning_rate": 0.0003838555667001003, + "loss": 1.3328, + "theoretical_loss": 3.5007061514993865, + "tokens_seen": 1732837376 + }, + { + "epoch": 0.53, + "learning_rate": 0.000383839518555667, + "loss": 1.436, + "theoretical_loss": 3.500694891401647, + "tokens_seen": 1732902912 + }, + { + "epoch": 0.53, + "learning_rate": 0.0003838234704112337, + "loss": 1.3524, + "theoretical_loss": 3.500683631848971, + "tokens_seen": 1732968448 + }, + { + "epoch": 0.53, + "learning_rate": 0.0003838074222668004, + "loss": 1.4494, + "theoretical_loss": 3.5006723728413123, + "tokens_seen": 1733033984 + }, + { + "epoch": 0.53, + "learning_rate": 0.0003837913741223671, + "loss": 1.4654, + "theoretical_loss": 3.500661114378623, + "tokens_seen": 1733099520 + }, + { + "epoch": 0.53, + "learning_rate": 0.0003837753259779338, + "loss": 1.3436, + "theoretical_loss": 3.500649856460856, + "tokens_seen": 1733165056 + }, + { + "epoch": 0.53, + "learning_rate": 0.0003837592778335005, + "loss": 1.4779, + "theoretical_loss": 3.5006385990879663, + "tokens_seen": 1733230592 + }, + { + "epoch": 0.53, + "learning_rate": 0.0003837432296890672, + "loss": 1.3539, + "theoretical_loss": 3.500627342259904, + "tokens_seen": 1733296128 + }, + { + "epoch": 0.53, + "learning_rate": 0.0003837271815446339, + "loss": 1.3173, + "theoretical_loss": 3.5006160859766244, + "tokens_seen": 1733361664 + }, + { + "epoch": 0.53, + "objective/train/docs_used": 632954, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.3361942768096924, + "objective/train/theoretical_loss": 3.5006048302380792, + "objective/train/tokens_used": 1753887200, + "theoretical_loss": 3.5006048302380792, + "tokens_seen": 1733427200 + }, + { + "epoch": 0.53, + "learning_rate": 0.0003837111334002006, + "loss": 1.3455, + "theoretical_loss": 3.5006048302380792, + "tokens_seen": 1733427200 + }, + { + "epoch": 0.53, + "learning_rate": 0.0003836950852557673, + "loss": 1.3342, + "theoretical_loss": 3.5005935750442223, + "tokens_seen": 1733492736 + }, + { + "epoch": 0.53, + "learning_rate": 0.000383679037111334, + "loss": 1.3629, + "theoretical_loss": 3.5005823203950066, + "tokens_seen": 1733558272 + }, + { + "epoch": 0.53, + "learning_rate": 0.0003836629889669007, + "loss": 1.3104, + "theoretical_loss": 3.5005710662903846, + "tokens_seen": 1733623808 + }, + { + "epoch": 0.53, + "learning_rate": 0.0003836469408224674, + "loss": 1.3324, + "theoretical_loss": 3.50055981273031, + "tokens_seen": 1733689344 + }, + { + "epoch": 0.53, + "learning_rate": 0.0003836308926780341, + "loss": 1.3243, + "theoretical_loss": 3.500548559714735, + "tokens_seen": 1733754880 + }, + { + "epoch": 0.53, + "learning_rate": 0.0003836148445336008, + "loss": 1.4258, + "theoretical_loss": 3.500537307243614, + "tokens_seen": 1733820416 + }, + { + "epoch": 0.53, + "learning_rate": 0.0003835987963891675, + "loss": 1.2693, + "theoretical_loss": 3.5005260553168993, + "tokens_seen": 1733885952 + }, + { + "epoch": 0.53, + "learning_rate": 0.0003835827482447342, + "loss": 1.3704, + "theoretical_loss": 3.5005148039345437, + "tokens_seen": 1733951488 + }, + { + "epoch": 0.53, + "learning_rate": 0.0003835667001003009, + "loss": 1.372, + "theoretical_loss": 3.500503553096501, + "tokens_seen": 1734017024 + }, + { + "epoch": 0.53, + "learning_rate": 0.0003835667001003009, + "loss": 1.2746, + "theoretical_loss": 3.5004923028027237, + "tokens_seen": 1734082560 + }, + { + "epoch": 0.53, + "learning_rate": 0.00038355065195586763, + "loss": 1.4315, + "theoretical_loss": 3.500481053053165, + "tokens_seen": 1734148096 + }, + { + "epoch": 0.53, + "learning_rate": 0.00038353460381143433, + "loss": 1.2793, + "theoretical_loss": 3.5004698038477784, + "tokens_seen": 1734213632 + }, + { + "epoch": 0.53, + "learning_rate": 0.00038351855566700103, + "loss": 1.3656, + "theoretical_loss": 3.500458555186517, + "tokens_seen": 1734279168 + }, + { + "epoch": 0.53, + "learning_rate": 0.00038350250752256773, + "loss": 1.4604, + "theoretical_loss": 3.500447307069333, + "tokens_seen": 1734344704 + }, + { + "epoch": 0.53, + "learning_rate": 0.00038348645937813443, + "loss": 1.3039, + "theoretical_loss": 3.500436059496181, + "tokens_seen": 1734410240 + }, + { + "epoch": 0.53, + "learning_rate": 0.00038347041123370113, + "loss": 1.3334, + "theoretical_loss": 3.5004248124670125, + "tokens_seen": 1734475776 + }, + { + "epoch": 0.53, + "learning_rate": 0.00038345436308926783, + "loss": 1.3909, + "theoretical_loss": 3.5004135659817823, + "tokens_seen": 1734541312 + }, + { + "epoch": 0.53, + "learning_rate": 0.00038343831494483453, + "loss": 1.3966, + "theoretical_loss": 3.500402320040442, + "tokens_seen": 1734606848 + }, + { + "epoch": 0.53, + "learning_rate": 0.00038342226680040123, + "loss": 1.3792, + "theoretical_loss": 3.5003910746429456, + "tokens_seen": 1734672384 + }, + { + "epoch": 0.53, + "learning_rate": 0.00038340621865596793, + "loss": 1.3199, + "theoretical_loss": 3.5003798297892468, + "tokens_seen": 1734737920 + }, + { + "epoch": 0.53, + "learning_rate": 0.00038339017051153463, + "loss": 1.3211, + "theoretical_loss": 3.500368585479297, + "tokens_seen": 1734803456 + }, + { + "epoch": 0.53, + "learning_rate": 0.00038337412236710133, + "loss": 1.4729, + "theoretical_loss": 3.500357341713051, + "tokens_seen": 1734868992 + }, + { + "epoch": 0.53, + "learning_rate": 0.00038335807422266804, + "loss": 1.4818, + "theoretical_loss": 3.5003460984904615, + "tokens_seen": 1734934528 + }, + { + "epoch": 0.53, + "learning_rate": 0.00038334202607823474, + "loss": 1.3493, + "theoretical_loss": 3.5003348558114817, + "tokens_seen": 1735000064 + }, + { + "epoch": 0.53, + "learning_rate": 0.00038332597793380144, + "loss": 1.3579, + "theoretical_loss": 3.5003236136760645, + "tokens_seen": 1735065600 + }, + { + "epoch": 0.53, + "learning_rate": 0.00038330992978936814, + "loss": 1.389, + "theoretical_loss": 3.5003123720841636, + "tokens_seen": 1735131136 + }, + { + "epoch": 0.53, + "learning_rate": 0.00038329388164493484, + "loss": 1.3642, + "theoretical_loss": 3.500301131035731, + "tokens_seen": 1735196672 + }, + { + "epoch": 0.53, + "learning_rate": 0.00038327783350050154, + "loss": 1.3382, + "theoretical_loss": 3.5002898905307216, + "tokens_seen": 1735262208 + }, + { + "epoch": 0.53, + "learning_rate": 0.00038326178535606824, + "loss": 1.433, + "theoretical_loss": 3.5002786505690873, + "tokens_seen": 1735327744 + }, + { + "epoch": 0.53, + "learning_rate": 0.00038324573721163494, + "loss": 1.3905, + "theoretical_loss": 3.5002674111507823, + "tokens_seen": 1735393280 + }, + { + "epoch": 0.53, + "learning_rate": 0.00038322968906720164, + "loss": 1.3913, + "theoretical_loss": 3.5002561722757592, + "tokens_seen": 1735458816 + }, + { + "epoch": 0.53, + "learning_rate": 0.00038321364092276834, + "loss": 1.392, + "theoretical_loss": 3.500244933943971, + "tokens_seen": 1735524352 + }, + { + "epoch": 0.53, + "learning_rate": 0.00038319759277833504, + "loss": 1.369, + "theoretical_loss": 3.5002336961553713, + "tokens_seen": 1735589888 + }, + { + "epoch": 0.53, + "learning_rate": 0.00038318154463390174, + "loss": 1.3187, + "theoretical_loss": 3.5002224589099136, + "tokens_seen": 1735655424 + }, + { + "epoch": 0.53, + "learning_rate": 0.00038316549648946844, + "loss": 1.3199, + "theoretical_loss": 3.5002112222075508, + "tokens_seen": 1735720960 + }, + { + "epoch": 0.53, + "learning_rate": 0.00038314944834503514, + "loss": 1.3449, + "theoretical_loss": 3.5001999860482362, + "tokens_seen": 1735786496 + }, + { + "epoch": 0.53, + "learning_rate": 0.00038313340020060184, + "loss": 1.3757, + "theoretical_loss": 3.500188750431923, + "tokens_seen": 1735852032 + }, + { + "epoch": 0.53, + "learning_rate": 0.00038311735205616854, + "loss": 1.3485, + "theoretical_loss": 3.5001775153585646, + "tokens_seen": 1735917568 + }, + { + "epoch": 0.53, + "learning_rate": 0.00038310130391173525, + "loss": 1.4501, + "theoretical_loss": 3.5001662808281138, + "tokens_seen": 1735983104 + }, + { + "epoch": 0.53, + "learning_rate": 0.00038308525576730195, + "loss": 1.3668, + "theoretical_loss": 3.5001550468405247, + "tokens_seen": 1736048640 + }, + { + "epoch": 0.53, + "learning_rate": 0.00038306920762286865, + "loss": 1.305, + "theoretical_loss": 3.5001438133957494, + "tokens_seen": 1736114176 + }, + { + "epoch": 0.53, + "learning_rate": 0.0003830531594784353, + "loss": 1.3909, + "theoretical_loss": 3.5001325804937427, + "tokens_seen": 1736179712 + }, + { + "epoch": 0.53, + "learning_rate": 0.000383037111334002, + "loss": 1.402, + "theoretical_loss": 3.500121348134457, + "tokens_seen": 1736245248 + }, + { + "epoch": 0.53, + "learning_rate": 0.0003830210631895687, + "loss": 1.2602, + "theoretical_loss": 3.5001101163178454, + "tokens_seen": 1736310784 + }, + { + "epoch": 0.53, + "learning_rate": 0.0003830050150451354, + "loss": 1.3519, + "theoretical_loss": 3.5000988850438617, + "tokens_seen": 1736376320 + }, + { + "epoch": 0.53, + "learning_rate": 0.0003829889669007021, + "loss": 1.4119, + "theoretical_loss": 3.500087654312459, + "tokens_seen": 1736441856 + }, + { + "epoch": 0.53, + "learning_rate": 0.0003829729187562688, + "loss": 1.3056, + "theoretical_loss": 3.5000764241235904, + "tokens_seen": 1736507392 + }, + { + "epoch": 0.53, + "learning_rate": 0.0003829568706118355, + "loss": 1.3016, + "theoretical_loss": 3.500065194477209, + "tokens_seen": 1736572928 + }, + { + "epoch": 0.53, + "learning_rate": 0.0003829408224674022, + "loss": 1.3816, + "theoretical_loss": 3.5000539653732696, + "tokens_seen": 1736638464 + }, + { + "epoch": 0.53, + "objective/train/docs_used": 634251, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.373918056488037, + "objective/train/theoretical_loss": 3.5000427368117233, + "objective/train/tokens_used": 1757164000, + "theoretical_loss": 3.5000427368117233, + "tokens_seen": 1736704000 + }, + { + "epoch": 0.53, + "learning_rate": 0.0003829247743229689, + "loss": 1.446, + "theoretical_loss": 3.5000427368117233, + "tokens_seen": 1736704000 + }, + { + "epoch": 0.53, + "learning_rate": 0.0003829087261785356, + "loss": 1.3633, + "theoretical_loss": 3.5000315087925253, + "tokens_seen": 1736769536 + }, + { + "epoch": 0.53, + "learning_rate": 0.0003828926780341023, + "loss": 1.4103, + "theoretical_loss": 3.5000202813156287, + "tokens_seen": 1736835072 + }, + { + "epoch": 0.53, + "learning_rate": 0.000382876629889669, + "loss": 1.3835, + "theoretical_loss": 3.5000090543809854, + "tokens_seen": 1736900608 + }, + { + "epoch": 0.53, + "learning_rate": 0.0003828605817452357, + "loss": 1.4208, + "theoretical_loss": 3.4999978279885506, + "tokens_seen": 1736966144 + }, + { + "epoch": 0.53, + "learning_rate": 0.0003828445336008024, + "loss": 1.335, + "theoretical_loss": 3.4999866021382764, + "tokens_seen": 1737031680 + }, + { + "epoch": 0.53, + "learning_rate": 0.0003828284854563691, + "loss": 1.4071, + "theoretical_loss": 3.4999753768301165, + "tokens_seen": 1737097216 + }, + { + "epoch": 0.53, + "learning_rate": 0.0003828124373119358, + "loss": 1.3394, + "theoretical_loss": 3.499964152064025, + "tokens_seen": 1737162752 + }, + { + "epoch": 0.53, + "learning_rate": 0.0003827963891675025, + "loss": 1.4081, + "theoretical_loss": 3.4999529278399537, + "tokens_seen": 1737228288 + }, + { + "epoch": 0.53, + "learning_rate": 0.0003827803410230692, + "loss": 1.4089, + "theoretical_loss": 3.499941704157857, + "tokens_seen": 1737293824 + }, + { + "epoch": 0.53, + "learning_rate": 0.0003827642928786359, + "loss": 1.3743, + "theoretical_loss": 3.4999304810176888, + "tokens_seen": 1737359360 + }, + { + "epoch": 0.53, + "learning_rate": 0.0003827482447342026, + "loss": 1.3529, + "theoretical_loss": 3.499919258419401, + "tokens_seen": 1737424896 + }, + { + "epoch": 0.53, + "learning_rate": 0.0003827321965897693, + "loss": 1.3054, + "theoretical_loss": 3.4999080363629487, + "tokens_seen": 1737490432 + }, + { + "epoch": 0.53, + "learning_rate": 0.000382716148445336, + "loss": 1.2204, + "theoretical_loss": 3.4998968148482845, + "tokens_seen": 1737555968 + }, + { + "epoch": 0.53, + "learning_rate": 0.0003827001003009027, + "loss": 1.2942, + "theoretical_loss": 3.4998855938753612, + "tokens_seen": 1737621504 + }, + { + "epoch": 0.53, + "learning_rate": 0.0003826840521564694, + "loss": 1.3862, + "theoretical_loss": 3.499874373444133, + "tokens_seen": 1737687040 + }, + { + "epoch": 0.53, + "learning_rate": 0.0003826680040120361, + "loss": 1.414, + "theoretical_loss": 3.4998631535545535, + "tokens_seen": 1737752576 + }, + { + "epoch": 0.53, + "learning_rate": 0.0003826519558676028, + "loss": 1.4308, + "theoretical_loss": 3.4998519342065753, + "tokens_seen": 1737818112 + }, + { + "epoch": 0.53, + "learning_rate": 0.0003826359077231695, + "loss": 1.309, + "theoretical_loss": 3.4998407154001523, + "tokens_seen": 1737883648 + }, + { + "epoch": 0.53, + "learning_rate": 0.0003826198595787362, + "loss": 1.2648, + "theoretical_loss": 3.499829497135238, + "tokens_seen": 1737949184 + }, + { + "epoch": 0.53, + "learning_rate": 0.0003826038114343029, + "loss": 1.3533, + "theoretical_loss": 3.4998182794117856, + "tokens_seen": 1738014720 + }, + { + "epoch": 0.53, + "learning_rate": 0.0003825877632898696, + "loss": 1.4204, + "theoretical_loss": 3.499807062229749, + "tokens_seen": 1738080256 + }, + { + "epoch": 0.53, + "learning_rate": 0.00038257171514543637, + "loss": 1.3383, + "theoretical_loss": 3.4997958455890816, + "tokens_seen": 1738145792 + }, + { + "epoch": 0.53, + "learning_rate": 0.00038255566700100307, + "loss": 1.3415, + "theoretical_loss": 3.4997846294897363, + "tokens_seen": 1738211328 + }, + { + "epoch": 0.53, + "learning_rate": 0.00038253961885656977, + "loss": 1.3624, + "theoretical_loss": 3.499773413931667, + "tokens_seen": 1738276864 + }, + { + "epoch": 0.53, + "learning_rate": 0.00038252357071213647, + "loss": 1.3862, + "theoretical_loss": 3.4997621989148273, + "tokens_seen": 1738342400 + }, + { + "epoch": 0.53, + "learning_rate": 0.00038250752256770317, + "loss": 1.2914, + "theoretical_loss": 3.49975098443917, + "tokens_seen": 1738407936 + }, + { + "epoch": 0.53, + "learning_rate": 0.00038249147442326987, + "loss": 1.3393, + "theoretical_loss": 3.4997397705046493, + "tokens_seen": 1738473472 + }, + { + "epoch": 0.53, + "learning_rate": 0.00038247542627883657, + "loss": 1.3463, + "theoretical_loss": 3.4997285571112178, + "tokens_seen": 1738539008 + }, + { + "epoch": 0.53, + "learning_rate": 0.00038245937813440327, + "loss": 1.2923, + "theoretical_loss": 3.4997173442588307, + "tokens_seen": 1738604544 + }, + { + "epoch": 0.53, + "learning_rate": 0.00038244332998996997, + "loss": 1.3035, + "theoretical_loss": 3.4997061319474394, + "tokens_seen": 1738670080 + }, + { + "epoch": 0.53, + "learning_rate": 0.0003824272818455366, + "loss": 1.2809, + "theoretical_loss": 3.4996949201769993, + "tokens_seen": 1738735616 + }, + { + "epoch": 0.53, + "learning_rate": 0.0003824112337011033, + "loss": 1.3708, + "theoretical_loss": 3.4996837089474626, + "tokens_seen": 1738801152 + }, + { + "epoch": 0.53, + "learning_rate": 0.00038239518555667, + "loss": 1.3128, + "theoretical_loss": 3.4996724982587835, + "tokens_seen": 1738866688 + }, + { + "epoch": 0.53, + "learning_rate": 0.0003823791374122367, + "loss": 1.2722, + "theoretical_loss": 3.4996612881109153, + "tokens_seen": 1738932224 + }, + { + "epoch": 0.53, + "learning_rate": 0.0003823630892678034, + "loss": 1.3303, + "theoretical_loss": 3.4996500785038114, + "tokens_seen": 1738997760 + }, + { + "epoch": 0.53, + "learning_rate": 0.0003823470411233701, + "loss": 1.3723, + "theoretical_loss": 3.499638869437425, + "tokens_seen": 1739063296 + }, + { + "epoch": 0.53, + "learning_rate": 0.0003823309929789368, + "loss": 1.4472, + "theoretical_loss": 3.4996276609117105, + "tokens_seen": 1739128832 + }, + { + "epoch": 0.53, + "learning_rate": 0.0003823149448345035, + "loss": 1.4118, + "theoretical_loss": 3.499616452926621, + "tokens_seen": 1739194368 + }, + { + "epoch": 0.53, + "learning_rate": 0.0003822988966900702, + "loss": 1.3578, + "theoretical_loss": 3.4996052454821105, + "tokens_seen": 1739259904 + }, + { + "epoch": 0.53, + "learning_rate": 0.0003822828485456369, + "loss": 1.3623, + "theoretical_loss": 3.499594038578132, + "tokens_seen": 1739325440 + }, + { + "epoch": 0.53, + "learning_rate": 0.0003822668004012036, + "loss": 1.3867, + "theoretical_loss": 3.4995828322146387, + "tokens_seen": 1739390976 + }, + { + "epoch": 0.53, + "learning_rate": 0.0003822507522567703, + "loss": 1.4626, + "theoretical_loss": 3.4995716263915853, + "tokens_seen": 1739456512 + }, + { + "epoch": 0.53, + "learning_rate": 0.000382234704112337, + "loss": 1.3628, + "theoretical_loss": 3.4995604211089244, + "tokens_seen": 1739522048 + }, + { + "epoch": 0.53, + "learning_rate": 0.0003822186559679037, + "loss": 1.3797, + "theoretical_loss": 3.49954921636661, + "tokens_seen": 1739587584 + }, + { + "epoch": 0.53, + "learning_rate": 0.0003822026078234704, + "loss": 1.2928, + "theoretical_loss": 3.499538012164596, + "tokens_seen": 1739653120 + }, + { + "epoch": 0.53, + "learning_rate": 0.00038218655967903713, + "loss": 1.4461, + "theoretical_loss": 3.499526808502835, + "tokens_seen": 1739718656 + }, + { + "epoch": 0.53, + "learning_rate": 0.00038217051153460383, + "loss": 1.3561, + "theoretical_loss": 3.4995156053812817, + "tokens_seen": 1739784192 + }, + { + "epoch": 0.53, + "learning_rate": 0.00038215446339017053, + "loss": 1.3479, + "theoretical_loss": 3.4995044027998894, + "tokens_seen": 1739849728 + }, + { + "epoch": 0.53, + "learning_rate": 0.00038213841524573723, + "loss": 1.2886, + "theoretical_loss": 3.499493200758611, + "tokens_seen": 1739915264 + }, + { + "epoch": 0.53, + "objective/train/docs_used": 635535, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.3093488216400146, + "objective/train/theoretical_loss": 3.499481999257401, + "objective/train/tokens_used": 1760440800, + "theoretical_loss": 3.499481999257401, + "tokens_seen": 1739980800 + }, + { + "epoch": 0.53, + "learning_rate": 0.00038212236710130393, + "loss": 1.3074, + "theoretical_loss": 3.499481999257401, + "tokens_seen": 1739980800 + }, + { + "epoch": 0.53, + "learning_rate": 0.00038210631895687063, + "loss": 1.3506, + "theoretical_loss": 3.4994707982962128, + "tokens_seen": 1740046336 + }, + { + "epoch": 0.53, + "learning_rate": 0.00038209027081243733, + "loss": 1.3099, + "theoretical_loss": 3.499459597875, + "tokens_seen": 1740111872 + }, + { + "epoch": 0.53, + "learning_rate": 0.00038207422266800403, + "loss": 1.3647, + "theoretical_loss": 3.4994483979937154, + "tokens_seen": 1740177408 + }, + { + "epoch": 0.53, + "learning_rate": 0.00038205817452357073, + "loss": 1.3228, + "theoretical_loss": 3.4994371986523145, + "tokens_seen": 1740242944 + }, + { + "epoch": 0.53, + "learning_rate": 0.00038204212637913743, + "loss": 1.3654, + "theoretical_loss": 3.499425999850749, + "tokens_seen": 1740308480 + }, + { + "epoch": 0.53, + "learning_rate": 0.00038202607823470413, + "loss": 1.2629, + "theoretical_loss": 3.4994148015889737, + "tokens_seen": 1740374016 + }, + { + "epoch": 0.53, + "learning_rate": 0.00038201003009027083, + "loss": 1.3177, + "theoretical_loss": 3.499403603866942, + "tokens_seen": 1740439552 + }, + { + "epoch": 0.53, + "learning_rate": 0.00038199398194583753, + "loss": 1.434, + "theoretical_loss": 3.499392406684607, + "tokens_seen": 1740505088 + }, + { + "epoch": 0.53, + "learning_rate": 0.00038197793380140424, + "loss": 1.3584, + "theoretical_loss": 3.4993812100419235, + "tokens_seen": 1740570624 + }, + { + "epoch": 0.53, + "learning_rate": 0.00038196188565697094, + "loss": 1.3447, + "theoretical_loss": 3.4993700139388446, + "tokens_seen": 1740636160 + }, + { + "epoch": 0.53, + "learning_rate": 0.00038194583751253764, + "loss": 1.3515, + "theoretical_loss": 3.4993588183753235, + "tokens_seen": 1740701696 + }, + { + "epoch": 0.53, + "learning_rate": 0.00038192978936810434, + "loss": 1.3876, + "theoretical_loss": 3.4993476233513148, + "tokens_seen": 1740767232 + }, + { + "epoch": 0.53, + "learning_rate": 0.00038191374122367104, + "loss": 1.2978, + "theoretical_loss": 3.499336428866771, + "tokens_seen": 1740832768 + }, + { + "epoch": 0.53, + "learning_rate": 0.00038189769307923774, + "loss": 1.3822, + "theoretical_loss": 3.4993252349216473, + "tokens_seen": 1740898304 + }, + { + "epoch": 0.53, + "learning_rate": 0.00038188164493480444, + "loss": 1.3262, + "theoretical_loss": 3.4993140415158956, + "tokens_seen": 1740963840 + }, + { + "epoch": 0.53, + "learning_rate": 0.00038186559679037114, + "loss": 1.2955, + "theoretical_loss": 3.4993028486494717, + "tokens_seen": 1741029376 + }, + { + "epoch": 0.53, + "learning_rate": 0.00038184954864593784, + "loss": 1.3149, + "theoretical_loss": 3.4992916563223275, + "tokens_seen": 1741094912 + }, + { + "epoch": 0.53, + "learning_rate": 0.00038183350050150454, + "loss": 1.3672, + "theoretical_loss": 3.4992804645344178, + "tokens_seen": 1741160448 + }, + { + "epoch": 0.53, + "learning_rate": 0.00038181745235707124, + "loss": 1.2266, + "theoretical_loss": 3.499269273285696, + "tokens_seen": 1741225984 + }, + { + "epoch": 0.53, + "learning_rate": 0.00038180140421263794, + "loss": 1.3973, + "theoretical_loss": 3.499258082576116, + "tokens_seen": 1741291520 + }, + { + "epoch": 0.53, + "learning_rate": 0.00038178535606820464, + "loss": 1.2973, + "theoretical_loss": 3.4992468924056306, + "tokens_seen": 1741357056 + }, + { + "epoch": 0.53, + "learning_rate": 0.00038176930792377134, + "loss": 1.3906, + "theoretical_loss": 3.499235702774195, + "tokens_seen": 1741422592 + }, + { + "epoch": 0.53, + "learning_rate": 0.00038175325977933804, + "loss": 1.2166, + "theoretical_loss": 3.499224513681762, + "tokens_seen": 1741488128 + }, + { + "epoch": 0.53, + "learning_rate": 0.00038173721163490474, + "loss": 1.4198, + "theoretical_loss": 3.4992133251282853, + "tokens_seen": 1741553664 + }, + { + "epoch": 0.53, + "learning_rate": 0.00038172116349047145, + "loss": 1.3005, + "theoretical_loss": 3.499202137113719, + "tokens_seen": 1741619200 + }, + { + "epoch": 0.53, + "learning_rate": 0.00038170511534603815, + "loss": 1.4039, + "theoretical_loss": 3.499190949638017, + "tokens_seen": 1741684736 + }, + { + "epoch": 0.53, + "learning_rate": 0.00038168906720160485, + "loss": 1.3796, + "theoretical_loss": 3.499179762701133, + "tokens_seen": 1741750272 + }, + { + "epoch": 0.53, + "learning_rate": 0.00038167301905717155, + "loss": 1.2885, + "theoretical_loss": 3.4991685763030205, + "tokens_seen": 1741815808 + }, + { + "epoch": 0.53, + "learning_rate": 0.0003816569709127382, + "loss": 1.3196, + "theoretical_loss": 3.4991573904436333, + "tokens_seen": 1741881344 + }, + { + "epoch": 0.53, + "learning_rate": 0.0003816409227683049, + "loss": 1.395, + "theoretical_loss": 3.499146205122926, + "tokens_seen": 1741946880 + }, + { + "epoch": 0.53, + "learning_rate": 0.0003816248746238716, + "loss": 1.3878, + "theoretical_loss": 3.4991350203408507, + "tokens_seen": 1742012416 + }, + { + "epoch": 0.53, + "learning_rate": 0.0003816088264794383, + "loss": 1.5152, + "theoretical_loss": 3.499123836097363, + "tokens_seen": 1742077952 + }, + { + "epoch": 0.53, + "learning_rate": 0.000381592778335005, + "loss": 1.3517, + "theoretical_loss": 3.499112652392416, + "tokens_seen": 1742143488 + }, + { + "epoch": 0.53, + "learning_rate": 0.0003815767301905717, + "loss": 1.3679, + "theoretical_loss": 3.499101469225963, + "tokens_seen": 1742209024 + }, + { + "epoch": 0.53, + "learning_rate": 0.0003815606820461384, + "loss": 1.3515, + "theoretical_loss": 3.4990902865979585, + "tokens_seen": 1742274560 + }, + { + "epoch": 0.53, + "learning_rate": 0.0003815446339017051, + "loss": 1.2677, + "theoretical_loss": 3.499079104508356, + "tokens_seen": 1742340096 + }, + { + "epoch": 0.53, + "learning_rate": 0.0003815285857572718, + "loss": 1.3923, + "theoretical_loss": 3.499067922957109, + "tokens_seen": 1742405632 + }, + { + "epoch": 0.53, + "learning_rate": 0.0003815125376128385, + "loss": 1.3049, + "theoretical_loss": 3.4990567419441723, + "tokens_seen": 1742471168 + }, + { + "epoch": 0.53, + "learning_rate": 0.0003814964894684052, + "loss": 1.3293, + "theoretical_loss": 3.499045561469499, + "tokens_seen": 1742536704 + }, + { + "epoch": 0.53, + "learning_rate": 0.0003814804413239719, + "loss": 1.3438, + "theoretical_loss": 3.4990343815330434, + "tokens_seen": 1742602240 + }, + { + "epoch": 0.53, + "learning_rate": 0.0003814643931795386, + "loss": 1.529, + "theoretical_loss": 3.4990232021347585, + "tokens_seen": 1742667776 + }, + { + "epoch": 0.53, + "learning_rate": 0.0003814483450351053, + "loss": 1.3334, + "theoretical_loss": 3.4990120232745987, + "tokens_seen": 1742733312 + }, + { + "epoch": 0.53, + "learning_rate": 0.000381432296890672, + "loss": 1.4556, + "theoretical_loss": 3.4990008449525183, + "tokens_seen": 1742798848 + }, + { + "epoch": 0.53, + "learning_rate": 0.0003814162487462387, + "loss": 1.3854, + "theoretical_loss": 3.4989896671684706, + "tokens_seen": 1742864384 + }, + { + "epoch": 0.53, + "learning_rate": 0.0003814002006018054, + "loss": 1.4211, + "theoretical_loss": 3.49897848992241, + "tokens_seen": 1742929920 + }, + { + "epoch": 0.53, + "learning_rate": 0.0003813841524573721, + "loss": 1.3277, + "theoretical_loss": 3.4989673132142896, + "tokens_seen": 1742995456 + }, + { + "epoch": 0.53, + "learning_rate": 0.00038136810431293886, + "loss": 1.4065, + "theoretical_loss": 3.498956137044064, + "tokens_seen": 1743060992 + }, + { + "epoch": 0.53, + "learning_rate": 0.00038135205616850556, + "loss": 1.3377, + "theoretical_loss": 3.498944961411687, + "tokens_seen": 1743126528 + }, + { + "epoch": 0.53, + "learning_rate": 0.00038133600802407226, + "loss": 1.4824, + "theoretical_loss": 3.4989337863171115, + "tokens_seen": 1743192064 + }, + { + "epoch": 0.53, + "objective/train/docs_used": 636817, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.3738594055175781, + "objective/train/theoretical_loss": 3.4989226117602925, + "objective/train/tokens_used": 1763717600, + "theoretical_loss": 3.4989226117602925, + "tokens_seen": 1743257600 + }, + { + "epoch": 0.53, + "learning_rate": 0.00038131995987963896, + "loss": 1.3613, + "theoretical_loss": 3.4989226117602925, + "tokens_seen": 1743257600 + }, + { + "epoch": 0.53, + "learning_rate": 0.00038130391173520566, + "loss": 1.3841, + "theoretical_loss": 3.498911437741184, + "tokens_seen": 1743323136 + }, + { + "epoch": 0.53, + "learning_rate": 0.00038128786359077236, + "loss": 1.4394, + "theoretical_loss": 3.4989002642597393, + "tokens_seen": 1743388672 + }, + { + "epoch": 0.53, + "learning_rate": 0.00038127181544633906, + "loss": 1.354, + "theoretical_loss": 3.4988890913159123, + "tokens_seen": 1743454208 + }, + { + "epoch": 0.53, + "learning_rate": 0.00038125576730190576, + "loss": 1.3447, + "theoretical_loss": 3.498877918909657, + "tokens_seen": 1743519744 + }, + { + "epoch": 0.53, + "learning_rate": 0.00038123971915747246, + "loss": 1.3379, + "theoretical_loss": 3.4988667470409283, + "tokens_seen": 1743585280 + }, + { + "epoch": 0.53, + "learning_rate": 0.00038122367101303917, + "loss": 1.388, + "theoretical_loss": 3.498855575709679, + "tokens_seen": 1743650816 + }, + { + "epoch": 0.53, + "learning_rate": 0.00038120762286860587, + "loss": 1.3845, + "theoretical_loss": 3.4988444049158627, + "tokens_seen": 1743716352 + }, + { + "epoch": 0.53, + "learning_rate": 0.00038119157472417257, + "loss": 1.3739, + "theoretical_loss": 3.4988332346594344, + "tokens_seen": 1743781888 + }, + { + "epoch": 0.53, + "learning_rate": 0.00038117552657973927, + "loss": 1.2292, + "theoretical_loss": 3.498822064940348, + "tokens_seen": 1743847424 + }, + { + "epoch": 0.53, + "learning_rate": 0.00038115947843530597, + "loss": 1.4854, + "theoretical_loss": 3.498810895758557, + "tokens_seen": 1743912960 + }, + { + "epoch": 0.53, + "learning_rate": 0.00038114343029087267, + "loss": 1.4092, + "theoretical_loss": 3.498799727114015, + "tokens_seen": 1743978496 + }, + { + "epoch": 0.53, + "learning_rate": 0.00038112738214643937, + "loss": 1.4147, + "theoretical_loss": 3.4987885590066767, + "tokens_seen": 1744044032 + }, + { + "epoch": 0.53, + "learning_rate": 0.00038111133400200607, + "loss": 1.3624, + "theoretical_loss": 3.498777391436496, + "tokens_seen": 1744109568 + }, + { + "epoch": 0.53, + "learning_rate": 0.00038109528585757277, + "loss": 1.2575, + "theoretical_loss": 3.4987662244034268, + "tokens_seen": 1744175104 + }, + { + "epoch": 0.53, + "learning_rate": 0.00038107923771313947, + "loss": 1.2358, + "theoretical_loss": 3.498755057907423, + "tokens_seen": 1744240640 + }, + { + "epoch": 0.53, + "learning_rate": 0.00038106318956870617, + "loss": 1.2356, + "theoretical_loss": 3.4987438919484384, + "tokens_seen": 1744306176 + }, + { + "epoch": 0.53, + "learning_rate": 0.00038104714142427287, + "loss": 1.3553, + "theoretical_loss": 3.498732726526427, + "tokens_seen": 1744371712 + }, + { + "epoch": 0.53, + "learning_rate": 0.0003810310932798396, + "loss": 1.3476, + "theoretical_loss": 3.4987215616413434, + "tokens_seen": 1744437248 + }, + { + "epoch": 0.53, + "learning_rate": 0.0003810150451354062, + "loss": 1.3057, + "theoretical_loss": 3.498710397293141, + "tokens_seen": 1744502784 + }, + { + "epoch": 0.53, + "learning_rate": 0.0003809989969909729, + "loss": 1.3584, + "theoretical_loss": 3.498699233481774, + "tokens_seen": 1744568320 + }, + { + "epoch": 0.53, + "learning_rate": 0.0003809829488465396, + "loss": 1.3228, + "theoretical_loss": 3.498688070207197, + "tokens_seen": 1744633856 + }, + { + "epoch": 0.53, + "learning_rate": 0.0003809669007021063, + "loss": 1.3921, + "theoretical_loss": 3.4986769074693624, + "tokens_seen": 1744699392 + }, + { + "epoch": 0.53, + "learning_rate": 0.000380950852557673, + "loss": 1.4404, + "theoretical_loss": 3.498665745268226, + "tokens_seen": 1744764928 + }, + { + "epoch": 0.53, + "learning_rate": 0.0003809348044132397, + "loss": 1.303, + "theoretical_loss": 3.498654583603741, + "tokens_seen": 1744830464 + }, + { + "epoch": 0.53, + "learning_rate": 0.0003809187562688064, + "loss": 1.3636, + "theoretical_loss": 3.498643422475862, + "tokens_seen": 1744896000 + }, + { + "epoch": 0.53, + "learning_rate": 0.0003809027081243731, + "loss": 1.5368, + "theoretical_loss": 3.4986322618845422, + "tokens_seen": 1744961536 + }, + { + "epoch": 0.53, + "learning_rate": 0.0003808866599799398, + "loss": 1.2788, + "theoretical_loss": 3.4986211018297357, + "tokens_seen": 1745027072 + }, + { + "epoch": 0.53, + "learning_rate": 0.0003808706118355065, + "loss": 1.3321, + "theoretical_loss": 3.498609942311398, + "tokens_seen": 1745092608 + }, + { + "epoch": 0.53, + "learning_rate": 0.0003808545636910732, + "loss": 1.2876, + "theoretical_loss": 3.4985987833294816, + "tokens_seen": 1745158144 + }, + { + "epoch": 0.53, + "learning_rate": 0.0003808385155466399, + "loss": 1.4598, + "theoretical_loss": 3.4985876248839407, + "tokens_seen": 1745223680 + }, + { + "epoch": 0.53, + "learning_rate": 0.0003808224674022066, + "loss": 1.3571, + "theoretical_loss": 3.49857646697473, + "tokens_seen": 1745289216 + }, + { + "epoch": 0.53, + "learning_rate": 0.00038080641925777333, + "loss": 1.3798, + "theoretical_loss": 3.498565309601804, + "tokens_seen": 1745354752 + }, + { + "epoch": 0.53, + "learning_rate": 0.00038079037111334003, + "loss": 1.3836, + "theoretical_loss": 3.498554152765115, + "tokens_seen": 1745420288 + }, + { + "epoch": 0.53, + "learning_rate": 0.00038077432296890673, + "loss": 1.2496, + "theoretical_loss": 3.498542996464619, + "tokens_seen": 1745485824 + }, + { + "epoch": 0.53, + "learning_rate": 0.00038075827482447343, + "loss": 1.4693, + "theoretical_loss": 3.4985318407002692, + "tokens_seen": 1745551360 + }, + { + "epoch": 0.53, + "learning_rate": 0.00038074222668004013, + "loss": 1.303, + "theoretical_loss": 3.4985206854720197, + "tokens_seen": 1745616896 + }, + { + "epoch": 0.53, + "learning_rate": 0.00038072617853560683, + "loss": 1.4274, + "theoretical_loss": 3.498509530779825, + "tokens_seen": 1745682432 + }, + { + "epoch": 0.53, + "learning_rate": 0.00038071013039117353, + "loss": 1.3862, + "theoretical_loss": 3.4984983766236386, + "tokens_seen": 1745747968 + }, + { + "epoch": 0.53, + "learning_rate": 0.00038069408224674023, + "loss": 1.4058, + "theoretical_loss": 3.498487223003415, + "tokens_seen": 1745813504 + }, + { + "epoch": 0.53, + "learning_rate": 0.00038067803410230693, + "loss": 1.3898, + "theoretical_loss": 3.498476069919109, + "tokens_seen": 1745879040 + }, + { + "epoch": 0.53, + "learning_rate": 0.00038066198595787363, + "loss": 1.3972, + "theoretical_loss": 3.498464917370673, + "tokens_seen": 1745944576 + }, + { + "epoch": 0.53, + "learning_rate": 0.00038064593781344033, + "loss": 1.3327, + "theoretical_loss": 3.498453765358063, + "tokens_seen": 1746010112 + }, + { + "epoch": 0.53, + "learning_rate": 0.00038062988966900703, + "loss": 1.4642, + "theoretical_loss": 3.498442613881232, + "tokens_seen": 1746075648 + }, + { + "epoch": 0.53, + "learning_rate": 0.00038061384152457374, + "loss": 1.3782, + "theoretical_loss": 3.4984314629401347, + "tokens_seen": 1746141184 + }, + { + "epoch": 0.53, + "learning_rate": 0.00038059779338014044, + "loss": 1.4065, + "theoretical_loss": 3.498420312534725, + "tokens_seen": 1746206720 + }, + { + "epoch": 0.53, + "learning_rate": 0.00038058174523570714, + "loss": 1.3437, + "theoretical_loss": 3.498409162664957, + "tokens_seen": 1746272256 + }, + { + "epoch": 0.53, + "learning_rate": 0.00038056569709127384, + "loss": 1.4418, + "theoretical_loss": 3.4983980133307853, + "tokens_seen": 1746337792 + }, + { + "epoch": 0.53, + "learning_rate": 0.00038054964894684054, + "loss": 1.2878, + "theoretical_loss": 3.498386864532163, + "tokens_seen": 1746403328 + }, + { + "epoch": 0.53, + "learning_rate": 0.00038053360080240724, + "loss": 1.2113, + "theoretical_loss": 3.498375716269046, + "tokens_seen": 1746468864 + }, + { + "epoch": 0.53, + "objective/train/docs_used": 637932, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.4134416580200195, + "objective/train/theoretical_loss": 3.498364568541387, + "objective/train/tokens_used": 1766994400, + "theoretical_loss": 3.498364568541387, + "tokens_seen": 1746534400 + }, + { + "epoch": 0.53, + "learning_rate": 0.00038051755265797394, + "loss": 1.3184, + "theoretical_loss": 3.498364568541387, + "tokens_seen": 1746534400 + }, + { + "epoch": 0.53, + "learning_rate": 0.00038050150451354064, + "loss": 1.4147, + "theoretical_loss": 3.4983534213491407, + "tokens_seen": 1746599936 + }, + { + "epoch": 0.53, + "learning_rate": 0.00038048545636910734, + "loss": 1.3074, + "theoretical_loss": 3.4983422746922614, + "tokens_seen": 1746665472 + }, + { + "epoch": 0.53, + "learning_rate": 0.00038046940822467404, + "loss": 1.387, + "theoretical_loss": 3.4983311285707033, + "tokens_seen": 1746731008 + }, + { + "epoch": 0.53, + "learning_rate": 0.00038045336008024074, + "loss": 1.3219, + "theoretical_loss": 3.4983199829844205, + "tokens_seen": 1746796544 + }, + { + "epoch": 0.53, + "learning_rate": 0.00038043731193580744, + "loss": 1.3603, + "theoretical_loss": 3.498308837933367, + "tokens_seen": 1746862080 + }, + { + "epoch": 0.53, + "learning_rate": 0.00038042126379137414, + "loss": 1.3785, + "theoretical_loss": 3.4982976934174976, + "tokens_seen": 1746927616 + }, + { + "epoch": 0.53, + "learning_rate": 0.00038040521564694084, + "loss": 1.275, + "theoretical_loss": 3.4982865494367656, + "tokens_seen": 1746993152 + }, + { + "epoch": 0.53, + "learning_rate": 0.00038038916750250754, + "loss": 1.2862, + "theoretical_loss": 3.4982754059911265, + "tokens_seen": 1747058688 + }, + { + "epoch": 0.53, + "learning_rate": 0.00038037311935807424, + "loss": 1.2283, + "theoretical_loss": 3.498264263080533, + "tokens_seen": 1747124224 + }, + { + "epoch": 0.53, + "learning_rate": 0.00038035707121364095, + "loss": 1.3111, + "theoretical_loss": 3.498253120704941, + "tokens_seen": 1747189760 + }, + { + "epoch": 0.53, + "learning_rate": 0.00038034102306920765, + "loss": 1.3198, + "theoretical_loss": 3.4982419788643035, + "tokens_seen": 1747255296 + }, + { + "epoch": 0.53, + "learning_rate": 0.00038032497492477435, + "loss": 1.4114, + "theoretical_loss": 3.498230837558576, + "tokens_seen": 1747320832 + }, + { + "epoch": 0.53, + "learning_rate": 0.00038030892678034105, + "loss": 1.3271, + "theoretical_loss": 3.498219696787711, + "tokens_seen": 1747386368 + }, + { + "epoch": 0.53, + "learning_rate": 0.00038029287863590775, + "loss": 1.4799, + "theoretical_loss": 3.498208556551664, + "tokens_seen": 1747451904 + }, + { + "epoch": 0.53, + "learning_rate": 0.00038027683049147445, + "loss": 1.3645, + "theoretical_loss": 3.4981974168503887, + "tokens_seen": 1747517440 + }, + { + "epoch": 0.53, + "learning_rate": 0.00038026078234704115, + "loss": 1.4741, + "theoretical_loss": 3.49818627768384, + "tokens_seen": 1747582976 + }, + { + "epoch": 0.53, + "learning_rate": 0.0003802447342026078, + "loss": 1.2428, + "theoretical_loss": 3.4981751390519715, + "tokens_seen": 1747648512 + }, + { + "epoch": 0.53, + "learning_rate": 0.0003802286860581745, + "loss": 1.3196, + "theoretical_loss": 3.498164000954738, + "tokens_seen": 1747714048 + }, + { + "epoch": 0.53, + "learning_rate": 0.0003802126379137412, + "loss": 1.3921, + "theoretical_loss": 3.498152863392094, + "tokens_seen": 1747779584 + }, + { + "epoch": 0.53, + "learning_rate": 0.0003801965897693079, + "loss": 1.3164, + "theoretical_loss": 3.498141726363993, + "tokens_seen": 1747845120 + }, + { + "epoch": 0.53, + "learning_rate": 0.0003801805416248746, + "loss": 1.3528, + "theoretical_loss": 3.4981305898703896, + "tokens_seen": 1747910656 + }, + { + "epoch": 0.53, + "learning_rate": 0.00038016449348044135, + "loss": 1.4356, + "theoretical_loss": 3.4981194539112384, + "tokens_seen": 1747976192 + }, + { + "epoch": 0.53, + "learning_rate": 0.00038014844533600805, + "loss": 1.2635, + "theoretical_loss": 3.4981083184864934, + "tokens_seen": 1748041728 + }, + { + "epoch": 0.53, + "learning_rate": 0.00038013239719157475, + "loss": 1.4344, + "theoretical_loss": 3.4980971835961094, + "tokens_seen": 1748107264 + }, + { + "epoch": 0.53, + "learning_rate": 0.00038011634904714145, + "loss": 1.3793, + "theoretical_loss": 3.49808604924004, + "tokens_seen": 1748172800 + }, + { + "epoch": 0.53, + "learning_rate": 0.00038010030090270816, + "loss": 1.2505, + "theoretical_loss": 3.49807491541824, + "tokens_seen": 1748238336 + }, + { + "epoch": 0.53, + "learning_rate": 0.00038008425275827486, + "loss": 1.3432, + "theoretical_loss": 3.4980637821306635, + "tokens_seen": 1748303872 + }, + { + "epoch": 0.53, + "learning_rate": 0.00038008425275827486, + "loss": 1.3104, + "theoretical_loss": 3.4980526493772652, + "tokens_seen": 1748369408 + }, + { + "epoch": 0.53, + "learning_rate": 0.00038006820461384156, + "loss": 1.1837, + "theoretical_loss": 3.498041517157999, + "tokens_seen": 1748434944 + }, + { + "epoch": 0.53, + "learning_rate": 0.00038005215646940826, + "loss": 1.321, + "theoretical_loss": 3.49803038547282, + "tokens_seen": 1748500480 + }, + { + "epoch": 0.53, + "learning_rate": 0.00038003610832497496, + "loss": 1.2578, + "theoretical_loss": 3.4980192543216813, + "tokens_seen": 1748566016 + }, + { + "epoch": 0.53, + "learning_rate": 0.00038002006018054166, + "loss": 1.3268, + "theoretical_loss": 3.4980081237045386, + "tokens_seen": 1748631552 + }, + { + "epoch": 0.53, + "learning_rate": 0.00038000401203610836, + "loss": 1.3688, + "theoretical_loss": 3.4979969936213458, + "tokens_seen": 1748697088 + }, + { + "epoch": 0.53, + "learning_rate": 0.00037998796389167506, + "loss": 1.4294, + "theoretical_loss": 3.497985864072057, + "tokens_seen": 1748762624 + }, + { + "epoch": 0.53, + "learning_rate": 0.00037997191574724176, + "loss": 1.326, + "theoretical_loss": 3.4979747350566264, + "tokens_seen": 1748828160 + }, + { + "epoch": 0.53, + "learning_rate": 0.00037995586760280846, + "loss": 1.4323, + "theoretical_loss": 3.4979636065750093, + "tokens_seen": 1748893696 + }, + { + "epoch": 0.53, + "learning_rate": 0.00037993981945837516, + "loss": 1.4318, + "theoretical_loss": 3.497952478627159, + "tokens_seen": 1748959232 + }, + { + "epoch": 0.53, + "learning_rate": 0.00037992377131394186, + "loss": 1.4382, + "theoretical_loss": 3.4979413512130306, + "tokens_seen": 1749024768 + }, + { + "epoch": 0.53, + "learning_rate": 0.00037990772316950856, + "loss": 1.3608, + "theoretical_loss": 3.4979302243325785, + "tokens_seen": 1749090304 + }, + { + "epoch": 0.53, + "learning_rate": 0.00037989167502507526, + "loss": 1.3511, + "theoretical_loss": 3.497919097985757, + "tokens_seen": 1749155840 + }, + { + "epoch": 0.53, + "learning_rate": 0.00037987562688064196, + "loss": 1.2838, + "theoretical_loss": 3.49790797217252, + "tokens_seen": 1749221376 + }, + { + "epoch": 0.53, + "learning_rate": 0.00037985957873620867, + "loss": 1.3217, + "theoretical_loss": 3.4978968468928224, + "tokens_seen": 1749286912 + }, + { + "epoch": 0.53, + "learning_rate": 0.00037984353059177537, + "loss": 1.2916, + "theoretical_loss": 3.497885722146619, + "tokens_seen": 1749352448 + }, + { + "epoch": 0.53, + "learning_rate": 0.00037982748244734207, + "loss": 1.307, + "theoretical_loss": 3.4978745979338637, + "tokens_seen": 1749417984 + }, + { + "epoch": 0.53, + "learning_rate": 0.00037981143430290877, + "loss": 1.378, + "theoretical_loss": 3.497863474254511, + "tokens_seen": 1749483520 + }, + { + "epoch": 0.53, + "learning_rate": 0.00037979538615847547, + "loss": 1.4046, + "theoretical_loss": 3.4978523511085156, + "tokens_seen": 1749549056 + }, + { + "epoch": 0.53, + "learning_rate": 0.00037977933801404217, + "loss": 1.3487, + "theoretical_loss": 3.4978412284958313, + "tokens_seen": 1749614592 + }, + { + "epoch": 0.53, + "learning_rate": 0.00037976328986960887, + "loss": 1.4421, + "theoretical_loss": 3.497830106416414, + "tokens_seen": 1749680128 + }, + { + "epoch": 0.53, + "learning_rate": 0.00037974724172517557, + "loss": 1.277, + "theoretical_loss": 3.497818984870216, + "tokens_seen": 1749745664 + }, + { + "epoch": 0.53, + "objective/train/docs_used": 639149, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.3450937271118164, + "objective/train/theoretical_loss": 3.4978078638571937, + "objective/train/tokens_used": 1770271200, + "theoretical_loss": 3.4978078638571937, + "tokens_seen": 1749811200 + }, + { + "epoch": 0.53, + "learning_rate": 0.00037973119358074227, + "loss": 1.3856, + "theoretical_loss": 3.4978078638571937, + "tokens_seen": 1749811200 + }, + { + "epoch": 0.53, + "learning_rate": 0.00037971514543630897, + "loss": 1.3275, + "theoretical_loss": 3.4977967433773003, + "tokens_seen": 1749876736 + }, + { + "epoch": 0.53, + "learning_rate": 0.00037969909729187567, + "loss": 1.3831, + "theoretical_loss": 3.497785623430491, + "tokens_seen": 1749942272 + }, + { + "epoch": 0.53, + "learning_rate": 0.00037968304914744237, + "loss": 1.4351, + "theoretical_loss": 3.4977745040167205, + "tokens_seen": 1750007808 + }, + { + "epoch": 0.53, + "learning_rate": 0.00037966700100300907, + "loss": 1.5051, + "theoretical_loss": 3.4977633851359426, + "tokens_seen": 1750073344 + }, + { + "epoch": 0.53, + "learning_rate": 0.0003796509528585758, + "loss": 1.4253, + "theoretical_loss": 3.497752266788112, + "tokens_seen": 1750138880 + }, + { + "epoch": 0.53, + "learning_rate": 0.0003796349047141425, + "loss": 1.3787, + "theoretical_loss": 3.497741148973183, + "tokens_seen": 1750204416 + }, + { + "epoch": 0.53, + "learning_rate": 0.0003796188565697091, + "loss": 1.4524, + "theoretical_loss": 3.497730031691111, + "tokens_seen": 1750269952 + }, + { + "epoch": 0.53, + "learning_rate": 0.0003796028084252758, + "loss": 1.1602, + "theoretical_loss": 3.4977189149418493, + "tokens_seen": 1750335488 + }, + { + "epoch": 0.53, + "learning_rate": 0.0003795867602808425, + "loss": 1.2701, + "theoretical_loss": 3.497707798725353, + "tokens_seen": 1750401024 + }, + { + "epoch": 0.53, + "learning_rate": 0.0003795707121364092, + "loss": 1.3321, + "theoretical_loss": 3.497696683041577, + "tokens_seen": 1750466560 + }, + { + "epoch": 0.53, + "learning_rate": 0.0003795546639919759, + "loss": 1.2891, + "theoretical_loss": 3.4976855678904752, + "tokens_seen": 1750532096 + }, + { + "epoch": 0.53, + "learning_rate": 0.0003795386158475426, + "loss": 1.3843, + "theoretical_loss": 3.497674453272002, + "tokens_seen": 1750597632 + }, + { + "epoch": 0.53, + "learning_rate": 0.0003795225677031093, + "loss": 1.3647, + "theoretical_loss": 3.497663339186113, + "tokens_seen": 1750663168 + }, + { + "epoch": 0.53, + "learning_rate": 0.000379506519558676, + "loss": 1.3057, + "theoretical_loss": 3.497652225632762, + "tokens_seen": 1750728704 + }, + { + "epoch": 0.53, + "learning_rate": 0.0003794904714142427, + "loss": 1.438, + "theoretical_loss": 3.4976411126119036, + "tokens_seen": 1750794240 + }, + { + "epoch": 0.53, + "learning_rate": 0.0003794744232698094, + "loss": 1.2661, + "theoretical_loss": 3.497630000123492, + "tokens_seen": 1750859776 + }, + { + "epoch": 0.53, + "learning_rate": 0.0003794583751253761, + "loss": 1.2521, + "theoretical_loss": 3.497618888167482, + "tokens_seen": 1750925312 + }, + { + "epoch": 0.53, + "learning_rate": 0.00037944232698094283, + "loss": 1.3655, + "theoretical_loss": 3.4976077767438287, + "tokens_seen": 1750990848 + }, + { + "epoch": 0.53, + "learning_rate": 0.00037942627883650953, + "loss": 1.3047, + "theoretical_loss": 3.497596665852486, + "tokens_seen": 1751056384 + }, + { + "epoch": 0.53, + "learning_rate": 0.00037941023069207623, + "loss": 1.3885, + "theoretical_loss": 3.497585555493409, + "tokens_seen": 1751121920 + }, + { + "epoch": 0.53, + "learning_rate": 0.00037939418254764293, + "loss": 1.3882, + "theoretical_loss": 3.497574445666552, + "tokens_seen": 1751187456 + }, + { + "epoch": 0.53, + "learning_rate": 0.00037937813440320963, + "loss": 1.3053, + "theoretical_loss": 3.4975633363718694, + "tokens_seen": 1751252992 + }, + { + "epoch": 0.53, + "learning_rate": 0.00037936208625877633, + "loss": 1.374, + "theoretical_loss": 3.497552227609316, + "tokens_seen": 1751318528 + }, + { + "epoch": 0.53, + "learning_rate": 0.00037934603811434303, + "loss": 1.4438, + "theoretical_loss": 3.4975411193788464, + "tokens_seen": 1751384064 + }, + { + "epoch": 0.53, + "learning_rate": 0.00037932998996990973, + "loss": 1.3686, + "theoretical_loss": 3.4975300116804156, + "tokens_seen": 1751449600 + }, + { + "epoch": 0.53, + "learning_rate": 0.00037931394182547643, + "loss": 1.3482, + "theoretical_loss": 3.497518904513978, + "tokens_seen": 1751515136 + }, + { + "epoch": 0.53, + "learning_rate": 0.00037929789368104313, + "loss": 1.3737, + "theoretical_loss": 3.4975077978794875, + "tokens_seen": 1751580672 + }, + { + "epoch": 0.53, + "learning_rate": 0.00037928184553660983, + "loss": 1.3468, + "theoretical_loss": 3.497496691776899, + "tokens_seen": 1751646208 + }, + { + "epoch": 0.53, + "learning_rate": 0.00037926579739217653, + "loss": 1.425, + "theoretical_loss": 3.497485586206168, + "tokens_seen": 1751711744 + }, + { + "epoch": 0.53, + "learning_rate": 0.00037924974924774323, + "loss": 1.3495, + "theoretical_loss": 3.497474481167248, + "tokens_seen": 1751777280 + }, + { + "epoch": 0.53, + "learning_rate": 0.00037923370110330994, + "loss": 1.4981, + "theoretical_loss": 3.4974633766600944, + "tokens_seen": 1751842816 + }, + { + "epoch": 0.53, + "learning_rate": 0.00037921765295887664, + "loss": 1.3211, + "theoretical_loss": 3.4974522726846615, + "tokens_seen": 1751908352 + }, + { + "epoch": 0.53, + "learning_rate": 0.00037920160481444334, + "loss": 1.3836, + "theoretical_loss": 3.497441169240904, + "tokens_seen": 1751973888 + }, + { + "epoch": 0.53, + "learning_rate": 0.00037918555667001004, + "loss": 1.451, + "theoretical_loss": 3.497430066328777, + "tokens_seen": 1752039424 + }, + { + "epoch": 0.53, + "learning_rate": 0.00037916950852557674, + "loss": 1.3987, + "theoretical_loss": 3.4974189639482347, + "tokens_seen": 1752104960 + }, + { + "epoch": 0.53, + "learning_rate": 0.00037915346038114344, + "loss": 1.4278, + "theoretical_loss": 3.497407862099232, + "tokens_seen": 1752170496 + }, + { + "epoch": 0.53, + "learning_rate": 0.00037913741223671014, + "loss": 1.4086, + "theoretical_loss": 3.497396760781723, + "tokens_seen": 1752236032 + }, + { + "epoch": 0.53, + "learning_rate": 0.00037912136409227684, + "loss": 1.5012, + "theoretical_loss": 3.497385659995663, + "tokens_seen": 1752301568 + }, + { + "epoch": 0.53, + "learning_rate": 0.00037910531594784354, + "loss": 1.3189, + "theoretical_loss": 3.497374559741006, + "tokens_seen": 1752367104 + }, + { + "epoch": 0.53, + "learning_rate": 0.00037908926780341024, + "loss": 1.3467, + "theoretical_loss": 3.4973634600177075, + "tokens_seen": 1752432640 + }, + { + "epoch": 0.53, + "learning_rate": 0.00037907321965897694, + "loss": 1.3602, + "theoretical_loss": 3.497352360825722, + "tokens_seen": 1752498176 + }, + { + "epoch": 0.53, + "learning_rate": 0.00037905717151454364, + "loss": 1.3544, + "theoretical_loss": 3.497341262165004, + "tokens_seen": 1752563712 + }, + { + "epoch": 0.53, + "learning_rate": 0.00037904112337011034, + "loss": 1.49, + "theoretical_loss": 3.4973301640355086, + "tokens_seen": 1752629248 + }, + { + "epoch": 0.53, + "learning_rate": 0.00037902507522567704, + "loss": 1.3388, + "theoretical_loss": 3.497319066437189, + "tokens_seen": 1752694784 + }, + { + "epoch": 0.53, + "learning_rate": 0.00037900902708124374, + "loss": 1.2817, + "theoretical_loss": 3.497307969370002, + "tokens_seen": 1752760320 + }, + { + "epoch": 0.53, + "learning_rate": 0.00037899297893681045, + "loss": 1.2411, + "theoretical_loss": 3.4972968728339016, + "tokens_seen": 1752825856 + }, + { + "epoch": 0.53, + "learning_rate": 0.00037897693079237715, + "loss": 1.3332, + "theoretical_loss": 3.497285776828842, + "tokens_seen": 1752891392 + }, + { + "epoch": 0.53, + "learning_rate": 0.00037896088264794385, + "loss": 1.3121, + "theoretical_loss": 3.4972746813547784, + "tokens_seen": 1752956928 + }, + { + "epoch": 0.53, + "learning_rate": 0.00037894483450351055, + "loss": 1.3191, + "theoretical_loss": 3.4972635864116652, + "tokens_seen": 1753022464 + }, + { + "epoch": 0.53, + "objective/train/docs_used": 640338, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.3276760578155518, + "objective/train/theoretical_loss": 3.4972524919994576, + "objective/train/tokens_used": 1773548000, + "theoretical_loss": 3.4972524919994576, + "tokens_seen": 1753088000 + }, + { + "epoch": 0.53, + "learning_rate": 0.00037892878635907725, + "loss": 1.3332, + "theoretical_loss": 3.4972524919994576, + "tokens_seen": 1753088000 + }, + { + "epoch": 0.53, + "learning_rate": 0.00037891273821464395, + "loss": 1.3347, + "theoretical_loss": 3.49724139811811, + "tokens_seen": 1753153536 + }, + { + "epoch": 0.53, + "learning_rate": 0.00037889669007021065, + "loss": 1.3752, + "theoretical_loss": 3.4972303047675775, + "tokens_seen": 1753219072 + }, + { + "epoch": 0.53, + "learning_rate": 0.00037888064192577735, + "loss": 1.3893, + "theoretical_loss": 3.4972192119478143, + "tokens_seen": 1753284608 + }, + { + "epoch": 0.53, + "learning_rate": 0.00037886459378134405, + "loss": 1.3918, + "theoretical_loss": 3.497208119658775, + "tokens_seen": 1753350144 + }, + { + "epoch": 0.53, + "learning_rate": 0.00037884854563691075, + "loss": 1.3362, + "theoretical_loss": 3.4971970279004156, + "tokens_seen": 1753415680 + }, + { + "epoch": 0.53, + "learning_rate": 0.00037883249749247745, + "loss": 1.4525, + "theoretical_loss": 3.4971859366726905, + "tokens_seen": 1753481216 + }, + { + "epoch": 0.53, + "learning_rate": 0.00037881644934804415, + "loss": 1.389, + "theoretical_loss": 3.497174845975553, + "tokens_seen": 1753546752 + }, + { + "epoch": 0.53, + "learning_rate": 0.00037880040120361085, + "loss": 1.3537, + "theoretical_loss": 3.49716375580896, + "tokens_seen": 1753612288 + }, + { + "epoch": 0.53, + "learning_rate": 0.00037878435305917755, + "loss": 1.423, + "theoretical_loss": 3.4971526661728647, + "tokens_seen": 1753677824 + }, + { + "epoch": 0.53, + "learning_rate": 0.00037876830491474425, + "loss": 1.387, + "theoretical_loss": 3.4971415770672225, + "tokens_seen": 1753743360 + }, + { + "epoch": 0.53, + "learning_rate": 0.00037875225677031095, + "loss": 1.4214, + "theoretical_loss": 3.4971304884919885, + "tokens_seen": 1753808896 + }, + { + "epoch": 0.53, + "learning_rate": 0.00037873620862587766, + "loss": 1.3204, + "theoretical_loss": 3.4971194004471173, + "tokens_seen": 1753874432 + }, + { + "epoch": 0.53, + "learning_rate": 0.00037872016048144436, + "loss": 1.3073, + "theoretical_loss": 3.4971083129325633, + "tokens_seen": 1753939968 + }, + { + "epoch": 0.53, + "learning_rate": 0.00037870411233701106, + "loss": 1.3515, + "theoretical_loss": 3.4970972259482824, + "tokens_seen": 1754005504 + }, + { + "epoch": 0.53, + "learning_rate": 0.00037868806419257776, + "loss": 1.3837, + "theoretical_loss": 3.497086139494228, + "tokens_seen": 1754071040 + }, + { + "epoch": 0.53, + "learning_rate": 0.00037867201604814446, + "loss": 1.4823, + "theoretical_loss": 3.497075053570356, + "tokens_seen": 1754136576 + }, + { + "epoch": 0.53, + "learning_rate": 0.00037865596790371116, + "loss": 1.5203, + "theoretical_loss": 3.4970639681766205, + "tokens_seen": 1754202112 + }, + { + "epoch": 0.53, + "learning_rate": 0.00037863991975927786, + "loss": 1.3115, + "theoretical_loss": 3.4970528833129766, + "tokens_seen": 1754267648 + }, + { + "epoch": 0.53, + "learning_rate": 0.00037862387161484456, + "loss": 1.431, + "theoretical_loss": 3.49704179897938, + "tokens_seen": 1754333184 + }, + { + "epoch": 0.53, + "learning_rate": 0.00037860782347041126, + "loss": 1.4305, + "theoretical_loss": 3.497030715175784, + "tokens_seen": 1754398720 + }, + { + "epoch": 0.53, + "learning_rate": 0.00037859177532597796, + "loss": 1.4065, + "theoretical_loss": 3.4970196319021447, + "tokens_seen": 1754464256 + }, + { + "epoch": 0.53, + "learning_rate": 0.00037857572718154466, + "loss": 1.3752, + "theoretical_loss": 3.4970085491584166, + "tokens_seen": 1754529792 + }, + { + "epoch": 0.53, + "learning_rate": 0.00037855967903711136, + "loss": 1.4746, + "theoretical_loss": 3.496997466944554, + "tokens_seen": 1754595328 + }, + { + "epoch": 0.53, + "learning_rate": 0.00037854363089267806, + "loss": 1.4328, + "theoretical_loss": 3.496986385260513, + "tokens_seen": 1754660864 + }, + { + "epoch": 0.53, + "learning_rate": 0.00037852758274824476, + "loss": 1.4172, + "theoretical_loss": 3.496975304106247, + "tokens_seen": 1754726400 + }, + { + "epoch": 0.53, + "learning_rate": 0.00037851153460381146, + "loss": 1.4078, + "theoretical_loss": 3.4969642234817124, + "tokens_seen": 1754791936 + }, + { + "epoch": 0.53, + "learning_rate": 0.00037849548645937816, + "loss": 1.393, + "theoretical_loss": 3.496953143386863, + "tokens_seen": 1754857472 + }, + { + "epoch": 0.53, + "learning_rate": 0.00037847943831494487, + "loss": 1.3362, + "theoretical_loss": 3.4969420638216535, + "tokens_seen": 1754923008 + }, + { + "epoch": 0.53, + "learning_rate": 0.00037846339017051157, + "loss": 1.3569, + "theoretical_loss": 3.49693098478604, + "tokens_seen": 1754988544 + }, + { + "epoch": 0.53, + "learning_rate": 0.00037844734202607827, + "loss": 1.3819, + "theoretical_loss": 3.496919906279977, + "tokens_seen": 1755054080 + }, + { + "epoch": 0.53, + "learning_rate": 0.00037843129388164497, + "loss": 1.3218, + "theoretical_loss": 3.496908828303419, + "tokens_seen": 1755119616 + }, + { + "epoch": 0.53, + "learning_rate": 0.00037841524573721167, + "loss": 1.4494, + "theoretical_loss": 3.496897750856321, + "tokens_seen": 1755185152 + }, + { + "epoch": 0.53, + "learning_rate": 0.00037839919759277837, + "loss": 1.3817, + "theoretical_loss": 3.496886673938638, + "tokens_seen": 1755250688 + }, + { + "epoch": 0.53, + "learning_rate": 0.00037838314944834507, + "loss": 1.3594, + "theoretical_loss": 3.496875597550325, + "tokens_seen": 1755316224 + }, + { + "epoch": 0.53, + "learning_rate": 0.00037836710130391177, + "loss": 1.3727, + "theoretical_loss": 3.4968645216913368, + "tokens_seen": 1755381760 + }, + { + "epoch": 0.53, + "learning_rate": 0.00037835105315947847, + "loss": 1.389, + "theoretical_loss": 3.4968534463616283, + "tokens_seen": 1755447296 + }, + { + "epoch": 0.53, + "learning_rate": 0.00037833500501504517, + "loss": 1.362, + "theoretical_loss": 3.496842371561155, + "tokens_seen": 1755512832 + }, + { + "epoch": 0.53, + "learning_rate": 0.00037831895687061187, + "loss": 1.32, + "theoretical_loss": 3.496831297289871, + "tokens_seen": 1755578368 + }, + { + "epoch": 0.53, + "learning_rate": 0.00037830290872617857, + "loss": 1.3908, + "theoretical_loss": 3.4968202235477324, + "tokens_seen": 1755643904 + }, + { + "epoch": 0.53, + "learning_rate": 0.0003782868605817453, + "loss": 1.3595, + "theoretical_loss": 3.496809150334693, + "tokens_seen": 1755709440 + }, + { + "epoch": 0.53, + "learning_rate": 0.000378270812437312, + "loss": 1.3616, + "theoretical_loss": 3.4967980776507086, + "tokens_seen": 1755774976 + }, + { + "epoch": 0.53, + "learning_rate": 0.0003782547642928787, + "loss": 1.2767, + "theoretical_loss": 3.4967870054957335, + "tokens_seen": 1755840512 + }, + { + "epoch": 0.53, + "learning_rate": 0.0003782387161484454, + "loss": 1.3692, + "theoretical_loss": 3.4967759338697233, + "tokens_seen": 1755906048 + }, + { + "epoch": 0.53, + "learning_rate": 0.0003782226680040121, + "loss": 1.3293, + "theoretical_loss": 3.496764862772632, + "tokens_seen": 1755971584 + }, + { + "epoch": 0.53, + "learning_rate": 0.0003782066198595787, + "loss": 1.3891, + "theoretical_loss": 3.496753792204416, + "tokens_seen": 1756037120 + }, + { + "epoch": 0.53, + "learning_rate": 0.0003781905717151454, + "loss": 1.4163, + "theoretical_loss": 3.4967427221650293, + "tokens_seen": 1756102656 + }, + { + "epoch": 0.53, + "learning_rate": 0.0003781745235707121, + "loss": 1.2759, + "theoretical_loss": 3.4967316526544274, + "tokens_seen": 1756168192 + }, + { + "epoch": 0.53, + "learning_rate": 0.0003781584754262788, + "loss": 1.3547, + "theoretical_loss": 3.496720583672565, + "tokens_seen": 1756233728 + }, + { + "epoch": 0.53, + "learning_rate": 0.0003781424272818455, + "loss": 1.3976, + "theoretical_loss": 3.496709515219397, + "tokens_seen": 1756299264 + }, + { + "epoch": 0.53, + "objective/train/docs_used": 641471, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.3039307594299316, + "objective/train/theoretical_loss": 3.4966984472948788, + "objective/train/tokens_used": 1776824800, + "theoretical_loss": 3.4966984472948788, + "tokens_seen": 1756364800 + }, + { + "epoch": 0.53, + "learning_rate": 0.0003781263791374122, + "loss": 1.2726, + "theoretical_loss": 3.4966984472948788, + "tokens_seen": 1756364800 + }, + { + "epoch": 0.53, + "learning_rate": 0.0003781103309929789, + "loss": 1.3752, + "theoretical_loss": 3.4966873798989653, + "tokens_seen": 1756430336 + }, + { + "epoch": 0.53, + "learning_rate": 0.0003780942828485456, + "loss": 1.5023, + "theoretical_loss": 3.4966763130316116, + "tokens_seen": 1756495872 + }, + { + "epoch": 0.53, + "learning_rate": 0.0003780782347041123, + "loss": 1.2679, + "theoretical_loss": 3.4966652466927726, + "tokens_seen": 1756561408 + }, + { + "epoch": 0.53, + "learning_rate": 0.00037806218655967903, + "loss": 1.3967, + "theoretical_loss": 3.496654180882403, + "tokens_seen": 1756626944 + }, + { + "epoch": 0.53, + "learning_rate": 0.00037804613841524573, + "loss": 1.3524, + "theoretical_loss": 3.4966431156004587, + "tokens_seen": 1756692480 + }, + { + "epoch": 0.53, + "learning_rate": 0.00037803009027081243, + "loss": 1.3032, + "theoretical_loss": 3.496632050846894, + "tokens_seen": 1756758016 + }, + { + "epoch": 0.53, + "learning_rate": 0.00037801404212637913, + "loss": 1.3763, + "theoretical_loss": 3.4966209866216644, + "tokens_seen": 1756823552 + }, + { + "epoch": 0.53, + "learning_rate": 0.00037799799398194583, + "loss": 1.395, + "theoretical_loss": 3.496609922924725, + "tokens_seen": 1756889088 + }, + { + "epoch": 0.53, + "learning_rate": 0.00037798194583751253, + "loss": 1.3946, + "theoretical_loss": 3.4965988597560305, + "tokens_seen": 1756954624 + }, + { + "epoch": 0.53, + "learning_rate": 0.00037796589769307923, + "loss": 1.4846, + "theoretical_loss": 3.496587797115536, + "tokens_seen": 1757020160 + }, + { + "epoch": 0.53, + "learning_rate": 0.00037794984954864593, + "loss": 1.3559, + "theoretical_loss": 3.496576735003197, + "tokens_seen": 1757085696 + }, + { + "epoch": 0.53, + "learning_rate": 0.00037793380140421263, + "loss": 1.313, + "theoretical_loss": 3.496565673418968, + "tokens_seen": 1757151232 + }, + { + "epoch": 0.53, + "learning_rate": 0.00037791775325977933, + "loss": 1.3763, + "theoretical_loss": 3.4965546123628046, + "tokens_seen": 1757216768 + }, + { + "epoch": 0.53, + "learning_rate": 0.00037790170511534603, + "loss": 1.3944, + "theoretical_loss": 3.4965435518346615, + "tokens_seen": 1757282304 + }, + { + "epoch": 0.53, + "learning_rate": 0.00037788565697091273, + "loss": 1.493, + "theoretical_loss": 3.496532491834494, + "tokens_seen": 1757347840 + }, + { + "epoch": 0.53, + "learning_rate": 0.00037786960882647944, + "loss": 1.3073, + "theoretical_loss": 3.4965214323622575, + "tokens_seen": 1757413376 + }, + { + "epoch": 0.53, + "learning_rate": 0.00037785356068204614, + "loss": 1.4504, + "theoretical_loss": 3.496510373417907, + "tokens_seen": 1757478912 + }, + { + "epoch": 0.53, + "learning_rate": 0.00037783751253761284, + "loss": 1.3442, + "theoretical_loss": 3.4964993150013974, + "tokens_seen": 1757544448 + }, + { + "epoch": 0.53, + "learning_rate": 0.00037782146439317954, + "loss": 1.4012, + "theoretical_loss": 3.4964882571126834, + "tokens_seen": 1757609984 + }, + { + "epoch": 0.53, + "learning_rate": 0.00037780541624874624, + "loss": 1.3808, + "theoretical_loss": 3.496477199751721, + "tokens_seen": 1757675520 + }, + { + "epoch": 0.53, + "learning_rate": 0.00037778936810431294, + "loss": 1.4413, + "theoretical_loss": 3.496466142918465, + "tokens_seen": 1757741056 + }, + { + "epoch": 0.53, + "learning_rate": 0.0003777733199598797, + "loss": 1.3728, + "theoretical_loss": 3.4964550866128703, + "tokens_seen": 1757806592 + }, + { + "epoch": 0.53, + "learning_rate": 0.0003777572718154464, + "loss": 1.3815, + "theoretical_loss": 3.4964440308348923, + "tokens_seen": 1757872128 + }, + { + "epoch": 0.53, + "learning_rate": 0.0003777412236710131, + "loss": 1.558, + "theoretical_loss": 3.4964329755844856, + "tokens_seen": 1757937664 + }, + { + "epoch": 0.53, + "learning_rate": 0.0003777251755265798, + "loss": 1.4597, + "theoretical_loss": 3.4964219208616067, + "tokens_seen": 1758003200 + }, + { + "epoch": 0.53, + "learning_rate": 0.0003777091273821465, + "loss": 1.4239, + "theoretical_loss": 3.49641086666621, + "tokens_seen": 1758068736 + }, + { + "epoch": 0.53, + "learning_rate": 0.0003776930792377132, + "loss": 1.3612, + "theoretical_loss": 3.4963998129982494, + "tokens_seen": 1758134272 + }, + { + "epoch": 0.53, + "learning_rate": 0.0003776770310932799, + "loss": 1.4283, + "theoretical_loss": 3.496388759857682, + "tokens_seen": 1758199808 + }, + { + "epoch": 0.53, + "learning_rate": 0.0003776609829488466, + "loss": 1.4452, + "theoretical_loss": 3.4963777072444624, + "tokens_seen": 1758265344 + }, + { + "epoch": 0.53, + "learning_rate": 0.0003776449348044133, + "loss": 1.3797, + "theoretical_loss": 3.4963666551585453, + "tokens_seen": 1758330880 + }, + { + "epoch": 0.53, + "learning_rate": 0.00037762888665998, + "loss": 1.3681, + "theoretical_loss": 3.496355603599887, + "tokens_seen": 1758396416 + }, + { + "epoch": 0.53, + "learning_rate": 0.0003776128385155467, + "loss": 1.3544, + "theoretical_loss": 3.496344552568441, + "tokens_seen": 1758461952 + }, + { + "epoch": 0.53, + "learning_rate": 0.0003775967903711134, + "loss": 1.4157, + "theoretical_loss": 3.4963335020641635, + "tokens_seen": 1758527488 + }, + { + "epoch": 0.53, + "learning_rate": 0.00037758074222668005, + "loss": 1.3061, + "theoretical_loss": 3.4963224520870098, + "tokens_seen": 1758593024 + }, + { + "epoch": 0.53, + "learning_rate": 0.00037756469408224675, + "loss": 1.3608, + "theoretical_loss": 3.496311402636935, + "tokens_seen": 1758658560 + }, + { + "epoch": 0.53, + "learning_rate": 0.00037754864593781345, + "loss": 1.3021, + "theoretical_loss": 3.4963003537138944, + "tokens_seen": 1758724096 + }, + { + "epoch": 0.53, + "learning_rate": 0.00037753259779338015, + "loss": 1.39, + "theoretical_loss": 3.4962893053178425, + "tokens_seen": 1758789632 + }, + { + "epoch": 0.53, + "learning_rate": 0.00037751654964894685, + "loss": 1.3048, + "theoretical_loss": 3.4962782574487354, + "tokens_seen": 1758855168 + }, + { + "epoch": 0.53, + "learning_rate": 0.00037750050150451355, + "loss": 1.2452, + "theoretical_loss": 3.496267210106528, + "tokens_seen": 1758920704 + }, + { + "epoch": 0.53, + "learning_rate": 0.00037748445336008025, + "loss": 1.4102, + "theoretical_loss": 3.496256163291176, + "tokens_seen": 1758986240 + }, + { + "epoch": 0.53, + "learning_rate": 0.00037746840521564695, + "loss": 1.3904, + "theoretical_loss": 3.496245117002634, + "tokens_seen": 1759051776 + }, + { + "epoch": 0.53, + "learning_rate": 0.00037745235707121365, + "loss": 1.3768, + "theoretical_loss": 3.496234071240857, + "tokens_seen": 1759117312 + }, + { + "epoch": 0.53, + "learning_rate": 0.00037743630892678035, + "loss": 1.4644, + "theoretical_loss": 3.4962230260058016, + "tokens_seen": 1759182848 + }, + { + "epoch": 0.53, + "learning_rate": 0.00037742026078234705, + "loss": 1.3025, + "theoretical_loss": 3.4962119812974217, + "tokens_seen": 1759248384 + }, + { + "epoch": 0.53, + "learning_rate": 0.00037740421263791375, + "loss": 1.4251, + "theoretical_loss": 3.496200937115673, + "tokens_seen": 1759313920 + }, + { + "epoch": 0.53, + "learning_rate": 0.00037738816449348045, + "loss": 1.3936, + "theoretical_loss": 3.496189893460511, + "tokens_seen": 1759379456 + }, + { + "epoch": 0.53, + "learning_rate": 0.00037737211634904715, + "loss": 1.4859, + "theoretical_loss": 3.4961788503318907, + "tokens_seen": 1759444992 + }, + { + "epoch": 0.53, + "learning_rate": 0.00037735606820461386, + "loss": 1.4186, + "theoretical_loss": 3.4961678077297673, + "tokens_seen": 1759510528 + }, + { + "epoch": 0.53, + "learning_rate": 0.00037734002006018056, + "loss": 1.4752, + "theoretical_loss": 3.4961567656540966, + "tokens_seen": 1759576064 + }, + { + "epoch": 0.53, + "objective/train/docs_used": 642679, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.3062949180603027, + "objective/train/theoretical_loss": 3.4961457241048337, + "objective/train/tokens_used": 1780101600, + "theoretical_loss": 3.4961457241048337, + "tokens_seen": 1759641600 + }, + { + "epoch": 0.53, + "learning_rate": 0.00037732397191574726, + "loss": 1.3828, + "theoretical_loss": 3.4961457241048337, + "tokens_seen": 1759641600 + }, + { + "epoch": 0.53, + "learning_rate": 0.00037730792377131396, + "loss": 1.3502, + "theoretical_loss": 3.4961346830819338, + "tokens_seen": 1759707136 + }, + { + "epoch": 0.53, + "learning_rate": 0.00037729187562688066, + "loss": 1.3081, + "theoretical_loss": 3.496123642585352, + "tokens_seen": 1759772672 + }, + { + "epoch": 0.53, + "learning_rate": 0.00037727582748244736, + "loss": 1.3739, + "theoretical_loss": 3.4961126026150433, + "tokens_seen": 1759838208 + }, + { + "epoch": 0.53, + "learning_rate": 0.00037725977933801406, + "loss": 1.3869, + "theoretical_loss": 3.4961015631709644, + "tokens_seen": 1759903744 + }, + { + "epoch": 0.53, + "learning_rate": 0.00037724373119358076, + "loss": 1.2866, + "theoretical_loss": 3.496090524253069, + "tokens_seen": 1759969280 + }, + { + "epoch": 0.53, + "learning_rate": 0.00037722768304914746, + "loss": 1.4326, + "theoretical_loss": 3.496079485861314, + "tokens_seen": 1760034816 + }, + { + "epoch": 0.53, + "learning_rate": 0.00037721163490471416, + "loss": 1.4314, + "theoretical_loss": 3.496068447995653, + "tokens_seen": 1760100352 + }, + { + "epoch": 0.53, + "learning_rate": 0.00037719558676028086, + "loss": 1.4415, + "theoretical_loss": 3.496057410656043, + "tokens_seen": 1760165888 + }, + { + "epoch": 0.53, + "learning_rate": 0.00037717953861584756, + "loss": 1.4315, + "theoretical_loss": 3.4960463738424385, + "tokens_seen": 1760231424 + }, + { + "epoch": 0.53, + "learning_rate": 0.00037716349047141426, + "loss": 1.364, + "theoretical_loss": 3.496035337554795, + "tokens_seen": 1760296960 + }, + { + "epoch": 0.53, + "learning_rate": 0.00037714744232698096, + "loss": 1.3732, + "theoretical_loss": 3.4960243017930672, + "tokens_seen": 1760362496 + }, + { + "epoch": 0.53, + "learning_rate": 0.00037713139418254766, + "loss": 1.395, + "theoretical_loss": 3.4960132665572115, + "tokens_seen": 1760428032 + }, + { + "epoch": 0.53, + "learning_rate": 0.00037711534603811437, + "loss": 1.4089, + "theoretical_loss": 3.496002231847182, + "tokens_seen": 1760493568 + }, + { + "epoch": 0.53, + "learning_rate": 0.00037709929789368107, + "loss": 1.3589, + "theoretical_loss": 3.495991197662936, + "tokens_seen": 1760559104 + }, + { + "epoch": 0.53, + "learning_rate": 0.00037708324974924777, + "loss": 1.391, + "theoretical_loss": 3.4959801640044272, + "tokens_seen": 1760624640 + }, + { + "epoch": 0.53, + "learning_rate": 0.00037706720160481447, + "loss": 1.2881, + "theoretical_loss": 3.495969130871612, + "tokens_seen": 1760690176 + }, + { + "epoch": 0.53, + "learning_rate": 0.00037705115346038117, + "loss": 1.4087, + "theoretical_loss": 3.4959580982644445, + "tokens_seen": 1760755712 + }, + { + "epoch": 0.53, + "learning_rate": 0.00037703510531594787, + "loss": 1.4169, + "theoretical_loss": 3.4959470661828815, + "tokens_seen": 1760821248 + }, + { + "epoch": 0.53, + "learning_rate": 0.00037701905717151457, + "loss": 1.3502, + "theoretical_loss": 3.495936034626878, + "tokens_seen": 1760886784 + }, + { + "epoch": 0.53, + "learning_rate": 0.00037700300902708127, + "loss": 1.3665, + "theoretical_loss": 3.4959250035963887, + "tokens_seen": 1760952320 + }, + { + "epoch": 0.53, + "learning_rate": 0.00037698696088264797, + "loss": 1.3916, + "theoretical_loss": 3.49591397309137, + "tokens_seen": 1761017856 + }, + { + "epoch": 0.53, + "learning_rate": 0.00037697091273821467, + "loss": 1.2761, + "theoretical_loss": 3.495902943111777, + "tokens_seen": 1761083392 + }, + { + "epoch": 0.53, + "learning_rate": 0.00037695486459378137, + "loss": 1.2954, + "theoretical_loss": 3.4958919136575646, + "tokens_seen": 1761148928 + }, + { + "epoch": 0.53, + "learning_rate": 0.00037693881644934807, + "loss": 1.3654, + "theoretical_loss": 3.4958808847286886, + "tokens_seen": 1761214464 + }, + { + "epoch": 0.53, + "learning_rate": 0.00037692276830491477, + "loss": 1.3515, + "theoretical_loss": 3.495869856325105, + "tokens_seen": 1761280000 + }, + { + "epoch": 0.53, + "learning_rate": 0.0003769067201604815, + "loss": 1.4432, + "theoretical_loss": 3.4958588284467678, + "tokens_seen": 1761345536 + }, + { + "epoch": 0.53, + "learning_rate": 0.0003768906720160482, + "loss": 1.4187, + "theoretical_loss": 3.4958478010936336, + "tokens_seen": 1761411072 + }, + { + "epoch": 0.53, + "learning_rate": 0.0003768746238716149, + "loss": 1.3506, + "theoretical_loss": 3.4958367742656575, + "tokens_seen": 1761476608 + }, + { + "epoch": 0.53, + "learning_rate": 0.0003768585757271816, + "loss": 1.3634, + "theoretical_loss": 3.4958257479627957, + "tokens_seen": 1761542144 + }, + { + "epoch": 0.53, + "learning_rate": 0.0003768425275827483, + "loss": 1.2591, + "theoretical_loss": 3.495814722185002, + "tokens_seen": 1761607680 + }, + { + "epoch": 0.53, + "learning_rate": 0.000376826479438315, + "loss": 1.4199, + "theoretical_loss": 3.4958036969322337, + "tokens_seen": 1761673216 + }, + { + "epoch": 0.53, + "learning_rate": 0.0003768104312938816, + "loss": 1.3497, + "theoretical_loss": 3.495792672204445, + "tokens_seen": 1761738752 + }, + { + "epoch": 0.53, + "learning_rate": 0.0003767943831494483, + "loss": 1.3699, + "theoretical_loss": 3.495781648001592, + "tokens_seen": 1761804288 + }, + { + "epoch": 0.53, + "learning_rate": 0.000376778335005015, + "loss": 1.3655, + "theoretical_loss": 3.4957706243236295, + "tokens_seen": 1761869824 + }, + { + "epoch": 0.53, + "learning_rate": 0.0003767622868605817, + "loss": 1.4491, + "theoretical_loss": 3.495759601170514, + "tokens_seen": 1761935360 + }, + { + "epoch": 0.53, + "learning_rate": 0.0003767462387161484, + "loss": 1.456, + "theoretical_loss": 3.4957485785422, + "tokens_seen": 1762000896 + }, + { + "epoch": 0.53, + "learning_rate": 0.0003767301905717151, + "loss": 1.3777, + "theoretical_loss": 3.4957375564386437, + "tokens_seen": 1762066432 + }, + { + "epoch": 0.53, + "learning_rate": 0.0003767141424272818, + "loss": 1.3397, + "theoretical_loss": 3.4957265348598, + "tokens_seen": 1762131968 + }, + { + "epoch": 0.53, + "learning_rate": 0.00037669809428284853, + "loss": 1.4146, + "theoretical_loss": 3.495715513805625, + "tokens_seen": 1762197504 + }, + { + "epoch": 0.53, + "learning_rate": 0.00037668204613841523, + "loss": 1.3632, + "theoretical_loss": 3.495704493276074, + "tokens_seen": 1762263040 + }, + { + "epoch": 0.53, + "learning_rate": 0.00037666599799398193, + "loss": 1.4422, + "theoretical_loss": 3.4956934732711025, + "tokens_seen": 1762328576 + }, + { + "epoch": 0.53, + "learning_rate": 0.00037664994984954863, + "loss": 1.3486, + "theoretical_loss": 3.4956824537906654, + "tokens_seen": 1762394112 + }, + { + "epoch": 0.53, + "learning_rate": 0.00037663390170511533, + "loss": 1.4089, + "theoretical_loss": 3.4956714348347195, + "tokens_seen": 1762459648 + }, + { + "epoch": 0.53, + "learning_rate": 0.00037661785356068203, + "loss": 1.3492, + "theoretical_loss": 3.4956604164032195, + "tokens_seen": 1762525184 + }, + { + "epoch": 0.53, + "learning_rate": 0.00037660180541624873, + "loss": 1.2527, + "theoretical_loss": 3.4956493984961208, + "tokens_seen": 1762590720 + }, + { + "epoch": 0.53, + "learning_rate": 0.00037658575727181543, + "loss": 1.3158, + "theoretical_loss": 3.4956383811133795, + "tokens_seen": 1762656256 + }, + { + "epoch": 0.53, + "learning_rate": 0.0003765697091273822, + "loss": 1.4155, + "theoretical_loss": 3.4956273642549514, + "tokens_seen": 1762721792 + }, + { + "epoch": 0.53, + "learning_rate": 0.0003765536609829489, + "loss": 1.2956, + "theoretical_loss": 3.495616347920791, + "tokens_seen": 1762787328 + }, + { + "epoch": 0.53, + "learning_rate": 0.0003765376128385156, + "loss": 1.3712, + "theoretical_loss": 3.4956053321108547, + "tokens_seen": 1762852864 + }, + { + "epoch": 0.53, + "objective/train/docs_used": 643830, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.179969310760498, + "objective/train/theoretical_loss": 3.4955943168250974, + "objective/train/tokens_used": 1783378400, + "theoretical_loss": 3.4955943168250974, + "tokens_seen": 1762918400 + }, + { + "epoch": 0.53, + "learning_rate": 0.0003765215646940823, + "loss": 1.2731, + "theoretical_loss": 3.4955943168250974, + "tokens_seen": 1762918400 + }, + { + "epoch": 0.53, + "learning_rate": 0.000376505516549649, + "loss": 1.477, + "theoretical_loss": 3.4955833020634755, + "tokens_seen": 1762983936 + }, + { + "epoch": 0.53, + "learning_rate": 0.0003764894684052157, + "loss": 1.5292, + "theoretical_loss": 3.4955722878259436, + "tokens_seen": 1763049472 + }, + { + "epoch": 0.53, + "learning_rate": 0.0003764734202607824, + "loss": 1.28, + "theoretical_loss": 3.495561274112458, + "tokens_seen": 1763115008 + }, + { + "epoch": 0.53, + "learning_rate": 0.0003764573721163491, + "loss": 1.4435, + "theoretical_loss": 3.4955502609229745, + "tokens_seen": 1763180544 + }, + { + "epoch": 0.53, + "learning_rate": 0.0003764413239719158, + "loss": 1.3032, + "theoretical_loss": 3.4955392482574474, + "tokens_seen": 1763246080 + }, + { + "epoch": 0.53, + "learning_rate": 0.0003764252758274825, + "loss": 1.4094, + "theoretical_loss": 3.4955282361158337, + "tokens_seen": 1763311616 + }, + { + "epoch": 0.53, + "learning_rate": 0.0003764092276830492, + "loss": 1.4343, + "theoretical_loss": 3.495517224498089, + "tokens_seen": 1763377152 + }, + { + "epoch": 0.53, + "learning_rate": 0.0003763931795386159, + "loss": 1.306, + "theoretical_loss": 3.495506213404168, + "tokens_seen": 1763442688 + }, + { + "epoch": 0.53, + "learning_rate": 0.0003763771313941826, + "loss": 1.3938, + "theoretical_loss": 3.495495202834027, + "tokens_seen": 1763508224 + }, + { + "epoch": 0.53, + "learning_rate": 0.0003763610832497493, + "loss": 1.31, + "theoretical_loss": 3.495484192787621, + "tokens_seen": 1763573760 + }, + { + "epoch": 0.53, + "learning_rate": 0.000376345035105316, + "loss": 1.2396, + "theoretical_loss": 3.495473183264906, + "tokens_seen": 1763639296 + }, + { + "epoch": 0.53, + "learning_rate": 0.0003763289869608827, + "loss": 1.3442, + "theoretical_loss": 3.495462174265838, + "tokens_seen": 1763704832 + }, + { + "epoch": 0.53, + "learning_rate": 0.0003763129388164494, + "loss": 1.3378, + "theoretical_loss": 3.4954511657903717, + "tokens_seen": 1763770368 + }, + { + "epoch": 0.53, + "learning_rate": 0.0003762968906720161, + "loss": 1.3331, + "theoretical_loss": 3.4954401578384635, + "tokens_seen": 1763835904 + }, + { + "epoch": 0.53, + "learning_rate": 0.0003762808425275828, + "loss": 1.4229, + "theoretical_loss": 3.495429150410069, + "tokens_seen": 1763901440 + }, + { + "epoch": 0.53, + "learning_rate": 0.0003762647943831495, + "loss": 1.3283, + "theoretical_loss": 3.4954181435051437, + "tokens_seen": 1763966976 + }, + { + "epoch": 0.53, + "learning_rate": 0.0003762487462387162, + "loss": 1.3575, + "theoretical_loss": 3.495407137123643, + "tokens_seen": 1764032512 + }, + { + "epoch": 0.53, + "learning_rate": 0.0003762326980942829, + "loss": 1.3299, + "theoretical_loss": 3.495396131265523, + "tokens_seen": 1764098048 + }, + { + "epoch": 0.53, + "learning_rate": 0.0003762166499498496, + "loss": 1.441, + "theoretical_loss": 3.495385125930739, + "tokens_seen": 1764163584 + }, + { + "epoch": 0.53, + "learning_rate": 0.0003762006018054163, + "loss": 1.377, + "theoretical_loss": 3.495374121119247, + "tokens_seen": 1764229120 + }, + { + "epoch": 0.53, + "learning_rate": 0.00037618455366098295, + "loss": 1.3563, + "theoretical_loss": 3.4953631168310024, + "tokens_seen": 1764294656 + }, + { + "epoch": 0.53, + "learning_rate": 0.00037616850551654965, + "loss": 1.2678, + "theoretical_loss": 3.4953521130659615, + "tokens_seen": 1764360192 + }, + { + "epoch": 0.53, + "learning_rate": 0.00037615245737211635, + "loss": 1.2918, + "theoretical_loss": 3.4953411098240785, + "tokens_seen": 1764425728 + }, + { + "epoch": 0.53, + "learning_rate": 0.00037613640922768305, + "loss": 1.306, + "theoretical_loss": 3.4953301071053113, + "tokens_seen": 1764491264 + }, + { + "epoch": 0.53, + "learning_rate": 0.00037612036108324975, + "loss": 1.2791, + "theoretical_loss": 3.4953191049096137, + "tokens_seen": 1764556800 + }, + { + "epoch": 0.53, + "learning_rate": 0.00037610431293881645, + "loss": 1.4335, + "theoretical_loss": 3.495308103236942, + "tokens_seen": 1764622336 + }, + { + "epoch": 0.53, + "learning_rate": 0.00037608826479438315, + "loss": 1.451, + "theoretical_loss": 3.4952971020872527, + "tokens_seen": 1764687872 + }, + { + "epoch": 0.53, + "learning_rate": 0.00037607221664994985, + "loss": 1.2847, + "theoretical_loss": 3.4952861014605006, + "tokens_seen": 1764753408 + }, + { + "epoch": 0.53, + "learning_rate": 0.00037605616850551655, + "loss": 1.4467, + "theoretical_loss": 3.4952751013566417, + "tokens_seen": 1764818944 + }, + { + "epoch": 0.53, + "learning_rate": 0.00037604012036108325, + "loss": 1.3215, + "theoretical_loss": 3.4952641017756316, + "tokens_seen": 1764884480 + }, + { + "epoch": 0.53, + "learning_rate": 0.00037602407221664995, + "loss": 1.3914, + "theoretical_loss": 3.495253102717426, + "tokens_seen": 1764950016 + }, + { + "epoch": 0.53, + "learning_rate": 0.00037600802407221665, + "loss": 1.3289, + "theoretical_loss": 3.495242104181981, + "tokens_seen": 1765015552 + }, + { + "epoch": 0.53, + "learning_rate": 0.00037599197592778336, + "loss": 1.3993, + "theoretical_loss": 3.495231106169252, + "tokens_seen": 1765081088 + }, + { + "epoch": 0.53, + "learning_rate": 0.00037597592778335006, + "loss": 1.3682, + "theoretical_loss": 3.495220108679195, + "tokens_seen": 1765146624 + }, + { + "epoch": 0.53, + "learning_rate": 0.00037595987963891676, + "loss": 1.448, + "theoretical_loss": 3.4952091117117656, + "tokens_seen": 1765212160 + }, + { + "epoch": 0.53, + "learning_rate": 0.00037594383149448346, + "loss": 1.4139, + "theoretical_loss": 3.49519811526692, + "tokens_seen": 1765277696 + }, + { + "epoch": 0.53, + "learning_rate": 0.00037592778335005016, + "loss": 1.4059, + "theoretical_loss": 3.495187119344613, + "tokens_seen": 1765343232 + }, + { + "epoch": 0.53, + "learning_rate": 0.00037591173520561686, + "loss": 1.3821, + "theoretical_loss": 3.495176123944801, + "tokens_seen": 1765408768 + }, + { + "epoch": 0.53, + "learning_rate": 0.00037589568706118356, + "loss": 1.4128, + "theoretical_loss": 3.4951651290674395, + "tokens_seen": 1765474304 + }, + { + "epoch": 0.54, + "learning_rate": 0.00037587963891675026, + "loss": 1.4287, + "theoretical_loss": 3.495154134712485, + "tokens_seen": 1765539840 + }, + { + "epoch": 0.54, + "learning_rate": 0.00037586359077231696, + "loss": 1.2873, + "theoretical_loss": 3.4951431408798928, + "tokens_seen": 1765605376 + }, + { + "epoch": 0.54, + "learning_rate": 0.00037584754262788366, + "loss": 1.2635, + "theoretical_loss": 3.4951321475696187, + "tokens_seen": 1765670912 + }, + { + "epoch": 0.54, + "learning_rate": 0.00037583149448345036, + "loss": 1.2746, + "theoretical_loss": 3.495121154781618, + "tokens_seen": 1765736448 + }, + { + "epoch": 0.54, + "learning_rate": 0.00037581544633901706, + "loss": 1.3855, + "theoretical_loss": 3.4951101625158474, + "tokens_seen": 1765801984 + }, + { + "epoch": 0.54, + "learning_rate": 0.00037579939819458376, + "loss": 1.3379, + "theoretical_loss": 3.495099170772262, + "tokens_seen": 1765867520 + }, + { + "epoch": 0.54, + "learning_rate": 0.00037578335005015046, + "loss": 1.428, + "theoretical_loss": 3.4950881795508177, + "tokens_seen": 1765933056 + }, + { + "epoch": 0.54, + "learning_rate": 0.00037576730190571716, + "loss": 1.32, + "theoretical_loss": 3.4950771888514716, + "tokens_seen": 1765998592 + }, + { + "epoch": 0.54, + "learning_rate": 0.00037575125376128386, + "loss": 1.4009, + "theoretical_loss": 3.495066198674177, + "tokens_seen": 1766064128 + }, + { + "epoch": 0.54, + "learning_rate": 0.00037573520561685057, + "loss": 1.4094, + "theoretical_loss": 3.495055209018892, + "tokens_seen": 1766129664 + }, + { + "epoch": 0.54, + "objective/train/docs_used": 644977, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.5683029890060425, + "objective/train/theoretical_loss": 3.495044219885572, + "objective/train/tokens_used": 1786655200, + "theoretical_loss": 3.495044219885572, + "tokens_seen": 1766195200 + }, + { + "epoch": 0.54, + "learning_rate": 0.00037571915747241727, + "loss": 1.4377, + "theoretical_loss": 3.495044219885572, + "tokens_seen": 1766195200 + }, + { + "epoch": 0.54, + "learning_rate": 0.00037570310932798397, + "loss": 1.2217, + "theoretical_loss": 3.4950332312741716, + "tokens_seen": 1766260736 + }, + { + "epoch": 0.54, + "learning_rate": 0.00037568706118355067, + "loss": 1.4342, + "theoretical_loss": 3.495022243184648, + "tokens_seen": 1766326272 + }, + { + "epoch": 0.54, + "learning_rate": 0.00037567101303911737, + "loss": 1.3413, + "theoretical_loss": 3.495011255616957, + "tokens_seen": 1766391808 + }, + { + "epoch": 0.54, + "learning_rate": 0.00037565496489468407, + "loss": 1.369, + "theoretical_loss": 3.4950002685710535, + "tokens_seen": 1766457344 + }, + { + "epoch": 0.54, + "learning_rate": 0.00037563891675025077, + "loss": 1.4122, + "theoretical_loss": 3.4949892820468937, + "tokens_seen": 1766522880 + }, + { + "epoch": 0.54, + "learning_rate": 0.00037562286860581747, + "loss": 1.3938, + "theoretical_loss": 3.494978296044434, + "tokens_seen": 1766588416 + }, + { + "epoch": 0.54, + "learning_rate": 0.00037560682046138417, + "loss": 1.423, + "theoretical_loss": 3.49496731056363, + "tokens_seen": 1766653952 + }, + { + "epoch": 0.54, + "learning_rate": 0.00037559077231695087, + "loss": 1.3693, + "theoretical_loss": 3.494956325604437, + "tokens_seen": 1766719488 + }, + { + "epoch": 0.54, + "learning_rate": 0.00037557472417251757, + "loss": 1.4341, + "theoretical_loss": 3.4949453411668117, + "tokens_seen": 1766785024 + }, + { + "epoch": 0.54, + "learning_rate": 0.00037555867602808427, + "loss": 1.3605, + "theoretical_loss": 3.4949343572507097, + "tokens_seen": 1766850560 + }, + { + "epoch": 0.54, + "learning_rate": 0.000375542627883651, + "loss": 1.3207, + "theoretical_loss": 3.4949233738560874, + "tokens_seen": 1766916096 + }, + { + "epoch": 0.54, + "learning_rate": 0.0003755265797392177, + "loss": 1.3865, + "theoretical_loss": 3.4949123909828996, + "tokens_seen": 1766981632 + }, + { + "epoch": 0.54, + "learning_rate": 0.0003755105315947844, + "loss": 1.4956, + "theoretical_loss": 3.4949014086311028, + "tokens_seen": 1767047168 + }, + { + "epoch": 0.54, + "learning_rate": 0.0003754944834503511, + "loss": 1.3425, + "theoretical_loss": 3.4948904268006533, + "tokens_seen": 1767112704 + }, + { + "epoch": 0.54, + "learning_rate": 0.0003754784353059178, + "loss": 1.3924, + "theoretical_loss": 3.494879445491506, + "tokens_seen": 1767178240 + }, + { + "epoch": 0.54, + "learning_rate": 0.0003754623871614845, + "loss": 1.3511, + "theoretical_loss": 3.494868464703618, + "tokens_seen": 1767243776 + }, + { + "epoch": 0.54, + "learning_rate": 0.0003754463390170512, + "loss": 1.398, + "theoretical_loss": 3.4948574844369444, + "tokens_seen": 1767309312 + }, + { + "epoch": 0.54, + "learning_rate": 0.0003754302908726179, + "loss": 1.3868, + "theoretical_loss": 3.4948465046914414, + "tokens_seen": 1767374848 + }, + { + "epoch": 0.54, + "learning_rate": 0.0003754142427281845, + "loss": 1.4301, + "theoretical_loss": 3.494835525467065, + "tokens_seen": 1767440384 + }, + { + "epoch": 0.54, + "learning_rate": 0.0003753981945837512, + "loss": 1.269, + "theoretical_loss": 3.494824546763771, + "tokens_seen": 1767505920 + }, + { + "epoch": 0.54, + "learning_rate": 0.0003753821464393179, + "loss": 1.4545, + "theoretical_loss": 3.494813568581516, + "tokens_seen": 1767571456 + }, + { + "epoch": 0.54, + "learning_rate": 0.0003753660982948847, + "loss": 1.3161, + "theoretical_loss": 3.494802590920255, + "tokens_seen": 1767636992 + }, + { + "epoch": 0.54, + "learning_rate": 0.0003753500501504514, + "loss": 1.3598, + "theoretical_loss": 3.494791613779944, + "tokens_seen": 1767702528 + }, + { + "epoch": 0.54, + "learning_rate": 0.0003753340020060181, + "loss": 1.367, + "theoretical_loss": 3.49478063716054, + "tokens_seen": 1767768064 + }, + { + "epoch": 0.54, + "learning_rate": 0.0003753179538615848, + "loss": 1.1748, + "theoretical_loss": 3.494769661061998, + "tokens_seen": 1767833600 + }, + { + "epoch": 0.54, + "learning_rate": 0.0003753019057171515, + "loss": 1.31, + "theoretical_loss": 3.494758685484274, + "tokens_seen": 1767899136 + }, + { + "epoch": 0.54, + "learning_rate": 0.0003752858575727182, + "loss": 1.3417, + "theoretical_loss": 3.4947477104273243, + "tokens_seen": 1767964672 + }, + { + "epoch": 0.54, + "learning_rate": 0.0003752698094282849, + "loss": 1.3324, + "theoretical_loss": 3.494736735891105, + "tokens_seen": 1768030208 + }, + { + "epoch": 0.54, + "learning_rate": 0.0003752537612838516, + "loss": 1.3671, + "theoretical_loss": 3.494725761875572, + "tokens_seen": 1768095744 + }, + { + "epoch": 0.54, + "learning_rate": 0.0003752377131394183, + "loss": 1.3273, + "theoretical_loss": 3.494714788380681, + "tokens_seen": 1768161280 + }, + { + "epoch": 0.54, + "learning_rate": 0.000375221664994985, + "loss": 1.4975, + "theoretical_loss": 3.494703815406388, + "tokens_seen": 1768226816 + }, + { + "epoch": 0.54, + "learning_rate": 0.0003752056168505517, + "loss": 1.4146, + "theoretical_loss": 3.4946928429526496, + "tokens_seen": 1768292352 + }, + { + "epoch": 0.54, + "learning_rate": 0.0003751895687061184, + "loss": 1.2638, + "theoretical_loss": 3.494681871019422, + "tokens_seen": 1768357888 + }, + { + "epoch": 0.54, + "learning_rate": 0.0003751735205616851, + "loss": 1.3851, + "theoretical_loss": 3.4946708996066596, + "tokens_seen": 1768423424 + }, + { + "epoch": 0.54, + "learning_rate": 0.0003751574724172518, + "loss": 1.4267, + "theoretical_loss": 3.49465992871432, + "tokens_seen": 1768488960 + }, + { + "epoch": 0.54, + "learning_rate": 0.0003751414242728185, + "loss": 1.3656, + "theoretical_loss": 3.494648958342359, + "tokens_seen": 1768554496 + }, + { + "epoch": 0.54, + "learning_rate": 0.0003751253761283852, + "loss": 1.2609, + "theoretical_loss": 3.494637988490732, + "tokens_seen": 1768620032 + }, + { + "epoch": 0.54, + "learning_rate": 0.0003751093279839519, + "loss": 1.3657, + "theoretical_loss": 3.4946270191593958, + "tokens_seen": 1768685568 + }, + { + "epoch": 0.54, + "learning_rate": 0.0003750932798395186, + "loss": 1.3606, + "theoretical_loss": 3.4946160503483057, + "tokens_seen": 1768751104 + }, + { + "epoch": 0.54, + "learning_rate": 0.0003750772316950853, + "loss": 1.385, + "theoretical_loss": 3.494605082057418, + "tokens_seen": 1768816640 + }, + { + "epoch": 0.54, + "learning_rate": 0.000375061183550652, + "loss": 1.4, + "theoretical_loss": 3.4945941142866888, + "tokens_seen": 1768882176 + }, + { + "epoch": 0.54, + "learning_rate": 0.0003750451354062187, + "loss": 1.3992, + "theoretical_loss": 3.494583147036075, + "tokens_seen": 1768947712 + }, + { + "epoch": 0.54, + "learning_rate": 0.0003750290872617854, + "loss": 1.358, + "theoretical_loss": 3.494572180305531, + "tokens_seen": 1769013248 + }, + { + "epoch": 0.54, + "learning_rate": 0.0003750130391173521, + "loss": 1.4586, + "theoretical_loss": 3.494561214095014, + "tokens_seen": 1769078784 + }, + { + "epoch": 0.54, + "learning_rate": 0.0003749969909729188, + "loss": 1.4023, + "theoretical_loss": 3.4945502484044804, + "tokens_seen": 1769144320 + }, + { + "epoch": 0.54, + "learning_rate": 0.0003749809428284855, + "loss": 1.2415, + "theoretical_loss": 3.494539283233885, + "tokens_seen": 1769209856 + }, + { + "epoch": 0.54, + "learning_rate": 0.0003749648946840522, + "loss": 1.2918, + "theoretical_loss": 3.4945283185831846, + "tokens_seen": 1769275392 + }, + { + "epoch": 0.54, + "learning_rate": 0.0003749488465396189, + "loss": 1.3028, + "theoretical_loss": 3.4945173544523356, + "tokens_seen": 1769340928 + }, + { + "epoch": 0.54, + "learning_rate": 0.0003749327983951856, + "loss": 1.2571, + "theoretical_loss": 3.494506390841294, + "tokens_seen": 1769406464 + }, + { + "epoch": 0.54, + "objective/train/docs_used": 646001, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.2559581995010376, + "objective/train/theoretical_loss": 3.4944954277500155, + "objective/train/tokens_used": 1789932000, + "theoretical_loss": 3.4944954277500155, + "tokens_seen": 1769472000 + }, + { + "epoch": 0.54, + "learning_rate": 0.0003749167502507523, + "loss": 1.2687, + "theoretical_loss": 3.4944954277500155, + "tokens_seen": 1769472000 + }, + { + "epoch": 0.54, + "learning_rate": 0.000374900702106319, + "loss": 1.393, + "theoretical_loss": 3.4944844651784566, + "tokens_seen": 1769537536 + }, + { + "epoch": 0.54, + "learning_rate": 0.0003748846539618857, + "loss": 1.3511, + "theoretical_loss": 3.4944735031265726, + "tokens_seen": 1769603072 + }, + { + "epoch": 0.54, + "learning_rate": 0.0003748686058174524, + "loss": 1.3154, + "theoretical_loss": 3.494462541594321, + "tokens_seen": 1769668608 + }, + { + "epoch": 0.54, + "learning_rate": 0.0003748525576730191, + "loss": 1.2831, + "theoretical_loss": 3.4944515805816563, + "tokens_seen": 1769734144 + }, + { + "epoch": 0.54, + "learning_rate": 0.0003748365095285858, + "loss": 1.4571, + "theoretical_loss": 3.494440620088536, + "tokens_seen": 1769799680 + }, + { + "epoch": 0.54, + "learning_rate": 0.0003748204613841525, + "loss": 1.3491, + "theoretical_loss": 3.494429660114916, + "tokens_seen": 1769865216 + }, + { + "epoch": 0.54, + "learning_rate": 0.0003748044132397192, + "loss": 1.3386, + "theoretical_loss": 3.494418700660752, + "tokens_seen": 1769930752 + }, + { + "epoch": 0.54, + "learning_rate": 0.0003747883650952859, + "loss": 1.3537, + "theoretical_loss": 3.4944077417260004, + "tokens_seen": 1769996288 + }, + { + "epoch": 0.54, + "learning_rate": 0.00037477231695085255, + "loss": 1.4747, + "theoretical_loss": 3.4943967833106173, + "tokens_seen": 1770061824 + }, + { + "epoch": 0.54, + "learning_rate": 0.00037475626880641925, + "loss": 1.2761, + "theoretical_loss": 3.4943858254145583, + "tokens_seen": 1770127360 + }, + { + "epoch": 0.54, + "learning_rate": 0.00037474022066198595, + "loss": 1.4469, + "theoretical_loss": 3.494374868037781, + "tokens_seen": 1770192896 + }, + { + "epoch": 0.54, + "learning_rate": 0.00037472417251755265, + "loss": 1.4083, + "theoretical_loss": 3.49436391118024, + "tokens_seen": 1770258432 + }, + { + "epoch": 0.54, + "learning_rate": 0.00037470812437311935, + "loss": 1.3305, + "theoretical_loss": 3.4943529548418923, + "tokens_seen": 1770323968 + }, + { + "epoch": 0.54, + "learning_rate": 0.00037469207622868605, + "loss": 1.3921, + "theoretical_loss": 3.4943419990226943, + "tokens_seen": 1770389504 + }, + { + "epoch": 0.54, + "learning_rate": 0.00037467602808425275, + "loss": 1.2716, + "theoretical_loss": 3.4943310437226014, + "tokens_seen": 1770455040 + }, + { + "epoch": 0.54, + "learning_rate": 0.00037465997993981945, + "loss": 1.4951, + "theoretical_loss": 3.49432008894157, + "tokens_seen": 1770520576 + }, + { + "epoch": 0.54, + "learning_rate": 0.00037464393179538615, + "loss": 1.3309, + "theoretical_loss": 3.4943091346795567, + "tokens_seen": 1770586112 + }, + { + "epoch": 0.54, + "learning_rate": 0.00037462788365095286, + "loss": 1.444, + "theoretical_loss": 3.494298180936518, + "tokens_seen": 1770651648 + }, + { + "epoch": 0.54, + "learning_rate": 0.00037461183550651956, + "loss": 1.3646, + "theoretical_loss": 3.4942872277124093, + "tokens_seen": 1770717184 + }, + { + "epoch": 0.54, + "learning_rate": 0.00037459578736208626, + "loss": 1.2839, + "theoretical_loss": 3.494276275007187, + "tokens_seen": 1770782720 + }, + { + "epoch": 0.54, + "learning_rate": 0.00037457973921765296, + "loss": 1.3952, + "theoretical_loss": 3.494265322820808, + "tokens_seen": 1770848256 + }, + { + "epoch": 0.54, + "learning_rate": 0.00037456369107321966, + "loss": 1.3867, + "theoretical_loss": 3.494254371153227, + "tokens_seen": 1770913792 + }, + { + "epoch": 0.54, + "learning_rate": 0.00037454764292878636, + "loss": 1.3791, + "theoretical_loss": 3.4942434200044015, + "tokens_seen": 1770979328 + }, + { + "epoch": 0.54, + "learning_rate": 0.00037453159478435306, + "loss": 1.3558, + "theoretical_loss": 3.494232469374287, + "tokens_seen": 1771044864 + }, + { + "epoch": 0.54, + "learning_rate": 0.00037451554663991976, + "loss": 1.307, + "theoretical_loss": 3.494221519262841, + "tokens_seen": 1771110400 + }, + { + "epoch": 0.54, + "learning_rate": 0.00037449949849548646, + "loss": 1.3314, + "theoretical_loss": 3.4942105696700185, + "tokens_seen": 1771175936 + }, + { + "epoch": 0.54, + "learning_rate": 0.00037448345035105316, + "loss": 1.3472, + "theoretical_loss": 3.4941996205957766, + "tokens_seen": 1771241472 + }, + { + "epoch": 0.54, + "learning_rate": 0.00037446740220661986, + "loss": 1.3707, + "theoretical_loss": 3.4941886720400706, + "tokens_seen": 1771307008 + }, + { + "epoch": 0.54, + "learning_rate": 0.00037445135406218656, + "loss": 1.3539, + "theoretical_loss": 3.4941777240028573, + "tokens_seen": 1771372544 + }, + { + "epoch": 0.54, + "learning_rate": 0.00037443530591775326, + "loss": 1.3697, + "theoretical_loss": 3.494166776484093, + "tokens_seen": 1771438080 + }, + { + "epoch": 0.54, + "learning_rate": 0.00037441925777331996, + "loss": 1.354, + "theoretical_loss": 3.494155829483734, + "tokens_seen": 1771503616 + }, + { + "epoch": 0.54, + "learning_rate": 0.00037440320962888666, + "loss": 1.3195, + "theoretical_loss": 3.494144883001736, + "tokens_seen": 1771569152 + }, + { + "epoch": 0.54, + "learning_rate": 0.00037438716148445336, + "loss": 1.4046, + "theoretical_loss": 3.4941339370380566, + "tokens_seen": 1771634688 + }, + { + "epoch": 0.54, + "learning_rate": 0.00037437111334002007, + "loss": 1.4055, + "theoretical_loss": 3.4941229915926506, + "tokens_seen": 1771700224 + }, + { + "epoch": 0.54, + "learning_rate": 0.00037435506519558677, + "loss": 1.3627, + "theoretical_loss": 3.4941120466654745, + "tokens_seen": 1771765760 + }, + { + "epoch": 0.54, + "learning_rate": 0.00037433901705115347, + "loss": 1.4376, + "theoretical_loss": 3.4941011022564856, + "tokens_seen": 1771831296 + }, + { + "epoch": 0.54, + "learning_rate": 0.00037432296890672017, + "loss": 1.3656, + "theoretical_loss": 3.4940901583656396, + "tokens_seen": 1771896832 + }, + { + "epoch": 0.54, + "learning_rate": 0.00037430692076228687, + "loss": 1.3828, + "theoretical_loss": 3.494079214992893, + "tokens_seen": 1771962368 + }, + { + "epoch": 0.54, + "learning_rate": 0.00037429087261785357, + "loss": 1.3932, + "theoretical_loss": 3.4940682721382017, + "tokens_seen": 1772027904 + }, + { + "epoch": 0.54, + "learning_rate": 0.00037427482447342027, + "loss": 1.3657, + "theoretical_loss": 3.4940573298015223, + "tokens_seen": 1772093440 + }, + { + "epoch": 0.54, + "learning_rate": 0.00037425877632898697, + "loss": 1.3373, + "theoretical_loss": 3.494046387982811, + "tokens_seen": 1772158976 + }, + { + "epoch": 0.54, + "learning_rate": 0.00037424272818455367, + "loss": 1.3703, + "theoretical_loss": 3.494035446682024, + "tokens_seen": 1772224512 + }, + { + "epoch": 0.54, + "learning_rate": 0.00037422668004012037, + "loss": 1.3039, + "theoretical_loss": 3.494024505899118, + "tokens_seen": 1772290048 + }, + { + "epoch": 0.54, + "learning_rate": 0.00037421063189568707, + "loss": 1.4259, + "theoretical_loss": 3.4940135656340496, + "tokens_seen": 1772355584 + }, + { + "epoch": 0.54, + "learning_rate": 0.00037419458375125377, + "loss": 1.4185, + "theoretical_loss": 3.4940026258867745, + "tokens_seen": 1772421120 + }, + { + "epoch": 0.54, + "learning_rate": 0.00037417853560682047, + "loss": 1.3349, + "theoretical_loss": 3.4939916866572487, + "tokens_seen": 1772486656 + }, + { + "epoch": 0.54, + "learning_rate": 0.00037416248746238723, + "loss": 1.3835, + "theoretical_loss": 3.4939807479454297, + "tokens_seen": 1772552192 + }, + { + "epoch": 0.54, + "learning_rate": 0.0003741464393179539, + "loss": 1.2986, + "theoretical_loss": 3.493969809751273, + "tokens_seen": 1772617728 + }, + { + "epoch": 0.54, + "learning_rate": 0.0003741303911735206, + "loss": 1.3592, + "theoretical_loss": 3.4939588720747357, + "tokens_seen": 1772683264 + }, + { + "epoch": 0.54, + "objective/train/docs_used": 647211, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.5324468612670898, + "objective/train/theoretical_loss": 3.493947934915773, + "objective/train/tokens_used": 1793208800, + "theoretical_loss": 3.493947934915773, + "tokens_seen": 1772748800 + }, + { + "epoch": 0.54, + "learning_rate": 0.0003741143430290873, + "loss": 1.4457, + "theoretical_loss": 3.493947934915773, + "tokens_seen": 1772748800 + }, + { + "epoch": 0.54, + "learning_rate": 0.000374098294884654, + "loss": 1.3502, + "theoretical_loss": 3.493936998274343, + "tokens_seen": 1772814336 + }, + { + "epoch": 0.54, + "learning_rate": 0.0003740822467402207, + "loss": 1.2896, + "theoretical_loss": 3.4939260621504005, + "tokens_seen": 1772879872 + }, + { + "epoch": 0.54, + "learning_rate": 0.0003740661985957874, + "loss": 1.333, + "theoretical_loss": 3.493915126543902, + "tokens_seen": 1772945408 + }, + { + "epoch": 0.54, + "learning_rate": 0.0003740501504513541, + "loss": 1.4546, + "theoretical_loss": 3.493904191454805, + "tokens_seen": 1773010944 + }, + { + "epoch": 0.54, + "learning_rate": 0.0003740341023069208, + "loss": 1.4406, + "theoretical_loss": 3.493893256883065, + "tokens_seen": 1773076480 + }, + { + "epoch": 0.54, + "learning_rate": 0.0003740180541624875, + "loss": 1.2688, + "theoretical_loss": 3.4938823228286386, + "tokens_seen": 1773142016 + }, + { + "epoch": 0.54, + "learning_rate": 0.0003740020060180542, + "loss": 1.4484, + "theoretical_loss": 3.493871389291482, + "tokens_seen": 1773207552 + }, + { + "epoch": 0.54, + "learning_rate": 0.0003739859578736209, + "loss": 1.3405, + "theoretical_loss": 3.4938604562715527, + "tokens_seen": 1773273088 + }, + { + "epoch": 0.54, + "learning_rate": 0.0003739699097291876, + "loss": 1.2592, + "theoretical_loss": 3.4938495237688056, + "tokens_seen": 1773338624 + }, + { + "epoch": 0.54, + "learning_rate": 0.0003739538615847543, + "loss": 1.4516, + "theoretical_loss": 3.4938385917831978, + "tokens_seen": 1773404160 + }, + { + "epoch": 0.54, + "learning_rate": 0.000373937813440321, + "loss": 1.4195, + "theoretical_loss": 3.493827660314686, + "tokens_seen": 1773469696 + }, + { + "epoch": 0.54, + "learning_rate": 0.0003739217652958877, + "loss": 1.165, + "theoretical_loss": 3.4938167293632265, + "tokens_seen": 1773535232 + }, + { + "epoch": 0.54, + "learning_rate": 0.0003739057171514544, + "loss": 1.3649, + "theoretical_loss": 3.493805798928775, + "tokens_seen": 1773600768 + }, + { + "epoch": 0.54, + "learning_rate": 0.0003738896690070211, + "loss": 1.3653, + "theoretical_loss": 3.493794869011289, + "tokens_seen": 1773666304 + }, + { + "epoch": 0.54, + "learning_rate": 0.0003738736208625878, + "loss": 1.3194, + "theoretical_loss": 3.493783939610725, + "tokens_seen": 1773731840 + }, + { + "epoch": 0.54, + "learning_rate": 0.0003738575727181545, + "loss": 1.3436, + "theoretical_loss": 3.493773010727038, + "tokens_seen": 1773797376 + }, + { + "epoch": 0.54, + "learning_rate": 0.0003738415245737212, + "loss": 1.3517, + "theoretical_loss": 3.493762082360186, + "tokens_seen": 1773862912 + }, + { + "epoch": 0.54, + "learning_rate": 0.0003738254764292879, + "loss": 1.3102, + "theoretical_loss": 3.4937511545101243, + "tokens_seen": 1773928448 + }, + { + "epoch": 0.54, + "learning_rate": 0.0003738094282848546, + "loss": 1.4142, + "theoretical_loss": 3.4937402271768105, + "tokens_seen": 1773993984 + }, + { + "epoch": 0.54, + "learning_rate": 0.0003737933801404213, + "loss": 1.3399, + "theoretical_loss": 3.493729300360201, + "tokens_seen": 1774059520 + }, + { + "epoch": 0.54, + "learning_rate": 0.000373777331995988, + "loss": 1.2229, + "theoretical_loss": 3.4937183740602507, + "tokens_seen": 1774125056 + }, + { + "epoch": 0.54, + "learning_rate": 0.0003737612838515547, + "loss": 1.2696, + "theoretical_loss": 3.4937074482769175, + "tokens_seen": 1774190592 + }, + { + "epoch": 0.54, + "learning_rate": 0.0003737452357071214, + "loss": 1.4449, + "theoretical_loss": 3.4936965230101578, + "tokens_seen": 1774256128 + }, + { + "epoch": 0.54, + "learning_rate": 0.0003737291875626881, + "loss": 1.3038, + "theoretical_loss": 3.493685598259928, + "tokens_seen": 1774321664 + }, + { + "epoch": 0.54, + "learning_rate": 0.0003737131394182548, + "loss": 1.3744, + "theoretical_loss": 3.4936746740261846, + "tokens_seen": 1774387200 + }, + { + "epoch": 0.54, + "learning_rate": 0.0003736970912738215, + "loss": 1.3399, + "theoretical_loss": 3.4936637503088837, + "tokens_seen": 1774452736 + }, + { + "epoch": 0.54, + "learning_rate": 0.0003736810431293882, + "loss": 1.2579, + "theoretical_loss": 3.493652827107982, + "tokens_seen": 1774518272 + }, + { + "epoch": 0.54, + "learning_rate": 0.0003736649949849549, + "loss": 1.2513, + "theoretical_loss": 3.493641904423437, + "tokens_seen": 1774583808 + }, + { + "epoch": 0.54, + "learning_rate": 0.0003736489468405216, + "loss": 1.3273, + "theoretical_loss": 3.493630982255203, + "tokens_seen": 1774649344 + }, + { + "epoch": 0.54, + "learning_rate": 0.0003736328986960883, + "loss": 1.3529, + "theoretical_loss": 3.493620060603239, + "tokens_seen": 1774714880 + }, + { + "epoch": 0.54, + "learning_rate": 0.000373616850551655, + "loss": 1.3514, + "theoretical_loss": 3.4936091394674995, + "tokens_seen": 1774780416 + }, + { + "epoch": 0.54, + "learning_rate": 0.0003736008024072217, + "loss": 1.3911, + "theoretical_loss": 3.4935982188479424, + "tokens_seen": 1774845952 + }, + { + "epoch": 0.54, + "learning_rate": 0.0003735847542627884, + "loss": 1.3443, + "theoretical_loss": 3.493587298744524, + "tokens_seen": 1774911488 + }, + { + "epoch": 0.54, + "learning_rate": 0.0003735687061183551, + "loss": 1.3049, + "theoretical_loss": 3.4935763791572003, + "tokens_seen": 1774977024 + }, + { + "epoch": 0.54, + "learning_rate": 0.0003735526579739218, + "loss": 1.3014, + "theoretical_loss": 3.4935654600859283, + "tokens_seen": 1775042560 + }, + { + "epoch": 0.54, + "learning_rate": 0.0003735366098294885, + "loss": 1.3182, + "theoretical_loss": 3.4935545415306644, + "tokens_seen": 1775108096 + }, + { + "epoch": 0.54, + "learning_rate": 0.0003735205616850552, + "loss": 1.3422, + "theoretical_loss": 3.4935436234913655, + "tokens_seen": 1775173632 + }, + { + "epoch": 0.54, + "learning_rate": 0.0003735045135406219, + "loss": 1.4121, + "theoretical_loss": 3.4935327059679873, + "tokens_seen": 1775239168 + }, + { + "epoch": 0.54, + "learning_rate": 0.0003734884653961886, + "loss": 1.3312, + "theoretical_loss": 3.4935217889604875, + "tokens_seen": 1775304704 + }, + { + "epoch": 0.54, + "learning_rate": 0.0003734724172517553, + "loss": 1.4294, + "theoretical_loss": 3.493510872468822, + "tokens_seen": 1775370240 + }, + { + "epoch": 0.54, + "learning_rate": 0.000373456369107322, + "loss": 1.4387, + "theoretical_loss": 3.493499956492948, + "tokens_seen": 1775435776 + }, + { + "epoch": 0.54, + "learning_rate": 0.0003734403209628887, + "loss": 1.4399, + "theoretical_loss": 3.493489041032821, + "tokens_seen": 1775501312 + }, + { + "epoch": 0.54, + "learning_rate": 0.0003734242728184554, + "loss": 1.3135, + "theoretical_loss": 3.493478126088398, + "tokens_seen": 1775566848 + }, + { + "epoch": 0.54, + "learning_rate": 0.0003734082246740221, + "loss": 1.2846, + "theoretical_loss": 3.4934672116596364, + "tokens_seen": 1775632384 + }, + { + "epoch": 0.54, + "learning_rate": 0.0003733921765295888, + "loss": 1.249, + "theoretical_loss": 3.4934562977464925, + "tokens_seen": 1775697920 + }, + { + "epoch": 0.54, + "learning_rate": 0.00037337612838515545, + "loss": 1.3538, + "theoretical_loss": 3.493445384348922, + "tokens_seen": 1775763456 + }, + { + "epoch": 0.54, + "learning_rate": 0.00037336008024072215, + "loss": 1.3175, + "theoretical_loss": 3.493434471466882, + "tokens_seen": 1775828992 + }, + { + "epoch": 0.54, + "learning_rate": 0.00037334403209628885, + "loss": 1.3773, + "theoretical_loss": 3.49342355910033, + "tokens_seen": 1775894528 + }, + { + "epoch": 0.54, + "learning_rate": 0.00037332798395185555, + "loss": 1.3299, + "theoretical_loss": 3.4934126472492215, + "tokens_seen": 1775960064 + }, + { + "epoch": 0.54, + "objective/train/docs_used": 648417, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.3483312129974365, + "objective/train/theoretical_loss": 3.4934017359135137, + "objective/train/tokens_used": 1796485600, + "theoretical_loss": 3.4934017359135137, + "tokens_seen": 1776025600 + }, + { + "epoch": 0.54, + "learning_rate": 0.00037331193580742225, + "loss": 1.4166, + "theoretical_loss": 3.4934017359135137, + "tokens_seen": 1776025600 + }, + { + "epoch": 0.54, + "learning_rate": 0.00037329588766298895, + "loss": 1.3107, + "theoretical_loss": 3.4933908250931633, + "tokens_seen": 1776091136 + }, + { + "epoch": 0.54, + "learning_rate": 0.00037327983951855565, + "loss": 1.2507, + "theoretical_loss": 3.493379914788126, + "tokens_seen": 1776156672 + }, + { + "epoch": 0.54, + "learning_rate": 0.00037326379137412235, + "loss": 1.3321, + "theoretical_loss": 3.49336900499836, + "tokens_seen": 1776222208 + }, + { + "epoch": 0.54, + "learning_rate": 0.00037324774322968906, + "loss": 1.5189, + "theoretical_loss": 3.4933580957238206, + "tokens_seen": 1776287744 + }, + { + "epoch": 0.54, + "learning_rate": 0.00037323169508525576, + "loss": 1.2761, + "theoretical_loss": 3.4933471869644652, + "tokens_seen": 1776353280 + }, + { + "epoch": 0.54, + "learning_rate": 0.00037321564694082246, + "loss": 1.2778, + "theoretical_loss": 3.4933362787202507, + "tokens_seen": 1776418816 + }, + { + "epoch": 0.54, + "learning_rate": 0.00037319959879638916, + "loss": 1.4079, + "theoretical_loss": 3.4933253709911325, + "tokens_seen": 1776484352 + }, + { + "epoch": 0.54, + "learning_rate": 0.00037318355065195586, + "loss": 1.3863, + "theoretical_loss": 3.493314463777069, + "tokens_seen": 1776549888 + }, + { + "epoch": 0.54, + "learning_rate": 0.00037316750250752256, + "loss": 1.3622, + "theoretical_loss": 3.493303557078015, + "tokens_seen": 1776615424 + }, + { + "epoch": 0.54, + "learning_rate": 0.00037315145436308926, + "loss": 1.3334, + "theoretical_loss": 3.493292650893929, + "tokens_seen": 1776680960 + }, + { + "epoch": 0.54, + "learning_rate": 0.00037313540621865596, + "loss": 1.468, + "theoretical_loss": 3.4932817452247664, + "tokens_seen": 1776746496 + }, + { + "epoch": 0.54, + "learning_rate": 0.00037311935807422266, + "loss": 1.3362, + "theoretical_loss": 3.493270840070484, + "tokens_seen": 1776812032 + }, + { + "epoch": 0.54, + "learning_rate": 0.00037310330992978936, + "loss": 1.3834, + "theoretical_loss": 3.4932599354310403, + "tokens_seen": 1776877568 + }, + { + "epoch": 0.54, + "learning_rate": 0.00037308726178535606, + "loss": 1.3275, + "theoretical_loss": 3.493249031306389, + "tokens_seen": 1776943104 + }, + { + "epoch": 0.54, + "learning_rate": 0.00037307121364092276, + "loss": 1.413, + "theoretical_loss": 3.4932381276964897, + "tokens_seen": 1777008640 + }, + { + "epoch": 0.54, + "learning_rate": 0.00037305516549648946, + "loss": 1.4479, + "theoretical_loss": 3.4932272246012968, + "tokens_seen": 1777074176 + }, + { + "epoch": 0.54, + "learning_rate": 0.00037303911735205616, + "loss": 1.3417, + "theoretical_loss": 3.4932163220207686, + "tokens_seen": 1777139712 + }, + { + "epoch": 0.54, + "learning_rate": 0.00037302306920762286, + "loss": 1.2489, + "theoretical_loss": 3.493205419954861, + "tokens_seen": 1777205248 + }, + { + "epoch": 0.54, + "learning_rate": 0.00037300702106318956, + "loss": 1.2932, + "theoretical_loss": 3.4931945184035307, + "tokens_seen": 1777270784 + }, + { + "epoch": 0.54, + "learning_rate": 0.00037299097291875627, + "loss": 1.222, + "theoretical_loss": 3.493183617366735, + "tokens_seen": 1777336320 + }, + { + "epoch": 0.54, + "learning_rate": 0.00037297492477432297, + "loss": 1.2212, + "theoretical_loss": 3.4931727168444304, + "tokens_seen": 1777401856 + }, + { + "epoch": 0.54, + "learning_rate": 0.0003729588766298897, + "loss": 1.4585, + "theoretical_loss": 3.4931618168365737, + "tokens_seen": 1777467392 + }, + { + "epoch": 0.54, + "learning_rate": 0.0003729428284854564, + "loss": 1.3289, + "theoretical_loss": 3.4931509173431214, + "tokens_seen": 1777532928 + }, + { + "epoch": 0.54, + "learning_rate": 0.0003729267803410231, + "loss": 1.363, + "theoretical_loss": 3.4931400183640307, + "tokens_seen": 1777598464 + }, + { + "epoch": 0.54, + "learning_rate": 0.0003729107321965898, + "loss": 1.3691, + "theoretical_loss": 3.4931291198992582, + "tokens_seen": 1777664000 + }, + { + "epoch": 0.54, + "learning_rate": 0.0003728946840521565, + "loss": 1.2269, + "theoretical_loss": 3.49311822194876, + "tokens_seen": 1777729536 + }, + { + "epoch": 0.54, + "learning_rate": 0.0003728786359077232, + "loss": 1.2544, + "theoretical_loss": 3.493107324512494, + "tokens_seen": 1777795072 + }, + { + "epoch": 0.54, + "learning_rate": 0.0003728625877632899, + "loss": 1.3492, + "theoretical_loss": 3.493096427590416, + "tokens_seen": 1777860608 + }, + { + "epoch": 0.54, + "learning_rate": 0.0003728465396188566, + "loss": 1.3101, + "theoretical_loss": 3.4930855311824835, + "tokens_seen": 1777926144 + }, + { + "epoch": 0.54, + "learning_rate": 0.0003728304914744233, + "loss": 1.2783, + "theoretical_loss": 3.493074635288653, + "tokens_seen": 1777991680 + }, + { + "epoch": 0.54, + "learning_rate": 0.00037281444332999, + "loss": 1.4002, + "theoretical_loss": 3.493063739908881, + "tokens_seen": 1778057216 + }, + { + "epoch": 0.54, + "learning_rate": 0.00037279839518555673, + "loss": 1.3938, + "theoretical_loss": 3.493052845043125, + "tokens_seen": 1778122752 + }, + { + "epoch": 0.54, + "learning_rate": 0.00037278234704112343, + "loss": 1.355, + "theoretical_loss": 3.493041950691341, + "tokens_seen": 1778188288 + }, + { + "epoch": 0.54, + "learning_rate": 0.00037276629889669013, + "loss": 1.3541, + "theoretical_loss": 3.4930310568534866, + "tokens_seen": 1778253824 + }, + { + "epoch": 0.54, + "learning_rate": 0.0003727502507522568, + "loss": 1.4255, + "theoretical_loss": 3.4930201635295175, + "tokens_seen": 1778319360 + }, + { + "epoch": 0.54, + "learning_rate": 0.0003727342026078235, + "loss": 1.4352, + "theoretical_loss": 3.493009270719392, + "tokens_seen": 1778384896 + }, + { + "epoch": 0.54, + "learning_rate": 0.0003727181544633902, + "loss": 1.3157, + "theoretical_loss": 3.4929983784230663, + "tokens_seen": 1778450432 + }, + { + "epoch": 0.54, + "learning_rate": 0.0003727021063189569, + "loss": 1.3876, + "theoretical_loss": 3.492987486640496, + "tokens_seen": 1778515968 + }, + { + "epoch": 0.54, + "learning_rate": 0.0003726860581745236, + "loss": 1.332, + "theoretical_loss": 3.4929765953716405, + "tokens_seen": 1778581504 + }, + { + "epoch": 0.54, + "learning_rate": 0.0003726700100300903, + "loss": 1.4613, + "theoretical_loss": 3.4929657046164544, + "tokens_seen": 1778647040 + }, + { + "epoch": 0.54, + "learning_rate": 0.000372653961885657, + "loss": 1.4064, + "theoretical_loss": 3.4929548143748956, + "tokens_seen": 1778712576 + }, + { + "epoch": 0.54, + "learning_rate": 0.0003726379137412237, + "loss": 1.3463, + "theoretical_loss": 3.4929439246469203, + "tokens_seen": 1778778112 + }, + { + "epoch": 0.54, + "learning_rate": 0.0003726218655967904, + "loss": 1.2944, + "theoretical_loss": 3.492933035432486, + "tokens_seen": 1778843648 + }, + { + "epoch": 0.54, + "learning_rate": 0.0003726058174523571, + "loss": 1.2996, + "theoretical_loss": 3.492922146731549, + "tokens_seen": 1778909184 + }, + { + "epoch": 0.54, + "learning_rate": 0.0003725897693079238, + "loss": 1.3444, + "theoretical_loss": 3.4929112585440674, + "tokens_seen": 1778974720 + }, + { + "epoch": 0.54, + "learning_rate": 0.0003725737211634905, + "loss": 1.3194, + "theoretical_loss": 3.492900370869996, + "tokens_seen": 1779040256 + }, + { + "epoch": 0.54, + "learning_rate": 0.0003725576730190572, + "loss": 1.3879, + "theoretical_loss": 3.492889483709294, + "tokens_seen": 1779105792 + }, + { + "epoch": 0.54, + "learning_rate": 0.0003725416248746239, + "loss": 1.4119, + "theoretical_loss": 3.4928785970619165, + "tokens_seen": 1779171328 + }, + { + "epoch": 0.54, + "learning_rate": 0.0003725255767301906, + "loss": 1.3604, + "theoretical_loss": 3.492867710927821, + "tokens_seen": 1779236864 + }, + { + "epoch": 0.54, + "objective/train/docs_used": 649543, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.4393491744995117, + "objective/train/theoretical_loss": 3.4928568253069643, + "objective/train/tokens_used": 1799762400, + "theoretical_loss": 3.4928568253069643, + "tokens_seen": 1779302400 + }, + { + "epoch": 0.54, + "learning_rate": 0.0003725095285857573, + "loss": 1.3717, + "theoretical_loss": 3.4928568253069643, + "tokens_seen": 1779302400 + }, + { + "epoch": 0.54, + "learning_rate": 0.000372493480441324, + "loss": 1.3673, + "theoretical_loss": 3.4928459401993033, + "tokens_seen": 1779367936 + }, + { + "epoch": 0.54, + "learning_rate": 0.0003724774322968907, + "loss": 1.3151, + "theoretical_loss": 3.492835055604796, + "tokens_seen": 1779433472 + }, + { + "epoch": 0.54, + "learning_rate": 0.0003724613841524574, + "loss": 1.2684, + "theoretical_loss": 3.492824171523397, + "tokens_seen": 1779499008 + }, + { + "epoch": 0.54, + "learning_rate": 0.0003724453360080241, + "loss": 1.2824, + "theoretical_loss": 3.492813287955065, + "tokens_seen": 1779564544 + }, + { + "epoch": 0.54, + "learning_rate": 0.0003724292878635908, + "loss": 1.3572, + "theoretical_loss": 3.4928024048997566, + "tokens_seen": 1779630080 + }, + { + "epoch": 0.54, + "learning_rate": 0.0003724132397191575, + "loss": 1.4371, + "theoretical_loss": 3.492791522357429, + "tokens_seen": 1779695616 + }, + { + "epoch": 0.54, + "learning_rate": 0.0003723971915747242, + "loss": 1.2773, + "theoretical_loss": 3.4927806403280384, + "tokens_seen": 1779761152 + }, + { + "epoch": 0.54, + "learning_rate": 0.0003723811434302909, + "loss": 1.3497, + "theoretical_loss": 3.4927697588115416, + "tokens_seen": 1779826688 + }, + { + "epoch": 0.54, + "learning_rate": 0.0003723650952858576, + "loss": 1.3652, + "theoretical_loss": 3.492758877807897, + "tokens_seen": 1779892224 + }, + { + "epoch": 0.54, + "learning_rate": 0.0003723490471414243, + "loss": 1.4759, + "theoretical_loss": 3.4927479973170596, + "tokens_seen": 1779957760 + }, + { + "epoch": 0.54, + "learning_rate": 0.000372332998996991, + "loss": 1.4271, + "theoretical_loss": 3.4927371173389874, + "tokens_seen": 1780023296 + }, + { + "epoch": 0.54, + "learning_rate": 0.0003723169508525577, + "loss": 1.5673, + "theoretical_loss": 3.4927262378736375, + "tokens_seen": 1780088832 + }, + { + "epoch": 0.54, + "learning_rate": 0.0003723009027081244, + "loss": 1.3647, + "theoretical_loss": 3.4927153589209667, + "tokens_seen": 1780154368 + }, + { + "epoch": 0.54, + "learning_rate": 0.0003722848545636911, + "loss": 1.3088, + "theoretical_loss": 3.492704480480932, + "tokens_seen": 1780219904 + }, + { + "epoch": 0.54, + "learning_rate": 0.0003722688064192578, + "loss": 1.3897, + "theoretical_loss": 3.49269360255349, + "tokens_seen": 1780285440 + }, + { + "epoch": 0.54, + "learning_rate": 0.0003722527582748245, + "loss": 1.2123, + "theoretical_loss": 3.492682725138598, + "tokens_seen": 1780350976 + }, + { + "epoch": 0.54, + "learning_rate": 0.0003722367101303912, + "loss": 1.3821, + "theoretical_loss": 3.492671848236213, + "tokens_seen": 1780416512 + }, + { + "epoch": 0.54, + "learning_rate": 0.0003722206619859579, + "loss": 1.2955, + "theoretical_loss": 3.4926609718462918, + "tokens_seen": 1780482048 + }, + { + "epoch": 0.54, + "learning_rate": 0.0003722046138415246, + "loss": 1.3008, + "theoretical_loss": 3.4926500959687914, + "tokens_seen": 1780547584 + }, + { + "epoch": 0.54, + "learning_rate": 0.0003721885656970913, + "loss": 1.2872, + "theoretical_loss": 3.492639220603669, + "tokens_seen": 1780613120 + }, + { + "epoch": 0.54, + "learning_rate": 0.000372172517552658, + "loss": 1.3484, + "theoretical_loss": 3.4926283457508815, + "tokens_seen": 1780678656 + }, + { + "epoch": 0.54, + "learning_rate": 0.0003721564694082247, + "loss": 1.3483, + "theoretical_loss": 3.4926174714103864, + "tokens_seen": 1780744192 + }, + { + "epoch": 0.54, + "learning_rate": 0.0003721404212637914, + "loss": 1.3428, + "theoretical_loss": 3.4926065975821396, + "tokens_seen": 1780809728 + }, + { + "epoch": 0.54, + "learning_rate": 0.0003721243731193581, + "loss": 1.3992, + "theoretical_loss": 3.4925957242660988, + "tokens_seen": 1780875264 + }, + { + "epoch": 0.54, + "learning_rate": 0.0003721083249749248, + "loss": 1.3644, + "theoretical_loss": 3.4925848514622215, + "tokens_seen": 1780940800 + }, + { + "epoch": 0.54, + "learning_rate": 0.0003720922768304915, + "loss": 1.3971, + "theoretical_loss": 3.4925739791704635, + "tokens_seen": 1781006336 + }, + { + "epoch": 0.54, + "learning_rate": 0.0003720762286860582, + "loss": 1.3462, + "theoretical_loss": 3.492563107390783, + "tokens_seen": 1781071872 + }, + { + "epoch": 0.54, + "learning_rate": 0.0003720601805416249, + "loss": 1.2293, + "theoretical_loss": 3.4925522361231365, + "tokens_seen": 1781137408 + }, + { + "epoch": 0.54, + "learning_rate": 0.0003720441323971916, + "loss": 1.4799, + "theoretical_loss": 3.4925413653674813, + "tokens_seen": 1781202944 + }, + { + "epoch": 0.54, + "learning_rate": 0.0003720280842527583, + "loss": 1.3532, + "theoretical_loss": 3.4925304951237734, + "tokens_seen": 1781268480 + }, + { + "epoch": 0.54, + "learning_rate": 0.000372012036108325, + "loss": 1.1765, + "theoretical_loss": 3.4925196253919717, + "tokens_seen": 1781334016 + }, + { + "epoch": 0.54, + "learning_rate": 0.0003719959879638917, + "loss": 1.2812, + "theoretical_loss": 3.4925087561720325, + "tokens_seen": 1781399552 + }, + { + "epoch": 0.54, + "learning_rate": 0.0003719799398194584, + "loss": 1.4457, + "theoretical_loss": 3.4924978874639114, + "tokens_seen": 1781465088 + }, + { + "epoch": 0.54, + "learning_rate": 0.00037196389167502505, + "loss": 1.3087, + "theoretical_loss": 3.4924870192675677, + "tokens_seen": 1781530624 + }, + { + "epoch": 0.54, + "learning_rate": 0.00037194784353059175, + "loss": 1.4336, + "theoretical_loss": 3.4924761515829577, + "tokens_seen": 1781596160 + }, + { + "epoch": 0.54, + "learning_rate": 0.00037193179538615845, + "loss": 1.3229, + "theoretical_loss": 3.492465284410038, + "tokens_seen": 1781661696 + }, + { + "epoch": 0.54, + "learning_rate": 0.00037191574724172515, + "loss": 1.4223, + "theoretical_loss": 3.4924544177487657, + "tokens_seen": 1781727232 + }, + { + "epoch": 0.54, + "learning_rate": 0.00037189969909729185, + "loss": 1.2624, + "theoretical_loss": 3.4924435515990986, + "tokens_seen": 1781792768 + }, + { + "epoch": 0.54, + "learning_rate": 0.00037188365095285856, + "loss": 1.3628, + "theoretical_loss": 3.492432685960993, + "tokens_seen": 1781858304 + }, + { + "epoch": 0.54, + "learning_rate": 0.00037186760280842526, + "loss": 1.3948, + "theoretical_loss": 3.492421820834406, + "tokens_seen": 1781923840 + }, + { + "epoch": 0.54, + "learning_rate": 0.00037185155466399196, + "loss": 1.36, + "theoretical_loss": 3.492410956219296, + "tokens_seen": 1781989376 + }, + { + "epoch": 0.54, + "learning_rate": 0.00037183550651955866, + "loss": 1.4394, + "theoretical_loss": 3.492400092115619, + "tokens_seen": 1782054912 + }, + { + "epoch": 0.54, + "learning_rate": 0.00037181945837512536, + "loss": 1.4156, + "theoretical_loss": 3.492389228523332, + "tokens_seen": 1782120448 + }, + { + "epoch": 0.54, + "learning_rate": 0.00037180341023069206, + "loss": 1.2797, + "theoretical_loss": 3.4923783654423923, + "tokens_seen": 1782185984 + }, + { + "epoch": 0.54, + "learning_rate": 0.00037178736208625876, + "loss": 1.2576, + "theoretical_loss": 3.4923675028727574, + "tokens_seen": 1782251520 + }, + { + "epoch": 0.54, + "learning_rate": 0.0003717713139418255, + "loss": 1.2694, + "theoretical_loss": 3.4923566408143842, + "tokens_seen": 1782317056 + }, + { + "epoch": 0.54, + "learning_rate": 0.0003717552657973922, + "loss": 1.3455, + "theoretical_loss": 3.49234577926723, + "tokens_seen": 1782382592 + }, + { + "epoch": 0.54, + "learning_rate": 0.0003717392176529589, + "loss": 1.3764, + "theoretical_loss": 3.4923349182312515, + "tokens_seen": 1782448128 + }, + { + "epoch": 0.54, + "learning_rate": 0.0003717231695085256, + "loss": 1.411, + "theoretical_loss": 3.4923240577064067, + "tokens_seen": 1782513664 + }, + { + "epoch": 0.54, + "objective/train/docs_used": 650724, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.5353834629058838, + "objective/train/theoretical_loss": 3.4923131976926514, + "objective/train/tokens_used": 1803039200, + "theoretical_loss": 3.4923131976926514, + "tokens_seen": 1782579200 + }, + { + "epoch": 0.54, + "learning_rate": 0.0003717071213640923, + "loss": 1.4398, + "theoretical_loss": 3.4923131976926514, + "tokens_seen": 1782579200 + }, + { + "epoch": 0.54, + "learning_rate": 0.000371691073219659, + "loss": 1.3384, + "theoretical_loss": 3.492302338189944, + "tokens_seen": 1782644736 + }, + { + "epoch": 0.54, + "learning_rate": 0.0003716750250752257, + "loss": 1.3941, + "theoretical_loss": 3.492291479198241, + "tokens_seen": 1782710272 + }, + { + "epoch": 0.54, + "learning_rate": 0.0003716589769307924, + "loss": 1.3669, + "theoretical_loss": 3.4922806207174997, + "tokens_seen": 1782775808 + }, + { + "epoch": 0.54, + "learning_rate": 0.0003716429287863591, + "loss": 1.3591, + "theoretical_loss": 3.4922697627476778, + "tokens_seen": 1782841344 + }, + { + "epoch": 0.54, + "learning_rate": 0.0003716268806419258, + "loss": 1.4082, + "theoretical_loss": 3.4922589052887316, + "tokens_seen": 1782906880 + }, + { + "epoch": 0.54, + "learning_rate": 0.0003716108324974925, + "loss": 1.4124, + "theoretical_loss": 3.492248048340619, + "tokens_seen": 1782972416 + }, + { + "epoch": 0.54, + "learning_rate": 0.0003715947843530592, + "loss": 1.4556, + "theoretical_loss": 3.492237191903297, + "tokens_seen": 1783037952 + }, + { + "epoch": 0.54, + "learning_rate": 0.0003715787362086259, + "loss": 1.3691, + "theoretical_loss": 3.4922263359767225, + "tokens_seen": 1783103488 + }, + { + "epoch": 0.54, + "learning_rate": 0.0003715626880641926, + "loss": 1.4327, + "theoretical_loss": 3.492215480560853, + "tokens_seen": 1783169024 + }, + { + "epoch": 0.54, + "learning_rate": 0.0003715466399197593, + "loss": 1.401, + "theoretical_loss": 3.4922046256556456, + "tokens_seen": 1783234560 + }, + { + "epoch": 0.54, + "learning_rate": 0.000371530591775326, + "loss": 1.3122, + "theoretical_loss": 3.4921937712610576, + "tokens_seen": 1783300096 + }, + { + "epoch": 0.54, + "learning_rate": 0.0003715145436308927, + "loss": 1.55, + "theoretical_loss": 3.492182917377046, + "tokens_seen": 1783365632 + }, + { + "epoch": 0.54, + "learning_rate": 0.0003714984954864594, + "loss": 1.3689, + "theoretical_loss": 3.492172064003568, + "tokens_seen": 1783431168 + }, + { + "epoch": 0.54, + "learning_rate": 0.0003714824473420261, + "loss": 1.3569, + "theoretical_loss": 3.492161211140581, + "tokens_seen": 1783496704 + }, + { + "epoch": 0.54, + "learning_rate": 0.0003714663991975928, + "loss": 1.3757, + "theoretical_loss": 3.4921503587880425, + "tokens_seen": 1783562240 + }, + { + "epoch": 0.54, + "learning_rate": 0.0003714503510531595, + "loss": 1.4135, + "theoretical_loss": 3.4921395069459096, + "tokens_seen": 1783627776 + }, + { + "epoch": 0.54, + "learning_rate": 0.00037143430290872623, + "loss": 1.3362, + "theoretical_loss": 3.492128655614139, + "tokens_seen": 1783693312 + }, + { + "epoch": 0.54, + "learning_rate": 0.00037141825476429293, + "loss": 1.3401, + "theoretical_loss": 3.492117804792689, + "tokens_seen": 1783758848 + }, + { + "epoch": 0.54, + "learning_rate": 0.00037140220661985963, + "loss": 1.3656, + "theoretical_loss": 3.4921069544815153, + "tokens_seen": 1783824384 + }, + { + "epoch": 0.54, + "learning_rate": 0.00037138615847542633, + "loss": 1.4017, + "theoretical_loss": 3.4920961046805763, + "tokens_seen": 1783889920 + }, + { + "epoch": 0.54, + "learning_rate": 0.00037137011033099303, + "loss": 1.323, + "theoretical_loss": 3.4920852553898296, + "tokens_seen": 1783955456 + }, + { + "epoch": 0.54, + "learning_rate": 0.00037135406218655973, + "loss": 1.3631, + "theoretical_loss": 3.4920744066092313, + "tokens_seen": 1784020992 + }, + { + "epoch": 0.54, + "learning_rate": 0.0003713380140421264, + "loss": 1.3849, + "theoretical_loss": 3.4920635583387396, + "tokens_seen": 1784086528 + }, + { + "epoch": 0.54, + "learning_rate": 0.0003713219658976931, + "loss": 1.2751, + "theoretical_loss": 3.492052710578311, + "tokens_seen": 1784152064 + }, + { + "epoch": 0.54, + "learning_rate": 0.0003713059177532598, + "loss": 1.3193, + "theoretical_loss": 3.492041863327904, + "tokens_seen": 1784217600 + }, + { + "epoch": 0.54, + "learning_rate": 0.0003712898696088265, + "loss": 1.3343, + "theoretical_loss": 3.4920310165874744, + "tokens_seen": 1784283136 + }, + { + "epoch": 0.54, + "learning_rate": 0.0003712738214643932, + "loss": 1.3965, + "theoretical_loss": 3.49202017035698, + "tokens_seen": 1784348672 + }, + { + "epoch": 0.54, + "learning_rate": 0.0003712577733199599, + "loss": 1.3395, + "theoretical_loss": 3.492009324636379, + "tokens_seen": 1784414208 + }, + { + "epoch": 0.54, + "learning_rate": 0.0003712417251755266, + "loss": 1.3972, + "theoretical_loss": 3.491998479425628, + "tokens_seen": 1784479744 + }, + { + "epoch": 0.54, + "learning_rate": 0.0003712256770310933, + "loss": 1.2491, + "theoretical_loss": 3.491987634724684, + "tokens_seen": 1784545280 + }, + { + "epoch": 0.54, + "learning_rate": 0.00037120962888666, + "loss": 1.3307, + "theoretical_loss": 3.4919767905335046, + "tokens_seen": 1784610816 + }, + { + "epoch": 0.54, + "learning_rate": 0.0003711935807422267, + "loss": 1.3658, + "theoretical_loss": 3.491965946852047, + "tokens_seen": 1784676352 + }, + { + "epoch": 0.54, + "learning_rate": 0.0003711775325977934, + "loss": 1.3497, + "theoretical_loss": 3.491955103680269, + "tokens_seen": 1784741888 + }, + { + "epoch": 0.54, + "learning_rate": 0.0003711614844533601, + "loss": 1.5218, + "theoretical_loss": 3.4919442610181277, + "tokens_seen": 1784807424 + }, + { + "epoch": 0.54, + "learning_rate": 0.0003711454363089268, + "loss": 1.359, + "theoretical_loss": 3.49193341886558, + "tokens_seen": 1784872960 + }, + { + "epoch": 0.54, + "learning_rate": 0.0003711293881644935, + "loss": 1.4921, + "theoretical_loss": 3.491922577222584, + "tokens_seen": 1784938496 + }, + { + "epoch": 0.54, + "learning_rate": 0.0003711133400200602, + "loss": 1.3886, + "theoretical_loss": 3.4919117360890963, + "tokens_seen": 1785004032 + }, + { + "epoch": 0.54, + "learning_rate": 0.0003710972918756269, + "loss": 1.4477, + "theoretical_loss": 3.4919008954650748, + "tokens_seen": 1785069568 + }, + { + "epoch": 0.54, + "learning_rate": 0.0003710812437311936, + "loss": 1.4769, + "theoretical_loss": 3.491890055350477, + "tokens_seen": 1785135104 + }, + { + "epoch": 0.54, + "learning_rate": 0.0003710651955867603, + "loss": 1.3383, + "theoretical_loss": 3.491879215745259, + "tokens_seen": 1785200640 + }, + { + "epoch": 0.54, + "learning_rate": 0.000371049147442327, + "loss": 1.2994, + "theoretical_loss": 3.4918683766493794, + "tokens_seen": 1785266176 + }, + { + "epoch": 0.54, + "learning_rate": 0.0003710330992978937, + "loss": 1.3761, + "theoretical_loss": 3.491857538062795, + "tokens_seen": 1785331712 + }, + { + "epoch": 0.54, + "learning_rate": 0.0003710170511534604, + "loss": 1.3066, + "theoretical_loss": 3.491846699985464, + "tokens_seen": 1785397248 + }, + { + "epoch": 0.54, + "learning_rate": 0.0003710010030090271, + "loss": 1.3716, + "theoretical_loss": 3.491835862417343, + "tokens_seen": 1785462784 + }, + { + "epoch": 0.54, + "learning_rate": 0.0003709849548645938, + "loss": 1.3727, + "theoretical_loss": 3.491825025358389, + "tokens_seen": 1785528320 + }, + { + "epoch": 0.54, + "learning_rate": 0.0003709689067201605, + "loss": 1.238, + "theoretical_loss": 3.4918141888085605, + "tokens_seen": 1785593856 + }, + { + "epoch": 0.54, + "learning_rate": 0.0003709528585757272, + "loss": 1.3441, + "theoretical_loss": 3.4918033527678145, + "tokens_seen": 1785659392 + }, + { + "epoch": 0.54, + "learning_rate": 0.0003709368104312939, + "loss": 1.3235, + "theoretical_loss": 3.491792517236108, + "tokens_seen": 1785724928 + }, + { + "epoch": 0.54, + "learning_rate": 0.0003709207622868606, + "loss": 1.2784, + "theoretical_loss": 3.4917816822133982, + "tokens_seen": 1785790464 + }, + { + "epoch": 0.54, + "objective/train/docs_used": 652000, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.5574352741241455, + "objective/train/theoretical_loss": 3.4917708476996436, + "objective/train/tokens_used": 1806316000, + "theoretical_loss": 3.4917708476996436, + "tokens_seen": 1785856000 + }, + { + "epoch": 0.54, + "learning_rate": 0.0003709047141424273, + "loss": 1.4739, + "theoretical_loss": 3.4917708476996436, + "tokens_seen": 1785856000 + }, + { + "epoch": 0.54, + "learning_rate": 0.000370888665997994, + "loss": 1.3607, + "theoretical_loss": 3.491760013694801, + "tokens_seen": 1785921536 + }, + { + "epoch": 0.54, + "learning_rate": 0.0003708726178535607, + "loss": 1.439, + "theoretical_loss": 3.491749180198828, + "tokens_seen": 1785987072 + }, + { + "epoch": 0.54, + "learning_rate": 0.0003708565697091274, + "loss": 1.3135, + "theoretical_loss": 3.4917383472116814, + "tokens_seen": 1786052608 + }, + { + "epoch": 0.54, + "learning_rate": 0.0003708405215646941, + "loss": 1.3404, + "theoretical_loss": 3.491727514733319, + "tokens_seen": 1786118144 + }, + { + "epoch": 0.54, + "learning_rate": 0.0003708244734202608, + "loss": 1.392, + "theoretical_loss": 3.4917166827636987, + "tokens_seen": 1786183680 + }, + { + "epoch": 0.54, + "learning_rate": 0.0003708084252758275, + "loss": 1.3048, + "theoretical_loss": 3.491705851302777, + "tokens_seen": 1786249216 + }, + { + "epoch": 0.54, + "learning_rate": 0.0003707923771313942, + "loss": 1.4122, + "theoretical_loss": 3.4916950203505124, + "tokens_seen": 1786314752 + }, + { + "epoch": 0.54, + "learning_rate": 0.0003707763289869609, + "loss": 1.2626, + "theoretical_loss": 3.491684189906862, + "tokens_seen": 1786380288 + }, + { + "epoch": 0.54, + "learning_rate": 0.0003707602808425276, + "loss": 1.4174, + "theoretical_loss": 3.4916733599717826, + "tokens_seen": 1786445824 + }, + { + "epoch": 0.54, + "learning_rate": 0.0003707442326980943, + "loss": 1.3597, + "theoretical_loss": 3.4916625305452325, + "tokens_seen": 1786511360 + }, + { + "epoch": 0.54, + "learning_rate": 0.000370728184553661, + "loss": 1.3937, + "theoretical_loss": 3.491651701627169, + "tokens_seen": 1786576896 + }, + { + "epoch": 0.54, + "learning_rate": 0.0003707121364092277, + "loss": 1.2999, + "theoretical_loss": 3.4916408732175492, + "tokens_seen": 1786642432 + }, + { + "epoch": 0.54, + "learning_rate": 0.0003706960882647944, + "loss": 1.2318, + "theoretical_loss": 3.491630045316331, + "tokens_seen": 1786707968 + }, + { + "epoch": 0.54, + "learning_rate": 0.0003706800401203611, + "loss": 1.4873, + "theoretical_loss": 3.4916192179234713, + "tokens_seen": 1786773504 + }, + { + "epoch": 0.54, + "learning_rate": 0.0003706639919759278, + "loss": 1.5105, + "theoretical_loss": 3.491608391038928, + "tokens_seen": 1786839040 + }, + { + "epoch": 0.54, + "learning_rate": 0.0003706479438314945, + "loss": 1.4312, + "theoretical_loss": 3.4915975646626594, + "tokens_seen": 1786904576 + }, + { + "epoch": 0.54, + "learning_rate": 0.0003706318956870612, + "loss": 1.3497, + "theoretical_loss": 3.491586738794621, + "tokens_seen": 1786970112 + }, + { + "epoch": 0.54, + "learning_rate": 0.0003706158475426279, + "loss": 1.372, + "theoretical_loss": 3.491575913434772, + "tokens_seen": 1787035648 + }, + { + "epoch": 0.54, + "learning_rate": 0.0003705997993981946, + "loss": 1.3081, + "theoretical_loss": 3.4915650885830694, + "tokens_seen": 1787101184 + }, + { + "epoch": 0.54, + "learning_rate": 0.0003705837512537613, + "loss": 1.4022, + "theoretical_loss": 3.4915542642394706, + "tokens_seen": 1787166720 + }, + { + "epoch": 0.54, + "learning_rate": 0.000370567703109328, + "loss": 1.3437, + "theoretical_loss": 3.4915434404039334, + "tokens_seen": 1787232256 + }, + { + "epoch": 0.54, + "learning_rate": 0.0003705516549648947, + "loss": 1.3221, + "theoretical_loss": 3.491532617076415, + "tokens_seen": 1787297792 + }, + { + "epoch": 0.54, + "learning_rate": 0.0003705356068204614, + "loss": 1.3858, + "theoretical_loss": 3.491521794256873, + "tokens_seen": 1787363328 + }, + { + "epoch": 0.54, + "learning_rate": 0.0003705195586760281, + "loss": 1.3764, + "theoretical_loss": 3.4915109719452646, + "tokens_seen": 1787428864 + }, + { + "epoch": 0.54, + "learning_rate": 0.0003705035105315948, + "loss": 1.1663, + "theoretical_loss": 3.4915001501415484, + "tokens_seen": 1787494400 + }, + { + "epoch": 0.54, + "learning_rate": 0.0003704874623871615, + "loss": 1.3316, + "theoretical_loss": 3.491489328845681, + "tokens_seen": 1787559936 + }, + { + "epoch": 0.54, + "learning_rate": 0.0003704714142427282, + "loss": 1.5149, + "theoretical_loss": 3.4914785080576207, + "tokens_seen": 1787625472 + }, + { + "epoch": 0.54, + "learning_rate": 0.0003704553660982949, + "loss": 1.3804, + "theoretical_loss": 3.491467687777324, + "tokens_seen": 1787691008 + }, + { + "epoch": 0.54, + "learning_rate": 0.0003704393179538616, + "loss": 1.3621, + "theoretical_loss": 3.4914568680047493, + "tokens_seen": 1787756544 + }, + { + "epoch": 0.54, + "learning_rate": 0.0003704232698094283, + "loss": 1.4149, + "theoretical_loss": 3.4914460487398538, + "tokens_seen": 1787822080 + }, + { + "epoch": 0.54, + "learning_rate": 0.000370407221664995, + "loss": 1.3638, + "theoretical_loss": 3.491435229982595, + "tokens_seen": 1787887616 + }, + { + "epoch": 0.54, + "learning_rate": 0.0003703911735205617, + "loss": 1.2556, + "theoretical_loss": 3.4914244117329307, + "tokens_seen": 1787953152 + }, + { + "epoch": 0.54, + "learning_rate": 0.0003703751253761284, + "loss": 1.3951, + "theoretical_loss": 3.491413593990819, + "tokens_seen": 1788018688 + }, + { + "epoch": 0.54, + "learning_rate": 0.0003703590772316951, + "loss": 1.278, + "theoretical_loss": 3.4914027767562166, + "tokens_seen": 1788084224 + }, + { + "epoch": 0.54, + "learning_rate": 0.0003703430290872618, + "loss": 1.3359, + "theoretical_loss": 3.4913919600290813, + "tokens_seen": 1788149760 + }, + { + "epoch": 0.54, + "learning_rate": 0.0003703269809428285, + "loss": 1.3594, + "theoretical_loss": 3.491381143809371, + "tokens_seen": 1788215296 + }, + { + "epoch": 0.54, + "learning_rate": 0.0003703109327983952, + "loss": 1.3914, + "theoretical_loss": 3.4913703280970427, + "tokens_seen": 1788280832 + }, + { + "epoch": 0.54, + "learning_rate": 0.0003702948846539619, + "loss": 1.3431, + "theoretical_loss": 3.4913595128920547, + "tokens_seen": 1788346368 + }, + { + "epoch": 0.54, + "learning_rate": 0.0003702788365095286, + "loss": 1.2983, + "theoretical_loss": 3.4913486981943644, + "tokens_seen": 1788411904 + }, + { + "epoch": 0.54, + "learning_rate": 0.0003702627883650953, + "loss": 1.3954, + "theoretical_loss": 3.4913378840039293, + "tokens_seen": 1788477440 + }, + { + "epoch": 0.54, + "learning_rate": 0.000370246740220662, + "loss": 1.3112, + "theoretical_loss": 3.491327070320707, + "tokens_seen": 1788542976 + }, + { + "epoch": 0.54, + "learning_rate": 0.0003702306920762287, + "loss": 1.2869, + "theoretical_loss": 3.491316257144655, + "tokens_seen": 1788608512 + }, + { + "epoch": 0.54, + "learning_rate": 0.0003702146439317954, + "loss": 1.3185, + "theoretical_loss": 3.4913054444757314, + "tokens_seen": 1788674048 + }, + { + "epoch": 0.54, + "learning_rate": 0.0003701985957873621, + "loss": 1.3638, + "theoretical_loss": 3.4912946323138936, + "tokens_seen": 1788739584 + }, + { + "epoch": 0.54, + "learning_rate": 0.0003701825476429288, + "loss": 1.363, + "theoretical_loss": 3.491283820659099, + "tokens_seen": 1788805120 + }, + { + "epoch": 0.54, + "learning_rate": 0.0003701664994984955, + "loss": 1.3045, + "theoretical_loss": 3.491273009511306, + "tokens_seen": 1788870656 + }, + { + "epoch": 0.54, + "learning_rate": 0.0003701504513540622, + "loss": 1.3337, + "theoretical_loss": 3.491262198870471, + "tokens_seen": 1788936192 + }, + { + "epoch": 0.54, + "learning_rate": 0.0003701344032096289, + "loss": 1.4119, + "theoretical_loss": 3.491251388736553, + "tokens_seen": 1789001728 + }, + { + "epoch": 0.54, + "learning_rate": 0.0003701183550651956, + "loss": 1.328, + "theoretical_loss": 3.4912405791095082, + "tokens_seen": 1789067264 + }, + { + "epoch": 0.54, + "objective/train/docs_used": 653179, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.4469397068023682, + "objective/train/theoretical_loss": 3.491229769989296, + "objective/train/tokens_used": 1809592800, + "theoretical_loss": 3.491229769989296, + "tokens_seen": 1789132800 + }, + { + "epoch": 0.54, + "learning_rate": 0.0003701023069207623, + "loss": 1.3264, + "theoretical_loss": 3.491229769989296, + "tokens_seen": 1789132800 + }, + { + "epoch": 0.54, + "learning_rate": 0.000370086258776329, + "loss": 1.3513, + "theoretical_loss": 3.4912189613758726, + "tokens_seen": 1789198336 + }, + { + "epoch": 0.54, + "learning_rate": 0.0003700702106318957, + "loss": 1.3018, + "theoretical_loss": 3.4912081532691963, + "tokens_seen": 1789263872 + }, + { + "epoch": 0.54, + "learning_rate": 0.00037005416248746243, + "loss": 1.379, + "theoretical_loss": 3.491197345669225, + "tokens_seen": 1789329408 + }, + { + "epoch": 0.54, + "learning_rate": 0.00037003811434302913, + "loss": 1.3479, + "theoretical_loss": 3.4911865385759158, + "tokens_seen": 1789394944 + }, + { + "epoch": 0.54, + "learning_rate": 0.00037002206619859583, + "loss": 1.3833, + "theoretical_loss": 3.4911757319892267, + "tokens_seen": 1789460480 + }, + { + "epoch": 0.54, + "learning_rate": 0.00037000601805416253, + "loss": 1.259, + "theoretical_loss": 3.4911649259091155, + "tokens_seen": 1789526016 + }, + { + "epoch": 0.54, + "learning_rate": 0.00036998996990972923, + "loss": 1.4167, + "theoretical_loss": 3.4911541203355396, + "tokens_seen": 1789591552 + }, + { + "epoch": 0.54, + "learning_rate": 0.00036997392176529593, + "loss": 1.3233, + "theoretical_loss": 3.4911433152684572, + "tokens_seen": 1789657088 + }, + { + "epoch": 0.54, + "learning_rate": 0.00036995787362086263, + "loss": 1.4712, + "theoretical_loss": 3.4911325107078257, + "tokens_seen": 1789722624 + }, + { + "epoch": 0.54, + "learning_rate": 0.0003699418254764293, + "loss": 1.3853, + "theoretical_loss": 3.491121706653603, + "tokens_seen": 1789788160 + }, + { + "epoch": 0.54, + "learning_rate": 0.000369925777331996, + "loss": 1.3761, + "theoretical_loss": 3.4911109031057457, + "tokens_seen": 1789853696 + }, + { + "epoch": 0.54, + "learning_rate": 0.0003699097291875627, + "loss": 1.4788, + "theoretical_loss": 3.4911001000642132, + "tokens_seen": 1789919232 + }, + { + "epoch": 0.54, + "learning_rate": 0.0003698936810431294, + "loss": 1.2964, + "theoretical_loss": 3.491089297528963, + "tokens_seen": 1789984768 + }, + { + "epoch": 0.54, + "learning_rate": 0.0003698776328986961, + "loss": 1.4116, + "theoretical_loss": 3.4910784954999516, + "tokens_seen": 1790050304 + }, + { + "epoch": 0.54, + "learning_rate": 0.0003698615847542628, + "loss": 1.2578, + "theoretical_loss": 3.491067693977138, + "tokens_seen": 1790115840 + }, + { + "epoch": 0.54, + "learning_rate": 0.0003698455366098295, + "loss": 1.3791, + "theoretical_loss": 3.4910568929604793, + "tokens_seen": 1790181376 + }, + { + "epoch": 0.54, + "learning_rate": 0.0003698294884653962, + "loss": 1.4187, + "theoretical_loss": 3.491046092449933, + "tokens_seen": 1790246912 + }, + { + "epoch": 0.54, + "learning_rate": 0.0003698134403209629, + "loss": 1.2172, + "theoretical_loss": 3.4910352924454577, + "tokens_seen": 1790312448 + }, + { + "epoch": 0.54, + "learning_rate": 0.0003697973921765296, + "loss": 1.2457, + "theoretical_loss": 3.4910244929470107, + "tokens_seen": 1790377984 + }, + { + "epoch": 0.54, + "learning_rate": 0.0003697813440320963, + "loss": 1.438, + "theoretical_loss": 3.4910136939545495, + "tokens_seen": 1790443520 + }, + { + "epoch": 0.54, + "learning_rate": 0.000369765295887663, + "loss": 1.3129, + "theoretical_loss": 3.4910028954680326, + "tokens_seen": 1790509056 + }, + { + "epoch": 0.54, + "learning_rate": 0.0003697492477432297, + "loss": 1.3758, + "theoretical_loss": 3.490992097487417, + "tokens_seen": 1790574592 + }, + { + "epoch": 0.54, + "learning_rate": 0.0003697331995987964, + "loss": 1.2745, + "theoretical_loss": 3.490981300012661, + "tokens_seen": 1790640128 + }, + { + "epoch": 0.54, + "learning_rate": 0.0003697171514543631, + "loss": 1.1911, + "theoretical_loss": 3.490970503043722, + "tokens_seen": 1790705664 + }, + { + "epoch": 0.54, + "learning_rate": 0.0003697011033099298, + "loss": 1.3488, + "theoretical_loss": 3.4909597065805587, + "tokens_seen": 1790771200 + }, + { + "epoch": 0.54, + "learning_rate": 0.0003696850551654965, + "loss": 1.4224, + "theoretical_loss": 3.4909489106231275, + "tokens_seen": 1790836736 + }, + { + "epoch": 0.54, + "learning_rate": 0.0003696690070210632, + "loss": 1.3019, + "theoretical_loss": 3.490938115171387, + "tokens_seen": 1790902272 + }, + { + "epoch": 0.54, + "learning_rate": 0.0003696529588766299, + "loss": 1.3125, + "theoretical_loss": 3.4909273202252953, + "tokens_seen": 1790967808 + }, + { + "epoch": 0.54, + "learning_rate": 0.0003696369107321966, + "loss": 1.4496, + "theoretical_loss": 3.49091652578481, + "tokens_seen": 1791033344 + }, + { + "epoch": 0.54, + "learning_rate": 0.0003696208625877633, + "loss": 1.3408, + "theoretical_loss": 3.490905731849888, + "tokens_seen": 1791098880 + }, + { + "epoch": 0.54, + "learning_rate": 0.00036960481444333, + "loss": 1.3152, + "theoretical_loss": 3.4908949384204884, + "tokens_seen": 1791164416 + }, + { + "epoch": 0.54, + "learning_rate": 0.0003695887662988967, + "loss": 1.2946, + "theoretical_loss": 3.490884145496568, + "tokens_seen": 1791229952 + }, + { + "epoch": 0.54, + "learning_rate": 0.0003695727181544634, + "loss": 1.4392, + "theoretical_loss": 3.4908733530780856, + "tokens_seen": 1791295488 + }, + { + "epoch": 0.54, + "learning_rate": 0.0003695566700100301, + "loss": 1.327, + "theoretical_loss": 3.490862561164999, + "tokens_seen": 1791361024 + }, + { + "epoch": 0.54, + "learning_rate": 0.0003695406218655968, + "loss": 1.428, + "theoretical_loss": 3.490851769757265, + "tokens_seen": 1791426560 + }, + { + "epoch": 0.54, + "learning_rate": 0.0003695245737211635, + "loss": 1.3229, + "theoretical_loss": 3.490840978854842, + "tokens_seen": 1791492096 + }, + { + "epoch": 0.54, + "learning_rate": 0.0003695085255767302, + "loss": 1.347, + "theoretical_loss": 3.4908301884576884, + "tokens_seen": 1791557632 + }, + { + "epoch": 0.54, + "learning_rate": 0.0003694924774322969, + "loss": 1.3253, + "theoretical_loss": 3.4908193985657614, + "tokens_seen": 1791623168 + }, + { + "epoch": 0.54, + "learning_rate": 0.0003694764292878636, + "loss": 1.1583, + "theoretical_loss": 3.490808609179019, + "tokens_seen": 1791688704 + }, + { + "epoch": 0.54, + "learning_rate": 0.0003694603811434303, + "loss": 1.3554, + "theoretical_loss": 3.490797820297419, + "tokens_seen": 1791754240 + }, + { + "epoch": 0.54, + "learning_rate": 0.000369444332998997, + "loss": 1.3202, + "theoretical_loss": 3.4907870319209198, + "tokens_seen": 1791819776 + }, + { + "epoch": 0.54, + "learning_rate": 0.0003694282848545637, + "loss": 1.3408, + "theoretical_loss": 3.4907762440494787, + "tokens_seen": 1791885312 + }, + { + "epoch": 0.54, + "learning_rate": 0.0003694122367101304, + "loss": 1.4077, + "theoretical_loss": 3.4907654566830537, + "tokens_seen": 1791950848 + }, + { + "epoch": 0.54, + "learning_rate": 0.0003693961885656971, + "loss": 1.282, + "theoretical_loss": 3.4907546698216025, + "tokens_seen": 1792016384 + }, + { + "epoch": 0.54, + "learning_rate": 0.0003693801404212638, + "loss": 1.3133, + "theoretical_loss": 3.4907438834650844, + "tokens_seen": 1792081920 + }, + { + "epoch": 0.54, + "learning_rate": 0.00036936409227683055, + "loss": 1.364, + "theoretical_loss": 3.4907330976134547, + "tokens_seen": 1792147456 + }, + { + "epoch": 0.54, + "learning_rate": 0.00036934804413239726, + "loss": 1.2876, + "theoretical_loss": 3.4907223122666737, + "tokens_seen": 1792212992 + }, + { + "epoch": 0.54, + "learning_rate": 0.00036933199598796396, + "loss": 1.3739, + "theoretical_loss": 3.490711527424698, + "tokens_seen": 1792278528 + }, + { + "epoch": 0.54, + "learning_rate": 0.00036931594784353066, + "loss": 1.319, + "theoretical_loss": 3.490700743087486, + "tokens_seen": 1792344064 + }, + { + "epoch": 0.54, + "objective/train/docs_used": 654198, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.1911827325820923, + "objective/train/theoretical_loss": 3.4906899592549956, + "objective/train/tokens_used": 1812869600, + "theoretical_loss": 3.4906899592549956, + "tokens_seen": 1792409600 + }, + { + "epoch": 0.54, + "learning_rate": 0.0003692998996990973, + "loss": 1.2108, + "theoretical_loss": 3.4906899592549956, + "tokens_seen": 1792409600 + }, + { + "epoch": 0.54, + "learning_rate": 0.000369283851554664, + "loss": 1.3153, + "theoretical_loss": 3.4906791759271845, + "tokens_seen": 1792475136 + }, + { + "epoch": 0.54, + "learning_rate": 0.0003692678034102307, + "loss": 1.3045, + "theoretical_loss": 3.490668393104011, + "tokens_seen": 1792540672 + }, + { + "epoch": 0.54, + "learning_rate": 0.0003692517552657974, + "loss": 1.2795, + "theoretical_loss": 3.4906576107854326, + "tokens_seen": 1792606208 + }, + { + "epoch": 0.54, + "learning_rate": 0.0003692357071213641, + "loss": 1.4128, + "theoretical_loss": 3.490646828971407, + "tokens_seen": 1792671744 + }, + { + "epoch": 0.54, + "learning_rate": 0.0003692196589769308, + "loss": 1.4097, + "theoretical_loss": 3.4906360476618934, + "tokens_seen": 1792737280 + }, + { + "epoch": 0.54, + "learning_rate": 0.0003692036108324975, + "loss": 1.308, + "theoretical_loss": 3.4906252668568483, + "tokens_seen": 1792802816 + }, + { + "epoch": 0.54, + "learning_rate": 0.0003691875626880642, + "loss": 1.249, + "theoretical_loss": 3.4906144865562307, + "tokens_seen": 1792868352 + }, + { + "epoch": 0.54, + "learning_rate": 0.0003691715145436309, + "loss": 1.3953, + "theoretical_loss": 3.490603706759998, + "tokens_seen": 1792933888 + }, + { + "epoch": 0.54, + "learning_rate": 0.0003691554663991976, + "loss": 1.3603, + "theoretical_loss": 3.490592927468108, + "tokens_seen": 1792999424 + }, + { + "epoch": 0.54, + "learning_rate": 0.0003691394182547643, + "loss": 1.3255, + "theoretical_loss": 3.4905821486805197, + "tokens_seen": 1793064960 + }, + { + "epoch": 0.54, + "learning_rate": 0.000369123370110331, + "loss": 1.2931, + "theoretical_loss": 3.49057137039719, + "tokens_seen": 1793130496 + }, + { + "epoch": 0.54, + "learning_rate": 0.0003691073219658977, + "loss": 1.3074, + "theoretical_loss": 3.490560592618077, + "tokens_seen": 1793196032 + }, + { + "epoch": 0.54, + "learning_rate": 0.0003690912738214644, + "loss": 1.1985, + "theoretical_loss": 3.4905498153431394, + "tokens_seen": 1793261568 + }, + { + "epoch": 0.54, + "learning_rate": 0.0003690752256770311, + "loss": 1.3273, + "theoretical_loss": 3.4905390385723347, + "tokens_seen": 1793327104 + }, + { + "epoch": 0.54, + "learning_rate": 0.0003690591775325978, + "loss": 1.3384, + "theoretical_loss": 3.49052826230562, + "tokens_seen": 1793392640 + }, + { + "epoch": 0.54, + "learning_rate": 0.0003690431293881645, + "loss": 1.4225, + "theoretical_loss": 3.490517486542956, + "tokens_seen": 1793458176 + }, + { + "epoch": 0.54, + "learning_rate": 0.0003690270812437312, + "loss": 1.2845, + "theoretical_loss": 3.490506711284297, + "tokens_seen": 1793523712 + }, + { + "epoch": 0.54, + "learning_rate": 0.0003690110330992979, + "loss": 1.1193, + "theoretical_loss": 3.490495936529604, + "tokens_seen": 1793589248 + }, + { + "epoch": 0.54, + "learning_rate": 0.0003689949849548646, + "loss": 1.2351, + "theoretical_loss": 3.490485162278834, + "tokens_seen": 1793654784 + }, + { + "epoch": 0.54, + "learning_rate": 0.0003689789368104313, + "loss": 1.3129, + "theoretical_loss": 3.490474388531945, + "tokens_seen": 1793720320 + }, + { + "epoch": 0.54, + "learning_rate": 0.000368962888665998, + "loss": 1.3996, + "theoretical_loss": 3.4904636152888946, + "tokens_seen": 1793785856 + }, + { + "epoch": 0.54, + "learning_rate": 0.0003689468405215647, + "loss": 1.2434, + "theoretical_loss": 3.4904528425496415, + "tokens_seen": 1793851392 + }, + { + "epoch": 0.54, + "learning_rate": 0.0003689307923771314, + "loss": 1.3906, + "theoretical_loss": 3.490442070314143, + "tokens_seen": 1793916928 + }, + { + "epoch": 0.54, + "learning_rate": 0.0003689147442326981, + "loss": 1.4425, + "theoretical_loss": 3.490431298582358, + "tokens_seen": 1793982464 + }, + { + "epoch": 0.54, + "learning_rate": 0.0003688986960882648, + "loss": 1.2974, + "theoretical_loss": 3.490420527354244, + "tokens_seen": 1794048000 + }, + { + "epoch": 0.54, + "learning_rate": 0.0003688826479438315, + "loss": 1.3127, + "theoretical_loss": 3.49040975662976, + "tokens_seen": 1794113536 + }, + { + "epoch": 0.54, + "learning_rate": 0.0003688665997993982, + "loss": 1.2991, + "theoretical_loss": 3.4903989864088625, + "tokens_seen": 1794179072 + }, + { + "epoch": 0.54, + "learning_rate": 0.0003688505516549649, + "loss": 1.3653, + "theoretical_loss": 3.4903882166915103, + "tokens_seen": 1794244608 + }, + { + "epoch": 0.54, + "learning_rate": 0.0003688345035105316, + "loss": 1.4447, + "theoretical_loss": 3.4903774474776617, + "tokens_seen": 1794310144 + }, + { + "epoch": 0.54, + "learning_rate": 0.0003688184553660983, + "loss": 1.3283, + "theoretical_loss": 3.4903666787672747, + "tokens_seen": 1794375680 + }, + { + "epoch": 0.54, + "learning_rate": 0.000368802407221665, + "loss": 1.2786, + "theoretical_loss": 3.4903559105603073, + "tokens_seen": 1794441216 + }, + { + "epoch": 0.54, + "learning_rate": 0.0003687863590772317, + "loss": 1.3262, + "theoretical_loss": 3.490345142856717, + "tokens_seen": 1794506752 + }, + { + "epoch": 0.54, + "learning_rate": 0.0003687703109327984, + "loss": 1.4731, + "theoretical_loss": 3.4903343756564627, + "tokens_seen": 1794572288 + }, + { + "epoch": 0.54, + "learning_rate": 0.0003687542627883651, + "loss": 1.3669, + "theoretical_loss": 3.4903236089595024, + "tokens_seen": 1794637824 + }, + { + "epoch": 0.54, + "learning_rate": 0.0003687382146439318, + "loss": 1.2537, + "theoretical_loss": 3.4903128427657935, + "tokens_seen": 1794703360 + }, + { + "epoch": 0.54, + "learning_rate": 0.0003687221664994985, + "loss": 1.3971, + "theoretical_loss": 3.4903020770752953, + "tokens_seen": 1794768896 + }, + { + "epoch": 0.54, + "learning_rate": 0.0003687061183550652, + "loss": 1.3603, + "theoretical_loss": 3.4902913118879644, + "tokens_seen": 1794834432 + }, + { + "epoch": 0.54, + "learning_rate": 0.00036869007021063193, + "loss": 1.2982, + "theoretical_loss": 3.4902805472037604, + "tokens_seen": 1794899968 + }, + { + "epoch": 0.54, + "learning_rate": 0.00036867402206619863, + "loss": 1.3431, + "theoretical_loss": 3.4902697830226406, + "tokens_seen": 1794965504 + }, + { + "epoch": 0.54, + "learning_rate": 0.00036865797392176533, + "loss": 1.3773, + "theoretical_loss": 3.490259019344563, + "tokens_seen": 1795031040 + }, + { + "epoch": 0.54, + "learning_rate": 0.00036864192577733203, + "loss": 1.3446, + "theoretical_loss": 3.490248256169486, + "tokens_seen": 1795096576 + }, + { + "epoch": 0.54, + "learning_rate": 0.00036862587763289873, + "loss": 1.3234, + "theoretical_loss": 3.490237493497368, + "tokens_seen": 1795162112 + }, + { + "epoch": 0.54, + "learning_rate": 0.00036860982948846543, + "loss": 1.3824, + "theoretical_loss": 3.4902267313281667, + "tokens_seen": 1795227648 + }, + { + "epoch": 0.54, + "learning_rate": 0.00036859378134403213, + "loss": 1.4168, + "theoretical_loss": 3.4902159696618402, + "tokens_seen": 1795293184 + }, + { + "epoch": 0.54, + "learning_rate": 0.00036857773319959883, + "loss": 1.3788, + "theoretical_loss": 3.490205208498347, + "tokens_seen": 1795358720 + }, + { + "epoch": 0.54, + "learning_rate": 0.00036856168505516553, + "loss": 1.3216, + "theoretical_loss": 3.490194447837645, + "tokens_seen": 1795424256 + }, + { + "epoch": 0.54, + "learning_rate": 0.00036854563691073223, + "loss": 1.4128, + "theoretical_loss": 3.490183687679693, + "tokens_seen": 1795489792 + }, + { + "epoch": 0.54, + "learning_rate": 0.0003685295887662989, + "loss": 1.3224, + "theoretical_loss": 3.490172928024448, + "tokens_seen": 1795555328 + }, + { + "epoch": 0.54, + "learning_rate": 0.0003685135406218656, + "loss": 1.4395, + "theoretical_loss": 3.490162168871869, + "tokens_seen": 1795620864 + }, + { + "epoch": 0.54, + "objective/train/docs_used": 655419, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.3811092376708984, + "objective/train/theoretical_loss": 3.4901514102219138, + "objective/train/tokens_used": 1816146400, + "theoretical_loss": 3.4901514102219138, + "tokens_seen": 1795686400 + }, + { + "epoch": 0.54, + "learning_rate": 0.0003684974924774323, + "loss": 1.3184, + "theoretical_loss": 3.4901514102219138, + "tokens_seen": 1795686400 + }, + { + "epoch": 0.54, + "learning_rate": 0.000368481444332999, + "loss": 1.3274, + "theoretical_loss": 3.490140652074541, + "tokens_seen": 1795751936 + }, + { + "epoch": 0.54, + "learning_rate": 0.0003684653961885657, + "loss": 1.3271, + "theoretical_loss": 3.4901298944297086, + "tokens_seen": 1795817472 + }, + { + "epoch": 0.54, + "learning_rate": 0.0003684493480441324, + "loss": 1.294, + "theoretical_loss": 3.4901191372873743, + "tokens_seen": 1795883008 + }, + { + "epoch": 0.54, + "learning_rate": 0.0003684332998996991, + "loss": 1.4404, + "theoretical_loss": 3.490108380647497, + "tokens_seen": 1795948544 + }, + { + "epoch": 0.54, + "learning_rate": 0.0003684172517552658, + "loss": 1.4558, + "theoretical_loss": 3.4900976245100344, + "tokens_seen": 1796014080 + }, + { + "epoch": 0.54, + "learning_rate": 0.0003684012036108325, + "loss": 1.3347, + "theoretical_loss": 3.4900868688749442, + "tokens_seen": 1796079616 + }, + { + "epoch": 0.54, + "learning_rate": 0.0003683851554663992, + "loss": 1.3013, + "theoretical_loss": 3.4900761137421865, + "tokens_seen": 1796145152 + }, + { + "epoch": 0.54, + "learning_rate": 0.0003683691073219659, + "loss": 1.379, + "theoretical_loss": 3.4900653591117177, + "tokens_seen": 1796210688 + }, + { + "epoch": 0.54, + "learning_rate": 0.0003683530591775326, + "loss": 1.3936, + "theoretical_loss": 3.490054604983497, + "tokens_seen": 1796276224 + }, + { + "epoch": 0.54, + "learning_rate": 0.0003683370110330993, + "loss": 1.2265, + "theoretical_loss": 3.4900438513574814, + "tokens_seen": 1796341760 + }, + { + "epoch": 0.54, + "learning_rate": 0.000368320962888666, + "loss": 1.5006, + "theoretical_loss": 3.490033098233631, + "tokens_seen": 1796407296 + }, + { + "epoch": 0.54, + "learning_rate": 0.0003683049147442327, + "loss": 1.4218, + "theoretical_loss": 3.4900223456119024, + "tokens_seen": 1796472832 + }, + { + "epoch": 0.54, + "learning_rate": 0.0003682888665997994, + "loss": 1.3514, + "theoretical_loss": 3.4900115934922544, + "tokens_seen": 1796538368 + }, + { + "epoch": 0.54, + "learning_rate": 0.0003682728184553661, + "loss": 1.545, + "theoretical_loss": 3.4900008418746458, + "tokens_seen": 1796603904 + }, + { + "epoch": 0.54, + "learning_rate": 0.0003682567703109328, + "loss": 1.3487, + "theoretical_loss": 3.4899900907590338, + "tokens_seen": 1796669440 + }, + { + "epoch": 0.54, + "learning_rate": 0.0003682407221664995, + "loss": 1.3188, + "theoretical_loss": 3.4899793401453776, + "tokens_seen": 1796734976 + }, + { + "epoch": 0.54, + "learning_rate": 0.0003682246740220662, + "loss": 1.428, + "theoretical_loss": 3.4899685900336346, + "tokens_seen": 1796800512 + }, + { + "epoch": 0.54, + "learning_rate": 0.0003682086258776329, + "loss": 1.3618, + "theoretical_loss": 3.489957840423764, + "tokens_seen": 1796866048 + }, + { + "epoch": 0.54, + "learning_rate": 0.0003681925777331996, + "loss": 1.2462, + "theoretical_loss": 3.489947091315723, + "tokens_seen": 1796931584 + }, + { + "epoch": 0.54, + "learning_rate": 0.0003681765295887663, + "loss": 1.3913, + "theoretical_loss": 3.4899363427094707, + "tokens_seen": 1796997120 + }, + { + "epoch": 0.54, + "learning_rate": 0.00036816048144433305, + "loss": 1.2215, + "theoretical_loss": 3.489925594604965, + "tokens_seen": 1797062656 + }, + { + "epoch": 0.54, + "learning_rate": 0.00036814443329989975, + "loss": 1.432, + "theoretical_loss": 3.4899148470021646, + "tokens_seen": 1797128192 + }, + { + "epoch": 0.54, + "learning_rate": 0.00036812838515546645, + "loss": 1.2961, + "theoretical_loss": 3.489904099901027, + "tokens_seen": 1797193728 + }, + { + "epoch": 0.54, + "learning_rate": 0.00036811233701103315, + "loss": 1.2787, + "theoretical_loss": 3.4898933533015115, + "tokens_seen": 1797259264 + }, + { + "epoch": 0.54, + "learning_rate": 0.00036809628886659985, + "loss": 1.2807, + "theoretical_loss": 3.489882607203576, + "tokens_seen": 1797324800 + }, + { + "epoch": 0.54, + "learning_rate": 0.00036808024072216655, + "loss": 1.3259, + "theoretical_loss": 3.489871861607178, + "tokens_seen": 1797390336 + }, + { + "epoch": 0.54, + "learning_rate": 0.00036806419257773325, + "loss": 1.2927, + "theoretical_loss": 3.4898611165122766, + "tokens_seen": 1797455872 + }, + { + "epoch": 0.54, + "learning_rate": 0.00036804814443329995, + "loss": 1.4338, + "theoretical_loss": 3.4898503719188305, + "tokens_seen": 1797521408 + }, + { + "epoch": 0.54, + "learning_rate": 0.00036803209628886665, + "loss": 1.4626, + "theoretical_loss": 3.489839627826797, + "tokens_seen": 1797586944 + }, + { + "epoch": 0.54, + "learning_rate": 0.00036801604814443335, + "loss": 1.3465, + "theoretical_loss": 3.489828884236135, + "tokens_seen": 1797652480 + }, + { + "epoch": 0.54, + "learning_rate": 0.00036800000000000005, + "loss": 1.3393, + "theoretical_loss": 3.489818141146803, + "tokens_seen": 1797718016 + }, + { + "epoch": 0.54, + "learning_rate": 0.00036798395185556676, + "loss": 1.3588, + "theoretical_loss": 3.489807398558759, + "tokens_seen": 1797783552 + }, + { + "epoch": 0.54, + "learning_rate": 0.00036796790371113346, + "loss": 1.2236, + "theoretical_loss": 3.4897966564719614, + "tokens_seen": 1797849088 + }, + { + "epoch": 0.54, + "learning_rate": 0.00036795185556670016, + "loss": 1.3965, + "theoretical_loss": 3.489785914886369, + "tokens_seen": 1797914624 + }, + { + "epoch": 0.54, + "learning_rate": 0.00036793580742226686, + "loss": 1.3264, + "theoretical_loss": 3.489775173801939, + "tokens_seen": 1797980160 + }, + { + "epoch": 0.54, + "learning_rate": 0.00036791975927783356, + "loss": 1.2841, + "theoretical_loss": 3.489764433218631, + "tokens_seen": 1798045696 + }, + { + "epoch": 0.54, + "learning_rate": 0.0003679037111334002, + "loss": 1.3328, + "theoretical_loss": 3.489753693136402, + "tokens_seen": 1798111232 + }, + { + "epoch": 0.54, + "learning_rate": 0.0003678876629889669, + "loss": 1.2709, + "theoretical_loss": 3.4897429535552122, + "tokens_seen": 1798176768 + }, + { + "epoch": 0.54, + "learning_rate": 0.0003678716148445336, + "loss": 1.3864, + "theoretical_loss": 3.4897322144750182, + "tokens_seen": 1798242304 + }, + { + "epoch": 0.54, + "learning_rate": 0.0003678555667001003, + "loss": 1.3529, + "theoretical_loss": 3.4897214758957795, + "tokens_seen": 1798307840 + }, + { + "epoch": 0.54, + "learning_rate": 0.000367839518555667, + "loss": 1.4095, + "theoretical_loss": 3.489710737817454, + "tokens_seen": 1798373376 + }, + { + "epoch": 0.54, + "learning_rate": 0.0003678234704112337, + "loss": 1.3466, + "theoretical_loss": 3.48970000024, + "tokens_seen": 1798438912 + }, + { + "epoch": 0.55, + "learning_rate": 0.0003678074222668004, + "loss": 1.4123, + "theoretical_loss": 3.4896892631633762, + "tokens_seen": 1798504448 + }, + { + "epoch": 0.55, + "learning_rate": 0.0003677913741223671, + "loss": 1.3455, + "theoretical_loss": 3.489678526587541, + "tokens_seen": 1798569984 + }, + { + "epoch": 0.55, + "learning_rate": 0.0003677753259779338, + "loss": 1.2666, + "theoretical_loss": 3.4896677905124527, + "tokens_seen": 1798635520 + }, + { + "epoch": 0.55, + "learning_rate": 0.0003677592778335005, + "loss": 1.4008, + "theoretical_loss": 3.4896570549380694, + "tokens_seen": 1798701056 + }, + { + "epoch": 0.55, + "learning_rate": 0.0003677432296890672, + "loss": 1.3595, + "theoretical_loss": 3.4896463198643497, + "tokens_seen": 1798766592 + }, + { + "epoch": 0.55, + "learning_rate": 0.0003677271815446339, + "loss": 1.5136, + "theoretical_loss": 3.4896355852912526, + "tokens_seen": 1798832128 + }, + { + "epoch": 0.55, + "learning_rate": 0.0003677111334002006, + "loss": 1.3767, + "theoretical_loss": 3.4896248512187356, + "tokens_seen": 1798897664 + }, + { + "epoch": 0.55, + "objective/train/docs_used": 656735, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.3373134136199951, + "objective/train/theoretical_loss": 3.4896141176467577, + "objective/train/tokens_used": 1819423200, + "theoretical_loss": 3.4896141176467577, + "tokens_seen": 1798963200 + }, + { + "epoch": 0.55, + "learning_rate": 0.0003676950852557673, + "loss": 1.2964, + "theoretical_loss": 3.4896141176467577, + "tokens_seen": 1798963200 + }, + { + "epoch": 0.55, + "learning_rate": 0.000367679037111334, + "loss": 1.2829, + "theoretical_loss": 3.4896033845752763, + "tokens_seen": 1799028736 + }, + { + "epoch": 0.55, + "learning_rate": 0.0003676629889669007, + "loss": 1.2733, + "theoretical_loss": 3.4895926520042515, + "tokens_seen": 1799094272 + }, + { + "epoch": 0.55, + "learning_rate": 0.0003676469408224674, + "loss": 1.4025, + "theoretical_loss": 3.4895819199336406, + "tokens_seen": 1799159808 + }, + { + "epoch": 0.55, + "learning_rate": 0.0003676308926780341, + "loss": 1.3494, + "theoretical_loss": 3.489571188363403, + "tokens_seen": 1799225344 + }, + { + "epoch": 0.55, + "learning_rate": 0.0003676148445336008, + "loss": 1.2366, + "theoretical_loss": 3.4895604572934955, + "tokens_seen": 1799290880 + }, + { + "epoch": 0.55, + "learning_rate": 0.0003675987963891675, + "loss": 1.5152, + "theoretical_loss": 3.489549726723878, + "tokens_seen": 1799356416 + }, + { + "epoch": 0.55, + "learning_rate": 0.0003675827482447342, + "loss": 1.3482, + "theoretical_loss": 3.4895389966545087, + "tokens_seen": 1799421952 + }, + { + "epoch": 0.55, + "learning_rate": 0.0003675667001003009, + "loss": 1.305, + "theoretical_loss": 3.489528267085346, + "tokens_seen": 1799487488 + }, + { + "epoch": 0.55, + "learning_rate": 0.0003675506519558676, + "loss": 1.3872, + "theoretical_loss": 3.4895175380163472, + "tokens_seen": 1799553024 + }, + { + "epoch": 0.55, + "learning_rate": 0.0003675346038114343, + "loss": 1.4722, + "theoretical_loss": 3.4895068094474726, + "tokens_seen": 1799618560 + }, + { + "epoch": 0.55, + "learning_rate": 0.000367518555667001, + "loss": 1.367, + "theoretical_loss": 3.4894960813786797, + "tokens_seen": 1799684096 + }, + { + "epoch": 0.55, + "learning_rate": 0.0003675025075225677, + "loss": 1.3561, + "theoretical_loss": 3.4894853538099273, + "tokens_seen": 1799749632 + }, + { + "epoch": 0.55, + "learning_rate": 0.0003674864593781344, + "loss": 1.3345, + "theoretical_loss": 3.489474626741174, + "tokens_seen": 1799815168 + }, + { + "epoch": 0.55, + "learning_rate": 0.0003674704112337011, + "loss": 1.3365, + "theoretical_loss": 3.4894639001723773, + "tokens_seen": 1799880704 + }, + { + "epoch": 0.55, + "learning_rate": 0.0003674543630892678, + "loss": 1.4865, + "theoretical_loss": 3.4894531741034966, + "tokens_seen": 1799946240 + }, + { + "epoch": 0.55, + "learning_rate": 0.0003674383149448345, + "loss": 1.314, + "theoretical_loss": 3.489442448534491, + "tokens_seen": 1800011776 + }, + { + "epoch": 0.55, + "learning_rate": 0.0003674222668004012, + "loss": 1.3156, + "theoretical_loss": 3.489431723465317, + "tokens_seen": 1800077312 + }, + { + "epoch": 0.55, + "learning_rate": 0.0003674062186559679, + "loss": 1.2958, + "theoretical_loss": 3.489420998895935, + "tokens_seen": 1800142848 + }, + { + "epoch": 0.55, + "learning_rate": 0.0003673901705115346, + "loss": 1.3793, + "theoretical_loss": 3.489410274826303, + "tokens_seen": 1800208384 + }, + { + "epoch": 0.55, + "learning_rate": 0.0003673741223671013, + "loss": 1.3037, + "theoretical_loss": 3.489399551256379, + "tokens_seen": 1800273920 + }, + { + "epoch": 0.55, + "learning_rate": 0.000367358074222668, + "loss": 1.4853, + "theoretical_loss": 3.4893888281861223, + "tokens_seen": 1800339456 + }, + { + "epoch": 0.55, + "learning_rate": 0.0003673420260782347, + "loss": 1.3733, + "theoretical_loss": 3.489378105615491, + "tokens_seen": 1800404992 + }, + { + "epoch": 0.55, + "learning_rate": 0.0003673259779338014, + "loss": 1.265, + "theoretical_loss": 3.489367383544443, + "tokens_seen": 1800470528 + }, + { + "epoch": 0.55, + "learning_rate": 0.00036730992978936813, + "loss": 1.3348, + "theoretical_loss": 3.489356661972938, + "tokens_seen": 1800536064 + }, + { + "epoch": 0.55, + "learning_rate": 0.00036729388164493483, + "loss": 1.3444, + "theoretical_loss": 3.489345940900934, + "tokens_seen": 1800601600 + }, + { + "epoch": 0.55, + "learning_rate": 0.00036727783350050153, + "loss": 1.3584, + "theoretical_loss": 3.48933522032839, + "tokens_seen": 1800667136 + }, + { + "epoch": 0.55, + "learning_rate": 0.00036726178535606823, + "loss": 1.4682, + "theoretical_loss": 3.4893245002552638, + "tokens_seen": 1800732672 + }, + { + "epoch": 0.55, + "learning_rate": 0.00036724573721163493, + "loss": 1.3812, + "theoretical_loss": 3.489313780681514, + "tokens_seen": 1800798208 + }, + { + "epoch": 0.55, + "learning_rate": 0.00036722968906720163, + "loss": 1.2204, + "theoretical_loss": 3.4893030616070995, + "tokens_seen": 1800863744 + }, + { + "epoch": 0.55, + "learning_rate": 0.00036721364092276833, + "loss": 1.3192, + "theoretical_loss": 3.4892923430319795, + "tokens_seen": 1800929280 + }, + { + "epoch": 0.55, + "learning_rate": 0.00036719759277833503, + "loss": 1.4669, + "theoretical_loss": 3.4892816249561114, + "tokens_seen": 1800994816 + }, + { + "epoch": 0.55, + "learning_rate": 0.00036718154463390173, + "loss": 1.177, + "theoretical_loss": 3.4892709073794546, + "tokens_seen": 1801060352 + }, + { + "epoch": 0.55, + "learning_rate": 0.00036716549648946843, + "loss": 1.3267, + "theoretical_loss": 3.489260190301967, + "tokens_seen": 1801125888 + }, + { + "epoch": 0.55, + "learning_rate": 0.00036714944834503513, + "loss": 1.3355, + "theoretical_loss": 3.489249473723608, + "tokens_seen": 1801191424 + }, + { + "epoch": 0.55, + "learning_rate": 0.0003671334002006018, + "loss": 1.3324, + "theoretical_loss": 3.4892387576443356, + "tokens_seen": 1801256960 + }, + { + "epoch": 0.55, + "learning_rate": 0.0003671173520561685, + "loss": 1.5174, + "theoretical_loss": 3.4892280420641084, + "tokens_seen": 1801322496 + }, + { + "epoch": 0.55, + "learning_rate": 0.0003671013039117352, + "loss": 1.3441, + "theoretical_loss": 3.4892173269828852, + "tokens_seen": 1801388032 + }, + { + "epoch": 0.55, + "learning_rate": 0.0003670852557673019, + "loss": 1.2781, + "theoretical_loss": 3.489206612400625, + "tokens_seen": 1801453568 + }, + { + "epoch": 0.55, + "learning_rate": 0.0003670692076228686, + "loss": 1.34, + "theoretical_loss": 3.489195898317286, + "tokens_seen": 1801519104 + }, + { + "epoch": 0.55, + "learning_rate": 0.0003670531594784353, + "loss": 1.3053, + "theoretical_loss": 3.4891851847328264, + "tokens_seen": 1801584640 + }, + { + "epoch": 0.55, + "learning_rate": 0.000367037111334002, + "loss": 1.1845, + "theoretical_loss": 3.489174471647205, + "tokens_seen": 1801650176 + }, + { + "epoch": 0.55, + "learning_rate": 0.0003670210631895687, + "loss": 1.3557, + "theoretical_loss": 3.4891637590603812, + "tokens_seen": 1801715712 + }, + { + "epoch": 0.55, + "learning_rate": 0.0003670050150451354, + "loss": 1.3336, + "theoretical_loss": 3.489153046972313, + "tokens_seen": 1801781248 + }, + { + "epoch": 0.55, + "learning_rate": 0.0003669889669007021, + "loss": 1.3105, + "theoretical_loss": 3.4891423353829594, + "tokens_seen": 1801846784 + }, + { + "epoch": 0.55, + "learning_rate": 0.0003669729187562688, + "loss": 1.2343, + "theoretical_loss": 3.4891316242922787, + "tokens_seen": 1801912320 + }, + { + "epoch": 0.55, + "learning_rate": 0.00036695687061183554, + "loss": 1.3511, + "theoretical_loss": 3.489120913700229, + "tokens_seen": 1801977856 + }, + { + "epoch": 0.55, + "learning_rate": 0.00036694082246740224, + "loss": 1.2441, + "theoretical_loss": 3.4891102036067703, + "tokens_seen": 1802043392 + }, + { + "epoch": 0.55, + "learning_rate": 0.00036692477432296894, + "loss": 1.2925, + "theoretical_loss": 3.4890994940118603, + "tokens_seen": 1802108928 + }, + { + "epoch": 0.55, + "learning_rate": 0.00036690872617853564, + "loss": 1.4033, + "theoretical_loss": 3.4890887849154577, + "tokens_seen": 1802174464 + }, + { + "epoch": 0.55, + "objective/train/docs_used": 657910, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.4413325786590576, + "objective/train/theoretical_loss": 3.4890780763175218, + "objective/train/tokens_used": 1822700000, + "theoretical_loss": 3.4890780763175218, + "tokens_seen": 1802240000 + }, + { + "epoch": 0.55, + "learning_rate": 0.00036689267803410234, + "loss": 1.4225, + "theoretical_loss": 3.4890780763175218, + "tokens_seen": 1802240000 + }, + { + "epoch": 0.55, + "learning_rate": 0.00036687662988966904, + "loss": 1.3403, + "theoretical_loss": 3.4890673682180107, + "tokens_seen": 1802305536 + }, + { + "epoch": 0.55, + "learning_rate": 0.00036686058174523575, + "loss": 1.4026, + "theoretical_loss": 3.489056660616883, + "tokens_seen": 1802371072 + }, + { + "epoch": 0.55, + "learning_rate": 0.00036684453360080245, + "loss": 1.3043, + "theoretical_loss": 3.489045953514098, + "tokens_seen": 1802436608 + }, + { + "epoch": 0.55, + "learning_rate": 0.00036682848545636915, + "loss": 1.4061, + "theoretical_loss": 3.489035246909614, + "tokens_seen": 1802502144 + }, + { + "epoch": 0.55, + "learning_rate": 0.00036681243731193585, + "loss": 1.3404, + "theoretical_loss": 3.4890245408033893, + "tokens_seen": 1802567680 + }, + { + "epoch": 0.55, + "learning_rate": 0.00036679638916750255, + "loss": 1.3809, + "theoretical_loss": 3.489013835195383, + "tokens_seen": 1802633216 + }, + { + "epoch": 0.55, + "learning_rate": 0.00036678034102306925, + "loss": 1.3126, + "theoretical_loss": 3.489003130085554, + "tokens_seen": 1802698752 + }, + { + "epoch": 0.55, + "learning_rate": 0.00036676429287863595, + "loss": 1.2525, + "theoretical_loss": 3.488992425473861, + "tokens_seen": 1802764288 + }, + { + "epoch": 0.55, + "learning_rate": 0.00036674824473420265, + "loss": 1.4645, + "theoretical_loss": 3.4889817213602625, + "tokens_seen": 1802829824 + }, + { + "epoch": 0.55, + "learning_rate": 0.00036673219658976935, + "loss": 1.4319, + "theoretical_loss": 3.488971017744717, + "tokens_seen": 1802895360 + }, + { + "epoch": 0.55, + "learning_rate": 0.00036671614844533605, + "loss": 1.3309, + "theoretical_loss": 3.4889603146271835, + "tokens_seen": 1802960896 + }, + { + "epoch": 0.55, + "learning_rate": 0.00036670010030090275, + "loss": 1.4147, + "theoretical_loss": 3.488949612007621, + "tokens_seen": 1803026432 + }, + { + "epoch": 0.55, + "learning_rate": 0.00036668405215646945, + "loss": 1.3743, + "theoretical_loss": 3.488938909885988, + "tokens_seen": 1803091968 + }, + { + "epoch": 0.55, + "learning_rate": 0.00036666800401203615, + "loss": 1.3032, + "theoretical_loss": 3.4889282082622426, + "tokens_seen": 1803157504 + }, + { + "epoch": 0.55, + "learning_rate": 0.00036665195586760285, + "loss": 1.3152, + "theoretical_loss": 3.4889175071363443, + "tokens_seen": 1803223040 + }, + { + "epoch": 0.55, + "learning_rate": 0.00036663590772316955, + "loss": 1.319, + "theoretical_loss": 3.4889068065082522, + "tokens_seen": 1803288576 + }, + { + "epoch": 0.55, + "learning_rate": 0.00036661985957873625, + "loss": 1.2965, + "theoretical_loss": 3.488896106377924, + "tokens_seen": 1803354112 + }, + { + "epoch": 0.55, + "learning_rate": 0.00036660381143430296, + "loss": 1.3165, + "theoretical_loss": 3.488885406745319, + "tokens_seen": 1803419648 + }, + { + "epoch": 0.55, + "learning_rate": 0.00036658776328986966, + "loss": 1.4887, + "theoretical_loss": 3.4888747076103956, + "tokens_seen": 1803485184 + }, + { + "epoch": 0.55, + "learning_rate": 0.00036657171514543636, + "loss": 1.2674, + "theoretical_loss": 3.4888640089731133, + "tokens_seen": 1803550720 + }, + { + "epoch": 0.55, + "learning_rate": 0.00036655566700100306, + "loss": 1.1985, + "theoretical_loss": 3.4888533108334303, + "tokens_seen": 1803616256 + }, + { + "epoch": 0.55, + "learning_rate": 0.00036653961885656976, + "loss": 1.2358, + "theoretical_loss": 3.4888426131913057, + "tokens_seen": 1803681792 + }, + { + "epoch": 0.55, + "learning_rate": 0.00036652357071213646, + "loss": 1.3381, + "theoretical_loss": 3.4888319160466983, + "tokens_seen": 1803747328 + }, + { + "epoch": 0.55, + "learning_rate": 0.0003665075225677031, + "loss": 1.3443, + "theoretical_loss": 3.488821219399566, + "tokens_seen": 1803812864 + }, + { + "epoch": 0.55, + "learning_rate": 0.0003664914744232698, + "loss": 1.4359, + "theoretical_loss": 3.488810523249869, + "tokens_seen": 1803878400 + }, + { + "epoch": 0.55, + "learning_rate": 0.0003664754262788365, + "loss": 1.355, + "theoretical_loss": 3.488799827597565, + "tokens_seen": 1803943936 + }, + { + "epoch": 0.55, + "learning_rate": 0.0003664593781344032, + "loss": 1.3902, + "theoretical_loss": 3.4887891324426135, + "tokens_seen": 1804009472 + }, + { + "epoch": 0.55, + "learning_rate": 0.0003664433299899699, + "loss": 1.3632, + "theoretical_loss": 3.4887784377849727, + "tokens_seen": 1804075008 + }, + { + "epoch": 0.55, + "learning_rate": 0.0003664272818455366, + "loss": 1.4452, + "theoretical_loss": 3.488767743624602, + "tokens_seen": 1804140544 + }, + { + "epoch": 0.55, + "learning_rate": 0.0003664112337011033, + "loss": 1.3295, + "theoretical_loss": 3.4887570499614595, + "tokens_seen": 1804206080 + }, + { + "epoch": 0.55, + "learning_rate": 0.00036639518555667, + "loss": 1.4685, + "theoretical_loss": 3.488746356795505, + "tokens_seen": 1804271616 + }, + { + "epoch": 0.55, + "learning_rate": 0.0003663791374122367, + "loss": 1.2856, + "theoretical_loss": 3.488735664126697, + "tokens_seen": 1804337152 + }, + { + "epoch": 0.55, + "learning_rate": 0.0003663630892678034, + "loss": 1.4154, + "theoretical_loss": 3.4887249719549933, + "tokens_seen": 1804402688 + }, + { + "epoch": 0.55, + "learning_rate": 0.0003663470411233701, + "loss": 1.4412, + "theoretical_loss": 3.488714280280354, + "tokens_seen": 1804468224 + }, + { + "epoch": 0.55, + "learning_rate": 0.0003663309929789368, + "loss": 1.348, + "theoretical_loss": 3.4887035891027374, + "tokens_seen": 1804533760 + }, + { + "epoch": 0.55, + "learning_rate": 0.0003663149448345035, + "loss": 1.3594, + "theoretical_loss": 3.4886928984221024, + "tokens_seen": 1804599296 + }, + { + "epoch": 0.55, + "learning_rate": 0.0003662988966900702, + "loss": 1.3441, + "theoretical_loss": 3.4886822082384077, + "tokens_seen": 1804664832 + }, + { + "epoch": 0.55, + "learning_rate": 0.0003662828485456369, + "loss": 1.3777, + "theoretical_loss": 3.4886715185516124, + "tokens_seen": 1804730368 + }, + { + "epoch": 0.55, + "learning_rate": 0.0003662668004012036, + "loss": 1.3479, + "theoretical_loss": 3.4886608293616757, + "tokens_seen": 1804795904 + }, + { + "epoch": 0.55, + "learning_rate": 0.0003662507522567703, + "loss": 1.2276, + "theoretical_loss": 3.4886501406685557, + "tokens_seen": 1804861440 + }, + { + "epoch": 0.55, + "learning_rate": 0.000366234704112337, + "loss": 1.272, + "theoretical_loss": 3.4886394524722117, + "tokens_seen": 1804926976 + }, + { + "epoch": 0.55, + "learning_rate": 0.0003662186559679037, + "loss": 1.3448, + "theoretical_loss": 3.488628764772603, + "tokens_seen": 1804992512 + }, + { + "epoch": 0.55, + "learning_rate": 0.0003662026078234704, + "loss": 1.3449, + "theoretical_loss": 3.4886180775696873, + "tokens_seen": 1805058048 + }, + { + "epoch": 0.55, + "learning_rate": 0.0003661865596790371, + "loss": 1.3267, + "theoretical_loss": 3.4886073908634243, + "tokens_seen": 1805123584 + }, + { + "epoch": 0.55, + "learning_rate": 0.0003661705115346038, + "loss": 1.457, + "theoretical_loss": 3.488596704653773, + "tokens_seen": 1805189120 + }, + { + "epoch": 0.55, + "learning_rate": 0.0003661544633901705, + "loss": 1.2944, + "theoretical_loss": 3.488586018940692, + "tokens_seen": 1805254656 + }, + { + "epoch": 0.55, + "learning_rate": 0.0003661384152457372, + "loss": 1.35, + "theoretical_loss": 3.4885753337241403, + "tokens_seen": 1805320192 + }, + { + "epoch": 0.55, + "learning_rate": 0.0003661223671013039, + "loss": 1.3733, + "theoretical_loss": 3.4885646490040765, + "tokens_seen": 1805385728 + }, + { + "epoch": 0.55, + "learning_rate": 0.0003661063189568706, + "loss": 1.4049, + "theoretical_loss": 3.48855396478046, + "tokens_seen": 1805451264 + }, + { + "epoch": 0.55, + "objective/train/docs_used": 658936, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.2056217193603516, + "objective/train/theoretical_loss": 3.4885432810532495, + "objective/train/tokens_used": 1825976800, + "theoretical_loss": 3.4885432810532495, + "tokens_seen": 1805516800 + }, + { + "epoch": 0.55, + "learning_rate": 0.0003660902708124373, + "loss": 1.3752, + "theoretical_loss": 3.4885432810532495, + "tokens_seen": 1805516800 + }, + { + "epoch": 0.55, + "learning_rate": 0.000366074222668004, + "loss": 1.4137, + "theoretical_loss": 3.4885325978224038, + "tokens_seen": 1805582336 + }, + { + "epoch": 0.55, + "learning_rate": 0.0003660581745235707, + "loss": 1.3251, + "theoretical_loss": 3.488521915087882, + "tokens_seen": 1805647872 + }, + { + "epoch": 0.55, + "learning_rate": 0.0003660421263791374, + "loss": 1.4614, + "theoretical_loss": 3.4885112328496426, + "tokens_seen": 1805713408 + }, + { + "epoch": 0.55, + "learning_rate": 0.0003660260782347041, + "loss": 1.275, + "theoretical_loss": 3.4885005511076455, + "tokens_seen": 1805778944 + }, + { + "epoch": 0.55, + "learning_rate": 0.0003660100300902708, + "loss": 1.4443, + "theoretical_loss": 3.4884898698618487, + "tokens_seen": 1805844480 + }, + { + "epoch": 0.55, + "learning_rate": 0.0003659939819458375, + "loss": 1.3551, + "theoretical_loss": 3.4884791891122116, + "tokens_seen": 1805910016 + }, + { + "epoch": 0.55, + "learning_rate": 0.0003659779338014042, + "loss": 1.3895, + "theoretical_loss": 3.488468508858693, + "tokens_seen": 1805975552 + }, + { + "epoch": 0.55, + "learning_rate": 0.0003659618856569709, + "loss": 1.3288, + "theoretical_loss": 3.488457829101252, + "tokens_seen": 1806041088 + }, + { + "epoch": 0.55, + "learning_rate": 0.00036594583751253763, + "loss": 1.3073, + "theoretical_loss": 3.488447149839847, + "tokens_seen": 1806106624 + }, + { + "epoch": 0.55, + "learning_rate": 0.00036592978936810433, + "loss": 1.2741, + "theoretical_loss": 3.4884364710744378, + "tokens_seen": 1806172160 + }, + { + "epoch": 0.55, + "learning_rate": 0.00036591374122367103, + "loss": 1.2421, + "theoretical_loss": 3.488425792804983, + "tokens_seen": 1806237696 + }, + { + "epoch": 0.55, + "learning_rate": 0.00036589769307923773, + "loss": 1.3089, + "theoretical_loss": 3.488415115031441, + "tokens_seen": 1806303232 + }, + { + "epoch": 0.55, + "learning_rate": 0.00036588164493480443, + "loss": 1.3321, + "theoretical_loss": 3.488404437753772, + "tokens_seen": 1806368768 + }, + { + "epoch": 0.55, + "learning_rate": 0.00036586559679037113, + "loss": 1.3255, + "theoretical_loss": 3.4883937609719338, + "tokens_seen": 1806434304 + }, + { + "epoch": 0.55, + "learning_rate": 0.00036584954864593783, + "loss": 1.3477, + "theoretical_loss": 3.4883830846858856, + "tokens_seen": 1806499840 + }, + { + "epoch": 0.55, + "learning_rate": 0.00036583350050150453, + "loss": 1.3415, + "theoretical_loss": 3.4883724088955876, + "tokens_seen": 1806565376 + }, + { + "epoch": 0.55, + "learning_rate": 0.00036581745235707123, + "loss": 1.2945, + "theoretical_loss": 3.488361733600997, + "tokens_seen": 1806630912 + }, + { + "epoch": 0.55, + "learning_rate": 0.00036580140421263793, + "loss": 1.4306, + "theoretical_loss": 3.488351058802074, + "tokens_seen": 1806696448 + }, + { + "epoch": 0.55, + "learning_rate": 0.00036578535606820463, + "loss": 1.3003, + "theoretical_loss": 3.4883403844987777, + "tokens_seen": 1806761984 + }, + { + "epoch": 0.55, + "learning_rate": 0.00036576930792377133, + "loss": 1.3509, + "theoretical_loss": 3.4883297106910662, + "tokens_seen": 1806827520 + }, + { + "epoch": 0.55, + "learning_rate": 0.00036575325977933803, + "loss": 1.3656, + "theoretical_loss": 3.4883190373788997, + "tokens_seen": 1806893056 + }, + { + "epoch": 0.55, + "learning_rate": 0.00036573721163490474, + "loss": 1.4072, + "theoretical_loss": 3.4883083645622355, + "tokens_seen": 1806958592 + }, + { + "epoch": 0.55, + "learning_rate": 0.00036572116349047144, + "loss": 1.3288, + "theoretical_loss": 3.4882976922410345, + "tokens_seen": 1807024128 + }, + { + "epoch": 0.55, + "learning_rate": 0.00036570511534603814, + "loss": 1.4308, + "theoretical_loss": 3.4882870204152545, + "tokens_seen": 1807089664 + }, + { + "epoch": 0.55, + "learning_rate": 0.00036568906720160484, + "loss": 1.3781, + "theoretical_loss": 3.488276349084855, + "tokens_seen": 1807155200 + }, + { + "epoch": 0.55, + "learning_rate": 0.00036567301905717154, + "loss": 1.4054, + "theoretical_loss": 3.488265678249795, + "tokens_seen": 1807220736 + }, + { + "epoch": 0.55, + "learning_rate": 0.00036565697091273824, + "loss": 1.3751, + "theoretical_loss": 3.4882550079100336, + "tokens_seen": 1807286272 + }, + { + "epoch": 0.55, + "learning_rate": 0.00036564092276830494, + "loss": 1.3133, + "theoretical_loss": 3.4882443380655297, + "tokens_seen": 1807351808 + }, + { + "epoch": 0.55, + "learning_rate": 0.00036562487462387164, + "loss": 1.3729, + "theoretical_loss": 3.4882336687162425, + "tokens_seen": 1807417344 + }, + { + "epoch": 0.55, + "learning_rate": 0.00036560882647943834, + "loss": 1.3346, + "theoretical_loss": 3.4882229998621304, + "tokens_seen": 1807482880 + }, + { + "epoch": 0.55, + "learning_rate": 0.00036559277833500504, + "loss": 1.292, + "theoretical_loss": 3.4882123315031537, + "tokens_seen": 1807548416 + }, + { + "epoch": 0.55, + "learning_rate": 0.00036557673019057174, + "loss": 1.2581, + "theoretical_loss": 3.4882016636392708, + "tokens_seen": 1807613952 + }, + { + "epoch": 0.55, + "learning_rate": 0.00036556068204613844, + "loss": 1.3559, + "theoretical_loss": 3.4881909962704407, + "tokens_seen": 1807679488 + }, + { + "epoch": 0.55, + "learning_rate": 0.00036554463390170514, + "loss": 1.4812, + "theoretical_loss": 3.4881803293966227, + "tokens_seen": 1807745024 + }, + { + "epoch": 0.55, + "learning_rate": 0.00036552858575727184, + "loss": 1.3754, + "theoretical_loss": 3.488169663017775, + "tokens_seen": 1807810560 + }, + { + "epoch": 0.55, + "learning_rate": 0.00036551253761283854, + "loss": 1.435, + "theoretical_loss": 3.4881589971338585, + "tokens_seen": 1807876096 + }, + { + "epoch": 0.55, + "learning_rate": 0.00036549648946840525, + "loss": 1.4032, + "theoretical_loss": 3.4881483317448305, + "tokens_seen": 1807941632 + }, + { + "epoch": 0.55, + "learning_rate": 0.00036548044132397195, + "loss": 1.3103, + "theoretical_loss": 3.488137666850651, + "tokens_seen": 1808007168 + }, + { + "epoch": 0.55, + "learning_rate": 0.00036546439317953865, + "loss": 1.3509, + "theoretical_loss": 3.488127002451279, + "tokens_seen": 1808072704 + }, + { + "epoch": 0.55, + "learning_rate": 0.00036544834503510535, + "loss": 1.3407, + "theoretical_loss": 3.488116338546674, + "tokens_seen": 1808138240 + }, + { + "epoch": 0.55, + "learning_rate": 0.00036543229689067205, + "loss": 1.3105, + "theoretical_loss": 3.488105675136794, + "tokens_seen": 1808203776 + }, + { + "epoch": 0.55, + "learning_rate": 0.00036541624874623875, + "loss": 1.3854, + "theoretical_loss": 3.4880950122215992, + "tokens_seen": 1808269312 + }, + { + "epoch": 0.55, + "learning_rate": 0.00036540020060180545, + "loss": 1.2812, + "theoretical_loss": 3.4880843498010483, + "tokens_seen": 1808334848 + }, + { + "epoch": 0.55, + "learning_rate": 0.00036538415245737215, + "loss": 1.3102, + "theoretical_loss": 3.4880736878751, + "tokens_seen": 1808400384 + }, + { + "epoch": 0.55, + "learning_rate": 0.00036536810431293885, + "loss": 1.4735, + "theoretical_loss": 3.4880630264437142, + "tokens_seen": 1808465920 + }, + { + "epoch": 0.55, + "learning_rate": 0.00036535205616850555, + "loss": 1.2904, + "theoretical_loss": 3.48805236550685, + "tokens_seen": 1808531456 + }, + { + "epoch": 0.55, + "learning_rate": 0.00036533600802407225, + "loss": 1.4909, + "theoretical_loss": 3.488041705064466, + "tokens_seen": 1808596992 + }, + { + "epoch": 0.55, + "learning_rate": 0.00036531995987963895, + "loss": 1.3018, + "theoretical_loss": 3.4880310451165215, + "tokens_seen": 1808662528 + }, + { + "epoch": 0.55, + "learning_rate": 0.00036530391173520565, + "loss": 1.3519, + "theoretical_loss": 3.488020385662976, + "tokens_seen": 1808728064 + }, + { + "epoch": 0.55, + "objective/train/docs_used": 660069, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.3367552757263184, + "objective/train/theoretical_loss": 3.4880097267037877, + "objective/train/tokens_used": 1829253600, + "theoretical_loss": 3.4880097267037877, + "tokens_seen": 1808793600 + }, + { + "epoch": 0.55, + "learning_rate": 0.00036528786359077235, + "loss": 1.3482, + "theoretical_loss": 3.4880097267037877, + "tokens_seen": 1808793600 + }, + { + "epoch": 0.55, + "learning_rate": 0.00036527181544633905, + "loss": 1.3757, + "theoretical_loss": 3.4879990682389175, + "tokens_seen": 1808859136 + }, + { + "epoch": 0.55, + "learning_rate": 0.00036525576730190575, + "loss": 1.3424, + "theoretical_loss": 3.487988410268323, + "tokens_seen": 1808924672 + }, + { + "epoch": 0.55, + "learning_rate": 0.00036523971915747246, + "loss": 1.2677, + "theoretical_loss": 3.4879777527919638, + "tokens_seen": 1808990208 + }, + { + "epoch": 0.55, + "learning_rate": 0.00036522367101303916, + "loss": 1.3388, + "theoretical_loss": 3.4879670958097995, + "tokens_seen": 1809055744 + }, + { + "epoch": 0.55, + "learning_rate": 0.00036520762286860586, + "loss": 1.3344, + "theoretical_loss": 3.4879564393217892, + "tokens_seen": 1809121280 + }, + { + "epoch": 0.55, + "learning_rate": 0.00036519157472417256, + "loss": 1.3858, + "theoretical_loss": 3.4879457833278913, + "tokens_seen": 1809186816 + }, + { + "epoch": 0.55, + "learning_rate": 0.00036517552657973926, + "loss": 1.4014, + "theoretical_loss": 3.4879351278280657, + "tokens_seen": 1809252352 + }, + { + "epoch": 0.55, + "learning_rate": 0.00036515947843530596, + "loss": 1.3517, + "theoretical_loss": 3.4879244728222716, + "tokens_seen": 1809317888 + }, + { + "epoch": 0.55, + "learning_rate": 0.00036514343029087266, + "loss": 1.3159, + "theoretical_loss": 3.487913818310468, + "tokens_seen": 1809383424 + }, + { + "epoch": 0.55, + "learning_rate": 0.00036512738214643936, + "loss": 1.2865, + "theoretical_loss": 3.4879031642926144, + "tokens_seen": 1809448960 + }, + { + "epoch": 0.55, + "learning_rate": 0.00036511133400200606, + "loss": 1.249, + "theoretical_loss": 3.4878925107686696, + "tokens_seen": 1809514496 + }, + { + "epoch": 0.55, + "learning_rate": 0.0003650952858575727, + "loss": 1.3569, + "theoretical_loss": 3.487881857738593, + "tokens_seen": 1809580032 + }, + { + "epoch": 0.55, + "learning_rate": 0.0003650792377131394, + "loss": 1.33, + "theoretical_loss": 3.4878712052023433, + "tokens_seen": 1809645568 + }, + { + "epoch": 0.55, + "learning_rate": 0.0003650631895687061, + "loss": 1.2825, + "theoretical_loss": 3.487860553159881, + "tokens_seen": 1809711104 + }, + { + "epoch": 0.55, + "learning_rate": 0.0003650471414242728, + "loss": 1.2459, + "theoretical_loss": 3.487849901611164, + "tokens_seen": 1809776640 + }, + { + "epoch": 0.55, + "learning_rate": 0.0003650310932798395, + "loss": 1.1957, + "theoretical_loss": 3.487839250556153, + "tokens_seen": 1809842176 + }, + { + "epoch": 0.55, + "learning_rate": 0.0003650150451354062, + "loss": 1.4162, + "theoretical_loss": 3.4878285999948053, + "tokens_seen": 1809907712 + }, + { + "epoch": 0.55, + "learning_rate": 0.0003649989969909729, + "loss": 1.3784, + "theoretical_loss": 3.4878179499270816, + "tokens_seen": 1809973248 + }, + { + "epoch": 0.55, + "learning_rate": 0.0003649829488465396, + "loss": 1.3499, + "theoretical_loss": 3.487807300352941, + "tokens_seen": 1810038784 + }, + { + "epoch": 0.55, + "learning_rate": 0.0003649669007021063, + "loss": 1.3265, + "theoretical_loss": 3.487796651272342, + "tokens_seen": 1810104320 + }, + { + "epoch": 0.55, + "learning_rate": 0.000364950852557673, + "loss": 1.298, + "theoretical_loss": 3.4877860026852443, + "tokens_seen": 1810169856 + }, + { + "epoch": 0.55, + "learning_rate": 0.0003649348044132397, + "loss": 1.3068, + "theoretical_loss": 3.4877753545916077, + "tokens_seen": 1810235392 + }, + { + "epoch": 0.55, + "learning_rate": 0.0003649187562688064, + "loss": 1.3903, + "theoretical_loss": 3.487764706991391, + "tokens_seen": 1810300928 + }, + { + "epoch": 0.55, + "learning_rate": 0.0003649027081243731, + "loss": 1.378, + "theoretical_loss": 3.487754059884553, + "tokens_seen": 1810366464 + }, + { + "epoch": 0.55, + "learning_rate": 0.0003648866599799398, + "loss": 1.4351, + "theoretical_loss": 3.487743413271054, + "tokens_seen": 1810432000 + }, + { + "epoch": 0.55, + "learning_rate": 0.0003648706118355065, + "loss": 1.3326, + "theoretical_loss": 3.487732767150852, + "tokens_seen": 1810497536 + }, + { + "epoch": 0.55, + "learning_rate": 0.0003648545636910732, + "loss": 1.3132, + "theoretical_loss": 3.487722121523907, + "tokens_seen": 1810563072 + }, + { + "epoch": 0.55, + "learning_rate": 0.0003648385155466399, + "loss": 1.4261, + "theoretical_loss": 3.487711476390179, + "tokens_seen": 1810628608 + }, + { + "epoch": 0.55, + "learning_rate": 0.0003648224674022066, + "loss": 1.3621, + "theoretical_loss": 3.4877008317496268, + "tokens_seen": 1810694144 + }, + { + "epoch": 0.55, + "learning_rate": 0.0003648064192577733, + "loss": 1.3062, + "theoretical_loss": 3.4876901876022086, + "tokens_seen": 1810759680 + }, + { + "epoch": 0.55, + "learning_rate": 0.00036479037111334, + "loss": 1.3306, + "theoretical_loss": 3.487679543947885, + "tokens_seen": 1810825216 + }, + { + "epoch": 0.55, + "learning_rate": 0.0003647743229689067, + "loss": 1.3173, + "theoretical_loss": 3.4876689007866153, + "tokens_seen": 1810890752 + }, + { + "epoch": 0.55, + "learning_rate": 0.0003647582748244734, + "loss": 1.3081, + "theoretical_loss": 3.4876582581183575, + "tokens_seen": 1810956288 + }, + { + "epoch": 0.55, + "learning_rate": 0.0003647422266800401, + "loss": 1.3323, + "theoretical_loss": 3.4876476159430725, + "tokens_seen": 1811021824 + }, + { + "epoch": 0.55, + "learning_rate": 0.0003647261785356068, + "loss": 1.3046, + "theoretical_loss": 3.4876369742607194, + "tokens_seen": 1811087360 + }, + { + "epoch": 0.55, + "learning_rate": 0.0003647101303911735, + "loss": 1.3187, + "theoretical_loss": 3.4876263330712565, + "tokens_seen": 1811152896 + }, + { + "epoch": 0.55, + "learning_rate": 0.0003646940822467402, + "loss": 1.3544, + "theoretical_loss": 3.4876156923746437, + "tokens_seen": 1811218432 + }, + { + "epoch": 0.55, + "learning_rate": 0.0003646780341023069, + "loss": 1.3307, + "theoretical_loss": 3.487605052170841, + "tokens_seen": 1811283968 + }, + { + "epoch": 0.55, + "learning_rate": 0.0003646619859578736, + "loss": 1.4127, + "theoretical_loss": 3.4875944124598064, + "tokens_seen": 1811349504 + }, + { + "epoch": 0.55, + "learning_rate": 0.0003646459378134403, + "loss": 1.3418, + "theoretical_loss": 3.487583773241501, + "tokens_seen": 1811415040 + }, + { + "epoch": 0.55, + "learning_rate": 0.000364629889669007, + "loss": 1.351, + "theoretical_loss": 3.4875731345158822, + "tokens_seen": 1811480576 + }, + { + "epoch": 0.55, + "learning_rate": 0.0003646138415245737, + "loss": 1.324, + "theoretical_loss": 3.4875624962829104, + "tokens_seen": 1811546112 + }, + { + "epoch": 0.55, + "learning_rate": 0.0003645977933801404, + "loss": 1.2559, + "theoretical_loss": 3.4875518585425453, + "tokens_seen": 1811611648 + }, + { + "epoch": 0.55, + "learning_rate": 0.0003645817452357071, + "loss": 1.3342, + "theoretical_loss": 3.487541221294746, + "tokens_seen": 1811677184 + }, + { + "epoch": 0.55, + "learning_rate": 0.00036456569709127383, + "loss": 1.2929, + "theoretical_loss": 3.4875305845394715, + "tokens_seen": 1811742720 + }, + { + "epoch": 0.55, + "learning_rate": 0.0003645496489468406, + "loss": 1.3704, + "theoretical_loss": 3.487519948276681, + "tokens_seen": 1811808256 + }, + { + "epoch": 0.55, + "learning_rate": 0.0003645336008024073, + "loss": 1.2748, + "theoretical_loss": 3.487509312506335, + "tokens_seen": 1811873792 + }, + { + "epoch": 0.55, + "learning_rate": 0.000364517552657974, + "loss": 1.4002, + "theoretical_loss": 3.4874986772283916, + "tokens_seen": 1811939328 + }, + { + "epoch": 0.55, + "learning_rate": 0.0003645015045135407, + "loss": 1.3292, + "theoretical_loss": 3.487488042442811, + "tokens_seen": 1812004864 + }, + { + "epoch": 0.55, + "objective/train/docs_used": 661289, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.2686200141906738, + "objective/train/theoretical_loss": 3.4874774081495525, + "objective/train/tokens_used": 1832530400, + "theoretical_loss": 3.4874774081495525, + "tokens_seen": 1812070400 + }, + { + "epoch": 0.55, + "learning_rate": 0.0003644854563691074, + "loss": 1.3145, + "theoretical_loss": 3.4874774081495525, + "tokens_seen": 1812070400 + }, + { + "epoch": 0.55, + "learning_rate": 0.00036446940822467403, + "loss": 1.2834, + "theoretical_loss": 3.487466774348575, + "tokens_seen": 1812135936 + }, + { + "epoch": 0.55, + "learning_rate": 0.00036445336008024073, + "loss": 1.3604, + "theoretical_loss": 3.4874561410398393, + "tokens_seen": 1812201472 + }, + { + "epoch": 0.55, + "learning_rate": 0.00036443731193580743, + "loss": 1.2502, + "theoretical_loss": 3.4874455082233027, + "tokens_seen": 1812267008 + }, + { + "epoch": 0.55, + "learning_rate": 0.00036442126379137413, + "loss": 1.3134, + "theoretical_loss": 3.487434875898926, + "tokens_seen": 1812332544 + }, + { + "epoch": 0.55, + "learning_rate": 0.00036440521564694083, + "loss": 1.3641, + "theoretical_loss": 3.4874242440666685, + "tokens_seen": 1812398080 + }, + { + "epoch": 0.55, + "learning_rate": 0.00036438916750250753, + "loss": 1.5003, + "theoretical_loss": 3.487413612726489, + "tokens_seen": 1812463616 + }, + { + "epoch": 0.55, + "learning_rate": 0.00036437311935807424, + "loss": 1.385, + "theoretical_loss": 3.4874029818783487, + "tokens_seen": 1812529152 + }, + { + "epoch": 0.55, + "learning_rate": 0.00036435707121364094, + "loss": 1.3482, + "theoretical_loss": 3.4873923515222045, + "tokens_seen": 1812594688 + }, + { + "epoch": 0.55, + "learning_rate": 0.00036434102306920764, + "loss": 1.3707, + "theoretical_loss": 3.487381721658018, + "tokens_seen": 1812660224 + }, + { + "epoch": 0.55, + "learning_rate": 0.00036432497492477434, + "loss": 1.3323, + "theoretical_loss": 3.487371092285747, + "tokens_seen": 1812725760 + }, + { + "epoch": 0.55, + "learning_rate": 0.00036430892678034104, + "loss": 1.3071, + "theoretical_loss": 3.487360463405352, + "tokens_seen": 1812791296 + }, + { + "epoch": 0.55, + "learning_rate": 0.00036429287863590774, + "loss": 1.3006, + "theoretical_loss": 3.4873498350167917, + "tokens_seen": 1812856832 + }, + { + "epoch": 0.55, + "learning_rate": 0.00036427683049147444, + "loss": 1.2599, + "theoretical_loss": 3.4873392071200264, + "tokens_seen": 1812922368 + }, + { + "epoch": 0.55, + "learning_rate": 0.00036426078234704114, + "loss": 1.4307, + "theoretical_loss": 3.4873285797150153, + "tokens_seen": 1812987904 + }, + { + "epoch": 0.55, + "learning_rate": 0.00036424473420260784, + "loss": 1.4156, + "theoretical_loss": 3.4873179528017175, + "tokens_seen": 1813053440 + }, + { + "epoch": 0.55, + "learning_rate": 0.00036422868605817454, + "loss": 1.4513, + "theoretical_loss": 3.487307326380093, + "tokens_seen": 1813118976 + }, + { + "epoch": 0.55, + "learning_rate": 0.00036421263791374124, + "loss": 1.2875, + "theoretical_loss": 3.487296700450101, + "tokens_seen": 1813184512 + }, + { + "epoch": 0.55, + "learning_rate": 0.00036419658976930794, + "loss": 1.3015, + "theoretical_loss": 3.487286075011701, + "tokens_seen": 1813250048 + }, + { + "epoch": 0.55, + "learning_rate": 0.00036418054162487464, + "loss": 1.3012, + "theoretical_loss": 3.4872754500648524, + "tokens_seen": 1813315584 + }, + { + "epoch": 0.55, + "learning_rate": 0.00036416449348044134, + "loss": 1.1548, + "theoretical_loss": 3.4872648256095147, + "tokens_seen": 1813381120 + }, + { + "epoch": 0.55, + "learning_rate": 0.00036414844533600804, + "loss": 1.4051, + "theoretical_loss": 3.487254201645648, + "tokens_seen": 1813446656 + }, + { + "epoch": 0.55, + "learning_rate": 0.00036413239719157474, + "loss": 1.4218, + "theoretical_loss": 3.4872435781732105, + "tokens_seen": 1813512192 + }, + { + "epoch": 0.55, + "learning_rate": 0.00036411634904714145, + "loss": 1.3284, + "theoretical_loss": 3.487232955192163, + "tokens_seen": 1813577728 + }, + { + "epoch": 0.55, + "learning_rate": 0.00036410030090270815, + "loss": 1.4187, + "theoretical_loss": 3.487222332702464, + "tokens_seen": 1813643264 + }, + { + "epoch": 0.55, + "learning_rate": 0.00036408425275827485, + "loss": 1.3461, + "theoretical_loss": 3.487211710704074, + "tokens_seen": 1813708800 + }, + { + "epoch": 0.55, + "learning_rate": 0.00036406820461384155, + "loss": 1.3195, + "theoretical_loss": 3.487201089196952, + "tokens_seen": 1813774336 + }, + { + "epoch": 0.55, + "learning_rate": 0.00036405215646940825, + "loss": 1.3243, + "theoretical_loss": 3.4871904681810575, + "tokens_seen": 1813839872 + }, + { + "epoch": 0.55, + "learning_rate": 0.00036403610832497495, + "loss": 1.4353, + "theoretical_loss": 3.48717984765635, + "tokens_seen": 1813905408 + }, + { + "epoch": 0.55, + "learning_rate": 0.00036402006018054165, + "loss": 1.3907, + "theoretical_loss": 3.4871692276227897, + "tokens_seen": 1813970944 + }, + { + "epoch": 0.55, + "learning_rate": 0.00036400401203610835, + "loss": 1.4147, + "theoretical_loss": 3.487158608080335, + "tokens_seen": 1814036480 + }, + { + "epoch": 0.55, + "learning_rate": 0.00036398796389167505, + "loss": 1.4957, + "theoretical_loss": 3.487147989028946, + "tokens_seen": 1814102016 + }, + { + "epoch": 0.55, + "learning_rate": 0.00036397191574724175, + "loss": 1.3313, + "theoretical_loss": 3.4871373704685826, + "tokens_seen": 1814167552 + }, + { + "epoch": 0.55, + "learning_rate": 0.00036395586760280845, + "loss": 1.4293, + "theoretical_loss": 3.487126752399204, + "tokens_seen": 1814233088 + }, + { + "epoch": 0.55, + "learning_rate": 0.00036393981945837515, + "loss": 1.3969, + "theoretical_loss": 3.48711613482077, + "tokens_seen": 1814298624 + }, + { + "epoch": 0.55, + "learning_rate": 0.00036392377131394185, + "loss": 1.4151, + "theoretical_loss": 3.48710551773324, + "tokens_seen": 1814364160 + }, + { + "epoch": 0.55, + "learning_rate": 0.00036390772316950855, + "loss": 1.247, + "theoretical_loss": 3.4870949011365733, + "tokens_seen": 1814429696 + }, + { + "epoch": 0.55, + "learning_rate": 0.00036389167502507525, + "loss": 1.3653, + "theoretical_loss": 3.48708428503073, + "tokens_seen": 1814495232 + }, + { + "epoch": 0.55, + "learning_rate": 0.00036387562688064195, + "loss": 1.3068, + "theoretical_loss": 3.487073669415669, + "tokens_seen": 1814560768 + }, + { + "epoch": 0.55, + "learning_rate": 0.00036385957873620866, + "loss": 1.2617, + "theoretical_loss": 3.487063054291351, + "tokens_seen": 1814626304 + }, + { + "epoch": 0.55, + "learning_rate": 0.00036384353059177536, + "loss": 1.392, + "theoretical_loss": 3.4870524396577345, + "tokens_seen": 1814691840 + }, + { + "epoch": 0.55, + "learning_rate": 0.00036382748244734206, + "loss": 1.2418, + "theoretical_loss": 3.4870418255147793, + "tokens_seen": 1814757376 + }, + { + "epoch": 0.55, + "learning_rate": 0.00036381143430290876, + "loss": 1.2803, + "theoretical_loss": 3.487031211862446, + "tokens_seen": 1814822912 + }, + { + "epoch": 0.55, + "learning_rate": 0.00036379538615847546, + "loss": 1.3209, + "theoretical_loss": 3.4870205987006924, + "tokens_seen": 1814888448 + }, + { + "epoch": 0.55, + "learning_rate": 0.00036377933801404216, + "loss": 1.3981, + "theoretical_loss": 3.48700998602948, + "tokens_seen": 1814953984 + }, + { + "epoch": 0.55, + "learning_rate": 0.00036376328986960886, + "loss": 1.391, + "theoretical_loss": 3.486999373848767, + "tokens_seen": 1815019520 + }, + { + "epoch": 0.55, + "learning_rate": 0.00036374724172517556, + "loss": 1.4263, + "theoretical_loss": 3.4869887621585134, + "tokens_seen": 1815085056 + }, + { + "epoch": 0.55, + "learning_rate": 0.00036373119358074226, + "loss": 1.3762, + "theoretical_loss": 3.486978150958679, + "tokens_seen": 1815150592 + }, + { + "epoch": 0.55, + "learning_rate": 0.00036371514543630896, + "loss": 1.266, + "theoretical_loss": 3.486967540249224, + "tokens_seen": 1815216128 + }, + { + "epoch": 0.55, + "learning_rate": 0.0003636990972918756, + "loss": 1.4238, + "theoretical_loss": 3.486956930030107, + "tokens_seen": 1815281664 + }, + { + "epoch": 0.55, + "objective/train/docs_used": 662349, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.2700244188308716, + "objective/train/theoretical_loss": 3.4869463203012883, + "objective/train/tokens_used": 1835807200, + "theoretical_loss": 3.4869463203012883, + "tokens_seen": 1815347200 + }, + { + "epoch": 0.55, + "learning_rate": 0.0003636830491474423, + "loss": 1.3109, + "theoretical_loss": 3.4869463203012883, + "tokens_seen": 1815347200 + }, + { + "epoch": 0.55, + "learning_rate": 0.000363667001003009, + "loss": 1.4094, + "theoretical_loss": 3.486935711062727, + "tokens_seen": 1815412736 + }, + { + "epoch": 0.55, + "learning_rate": 0.0003636509528585757, + "loss": 1.331, + "theoretical_loss": 3.4869251023143835, + "tokens_seen": 1815478272 + }, + { + "epoch": 0.55, + "learning_rate": 0.0003636349047141424, + "loss": 1.3785, + "theoretical_loss": 3.486914494056217, + "tokens_seen": 1815543808 + }, + { + "epoch": 0.55, + "learning_rate": 0.0003636188565697091, + "loss": 1.5092, + "theoretical_loss": 3.4869038862881867, + "tokens_seen": 1815609344 + }, + { + "epoch": 0.55, + "learning_rate": 0.0003636028084252758, + "loss": 1.3801, + "theoretical_loss": 3.4868932790102534, + "tokens_seen": 1815674880 + }, + { + "epoch": 0.55, + "learning_rate": 0.0003635867602808425, + "loss": 1.3938, + "theoretical_loss": 3.4868826722223756, + "tokens_seen": 1815740416 + }, + { + "epoch": 0.55, + "learning_rate": 0.0003635707121364092, + "loss": 1.2151, + "theoretical_loss": 3.486872065924514, + "tokens_seen": 1815805952 + }, + { + "epoch": 0.55, + "learning_rate": 0.0003635546639919759, + "loss": 1.2722, + "theoretical_loss": 3.486861460116627, + "tokens_seen": 1815871488 + }, + { + "epoch": 0.55, + "learning_rate": 0.0003635386158475426, + "loss": 1.3468, + "theoretical_loss": 3.4868508547986754, + "tokens_seen": 1815937024 + }, + { + "epoch": 0.55, + "learning_rate": 0.0003635225677031093, + "loss": 1.3957, + "theoretical_loss": 3.486840249970619, + "tokens_seen": 1816002560 + }, + { + "epoch": 0.55, + "learning_rate": 0.000363506519558676, + "loss": 1.3503, + "theoretical_loss": 3.4868296456324166, + "tokens_seen": 1816068096 + }, + { + "epoch": 0.55, + "learning_rate": 0.0003634904714142427, + "loss": 1.4984, + "theoretical_loss": 3.4868190417840284, + "tokens_seen": 1816133632 + }, + { + "epoch": 0.55, + "learning_rate": 0.0003634744232698094, + "loss": 1.3724, + "theoretical_loss": 3.486808438425414, + "tokens_seen": 1816199168 + }, + { + "epoch": 0.55, + "learning_rate": 0.0003634583751253761, + "loss": 1.3212, + "theoretical_loss": 3.4867978355565334, + "tokens_seen": 1816264704 + }, + { + "epoch": 0.55, + "learning_rate": 0.0003634423269809428, + "loss": 1.2733, + "theoretical_loss": 3.4867872331773455, + "tokens_seen": 1816330240 + }, + { + "epoch": 0.55, + "learning_rate": 0.0003634262788365095, + "loss": 1.2544, + "theoretical_loss": 3.4867766312878112, + "tokens_seen": 1816395776 + }, + { + "epoch": 0.55, + "learning_rate": 0.0003634102306920762, + "loss": 1.3136, + "theoretical_loss": 3.486766029887889, + "tokens_seen": 1816461312 + }, + { + "epoch": 0.55, + "learning_rate": 0.0003633941825476429, + "loss": 1.3253, + "theoretical_loss": 3.4867554289775398, + "tokens_seen": 1816526848 + }, + { + "epoch": 0.55, + "learning_rate": 0.0003633781344032096, + "loss": 1.3697, + "theoretical_loss": 3.4867448285567226, + "tokens_seen": 1816592384 + }, + { + "epoch": 0.55, + "learning_rate": 0.0003633620862587764, + "loss": 1.3705, + "theoretical_loss": 3.486734228625397, + "tokens_seen": 1816657920 + }, + { + "epoch": 0.55, + "learning_rate": 0.0003633460381143431, + "loss": 1.3398, + "theoretical_loss": 3.486723629183523, + "tokens_seen": 1816723456 + }, + { + "epoch": 0.55, + "learning_rate": 0.0003633299899699098, + "loss": 1.2872, + "theoretical_loss": 3.4867130302310607, + "tokens_seen": 1816788992 + }, + { + "epoch": 0.55, + "learning_rate": 0.0003633139418254765, + "loss": 1.391, + "theoretical_loss": 3.486702431767969, + "tokens_seen": 1816854528 + }, + { + "epoch": 0.55, + "learning_rate": 0.0003632978936810432, + "loss": 1.3835, + "theoretical_loss": 3.4866918337942088, + "tokens_seen": 1816920064 + }, + { + "epoch": 0.55, + "learning_rate": 0.0003632818455366099, + "loss": 1.3452, + "theoretical_loss": 3.486681236309739, + "tokens_seen": 1816985600 + }, + { + "epoch": 0.55, + "learning_rate": 0.0003632657973921766, + "loss": 1.3176, + "theoretical_loss": 3.4866706393145197, + "tokens_seen": 1817051136 + }, + { + "epoch": 0.55, + "learning_rate": 0.0003632497492477433, + "loss": 1.326, + "theoretical_loss": 3.48666004280851, + "tokens_seen": 1817116672 + }, + { + "epoch": 0.55, + "learning_rate": 0.00036323370110331, + "loss": 1.3048, + "theoretical_loss": 3.4866494467916707, + "tokens_seen": 1817182208 + }, + { + "epoch": 0.55, + "learning_rate": 0.0003632176529588767, + "loss": 1.3152, + "theoretical_loss": 3.486638851263961, + "tokens_seen": 1817247744 + }, + { + "epoch": 0.55, + "learning_rate": 0.0003632016048144434, + "loss": 1.4263, + "theoretical_loss": 3.486628256225341, + "tokens_seen": 1817313280 + }, + { + "epoch": 0.55, + "learning_rate": 0.0003631855566700101, + "loss": 1.2705, + "theoretical_loss": 3.48661766167577, + "tokens_seen": 1817378816 + }, + { + "epoch": 0.55, + "learning_rate": 0.0003631695085255768, + "loss": 1.361, + "theoretical_loss": 3.486607067615208, + "tokens_seen": 1817444352 + }, + { + "epoch": 0.55, + "learning_rate": 0.0003631534603811435, + "loss": 1.3616, + "theoretical_loss": 3.486596474043615, + "tokens_seen": 1817509888 + }, + { + "epoch": 0.55, + "learning_rate": 0.0003631374122367102, + "loss": 1.3164, + "theoretical_loss": 3.4865858809609507, + "tokens_seen": 1817575424 + }, + { + "epoch": 0.55, + "learning_rate": 0.0003631213640922769, + "loss": 1.4273, + "theoretical_loss": 3.486575288367175, + "tokens_seen": 1817640960 + }, + { + "epoch": 0.55, + "learning_rate": 0.0003631053159478436, + "loss": 1.3527, + "theoretical_loss": 3.4865646962622474, + "tokens_seen": 1817706496 + }, + { + "epoch": 0.55, + "learning_rate": 0.0003630892678034103, + "loss": 1.467, + "theoretical_loss": 3.486554104646128, + "tokens_seen": 1817772032 + }, + { + "epoch": 0.55, + "learning_rate": 0.000363073219658977, + "loss": 1.349, + "theoretical_loss": 3.486543513518776, + "tokens_seen": 1817837568 + }, + { + "epoch": 0.55, + "learning_rate": 0.00036305717151454363, + "loss": 1.3676, + "theoretical_loss": 3.4865329228801523, + "tokens_seen": 1817903104 + }, + { + "epoch": 0.55, + "learning_rate": 0.00036304112337011033, + "loss": 1.323, + "theoretical_loss": 3.486522332730216, + "tokens_seen": 1817968640 + }, + { + "epoch": 0.55, + "learning_rate": 0.00036302507522567703, + "loss": 1.2381, + "theoretical_loss": 3.4865117430689274, + "tokens_seen": 1818034176 + }, + { + "epoch": 0.55, + "learning_rate": 0.00036300902708124374, + "loss": 1.3842, + "theoretical_loss": 3.4865011538962456, + "tokens_seen": 1818099712 + }, + { + "epoch": 0.55, + "learning_rate": 0.00036299297893681044, + "loss": 1.3926, + "theoretical_loss": 3.486490565212131, + "tokens_seen": 1818165248 + }, + { + "epoch": 0.55, + "learning_rate": 0.00036297693079237714, + "loss": 1.2345, + "theoretical_loss": 3.486479977016544, + "tokens_seen": 1818230784 + }, + { + "epoch": 0.55, + "learning_rate": 0.00036296088264794384, + "loss": 1.2939, + "theoretical_loss": 3.486469389309443, + "tokens_seen": 1818296320 + }, + { + "epoch": 0.55, + "learning_rate": 0.00036294483450351054, + "loss": 1.3575, + "theoretical_loss": 3.486458802090789, + "tokens_seen": 1818361856 + }, + { + "epoch": 0.55, + "learning_rate": 0.00036292878635907724, + "loss": 1.4006, + "theoretical_loss": 3.4864482153605416, + "tokens_seen": 1818427392 + }, + { + "epoch": 0.55, + "learning_rate": 0.00036291273821464394, + "loss": 1.4378, + "theoretical_loss": 3.4864376291186607, + "tokens_seen": 1818492928 + }, + { + "epoch": 0.55, + "learning_rate": 0.00036289669007021064, + "loss": 1.2493, + "theoretical_loss": 3.4864270433651052, + "tokens_seen": 1818558464 + }, + { + "epoch": 0.55, + "objective/train/docs_used": 663450, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.3421754837036133, + "objective/train/theoretical_loss": 3.486416458099837, + "objective/train/tokens_used": 1839084000, + "theoretical_loss": 3.486416458099837, + "tokens_seen": 1818624000 + }, + { + "epoch": 0.55, + "learning_rate": 0.00036288064192577734, + "loss": 1.3546, + "theoretical_loss": 3.486416458099837, + "tokens_seen": 1818624000 + }, + { + "epoch": 0.55, + "learning_rate": 0.00036286459378134404, + "loss": 1.3871, + "theoretical_loss": 3.486405873322814, + "tokens_seen": 1818689536 + }, + { + "epoch": 0.55, + "learning_rate": 0.00036284854563691074, + "loss": 1.3753, + "theoretical_loss": 3.4863952890339975, + "tokens_seen": 1818755072 + }, + { + "epoch": 0.55, + "learning_rate": 0.00036283249749247744, + "loss": 1.3349, + "theoretical_loss": 3.486384705233346, + "tokens_seen": 1818820608 + }, + { + "epoch": 0.55, + "learning_rate": 0.00036281644934804414, + "loss": 1.32, + "theoretical_loss": 3.4863741219208206, + "tokens_seen": 1818886144 + }, + { + "epoch": 0.55, + "learning_rate": 0.00036280040120361084, + "loss": 1.3831, + "theoretical_loss": 3.486363539096381, + "tokens_seen": 1818951680 + }, + { + "epoch": 0.55, + "learning_rate": 0.00036278435305917754, + "loss": 1.3655, + "theoretical_loss": 3.486352956759987, + "tokens_seen": 1819017216 + }, + { + "epoch": 0.55, + "learning_rate": 0.00036276830491474424, + "loss": 1.2957, + "theoretical_loss": 3.4863423749115983, + "tokens_seen": 1819082752 + }, + { + "epoch": 0.55, + "learning_rate": 0.00036275225677031095, + "loss": 1.3002, + "theoretical_loss": 3.486331793551175, + "tokens_seen": 1819148288 + }, + { + "epoch": 0.55, + "learning_rate": 0.00036273620862587765, + "loss": 1.3658, + "theoretical_loss": 3.4863212126786762, + "tokens_seen": 1819213824 + }, + { + "epoch": 0.55, + "learning_rate": 0.00036272016048144435, + "loss": 1.3769, + "theoretical_loss": 3.4863106322940633, + "tokens_seen": 1819279360 + }, + { + "epoch": 0.55, + "learning_rate": 0.00036270411233701105, + "loss": 1.4097, + "theoretical_loss": 3.4863000523972953, + "tokens_seen": 1819344896 + }, + { + "epoch": 0.55, + "learning_rate": 0.00036268806419257775, + "loss": 1.3957, + "theoretical_loss": 3.4862894729883322, + "tokens_seen": 1819410432 + }, + { + "epoch": 0.55, + "learning_rate": 0.00036267201604814445, + "loss": 1.3832, + "theoretical_loss": 3.4862788940671345, + "tokens_seen": 1819475968 + }, + { + "epoch": 0.55, + "learning_rate": 0.00036265596790371115, + "loss": 1.3782, + "theoretical_loss": 3.4862683156336614, + "tokens_seen": 1819541504 + }, + { + "epoch": 0.55, + "learning_rate": 0.00036263991975927785, + "loss": 1.3336, + "theoretical_loss": 3.486257737687873, + "tokens_seen": 1819607040 + }, + { + "epoch": 0.55, + "learning_rate": 0.00036262387161484455, + "loss": 1.3223, + "theoretical_loss": 3.4862471602297296, + "tokens_seen": 1819672576 + }, + { + "epoch": 0.55, + "learning_rate": 0.00036260782347041125, + "loss": 1.3768, + "theoretical_loss": 3.486236583259191, + "tokens_seen": 1819738112 + }, + { + "epoch": 0.55, + "learning_rate": 0.00036259177532597795, + "loss": 1.3056, + "theoretical_loss": 3.4862260067762176, + "tokens_seen": 1819803648 + }, + { + "epoch": 0.55, + "learning_rate": 0.00036257572718154465, + "loss": 1.2851, + "theoretical_loss": 3.486215430780769, + "tokens_seen": 1819869184 + }, + { + "epoch": 0.55, + "learning_rate": 0.00036255967903711135, + "loss": 1.4769, + "theoretical_loss": 3.486204855272804, + "tokens_seen": 1819934720 + }, + { + "epoch": 0.55, + "learning_rate": 0.00036254363089267805, + "loss": 1.2787, + "theoretical_loss": 3.4861942802522843, + "tokens_seen": 1820000256 + }, + { + "epoch": 0.55, + "learning_rate": 0.00036252758274824475, + "loss": 1.3184, + "theoretical_loss": 3.486183705719169, + "tokens_seen": 1820065792 + }, + { + "epoch": 0.55, + "learning_rate": 0.00036251153460381145, + "loss": 1.4453, + "theoretical_loss": 3.486173131673419, + "tokens_seen": 1820131328 + }, + { + "epoch": 0.55, + "learning_rate": 0.00036249548645937816, + "loss": 1.2826, + "theoretical_loss": 3.4861625581149926, + "tokens_seen": 1820196864 + }, + { + "epoch": 0.55, + "learning_rate": 0.00036247943831494486, + "loss": 1.2336, + "theoretical_loss": 3.4861519850438514, + "tokens_seen": 1820262400 + }, + { + "epoch": 0.55, + "learning_rate": 0.00036246339017051156, + "loss": 1.4404, + "theoretical_loss": 3.486141412459955, + "tokens_seen": 1820327936 + }, + { + "epoch": 0.55, + "learning_rate": 0.00036244734202607826, + "loss": 1.1907, + "theoretical_loss": 3.486130840363263, + "tokens_seen": 1820393472 + }, + { + "epoch": 0.55, + "learning_rate": 0.00036243129388164496, + "loss": 1.3732, + "theoretical_loss": 3.4861202687537354, + "tokens_seen": 1820459008 + }, + { + "epoch": 0.55, + "learning_rate": 0.00036241524573721166, + "loss": 1.3492, + "theoretical_loss": 3.486109697631333, + "tokens_seen": 1820524544 + }, + { + "epoch": 0.55, + "learning_rate": 0.00036239919759277836, + "loss": 1.4367, + "theoretical_loss": 3.486099126996015, + "tokens_seen": 1820590080 + }, + { + "epoch": 0.55, + "learning_rate": 0.00036238314944834506, + "loss": 1.5474, + "theoretical_loss": 3.4860885568477413, + "tokens_seen": 1820655616 + }, + { + "epoch": 0.55, + "learning_rate": 0.00036236710130391176, + "loss": 1.3375, + "theoretical_loss": 3.486077987186473, + "tokens_seen": 1820721152 + }, + { + "epoch": 0.55, + "learning_rate": 0.00036235105315947846, + "loss": 1.3493, + "theoretical_loss": 3.486067418012169, + "tokens_seen": 1820786688 + }, + { + "epoch": 0.55, + "learning_rate": 0.00036233500501504516, + "loss": 1.299, + "theoretical_loss": 3.4860568493247897, + "tokens_seen": 1820852224 + }, + { + "epoch": 0.55, + "learning_rate": 0.00036231895687061186, + "loss": 1.3589, + "theoretical_loss": 3.4860462811242954, + "tokens_seen": 1820917760 + }, + { + "epoch": 0.55, + "learning_rate": 0.00036230290872617856, + "loss": 1.3805, + "theoretical_loss": 3.4860357134106463, + "tokens_seen": 1820983296 + }, + { + "epoch": 0.55, + "learning_rate": 0.0003622868605817452, + "loss": 1.3383, + "theoretical_loss": 3.4860251461838017, + "tokens_seen": 1821048832 + }, + { + "epoch": 0.55, + "learning_rate": 0.0003622708124373119, + "loss": 1.5006, + "theoretical_loss": 3.4860145794437223, + "tokens_seen": 1821114368 + }, + { + "epoch": 0.55, + "learning_rate": 0.0003622547642928786, + "loss": 1.3751, + "theoretical_loss": 3.486004013190368, + "tokens_seen": 1821179904 + }, + { + "epoch": 0.55, + "learning_rate": 0.0003622387161484453, + "loss": 1.3878, + "theoretical_loss": 3.4859934474236987, + "tokens_seen": 1821245440 + }, + { + "epoch": 0.55, + "learning_rate": 0.000362222668004012, + "loss": 1.354, + "theoretical_loss": 3.4859828821436745, + "tokens_seen": 1821310976 + }, + { + "epoch": 0.55, + "learning_rate": 0.0003622066198595787, + "loss": 1.4532, + "theoretical_loss": 3.4859723173502557, + "tokens_seen": 1821376512 + }, + { + "epoch": 0.55, + "learning_rate": 0.0003621905717151454, + "loss": 1.327, + "theoretical_loss": 3.4859617530434024, + "tokens_seen": 1821442048 + }, + { + "epoch": 0.55, + "learning_rate": 0.0003621745235707121, + "loss": 1.42, + "theoretical_loss": 3.485951189223074, + "tokens_seen": 1821507584 + }, + { + "epoch": 0.55, + "learning_rate": 0.00036215847542627887, + "loss": 1.3605, + "theoretical_loss": 3.4859406258892314, + "tokens_seen": 1821573120 + }, + { + "epoch": 0.55, + "learning_rate": 0.00036214242728184557, + "loss": 1.2305, + "theoretical_loss": 3.4859300630418346, + "tokens_seen": 1821638656 + }, + { + "epoch": 0.55, + "learning_rate": 0.00036212637913741227, + "loss": 1.2696, + "theoretical_loss": 3.4859195006808434, + "tokens_seen": 1821704192 + }, + { + "epoch": 0.55, + "learning_rate": 0.00036211033099297897, + "loss": 1.3914, + "theoretical_loss": 3.485908938806218, + "tokens_seen": 1821769728 + }, + { + "epoch": 0.55, + "learning_rate": 0.00036209428284854567, + "loss": 1.3876, + "theoretical_loss": 3.485898377417919, + "tokens_seen": 1821835264 + }, + { + "epoch": 0.55, + "objective/train/docs_used": 664649, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.2098219394683838, + "objective/train/theoretical_loss": 3.4858878165159055, + "objective/train/tokens_used": 1842360800, + "theoretical_loss": 3.4858878165159055, + "tokens_seen": 1821900800 + }, + { + "epoch": 0.55, + "learning_rate": 0.00036207823470411237, + "loss": 1.2878, + "theoretical_loss": 3.4858878165159055, + "tokens_seen": 1821900800 + }, + { + "epoch": 0.55, + "learning_rate": 0.00036206218655967907, + "loss": 1.267, + "theoretical_loss": 3.485877256100138, + "tokens_seen": 1821966336 + }, + { + "epoch": 0.55, + "learning_rate": 0.0003620461384152458, + "loss": 1.4073, + "theoretical_loss": 3.4858666961705773, + "tokens_seen": 1822031872 + }, + { + "epoch": 0.55, + "learning_rate": 0.0003620300902708125, + "loss": 1.4118, + "theoretical_loss": 3.4858561367271825, + "tokens_seen": 1822097408 + }, + { + "epoch": 0.55, + "learning_rate": 0.0003620140421263792, + "loss": 1.3909, + "theoretical_loss": 3.485845577769915, + "tokens_seen": 1822162944 + }, + { + "epoch": 0.55, + "learning_rate": 0.0003619979939819459, + "loss": 1.511, + "theoretical_loss": 3.4858350192987335, + "tokens_seen": 1822228480 + }, + { + "epoch": 0.55, + "learning_rate": 0.0003619819458375126, + "loss": 1.4548, + "theoretical_loss": 3.485824461313599, + "tokens_seen": 1822294016 + }, + { + "epoch": 0.55, + "learning_rate": 0.0003619658976930793, + "loss": 1.3202, + "theoretical_loss": 3.4858139038144715, + "tokens_seen": 1822359552 + }, + { + "epoch": 0.55, + "learning_rate": 0.000361949849548646, + "loss": 1.3718, + "theoretical_loss": 3.485803346801311, + "tokens_seen": 1822425088 + }, + { + "epoch": 0.55, + "learning_rate": 0.0003619338014042127, + "loss": 1.3747, + "theoretical_loss": 3.485792790274078, + "tokens_seen": 1822490624 + }, + { + "epoch": 0.55, + "learning_rate": 0.0003619177532597794, + "loss": 1.3658, + "theoretical_loss": 3.4857822342327323, + "tokens_seen": 1822556160 + }, + { + "epoch": 0.55, + "learning_rate": 0.0003619017051153461, + "loss": 1.2253, + "theoretical_loss": 3.4857716786772346, + "tokens_seen": 1822621696 + }, + { + "epoch": 0.55, + "learning_rate": 0.0003618856569709128, + "loss": 1.4676, + "theoretical_loss": 3.4857611236075443, + "tokens_seen": 1822687232 + }, + { + "epoch": 0.55, + "learning_rate": 0.0003618696088264795, + "loss": 1.4349, + "theoretical_loss": 3.4857505690236223, + "tokens_seen": 1822752768 + }, + { + "epoch": 0.55, + "learning_rate": 0.0003618535606820462, + "loss": 1.3549, + "theoretical_loss": 3.485740014925428, + "tokens_seen": 1822818304 + }, + { + "epoch": 0.55, + "learning_rate": 0.0003618375125376129, + "loss": 1.3275, + "theoretical_loss": 3.485729461312922, + "tokens_seen": 1822883840 + }, + { + "epoch": 0.55, + "learning_rate": 0.0003618214643931796, + "loss": 1.4069, + "theoretical_loss": 3.4857189081860644, + "tokens_seen": 1822949376 + }, + { + "epoch": 0.55, + "learning_rate": 0.0003618054162487463, + "loss": 1.36, + "theoretical_loss": 3.485708355544816, + "tokens_seen": 1823014912 + }, + { + "epoch": 0.55, + "learning_rate": 0.000361789368104313, + "loss": 1.3591, + "theoretical_loss": 3.4856978033891366, + "tokens_seen": 1823080448 + }, + { + "epoch": 0.55, + "learning_rate": 0.0003617733199598797, + "loss": 1.2459, + "theoretical_loss": 3.4856872517189856, + "tokens_seen": 1823145984 + }, + { + "epoch": 0.55, + "learning_rate": 0.0003617572718154464, + "loss": 1.2851, + "theoretical_loss": 3.4856767005343245, + "tokens_seen": 1823211520 + }, + { + "epoch": 0.55, + "learning_rate": 0.0003617412236710131, + "loss": 1.4853, + "theoretical_loss": 3.4856661498351125, + "tokens_seen": 1823277056 + }, + { + "epoch": 0.55, + "learning_rate": 0.0003617251755265798, + "loss": 1.4275, + "theoretical_loss": 3.4856555996213103, + "tokens_seen": 1823342592 + }, + { + "epoch": 0.55, + "learning_rate": 0.0003617091273821465, + "loss": 1.4541, + "theoretical_loss": 3.485645049892878, + "tokens_seen": 1823408128 + }, + { + "epoch": 0.55, + "learning_rate": 0.0003616930792377132, + "loss": 1.4302, + "theoretical_loss": 3.485634500649777, + "tokens_seen": 1823473664 + }, + { + "epoch": 0.55, + "learning_rate": 0.0003616770310932799, + "loss": 1.3086, + "theoretical_loss": 3.4856239518919647, + "tokens_seen": 1823539200 + }, + { + "epoch": 0.55, + "learning_rate": 0.00036166098294884653, + "loss": 1.4426, + "theoretical_loss": 3.4856134036194044, + "tokens_seen": 1823604736 + }, + { + "epoch": 0.55, + "learning_rate": 0.00036164493480441323, + "loss": 1.4155, + "theoretical_loss": 3.485602855832054, + "tokens_seen": 1823670272 + }, + { + "epoch": 0.55, + "learning_rate": 0.00036162888665997994, + "loss": 1.3122, + "theoretical_loss": 3.4855923085298754, + "tokens_seen": 1823735808 + }, + { + "epoch": 0.55, + "learning_rate": 0.00036161283851554664, + "loss": 1.4086, + "theoretical_loss": 3.4855817617128277, + "tokens_seen": 1823801344 + }, + { + "epoch": 0.55, + "learning_rate": 0.00036159679037111334, + "loss": 1.3804, + "theoretical_loss": 3.4855712153808724, + "tokens_seen": 1823866880 + }, + { + "epoch": 0.55, + "learning_rate": 0.00036158074222668004, + "loss": 1.3253, + "theoretical_loss": 3.4855606695339683, + "tokens_seen": 1823932416 + }, + { + "epoch": 0.55, + "learning_rate": 0.00036156469408224674, + "loss": 1.3013, + "theoretical_loss": 3.485550124172077, + "tokens_seen": 1823997952 + }, + { + "epoch": 0.55, + "learning_rate": 0.00036154864593781344, + "loss": 1.3901, + "theoretical_loss": 3.4855395792951573, + "tokens_seen": 1824063488 + }, + { + "epoch": 0.55, + "learning_rate": 0.00036153259779338014, + "loss": 1.3932, + "theoretical_loss": 3.4855290349031707, + "tokens_seen": 1824129024 + }, + { + "epoch": 0.55, + "learning_rate": 0.00036151654964894684, + "loss": 1.3707, + "theoretical_loss": 3.4855184909960775, + "tokens_seen": 1824194560 + }, + { + "epoch": 0.55, + "learning_rate": 0.00036150050150451354, + "loss": 1.4263, + "theoretical_loss": 3.4855079475738373, + "tokens_seen": 1824260096 + }, + { + "epoch": 0.55, + "learning_rate": 0.00036148445336008024, + "loss": 1.1592, + "theoretical_loss": 3.4854974046364102, + "tokens_seen": 1824325632 + }, + { + "epoch": 0.55, + "learning_rate": 0.00036146840521564694, + "loss": 1.4077, + "theoretical_loss": 3.485486862183757, + "tokens_seen": 1824391168 + }, + { + "epoch": 0.55, + "learning_rate": 0.00036145235707121364, + "loss": 1.4906, + "theoretical_loss": 3.4854763202158385, + "tokens_seen": 1824456704 + }, + { + "epoch": 0.55, + "learning_rate": 0.00036143630892678034, + "loss": 1.3635, + "theoretical_loss": 3.485465778732614, + "tokens_seen": 1824522240 + }, + { + "epoch": 0.55, + "learning_rate": 0.00036142026078234704, + "loss": 1.3964, + "theoretical_loss": 3.485455237734045, + "tokens_seen": 1824587776 + }, + { + "epoch": 0.55, + "learning_rate": 0.00036140421263791374, + "loss": 1.2847, + "theoretical_loss": 3.4854446972200903, + "tokens_seen": 1824653312 + }, + { + "epoch": 0.55, + "learning_rate": 0.00036138816449348044, + "loss": 1.3522, + "theoretical_loss": 3.485434157190711, + "tokens_seen": 1824718848 + }, + { + "epoch": 0.55, + "learning_rate": 0.00036137211634904715, + "loss": 1.2144, + "theoretical_loss": 3.4854236176458677, + "tokens_seen": 1824784384 + }, + { + "epoch": 0.55, + "learning_rate": 0.00036135606820461385, + "loss": 1.3556, + "theoretical_loss": 3.4854130785855206, + "tokens_seen": 1824849920 + }, + { + "epoch": 0.55, + "learning_rate": 0.00036134002006018055, + "loss": 1.2736, + "theoretical_loss": 3.4854025400096296, + "tokens_seen": 1824915456 + }, + { + "epoch": 0.55, + "learning_rate": 0.00036132397191574725, + "loss": 1.3783, + "theoretical_loss": 3.485392001918155, + "tokens_seen": 1824980992 + }, + { + "epoch": 0.55, + "learning_rate": 0.00036130792377131395, + "loss": 1.3065, + "theoretical_loss": 3.485381464311058, + "tokens_seen": 1825046528 + }, + { + "epoch": 0.55, + "learning_rate": 0.00036129187562688065, + "loss": 1.2828, + "theoretical_loss": 3.485370927188298, + "tokens_seen": 1825112064 + }, + { + "epoch": 0.55, + "objective/train/docs_used": 665640, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.2097499370574951, + "objective/train/theoretical_loss": 3.4853603905498356, + "objective/train/tokens_used": 1845637600, + "theoretical_loss": 3.4853603905498356, + "tokens_seen": 1825177600 + }, + { + "epoch": 0.55, + "learning_rate": 0.00036127582748244735, + "loss": 1.3397, + "theoretical_loss": 3.4853603905498356, + "tokens_seen": 1825177600 + }, + { + "epoch": 0.55, + "learning_rate": 0.00036125977933801405, + "loss": 1.3892, + "theoretical_loss": 3.485349854395632, + "tokens_seen": 1825243136 + }, + { + "epoch": 0.55, + "learning_rate": 0.00036124373119358075, + "loss": 1.4205, + "theoretical_loss": 3.4853393187256456, + "tokens_seen": 1825308672 + }, + { + "epoch": 0.55, + "learning_rate": 0.00036122768304914745, + "loss": 1.4524, + "theoretical_loss": 3.4853287835398388, + "tokens_seen": 1825374208 + }, + { + "epoch": 0.55, + "learning_rate": 0.00036121163490471415, + "loss": 1.3945, + "theoretical_loss": 3.485318248838171, + "tokens_seen": 1825439744 + }, + { + "epoch": 0.55, + "learning_rate": 0.00036119558676028085, + "loss": 1.2343, + "theoretical_loss": 3.485307714620603, + "tokens_seen": 1825505280 + }, + { + "epoch": 0.55, + "learning_rate": 0.00036117953861584755, + "loss": 1.3361, + "theoretical_loss": 3.4852971808870947, + "tokens_seen": 1825570816 + }, + { + "epoch": 0.55, + "learning_rate": 0.00036116349047141425, + "loss": 1.3435, + "theoretical_loss": 3.4852866476376065, + "tokens_seen": 1825636352 + }, + { + "epoch": 0.55, + "learning_rate": 0.00036114744232698095, + "loss": 1.3359, + "theoretical_loss": 3.485276114872099, + "tokens_seen": 1825701888 + }, + { + "epoch": 0.55, + "learning_rate": 0.00036113139418254766, + "loss": 1.3326, + "theoretical_loss": 3.4852655825905328, + "tokens_seen": 1825767424 + }, + { + "epoch": 0.55, + "learning_rate": 0.00036111534603811436, + "loss": 1.415, + "theoretical_loss": 3.485255050792868, + "tokens_seen": 1825832960 + }, + { + "epoch": 0.55, + "learning_rate": 0.00036109929789368106, + "loss": 1.353, + "theoretical_loss": 3.485244519479065, + "tokens_seen": 1825898496 + }, + { + "epoch": 0.55, + "learning_rate": 0.00036108324974924776, + "loss": 1.3976, + "theoretical_loss": 3.4852339886490844, + "tokens_seen": 1825964032 + }, + { + "epoch": 0.55, + "learning_rate": 0.00036106720160481446, + "loss": 1.4183, + "theoretical_loss": 3.4852234583028863, + "tokens_seen": 1826029568 + }, + { + "epoch": 0.55, + "learning_rate": 0.00036105115346038116, + "loss": 1.3423, + "theoretical_loss": 3.485212928440431, + "tokens_seen": 1826095104 + }, + { + "epoch": 0.55, + "learning_rate": 0.00036103510531594786, + "loss": 1.3113, + "theoretical_loss": 3.48520239906168, + "tokens_seen": 1826160640 + }, + { + "epoch": 0.55, + "learning_rate": 0.00036101905717151456, + "loss": 1.3801, + "theoretical_loss": 3.4851918701665925, + "tokens_seen": 1826226176 + }, + { + "epoch": 0.55, + "learning_rate": 0.00036100300902708126, + "loss": 1.5092, + "theoretical_loss": 3.485181341755129, + "tokens_seen": 1826291712 + }, + { + "epoch": 0.55, + "learning_rate": 0.00036098696088264796, + "loss": 1.3518, + "theoretical_loss": 3.4851708138272506, + "tokens_seen": 1826357248 + }, + { + "epoch": 0.55, + "learning_rate": 0.00036097091273821466, + "loss": 1.3202, + "theoretical_loss": 3.4851602863829174, + "tokens_seen": 1826422784 + }, + { + "epoch": 0.55, + "learning_rate": 0.00036095486459378136, + "loss": 1.3077, + "theoretical_loss": 3.48514975942209, + "tokens_seen": 1826488320 + }, + { + "epoch": 0.55, + "learning_rate": 0.00036093881644934806, + "loss": 1.318, + "theoretical_loss": 3.4851392329447286, + "tokens_seen": 1826553856 + }, + { + "epoch": 0.55, + "learning_rate": 0.00036092276830491476, + "loss": 1.3138, + "theoretical_loss": 3.4851287069507935, + "tokens_seen": 1826619392 + }, + { + "epoch": 0.55, + "learning_rate": 0.00036090672016048146, + "loss": 1.4136, + "theoretical_loss": 3.485118181440246, + "tokens_seen": 1826684928 + }, + { + "epoch": 0.55, + "learning_rate": 0.00036089067201604816, + "loss": 1.1416, + "theoretical_loss": 3.4851076564130454, + "tokens_seen": 1826750464 + }, + { + "epoch": 0.55, + "learning_rate": 0.00036087462387161487, + "loss": 1.2883, + "theoretical_loss": 3.485097131869153, + "tokens_seen": 1826816000 + }, + { + "epoch": 0.55, + "learning_rate": 0.00036085857572718157, + "loss": 1.3585, + "theoretical_loss": 3.4850866078085287, + "tokens_seen": 1826881536 + }, + { + "epoch": 0.55, + "learning_rate": 0.00036084252758274827, + "loss": 1.2632, + "theoretical_loss": 3.485076084231134, + "tokens_seen": 1826947072 + }, + { + "epoch": 0.55, + "learning_rate": 0.00036082647943831497, + "loss": 1.4349, + "theoretical_loss": 3.485065561136928, + "tokens_seen": 1827012608 + }, + { + "epoch": 0.55, + "learning_rate": 0.00036081043129388167, + "loss": 1.292, + "theoretical_loss": 3.4850550385258723, + "tokens_seen": 1827078144 + }, + { + "epoch": 0.55, + "learning_rate": 0.00036079438314944837, + "loss": 1.2639, + "theoretical_loss": 3.4850445163979265, + "tokens_seen": 1827143680 + }, + { + "epoch": 0.55, + "learning_rate": 0.00036077833500501507, + "loss": 1.3413, + "theoretical_loss": 3.485033994753052, + "tokens_seen": 1827209216 + }, + { + "epoch": 0.55, + "learning_rate": 0.00036076228686058177, + "loss": 1.4233, + "theoretical_loss": 3.4850234735912085, + "tokens_seen": 1827274752 + }, + { + "epoch": 0.55, + "learning_rate": 0.00036074623871614847, + "loss": 1.3558, + "theoretical_loss": 3.4850129529123564, + "tokens_seen": 1827340288 + }, + { + "epoch": 0.55, + "learning_rate": 0.00036073019057171517, + "loss": 1.2079, + "theoretical_loss": 3.4850024327164575, + "tokens_seen": 1827405824 + }, + { + "epoch": 0.55, + "learning_rate": 0.00036071414242728187, + "loss": 1.2922, + "theoretical_loss": 3.484991913003471, + "tokens_seen": 1827471360 + }, + { + "epoch": 0.55, + "learning_rate": 0.00036069809428284857, + "loss": 1.4436, + "theoretical_loss": 3.484981393773358, + "tokens_seen": 1827536896 + }, + { + "epoch": 0.55, + "learning_rate": 0.00036068204613841527, + "loss": 1.4214, + "theoretical_loss": 3.484970875026079, + "tokens_seen": 1827602432 + }, + { + "epoch": 0.55, + "learning_rate": 0.000360665997993982, + "loss": 1.4078, + "theoretical_loss": 3.4849603567615937, + "tokens_seen": 1827667968 + }, + { + "epoch": 0.55, + "learning_rate": 0.0003606499498495487, + "loss": 1.4517, + "theoretical_loss": 3.4849498389798637, + "tokens_seen": 1827733504 + }, + { + "epoch": 0.55, + "learning_rate": 0.0003606339017051154, + "loss": 1.4436, + "theoretical_loss": 3.4849393216808497, + "tokens_seen": 1827799040 + }, + { + "epoch": 0.55, + "learning_rate": 0.0003606178535606821, + "loss": 1.3767, + "theoretical_loss": 3.484928804864511, + "tokens_seen": 1827864576 + }, + { + "epoch": 0.55, + "learning_rate": 0.0003606018054162488, + "loss": 1.3351, + "theoretical_loss": 3.4849182885308094, + "tokens_seen": 1827930112 + }, + { + "epoch": 0.55, + "learning_rate": 0.0003605857572718155, + "loss": 1.3161, + "theoretical_loss": 3.484907772679705, + "tokens_seen": 1827995648 + }, + { + "epoch": 0.55, + "learning_rate": 0.0003605697091273822, + "loss": 1.2093, + "theoretical_loss": 3.4848972573111574, + "tokens_seen": 1828061184 + }, + { + "epoch": 0.55, + "learning_rate": 0.0003605536609829489, + "loss": 1.4068, + "theoretical_loss": 3.484886742425129, + "tokens_seen": 1828126720 + }, + { + "epoch": 0.55, + "learning_rate": 0.0003605376128385156, + "loss": 1.3196, + "theoretical_loss": 3.4848762280215784, + "tokens_seen": 1828192256 + }, + { + "epoch": 0.55, + "learning_rate": 0.0003605215646940823, + "loss": 1.3083, + "theoretical_loss": 3.4848657141004677, + "tokens_seen": 1828257792 + }, + { + "epoch": 0.55, + "learning_rate": 0.000360505516549649, + "loss": 1.4514, + "theoretical_loss": 3.484855200661757, + "tokens_seen": 1828323328 + }, + { + "epoch": 0.55, + "learning_rate": 0.0003604894684052157, + "loss": 1.369, + "theoretical_loss": 3.4848446877054062, + "tokens_seen": 1828388864 + }, + { + "epoch": 0.55, + "objective/train/docs_used": 666768, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.2645268440246582, + "objective/train/theoretical_loss": 3.484834175231377, + "objective/train/tokens_used": 1848914400, + "theoretical_loss": 3.484834175231377, + "tokens_seen": 1828454400 + }, + { + "epoch": 0.55, + "learning_rate": 0.0003604734202607824, + "loss": 1.3602, + "theoretical_loss": 3.484834175231377, + "tokens_seen": 1828454400 + }, + { + "epoch": 0.55, + "learning_rate": 0.0003604573721163491, + "loss": 1.3123, + "theoretical_loss": 3.4848236632396294, + "tokens_seen": 1828519936 + }, + { + "epoch": 0.55, + "learning_rate": 0.0003604413239719158, + "loss": 1.3395, + "theoretical_loss": 3.484813151730124, + "tokens_seen": 1828585472 + }, + { + "epoch": 0.55, + "learning_rate": 0.0003604252758274825, + "loss": 1.403, + "theoretical_loss": 3.484802640702821, + "tokens_seen": 1828651008 + }, + { + "epoch": 0.55, + "learning_rate": 0.0003604092276830492, + "loss": 1.3092, + "theoretical_loss": 3.484792130157682, + "tokens_seen": 1828716544 + }, + { + "epoch": 0.55, + "learning_rate": 0.0003603931795386159, + "loss": 1.2459, + "theoretical_loss": 3.484781620094667, + "tokens_seen": 1828782080 + }, + { + "epoch": 0.55, + "learning_rate": 0.0003603771313941826, + "loss": 1.3506, + "theoretical_loss": 3.4847711105137367, + "tokens_seen": 1828847616 + }, + { + "epoch": 0.55, + "learning_rate": 0.0003603610832497493, + "loss": 1.2601, + "theoretical_loss": 3.4847606014148518, + "tokens_seen": 1828913152 + }, + { + "epoch": 0.55, + "learning_rate": 0.000360345035105316, + "loss": 1.4108, + "theoretical_loss": 3.484750092797972, + "tokens_seen": 1828978688 + }, + { + "epoch": 0.55, + "learning_rate": 0.0003603289869608827, + "loss": 1.3919, + "theoretical_loss": 3.4847395846630596, + "tokens_seen": 1829044224 + }, + { + "epoch": 0.55, + "learning_rate": 0.0003603129388164494, + "loss": 1.3801, + "theoretical_loss": 3.4847290770100736, + "tokens_seen": 1829109760 + }, + { + "epoch": 0.55, + "learning_rate": 0.0003602968906720161, + "loss": 1.2025, + "theoretical_loss": 3.484718569838976, + "tokens_seen": 1829175296 + }, + { + "epoch": 0.55, + "learning_rate": 0.0003602808425275828, + "loss": 1.349, + "theoretical_loss": 3.484708063149726, + "tokens_seen": 1829240832 + }, + { + "epoch": 0.55, + "learning_rate": 0.00036026479438314944, + "loss": 1.2838, + "theoretical_loss": 3.484697556942286, + "tokens_seen": 1829306368 + }, + { + "epoch": 0.55, + "learning_rate": 0.00036024874623871614, + "loss": 1.2825, + "theoretical_loss": 3.484687051216615, + "tokens_seen": 1829371904 + }, + { + "epoch": 0.55, + "learning_rate": 0.00036023269809428284, + "loss": 1.3933, + "theoretical_loss": 3.4846765459726745, + "tokens_seen": 1829437440 + }, + { + "epoch": 0.55, + "learning_rate": 0.00036021664994984954, + "loss": 1.3313, + "theoretical_loss": 3.484666041210425, + "tokens_seen": 1829502976 + }, + { + "epoch": 0.55, + "learning_rate": 0.00036020060180541624, + "loss": 1.4457, + "theoretical_loss": 3.484655536929827, + "tokens_seen": 1829568512 + }, + { + "epoch": 0.55, + "learning_rate": 0.00036018455366098294, + "loss": 1.3814, + "theoretical_loss": 3.4846450331308416, + "tokens_seen": 1829634048 + }, + { + "epoch": 0.55, + "learning_rate": 0.00036016850551654964, + "loss": 1.3014, + "theoretical_loss": 3.484634529813429, + "tokens_seen": 1829699584 + }, + { + "epoch": 0.55, + "learning_rate": 0.00036015245737211634, + "loss": 1.3115, + "theoretical_loss": 3.48462402697755, + "tokens_seen": 1829765120 + }, + { + "epoch": 0.55, + "learning_rate": 0.00036013640922768304, + "loss": 1.2738, + "theoretical_loss": 3.4846135246231658, + "tokens_seen": 1829830656 + }, + { + "epoch": 0.55, + "learning_rate": 0.00036012036108324974, + "loss": 1.5265, + "theoretical_loss": 3.4846030227502363, + "tokens_seen": 1829896192 + }, + { + "epoch": 0.55, + "learning_rate": 0.00036010431293881644, + "loss": 1.361, + "theoretical_loss": 3.484592521358722, + "tokens_seen": 1829961728 + }, + { + "epoch": 0.55, + "learning_rate": 0.00036008826479438314, + "loss": 1.2802, + "theoretical_loss": 3.4845820204485847, + "tokens_seen": 1830027264 + }, + { + "epoch": 0.55, + "learning_rate": 0.00036007221664994984, + "loss": 1.3255, + "theoretical_loss": 3.4845715200197844, + "tokens_seen": 1830092800 + }, + { + "epoch": 0.55, + "learning_rate": 0.00036005616850551654, + "loss": 1.2894, + "theoretical_loss": 3.484561020072282, + "tokens_seen": 1830158336 + }, + { + "epoch": 0.55, + "learning_rate": 0.00036004012036108324, + "loss": 1.3131, + "theoretical_loss": 3.484550520606038, + "tokens_seen": 1830223872 + }, + { + "epoch": 0.55, + "learning_rate": 0.00036002407221664994, + "loss": 1.3942, + "theoretical_loss": 3.484540021621013, + "tokens_seen": 1830289408 + }, + { + "epoch": 0.55, + "learning_rate": 0.00036000802407221665, + "loss": 1.4174, + "theoretical_loss": 3.484529523117168, + "tokens_seen": 1830354944 + }, + { + "epoch": 0.55, + "learning_rate": 0.00035999197592778335, + "loss": 1.3626, + "theoretical_loss": 3.4845190250944635, + "tokens_seen": 1830420480 + }, + { + "epoch": 0.55, + "learning_rate": 0.00035997592778335005, + "loss": 1.3776, + "theoretical_loss": 3.484508527552861, + "tokens_seen": 1830486016 + }, + { + "epoch": 0.55, + "learning_rate": 0.00035995987963891675, + "loss": 1.2884, + "theoretical_loss": 3.48449803049232, + "tokens_seen": 1830551552 + }, + { + "epoch": 0.55, + "learning_rate": 0.00035994383149448345, + "loss": 1.3879, + "theoretical_loss": 3.4844875339128016, + "tokens_seen": 1830617088 + }, + { + "epoch": 0.55, + "learning_rate": 0.00035992778335005015, + "loss": 1.2097, + "theoretical_loss": 3.4844770378142673, + "tokens_seen": 1830682624 + }, + { + "epoch": 0.55, + "learning_rate": 0.00035991173520561685, + "loss": 1.4095, + "theoretical_loss": 3.484466542196677, + "tokens_seen": 1830748160 + }, + { + "epoch": 0.55, + "learning_rate": 0.00035989568706118355, + "loss": 1.3039, + "theoretical_loss": 3.484456047059992, + "tokens_seen": 1830813696 + }, + { + "epoch": 0.55, + "learning_rate": 0.00035987963891675025, + "loss": 1.3484, + "theoretical_loss": 3.484445552404173, + "tokens_seen": 1830879232 + }, + { + "epoch": 0.55, + "learning_rate": 0.00035986359077231695, + "loss": 1.2803, + "theoretical_loss": 3.48443505822918, + "tokens_seen": 1830944768 + }, + { + "epoch": 0.55, + "learning_rate": 0.00035984754262788365, + "loss": 1.5008, + "theoretical_loss": 3.484424564534975, + "tokens_seen": 1831010304 + }, + { + "epoch": 0.55, + "learning_rate": 0.00035983149448345035, + "loss": 1.3196, + "theoretical_loss": 3.4844140713215177, + "tokens_seen": 1831075840 + }, + { + "epoch": 0.55, + "learning_rate": 0.00035981544633901705, + "loss": 1.4326, + "theoretical_loss": 3.4844035785887693, + "tokens_seen": 1831141376 + }, + { + "epoch": 0.55, + "learning_rate": 0.00035979939819458375, + "loss": 1.2997, + "theoretical_loss": 3.4843930863366905, + "tokens_seen": 1831206912 + }, + { + "epoch": 0.55, + "learning_rate": 0.00035978335005015045, + "loss": 1.2944, + "theoretical_loss": 3.4843825945652425, + "tokens_seen": 1831272448 + }, + { + "epoch": 0.55, + "learning_rate": 0.00035976730190571715, + "loss": 1.266, + "theoretical_loss": 3.484372103274385, + "tokens_seen": 1831337984 + }, + { + "epoch": 0.55, + "learning_rate": 0.0003597512537612839, + "loss": 1.3496, + "theoretical_loss": 3.4843616124640797, + "tokens_seen": 1831403520 + }, + { + "epoch": 0.55, + "learning_rate": 0.0003597352056168506, + "loss": 1.3129, + "theoretical_loss": 3.4843511221342873, + "tokens_seen": 1831469056 + }, + { + "epoch": 0.56, + "learning_rate": 0.0003597191574724173, + "loss": 1.335, + "theoretical_loss": 3.484340632284969, + "tokens_seen": 1831534592 + }, + { + "epoch": 0.56, + "learning_rate": 0.000359703109327984, + "loss": 1.2391, + "theoretical_loss": 3.4843301429160842, + "tokens_seen": 1831600128 + }, + { + "epoch": 0.56, + "learning_rate": 0.0003596870611835507, + "loss": 1.3502, + "theoretical_loss": 3.484319654027595, + "tokens_seen": 1831665664 + }, + { + "epoch": 0.56, + "objective/train/docs_used": 668070, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.2816635370254517, + "objective/train/theoretical_loss": 3.484309165619462, + "objective/train/tokens_used": 1852191200, + "theoretical_loss": 3.484309165619462, + "tokens_seen": 1831731200 + }, + { + "epoch": 0.56, + "learning_rate": 0.0003596710130391174, + "loss": 1.3336, + "theoretical_loss": 3.484309165619462, + "tokens_seen": 1831731200 + }, + { + "epoch": 0.56, + "learning_rate": 0.0003596549648946841, + "loss": 1.2727, + "theoretical_loss": 3.484298677691645, + "tokens_seen": 1831796736 + }, + { + "epoch": 0.56, + "learning_rate": 0.0003596389167502508, + "loss": 1.3452, + "theoretical_loss": 3.484288190244106, + "tokens_seen": 1831862272 + }, + { + "epoch": 0.56, + "learning_rate": 0.00035962286860581746, + "loss": 1.3036, + "theoretical_loss": 3.484277703276806, + "tokens_seen": 1831927808 + }, + { + "epoch": 0.56, + "learning_rate": 0.00035960682046138416, + "loss": 1.446, + "theoretical_loss": 3.484267216789705, + "tokens_seen": 1831993344 + }, + { + "epoch": 0.56, + "learning_rate": 0.00035959077231695086, + "loss": 1.3777, + "theoretical_loss": 3.4842567307827643, + "tokens_seen": 1832058880 + }, + { + "epoch": 0.56, + "learning_rate": 0.00035957472417251756, + "loss": 1.2768, + "theoretical_loss": 3.484246245255944, + "tokens_seen": 1832124416 + }, + { + "epoch": 0.56, + "learning_rate": 0.00035955867602808426, + "loss": 1.3239, + "theoretical_loss": 3.484235760209206, + "tokens_seen": 1832189952 + }, + { + "epoch": 0.56, + "learning_rate": 0.00035954262788365096, + "loss": 1.3875, + "theoretical_loss": 3.4842252756425105, + "tokens_seen": 1832255488 + }, + { + "epoch": 0.56, + "learning_rate": 0.00035952657973921766, + "loss": 1.3332, + "theoretical_loss": 3.4842147915558184, + "tokens_seen": 1832321024 + }, + { + "epoch": 0.56, + "learning_rate": 0.00035951053159478436, + "loss": 1.3705, + "theoretical_loss": 3.4842043079490908, + "tokens_seen": 1832386560 + }, + { + "epoch": 0.56, + "learning_rate": 0.00035949448345035107, + "loss": 1.3925, + "theoretical_loss": 3.4841938248222886, + "tokens_seen": 1832452096 + }, + { + "epoch": 0.56, + "learning_rate": 0.00035947843530591777, + "loss": 1.3814, + "theoretical_loss": 3.4841833421753723, + "tokens_seen": 1832517632 + }, + { + "epoch": 0.56, + "learning_rate": 0.00035946238716148447, + "loss": 1.496, + "theoretical_loss": 3.484172860008303, + "tokens_seen": 1832583168 + }, + { + "epoch": 0.56, + "learning_rate": 0.00035944633901705117, + "loss": 1.3648, + "theoretical_loss": 3.4841623783210416, + "tokens_seen": 1832648704 + }, + { + "epoch": 0.56, + "learning_rate": 0.00035943029087261787, + "loss": 1.3528, + "theoretical_loss": 3.4841518971135486, + "tokens_seen": 1832714240 + }, + { + "epoch": 0.56, + "learning_rate": 0.00035941424272818457, + "loss": 1.2678, + "theoretical_loss": 3.4841414163857856, + "tokens_seen": 1832779776 + }, + { + "epoch": 0.56, + "learning_rate": 0.00035939819458375127, + "loss": 1.3515, + "theoretical_loss": 3.484130936137713, + "tokens_seen": 1832845312 + }, + { + "epoch": 0.56, + "learning_rate": 0.00035938214643931797, + "loss": 1.2538, + "theoretical_loss": 3.4841204563692916, + "tokens_seen": 1832910848 + }, + { + "epoch": 0.56, + "learning_rate": 0.00035936609829488467, + "loss": 1.2855, + "theoretical_loss": 3.484109977080483, + "tokens_seen": 1832976384 + }, + { + "epoch": 0.56, + "learning_rate": 0.00035935005015045137, + "loss": 1.2514, + "theoretical_loss": 3.484099498271247, + "tokens_seen": 1833041920 + }, + { + "epoch": 0.56, + "learning_rate": 0.00035933400200601807, + "loss": 1.4282, + "theoretical_loss": 3.4840890199415453, + "tokens_seen": 1833107456 + }, + { + "epoch": 0.56, + "learning_rate": 0.00035931795386158477, + "loss": 1.3867, + "theoretical_loss": 3.4840785420913387, + "tokens_seen": 1833172992 + }, + { + "epoch": 0.56, + "learning_rate": 0.0003593019057171515, + "loss": 1.3642, + "theoretical_loss": 3.484068064720588, + "tokens_seen": 1833238528 + }, + { + "epoch": 0.56, + "learning_rate": 0.0003592858575727182, + "loss": 1.2862, + "theoretical_loss": 3.4840575878292546, + "tokens_seen": 1833304064 + }, + { + "epoch": 0.56, + "learning_rate": 0.0003592698094282849, + "loss": 1.4268, + "theoretical_loss": 3.484047111417298, + "tokens_seen": 1833369600 + }, + { + "epoch": 0.56, + "learning_rate": 0.0003592537612838516, + "loss": 1.4003, + "theoretical_loss": 3.484036635484681, + "tokens_seen": 1833435136 + }, + { + "epoch": 0.56, + "learning_rate": 0.0003592377131394183, + "loss": 1.3861, + "theoretical_loss": 3.4840261600313633, + "tokens_seen": 1833500672 + }, + { + "epoch": 0.56, + "learning_rate": 0.000359221664994985, + "loss": 1.3791, + "theoretical_loss": 3.4840156850573063, + "tokens_seen": 1833566208 + }, + { + "epoch": 0.56, + "learning_rate": 0.0003592056168505517, + "loss": 1.4106, + "theoretical_loss": 3.4840052105624704, + "tokens_seen": 1833631744 + }, + { + "epoch": 0.56, + "learning_rate": 0.0003591895687061184, + "loss": 1.3964, + "theoretical_loss": 3.4839947365468174, + "tokens_seen": 1833697280 + }, + { + "epoch": 0.56, + "learning_rate": 0.0003591735205616851, + "loss": 1.3761, + "theoretical_loss": 3.483984263010308, + "tokens_seen": 1833762816 + }, + { + "epoch": 0.56, + "learning_rate": 0.0003591574724172518, + "loss": 1.1651, + "theoretical_loss": 3.483973789952903, + "tokens_seen": 1833828352 + }, + { + "epoch": 0.56, + "learning_rate": 0.0003591414242728185, + "loss": 1.4269, + "theoretical_loss": 3.4839633173745628, + "tokens_seen": 1833893888 + }, + { + "epoch": 0.56, + "learning_rate": 0.0003591253761283852, + "loss": 1.3593, + "theoretical_loss": 3.4839528452752493, + "tokens_seen": 1833959424 + }, + { + "epoch": 0.56, + "learning_rate": 0.0003591093279839519, + "loss": 1.3737, + "theoretical_loss": 3.483942373654923, + "tokens_seen": 1834024960 + }, + { + "epoch": 0.56, + "learning_rate": 0.0003590932798395186, + "loss": 1.329, + "theoretical_loss": 3.483931902513545, + "tokens_seen": 1834090496 + }, + { + "epoch": 0.56, + "learning_rate": 0.0003590772316950853, + "loss": 1.3057, + "theoretical_loss": 3.4839214318510763, + "tokens_seen": 1834156032 + }, + { + "epoch": 0.56, + "learning_rate": 0.000359061183550652, + "loss": 1.302, + "theoretical_loss": 3.483910961667478, + "tokens_seen": 1834221568 + }, + { + "epoch": 0.56, + "learning_rate": 0.0003590451354062187, + "loss": 1.2569, + "theoretical_loss": 3.483900491962711, + "tokens_seen": 1834287104 + }, + { + "epoch": 0.56, + "learning_rate": 0.0003590290872617854, + "loss": 1.4218, + "theoretical_loss": 3.4838900227367358, + "tokens_seen": 1834352640 + }, + { + "epoch": 0.56, + "learning_rate": 0.0003590130391173521, + "loss": 1.4108, + "theoretical_loss": 3.4838795539895138, + "tokens_seen": 1834418176 + }, + { + "epoch": 0.56, + "learning_rate": 0.0003589969909729188, + "loss": 1.3624, + "theoretical_loss": 3.483869085721006, + "tokens_seen": 1834483712 + }, + { + "epoch": 0.56, + "learning_rate": 0.0003589809428284855, + "loss": 1.3157, + "theoretical_loss": 3.483858617931174, + "tokens_seen": 1834549248 + }, + { + "epoch": 0.56, + "learning_rate": 0.0003589648946840522, + "loss": 1.2358, + "theoretical_loss": 3.483848150619978, + "tokens_seen": 1834614784 + }, + { + "epoch": 0.56, + "learning_rate": 0.0003589488465396189, + "loss": 1.3048, + "theoretical_loss": 3.483837683787379, + "tokens_seen": 1834680320 + }, + { + "epoch": 0.56, + "learning_rate": 0.0003589327983951856, + "loss": 1.4425, + "theoretical_loss": 3.4838272174333387, + "tokens_seen": 1834745856 + }, + { + "epoch": 0.56, + "learning_rate": 0.0003589167502507523, + "loss": 1.3737, + "theoretical_loss": 3.4838167515578173, + "tokens_seen": 1834811392 + }, + { + "epoch": 0.56, + "learning_rate": 0.000358900702106319, + "loss": 1.3723, + "theoretical_loss": 3.4838062861607764, + "tokens_seen": 1834876928 + }, + { + "epoch": 0.56, + "learning_rate": 0.0003588846539618857, + "loss": 1.3676, + "theoretical_loss": 3.483795821242177, + "tokens_seen": 1834942464 + }, + { + "epoch": 0.56, + "objective/train/docs_used": 669245, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.311979055404663, + "objective/train/theoretical_loss": 3.48378535680198, + "objective/train/tokens_used": 1855468000, + "theoretical_loss": 3.48378535680198, + "tokens_seen": 1835008000 + }, + { + "epoch": 0.56, + "learning_rate": 0.0003588686058174524, + "loss": 1.272, + "theoretical_loss": 3.48378535680198, + "tokens_seen": 1835008000 + }, + { + "epoch": 0.56, + "learning_rate": 0.00035885255767301904, + "loss": 1.3226, + "theoretical_loss": 3.4837748928401466, + "tokens_seen": 1835073536 + }, + { + "epoch": 0.56, + "learning_rate": 0.00035883650952858574, + "loss": 1.3259, + "theoretical_loss": 3.4837644293566377, + "tokens_seen": 1835139072 + }, + { + "epoch": 0.56, + "learning_rate": 0.00035882046138415244, + "loss": 1.405, + "theoretical_loss": 3.483753966351414, + "tokens_seen": 1835204608 + }, + { + "epoch": 0.56, + "learning_rate": 0.00035880441323971914, + "loss": 1.3614, + "theoretical_loss": 3.4837435038244378, + "tokens_seen": 1835270144 + }, + { + "epoch": 0.56, + "learning_rate": 0.00035878836509528584, + "loss": 1.3709, + "theoretical_loss": 3.4837330417756682, + "tokens_seen": 1835335680 + }, + { + "epoch": 0.56, + "learning_rate": 0.00035877231695085254, + "loss": 1.2832, + "theoretical_loss": 3.483722580205068, + "tokens_seen": 1835401216 + }, + { + "epoch": 0.56, + "learning_rate": 0.00035875626880641924, + "loss": 1.3965, + "theoretical_loss": 3.4837121191125977, + "tokens_seen": 1835466752 + }, + { + "epoch": 0.56, + "learning_rate": 0.00035874022066198594, + "loss": 1.2373, + "theoretical_loss": 3.483701658498218, + "tokens_seen": 1835532288 + }, + { + "epoch": 0.56, + "learning_rate": 0.00035872417251755264, + "loss": 1.3804, + "theoretical_loss": 3.483691198361891, + "tokens_seen": 1835597824 + }, + { + "epoch": 0.56, + "learning_rate": 0.00035870812437311934, + "loss": 1.4374, + "theoretical_loss": 3.483680738703576, + "tokens_seen": 1835663360 + }, + { + "epoch": 0.56, + "learning_rate": 0.00035869207622868604, + "loss": 1.3351, + "theoretical_loss": 3.4836702795232357, + "tokens_seen": 1835728896 + }, + { + "epoch": 0.56, + "learning_rate": 0.00035867602808425274, + "loss": 1.2876, + "theoretical_loss": 3.4836598208208307, + "tokens_seen": 1835794432 + }, + { + "epoch": 0.56, + "learning_rate": 0.00035865997993981944, + "loss": 1.4057, + "theoretical_loss": 3.483649362596322, + "tokens_seen": 1835859968 + }, + { + "epoch": 0.56, + "learning_rate": 0.00035864393179538614, + "loss": 1.366, + "theoretical_loss": 3.4836389048496708, + "tokens_seen": 1835925504 + }, + { + "epoch": 0.56, + "learning_rate": 0.00035862788365095285, + "loss": 1.3862, + "theoretical_loss": 3.4836284475808386, + "tokens_seen": 1835991040 + }, + { + "epoch": 0.56, + "learning_rate": 0.00035861183550651955, + "loss": 1.5059, + "theoretical_loss": 3.4836179907897855, + "tokens_seen": 1836056576 + }, + { + "epoch": 0.56, + "learning_rate": 0.00035859578736208625, + "loss": 1.3616, + "theoretical_loss": 3.4836075344764734, + "tokens_seen": 1836122112 + }, + { + "epoch": 0.56, + "learning_rate": 0.00035857973921765295, + "loss": 1.3595, + "theoretical_loss": 3.483597078640863, + "tokens_seen": 1836187648 + }, + { + "epoch": 0.56, + "learning_rate": 0.00035856369107321965, + "loss": 1.3361, + "theoretical_loss": 3.4835866232829167, + "tokens_seen": 1836253184 + }, + { + "epoch": 0.56, + "learning_rate": 0.0003585476429287864, + "loss": 1.3217, + "theoretical_loss": 3.4835761684025934, + "tokens_seen": 1836318720 + }, + { + "epoch": 0.56, + "learning_rate": 0.0003585315947843531, + "loss": 1.4406, + "theoretical_loss": 3.483565713999856, + "tokens_seen": 1836384256 + }, + { + "epoch": 0.56, + "learning_rate": 0.0003585155466399198, + "loss": 1.4065, + "theoretical_loss": 3.483555260074665, + "tokens_seen": 1836449792 + }, + { + "epoch": 0.56, + "learning_rate": 0.0003584994984954865, + "loss": 1.3235, + "theoretical_loss": 3.4835448066269814, + "tokens_seen": 1836515328 + }, + { + "epoch": 0.56, + "learning_rate": 0.0003584834503510532, + "loss": 1.3237, + "theoretical_loss": 3.483534353656767, + "tokens_seen": 1836580864 + }, + { + "epoch": 0.56, + "learning_rate": 0.0003584674022066199, + "loss": 1.269, + "theoretical_loss": 3.4835239011639825, + "tokens_seen": 1836646400 + }, + { + "epoch": 0.56, + "learning_rate": 0.0003584513540621866, + "loss": 1.1991, + "theoretical_loss": 3.483513449148589, + "tokens_seen": 1836711936 + }, + { + "epoch": 0.56, + "learning_rate": 0.0003584353059177533, + "loss": 1.3951, + "theoretical_loss": 3.4835029976105476, + "tokens_seen": 1836777472 + }, + { + "epoch": 0.56, + "learning_rate": 0.00035841925777332, + "loss": 1.2675, + "theoretical_loss": 3.4834925465498197, + "tokens_seen": 1836843008 + }, + { + "epoch": 0.56, + "learning_rate": 0.0003584032096288867, + "loss": 1.2354, + "theoretical_loss": 3.483482095966366, + "tokens_seen": 1836908544 + }, + { + "epoch": 0.56, + "learning_rate": 0.0003583871614844534, + "loss": 1.3883, + "theoretical_loss": 3.4834716458601482, + "tokens_seen": 1836974080 + }, + { + "epoch": 0.56, + "learning_rate": 0.0003583711133400201, + "loss": 1.2458, + "theoretical_loss": 3.483461196231128, + "tokens_seen": 1837039616 + }, + { + "epoch": 0.56, + "learning_rate": 0.0003583550651955868, + "loss": 1.297, + "theoretical_loss": 3.4834507470792655, + "tokens_seen": 1837105152 + }, + { + "epoch": 0.56, + "learning_rate": 0.0003583390170511535, + "loss": 1.2994, + "theoretical_loss": 3.4834402984045223, + "tokens_seen": 1837170688 + }, + { + "epoch": 0.56, + "learning_rate": 0.0003583229689067202, + "loss": 1.3285, + "theoretical_loss": 3.4834298502068597, + "tokens_seen": 1837236224 + }, + { + "epoch": 0.56, + "learning_rate": 0.0003583069207622869, + "loss": 1.3031, + "theoretical_loss": 3.4834194024862386, + "tokens_seen": 1837301760 + }, + { + "epoch": 0.56, + "learning_rate": 0.0003582908726178536, + "loss": 1.3683, + "theoretical_loss": 3.483408955242621, + "tokens_seen": 1837367296 + }, + { + "epoch": 0.56, + "learning_rate": 0.0003582748244734203, + "loss": 1.3493, + "theoretical_loss": 3.483398508475967, + "tokens_seen": 1837432832 + }, + { + "epoch": 0.56, + "learning_rate": 0.000358258776328987, + "loss": 1.375, + "theoretical_loss": 3.4833880621862385, + "tokens_seen": 1837498368 + }, + { + "epoch": 0.56, + "learning_rate": 0.0003582427281845537, + "loss": 1.4452, + "theoretical_loss": 3.4833776163733967, + "tokens_seen": 1837563904 + }, + { + "epoch": 0.56, + "learning_rate": 0.00035822668004012036, + "loss": 1.311, + "theoretical_loss": 3.4833671710374023, + "tokens_seen": 1837629440 + }, + { + "epoch": 0.56, + "learning_rate": 0.00035821063189568706, + "loss": 1.2701, + "theoretical_loss": 3.4833567261782177, + "tokens_seen": 1837694976 + }, + { + "epoch": 0.56, + "learning_rate": 0.00035819458375125376, + "loss": 1.29, + "theoretical_loss": 3.4833462817958027, + "tokens_seen": 1837760512 + }, + { + "epoch": 0.56, + "learning_rate": 0.00035817853560682046, + "loss": 1.376, + "theoretical_loss": 3.4833358378901194, + "tokens_seen": 1837826048 + }, + { + "epoch": 0.56, + "learning_rate": 0.00035816248746238716, + "loss": 1.2917, + "theoretical_loss": 3.4833253944611284, + "tokens_seen": 1837891584 + }, + { + "epoch": 0.56, + "learning_rate": 0.00035814643931795386, + "loss": 1.2234, + "theoretical_loss": 3.483314951508792, + "tokens_seen": 1837957120 + }, + { + "epoch": 0.56, + "learning_rate": 0.00035813039117352057, + "loss": 1.3473, + "theoretical_loss": 3.4833045090330703, + "tokens_seen": 1838022656 + }, + { + "epoch": 0.56, + "learning_rate": 0.00035811434302908727, + "loss": 1.3995, + "theoretical_loss": 3.483294067033926, + "tokens_seen": 1838088192 + }, + { + "epoch": 0.56, + "learning_rate": 0.00035809829488465397, + "loss": 1.3531, + "theoretical_loss": 3.4832836255113184, + "tokens_seen": 1838153728 + }, + { + "epoch": 0.56, + "learning_rate": 0.00035808224674022067, + "loss": 1.3184, + "theoretical_loss": 3.48327318446521, + "tokens_seen": 1838219264 + }, + { + "epoch": 0.56, + "objective/train/docs_used": 670495, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.3886833190917969, + "objective/train/theoretical_loss": 3.483262743895562, + "objective/train/tokens_used": 1858744800, + "theoretical_loss": 3.483262743895562, + "tokens_seen": 1838284800 + }, + { + "epoch": 0.56, + "learning_rate": 0.00035806619859578737, + "loss": 1.4114, + "theoretical_loss": 3.483262743895562, + "tokens_seen": 1838284800 + }, + { + "epoch": 0.56, + "learning_rate": 0.00035805015045135407, + "loss": 1.305, + "theoretical_loss": 3.4832523038023355, + "tokens_seen": 1838350336 + }, + { + "epoch": 0.56, + "learning_rate": 0.00035803410230692077, + "loss": 1.3742, + "theoretical_loss": 3.483241864185492, + "tokens_seen": 1838415872 + }, + { + "epoch": 0.56, + "learning_rate": 0.00035801805416248747, + "loss": 1.42, + "theoretical_loss": 3.483231425044993, + "tokens_seen": 1838481408 + }, + { + "epoch": 0.56, + "learning_rate": 0.00035800200601805417, + "loss": 1.3144, + "theoretical_loss": 3.4832209863807986, + "tokens_seen": 1838546944 + }, + { + "epoch": 0.56, + "learning_rate": 0.00035798595787362087, + "loss": 1.3476, + "theoretical_loss": 3.4832105481928712, + "tokens_seen": 1838612480 + }, + { + "epoch": 0.56, + "learning_rate": 0.00035796990972918757, + "loss": 1.4811, + "theoretical_loss": 3.483200110481172, + "tokens_seen": 1838678016 + }, + { + "epoch": 0.56, + "learning_rate": 0.00035795386158475427, + "loss": 1.389, + "theoretical_loss": 3.4831896732456613, + "tokens_seen": 1838743552 + }, + { + "epoch": 0.56, + "learning_rate": 0.000357937813440321, + "loss": 1.4287, + "theoretical_loss": 3.483179236486302, + "tokens_seen": 1838809088 + }, + { + "epoch": 0.56, + "learning_rate": 0.0003579217652958877, + "loss": 1.4573, + "theoretical_loss": 3.483168800203054, + "tokens_seen": 1838874624 + }, + { + "epoch": 0.56, + "learning_rate": 0.0003579057171514544, + "loss": 1.3807, + "theoretical_loss": 3.4831583643958797, + "tokens_seen": 1838940160 + }, + { + "epoch": 0.56, + "learning_rate": 0.0003578896690070211, + "loss": 1.363, + "theoretical_loss": 3.4831479290647396, + "tokens_seen": 1839005696 + }, + { + "epoch": 0.56, + "learning_rate": 0.0003578736208625878, + "loss": 1.261, + "theoretical_loss": 3.483137494209595, + "tokens_seen": 1839071232 + }, + { + "epoch": 0.56, + "learning_rate": 0.0003578575727181545, + "loss": 1.3732, + "theoretical_loss": 3.4831270598304087, + "tokens_seen": 1839136768 + }, + { + "epoch": 0.56, + "learning_rate": 0.0003578415245737212, + "loss": 1.3524, + "theoretical_loss": 3.48311662592714, + "tokens_seen": 1839202304 + }, + { + "epoch": 0.56, + "learning_rate": 0.0003578254764292879, + "loss": 1.2583, + "theoretical_loss": 3.4831061924997515, + "tokens_seen": 1839267840 + }, + { + "epoch": 0.56, + "learning_rate": 0.0003578094282848546, + "loss": 1.3646, + "theoretical_loss": 3.4830957595482035, + "tokens_seen": 1839333376 + }, + { + "epoch": 0.56, + "learning_rate": 0.0003577933801404213, + "loss": 1.3465, + "theoretical_loss": 3.483085327072459, + "tokens_seen": 1839398912 + }, + { + "epoch": 0.56, + "learning_rate": 0.000357777331995988, + "loss": 1.3503, + "theoretical_loss": 3.4830748950724777, + "tokens_seen": 1839464448 + }, + { + "epoch": 0.56, + "learning_rate": 0.0003577612838515547, + "loss": 1.2451, + "theoretical_loss": 3.4830644635482217, + "tokens_seen": 1839529984 + }, + { + "epoch": 0.56, + "learning_rate": 0.0003577452357071214, + "loss": 1.3212, + "theoretical_loss": 3.483054032499653, + "tokens_seen": 1839595520 + }, + { + "epoch": 0.56, + "learning_rate": 0.0003577291875626881, + "loss": 1.3635, + "theoretical_loss": 3.483043601926731, + "tokens_seen": 1839661056 + }, + { + "epoch": 0.56, + "learning_rate": 0.0003577131394182548, + "loss": 1.3597, + "theoretical_loss": 3.483033171829419, + "tokens_seen": 1839726592 + }, + { + "epoch": 0.56, + "learning_rate": 0.0003576970912738215, + "loss": 1.3804, + "theoretical_loss": 3.4830227422076776, + "tokens_seen": 1839792128 + }, + { + "epoch": 0.56, + "learning_rate": 0.0003576810431293882, + "loss": 1.3899, + "theoretical_loss": 3.4830123130614687, + "tokens_seen": 1839857664 + }, + { + "epoch": 0.56, + "learning_rate": 0.0003576649949849549, + "loss": 1.3816, + "theoretical_loss": 3.4830018843907524, + "tokens_seen": 1839923200 + }, + { + "epoch": 0.56, + "learning_rate": 0.0003576489468405216, + "loss": 1.3497, + "theoretical_loss": 3.4829914561954913, + "tokens_seen": 1839988736 + }, + { + "epoch": 0.56, + "learning_rate": 0.0003576328986960883, + "loss": 1.3832, + "theoretical_loss": 3.4829810284756464, + "tokens_seen": 1840054272 + }, + { + "epoch": 0.56, + "learning_rate": 0.000357616850551655, + "loss": 1.4781, + "theoretical_loss": 3.4829706012311794, + "tokens_seen": 1840119808 + }, + { + "epoch": 0.56, + "learning_rate": 0.0003576008024072217, + "loss": 1.3655, + "theoretical_loss": 3.4829601744620513, + "tokens_seen": 1840185344 + }, + { + "epoch": 0.56, + "learning_rate": 0.0003575847542627884, + "loss": 1.2826, + "theoretical_loss": 3.4829497481682234, + "tokens_seen": 1840250880 + }, + { + "epoch": 0.56, + "learning_rate": 0.0003575687061183551, + "loss": 1.3116, + "theoretical_loss": 3.482939322349657, + "tokens_seen": 1840316416 + }, + { + "epoch": 0.56, + "learning_rate": 0.0003575526579739218, + "loss": 1.3434, + "theoretical_loss": 3.482928897006315, + "tokens_seen": 1840381952 + }, + { + "epoch": 0.56, + "learning_rate": 0.0003575366098294885, + "loss": 1.3224, + "theoretical_loss": 3.4829184721381568, + "tokens_seen": 1840447488 + }, + { + "epoch": 0.56, + "learning_rate": 0.0003575205616850552, + "loss": 1.2791, + "theoretical_loss": 3.482908047745145, + "tokens_seen": 1840513024 + }, + { + "epoch": 0.56, + "learning_rate": 0.0003575045135406219, + "loss": 1.4242, + "theoretical_loss": 3.48289762382724, + "tokens_seen": 1840578560 + }, + { + "epoch": 0.56, + "learning_rate": 0.0003574884653961886, + "loss": 1.3231, + "theoretical_loss": 3.4828872003844045, + "tokens_seen": 1840644096 + }, + { + "epoch": 0.56, + "learning_rate": 0.0003574724172517553, + "loss": 1.3433, + "theoretical_loss": 3.482876777416599, + "tokens_seen": 1840709632 + }, + { + "epoch": 0.56, + "learning_rate": 0.00035745636910732194, + "loss": 1.4082, + "theoretical_loss": 3.482866354923786, + "tokens_seen": 1840775168 + }, + { + "epoch": 0.56, + "learning_rate": 0.00035744032096288864, + "loss": 1.292, + "theoretical_loss": 3.4828559329059257, + "tokens_seen": 1840840704 + }, + { + "epoch": 0.56, + "learning_rate": 0.00035742427281845534, + "loss": 1.2483, + "theoretical_loss": 3.48284551136298, + "tokens_seen": 1840906240 + }, + { + "epoch": 0.56, + "learning_rate": 0.00035740822467402204, + "loss": 1.4437, + "theoretical_loss": 3.482835090294911, + "tokens_seen": 1840971776 + }, + { + "epoch": 0.56, + "learning_rate": 0.00035739217652958874, + "loss": 1.4019, + "theoretical_loss": 3.482824669701679, + "tokens_seen": 1841037312 + }, + { + "epoch": 0.56, + "learning_rate": 0.00035737612838515544, + "loss": 1.4639, + "theoretical_loss": 3.4828142495832464, + "tokens_seen": 1841102848 + }, + { + "epoch": 0.56, + "learning_rate": 0.00035736008024072214, + "loss": 1.4117, + "theoretical_loss": 3.4828038299395745, + "tokens_seen": 1841168384 + }, + { + "epoch": 0.56, + "learning_rate": 0.0003573440320962889, + "loss": 1.2803, + "theoretical_loss": 3.482793410770624, + "tokens_seen": 1841233920 + }, + { + "epoch": 0.56, + "learning_rate": 0.0003573279839518556, + "loss": 1.2377, + "theoretical_loss": 3.4827829920763573, + "tokens_seen": 1841299456 + }, + { + "epoch": 0.56, + "learning_rate": 0.0003573119358074223, + "loss": 1.3288, + "theoretical_loss": 3.482772573856736, + "tokens_seen": 1841364992 + }, + { + "epoch": 0.56, + "learning_rate": 0.000357295887662989, + "loss": 1.4077, + "theoretical_loss": 3.4827621561117206, + "tokens_seen": 1841430528 + }, + { + "epoch": 0.56, + "learning_rate": 0.0003572798395185557, + "loss": 1.3372, + "theoretical_loss": 3.4827517388412734, + "tokens_seen": 1841496064 + }, + { + "epoch": 0.56, + "objective/train/docs_used": 671743, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.5243488550186157, + "objective/train/theoretical_loss": 3.4827413220453556, + "objective/train/tokens_used": 1862021600, + "theoretical_loss": 3.4827413220453556, + "tokens_seen": 1841561600 + }, + { + "epoch": 0.56, + "learning_rate": 0.0003572637913741224, + "loss": 1.3479, + "theoretical_loss": 3.4827413220453556, + "tokens_seen": 1841561600 + }, + { + "epoch": 0.56, + "learning_rate": 0.0003572477432296891, + "loss": 1.3024, + "theoretical_loss": 3.482730905723929, + "tokens_seen": 1841627136 + }, + { + "epoch": 0.56, + "learning_rate": 0.0003572316950852558, + "loss": 1.2607, + "theoretical_loss": 3.482720489876954, + "tokens_seen": 1841692672 + }, + { + "epoch": 0.56, + "learning_rate": 0.0003572156469408225, + "loss": 1.2864, + "theoretical_loss": 3.4827100745043937, + "tokens_seen": 1841758208 + }, + { + "epoch": 0.56, + "learning_rate": 0.0003571995987963892, + "loss": 1.2983, + "theoretical_loss": 3.4826996596062085, + "tokens_seen": 1841823744 + }, + { + "epoch": 0.56, + "learning_rate": 0.0003571835506519559, + "loss": 1.2655, + "theoretical_loss": 3.4826892451823603, + "tokens_seen": 1841889280 + }, + { + "epoch": 0.56, + "learning_rate": 0.0003571675025075226, + "loss": 1.3766, + "theoretical_loss": 3.4826788312328105, + "tokens_seen": 1841954816 + }, + { + "epoch": 0.56, + "learning_rate": 0.0003571514543630893, + "loss": 1.4318, + "theoretical_loss": 3.4826684177575213, + "tokens_seen": 1842020352 + }, + { + "epoch": 0.56, + "learning_rate": 0.000357135406218656, + "loss": 1.3506, + "theoretical_loss": 3.482658004756453, + "tokens_seen": 1842085888 + }, + { + "epoch": 0.56, + "learning_rate": 0.0003571193580742227, + "loss": 1.3357, + "theoretical_loss": 3.4826475922295677, + "tokens_seen": 1842151424 + }, + { + "epoch": 0.56, + "learning_rate": 0.0003571033099297894, + "loss": 1.4066, + "theoretical_loss": 3.4826371801768277, + "tokens_seen": 1842216960 + }, + { + "epoch": 0.56, + "learning_rate": 0.0003570872617853561, + "loss": 1.3331, + "theoretical_loss": 3.4826267685981933, + "tokens_seen": 1842282496 + }, + { + "epoch": 0.56, + "learning_rate": 0.0003570712136409228, + "loss": 1.3081, + "theoretical_loss": 3.482616357493627, + "tokens_seen": 1842348032 + }, + { + "epoch": 0.56, + "learning_rate": 0.0003570551654964895, + "loss": 1.3783, + "theoretical_loss": 3.4826059468630897, + "tokens_seen": 1842413568 + }, + { + "epoch": 0.56, + "learning_rate": 0.0003570391173520562, + "loss": 1.2817, + "theoretical_loss": 3.482595536706543, + "tokens_seen": 1842479104 + }, + { + "epoch": 0.56, + "learning_rate": 0.0003570230692076229, + "loss": 1.4773, + "theoretical_loss": 3.482585127023949, + "tokens_seen": 1842544640 + }, + { + "epoch": 0.56, + "learning_rate": 0.0003570070210631896, + "loss": 1.3424, + "theoretical_loss": 3.482574717815269, + "tokens_seen": 1842610176 + }, + { + "epoch": 0.56, + "learning_rate": 0.0003569909729187563, + "loss": 1.5127, + "theoretical_loss": 3.4825643090804643, + "tokens_seen": 1842675712 + }, + { + "epoch": 0.56, + "learning_rate": 0.000356974924774323, + "loss": 1.3468, + "theoretical_loss": 3.4825539008194966, + "tokens_seen": 1842741248 + }, + { + "epoch": 0.56, + "learning_rate": 0.0003569588766298897, + "loss": 1.291, + "theoretical_loss": 3.4825434930323276, + "tokens_seen": 1842806784 + }, + { + "epoch": 0.56, + "learning_rate": 0.0003569428284854564, + "loss": 1.3674, + "theoretical_loss": 3.482533085718919, + "tokens_seen": 1842872320 + }, + { + "epoch": 0.56, + "learning_rate": 0.0003569267803410231, + "loss": 1.3623, + "theoretical_loss": 3.4825226788792323, + "tokens_seen": 1842937856 + }, + { + "epoch": 0.56, + "learning_rate": 0.0003569107321965898, + "loss": 1.3158, + "theoretical_loss": 3.4825122725132287, + "tokens_seen": 1843003392 + }, + { + "epoch": 0.56, + "learning_rate": 0.0003568946840521565, + "loss": 1.3579, + "theoretical_loss": 3.4825018666208702, + "tokens_seen": 1843068928 + }, + { + "epoch": 0.56, + "learning_rate": 0.0003568786359077232, + "loss": 1.3873, + "theoretical_loss": 3.4824914612021187, + "tokens_seen": 1843134464 + }, + { + "epoch": 0.56, + "learning_rate": 0.0003568625877632899, + "loss": 1.3062, + "theoretical_loss": 3.482481056256935, + "tokens_seen": 1843200000 + }, + { + "epoch": 0.56, + "learning_rate": 0.0003568465396188566, + "loss": 1.3528, + "theoretical_loss": 3.4824706517852815, + "tokens_seen": 1843265536 + }, + { + "epoch": 0.56, + "learning_rate": 0.0003568304914744233, + "loss": 1.3069, + "theoretical_loss": 3.4824602477871194, + "tokens_seen": 1843331072 + }, + { + "epoch": 0.56, + "learning_rate": 0.00035681444332998996, + "loss": 1.3475, + "theoretical_loss": 3.48244984426241, + "tokens_seen": 1843396608 + }, + { + "epoch": 0.56, + "learning_rate": 0.00035679839518555666, + "loss": 1.3902, + "theoretical_loss": 3.482439441211116, + "tokens_seen": 1843462144 + }, + { + "epoch": 0.56, + "learning_rate": 0.00035678234704112336, + "loss": 1.2054, + "theoretical_loss": 3.4824290386331977, + "tokens_seen": 1843527680 + }, + { + "epoch": 0.56, + "learning_rate": 0.00035676629889669007, + "loss": 1.211, + "theoretical_loss": 3.4824186365286174, + "tokens_seen": 1843593216 + }, + { + "epoch": 0.56, + "learning_rate": 0.00035675025075225677, + "loss": 1.3662, + "theoretical_loss": 3.4824082348973366, + "tokens_seen": 1843658752 + }, + { + "epoch": 0.56, + "learning_rate": 0.00035673420260782347, + "loss": 1.325, + "theoretical_loss": 3.4823978337393173, + "tokens_seen": 1843724288 + }, + { + "epoch": 0.56, + "learning_rate": 0.00035671815446339017, + "loss": 1.2922, + "theoretical_loss": 3.4823874330545213, + "tokens_seen": 1843789824 + }, + { + "epoch": 0.56, + "learning_rate": 0.00035670210631895687, + "loss": 1.3258, + "theoretical_loss": 3.4823770328429093, + "tokens_seen": 1843855360 + }, + { + "epoch": 0.56, + "learning_rate": 0.00035668605817452357, + "loss": 1.3009, + "theoretical_loss": 3.4823666331044434, + "tokens_seen": 1843920896 + }, + { + "epoch": 0.56, + "learning_rate": 0.00035667001003009027, + "loss": 1.347, + "theoretical_loss": 3.4823562338390857, + "tokens_seen": 1843986432 + }, + { + "epoch": 0.56, + "learning_rate": 0.00035665396188565697, + "loss": 1.2575, + "theoretical_loss": 3.4823458350467975, + "tokens_seen": 1844051968 + }, + { + "epoch": 0.56, + "learning_rate": 0.00035663791374122367, + "loss": 1.3046, + "theoretical_loss": 3.4823354367275403, + "tokens_seen": 1844117504 + }, + { + "epoch": 0.56, + "learning_rate": 0.00035662186559679037, + "loss": 1.3352, + "theoretical_loss": 3.482325038881276, + "tokens_seen": 1844183040 + }, + { + "epoch": 0.56, + "learning_rate": 0.00035660581745235707, + "loss": 1.3438, + "theoretical_loss": 3.482314641507966, + "tokens_seen": 1844248576 + }, + { + "epoch": 0.56, + "learning_rate": 0.00035658976930792377, + "loss": 1.3673, + "theoretical_loss": 3.4823042446075725, + "tokens_seen": 1844314112 + }, + { + "epoch": 0.56, + "learning_rate": 0.00035657372116349047, + "loss": 1.2178, + "theoretical_loss": 3.4822938481800567, + "tokens_seen": 1844379648 + }, + { + "epoch": 0.56, + "learning_rate": 0.0003565576730190572, + "loss": 1.3323, + "theoretical_loss": 3.4822834522253805, + "tokens_seen": 1844445184 + }, + { + "epoch": 0.56, + "learning_rate": 0.0003565416248746239, + "loss": 1.1665, + "theoretical_loss": 3.482273056743506, + "tokens_seen": 1844510720 + }, + { + "epoch": 0.56, + "learning_rate": 0.0003565255767301906, + "loss": 1.2354, + "theoretical_loss": 3.482262661734394, + "tokens_seen": 1844576256 + }, + { + "epoch": 0.56, + "learning_rate": 0.0003565095285857573, + "loss": 1.5354, + "theoretical_loss": 3.4822522671980067, + "tokens_seen": 1844641792 + }, + { + "epoch": 0.56, + "learning_rate": 0.000356493480441324, + "loss": 1.3525, + "theoretical_loss": 3.4822418731343063, + "tokens_seen": 1844707328 + }, + { + "epoch": 0.56, + "learning_rate": 0.0003564774322968907, + "loss": 1.3162, + "theoretical_loss": 3.4822314795432536, + "tokens_seen": 1844772864 + }, + { + "epoch": 0.56, + "objective/train/docs_used": 672951, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.3834693431854248, + "objective/train/theoretical_loss": 3.4822210864248113, + "objective/train/tokens_used": 1865298400, + "theoretical_loss": 3.4822210864248113, + "tokens_seen": 1844838400 + }, + { + "epoch": 0.56, + "learning_rate": 0.0003564613841524574, + "loss": 1.4434, + "theoretical_loss": 3.4822210864248113, + "tokens_seen": 1844838400 + }, + { + "epoch": 0.56, + "learning_rate": 0.0003564453360080241, + "loss": 1.3819, + "theoretical_loss": 3.4822106937789394, + "tokens_seen": 1844903936 + }, + { + "epoch": 0.56, + "learning_rate": 0.0003564292878635908, + "loss": 1.3925, + "theoretical_loss": 3.4822003016056016, + "tokens_seen": 1844969472 + }, + { + "epoch": 0.56, + "learning_rate": 0.0003564132397191575, + "loss": 1.3739, + "theoretical_loss": 3.4821899099047586, + "tokens_seen": 1845035008 + }, + { + "epoch": 0.56, + "learning_rate": 0.0003563971915747242, + "loss": 1.2526, + "theoretical_loss": 3.4821795186763724, + "tokens_seen": 1845100544 + }, + { + "epoch": 0.56, + "learning_rate": 0.0003563811434302909, + "loss": 1.1892, + "theoretical_loss": 3.482169127920405, + "tokens_seen": 1845166080 + }, + { + "epoch": 0.56, + "learning_rate": 0.0003563650952858576, + "loss": 1.3415, + "theoretical_loss": 3.4821587376368175, + "tokens_seen": 1845231616 + }, + { + "epoch": 0.56, + "learning_rate": 0.0003563490471414243, + "loss": 1.3703, + "theoretical_loss": 3.482148347825572, + "tokens_seen": 1845297152 + }, + { + "epoch": 0.56, + "learning_rate": 0.000356332998996991, + "loss": 1.2323, + "theoretical_loss": 3.4821379584866303, + "tokens_seen": 1845362688 + }, + { + "epoch": 0.56, + "learning_rate": 0.0003563169508525577, + "loss": 1.3615, + "theoretical_loss": 3.482127569619954, + "tokens_seen": 1845428224 + }, + { + "epoch": 0.56, + "learning_rate": 0.0003563009027081244, + "loss": 1.2307, + "theoretical_loss": 3.482117181225505, + "tokens_seen": 1845493760 + }, + { + "epoch": 0.56, + "learning_rate": 0.0003562848545636911, + "loss": 1.3652, + "theoretical_loss": 3.482106793303245, + "tokens_seen": 1845559296 + }, + { + "epoch": 0.56, + "learning_rate": 0.0003562688064192578, + "loss": 1.4163, + "theoretical_loss": 3.4820964058531354, + "tokens_seen": 1845624832 + }, + { + "epoch": 0.56, + "learning_rate": 0.0003562527582748245, + "loss": 1.4229, + "theoretical_loss": 3.482086018875139, + "tokens_seen": 1845690368 + }, + { + "epoch": 0.56, + "learning_rate": 0.0003562367101303912, + "loss": 1.3965, + "theoretical_loss": 3.4820756323692166, + "tokens_seen": 1845755904 + }, + { + "epoch": 0.56, + "learning_rate": 0.0003562206619859579, + "loss": 1.3925, + "theoretical_loss": 3.4820652463353303, + "tokens_seen": 1845821440 + }, + { + "epoch": 0.56, + "learning_rate": 0.0003562046138415246, + "loss": 1.3799, + "theoretical_loss": 3.4820548607734425, + "tokens_seen": 1845886976 + }, + { + "epoch": 0.56, + "learning_rate": 0.0003561885656970913, + "loss": 1.2469, + "theoretical_loss": 3.4820444756835136, + "tokens_seen": 1845952512 + }, + { + "epoch": 0.56, + "learning_rate": 0.000356172517552658, + "loss": 1.4526, + "theoretical_loss": 3.4820340910655068, + "tokens_seen": 1846018048 + }, + { + "epoch": 0.56, + "learning_rate": 0.0003561564694082247, + "loss": 1.2652, + "theoretical_loss": 3.482023706919383, + "tokens_seen": 1846083584 + }, + { + "epoch": 0.56, + "learning_rate": 0.0003561404212637914, + "loss": 1.3755, + "theoretical_loss": 3.482013323245104, + "tokens_seen": 1846149120 + }, + { + "epoch": 0.56, + "learning_rate": 0.0003561243731193581, + "loss": 1.3895, + "theoretical_loss": 3.4820029400426327, + "tokens_seen": 1846214656 + }, + { + "epoch": 0.56, + "learning_rate": 0.0003561083249749248, + "loss": 1.3055, + "theoretical_loss": 3.481992557311929, + "tokens_seen": 1846280192 + }, + { + "epoch": 0.56, + "learning_rate": 0.0003560922768304915, + "loss": 1.2411, + "theoretical_loss": 3.481982175052957, + "tokens_seen": 1846345728 + }, + { + "epoch": 0.56, + "learning_rate": 0.0003560762286860582, + "loss": 1.3413, + "theoretical_loss": 3.481971793265677, + "tokens_seen": 1846411264 + }, + { + "epoch": 0.56, + "learning_rate": 0.0003560601805416249, + "loss": 1.4999, + "theoretical_loss": 3.481961411950051, + "tokens_seen": 1846476800 + }, + { + "epoch": 0.56, + "learning_rate": 0.0003560441323971916, + "loss": 1.3178, + "theoretical_loss": 3.4819510311060418, + "tokens_seen": 1846542336 + }, + { + "epoch": 0.56, + "learning_rate": 0.0003560280842527583, + "loss": 1.302, + "theoretical_loss": 3.4819406507336095, + "tokens_seen": 1846607872 + }, + { + "epoch": 0.56, + "learning_rate": 0.000356012036108325, + "loss": 1.2948, + "theoretical_loss": 3.4819302708327173, + "tokens_seen": 1846673408 + }, + { + "epoch": 0.56, + "learning_rate": 0.0003559959879638917, + "loss": 1.3396, + "theoretical_loss": 3.4819198914033267, + "tokens_seen": 1846738944 + }, + { + "epoch": 0.56, + "learning_rate": 0.0003559799398194584, + "loss": 1.3098, + "theoretical_loss": 3.4819095124454, + "tokens_seen": 1846804480 + }, + { + "epoch": 0.56, + "learning_rate": 0.0003559638916750251, + "loss": 1.4215, + "theoretical_loss": 3.4818991339588976, + "tokens_seen": 1846870016 + }, + { + "epoch": 0.56, + "learning_rate": 0.0003559478435305918, + "loss": 1.3184, + "theoretical_loss": 3.4818887559437828, + "tokens_seen": 1846935552 + }, + { + "epoch": 0.56, + "learning_rate": 0.0003559317953861585, + "loss": 1.3043, + "theoretical_loss": 3.481878378400017, + "tokens_seen": 1847001088 + }, + { + "epoch": 0.56, + "learning_rate": 0.0003559157472417252, + "loss": 1.333, + "theoretical_loss": 3.4818680013275625, + "tokens_seen": 1847066624 + }, + { + "epoch": 0.56, + "learning_rate": 0.0003558996990972919, + "loss": 1.3678, + "theoretical_loss": 3.4818576247263806, + "tokens_seen": 1847132160 + }, + { + "epoch": 0.56, + "learning_rate": 0.0003558836509528586, + "loss": 1.3843, + "theoretical_loss": 3.481847248596433, + "tokens_seen": 1847197696 + }, + { + "epoch": 0.56, + "learning_rate": 0.0003558676028084253, + "loss": 1.2731, + "theoretical_loss": 3.4818368729376816, + "tokens_seen": 1847263232 + }, + { + "epoch": 0.56, + "learning_rate": 0.000355851554663992, + "loss": 1.3644, + "theoretical_loss": 3.4818264977500895, + "tokens_seen": 1847328768 + }, + { + "epoch": 0.56, + "learning_rate": 0.0003558355065195587, + "loss": 1.2573, + "theoretical_loss": 3.4818161230336173, + "tokens_seen": 1847394304 + }, + { + "epoch": 0.56, + "learning_rate": 0.0003558194583751254, + "loss": 1.2628, + "theoretical_loss": 3.481805748788227, + "tokens_seen": 1847459840 + }, + { + "epoch": 0.56, + "learning_rate": 0.0003558034102306921, + "loss": 1.3653, + "theoretical_loss": 3.4817953750138813, + "tokens_seen": 1847525376 + }, + { + "epoch": 0.56, + "learning_rate": 0.0003557873620862588, + "loss": 1.3627, + "theoretical_loss": 3.481785001710541, + "tokens_seen": 1847590912 + }, + { + "epoch": 0.56, + "learning_rate": 0.0003557713139418255, + "loss": 1.2312, + "theoretical_loss": 3.481774628878169, + "tokens_seen": 1847656448 + }, + { + "epoch": 0.56, + "learning_rate": 0.0003557552657973922, + "loss": 1.3259, + "theoretical_loss": 3.481764256516726, + "tokens_seen": 1847721984 + }, + { + "epoch": 0.56, + "learning_rate": 0.0003557392176529589, + "loss": 1.3367, + "theoretical_loss": 3.4817538846261757, + "tokens_seen": 1847787520 + }, + { + "epoch": 0.56, + "learning_rate": 0.0003557231695085256, + "loss": 1.2157, + "theoretical_loss": 3.481743513206479, + "tokens_seen": 1847853056 + }, + { + "epoch": 0.56, + "learning_rate": 0.0003557071213640923, + "loss": 1.4417, + "theoretical_loss": 3.4817331422575974, + "tokens_seen": 1847918592 + }, + { + "epoch": 0.56, + "learning_rate": 0.000355691073219659, + "loss": 1.3994, + "theoretical_loss": 3.4817227717794936, + "tokens_seen": 1847984128 + }, + { + "epoch": 0.56, + "learning_rate": 0.0003556750250752257, + "loss": 1.3638, + "theoretical_loss": 3.4817124017721293, + "tokens_seen": 1848049664 + }, + { + "epoch": 0.56, + "objective/train/docs_used": 673981, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.3666971921920776, + "objective/train/theoretical_loss": 3.481702032235466, + "objective/train/tokens_used": 1868575200, + "theoretical_loss": 3.481702032235466, + "tokens_seen": 1848115200 + }, + { + "epoch": 0.56, + "learning_rate": 0.0003556589769307924, + "loss": 1.3726, + "theoretical_loss": 3.481702032235466, + "tokens_seen": 1848115200 + }, + { + "epoch": 0.56, + "learning_rate": 0.0003556429287863591, + "loss": 1.3588, + "theoretical_loss": 3.4816916631694665, + "tokens_seen": 1848180736 + }, + { + "epoch": 0.56, + "learning_rate": 0.0003556268806419258, + "loss": 1.4323, + "theoretical_loss": 3.481681294574092, + "tokens_seen": 1848246272 + }, + { + "epoch": 0.56, + "learning_rate": 0.0003556108324974925, + "loss": 1.3529, + "theoretical_loss": 3.4816709264493046, + "tokens_seen": 1848311808 + }, + { + "epoch": 0.56, + "learning_rate": 0.0003555947843530592, + "loss": 1.4312, + "theoretical_loss": 3.481660558795067, + "tokens_seen": 1848377344 + }, + { + "epoch": 0.56, + "learning_rate": 0.0003555787362086259, + "loss": 1.5316, + "theoretical_loss": 3.4816501916113394, + "tokens_seen": 1848442880 + }, + { + "epoch": 0.56, + "learning_rate": 0.0003555626880641926, + "loss": 1.412, + "theoretical_loss": 3.4816398248980858, + "tokens_seen": 1848508416 + }, + { + "epoch": 0.56, + "learning_rate": 0.0003555466399197593, + "loss": 1.3074, + "theoretical_loss": 3.481629458655267, + "tokens_seen": 1848573952 + }, + { + "epoch": 0.56, + "learning_rate": 0.000355530591775326, + "loss": 1.3057, + "theoretical_loss": 3.4816190928828457, + "tokens_seen": 1848639488 + }, + { + "epoch": 0.56, + "learning_rate": 0.0003555145436308927, + "loss": 1.4298, + "theoretical_loss": 3.4816087275807828, + "tokens_seen": 1848705024 + }, + { + "epoch": 0.56, + "learning_rate": 0.0003554984954864594, + "loss": 1.3769, + "theoretical_loss": 3.481598362749041, + "tokens_seen": 1848770560 + }, + { + "epoch": 0.56, + "learning_rate": 0.0003554824473420261, + "loss": 1.3994, + "theoretical_loss": 3.4815879983875826, + "tokens_seen": 1848836096 + }, + { + "epoch": 0.56, + "learning_rate": 0.0003554663991975928, + "loss": 1.2866, + "theoretical_loss": 3.481577634496369, + "tokens_seen": 1848901632 + }, + { + "epoch": 0.56, + "learning_rate": 0.0003554503510531595, + "loss": 1.4843, + "theoretical_loss": 3.4815672710753622, + "tokens_seen": 1848967168 + }, + { + "epoch": 0.56, + "learning_rate": 0.0003554343029087262, + "loss": 1.4021, + "theoretical_loss": 3.4815569081245243, + "tokens_seen": 1849032704 + }, + { + "epoch": 0.56, + "learning_rate": 0.00035541825476429286, + "loss": 1.4494, + "theoretical_loss": 3.481546545643818, + "tokens_seen": 1849098240 + }, + { + "epoch": 0.56, + "learning_rate": 0.00035540220661985956, + "loss": 1.3785, + "theoretical_loss": 3.4815361836332044, + "tokens_seen": 1849163776 + }, + { + "epoch": 0.56, + "learning_rate": 0.00035538615847542627, + "loss": 1.4435, + "theoretical_loss": 3.481525822092646, + "tokens_seen": 1849229312 + }, + { + "epoch": 0.56, + "learning_rate": 0.00035537011033099297, + "loss": 1.4009, + "theoretical_loss": 3.481515461022104, + "tokens_seen": 1849294848 + }, + { + "epoch": 0.56, + "learning_rate": 0.00035535406218655967, + "loss": 1.3058, + "theoretical_loss": 3.4815051004215416, + "tokens_seen": 1849360384 + }, + { + "epoch": 0.56, + "learning_rate": 0.00035533801404212637, + "loss": 1.3658, + "theoretical_loss": 3.4814947402909207, + "tokens_seen": 1849425920 + }, + { + "epoch": 0.56, + "learning_rate": 0.00035532196589769307, + "loss": 1.3441, + "theoretical_loss": 3.481484380630202, + "tokens_seen": 1849491456 + }, + { + "epoch": 0.56, + "learning_rate": 0.00035530591775325977, + "loss": 1.4565, + "theoretical_loss": 3.4814740214393494, + "tokens_seen": 1849556992 + }, + { + "epoch": 0.56, + "learning_rate": 0.00035528986960882647, + "loss": 1.3694, + "theoretical_loss": 3.4814636627183235, + "tokens_seen": 1849622528 + }, + { + "epoch": 0.56, + "learning_rate": 0.00035527382146439317, + "loss": 1.4105, + "theoretical_loss": 3.4814533044670872, + "tokens_seen": 1849688064 + }, + { + "epoch": 0.56, + "learning_rate": 0.00035525777331995987, + "loss": 1.2905, + "theoretical_loss": 3.481442946685602, + "tokens_seen": 1849753600 + }, + { + "epoch": 0.56, + "learning_rate": 0.00035524172517552657, + "loss": 1.4006, + "theoretical_loss": 3.4814325893738296, + "tokens_seen": 1849819136 + }, + { + "epoch": 0.56, + "learning_rate": 0.00035522567703109327, + "loss": 1.3203, + "theoretical_loss": 3.481422232531733, + "tokens_seen": 1849884672 + }, + { + "epoch": 0.56, + "learning_rate": 0.00035520962888665997, + "loss": 1.3784, + "theoretical_loss": 3.4814118761592745, + "tokens_seen": 1849950208 + }, + { + "epoch": 0.56, + "learning_rate": 0.0003551935807422267, + "loss": 1.4426, + "theoretical_loss": 3.481401520256415, + "tokens_seen": 1850015744 + }, + { + "epoch": 0.56, + "learning_rate": 0.0003551775325977934, + "loss": 1.4082, + "theoretical_loss": 3.4813911648231173, + "tokens_seen": 1850081280 + }, + { + "epoch": 0.56, + "learning_rate": 0.0003551614844533601, + "loss": 1.1973, + "theoretical_loss": 3.481380809859343, + "tokens_seen": 1850146816 + }, + { + "epoch": 0.56, + "learning_rate": 0.0003551454363089268, + "loss": 1.3947, + "theoretical_loss": 3.4813704553650546, + "tokens_seen": 1850212352 + }, + { + "epoch": 0.56, + "learning_rate": 0.0003551293881644935, + "loss": 1.2462, + "theoretical_loss": 3.481360101340214, + "tokens_seen": 1850277888 + }, + { + "epoch": 0.56, + "learning_rate": 0.0003551133400200602, + "loss": 1.3804, + "theoretical_loss": 3.4813497477847837, + "tokens_seen": 1850343424 + }, + { + "epoch": 0.56, + "learning_rate": 0.0003550972918756269, + "loss": 1.3236, + "theoretical_loss": 3.4813393946987254, + "tokens_seen": 1850408960 + }, + { + "epoch": 0.56, + "learning_rate": 0.0003550812437311936, + "loss": 1.3307, + "theoretical_loss": 3.481329042082001, + "tokens_seen": 1850474496 + }, + { + "epoch": 0.56, + "learning_rate": 0.0003550651955867603, + "loss": 1.3711, + "theoretical_loss": 3.481318689934573, + "tokens_seen": 1850540032 + }, + { + "epoch": 0.56, + "learning_rate": 0.000355049147442327, + "loss": 1.3243, + "theoretical_loss": 3.481308338256403, + "tokens_seen": 1850605568 + }, + { + "epoch": 0.56, + "learning_rate": 0.0003550330992978937, + "loss": 1.4446, + "theoretical_loss": 3.481297987047454, + "tokens_seen": 1850671104 + }, + { + "epoch": 0.56, + "learning_rate": 0.0003550170511534604, + "loss": 1.462, + "theoretical_loss": 3.4812876363076874, + "tokens_seen": 1850736640 + }, + { + "epoch": 0.56, + "learning_rate": 0.0003550010030090271, + "loss": 1.3368, + "theoretical_loss": 3.4812772860370655, + "tokens_seen": 1850802176 + }, + { + "epoch": 0.56, + "learning_rate": 0.0003549849548645938, + "loss": 1.4463, + "theoretical_loss": 3.4812669362355506, + "tokens_seen": 1850867712 + }, + { + "epoch": 0.56, + "learning_rate": 0.0003549689067201605, + "loss": 1.4294, + "theoretical_loss": 3.4812565869031045, + "tokens_seen": 1850933248 + }, + { + "epoch": 0.56, + "learning_rate": 0.00035495285857572724, + "loss": 1.4274, + "theoretical_loss": 3.48124623803969, + "tokens_seen": 1850998784 + }, + { + "epoch": 0.56, + "learning_rate": 0.00035493681043129394, + "loss": 1.2907, + "theoretical_loss": 3.481235889645268, + "tokens_seen": 1851064320 + }, + { + "epoch": 0.56, + "learning_rate": 0.00035492076228686064, + "loss": 1.4082, + "theoretical_loss": 3.4812255417198017, + "tokens_seen": 1851129856 + }, + { + "epoch": 0.56, + "learning_rate": 0.00035490471414242734, + "loss": 1.3223, + "theoretical_loss": 3.4812151942632523, + "tokens_seen": 1851195392 + }, + { + "epoch": 0.56, + "learning_rate": 0.00035488866599799404, + "loss": 1.2853, + "theoretical_loss": 3.4812048472755834, + "tokens_seen": 1851260928 + }, + { + "epoch": 0.56, + "learning_rate": 0.00035487261785356074, + "loss": 1.2391, + "theoretical_loss": 3.4811945007567564, + "tokens_seen": 1851326464 + }, + { + "epoch": 0.56, + "objective/train/docs_used": 675199, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.413558006286621, + "objective/train/theoretical_loss": 3.4811841547067326, + "objective/train/tokens_used": 1871852000, + "theoretical_loss": 3.4811841547067326, + "tokens_seen": 1851392000 + }, + { + "epoch": 0.56, + "learning_rate": 0.00035485656970912744, + "loss": 1.3081, + "theoretical_loss": 3.4811841547067326, + "tokens_seen": 1851392000 + }, + { + "epoch": 0.56, + "learning_rate": 0.00035484052156469414, + "loss": 1.4362, + "theoretical_loss": 3.4811738091254756, + "tokens_seen": 1851457536 + }, + { + "epoch": 0.56, + "learning_rate": 0.00035482447342026084, + "loss": 1.3023, + "theoretical_loss": 3.4811634640129467, + "tokens_seen": 1851523072 + }, + { + "epoch": 0.56, + "learning_rate": 0.00035480842527582754, + "loss": 1.4018, + "theoretical_loss": 3.4811531193691083, + "tokens_seen": 1851588608 + }, + { + "epoch": 0.56, + "learning_rate": 0.0003547923771313942, + "loss": 1.3892, + "theoretical_loss": 3.4811427751939226, + "tokens_seen": 1851654144 + }, + { + "epoch": 0.56, + "learning_rate": 0.0003547763289869609, + "loss": 1.4094, + "theoretical_loss": 3.481132431487352, + "tokens_seen": 1851719680 + }, + { + "epoch": 0.56, + "learning_rate": 0.0003547602808425276, + "loss": 1.2283, + "theoretical_loss": 3.481122088249358, + "tokens_seen": 1851785216 + }, + { + "epoch": 0.56, + "learning_rate": 0.0003547442326980943, + "loss": 1.272, + "theoretical_loss": 3.481111745479904, + "tokens_seen": 1851850752 + }, + { + "epoch": 0.56, + "learning_rate": 0.000354728184553661, + "loss": 1.3315, + "theoretical_loss": 3.4811014031789504, + "tokens_seen": 1851916288 + }, + { + "epoch": 0.56, + "learning_rate": 0.0003547121364092277, + "loss": 1.3201, + "theoretical_loss": 3.4810910613464614, + "tokens_seen": 1851981824 + }, + { + "epoch": 0.56, + "learning_rate": 0.0003546960882647944, + "loss": 1.3985, + "theoretical_loss": 3.4810807199823977, + "tokens_seen": 1852047360 + }, + { + "epoch": 0.56, + "learning_rate": 0.0003546800401203611, + "loss": 1.3131, + "theoretical_loss": 3.481070379086722, + "tokens_seen": 1852112896 + }, + { + "epoch": 0.56, + "learning_rate": 0.0003546639919759278, + "loss": 1.4133, + "theoretical_loss": 3.481060038659397, + "tokens_seen": 1852178432 + }, + { + "epoch": 0.56, + "learning_rate": 0.0003546479438314945, + "loss": 1.4268, + "theoretical_loss": 3.4810496987003843, + "tokens_seen": 1852243968 + }, + { + "epoch": 0.56, + "learning_rate": 0.0003546318956870612, + "loss": 1.1894, + "theoretical_loss": 3.481039359209646, + "tokens_seen": 1852309504 + }, + { + "epoch": 0.56, + "learning_rate": 0.0003546158475426279, + "loss": 1.3763, + "theoretical_loss": 3.4810290201871448, + "tokens_seen": 1852375040 + }, + { + "epoch": 0.56, + "learning_rate": 0.0003545997993981946, + "loss": 1.3436, + "theoretical_loss": 3.481018681632843, + "tokens_seen": 1852440576 + }, + { + "epoch": 0.56, + "learning_rate": 0.0003545837512537613, + "loss": 1.3811, + "theoretical_loss": 3.4810083435467023, + "tokens_seen": 1852506112 + }, + { + "epoch": 0.56, + "learning_rate": 0.000354567703109328, + "loss": 1.3279, + "theoretical_loss": 3.480998005928685, + "tokens_seen": 1852571648 + }, + { + "epoch": 0.56, + "learning_rate": 0.0003545516549648947, + "loss": 1.3202, + "theoretical_loss": 3.480987668778754, + "tokens_seen": 1852637184 + }, + { + "epoch": 0.56, + "learning_rate": 0.0003545356068204614, + "loss": 1.2696, + "theoretical_loss": 3.480977332096871, + "tokens_seen": 1852702720 + }, + { + "epoch": 0.56, + "learning_rate": 0.0003545195586760281, + "loss": 1.4083, + "theoretical_loss": 3.4809669958829983, + "tokens_seen": 1852768256 + }, + { + "epoch": 0.56, + "learning_rate": 0.0003545035105315948, + "loss": 1.3731, + "theoretical_loss": 3.4809566601370983, + "tokens_seen": 1852833792 + }, + { + "epoch": 0.56, + "learning_rate": 0.0003544874623871615, + "loss": 1.3523, + "theoretical_loss": 3.4809463248591332, + "tokens_seen": 1852899328 + }, + { + "epoch": 0.56, + "learning_rate": 0.0003544714142427282, + "loss": 1.3292, + "theoretical_loss": 3.4809359900490655, + "tokens_seen": 1852964864 + }, + { + "epoch": 0.56, + "learning_rate": 0.0003544553660982949, + "loss": 1.423, + "theoretical_loss": 3.480925655706857, + "tokens_seen": 1853030400 + }, + { + "epoch": 0.56, + "learning_rate": 0.0003544393179538616, + "loss": 1.3117, + "theoretical_loss": 3.48091532183247, + "tokens_seen": 1853095936 + }, + { + "epoch": 0.56, + "learning_rate": 0.0003544232698094283, + "loss": 1.2452, + "theoretical_loss": 3.480904988425867, + "tokens_seen": 1853161472 + }, + { + "epoch": 0.56, + "learning_rate": 0.000354407221664995, + "loss": 1.3155, + "theoretical_loss": 3.4808946554870106, + "tokens_seen": 1853227008 + }, + { + "epoch": 0.56, + "learning_rate": 0.0003543911735205617, + "loss": 1.2649, + "theoretical_loss": 3.4808843230158626, + "tokens_seen": 1853292544 + }, + { + "epoch": 0.56, + "learning_rate": 0.0003543751253761284, + "loss": 1.3615, + "theoretical_loss": 3.480873991012385, + "tokens_seen": 1853358080 + }, + { + "epoch": 0.56, + "learning_rate": 0.0003543590772316951, + "loss": 1.2738, + "theoretical_loss": 3.480863659476541, + "tokens_seen": 1853423616 + }, + { + "epoch": 0.56, + "learning_rate": 0.0003543430290872618, + "loss": 1.4895, + "theoretical_loss": 3.4808533284082928, + "tokens_seen": 1853489152 + }, + { + "epoch": 0.56, + "learning_rate": 0.0003543269809428285, + "loss": 1.3737, + "theoretical_loss": 3.480842997807602, + "tokens_seen": 1853554688 + }, + { + "epoch": 0.56, + "learning_rate": 0.0003543109327983952, + "loss": 1.3266, + "theoretical_loss": 3.480832667674431, + "tokens_seen": 1853620224 + }, + { + "epoch": 0.56, + "learning_rate": 0.0003542948846539619, + "loss": 1.343, + "theoretical_loss": 3.4808223380087426, + "tokens_seen": 1853685760 + }, + { + "epoch": 0.56, + "learning_rate": 0.0003542788365095286, + "loss": 1.2846, + "theoretical_loss": 3.480812008810499, + "tokens_seen": 1853751296 + }, + { + "epoch": 0.56, + "learning_rate": 0.0003542627883650953, + "loss": 1.3679, + "theoretical_loss": 3.4808016800796624, + "tokens_seen": 1853816832 + }, + { + "epoch": 0.56, + "learning_rate": 0.000354246740220662, + "loss": 1.2446, + "theoretical_loss": 3.480791351816195, + "tokens_seen": 1853882368 + }, + { + "epoch": 0.56, + "learning_rate": 0.0003542306920762287, + "loss": 1.3467, + "theoretical_loss": 3.480781024020059, + "tokens_seen": 1853947904 + }, + { + "epoch": 0.56, + "learning_rate": 0.0003542146439317954, + "loss": 1.3251, + "theoretical_loss": 3.4807706966912173, + "tokens_seen": 1854013440 + }, + { + "epoch": 0.56, + "learning_rate": 0.0003541985957873621, + "loss": 1.406, + "theoretical_loss": 3.480760369829632, + "tokens_seen": 1854078976 + }, + { + "epoch": 0.56, + "learning_rate": 0.0003541825476429288, + "loss": 1.4024, + "theoretical_loss": 3.4807500434352647, + "tokens_seen": 1854144512 + }, + { + "epoch": 0.56, + "learning_rate": 0.0003541664994984955, + "loss": 1.3218, + "theoretical_loss": 3.4807397175080794, + "tokens_seen": 1854210048 + }, + { + "epoch": 0.56, + "learning_rate": 0.0003541504513540622, + "loss": 1.3965, + "theoretical_loss": 3.480729392048037, + "tokens_seen": 1854275584 + }, + { + "epoch": 0.56, + "learning_rate": 0.0003541344032096289, + "loss": 1.3296, + "theoretical_loss": 3.4807190670551007, + "tokens_seen": 1854341120 + }, + { + "epoch": 0.56, + "learning_rate": 0.0003541183550651956, + "loss": 1.4, + "theoretical_loss": 3.480708742529232, + "tokens_seen": 1854406656 + }, + { + "epoch": 0.56, + "learning_rate": 0.0003541023069207623, + "loss": 1.4315, + "theoretical_loss": 3.480698418470394, + "tokens_seen": 1854472192 + }, + { + "epoch": 0.56, + "learning_rate": 0.000354086258776329, + "loss": 1.3526, + "theoretical_loss": 3.480688094878549, + "tokens_seen": 1854537728 + }, + { + "epoch": 0.56, + "learning_rate": 0.0003540702106318957, + "loss": 1.3755, + "theoretical_loss": 3.480677771753659, + "tokens_seen": 1854603264 + }, + { + "epoch": 0.56, + "objective/train/docs_used": 676483, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.4653351306915283, + "objective/train/theoretical_loss": 3.4806674490956864, + "objective/train/tokens_used": 1875128800, + "theoretical_loss": 3.4806674490956864, + "tokens_seen": 1854668800 + }, + { + "epoch": 0.56, + "learning_rate": 0.0003540541624874624, + "loss": 1.5117, + "theoretical_loss": 3.4806674490956864, + "tokens_seen": 1854668800 + }, + { + "epoch": 0.56, + "learning_rate": 0.0003540381143430291, + "loss": 1.292, + "theoretical_loss": 3.4806571269045943, + "tokens_seen": 1854734336 + }, + { + "epoch": 0.56, + "learning_rate": 0.00035402206619859577, + "loss": 1.4103, + "theoretical_loss": 3.4806468051803443, + "tokens_seen": 1854799872 + }, + { + "epoch": 0.56, + "learning_rate": 0.00035400601805416247, + "loss": 1.3712, + "theoretical_loss": 3.480636483922899, + "tokens_seen": 1854865408 + }, + { + "epoch": 0.56, + "learning_rate": 0.00035398996990972917, + "loss": 1.3513, + "theoretical_loss": 3.480626163132221, + "tokens_seen": 1854930944 + }, + { + "epoch": 0.56, + "learning_rate": 0.00035397392176529587, + "loss": 1.3402, + "theoretical_loss": 3.4806158428082723, + "tokens_seen": 1854996480 + }, + { + "epoch": 0.56, + "learning_rate": 0.00035395787362086257, + "loss": 1.3312, + "theoretical_loss": 3.480605522951016, + "tokens_seen": 1855062016 + }, + { + "epoch": 0.56, + "learning_rate": 0.00035394182547642927, + "loss": 1.4306, + "theoretical_loss": 3.480595203560414, + "tokens_seen": 1855127552 + }, + { + "epoch": 0.56, + "learning_rate": 0.00035392577733199597, + "loss": 1.3459, + "theoretical_loss": 3.4805848846364285, + "tokens_seen": 1855193088 + }, + { + "epoch": 0.56, + "learning_rate": 0.00035390972918756267, + "loss": 1.2953, + "theoretical_loss": 3.480574566179022, + "tokens_seen": 1855258624 + }, + { + "epoch": 0.56, + "learning_rate": 0.00035389368104312937, + "loss": 1.3037, + "theoretical_loss": 3.4805642481881574, + "tokens_seen": 1855324160 + }, + { + "epoch": 0.56, + "learning_rate": 0.00035387763289869607, + "loss": 1.2825, + "theoretical_loss": 3.480553930663797, + "tokens_seen": 1855389696 + }, + { + "epoch": 0.56, + "learning_rate": 0.00035386158475426277, + "loss": 1.389, + "theoretical_loss": 3.480543613605903, + "tokens_seen": 1855455232 + }, + { + "epoch": 0.56, + "learning_rate": 0.00035384553660982947, + "loss": 1.3936, + "theoretical_loss": 3.4805332970144383, + "tokens_seen": 1855520768 + }, + { + "epoch": 0.56, + "learning_rate": 0.00035382948846539617, + "loss": 1.2633, + "theoretical_loss": 3.480522980889364, + "tokens_seen": 1855586304 + }, + { + "epoch": 0.56, + "learning_rate": 0.0003538134403209629, + "loss": 1.3031, + "theoretical_loss": 3.4805126652306444, + "tokens_seen": 1855651840 + }, + { + "epoch": 0.56, + "learning_rate": 0.0003537973921765296, + "loss": 1.2896, + "theoretical_loss": 3.4805023500382406, + "tokens_seen": 1855717376 + }, + { + "epoch": 0.56, + "learning_rate": 0.0003537813440320963, + "loss": 1.3011, + "theoretical_loss": 3.480492035312116, + "tokens_seen": 1855782912 + }, + { + "epoch": 0.56, + "learning_rate": 0.000353765295887663, + "loss": 1.3266, + "theoretical_loss": 3.480481721052232, + "tokens_seen": 1855848448 + }, + { + "epoch": 0.56, + "learning_rate": 0.00035374924774322973, + "loss": 1.4276, + "theoretical_loss": 3.480471407258552, + "tokens_seen": 1855913984 + }, + { + "epoch": 0.56, + "learning_rate": 0.00035373319959879643, + "loss": 1.3676, + "theoretical_loss": 3.480461093931037, + "tokens_seen": 1855979520 + }, + { + "epoch": 0.56, + "learning_rate": 0.00035371715145436313, + "loss": 1.4739, + "theoretical_loss": 3.480450781069652, + "tokens_seen": 1856045056 + }, + { + "epoch": 0.56, + "learning_rate": 0.00035370110330992983, + "loss": 1.3465, + "theoretical_loss": 3.4804404686743577, + "tokens_seen": 1856110592 + }, + { + "epoch": 0.56, + "learning_rate": 0.00035368505516549653, + "loss": 1.2906, + "theoretical_loss": 3.4804301567451166, + "tokens_seen": 1856176128 + }, + { + "epoch": 0.56, + "learning_rate": 0.00035366900702106323, + "loss": 1.2801, + "theoretical_loss": 3.4804198452818915, + "tokens_seen": 1856241664 + }, + { + "epoch": 0.56, + "learning_rate": 0.00035365295887662993, + "loss": 1.2397, + "theoretical_loss": 3.480409534284645, + "tokens_seen": 1856307200 + }, + { + "epoch": 0.56, + "learning_rate": 0.00035363691073219663, + "loss": 1.3315, + "theoretical_loss": 3.4803992237533397, + "tokens_seen": 1856372736 + }, + { + "epoch": 0.56, + "learning_rate": 0.00035362086258776334, + "loss": 1.2727, + "theoretical_loss": 3.4803889136879373, + "tokens_seen": 1856438272 + }, + { + "epoch": 0.56, + "learning_rate": 0.00035360481444333004, + "loss": 1.2943, + "theoretical_loss": 3.4803786040884015, + "tokens_seen": 1856503808 + }, + { + "epoch": 0.56, + "learning_rate": 0.00035358876629889674, + "loss": 1.3884, + "theoretical_loss": 3.480368294954694, + "tokens_seen": 1856569344 + }, + { + "epoch": 0.56, + "learning_rate": 0.00035357271815446344, + "loss": 1.3772, + "theoretical_loss": 3.4803579862867773, + "tokens_seen": 1856634880 + }, + { + "epoch": 0.56, + "learning_rate": 0.00035355667001003014, + "loss": 1.3574, + "theoretical_loss": 3.4803476780846143, + "tokens_seen": 1856700416 + }, + { + "epoch": 0.56, + "learning_rate": 0.00035354062186559684, + "loss": 1.4489, + "theoretical_loss": 3.4803373703481673, + "tokens_seen": 1856765952 + }, + { + "epoch": 0.56, + "learning_rate": 0.00035352457372116354, + "loss": 1.3059, + "theoretical_loss": 3.4803270630773984, + "tokens_seen": 1856831488 + }, + { + "epoch": 0.56, + "learning_rate": 0.00035350852557673024, + "loss": 1.2719, + "theoretical_loss": 3.480316756272271, + "tokens_seen": 1856897024 + }, + { + "epoch": 0.56, + "learning_rate": 0.00035349247743229694, + "loss": 1.3695, + "theoretical_loss": 3.4803064499327476, + "tokens_seen": 1856962560 + }, + { + "epoch": 0.56, + "learning_rate": 0.00035347642928786364, + "loss": 1.3479, + "theoretical_loss": 3.48029614405879, + "tokens_seen": 1857028096 + }, + { + "epoch": 0.56, + "learning_rate": 0.00035346038114343034, + "loss": 1.4102, + "theoretical_loss": 3.480285838650361, + "tokens_seen": 1857093632 + }, + { + "epoch": 0.56, + "learning_rate": 0.00035344433299899704, + "loss": 1.3818, + "theoretical_loss": 3.480275533707423, + "tokens_seen": 1857159168 + }, + { + "epoch": 0.56, + "learning_rate": 0.00035342828485456374, + "loss": 1.2104, + "theoretical_loss": 3.480265229229939, + "tokens_seen": 1857224704 + }, + { + "epoch": 0.56, + "learning_rate": 0.00035341223671013044, + "loss": 1.3472, + "theoretical_loss": 3.4802549252178716, + "tokens_seen": 1857290240 + }, + { + "epoch": 0.56, + "learning_rate": 0.00035339618856569714, + "loss": 1.4107, + "theoretical_loss": 3.4802446216711824, + "tokens_seen": 1857355776 + }, + { + "epoch": 0.56, + "learning_rate": 0.0003533801404212638, + "loss": 1.3458, + "theoretical_loss": 3.4802343185898352, + "tokens_seen": 1857421312 + }, + { + "epoch": 0.56, + "learning_rate": 0.0003533640922768305, + "loss": 1.3492, + "theoretical_loss": 3.480224015973792, + "tokens_seen": 1857486848 + }, + { + "epoch": 0.56, + "learning_rate": 0.0003533480441323972, + "loss": 1.4026, + "theoretical_loss": 3.480213713823015, + "tokens_seen": 1857552384 + }, + { + "epoch": 0.56, + "learning_rate": 0.0003533319959879639, + "loss": 1.3613, + "theoretical_loss": 3.4802034121374676, + "tokens_seen": 1857617920 + }, + { + "epoch": 0.56, + "learning_rate": 0.0003533159478435306, + "loss": 1.2559, + "theoretical_loss": 3.480193110917112, + "tokens_seen": 1857683456 + }, + { + "epoch": 0.56, + "learning_rate": 0.0003532998996990973, + "loss": 1.3869, + "theoretical_loss": 3.48018281016191, + "tokens_seen": 1857748992 + }, + { + "epoch": 0.56, + "learning_rate": 0.000353283851554664, + "loss": 1.3835, + "theoretical_loss": 3.4801725098718257, + "tokens_seen": 1857814528 + }, + { + "epoch": 0.56, + "learning_rate": 0.0003532678034102307, + "loss": 1.2436, + "theoretical_loss": 3.480162210046821, + "tokens_seen": 1857880064 + }, + { + "epoch": 0.56, + "objective/train/docs_used": 677697, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.2340425252914429, + "objective/train/theoretical_loss": 3.4801519106868577, + "objective/train/tokens_used": 1878405600, + "theoretical_loss": 3.4801519106868577, + "tokens_seen": 1857945600 + }, + { + "epoch": 0.56, + "learning_rate": 0.0003532517552657974, + "loss": 1.3111, + "theoretical_loss": 3.4801519106868577, + "tokens_seen": 1857945600 + }, + { + "epoch": 0.56, + "learning_rate": 0.0003532357071213641, + "loss": 1.2803, + "theoretical_loss": 3.480141611791899, + "tokens_seen": 1858011136 + }, + { + "epoch": 0.56, + "learning_rate": 0.0003532196589769308, + "loss": 1.3168, + "theoretical_loss": 3.480131313361908, + "tokens_seen": 1858076672 + }, + { + "epoch": 0.56, + "learning_rate": 0.0003532036108324975, + "loss": 1.3542, + "theoretical_loss": 3.480121015396847, + "tokens_seen": 1858142208 + }, + { + "epoch": 0.56, + "learning_rate": 0.0003531875626880642, + "loss": 1.3603, + "theoretical_loss": 3.480110717896679, + "tokens_seen": 1858207744 + }, + { + "epoch": 0.56, + "learning_rate": 0.0003531715145436309, + "loss": 1.3786, + "theoretical_loss": 3.480100420861365, + "tokens_seen": 1858273280 + }, + { + "epoch": 0.56, + "learning_rate": 0.0003531554663991976, + "loss": 1.3411, + "theoretical_loss": 3.4800901242908693, + "tokens_seen": 1858338816 + }, + { + "epoch": 0.56, + "learning_rate": 0.0003531394182547643, + "loss": 1.304, + "theoretical_loss": 3.480079828185154, + "tokens_seen": 1858404352 + }, + { + "epoch": 0.56, + "learning_rate": 0.000353123370110331, + "loss": 1.3053, + "theoretical_loss": 3.480069532544182, + "tokens_seen": 1858469888 + }, + { + "epoch": 0.56, + "learning_rate": 0.0003531073219658977, + "loss": 1.2489, + "theoretical_loss": 3.4800592373679153, + "tokens_seen": 1858535424 + }, + { + "epoch": 0.56, + "learning_rate": 0.0003530912738214644, + "loss": 1.2896, + "theoretical_loss": 3.4800489426563166, + "tokens_seen": 1858600960 + }, + { + "epoch": 0.56, + "learning_rate": 0.0003530752256770311, + "loss": 1.4238, + "theoretical_loss": 3.4800386484093497, + "tokens_seen": 1858666496 + }, + { + "epoch": 0.56, + "learning_rate": 0.0003530591775325978, + "loss": 1.3887, + "theoretical_loss": 3.4800283546269757, + "tokens_seen": 1858732032 + }, + { + "epoch": 0.56, + "learning_rate": 0.0003530431293881645, + "loss": 1.3562, + "theoretical_loss": 3.480018061309158, + "tokens_seen": 1858797568 + }, + { + "epoch": 0.56, + "learning_rate": 0.0003530270812437312, + "loss": 1.3824, + "theoretical_loss": 3.4800077684558595, + "tokens_seen": 1858863104 + }, + { + "epoch": 0.56, + "learning_rate": 0.0003530110330992979, + "loss": 1.256, + "theoretical_loss": 3.4799974760670427, + "tokens_seen": 1858928640 + }, + { + "epoch": 0.56, + "learning_rate": 0.0003529949849548646, + "loss": 1.275, + "theoretical_loss": 3.4799871841426695, + "tokens_seen": 1858994176 + }, + { + "epoch": 0.56, + "learning_rate": 0.0003529789368104313, + "loss": 1.2686, + "theoretical_loss": 3.4799768926827035, + "tokens_seen": 1859059712 + }, + { + "epoch": 0.56, + "learning_rate": 0.000352962888665998, + "loss": 1.2542, + "theoretical_loss": 3.4799666016871074, + "tokens_seen": 1859125248 + }, + { + "epoch": 0.56, + "learning_rate": 0.0003529468405215647, + "loss": 1.3256, + "theoretical_loss": 3.479956311155843, + "tokens_seen": 1859190784 + }, + { + "epoch": 0.56, + "learning_rate": 0.0003529307923771314, + "loss": 1.3307, + "theoretical_loss": 3.4799460210888737, + "tokens_seen": 1859256320 + }, + { + "epoch": 0.56, + "learning_rate": 0.0003529147442326981, + "loss": 1.302, + "theoretical_loss": 3.4799357314861625, + "tokens_seen": 1859321856 + }, + { + "epoch": 0.56, + "learning_rate": 0.0003528986960882648, + "loss": 1.3149, + "theoretical_loss": 3.479925442347671, + "tokens_seen": 1859387392 + }, + { + "epoch": 0.56, + "learning_rate": 0.0003528826479438315, + "loss": 1.3149, + "theoretical_loss": 3.479915153673363, + "tokens_seen": 1859452928 + }, + { + "epoch": 0.56, + "learning_rate": 0.0003528665997993982, + "loss": 1.2398, + "theoretical_loss": 3.4799048654632, + "tokens_seen": 1859518464 + }, + { + "epoch": 0.56, + "learning_rate": 0.0003528505516549649, + "loss": 1.3572, + "theoretical_loss": 3.479894577717146, + "tokens_seen": 1859584000 + }, + { + "epoch": 0.56, + "learning_rate": 0.0003528345035105316, + "loss": 1.2777, + "theoretical_loss": 3.479884290435163, + "tokens_seen": 1859649536 + }, + { + "epoch": 0.56, + "learning_rate": 0.0003528184553660983, + "loss": 1.321, + "theoretical_loss": 3.4798740036172138, + "tokens_seen": 1859715072 + }, + { + "epoch": 0.56, + "learning_rate": 0.000352802407221665, + "loss": 1.4605, + "theoretical_loss": 3.479863717263261, + "tokens_seen": 1859780608 + }, + { + "epoch": 0.56, + "learning_rate": 0.0003527863590772317, + "loss": 1.2565, + "theoretical_loss": 3.4798534313732676, + "tokens_seen": 1859846144 + }, + { + "epoch": 0.56, + "learning_rate": 0.0003527703109327984, + "loss": 1.462, + "theoretical_loss": 3.4798431459471963, + "tokens_seen": 1859911680 + }, + { + "epoch": 0.56, + "learning_rate": 0.0003527542627883651, + "loss": 1.2945, + "theoretical_loss": 3.4798328609850095, + "tokens_seen": 1859977216 + }, + { + "epoch": 0.56, + "learning_rate": 0.0003527382146439318, + "loss": 1.3174, + "theoretical_loss": 3.4798225764866704, + "tokens_seen": 1860042752 + }, + { + "epoch": 0.56, + "learning_rate": 0.0003527221664994985, + "loss": 1.2265, + "theoretical_loss": 3.4798122924521415, + "tokens_seen": 1860108288 + }, + { + "epoch": 0.56, + "learning_rate": 0.0003527061183550652, + "loss": 1.3338, + "theoretical_loss": 3.4798020088813852, + "tokens_seen": 1860173824 + }, + { + "epoch": 0.56, + "learning_rate": 0.0003526900702106319, + "loss": 1.3666, + "theoretical_loss": 3.4797917257743647, + "tokens_seen": 1860239360 + }, + { + "epoch": 0.56, + "learning_rate": 0.0003526740220661986, + "loss": 1.3702, + "theoretical_loss": 3.479781443131043, + "tokens_seen": 1860304896 + }, + { + "epoch": 0.56, + "learning_rate": 0.0003526579739217653, + "loss": 1.3177, + "theoretical_loss": 3.479771160951382, + "tokens_seen": 1860370432 + }, + { + "epoch": 0.56, + "learning_rate": 0.000352641925777332, + "loss": 1.4216, + "theoretical_loss": 3.479760879235345, + "tokens_seen": 1860435968 + }, + { + "epoch": 0.56, + "learning_rate": 0.0003526258776328987, + "loss": 1.3065, + "theoretical_loss": 3.4797505979828953, + "tokens_seen": 1860501504 + }, + { + "epoch": 0.56, + "learning_rate": 0.00035260982948846537, + "loss": 1.4734, + "theoretical_loss": 3.4797403171939942, + "tokens_seen": 1860567040 + }, + { + "epoch": 0.56, + "learning_rate": 0.00035259378134403207, + "loss": 1.3446, + "theoretical_loss": 3.479730036868606, + "tokens_seen": 1860632576 + }, + { + "epoch": 0.56, + "learning_rate": 0.00035257773319959877, + "loss": 1.3087, + "theoretical_loss": 3.479719757006693, + "tokens_seen": 1860698112 + }, + { + "epoch": 0.56, + "learning_rate": 0.00035256168505516547, + "loss": 1.3297, + "theoretical_loss": 3.4797094776082176, + "tokens_seen": 1860763648 + }, + { + "epoch": 0.56, + "learning_rate": 0.0003525456369107322, + "loss": 1.3352, + "theoretical_loss": 3.4796991986731425, + "tokens_seen": 1860829184 + }, + { + "epoch": 0.56, + "learning_rate": 0.0003525295887662989, + "loss": 1.3039, + "theoretical_loss": 3.479688920201431, + "tokens_seen": 1860894720 + }, + { + "epoch": 0.56, + "learning_rate": 0.0003525135406218656, + "loss": 1.3, + "theoretical_loss": 3.4796786421930457, + "tokens_seen": 1860960256 + }, + { + "epoch": 0.56, + "learning_rate": 0.0003524974924774323, + "loss": 1.4169, + "theoretical_loss": 3.4796683646479494, + "tokens_seen": 1861025792 + }, + { + "epoch": 0.56, + "learning_rate": 0.000352481444332999, + "loss": 1.4366, + "theoretical_loss": 3.479658087566105, + "tokens_seen": 1861091328 + }, + { + "epoch": 0.56, + "learning_rate": 0.0003524653961885657, + "loss": 1.3168, + "theoretical_loss": 3.4796478109474753, + "tokens_seen": 1861156864 + }, + { + "epoch": 0.56, + "objective/train/docs_used": 678968, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.3742542266845703, + "objective/train/theoretical_loss": 3.4796375347920225, + "objective/train/tokens_used": 1881682400, + "theoretical_loss": 3.4796375347920225, + "tokens_seen": 1861222400 + }, + { + "epoch": 0.56, + "learning_rate": 0.00035244934804413243, + "loss": 1.3313, + "theoretical_loss": 3.4796375347920225, + "tokens_seen": 1861222400 + }, + { + "epoch": 0.56, + "learning_rate": 0.00035243329989969913, + "loss": 1.3698, + "theoretical_loss": 3.4796272590997104, + "tokens_seen": 1861287936 + }, + { + "epoch": 0.56, + "learning_rate": 0.00035241725175526583, + "loss": 1.2618, + "theoretical_loss": 3.4796169838705016, + "tokens_seen": 1861353472 + }, + { + "epoch": 0.56, + "learning_rate": 0.00035240120361083253, + "loss": 1.3591, + "theoretical_loss": 3.479606709104358, + "tokens_seen": 1861419008 + }, + { + "epoch": 0.56, + "learning_rate": 0.00035238515546639923, + "loss": 1.2998, + "theoretical_loss": 3.4795964348012434, + "tokens_seen": 1861484544 + }, + { + "epoch": 0.56, + "learning_rate": 0.00035236910732196593, + "loss": 1.3242, + "theoretical_loss": 3.4795861609611203, + "tokens_seen": 1861550080 + }, + { + "epoch": 0.56, + "learning_rate": 0.00035235305917753263, + "loss": 1.356, + "theoretical_loss": 3.4795758875839518, + "tokens_seen": 1861615616 + }, + { + "epoch": 0.56, + "learning_rate": 0.00035233701103309933, + "loss": 1.3503, + "theoretical_loss": 3.4795656146697, + "tokens_seen": 1861681152 + }, + { + "epoch": 0.56, + "learning_rate": 0.00035232096288866603, + "loss": 1.3729, + "theoretical_loss": 3.4795553422183287, + "tokens_seen": 1861746688 + }, + { + "epoch": 0.56, + "learning_rate": 0.00035230491474423273, + "loss": 1.4324, + "theoretical_loss": 3.4795450702298005, + "tokens_seen": 1861812224 + }, + { + "epoch": 0.56, + "learning_rate": 0.00035228886659979943, + "loss": 1.463, + "theoretical_loss": 3.4795347987040777, + "tokens_seen": 1861877760 + }, + { + "epoch": 0.56, + "learning_rate": 0.00035227281845536613, + "loss": 1.3345, + "theoretical_loss": 3.4795245276411233, + "tokens_seen": 1861943296 + }, + { + "epoch": 0.56, + "learning_rate": 0.00035225677031093283, + "loss": 1.385, + "theoretical_loss": 3.4795142570409006, + "tokens_seen": 1862008832 + }, + { + "epoch": 0.56, + "learning_rate": 0.00035224072216649954, + "loss": 1.2754, + "theoretical_loss": 3.4795039869033726, + "tokens_seen": 1862074368 + }, + { + "epoch": 0.56, + "learning_rate": 0.00035222467402206624, + "loss": 1.2024, + "theoretical_loss": 3.4794937172285016, + "tokens_seen": 1862139904 + }, + { + "epoch": 0.56, + "learning_rate": 0.00035220862587763294, + "loss": 1.3339, + "theoretical_loss": 3.479483448016251, + "tokens_seen": 1862205440 + }, + { + "epoch": 0.56, + "learning_rate": 0.00035219257773319964, + "loss": 1.3643, + "theoretical_loss": 3.479473179266583, + "tokens_seen": 1862270976 + }, + { + "epoch": 0.56, + "learning_rate": 0.00035217652958876634, + "loss": 1.3708, + "theoretical_loss": 3.4794629109794606, + "tokens_seen": 1862336512 + }, + { + "epoch": 0.56, + "learning_rate": 0.00035216048144433304, + "loss": 1.3446, + "theoretical_loss": 3.479452643154848, + "tokens_seen": 1862402048 + }, + { + "epoch": 0.56, + "learning_rate": 0.00035214443329989974, + "loss": 1.3764, + "theoretical_loss": 3.479442375792706, + "tokens_seen": 1862467584 + }, + { + "epoch": 0.56, + "learning_rate": 0.00035212838515546644, + "loss": 1.3778, + "theoretical_loss": 3.4794321088929987, + "tokens_seen": 1862533120 + }, + { + "epoch": 0.56, + "learning_rate": 0.00035211233701103314, + "loss": 1.3051, + "theoretical_loss": 3.4794218424556895, + "tokens_seen": 1862598656 + }, + { + "epoch": 0.56, + "learning_rate": 0.00035209628886659984, + "loss": 1.2998, + "theoretical_loss": 3.4794115764807403, + "tokens_seen": 1862664192 + }, + { + "epoch": 0.56, + "learning_rate": 0.00035208024072216654, + "loss": 1.3512, + "theoretical_loss": 3.479401310968114, + "tokens_seen": 1862729728 + }, + { + "epoch": 0.56, + "learning_rate": 0.00035206419257773324, + "loss": 1.3836, + "theoretical_loss": 3.479391045917774, + "tokens_seen": 1862795264 + }, + { + "epoch": 0.56, + "learning_rate": 0.00035204814443329994, + "loss": 1.366, + "theoretical_loss": 3.4793807813296835, + "tokens_seen": 1862860800 + }, + { + "epoch": 0.56, + "learning_rate": 0.00035203209628886664, + "loss": 1.2987, + "theoretical_loss": 3.479370517203805, + "tokens_seen": 1862926336 + }, + { + "epoch": 0.56, + "learning_rate": 0.00035201604814443334, + "loss": 1.3607, + "theoretical_loss": 3.4793602535401016, + "tokens_seen": 1862991872 + }, + { + "epoch": 0.56, + "learning_rate": 0.00035200000000000005, + "loss": 1.3431, + "theoretical_loss": 3.4793499903385356, + "tokens_seen": 1863057408 + }, + { + "epoch": 0.56, + "learning_rate": 0.0003519839518555667, + "loss": 1.3694, + "theoretical_loss": 3.4793397275990703, + "tokens_seen": 1863122944 + }, + { + "epoch": 0.56, + "learning_rate": 0.0003519679037111334, + "loss": 1.3727, + "theoretical_loss": 3.479329465321669, + "tokens_seen": 1863188480 + }, + { + "epoch": 0.56, + "learning_rate": 0.0003519518555667001, + "loss": 1.3574, + "theoretical_loss": 3.4793192035062948, + "tokens_seen": 1863254016 + }, + { + "epoch": 0.56, + "learning_rate": 0.0003519358074222668, + "loss": 1.4724, + "theoretical_loss": 3.47930894215291, + "tokens_seen": 1863319552 + }, + { + "epoch": 0.56, + "learning_rate": 0.0003519197592778335, + "loss": 1.4356, + "theoretical_loss": 3.4792986812614775, + "tokens_seen": 1863385088 + }, + { + "epoch": 0.56, + "learning_rate": 0.0003519037111334002, + "loss": 1.2173, + "theoretical_loss": 3.479288420831961, + "tokens_seen": 1863450624 + }, + { + "epoch": 0.56, + "learning_rate": 0.0003518876629889669, + "loss": 1.3874, + "theoretical_loss": 3.479278160864323, + "tokens_seen": 1863516160 + }, + { + "epoch": 0.56, + "learning_rate": 0.0003518716148445336, + "loss": 1.2862, + "theoretical_loss": 3.4792679013585266, + "tokens_seen": 1863581696 + }, + { + "epoch": 0.56, + "learning_rate": 0.0003518555667001003, + "loss": 1.4193, + "theoretical_loss": 3.4792576423145345, + "tokens_seen": 1863647232 + }, + { + "epoch": 0.56, + "learning_rate": 0.000351839518555667, + "loss": 1.353, + "theoretical_loss": 3.4792473837323095, + "tokens_seen": 1863712768 + }, + { + "epoch": 0.56, + "learning_rate": 0.0003518234704112337, + "loss": 1.366, + "theoretical_loss": 3.4792371256118155, + "tokens_seen": 1863778304 + }, + { + "epoch": 0.56, + "learning_rate": 0.0003518074222668004, + "loss": 1.273, + "theoretical_loss": 3.479226867953014, + "tokens_seen": 1863843840 + }, + { + "epoch": 0.56, + "learning_rate": 0.0003517913741223671, + "loss": 1.3672, + "theoretical_loss": 3.47921661075587, + "tokens_seen": 1863909376 + }, + { + "epoch": 0.56, + "learning_rate": 0.0003517753259779338, + "loss": 1.3987, + "theoretical_loss": 3.4792063540203446, + "tokens_seen": 1863974912 + }, + { + "epoch": 0.56, + "learning_rate": 0.0003517592778335005, + "loss": 1.3689, + "theoretical_loss": 3.479196097746402, + "tokens_seen": 1864040448 + }, + { + "epoch": 0.56, + "learning_rate": 0.0003517432296890672, + "loss": 1.4215, + "theoretical_loss": 3.4791858419340045, + "tokens_seen": 1864105984 + }, + { + "epoch": 0.56, + "learning_rate": 0.0003517271815446339, + "loss": 1.3255, + "theoretical_loss": 3.479175586583116, + "tokens_seen": 1864171520 + }, + { + "epoch": 0.56, + "learning_rate": 0.0003517111334002006, + "loss": 1.3832, + "theoretical_loss": 3.4791653316936983, + "tokens_seen": 1864237056 + }, + { + "epoch": 0.56, + "learning_rate": 0.0003516950852557673, + "loss": 1.3261, + "theoretical_loss": 3.479155077265715, + "tokens_seen": 1864302592 + }, + { + "epoch": 0.56, + "learning_rate": 0.000351679037111334, + "loss": 1.3871, + "theoretical_loss": 3.479144823299129, + "tokens_seen": 1864368128 + }, + { + "epoch": 0.56, + "learning_rate": 0.0003516629889669007, + "loss": 1.2134, + "theoretical_loss": 3.479134569793904, + "tokens_seen": 1864433664 + }, + { + "epoch": 0.56, + "objective/train/docs_used": 680280, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.1466598510742188, + "objective/train/theoretical_loss": 3.479124316750002, + "objective/train/tokens_used": 1884959200, + "theoretical_loss": 3.479124316750002, + "tokens_seen": 1864499200 + }, + { + "epoch": 0.56, + "learning_rate": 0.0003516469408224674, + "loss": 1.2356, + "theoretical_loss": 3.479124316750002, + "tokens_seen": 1864499200 + }, + { + "epoch": 0.57, + "learning_rate": 0.0003516308926780341, + "loss": 1.3644, + "theoretical_loss": 3.479114064167387, + "tokens_seen": 1864564736 + }, + { + "epoch": 0.57, + "learning_rate": 0.0003516148445336008, + "loss": 1.432, + "theoretical_loss": 3.479103812046021, + "tokens_seen": 1864630272 + }, + { + "epoch": 0.57, + "learning_rate": 0.0003515987963891675, + "loss": 1.2273, + "theoretical_loss": 3.4790935603858677, + "tokens_seen": 1864695808 + }, + { + "epoch": 0.57, + "learning_rate": 0.0003515827482447342, + "loss": 1.2993, + "theoretical_loss": 3.4790833091868905, + "tokens_seen": 1864761344 + }, + { + "epoch": 0.57, + "learning_rate": 0.0003515667001003009, + "loss": 1.2828, + "theoretical_loss": 3.4790730584490515, + "tokens_seen": 1864826880 + }, + { + "epoch": 0.57, + "learning_rate": 0.0003515506519558676, + "loss": 1.2864, + "theoretical_loss": 3.479062808172314, + "tokens_seen": 1864892416 + }, + { + "epoch": 0.57, + "learning_rate": 0.0003515346038114343, + "loss": 1.3587, + "theoretical_loss": 3.479052558356642, + "tokens_seen": 1864957952 + }, + { + "epoch": 0.57, + "learning_rate": 0.000351518555667001, + "loss": 1.3726, + "theoretical_loss": 3.479042309001997, + "tokens_seen": 1865023488 + }, + { + "epoch": 0.57, + "learning_rate": 0.0003515025075225677, + "loss": 1.4105, + "theoretical_loss": 3.4790320601083433, + "tokens_seen": 1865089024 + }, + { + "epoch": 0.57, + "learning_rate": 0.0003514864593781344, + "loss": 1.2784, + "theoretical_loss": 3.4790218116756435, + "tokens_seen": 1865154560 + }, + { + "epoch": 0.57, + "learning_rate": 0.0003514704112337011, + "loss": 1.3341, + "theoretical_loss": 3.4790115637038608, + "tokens_seen": 1865220096 + }, + { + "epoch": 0.57, + "learning_rate": 0.0003514543630892678, + "loss": 1.3212, + "theoretical_loss": 3.4790013161929583, + "tokens_seen": 1865285632 + }, + { + "epoch": 0.57, + "learning_rate": 0.0003514383149448345, + "loss": 1.3672, + "theoretical_loss": 3.478991069142899, + "tokens_seen": 1865351168 + }, + { + "epoch": 0.57, + "learning_rate": 0.0003514222668004012, + "loss": 1.3866, + "theoretical_loss": 3.4789808225536456, + "tokens_seen": 1865416704 + }, + { + "epoch": 0.57, + "learning_rate": 0.0003514062186559679, + "loss": 1.4063, + "theoretical_loss": 3.478970576425162, + "tokens_seen": 1865482240 + }, + { + "epoch": 0.57, + "learning_rate": 0.0003513901705115346, + "loss": 1.3853, + "theoretical_loss": 3.4789603307574106, + "tokens_seen": 1865547776 + }, + { + "epoch": 0.57, + "learning_rate": 0.0003513741223671013, + "loss": 1.3735, + "theoretical_loss": 3.478950085550355, + "tokens_seen": 1865613312 + }, + { + "epoch": 0.57, + "learning_rate": 0.000351358074222668, + "loss": 1.3498, + "theoretical_loss": 3.478939840803958, + "tokens_seen": 1865678848 + }, + { + "epoch": 0.57, + "learning_rate": 0.0003513420260782347, + "loss": 1.3789, + "theoretical_loss": 3.478929596518183, + "tokens_seen": 1865744384 + }, + { + "epoch": 0.57, + "learning_rate": 0.0003513259779338014, + "loss": 1.2828, + "theoretical_loss": 3.4789193526929925, + "tokens_seen": 1865809920 + }, + { + "epoch": 0.57, + "learning_rate": 0.0003513099297893681, + "loss": 1.3756, + "theoretical_loss": 3.4789091093283497, + "tokens_seen": 1865875456 + }, + { + "epoch": 0.57, + "learning_rate": 0.0003512938816449348, + "loss": 1.3144, + "theoretical_loss": 3.4788988664242186, + "tokens_seen": 1865940992 + }, + { + "epoch": 0.57, + "learning_rate": 0.0003512778335005015, + "loss": 1.3171, + "theoretical_loss": 3.478888623980562, + "tokens_seen": 1866006528 + }, + { + "epoch": 0.57, + "learning_rate": 0.0003512617853560682, + "loss": 1.4514, + "theoretical_loss": 3.478878381997342, + "tokens_seen": 1866072064 + }, + { + "epoch": 0.57, + "learning_rate": 0.0003512457372116349, + "loss": 1.3992, + "theoretical_loss": 3.478868140474523, + "tokens_seen": 1866137600 + }, + { + "epoch": 0.57, + "learning_rate": 0.0003512296890672016, + "loss": 1.3806, + "theoretical_loss": 3.478857899412067, + "tokens_seen": 1866203136 + }, + { + "epoch": 0.57, + "learning_rate": 0.0003512136409227683, + "loss": 1.3474, + "theoretical_loss": 3.478847658809938, + "tokens_seen": 1866268672 + }, + { + "epoch": 0.57, + "learning_rate": 0.000351197592778335, + "loss": 1.3843, + "theoretical_loss": 3.4788374186680997, + "tokens_seen": 1866334208 + }, + { + "epoch": 0.57, + "learning_rate": 0.0003511815446339017, + "loss": 1.4097, + "theoretical_loss": 3.4788271789865135, + "tokens_seen": 1866399744 + }, + { + "epoch": 0.57, + "learning_rate": 0.0003511654964894684, + "loss": 1.3283, + "theoretical_loss": 3.478816939765144, + "tokens_seen": 1866465280 + }, + { + "epoch": 0.57, + "learning_rate": 0.0003511494483450351, + "loss": 1.3215, + "theoretical_loss": 3.478806701003954, + "tokens_seen": 1866530816 + }, + { + "epoch": 0.57, + "learning_rate": 0.0003511334002006018, + "loss": 1.2534, + "theoretical_loss": 3.478796462702906, + "tokens_seen": 1866596352 + }, + { + "epoch": 0.57, + "learning_rate": 0.0003511173520561685, + "loss": 1.3414, + "theoretical_loss": 3.4787862248619645, + "tokens_seen": 1866661888 + }, + { + "epoch": 0.57, + "learning_rate": 0.0003511013039117352, + "loss": 1.319, + "theoretical_loss": 3.4787759874810913, + "tokens_seen": 1866727424 + }, + { + "epoch": 0.57, + "learning_rate": 0.0003510852557673019, + "loss": 1.4489, + "theoretical_loss": 3.4787657505602505, + "tokens_seen": 1866792960 + }, + { + "epoch": 0.57, + "learning_rate": 0.00035106920762286863, + "loss": 1.2713, + "theoretical_loss": 3.478755514099405, + "tokens_seen": 1866858496 + }, + { + "epoch": 0.57, + "learning_rate": 0.00035105315947843533, + "loss": 1.2741, + "theoretical_loss": 3.4787452780985175, + "tokens_seen": 1866924032 + }, + { + "epoch": 0.57, + "learning_rate": 0.00035103711133400203, + "loss": 1.4004, + "theoretical_loss": 3.4787350425575516, + "tokens_seen": 1866989568 + }, + { + "epoch": 0.57, + "learning_rate": 0.00035102106318956873, + "loss": 1.2977, + "theoretical_loss": 3.478724807476471, + "tokens_seen": 1867055104 + }, + { + "epoch": 0.57, + "learning_rate": 0.00035100501504513543, + "loss": 1.3828, + "theoretical_loss": 3.478714572855238, + "tokens_seen": 1867120640 + }, + { + "epoch": 0.57, + "learning_rate": 0.00035098896690070213, + "loss": 1.2958, + "theoretical_loss": 3.4787043386938166, + "tokens_seen": 1867186176 + }, + { + "epoch": 0.57, + "learning_rate": 0.00035097291875626883, + "loss": 1.3021, + "theoretical_loss": 3.4786941049921687, + "tokens_seen": 1867251712 + }, + { + "epoch": 0.57, + "learning_rate": 0.00035095687061183553, + "loss": 1.4947, + "theoretical_loss": 3.4786838717502593, + "tokens_seen": 1867317248 + }, + { + "epoch": 0.57, + "learning_rate": 0.00035094082246740223, + "loss": 1.3343, + "theoretical_loss": 3.4786736389680506, + "tokens_seen": 1867382784 + }, + { + "epoch": 0.57, + "learning_rate": 0.00035092477432296893, + "loss": 1.3579, + "theoretical_loss": 3.4786634066455058, + "tokens_seen": 1867448320 + }, + { + "epoch": 0.57, + "learning_rate": 0.00035090872617853563, + "loss": 1.3183, + "theoretical_loss": 3.478653174782588, + "tokens_seen": 1867513856 + }, + { + "epoch": 0.57, + "learning_rate": 0.00035089267803410233, + "loss": 1.3253, + "theoretical_loss": 3.4786429433792607, + "tokens_seen": 1867579392 + }, + { + "epoch": 0.57, + "learning_rate": 0.00035087662988966904, + "loss": 1.3537, + "theoretical_loss": 3.4786327124354877, + "tokens_seen": 1867644928 + }, + { + "epoch": 0.57, + "learning_rate": 0.00035086058174523574, + "loss": 1.3035, + "theoretical_loss": 3.4786224819512306, + "tokens_seen": 1867710464 + }, + { + "epoch": 0.57, + "objective/train/docs_used": 681429, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.2997679710388184, + "objective/train/theoretical_loss": 3.478612251926455, + "objective/train/tokens_used": 1888236000, + "theoretical_loss": 3.478612251926455, + "tokens_seen": 1867776000 + }, + { + "epoch": 0.57, + "learning_rate": 0.00035084453360080244, + "loss": 1.3004, + "theoretical_loss": 3.478612251926455, + "tokens_seen": 1867776000 + }, + { + "epoch": 0.57, + "learning_rate": 0.00035082848545636914, + "loss": 1.3008, + "theoretical_loss": 3.478602022361122, + "tokens_seen": 1867841536 + }, + { + "epoch": 0.57, + "learning_rate": 0.00035081243731193584, + "loss": 1.4045, + "theoretical_loss": 3.4785917932551955, + "tokens_seen": 1867907072 + }, + { + "epoch": 0.57, + "learning_rate": 0.00035079638916750254, + "loss": 1.3206, + "theoretical_loss": 3.478581564608639, + "tokens_seen": 1867972608 + }, + { + "epoch": 0.57, + "learning_rate": 0.00035078034102306924, + "loss": 1.3723, + "theoretical_loss": 3.4785713364214166, + "tokens_seen": 1868038144 + }, + { + "epoch": 0.57, + "learning_rate": 0.00035076429287863594, + "loss": 1.354, + "theoretical_loss": 3.47856110869349, + "tokens_seen": 1868103680 + }, + { + "epoch": 0.57, + "learning_rate": 0.00035074824473420264, + "loss": 1.3079, + "theoretical_loss": 3.4785508814248227, + "tokens_seen": 1868169216 + }, + { + "epoch": 0.57, + "learning_rate": 0.00035073219658976934, + "loss": 1.2913, + "theoretical_loss": 3.478540654615379, + "tokens_seen": 1868234752 + }, + { + "epoch": 0.57, + "learning_rate": 0.00035071614844533604, + "loss": 1.2968, + "theoretical_loss": 3.4785304282651213, + "tokens_seen": 1868300288 + }, + { + "epoch": 0.57, + "learning_rate": 0.00035070010030090274, + "loss": 1.2674, + "theoretical_loss": 3.478520202374013, + "tokens_seen": 1868365824 + }, + { + "epoch": 0.57, + "learning_rate": 0.00035068405215646944, + "loss": 1.3756, + "theoretical_loss": 3.478509976942018, + "tokens_seen": 1868431360 + }, + { + "epoch": 0.57, + "learning_rate": 0.00035066800401203614, + "loss": 1.2753, + "theoretical_loss": 3.478499751969098, + "tokens_seen": 1868496896 + }, + { + "epoch": 0.57, + "learning_rate": 0.00035065195586760284, + "loss": 1.2908, + "theoretical_loss": 3.4784895274552183, + "tokens_seen": 1868562432 + }, + { + "epoch": 0.57, + "learning_rate": 0.00035063590772316954, + "loss": 1.3543, + "theoretical_loss": 3.478479303400341, + "tokens_seen": 1868627968 + }, + { + "epoch": 0.57, + "learning_rate": 0.00035061985957873625, + "loss": 1.5043, + "theoretical_loss": 3.4784690798044298, + "tokens_seen": 1868693504 + }, + { + "epoch": 0.57, + "learning_rate": 0.00035060381143430295, + "loss": 1.3159, + "theoretical_loss": 3.478458856667448, + "tokens_seen": 1868759040 + }, + { + "epoch": 0.57, + "learning_rate": 0.00035058776328986965, + "loss": 1.2948, + "theoretical_loss": 3.478448633989358, + "tokens_seen": 1868824576 + }, + { + "epoch": 0.57, + "learning_rate": 0.0003505717151454363, + "loss": 1.4671, + "theoretical_loss": 3.4784384117701244, + "tokens_seen": 1868890112 + }, + { + "epoch": 0.57, + "learning_rate": 0.000350555667001003, + "loss": 1.4299, + "theoretical_loss": 3.4784281900097103, + "tokens_seen": 1868955648 + }, + { + "epoch": 0.57, + "learning_rate": 0.0003505396188565697, + "loss": 1.2735, + "theoretical_loss": 3.4784179687080785, + "tokens_seen": 1869021184 + }, + { + "epoch": 0.57, + "learning_rate": 0.0003505235707121364, + "loss": 1.3688, + "theoretical_loss": 3.4784077478651922, + "tokens_seen": 1869086720 + }, + { + "epoch": 0.57, + "learning_rate": 0.0003505075225677031, + "loss": 1.3499, + "theoretical_loss": 3.4783975274810155, + "tokens_seen": 1869152256 + }, + { + "epoch": 0.57, + "learning_rate": 0.0003504914744232698, + "loss": 1.4425, + "theoretical_loss": 3.478387307555511, + "tokens_seen": 1869217792 + }, + { + "epoch": 0.57, + "learning_rate": 0.0003504754262788365, + "loss": 1.3243, + "theoretical_loss": 3.478377088088642, + "tokens_seen": 1869283328 + }, + { + "epoch": 0.57, + "learning_rate": 0.0003504593781344032, + "loss": 1.3484, + "theoretical_loss": 3.4783668690803733, + "tokens_seen": 1869348864 + }, + { + "epoch": 0.57, + "learning_rate": 0.0003504433299899699, + "loss": 1.3289, + "theoretical_loss": 3.478356650530666, + "tokens_seen": 1869414400 + }, + { + "epoch": 0.57, + "learning_rate": 0.0003504272818455366, + "loss": 1.4099, + "theoretical_loss": 3.4783464324394853, + "tokens_seen": 1869479936 + }, + { + "epoch": 0.57, + "learning_rate": 0.0003504112337011033, + "loss": 1.3651, + "theoretical_loss": 3.4783362148067933, + "tokens_seen": 1869545472 + }, + { + "epoch": 0.57, + "learning_rate": 0.00035039518555667, + "loss": 1.3754, + "theoretical_loss": 3.4783259976325542, + "tokens_seen": 1869611008 + }, + { + "epoch": 0.57, + "learning_rate": 0.0003503791374122367, + "loss": 1.3919, + "theoretical_loss": 3.4783157809167307, + "tokens_seen": 1869676544 + }, + { + "epoch": 0.57, + "learning_rate": 0.0003503630892678034, + "loss": 1.4171, + "theoretical_loss": 3.4783055646592866, + "tokens_seen": 1869742080 + }, + { + "epoch": 0.57, + "learning_rate": 0.0003503470411233701, + "loss": 1.4117, + "theoretical_loss": 3.4782953488601853, + "tokens_seen": 1869807616 + }, + { + "epoch": 0.57, + "learning_rate": 0.0003503309929789368, + "loss": 1.3966, + "theoretical_loss": 3.47828513351939, + "tokens_seen": 1869873152 + }, + { + "epoch": 0.57, + "learning_rate": 0.0003503149448345035, + "loss": 1.322, + "theoretical_loss": 3.478274918636864, + "tokens_seen": 1869938688 + }, + { + "epoch": 0.57, + "learning_rate": 0.0003502988966900702, + "loss": 1.473, + "theoretical_loss": 3.478264704212571, + "tokens_seen": 1870004224 + }, + { + "epoch": 0.57, + "learning_rate": 0.0003502828485456369, + "loss": 1.3224, + "theoretical_loss": 3.4782544902464743, + "tokens_seen": 1870069760 + }, + { + "epoch": 0.57, + "learning_rate": 0.0003502668004012036, + "loss": 1.3996, + "theoretical_loss": 3.4782442767385366, + "tokens_seen": 1870135296 + }, + { + "epoch": 0.57, + "learning_rate": 0.0003502507522567703, + "loss": 1.3689, + "theoretical_loss": 3.4782340636887223, + "tokens_seen": 1870200832 + }, + { + "epoch": 0.57, + "learning_rate": 0.000350234704112337, + "loss": 1.3515, + "theoretical_loss": 3.4782238510969945, + "tokens_seen": 1870266368 + }, + { + "epoch": 0.57, + "learning_rate": 0.0003502186559679037, + "loss": 1.3096, + "theoretical_loss": 3.4782136389633163, + "tokens_seen": 1870331904 + }, + { + "epoch": 0.57, + "learning_rate": 0.0003502026078234704, + "loss": 1.379, + "theoretical_loss": 3.478203427287651, + "tokens_seen": 1870397440 + }, + { + "epoch": 0.57, + "learning_rate": 0.0003501865596790371, + "loss": 1.4152, + "theoretical_loss": 3.4781932160699625, + "tokens_seen": 1870462976 + }, + { + "epoch": 0.57, + "learning_rate": 0.0003501705115346038, + "loss": 1.3268, + "theoretical_loss": 3.478183005310214, + "tokens_seen": 1870528512 + }, + { + "epoch": 0.57, + "learning_rate": 0.0003501544633901705, + "loss": 1.2435, + "theoretical_loss": 3.478172795008369, + "tokens_seen": 1870594048 + }, + { + "epoch": 0.57, + "learning_rate": 0.00035013841524573726, + "loss": 1.3441, + "theoretical_loss": 3.478162585164391, + "tokens_seen": 1870659584 + }, + { + "epoch": 0.57, + "learning_rate": 0.00035012236710130397, + "loss": 1.2523, + "theoretical_loss": 3.478152375778243, + "tokens_seen": 1870725120 + }, + { + "epoch": 0.57, + "learning_rate": 0.00035010631895687067, + "loss": 1.3139, + "theoretical_loss": 3.4781421668498886, + "tokens_seen": 1870790656 + }, + { + "epoch": 0.57, + "learning_rate": 0.00035009027081243737, + "loss": 1.4147, + "theoretical_loss": 3.4781319583792913, + "tokens_seen": 1870856192 + }, + { + "epoch": 0.57, + "learning_rate": 0.00035007422266800407, + "loss": 1.2522, + "theoretical_loss": 3.478121750366415, + "tokens_seen": 1870921728 + }, + { + "epoch": 0.57, + "learning_rate": 0.00035005817452357077, + "loss": 1.493, + "theoretical_loss": 3.478111542811222, + "tokens_seen": 1870987264 + }, + { + "epoch": 0.57, + "objective/train/docs_used": 682627, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.3513152599334717, + "objective/train/theoretical_loss": 3.4781013357136774, + "objective/train/tokens_used": 1891512800, + "theoretical_loss": 3.4781013357136774, + "tokens_seen": 1871052800 + }, + { + "epoch": 0.57, + "learning_rate": 0.00035004212637913747, + "loss": 1.3178, + "theoretical_loss": 3.4781013357136774, + "tokens_seen": 1871052800 + }, + { + "epoch": 0.57, + "learning_rate": 0.00035002607823470417, + "loss": 1.3713, + "theoretical_loss": 3.4780911290737433, + "tokens_seen": 1871118336 + }, + { + "epoch": 0.57, + "learning_rate": 0.00035001003009027087, + "loss": 1.4849, + "theoretical_loss": 3.478080922891383, + "tokens_seen": 1871183872 + }, + { + "epoch": 0.57, + "learning_rate": 0.00034999398194583757, + "loss": 1.3639, + "theoretical_loss": 3.4780707171665615, + "tokens_seen": 1871249408 + }, + { + "epoch": 0.57, + "learning_rate": 0.00034997793380140427, + "loss": 1.4093, + "theoretical_loss": 3.478060511899241, + "tokens_seen": 1871314944 + }, + { + "epoch": 0.57, + "learning_rate": 0.00034996188565697097, + "loss": 1.4505, + "theoretical_loss": 3.478050307089385, + "tokens_seen": 1871380480 + }, + { + "epoch": 0.57, + "learning_rate": 0.0003499458375125376, + "loss": 1.3633, + "theoretical_loss": 3.4780401027369576, + "tokens_seen": 1871446016 + }, + { + "epoch": 0.57, + "learning_rate": 0.0003499297893681043, + "loss": 1.3199, + "theoretical_loss": 3.478029898841922, + "tokens_seen": 1871511552 + }, + { + "epoch": 0.57, + "learning_rate": 0.000349913741223671, + "loss": 1.3177, + "theoretical_loss": 3.4780196954042415, + "tokens_seen": 1871577088 + }, + { + "epoch": 0.57, + "learning_rate": 0.0003498976930792377, + "loss": 1.2605, + "theoretical_loss": 3.4780094924238796, + "tokens_seen": 1871642624 + }, + { + "epoch": 0.57, + "learning_rate": 0.0003498816449348044, + "loss": 1.3714, + "theoretical_loss": 3.4779992899008, + "tokens_seen": 1871708160 + }, + { + "epoch": 0.57, + "learning_rate": 0.0003498655967903711, + "loss": 1.3677, + "theoretical_loss": 3.477989087834966, + "tokens_seen": 1871773696 + }, + { + "epoch": 0.57, + "learning_rate": 0.0003498495486459378, + "loss": 1.4031, + "theoretical_loss": 3.477978886226341, + "tokens_seen": 1871839232 + }, + { + "epoch": 0.57, + "learning_rate": 0.0003498335005015045, + "loss": 1.4619, + "theoretical_loss": 3.477968685074889, + "tokens_seen": 1871904768 + }, + { + "epoch": 0.57, + "learning_rate": 0.0003498174523570712, + "loss": 1.3978, + "theoretical_loss": 3.477958484380573, + "tokens_seen": 1871970304 + }, + { + "epoch": 0.57, + "learning_rate": 0.0003498014042126379, + "loss": 1.323, + "theoretical_loss": 3.477948284143357, + "tokens_seen": 1872035840 + }, + { + "epoch": 0.57, + "learning_rate": 0.0003497853560682046, + "loss": 1.4518, + "theoretical_loss": 3.477938084363204, + "tokens_seen": 1872101376 + }, + { + "epoch": 0.57, + "learning_rate": 0.0003497693079237713, + "loss": 1.3273, + "theoretical_loss": 3.477927885040078, + "tokens_seen": 1872166912 + }, + { + "epoch": 0.57, + "learning_rate": 0.000349753259779338, + "loss": 1.4119, + "theoretical_loss": 3.477917686173942, + "tokens_seen": 1872232448 + }, + { + "epoch": 0.57, + "learning_rate": 0.0003497372116349047, + "loss": 1.3906, + "theoretical_loss": 3.4779074877647593, + "tokens_seen": 1872297984 + }, + { + "epoch": 0.57, + "learning_rate": 0.0003497211634904714, + "loss": 1.4292, + "theoretical_loss": 3.4778972898124945, + "tokens_seen": 1872363520 + }, + { + "epoch": 0.57, + "learning_rate": 0.00034970511534603813, + "loss": 1.3561, + "theoretical_loss": 3.477887092317111, + "tokens_seen": 1872429056 + }, + { + "epoch": 0.57, + "learning_rate": 0.00034968906720160483, + "loss": 1.3049, + "theoretical_loss": 3.4778768952785715, + "tokens_seen": 1872494592 + }, + { + "epoch": 0.57, + "learning_rate": 0.00034967301905717153, + "loss": 1.3557, + "theoretical_loss": 3.4778666986968396, + "tokens_seen": 1872560128 + }, + { + "epoch": 0.57, + "learning_rate": 0.00034965697091273823, + "loss": 1.3363, + "theoretical_loss": 3.4778565025718797, + "tokens_seen": 1872625664 + }, + { + "epoch": 0.57, + "learning_rate": 0.00034964092276830493, + "loss": 1.3826, + "theoretical_loss": 3.477846306903655, + "tokens_seen": 1872691200 + }, + { + "epoch": 0.57, + "learning_rate": 0.00034962487462387163, + "loss": 1.304, + "theoretical_loss": 3.4778361116921284, + "tokens_seen": 1872756736 + }, + { + "epoch": 0.57, + "learning_rate": 0.00034960882647943833, + "loss": 1.3283, + "theoretical_loss": 3.477825916937264, + "tokens_seen": 1872822272 + }, + { + "epoch": 0.57, + "learning_rate": 0.00034959277833500503, + "loss": 1.4191, + "theoretical_loss": 3.4778157226390256, + "tokens_seen": 1872887808 + }, + { + "epoch": 0.57, + "learning_rate": 0.00034957673019057173, + "loss": 1.2748, + "theoretical_loss": 3.4778055287973766, + "tokens_seen": 1872953344 + }, + { + "epoch": 0.57, + "learning_rate": 0.00034956068204613843, + "loss": 1.382, + "theoretical_loss": 3.4777953354122806, + "tokens_seen": 1873018880 + }, + { + "epoch": 0.57, + "learning_rate": 0.00034954463390170513, + "loss": 1.3904, + "theoretical_loss": 3.4777851424837007, + "tokens_seen": 1873084416 + }, + { + "epoch": 0.57, + "learning_rate": 0.00034952858575727183, + "loss": 1.3218, + "theoretical_loss": 3.477774950011601, + "tokens_seen": 1873149952 + }, + { + "epoch": 0.57, + "learning_rate": 0.00034951253761283853, + "loss": 1.4486, + "theoretical_loss": 3.4777647579959448, + "tokens_seen": 1873215488 + }, + { + "epoch": 0.57, + "learning_rate": 0.00034949648946840524, + "loss": 1.3811, + "theoretical_loss": 3.477754566436696, + "tokens_seen": 1873281024 + }, + { + "epoch": 0.57, + "learning_rate": 0.00034948044132397194, + "loss": 1.3707, + "theoretical_loss": 3.4777443753338178, + "tokens_seen": 1873346560 + }, + { + "epoch": 0.57, + "learning_rate": 0.00034946439317953864, + "loss": 1.2802, + "theoretical_loss": 3.4777341846872742, + "tokens_seen": 1873412096 + }, + { + "epoch": 0.57, + "learning_rate": 0.00034944834503510534, + "loss": 1.33, + "theoretical_loss": 3.477723994497029, + "tokens_seen": 1873477632 + }, + { + "epoch": 0.57, + "learning_rate": 0.00034943229689067204, + "loss": 1.3852, + "theoretical_loss": 3.477713804763045, + "tokens_seen": 1873543168 + }, + { + "epoch": 0.57, + "learning_rate": 0.00034941624874623874, + "loss": 1.4491, + "theoretical_loss": 3.4777036154852863, + "tokens_seen": 1873608704 + }, + { + "epoch": 0.57, + "learning_rate": 0.00034940020060180544, + "loss": 1.2975, + "theoretical_loss": 3.4776934266637167, + "tokens_seen": 1873674240 + }, + { + "epoch": 0.57, + "learning_rate": 0.00034938415245737214, + "loss": 1.3091, + "theoretical_loss": 3.477683238298299, + "tokens_seen": 1873739776 + }, + { + "epoch": 0.57, + "learning_rate": 0.00034936810431293884, + "loss": 1.303, + "theoretical_loss": 3.477673050388998, + "tokens_seen": 1873805312 + }, + { + "epoch": 0.57, + "learning_rate": 0.00034935205616850554, + "loss": 1.453, + "theoretical_loss": 3.4776628629357766, + "tokens_seen": 1873870848 + }, + { + "epoch": 0.57, + "learning_rate": 0.00034933600802407224, + "loss": 1.2876, + "theoretical_loss": 3.4776526759385984, + "tokens_seen": 1873936384 + }, + { + "epoch": 0.57, + "learning_rate": 0.00034931995987963894, + "loss": 1.3555, + "theoretical_loss": 3.477642489397428, + "tokens_seen": 1874001920 + }, + { + "epoch": 0.57, + "learning_rate": 0.00034930391173520564, + "loss": 1.2558, + "theoretical_loss": 3.477632303312227, + "tokens_seen": 1874067456 + }, + { + "epoch": 0.57, + "learning_rate": 0.00034928786359077234, + "loss": 1.3734, + "theoretical_loss": 3.4776221176829614, + "tokens_seen": 1874132992 + }, + { + "epoch": 0.57, + "learning_rate": 0.00034927181544633904, + "loss": 1.3268, + "theoretical_loss": 3.477611932509593, + "tokens_seen": 1874198528 + }, + { + "epoch": 0.57, + "learning_rate": 0.00034925576730190575, + "loss": 1.444, + "theoretical_loss": 3.4776017477920864, + "tokens_seen": 1874264064 + }, + { + "epoch": 0.57, + "objective/train/docs_used": 683764, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.4641492366790771, + "objective/train/theoretical_loss": 3.477591563530405, + "objective/train/tokens_used": 1894789600, + "theoretical_loss": 3.477591563530405, + "tokens_seen": 1874329600 + }, + { + "epoch": 0.57, + "learning_rate": 0.00034923971915747245, + "loss": 1.4265, + "theoretical_loss": 3.477591563530405, + "tokens_seen": 1874329600 + }, + { + "epoch": 0.57, + "learning_rate": 0.00034922367101303915, + "loss": 1.3094, + "theoretical_loss": 3.477581379724513, + "tokens_seen": 1874395136 + }, + { + "epoch": 0.57, + "learning_rate": 0.00034920762286860585, + "loss": 1.4762, + "theoretical_loss": 3.477571196374373, + "tokens_seen": 1874460672 + }, + { + "epoch": 0.57, + "learning_rate": 0.00034919157472417255, + "loss": 1.4586, + "theoretical_loss": 3.4775610134799497, + "tokens_seen": 1874526208 + }, + { + "epoch": 0.57, + "learning_rate": 0.0003491755265797392, + "loss": 1.4294, + "theoretical_loss": 3.477550831041206, + "tokens_seen": 1874591744 + }, + { + "epoch": 0.57, + "learning_rate": 0.0003491594784353059, + "loss": 1.3112, + "theoretical_loss": 3.4775406490581062, + "tokens_seen": 1874657280 + }, + { + "epoch": 0.57, + "learning_rate": 0.0003491434302908726, + "loss": 1.4629, + "theoretical_loss": 3.4775304675306136, + "tokens_seen": 1874722816 + }, + { + "epoch": 0.57, + "learning_rate": 0.0003491273821464393, + "loss": 1.4164, + "theoretical_loss": 3.4775202864586916, + "tokens_seen": 1874788352 + }, + { + "epoch": 0.57, + "learning_rate": 0.000349111334002006, + "loss": 1.3338, + "theoretical_loss": 3.477510105842305, + "tokens_seen": 1874853888 + }, + { + "epoch": 0.57, + "learning_rate": 0.0003490952858575727, + "loss": 1.3227, + "theoretical_loss": 3.4774999256814163, + "tokens_seen": 1874919424 + }, + { + "epoch": 0.57, + "learning_rate": 0.0003490792377131394, + "loss": 1.2938, + "theoretical_loss": 3.4774897459759897, + "tokens_seen": 1874984960 + }, + { + "epoch": 0.57, + "learning_rate": 0.0003490631895687061, + "loss": 1.3868, + "theoretical_loss": 3.477479566725989, + "tokens_seen": 1875050496 + }, + { + "epoch": 0.57, + "learning_rate": 0.0003490471414242728, + "loss": 1.4008, + "theoretical_loss": 3.477469387931378, + "tokens_seen": 1875116032 + }, + { + "epoch": 0.57, + "learning_rate": 0.0003490310932798395, + "loss": 1.2785, + "theoretical_loss": 3.4774592095921197, + "tokens_seen": 1875181568 + }, + { + "epoch": 0.57, + "learning_rate": 0.0003490150451354062, + "loss": 1.4059, + "theoretical_loss": 3.4774490317081783, + "tokens_seen": 1875247104 + }, + { + "epoch": 0.57, + "learning_rate": 0.0003489989969909729, + "loss": 1.2415, + "theoretical_loss": 3.477438854279518, + "tokens_seen": 1875312640 + }, + { + "epoch": 0.57, + "learning_rate": 0.0003489829488465396, + "loss": 1.307, + "theoretical_loss": 3.4774286773061016, + "tokens_seen": 1875378176 + }, + { + "epoch": 0.57, + "learning_rate": 0.0003489669007021063, + "loss": 1.4511, + "theoretical_loss": 3.477418500787894, + "tokens_seen": 1875443712 + }, + { + "epoch": 0.57, + "learning_rate": 0.000348950852557673, + "loss": 1.3119, + "theoretical_loss": 3.4774083247248573, + "tokens_seen": 1875509248 + }, + { + "epoch": 0.57, + "learning_rate": 0.00034893480441323976, + "loss": 1.5403, + "theoretical_loss": 3.477398149116957, + "tokens_seen": 1875574784 + }, + { + "epoch": 0.57, + "learning_rate": 0.00034891875626880646, + "loss": 1.396, + "theoretical_loss": 3.4773879739641553, + "tokens_seen": 1875640320 + }, + { + "epoch": 0.57, + "learning_rate": 0.00034890270812437316, + "loss": 1.408, + "theoretical_loss": 3.477377799266417, + "tokens_seen": 1875705856 + }, + { + "epoch": 0.57, + "learning_rate": 0.00034888665997993986, + "loss": 1.3827, + "theoretical_loss": 3.4773676250237058, + "tokens_seen": 1875771392 + }, + { + "epoch": 0.57, + "learning_rate": 0.00034887061183550656, + "loss": 1.3352, + "theoretical_loss": 3.477357451235984, + "tokens_seen": 1875836928 + }, + { + "epoch": 0.57, + "learning_rate": 0.00034885456369107326, + "loss": 1.3733, + "theoretical_loss": 3.4773472779032177, + "tokens_seen": 1875902464 + }, + { + "epoch": 0.57, + "learning_rate": 0.00034883851554663996, + "loss": 1.5263, + "theoretical_loss": 3.477337105025369, + "tokens_seen": 1875968000 + }, + { + "epoch": 0.57, + "learning_rate": 0.00034882246740220666, + "loss": 1.4277, + "theoretical_loss": 3.477326932602402, + "tokens_seen": 1876033536 + }, + { + "epoch": 0.57, + "learning_rate": 0.00034880641925777336, + "loss": 1.3115, + "theoretical_loss": 3.477316760634281, + "tokens_seen": 1876099072 + }, + { + "epoch": 0.57, + "learning_rate": 0.00034879037111334006, + "loss": 1.4124, + "theoretical_loss": 3.4773065891209693, + "tokens_seen": 1876164608 + }, + { + "epoch": 0.57, + "learning_rate": 0.00034877432296890676, + "loss": 1.2851, + "theoretical_loss": 3.477296418062431, + "tokens_seen": 1876230144 + }, + { + "epoch": 0.57, + "learning_rate": 0.00034875827482447346, + "loss": 1.3766, + "theoretical_loss": 3.477286247458629, + "tokens_seen": 1876295680 + }, + { + "epoch": 0.57, + "learning_rate": 0.00034874222668004017, + "loss": 1.3531, + "theoretical_loss": 3.477276077309528, + "tokens_seen": 1876361216 + }, + { + "epoch": 0.57, + "learning_rate": 0.00034872617853560687, + "loss": 1.3662, + "theoretical_loss": 3.4772659076150916, + "tokens_seen": 1876426752 + }, + { + "epoch": 0.57, + "learning_rate": 0.00034871013039117357, + "loss": 1.3219, + "theoretical_loss": 3.477255738375283, + "tokens_seen": 1876492288 + }, + { + "epoch": 0.57, + "learning_rate": 0.00034869408224674027, + "loss": 1.3198, + "theoretical_loss": 3.4772455695900675, + "tokens_seen": 1876557824 + }, + { + "epoch": 0.57, + "learning_rate": 0.00034867803410230697, + "loss": 1.3441, + "theoretical_loss": 3.4772354012594073, + "tokens_seen": 1876623360 + }, + { + "epoch": 0.57, + "learning_rate": 0.00034866198595787367, + "loss": 1.3325, + "theoretical_loss": 3.4772252333832667, + "tokens_seen": 1876688896 + }, + { + "epoch": 0.57, + "learning_rate": 0.00034864593781344037, + "loss": 1.3631, + "theoretical_loss": 3.4772150659616097, + "tokens_seen": 1876754432 + }, + { + "epoch": 0.57, + "learning_rate": 0.00034862988966900707, + "loss": 1.3325, + "theoretical_loss": 3.4772048989944, + "tokens_seen": 1876819968 + }, + { + "epoch": 0.57, + "learning_rate": 0.00034861384152457377, + "loss": 1.4229, + "theoretical_loss": 3.4771947324816024, + "tokens_seen": 1876885504 + }, + { + "epoch": 0.57, + "learning_rate": 0.00034859779338014047, + "loss": 1.3248, + "theoretical_loss": 3.477184566423179, + "tokens_seen": 1876951040 + }, + { + "epoch": 0.57, + "learning_rate": 0.00034858174523570717, + "loss": 1.2899, + "theoretical_loss": 3.477174400819094, + "tokens_seen": 1877016576 + }, + { + "epoch": 0.57, + "learning_rate": 0.00034856569709127387, + "loss": 1.3736, + "theoretical_loss": 3.4771642356693127, + "tokens_seen": 1877082112 + }, + { + "epoch": 0.57, + "learning_rate": 0.0003485496489468405, + "loss": 1.2826, + "theoretical_loss": 3.477154070973797, + "tokens_seen": 1877147648 + }, + { + "epoch": 0.57, + "learning_rate": 0.0003485336008024072, + "loss": 1.3795, + "theoretical_loss": 3.477143906732512, + "tokens_seen": 1877213184 + }, + { + "epoch": 0.57, + "learning_rate": 0.0003485175526579739, + "loss": 1.4187, + "theoretical_loss": 3.477133742945421, + "tokens_seen": 1877278720 + }, + { + "epoch": 0.57, + "learning_rate": 0.0003485015045135406, + "loss": 1.2944, + "theoretical_loss": 3.4771235796124884, + "tokens_seen": 1877344256 + }, + { + "epoch": 0.57, + "learning_rate": 0.0003484854563691073, + "loss": 1.3801, + "theoretical_loss": 3.4771134167336775, + "tokens_seen": 1877409792 + }, + { + "epoch": 0.57, + "learning_rate": 0.000348469408224674, + "loss": 1.4088, + "theoretical_loss": 3.477103254308952, + "tokens_seen": 1877475328 + }, + { + "epoch": 0.57, + "learning_rate": 0.0003484533600802407, + "loss": 1.2993, + "theoretical_loss": 3.4770930923382766, + "tokens_seen": 1877540864 + }, + { + "epoch": 0.57, + "objective/train/docs_used": 684931, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.3768601417541504, + "objective/train/theoretical_loss": 3.4770829308216147, + "objective/train/tokens_used": 1898066400, + "theoretical_loss": 3.4770829308216147, + "tokens_seen": 1877606400 + }, + { + "epoch": 0.57, + "learning_rate": 0.0003484373119358074, + "loss": 1.4445, + "theoretical_loss": 3.4770829308216147, + "tokens_seen": 1877606400 + }, + { + "epoch": 0.57, + "learning_rate": 0.0003484212637913741, + "loss": 1.3171, + "theoretical_loss": 3.47707276975893, + "tokens_seen": 1877671936 + }, + { + "epoch": 0.57, + "learning_rate": 0.0003484052156469408, + "loss": 1.3157, + "theoretical_loss": 3.4770626091501864, + "tokens_seen": 1877737472 + }, + { + "epoch": 0.57, + "learning_rate": 0.0003483891675025075, + "loss": 1.4614, + "theoretical_loss": 3.477052448995348, + "tokens_seen": 1877803008 + }, + { + "epoch": 0.57, + "learning_rate": 0.0003483731193580742, + "loss": 1.4745, + "theoretical_loss": 3.4770422892943786, + "tokens_seen": 1877868544 + }, + { + "epoch": 0.57, + "learning_rate": 0.0003483570712136409, + "loss": 1.3243, + "theoretical_loss": 3.4770321300472418, + "tokens_seen": 1877934080 + }, + { + "epoch": 0.57, + "learning_rate": 0.00034834102306920763, + "loss": 1.4875, + "theoretical_loss": 3.477021971253902, + "tokens_seen": 1877999616 + }, + { + "epoch": 0.57, + "learning_rate": 0.00034832497492477433, + "loss": 1.4784, + "theoretical_loss": 3.4770118129143226, + "tokens_seen": 1878065152 + }, + { + "epoch": 0.57, + "learning_rate": 0.00034830892678034103, + "loss": 1.3779, + "theoretical_loss": 3.4770016550284684, + "tokens_seen": 1878130688 + }, + { + "epoch": 0.57, + "learning_rate": 0.00034829287863590773, + "loss": 1.4038, + "theoretical_loss": 3.476991497596302, + "tokens_seen": 1878196224 + }, + { + "epoch": 0.57, + "learning_rate": 0.00034827683049147443, + "loss": 1.3835, + "theoretical_loss": 3.476981340617788, + "tokens_seen": 1878261760 + }, + { + "epoch": 0.57, + "learning_rate": 0.00034826078234704113, + "loss": 1.3529, + "theoretical_loss": 3.47697118409289, + "tokens_seen": 1878327296 + }, + { + "epoch": 0.57, + "learning_rate": 0.00034824473420260783, + "loss": 1.4241, + "theoretical_loss": 3.476961028021573, + "tokens_seen": 1878392832 + }, + { + "epoch": 0.57, + "learning_rate": 0.00034822868605817453, + "loss": 1.3776, + "theoretical_loss": 3.4769508724038, + "tokens_seen": 1878458368 + }, + { + "epoch": 0.57, + "learning_rate": 0.00034821263791374123, + "loss": 1.418, + "theoretical_loss": 3.4769407172395344, + "tokens_seen": 1878523904 + }, + { + "epoch": 0.57, + "learning_rate": 0.00034819658976930793, + "loss": 1.4773, + "theoretical_loss": 3.476930562528741, + "tokens_seen": 1878589440 + }, + { + "epoch": 0.57, + "learning_rate": 0.00034818054162487463, + "loss": 1.4518, + "theoretical_loss": 3.4769204082713836, + "tokens_seen": 1878654976 + }, + { + "epoch": 0.57, + "learning_rate": 0.00034816449348044133, + "loss": 1.4431, + "theoretical_loss": 3.476910254467426, + "tokens_seen": 1878720512 + }, + { + "epoch": 0.57, + "learning_rate": 0.00034814844533600803, + "loss": 1.4179, + "theoretical_loss": 3.476900101116832, + "tokens_seen": 1878786048 + }, + { + "epoch": 0.57, + "learning_rate": 0.00034813239719157474, + "loss": 1.3609, + "theoretical_loss": 3.476889948219566, + "tokens_seen": 1878851584 + }, + { + "epoch": 0.57, + "learning_rate": 0.00034811634904714144, + "loss": 1.3164, + "theoretical_loss": 3.4768797957755915, + "tokens_seen": 1878917120 + }, + { + "epoch": 0.57, + "learning_rate": 0.00034810030090270814, + "loss": 1.3289, + "theoretical_loss": 3.4768696437848723, + "tokens_seen": 1878982656 + }, + { + "epoch": 0.57, + "learning_rate": 0.00034808425275827484, + "loss": 1.2559, + "theoretical_loss": 3.4768594922473732, + "tokens_seen": 1879048192 + }, + { + "epoch": 0.57, + "learning_rate": 0.00034806820461384154, + "loss": 1.4302, + "theoretical_loss": 3.4768493411630574, + "tokens_seen": 1879113728 + }, + { + "epoch": 0.57, + "learning_rate": 0.00034805215646940824, + "loss": 1.3647, + "theoretical_loss": 3.476839190531889, + "tokens_seen": 1879179264 + }, + { + "epoch": 0.57, + "learning_rate": 0.00034803610832497494, + "loss": 1.4103, + "theoretical_loss": 3.476829040353832, + "tokens_seen": 1879244800 + }, + { + "epoch": 0.57, + "learning_rate": 0.00034802006018054164, + "loss": 1.3313, + "theoretical_loss": 3.4768188906288504, + "tokens_seen": 1879310336 + }, + { + "epoch": 0.57, + "learning_rate": 0.00034800401203610834, + "loss": 1.3509, + "theoretical_loss": 3.4768087413569084, + "tokens_seen": 1879375872 + }, + { + "epoch": 0.57, + "learning_rate": 0.00034798796389167504, + "loss": 1.351, + "theoretical_loss": 3.4767985925379703, + "tokens_seen": 1879441408 + }, + { + "epoch": 0.57, + "learning_rate": 0.00034797191574724174, + "loss": 1.3372, + "theoretical_loss": 3.4767884441719987, + "tokens_seen": 1879506944 + }, + { + "epoch": 0.57, + "learning_rate": 0.00034795586760280844, + "loss": 1.3427, + "theoretical_loss": 3.4767782962589586, + "tokens_seen": 1879572480 + }, + { + "epoch": 0.57, + "learning_rate": 0.00034793981945837514, + "loss": 1.3467, + "theoretical_loss": 3.476768148798814, + "tokens_seen": 1879638016 + }, + { + "epoch": 0.57, + "learning_rate": 0.00034792377131394184, + "loss": 1.419, + "theoretical_loss": 3.476758001791529, + "tokens_seen": 1879703552 + }, + { + "epoch": 0.57, + "learning_rate": 0.00034790772316950854, + "loss": 1.3576, + "theoretical_loss": 3.4767478552370674, + "tokens_seen": 1879769088 + }, + { + "epoch": 0.57, + "learning_rate": 0.00034789167502507524, + "loss": 1.389, + "theoretical_loss": 3.4767377091353926, + "tokens_seen": 1879834624 + }, + { + "epoch": 0.57, + "learning_rate": 0.00034787562688064195, + "loss": 1.3637, + "theoretical_loss": 3.4767275634864694, + "tokens_seen": 1879900160 + }, + { + "epoch": 0.57, + "learning_rate": 0.00034785957873620865, + "loss": 1.3989, + "theoretical_loss": 3.476717418290262, + "tokens_seen": 1879965696 + }, + { + "epoch": 0.57, + "learning_rate": 0.00034784353059177535, + "loss": 1.3779, + "theoretical_loss": 3.476707273546733, + "tokens_seen": 1880031232 + }, + { + "epoch": 0.57, + "learning_rate": 0.00034782748244734205, + "loss": 1.3379, + "theoretical_loss": 3.4766971292558484, + "tokens_seen": 1880096768 + }, + { + "epoch": 0.57, + "learning_rate": 0.00034781143430290875, + "loss": 1.3949, + "theoretical_loss": 3.476686985417571, + "tokens_seen": 1880162304 + }, + { + "epoch": 0.57, + "learning_rate": 0.00034779538615847545, + "loss": 1.3104, + "theoretical_loss": 3.476676842031865, + "tokens_seen": 1880227840 + }, + { + "epoch": 0.57, + "learning_rate": 0.0003477793380140421, + "loss": 1.3689, + "theoretical_loss": 3.4766666990986947, + "tokens_seen": 1880293376 + }, + { + "epoch": 0.57, + "learning_rate": 0.0003477632898696088, + "loss": 1.2845, + "theoretical_loss": 3.476656556618024, + "tokens_seen": 1880358912 + }, + { + "epoch": 0.57, + "learning_rate": 0.00034774724172517555, + "loss": 1.2982, + "theoretical_loss": 3.4766464145898164, + "tokens_seen": 1880424448 + }, + { + "epoch": 0.57, + "learning_rate": 0.00034773119358074225, + "loss": 1.3326, + "theoretical_loss": 3.4766362730140368, + "tokens_seen": 1880489984 + }, + { + "epoch": 0.57, + "learning_rate": 0.00034771514543630895, + "loss": 1.4343, + "theoretical_loss": 3.476626131890649, + "tokens_seen": 1880555520 + }, + { + "epoch": 0.57, + "learning_rate": 0.00034769909729187565, + "loss": 1.318, + "theoretical_loss": 3.476615991219617, + "tokens_seen": 1880621056 + }, + { + "epoch": 0.57, + "learning_rate": 0.00034768304914744235, + "loss": 1.3708, + "theoretical_loss": 3.4766058510009046, + "tokens_seen": 1880686592 + }, + { + "epoch": 0.57, + "learning_rate": 0.00034766700100300905, + "loss": 1.3264, + "theoretical_loss": 3.476595711234476, + "tokens_seen": 1880752128 + }, + { + "epoch": 0.57, + "learning_rate": 0.00034765095285857575, + "loss": 1.2724, + "theoretical_loss": 3.476585571920295, + "tokens_seen": 1880817664 + }, + { + "epoch": 0.57, + "objective/train/docs_used": 686045, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.2848793268203735, + "objective/train/theoretical_loss": 3.476575433058327, + "objective/train/tokens_used": 1901343200, + "theoretical_loss": 3.476575433058327, + "tokens_seen": 1880883200 + }, + { + "epoch": 0.57, + "learning_rate": 0.00034763490471414246, + "loss": 1.3524, + "theoretical_loss": 3.476575433058327, + "tokens_seen": 1880883200 + }, + { + "epoch": 0.57, + "learning_rate": 0.00034761885656970916, + "loss": 1.4421, + "theoretical_loss": 3.476565294648535, + "tokens_seen": 1880948736 + }, + { + "epoch": 0.57, + "learning_rate": 0.00034760280842527586, + "loss": 1.3353, + "theoretical_loss": 3.4765551566908823, + "tokens_seen": 1881014272 + }, + { + "epoch": 0.57, + "learning_rate": 0.00034758676028084256, + "loss": 1.4784, + "theoretical_loss": 3.476545019185334, + "tokens_seen": 1881079808 + }, + { + "epoch": 0.57, + "learning_rate": 0.00034757071213640926, + "loss": 1.2889, + "theoretical_loss": 3.4765348821318547, + "tokens_seen": 1881145344 + }, + { + "epoch": 0.57, + "learning_rate": 0.00034755466399197596, + "loss": 1.3771, + "theoretical_loss": 3.4765247455304076, + "tokens_seen": 1881210880 + }, + { + "epoch": 0.57, + "learning_rate": 0.00034753861584754266, + "loss": 1.3574, + "theoretical_loss": 3.476514609380957, + "tokens_seen": 1881276416 + }, + { + "epoch": 0.57, + "learning_rate": 0.00034752256770310936, + "loss": 1.2318, + "theoretical_loss": 3.476504473683467, + "tokens_seen": 1881341952 + }, + { + "epoch": 0.57, + "learning_rate": 0.00034750651955867606, + "loss": 1.2743, + "theoretical_loss": 3.4764943384379015, + "tokens_seen": 1881407488 + }, + { + "epoch": 0.57, + "learning_rate": 0.00034749047141424276, + "loss": 1.2517, + "theoretical_loss": 3.476484203644225, + "tokens_seen": 1881473024 + }, + { + "epoch": 0.57, + "learning_rate": 0.00034747442326980946, + "loss": 1.3475, + "theoretical_loss": 3.4764740693024017, + "tokens_seen": 1881538560 + }, + { + "epoch": 0.57, + "learning_rate": 0.00034745837512537616, + "loss": 1.3539, + "theoretical_loss": 3.4764639354123954, + "tokens_seen": 1881604096 + }, + { + "epoch": 0.57, + "learning_rate": 0.00034744232698094286, + "loss": 1.3165, + "theoretical_loss": 3.47645380197417, + "tokens_seen": 1881669632 + }, + { + "epoch": 0.57, + "learning_rate": 0.00034742627883650956, + "loss": 1.325, + "theoretical_loss": 3.4764436689876903, + "tokens_seen": 1881735168 + }, + { + "epoch": 0.57, + "learning_rate": 0.00034741023069207626, + "loss": 1.3016, + "theoretical_loss": 3.47643353645292, + "tokens_seen": 1881800704 + }, + { + "epoch": 0.57, + "learning_rate": 0.00034739418254764296, + "loss": 1.3014, + "theoretical_loss": 3.476423404369823, + "tokens_seen": 1881866240 + }, + { + "epoch": 0.57, + "learning_rate": 0.00034737813440320967, + "loss": 1.3456, + "theoretical_loss": 3.4764132727383643, + "tokens_seen": 1881931776 + }, + { + "epoch": 0.57, + "learning_rate": 0.00034736208625877637, + "loss": 1.4962, + "theoretical_loss": 3.4764031415585066, + "tokens_seen": 1881997312 + }, + { + "epoch": 0.57, + "learning_rate": 0.00034734603811434307, + "loss": 1.3128, + "theoretical_loss": 3.4763930108302157, + "tokens_seen": 1882062848 + }, + { + "epoch": 0.57, + "learning_rate": 0.00034732998996990977, + "loss": 1.4668, + "theoretical_loss": 3.4763828805534547, + "tokens_seen": 1882128384 + }, + { + "epoch": 0.57, + "learning_rate": 0.00034731394182547647, + "loss": 1.2801, + "theoretical_loss": 3.4763727507281876, + "tokens_seen": 1882193920 + }, + { + "epoch": 0.57, + "learning_rate": 0.00034729789368104317, + "loss": 1.3105, + "theoretical_loss": 3.4763626213543795, + "tokens_seen": 1882259456 + }, + { + "epoch": 0.57, + "learning_rate": 0.00034728184553660987, + "loss": 1.318, + "theoretical_loss": 3.476352492431994, + "tokens_seen": 1882324992 + }, + { + "epoch": 0.57, + "learning_rate": 0.00034726579739217657, + "loss": 1.3059, + "theoretical_loss": 3.4763423639609954, + "tokens_seen": 1882390528 + }, + { + "epoch": 0.57, + "learning_rate": 0.00034724974924774327, + "loss": 1.3997, + "theoretical_loss": 3.4763322359413475, + "tokens_seen": 1882456064 + }, + { + "epoch": 0.57, + "learning_rate": 0.00034723370110330997, + "loss": 1.4203, + "theoretical_loss": 3.4763221083730147, + "tokens_seen": 1882521600 + }, + { + "epoch": 0.57, + "learning_rate": 0.00034721765295887667, + "loss": 1.3359, + "theoretical_loss": 3.4763119812559613, + "tokens_seen": 1882587136 + }, + { + "epoch": 0.57, + "learning_rate": 0.00034720160481444337, + "loss": 1.3981, + "theoretical_loss": 3.4763018545901514, + "tokens_seen": 1882652672 + }, + { + "epoch": 0.57, + "learning_rate": 0.00034718555667001007, + "loss": 1.4468, + "theoretical_loss": 3.476291728375549, + "tokens_seen": 1882718208 + }, + { + "epoch": 0.57, + "learning_rate": 0.0003471695085255768, + "loss": 1.5006, + "theoretical_loss": 3.476281602612119, + "tokens_seen": 1882783744 + }, + { + "epoch": 0.57, + "learning_rate": 0.0003471534603811435, + "loss": 1.3888, + "theoretical_loss": 3.476271477299825, + "tokens_seen": 1882849280 + }, + { + "epoch": 0.57, + "learning_rate": 0.0003471374122367101, + "loss": 1.2875, + "theoretical_loss": 3.476261352438631, + "tokens_seen": 1882914816 + }, + { + "epoch": 0.57, + "learning_rate": 0.0003471213640922768, + "loss": 1.2834, + "theoretical_loss": 3.4762512280285014, + "tokens_seen": 1882980352 + }, + { + "epoch": 0.57, + "learning_rate": 0.0003471053159478435, + "loss": 1.3731, + "theoretical_loss": 3.476241104069401, + "tokens_seen": 1883045888 + }, + { + "epoch": 0.57, + "learning_rate": 0.0003470892678034102, + "loss": 1.3755, + "theoretical_loss": 3.4762309805612928, + "tokens_seen": 1883111424 + }, + { + "epoch": 0.57, + "learning_rate": 0.0003470732196589769, + "loss": 1.2293, + "theoretical_loss": 3.4762208575041424, + "tokens_seen": 1883176960 + }, + { + "epoch": 0.57, + "learning_rate": 0.0003470571715145436, + "loss": 1.3683, + "theoretical_loss": 3.476210734897913, + "tokens_seen": 1883242496 + }, + { + "epoch": 0.57, + "learning_rate": 0.0003470411233701103, + "loss": 1.2385, + "theoretical_loss": 3.476200612742569, + "tokens_seen": 1883308032 + }, + { + "epoch": 0.57, + "learning_rate": 0.000347025075225677, + "loss": 1.3983, + "theoretical_loss": 3.476190491038075, + "tokens_seen": 1883373568 + }, + { + "epoch": 0.57, + "learning_rate": 0.0003470090270812437, + "loss": 1.485, + "theoretical_loss": 3.476180369784395, + "tokens_seen": 1883439104 + }, + { + "epoch": 0.57, + "learning_rate": 0.0003469929789368104, + "loss": 1.3033, + "theoretical_loss": 3.476170248981493, + "tokens_seen": 1883504640 + }, + { + "epoch": 0.57, + "learning_rate": 0.0003469769307923771, + "loss": 1.3416, + "theoretical_loss": 3.476160128629334, + "tokens_seen": 1883570176 + }, + { + "epoch": 0.57, + "learning_rate": 0.00034696088264794383, + "loss": 1.3447, + "theoretical_loss": 3.476150008727881, + "tokens_seen": 1883635712 + }, + { + "epoch": 0.57, + "learning_rate": 0.00034694483450351053, + "loss": 1.3791, + "theoretical_loss": 3.4761398892771, + "tokens_seen": 1883701248 + }, + { + "epoch": 0.57, + "learning_rate": 0.00034692878635907723, + "loss": 1.3777, + "theoretical_loss": 3.4761297702769536, + "tokens_seen": 1883766784 + }, + { + "epoch": 0.57, + "learning_rate": 0.00034691273821464393, + "loss": 1.3707, + "theoretical_loss": 3.476119651727407, + "tokens_seen": 1883832320 + }, + { + "epoch": 0.57, + "learning_rate": 0.00034689669007021063, + "loss": 1.402, + "theoretical_loss": 3.476109533628424, + "tokens_seen": 1883897856 + }, + { + "epoch": 0.57, + "learning_rate": 0.00034688064192577733, + "loss": 1.3581, + "theoretical_loss": 3.4760994159799687, + "tokens_seen": 1883963392 + }, + { + "epoch": 0.57, + "learning_rate": 0.00034686459378134403, + "loss": 1.4172, + "theoretical_loss": 3.4760892987820062, + "tokens_seen": 1884028928 + }, + { + "epoch": 0.57, + "learning_rate": 0.00034684854563691073, + "loss": 1.3499, + "theoretical_loss": 3.4760791820345, + "tokens_seen": 1884094464 + }, + { + "epoch": 0.57, + "objective/train/docs_used": 687181, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.3734180927276611, + "objective/train/theoretical_loss": 3.4760690657374145, + "objective/train/tokens_used": 1904620000, + "theoretical_loss": 3.4760690657374145, + "tokens_seen": 1884160000 + }, + { + "epoch": 0.57, + "learning_rate": 0.00034683249749247743, + "loss": 1.441, + "theoretical_loss": 3.4760690657374145, + "tokens_seen": 1884160000 + }, + { + "epoch": 0.57, + "learning_rate": 0.00034681644934804413, + "loss": 1.3207, + "theoretical_loss": 3.476058949890714, + "tokens_seen": 1884225536 + }, + { + "epoch": 0.57, + "learning_rate": 0.00034680040120361083, + "loss": 1.4576, + "theoretical_loss": 3.476048834494363, + "tokens_seen": 1884291072 + }, + { + "epoch": 0.57, + "learning_rate": 0.00034678435305917753, + "loss": 1.4026, + "theoretical_loss": 3.4760387195483258, + "tokens_seen": 1884356608 + }, + { + "epoch": 0.57, + "learning_rate": 0.00034676830491474424, + "loss": 1.3429, + "theoretical_loss": 3.4760286050525666, + "tokens_seen": 1884422144 + }, + { + "epoch": 0.57, + "learning_rate": 0.00034675225677031094, + "loss": 1.3602, + "theoretical_loss": 3.47601849100705, + "tokens_seen": 1884487680 + }, + { + "epoch": 0.57, + "learning_rate": 0.00034673620862587764, + "loss": 1.3, + "theoretical_loss": 3.4760083774117394, + "tokens_seen": 1884553216 + }, + { + "epoch": 0.57, + "learning_rate": 0.00034672016048144434, + "loss": 1.4229, + "theoretical_loss": 3.4759982642666003, + "tokens_seen": 1884618752 + }, + { + "epoch": 0.57, + "learning_rate": 0.00034670411233701104, + "loss": 1.4452, + "theoretical_loss": 3.475988151571596, + "tokens_seen": 1884684288 + }, + { + "epoch": 0.57, + "learning_rate": 0.00034668806419257774, + "loss": 1.3992, + "theoretical_loss": 3.475978039326691, + "tokens_seen": 1884749824 + }, + { + "epoch": 0.57, + "learning_rate": 0.00034667201604814444, + "loss": 1.4708, + "theoretical_loss": 3.4759679275318502, + "tokens_seen": 1884815360 + }, + { + "epoch": 0.57, + "learning_rate": 0.00034665596790371114, + "loss": 1.4328, + "theoretical_loss": 3.4759578161870373, + "tokens_seen": 1884880896 + }, + { + "epoch": 0.57, + "learning_rate": 0.00034663991975927784, + "loss": 1.3026, + "theoretical_loss": 3.475947705292217, + "tokens_seen": 1884946432 + }, + { + "epoch": 0.57, + "learning_rate": 0.00034662387161484454, + "loss": 1.3572, + "theoretical_loss": 3.4759375948473537, + "tokens_seen": 1885011968 + }, + { + "epoch": 0.57, + "learning_rate": 0.00034660782347041124, + "loss": 1.5017, + "theoretical_loss": 3.475927484852411, + "tokens_seen": 1885077504 + }, + { + "epoch": 0.57, + "learning_rate": 0.00034659177532597794, + "loss": 1.4279, + "theoretical_loss": 3.475917375307354, + "tokens_seen": 1885143040 + }, + { + "epoch": 0.57, + "learning_rate": 0.00034657572718154464, + "loss": 1.4133, + "theoretical_loss": 3.475907266212147, + "tokens_seen": 1885208576 + }, + { + "epoch": 0.57, + "learning_rate": 0.00034655967903711134, + "loss": 1.3306, + "theoretical_loss": 3.475897157566754, + "tokens_seen": 1885274112 + }, + { + "epoch": 0.57, + "learning_rate": 0.0003465436308926781, + "loss": 1.4069, + "theoretical_loss": 3.4758870493711393, + "tokens_seen": 1885339648 + }, + { + "epoch": 0.57, + "learning_rate": 0.0003465275827482448, + "loss": 1.2748, + "theoretical_loss": 3.475876941625268, + "tokens_seen": 1885405184 + }, + { + "epoch": 0.57, + "learning_rate": 0.00034651153460381145, + "loss": 1.2797, + "theoretical_loss": 3.4758668343291035, + "tokens_seen": 1885470720 + }, + { + "epoch": 0.57, + "learning_rate": 0.00034649548645937815, + "loss": 1.3359, + "theoretical_loss": 3.4758567274826104, + "tokens_seen": 1885536256 + }, + { + "epoch": 0.57, + "learning_rate": 0.00034647943831494485, + "loss": 1.3818, + "theoretical_loss": 3.475846621085754, + "tokens_seen": 1885601792 + }, + { + "epoch": 0.57, + "learning_rate": 0.00034646339017051155, + "loss": 1.335, + "theoretical_loss": 3.4758365151384973, + "tokens_seen": 1885667328 + }, + { + "epoch": 0.57, + "learning_rate": 0.00034644734202607825, + "loss": 1.3201, + "theoretical_loss": 3.4758264096408054, + "tokens_seen": 1885732864 + }, + { + "epoch": 0.57, + "learning_rate": 0.00034643129388164495, + "loss": 1.3979, + "theoretical_loss": 3.4758163045926427, + "tokens_seen": 1885798400 + }, + { + "epoch": 0.57, + "learning_rate": 0.00034641524573721165, + "loss": 1.3188, + "theoretical_loss": 3.475806199993973, + "tokens_seen": 1885863936 + }, + { + "epoch": 0.57, + "learning_rate": 0.00034639919759277835, + "loss": 1.4284, + "theoretical_loss": 3.475796095844762, + "tokens_seen": 1885929472 + }, + { + "epoch": 0.57, + "learning_rate": 0.00034638314944834505, + "loss": 1.4198, + "theoretical_loss": 3.4757859921449725, + "tokens_seen": 1885995008 + }, + { + "epoch": 0.57, + "learning_rate": 0.00034636710130391175, + "loss": 1.3878, + "theoretical_loss": 3.47577588889457, + "tokens_seen": 1886060544 + }, + { + "epoch": 0.57, + "learning_rate": 0.00034635105315947845, + "loss": 1.4087, + "theoretical_loss": 3.4757657860935183, + "tokens_seen": 1886126080 + }, + { + "epoch": 0.57, + "learning_rate": 0.00034633500501504515, + "loss": 1.4124, + "theoretical_loss": 3.475755683741782, + "tokens_seen": 1886191616 + }, + { + "epoch": 0.57, + "learning_rate": 0.00034631895687061185, + "loss": 1.3801, + "theoretical_loss": 3.4757455818393255, + "tokens_seen": 1886257152 + }, + { + "epoch": 0.57, + "learning_rate": 0.00034630290872617855, + "loss": 1.3265, + "theoretical_loss": 3.4757354803861134, + "tokens_seen": 1886322688 + }, + { + "epoch": 0.57, + "learning_rate": 0.00034628686058174525, + "loss": 1.3354, + "theoretical_loss": 3.4757253793821095, + "tokens_seen": 1886388224 + }, + { + "epoch": 0.57, + "learning_rate": 0.00034627081243731195, + "loss": 1.3242, + "theoretical_loss": 3.475715278827279, + "tokens_seen": 1886453760 + }, + { + "epoch": 0.57, + "learning_rate": 0.00034625476429287866, + "loss": 1.3277, + "theoretical_loss": 3.4757051787215865, + "tokens_seen": 1886519296 + }, + { + "epoch": 0.57, + "learning_rate": 0.00034623871614844536, + "loss": 1.4355, + "theoretical_loss": 3.4756950790649954, + "tokens_seen": 1886584832 + }, + { + "epoch": 0.57, + "learning_rate": 0.00034622266800401206, + "loss": 1.3102, + "theoretical_loss": 3.4756849798574705, + "tokens_seen": 1886650368 + }, + { + "epoch": 0.57, + "learning_rate": 0.00034620661985957876, + "loss": 1.3653, + "theoretical_loss": 3.4756748810989766, + "tokens_seen": 1886715904 + }, + { + "epoch": 0.57, + "learning_rate": 0.00034619057171514546, + "loss": 1.395, + "theoretical_loss": 3.4756647827894778, + "tokens_seen": 1886781440 + }, + { + "epoch": 0.57, + "learning_rate": 0.00034617452357071216, + "loss": 1.2759, + "theoretical_loss": 3.475654684928939, + "tokens_seen": 1886846976 + }, + { + "epoch": 0.57, + "learning_rate": 0.00034615847542627886, + "loss": 1.3226, + "theoretical_loss": 3.475644587517324, + "tokens_seen": 1886912512 + }, + { + "epoch": 0.57, + "learning_rate": 0.00034614242728184556, + "loss": 1.3117, + "theoretical_loss": 3.4756344905545977, + "tokens_seen": 1886978048 + }, + { + "epoch": 0.57, + "learning_rate": 0.00034612637913741226, + "loss": 1.3553, + "theoretical_loss": 3.4756243940407243, + "tokens_seen": 1887043584 + }, + { + "epoch": 0.57, + "learning_rate": 0.00034611033099297896, + "loss": 1.3197, + "theoretical_loss": 3.4756142979756683, + "tokens_seen": 1887109120 + }, + { + "epoch": 0.57, + "learning_rate": 0.00034609428284854566, + "loss": 1.4064, + "theoretical_loss": 3.4756042023593947, + "tokens_seen": 1887174656 + }, + { + "epoch": 0.57, + "learning_rate": 0.00034607823470411236, + "loss": 1.2532, + "theoretical_loss": 3.475594107191867, + "tokens_seen": 1887240192 + }, + { + "epoch": 0.57, + "learning_rate": 0.00034606218655967906, + "loss": 1.3336, + "theoretical_loss": 3.4755840124730506, + "tokens_seen": 1887305728 + }, + { + "epoch": 0.57, + "learning_rate": 0.00034604613841524576, + "loss": 1.4583, + "theoretical_loss": 3.475573918202909, + "tokens_seen": 1887371264 + }, + { + "epoch": 0.57, + "objective/train/docs_used": 688313, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.5138492584228516, + "objective/train/theoretical_loss": 3.4755638243814078, + "objective/train/tokens_used": 1907896800, + "theoretical_loss": 3.4755638243814078, + "tokens_seen": 1887436800 + }, + { + "epoch": 0.57, + "learning_rate": 0.00034603009027081246, + "loss": 1.3617, + "theoretical_loss": 3.4755638243814078, + "tokens_seen": 1887436800 + }, + { + "epoch": 0.57, + "learning_rate": 0.00034601404212637916, + "loss": 1.4441, + "theoretical_loss": 3.475553731008511, + "tokens_seen": 1887502336 + }, + { + "epoch": 0.57, + "learning_rate": 0.00034599799398194587, + "loss": 1.3926, + "theoretical_loss": 3.475543638084183, + "tokens_seen": 1887567872 + }, + { + "epoch": 0.57, + "learning_rate": 0.00034598194583751257, + "loss": 1.3395, + "theoretical_loss": 3.4755335456083873, + "tokens_seen": 1887633408 + }, + { + "epoch": 0.57, + "learning_rate": 0.00034596589769307927, + "loss": 1.4177, + "theoretical_loss": 3.4755234535810904, + "tokens_seen": 1887698944 + }, + { + "epoch": 0.57, + "learning_rate": 0.00034594984954864597, + "loss": 1.3252, + "theoretical_loss": 3.4755133620022556, + "tokens_seen": 1887764480 + }, + { + "epoch": 0.57, + "learning_rate": 0.00034593380140421267, + "loss": 1.3068, + "theoretical_loss": 3.4755032708718474, + "tokens_seen": 1887830016 + }, + { + "epoch": 0.57, + "learning_rate": 0.00034591775325977937, + "loss": 1.3951, + "theoretical_loss": 3.4754931801898308, + "tokens_seen": 1887895552 + }, + { + "epoch": 0.57, + "learning_rate": 0.00034590170511534607, + "loss": 1.2466, + "theoretical_loss": 3.4754830899561697, + "tokens_seen": 1887961088 + }, + { + "epoch": 0.57, + "learning_rate": 0.00034588565697091277, + "loss": 1.3049, + "theoretical_loss": 3.4754730001708287, + "tokens_seen": 1888026624 + }, + { + "epoch": 0.57, + "learning_rate": 0.00034586960882647947, + "loss": 1.307, + "theoretical_loss": 3.475462910833773, + "tokens_seen": 1888092160 + }, + { + "epoch": 0.57, + "learning_rate": 0.00034585356068204617, + "loss": 1.3426, + "theoretical_loss": 3.475452821944967, + "tokens_seen": 1888157696 + }, + { + "epoch": 0.57, + "learning_rate": 0.00034583751253761287, + "loss": 1.4109, + "theoretical_loss": 3.4754427335043743, + "tokens_seen": 1888223232 + }, + { + "epoch": 0.57, + "learning_rate": 0.00034582146439317957, + "loss": 1.2866, + "theoretical_loss": 3.47543264551196, + "tokens_seen": 1888288768 + }, + { + "epoch": 0.57, + "learning_rate": 0.0003458054162487463, + "loss": 1.3601, + "theoretical_loss": 3.475422557967689, + "tokens_seen": 1888354304 + }, + { + "epoch": 0.57, + "learning_rate": 0.000345789368104313, + "loss": 1.3404, + "theoretical_loss": 3.4754124708715257, + "tokens_seen": 1888419840 + }, + { + "epoch": 0.57, + "learning_rate": 0.0003457733199598797, + "loss": 1.3833, + "theoretical_loss": 3.4754023842234343, + "tokens_seen": 1888485376 + }, + { + "epoch": 0.57, + "learning_rate": 0.0003457572718154464, + "loss": 1.1914, + "theoretical_loss": 3.4753922980233796, + "tokens_seen": 1888550912 + }, + { + "epoch": 0.57, + "learning_rate": 0.000345741223671013, + "loss": 1.3027, + "theoretical_loss": 3.4753822122713256, + "tokens_seen": 1888616448 + }, + { + "epoch": 0.57, + "learning_rate": 0.0003457251755265797, + "loss": 1.2199, + "theoretical_loss": 3.4753721269672373, + "tokens_seen": 1888681984 + }, + { + "epoch": 0.57, + "learning_rate": 0.0003457091273821464, + "loss": 1.2593, + "theoretical_loss": 3.4753620421110796, + "tokens_seen": 1888747520 + }, + { + "epoch": 0.57, + "learning_rate": 0.0003456930792377131, + "loss": 1.3293, + "theoretical_loss": 3.4753519577028165, + "tokens_seen": 1888813056 + }, + { + "epoch": 0.57, + "learning_rate": 0.0003456770310932798, + "loss": 1.2613, + "theoretical_loss": 3.475341873742413, + "tokens_seen": 1888878592 + }, + { + "epoch": 0.57, + "learning_rate": 0.0003456609829488465, + "loss": 1.3652, + "theoretical_loss": 3.4753317902298337, + "tokens_seen": 1888944128 + }, + { + "epoch": 0.57, + "learning_rate": 0.0003456449348044132, + "loss": 1.3145, + "theoretical_loss": 3.4753217071650426, + "tokens_seen": 1889009664 + }, + { + "epoch": 0.57, + "learning_rate": 0.0003456288866599799, + "loss": 1.2601, + "theoretical_loss": 3.475311624548005, + "tokens_seen": 1889075200 + }, + { + "epoch": 0.57, + "learning_rate": 0.0003456128385155466, + "loss": 1.3323, + "theoretical_loss": 3.4753015423786846, + "tokens_seen": 1889140736 + }, + { + "epoch": 0.57, + "learning_rate": 0.00034559679037111333, + "loss": 1.3817, + "theoretical_loss": 3.475291460657047, + "tokens_seen": 1889206272 + }, + { + "epoch": 0.57, + "learning_rate": 0.00034558074222668003, + "loss": 1.3137, + "theoretical_loss": 3.475281379383056, + "tokens_seen": 1889271808 + }, + { + "epoch": 0.57, + "learning_rate": 0.00034556469408224673, + "loss": 1.4302, + "theoretical_loss": 3.4752712985566765, + "tokens_seen": 1889337344 + }, + { + "epoch": 0.57, + "learning_rate": 0.00034554864593781343, + "loss": 1.3591, + "theoretical_loss": 3.475261218177873, + "tokens_seen": 1889402880 + }, + { + "epoch": 0.57, + "learning_rate": 0.00034553259779338013, + "loss": 1.2381, + "theoretical_loss": 3.4752511382466107, + "tokens_seen": 1889468416 + }, + { + "epoch": 0.57, + "learning_rate": 0.00034551654964894683, + "loss": 1.315, + "theoretical_loss": 3.4752410587628533, + "tokens_seen": 1889533952 + }, + { + "epoch": 0.57, + "learning_rate": 0.00034550050150451353, + "loss": 1.4041, + "theoretical_loss": 3.475230979726566, + "tokens_seen": 1889599488 + }, + { + "epoch": 0.57, + "learning_rate": 0.00034548445336008023, + "loss": 1.3444, + "theoretical_loss": 3.475220901137713, + "tokens_seen": 1889665024 + }, + { + "epoch": 0.57, + "learning_rate": 0.00034546840521564693, + "loss": 1.3591, + "theoretical_loss": 3.4752108229962593, + "tokens_seen": 1889730560 + }, + { + "epoch": 0.57, + "learning_rate": 0.00034545235707121363, + "loss": 1.3302, + "theoretical_loss": 3.4752007453021694, + "tokens_seen": 1889796096 + }, + { + "epoch": 0.57, + "learning_rate": 0.00034543630892678033, + "loss": 1.4167, + "theoretical_loss": 3.475190668055408, + "tokens_seen": 1889861632 + }, + { + "epoch": 0.57, + "learning_rate": 0.00034542026078234703, + "loss": 1.3339, + "theoretical_loss": 3.4751805912559393, + "tokens_seen": 1889927168 + }, + { + "epoch": 0.57, + "learning_rate": 0.00034540421263791373, + "loss": 1.4084, + "theoretical_loss": 3.475170514903729, + "tokens_seen": 1889992704 + }, + { + "epoch": 0.57, + "learning_rate": 0.00034538816449348044, + "loss": 1.4455, + "theoretical_loss": 3.4751604389987403, + "tokens_seen": 1890058240 + }, + { + "epoch": 0.57, + "learning_rate": 0.00034537211634904714, + "loss": 1.2176, + "theoretical_loss": 3.475150363540939, + "tokens_seen": 1890123776 + }, + { + "epoch": 0.57, + "learning_rate": 0.00034535606820461384, + "loss": 1.3034, + "theoretical_loss": 3.475140288530289, + "tokens_seen": 1890189312 + }, + { + "epoch": 0.57, + "learning_rate": 0.0003453400200601806, + "loss": 1.3677, + "theoretical_loss": 3.4751302139667555, + "tokens_seen": 1890254848 + }, + { + "epoch": 0.57, + "learning_rate": 0.0003453239719157473, + "loss": 1.3483, + "theoretical_loss": 3.4751201398503024, + "tokens_seen": 1890320384 + }, + { + "epoch": 0.57, + "learning_rate": 0.000345307923771314, + "loss": 1.2837, + "theoretical_loss": 3.4751100661808954, + "tokens_seen": 1890385920 + }, + { + "epoch": 0.57, + "learning_rate": 0.0003452918756268807, + "loss": 1.384, + "theoretical_loss": 3.4750999929584987, + "tokens_seen": 1890451456 + }, + { + "epoch": 0.57, + "learning_rate": 0.0003452758274824474, + "loss": 1.2557, + "theoretical_loss": 3.475089920183077, + "tokens_seen": 1890516992 + }, + { + "epoch": 0.57, + "learning_rate": 0.0003452597793380141, + "loss": 1.4911, + "theoretical_loss": 3.4750798478545946, + "tokens_seen": 1890582528 + }, + { + "epoch": 0.57, + "learning_rate": 0.0003452437311935808, + "loss": 1.3047, + "theoretical_loss": 3.4750697759730165, + "tokens_seen": 1890648064 + }, + { + "epoch": 0.57, + "objective/train/docs_used": 689504, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.3138729333877563, + "objective/train/theoretical_loss": 3.4750597045383076, + "objective/train/tokens_used": 1911173600, + "theoretical_loss": 3.4750597045383076, + "tokens_seen": 1890713600 + }, + { + "epoch": 0.57, + "learning_rate": 0.0003452276830491475, + "loss": 1.29, + "theoretical_loss": 3.4750597045383076, + "tokens_seen": 1890713600 + }, + { + "epoch": 0.57, + "learning_rate": 0.0003452116349047142, + "loss": 1.2711, + "theoretical_loss": 3.475049633550432, + "tokens_seen": 1890779136 + }, + { + "epoch": 0.57, + "learning_rate": 0.0003451955867602809, + "loss": 1.2497, + "theoretical_loss": 3.475039563009355, + "tokens_seen": 1890844672 + }, + { + "epoch": 0.57, + "learning_rate": 0.0003451795386158476, + "loss": 1.291, + "theoretical_loss": 3.4750294929150405, + "tokens_seen": 1890910208 + }, + { + "epoch": 0.57, + "learning_rate": 0.0003451634904714143, + "loss": 1.4086, + "theoretical_loss": 3.4750194232674545, + "tokens_seen": 1890975744 + }, + { + "epoch": 0.57, + "learning_rate": 0.000345147442326981, + "loss": 1.2735, + "theoretical_loss": 3.4750093540665605, + "tokens_seen": 1891041280 + }, + { + "epoch": 0.57, + "learning_rate": 0.0003451313941825477, + "loss": 1.3073, + "theoretical_loss": 3.4749992853123235, + "tokens_seen": 1891106816 + }, + { + "epoch": 0.57, + "learning_rate": 0.00034511534603811435, + "loss": 1.3258, + "theoretical_loss": 3.4749892170047083, + "tokens_seen": 1891172352 + }, + { + "epoch": 0.57, + "learning_rate": 0.00034509929789368105, + "loss": 1.4309, + "theoretical_loss": 3.47497914914368, + "tokens_seen": 1891237888 + }, + { + "epoch": 0.57, + "learning_rate": 0.00034508324974924775, + "loss": 1.3753, + "theoretical_loss": 3.474969081729203, + "tokens_seen": 1891303424 + }, + { + "epoch": 0.57, + "learning_rate": 0.00034506720160481445, + "loss": 1.3347, + "theoretical_loss": 3.4749590147612417, + "tokens_seen": 1891368960 + }, + { + "epoch": 0.57, + "learning_rate": 0.00034505115346038115, + "loss": 1.3117, + "theoretical_loss": 3.4749489482397613, + "tokens_seen": 1891434496 + }, + { + "epoch": 0.57, + "learning_rate": 0.00034503510531594785, + "loss": 1.3222, + "theoretical_loss": 3.474938882164726, + "tokens_seen": 1891500032 + }, + { + "epoch": 0.57, + "learning_rate": 0.00034501905717151455, + "loss": 1.3904, + "theoretical_loss": 3.474928816536101, + "tokens_seen": 1891565568 + }, + { + "epoch": 0.57, + "learning_rate": 0.00034500300902708125, + "loss": 1.4101, + "theoretical_loss": 3.4749187513538513, + "tokens_seen": 1891631104 + }, + { + "epoch": 0.57, + "learning_rate": 0.00034498696088264795, + "loss": 1.2906, + "theoretical_loss": 3.4749086866179413, + "tokens_seen": 1891696640 + }, + { + "epoch": 0.57, + "learning_rate": 0.00034497091273821465, + "loss": 1.329, + "theoretical_loss": 3.474898622328335, + "tokens_seen": 1891762176 + }, + { + "epoch": 0.57, + "learning_rate": 0.00034495486459378135, + "loss": 1.3094, + "theoretical_loss": 3.4748885584849987, + "tokens_seen": 1891827712 + }, + { + "epoch": 0.57, + "learning_rate": 0.00034493881644934805, + "loss": 1.3368, + "theoretical_loss": 3.4748784950878955, + "tokens_seen": 1891893248 + }, + { + "epoch": 0.57, + "learning_rate": 0.00034492276830491475, + "loss": 1.3859, + "theoretical_loss": 3.474868432136991, + "tokens_seen": 1891958784 + }, + { + "epoch": 0.57, + "learning_rate": 0.00034490672016048145, + "loss": 1.3338, + "theoretical_loss": 3.4748583696322504, + "tokens_seen": 1892024320 + }, + { + "epoch": 0.57, + "learning_rate": 0.00034489067201604816, + "loss": 1.3444, + "theoretical_loss": 3.474848307573638, + "tokens_seen": 1892089856 + }, + { + "epoch": 0.57, + "learning_rate": 0.00034487462387161486, + "loss": 1.3264, + "theoretical_loss": 3.4748382459611182, + "tokens_seen": 1892155392 + }, + { + "epoch": 0.57, + "learning_rate": 0.00034485857572718156, + "loss": 1.3606, + "theoretical_loss": 3.474828184794656, + "tokens_seen": 1892220928 + }, + { + "epoch": 0.57, + "learning_rate": 0.00034484252758274826, + "loss": 1.4234, + "theoretical_loss": 3.4748181240742166, + "tokens_seen": 1892286464 + }, + { + "epoch": 0.57, + "learning_rate": 0.00034482647943831496, + "loss": 1.3101, + "theoretical_loss": 3.474808063799764, + "tokens_seen": 1892352000 + }, + { + "epoch": 0.57, + "learning_rate": 0.00034481043129388166, + "loss": 1.4042, + "theoretical_loss": 3.474798003971264, + "tokens_seen": 1892417536 + }, + { + "epoch": 0.57, + "learning_rate": 0.00034479438314944836, + "loss": 1.3709, + "theoretical_loss": 3.474787944588681, + "tokens_seen": 1892483072 + }, + { + "epoch": 0.57, + "learning_rate": 0.00034477833500501506, + "loss": 1.4084, + "theoretical_loss": 3.4747778856519793, + "tokens_seen": 1892548608 + }, + { + "epoch": 0.57, + "learning_rate": 0.00034476228686058176, + "loss": 1.4141, + "theoretical_loss": 3.4747678271611244, + "tokens_seen": 1892614144 + }, + { + "epoch": 0.57, + "learning_rate": 0.00034474623871614846, + "loss": 1.3253, + "theoretical_loss": 3.4747577691160805, + "tokens_seen": 1892679680 + }, + { + "epoch": 0.57, + "learning_rate": 0.00034473019057171516, + "loss": 1.2816, + "theoretical_loss": 3.4747477115168124, + "tokens_seen": 1892745216 + }, + { + "epoch": 0.57, + "learning_rate": 0.00034471414242728186, + "loss": 1.3669, + "theoretical_loss": 3.4747376543632855, + "tokens_seen": 1892810752 + }, + { + "epoch": 0.57, + "learning_rate": 0.00034469809428284856, + "loss": 1.3866, + "theoretical_loss": 3.474727597655464, + "tokens_seen": 1892876288 + }, + { + "epoch": 0.57, + "learning_rate": 0.00034468204613841526, + "loss": 1.2461, + "theoretical_loss": 3.4747175413933133, + "tokens_seen": 1892941824 + }, + { + "epoch": 0.57, + "learning_rate": 0.00034466599799398196, + "loss": 1.2983, + "theoretical_loss": 3.4747074855767974, + "tokens_seen": 1893007360 + }, + { + "epoch": 0.57, + "learning_rate": 0.00034464994984954866, + "loss": 1.3331, + "theoretical_loss": 3.4746974302058824, + "tokens_seen": 1893072896 + }, + { + "epoch": 0.57, + "learning_rate": 0.00034463390170511537, + "loss": 1.3189, + "theoretical_loss": 3.4746873752805314, + "tokens_seen": 1893138432 + }, + { + "epoch": 0.57, + "learning_rate": 0.00034461785356068207, + "loss": 1.3803, + "theoretical_loss": 3.474677320800711, + "tokens_seen": 1893203968 + }, + { + "epoch": 0.57, + "learning_rate": 0.00034460180541624877, + "loss": 1.3192, + "theoretical_loss": 3.474667266766385, + "tokens_seen": 1893269504 + }, + { + "epoch": 0.57, + "learning_rate": 0.00034458575727181547, + "loss": 1.4146, + "theoretical_loss": 3.4746572131775184, + "tokens_seen": 1893335040 + }, + { + "epoch": 0.57, + "learning_rate": 0.00034456970912738217, + "loss": 1.3985, + "theoretical_loss": 3.4746471600340763, + "tokens_seen": 1893400576 + }, + { + "epoch": 0.57, + "learning_rate": 0.00034455366098294887, + "loss": 1.3513, + "theoretical_loss": 3.474637107336023, + "tokens_seen": 1893466112 + }, + { + "epoch": 0.57, + "learning_rate": 0.00034453761283851557, + "loss": 1.3905, + "theoretical_loss": 3.474627055083324, + "tokens_seen": 1893531648 + }, + { + "epoch": 0.57, + "learning_rate": 0.00034452156469408227, + "loss": 1.3296, + "theoretical_loss": 3.4746170032759442, + "tokens_seen": 1893597184 + }, + { + "epoch": 0.57, + "learning_rate": 0.00034450551654964897, + "loss": 1.3241, + "theoretical_loss": 3.474606951913848, + "tokens_seen": 1893662720 + }, + { + "epoch": 0.57, + "learning_rate": 0.00034448946840521567, + "loss": 1.3191, + "theoretical_loss": 3.4745969009969997, + "tokens_seen": 1893728256 + }, + { + "epoch": 0.57, + "learning_rate": 0.00034447342026078237, + "loss": 1.418, + "theoretical_loss": 3.4745868505253656, + "tokens_seen": 1893793792 + }, + { + "epoch": 0.57, + "learning_rate": 0.00034445737211634907, + "loss": 1.2954, + "theoretical_loss": 3.4745768004989097, + "tokens_seen": 1893859328 + }, + { + "epoch": 0.57, + "learning_rate": 0.00034444132397191577, + "loss": 1.3793, + "theoretical_loss": 3.474566750917597, + "tokens_seen": 1893924864 + }, + { + "epoch": 0.57, + "objective/train/docs_used": 690600, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.3900630474090576, + "objective/train/theoretical_loss": 3.474556701781392, + "objective/train/tokens_used": 1914450400, + "theoretical_loss": 3.474556701781392, + "tokens_seen": 1893990400 + }, + { + "epoch": 0.57, + "learning_rate": 0.0003444252758274825, + "loss": 1.3223, + "theoretical_loss": 3.474556701781392, + "tokens_seen": 1893990400 + }, + { + "epoch": 0.57, + "learning_rate": 0.0003444092276830492, + "loss": 1.4117, + "theoretical_loss": 3.4745466530902607, + "tokens_seen": 1894055936 + }, + { + "epoch": 0.57, + "learning_rate": 0.0003443931795386159, + "loss": 1.3662, + "theoretical_loss": 3.474536604844167, + "tokens_seen": 1894121472 + }, + { + "epoch": 0.57, + "learning_rate": 0.0003443771313941826, + "loss": 1.273, + "theoretical_loss": 3.4745265570430757, + "tokens_seen": 1894187008 + }, + { + "epoch": 0.57, + "learning_rate": 0.0003443610832497493, + "loss": 1.4394, + "theoretical_loss": 3.4745165096869526, + "tokens_seen": 1894252544 + }, + { + "epoch": 0.57, + "learning_rate": 0.0003443450351053159, + "loss": 1.4376, + "theoretical_loss": 3.4745064627757616, + "tokens_seen": 1894318080 + }, + { + "epoch": 0.57, + "learning_rate": 0.0003443289869608826, + "loss": 1.3021, + "theoretical_loss": 3.4744964163094685, + "tokens_seen": 1894383616 + }, + { + "epoch": 0.57, + "learning_rate": 0.0003443129388164493, + "loss": 1.3791, + "theoretical_loss": 3.4744863702880378, + "tokens_seen": 1894449152 + }, + { + "epoch": 0.57, + "learning_rate": 0.000344296890672016, + "loss": 1.2837, + "theoretical_loss": 3.474476324711434, + "tokens_seen": 1894514688 + }, + { + "epoch": 0.57, + "learning_rate": 0.0003442808425275827, + "loss": 1.4148, + "theoretical_loss": 3.474466279579623, + "tokens_seen": 1894580224 + }, + { + "epoch": 0.57, + "learning_rate": 0.0003442647943831494, + "loss": 1.3502, + "theoretical_loss": 3.4744562348925685, + "tokens_seen": 1894645760 + }, + { + "epoch": 0.57, + "learning_rate": 0.0003442487462387161, + "loss": 1.3045, + "theoretical_loss": 3.4744461906502364, + "tokens_seen": 1894711296 + }, + { + "epoch": 0.57, + "learning_rate": 0.0003442326980942828, + "loss": 1.3615, + "theoretical_loss": 3.4744361468525913, + "tokens_seen": 1894776832 + }, + { + "epoch": 0.57, + "learning_rate": 0.00034421664994984953, + "loss": 1.3116, + "theoretical_loss": 3.474426103499598, + "tokens_seen": 1894842368 + }, + { + "epoch": 0.57, + "learning_rate": 0.00034420060180541623, + "loss": 1.3343, + "theoretical_loss": 3.4744160605912215, + "tokens_seen": 1894907904 + }, + { + "epoch": 0.57, + "learning_rate": 0.00034418455366098293, + "loss": 1.2419, + "theoretical_loss": 3.4744060181274268, + "tokens_seen": 1894973440 + }, + { + "epoch": 0.57, + "learning_rate": 0.00034416850551654963, + "loss": 1.3333, + "theoretical_loss": 3.474395976108179, + "tokens_seen": 1895038976 + }, + { + "epoch": 0.57, + "learning_rate": 0.00034415245737211633, + "loss": 1.3948, + "theoretical_loss": 3.474385934533443, + "tokens_seen": 1895104512 + }, + { + "epoch": 0.57, + "learning_rate": 0.0003441364092276831, + "loss": 1.2736, + "theoretical_loss": 3.4743758934031836, + "tokens_seen": 1895170048 + }, + { + "epoch": 0.57, + "learning_rate": 0.0003441203610832498, + "loss": 1.406, + "theoretical_loss": 3.4743658527173658, + "tokens_seen": 1895235584 + }, + { + "epoch": 0.57, + "learning_rate": 0.0003441043129388165, + "loss": 1.3433, + "theoretical_loss": 3.474355812475954, + "tokens_seen": 1895301120 + }, + { + "epoch": 0.57, + "learning_rate": 0.0003440882647943832, + "loss": 1.2816, + "theoretical_loss": 3.4743457726789146, + "tokens_seen": 1895366656 + }, + { + "epoch": 0.57, + "learning_rate": 0.0003440722166499499, + "loss": 1.3852, + "theoretical_loss": 3.4743357333262113, + "tokens_seen": 1895432192 + }, + { + "epoch": 0.57, + "learning_rate": 0.0003440561685055166, + "loss": 1.3357, + "theoretical_loss": 3.4743256944178094, + "tokens_seen": 1895497728 + }, + { + "epoch": 0.57, + "learning_rate": 0.0003440401203610833, + "loss": 1.3136, + "theoretical_loss": 3.4743156559536743, + "tokens_seen": 1895563264 + }, + { + "epoch": 0.57, + "learning_rate": 0.00034402407221665, + "loss": 1.3784, + "theoretical_loss": 3.47430561793377, + "tokens_seen": 1895628800 + }, + { + "epoch": 0.57, + "learning_rate": 0.0003440080240722167, + "loss": 1.4467, + "theoretical_loss": 3.4742955803580626, + "tokens_seen": 1895694336 + }, + { + "epoch": 0.57, + "learning_rate": 0.0003439919759277834, + "loss": 1.4874, + "theoretical_loss": 3.4742855432265163, + "tokens_seen": 1895759872 + }, + { + "epoch": 0.57, + "learning_rate": 0.0003439759277833501, + "loss": 1.3521, + "theoretical_loss": 3.4742755065390964, + "tokens_seen": 1895825408 + }, + { + "epoch": 0.57, + "learning_rate": 0.0003439598796389168, + "loss": 1.4356, + "theoretical_loss": 3.4742654702957685, + "tokens_seen": 1895890944 + }, + { + "epoch": 0.57, + "learning_rate": 0.0003439438314944835, + "loss": 1.2878, + "theoretical_loss": 3.4742554344964964, + "tokens_seen": 1895956480 + }, + { + "epoch": 0.57, + "learning_rate": 0.0003439277833500502, + "loss": 1.3403, + "theoretical_loss": 3.474245399141246, + "tokens_seen": 1896022016 + }, + { + "epoch": 0.57, + "learning_rate": 0.0003439117352056169, + "loss": 1.2981, + "theoretical_loss": 3.4742353642299815, + "tokens_seen": 1896087552 + }, + { + "epoch": 0.57, + "learning_rate": 0.0003438956870611836, + "loss": 1.4133, + "theoretical_loss": 3.474225329762669, + "tokens_seen": 1896153088 + }, + { + "epoch": 0.57, + "learning_rate": 0.0003438796389167503, + "loss": 1.2743, + "theoretical_loss": 3.4742152957392722, + "tokens_seen": 1896218624 + }, + { + "epoch": 0.57, + "learning_rate": 0.000343863590772317, + "loss": 1.3282, + "theoretical_loss": 3.4742052621597574, + "tokens_seen": 1896284160 + }, + { + "epoch": 0.57, + "learning_rate": 0.0003438475426278837, + "loss": 1.3565, + "theoretical_loss": 3.4741952290240894, + "tokens_seen": 1896349696 + }, + { + "epoch": 0.57, + "learning_rate": 0.0003438314944834504, + "loss": 1.3868, + "theoretical_loss": 3.4741851963322326, + "tokens_seen": 1896415232 + }, + { + "epoch": 0.57, + "learning_rate": 0.0003438154463390171, + "loss": 1.4599, + "theoretical_loss": 3.4741751640841523, + "tokens_seen": 1896480768 + }, + { + "epoch": 0.57, + "learning_rate": 0.0003437993981945838, + "loss": 1.3185, + "theoretical_loss": 3.474165132279813, + "tokens_seen": 1896546304 + }, + { + "epoch": 0.57, + "learning_rate": 0.0003437833500501505, + "loss": 1.2874, + "theoretical_loss": 3.474155100919181, + "tokens_seen": 1896611840 + }, + { + "epoch": 0.57, + "learning_rate": 0.0003437673019057172, + "loss": 1.3655, + "theoretical_loss": 3.47414507000222, + "tokens_seen": 1896677376 + }, + { + "epoch": 0.57, + "learning_rate": 0.0003437512537612839, + "loss": 1.247, + "theoretical_loss": 3.4741350395288966, + "tokens_seen": 1896742912 + }, + { + "epoch": 0.57, + "learning_rate": 0.0003437352056168506, + "loss": 1.3629, + "theoretical_loss": 3.4741250094991747, + "tokens_seen": 1896808448 + }, + { + "epoch": 0.57, + "learning_rate": 0.0003437191574724173, + "loss": 1.4054, + "theoretical_loss": 3.4741149799130193, + "tokens_seen": 1896873984 + }, + { + "epoch": 0.57, + "learning_rate": 0.00034370310932798395, + "loss": 1.3099, + "theoretical_loss": 3.474104950770396, + "tokens_seen": 1896939520 + }, + { + "epoch": 0.57, + "learning_rate": 0.00034368706118355065, + "loss": 1.3109, + "theoretical_loss": 3.474094922071269, + "tokens_seen": 1897005056 + }, + { + "epoch": 0.57, + "learning_rate": 0.00034367101303911735, + "loss": 1.4505, + "theoretical_loss": 3.474084893815605, + "tokens_seen": 1897070592 + }, + { + "epoch": 0.57, + "learning_rate": 0.00034365496489468405, + "loss": 1.3371, + "theoretical_loss": 3.4740748660033676, + "tokens_seen": 1897136128 + }, + { + "epoch": 0.57, + "learning_rate": 0.00034363891675025075, + "loss": 1.4562, + "theoretical_loss": 3.474064838634522, + "tokens_seen": 1897201664 + }, + { + "epoch": 0.57, + "objective/train/docs_used": 691800, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.4519823789596558, + "objective/train/theoretical_loss": 3.474054811709034, + "objective/train/tokens_used": 1917727200, + "theoretical_loss": 3.474054811709034, + "tokens_seen": 1897267200 + }, + { + "epoch": 0.57, + "learning_rate": 0.00034362286860581745, + "loss": 1.3866, + "theoretical_loss": 3.474054811709034, + "tokens_seen": 1897267200 + }, + { + "epoch": 0.57, + "learning_rate": 0.00034360682046138415, + "loss": 1.4234, + "theoretical_loss": 3.4740447852268685, + "tokens_seen": 1897332736 + }, + { + "epoch": 0.57, + "learning_rate": 0.00034359077231695085, + "loss": 1.3754, + "theoretical_loss": 3.4740347591879903, + "tokens_seen": 1897398272 + }, + { + "epoch": 0.57, + "learning_rate": 0.00034357472417251755, + "loss": 1.283, + "theoretical_loss": 3.4740247335923646, + "tokens_seen": 1897463808 + }, + { + "epoch": 0.58, + "learning_rate": 0.00034355867602808425, + "loss": 1.3717, + "theoretical_loss": 3.474014708439956, + "tokens_seen": 1897529344 + }, + { + "epoch": 0.58, + "learning_rate": 0.00034354262788365095, + "loss": 1.3528, + "theoretical_loss": 3.4740046837307306, + "tokens_seen": 1897594880 + }, + { + "epoch": 0.58, + "learning_rate": 0.00034352657973921765, + "loss": 1.3551, + "theoretical_loss": 3.473994659464653, + "tokens_seen": 1897660416 + }, + { + "epoch": 0.58, + "learning_rate": 0.00034351053159478436, + "loss": 1.339, + "theoretical_loss": 3.4739846356416875, + "tokens_seen": 1897725952 + }, + { + "epoch": 0.58, + "learning_rate": 0.00034349448345035106, + "loss": 1.3029, + "theoretical_loss": 3.473974612261801, + "tokens_seen": 1897791488 + }, + { + "epoch": 0.58, + "learning_rate": 0.00034347843530591776, + "loss": 1.2216, + "theoretical_loss": 3.4739645893249573, + "tokens_seen": 1897857024 + }, + { + "epoch": 0.58, + "learning_rate": 0.00034346238716148446, + "loss": 1.4184, + "theoretical_loss": 3.4739545668311216, + "tokens_seen": 1897922560 + }, + { + "epoch": 0.58, + "learning_rate": 0.00034344633901705116, + "loss": 1.2287, + "theoretical_loss": 3.4739445447802595, + "tokens_seen": 1897988096 + }, + { + "epoch": 0.58, + "learning_rate": 0.00034343029087261786, + "loss": 1.3643, + "theoretical_loss": 3.473934523172336, + "tokens_seen": 1898053632 + }, + { + "epoch": 0.58, + "learning_rate": 0.00034341424272818456, + "loss": 1.3722, + "theoretical_loss": 3.473924502007316, + "tokens_seen": 1898119168 + }, + { + "epoch": 0.58, + "learning_rate": 0.00034339819458375126, + "loss": 1.3405, + "theoretical_loss": 3.4739144812851643, + "tokens_seen": 1898184704 + }, + { + "epoch": 0.58, + "learning_rate": 0.00034338214643931796, + "loss": 1.2137, + "theoretical_loss": 3.473904461005847, + "tokens_seen": 1898250240 + }, + { + "epoch": 0.58, + "learning_rate": 0.00034336609829488466, + "loss": 1.2939, + "theoretical_loss": 3.473894441169329, + "tokens_seen": 1898315776 + }, + { + "epoch": 0.58, + "learning_rate": 0.00034335005015045136, + "loss": 1.3099, + "theoretical_loss": 3.4738844217755744, + "tokens_seen": 1898381312 + }, + { + "epoch": 0.58, + "learning_rate": 0.00034333400200601806, + "loss": 1.3342, + "theoretical_loss": 3.47387440282455, + "tokens_seen": 1898446848 + }, + { + "epoch": 0.58, + "learning_rate": 0.00034331795386158476, + "loss": 1.3312, + "theoretical_loss": 3.4738643843162196, + "tokens_seen": 1898512384 + }, + { + "epoch": 0.58, + "learning_rate": 0.00034330190571715146, + "loss": 1.4467, + "theoretical_loss": 3.4738543662505488, + "tokens_seen": 1898577920 + }, + { + "epoch": 0.58, + "learning_rate": 0.00034328585757271816, + "loss": 1.3472, + "theoretical_loss": 3.473844348627503, + "tokens_seen": 1898643456 + }, + { + "epoch": 0.58, + "learning_rate": 0.00034326980942828487, + "loss": 1.3341, + "theoretical_loss": 3.473834331447047, + "tokens_seen": 1898708992 + }, + { + "epoch": 0.58, + "learning_rate": 0.00034325376128385157, + "loss": 1.2564, + "theoretical_loss": 3.4738243147091463, + "tokens_seen": 1898774528 + }, + { + "epoch": 0.58, + "learning_rate": 0.00034323771313941827, + "loss": 1.2595, + "theoretical_loss": 3.4738142984137657, + "tokens_seen": 1898840064 + }, + { + "epoch": 0.58, + "learning_rate": 0.00034322166499498497, + "loss": 1.3084, + "theoretical_loss": 3.473804282560871, + "tokens_seen": 1898905600 + }, + { + "epoch": 0.58, + "learning_rate": 0.00034320561685055167, + "loss": 1.3279, + "theoretical_loss": 3.473794267150427, + "tokens_seen": 1898971136 + }, + { + "epoch": 0.58, + "learning_rate": 0.00034318956870611837, + "loss": 1.2422, + "theoretical_loss": 3.4737842521823983, + "tokens_seen": 1899036672 + }, + { + "epoch": 0.58, + "learning_rate": 0.00034317352056168507, + "loss": 1.3637, + "theoretical_loss": 3.473774237656751, + "tokens_seen": 1899102208 + }, + { + "epoch": 0.58, + "learning_rate": 0.00034315747241725177, + "loss": 1.3013, + "theoretical_loss": 3.47376422357345, + "tokens_seen": 1899167744 + }, + { + "epoch": 0.58, + "learning_rate": 0.00034314142427281847, + "loss": 1.4541, + "theoretical_loss": 3.4737542099324603, + "tokens_seen": 1899233280 + }, + { + "epoch": 0.58, + "learning_rate": 0.00034312537612838517, + "loss": 1.351, + "theoretical_loss": 3.4737441967337475, + "tokens_seen": 1899298816 + }, + { + "epoch": 0.58, + "learning_rate": 0.00034310932798395187, + "loss": 1.3655, + "theoretical_loss": 3.4737341839772764, + "tokens_seen": 1899364352 + }, + { + "epoch": 0.58, + "learning_rate": 0.00034309327983951857, + "loss": 1.3537, + "theoretical_loss": 3.4737241716630125, + "tokens_seen": 1899429888 + }, + { + "epoch": 0.58, + "learning_rate": 0.00034307723169508527, + "loss": 1.2751, + "theoretical_loss": 3.4737141597909207, + "tokens_seen": 1899495424 + }, + { + "epoch": 0.58, + "learning_rate": 0.000343061183550652, + "loss": 1.312, + "theoretical_loss": 3.4737041483609663, + "tokens_seen": 1899560960 + }, + { + "epoch": 0.58, + "learning_rate": 0.0003430451354062187, + "loss": 1.2581, + "theoretical_loss": 3.4736941373731147, + "tokens_seen": 1899626496 + }, + { + "epoch": 0.58, + "learning_rate": 0.0003430290872617854, + "loss": 1.3255, + "theoretical_loss": 3.4736841268273313, + "tokens_seen": 1899692032 + }, + { + "epoch": 0.58, + "learning_rate": 0.0003430130391173521, + "loss": 1.4181, + "theoretical_loss": 3.473674116723581, + "tokens_seen": 1899757568 + }, + { + "epoch": 0.58, + "learning_rate": 0.0003429969909729188, + "loss": 1.326, + "theoretical_loss": 3.473664107061829, + "tokens_seen": 1899823104 + }, + { + "epoch": 0.58, + "learning_rate": 0.0003429809428284855, + "loss": 1.2749, + "theoretical_loss": 3.4736540978420405, + "tokens_seen": 1899888640 + }, + { + "epoch": 0.58, + "learning_rate": 0.0003429648946840522, + "loss": 1.3099, + "theoretical_loss": 3.473644089064181, + "tokens_seen": 1899954176 + }, + { + "epoch": 0.58, + "learning_rate": 0.0003429488465396189, + "loss": 1.4157, + "theoretical_loss": 3.473634080728216, + "tokens_seen": 1900019712 + }, + { + "epoch": 0.58, + "learning_rate": 0.0003429327983951856, + "loss": 1.3277, + "theoretical_loss": 3.47362407283411, + "tokens_seen": 1900085248 + }, + { + "epoch": 0.58, + "learning_rate": 0.0003429167502507523, + "loss": 1.3917, + "theoretical_loss": 3.4736140653818284, + "tokens_seen": 1900150784 + }, + { + "epoch": 0.58, + "learning_rate": 0.000342900702106319, + "loss": 1.3576, + "theoretical_loss": 3.473604058371337, + "tokens_seen": 1900216320 + }, + { + "epoch": 0.58, + "learning_rate": 0.0003428846539618857, + "loss": 1.3243, + "theoretical_loss": 3.4735940518026007, + "tokens_seen": 1900281856 + }, + { + "epoch": 0.58, + "learning_rate": 0.0003428686058174524, + "loss": 1.3414, + "theoretical_loss": 3.473584045675585, + "tokens_seen": 1900347392 + }, + { + "epoch": 0.58, + "learning_rate": 0.0003428525576730191, + "loss": 1.3988, + "theoretical_loss": 3.473574039990255, + "tokens_seen": 1900412928 + }, + { + "epoch": 0.58, + "learning_rate": 0.0003428365095285858, + "loss": 1.2832, + "theoretical_loss": 3.473564034746576, + "tokens_seen": 1900478464 + }, + { + "epoch": 0.58, + "objective/train/docs_used": 693007, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.2796458005905151, + "objective/train/theoretical_loss": 3.4735540299445127, + "objective/train/tokens_used": 1921004000, + "theoretical_loss": 3.4735540299445127, + "tokens_seen": 1900544000 + }, + { + "epoch": 0.58, + "learning_rate": 0.0003428204613841525, + "loss": 1.3078, + "theoretical_loss": 3.4735540299445127, + "tokens_seen": 1900544000 + }, + { + "epoch": 0.58, + "learning_rate": 0.0003428044132397192, + "loss": 1.2822, + "theoretical_loss": 3.4735440255840313, + "tokens_seen": 1900609536 + }, + { + "epoch": 0.58, + "learning_rate": 0.0003427883650952859, + "loss": 1.2625, + "theoretical_loss": 3.4735340216650967, + "tokens_seen": 1900675072 + }, + { + "epoch": 0.58, + "learning_rate": 0.0003427723169508526, + "loss": 1.3475, + "theoretical_loss": 3.4735240181876743, + "tokens_seen": 1900740608 + }, + { + "epoch": 0.58, + "learning_rate": 0.0003427562688064193, + "loss": 1.3717, + "theoretical_loss": 3.473514015151729, + "tokens_seen": 1900806144 + }, + { + "epoch": 0.58, + "learning_rate": 0.000342740220661986, + "loss": 1.397, + "theoretical_loss": 3.4735040125572265, + "tokens_seen": 1900871680 + }, + { + "epoch": 0.58, + "learning_rate": 0.0003427241725175527, + "loss": 1.3445, + "theoretical_loss": 3.4734940104041323, + "tokens_seen": 1900937216 + }, + { + "epoch": 0.58, + "learning_rate": 0.0003427081243731194, + "loss": 1.2989, + "theoretical_loss": 3.473484008692411, + "tokens_seen": 1901002752 + }, + { + "epoch": 0.58, + "learning_rate": 0.0003426920762286861, + "loss": 1.2937, + "theoretical_loss": 3.4734740074220283, + "tokens_seen": 1901068288 + }, + { + "epoch": 0.58, + "learning_rate": 0.0003426760280842528, + "loss": 1.349, + "theoretical_loss": 3.4734640065929496, + "tokens_seen": 1901133824 + }, + { + "epoch": 0.58, + "learning_rate": 0.0003426599799398195, + "loss": 1.3265, + "theoretical_loss": 3.47345400620514, + "tokens_seen": 1901199360 + }, + { + "epoch": 0.58, + "learning_rate": 0.0003426439317953862, + "loss": 1.3575, + "theoretical_loss": 3.473444006258565, + "tokens_seen": 1901264896 + }, + { + "epoch": 0.58, + "learning_rate": 0.0003426278836509529, + "loss": 1.3316, + "theoretical_loss": 3.47343400675319, + "tokens_seen": 1901330432 + }, + { + "epoch": 0.58, + "learning_rate": 0.0003426118355065196, + "loss": 1.3485, + "theoretical_loss": 3.47342400768898, + "tokens_seen": 1901395968 + }, + { + "epoch": 0.58, + "learning_rate": 0.0003425957873620863, + "loss": 1.3963, + "theoretical_loss": 3.473414009065901, + "tokens_seen": 1901461504 + }, + { + "epoch": 0.58, + "learning_rate": 0.000342579739217653, + "loss": 1.377, + "theoretical_loss": 3.4734040108839173, + "tokens_seen": 1901527040 + }, + { + "epoch": 0.58, + "learning_rate": 0.0003425636910732197, + "loss": 1.3249, + "theoretical_loss": 3.473394013142995, + "tokens_seen": 1901592576 + }, + { + "epoch": 0.58, + "learning_rate": 0.0003425476429287864, + "loss": 1.2899, + "theoretical_loss": 3.4733840158430995, + "tokens_seen": 1901658112 + }, + { + "epoch": 0.58, + "learning_rate": 0.0003425315947843531, + "loss": 1.3904, + "theoretical_loss": 3.4733740189841953, + "tokens_seen": 1901723648 + }, + { + "epoch": 0.58, + "learning_rate": 0.0003425155466399198, + "loss": 1.3324, + "theoretical_loss": 3.4733640225662485, + "tokens_seen": 1901789184 + }, + { + "epoch": 0.58, + "learning_rate": 0.0003424994984954865, + "loss": 1.4207, + "theoretical_loss": 3.4733540265892247, + "tokens_seen": 1901854720 + }, + { + "epoch": 0.58, + "learning_rate": 0.0003424834503510532, + "loss": 1.3671, + "theoretical_loss": 3.473344031053089, + "tokens_seen": 1901920256 + }, + { + "epoch": 0.58, + "learning_rate": 0.0003424674022066199, + "loss": 1.352, + "theoretical_loss": 3.473334035957806, + "tokens_seen": 1901985792 + }, + { + "epoch": 0.58, + "learning_rate": 0.0003424513540621866, + "loss": 1.3999, + "theoretical_loss": 3.473324041303342, + "tokens_seen": 1902051328 + }, + { + "epoch": 0.58, + "learning_rate": 0.0003424353059177533, + "loss": 1.331, + "theoretical_loss": 3.4733140470896617, + "tokens_seen": 1902116864 + }, + { + "epoch": 0.58, + "learning_rate": 0.00034241925777332, + "loss": 1.4014, + "theoretical_loss": 3.473304053316731, + "tokens_seen": 1902182400 + }, + { + "epoch": 0.58, + "learning_rate": 0.0003424032096288867, + "loss": 1.2526, + "theoretical_loss": 3.4732940599845152, + "tokens_seen": 1902247936 + }, + { + "epoch": 0.58, + "learning_rate": 0.0003423871614844534, + "loss": 1.4096, + "theoretical_loss": 3.4732840670929797, + "tokens_seen": 1902313472 + }, + { + "epoch": 0.58, + "learning_rate": 0.0003423711133400201, + "loss": 1.2959, + "theoretical_loss": 3.47327407464209, + "tokens_seen": 1902379008 + }, + { + "epoch": 0.58, + "learning_rate": 0.0003423550651955868, + "loss": 1.4568, + "theoretical_loss": 3.4732640826318106, + "tokens_seen": 1902444544 + }, + { + "epoch": 0.58, + "learning_rate": 0.0003423390170511535, + "loss": 1.4152, + "theoretical_loss": 3.473254091062108, + "tokens_seen": 1902510080 + }, + { + "epoch": 0.58, + "learning_rate": 0.0003423229689067202, + "loss": 1.3739, + "theoretical_loss": 3.473244099932947, + "tokens_seen": 1902575616 + }, + { + "epoch": 0.58, + "learning_rate": 0.00034230692076228685, + "loss": 1.3469, + "theoretical_loss": 3.4732341092442933, + "tokens_seen": 1902641152 + }, + { + "epoch": 0.58, + "learning_rate": 0.00034229087261785355, + "loss": 1.3364, + "theoretical_loss": 3.4732241189961117, + "tokens_seen": 1902706688 + }, + { + "epoch": 0.58, + "learning_rate": 0.00034227482447342025, + "loss": 1.3007, + "theoretical_loss": 3.473214129188369, + "tokens_seen": 1902772224 + }, + { + "epoch": 0.58, + "learning_rate": 0.00034225877632898695, + "loss": 1.3352, + "theoretical_loss": 3.473204139821029, + "tokens_seen": 1902837760 + }, + { + "epoch": 0.58, + "learning_rate": 0.00034224272818455365, + "loss": 1.3602, + "theoretical_loss": 3.473194150894058, + "tokens_seen": 1902903296 + }, + { + "epoch": 0.58, + "learning_rate": 0.00034222668004012035, + "loss": 1.4447, + "theoretical_loss": 3.473184162407421, + "tokens_seen": 1902968832 + }, + { + "epoch": 0.58, + "learning_rate": 0.00034221063189568705, + "loss": 1.3402, + "theoretical_loss": 3.473174174361084, + "tokens_seen": 1903034368 + }, + { + "epoch": 0.58, + "learning_rate": 0.00034219458375125375, + "loss": 1.2414, + "theoretical_loss": 3.473164186755012, + "tokens_seen": 1903099904 + }, + { + "epoch": 0.58, + "learning_rate": 0.00034217853560682045, + "loss": 1.3925, + "theoretical_loss": 3.47315419958917, + "tokens_seen": 1903165440 + }, + { + "epoch": 0.58, + "learning_rate": 0.00034216248746238715, + "loss": 1.2978, + "theoretical_loss": 3.4731442128635246, + "tokens_seen": 1903230976 + }, + { + "epoch": 0.58, + "learning_rate": 0.00034214643931795386, + "loss": 1.5025, + "theoretical_loss": 3.4731342265780407, + "tokens_seen": 1903296512 + }, + { + "epoch": 0.58, + "learning_rate": 0.00034213039117352056, + "loss": 1.4197, + "theoretical_loss": 3.473124240732683, + "tokens_seen": 1903362048 + }, + { + "epoch": 0.58, + "learning_rate": 0.00034211434302908726, + "loss": 1.4438, + "theoretical_loss": 3.473114255327418, + "tokens_seen": 1903427584 + }, + { + "epoch": 0.58, + "learning_rate": 0.00034209829488465396, + "loss": 1.2652, + "theoretical_loss": 3.473104270362211, + "tokens_seen": 1903493120 + }, + { + "epoch": 0.58, + "learning_rate": 0.00034208224674022066, + "loss": 1.44, + "theoretical_loss": 3.4730942858370266, + "tokens_seen": 1903558656 + }, + { + "epoch": 0.58, + "learning_rate": 0.00034206619859578736, + "loss": 1.3593, + "theoretical_loss": 3.473084301751831, + "tokens_seen": 1903624192 + }, + { + "epoch": 0.58, + "learning_rate": 0.00034205015045135406, + "loss": 1.2975, + "theoretical_loss": 3.47307431810659, + "tokens_seen": 1903689728 + }, + { + "epoch": 0.58, + "learning_rate": 0.00034203410230692076, + "loss": 1.3784, + "theoretical_loss": 3.4730643349012684, + "tokens_seen": 1903755264 + }, + { + "epoch": 0.58, + "objective/train/docs_used": 694210, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.3429018259048462, + "objective/train/theoretical_loss": 3.4730543521358315, + "objective/train/tokens_used": 1924280800, + "theoretical_loss": 3.4730543521358315, + "tokens_seen": 1903820800 + }, + { + "epoch": 0.58, + "learning_rate": 0.00034201805416248746, + "loss": 1.3562, + "theoretical_loss": 3.4730543521358315, + "tokens_seen": 1903820800 + }, + { + "epoch": 0.58, + "learning_rate": 0.00034200200601805416, + "loss": 1.3932, + "theoretical_loss": 3.4730443698102453, + "tokens_seen": 1903886336 + }, + { + "epoch": 0.58, + "learning_rate": 0.00034198595787362086, + "loss": 1.3493, + "theoretical_loss": 3.473034387924475, + "tokens_seen": 1903951872 + }, + { + "epoch": 0.58, + "learning_rate": 0.00034196990972918756, + "loss": 1.2341, + "theoretical_loss": 3.473024406478487, + "tokens_seen": 1904017408 + }, + { + "epoch": 0.58, + "learning_rate": 0.00034195386158475426, + "loss": 1.3648, + "theoretical_loss": 3.473014425472245, + "tokens_seen": 1904082944 + }, + { + "epoch": 0.58, + "learning_rate": 0.00034193781344032096, + "loss": 1.3376, + "theoretical_loss": 3.4730044449057162, + "tokens_seen": 1904148480 + }, + { + "epoch": 0.58, + "learning_rate": 0.00034192176529588766, + "loss": 1.4038, + "theoretical_loss": 3.472994464778865, + "tokens_seen": 1904214016 + }, + { + "epoch": 0.58, + "learning_rate": 0.00034190571715145436, + "loss": 1.37, + "theoretical_loss": 3.472984485091657, + "tokens_seen": 1904279552 + }, + { + "epoch": 0.58, + "learning_rate": 0.00034188966900702107, + "loss": 1.3966, + "theoretical_loss": 3.4729745058440584, + "tokens_seen": 1904345088 + }, + { + "epoch": 0.58, + "learning_rate": 0.00034187362086258777, + "loss": 1.2782, + "theoretical_loss": 3.4729645270360345, + "tokens_seen": 1904410624 + }, + { + "epoch": 0.58, + "learning_rate": 0.00034185757271815447, + "loss": 1.3519, + "theoretical_loss": 3.47295454866755, + "tokens_seen": 1904476160 + }, + { + "epoch": 0.58, + "learning_rate": 0.00034184152457372117, + "loss": 1.297, + "theoretical_loss": 3.472944570738571, + "tokens_seen": 1904541696 + }, + { + "epoch": 0.58, + "learning_rate": 0.00034182547642928787, + "loss": 1.3716, + "theoretical_loss": 3.4729345932490636, + "tokens_seen": 1904607232 + }, + { + "epoch": 0.58, + "learning_rate": 0.00034180942828485457, + "loss": 1.4422, + "theoretical_loss": 3.472924616198993, + "tokens_seen": 1904672768 + }, + { + "epoch": 0.58, + "learning_rate": 0.00034179338014042127, + "loss": 1.2918, + "theoretical_loss": 3.472914639588324, + "tokens_seen": 1904738304 + }, + { + "epoch": 0.58, + "learning_rate": 0.00034177733199598797, + "loss": 1.3265, + "theoretical_loss": 3.4729046634170224, + "tokens_seen": 1904803840 + }, + { + "epoch": 0.58, + "learning_rate": 0.00034176128385155467, + "loss": 1.4272, + "theoretical_loss": 3.4728946876850544, + "tokens_seen": 1904869376 + }, + { + "epoch": 0.58, + "learning_rate": 0.00034174523570712137, + "loss": 1.4003, + "theoretical_loss": 3.4728847123923847, + "tokens_seen": 1904934912 + }, + { + "epoch": 0.58, + "learning_rate": 0.0003417291875626881, + "loss": 1.2483, + "theoretical_loss": 3.4728747375389792, + "tokens_seen": 1905000448 + }, + { + "epoch": 0.58, + "learning_rate": 0.0003417131394182548, + "loss": 1.4021, + "theoretical_loss": 3.4728647631248037, + "tokens_seen": 1905065984 + }, + { + "epoch": 0.58, + "learning_rate": 0.00034169709127382153, + "loss": 1.323, + "theoretical_loss": 3.4728547891498236, + "tokens_seen": 1905131520 + }, + { + "epoch": 0.58, + "learning_rate": 0.00034168104312938823, + "loss": 1.4082, + "theoretical_loss": 3.4728448156140046, + "tokens_seen": 1905197056 + }, + { + "epoch": 0.58, + "learning_rate": 0.0003416649949849549, + "loss": 1.3286, + "theoretical_loss": 3.4728348425173117, + "tokens_seen": 1905262592 + }, + { + "epoch": 0.58, + "learning_rate": 0.0003416489468405216, + "loss": 1.309, + "theoretical_loss": 3.4728248698597106, + "tokens_seen": 1905328128 + }, + { + "epoch": 0.58, + "learning_rate": 0.0003416328986960883, + "loss": 1.34, + "theoretical_loss": 3.4728148976411677, + "tokens_seen": 1905393664 + }, + { + "epoch": 0.58, + "learning_rate": 0.000341616850551655, + "loss": 1.4107, + "theoretical_loss": 3.472804925861648, + "tokens_seen": 1905459200 + }, + { + "epoch": 0.58, + "learning_rate": 0.0003416008024072217, + "loss": 1.381, + "theoretical_loss": 3.472794954521116, + "tokens_seen": 1905524736 + }, + { + "epoch": 0.58, + "learning_rate": 0.0003415847542627884, + "loss": 1.41, + "theoretical_loss": 3.472784983619539, + "tokens_seen": 1905590272 + }, + { + "epoch": 0.58, + "learning_rate": 0.0003415687061183551, + "loss": 1.3178, + "theoretical_loss": 3.4727750131568818, + "tokens_seen": 1905655808 + }, + { + "epoch": 0.58, + "learning_rate": 0.0003415526579739218, + "loss": 1.3517, + "theoretical_loss": 3.47276504313311, + "tokens_seen": 1905721344 + }, + { + "epoch": 0.58, + "learning_rate": 0.0003415366098294885, + "loss": 1.449, + "theoretical_loss": 3.472755073548189, + "tokens_seen": 1905786880 + }, + { + "epoch": 0.58, + "learning_rate": 0.0003415205616850552, + "loss": 1.3536, + "theoretical_loss": 3.472745104402085, + "tokens_seen": 1905852416 + }, + { + "epoch": 0.58, + "learning_rate": 0.0003415045135406219, + "loss": 1.3432, + "theoretical_loss": 3.472735135694763, + "tokens_seen": 1905917952 + }, + { + "epoch": 0.58, + "learning_rate": 0.0003414884653961886, + "loss": 1.2777, + "theoretical_loss": 3.472725167426189, + "tokens_seen": 1905983488 + }, + { + "epoch": 0.58, + "learning_rate": 0.0003414724172517553, + "loss": 1.2986, + "theoretical_loss": 3.472715199596329, + "tokens_seen": 1906049024 + }, + { + "epoch": 0.58, + "learning_rate": 0.000341456369107322, + "loss": 1.4462, + "theoretical_loss": 3.472705232205147, + "tokens_seen": 1906114560 + }, + { + "epoch": 0.58, + "learning_rate": 0.0003414403209628887, + "loss": 1.3913, + "theoretical_loss": 3.4726952652526104, + "tokens_seen": 1906180096 + }, + { + "epoch": 0.58, + "learning_rate": 0.0003414242728184554, + "loss": 1.3897, + "theoretical_loss": 3.472685298738684, + "tokens_seen": 1906245632 + }, + { + "epoch": 0.58, + "learning_rate": 0.0003414082246740221, + "loss": 1.3804, + "theoretical_loss": 3.472675332663333, + "tokens_seen": 1906311168 + }, + { + "epoch": 0.58, + "learning_rate": 0.0003413921765295888, + "loss": 1.3753, + "theoretical_loss": 3.472665367026524, + "tokens_seen": 1906376704 + }, + { + "epoch": 0.58, + "learning_rate": 0.0003413761283851555, + "loss": 1.2595, + "theoretical_loss": 3.472655401828222, + "tokens_seen": 1906442240 + }, + { + "epoch": 0.58, + "learning_rate": 0.0003413600802407222, + "loss": 1.2565, + "theoretical_loss": 3.4726454370683926, + "tokens_seen": 1906507776 + }, + { + "epoch": 0.58, + "learning_rate": 0.0003413440320962889, + "loss": 1.3371, + "theoretical_loss": 3.472635472747002, + "tokens_seen": 1906573312 + }, + { + "epoch": 0.58, + "learning_rate": 0.0003413279839518556, + "loss": 1.4592, + "theoretical_loss": 3.4726255088640148, + "tokens_seen": 1906638848 + }, + { + "epoch": 0.58, + "learning_rate": 0.0003413119358074223, + "loss": 1.2761, + "theoretical_loss": 3.472615545419398, + "tokens_seen": 1906704384 + }, + { + "epoch": 0.58, + "learning_rate": 0.000341295887662989, + "loss": 1.2887, + "theoretical_loss": 3.472605582413116, + "tokens_seen": 1906769920 + }, + { + "epoch": 0.58, + "learning_rate": 0.0003412798395185557, + "loss": 1.4174, + "theoretical_loss": 3.4725956198451353, + "tokens_seen": 1906835456 + }, + { + "epoch": 0.58, + "learning_rate": 0.0003412637913741224, + "loss": 1.419, + "theoretical_loss": 3.4725856577154213, + "tokens_seen": 1906900992 + }, + { + "epoch": 0.58, + "learning_rate": 0.0003412477432296891, + "loss": 1.3333, + "theoretical_loss": 3.47257569602394, + "tokens_seen": 1906966528 + }, + { + "epoch": 0.58, + "learning_rate": 0.0003412316950852558, + "loss": 1.3449, + "theoretical_loss": 3.4725657347706558, + "tokens_seen": 1907032064 + }, + { + "epoch": 0.58, + "objective/train/docs_used": 695522, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.3173519372940063, + "objective/train/theoretical_loss": 3.4725557739555355, + "objective/train/tokens_used": 1927557600, + "theoretical_loss": 3.4725557739555355, + "tokens_seen": 1907097600 + }, + { + "epoch": 0.58, + "learning_rate": 0.0003412156469408225, + "loss": 1.3111, + "theoretical_loss": 3.4725557739555355, + "tokens_seen": 1907097600 + }, + { + "epoch": 0.58, + "learning_rate": 0.0003411995987963892, + "loss": 1.3029, + "theoretical_loss": 3.4725458135785447, + "tokens_seen": 1907163136 + }, + { + "epoch": 0.58, + "learning_rate": 0.0003411835506519559, + "loss": 1.2509, + "theoretical_loss": 3.4725358536396485, + "tokens_seen": 1907228672 + }, + { + "epoch": 0.58, + "learning_rate": 0.0003411675025075226, + "loss": 1.4025, + "theoretical_loss": 3.4725258941388137, + "tokens_seen": 1907294208 + }, + { + "epoch": 0.58, + "learning_rate": 0.0003411514543630893, + "loss": 1.2963, + "theoretical_loss": 3.4725159350760046, + "tokens_seen": 1907359744 + }, + { + "epoch": 0.58, + "learning_rate": 0.000341135406218656, + "loss": 1.3209, + "theoretical_loss": 3.472505976451188, + "tokens_seen": 1907425280 + }, + { + "epoch": 0.58, + "learning_rate": 0.0003411193580742227, + "loss": 1.2621, + "theoretical_loss": 3.472496018264329, + "tokens_seen": 1907490816 + }, + { + "epoch": 0.58, + "learning_rate": 0.0003411033099297894, + "loss": 1.3686, + "theoretical_loss": 3.4724860605153935, + "tokens_seen": 1907556352 + }, + { + "epoch": 0.58, + "learning_rate": 0.0003410872617853561, + "loss": 1.2712, + "theoretical_loss": 3.4724761032043467, + "tokens_seen": 1907621888 + }, + { + "epoch": 0.58, + "learning_rate": 0.0003410712136409228, + "loss": 1.3619, + "theoretical_loss": 3.472466146331155, + "tokens_seen": 1907687424 + }, + { + "epoch": 0.58, + "learning_rate": 0.0003410551654964895, + "loss": 1.355, + "theoretical_loss": 3.4724561898957838, + "tokens_seen": 1907752960 + }, + { + "epoch": 0.58, + "learning_rate": 0.0003410391173520562, + "loss": 1.2468, + "theoretical_loss": 3.4724462338981983, + "tokens_seen": 1907818496 + }, + { + "epoch": 0.58, + "learning_rate": 0.0003410230692076229, + "loss": 1.3708, + "theoretical_loss": 3.4724362783383658, + "tokens_seen": 1907884032 + }, + { + "epoch": 0.58, + "learning_rate": 0.0003410070210631896, + "loss": 1.4158, + "theoretical_loss": 3.47242632321625, + "tokens_seen": 1907949568 + }, + { + "epoch": 0.58, + "learning_rate": 0.0003409909729187563, + "loss": 1.2552, + "theoretical_loss": 3.4724163685318183, + "tokens_seen": 1908015104 + }, + { + "epoch": 0.58, + "learning_rate": 0.000340974924774323, + "loss": 1.3529, + "theoretical_loss": 3.472406414285035, + "tokens_seen": 1908080640 + }, + { + "epoch": 0.58, + "learning_rate": 0.0003409588766298897, + "loss": 1.3222, + "theoretical_loss": 3.472396460475867, + "tokens_seen": 1908146176 + }, + { + "epoch": 0.58, + "learning_rate": 0.0003409428284854564, + "loss": 1.34, + "theoretical_loss": 3.4723865071042797, + "tokens_seen": 1908211712 + }, + { + "epoch": 0.58, + "learning_rate": 0.0003409267803410231, + "loss": 1.2877, + "theoretical_loss": 3.4723765541702383, + "tokens_seen": 1908277248 + }, + { + "epoch": 0.58, + "learning_rate": 0.0003409107321965898, + "loss": 1.4248, + "theoretical_loss": 3.472366601673709, + "tokens_seen": 1908342784 + }, + { + "epoch": 0.58, + "learning_rate": 0.00034089468405215645, + "loss": 1.2505, + "theoretical_loss": 3.4723566496146576, + "tokens_seen": 1908408320 + }, + { + "epoch": 0.58, + "learning_rate": 0.00034087863590772315, + "loss": 1.3646, + "theoretical_loss": 3.47234669799305, + "tokens_seen": 1908473856 + }, + { + "epoch": 0.58, + "learning_rate": 0.00034086258776328985, + "loss": 1.3418, + "theoretical_loss": 3.4723367468088515, + "tokens_seen": 1908539392 + }, + { + "epoch": 0.58, + "learning_rate": 0.00034084653961885655, + "loss": 1.3022, + "theoretical_loss": 3.4723267960620277, + "tokens_seen": 1908604928 + }, + { + "epoch": 0.58, + "learning_rate": 0.00034083049147442325, + "loss": 1.4282, + "theoretical_loss": 3.4723168457525446, + "tokens_seen": 1908670464 + }, + { + "epoch": 0.58, + "learning_rate": 0.00034081444332998995, + "loss": 1.2353, + "theoretical_loss": 3.4723068958803687, + "tokens_seen": 1908736000 + }, + { + "epoch": 0.58, + "learning_rate": 0.00034079839518555665, + "loss": 1.3331, + "theoretical_loss": 3.4722969464454647, + "tokens_seen": 1908801536 + }, + { + "epoch": 0.58, + "learning_rate": 0.00034078234704112335, + "loss": 1.4451, + "theoretical_loss": 3.472286997447799, + "tokens_seen": 1908867072 + }, + { + "epoch": 0.58, + "learning_rate": 0.00034076629889669006, + "loss": 1.3365, + "theoretical_loss": 3.472277048887337, + "tokens_seen": 1908932608 + }, + { + "epoch": 0.58, + "learning_rate": 0.00034075025075225676, + "loss": 1.3869, + "theoretical_loss": 3.4722671007640447, + "tokens_seen": 1908998144 + }, + { + "epoch": 0.58, + "learning_rate": 0.00034073420260782346, + "loss": 1.3467, + "theoretical_loss": 3.472257153077888, + "tokens_seen": 1909063680 + }, + { + "epoch": 0.58, + "learning_rate": 0.00034071815446339016, + "loss": 1.2811, + "theoretical_loss": 3.4722472058288325, + "tokens_seen": 1909129216 + }, + { + "epoch": 0.58, + "learning_rate": 0.00034070210631895686, + "loss": 1.3987, + "theoretical_loss": 3.472237259016844, + "tokens_seen": 1909194752 + }, + { + "epoch": 0.58, + "learning_rate": 0.00034068605817452356, + "loss": 1.2953, + "theoretical_loss": 3.472227312641888, + "tokens_seen": 1909260288 + }, + { + "epoch": 0.58, + "learning_rate": 0.00034067001003009026, + "loss": 1.4001, + "theoretical_loss": 3.472217366703931, + "tokens_seen": 1909325824 + }, + { + "epoch": 0.58, + "learning_rate": 0.00034065396188565696, + "loss": 1.3926, + "theoretical_loss": 3.4722074212029383, + "tokens_seen": 1909391360 + }, + { + "epoch": 0.58, + "learning_rate": 0.00034063791374122366, + "loss": 1.4184, + "theoretical_loss": 3.4721974761388754, + "tokens_seen": 1909456896 + }, + { + "epoch": 0.58, + "learning_rate": 0.00034062186559679036, + "loss": 1.4603, + "theoretical_loss": 3.4721875315117092, + "tokens_seen": 1909522432 + }, + { + "epoch": 0.58, + "learning_rate": 0.00034060581745235706, + "loss": 1.3182, + "theoretical_loss": 3.4721775873214042, + "tokens_seen": 1909587968 + }, + { + "epoch": 0.58, + "learning_rate": 0.00034058976930792376, + "loss": 1.3427, + "theoretical_loss": 3.472167643567927, + "tokens_seen": 1909653504 + }, + { + "epoch": 0.58, + "learning_rate": 0.00034057372116349046, + "loss": 1.3369, + "theoretical_loss": 3.472157700251244, + "tokens_seen": 1909719040 + }, + { + "epoch": 0.58, + "learning_rate": 0.00034055767301905716, + "loss": 1.4083, + "theoretical_loss": 3.472147757371319, + "tokens_seen": 1909784576 + }, + { + "epoch": 0.58, + "learning_rate": 0.00034054162487462386, + "loss": 1.2761, + "theoretical_loss": 3.47213781492812, + "tokens_seen": 1909850112 + }, + { + "epoch": 0.58, + "learning_rate": 0.0003405255767301906, + "loss": 1.3989, + "theoretical_loss": 3.472127872921612, + "tokens_seen": 1909915648 + }, + { + "epoch": 0.58, + "learning_rate": 0.0003405095285857573, + "loss": 1.4129, + "theoretical_loss": 3.4721179313517605, + "tokens_seen": 1909981184 + }, + { + "epoch": 0.58, + "learning_rate": 0.000340493480441324, + "loss": 1.3302, + "theoretical_loss": 3.4721079902185314, + "tokens_seen": 1910046720 + }, + { + "epoch": 0.58, + "learning_rate": 0.0003404774322968907, + "loss": 1.264, + "theoretical_loss": 3.4720980495218914, + "tokens_seen": 1910112256 + }, + { + "epoch": 0.58, + "learning_rate": 0.0003404613841524574, + "loss": 1.2996, + "theoretical_loss": 3.4720881092618052, + "tokens_seen": 1910177792 + }, + { + "epoch": 0.58, + "learning_rate": 0.0003404453360080241, + "loss": 1.4011, + "theoretical_loss": 3.472078169438239, + "tokens_seen": 1910243328 + }, + { + "epoch": 0.58, + "learning_rate": 0.0003404292878635908, + "loss": 1.3293, + "theoretical_loss": 3.4720682300511596, + "tokens_seen": 1910308864 + }, + { + "epoch": 0.58, + "objective/train/docs_used": 696705, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.2611422538757324, + "objective/train/theoretical_loss": 3.4720582911005318, + "objective/train/tokens_used": 1930834400, + "theoretical_loss": 3.4720582911005318, + "tokens_seen": 1910374400 + }, + { + "epoch": 0.58, + "learning_rate": 0.0003404132397191575, + "loss": 1.2884, + "theoretical_loss": 3.4720582911005318, + "tokens_seen": 1910374400 + }, + { + "epoch": 0.58, + "learning_rate": 0.0003403971915747242, + "loss": 1.3447, + "theoretical_loss": 3.4720483525863215, + "tokens_seen": 1910439936 + }, + { + "epoch": 0.58, + "learning_rate": 0.0003403811434302909, + "loss": 1.3726, + "theoretical_loss": 3.472038414508495, + "tokens_seen": 1910505472 + }, + { + "epoch": 0.58, + "learning_rate": 0.0003403650952858576, + "loss": 1.3774, + "theoretical_loss": 3.4720284768670178, + "tokens_seen": 1910571008 + }, + { + "epoch": 0.58, + "learning_rate": 0.0003403490471414243, + "loss": 1.2885, + "theoretical_loss": 3.4720185396618564, + "tokens_seen": 1910636544 + }, + { + "epoch": 0.58, + "learning_rate": 0.000340332998996991, + "loss": 1.2255, + "theoretical_loss": 3.472008602892976, + "tokens_seen": 1910702080 + }, + { + "epoch": 0.58, + "learning_rate": 0.00034031695085255773, + "loss": 1.3124, + "theoretical_loss": 3.4719986665603426, + "tokens_seen": 1910767616 + }, + { + "epoch": 0.58, + "learning_rate": 0.00034030090270812443, + "loss": 1.4063, + "theoretical_loss": 3.471988730663922, + "tokens_seen": 1910833152 + }, + { + "epoch": 0.58, + "learning_rate": 0.00034028485456369113, + "loss": 1.474, + "theoretical_loss": 3.4719787952036807, + "tokens_seen": 1910898688 + }, + { + "epoch": 0.58, + "learning_rate": 0.0003402688064192578, + "loss": 1.3582, + "theoretical_loss": 3.471968860179584, + "tokens_seen": 1910964224 + }, + { + "epoch": 0.58, + "learning_rate": 0.0003402527582748245, + "loss": 1.3884, + "theoretical_loss": 3.4719589255915984, + "tokens_seen": 1911029760 + }, + { + "epoch": 0.58, + "learning_rate": 0.0003402367101303912, + "loss": 1.4563, + "theoretical_loss": 3.471948991439689, + "tokens_seen": 1911095296 + }, + { + "epoch": 0.58, + "learning_rate": 0.0003402206619859579, + "loss": 1.397, + "theoretical_loss": 3.471939057723822, + "tokens_seen": 1911160832 + }, + { + "epoch": 0.58, + "learning_rate": 0.0003402046138415246, + "loss": 1.3309, + "theoretical_loss": 3.471929124443964, + "tokens_seen": 1911226368 + }, + { + "epoch": 0.58, + "learning_rate": 0.0003401885656970913, + "loss": 1.3931, + "theoretical_loss": 3.4719191916000796, + "tokens_seen": 1911291904 + }, + { + "epoch": 0.58, + "learning_rate": 0.000340172517552658, + "loss": 1.3864, + "theoretical_loss": 3.471909259192136, + "tokens_seen": 1911357440 + }, + { + "epoch": 0.58, + "learning_rate": 0.0003401564694082247, + "loss": 1.349, + "theoretical_loss": 3.471899327220098, + "tokens_seen": 1911422976 + }, + { + "epoch": 0.58, + "learning_rate": 0.0003401404212637914, + "loss": 1.3912, + "theoretical_loss": 3.4718893956839327, + "tokens_seen": 1911488512 + }, + { + "epoch": 0.58, + "learning_rate": 0.0003401243731193581, + "loss": 1.3041, + "theoretical_loss": 3.471879464583605, + "tokens_seen": 1911554048 + }, + { + "epoch": 0.58, + "learning_rate": 0.0003401083249749248, + "loss": 1.3849, + "theoretical_loss": 3.4718695339190813, + "tokens_seen": 1911619584 + }, + { + "epoch": 0.58, + "learning_rate": 0.0003400922768304915, + "loss": 1.2206, + "theoretical_loss": 3.4718596036903273, + "tokens_seen": 1911685120 + }, + { + "epoch": 0.58, + "learning_rate": 0.0003400762286860582, + "loss": 1.3363, + "theoretical_loss": 3.4718496738973097, + "tokens_seen": 1911750656 + }, + { + "epoch": 0.58, + "learning_rate": 0.0003400601805416249, + "loss": 1.3074, + "theoretical_loss": 3.4718397445399933, + "tokens_seen": 1911816192 + }, + { + "epoch": 0.58, + "learning_rate": 0.0003400441323971916, + "loss": 1.3245, + "theoretical_loss": 3.471829815618345, + "tokens_seen": 1911881728 + }, + { + "epoch": 0.58, + "learning_rate": 0.0003400280842527583, + "loss": 1.3399, + "theoretical_loss": 3.4718198871323303, + "tokens_seen": 1911947264 + }, + { + "epoch": 0.58, + "learning_rate": 0.000340012036108325, + "loss": 1.3358, + "theoretical_loss": 3.471809959081915, + "tokens_seen": 1912012800 + }, + { + "epoch": 0.58, + "learning_rate": 0.0003399959879638917, + "loss": 1.5352, + "theoretical_loss": 3.4718000314670654, + "tokens_seen": 1912078336 + }, + { + "epoch": 0.58, + "learning_rate": 0.0003399799398194584, + "loss": 1.2768, + "theoretical_loss": 3.4717901042877477, + "tokens_seen": 1912143872 + }, + { + "epoch": 0.58, + "learning_rate": 0.0003399638916750251, + "loss": 1.2903, + "theoretical_loss": 3.471780177543927, + "tokens_seen": 1912209408 + }, + { + "epoch": 0.58, + "learning_rate": 0.0003399478435305918, + "loss": 1.4377, + "theoretical_loss": 3.47177025123557, + "tokens_seen": 1912274944 + }, + { + "epoch": 0.58, + "learning_rate": 0.0003399317953861585, + "loss": 1.4574, + "theoretical_loss": 3.4717603253626423, + "tokens_seen": 1912340480 + }, + { + "epoch": 0.58, + "learning_rate": 0.0003399157472417252, + "loss": 1.2975, + "theoretical_loss": 3.47175039992511, + "tokens_seen": 1912406016 + }, + { + "epoch": 0.58, + "learning_rate": 0.0003398996990972919, + "loss": 1.4819, + "theoretical_loss": 3.4717404749229397, + "tokens_seen": 1912471552 + }, + { + "epoch": 0.58, + "learning_rate": 0.0003398836509528586, + "loss": 1.315, + "theoretical_loss": 3.471730550356096, + "tokens_seen": 1912537088 + }, + { + "epoch": 0.58, + "learning_rate": 0.0003398676028084253, + "loss": 1.3597, + "theoretical_loss": 3.4717206262245464, + "tokens_seen": 1912602624 + }, + { + "epoch": 0.58, + "learning_rate": 0.000339851554663992, + "loss": 1.2333, + "theoretical_loss": 3.471710702528256, + "tokens_seen": 1912668160 + }, + { + "epoch": 0.58, + "learning_rate": 0.0003398355065195587, + "loss": 1.4, + "theoretical_loss": 3.4717007792671906, + "tokens_seen": 1912733696 + }, + { + "epoch": 0.58, + "learning_rate": 0.0003398194583751254, + "loss": 1.3639, + "theoretical_loss": 3.4716908564413167, + "tokens_seen": 1912799232 + }, + { + "epoch": 0.58, + "learning_rate": 0.0003398034102306921, + "loss": 1.3272, + "theoretical_loss": 3.4716809340506, + "tokens_seen": 1912864768 + }, + { + "epoch": 0.58, + "learning_rate": 0.0003397873620862588, + "loss": 1.3314, + "theoretical_loss": 3.4716710120950074, + "tokens_seen": 1912930304 + }, + { + "epoch": 0.58, + "learning_rate": 0.0003397713139418255, + "loss": 1.2824, + "theoretical_loss": 3.4716610905745036, + "tokens_seen": 1912995840 + }, + { + "epoch": 0.58, + "learning_rate": 0.0003397552657973922, + "loss": 1.414, + "theoretical_loss": 3.4716511694890553, + "tokens_seen": 1913061376 + }, + { + "epoch": 0.58, + "learning_rate": 0.0003397392176529589, + "loss": 1.343, + "theoretical_loss": 3.471641248838629, + "tokens_seen": 1913126912 + }, + { + "epoch": 0.58, + "learning_rate": 0.0003397231695085256, + "loss": 1.3511, + "theoretical_loss": 3.4716313286231895, + "tokens_seen": 1913192448 + }, + { + "epoch": 0.58, + "learning_rate": 0.0003397071213640923, + "loss": 1.1532, + "theoretical_loss": 3.4716214088427035, + "tokens_seen": 1913257984 + }, + { + "epoch": 0.58, + "learning_rate": 0.000339691073219659, + "loss": 1.2781, + "theoretical_loss": 3.4716114894971373, + "tokens_seen": 1913323520 + }, + { + "epoch": 0.58, + "learning_rate": 0.0003396750250752257, + "loss": 1.4015, + "theoretical_loss": 3.471601570586456, + "tokens_seen": 1913389056 + }, + { + "epoch": 0.58, + "learning_rate": 0.0003396589769307924, + "loss": 1.3776, + "theoretical_loss": 3.471591652110627, + "tokens_seen": 1913454592 + }, + { + "epoch": 0.58, + "learning_rate": 0.0003396429287863591, + "loss": 1.4834, + "theoretical_loss": 3.471581734069616, + "tokens_seen": 1913520128 + }, + { + "epoch": 0.58, + "learning_rate": 0.0003396268806419258, + "loss": 1.3485, + "theoretical_loss": 3.471571816463388, + "tokens_seen": 1913585664 + }, + { + "epoch": 0.58, + "objective/train/docs_used": 697951, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.4673137664794922, + "objective/train/theoretical_loss": 3.4715618992919097, + "objective/train/tokens_used": 1934111200, + "theoretical_loss": 3.4715618992919097, + "tokens_seen": 1913651200 + }, + { + "epoch": 0.58, + "learning_rate": 0.0003396108324974925, + "loss": 1.3926, + "theoretical_loss": 3.4715618992919097, + "tokens_seen": 1913651200 + }, + { + "epoch": 0.58, + "learning_rate": 0.0003395947843530592, + "loss": 1.2802, + "theoretical_loss": 3.4715519825551477, + "tokens_seen": 1913716736 + }, + { + "epoch": 0.58, + "learning_rate": 0.0003395787362086259, + "loss": 1.3415, + "theoretical_loss": 3.4715420662530674, + "tokens_seen": 1913782272 + }, + { + "epoch": 0.58, + "learning_rate": 0.0003395626880641926, + "loss": 1.3851, + "theoretical_loss": 3.471532150385635, + "tokens_seen": 1913847808 + }, + { + "epoch": 0.58, + "learning_rate": 0.0003395466399197593, + "loss": 1.3753, + "theoretical_loss": 3.4715222349528165, + "tokens_seen": 1913913344 + }, + { + "epoch": 0.58, + "learning_rate": 0.000339530591775326, + "loss": 1.3494, + "theoretical_loss": 3.471512319954578, + "tokens_seen": 1913978880 + }, + { + "epoch": 0.58, + "learning_rate": 0.0003395145436308927, + "loss": 1.3468, + "theoretical_loss": 3.4715024053908854, + "tokens_seen": 1914044416 + }, + { + "epoch": 0.58, + "learning_rate": 0.00033949849548645935, + "loss": 1.3418, + "theoretical_loss": 3.4714924912617056, + "tokens_seen": 1914109952 + }, + { + "epoch": 0.58, + "learning_rate": 0.00033948244734202605, + "loss": 1.4, + "theoretical_loss": 3.471482577567004, + "tokens_seen": 1914175488 + }, + { + "epoch": 0.58, + "learning_rate": 0.00033946639919759275, + "loss": 1.2682, + "theoretical_loss": 3.4714726643067464, + "tokens_seen": 1914241024 + }, + { + "epoch": 0.58, + "learning_rate": 0.00033945035105315945, + "loss": 1.2447, + "theoretical_loss": 3.4714627514808996, + "tokens_seen": 1914306560 + }, + { + "epoch": 0.58, + "learning_rate": 0.00033943430290872615, + "loss": 1.4181, + "theoretical_loss": 3.4714528390894293, + "tokens_seen": 1914372096 + }, + { + "epoch": 0.58, + "learning_rate": 0.00033941825476429285, + "loss": 1.4393, + "theoretical_loss": 3.4714429271323013, + "tokens_seen": 1914437632 + }, + { + "epoch": 0.58, + "learning_rate": 0.00033940220661985956, + "loss": 1.4539, + "theoretical_loss": 3.471433015609482, + "tokens_seen": 1914503168 + }, + { + "epoch": 0.58, + "learning_rate": 0.00033938615847542626, + "loss": 1.3073, + "theoretical_loss": 3.4714231045209383, + "tokens_seen": 1914568704 + }, + { + "epoch": 0.58, + "learning_rate": 0.00033937011033099296, + "loss": 1.3324, + "theoretical_loss": 3.471413193866635, + "tokens_seen": 1914634240 + }, + { + "epoch": 0.58, + "learning_rate": 0.00033935406218655966, + "loss": 1.2097, + "theoretical_loss": 3.4714032836465387, + "tokens_seen": 1914699776 + }, + { + "epoch": 0.58, + "learning_rate": 0.0003393380140421264, + "loss": 1.4186, + "theoretical_loss": 3.4713933738606157, + "tokens_seen": 1914765312 + }, + { + "epoch": 0.58, + "learning_rate": 0.0003393219658976931, + "loss": 1.3896, + "theoretical_loss": 3.4713834645088317, + "tokens_seen": 1914830848 + }, + { + "epoch": 0.58, + "learning_rate": 0.0003393059177532598, + "loss": 1.3904, + "theoretical_loss": 3.471373555591154, + "tokens_seen": 1914896384 + }, + { + "epoch": 0.58, + "learning_rate": 0.0003392898696088265, + "loss": 1.4219, + "theoretical_loss": 3.4713636471075473, + "tokens_seen": 1914961920 + }, + { + "epoch": 0.58, + "learning_rate": 0.0003392738214643932, + "loss": 1.3638, + "theoretical_loss": 3.471353739057978, + "tokens_seen": 1915027456 + }, + { + "epoch": 0.58, + "learning_rate": 0.0003392577733199599, + "loss": 1.3047, + "theoretical_loss": 3.4713438314424128, + "tokens_seen": 1915092992 + }, + { + "epoch": 0.58, + "learning_rate": 0.0003392417251755266, + "loss": 1.34, + "theoretical_loss": 3.4713339242608177, + "tokens_seen": 1915158528 + }, + { + "epoch": 0.58, + "learning_rate": 0.0003392256770310933, + "loss": 1.4361, + "theoretical_loss": 3.471324017513159, + "tokens_seen": 1915224064 + }, + { + "epoch": 0.58, + "learning_rate": 0.00033920962888666, + "loss": 1.4068, + "theoretical_loss": 3.471314111199402, + "tokens_seen": 1915289600 + }, + { + "epoch": 0.58, + "learning_rate": 0.0003391935807422267, + "loss": 1.3484, + "theoretical_loss": 3.4713042053195133, + "tokens_seen": 1915355136 + }, + { + "epoch": 0.58, + "learning_rate": 0.0003391775325977934, + "loss": 1.3515, + "theoretical_loss": 3.4712942998734597, + "tokens_seen": 1915420672 + }, + { + "epoch": 0.58, + "learning_rate": 0.0003391614844533601, + "loss": 1.3498, + "theoretical_loss": 3.4712843948612067, + "tokens_seen": 1915486208 + }, + { + "epoch": 0.58, + "learning_rate": 0.0003391454363089268, + "loss": 1.3007, + "theoretical_loss": 3.47127449028272, + "tokens_seen": 1915551744 + }, + { + "epoch": 0.58, + "learning_rate": 0.0003391293881644935, + "loss": 1.3772, + "theoretical_loss": 3.4712645861379667, + "tokens_seen": 1915617280 + }, + { + "epoch": 0.58, + "learning_rate": 0.0003391133400200602, + "loss": 1.3078, + "theoretical_loss": 3.4712546824269124, + "tokens_seen": 1915682816 + }, + { + "epoch": 0.58, + "learning_rate": 0.0003390972918756269, + "loss": 1.3949, + "theoretical_loss": 3.471244779149524, + "tokens_seen": 1915748352 + }, + { + "epoch": 0.58, + "learning_rate": 0.0003390812437311936, + "loss": 1.3649, + "theoretical_loss": 3.4712348763057665, + "tokens_seen": 1915813888 + }, + { + "epoch": 0.58, + "learning_rate": 0.0003390651955867603, + "loss": 1.3375, + "theoretical_loss": 3.471224973895607, + "tokens_seen": 1915879424 + }, + { + "epoch": 0.58, + "learning_rate": 0.000339049147442327, + "loss": 1.3166, + "theoretical_loss": 3.471215071919012, + "tokens_seen": 1915944960 + }, + { + "epoch": 0.58, + "learning_rate": 0.0003390330992978937, + "loss": 1.267, + "theoretical_loss": 3.4712051703759466, + "tokens_seen": 1916010496 + }, + { + "epoch": 0.58, + "learning_rate": 0.0003390170511534604, + "loss": 1.1985, + "theoretical_loss": 3.4711952692663774, + "tokens_seen": 1916076032 + }, + { + "epoch": 0.58, + "learning_rate": 0.0003390010030090271, + "loss": 1.3789, + "theoretical_loss": 3.471185368590271, + "tokens_seen": 1916141568 + }, + { + "epoch": 0.58, + "learning_rate": 0.0003389849548645938, + "loss": 1.3347, + "theoretical_loss": 3.4711754683475933, + "tokens_seen": 1916207104 + }, + { + "epoch": 0.58, + "learning_rate": 0.0003389689067201605, + "loss": 1.2849, + "theoretical_loss": 3.47116556853831, + "tokens_seen": 1916272640 + }, + { + "epoch": 0.58, + "learning_rate": 0.00033895285857572723, + "loss": 1.4192, + "theoretical_loss": 3.471155669162388, + "tokens_seen": 1916338176 + }, + { + "epoch": 0.58, + "learning_rate": 0.00033893681043129393, + "loss": 1.3542, + "theoretical_loss": 3.4711457702197936, + "tokens_seen": 1916403712 + }, + { + "epoch": 0.58, + "learning_rate": 0.00033892076228686063, + "loss": 1.3059, + "theoretical_loss": 3.4711358717104925, + "tokens_seen": 1916469248 + }, + { + "epoch": 0.58, + "learning_rate": 0.00033890471414242733, + "loss": 1.3996, + "theoretical_loss": 3.471125973634451, + "tokens_seen": 1916534784 + }, + { + "epoch": 0.58, + "learning_rate": 0.00033888866599799403, + "loss": 1.2064, + "theoretical_loss": 3.4711160759916355, + "tokens_seen": 1916600320 + }, + { + "epoch": 0.58, + "learning_rate": 0.0003388726178535607, + "loss": 1.3075, + "theoretical_loss": 3.471106178782012, + "tokens_seen": 1916665856 + }, + { + "epoch": 0.58, + "learning_rate": 0.0003388565697091274, + "loss": 1.3615, + "theoretical_loss": 3.4710962820055475, + "tokens_seen": 1916731392 + }, + { + "epoch": 0.58, + "learning_rate": 0.0003388405215646941, + "loss": 1.4218, + "theoretical_loss": 3.4710863856622076, + "tokens_seen": 1916796928 + }, + { + "epoch": 0.58, + "learning_rate": 0.0003388244734202608, + "loss": 1.2611, + "theoretical_loss": 3.471076489751958, + "tokens_seen": 1916862464 + }, + { + "epoch": 0.58, + "objective/train/docs_used": 699035, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.2901039123535156, + "objective/train/theoretical_loss": 3.471066594274766, + "objective/train/tokens_used": 1937388000, + "theoretical_loss": 3.471066594274766, + "tokens_seen": 1916928000 + }, + { + "epoch": 0.58, + "learning_rate": 0.0003388084252758275, + "loss": 1.3659, + "theoretical_loss": 3.471066594274766, + "tokens_seen": 1916928000 + }, + { + "epoch": 0.58, + "learning_rate": 0.0003387923771313942, + "loss": 1.3707, + "theoretical_loss": 3.4710566992305973, + "tokens_seen": 1916993536 + }, + { + "epoch": 0.58, + "learning_rate": 0.0003387763289869609, + "loss": 1.3481, + "theoretical_loss": 3.4710468046194185, + "tokens_seen": 1917059072 + }, + { + "epoch": 0.58, + "learning_rate": 0.0003387602808425276, + "loss": 1.4402, + "theoretical_loss": 3.471036910441195, + "tokens_seen": 1917124608 + }, + { + "epoch": 0.58, + "learning_rate": 0.0003387442326980943, + "loss": 1.2857, + "theoretical_loss": 3.471027016695894, + "tokens_seen": 1917190144 + }, + { + "epoch": 0.58, + "learning_rate": 0.000338728184553661, + "loss": 1.2566, + "theoretical_loss": 3.471017123383481, + "tokens_seen": 1917255680 + }, + { + "epoch": 0.58, + "learning_rate": 0.0003387121364092277, + "loss": 1.3419, + "theoretical_loss": 3.4710072305039232, + "tokens_seen": 1917321216 + }, + { + "epoch": 0.58, + "learning_rate": 0.0003386960882647944, + "loss": 1.3463, + "theoretical_loss": 3.470997338057186, + "tokens_seen": 1917386752 + }, + { + "epoch": 0.58, + "learning_rate": 0.0003386800401203611, + "loss": 1.3886, + "theoretical_loss": 3.470987446043236, + "tokens_seen": 1917452288 + }, + { + "epoch": 0.58, + "learning_rate": 0.0003386639919759278, + "loss": 1.3533, + "theoretical_loss": 3.4709775544620394, + "tokens_seen": 1917517824 + }, + { + "epoch": 0.58, + "learning_rate": 0.0003386479438314945, + "loss": 1.3158, + "theoretical_loss": 3.4709676633135627, + "tokens_seen": 1917583360 + }, + { + "epoch": 0.58, + "learning_rate": 0.0003386318956870612, + "loss": 1.3305, + "theoretical_loss": 3.470957772597772, + "tokens_seen": 1917648896 + }, + { + "epoch": 0.58, + "learning_rate": 0.0003386158475426279, + "loss": 1.4654, + "theoretical_loss": 3.4709478823146336, + "tokens_seen": 1917714432 + }, + { + "epoch": 0.58, + "learning_rate": 0.0003385997993981946, + "loss": 1.3605, + "theoretical_loss": 3.4709379924641137, + "tokens_seen": 1917779968 + }, + { + "epoch": 0.58, + "learning_rate": 0.0003385837512537613, + "loss": 1.4537, + "theoretical_loss": 3.470928103046179, + "tokens_seen": 1917845504 + }, + { + "epoch": 0.58, + "learning_rate": 0.000338567703109328, + "loss": 1.3816, + "theoretical_loss": 3.470918214060795, + "tokens_seen": 1917911040 + }, + { + "epoch": 0.58, + "learning_rate": 0.0003385516549648947, + "loss": 1.3312, + "theoretical_loss": 3.4709083255079287, + "tokens_seen": 1917976576 + }, + { + "epoch": 0.58, + "learning_rate": 0.0003385356068204614, + "loss": 1.2876, + "theoretical_loss": 3.4708984373875467, + "tokens_seen": 1918042112 + }, + { + "epoch": 0.58, + "learning_rate": 0.0003385195586760281, + "loss": 1.2439, + "theoretical_loss": 3.4708885496996142, + "tokens_seen": 1918107648 + }, + { + "epoch": 0.58, + "learning_rate": 0.0003385035105315948, + "loss": 1.4324, + "theoretical_loss": 3.4708786624440986, + "tokens_seen": 1918173184 + }, + { + "epoch": 0.58, + "learning_rate": 0.0003384874623871615, + "loss": 1.3283, + "theoretical_loss": 3.470868775620965, + "tokens_seen": 1918238720 + }, + { + "epoch": 0.58, + "learning_rate": 0.0003384714142427282, + "loss": 1.2549, + "theoretical_loss": 3.470858889230181, + "tokens_seen": 1918304256 + }, + { + "epoch": 0.58, + "learning_rate": 0.0003384553660982949, + "loss": 1.4523, + "theoretical_loss": 3.470849003271712, + "tokens_seen": 1918369792 + }, + { + "epoch": 0.58, + "learning_rate": 0.0003384393179538616, + "loss": 1.2934, + "theoretical_loss": 3.470839117745525, + "tokens_seen": 1918435328 + }, + { + "epoch": 0.58, + "learning_rate": 0.0003384232698094283, + "loss": 1.3694, + "theoretical_loss": 3.470829232651586, + "tokens_seen": 1918500864 + }, + { + "epoch": 0.58, + "learning_rate": 0.000338407221664995, + "loss": 1.2799, + "theoretical_loss": 3.4708193479898615, + "tokens_seen": 1918566400 + }, + { + "epoch": 0.58, + "learning_rate": 0.0003383911735205617, + "loss": 1.3137, + "theoretical_loss": 3.4708094637603177, + "tokens_seen": 1918631936 + }, + { + "epoch": 0.58, + "learning_rate": 0.0003383751253761284, + "loss": 1.4018, + "theoretical_loss": 3.4707995799629208, + "tokens_seen": 1918697472 + }, + { + "epoch": 0.58, + "learning_rate": 0.0003383590772316951, + "loss": 1.4, + "theoretical_loss": 3.4707896965976373, + "tokens_seen": 1918763008 + }, + { + "epoch": 0.58, + "learning_rate": 0.0003383430290872618, + "loss": 1.3148, + "theoretical_loss": 3.470779813664434, + "tokens_seen": 1918828544 + }, + { + "epoch": 0.58, + "learning_rate": 0.0003383269809428285, + "loss": 1.3073, + "theoretical_loss": 3.4707699311632765, + "tokens_seen": 1918894080 + }, + { + "epoch": 0.58, + "learning_rate": 0.0003383109327983952, + "loss": 1.3476, + "theoretical_loss": 3.4707600490941313, + "tokens_seen": 1918959616 + }, + { + "epoch": 0.58, + "learning_rate": 0.0003382948846539619, + "loss": 1.3562, + "theoretical_loss": 3.470750167456965, + "tokens_seen": 1919025152 + }, + { + "epoch": 0.58, + "learning_rate": 0.0003382788365095286, + "loss": 1.346, + "theoretical_loss": 3.4707402862517442, + "tokens_seen": 1919090688 + }, + { + "epoch": 0.58, + "learning_rate": 0.0003382627883650953, + "loss": 1.2851, + "theoretical_loss": 3.4707304054784345, + "tokens_seen": 1919156224 + }, + { + "epoch": 0.58, + "learning_rate": 0.000338246740220662, + "loss": 1.2703, + "theoretical_loss": 3.470720525137003, + "tokens_seen": 1919221760 + }, + { + "epoch": 0.58, + "learning_rate": 0.0003382306920762287, + "loss": 1.3079, + "theoretical_loss": 3.4707106452274155, + "tokens_seen": 1919287296 + }, + { + "epoch": 0.58, + "learning_rate": 0.0003382146439317954, + "loss": 1.3328, + "theoretical_loss": 3.4707007657496387, + "tokens_seen": 1919352832 + }, + { + "epoch": 0.58, + "learning_rate": 0.0003381985957873621, + "loss": 1.3132, + "theoretical_loss": 3.4706908867036397, + "tokens_seen": 1919418368 + }, + { + "epoch": 0.58, + "learning_rate": 0.0003381825476429288, + "loss": 1.3031, + "theoretical_loss": 3.4706810080893833, + "tokens_seen": 1919483904 + }, + { + "epoch": 0.58, + "learning_rate": 0.0003381664994984955, + "loss": 1.4694, + "theoretical_loss": 3.4706711299068376, + "tokens_seen": 1919549440 + }, + { + "epoch": 0.58, + "learning_rate": 0.0003381504513540622, + "loss": 1.3049, + "theoretical_loss": 3.4706612521559674, + "tokens_seen": 1919614976 + }, + { + "epoch": 0.58, + "learning_rate": 0.0003381344032096289, + "loss": 1.355, + "theoretical_loss": 3.47065137483674, + "tokens_seen": 1919680512 + }, + { + "epoch": 0.58, + "learning_rate": 0.0003381183550651956, + "loss": 1.2301, + "theoretical_loss": 3.4706414979491216, + "tokens_seen": 1919746048 + }, + { + "epoch": 0.58, + "learning_rate": 0.0003381023069207623, + "loss": 1.3555, + "theoretical_loss": 3.4706316214930792, + "tokens_seen": 1919811584 + }, + { + "epoch": 0.58, + "learning_rate": 0.000338086258776329, + "loss": 1.3722, + "theoretical_loss": 3.470621745468578, + "tokens_seen": 1919877120 + }, + { + "epoch": 0.58, + "learning_rate": 0.0003380702106318957, + "loss": 1.4008, + "theoretical_loss": 3.4706118698755857, + "tokens_seen": 1919942656 + }, + { + "epoch": 0.58, + "learning_rate": 0.0003380541624874624, + "loss": 1.3543, + "theoretical_loss": 3.470601994714068, + "tokens_seen": 1920008192 + }, + { + "epoch": 0.58, + "learning_rate": 0.0003380381143430291, + "loss": 1.3664, + "theoretical_loss": 3.4705921199839906, + "tokens_seen": 1920073728 + }, + { + "epoch": 0.58, + "learning_rate": 0.0003380220661985958, + "loss": 1.3786, + "theoretical_loss": 3.4705822456853213, + "tokens_seen": 1920139264 + }, + { + "epoch": 0.58, + "objective/train/docs_used": 700307, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.4534680843353271, + "objective/train/theoretical_loss": 3.4705723718180264, + "objective/train/tokens_used": 1940664800, + "theoretical_loss": 3.4705723718180264, + "tokens_seen": 1920204800 + }, + { + "epoch": 0.58, + "learning_rate": 0.0003380060180541625, + "loss": 1.4545, + "theoretical_loss": 3.4705723718180264, + "tokens_seen": 1920204800 + }, + { + "epoch": 0.58, + "learning_rate": 0.0003379899699097292, + "loss": 1.3468, + "theoretical_loss": 3.4705624983820718, + "tokens_seen": 1920270336 + }, + { + "epoch": 0.58, + "learning_rate": 0.0003379739217652959, + "loss": 1.3443, + "theoretical_loss": 3.4705526253774233, + "tokens_seen": 1920335872 + }, + { + "epoch": 0.58, + "learning_rate": 0.0003379578736208626, + "loss": 1.3567, + "theoretical_loss": 3.470542752804049, + "tokens_seen": 1920401408 + }, + { + "epoch": 0.58, + "learning_rate": 0.0003379418254764293, + "loss": 1.3571, + "theoretical_loss": 3.4705328806619136, + "tokens_seen": 1920466944 + }, + { + "epoch": 0.58, + "learning_rate": 0.000337925777331996, + "loss": 1.3733, + "theoretical_loss": 3.470523008950985, + "tokens_seen": 1920532480 + }, + { + "epoch": 0.58, + "learning_rate": 0.0003379097291875627, + "loss": 1.2645, + "theoretical_loss": 3.470513137671229, + "tokens_seen": 1920598016 + }, + { + "epoch": 0.58, + "learning_rate": 0.0003378936810431294, + "loss": 1.2585, + "theoretical_loss": 3.470503266822612, + "tokens_seen": 1920663552 + }, + { + "epoch": 0.58, + "learning_rate": 0.0003378776328986961, + "loss": 1.4414, + "theoretical_loss": 3.4704933964051, + "tokens_seen": 1920729088 + }, + { + "epoch": 0.58, + "learning_rate": 0.0003378615847542628, + "loss": 1.2618, + "theoretical_loss": 3.470483526418661, + "tokens_seen": 1920794624 + }, + { + "epoch": 0.58, + "learning_rate": 0.0003378455366098295, + "loss": 1.2519, + "theoretical_loss": 3.47047365686326, + "tokens_seen": 1920860160 + }, + { + "epoch": 0.58, + "learning_rate": 0.0003378294884653962, + "loss": 1.4044, + "theoretical_loss": 3.470463787738864, + "tokens_seen": 1920925696 + }, + { + "epoch": 0.58, + "learning_rate": 0.0003378134403209629, + "loss": 1.3261, + "theoretical_loss": 3.470453919045439, + "tokens_seen": 1920991232 + }, + { + "epoch": 0.58, + "learning_rate": 0.0003377973921765296, + "loss": 1.3936, + "theoretical_loss": 3.4704440507829526, + "tokens_seen": 1921056768 + }, + { + "epoch": 0.58, + "learning_rate": 0.0003377813440320963, + "loss": 1.1962, + "theoretical_loss": 3.4704341829513696, + "tokens_seen": 1921122304 + }, + { + "epoch": 0.58, + "learning_rate": 0.000337765295887663, + "loss": 1.4176, + "theoretical_loss": 3.4704243155506584, + "tokens_seen": 1921187840 + }, + { + "epoch": 0.58, + "learning_rate": 0.0003377492477432297, + "loss": 1.4423, + "theoretical_loss": 3.470414448580784, + "tokens_seen": 1921253376 + }, + { + "epoch": 0.58, + "learning_rate": 0.0003377331995987964, + "loss": 1.2672, + "theoretical_loss": 3.4704045820417138, + "tokens_seen": 1921318912 + }, + { + "epoch": 0.58, + "learning_rate": 0.0003377171514543631, + "loss": 1.3098, + "theoretical_loss": 3.470394715933414, + "tokens_seen": 1921384448 + }, + { + "epoch": 0.58, + "learning_rate": 0.0003377011033099298, + "loss": 1.2849, + "theoretical_loss": 3.4703848502558508, + "tokens_seen": 1921449984 + }, + { + "epoch": 0.58, + "learning_rate": 0.0003376850551654965, + "loss": 1.374, + "theoretical_loss": 3.4703749850089913, + "tokens_seen": 1921515520 + }, + { + "epoch": 0.58, + "learning_rate": 0.0003376690070210632, + "loss": 1.437, + "theoretical_loss": 3.4703651201928016, + "tokens_seen": 1921581056 + }, + { + "epoch": 0.58, + "learning_rate": 0.0003376529588766299, + "loss": 1.2215, + "theoretical_loss": 3.470355255807248, + "tokens_seen": 1921646592 + }, + { + "epoch": 0.58, + "learning_rate": 0.0003376369107321966, + "loss": 1.3588, + "theoretical_loss": 3.4703453918522973, + "tokens_seen": 1921712128 + }, + { + "epoch": 0.58, + "learning_rate": 0.0003376208625877633, + "loss": 1.2612, + "theoretical_loss": 3.470335528327916, + "tokens_seen": 1921777664 + }, + { + "epoch": 0.58, + "learning_rate": 0.00033760481444333, + "loss": 1.3376, + "theoretical_loss": 3.470325665234071, + "tokens_seen": 1921843200 + }, + { + "epoch": 0.58, + "learning_rate": 0.0003375887662988967, + "loss": 1.3393, + "theoretical_loss": 3.4703158025707284, + "tokens_seen": 1921908736 + }, + { + "epoch": 0.58, + "learning_rate": 0.00033757271815446343, + "loss": 1.3479, + "theoretical_loss": 3.4703059403378544, + "tokens_seen": 1921974272 + }, + { + "epoch": 0.58, + "learning_rate": 0.00033755667001003013, + "loss": 1.3251, + "theoretical_loss": 3.4702960785354158, + "tokens_seen": 1922039808 + }, + { + "epoch": 0.58, + "learning_rate": 0.00033754062186559683, + "loss": 1.3921, + "theoretical_loss": 3.4702862171633795, + "tokens_seen": 1922105344 + }, + { + "epoch": 0.58, + "learning_rate": 0.00033752457372116353, + "loss": 1.3905, + "theoretical_loss": 3.4702763562217123, + "tokens_seen": 1922170880 + }, + { + "epoch": 0.58, + "learning_rate": 0.00033750852557673023, + "loss": 1.2243, + "theoretical_loss": 3.47026649571038, + "tokens_seen": 1922236416 + }, + { + "epoch": 0.58, + "learning_rate": 0.00033749247743229693, + "loss": 1.3167, + "theoretical_loss": 3.470256635629349, + "tokens_seen": 1922301952 + }, + { + "epoch": 0.58, + "learning_rate": 0.00033747642928786363, + "loss": 1.3623, + "theoretical_loss": 3.470246775978586, + "tokens_seen": 1922367488 + }, + { + "epoch": 0.58, + "learning_rate": 0.0003374603811434303, + "loss": 1.376, + "theoretical_loss": 3.4702369167580587, + "tokens_seen": 1922433024 + }, + { + "epoch": 0.58, + "learning_rate": 0.000337444332998997, + "loss": 1.3403, + "theoretical_loss": 3.4702270579677323, + "tokens_seen": 1922498560 + }, + { + "epoch": 0.58, + "learning_rate": 0.0003374282848545637, + "loss": 1.3167, + "theoretical_loss": 3.470217199607574, + "tokens_seen": 1922564096 + }, + { + "epoch": 0.58, + "learning_rate": 0.0003374122367101304, + "loss": 1.3768, + "theoretical_loss": 3.4702073416775496, + "tokens_seen": 1922629632 + }, + { + "epoch": 0.58, + "learning_rate": 0.0003373961885656971, + "loss": 1.4222, + "theoretical_loss": 3.470197484177627, + "tokens_seen": 1922695168 + }, + { + "epoch": 0.58, + "learning_rate": 0.0003373801404212638, + "loss": 1.42, + "theoretical_loss": 3.470187627107772, + "tokens_seen": 1922760704 + }, + { + "epoch": 0.58, + "learning_rate": 0.0003373640922768305, + "loss": 1.3898, + "theoretical_loss": 3.470177770467951, + "tokens_seen": 1922826240 + }, + { + "epoch": 0.58, + "learning_rate": 0.0003373480441323972, + "loss": 1.4084, + "theoretical_loss": 3.4701679142581305, + "tokens_seen": 1922891776 + }, + { + "epoch": 0.58, + "learning_rate": 0.0003373319959879639, + "loss": 1.2614, + "theoretical_loss": 3.470158058478278, + "tokens_seen": 1922957312 + }, + { + "epoch": 0.58, + "learning_rate": 0.0003373159478435306, + "loss": 1.343, + "theoretical_loss": 3.4701482031283595, + "tokens_seen": 1923022848 + }, + { + "epoch": 0.58, + "learning_rate": 0.0003372998996990973, + "loss": 1.3145, + "theoretical_loss": 3.4701383482083408, + "tokens_seen": 1923088384 + }, + { + "epoch": 0.58, + "learning_rate": 0.000337283851554664, + "loss": 1.3211, + "theoretical_loss": 3.4701284937181898, + "tokens_seen": 1923153920 + }, + { + "epoch": 0.58, + "learning_rate": 0.0003372678034102307, + "loss": 1.3898, + "theoretical_loss": 3.4701186396578727, + "tokens_seen": 1923219456 + }, + { + "epoch": 0.58, + "learning_rate": 0.0003372517552657974, + "loss": 1.3527, + "theoretical_loss": 3.470108786027356, + "tokens_seen": 1923284992 + }, + { + "epoch": 0.58, + "learning_rate": 0.0003372357071213641, + "loss": 1.5058, + "theoretical_loss": 3.4700989328266063, + "tokens_seen": 1923350528 + }, + { + "epoch": 0.58, + "learning_rate": 0.0003372196589769308, + "loss": 1.3597, + "theoretical_loss": 3.4700890800555895, + "tokens_seen": 1923416064 + }, + { + "epoch": 0.58, + "objective/train/docs_used": 701572, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.3299509286880493, + "objective/train/theoretical_loss": 3.470079227714274, + "objective/train/tokens_used": 1943941600, + "theoretical_loss": 3.470079227714274, + "tokens_seen": 1923481600 + }, + { + "epoch": 0.58, + "learning_rate": 0.0003372036108324975, + "loss": 1.3413, + "theoretical_loss": 3.470079227714274, + "tokens_seen": 1923481600 + }, + { + "epoch": 0.58, + "learning_rate": 0.0003371875626880642, + "loss": 1.2653, + "theoretical_loss": 3.4700693758026246, + "tokens_seen": 1923547136 + }, + { + "epoch": 0.58, + "learning_rate": 0.0003371715145436309, + "loss": 1.3624, + "theoretical_loss": 3.470059524320609, + "tokens_seen": 1923612672 + }, + { + "epoch": 0.58, + "learning_rate": 0.0003371554663991976, + "loss": 1.3499, + "theoretical_loss": 3.4700496732681936, + "tokens_seen": 1923678208 + }, + { + "epoch": 0.58, + "learning_rate": 0.0003371394182547643, + "loss": 1.3143, + "theoretical_loss": 3.4700398226453446, + "tokens_seen": 1923743744 + }, + { + "epoch": 0.58, + "learning_rate": 0.000337123370110331, + "loss": 1.352, + "theoretical_loss": 3.470029972452029, + "tokens_seen": 1923809280 + }, + { + "epoch": 0.58, + "learning_rate": 0.0003371073219658977, + "loss": 1.5427, + "theoretical_loss": 3.4700201226882132, + "tokens_seen": 1923874816 + }, + { + "epoch": 0.58, + "learning_rate": 0.0003370912738214644, + "loss": 1.2698, + "theoretical_loss": 3.4700102733538642, + "tokens_seen": 1923940352 + }, + { + "epoch": 0.58, + "learning_rate": 0.0003370752256770311, + "loss": 1.3556, + "theoretical_loss": 3.470000424448949, + "tokens_seen": 1924005888 + }, + { + "epoch": 0.58, + "learning_rate": 0.0003370591775325978, + "loss": 1.3817, + "theoretical_loss": 3.4699905759734335, + "tokens_seen": 1924071424 + }, + { + "epoch": 0.58, + "learning_rate": 0.0003370431293881645, + "loss": 1.3093, + "theoretical_loss": 3.4699807279272843, + "tokens_seen": 1924136960 + }, + { + "epoch": 0.58, + "learning_rate": 0.0003370270812437312, + "loss": 1.3627, + "theoretical_loss": 3.4699708803104685, + "tokens_seen": 1924202496 + }, + { + "epoch": 0.58, + "learning_rate": 0.0003370110330992979, + "loss": 1.257, + "theoretical_loss": 3.469961033122953, + "tokens_seen": 1924268032 + }, + { + "epoch": 0.58, + "learning_rate": 0.0003369949849548646, + "loss": 1.2306, + "theoretical_loss": 3.4699511863647032, + "tokens_seen": 1924333568 + }, + { + "epoch": 0.58, + "learning_rate": 0.0003369789368104313, + "loss": 1.3135, + "theoretical_loss": 3.469941340035687, + "tokens_seen": 1924399104 + }, + { + "epoch": 0.58, + "learning_rate": 0.000336962888665998, + "loss": 1.2946, + "theoretical_loss": 3.4699314941358708, + "tokens_seen": 1924464640 + }, + { + "epoch": 0.58, + "learning_rate": 0.0003369468405215647, + "loss": 1.2858, + "theoretical_loss": 3.4699216486652213, + "tokens_seen": 1924530176 + }, + { + "epoch": 0.58, + "learning_rate": 0.00033693079237713145, + "loss": 1.3153, + "theoretical_loss": 3.469911803623705, + "tokens_seen": 1924595712 + }, + { + "epoch": 0.58, + "learning_rate": 0.00033691474423269815, + "loss": 1.4142, + "theoretical_loss": 3.469901959011289, + "tokens_seen": 1924661248 + }, + { + "epoch": 0.58, + "learning_rate": 0.00033689869608826485, + "loss": 1.1549, + "theoretical_loss": 3.4698921148279394, + "tokens_seen": 1924726784 + }, + { + "epoch": 0.58, + "learning_rate": 0.00033688264794383155, + "loss": 1.3527, + "theoretical_loss": 3.4698822710736232, + "tokens_seen": 1924792320 + }, + { + "epoch": 0.58, + "learning_rate": 0.00033686659979939826, + "loss": 1.309, + "theoretical_loss": 3.4698724277483066, + "tokens_seen": 1924857856 + }, + { + "epoch": 0.58, + "learning_rate": 0.00033685055165496496, + "loss": 1.3076, + "theoretical_loss": 3.469862584851957, + "tokens_seen": 1924923392 + }, + { + "epoch": 0.58, + "learning_rate": 0.0003368345035105316, + "loss": 1.3434, + "theoretical_loss": 3.469852742384541, + "tokens_seen": 1924988928 + }, + { + "epoch": 0.58, + "learning_rate": 0.0003368184553660983, + "loss": 1.4105, + "theoretical_loss": 3.469842900346025, + "tokens_seen": 1925054464 + }, + { + "epoch": 0.58, + "learning_rate": 0.000336802407221665, + "loss": 1.3089, + "theoretical_loss": 3.469833058736376, + "tokens_seen": 1925120000 + }, + { + "epoch": 0.58, + "learning_rate": 0.0003367863590772317, + "loss": 1.3686, + "theoretical_loss": 3.4698232175555606, + "tokens_seen": 1925185536 + }, + { + "epoch": 0.58, + "learning_rate": 0.0003367703109327984, + "loss": 1.3191, + "theoretical_loss": 3.469813376803545, + "tokens_seen": 1925251072 + }, + { + "epoch": 0.58, + "learning_rate": 0.0003367542627883651, + "loss": 1.2782, + "theoretical_loss": 3.469803536480297, + "tokens_seen": 1925316608 + }, + { + "epoch": 0.58, + "learning_rate": 0.0003367382146439318, + "loss": 1.3221, + "theoretical_loss": 3.469793696585782, + "tokens_seen": 1925382144 + }, + { + "epoch": 0.58, + "learning_rate": 0.0003367221664994985, + "loss": 1.4073, + "theoretical_loss": 3.4697838571199684, + "tokens_seen": 1925447680 + }, + { + "epoch": 0.58, + "learning_rate": 0.0003367061183550652, + "loss": 1.2729, + "theoretical_loss": 3.469774018082821, + "tokens_seen": 1925513216 + }, + { + "epoch": 0.58, + "learning_rate": 0.0003366900702106319, + "loss": 1.297, + "theoretical_loss": 3.4697641794743084, + "tokens_seen": 1925578752 + }, + { + "epoch": 0.58, + "learning_rate": 0.0003366740220661986, + "loss": 1.2962, + "theoretical_loss": 3.469754341294396, + "tokens_seen": 1925644288 + }, + { + "epoch": 0.58, + "learning_rate": 0.0003366579739217653, + "loss": 1.3058, + "theoretical_loss": 3.469744503543051, + "tokens_seen": 1925709824 + }, + { + "epoch": 0.58, + "learning_rate": 0.000336641925777332, + "loss": 1.3417, + "theoretical_loss": 3.46973466622024, + "tokens_seen": 1925775360 + }, + { + "epoch": 0.58, + "learning_rate": 0.0003366258776328987, + "loss": 1.3663, + "theoretical_loss": 3.4697248293259304, + "tokens_seen": 1925840896 + }, + { + "epoch": 0.58, + "learning_rate": 0.0003366098294884654, + "loss": 1.4096, + "theoretical_loss": 3.469714992860088, + "tokens_seen": 1925906432 + }, + { + "epoch": 0.58, + "learning_rate": 0.0003365937813440321, + "loss": 1.3824, + "theoretical_loss": 3.4697051568226804, + "tokens_seen": 1925971968 + }, + { + "epoch": 0.58, + "learning_rate": 0.0003365777331995988, + "loss": 1.3462, + "theoretical_loss": 3.4696953212136736, + "tokens_seen": 1926037504 + }, + { + "epoch": 0.58, + "learning_rate": 0.0003365616850551655, + "loss": 1.2711, + "theoretical_loss": 3.469685486033035, + "tokens_seen": 1926103040 + }, + { + "epoch": 0.58, + "learning_rate": 0.0003365456369107322, + "loss": 1.3335, + "theoretical_loss": 3.469675651280731, + "tokens_seen": 1926168576 + }, + { + "epoch": 0.58, + "learning_rate": 0.0003365295887662989, + "loss": 1.2787, + "theoretical_loss": 3.4696658169567285, + "tokens_seen": 1926234112 + }, + { + "epoch": 0.58, + "learning_rate": 0.0003365135406218656, + "loss": 1.3535, + "theoretical_loss": 3.469655983060994, + "tokens_seen": 1926299648 + }, + { + "epoch": 0.58, + "learning_rate": 0.0003364974924774323, + "loss": 1.3238, + "theoretical_loss": 3.469646149593495, + "tokens_seen": 1926365184 + }, + { + "epoch": 0.58, + "learning_rate": 0.000336481444332999, + "loss": 1.3082, + "theoretical_loss": 3.4696363165541975, + "tokens_seen": 1926430720 + }, + { + "epoch": 0.58, + "learning_rate": 0.0003364653961885657, + "loss": 1.3447, + "theoretical_loss": 3.4696264839430686, + "tokens_seen": 1926496256 + }, + { + "epoch": 0.58, + "learning_rate": 0.0003364493480441324, + "loss": 1.4328, + "theoretical_loss": 3.4696166517600755, + "tokens_seen": 1926561792 + }, + { + "epoch": 0.58, + "learning_rate": 0.0003364332998996991, + "loss": 1.3341, + "theoretical_loss": 3.4696068200051844, + "tokens_seen": 1926627328 + }, + { + "epoch": 0.58, + "learning_rate": 0.0003364172517552658, + "loss": 1.3937, + "theoretical_loss": 3.469596988678362, + "tokens_seen": 1926692864 + }, + { + "epoch": 0.58, + "objective/train/docs_used": 702709, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.5514845848083496, + "objective/train/theoretical_loss": 3.469587157779576, + "objective/train/tokens_used": 1947218400, + "theoretical_loss": 3.469587157779576, + "tokens_seen": 1926758400 + }, + { + "epoch": 0.58, + "learning_rate": 0.0003364012036108325, + "loss": 1.3889, + "theoretical_loss": 3.469587157779576, + "tokens_seen": 1926758400 + }, + { + "epoch": 0.58, + "learning_rate": 0.0003363851554663992, + "loss": 1.3354, + "theoretical_loss": 3.4695773273087918, + "tokens_seen": 1926823936 + }, + { + "epoch": 0.58, + "learning_rate": 0.0003363691073219659, + "loss": 1.3356, + "theoretical_loss": 3.4695674972659774, + "tokens_seen": 1926889472 + }, + { + "epoch": 0.58, + "learning_rate": 0.0003363530591775326, + "loss": 1.2111, + "theoretical_loss": 3.469557667651099, + "tokens_seen": 1926955008 + }, + { + "epoch": 0.58, + "learning_rate": 0.0003363370110330993, + "loss": 1.354, + "theoretical_loss": 3.4695478384641243, + "tokens_seen": 1927020544 + }, + { + "epoch": 0.58, + "learning_rate": 0.000336320962888666, + "loss": 1.385, + "theoretical_loss": 3.469538009705019, + "tokens_seen": 1927086080 + }, + { + "epoch": 0.58, + "learning_rate": 0.0003363049147442327, + "loss": 1.2473, + "theoretical_loss": 3.46952818137375, + "tokens_seen": 1927151616 + }, + { + "epoch": 0.58, + "learning_rate": 0.0003362888665997994, + "loss": 1.4223, + "theoretical_loss": 3.4695183534702854, + "tokens_seen": 1927217152 + }, + { + "epoch": 0.58, + "learning_rate": 0.0003362728184553661, + "loss": 1.4016, + "theoretical_loss": 3.4695085259945904, + "tokens_seen": 1927282688 + }, + { + "epoch": 0.58, + "learning_rate": 0.0003362567703109328, + "loss": 1.4135, + "theoretical_loss": 3.4694986989466328, + "tokens_seen": 1927348224 + }, + { + "epoch": 0.58, + "learning_rate": 0.0003362407221664995, + "loss": 1.3857, + "theoretical_loss": 3.4694888723263793, + "tokens_seen": 1927413760 + }, + { + "epoch": 0.58, + "learning_rate": 0.0003362246740220662, + "loss": 1.3974, + "theoretical_loss": 3.469479046133797, + "tokens_seen": 1927479296 + }, + { + "epoch": 0.58, + "learning_rate": 0.00033620862587763293, + "loss": 1.3812, + "theoretical_loss": 3.469469220368852, + "tokens_seen": 1927544832 + }, + { + "epoch": 0.58, + "learning_rate": 0.00033619257773319963, + "loss": 1.2497, + "theoretical_loss": 3.4694593950315116, + "tokens_seen": 1927610368 + }, + { + "epoch": 0.58, + "learning_rate": 0.00033617652958876633, + "loss": 1.3703, + "theoretical_loss": 3.4694495701217427, + "tokens_seen": 1927675904 + }, + { + "epoch": 0.58, + "learning_rate": 0.00033616048144433303, + "loss": 1.3101, + "theoretical_loss": 3.4694397456395123, + "tokens_seen": 1927741440 + }, + { + "epoch": 0.58, + "learning_rate": 0.00033614443329989973, + "loss": 1.419, + "theoretical_loss": 3.469429921584787, + "tokens_seen": 1927806976 + }, + { + "epoch": 0.58, + "learning_rate": 0.00033612838515546643, + "loss": 1.332, + "theoretical_loss": 3.469420097957533, + "tokens_seen": 1927872512 + }, + { + "epoch": 0.58, + "learning_rate": 0.00033611233701103313, + "loss": 1.3682, + "theoretical_loss": 3.4694102747577187, + "tokens_seen": 1927938048 + }, + { + "epoch": 0.58, + "learning_rate": 0.00033609628886659983, + "loss": 1.3289, + "theoretical_loss": 3.46940045198531, + "tokens_seen": 1928003584 + }, + { + "epoch": 0.58, + "learning_rate": 0.00033608024072216653, + "loss": 1.2723, + "theoretical_loss": 3.4693906296402734, + "tokens_seen": 1928069120 + }, + { + "epoch": 0.58, + "learning_rate": 0.0003360641925777332, + "loss": 1.3291, + "theoretical_loss": 3.469380807722577, + "tokens_seen": 1928134656 + }, + { + "epoch": 0.58, + "learning_rate": 0.0003360481444332999, + "loss": 1.4693, + "theoretical_loss": 3.469370986232186, + "tokens_seen": 1928200192 + }, + { + "epoch": 0.58, + "learning_rate": 0.0003360320962888666, + "loss": 1.3534, + "theoretical_loss": 3.4693611651690692, + "tokens_seen": 1928265728 + }, + { + "epoch": 0.58, + "learning_rate": 0.0003360160481444333, + "loss": 1.3871, + "theoretical_loss": 3.4693513445331923, + "tokens_seen": 1928331264 + }, + { + "epoch": 0.58, + "learning_rate": 0.000336, + "loss": 1.3404, + "theoretical_loss": 3.4693415243245225, + "tokens_seen": 1928396800 + }, + { + "epoch": 0.58, + "learning_rate": 0.0003359839518555667, + "loss": 1.2778, + "theoretical_loss": 3.4693317045430265, + "tokens_seen": 1928462336 + }, + { + "epoch": 0.58, + "learning_rate": 0.0003359679037111334, + "loss": 1.4152, + "theoretical_loss": 3.4693218851886716, + "tokens_seen": 1928527872 + }, + { + "epoch": 0.58, + "learning_rate": 0.0003359518555667001, + "loss": 1.3101, + "theoretical_loss": 3.4693120662614243, + "tokens_seen": 1928593408 + }, + { + "epoch": 0.58, + "learning_rate": 0.0003359358074222668, + "loss": 1.4656, + "theoretical_loss": 3.4693022477612514, + "tokens_seen": 1928658944 + }, + { + "epoch": 0.58, + "learning_rate": 0.0003359197592778335, + "loss": 1.4135, + "theoretical_loss": 3.4692924296881205, + "tokens_seen": 1928724480 + }, + { + "epoch": 0.58, + "learning_rate": 0.0003359037111334002, + "loss": 1.4184, + "theoretical_loss": 3.469282612041998, + "tokens_seen": 1928790016 + }, + { + "epoch": 0.58, + "learning_rate": 0.0003358876629889669, + "loss": 1.3334, + "theoretical_loss": 3.469272794822851, + "tokens_seen": 1928855552 + }, + { + "epoch": 0.58, + "learning_rate": 0.0003358716148445336, + "loss": 1.4178, + "theoretical_loss": 3.4692629780306463, + "tokens_seen": 1928921088 + }, + { + "epoch": 0.58, + "learning_rate": 0.0003358555667001003, + "loss": 1.3419, + "theoretical_loss": 3.4692531616653506, + "tokens_seen": 1928986624 + }, + { + "epoch": 0.58, + "learning_rate": 0.000335839518555667, + "loss": 1.3458, + "theoretical_loss": 3.469243345726931, + "tokens_seen": 1929052160 + }, + { + "epoch": 0.58, + "learning_rate": 0.0003358234704112337, + "loss": 1.263, + "theoretical_loss": 3.469233530215355, + "tokens_seen": 1929117696 + }, + { + "epoch": 0.58, + "learning_rate": 0.0003358074222668004, + "loss": 1.2528, + "theoretical_loss": 3.4692237151305894, + "tokens_seen": 1929183232 + }, + { + "epoch": 0.58, + "learning_rate": 0.0003357913741223671, + "loss": 1.2691, + "theoretical_loss": 3.4692139004726004, + "tokens_seen": 1929248768 + }, + { + "epoch": 0.58, + "learning_rate": 0.0003357753259779338, + "loss": 1.3525, + "theoretical_loss": 3.469204086241355, + "tokens_seen": 1929314304 + }, + { + "epoch": 0.58, + "learning_rate": 0.0003357592778335005, + "loss": 1.3635, + "theoretical_loss": 3.4691942724368214, + "tokens_seen": 1929379840 + }, + { + "epoch": 0.58, + "learning_rate": 0.0003357432296890672, + "loss": 1.4902, + "theoretical_loss": 3.469184459058965, + "tokens_seen": 1929445376 + }, + { + "epoch": 0.58, + "learning_rate": 0.00033572718154463395, + "loss": 1.4251, + "theoretical_loss": 3.4691746461077537, + "tokens_seen": 1929510912 + }, + { + "epoch": 0.58, + "learning_rate": 0.00033571113340020065, + "loss": 1.297, + "theoretical_loss": 3.4691648335831538, + "tokens_seen": 1929576448 + }, + { + "epoch": 0.58, + "learning_rate": 0.00033569508525576735, + "loss": 1.4163, + "theoretical_loss": 3.4691550214851334, + "tokens_seen": 1929641984 + }, + { + "epoch": 0.58, + "learning_rate": 0.00033567903711133405, + "loss": 1.3329, + "theoretical_loss": 3.469145209813658, + "tokens_seen": 1929707520 + }, + { + "epoch": 0.58, + "learning_rate": 0.00033566298896690075, + "loss": 1.3844, + "theoretical_loss": 3.469135398568696, + "tokens_seen": 1929773056 + }, + { + "epoch": 0.58, + "learning_rate": 0.00033564694082246745, + "loss": 1.4242, + "theoretical_loss": 3.469125587750213, + "tokens_seen": 1929838592 + }, + { + "epoch": 0.58, + "learning_rate": 0.00033563089267803415, + "loss": 1.2826, + "theoretical_loss": 3.469115777358177, + "tokens_seen": 1929904128 + }, + { + "epoch": 0.58, + "learning_rate": 0.00033561484453360085, + "loss": 1.2809, + "theoretical_loss": 3.4691059673925544, + "tokens_seen": 1929969664 + }, + { + "epoch": 0.58, + "objective/train/docs_used": 703903, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.3803770542144775, + "objective/train/theoretical_loss": 3.4690961578533126, + "objective/train/tokens_used": 1950495200, + "theoretical_loss": 3.4690961578533126, + "tokens_seen": 1930035200 + }, + { + "epoch": 0.58, + "learning_rate": 0.00033559879638916755, + "loss": 1.3522, + "theoretical_loss": 3.4690961578533126, + "tokens_seen": 1930035200 + }, + { + "epoch": 0.58, + "learning_rate": 0.00033558274824473425, + "loss": 1.3236, + "theoretical_loss": 3.4690863487404187, + "tokens_seen": 1930100736 + }, + { + "epoch": 0.58, + "learning_rate": 0.00033556670010030095, + "loss": 1.4344, + "theoretical_loss": 3.4690765400538393, + "tokens_seen": 1930166272 + }, + { + "epoch": 0.58, + "learning_rate": 0.00033555065195586765, + "loss": 1.3958, + "theoretical_loss": 3.469066731793541, + "tokens_seen": 1930231808 + }, + { + "epoch": 0.58, + "learning_rate": 0.00033553460381143435, + "loss": 1.197, + "theoretical_loss": 3.4690569239594917, + "tokens_seen": 1930297344 + }, + { + "epoch": 0.58, + "learning_rate": 0.00033551855566700105, + "loss": 1.3954, + "theoretical_loss": 3.469047116551658, + "tokens_seen": 1930362880 + }, + { + "epoch": 0.58, + "learning_rate": 0.00033550250752256776, + "loss": 1.3318, + "theoretical_loss": 3.469037309570007, + "tokens_seen": 1930428416 + }, + { + "epoch": 0.58, + "learning_rate": 0.00033548645937813446, + "loss": 1.3435, + "theoretical_loss": 3.4690275030145052, + "tokens_seen": 1930493952 + }, + { + "epoch": 0.59, + "learning_rate": 0.00033547041123370116, + "loss": 1.4031, + "theoretical_loss": 3.4690176968851203, + "tokens_seen": 1930559488 + }, + { + "epoch": 0.59, + "learning_rate": 0.00033545436308926786, + "loss": 1.305, + "theoretical_loss": 3.4690078911818194, + "tokens_seen": 1930625024 + }, + { + "epoch": 0.59, + "learning_rate": 0.00033543831494483456, + "loss": 1.3031, + "theoretical_loss": 3.468998085904569, + "tokens_seen": 1930690560 + }, + { + "epoch": 0.59, + "learning_rate": 0.0003354222668004012, + "loss": 1.407, + "theoretical_loss": 3.4689882810533366, + "tokens_seen": 1930756096 + }, + { + "epoch": 0.59, + "learning_rate": 0.0003354062186559679, + "loss": 1.3683, + "theoretical_loss": 3.4689784766280884, + "tokens_seen": 1930821632 + }, + { + "epoch": 0.59, + "learning_rate": 0.0003353901705115346, + "loss": 1.307, + "theoretical_loss": 3.4689686726287925, + "tokens_seen": 1930887168 + }, + { + "epoch": 0.59, + "learning_rate": 0.0003353741223671013, + "loss": 1.3885, + "theoretical_loss": 3.468958869055415, + "tokens_seen": 1930952704 + }, + { + "epoch": 0.59, + "learning_rate": 0.000335358074222668, + "loss": 1.3591, + "theoretical_loss": 3.468949065907924, + "tokens_seen": 1931018240 + }, + { + "epoch": 0.59, + "learning_rate": 0.0003353420260782347, + "loss": 1.2328, + "theoretical_loss": 3.468939263186285, + "tokens_seen": 1931083776 + }, + { + "epoch": 0.59, + "learning_rate": 0.0003353259779338014, + "loss": 1.3874, + "theoretical_loss": 3.4689294608904664, + "tokens_seen": 1931149312 + }, + { + "epoch": 0.59, + "learning_rate": 0.0003353099297893681, + "loss": 1.2883, + "theoretical_loss": 3.468919659020435, + "tokens_seen": 1931214848 + }, + { + "epoch": 0.59, + "learning_rate": 0.0003352938816449348, + "loss": 1.3568, + "theoretical_loss": 3.4689098575761577, + "tokens_seen": 1931280384 + }, + { + "epoch": 0.59, + "learning_rate": 0.0003352778335005015, + "loss": 1.382, + "theoretical_loss": 3.4689000565576014, + "tokens_seen": 1931345920 + }, + { + "epoch": 0.59, + "learning_rate": 0.0003352617853560682, + "loss": 1.3846, + "theoretical_loss": 3.468890255964733, + "tokens_seen": 1931411456 + }, + { + "epoch": 0.59, + "learning_rate": 0.0003352457372116349, + "loss": 1.353, + "theoretical_loss": 3.46888045579752, + "tokens_seen": 1931476992 + }, + { + "epoch": 0.59, + "learning_rate": 0.0003352296890672016, + "loss": 1.3738, + "theoretical_loss": 3.46887065605593, + "tokens_seen": 1931542528 + }, + { + "epoch": 0.59, + "learning_rate": 0.0003352136409227683, + "loss": 1.3848, + "theoretical_loss": 3.468860856739929, + "tokens_seen": 1931608064 + }, + { + "epoch": 0.59, + "learning_rate": 0.000335197592778335, + "loss": 1.3909, + "theoretical_loss": 3.468851057849484, + "tokens_seen": 1931673600 + }, + { + "epoch": 0.59, + "learning_rate": 0.0003351815446339017, + "loss": 1.3633, + "theoretical_loss": 3.468841259384563, + "tokens_seen": 1931739136 + }, + { + "epoch": 0.59, + "learning_rate": 0.0003351654964894684, + "loss": 1.441, + "theoretical_loss": 3.4688314613451325, + "tokens_seen": 1931804672 + }, + { + "epoch": 0.59, + "learning_rate": 0.0003351494483450351, + "loss": 1.3084, + "theoretical_loss": 3.46882166373116, + "tokens_seen": 1931870208 + }, + { + "epoch": 0.59, + "learning_rate": 0.0003351334002006018, + "loss": 1.4019, + "theoretical_loss": 3.4688118665426124, + "tokens_seen": 1931935744 + }, + { + "epoch": 0.59, + "learning_rate": 0.0003351173520561685, + "loss": 1.3672, + "theoretical_loss": 3.4688020697794566, + "tokens_seen": 1932001280 + }, + { + "epoch": 0.59, + "learning_rate": 0.0003351013039117352, + "loss": 1.3681, + "theoretical_loss": 3.46879227344166, + "tokens_seen": 1932066816 + }, + { + "epoch": 0.59, + "learning_rate": 0.0003350852557673019, + "loss": 1.36, + "theoretical_loss": 3.4687824775291896, + "tokens_seen": 1932132352 + }, + { + "epoch": 0.59, + "learning_rate": 0.0003350692076228686, + "loss": 1.2753, + "theoretical_loss": 3.4687726820420117, + "tokens_seen": 1932197888 + }, + { + "epoch": 0.59, + "learning_rate": 0.0003350531594784353, + "loss": 1.2837, + "theoretical_loss": 3.4687628869800946, + "tokens_seen": 1932263424 + }, + { + "epoch": 0.59, + "learning_rate": 0.000335037111334002, + "loss": 1.2753, + "theoretical_loss": 3.4687530923434053, + "tokens_seen": 1932328960 + }, + { + "epoch": 0.59, + "learning_rate": 0.0003350210631895687, + "loss": 1.3443, + "theoretical_loss": 3.46874329813191, + "tokens_seen": 1932394496 + }, + { + "epoch": 0.59, + "learning_rate": 0.0003350050150451354, + "loss": 1.3717, + "theoretical_loss": 3.468733504345577, + "tokens_seen": 1932460032 + }, + { + "epoch": 0.59, + "learning_rate": 0.0003349889669007021, + "loss": 1.3872, + "theoretical_loss": 3.468723710984373, + "tokens_seen": 1932525568 + }, + { + "epoch": 0.59, + "learning_rate": 0.0003349729187562688, + "loss": 1.3117, + "theoretical_loss": 3.4687139180482647, + "tokens_seen": 1932591104 + }, + { + "epoch": 0.59, + "learning_rate": 0.0003349568706118355, + "loss": 1.2904, + "theoretical_loss": 3.4687041255372195, + "tokens_seen": 1932656640 + }, + { + "epoch": 0.59, + "learning_rate": 0.0003349408224674022, + "loss": 1.3796, + "theoretical_loss": 3.4686943334512046, + "tokens_seen": 1932722176 + }, + { + "epoch": 0.59, + "learning_rate": 0.0003349247743229689, + "loss": 1.4053, + "theoretical_loss": 3.4686845417901866, + "tokens_seen": 1932787712 + }, + { + "epoch": 0.59, + "learning_rate": 0.0003349087261785356, + "loss": 1.3863, + "theoretical_loss": 3.4686747505541335, + "tokens_seen": 1932853248 + }, + { + "epoch": 0.59, + "learning_rate": 0.0003348926780341023, + "loss": 1.3965, + "theoretical_loss": 3.4686649597430126, + "tokens_seen": 1932918784 + }, + { + "epoch": 0.59, + "learning_rate": 0.000334876629889669, + "loss": 1.4876, + "theoretical_loss": 3.46865516935679, + "tokens_seen": 1932984320 + }, + { + "epoch": 0.59, + "learning_rate": 0.0003348605817452357, + "loss": 1.1866, + "theoretical_loss": 3.4686453793954337, + "tokens_seen": 1933049856 + }, + { + "epoch": 0.59, + "learning_rate": 0.0003348445336008024, + "loss": 1.2358, + "theoretical_loss": 3.46863558985891, + "tokens_seen": 1933115392 + }, + { + "epoch": 0.59, + "learning_rate": 0.00033482848545636913, + "loss": 1.255, + "theoretical_loss": 3.468625800747187, + "tokens_seen": 1933180928 + }, + { + "epoch": 0.59, + "learning_rate": 0.00033481243731193583, + "loss": 1.3929, + "theoretical_loss": 3.4686160120602314, + "tokens_seen": 1933246464 + }, + { + "epoch": 0.59, + "objective/train/docs_used": 705162, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.3316068649291992, + "objective/train/theoretical_loss": 3.4686062237980106, + "objective/train/tokens_used": 1953772000, + "theoretical_loss": 3.4686062237980106, + "tokens_seen": 1933312000 + }, + { + "epoch": 0.59, + "learning_rate": 0.00033479638916750253, + "loss": 1.3462, + "theoretical_loss": 3.4686062237980106, + "tokens_seen": 1933312000 + }, + { + "epoch": 0.59, + "learning_rate": 0.00033478034102306923, + "loss": 1.1544, + "theoretical_loss": 3.468596435960491, + "tokens_seen": 1933377536 + }, + { + "epoch": 0.59, + "learning_rate": 0.00033476429287863593, + "loss": 1.2861, + "theoretical_loss": 3.468586648547641, + "tokens_seen": 1933443072 + }, + { + "epoch": 0.59, + "learning_rate": 0.00033474824473420263, + "loss": 1.3536, + "theoretical_loss": 3.468576861559427, + "tokens_seen": 1933508608 + }, + { + "epoch": 0.59, + "learning_rate": 0.00033473219658976933, + "loss": 1.463, + "theoretical_loss": 3.468567074995817, + "tokens_seen": 1933574144 + }, + { + "epoch": 0.59, + "learning_rate": 0.00033471614844533603, + "loss": 1.3701, + "theoretical_loss": 3.4685572888567764, + "tokens_seen": 1933639680 + }, + { + "epoch": 0.59, + "learning_rate": 0.00033471614844533603, + "loss": 1.5572, + "theoretical_loss": 3.4685475031422746, + "tokens_seen": 1933705216 + }, + { + "epoch": 0.59, + "learning_rate": 0.00033470010030090273, + "loss": 1.3941, + "theoretical_loss": 3.468537717852277, + "tokens_seen": 1933770752 + }, + { + "epoch": 0.59, + "learning_rate": 0.00033468405215646943, + "loss": 1.3971, + "theoretical_loss": 3.4685279329867518, + "tokens_seen": 1933836288 + }, + { + "epoch": 0.59, + "learning_rate": 0.00033466800401203613, + "loss": 1.2514, + "theoretical_loss": 3.468518148545666, + "tokens_seen": 1933901824 + }, + { + "epoch": 0.59, + "learning_rate": 0.0003346519558676028, + "loss": 1.3725, + "theoretical_loss": 3.468508364528987, + "tokens_seen": 1933967360 + }, + { + "epoch": 0.59, + "learning_rate": 0.0003346359077231695, + "loss": 1.326, + "theoretical_loss": 3.4684985809366813, + "tokens_seen": 1934032896 + }, + { + "epoch": 0.59, + "learning_rate": 0.0003346198595787362, + "loss": 1.2186, + "theoretical_loss": 3.4684887977687167, + "tokens_seen": 1934098432 + }, + { + "epoch": 0.59, + "learning_rate": 0.0003346038114343029, + "loss": 1.3695, + "theoretical_loss": 3.4684790150250606, + "tokens_seen": 1934163968 + }, + { + "epoch": 0.59, + "learning_rate": 0.0003345877632898696, + "loss": 1.3245, + "theoretical_loss": 3.4684692327056794, + "tokens_seen": 1934229504 + }, + { + "epoch": 0.59, + "learning_rate": 0.0003345717151454363, + "loss": 1.3877, + "theoretical_loss": 3.468459450810541, + "tokens_seen": 1934295040 + }, + { + "epoch": 0.59, + "learning_rate": 0.000334555667001003, + "loss": 1.4042, + "theoretical_loss": 3.4684496693396127, + "tokens_seen": 1934360576 + }, + { + "epoch": 0.59, + "learning_rate": 0.0003345396188565697, + "loss": 1.2915, + "theoretical_loss": 3.4684398882928615, + "tokens_seen": 1934426112 + }, + { + "epoch": 0.59, + "learning_rate": 0.00033452357071213644, + "loss": 1.3822, + "theoretical_loss": 3.4684301076702546, + "tokens_seen": 1934491648 + }, + { + "epoch": 0.59, + "learning_rate": 0.00033450752256770314, + "loss": 1.3491, + "theoretical_loss": 3.468420327471759, + "tokens_seen": 1934557184 + }, + { + "epoch": 0.59, + "learning_rate": 0.00033449147442326984, + "loss": 1.3787, + "theoretical_loss": 3.4684105476973426, + "tokens_seen": 1934622720 + }, + { + "epoch": 0.59, + "learning_rate": 0.00033447542627883654, + "loss": 1.3243, + "theoretical_loss": 3.4684007683469718, + "tokens_seen": 1934688256 + }, + { + "epoch": 0.59, + "learning_rate": 0.00033445937813440324, + "loss": 1.32, + "theoretical_loss": 3.468390989420615, + "tokens_seen": 1934753792 + }, + { + "epoch": 0.59, + "learning_rate": 0.00033444332998996994, + "loss": 1.323, + "theoretical_loss": 3.4683812109182384, + "tokens_seen": 1934819328 + }, + { + "epoch": 0.59, + "learning_rate": 0.00033442728184553664, + "loss": 1.4057, + "theoretical_loss": 3.4683714328398096, + "tokens_seen": 1934884864 + }, + { + "epoch": 0.59, + "learning_rate": 0.00033441123370110334, + "loss": 1.3995, + "theoretical_loss": 3.4683616551852956, + "tokens_seen": 1934950400 + }, + { + "epoch": 0.59, + "learning_rate": 0.00033439518555667004, + "loss": 1.4085, + "theoretical_loss": 3.4683518779546647, + "tokens_seen": 1935015936 + }, + { + "epoch": 0.59, + "learning_rate": 0.00033437913741223675, + "loss": 1.3815, + "theoretical_loss": 3.4683421011478828, + "tokens_seen": 1935081472 + }, + { + "epoch": 0.59, + "learning_rate": 0.00033436308926780345, + "loss": 1.4023, + "theoretical_loss": 3.468332324764918, + "tokens_seen": 1935147008 + }, + { + "epoch": 0.59, + "learning_rate": 0.00033434704112337015, + "loss": 1.2796, + "theoretical_loss": 3.4683225488057374, + "tokens_seen": 1935212544 + }, + { + "epoch": 0.59, + "learning_rate": 0.00033433099297893685, + "loss": 1.3043, + "theoretical_loss": 3.468312773270308, + "tokens_seen": 1935278080 + }, + { + "epoch": 0.59, + "learning_rate": 0.00033431494483450355, + "loss": 1.3887, + "theoretical_loss": 3.4683029981585975, + "tokens_seen": 1935343616 + }, + { + "epoch": 0.59, + "learning_rate": 0.00033429889669007025, + "loss": 1.3611, + "theoretical_loss": 3.468293223470573, + "tokens_seen": 1935409152 + }, + { + "epoch": 0.59, + "learning_rate": 0.00033428284854563695, + "loss": 1.2935, + "theoretical_loss": 3.4682834492062016, + "tokens_seen": 1935474688 + }, + { + "epoch": 0.59, + "learning_rate": 0.00033426680040120365, + "loss": 1.3382, + "theoretical_loss": 3.4682736753654515, + "tokens_seen": 1935540224 + }, + { + "epoch": 0.59, + "learning_rate": 0.00033425075225677035, + "loss": 1.2244, + "theoretical_loss": 3.468263901948289, + "tokens_seen": 1935605760 + }, + { + "epoch": 0.59, + "learning_rate": 0.00033423470411233705, + "loss": 1.2584, + "theoretical_loss": 3.468254128954681, + "tokens_seen": 1935671296 + }, + { + "epoch": 0.59, + "learning_rate": 0.00033421865596790375, + "loss": 1.2974, + "theoretical_loss": 3.468244356384596, + "tokens_seen": 1935736832 + }, + { + "epoch": 0.59, + "learning_rate": 0.00033420260782347045, + "loss": 1.3074, + "theoretical_loss": 3.468234584238001, + "tokens_seen": 1935802368 + }, + { + "epoch": 0.59, + "learning_rate": 0.00033418655967903715, + "loss": 1.392, + "theoretical_loss": 3.468224812514863, + "tokens_seen": 1935867904 + }, + { + "epoch": 0.59, + "learning_rate": 0.00033417051153460385, + "loss": 1.2698, + "theoretical_loss": 3.4682150412151493, + "tokens_seen": 1935933440 + }, + { + "epoch": 0.59, + "learning_rate": 0.00033415446339017055, + "loss": 1.3468, + "theoretical_loss": 3.468205270338828, + "tokens_seen": 1935998976 + }, + { + "epoch": 0.59, + "learning_rate": 0.00033413841524573726, + "loss": 1.3024, + "theoretical_loss": 3.468195499885865, + "tokens_seen": 1936064512 + }, + { + "epoch": 0.59, + "learning_rate": 0.00033412236710130396, + "loss": 1.3549, + "theoretical_loss": 3.468185729856229, + "tokens_seen": 1936130048 + }, + { + "epoch": 0.59, + "learning_rate": 0.00033410631895687066, + "loss": 1.4147, + "theoretical_loss": 3.4681759602498863, + "tokens_seen": 1936195584 + }, + { + "epoch": 0.59, + "learning_rate": 0.00033409027081243736, + "loss": 1.4507, + "theoretical_loss": 3.4681661910668047, + "tokens_seen": 1936261120 + }, + { + "epoch": 0.59, + "learning_rate": 0.00033407422266800406, + "loss": 1.3873, + "theoretical_loss": 3.4681564223069516, + "tokens_seen": 1936326656 + }, + { + "epoch": 0.59, + "learning_rate": 0.00033405817452357076, + "loss": 1.5456, + "theoretical_loss": 3.4681466539702948, + "tokens_seen": 1936392192 + }, + { + "epoch": 0.59, + "learning_rate": 0.00033404212637913746, + "loss": 1.4059, + "theoretical_loss": 3.4681368860568, + "tokens_seen": 1936457728 + }, + { + "epoch": 0.59, + "learning_rate": 0.0003340260782347041, + "loss": 1.4352, + "theoretical_loss": 3.4681271185664366, + "tokens_seen": 1936523264 + }, + { + "epoch": 0.59, + "objective/train/docs_used": 706032, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.280336856842041, + "objective/train/theoretical_loss": 3.4681173514991706, + "objective/train/tokens_used": 1957048800, + "theoretical_loss": 3.4681173514991706, + "tokens_seen": 1936588800 + }, + { + "epoch": 0.59, + "learning_rate": 0.0003340100300902708, + "loss": 1.3274, + "theoretical_loss": 3.4681173514991706, + "tokens_seen": 1936588800 + }, + { + "epoch": 0.59, + "learning_rate": 0.0003339939819458375, + "loss": 1.315, + "theoretical_loss": 3.4681075848549696, + "tokens_seen": 1936654336 + }, + { + "epoch": 0.59, + "learning_rate": 0.0003339779338014042, + "loss": 1.3441, + "theoretical_loss": 3.4680978186338014, + "tokens_seen": 1936719872 + }, + { + "epoch": 0.59, + "learning_rate": 0.0003339618856569709, + "loss": 1.3813, + "theoretical_loss": 3.4680880528356326, + "tokens_seen": 1936785408 + }, + { + "epoch": 0.59, + "learning_rate": 0.0003339458375125376, + "loss": 1.4026, + "theoretical_loss": 3.4680782874604317, + "tokens_seen": 1936850944 + }, + { + "epoch": 0.59, + "learning_rate": 0.0003339297893681043, + "loss": 1.3552, + "theoretical_loss": 3.4680685225081653, + "tokens_seen": 1936916480 + }, + { + "epoch": 0.59, + "learning_rate": 0.000333913741223671, + "loss": 1.3149, + "theoretical_loss": 3.4680587579788007, + "tokens_seen": 1936982016 + }, + { + "epoch": 0.59, + "learning_rate": 0.0003338976930792377, + "loss": 1.3102, + "theoretical_loss": 3.468048993872306, + "tokens_seen": 1937047552 + }, + { + "epoch": 0.59, + "learning_rate": 0.0003338816449348044, + "loss": 1.3118, + "theoretical_loss": 3.468039230188647, + "tokens_seen": 1937113088 + }, + { + "epoch": 0.59, + "learning_rate": 0.0003338655967903711, + "loss": 1.3224, + "theoretical_loss": 3.468029466927793, + "tokens_seen": 1937178624 + }, + { + "epoch": 0.59, + "learning_rate": 0.0003338495486459378, + "loss": 1.3863, + "theoretical_loss": 3.46801970408971, + "tokens_seen": 1937244160 + }, + { + "epoch": 0.59, + "learning_rate": 0.0003338335005015045, + "loss": 1.3485, + "theoretical_loss": 3.468009941674366, + "tokens_seen": 1937309696 + }, + { + "epoch": 0.59, + "learning_rate": 0.0003338174523570712, + "loss": 1.3925, + "theoretical_loss": 3.468000179681728, + "tokens_seen": 1937375232 + }, + { + "epoch": 0.59, + "learning_rate": 0.0003338014042126379, + "loss": 1.3431, + "theoretical_loss": 3.4679904181117642, + "tokens_seen": 1937440768 + }, + { + "epoch": 0.59, + "learning_rate": 0.0003337853560682046, + "loss": 1.3613, + "theoretical_loss": 3.467980656964442, + "tokens_seen": 1937506304 + }, + { + "epoch": 0.59, + "learning_rate": 0.0003337693079237713, + "loss": 1.4148, + "theoretical_loss": 3.467970896239727, + "tokens_seen": 1937571840 + }, + { + "epoch": 0.59, + "learning_rate": 0.000333753259779338, + "loss": 1.3601, + "theoretical_loss": 3.467961135937589, + "tokens_seen": 1937637376 + }, + { + "epoch": 0.59, + "learning_rate": 0.0003337372116349047, + "loss": 1.3948, + "theoretical_loss": 3.467951376057994, + "tokens_seen": 1937702912 + }, + { + "epoch": 0.59, + "learning_rate": 0.0003337211634904714, + "loss": 1.3455, + "theoretical_loss": 3.4679416166009096, + "tokens_seen": 1937768448 + }, + { + "epoch": 0.59, + "learning_rate": 0.0003337051153460381, + "loss": 1.3376, + "theoretical_loss": 3.467931857566304, + "tokens_seen": 1937833984 + }, + { + "epoch": 0.59, + "learning_rate": 0.0003336890672016048, + "loss": 1.2677, + "theoretical_loss": 3.4679220989541433, + "tokens_seen": 1937899520 + }, + { + "epoch": 0.59, + "learning_rate": 0.0003336730190571715, + "loss": 1.3736, + "theoretical_loss": 3.4679123407643955, + "tokens_seen": 1937965056 + }, + { + "epoch": 0.59, + "learning_rate": 0.0003336569709127382, + "loss": 1.3535, + "theoretical_loss": 3.467902582997029, + "tokens_seen": 1938030592 + }, + { + "epoch": 0.59, + "learning_rate": 0.0003336409227683049, + "loss": 1.3222, + "theoretical_loss": 3.46789282565201, + "tokens_seen": 1938096128 + }, + { + "epoch": 0.59, + "learning_rate": 0.0003336248746238716, + "loss": 1.4582, + "theoretical_loss": 3.467883068729306, + "tokens_seen": 1938161664 + }, + { + "epoch": 0.59, + "learning_rate": 0.0003336088264794383, + "loss": 1.3143, + "theoretical_loss": 3.467873312228885, + "tokens_seen": 1938227200 + }, + { + "epoch": 0.59, + "learning_rate": 0.000333592778335005, + "loss": 1.3699, + "theoretical_loss": 3.4678635561507143, + "tokens_seen": 1938292736 + }, + { + "epoch": 0.59, + "learning_rate": 0.0003335767301905717, + "loss": 1.3567, + "theoretical_loss": 3.467853800494761, + "tokens_seen": 1938358272 + }, + { + "epoch": 0.59, + "learning_rate": 0.0003335606820461384, + "loss": 1.3622, + "theoretical_loss": 3.467844045260993, + "tokens_seen": 1938423808 + }, + { + "epoch": 0.59, + "learning_rate": 0.0003335446339017051, + "loss": 1.3221, + "theoretical_loss": 3.4678342904493777, + "tokens_seen": 1938489344 + }, + { + "epoch": 0.59, + "learning_rate": 0.0003335285857572718, + "loss": 1.3242, + "theoretical_loss": 3.467824536059882, + "tokens_seen": 1938554880 + }, + { + "epoch": 0.59, + "learning_rate": 0.0003335125376128385, + "loss": 1.3582, + "theoretical_loss": 3.4678147820924745, + "tokens_seen": 1938620416 + }, + { + "epoch": 0.59, + "learning_rate": 0.0003334964894684052, + "loss": 1.2934, + "theoretical_loss": 3.4678050285471214, + "tokens_seen": 1938685952 + }, + { + "epoch": 0.59, + "learning_rate": 0.0003334804413239719, + "loss": 1.3101, + "theoretical_loss": 3.467795275423791, + "tokens_seen": 1938751488 + }, + { + "epoch": 0.59, + "learning_rate": 0.00033346439317953863, + "loss": 1.3736, + "theoretical_loss": 3.4677855227224508, + "tokens_seen": 1938817024 + }, + { + "epoch": 0.59, + "learning_rate": 0.00033344834503510533, + "loss": 1.3968, + "theoretical_loss": 3.4677757704430676, + "tokens_seen": 1938882560 + }, + { + "epoch": 0.59, + "learning_rate": 0.00033343229689067203, + "loss": 1.2575, + "theoretical_loss": 3.4677660185856096, + "tokens_seen": 1938948096 + }, + { + "epoch": 0.59, + "learning_rate": 0.00033341624874623873, + "loss": 1.3443, + "theoretical_loss": 3.467756267150044, + "tokens_seen": 1939013632 + }, + { + "epoch": 0.59, + "learning_rate": 0.00033340020060180543, + "loss": 1.366, + "theoretical_loss": 3.467746516136337, + "tokens_seen": 1939079168 + }, + { + "epoch": 0.59, + "learning_rate": 0.00033338415245737213, + "loss": 1.2674, + "theoretical_loss": 3.467736765544459, + "tokens_seen": 1939144704 + }, + { + "epoch": 0.59, + "learning_rate": 0.00033336810431293883, + "loss": 1.3544, + "theoretical_loss": 3.467727015374375, + "tokens_seen": 1939210240 + }, + { + "epoch": 0.59, + "learning_rate": 0.00033335205616850553, + "loss": 1.3592, + "theoretical_loss": 3.467717265626054, + "tokens_seen": 1939275776 + }, + { + "epoch": 0.59, + "learning_rate": 0.00033333600802407223, + "loss": 1.401, + "theoretical_loss": 3.4677075162994617, + "tokens_seen": 1939341312 + }, + { + "epoch": 0.59, + "learning_rate": 0.00033331995987963893, + "loss": 1.3605, + "theoretical_loss": 3.4676977673945677, + "tokens_seen": 1939406848 + }, + { + "epoch": 0.59, + "learning_rate": 0.00033330391173520563, + "loss": 1.3936, + "theoretical_loss": 3.467688018911338, + "tokens_seen": 1939472384 + }, + { + "epoch": 0.59, + "learning_rate": 0.00033328786359077233, + "loss": 1.3555, + "theoretical_loss": 3.467678270849741, + "tokens_seen": 1939537920 + }, + { + "epoch": 0.59, + "learning_rate": 0.00033327181544633904, + "loss": 1.3062, + "theoretical_loss": 3.467668523209744, + "tokens_seen": 1939603456 + }, + { + "epoch": 0.59, + "learning_rate": 0.00033325576730190574, + "loss": 1.3574, + "theoretical_loss": 3.467658775991315, + "tokens_seen": 1939668992 + }, + { + "epoch": 0.59, + "learning_rate": 0.00033323971915747244, + "loss": 1.2953, + "theoretical_loss": 3.46764902919442, + "tokens_seen": 1939734528 + }, + { + "epoch": 0.59, + "learning_rate": 0.00033322367101303914, + "loss": 1.3469, + "theoretical_loss": 3.4676392828190274, + "tokens_seen": 1939800064 + }, + { + "epoch": 0.59, + "objective/train/docs_used": 707259, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.3778421878814697, + "objective/train/theoretical_loss": 3.4676295368651053, + "objective/train/tokens_used": 1960325600, + "theoretical_loss": 3.4676295368651053, + "tokens_seen": 1939865600 + }, + { + "epoch": 0.59, + "learning_rate": 0.00033320762286860584, + "loss": 1.3677, + "theoretical_loss": 3.4676295368651053, + "tokens_seen": 1939865600 + }, + { + "epoch": 0.59, + "learning_rate": 0.00033319157472417254, + "loss": 1.3072, + "theoretical_loss": 3.4676197913326208, + "tokens_seen": 1939931136 + }, + { + "epoch": 0.59, + "learning_rate": 0.00033317552657973924, + "loss": 1.336, + "theoretical_loss": 3.467610046221541, + "tokens_seen": 1939996672 + }, + { + "epoch": 0.59, + "learning_rate": 0.00033315947843530594, + "loss": 1.3979, + "theoretical_loss": 3.467600301531834, + "tokens_seen": 1940062208 + }, + { + "epoch": 0.59, + "learning_rate": 0.00033314343029087264, + "loss": 1.3238, + "theoretical_loss": 3.467590557263467, + "tokens_seen": 1940127744 + }, + { + "epoch": 0.59, + "learning_rate": 0.00033312738214643934, + "loss": 1.2766, + "theoretical_loss": 3.467580813416408, + "tokens_seen": 1940193280 + }, + { + "epoch": 0.59, + "learning_rate": 0.00033311133400200604, + "loss": 1.3718, + "theoretical_loss": 3.4675710699906244, + "tokens_seen": 1940258816 + }, + { + "epoch": 0.59, + "learning_rate": 0.00033309528585757274, + "loss": 1.3294, + "theoretical_loss": 3.4675613269860834, + "tokens_seen": 1940324352 + }, + { + "epoch": 0.59, + "learning_rate": 0.00033307923771313944, + "loss": 1.3579, + "theoretical_loss": 3.4675515844027522, + "tokens_seen": 1940389888 + }, + { + "epoch": 0.59, + "learning_rate": 0.00033306318956870614, + "loss": 1.3906, + "theoretical_loss": 3.4675418422405997, + "tokens_seen": 1940455424 + }, + { + "epoch": 0.59, + "learning_rate": 0.00033304714142427284, + "loss": 1.3999, + "theoretical_loss": 3.467532100499593, + "tokens_seen": 1940520960 + }, + { + "epoch": 0.59, + "learning_rate": 0.00033303109327983954, + "loss": 1.3218, + "theoretical_loss": 3.4675223591796986, + "tokens_seen": 1940586496 + }, + { + "epoch": 0.59, + "learning_rate": 0.00033301504513540625, + "loss": 1.3334, + "theoretical_loss": 3.4675126182808853, + "tokens_seen": 1940652032 + }, + { + "epoch": 0.59, + "learning_rate": 0.00033299899699097295, + "loss": 1.3403, + "theoretical_loss": 3.4675028778031205, + "tokens_seen": 1940717568 + }, + { + "epoch": 0.59, + "learning_rate": 0.00033298294884653965, + "loss": 1.32, + "theoretical_loss": 3.467493137746371, + "tokens_seen": 1940783104 + }, + { + "epoch": 0.59, + "learning_rate": 0.00033296690070210635, + "loss": 1.2315, + "theoretical_loss": 3.4674833981106055, + "tokens_seen": 1940848640 + }, + { + "epoch": 0.59, + "learning_rate": 0.00033295085255767305, + "loss": 1.3081, + "theoretical_loss": 3.4674736588957904, + "tokens_seen": 1940914176 + }, + { + "epoch": 0.59, + "learning_rate": 0.00033293480441323975, + "loss": 1.327, + "theoretical_loss": 3.4674639201018946, + "tokens_seen": 1940979712 + }, + { + "epoch": 0.59, + "learning_rate": 0.00033291875626880645, + "loss": 1.3805, + "theoretical_loss": 3.4674541817288844, + "tokens_seen": 1941045248 + }, + { + "epoch": 0.59, + "learning_rate": 0.00033290270812437315, + "loss": 1.3642, + "theoretical_loss": 3.4674444437767282, + "tokens_seen": 1941110784 + }, + { + "epoch": 0.59, + "learning_rate": 0.00033288665997993985, + "loss": 1.3623, + "theoretical_loss": 3.4674347062453936, + "tokens_seen": 1941176320 + }, + { + "epoch": 0.59, + "learning_rate": 0.00033287061183550655, + "loss": 1.3163, + "theoretical_loss": 3.467424969134848, + "tokens_seen": 1941241856 + }, + { + "epoch": 0.59, + "learning_rate": 0.00033285456369107325, + "loss": 1.287, + "theoretical_loss": 3.4674152324450587, + "tokens_seen": 1941307392 + }, + { + "epoch": 0.59, + "learning_rate": 0.00033283851554663995, + "loss": 1.3739, + "theoretical_loss": 3.467405496175994, + "tokens_seen": 1941372928 + }, + { + "epoch": 0.59, + "learning_rate": 0.00033282246740220665, + "loss": 1.2539, + "theoretical_loss": 3.467395760327621, + "tokens_seen": 1941438464 + }, + { + "epoch": 0.59, + "learning_rate": 0.00033280641925777335, + "loss": 1.4524, + "theoretical_loss": 3.4673860248999073, + "tokens_seen": 1941504000 + }, + { + "epoch": 0.59, + "learning_rate": 0.00033279037111334005, + "loss": 1.373, + "theoretical_loss": 3.467376289892821, + "tokens_seen": 1941569536 + }, + { + "epoch": 0.59, + "learning_rate": 0.00033277432296890675, + "loss": 1.4005, + "theoretical_loss": 3.4673665553063295, + "tokens_seen": 1941635072 + }, + { + "epoch": 0.59, + "learning_rate": 0.00033275827482447346, + "loss": 1.3279, + "theoretical_loss": 3.4673568211404002, + "tokens_seen": 1941700608 + }, + { + "epoch": 0.59, + "learning_rate": 0.00033274222668004016, + "loss": 1.2181, + "theoretical_loss": 3.4673470873950007, + "tokens_seen": 1941766144 + }, + { + "epoch": 0.59, + "learning_rate": 0.00033272617853560686, + "loss": 1.2267, + "theoretical_loss": 3.467337354070099, + "tokens_seen": 1941831680 + }, + { + "epoch": 0.59, + "learning_rate": 0.00033271013039117356, + "loss": 1.3045, + "theoretical_loss": 3.467327621165663, + "tokens_seen": 1941897216 + }, + { + "epoch": 0.59, + "learning_rate": 0.00033269408224674026, + "loss": 1.3422, + "theoretical_loss": 3.46731788868166, + "tokens_seen": 1941962752 + }, + { + "epoch": 0.59, + "learning_rate": 0.00033267803410230696, + "loss": 1.3328, + "theoretical_loss": 3.4673081566180572, + "tokens_seen": 1942028288 + }, + { + "epoch": 0.59, + "learning_rate": 0.00033266198595787366, + "loss": 1.293, + "theoretical_loss": 3.467298424974823, + "tokens_seen": 1942093824 + }, + { + "epoch": 0.59, + "learning_rate": 0.00033264593781344036, + "loss": 1.3034, + "theoretical_loss": 3.4672886937519243, + "tokens_seen": 1942159360 + }, + { + "epoch": 0.59, + "learning_rate": 0.000332629889669007, + "loss": 1.2758, + "theoretical_loss": 3.467278962949329, + "tokens_seen": 1942224896 + }, + { + "epoch": 0.59, + "learning_rate": 0.0003326138415245737, + "loss": 1.3598, + "theoretical_loss": 3.4672692325670056, + "tokens_seen": 1942290432 + }, + { + "epoch": 0.59, + "learning_rate": 0.0003325977933801404, + "loss": 1.2922, + "theoretical_loss": 3.4672595026049207, + "tokens_seen": 1942355968 + }, + { + "epoch": 0.59, + "learning_rate": 0.0003325817452357071, + "loss": 1.2849, + "theoretical_loss": 3.467249773063042, + "tokens_seen": 1942421504 + }, + { + "epoch": 0.59, + "learning_rate": 0.0003325656970912738, + "loss": 1.3296, + "theoretical_loss": 3.467240043941338, + "tokens_seen": 1942487040 + }, + { + "epoch": 0.59, + "learning_rate": 0.0003325496489468405, + "loss": 1.4222, + "theoretical_loss": 3.4672303152397763, + "tokens_seen": 1942552576 + }, + { + "epoch": 0.59, + "learning_rate": 0.0003325336008024072, + "loss": 1.3582, + "theoretical_loss": 3.4672205869583235, + "tokens_seen": 1942618112 + }, + { + "epoch": 0.59, + "learning_rate": 0.0003325175526579739, + "loss": 1.4114, + "theoretical_loss": 3.467210859096948, + "tokens_seen": 1942683648 + }, + { + "epoch": 0.59, + "learning_rate": 0.0003325015045135406, + "loss": 1.2792, + "theoretical_loss": 3.467201131655618, + "tokens_seen": 1942749184 + }, + { + "epoch": 0.59, + "learning_rate": 0.0003324854563691073, + "loss": 1.3696, + "theoretical_loss": 3.467191404634301, + "tokens_seen": 1942814720 + }, + { + "epoch": 0.59, + "learning_rate": 0.000332469408224674, + "loss": 1.3479, + "theoretical_loss": 3.467181678032963, + "tokens_seen": 1942880256 + }, + { + "epoch": 0.59, + "learning_rate": 0.0003324533600802407, + "loss": 1.2779, + "theoretical_loss": 3.4671719518515745, + "tokens_seen": 1942945792 + }, + { + "epoch": 0.59, + "learning_rate": 0.0003324373119358074, + "loss": 1.2727, + "theoretical_loss": 3.4671622260901014, + "tokens_seen": 1943011328 + }, + { + "epoch": 0.59, + "learning_rate": 0.0003324212637913741, + "loss": 1.4252, + "theoretical_loss": 3.4671525007485116, + "tokens_seen": 1943076864 + }, + { + "epoch": 0.59, + "objective/train/docs_used": 708379, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.3107712268829346, + "objective/train/theoretical_loss": 3.4671427758267725, + "objective/train/tokens_used": 1963602400, + "theoretical_loss": 3.4671427758267725, + "tokens_seen": 1943142400 + }, + { + "epoch": 0.59, + "learning_rate": 0.0003324052156469408, + "loss": 1.2629, + "theoretical_loss": 3.4671427758267725, + "tokens_seen": 1943142400 + }, + { + "epoch": 0.59, + "learning_rate": 0.0003323891675025075, + "loss": 1.4256, + "theoretical_loss": 3.4671330513248533, + "tokens_seen": 1943207936 + }, + { + "epoch": 0.59, + "learning_rate": 0.0003323731193580742, + "loss": 1.415, + "theoretical_loss": 3.46712332724272, + "tokens_seen": 1943273472 + }, + { + "epoch": 0.59, + "learning_rate": 0.0003323570712136409, + "loss": 1.3862, + "theoretical_loss": 3.4671136035803416, + "tokens_seen": 1943339008 + }, + { + "epoch": 0.59, + "learning_rate": 0.0003323410230692076, + "loss": 1.2965, + "theoretical_loss": 3.4671038803376852, + "tokens_seen": 1943404544 + }, + { + "epoch": 0.59, + "learning_rate": 0.0003323249749247743, + "loss": 1.3198, + "theoretical_loss": 3.4670941575147185, + "tokens_seen": 1943470080 + }, + { + "epoch": 0.59, + "learning_rate": 0.000332308926780341, + "loss": 1.3206, + "theoretical_loss": 3.4670844351114094, + "tokens_seen": 1943535616 + }, + { + "epoch": 0.59, + "learning_rate": 0.0003322928786359077, + "loss": 1.3612, + "theoretical_loss": 3.4670747131277255, + "tokens_seen": 1943601152 + }, + { + "epoch": 0.59, + "learning_rate": 0.0003322768304914744, + "loss": 1.3786, + "theoretical_loss": 3.467064991563635, + "tokens_seen": 1943666688 + }, + { + "epoch": 0.59, + "learning_rate": 0.0003322607823470411, + "loss": 1.3725, + "theoretical_loss": 3.4670552704191047, + "tokens_seen": 1943732224 + }, + { + "epoch": 0.59, + "learning_rate": 0.0003322447342026078, + "loss": 1.4725, + "theoretical_loss": 3.4670455496941033, + "tokens_seen": 1943797760 + }, + { + "epoch": 0.59, + "learning_rate": 0.0003322286860581745, + "loss": 1.3224, + "theoretical_loss": 3.467035829388598, + "tokens_seen": 1943863296 + }, + { + "epoch": 0.59, + "learning_rate": 0.0003322126379137412, + "loss": 1.2773, + "theoretical_loss": 3.467026109502557, + "tokens_seen": 1943928832 + }, + { + "epoch": 0.59, + "learning_rate": 0.0003321965897693079, + "loss": 1.3599, + "theoretical_loss": 3.4670163900359476, + "tokens_seen": 1943994368 + }, + { + "epoch": 0.59, + "learning_rate": 0.0003321805416248746, + "loss": 1.4292, + "theoretical_loss": 3.4670066709887375, + "tokens_seen": 1944059904 + }, + { + "epoch": 0.59, + "learning_rate": 0.0003321644934804413, + "loss": 1.3065, + "theoretical_loss": 3.466996952360895, + "tokens_seen": 1944125440 + }, + { + "epoch": 0.59, + "learning_rate": 0.000332148445336008, + "loss": 1.2734, + "theoretical_loss": 3.4669872341523877, + "tokens_seen": 1944190976 + }, + { + "epoch": 0.59, + "learning_rate": 0.0003321323971915747, + "loss": 1.1265, + "theoretical_loss": 3.4669775163631833, + "tokens_seen": 1944256512 + }, + { + "epoch": 0.59, + "learning_rate": 0.0003321163490471415, + "loss": 1.4537, + "theoretical_loss": 3.466967798993249, + "tokens_seen": 1944322048 + }, + { + "epoch": 0.59, + "learning_rate": 0.0003321003009027082, + "loss": 1.3144, + "theoretical_loss": 3.466958082042553, + "tokens_seen": 1944387584 + }, + { + "epoch": 0.59, + "learning_rate": 0.0003320842527582749, + "loss": 1.3232, + "theoretical_loss": 3.466948365511064, + "tokens_seen": 1944453120 + }, + { + "epoch": 0.59, + "learning_rate": 0.0003320682046138416, + "loss": 1.4812, + "theoretical_loss": 3.466938649398749, + "tokens_seen": 1944518656 + }, + { + "epoch": 0.59, + "learning_rate": 0.0003320521564694083, + "loss": 1.3478, + "theoretical_loss": 3.4669289337055753, + "tokens_seen": 1944584192 + }, + { + "epoch": 0.59, + "learning_rate": 0.000332036108324975, + "loss": 1.3788, + "theoretical_loss": 3.466919218431511, + "tokens_seen": 1944649728 + }, + { + "epoch": 0.59, + "learning_rate": 0.0003320200601805417, + "loss": 1.3286, + "theoretical_loss": 3.4669095035765247, + "tokens_seen": 1944715264 + }, + { + "epoch": 0.59, + "learning_rate": 0.0003320040120361084, + "loss": 1.3754, + "theoretical_loss": 3.466899789140583, + "tokens_seen": 1944780800 + }, + { + "epoch": 0.59, + "learning_rate": 0.00033198796389167503, + "loss": 1.334, + "theoretical_loss": 3.466890075123654, + "tokens_seen": 1944846336 + }, + { + "epoch": 0.59, + "learning_rate": 0.00033197191574724173, + "loss": 1.3928, + "theoretical_loss": 3.466880361525707, + "tokens_seen": 1944911872 + }, + { + "epoch": 0.59, + "learning_rate": 0.00033195586760280843, + "loss": 1.3293, + "theoretical_loss": 3.466870648346707, + "tokens_seen": 1944977408 + }, + { + "epoch": 0.59, + "learning_rate": 0.00033193981945837513, + "loss": 1.3007, + "theoretical_loss": 3.4668609355866247, + "tokens_seen": 1945042944 + }, + { + "epoch": 0.59, + "learning_rate": 0.00033192377131394183, + "loss": 1.3207, + "theoretical_loss": 3.466851223245426, + "tokens_seen": 1945108480 + }, + { + "epoch": 0.59, + "learning_rate": 0.00033190772316950853, + "loss": 1.3753, + "theoretical_loss": 3.4668415113230795, + "tokens_seen": 1945174016 + }, + { + "epoch": 0.59, + "learning_rate": 0.00033189167502507524, + "loss": 1.3894, + "theoretical_loss": 3.466831799819553, + "tokens_seen": 1945239552 + }, + { + "epoch": 0.59, + "learning_rate": 0.00033187562688064194, + "loss": 1.1942, + "theoretical_loss": 3.466822088734814, + "tokens_seen": 1945305088 + }, + { + "epoch": 0.59, + "learning_rate": 0.00033185957873620864, + "loss": 1.3764, + "theoretical_loss": 3.4668123780688305, + "tokens_seen": 1945370624 + }, + { + "epoch": 0.59, + "learning_rate": 0.00033184353059177534, + "loss": 1.3645, + "theoretical_loss": 3.4668026678215704, + "tokens_seen": 1945436160 + }, + { + "epoch": 0.59, + "learning_rate": 0.00033182748244734204, + "loss": 1.3195, + "theoretical_loss": 3.4667929579930012, + "tokens_seen": 1945501696 + }, + { + "epoch": 0.59, + "learning_rate": 0.00033181143430290874, + "loss": 1.3328, + "theoretical_loss": 3.466783248583092, + "tokens_seen": 1945567232 + }, + { + "epoch": 0.59, + "learning_rate": 0.00033179538615847544, + "loss": 1.2852, + "theoretical_loss": 3.466773539591809, + "tokens_seen": 1945632768 + }, + { + "epoch": 0.59, + "learning_rate": 0.00033177933801404214, + "loss": 1.2039, + "theoretical_loss": 3.4667638310191204, + "tokens_seen": 1945698304 + }, + { + "epoch": 0.59, + "learning_rate": 0.00033176328986960884, + "loss": 1.4006, + "theoretical_loss": 3.4667541228649954, + "tokens_seen": 1945763840 + }, + { + "epoch": 0.59, + "learning_rate": 0.00033174724172517554, + "loss": 1.3491, + "theoretical_loss": 3.4667444151294005, + "tokens_seen": 1945829376 + }, + { + "epoch": 0.59, + "learning_rate": 0.00033173119358074224, + "loss": 1.2612, + "theoretical_loss": 3.4667347078123036, + "tokens_seen": 1945894912 + }, + { + "epoch": 0.59, + "learning_rate": 0.00033171514543630894, + "loss": 1.3889, + "theoretical_loss": 3.466725000913673, + "tokens_seen": 1945960448 + }, + { + "epoch": 0.59, + "learning_rate": 0.00033169909729187564, + "loss": 1.3587, + "theoretical_loss": 3.4667152944334765, + "tokens_seen": 1946025984 + }, + { + "epoch": 0.59, + "learning_rate": 0.00033168304914744234, + "loss": 1.31, + "theoretical_loss": 3.466705588371682, + "tokens_seen": 1946091520 + }, + { + "epoch": 0.59, + "learning_rate": 0.00033166700100300904, + "loss": 1.251, + "theoretical_loss": 3.4666958827282572, + "tokens_seen": 1946157056 + }, + { + "epoch": 0.59, + "learning_rate": 0.00033165095285857574, + "loss": 1.3018, + "theoretical_loss": 3.46668617750317, + "tokens_seen": 1946222592 + }, + { + "epoch": 0.59, + "learning_rate": 0.00033163490471414245, + "loss": 1.4172, + "theoretical_loss": 3.4666764726963883, + "tokens_seen": 1946288128 + }, + { + "epoch": 0.59, + "learning_rate": 0.00033161885656970915, + "loss": 1.3424, + "theoretical_loss": 3.4666667683078805, + "tokens_seen": 1946353664 + }, + { + "epoch": 0.59, + "objective/train/docs_used": 709628, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.3471348285675049, + "objective/train/theoretical_loss": 3.466657064337614, + "objective/train/tokens_used": 1966879200, + "theoretical_loss": 3.466657064337614, + "tokens_seen": 1946419200 + }, + { + "epoch": 0.59, + "learning_rate": 0.00033160280842527585, + "loss": 1.302, + "theoretical_loss": 3.466657064337614, + "tokens_seen": 1946419200 + }, + { + "epoch": 0.59, + "learning_rate": 0.00033158676028084255, + "loss": 1.3149, + "theoretical_loss": 3.466647360785556, + "tokens_seen": 1946484736 + }, + { + "epoch": 0.59, + "learning_rate": 0.00033157071213640925, + "loss": 1.4229, + "theoretical_loss": 3.4666376576516758, + "tokens_seen": 1946550272 + }, + { + "epoch": 0.59, + "learning_rate": 0.00033155466399197595, + "loss": 1.2616, + "theoretical_loss": 3.4666279549359404, + "tokens_seen": 1946615808 + }, + { + "epoch": 0.59, + "learning_rate": 0.00033153861584754265, + "loss": 1.3445, + "theoretical_loss": 3.4666182526383182, + "tokens_seen": 1946681344 + }, + { + "epoch": 0.59, + "learning_rate": 0.00033152256770310935, + "loss": 1.223, + "theoretical_loss": 3.4666085507587767, + "tokens_seen": 1946746880 + }, + { + "epoch": 0.59, + "learning_rate": 0.00033150651955867605, + "loss": 1.4175, + "theoretical_loss": 3.466598849297284, + "tokens_seen": 1946812416 + }, + { + "epoch": 0.59, + "learning_rate": 0.00033149047141424275, + "loss": 1.4318, + "theoretical_loss": 3.466589148253808, + "tokens_seen": 1946877952 + }, + { + "epoch": 0.59, + "learning_rate": 0.00033147442326980945, + "loss": 1.3934, + "theoretical_loss": 3.466579447628316, + "tokens_seen": 1946943488 + }, + { + "epoch": 0.59, + "learning_rate": 0.00033145837512537615, + "loss": 1.4441, + "theoretical_loss": 3.4665697474207775, + "tokens_seen": 1947009024 + }, + { + "epoch": 0.59, + "learning_rate": 0.00033144232698094285, + "loss": 1.3661, + "theoretical_loss": 3.466560047631159, + "tokens_seen": 1947074560 + }, + { + "epoch": 0.59, + "learning_rate": 0.00033142627883650955, + "loss": 1.2631, + "theoretical_loss": 3.466550348259428, + "tokens_seen": 1947140096 + }, + { + "epoch": 0.59, + "learning_rate": 0.00033141023069207625, + "loss": 1.375, + "theoretical_loss": 3.4665406493055544, + "tokens_seen": 1947205632 + }, + { + "epoch": 0.59, + "learning_rate": 0.00033139418254764296, + "loss": 1.2844, + "theoretical_loss": 3.4665309507695046, + "tokens_seen": 1947271168 + }, + { + "epoch": 0.59, + "learning_rate": 0.00033137813440320966, + "loss": 1.4297, + "theoretical_loss": 3.466521252651247, + "tokens_seen": 1947336704 + }, + { + "epoch": 0.59, + "learning_rate": 0.00033136208625877636, + "loss": 1.2871, + "theoretical_loss": 3.46651155495075, + "tokens_seen": 1947402240 + }, + { + "epoch": 0.59, + "learning_rate": 0.00033134603811434306, + "loss": 1.2475, + "theoretical_loss": 3.4665018576679802, + "tokens_seen": 1947467776 + }, + { + "epoch": 0.59, + "learning_rate": 0.00033132998996990976, + "loss": 1.3277, + "theoretical_loss": 3.466492160802907, + "tokens_seen": 1947533312 + }, + { + "epoch": 0.59, + "learning_rate": 0.00033131394182547646, + "loss": 1.3806, + "theoretical_loss": 3.466482464355498, + "tokens_seen": 1947598848 + }, + { + "epoch": 0.59, + "learning_rate": 0.00033129789368104316, + "loss": 1.3682, + "theoretical_loss": 3.4664727683257204, + "tokens_seen": 1947664384 + }, + { + "epoch": 0.59, + "learning_rate": 0.00033128184553660986, + "loss": 1.3166, + "theoretical_loss": 3.466463072713543, + "tokens_seen": 1947729920 + }, + { + "epoch": 0.59, + "learning_rate": 0.00033126579739217656, + "loss": 1.3357, + "theoretical_loss": 3.466453377518933, + "tokens_seen": 1947795456 + }, + { + "epoch": 0.59, + "learning_rate": 0.00033124974924774326, + "loss": 1.2928, + "theoretical_loss": 3.466443682741859, + "tokens_seen": 1947860992 + }, + { + "epoch": 0.59, + "learning_rate": 0.00033123370110330996, + "loss": 1.3628, + "theoretical_loss": 3.466433988382289, + "tokens_seen": 1947926528 + }, + { + "epoch": 0.59, + "learning_rate": 0.0003312176529588766, + "loss": 1.2923, + "theoretical_loss": 3.4664242944401904, + "tokens_seen": 1947992064 + }, + { + "epoch": 0.59, + "learning_rate": 0.0003312016048144433, + "loss": 1.3012, + "theoretical_loss": 3.466414600915532, + "tokens_seen": 1948057600 + }, + { + "epoch": 0.59, + "learning_rate": 0.00033118555667001, + "loss": 1.3858, + "theoretical_loss": 3.466404907808281, + "tokens_seen": 1948123136 + }, + { + "epoch": 0.59, + "learning_rate": 0.0003311695085255767, + "loss": 1.357, + "theoretical_loss": 3.466395215118406, + "tokens_seen": 1948188672 + }, + { + "epoch": 0.59, + "learning_rate": 0.0003311534603811434, + "loss": 1.4181, + "theoretical_loss": 3.466385522845874, + "tokens_seen": 1948254208 + }, + { + "epoch": 0.59, + "learning_rate": 0.0003311374122367101, + "loss": 1.341, + "theoretical_loss": 3.4663758309906543, + "tokens_seen": 1948319744 + }, + { + "epoch": 0.59, + "learning_rate": 0.0003311213640922768, + "loss": 1.3153, + "theoretical_loss": 3.466366139552714, + "tokens_seen": 1948385280 + }, + { + "epoch": 0.59, + "learning_rate": 0.0003311053159478435, + "loss": 1.3811, + "theoretical_loss": 3.4663564485320215, + "tokens_seen": 1948450816 + }, + { + "epoch": 0.59, + "learning_rate": 0.0003310892678034102, + "loss": 1.3731, + "theoretical_loss": 3.466346757928545, + "tokens_seen": 1948516352 + }, + { + "epoch": 0.59, + "learning_rate": 0.0003310732196589769, + "loss": 1.3102, + "theoretical_loss": 3.466337067742252, + "tokens_seen": 1948581888 + }, + { + "epoch": 0.59, + "learning_rate": 0.0003310571715145436, + "loss": 1.2387, + "theoretical_loss": 3.4663273779731103, + "tokens_seen": 1948647424 + }, + { + "epoch": 0.59, + "learning_rate": 0.0003310411233701103, + "loss": 1.4669, + "theoretical_loss": 3.4663176886210887, + "tokens_seen": 1948712960 + }, + { + "epoch": 0.59, + "learning_rate": 0.000331025075225677, + "loss": 1.3656, + "theoretical_loss": 3.4663079996861548, + "tokens_seen": 1948778496 + }, + { + "epoch": 0.59, + "learning_rate": 0.0003310090270812437, + "loss": 1.2339, + "theoretical_loss": 3.4662983111682766, + "tokens_seen": 1948844032 + }, + { + "epoch": 0.59, + "learning_rate": 0.0003309929789368104, + "loss": 1.3928, + "theoretical_loss": 3.466288623067422, + "tokens_seen": 1948909568 + }, + { + "epoch": 0.59, + "learning_rate": 0.0003309769307923771, + "loss": 1.4357, + "theoretical_loss": 3.4662789353835595, + "tokens_seen": 1948975104 + }, + { + "epoch": 0.59, + "learning_rate": 0.0003309608826479438, + "loss": 1.4151, + "theoretical_loss": 3.4662692481166566, + "tokens_seen": 1949040640 + }, + { + "epoch": 0.59, + "learning_rate": 0.0003309448345035105, + "loss": 1.2648, + "theoretical_loss": 3.466259561266682, + "tokens_seen": 1949106176 + }, + { + "epoch": 0.59, + "learning_rate": 0.0003309287863590773, + "loss": 1.2449, + "theoretical_loss": 3.466249874833603, + "tokens_seen": 1949171712 + }, + { + "epoch": 0.59, + "learning_rate": 0.000330912738214644, + "loss": 1.2905, + "theoretical_loss": 3.4662401888173875, + "tokens_seen": 1949237248 + }, + { + "epoch": 0.59, + "learning_rate": 0.0003308966900702107, + "loss": 1.2093, + "theoretical_loss": 3.4662305032180045, + "tokens_seen": 1949302784 + }, + { + "epoch": 0.59, + "learning_rate": 0.0003308806419257774, + "loss": 1.4546, + "theoretical_loss": 3.4662208180354215, + "tokens_seen": 1949368320 + }, + { + "epoch": 0.59, + "learning_rate": 0.0003308645937813441, + "loss": 1.3667, + "theoretical_loss": 3.4662111332696064, + "tokens_seen": 1949433856 + }, + { + "epoch": 0.59, + "learning_rate": 0.0003308485456369108, + "loss": 1.3274, + "theoretical_loss": 3.4662014489205277, + "tokens_seen": 1949499392 + }, + { + "epoch": 0.59, + "learning_rate": 0.0003308324974924775, + "loss": 1.2118, + "theoretical_loss": 3.466191764988153, + "tokens_seen": 1949564928 + }, + { + "epoch": 0.59, + "learning_rate": 0.0003308164493480442, + "loss": 1.4739, + "theoretical_loss": 3.466182081472451, + "tokens_seen": 1949630464 + }, + { + "epoch": 0.59, + "objective/train/docs_used": 710798, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.3615260124206543, + "objective/train/theoretical_loss": 3.4661723983733888, + "objective/train/tokens_used": 1970156000, + "theoretical_loss": 3.4661723983733888, + "tokens_seen": 1949696000 + }, + { + "epoch": 0.59, + "learning_rate": 0.0003308004012036109, + "loss": 1.3514, + "theoretical_loss": 3.4661723983733888, + "tokens_seen": 1949696000 + }, + { + "epoch": 0.59, + "learning_rate": 0.0003307843530591776, + "loss": 1.3816, + "theoretical_loss": 3.4661627156909356, + "tokens_seen": 1949761536 + }, + { + "epoch": 0.59, + "learning_rate": 0.0003307683049147443, + "loss": 1.3777, + "theoretical_loss": 3.466153033425058, + "tokens_seen": 1949827072 + }, + { + "epoch": 0.59, + "learning_rate": 0.000330752256770311, + "loss": 1.4468, + "theoretical_loss": 3.4661433515757256, + "tokens_seen": 1949892608 + }, + { + "epoch": 0.59, + "learning_rate": 0.0003307362086258777, + "loss": 1.287, + "theoretical_loss": 3.4661336701429057, + "tokens_seen": 1949958144 + }, + { + "epoch": 0.59, + "learning_rate": 0.0003307201604814444, + "loss": 1.3123, + "theoretical_loss": 3.4661239891265665, + "tokens_seen": 1950023680 + }, + { + "epoch": 0.59, + "learning_rate": 0.0003307041123370111, + "loss": 1.142, + "theoretical_loss": 3.4661143085266763, + "tokens_seen": 1950089216 + }, + { + "epoch": 0.59, + "learning_rate": 0.0003306880641925778, + "loss": 1.3905, + "theoretical_loss": 3.466104628343203, + "tokens_seen": 1950154752 + }, + { + "epoch": 0.59, + "learning_rate": 0.0003306720160481445, + "loss": 1.3069, + "theoretical_loss": 3.4660949485761146, + "tokens_seen": 1950220288 + }, + { + "epoch": 0.59, + "learning_rate": 0.0003306559679037112, + "loss": 1.3563, + "theoretical_loss": 3.466085269225379, + "tokens_seen": 1950285824 + }, + { + "epoch": 0.59, + "learning_rate": 0.0003306399197592779, + "loss": 1.3547, + "theoretical_loss": 3.4660755902909646, + "tokens_seen": 1950351360 + }, + { + "epoch": 0.59, + "learning_rate": 0.0003306238716148446, + "loss": 1.4487, + "theoretical_loss": 3.46606591177284, + "tokens_seen": 1950416896 + }, + { + "epoch": 0.59, + "learning_rate": 0.0003306078234704113, + "loss": 1.4133, + "theoretical_loss": 3.4660562336709724, + "tokens_seen": 1950482432 + }, + { + "epoch": 0.59, + "learning_rate": 0.00033059177532597793, + "loss": 1.3387, + "theoretical_loss": 3.4660465559853306, + "tokens_seen": 1950547968 + }, + { + "epoch": 0.59, + "learning_rate": 0.00033057572718154463, + "loss": 1.3308, + "theoretical_loss": 3.466036878715882, + "tokens_seen": 1950613504 + }, + { + "epoch": 0.59, + "learning_rate": 0.00033055967903711133, + "loss": 1.4054, + "theoretical_loss": 3.4660272018625955, + "tokens_seen": 1950679040 + }, + { + "epoch": 0.59, + "learning_rate": 0.00033054363089267803, + "loss": 1.2682, + "theoretical_loss": 3.466017525425439, + "tokens_seen": 1950744576 + }, + { + "epoch": 0.59, + "learning_rate": 0.00033052758274824474, + "loss": 1.282, + "theoretical_loss": 3.4660078494043804, + "tokens_seen": 1950810112 + }, + { + "epoch": 0.59, + "learning_rate": 0.00033051153460381144, + "loss": 1.2917, + "theoretical_loss": 3.465998173799388, + "tokens_seen": 1950875648 + }, + { + "epoch": 0.59, + "learning_rate": 0.00033049548645937814, + "loss": 1.3091, + "theoretical_loss": 3.4659884986104297, + "tokens_seen": 1950941184 + }, + { + "epoch": 0.59, + "learning_rate": 0.00033047943831494484, + "loss": 1.3178, + "theoretical_loss": 3.4659788238374736, + "tokens_seen": 1951006720 + }, + { + "epoch": 0.59, + "learning_rate": 0.00033046339017051154, + "loss": 1.4366, + "theoretical_loss": 3.4659691494804887, + "tokens_seen": 1951072256 + }, + { + "epoch": 0.59, + "learning_rate": 0.00033044734202607824, + "loss": 1.4231, + "theoretical_loss": 3.4659594755394423, + "tokens_seen": 1951137792 + }, + { + "epoch": 0.59, + "learning_rate": 0.00033043129388164494, + "loss": 1.3738, + "theoretical_loss": 3.4659498020143023, + "tokens_seen": 1951203328 + }, + { + "epoch": 0.59, + "learning_rate": 0.00033041524573721164, + "loss": 1.3794, + "theoretical_loss": 3.4659401289050376, + "tokens_seen": 1951268864 + }, + { + "epoch": 0.59, + "learning_rate": 0.00033039919759277834, + "loss": 1.2255, + "theoretical_loss": 3.4659304562116158, + "tokens_seen": 1951334400 + }, + { + "epoch": 0.59, + "learning_rate": 0.00033038314944834504, + "loss": 1.2764, + "theoretical_loss": 3.4659207839340054, + "tokens_seen": 1951399936 + }, + { + "epoch": 0.59, + "learning_rate": 0.00033036710130391174, + "loss": 1.2782, + "theoretical_loss": 3.465911112072175, + "tokens_seen": 1951465472 + }, + { + "epoch": 0.59, + "learning_rate": 0.00033035105315947844, + "loss": 1.3258, + "theoretical_loss": 3.4659014406260913, + "tokens_seen": 1951531008 + }, + { + "epoch": 0.59, + "learning_rate": 0.00033033500501504514, + "loss": 1.3254, + "theoretical_loss": 3.465891769595724, + "tokens_seen": 1951596544 + }, + { + "epoch": 0.59, + "learning_rate": 0.00033031895687061184, + "loss": 1.3663, + "theoretical_loss": 3.465882098981041, + "tokens_seen": 1951662080 + }, + { + "epoch": 0.59, + "learning_rate": 0.00033030290872617854, + "loss": 1.3886, + "theoretical_loss": 3.465872428782009, + "tokens_seen": 1951727616 + }, + { + "epoch": 0.59, + "learning_rate": 0.00033028686058174524, + "loss": 1.4467, + "theoretical_loss": 3.465862758998598, + "tokens_seen": 1951793152 + }, + { + "epoch": 0.59, + "learning_rate": 0.00033027081243731195, + "loss": 1.2927, + "theoretical_loss": 3.4658530896307758, + "tokens_seen": 1951858688 + }, + { + "epoch": 0.59, + "learning_rate": 0.00033025476429287865, + "loss": 1.2983, + "theoretical_loss": 3.46584342067851, + "tokens_seen": 1951924224 + }, + { + "epoch": 0.59, + "learning_rate": 0.00033023871614844535, + "loss": 1.296, + "theoretical_loss": 3.465833752141769, + "tokens_seen": 1951989760 + }, + { + "epoch": 0.59, + "learning_rate": 0.00033022266800401205, + "loss": 1.3516, + "theoretical_loss": 3.465824084020521, + "tokens_seen": 1952055296 + }, + { + "epoch": 0.59, + "learning_rate": 0.00033020661985957875, + "loss": 1.3272, + "theoretical_loss": 3.465814416314734, + "tokens_seen": 1952120832 + }, + { + "epoch": 0.59, + "learning_rate": 0.00033019057171514545, + "loss": 1.3116, + "theoretical_loss": 3.465804749024377, + "tokens_seen": 1952186368 + }, + { + "epoch": 0.59, + "learning_rate": 0.00033017452357071215, + "loss": 1.3137, + "theoretical_loss": 3.465795082149418, + "tokens_seen": 1952251904 + }, + { + "epoch": 0.59, + "learning_rate": 0.00033015847542627885, + "loss": 1.3638, + "theoretical_loss": 3.4657854156898242, + "tokens_seen": 1952317440 + }, + { + "epoch": 0.59, + "learning_rate": 0.00033014242728184555, + "loss": 1.4668, + "theoretical_loss": 3.4657757496455646, + "tokens_seen": 1952382976 + }, + { + "epoch": 0.59, + "learning_rate": 0.00033012637913741225, + "loss": 1.393, + "theoretical_loss": 3.4657660840166074, + "tokens_seen": 1952448512 + }, + { + "epoch": 0.59, + "learning_rate": 0.00033011033099297895, + "loss": 1.3359, + "theoretical_loss": 3.4657564188029206, + "tokens_seen": 1952514048 + }, + { + "epoch": 0.59, + "learning_rate": 0.00033009428284854565, + "loss": 1.3788, + "theoretical_loss": 3.465746754004473, + "tokens_seen": 1952579584 + }, + { + "epoch": 0.59, + "learning_rate": 0.00033007823470411235, + "loss": 1.3614, + "theoretical_loss": 3.465737089621232, + "tokens_seen": 1952645120 + }, + { + "epoch": 0.59, + "learning_rate": 0.00033006218655967905, + "loss": 1.3281, + "theoretical_loss": 3.4657274256531663, + "tokens_seen": 1952710656 + }, + { + "epoch": 0.59, + "learning_rate": 0.00033004613841524575, + "loss": 1.2929, + "theoretical_loss": 3.465717762100244, + "tokens_seen": 1952776192 + }, + { + "epoch": 0.59, + "learning_rate": 0.00033003009027081245, + "loss": 1.3325, + "theoretical_loss": 3.465708098962433, + "tokens_seen": 1952841728 + }, + { + "epoch": 0.59, + "learning_rate": 0.00033001404212637916, + "loss": 1.5077, + "theoretical_loss": 3.465698436239702, + "tokens_seen": 1952907264 + }, + { + "epoch": 0.59, + "objective/train/docs_used": 712021, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.3029097318649292, + "objective/train/theoretical_loss": 3.4656887739320195, + "objective/train/tokens_used": 1973432800, + "theoretical_loss": 3.4656887739320195, + "tokens_seen": 1952972800 + }, + { + "epoch": 0.59, + "learning_rate": 0.00032999799398194586, + "loss": 1.285, + "theoretical_loss": 3.4656887739320195, + "tokens_seen": 1952972800 + }, + { + "epoch": 0.59, + "learning_rate": 0.00032998194583751256, + "loss": 1.374, + "theoretical_loss": 3.465679112039353, + "tokens_seen": 1953038336 + }, + { + "epoch": 0.59, + "learning_rate": 0.00032996589769307926, + "loss": 1.321, + "theoretical_loss": 3.4656694505616716, + "tokens_seen": 1953103872 + }, + { + "epoch": 0.59, + "learning_rate": 0.00032994984954864596, + "loss": 1.3027, + "theoretical_loss": 3.4656597894989427, + "tokens_seen": 1953169408 + }, + { + "epoch": 0.59, + "learning_rate": 0.00032993380140421266, + "loss": 1.3181, + "theoretical_loss": 3.465650128851135, + "tokens_seen": 1953234944 + }, + { + "epoch": 0.59, + "learning_rate": 0.00032991775325977936, + "loss": 1.2732, + "theoretical_loss": 3.4656404686182167, + "tokens_seen": 1953300480 + }, + { + "epoch": 0.59, + "learning_rate": 0.00032990170511534606, + "loss": 1.3708, + "theoretical_loss": 3.465630808800156, + "tokens_seen": 1953366016 + }, + { + "epoch": 0.59, + "learning_rate": 0.00032988565697091276, + "loss": 1.2528, + "theoretical_loss": 3.4656211493969216, + "tokens_seen": 1953431552 + }, + { + "epoch": 0.59, + "learning_rate": 0.00032986960882647946, + "loss": 1.3267, + "theoretical_loss": 3.465611490408481, + "tokens_seen": 1953497088 + }, + { + "epoch": 0.59, + "learning_rate": 0.00032985356068204616, + "loss": 1.3939, + "theoretical_loss": 3.465601831834803, + "tokens_seen": 1953562624 + }, + { + "epoch": 0.59, + "learning_rate": 0.00032983751253761286, + "loss": 1.3873, + "theoretical_loss": 3.465592173675856, + "tokens_seen": 1953628160 + }, + { + "epoch": 0.59, + "learning_rate": 0.0003298214643931795, + "loss": 1.3165, + "theoretical_loss": 3.465582515931608, + "tokens_seen": 1953693696 + }, + { + "epoch": 0.59, + "learning_rate": 0.0003298054162487462, + "loss": 1.3679, + "theoretical_loss": 3.4655728586020267, + "tokens_seen": 1953759232 + }, + { + "epoch": 0.59, + "learning_rate": 0.0003297893681043129, + "loss": 1.4369, + "theoretical_loss": 3.4655632016870817, + "tokens_seen": 1953824768 + }, + { + "epoch": 0.59, + "learning_rate": 0.0003297733199598796, + "loss": 1.2661, + "theoretical_loss": 3.46555354518674, + "tokens_seen": 1953890304 + }, + { + "epoch": 0.59, + "learning_rate": 0.0003297572718154463, + "loss": 1.3909, + "theoretical_loss": 3.465543889100971, + "tokens_seen": 1953955840 + }, + { + "epoch": 0.59, + "learning_rate": 0.000329741223671013, + "loss": 1.4307, + "theoretical_loss": 3.465534233429742, + "tokens_seen": 1954021376 + }, + { + "epoch": 0.59, + "learning_rate": 0.00032972517552657977, + "loss": 1.2781, + "theoretical_loss": 3.4655245781730226, + "tokens_seen": 1954086912 + }, + { + "epoch": 0.59, + "learning_rate": 0.00032970912738214647, + "loss": 1.3785, + "theoretical_loss": 3.4655149233307796, + "tokens_seen": 1954152448 + }, + { + "epoch": 0.59, + "learning_rate": 0.00032969307923771317, + "loss": 1.4484, + "theoretical_loss": 3.465505268902982, + "tokens_seen": 1954217984 + }, + { + "epoch": 0.59, + "learning_rate": 0.00032967703109327987, + "loss": 1.2934, + "theoretical_loss": 3.4654956148895986, + "tokens_seen": 1954283520 + }, + { + "epoch": 0.59, + "learning_rate": 0.00032966098294884657, + "loss": 1.2658, + "theoretical_loss": 3.4654859612905966, + "tokens_seen": 1954349056 + }, + { + "epoch": 0.59, + "learning_rate": 0.00032964493480441327, + "loss": 1.3655, + "theoretical_loss": 3.465476308105945, + "tokens_seen": 1954414592 + }, + { + "epoch": 0.59, + "learning_rate": 0.00032962888665997997, + "loss": 1.3482, + "theoretical_loss": 3.4654666553356126, + "tokens_seen": 1954480128 + }, + { + "epoch": 0.59, + "learning_rate": 0.00032961283851554667, + "loss": 1.4062, + "theoretical_loss": 3.4654570029795666, + "tokens_seen": 1954545664 + }, + { + "epoch": 0.59, + "learning_rate": 0.00032959679037111337, + "loss": 1.3025, + "theoretical_loss": 3.465447351037776, + "tokens_seen": 1954611200 + }, + { + "epoch": 0.59, + "learning_rate": 0.00032958074222668007, + "loss": 1.3776, + "theoretical_loss": 3.4654376995102094, + "tokens_seen": 1954676736 + }, + { + "epoch": 0.59, + "learning_rate": 0.0003295646940822468, + "loss": 1.4673, + "theoretical_loss": 3.4654280483968347, + "tokens_seen": 1954742272 + }, + { + "epoch": 0.59, + "learning_rate": 0.0003295486459378135, + "loss": 1.3301, + "theoretical_loss": 3.4654183976976203, + "tokens_seen": 1954807808 + }, + { + "epoch": 0.59, + "learning_rate": 0.0003295325977933802, + "loss": 1.3044, + "theoretical_loss": 3.4654087474125346, + "tokens_seen": 1954873344 + }, + { + "epoch": 0.59, + "learning_rate": 0.0003295165496489469, + "loss": 1.4279, + "theoretical_loss": 3.4653990975415456, + "tokens_seen": 1954938880 + }, + { + "epoch": 0.59, + "learning_rate": 0.0003295005015045136, + "loss": 1.374, + "theoretical_loss": 3.4653894480846223, + "tokens_seen": 1955004416 + }, + { + "epoch": 0.59, + "learning_rate": 0.0003294844533600803, + "loss": 1.3472, + "theoretical_loss": 3.4653797990417328, + "tokens_seen": 1955069952 + }, + { + "epoch": 0.59, + "learning_rate": 0.000329468405215647, + "loss": 1.3438, + "theoretical_loss": 3.465370150412845, + "tokens_seen": 1955135488 + }, + { + "epoch": 0.59, + "learning_rate": 0.0003294523570712137, + "loss": 1.3947, + "theoretical_loss": 3.465360502197928, + "tokens_seen": 1955201024 + }, + { + "epoch": 0.59, + "learning_rate": 0.0003294363089267804, + "loss": 1.3431, + "theoretical_loss": 3.4653508543969496, + "tokens_seen": 1955266560 + }, + { + "epoch": 0.59, + "learning_rate": 0.0003294202607823471, + "loss": 1.3202, + "theoretical_loss": 3.465341207009878, + "tokens_seen": 1955332096 + }, + { + "epoch": 0.59, + "learning_rate": 0.0003294042126379138, + "loss": 1.4321, + "theoretical_loss": 3.4653315600366827, + "tokens_seen": 1955397632 + }, + { + "epoch": 0.59, + "learning_rate": 0.0003293881644934805, + "loss": 1.3793, + "theoretical_loss": 3.465321913477331, + "tokens_seen": 1955463168 + }, + { + "epoch": 0.59, + "learning_rate": 0.0003293721163490472, + "loss": 1.3777, + "theoretical_loss": 3.4653122673317913, + "tokens_seen": 1955528704 + }, + { + "epoch": 0.59, + "learning_rate": 0.0003293560682046139, + "loss": 1.3339, + "theoretical_loss": 3.4653026216000327, + "tokens_seen": 1955594240 + }, + { + "epoch": 0.59, + "learning_rate": 0.0003293400200601806, + "loss": 1.4501, + "theoretical_loss": 3.465292976282023, + "tokens_seen": 1955659776 + }, + { + "epoch": 0.59, + "learning_rate": 0.0003293239719157473, + "loss": 1.2681, + "theoretical_loss": 3.465283331377731, + "tokens_seen": 1955725312 + }, + { + "epoch": 0.59, + "learning_rate": 0.000329307923771314, + "loss": 1.3202, + "theoretical_loss": 3.465273686887124, + "tokens_seen": 1955790848 + }, + { + "epoch": 0.59, + "learning_rate": 0.0003292918756268807, + "loss": 1.3392, + "theoretical_loss": 3.4652640428101718, + "tokens_seen": 1955856384 + }, + { + "epoch": 0.59, + "learning_rate": 0.0003292758274824474, + "loss": 1.3683, + "theoretical_loss": 3.4652543991468425, + "tokens_seen": 1955921920 + }, + { + "epoch": 0.59, + "learning_rate": 0.0003292597793380141, + "loss": 1.3496, + "theoretical_loss": 3.4652447558971042, + "tokens_seen": 1955987456 + }, + { + "epoch": 0.59, + "learning_rate": 0.0003292437311935808, + "loss": 1.3503, + "theoretical_loss": 3.465235113060925, + "tokens_seen": 1956052992 + }, + { + "epoch": 0.59, + "learning_rate": 0.0003292276830491475, + "loss": 1.3564, + "theoretical_loss": 3.4652254706382735, + "tokens_seen": 1956118528 + }, + { + "epoch": 0.59, + "learning_rate": 0.0003292116349047142, + "loss": 1.3751, + "theoretical_loss": 3.4652158286291184, + "tokens_seen": 1956184064 + }, + { + "epoch": 0.59, + "objective/train/docs_used": 713197, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.291121244430542, + "objective/train/theoretical_loss": 3.465206187033428, + "objective/train/tokens_used": 1976709600, + "theoretical_loss": 3.465206187033428, + "tokens_seen": 1956249600 + }, + { + "epoch": 0.59, + "learning_rate": 0.0003291955867602809, + "loss": 1.3107, + "theoretical_loss": 3.465206187033428, + "tokens_seen": 1956249600 + }, + { + "epoch": 0.59, + "learning_rate": 0.00032917953861584753, + "loss": 1.3776, + "theoretical_loss": 3.465196545851171, + "tokens_seen": 1956315136 + }, + { + "epoch": 0.59, + "learning_rate": 0.00032916349047141423, + "loss": 1.3894, + "theoretical_loss": 3.465186905082315, + "tokens_seen": 1956380672 + }, + { + "epoch": 0.59, + "learning_rate": 0.00032914744232698094, + "loss": 1.3929, + "theoretical_loss": 3.4651772647268295, + "tokens_seen": 1956446208 + }, + { + "epoch": 0.59, + "learning_rate": 0.00032913139418254764, + "loss": 1.2254, + "theoretical_loss": 3.465167624784682, + "tokens_seen": 1956511744 + }, + { + "epoch": 0.59, + "learning_rate": 0.00032911534603811434, + "loss": 1.4819, + "theoretical_loss": 3.4651579852558414, + "tokens_seen": 1956577280 + }, + { + "epoch": 0.59, + "learning_rate": 0.00032909929789368104, + "loss": 1.3671, + "theoretical_loss": 3.465148346140276, + "tokens_seen": 1956642816 + }, + { + "epoch": 0.59, + "learning_rate": 0.00032908324974924774, + "loss": 1.2881, + "theoretical_loss": 3.4651387074379545, + "tokens_seen": 1956708352 + }, + { + "epoch": 0.59, + "learning_rate": 0.00032906720160481444, + "loss": 1.3389, + "theoretical_loss": 3.465129069148845, + "tokens_seen": 1956773888 + }, + { + "epoch": 0.59, + "learning_rate": 0.00032905115346038114, + "loss": 1.4365, + "theoretical_loss": 3.4651194312729157, + "tokens_seen": 1956839424 + }, + { + "epoch": 0.59, + "learning_rate": 0.00032903510531594784, + "loss": 1.2977, + "theoretical_loss": 3.465109793810136, + "tokens_seen": 1956904960 + }, + { + "epoch": 0.59, + "learning_rate": 0.00032901905717151454, + "loss": 1.3908, + "theoretical_loss": 3.4651001567604736, + "tokens_seen": 1956970496 + }, + { + "epoch": 0.59, + "learning_rate": 0.00032900300902708124, + "loss": 1.3899, + "theoretical_loss": 3.465090520123897, + "tokens_seen": 1957036032 + }, + { + "epoch": 0.59, + "learning_rate": 0.00032898696088264794, + "loss": 1.4256, + "theoretical_loss": 3.465080883900375, + "tokens_seen": 1957101568 + }, + { + "epoch": 0.59, + "learning_rate": 0.00032897091273821464, + "loss": 1.3159, + "theoretical_loss": 3.4650712480898758, + "tokens_seen": 1957167104 + }, + { + "epoch": 0.59, + "learning_rate": 0.00032895486459378134, + "loss": 1.3318, + "theoretical_loss": 3.4650616126923675, + "tokens_seen": 1957232640 + }, + { + "epoch": 0.59, + "learning_rate": 0.00032893881644934804, + "loss": 1.3868, + "theoretical_loss": 3.4650519777078195, + "tokens_seen": 1957298176 + }, + { + "epoch": 0.59, + "learning_rate": 0.00032892276830491474, + "loss": 1.2627, + "theoretical_loss": 3.4650423431362, + "tokens_seen": 1957363712 + }, + { + "epoch": 0.59, + "learning_rate": 0.00032890672016048145, + "loss": 1.2821, + "theoretical_loss": 3.4650327089774766, + "tokens_seen": 1957429248 + }, + { + "epoch": 0.59, + "learning_rate": 0.00032889067201604815, + "loss": 1.5082, + "theoretical_loss": 3.4650230752316187, + "tokens_seen": 1957494784 + }, + { + "epoch": 0.59, + "learning_rate": 0.00032887462387161485, + "loss": 1.3992, + "theoretical_loss": 3.465013441898595, + "tokens_seen": 1957560320 + }, + { + "epoch": 0.59, + "learning_rate": 0.00032885857572718155, + "loss": 1.4061, + "theoretical_loss": 3.4650038089783726, + "tokens_seen": 1957625856 + }, + { + "epoch": 0.59, + "learning_rate": 0.00032884252758274825, + "loss": 1.3579, + "theoretical_loss": 3.4649941764709213, + "tokens_seen": 1957691392 + }, + { + "epoch": 0.59, + "learning_rate": 0.00032882647943831495, + "loss": 1.4746, + "theoretical_loss": 3.4649845443762093, + "tokens_seen": 1957756928 + }, + { + "epoch": 0.59, + "learning_rate": 0.00032881043129388165, + "loss": 1.4397, + "theoretical_loss": 3.4649749126942053, + "tokens_seen": 1957822464 + }, + { + "epoch": 0.59, + "learning_rate": 0.00032879438314944835, + "loss": 1.3328, + "theoretical_loss": 3.4649652814248775, + "tokens_seen": 1957888000 + }, + { + "epoch": 0.59, + "learning_rate": 0.00032877833500501505, + "loss": 1.2934, + "theoretical_loss": 3.4649556505681938, + "tokens_seen": 1957953536 + }, + { + "epoch": 0.59, + "learning_rate": 0.00032876228686058175, + "loss": 1.4069, + "theoretical_loss": 3.4649460201241236, + "tokens_seen": 1958019072 + }, + { + "epoch": 0.59, + "learning_rate": 0.00032874623871614845, + "loss": 1.3118, + "theoretical_loss": 3.4649363900926353, + "tokens_seen": 1958084608 + }, + { + "epoch": 0.59, + "learning_rate": 0.00032873019057171515, + "loss": 1.3073, + "theoretical_loss": 3.4649267604736966, + "tokens_seen": 1958150144 + }, + { + "epoch": 0.59, + "learning_rate": 0.00032871414242728185, + "loss": 1.3841, + "theoretical_loss": 3.4649171312672777, + "tokens_seen": 1958215680 + }, + { + "epoch": 0.59, + "learning_rate": 0.00032869809428284855, + "loss": 1.3329, + "theoretical_loss": 3.4649075024733453, + "tokens_seen": 1958281216 + }, + { + "epoch": 0.59, + "learning_rate": 0.00032868204613841525, + "loss": 1.3286, + "theoretical_loss": 3.464897874091869, + "tokens_seen": 1958346752 + }, + { + "epoch": 0.59, + "learning_rate": 0.00032866599799398195, + "loss": 1.2808, + "theoretical_loss": 3.4648882461228174, + "tokens_seen": 1958412288 + }, + { + "epoch": 0.59, + "learning_rate": 0.00032864994984954866, + "loss": 1.3523, + "theoretical_loss": 3.464878618566158, + "tokens_seen": 1958477824 + }, + { + "epoch": 0.59, + "learning_rate": 0.00032863390170511536, + "loss": 1.3392, + "theoretical_loss": 3.4648689914218602, + "tokens_seen": 1958543360 + }, + { + "epoch": 0.59, + "learning_rate": 0.00032861785356068206, + "loss": 1.4057, + "theoretical_loss": 3.4648593646898926, + "tokens_seen": 1958608896 + }, + { + "epoch": 0.59, + "learning_rate": 0.00032860180541624876, + "loss": 1.3355, + "theoretical_loss": 3.4648497383702237, + "tokens_seen": 1958674432 + }, + { + "epoch": 0.59, + "learning_rate": 0.00032858575727181546, + "loss": 1.4527, + "theoretical_loss": 3.464840112462821, + "tokens_seen": 1958739968 + }, + { + "epoch": 0.59, + "learning_rate": 0.00032856970912738216, + "loss": 1.2634, + "theoretical_loss": 3.4648304869676547, + "tokens_seen": 1958805504 + }, + { + "epoch": 0.59, + "learning_rate": 0.00032855366098294886, + "loss": 1.3501, + "theoretical_loss": 3.464820861884692, + "tokens_seen": 1958871040 + }, + { + "epoch": 0.59, + "learning_rate": 0.00032853761283851556, + "loss": 1.3376, + "theoretical_loss": 3.464811237213903, + "tokens_seen": 1958936576 + }, + { + "epoch": 0.59, + "learning_rate": 0.00032852156469408226, + "loss": 1.3775, + "theoretical_loss": 3.464801612955254, + "tokens_seen": 1959002112 + }, + { + "epoch": 0.59, + "learning_rate": 0.00032850551654964896, + "loss": 1.3492, + "theoretical_loss": 3.4647919891087158, + "tokens_seen": 1959067648 + }, + { + "epoch": 0.59, + "learning_rate": 0.00032848946840521566, + "loss": 1.3888, + "theoretical_loss": 3.4647823656742553, + "tokens_seen": 1959133184 + }, + { + "epoch": 0.59, + "learning_rate": 0.00032847342026078236, + "loss": 1.366, + "theoretical_loss": 3.464772742651842, + "tokens_seen": 1959198720 + }, + { + "epoch": 0.59, + "learning_rate": 0.00032845737211634906, + "loss": 1.3763, + "theoretical_loss": 3.464763120041445, + "tokens_seen": 1959264256 + }, + { + "epoch": 0.59, + "learning_rate": 0.00032844132397191576, + "loss": 1.301, + "theoretical_loss": 3.4647534978430308, + "tokens_seen": 1959329792 + }, + { + "epoch": 0.59, + "learning_rate": 0.00032842527582748246, + "loss": 1.3663, + "theoretical_loss": 3.4647438760565703, + "tokens_seen": 1959395328 + }, + { + "epoch": 0.59, + "learning_rate": 0.00032840922768304916, + "loss": 1.2831, + "theoretical_loss": 3.464734254682031, + "tokens_seen": 1959460864 + }, + { + "epoch": 0.59, + "objective/train/docs_used": 714425, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.4263639450073242, + "objective/train/theoretical_loss": 3.4647246337193813, + "objective/train/tokens_used": 1979986400, + "theoretical_loss": 3.4647246337193813, + "tokens_seen": 1959526400 + }, + { + "epoch": 0.59, + "learning_rate": 0.00032839317953861587, + "loss": 1.4035, + "theoretical_loss": 3.4647246337193813, + "tokens_seen": 1959526400 + }, + { + "epoch": 0.59, + "learning_rate": 0.00032837713139418257, + "loss": 1.3873, + "theoretical_loss": 3.4647150131685898, + "tokens_seen": 1959591936 + }, + { + "epoch": 0.59, + "learning_rate": 0.00032836108324974927, + "loss": 1.3399, + "theoretical_loss": 3.464705393029626, + "tokens_seen": 1959657472 + }, + { + "epoch": 0.59, + "learning_rate": 0.00032834503510531597, + "loss": 1.396, + "theoretical_loss": 3.4646957733024575, + "tokens_seen": 1959723008 + }, + { + "epoch": 0.59, + "learning_rate": 0.00032832898696088267, + "loss": 1.3669, + "theoretical_loss": 3.4646861539870537, + "tokens_seen": 1959788544 + }, + { + "epoch": 0.59, + "learning_rate": 0.00032831293881644937, + "loss": 1.342, + "theoretical_loss": 3.4646765350833824, + "tokens_seen": 1959854080 + }, + { + "epoch": 0.59, + "learning_rate": 0.00032829689067201607, + "loss": 1.2904, + "theoretical_loss": 3.464666916591413, + "tokens_seen": 1959919616 + }, + { + "epoch": 0.59, + "learning_rate": 0.00032828084252758277, + "loss": 1.2637, + "theoretical_loss": 3.4646572985111126, + "tokens_seen": 1959985152 + }, + { + "epoch": 0.59, + "learning_rate": 0.00032826479438314947, + "loss": 1.4261, + "theoretical_loss": 3.464647680842452, + "tokens_seen": 1960050688 + }, + { + "epoch": 0.59, + "learning_rate": 0.00032824874623871617, + "loss": 1.2832, + "theoretical_loss": 3.464638063585398, + "tokens_seen": 1960116224 + }, + { + "epoch": 0.59, + "learning_rate": 0.00032823269809428287, + "loss": 1.3714, + "theoretical_loss": 3.464628446739921, + "tokens_seen": 1960181760 + }, + { + "epoch": 0.59, + "learning_rate": 0.00032821664994984957, + "loss": 1.4204, + "theoretical_loss": 3.464618830305988, + "tokens_seen": 1960247296 + }, + { + "epoch": 0.59, + "learning_rate": 0.0003282006018054163, + "loss": 1.4332, + "theoretical_loss": 3.464609214283568, + "tokens_seen": 1960312832 + }, + { + "epoch": 0.59, + "learning_rate": 0.000328184553660983, + "loss": 1.3643, + "theoretical_loss": 3.4645995986726303, + "tokens_seen": 1960378368 + }, + { + "epoch": 0.59, + "learning_rate": 0.0003281685055165497, + "loss": 1.3866, + "theoretical_loss": 3.464589983473143, + "tokens_seen": 1960443904 + }, + { + "epoch": 0.59, + "learning_rate": 0.0003281524573721164, + "loss": 1.4486, + "theoretical_loss": 3.4645803686850747, + "tokens_seen": 1960509440 + }, + { + "epoch": 0.59, + "learning_rate": 0.0003281364092276831, + "loss": 1.3864, + "theoretical_loss": 3.464570754308394, + "tokens_seen": 1960574976 + }, + { + "epoch": 0.59, + "learning_rate": 0.0003281203610832498, + "loss": 1.2781, + "theoretical_loss": 3.4645611403430703, + "tokens_seen": 1960640512 + }, + { + "epoch": 0.59, + "learning_rate": 0.0003281043129388165, + "loss": 1.5177, + "theoretical_loss": 3.464551526789071, + "tokens_seen": 1960706048 + }, + { + "epoch": 0.59, + "learning_rate": 0.0003280882647943832, + "loss": 1.3544, + "theoretical_loss": 3.464541913646366, + "tokens_seen": 1960771584 + }, + { + "epoch": 0.59, + "learning_rate": 0.0003280722166499499, + "loss": 1.3982, + "theoretical_loss": 3.4645323009149234, + "tokens_seen": 1960837120 + }, + { + "epoch": 0.59, + "learning_rate": 0.0003280561685055166, + "loss": 1.3439, + "theoretical_loss": 3.4645226885947116, + "tokens_seen": 1960902656 + }, + { + "epoch": 0.59, + "learning_rate": 0.0003280401203610833, + "loss": 1.4295, + "theoretical_loss": 3.4645130766856993, + "tokens_seen": 1960968192 + }, + { + "epoch": 0.59, + "learning_rate": 0.00032802407221665, + "loss": 1.3088, + "theoretical_loss": 3.464503465187856, + "tokens_seen": 1961033728 + }, + { + "epoch": 0.59, + "learning_rate": 0.0003280080240722167, + "loss": 1.3854, + "theoretical_loss": 3.464493854101149, + "tokens_seen": 1961099264 + }, + { + "epoch": 0.59, + "learning_rate": 0.0003279919759277834, + "loss": 1.3472, + "theoretical_loss": 3.464484243425548, + "tokens_seen": 1961164800 + }, + { + "epoch": 0.59, + "learning_rate": 0.0003279759277833501, + "loss": 1.4641, + "theoretical_loss": 3.4644746331610214, + "tokens_seen": 1961230336 + }, + { + "epoch": 0.59, + "learning_rate": 0.0003279598796389168, + "loss": 1.3501, + "theoretical_loss": 3.464465023307538, + "tokens_seen": 1961295872 + }, + { + "epoch": 0.59, + "learning_rate": 0.0003279438314944835, + "loss": 1.3078, + "theoretical_loss": 3.4644554138650667, + "tokens_seen": 1961361408 + }, + { + "epoch": 0.59, + "learning_rate": 0.0003279277833500502, + "loss": 1.3482, + "theoretical_loss": 3.4644458048335753, + "tokens_seen": 1961426944 + }, + { + "epoch": 0.59, + "learning_rate": 0.0003279117352056169, + "loss": 1.3391, + "theoretical_loss": 3.4644361962130334, + "tokens_seen": 1961492480 + }, + { + "epoch": 0.59, + "learning_rate": 0.0003278956870611836, + "loss": 1.4321, + "theoretical_loss": 3.464426588003409, + "tokens_seen": 1961558016 + }, + { + "epoch": 0.59, + "learning_rate": 0.0003278796389167503, + "loss": 1.405, + "theoretical_loss": 3.4644169802046716, + "tokens_seen": 1961623552 + }, + { + "epoch": 0.59, + "learning_rate": 0.000327863590772317, + "loss": 1.4236, + "theoretical_loss": 3.4644073728167895, + "tokens_seen": 1961689088 + }, + { + "epoch": 0.59, + "learning_rate": 0.0003278475426278837, + "loss": 1.3147, + "theoretical_loss": 3.4643977658397307, + "tokens_seen": 1961754624 + }, + { + "epoch": 0.59, + "learning_rate": 0.0003278314944834504, + "loss": 1.2995, + "theoretical_loss": 3.464388159273465, + "tokens_seen": 1961820160 + }, + { + "epoch": 0.59, + "learning_rate": 0.0003278154463390171, + "loss": 1.4159, + "theoretical_loss": 3.4643785531179607, + "tokens_seen": 1961885696 + }, + { + "epoch": 0.59, + "learning_rate": 0.0003277993981945838, + "loss": 1.3528, + "theoretical_loss": 3.4643689473731865, + "tokens_seen": 1961951232 + }, + { + "epoch": 0.59, + "learning_rate": 0.00032778335005015044, + "loss": 1.252, + "theoretical_loss": 3.464359342039111, + "tokens_seen": 1962016768 + }, + { + "epoch": 0.59, + "learning_rate": 0.00032776730190571714, + "loss": 1.2942, + "theoretical_loss": 3.464349737115703, + "tokens_seen": 1962082304 + }, + { + "epoch": 0.59, + "learning_rate": 0.00032775125376128384, + "loss": 1.3317, + "theoretical_loss": 3.4643401326029313, + "tokens_seen": 1962147840 + }, + { + "epoch": 0.59, + "learning_rate": 0.00032773520561685054, + "loss": 1.2843, + "theoretical_loss": 3.464330528500765, + "tokens_seen": 1962213376 + }, + { + "epoch": 0.59, + "learning_rate": 0.00032771915747241724, + "loss": 1.3514, + "theoretical_loss": 3.464320924809172, + "tokens_seen": 1962278912 + }, + { + "epoch": 0.59, + "learning_rate": 0.00032770310932798394, + "loss": 1.337, + "theoretical_loss": 3.464311321528122, + "tokens_seen": 1962344448 + }, + { + "epoch": 0.59, + "learning_rate": 0.00032768706118355064, + "loss": 1.3722, + "theoretical_loss": 3.464301718657582, + "tokens_seen": 1962409984 + }, + { + "epoch": 0.59, + "learning_rate": 0.00032767101303911734, + "loss": 1.3896, + "theoretical_loss": 3.464292116197523, + "tokens_seen": 1962475520 + }, + { + "epoch": 0.59, + "learning_rate": 0.00032765496489468404, + "loss": 1.3831, + "theoretical_loss": 3.4642825141479126, + "tokens_seen": 1962541056 + }, + { + "epoch": 0.59, + "learning_rate": 0.00032763891675025074, + "loss": 1.3213, + "theoretical_loss": 3.4642729125087195, + "tokens_seen": 1962606592 + }, + { + "epoch": 0.59, + "learning_rate": 0.00032762286860581744, + "loss": 1.3578, + "theoretical_loss": 3.464263311279912, + "tokens_seen": 1962672128 + }, + { + "epoch": 0.59, + "learning_rate": 0.00032760682046138414, + "loss": 1.3682, + "theoretical_loss": 3.4642537104614606, + "tokens_seen": 1962737664 + }, + { + "epoch": 0.59, + "objective/train/docs_used": 715564, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.2119592428207397, + "objective/train/theoretical_loss": 3.464244110053332, + "objective/train/tokens_used": 1983263200, + "theoretical_loss": 3.464244110053332, + "tokens_seen": 1962803200 + }, + { + "epoch": 0.59, + "learning_rate": 0.00032759077231695084, + "loss": 1.3384, + "theoretical_loss": 3.464244110053332, + "tokens_seen": 1962803200 + }, + { + "epoch": 0.59, + "learning_rate": 0.00032757472417251754, + "loss": 1.425, + "theoretical_loss": 3.4642345100554963, + "tokens_seen": 1962868736 + }, + { + "epoch": 0.59, + "learning_rate": 0.00032755867602808424, + "loss": 1.3572, + "theoretical_loss": 3.4642249104679217, + "tokens_seen": 1962934272 + }, + { + "epoch": 0.59, + "learning_rate": 0.00032754262788365094, + "loss": 1.3418, + "theoretical_loss": 3.464215311290577, + "tokens_seen": 1962999808 + }, + { + "epoch": 0.59, + "learning_rate": 0.00032752657973921765, + "loss": 1.2961, + "theoretical_loss": 3.4642057125234316, + "tokens_seen": 1963065344 + }, + { + "epoch": 0.59, + "learning_rate": 0.00032751053159478435, + "loss": 1.4627, + "theoretical_loss": 3.4641961141664535, + "tokens_seen": 1963130880 + }, + { + "epoch": 0.59, + "learning_rate": 0.00032749448345035105, + "loss": 1.4435, + "theoretical_loss": 3.4641865162196117, + "tokens_seen": 1963196416 + }, + { + "epoch": 0.59, + "learning_rate": 0.00032747843530591775, + "loss": 1.4362, + "theoretical_loss": 3.464176918682875, + "tokens_seen": 1963261952 + }, + { + "epoch": 0.59, + "learning_rate": 0.00032746238716148445, + "loss": 1.3862, + "theoretical_loss": 3.464167321556212, + "tokens_seen": 1963327488 + }, + { + "epoch": 0.59, + "learning_rate": 0.00032744633901705115, + "loss": 1.3815, + "theoretical_loss": 3.464157724839592, + "tokens_seen": 1963393024 + }, + { + "epoch": 0.59, + "learning_rate": 0.00032743029087261785, + "loss": 1.4079, + "theoretical_loss": 3.4641481285329836, + "tokens_seen": 1963458560 + }, + { + "epoch": 0.6, + "learning_rate": 0.00032741424272818455, + "loss": 1.3896, + "theoretical_loss": 3.4641385326363556, + "tokens_seen": 1963524096 + }, + { + "epoch": 0.6, + "learning_rate": 0.00032739819458375125, + "loss": 1.3724, + "theoretical_loss": 3.4641289371496766, + "tokens_seen": 1963589632 + }, + { + "epoch": 0.6, + "learning_rate": 0.00032738214643931795, + "loss": 1.4214, + "theoretical_loss": 3.464119342072915, + "tokens_seen": 1963655168 + }, + { + "epoch": 0.6, + "learning_rate": 0.00032736609829488465, + "loss": 1.3618, + "theoretical_loss": 3.4641097474060407, + "tokens_seen": 1963720704 + }, + { + "epoch": 0.6, + "learning_rate": 0.00032735005015045135, + "loss": 1.2906, + "theoretical_loss": 3.4641001531490216, + "tokens_seen": 1963786240 + }, + { + "epoch": 0.6, + "learning_rate": 0.00032733400200601805, + "loss": 1.3001, + "theoretical_loss": 3.464090559301827, + "tokens_seen": 1963851776 + }, + { + "epoch": 0.6, + "learning_rate": 0.0003273179538615848, + "loss": 1.354, + "theoretical_loss": 3.464080965864426, + "tokens_seen": 1963917312 + }, + { + "epoch": 0.6, + "learning_rate": 0.0003273019057171515, + "loss": 1.353, + "theoretical_loss": 3.4640713728367865, + "tokens_seen": 1963982848 + }, + { + "epoch": 0.6, + "learning_rate": 0.0003272858575727182, + "loss": 1.3548, + "theoretical_loss": 3.464061780218878, + "tokens_seen": 1964048384 + }, + { + "epoch": 0.6, + "learning_rate": 0.0003272698094282849, + "loss": 1.3475, + "theoretical_loss": 3.464052188010669, + "tokens_seen": 1964113920 + }, + { + "epoch": 0.6, + "learning_rate": 0.0003272537612838516, + "loss": 1.3655, + "theoretical_loss": 3.4640425962121286, + "tokens_seen": 1964179456 + }, + { + "epoch": 0.6, + "learning_rate": 0.0003272377131394183, + "loss": 1.2828, + "theoretical_loss": 3.464033004823225, + "tokens_seen": 1964244992 + }, + { + "epoch": 0.6, + "learning_rate": 0.000327221664994985, + "loss": 1.3308, + "theoretical_loss": 3.464023413843928, + "tokens_seen": 1964310528 + }, + { + "epoch": 0.6, + "learning_rate": 0.0003272056168505517, + "loss": 1.3551, + "theoretical_loss": 3.4640138232742066, + "tokens_seen": 1964376064 + }, + { + "epoch": 0.6, + "learning_rate": 0.0003271895687061184, + "loss": 1.3795, + "theoretical_loss": 3.464004233114028, + "tokens_seen": 1964441600 + }, + { + "epoch": 0.6, + "learning_rate": 0.0003271735205616851, + "loss": 1.3094, + "theoretical_loss": 3.4639946433633626, + "tokens_seen": 1964507136 + }, + { + "epoch": 0.6, + "learning_rate": 0.00032715747241725176, + "loss": 1.4006, + "theoretical_loss": 3.463985054022179, + "tokens_seen": 1964572672 + }, + { + "epoch": 0.6, + "learning_rate": 0.00032714142427281846, + "loss": 1.2755, + "theoretical_loss": 3.4639754650904453, + "tokens_seen": 1964638208 + }, + { + "epoch": 0.6, + "learning_rate": 0.00032712537612838516, + "loss": 1.3294, + "theoretical_loss": 3.463965876568131, + "tokens_seen": 1964703744 + }, + { + "epoch": 0.6, + "learning_rate": 0.00032710932798395186, + "loss": 1.4157, + "theoretical_loss": 3.463956288455205, + "tokens_seen": 1964769280 + }, + { + "epoch": 0.6, + "learning_rate": 0.00032709327983951856, + "loss": 1.2936, + "theoretical_loss": 3.463946700751636, + "tokens_seen": 1964834816 + }, + { + "epoch": 0.6, + "learning_rate": 0.00032707723169508526, + "loss": 1.3177, + "theoretical_loss": 3.4639371134573924, + "tokens_seen": 1964900352 + }, + { + "epoch": 0.6, + "learning_rate": 0.00032706118355065196, + "loss": 1.2839, + "theoretical_loss": 3.463927526572444, + "tokens_seen": 1964965888 + }, + { + "epoch": 0.6, + "learning_rate": 0.00032704513540621866, + "loss": 1.3827, + "theoretical_loss": 3.463917940096759, + "tokens_seen": 1965031424 + }, + { + "epoch": 0.6, + "learning_rate": 0.00032702908726178537, + "loss": 1.3287, + "theoretical_loss": 3.4639083540303064, + "tokens_seen": 1965096960 + }, + { + "epoch": 0.6, + "learning_rate": 0.00032701303911735207, + "loss": 1.3203, + "theoretical_loss": 3.4638987683730553, + "tokens_seen": 1965162496 + }, + { + "epoch": 0.6, + "learning_rate": 0.00032699699097291877, + "loss": 1.4069, + "theoretical_loss": 3.4638891831249747, + "tokens_seen": 1965228032 + }, + { + "epoch": 0.6, + "learning_rate": 0.00032698094282848547, + "loss": 1.3413, + "theoretical_loss": 3.463879598286033, + "tokens_seen": 1965293568 + }, + { + "epoch": 0.6, + "learning_rate": 0.00032696489468405217, + "loss": 1.4123, + "theoretical_loss": 3.4638700138561993, + "tokens_seen": 1965359104 + }, + { + "epoch": 0.6, + "learning_rate": 0.00032694884653961887, + "loss": 1.2214, + "theoretical_loss": 3.4638604298354427, + "tokens_seen": 1965424640 + }, + { + "epoch": 0.6, + "learning_rate": 0.00032693279839518557, + "loss": 1.3301, + "theoretical_loss": 3.463850846223732, + "tokens_seen": 1965490176 + }, + { + "epoch": 0.6, + "learning_rate": 0.00032691675025075227, + "loss": 1.3984, + "theoretical_loss": 3.4638412630210356, + "tokens_seen": 1965555712 + }, + { + "epoch": 0.6, + "learning_rate": 0.00032690070210631897, + "loss": 1.4078, + "theoretical_loss": 3.4638316802273232, + "tokens_seen": 1965621248 + }, + { + "epoch": 0.6, + "learning_rate": 0.00032688465396188567, + "loss": 1.2833, + "theoretical_loss": 3.463822097842564, + "tokens_seen": 1965686784 + }, + { + "epoch": 0.6, + "learning_rate": 0.00032686860581745237, + "loss": 1.3423, + "theoretical_loss": 3.4638125158667252, + "tokens_seen": 1965752320 + }, + { + "epoch": 0.6, + "learning_rate": 0.00032685255767301907, + "loss": 1.4457, + "theoretical_loss": 3.4638029342997774, + "tokens_seen": 1965817856 + }, + { + "epoch": 0.6, + "learning_rate": 0.00032683650952858577, + "loss": 1.3487, + "theoretical_loss": 3.4637933531416882, + "tokens_seen": 1965883392 + }, + { + "epoch": 0.6, + "learning_rate": 0.0003268204613841525, + "loss": 1.3398, + "theoretical_loss": 3.4637837723924276, + "tokens_seen": 1965948928 + }, + { + "epoch": 0.6, + "learning_rate": 0.0003268044132397192, + "loss": 1.2299, + "theoretical_loss": 3.4637741920519645, + "tokens_seen": 1966014464 + }, + { + "debugging/Compilability": 1.0, + "debugging/distinct-1-grams": 0.7394139103740143, + "debugging/entropy-1-grams": 5.1329642330953185, + "debugging/length": 486.55555555555554, + "debugging/num_segments": 9, + "debugging/score": 0.00915515707107717, + "debugging/score_std": 0.01661134296838666, + "epoch": 0.6, + "objective/train/docs_used": 716760, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.4669904708862305, + "objective/train/theoretical_loss": 3.4637646121202676, + "objective/train/tokens_used": 1986540000, + "theoretical_loss": 3.4637646121202676, + "tokens_seen": 1966080000 + }, + { + "epoch": 0.6, + "learning_rate": 0.0003267883650952859, + "loss": 1.4407, + "theoretical_loss": 3.4637646121202676, + "tokens_seen": 1966080000 + }, + { + "epoch": 0.6, + "learning_rate": 0.0003267723169508526, + "loss": 1.3419, + "theoretical_loss": 3.463755032597305, + "tokens_seen": 1966145536 + }, + { + "epoch": 0.6, + "learning_rate": 0.0003267562688064193, + "loss": 1.2799, + "theoretical_loss": 3.463745453483047, + "tokens_seen": 1966211072 + }, + { + "epoch": 0.6, + "learning_rate": 0.000326740220661986, + "loss": 1.3182, + "theoretical_loss": 3.463735874777462, + "tokens_seen": 1966276608 + }, + { + "epoch": 0.6, + "learning_rate": 0.0003267241725175527, + "loss": 1.292, + "theoretical_loss": 3.4637262964805187, + "tokens_seen": 1966342144 + }, + { + "epoch": 0.6, + "learning_rate": 0.0003267081243731194, + "loss": 1.4168, + "theoretical_loss": 3.4637167185921856, + "tokens_seen": 1966407680 + }, + { + "epoch": 0.6, + "learning_rate": 0.0003266920762286861, + "loss": 1.3113, + "theoretical_loss": 3.463707141112433, + "tokens_seen": 1966473216 + }, + { + "epoch": 0.6, + "learning_rate": 0.0003266760280842528, + "loss": 1.3633, + "theoretical_loss": 3.4636975640412286, + "tokens_seen": 1966538752 + }, + { + "epoch": 0.6, + "learning_rate": 0.0003266599799398195, + "loss": 1.3722, + "theoretical_loss": 3.4636879873785427, + "tokens_seen": 1966604288 + }, + { + "epoch": 0.6, + "learning_rate": 0.0003266439317953862, + "loss": 1.3269, + "theoretical_loss": 3.4636784111243424, + "tokens_seen": 1966669824 + }, + { + "epoch": 0.6, + "learning_rate": 0.0003266278836509529, + "loss": 1.3533, + "theoretical_loss": 3.4636688352785985, + "tokens_seen": 1966735360 + }, + { + "epoch": 0.6, + "learning_rate": 0.0003266118355065196, + "loss": 1.3332, + "theoretical_loss": 3.463659259841279, + "tokens_seen": 1966800896 + }, + { + "epoch": 0.6, + "learning_rate": 0.0003265957873620863, + "loss": 1.3228, + "theoretical_loss": 3.463649684812353, + "tokens_seen": 1966866432 + }, + { + "epoch": 0.6, + "learning_rate": 0.000326579739217653, + "loss": 1.3678, + "theoretical_loss": 3.463640110191789, + "tokens_seen": 1966931968 + }, + { + "epoch": 0.6, + "learning_rate": 0.0003265636910732197, + "loss": 1.3488, + "theoretical_loss": 3.463630535979557, + "tokens_seen": 1966997504 + }, + { + "epoch": 0.6, + "learning_rate": 0.0003265476429287864, + "loss": 1.4806, + "theoretical_loss": 3.463620962175625, + "tokens_seen": 1967063040 + }, + { + "epoch": 0.6, + "learning_rate": 0.0003265315947843531, + "loss": 1.2433, + "theoretical_loss": 3.4636113887799627, + "tokens_seen": 1967128576 + }, + { + "epoch": 0.6, + "learning_rate": 0.0003265155466399198, + "loss": 1.2589, + "theoretical_loss": 3.463601815792539, + "tokens_seen": 1967194112 + }, + { + "epoch": 0.6, + "learning_rate": 0.0003264994984954865, + "loss": 1.4535, + "theoretical_loss": 3.4635922432133226, + "tokens_seen": 1967259648 + }, + { + "epoch": 0.6, + "learning_rate": 0.0003264834503510532, + "loss": 1.3527, + "theoretical_loss": 3.463582671042283, + "tokens_seen": 1967325184 + }, + { + "epoch": 0.6, + "learning_rate": 0.0003264674022066199, + "loss": 1.3933, + "theoretical_loss": 3.4635730992793885, + "tokens_seen": 1967390720 + }, + { + "epoch": 0.6, + "learning_rate": 0.0003264513540621866, + "loss": 1.4197, + "theoretical_loss": 3.463563527924608, + "tokens_seen": 1967456256 + }, + { + "epoch": 0.6, + "learning_rate": 0.0003264353059177533, + "loss": 1.4437, + "theoretical_loss": 3.4635539569779117, + "tokens_seen": 1967521792 + }, + { + "epoch": 0.6, + "learning_rate": 0.00032641925777332, + "loss": 1.3781, + "theoretical_loss": 3.463544386439268, + "tokens_seen": 1967587328 + }, + { + "epoch": 0.6, + "learning_rate": 0.0003264032096288867, + "loss": 1.3063, + "theoretical_loss": 3.463534816308645, + "tokens_seen": 1967652864 + }, + { + "epoch": 0.6, + "learning_rate": 0.00032638716148445334, + "loss": 1.3615, + "theoretical_loss": 3.4635252465860127, + "tokens_seen": 1967718400 + }, + { + "epoch": 0.6, + "learning_rate": 0.00032637111334002004, + "loss": 1.5054, + "theoretical_loss": 3.46351567727134, + "tokens_seen": 1967783936 + }, + { + "epoch": 0.6, + "learning_rate": 0.00032635506519558674, + "loss": 1.3241, + "theoretical_loss": 3.4635061083645957, + "tokens_seen": 1967849472 + }, + { + "epoch": 0.6, + "learning_rate": 0.00032633901705115344, + "loss": 1.2677, + "theoretical_loss": 3.463496539865749, + "tokens_seen": 1967915008 + }, + { + "epoch": 0.6, + "learning_rate": 0.00032632296890672014, + "loss": 1.2836, + "theoretical_loss": 3.463486971774769, + "tokens_seen": 1967980544 + }, + { + "epoch": 0.6, + "learning_rate": 0.00032630692076228684, + "loss": 1.4781, + "theoretical_loss": 3.4634774040916243, + "tokens_seen": 1968046080 + }, + { + "epoch": 0.6, + "learning_rate": 0.00032629087261785354, + "loss": 1.4025, + "theoretical_loss": 3.463467836816285, + "tokens_seen": 1968111616 + }, + { + "epoch": 0.6, + "learning_rate": 0.00032627482447342024, + "loss": 1.4564, + "theoretical_loss": 3.4634582699487186, + "tokens_seen": 1968177152 + }, + { + "epoch": 0.6, + "learning_rate": 0.00032625877632898694, + "loss": 1.3415, + "theoretical_loss": 3.463448703488895, + "tokens_seen": 1968242688 + }, + { + "epoch": 0.6, + "learning_rate": 0.00032624272818455364, + "loss": 1.3685, + "theoretical_loss": 3.4634391374367834, + "tokens_seen": 1968308224 + }, + { + "epoch": 0.6, + "learning_rate": 0.00032622668004012034, + "loss": 1.3889, + "theoretical_loss": 3.4634295717923527, + "tokens_seen": 1968373760 + }, + { + "epoch": 0.6, + "learning_rate": 0.00032621063189568704, + "loss": 1.3986, + "theoretical_loss": 3.463420006555572, + "tokens_seen": 1968439296 + }, + { + "epoch": 0.6, + "learning_rate": 0.00032619458375125374, + "loss": 1.3325, + "theoretical_loss": 3.4634104417264098, + "tokens_seen": 1968504832 + }, + { + "epoch": 0.6, + "learning_rate": 0.00032617853560682044, + "loss": 1.2936, + "theoretical_loss": 3.463400877304836, + "tokens_seen": 1968570368 + }, + { + "epoch": 0.6, + "learning_rate": 0.00032616248746238715, + "loss": 1.3126, + "theoretical_loss": 3.4633913132908187, + "tokens_seen": 1968635904 + }, + { + "epoch": 0.6, + "learning_rate": 0.00032614643931795385, + "loss": 1.3546, + "theoretical_loss": 3.463381749684328, + "tokens_seen": 1968701440 + }, + { + "epoch": 0.6, + "learning_rate": 0.00032613039117352055, + "loss": 1.3737, + "theoretical_loss": 3.463372186485332, + "tokens_seen": 1968766976 + }, + { + "epoch": 0.6, + "learning_rate": 0.0003261143430290873, + "loss": 1.3293, + "theoretical_loss": 3.4633626236938007, + "tokens_seen": 1968832512 + }, + { + "epoch": 0.6, + "learning_rate": 0.000326098294884654, + "loss": 1.3753, + "theoretical_loss": 3.4633530613097028, + "tokens_seen": 1968898048 + }, + { + "epoch": 0.6, + "learning_rate": 0.0003260822467402207, + "loss": 1.2939, + "theoretical_loss": 3.463343499333007, + "tokens_seen": 1968963584 + }, + { + "epoch": 0.6, + "learning_rate": 0.0003260661985957874, + "loss": 1.3118, + "theoretical_loss": 3.463333937763683, + "tokens_seen": 1969029120 + }, + { + "epoch": 0.6, + "learning_rate": 0.0003260501504513541, + "loss": 1.2869, + "theoretical_loss": 3.4633243766016992, + "tokens_seen": 1969094656 + }, + { + "epoch": 0.6, + "learning_rate": 0.0003260341023069208, + "loss": 1.2744, + "theoretical_loss": 3.463314815847026, + "tokens_seen": 1969160192 + }, + { + "epoch": 0.6, + "learning_rate": 0.0003260180541624875, + "loss": 1.3373, + "theoretical_loss": 3.4633052554996304, + "tokens_seen": 1969225728 + }, + { + "epoch": 0.6, + "learning_rate": 0.0003260020060180542, + "loss": 1.4276, + "theoretical_loss": 3.4632956955594834, + "tokens_seen": 1969291264 + }, + { + "epoch": 0.6, + "objective/train/docs_used": 718017, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.4253120422363281, + "objective/train/theoretical_loss": 3.4632861360265528, + "objective/train/tokens_used": 1989816800, + "theoretical_loss": 3.4632861360265528, + "tokens_seen": 1969356800 + }, + { + "epoch": 0.6, + "learning_rate": 0.0003259859578736209, + "loss": 1.355, + "theoretical_loss": 3.4632861360265528, + "tokens_seen": 1969356800 + }, + { + "epoch": 0.6, + "learning_rate": 0.0003259699097291876, + "loss": 1.3445, + "theoretical_loss": 3.463276576900809, + "tokens_seen": 1969422336 + }, + { + "epoch": 0.6, + "learning_rate": 0.0003259538615847543, + "loss": 1.3304, + "theoretical_loss": 3.4632670181822203, + "tokens_seen": 1969487872 + }, + { + "epoch": 0.6, + "learning_rate": 0.000325937813440321, + "loss": 1.4885, + "theoretical_loss": 3.4632574598707553, + "tokens_seen": 1969553408 + }, + { + "epoch": 0.6, + "learning_rate": 0.0003259217652958877, + "loss": 1.3826, + "theoretical_loss": 3.463247901966384, + "tokens_seen": 1969618944 + }, + { + "epoch": 0.6, + "learning_rate": 0.0003259057171514544, + "loss": 1.2627, + "theoretical_loss": 3.463238344469075, + "tokens_seen": 1969684480 + }, + { + "epoch": 0.6, + "learning_rate": 0.0003258896690070211, + "loss": 1.258, + "theoretical_loss": 3.463228787378798, + "tokens_seen": 1969750016 + }, + { + "epoch": 0.6, + "learning_rate": 0.0003258736208625878, + "loss": 1.3365, + "theoretical_loss": 3.463219230695522, + "tokens_seen": 1969815552 + }, + { + "epoch": 0.6, + "learning_rate": 0.0003258575727181545, + "loss": 1.4455, + "theoretical_loss": 3.4632096744192156, + "tokens_seen": 1969881088 + }, + { + "epoch": 0.6, + "learning_rate": 0.0003258415245737212, + "loss": 1.3212, + "theoretical_loss": 3.4632001185498478, + "tokens_seen": 1969946624 + }, + { + "epoch": 0.6, + "learning_rate": 0.0003258254764292879, + "loss": 1.364, + "theoretical_loss": 3.4631905630873883, + "tokens_seen": 1970012160 + }, + { + "epoch": 0.6, + "learning_rate": 0.0003258094282848546, + "loss": 1.221, + "theoretical_loss": 3.4631810080318064, + "tokens_seen": 1970077696 + }, + { + "epoch": 0.6, + "learning_rate": 0.0003257933801404213, + "loss": 1.4334, + "theoretical_loss": 3.4631714533830706, + "tokens_seen": 1970143232 + }, + { + "epoch": 0.6, + "learning_rate": 0.000325777331995988, + "loss": 1.2151, + "theoretical_loss": 3.4631618991411512, + "tokens_seen": 1970208768 + }, + { + "epoch": 0.6, + "learning_rate": 0.0003257612838515547, + "loss": 1.4011, + "theoretical_loss": 3.463152345306016, + "tokens_seen": 1970274304 + }, + { + "epoch": 0.6, + "learning_rate": 0.00032574523570712136, + "loss": 1.44, + "theoretical_loss": 3.463142791877634, + "tokens_seen": 1970339840 + }, + { + "epoch": 0.6, + "learning_rate": 0.00032572918756268806, + "loss": 1.3109, + "theoretical_loss": 3.463133238855976, + "tokens_seen": 1970405376 + }, + { + "epoch": 0.6, + "learning_rate": 0.00032571313941825476, + "loss": 1.3215, + "theoretical_loss": 3.46312368624101, + "tokens_seen": 1970470912 + }, + { + "epoch": 0.6, + "learning_rate": 0.00032569709127382146, + "loss": 1.1986, + "theoretical_loss": 3.463114134032705, + "tokens_seen": 1970536448 + }, + { + "epoch": 0.6, + "learning_rate": 0.00032568104312938816, + "loss": 1.3417, + "theoretical_loss": 3.463104582231031, + "tokens_seen": 1970601984 + }, + { + "epoch": 0.6, + "learning_rate": 0.00032566499498495486, + "loss": 1.3092, + "theoretical_loss": 3.4630950308359565, + "tokens_seen": 1970667520 + }, + { + "epoch": 0.6, + "learning_rate": 0.00032564894684052157, + "loss": 1.3647, + "theoretical_loss": 3.4630854798474506, + "tokens_seen": 1970733056 + }, + { + "epoch": 0.6, + "learning_rate": 0.00032563289869608827, + "loss": 1.3468, + "theoretical_loss": 3.463075929265483, + "tokens_seen": 1970798592 + }, + { + "epoch": 0.6, + "learning_rate": 0.00032561685055165497, + "loss": 1.2952, + "theoretical_loss": 3.4630663790900225, + "tokens_seen": 1970864128 + }, + { + "epoch": 0.6, + "learning_rate": 0.00032560080240722167, + "loss": 1.3136, + "theoretical_loss": 3.4630568293210384, + "tokens_seen": 1970929664 + }, + { + "epoch": 0.6, + "learning_rate": 0.00032558475426278837, + "loss": 1.3096, + "theoretical_loss": 3.4630472799585, + "tokens_seen": 1970995200 + }, + { + "epoch": 0.6, + "learning_rate": 0.00032556870611835507, + "loss": 1.4248, + "theoretical_loss": 3.4630377310023763, + "tokens_seen": 1971060736 + }, + { + "epoch": 0.6, + "learning_rate": 0.00032555265797392177, + "loss": 1.3946, + "theoretical_loss": 3.4630281824526365, + "tokens_seen": 1971126272 + }, + { + "epoch": 0.6, + "learning_rate": 0.00032553660982948847, + "loss": 1.3127, + "theoretical_loss": 3.4630186343092495, + "tokens_seen": 1971191808 + }, + { + "epoch": 0.6, + "learning_rate": 0.00032552056168505517, + "loss": 1.4562, + "theoretical_loss": 3.4630090865721854, + "tokens_seen": 1971257344 + }, + { + "epoch": 0.6, + "learning_rate": 0.00032550451354062187, + "loss": 1.3021, + "theoretical_loss": 3.462999539241413, + "tokens_seen": 1971322880 + }, + { + "epoch": 0.6, + "learning_rate": 0.00032548846539618857, + "loss": 1.4085, + "theoretical_loss": 3.462989992316901, + "tokens_seen": 1971388416 + }, + { + "epoch": 0.6, + "learning_rate": 0.00032547241725175527, + "loss": 1.4694, + "theoretical_loss": 3.4629804457986184, + "tokens_seen": 1971453952 + }, + { + "epoch": 0.6, + "learning_rate": 0.000325456369107322, + "loss": 1.3666, + "theoretical_loss": 3.462970899686536, + "tokens_seen": 1971519488 + }, + { + "epoch": 0.6, + "learning_rate": 0.0003254403209628887, + "loss": 1.3434, + "theoretical_loss": 3.462961353980621, + "tokens_seen": 1971585024 + }, + { + "epoch": 0.6, + "learning_rate": 0.0003254242728184554, + "loss": 1.3235, + "theoretical_loss": 3.4629518086808444, + "tokens_seen": 1971650560 + }, + { + "epoch": 0.6, + "learning_rate": 0.0003254082246740221, + "loss": 1.4053, + "theoretical_loss": 3.4629422637871743, + "tokens_seen": 1971716096 + }, + { + "epoch": 0.6, + "learning_rate": 0.0003253921765295888, + "loss": 1.319, + "theoretical_loss": 3.4629327192995802, + "tokens_seen": 1971781632 + }, + { + "epoch": 0.6, + "learning_rate": 0.0003253761283851555, + "loss": 1.409, + "theoretical_loss": 3.4629231752180316, + "tokens_seen": 1971847168 + }, + { + "epoch": 0.6, + "learning_rate": 0.0003253600802407222, + "loss": 1.3796, + "theoretical_loss": 3.4629136315424973, + "tokens_seen": 1971912704 + }, + { + "epoch": 0.6, + "learning_rate": 0.0003253440320962889, + "loss": 1.4866, + "theoretical_loss": 3.4629040882729467, + "tokens_seen": 1971978240 + }, + { + "epoch": 0.6, + "learning_rate": 0.0003253279839518556, + "loss": 1.4318, + "theoretical_loss": 3.462894545409349, + "tokens_seen": 1972043776 + }, + { + "epoch": 0.6, + "learning_rate": 0.0003253119358074223, + "loss": 1.2364, + "theoretical_loss": 3.462885002951674, + "tokens_seen": 1972109312 + }, + { + "epoch": 0.6, + "learning_rate": 0.000325295887662989, + "loss": 1.2774, + "theoretical_loss": 3.46287546089989, + "tokens_seen": 1972174848 + }, + { + "epoch": 0.6, + "learning_rate": 0.0003252798395185557, + "loss": 1.3172, + "theoretical_loss": 3.462865919253967, + "tokens_seen": 1972240384 + }, + { + "epoch": 0.6, + "learning_rate": 0.0003252637913741224, + "loss": 1.3752, + "theoretical_loss": 3.462856378013874, + "tokens_seen": 1972305920 + }, + { + "epoch": 0.6, + "learning_rate": 0.0003252477432296891, + "loss": 1.2345, + "theoretical_loss": 3.4628468371795798, + "tokens_seen": 1972371456 + }, + { + "epoch": 0.6, + "learning_rate": 0.0003252316950852558, + "loss": 1.3294, + "theoretical_loss": 3.4628372967510543, + "tokens_seen": 1972436992 + }, + { + "epoch": 0.6, + "learning_rate": 0.0003252156469408225, + "loss": 1.3152, + "theoretical_loss": 3.4628277567282666, + "tokens_seen": 1972502528 + }, + { + "epoch": 0.6, + "learning_rate": 0.0003251995987963892, + "loss": 1.2439, + "theoretical_loss": 3.4628182171111863, + "tokens_seen": 1972568064 + }, + { + "epoch": 0.6, + "objective/train/docs_used": 719209, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.242699384689331, + "objective/train/theoretical_loss": 3.4628086778997815, + "objective/train/tokens_used": 1993093600, + "theoretical_loss": 3.4628086778997815, + "tokens_seen": 1972633600 + }, + { + "epoch": 0.6, + "learning_rate": 0.0003251835506519559, + "loss": 1.3021, + "theoretical_loss": 3.4628086778997815, + "tokens_seen": 1972633600 + }, + { + "epoch": 0.6, + "learning_rate": 0.0003251675025075226, + "loss": 1.2194, + "theoretical_loss": 3.462799139094023, + "tokens_seen": 1972699136 + }, + { + "epoch": 0.6, + "learning_rate": 0.0003251514543630893, + "loss": 1.39, + "theoretical_loss": 3.462789600693879, + "tokens_seen": 1972764672 + }, + { + "epoch": 0.6, + "learning_rate": 0.000325135406218656, + "loss": 1.3135, + "theoretical_loss": 3.462780062699319, + "tokens_seen": 1972830208 + }, + { + "epoch": 0.6, + "learning_rate": 0.0003251193580742227, + "loss": 1.4013, + "theoretical_loss": 3.4627705251103125, + "tokens_seen": 1972895744 + }, + { + "epoch": 0.6, + "learning_rate": 0.0003251033099297894, + "loss": 1.2655, + "theoretical_loss": 3.4627609879268286, + "tokens_seen": 1972961280 + }, + { + "epoch": 0.6, + "learning_rate": 0.0003250872617853561, + "loss": 1.3995, + "theoretical_loss": 3.4627514511488364, + "tokens_seen": 1973026816 + }, + { + "epoch": 0.6, + "learning_rate": 0.0003250712136409228, + "loss": 1.3862, + "theoretical_loss": 3.4627419147763057, + "tokens_seen": 1973092352 + }, + { + "epoch": 0.6, + "learning_rate": 0.0003250551654964895, + "loss": 1.3607, + "theoretical_loss": 3.462732378809206, + "tokens_seen": 1973157888 + }, + { + "epoch": 0.6, + "learning_rate": 0.0003250391173520562, + "loss": 1.3773, + "theoretical_loss": 3.4627228432475055, + "tokens_seen": 1973223424 + }, + { + "epoch": 0.6, + "learning_rate": 0.0003250230692076229, + "loss": 1.3767, + "theoretical_loss": 3.4627133080911743, + "tokens_seen": 1973288960 + }, + { + "epoch": 0.6, + "learning_rate": 0.0003250070210631896, + "loss": 1.362, + "theoretical_loss": 3.4627037733401815, + "tokens_seen": 1973354496 + }, + { + "epoch": 0.6, + "learning_rate": 0.0003249909729187563, + "loss": 1.3463, + "theoretical_loss": 3.462694238994497, + "tokens_seen": 1973420032 + }, + { + "epoch": 0.6, + "learning_rate": 0.00032497492477432294, + "loss": 1.3507, + "theoretical_loss": 3.462684705054089, + "tokens_seen": 1973485568 + }, + { + "epoch": 0.6, + "learning_rate": 0.00032495887662988964, + "loss": 1.3326, + "theoretical_loss": 3.4626751715189275, + "tokens_seen": 1973551104 + }, + { + "epoch": 0.6, + "learning_rate": 0.00032494282848545634, + "loss": 1.2502, + "theoretical_loss": 3.4626656383889816, + "tokens_seen": 1973616640 + }, + { + "epoch": 0.6, + "learning_rate": 0.00032492678034102304, + "loss": 1.3943, + "theoretical_loss": 3.462656105664221, + "tokens_seen": 1973682176 + }, + { + "epoch": 0.6, + "learning_rate": 0.0003249107321965898, + "loss": 1.3622, + "theoretical_loss": 3.4626465733446143, + "tokens_seen": 1973747712 + }, + { + "epoch": 0.6, + "learning_rate": 0.0003248946840521565, + "loss": 1.2904, + "theoretical_loss": 3.4626370414301317, + "tokens_seen": 1973813248 + }, + { + "epoch": 0.6, + "learning_rate": 0.0003248786359077232, + "loss": 1.413, + "theoretical_loss": 3.462627509920742, + "tokens_seen": 1973878784 + }, + { + "epoch": 0.6, + "learning_rate": 0.0003248625877632899, + "loss": 1.2934, + "theoretical_loss": 3.4626179788164144, + "tokens_seen": 1973944320 + }, + { + "epoch": 0.6, + "learning_rate": 0.0003248465396188566, + "loss": 1.3663, + "theoretical_loss": 3.462608448117119, + "tokens_seen": 1974009856 + }, + { + "epoch": 0.6, + "learning_rate": 0.0003248304914744233, + "loss": 1.302, + "theoretical_loss": 3.4625989178228243, + "tokens_seen": 1974075392 + }, + { + "epoch": 0.6, + "learning_rate": 0.00032481444332999, + "loss": 1.3481, + "theoretical_loss": 3.4625893879335, + "tokens_seen": 1974140928 + }, + { + "epoch": 0.6, + "learning_rate": 0.0003247983951855567, + "loss": 1.4042, + "theoretical_loss": 3.462579858449115, + "tokens_seen": 1974206464 + }, + { + "epoch": 0.6, + "learning_rate": 0.0003247823470411234, + "loss": 1.434, + "theoretical_loss": 3.4625703293696395, + "tokens_seen": 1974272000 + }, + { + "epoch": 0.6, + "learning_rate": 0.0003247662988966901, + "loss": 1.3575, + "theoretical_loss": 3.4625608006950426, + "tokens_seen": 1974337536 + }, + { + "epoch": 0.6, + "learning_rate": 0.0003247502507522568, + "loss": 1.3468, + "theoretical_loss": 3.4625512724252934, + "tokens_seen": 1974403072 + }, + { + "epoch": 0.6, + "learning_rate": 0.0003247342026078235, + "loss": 1.2925, + "theoretical_loss": 3.4625417445603617, + "tokens_seen": 1974468608 + }, + { + "epoch": 0.6, + "learning_rate": 0.0003247181544633902, + "loss": 1.3208, + "theoretical_loss": 3.4625322171002155, + "tokens_seen": 1974534144 + }, + { + "epoch": 0.6, + "learning_rate": 0.0003247021063189569, + "loss": 1.2971, + "theoretical_loss": 3.462522690044826, + "tokens_seen": 1974599680 + }, + { + "epoch": 0.6, + "learning_rate": 0.0003246860581745236, + "loss": 1.3381, + "theoretical_loss": 3.4625131633941617, + "tokens_seen": 1974665216 + }, + { + "epoch": 0.6, + "learning_rate": 0.0003246700100300903, + "loss": 1.3172, + "theoretical_loss": 3.462503637148192, + "tokens_seen": 1974730752 + }, + { + "epoch": 0.6, + "learning_rate": 0.000324653961885657, + "loss": 1.4019, + "theoretical_loss": 3.462494111306886, + "tokens_seen": 1974796288 + }, + { + "epoch": 0.6, + "learning_rate": 0.0003246379137412237, + "loss": 1.3842, + "theoretical_loss": 3.4624845858702136, + "tokens_seen": 1974861824 + }, + { + "epoch": 0.6, + "learning_rate": 0.0003246218655967904, + "loss": 1.3642, + "theoretical_loss": 3.4624750608381434, + "tokens_seen": 1974927360 + }, + { + "epoch": 0.6, + "learning_rate": 0.0003246058174523571, + "loss": 1.3782, + "theoretical_loss": 3.462465536210646, + "tokens_seen": 1974992896 + }, + { + "epoch": 0.6, + "learning_rate": 0.0003245897693079238, + "loss": 1.3777, + "theoretical_loss": 3.4624560119876904, + "tokens_seen": 1975058432 + }, + { + "epoch": 0.6, + "learning_rate": 0.0003245737211634905, + "loss": 1.3078, + "theoretical_loss": 3.462446488169245, + "tokens_seen": 1975123968 + }, + { + "epoch": 0.6, + "learning_rate": 0.0003245576730190572, + "loss": 1.2788, + "theoretical_loss": 3.4624369647552804, + "tokens_seen": 1975189504 + }, + { + "epoch": 0.6, + "learning_rate": 0.0003245416248746239, + "loss": 1.4038, + "theoretical_loss": 3.462427441745765, + "tokens_seen": 1975255040 + }, + { + "epoch": 0.6, + "learning_rate": 0.0003245255767301906, + "loss": 1.4484, + "theoretical_loss": 3.4624179191406697, + "tokens_seen": 1975320576 + }, + { + "epoch": 0.6, + "learning_rate": 0.0003245095285857573, + "loss": 1.3986, + "theoretical_loss": 3.4624083969399626, + "tokens_seen": 1975386112 + }, + { + "epoch": 0.6, + "learning_rate": 0.000324493480441324, + "loss": 1.3797, + "theoretical_loss": 3.462398875143613, + "tokens_seen": 1975451648 + }, + { + "epoch": 0.6, + "learning_rate": 0.0003244774322968907, + "loss": 1.4617, + "theoretical_loss": 3.4623893537515915, + "tokens_seen": 1975517184 + }, + { + "epoch": 0.6, + "learning_rate": 0.0003244613841524574, + "loss": 1.3331, + "theoretical_loss": 3.462379832763866, + "tokens_seen": 1975582720 + }, + { + "epoch": 0.6, + "learning_rate": 0.0003244453360080241, + "loss": 1.3358, + "theoretical_loss": 3.4623703121804077, + "tokens_seen": 1975648256 + }, + { + "epoch": 0.6, + "learning_rate": 0.0003244292878635908, + "loss": 1.3687, + "theoretical_loss": 3.462360792001184, + "tokens_seen": 1975713792 + }, + { + "epoch": 0.6, + "learning_rate": 0.0003244132397191575, + "loss": 1.464, + "theoretical_loss": 3.462351272226166, + "tokens_seen": 1975779328 + }, + { + "epoch": 0.6, + "learning_rate": 0.0003243971915747242, + "loss": 1.3881, + "theoretical_loss": 3.4623417528553224, + "tokens_seen": 1975844864 + }, + { + "epoch": 0.6, + "objective/train/docs_used": 720450, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.5214223861694336, + "objective/train/theoretical_loss": 3.462332233888623, + "objective/train/tokens_used": 1996370400, + "theoretical_loss": 3.462332233888623, + "tokens_seen": 1975910400 + }, + { + "epoch": 0.6, + "learning_rate": 0.0003243811434302909, + "loss": 1.4929, + "theoretical_loss": 3.462332233888623, + "tokens_seen": 1975910400 + }, + { + "epoch": 0.6, + "learning_rate": 0.0003243650952858576, + "loss": 1.3909, + "theoretical_loss": 3.462322715326037, + "tokens_seen": 1975975936 + }, + { + "epoch": 0.6, + "learning_rate": 0.00032434904714142426, + "loss": 1.2866, + "theoretical_loss": 3.462313197167534, + "tokens_seen": 1976041472 + }, + { + "epoch": 0.6, + "learning_rate": 0.00032433299899699096, + "loss": 1.3454, + "theoretical_loss": 3.4623036794130826, + "tokens_seen": 1976107008 + }, + { + "epoch": 0.6, + "learning_rate": 0.00032431695085255766, + "loss": 1.3388, + "theoretical_loss": 3.462294162062653, + "tokens_seen": 1976172544 + }, + { + "epoch": 0.6, + "learning_rate": 0.00032430090270812436, + "loss": 1.2605, + "theoretical_loss": 3.462284645116215, + "tokens_seen": 1976238080 + }, + { + "epoch": 0.6, + "learning_rate": 0.00032428485456369107, + "loss": 1.3728, + "theoretical_loss": 3.4622751285737374, + "tokens_seen": 1976303616 + }, + { + "epoch": 0.6, + "learning_rate": 0.00032426880641925777, + "loss": 1.2134, + "theoretical_loss": 3.46226561243519, + "tokens_seen": 1976369152 + }, + { + "epoch": 0.6, + "learning_rate": 0.00032425275827482447, + "loss": 1.336, + "theoretical_loss": 3.462256096700542, + "tokens_seen": 1976434688 + }, + { + "epoch": 0.6, + "learning_rate": 0.00032423671013039117, + "loss": 1.254, + "theoretical_loss": 3.462246581369763, + "tokens_seen": 1976500224 + }, + { + "epoch": 0.6, + "learning_rate": 0.00032422066198595787, + "loss": 1.3446, + "theoretical_loss": 3.462237066442823, + "tokens_seen": 1976565760 + }, + { + "epoch": 0.6, + "learning_rate": 0.00032420461384152457, + "loss": 1.4392, + "theoretical_loss": 3.46222755191969, + "tokens_seen": 1976631296 + }, + { + "epoch": 0.6, + "learning_rate": 0.00032418856569709127, + "loss": 1.3863, + "theoretical_loss": 3.4622180378003353, + "tokens_seen": 1976696832 + }, + { + "epoch": 0.6, + "learning_rate": 0.00032417251755265797, + "loss": 1.3501, + "theoretical_loss": 3.462208524084727, + "tokens_seen": 1976762368 + }, + { + "epoch": 0.6, + "learning_rate": 0.00032415646940822467, + "loss": 1.2344, + "theoretical_loss": 3.4621990107728355, + "tokens_seen": 1976827904 + }, + { + "epoch": 0.6, + "learning_rate": 0.00032414042126379137, + "loss": 1.3974, + "theoretical_loss": 3.46218949786463, + "tokens_seen": 1976893440 + }, + { + "epoch": 0.6, + "learning_rate": 0.00032412437311935807, + "loss": 1.3721, + "theoretical_loss": 3.462179985360079, + "tokens_seen": 1976958976 + }, + { + "epoch": 0.6, + "learning_rate": 0.00032410832497492477, + "loss": 1.3989, + "theoretical_loss": 3.4621704732591536, + "tokens_seen": 1977024512 + }, + { + "epoch": 0.6, + "learning_rate": 0.00032409227683049147, + "loss": 1.3749, + "theoretical_loss": 3.462160961561822, + "tokens_seen": 1977090048 + }, + { + "epoch": 0.6, + "learning_rate": 0.0003240762286860582, + "loss": 1.3136, + "theoretical_loss": 3.4621514502680553, + "tokens_seen": 1977155584 + }, + { + "epoch": 0.6, + "learning_rate": 0.0003240601805416249, + "loss": 1.3539, + "theoretical_loss": 3.4621419393778208, + "tokens_seen": 1977221120 + }, + { + "epoch": 0.6, + "learning_rate": 0.0003240441323971916, + "loss": 1.2983, + "theoretical_loss": 3.4621324288910897, + "tokens_seen": 1977286656 + }, + { + "epoch": 0.6, + "learning_rate": 0.0003240280842527583, + "loss": 1.4355, + "theoretical_loss": 3.462122918807831, + "tokens_seen": 1977352192 + }, + { + "epoch": 0.6, + "learning_rate": 0.000324012036108325, + "loss": 1.3416, + "theoretical_loss": 3.4621134091280137, + "tokens_seen": 1977417728 + }, + { + "epoch": 0.6, + "learning_rate": 0.0003239959879638917, + "loss": 1.3173, + "theoretical_loss": 3.4621038998516083, + "tokens_seen": 1977483264 + }, + { + "epoch": 0.6, + "learning_rate": 0.0003239799398194584, + "loss": 1.3883, + "theoretical_loss": 3.4620943909785837, + "tokens_seen": 1977548800 + }, + { + "epoch": 0.6, + "learning_rate": 0.0003239638916750251, + "loss": 1.459, + "theoretical_loss": 3.4620848825089094, + "tokens_seen": 1977614336 + }, + { + "epoch": 0.6, + "learning_rate": 0.0003239478435305918, + "loss": 1.3927, + "theoretical_loss": 3.4620753744425556, + "tokens_seen": 1977679872 + }, + { + "epoch": 0.6, + "learning_rate": 0.0003239317953861585, + "loss": 1.2604, + "theoretical_loss": 3.462065866779491, + "tokens_seen": 1977745408 + }, + { + "epoch": 0.6, + "learning_rate": 0.0003239157472417252, + "loss": 1.4032, + "theoretical_loss": 3.4620563595196847, + "tokens_seen": 1977810944 + }, + { + "epoch": 0.6, + "learning_rate": 0.0003238996990972919, + "loss": 1.3786, + "theoretical_loss": 3.462046852663108, + "tokens_seen": 1977876480 + }, + { + "epoch": 0.6, + "learning_rate": 0.0003238836509528586, + "loss": 1.3218, + "theoretical_loss": 3.462037346209729, + "tokens_seen": 1977942016 + }, + { + "epoch": 0.6, + "learning_rate": 0.0003238676028084253, + "loss": 1.3834, + "theoretical_loss": 3.462027840159517, + "tokens_seen": 1978007552 + }, + { + "epoch": 0.6, + "learning_rate": 0.000323851554663992, + "loss": 1.3178, + "theoretical_loss": 3.462018334512443, + "tokens_seen": 1978073088 + }, + { + "epoch": 0.6, + "learning_rate": 0.0003238355065195587, + "loss": 1.3488, + "theoretical_loss": 3.462008829268475, + "tokens_seen": 1978138624 + }, + { + "epoch": 0.6, + "learning_rate": 0.0003238194583751254, + "loss": 1.3512, + "theoretical_loss": 3.461999324427584, + "tokens_seen": 1978204160 + }, + { + "epoch": 0.6, + "learning_rate": 0.0003238034102306921, + "loss": 1.3543, + "theoretical_loss": 3.4619898199897388, + "tokens_seen": 1978269696 + }, + { + "epoch": 0.6, + "learning_rate": 0.0003237873620862588, + "loss": 1.2205, + "theoretical_loss": 3.461980315954909, + "tokens_seen": 1978335232 + }, + { + "epoch": 0.6, + "learning_rate": 0.0003237713139418255, + "loss": 1.3172, + "theoretical_loss": 3.461970812323064, + "tokens_seen": 1978400768 + }, + { + "epoch": 0.6, + "learning_rate": 0.0003237552657973922, + "loss": 1.3386, + "theoretical_loss": 3.461961309094173, + "tokens_seen": 1978466304 + }, + { + "epoch": 0.6, + "learning_rate": 0.0003237392176529589, + "loss": 1.4121, + "theoretical_loss": 3.4619518062682064, + "tokens_seen": 1978531840 + }, + { + "epoch": 0.6, + "learning_rate": 0.0003237231695085256, + "loss": 1.2639, + "theoretical_loss": 3.4619423038451336, + "tokens_seen": 1978597376 + }, + { + "epoch": 0.6, + "learning_rate": 0.0003237071213640923, + "loss": 1.4032, + "theoretical_loss": 3.461932801824924, + "tokens_seen": 1978662912 + }, + { + "epoch": 0.6, + "learning_rate": 0.000323691073219659, + "loss": 1.2698, + "theoretical_loss": 3.4619233002075473, + "tokens_seen": 1978728448 + }, + { + "epoch": 0.6, + "learning_rate": 0.0003236750250752257, + "loss": 1.3876, + "theoretical_loss": 3.4619137989929727, + "tokens_seen": 1978793984 + }, + { + "epoch": 0.6, + "learning_rate": 0.0003236589769307924, + "loss": 1.3731, + "theoretical_loss": 3.46190429818117, + "tokens_seen": 1978859520 + }, + { + "epoch": 0.6, + "learning_rate": 0.0003236429287863591, + "loss": 1.3298, + "theoretical_loss": 3.4618947977721097, + "tokens_seen": 1978925056 + }, + { + "epoch": 0.6, + "learning_rate": 0.0003236268806419258, + "loss": 1.1826, + "theoretical_loss": 3.4618852977657593, + "tokens_seen": 1978990592 + }, + { + "epoch": 0.6, + "learning_rate": 0.0003236108324974925, + "loss": 1.3185, + "theoretical_loss": 3.4618757981620907, + "tokens_seen": 1979056128 + }, + { + "epoch": 0.6, + "learning_rate": 0.0003235947843530592, + "loss": 1.4312, + "theoretical_loss": 3.461866298961072, + "tokens_seen": 1979121664 + }, + { + "epoch": 0.6, + "objective/train/docs_used": 721537, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.3898316621780396, + "objective/train/theoretical_loss": 3.461856800162673, + "objective/train/tokens_used": 1999647200, + "theoretical_loss": 3.461856800162673, + "tokens_seen": 1979187200 + }, + { + "epoch": 0.6, + "learning_rate": 0.0003235787362086259, + "loss": 1.3957, + "theoretical_loss": 3.461856800162673, + "tokens_seen": 1979187200 + }, + { + "epoch": 0.6, + "learning_rate": 0.0003235626880641926, + "loss": 1.3699, + "theoretical_loss": 3.461847301766864, + "tokens_seen": 1979252736 + }, + { + "epoch": 0.6, + "learning_rate": 0.0003235466399197593, + "loss": 1.3722, + "theoretical_loss": 3.461837803773614, + "tokens_seen": 1979318272 + }, + { + "epoch": 0.6, + "learning_rate": 0.000323530591775326, + "loss": 1.3325, + "theoretical_loss": 3.461828306182893, + "tokens_seen": 1979383808 + }, + { + "epoch": 0.6, + "learning_rate": 0.0003235145436308927, + "loss": 1.3487, + "theoretical_loss": 3.4618188089946695, + "tokens_seen": 1979449344 + }, + { + "epoch": 0.6, + "learning_rate": 0.0003234984954864594, + "loss": 1.3778, + "theoretical_loss": 3.461809312208915, + "tokens_seen": 1979514880 + }, + { + "epoch": 0.6, + "learning_rate": 0.0003234824473420261, + "loss": 1.3699, + "theoretical_loss": 3.4617998158255974, + "tokens_seen": 1979580416 + }, + { + "epoch": 0.6, + "learning_rate": 0.0003234663991975928, + "loss": 1.4034, + "theoretical_loss": 3.4617903198446878, + "tokens_seen": 1979645952 + }, + { + "epoch": 0.6, + "learning_rate": 0.0003234503510531595, + "loss": 1.3234, + "theoretical_loss": 3.4617808242661545, + "tokens_seen": 1979711488 + }, + { + "epoch": 0.6, + "learning_rate": 0.0003234343029087262, + "loss": 1.3911, + "theoretical_loss": 3.461771329089968, + "tokens_seen": 1979777024 + }, + { + "epoch": 0.6, + "learning_rate": 0.0003234182547642929, + "loss": 1.4964, + "theoretical_loss": 3.461761834316097, + "tokens_seen": 1979842560 + }, + { + "epoch": 0.6, + "learning_rate": 0.0003234022066198596, + "loss": 1.3495, + "theoretical_loss": 3.4617523399445123, + "tokens_seen": 1979908096 + }, + { + "epoch": 0.6, + "learning_rate": 0.0003233861584754263, + "loss": 1.4598, + "theoretical_loss": 3.4617428459751833, + "tokens_seen": 1979973632 + }, + { + "epoch": 0.6, + "learning_rate": 0.000323370110330993, + "loss": 1.3497, + "theoretical_loss": 3.4617333524080784, + "tokens_seen": 1980039168 + }, + { + "epoch": 0.6, + "learning_rate": 0.0003233540621865597, + "loss": 1.4096, + "theoretical_loss": 3.461723859243169, + "tokens_seen": 1980104704 + }, + { + "epoch": 0.6, + "learning_rate": 0.0003233380140421264, + "loss": 1.384, + "theoretical_loss": 3.461714366480424, + "tokens_seen": 1980170240 + }, + { + "epoch": 0.6, + "learning_rate": 0.0003233219658976931, + "loss": 1.301, + "theoretical_loss": 3.461704874119813, + "tokens_seen": 1980235776 + }, + { + "epoch": 0.6, + "learning_rate": 0.0003233059177532598, + "loss": 1.3552, + "theoretical_loss": 3.461695382161305, + "tokens_seen": 1980301312 + }, + { + "epoch": 0.6, + "learning_rate": 0.0003232898696088265, + "loss": 1.3756, + "theoretical_loss": 3.461685890604871, + "tokens_seen": 1980366848 + }, + { + "epoch": 0.6, + "learning_rate": 0.0003232738214643932, + "loss": 1.3239, + "theoretical_loss": 3.4616763994504796, + "tokens_seen": 1980432384 + }, + { + "epoch": 0.6, + "learning_rate": 0.0003232577733199599, + "loss": 1.4217, + "theoretical_loss": 3.461666908698101, + "tokens_seen": 1980497920 + }, + { + "epoch": 0.6, + "learning_rate": 0.0003232417251755266, + "loss": 1.4296, + "theoretical_loss": 3.4616574183477047, + "tokens_seen": 1980563456 + }, + { + "epoch": 0.6, + "learning_rate": 0.0003232256770310933, + "loss": 1.3751, + "theoretical_loss": 3.4616479283992607, + "tokens_seen": 1980628992 + }, + { + "epoch": 0.6, + "learning_rate": 0.00032320962888666, + "loss": 1.2366, + "theoretical_loss": 3.461638438852738, + "tokens_seen": 1980694528 + }, + { + "epoch": 0.6, + "learning_rate": 0.0003231935807422267, + "loss": 1.3261, + "theoretical_loss": 3.461628949708106, + "tokens_seen": 1980760064 + }, + { + "epoch": 0.6, + "learning_rate": 0.0003231775325977934, + "loss": 1.3363, + "theoretical_loss": 3.461619460965336, + "tokens_seen": 1980825600 + }, + { + "epoch": 0.6, + "learning_rate": 0.0003231614844533601, + "loss": 1.3165, + "theoretical_loss": 3.4616099726243963, + "tokens_seen": 1980891136 + }, + { + "epoch": 0.6, + "learning_rate": 0.0003231454363089268, + "loss": 1.4428, + "theoretical_loss": 3.4616004846852575, + "tokens_seen": 1980956672 + }, + { + "epoch": 0.6, + "learning_rate": 0.0003231293881644935, + "loss": 1.3667, + "theoretical_loss": 3.461590997147889, + "tokens_seen": 1981022208 + }, + { + "epoch": 0.6, + "learning_rate": 0.0003231133400200602, + "loss": 1.4206, + "theoretical_loss": 3.4615815100122593, + "tokens_seen": 1981087744 + }, + { + "epoch": 0.6, + "learning_rate": 0.0003230972918756269, + "loss": 1.2927, + "theoretical_loss": 3.4615720232783396, + "tokens_seen": 1981153280 + }, + { + "epoch": 0.6, + "learning_rate": 0.0003230812437311936, + "loss": 1.2191, + "theoretical_loss": 3.461562536946099, + "tokens_seen": 1981218816 + }, + { + "epoch": 0.6, + "learning_rate": 0.0003230651955867603, + "loss": 1.427, + "theoretical_loss": 3.4615530510155077, + "tokens_seen": 1981284352 + }, + { + "epoch": 0.6, + "learning_rate": 0.000323049147442327, + "loss": 1.2791, + "theoretical_loss": 3.4615435654865347, + "tokens_seen": 1981349888 + }, + { + "epoch": 0.6, + "learning_rate": 0.0003230330992978937, + "loss": 1.4044, + "theoretical_loss": 3.46153408035915, + "tokens_seen": 1981415424 + }, + { + "epoch": 0.6, + "learning_rate": 0.0003230170511534604, + "loss": 1.3104, + "theoretical_loss": 3.4615245956333234, + "tokens_seen": 1981480960 + }, + { + "epoch": 0.6, + "learning_rate": 0.0003230010030090271, + "loss": 1.3197, + "theoretical_loss": 3.4615151113090246, + "tokens_seen": 1981546496 + }, + { + "epoch": 0.6, + "learning_rate": 0.0003229849548645938, + "loss": 1.3571, + "theoretical_loss": 3.4615056273862232, + "tokens_seen": 1981612032 + }, + { + "epoch": 0.6, + "learning_rate": 0.0003229689067201605, + "loss": 1.4557, + "theoretical_loss": 3.4614961438648892, + "tokens_seen": 1981677568 + }, + { + "epoch": 0.6, + "learning_rate": 0.00032295285857572716, + "loss": 1.2457, + "theoretical_loss": 3.461486660744992, + "tokens_seen": 1981743104 + }, + { + "epoch": 0.6, + "learning_rate": 0.00032293681043129386, + "loss": 1.3039, + "theoretical_loss": 3.4614771780265015, + "tokens_seen": 1981808640 + }, + { + "epoch": 0.6, + "learning_rate": 0.00032292076228686056, + "loss": 1.2661, + "theoretical_loss": 3.4614676957093873, + "tokens_seen": 1981874176 + }, + { + "epoch": 0.6, + "learning_rate": 0.00032290471414242727, + "loss": 1.2695, + "theoretical_loss": 3.4614582137936196, + "tokens_seen": 1981939712 + }, + { + "epoch": 0.6, + "learning_rate": 0.00032288866599799397, + "loss": 1.3642, + "theoretical_loss": 3.4614487322791674, + "tokens_seen": 1982005248 + }, + { + "epoch": 0.6, + "learning_rate": 0.00032287261785356067, + "loss": 1.2737, + "theoretical_loss": 3.4614392511660013, + "tokens_seen": 1982070784 + }, + { + "epoch": 0.6, + "learning_rate": 0.00032285656970912737, + "loss": 1.2605, + "theoretical_loss": 3.46142977045409, + "tokens_seen": 1982136320 + }, + { + "epoch": 0.6, + "learning_rate": 0.00032284052156469407, + "loss": 1.3495, + "theoretical_loss": 3.461420290143404, + "tokens_seen": 1982201856 + }, + { + "epoch": 0.6, + "learning_rate": 0.00032282447342026077, + "loss": 1.2952, + "theoretical_loss": 3.461410810233913, + "tokens_seen": 1982267392 + }, + { + "epoch": 0.6, + "learning_rate": 0.00032280842527582747, + "loss": 1.361, + "theoretical_loss": 3.461401330725587, + "tokens_seen": 1982332928 + }, + { + "epoch": 0.6, + "learning_rate": 0.00032279237713139417, + "loss": 1.3916, + "theoretical_loss": 3.4613918516183952, + "tokens_seen": 1982398464 + }, + { + "epoch": 0.6, + "objective/train/docs_used": 722739, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.3152892589569092, + "objective/train/theoretical_loss": 3.4613823729123077, + "objective/train/tokens_used": 2002924000, + "theoretical_loss": 3.4613823729123077, + "tokens_seen": 1982464000 + }, + { + "epoch": 0.6, + "learning_rate": 0.00032277632898696087, + "loss": 1.3342, + "theoretical_loss": 3.4613823729123077, + "tokens_seen": 1982464000 + }, + { + "epoch": 0.6, + "learning_rate": 0.00032276028084252757, + "loss": 1.3698, + "theoretical_loss": 3.461372894607294, + "tokens_seen": 1982529536 + }, + { + "epoch": 0.6, + "learning_rate": 0.00032274423269809427, + "loss": 1.3237, + "theoretical_loss": 3.461363416703324, + "tokens_seen": 1982595072 + }, + { + "epoch": 0.6, + "learning_rate": 0.00032272818455366097, + "loss": 1.4626, + "theoretical_loss": 3.461353939200367, + "tokens_seen": 1982660608 + }, + { + "epoch": 0.6, + "learning_rate": 0.0003227121364092277, + "loss": 1.3733, + "theoretical_loss": 3.4613444620983937, + "tokens_seen": 1982726144 + }, + { + "epoch": 0.6, + "learning_rate": 0.0003226960882647944, + "loss": 1.3034, + "theoretical_loss": 3.461334985397374, + "tokens_seen": 1982791680 + }, + { + "epoch": 0.6, + "learning_rate": 0.0003226800401203611, + "loss": 1.3335, + "theoretical_loss": 3.4613255090972768, + "tokens_seen": 1982857216 + }, + { + "epoch": 0.6, + "learning_rate": 0.0003226639919759278, + "loss": 1.3349, + "theoretical_loss": 3.461316033198072, + "tokens_seen": 1982922752 + }, + { + "epoch": 0.6, + "learning_rate": 0.0003226479438314945, + "loss": 1.2465, + "theoretical_loss": 3.4613065576997295, + "tokens_seen": 1982988288 + }, + { + "epoch": 0.6, + "learning_rate": 0.0003226318956870612, + "loss": 1.3821, + "theoretical_loss": 3.4612970826022194, + "tokens_seen": 1983053824 + }, + { + "epoch": 0.6, + "learning_rate": 0.0003226158475426279, + "loss": 1.4253, + "theoretical_loss": 3.4612876079055113, + "tokens_seen": 1983119360 + }, + { + "epoch": 0.6, + "learning_rate": 0.0003225997993981946, + "loss": 1.4297, + "theoretical_loss": 3.461278133609575, + "tokens_seen": 1983184896 + }, + { + "epoch": 0.6, + "learning_rate": 0.0003225837512537613, + "loss": 1.3929, + "theoretical_loss": 3.461268659714381, + "tokens_seen": 1983250432 + }, + { + "epoch": 0.6, + "learning_rate": 0.000322567703109328, + "loss": 1.2678, + "theoretical_loss": 3.4612591862198974, + "tokens_seen": 1983315968 + }, + { + "epoch": 0.6, + "learning_rate": 0.0003225516549648947, + "loss": 1.4294, + "theoretical_loss": 3.4612497131260955, + "tokens_seen": 1983381504 + }, + { + "epoch": 0.6, + "learning_rate": 0.0003225356068204614, + "loss": 1.3853, + "theoretical_loss": 3.4612402404329448, + "tokens_seen": 1983447040 + }, + { + "epoch": 0.6, + "learning_rate": 0.00032251955867602813, + "loss": 1.3783, + "theoretical_loss": 3.461230768140415, + "tokens_seen": 1983512576 + }, + { + "epoch": 0.6, + "learning_rate": 0.00032250351053159484, + "loss": 1.3133, + "theoretical_loss": 3.4612212962484756, + "tokens_seen": 1983578112 + }, + { + "epoch": 0.6, + "learning_rate": 0.00032248746238716154, + "loss": 1.3651, + "theoretical_loss": 3.461211824757097, + "tokens_seen": 1983643648 + }, + { + "epoch": 0.6, + "learning_rate": 0.00032247141424272824, + "loss": 1.3825, + "theoretical_loss": 3.4612023536662484, + "tokens_seen": 1983709184 + }, + { + "epoch": 0.6, + "learning_rate": 0.00032245536609829494, + "loss": 1.3648, + "theoretical_loss": 3.4611928829759004, + "tokens_seen": 1983774720 + }, + { + "epoch": 0.6, + "learning_rate": 0.00032243931795386164, + "loss": 1.3071, + "theoretical_loss": 3.4611834126860224, + "tokens_seen": 1983840256 + }, + { + "epoch": 0.6, + "learning_rate": 0.00032242326980942834, + "loss": 1.3722, + "theoretical_loss": 3.4611739427965835, + "tokens_seen": 1983905792 + }, + { + "epoch": 0.6, + "learning_rate": 0.00032240722166499504, + "loss": 1.2919, + "theoretical_loss": 3.461164473307555, + "tokens_seen": 1983971328 + }, + { + "epoch": 0.6, + "learning_rate": 0.00032239117352056174, + "loss": 1.3494, + "theoretical_loss": 3.461155004218906, + "tokens_seen": 1984036864 + }, + { + "epoch": 0.6, + "learning_rate": 0.00032237512537612844, + "loss": 1.4227, + "theoretical_loss": 3.4611455355306067, + "tokens_seen": 1984102400 + }, + { + "epoch": 0.6, + "learning_rate": 0.00032235907723169514, + "loss": 1.3839, + "theoretical_loss": 3.461136067242626, + "tokens_seen": 1984167936 + }, + { + "epoch": 0.6, + "learning_rate": 0.00032234302908726184, + "loss": 1.2755, + "theoretical_loss": 3.4611265993549347, + "tokens_seen": 1984233472 + }, + { + "epoch": 0.6, + "learning_rate": 0.00032232698094282854, + "loss": 1.3138, + "theoretical_loss": 3.461117131867502, + "tokens_seen": 1984299008 + }, + { + "epoch": 0.6, + "learning_rate": 0.0003223109327983952, + "loss": 1.3837, + "theoretical_loss": 3.4611076647802985, + "tokens_seen": 1984364544 + }, + { + "epoch": 0.6, + "learning_rate": 0.0003222948846539619, + "loss": 1.3685, + "theoretical_loss": 3.4610981980932936, + "tokens_seen": 1984430080 + }, + { + "epoch": 0.6, + "learning_rate": 0.0003222788365095286, + "loss": 1.4251, + "theoretical_loss": 3.461088731806457, + "tokens_seen": 1984495616 + }, + { + "epoch": 0.6, + "learning_rate": 0.0003222627883650953, + "loss": 1.3792, + "theoretical_loss": 3.461079265919759, + "tokens_seen": 1984561152 + }, + { + "epoch": 0.6, + "learning_rate": 0.000322246740220662, + "loss": 1.3985, + "theoretical_loss": 3.4610698004331693, + "tokens_seen": 1984626688 + }, + { + "epoch": 0.6, + "learning_rate": 0.0003222306920762287, + "loss": 1.2544, + "theoretical_loss": 3.461060335346658, + "tokens_seen": 1984692224 + }, + { + "epoch": 0.6, + "learning_rate": 0.0003222146439317954, + "loss": 1.4019, + "theoretical_loss": 3.461050870660194, + "tokens_seen": 1984757760 + }, + { + "epoch": 0.6, + "learning_rate": 0.0003221985957873621, + "loss": 1.4064, + "theoretical_loss": 3.4610414063737487, + "tokens_seen": 1984823296 + }, + { + "epoch": 0.6, + "learning_rate": 0.0003221825476429288, + "loss": 1.3039, + "theoretical_loss": 3.46103194248729, + "tokens_seen": 1984888832 + }, + { + "epoch": 0.6, + "learning_rate": 0.0003221664994984955, + "loss": 1.2697, + "theoretical_loss": 3.4610224790007904, + "tokens_seen": 1984954368 + }, + { + "epoch": 0.6, + "learning_rate": 0.0003221504513540622, + "loss": 1.3221, + "theoretical_loss": 3.461013015914218, + "tokens_seen": 1985019904 + }, + { + "epoch": 0.6, + "learning_rate": 0.0003221344032096289, + "loss": 1.3264, + "theoretical_loss": 3.4610035532275427, + "tokens_seen": 1985085440 + }, + { + "epoch": 0.6, + "learning_rate": 0.0003221183550651956, + "loss": 1.3441, + "theoretical_loss": 3.4609940909407353, + "tokens_seen": 1985150976 + }, + { + "epoch": 0.6, + "learning_rate": 0.0003221023069207623, + "loss": 1.5204, + "theoretical_loss": 3.4609846290537645, + "tokens_seen": 1985216512 + }, + { + "epoch": 0.6, + "learning_rate": 0.000322086258776329, + "loss": 1.2692, + "theoretical_loss": 3.4609751675666014, + "tokens_seen": 1985282048 + }, + { + "epoch": 0.6, + "learning_rate": 0.0003220702106318957, + "loss": 1.3546, + "theoretical_loss": 3.4609657064792154, + "tokens_seen": 1985347584 + }, + { + "epoch": 0.6, + "learning_rate": 0.0003220541624874624, + "loss": 1.2515, + "theoretical_loss": 3.460956245791576, + "tokens_seen": 1985413120 + }, + { + "epoch": 0.6, + "learning_rate": 0.0003220381143430291, + "loss": 1.4169, + "theoretical_loss": 3.460946785503654, + "tokens_seen": 1985478656 + }, + { + "epoch": 0.6, + "learning_rate": 0.0003220220661985958, + "loss": 1.3325, + "theoretical_loss": 3.4609373256154186, + "tokens_seen": 1985544192 + }, + { + "epoch": 0.6, + "learning_rate": 0.0003220060180541625, + "loss": 1.3587, + "theoretical_loss": 3.4609278661268403, + "tokens_seen": 1985609728 + }, + { + "epoch": 0.6, + "learning_rate": 0.0003219899699097292, + "loss": 1.2816, + "theoretical_loss": 3.4609184070378882, + "tokens_seen": 1985675264 + }, + { + "epoch": 0.6, + "objective/train/docs_used": 723995, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.335287094116211, + "objective/train/theoretical_loss": 3.460908948348533, + "objective/train/tokens_used": 2006200800, + "theoretical_loss": 3.460908948348533, + "tokens_seen": 1985740800 + }, + { + "epoch": 0.6, + "learning_rate": 0.0003219739217652959, + "loss": 1.3433, + "theoretical_loss": 3.460908948348533, + "tokens_seen": 1985740800 + }, + { + "epoch": 0.6, + "learning_rate": 0.0003219578736208626, + "loss": 1.3146, + "theoretical_loss": 3.460899490058744, + "tokens_seen": 1985806336 + }, + { + "epoch": 0.6, + "learning_rate": 0.0003219418254764293, + "loss": 1.3866, + "theoretical_loss": 3.460890032168492, + "tokens_seen": 1985871872 + }, + { + "epoch": 0.6, + "learning_rate": 0.000321925777331996, + "loss": 1.3376, + "theoretical_loss": 3.4608805746777467, + "tokens_seen": 1985937408 + }, + { + "epoch": 0.6, + "learning_rate": 0.0003219097291875627, + "loss": 1.3191, + "theoretical_loss": 3.460871117586477, + "tokens_seen": 1986002944 + }, + { + "epoch": 0.6, + "learning_rate": 0.0003218936810431294, + "loss": 1.3615, + "theoretical_loss": 3.460861660894654, + "tokens_seen": 1986068480 + }, + { + "epoch": 0.6, + "learning_rate": 0.0003218776328986961, + "loss": 1.3447, + "theoretical_loss": 3.460852204602247, + "tokens_seen": 1986134016 + }, + { + "epoch": 0.6, + "learning_rate": 0.0003218615847542628, + "loss": 1.3558, + "theoretical_loss": 3.4608427487092266, + "tokens_seen": 1986199552 + }, + { + "epoch": 0.6, + "learning_rate": 0.0003218455366098295, + "loss": 1.3147, + "theoretical_loss": 3.460833293215562, + "tokens_seen": 1986265088 + }, + { + "epoch": 0.6, + "learning_rate": 0.0003218294884653962, + "loss": 1.4687, + "theoretical_loss": 3.4608238381212235, + "tokens_seen": 1986330624 + }, + { + "epoch": 0.6, + "learning_rate": 0.0003218134403209629, + "loss": 1.3071, + "theoretical_loss": 3.4608143834261815, + "tokens_seen": 1986396160 + }, + { + "epoch": 0.6, + "learning_rate": 0.0003217973921765296, + "loss": 1.3942, + "theoretical_loss": 3.4608049291304055, + "tokens_seen": 1986461696 + }, + { + "epoch": 0.6, + "learning_rate": 0.0003217813440320963, + "loss": 1.2881, + "theoretical_loss": 3.4607954752338657, + "tokens_seen": 1986527232 + }, + { + "epoch": 0.6, + "learning_rate": 0.000321765295887663, + "loss": 1.3264, + "theoretical_loss": 3.4607860217365314, + "tokens_seen": 1986592768 + }, + { + "epoch": 0.6, + "learning_rate": 0.0003217492477432297, + "loss": 1.3119, + "theoretical_loss": 3.460776568638373, + "tokens_seen": 1986658304 + }, + { + "epoch": 0.6, + "learning_rate": 0.0003217331995987964, + "loss": 1.3623, + "theoretical_loss": 3.460767115939361, + "tokens_seen": 1986723840 + }, + { + "epoch": 0.6, + "learning_rate": 0.0003217171514543631, + "loss": 1.2906, + "theoretical_loss": 3.4607576636394644, + "tokens_seen": 1986789376 + }, + { + "epoch": 0.6, + "learning_rate": 0.0003217011033099298, + "loss": 1.396, + "theoretical_loss": 3.460748211738654, + "tokens_seen": 1986854912 + }, + { + "epoch": 0.6, + "learning_rate": 0.0003216850551654965, + "loss": 1.312, + "theoretical_loss": 3.4607387602368997, + "tokens_seen": 1986920448 + }, + { + "epoch": 0.6, + "learning_rate": 0.0003216690070210632, + "loss": 1.3685, + "theoretical_loss": 3.4607293091341713, + "tokens_seen": 1986985984 + }, + { + "epoch": 0.6, + "learning_rate": 0.0003216529588766299, + "loss": 1.3513, + "theoretical_loss": 3.4607198584304384, + "tokens_seen": 1987051520 + }, + { + "epoch": 0.6, + "learning_rate": 0.0003216369107321966, + "loss": 1.3423, + "theoretical_loss": 3.4607104081256717, + "tokens_seen": 1987117056 + }, + { + "epoch": 0.6, + "learning_rate": 0.0003216208625877633, + "loss": 1.4155, + "theoretical_loss": 3.4607009582198405, + "tokens_seen": 1987182592 + }, + { + "epoch": 0.6, + "learning_rate": 0.00032160481444333, + "loss": 1.2423, + "theoretical_loss": 3.4606915087129155, + "tokens_seen": 1987248128 + }, + { + "epoch": 0.6, + "learning_rate": 0.0003215887662988967, + "loss": 1.3801, + "theoretical_loss": 3.460682059604866, + "tokens_seen": 1987313664 + }, + { + "epoch": 0.6, + "learning_rate": 0.0003215727181544634, + "loss": 1.3639, + "theoretical_loss": 3.4606726108956627, + "tokens_seen": 1987379200 + }, + { + "epoch": 0.6, + "learning_rate": 0.0003215566700100301, + "loss": 1.4021, + "theoretical_loss": 3.4606631625852757, + "tokens_seen": 1987444736 + }, + { + "epoch": 0.6, + "learning_rate": 0.00032154062186559677, + "loss": 1.3661, + "theoretical_loss": 3.460653714673674, + "tokens_seen": 1987510272 + }, + { + "epoch": 0.6, + "learning_rate": 0.00032152457372116347, + "loss": 1.297, + "theoretical_loss": 3.460644267160829, + "tokens_seen": 1987575808 + }, + { + "epoch": 0.6, + "learning_rate": 0.00032150852557673017, + "loss": 1.4924, + "theoretical_loss": 3.4606348200467094, + "tokens_seen": 1987641344 + }, + { + "epoch": 0.6, + "learning_rate": 0.00032149247743229687, + "loss": 1.4349, + "theoretical_loss": 3.4606253733312857, + "tokens_seen": 1987706880 + }, + { + "epoch": 0.6, + "learning_rate": 0.00032147642928786357, + "loss": 1.3891, + "theoretical_loss": 3.460615927014528, + "tokens_seen": 1987772416 + }, + { + "epoch": 0.6, + "learning_rate": 0.00032146038114343027, + "loss": 1.3406, + "theoretical_loss": 3.4606064810964066, + "tokens_seen": 1987837952 + }, + { + "epoch": 0.6, + "learning_rate": 0.00032144433299899697, + "loss": 1.2691, + "theoretical_loss": 3.4605970355768916, + "tokens_seen": 1987903488 + }, + { + "epoch": 0.6, + "learning_rate": 0.00032142828485456367, + "loss": 1.4156, + "theoretical_loss": 3.460587590455952, + "tokens_seen": 1987969024 + }, + { + "epoch": 0.6, + "learning_rate": 0.00032141223671013037, + "loss": 1.226, + "theoretical_loss": 3.4605781457335594, + "tokens_seen": 1988034560 + }, + { + "epoch": 0.6, + "learning_rate": 0.00032139618856569707, + "loss": 1.2576, + "theoretical_loss": 3.4605687014096826, + "tokens_seen": 1988100096 + }, + { + "epoch": 0.6, + "learning_rate": 0.00032138014042126377, + "loss": 1.4222, + "theoretical_loss": 3.4605592574842916, + "tokens_seen": 1988165632 + }, + { + "epoch": 0.6, + "learning_rate": 0.00032136409227683047, + "loss": 1.2428, + "theoretical_loss": 3.460549813957358, + "tokens_seen": 1988231168 + }, + { + "epoch": 0.6, + "learning_rate": 0.00032134804413239717, + "loss": 1.358, + "theoretical_loss": 3.46054037082885, + "tokens_seen": 1988296704 + }, + { + "epoch": 0.6, + "learning_rate": 0.0003213319959879639, + "loss": 1.2641, + "theoretical_loss": 3.4605309280987386, + "tokens_seen": 1988362240 + }, + { + "epoch": 0.6, + "learning_rate": 0.00032131594784353063, + "loss": 1.3151, + "theoretical_loss": 3.4605214857669937, + "tokens_seen": 1988427776 + }, + { + "epoch": 0.6, + "learning_rate": 0.00032129989969909733, + "loss": 1.3595, + "theoretical_loss": 3.4605120438335852, + "tokens_seen": 1988493312 + }, + { + "epoch": 0.6, + "learning_rate": 0.00032128385155466403, + "loss": 1.3758, + "theoretical_loss": 3.4605026022984835, + "tokens_seen": 1988558848 + }, + { + "epoch": 0.6, + "learning_rate": 0.00032126780341023073, + "loss": 1.1608, + "theoretical_loss": 3.4604931611616587, + "tokens_seen": 1988624384 + }, + { + "epoch": 0.6, + "learning_rate": 0.00032125175526579743, + "loss": 1.2506, + "theoretical_loss": 3.4604837204230803, + "tokens_seen": 1988689920 + }, + { + "epoch": 0.6, + "learning_rate": 0.00032123570712136413, + "loss": 1.3716, + "theoretical_loss": 3.460474280082719, + "tokens_seen": 1988755456 + }, + { + "epoch": 0.6, + "learning_rate": 0.00032121965897693083, + "loss": 1.3914, + "theoretical_loss": 3.4604648401405447, + "tokens_seen": 1988820992 + }, + { + "epoch": 0.6, + "learning_rate": 0.00032120361083249753, + "loss": 1.2912, + "theoretical_loss": 3.460455400596527, + "tokens_seen": 1988886528 + }, + { + "epoch": 0.6, + "learning_rate": 0.00032118756268806423, + "loss": 1.3144, + "theoretical_loss": 3.460445961450637, + "tokens_seen": 1988952064 + }, + { + "epoch": 0.6, + "objective/train/docs_used": 725238, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.4862852096557617, + "objective/train/theoretical_loss": 3.4604365227028437, + "objective/train/tokens_used": 2009477600, + "theoretical_loss": 3.4604365227028437, + "tokens_seen": 1989017600 + }, + { + "epoch": 0.6, + "learning_rate": 0.00032117151454363093, + "loss": 1.4255, + "theoretical_loss": 3.4604365227028437, + "tokens_seen": 1989017600 + }, + { + "epoch": 0.6, + "learning_rate": 0.00032115546639919763, + "loss": 1.3646, + "theoretical_loss": 3.460427084353118, + "tokens_seen": 1989083136 + }, + { + "epoch": 0.6, + "learning_rate": 0.00032113941825476434, + "loss": 1.379, + "theoretical_loss": 3.4604176464014293, + "tokens_seen": 1989148672 + }, + { + "epoch": 0.6, + "learning_rate": 0.00032112337011033104, + "loss": 1.3541, + "theoretical_loss": 3.4604082088477486, + "tokens_seen": 1989214208 + }, + { + "epoch": 0.6, + "learning_rate": 0.00032110732196589774, + "loss": 1.4063, + "theoretical_loss": 3.460398771692045, + "tokens_seen": 1989279744 + }, + { + "epoch": 0.6, + "learning_rate": 0.00032109127382146444, + "loss": 1.2615, + "theoretical_loss": 3.4603893349342894, + "tokens_seen": 1989345280 + }, + { + "epoch": 0.6, + "learning_rate": 0.00032107522567703114, + "loss": 1.3605, + "theoretical_loss": 3.460379898574452, + "tokens_seen": 1989410816 + }, + { + "epoch": 0.6, + "learning_rate": 0.00032105917753259784, + "loss": 1.3337, + "theoretical_loss": 3.4603704626125023, + "tokens_seen": 1989476352 + }, + { + "epoch": 0.6, + "learning_rate": 0.00032104312938816454, + "loss": 1.4202, + "theoretical_loss": 3.4603610270484104, + "tokens_seen": 1989541888 + }, + { + "epoch": 0.6, + "learning_rate": 0.00032102708124373124, + "loss": 1.3664, + "theoretical_loss": 3.4603515918821466, + "tokens_seen": 1989607424 + }, + { + "epoch": 0.6, + "learning_rate": 0.00032101103309929794, + "loss": 1.332, + "theoretical_loss": 3.460342157113681, + "tokens_seen": 1989672960 + }, + { + "epoch": 0.6, + "learning_rate": 0.00032099498495486464, + "loss": 1.3038, + "theoretical_loss": 3.4603327227429843, + "tokens_seen": 1989738496 + }, + { + "epoch": 0.6, + "learning_rate": 0.00032097893681043134, + "loss": 1.3255, + "theoretical_loss": 3.460323288770026, + "tokens_seen": 1989804032 + }, + { + "epoch": 0.6, + "learning_rate": 0.00032096288866599804, + "loss": 1.298, + "theoretical_loss": 3.460313855194776, + "tokens_seen": 1989869568 + }, + { + "epoch": 0.6, + "learning_rate": 0.00032094684052156474, + "loss": 1.2696, + "theoretical_loss": 3.460304422017205, + "tokens_seen": 1989935104 + }, + { + "epoch": 0.6, + "learning_rate": 0.00032093079237713144, + "loss": 1.376, + "theoretical_loss": 3.4602949892372834, + "tokens_seen": 1990000640 + }, + { + "epoch": 0.6, + "learning_rate": 0.0003209147442326981, + "loss": 1.4992, + "theoretical_loss": 3.4602855568549806, + "tokens_seen": 1990066176 + }, + { + "epoch": 0.6, + "learning_rate": 0.0003208986960882648, + "loss": 1.327, + "theoretical_loss": 3.4602761248702665, + "tokens_seen": 1990131712 + }, + { + "epoch": 0.6, + "learning_rate": 0.0003208826479438315, + "loss": 1.3824, + "theoretical_loss": 3.4602666932831125, + "tokens_seen": 1990197248 + }, + { + "epoch": 0.6, + "learning_rate": 0.0003208665997993982, + "loss": 1.3329, + "theoretical_loss": 3.4602572620934877, + "tokens_seen": 1990262784 + }, + { + "epoch": 0.6, + "learning_rate": 0.0003208505516549649, + "loss": 1.3918, + "theoretical_loss": 3.4602478313013627, + "tokens_seen": 1990328320 + }, + { + "epoch": 0.6, + "learning_rate": 0.0003208345035105316, + "loss": 1.2524, + "theoretical_loss": 3.4602384009067073, + "tokens_seen": 1990393856 + }, + { + "epoch": 0.6, + "learning_rate": 0.0003208184553660983, + "loss": 1.3496, + "theoretical_loss": 3.460228970909492, + "tokens_seen": 1990459392 + }, + { + "epoch": 0.6, + "learning_rate": 0.000320802407221665, + "loss": 1.3171, + "theoretical_loss": 3.4602195413096877, + "tokens_seen": 1990524928 + }, + { + "epoch": 0.6, + "learning_rate": 0.0003207863590772317, + "loss": 1.3977, + "theoretical_loss": 3.460210112107263, + "tokens_seen": 1990590464 + }, + { + "epoch": 0.6, + "learning_rate": 0.0003207703109327984, + "loss": 1.3679, + "theoretical_loss": 3.4602006833021886, + "tokens_seen": 1990656000 + }, + { + "epoch": 0.6, + "learning_rate": 0.0003207542627883651, + "loss": 1.3713, + "theoretical_loss": 3.4601912548944354, + "tokens_seen": 1990721536 + }, + { + "epoch": 0.6, + "learning_rate": 0.0003207382146439318, + "loss": 1.3774, + "theoretical_loss": 3.4601818268839732, + "tokens_seen": 1990787072 + }, + { + "epoch": 0.6, + "learning_rate": 0.0003207221664994985, + "loss": 1.2485, + "theoretical_loss": 3.4601723992707716, + "tokens_seen": 1990852608 + }, + { + "epoch": 0.6, + "learning_rate": 0.0003207061183550652, + "loss": 1.304, + "theoretical_loss": 3.4601629720548015, + "tokens_seen": 1990918144 + }, + { + "epoch": 0.6, + "learning_rate": 0.0003206900702106319, + "loss": 1.3443, + "theoretical_loss": 3.460153545236033, + "tokens_seen": 1990983680 + }, + { + "epoch": 0.6, + "learning_rate": 0.0003206740220661986, + "loss": 1.3332, + "theoretical_loss": 3.4601441188144353, + "tokens_seen": 1991049216 + }, + { + "epoch": 0.6, + "learning_rate": 0.0003206579739217653, + "loss": 1.3377, + "theoretical_loss": 3.46013469278998, + "tokens_seen": 1991114752 + }, + { + "epoch": 0.6, + "learning_rate": 0.000320641925777332, + "loss": 1.3258, + "theoretical_loss": 3.4601252671626366, + "tokens_seen": 1991180288 + }, + { + "epoch": 0.6, + "learning_rate": 0.0003206258776328987, + "loss": 1.3584, + "theoretical_loss": 3.460115841932376, + "tokens_seen": 1991245824 + }, + { + "epoch": 0.6, + "learning_rate": 0.0003206098294884654, + "loss": 1.2959, + "theoretical_loss": 3.4601064170991673, + "tokens_seen": 1991311360 + }, + { + "epoch": 0.6, + "learning_rate": 0.0003205937813440321, + "loss": 1.4052, + "theoretical_loss": 3.460096992662981, + "tokens_seen": 1991376896 + }, + { + "epoch": 0.6, + "learning_rate": 0.0003205777331995988, + "loss": 1.3702, + "theoretical_loss": 3.4600875686237877, + "tokens_seen": 1991442432 + }, + { + "epoch": 0.6, + "learning_rate": 0.0003205616850551655, + "loss": 1.3523, + "theoretical_loss": 3.4600781449815576, + "tokens_seen": 1991507968 + }, + { + "epoch": 0.6, + "learning_rate": 0.0003205456369107322, + "loss": 1.3679, + "theoretical_loss": 3.4600687217362607, + "tokens_seen": 1991573504 + }, + { + "epoch": 0.6, + "learning_rate": 0.0003205295887662989, + "loss": 1.3866, + "theoretical_loss": 3.4600592988878676, + "tokens_seen": 1991639040 + }, + { + "epoch": 0.6, + "learning_rate": 0.0003205135406218656, + "loss": 1.3055, + "theoretical_loss": 3.4600498764363476, + "tokens_seen": 1991704576 + }, + { + "epoch": 0.6, + "learning_rate": 0.0003204974924774323, + "loss": 1.3522, + "theoretical_loss": 3.460040454381672, + "tokens_seen": 1991770112 + }, + { + "epoch": 0.6, + "learning_rate": 0.000320481444332999, + "loss": 1.4445, + "theoretical_loss": 3.46003103272381, + "tokens_seen": 1991835648 + }, + { + "epoch": 0.6, + "learning_rate": 0.0003204653961885657, + "loss": 1.3311, + "theoretical_loss": 3.460021611462733, + "tokens_seen": 1991901184 + }, + { + "epoch": 0.6, + "learning_rate": 0.0003204493480441324, + "loss": 1.4659, + "theoretical_loss": 3.4600121905984103, + "tokens_seen": 1991966720 + }, + { + "epoch": 0.6, + "learning_rate": 0.0003204332998996991, + "loss": 1.3263, + "theoretical_loss": 3.4600027701308127, + "tokens_seen": 1992032256 + }, + { + "epoch": 0.6, + "learning_rate": 0.0003204172517552658, + "loss": 1.4096, + "theoretical_loss": 3.45999335005991, + "tokens_seen": 1992097792 + }, + { + "epoch": 0.6, + "learning_rate": 0.0003204012036108325, + "loss": 1.277, + "theoretical_loss": 3.4599839303856728, + "tokens_seen": 1992163328 + }, + { + "epoch": 0.6, + "learning_rate": 0.0003203851554663992, + "loss": 1.3577, + "theoretical_loss": 3.4599745111080713, + "tokens_seen": 1992228864 + }, + { + "epoch": 0.6, + "objective/train/docs_used": 726441, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.3859813213348389, + "objective/train/theoretical_loss": 3.4599650922270753, + "objective/train/tokens_used": 2012754400, + "theoretical_loss": 3.4599650922270753, + "tokens_seen": 1992294400 + }, + { + "epoch": 0.6, + "learning_rate": 0.0003203691073219659, + "loss": 1.3126, + "theoretical_loss": 3.4599650922270753, + "tokens_seen": 1992294400 + }, + { + "epoch": 0.6, + "learning_rate": 0.0003203530591775326, + "loss": 1.3231, + "theoretical_loss": 3.4599556737426553, + "tokens_seen": 1992359936 + }, + { + "epoch": 0.6, + "learning_rate": 0.0003203370110330993, + "loss": 1.3812, + "theoretical_loss": 3.459946255654782, + "tokens_seen": 1992425472 + }, + { + "epoch": 0.6, + "learning_rate": 0.000320320962888666, + "loss": 1.3539, + "theoretical_loss": 3.4599368379634257, + "tokens_seen": 1992491008 + }, + { + "epoch": 0.6, + "learning_rate": 0.0003203049147442327, + "loss": 1.3103, + "theoretical_loss": 3.4599274206685555, + "tokens_seen": 1992556544 + }, + { + "epoch": 0.6, + "learning_rate": 0.0003202888665997994, + "loss": 1.3308, + "theoretical_loss": 3.4599180037701434, + "tokens_seen": 1992622080 + }, + { + "epoch": 0.6, + "learning_rate": 0.0003202728184553661, + "loss": 1.4208, + "theoretical_loss": 3.459908587268158, + "tokens_seen": 1992687616 + }, + { + "epoch": 0.6, + "learning_rate": 0.0003202567703109328, + "loss": 1.307, + "theoretical_loss": 3.4598991711625704, + "tokens_seen": 1992753152 + }, + { + "epoch": 0.6, + "learning_rate": 0.0003202407221664995, + "loss": 1.3207, + "theoretical_loss": 3.459889755453351, + "tokens_seen": 1992818688 + }, + { + "epoch": 0.6, + "learning_rate": 0.0003202246740220662, + "loss": 1.2718, + "theoretical_loss": 3.45988034014047, + "tokens_seen": 1992884224 + }, + { + "epoch": 0.6, + "learning_rate": 0.0003202086258776329, + "loss": 1.2881, + "theoretical_loss": 3.459870925223897, + "tokens_seen": 1992949760 + }, + { + "epoch": 0.6, + "learning_rate": 0.0003201925777331996, + "loss": 1.3798, + "theoretical_loss": 3.4598615107036035, + "tokens_seen": 1993015296 + }, + { + "epoch": 0.6, + "learning_rate": 0.0003201765295887663, + "loss": 1.3784, + "theoretical_loss": 3.4598520965795587, + "tokens_seen": 1993080832 + }, + { + "epoch": 0.6, + "learning_rate": 0.000320160481444333, + "loss": 1.2576, + "theoretical_loss": 3.4598426828517335, + "tokens_seen": 1993146368 + }, + { + "epoch": 0.6, + "learning_rate": 0.00032014443329989967, + "loss": 1.3374, + "theoretical_loss": 3.4598332695200984, + "tokens_seen": 1993211904 + }, + { + "epoch": 0.6, + "learning_rate": 0.00032012838515546637, + "loss": 1.3854, + "theoretical_loss": 3.459823856584623, + "tokens_seen": 1993277440 + }, + { + "epoch": 0.6, + "learning_rate": 0.0003201123370110331, + "loss": 1.3419, + "theoretical_loss": 3.4598144440452776, + "tokens_seen": 1993342976 + }, + { + "epoch": 0.6, + "learning_rate": 0.0003200962888665998, + "loss": 1.2437, + "theoretical_loss": 3.459805031902033, + "tokens_seen": 1993408512 + }, + { + "epoch": 0.6, + "learning_rate": 0.0003200802407221665, + "loss": 1.3784, + "theoretical_loss": 3.459795620154859, + "tokens_seen": 1993474048 + }, + { + "epoch": 0.6, + "learning_rate": 0.0003200641925777332, + "loss": 1.2876, + "theoretical_loss": 3.4597862088037266, + "tokens_seen": 1993539584 + }, + { + "epoch": 0.6, + "learning_rate": 0.0003200481444332999, + "loss": 1.3801, + "theoretical_loss": 3.4597767978486065, + "tokens_seen": 1993605120 + }, + { + "epoch": 0.6, + "learning_rate": 0.0003200320962888666, + "loss": 1.3045, + "theoretical_loss": 3.4597673872894674, + "tokens_seen": 1993670656 + }, + { + "epoch": 0.6, + "learning_rate": 0.0003200160481444333, + "loss": 1.4689, + "theoretical_loss": 3.459757977126281, + "tokens_seen": 1993736192 + }, + { + "epoch": 0.6, + "learning_rate": 0.00032, + "loss": 1.3792, + "theoretical_loss": 3.4597485673590165, + "tokens_seen": 1993801728 + }, + { + "epoch": 0.6, + "learning_rate": 0.0003199839518555667, + "loss": 1.3809, + "theoretical_loss": 3.459739157987645, + "tokens_seen": 1993867264 + }, + { + "epoch": 0.6, + "learning_rate": 0.00031996790371113343, + "loss": 1.4135, + "theoretical_loss": 3.459729749012137, + "tokens_seen": 1993932800 + }, + { + "epoch": 0.6, + "learning_rate": 0.00031995185556670013, + "loss": 1.3429, + "theoretical_loss": 3.4597203404324626, + "tokens_seen": 1993998336 + }, + { + "epoch": 0.6, + "learning_rate": 0.00031993580742226683, + "loss": 1.3064, + "theoretical_loss": 3.459710932248592, + "tokens_seen": 1994063872 + }, + { + "epoch": 0.6, + "learning_rate": 0.00031991975927783353, + "loss": 1.258, + "theoretical_loss": 3.4597015244604954, + "tokens_seen": 1994129408 + }, + { + "epoch": 0.6, + "learning_rate": 0.00031990371113340023, + "loss": 1.4218, + "theoretical_loss": 3.459692117068144, + "tokens_seen": 1994194944 + }, + { + "epoch": 0.6, + "learning_rate": 0.00031988766298896693, + "loss": 1.3124, + "theoretical_loss": 3.459682710071507, + "tokens_seen": 1994260480 + }, + { + "epoch": 0.6, + "learning_rate": 0.00031987161484453363, + "loss": 1.3377, + "theoretical_loss": 3.459673303470555, + "tokens_seen": 1994326016 + }, + { + "epoch": 0.6, + "learning_rate": 0.00031985556670010033, + "loss": 1.3263, + "theoretical_loss": 3.4596638972652594, + "tokens_seen": 1994391552 + }, + { + "epoch": 0.6, + "learning_rate": 0.00031983951855566703, + "loss": 1.3543, + "theoretical_loss": 3.4596544914555887, + "tokens_seen": 1994457088 + }, + { + "epoch": 0.6, + "learning_rate": 0.00031982347041123373, + "loss": 1.2669, + "theoretical_loss": 3.4596450860415153, + "tokens_seen": 1994522624 + }, + { + "epoch": 0.6, + "learning_rate": 0.00031980742226680043, + "loss": 1.3733, + "theoretical_loss": 3.4596356810230082, + "tokens_seen": 1994588160 + }, + { + "epoch": 0.6, + "learning_rate": 0.00031979137412236713, + "loss": 1.3987, + "theoretical_loss": 3.459626276400038, + "tokens_seen": 1994653696 + }, + { + "epoch": 0.6, + "learning_rate": 0.00031977532597793384, + "loss": 1.2826, + "theoretical_loss": 3.4596168721725755, + "tokens_seen": 1994719232 + }, + { + "epoch": 0.6, + "learning_rate": 0.00031975927783350054, + "loss": 1.3179, + "theoretical_loss": 3.459607468340591, + "tokens_seen": 1994784768 + }, + { + "epoch": 0.6, + "learning_rate": 0.00031974322968906724, + "loss": 1.3459, + "theoretical_loss": 3.4595980649040543, + "tokens_seen": 1994850304 + }, + { + "epoch": 0.6, + "learning_rate": 0.00031972718154463394, + "loss": 1.3141, + "theoretical_loss": 3.459588661862936, + "tokens_seen": 1994915840 + }, + { + "epoch": 0.6, + "learning_rate": 0.00031971113340020064, + "loss": 1.3969, + "theoretical_loss": 3.459579259217207, + "tokens_seen": 1994981376 + }, + { + "epoch": 0.6, + "learning_rate": 0.00031969508525576734, + "loss": 1.3475, + "theoretical_loss": 3.459569856966837, + "tokens_seen": 1995046912 + }, + { + "epoch": 0.6, + "learning_rate": 0.00031967903711133404, + "loss": 1.376, + "theoretical_loss": 3.4595604551117973, + "tokens_seen": 1995112448 + }, + { + "epoch": 0.6, + "learning_rate": 0.00031966298896690074, + "loss": 1.3392, + "theoretical_loss": 3.4595510536520573, + "tokens_seen": 1995177984 + }, + { + "epoch": 0.6, + "learning_rate": 0.00031964694082246744, + "loss": 1.3322, + "theoretical_loss": 3.459541652587588, + "tokens_seen": 1995243520 + }, + { + "epoch": 0.6, + "learning_rate": 0.00031963089267803414, + "loss": 1.3497, + "theoretical_loss": 3.4595322519183593, + "tokens_seen": 1995309056 + }, + { + "epoch": 0.6, + "learning_rate": 0.00031961484453360084, + "loss": 1.3268, + "theoretical_loss": 3.4595228516443424, + "tokens_seen": 1995374592 + }, + { + "epoch": 0.6, + "learning_rate": 0.00031959879638916754, + "loss": 1.2995, + "theoretical_loss": 3.459513451765507, + "tokens_seen": 1995440128 + }, + { + "epoch": 0.6, + "learning_rate": 0.00031958274824473424, + "loss": 1.4152, + "theoretical_loss": 3.4595040522818232, + "tokens_seen": 1995505664 + }, + { + "epoch": 0.6, + "objective/train/docs_used": 727673, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.1932373046875, + "objective/train/theoretical_loss": 3.4594946531932624, + "objective/train/tokens_used": 2016031200, + "theoretical_loss": 3.4594946531932624, + "tokens_seen": 1995571200 + }, + { + "epoch": 0.6, + "learning_rate": 0.00031956670010030094, + "loss": 1.2531, + "theoretical_loss": 3.4594946531932624, + "tokens_seen": 1995571200 + }, + { + "epoch": 0.6, + "learning_rate": 0.00031955065195586764, + "loss": 1.3751, + "theoretical_loss": 3.459485254499795, + "tokens_seen": 1995636736 + }, + { + "epoch": 0.6, + "learning_rate": 0.00031953460381143434, + "loss": 1.306, + "theoretical_loss": 3.4594758562013905, + "tokens_seen": 1995702272 + }, + { + "epoch": 0.6, + "learning_rate": 0.00031951855566700105, + "loss": 1.4128, + "theoretical_loss": 3.45946645829802, + "tokens_seen": 1995767808 + }, + { + "epoch": 0.6, + "learning_rate": 0.0003195025075225677, + "loss": 1.3249, + "theoretical_loss": 3.4594570607896538, + "tokens_seen": 1995833344 + }, + { + "epoch": 0.6, + "learning_rate": 0.0003194864593781344, + "loss": 1.3346, + "theoretical_loss": 3.4594476636762614, + "tokens_seen": 1995898880 + }, + { + "epoch": 0.6, + "learning_rate": 0.0003194704112337011, + "loss": 1.4391, + "theoretical_loss": 3.459438266957815, + "tokens_seen": 1995964416 + }, + { + "epoch": 0.6, + "learning_rate": 0.0003194543630892678, + "loss": 1.3801, + "theoretical_loss": 3.4594288706342837, + "tokens_seen": 1996029952 + }, + { + "epoch": 0.6, + "learning_rate": 0.0003194383149448345, + "loss": 1.3295, + "theoretical_loss": 3.459419474705639, + "tokens_seen": 1996095488 + }, + { + "epoch": 0.6, + "learning_rate": 0.0003194222668004012, + "loss": 1.4303, + "theoretical_loss": 3.4594100791718496, + "tokens_seen": 1996161024 + }, + { + "epoch": 0.6, + "learning_rate": 0.0003194062186559679, + "loss": 1.2945, + "theoretical_loss": 3.4594006840328877, + "tokens_seen": 1996226560 + }, + { + "epoch": 0.6, + "learning_rate": 0.0003193901705115346, + "loss": 1.3017, + "theoretical_loss": 3.459391289288723, + "tokens_seen": 1996292096 + }, + { + "epoch": 0.6, + "learning_rate": 0.0003193741223671013, + "loss": 1.3844, + "theoretical_loss": 3.459381894939326, + "tokens_seen": 1996357632 + }, + { + "epoch": 0.6, + "learning_rate": 0.000319358074222668, + "loss": 1.3219, + "theoretical_loss": 3.459372500984667, + "tokens_seen": 1996423168 + }, + { + "epoch": 0.6, + "learning_rate": 0.0003193420260782347, + "loss": 1.329, + "theoretical_loss": 3.459363107424717, + "tokens_seen": 1996488704 + }, + { + "epoch": 0.61, + "learning_rate": 0.0003193259779338014, + "loss": 1.3318, + "theoretical_loss": 3.459353714259446, + "tokens_seen": 1996554240 + }, + { + "epoch": 0.61, + "learning_rate": 0.0003193099297893681, + "loss": 1.3632, + "theoretical_loss": 3.4593443214888246, + "tokens_seen": 1996619776 + }, + { + "epoch": 0.61, + "learning_rate": 0.0003192938816449348, + "loss": 1.4032, + "theoretical_loss": 3.4593349291128233, + "tokens_seen": 1996685312 + }, + { + "epoch": 0.61, + "learning_rate": 0.0003192778335005015, + "loss": 1.3735, + "theoretical_loss": 3.459325537131412, + "tokens_seen": 1996750848 + }, + { + "epoch": 0.61, + "learning_rate": 0.0003192617853560682, + "loss": 1.3725, + "theoretical_loss": 3.459316145544562, + "tokens_seen": 1996816384 + }, + { + "epoch": 0.61, + "learning_rate": 0.0003192457372116349, + "loss": 1.3266, + "theoretical_loss": 3.4593067543522436, + "tokens_seen": 1996881920 + }, + { + "epoch": 0.61, + "learning_rate": 0.0003192296890672016, + "loss": 1.2452, + "theoretical_loss": 3.459297363554427, + "tokens_seen": 1996947456 + }, + { + "epoch": 0.61, + "learning_rate": 0.0003192136409227683, + "loss": 1.3948, + "theoretical_loss": 3.4592879731510826, + "tokens_seen": 1997012992 + }, + { + "epoch": 0.61, + "learning_rate": 0.000319197592778335, + "loss": 1.4199, + "theoretical_loss": 3.459278583142181, + "tokens_seen": 1997078528 + }, + { + "epoch": 0.61, + "learning_rate": 0.0003191815446339017, + "loss": 1.2815, + "theoretical_loss": 3.4592691935276934, + "tokens_seen": 1997144064 + }, + { + "epoch": 0.61, + "learning_rate": 0.0003191654964894684, + "loss": 1.3558, + "theoretical_loss": 3.4592598043075893, + "tokens_seen": 1997209600 + }, + { + "epoch": 0.61, + "learning_rate": 0.0003191494483450351, + "loss": 1.2372, + "theoretical_loss": 3.4592504154818395, + "tokens_seen": 1997275136 + }, + { + "epoch": 0.61, + "learning_rate": 0.0003191334002006018, + "loss": 1.4338, + "theoretical_loss": 3.459241027050415, + "tokens_seen": 1997340672 + }, + { + "epoch": 0.61, + "learning_rate": 0.0003191173520561685, + "loss": 1.2888, + "theoretical_loss": 3.4592316390132853, + "tokens_seen": 1997406208 + }, + { + "epoch": 0.61, + "learning_rate": 0.0003191013039117352, + "loss": 1.2915, + "theoretical_loss": 3.459222251370422, + "tokens_seen": 1997471744 + }, + { + "epoch": 0.61, + "learning_rate": 0.0003190852557673019, + "loss": 1.4353, + "theoretical_loss": 3.4592128641217945, + "tokens_seen": 1997537280 + }, + { + "epoch": 0.61, + "learning_rate": 0.0003190692076228686, + "loss": 1.5266, + "theoretical_loss": 3.459203477267374, + "tokens_seen": 1997602816 + }, + { + "epoch": 0.61, + "learning_rate": 0.0003190531594784353, + "loss": 1.3657, + "theoretical_loss": 3.459194090807131, + "tokens_seen": 1997668352 + }, + { + "epoch": 0.61, + "learning_rate": 0.000319037111334002, + "loss": 1.336, + "theoretical_loss": 3.4591847047410362, + "tokens_seen": 1997733888 + }, + { + "epoch": 0.61, + "learning_rate": 0.0003190210631895687, + "loss": 1.2984, + "theoretical_loss": 3.4591753190690593, + "tokens_seen": 1997799424 + }, + { + "epoch": 0.61, + "learning_rate": 0.0003190050150451354, + "loss": 1.3581, + "theoretical_loss": 3.459165933791172, + "tokens_seen": 1997864960 + }, + { + "epoch": 0.61, + "learning_rate": 0.0003189889669007021, + "loss": 1.4721, + "theoretical_loss": 3.4591565489073437, + "tokens_seen": 1997930496 + }, + { + "epoch": 0.61, + "learning_rate": 0.0003189729187562688, + "loss": 1.3205, + "theoretical_loss": 3.4591471644175456, + "tokens_seen": 1997996032 + }, + { + "epoch": 0.61, + "learning_rate": 0.0003189568706118355, + "loss": 1.2691, + "theoretical_loss": 3.4591377803217482, + "tokens_seen": 1998061568 + }, + { + "epoch": 0.61, + "learning_rate": 0.0003189408224674022, + "loss": 1.4785, + "theoretical_loss": 3.4591283966199216, + "tokens_seen": 1998127104 + }, + { + "epoch": 0.61, + "learning_rate": 0.0003189247743229689, + "loss": 1.3951, + "theoretical_loss": 3.4591190133120366, + "tokens_seen": 1998192640 + }, + { + "epoch": 0.61, + "learning_rate": 0.00031890872617853567, + "loss": 1.4023, + "theoretical_loss": 3.459109630398064, + "tokens_seen": 1998258176 + }, + { + "epoch": 0.61, + "learning_rate": 0.00031889267803410237, + "loss": 1.3449, + "theoretical_loss": 3.4591002478779744, + "tokens_seen": 1998323712 + }, + { + "epoch": 0.61, + "learning_rate": 0.000318876629889669, + "loss": 1.2325, + "theoretical_loss": 3.4590908657517376, + "tokens_seen": 1998389248 + }, + { + "epoch": 0.61, + "learning_rate": 0.0003188605817452357, + "loss": 1.2903, + "theoretical_loss": 3.4590814840193245, + "tokens_seen": 1998454784 + }, + { + "epoch": 0.61, + "learning_rate": 0.0003188445336008024, + "loss": 1.3507, + "theoretical_loss": 3.459072102680706, + "tokens_seen": 1998520320 + }, + { + "epoch": 0.61, + "learning_rate": 0.0003188284854563691, + "loss": 1.4301, + "theoretical_loss": 3.4590627217358523, + "tokens_seen": 1998585856 + }, + { + "epoch": 0.61, + "learning_rate": 0.0003188124373119358, + "loss": 1.3117, + "theoretical_loss": 3.4590533411847346, + "tokens_seen": 1998651392 + }, + { + "epoch": 0.61, + "learning_rate": 0.0003187963891675025, + "loss": 1.405, + "theoretical_loss": 3.459043961027322, + "tokens_seen": 1998716928 + }, + { + "epoch": 0.61, + "learning_rate": 0.0003187803410230692, + "loss": 1.3276, + "theoretical_loss": 3.4590345812635865, + "tokens_seen": 1998782464 + }, + { + "epoch": 0.61, + "objective/train/docs_used": 728818, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.3868350982666016, + "objective/train/theoretical_loss": 3.459025201893499, + "objective/train/tokens_used": 2019308000, + "theoretical_loss": 3.459025201893499, + "tokens_seen": 1998848000 + }, + { + "epoch": 0.61, + "learning_rate": 0.0003187642928786359, + "loss": 1.3489, + "theoretical_loss": 3.459025201893499, + "tokens_seen": 1998848000 + }, + { + "epoch": 0.61, + "learning_rate": 0.0003187482447342026, + "loss": 1.2908, + "theoretical_loss": 3.459015822917028, + "tokens_seen": 1998913536 + }, + { + "epoch": 0.61, + "learning_rate": 0.0003187321965897693, + "loss": 1.3846, + "theoretical_loss": 3.459006444334146, + "tokens_seen": 1998979072 + }, + { + "epoch": 0.61, + "learning_rate": 0.000318716148445336, + "loss": 1.25, + "theoretical_loss": 3.4589970661448226, + "tokens_seen": 1999044608 + }, + { + "epoch": 0.61, + "learning_rate": 0.0003187001003009027, + "loss": 1.4029, + "theoretical_loss": 3.458987688349029, + "tokens_seen": 1999110144 + }, + { + "epoch": 0.61, + "learning_rate": 0.0003186840521564694, + "loss": 1.3635, + "theoretical_loss": 3.458978310946735, + "tokens_seen": 1999175680 + }, + { + "epoch": 0.61, + "learning_rate": 0.0003186680040120361, + "loss": 1.3954, + "theoretical_loss": 3.4589689339379124, + "tokens_seen": 1999241216 + }, + { + "epoch": 0.61, + "learning_rate": 0.0003186519558676028, + "loss": 1.2464, + "theoretical_loss": 3.4589595573225305, + "tokens_seen": 1999306752 + }, + { + "epoch": 0.61, + "learning_rate": 0.0003186359077231695, + "loss": 1.3428, + "theoretical_loss": 3.458950181100561, + "tokens_seen": 1999372288 + }, + { + "epoch": 0.61, + "learning_rate": 0.0003186198595787362, + "loss": 1.2875, + "theoretical_loss": 3.4589408052719737, + "tokens_seen": 1999437824 + }, + { + "epoch": 0.61, + "learning_rate": 0.00031860381143430293, + "loss": 1.4004, + "theoretical_loss": 3.4589314298367393, + "tokens_seen": 1999503360 + }, + { + "epoch": 0.61, + "learning_rate": 0.00031858776328986963, + "loss": 1.3009, + "theoretical_loss": 3.4589220547948285, + "tokens_seen": 1999568896 + }, + { + "epoch": 0.61, + "learning_rate": 0.00031857171514543633, + "loss": 1.2807, + "theoretical_loss": 3.4589126801462124, + "tokens_seen": 1999634432 + }, + { + "epoch": 0.61, + "learning_rate": 0.00031855566700100303, + "loss": 1.2726, + "theoretical_loss": 3.458903305890861, + "tokens_seen": 1999699968 + }, + { + "epoch": 0.61, + "learning_rate": 0.00031853961885656973, + "loss": 1.4695, + "theoretical_loss": 3.4588939320287455, + "tokens_seen": 1999765504 + }, + { + "epoch": 0.61, + "learning_rate": 0.00031852357071213643, + "loss": 1.5345, + "theoretical_loss": 3.458884558559835, + "tokens_seen": 1999831040 + }, + { + "epoch": 0.61, + "learning_rate": 0.00031850752256770313, + "loss": 1.4048, + "theoretical_loss": 3.4588751854841027, + "tokens_seen": 1999896576 + }, + { + "epoch": 0.61, + "learning_rate": 0.00031849147442326983, + "loss": 1.3142, + "theoretical_loss": 3.4588658128015166, + "tokens_seen": 1999962112 + }, + { + "epoch": 0.61, + "learning_rate": 0.00031847542627883653, + "loss": 1.4008, + "theoretical_loss": 3.458856440512049, + "tokens_seen": 2000027648 + }, + { + "epoch": 0.61, + "learning_rate": 0.00031845937813440323, + "loss": 1.4021, + "theoretical_loss": 3.45884706861567, + "tokens_seen": 2000093184 + }, + { + "epoch": 0.61, + "learning_rate": 0.00031844332998996993, + "loss": 1.3877, + "theoretical_loss": 3.4588376971123505, + "tokens_seen": 2000158720 + }, + { + "epoch": 0.61, + "learning_rate": 0.00031842728184553663, + "loss": 1.3203, + "theoretical_loss": 3.4588283260020605, + "tokens_seen": 2000224256 + }, + { + "epoch": 0.61, + "learning_rate": 0.00031841123370110333, + "loss": 1.3939, + "theoretical_loss": 3.4588189552847712, + "tokens_seen": 2000289792 + }, + { + "epoch": 0.61, + "learning_rate": 0.00031839518555667004, + "loss": 1.3924, + "theoretical_loss": 3.458809584960453, + "tokens_seen": 2000355328 + }, + { + "epoch": 0.61, + "learning_rate": 0.00031837913741223674, + "loss": 1.2534, + "theoretical_loss": 3.458800215029077, + "tokens_seen": 2000420864 + }, + { + "epoch": 0.61, + "learning_rate": 0.00031836308926780344, + "loss": 1.3236, + "theoretical_loss": 3.458790845490613, + "tokens_seen": 2000486400 + }, + { + "epoch": 0.61, + "learning_rate": 0.00031834704112337014, + "loss": 1.2222, + "theoretical_loss": 3.458781476345033, + "tokens_seen": 2000551936 + }, + { + "epoch": 0.61, + "learning_rate": 0.00031833099297893684, + "loss": 1.3806, + "theoretical_loss": 3.4587721075923055, + "tokens_seen": 2000617472 + }, + { + "epoch": 0.61, + "learning_rate": 0.00031831494483450354, + "loss": 1.4379, + "theoretical_loss": 3.4587627392324034, + "tokens_seen": 2000683008 + }, + { + "epoch": 0.61, + "learning_rate": 0.00031829889669007024, + "loss": 1.3895, + "theoretical_loss": 3.458753371265296, + "tokens_seen": 2000748544 + }, + { + "epoch": 0.61, + "learning_rate": 0.00031828284854563694, + "loss": 1.257, + "theoretical_loss": 3.4587440036909545, + "tokens_seen": 2000814080 + }, + { + "epoch": 0.61, + "learning_rate": 0.00031826680040120364, + "loss": 1.4182, + "theoretical_loss": 3.4587346365093494, + "tokens_seen": 2000879616 + }, + { + "epoch": 0.61, + "learning_rate": 0.00031825075225677034, + "loss": 1.3775, + "theoretical_loss": 3.4587252697204516, + "tokens_seen": 2000945152 + }, + { + "epoch": 0.61, + "learning_rate": 0.00031823470411233704, + "loss": 1.3426, + "theoretical_loss": 3.458715903324231, + "tokens_seen": 2001010688 + }, + { + "epoch": 0.61, + "learning_rate": 0.00031821865596790374, + "loss": 1.2998, + "theoretical_loss": 3.45870653732066, + "tokens_seen": 2001076224 + }, + { + "epoch": 0.61, + "learning_rate": 0.00031820260782347044, + "loss": 1.2699, + "theoretical_loss": 3.4586971717097077, + "tokens_seen": 2001141760 + }, + { + "epoch": 0.61, + "learning_rate": 0.00031818655967903714, + "loss": 1.3019, + "theoretical_loss": 3.4586878064913447, + "tokens_seen": 2001207296 + }, + { + "epoch": 0.61, + "learning_rate": 0.00031817051153460384, + "loss": 1.2665, + "theoretical_loss": 3.4586784416655423, + "tokens_seen": 2001272832 + }, + { + "epoch": 0.61, + "learning_rate": 0.00031815446339017054, + "loss": 1.3242, + "theoretical_loss": 3.4586690772322717, + "tokens_seen": 2001338368 + }, + { + "epoch": 0.61, + "learning_rate": 0.00031813841524573725, + "loss": 1.3591, + "theoretical_loss": 3.4586597131915027, + "tokens_seen": 2001403904 + }, + { + "epoch": 0.61, + "learning_rate": 0.00031812236710130395, + "loss": 1.4278, + "theoretical_loss": 3.4586503495432064, + "tokens_seen": 2001469440 + }, + { + "epoch": 0.61, + "learning_rate": 0.0003181063189568706, + "loss": 1.3907, + "theoretical_loss": 3.458640986287353, + "tokens_seen": 2001534976 + }, + { + "epoch": 0.61, + "learning_rate": 0.0003180902708124373, + "loss": 1.3371, + "theoretical_loss": 3.4586316234239143, + "tokens_seen": 2001600512 + }, + { + "epoch": 0.61, + "learning_rate": 0.000318074222668004, + "loss": 1.3939, + "theoretical_loss": 3.45862226095286, + "tokens_seen": 2001666048 + }, + { + "epoch": 0.61, + "learning_rate": 0.0003180581745235707, + "loss": 1.3598, + "theoretical_loss": 3.458612898874161, + "tokens_seen": 2001731584 + }, + { + "epoch": 0.61, + "learning_rate": 0.0003180421263791374, + "loss": 1.3375, + "theoretical_loss": 3.458603537187788, + "tokens_seen": 2001797120 + }, + { + "epoch": 0.61, + "learning_rate": 0.0003180260782347041, + "loss": 1.3271, + "theoretical_loss": 3.458594175893712, + "tokens_seen": 2001862656 + }, + { + "epoch": 0.61, + "learning_rate": 0.0003180100300902708, + "loss": 1.4165, + "theoretical_loss": 3.458584814991904, + "tokens_seen": 2001928192 + }, + { + "epoch": 0.61, + "learning_rate": 0.0003179939819458375, + "loss": 1.3313, + "theoretical_loss": 3.4585754544823337, + "tokens_seen": 2001993728 + }, + { + "epoch": 0.61, + "learning_rate": 0.0003179779338014042, + "loss": 1.2529, + "theoretical_loss": 3.458566094364973, + "tokens_seen": 2002059264 + }, + { + "epoch": 0.61, + "objective/train/docs_used": 729965, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.2941763401031494, + "objective/train/theoretical_loss": 3.4585567346397914, + "objective/train/tokens_used": 2022584800, + "theoretical_loss": 3.4585567346397914, + "tokens_seen": 2002124800 + }, + { + "epoch": 0.61, + "learning_rate": 0.0003179618856569709, + "loss": 1.3313, + "theoretical_loss": 3.4585567346397914, + "tokens_seen": 2002124800 + }, + { + "epoch": 0.61, + "learning_rate": 0.0003179458375125376, + "loss": 1.2633, + "theoretical_loss": 3.458547375306761, + "tokens_seen": 2002190336 + }, + { + "epoch": 0.61, + "learning_rate": 0.0003179297893681043, + "loss": 1.4007, + "theoretical_loss": 3.4585380163658512, + "tokens_seen": 2002255872 + }, + { + "epoch": 0.61, + "learning_rate": 0.000317913741223671, + "loss": 1.3159, + "theoretical_loss": 3.4585286578170336, + "tokens_seen": 2002321408 + }, + { + "epoch": 0.61, + "learning_rate": 0.0003178976930792377, + "loss": 1.3528, + "theoretical_loss": 3.4585192996602787, + "tokens_seen": 2002386944 + }, + { + "epoch": 0.61, + "learning_rate": 0.0003178816449348044, + "loss": 1.2556, + "theoretical_loss": 3.4585099418955574, + "tokens_seen": 2002452480 + }, + { + "epoch": 0.61, + "learning_rate": 0.0003178655967903711, + "loss": 1.2867, + "theoretical_loss": 3.45850058452284, + "tokens_seen": 2002518016 + }, + { + "epoch": 0.61, + "learning_rate": 0.0003178495486459378, + "loss": 1.3726, + "theoretical_loss": 3.4584912275420976, + "tokens_seen": 2002583552 + }, + { + "epoch": 0.61, + "learning_rate": 0.0003178335005015045, + "loss": 1.4213, + "theoretical_loss": 3.458481870953301, + "tokens_seen": 2002649088 + }, + { + "epoch": 0.61, + "learning_rate": 0.0003178174523570712, + "loss": 1.3275, + "theoretical_loss": 3.4584725147564206, + "tokens_seen": 2002714624 + }, + { + "epoch": 0.61, + "learning_rate": 0.0003178014042126379, + "loss": 1.3791, + "theoretical_loss": 3.4584631589514276, + "tokens_seen": 2002780160 + }, + { + "epoch": 0.61, + "learning_rate": 0.0003177853560682046, + "loss": 1.3339, + "theoretical_loss": 3.4584538035382923, + "tokens_seen": 2002845696 + }, + { + "epoch": 0.61, + "learning_rate": 0.0003177693079237713, + "loss": 1.3407, + "theoretical_loss": 3.4584444485169863, + "tokens_seen": 2002911232 + }, + { + "epoch": 0.61, + "learning_rate": 0.000317753259779338, + "loss": 1.3203, + "theoretical_loss": 3.4584350938874797, + "tokens_seen": 2002976768 + }, + { + "epoch": 0.61, + "learning_rate": 0.0003177372116349047, + "loss": 1.2535, + "theoretical_loss": 3.4584257396497433, + "tokens_seen": 2003042304 + }, + { + "epoch": 0.61, + "learning_rate": 0.0003177211634904714, + "loss": 1.3184, + "theoretical_loss": 3.4584163858037478, + "tokens_seen": 2003107840 + }, + { + "epoch": 0.61, + "learning_rate": 0.00031770511534603816, + "loss": 1.288, + "theoretical_loss": 3.458407032349464, + "tokens_seen": 2003173376 + }, + { + "epoch": 0.61, + "learning_rate": 0.00031768906720160486, + "loss": 1.4084, + "theoretical_loss": 3.458397679286863, + "tokens_seen": 2003238912 + }, + { + "epoch": 0.61, + "learning_rate": 0.00031767301905717156, + "loss": 1.4002, + "theoretical_loss": 3.4583883266159154, + "tokens_seen": 2003304448 + }, + { + "epoch": 0.61, + "learning_rate": 0.00031765697091273826, + "loss": 1.2386, + "theoretical_loss": 3.458378974336592, + "tokens_seen": 2003369984 + }, + { + "epoch": 0.61, + "learning_rate": 0.00031764092276830497, + "loss": 1.4688, + "theoretical_loss": 3.4583696224488634, + "tokens_seen": 2003435520 + }, + { + "epoch": 0.61, + "learning_rate": 0.00031762487462387167, + "loss": 1.4541, + "theoretical_loss": 3.458360270952701, + "tokens_seen": 2003501056 + }, + { + "epoch": 0.61, + "learning_rate": 0.00031760882647943837, + "loss": 1.3776, + "theoretical_loss": 3.4583509198480744, + "tokens_seen": 2003566592 + }, + { + "epoch": 0.61, + "learning_rate": 0.00031759277833500507, + "loss": 1.3118, + "theoretical_loss": 3.458341569134956, + "tokens_seen": 2003632128 + }, + { + "epoch": 0.61, + "learning_rate": 0.00031757673019057177, + "loss": 1.3777, + "theoretical_loss": 3.4583322188133154, + "tokens_seen": 2003697664 + }, + { + "epoch": 0.61, + "learning_rate": 0.00031756068204613847, + "loss": 1.3489, + "theoretical_loss": 3.458322868883124, + "tokens_seen": 2003763200 + }, + { + "epoch": 0.61, + "learning_rate": 0.00031754463390170517, + "loss": 1.38, + "theoretical_loss": 3.458313519344352, + "tokens_seen": 2003828736 + }, + { + "epoch": 0.61, + "learning_rate": 0.00031752858575727187, + "loss": 1.3787, + "theoretical_loss": 3.4583041701969712, + "tokens_seen": 2003894272 + }, + { + "epoch": 0.61, + "learning_rate": 0.00031751253761283857, + "loss": 1.3355, + "theoretical_loss": 3.4582948214409512, + "tokens_seen": 2003959808 + }, + { + "epoch": 0.61, + "learning_rate": 0.00031749648946840527, + "loss": 1.386, + "theoretical_loss": 3.4582854730762635, + "tokens_seen": 2004025344 + }, + { + "epoch": 0.61, + "learning_rate": 0.0003174804413239719, + "loss": 1.3776, + "theoretical_loss": 3.4582761251028793, + "tokens_seen": 2004090880 + }, + { + "epoch": 0.61, + "learning_rate": 0.0003174643931795386, + "loss": 1.4373, + "theoretical_loss": 3.4582667775207687, + "tokens_seen": 2004156416 + }, + { + "epoch": 0.61, + "learning_rate": 0.0003174483450351053, + "loss": 1.3434, + "theoretical_loss": 3.458257430329903, + "tokens_seen": 2004221952 + }, + { + "epoch": 0.61, + "learning_rate": 0.000317432296890672, + "loss": 1.4144, + "theoretical_loss": 3.4582480835302527, + "tokens_seen": 2004287488 + }, + { + "epoch": 0.61, + "learning_rate": 0.0003174162487462387, + "loss": 1.2638, + "theoretical_loss": 3.458238737121789, + "tokens_seen": 2004353024 + }, + { + "epoch": 0.61, + "learning_rate": 0.0003174002006018054, + "loss": 1.3739, + "theoretical_loss": 3.4582293911044824, + "tokens_seen": 2004418560 + }, + { + "epoch": 0.61, + "learning_rate": 0.0003173841524573721, + "loss": 1.3303, + "theoretical_loss": 3.4582200454783036, + "tokens_seen": 2004484096 + }, + { + "epoch": 0.61, + "learning_rate": 0.0003173681043129388, + "loss": 1.3139, + "theoretical_loss": 3.4582107002432236, + "tokens_seen": 2004549632 + }, + { + "epoch": 0.61, + "learning_rate": 0.0003173520561685055, + "loss": 1.2695, + "theoretical_loss": 3.458201355399214, + "tokens_seen": 2004615168 + }, + { + "epoch": 0.61, + "learning_rate": 0.0003173360080240722, + "loss": 1.3889, + "theoretical_loss": 3.4581920109462443, + "tokens_seen": 2004680704 + }, + { + "epoch": 0.61, + "learning_rate": 0.0003173199598796389, + "loss": 1.35, + "theoretical_loss": 3.4581826668842863, + "tokens_seen": 2004746240 + }, + { + "epoch": 0.61, + "learning_rate": 0.0003173039117352056, + "loss": 1.4303, + "theoretical_loss": 3.4581733232133107, + "tokens_seen": 2004811776 + }, + { + "epoch": 0.61, + "learning_rate": 0.0003172878635907723, + "loss": 1.277, + "theoretical_loss": 3.4581639799332886, + "tokens_seen": 2004877312 + }, + { + "epoch": 0.61, + "learning_rate": 0.000317271815446339, + "loss": 1.2467, + "theoretical_loss": 3.4581546370441902, + "tokens_seen": 2004942848 + }, + { + "epoch": 0.61, + "learning_rate": 0.0003172557673019057, + "loss": 1.2803, + "theoretical_loss": 3.4581452945459867, + "tokens_seen": 2005008384 + }, + { + "epoch": 0.61, + "learning_rate": 0.0003172397191574724, + "loss": 1.2575, + "theoretical_loss": 3.4581359524386492, + "tokens_seen": 2005073920 + }, + { + "epoch": 0.61, + "learning_rate": 0.00031722367101303913, + "loss": 1.3814, + "theoretical_loss": 3.4581266107221484, + "tokens_seen": 2005139456 + }, + { + "epoch": 0.61, + "learning_rate": 0.00031720762286860583, + "loss": 1.2818, + "theoretical_loss": 3.458117269396455, + "tokens_seen": 2005204992 + }, + { + "epoch": 0.61, + "learning_rate": 0.00031719157472417253, + "loss": 1.3629, + "theoretical_loss": 3.4581079284615397, + "tokens_seen": 2005270528 + }, + { + "epoch": 0.61, + "learning_rate": 0.00031717552657973923, + "loss": 1.3229, + "theoretical_loss": 3.458098587917374, + "tokens_seen": 2005336064 + }, + { + "epoch": 0.61, + "objective/train/docs_used": 731237, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.330771803855896, + "objective/train/theoretical_loss": 3.4580892477639282, + "objective/train/tokens_used": 2025861600, + "theoretical_loss": 3.4580892477639282, + "tokens_seen": 2005401600 + }, + { + "epoch": 0.61, + "learning_rate": 0.00031715947843530593, + "loss": 1.3507, + "theoretical_loss": 3.4580892477639282, + "tokens_seen": 2005401600 + }, + { + "epoch": 0.61, + "learning_rate": 0.00031714343029087263, + "loss": 1.3794, + "theoretical_loss": 3.4580799080011735, + "tokens_seen": 2005467136 + }, + { + "epoch": 0.61, + "learning_rate": 0.00031712738214643933, + "loss": 1.3705, + "theoretical_loss": 3.458070568629081, + "tokens_seen": 2005532672 + }, + { + "epoch": 0.61, + "learning_rate": 0.00031711133400200603, + "loss": 1.2326, + "theoretical_loss": 3.4580612296476207, + "tokens_seen": 2005598208 + }, + { + "epoch": 0.61, + "learning_rate": 0.00031709528585757273, + "loss": 1.3457, + "theoretical_loss": 3.458051891056765, + "tokens_seen": 2005663744 + }, + { + "epoch": 0.61, + "learning_rate": 0.00031707923771313943, + "loss": 1.3361, + "theoretical_loss": 3.458042552856484, + "tokens_seen": 2005729280 + }, + { + "epoch": 0.61, + "learning_rate": 0.00031706318956870613, + "loss": 1.3822, + "theoretical_loss": 3.458033215046748, + "tokens_seen": 2005794816 + }, + { + "epoch": 0.61, + "learning_rate": 0.00031704714142427283, + "loss": 1.3561, + "theoretical_loss": 3.4580238776275283, + "tokens_seen": 2005860352 + }, + { + "epoch": 0.61, + "learning_rate": 0.00031703109327983954, + "loss": 1.2983, + "theoretical_loss": 3.4580145405987963, + "tokens_seen": 2005925888 + }, + { + "epoch": 0.61, + "learning_rate": 0.00031701504513540624, + "loss": 1.3605, + "theoretical_loss": 3.4580052039605222, + "tokens_seen": 2005991424 + }, + { + "epoch": 0.61, + "learning_rate": 0.00031699899699097294, + "loss": 1.2896, + "theoretical_loss": 3.457995867712678, + "tokens_seen": 2006056960 + }, + { + "epoch": 0.61, + "learning_rate": 0.00031698294884653964, + "loss": 1.3262, + "theoretical_loss": 3.4579865318552336, + "tokens_seen": 2006122496 + }, + { + "epoch": 0.61, + "learning_rate": 0.00031696690070210634, + "loss": 1.3121, + "theoretical_loss": 3.45797719638816, + "tokens_seen": 2006188032 + }, + { + "epoch": 0.61, + "learning_rate": 0.00031695085255767304, + "loss": 1.2533, + "theoretical_loss": 3.4579678613114284, + "tokens_seen": 2006253568 + }, + { + "epoch": 0.61, + "learning_rate": 0.00031693480441323974, + "loss": 1.291, + "theoretical_loss": 3.45795852662501, + "tokens_seen": 2006319104 + }, + { + "epoch": 0.61, + "learning_rate": 0.00031691875626880644, + "loss": 1.3407, + "theoretical_loss": 3.457949192328875, + "tokens_seen": 2006384640 + }, + { + "epoch": 0.61, + "learning_rate": 0.00031690270812437314, + "loss": 1.2655, + "theoretical_loss": 3.4579398584229954, + "tokens_seen": 2006450176 + }, + { + "epoch": 0.61, + "learning_rate": 0.00031688665997993984, + "loss": 1.4005, + "theoretical_loss": 3.457930524907341, + "tokens_seen": 2006515712 + }, + { + "epoch": 0.61, + "learning_rate": 0.00031687061183550654, + "loss": 1.4501, + "theoretical_loss": 3.457921191781883, + "tokens_seen": 2006581248 + }, + { + "epoch": 0.61, + "learning_rate": 0.00031685456369107324, + "loss": 1.2464, + "theoretical_loss": 3.457911859046593, + "tokens_seen": 2006646784 + }, + { + "epoch": 0.61, + "learning_rate": 0.00031683851554663994, + "loss": 1.294, + "theoretical_loss": 3.4579025267014414, + "tokens_seen": 2006712320 + }, + { + "epoch": 0.61, + "learning_rate": 0.00031682246740220664, + "loss": 1.3178, + "theoretical_loss": 3.4578931947464, + "tokens_seen": 2006777856 + }, + { + "epoch": 0.61, + "learning_rate": 0.00031680641925777334, + "loss": 1.4713, + "theoretical_loss": 3.4578838631814377, + "tokens_seen": 2006843392 + }, + { + "epoch": 0.61, + "learning_rate": 0.00031679037111334004, + "loss": 1.4395, + "theoretical_loss": 3.457874532006527, + "tokens_seen": 2006908928 + }, + { + "epoch": 0.61, + "learning_rate": 0.00031677432296890675, + "loss": 1.3059, + "theoretical_loss": 3.4578652012216393, + "tokens_seen": 2006974464 + }, + { + "epoch": 0.61, + "learning_rate": 0.00031675827482447345, + "loss": 1.4118, + "theoretical_loss": 3.457855870826745, + "tokens_seen": 2007040000 + }, + { + "epoch": 0.61, + "learning_rate": 0.00031674222668004015, + "loss": 1.3029, + "theoretical_loss": 3.4578465408218144, + "tokens_seen": 2007105536 + }, + { + "epoch": 0.61, + "learning_rate": 0.00031672617853560685, + "loss": 1.3738, + "theoretical_loss": 3.4578372112068196, + "tokens_seen": 2007171072 + }, + { + "epoch": 0.61, + "learning_rate": 0.0003167101303911735, + "loss": 1.4323, + "theoretical_loss": 3.4578278819817303, + "tokens_seen": 2007236608 + }, + { + "epoch": 0.61, + "learning_rate": 0.0003166940822467402, + "loss": 1.3557, + "theoretical_loss": 3.4578185531465184, + "tokens_seen": 2007302144 + }, + { + "epoch": 0.61, + "learning_rate": 0.0003166780341023069, + "loss": 1.4178, + "theoretical_loss": 3.457809224701155, + "tokens_seen": 2007367680 + }, + { + "epoch": 0.61, + "learning_rate": 0.0003166619859578736, + "loss": 1.3205, + "theoretical_loss": 3.4577998966456107, + "tokens_seen": 2007433216 + }, + { + "epoch": 0.61, + "learning_rate": 0.0003166459378134403, + "loss": 1.4444, + "theoretical_loss": 3.4577905689798563, + "tokens_seen": 2007498752 + }, + { + "epoch": 0.61, + "learning_rate": 0.000316629889669007, + "loss": 1.2763, + "theoretical_loss": 3.4577812417038634, + "tokens_seen": 2007564288 + }, + { + "epoch": 0.61, + "learning_rate": 0.0003166138415245737, + "loss": 1.4567, + "theoretical_loss": 3.4577719148176023, + "tokens_seen": 2007629824 + }, + { + "epoch": 0.61, + "learning_rate": 0.0003165977933801404, + "loss": 1.3034, + "theoretical_loss": 3.457762588321044, + "tokens_seen": 2007695360 + }, + { + "epoch": 0.61, + "learning_rate": 0.0003165817452357071, + "loss": 1.385, + "theoretical_loss": 3.4577532622141605, + "tokens_seen": 2007760896 + }, + { + "epoch": 0.61, + "learning_rate": 0.0003165656970912738, + "loss": 1.2535, + "theoretical_loss": 3.4577439364969216, + "tokens_seen": 2007826432 + }, + { + "epoch": 0.61, + "learning_rate": 0.0003165496489468405, + "loss": 1.3144, + "theoretical_loss": 3.4577346111692995, + "tokens_seen": 2007891968 + }, + { + "epoch": 0.61, + "learning_rate": 0.0003165336008024072, + "loss": 1.3404, + "theoretical_loss": 3.4577252862312635, + "tokens_seen": 2007957504 + }, + { + "epoch": 0.61, + "learning_rate": 0.0003165175526579739, + "loss": 1.2365, + "theoretical_loss": 3.457715961682786, + "tokens_seen": 2008023040 + }, + { + "epoch": 0.61, + "learning_rate": 0.00031650150451354066, + "loss": 1.2838, + "theoretical_loss": 3.457706637523838, + "tokens_seen": 2008088576 + }, + { + "epoch": 0.61, + "learning_rate": 0.00031648545636910736, + "loss": 1.3221, + "theoretical_loss": 3.45769731375439, + "tokens_seen": 2008154112 + }, + { + "epoch": 0.61, + "learning_rate": 0.00031646940822467406, + "loss": 1.3446, + "theoretical_loss": 3.457687990374413, + "tokens_seen": 2008219648 + }, + { + "epoch": 0.61, + "learning_rate": 0.00031645336008024076, + "loss": 1.3128, + "theoretical_loss": 3.4576786673838784, + "tokens_seen": 2008285184 + }, + { + "epoch": 0.61, + "learning_rate": 0.00031643731193580746, + "loss": 1.3442, + "theoretical_loss": 3.457669344782757, + "tokens_seen": 2008350720 + }, + { + "epoch": 0.61, + "learning_rate": 0.00031642126379137416, + "loss": 1.3831, + "theoretical_loss": 3.4576600225710195, + "tokens_seen": 2008416256 + }, + { + "epoch": 0.61, + "learning_rate": 0.00031640521564694086, + "loss": 1.2667, + "theoretical_loss": 3.457650700748638, + "tokens_seen": 2008481792 + }, + { + "epoch": 0.61, + "learning_rate": 0.00031638916750250756, + "loss": 1.3691, + "theoretical_loss": 3.4576413793155822, + "tokens_seen": 2008547328 + }, + { + "epoch": 0.61, + "learning_rate": 0.00031637311935807426, + "loss": 1.3368, + "theoretical_loss": 3.457632058271824, + "tokens_seen": 2008612864 + }, + { + "epoch": 0.61, + "objective/train/docs_used": 732466, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.3434908390045166, + "objective/train/theoretical_loss": 3.457622737617334, + "objective/train/tokens_used": 2029138400, + "theoretical_loss": 3.457622737617334, + "tokens_seen": 2008678400 + }, + { + "epoch": 0.61, + "learning_rate": 0.00031635707121364096, + "loss": 1.3314, + "theoretical_loss": 3.457622737617334, + "tokens_seen": 2008678400 + }, + { + "epoch": 0.61, + "learning_rate": 0.00031634102306920766, + "loss": 1.4244, + "theoretical_loss": 3.4576134173520834, + "tokens_seen": 2008743936 + }, + { + "epoch": 0.61, + "learning_rate": 0.00031632497492477436, + "loss": 1.4256, + "theoretical_loss": 3.457604097476043, + "tokens_seen": 2008809472 + }, + { + "epoch": 0.61, + "learning_rate": 0.00031630892678034106, + "loss": 1.2959, + "theoretical_loss": 3.457594777989185, + "tokens_seen": 2008875008 + }, + { + "epoch": 0.61, + "learning_rate": 0.00031629287863590776, + "loss": 1.3507, + "theoretical_loss": 3.457585458891479, + "tokens_seen": 2008940544 + }, + { + "epoch": 0.61, + "learning_rate": 0.00031627683049147446, + "loss": 1.403, + "theoretical_loss": 3.457576140182897, + "tokens_seen": 2009006080 + }, + { + "epoch": 0.61, + "learning_rate": 0.00031626078234704117, + "loss": 1.3273, + "theoretical_loss": 3.4575668218634092, + "tokens_seen": 2009071616 + }, + { + "epoch": 0.61, + "learning_rate": 0.00031624473420260787, + "loss": 1.3662, + "theoretical_loss": 3.4575575039329873, + "tokens_seen": 2009137152 + }, + { + "epoch": 0.61, + "learning_rate": 0.00031622868605817457, + "loss": 1.3196, + "theoretical_loss": 3.457548186391602, + "tokens_seen": 2009202688 + }, + { + "epoch": 0.61, + "learning_rate": 0.00031621263791374127, + "loss": 1.3423, + "theoretical_loss": 3.457538869239225, + "tokens_seen": 2009268224 + }, + { + "epoch": 0.61, + "learning_rate": 0.00031619658976930797, + "loss": 1.3943, + "theoretical_loss": 3.457529552475827, + "tokens_seen": 2009333760 + }, + { + "epoch": 0.61, + "learning_rate": 0.00031618054162487467, + "loss": 1.3238, + "theoretical_loss": 3.457520236101379, + "tokens_seen": 2009399296 + }, + { + "epoch": 0.61, + "learning_rate": 0.00031616449348044137, + "loss": 1.3508, + "theoretical_loss": 3.457510920115852, + "tokens_seen": 2009464832 + }, + { + "epoch": 0.61, + "learning_rate": 0.00031614844533600807, + "loss": 1.4439, + "theoretical_loss": 3.4575016045192175, + "tokens_seen": 2009530368 + }, + { + "epoch": 0.61, + "learning_rate": 0.00031613239719157477, + "loss": 1.277, + "theoretical_loss": 3.457492289311446, + "tokens_seen": 2009595904 + }, + { + "epoch": 0.61, + "learning_rate": 0.00031611634904714147, + "loss": 1.2866, + "theoretical_loss": 3.4574829744925086, + "tokens_seen": 2009661440 + }, + { + "epoch": 0.61, + "learning_rate": 0.00031610030090270817, + "loss": 1.4026, + "theoretical_loss": 3.4574736600623774, + "tokens_seen": 2009726976 + }, + { + "epoch": 0.61, + "learning_rate": 0.00031608425275827487, + "loss": 1.2882, + "theoretical_loss": 3.457464346021022, + "tokens_seen": 2009792512 + }, + { + "epoch": 0.61, + "learning_rate": 0.0003160682046138415, + "loss": 1.2277, + "theoretical_loss": 3.4574550323684146, + "tokens_seen": 2009858048 + }, + { + "epoch": 0.61, + "learning_rate": 0.0003160521564694082, + "loss": 1.3026, + "theoretical_loss": 3.457445719104526, + "tokens_seen": 2009923584 + }, + { + "epoch": 0.61, + "learning_rate": 0.0003160361083249749, + "loss": 1.3889, + "theoretical_loss": 3.457436406229327, + "tokens_seen": 2009989120 + }, + { + "epoch": 0.61, + "learning_rate": 0.0003160200601805416, + "loss": 1.2679, + "theoretical_loss": 3.457427093742789, + "tokens_seen": 2010054656 + }, + { + "epoch": 0.61, + "learning_rate": 0.0003160040120361083, + "loss": 1.3347, + "theoretical_loss": 3.4574177816448834, + "tokens_seen": 2010120192 + }, + { + "epoch": 0.61, + "learning_rate": 0.000315987963891675, + "loss": 1.3168, + "theoretical_loss": 3.4574084699355807, + "tokens_seen": 2010185728 + }, + { + "epoch": 0.61, + "learning_rate": 0.0003159719157472417, + "loss": 1.3264, + "theoretical_loss": 3.4573991586148525, + "tokens_seen": 2010251264 + }, + { + "epoch": 0.61, + "learning_rate": 0.0003159558676028084, + "loss": 1.4303, + "theoretical_loss": 3.4573898476826694, + "tokens_seen": 2010316800 + }, + { + "epoch": 0.61, + "learning_rate": 0.0003159398194583751, + "loss": 1.3332, + "theoretical_loss": 3.457380537139003, + "tokens_seen": 2010382336 + }, + { + "epoch": 0.61, + "learning_rate": 0.0003159237713139418, + "loss": 1.2753, + "theoretical_loss": 3.457371226983824, + "tokens_seen": 2010447872 + }, + { + "epoch": 0.61, + "learning_rate": 0.0003159077231695085, + "loss": 1.3581, + "theoretical_loss": 3.457361917217104, + "tokens_seen": 2010513408 + }, + { + "epoch": 0.61, + "learning_rate": 0.0003158916750250752, + "loss": 1.2543, + "theoretical_loss": 3.457352607838814, + "tokens_seen": 2010578944 + }, + { + "epoch": 0.61, + "learning_rate": 0.0003158756268806419, + "loss": 1.4227, + "theoretical_loss": 3.457343298848925, + "tokens_seen": 2010644480 + }, + { + "epoch": 0.61, + "learning_rate": 0.00031585957873620863, + "loss": 1.2607, + "theoretical_loss": 3.4573339902474083, + "tokens_seen": 2010710016 + }, + { + "epoch": 0.61, + "learning_rate": 0.00031584353059177533, + "loss": 1.3825, + "theoretical_loss": 3.4573246820342347, + "tokens_seen": 2010775552 + }, + { + "epoch": 0.61, + "learning_rate": 0.00031582748244734203, + "loss": 1.2344, + "theoretical_loss": 3.457315374209376, + "tokens_seen": 2010841088 + }, + { + "epoch": 0.61, + "learning_rate": 0.00031581143430290873, + "loss": 1.3794, + "theoretical_loss": 3.4573060667728024, + "tokens_seen": 2010906624 + }, + { + "epoch": 0.61, + "learning_rate": 0.00031579538615847543, + "loss": 1.2892, + "theoretical_loss": 3.4572967597244855, + "tokens_seen": 2010972160 + }, + { + "epoch": 0.61, + "learning_rate": 0.00031577933801404213, + "loss": 1.3463, + "theoretical_loss": 3.4572874530643967, + "tokens_seen": 2011037696 + }, + { + "epoch": 0.61, + "learning_rate": 0.00031576328986960883, + "loss": 1.4035, + "theoretical_loss": 3.4572781467925067, + "tokens_seen": 2011103232 + }, + { + "epoch": 0.61, + "learning_rate": 0.00031574724172517553, + "loss": 1.3728, + "theoretical_loss": 3.4572688409087875, + "tokens_seen": 2011168768 + }, + { + "epoch": 0.61, + "learning_rate": 0.00031573119358074223, + "loss": 1.3526, + "theoretical_loss": 3.4572595354132094, + "tokens_seen": 2011234304 + }, + { + "epoch": 0.61, + "learning_rate": 0.00031571514543630893, + "loss": 1.3667, + "theoretical_loss": 3.4572502303057435, + "tokens_seen": 2011299840 + }, + { + "epoch": 0.61, + "learning_rate": 0.00031569909729187563, + "loss": 1.251, + "theoretical_loss": 3.457240925586362, + "tokens_seen": 2011365376 + }, + { + "epoch": 0.61, + "learning_rate": 0.00031568304914744233, + "loss": 1.405, + "theoretical_loss": 3.457231621255035, + "tokens_seen": 2011430912 + }, + { + "epoch": 0.61, + "learning_rate": 0.00031566700100300903, + "loss": 1.4096, + "theoretical_loss": 3.457222317311734, + "tokens_seen": 2011496448 + }, + { + "epoch": 0.61, + "learning_rate": 0.00031565095285857574, + "loss": 1.2944, + "theoretical_loss": 3.45721301375643, + "tokens_seen": 2011561984 + }, + { + "epoch": 0.61, + "learning_rate": 0.00031563490471414244, + "loss": 1.3121, + "theoretical_loss": 3.457203710589095, + "tokens_seen": 2011627520 + }, + { + "epoch": 0.61, + "learning_rate": 0.00031561885656970914, + "loss": 1.3074, + "theoretical_loss": 3.457194407809699, + "tokens_seen": 2011693056 + }, + { + "epoch": 0.61, + "learning_rate": 0.00031560280842527584, + "loss": 1.356, + "theoretical_loss": 3.4571851054182146, + "tokens_seen": 2011758592 + }, + { + "epoch": 0.61, + "learning_rate": 0.00031558676028084254, + "loss": 1.342, + "theoretical_loss": 3.4571758034146116, + "tokens_seen": 2011824128 + }, + { + "epoch": 0.61, + "learning_rate": 0.00031557071213640924, + "loss": 1.357, + "theoretical_loss": 3.457166501798862, + "tokens_seen": 2011889664 + }, + { + "epoch": 0.61, + "objective/train/docs_used": 733698, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.4806125164031982, + "objective/train/theoretical_loss": 3.4571572005709363, + "objective/train/tokens_used": 2032415200, + "theoretical_loss": 3.4571572005709363, + "tokens_seen": 2011955200 + }, + { + "epoch": 0.61, + "learning_rate": 0.00031555466399197594, + "loss": 1.444, + "theoretical_loss": 3.4571572005709363, + "tokens_seen": 2011955200 + }, + { + "epoch": 0.61, + "learning_rate": 0.00031553861584754264, + "loss": 1.3471, + "theoretical_loss": 3.4571478997308063, + "tokens_seen": 2012020736 + }, + { + "epoch": 0.61, + "learning_rate": 0.00031552256770310934, + "loss": 1.4166, + "theoretical_loss": 3.4571385992784434, + "tokens_seen": 2012086272 + }, + { + "epoch": 0.61, + "learning_rate": 0.00031550651955867604, + "loss": 1.2324, + "theoretical_loss": 3.4571292992138183, + "tokens_seen": 2012151808 + }, + { + "epoch": 0.61, + "learning_rate": 0.00031549047141424274, + "loss": 1.4207, + "theoretical_loss": 3.457119999536902, + "tokens_seen": 2012217344 + }, + { + "epoch": 0.61, + "learning_rate": 0.00031547442326980944, + "loss": 1.3927, + "theoretical_loss": 3.4571107002476666, + "tokens_seen": 2012282880 + }, + { + "epoch": 0.61, + "learning_rate": 0.00031545837512537614, + "loss": 1.3202, + "theoretical_loss": 3.457101401346083, + "tokens_seen": 2012348416 + }, + { + "epoch": 0.61, + "learning_rate": 0.00031544232698094284, + "loss": 1.398, + "theoretical_loss": 3.4570921028321218, + "tokens_seen": 2012413952 + }, + { + "epoch": 0.61, + "learning_rate": 0.00031542627883650954, + "loss": 1.3941, + "theoretical_loss": 3.4570828047057542, + "tokens_seen": 2012479488 + }, + { + "epoch": 0.61, + "learning_rate": 0.00031541023069207625, + "loss": 1.2986, + "theoretical_loss": 3.4570735069669523, + "tokens_seen": 2012545024 + }, + { + "epoch": 0.61, + "learning_rate": 0.00031539418254764295, + "loss": 1.3401, + "theoretical_loss": 3.457064209615687, + "tokens_seen": 2012610560 + }, + { + "epoch": 0.61, + "learning_rate": 0.00031537813440320965, + "loss": 1.4006, + "theoretical_loss": 3.4570549126519294, + "tokens_seen": 2012676096 + }, + { + "epoch": 0.61, + "learning_rate": 0.00031536208625877635, + "loss": 1.4349, + "theoretical_loss": 3.457045616075651, + "tokens_seen": 2012741632 + }, + { + "epoch": 0.61, + "learning_rate": 0.00031534603811434305, + "loss": 1.2996, + "theoretical_loss": 3.457036319886822, + "tokens_seen": 2012807168 + }, + { + "epoch": 0.61, + "learning_rate": 0.00031532998996990975, + "loss": 1.3955, + "theoretical_loss": 3.4570270240854146, + "tokens_seen": 2012872704 + }, + { + "epoch": 0.61, + "learning_rate": 0.00031531394182547645, + "loss": 1.4624, + "theoretical_loss": 3.4570177286714, + "tokens_seen": 2012938240 + }, + { + "epoch": 0.61, + "learning_rate": 0.00031529789368104315, + "loss": 1.3806, + "theoretical_loss": 3.4570084336447495, + "tokens_seen": 2013003776 + }, + { + "epoch": 0.61, + "learning_rate": 0.00031528184553660985, + "loss": 1.2817, + "theoretical_loss": 3.4569991390054335, + "tokens_seen": 2013069312 + }, + { + "epoch": 0.61, + "learning_rate": 0.00031526579739217655, + "loss": 1.3795, + "theoretical_loss": 3.4569898447534246, + "tokens_seen": 2013134848 + }, + { + "epoch": 0.61, + "learning_rate": 0.00031524974924774325, + "loss": 1.4027, + "theoretical_loss": 3.4569805508886935, + "tokens_seen": 2013200384 + }, + { + "epoch": 0.61, + "learning_rate": 0.00031523370110330995, + "loss": 1.2561, + "theoretical_loss": 3.456971257411211, + "tokens_seen": 2013265920 + }, + { + "epoch": 0.61, + "learning_rate": 0.00031521765295887665, + "loss": 1.3348, + "theoretical_loss": 3.456961964320948, + "tokens_seen": 2013331456 + }, + { + "epoch": 0.61, + "learning_rate": 0.00031520160481444335, + "loss": 1.422, + "theoretical_loss": 3.456952671617877, + "tokens_seen": 2013396992 + }, + { + "epoch": 0.61, + "learning_rate": 0.00031518555667001005, + "loss": 1.3163, + "theoretical_loss": 3.4569433793019693, + "tokens_seen": 2013462528 + }, + { + "epoch": 0.61, + "learning_rate": 0.00031516950852557675, + "loss": 1.3048, + "theoretical_loss": 3.4569340873731944, + "tokens_seen": 2013528064 + }, + { + "epoch": 0.61, + "learning_rate": 0.00031515346038114346, + "loss": 1.3211, + "theoretical_loss": 3.4569247958315255, + "tokens_seen": 2013593600 + }, + { + "epoch": 0.61, + "learning_rate": 0.00031513741223671016, + "loss": 1.3964, + "theoretical_loss": 3.4569155046769326, + "tokens_seen": 2013659136 + }, + { + "epoch": 0.61, + "learning_rate": 0.00031512136409227686, + "loss": 1.2847, + "theoretical_loss": 3.456906213909388, + "tokens_seen": 2013724672 + }, + { + "epoch": 0.61, + "learning_rate": 0.00031510531594784356, + "loss": 1.3604, + "theoretical_loss": 3.4568969235288627, + "tokens_seen": 2013790208 + }, + { + "epoch": 0.61, + "learning_rate": 0.00031508926780341026, + "loss": 1.2704, + "theoretical_loss": 3.456887633535327, + "tokens_seen": 2013855744 + }, + { + "epoch": 0.61, + "learning_rate": 0.00031507321965897696, + "loss": 1.3055, + "theoretical_loss": 3.456878343928753, + "tokens_seen": 2013921280 + }, + { + "epoch": 0.61, + "learning_rate": 0.00031505717151454366, + "loss": 1.3253, + "theoretical_loss": 3.456869054709112, + "tokens_seen": 2013986816 + }, + { + "epoch": 0.61, + "learning_rate": 0.00031504112337011036, + "loss": 1.333, + "theoretical_loss": 3.4568597658763753, + "tokens_seen": 2014052352 + }, + { + "epoch": 0.61, + "learning_rate": 0.00031502507522567706, + "loss": 1.308, + "theoretical_loss": 3.4568504774305144, + "tokens_seen": 2014117888 + }, + { + "epoch": 0.61, + "learning_rate": 0.00031500902708124376, + "loss": 1.3445, + "theoretical_loss": 3.4568411893715, + "tokens_seen": 2014183424 + }, + { + "epoch": 0.61, + "learning_rate": 0.00031499297893681046, + "loss": 1.4328, + "theoretical_loss": 3.456831901699304, + "tokens_seen": 2014248960 + }, + { + "epoch": 0.61, + "learning_rate": 0.00031497693079237716, + "loss": 1.3715, + "theoretical_loss": 3.4568226144138974, + "tokens_seen": 2014314496 + }, + { + "epoch": 0.61, + "learning_rate": 0.00031496088264794386, + "loss": 1.3574, + "theoretical_loss": 3.4568133275152513, + "tokens_seen": 2014380032 + }, + { + "epoch": 0.61, + "learning_rate": 0.00031494483450351056, + "loss": 1.3821, + "theoretical_loss": 3.456804041003337, + "tokens_seen": 2014445568 + }, + { + "epoch": 0.61, + "learning_rate": 0.00031492878635907726, + "loss": 1.24, + "theoretical_loss": 3.4567947548781266, + "tokens_seen": 2014511104 + }, + { + "epoch": 0.61, + "learning_rate": 0.00031491273821464396, + "loss": 1.3498, + "theoretical_loss": 3.4567854691395903, + "tokens_seen": 2014576640 + }, + { + "epoch": 0.61, + "learning_rate": 0.00031489669007021067, + "loss": 1.3428, + "theoretical_loss": 3.4567761837877002, + "tokens_seen": 2014642176 + }, + { + "epoch": 0.61, + "learning_rate": 0.00031488064192577737, + "loss": 1.3949, + "theoretical_loss": 3.456766898822428, + "tokens_seen": 2014707712 + }, + { + "epoch": 0.61, + "learning_rate": 0.00031486459378134407, + "loss": 1.2603, + "theoretical_loss": 3.456757614243744, + "tokens_seen": 2014773248 + }, + { + "epoch": 0.61, + "learning_rate": 0.00031484854563691077, + "loss": 1.3322, + "theoretical_loss": 3.45674833005162, + "tokens_seen": 2014838784 + }, + { + "epoch": 0.61, + "learning_rate": 0.00031483249749247747, + "loss": 1.2384, + "theoretical_loss": 3.4567390462460272, + "tokens_seen": 2014904320 + }, + { + "epoch": 0.61, + "learning_rate": 0.00031481644934804417, + "loss": 1.3439, + "theoretical_loss": 3.456729762826937, + "tokens_seen": 2014969856 + }, + { + "epoch": 0.61, + "learning_rate": 0.00031480040120361087, + "loss": 1.3428, + "theoretical_loss": 3.456720479794321, + "tokens_seen": 2015035392 + }, + { + "epoch": 0.61, + "learning_rate": 0.00031478435305917757, + "loss": 1.3998, + "theoretical_loss": 3.45671119714815, + "tokens_seen": 2015100928 + }, + { + "epoch": 0.61, + "learning_rate": 0.00031476830491474427, + "loss": 1.3401, + "theoretical_loss": 3.456701914888396, + "tokens_seen": 2015166464 + }, + { + "epoch": 0.61, + "objective/train/docs_used": 734924, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.4040330648422241, + "objective/train/theoretical_loss": 3.45669263301503, + "objective/train/tokens_used": 2035692000, + "theoretical_loss": 3.45669263301503, + "tokens_seen": 2015232000 + }, + { + "epoch": 0.61, + "learning_rate": 0.00031475225677031097, + "loss": 1.3675, + "theoretical_loss": 3.45669263301503, + "tokens_seen": 2015232000 + }, + { + "epoch": 0.61, + "learning_rate": 0.00031473620862587767, + "loss": 1.2432, + "theoretical_loss": 3.4566833515280235, + "tokens_seen": 2015297536 + }, + { + "epoch": 0.61, + "learning_rate": 0.00031472016048144437, + "loss": 1.3783, + "theoretical_loss": 3.456674070427347, + "tokens_seen": 2015363072 + }, + { + "epoch": 0.61, + "learning_rate": 0.0003147041123370111, + "loss": 1.3833, + "theoretical_loss": 3.456664789712973, + "tokens_seen": 2015428608 + }, + { + "epoch": 0.61, + "learning_rate": 0.0003146880641925778, + "loss": 1.3323, + "theoretical_loss": 3.4566555093848725, + "tokens_seen": 2015494144 + }, + { + "epoch": 0.61, + "learning_rate": 0.0003146720160481444, + "loss": 1.2511, + "theoretical_loss": 3.456646229443017, + "tokens_seen": 2015559680 + }, + { + "epoch": 0.61, + "learning_rate": 0.0003146559679037111, + "loss": 1.3802, + "theoretical_loss": 3.4566369498873777, + "tokens_seen": 2015625216 + }, + { + "epoch": 0.61, + "learning_rate": 0.0003146399197592778, + "loss": 1.2598, + "theoretical_loss": 3.4566276707179258, + "tokens_seen": 2015690752 + }, + { + "epoch": 0.61, + "learning_rate": 0.0003146238716148445, + "loss": 1.2577, + "theoretical_loss": 3.456618391934633, + "tokens_seen": 2015756288 + }, + { + "epoch": 0.61, + "learning_rate": 0.0003146078234704112, + "loss": 1.3645, + "theoretical_loss": 3.45660911353747, + "tokens_seen": 2015821824 + }, + { + "epoch": 0.61, + "learning_rate": 0.0003145917753259779, + "loss": 1.2245, + "theoretical_loss": 3.4565998355264087, + "tokens_seen": 2015887360 + }, + { + "epoch": 0.61, + "learning_rate": 0.0003145757271815446, + "loss": 1.3801, + "theoretical_loss": 3.456590557901421, + "tokens_seen": 2015952896 + }, + { + "epoch": 0.61, + "learning_rate": 0.0003145596790371113, + "loss": 1.3355, + "theoretical_loss": 3.456581280662477, + "tokens_seen": 2016018432 + }, + { + "epoch": 0.61, + "learning_rate": 0.000314543630892678, + "loss": 1.3127, + "theoretical_loss": 3.4565720038095495, + "tokens_seen": 2016083968 + }, + { + "epoch": 0.61, + "learning_rate": 0.0003145275827482447, + "loss": 1.3989, + "theoretical_loss": 3.456562727342609, + "tokens_seen": 2016149504 + }, + { + "epoch": 0.61, + "learning_rate": 0.0003145115346038114, + "loss": 1.3396, + "theoretical_loss": 3.4565534512616267, + "tokens_seen": 2016215040 + }, + { + "epoch": 0.61, + "learning_rate": 0.0003144954864593781, + "loss": 1.395, + "theoretical_loss": 3.4565441755665747, + "tokens_seen": 2016280576 + }, + { + "epoch": 0.61, + "learning_rate": 0.00031447943831494483, + "loss": 1.2703, + "theoretical_loss": 3.4565349002574246, + "tokens_seen": 2016346112 + }, + { + "epoch": 0.61, + "learning_rate": 0.00031446339017051153, + "loss": 1.2925, + "theoretical_loss": 3.456525625334147, + "tokens_seen": 2016411648 + }, + { + "epoch": 0.61, + "learning_rate": 0.00031444734202607823, + "loss": 1.2969, + "theoretical_loss": 3.456516350796713, + "tokens_seen": 2016477184 + }, + { + "epoch": 0.61, + "learning_rate": 0.00031443129388164493, + "loss": 1.275, + "theoretical_loss": 3.4565070766450954, + "tokens_seen": 2016542720 + }, + { + "epoch": 0.61, + "learning_rate": 0.00031441524573721163, + "loss": 1.2943, + "theoretical_loss": 3.4564978028792646, + "tokens_seen": 2016608256 + }, + { + "epoch": 0.61, + "learning_rate": 0.00031439919759277833, + "loss": 1.2691, + "theoretical_loss": 3.4564885294991923, + "tokens_seen": 2016673792 + }, + { + "epoch": 0.61, + "learning_rate": 0.00031438314944834503, + "loss": 1.3916, + "theoretical_loss": 3.4564792565048497, + "tokens_seen": 2016739328 + }, + { + "epoch": 0.61, + "learning_rate": 0.00031436710130391173, + "loss": 1.3659, + "theoretical_loss": 3.4564699838962083, + "tokens_seen": 2016804864 + }, + { + "epoch": 0.61, + "learning_rate": 0.00031435105315947843, + "loss": 1.3288, + "theoretical_loss": 3.45646071167324, + "tokens_seen": 2016870400 + }, + { + "epoch": 0.61, + "learning_rate": 0.00031433500501504513, + "loss": 1.3053, + "theoretical_loss": 3.4564514398359156, + "tokens_seen": 2016935936 + }, + { + "epoch": 0.61, + "learning_rate": 0.00031431895687061183, + "loss": 1.3535, + "theoretical_loss": 3.456442168384207, + "tokens_seen": 2017001472 + }, + { + "epoch": 0.61, + "learning_rate": 0.00031430290872617853, + "loss": 1.31, + "theoretical_loss": 3.456432897318085, + "tokens_seen": 2017067008 + }, + { + "epoch": 0.61, + "learning_rate": 0.00031428686058174524, + "loss": 1.4448, + "theoretical_loss": 3.456423626637522, + "tokens_seen": 2017132544 + }, + { + "epoch": 0.61, + "learning_rate": 0.00031427081243731194, + "loss": 1.3452, + "theoretical_loss": 3.4564143563424885, + "tokens_seen": 2017198080 + }, + { + "epoch": 0.61, + "learning_rate": 0.00031425476429287864, + "loss": 1.4335, + "theoretical_loss": 3.456405086432956, + "tokens_seen": 2017263616 + }, + { + "epoch": 0.61, + "learning_rate": 0.00031423871614844534, + "loss": 1.4578, + "theoretical_loss": 3.456395816908897, + "tokens_seen": 2017329152 + }, + { + "epoch": 0.61, + "learning_rate": 0.00031422266800401204, + "loss": 1.2949, + "theoretical_loss": 3.4563865477702818, + "tokens_seen": 2017394688 + }, + { + "epoch": 0.61, + "learning_rate": 0.00031420661985957874, + "loss": 1.295, + "theoretical_loss": 3.4563772790170826, + "tokens_seen": 2017460224 + }, + { + "epoch": 0.61, + "learning_rate": 0.00031419057171514544, + "loss": 1.3034, + "theoretical_loss": 3.4563680106492702, + "tokens_seen": 2017525760 + }, + { + "epoch": 0.61, + "learning_rate": 0.00031417452357071214, + "loss": 1.3706, + "theoretical_loss": 3.4563587426668168, + "tokens_seen": 2017591296 + }, + { + "epoch": 0.61, + "learning_rate": 0.00031415847542627884, + "loss": 1.4375, + "theoretical_loss": 3.456349475069693, + "tokens_seen": 2017656832 + }, + { + "epoch": 0.61, + "learning_rate": 0.00031414242728184554, + "loss": 1.2916, + "theoretical_loss": 3.4563402078578704, + "tokens_seen": 2017722368 + }, + { + "epoch": 0.61, + "learning_rate": 0.00031412637913741224, + "loss": 1.2982, + "theoretical_loss": 3.4563309410313217, + "tokens_seen": 2017787904 + }, + { + "epoch": 0.61, + "learning_rate": 0.000314110330992979, + "loss": 1.3456, + "theoretical_loss": 3.4563216745900167, + "tokens_seen": 2017853440 + }, + { + "epoch": 0.61, + "learning_rate": 0.0003140942828485457, + "loss": 1.3592, + "theoretical_loss": 3.456312408533928, + "tokens_seen": 2017918976 + }, + { + "epoch": 0.61, + "learning_rate": 0.0003140782347041124, + "loss": 1.3418, + "theoretical_loss": 3.456303142863027, + "tokens_seen": 2017984512 + }, + { + "epoch": 0.61, + "learning_rate": 0.0003140621865596791, + "loss": 1.2581, + "theoretical_loss": 3.4562938775772842, + "tokens_seen": 2018050048 + }, + { + "epoch": 0.61, + "learning_rate": 0.0003140461384152458, + "loss": 1.2851, + "theoretical_loss": 3.456284612676672, + "tokens_seen": 2018115584 + }, + { + "epoch": 0.61, + "learning_rate": 0.00031403009027081245, + "loss": 1.3552, + "theoretical_loss": 3.456275348161162, + "tokens_seen": 2018181120 + }, + { + "epoch": 0.61, + "learning_rate": 0.00031401404212637915, + "loss": 1.3387, + "theoretical_loss": 3.456266084030725, + "tokens_seen": 2018246656 + }, + { + "epoch": 0.61, + "learning_rate": 0.00031399799398194585, + "loss": 1.2744, + "theoretical_loss": 3.4562568202853328, + "tokens_seen": 2018312192 + }, + { + "epoch": 0.61, + "learning_rate": 0.00031398194583751255, + "loss": 1.3711, + "theoretical_loss": 3.456247556924957, + "tokens_seen": 2018377728 + }, + { + "epoch": 0.61, + "learning_rate": 0.00031396589769307925, + "loss": 1.4117, + "theoretical_loss": 3.456238293949569, + "tokens_seen": 2018443264 + }, + { + "epoch": 0.61, + "objective/train/docs_used": 736110, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.2961543798446655, + "objective/train/theoretical_loss": 3.4562290313591406, + "objective/train/tokens_used": 2038968800, + "theoretical_loss": 3.4562290313591406, + "tokens_seen": 2018508800 + }, + { + "epoch": 0.61, + "learning_rate": 0.00031394984954864595, + "loss": 1.334, + "theoretical_loss": 3.4562290313591406, + "tokens_seen": 2018508800 + }, + { + "epoch": 0.61, + "learning_rate": 0.00031393380140421265, + "loss": 1.37, + "theoretical_loss": 3.4562197691536425, + "tokens_seen": 2018574336 + }, + { + "epoch": 0.61, + "learning_rate": 0.00031391775325977935, + "loss": 1.3736, + "theoretical_loss": 3.456210507333047, + "tokens_seen": 2018639872 + }, + { + "epoch": 0.61, + "learning_rate": 0.00031390170511534605, + "loss": 1.3229, + "theoretical_loss": 3.4562012458973257, + "tokens_seen": 2018705408 + }, + { + "epoch": 0.61, + "learning_rate": 0.00031388565697091275, + "loss": 1.2764, + "theoretical_loss": 3.4561919848464493, + "tokens_seen": 2018770944 + }, + { + "epoch": 0.61, + "learning_rate": 0.00031386960882647945, + "loss": 1.2903, + "theoretical_loss": 3.4561827241803895, + "tokens_seen": 2018836480 + }, + { + "epoch": 0.61, + "learning_rate": 0.00031385356068204615, + "loss": 1.3161, + "theoretical_loss": 3.456173463899119, + "tokens_seen": 2018902016 + }, + { + "epoch": 0.61, + "learning_rate": 0.00031383751253761285, + "loss": 1.3818, + "theoretical_loss": 3.4561642040026075, + "tokens_seen": 2018967552 + }, + { + "epoch": 0.61, + "learning_rate": 0.00031382146439317955, + "loss": 1.3069, + "theoretical_loss": 3.4561549444908275, + "tokens_seen": 2019033088 + }, + { + "epoch": 0.61, + "learning_rate": 0.00031380541624874625, + "loss": 1.2449, + "theoretical_loss": 3.456145685363751, + "tokens_seen": 2019098624 + }, + { + "epoch": 0.61, + "learning_rate": 0.00031378936810431295, + "loss": 1.2158, + "theoretical_loss": 3.456136426621349, + "tokens_seen": 2019164160 + }, + { + "epoch": 0.61, + "learning_rate": 0.00031377331995987966, + "loss": 1.3147, + "theoretical_loss": 3.456127168263593, + "tokens_seen": 2019229696 + }, + { + "epoch": 0.61, + "learning_rate": 0.00031375727181544636, + "loss": 1.3593, + "theoretical_loss": 3.4561179102904536, + "tokens_seen": 2019295232 + }, + { + "epoch": 0.61, + "learning_rate": 0.00031374122367101306, + "loss": 1.2817, + "theoretical_loss": 3.4561086527019045, + "tokens_seen": 2019360768 + }, + { + "epoch": 0.61, + "learning_rate": 0.00031372517552657976, + "loss": 1.3173, + "theoretical_loss": 3.4560993954979153, + "tokens_seen": 2019426304 + }, + { + "epoch": 0.61, + "learning_rate": 0.00031370912738214646, + "loss": 1.3672, + "theoretical_loss": 3.4560901386784586, + "tokens_seen": 2019491840 + }, + { + "epoch": 0.61, + "learning_rate": 0.00031369307923771316, + "loss": 1.332, + "theoretical_loss": 3.456080882243506, + "tokens_seen": 2019557376 + }, + { + "epoch": 0.61, + "learning_rate": 0.00031367703109327986, + "loss": 1.2859, + "theoretical_loss": 3.456071626193028, + "tokens_seen": 2019622912 + }, + { + "epoch": 0.61, + "learning_rate": 0.00031366098294884656, + "loss": 1.4525, + "theoretical_loss": 3.4560623705269973, + "tokens_seen": 2019688448 + }, + { + "epoch": 0.61, + "learning_rate": 0.00031364493480441326, + "loss": 1.4155, + "theoretical_loss": 3.4560531152453846, + "tokens_seen": 2019753984 + }, + { + "epoch": 0.61, + "learning_rate": 0.00031362888665997996, + "loss": 1.4112, + "theoretical_loss": 3.4560438603481622, + "tokens_seen": 2019819520 + }, + { + "epoch": 0.61, + "learning_rate": 0.00031361283851554666, + "loss": 1.2966, + "theoretical_loss": 3.4560346058353018, + "tokens_seen": 2019885056 + }, + { + "epoch": 0.61, + "learning_rate": 0.00031359679037111336, + "loss": 1.4864, + "theoretical_loss": 3.4560253517067734, + "tokens_seen": 2019950592 + }, + { + "epoch": 0.61, + "learning_rate": 0.00031358074222668006, + "loss": 1.3401, + "theoretical_loss": 3.4560160979625505, + "tokens_seen": 2020016128 + }, + { + "epoch": 0.61, + "learning_rate": 0.00031356469408224676, + "loss": 1.4207, + "theoretical_loss": 3.4560068446026033, + "tokens_seen": 2020081664 + }, + { + "epoch": 0.61, + "learning_rate": 0.00031354864593781346, + "loss": 1.2824, + "theoretical_loss": 3.455997591626905, + "tokens_seen": 2020147200 + }, + { + "epoch": 0.61, + "learning_rate": 0.00031353259779338017, + "loss": 1.3109, + "theoretical_loss": 3.455988339035425, + "tokens_seen": 2020212736 + }, + { + "epoch": 0.61, + "learning_rate": 0.00031351654964894687, + "loss": 1.3334, + "theoretical_loss": 3.455979086828136, + "tokens_seen": 2020278272 + }, + { + "epoch": 0.61, + "learning_rate": 0.00031350050150451357, + "loss": 1.3066, + "theoretical_loss": 3.45596983500501, + "tokens_seen": 2020343808 + }, + { + "epoch": 0.61, + "learning_rate": 0.00031348445336008027, + "loss": 1.2887, + "theoretical_loss": 3.455960583566018, + "tokens_seen": 2020409344 + }, + { + "epoch": 0.61, + "learning_rate": 0.00031346840521564697, + "loss": 1.3767, + "theoretical_loss": 3.4559513325111317, + "tokens_seen": 2020474880 + }, + { + "epoch": 0.61, + "learning_rate": 0.00031345235707121367, + "loss": 1.3548, + "theoretical_loss": 3.4559420818403233, + "tokens_seen": 2020540416 + }, + { + "epoch": 0.61, + "learning_rate": 0.00031343630892678037, + "loss": 1.2932, + "theoretical_loss": 3.455932831553563, + "tokens_seen": 2020605952 + }, + { + "epoch": 0.61, + "learning_rate": 0.00031342026078234707, + "loss": 1.3468, + "theoretical_loss": 3.455923581650824, + "tokens_seen": 2020671488 + }, + { + "epoch": 0.61, + "learning_rate": 0.00031340421263791377, + "loss": 1.3017, + "theoretical_loss": 3.4559143321320764, + "tokens_seen": 2020737024 + }, + { + "epoch": 0.61, + "learning_rate": 0.00031338816449348047, + "loss": 1.3138, + "theoretical_loss": 3.455905082997293, + "tokens_seen": 2020802560 + }, + { + "epoch": 0.61, + "learning_rate": 0.00031337211634904717, + "loss": 1.3584, + "theoretical_loss": 3.4558958342464448, + "tokens_seen": 2020868096 + }, + { + "epoch": 0.61, + "learning_rate": 0.00031335606820461387, + "loss": 1.4462, + "theoretical_loss": 3.4558865858795036, + "tokens_seen": 2020933632 + }, + { + "epoch": 0.61, + "learning_rate": 0.00031334002006018057, + "loss": 1.4346, + "theoretical_loss": 3.4558773378964407, + "tokens_seen": 2020999168 + }, + { + "epoch": 0.61, + "learning_rate": 0.0003133239719157473, + "loss": 1.3373, + "theoretical_loss": 3.4558680902972285, + "tokens_seen": 2021064704 + }, + { + "epoch": 0.61, + "learning_rate": 0.000313307923771314, + "loss": 1.2792, + "theoretical_loss": 3.455858843081838, + "tokens_seen": 2021130240 + }, + { + "epoch": 0.61, + "learning_rate": 0.0003132918756268807, + "loss": 1.2682, + "theoretical_loss": 3.455849596250241, + "tokens_seen": 2021195776 + }, + { + "epoch": 0.61, + "learning_rate": 0.0003132758274824474, + "loss": 1.3294, + "theoretical_loss": 3.455840349802409, + "tokens_seen": 2021261312 + }, + { + "epoch": 0.61, + "learning_rate": 0.000313259779338014, + "loss": 1.3393, + "theoretical_loss": 3.455831103738314, + "tokens_seen": 2021326848 + }, + { + "epoch": 0.61, + "learning_rate": 0.0003132437311935807, + "loss": 1.2609, + "theoretical_loss": 3.4558218580579263, + "tokens_seen": 2021392384 + }, + { + "epoch": 0.61, + "learning_rate": 0.0003132276830491474, + "loss": 1.2913, + "theoretical_loss": 3.4558126127612194, + "tokens_seen": 2021457920 + }, + { + "epoch": 0.61, + "learning_rate": 0.0003132116349047141, + "loss": 1.4241, + "theoretical_loss": 3.4558033678481643, + "tokens_seen": 2021523456 + }, + { + "epoch": 0.61, + "learning_rate": 0.0003131955867602808, + "loss": 1.3863, + "theoretical_loss": 3.455794123318732, + "tokens_seen": 2021588992 + }, + { + "epoch": 0.61, + "learning_rate": 0.0003131795386158475, + "loss": 1.4119, + "theoretical_loss": 3.455784879172895, + "tokens_seen": 2021654528 + }, + { + "epoch": 0.61, + "learning_rate": 0.0003131634904714142, + "loss": 1.3181, + "theoretical_loss": 3.4557756354106246, + "tokens_seen": 2021720064 + }, + { + "epoch": 0.61, + "objective/train/docs_used": 737310, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.3218255043029785, + "objective/train/theoretical_loss": 3.455766392031892, + "objective/train/tokens_used": 2042245600, + "theoretical_loss": 3.455766392031892, + "tokens_seen": 2021785600 + }, + { + "epoch": 0.61, + "learning_rate": 0.0003131474423269809, + "loss": 1.3584, + "theoretical_loss": 3.455766392031892, + "tokens_seen": 2021785600 + }, + { + "epoch": 0.61, + "learning_rate": 0.0003131313941825476, + "loss": 1.2926, + "theoretical_loss": 3.4557571490366694, + "tokens_seen": 2021851136 + }, + { + "epoch": 0.61, + "learning_rate": 0.00031311534603811433, + "loss": 1.1553, + "theoretical_loss": 3.4557479064249286, + "tokens_seen": 2021916672 + }, + { + "epoch": 0.61, + "learning_rate": 0.00031309929789368103, + "loss": 1.2153, + "theoretical_loss": 3.45573866419664, + "tokens_seen": 2021982208 + }, + { + "epoch": 0.61, + "learning_rate": 0.00031308324974924773, + "loss": 1.3387, + "theoretical_loss": 3.4557294223517774, + "tokens_seen": 2022047744 + }, + { + "epoch": 0.61, + "learning_rate": 0.00031306720160481443, + "loss": 1.4859, + "theoretical_loss": 3.455720180890311, + "tokens_seen": 2022113280 + }, + { + "epoch": 0.61, + "learning_rate": 0.00031305115346038113, + "loss": 1.4141, + "theoretical_loss": 3.455710939812213, + "tokens_seen": 2022178816 + }, + { + "epoch": 0.61, + "learning_rate": 0.00031303510531594783, + "loss": 1.2941, + "theoretical_loss": 3.455701699117454, + "tokens_seen": 2022244352 + }, + { + "epoch": 0.61, + "learning_rate": 0.00031301905717151453, + "loss": 1.3952, + "theoretical_loss": 3.4556924588060074, + "tokens_seen": 2022309888 + }, + { + "epoch": 0.61, + "learning_rate": 0.00031300300902708123, + "loss": 1.2864, + "theoretical_loss": 3.4556832188778435, + "tokens_seen": 2022375424 + }, + { + "epoch": 0.61, + "learning_rate": 0.00031298696088264793, + "loss": 1.3862, + "theoretical_loss": 3.4556739793329347, + "tokens_seen": 2022440960 + }, + { + "epoch": 0.61, + "learning_rate": 0.00031297091273821463, + "loss": 1.2751, + "theoretical_loss": 3.455664740171253, + "tokens_seen": 2022506496 + }, + { + "epoch": 0.61, + "learning_rate": 0.00031295486459378133, + "loss": 1.3713, + "theoretical_loss": 3.4556555013927683, + "tokens_seen": 2022572032 + }, + { + "epoch": 0.61, + "learning_rate": 0.00031293881644934803, + "loss": 1.2286, + "theoretical_loss": 3.4556462629974547, + "tokens_seen": 2022637568 + }, + { + "epoch": 0.61, + "learning_rate": 0.00031292276830491473, + "loss": 1.4138, + "theoretical_loss": 3.455637024985282, + "tokens_seen": 2022703104 + }, + { + "epoch": 0.61, + "learning_rate": 0.0003129067201604815, + "loss": 1.2298, + "theoretical_loss": 3.455627787356223, + "tokens_seen": 2022768640 + }, + { + "epoch": 0.61, + "learning_rate": 0.0003128906720160482, + "loss": 1.333, + "theoretical_loss": 3.455618550110249, + "tokens_seen": 2022834176 + }, + { + "epoch": 0.61, + "learning_rate": 0.0003128746238716149, + "loss": 1.3199, + "theoretical_loss": 3.455609313247332, + "tokens_seen": 2022899712 + }, + { + "epoch": 0.61, + "learning_rate": 0.0003128585757271816, + "loss": 1.3463, + "theoretical_loss": 3.4556000767674435, + "tokens_seen": 2022965248 + }, + { + "epoch": 0.61, + "learning_rate": 0.0003128425275827483, + "loss": 1.4527, + "theoretical_loss": 3.4555908406705544, + "tokens_seen": 2023030784 + }, + { + "epoch": 0.61, + "learning_rate": 0.000312826479438315, + "loss": 1.2345, + "theoretical_loss": 3.455581604956638, + "tokens_seen": 2023096320 + }, + { + "epoch": 0.61, + "learning_rate": 0.0003128104312938817, + "loss": 1.2787, + "theoretical_loss": 3.455572369625665, + "tokens_seen": 2023161856 + }, + { + "epoch": 0.61, + "learning_rate": 0.0003127943831494484, + "loss": 1.4041, + "theoretical_loss": 3.455563134677607, + "tokens_seen": 2023227392 + }, + { + "epoch": 0.61, + "learning_rate": 0.0003127783350050151, + "loss": 1.2991, + "theoretical_loss": 3.4555539001124362, + "tokens_seen": 2023292928 + }, + { + "epoch": 0.61, + "learning_rate": 0.0003127622868605818, + "loss": 1.315, + "theoretical_loss": 3.455544665930124, + "tokens_seen": 2023358464 + }, + { + "epoch": 0.61, + "learning_rate": 0.0003127462387161485, + "loss": 1.4057, + "theoretical_loss": 3.4555354321306426, + "tokens_seen": 2023424000 + }, + { + "epoch": 0.61, + "learning_rate": 0.0003127301905717152, + "loss": 1.3194, + "theoretical_loss": 3.4555261987139634, + "tokens_seen": 2023489536 + }, + { + "epoch": 0.61, + "learning_rate": 0.0003127141424272819, + "loss": 1.3715, + "theoretical_loss": 3.455516965680058, + "tokens_seen": 2023555072 + }, + { + "epoch": 0.61, + "learning_rate": 0.0003126980942828486, + "loss": 1.367, + "theoretical_loss": 3.4555077330288984, + "tokens_seen": 2023620608 + }, + { + "epoch": 0.61, + "learning_rate": 0.0003126820461384153, + "loss": 1.3853, + "theoretical_loss": 3.455498500760456, + "tokens_seen": 2023686144 + }, + { + "epoch": 0.61, + "learning_rate": 0.000312665997993982, + "loss": 1.2961, + "theoretical_loss": 3.455489268874703, + "tokens_seen": 2023751680 + }, + { + "epoch": 0.61, + "learning_rate": 0.0003126499498495487, + "loss": 1.2963, + "theoretical_loss": 3.455480037371611, + "tokens_seen": 2023817216 + }, + { + "epoch": 0.61, + "learning_rate": 0.00031263390170511535, + "loss": 1.3858, + "theoretical_loss": 3.455470806251151, + "tokens_seen": 2023882752 + }, + { + "epoch": 0.61, + "learning_rate": 0.00031261785356068205, + "loss": 1.3944, + "theoretical_loss": 3.4554615755132962, + "tokens_seen": 2023948288 + }, + { + "epoch": 0.61, + "learning_rate": 0.00031260180541624875, + "loss": 1.3341, + "theoretical_loss": 3.455452345158017, + "tokens_seen": 2024013824 + }, + { + "epoch": 0.61, + "learning_rate": 0.00031258575727181545, + "loss": 1.3395, + "theoretical_loss": 3.455443115185286, + "tokens_seen": 2024079360 + }, + { + "epoch": 0.61, + "learning_rate": 0.00031256970912738215, + "loss": 1.3995, + "theoretical_loss": 3.455433885595075, + "tokens_seen": 2024144896 + }, + { + "epoch": 0.61, + "learning_rate": 0.00031255366098294885, + "loss": 1.4612, + "theoretical_loss": 3.455424656387355, + "tokens_seen": 2024210432 + }, + { + "epoch": 0.61, + "learning_rate": 0.00031253761283851555, + "loss": 1.3414, + "theoretical_loss": 3.455415427562098, + "tokens_seen": 2024275968 + }, + { + "epoch": 0.61, + "learning_rate": 0.00031252156469408225, + "loss": 1.3632, + "theoretical_loss": 3.4554061991192766, + "tokens_seen": 2024341504 + }, + { + "epoch": 0.61, + "learning_rate": 0.00031250551654964895, + "loss": 1.3586, + "theoretical_loss": 3.4553969710588612, + "tokens_seen": 2024407040 + }, + { + "epoch": 0.61, + "learning_rate": 0.00031248946840521565, + "loss": 1.3049, + "theoretical_loss": 3.4553877433808253, + "tokens_seen": 2024472576 + }, + { + "epoch": 0.61, + "learning_rate": 0.00031247342026078235, + "loss": 1.3313, + "theoretical_loss": 3.4553785160851387, + "tokens_seen": 2024538112 + }, + { + "epoch": 0.61, + "learning_rate": 0.00031245737211634905, + "loss": 1.3445, + "theoretical_loss": 3.4553692891717747, + "tokens_seen": 2024603648 + }, + { + "epoch": 0.61, + "learning_rate": 0.00031244132397191575, + "loss": 1.3094, + "theoretical_loss": 3.455360062640705, + "tokens_seen": 2024669184 + }, + { + "epoch": 0.61, + "learning_rate": 0.00031242527582748245, + "loss": 1.2937, + "theoretical_loss": 3.4553508364919, + "tokens_seen": 2024734720 + }, + { + "epoch": 0.61, + "learning_rate": 0.00031240922768304916, + "loss": 1.424, + "theoretical_loss": 3.4553416107253327, + "tokens_seen": 2024800256 + }, + { + "epoch": 0.61, + "learning_rate": 0.00031239317953861586, + "loss": 1.4162, + "theoretical_loss": 3.455332385340975, + "tokens_seen": 2024865792 + }, + { + "epoch": 0.61, + "learning_rate": 0.00031237713139418256, + "loss": 1.2994, + "theoretical_loss": 3.455323160338798, + "tokens_seen": 2024931328 + }, + { + "epoch": 0.61, + "learning_rate": 0.00031236108324974926, + "loss": 1.2846, + "theoretical_loss": 3.455313935718774, + "tokens_seen": 2024996864 + }, + { + "epoch": 0.61, + "objective/train/docs_used": 738608, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.2674247026443481, + "objective/train/theoretical_loss": 3.4553047114808746, + "objective/train/tokens_used": 2045522400, + "theoretical_loss": 3.4553047114808746, + "tokens_seen": 2025062400 + }, + { + "epoch": 0.61, + "learning_rate": 0.00031234503510531596, + "loss": 1.2741, + "theoretical_loss": 3.4553047114808746, + "tokens_seen": 2025062400 + }, + { + "epoch": 0.61, + "learning_rate": 0.00031232898696088266, + "loss": 1.3973, + "theoretical_loss": 3.455295487625072, + "tokens_seen": 2025127936 + }, + { + "epoch": 0.61, + "learning_rate": 0.00031231293881644936, + "loss": 1.2434, + "theoretical_loss": 3.4552862641513373, + "tokens_seen": 2025193472 + }, + { + "epoch": 0.61, + "learning_rate": 0.00031229689067201606, + "loss": 1.3197, + "theoretical_loss": 3.4552770410596425, + "tokens_seen": 2025259008 + }, + { + "epoch": 0.61, + "learning_rate": 0.00031228084252758276, + "loss": 1.3029, + "theoretical_loss": 3.4552678183499594, + "tokens_seen": 2025324544 + }, + { + "epoch": 0.61, + "learning_rate": 0.00031226479438314946, + "loss": 1.2997, + "theoretical_loss": 3.4552585960222606, + "tokens_seen": 2025390080 + }, + { + "epoch": 0.61, + "learning_rate": 0.00031224874623871616, + "loss": 1.2604, + "theoretical_loss": 3.455249374076517, + "tokens_seen": 2025455616 + }, + { + "epoch": 0.61, + "learning_rate": 0.00031223269809428286, + "loss": 1.3121, + "theoretical_loss": 3.455240152512701, + "tokens_seen": 2025521152 + }, + { + "epoch": 0.61, + "learning_rate": 0.00031221664994984956, + "loss": 1.4, + "theoretical_loss": 3.455230931330784, + "tokens_seen": 2025586688 + }, + { + "epoch": 0.61, + "learning_rate": 0.00031220060180541626, + "loss": 1.3211, + "theoretical_loss": 3.4552217105307377, + "tokens_seen": 2025652224 + }, + { + "epoch": 0.61, + "learning_rate": 0.00031218455366098296, + "loss": 1.3548, + "theoretical_loss": 3.4552124901125345, + "tokens_seen": 2025717760 + }, + { + "epoch": 0.61, + "learning_rate": 0.00031216850551654966, + "loss": 1.3319, + "theoretical_loss": 3.4552032700761455, + "tokens_seen": 2025783296 + }, + { + "epoch": 0.61, + "learning_rate": 0.00031215245737211637, + "loss": 1.3492, + "theoretical_loss": 3.455194050421544, + "tokens_seen": 2025848832 + }, + { + "epoch": 0.61, + "learning_rate": 0.00031213640922768307, + "loss": 1.3217, + "theoretical_loss": 3.4551848311487, + "tokens_seen": 2025914368 + }, + { + "epoch": 0.61, + "learning_rate": 0.00031212036108324977, + "loss": 1.3385, + "theoretical_loss": 3.4551756122575865, + "tokens_seen": 2025979904 + }, + { + "epoch": 0.61, + "learning_rate": 0.00031210431293881647, + "loss": 1.2945, + "theoretical_loss": 3.4551663937481747, + "tokens_seen": 2026045440 + }, + { + "epoch": 0.61, + "learning_rate": 0.00031208826479438317, + "loss": 1.2257, + "theoretical_loss": 3.455157175620437, + "tokens_seen": 2026110976 + }, + { + "epoch": 0.61, + "learning_rate": 0.00031207221664994987, + "loss": 1.247, + "theoretical_loss": 3.455147957874345, + "tokens_seen": 2026176512 + }, + { + "epoch": 0.61, + "learning_rate": 0.00031205616850551657, + "loss": 1.2378, + "theoretical_loss": 3.45513874050987, + "tokens_seen": 2026242048 + }, + { + "epoch": 0.61, + "learning_rate": 0.00031204012036108327, + "loss": 1.327, + "theoretical_loss": 3.4551295235269857, + "tokens_seen": 2026307584 + }, + { + "epoch": 0.61, + "learning_rate": 0.00031202407221664997, + "loss": 1.3274, + "theoretical_loss": 3.4551203069256617, + "tokens_seen": 2026373120 + }, + { + "epoch": 0.61, + "learning_rate": 0.00031200802407221667, + "loss": 1.2948, + "theoretical_loss": 3.4551110907058713, + "tokens_seen": 2026438656 + }, + { + "epoch": 0.61, + "learning_rate": 0.00031199197592778337, + "loss": 1.3126, + "theoretical_loss": 3.455101874867586, + "tokens_seen": 2026504192 + }, + { + "epoch": 0.61, + "learning_rate": 0.00031197592778335007, + "loss": 1.308, + "theoretical_loss": 3.455092659410777, + "tokens_seen": 2026569728 + }, + { + "epoch": 0.61, + "learning_rate": 0.0003119598796389168, + "loss": 1.4006, + "theoretical_loss": 3.4550834443354175, + "tokens_seen": 2026635264 + }, + { + "epoch": 0.61, + "learning_rate": 0.0003119438314944835, + "loss": 1.259, + "theoretical_loss": 3.4550742296414785, + "tokens_seen": 2026700800 + }, + { + "epoch": 0.61, + "learning_rate": 0.0003119277833500502, + "loss": 1.3167, + "theoretical_loss": 3.455065015328932, + "tokens_seen": 2026766336 + }, + { + "epoch": 0.61, + "learning_rate": 0.0003119117352056169, + "loss": 1.3127, + "theoretical_loss": 3.4550558013977497, + "tokens_seen": 2026831872 + }, + { + "epoch": 0.61, + "learning_rate": 0.0003118956870611836, + "loss": 1.4521, + "theoretical_loss": 3.455046587847904, + "tokens_seen": 2026897408 + }, + { + "epoch": 0.61, + "learning_rate": 0.0003118796389167503, + "loss": 1.3316, + "theoretical_loss": 3.455037374679366, + "tokens_seen": 2026962944 + }, + { + "epoch": 0.61, + "learning_rate": 0.0003118635907723169, + "loss": 1.3109, + "theoretical_loss": 3.455028161892108, + "tokens_seen": 2027028480 + }, + { + "epoch": 0.61, + "learning_rate": 0.0003118475426278836, + "loss": 1.4285, + "theoretical_loss": 3.455018949486103, + "tokens_seen": 2027094016 + }, + { + "epoch": 0.61, + "learning_rate": 0.0003118314944834503, + "loss": 1.3735, + "theoretical_loss": 3.455009737461321, + "tokens_seen": 2027159552 + }, + { + "epoch": 0.61, + "learning_rate": 0.000311815446339017, + "loss": 1.2909, + "theoretical_loss": 3.455000525817735, + "tokens_seen": 2027225088 + }, + { + "epoch": 0.61, + "learning_rate": 0.0003117993981945837, + "loss": 1.3092, + "theoretical_loss": 3.4549913145553166, + "tokens_seen": 2027290624 + }, + { + "epoch": 0.61, + "learning_rate": 0.0003117833500501504, + "loss": 1.4636, + "theoretical_loss": 3.4549821036740385, + "tokens_seen": 2027356160 + }, + { + "epoch": 0.61, + "learning_rate": 0.0003117673019057171, + "loss": 1.3021, + "theoretical_loss": 3.454972893173871, + "tokens_seen": 2027421696 + }, + { + "epoch": 0.61, + "learning_rate": 0.0003117512537612838, + "loss": 1.3407, + "theoretical_loss": 3.454963683054787, + "tokens_seen": 2027487232 + }, + { + "epoch": 0.61, + "learning_rate": 0.00031173520561685053, + "loss": 1.365, + "theoretical_loss": 3.4549544733167585, + "tokens_seen": 2027552768 + }, + { + "epoch": 0.61, + "learning_rate": 0.00031171915747241723, + "loss": 1.3704, + "theoretical_loss": 3.4549452639597575, + "tokens_seen": 2027618304 + }, + { + "epoch": 0.61, + "learning_rate": 0.000311703109327984, + "loss": 1.4362, + "theoretical_loss": 3.4549360549837553, + "tokens_seen": 2027683840 + }, + { + "epoch": 0.61, + "learning_rate": 0.0003116870611835507, + "loss": 1.2599, + "theoretical_loss": 3.4549268463887244, + "tokens_seen": 2027749376 + }, + { + "epoch": 0.61, + "learning_rate": 0.0003116710130391174, + "loss": 1.3733, + "theoretical_loss": 3.4549176381746367, + "tokens_seen": 2027814912 + }, + { + "epoch": 0.61, + "learning_rate": 0.0003116549648946841, + "loss": 1.4011, + "theoretical_loss": 3.4549084303414634, + "tokens_seen": 2027880448 + }, + { + "epoch": 0.61, + "learning_rate": 0.0003116389167502508, + "loss": 1.3173, + "theoretical_loss": 3.4548992228891775, + "tokens_seen": 2027945984 + }, + { + "epoch": 0.61, + "learning_rate": 0.0003116228686058175, + "loss": 1.466, + "theoretical_loss": 3.45489001581775, + "tokens_seen": 2028011520 + }, + { + "epoch": 0.61, + "learning_rate": 0.0003116068204613842, + "loss": 1.3225, + "theoretical_loss": 3.4548808091271535, + "tokens_seen": 2028077056 + }, + { + "epoch": 0.61, + "learning_rate": 0.0003115907723169509, + "loss": 1.3242, + "theoretical_loss": 3.4548716028173594, + "tokens_seen": 2028142592 + }, + { + "epoch": 0.61, + "learning_rate": 0.0003115747241725176, + "loss": 1.2694, + "theoretical_loss": 3.4548623968883403, + "tokens_seen": 2028208128 + }, + { + "epoch": 0.61, + "learning_rate": 0.0003115586760280843, + "loss": 1.2076, + "theoretical_loss": 3.4548531913400677, + "tokens_seen": 2028273664 + }, + { + "epoch": 0.61, + "objective/train/docs_used": 739750, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.3667430877685547, + "objective/train/theoretical_loss": 3.4548439861725138, + "objective/train/tokens_used": 2048799200, + "theoretical_loss": 3.4548439861725138, + "tokens_seen": 2028339200 + }, + { + "epoch": 0.61, + "learning_rate": 0.000311542627883651, + "loss": 1.3062, + "theoretical_loss": 3.4548439861725138, + "tokens_seen": 2028339200 + }, + { + "epoch": 0.61, + "learning_rate": 0.0003115265797392177, + "loss": 1.2812, + "theoretical_loss": 3.45483478138565, + "tokens_seen": 2028404736 + }, + { + "epoch": 0.61, + "learning_rate": 0.0003115105315947844, + "loss": 1.351, + "theoretical_loss": 3.4548255769794487, + "tokens_seen": 2028470272 + }, + { + "epoch": 0.61, + "learning_rate": 0.0003114944834503511, + "loss": 1.3848, + "theoretical_loss": 3.454816372953882, + "tokens_seen": 2028535808 + }, + { + "epoch": 0.61, + "learning_rate": 0.0003114784353059178, + "loss": 1.2922, + "theoretical_loss": 3.4548071693089217, + "tokens_seen": 2028601344 + }, + { + "epoch": 0.61, + "learning_rate": 0.0003114623871614845, + "loss": 1.1887, + "theoretical_loss": 3.4547979660445396, + "tokens_seen": 2028666880 + }, + { + "epoch": 0.61, + "learning_rate": 0.0003114463390170512, + "loss": 1.3431, + "theoretical_loss": 3.4547887631607077, + "tokens_seen": 2028732416 + }, + { + "epoch": 0.61, + "learning_rate": 0.0003114302908726179, + "loss": 1.385, + "theoretical_loss": 3.4547795606573986, + "tokens_seen": 2028797952 + }, + { + "epoch": 0.61, + "learning_rate": 0.0003114142427281846, + "loss": 1.3568, + "theoretical_loss": 3.4547703585345833, + "tokens_seen": 2028863488 + }, + { + "epoch": 0.61, + "learning_rate": 0.0003113981945837513, + "loss": 1.3253, + "theoretical_loss": 3.4547611567922347, + "tokens_seen": 2028929024 + }, + { + "epoch": 0.61, + "learning_rate": 0.000311382146439318, + "loss": 1.2525, + "theoretical_loss": 3.4547519554303237, + "tokens_seen": 2028994560 + }, + { + "epoch": 0.61, + "learning_rate": 0.0003113660982948847, + "loss": 1.3022, + "theoretical_loss": 3.454742754448823, + "tokens_seen": 2029060096 + }, + { + "epoch": 0.61, + "learning_rate": 0.0003113500501504514, + "loss": 1.2945, + "theoretical_loss": 3.4547335538477046, + "tokens_seen": 2029125632 + }, + { + "epoch": 0.61, + "learning_rate": 0.0003113340020060181, + "loss": 1.2632, + "theoretical_loss": 3.4547243536269407, + "tokens_seen": 2029191168 + }, + { + "epoch": 0.61, + "learning_rate": 0.0003113179538615848, + "loss": 1.3276, + "theoretical_loss": 3.454715153786503, + "tokens_seen": 2029256704 + }, + { + "epoch": 0.61, + "learning_rate": 0.0003113019057171515, + "loss": 1.299, + "theoretical_loss": 3.454705954326363, + "tokens_seen": 2029322240 + }, + { + "epoch": 0.61, + "learning_rate": 0.0003112858575727182, + "loss": 1.3399, + "theoretical_loss": 3.454696755246493, + "tokens_seen": 2029387776 + }, + { + "epoch": 0.61, + "learning_rate": 0.0003112698094282849, + "loss": 1.2597, + "theoretical_loss": 3.4546875565468658, + "tokens_seen": 2029453312 + }, + { + "epoch": 0.62, + "learning_rate": 0.0003112537612838516, + "loss": 1.4382, + "theoretical_loss": 3.4546783582274525, + "tokens_seen": 2029518848 + }, + { + "epoch": 0.62, + "learning_rate": 0.00031123771313941825, + "loss": 1.318, + "theoretical_loss": 3.4546691602882253, + "tokens_seen": 2029584384 + }, + { + "epoch": 0.62, + "learning_rate": 0.00031122166499498495, + "loss": 1.3536, + "theoretical_loss": 3.4546599627291563, + "tokens_seen": 2029649920 + }, + { + "epoch": 0.62, + "learning_rate": 0.00031120561685055165, + "loss": 1.2757, + "theoretical_loss": 3.454650765550218, + "tokens_seen": 2029715456 + }, + { + "epoch": 0.62, + "learning_rate": 0.00031118956870611835, + "loss": 1.362, + "theoretical_loss": 3.4546415687513807, + "tokens_seen": 2029780992 + }, + { + "epoch": 0.62, + "learning_rate": 0.00031117352056168505, + "loss": 1.3287, + "theoretical_loss": 3.4546323723326187, + "tokens_seen": 2029846528 + }, + { + "epoch": 0.62, + "learning_rate": 0.00031115747241725175, + "loss": 1.3409, + "theoretical_loss": 3.454623176293903, + "tokens_seen": 2029912064 + }, + { + "epoch": 0.62, + "learning_rate": 0.00031114142427281845, + "loss": 1.4055, + "theoretical_loss": 3.4546139806352048, + "tokens_seen": 2029977600 + }, + { + "epoch": 0.62, + "learning_rate": 0.00031112537612838515, + "loss": 1.4365, + "theoretical_loss": 3.4546047853564974, + "tokens_seen": 2030043136 + }, + { + "epoch": 0.62, + "learning_rate": 0.00031110932798395185, + "loss": 1.3595, + "theoretical_loss": 3.4545955904577523, + "tokens_seen": 2030108672 + }, + { + "epoch": 0.62, + "learning_rate": 0.00031109327983951855, + "loss": 1.3158, + "theoretical_loss": 3.4545863959389416, + "tokens_seen": 2030174208 + }, + { + "epoch": 0.62, + "learning_rate": 0.00031107723169508525, + "loss": 1.2439, + "theoretical_loss": 3.4545772018000376, + "tokens_seen": 2030239744 + }, + { + "epoch": 0.62, + "learning_rate": 0.00031106118355065195, + "loss": 1.2677, + "theoretical_loss": 3.4545680080410115, + "tokens_seen": 2030305280 + }, + { + "epoch": 0.62, + "learning_rate": 0.00031104513540621866, + "loss": 1.357, + "theoretical_loss": 3.4545588146618362, + "tokens_seen": 2030370816 + }, + { + "epoch": 0.62, + "learning_rate": 0.00031102908726178536, + "loss": 1.4008, + "theoretical_loss": 3.4545496216624834, + "tokens_seen": 2030436352 + }, + { + "epoch": 0.62, + "learning_rate": 0.00031101303911735206, + "loss": 1.4644, + "theoretical_loss": 3.454540429042925, + "tokens_seen": 2030501888 + }, + { + "epoch": 0.62, + "learning_rate": 0.00031099699097291876, + "loss": 1.3659, + "theoretical_loss": 3.454531236803134, + "tokens_seen": 2030567424 + }, + { + "epoch": 0.62, + "learning_rate": 0.00031098094282848546, + "loss": 1.3507, + "theoretical_loss": 3.4545220449430807, + "tokens_seen": 2030632960 + }, + { + "epoch": 0.62, + "learning_rate": 0.00031096489468405216, + "loss": 1.4235, + "theoretical_loss": 3.4545128534627385, + "tokens_seen": 2030698496 + }, + { + "epoch": 0.62, + "learning_rate": 0.00031094884653961886, + "loss": 1.2718, + "theoretical_loss": 3.4545036623620793, + "tokens_seen": 2030764032 + }, + { + "epoch": 0.62, + "learning_rate": 0.00031093279839518556, + "loss": 1.3604, + "theoretical_loss": 3.454494471641075, + "tokens_seen": 2030829568 + }, + { + "epoch": 0.62, + "learning_rate": 0.00031091675025075226, + "loss": 1.4028, + "theoretical_loss": 3.4544852812996973, + "tokens_seen": 2030895104 + }, + { + "epoch": 0.62, + "learning_rate": 0.00031090070210631896, + "loss": 1.3524, + "theoretical_loss": 3.454476091337919, + "tokens_seen": 2030960640 + }, + { + "epoch": 0.62, + "learning_rate": 0.00031088465396188566, + "loss": 1.3165, + "theoretical_loss": 3.4544669017557115, + "tokens_seen": 2031026176 + }, + { + "epoch": 0.62, + "learning_rate": 0.00031086860581745236, + "loss": 1.3858, + "theoretical_loss": 3.4544577125530473, + "tokens_seen": 2031091712 + }, + { + "epoch": 0.62, + "learning_rate": 0.00031085255767301906, + "loss": 1.3057, + "theoretical_loss": 3.4544485237298987, + "tokens_seen": 2031157248 + }, + { + "epoch": 0.62, + "learning_rate": 0.00031083650952858576, + "loss": 1.2862, + "theoretical_loss": 3.454439335286237, + "tokens_seen": 2031222784 + }, + { + "epoch": 0.62, + "learning_rate": 0.00031082046138415246, + "loss": 1.3183, + "theoretical_loss": 3.454430147222035, + "tokens_seen": 2031288320 + }, + { + "epoch": 0.62, + "learning_rate": 0.00031080441323971916, + "loss": 1.292, + "theoretical_loss": 3.4544209595372637, + "tokens_seen": 2031353856 + }, + { + "epoch": 0.62, + "learning_rate": 0.00031078836509528587, + "loss": 1.2418, + "theoretical_loss": 3.4544117722318965, + "tokens_seen": 2031419392 + }, + { + "epoch": 0.62, + "learning_rate": 0.00031077231695085257, + "loss": 1.297, + "theoretical_loss": 3.4544025853059055, + "tokens_seen": 2031484928 + }, + { + "epoch": 0.62, + "learning_rate": 0.00031075626880641927, + "loss": 1.4024, + "theoretical_loss": 3.4543933987592617, + "tokens_seen": 2031550464 + }, + { + "epoch": 0.62, + "objective/train/docs_used": 741020, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.3118927478790283, + "objective/train/theoretical_loss": 3.454384212591938, + "objective/train/tokens_used": 2052076000, + "theoretical_loss": 3.454384212591938, + "tokens_seen": 2031616000 + }, + { + "epoch": 0.62, + "learning_rate": 0.00031074022066198597, + "loss": 1.3195, + "theoretical_loss": 3.454384212591938, + "tokens_seen": 2031616000 + }, + { + "epoch": 0.62, + "learning_rate": 0.00031072417251755267, + "loss": 1.2888, + "theoretical_loss": 3.4543750268039064, + "tokens_seen": 2031681536 + }, + { + "epoch": 0.62, + "learning_rate": 0.00031070812437311937, + "loss": 1.3178, + "theoretical_loss": 3.4543658413951386, + "tokens_seen": 2031747072 + }, + { + "epoch": 0.62, + "learning_rate": 0.00031069207622868607, + "loss": 1.3993, + "theoretical_loss": 3.454356656365607, + "tokens_seen": 2031812608 + }, + { + "epoch": 0.62, + "learning_rate": 0.00031067602808425277, + "loss": 1.2382, + "theoretical_loss": 3.454347471715284, + "tokens_seen": 2031878144 + }, + { + "epoch": 0.62, + "learning_rate": 0.00031065997993981947, + "loss": 1.332, + "theoretical_loss": 3.4543382874441413, + "tokens_seen": 2031943680 + }, + { + "epoch": 0.62, + "learning_rate": 0.00031064393179538617, + "loss": 1.3837, + "theoretical_loss": 3.454329103552151, + "tokens_seen": 2032009216 + }, + { + "epoch": 0.62, + "learning_rate": 0.00031062788365095287, + "loss": 1.2819, + "theoretical_loss": 3.454319920039285, + "tokens_seen": 2032074752 + }, + { + "epoch": 0.62, + "learning_rate": 0.00031061183550651957, + "loss": 1.3904, + "theoretical_loss": 3.4543107369055166, + "tokens_seen": 2032140288 + }, + { + "epoch": 0.62, + "learning_rate": 0.00031059578736208627, + "loss": 1.3432, + "theoretical_loss": 3.4543015541508164, + "tokens_seen": 2032205824 + }, + { + "epoch": 0.62, + "learning_rate": 0.000310579739217653, + "loss": 1.2244, + "theoretical_loss": 3.454292371775158, + "tokens_seen": 2032271360 + }, + { + "epoch": 0.62, + "learning_rate": 0.0003105636910732197, + "loss": 1.2759, + "theoretical_loss": 3.4542831897785122, + "tokens_seen": 2032336896 + }, + { + "epoch": 0.62, + "learning_rate": 0.0003105476429287864, + "loss": 1.3644, + "theoretical_loss": 3.4542740081608514, + "tokens_seen": 2032402432 + }, + { + "epoch": 0.62, + "learning_rate": 0.0003105315947843531, + "loss": 1.2883, + "theoretical_loss": 3.4542648269221488, + "tokens_seen": 2032467968 + }, + { + "epoch": 0.62, + "learning_rate": 0.0003105155466399198, + "loss": 1.4147, + "theoretical_loss": 3.454255646062375, + "tokens_seen": 2032533504 + }, + { + "epoch": 0.62, + "learning_rate": 0.0003104994984954865, + "loss": 1.3118, + "theoretical_loss": 3.4542464655815035, + "tokens_seen": 2032599040 + }, + { + "epoch": 0.62, + "learning_rate": 0.0003104834503510532, + "loss": 1.3498, + "theoretical_loss": 3.454237285479506, + "tokens_seen": 2032664576 + }, + { + "epoch": 0.62, + "learning_rate": 0.0003104674022066199, + "loss": 1.202, + "theoretical_loss": 3.454228105756354, + "tokens_seen": 2032730112 + }, + { + "epoch": 0.62, + "learning_rate": 0.0003104513540621866, + "loss": 1.2543, + "theoretical_loss": 3.4542189264120204, + "tokens_seen": 2032795648 + }, + { + "epoch": 0.62, + "learning_rate": 0.0003104353059177533, + "loss": 1.3154, + "theoretical_loss": 3.4542097474464772, + "tokens_seen": 2032861184 + }, + { + "epoch": 0.62, + "learning_rate": 0.00031041925777332, + "loss": 1.3075, + "theoretical_loss": 3.4542005688596964, + "tokens_seen": 2032926720 + }, + { + "epoch": 0.62, + "learning_rate": 0.0003104032096288867, + "loss": 1.3453, + "theoretical_loss": 3.4541913906516504, + "tokens_seen": 2032992256 + }, + { + "epoch": 0.62, + "learning_rate": 0.0003103871614844534, + "loss": 1.435, + "theoretical_loss": 3.454182212822311, + "tokens_seen": 2033057792 + }, + { + "epoch": 0.62, + "learning_rate": 0.0003103711133400201, + "loss": 1.3443, + "theoretical_loss": 3.4541730353716504, + "tokens_seen": 2033123328 + }, + { + "epoch": 0.62, + "learning_rate": 0.0003103550651955868, + "loss": 1.3835, + "theoretical_loss": 3.4541638582996415, + "tokens_seen": 2033188864 + }, + { + "epoch": 0.62, + "learning_rate": 0.0003103390170511535, + "loss": 1.2885, + "theoretical_loss": 3.454154681606256, + "tokens_seen": 2033254400 + }, + { + "epoch": 0.62, + "learning_rate": 0.0003103229689067202, + "loss": 1.4965, + "theoretical_loss": 3.454145505291465, + "tokens_seen": 2033319936 + }, + { + "epoch": 0.62, + "learning_rate": 0.0003103069207622869, + "loss": 1.4489, + "theoretical_loss": 3.4541363293552427, + "tokens_seen": 2033385472 + }, + { + "epoch": 0.62, + "learning_rate": 0.0003102908726178536, + "loss": 1.355, + "theoretical_loss": 3.4541271537975593, + "tokens_seen": 2033451008 + }, + { + "epoch": 0.62, + "learning_rate": 0.0003102748244734203, + "loss": 1.2901, + "theoretical_loss": 3.4541179786183887, + "tokens_seen": 2033516544 + }, + { + "epoch": 0.62, + "learning_rate": 0.000310258776328987, + "loss": 1.3929, + "theoretical_loss": 3.454108803817702, + "tokens_seen": 2033582080 + }, + { + "epoch": 0.62, + "learning_rate": 0.0003102427281845537, + "loss": 1.2952, + "theoretical_loss": 3.454099629395472, + "tokens_seen": 2033647616 + }, + { + "epoch": 0.62, + "learning_rate": 0.0003102266800401204, + "loss": 1.3703, + "theoretical_loss": 3.45409045535167, + "tokens_seen": 2033713152 + }, + { + "epoch": 0.62, + "learning_rate": 0.0003102106318956871, + "loss": 1.2875, + "theoretical_loss": 3.4540812816862694, + "tokens_seen": 2033778688 + }, + { + "epoch": 0.62, + "learning_rate": 0.0003101945837512538, + "loss": 1.2673, + "theoretical_loss": 3.4540721083992416, + "tokens_seen": 2033844224 + }, + { + "epoch": 0.62, + "learning_rate": 0.0003101785356068205, + "loss": 1.2646, + "theoretical_loss": 3.454062935490559, + "tokens_seen": 2033909760 + }, + { + "epoch": 0.62, + "learning_rate": 0.0003101624874623872, + "loss": 1.3142, + "theoretical_loss": 3.454053762960194, + "tokens_seen": 2033975296 + }, + { + "epoch": 0.62, + "learning_rate": 0.0003101464393179539, + "loss": 1.3057, + "theoretical_loss": 3.4540445908081185, + "tokens_seen": 2034040832 + }, + { + "epoch": 0.62, + "learning_rate": 0.0003101303911735206, + "loss": 1.4243, + "theoretical_loss": 3.454035419034305, + "tokens_seen": 2034106368 + }, + { + "epoch": 0.62, + "learning_rate": 0.0003101143430290873, + "loss": 1.2856, + "theoretical_loss": 3.4540262476387253, + "tokens_seen": 2034171904 + }, + { + "epoch": 0.62, + "learning_rate": 0.000310098294884654, + "loss": 1.3413, + "theoretical_loss": 3.4540170766213523, + "tokens_seen": 2034237440 + }, + { + "epoch": 0.62, + "learning_rate": 0.0003100822467402207, + "loss": 1.3721, + "theoretical_loss": 3.4540079059821576, + "tokens_seen": 2034302976 + }, + { + "epoch": 0.62, + "learning_rate": 0.0003100661985957874, + "loss": 1.4041, + "theoretical_loss": 3.453998735721113, + "tokens_seen": 2034368512 + }, + { + "epoch": 0.62, + "learning_rate": 0.0003100501504513541, + "loss": 1.2783, + "theoretical_loss": 3.453989565838192, + "tokens_seen": 2034434048 + }, + { + "epoch": 0.62, + "learning_rate": 0.0003100341023069208, + "loss": 1.3247, + "theoretical_loss": 3.4539803963333657, + "tokens_seen": 2034499584 + }, + { + "epoch": 0.62, + "learning_rate": 0.0003100180541624875, + "loss": 1.2942, + "theoretical_loss": 3.4539712272066074, + "tokens_seen": 2034565120 + }, + { + "epoch": 0.62, + "learning_rate": 0.0003100020060180542, + "loss": 1.4269, + "theoretical_loss": 3.453962058457888, + "tokens_seen": 2034630656 + }, + { + "epoch": 0.62, + "learning_rate": 0.0003099859578736209, + "loss": 1.2582, + "theoretical_loss": 3.453952890087181, + "tokens_seen": 2034696192 + }, + { + "epoch": 0.62, + "learning_rate": 0.0003099699097291876, + "loss": 1.27, + "theoretical_loss": 3.4539437220944578, + "tokens_seen": 2034761728 + }, + { + "epoch": 0.62, + "learning_rate": 0.0003099538615847543, + "loss": 1.3157, + "theoretical_loss": 3.453934554479691, + "tokens_seen": 2034827264 + }, + { + "epoch": 0.62, + "objective/train/docs_used": 742098, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.4511041641235352, + "objective/train/theoretical_loss": 3.453925387242853, + "objective/train/tokens_used": 2055352800, + "theoretical_loss": 3.453925387242853, + "tokens_seen": 2034892800 + }, + { + "epoch": 0.62, + "learning_rate": 0.000309937813440321, + "loss": 1.3584, + "theoretical_loss": 3.453925387242853, + "tokens_seen": 2034892800 + }, + { + "epoch": 0.62, + "learning_rate": 0.0003099217652958877, + "loss": 1.3787, + "theoretical_loss": 3.453916220383916, + "tokens_seen": 2034958336 + }, + { + "epoch": 0.62, + "learning_rate": 0.0003099057171514544, + "loss": 1.2745, + "theoretical_loss": 3.4539070539028516, + "tokens_seen": 2035023872 + }, + { + "epoch": 0.62, + "learning_rate": 0.0003098896690070211, + "loss": 1.3845, + "theoretical_loss": 3.4538978877996325, + "tokens_seen": 2035089408 + }, + { + "epoch": 0.62, + "learning_rate": 0.0003098736208625878, + "loss": 1.265, + "theoretical_loss": 3.4538887220742316, + "tokens_seen": 2035154944 + }, + { + "epoch": 0.62, + "learning_rate": 0.0003098575727181545, + "loss": 1.455, + "theoretical_loss": 3.45387955672662, + "tokens_seen": 2035220480 + }, + { + "epoch": 0.62, + "learning_rate": 0.0003098415245737212, + "loss": 1.3191, + "theoretical_loss": 3.4538703917567712, + "tokens_seen": 2035286016 + }, + { + "epoch": 0.62, + "learning_rate": 0.00030982547642928785, + "loss": 1.3413, + "theoretical_loss": 3.4538612271646563, + "tokens_seen": 2035351552 + }, + { + "epoch": 0.62, + "learning_rate": 0.00030980942828485455, + "loss": 1.3059, + "theoretical_loss": 3.453852062950248, + "tokens_seen": 2035417088 + }, + { + "epoch": 0.62, + "learning_rate": 0.00030979338014042125, + "loss": 1.3696, + "theoretical_loss": 3.453842899113519, + "tokens_seen": 2035482624 + }, + { + "epoch": 0.62, + "learning_rate": 0.00030977733199598795, + "loss": 1.3955, + "theoretical_loss": 3.453833735654441, + "tokens_seen": 2035548160 + }, + { + "epoch": 0.62, + "learning_rate": 0.00030976128385155465, + "loss": 1.2733, + "theoretical_loss": 3.453824572572987, + "tokens_seen": 2035613696 + }, + { + "epoch": 0.62, + "learning_rate": 0.00030974523570712135, + "loss": 1.2382, + "theoretical_loss": 3.453815409869128, + "tokens_seen": 2035679232 + }, + { + "epoch": 0.62, + "learning_rate": 0.00030972918756268805, + "loss": 1.2447, + "theoretical_loss": 3.4538062475428375, + "tokens_seen": 2035744768 + }, + { + "epoch": 0.62, + "learning_rate": 0.00030971313941825475, + "loss": 1.3268, + "theoretical_loss": 3.453797085594087, + "tokens_seen": 2035810304 + }, + { + "epoch": 0.62, + "learning_rate": 0.00030969709127382145, + "loss": 1.3743, + "theoretical_loss": 3.4537879240228495, + "tokens_seen": 2035875840 + }, + { + "epoch": 0.62, + "learning_rate": 0.00030968104312938815, + "loss": 1.3311, + "theoretical_loss": 3.453778762829097, + "tokens_seen": 2035941376 + }, + { + "epoch": 0.62, + "learning_rate": 0.00030966499498495486, + "loss": 1.323, + "theoretical_loss": 3.4537696020128017, + "tokens_seen": 2036006912 + }, + { + "epoch": 0.62, + "learning_rate": 0.00030964894684052156, + "loss": 1.3705, + "theoretical_loss": 3.4537604415739356, + "tokens_seen": 2036072448 + }, + { + "epoch": 0.62, + "learning_rate": 0.00030963289869608826, + "loss": 1.3225, + "theoretical_loss": 3.4537512815124716, + "tokens_seen": 2036137984 + }, + { + "epoch": 0.62, + "learning_rate": 0.00030961685055165496, + "loss": 1.375, + "theoretical_loss": 3.4537421218283817, + "tokens_seen": 2036203520 + }, + { + "epoch": 0.62, + "learning_rate": 0.00030960080240722166, + "loss": 1.4035, + "theoretical_loss": 3.4537329625216384, + "tokens_seen": 2036269056 + }, + { + "epoch": 0.62, + "learning_rate": 0.00030958475426278836, + "loss": 1.3986, + "theoretical_loss": 3.4537238035922133, + "tokens_seen": 2036334592 + }, + { + "epoch": 0.62, + "learning_rate": 0.00030956870611835506, + "loss": 1.3697, + "theoretical_loss": 3.4537146450400797, + "tokens_seen": 2036400128 + }, + { + "epoch": 0.62, + "learning_rate": 0.00030955265797392176, + "loss": 1.3635, + "theoretical_loss": 3.4537054868652097, + "tokens_seen": 2036465664 + }, + { + "epoch": 0.62, + "learning_rate": 0.00030953660982948846, + "loss": 1.3431, + "theoretical_loss": 3.4536963290675753, + "tokens_seen": 2036531200 + }, + { + "epoch": 0.62, + "learning_rate": 0.00030952056168505516, + "loss": 1.3052, + "theoretical_loss": 3.453687171647149, + "tokens_seen": 2036596736 + }, + { + "epoch": 0.62, + "learning_rate": 0.00030950451354062186, + "loss": 1.2589, + "theoretical_loss": 3.4536780146039026, + "tokens_seen": 2036662272 + }, + { + "epoch": 0.62, + "learning_rate": 0.00030948846539618856, + "loss": 1.3407, + "theoretical_loss": 3.453668857937809, + "tokens_seen": 2036727808 + }, + { + "epoch": 0.62, + "learning_rate": 0.00030947241725175526, + "loss": 1.3758, + "theoretical_loss": 3.4536597016488404, + "tokens_seen": 2036793344 + }, + { + "epoch": 0.62, + "learning_rate": 0.00030945636910732196, + "loss": 1.4059, + "theoretical_loss": 3.4536505457369695, + "tokens_seen": 2036858880 + }, + { + "epoch": 0.62, + "learning_rate": 0.00030944032096288866, + "loss": 1.35, + "theoretical_loss": 3.453641390202168, + "tokens_seen": 2036924416 + }, + { + "epoch": 0.62, + "learning_rate": 0.00030942427281845536, + "loss": 1.3213, + "theoretical_loss": 3.4536322350444086, + "tokens_seen": 2036989952 + }, + { + "epoch": 0.62, + "learning_rate": 0.00030940822467402207, + "loss": 1.3915, + "theoretical_loss": 3.4536230802636636, + "tokens_seen": 2037055488 + }, + { + "epoch": 0.62, + "learning_rate": 0.00030939217652958877, + "loss": 1.392, + "theoretical_loss": 3.4536139258599055, + "tokens_seen": 2037121024 + }, + { + "epoch": 0.62, + "learning_rate": 0.00030937612838515547, + "loss": 1.3269, + "theoretical_loss": 3.453604771833106, + "tokens_seen": 2037186560 + }, + { + "epoch": 0.62, + "learning_rate": 0.00030936008024072217, + "loss": 1.266, + "theoretical_loss": 3.4535956181832383, + "tokens_seen": 2037252096 + }, + { + "epoch": 0.62, + "learning_rate": 0.00030934403209628887, + "loss": 1.3469, + "theoretical_loss": 3.4535864649102743, + "tokens_seen": 2037317632 + }, + { + "epoch": 0.62, + "learning_rate": 0.00030932798395185557, + "loss": 1.3824, + "theoretical_loss": 3.453577312014186, + "tokens_seen": 2037383168 + }, + { + "epoch": 0.62, + "learning_rate": 0.00030931193580742227, + "loss": 1.3107, + "theoretical_loss": 3.4535681594949468, + "tokens_seen": 2037448704 + }, + { + "epoch": 0.62, + "learning_rate": 0.000309295887662989, + "loss": 1.2726, + "theoretical_loss": 3.4535590073525286, + "tokens_seen": 2037514240 + }, + { + "epoch": 0.62, + "learning_rate": 0.0003092798395185557, + "loss": 1.313, + "theoretical_loss": 3.4535498555869033, + "tokens_seen": 2037579776 + }, + { + "epoch": 0.62, + "learning_rate": 0.0003092637913741224, + "loss": 1.2403, + "theoretical_loss": 3.4535407041980433, + "tokens_seen": 2037645312 + }, + { + "epoch": 0.62, + "learning_rate": 0.0003092477432296891, + "loss": 1.429, + "theoretical_loss": 3.453531553185922, + "tokens_seen": 2037710848 + }, + { + "epoch": 0.62, + "learning_rate": 0.0003092316950852558, + "loss": 1.3572, + "theoretical_loss": 3.45352240255051, + "tokens_seen": 2037776384 + }, + { + "epoch": 0.62, + "learning_rate": 0.00030921564694082253, + "loss": 1.3253, + "theoretical_loss": 3.4535132522917813, + "tokens_seen": 2037841920 + }, + { + "epoch": 0.62, + "learning_rate": 0.0003091995987963892, + "loss": 1.2837, + "theoretical_loss": 3.4535041024097075, + "tokens_seen": 2037907456 + }, + { + "epoch": 0.62, + "learning_rate": 0.0003091835506519559, + "loss": 1.3752, + "theoretical_loss": 3.4534949529042613, + "tokens_seen": 2037972992 + }, + { + "epoch": 0.62, + "learning_rate": 0.0003091675025075226, + "loss": 1.2765, + "theoretical_loss": 3.4534858037754153, + "tokens_seen": 2038038528 + }, + { + "epoch": 0.62, + "learning_rate": 0.0003091514543630893, + "loss": 1.3586, + "theoretical_loss": 3.453476655023141, + "tokens_seen": 2038104064 + }, + { + "epoch": 0.62, + "objective/train/docs_used": 743253, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.2579412460327148, + "objective/train/theoretical_loss": 3.453467506647412, + "objective/train/tokens_used": 2058629600, + "theoretical_loss": 3.453467506647412, + "tokens_seen": 2038169600 + }, + { + "epoch": 0.62, + "learning_rate": 0.000309135406218656, + "loss": 1.3179, + "theoretical_loss": 3.453467506647412, + "tokens_seen": 2038169600 + }, + { + "epoch": 0.62, + "learning_rate": 0.0003091193580742227, + "loss": 1.3659, + "theoretical_loss": 3.4534583586481995, + "tokens_seen": 2038235136 + }, + { + "epoch": 0.62, + "learning_rate": 0.0003091033099297894, + "loss": 1.3127, + "theoretical_loss": 3.453449211025476, + "tokens_seen": 2038300672 + }, + { + "epoch": 0.62, + "learning_rate": 0.0003090872617853561, + "loss": 1.321, + "theoretical_loss": 3.4534400637792153, + "tokens_seen": 2038366208 + }, + { + "epoch": 0.62, + "learning_rate": 0.0003090712136409228, + "loss": 1.2875, + "theoretical_loss": 3.4534309169093884, + "tokens_seen": 2038431744 + }, + { + "epoch": 0.62, + "learning_rate": 0.0003090551654964895, + "loss": 1.287, + "theoretical_loss": 3.453421770415968, + "tokens_seen": 2038497280 + }, + { + "epoch": 0.62, + "learning_rate": 0.0003090391173520562, + "loss": 1.2985, + "theoretical_loss": 3.453412624298927, + "tokens_seen": 2038562816 + }, + { + "epoch": 0.62, + "learning_rate": 0.0003090230692076229, + "loss": 1.4246, + "theoretical_loss": 3.4534034785582373, + "tokens_seen": 2038628352 + }, + { + "epoch": 0.62, + "learning_rate": 0.0003090070210631896, + "loss": 1.3865, + "theoretical_loss": 3.4533943331938715, + "tokens_seen": 2038693888 + }, + { + "epoch": 0.62, + "learning_rate": 0.0003089909729187563, + "loss": 1.3719, + "theoretical_loss": 3.453385188205802, + "tokens_seen": 2038759424 + }, + { + "epoch": 0.62, + "learning_rate": 0.000308974924774323, + "loss": 1.4186, + "theoretical_loss": 3.4533760435940013, + "tokens_seen": 2038824960 + }, + { + "epoch": 0.62, + "learning_rate": 0.0003089588766298897, + "loss": 1.2257, + "theoretical_loss": 3.453366899358442, + "tokens_seen": 2038890496 + }, + { + "epoch": 0.62, + "learning_rate": 0.0003089428284854564, + "loss": 1.3634, + "theoretical_loss": 3.4533577554990957, + "tokens_seen": 2038956032 + }, + { + "epoch": 0.62, + "learning_rate": 0.0003089267803410231, + "loss": 1.4037, + "theoretical_loss": 3.453348612015936, + "tokens_seen": 2039021568 + }, + { + "epoch": 0.62, + "learning_rate": 0.0003089107321965898, + "loss": 1.3484, + "theoretical_loss": 3.453339468908934, + "tokens_seen": 2039087104 + }, + { + "epoch": 0.62, + "learning_rate": 0.0003088946840521565, + "loss": 1.3853, + "theoretical_loss": 3.4533303261780635, + "tokens_seen": 2039152640 + }, + { + "epoch": 0.62, + "learning_rate": 0.0003088786359077232, + "loss": 1.3479, + "theoretical_loss": 3.453321183823296, + "tokens_seen": 2039218176 + }, + { + "epoch": 0.62, + "learning_rate": 0.0003088625877632899, + "loss": 1.3028, + "theoretical_loss": 3.4533120418446046, + "tokens_seen": 2039283712 + }, + { + "epoch": 0.62, + "learning_rate": 0.0003088465396188566, + "loss": 1.275, + "theoretical_loss": 3.453302900241961, + "tokens_seen": 2039349248 + }, + { + "epoch": 0.62, + "learning_rate": 0.0003088304914744233, + "loss": 1.3356, + "theoretical_loss": 3.453293759015338, + "tokens_seen": 2039414784 + }, + { + "epoch": 0.62, + "learning_rate": 0.00030881444332999, + "loss": 1.321, + "theoretical_loss": 3.4532846181647088, + "tokens_seen": 2039480320 + }, + { + "epoch": 0.62, + "learning_rate": 0.0003087983951855567, + "loss": 1.3011, + "theoretical_loss": 3.4532754776900445, + "tokens_seen": 2039545856 + }, + { + "epoch": 0.62, + "learning_rate": 0.0003087823470411234, + "loss": 1.2394, + "theoretical_loss": 3.4532663375913186, + "tokens_seen": 2039611392 + }, + { + "epoch": 0.62, + "learning_rate": 0.0003087662988966901, + "loss": 1.2996, + "theoretical_loss": 3.4532571978685027, + "tokens_seen": 2039676928 + }, + { + "epoch": 0.62, + "learning_rate": 0.0003087502507522568, + "loss": 1.3874, + "theoretical_loss": 3.45324805852157, + "tokens_seen": 2039742464 + }, + { + "epoch": 0.62, + "learning_rate": 0.0003087342026078235, + "loss": 1.3259, + "theoretical_loss": 3.4532389195504933, + "tokens_seen": 2039808000 + }, + { + "epoch": 0.62, + "learning_rate": 0.0003087181544633902, + "loss": 1.4397, + "theoretical_loss": 3.453229780955244, + "tokens_seen": 2039873536 + }, + { + "epoch": 0.62, + "learning_rate": 0.0003087021063189569, + "loss": 1.3543, + "theoretical_loss": 3.4532206427357948, + "tokens_seen": 2039939072 + }, + { + "epoch": 0.62, + "learning_rate": 0.0003086860581745236, + "loss": 1.3662, + "theoretical_loss": 3.4532115048921184, + "tokens_seen": 2040004608 + }, + { + "epoch": 0.62, + "learning_rate": 0.0003086700100300903, + "loss": 1.3127, + "theoretical_loss": 3.4532023674241876, + "tokens_seen": 2040070144 + }, + { + "epoch": 0.62, + "learning_rate": 0.000308653961885657, + "loss": 1.4379, + "theoretical_loss": 3.4531932303319746, + "tokens_seen": 2040135680 + }, + { + "epoch": 0.62, + "learning_rate": 0.0003086379137412237, + "loss": 1.3315, + "theoretical_loss": 3.4531840936154516, + "tokens_seen": 2040201216 + }, + { + "epoch": 0.62, + "learning_rate": 0.0003086218655967904, + "loss": 1.3038, + "theoretical_loss": 3.4531749572745913, + "tokens_seen": 2040266752 + }, + { + "epoch": 0.62, + "learning_rate": 0.0003086058174523571, + "loss": 1.4108, + "theoretical_loss": 3.4531658213093666, + "tokens_seen": 2040332288 + }, + { + "epoch": 0.62, + "learning_rate": 0.0003085897693079238, + "loss": 1.2484, + "theoretical_loss": 3.4531566857197493, + "tokens_seen": 2040397824 + }, + { + "epoch": 0.62, + "learning_rate": 0.0003085737211634905, + "loss": 1.2833, + "theoretical_loss": 3.4531475505057125, + "tokens_seen": 2040463360 + }, + { + "epoch": 0.62, + "learning_rate": 0.0003085576730190572, + "loss": 1.3041, + "theoretical_loss": 3.453138415667228, + "tokens_seen": 2040528896 + }, + { + "epoch": 0.62, + "learning_rate": 0.0003085416248746239, + "loss": 1.4033, + "theoretical_loss": 3.453129281204269, + "tokens_seen": 2040594432 + }, + { + "epoch": 0.62, + "learning_rate": 0.0003085255767301906, + "loss": 1.4216, + "theoretical_loss": 3.4531201471168074, + "tokens_seen": 2040659968 + }, + { + "epoch": 0.62, + "learning_rate": 0.0003085095285857573, + "loss": 1.3407, + "theoretical_loss": 3.4531110134048166, + "tokens_seen": 2040725504 + }, + { + "epoch": 0.62, + "learning_rate": 0.000308493480441324, + "loss": 1.3299, + "theoretical_loss": 3.453101880068268, + "tokens_seen": 2040791040 + }, + { + "epoch": 0.62, + "learning_rate": 0.0003084774322968907, + "loss": 1.3768, + "theoretical_loss": 3.4530927471071347, + "tokens_seen": 2040856576 + }, + { + "epoch": 0.62, + "learning_rate": 0.0003084613841524574, + "loss": 1.3136, + "theoretical_loss": 3.4530836145213892, + "tokens_seen": 2040922112 + }, + { + "epoch": 0.62, + "learning_rate": 0.0003084453360080241, + "loss": 1.3649, + "theoretical_loss": 3.453074482311004, + "tokens_seen": 2040987648 + }, + { + "epoch": 0.62, + "learning_rate": 0.00030842928786359075, + "loss": 1.3297, + "theoretical_loss": 3.453065350475952, + "tokens_seen": 2041053184 + }, + { + "epoch": 0.62, + "learning_rate": 0.00030841323971915745, + "loss": 1.2486, + "theoretical_loss": 3.4530562190162044, + "tokens_seen": 2041118720 + }, + { + "epoch": 0.62, + "learning_rate": 0.00030839719157472415, + "loss": 1.2816, + "theoretical_loss": 3.4530470879317354, + "tokens_seen": 2041184256 + }, + { + "epoch": 0.62, + "learning_rate": 0.00030838114343029085, + "loss": 1.3948, + "theoretical_loss": 3.4530379572225165, + "tokens_seen": 2041249792 + }, + { + "epoch": 0.62, + "learning_rate": 0.00030836509528585755, + "loss": 1.2958, + "theoretical_loss": 3.4530288268885205, + "tokens_seen": 2041315328 + }, + { + "epoch": 0.62, + "learning_rate": 0.00030834904714142425, + "loss": 1.3542, + "theoretical_loss": 3.45301969692972, + "tokens_seen": 2041380864 + }, + { + "epoch": 0.62, + "objective/train/docs_used": 744490, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.4903392791748047, + "objective/train/theoretical_loss": 3.4530105673460874, + "objective/train/tokens_used": 2061906400, + "theoretical_loss": 3.4530105673460874, + "tokens_seen": 2041446400 + }, + { + "epoch": 0.62, + "learning_rate": 0.00030833299899699095, + "loss": 1.4212, + "theoretical_loss": 3.4530105673460874, + "tokens_seen": 2041446400 + }, + { + "epoch": 0.62, + "learning_rate": 0.00030831695085255765, + "loss": 1.3473, + "theoretical_loss": 3.4530014381375955, + "tokens_seen": 2041511936 + }, + { + "epoch": 0.62, + "learning_rate": 0.00030830090270812436, + "loss": 1.3333, + "theoretical_loss": 3.4529923093042165, + "tokens_seen": 2041577472 + }, + { + "epoch": 0.62, + "learning_rate": 0.00030828485456369106, + "loss": 1.2889, + "theoretical_loss": 3.452983180845923, + "tokens_seen": 2041643008 + }, + { + "epoch": 0.62, + "learning_rate": 0.00030826880641925776, + "loss": 1.3932, + "theoretical_loss": 3.4529740527626878, + "tokens_seen": 2041708544 + }, + { + "epoch": 0.62, + "learning_rate": 0.00030825275827482446, + "loss": 1.4098, + "theoretical_loss": 3.4529649250544834, + "tokens_seen": 2041774080 + }, + { + "epoch": 0.62, + "learning_rate": 0.00030823671013039116, + "loss": 1.2378, + "theoretical_loss": 3.452955797721282, + "tokens_seen": 2041839616 + }, + { + "epoch": 0.62, + "learning_rate": 0.00030822066198595786, + "loss": 1.4066, + "theoretical_loss": 3.4529466707630565, + "tokens_seen": 2041905152 + }, + { + "epoch": 0.62, + "learning_rate": 0.00030820461384152456, + "loss": 1.2183, + "theoretical_loss": 3.45293754417978, + "tokens_seen": 2041970688 + }, + { + "epoch": 0.62, + "learning_rate": 0.00030818856569709126, + "loss": 1.2905, + "theoretical_loss": 3.4529284179714237, + "tokens_seen": 2042036224 + }, + { + "epoch": 0.62, + "learning_rate": 0.00030817251755265796, + "loss": 1.2242, + "theoretical_loss": 3.4529192921379606, + "tokens_seen": 2042101760 + }, + { + "epoch": 0.62, + "learning_rate": 0.00030815646940822466, + "loss": 1.3036, + "theoretical_loss": 3.4529101666793642, + "tokens_seen": 2042167296 + }, + { + "epoch": 0.62, + "learning_rate": 0.00030814042126379136, + "loss": 1.366, + "theoretical_loss": 3.4529010415956063, + "tokens_seen": 2042232832 + }, + { + "epoch": 0.62, + "learning_rate": 0.00030812437311935806, + "loss": 1.395, + "theoretical_loss": 3.4528919168866596, + "tokens_seen": 2042298368 + }, + { + "epoch": 0.62, + "learning_rate": 0.00030810832497492476, + "loss": 1.2571, + "theoretical_loss": 3.4528827925524967, + "tokens_seen": 2042363904 + }, + { + "epoch": 0.62, + "learning_rate": 0.0003080922768304915, + "loss": 1.3561, + "theoretical_loss": 3.4528736685930905, + "tokens_seen": 2042429440 + }, + { + "epoch": 0.62, + "learning_rate": 0.0003080762286860582, + "loss": 1.3254, + "theoretical_loss": 3.4528645450084126, + "tokens_seen": 2042494976 + }, + { + "epoch": 0.62, + "learning_rate": 0.0003080601805416249, + "loss": 1.3151, + "theoretical_loss": 3.4528554217984366, + "tokens_seen": 2042560512 + }, + { + "epoch": 0.62, + "learning_rate": 0.0003080441323971916, + "loss": 1.3241, + "theoretical_loss": 3.4528462989631348, + "tokens_seen": 2042626048 + }, + { + "epoch": 0.62, + "learning_rate": 0.0003080280842527583, + "loss": 1.4005, + "theoretical_loss": 3.45283717650248, + "tokens_seen": 2042691584 + }, + { + "epoch": 0.62, + "learning_rate": 0.000308012036108325, + "loss": 1.2727, + "theoretical_loss": 3.452828054416444, + "tokens_seen": 2042757120 + }, + { + "epoch": 0.62, + "learning_rate": 0.0003079959879638917, + "loss": 1.368, + "theoretical_loss": 3.452818932705, + "tokens_seen": 2042822656 + }, + { + "epoch": 0.62, + "learning_rate": 0.0003079799398194584, + "loss": 1.4015, + "theoretical_loss": 3.452809811368121, + "tokens_seen": 2042888192 + }, + { + "epoch": 0.62, + "learning_rate": 0.0003079638916750251, + "loss": 1.2854, + "theoretical_loss": 3.4528006904057786, + "tokens_seen": 2042953728 + }, + { + "epoch": 0.62, + "learning_rate": 0.0003079478435305918, + "loss": 1.3311, + "theoretical_loss": 3.452791569817946, + "tokens_seen": 2043019264 + }, + { + "epoch": 0.62, + "learning_rate": 0.0003079317953861585, + "loss": 1.3346, + "theoretical_loss": 3.4527824496045962, + "tokens_seen": 2043084800 + }, + { + "epoch": 0.62, + "learning_rate": 0.0003079157472417252, + "loss": 1.2293, + "theoretical_loss": 3.452773329765701, + "tokens_seen": 2043150336 + }, + { + "epoch": 0.62, + "learning_rate": 0.0003078996990972919, + "loss": 1.2555, + "theoretical_loss": 3.4527642103012335, + "tokens_seen": 2043215872 + }, + { + "epoch": 0.62, + "learning_rate": 0.0003078836509528586, + "loss": 1.3583, + "theoretical_loss": 3.452755091211166, + "tokens_seen": 2043281408 + }, + { + "epoch": 0.62, + "learning_rate": 0.0003078676028084253, + "loss": 1.2801, + "theoretical_loss": 3.452745972495472, + "tokens_seen": 2043346944 + }, + { + "epoch": 0.62, + "learning_rate": 0.000307851554663992, + "loss": 1.3419, + "theoretical_loss": 3.4527368541541223, + "tokens_seen": 2043412480 + }, + { + "epoch": 0.62, + "learning_rate": 0.00030783550651955873, + "loss": 1.4068, + "theoretical_loss": 3.452727736187091, + "tokens_seen": 2043478016 + }, + { + "epoch": 0.62, + "learning_rate": 0.00030781945837512543, + "loss": 1.2471, + "theoretical_loss": 3.452718618594351, + "tokens_seen": 2043543552 + }, + { + "epoch": 0.62, + "learning_rate": 0.00030780341023069213, + "loss": 1.3198, + "theoretical_loss": 3.4527095013758737, + "tokens_seen": 2043609088 + }, + { + "epoch": 0.62, + "learning_rate": 0.0003077873620862588, + "loss": 1.2941, + "theoretical_loss": 3.452700384531633, + "tokens_seen": 2043674624 + }, + { + "epoch": 0.62, + "learning_rate": 0.0003077713139418255, + "loss": 1.2761, + "theoretical_loss": 3.4526912680616006, + "tokens_seen": 2043740160 + }, + { + "epoch": 0.62, + "learning_rate": 0.0003077552657973922, + "loss": 1.3325, + "theoretical_loss": 3.4526821519657496, + "tokens_seen": 2043805696 + }, + { + "epoch": 0.62, + "learning_rate": 0.0003077392176529589, + "loss": 1.447, + "theoretical_loss": 3.452673036244052, + "tokens_seen": 2043871232 + }, + { + "epoch": 0.62, + "learning_rate": 0.0003077231695085256, + "loss": 1.2566, + "theoretical_loss": 3.452663920896481, + "tokens_seen": 2043936768 + }, + { + "epoch": 0.62, + "learning_rate": 0.0003077071213640923, + "loss": 1.4203, + "theoretical_loss": 3.45265480592301, + "tokens_seen": 2044002304 + }, + { + "epoch": 0.62, + "learning_rate": 0.000307691073219659, + "loss": 1.3598, + "theoretical_loss": 3.4526456913236103, + "tokens_seen": 2044067840 + }, + { + "epoch": 0.62, + "learning_rate": 0.0003076750250752257, + "loss": 1.266, + "theoretical_loss": 3.4526365770982546, + "tokens_seen": 2044133376 + }, + { + "epoch": 0.62, + "learning_rate": 0.0003076589769307924, + "loss": 1.3234, + "theoretical_loss": 3.452627463246917, + "tokens_seen": 2044198912 + }, + { + "epoch": 0.62, + "learning_rate": 0.0003076429287863591, + "loss": 1.3091, + "theoretical_loss": 3.452618349769568, + "tokens_seen": 2044264448 + }, + { + "epoch": 0.62, + "learning_rate": 0.0003076268806419258, + "loss": 1.3848, + "theoretical_loss": 3.4526092366661825, + "tokens_seen": 2044329984 + }, + { + "epoch": 0.62, + "learning_rate": 0.0003076108324974925, + "loss": 1.3187, + "theoretical_loss": 3.4526001239367323, + "tokens_seen": 2044395520 + }, + { + "epoch": 0.62, + "learning_rate": 0.0003075947843530592, + "loss": 1.3902, + "theoretical_loss": 3.4525910115811893, + "tokens_seen": 2044461056 + }, + { + "epoch": 0.62, + "learning_rate": 0.0003075787362086259, + "loss": 1.3328, + "theoretical_loss": 3.4525818995995268, + "tokens_seen": 2044526592 + }, + { + "epoch": 0.62, + "learning_rate": 0.0003075626880641926, + "loss": 1.3538, + "theoretical_loss": 3.452572787991718, + "tokens_seen": 2044592128 + }, + { + "epoch": 0.62, + "learning_rate": 0.0003075466399197593, + "loss": 1.4036, + "theoretical_loss": 3.452563676757735, + "tokens_seen": 2044657664 + }, + { + "epoch": 0.62, + "objective/train/docs_used": 745676, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.109898328781128, + "objective/train/theoretical_loss": 3.4525545658975494, + "objective/train/tokens_used": 2065183200, + "theoretical_loss": 3.4525545658975494, + "tokens_seen": 2044723200 + }, + { + "epoch": 0.62, + "learning_rate": 0.000307530591775326, + "loss": 1.2629, + "theoretical_loss": 3.4525545658975494, + "tokens_seen": 2044723200 + }, + { + "epoch": 0.62, + "learning_rate": 0.0003075145436308927, + "loss": 1.3352, + "theoretical_loss": 3.452545455411136, + "tokens_seen": 2044788736 + }, + { + "epoch": 0.62, + "learning_rate": 0.0003074984954864594, + "loss": 1.3121, + "theoretical_loss": 3.4525363452984665, + "tokens_seen": 2044854272 + }, + { + "epoch": 0.62, + "learning_rate": 0.0003074824473420261, + "loss": 1.3914, + "theoretical_loss": 3.452527235559513, + "tokens_seen": 2044919808 + }, + { + "epoch": 0.62, + "learning_rate": 0.0003074663991975928, + "loss": 1.3274, + "theoretical_loss": 3.4525181261942492, + "tokens_seen": 2044985344 + }, + { + "epoch": 0.62, + "learning_rate": 0.0003074503510531595, + "loss": 1.4314, + "theoretical_loss": 3.452509017202648, + "tokens_seen": 2045050880 + }, + { + "epoch": 0.62, + "learning_rate": 0.0003074343029087262, + "loss": 1.4198, + "theoretical_loss": 3.452499908584681, + "tokens_seen": 2045116416 + }, + { + "epoch": 0.62, + "learning_rate": 0.0003074182547642929, + "loss": 1.3409, + "theoretical_loss": 3.4524908003403207, + "tokens_seen": 2045181952 + }, + { + "epoch": 0.62, + "learning_rate": 0.0003074022066198596, + "loss": 1.3675, + "theoretical_loss": 3.452481692469541, + "tokens_seen": 2045247488 + }, + { + "epoch": 0.62, + "learning_rate": 0.0003073861584754263, + "loss": 1.2845, + "theoretical_loss": 3.4524725849723144, + "tokens_seen": 2045313024 + }, + { + "epoch": 0.62, + "learning_rate": 0.000307370110330993, + "loss": 1.3141, + "theoretical_loss": 3.452463477848613, + "tokens_seen": 2045378560 + }, + { + "epoch": 0.62, + "learning_rate": 0.0003073540621865597, + "loss": 1.4199, + "theoretical_loss": 3.4524543710984097, + "tokens_seen": 2045444096 + }, + { + "epoch": 0.62, + "learning_rate": 0.0003073380140421264, + "loss": 1.2449, + "theoretical_loss": 3.4524452647216775, + "tokens_seen": 2045509632 + }, + { + "epoch": 0.62, + "learning_rate": 0.0003073219658976931, + "loss": 1.2933, + "theoretical_loss": 3.452436158718389, + "tokens_seen": 2045575168 + }, + { + "epoch": 0.62, + "learning_rate": 0.0003073059177532598, + "loss": 1.2532, + "theoretical_loss": 3.452427053088517, + "tokens_seen": 2045640704 + }, + { + "epoch": 0.62, + "learning_rate": 0.0003072898696088265, + "loss": 1.2797, + "theoretical_loss": 3.452417947832034, + "tokens_seen": 2045706240 + }, + { + "epoch": 0.62, + "learning_rate": 0.0003072738214643932, + "loss": 1.2216, + "theoretical_loss": 3.452408842948912, + "tokens_seen": 2045771776 + }, + { + "epoch": 0.62, + "learning_rate": 0.0003072577733199599, + "loss": 1.2404, + "theoretical_loss": 3.4523997384391256, + "tokens_seen": 2045837312 + }, + { + "epoch": 0.62, + "learning_rate": 0.0003072417251755266, + "loss": 1.2669, + "theoretical_loss": 3.4523906343026463, + "tokens_seen": 2045902848 + }, + { + "epoch": 0.62, + "learning_rate": 0.0003072256770310933, + "loss": 1.2688, + "theoretical_loss": 3.4523815305394465, + "tokens_seen": 2045968384 + }, + { + "epoch": 0.62, + "learning_rate": 0.00030720962888666, + "loss": 1.2748, + "theoretical_loss": 3.4523724271495, + "tokens_seen": 2046033920 + }, + { + "epoch": 0.62, + "learning_rate": 0.0003071935807422267, + "loss": 1.3288, + "theoretical_loss": 3.452363324132779, + "tokens_seen": 2046099456 + }, + { + "epoch": 0.62, + "learning_rate": 0.0003071775325977934, + "loss": 1.33, + "theoretical_loss": 3.452354221489256, + "tokens_seen": 2046164992 + }, + { + "epoch": 0.62, + "learning_rate": 0.0003071614844533601, + "loss": 1.283, + "theoretical_loss": 3.452345119218904, + "tokens_seen": 2046230528 + }, + { + "epoch": 0.62, + "learning_rate": 0.0003071454363089268, + "loss": 1.3816, + "theoretical_loss": 3.452336017321696, + "tokens_seen": 2046296064 + }, + { + "epoch": 0.62, + "learning_rate": 0.0003071293881644935, + "loss": 1.3148, + "theoretical_loss": 3.452326915797604, + "tokens_seen": 2046361600 + }, + { + "epoch": 0.62, + "learning_rate": 0.0003071133400200602, + "loss": 1.3059, + "theoretical_loss": 3.4523178146466016, + "tokens_seen": 2046427136 + }, + { + "epoch": 0.62, + "learning_rate": 0.0003070972918756269, + "loss": 1.3631, + "theoretical_loss": 3.4523087138686614, + "tokens_seen": 2046492672 + }, + { + "epoch": 0.62, + "learning_rate": 0.0003070812437311936, + "loss": 1.3387, + "theoretical_loss": 3.452299613463756, + "tokens_seen": 2046558208 + }, + { + "epoch": 0.62, + "learning_rate": 0.0003070651955867603, + "loss": 1.3733, + "theoretical_loss": 3.4522905134318576, + "tokens_seen": 2046623744 + }, + { + "epoch": 0.62, + "learning_rate": 0.000307049147442327, + "loss": 1.2744, + "theoretical_loss": 3.4522814137729396, + "tokens_seen": 2046689280 + }, + { + "epoch": 0.62, + "learning_rate": 0.0003070330992978937, + "loss": 1.3312, + "theoretical_loss": 3.4522723144869754, + "tokens_seen": 2046754816 + }, + { + "epoch": 0.62, + "learning_rate": 0.00030701705115346035, + "loss": 1.2764, + "theoretical_loss": 3.4522632155739363, + "tokens_seen": 2046820352 + }, + { + "epoch": 0.62, + "learning_rate": 0.00030700100300902705, + "loss": 1.2855, + "theoretical_loss": 3.4522541170337964, + "tokens_seen": 2046885888 + }, + { + "epoch": 0.62, + "learning_rate": 0.00030698495486459375, + "loss": 1.3457, + "theoretical_loss": 3.4522450188665275, + "tokens_seen": 2046951424 + }, + { + "epoch": 0.62, + "learning_rate": 0.00030696890672016045, + "loss": 1.2783, + "theoretical_loss": 3.452235921072103, + "tokens_seen": 2047016960 + }, + { + "epoch": 0.62, + "learning_rate": 0.00030695285857572715, + "loss": 1.1917, + "theoretical_loss": 3.452226823650495, + "tokens_seen": 2047082496 + }, + { + "epoch": 0.62, + "learning_rate": 0.00030693681043129385, + "loss": 1.3787, + "theoretical_loss": 3.4522177266016776, + "tokens_seen": 2047148032 + }, + { + "epoch": 0.62, + "learning_rate": 0.00030692076228686056, + "loss": 1.3063, + "theoretical_loss": 3.452208629925622, + "tokens_seen": 2047213568 + }, + { + "epoch": 0.62, + "learning_rate": 0.0003069047141424273, + "loss": 1.2537, + "theoretical_loss": 3.452199533622302, + "tokens_seen": 2047279104 + }, + { + "epoch": 0.62, + "learning_rate": 0.000306888665997994, + "loss": 1.4697, + "theoretical_loss": 3.4521904376916903, + "tokens_seen": 2047344640 + }, + { + "epoch": 0.62, + "learning_rate": 0.0003068726178535607, + "loss": 1.3308, + "theoretical_loss": 3.4521813421337595, + "tokens_seen": 2047410176 + }, + { + "epoch": 0.62, + "learning_rate": 0.0003068565697091274, + "loss": 1.3299, + "theoretical_loss": 3.4521722469484826, + "tokens_seen": 2047475712 + }, + { + "epoch": 0.62, + "learning_rate": 0.0003068405215646941, + "loss": 1.2961, + "theoretical_loss": 3.4521631521358316, + "tokens_seen": 2047541248 + }, + { + "epoch": 0.62, + "learning_rate": 0.0003068244734202608, + "loss": 1.2426, + "theoretical_loss": 3.4521540576957808, + "tokens_seen": 2047606784 + }, + { + "epoch": 0.62, + "learning_rate": 0.0003068084252758275, + "loss": 1.3354, + "theoretical_loss": 3.4521449636283013, + "tokens_seen": 2047672320 + }, + { + "epoch": 0.62, + "learning_rate": 0.0003067923771313942, + "loss": 1.3277, + "theoretical_loss": 3.4521358699333677, + "tokens_seen": 2047737856 + }, + { + "epoch": 0.62, + "learning_rate": 0.0003067763289869609, + "loss": 1.3577, + "theoretical_loss": 3.4521267766109514, + "tokens_seen": 2047803392 + }, + { + "epoch": 0.62, + "learning_rate": 0.0003067602808425276, + "loss": 1.3513, + "theoretical_loss": 3.452117683661026, + "tokens_seen": 2047868928 + }, + { + "epoch": 0.62, + "learning_rate": 0.0003067442326980943, + "loss": 1.2324, + "theoretical_loss": 3.4521085910835634, + "tokens_seen": 2047934464 + }, + { + "epoch": 0.62, + "objective/train/docs_used": 746888, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.2309513092041016, + "objective/train/theoretical_loss": 3.4520994988785376, + "objective/train/tokens_used": 2068460000, + "theoretical_loss": 3.4520994988785376, + "tokens_seen": 2048000000 + }, + { + "epoch": 0.62, + "learning_rate": 0.000306728184553661, + "loss": 1.1919, + "theoretical_loss": 3.4520994988785376, + "tokens_seen": 2048000000 + }, + { + "epoch": 0.62, + "learning_rate": 0.0003067121364092277, + "loss": 1.3134, + "theoretical_loss": 3.4520904070459206, + "tokens_seen": 2048065536 + }, + { + "epoch": 0.62, + "learning_rate": 0.0003066960882647944, + "loss": 1.2305, + "theoretical_loss": 3.4520813155856858, + "tokens_seen": 2048131072 + }, + { + "epoch": 0.62, + "learning_rate": 0.0003066800401203611, + "loss": 1.4252, + "theoretical_loss": 3.4520722244978055, + "tokens_seen": 2048196608 + }, + { + "epoch": 0.62, + "learning_rate": 0.0003066639919759278, + "loss": 1.3542, + "theoretical_loss": 3.4520631337822527, + "tokens_seen": 2048262144 + }, + { + "epoch": 0.62, + "learning_rate": 0.0003066479438314945, + "loss": 1.4328, + "theoretical_loss": 3.4520540434390012, + "tokens_seen": 2048327680 + }, + { + "epoch": 0.62, + "learning_rate": 0.0003066318956870612, + "loss": 1.3597, + "theoretical_loss": 3.4520449534680218, + "tokens_seen": 2048393216 + }, + { + "epoch": 0.62, + "learning_rate": 0.0003066158475426279, + "loss": 1.3023, + "theoretical_loss": 3.4520358638692894, + "tokens_seen": 2048458752 + }, + { + "epoch": 0.62, + "learning_rate": 0.0003065997993981946, + "loss": 1.2927, + "theoretical_loss": 3.4520267746427757, + "tokens_seen": 2048524288 + }, + { + "epoch": 0.62, + "learning_rate": 0.0003065837512537613, + "loss": 1.4309, + "theoretical_loss": 3.4520176857884532, + "tokens_seen": 2048589824 + }, + { + "epoch": 0.62, + "learning_rate": 0.000306567703109328, + "loss": 1.3237, + "theoretical_loss": 3.452008597306296, + "tokens_seen": 2048655360 + }, + { + "epoch": 0.62, + "learning_rate": 0.0003065516549648947, + "loss": 1.3231, + "theoretical_loss": 3.451999509196276, + "tokens_seen": 2048720896 + }, + { + "epoch": 0.62, + "learning_rate": 0.0003065356068204614, + "loss": 1.3299, + "theoretical_loss": 3.451990421458367, + "tokens_seen": 2048786432 + }, + { + "epoch": 0.62, + "learning_rate": 0.0003065195586760281, + "loss": 1.2842, + "theoretical_loss": 3.4519813340925407, + "tokens_seen": 2048851968 + }, + { + "epoch": 0.62, + "learning_rate": 0.0003065035105315948, + "loss": 1.4642, + "theoretical_loss": 3.451972247098771, + "tokens_seen": 2048917504 + }, + { + "epoch": 0.62, + "learning_rate": 0.0003064874623871615, + "loss": 1.3443, + "theoretical_loss": 3.4519631604770296, + "tokens_seen": 2048983040 + }, + { + "epoch": 0.62, + "learning_rate": 0.00030647141424272823, + "loss": 1.4611, + "theoretical_loss": 3.4519540742272907, + "tokens_seen": 2049048576 + }, + { + "epoch": 0.62, + "learning_rate": 0.00030645536609829493, + "loss": 1.3095, + "theoretical_loss": 3.451944988349526, + "tokens_seen": 2049114112 + }, + { + "epoch": 0.62, + "learning_rate": 0.00030643931795386163, + "loss": 1.3307, + "theoretical_loss": 3.451935902843709, + "tokens_seen": 2049179648 + }, + { + "epoch": 0.62, + "learning_rate": 0.00030642326980942833, + "loss": 1.3117, + "theoretical_loss": 3.4519268177098126, + "tokens_seen": 2049245184 + }, + { + "epoch": 0.62, + "learning_rate": 0.00030640722166499503, + "loss": 1.2311, + "theoretical_loss": 3.4519177329478095, + "tokens_seen": 2049310720 + }, + { + "epoch": 0.62, + "learning_rate": 0.0003063911735205617, + "loss": 1.425, + "theoretical_loss": 3.451908648557673, + "tokens_seen": 2049376256 + }, + { + "epoch": 0.62, + "learning_rate": 0.0003063751253761284, + "loss": 1.3805, + "theoretical_loss": 3.451899564539375, + "tokens_seen": 2049441792 + }, + { + "epoch": 0.62, + "learning_rate": 0.0003063590772316951, + "loss": 1.3377, + "theoretical_loss": 3.451890480892889, + "tokens_seen": 2049507328 + }, + { + "epoch": 0.62, + "learning_rate": 0.0003063430290872618, + "loss": 1.3713, + "theoretical_loss": 3.4518813976181884, + "tokens_seen": 2049572864 + }, + { + "epoch": 0.62, + "learning_rate": 0.0003063269809428285, + "loss": 1.4175, + "theoretical_loss": 3.4518723147152457, + "tokens_seen": 2049638400 + }, + { + "epoch": 0.62, + "learning_rate": 0.0003063109327983952, + "loss": 1.3253, + "theoretical_loss": 3.4518632321840332, + "tokens_seen": 2049703936 + }, + { + "epoch": 0.62, + "learning_rate": 0.0003062948846539619, + "loss": 1.3548, + "theoretical_loss": 3.4518541500245243, + "tokens_seen": 2049769472 + }, + { + "epoch": 0.62, + "learning_rate": 0.0003062788365095286, + "loss": 1.3426, + "theoretical_loss": 3.4518450682366923, + "tokens_seen": 2049835008 + }, + { + "epoch": 0.62, + "learning_rate": 0.0003062627883650953, + "loss": 1.3649, + "theoretical_loss": 3.4518359868205097, + "tokens_seen": 2049900544 + }, + { + "epoch": 0.62, + "learning_rate": 0.000306246740220662, + "loss": 1.3479, + "theoretical_loss": 3.4518269057759494, + "tokens_seen": 2049966080 + }, + { + "epoch": 0.62, + "learning_rate": 0.0003062306920762287, + "loss": 1.4155, + "theoretical_loss": 3.4518178251029843, + "tokens_seen": 2050031616 + }, + { + "epoch": 0.62, + "learning_rate": 0.0003062146439317954, + "loss": 1.3952, + "theoretical_loss": 3.4518087448015873, + "tokens_seen": 2050097152 + }, + { + "epoch": 0.62, + "learning_rate": 0.0003061985957873621, + "loss": 1.3461, + "theoretical_loss": 3.4517996648717313, + "tokens_seen": 2050162688 + }, + { + "epoch": 0.62, + "learning_rate": 0.0003061825476429288, + "loss": 1.4501, + "theoretical_loss": 3.4517905853133897, + "tokens_seen": 2050228224 + }, + { + "epoch": 0.62, + "learning_rate": 0.0003061664994984955, + "loss": 1.3535, + "theoretical_loss": 3.4517815061265344, + "tokens_seen": 2050293760 + }, + { + "epoch": 0.62, + "learning_rate": 0.0003061504513540622, + "loss": 1.3702, + "theoretical_loss": 3.4517724273111394, + "tokens_seen": 2050359296 + }, + { + "epoch": 0.62, + "learning_rate": 0.0003061344032096289, + "loss": 1.3576, + "theoretical_loss": 3.4517633488671766, + "tokens_seen": 2050424832 + }, + { + "epoch": 0.62, + "learning_rate": 0.0003061183550651956, + "loss": 1.2544, + "theoretical_loss": 3.4517542707946203, + "tokens_seen": 2050490368 + }, + { + "epoch": 0.62, + "learning_rate": 0.0003061023069207623, + "loss": 1.357, + "theoretical_loss": 3.4517451930934424, + "tokens_seen": 2050555904 + }, + { + "epoch": 0.62, + "learning_rate": 0.000306086258776329, + "loss": 1.2957, + "theoretical_loss": 3.451736115763616, + "tokens_seen": 2050621440 + }, + { + "epoch": 0.62, + "learning_rate": 0.0003060702106318957, + "loss": 1.3398, + "theoretical_loss": 3.451727038805114, + "tokens_seen": 2050686976 + }, + { + "epoch": 0.62, + "learning_rate": 0.0003060541624874624, + "loss": 1.2698, + "theoretical_loss": 3.4517179622179093, + "tokens_seen": 2050752512 + }, + { + "epoch": 0.62, + "learning_rate": 0.0003060381143430291, + "loss": 1.3397, + "theoretical_loss": 3.451708886001976, + "tokens_seen": 2050818048 + }, + { + "epoch": 0.62, + "learning_rate": 0.0003060220661985958, + "loss": 1.3462, + "theoretical_loss": 3.451699810157285, + "tokens_seen": 2050883584 + }, + { + "epoch": 0.62, + "learning_rate": 0.0003060060180541625, + "loss": 1.3318, + "theoretical_loss": 3.451690734683811, + "tokens_seen": 2050949120 + }, + { + "epoch": 0.62, + "learning_rate": 0.0003059899699097292, + "loss": 1.2677, + "theoretical_loss": 3.4516816595815256, + "tokens_seen": 2051014656 + }, + { + "epoch": 0.62, + "learning_rate": 0.0003059739217652959, + "loss": 1.3808, + "theoretical_loss": 3.451672584850403, + "tokens_seen": 2051080192 + }, + { + "epoch": 0.62, + "learning_rate": 0.0003059578736208626, + "loss": 1.3181, + "theoretical_loss": 3.451663510490415, + "tokens_seen": 2051145728 + }, + { + "epoch": 0.62, + "learning_rate": 0.0003059418254764293, + "loss": 1.3142, + "theoretical_loss": 3.451654436501536, + "tokens_seen": 2051211264 + }, + { + "epoch": 0.62, + "objective/train/docs_used": 748156, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.2556326389312744, + "objective/train/theoretical_loss": 3.451645362883738, + "objective/train/tokens_used": 2071736800, + "theoretical_loss": 3.451645362883738, + "tokens_seen": 2051276800 + }, + { + "epoch": 0.62, + "learning_rate": 0.000305925777331996, + "loss": 1.3606, + "theoretical_loss": 3.451645362883738, + "tokens_seen": 2051276800 + }, + { + "epoch": 0.62, + "learning_rate": 0.0003059097291875627, + "loss": 1.2988, + "theoretical_loss": 3.451636289636993, + "tokens_seen": 2051342336 + }, + { + "epoch": 0.62, + "learning_rate": 0.0003058936810431294, + "loss": 1.4615, + "theoretical_loss": 3.4516272167612767, + "tokens_seen": 2051407872 + }, + { + "epoch": 0.62, + "learning_rate": 0.0003058776328986961, + "loss": 1.3509, + "theoretical_loss": 3.4516181442565594, + "tokens_seen": 2051473408 + }, + { + "epoch": 0.62, + "learning_rate": 0.0003058615847542628, + "loss": 1.3894, + "theoretical_loss": 3.4516090721228156, + "tokens_seen": 2051538944 + }, + { + "epoch": 0.62, + "learning_rate": 0.0003058455366098295, + "loss": 1.4553, + "theoretical_loss": 3.451600000360017, + "tokens_seen": 2051604480 + }, + { + "epoch": 0.62, + "learning_rate": 0.0003058294884653962, + "loss": 1.41, + "theoretical_loss": 3.4515909289681383, + "tokens_seen": 2051670016 + }, + { + "epoch": 0.62, + "learning_rate": 0.0003058134403209629, + "loss": 1.347, + "theoretical_loss": 3.451581857947151, + "tokens_seen": 2051735552 + }, + { + "epoch": 0.62, + "learning_rate": 0.0003057973921765296, + "loss": 1.2269, + "theoretical_loss": 3.451572787297029, + "tokens_seen": 2051801088 + }, + { + "epoch": 0.62, + "learning_rate": 0.0003057813440320963, + "loss": 1.2921, + "theoretical_loss": 3.451563717017745, + "tokens_seen": 2051866624 + }, + { + "epoch": 0.62, + "learning_rate": 0.000305765295887663, + "loss": 1.2395, + "theoretical_loss": 3.4515546471092717, + "tokens_seen": 2051932160 + }, + { + "epoch": 0.62, + "learning_rate": 0.0003057492477432297, + "loss": 1.2544, + "theoretical_loss": 3.4515455775715824, + "tokens_seen": 2051997696 + }, + { + "epoch": 0.62, + "learning_rate": 0.0003057331995987964, + "loss": 1.4627, + "theoretical_loss": 3.4515365084046503, + "tokens_seen": 2052063232 + }, + { + "epoch": 0.62, + "learning_rate": 0.0003057171514543631, + "loss": 1.4036, + "theoretical_loss": 3.451527439608448, + "tokens_seen": 2052128768 + }, + { + "epoch": 0.62, + "learning_rate": 0.0003057011033099298, + "loss": 1.3223, + "theoretical_loss": 3.4515183711829485, + "tokens_seen": 2052194304 + }, + { + "epoch": 0.62, + "learning_rate": 0.0003056850551654965, + "loss": 1.3024, + "theoretical_loss": 3.451509303128125, + "tokens_seen": 2052259840 + }, + { + "epoch": 0.62, + "learning_rate": 0.0003056690070210632, + "loss": 1.3813, + "theoretical_loss": 3.451500235443951, + "tokens_seen": 2052325376 + }, + { + "epoch": 0.62, + "learning_rate": 0.0003056529588766299, + "loss": 1.3032, + "theoretical_loss": 3.4514911681303984, + "tokens_seen": 2052390912 + }, + { + "epoch": 0.62, + "learning_rate": 0.0003056369107321966, + "loss": 1.3249, + "theoretical_loss": 3.451482101187441, + "tokens_seen": 2052456448 + }, + { + "epoch": 0.62, + "learning_rate": 0.0003056208625877633, + "loss": 1.2838, + "theoretical_loss": 3.4514730346150517, + "tokens_seen": 2052521984 + }, + { + "epoch": 0.62, + "learning_rate": 0.00030560481444333, + "loss": 1.4149, + "theoretical_loss": 3.4514639684132034, + "tokens_seen": 2052587520 + }, + { + "epoch": 0.62, + "learning_rate": 0.0003055887662988967, + "loss": 1.3089, + "theoretical_loss": 3.4514549025818693, + "tokens_seen": 2052653056 + }, + { + "epoch": 0.62, + "learning_rate": 0.0003055727181544634, + "loss": 1.3138, + "theoretical_loss": 3.451445837121022, + "tokens_seen": 2052718592 + }, + { + "epoch": 0.62, + "learning_rate": 0.0003055566700100301, + "loss": 1.2312, + "theoretical_loss": 3.451436772030635, + "tokens_seen": 2052784128 + }, + { + "epoch": 0.62, + "learning_rate": 0.0003055406218655968, + "loss": 1.3111, + "theoretical_loss": 3.451427707310682, + "tokens_seen": 2052849664 + }, + { + "epoch": 0.62, + "learning_rate": 0.0003055245737211635, + "loss": 1.2537, + "theoretical_loss": 3.451418642961134, + "tokens_seen": 2052915200 + }, + { + "epoch": 0.62, + "learning_rate": 0.0003055085255767302, + "loss": 1.3502, + "theoretical_loss": 3.4514095789819654, + "tokens_seen": 2052980736 + }, + { + "epoch": 0.62, + "learning_rate": 0.0003054924774322969, + "loss": 1.3633, + "theoretical_loss": 3.45140051537315, + "tokens_seen": 2053046272 + }, + { + "epoch": 0.62, + "learning_rate": 0.0003054764292878636, + "loss": 1.2858, + "theoretical_loss": 3.451391452134659, + "tokens_seen": 2053111808 + }, + { + "epoch": 0.62, + "learning_rate": 0.0003054603811434303, + "loss": 1.2883, + "theoretical_loss": 3.451382389266467, + "tokens_seen": 2053177344 + }, + { + "epoch": 0.62, + "learning_rate": 0.000305444332998997, + "loss": 1.3351, + "theoretical_loss": 3.4513733267685462, + "tokens_seen": 2053242880 + }, + { + "epoch": 0.62, + "learning_rate": 0.0003054282848545637, + "loss": 1.3504, + "theoretical_loss": 3.45136426464087, + "tokens_seen": 2053308416 + }, + { + "epoch": 0.62, + "learning_rate": 0.0003054122367101304, + "loss": 1.3681, + "theoretical_loss": 3.451355202883411, + "tokens_seen": 2053373952 + }, + { + "epoch": 0.62, + "learning_rate": 0.0003053961885656971, + "loss": 1.3052, + "theoretical_loss": 3.4513461414961433, + "tokens_seen": 2053439488 + }, + { + "epoch": 0.62, + "learning_rate": 0.0003053801404212638, + "loss": 1.3634, + "theoretical_loss": 3.451337080479039, + "tokens_seen": 2053505024 + }, + { + "epoch": 0.62, + "learning_rate": 0.0003053640922768305, + "loss": 1.3583, + "theoretical_loss": 3.4513280198320713, + "tokens_seen": 2053570560 + }, + { + "epoch": 0.62, + "learning_rate": 0.0003053480441323972, + "loss": 1.4127, + "theoretical_loss": 3.4513189595552136, + "tokens_seen": 2053636096 + }, + { + "epoch": 0.62, + "learning_rate": 0.0003053319959879639, + "loss": 1.2961, + "theoretical_loss": 3.4513098996484386, + "tokens_seen": 2053701632 + }, + { + "epoch": 0.62, + "learning_rate": 0.0003053159478435306, + "loss": 1.2637, + "theoretical_loss": 3.4513008401117196, + "tokens_seen": 2053767168 + }, + { + "epoch": 0.62, + "learning_rate": 0.0003052998996990973, + "loss": 1.2625, + "theoretical_loss": 3.45129178094503, + "tokens_seen": 2053832704 + }, + { + "epoch": 0.62, + "learning_rate": 0.000305283851554664, + "loss": 1.3034, + "theoretical_loss": 3.451282722148342, + "tokens_seen": 2053898240 + }, + { + "epoch": 0.62, + "learning_rate": 0.0003052678034102307, + "loss": 1.3338, + "theoretical_loss": 3.4512736637216292, + "tokens_seen": 2053963776 + }, + { + "epoch": 0.62, + "learning_rate": 0.0003052517552657974, + "loss": 1.2917, + "theoretical_loss": 3.451264605664865, + "tokens_seen": 2054029312 + }, + { + "epoch": 0.62, + "learning_rate": 0.0003052357071213641, + "loss": 1.384, + "theoretical_loss": 3.4512555479780223, + "tokens_seen": 2054094848 + }, + { + "epoch": 0.62, + "learning_rate": 0.0003052196589769308, + "loss": 1.3228, + "theoretical_loss": 3.451246490661074, + "tokens_seen": 2054160384 + }, + { + "epoch": 0.62, + "learning_rate": 0.0003052036108324975, + "loss": 1.3207, + "theoretical_loss": 3.451237433713993, + "tokens_seen": 2054225920 + }, + { + "epoch": 0.62, + "learning_rate": 0.0003051875626880642, + "loss": 1.307, + "theoretical_loss": 3.451228377136753, + "tokens_seen": 2054291456 + }, + { + "epoch": 0.62, + "learning_rate": 0.0003051715145436309, + "loss": 1.2647, + "theoretical_loss": 3.4512193209293267, + "tokens_seen": 2054356992 + }, + { + "epoch": 0.62, + "learning_rate": 0.0003051554663991976, + "loss": 1.2878, + "theoretical_loss": 3.451210265091687, + "tokens_seen": 2054422528 + }, + { + "epoch": 0.62, + "learning_rate": 0.0003051394182547643, + "loss": 1.3337, + "theoretical_loss": 3.4512012096238074, + "tokens_seen": 2054488064 + }, + { + "epoch": 0.62, + "objective/train/docs_used": 749325, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.2915825843811035, + "objective/train/theoretical_loss": 3.4511921545256605, + "objective/train/tokens_used": 2075013600, + "theoretical_loss": 3.4511921545256605, + "tokens_seen": 2054553600 + }, + { + "epoch": 0.62, + "learning_rate": 0.000305123370110331, + "loss": 1.3711, + "theoretical_loss": 3.4511921545256605, + "tokens_seen": 2054553600 + }, + { + "epoch": 0.62, + "learning_rate": 0.00030510732196589773, + "loss": 1.332, + "theoretical_loss": 3.4511830997972206, + "tokens_seen": 2054619136 + }, + { + "epoch": 0.62, + "learning_rate": 0.00030509127382146443, + "loss": 1.2836, + "theoretical_loss": 3.451174045438459, + "tokens_seen": 2054684672 + }, + { + "epoch": 0.62, + "learning_rate": 0.00030507522567703113, + "loss": 1.4343, + "theoretical_loss": 3.4511649914493505, + "tokens_seen": 2054750208 + }, + { + "epoch": 0.62, + "learning_rate": 0.00030505917753259783, + "loss": 1.3863, + "theoretical_loss": 3.451155937829868, + "tokens_seen": 2054815744 + }, + { + "epoch": 0.62, + "learning_rate": 0.00030504312938816453, + "loss": 1.2806, + "theoretical_loss": 3.451146884579983, + "tokens_seen": 2054881280 + }, + { + "epoch": 0.62, + "learning_rate": 0.00030502708124373123, + "loss": 1.3893, + "theoretical_loss": 3.451137831699671, + "tokens_seen": 2054946816 + }, + { + "epoch": 0.62, + "learning_rate": 0.00030501103309929793, + "loss": 1.2235, + "theoretical_loss": 3.451128779188903, + "tokens_seen": 2055012352 + }, + { + "epoch": 0.62, + "learning_rate": 0.0003049949849548646, + "loss": 1.378, + "theoretical_loss": 3.4511197270476535, + "tokens_seen": 2055077888 + }, + { + "epoch": 0.62, + "learning_rate": 0.0003049789368104313, + "loss": 1.3225, + "theoretical_loss": 3.4511106752758947, + "tokens_seen": 2055143424 + }, + { + "epoch": 0.62, + "learning_rate": 0.000304962888665998, + "loss": 1.3647, + "theoretical_loss": 3.4511016238736008, + "tokens_seen": 2055208960 + }, + { + "epoch": 0.62, + "learning_rate": 0.0003049468405215647, + "loss": 1.2442, + "theoretical_loss": 3.4510925728407438, + "tokens_seen": 2055274496 + }, + { + "epoch": 0.62, + "learning_rate": 0.0003049307923771314, + "loss": 1.2766, + "theoretical_loss": 3.4510835221772975, + "tokens_seen": 2055340032 + }, + { + "epoch": 0.62, + "learning_rate": 0.0003049147442326981, + "loss": 1.4054, + "theoretical_loss": 3.451074471883236, + "tokens_seen": 2055405568 + }, + { + "epoch": 0.62, + "learning_rate": 0.0003048986960882648, + "loss": 1.3607, + "theoretical_loss": 3.45106542195853, + "tokens_seen": 2055471104 + }, + { + "epoch": 0.62, + "learning_rate": 0.0003048826479438315, + "loss": 1.4301, + "theoretical_loss": 3.4510563724031544, + "tokens_seen": 2055536640 + }, + { + "epoch": 0.62, + "learning_rate": 0.0003048665997993982, + "loss": 1.3182, + "theoretical_loss": 3.4510473232170824, + "tokens_seen": 2055602176 + }, + { + "epoch": 0.62, + "learning_rate": 0.0003048505516549649, + "loss": 1.3446, + "theoretical_loss": 3.451038274400286, + "tokens_seen": 2055667712 + }, + { + "epoch": 0.62, + "learning_rate": 0.0003048345035105316, + "loss": 1.295, + "theoretical_loss": 3.4510292259527398, + "tokens_seen": 2055733248 + }, + { + "epoch": 0.62, + "learning_rate": 0.0003048184553660983, + "loss": 1.3393, + "theoretical_loss": 3.4510201778744163, + "tokens_seen": 2055798784 + }, + { + "epoch": 0.62, + "learning_rate": 0.000304802407221665, + "loss": 1.4096, + "theoretical_loss": 3.451011130165288, + "tokens_seen": 2055864320 + }, + { + "epoch": 0.62, + "learning_rate": 0.0003047863590772317, + "loss": 1.2537, + "theoretical_loss": 3.451002082825329, + "tokens_seen": 2055929856 + }, + { + "epoch": 0.62, + "learning_rate": 0.0003047703109327984, + "loss": 1.3905, + "theoretical_loss": 3.4509930358545127, + "tokens_seen": 2055995392 + }, + { + "epoch": 0.62, + "learning_rate": 0.0003047542627883651, + "loss": 1.3214, + "theoretical_loss": 3.450983989252811, + "tokens_seen": 2056060928 + }, + { + "epoch": 0.62, + "learning_rate": 0.0003047382146439318, + "loss": 1.3765, + "theoretical_loss": 3.4509749430201984, + "tokens_seen": 2056126464 + }, + { + "epoch": 0.62, + "learning_rate": 0.0003047221664994985, + "loss": 1.3524, + "theoretical_loss": 3.450965897156647, + "tokens_seen": 2056192000 + }, + { + "epoch": 0.62, + "learning_rate": 0.0003047061183550652, + "loss": 1.1767, + "theoretical_loss": 3.4509568516621307, + "tokens_seen": 2056257536 + }, + { + "epoch": 0.62, + "learning_rate": 0.0003046900702106319, + "loss": 1.3491, + "theoretical_loss": 3.450947806536622, + "tokens_seen": 2056323072 + }, + { + "epoch": 0.62, + "learning_rate": 0.0003046740220661986, + "loss": 1.3793, + "theoretical_loss": 3.4509387617800953, + "tokens_seen": 2056388608 + }, + { + "epoch": 0.62, + "learning_rate": 0.0003046579739217653, + "loss": 1.3523, + "theoretical_loss": 3.4509297173925226, + "tokens_seen": 2056454144 + }, + { + "epoch": 0.62, + "learning_rate": 0.000304641925777332, + "loss": 1.4106, + "theoretical_loss": 3.4509206733738775, + "tokens_seen": 2056519680 + }, + { + "epoch": 0.62, + "learning_rate": 0.0003046258776328987, + "loss": 1.2816, + "theoretical_loss": 3.4509116297241333, + "tokens_seen": 2056585216 + }, + { + "epoch": 0.62, + "learning_rate": 0.0003046098294884654, + "loss": 1.3733, + "theoretical_loss": 3.450902586443263, + "tokens_seen": 2056650752 + }, + { + "epoch": 0.62, + "learning_rate": 0.0003045937813440321, + "loss": 1.3786, + "theoretical_loss": 3.4508935435312402, + "tokens_seen": 2056716288 + }, + { + "epoch": 0.62, + "learning_rate": 0.0003045777331995988, + "loss": 1.3394, + "theoretical_loss": 3.450884500988038, + "tokens_seen": 2056781824 + }, + { + "epoch": 0.62, + "learning_rate": 0.0003045616850551655, + "loss": 1.3965, + "theoretical_loss": 3.450875458813629, + "tokens_seen": 2056847360 + }, + { + "epoch": 0.62, + "learning_rate": 0.0003045456369107322, + "loss": 1.391, + "theoretical_loss": 3.450866417007987, + "tokens_seen": 2056912896 + }, + { + "epoch": 0.62, + "learning_rate": 0.0003045295887662989, + "loss": 1.4371, + "theoretical_loss": 3.450857375571085, + "tokens_seen": 2056978432 + }, + { + "epoch": 0.62, + "learning_rate": 0.0003045135406218656, + "loss": 1.3459, + "theoretical_loss": 3.450848334502896, + "tokens_seen": 2057043968 + }, + { + "epoch": 0.62, + "learning_rate": 0.00030449749247743235, + "loss": 1.375, + "theoretical_loss": 3.4508392938033943, + "tokens_seen": 2057109504 + }, + { + "epoch": 0.62, + "learning_rate": 0.00030448144433299905, + "loss": 1.3842, + "theoretical_loss": 3.4508302534725512, + "tokens_seen": 2057175040 + }, + { + "epoch": 0.62, + "learning_rate": 0.00030446539618856575, + "loss": 1.3228, + "theoretical_loss": 3.4508212135103418, + "tokens_seen": 2057240576 + }, + { + "epoch": 0.62, + "learning_rate": 0.00030444934804413245, + "loss": 1.2959, + "theoretical_loss": 3.4508121739167383, + "tokens_seen": 2057306112 + }, + { + "epoch": 0.62, + "learning_rate": 0.00030443329989969915, + "loss": 1.4086, + "theoretical_loss": 3.4508031346917143, + "tokens_seen": 2057371648 + }, + { + "epoch": 0.62, + "learning_rate": 0.00030441725175526585, + "loss": 1.3291, + "theoretical_loss": 3.450794095835243, + "tokens_seen": 2057437184 + }, + { + "epoch": 0.62, + "learning_rate": 0.00030440120361083256, + "loss": 1.4129, + "theoretical_loss": 3.4507850573472973, + "tokens_seen": 2057502720 + }, + { + "epoch": 0.62, + "learning_rate": 0.00030438515546639926, + "loss": 1.3181, + "theoretical_loss": 3.4507760192278507, + "tokens_seen": 2057568256 + }, + { + "epoch": 0.62, + "learning_rate": 0.00030436910732196596, + "loss": 1.2927, + "theoretical_loss": 3.4507669814768764, + "tokens_seen": 2057633792 + }, + { + "epoch": 0.62, + "learning_rate": 0.0003043530591775326, + "loss": 1.3495, + "theoretical_loss": 3.450757944094348, + "tokens_seen": 2057699328 + }, + { + "epoch": 0.62, + "learning_rate": 0.0003043370110330993, + "loss": 1.3554, + "theoretical_loss": 3.450748907080238, + "tokens_seen": 2057764864 + }, + { + "epoch": 0.62, + "objective/train/docs_used": 750571, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.4228017330169678, + "objective/train/theoretical_loss": 3.4507398704345205, + "objective/train/tokens_used": 2078290400, + "theoretical_loss": 3.4507398704345205, + "tokens_seen": 2057830400 + }, + { + "epoch": 0.62, + "learning_rate": 0.000304320962888666, + "loss": 1.4172, + "theoretical_loss": 3.4507398704345205, + "tokens_seen": 2057830400 + }, + { + "epoch": 0.62, + "learning_rate": 0.0003043049147442327, + "loss": 1.3472, + "theoretical_loss": 3.4507308341571674, + "tokens_seen": 2057895936 + }, + { + "epoch": 0.62, + "learning_rate": 0.0003042888665997994, + "loss": 1.3485, + "theoretical_loss": 3.4507217982481535, + "tokens_seen": 2057961472 + }, + { + "epoch": 0.62, + "learning_rate": 0.0003042728184553661, + "loss": 1.2327, + "theoretical_loss": 3.450712762707452, + "tokens_seen": 2058027008 + }, + { + "epoch": 0.62, + "learning_rate": 0.0003042567703109328, + "loss": 1.3974, + "theoretical_loss": 3.4507037275350347, + "tokens_seen": 2058092544 + }, + { + "epoch": 0.62, + "learning_rate": 0.0003042407221664995, + "loss": 1.3644, + "theoretical_loss": 3.450694692730876, + "tokens_seen": 2058158080 + }, + { + "epoch": 0.62, + "learning_rate": 0.0003042246740220662, + "loss": 1.3059, + "theoretical_loss": 3.450685658294949, + "tokens_seen": 2058223616 + }, + { + "epoch": 0.62, + "learning_rate": 0.0003042086258776329, + "loss": 1.3592, + "theoretical_loss": 3.450676624227227, + "tokens_seen": 2058289152 + }, + { + "epoch": 0.62, + "learning_rate": 0.0003041925777331996, + "loss": 1.3655, + "theoretical_loss": 3.450667590527683, + "tokens_seen": 2058354688 + }, + { + "epoch": 0.62, + "learning_rate": 0.0003041765295887663, + "loss": 1.3218, + "theoretical_loss": 3.45065855719629, + "tokens_seen": 2058420224 + }, + { + "epoch": 0.62, + "learning_rate": 0.000304160481444333, + "loss": 1.3032, + "theoretical_loss": 3.4506495242330226, + "tokens_seen": 2058485760 + }, + { + "epoch": 0.62, + "learning_rate": 0.0003041444332998997, + "loss": 1.3762, + "theoretical_loss": 3.4506404916378526, + "tokens_seen": 2058551296 + }, + { + "epoch": 0.62, + "learning_rate": 0.0003041283851554664, + "loss": 1.2797, + "theoretical_loss": 3.450631459410754, + "tokens_seen": 2058616832 + }, + { + "epoch": 0.62, + "learning_rate": 0.0003041123370110331, + "loss": 1.4133, + "theoretical_loss": 3.4506224275517, + "tokens_seen": 2058682368 + }, + { + "epoch": 0.62, + "learning_rate": 0.0003040962888665998, + "loss": 1.4011, + "theoretical_loss": 3.4506133960606635, + "tokens_seen": 2058747904 + }, + { + "epoch": 0.62, + "learning_rate": 0.0003040802407221665, + "loss": 1.3833, + "theoretical_loss": 3.450604364937618, + "tokens_seen": 2058813440 + }, + { + "epoch": 0.62, + "learning_rate": 0.0003040641925777332, + "loss": 1.3603, + "theoretical_loss": 3.4505953341825375, + "tokens_seen": 2058878976 + }, + { + "epoch": 0.62, + "learning_rate": 0.0003040481444332999, + "loss": 1.3687, + "theoretical_loss": 3.4505863037953945, + "tokens_seen": 2058944512 + }, + { + "epoch": 0.62, + "learning_rate": 0.0003040320962888666, + "loss": 1.3693, + "theoretical_loss": 3.4505772737761626, + "tokens_seen": 2059010048 + }, + { + "epoch": 0.62, + "learning_rate": 0.0003040160481444333, + "loss": 1.2686, + "theoretical_loss": 3.4505682441248147, + "tokens_seen": 2059075584 + }, + { + "epoch": 0.62, + "learning_rate": 0.000304, + "loss": 1.3484, + "theoretical_loss": 3.4505592148413244, + "tokens_seen": 2059141120 + }, + { + "epoch": 0.62, + "learning_rate": 0.0003039839518555667, + "loss": 1.3314, + "theoretical_loss": 3.4505501859256653, + "tokens_seen": 2059206656 + }, + { + "epoch": 0.62, + "learning_rate": 0.0003039679037111334, + "loss": 1.2611, + "theoretical_loss": 3.45054115737781, + "tokens_seen": 2059272192 + }, + { + "epoch": 0.62, + "learning_rate": 0.0003039518555667001, + "loss": 1.2637, + "theoretical_loss": 3.450532129197733, + "tokens_seen": 2059337728 + }, + { + "epoch": 0.62, + "learning_rate": 0.0003039358074222668, + "loss": 1.3733, + "theoretical_loss": 3.4505231013854063, + "tokens_seen": 2059403264 + }, + { + "epoch": 0.62, + "learning_rate": 0.0003039197592778335, + "loss": 1.3938, + "theoretical_loss": 3.450514073940804, + "tokens_seen": 2059468800 + }, + { + "epoch": 0.62, + "learning_rate": 0.0003039037111334002, + "loss": 1.3577, + "theoretical_loss": 3.450505046863899, + "tokens_seen": 2059534336 + }, + { + "epoch": 0.62, + "learning_rate": 0.0003038876629889669, + "loss": 1.3313, + "theoretical_loss": 3.450496020154665, + "tokens_seen": 2059599872 + }, + { + "epoch": 0.62, + "learning_rate": 0.0003038716148445336, + "loss": 1.436, + "theoretical_loss": 3.450486993813075, + "tokens_seen": 2059665408 + }, + { + "epoch": 0.62, + "learning_rate": 0.0003038555667001003, + "loss": 1.1995, + "theoretical_loss": 3.4504779678391024, + "tokens_seen": 2059730944 + }, + { + "epoch": 0.62, + "learning_rate": 0.000303839518555667, + "loss": 1.4048, + "theoretical_loss": 3.4504689422327206, + "tokens_seen": 2059796480 + }, + { + "epoch": 0.62, + "learning_rate": 0.0003038234704112337, + "loss": 1.3615, + "theoretical_loss": 3.4504599169939034, + "tokens_seen": 2059862016 + }, + { + "epoch": 0.62, + "learning_rate": 0.0003038074222668004, + "loss": 1.3178, + "theoretical_loss": 3.450450892122623, + "tokens_seen": 2059927552 + }, + { + "epoch": 0.62, + "learning_rate": 0.0003037913741223671, + "loss": 1.3366, + "theoretical_loss": 3.4504418676188537, + "tokens_seen": 2059993088 + }, + { + "epoch": 0.62, + "learning_rate": 0.0003037753259779338, + "loss": 1.3213, + "theoretical_loss": 3.4504328434825684, + "tokens_seen": 2060058624 + }, + { + "epoch": 0.62, + "learning_rate": 0.0003037592778335005, + "loss": 1.3006, + "theoretical_loss": 3.450423819713741, + "tokens_seen": 2060124160 + }, + { + "epoch": 0.62, + "learning_rate": 0.0003037432296890672, + "loss": 1.3348, + "theoretical_loss": 3.4504147963123444, + "tokens_seen": 2060189696 + }, + { + "epoch": 0.62, + "learning_rate": 0.00030372718154463393, + "loss": 1.3903, + "theoretical_loss": 3.450405773278352, + "tokens_seen": 2060255232 + }, + { + "epoch": 0.62, + "learning_rate": 0.00030371113340020063, + "loss": 1.4427, + "theoretical_loss": 3.450396750611737, + "tokens_seen": 2060320768 + }, + { + "epoch": 0.62, + "learning_rate": 0.00030369508525576733, + "loss": 1.2076, + "theoretical_loss": 3.4503877283124726, + "tokens_seen": 2060386304 + }, + { + "epoch": 0.62, + "learning_rate": 0.00030367903711133403, + "loss": 1.2921, + "theoretical_loss": 3.4503787063805333, + "tokens_seen": 2060451840 + }, + { + "epoch": 0.62, + "learning_rate": 0.00030366298896690073, + "loss": 1.3631, + "theoretical_loss": 3.450369684815891, + "tokens_seen": 2060517376 + }, + { + "epoch": 0.62, + "learning_rate": 0.00030364694082246743, + "loss": 1.2469, + "theoretical_loss": 3.45036066361852, + "tokens_seen": 2060582912 + }, + { + "epoch": 0.62, + "learning_rate": 0.00030363089267803413, + "loss": 1.3526, + "theoretical_loss": 3.450351642788393, + "tokens_seen": 2060648448 + }, + { + "epoch": 0.62, + "learning_rate": 0.00030361484453360083, + "loss": 1.2893, + "theoretical_loss": 3.4503426223254845, + "tokens_seen": 2060713984 + }, + { + "epoch": 0.62, + "learning_rate": 0.00030359879638916753, + "loss": 1.2261, + "theoretical_loss": 3.450333602229766, + "tokens_seen": 2060779520 + }, + { + "epoch": 0.62, + "learning_rate": 0.0003035827482447342, + "loss": 1.3376, + "theoretical_loss": 3.4503245825012128, + "tokens_seen": 2060845056 + }, + { + "epoch": 0.62, + "learning_rate": 0.0003035667001003009, + "loss": 1.2436, + "theoretical_loss": 3.4503155631397977, + "tokens_seen": 2060910592 + }, + { + "epoch": 0.62, + "learning_rate": 0.0003035506519558676, + "loss": 1.3697, + "theoretical_loss": 3.4503065441454934, + "tokens_seen": 2060976128 + }, + { + "epoch": 0.62, + "learning_rate": 0.0003035346038114343, + "loss": 1.371, + "theoretical_loss": 3.4502975255182733, + "tokens_seen": 2061041664 + }, + { + "epoch": 0.62, + "objective/train/docs_used": 751685, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.3872287273406982, + "objective/train/theoretical_loss": 3.450288507258112, + "objective/train/tokens_used": 2081567200, + "theoretical_loss": 3.450288507258112, + "tokens_seen": 2061107200 + }, + { + "epoch": 0.62, + "learning_rate": 0.000303518555667001, + "loss": 1.3497, + "theoretical_loss": 3.450288507258112, + "tokens_seen": 2061107200 + }, + { + "epoch": 0.62, + "learning_rate": 0.0003035025075225677, + "loss": 1.4091, + "theoretical_loss": 3.4502794893649815, + "tokens_seen": 2061172736 + }, + { + "epoch": 0.62, + "learning_rate": 0.0003034864593781344, + "loss": 1.271, + "theoretical_loss": 3.4502704718388566, + "tokens_seen": 2061238272 + }, + { + "epoch": 0.62, + "learning_rate": 0.0003034704112337011, + "loss": 1.3904, + "theoretical_loss": 3.4502614546797092, + "tokens_seen": 2061303808 + }, + { + "epoch": 0.62, + "learning_rate": 0.0003034543630892678, + "loss": 1.298, + "theoretical_loss": 3.4502524378875137, + "tokens_seen": 2061369344 + }, + { + "epoch": 0.62, + "learning_rate": 0.0003034383149448345, + "loss": 1.3723, + "theoretical_loss": 3.4502434214622433, + "tokens_seen": 2061434880 + }, + { + "epoch": 0.62, + "learning_rate": 0.0003034222668004012, + "loss": 1.3392, + "theoretical_loss": 3.4502344054038714, + "tokens_seen": 2061500416 + }, + { + "epoch": 0.62, + "learning_rate": 0.0003034062186559679, + "loss": 1.2737, + "theoretical_loss": 3.4502253897123705, + "tokens_seen": 2061565952 + }, + { + "epoch": 0.62, + "learning_rate": 0.0003033901705115346, + "loss": 1.4893, + "theoretical_loss": 3.4502163743877157, + "tokens_seen": 2061631488 + }, + { + "epoch": 0.62, + "learning_rate": 0.0003033741223671013, + "loss": 1.3557, + "theoretical_loss": 3.4502073594298794, + "tokens_seen": 2061697024 + }, + { + "epoch": 0.62, + "learning_rate": 0.000303358074222668, + "loss": 1.3415, + "theoretical_loss": 3.450198344838835, + "tokens_seen": 2061762560 + }, + { + "epoch": 0.62, + "learning_rate": 0.0003033420260782347, + "loss": 1.3662, + "theoretical_loss": 3.450189330614556, + "tokens_seen": 2061828096 + }, + { + "epoch": 0.62, + "learning_rate": 0.0003033259779338014, + "loss": 1.3094, + "theoretical_loss": 3.4501803167570158, + "tokens_seen": 2061893632 + }, + { + "epoch": 0.62, + "learning_rate": 0.0003033099297893681, + "loss": 1.4039, + "theoretical_loss": 3.4501713032661883, + "tokens_seen": 2061959168 + }, + { + "epoch": 0.62, + "learning_rate": 0.00030329388164493484, + "loss": 1.3439, + "theoretical_loss": 3.450162290142046, + "tokens_seen": 2062024704 + }, + { + "epoch": 0.62, + "learning_rate": 0.00030327783350050155, + "loss": 1.3914, + "theoretical_loss": 3.4501532773845636, + "tokens_seen": 2062090240 + }, + { + "epoch": 0.62, + "learning_rate": 0.00030326178535606825, + "loss": 1.3383, + "theoretical_loss": 3.4501442649937135, + "tokens_seen": 2062155776 + }, + { + "epoch": 0.62, + "learning_rate": 0.00030324573721163495, + "loss": 1.2616, + "theoretical_loss": 3.4501352529694693, + "tokens_seen": 2062221312 + }, + { + "epoch": 0.62, + "learning_rate": 0.00030322968906720165, + "loss": 1.3133, + "theoretical_loss": 3.450126241311805, + "tokens_seen": 2062286848 + }, + { + "epoch": 0.62, + "learning_rate": 0.00030321364092276835, + "loss": 1.3163, + "theoretical_loss": 3.450117230020693, + "tokens_seen": 2062352384 + }, + { + "epoch": 0.62, + "learning_rate": 0.00030319759277833505, + "loss": 1.345, + "theoretical_loss": 3.4501082190961077, + "tokens_seen": 2062417920 + }, + { + "epoch": 0.62, + "learning_rate": 0.00030318154463390175, + "loss": 1.2908, + "theoretical_loss": 3.450099208538022, + "tokens_seen": 2062483456 + }, + { + "epoch": 0.63, + "learning_rate": 0.00030316549648946845, + "loss": 1.4538, + "theoretical_loss": 3.45009019834641, + "tokens_seen": 2062548992 + }, + { + "epoch": 0.63, + "learning_rate": 0.00030314944834503515, + "loss": 1.3306, + "theoretical_loss": 3.450081188521244, + "tokens_seen": 2062614528 + }, + { + "epoch": 0.63, + "learning_rate": 0.00030313340020060185, + "loss": 1.3386, + "theoretical_loss": 3.4500721790624986, + "tokens_seen": 2062680064 + }, + { + "epoch": 0.63, + "learning_rate": 0.00030311735205616855, + "loss": 1.2956, + "theoretical_loss": 3.4500631699701465, + "tokens_seen": 2062745600 + }, + { + "epoch": 0.63, + "learning_rate": 0.00030310130391173525, + "loss": 1.2878, + "theoretical_loss": 3.450054161244162, + "tokens_seen": 2062811136 + }, + { + "epoch": 0.63, + "learning_rate": 0.00030308525576730195, + "loss": 1.2348, + "theoretical_loss": 3.4500451528845177, + "tokens_seen": 2062876672 + }, + { + "epoch": 0.63, + "learning_rate": 0.00030306920762286865, + "loss": 1.3246, + "theoretical_loss": 3.450036144891188, + "tokens_seen": 2062942208 + }, + { + "epoch": 0.63, + "learning_rate": 0.00030305315947843535, + "loss": 1.3168, + "theoretical_loss": 3.450027137264145, + "tokens_seen": 2063007744 + }, + { + "epoch": 0.63, + "learning_rate": 0.00030303711133400205, + "loss": 1.4141, + "theoretical_loss": 3.450018130003363, + "tokens_seen": 2063073280 + }, + { + "epoch": 0.63, + "learning_rate": 0.00030302106318956876, + "loss": 1.315, + "theoretical_loss": 3.4500091231088157, + "tokens_seen": 2063138816 + }, + { + "epoch": 0.63, + "learning_rate": 0.00030300501504513546, + "loss": 1.4842, + "theoretical_loss": 3.450000116580476, + "tokens_seen": 2063204352 + }, + { + "epoch": 0.63, + "learning_rate": 0.00030298896690070216, + "loss": 1.3539, + "theoretical_loss": 3.449991110418318, + "tokens_seen": 2063269888 + }, + { + "epoch": 0.63, + "learning_rate": 0.00030297291875626886, + "loss": 1.3123, + "theoretical_loss": 3.4499821046223147, + "tokens_seen": 2063335424 + }, + { + "epoch": 0.63, + "learning_rate": 0.0003029568706118355, + "loss": 1.2573, + "theoretical_loss": 3.44997309919244, + "tokens_seen": 2063400960 + }, + { + "epoch": 0.63, + "learning_rate": 0.0003029408224674022, + "loss": 1.4147, + "theoretical_loss": 3.449964094128667, + "tokens_seen": 2063466496 + }, + { + "epoch": 0.63, + "learning_rate": 0.0003029247743229689, + "loss": 1.4247, + "theoretical_loss": 3.4499550894309694, + "tokens_seen": 2063532032 + }, + { + "epoch": 0.63, + "learning_rate": 0.0003029087261785356, + "loss": 1.3625, + "theoretical_loss": 3.44994608509932, + "tokens_seen": 2063597568 + }, + { + "epoch": 0.63, + "learning_rate": 0.0003028926780341023, + "loss": 1.2597, + "theoretical_loss": 3.449937081133694, + "tokens_seen": 2063663104 + }, + { + "epoch": 0.63, + "learning_rate": 0.000302876629889669, + "loss": 1.3989, + "theoretical_loss": 3.4499280775340626, + "tokens_seen": 2063728640 + }, + { + "epoch": 0.63, + "learning_rate": 0.0003028605817452357, + "loss": 1.3519, + "theoretical_loss": 3.449919074300401, + "tokens_seen": 2063794176 + }, + { + "epoch": 0.63, + "learning_rate": 0.0003028445336008024, + "loss": 1.3206, + "theoretical_loss": 3.449910071432682, + "tokens_seen": 2063859712 + }, + { + "epoch": 0.63, + "learning_rate": 0.0003028284854563691, + "loss": 1.2401, + "theoretical_loss": 3.4499010689308798, + "tokens_seen": 2063925248 + }, + { + "epoch": 0.63, + "learning_rate": 0.0003028124373119358, + "loss": 1.324, + "theoretical_loss": 3.449892066794967, + "tokens_seen": 2063990784 + }, + { + "epoch": 0.63, + "learning_rate": 0.0003027963891675025, + "loss": 1.3218, + "theoretical_loss": 3.4498830650249177, + "tokens_seen": 2064056320 + }, + { + "epoch": 0.63, + "learning_rate": 0.0003027803410230692, + "loss": 1.4567, + "theoretical_loss": 3.4498740636207055, + "tokens_seen": 2064121856 + }, + { + "epoch": 0.63, + "learning_rate": 0.0003027642928786359, + "loss": 1.4026, + "theoretical_loss": 3.4498650625823033, + "tokens_seen": 2064187392 + }, + { + "epoch": 0.63, + "learning_rate": 0.0003027482447342026, + "loss": 1.322, + "theoretical_loss": 3.449856061909685, + "tokens_seen": 2064252928 + }, + { + "epoch": 0.63, + "learning_rate": 0.0003027321965897693, + "loss": 1.5006, + "theoretical_loss": 3.4498470616028243, + "tokens_seen": 2064318464 + }, + { + "epoch": 0.63, + "objective/train/docs_used": 752866, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.312256932258606, + "objective/train/theoretical_loss": 3.4498380616616946, + "objective/train/tokens_used": 2084844000, + "theoretical_loss": 3.4498380616616946, + "tokens_seen": 2064384000 + }, + { + "epoch": 0.63, + "learning_rate": 0.000302716148445336, + "loss": 1.4159, + "theoretical_loss": 3.4498380616616946, + "tokens_seen": 2064384000 + }, + { + "epoch": 0.63, + "learning_rate": 0.0003027001003009027, + "loss": 1.3139, + "theoretical_loss": 3.449829062086269, + "tokens_seen": 2064449536 + }, + { + "epoch": 0.63, + "learning_rate": 0.0003026840521564694, + "loss": 1.2993, + "theoretical_loss": 3.4498200628765217, + "tokens_seen": 2064515072 + }, + { + "epoch": 0.63, + "learning_rate": 0.0003026680040120361, + "loss": 1.2721, + "theoretical_loss": 3.449811064032426, + "tokens_seen": 2064580608 + }, + { + "epoch": 0.63, + "learning_rate": 0.0003026519558676028, + "loss": 1.3625, + "theoretical_loss": 3.449802065553955, + "tokens_seen": 2064646144 + }, + { + "epoch": 0.63, + "learning_rate": 0.0003026359077231695, + "loss": 1.4185, + "theoretical_loss": 3.4497930674410826, + "tokens_seen": 2064711680 + }, + { + "epoch": 0.63, + "learning_rate": 0.0003026198595787362, + "loss": 1.3359, + "theoretical_loss": 3.4497840696937825, + "tokens_seen": 2064777216 + }, + { + "epoch": 0.63, + "learning_rate": 0.0003026038114343029, + "loss": 1.2905, + "theoretical_loss": 3.449775072312028, + "tokens_seen": 2064842752 + }, + { + "epoch": 0.63, + "learning_rate": 0.0003025877632898696, + "loss": 1.3156, + "theoretical_loss": 3.449766075295793, + "tokens_seen": 2064908288 + }, + { + "epoch": 0.63, + "learning_rate": 0.0003025717151454363, + "loss": 1.3447, + "theoretical_loss": 3.449757078645051, + "tokens_seen": 2064973824 + }, + { + "epoch": 0.63, + "learning_rate": 0.000302555667001003, + "loss": 1.4253, + "theoretical_loss": 3.4497480823597746, + "tokens_seen": 2065039360 + }, + { + "epoch": 0.63, + "learning_rate": 0.0003025396188565697, + "loss": 1.3948, + "theoretical_loss": 3.449739086439939, + "tokens_seen": 2065104896 + }, + { + "epoch": 0.63, + "learning_rate": 0.0003025235707121364, + "loss": 1.3631, + "theoretical_loss": 3.449730090885516, + "tokens_seen": 2065170432 + }, + { + "epoch": 0.63, + "learning_rate": 0.0003025075225677031, + "loss": 1.399, + "theoretical_loss": 3.4497210956964803, + "tokens_seen": 2065235968 + }, + { + "epoch": 0.63, + "learning_rate": 0.0003024914744232698, + "loss": 1.267, + "theoretical_loss": 3.4497121008728056, + "tokens_seen": 2065301504 + }, + { + "epoch": 0.63, + "learning_rate": 0.0003024754262788365, + "loss": 1.267, + "theoretical_loss": 3.4497031064144648, + "tokens_seen": 2065367040 + }, + { + "epoch": 0.63, + "learning_rate": 0.0003024593781344032, + "loss": 1.2838, + "theoretical_loss": 3.4496941123214313, + "tokens_seen": 2065432576 + }, + { + "epoch": 0.63, + "learning_rate": 0.0003024433299899699, + "loss": 1.2455, + "theoretical_loss": 3.44968511859368, + "tokens_seen": 2065498112 + }, + { + "epoch": 0.63, + "learning_rate": 0.0003024272818455366, + "loss": 1.455, + "theoretical_loss": 3.449676125231183, + "tokens_seen": 2065563648 + }, + { + "epoch": 0.63, + "learning_rate": 0.0003024112337011033, + "loss": 1.4336, + "theoretical_loss": 3.4496671322339143, + "tokens_seen": 2065629184 + }, + { + "epoch": 0.63, + "learning_rate": 0.00030239518555667, + "loss": 1.3135, + "theoretical_loss": 3.4496581396018478, + "tokens_seen": 2065694720 + }, + { + "epoch": 0.63, + "learning_rate": 0.0003023791374122367, + "loss": 1.3673, + "theoretical_loss": 3.4496491473349566, + "tokens_seen": 2065760256 + }, + { + "epoch": 0.63, + "learning_rate": 0.00030236308926780343, + "loss": 1.2907, + "theoretical_loss": 3.4496401554332152, + "tokens_seen": 2065825792 + }, + { + "epoch": 0.63, + "learning_rate": 0.00030234704112337013, + "loss": 1.3698, + "theoretical_loss": 3.449631163896596, + "tokens_seen": 2065891328 + }, + { + "epoch": 0.63, + "learning_rate": 0.00030233099297893683, + "loss": 1.2658, + "theoretical_loss": 3.449622172725074, + "tokens_seen": 2065956864 + }, + { + "epoch": 0.63, + "learning_rate": 0.00030231494483450353, + "loss": 1.2424, + "theoretical_loss": 3.4496131819186213, + "tokens_seen": 2066022400 + }, + { + "epoch": 0.63, + "learning_rate": 0.00030229889669007023, + "loss": 1.3233, + "theoretical_loss": 3.449604191477212, + "tokens_seen": 2066087936 + }, + { + "epoch": 0.63, + "learning_rate": 0.00030228284854563693, + "loss": 1.3795, + "theoretical_loss": 3.4495952014008204, + "tokens_seen": 2066153472 + }, + { + "epoch": 0.63, + "learning_rate": 0.00030226680040120363, + "loss": 1.2622, + "theoretical_loss": 3.44958621168942, + "tokens_seen": 2066219008 + }, + { + "epoch": 0.63, + "learning_rate": 0.00030225075225677033, + "loss": 1.3859, + "theoretical_loss": 3.449577222342983, + "tokens_seen": 2066284544 + }, + { + "epoch": 0.63, + "learning_rate": 0.00030223470411233703, + "loss": 1.3748, + "theoretical_loss": 3.449568233361484, + "tokens_seen": 2066350080 + }, + { + "epoch": 0.63, + "learning_rate": 0.00030221865596790373, + "loss": 1.2621, + "theoretical_loss": 3.449559244744897, + "tokens_seen": 2066415616 + }, + { + "epoch": 0.63, + "learning_rate": 0.00030220260782347043, + "loss": 1.2807, + "theoretical_loss": 3.449550256493195, + "tokens_seen": 2066481152 + }, + { + "epoch": 0.63, + "learning_rate": 0.0003021865596790371, + "loss": 1.3778, + "theoretical_loss": 3.4495412686063522, + "tokens_seen": 2066546688 + }, + { + "epoch": 0.63, + "learning_rate": 0.0003021705115346038, + "loss": 1.4139, + "theoretical_loss": 3.449532281084342, + "tokens_seen": 2066612224 + }, + { + "epoch": 0.63, + "learning_rate": 0.0003021544633901705, + "loss": 1.3426, + "theoretical_loss": 3.449523293927137, + "tokens_seen": 2066677760 + }, + { + "epoch": 0.63, + "learning_rate": 0.0003021384152457372, + "loss": 1.4791, + "theoretical_loss": 3.4495143071347125, + "tokens_seen": 2066743296 + }, + { + "epoch": 0.63, + "learning_rate": 0.0003021223671013039, + "loss": 1.3325, + "theoretical_loss": 3.449505320707041, + "tokens_seen": 2066808832 + }, + { + "epoch": 0.63, + "learning_rate": 0.0003021063189568706, + "loss": 1.3569, + "theoretical_loss": 3.4494963346440968, + "tokens_seen": 2066874368 + }, + { + "epoch": 0.63, + "learning_rate": 0.00030209027081243734, + "loss": 1.318, + "theoretical_loss": 3.449487348945853, + "tokens_seen": 2066939904 + }, + { + "epoch": 0.63, + "learning_rate": 0.00030207422266800404, + "loss": 1.2878, + "theoretical_loss": 3.449478363612283, + "tokens_seen": 2067005440 + }, + { + "epoch": 0.63, + "learning_rate": 0.00030205817452357074, + "loss": 1.3647, + "theoretical_loss": 3.4494693786433612, + "tokens_seen": 2067070976 + }, + { + "epoch": 0.63, + "learning_rate": 0.00030204212637913744, + "loss": 1.3493, + "theoretical_loss": 3.449460394039061, + "tokens_seen": 2067136512 + }, + { + "epoch": 0.63, + "learning_rate": 0.00030202607823470414, + "loss": 1.252, + "theoretical_loss": 3.449451409799356, + "tokens_seen": 2067202048 + }, + { + "epoch": 0.63, + "learning_rate": 0.00030201003009027084, + "loss": 1.3646, + "theoretical_loss": 3.4494424259242193, + "tokens_seen": 2067267584 + }, + { + "epoch": 0.63, + "learning_rate": 0.00030199398194583754, + "loss": 1.3093, + "theoretical_loss": 3.4494334424136257, + "tokens_seen": 2067333120 + }, + { + "epoch": 0.63, + "learning_rate": 0.00030197793380140424, + "loss": 1.4151, + "theoretical_loss": 3.4494244592675476, + "tokens_seen": 2067398656 + }, + { + "epoch": 0.63, + "learning_rate": 0.00030196188565697094, + "loss": 1.3514, + "theoretical_loss": 3.4494154764859593, + "tokens_seen": 2067464192 + }, + { + "epoch": 0.63, + "learning_rate": 0.00030194583751253764, + "loss": 1.393, + "theoretical_loss": 3.449406494068835, + "tokens_seen": 2067529728 + }, + { + "epoch": 0.63, + "learning_rate": 0.00030192978936810434, + "loss": 1.4221, + "theoretical_loss": 3.449397512016147, + "tokens_seen": 2067595264 + }, + { + "epoch": 0.63, + "objective/train/docs_used": 753921, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.4945449829101562, + "objective/train/theoretical_loss": 3.4493885303278704, + "objective/train/tokens_used": 2088120800, + "theoretical_loss": 3.4493885303278704, + "tokens_seen": 2067660800 + }, + { + "epoch": 0.63, + "learning_rate": 0.00030191374122367105, + "loss": 1.4033, + "theoretical_loss": 3.4493885303278704, + "tokens_seen": 2067660800 + }, + { + "epoch": 0.63, + "learning_rate": 0.00030189769307923775, + "loss": 1.3908, + "theoretical_loss": 3.449379549003978, + "tokens_seen": 2067726336 + }, + { + "epoch": 0.63, + "learning_rate": 0.00030188164493480445, + "loss": 1.4069, + "theoretical_loss": 3.4493705680444435, + "tokens_seen": 2067791872 + }, + { + "epoch": 0.63, + "learning_rate": 0.00030186559679037115, + "loss": 1.4728, + "theoretical_loss": 3.449361587449241, + "tokens_seen": 2067857408 + }, + { + "epoch": 0.63, + "learning_rate": 0.00030184954864593785, + "loss": 1.3575, + "theoretical_loss": 3.4493526072183434, + "tokens_seen": 2067922944 + }, + { + "epoch": 0.63, + "learning_rate": 0.00030183350050150455, + "loss": 1.4244, + "theoretical_loss": 3.4493436273517255, + "tokens_seen": 2067988480 + }, + { + "epoch": 0.63, + "learning_rate": 0.00030181745235707125, + "loss": 1.2756, + "theoretical_loss": 3.4493346478493603, + "tokens_seen": 2068054016 + }, + { + "epoch": 0.63, + "learning_rate": 0.00030180140421263795, + "loss": 1.3367, + "theoretical_loss": 3.4493256687112215, + "tokens_seen": 2068119552 + }, + { + "epoch": 0.63, + "learning_rate": 0.00030178535606820465, + "loss": 1.3444, + "theoretical_loss": 3.449316689937282, + "tokens_seen": 2068185088 + }, + { + "epoch": 0.63, + "learning_rate": 0.00030176930792377135, + "loss": 1.3626, + "theoretical_loss": 3.4493077115275175, + "tokens_seen": 2068250624 + }, + { + "epoch": 0.63, + "learning_rate": 0.00030175325977933805, + "loss": 1.2831, + "theoretical_loss": 3.4492987334819, + "tokens_seen": 2068316160 + }, + { + "epoch": 0.63, + "learning_rate": 0.00030173721163490475, + "loss": 1.3685, + "theoretical_loss": 3.449289755800404, + "tokens_seen": 2068381696 + }, + { + "epoch": 0.63, + "learning_rate": 0.00030172116349047145, + "loss": 1.2187, + "theoretical_loss": 3.4492807784830024, + "tokens_seen": 2068447232 + }, + { + "epoch": 0.63, + "learning_rate": 0.00030170511534603815, + "loss": 1.3361, + "theoretical_loss": 3.4492718015296697, + "tokens_seen": 2068512768 + }, + { + "epoch": 0.63, + "learning_rate": 0.00030168906720160485, + "loss": 1.3416, + "theoretical_loss": 3.449262824940379, + "tokens_seen": 2068578304 + }, + { + "epoch": 0.63, + "learning_rate": 0.00030167301905717155, + "loss": 1.4203, + "theoretical_loss": 3.449253848715105, + "tokens_seen": 2068643840 + }, + { + "epoch": 0.63, + "learning_rate": 0.00030165697091273826, + "loss": 1.2832, + "theoretical_loss": 3.4492448728538196, + "tokens_seen": 2068709376 + }, + { + "epoch": 0.63, + "learning_rate": 0.00030164092276830496, + "loss": 1.3589, + "theoretical_loss": 3.4492358973564987, + "tokens_seen": 2068774912 + }, + { + "epoch": 0.63, + "learning_rate": 0.00030162487462387166, + "loss": 1.3416, + "theoretical_loss": 3.4492269222231147, + "tokens_seen": 2068840448 + }, + { + "epoch": 0.63, + "learning_rate": 0.00030160882647943836, + "loss": 1.3461, + "theoretical_loss": 3.4492179474536413, + "tokens_seen": 2068905984 + }, + { + "epoch": 0.63, + "learning_rate": 0.00030159277833500506, + "loss": 1.3647, + "theoretical_loss": 3.4492089730480524, + "tokens_seen": 2068971520 + }, + { + "epoch": 0.63, + "learning_rate": 0.00030157673019057176, + "loss": 1.3841, + "theoretical_loss": 3.4491999990063222, + "tokens_seen": 2069037056 + }, + { + "epoch": 0.63, + "learning_rate": 0.00030156068204613846, + "loss": 1.3643, + "theoretical_loss": 3.4491910253284237, + "tokens_seen": 2069102592 + }, + { + "epoch": 0.63, + "learning_rate": 0.0003015446339017051, + "loss": 1.2341, + "theoretical_loss": 3.449182052014331, + "tokens_seen": 2069168128 + }, + { + "epoch": 0.63, + "learning_rate": 0.0003015285857572718, + "loss": 1.3643, + "theoretical_loss": 3.4491730790640176, + "tokens_seen": 2069233664 + }, + { + "epoch": 0.63, + "learning_rate": 0.0003015125376128385, + "loss": 1.4422, + "theoretical_loss": 3.4491641064774576, + "tokens_seen": 2069299200 + }, + { + "epoch": 0.63, + "learning_rate": 0.0003014964894684052, + "loss": 1.3937, + "theoretical_loss": 3.449155134254625, + "tokens_seen": 2069364736 + }, + { + "epoch": 0.63, + "learning_rate": 0.0003014804413239719, + "loss": 1.3414, + "theoretical_loss": 3.4491461623954924, + "tokens_seen": 2069430272 + }, + { + "epoch": 0.63, + "learning_rate": 0.0003014643931795386, + "loss": 1.3151, + "theoretical_loss": 3.449137190900035, + "tokens_seen": 2069495808 + }, + { + "epoch": 0.63, + "learning_rate": 0.0003014483450351053, + "loss": 1.3336, + "theoretical_loss": 3.449128219768225, + "tokens_seen": 2069561344 + }, + { + "epoch": 0.63, + "learning_rate": 0.000301432296890672, + "loss": 1.3648, + "theoretical_loss": 3.449119249000037, + "tokens_seen": 2069626880 + }, + { + "epoch": 0.63, + "learning_rate": 0.0003014162487462387, + "loss": 1.3869, + "theoretical_loss": 3.4491102785954446, + "tokens_seen": 2069692416 + }, + { + "epoch": 0.63, + "learning_rate": 0.0003014002006018054, + "loss": 1.3417, + "theoretical_loss": 3.4491013085544218, + "tokens_seen": 2069757952 + }, + { + "epoch": 0.63, + "learning_rate": 0.0003013841524573721, + "loss": 1.3728, + "theoretical_loss": 3.449092338876942, + "tokens_seen": 2069823488 + }, + { + "epoch": 0.63, + "learning_rate": 0.0003013681043129388, + "loss": 1.2129, + "theoretical_loss": 3.4490833695629792, + "tokens_seen": 2069889024 + }, + { + "epoch": 0.63, + "learning_rate": 0.0003013520561685055, + "loss": 1.3518, + "theoretical_loss": 3.449074400612507, + "tokens_seen": 2069954560 + }, + { + "epoch": 0.63, + "learning_rate": 0.0003013360080240722, + "loss": 1.3492, + "theoretical_loss": 3.449065432025499, + "tokens_seen": 2070020096 + }, + { + "epoch": 0.63, + "learning_rate": 0.0003013199598796389, + "loss": 1.2381, + "theoretical_loss": 3.44905646380193, + "tokens_seen": 2070085632 + }, + { + "epoch": 0.63, + "learning_rate": 0.0003013039117352056, + "loss": 1.3693, + "theoretical_loss": 3.4490474959417723, + "tokens_seen": 2070151168 + }, + { + "epoch": 0.63, + "learning_rate": 0.0003012878635907723, + "loss": 1.3975, + "theoretical_loss": 3.4490385284450005, + "tokens_seen": 2070216704 + }, + { + "epoch": 0.63, + "learning_rate": 0.000301271815446339, + "loss": 1.3696, + "theoretical_loss": 3.4490295613115878, + "tokens_seen": 2070282240 + }, + { + "epoch": 0.63, + "learning_rate": 0.0003012557673019057, + "loss": 1.4032, + "theoretical_loss": 3.449020594541509, + "tokens_seen": 2070347776 + }, + { + "epoch": 0.63, + "learning_rate": 0.0003012397191574724, + "loss": 1.3812, + "theoretical_loss": 3.4490116281347367, + "tokens_seen": 2070413312 + }, + { + "epoch": 0.63, + "learning_rate": 0.0003012236710130391, + "loss": 1.2741, + "theoretical_loss": 3.449002662091246, + "tokens_seen": 2070478848 + }, + { + "epoch": 0.63, + "learning_rate": 0.0003012076228686058, + "loss": 1.2549, + "theoretical_loss": 3.4489936964110095, + "tokens_seen": 2070544384 + }, + { + "epoch": 0.63, + "learning_rate": 0.0003011915747241725, + "loss": 1.4449, + "theoretical_loss": 3.448984731094001, + "tokens_seen": 2070609920 + }, + { + "epoch": 0.63, + "learning_rate": 0.0003011755265797392, + "loss": 1.324, + "theoretical_loss": 3.448975766140195, + "tokens_seen": 2070675456 + }, + { + "epoch": 0.63, + "learning_rate": 0.0003011594784353059, + "loss": 1.3459, + "theoretical_loss": 3.448966801549565, + "tokens_seen": 2070740992 + }, + { + "epoch": 0.63, + "learning_rate": 0.0003011434302908726, + "loss": 1.3285, + "theoretical_loss": 3.448957837322085, + "tokens_seen": 2070806528 + }, + { + "epoch": 0.63, + "learning_rate": 0.0003011273821464393, + "loss": 1.395, + "theoretical_loss": 3.448948873457728, + "tokens_seen": 2070872064 + }, + { + "epoch": 0.63, + "objective/train/docs_used": 755047, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.2775264978408813, + "objective/train/theoretical_loss": 3.4489399099564695, + "objective/train/tokens_used": 2091397600, + "theoretical_loss": 3.4489399099564695, + "tokens_seen": 2070937600 + }, + { + "epoch": 0.63, + "learning_rate": 0.000301111334002006, + "loss": 1.3046, + "theoretical_loss": 3.4489399099564695, + "tokens_seen": 2070937600 + }, + { + "epoch": 0.63, + "learning_rate": 0.0003010952858575727, + "loss": 1.3894, + "theoretical_loss": 3.4489309468182814, + "tokens_seen": 2071003136 + }, + { + "epoch": 0.63, + "learning_rate": 0.0003010792377131394, + "loss": 1.3746, + "theoretical_loss": 3.448921984043138, + "tokens_seen": 2071068672 + }, + { + "epoch": 0.63, + "learning_rate": 0.0003010631895687061, + "loss": 1.4446, + "theoretical_loss": 3.4489130216310135, + "tokens_seen": 2071134208 + }, + { + "epoch": 0.63, + "learning_rate": 0.0003010471414242728, + "loss": 1.3471, + "theoretical_loss": 3.4489040595818823, + "tokens_seen": 2071199744 + }, + { + "epoch": 0.63, + "learning_rate": 0.0003010310932798395, + "loss": 1.3704, + "theoretical_loss": 3.4488950978957167, + "tokens_seen": 2071265280 + }, + { + "epoch": 0.63, + "learning_rate": 0.0003010150451354062, + "loss": 1.3369, + "theoretical_loss": 3.448886136572492, + "tokens_seen": 2071330816 + }, + { + "epoch": 0.63, + "learning_rate": 0.0003009989969909729, + "loss": 1.3504, + "theoretical_loss": 3.448877175612181, + "tokens_seen": 2071396352 + }, + { + "epoch": 0.63, + "learning_rate": 0.00030098294884653963, + "loss": 1.291, + "theoretical_loss": 3.448868215014758, + "tokens_seen": 2071461888 + }, + { + "epoch": 0.63, + "learning_rate": 0.00030096690070210633, + "loss": 1.3046, + "theoretical_loss": 3.4488592547801966, + "tokens_seen": 2071527424 + }, + { + "epoch": 0.63, + "learning_rate": 0.00030095085255767303, + "loss": 1.3493, + "theoretical_loss": 3.4488502949084703, + "tokens_seen": 2071592960 + }, + { + "epoch": 0.63, + "learning_rate": 0.00030093480441323973, + "loss": 1.2876, + "theoretical_loss": 3.448841335399554, + "tokens_seen": 2071658496 + }, + { + "epoch": 0.63, + "learning_rate": 0.00030091875626880643, + "loss": 1.4392, + "theoretical_loss": 3.4488323762534208, + "tokens_seen": 2071724032 + }, + { + "epoch": 0.63, + "learning_rate": 0.00030090270812437313, + "loss": 1.3857, + "theoretical_loss": 3.4488234174700443, + "tokens_seen": 2071789568 + }, + { + "epoch": 0.63, + "learning_rate": 0.00030088665997993983, + "loss": 1.3628, + "theoretical_loss": 3.448814459049399, + "tokens_seen": 2071855104 + }, + { + "epoch": 0.63, + "learning_rate": 0.00030087061183550653, + "loss": 1.416, + "theoretical_loss": 3.448805500991458, + "tokens_seen": 2071920640 + }, + { + "epoch": 0.63, + "learning_rate": 0.00030085456369107323, + "loss": 1.423, + "theoretical_loss": 3.4487965432961962, + "tokens_seen": 2071986176 + }, + { + "epoch": 0.63, + "learning_rate": 0.00030083851554663993, + "loss": 1.2912, + "theoretical_loss": 3.4487875859635864, + "tokens_seen": 2072051712 + }, + { + "epoch": 0.63, + "learning_rate": 0.00030082246740220663, + "loss": 1.4277, + "theoretical_loss": 3.4487786289936024, + "tokens_seen": 2072117248 + }, + { + "epoch": 0.63, + "learning_rate": 0.00030080641925777333, + "loss": 1.3887, + "theoretical_loss": 3.448769672386219, + "tokens_seen": 2072182784 + }, + { + "epoch": 0.63, + "learning_rate": 0.00030079037111334004, + "loss": 1.2665, + "theoretical_loss": 3.4487607161414093, + "tokens_seen": 2072248320 + }, + { + "epoch": 0.63, + "learning_rate": 0.00030077432296890674, + "loss": 1.2719, + "theoretical_loss": 3.4487517602591478, + "tokens_seen": 2072313856 + }, + { + "epoch": 0.63, + "learning_rate": 0.00030075827482447344, + "loss": 1.3369, + "theoretical_loss": 3.4487428047394073, + "tokens_seen": 2072379392 + }, + { + "epoch": 0.63, + "learning_rate": 0.00030074222668004014, + "loss": 1.3809, + "theoretical_loss": 3.4487338495821627, + "tokens_seen": 2072444928 + }, + { + "epoch": 0.63, + "learning_rate": 0.00030072617853560684, + "loss": 1.3048, + "theoretical_loss": 3.4487248947873876, + "tokens_seen": 2072510464 + }, + { + "epoch": 0.63, + "learning_rate": 0.00030071013039117354, + "loss": 1.3847, + "theoretical_loss": 3.4487159403550556, + "tokens_seen": 2072576000 + }, + { + "epoch": 0.63, + "learning_rate": 0.00030069408224674024, + "loss": 1.228, + "theoretical_loss": 3.4487069862851407, + "tokens_seen": 2072641536 + }, + { + "epoch": 0.63, + "learning_rate": 0.00030067803410230694, + "loss": 1.3031, + "theoretical_loss": 3.448698032577617, + "tokens_seen": 2072707072 + }, + { + "epoch": 0.63, + "learning_rate": 0.00030066198595787364, + "loss": 1.388, + "theoretical_loss": 3.448689079232458, + "tokens_seen": 2072772608 + }, + { + "epoch": 0.63, + "learning_rate": 0.00030064593781344034, + "loss": 1.2783, + "theoretical_loss": 3.448680126249638, + "tokens_seen": 2072838144 + }, + { + "epoch": 0.63, + "learning_rate": 0.00030062988966900704, + "loss": 1.2925, + "theoretical_loss": 3.44867117362913, + "tokens_seen": 2072903680 + }, + { + "epoch": 0.63, + "learning_rate": 0.00030061384152457374, + "loss": 1.3019, + "theoretical_loss": 3.448662221370909, + "tokens_seen": 2072969216 + }, + { + "epoch": 0.63, + "learning_rate": 0.00030059779338014044, + "loss": 1.319, + "theoretical_loss": 3.4486532694749483, + "tokens_seen": 2073034752 + }, + { + "epoch": 0.63, + "learning_rate": 0.00030058174523570714, + "loss": 1.481, + "theoretical_loss": 3.4486443179412216, + "tokens_seen": 2073100288 + }, + { + "epoch": 0.63, + "learning_rate": 0.00030056569709127384, + "loss": 1.328, + "theoretical_loss": 3.4486353667697034, + "tokens_seen": 2073165824 + }, + { + "epoch": 0.63, + "learning_rate": 0.00030054964894684054, + "loss": 1.3278, + "theoretical_loss": 3.4486264159603675, + "tokens_seen": 2073231360 + }, + { + "epoch": 0.63, + "learning_rate": 0.00030053360080240725, + "loss": 1.4083, + "theoretical_loss": 3.448617465513187, + "tokens_seen": 2073296896 + }, + { + "epoch": 0.63, + "learning_rate": 0.00030051755265797395, + "loss": 1.3404, + "theoretical_loss": 3.4486085154281367, + "tokens_seen": 2073362432 + }, + { + "epoch": 0.63, + "learning_rate": 0.00030050150451354065, + "loss": 1.3507, + "theoretical_loss": 3.4485995657051904, + "tokens_seen": 2073427968 + }, + { + "epoch": 0.63, + "learning_rate": 0.00030048545636910735, + "loss": 1.3343, + "theoretical_loss": 3.4485906163443216, + "tokens_seen": 2073493504 + }, + { + "epoch": 0.63, + "learning_rate": 0.00030046940822467405, + "loss": 1.383, + "theoretical_loss": 3.448581667345504, + "tokens_seen": 2073559040 + }, + { + "epoch": 0.63, + "learning_rate": 0.00030045336008024075, + "loss": 1.5243, + "theoretical_loss": 3.4485727187087125, + "tokens_seen": 2073624576 + }, + { + "epoch": 0.63, + "learning_rate": 0.00030043731193580745, + "loss": 1.2901, + "theoretical_loss": 3.4485637704339203, + "tokens_seen": 2073690112 + }, + { + "epoch": 0.63, + "learning_rate": 0.00030042126379137415, + "loss": 1.2986, + "theoretical_loss": 3.4485548225211007, + "tokens_seen": 2073755648 + }, + { + "epoch": 0.63, + "learning_rate": 0.00030040521564694085, + "loss": 1.2905, + "theoretical_loss": 3.448545874970229, + "tokens_seen": 2073821184 + }, + { + "epoch": 0.63, + "learning_rate": 0.00030038916750250755, + "loss": 1.4499, + "theoretical_loss": 3.4485369277812783, + "tokens_seen": 2073886720 + }, + { + "epoch": 0.63, + "learning_rate": 0.00030037311935807425, + "loss": 1.3634, + "theoretical_loss": 3.448527980954223, + "tokens_seen": 2073952256 + }, + { + "epoch": 0.63, + "learning_rate": 0.00030035707121364095, + "loss": 1.2488, + "theoretical_loss": 3.4485190344890366, + "tokens_seen": 2074017792 + }, + { + "epoch": 0.63, + "learning_rate": 0.00030034102306920765, + "loss": 1.4097, + "theoretical_loss": 3.448510088385693, + "tokens_seen": 2074083328 + }, + { + "epoch": 0.63, + "learning_rate": 0.00030032497492477435, + "loss": 1.3345, + "theoretical_loss": 3.448501142644166, + "tokens_seen": 2074148864 + }, + { + "epoch": 0.63, + "objective/train/docs_used": 756110, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.3042948246002197, + "objective/train/theoretical_loss": 3.4484921972644305, + "objective/train/tokens_used": 2094674400, + "theoretical_loss": 3.4484921972644305, + "tokens_seen": 2074214400 + }, + { + "epoch": 0.63, + "learning_rate": 0.00030030892678034105, + "loss": 1.3737, + "theoretical_loss": 3.4484921972644305, + "tokens_seen": 2074214400 + }, + { + "epoch": 0.63, + "learning_rate": 0.00030029287863590775, + "loss": 1.2446, + "theoretical_loss": 3.4484832522464597, + "tokens_seen": 2074279936 + }, + { + "epoch": 0.63, + "learning_rate": 0.00030027683049147446, + "loss": 1.316, + "theoretical_loss": 3.448474307590227, + "tokens_seen": 2074345472 + }, + { + "epoch": 0.63, + "learning_rate": 0.00030026078234704116, + "loss": 1.32, + "theoretical_loss": 3.448465363295707, + "tokens_seen": 2074411008 + }, + { + "epoch": 0.63, + "learning_rate": 0.00030024473420260786, + "loss": 1.26, + "theoretical_loss": 3.4484564193628744, + "tokens_seen": 2074476544 + }, + { + "epoch": 0.63, + "learning_rate": 0.00030022868605817456, + "loss": 1.297, + "theoretical_loss": 3.4484474757917014, + "tokens_seen": 2074542080 + }, + { + "epoch": 0.63, + "learning_rate": 0.00030021263791374126, + "loss": 1.2439, + "theoretical_loss": 3.4484385325821636, + "tokens_seen": 2074607616 + }, + { + "epoch": 0.63, + "learning_rate": 0.00030019658976930796, + "loss": 1.3138, + "theoretical_loss": 3.4484295897342334, + "tokens_seen": 2074673152 + }, + { + "epoch": 0.63, + "learning_rate": 0.00030018054162487466, + "loss": 1.272, + "theoretical_loss": 3.448420647247886, + "tokens_seen": 2074738688 + }, + { + "epoch": 0.63, + "learning_rate": 0.00030016449348044136, + "loss": 1.4179, + "theoretical_loss": 3.448411705123095, + "tokens_seen": 2074804224 + }, + { + "epoch": 0.63, + "learning_rate": 0.000300148445336008, + "loss": 1.4949, + "theoretical_loss": 3.4484027633598346, + "tokens_seen": 2074869760 + }, + { + "epoch": 0.63, + "learning_rate": 0.0003001323971915747, + "loss": 1.3908, + "theoretical_loss": 3.4483938219580783, + "tokens_seen": 2074935296 + }, + { + "epoch": 0.63, + "learning_rate": 0.0003001163490471414, + "loss": 1.3107, + "theoretical_loss": 3.4483848809178, + "tokens_seen": 2075000832 + }, + { + "epoch": 0.63, + "learning_rate": 0.0003001003009027081, + "loss": 1.3536, + "theoretical_loss": 3.448375940238974, + "tokens_seen": 2075066368 + }, + { + "epoch": 0.63, + "learning_rate": 0.0003000842527582748, + "loss": 1.3385, + "theoretical_loss": 3.448366999921575, + "tokens_seen": 2075131904 + }, + { + "epoch": 0.63, + "learning_rate": 0.0003000682046138415, + "loss": 1.2879, + "theoretical_loss": 3.448358059965575, + "tokens_seen": 2075197440 + }, + { + "epoch": 0.63, + "learning_rate": 0.0003000521564694082, + "loss": 1.3537, + "theoretical_loss": 3.44834912037095, + "tokens_seen": 2075262976 + }, + { + "epoch": 0.63, + "learning_rate": 0.0003000361083249749, + "loss": 1.3082, + "theoretical_loss": 3.4483401811376724, + "tokens_seen": 2075328512 + }, + { + "epoch": 0.63, + "learning_rate": 0.0003000200601805416, + "loss": 1.331, + "theoretical_loss": 3.4483312422657173, + "tokens_seen": 2075394048 + }, + { + "epoch": 0.63, + "learning_rate": 0.0003000040120361083, + "loss": 1.363, + "theoretical_loss": 3.4483223037550585, + "tokens_seen": 2075459584 + }, + { + "epoch": 0.63, + "learning_rate": 0.000299987963891675, + "loss": 1.3889, + "theoretical_loss": 3.4483133656056695, + "tokens_seen": 2075525120 + }, + { + "epoch": 0.63, + "learning_rate": 0.0002999719157472417, + "loss": 1.3987, + "theoretical_loss": 3.4483044278175248, + "tokens_seen": 2075590656 + }, + { + "epoch": 0.63, + "learning_rate": 0.0002999558676028084, + "loss": 1.3102, + "theoretical_loss": 3.448295490390598, + "tokens_seen": 2075656192 + }, + { + "epoch": 0.63, + "learning_rate": 0.0002999398194583751, + "loss": 1.4296, + "theoretical_loss": 3.448286553324863, + "tokens_seen": 2075721728 + }, + { + "epoch": 0.63, + "learning_rate": 0.0002999237713139418, + "loss": 1.3725, + "theoretical_loss": 3.4482776166202944, + "tokens_seen": 2075787264 + }, + { + "epoch": 0.63, + "learning_rate": 0.0002999077231695085, + "loss": 1.2878, + "theoretical_loss": 3.4482686802768656, + "tokens_seen": 2075852800 + }, + { + "epoch": 0.63, + "learning_rate": 0.0002998916750250752, + "loss": 1.4938, + "theoretical_loss": 3.4482597442945515, + "tokens_seen": 2075918336 + }, + { + "epoch": 0.63, + "learning_rate": 0.0002998756268806419, + "loss": 1.326, + "theoretical_loss": 3.4482508086733246, + "tokens_seen": 2075983872 + }, + { + "epoch": 0.63, + "learning_rate": 0.0002998595787362086, + "loss": 1.34, + "theoretical_loss": 3.448241873413161, + "tokens_seen": 2076049408 + }, + { + "epoch": 0.63, + "learning_rate": 0.0002998435305917753, + "loss": 1.3423, + "theoretical_loss": 3.4482329385140327, + "tokens_seen": 2076114944 + }, + { + "epoch": 0.63, + "learning_rate": 0.000299827482447342, + "loss": 1.3502, + "theoretical_loss": 3.4482240039759144, + "tokens_seen": 2076180480 + }, + { + "epoch": 0.63, + "learning_rate": 0.0002998114343029087, + "loss": 1.2996, + "theoretical_loss": 3.4482150697987803, + "tokens_seen": 2076246016 + }, + { + "epoch": 0.63, + "learning_rate": 0.0002997953861584754, + "loss": 1.4046, + "theoretical_loss": 3.4482061359826046, + "tokens_seen": 2076311552 + }, + { + "epoch": 0.63, + "learning_rate": 0.0002997793380140421, + "loss": 1.3429, + "theoretical_loss": 3.448197202527361, + "tokens_seen": 2076377088 + }, + { + "epoch": 0.63, + "learning_rate": 0.0002997632898696088, + "loss": 1.2929, + "theoretical_loss": 3.4481882694330235, + "tokens_seen": 2076442624 + }, + { + "epoch": 0.63, + "learning_rate": 0.0002997472417251755, + "loss": 1.2521, + "theoretical_loss": 3.4481793366995666, + "tokens_seen": 2076508160 + }, + { + "epoch": 0.63, + "learning_rate": 0.0002997311935807422, + "loss": 1.4149, + "theoretical_loss": 3.448170404326964, + "tokens_seen": 2076573696 + }, + { + "epoch": 0.63, + "learning_rate": 0.0002997151454363089, + "loss": 1.2985, + "theoretical_loss": 3.4481614723151894, + "tokens_seen": 2076639232 + }, + { + "epoch": 0.63, + "learning_rate": 0.0002996990972918757, + "loss": 1.3, + "theoretical_loss": 3.448152540664217, + "tokens_seen": 2076704768 + }, + { + "epoch": 0.63, + "learning_rate": 0.0002996830491474424, + "loss": 1.2455, + "theoretical_loss": 3.4481436093740214, + "tokens_seen": 2076770304 + }, + { + "epoch": 0.63, + "learning_rate": 0.0002996670010030091, + "loss": 1.2128, + "theoretical_loss": 3.448134678444576, + "tokens_seen": 2076835840 + }, + { + "epoch": 0.63, + "learning_rate": 0.0002996509528585758, + "loss": 1.3818, + "theoretical_loss": 3.4481257478758556, + "tokens_seen": 2076901376 + }, + { + "epoch": 0.63, + "learning_rate": 0.0002996349047141425, + "loss": 1.2078, + "theoretical_loss": 3.4481168176678327, + "tokens_seen": 2076966912 + }, + { + "epoch": 0.63, + "learning_rate": 0.0002996188565697092, + "loss": 1.297, + "theoretical_loss": 3.448107887820483, + "tokens_seen": 2077032448 + }, + { + "epoch": 0.63, + "learning_rate": 0.0002996028084252759, + "loss": 1.3188, + "theoretical_loss": 3.44809895833378, + "tokens_seen": 2077097984 + }, + { + "epoch": 0.63, + "learning_rate": 0.0002995867602808426, + "loss": 1.3185, + "theoretical_loss": 3.4480900292076977, + "tokens_seen": 2077163520 + }, + { + "epoch": 0.63, + "learning_rate": 0.0002995707121364093, + "loss": 1.3122, + "theoretical_loss": 3.4480811004422103, + "tokens_seen": 2077229056 + }, + { + "epoch": 0.63, + "learning_rate": 0.000299554663991976, + "loss": 1.3954, + "theoretical_loss": 3.4480721720372913, + "tokens_seen": 2077294592 + }, + { + "epoch": 0.63, + "learning_rate": 0.0002995386158475427, + "loss": 1.1665, + "theoretical_loss": 3.448063243992915, + "tokens_seen": 2077360128 + }, + { + "epoch": 0.63, + "learning_rate": 0.00029952256770310933, + "loss": 1.3376, + "theoretical_loss": 3.4480543163090562, + "tokens_seen": 2077425664 + }, + { + "epoch": 0.63, + "objective/train/docs_used": 757304, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.4160685539245605, + "objective/train/theoretical_loss": 3.4480453889856886, + "objective/train/tokens_used": 2097951200, + "theoretical_loss": 3.4480453889856886, + "tokens_seen": 2077491200 + }, + { + "epoch": 0.63, + "learning_rate": 0.00029950651955867603, + "loss": 1.3429, + "theoretical_loss": 3.4480453889856886, + "tokens_seen": 2077491200 + }, + { + "epoch": 0.63, + "learning_rate": 0.00029949047141424273, + "loss": 1.3254, + "theoretical_loss": 3.4480364620227855, + "tokens_seen": 2077556736 + }, + { + "epoch": 0.63, + "learning_rate": 0.00029947442326980943, + "loss": 1.2745, + "theoretical_loss": 3.448027535420322, + "tokens_seen": 2077622272 + }, + { + "epoch": 0.63, + "learning_rate": 0.00029945837512537613, + "loss": 1.3723, + "theoretical_loss": 3.448018609178271, + "tokens_seen": 2077687808 + }, + { + "epoch": 0.63, + "learning_rate": 0.00029944232698094283, + "loss": 1.4462, + "theoretical_loss": 3.448009683296608, + "tokens_seen": 2077753344 + }, + { + "epoch": 0.63, + "learning_rate": 0.00029942627883650953, + "loss": 1.361, + "theoretical_loss": 3.448000757775306, + "tokens_seen": 2077818880 + }, + { + "epoch": 0.63, + "learning_rate": 0.00029941023069207624, + "loss": 1.3416, + "theoretical_loss": 3.4479918326143397, + "tokens_seen": 2077884416 + }, + { + "epoch": 0.63, + "learning_rate": 0.00029939418254764294, + "loss": 1.3856, + "theoretical_loss": 3.447982907813683, + "tokens_seen": 2077949952 + }, + { + "epoch": 0.63, + "learning_rate": 0.00029937813440320964, + "loss": 1.3696, + "theoretical_loss": 3.44797398337331, + "tokens_seen": 2078015488 + }, + { + "epoch": 0.63, + "learning_rate": 0.00029936208625877634, + "loss": 1.2842, + "theoretical_loss": 3.4479650592931947, + "tokens_seen": 2078081024 + }, + { + "epoch": 0.63, + "learning_rate": 0.00029934603811434304, + "loss": 1.2991, + "theoretical_loss": 3.4479561355733113, + "tokens_seen": 2078146560 + }, + { + "epoch": 0.63, + "learning_rate": 0.00029932998996990974, + "loss": 1.2553, + "theoretical_loss": 3.447947212213634, + "tokens_seen": 2078212096 + }, + { + "epoch": 0.63, + "learning_rate": 0.00029931394182547644, + "loss": 1.2731, + "theoretical_loss": 3.447938289214137, + "tokens_seen": 2078277632 + }, + { + "epoch": 0.63, + "learning_rate": 0.00029929789368104314, + "loss": 1.2452, + "theoretical_loss": 3.447929366574793, + "tokens_seen": 2078343168 + }, + { + "epoch": 0.63, + "learning_rate": 0.00029928184553660984, + "loss": 1.3254, + "theoretical_loss": 3.447920444295578, + "tokens_seen": 2078408704 + }, + { + "epoch": 0.63, + "learning_rate": 0.00029926579739217654, + "loss": 1.2809, + "theoretical_loss": 3.4479115223764656, + "tokens_seen": 2078474240 + }, + { + "epoch": 0.63, + "learning_rate": 0.00029924974924774324, + "loss": 1.2854, + "theoretical_loss": 3.4479026008174296, + "tokens_seen": 2078539776 + }, + { + "epoch": 0.63, + "learning_rate": 0.00029923370110330994, + "loss": 1.3137, + "theoretical_loss": 3.447893679618444, + "tokens_seen": 2078605312 + }, + { + "epoch": 0.63, + "learning_rate": 0.00029921765295887664, + "loss": 1.3435, + "theoretical_loss": 3.447884758779483, + "tokens_seen": 2078670848 + }, + { + "epoch": 0.63, + "learning_rate": 0.00029920160481444334, + "loss": 1.3819, + "theoretical_loss": 3.4478758383005212, + "tokens_seen": 2078736384 + }, + { + "epoch": 0.63, + "learning_rate": 0.00029918555667001004, + "loss": 1.2871, + "theoretical_loss": 3.447866918181532, + "tokens_seen": 2078801920 + }, + { + "epoch": 0.63, + "learning_rate": 0.00029916950852557675, + "loss": 1.2839, + "theoretical_loss": 3.4478579984224904, + "tokens_seen": 2078867456 + }, + { + "epoch": 0.63, + "learning_rate": 0.00029915346038114345, + "loss": 1.349, + "theoretical_loss": 3.44784907902337, + "tokens_seen": 2078932992 + }, + { + "epoch": 0.63, + "learning_rate": 0.00029913741223671015, + "loss": 1.4061, + "theoretical_loss": 3.4478401599841444, + "tokens_seen": 2078998528 + }, + { + "epoch": 0.63, + "learning_rate": 0.00029912136409227685, + "loss": 1.3713, + "theoretical_loss": 3.447831241304789, + "tokens_seen": 2079064064 + }, + { + "epoch": 0.63, + "learning_rate": 0.00029910531594784355, + "loss": 1.2325, + "theoretical_loss": 3.4478223229852767, + "tokens_seen": 2079129600 + }, + { + "epoch": 0.63, + "learning_rate": 0.00029908926780341025, + "loss": 1.2533, + "theoretical_loss": 3.447813405025582, + "tokens_seen": 2079195136 + }, + { + "epoch": 0.63, + "learning_rate": 0.00029907321965897695, + "loss": 1.3105, + "theoretical_loss": 3.44780448742568, + "tokens_seen": 2079260672 + }, + { + "epoch": 0.63, + "learning_rate": 0.00029905717151454365, + "loss": 1.3727, + "theoretical_loss": 3.4477955701855434, + "tokens_seen": 2079326208 + }, + { + "epoch": 0.63, + "learning_rate": 0.00029904112337011035, + "loss": 1.2444, + "theoretical_loss": 3.447786653305147, + "tokens_seen": 2079391744 + }, + { + "epoch": 0.63, + "learning_rate": 0.00029902507522567705, + "loss": 1.2531, + "theoretical_loss": 3.4477777367844653, + "tokens_seen": 2079457280 + }, + { + "epoch": 0.63, + "learning_rate": 0.00029900902708124375, + "loss": 1.2586, + "theoretical_loss": 3.4477688206234722, + "tokens_seen": 2079522816 + }, + { + "epoch": 0.63, + "learning_rate": 0.00029899297893681045, + "loss": 1.3128, + "theoretical_loss": 3.4477599048221412, + "tokens_seen": 2079588352 + }, + { + "epoch": 0.63, + "learning_rate": 0.00029897693079237715, + "loss": 1.2813, + "theoretical_loss": 3.4477509893804474, + "tokens_seen": 2079653888 + }, + { + "epoch": 0.63, + "learning_rate": 0.00029896088264794385, + "loss": 1.308, + "theoretical_loss": 3.4477420742983647, + "tokens_seen": 2079719424 + }, + { + "epoch": 0.63, + "learning_rate": 0.00029894483450351055, + "loss": 1.2662, + "theoretical_loss": 3.4477331595758667, + "tokens_seen": 2079784960 + }, + { + "epoch": 0.63, + "learning_rate": 0.00029892878635907725, + "loss": 1.386, + "theoretical_loss": 3.447724245212928, + "tokens_seen": 2079850496 + }, + { + "epoch": 0.63, + "learning_rate": 0.00029891273821464396, + "loss": 1.321, + "theoretical_loss": 3.4477153312095234, + "tokens_seen": 2079916032 + }, + { + "epoch": 0.63, + "learning_rate": 0.00029889669007021066, + "loss": 1.2603, + "theoretical_loss": 3.447706417565626, + "tokens_seen": 2079981568 + }, + { + "epoch": 0.63, + "learning_rate": 0.00029888064192577736, + "loss": 1.2058, + "theoretical_loss": 3.4476975042812104, + "tokens_seen": 2080047104 + }, + { + "epoch": 0.63, + "learning_rate": 0.00029886459378134406, + "loss": 1.2402, + "theoretical_loss": 3.4476885913562514, + "tokens_seen": 2080112640 + }, + { + "epoch": 0.63, + "learning_rate": 0.00029884854563691076, + "loss": 1.3119, + "theoretical_loss": 3.447679678790722, + "tokens_seen": 2080178176 + }, + { + "epoch": 0.63, + "learning_rate": 0.00029883249749247746, + "loss": 1.3106, + "theoretical_loss": 3.4476707665845967, + "tokens_seen": 2080243712 + }, + { + "epoch": 0.63, + "learning_rate": 0.00029881644934804416, + "loss": 1.3497, + "theoretical_loss": 3.4476618547378504, + "tokens_seen": 2080309248 + }, + { + "epoch": 0.63, + "learning_rate": 0.00029880040120361086, + "loss": 1.4102, + "theoretical_loss": 3.447652943250457, + "tokens_seen": 2080374784 + }, + { + "epoch": 0.63, + "learning_rate": 0.00029878435305917756, + "loss": 1.3641, + "theoretical_loss": 3.44764403212239, + "tokens_seen": 2080440320 + }, + { + "epoch": 0.63, + "learning_rate": 0.00029876830491474426, + "loss": 1.3923, + "theoretical_loss": 3.4476351213536245, + "tokens_seen": 2080505856 + }, + { + "epoch": 0.63, + "learning_rate": 0.0002987522567703109, + "loss": 1.3004, + "theoretical_loss": 3.447626210944134, + "tokens_seen": 2080571392 + }, + { + "epoch": 0.63, + "learning_rate": 0.0002987362086258776, + "loss": 1.2701, + "theoretical_loss": 3.447617300893893, + "tokens_seen": 2080636928 + }, + { + "epoch": 0.63, + "learning_rate": 0.0002987201604814443, + "loss": 1.4471, + "theoretical_loss": 3.447608391202876, + "tokens_seen": 2080702464 + }, + { + "epoch": 0.63, + "objective/train/docs_used": 758504, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.2684274911880493, + "objective/train/theoretical_loss": 3.447599481871057, + "objective/train/tokens_used": 2101228000, + "theoretical_loss": 3.447599481871057, + "tokens_seen": 2080768000 + }, + { + "epoch": 0.63, + "learning_rate": 0.000298704112337011, + "loss": 1.3273, + "theoretical_loss": 3.447599481871057, + "tokens_seen": 2080768000 + }, + { + "epoch": 0.63, + "learning_rate": 0.0002986880641925777, + "loss": 1.2923, + "theoretical_loss": 3.44759057289841, + "tokens_seen": 2080833536 + }, + { + "epoch": 0.63, + "learning_rate": 0.0002986720160481444, + "loss": 1.37, + "theoretical_loss": 3.4475816642849093, + "tokens_seen": 2080899072 + }, + { + "epoch": 0.63, + "learning_rate": 0.0002986559679037111, + "loss": 1.2734, + "theoretical_loss": 3.447572756030529, + "tokens_seen": 2080964608 + }, + { + "epoch": 0.63, + "learning_rate": 0.0002986559679037111, + "loss": 1.4919, + "theoretical_loss": 3.4475638481352435, + "tokens_seen": 2081030144 + }, + { + "epoch": 0.63, + "learning_rate": 0.0002986399197592778, + "loss": 1.2843, + "theoretical_loss": 3.4475549405990273, + "tokens_seen": 2081095680 + }, + { + "epoch": 0.63, + "learning_rate": 0.0002986238716148445, + "loss": 1.3243, + "theoretical_loss": 3.4475460334218537, + "tokens_seen": 2081161216 + }, + { + "epoch": 0.63, + "learning_rate": 0.0002986078234704112, + "loss": 1.2103, + "theoretical_loss": 3.447537126603698, + "tokens_seen": 2081226752 + }, + { + "epoch": 0.63, + "learning_rate": 0.0002985917753259779, + "loss": 1.243, + "theoretical_loss": 3.447528220144534, + "tokens_seen": 2081292288 + }, + { + "epoch": 0.63, + "learning_rate": 0.0002985757271815446, + "loss": 1.2601, + "theoretical_loss": 3.447519314044335, + "tokens_seen": 2081357824 + }, + { + "epoch": 0.63, + "learning_rate": 0.0002985596790371113, + "loss": 1.3234, + "theoretical_loss": 3.447510408303077, + "tokens_seen": 2081423360 + }, + { + "epoch": 0.63, + "learning_rate": 0.000298543630892678, + "loss": 1.3286, + "theoretical_loss": 3.4475015029207334, + "tokens_seen": 2081488896 + }, + { + "epoch": 0.63, + "learning_rate": 0.0002985275827482447, + "loss": 1.2268, + "theoretical_loss": 3.447492597897278, + "tokens_seen": 2081554432 + }, + { + "epoch": 0.63, + "learning_rate": 0.0002985115346038114, + "loss": 1.3315, + "theoretical_loss": 3.4474836932326856, + "tokens_seen": 2081619968 + }, + { + "epoch": 0.63, + "learning_rate": 0.00029849548645937817, + "loss": 1.3339, + "theoretical_loss": 3.4474747889269297, + "tokens_seen": 2081685504 + }, + { + "epoch": 0.63, + "learning_rate": 0.00029847943831494487, + "loss": 1.3145, + "theoretical_loss": 3.4474658849799855, + "tokens_seen": 2081751040 + }, + { + "epoch": 0.63, + "learning_rate": 0.0002984633901705116, + "loss": 1.5039, + "theoretical_loss": 3.447456981391827, + "tokens_seen": 2081816576 + }, + { + "epoch": 0.63, + "learning_rate": 0.0002984473420260783, + "loss": 1.3841, + "theoretical_loss": 3.447448078162428, + "tokens_seen": 2081882112 + }, + { + "epoch": 0.63, + "learning_rate": 0.000298431293881645, + "loss": 1.3687, + "theoretical_loss": 3.4474391752917635, + "tokens_seen": 2081947648 + }, + { + "epoch": 0.63, + "learning_rate": 0.0002984152457372117, + "loss": 1.3482, + "theoretical_loss": 3.4474302727798065, + "tokens_seen": 2082013184 + }, + { + "epoch": 0.63, + "learning_rate": 0.0002983991975927784, + "loss": 1.3291, + "theoretical_loss": 3.4474213706265324, + "tokens_seen": 2082078720 + }, + { + "epoch": 0.63, + "learning_rate": 0.0002983831494483451, + "loss": 1.3649, + "theoretical_loss": 3.4474124688319154, + "tokens_seen": 2082144256 + }, + { + "epoch": 0.63, + "learning_rate": 0.0002983671013039118, + "loss": 1.3388, + "theoretical_loss": 3.4474035673959293, + "tokens_seen": 2082209792 + }, + { + "epoch": 0.63, + "learning_rate": 0.0002983510531594785, + "loss": 1.2216, + "theoretical_loss": 3.4473946663185484, + "tokens_seen": 2082275328 + }, + { + "epoch": 0.63, + "learning_rate": 0.0002983350050150452, + "loss": 1.3459, + "theoretical_loss": 3.4473857655997473, + "tokens_seen": 2082340864 + }, + { + "epoch": 0.63, + "learning_rate": 0.0002983189568706119, + "loss": 1.4389, + "theoretical_loss": 3.4473768652395, + "tokens_seen": 2082406400 + }, + { + "epoch": 0.63, + "learning_rate": 0.0002983029087261786, + "loss": 1.3898, + "theoretical_loss": 3.4473679652377807, + "tokens_seen": 2082471936 + }, + { + "epoch": 0.63, + "learning_rate": 0.0002982868605817453, + "loss": 1.3081, + "theoretical_loss": 3.447359065594564, + "tokens_seen": 2082537472 + }, + { + "epoch": 0.63, + "learning_rate": 0.000298270812437312, + "loss": 1.3388, + "theoretical_loss": 3.447350166309824, + "tokens_seen": 2082603008 + }, + { + "epoch": 0.63, + "learning_rate": 0.0002982547642928787, + "loss": 1.1657, + "theoretical_loss": 3.4473412673835346, + "tokens_seen": 2082668544 + }, + { + "epoch": 0.63, + "learning_rate": 0.0002982387161484454, + "loss": 1.492, + "theoretical_loss": 3.4473323688156707, + "tokens_seen": 2082734080 + }, + { + "epoch": 0.63, + "learning_rate": 0.0002982226680040121, + "loss": 1.4902, + "theoretical_loss": 3.4473234706062064, + "tokens_seen": 2082799616 + }, + { + "epoch": 0.63, + "learning_rate": 0.0002982066198595788, + "loss": 1.2114, + "theoretical_loss": 3.4473145727551158, + "tokens_seen": 2082865152 + }, + { + "epoch": 0.63, + "learning_rate": 0.0002981905717151455, + "loss": 1.3479, + "theoretical_loss": 3.447305675262373, + "tokens_seen": 2082930688 + }, + { + "epoch": 0.63, + "learning_rate": 0.0002981745235707122, + "loss": 1.2849, + "theoretical_loss": 3.4472967781279533, + "tokens_seen": 2082996224 + }, + { + "epoch": 0.63, + "learning_rate": 0.0002981584754262789, + "loss": 1.299, + "theoretical_loss": 3.44728788135183, + "tokens_seen": 2083061760 + }, + { + "epoch": 0.63, + "learning_rate": 0.0002981424272818456, + "loss": 1.224, + "theoretical_loss": 3.4472789849339778, + "tokens_seen": 2083127296 + }, + { + "epoch": 0.63, + "learning_rate": 0.0002981263791374123, + "loss": 1.294, + "theoretical_loss": 3.447270088874371, + "tokens_seen": 2083192832 + }, + { + "epoch": 0.63, + "learning_rate": 0.00029811033099297893, + "loss": 1.3349, + "theoretical_loss": 3.4472611931729835, + "tokens_seen": 2083258368 + }, + { + "epoch": 0.63, + "learning_rate": 0.00029809428284854563, + "loss": 1.3948, + "theoretical_loss": 3.4472522978297904, + "tokens_seen": 2083323904 + }, + { + "epoch": 0.63, + "learning_rate": 0.00029807823470411233, + "loss": 1.4345, + "theoretical_loss": 3.447243402844765, + "tokens_seen": 2083389440 + }, + { + "epoch": 0.63, + "learning_rate": 0.00029806218655967903, + "loss": 1.2832, + "theoretical_loss": 3.447234508217883, + "tokens_seen": 2083454976 + }, + { + "epoch": 0.63, + "learning_rate": 0.00029804613841524574, + "loss": 1.2775, + "theoretical_loss": 3.447225613949117, + "tokens_seen": 2083520512 + }, + { + "epoch": 0.63, + "learning_rate": 0.00029803009027081244, + "loss": 1.3, + "theoretical_loss": 3.4472167200384423, + "tokens_seen": 2083586048 + }, + { + "epoch": 0.63, + "learning_rate": 0.00029801404212637914, + "loss": 1.3204, + "theoretical_loss": 3.4472078264858332, + "tokens_seen": 2083651584 + }, + { + "epoch": 0.63, + "learning_rate": 0.00029799799398194584, + "loss": 1.32, + "theoretical_loss": 3.447198933291264, + "tokens_seen": 2083717120 + }, + { + "epoch": 0.63, + "learning_rate": 0.00029798194583751254, + "loss": 1.4048, + "theoretical_loss": 3.447190040454709, + "tokens_seen": 2083782656 + }, + { + "epoch": 0.63, + "learning_rate": 0.00029796589769307924, + "loss": 1.3539, + "theoretical_loss": 3.447181147976142, + "tokens_seen": 2083848192 + }, + { + "epoch": 0.63, + "learning_rate": 0.00029794984954864594, + "loss": 1.3622, + "theoretical_loss": 3.4471722558555387, + "tokens_seen": 2083913728 + }, + { + "epoch": 0.63, + "learning_rate": 0.00029793380140421264, + "loss": 1.3948, + "theoretical_loss": 3.447163364092872, + "tokens_seen": 2083979264 + }, + { + "epoch": 0.63, + "objective/train/docs_used": 759560, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.3500242233276367, + "objective/train/theoretical_loss": 3.4471544726881165, + "objective/train/tokens_used": 2104504800, + "theoretical_loss": 3.4471544726881165, + "tokens_seen": 2084044800 + }, + { + "epoch": 0.63, + "learning_rate": 0.00029791775325977934, + "loss": 1.4788, + "theoretical_loss": 3.4471544726881165, + "tokens_seen": 2084044800 + }, + { + "epoch": 0.63, + "learning_rate": 0.00029790170511534604, + "loss": 1.2704, + "theoretical_loss": 3.4471455816412475, + "tokens_seen": 2084110336 + }, + { + "epoch": 0.63, + "learning_rate": 0.00029788565697091274, + "loss": 1.2788, + "theoretical_loss": 3.447136690952238, + "tokens_seen": 2084175872 + }, + { + "epoch": 0.63, + "learning_rate": 0.00029786960882647944, + "loss": 1.2808, + "theoretical_loss": 3.4471278006210637, + "tokens_seen": 2084241408 + }, + { + "epoch": 0.63, + "learning_rate": 0.00029785356068204614, + "loss": 1.4464, + "theoretical_loss": 3.447118910647698, + "tokens_seen": 2084306944 + }, + { + "epoch": 0.63, + "learning_rate": 0.00029783751253761284, + "loss": 1.2539, + "theoretical_loss": 3.447110021032115, + "tokens_seen": 2084372480 + }, + { + "epoch": 0.63, + "learning_rate": 0.00029782146439317954, + "loss": 1.3561, + "theoretical_loss": 3.44710113177429, + "tokens_seen": 2084438016 + }, + { + "epoch": 0.63, + "learning_rate": 0.00029780541624874624, + "loss": 1.2898, + "theoretical_loss": 3.447092242874197, + "tokens_seen": 2084503552 + }, + { + "epoch": 0.63, + "learning_rate": 0.00029778936810431295, + "loss": 1.48, + "theoretical_loss": 3.4470833543318102, + "tokens_seen": 2084569088 + }, + { + "epoch": 0.63, + "learning_rate": 0.00029777331995987965, + "loss": 1.253, + "theoretical_loss": 3.4470744661471038, + "tokens_seen": 2084634624 + }, + { + "epoch": 0.63, + "learning_rate": 0.00029775727181544635, + "loss": 1.4069, + "theoretical_loss": 3.4470655783200526, + "tokens_seen": 2084700160 + }, + { + "epoch": 0.63, + "learning_rate": 0.00029774122367101305, + "loss": 1.2733, + "theoretical_loss": 3.447056690850631, + "tokens_seen": 2084765696 + }, + { + "epoch": 0.63, + "learning_rate": 0.00029772517552657975, + "loss": 1.4051, + "theoretical_loss": 3.447047803738812, + "tokens_seen": 2084831232 + }, + { + "epoch": 0.63, + "learning_rate": 0.00029770912738214645, + "loss": 1.2453, + "theoretical_loss": 3.447038916984572, + "tokens_seen": 2084896768 + }, + { + "epoch": 0.63, + "learning_rate": 0.00029769307923771315, + "loss": 1.2583, + "theoretical_loss": 3.4470300305878845, + "tokens_seen": 2084962304 + }, + { + "epoch": 0.63, + "learning_rate": 0.00029767703109327985, + "loss": 1.2609, + "theoretical_loss": 3.4470211445487235, + "tokens_seen": 2085027840 + }, + { + "epoch": 0.63, + "learning_rate": 0.00029766098294884655, + "loss": 1.4162, + "theoretical_loss": 3.447012258867064, + "tokens_seen": 2085093376 + }, + { + "epoch": 0.63, + "learning_rate": 0.00029764493480441325, + "loss": 1.2447, + "theoretical_loss": 3.44700337354288, + "tokens_seen": 2085158912 + }, + { + "epoch": 0.63, + "learning_rate": 0.00029762888665997995, + "loss": 1.2819, + "theoretical_loss": 3.446994488576146, + "tokens_seen": 2085224448 + }, + { + "epoch": 0.63, + "learning_rate": 0.00029761283851554665, + "loss": 1.2356, + "theoretical_loss": 3.4469856039668363, + "tokens_seen": 2085289984 + }, + { + "epoch": 0.63, + "learning_rate": 0.00029759679037111335, + "loss": 1.3484, + "theoretical_loss": 3.4469767197149253, + "tokens_seen": 2085355520 + }, + { + "epoch": 0.63, + "learning_rate": 0.00029758074222668005, + "loss": 1.3065, + "theoretical_loss": 3.446967835820388, + "tokens_seen": 2085421056 + }, + { + "epoch": 0.63, + "learning_rate": 0.00029756469408224675, + "loss": 1.1947, + "theoretical_loss": 3.4469589522831976, + "tokens_seen": 2085486592 + }, + { + "epoch": 0.63, + "learning_rate": 0.00029754864593781345, + "loss": 1.2925, + "theoretical_loss": 3.4469500691033295, + "tokens_seen": 2085552128 + }, + { + "epoch": 0.63, + "learning_rate": 0.00029753259779338016, + "loss": 1.3931, + "theoretical_loss": 3.4469411862807577, + "tokens_seen": 2085617664 + }, + { + "epoch": 0.63, + "learning_rate": 0.00029751654964894686, + "loss": 1.4087, + "theoretical_loss": 3.4469323038154567, + "tokens_seen": 2085683200 + }, + { + "epoch": 0.63, + "learning_rate": 0.00029750050150451356, + "loss": 1.3496, + "theoretical_loss": 3.4469234217074005, + "tokens_seen": 2085748736 + }, + { + "epoch": 0.63, + "learning_rate": 0.00029748445336008026, + "loss": 1.3511, + "theoretical_loss": 3.4469145399565644, + "tokens_seen": 2085814272 + }, + { + "epoch": 0.63, + "learning_rate": 0.00029746840521564696, + "loss": 1.3966, + "theoretical_loss": 3.4469056585629216, + "tokens_seen": 2085879808 + }, + { + "epoch": 0.63, + "learning_rate": 0.00029745235707121366, + "loss": 1.241, + "theoretical_loss": 3.4468967775264474, + "tokens_seen": 2085945344 + }, + { + "epoch": 0.63, + "learning_rate": 0.00029743630892678036, + "loss": 1.3589, + "theoretical_loss": 3.4468878968471164, + "tokens_seen": 2086010880 + }, + { + "epoch": 0.63, + "learning_rate": 0.00029742026078234706, + "loss": 1.4351, + "theoretical_loss": 3.4468790165249024, + "tokens_seen": 2086076416 + }, + { + "epoch": 0.63, + "learning_rate": 0.00029740421263791376, + "loss": 1.3797, + "theoretical_loss": 3.44687013655978, + "tokens_seen": 2086141952 + }, + { + "epoch": 0.63, + "learning_rate": 0.00029738816449348046, + "loss": 1.4322, + "theoretical_loss": 3.4468612569517236, + "tokens_seen": 2086207488 + }, + { + "epoch": 0.63, + "learning_rate": 0.00029737211634904716, + "loss": 1.332, + "theoretical_loss": 3.446852377700708, + "tokens_seen": 2086273024 + }, + { + "epoch": 0.63, + "learning_rate": 0.00029735606820461386, + "loss": 1.3283, + "theoretical_loss": 3.4468434988067065, + "tokens_seen": 2086338560 + }, + { + "epoch": 0.63, + "learning_rate": 0.0002973400200601805, + "loss": 1.3466, + "theoretical_loss": 3.446834620269695, + "tokens_seen": 2086404096 + }, + { + "epoch": 0.63, + "learning_rate": 0.0002973239719157472, + "loss": 1.3172, + "theoretical_loss": 3.446825742089647, + "tokens_seen": 2086469632 + }, + { + "epoch": 0.63, + "learning_rate": 0.0002973079237713139, + "loss": 1.2448, + "theoretical_loss": 3.4468168642665376, + "tokens_seen": 2086535168 + }, + { + "epoch": 0.63, + "learning_rate": 0.00029729187562688067, + "loss": 1.2853, + "theoretical_loss": 3.4468079868003407, + "tokens_seen": 2086600704 + }, + { + "epoch": 0.63, + "learning_rate": 0.00029727582748244737, + "loss": 1.3176, + "theoretical_loss": 3.446799109691031, + "tokens_seen": 2086666240 + }, + { + "epoch": 0.63, + "learning_rate": 0.00029725977933801407, + "loss": 1.4111, + "theoretical_loss": 3.446790232938582, + "tokens_seen": 2086731776 + }, + { + "epoch": 0.63, + "learning_rate": 0.00029724373119358077, + "loss": 1.2579, + "theoretical_loss": 3.44678135654297, + "tokens_seen": 2086797312 + }, + { + "epoch": 0.63, + "learning_rate": 0.00029722768304914747, + "loss": 1.2586, + "theoretical_loss": 3.446772480504168, + "tokens_seen": 2086862848 + }, + { + "epoch": 0.63, + "learning_rate": 0.00029721163490471417, + "loss": 1.2359, + "theoretical_loss": 3.4467636048221513, + "tokens_seen": 2086928384 + }, + { + "epoch": 0.63, + "learning_rate": 0.00029719558676028087, + "loss": 1.308, + "theoretical_loss": 3.4467547294968934, + "tokens_seen": 2086993920 + }, + { + "epoch": 0.63, + "learning_rate": 0.00029717953861584757, + "loss": 1.3313, + "theoretical_loss": 3.44674585452837, + "tokens_seen": 2087059456 + }, + { + "epoch": 0.63, + "learning_rate": 0.00029716349047141427, + "loss": 1.2665, + "theoretical_loss": 3.446736979916554, + "tokens_seen": 2087124992 + }, + { + "epoch": 0.63, + "learning_rate": 0.00029714744232698097, + "loss": 1.3406, + "theoretical_loss": 3.446728105661421, + "tokens_seen": 2087190528 + }, + { + "epoch": 0.63, + "learning_rate": 0.00029713139418254767, + "loss": 1.2383, + "theoretical_loss": 3.4467192317629456, + "tokens_seen": 2087256064 + }, + { + "epoch": 0.63, + "objective/train/docs_used": 760861, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.26979660987854, + "objective/train/theoretical_loss": 3.446710358221102, + "objective/train/tokens_used": 2107781600, + "theoretical_loss": 3.446710358221102, + "tokens_seen": 2087321600 + }, + { + "epoch": 0.63, + "learning_rate": 0.00029711534603811437, + "loss": 1.3055, + "theoretical_loss": 3.446710358221102, + "tokens_seen": 2087321600 + }, + { + "epoch": 0.63, + "learning_rate": 0.00029709929789368107, + "loss": 1.3154, + "theoretical_loss": 3.446701485035864, + "tokens_seen": 2087387136 + }, + { + "epoch": 0.63, + "learning_rate": 0.0002970832497492478, + "loss": 1.4046, + "theoretical_loss": 3.4466926122072064, + "tokens_seen": 2087452672 + }, + { + "epoch": 0.63, + "learning_rate": 0.0002970672016048145, + "loss": 1.2784, + "theoretical_loss": 3.4466837397351044, + "tokens_seen": 2087518208 + }, + { + "epoch": 0.63, + "learning_rate": 0.0002970511534603812, + "loss": 1.389, + "theoretical_loss": 3.4466748676195316, + "tokens_seen": 2087583744 + }, + { + "epoch": 0.63, + "learning_rate": 0.0002970351053159479, + "loss": 1.2895, + "theoretical_loss": 3.4466659958604633, + "tokens_seen": 2087649280 + }, + { + "epoch": 0.63, + "learning_rate": 0.0002970190571715146, + "loss": 1.3813, + "theoretical_loss": 3.4466571244578734, + "tokens_seen": 2087714816 + }, + { + "epoch": 0.63, + "learning_rate": 0.0002970030090270813, + "loss": 1.2996, + "theoretical_loss": 3.446648253411736, + "tokens_seen": 2087780352 + }, + { + "epoch": 0.63, + "learning_rate": 0.000296986960882648, + "loss": 1.376, + "theoretical_loss": 3.4466393827220267, + "tokens_seen": 2087845888 + }, + { + "epoch": 0.63, + "learning_rate": 0.0002969709127382147, + "loss": 1.3468, + "theoretical_loss": 3.4466305123887193, + "tokens_seen": 2087911424 + }, + { + "epoch": 0.63, + "learning_rate": 0.0002969548645937814, + "loss": 1.2424, + "theoretical_loss": 3.446621642411788, + "tokens_seen": 2087976960 + }, + { + "epoch": 0.63, + "learning_rate": 0.0002969388164493481, + "loss": 1.2921, + "theoretical_loss": 3.446612772791208, + "tokens_seen": 2088042496 + }, + { + "epoch": 0.63, + "learning_rate": 0.0002969227683049148, + "loss": 1.3689, + "theoretical_loss": 3.4466039035269533, + "tokens_seen": 2088108032 + }, + { + "epoch": 0.63, + "learning_rate": 0.0002969067201604815, + "loss": 1.3376, + "theoretical_loss": 3.4465950346189986, + "tokens_seen": 2088173568 + }, + { + "epoch": 0.63, + "learning_rate": 0.0002968906720160482, + "loss": 1.3279, + "theoretical_loss": 3.446586166067319, + "tokens_seen": 2088239104 + }, + { + "epoch": 0.63, + "learning_rate": 0.0002968746238716149, + "loss": 1.3413, + "theoretical_loss": 3.4465772978718876, + "tokens_seen": 2088304640 + }, + { + "epoch": 0.63, + "learning_rate": 0.0002968585757271816, + "loss": 1.3954, + "theoretical_loss": 3.44656843003268, + "tokens_seen": 2088370176 + }, + { + "epoch": 0.63, + "learning_rate": 0.0002968425275827483, + "loss": 1.3645, + "theoretical_loss": 3.4465595625496706, + "tokens_seen": 2088435712 + }, + { + "epoch": 0.63, + "learning_rate": 0.000296826479438315, + "loss": 1.3219, + "theoretical_loss": 3.4465506954228338, + "tokens_seen": 2088501248 + }, + { + "epoch": 0.63, + "learning_rate": 0.0002968104312938817, + "loss": 1.3684, + "theoretical_loss": 3.4465418286521436, + "tokens_seen": 2088566784 + }, + { + "epoch": 0.63, + "learning_rate": 0.0002967943831494484, + "loss": 1.4598, + "theoretical_loss": 3.4465329622375753, + "tokens_seen": 2088632320 + }, + { + "epoch": 0.63, + "learning_rate": 0.0002967783350050151, + "loss": 1.2744, + "theoretical_loss": 3.446524096179103, + "tokens_seen": 2088697856 + }, + { + "epoch": 0.63, + "learning_rate": 0.0002967622868605818, + "loss": 1.2989, + "theoretical_loss": 3.4465152304767015, + "tokens_seen": 2088763392 + }, + { + "epoch": 0.63, + "learning_rate": 0.0002967462387161485, + "loss": 1.3597, + "theoretical_loss": 3.446506365130345, + "tokens_seen": 2088828928 + }, + { + "epoch": 0.63, + "learning_rate": 0.0002967301905717152, + "loss": 1.3715, + "theoretical_loss": 3.446497500140008, + "tokens_seen": 2088894464 + }, + { + "epoch": 0.63, + "learning_rate": 0.00029671414242728183, + "loss": 1.3074, + "theoretical_loss": 3.4464886355056654, + "tokens_seen": 2088960000 + }, + { + "epoch": 0.63, + "learning_rate": 0.00029669809428284853, + "loss": 1.3108, + "theoretical_loss": 3.446479771227292, + "tokens_seen": 2089025536 + }, + { + "epoch": 0.63, + "learning_rate": 0.00029668204613841524, + "loss": 1.3516, + "theoretical_loss": 3.446470907304861, + "tokens_seen": 2089091072 + }, + { + "epoch": 0.63, + "learning_rate": 0.00029666599799398194, + "loss": 1.3377, + "theoretical_loss": 3.4464620437383484, + "tokens_seen": 2089156608 + }, + { + "epoch": 0.63, + "learning_rate": 0.00029664994984954864, + "loss": 1.4211, + "theoretical_loss": 3.446453180527728, + "tokens_seen": 2089222144 + }, + { + "epoch": 0.63, + "learning_rate": 0.00029663390170511534, + "loss": 1.3225, + "theoretical_loss": 3.4464443176729747, + "tokens_seen": 2089287680 + }, + { + "epoch": 0.63, + "learning_rate": 0.00029661785356068204, + "loss": 1.4562, + "theoretical_loss": 3.446435455174063, + "tokens_seen": 2089353216 + }, + { + "epoch": 0.63, + "learning_rate": 0.00029660180541624874, + "loss": 1.3316, + "theoretical_loss": 3.446426593030967, + "tokens_seen": 2089418752 + }, + { + "epoch": 0.63, + "learning_rate": 0.00029658575727181544, + "loss": 1.3529, + "theoretical_loss": 3.4464177312436615, + "tokens_seen": 2089484288 + }, + { + "epoch": 0.63, + "learning_rate": 0.00029656970912738214, + "loss": 1.2984, + "theoretical_loss": 3.446408869812122, + "tokens_seen": 2089549824 + }, + { + "epoch": 0.63, + "learning_rate": 0.00029655366098294884, + "loss": 1.4236, + "theoretical_loss": 3.446400008736321, + "tokens_seen": 2089615360 + }, + { + "epoch": 0.63, + "learning_rate": 0.00029653761283851554, + "loss": 1.3769, + "theoretical_loss": 3.446391148016235, + "tokens_seen": 2089680896 + }, + { + "epoch": 0.63, + "learning_rate": 0.00029652156469408224, + "loss": 1.2754, + "theoretical_loss": 3.4463822876518377, + "tokens_seen": 2089746432 + }, + { + "epoch": 0.63, + "learning_rate": 0.00029650551654964894, + "loss": 1.2954, + "theoretical_loss": 3.4463734276431035, + "tokens_seen": 2089811968 + }, + { + "epoch": 0.63, + "learning_rate": 0.00029648946840521564, + "loss": 1.3416, + "theoretical_loss": 3.446364567990008, + "tokens_seen": 2089877504 + }, + { + "epoch": 0.63, + "learning_rate": 0.00029647342026078234, + "loss": 1.3429, + "theoretical_loss": 3.4463557086925243, + "tokens_seen": 2089943040 + }, + { + "epoch": 0.63, + "learning_rate": 0.00029645737211634904, + "loss": 1.2992, + "theoretical_loss": 3.446346849750628, + "tokens_seen": 2090008576 + }, + { + "epoch": 0.63, + "learning_rate": 0.00029644132397191574, + "loss": 1.329, + "theoretical_loss": 3.4463379911642935, + "tokens_seen": 2090074112 + }, + { + "epoch": 0.63, + "learning_rate": 0.00029642527582748245, + "loss": 1.3095, + "theoretical_loss": 3.446329132933495, + "tokens_seen": 2090139648 + }, + { + "epoch": 0.63, + "learning_rate": 0.00029640922768304915, + "loss": 1.3458, + "theoretical_loss": 3.4463202750582074, + "tokens_seen": 2090205184 + }, + { + "epoch": 0.63, + "learning_rate": 0.00029639317953861585, + "loss": 1.3203, + "theoretical_loss": 3.4463114175384053, + "tokens_seen": 2090270720 + }, + { + "epoch": 0.63, + "learning_rate": 0.00029637713139418255, + "loss": 1.2865, + "theoretical_loss": 3.446302560374064, + "tokens_seen": 2090336256 + }, + { + "epoch": 0.63, + "learning_rate": 0.00029636108324974925, + "loss": 1.3797, + "theoretical_loss": 3.446293703565156, + "tokens_seen": 2090401792 + }, + { + "epoch": 0.63, + "learning_rate": 0.00029634503510531595, + "loss": 1.3526, + "theoretical_loss": 3.446284847111658, + "tokens_seen": 2090467328 + }, + { + "epoch": 0.63, + "learning_rate": 0.00029632898696088265, + "loss": 1.3281, + "theoretical_loss": 3.4462759910135436, + "tokens_seen": 2090532864 + }, + { + "epoch": 0.63, + "objective/train/docs_used": 762048, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.3908305168151855, + "objective/train/theoretical_loss": 3.4462671352707877, + "objective/train/tokens_used": 2111058400, + "theoretical_loss": 3.4462671352707877, + "tokens_seen": 2090598400 + }, + { + "epoch": 0.63, + "learning_rate": 0.00029631293881644935, + "loss": 1.4013, + "theoretical_loss": 3.4462671352707877, + "tokens_seen": 2090598400 + }, + { + "epoch": 0.63, + "learning_rate": 0.00029629689067201605, + "loss": 1.3017, + "theoretical_loss": 3.4462582798833648, + "tokens_seen": 2090663936 + }, + { + "epoch": 0.63, + "learning_rate": 0.00029628084252758275, + "loss": 1.3417, + "theoretical_loss": 3.4462494248512496, + "tokens_seen": 2090729472 + }, + { + "epoch": 0.63, + "learning_rate": 0.00029626479438314945, + "loss": 1.289, + "theoretical_loss": 3.4462405701744165, + "tokens_seen": 2090795008 + }, + { + "epoch": 0.63, + "learning_rate": 0.00029624874623871615, + "loss": 1.4102, + "theoretical_loss": 3.44623171585284, + "tokens_seen": 2090860544 + }, + { + "epoch": 0.63, + "learning_rate": 0.00029623269809428285, + "loss": 1.3425, + "theoretical_loss": 3.4462228618864956, + "tokens_seen": 2090926080 + }, + { + "epoch": 0.63, + "learning_rate": 0.00029621664994984955, + "loss": 1.2265, + "theoretical_loss": 3.4462140082753567, + "tokens_seen": 2090991616 + }, + { + "epoch": 0.63, + "learning_rate": 0.00029620060180541625, + "loss": 1.5147, + "theoretical_loss": 3.446205155019399, + "tokens_seen": 2091057152 + }, + { + "epoch": 0.63, + "learning_rate": 0.00029618455366098295, + "loss": 1.3411, + "theoretical_loss": 3.4461963021185964, + "tokens_seen": 2091122688 + }, + { + "epoch": 0.63, + "learning_rate": 0.00029616850551654966, + "loss": 1.363, + "theoretical_loss": 3.4461874495729234, + "tokens_seen": 2091188224 + }, + { + "epoch": 0.63, + "learning_rate": 0.00029615245737211636, + "loss": 1.3277, + "theoretical_loss": 3.4461785973823553, + "tokens_seen": 2091253760 + }, + { + "epoch": 0.63, + "learning_rate": 0.00029613640922768306, + "loss": 1.2909, + "theoretical_loss": 3.4461697455468663, + "tokens_seen": 2091319296 + }, + { + "epoch": 0.63, + "learning_rate": 0.00029612036108324976, + "loss": 1.3095, + "theoretical_loss": 3.4461608940664314, + "tokens_seen": 2091384832 + }, + { + "epoch": 0.63, + "learning_rate": 0.00029610431293881646, + "loss": 1.2936, + "theoretical_loss": 3.446152042941024, + "tokens_seen": 2091450368 + }, + { + "epoch": 0.63, + "learning_rate": 0.00029608826479438316, + "loss": 1.2506, + "theoretical_loss": 3.446143192170621, + "tokens_seen": 2091515904 + }, + { + "epoch": 0.63, + "learning_rate": 0.00029607221664994986, + "loss": 1.2975, + "theoretical_loss": 3.4461343417551946, + "tokens_seen": 2091581440 + }, + { + "epoch": 0.63, + "learning_rate": 0.00029605616850551656, + "loss": 1.3262, + "theoretical_loss": 3.446125491694721, + "tokens_seen": 2091646976 + }, + { + "epoch": 0.63, + "learning_rate": 0.00029604012036108326, + "loss": 1.4121, + "theoretical_loss": 3.4461166419891747, + "tokens_seen": 2091712512 + }, + { + "epoch": 0.63, + "learning_rate": 0.00029602407221664996, + "loss": 1.3126, + "theoretical_loss": 3.4461077926385295, + "tokens_seen": 2091778048 + }, + { + "epoch": 0.63, + "learning_rate": 0.00029600802407221666, + "loss": 1.3262, + "theoretical_loss": 3.446098943642761, + "tokens_seen": 2091843584 + }, + { + "epoch": 0.63, + "learning_rate": 0.00029599197592778336, + "loss": 1.3486, + "theoretical_loss": 3.446090095001843, + "tokens_seen": 2091909120 + }, + { + "epoch": 0.63, + "learning_rate": 0.00029597592778335006, + "loss": 1.2947, + "theoretical_loss": 3.4460812467157513, + "tokens_seen": 2091974656 + }, + { + "epoch": 0.63, + "learning_rate": 0.00029595987963891676, + "loss": 1.3657, + "theoretical_loss": 3.4460723987844593, + "tokens_seen": 2092040192 + }, + { + "epoch": 0.63, + "learning_rate": 0.00029594383149448346, + "loss": 1.4042, + "theoretical_loss": 3.446063551207942, + "tokens_seen": 2092105728 + }, + { + "epoch": 0.63, + "learning_rate": 0.00029592778335005016, + "loss": 1.2949, + "theoretical_loss": 3.446054703986175, + "tokens_seen": 2092171264 + }, + { + "epoch": 0.63, + "learning_rate": 0.00029591173520561687, + "loss": 1.3317, + "theoretical_loss": 3.446045857119132, + "tokens_seen": 2092236800 + }, + { + "epoch": 0.63, + "learning_rate": 0.00029589568706118357, + "loss": 1.4057, + "theoretical_loss": 3.446037010606788, + "tokens_seen": 2092302336 + }, + { + "epoch": 0.63, + "learning_rate": 0.00029587963891675027, + "loss": 1.2231, + "theoretical_loss": 3.446028164449117, + "tokens_seen": 2092367872 + }, + { + "epoch": 0.63, + "learning_rate": 0.00029586359077231697, + "loss": 1.3863, + "theoretical_loss": 3.446019318646095, + "tokens_seen": 2092433408 + }, + { + "epoch": 0.63, + "learning_rate": 0.00029584754262788367, + "loss": 1.3465, + "theoretical_loss": 3.4460104731976955, + "tokens_seen": 2092498944 + }, + { + "epoch": 0.63, + "learning_rate": 0.00029583149448345037, + "loss": 1.3454, + "theoretical_loss": 3.4460016281038937, + "tokens_seen": 2092564480 + }, + { + "epoch": 0.63, + "learning_rate": 0.00029581544633901707, + "loss": 1.3148, + "theoretical_loss": 3.445992783364664, + "tokens_seen": 2092630016 + }, + { + "epoch": 0.63, + "learning_rate": 0.00029579939819458377, + "loss": 1.2801, + "theoretical_loss": 3.445983938979982, + "tokens_seen": 2092695552 + }, + { + "epoch": 0.63, + "learning_rate": 0.00029578335005015047, + "loss": 1.3163, + "theoretical_loss": 3.4459750949498207, + "tokens_seen": 2092761088 + }, + { + "epoch": 0.63, + "learning_rate": 0.00029576730190571717, + "loss": 1.2648, + "theoretical_loss": 3.4459662512741565, + "tokens_seen": 2092826624 + }, + { + "epoch": 0.63, + "learning_rate": 0.00029575125376128387, + "loss": 1.3087, + "theoretical_loss": 3.445957407952963, + "tokens_seen": 2092892160 + }, + { + "epoch": 0.63, + "learning_rate": 0.00029573520561685057, + "loss": 1.3511, + "theoretical_loss": 3.4459485649862147, + "tokens_seen": 2092957696 + }, + { + "epoch": 0.63, + "learning_rate": 0.0002957191574724173, + "loss": 1.2622, + "theoretical_loss": 3.4459397223738875, + "tokens_seen": 2093023232 + }, + { + "epoch": 0.63, + "learning_rate": 0.000295703109327984, + "loss": 1.3216, + "theoretical_loss": 3.445930880115955, + "tokens_seen": 2093088768 + }, + { + "epoch": 0.63, + "learning_rate": 0.0002956870611835507, + "loss": 1.3372, + "theoretical_loss": 3.4459220382123927, + "tokens_seen": 2093154304 + }, + { + "epoch": 0.63, + "learning_rate": 0.0002956710130391174, + "loss": 1.3207, + "theoretical_loss": 3.445913196663175, + "tokens_seen": 2093219840 + }, + { + "epoch": 0.63, + "learning_rate": 0.0002956549648946841, + "loss": 1.2409, + "theoretical_loss": 3.4459043554682762, + "tokens_seen": 2093285376 + }, + { + "epoch": 0.63, + "learning_rate": 0.0002956389167502508, + "loss": 1.1877, + "theoretical_loss": 3.4458955146276713, + "tokens_seen": 2093350912 + }, + { + "epoch": 0.63, + "learning_rate": 0.0002956228686058175, + "loss": 1.3117, + "theoretical_loss": 3.4458866741413354, + "tokens_seen": 2093416448 + }, + { + "epoch": 0.63, + "learning_rate": 0.0002956068204613842, + "loss": 1.2646, + "theoretical_loss": 3.445877834009243, + "tokens_seen": 2093481984 + }, + { + "epoch": 0.63, + "learning_rate": 0.0002955907723169509, + "loss": 1.26, + "theoretical_loss": 3.4458689942313683, + "tokens_seen": 2093547520 + }, + { + "epoch": 0.63, + "learning_rate": 0.0002955747241725176, + "loss": 1.4079, + "theoretical_loss": 3.4458601548076864, + "tokens_seen": 2093613056 + }, + { + "epoch": 0.63, + "learning_rate": 0.0002955586760280843, + "loss": 1.3868, + "theoretical_loss": 3.445851315738172, + "tokens_seen": 2093678592 + }, + { + "epoch": 0.63, + "learning_rate": 0.000295542627883651, + "loss": 1.322, + "theoretical_loss": 3.4458424770227998, + "tokens_seen": 2093744128 + }, + { + "epoch": 0.63, + "learning_rate": 0.0002955265797392177, + "loss": 1.2809, + "theoretical_loss": 3.445833638661545, + "tokens_seen": 2093809664 + }, + { + "epoch": 0.63, + "objective/train/docs_used": 763180, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.337571144104004, + "objective/train/theoretical_loss": 3.4458248006543815, + "objective/train/tokens_used": 2114335200, + "theoretical_loss": 3.4458248006543815, + "tokens_seen": 2093875200 + }, + { + "epoch": 0.63, + "learning_rate": 0.0002955105315947844, + "loss": 1.3111, + "theoretical_loss": 3.4458248006543815, + "tokens_seen": 2093875200 + }, + { + "epoch": 0.63, + "learning_rate": 0.0002954944834503511, + "loss": 1.2775, + "theoretical_loss": 3.4458159630012846, + "tokens_seen": 2093940736 + }, + { + "epoch": 0.63, + "learning_rate": 0.0002954784353059178, + "loss": 1.3462, + "theoretical_loss": 3.445807125702229, + "tokens_seen": 2094006272 + }, + { + "epoch": 0.63, + "learning_rate": 0.0002954623871614845, + "loss": 1.2822, + "theoretical_loss": 3.4457982887571887, + "tokens_seen": 2094071808 + }, + { + "epoch": 0.63, + "learning_rate": 0.0002954463390170512, + "loss": 1.331, + "theoretical_loss": 3.4457894521661396, + "tokens_seen": 2094137344 + }, + { + "epoch": 0.63, + "learning_rate": 0.0002954302908726179, + "loss": 1.3238, + "theoretical_loss": 3.445780615929056, + "tokens_seen": 2094202880 + }, + { + "epoch": 0.63, + "learning_rate": 0.0002954142427281846, + "loss": 1.2738, + "theoretical_loss": 3.445771780045912, + "tokens_seen": 2094268416 + }, + { + "epoch": 0.63, + "learning_rate": 0.0002953981945837513, + "loss": 1.3604, + "theoretical_loss": 3.4457629445166833, + "tokens_seen": 2094333952 + }, + { + "epoch": 0.63, + "learning_rate": 0.000295382146439318, + "loss": 1.2753, + "theoretical_loss": 3.445754109341344, + "tokens_seen": 2094399488 + }, + { + "epoch": 0.63, + "learning_rate": 0.0002953660982948847, + "loss": 1.371, + "theoretical_loss": 3.4457452745198696, + "tokens_seen": 2094465024 + }, + { + "epoch": 0.63, + "learning_rate": 0.0002953500501504514, + "loss": 1.3281, + "theoretical_loss": 3.445736440052234, + "tokens_seen": 2094530560 + }, + { + "epoch": 0.63, + "learning_rate": 0.0002953340020060181, + "loss": 1.3328, + "theoretical_loss": 3.445727605938412, + "tokens_seen": 2094596096 + }, + { + "epoch": 0.63, + "learning_rate": 0.00029531795386158473, + "loss": 1.3472, + "theoretical_loss": 3.445718772178379, + "tokens_seen": 2094661632 + }, + { + "epoch": 0.63, + "learning_rate": 0.00029530190571715144, + "loss": 1.2897, + "theoretical_loss": 3.445709938772109, + "tokens_seen": 2094727168 + }, + { + "epoch": 0.63, + "learning_rate": 0.00029528585757271814, + "loss": 1.254, + "theoretical_loss": 3.4457011057195777, + "tokens_seen": 2094792704 + }, + { + "epoch": 0.63, + "learning_rate": 0.00029526980942828484, + "loss": 1.3075, + "theoretical_loss": 3.445692273020759, + "tokens_seen": 2094858240 + }, + { + "epoch": 0.63, + "learning_rate": 0.00029525376128385154, + "loss": 1.3674, + "theoretical_loss": 3.4456834406756283, + "tokens_seen": 2094923776 + }, + { + "epoch": 0.63, + "learning_rate": 0.00029523771313941824, + "loss": 1.2343, + "theoretical_loss": 3.44567460868416, + "tokens_seen": 2094989312 + }, + { + "epoch": 0.63, + "learning_rate": 0.00029522166499498494, + "loss": 1.3365, + "theoretical_loss": 3.4456657770463286, + "tokens_seen": 2095054848 + }, + { + "epoch": 0.63, + "learning_rate": 0.00029520561685055164, + "loss": 1.3448, + "theoretical_loss": 3.44565694576211, + "tokens_seen": 2095120384 + }, + { + "epoch": 0.63, + "learning_rate": 0.00029518956870611834, + "loss": 1.3429, + "theoretical_loss": 3.4456481148314775, + "tokens_seen": 2095185920 + }, + { + "epoch": 0.63, + "learning_rate": 0.00029517352056168504, + "loss": 1.3789, + "theoretical_loss": 3.445639284254407, + "tokens_seen": 2095251456 + }, + { + "epoch": 0.63, + "learning_rate": 0.00029515747241725174, + "loss": 1.3239, + "theoretical_loss": 3.445630454030873, + "tokens_seen": 2095316992 + }, + { + "epoch": 0.63, + "learning_rate": 0.00029514142427281844, + "loss": 1.3158, + "theoretical_loss": 3.44562162416085, + "tokens_seen": 2095382528 + }, + { + "epoch": 0.63, + "learning_rate": 0.00029512537612838514, + "loss": 1.3115, + "theoretical_loss": 3.4456127946443127, + "tokens_seen": 2095448064 + }, + { + "epoch": 0.64, + "learning_rate": 0.00029510932798395184, + "loss": 1.3118, + "theoretical_loss": 3.4456039654812365, + "tokens_seen": 2095513600 + }, + { + "epoch": 0.64, + "learning_rate": 0.00029509327983951854, + "loss": 1.2772, + "theoretical_loss": 3.445595136671596, + "tokens_seen": 2095579136 + }, + { + "epoch": 0.64, + "learning_rate": 0.00029507723169508524, + "loss": 1.4305, + "theoretical_loss": 3.4455863082153657, + "tokens_seen": 2095644672 + }, + { + "epoch": 0.64, + "learning_rate": 0.00029506118355065194, + "loss": 1.3971, + "theoretical_loss": 3.4455774801125205, + "tokens_seen": 2095710208 + }, + { + "epoch": 0.64, + "learning_rate": 0.00029504513540621865, + "loss": 1.3407, + "theoretical_loss": 3.4455686523630358, + "tokens_seen": 2095775744 + }, + { + "epoch": 0.64, + "learning_rate": 0.00029502908726178535, + "loss": 1.3448, + "theoretical_loss": 3.4455598249668853, + "tokens_seen": 2095841280 + }, + { + "epoch": 0.64, + "learning_rate": 0.00029501303911735205, + "loss": 1.2966, + "theoretical_loss": 3.4455509979240446, + "tokens_seen": 2095906816 + }, + { + "epoch": 0.64, + "learning_rate": 0.00029499699097291875, + "loss": 1.3537, + "theoretical_loss": 3.4455421712344885, + "tokens_seen": 2095972352 + }, + { + "epoch": 0.64, + "learning_rate": 0.00029498094282848545, + "loss": 1.2924, + "theoretical_loss": 3.445533344898191, + "tokens_seen": 2096037888 + }, + { + "epoch": 0.64, + "learning_rate": 0.00029496489468405215, + "loss": 1.2963, + "theoretical_loss": 3.4455245189151285, + "tokens_seen": 2096103424 + }, + { + "epoch": 0.64, + "learning_rate": 0.00029494884653961885, + "loss": 1.385, + "theoretical_loss": 3.445515693285274, + "tokens_seen": 2096168960 + }, + { + "epoch": 0.64, + "learning_rate": 0.00029493279839518555, + "loss": 1.3448, + "theoretical_loss": 3.4455068680086036, + "tokens_seen": 2096234496 + }, + { + "epoch": 0.64, + "learning_rate": 0.00029491675025075225, + "loss": 1.3706, + "theoretical_loss": 3.4454980430850917, + "tokens_seen": 2096300032 + }, + { + "epoch": 0.64, + "learning_rate": 0.00029490070210631895, + "loss": 1.4694, + "theoretical_loss": 3.4454892185147132, + "tokens_seen": 2096365568 + }, + { + "epoch": 0.64, + "learning_rate": 0.0002948846539618857, + "loss": 1.3267, + "theoretical_loss": 3.4454803942974426, + "tokens_seen": 2096431104 + }, + { + "epoch": 0.64, + "learning_rate": 0.0002948686058174524, + "loss": 1.3694, + "theoretical_loss": 3.4454715704332552, + "tokens_seen": 2096496640 + }, + { + "epoch": 0.64, + "learning_rate": 0.0002948525576730191, + "loss": 1.4201, + "theoretical_loss": 3.445462746922126, + "tokens_seen": 2096562176 + }, + { + "epoch": 0.64, + "learning_rate": 0.0002948365095285858, + "loss": 1.3442, + "theoretical_loss": 3.4454539237640285, + "tokens_seen": 2096627712 + }, + { + "epoch": 0.64, + "learning_rate": 0.0002948204613841525, + "loss": 1.4022, + "theoretical_loss": 3.4454451009589393, + "tokens_seen": 2096693248 + }, + { + "epoch": 0.64, + "learning_rate": 0.0002948044132397192, + "loss": 1.3709, + "theoretical_loss": 3.445436278506832, + "tokens_seen": 2096758784 + }, + { + "epoch": 0.64, + "learning_rate": 0.0002947883650952859, + "loss": 1.3291, + "theoretical_loss": 3.4454274564076823, + "tokens_seen": 2096824320 + }, + { + "epoch": 0.64, + "learning_rate": 0.0002947723169508526, + "loss": 1.3853, + "theoretical_loss": 3.445418634661465, + "tokens_seen": 2096889856 + }, + { + "epoch": 0.64, + "learning_rate": 0.0002947562688064193, + "loss": 1.3381, + "theoretical_loss": 3.445409813268154, + "tokens_seen": 2096955392 + }, + { + "epoch": 0.64, + "learning_rate": 0.000294740220661986, + "loss": 1.3137, + "theoretical_loss": 3.4454009922277242, + "tokens_seen": 2097020928 + }, + { + "epoch": 0.64, + "learning_rate": 0.0002947241725175527, + "loss": 1.2727, + "theoretical_loss": 3.445392171540152, + "tokens_seen": 2097086464 + }, + { + "epoch": 0.64, + "objective/train/docs_used": 764413, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.4654735326766968, + "objective/train/theoretical_loss": 3.445383351205411, + "objective/train/tokens_used": 2117612000, + "theoretical_loss": 3.445383351205411, + "tokens_seen": 2097152000 + }, + { + "epoch": 0.64, + "learning_rate": 0.0002947081243731194, + "loss": 1.3947, + "theoretical_loss": 3.445383351205411, + "tokens_seen": 2097152000 + }, + { + "epoch": 0.64, + "learning_rate": 0.0002946920762286861, + "loss": 1.3034, + "theoretical_loss": 3.445374531223476, + "tokens_seen": 2097217536 + }, + { + "epoch": 0.64, + "learning_rate": 0.00029467602808425276, + "loss": 1.2391, + "theoretical_loss": 3.445365711594323, + "tokens_seen": 2097283072 + }, + { + "epoch": 0.64, + "learning_rate": 0.00029465997993981946, + "loss": 1.3249, + "theoretical_loss": 3.445356892317925, + "tokens_seen": 2097348608 + }, + { + "epoch": 0.64, + "learning_rate": 0.00029464393179538616, + "loss": 1.4631, + "theoretical_loss": 3.4453480733942587, + "tokens_seen": 2097414144 + }, + { + "epoch": 0.64, + "learning_rate": 0.00029462788365095286, + "loss": 1.294, + "theoretical_loss": 3.445339254823298, + "tokens_seen": 2097479680 + }, + { + "epoch": 0.64, + "learning_rate": 0.00029461183550651956, + "loss": 1.3278, + "theoretical_loss": 3.445330436605018, + "tokens_seen": 2097545216 + }, + { + "epoch": 0.64, + "learning_rate": 0.00029459578736208626, + "loss": 1.3856, + "theoretical_loss": 3.4453216187393934, + "tokens_seen": 2097610752 + }, + { + "epoch": 0.64, + "learning_rate": 0.00029457973921765296, + "loss": 1.4326, + "theoretical_loss": 3.4453128012263994, + "tokens_seen": 2097676288 + }, + { + "epoch": 0.64, + "learning_rate": 0.00029456369107321966, + "loss": 1.3176, + "theoretical_loss": 3.445303984066011, + "tokens_seen": 2097741824 + }, + { + "epoch": 0.64, + "learning_rate": 0.00029454764292878637, + "loss": 1.3734, + "theoretical_loss": 3.445295167258202, + "tokens_seen": 2097807360 + }, + { + "epoch": 0.64, + "learning_rate": 0.00029453159478435307, + "loss": 1.3409, + "theoretical_loss": 3.4452863508029488, + "tokens_seen": 2097872896 + }, + { + "epoch": 0.64, + "learning_rate": 0.00029451554663991977, + "loss": 1.338, + "theoretical_loss": 3.4452775347002254, + "tokens_seen": 2097938432 + }, + { + "epoch": 0.64, + "learning_rate": 0.00029449949849548647, + "loss": 1.3045, + "theoretical_loss": 3.4452687189500066, + "tokens_seen": 2098003968 + }, + { + "epoch": 0.64, + "learning_rate": 0.00029448345035105317, + "loss": 1.2749, + "theoretical_loss": 3.4452599035522677, + "tokens_seen": 2098069504 + }, + { + "epoch": 0.64, + "learning_rate": 0.00029446740220661987, + "loss": 1.1998, + "theoretical_loss": 3.445251088506984, + "tokens_seen": 2098135040 + }, + { + "epoch": 0.64, + "learning_rate": 0.00029445135406218657, + "loss": 1.2867, + "theoretical_loss": 3.445242273814129, + "tokens_seen": 2098200576 + }, + { + "epoch": 0.64, + "learning_rate": 0.00029443530591775327, + "loss": 1.4168, + "theoretical_loss": 3.4452334594736786, + "tokens_seen": 2098266112 + }, + { + "epoch": 0.64, + "learning_rate": 0.00029441925777331997, + "loss": 1.3701, + "theoretical_loss": 3.4452246454856077, + "tokens_seen": 2098331648 + }, + { + "epoch": 0.64, + "learning_rate": 0.00029440320962888667, + "loss": 1.384, + "theoretical_loss": 3.4452158318498913, + "tokens_seen": 2098397184 + }, + { + "epoch": 0.64, + "learning_rate": 0.00029438716148445337, + "loss": 1.3056, + "theoretical_loss": 3.4452070185665042, + "tokens_seen": 2098462720 + }, + { + "epoch": 0.64, + "learning_rate": 0.00029437111334002007, + "loss": 1.3614, + "theoretical_loss": 3.445198205635421, + "tokens_seen": 2098528256 + }, + { + "epoch": 0.64, + "learning_rate": 0.00029435506519558677, + "loss": 1.251, + "theoretical_loss": 3.4451893930566166, + "tokens_seen": 2098593792 + }, + { + "epoch": 0.64, + "learning_rate": 0.0002943390170511535, + "loss": 1.221, + "theoretical_loss": 3.4451805808300664, + "tokens_seen": 2098659328 + }, + { + "epoch": 0.64, + "learning_rate": 0.0002943229689067202, + "loss": 1.458, + "theoretical_loss": 3.4451717689557446, + "tokens_seen": 2098724864 + }, + { + "epoch": 0.64, + "learning_rate": 0.0002943069207622869, + "loss": 1.38, + "theoretical_loss": 3.4451629574336273, + "tokens_seen": 2098790400 + }, + { + "epoch": 0.64, + "learning_rate": 0.0002942908726178536, + "loss": 1.3417, + "theoretical_loss": 3.4451541462636883, + "tokens_seen": 2098855936 + }, + { + "epoch": 0.64, + "learning_rate": 0.0002942748244734203, + "loss": 1.3062, + "theoretical_loss": 3.445145335445903, + "tokens_seen": 2098921472 + }, + { + "epoch": 0.64, + "learning_rate": 0.000294258776328987, + "loss": 1.3559, + "theoretical_loss": 3.4451365249802466, + "tokens_seen": 2098987008 + }, + { + "epoch": 0.64, + "learning_rate": 0.0002942427281845537, + "loss": 1.3072, + "theoretical_loss": 3.4451277148666932, + "tokens_seen": 2099052544 + }, + { + "epoch": 0.64, + "learning_rate": 0.0002942266800401204, + "loss": 1.3929, + "theoretical_loss": 3.4451189051052182, + "tokens_seen": 2099118080 + }, + { + "epoch": 0.64, + "learning_rate": 0.0002942106318956871, + "loss": 1.4188, + "theoretical_loss": 3.4451100956957967, + "tokens_seen": 2099183616 + }, + { + "epoch": 0.64, + "learning_rate": 0.0002941945837512538, + "loss": 1.3131, + "theoretical_loss": 3.4451012866384034, + "tokens_seen": 2099249152 + }, + { + "epoch": 0.64, + "learning_rate": 0.0002941785356068205, + "loss": 1.3792, + "theoretical_loss": 3.445092477933014, + "tokens_seen": 2099314688 + }, + { + "epoch": 0.64, + "learning_rate": 0.0002941624874623872, + "loss": 1.3706, + "theoretical_loss": 3.445083669579602, + "tokens_seen": 2099380224 + }, + { + "epoch": 0.64, + "learning_rate": 0.0002941464393179539, + "loss": 1.3016, + "theoretical_loss": 3.4450748615781435, + "tokens_seen": 2099445760 + }, + { + "epoch": 0.64, + "learning_rate": 0.0002941303911735206, + "loss": 1.3658, + "theoretical_loss": 3.445066053928613, + "tokens_seen": 2099511296 + }, + { + "epoch": 0.64, + "learning_rate": 0.0002941143430290873, + "loss": 1.3507, + "theoretical_loss": 3.4450572466309852, + "tokens_seen": 2099576832 + }, + { + "epoch": 0.64, + "learning_rate": 0.000294098294884654, + "loss": 1.3068, + "theoretical_loss": 3.445048439685236, + "tokens_seen": 2099642368 + }, + { + "epoch": 0.64, + "learning_rate": 0.0002940822467402207, + "loss": 1.2395, + "theoretical_loss": 3.4450396330913398, + "tokens_seen": 2099707904 + }, + { + "epoch": 0.64, + "learning_rate": 0.0002940661985957874, + "loss": 1.3235, + "theoretical_loss": 3.4450308268492713, + "tokens_seen": 2099773440 + }, + { + "epoch": 0.64, + "learning_rate": 0.0002940501504513541, + "loss": 1.2755, + "theoretical_loss": 3.4450220209590054, + "tokens_seen": 2099838976 + }, + { + "epoch": 0.64, + "learning_rate": 0.0002940341023069208, + "loss": 1.1603, + "theoretical_loss": 3.4450132154205177, + "tokens_seen": 2099904512 + }, + { + "epoch": 0.64, + "learning_rate": 0.0002940180541624875, + "loss": 1.2549, + "theoretical_loss": 3.445004410233783, + "tokens_seen": 2099970048 + }, + { + "epoch": 0.64, + "learning_rate": 0.0002940020060180542, + "loss": 1.325, + "theoretical_loss": 3.444995605398776, + "tokens_seen": 2100035584 + }, + { + "epoch": 0.64, + "learning_rate": 0.0002939859578736209, + "loss": 1.336, + "theoretical_loss": 3.4449868009154714, + "tokens_seen": 2100101120 + }, + { + "epoch": 0.64, + "learning_rate": 0.0002939699097291876, + "loss": 1.4098, + "theoretical_loss": 3.444977996783845, + "tokens_seen": 2100166656 + }, + { + "epoch": 0.64, + "learning_rate": 0.0002939538615847543, + "loss": 1.4189, + "theoretical_loss": 3.444969193003871, + "tokens_seen": 2100232192 + }, + { + "epoch": 0.64, + "learning_rate": 0.000293937813440321, + "loss": 1.2589, + "theoretical_loss": 3.4449603895755248, + "tokens_seen": 2100297728 + }, + { + "epoch": 0.64, + "learning_rate": 0.0002939217652958877, + "loss": 1.3448, + "theoretical_loss": 3.4449515864987816, + "tokens_seen": 2100363264 + }, + { + "epoch": 0.64, + "objective/train/docs_used": 765555, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.1877882480621338, + "objective/train/theoretical_loss": 3.4449427837736155, + "objective/train/tokens_used": 2120888800, + "theoretical_loss": 3.4449427837736155, + "tokens_seen": 2100428800 + }, + { + "epoch": 0.64, + "learning_rate": 0.00029390571715145434, + "loss": 1.2621, + "theoretical_loss": 3.4449427837736155, + "tokens_seen": 2100428800 + }, + { + "epoch": 0.64, + "learning_rate": 0.00029388966900702104, + "loss": 1.1718, + "theoretical_loss": 3.4449339814000024, + "tokens_seen": 2100494336 + }, + { + "epoch": 0.64, + "learning_rate": 0.00029387362086258774, + "loss": 1.3319, + "theoretical_loss": 3.4449251793779174, + "tokens_seen": 2100559872 + }, + { + "epoch": 0.64, + "learning_rate": 0.00029385757271815444, + "loss": 1.3087, + "theoretical_loss": 3.4449163777073344, + "tokens_seen": 2100625408 + }, + { + "epoch": 0.64, + "learning_rate": 0.00029384152457372114, + "loss": 1.2781, + "theoretical_loss": 3.4449075763882293, + "tokens_seen": 2100690944 + }, + { + "epoch": 0.64, + "learning_rate": 0.00029382547642928784, + "loss": 1.2634, + "theoretical_loss": 3.444898775420577, + "tokens_seen": 2100756480 + }, + { + "epoch": 0.64, + "learning_rate": 0.00029380942828485454, + "loss": 1.271, + "theoretical_loss": 3.4448899748043518, + "tokens_seen": 2100822016 + }, + { + "epoch": 0.64, + "learning_rate": 0.00029379338014042124, + "loss": 1.2362, + "theoretical_loss": 3.44488117453953, + "tokens_seen": 2100887552 + }, + { + "epoch": 0.64, + "learning_rate": 0.00029377733199598794, + "loss": 1.2735, + "theoretical_loss": 3.4448723746260854, + "tokens_seen": 2100953088 + }, + { + "epoch": 0.64, + "learning_rate": 0.00029376128385155464, + "loss": 1.3205, + "theoretical_loss": 3.4448635750639935, + "tokens_seen": 2101018624 + }, + { + "epoch": 0.64, + "learning_rate": 0.00029374523570712134, + "loss": 1.3086, + "theoretical_loss": 3.44485477585323, + "tokens_seen": 2101084160 + }, + { + "epoch": 0.64, + "learning_rate": 0.00029372918756268804, + "loss": 1.2864, + "theoretical_loss": 3.4448459769937685, + "tokens_seen": 2101149696 + }, + { + "epoch": 0.64, + "learning_rate": 0.00029371313941825474, + "loss": 1.2677, + "theoretical_loss": 3.4448371784855847, + "tokens_seen": 2101215232 + }, + { + "epoch": 0.64, + "learning_rate": 0.00029369709127382144, + "loss": 1.2464, + "theoretical_loss": 3.4448283803286537, + "tokens_seen": 2101280768 + }, + { + "epoch": 0.64, + "learning_rate": 0.0002936810431293882, + "loss": 1.2687, + "theoretical_loss": 3.4448195825229506, + "tokens_seen": 2101346304 + }, + { + "epoch": 0.64, + "learning_rate": 0.0002936649949849549, + "loss": 1.2992, + "theoretical_loss": 3.4448107850684506, + "tokens_seen": 2101411840 + }, + { + "epoch": 0.64, + "learning_rate": 0.0002936489468405216, + "loss": 1.2097, + "theoretical_loss": 3.444801987965128, + "tokens_seen": 2101477376 + }, + { + "epoch": 0.64, + "learning_rate": 0.0002936328986960883, + "loss": 1.363, + "theoretical_loss": 3.444793191212958, + "tokens_seen": 2101542912 + }, + { + "epoch": 0.64, + "learning_rate": 0.000293616850551655, + "loss": 1.3032, + "theoretical_loss": 3.444784394811917, + "tokens_seen": 2101608448 + }, + { + "epoch": 0.64, + "learning_rate": 0.0002936008024072217, + "loss": 1.355, + "theoretical_loss": 3.4447755987619777, + "tokens_seen": 2101673984 + }, + { + "epoch": 0.64, + "learning_rate": 0.0002935847542627884, + "loss": 1.2869, + "theoretical_loss": 3.444766803063117, + "tokens_seen": 2101739520 + }, + { + "epoch": 0.64, + "learning_rate": 0.0002935687061183551, + "loss": 1.2362, + "theoretical_loss": 3.4447580077153095, + "tokens_seen": 2101805056 + }, + { + "epoch": 0.64, + "learning_rate": 0.0002935526579739218, + "loss": 1.3146, + "theoretical_loss": 3.4447492127185293, + "tokens_seen": 2101870592 + }, + { + "epoch": 0.64, + "learning_rate": 0.0002935366098294885, + "loss": 1.21, + "theoretical_loss": 3.4447404180727528, + "tokens_seen": 2101936128 + }, + { + "epoch": 0.64, + "learning_rate": 0.0002935205616850552, + "loss": 1.2958, + "theoretical_loss": 3.444731623777954, + "tokens_seen": 2102001664 + }, + { + "epoch": 0.64, + "learning_rate": 0.0002935045135406219, + "loss": 1.4428, + "theoretical_loss": 3.4447228298341086, + "tokens_seen": 2102067200 + }, + { + "epoch": 0.64, + "learning_rate": 0.0002934884653961886, + "loss": 1.3191, + "theoretical_loss": 3.4447140362411917, + "tokens_seen": 2102132736 + }, + { + "epoch": 0.64, + "learning_rate": 0.0002934724172517553, + "loss": 1.3162, + "theoretical_loss": 3.444705242999178, + "tokens_seen": 2102198272 + }, + { + "epoch": 0.64, + "learning_rate": 0.000293456369107322, + "loss": 1.2745, + "theoretical_loss": 3.4446964501080424, + "tokens_seen": 2102263808 + }, + { + "epoch": 0.64, + "learning_rate": 0.0002934403209628887, + "loss": 1.3149, + "theoretical_loss": 3.4446876575677603, + "tokens_seen": 2102329344 + }, + { + "epoch": 0.64, + "learning_rate": 0.0002934242728184554, + "loss": 1.2492, + "theoretical_loss": 3.444678865378307, + "tokens_seen": 2102394880 + }, + { + "epoch": 0.64, + "learning_rate": 0.0002934082246740221, + "loss": 1.2467, + "theoretical_loss": 3.444670073539657, + "tokens_seen": 2102460416 + }, + { + "epoch": 0.64, + "learning_rate": 0.0002933921765295888, + "loss": 1.324, + "theoretical_loss": 3.4446612820517855, + "tokens_seen": 2102525952 + }, + { + "epoch": 0.64, + "learning_rate": 0.0002933761283851555, + "loss": 1.2722, + "theoretical_loss": 3.4446524909146676, + "tokens_seen": 2102591488 + }, + { + "epoch": 0.64, + "learning_rate": 0.0002933600802407222, + "loss": 1.223, + "theoretical_loss": 3.444643700128279, + "tokens_seen": 2102657024 + }, + { + "epoch": 0.64, + "learning_rate": 0.0002933440320962889, + "loss": 1.2854, + "theoretical_loss": 3.444634909692594, + "tokens_seen": 2102722560 + }, + { + "epoch": 0.64, + "learning_rate": 0.0002933279839518556, + "loss": 1.2776, + "theoretical_loss": 3.444626119607588, + "tokens_seen": 2102788096 + }, + { + "epoch": 0.64, + "learning_rate": 0.0002933119358074223, + "loss": 1.316, + "theoretical_loss": 3.4446173298732354, + "tokens_seen": 2102853632 + }, + { + "epoch": 0.64, + "learning_rate": 0.000293295887662989, + "loss": 1.3289, + "theoretical_loss": 3.4446085404895124, + "tokens_seen": 2102919168 + }, + { + "epoch": 0.64, + "learning_rate": 0.00029327983951855566, + "loss": 1.2914, + "theoretical_loss": 3.4445997514563937, + "tokens_seen": 2102984704 + }, + { + "epoch": 0.64, + "learning_rate": 0.00029326379137412236, + "loss": 1.306, + "theoretical_loss": 3.4445909627738542, + "tokens_seen": 2103050240 + }, + { + "epoch": 0.64, + "learning_rate": 0.00029324774322968906, + "loss": 1.3529, + "theoretical_loss": 3.444582174441869, + "tokens_seen": 2103115776 + }, + { + "epoch": 0.64, + "learning_rate": 0.00029323169508525576, + "loss": 1.2349, + "theoretical_loss": 3.444573386460413, + "tokens_seen": 2103181312 + }, + { + "epoch": 0.64, + "learning_rate": 0.00029321564694082246, + "loss": 1.389, + "theoretical_loss": 3.444564598829462, + "tokens_seen": 2103246848 + }, + { + "epoch": 0.64, + "learning_rate": 0.00029319959879638916, + "loss": 1.28, + "theoretical_loss": 3.4445558115489905, + "tokens_seen": 2103312384 + }, + { + "epoch": 0.64, + "learning_rate": 0.00029318355065195586, + "loss": 1.3827, + "theoretical_loss": 3.4445470246189736, + "tokens_seen": 2103377920 + }, + { + "epoch": 0.64, + "learning_rate": 0.00029316750250752257, + "loss": 1.158, + "theoretical_loss": 3.4445382380393865, + "tokens_seen": 2103443456 + }, + { + "epoch": 0.64, + "learning_rate": 0.00029315145436308927, + "loss": 1.2308, + "theoretical_loss": 3.4445294518102045, + "tokens_seen": 2103508992 + }, + { + "epoch": 0.64, + "learning_rate": 0.00029313540621865597, + "loss": 1.3617, + "theoretical_loss": 3.4445206659314027, + "tokens_seen": 2103574528 + }, + { + "epoch": 0.64, + "learning_rate": 0.00029311935807422267, + "loss": 1.434, + "theoretical_loss": 3.4445118804029558, + "tokens_seen": 2103640064 + }, + { + "epoch": 0.64, + "objective/train/docs_used": 766743, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.4091877937316895, + "objective/train/theoretical_loss": 3.444503095224839, + "objective/train/tokens_used": 2124165600, + "theoretical_loss": 3.444503095224839, + "tokens_seen": 2103705600 + }, + { + "epoch": 0.64, + "learning_rate": 0.00029310330992978937, + "loss": 1.439, + "theoretical_loss": 3.444503095224839, + "tokens_seen": 2103705600 + }, + { + "epoch": 0.64, + "learning_rate": 0.00029308726178535607, + "loss": 1.3088, + "theoretical_loss": 3.4444943103970282, + "tokens_seen": 2103771136 + }, + { + "epoch": 0.64, + "learning_rate": 0.00029307121364092277, + "loss": 1.4625, + "theoretical_loss": 3.4444855259194975, + "tokens_seen": 2103836672 + }, + { + "epoch": 0.64, + "learning_rate": 0.00029305516549648947, + "loss": 1.1766, + "theoretical_loss": 3.444476741792223, + "tokens_seen": 2103902208 + }, + { + "epoch": 0.64, + "learning_rate": 0.00029303911735205617, + "loss": 1.3133, + "theoretical_loss": 3.4444679580151787, + "tokens_seen": 2103967744 + }, + { + "epoch": 0.64, + "learning_rate": 0.00029302306920762287, + "loss": 1.1887, + "theoretical_loss": 3.4444591745883404, + "tokens_seen": 2104033280 + }, + { + "epoch": 0.64, + "learning_rate": 0.00029300702106318957, + "loss": 1.2188, + "theoretical_loss": 3.4444503915116833, + "tokens_seen": 2104098816 + }, + { + "epoch": 0.64, + "learning_rate": 0.00029299097291875627, + "loss": 1.2868, + "theoretical_loss": 3.4444416087851826, + "tokens_seen": 2104164352 + }, + { + "epoch": 0.64, + "learning_rate": 0.000292974924774323, + "loss": 1.238, + "theoretical_loss": 3.4444328264088124, + "tokens_seen": 2104229888 + }, + { + "epoch": 0.64, + "learning_rate": 0.0002929588766298897, + "loss": 1.2179, + "theoretical_loss": 3.44442404438255, + "tokens_seen": 2104295424 + }, + { + "epoch": 0.64, + "learning_rate": 0.0002929428284854564, + "loss": 1.3212, + "theoretical_loss": 3.444415262706368, + "tokens_seen": 2104360960 + }, + { + "epoch": 0.64, + "learning_rate": 0.0002929267803410231, + "loss": 1.2768, + "theoretical_loss": 3.444406481380243, + "tokens_seen": 2104426496 + }, + { + "epoch": 0.64, + "learning_rate": 0.0002929107321965898, + "loss": 1.2377, + "theoretical_loss": 3.44439770040415, + "tokens_seen": 2104492032 + }, + { + "epoch": 0.64, + "learning_rate": 0.0002928946840521565, + "loss": 1.3014, + "theoretical_loss": 3.444388919778064, + "tokens_seen": 2104557568 + }, + { + "epoch": 0.64, + "learning_rate": 0.0002928786359077232, + "loss": 1.2744, + "theoretical_loss": 3.44438013950196, + "tokens_seen": 2104623104 + }, + { + "epoch": 0.64, + "learning_rate": 0.0002928625877632899, + "loss": 1.2484, + "theoretical_loss": 3.4443713595758134, + "tokens_seen": 2104688640 + }, + { + "epoch": 0.64, + "learning_rate": 0.0002928465396188566, + "loss": 1.3313, + "theoretical_loss": 3.4443625799995994, + "tokens_seen": 2104754176 + }, + { + "epoch": 0.64, + "learning_rate": 0.0002928304914744233, + "loss": 1.2537, + "theoretical_loss": 3.444353800773293, + "tokens_seen": 2104819712 + }, + { + "epoch": 0.64, + "learning_rate": 0.00029281444332999, + "loss": 1.3198, + "theoretical_loss": 3.4443450218968694, + "tokens_seen": 2104885248 + }, + { + "epoch": 0.64, + "learning_rate": 0.0002927983951855567, + "loss": 1.2795, + "theoretical_loss": 3.4443362433703038, + "tokens_seen": 2104950784 + }, + { + "epoch": 0.64, + "learning_rate": 0.0002927823470411234, + "loss": 1.4201, + "theoretical_loss": 3.444327465193571, + "tokens_seen": 2105016320 + }, + { + "epoch": 0.64, + "learning_rate": 0.0002927662988966901, + "loss": 1.1531, + "theoretical_loss": 3.444318687366647, + "tokens_seen": 2105081856 + }, + { + "epoch": 0.64, + "learning_rate": 0.0002927502507522568, + "loss": 1.2386, + "theoretical_loss": 3.4443099098895065, + "tokens_seen": 2105147392 + }, + { + "epoch": 0.64, + "learning_rate": 0.0002927342026078235, + "loss": 1.2793, + "theoretical_loss": 3.4443011327621242, + "tokens_seen": 2105212928 + }, + { + "epoch": 0.64, + "learning_rate": 0.0002927181544633902, + "loss": 1.238, + "theoretical_loss": 3.4442923559844756, + "tokens_seen": 2105278464 + }, + { + "epoch": 0.64, + "learning_rate": 0.0002927021063189569, + "loss": 1.3203, + "theoretical_loss": 3.4442835795565365, + "tokens_seen": 2105344000 + }, + { + "epoch": 0.64, + "learning_rate": 0.0002926860581745236, + "loss": 1.3796, + "theoretical_loss": 3.4442748034782813, + "tokens_seen": 2105409536 + }, + { + "epoch": 0.64, + "learning_rate": 0.0002926700100300903, + "loss": 1.2811, + "theoretical_loss": 3.444266027749685, + "tokens_seen": 2105475072 + }, + { + "epoch": 0.64, + "learning_rate": 0.000292653961885657, + "loss": 1.3151, + "theoretical_loss": 3.4442572523707238, + "tokens_seen": 2105540608 + }, + { + "epoch": 0.64, + "learning_rate": 0.0002926379137412237, + "loss": 1.3984, + "theoretical_loss": 3.4442484773413726, + "tokens_seen": 2105606144 + }, + { + "epoch": 0.64, + "learning_rate": 0.0002926218655967904, + "loss": 1.3693, + "theoretical_loss": 3.444239702661606, + "tokens_seen": 2105671680 + }, + { + "epoch": 0.64, + "learning_rate": 0.0002926058174523571, + "loss": 1.288, + "theoretical_loss": 3.4442309283313994, + "tokens_seen": 2105737216 + }, + { + "epoch": 0.64, + "learning_rate": 0.0002925897693079238, + "loss": 1.2574, + "theoretical_loss": 3.444222154350728, + "tokens_seen": 2105802752 + }, + { + "epoch": 0.64, + "learning_rate": 0.0002925737211634905, + "loss": 1.1771, + "theoretical_loss": 3.4442133807195674, + "tokens_seen": 2105868288 + }, + { + "epoch": 0.64, + "learning_rate": 0.0002925576730190572, + "loss": 1.3607, + "theoretical_loss": 3.444204607437892, + "tokens_seen": 2105933824 + }, + { + "epoch": 0.64, + "learning_rate": 0.0002925416248746239, + "loss": 1.3174, + "theoretical_loss": 3.4441958345056785, + "tokens_seen": 2105999360 + }, + { + "epoch": 0.64, + "learning_rate": 0.0002925255767301906, + "loss": 1.221, + "theoretical_loss": 3.4441870619229, + "tokens_seen": 2106064896 + }, + { + "epoch": 0.64, + "learning_rate": 0.00029250952858575724, + "loss": 1.2485, + "theoretical_loss": 3.4441782896895337, + "tokens_seen": 2106130432 + }, + { + "epoch": 0.64, + "learning_rate": 0.00029249348044132394, + "loss": 1.3084, + "theoretical_loss": 3.4441695178055536, + "tokens_seen": 2106195968 + }, + { + "epoch": 0.64, + "learning_rate": 0.0002924774322968907, + "loss": 1.3029, + "theoretical_loss": 3.444160746270935, + "tokens_seen": 2106261504 + }, + { + "epoch": 0.64, + "learning_rate": 0.0002924613841524574, + "loss": 1.3359, + "theoretical_loss": 3.4441519750856537, + "tokens_seen": 2106327040 + }, + { + "epoch": 0.64, + "learning_rate": 0.0002924453360080241, + "loss": 1.2562, + "theoretical_loss": 3.4441432042496842, + "tokens_seen": 2106392576 + }, + { + "epoch": 0.64, + "learning_rate": 0.0002924292878635908, + "loss": 1.4302, + "theoretical_loss": 3.4441344337630024, + "tokens_seen": 2106458112 + }, + { + "epoch": 0.64, + "learning_rate": 0.0002924132397191575, + "loss": 1.402, + "theoretical_loss": 3.4441256636255835, + "tokens_seen": 2106523648 + }, + { + "epoch": 0.64, + "learning_rate": 0.0002923971915747242, + "loss": 1.3421, + "theoretical_loss": 3.444116893837402, + "tokens_seen": 2106589184 + }, + { + "epoch": 0.64, + "learning_rate": 0.0002923811434302909, + "loss": 1.2814, + "theoretical_loss": 3.4441081243984337, + "tokens_seen": 2106654720 + }, + { + "epoch": 0.64, + "learning_rate": 0.0002923650952858576, + "loss": 1.4145, + "theoretical_loss": 3.444099355308654, + "tokens_seen": 2106720256 + }, + { + "epoch": 0.64, + "learning_rate": 0.0002923490471414243, + "loss": 1.3646, + "theoretical_loss": 3.444090586568038, + "tokens_seen": 2106785792 + }, + { + "epoch": 0.64, + "learning_rate": 0.000292332998996991, + "loss": 1.3547, + "theoretical_loss": 3.44408181817656, + "tokens_seen": 2106851328 + }, + { + "epoch": 0.64, + "learning_rate": 0.0002923169508525577, + "loss": 1.3916, + "theoretical_loss": 3.4440730501341967, + "tokens_seen": 2106916864 + }, + { + "epoch": 0.64, + "objective/train/docs_used": 767939, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.3456382751464844, + "objective/train/theoretical_loss": 3.444064282440922, + "objective/train/tokens_used": 2127442400, + "theoretical_loss": 3.444064282440922, + "tokens_seen": 2106982400 + }, + { + "epoch": 0.64, + "learning_rate": 0.0002923009027081244, + "loss": 1.3481, + "theoretical_loss": 3.444064282440922, + "tokens_seen": 2106982400 + }, + { + "epoch": 0.64, + "learning_rate": 0.0002922848545636911, + "loss": 1.3594, + "theoretical_loss": 3.4440555150967125, + "tokens_seen": 2107047936 + }, + { + "epoch": 0.64, + "learning_rate": 0.0002922688064192578, + "loss": 1.362, + "theoretical_loss": 3.444046748101542, + "tokens_seen": 2107113472 + }, + { + "epoch": 0.64, + "learning_rate": 0.0002922527582748245, + "loss": 1.3122, + "theoretical_loss": 3.444037981455387, + "tokens_seen": 2107179008 + }, + { + "epoch": 0.64, + "learning_rate": 0.0002922367101303912, + "loss": 1.3486, + "theoretical_loss": 3.4440292151582224, + "tokens_seen": 2107244544 + }, + { + "epoch": 0.64, + "learning_rate": 0.0002922206619859579, + "loss": 1.4271, + "theoretical_loss": 3.444020449210023, + "tokens_seen": 2107310080 + }, + { + "epoch": 0.64, + "learning_rate": 0.0002922046138415246, + "loss": 1.2492, + "theoretical_loss": 3.4440116836107646, + "tokens_seen": 2107375616 + }, + { + "epoch": 0.64, + "learning_rate": 0.0002921885656970913, + "loss": 1.4086, + "theoretical_loss": 3.444002918360422, + "tokens_seen": 2107441152 + }, + { + "epoch": 0.64, + "learning_rate": 0.000292172517552658, + "loss": 1.3217, + "theoretical_loss": 3.443994153458971, + "tokens_seen": 2107506688 + }, + { + "epoch": 0.64, + "learning_rate": 0.0002921564694082247, + "loss": 1.2822, + "theoretical_loss": 3.4439853889063863, + "tokens_seen": 2107572224 + }, + { + "epoch": 0.64, + "learning_rate": 0.0002921404212637914, + "loss": 1.427, + "theoretical_loss": 3.4439766247026435, + "tokens_seen": 2107637760 + }, + { + "epoch": 0.64, + "learning_rate": 0.0002921243731193581, + "loss": 1.2665, + "theoretical_loss": 3.4439678608477178, + "tokens_seen": 2107703296 + }, + { + "epoch": 0.64, + "learning_rate": 0.0002921083249749248, + "loss": 1.3068, + "theoretical_loss": 3.443959097341584, + "tokens_seen": 2107768832 + }, + { + "epoch": 0.64, + "learning_rate": 0.0002920922768304915, + "loss": 1.3153, + "theoretical_loss": 3.4439503341842186, + "tokens_seen": 2107834368 + }, + { + "epoch": 0.64, + "learning_rate": 0.0002920762286860582, + "loss": 1.3676, + "theoretical_loss": 3.443941571375596, + "tokens_seen": 2107899904 + }, + { + "epoch": 0.64, + "learning_rate": 0.0002920601805416249, + "loss": 1.3687, + "theoretical_loss": 3.443932808915691, + "tokens_seen": 2107965440 + }, + { + "epoch": 0.64, + "learning_rate": 0.0002920441323971916, + "loss": 1.3076, + "theoretical_loss": 3.4439240468044803, + "tokens_seen": 2108030976 + }, + { + "epoch": 0.64, + "learning_rate": 0.0002920280842527583, + "loss": 1.2477, + "theoretical_loss": 3.4439152850419377, + "tokens_seen": 2108096512 + }, + { + "epoch": 0.64, + "learning_rate": 0.000292012036108325, + "loss": 1.2718, + "theoretical_loss": 3.4439065236280393, + "tokens_seen": 2108162048 + }, + { + "epoch": 0.64, + "learning_rate": 0.0002919959879638917, + "loss": 1.3501, + "theoretical_loss": 3.4438977625627603, + "tokens_seen": 2108227584 + }, + { + "epoch": 0.64, + "learning_rate": 0.0002919799398194584, + "loss": 1.3735, + "theoretical_loss": 3.4438890018460757, + "tokens_seen": 2108293120 + }, + { + "epoch": 0.64, + "learning_rate": 0.0002919638916750251, + "loss": 1.4155, + "theoretical_loss": 3.4438802414779612, + "tokens_seen": 2108358656 + }, + { + "epoch": 0.64, + "learning_rate": 0.0002919478435305918, + "loss": 1.3143, + "theoretical_loss": 3.443871481458392, + "tokens_seen": 2108424192 + }, + { + "epoch": 0.64, + "learning_rate": 0.0002919317953861585, + "loss": 1.3035, + "theoretical_loss": 3.4438627217873434, + "tokens_seen": 2108489728 + }, + { + "epoch": 0.64, + "learning_rate": 0.0002919157472417252, + "loss": 1.3561, + "theoretical_loss": 3.4438539624647904, + "tokens_seen": 2108555264 + }, + { + "epoch": 0.64, + "learning_rate": 0.0002918996990972919, + "loss": 1.2838, + "theoretical_loss": 3.4438452034907083, + "tokens_seen": 2108620800 + }, + { + "epoch": 0.64, + "learning_rate": 0.0002918836509528586, + "loss": 1.3219, + "theoretical_loss": 3.443836444865073, + "tokens_seen": 2108686336 + }, + { + "epoch": 0.64, + "learning_rate": 0.00029186760280842526, + "loss": 1.3323, + "theoretical_loss": 3.443827686587859, + "tokens_seen": 2108751872 + }, + { + "epoch": 0.64, + "learning_rate": 0.00029185155466399196, + "loss": 1.2786, + "theoretical_loss": 3.443818928659043, + "tokens_seen": 2108817408 + }, + { + "epoch": 0.64, + "learning_rate": 0.00029183550651955866, + "loss": 1.3652, + "theoretical_loss": 3.4438101710785984, + "tokens_seen": 2108882944 + }, + { + "epoch": 0.64, + "learning_rate": 0.00029181945837512536, + "loss": 1.3276, + "theoretical_loss": 3.443801413846502, + "tokens_seen": 2108948480 + }, + { + "epoch": 0.64, + "learning_rate": 0.00029180341023069207, + "loss": 1.275, + "theoretical_loss": 3.4437926569627284, + "tokens_seen": 2109014016 + }, + { + "epoch": 0.64, + "learning_rate": 0.00029178736208625877, + "loss": 1.3484, + "theoretical_loss": 3.4437839004272535, + "tokens_seen": 2109079552 + }, + { + "epoch": 0.64, + "learning_rate": 0.00029177131394182547, + "loss": 1.3133, + "theoretical_loss": 3.4437751442400515, + "tokens_seen": 2109145088 + }, + { + "epoch": 0.64, + "learning_rate": 0.00029175526579739217, + "loss": 1.3716, + "theoretical_loss": 3.443766388401099, + "tokens_seen": 2109210624 + }, + { + "epoch": 0.64, + "learning_rate": 0.00029173921765295887, + "loss": 1.3929, + "theoretical_loss": 3.443757632910371, + "tokens_seen": 2109276160 + }, + { + "epoch": 0.64, + "learning_rate": 0.00029172316950852557, + "loss": 1.2671, + "theoretical_loss": 3.4437488777678418, + "tokens_seen": 2109341696 + }, + { + "epoch": 0.64, + "learning_rate": 0.00029170712136409227, + "loss": 1.2831, + "theoretical_loss": 3.4437401229734883, + "tokens_seen": 2109407232 + }, + { + "epoch": 0.64, + "learning_rate": 0.00029169107321965897, + "loss": 1.2247, + "theoretical_loss": 3.443731368527285, + "tokens_seen": 2109472768 + }, + { + "epoch": 0.64, + "learning_rate": 0.00029167502507522567, + "loss": 1.3598, + "theoretical_loss": 3.4437226144292072, + "tokens_seen": 2109538304 + }, + { + "epoch": 0.64, + "learning_rate": 0.00029165897693079237, + "loss": 1.5248, + "theoretical_loss": 3.4437138606792304, + "tokens_seen": 2109603840 + }, + { + "epoch": 0.64, + "learning_rate": 0.00029164292878635907, + "loss": 1.3411, + "theoretical_loss": 3.4437051072773297, + "tokens_seen": 2109669376 + }, + { + "epoch": 0.64, + "learning_rate": 0.00029162688064192577, + "loss": 1.2383, + "theoretical_loss": 3.4436963542234813, + "tokens_seen": 2109734912 + }, + { + "epoch": 0.64, + "learning_rate": 0.0002916108324974925, + "loss": 1.42, + "theoretical_loss": 3.4436876015176594, + "tokens_seen": 2109800448 + }, + { + "epoch": 0.64, + "learning_rate": 0.0002915947843530592, + "loss": 1.3357, + "theoretical_loss": 3.4436788491598405, + "tokens_seen": 2109865984 + }, + { + "epoch": 0.64, + "learning_rate": 0.0002915787362086259, + "loss": 1.2698, + "theoretical_loss": 3.4436700971499983, + "tokens_seen": 2109931520 + }, + { + "epoch": 0.64, + "learning_rate": 0.0002915626880641926, + "loss": 1.2108, + "theoretical_loss": 3.4436613454881098, + "tokens_seen": 2109997056 + }, + { + "epoch": 0.64, + "learning_rate": 0.0002915466399197593, + "loss": 1.2729, + "theoretical_loss": 3.44365259417415, + "tokens_seen": 2110062592 + }, + { + "epoch": 0.64, + "learning_rate": 0.000291530591775326, + "loss": 1.349, + "theoretical_loss": 3.443643843208094, + "tokens_seen": 2110128128 + }, + { + "epoch": 0.64, + "learning_rate": 0.0002915145436308927, + "loss": 1.3599, + "theoretical_loss": 3.4436350925899166, + "tokens_seen": 2110193664 + }, + { + "epoch": 0.64, + "objective/train/docs_used": 769058, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.1890474557876587, + "objective/train/theoretical_loss": 3.443626342319594, + "objective/train/tokens_used": 2130719200, + "theoretical_loss": 3.443626342319594, + "tokens_seen": 2110259200 + }, + { + "epoch": 0.64, + "learning_rate": 0.0002914984954864594, + "loss": 1.3197, + "theoretical_loss": 3.443626342319594, + "tokens_seen": 2110259200 + }, + { + "epoch": 0.64, + "learning_rate": 0.0002914824473420261, + "loss": 1.3093, + "theoretical_loss": 3.4436175923971017, + "tokens_seen": 2110324736 + }, + { + "epoch": 0.64, + "learning_rate": 0.0002914663991975928, + "loss": 1.3245, + "theoretical_loss": 3.4436088428224143, + "tokens_seen": 2110390272 + }, + { + "epoch": 0.64, + "learning_rate": 0.0002914503510531595, + "loss": 1.3413, + "theoretical_loss": 3.443600093595508, + "tokens_seen": 2110455808 + }, + { + "epoch": 0.64, + "learning_rate": 0.0002914343029087262, + "loss": 1.3581, + "theoretical_loss": 3.4435913447163573, + "tokens_seen": 2110521344 + }, + { + "epoch": 0.64, + "learning_rate": 0.0002914182547642929, + "loss": 1.2607, + "theoretical_loss": 3.443582596184938, + "tokens_seen": 2110586880 + }, + { + "epoch": 0.64, + "learning_rate": 0.0002914022066198596, + "loss": 1.4804, + "theoretical_loss": 3.4435738480012255, + "tokens_seen": 2110652416 + }, + { + "epoch": 0.64, + "learning_rate": 0.0002913861584754263, + "loss": 1.2866, + "theoretical_loss": 3.4435651001651957, + "tokens_seen": 2110717952 + }, + { + "epoch": 0.64, + "learning_rate": 0.000291370110330993, + "loss": 1.2207, + "theoretical_loss": 3.443556352676823, + "tokens_seen": 2110783488 + }, + { + "epoch": 0.64, + "learning_rate": 0.0002913540621865597, + "loss": 1.1857, + "theoretical_loss": 3.443547605536083, + "tokens_seen": 2110849024 + }, + { + "epoch": 0.64, + "learning_rate": 0.0002913380140421264, + "loss": 1.3607, + "theoretical_loss": 3.443538858742952, + "tokens_seen": 2110914560 + }, + { + "epoch": 0.64, + "learning_rate": 0.0002913219658976931, + "loss": 1.2519, + "theoretical_loss": 3.4435301122974047, + "tokens_seen": 2110980096 + }, + { + "epoch": 0.64, + "learning_rate": 0.0002913059177532598, + "loss": 1.2222, + "theoretical_loss": 3.443521366199416, + "tokens_seen": 2111045632 + }, + { + "epoch": 0.64, + "learning_rate": 0.00029128986960882654, + "loss": 1.2828, + "theoretical_loss": 3.4435126204489626, + "tokens_seen": 2111111168 + }, + { + "epoch": 0.64, + "learning_rate": 0.00029127382146439324, + "loss": 1.2705, + "theoretical_loss": 3.4435038750460185, + "tokens_seen": 2111176704 + }, + { + "epoch": 0.64, + "learning_rate": 0.00029125777331995994, + "loss": 1.2283, + "theoretical_loss": 3.44349512999056, + "tokens_seen": 2111242240 + }, + { + "epoch": 0.64, + "learning_rate": 0.0002912417251755266, + "loss": 1.2631, + "theoretical_loss": 3.4434863852825623, + "tokens_seen": 2111307776 + }, + { + "epoch": 0.64, + "learning_rate": 0.0002912256770310933, + "loss": 1.3096, + "theoretical_loss": 3.443477640922001, + "tokens_seen": 2111373312 + }, + { + "epoch": 0.64, + "learning_rate": 0.00029120962888666, + "loss": 1.3782, + "theoretical_loss": 3.4434688969088514, + "tokens_seen": 2111438848 + }, + { + "epoch": 0.64, + "learning_rate": 0.0002911935807422267, + "loss": 1.3234, + "theoretical_loss": 3.4434601532430884, + "tokens_seen": 2111504384 + }, + { + "epoch": 0.64, + "learning_rate": 0.0002911775325977934, + "loss": 1.2725, + "theoretical_loss": 3.4434514099246876, + "tokens_seen": 2111569920 + }, + { + "epoch": 0.64, + "learning_rate": 0.0002911614844533601, + "loss": 1.3493, + "theoretical_loss": 3.443442666953625, + "tokens_seen": 2111635456 + }, + { + "epoch": 0.64, + "learning_rate": 0.0002911454363089268, + "loss": 1.3883, + "theoretical_loss": 3.4434339243298755, + "tokens_seen": 2111700992 + }, + { + "epoch": 0.64, + "learning_rate": 0.0002911293881644935, + "loss": 1.369, + "theoretical_loss": 3.4434251820534145, + "tokens_seen": 2111766528 + }, + { + "epoch": 0.64, + "learning_rate": 0.0002911133400200602, + "loss": 1.3171, + "theoretical_loss": 3.443416440124218, + "tokens_seen": 2111832064 + }, + { + "epoch": 0.64, + "learning_rate": 0.0002910972918756269, + "loss": 1.3037, + "theoretical_loss": 3.443407698542261, + "tokens_seen": 2111897600 + }, + { + "epoch": 0.64, + "learning_rate": 0.0002910812437311936, + "loss": 1.2776, + "theoretical_loss": 3.443398957307519, + "tokens_seen": 2111963136 + }, + { + "epoch": 0.64, + "learning_rate": 0.0002910651955867603, + "loss": 1.3331, + "theoretical_loss": 3.4433902164199672, + "tokens_seen": 2112028672 + }, + { + "epoch": 0.64, + "learning_rate": 0.000291049147442327, + "loss": 1.3471, + "theoretical_loss": 3.4433814758795815, + "tokens_seen": 2112094208 + }, + { + "epoch": 0.64, + "learning_rate": 0.0002910330992978937, + "loss": 1.3017, + "theoretical_loss": 3.443372735686337, + "tokens_seen": 2112159744 + }, + { + "epoch": 0.64, + "learning_rate": 0.0002910170511534604, + "loss": 1.3466, + "theoretical_loss": 3.443363995840209, + "tokens_seen": 2112225280 + }, + { + "epoch": 0.64, + "learning_rate": 0.0002910010030090271, + "loss": 1.1883, + "theoretical_loss": 3.4433552563411736, + "tokens_seen": 2112290816 + }, + { + "epoch": 0.64, + "learning_rate": 0.0002909849548645938, + "loss": 1.2883, + "theoretical_loss": 3.4433465171892057, + "tokens_seen": 2112356352 + }, + { + "epoch": 0.64, + "learning_rate": 0.0002909689067201605, + "loss": 1.3399, + "theoretical_loss": 3.443337778384281, + "tokens_seen": 2112421888 + }, + { + "epoch": 0.64, + "learning_rate": 0.0002909528585757272, + "loss": 1.3346, + "theoretical_loss": 3.4433290399263745, + "tokens_seen": 2112487424 + }, + { + "epoch": 0.64, + "learning_rate": 0.0002909368104312939, + "loss": 1.2708, + "theoretical_loss": 3.4433203018154623, + "tokens_seen": 2112552960 + }, + { + "epoch": 0.64, + "learning_rate": 0.0002909207622868606, + "loss": 1.3638, + "theoretical_loss": 3.443311564051519, + "tokens_seen": 2112618496 + }, + { + "epoch": 0.64, + "learning_rate": 0.0002909047141424273, + "loss": 1.4196, + "theoretical_loss": 3.443302826634521, + "tokens_seen": 2112684032 + }, + { + "epoch": 0.64, + "learning_rate": 0.000290888665997994, + "loss": 1.2641, + "theoretical_loss": 3.4432940895644433, + "tokens_seen": 2112749568 + }, + { + "epoch": 0.64, + "learning_rate": 0.0002908726178535607, + "loss": 1.2508, + "theoretical_loss": 3.4432853528412615, + "tokens_seen": 2112815104 + }, + { + "epoch": 0.64, + "learning_rate": 0.0002908565697091274, + "loss": 1.3148, + "theoretical_loss": 3.443276616464951, + "tokens_seen": 2112880640 + }, + { + "epoch": 0.64, + "learning_rate": 0.0002908405215646941, + "loss": 1.3674, + "theoretical_loss": 3.443267880435487, + "tokens_seen": 2112946176 + }, + { + "epoch": 0.64, + "learning_rate": 0.0002908244734202608, + "loss": 1.2768, + "theoretical_loss": 3.4432591447528456, + "tokens_seen": 2113011712 + }, + { + "epoch": 0.64, + "learning_rate": 0.0002908084252758275, + "loss": 1.34, + "theoretical_loss": 3.4432504094170016, + "tokens_seen": 2113077248 + }, + { + "epoch": 0.64, + "learning_rate": 0.0002907923771313942, + "loss": 1.2979, + "theoretical_loss": 3.443241674427931, + "tokens_seen": 2113142784 + }, + { + "epoch": 0.64, + "learning_rate": 0.0002907763289869609, + "loss": 1.3255, + "theoretical_loss": 3.443232939785609, + "tokens_seen": 2113208320 + }, + { + "epoch": 0.64, + "learning_rate": 0.0002907602808425276, + "loss": 1.1011, + "theoretical_loss": 3.443224205490011, + "tokens_seen": 2113273856 + }, + { + "epoch": 0.64, + "learning_rate": 0.0002907442326980943, + "loss": 1.3015, + "theoretical_loss": 3.443215471541113, + "tokens_seen": 2113339392 + }, + { + "epoch": 0.64, + "learning_rate": 0.000290728184553661, + "loss": 1.3944, + "theoretical_loss": 3.44320673793889, + "tokens_seen": 2113404928 + }, + { + "epoch": 0.64, + "learning_rate": 0.0002907121364092277, + "loss": 1.2436, + "theoretical_loss": 3.4431980046833175, + "tokens_seen": 2113470464 + }, + { + "epoch": 0.64, + "objective/train/docs_used": 770311, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.2672467231750488, + "objective/train/theoretical_loss": 3.4431892717743713, + "objective/train/tokens_used": 2133996000, + "theoretical_loss": 3.4431892717743713, + "tokens_seen": 2113536000 + }, + { + "epoch": 0.64, + "learning_rate": 0.0002906960882647944, + "loss": 1.2357, + "theoretical_loss": 3.4431892717743713, + "tokens_seen": 2113536000 + }, + { + "epoch": 0.64, + "learning_rate": 0.0002906800401203611, + "loss": 1.3174, + "theoretical_loss": 3.443180539212026, + "tokens_seen": 2113601536 + }, + { + "epoch": 0.64, + "learning_rate": 0.0002906639919759278, + "loss": 1.2791, + "theoretical_loss": 3.4431718069962587, + "tokens_seen": 2113667072 + }, + { + "epoch": 0.64, + "learning_rate": 0.0002906479438314945, + "loss": 1.2428, + "theoretical_loss": 3.4431630751270434, + "tokens_seen": 2113732608 + }, + { + "epoch": 0.64, + "learning_rate": 0.0002906318956870612, + "loss": 1.3552, + "theoretical_loss": 3.4431543436043563, + "tokens_seen": 2113798144 + }, + { + "epoch": 0.64, + "learning_rate": 0.0002906158475426279, + "loss": 1.4474, + "theoretical_loss": 3.4431456124281734, + "tokens_seen": 2113863680 + }, + { + "epoch": 0.64, + "learning_rate": 0.0002905997993981946, + "loss": 1.2855, + "theoretical_loss": 3.443136881598469, + "tokens_seen": 2113929216 + }, + { + "epoch": 0.64, + "learning_rate": 0.0002905837512537613, + "loss": 1.2959, + "theoretical_loss": 3.443128151115219, + "tokens_seen": 2113994752 + }, + { + "epoch": 0.64, + "learning_rate": 0.000290567703109328, + "loss": 1.3273, + "theoretical_loss": 3.4431194209784, + "tokens_seen": 2114060288 + }, + { + "epoch": 0.64, + "learning_rate": 0.0002905516549648947, + "loss": 1.3993, + "theoretical_loss": 3.443110691187986, + "tokens_seen": 2114125824 + }, + { + "epoch": 0.64, + "learning_rate": 0.0002905356068204614, + "loss": 1.3122, + "theoretical_loss": 3.4431019617439533, + "tokens_seen": 2114191360 + }, + { + "epoch": 0.64, + "learning_rate": 0.0002905195586760281, + "loss": 1.3012, + "theoretical_loss": 3.4430932326462775, + "tokens_seen": 2114256896 + }, + { + "epoch": 0.64, + "learning_rate": 0.0002905035105315948, + "loss": 1.2453, + "theoretical_loss": 3.4430845038949336, + "tokens_seen": 2114322432 + }, + { + "epoch": 0.64, + "learning_rate": 0.0002904874623871615, + "loss": 1.3953, + "theoretical_loss": 3.4430757754898975, + "tokens_seen": 2114387968 + }, + { + "epoch": 0.64, + "learning_rate": 0.00029047141424272816, + "loss": 1.2784, + "theoretical_loss": 3.4430670474311444, + "tokens_seen": 2114453504 + }, + { + "epoch": 0.64, + "learning_rate": 0.00029045536609829486, + "loss": 1.444, + "theoretical_loss": 3.4430583197186504, + "tokens_seen": 2114519040 + }, + { + "epoch": 0.64, + "learning_rate": 0.00029043931795386157, + "loss": 1.4225, + "theoretical_loss": 3.443049592352391, + "tokens_seen": 2114584576 + }, + { + "epoch": 0.64, + "learning_rate": 0.00029042326980942827, + "loss": 1.3514, + "theoretical_loss": 3.443040865332341, + "tokens_seen": 2114650112 + }, + { + "epoch": 0.64, + "learning_rate": 0.00029040722166499497, + "loss": 1.3478, + "theoretical_loss": 3.443032138658477, + "tokens_seen": 2114715648 + }, + { + "epoch": 0.64, + "learning_rate": 0.00029039117352056167, + "loss": 1.3194, + "theoretical_loss": 3.4430234123307732, + "tokens_seen": 2114781184 + }, + { + "epoch": 0.64, + "learning_rate": 0.00029037512537612837, + "loss": 1.361, + "theoretical_loss": 3.443014686349206, + "tokens_seen": 2114846720 + }, + { + "epoch": 0.64, + "learning_rate": 0.00029035907723169507, + "loss": 1.2555, + "theoretical_loss": 3.4430059607137515, + "tokens_seen": 2114912256 + }, + { + "epoch": 0.64, + "learning_rate": 0.00029034302908726177, + "loss": 1.4262, + "theoretical_loss": 3.442997235424384, + "tokens_seen": 2114977792 + }, + { + "epoch": 0.64, + "learning_rate": 0.00029032698094282847, + "loss": 1.3649, + "theoretical_loss": 3.4429885104810793, + "tokens_seen": 2115043328 + }, + { + "epoch": 0.64, + "learning_rate": 0.00029031093279839517, + "loss": 1.2885, + "theoretical_loss": 3.442979785883814, + "tokens_seen": 2115108864 + }, + { + "epoch": 0.64, + "learning_rate": 0.00029029488465396187, + "loss": 1.2813, + "theoretical_loss": 3.442971061632562, + "tokens_seen": 2115174400 + }, + { + "epoch": 0.64, + "learning_rate": 0.00029027883650952857, + "loss": 1.288, + "theoretical_loss": 3.442962337727301, + "tokens_seen": 2115239936 + }, + { + "epoch": 0.64, + "learning_rate": 0.00029026278836509527, + "loss": 1.4021, + "theoretical_loss": 3.4429536141680046, + "tokens_seen": 2115305472 + }, + { + "epoch": 0.64, + "learning_rate": 0.00029024674022066197, + "loss": 1.252, + "theoretical_loss": 3.442944890954649, + "tokens_seen": 2115371008 + }, + { + "epoch": 0.64, + "learning_rate": 0.0002902306920762287, + "loss": 1.2798, + "theoretical_loss": 3.4429361680872104, + "tokens_seen": 2115436544 + }, + { + "epoch": 0.64, + "learning_rate": 0.0002902146439317954, + "loss": 1.3884, + "theoretical_loss": 3.442927445565663, + "tokens_seen": 2115502080 + }, + { + "epoch": 0.64, + "learning_rate": 0.0002901985957873621, + "loss": 1.2833, + "theoretical_loss": 3.442918723389984, + "tokens_seen": 2115567616 + }, + { + "epoch": 0.64, + "learning_rate": 0.0002901825476429288, + "loss": 1.3919, + "theoretical_loss": 3.442910001560148, + "tokens_seen": 2115633152 + }, + { + "epoch": 0.64, + "learning_rate": 0.0002901664994984955, + "loss": 1.429, + "theoretical_loss": 3.4429012800761303, + "tokens_seen": 2115698688 + }, + { + "epoch": 0.64, + "learning_rate": 0.0002901504513540622, + "loss": 1.2256, + "theoretical_loss": 3.442892558937907, + "tokens_seen": 2115764224 + }, + { + "epoch": 0.64, + "learning_rate": 0.0002901344032096289, + "loss": 1.2752, + "theoretical_loss": 3.4428838381454545, + "tokens_seen": 2115829760 + }, + { + "epoch": 0.64, + "learning_rate": 0.0002901183550651956, + "loss": 1.2736, + "theoretical_loss": 3.4428751176987467, + "tokens_seen": 2115895296 + }, + { + "epoch": 0.64, + "learning_rate": 0.0002901023069207623, + "loss": 1.2499, + "theoretical_loss": 3.4428663975977596, + "tokens_seen": 2115960832 + }, + { + "epoch": 0.64, + "learning_rate": 0.00029008625877632903, + "loss": 1.2447, + "theoretical_loss": 3.44285767784247, + "tokens_seen": 2116026368 + }, + { + "epoch": 0.64, + "learning_rate": 0.00029007021063189573, + "loss": 1.2961, + "theoretical_loss": 3.4428489584328523, + "tokens_seen": 2116091904 + }, + { + "epoch": 0.64, + "learning_rate": 0.00029005416248746243, + "loss": 1.23, + "theoretical_loss": 3.4428402393688824, + "tokens_seen": 2116157440 + }, + { + "epoch": 0.64, + "learning_rate": 0.00029003811434302914, + "loss": 1.3096, + "theoretical_loss": 3.442831520650536, + "tokens_seen": 2116222976 + }, + { + "epoch": 0.64, + "learning_rate": 0.00029002206619859584, + "loss": 1.1914, + "theoretical_loss": 3.4428228022777887, + "tokens_seen": 2116288512 + }, + { + "epoch": 0.64, + "learning_rate": 0.00029000601805416254, + "loss": 1.3036, + "theoretical_loss": 3.442814084250616, + "tokens_seen": 2116354048 + }, + { + "epoch": 0.64, + "learning_rate": 0.00028998996990972924, + "loss": 1.2284, + "theoretical_loss": 3.4428053665689937, + "tokens_seen": 2116419584 + }, + { + "epoch": 0.64, + "learning_rate": 0.00028997392176529594, + "loss": 1.2929, + "theoretical_loss": 3.4427966492328967, + "tokens_seen": 2116485120 + }, + { + "epoch": 0.64, + "learning_rate": 0.00028995787362086264, + "loss": 1.3515, + "theoretical_loss": 3.4427879322423016, + "tokens_seen": 2116550656 + }, + { + "epoch": 0.64, + "learning_rate": 0.00028994182547642934, + "loss": 1.2467, + "theoretical_loss": 3.4427792155971835, + "tokens_seen": 2116616192 + }, + { + "epoch": 0.64, + "learning_rate": 0.00028992577733199604, + "loss": 1.3562, + "theoretical_loss": 3.442770499297518, + "tokens_seen": 2116681728 + }, + { + "epoch": 0.64, + "learning_rate": 0.00028990972918756274, + "loss": 1.2079, + "theoretical_loss": 3.442761783343281, + "tokens_seen": 2116747264 + }, + { + "epoch": 0.64, + "objective/train/docs_used": 771474, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.2582566738128662, + "objective/train/theoretical_loss": 3.442753067734447, + "objective/train/tokens_used": 2137272800, + "theoretical_loss": 3.442753067734447, + "tokens_seen": 2116812800 + }, + { + "epoch": 0.64, + "learning_rate": 0.00028989368104312944, + "loss": 1.2961, + "theoretical_loss": 3.442753067734447, + "tokens_seen": 2116812800 + }, + { + "epoch": 0.64, + "learning_rate": 0.00028987763289869614, + "loss": 1.2753, + "theoretical_loss": 3.4427443524709935, + "tokens_seen": 2116878336 + }, + { + "epoch": 0.64, + "learning_rate": 0.00028986158475426284, + "loss": 1.3487, + "theoretical_loss": 3.442735637552895, + "tokens_seen": 2116943872 + }, + { + "epoch": 0.64, + "learning_rate": 0.0002898455366098295, + "loss": 1.2828, + "theoretical_loss": 3.442726922980127, + "tokens_seen": 2117009408 + }, + { + "epoch": 0.64, + "learning_rate": 0.0002898294884653962, + "loss": 1.3401, + "theoretical_loss": 3.442718208752665, + "tokens_seen": 2117074944 + }, + { + "epoch": 0.64, + "learning_rate": 0.0002898134403209629, + "loss": 1.2949, + "theoretical_loss": 3.4427094948704857, + "tokens_seen": 2117140480 + }, + { + "epoch": 0.64, + "learning_rate": 0.0002897973921765296, + "loss": 1.2933, + "theoretical_loss": 3.442700781333564, + "tokens_seen": 2117206016 + }, + { + "epoch": 0.64, + "learning_rate": 0.0002897813440320963, + "loss": 1.2996, + "theoretical_loss": 3.4426920681418753, + "tokens_seen": 2117271552 + }, + { + "epoch": 0.64, + "learning_rate": 0.000289765295887663, + "loss": 1.3841, + "theoretical_loss": 3.4426833552953955, + "tokens_seen": 2117337088 + }, + { + "epoch": 0.64, + "learning_rate": 0.0002897492477432297, + "loss": 1.3061, + "theoretical_loss": 3.4426746427941004, + "tokens_seen": 2117402624 + }, + { + "epoch": 0.64, + "learning_rate": 0.0002897331995987964, + "loss": 1.3527, + "theoretical_loss": 3.4426659306379657, + "tokens_seen": 2117468160 + }, + { + "epoch": 0.64, + "learning_rate": 0.0002897171514543631, + "loss": 1.3522, + "theoretical_loss": 3.4426572188269664, + "tokens_seen": 2117533696 + }, + { + "epoch": 0.64, + "learning_rate": 0.0002897011033099298, + "loss": 1.3923, + "theoretical_loss": 3.4426485073610786, + "tokens_seen": 2117599232 + }, + { + "epoch": 0.64, + "learning_rate": 0.0002896850551654965, + "loss": 1.3305, + "theoretical_loss": 3.442639796240278, + "tokens_seen": 2117664768 + }, + { + "epoch": 0.64, + "learning_rate": 0.0002896690070210632, + "loss": 1.3487, + "theoretical_loss": 3.4426310854645408, + "tokens_seen": 2117730304 + }, + { + "epoch": 0.64, + "learning_rate": 0.0002896529588766299, + "loss": 1.3322, + "theoretical_loss": 3.4426223750338414, + "tokens_seen": 2117795840 + }, + { + "epoch": 0.64, + "learning_rate": 0.0002896369107321966, + "loss": 1.2357, + "theoretical_loss": 3.442613664948156, + "tokens_seen": 2117861376 + }, + { + "epoch": 0.64, + "learning_rate": 0.0002896208625877633, + "loss": 1.337, + "theoretical_loss": 3.442604955207461, + "tokens_seen": 2117926912 + }, + { + "epoch": 0.64, + "learning_rate": 0.00028960481444333, + "loss": 1.3019, + "theoretical_loss": 3.442596245811731, + "tokens_seen": 2117992448 + }, + { + "epoch": 0.64, + "learning_rate": 0.0002895887662988967, + "loss": 1.3503, + "theoretical_loss": 3.442587536760942, + "tokens_seen": 2118057984 + }, + { + "epoch": 0.64, + "learning_rate": 0.0002895727181544634, + "loss": 1.32, + "theoretical_loss": 3.4425788280550695, + "tokens_seen": 2118123520 + }, + { + "epoch": 0.64, + "learning_rate": 0.0002895566700100301, + "loss": 1.2352, + "theoretical_loss": 3.44257011969409, + "tokens_seen": 2118189056 + }, + { + "epoch": 0.64, + "learning_rate": 0.0002895406218655968, + "loss": 1.2456, + "theoretical_loss": 3.4425614116779784, + "tokens_seen": 2118254592 + }, + { + "epoch": 0.64, + "learning_rate": 0.0002895245737211635, + "loss": 1.3395, + "theoretical_loss": 3.44255270400671, + "tokens_seen": 2118320128 + }, + { + "epoch": 0.64, + "learning_rate": 0.0002895085255767302, + "loss": 1.2597, + "theoretical_loss": 3.4425439966802616, + "tokens_seen": 2118385664 + }, + { + "epoch": 0.64, + "learning_rate": 0.0002894924774322969, + "loss": 1.2822, + "theoretical_loss": 3.4425352896986086, + "tokens_seen": 2118451200 + }, + { + "epoch": 0.64, + "learning_rate": 0.0002894764292878636, + "loss": 1.3359, + "theoretical_loss": 3.442526583061726, + "tokens_seen": 2118516736 + }, + { + "epoch": 0.64, + "learning_rate": 0.0002894603811434303, + "loss": 1.2261, + "theoretical_loss": 3.44251787676959, + "tokens_seen": 2118582272 + }, + { + "epoch": 0.64, + "learning_rate": 0.000289444332998997, + "loss": 1.3445, + "theoretical_loss": 3.442509170822176, + "tokens_seen": 2118647808 + }, + { + "epoch": 0.64, + "learning_rate": 0.0002894282848545637, + "loss": 1.2305, + "theoretical_loss": 3.4425004652194593, + "tokens_seen": 2118713344 + }, + { + "epoch": 0.64, + "learning_rate": 0.0002894122367101304, + "loss": 1.4006, + "theoretical_loss": 3.442491759961417, + "tokens_seen": 2118778880 + }, + { + "epoch": 0.64, + "learning_rate": 0.0002893961885656971, + "loss": 1.3565, + "theoretical_loss": 3.4424830550480237, + "tokens_seen": 2118844416 + }, + { + "epoch": 0.64, + "learning_rate": 0.0002893801404212638, + "loss": 1.3024, + "theoretical_loss": 3.442474350479255, + "tokens_seen": 2118909952 + }, + { + "epoch": 0.64, + "learning_rate": 0.0002893640922768305, + "loss": 1.3865, + "theoretical_loss": 3.4424656462550876, + "tokens_seen": 2118975488 + }, + { + "epoch": 0.64, + "learning_rate": 0.0002893480441323972, + "loss": 1.3723, + "theoretical_loss": 3.442456942375496, + "tokens_seen": 2119041024 + }, + { + "epoch": 0.64, + "learning_rate": 0.0002893319959879639, + "loss": 1.3, + "theoretical_loss": 3.4424482388404565, + "tokens_seen": 2119106560 + }, + { + "epoch": 0.64, + "learning_rate": 0.0002893159478435306, + "loss": 1.2805, + "theoretical_loss": 3.442439535649945, + "tokens_seen": 2119172096 + }, + { + "epoch": 0.64, + "learning_rate": 0.0002892998996990973, + "loss": 1.2683, + "theoretical_loss": 3.442430832803937, + "tokens_seen": 2119237632 + }, + { + "epoch": 0.64, + "learning_rate": 0.000289283851554664, + "loss": 1.3736, + "theoretical_loss": 3.4424221303024076, + "tokens_seen": 2119303168 + }, + { + "epoch": 0.64, + "learning_rate": 0.0002892678034102307, + "loss": 1.3763, + "theoretical_loss": 3.4424134281453336, + "tokens_seen": 2119368704 + }, + { + "epoch": 0.64, + "learning_rate": 0.0002892517552657974, + "loss": 1.3535, + "theoretical_loss": 3.44240472633269, + "tokens_seen": 2119434240 + }, + { + "epoch": 0.64, + "learning_rate": 0.0002892357071213641, + "loss": 1.3083, + "theoretical_loss": 3.4423960248644523, + "tokens_seen": 2119499776 + }, + { + "epoch": 0.64, + "learning_rate": 0.0002892196589769308, + "loss": 1.3174, + "theoretical_loss": 3.442387323740597, + "tokens_seen": 2119565312 + }, + { + "epoch": 0.64, + "learning_rate": 0.0002892036108324975, + "loss": 1.4258, + "theoretical_loss": 3.4423786229610998, + "tokens_seen": 2119630848 + }, + { + "epoch": 0.64, + "learning_rate": 0.0002891875626880642, + "loss": 1.386, + "theoretical_loss": 3.442369922525936, + "tokens_seen": 2119696384 + }, + { + "epoch": 0.64, + "learning_rate": 0.0002891715145436309, + "loss": 1.2977, + "theoretical_loss": 3.442361222435081, + "tokens_seen": 2119761920 + }, + { + "epoch": 0.64, + "learning_rate": 0.0002891554663991976, + "loss": 1.3097, + "theoretical_loss": 3.4423525226885108, + "tokens_seen": 2119827456 + }, + { + "epoch": 0.64, + "learning_rate": 0.0002891394182547643, + "loss": 1.2141, + "theoretical_loss": 3.442343823286201, + "tokens_seen": 2119892992 + }, + { + "epoch": 0.64, + "learning_rate": 0.000289123370110331, + "loss": 1.3688, + "theoretical_loss": 3.4423351242281286, + "tokens_seen": 2119958528 + }, + { + "epoch": 0.64, + "learning_rate": 0.0002891073219658977, + "loss": 1.3261, + "theoretical_loss": 3.442326425514268, + "tokens_seen": 2120024064 + }, + { + "epoch": 0.64, + "objective/train/docs_used": 772695, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.3367173671722412, + "objective/train/theoretical_loss": 3.442317727144595, + "objective/train/tokens_used": 2140549600, + "theoretical_loss": 3.442317727144595, + "tokens_seen": 2120089600 + }, + { + "epoch": 0.64, + "learning_rate": 0.0002890912738214644, + "loss": 1.3564, + "theoretical_loss": 3.442317727144595, + "tokens_seen": 2120089600 + }, + { + "epoch": 0.64, + "learning_rate": 0.00028907522567703106, + "loss": 1.3254, + "theoretical_loss": 3.4423090291190857, + "tokens_seen": 2120155136 + }, + { + "epoch": 0.64, + "learning_rate": 0.00028905917753259777, + "loss": 1.3369, + "theoretical_loss": 3.442300331437716, + "tokens_seen": 2120220672 + }, + { + "epoch": 0.64, + "learning_rate": 0.00028904312938816447, + "loss": 1.3361, + "theoretical_loss": 3.442291634100461, + "tokens_seen": 2120286208 + }, + { + "epoch": 0.64, + "learning_rate": 0.00028902708124373117, + "loss": 1.2742, + "theoretical_loss": 3.4422829371072967, + "tokens_seen": 2120351744 + }, + { + "epoch": 0.64, + "learning_rate": 0.00028901103309929787, + "loss": 1.3667, + "theoretical_loss": 3.4422742404581994, + "tokens_seen": 2120417280 + }, + { + "epoch": 0.64, + "learning_rate": 0.00028899498495486457, + "loss": 1.2526, + "theoretical_loss": 3.4422655441531447, + "tokens_seen": 2120482816 + }, + { + "epoch": 0.64, + "learning_rate": 0.00028897893681043127, + "loss": 1.3521, + "theoretical_loss": 3.442256848192107, + "tokens_seen": 2120548352 + }, + { + "epoch": 0.64, + "learning_rate": 0.00028896288866599797, + "loss": 1.3174, + "theoretical_loss": 3.442248152575065, + "tokens_seen": 2120613888 + }, + { + "epoch": 0.64, + "learning_rate": 0.00028894684052156467, + "loss": 1.2887, + "theoretical_loss": 3.4422394573019908, + "tokens_seen": 2120679424 + }, + { + "epoch": 0.64, + "learning_rate": 0.00028893079237713137, + "loss": 1.2617, + "theoretical_loss": 3.442230762372863, + "tokens_seen": 2120744960 + }, + { + "epoch": 0.64, + "learning_rate": 0.00028891474423269807, + "loss": 1.3446, + "theoretical_loss": 3.4422220677876565, + "tokens_seen": 2120810496 + }, + { + "epoch": 0.64, + "learning_rate": 0.00028889869608826477, + "loss": 1.218, + "theoretical_loss": 3.4422133735463465, + "tokens_seen": 2120876032 + }, + { + "epoch": 0.64, + "learning_rate": 0.0002888826479438315, + "loss": 1.2738, + "theoretical_loss": 3.4422046796489094, + "tokens_seen": 2120941568 + }, + { + "epoch": 0.64, + "learning_rate": 0.00028886659979939823, + "loss": 1.4033, + "theoretical_loss": 3.442195986095321, + "tokens_seen": 2121007104 + }, + { + "epoch": 0.64, + "learning_rate": 0.00028885055165496493, + "loss": 1.2827, + "theoretical_loss": 3.4421872928855564, + "tokens_seen": 2121072640 + }, + { + "epoch": 0.64, + "learning_rate": 0.00028883450351053163, + "loss": 1.322, + "theoretical_loss": 3.442178600019592, + "tokens_seen": 2121138176 + }, + { + "epoch": 0.64, + "learning_rate": 0.00028881845536609833, + "loss": 1.2247, + "theoretical_loss": 3.4421699074974033, + "tokens_seen": 2121203712 + }, + { + "epoch": 0.64, + "learning_rate": 0.00028880240722166503, + "loss": 1.2648, + "theoretical_loss": 3.4421612153189667, + "tokens_seen": 2121269248 + }, + { + "epoch": 0.64, + "learning_rate": 0.00028878635907723173, + "loss": 1.316, + "theoretical_loss": 3.442152523484257, + "tokens_seen": 2121334784 + }, + { + "epoch": 0.64, + "learning_rate": 0.00028877031093279843, + "loss": 1.4174, + "theoretical_loss": 3.442143831993251, + "tokens_seen": 2121400320 + }, + { + "epoch": 0.64, + "learning_rate": 0.00028875426278836513, + "loss": 1.2428, + "theoretical_loss": 3.4421351408459238, + "tokens_seen": 2121465856 + }, + { + "epoch": 0.64, + "learning_rate": 0.00028873821464393183, + "loss": 1.321, + "theoretical_loss": 3.442126450042251, + "tokens_seen": 2121531392 + }, + { + "epoch": 0.64, + "learning_rate": 0.00028872216649949853, + "loss": 1.2758, + "theoretical_loss": 3.4421177595822092, + "tokens_seen": 2121596928 + }, + { + "epoch": 0.64, + "learning_rate": 0.00028870611835506523, + "loss": 1.3226, + "theoretical_loss": 3.4421090694657734, + "tokens_seen": 2121662464 + }, + { + "epoch": 0.64, + "learning_rate": 0.00028869007021063193, + "loss": 1.4781, + "theoretical_loss": 3.4421003796929197, + "tokens_seen": 2121728000 + }, + { + "epoch": 0.64, + "learning_rate": 0.00028867402206619863, + "loss": 1.3995, + "theoretical_loss": 3.4420916902636245, + "tokens_seen": 2121793536 + }, + { + "epoch": 0.64, + "learning_rate": 0.00028865797392176534, + "loss": 1.3402, + "theoretical_loss": 3.442083001177863, + "tokens_seen": 2121859072 + }, + { + "epoch": 0.64, + "learning_rate": 0.00028864192577733204, + "loss": 1.339, + "theoretical_loss": 3.4420743124356106, + "tokens_seen": 2121924608 + }, + { + "epoch": 0.64, + "learning_rate": 0.00028862587763289874, + "loss": 1.4769, + "theoretical_loss": 3.442065624036844, + "tokens_seen": 2121990144 + }, + { + "epoch": 0.64, + "learning_rate": 0.00028860982948846544, + "loss": 1.2984, + "theoretical_loss": 3.4420569359815385, + "tokens_seen": 2122055680 + }, + { + "epoch": 0.64, + "learning_rate": 0.00028859378134403214, + "loss": 1.3873, + "theoretical_loss": 3.4420482482696704, + "tokens_seen": 2122121216 + }, + { + "epoch": 0.64, + "learning_rate": 0.00028857773319959884, + "loss": 1.2803, + "theoretical_loss": 3.4420395609012147, + "tokens_seen": 2122186752 + }, + { + "epoch": 0.64, + "learning_rate": 0.00028856168505516554, + "loss": 1.3693, + "theoretical_loss": 3.4420308738761474, + "tokens_seen": 2122252288 + }, + { + "epoch": 0.64, + "learning_rate": 0.00028854563691073224, + "loss": 1.3301, + "theoretical_loss": 3.4420221871944454, + "tokens_seen": 2122317824 + }, + { + "epoch": 0.64, + "learning_rate": 0.00028852958876629894, + "loss": 1.2185, + "theoretical_loss": 3.442013500856083, + "tokens_seen": 2122383360 + }, + { + "epoch": 0.64, + "learning_rate": 0.00028851354062186564, + "loss": 1.3924, + "theoretical_loss": 3.442004814861037, + "tokens_seen": 2122448896 + }, + { + "epoch": 0.64, + "learning_rate": 0.00028849749247743234, + "loss": 1.3611, + "theoretical_loss": 3.441996129209283, + "tokens_seen": 2122514432 + }, + { + "epoch": 0.64, + "learning_rate": 0.00028848144433299904, + "loss": 1.2533, + "theoretical_loss": 3.4419874439007967, + "tokens_seen": 2122579968 + }, + { + "epoch": 0.64, + "learning_rate": 0.00028846539618856574, + "loss": 1.3451, + "theoretical_loss": 3.441978758935554, + "tokens_seen": 2122645504 + }, + { + "epoch": 0.64, + "learning_rate": 0.00028844934804413244, + "loss": 1.2931, + "theoretical_loss": 3.441970074313531, + "tokens_seen": 2122711040 + }, + { + "epoch": 0.64, + "learning_rate": 0.0002884332998996991, + "loss": 1.3634, + "theoretical_loss": 3.4419613900347032, + "tokens_seen": 2122776576 + }, + { + "epoch": 0.64, + "learning_rate": 0.0002884172517552658, + "loss": 1.2775, + "theoretical_loss": 3.4419527060990465, + "tokens_seen": 2122842112 + }, + { + "epoch": 0.64, + "learning_rate": 0.0002884012036108325, + "loss": 1.3926, + "theoretical_loss": 3.4419440225065374, + "tokens_seen": 2122907648 + }, + { + "epoch": 0.64, + "learning_rate": 0.0002883851554663992, + "loss": 1.3364, + "theoretical_loss": 3.4419353392571503, + "tokens_seen": 2122973184 + }, + { + "epoch": 0.64, + "learning_rate": 0.0002883691073219659, + "loss": 1.3794, + "theoretical_loss": 3.441926656350862, + "tokens_seen": 2123038720 + }, + { + "epoch": 0.64, + "learning_rate": 0.0002883530591775326, + "loss": 1.3562, + "theoretical_loss": 3.441917973787649, + "tokens_seen": 2123104256 + }, + { + "epoch": 0.64, + "learning_rate": 0.0002883370110330993, + "loss": 1.2309, + "theoretical_loss": 3.441909291567486, + "tokens_seen": 2123169792 + }, + { + "epoch": 0.64, + "learning_rate": 0.000288320962888666, + "loss": 1.2751, + "theoretical_loss": 3.4419006096903493, + "tokens_seen": 2123235328 + }, + { + "epoch": 0.64, + "learning_rate": 0.0002883049147442327, + "loss": 1.2616, + "theoretical_loss": 3.441891928156215, + "tokens_seen": 2123300864 + }, + { + "epoch": 0.64, + "objective/train/docs_used": 773937, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.257105827331543, + "objective/train/theoretical_loss": 3.441883246965058, + "objective/train/tokens_used": 2143826400, + "theoretical_loss": 3.441883246965058, + "tokens_seen": 2123366400 + }, + { + "epoch": 0.64, + "learning_rate": 0.0002882888665997994, + "loss": 1.2656, + "theoretical_loss": 3.441883246965058, + "tokens_seen": 2123366400 + }, + { + "epoch": 0.64, + "learning_rate": 0.0002882728184553661, + "loss": 1.1998, + "theoretical_loss": 3.441874566116855, + "tokens_seen": 2123431936 + }, + { + "epoch": 0.64, + "learning_rate": 0.0002882567703109328, + "loss": 1.2992, + "theoretical_loss": 3.4418658856115822, + "tokens_seen": 2123497472 + }, + { + "epoch": 0.64, + "learning_rate": 0.0002882407221664995, + "loss": 1.2658, + "theoretical_loss": 3.4418572054492147, + "tokens_seen": 2123563008 + }, + { + "epoch": 0.64, + "learning_rate": 0.0002882246740220662, + "loss": 1.2403, + "theoretical_loss": 3.441848525629729, + "tokens_seen": 2123628544 + }, + { + "epoch": 0.64, + "learning_rate": 0.0002882086258776329, + "loss": 1.3316, + "theoretical_loss": 3.4418398461531003, + "tokens_seen": 2123694080 + }, + { + "epoch": 0.64, + "learning_rate": 0.0002881925777331996, + "loss": 1.2709, + "theoretical_loss": 3.4418311670193047, + "tokens_seen": 2123759616 + }, + { + "epoch": 0.64, + "learning_rate": 0.0002881765295887663, + "loss": 1.2816, + "theoretical_loss": 3.441822488228319, + "tokens_seen": 2123825152 + }, + { + "epoch": 0.64, + "learning_rate": 0.000288160481444333, + "loss": 1.3662, + "theoretical_loss": 3.4418138097801174, + "tokens_seen": 2123890688 + }, + { + "epoch": 0.64, + "learning_rate": 0.0002881444332998997, + "loss": 1.3958, + "theoretical_loss": 3.441805131674677, + "tokens_seen": 2123956224 + }, + { + "epoch": 0.64, + "learning_rate": 0.0002881283851554664, + "loss": 1.3683, + "theoretical_loss": 3.4417964539119734, + "tokens_seen": 2124021760 + }, + { + "epoch": 0.64, + "learning_rate": 0.0002881123370110331, + "loss": 1.2767, + "theoretical_loss": 3.4417877764919824, + "tokens_seen": 2124087296 + }, + { + "epoch": 0.64, + "learning_rate": 0.0002880962888665998, + "loss": 1.3768, + "theoretical_loss": 3.44177909941468, + "tokens_seen": 2124152832 + }, + { + "epoch": 0.64, + "learning_rate": 0.0002880802407221665, + "loss": 1.4007, + "theoretical_loss": 3.441770422680042, + "tokens_seen": 2124218368 + }, + { + "epoch": 0.64, + "learning_rate": 0.0002880641925777332, + "loss": 1.3868, + "theoretical_loss": 3.4417617462880443, + "tokens_seen": 2124283904 + }, + { + "epoch": 0.64, + "learning_rate": 0.0002880481444332999, + "loss": 1.4397, + "theoretical_loss": 3.4417530702386627, + "tokens_seen": 2124349440 + }, + { + "epoch": 0.64, + "learning_rate": 0.0002880320962888666, + "loss": 1.3301, + "theoretical_loss": 3.4417443945318738, + "tokens_seen": 2124414976 + }, + { + "epoch": 0.64, + "learning_rate": 0.0002880160481444333, + "loss": 1.2839, + "theoretical_loss": 3.441735719167652, + "tokens_seen": 2124480512 + }, + { + "epoch": 0.64, + "learning_rate": 0.000288, + "loss": 1.2249, + "theoretical_loss": 3.441727044145975, + "tokens_seen": 2124546048 + }, + { + "epoch": 0.64, + "learning_rate": 0.0002879839518555667, + "loss": 1.3447, + "theoretical_loss": 3.441718369466817, + "tokens_seen": 2124611584 + }, + { + "epoch": 0.64, + "learning_rate": 0.0002879679037111334, + "loss": 1.2656, + "theoretical_loss": 3.4417096951301556, + "tokens_seen": 2124677120 + }, + { + "epoch": 0.64, + "learning_rate": 0.0002879518555667001, + "loss": 1.3493, + "theoretical_loss": 3.4417010211359655, + "tokens_seen": 2124742656 + }, + { + "epoch": 0.64, + "learning_rate": 0.0002879358074222668, + "loss": 1.3256, + "theoretical_loss": 3.4416923474842234, + "tokens_seen": 2124808192 + }, + { + "epoch": 0.64, + "learning_rate": 0.0002879197592778335, + "loss": 1.2833, + "theoretical_loss": 3.441683674174904, + "tokens_seen": 2124873728 + }, + { + "epoch": 0.64, + "learning_rate": 0.0002879037111334002, + "loss": 1.3088, + "theoretical_loss": 3.441675001207985, + "tokens_seen": 2124939264 + }, + { + "epoch": 0.64, + "learning_rate": 0.0002878876629889669, + "loss": 1.1897, + "theoretical_loss": 3.44166632858344, + "tokens_seen": 2125004800 + }, + { + "epoch": 0.64, + "learning_rate": 0.0002878716148445336, + "loss": 1.28, + "theoretical_loss": 3.4416576563012473, + "tokens_seen": 2125070336 + }, + { + "epoch": 0.64, + "learning_rate": 0.0002878555667001003, + "loss": 1.4488, + "theoretical_loss": 3.4416489843613816, + "tokens_seen": 2125135872 + }, + { + "epoch": 0.64, + "learning_rate": 0.000287839518555667, + "loss": 1.2691, + "theoretical_loss": 3.4416403127638193, + "tokens_seen": 2125201408 + }, + { + "epoch": 0.64, + "learning_rate": 0.0002878234704112337, + "loss": 1.3481, + "theoretical_loss": 3.441631641508536, + "tokens_seen": 2125266944 + }, + { + "epoch": 0.64, + "learning_rate": 0.0002878074222668004, + "loss": 1.3563, + "theoretical_loss": 3.4416229705955077, + "tokens_seen": 2125332480 + }, + { + "epoch": 0.64, + "learning_rate": 0.0002877913741223671, + "loss": 1.3632, + "theoretical_loss": 3.44161430002471, + "tokens_seen": 2125398016 + }, + { + "epoch": 0.64, + "learning_rate": 0.0002877753259779338, + "loss": 1.4531, + "theoretical_loss": 3.4416056297961193, + "tokens_seen": 2125463552 + }, + { + "epoch": 0.64, + "learning_rate": 0.0002877592778335005, + "loss": 1.303, + "theoretical_loss": 3.4415969599097114, + "tokens_seen": 2125529088 + }, + { + "epoch": 0.64, + "learning_rate": 0.0002877432296890672, + "loss": 1.3241, + "theoretical_loss": 3.4415882903654627, + "tokens_seen": 2125594624 + }, + { + "epoch": 0.64, + "learning_rate": 0.0002877271815446339, + "loss": 1.3241, + "theoretical_loss": 3.4415796211633483, + "tokens_seen": 2125660160 + }, + { + "epoch": 0.64, + "learning_rate": 0.0002877111334002006, + "loss": 1.3607, + "theoretical_loss": 3.4415709523033446, + "tokens_seen": 2125725696 + }, + { + "epoch": 0.64, + "learning_rate": 0.0002876950852557673, + "loss": 1.3691, + "theoretical_loss": 3.441562283785428, + "tokens_seen": 2125791232 + }, + { + "epoch": 0.64, + "learning_rate": 0.000287679037111334, + "loss": 1.2983, + "theoretical_loss": 3.4415536156095734, + "tokens_seen": 2125856768 + }, + { + "epoch": 0.64, + "learning_rate": 0.0002876629889669007, + "loss": 1.3724, + "theoretical_loss": 3.4415449477757574, + "tokens_seen": 2125922304 + }, + { + "epoch": 0.64, + "learning_rate": 0.0002876469408224674, + "loss": 1.2916, + "theoretical_loss": 3.441536280283956, + "tokens_seen": 2125987840 + }, + { + "epoch": 0.64, + "learning_rate": 0.0002876308926780341, + "loss": 1.2446, + "theoretical_loss": 3.441527613134145, + "tokens_seen": 2126053376 + }, + { + "epoch": 0.64, + "learning_rate": 0.0002876148445336008, + "loss": 1.3185, + "theoretical_loss": 3.4415189463263007, + "tokens_seen": 2126118912 + }, + { + "epoch": 0.64, + "learning_rate": 0.0002875987963891675, + "loss": 1.1774, + "theoretical_loss": 3.441510279860399, + "tokens_seen": 2126184448 + }, + { + "epoch": 0.64, + "learning_rate": 0.0002875827482447342, + "loss": 1.3375, + "theoretical_loss": 3.441501613736415, + "tokens_seen": 2126249984 + }, + { + "epoch": 0.64, + "learning_rate": 0.0002875667001003009, + "loss": 1.2488, + "theoretical_loss": 3.441492947954326, + "tokens_seen": 2126315520 + }, + { + "epoch": 0.64, + "learning_rate": 0.0002875506519558676, + "loss": 1.2706, + "theoretical_loss": 3.4414842825141063, + "tokens_seen": 2126381056 + }, + { + "epoch": 0.64, + "learning_rate": 0.0002875346038114343, + "loss": 1.2877, + "theoretical_loss": 3.4414756174157337, + "tokens_seen": 2126446592 + }, + { + "epoch": 0.64, + "learning_rate": 0.000287518555667001, + "loss": 1.3261, + "theoretical_loss": 3.4414669526591832, + "tokens_seen": 2126512128 + }, + { + "epoch": 0.64, + "learning_rate": 0.0002875025075225677, + "loss": 1.4179, + "theoretical_loss": 3.441458288244431, + "tokens_seen": 2126577664 + }, + { + "epoch": 0.64, + "objective/train/docs_used": 774962, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.2357064485549927, + "objective/train/theoretical_loss": 3.441449624171453, + "objective/train/tokens_used": 2147103200, + "theoretical_loss": 3.441449624171453, + "tokens_seen": 2126643200 + }, + { + "epoch": 0.64, + "learning_rate": 0.00028748645937813443, + "loss": 1.2616, + "theoretical_loss": 3.441449624171453, + "tokens_seen": 2126643200 + }, + { + "epoch": 0.64, + "learning_rate": 0.00028747041123370113, + "loss": 1.316, + "theoretical_loss": 3.441440960440225, + "tokens_seen": 2126708736 + }, + { + "epoch": 0.64, + "learning_rate": 0.00028745436308926783, + "loss": 1.2837, + "theoretical_loss": 3.4414322970507234, + "tokens_seen": 2126774272 + }, + { + "epoch": 0.64, + "learning_rate": 0.00028743831494483453, + "loss": 1.2503, + "theoretical_loss": 3.4414236340029243, + "tokens_seen": 2126839808 + }, + { + "epoch": 0.64, + "learning_rate": 0.00028742226680040123, + "loss": 1.2783, + "theoretical_loss": 3.441414971296803, + "tokens_seen": 2126905344 + }, + { + "epoch": 0.64, + "learning_rate": 0.00028740621865596793, + "loss": 1.281, + "theoretical_loss": 3.4414063089323363, + "tokens_seen": 2126970880 + }, + { + "epoch": 0.64, + "learning_rate": 0.00028739017051153463, + "loss": 1.3301, + "theoretical_loss": 3.4413976469094996, + "tokens_seen": 2127036416 + }, + { + "epoch": 0.64, + "learning_rate": 0.00028737412236710133, + "loss": 1.3283, + "theoretical_loss": 3.4413889852282686, + "tokens_seen": 2127101952 + }, + { + "epoch": 0.64, + "learning_rate": 0.00028735807422266803, + "loss": 1.3549, + "theoretical_loss": 3.4413803238886205, + "tokens_seen": 2127167488 + }, + { + "epoch": 0.64, + "learning_rate": 0.00028734202607823473, + "loss": 1.3012, + "theoretical_loss": 3.44137166289053, + "tokens_seen": 2127233024 + }, + { + "epoch": 0.64, + "learning_rate": 0.00028732597793380143, + "loss": 1.2536, + "theoretical_loss": 3.4413630022339743, + "tokens_seen": 2127298560 + }, + { + "epoch": 0.64, + "learning_rate": 0.00028730992978936813, + "loss": 1.3018, + "theoretical_loss": 3.4413543419189283, + "tokens_seen": 2127364096 + }, + { + "epoch": 0.64, + "learning_rate": 0.00028729388164493484, + "loss": 1.3648, + "theoretical_loss": 3.4413456819453687, + "tokens_seen": 2127429632 + }, + { + "epoch": 0.64, + "learning_rate": 0.00028727783350050154, + "loss": 1.3832, + "theoretical_loss": 3.441337022313272, + "tokens_seen": 2127495168 + }, + { + "epoch": 0.64, + "learning_rate": 0.00028726178535606824, + "loss": 1.2536, + "theoretical_loss": 3.441328363022613, + "tokens_seen": 2127560704 + }, + { + "epoch": 0.64, + "learning_rate": 0.00028724573721163494, + "loss": 1.3568, + "theoretical_loss": 3.441319704073368, + "tokens_seen": 2127626240 + }, + { + "epoch": 0.64, + "learning_rate": 0.00028722968906720164, + "loss": 1.3279, + "theoretical_loss": 3.441311045465514, + "tokens_seen": 2127691776 + }, + { + "epoch": 0.64, + "learning_rate": 0.00028721364092276834, + "loss": 1.2898, + "theoretical_loss": 3.4413023871990256, + "tokens_seen": 2127757312 + }, + { + "epoch": 0.64, + "learning_rate": 0.00028719759277833504, + "loss": 1.3412, + "theoretical_loss": 3.44129372927388, + "tokens_seen": 2127822848 + }, + { + "epoch": 0.64, + "learning_rate": 0.00028718154463390174, + "loss": 1.3219, + "theoretical_loss": 3.4412850716900527, + "tokens_seen": 2127888384 + }, + { + "epoch": 0.64, + "learning_rate": 0.00028716549648946844, + "loss": 1.2676, + "theoretical_loss": 3.4412764144475196, + "tokens_seen": 2127953920 + }, + { + "epoch": 0.64, + "learning_rate": 0.00028714944834503514, + "loss": 1.2466, + "theoretical_loss": 3.4412677575462576, + "tokens_seen": 2128019456 + }, + { + "epoch": 0.64, + "learning_rate": 0.00028713340020060184, + "loss": 1.2424, + "theoretical_loss": 3.441259100986241, + "tokens_seen": 2128084992 + }, + { + "epoch": 0.64, + "learning_rate": 0.00028711735205616854, + "loss": 1.3614, + "theoretical_loss": 3.441250444767448, + "tokens_seen": 2128150528 + }, + { + "epoch": 0.64, + "learning_rate": 0.00028710130391173524, + "loss": 1.4711, + "theoretical_loss": 3.441241788889853, + "tokens_seen": 2128216064 + }, + { + "epoch": 0.64, + "learning_rate": 0.00028708525576730194, + "loss": 1.2873, + "theoretical_loss": 3.441233133353433, + "tokens_seen": 2128281600 + }, + { + "epoch": 0.64, + "learning_rate": 0.00028706920762286864, + "loss": 1.3118, + "theoretical_loss": 3.441224478158164, + "tokens_seen": 2128347136 + }, + { + "epoch": 0.64, + "learning_rate": 0.00028705315947843534, + "loss": 1.2873, + "theoretical_loss": 3.441215823304021, + "tokens_seen": 2128412672 + }, + { + "epoch": 0.64, + "learning_rate": 0.000287037111334002, + "loss": 1.3658, + "theoretical_loss": 3.441207168790981, + "tokens_seen": 2128478208 + }, + { + "epoch": 0.65, + "learning_rate": 0.0002870210631895687, + "loss": 1.3041, + "theoretical_loss": 3.4411985146190203, + "tokens_seen": 2128543744 + }, + { + "epoch": 0.65, + "learning_rate": 0.0002870050150451354, + "loss": 1.442, + "theoretical_loss": 3.4411898607881137, + "tokens_seen": 2128609280 + }, + { + "epoch": 0.65, + "learning_rate": 0.0002869889669007021, + "loss": 1.2384, + "theoretical_loss": 3.4411812072982384, + "tokens_seen": 2128674816 + }, + { + "epoch": 0.65, + "learning_rate": 0.0002869729187562688, + "loss": 1.3309, + "theoretical_loss": 3.4411725541493703, + "tokens_seen": 2128740352 + }, + { + "epoch": 0.65, + "learning_rate": 0.0002869568706118355, + "loss": 1.2958, + "theoretical_loss": 3.441163901341485, + "tokens_seen": 2128805888 + }, + { + "epoch": 0.65, + "learning_rate": 0.0002869408224674022, + "loss": 1.2729, + "theoretical_loss": 3.4411552488745594, + "tokens_seen": 2128871424 + }, + { + "epoch": 0.65, + "learning_rate": 0.0002869247743229689, + "loss": 1.3208, + "theoretical_loss": 3.441146596748568, + "tokens_seen": 2128936960 + }, + { + "epoch": 0.65, + "learning_rate": 0.0002869087261785356, + "loss": 1.2622, + "theoretical_loss": 3.4411379449634887, + "tokens_seen": 2129002496 + }, + { + "epoch": 0.65, + "learning_rate": 0.0002868926780341023, + "loss": 1.3746, + "theoretical_loss": 3.4411292935192965, + "tokens_seen": 2129068032 + }, + { + "epoch": 0.65, + "learning_rate": 0.000286876629889669, + "loss": 1.424, + "theoretical_loss": 3.4411206424159677, + "tokens_seen": 2129133568 + }, + { + "epoch": 0.65, + "learning_rate": 0.0002868605817452357, + "loss": 1.3368, + "theoretical_loss": 3.4411119916534787, + "tokens_seen": 2129199104 + }, + { + "epoch": 0.65, + "learning_rate": 0.0002868445336008024, + "loss": 1.3273, + "theoretical_loss": 3.4411033412318046, + "tokens_seen": 2129264640 + }, + { + "epoch": 0.65, + "learning_rate": 0.0002868284854563691, + "loss": 1.4297, + "theoretical_loss": 3.441094691150923, + "tokens_seen": 2129330176 + }, + { + "epoch": 0.65, + "learning_rate": 0.0002868124373119358, + "loss": 1.2723, + "theoretical_loss": 3.441086041410809, + "tokens_seen": 2129395712 + }, + { + "epoch": 0.65, + "learning_rate": 0.0002867963891675025, + "loss": 1.3333, + "theoretical_loss": 3.4410773920114384, + "tokens_seen": 2129461248 + }, + { + "epoch": 0.65, + "learning_rate": 0.0002867803410230692, + "loss": 1.4785, + "theoretical_loss": 3.441068742952788, + "tokens_seen": 2129526784 + }, + { + "epoch": 0.65, + "learning_rate": 0.0002867642928786359, + "loss": 1.2387, + "theoretical_loss": 3.4410600942348335, + "tokens_seen": 2129592320 + }, + { + "epoch": 0.65, + "learning_rate": 0.0002867482447342026, + "loss": 1.4241, + "theoretical_loss": 3.4410514458575516, + "tokens_seen": 2129657856 + }, + { + "epoch": 0.65, + "learning_rate": 0.0002867321965897693, + "loss": 1.2336, + "theoretical_loss": 3.4410427978209173, + "tokens_seen": 2129723392 + }, + { + "epoch": 0.65, + "learning_rate": 0.000286716148445336, + "loss": 1.2435, + "theoretical_loss": 3.4410341501249078, + "tokens_seen": 2129788928 + }, + { + "epoch": 0.65, + "learning_rate": 0.0002867001003009027, + "loss": 1.3429, + "theoretical_loss": 3.4410255027694987, + "tokens_seen": 2129854464 + }, + { + "epoch": 0.65, + "objective/train/docs_used": 775953, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.2863333225250244, + "objective/train/theoretical_loss": 3.441016855754666, + "objective/train/tokens_used": 2150380000, + "theoretical_loss": 3.441016855754666, + "tokens_seen": 2129920000 + }, + { + "epoch": 0.65, + "learning_rate": 0.0002866840521564694, + "loss": 1.3548, + "theoretical_loss": 3.441016855754666, + "tokens_seen": 2129920000 + }, + { + "epoch": 0.65, + "learning_rate": 0.0002866680040120361, + "loss": 1.3264, + "theoretical_loss": 3.441008209080386, + "tokens_seen": 2129985536 + }, + { + "epoch": 0.65, + "learning_rate": 0.0002866519558676028, + "loss": 1.3601, + "theoretical_loss": 3.4409995627466348, + "tokens_seen": 2130051072 + }, + { + "epoch": 0.65, + "learning_rate": 0.0002866359077231695, + "loss": 1.2638, + "theoretical_loss": 3.440990916753388, + "tokens_seen": 2130116608 + }, + { + "epoch": 0.65, + "learning_rate": 0.0002866198595787362, + "loss": 1.3454, + "theoretical_loss": 3.440982271100623, + "tokens_seen": 2130182144 + }, + { + "epoch": 0.65, + "learning_rate": 0.0002866038114343029, + "loss": 1.3669, + "theoretical_loss": 3.440973625788314, + "tokens_seen": 2130247680 + }, + { + "epoch": 0.65, + "learning_rate": 0.0002865877632898696, + "loss": 1.3528, + "theoretical_loss": 3.4409649808164393, + "tokens_seen": 2130313216 + }, + { + "epoch": 0.65, + "learning_rate": 0.0002865717151454363, + "loss": 1.3158, + "theoretical_loss": 3.440956336184974, + "tokens_seen": 2130378752 + }, + { + "epoch": 0.65, + "learning_rate": 0.000286555667001003, + "loss": 1.3364, + "theoretical_loss": 3.4409476918938937, + "tokens_seen": 2130444288 + }, + { + "epoch": 0.65, + "learning_rate": 0.0002865396188565697, + "loss": 1.3571, + "theoretical_loss": 3.440939047943175, + "tokens_seen": 2130509824 + }, + { + "epoch": 0.65, + "learning_rate": 0.0002865235707121364, + "loss": 1.3701, + "theoretical_loss": 3.4409304043327937, + "tokens_seen": 2130575360 + }, + { + "epoch": 0.65, + "learning_rate": 0.0002865075225677031, + "loss": 1.431, + "theoretical_loss": 3.4409217610627265, + "tokens_seen": 2130640896 + }, + { + "epoch": 0.65, + "learning_rate": 0.0002864914744232698, + "loss": 1.4085, + "theoretical_loss": 3.4409131181329498, + "tokens_seen": 2130706432 + }, + { + "epoch": 0.65, + "learning_rate": 0.00028647542627883657, + "loss": 1.2719, + "theoretical_loss": 3.4409044755434386, + "tokens_seen": 2130771968 + }, + { + "epoch": 0.65, + "learning_rate": 0.00028645937813440327, + "loss": 1.2201, + "theoretical_loss": 3.44089583329417, + "tokens_seen": 2130837504 + }, + { + "epoch": 0.65, + "learning_rate": 0.00028644332998996997, + "loss": 1.3944, + "theoretical_loss": 3.4408871913851193, + "tokens_seen": 2130903040 + }, + { + "epoch": 0.65, + "learning_rate": 0.00028642728184553667, + "loss": 1.3528, + "theoretical_loss": 3.4408785498162637, + "tokens_seen": 2130968576 + }, + { + "epoch": 0.65, + "learning_rate": 0.00028641123370110337, + "loss": 1.3467, + "theoretical_loss": 3.4408699085875787, + "tokens_seen": 2131034112 + }, + { + "epoch": 0.65, + "learning_rate": 0.00028639518555667, + "loss": 1.3291, + "theoretical_loss": 3.44086126769904, + "tokens_seen": 2131099648 + }, + { + "epoch": 0.65, + "learning_rate": 0.0002863791374122367, + "loss": 1.2203, + "theoretical_loss": 3.4408526271506252, + "tokens_seen": 2131165184 + }, + { + "epoch": 0.65, + "learning_rate": 0.0002863630892678034, + "loss": 1.3314, + "theoretical_loss": 3.4408439869423093, + "tokens_seen": 2131230720 + }, + { + "epoch": 0.65, + "learning_rate": 0.0002863470411233701, + "loss": 1.3056, + "theoretical_loss": 3.440835347074068, + "tokens_seen": 2131296256 + }, + { + "epoch": 0.65, + "learning_rate": 0.0002863309929789368, + "loss": 1.3253, + "theoretical_loss": 3.440826707545879, + "tokens_seen": 2131361792 + }, + { + "epoch": 0.65, + "learning_rate": 0.0002863149448345035, + "loss": 1.2438, + "theoretical_loss": 3.4408180683577174, + "tokens_seen": 2131427328 + }, + { + "epoch": 0.65, + "learning_rate": 0.0002862988966900702, + "loss": 1.2666, + "theoretical_loss": 3.4408094295095593, + "tokens_seen": 2131492864 + }, + { + "epoch": 0.65, + "learning_rate": 0.0002862828485456369, + "loss": 1.3595, + "theoretical_loss": 3.4408007910013816, + "tokens_seen": 2131558400 + }, + { + "epoch": 0.65, + "learning_rate": 0.0002862668004012036, + "loss": 1.2238, + "theoretical_loss": 3.4407921528331595, + "tokens_seen": 2131623936 + }, + { + "epoch": 0.65, + "learning_rate": 0.0002862507522567703, + "loss": 1.3451, + "theoretical_loss": 3.4407835150048705, + "tokens_seen": 2131689472 + }, + { + "epoch": 0.65, + "learning_rate": 0.000286234704112337, + "loss": 1.2379, + "theoretical_loss": 3.440774877516489, + "tokens_seen": 2131755008 + }, + { + "epoch": 0.65, + "learning_rate": 0.0002862186559679037, + "loss": 1.2971, + "theoretical_loss": 3.440766240367993, + "tokens_seen": 2131820544 + }, + { + "epoch": 0.65, + "learning_rate": 0.0002862026078234704, + "loss": 1.3242, + "theoretical_loss": 3.4407576035593572, + "tokens_seen": 2131886080 + }, + { + "epoch": 0.65, + "learning_rate": 0.0002861865596790371, + "loss": 1.3771, + "theoretical_loss": 3.440748967090559, + "tokens_seen": 2131951616 + }, + { + "epoch": 0.65, + "learning_rate": 0.0002861705115346038, + "loss": 1.3207, + "theoretical_loss": 3.440740330961573, + "tokens_seen": 2132017152 + }, + { + "epoch": 0.65, + "learning_rate": 0.0002861544633901705, + "loss": 1.4103, + "theoretical_loss": 3.440731695172377, + "tokens_seen": 2132082688 + }, + { + "epoch": 0.65, + "learning_rate": 0.0002861384152457372, + "loss": 1.3132, + "theoretical_loss": 3.440723059722947, + "tokens_seen": 2132148224 + }, + { + "epoch": 0.65, + "learning_rate": 0.00028612236710130393, + "loss": 1.3408, + "theoretical_loss": 3.440714424613258, + "tokens_seen": 2132213760 + }, + { + "epoch": 0.65, + "learning_rate": 0.00028610631895687063, + "loss": 1.3596, + "theoretical_loss": 3.440705789843287, + "tokens_seen": 2132279296 + }, + { + "epoch": 0.65, + "learning_rate": 0.00028609027081243733, + "loss": 1.3445, + "theoretical_loss": 3.4406971554130106, + "tokens_seen": 2132344832 + }, + { + "epoch": 0.65, + "learning_rate": 0.00028607422266800403, + "loss": 1.2612, + "theoretical_loss": 3.440688521322404, + "tokens_seen": 2132410368 + }, + { + "epoch": 0.65, + "learning_rate": 0.00028605817452357073, + "loss": 1.3984, + "theoretical_loss": 3.440679887571444, + "tokens_seen": 2132475904 + }, + { + "epoch": 0.65, + "learning_rate": 0.00028604212637913743, + "loss": 1.4567, + "theoretical_loss": 3.440671254160107, + "tokens_seen": 2132541440 + }, + { + "epoch": 0.65, + "learning_rate": 0.00028602607823470413, + "loss": 1.297, + "theoretical_loss": 3.440662621088369, + "tokens_seen": 2132606976 + }, + { + "epoch": 0.65, + "learning_rate": 0.00028601003009027083, + "loss": 1.3296, + "theoretical_loss": 3.440653988356206, + "tokens_seen": 2132672512 + }, + { + "epoch": 0.65, + "learning_rate": 0.00028599398194583753, + "loss": 1.3584, + "theoretical_loss": 3.440645355963594, + "tokens_seen": 2132738048 + }, + { + "epoch": 0.65, + "learning_rate": 0.00028597793380140423, + "loss": 1.4006, + "theoretical_loss": 3.44063672391051, + "tokens_seen": 2132803584 + }, + { + "epoch": 0.65, + "learning_rate": 0.00028596188565697093, + "loss": 1.2276, + "theoretical_loss": 3.44062809219693, + "tokens_seen": 2132869120 + }, + { + "epoch": 0.65, + "learning_rate": 0.00028594583751253763, + "loss": 1.3169, + "theoretical_loss": 3.4406194608228295, + "tokens_seen": 2132934656 + }, + { + "epoch": 0.65, + "learning_rate": 0.00028592978936810433, + "loss": 1.2281, + "theoretical_loss": 3.440610829788185, + "tokens_seen": 2133000192 + }, + { + "epoch": 0.65, + "learning_rate": 0.00028591374122367104, + "loss": 1.2741, + "theoretical_loss": 3.4406021990929734, + "tokens_seen": 2133065728 + }, + { + "epoch": 0.65, + "learning_rate": 0.00028589769307923774, + "loss": 1.2562, + "theoretical_loss": 3.44059356873717, + "tokens_seen": 2133131264 + }, + { + "epoch": 0.65, + "objective/train/docs_used": 777226, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.2317841053009033, + "objective/train/theoretical_loss": 3.440584938720752, + "objective/train/tokens_used": 2153656800, + "theoretical_loss": 3.440584938720752, + "tokens_seen": 2133196800 + }, + { + "epoch": 0.65, + "learning_rate": 0.00028588164493480444, + "loss": 1.2536, + "theoretical_loss": 3.440584938720752, + "tokens_seen": 2133196800 + }, + { + "epoch": 0.65, + "learning_rate": 0.00028586559679037114, + "loss": 1.2629, + "theoretical_loss": 3.4405763090436947, + "tokens_seen": 2133262336 + }, + { + "epoch": 0.65, + "learning_rate": 0.00028584954864593784, + "loss": 1.2763, + "theoretical_loss": 3.4405676797059748, + "tokens_seen": 2133327872 + }, + { + "epoch": 0.65, + "learning_rate": 0.00028583350050150454, + "loss": 1.2267, + "theoretical_loss": 3.440559050707569, + "tokens_seen": 2133393408 + }, + { + "epoch": 0.65, + "learning_rate": 0.00028581745235707124, + "loss": 1.2966, + "theoretical_loss": 3.440550422048452, + "tokens_seen": 2133458944 + }, + { + "epoch": 0.65, + "learning_rate": 0.00028580140421263794, + "loss": 1.2881, + "theoretical_loss": 3.440541793728602, + "tokens_seen": 2133524480 + }, + { + "epoch": 0.65, + "learning_rate": 0.00028578535606820464, + "loss": 1.2455, + "theoretical_loss": 3.440533165747994, + "tokens_seen": 2133590016 + }, + { + "epoch": 0.65, + "learning_rate": 0.00028576930792377134, + "loss": 1.3042, + "theoretical_loss": 3.4405245381066045, + "tokens_seen": 2133655552 + }, + { + "epoch": 0.65, + "learning_rate": 0.00028575325977933804, + "loss": 1.3881, + "theoretical_loss": 3.440515910804409, + "tokens_seen": 2133721088 + }, + { + "epoch": 0.65, + "learning_rate": 0.00028573721163490474, + "loss": 1.2685, + "theoretical_loss": 3.4405072838413853, + "tokens_seen": 2133786624 + }, + { + "epoch": 0.65, + "learning_rate": 0.00028572116349047144, + "loss": 1.33, + "theoretical_loss": 3.4404986572175087, + "tokens_seen": 2133852160 + }, + { + "epoch": 0.65, + "learning_rate": 0.00028570511534603814, + "loss": 1.3222, + "theoretical_loss": 3.440490030932756, + "tokens_seen": 2133917696 + }, + { + "epoch": 0.65, + "learning_rate": 0.00028568906720160484, + "loss": 1.2742, + "theoretical_loss": 3.4404814049871026, + "tokens_seen": 2133983232 + }, + { + "epoch": 0.65, + "learning_rate": 0.00028568906720160484, + "loss": 1.2374, + "theoretical_loss": 3.440472779380525, + "tokens_seen": 2134048768 + }, + { + "epoch": 0.65, + "learning_rate": 0.00028567301905717155, + "loss": 1.2585, + "theoretical_loss": 3.440464154113, + "tokens_seen": 2134114304 + }, + { + "epoch": 0.65, + "learning_rate": 0.00028565697091273825, + "loss": 1.3388, + "theoretical_loss": 3.440455529184504, + "tokens_seen": 2134179840 + }, + { + "epoch": 0.65, + "learning_rate": 0.00028564092276830495, + "loss": 1.2921, + "theoretical_loss": 3.440446904595012, + "tokens_seen": 2134245376 + }, + { + "epoch": 0.65, + "learning_rate": 0.0002856248746238716, + "loss": 1.3202, + "theoretical_loss": 3.4404382803445017, + "tokens_seen": 2134310912 + }, + { + "epoch": 0.65, + "learning_rate": 0.0002856088264794383, + "loss": 1.3883, + "theoretical_loss": 3.4404296564329484, + "tokens_seen": 2134376448 + }, + { + "epoch": 0.65, + "learning_rate": 0.000285592778335005, + "loss": 1.3736, + "theoretical_loss": 3.440421032860329, + "tokens_seen": 2134441984 + }, + { + "epoch": 0.65, + "learning_rate": 0.0002855767301905717, + "loss": 1.2836, + "theoretical_loss": 3.440412409626619, + "tokens_seen": 2134507520 + }, + { + "epoch": 0.65, + "learning_rate": 0.0002855606820461384, + "loss": 1.4284, + "theoretical_loss": 3.440403786731795, + "tokens_seen": 2134573056 + }, + { + "epoch": 0.65, + "learning_rate": 0.0002855446339017051, + "loss": 1.3154, + "theoretical_loss": 3.4403951641758344, + "tokens_seen": 2134638592 + }, + { + "epoch": 0.65, + "learning_rate": 0.0002855285857572718, + "loss": 1.3063, + "theoretical_loss": 3.440386541958712, + "tokens_seen": 2134704128 + }, + { + "epoch": 0.65, + "learning_rate": 0.0002855125376128385, + "loss": 1.403, + "theoretical_loss": 3.4403779200804046, + "tokens_seen": 2134769664 + }, + { + "epoch": 0.65, + "learning_rate": 0.0002854964894684052, + "loss": 1.3026, + "theoretical_loss": 3.4403692985408885, + "tokens_seen": 2134835200 + }, + { + "epoch": 0.65, + "learning_rate": 0.0002854804413239719, + "loss": 1.3415, + "theoretical_loss": 3.4403606773401396, + "tokens_seen": 2134900736 + }, + { + "epoch": 0.65, + "learning_rate": 0.0002854643931795386, + "loss": 1.3655, + "theoretical_loss": 3.440352056478135, + "tokens_seen": 2134966272 + }, + { + "epoch": 0.65, + "learning_rate": 0.0002854483450351053, + "loss": 1.2198, + "theoretical_loss": 3.4403434359548504, + "tokens_seen": 2135031808 + }, + { + "epoch": 0.65, + "learning_rate": 0.000285432296890672, + "loss": 1.3599, + "theoretical_loss": 3.4403348157702625, + "tokens_seen": 2135097344 + }, + { + "epoch": 0.65, + "learning_rate": 0.0002854162487462387, + "loss": 1.37, + "theoretical_loss": 3.440326195924347, + "tokens_seen": 2135162880 + }, + { + "epoch": 0.65, + "learning_rate": 0.0002854002006018054, + "loss": 1.3055, + "theoretical_loss": 3.4403175764170806, + "tokens_seen": 2135228416 + }, + { + "epoch": 0.65, + "learning_rate": 0.0002853841524573721, + "loss": 1.4438, + "theoretical_loss": 3.4403089572484395, + "tokens_seen": 2135293952 + }, + { + "epoch": 0.65, + "learning_rate": 0.0002853681043129388, + "loss": 1.2289, + "theoretical_loss": 3.4403003384184005, + "tokens_seen": 2135359488 + }, + { + "epoch": 0.65, + "learning_rate": 0.0002853520561685055, + "loss": 1.2352, + "theoretical_loss": 3.4402917199269396, + "tokens_seen": 2135425024 + }, + { + "epoch": 0.65, + "learning_rate": 0.0002853360080240722, + "loss": 1.2551, + "theoretical_loss": 3.4402831017740323, + "tokens_seen": 2135490560 + }, + { + "epoch": 0.65, + "learning_rate": 0.0002853199598796389, + "loss": 1.1581, + "theoretical_loss": 3.440274483959656, + "tokens_seen": 2135556096 + }, + { + "epoch": 0.65, + "learning_rate": 0.0002853039117352056, + "loss": 1.3127, + "theoretical_loss": 3.440265866483786, + "tokens_seen": 2135621632 + }, + { + "epoch": 0.65, + "learning_rate": 0.0002852878635907723, + "loss": 1.2403, + "theoretical_loss": 3.4402572493464, + "tokens_seen": 2135687168 + }, + { + "epoch": 0.65, + "learning_rate": 0.00028527181544633906, + "loss": 1.3796, + "theoretical_loss": 3.440248632547473, + "tokens_seen": 2135752704 + }, + { + "epoch": 0.65, + "learning_rate": 0.00028525576730190576, + "loss": 1.3497, + "theoretical_loss": 3.440240016086982, + "tokens_seen": 2135818240 + }, + { + "epoch": 0.65, + "learning_rate": 0.00028523971915747246, + "loss": 1.3141, + "theoretical_loss": 3.4402313999649037, + "tokens_seen": 2135883776 + }, + { + "epoch": 0.65, + "learning_rate": 0.00028522367101303916, + "loss": 1.2446, + "theoretical_loss": 3.440222784181213, + "tokens_seen": 2135949312 + }, + { + "epoch": 0.65, + "learning_rate": 0.00028520762286860586, + "loss": 1.3101, + "theoretical_loss": 3.4402141687358876, + "tokens_seen": 2136014848 + }, + { + "epoch": 0.65, + "learning_rate": 0.00028519157472417256, + "loss": 1.1846, + "theoretical_loss": 3.4402055536289033, + "tokens_seen": 2136080384 + }, + { + "epoch": 0.65, + "learning_rate": 0.00028517552657973926, + "loss": 1.2582, + "theoretical_loss": 3.4401969388602365, + "tokens_seen": 2136145920 + }, + { + "epoch": 0.65, + "learning_rate": 0.00028515947843530597, + "loss": 1.3053, + "theoretical_loss": 3.4401883244298634, + "tokens_seen": 2136211456 + }, + { + "epoch": 0.65, + "learning_rate": 0.00028514343029087267, + "loss": 1.3304, + "theoretical_loss": 3.4401797103377607, + "tokens_seen": 2136276992 + }, + { + "epoch": 0.65, + "learning_rate": 0.00028512738214643937, + "loss": 1.2307, + "theoretical_loss": 3.4401710965839047, + "tokens_seen": 2136342528 + }, + { + "epoch": 0.65, + "learning_rate": 0.00028511133400200607, + "loss": 1.2961, + "theoretical_loss": 3.440162483168271, + "tokens_seen": 2136408064 + }, + { + "epoch": 0.65, + "objective/train/docs_used": 778288, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.3133082389831543, + "objective/train/theoretical_loss": 3.440153870090837, + "objective/train/tokens_used": 2156933600, + "theoretical_loss": 3.440153870090837, + "tokens_seen": 2136473600 + }, + { + "epoch": 0.65, + "learning_rate": 0.00028509528585757277, + "loss": 1.3427, + "theoretical_loss": 3.440153870090837, + "tokens_seen": 2136473600 + }, + { + "epoch": 0.65, + "learning_rate": 0.00028507923771313947, + "loss": 1.3725, + "theoretical_loss": 3.4401452573515785, + "tokens_seen": 2136539136 + }, + { + "epoch": 0.65, + "learning_rate": 0.00028506318956870617, + "loss": 1.2216, + "theoretical_loss": 3.4401366449504716, + "tokens_seen": 2136604672 + }, + { + "epoch": 0.65, + "learning_rate": 0.00028504714142427287, + "loss": 1.4451, + "theoretical_loss": 3.440128032887493, + "tokens_seen": 2136670208 + }, + { + "epoch": 0.65, + "learning_rate": 0.00028503109327983957, + "loss": 1.2324, + "theoretical_loss": 3.4401194211626187, + "tokens_seen": 2136735744 + }, + { + "epoch": 0.65, + "learning_rate": 0.00028501504513540627, + "loss": 1.2994, + "theoretical_loss": 3.4401108097758257, + "tokens_seen": 2136801280 + }, + { + "epoch": 0.65, + "learning_rate": 0.0002849989969909729, + "loss": 1.2438, + "theoretical_loss": 3.4401021987270903, + "tokens_seen": 2136866816 + }, + { + "epoch": 0.65, + "learning_rate": 0.0002849829488465396, + "loss": 1.3536, + "theoretical_loss": 3.440093588016388, + "tokens_seen": 2136932352 + }, + { + "epoch": 0.65, + "learning_rate": 0.0002849669007021063, + "loss": 1.2754, + "theoretical_loss": 3.4400849776436964, + "tokens_seen": 2136997888 + }, + { + "epoch": 0.65, + "learning_rate": 0.000284950852557673, + "loss": 1.372, + "theoretical_loss": 3.4400763676089907, + "tokens_seen": 2137063424 + }, + { + "epoch": 0.65, + "learning_rate": 0.0002849348044132397, + "loss": 1.3086, + "theoretical_loss": 3.4400677579122476, + "tokens_seen": 2137128960 + }, + { + "epoch": 0.65, + "learning_rate": 0.0002849187562688064, + "loss": 1.2866, + "theoretical_loss": 3.440059148553444, + "tokens_seen": 2137194496 + }, + { + "epoch": 0.65, + "learning_rate": 0.0002849027081243731, + "loss": 1.2955, + "theoretical_loss": 3.440050539532556, + "tokens_seen": 2137260032 + }, + { + "epoch": 0.65, + "learning_rate": 0.0002848866599799398, + "loss": 1.4073, + "theoretical_loss": 3.44004193084956, + "tokens_seen": 2137325568 + }, + { + "epoch": 0.65, + "learning_rate": 0.0002848706118355065, + "loss": 1.3538, + "theoretical_loss": 3.4400333225044317, + "tokens_seen": 2137391104 + }, + { + "epoch": 0.65, + "learning_rate": 0.0002848545636910732, + "loss": 1.3301, + "theoretical_loss": 3.4400247144971488, + "tokens_seen": 2137456640 + }, + { + "epoch": 0.65, + "learning_rate": 0.0002848385155466399, + "loss": 1.3308, + "theoretical_loss": 3.440016106827686, + "tokens_seen": 2137522176 + }, + { + "epoch": 0.65, + "learning_rate": 0.0002848224674022066, + "loss": 1.3241, + "theoretical_loss": 3.440007499496021, + "tokens_seen": 2137587712 + }, + { + "epoch": 0.65, + "learning_rate": 0.0002848064192577733, + "loss": 1.2633, + "theoretical_loss": 3.43999889250213, + "tokens_seen": 2137653248 + }, + { + "epoch": 0.65, + "learning_rate": 0.00028479037111334, + "loss": 1.3408, + "theoretical_loss": 3.4399902858459894, + "tokens_seen": 2137718784 + }, + { + "epoch": 0.65, + "learning_rate": 0.0002847743229689067, + "loss": 1.3761, + "theoretical_loss": 3.439981679527575, + "tokens_seen": 2137784320 + }, + { + "epoch": 0.65, + "learning_rate": 0.0002847582748244734, + "loss": 1.3479, + "theoretical_loss": 3.439973073546864, + "tokens_seen": 2137849856 + }, + { + "epoch": 0.65, + "learning_rate": 0.00028474222668004013, + "loss": 1.2811, + "theoretical_loss": 3.439964467903832, + "tokens_seen": 2137915392 + }, + { + "epoch": 0.65, + "learning_rate": 0.00028472617853560683, + "loss": 1.3206, + "theoretical_loss": 3.439955862598456, + "tokens_seen": 2137980928 + }, + { + "epoch": 0.65, + "learning_rate": 0.00028471013039117353, + "loss": 1.3042, + "theoretical_loss": 3.4399472576307124, + "tokens_seen": 2138046464 + }, + { + "epoch": 0.65, + "learning_rate": 0.00028469408224674023, + "loss": 1.3033, + "theoretical_loss": 3.439938653000577, + "tokens_seen": 2138112000 + }, + { + "epoch": 0.65, + "learning_rate": 0.00028467803410230693, + "loss": 1.2761, + "theoretical_loss": 3.4399300487080264, + "tokens_seen": 2138177536 + }, + { + "epoch": 0.65, + "learning_rate": 0.00028466198595787363, + "loss": 1.4058, + "theoretical_loss": 3.4399214447530375, + "tokens_seen": 2138243072 + }, + { + "epoch": 0.65, + "learning_rate": 0.00028464593781344033, + "loss": 1.417, + "theoretical_loss": 3.4399128411355866, + "tokens_seen": 2138308608 + }, + { + "epoch": 0.65, + "learning_rate": 0.00028462988966900703, + "loss": 1.3181, + "theoretical_loss": 3.4399042378556497, + "tokens_seen": 2138374144 + }, + { + "epoch": 0.65, + "learning_rate": 0.00028461384152457373, + "loss": 1.3408, + "theoretical_loss": 3.4398956349132037, + "tokens_seen": 2138439680 + }, + { + "epoch": 0.65, + "learning_rate": 0.00028459779338014043, + "loss": 1.2811, + "theoretical_loss": 3.4398870323082242, + "tokens_seen": 2138505216 + }, + { + "epoch": 0.65, + "learning_rate": 0.00028458174523570713, + "loss": 1.3704, + "theoretical_loss": 3.4398784300406886, + "tokens_seen": 2138570752 + }, + { + "epoch": 0.65, + "learning_rate": 0.00028456569709127383, + "loss": 1.5075, + "theoretical_loss": 3.439869828110573, + "tokens_seen": 2138636288 + }, + { + "epoch": 0.65, + "learning_rate": 0.00028454964894684054, + "loss": 1.2496, + "theoretical_loss": 3.4398612265178534, + "tokens_seen": 2138701824 + }, + { + "epoch": 0.65, + "learning_rate": 0.00028453360080240724, + "loss": 1.3546, + "theoretical_loss": 3.439852625262507, + "tokens_seen": 2138767360 + }, + { + "epoch": 0.65, + "learning_rate": 0.00028451755265797394, + "loss": 1.3363, + "theoretical_loss": 3.4398440243445094, + "tokens_seen": 2138832896 + }, + { + "epoch": 0.65, + "learning_rate": 0.00028450150451354064, + "loss": 1.2922, + "theoretical_loss": 3.4398354237638378, + "tokens_seen": 2138898432 + }, + { + "epoch": 0.65, + "learning_rate": 0.00028448545636910734, + "loss": 1.3106, + "theoretical_loss": 3.439826823520468, + "tokens_seen": 2138963968 + }, + { + "epoch": 0.65, + "learning_rate": 0.00028446940822467404, + "loss": 1.3656, + "theoretical_loss": 3.4398182236143766, + "tokens_seen": 2139029504 + }, + { + "epoch": 0.65, + "learning_rate": 0.00028445336008024074, + "loss": 1.2584, + "theoretical_loss": 3.43980962404554, + "tokens_seen": 2139095040 + }, + { + "epoch": 0.65, + "learning_rate": 0.00028443731193580744, + "loss": 1.317, + "theoretical_loss": 3.439801024813935, + "tokens_seen": 2139160576 + }, + { + "epoch": 0.65, + "learning_rate": 0.00028442126379137414, + "loss": 1.3842, + "theoretical_loss": 3.439792425919538, + "tokens_seen": 2139226112 + }, + { + "epoch": 0.65, + "learning_rate": 0.00028440521564694084, + "loss": 1.4374, + "theoretical_loss": 3.439783827362325, + "tokens_seen": 2139291648 + }, + { + "epoch": 0.65, + "learning_rate": 0.00028438916750250754, + "loss": 1.3478, + "theoretical_loss": 3.439775229142273, + "tokens_seen": 2139357184 + }, + { + "epoch": 0.65, + "learning_rate": 0.00028437311935807424, + "loss": 1.3884, + "theoretical_loss": 3.439766631259358, + "tokens_seen": 2139422720 + }, + { + "epoch": 0.65, + "learning_rate": 0.00028435707121364094, + "loss": 1.2092, + "theoretical_loss": 3.4397580337135567, + "tokens_seen": 2139488256 + }, + { + "epoch": 0.65, + "learning_rate": 0.00028434102306920764, + "loss": 1.399, + "theoretical_loss": 3.4397494365048455, + "tokens_seen": 2139553792 + }, + { + "epoch": 0.65, + "learning_rate": 0.00028432497492477434, + "loss": 1.3168, + "theoretical_loss": 3.4397408396332008, + "tokens_seen": 2139619328 + }, + { + "epoch": 0.65, + "learning_rate": 0.00028430892678034104, + "loss": 1.4007, + "theoretical_loss": 3.4397322430985993, + "tokens_seen": 2139684864 + }, + { + "epoch": 0.65, + "objective/train/docs_used": 779442, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.3021600246429443, + "objective/train/theoretical_loss": 3.439723646901017, + "objective/train/tokens_used": 2160210400, + "theoretical_loss": 3.439723646901017, + "tokens_seen": 2139750400 + }, + { + "epoch": 0.65, + "learning_rate": 0.00028429287863590775, + "loss": 1.3278, + "theoretical_loss": 3.439723646901017, + "tokens_seen": 2139750400 + }, + { + "epoch": 0.65, + "learning_rate": 0.00028427683049147445, + "loss": 1.3701, + "theoretical_loss": 3.43971505104043, + "tokens_seen": 2139815936 + }, + { + "epoch": 0.65, + "learning_rate": 0.00028426078234704115, + "loss": 1.1988, + "theoretical_loss": 3.4397064555168164, + "tokens_seen": 2139881472 + }, + { + "epoch": 0.65, + "learning_rate": 0.00028424473420260785, + "loss": 1.3682, + "theoretical_loss": 3.439697860330152, + "tokens_seen": 2139947008 + }, + { + "epoch": 0.65, + "learning_rate": 0.0002842286860581745, + "loss": 1.2953, + "theoretical_loss": 3.439689265480412, + "tokens_seen": 2140012544 + }, + { + "epoch": 0.65, + "learning_rate": 0.0002842126379137412, + "loss": 1.4826, + "theoretical_loss": 3.439680670967574, + "tokens_seen": 2140078080 + }, + { + "epoch": 0.65, + "learning_rate": 0.0002841965897693079, + "loss": 1.2858, + "theoretical_loss": 3.4396720767916147, + "tokens_seen": 2140143616 + }, + { + "epoch": 0.65, + "learning_rate": 0.0002841805416248746, + "loss": 1.3029, + "theoretical_loss": 3.4396634829525095, + "tokens_seen": 2140209152 + }, + { + "epoch": 0.65, + "learning_rate": 0.0002841644934804413, + "loss": 1.2073, + "theoretical_loss": 3.439654889450236, + "tokens_seen": 2140274688 + }, + { + "epoch": 0.65, + "learning_rate": 0.000284148445336008, + "loss": 1.2877, + "theoretical_loss": 3.4396462962847707, + "tokens_seen": 2140340224 + }, + { + "epoch": 0.65, + "learning_rate": 0.0002841323971915747, + "loss": 1.3442, + "theoretical_loss": 3.4396377034560888, + "tokens_seen": 2140405760 + }, + { + "epoch": 0.65, + "learning_rate": 0.0002841163490471414, + "loss": 1.2444, + "theoretical_loss": 3.4396291109641677, + "tokens_seen": 2140471296 + }, + { + "epoch": 0.65, + "learning_rate": 0.0002841003009027081, + "loss": 1.185, + "theoretical_loss": 3.4396205188089843, + "tokens_seen": 2140536832 + }, + { + "epoch": 0.65, + "learning_rate": 0.0002840842527582748, + "loss": 1.3662, + "theoretical_loss": 3.4396119269905143, + "tokens_seen": 2140602368 + }, + { + "epoch": 0.65, + "learning_rate": 0.00028406820461384155, + "loss": 1.3046, + "theoretical_loss": 3.4396033355087345, + "tokens_seen": 2140667904 + }, + { + "epoch": 0.65, + "learning_rate": 0.00028405215646940825, + "loss": 1.267, + "theoretical_loss": 3.439594744363622, + "tokens_seen": 2140733440 + }, + { + "epoch": 0.65, + "learning_rate": 0.00028403610832497496, + "loss": 1.2729, + "theoretical_loss": 3.439586153555152, + "tokens_seen": 2140798976 + }, + { + "epoch": 0.65, + "learning_rate": 0.00028402006018054166, + "loss": 1.3278, + "theoretical_loss": 3.4395775630833016, + "tokens_seen": 2140864512 + }, + { + "epoch": 0.65, + "learning_rate": 0.00028400401203610836, + "loss": 1.3554, + "theoretical_loss": 3.439568972948048, + "tokens_seen": 2140930048 + }, + { + "epoch": 0.65, + "learning_rate": 0.00028398796389167506, + "loss": 1.2707, + "theoretical_loss": 3.439560383149367, + "tokens_seen": 2140995584 + }, + { + "epoch": 0.65, + "learning_rate": 0.00028397191574724176, + "loss": 1.2512, + "theoretical_loss": 3.439551793687235, + "tokens_seen": 2141061120 + }, + { + "epoch": 0.65, + "learning_rate": 0.00028395586760280846, + "loss": 1.3031, + "theoretical_loss": 3.439543204561629, + "tokens_seen": 2141126656 + }, + { + "epoch": 0.65, + "learning_rate": 0.00028393981945837516, + "loss": 1.2499, + "theoretical_loss": 3.439534615772525, + "tokens_seen": 2141192192 + }, + { + "epoch": 0.65, + "learning_rate": 0.00028392377131394186, + "loss": 1.3216, + "theoretical_loss": 3.4395260273199, + "tokens_seen": 2141257728 + }, + { + "epoch": 0.65, + "learning_rate": 0.00028390772316950856, + "loss": 1.1761, + "theoretical_loss": 3.4395174392037298, + "tokens_seen": 2141323264 + }, + { + "epoch": 0.65, + "learning_rate": 0.00028389167502507526, + "loss": 1.3167, + "theoretical_loss": 3.439508851423992, + "tokens_seen": 2141388800 + }, + { + "epoch": 0.65, + "learning_rate": 0.00028387562688064196, + "loss": 1.3554, + "theoretical_loss": 3.4395002639806624, + "tokens_seen": 2141454336 + }, + { + "epoch": 0.65, + "learning_rate": 0.00028385957873620866, + "loss": 1.4307, + "theoretical_loss": 3.4394916768737174, + "tokens_seen": 2141519872 + }, + { + "epoch": 0.65, + "learning_rate": 0.00028384353059177536, + "loss": 1.309, + "theoretical_loss": 3.4394830901031344, + "tokens_seen": 2141585408 + }, + { + "epoch": 0.65, + "learning_rate": 0.00028382748244734206, + "loss": 1.215, + "theoretical_loss": 3.4394745036688885, + "tokens_seen": 2141650944 + }, + { + "epoch": 0.65, + "learning_rate": 0.00028381143430290876, + "loss": 1.2986, + "theoretical_loss": 3.439465917570958, + "tokens_seen": 2141716480 + }, + { + "epoch": 0.65, + "learning_rate": 0.00028379538615847547, + "loss": 1.2463, + "theoretical_loss": 3.439457331809318, + "tokens_seen": 2141782016 + }, + { + "epoch": 0.65, + "learning_rate": 0.00028377933801404217, + "loss": 1.3976, + "theoretical_loss": 3.4394487463839454, + "tokens_seen": 2141847552 + }, + { + "epoch": 0.65, + "learning_rate": 0.00028376328986960887, + "loss": 1.2135, + "theoretical_loss": 3.4394401612948173, + "tokens_seen": 2141913088 + }, + { + "epoch": 0.65, + "learning_rate": 0.00028374724172517557, + "loss": 1.2793, + "theoretical_loss": 3.4394315765419097, + "tokens_seen": 2141978624 + }, + { + "epoch": 0.65, + "learning_rate": 0.00028373119358074227, + "loss": 1.3629, + "theoretical_loss": 3.4394229921251993, + "tokens_seen": 2142044160 + }, + { + "epoch": 0.65, + "learning_rate": 0.00028371514543630897, + "loss": 1.4164, + "theoretical_loss": 3.4394144080446623, + "tokens_seen": 2142109696 + }, + { + "epoch": 0.65, + "learning_rate": 0.00028369909729187567, + "loss": 1.2687, + "theoretical_loss": 3.439405824300276, + "tokens_seen": 2142175232 + }, + { + "epoch": 0.65, + "learning_rate": 0.00028368304914744237, + "loss": 1.3159, + "theoretical_loss": 3.439397240892016, + "tokens_seen": 2142240768 + }, + { + "epoch": 0.65, + "learning_rate": 0.00028366700100300907, + "loss": 1.3775, + "theoretical_loss": 3.43938865781986, + "tokens_seen": 2142306304 + }, + { + "epoch": 0.65, + "learning_rate": 0.00028365095285857577, + "loss": 1.2273, + "theoretical_loss": 3.4393800750837835, + "tokens_seen": 2142371840 + }, + { + "epoch": 0.65, + "learning_rate": 0.00028363490471414247, + "loss": 1.1701, + "theoretical_loss": 3.4393714926837635, + "tokens_seen": 2142437376 + }, + { + "epoch": 0.65, + "learning_rate": 0.00028361885656970917, + "loss": 1.3189, + "theoretical_loss": 3.439362910619777, + "tokens_seen": 2142502912 + }, + { + "epoch": 0.65, + "learning_rate": 0.0002836028084252758, + "loss": 1.3202, + "theoretical_loss": 3.4393543288918, + "tokens_seen": 2142568448 + }, + { + "epoch": 0.65, + "learning_rate": 0.0002835867602808425, + "loss": 1.3829, + "theoretical_loss": 3.439345747499809, + "tokens_seen": 2142633984 + }, + { + "epoch": 0.65, + "learning_rate": 0.0002835707121364092, + "loss": 1.1637, + "theoretical_loss": 3.4393371664437806, + "tokens_seen": 2142699520 + }, + { + "epoch": 0.65, + "learning_rate": 0.0002835546639919759, + "loss": 1.342, + "theoretical_loss": 3.4393285857236915, + "tokens_seen": 2142765056 + }, + { + "epoch": 0.65, + "learning_rate": 0.0002835386158475426, + "loss": 1.3543, + "theoretical_loss": 3.4393200053395185, + "tokens_seen": 2142830592 + }, + { + "epoch": 0.65, + "learning_rate": 0.0002835225677031093, + "loss": 1.2983, + "theoretical_loss": 3.4393114252912387, + "tokens_seen": 2142896128 + }, + { + "epoch": 0.65, + "learning_rate": 0.000283506519558676, + "loss": 1.4445, + "theoretical_loss": 3.439302845578827, + "tokens_seen": 2142961664 + }, + { + "epoch": 0.65, + "objective/train/docs_used": 780500, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.162071943283081, + "objective/train/theoretical_loss": 3.4392942662022614, + "objective/train/tokens_used": 2163487200, + "theoretical_loss": 3.4392942662022614, + "tokens_seen": 2143027200 + }, + { + "epoch": 0.65, + "learning_rate": 0.0002834904714142427, + "loss": 1.284, + "theoretical_loss": 3.4392942662022614, + "tokens_seen": 2143027200 + }, + { + "epoch": 0.65, + "learning_rate": 0.0002834744232698094, + "loss": 1.3601, + "theoretical_loss": 3.439285687161518, + "tokens_seen": 2143092736 + }, + { + "epoch": 0.65, + "learning_rate": 0.0002834583751253761, + "loss": 1.4014, + "theoretical_loss": 3.439277108456573, + "tokens_seen": 2143158272 + }, + { + "epoch": 0.65, + "learning_rate": 0.0002834423269809428, + "loss": 1.1994, + "theoretical_loss": 3.439268530087404, + "tokens_seen": 2143223808 + }, + { + "epoch": 0.65, + "learning_rate": 0.0002834262788365095, + "loss": 1.3696, + "theoretical_loss": 3.4392599520539866, + "tokens_seen": 2143289344 + }, + { + "epoch": 0.65, + "learning_rate": 0.0002834102306920762, + "loss": 1.2234, + "theoretical_loss": 3.439251374356298, + "tokens_seen": 2143354880 + }, + { + "epoch": 0.65, + "learning_rate": 0.0002833941825476429, + "loss": 1.3661, + "theoretical_loss": 3.4392427969943147, + "tokens_seen": 2143420416 + }, + { + "epoch": 0.65, + "learning_rate": 0.00028337813440320963, + "loss": 1.1598, + "theoretical_loss": 3.4392342199680126, + "tokens_seen": 2143485952 + }, + { + "epoch": 0.65, + "learning_rate": 0.00028336208625877633, + "loss": 1.343, + "theoretical_loss": 3.4392256432773696, + "tokens_seen": 2143551488 + }, + { + "epoch": 0.65, + "learning_rate": 0.00028334603811434303, + "loss": 1.2524, + "theoretical_loss": 3.439217066922361, + "tokens_seen": 2143617024 + }, + { + "epoch": 0.65, + "learning_rate": 0.00028332998996990973, + "loss": 1.3231, + "theoretical_loss": 3.4392084909029643, + "tokens_seen": 2143682560 + }, + { + "epoch": 0.65, + "learning_rate": 0.00028331394182547643, + "loss": 1.3863, + "theoretical_loss": 3.439199915219156, + "tokens_seen": 2143748096 + }, + { + "epoch": 0.65, + "learning_rate": 0.00028329789368104313, + "loss": 1.2567, + "theoretical_loss": 3.4391913398709124, + "tokens_seen": 2143813632 + }, + { + "epoch": 0.65, + "learning_rate": 0.00028328184553660983, + "loss": 1.2154, + "theoretical_loss": 3.4391827648582103, + "tokens_seen": 2143879168 + }, + { + "epoch": 0.65, + "learning_rate": 0.00028326579739217653, + "loss": 1.2545, + "theoretical_loss": 3.439174190181026, + "tokens_seen": 2143944704 + }, + { + "epoch": 0.65, + "learning_rate": 0.00028324974924774323, + "loss": 1.3278, + "theoretical_loss": 3.439165615839336, + "tokens_seen": 2144010240 + }, + { + "epoch": 0.65, + "learning_rate": 0.00028323370110330993, + "loss": 1.2349, + "theoretical_loss": 3.439157041833118, + "tokens_seen": 2144075776 + }, + { + "epoch": 0.65, + "learning_rate": 0.00028321765295887663, + "loss": 1.3316, + "theoretical_loss": 3.4391484681623474, + "tokens_seen": 2144141312 + }, + { + "epoch": 0.65, + "learning_rate": 0.00028320160481444333, + "loss": 1.4004, + "theoretical_loss": 3.4391398948270018, + "tokens_seen": 2144206848 + }, + { + "epoch": 0.65, + "learning_rate": 0.00028318555667001004, + "loss": 1.3361, + "theoretical_loss": 3.439131321827057, + "tokens_seen": 2144272384 + }, + { + "epoch": 0.65, + "learning_rate": 0.00028316950852557674, + "loss": 1.3387, + "theoretical_loss": 3.4391227491624896, + "tokens_seen": 2144337920 + }, + { + "epoch": 0.65, + "learning_rate": 0.00028315346038114344, + "loss": 1.2995, + "theoretical_loss": 3.4391141768332774, + "tokens_seen": 2144403456 + }, + { + "epoch": 0.65, + "learning_rate": 0.00028313741223671014, + "loss": 1.3763, + "theoretical_loss": 3.439105604839396, + "tokens_seen": 2144468992 + }, + { + "epoch": 0.65, + "learning_rate": 0.00028312136409227684, + "loss": 1.2697, + "theoretical_loss": 3.439097033180822, + "tokens_seen": 2144534528 + }, + { + "epoch": 0.65, + "learning_rate": 0.00028310531594784354, + "loss": 1.1998, + "theoretical_loss": 3.4390884618575326, + "tokens_seen": 2144600064 + }, + { + "epoch": 0.65, + "learning_rate": 0.00028308926780341024, + "loss": 1.2836, + "theoretical_loss": 3.4390798908695035, + "tokens_seen": 2144665600 + }, + { + "epoch": 0.65, + "learning_rate": 0.00028307321965897694, + "loss": 1.1901, + "theoretical_loss": 3.439071320216713, + "tokens_seen": 2144731136 + }, + { + "epoch": 0.65, + "learning_rate": 0.00028305717151454364, + "loss": 1.3505, + "theoretical_loss": 3.4390627498991355, + "tokens_seen": 2144796672 + }, + { + "epoch": 0.65, + "learning_rate": 0.00028304112337011034, + "loss": 1.4531, + "theoretical_loss": 3.43905417991675, + "tokens_seen": 2144862208 + }, + { + "epoch": 0.65, + "learning_rate": 0.00028302507522567704, + "loss": 1.3034, + "theoretical_loss": 3.4390456102695315, + "tokens_seen": 2144927744 + }, + { + "epoch": 0.65, + "learning_rate": 0.00028300902708124374, + "loss": 1.2365, + "theoretical_loss": 3.4390370409574573, + "tokens_seen": 2144993280 + }, + { + "epoch": 0.65, + "learning_rate": 0.00028299297893681044, + "loss": 1.2643, + "theoretical_loss": 3.439028471980504, + "tokens_seen": 2145058816 + }, + { + "epoch": 0.65, + "learning_rate": 0.00028297693079237714, + "loss": 1.2704, + "theoretical_loss": 3.439019903338648, + "tokens_seen": 2145124352 + }, + { + "epoch": 0.65, + "learning_rate": 0.00028296088264794384, + "loss": 1.3636, + "theoretical_loss": 3.439011335031866, + "tokens_seen": 2145189888 + }, + { + "epoch": 0.65, + "learning_rate": 0.00028294483450351054, + "loss": 1.3394, + "theoretical_loss": 3.4390027670601353, + "tokens_seen": 2145255424 + }, + { + "epoch": 0.65, + "learning_rate": 0.00028292878635907725, + "loss": 1.3199, + "theoretical_loss": 3.4389941994234317, + "tokens_seen": 2145320960 + }, + { + "epoch": 0.65, + "learning_rate": 0.00028291273821464395, + "loss": 1.3258, + "theoretical_loss": 3.4389856321217325, + "tokens_seen": 2145386496 + }, + { + "epoch": 0.65, + "learning_rate": 0.00028289669007021065, + "loss": 1.3241, + "theoretical_loss": 3.4389770651550142, + "tokens_seen": 2145452032 + }, + { + "epoch": 0.65, + "learning_rate": 0.00028288064192577735, + "loss": 1.2148, + "theoretical_loss": 3.438968498523253, + "tokens_seen": 2145517568 + }, + { + "epoch": 0.65, + "learning_rate": 0.00028286459378134405, + "loss": 1.3043, + "theoretical_loss": 3.4389599322264264, + "tokens_seen": 2145583104 + }, + { + "epoch": 0.65, + "learning_rate": 0.00028284854563691075, + "loss": 1.357, + "theoretical_loss": 3.43895136626451, + "tokens_seen": 2145648640 + }, + { + "epoch": 0.65, + "learning_rate": 0.00028283249749247745, + "loss": 1.3084, + "theoretical_loss": 3.4389428006374816, + "tokens_seen": 2145714176 + }, + { + "epoch": 0.65, + "learning_rate": 0.00028281644934804415, + "loss": 1.3024, + "theoretical_loss": 3.438934235345317, + "tokens_seen": 2145779712 + }, + { + "epoch": 0.65, + "learning_rate": 0.00028280040120361085, + "loss": 1.291, + "theoretical_loss": 3.4389256703879942, + "tokens_seen": 2145845248 + }, + { + "epoch": 0.65, + "learning_rate": 0.00028278435305917755, + "loss": 1.2092, + "theoretical_loss": 3.438917105765488, + "tokens_seen": 2145910784 + }, + { + "epoch": 0.65, + "learning_rate": 0.00028276830491474425, + "loss": 1.342, + "theoretical_loss": 3.438908541477776, + "tokens_seen": 2145976320 + }, + { + "epoch": 0.65, + "learning_rate": 0.00028275225677031095, + "loss": 1.3021, + "theoretical_loss": 3.4388999775248354, + "tokens_seen": 2146041856 + }, + { + "epoch": 0.65, + "learning_rate": 0.00028273620862587765, + "loss": 1.3782, + "theoretical_loss": 3.438891413906642, + "tokens_seen": 2146107392 + }, + { + "epoch": 0.65, + "learning_rate": 0.00028272016048144435, + "loss": 1.3864, + "theoretical_loss": 3.4388828506231737, + "tokens_seen": 2146172928 + }, + { + "epoch": 0.65, + "learning_rate": 0.00028270411233701105, + "loss": 1.3311, + "theoretical_loss": 3.438874287674406, + "tokens_seen": 2146238464 + }, + { + "epoch": 0.65, + "objective/train/docs_used": 781712, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.2351510524749756, + "objective/train/theoretical_loss": 3.4388657250603156, + "objective/train/tokens_used": 2166764000, + "theoretical_loss": 3.4388657250603156, + "tokens_seen": 2146304000 + }, + { + "epoch": 0.65, + "learning_rate": 0.00028268806419257775, + "loss": 1.2635, + "theoretical_loss": 3.4388657250603156, + "tokens_seen": 2146304000 + }, + { + "epoch": 0.65, + "learning_rate": 0.00028267201604814446, + "loss": 1.2722, + "theoretical_loss": 3.43885716278088, + "tokens_seen": 2146369536 + }, + { + "epoch": 0.65, + "learning_rate": 0.00028265596790371116, + "loss": 1.2641, + "theoretical_loss": 3.4388486008360752, + "tokens_seen": 2146435072 + }, + { + "epoch": 0.65, + "learning_rate": 0.00028263991975927786, + "loss": 1.2178, + "theoretical_loss": 3.4388400392258784, + "tokens_seen": 2146500608 + }, + { + "epoch": 0.65, + "learning_rate": 0.00028262387161484456, + "loss": 1.2726, + "theoretical_loss": 3.4388314779502664, + "tokens_seen": 2146566144 + }, + { + "epoch": 0.65, + "learning_rate": 0.00028260782347041126, + "loss": 1.3878, + "theoretical_loss": 3.4388229170092153, + "tokens_seen": 2146631680 + }, + { + "epoch": 0.65, + "learning_rate": 0.00028259177532597796, + "loss": 1.2906, + "theoretical_loss": 3.4388143564027023, + "tokens_seen": 2146697216 + }, + { + "epoch": 0.65, + "learning_rate": 0.00028257572718154466, + "loss": 1.3457, + "theoretical_loss": 3.438805796130704, + "tokens_seen": 2146762752 + }, + { + "epoch": 0.65, + "learning_rate": 0.00028255967903711136, + "loss": 1.2834, + "theoretical_loss": 3.4387972361931967, + "tokens_seen": 2146828288 + }, + { + "epoch": 0.65, + "learning_rate": 0.00028254363089267806, + "loss": 1.2557, + "theoretical_loss": 3.438788676590158, + "tokens_seen": 2146893824 + }, + { + "epoch": 0.65, + "learning_rate": 0.00028252758274824476, + "loss": 1.3509, + "theoretical_loss": 3.4387801173215635, + "tokens_seen": 2146959360 + }, + { + "epoch": 0.65, + "learning_rate": 0.00028251153460381146, + "loss": 1.3666, + "theoretical_loss": 3.4387715583873915, + "tokens_seen": 2147024896 + }, + { + "epoch": 0.65, + "learning_rate": 0.00028249548645937816, + "loss": 1.3251, + "theoretical_loss": 3.438762999787617, + "tokens_seen": 2147090432 + }, + { + "epoch": 0.65, + "learning_rate": 0.00028247943831494486, + "loss": 1.3806, + "theoretical_loss": 3.4387544415222178, + "tokens_seen": 2147155968 + }, + { + "epoch": 0.65, + "learning_rate": 0.00028246339017051156, + "loss": 1.4316, + "theoretical_loss": 3.43874588359117, + "tokens_seen": 2147221504 + }, + { + "epoch": 0.65, + "learning_rate": 0.00028244734202607826, + "loss": 1.2569, + "theoretical_loss": 3.4387373259944507, + "tokens_seen": 2147287040 + }, + { + "epoch": 0.65, + "learning_rate": 0.00028243129388164496, + "loss": 1.332, + "theoretical_loss": 3.4387287687320374, + "tokens_seen": 2147352576 + }, + { + "epoch": 0.65, + "learning_rate": 0.00028241524573721167, + "loss": 1.2431, + "theoretical_loss": 3.438720211803905, + "tokens_seen": 2147418112 + }, + { + "epoch": 0.65, + "learning_rate": 0.00028239919759277837, + "loss": 1.3091, + "theoretical_loss": 3.4387116552100316, + "tokens_seen": 2147483648 + }, + { + "epoch": 0.65, + "learning_rate": 0.00028238314944834507, + "loss": 1.2102, + "theoretical_loss": 3.438703098950394, + "tokens_seen": 2147549184 + }, + { + "epoch": 0.65, + "learning_rate": 0.00028236710130391177, + "loss": 1.1489, + "theoretical_loss": 3.4386945430249685, + "tokens_seen": 2147614720 + }, + { + "epoch": 0.65, + "learning_rate": 0.00028235105315947847, + "loss": 1.1439, + "theoretical_loss": 3.4386859874337317, + "tokens_seen": 2147680256 + }, + { + "epoch": 0.65, + "learning_rate": 0.00028233500501504517, + "loss": 1.3917, + "theoretical_loss": 3.4386774321766604, + "tokens_seen": 2147745792 + }, + { + "epoch": 0.65, + "learning_rate": 0.00028231895687061187, + "loss": 1.2881, + "theoretical_loss": 3.4386688772537317, + "tokens_seen": 2147811328 + }, + { + "epoch": 0.65, + "learning_rate": 0.00028230290872617857, + "loss": 1.434, + "theoretical_loss": 3.438660322664922, + "tokens_seen": 2147876864 + }, + { + "epoch": 0.65, + "learning_rate": 0.00028228686058174527, + "loss": 1.3391, + "theoretical_loss": 3.4386517684102085, + "tokens_seen": 2147942400 + }, + { + "epoch": 0.65, + "learning_rate": 0.00028227081243731197, + "loss": 1.4192, + "theoretical_loss": 3.4386432144895673, + "tokens_seen": 2148007936 + }, + { + "epoch": 0.65, + "learning_rate": 0.00028225476429287867, + "loss": 1.3183, + "theoretical_loss": 3.4386346609029763, + "tokens_seen": 2148073472 + }, + { + "epoch": 0.65, + "learning_rate": 0.00028223871614844537, + "loss": 1.3679, + "theoretical_loss": 3.4386261076504105, + "tokens_seen": 2148139008 + }, + { + "epoch": 0.65, + "learning_rate": 0.0002822226680040121, + "loss": 1.404, + "theoretical_loss": 3.4386175547318487, + "tokens_seen": 2148204544 + }, + { + "epoch": 0.65, + "learning_rate": 0.0002822066198595788, + "loss": 1.1702, + "theoretical_loss": 3.438609002147266, + "tokens_seen": 2148270080 + }, + { + "epoch": 0.65, + "learning_rate": 0.0002821905717151454, + "loss": 1.2224, + "theoretical_loss": 3.43860044989664, + "tokens_seen": 2148335616 + }, + { + "epoch": 0.65, + "learning_rate": 0.0002821745235707121, + "loss": 1.3442, + "theoretical_loss": 3.438591897979947, + "tokens_seen": 2148401152 + }, + { + "epoch": 0.65, + "learning_rate": 0.0002821584754262788, + "loss": 1.2637, + "theoretical_loss": 3.4385833463971647, + "tokens_seen": 2148466688 + }, + { + "epoch": 0.65, + "learning_rate": 0.0002821424272818455, + "loss": 1.334, + "theoretical_loss": 3.4385747951482686, + "tokens_seen": 2148532224 + }, + { + "epoch": 0.65, + "learning_rate": 0.0002821263791374122, + "loss": 1.2637, + "theoretical_loss": 3.438566244233236, + "tokens_seen": 2148597760 + }, + { + "epoch": 0.65, + "learning_rate": 0.0002821103309929789, + "loss": 1.3767, + "theoretical_loss": 3.4385576936520446, + "tokens_seen": 2148663296 + }, + { + "epoch": 0.65, + "learning_rate": 0.0002820942828485456, + "loss": 1.2898, + "theoretical_loss": 3.4385491434046696, + "tokens_seen": 2148728832 + }, + { + "epoch": 0.65, + "learning_rate": 0.0002820782347041123, + "loss": 1.3713, + "theoretical_loss": 3.438540593491089, + "tokens_seen": 2148794368 + }, + { + "epoch": 0.65, + "learning_rate": 0.000282062186559679, + "loss": 1.2883, + "theoretical_loss": 3.4385320439112794, + "tokens_seen": 2148859904 + }, + { + "epoch": 0.65, + "learning_rate": 0.0002820461384152457, + "loss": 1.3754, + "theoretical_loss": 3.438523494665217, + "tokens_seen": 2148925440 + }, + { + "epoch": 0.65, + "learning_rate": 0.0002820300902708124, + "loss": 1.3308, + "theoretical_loss": 3.438514945752879, + "tokens_seen": 2148990976 + }, + { + "epoch": 0.65, + "learning_rate": 0.00028201404212637913, + "loss": 1.3562, + "theoretical_loss": 3.4385063971742422, + "tokens_seen": 2149056512 + }, + { + "epoch": 0.65, + "learning_rate": 0.00028199799398194583, + "loss": 1.2782, + "theoretical_loss": 3.4384978489292832, + "tokens_seen": 2149122048 + }, + { + "epoch": 0.65, + "learning_rate": 0.00028198194583751253, + "loss": 1.2839, + "theoretical_loss": 3.438489301017979, + "tokens_seen": 2149187584 + }, + { + "epoch": 0.65, + "learning_rate": 0.00028196589769307923, + "loss": 1.3484, + "theoretical_loss": 3.4384807534403063, + "tokens_seen": 2149253120 + }, + { + "epoch": 0.65, + "learning_rate": 0.00028194984954864593, + "loss": 1.2459, + "theoretical_loss": 3.4384722061962423, + "tokens_seen": 2149318656 + }, + { + "epoch": 0.65, + "learning_rate": 0.00028193380140421263, + "loss": 1.2017, + "theoretical_loss": 3.438463659285763, + "tokens_seen": 2149384192 + }, + { + "epoch": 0.65, + "learning_rate": 0.00028191775325977933, + "loss": 1.3803, + "theoretical_loss": 3.4384551127088456, + "tokens_seen": 2149449728 + }, + { + "epoch": 0.65, + "learning_rate": 0.00028190170511534603, + "loss": 1.267, + "theoretical_loss": 3.4384465664654673, + "tokens_seen": 2149515264 + }, + { + "epoch": 0.65, + "objective/train/docs_used": 782928, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.150974988937378, + "objective/train/theoretical_loss": 3.4384380205556044, + "objective/train/tokens_used": 2170040800, + "theoretical_loss": 3.4384380205556044, + "tokens_seen": 2149580800 + }, + { + "epoch": 0.65, + "learning_rate": 0.00028188565697091273, + "loss": 1.1715, + "theoretical_loss": 3.4384380205556044, + "tokens_seen": 2149580800 + }, + { + "epoch": 0.65, + "learning_rate": 0.00028186960882647943, + "loss": 1.2192, + "theoretical_loss": 3.4384294749792343, + "tokens_seen": 2149646336 + }, + { + "epoch": 0.65, + "learning_rate": 0.00028185356068204613, + "loss": 1.232, + "theoretical_loss": 3.438420929736333, + "tokens_seen": 2149711872 + }, + { + "epoch": 0.65, + "learning_rate": 0.00028183751253761283, + "loss": 1.37, + "theoretical_loss": 3.4384123848268784, + "tokens_seen": 2149777408 + }, + { + "epoch": 0.65, + "learning_rate": 0.00028182146439317953, + "loss": 1.3383, + "theoretical_loss": 3.438403840250846, + "tokens_seen": 2149842944 + }, + { + "epoch": 0.65, + "learning_rate": 0.00028180541624874624, + "loss": 1.3158, + "theoretical_loss": 3.4383952960082134, + "tokens_seen": 2149908480 + }, + { + "epoch": 0.65, + "learning_rate": 0.00028178936810431294, + "loss": 1.2264, + "theoretical_loss": 3.4383867520989577, + "tokens_seen": 2149974016 + }, + { + "epoch": 0.65, + "learning_rate": 0.00028177331995987964, + "loss": 1.3617, + "theoretical_loss": 3.438378208523055, + "tokens_seen": 2150039552 + }, + { + "epoch": 0.65, + "learning_rate": 0.00028175727181544634, + "loss": 1.4597, + "theoretical_loss": 3.4383696652804825, + "tokens_seen": 2150105088 + }, + { + "epoch": 0.65, + "learning_rate": 0.00028174122367101304, + "loss": 1.1991, + "theoretical_loss": 3.438361122371217, + "tokens_seen": 2150170624 + }, + { + "epoch": 0.65, + "learning_rate": 0.00028172517552657974, + "loss": 1.2128, + "theoretical_loss": 3.438352579795236, + "tokens_seen": 2150236160 + }, + { + "epoch": 0.65, + "learning_rate": 0.00028170912738214644, + "loss": 1.3434, + "theoretical_loss": 3.4383440375525147, + "tokens_seen": 2150301696 + }, + { + "epoch": 0.65, + "learning_rate": 0.00028169307923771314, + "loss": 1.4006, + "theoretical_loss": 3.438335495643032, + "tokens_seen": 2150367232 + }, + { + "epoch": 0.65, + "learning_rate": 0.0002816770310932799, + "loss": 1.3572, + "theoretical_loss": 3.4383269540667625, + "tokens_seen": 2150432768 + }, + { + "epoch": 0.65, + "learning_rate": 0.0002816609829488466, + "loss": 1.3769, + "theoretical_loss": 3.438318412823685, + "tokens_seen": 2150498304 + }, + { + "epoch": 0.65, + "learning_rate": 0.0002816449348044133, + "loss": 1.3128, + "theoretical_loss": 3.4383098719137757, + "tokens_seen": 2150563840 + }, + { + "epoch": 0.65, + "learning_rate": 0.00028162888665998, + "loss": 1.3676, + "theoretical_loss": 3.438301331337011, + "tokens_seen": 2150629376 + }, + { + "epoch": 0.65, + "learning_rate": 0.0002816128385155467, + "loss": 1.2975, + "theoretical_loss": 3.4382927910933683, + "tokens_seen": 2150694912 + }, + { + "epoch": 0.65, + "learning_rate": 0.0002815967903711134, + "loss": 1.2717, + "theoretical_loss": 3.438284251182824, + "tokens_seen": 2150760448 + }, + { + "epoch": 0.65, + "learning_rate": 0.0002815807422266801, + "loss": 1.2828, + "theoretical_loss": 3.438275711605356, + "tokens_seen": 2150825984 + }, + { + "epoch": 0.65, + "learning_rate": 0.00028156469408224674, + "loss": 1.2982, + "theoretical_loss": 3.4382671723609395, + "tokens_seen": 2150891520 + }, + { + "epoch": 0.65, + "learning_rate": 0.00028154864593781345, + "loss": 1.3189, + "theoretical_loss": 3.4382586334495526, + "tokens_seen": 2150957056 + }, + { + "epoch": 0.65, + "learning_rate": 0.00028153259779338015, + "loss": 1.4138, + "theoretical_loss": 3.4382500948711723, + "tokens_seen": 2151022592 + }, + { + "epoch": 0.65, + "learning_rate": 0.00028151654964894685, + "loss": 1.2524, + "theoretical_loss": 3.4382415566257745, + "tokens_seen": 2151088128 + }, + { + "epoch": 0.65, + "learning_rate": 0.00028150050150451355, + "loss": 1.3526, + "theoretical_loss": 3.438233018713336, + "tokens_seen": 2151153664 + }, + { + "epoch": 0.65, + "learning_rate": 0.00028148445336008025, + "loss": 1.2912, + "theoretical_loss": 3.438224481133835, + "tokens_seen": 2151219200 + }, + { + "epoch": 0.65, + "learning_rate": 0.00028146840521564695, + "loss": 1.1802, + "theoretical_loss": 3.438215943887247, + "tokens_seen": 2151284736 + }, + { + "epoch": 0.65, + "learning_rate": 0.00028145235707121365, + "loss": 1.3634, + "theoretical_loss": 3.43820740697355, + "tokens_seen": 2151350272 + }, + { + "epoch": 0.65, + "learning_rate": 0.00028143630892678035, + "loss": 1.3755, + "theoretical_loss": 3.4381988703927204, + "tokens_seen": 2151415808 + }, + { + "epoch": 0.65, + "learning_rate": 0.00028142026078234705, + "loss": 1.3761, + "theoretical_loss": 3.4381903341447346, + "tokens_seen": 2151481344 + }, + { + "epoch": 0.65, + "learning_rate": 0.00028140421263791375, + "loss": 1.3013, + "theoretical_loss": 3.4381817982295697, + "tokens_seen": 2151546880 + }, + { + "epoch": 0.65, + "learning_rate": 0.00028138816449348045, + "loss": 1.3361, + "theoretical_loss": 3.4381732626472035, + "tokens_seen": 2151612416 + }, + { + "epoch": 0.65, + "learning_rate": 0.00028137211634904715, + "loss": 1.347, + "theoretical_loss": 3.438164727397612, + "tokens_seen": 2151677952 + }, + { + "epoch": 0.65, + "learning_rate": 0.00028135606820461385, + "loss": 1.3273, + "theoretical_loss": 3.438156192480772, + "tokens_seen": 2151743488 + }, + { + "epoch": 0.65, + "learning_rate": 0.00028134002006018055, + "loss": 1.3756, + "theoretical_loss": 3.438147657896661, + "tokens_seen": 2151809024 + }, + { + "epoch": 0.65, + "learning_rate": 0.00028132397191574725, + "loss": 1.3497, + "theoretical_loss": 3.4381391236452554, + "tokens_seen": 2151874560 + }, + { + "epoch": 0.65, + "learning_rate": 0.00028130792377131396, + "loss": 1.2626, + "theoretical_loss": 3.438130589726532, + "tokens_seen": 2151940096 + }, + { + "epoch": 0.65, + "learning_rate": 0.00028129187562688066, + "loss": 1.3879, + "theoretical_loss": 3.4381220561404686, + "tokens_seen": 2152005632 + }, + { + "epoch": 0.65, + "learning_rate": 0.00028127582748244736, + "loss": 1.356, + "theoretical_loss": 3.4381135228870408, + "tokens_seen": 2152071168 + }, + { + "epoch": 0.65, + "learning_rate": 0.00028125977933801406, + "loss": 1.3269, + "theoretical_loss": 3.438104989966227, + "tokens_seen": 2152136704 + }, + { + "epoch": 0.65, + "learning_rate": 0.00028124373119358076, + "loss": 1.3418, + "theoretical_loss": 3.4380964573780024, + "tokens_seen": 2152202240 + }, + { + "epoch": 0.65, + "learning_rate": 0.00028122768304914746, + "loss": 1.2785, + "theoretical_loss": 3.4380879251223453, + "tokens_seen": 2152267776 + }, + { + "epoch": 0.65, + "learning_rate": 0.00028121163490471416, + "loss": 1.3533, + "theoretical_loss": 3.438079393199232, + "tokens_seen": 2152333312 + }, + { + "epoch": 0.65, + "learning_rate": 0.00028119558676028086, + "loss": 1.3068, + "theoretical_loss": 3.4380708616086393, + "tokens_seen": 2152398848 + }, + { + "epoch": 0.65, + "learning_rate": 0.00028117953861584756, + "loss": 1.2873, + "theoretical_loss": 3.4380623303505446, + "tokens_seen": 2152464384 + }, + { + "epoch": 0.65, + "learning_rate": 0.00028116349047141426, + "loss": 1.2958, + "theoretical_loss": 3.4380537994249245, + "tokens_seen": 2152529920 + }, + { + "epoch": 0.65, + "learning_rate": 0.00028114744232698096, + "loss": 1.3131, + "theoretical_loss": 3.4380452688317558, + "tokens_seen": 2152595456 + }, + { + "epoch": 0.65, + "learning_rate": 0.00028113139418254766, + "loss": 1.349, + "theoretical_loss": 3.438036738571016, + "tokens_seen": 2152660992 + }, + { + "epoch": 0.65, + "learning_rate": 0.00028111534603811436, + "loss": 1.3651, + "theoretical_loss": 3.4380282086426814, + "tokens_seen": 2152726528 + }, + { + "epoch": 0.65, + "learning_rate": 0.00028109929789368106, + "loss": 1.3001, + "theoretical_loss": 3.4380196790467292, + "tokens_seen": 2152792064 + }, + { + "epoch": 0.65, + "objective/train/docs_used": 784071, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.3354017734527588, + "objective/train/theoretical_loss": 3.4380111497831365, + "objective/train/tokens_used": 2173317600, + "theoretical_loss": 3.4380111497831365, + "tokens_seen": 2152857600 + }, + { + "epoch": 0.65, + "learning_rate": 0.00028108324974924776, + "loss": 1.3848, + "theoretical_loss": 3.4380111497831365, + "tokens_seen": 2152857600 + }, + { + "epoch": 0.65, + "learning_rate": 0.00028106720160481446, + "loss": 1.2784, + "theoretical_loss": 3.4380026208518792, + "tokens_seen": 2152923136 + }, + { + "epoch": 0.65, + "learning_rate": 0.00028105115346038117, + "loss": 1.2926, + "theoretical_loss": 3.4379940922529357, + "tokens_seen": 2152988672 + }, + { + "epoch": 0.65, + "learning_rate": 0.00028103510531594787, + "loss": 1.2926, + "theoretical_loss": 3.4379855639862824, + "tokens_seen": 2153054208 + }, + { + "epoch": 0.65, + "learning_rate": 0.00028101905717151457, + "loss": 1.4418, + "theoretical_loss": 3.4379770360518958, + "tokens_seen": 2153119744 + }, + { + "epoch": 0.65, + "learning_rate": 0.00028100300902708127, + "loss": 1.2292, + "theoretical_loss": 3.437968508449753, + "tokens_seen": 2153185280 + }, + { + "epoch": 0.65, + "learning_rate": 0.00028098696088264797, + "loss": 1.3156, + "theoretical_loss": 3.437959981179832, + "tokens_seen": 2153250816 + }, + { + "epoch": 0.65, + "learning_rate": 0.00028097091273821467, + "loss": 1.3737, + "theoretical_loss": 3.437951454242108, + "tokens_seen": 2153316352 + }, + { + "epoch": 0.65, + "learning_rate": 0.00028095486459378137, + "loss": 1.3203, + "theoretical_loss": 3.437942927636559, + "tokens_seen": 2153381888 + }, + { + "epoch": 0.65, + "learning_rate": 0.00028093881644934807, + "loss": 1.4064, + "theoretical_loss": 3.4379344013631616, + "tokens_seen": 2153447424 + }, + { + "epoch": 0.65, + "learning_rate": 0.00028092276830491477, + "loss": 1.3499, + "theoretical_loss": 3.4379258754218935, + "tokens_seen": 2153512960 + }, + { + "epoch": 0.65, + "learning_rate": 0.00028090672016048147, + "loss": 1.413, + "theoretical_loss": 3.4379173498127304, + "tokens_seen": 2153578496 + }, + { + "epoch": 0.65, + "learning_rate": 0.00028089067201604817, + "loss": 1.1789, + "theoretical_loss": 3.4379088245356506, + "tokens_seen": 2153644032 + }, + { + "epoch": 0.65, + "learning_rate": 0.00028087462387161487, + "loss": 1.1897, + "theoretical_loss": 3.43790029959063, + "tokens_seen": 2153709568 + }, + { + "epoch": 0.65, + "learning_rate": 0.00028085857572718157, + "loss": 1.2149, + "theoretical_loss": 3.437891774977646, + "tokens_seen": 2153775104 + }, + { + "epoch": 0.65, + "learning_rate": 0.0002808425275827483, + "loss": 1.2976, + "theoretical_loss": 3.437883250696675, + "tokens_seen": 2153840640 + }, + { + "epoch": 0.65, + "learning_rate": 0.000280826479438315, + "loss": 1.2966, + "theoretical_loss": 3.4378747267476957, + "tokens_seen": 2153906176 + }, + { + "epoch": 0.65, + "learning_rate": 0.0002808104312938817, + "loss": 1.3257, + "theoretical_loss": 3.4378662031306826, + "tokens_seen": 2153971712 + }, + { + "epoch": 0.65, + "learning_rate": 0.0002807943831494483, + "loss": 1.3481, + "theoretical_loss": 3.4378576798456146, + "tokens_seen": 2154037248 + }, + { + "epoch": 0.65, + "learning_rate": 0.000280778335005015, + "loss": 1.2683, + "theoretical_loss": 3.437849156892468, + "tokens_seen": 2154102784 + }, + { + "epoch": 0.65, + "learning_rate": 0.0002807622868605817, + "loss": 1.1724, + "theoretical_loss": 3.4378406342712196, + "tokens_seen": 2154168320 + }, + { + "epoch": 0.65, + "learning_rate": 0.0002807462387161484, + "loss": 1.3495, + "theoretical_loss": 3.4378321119818462, + "tokens_seen": 2154233856 + }, + { + "epoch": 0.65, + "learning_rate": 0.0002807301905717151, + "loss": 1.3254, + "theoretical_loss": 3.4378235900243257, + "tokens_seen": 2154299392 + }, + { + "epoch": 0.65, + "learning_rate": 0.0002807141424272818, + "loss": 1.2103, + "theoretical_loss": 3.4378150683986344, + "tokens_seen": 2154364928 + }, + { + "epoch": 0.65, + "learning_rate": 0.0002806980942828485, + "loss": 1.3572, + "theoretical_loss": 3.437806547104749, + "tokens_seen": 2154430464 + }, + { + "epoch": 0.65, + "learning_rate": 0.0002806820461384152, + "loss": 1.347, + "theoretical_loss": 3.4377980261426475, + "tokens_seen": 2154496000 + }, + { + "epoch": 0.65, + "learning_rate": 0.0002806659979939819, + "loss": 1.2204, + "theoretical_loss": 3.437789505512306, + "tokens_seen": 2154561536 + }, + { + "epoch": 0.65, + "learning_rate": 0.0002806499498495486, + "loss": 1.2818, + "theoretical_loss": 3.437780985213702, + "tokens_seen": 2154627072 + }, + { + "epoch": 0.65, + "learning_rate": 0.00028063390170511533, + "loss": 1.242, + "theoretical_loss": 3.4377724652468116, + "tokens_seen": 2154692608 + }, + { + "epoch": 0.65, + "learning_rate": 0.00028061785356068203, + "loss": 1.3577, + "theoretical_loss": 3.4377639456116134, + "tokens_seen": 2154758144 + }, + { + "epoch": 0.65, + "learning_rate": 0.00028060180541624873, + "loss": 1.3546, + "theoretical_loss": 3.4377554263080827, + "tokens_seen": 2154823680 + }, + { + "epoch": 0.65, + "learning_rate": 0.00028058575727181543, + "loss": 1.3323, + "theoretical_loss": 3.4377469073361975, + "tokens_seen": 2154889216 + }, + { + "epoch": 0.65, + "learning_rate": 0.00028056970912738213, + "loss": 1.2743, + "theoretical_loss": 3.4377383886959345, + "tokens_seen": 2154954752 + }, + { + "epoch": 0.65, + "learning_rate": 0.00028055366098294883, + "loss": 1.2918, + "theoretical_loss": 3.4377298703872707, + "tokens_seen": 2155020288 + }, + { + "epoch": 0.65, + "learning_rate": 0.00028053761283851553, + "loss": 1.3911, + "theoretical_loss": 3.4377213524101835, + "tokens_seen": 2155085824 + }, + { + "epoch": 0.65, + "learning_rate": 0.00028052156469408223, + "loss": 1.195, + "theoretical_loss": 3.4377128347646497, + "tokens_seen": 2155151360 + }, + { + "epoch": 0.65, + "learning_rate": 0.00028050551654964893, + "loss": 1.2815, + "theoretical_loss": 3.4377043174506454, + "tokens_seen": 2155216896 + }, + { + "epoch": 0.65, + "learning_rate": 0.00028048946840521563, + "loss": 1.3291, + "theoretical_loss": 3.4376958004681493, + "tokens_seen": 2155282432 + }, + { + "epoch": 0.65, + "learning_rate": 0.0002804734202607824, + "loss": 1.3475, + "theoretical_loss": 3.4376872838171373, + "tokens_seen": 2155347968 + }, + { + "epoch": 0.65, + "learning_rate": 0.0002804573721163491, + "loss": 1.2946, + "theoretical_loss": 3.4376787674975864, + "tokens_seen": 2155413504 + }, + { + "epoch": 0.65, + "learning_rate": 0.0002804413239719158, + "loss": 1.2624, + "theoretical_loss": 3.437670251509474, + "tokens_seen": 2155479040 + }, + { + "epoch": 0.65, + "learning_rate": 0.0002804252758274825, + "loss": 1.2613, + "theoretical_loss": 3.4376617358527772, + "tokens_seen": 2155544576 + }, + { + "epoch": 0.65, + "learning_rate": 0.0002804092276830492, + "loss": 1.4026, + "theoretical_loss": 3.4376532205274724, + "tokens_seen": 2155610112 + }, + { + "epoch": 0.65, + "learning_rate": 0.0002803931795386159, + "loss": 1.3498, + "theoretical_loss": 3.437644705533537, + "tokens_seen": 2155675648 + }, + { + "epoch": 0.65, + "learning_rate": 0.0002803771313941826, + "loss": 1.3897, + "theoretical_loss": 3.4376361908709487, + "tokens_seen": 2155741184 + }, + { + "epoch": 0.65, + "learning_rate": 0.0002803610832497493, + "loss": 1.2292, + "theoretical_loss": 3.437627676539684, + "tokens_seen": 2155806720 + }, + { + "epoch": 0.65, + "learning_rate": 0.000280345035105316, + "loss": 1.2598, + "theoretical_loss": 3.437619162539719, + "tokens_seen": 2155872256 + }, + { + "epoch": 0.65, + "learning_rate": 0.0002803289869608827, + "loss": 1.2983, + "theoretical_loss": 3.4376106488710327, + "tokens_seen": 2155937792 + }, + { + "epoch": 0.65, + "learning_rate": 0.0002803129388164494, + "loss": 1.3235, + "theoretical_loss": 3.4376021355336004, + "tokens_seen": 2156003328 + }, + { + "epoch": 0.65, + "learning_rate": 0.0002802968906720161, + "loss": 1.3894, + "theoretical_loss": 3.4375936225273995, + "tokens_seen": 2156068864 + }, + { + "epoch": 0.65, + "objective/train/docs_used": 785270, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.3148300647735596, + "objective/train/theoretical_loss": 3.4375851098524075, + "objective/train/tokens_used": 2176594400, + "theoretical_loss": 3.4375851098524075, + "tokens_seen": 2156134400 + }, + { + "epoch": 0.65, + "learning_rate": 0.0002802808425275828, + "loss": 1.327, + "theoretical_loss": 3.4375851098524075, + "tokens_seen": 2156134400 + }, + { + "epoch": 0.65, + "learning_rate": 0.0002802647943831495, + "loss": 1.2891, + "theoretical_loss": 3.4375765975086017, + "tokens_seen": 2156199936 + }, + { + "epoch": 0.65, + "learning_rate": 0.0002802487462387162, + "loss": 1.268, + "theoretical_loss": 3.4375680854959585, + "tokens_seen": 2156265472 + }, + { + "epoch": 0.65, + "learning_rate": 0.0002802326980942829, + "loss": 1.3356, + "theoretical_loss": 3.437559573814455, + "tokens_seen": 2156331008 + }, + { + "epoch": 0.65, + "learning_rate": 0.0002802166499498496, + "loss": 1.3244, + "theoretical_loss": 3.4375510624640686, + "tokens_seen": 2156396544 + }, + { + "epoch": 0.65, + "learning_rate": 0.0002802006018054163, + "loss": 1.3032, + "theoretical_loss": 3.437542551444776, + "tokens_seen": 2156462080 + }, + { + "epoch": 0.65, + "learning_rate": 0.000280184553660983, + "loss": 1.2884, + "theoretical_loss": 3.437534040756555, + "tokens_seen": 2156527616 + }, + { + "epoch": 0.65, + "learning_rate": 0.0002801685055165497, + "loss": 1.2569, + "theoretical_loss": 3.4375255303993812, + "tokens_seen": 2156593152 + }, + { + "epoch": 0.65, + "learning_rate": 0.00028015245737211635, + "loss": 1.2693, + "theoretical_loss": 3.4375170203732335, + "tokens_seen": 2156658688 + }, + { + "epoch": 0.65, + "learning_rate": 0.00028013640922768305, + "loss": 1.4133, + "theoretical_loss": 3.4375085106780876, + "tokens_seen": 2156724224 + }, + { + "epoch": 0.65, + "learning_rate": 0.00028012036108324975, + "loss": 1.2504, + "theoretical_loss": 3.437500001313921, + "tokens_seen": 2156789760 + }, + { + "epoch": 0.65, + "learning_rate": 0.00028010431293881645, + "loss": 1.2877, + "theoretical_loss": 3.437491492280711, + "tokens_seen": 2156855296 + }, + { + "epoch": 0.65, + "learning_rate": 0.00028008826479438315, + "loss": 1.2767, + "theoretical_loss": 3.437482983578434, + "tokens_seen": 2156920832 + }, + { + "epoch": 0.65, + "learning_rate": 0.00028007221664994985, + "loss": 1.3571, + "theoretical_loss": 3.437474475207068, + "tokens_seen": 2156986368 + }, + { + "epoch": 0.65, + "learning_rate": 0.00028005616850551655, + "loss": 1.3544, + "theoretical_loss": 3.437465967166589, + "tokens_seen": 2157051904 + }, + { + "epoch": 0.65, + "learning_rate": 0.00028004012036108325, + "loss": 1.2748, + "theoretical_loss": 3.4374574594569753, + "tokens_seen": 2157117440 + }, + { + "epoch": 0.65, + "learning_rate": 0.00028002407221664995, + "loss": 1.2562, + "theoretical_loss": 3.437448952078203, + "tokens_seen": 2157182976 + }, + { + "epoch": 0.65, + "learning_rate": 0.00028000802407221665, + "loss": 1.3519, + "theoretical_loss": 3.4374404450302496, + "tokens_seen": 2157248512 + }, + { + "epoch": 0.65, + "learning_rate": 0.00027999197592778335, + "loss": 1.2446, + "theoretical_loss": 3.4374319383130922, + "tokens_seen": 2157314048 + }, + { + "epoch": 0.65, + "learning_rate": 0.00027997592778335005, + "loss": 1.2139, + "theoretical_loss": 3.4374234319267076, + "tokens_seen": 2157379584 + }, + { + "epoch": 0.65, + "learning_rate": 0.00027995987963891675, + "loss": 1.2771, + "theoretical_loss": 3.4374149258710736, + "tokens_seen": 2157445120 + }, + { + "epoch": 0.65, + "learning_rate": 0.00027994383149448345, + "loss": 1.3427, + "theoretical_loss": 3.437406420146167, + "tokens_seen": 2157510656 + }, + { + "epoch": 0.65, + "learning_rate": 0.00027992778335005016, + "loss": 1.2734, + "theoretical_loss": 3.4373979147519638, + "tokens_seen": 2157576192 + }, + { + "epoch": 0.65, + "learning_rate": 0.00027991173520561686, + "loss": 1.3306, + "theoretical_loss": 3.4373894096884423, + "tokens_seen": 2157641728 + }, + { + "epoch": 0.65, + "learning_rate": 0.00027989568706118356, + "loss": 1.2529, + "theoretical_loss": 3.4373809049555795, + "tokens_seen": 2157707264 + }, + { + "epoch": 0.65, + "learning_rate": 0.00027987963891675026, + "loss": 1.2196, + "theoretical_loss": 3.437372400553352, + "tokens_seen": 2157772800 + }, + { + "epoch": 0.65, + "learning_rate": 0.00027986359077231696, + "loss": 1.271, + "theoretical_loss": 3.4373638964817377, + "tokens_seen": 2157838336 + }, + { + "epoch": 0.65, + "learning_rate": 0.00027984754262788366, + "loss": 1.409, + "theoretical_loss": 3.437355392740713, + "tokens_seen": 2157903872 + }, + { + "epoch": 0.65, + "learning_rate": 0.00027983149448345036, + "loss": 1.3124, + "theoretical_loss": 3.437346889330255, + "tokens_seen": 2157969408 + }, + { + "epoch": 0.65, + "learning_rate": 0.00027981544633901706, + "loss": 1.3441, + "theoretical_loss": 3.4373383862503415, + "tokens_seen": 2158034944 + }, + { + "epoch": 0.65, + "learning_rate": 0.00027979939819458376, + "loss": 1.2618, + "theoretical_loss": 3.4373298835009485, + "tokens_seen": 2158100480 + }, + { + "epoch": 0.65, + "learning_rate": 0.00027978335005015046, + "loss": 1.3626, + "theoretical_loss": 3.437321381082054, + "tokens_seen": 2158166016 + }, + { + "epoch": 0.65, + "learning_rate": 0.00027976730190571716, + "loss": 1.2739, + "theoretical_loss": 3.437312878993635, + "tokens_seen": 2158231552 + }, + { + "epoch": 0.65, + "learning_rate": 0.00027975125376128386, + "loss": 1.3407, + "theoretical_loss": 3.4373043772356686, + "tokens_seen": 2158297088 + }, + { + "epoch": 0.65, + "learning_rate": 0.00027973520561685056, + "loss": 1.3134, + "theoretical_loss": 3.4372958758081316, + "tokens_seen": 2158362624 + }, + { + "epoch": 0.65, + "learning_rate": 0.00027971915747241726, + "loss": 1.2824, + "theoretical_loss": 3.4372873747110013, + "tokens_seen": 2158428160 + }, + { + "epoch": 0.65, + "learning_rate": 0.00027970310932798396, + "loss": 1.3351, + "theoretical_loss": 3.4372788739442552, + "tokens_seen": 2158493696 + }, + { + "epoch": 0.65, + "learning_rate": 0.00027968706118355066, + "loss": 1.3337, + "theoretical_loss": 3.43727037350787, + "tokens_seen": 2158559232 + }, + { + "epoch": 0.65, + "learning_rate": 0.00027967101303911737, + "loss": 1.3992, + "theoretical_loss": 3.4372618734018228, + "tokens_seen": 2158624768 + }, + { + "epoch": 0.65, + "learning_rate": 0.00027965496489468407, + "loss": 1.2777, + "theoretical_loss": 3.437253373626091, + "tokens_seen": 2158690304 + }, + { + "epoch": 0.65, + "learning_rate": 0.00027963891675025077, + "loss": 1.3359, + "theoretical_loss": 3.437244874180651, + "tokens_seen": 2158755840 + }, + { + "epoch": 0.65, + "learning_rate": 0.00027962286860581747, + "loss": 1.2835, + "theoretical_loss": 3.4372363750654813, + "tokens_seen": 2158821376 + }, + { + "epoch": 0.65, + "learning_rate": 0.00027960682046138417, + "loss": 1.2981, + "theoretical_loss": 3.4372278762805584, + "tokens_seen": 2158886912 + }, + { + "epoch": 0.65, + "learning_rate": 0.00027959077231695087, + "loss": 1.2862, + "theoretical_loss": 3.437219377825859, + "tokens_seen": 2158952448 + }, + { + "epoch": 0.65, + "learning_rate": 0.00027957472417251757, + "loss": 1.3196, + "theoretical_loss": 3.43721087970136, + "tokens_seen": 2159017984 + }, + { + "epoch": 0.65, + "learning_rate": 0.00027955867602808427, + "loss": 1.2358, + "theoretical_loss": 3.43720238190704, + "tokens_seen": 2159083520 + }, + { + "epoch": 0.65, + "learning_rate": 0.00027954262788365097, + "loss": 1.3127, + "theoretical_loss": 3.437193884442875, + "tokens_seen": 2159149056 + }, + { + "epoch": 0.65, + "learning_rate": 0.00027952657973921767, + "loss": 1.3886, + "theoretical_loss": 3.4371853873088423, + "tokens_seen": 2159214592 + }, + { + "epoch": 0.65, + "learning_rate": 0.00027951053159478437, + "loss": 1.3254, + "theoretical_loss": 3.4371768905049196, + "tokens_seen": 2159280128 + }, + { + "epoch": 0.65, + "learning_rate": 0.00027949448345035107, + "loss": 1.3108, + "theoretical_loss": 3.437168394031083, + "tokens_seen": 2159345664 + }, + { + "epoch": 0.65, + "objective/train/docs_used": 786428, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.31767737865448, + "objective/train/theoretical_loss": 3.4371598978873106, + "objective/train/tokens_used": 2179871200, + "theoretical_loss": 3.4371598978873106, + "tokens_seen": 2159411200 + }, + { + "epoch": 0.65, + "learning_rate": 0.0002794784353059178, + "loss": 1.2518, + "theoretical_loss": 3.4371598978873106, + "tokens_seen": 2159411200 + }, + { + "epoch": 0.65, + "learning_rate": 0.0002794623871614845, + "loss": 1.2612, + "theoretical_loss": 3.4371514020735794, + "tokens_seen": 2159476736 + }, + { + "epoch": 0.65, + "learning_rate": 0.0002794463390170512, + "loss": 1.3445, + "theoretical_loss": 3.437142906589866, + "tokens_seen": 2159542272 + }, + { + "epoch": 0.65, + "learning_rate": 0.0002794302908726179, + "loss": 1.2967, + "theoretical_loss": 3.437134411436148, + "tokens_seen": 2159607808 + }, + { + "epoch": 0.65, + "learning_rate": 0.0002794142427281846, + "loss": 1.2189, + "theoretical_loss": 3.437125916612403, + "tokens_seen": 2159673344 + }, + { + "epoch": 0.65, + "learning_rate": 0.0002793981945837513, + "loss": 1.2784, + "theoretical_loss": 3.4371174221186074, + "tokens_seen": 2159738880 + }, + { + "epoch": 0.65, + "learning_rate": 0.0002793821464393179, + "loss": 1.2779, + "theoretical_loss": 3.437108927954739, + "tokens_seen": 2159804416 + }, + { + "epoch": 0.65, + "learning_rate": 0.0002793660982948846, + "loss": 1.3293, + "theoretical_loss": 3.4371004341207745, + "tokens_seen": 2159869952 + }, + { + "epoch": 0.65, + "learning_rate": 0.0002793500501504513, + "loss": 1.3084, + "theoretical_loss": 3.4370919406166913, + "tokens_seen": 2159935488 + }, + { + "epoch": 0.65, + "learning_rate": 0.000279334002006018, + "loss": 1.2668, + "theoretical_loss": 3.4370834474424665, + "tokens_seen": 2160001024 + }, + { + "epoch": 0.65, + "learning_rate": 0.0002793179538615847, + "loss": 1.3798, + "theoretical_loss": 3.437074954598077, + "tokens_seen": 2160066560 + }, + { + "epoch": 0.65, + "learning_rate": 0.0002793019057171514, + "loss": 1.2041, + "theoretical_loss": 3.437066462083501, + "tokens_seen": 2160132096 + }, + { + "epoch": 0.65, + "learning_rate": 0.0002792858575727181, + "loss": 1.2736, + "theoretical_loss": 3.4370579698987145, + "tokens_seen": 2160197632 + }, + { + "epoch": 0.65, + "learning_rate": 0.0002792698094282849, + "loss": 1.3032, + "theoretical_loss": 3.4370494780436953, + "tokens_seen": 2160263168 + }, + { + "epoch": 0.65, + "learning_rate": 0.0002792537612838516, + "loss": 1.2857, + "theoretical_loss": 3.4370409865184204, + "tokens_seen": 2160328704 + }, + { + "epoch": 0.65, + "learning_rate": 0.0002792377131394183, + "loss": 1.2566, + "theoretical_loss": 3.437032495322867, + "tokens_seen": 2160394240 + }, + { + "epoch": 0.65, + "learning_rate": 0.000279221664994985, + "loss": 1.3183, + "theoretical_loss": 3.437024004457012, + "tokens_seen": 2160459776 + }, + { + "epoch": 0.65, + "learning_rate": 0.0002792056168505517, + "loss": 1.2656, + "theoretical_loss": 3.4370155139208336, + "tokens_seen": 2160525312 + }, + { + "epoch": 0.65, + "learning_rate": 0.0002791895687061184, + "loss": 1.227, + "theoretical_loss": 3.437007023714308, + "tokens_seen": 2160590848 + }, + { + "epoch": 0.65, + "learning_rate": 0.0002791735205616851, + "loss": 1.2762, + "theoretical_loss": 3.436998533837413, + "tokens_seen": 2160656384 + }, + { + "epoch": 0.65, + "learning_rate": 0.0002791574724172518, + "loss": 1.3158, + "theoretical_loss": 3.436990044290125, + "tokens_seen": 2160721920 + }, + { + "epoch": 0.65, + "learning_rate": 0.0002791414242728185, + "loss": 1.3213, + "theoretical_loss": 3.4369815550724225, + "tokens_seen": 2160787456 + }, + { + "epoch": 0.65, + "learning_rate": 0.0002791253761283852, + "loss": 1.2352, + "theoretical_loss": 3.4369730661842817, + "tokens_seen": 2160852992 + }, + { + "epoch": 0.65, + "learning_rate": 0.0002791093279839519, + "loss": 1.3856, + "theoretical_loss": 3.43696457762568, + "tokens_seen": 2160918528 + }, + { + "epoch": 0.65, + "learning_rate": 0.0002790932798395186, + "loss": 1.3664, + "theoretical_loss": 3.4369560893965945, + "tokens_seen": 2160984064 + }, + { + "epoch": 0.65, + "learning_rate": 0.0002790772316950853, + "loss": 1.3543, + "theoretical_loss": 3.436947601497003, + "tokens_seen": 2161049600 + }, + { + "epoch": 0.65, + "learning_rate": 0.000279061183550652, + "loss": 1.2518, + "theoretical_loss": 3.436939113926882, + "tokens_seen": 2161115136 + }, + { + "epoch": 0.65, + "learning_rate": 0.0002790451354062187, + "loss": 1.3491, + "theoretical_loss": 3.436930626686209, + "tokens_seen": 2161180672 + }, + { + "epoch": 0.65, + "learning_rate": 0.0002790290872617854, + "loss": 1.277, + "theoretical_loss": 3.436922139774962, + "tokens_seen": 2161246208 + }, + { + "epoch": 0.65, + "learning_rate": 0.0002790130391173521, + "loss": 1.4596, + "theoretical_loss": 3.436913653193117, + "tokens_seen": 2161311744 + }, + { + "epoch": 0.65, + "learning_rate": 0.0002789969909729188, + "loss": 1.2985, + "theoretical_loss": 3.4369051669406514, + "tokens_seen": 2161377280 + }, + { + "epoch": 0.65, + "learning_rate": 0.0002789809428284855, + "loss": 1.2767, + "theoretical_loss": 3.436896681017543, + "tokens_seen": 2161442816 + }, + { + "epoch": 0.66, + "learning_rate": 0.0002789648946840522, + "loss": 1.3739, + "theoretical_loss": 3.436888195423769, + "tokens_seen": 2161508352 + }, + { + "epoch": 0.66, + "learning_rate": 0.0002789488465396189, + "loss": 1.2327, + "theoretical_loss": 3.4368797101593063, + "tokens_seen": 2161573888 + }, + { + "epoch": 0.66, + "learning_rate": 0.0002789327983951856, + "loss": 1.2641, + "theoretical_loss": 3.436871225224132, + "tokens_seen": 2161639424 + }, + { + "epoch": 0.66, + "learning_rate": 0.0002789167502507523, + "loss": 1.2218, + "theoretical_loss": 3.4368627406182237, + "tokens_seen": 2161704960 + }, + { + "epoch": 0.66, + "learning_rate": 0.000278900702106319, + "loss": 1.3477, + "theoretical_loss": 3.436854256341559, + "tokens_seen": 2161770496 + }, + { + "epoch": 0.66, + "learning_rate": 0.0002788846539618857, + "loss": 1.1856, + "theoretical_loss": 3.436845772394114, + "tokens_seen": 2161836032 + }, + { + "epoch": 0.66, + "learning_rate": 0.0002788686058174524, + "loss": 1.3074, + "theoretical_loss": 3.4368372887758674, + "tokens_seen": 2161901568 + }, + { + "epoch": 0.66, + "learning_rate": 0.0002788525576730191, + "loss": 1.2625, + "theoretical_loss": 3.436828805486795, + "tokens_seen": 2161967104 + }, + { + "epoch": 0.66, + "learning_rate": 0.0002788365095285858, + "loss": 1.3279, + "theoretical_loss": 3.436820322526875, + "tokens_seen": 2162032640 + }, + { + "epoch": 0.66, + "learning_rate": 0.0002788204613841525, + "loss": 1.3552, + "theoretical_loss": 3.4368118398960843, + "tokens_seen": 2162098176 + }, + { + "epoch": 0.66, + "learning_rate": 0.0002788044132397192, + "loss": 1.3104, + "theoretical_loss": 3.4368033575944, + "tokens_seen": 2162163712 + }, + { + "epoch": 0.66, + "learning_rate": 0.0002787883650952859, + "loss": 1.4236, + "theoretical_loss": 3.4367948756218, + "tokens_seen": 2162229248 + }, + { + "epoch": 0.66, + "learning_rate": 0.0002787723169508526, + "loss": 1.377, + "theoretical_loss": 3.4367863939782612, + "tokens_seen": 2162294784 + }, + { + "epoch": 0.66, + "learning_rate": 0.00027875626880641925, + "loss": 1.2847, + "theoretical_loss": 3.4367779126637603, + "tokens_seen": 2162360320 + }, + { + "epoch": 0.66, + "learning_rate": 0.00027874022066198595, + "loss": 1.2927, + "theoretical_loss": 3.4367694316782753, + "tokens_seen": 2162425856 + }, + { + "epoch": 0.66, + "learning_rate": 0.00027872417251755265, + "loss": 1.1986, + "theoretical_loss": 3.4367609510217836, + "tokens_seen": 2162491392 + }, + { + "epoch": 0.66, + "learning_rate": 0.00027870812437311935, + "loss": 1.3308, + "theoretical_loss": 3.4367524706942616, + "tokens_seen": 2162556928 + }, + { + "epoch": 0.66, + "learning_rate": 0.00027869207622868605, + "loss": 1.3204, + "theoretical_loss": 3.436743990695687, + "tokens_seen": 2162622464 + }, + { + "epoch": 0.66, + "objective/train/docs_used": 787745, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.4408984184265137, + "objective/train/theoretical_loss": 3.4367355110260376, + "objective/train/tokens_used": 2183148000, + "theoretical_loss": 3.4367355110260376, + "tokens_seen": 2162688000 + }, + { + "epoch": 0.66, + "learning_rate": 0.00027867602808425275, + "loss": 1.4219, + "theoretical_loss": 3.4367355110260376, + "tokens_seen": 2162688000 + }, + { + "epoch": 0.66, + "learning_rate": 0.00027865997993981945, + "loss": 1.3995, + "theoretical_loss": 3.4367270316852903, + "tokens_seen": 2162753536 + }, + { + "epoch": 0.66, + "learning_rate": 0.00027864393179538615, + "loss": 1.3531, + "theoretical_loss": 3.4367185526734216, + "tokens_seen": 2162819072 + }, + { + "epoch": 0.66, + "learning_rate": 0.00027862788365095285, + "loss": 1.2603, + "theoretical_loss": 3.43671007399041, + "tokens_seen": 2162884608 + }, + { + "epoch": 0.66, + "learning_rate": 0.00027861183550651955, + "loss": 1.3369, + "theoretical_loss": 3.436701595636232, + "tokens_seen": 2162950144 + }, + { + "epoch": 0.66, + "learning_rate": 0.00027859578736208625, + "loss": 1.4936, + "theoretical_loss": 3.4366931176108655, + "tokens_seen": 2163015680 + }, + { + "epoch": 0.66, + "learning_rate": 0.00027857973921765295, + "loss": 1.3649, + "theoretical_loss": 3.4366846399142874, + "tokens_seen": 2163081216 + }, + { + "epoch": 0.66, + "learning_rate": 0.00027856369107321966, + "loss": 1.3479, + "theoretical_loss": 3.436676162546475, + "tokens_seen": 2163146752 + }, + { + "epoch": 0.66, + "learning_rate": 0.00027854764292878636, + "loss": 1.2446, + "theoretical_loss": 3.4366676855074054, + "tokens_seen": 2163212288 + }, + { + "epoch": 0.66, + "learning_rate": 0.00027853159478435306, + "loss": 1.3831, + "theoretical_loss": 3.4366592087970558, + "tokens_seen": 2163277824 + }, + { + "epoch": 0.66, + "learning_rate": 0.00027851554663991976, + "loss": 1.2727, + "theoretical_loss": 3.436650732415404, + "tokens_seen": 2163343360 + }, + { + "epoch": 0.66, + "learning_rate": 0.00027849949849548646, + "loss": 1.3642, + "theoretical_loss": 3.4366422563624277, + "tokens_seen": 2163408896 + }, + { + "epoch": 0.66, + "learning_rate": 0.00027848345035105316, + "loss": 1.3081, + "theoretical_loss": 3.436633780638103, + "tokens_seen": 2163474432 + }, + { + "epoch": 0.66, + "learning_rate": 0.00027846740220661986, + "loss": 1.3628, + "theoretical_loss": 3.436625305242408, + "tokens_seen": 2163539968 + }, + { + "epoch": 0.66, + "learning_rate": 0.00027845135406218656, + "loss": 1.4115, + "theoretical_loss": 3.4366168301753195, + "tokens_seen": 2163605504 + }, + { + "epoch": 0.66, + "learning_rate": 0.00027843530591775326, + "loss": 1.3296, + "theoretical_loss": 3.4366083554368156, + "tokens_seen": 2163671040 + }, + { + "epoch": 0.66, + "learning_rate": 0.00027841925777331996, + "loss": 1.2725, + "theoretical_loss": 3.4365998810268725, + "tokens_seen": 2163736576 + }, + { + "epoch": 0.66, + "learning_rate": 0.00027840320962888666, + "loss": 1.2708, + "theoretical_loss": 3.4365914069454684, + "tokens_seen": 2163802112 + }, + { + "epoch": 0.66, + "learning_rate": 0.00027838716148445336, + "loss": 1.3562, + "theoretical_loss": 3.4365829331925806, + "tokens_seen": 2163867648 + }, + { + "epoch": 0.66, + "learning_rate": 0.00027837111334002006, + "loss": 1.2121, + "theoretical_loss": 3.436574459768186, + "tokens_seen": 2163933184 + }, + { + "epoch": 0.66, + "learning_rate": 0.00027835506519558676, + "loss": 1.3044, + "theoretical_loss": 3.436565986672262, + "tokens_seen": 2163998720 + }, + { + "epoch": 0.66, + "learning_rate": 0.00027833901705115346, + "loss": 1.2941, + "theoretical_loss": 3.436557513904786, + "tokens_seen": 2164064256 + }, + { + "epoch": 0.66, + "learning_rate": 0.00027832296890672016, + "loss": 1.2984, + "theoretical_loss": 3.436549041465735, + "tokens_seen": 2164129792 + }, + { + "epoch": 0.66, + "learning_rate": 0.00027830692076228687, + "loss": 1.2845, + "theoretical_loss": 3.436540569355087, + "tokens_seen": 2164195328 + }, + { + "epoch": 0.66, + "learning_rate": 0.00027829087261785357, + "loss": 1.3233, + "theoretical_loss": 3.436532097572819, + "tokens_seen": 2164260864 + }, + { + "epoch": 0.66, + "learning_rate": 0.00027827482447342027, + "loss": 1.2538, + "theoretical_loss": 3.436523626118908, + "tokens_seen": 2164326400 + }, + { + "epoch": 0.66, + "learning_rate": 0.00027825877632898697, + "loss": 1.3196, + "theoretical_loss": 3.436515154993332, + "tokens_seen": 2164391936 + }, + { + "epoch": 0.66, + "learning_rate": 0.00027824272818455367, + "loss": 1.35, + "theoretical_loss": 3.436506684196068, + "tokens_seen": 2164457472 + }, + { + "epoch": 0.66, + "learning_rate": 0.00027822668004012037, + "loss": 1.3111, + "theoretical_loss": 3.4364982137270927, + "tokens_seen": 2164523008 + }, + { + "epoch": 0.66, + "learning_rate": 0.00027821063189568707, + "loss": 1.3047, + "theoretical_loss": 3.4364897435863844, + "tokens_seen": 2164588544 + }, + { + "epoch": 0.66, + "learning_rate": 0.00027819458375125377, + "loss": 1.2159, + "theoretical_loss": 3.43648127377392, + "tokens_seen": 2164654080 + }, + { + "epoch": 0.66, + "learning_rate": 0.00027817853560682047, + "loss": 1.3283, + "theoretical_loss": 3.436472804289677, + "tokens_seen": 2164719616 + }, + { + "epoch": 0.66, + "learning_rate": 0.00027816248746238717, + "loss": 1.3607, + "theoretical_loss": 3.436464335133633, + "tokens_seen": 2164785152 + }, + { + "epoch": 0.66, + "learning_rate": 0.00027814643931795387, + "loss": 1.3455, + "theoretical_loss": 3.4364558663057645, + "tokens_seen": 2164850688 + }, + { + "epoch": 0.66, + "learning_rate": 0.00027813039117352057, + "loss": 1.32, + "theoretical_loss": 3.4364473978060492, + "tokens_seen": 2164916224 + }, + { + "epoch": 0.66, + "learning_rate": 0.0002781143430290873, + "loss": 1.289, + "theoretical_loss": 3.4364389296344653, + "tokens_seen": 2164981760 + }, + { + "epoch": 0.66, + "learning_rate": 0.000278098294884654, + "loss": 1.2452, + "theoretical_loss": 3.4364304617909887, + "tokens_seen": 2165047296 + }, + { + "epoch": 0.66, + "learning_rate": 0.0002780822467402207, + "loss": 1.3918, + "theoretical_loss": 3.436421994275598, + "tokens_seen": 2165112832 + }, + { + "epoch": 0.66, + "learning_rate": 0.0002780661985957874, + "loss": 1.3465, + "theoretical_loss": 3.43641352708827, + "tokens_seen": 2165178368 + }, + { + "epoch": 0.66, + "learning_rate": 0.0002780501504513541, + "loss": 1.301, + "theoretical_loss": 3.436405060228982, + "tokens_seen": 2165243904 + }, + { + "epoch": 0.66, + "learning_rate": 0.0002780341023069208, + "loss": 1.239, + "theoretical_loss": 3.4363965936977117, + "tokens_seen": 2165309440 + }, + { + "epoch": 0.66, + "learning_rate": 0.0002780180541624875, + "loss": 1.3534, + "theoretical_loss": 3.436388127494436, + "tokens_seen": 2165374976 + }, + { + "epoch": 0.66, + "learning_rate": 0.0002780020060180542, + "loss": 1.3122, + "theoretical_loss": 3.4363796616191324, + "tokens_seen": 2165440512 + }, + { + "epoch": 0.66, + "learning_rate": 0.0002779859578736209, + "loss": 1.2252, + "theoretical_loss": 3.4363711960717787, + "tokens_seen": 2165506048 + }, + { + "epoch": 0.66, + "learning_rate": 0.0002779699097291876, + "loss": 1.3637, + "theoretical_loss": 3.436362730852352, + "tokens_seen": 2165571584 + }, + { + "epoch": 0.66, + "learning_rate": 0.0002779538615847543, + "loss": 1.3227, + "theoretical_loss": 3.4363542659608295, + "tokens_seen": 2165637120 + }, + { + "epoch": 0.66, + "learning_rate": 0.000277937813440321, + "loss": 1.3041, + "theoretical_loss": 3.436345801397189, + "tokens_seen": 2165702656 + }, + { + "epoch": 0.66, + "learning_rate": 0.0002779217652958877, + "loss": 1.2434, + "theoretical_loss": 3.4363373371614068, + "tokens_seen": 2165768192 + }, + { + "epoch": 0.66, + "learning_rate": 0.0002779057171514544, + "loss": 1.3919, + "theoretical_loss": 3.4363288732534616, + "tokens_seen": 2165833728 + }, + { + "epoch": 0.66, + "learning_rate": 0.0002778896690070211, + "loss": 1.3333, + "theoretical_loss": 3.43632040967333, + "tokens_seen": 2165899264 + }, + { + "epoch": 0.66, + "objective/train/docs_used": 788845, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.2983404397964478, + "objective/train/theoretical_loss": 3.4363119464209904, + "objective/train/tokens_used": 2186424800, + "theoretical_loss": 3.4363119464209904, + "tokens_seen": 2165964800 + }, + { + "epoch": 0.66, + "learning_rate": 0.0002778736208625878, + "loss": 1.3434, + "theoretical_loss": 3.4363119464209904, + "tokens_seen": 2165964800 + }, + { + "epoch": 0.66, + "learning_rate": 0.0002778575727181545, + "loss": 1.3435, + "theoretical_loss": 3.4363034834964186, + "tokens_seen": 2166030336 + }, + { + "epoch": 0.66, + "learning_rate": 0.0002778415245737212, + "loss": 1.3573, + "theoretical_loss": 3.4362950208995935, + "tokens_seen": 2166095872 + }, + { + "epoch": 0.66, + "learning_rate": 0.0002778254764292879, + "loss": 1.4521, + "theoretical_loss": 3.436286558630491, + "tokens_seen": 2166161408 + }, + { + "epoch": 0.66, + "learning_rate": 0.0002778094282848546, + "loss": 1.5002, + "theoretical_loss": 3.4362780966890902, + "tokens_seen": 2166226944 + }, + { + "epoch": 0.66, + "learning_rate": 0.0002777933801404213, + "loss": 1.3893, + "theoretical_loss": 3.436269635075367, + "tokens_seen": 2166292480 + }, + { + "epoch": 0.66, + "learning_rate": 0.000277777331995988, + "loss": 1.3793, + "theoretical_loss": 3.4362611737893, + "tokens_seen": 2166358016 + }, + { + "epoch": 0.66, + "learning_rate": 0.0002777612838515547, + "loss": 1.2998, + "theoretical_loss": 3.436252712830865, + "tokens_seen": 2166423552 + }, + { + "epoch": 0.66, + "learning_rate": 0.0002777452357071214, + "loss": 1.2628, + "theoretical_loss": 3.436244252200041, + "tokens_seen": 2166489088 + }, + { + "epoch": 0.66, + "learning_rate": 0.0002777291875626881, + "loss": 1.2634, + "theoretical_loss": 3.436235791896805, + "tokens_seen": 2166554624 + }, + { + "epoch": 0.66, + "learning_rate": 0.0002777131394182548, + "loss": 1.3195, + "theoretical_loss": 3.436227331921134, + "tokens_seen": 2166620160 + }, + { + "epoch": 0.66, + "learning_rate": 0.0002776970912738215, + "loss": 1.4045, + "theoretical_loss": 3.4362188722730056, + "tokens_seen": 2166685696 + }, + { + "epoch": 0.66, + "learning_rate": 0.0002776810431293882, + "loss": 1.2093, + "theoretical_loss": 3.4362104129523976, + "tokens_seen": 2166751232 + }, + { + "epoch": 0.66, + "learning_rate": 0.0002776649949849549, + "loss": 1.3087, + "theoretical_loss": 3.4362019539592863, + "tokens_seen": 2166816768 + }, + { + "epoch": 0.66, + "learning_rate": 0.0002776489468405216, + "loss": 1.3425, + "theoretical_loss": 3.436193495293651, + "tokens_seen": 2166882304 + }, + { + "epoch": 0.66, + "learning_rate": 0.0002776328986960883, + "loss": 1.196, + "theoretical_loss": 3.4361850369554667, + "tokens_seen": 2166947840 + }, + { + "epoch": 0.66, + "learning_rate": 0.000277616850551655, + "loss": 1.4263, + "theoretical_loss": 3.4361765789447127, + "tokens_seen": 2167013376 + }, + { + "epoch": 0.66, + "learning_rate": 0.0002776008024072217, + "loss": 1.396, + "theoretical_loss": 3.4361681212613657, + "tokens_seen": 2167078912 + }, + { + "epoch": 0.66, + "learning_rate": 0.0002775847542627884, + "loss": 1.322, + "theoretical_loss": 3.4361596639054035, + "tokens_seen": 2167144448 + }, + { + "epoch": 0.66, + "learning_rate": 0.0002775687061183551, + "loss": 1.3809, + "theoretical_loss": 3.4361512068768034, + "tokens_seen": 2167209984 + }, + { + "epoch": 0.66, + "learning_rate": 0.0002775526579739218, + "loss": 1.2597, + "theoretical_loss": 3.4361427501755424, + "tokens_seen": 2167275520 + }, + { + "epoch": 0.66, + "learning_rate": 0.0002775366098294885, + "loss": 1.3124, + "theoretical_loss": 3.436134293801598, + "tokens_seen": 2167341056 + }, + { + "epoch": 0.66, + "learning_rate": 0.0002775205616850552, + "loss": 1.3791, + "theoretical_loss": 3.4361258377549486, + "tokens_seen": 2167406592 + }, + { + "epoch": 0.66, + "learning_rate": 0.0002775045135406219, + "loss": 1.4109, + "theoretical_loss": 3.4361173820355706, + "tokens_seen": 2167472128 + }, + { + "epoch": 0.66, + "learning_rate": 0.0002774884653961886, + "loss": 1.295, + "theoretical_loss": 3.4361089266434415, + "tokens_seen": 2167537664 + }, + { + "epoch": 0.66, + "learning_rate": 0.0002774724172517553, + "loss": 1.3092, + "theoretical_loss": 3.4361004715785395, + "tokens_seen": 2167603200 + }, + { + "epoch": 0.66, + "learning_rate": 0.000277456369107322, + "loss": 1.4354, + "theoretical_loss": 3.436092016840841, + "tokens_seen": 2167668736 + }, + { + "epoch": 0.66, + "learning_rate": 0.0002774403209628887, + "loss": 1.2406, + "theoretical_loss": 3.436083562430324, + "tokens_seen": 2167734272 + }, + { + "epoch": 0.66, + "learning_rate": 0.0002774242728184554, + "loss": 1.2845, + "theoretical_loss": 3.4360751083469663, + "tokens_seen": 2167799808 + }, + { + "epoch": 0.66, + "learning_rate": 0.0002774082246740221, + "loss": 1.3563, + "theoretical_loss": 3.4360666545907446, + "tokens_seen": 2167865344 + }, + { + "epoch": 0.66, + "learning_rate": 0.0002773921765295888, + "loss": 1.3374, + "theoretical_loss": 3.436058201161637, + "tokens_seen": 2167930880 + }, + { + "epoch": 0.66, + "learning_rate": 0.0002773761283851555, + "loss": 1.3976, + "theoretical_loss": 3.4360497480596206, + "tokens_seen": 2167996416 + }, + { + "epoch": 0.66, + "learning_rate": 0.00027736008024072215, + "loss": 1.3675, + "theoretical_loss": 3.4360412952846735, + "tokens_seen": 2168061952 + }, + { + "epoch": 0.66, + "learning_rate": 0.00027734403209628885, + "loss": 1.2867, + "theoretical_loss": 3.436032842836772, + "tokens_seen": 2168127488 + }, + { + "epoch": 0.66, + "learning_rate": 0.00027732798395185555, + "loss": 1.3424, + "theoretical_loss": 3.436024390715894, + "tokens_seen": 2168193024 + }, + { + "epoch": 0.66, + "learning_rate": 0.00027731193580742225, + "loss": 1.4485, + "theoretical_loss": 3.4360159389220177, + "tokens_seen": 2168258560 + }, + { + "epoch": 0.66, + "learning_rate": 0.00027729588766298895, + "loss": 1.3534, + "theoretical_loss": 3.4360074874551194, + "tokens_seen": 2168324096 + }, + { + "epoch": 0.66, + "learning_rate": 0.00027727983951855565, + "loss": 1.3255, + "theoretical_loss": 3.4359990363151773, + "tokens_seen": 2168389632 + }, + { + "epoch": 0.66, + "learning_rate": 0.00027726379137412235, + "loss": 1.3856, + "theoretical_loss": 3.4359905855021693, + "tokens_seen": 2168455168 + }, + { + "epoch": 0.66, + "learning_rate": 0.00027724774322968905, + "loss": 1.2535, + "theoretical_loss": 3.4359821350160717, + "tokens_seen": 2168520704 + }, + { + "epoch": 0.66, + "learning_rate": 0.00027723169508525575, + "loss": 1.3051, + "theoretical_loss": 3.435973684856863, + "tokens_seen": 2168586240 + }, + { + "epoch": 0.66, + "learning_rate": 0.00027721564694082245, + "loss": 1.3091, + "theoretical_loss": 3.4359652350245202, + "tokens_seen": 2168651776 + }, + { + "epoch": 0.66, + "learning_rate": 0.00027719959879638915, + "loss": 1.2748, + "theoretical_loss": 3.4359567855190205, + "tokens_seen": 2168717312 + }, + { + "epoch": 0.66, + "learning_rate": 0.00027718355065195586, + "loss": 1.2445, + "theoretical_loss": 3.435948336340342, + "tokens_seen": 2168782848 + }, + { + "epoch": 0.66, + "learning_rate": 0.00027716750250752256, + "loss": 1.4211, + "theoretical_loss": 3.4359398874884617, + "tokens_seen": 2168848384 + }, + { + "epoch": 0.66, + "learning_rate": 0.00027715145436308926, + "loss": 1.3012, + "theoretical_loss": 3.4359314389633573, + "tokens_seen": 2168913920 + }, + { + "epoch": 0.66, + "learning_rate": 0.00027713540621865596, + "loss": 1.2343, + "theoretical_loss": 3.4359229907650066, + "tokens_seen": 2168979456 + }, + { + "epoch": 0.66, + "learning_rate": 0.00027711935807422266, + "loss": 1.2738, + "theoretical_loss": 3.4359145428933866, + "tokens_seen": 2169044992 + }, + { + "epoch": 0.66, + "learning_rate": 0.00027710330992978936, + "loss": 1.2649, + "theoretical_loss": 3.435906095348475, + "tokens_seen": 2169110528 + }, + { + "epoch": 0.66, + "learning_rate": 0.00027708726178535606, + "loss": 1.3444, + "theoretical_loss": 3.435897648130249, + "tokens_seen": 2169176064 + }, + { + "epoch": 0.66, + "objective/train/docs_used": 790125, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.5030159950256348, + "objective/train/theoretical_loss": 3.4358892012386866, + "objective/train/tokens_used": 2189701600, + "theoretical_loss": 3.4358892012386866, + "tokens_seen": 2169241600 + }, + { + "epoch": 0.66, + "learning_rate": 0.00027707121364092276, + "loss": 1.3831, + "theoretical_loss": 3.4358892012386866, + "tokens_seen": 2169241600 + }, + { + "epoch": 0.66, + "learning_rate": 0.00027705516549648946, + "loss": 1.2636, + "theoretical_loss": 3.435880754673765, + "tokens_seen": 2169307136 + }, + { + "epoch": 0.66, + "learning_rate": 0.00027703911735205616, + "loss": 1.244, + "theoretical_loss": 3.435872308435462, + "tokens_seen": 2169372672 + }, + { + "epoch": 0.66, + "learning_rate": 0.00027702306920762286, + "loss": 1.2775, + "theoretical_loss": 3.435863862523754, + "tokens_seen": 2169438208 + }, + { + "epoch": 0.66, + "learning_rate": 0.00027700702106318956, + "loss": 1.3894, + "theoretical_loss": 3.43585541693862, + "tokens_seen": 2169503744 + }, + { + "epoch": 0.66, + "learning_rate": 0.00027699097291875626, + "loss": 1.3083, + "theoretical_loss": 3.435846971680037, + "tokens_seen": 2169569280 + }, + { + "epoch": 0.66, + "learning_rate": 0.00027697492477432296, + "loss": 1.2568, + "theoretical_loss": 3.435838526747982, + "tokens_seen": 2169634816 + }, + { + "epoch": 0.66, + "learning_rate": 0.00027695887662988966, + "loss": 1.2134, + "theoretical_loss": 3.4358300821424335, + "tokens_seen": 2169700352 + }, + { + "epoch": 0.66, + "learning_rate": 0.00027694282848545637, + "loss": 1.247, + "theoretical_loss": 3.4358216378633677, + "tokens_seen": 2169765888 + }, + { + "epoch": 0.66, + "learning_rate": 0.00027692678034102307, + "loss": 1.3165, + "theoretical_loss": 3.435813193910763, + "tokens_seen": 2169831424 + }, + { + "epoch": 0.66, + "learning_rate": 0.00027691073219658977, + "loss": 1.2746, + "theoretical_loss": 3.435804750284597, + "tokens_seen": 2169896960 + }, + { + "epoch": 0.66, + "learning_rate": 0.00027689468405215647, + "loss": 1.2843, + "theoretical_loss": 3.4357963069848467, + "tokens_seen": 2169962496 + }, + { + "epoch": 0.66, + "learning_rate": 0.00027687863590772317, + "loss": 1.3096, + "theoretical_loss": 3.4357878640114903, + "tokens_seen": 2170028032 + }, + { + "epoch": 0.66, + "learning_rate": 0.0002768625877632899, + "loss": 1.2587, + "theoretical_loss": 3.4357794213645043, + "tokens_seen": 2170093568 + }, + { + "epoch": 0.66, + "learning_rate": 0.0002768465396188566, + "loss": 1.3706, + "theoretical_loss": 3.435770979043867, + "tokens_seen": 2170159104 + }, + { + "epoch": 0.66, + "learning_rate": 0.0002768304914744233, + "loss": 1.2778, + "theoretical_loss": 3.4357625370495564, + "tokens_seen": 2170224640 + }, + { + "epoch": 0.66, + "learning_rate": 0.00027681444332999, + "loss": 1.3564, + "theoretical_loss": 3.4357540953815486, + "tokens_seen": 2170290176 + }, + { + "epoch": 0.66, + "learning_rate": 0.0002767983951855567, + "loss": 1.2501, + "theoretical_loss": 3.4357456540398226, + "tokens_seen": 2170355712 + }, + { + "epoch": 0.66, + "learning_rate": 0.0002767823470411234, + "loss": 1.4157, + "theoretical_loss": 3.435737213024355, + "tokens_seen": 2170421248 + }, + { + "epoch": 0.66, + "learning_rate": 0.0002767662988966901, + "loss": 1.3033, + "theoretical_loss": 3.4357287723351235, + "tokens_seen": 2170486784 + }, + { + "epoch": 0.66, + "learning_rate": 0.0002767502507522568, + "loss": 1.3408, + "theoretical_loss": 3.4357203319721057, + "tokens_seen": 2170552320 + }, + { + "epoch": 0.66, + "learning_rate": 0.00027673420260782353, + "loss": 1.2849, + "theoretical_loss": 3.4357118919352794, + "tokens_seen": 2170617856 + }, + { + "epoch": 0.66, + "learning_rate": 0.0002767181544633902, + "loss": 1.2653, + "theoretical_loss": 3.435703452224622, + "tokens_seen": 2170683392 + }, + { + "epoch": 0.66, + "learning_rate": 0.0002767021063189569, + "loss": 1.3203, + "theoretical_loss": 3.4356950128401107, + "tokens_seen": 2170748928 + }, + { + "epoch": 0.66, + "learning_rate": 0.0002766860581745236, + "loss": 1.2999, + "theoretical_loss": 3.4356865737817235, + "tokens_seen": 2170814464 + }, + { + "epoch": 0.66, + "learning_rate": 0.0002766700100300903, + "loss": 1.3015, + "theoretical_loss": 3.435678135049438, + "tokens_seen": 2170880000 + }, + { + "epoch": 0.66, + "learning_rate": 0.000276653961885657, + "loss": 1.3279, + "theoretical_loss": 3.4356696966432314, + "tokens_seen": 2170945536 + }, + { + "epoch": 0.66, + "learning_rate": 0.0002766379137412237, + "loss": 1.3024, + "theoretical_loss": 3.4356612585630812, + "tokens_seen": 2171011072 + }, + { + "epoch": 0.66, + "learning_rate": 0.0002766218655967904, + "loss": 1.4402, + "theoretical_loss": 3.4356528208089654, + "tokens_seen": 2171076608 + }, + { + "epoch": 0.66, + "learning_rate": 0.0002766058174523571, + "loss": 1.4024, + "theoretical_loss": 3.435644383380861, + "tokens_seen": 2171142144 + }, + { + "epoch": 0.66, + "learning_rate": 0.0002765897693079238, + "loss": 1.4219, + "theoretical_loss": 3.4356359462787465, + "tokens_seen": 2171207680 + }, + { + "epoch": 0.66, + "learning_rate": 0.0002765737211634905, + "loss": 1.335, + "theoretical_loss": 3.4356275095025985, + "tokens_seen": 2171273216 + }, + { + "epoch": 0.66, + "learning_rate": 0.0002765576730190572, + "loss": 1.3099, + "theoretical_loss": 3.4356190730523952, + "tokens_seen": 2171338752 + }, + { + "epoch": 0.66, + "learning_rate": 0.0002765416248746239, + "loss": 1.3647, + "theoretical_loss": 3.4356106369281134, + "tokens_seen": 2171404288 + }, + { + "epoch": 0.66, + "learning_rate": 0.0002765255767301906, + "loss": 1.4067, + "theoretical_loss": 3.435602201129732, + "tokens_seen": 2171469824 + }, + { + "epoch": 0.66, + "learning_rate": 0.0002765095285857573, + "loss": 1.2827, + "theoretical_loss": 3.4355937656572264, + "tokens_seen": 2171535360 + }, + { + "epoch": 0.66, + "learning_rate": 0.000276493480441324, + "loss": 1.2828, + "theoretical_loss": 3.435585330510577, + "tokens_seen": 2171600896 + }, + { + "epoch": 0.66, + "learning_rate": 0.0002764774322968907, + "loss": 1.3045, + "theoretical_loss": 3.435576895689759, + "tokens_seen": 2171666432 + }, + { + "epoch": 0.66, + "learning_rate": 0.0002764613841524574, + "loss": 1.3032, + "theoretical_loss": 3.435568461194751, + "tokens_seen": 2171731968 + }, + { + "epoch": 0.66, + "learning_rate": 0.0002764453360080241, + "loss": 1.1801, + "theoretical_loss": 3.4355600270255304, + "tokens_seen": 2171797504 + }, + { + "epoch": 0.66, + "learning_rate": 0.0002764292878635908, + "loss": 1.4988, + "theoretical_loss": 3.4355515931820753, + "tokens_seen": 2171863040 + }, + { + "epoch": 0.66, + "learning_rate": 0.0002764132397191575, + "loss": 1.3468, + "theoretical_loss": 3.4355431596643626, + "tokens_seen": 2171928576 + }, + { + "epoch": 0.66, + "learning_rate": 0.0002763971915747242, + "loss": 1.2743, + "theoretical_loss": 3.43553472647237, + "tokens_seen": 2171994112 + }, + { + "epoch": 0.66, + "learning_rate": 0.0002763811434302909, + "loss": 1.2453, + "theoretical_loss": 3.4355262936060753, + "tokens_seen": 2172059648 + }, + { + "epoch": 0.66, + "learning_rate": 0.0002763650952858576, + "loss": 1.4033, + "theoretical_loss": 3.435517861065456, + "tokens_seen": 2172125184 + }, + { + "epoch": 0.66, + "learning_rate": 0.0002763490471414243, + "loss": 1.2909, + "theoretical_loss": 3.43550942885049, + "tokens_seen": 2172190720 + }, + { + "epoch": 0.66, + "learning_rate": 0.000276332998996991, + "loss": 1.4402, + "theoretical_loss": 3.435500996961154, + "tokens_seen": 2172256256 + }, + { + "epoch": 0.66, + "learning_rate": 0.0002763169508525577, + "loss": 1.3114, + "theoretical_loss": 3.4354925653974266, + "tokens_seen": 2172321792 + }, + { + "epoch": 0.66, + "learning_rate": 0.0002763009027081244, + "loss": 1.362, + "theoretical_loss": 3.435484134159285, + "tokens_seen": 2172387328 + }, + { + "epoch": 0.66, + "learning_rate": 0.0002762848545636911, + "loss": 1.373, + "theoretical_loss": 3.435475703246707, + "tokens_seen": 2172452864 + }, + { + "epoch": 0.66, + "objective/train/docs_used": 791273, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.294751763343811, + "objective/train/theoretical_loss": 3.4354672726596696, + "objective/train/tokens_used": 2192978400, + "theoretical_loss": 3.4354672726596696, + "tokens_seen": 2172518400 + }, + { + "epoch": 0.66, + "learning_rate": 0.0002762688064192578, + "loss": 1.2619, + "theoretical_loss": 3.4354672726596696, + "tokens_seen": 2172518400 + }, + { + "epoch": 0.66, + "learning_rate": 0.0002762527582748245, + "loss": 1.3619, + "theoretical_loss": 3.4354588423981514, + "tokens_seen": 2172583936 + }, + { + "epoch": 0.66, + "learning_rate": 0.0002762367101303912, + "loss": 1.1559, + "theoretical_loss": 3.4354504124621292, + "tokens_seen": 2172649472 + }, + { + "epoch": 0.66, + "learning_rate": 0.0002762206619859579, + "loss": 1.411, + "theoretical_loss": 3.4354419828515805, + "tokens_seen": 2172715008 + }, + { + "epoch": 0.66, + "learning_rate": 0.0002762046138415246, + "loss": 1.2794, + "theoretical_loss": 3.435433553566484, + "tokens_seen": 2172780544 + }, + { + "epoch": 0.66, + "learning_rate": 0.0002761885656970913, + "loss": 1.2863, + "theoretical_loss": 3.435425124606816, + "tokens_seen": 2172846080 + }, + { + "epoch": 0.66, + "learning_rate": 0.000276172517552658, + "loss": 1.3347, + "theoretical_loss": 3.435416695972555, + "tokens_seen": 2172911616 + }, + { + "epoch": 0.66, + "learning_rate": 0.0002761564694082247, + "loss": 1.2981, + "theoretical_loss": 3.4354082676636786, + "tokens_seen": 2172977152 + }, + { + "epoch": 0.66, + "learning_rate": 0.0002761404212637914, + "loss": 1.3838, + "theoretical_loss": 3.4353998396801635, + "tokens_seen": 2173042688 + }, + { + "epoch": 0.66, + "learning_rate": 0.0002761243731193581, + "loss": 1.3908, + "theoretical_loss": 3.4353914120219886, + "tokens_seen": 2173108224 + }, + { + "epoch": 0.66, + "learning_rate": 0.0002761083249749248, + "loss": 1.3314, + "theoretical_loss": 3.435382984689131, + "tokens_seen": 2173173760 + }, + { + "epoch": 0.66, + "learning_rate": 0.0002760922768304915, + "loss": 1.4129, + "theoretical_loss": 3.4353745576815675, + "tokens_seen": 2173239296 + }, + { + "epoch": 0.66, + "learning_rate": 0.0002760762286860582, + "loss": 1.3568, + "theoretical_loss": 3.4353661309992773, + "tokens_seen": 2173304832 + }, + { + "epoch": 0.66, + "learning_rate": 0.0002760601805416249, + "loss": 1.3684, + "theoretical_loss": 3.4353577046422368, + "tokens_seen": 2173370368 + }, + { + "epoch": 0.66, + "learning_rate": 0.0002760441323971916, + "loss": 1.2933, + "theoretical_loss": 3.435349278610424, + "tokens_seen": 2173435904 + }, + { + "epoch": 0.66, + "learning_rate": 0.0002760280842527583, + "loss": 1.3593, + "theoretical_loss": 3.435340852903817, + "tokens_seen": 2173501440 + }, + { + "epoch": 0.66, + "learning_rate": 0.000276012036108325, + "loss": 1.3554, + "theoretical_loss": 3.4353324275223933, + "tokens_seen": 2173566976 + }, + { + "epoch": 0.66, + "learning_rate": 0.0002759959879638917, + "loss": 1.3101, + "theoretical_loss": 3.43532400246613, + "tokens_seen": 2173632512 + }, + { + "epoch": 0.66, + "learning_rate": 0.0002759799398194584, + "loss": 1.3026, + "theoretical_loss": 3.435315577735005, + "tokens_seen": 2173698048 + }, + { + "epoch": 0.66, + "learning_rate": 0.0002759638916750251, + "loss": 1.3136, + "theoretical_loss": 3.4353071533289956, + "tokens_seen": 2173763584 + }, + { + "epoch": 0.66, + "learning_rate": 0.00027594784353059175, + "loss": 1.3198, + "theoretical_loss": 3.43529872924808, + "tokens_seen": 2173829120 + }, + { + "epoch": 0.66, + "learning_rate": 0.00027593179538615845, + "loss": 1.3217, + "theoretical_loss": 3.435290305492236, + "tokens_seen": 2173894656 + }, + { + "epoch": 0.66, + "learning_rate": 0.00027591574724172515, + "loss": 1.2575, + "theoretical_loss": 3.4352818820614415, + "tokens_seen": 2173960192 + }, + { + "epoch": 0.66, + "learning_rate": 0.00027589969909729185, + "loss": 1.2331, + "theoretical_loss": 3.435273458955673, + "tokens_seen": 2174025728 + }, + { + "epoch": 0.66, + "learning_rate": 0.00027588365095285855, + "loss": 1.1826, + "theoretical_loss": 3.4352650361749086, + "tokens_seen": 2174091264 + }, + { + "epoch": 0.66, + "learning_rate": 0.00027586760280842525, + "loss": 1.3174, + "theoretical_loss": 3.4352566137191265, + "tokens_seen": 2174156800 + }, + { + "epoch": 0.66, + "learning_rate": 0.00027585155466399195, + "loss": 1.3167, + "theoretical_loss": 3.435248191588304, + "tokens_seen": 2174222336 + }, + { + "epoch": 0.66, + "learning_rate": 0.00027583550651955865, + "loss": 1.3511, + "theoretical_loss": 3.435239769782419, + "tokens_seen": 2174287872 + }, + { + "epoch": 0.66, + "learning_rate": 0.00027581945837512536, + "loss": 1.2212, + "theoretical_loss": 3.4352313483014485, + "tokens_seen": 2174353408 + }, + { + "epoch": 0.66, + "learning_rate": 0.00027580341023069206, + "loss": 1.2883, + "theoretical_loss": 3.4352229271453707, + "tokens_seen": 2174418944 + }, + { + "epoch": 0.66, + "learning_rate": 0.00027578736208625876, + "loss": 1.2147, + "theoretical_loss": 3.4352145063141633, + "tokens_seen": 2174484480 + }, + { + "epoch": 0.66, + "learning_rate": 0.00027577131394182546, + "loss": 1.2893, + "theoretical_loss": 3.435206085807804, + "tokens_seen": 2174550016 + }, + { + "epoch": 0.66, + "learning_rate": 0.00027575526579739216, + "loss": 1.3027, + "theoretical_loss": 3.43519766562627, + "tokens_seen": 2174615552 + }, + { + "epoch": 0.66, + "learning_rate": 0.00027573921765295886, + "loss": 1.2655, + "theoretical_loss": 3.43518924576954, + "tokens_seen": 2174681088 + }, + { + "epoch": 0.66, + "learning_rate": 0.00027572316950852556, + "loss": 1.2154, + "theoretical_loss": 3.4351808262375907, + "tokens_seen": 2174746624 + }, + { + "epoch": 0.66, + "learning_rate": 0.00027570712136409226, + "loss": 1.3167, + "theoretical_loss": 3.4351724070304, + "tokens_seen": 2174812160 + }, + { + "epoch": 0.66, + "learning_rate": 0.00027569107321965896, + "loss": 1.3389, + "theoretical_loss": 3.435163988147946, + "tokens_seen": 2174877696 + }, + { + "epoch": 0.66, + "learning_rate": 0.0002756750250752257, + "loss": 1.3131, + "theoretical_loss": 3.4351555695902056, + "tokens_seen": 2174943232 + }, + { + "epoch": 0.66, + "learning_rate": 0.0002756589769307924, + "loss": 1.3388, + "theoretical_loss": 3.435147151357157, + "tokens_seen": 2175008768 + }, + { + "epoch": 0.66, + "learning_rate": 0.0002756429287863591, + "loss": 1.2954, + "theoretical_loss": 3.4351387334487784, + "tokens_seen": 2175074304 + }, + { + "epoch": 0.66, + "learning_rate": 0.0002756268806419258, + "loss": 1.3364, + "theoretical_loss": 3.4351303158650466, + "tokens_seen": 2175139840 + }, + { + "epoch": 0.66, + "learning_rate": 0.0002756108324974925, + "loss": 1.2696, + "theoretical_loss": 3.4351218986059395, + "tokens_seen": 2175205376 + }, + { + "epoch": 0.66, + "learning_rate": 0.0002755947843530592, + "loss": 1.3747, + "theoretical_loss": 3.4351134816714355, + "tokens_seen": 2175270912 + }, + { + "epoch": 0.66, + "learning_rate": 0.0002755787362086259, + "loss": 1.3394, + "theoretical_loss": 3.4351050650615114, + "tokens_seen": 2175336448 + }, + { + "epoch": 0.66, + "learning_rate": 0.0002755626880641926, + "loss": 1.2746, + "theoretical_loss": 3.4350966487761454, + "tokens_seen": 2175401984 + }, + { + "epoch": 0.66, + "learning_rate": 0.0002755466399197593, + "loss": 1.2054, + "theoretical_loss": 3.435088232815315, + "tokens_seen": 2175467520 + }, + { + "epoch": 0.66, + "learning_rate": 0.000275530591775326, + "loss": 1.2175, + "theoretical_loss": 3.4350798171789982, + "tokens_seen": 2175533056 + }, + { + "epoch": 0.66, + "learning_rate": 0.0002755145436308927, + "loss": 1.2674, + "theoretical_loss": 3.435071401867172, + "tokens_seen": 2175598592 + }, + { + "epoch": 0.66, + "learning_rate": 0.0002754984954864594, + "loss": 1.314, + "theoretical_loss": 3.435062986879815, + "tokens_seen": 2175664128 + }, + { + "epoch": 0.66, + "learning_rate": 0.0002754824473420261, + "loss": 1.2199, + "theoretical_loss": 3.4350545722169046, + "tokens_seen": 2175729664 + }, + { + "epoch": 0.66, + "objective/train/docs_used": 792571, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.2568638324737549, + "objective/train/theoretical_loss": 3.4350461578784186, + "objective/train/tokens_used": 2196255200, + "theoretical_loss": 3.4350461578784186, + "tokens_seen": 2175795200 + }, + { + "epoch": 0.66, + "learning_rate": 0.0002754663991975928, + "loss": 1.2934, + "theoretical_loss": 3.4350461578784186, + "tokens_seen": 2175795200 + }, + { + "epoch": 0.66, + "learning_rate": 0.0002754503510531595, + "loss": 1.331, + "theoretical_loss": 3.435037743864334, + "tokens_seen": 2175860736 + }, + { + "epoch": 0.66, + "learning_rate": 0.0002754343029087262, + "loss": 1.2124, + "theoretical_loss": 3.4350293301746295, + "tokens_seen": 2175926272 + }, + { + "epoch": 0.66, + "learning_rate": 0.0002754182547642929, + "loss": 1.3482, + "theoretical_loss": 3.4350209168092825, + "tokens_seen": 2175991808 + }, + { + "epoch": 0.66, + "learning_rate": 0.0002754022066198596, + "loss": 1.4226, + "theoretical_loss": 3.43501250376827, + "tokens_seen": 2176057344 + }, + { + "epoch": 0.66, + "learning_rate": 0.0002753861584754263, + "loss": 1.3931, + "theoretical_loss": 3.4350040910515713, + "tokens_seen": 2176122880 + }, + { + "epoch": 0.66, + "learning_rate": 0.00027537011033099303, + "loss": 1.2629, + "theoretical_loss": 3.434995678659163, + "tokens_seen": 2176188416 + }, + { + "epoch": 0.66, + "learning_rate": 0.00027535406218655973, + "loss": 1.2586, + "theoretical_loss": 3.4349872665910226, + "tokens_seen": 2176253952 + }, + { + "epoch": 0.66, + "learning_rate": 0.00027533801404212643, + "loss": 1.198, + "theoretical_loss": 3.4349788548471287, + "tokens_seen": 2176319488 + }, + { + "epoch": 0.66, + "learning_rate": 0.0002753219658976931, + "loss": 1.3056, + "theoretical_loss": 3.4349704434274586, + "tokens_seen": 2176385024 + }, + { + "epoch": 0.66, + "learning_rate": 0.0002753059177532598, + "loss": 1.2266, + "theoretical_loss": 3.43496203233199, + "tokens_seen": 2176450560 + }, + { + "epoch": 0.66, + "learning_rate": 0.0002752898696088265, + "loss": 1.3153, + "theoretical_loss": 3.4349536215607004, + "tokens_seen": 2176516096 + }, + { + "epoch": 0.66, + "learning_rate": 0.0002752738214643932, + "loss": 1.2784, + "theoretical_loss": 3.4349452111135683, + "tokens_seen": 2176581632 + }, + { + "epoch": 0.66, + "learning_rate": 0.0002752577733199599, + "loss": 1.2884, + "theoretical_loss": 3.4349368009905707, + "tokens_seen": 2176647168 + }, + { + "epoch": 0.66, + "learning_rate": 0.0002752417251755266, + "loss": 1.2864, + "theoretical_loss": 3.4349283911916855, + "tokens_seen": 2176712704 + }, + { + "epoch": 0.66, + "learning_rate": 0.0002752256770310933, + "loss": 1.457, + "theoretical_loss": 3.4349199817168907, + "tokens_seen": 2176778240 + }, + { + "epoch": 0.66, + "learning_rate": 0.00027520962888666, + "loss": 1.2204, + "theoretical_loss": 3.4349115725661643, + "tokens_seen": 2176843776 + }, + { + "epoch": 0.66, + "learning_rate": 0.0002751935807422267, + "loss": 1.2467, + "theoretical_loss": 3.4349031637394836, + "tokens_seen": 2176909312 + }, + { + "epoch": 0.66, + "learning_rate": 0.0002751775325977934, + "loss": 1.3265, + "theoretical_loss": 3.4348947552368263, + "tokens_seen": 2176974848 + }, + { + "epoch": 0.66, + "learning_rate": 0.0002751614844533601, + "loss": 1.3786, + "theoretical_loss": 3.4348863470581703, + "tokens_seen": 2177040384 + }, + { + "epoch": 0.66, + "learning_rate": 0.0002751454363089268, + "loss": 1.3373, + "theoretical_loss": 3.434877939203494, + "tokens_seen": 2177105920 + }, + { + "epoch": 0.66, + "learning_rate": 0.0002751293881644935, + "loss": 1.3255, + "theoretical_loss": 3.4348695316727733, + "tokens_seen": 2177171456 + }, + { + "epoch": 0.66, + "learning_rate": 0.0002751133400200602, + "loss": 1.3489, + "theoretical_loss": 3.4348611244659883, + "tokens_seen": 2177236992 + }, + { + "epoch": 0.66, + "learning_rate": 0.0002750972918756269, + "loss": 1.2239, + "theoretical_loss": 3.4348527175831154, + "tokens_seen": 2177302528 + }, + { + "epoch": 0.66, + "learning_rate": 0.0002750812437311936, + "loss": 1.2036, + "theoretical_loss": 3.4348443110241327, + "tokens_seen": 2177368064 + }, + { + "epoch": 0.66, + "learning_rate": 0.0002750651955867603, + "loss": 1.3435, + "theoretical_loss": 3.4348359047890176, + "tokens_seen": 2177433600 + }, + { + "epoch": 0.66, + "learning_rate": 0.000275049147442327, + "loss": 1.2342, + "theoretical_loss": 3.4348274988777483, + "tokens_seen": 2177499136 + }, + { + "epoch": 0.66, + "learning_rate": 0.0002750330992978937, + "loss": 1.4016, + "theoretical_loss": 3.4348190932903027, + "tokens_seen": 2177564672 + }, + { + "epoch": 0.66, + "learning_rate": 0.0002750170511534604, + "loss": 1.3018, + "theoretical_loss": 3.4348106880266585, + "tokens_seen": 2177630208 + }, + { + "epoch": 0.66, + "learning_rate": 0.0002750010030090271, + "loss": 1.3837, + "theoretical_loss": 3.434802283086793, + "tokens_seen": 2177695744 + }, + { + "epoch": 0.66, + "learning_rate": 0.0002749849548645938, + "loss": 1.2307, + "theoretical_loss": 3.4347938784706846, + "tokens_seen": 2177761280 + }, + { + "epoch": 0.66, + "learning_rate": 0.0002749689067201605, + "loss": 1.2565, + "theoretical_loss": 3.4347854741783106, + "tokens_seen": 2177826816 + }, + { + "epoch": 0.66, + "learning_rate": 0.0002749528585757272, + "loss": 1.3372, + "theoretical_loss": 3.434777070209649, + "tokens_seen": 2177892352 + }, + { + "epoch": 0.66, + "learning_rate": 0.0002749368104312939, + "loss": 1.3938, + "theoretical_loss": 3.4347686665646777, + "tokens_seen": 2177957888 + }, + { + "epoch": 0.66, + "learning_rate": 0.0002749207622868606, + "loss": 1.3787, + "theoretical_loss": 3.4347602632433745, + "tokens_seen": 2178023424 + }, + { + "epoch": 0.66, + "learning_rate": 0.0002749047141424273, + "loss": 1.3747, + "theoretical_loss": 3.4347518602457168, + "tokens_seen": 2178088960 + }, + { + "epoch": 0.66, + "learning_rate": 0.000274888665997994, + "loss": 1.1818, + "theoretical_loss": 3.434743457571683, + "tokens_seen": 2178154496 + }, + { + "epoch": 0.66, + "learning_rate": 0.0002748726178535607, + "loss": 1.2495, + "theoretical_loss": 3.4347350552212506, + "tokens_seen": 2178220032 + }, + { + "epoch": 0.66, + "learning_rate": 0.0002748565697091274, + "loss": 1.3241, + "theoretical_loss": 3.434726653194397, + "tokens_seen": 2178285568 + }, + { + "epoch": 0.66, + "learning_rate": 0.0002748405215646941, + "loss": 1.4895, + "theoretical_loss": 3.4347182514911005, + "tokens_seen": 2178351104 + }, + { + "epoch": 0.66, + "learning_rate": 0.0002748244734202608, + "loss": 1.2615, + "theoretical_loss": 3.434709850111339, + "tokens_seen": 2178416640 + }, + { + "epoch": 0.66, + "learning_rate": 0.0002748084252758275, + "loss": 1.2824, + "theoretical_loss": 3.43470144905509, + "tokens_seen": 2178482176 + }, + { + "epoch": 0.66, + "learning_rate": 0.0002747923771313942, + "loss": 1.3116, + "theoretical_loss": 3.4346930483223312, + "tokens_seen": 2178547712 + }, + { + "epoch": 0.66, + "learning_rate": 0.0002747763289869609, + "loss": 1.3603, + "theoretical_loss": 3.4346846479130413, + "tokens_seen": 2178613248 + }, + { + "epoch": 0.66, + "learning_rate": 0.0002747602808425276, + "loss": 1.3432, + "theoretical_loss": 3.434676247827197, + "tokens_seen": 2178678784 + }, + { + "epoch": 0.66, + "learning_rate": 0.0002747442326980943, + "loss": 1.3436, + "theoretical_loss": 3.4346678480647768, + "tokens_seen": 2178744320 + }, + { + "epoch": 0.66, + "learning_rate": 0.000274728184553661, + "loss": 1.21, + "theoretical_loss": 3.434659448625758, + "tokens_seen": 2178809856 + }, + { + "epoch": 0.66, + "learning_rate": 0.0002747121364092277, + "loss": 1.322, + "theoretical_loss": 3.4346510495101183, + "tokens_seen": 2178875392 + }, + { + "epoch": 0.66, + "learning_rate": 0.0002746960882647944, + "loss": 1.1973, + "theoretical_loss": 3.4346426507178367, + "tokens_seen": 2178940928 + }, + { + "epoch": 0.66, + "learning_rate": 0.0002746800401203611, + "loss": 1.4195, + "theoretical_loss": 3.4346342522488897, + "tokens_seen": 2179006464 + }, + { + "epoch": 0.66, + "objective/train/docs_used": 793725, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.1476774215698242, + "objective/train/theoretical_loss": 3.434625854103256, + "objective/train/tokens_used": 2199532000, + "theoretical_loss": 3.434625854103256, + "tokens_seen": 2179072000 + }, + { + "epoch": 0.66, + "learning_rate": 0.0002746639919759278, + "loss": 1.2283, + "theoretical_loss": 3.434625854103256, + "tokens_seen": 2179072000 + }, + { + "epoch": 0.66, + "learning_rate": 0.0002746479438314945, + "loss": 1.28, + "theoretical_loss": 3.4346174562809133, + "tokens_seen": 2179137536 + }, + { + "epoch": 0.66, + "learning_rate": 0.0002746318956870612, + "loss": 1.318, + "theoretical_loss": 3.434609058781839, + "tokens_seen": 2179203072 + }, + { + "epoch": 0.66, + "learning_rate": 0.0002746158475426279, + "loss": 1.2623, + "theoretical_loss": 3.434600661606011, + "tokens_seen": 2179268608 + }, + { + "epoch": 0.66, + "learning_rate": 0.0002745997993981946, + "loss": 1.279, + "theoretical_loss": 3.4345922647534075, + "tokens_seen": 2179334144 + }, + { + "epoch": 0.66, + "learning_rate": 0.0002745837512537613, + "loss": 1.238, + "theoretical_loss": 3.4345838682240064, + "tokens_seen": 2179399680 + }, + { + "epoch": 0.66, + "learning_rate": 0.000274567703109328, + "loss": 1.297, + "theoretical_loss": 3.434575472017785, + "tokens_seen": 2179465216 + }, + { + "epoch": 0.66, + "learning_rate": 0.00027455165496489465, + "loss": 1.2583, + "theoretical_loss": 3.4345670761347216, + "tokens_seen": 2179530752 + }, + { + "epoch": 0.66, + "learning_rate": 0.00027453560682046135, + "loss": 1.3696, + "theoretical_loss": 3.434558680574794, + "tokens_seen": 2179596288 + }, + { + "epoch": 0.66, + "learning_rate": 0.00027451955867602805, + "loss": 1.2254, + "theoretical_loss": 3.4345502853379792, + "tokens_seen": 2179661824 + }, + { + "epoch": 0.66, + "learning_rate": 0.00027450351053159475, + "loss": 1.2445, + "theoretical_loss": 3.4345418904242564, + "tokens_seen": 2179727360 + }, + { + "epoch": 0.66, + "learning_rate": 0.00027448746238716145, + "loss": 1.3855, + "theoretical_loss": 3.4345334958336027, + "tokens_seen": 2179792896 + }, + { + "epoch": 0.66, + "learning_rate": 0.0002744714142427282, + "loss": 1.2381, + "theoretical_loss": 3.4345251015659963, + "tokens_seen": 2179858432 + }, + { + "epoch": 0.66, + "learning_rate": 0.0002744553660982949, + "loss": 1.3337, + "theoretical_loss": 3.4345167076214143, + "tokens_seen": 2179923968 + }, + { + "epoch": 0.66, + "learning_rate": 0.0002744393179538616, + "loss": 1.3061, + "theoretical_loss": 3.4345083139998356, + "tokens_seen": 2179989504 + }, + { + "epoch": 0.66, + "learning_rate": 0.0002744232698094283, + "loss": 1.2854, + "theoretical_loss": 3.4344999207012377, + "tokens_seen": 2180055040 + }, + { + "epoch": 0.66, + "learning_rate": 0.000274407221664995, + "loss": 1.31, + "theoretical_loss": 3.434491527725598, + "tokens_seen": 2180120576 + }, + { + "epoch": 0.66, + "learning_rate": 0.0002743911735205617, + "loss": 1.3424, + "theoretical_loss": 3.4344831350728944, + "tokens_seen": 2180186112 + }, + { + "epoch": 0.66, + "learning_rate": 0.0002743751253761284, + "loss": 1.2554, + "theoretical_loss": 3.434474742743106, + "tokens_seen": 2180251648 + }, + { + "epoch": 0.66, + "learning_rate": 0.0002743590772316951, + "loss": 1.266, + "theoretical_loss": 3.4344663507362085, + "tokens_seen": 2180317184 + }, + { + "epoch": 0.66, + "learning_rate": 0.0002743430290872618, + "loss": 1.3908, + "theoretical_loss": 3.4344579590521818, + "tokens_seen": 2180382720 + }, + { + "epoch": 0.66, + "learning_rate": 0.0002743269809428285, + "loss": 1.3577, + "theoretical_loss": 3.434449567691003, + "tokens_seen": 2180448256 + }, + { + "epoch": 0.66, + "learning_rate": 0.0002743109327983952, + "loss": 1.3081, + "theoretical_loss": 3.43444117665265, + "tokens_seen": 2180513792 + }, + { + "epoch": 0.66, + "learning_rate": 0.0002742948846539619, + "loss": 1.2518, + "theoretical_loss": 3.4344327859371004, + "tokens_seen": 2180579328 + }, + { + "epoch": 0.66, + "learning_rate": 0.0002742788365095286, + "loss": 1.34, + "theoretical_loss": 3.4344243955443323, + "tokens_seen": 2180644864 + }, + { + "epoch": 0.66, + "learning_rate": 0.0002742627883650953, + "loss": 1.2926, + "theoretical_loss": 3.4344160054743234, + "tokens_seen": 2180710400 + }, + { + "epoch": 0.66, + "learning_rate": 0.000274246740220662, + "loss": 1.1593, + "theoretical_loss": 3.434407615727052, + "tokens_seen": 2180775936 + }, + { + "epoch": 0.66, + "learning_rate": 0.0002742306920762287, + "loss": 1.2163, + "theoretical_loss": 3.434399226302496, + "tokens_seen": 2180841472 + }, + { + "epoch": 0.66, + "learning_rate": 0.0002742146439317954, + "loss": 1.3513, + "theoretical_loss": 3.434390837200633, + "tokens_seen": 2180907008 + }, + { + "epoch": 0.66, + "learning_rate": 0.0002741985957873621, + "loss": 1.3444, + "theoretical_loss": 3.4343824484214407, + "tokens_seen": 2180972544 + }, + { + "epoch": 0.66, + "learning_rate": 0.0002741825476429288, + "loss": 1.3648, + "theoretical_loss": 3.434374059964897, + "tokens_seen": 2181038080 + }, + { + "epoch": 0.66, + "learning_rate": 0.0002741664994984955, + "loss": 1.2746, + "theoretical_loss": 3.4343656718309807, + "tokens_seen": 2181103616 + }, + { + "epoch": 0.66, + "learning_rate": 0.0002741504513540622, + "loss": 1.2381, + "theoretical_loss": 3.4343572840196686, + "tokens_seen": 2181169152 + }, + { + "epoch": 0.66, + "learning_rate": 0.0002741344032096289, + "loss": 1.385, + "theoretical_loss": 3.434348896530939, + "tokens_seen": 2181234688 + }, + { + "epoch": 0.66, + "learning_rate": 0.0002741183550651956, + "loss": 1.3641, + "theoretical_loss": 3.4343405093647705, + "tokens_seen": 2181300224 + }, + { + "epoch": 0.66, + "learning_rate": 0.0002741023069207623, + "loss": 1.3955, + "theoretical_loss": 3.4343321225211394, + "tokens_seen": 2181365760 + }, + { + "epoch": 0.66, + "learning_rate": 0.000274086258776329, + "loss": 1.2933, + "theoretical_loss": 3.4343237360000254, + "tokens_seen": 2181431296 + }, + { + "epoch": 0.66, + "learning_rate": 0.0002740702106318957, + "loss": 1.4618, + "theoretical_loss": 3.434315349801405, + "tokens_seen": 2181496832 + }, + { + "epoch": 0.66, + "learning_rate": 0.0002740541624874624, + "loss": 1.3292, + "theoretical_loss": 3.434306963925257, + "tokens_seen": 2181562368 + }, + { + "epoch": 0.66, + "learning_rate": 0.0002740381143430291, + "loss": 1.3957, + "theoretical_loss": 3.4342985783715587, + "tokens_seen": 2181627904 + }, + { + "epoch": 0.66, + "learning_rate": 0.0002740220661985958, + "loss": 1.3169, + "theoretical_loss": 3.4342901931402885, + "tokens_seen": 2181693440 + }, + { + "epoch": 0.66, + "learning_rate": 0.0002740060180541625, + "loss": 1.321, + "theoretical_loss": 3.434281808231424, + "tokens_seen": 2181758976 + }, + { + "epoch": 0.66, + "learning_rate": 0.00027398996990972923, + "loss": 1.3079, + "theoretical_loss": 3.4342734236449433, + "tokens_seen": 2181824512 + }, + { + "epoch": 0.66, + "learning_rate": 0.00027397392176529593, + "loss": 1.359, + "theoretical_loss": 3.434265039380824, + "tokens_seen": 2181890048 + }, + { + "epoch": 0.66, + "learning_rate": 0.00027395787362086263, + "loss": 1.3293, + "theoretical_loss": 3.434256655439045, + "tokens_seen": 2181955584 + }, + { + "epoch": 0.66, + "learning_rate": 0.00027394182547642933, + "loss": 1.2998, + "theoretical_loss": 3.4342482718195826, + "tokens_seen": 2182021120 + }, + { + "epoch": 0.66, + "learning_rate": 0.000273925777331996, + "loss": 1.279, + "theoretical_loss": 3.4342398885224164, + "tokens_seen": 2182086656 + }, + { + "epoch": 0.66, + "learning_rate": 0.0002739097291875627, + "loss": 1.4941, + "theoretical_loss": 3.434231505547523, + "tokens_seen": 2182152192 + }, + { + "epoch": 0.66, + "learning_rate": 0.0002738936810431294, + "loss": 1.2785, + "theoretical_loss": 3.434223122894881, + "tokens_seen": 2182217728 + }, + { + "epoch": 0.66, + "learning_rate": 0.0002738776328986961, + "loss": 1.2864, + "theoretical_loss": 3.4342147405644683, + "tokens_seen": 2182283264 + }, + { + "epoch": 0.66, + "objective/train/docs_used": 794520, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.4507591724395752, + "objective/train/theoretical_loss": 3.434206358556263, + "objective/train/tokens_used": 2202808800, + "theoretical_loss": 3.434206358556263, + "tokens_seen": 2182348800 + }, + { + "epoch": 0.66, + "learning_rate": 0.0002738615847542628, + "loss": 1.4103, + "theoretical_loss": 3.434206358556263, + "tokens_seen": 2182348800 + }, + { + "epoch": 0.66, + "learning_rate": 0.0002738455366098295, + "loss": 1.358, + "theoretical_loss": 3.4341979768702426, + "tokens_seen": 2182414336 + }, + { + "epoch": 0.66, + "learning_rate": 0.0002738294884653962, + "loss": 1.4419, + "theoretical_loss": 3.4341895955063855, + "tokens_seen": 2182479872 + }, + { + "epoch": 0.66, + "learning_rate": 0.0002738134403209629, + "loss": 1.4434, + "theoretical_loss": 3.434181214464669, + "tokens_seen": 2182545408 + }, + { + "epoch": 0.66, + "learning_rate": 0.0002737973921765296, + "loss": 1.276, + "theoretical_loss": 3.434172833745072, + "tokens_seen": 2182610944 + }, + { + "epoch": 0.66, + "learning_rate": 0.0002737813440320963, + "loss": 1.2722, + "theoretical_loss": 3.4341644533475715, + "tokens_seen": 2182676480 + }, + { + "epoch": 0.66, + "learning_rate": 0.000273765295887663, + "loss": 1.3062, + "theoretical_loss": 3.434156073272146, + "tokens_seen": 2182742016 + }, + { + "epoch": 0.66, + "learning_rate": 0.0002737492477432297, + "loss": 1.3139, + "theoretical_loss": 3.434147693518773, + "tokens_seen": 2182807552 + }, + { + "epoch": 0.66, + "learning_rate": 0.0002737331995987964, + "loss": 1.2627, + "theoretical_loss": 3.434139314087431, + "tokens_seen": 2182873088 + }, + { + "epoch": 0.66, + "learning_rate": 0.0002737171514543631, + "loss": 1.3008, + "theoretical_loss": 3.4341309349780977, + "tokens_seen": 2182938624 + }, + { + "epoch": 0.66, + "learning_rate": 0.0002737011033099298, + "loss": 1.229, + "theoretical_loss": 3.434122556190751, + "tokens_seen": 2183004160 + }, + { + "epoch": 0.66, + "learning_rate": 0.0002736850551654965, + "loss": 1.2786, + "theoretical_loss": 3.4341141777253688, + "tokens_seen": 2183069696 + }, + { + "epoch": 0.66, + "learning_rate": 0.0002736690070210632, + "loss": 1.3348, + "theoretical_loss": 3.4341057995819293, + "tokens_seen": 2183135232 + }, + { + "epoch": 0.66, + "learning_rate": 0.0002736529588766299, + "loss": 1.3394, + "theoretical_loss": 3.4340974217604106, + "tokens_seen": 2183200768 + }, + { + "epoch": 0.66, + "learning_rate": 0.0002736369107321966, + "loss": 1.3076, + "theoretical_loss": 3.4340890442607903, + "tokens_seen": 2183266304 + }, + { + "epoch": 0.66, + "learning_rate": 0.0002736208625877633, + "loss": 1.3374, + "theoretical_loss": 3.4340806670830464, + "tokens_seen": 2183331840 + }, + { + "epoch": 0.66, + "learning_rate": 0.00027360481444333, + "loss": 1.3117, + "theoretical_loss": 3.434072290227157, + "tokens_seen": 2183397376 + }, + { + "epoch": 0.66, + "learning_rate": 0.0002735887662988967, + "loss": 1.3313, + "theoretical_loss": 3.4340639136931, + "tokens_seen": 2183462912 + }, + { + "epoch": 0.66, + "learning_rate": 0.0002735727181544634, + "loss": 1.3959, + "theoretical_loss": 3.4340555374808535, + "tokens_seen": 2183528448 + }, + { + "epoch": 0.66, + "learning_rate": 0.0002735566700100301, + "loss": 1.234, + "theoretical_loss": 3.4340471615903954, + "tokens_seen": 2183593984 + }, + { + "epoch": 0.66, + "learning_rate": 0.0002735406218655968, + "loss": 1.2316, + "theoretical_loss": 3.4340387860217034, + "tokens_seen": 2183659520 + }, + { + "epoch": 0.66, + "learning_rate": 0.0002735245737211635, + "loss": 1.2893, + "theoretical_loss": 3.434030410774756, + "tokens_seen": 2183725056 + }, + { + "epoch": 0.66, + "learning_rate": 0.0002735085255767302, + "loss": 1.3308, + "theoretical_loss": 3.434022035849531, + "tokens_seen": 2183790592 + }, + { + "epoch": 0.66, + "learning_rate": 0.0002734924774322969, + "loss": 1.143, + "theoretical_loss": 3.434013661246006, + "tokens_seen": 2183856128 + }, + { + "epoch": 0.66, + "learning_rate": 0.0002734764292878636, + "loss": 1.3381, + "theoretical_loss": 3.43400528696416, + "tokens_seen": 2183921664 + }, + { + "epoch": 0.66, + "learning_rate": 0.0002734603811434303, + "loss": 1.3072, + "theoretical_loss": 3.43399691300397, + "tokens_seen": 2183987200 + }, + { + "epoch": 0.66, + "learning_rate": 0.000273444332998997, + "loss": 1.3445, + "theoretical_loss": 3.433988539365414, + "tokens_seen": 2184052736 + }, + { + "epoch": 0.66, + "learning_rate": 0.0002734282848545637, + "loss": 1.2982, + "theoretical_loss": 3.43398016604847, + "tokens_seen": 2184118272 + }, + { + "epoch": 0.66, + "learning_rate": 0.0002734122367101304, + "loss": 1.2519, + "theoretical_loss": 3.433971793053117, + "tokens_seen": 2184183808 + }, + { + "epoch": 0.66, + "learning_rate": 0.0002733961885656971, + "loss": 1.4091, + "theoretical_loss": 3.433963420379332, + "tokens_seen": 2184249344 + }, + { + "epoch": 0.66, + "learning_rate": 0.0002733801404212638, + "loss": 1.2671, + "theoretical_loss": 3.4339550480270935, + "tokens_seen": 2184314880 + }, + { + "epoch": 0.66, + "learning_rate": 0.0002733640922768305, + "loss": 1.244, + "theoretical_loss": 3.433946675996379, + "tokens_seen": 2184380416 + }, + { + "epoch": 0.66, + "learning_rate": 0.0002733480441323972, + "loss": 1.3752, + "theoretical_loss": 3.4339383042871674, + "tokens_seen": 2184445952 + }, + { + "epoch": 0.66, + "learning_rate": 0.0002733319959879639, + "loss": 1.3071, + "theoretical_loss": 3.4339299328994355, + "tokens_seen": 2184511488 + }, + { + "epoch": 0.66, + "learning_rate": 0.0002733159478435306, + "loss": 1.2768, + "theoretical_loss": 3.433921561833162, + "tokens_seen": 2184577024 + }, + { + "epoch": 0.66, + "learning_rate": 0.0002732998996990973, + "loss": 1.2435, + "theoretical_loss": 3.433913191088325, + "tokens_seen": 2184642560 + }, + { + "epoch": 0.66, + "learning_rate": 0.000273283851554664, + "loss": 1.2892, + "theoretical_loss": 3.4339048206649023, + "tokens_seen": 2184708096 + }, + { + "epoch": 0.66, + "learning_rate": 0.0002732678034102307, + "loss": 1.3449, + "theoretical_loss": 3.433896450562872, + "tokens_seen": 2184773632 + }, + { + "epoch": 0.66, + "learning_rate": 0.0002732517552657974, + "loss": 1.2829, + "theoretical_loss": 3.4338880807822125, + "tokens_seen": 2184839168 + }, + { + "epoch": 0.66, + "learning_rate": 0.0002732357071213641, + "loss": 1.2705, + "theoretical_loss": 3.4338797113229007, + "tokens_seen": 2184904704 + }, + { + "epoch": 0.66, + "learning_rate": 0.0002732196589769308, + "loss": 1.2355, + "theoretical_loss": 3.433871342184916, + "tokens_seen": 2184970240 + }, + { + "epoch": 0.66, + "learning_rate": 0.0002732036108324975, + "loss": 1.3328, + "theoretical_loss": 3.4338629733682353, + "tokens_seen": 2185035776 + }, + { + "epoch": 0.66, + "learning_rate": 0.0002731875626880642, + "loss": 1.3005, + "theoretical_loss": 3.4338546048728373, + "tokens_seen": 2185101312 + }, + { + "epoch": 0.66, + "learning_rate": 0.0002731715145436309, + "loss": 1.304, + "theoretical_loss": 3.4338462366986997, + "tokens_seen": 2185166848 + }, + { + "epoch": 0.66, + "learning_rate": 0.0002731554663991976, + "loss": 1.2912, + "theoretical_loss": 3.433837868845801, + "tokens_seen": 2185232384 + }, + { + "epoch": 0.66, + "learning_rate": 0.0002731394182547643, + "loss": 1.2991, + "theoretical_loss": 3.4338295013141185, + "tokens_seen": 2185297920 + }, + { + "epoch": 0.66, + "learning_rate": 0.000273123370110331, + "loss": 1.263, + "theoretical_loss": 3.433821134103631, + "tokens_seen": 2185363456 + }, + { + "epoch": 0.66, + "learning_rate": 0.0002731073219658977, + "loss": 1.2914, + "theoretical_loss": 3.4338127672143157, + "tokens_seen": 2185428992 + }, + { + "epoch": 0.66, + "learning_rate": 0.0002730912738214644, + "loss": 1.336, + "theoretical_loss": 3.4338044006461512, + "tokens_seen": 2185494528 + }, + { + "epoch": 0.66, + "learning_rate": 0.0002730752256770311, + "loss": 1.2606, + "theoretical_loss": 3.4337960343991156, + "tokens_seen": 2185560064 + }, + { + "epoch": 0.66, + "objective/train/docs_used": 795634, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.23264741897583, + "objective/train/theoretical_loss": 3.4337876684731867, + "objective/train/tokens_used": 2206085600, + "theoretical_loss": 3.4337876684731867, + "tokens_seen": 2185625600 + }, + { + "epoch": 0.66, + "learning_rate": 0.0002730591775325978, + "loss": 1.3219, + "theoretical_loss": 3.4337876684731867, + "tokens_seen": 2185625600 + }, + { + "epoch": 0.66, + "learning_rate": 0.0002730431293881645, + "loss": 1.2656, + "theoretical_loss": 3.4337793028683428, + "tokens_seen": 2185691136 + }, + { + "epoch": 0.66, + "learning_rate": 0.0002730270812437312, + "loss": 1.279, + "theoretical_loss": 3.433770937584562, + "tokens_seen": 2185756672 + }, + { + "epoch": 0.66, + "learning_rate": 0.0002730110330992979, + "loss": 1.2864, + "theoretical_loss": 3.4337625726218217, + "tokens_seen": 2185822208 + }, + { + "epoch": 0.66, + "learning_rate": 0.0002729949849548646, + "loss": 1.2116, + "theoretical_loss": 3.4337542079801007, + "tokens_seen": 2185887744 + }, + { + "epoch": 0.66, + "learning_rate": 0.0002729789368104313, + "loss": 1.3773, + "theoretical_loss": 3.4337458436593766, + "tokens_seen": 2185953280 + }, + { + "epoch": 0.66, + "learning_rate": 0.000272962888665998, + "loss": 1.3719, + "theoretical_loss": 3.433737479659628, + "tokens_seen": 2186018816 + }, + { + "epoch": 0.66, + "learning_rate": 0.0002729468405215647, + "loss": 1.4075, + "theoretical_loss": 3.433729115980832, + "tokens_seen": 2186084352 + }, + { + "epoch": 0.66, + "learning_rate": 0.0002729307923771314, + "loss": 1.2119, + "theoretical_loss": 3.4337207526229676, + "tokens_seen": 2186149888 + }, + { + "epoch": 0.66, + "learning_rate": 0.0002729147442326981, + "loss": 1.3186, + "theoretical_loss": 3.433712389586012, + "tokens_seen": 2186215424 + }, + { + "epoch": 0.66, + "learning_rate": 0.0002728986960882648, + "loss": 1.2825, + "theoretical_loss": 3.4337040268699446, + "tokens_seen": 2186280960 + }, + { + "epoch": 0.66, + "learning_rate": 0.0002728826479438315, + "loss": 1.4457, + "theoretical_loss": 3.433695664474742, + "tokens_seen": 2186346496 + }, + { + "epoch": 0.66, + "learning_rate": 0.0002728665997993982, + "loss": 1.2775, + "theoretical_loss": 3.4336873024003833, + "tokens_seen": 2186412032 + }, + { + "epoch": 0.66, + "learning_rate": 0.0002728505516549649, + "loss": 1.3154, + "theoretical_loss": 3.433678940646846, + "tokens_seen": 2186477568 + }, + { + "epoch": 0.66, + "learning_rate": 0.0002728345035105316, + "loss": 1.3664, + "theoretical_loss": 3.4336705792141085, + "tokens_seen": 2186543104 + }, + { + "epoch": 0.66, + "learning_rate": 0.0002728184553660983, + "loss": 1.3426, + "theoretical_loss": 3.4336622181021488, + "tokens_seen": 2186608640 + }, + { + "epoch": 0.66, + "learning_rate": 0.000272802407221665, + "loss": 1.3397, + "theoretical_loss": 3.4336538573109445, + "tokens_seen": 2186674176 + }, + { + "epoch": 0.66, + "learning_rate": 0.0002727863590772317, + "loss": 1.4199, + "theoretical_loss": 3.4336454968404744, + "tokens_seen": 2186739712 + }, + { + "epoch": 0.66, + "learning_rate": 0.0002727703109327984, + "loss": 1.3594, + "theoretical_loss": 3.4336371366907166, + "tokens_seen": 2186805248 + }, + { + "epoch": 0.66, + "learning_rate": 0.0002727542627883651, + "loss": 1.3081, + "theoretical_loss": 3.4336287768616485, + "tokens_seen": 2186870784 + }, + { + "epoch": 0.66, + "learning_rate": 0.0002727382146439318, + "loss": 1.2999, + "theoretical_loss": 3.4336204173532483, + "tokens_seen": 2186936320 + }, + { + "epoch": 0.66, + "learning_rate": 0.0002727221664994985, + "loss": 1.336, + "theoretical_loss": 3.4336120581654948, + "tokens_seen": 2187001856 + }, + { + "epoch": 0.66, + "learning_rate": 0.0002727061183550652, + "loss": 1.2478, + "theoretical_loss": 3.4336036992983656, + "tokens_seen": 2187067392 + }, + { + "epoch": 0.66, + "learning_rate": 0.0002726900702106319, + "loss": 1.3662, + "theoretical_loss": 3.4335953407518387, + "tokens_seen": 2187132928 + }, + { + "epoch": 0.66, + "learning_rate": 0.0002726740220661986, + "loss": 1.2654, + "theoretical_loss": 3.4335869825258922, + "tokens_seen": 2187198464 + }, + { + "epoch": 0.66, + "learning_rate": 0.0002726579739217653, + "loss": 1.2834, + "theoretical_loss": 3.4335786246205044, + "tokens_seen": 2187264000 + }, + { + "epoch": 0.66, + "learning_rate": 0.000272641925777332, + "loss": 1.2653, + "theoretical_loss": 3.4335702670356536, + "tokens_seen": 2187329536 + }, + { + "epoch": 0.66, + "learning_rate": 0.00027262587763289873, + "loss": 1.3993, + "theoretical_loss": 3.4335619097713175, + "tokens_seen": 2187395072 + }, + { + "epoch": 0.66, + "learning_rate": 0.00027260982948846543, + "loss": 1.331, + "theoretical_loss": 3.4335535528274743, + "tokens_seen": 2187460608 + }, + { + "epoch": 0.66, + "learning_rate": 0.00027259378134403213, + "loss": 1.4803, + "theoretical_loss": 3.4335451962041024, + "tokens_seen": 2187526144 + }, + { + "epoch": 0.66, + "learning_rate": 0.00027257773319959883, + "loss": 1.3142, + "theoretical_loss": 3.4335368399011794, + "tokens_seen": 2187591680 + }, + { + "epoch": 0.66, + "learning_rate": 0.00027256168505516553, + "loss": 1.3322, + "theoretical_loss": 3.4335284839186837, + "tokens_seen": 2187657216 + }, + { + "epoch": 0.66, + "learning_rate": 0.00027254563691073223, + "loss": 1.3293, + "theoretical_loss": 3.4335201282565935, + "tokens_seen": 2187722752 + }, + { + "epoch": 0.66, + "learning_rate": 0.00027252958876629893, + "loss": 1.3106, + "theoretical_loss": 3.433511772914887, + "tokens_seen": 2187788288 + }, + { + "epoch": 0.66, + "learning_rate": 0.0002725135406218656, + "loss": 1.2446, + "theoretical_loss": 3.433503417893542, + "tokens_seen": 2187853824 + }, + { + "epoch": 0.66, + "learning_rate": 0.0002724974924774323, + "loss": 1.3477, + "theoretical_loss": 3.4334950631925363, + "tokens_seen": 2187919360 + }, + { + "epoch": 0.66, + "learning_rate": 0.000272481444332999, + "loss": 1.3498, + "theoretical_loss": 3.433486708811849, + "tokens_seen": 2187984896 + }, + { + "epoch": 0.66, + "learning_rate": 0.0002724653961885657, + "loss": 1.2824, + "theoretical_loss": 3.4334783547514576, + "tokens_seen": 2188050432 + }, + { + "epoch": 0.66, + "learning_rate": 0.0002724493480441324, + "loss": 1.1796, + "theoretical_loss": 3.43347000101134, + "tokens_seen": 2188115968 + }, + { + "epoch": 0.66, + "learning_rate": 0.0002724332998996991, + "loss": 1.2639, + "theoretical_loss": 3.4334616475914754, + "tokens_seen": 2188181504 + }, + { + "epoch": 0.66, + "learning_rate": 0.0002724172517552658, + "loss": 1.2878, + "theoretical_loss": 3.433453294491841, + "tokens_seen": 2188247040 + }, + { + "epoch": 0.66, + "learning_rate": 0.0002724012036108325, + "loss": 1.2914, + "theoretical_loss": 3.4334449417124144, + "tokens_seen": 2188312576 + }, + { + "epoch": 0.66, + "learning_rate": 0.0002723851554663992, + "loss": 1.3383, + "theoretical_loss": 3.433436589253175, + "tokens_seen": 2188378112 + }, + { + "epoch": 0.66, + "learning_rate": 0.0002723691073219659, + "loss": 1.233, + "theoretical_loss": 3.4334282371141, + "tokens_seen": 2188443648 + }, + { + "epoch": 0.66, + "learning_rate": 0.0002723530591775326, + "loss": 1.2687, + "theoretical_loss": 3.4334198852951685, + "tokens_seen": 2188509184 + }, + { + "epoch": 0.66, + "learning_rate": 0.0002723370110330993, + "loss": 1.3634, + "theoretical_loss": 3.4334115337963578, + "tokens_seen": 2188574720 + }, + { + "epoch": 0.66, + "learning_rate": 0.000272320962888666, + "loss": 1.441, + "theoretical_loss": 3.4334031826176465, + "tokens_seen": 2188640256 + }, + { + "epoch": 0.66, + "learning_rate": 0.0002723049147442327, + "loss": 1.4189, + "theoretical_loss": 3.433394831759012, + "tokens_seen": 2188705792 + }, + { + "epoch": 0.66, + "learning_rate": 0.0002722888665997994, + "loss": 1.3038, + "theoretical_loss": 3.4333864812204338, + "tokens_seen": 2188771328 + }, + { + "epoch": 0.66, + "learning_rate": 0.0002722728184553661, + "loss": 1.3062, + "theoretical_loss": 3.4333781310018887, + "tokens_seen": 2188836864 + }, + { + "epoch": 0.66, + "objective/train/docs_used": 796788, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.3082090616226196, + "objective/train/theoretical_loss": 3.4333697811033557, + "objective/train/tokens_used": 2209362400, + "theoretical_loss": 3.4333697811033557, + "tokens_seen": 2188902400 + }, + { + "epoch": 0.66, + "learning_rate": 0.0002722567703109328, + "loss": 1.2361, + "theoretical_loss": 3.4333697811033557, + "tokens_seen": 2188902400 + }, + { + "epoch": 0.66, + "learning_rate": 0.0002722407221664995, + "loss": 1.3007, + "theoretical_loss": 3.433361431524813, + "tokens_seen": 2188967936 + }, + { + "epoch": 0.66, + "learning_rate": 0.0002722246740220662, + "loss": 1.3148, + "theoretical_loss": 3.4333530822662377, + "tokens_seen": 2189033472 + }, + { + "epoch": 0.66, + "learning_rate": 0.0002722086258776329, + "loss": 1.3106, + "theoretical_loss": 3.4333447333276093, + "tokens_seen": 2189099008 + }, + { + "epoch": 0.66, + "learning_rate": 0.0002721925777331996, + "loss": 1.2665, + "theoretical_loss": 3.433336384708905, + "tokens_seen": 2189164544 + }, + { + "epoch": 0.66, + "learning_rate": 0.0002721765295887663, + "loss": 1.3041, + "theoretical_loss": 3.4333280364101038, + "tokens_seen": 2189230080 + }, + { + "epoch": 0.66, + "learning_rate": 0.000272160481444333, + "loss": 1.3788, + "theoretical_loss": 3.4333196884311827, + "tokens_seen": 2189295616 + }, + { + "epoch": 0.66, + "learning_rate": 0.0002721444332998997, + "loss": 1.2129, + "theoretical_loss": 3.433311340772121, + "tokens_seen": 2189361152 + }, + { + "epoch": 0.66, + "learning_rate": 0.0002721283851554664, + "loss": 1.3555, + "theoretical_loss": 3.4333029934328962, + "tokens_seen": 2189426688 + }, + { + "epoch": 0.66, + "learning_rate": 0.0002721123370110331, + "loss": 1.3611, + "theoretical_loss": 3.433294646413487, + "tokens_seen": 2189492224 + }, + { + "epoch": 0.66, + "learning_rate": 0.0002720962888665998, + "loss": 1.328, + "theoretical_loss": 3.433286299713871, + "tokens_seen": 2189557760 + }, + { + "epoch": 0.66, + "learning_rate": 0.0002720802407221665, + "loss": 1.3186, + "theoretical_loss": 3.4332779533340263, + "tokens_seen": 2189623296 + }, + { + "epoch": 0.66, + "learning_rate": 0.00027206419257773325, + "loss": 1.2522, + "theoretical_loss": 3.433269607273932, + "tokens_seen": 2189688832 + }, + { + "epoch": 0.66, + "learning_rate": 0.00027204814443329995, + "loss": 1.4726, + "theoretical_loss": 3.433261261533566, + "tokens_seen": 2189754368 + }, + { + "epoch": 0.66, + "learning_rate": 0.00027203209628886665, + "loss": 1.3675, + "theoretical_loss": 3.4332529161129055, + "tokens_seen": 2189819904 + }, + { + "epoch": 0.66, + "learning_rate": 0.00027201604814443335, + "loss": 1.4427, + "theoretical_loss": 3.4332445710119295, + "tokens_seen": 2189885440 + }, + { + "epoch": 0.66, + "learning_rate": 0.00027200000000000005, + "loss": 1.4, + "theoretical_loss": 3.4332362262306164, + "tokens_seen": 2189950976 + }, + { + "epoch": 0.66, + "learning_rate": 0.00027198395185556675, + "loss": 1.3934, + "theoretical_loss": 3.4332278817689437, + "tokens_seen": 2190016512 + }, + { + "epoch": 0.66, + "learning_rate": 0.00027196790371113345, + "loss": 1.2453, + "theoretical_loss": 3.43321953762689, + "tokens_seen": 2190082048 + }, + { + "epoch": 0.66, + "learning_rate": 0.00027195185556670015, + "loss": 1.3218, + "theoretical_loss": 3.433211193804434, + "tokens_seen": 2190147584 + }, + { + "epoch": 0.66, + "learning_rate": 0.00027193580742226685, + "loss": 1.3427, + "theoretical_loss": 3.4332028503015524, + "tokens_seen": 2190213120 + }, + { + "epoch": 0.66, + "learning_rate": 0.00027191975927783356, + "loss": 1.3649, + "theoretical_loss": 3.433194507118225, + "tokens_seen": 2190278656 + }, + { + "epoch": 0.66, + "learning_rate": 0.00027190371113340026, + "loss": 1.2765, + "theoretical_loss": 3.4331861642544292, + "tokens_seen": 2190344192 + }, + { + "epoch": 0.66, + "learning_rate": 0.0002718876629889669, + "loss": 1.2032, + "theoretical_loss": 3.433177821710143, + "tokens_seen": 2190409728 + }, + { + "epoch": 0.66, + "learning_rate": 0.0002718716148445336, + "loss": 1.2423, + "theoretical_loss": 3.4331694794853456, + "tokens_seen": 2190475264 + }, + { + "epoch": 0.66, + "learning_rate": 0.0002718555667001003, + "loss": 1.3959, + "theoretical_loss": 3.433161137580014, + "tokens_seen": 2190540800 + }, + { + "epoch": 0.66, + "learning_rate": 0.000271839518555667, + "loss": 1.3624, + "theoretical_loss": 3.4331527959941273, + "tokens_seen": 2190606336 + }, + { + "epoch": 0.66, + "learning_rate": 0.0002718234704112337, + "loss": 1.2503, + "theoretical_loss": 3.4331444547276635, + "tokens_seen": 2190671872 + }, + { + "epoch": 0.66, + "learning_rate": 0.0002718074222668004, + "loss": 1.4212, + "theoretical_loss": 3.4331361137806002, + "tokens_seen": 2190737408 + }, + { + "epoch": 0.66, + "learning_rate": 0.0002717913741223671, + "loss": 1.2555, + "theoretical_loss": 3.433127773152916, + "tokens_seen": 2190802944 + }, + { + "epoch": 0.66, + "learning_rate": 0.0002717753259779338, + "loss": 1.2332, + "theoretical_loss": 3.43311943284459, + "tokens_seen": 2190868480 + }, + { + "epoch": 0.66, + "learning_rate": 0.0002717592778335005, + "loss": 1.3428, + "theoretical_loss": 3.433111092855599, + "tokens_seen": 2190934016 + }, + { + "epoch": 0.66, + "learning_rate": 0.0002717432296890672, + "loss": 1.3246, + "theoretical_loss": 3.433102753185922, + "tokens_seen": 2190999552 + }, + { + "epoch": 0.66, + "learning_rate": 0.0002717271815446339, + "loss": 1.3137, + "theoretical_loss": 3.433094413835537, + "tokens_seen": 2191065088 + }, + { + "epoch": 0.66, + "learning_rate": 0.0002717111334002006, + "loss": 1.2454, + "theoretical_loss": 3.4330860748044225, + "tokens_seen": 2191130624 + }, + { + "epoch": 0.66, + "learning_rate": 0.0002716950852557673, + "loss": 1.2864, + "theoretical_loss": 3.4330777360925566, + "tokens_seen": 2191196160 + }, + { + "epoch": 0.66, + "learning_rate": 0.000271679037111334, + "loss": 1.464, + "theoretical_loss": 3.433069397699917, + "tokens_seen": 2191261696 + }, + { + "epoch": 0.66, + "learning_rate": 0.0002716629889669007, + "loss": 1.2451, + "theoretical_loss": 3.433061059626483, + "tokens_seen": 2191327232 + }, + { + "epoch": 0.66, + "learning_rate": 0.0002716469408224674, + "loss": 1.3115, + "theoretical_loss": 3.433052721872232, + "tokens_seen": 2191392768 + }, + { + "epoch": 0.66, + "learning_rate": 0.0002716308926780341, + "loss": 1.2503, + "theoretical_loss": 3.4330443844371423, + "tokens_seen": 2191458304 + }, + { + "epoch": 0.66, + "learning_rate": 0.0002716148445336008, + "loss": 1.3271, + "theoretical_loss": 3.4330360473211923, + "tokens_seen": 2191523840 + }, + { + "epoch": 0.66, + "learning_rate": 0.0002715987963891675, + "loss": 1.2327, + "theoretical_loss": 3.4330277105243603, + "tokens_seen": 2191589376 + }, + { + "epoch": 0.66, + "learning_rate": 0.0002715827482447342, + "loss": 1.259, + "theoretical_loss": 3.4330193740466246, + "tokens_seen": 2191654912 + }, + { + "epoch": 0.66, + "learning_rate": 0.0002715667001003009, + "loss": 1.2795, + "theoretical_loss": 3.4330110378879635, + "tokens_seen": 2191720448 + }, + { + "epoch": 0.66, + "learning_rate": 0.0002715506519558676, + "loss": 1.2429, + "theoretical_loss": 3.4330027020483547, + "tokens_seen": 2191785984 + }, + { + "epoch": 0.66, + "learning_rate": 0.0002715346038114343, + "loss": 1.3404, + "theoretical_loss": 3.432994366527777, + "tokens_seen": 2191851520 + }, + { + "epoch": 0.66, + "learning_rate": 0.000271518555667001, + "loss": 1.3949, + "theoretical_loss": 3.432986031326209, + "tokens_seen": 2191917056 + }, + { + "epoch": 0.66, + "learning_rate": 0.0002715025075225677, + "loss": 1.2457, + "theoretical_loss": 3.4329776964436274, + "tokens_seen": 2191982592 + }, + { + "epoch": 0.66, + "learning_rate": 0.0002714864593781344, + "loss": 1.2222, + "theoretical_loss": 3.432969361880012, + "tokens_seen": 2192048128 + }, + { + "epoch": 0.66, + "learning_rate": 0.0002714704112337011, + "loss": 1.4213, + "theoretical_loss": 3.432961027635341, + "tokens_seen": 2192113664 + }, + { + "epoch": 0.66, + "objective/train/docs_used": 798413, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.350172519683838, + "objective/train/theoretical_loss": 3.4329526937095913, + "objective/train/tokens_used": 2212639200, + "theoretical_loss": 3.4329526937095913, + "tokens_seen": 2192179200 + }, + { + "epoch": 0.66, + "learning_rate": 0.0002714543630892678, + "loss": 1.3107, + "theoretical_loss": 3.4329526937095913, + "tokens_seen": 2192179200 + }, + { + "epoch": 0.66, + "learning_rate": 0.0002714383149448345, + "loss": 1.413, + "theoretical_loss": 3.4329443601027427, + "tokens_seen": 2192244736 + }, + { + "epoch": 0.66, + "learning_rate": 0.0002714222668004012, + "loss": 1.3359, + "theoretical_loss": 3.432936026814773, + "tokens_seen": 2192310272 + }, + { + "epoch": 0.66, + "learning_rate": 0.0002714062186559679, + "loss": 1.3361, + "theoretical_loss": 3.43292769384566, + "tokens_seen": 2192375808 + }, + { + "epoch": 0.66, + "learning_rate": 0.0002713901705115346, + "loss": 1.3088, + "theoretical_loss": 3.432919361195382, + "tokens_seen": 2192441344 + }, + { + "epoch": 0.66, + "learning_rate": 0.0002713741223671013, + "loss": 1.3657, + "theoretical_loss": 3.4329110288639177, + "tokens_seen": 2192506880 + }, + { + "epoch": 0.66, + "learning_rate": 0.000271358074222668, + "loss": 1.206, + "theoretical_loss": 3.4329026968512455, + "tokens_seen": 2192572416 + }, + { + "epoch": 0.66, + "learning_rate": 0.0002713420260782347, + "loss": 1.3117, + "theoretical_loss": 3.432894365157343, + "tokens_seen": 2192637952 + }, + { + "epoch": 0.66, + "learning_rate": 0.0002713259779338014, + "loss": 1.3597, + "theoretical_loss": 3.4328860337821894, + "tokens_seen": 2192703488 + }, + { + "epoch": 0.66, + "learning_rate": 0.0002713099297893681, + "loss": 1.352, + "theoretical_loss": 3.432877702725762, + "tokens_seen": 2192769024 + }, + { + "epoch": 0.66, + "learning_rate": 0.0002712938816449348, + "loss": 1.3314, + "theoretical_loss": 3.4328693719880397, + "tokens_seen": 2192834560 + }, + { + "epoch": 0.66, + "learning_rate": 0.0002712778335005015, + "loss": 1.4422, + "theoretical_loss": 3.4328610415690006, + "tokens_seen": 2192900096 + }, + { + "epoch": 0.66, + "learning_rate": 0.0002712617853560682, + "loss": 1.2184, + "theoretical_loss": 3.432852711468623, + "tokens_seen": 2192965632 + }, + { + "epoch": 0.66, + "learning_rate": 0.00027124573721163493, + "loss": 1.2554, + "theoretical_loss": 3.432844381686885, + "tokens_seen": 2193031168 + }, + { + "epoch": 0.66, + "learning_rate": 0.00027122968906720163, + "loss": 1.45, + "theoretical_loss": 3.4328360522237653, + "tokens_seen": 2193096704 + }, + { + "epoch": 0.66, + "learning_rate": 0.00027121364092276833, + "loss": 1.2108, + "theoretical_loss": 3.4328277230792423, + "tokens_seen": 2193162240 + }, + { + "epoch": 0.66, + "learning_rate": 0.00027119759277833503, + "loss": 1.3211, + "theoretical_loss": 3.4328193942532934, + "tokens_seen": 2193227776 + }, + { + "epoch": 0.66, + "learning_rate": 0.00027118154463390173, + "loss": 1.2801, + "theoretical_loss": 3.4328110657458977, + "tokens_seen": 2193293312 + }, + { + "epoch": 0.66, + "learning_rate": 0.00027116549648946843, + "loss": 1.2863, + "theoretical_loss": 3.432802737557033, + "tokens_seen": 2193358848 + }, + { + "epoch": 0.66, + "learning_rate": 0.00027114944834503513, + "loss": 1.1795, + "theoretical_loss": 3.4327944096866783, + "tokens_seen": 2193424384 + }, + { + "epoch": 0.66, + "learning_rate": 0.00027113340020060183, + "loss": 1.3999, + "theoretical_loss": 3.432786082134811, + "tokens_seen": 2193489920 + }, + { + "epoch": 0.66, + "learning_rate": 0.0002711173520561685, + "loss": 1.3099, + "theoretical_loss": 3.43277775490141, + "tokens_seen": 2193555456 + }, + { + "epoch": 0.66, + "learning_rate": 0.0002711013039117352, + "loss": 1.3466, + "theoretical_loss": 3.432769427986454, + "tokens_seen": 2193620992 + }, + { + "epoch": 0.66, + "learning_rate": 0.0002710852557673019, + "loss": 1.2894, + "theoretical_loss": 3.43276110138992, + "tokens_seen": 2193686528 + }, + { + "epoch": 0.66, + "learning_rate": 0.0002710692076228686, + "loss": 1.3529, + "theoretical_loss": 3.4327527751117874, + "tokens_seen": 2193752064 + }, + { + "epoch": 0.66, + "learning_rate": 0.0002710531594784353, + "loss": 1.3415, + "theoretical_loss": 3.4327444491520342, + "tokens_seen": 2193817600 + }, + { + "epoch": 0.66, + "learning_rate": 0.000271037111334002, + "loss": 1.3219, + "theoretical_loss": 3.432736123510639, + "tokens_seen": 2193883136 + }, + { + "epoch": 0.66, + "learning_rate": 0.0002710210631895687, + "loss": 1.359, + "theoretical_loss": 3.4327277981875794, + "tokens_seen": 2193948672 + }, + { + "epoch": 0.66, + "learning_rate": 0.0002710050150451354, + "loss": 1.2775, + "theoretical_loss": 3.4327194731828348, + "tokens_seen": 2194014208 + }, + { + "epoch": 0.66, + "learning_rate": 0.0002709889669007021, + "loss": 1.3866, + "theoretical_loss": 3.432711148496382, + "tokens_seen": 2194079744 + }, + { + "epoch": 0.66, + "learning_rate": 0.0002709729187562688, + "loss": 1.4329, + "theoretical_loss": 3.4327028241282007, + "tokens_seen": 2194145280 + }, + { + "epoch": 0.66, + "learning_rate": 0.0002709568706118355, + "loss": 1.2641, + "theoretical_loss": 3.4326945000782687, + "tokens_seen": 2194210816 + }, + { + "epoch": 0.66, + "learning_rate": 0.0002709408224674022, + "loss": 1.3212, + "theoretical_loss": 3.4326861763465644, + "tokens_seen": 2194276352 + }, + { + "epoch": 0.66, + "learning_rate": 0.0002709247743229689, + "loss": 1.3922, + "theoretical_loss": 3.432677852933066, + "tokens_seen": 2194341888 + }, + { + "epoch": 0.66, + "learning_rate": 0.0002709087261785356, + "loss": 1.2983, + "theoretical_loss": 3.432669529837752, + "tokens_seen": 2194407424 + }, + { + "epoch": 0.66, + "learning_rate": 0.0002708926780341023, + "loss": 1.4072, + "theoretical_loss": 3.4326612070606006, + "tokens_seen": 2194472960 + }, + { + "epoch": 0.67, + "learning_rate": 0.000270876629889669, + "loss": 1.265, + "theoretical_loss": 3.4326528846015902, + "tokens_seen": 2194538496 + }, + { + "epoch": 0.67, + "learning_rate": 0.00027086058174523574, + "loss": 1.3202, + "theoretical_loss": 3.432644562460699, + "tokens_seen": 2194604032 + }, + { + "epoch": 0.67, + "learning_rate": 0.00027084453360080244, + "loss": 1.2583, + "theoretical_loss": 3.4326362406379056, + "tokens_seen": 2194669568 + }, + { + "epoch": 0.67, + "learning_rate": 0.00027082848545636914, + "loss": 1.2019, + "theoretical_loss": 3.4326279191331883, + "tokens_seen": 2194735104 + }, + { + "epoch": 0.67, + "learning_rate": 0.00027081243731193584, + "loss": 1.3966, + "theoretical_loss": 3.4326195979465255, + "tokens_seen": 2194800640 + }, + { + "epoch": 0.67, + "learning_rate": 0.00027079638916750255, + "loss": 1.3939, + "theoretical_loss": 3.432611277077895, + "tokens_seen": 2194866176 + }, + { + "epoch": 0.67, + "learning_rate": 0.00027078034102306925, + "loss": 1.2543, + "theoretical_loss": 3.4326029565272758, + "tokens_seen": 2194931712 + }, + { + "epoch": 0.67, + "learning_rate": 0.00027076429287863595, + "loss": 1.3909, + "theoretical_loss": 3.432594636294646, + "tokens_seen": 2194997248 + }, + { + "epoch": 0.67, + "learning_rate": 0.00027074824473420265, + "loss": 1.3159, + "theoretical_loss": 3.4325863163799837, + "tokens_seen": 2195062784 + }, + { + "epoch": 0.67, + "learning_rate": 0.00027073219658976935, + "loss": 1.3816, + "theoretical_loss": 3.4325779967832677, + "tokens_seen": 2195128320 + }, + { + "epoch": 0.67, + "learning_rate": 0.00027071614844533605, + "loss": 1.2756, + "theoretical_loss": 3.4325696775044765, + "tokens_seen": 2195193856 + }, + { + "epoch": 0.67, + "learning_rate": 0.00027070010030090275, + "loss": 1.2324, + "theoretical_loss": 3.4325613585435875, + "tokens_seen": 2195259392 + }, + { + "epoch": 0.67, + "learning_rate": 0.00027068405215646945, + "loss": 1.2882, + "theoretical_loss": 3.43255303990058, + "tokens_seen": 2195324928 + }, + { + "epoch": 0.67, + "learning_rate": 0.00027066800401203615, + "loss": 1.2026, + "theoretical_loss": 3.4325447215754323, + "tokens_seen": 2195390464 + }, + { + "epoch": 0.67, + "objective/train/docs_used": 799213, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.3696959018707275, + "objective/train/theoretical_loss": 3.4325364035681223, + "objective/train/tokens_used": 2215916000, + "theoretical_loss": 3.4325364035681223, + "tokens_seen": 2195456000 + }, + { + "epoch": 0.67, + "learning_rate": 0.00027065195586760285, + "loss": 1.3443, + "theoretical_loss": 3.4325364035681223, + "tokens_seen": 2195456000 + }, + { + "epoch": 0.67, + "learning_rate": 0.00027063590772316955, + "loss": 1.2846, + "theoretical_loss": 3.4325280858786287, + "tokens_seen": 2195521536 + }, + { + "epoch": 0.67, + "learning_rate": 0.00027061985957873625, + "loss": 1.3732, + "theoretical_loss": 3.4325197685069297, + "tokens_seen": 2195587072 + }, + { + "epoch": 0.67, + "learning_rate": 0.00027060381143430295, + "loss": 1.3361, + "theoretical_loss": 3.4325114514530037, + "tokens_seen": 2195652608 + }, + { + "epoch": 0.67, + "learning_rate": 0.00027058776328986965, + "loss": 1.3291, + "theoretical_loss": 3.4325031347168293, + "tokens_seen": 2195718144 + }, + { + "epoch": 0.67, + "learning_rate": 0.00027057171514543635, + "loss": 1.3851, + "theoretical_loss": 3.4324948182983843, + "tokens_seen": 2195783680 + }, + { + "epoch": 0.67, + "learning_rate": 0.00027055566700100305, + "loss": 1.3264, + "theoretical_loss": 3.432486502197648, + "tokens_seen": 2195849216 + }, + { + "epoch": 0.67, + "learning_rate": 0.00027053961885656976, + "loss": 1.3421, + "theoretical_loss": 3.4324781864145977, + "tokens_seen": 2195914752 + }, + { + "epoch": 0.67, + "learning_rate": 0.00027052357071213646, + "loss": 1.3531, + "theoretical_loss": 3.4324698709492125, + "tokens_seen": 2195980288 + }, + { + "epoch": 0.67, + "learning_rate": 0.00027050752256770316, + "loss": 1.4277, + "theoretical_loss": 3.432461555801471, + "tokens_seen": 2196045824 + }, + { + "epoch": 0.67, + "learning_rate": 0.00027049147442326986, + "loss": 1.2468, + "theoretical_loss": 3.432453240971351, + "tokens_seen": 2196111360 + }, + { + "epoch": 0.67, + "learning_rate": 0.0002704754262788365, + "loss": 1.2103, + "theoretical_loss": 3.432444926458831, + "tokens_seen": 2196176896 + }, + { + "epoch": 0.67, + "learning_rate": 0.0002704593781344032, + "loss": 1.3775, + "theoretical_loss": 3.4324366122638894, + "tokens_seen": 2196242432 + }, + { + "epoch": 0.67, + "learning_rate": 0.0002704433299899699, + "loss": 1.3568, + "theoretical_loss": 3.4324282983865047, + "tokens_seen": 2196307968 + }, + { + "epoch": 0.67, + "learning_rate": 0.0002704272818455366, + "loss": 1.3339, + "theoretical_loss": 3.4324199848266557, + "tokens_seen": 2196373504 + }, + { + "epoch": 0.67, + "learning_rate": 0.0002704112337011033, + "loss": 1.4153, + "theoretical_loss": 3.43241167158432, + "tokens_seen": 2196439040 + }, + { + "epoch": 0.67, + "learning_rate": 0.00027039518555667, + "loss": 1.3944, + "theoretical_loss": 3.432403358659476, + "tokens_seen": 2196504576 + }, + { + "epoch": 0.67, + "learning_rate": 0.0002703791374122367, + "loss": 1.3583, + "theoretical_loss": 3.432395046052103, + "tokens_seen": 2196570112 + }, + { + "epoch": 0.67, + "learning_rate": 0.0002703630892678034, + "loss": 1.3389, + "theoretical_loss": 3.432386733762179, + "tokens_seen": 2196635648 + }, + { + "epoch": 0.67, + "learning_rate": 0.0002703470411233701, + "loss": 1.3006, + "theoretical_loss": 3.432378421789682, + "tokens_seen": 2196701184 + }, + { + "epoch": 0.67, + "learning_rate": 0.0002703309929789368, + "loss": 1.2789, + "theoretical_loss": 3.432370110134591, + "tokens_seen": 2196766720 + }, + { + "epoch": 0.67, + "learning_rate": 0.0002703149448345035, + "loss": 1.2692, + "theoretical_loss": 3.432361798796884, + "tokens_seen": 2196832256 + }, + { + "epoch": 0.67, + "learning_rate": 0.0002702988966900702, + "loss": 1.371, + "theoretical_loss": 3.4323534877765396, + "tokens_seen": 2196897792 + }, + { + "epoch": 0.67, + "learning_rate": 0.0002702828485456369, + "loss": 1.307, + "theoretical_loss": 3.4323451770735356, + "tokens_seen": 2196963328 + }, + { + "epoch": 0.67, + "learning_rate": 0.0002702668004012036, + "loss": 1.2664, + "theoretical_loss": 3.4323368666878515, + "tokens_seen": 2197028864 + }, + { + "epoch": 0.67, + "learning_rate": 0.0002702507522567703, + "loss": 1.3025, + "theoretical_loss": 3.4323285566194652, + "tokens_seen": 2197094400 + }, + { + "epoch": 0.67, + "learning_rate": 0.000270234704112337, + "loss": 1.2662, + "theoretical_loss": 3.432320246868355, + "tokens_seen": 2197159936 + }, + { + "epoch": 0.67, + "learning_rate": 0.0002702186559679037, + "loss": 1.3374, + "theoretical_loss": 3.4323119374344992, + "tokens_seen": 2197225472 + }, + { + "epoch": 0.67, + "learning_rate": 0.0002702026078234704, + "loss": 1.1659, + "theoretical_loss": 3.4323036283178765, + "tokens_seen": 2197291008 + }, + { + "epoch": 0.67, + "learning_rate": 0.0002701865596790371, + "loss": 1.3448, + "theoretical_loss": 3.4322953195184653, + "tokens_seen": 2197356544 + }, + { + "epoch": 0.67, + "learning_rate": 0.0002701705115346038, + "loss": 1.3557, + "theoretical_loss": 3.432287011036244, + "tokens_seen": 2197422080 + }, + { + "epoch": 0.67, + "learning_rate": 0.0002701544633901705, + "loss": 1.2843, + "theoretical_loss": 3.432278702871191, + "tokens_seen": 2197487616 + }, + { + "epoch": 0.67, + "learning_rate": 0.0002701384152457372, + "loss": 1.1668, + "theoretical_loss": 3.432270395023285, + "tokens_seen": 2197553152 + }, + { + "epoch": 0.67, + "learning_rate": 0.0002701223671013039, + "loss": 1.3654, + "theoretical_loss": 3.432262087492504, + "tokens_seen": 2197618688 + }, + { + "epoch": 0.67, + "learning_rate": 0.0002701063189568706, + "loss": 1.4438, + "theoretical_loss": 3.4322537802788267, + "tokens_seen": 2197684224 + }, + { + "epoch": 0.67, + "learning_rate": 0.0002700902708124373, + "loss": 1.2213, + "theoretical_loss": 3.4322454733822316, + "tokens_seen": 2197749760 + }, + { + "epoch": 0.67, + "learning_rate": 0.000270074222668004, + "loss": 1.3714, + "theoretical_loss": 3.4322371668026967, + "tokens_seen": 2197815296 + }, + { + "epoch": 0.67, + "learning_rate": 0.0002700581745235707, + "loss": 1.2794, + "theoretical_loss": 3.432228860540201, + "tokens_seen": 2197880832 + }, + { + "epoch": 0.67, + "learning_rate": 0.0002700421263791374, + "loss": 1.3773, + "theoretical_loss": 3.4322205545947226, + "tokens_seen": 2197946368 + }, + { + "epoch": 0.67, + "learning_rate": 0.0002700260782347041, + "loss": 1.3893, + "theoretical_loss": 3.4322122489662403, + "tokens_seen": 2198011904 + }, + { + "epoch": 0.67, + "learning_rate": 0.0002700100300902708, + "loss": 1.3133, + "theoretical_loss": 3.432203943654732, + "tokens_seen": 2198077440 + }, + { + "epoch": 0.67, + "learning_rate": 0.0002699939819458375, + "loss": 1.3591, + "theoretical_loss": 3.432195638660177, + "tokens_seen": 2198142976 + }, + { + "epoch": 0.67, + "learning_rate": 0.0002699779338014042, + "loss": 1.271, + "theoretical_loss": 3.4321873339825526, + "tokens_seen": 2198208512 + }, + { + "epoch": 0.67, + "learning_rate": 0.0002699618856569709, + "loss": 1.3066, + "theoretical_loss": 3.432179029621838, + "tokens_seen": 2198274048 + }, + { + "epoch": 0.67, + "learning_rate": 0.0002699458375125376, + "loss": 1.3061, + "theoretical_loss": 3.432170725578012, + "tokens_seen": 2198339584 + }, + { + "epoch": 0.67, + "learning_rate": 0.0002699297893681043, + "loss": 1.277, + "theoretical_loss": 3.4321624218510522, + "tokens_seen": 2198405120 + }, + { + "epoch": 0.67, + "learning_rate": 0.000269913741223671, + "loss": 1.3069, + "theoretical_loss": 3.4321541184409377, + "tokens_seen": 2198470656 + }, + { + "epoch": 0.67, + "learning_rate": 0.0002698976930792377, + "loss": 1.2795, + "theoretical_loss": 3.4321458153476465, + "tokens_seen": 2198536192 + }, + { + "epoch": 0.67, + "learning_rate": 0.00026988164493480443, + "loss": 1.4245, + "theoretical_loss": 3.432137512571157, + "tokens_seen": 2198601728 + }, + { + "epoch": 0.67, + "learning_rate": 0.00026986559679037113, + "loss": 1.3109, + "theoretical_loss": 3.432129210111449, + "tokens_seen": 2198667264 + }, + { + "epoch": 0.67, + "objective/train/docs_used": 800708, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.2260329723358154, + "objective/train/theoretical_loss": 3.432120907968499, + "objective/train/tokens_used": 2219192800, + "theoretical_loss": 3.432120907968499, + "tokens_seen": 2198732800 + }, + { + "epoch": 0.67, + "learning_rate": 0.00026984954864593783, + "loss": 1.3247, + "theoretical_loss": 3.432120907968499, + "tokens_seen": 2198732800 + }, + { + "epoch": 0.67, + "learning_rate": 0.00026983350050150453, + "loss": 1.2912, + "theoretical_loss": 3.4321126061422866, + "tokens_seen": 2198798336 + }, + { + "epoch": 0.67, + "learning_rate": 0.00026981745235707123, + "loss": 1.3543, + "theoretical_loss": 3.43210430463279, + "tokens_seen": 2198863872 + }, + { + "epoch": 0.67, + "learning_rate": 0.00026980140421263793, + "loss": 1.3629, + "theoretical_loss": 3.4320960034399883, + "tokens_seen": 2198929408 + }, + { + "epoch": 0.67, + "learning_rate": 0.00026978535606820463, + "loss": 1.3505, + "theoretical_loss": 3.4320877025638588, + "tokens_seen": 2198994944 + }, + { + "epoch": 0.67, + "learning_rate": 0.00026976930792377133, + "loss": 1.311, + "theoretical_loss": 3.432079402004381, + "tokens_seen": 2199060480 + }, + { + "epoch": 0.67, + "learning_rate": 0.00026975325977933803, + "loss": 1.3193, + "theoretical_loss": 3.432071101761533, + "tokens_seen": 2199126016 + }, + { + "epoch": 0.67, + "learning_rate": 0.00026973721163490473, + "loss": 1.3495, + "theoretical_loss": 3.4320628018352934, + "tokens_seen": 2199191552 + }, + { + "epoch": 0.67, + "learning_rate": 0.00026972116349047143, + "loss": 1.3703, + "theoretical_loss": 3.4320545022256406, + "tokens_seen": 2199257088 + }, + { + "epoch": 0.67, + "learning_rate": 0.0002697051153460381, + "loss": 1.3877, + "theoretical_loss": 3.4320462029325527, + "tokens_seen": 2199322624 + }, + { + "epoch": 0.67, + "learning_rate": 0.0002696890672016048, + "loss": 1.2373, + "theoretical_loss": 3.432037903956009, + "tokens_seen": 2199388160 + }, + { + "epoch": 0.67, + "learning_rate": 0.0002696730190571715, + "loss": 1.3334, + "theoretical_loss": 3.432029605295987, + "tokens_seen": 2199453696 + }, + { + "epoch": 0.67, + "learning_rate": 0.00026965697091273824, + "loss": 1.3999, + "theoretical_loss": 3.4320213069524663, + "tokens_seen": 2199519232 + }, + { + "epoch": 0.67, + "learning_rate": 0.00026964092276830494, + "loss": 1.2937, + "theoretical_loss": 3.432013008925425, + "tokens_seen": 2199584768 + }, + { + "epoch": 0.67, + "learning_rate": 0.00026962487462387164, + "loss": 1.3293, + "theoretical_loss": 3.432004711214841, + "tokens_seen": 2199650304 + }, + { + "epoch": 0.67, + "learning_rate": 0.00026960882647943834, + "loss": 1.2319, + "theoretical_loss": 3.4319964138206935, + "tokens_seen": 2199715840 + }, + { + "epoch": 0.67, + "learning_rate": 0.00026959277833500504, + "loss": 1.2876, + "theoretical_loss": 3.4319881167429607, + "tokens_seen": 2199781376 + }, + { + "epoch": 0.67, + "learning_rate": 0.00026957673019057174, + "loss": 1.2808, + "theoretical_loss": 3.4319798199816214, + "tokens_seen": 2199846912 + }, + { + "epoch": 0.67, + "learning_rate": 0.00026956068204613844, + "loss": 1.3017, + "theoretical_loss": 3.431971523536654, + "tokens_seen": 2199912448 + }, + { + "epoch": 0.67, + "learning_rate": 0.00026954463390170514, + "loss": 1.318, + "theoretical_loss": 3.431963227408036, + "tokens_seen": 2199977984 + }, + { + "epoch": 0.67, + "learning_rate": 0.00026952858575727184, + "loss": 1.3308, + "theoretical_loss": 3.431954931595748, + "tokens_seen": 2200043520 + }, + { + "epoch": 0.67, + "learning_rate": 0.00026951253761283854, + "loss": 1.1911, + "theoretical_loss": 3.4319466360997666, + "tokens_seen": 2200109056 + }, + { + "epoch": 0.67, + "learning_rate": 0.00026949648946840524, + "loss": 1.3689, + "theoretical_loss": 3.431938340920071, + "tokens_seen": 2200174592 + }, + { + "epoch": 0.67, + "learning_rate": 0.00026948044132397194, + "loss": 1.3117, + "theoretical_loss": 3.43193004605664, + "tokens_seen": 2200240128 + }, + { + "epoch": 0.67, + "learning_rate": 0.00026946439317953864, + "loss": 1.3754, + "theoretical_loss": 3.431921751509452, + "tokens_seen": 2200305664 + }, + { + "epoch": 0.67, + "learning_rate": 0.00026944834503510534, + "loss": 1.3134, + "theoretical_loss": 3.4319134572784855, + "tokens_seen": 2200371200 + }, + { + "epoch": 0.67, + "learning_rate": 0.00026943229689067205, + "loss": 1.3425, + "theoretical_loss": 3.4319051633637185, + "tokens_seen": 2200436736 + }, + { + "epoch": 0.67, + "learning_rate": 0.00026941624874623875, + "loss": 1.3, + "theoretical_loss": 3.4318968697651306, + "tokens_seen": 2200502272 + }, + { + "epoch": 0.67, + "learning_rate": 0.00026940020060180545, + "loss": 1.2883, + "theoretical_loss": 3.431888576482699, + "tokens_seen": 2200567808 + }, + { + "epoch": 0.67, + "learning_rate": 0.00026938415245737215, + "loss": 1.3492, + "theoretical_loss": 3.4318802835164037, + "tokens_seen": 2200633344 + }, + { + "epoch": 0.67, + "learning_rate": 0.00026936810431293885, + "loss": 1.2277, + "theoretical_loss": 3.431871990866222, + "tokens_seen": 2200698880 + }, + { + "epoch": 0.67, + "learning_rate": 0.00026935205616850555, + "loss": 1.2526, + "theoretical_loss": 3.431863698532133, + "tokens_seen": 2200764416 + }, + { + "epoch": 0.67, + "learning_rate": 0.00026933600802407225, + "loss": 1.3517, + "theoretical_loss": 3.4318554065141154, + "tokens_seen": 2200829952 + }, + { + "epoch": 0.67, + "learning_rate": 0.00026931995987963895, + "loss": 1.2758, + "theoretical_loss": 3.4318471148121468, + "tokens_seen": 2200895488 + }, + { + "epoch": 0.67, + "learning_rate": 0.00026930391173520565, + "loss": 1.4001, + "theoretical_loss": 3.431838823426207, + "tokens_seen": 2200961024 + }, + { + "epoch": 0.67, + "learning_rate": 0.00026928786359077235, + "loss": 1.2741, + "theoretical_loss": 3.431830532356274, + "tokens_seen": 2201026560 + }, + { + "epoch": 0.67, + "learning_rate": 0.00026927181544633905, + "loss": 1.4307, + "theoretical_loss": 3.431822241602326, + "tokens_seen": 2201092096 + }, + { + "epoch": 0.67, + "learning_rate": 0.00026925576730190575, + "loss": 1.3578, + "theoretical_loss": 3.431813951164342, + "tokens_seen": 2201157632 + }, + { + "epoch": 0.67, + "learning_rate": 0.00026923971915747245, + "loss": 1.3025, + "theoretical_loss": 3.4318056610423007, + "tokens_seen": 2201223168 + }, + { + "epoch": 0.67, + "learning_rate": 0.00026922367101303915, + "loss": 1.315, + "theoretical_loss": 3.43179737123618, + "tokens_seen": 2201288704 + }, + { + "epoch": 0.67, + "learning_rate": 0.00026920762286860585, + "loss": 1.2428, + "theoretical_loss": 3.431789081745959, + "tokens_seen": 2201354240 + }, + { + "epoch": 0.67, + "learning_rate": 0.00026919157472417255, + "loss": 1.3461, + "theoretical_loss": 3.431780792571616, + "tokens_seen": 2201419776 + }, + { + "epoch": 0.67, + "learning_rate": 0.00026917552657973926, + "loss": 1.2962, + "theoretical_loss": 3.43177250371313, + "tokens_seen": 2201485312 + }, + { + "epoch": 0.67, + "learning_rate": 0.00026915947843530596, + "loss": 1.3238, + "theoretical_loss": 3.431764215170479, + "tokens_seen": 2201550848 + }, + { + "epoch": 0.67, + "learning_rate": 0.00026914343029087266, + "loss": 1.1853, + "theoretical_loss": 3.4317559269436417, + "tokens_seen": 2201616384 + }, + { + "epoch": 0.67, + "learning_rate": 0.00026912738214643936, + "loss": 1.2952, + "theoretical_loss": 3.431747639032597, + "tokens_seen": 2201681920 + }, + { + "epoch": 0.67, + "learning_rate": 0.00026911133400200606, + "loss": 1.2809, + "theoretical_loss": 3.431739351437323, + "tokens_seen": 2201747456 + }, + { + "epoch": 0.67, + "learning_rate": 0.00026909528585757276, + "loss": 1.1795, + "theoretical_loss": 3.4317310641577987, + "tokens_seen": 2201812992 + }, + { + "epoch": 0.67, + "learning_rate": 0.0002690792377131394, + "loss": 1.2305, + "theoretical_loss": 3.431722777194002, + "tokens_seen": 2201878528 + }, + { + "epoch": 0.67, + "learning_rate": 0.0002690631895687061, + "loss": 1.3803, + "theoretical_loss": 3.4317144905459127, + "tokens_seen": 2201944064 + }, + { + "epoch": 0.67, + "objective/train/docs_used": 801578, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.3714072704315186, + "objective/train/theoretical_loss": 3.4317062042135076, + "objective/train/tokens_used": 2222469600, + "theoretical_loss": 3.4317062042135076, + "tokens_seen": 2202009600 + }, + { + "epoch": 0.67, + "learning_rate": 0.0002690471414242728, + "loss": 1.3512, + "theoretical_loss": 3.4317062042135076, + "tokens_seen": 2202009600 + }, + { + "epoch": 0.67, + "learning_rate": 0.0002690310932798395, + "loss": 1.2582, + "theoretical_loss": 3.4316979181967673, + "tokens_seen": 2202075136 + }, + { + "epoch": 0.67, + "learning_rate": 0.0002690150451354062, + "loss": 1.2999, + "theoretical_loss": 3.4316896324956687, + "tokens_seen": 2202140672 + }, + { + "epoch": 0.67, + "learning_rate": 0.0002689989969909729, + "loss": 1.2986, + "theoretical_loss": 3.4316813471101915, + "tokens_seen": 2202206208 + }, + { + "epoch": 0.67, + "learning_rate": 0.0002689829488465396, + "loss": 1.38, + "theoretical_loss": 3.4316730620403133, + "tokens_seen": 2202271744 + }, + { + "epoch": 0.67, + "learning_rate": 0.0002689669007021063, + "loss": 1.2758, + "theoretical_loss": 3.431664777286014, + "tokens_seen": 2202337280 + }, + { + "epoch": 0.67, + "learning_rate": 0.000268950852557673, + "loss": 1.3749, + "theoretical_loss": 3.431656492847271, + "tokens_seen": 2202402816 + }, + { + "epoch": 0.67, + "learning_rate": 0.0002689348044132397, + "loss": 1.3186, + "theoretical_loss": 3.431648208724063, + "tokens_seen": 2202468352 + }, + { + "epoch": 0.67, + "learning_rate": 0.0002689187562688064, + "loss": 1.3137, + "theoretical_loss": 3.4316399249163694, + "tokens_seen": 2202533888 + }, + { + "epoch": 0.67, + "learning_rate": 0.0002689027081243731, + "loss": 1.3372, + "theoretical_loss": 3.431631641424168, + "tokens_seen": 2202599424 + }, + { + "epoch": 0.67, + "learning_rate": 0.0002688866599799398, + "loss": 1.2694, + "theoretical_loss": 3.4316233582474376, + "tokens_seen": 2202664960 + }, + { + "epoch": 0.67, + "learning_rate": 0.0002688706118355065, + "loss": 1.2618, + "theoretical_loss": 3.431615075386157, + "tokens_seen": 2202730496 + }, + { + "epoch": 0.67, + "learning_rate": 0.0002688545636910732, + "loss": 1.2915, + "theoretical_loss": 3.431606792840305, + "tokens_seen": 2202796032 + }, + { + "epoch": 0.67, + "learning_rate": 0.0002688385155466399, + "loss": 1.3908, + "theoretical_loss": 3.4315985106098594, + "tokens_seen": 2202861568 + }, + { + "epoch": 0.67, + "learning_rate": 0.0002688224674022066, + "loss": 1.4121, + "theoretical_loss": 3.4315902286948, + "tokens_seen": 2202927104 + }, + { + "epoch": 0.67, + "learning_rate": 0.0002688064192577733, + "loss": 1.3482, + "theoretical_loss": 3.431581947095104, + "tokens_seen": 2202992640 + }, + { + "epoch": 0.67, + "learning_rate": 0.00026879037111334, + "loss": 1.3726, + "theoretical_loss": 3.431573665810751, + "tokens_seen": 2203058176 + }, + { + "epoch": 0.67, + "learning_rate": 0.0002687743229689067, + "loss": 1.3362, + "theoretical_loss": 3.43156538484172, + "tokens_seen": 2203123712 + }, + { + "epoch": 0.67, + "learning_rate": 0.0002687582748244734, + "loss": 1.2755, + "theoretical_loss": 3.4315571041879878, + "tokens_seen": 2203189248 + }, + { + "epoch": 0.67, + "learning_rate": 0.0002687422266800401, + "loss": 1.2096, + "theoretical_loss": 3.431548823849535, + "tokens_seen": 2203254784 + }, + { + "epoch": 0.67, + "learning_rate": 0.0002687261785356068, + "loss": 1.283, + "theoretical_loss": 3.431540543826339, + "tokens_seen": 2203320320 + }, + { + "epoch": 0.67, + "learning_rate": 0.0002687101303911735, + "loss": 1.2794, + "theoretical_loss": 3.431532264118379, + "tokens_seen": 2203385856 + }, + { + "epoch": 0.67, + "learning_rate": 0.0002686940822467402, + "loss": 1.3411, + "theoretical_loss": 3.4315239847256334, + "tokens_seen": 2203451392 + }, + { + "epoch": 0.67, + "learning_rate": 0.0002686780341023069, + "loss": 1.3742, + "theoretical_loss": 3.4315157056480805, + "tokens_seen": 2203516928 + }, + { + "epoch": 0.67, + "learning_rate": 0.0002686619859578736, + "loss": 1.3105, + "theoretical_loss": 3.4315074268856995, + "tokens_seen": 2203582464 + }, + { + "epoch": 0.67, + "learning_rate": 0.0002686459378134403, + "loss": 1.4027, + "theoretical_loss": 3.431499148438469, + "tokens_seen": 2203648000 + }, + { + "epoch": 0.67, + "learning_rate": 0.000268629889669007, + "loss": 1.2076, + "theoretical_loss": 3.4314908703063676, + "tokens_seen": 2203713536 + }, + { + "epoch": 0.67, + "learning_rate": 0.0002686138415245737, + "loss": 1.3475, + "theoretical_loss": 3.4314825924893735, + "tokens_seen": 2203779072 + }, + { + "epoch": 0.67, + "learning_rate": 0.0002685977933801404, + "loss": 1.4434, + "theoretical_loss": 3.4314743149874656, + "tokens_seen": 2203844608 + }, + { + "epoch": 0.67, + "learning_rate": 0.0002685817452357071, + "loss": 1.2587, + "theoretical_loss": 3.4314660378006225, + "tokens_seen": 2203910144 + }, + { + "epoch": 0.67, + "learning_rate": 0.0002685656970912738, + "loss": 1.2265, + "theoretical_loss": 3.431457760928823, + "tokens_seen": 2203975680 + }, + { + "epoch": 0.67, + "learning_rate": 0.0002685496489468405, + "loss": 1.3341, + "theoretical_loss": 3.4314494843720458, + "tokens_seen": 2204041216 + }, + { + "epoch": 0.67, + "learning_rate": 0.0002685336008024072, + "loss": 1.3639, + "theoretical_loss": 3.431441208130269, + "tokens_seen": 2204106752 + }, + { + "epoch": 0.67, + "learning_rate": 0.00026851755265797393, + "loss": 1.3217, + "theoretical_loss": 3.431432932203472, + "tokens_seen": 2204172288 + }, + { + "epoch": 0.67, + "learning_rate": 0.00026850150451354063, + "loss": 1.3349, + "theoretical_loss": 3.4314246565916324, + "tokens_seen": 2204237824 + }, + { + "epoch": 0.67, + "learning_rate": 0.00026848545636910733, + "loss": 1.3274, + "theoretical_loss": 3.4314163812947305, + "tokens_seen": 2204303360 + }, + { + "epoch": 0.67, + "learning_rate": 0.00026846940822467403, + "loss": 1.4396, + "theoretical_loss": 3.4314081063127433, + "tokens_seen": 2204368896 + }, + { + "epoch": 0.67, + "learning_rate": 0.00026845336008024073, + "loss": 1.2448, + "theoretical_loss": 3.4313998316456504, + "tokens_seen": 2204434432 + }, + { + "epoch": 0.67, + "learning_rate": 0.00026843731193580743, + "loss": 1.3431, + "theoretical_loss": 3.4313915572934297, + "tokens_seen": 2204499968 + }, + { + "epoch": 0.67, + "learning_rate": 0.00026842126379137413, + "loss": 1.2853, + "theoretical_loss": 3.431383283256061, + "tokens_seen": 2204565504 + }, + { + "epoch": 0.67, + "learning_rate": 0.00026840521564694083, + "loss": 1.2834, + "theoretical_loss": 3.431375009533522, + "tokens_seen": 2204631040 + }, + { + "epoch": 0.67, + "learning_rate": 0.00026838916750250753, + "loss": 1.4363, + "theoretical_loss": 3.4313667361257916, + "tokens_seen": 2204696576 + }, + { + "epoch": 0.67, + "learning_rate": 0.00026837311935807423, + "loss": 1.2967, + "theoretical_loss": 3.4313584630328484, + "tokens_seen": 2204762112 + }, + { + "epoch": 0.67, + "learning_rate": 0.00026835707121364093, + "loss": 1.3542, + "theoretical_loss": 3.4313501902546717, + "tokens_seen": 2204827648 + }, + { + "epoch": 0.67, + "learning_rate": 0.00026834102306920763, + "loss": 1.3291, + "theoretical_loss": 3.431341917791239, + "tokens_seen": 2204893184 + }, + { + "epoch": 0.67, + "learning_rate": 0.00026832497492477433, + "loss": 1.364, + "theoretical_loss": 3.43133364564253, + "tokens_seen": 2204958720 + }, + { + "epoch": 0.67, + "learning_rate": 0.00026830892678034104, + "loss": 1.3309, + "theoretical_loss": 3.431325373808523, + "tokens_seen": 2205024256 + }, + { + "epoch": 0.67, + "learning_rate": 0.00026829287863590774, + "loss": 1.2356, + "theoretical_loss": 3.4313171022891966, + "tokens_seen": 2205089792 + }, + { + "epoch": 0.67, + "learning_rate": 0.00026827683049147444, + "loss": 1.302, + "theoretical_loss": 3.4313088310845297, + "tokens_seen": 2205155328 + }, + { + "epoch": 0.67, + "learning_rate": 0.00026826078234704114, + "loss": 1.3485, + "theoretical_loss": 3.431300560194501, + "tokens_seen": 2205220864 + }, + { + "epoch": 0.67, + "objective/train/docs_used": 802757, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.3230600357055664, + "objective/train/theoretical_loss": 3.4312922896190887, + "objective/train/tokens_used": 2225746400, + "theoretical_loss": 3.4312922896190887, + "tokens_seen": 2205286400 + }, + { + "epoch": 0.67, + "learning_rate": 0.00026824473420260784, + "loss": 1.4036, + "theoretical_loss": 3.4312922896190887, + "tokens_seen": 2205286400 + }, + { + "epoch": 0.67, + "learning_rate": 0.00026822868605817454, + "loss": 1.354, + "theoretical_loss": 3.4312840193582717, + "tokens_seen": 2205351936 + }, + { + "epoch": 0.67, + "learning_rate": 0.00026821263791374124, + "loss": 1.3093, + "theoretical_loss": 3.431275749412029, + "tokens_seen": 2205417472 + }, + { + "epoch": 0.67, + "learning_rate": 0.00026819658976930794, + "loss": 1.3418, + "theoretical_loss": 3.431267479780339, + "tokens_seen": 2205483008 + }, + { + "epoch": 0.67, + "learning_rate": 0.00026818054162487464, + "loss": 1.4537, + "theoretical_loss": 3.4312592104631805, + "tokens_seen": 2205548544 + }, + { + "epoch": 0.67, + "learning_rate": 0.00026816449348044134, + "loss": 1.2242, + "theoretical_loss": 3.431250941460532, + "tokens_seen": 2205614080 + }, + { + "epoch": 0.67, + "learning_rate": 0.00026814844533600804, + "loss": 1.3035, + "theoretical_loss": 3.431242672772373, + "tokens_seen": 2205679616 + }, + { + "epoch": 0.67, + "learning_rate": 0.00026813239719157474, + "loss": 1.2172, + "theoretical_loss": 3.4312344043986807, + "tokens_seen": 2205745152 + }, + { + "epoch": 0.67, + "learning_rate": 0.00026811634904714144, + "loss": 1.2763, + "theoretical_loss": 3.431226136339435, + "tokens_seen": 2205810688 + }, + { + "epoch": 0.67, + "learning_rate": 0.00026810030090270814, + "loss": 1.3455, + "theoretical_loss": 3.431217868594614, + "tokens_seen": 2205876224 + }, + { + "epoch": 0.67, + "learning_rate": 0.00026808425275827484, + "loss": 1.2975, + "theoretical_loss": 3.431209601164197, + "tokens_seen": 2205941760 + }, + { + "epoch": 0.67, + "learning_rate": 0.00026806820461384154, + "loss": 1.2969, + "theoretical_loss": 3.431201334048162, + "tokens_seen": 2206007296 + }, + { + "epoch": 0.67, + "learning_rate": 0.00026805215646940825, + "loss": 1.279, + "theoretical_loss": 3.4311930672464888, + "tokens_seen": 2206072832 + }, + { + "epoch": 0.67, + "learning_rate": 0.00026803610832497495, + "loss": 1.3929, + "theoretical_loss": 3.4311848007591546, + "tokens_seen": 2206138368 + }, + { + "epoch": 0.67, + "learning_rate": 0.00026802006018054165, + "loss": 1.35, + "theoretical_loss": 3.431176534586139, + "tokens_seen": 2206203904 + }, + { + "epoch": 0.67, + "learning_rate": 0.00026800401203610835, + "loss": 1.3097, + "theoretical_loss": 3.4311682687274208, + "tokens_seen": 2206269440 + }, + { + "epoch": 0.67, + "learning_rate": 0.00026798796389167505, + "loss": 1.405, + "theoretical_loss": 3.431160003182978, + "tokens_seen": 2206334976 + }, + { + "epoch": 0.67, + "learning_rate": 0.00026797191574724175, + "loss": 1.3229, + "theoretical_loss": 3.4311517379527903, + "tokens_seen": 2206400512 + }, + { + "epoch": 0.67, + "learning_rate": 0.00026795586760280845, + "loss": 1.2852, + "theoretical_loss": 3.431143473036836, + "tokens_seen": 2206466048 + }, + { + "epoch": 0.67, + "learning_rate": 0.00026793981945837515, + "loss": 1.2381, + "theoretical_loss": 3.4311352084350935, + "tokens_seen": 2206531584 + }, + { + "epoch": 0.67, + "learning_rate": 0.00026792377131394185, + "loss": 1.3133, + "theoretical_loss": 3.431126944147542, + "tokens_seen": 2206597120 + }, + { + "epoch": 0.67, + "learning_rate": 0.00026790772316950855, + "loss": 1.3645, + "theoretical_loss": 3.4311186801741598, + "tokens_seen": 2206662656 + }, + { + "epoch": 0.67, + "learning_rate": 0.00026789167502507525, + "loss": 1.3701, + "theoretical_loss": 3.4311104165149255, + "tokens_seen": 2206728192 + }, + { + "epoch": 0.67, + "learning_rate": 0.00026787562688064195, + "loss": 1.2738, + "theoretical_loss": 3.4311021531698187, + "tokens_seen": 2206793728 + }, + { + "epoch": 0.67, + "learning_rate": 0.00026785957873620865, + "loss": 1.3474, + "theoretical_loss": 3.4310938901388175, + "tokens_seen": 2206859264 + }, + { + "epoch": 0.67, + "learning_rate": 0.00026784353059177535, + "loss": 1.2606, + "theoretical_loss": 3.4310856274219006, + "tokens_seen": 2206924800 + }, + { + "epoch": 0.67, + "learning_rate": 0.00026782748244734205, + "loss": 1.2749, + "theoretical_loss": 3.4310773650190467, + "tokens_seen": 2206990336 + }, + { + "epoch": 0.67, + "learning_rate": 0.00026781143430290876, + "loss": 1.4263, + "theoretical_loss": 3.431069102930235, + "tokens_seen": 2207055872 + }, + { + "epoch": 0.67, + "learning_rate": 0.00026779538615847546, + "loss": 1.2771, + "theoretical_loss": 3.4310608411554435, + "tokens_seen": 2207121408 + }, + { + "epoch": 0.67, + "learning_rate": 0.00026777933801404216, + "loss": 1.2498, + "theoretical_loss": 3.4310525796946516, + "tokens_seen": 2207186944 + }, + { + "epoch": 0.67, + "learning_rate": 0.00026776328986960886, + "loss": 1.3341, + "theoretical_loss": 3.431044318547838, + "tokens_seen": 2207252480 + }, + { + "epoch": 0.67, + "learning_rate": 0.00026774724172517556, + "loss": 1.3649, + "theoretical_loss": 3.431036057714981, + "tokens_seen": 2207318016 + }, + { + "epoch": 0.67, + "learning_rate": 0.00026773119358074226, + "loss": 1.3528, + "theoretical_loss": 3.4310277971960597, + "tokens_seen": 2207383552 + }, + { + "epoch": 0.67, + "learning_rate": 0.00026771514543630896, + "loss": 1.3689, + "theoretical_loss": 3.4310195369910526, + "tokens_seen": 2207449088 + }, + { + "epoch": 0.67, + "learning_rate": 0.00026769909729187566, + "loss": 1.3071, + "theoretical_loss": 3.431011277099939, + "tokens_seen": 2207514624 + }, + { + "epoch": 0.67, + "learning_rate": 0.0002676830491474423, + "loss": 1.3327, + "theoretical_loss": 3.431003017522697, + "tokens_seen": 2207580160 + }, + { + "epoch": 0.67, + "learning_rate": 0.000267667001003009, + "loss": 1.4543, + "theoretical_loss": 3.4309947582593057, + "tokens_seen": 2207645696 + }, + { + "epoch": 0.67, + "learning_rate": 0.0002676509528585757, + "loss": 1.3705, + "theoretical_loss": 3.4309864993097436, + "tokens_seen": 2207711232 + }, + { + "epoch": 0.67, + "learning_rate": 0.0002676349047141424, + "loss": 1.2157, + "theoretical_loss": 3.4309782406739897, + "tokens_seen": 2207776768 + }, + { + "epoch": 0.67, + "learning_rate": 0.0002676188565697091, + "loss": 1.2859, + "theoretical_loss": 3.430969982352023, + "tokens_seen": 2207842304 + }, + { + "epoch": 0.67, + "learning_rate": 0.0002676028084252758, + "loss": 1.3283, + "theoretical_loss": 3.4309617243438217, + "tokens_seen": 2207907840 + }, + { + "epoch": 0.67, + "learning_rate": 0.0002675867602808425, + "loss": 1.3112, + "theoretical_loss": 3.4309534666493646, + "tokens_seen": 2207973376 + }, + { + "epoch": 0.67, + "learning_rate": 0.0002675707121364092, + "loss": 1.3646, + "theoretical_loss": 3.4309452092686312, + "tokens_seen": 2208038912 + }, + { + "epoch": 0.67, + "learning_rate": 0.0002675546639919759, + "loss": 1.3874, + "theoretical_loss": 3.4309369522015993, + "tokens_seen": 2208104448 + }, + { + "epoch": 0.67, + "learning_rate": 0.0002675386158475426, + "loss": 1.2996, + "theoretical_loss": 3.4309286954482485, + "tokens_seen": 2208169984 + }, + { + "epoch": 0.67, + "learning_rate": 0.0002675225677031093, + "loss": 1.4776, + "theoretical_loss": 3.430920439008557, + "tokens_seen": 2208235520 + }, + { + "epoch": 0.67, + "learning_rate": 0.000267506519558676, + "loss": 1.3263, + "theoretical_loss": 3.430912182882504, + "tokens_seen": 2208301056 + }, + { + "epoch": 0.67, + "learning_rate": 0.0002674904714142427, + "loss": 1.3173, + "theoretical_loss": 3.4309039270700676, + "tokens_seen": 2208366592 + }, + { + "epoch": 0.67, + "learning_rate": 0.0002674744232698094, + "loss": 1.3447, + "theoretical_loss": 3.4308956715712275, + "tokens_seen": 2208432128 + }, + { + "epoch": 0.67, + "learning_rate": 0.0002674583751253761, + "loss": 1.3704, + "theoretical_loss": 3.430887416385962, + "tokens_seen": 2208497664 + }, + { + "epoch": 0.67, + "objective/train/docs_used": 804023, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.3815486431121826, + "objective/train/theoretical_loss": 3.4308791615142495, + "objective/train/tokens_used": 2229023200, + "theoretical_loss": 3.4308791615142495, + "tokens_seen": 2208563200 + }, + { + "epoch": 0.67, + "learning_rate": 0.0002674423269809428, + "loss": 1.3519, + "theoretical_loss": 3.4308791615142495, + "tokens_seen": 2208563200 + }, + { + "epoch": 0.67, + "learning_rate": 0.0002674262788365095, + "loss": 1.2627, + "theoretical_loss": 3.43087090695607, + "tokens_seen": 2208628736 + }, + { + "epoch": 0.67, + "learning_rate": 0.0002674102306920762, + "loss": 1.3878, + "theoretical_loss": 3.430862652711401, + "tokens_seen": 2208694272 + }, + { + "epoch": 0.67, + "learning_rate": 0.0002673941825476429, + "loss": 1.2905, + "theoretical_loss": 3.4308543987802214, + "tokens_seen": 2208759808 + }, + { + "epoch": 0.67, + "learning_rate": 0.0002673781344032096, + "loss": 1.3644, + "theoretical_loss": 3.4308461451625107, + "tokens_seen": 2208825344 + }, + { + "epoch": 0.67, + "learning_rate": 0.0002673620862587763, + "loss": 1.2673, + "theoretical_loss": 3.430837891858247, + "tokens_seen": 2208890880 + }, + { + "epoch": 0.67, + "learning_rate": 0.000267346038114343, + "loss": 1.3005, + "theoretical_loss": 3.43082963886741, + "tokens_seen": 2208956416 + }, + { + "epoch": 0.67, + "learning_rate": 0.0002673299899699097, + "loss": 1.1997, + "theoretical_loss": 3.430821386189978, + "tokens_seen": 2209021952 + }, + { + "epoch": 0.67, + "learning_rate": 0.0002673139418254764, + "loss": 1.2959, + "theoretical_loss": 3.4308131338259296, + "tokens_seen": 2209087488 + }, + { + "epoch": 0.67, + "learning_rate": 0.0002672978936810431, + "loss": 1.3645, + "theoretical_loss": 3.4308048817752437, + "tokens_seen": 2209153024 + }, + { + "epoch": 0.67, + "learning_rate": 0.0002672818455366098, + "loss": 1.3372, + "theoretical_loss": 3.430796630037899, + "tokens_seen": 2209218560 + }, + { + "epoch": 0.67, + "learning_rate": 0.0002672657973921766, + "loss": 1.3788, + "theoretical_loss": 3.430788378613875, + "tokens_seen": 2209284096 + }, + { + "epoch": 0.67, + "learning_rate": 0.0002672497492477433, + "loss": 1.2945, + "theoretical_loss": 3.4307801275031498, + "tokens_seen": 2209349632 + }, + { + "epoch": 0.67, + "learning_rate": 0.00026723370110331, + "loss": 1.3752, + "theoretical_loss": 3.4307718767057023, + "tokens_seen": 2209415168 + }, + { + "epoch": 0.67, + "learning_rate": 0.0002672176529588767, + "loss": 1.3715, + "theoretical_loss": 3.4307636262215113, + "tokens_seen": 2209480704 + }, + { + "epoch": 0.67, + "learning_rate": 0.0002672016048144434, + "loss": 1.2612, + "theoretical_loss": 3.4307553760505556, + "tokens_seen": 2209546240 + }, + { + "epoch": 0.67, + "learning_rate": 0.0002671855566700101, + "loss": 1.2937, + "theoretical_loss": 3.4307471261928146, + "tokens_seen": 2209611776 + }, + { + "epoch": 0.67, + "learning_rate": 0.0002671695085255768, + "loss": 1.3953, + "theoretical_loss": 3.4307388766482667, + "tokens_seen": 2209677312 + }, + { + "epoch": 0.67, + "learning_rate": 0.0002671534603811435, + "loss": 1.2854, + "theoretical_loss": 3.43073062741689, + "tokens_seen": 2209742848 + }, + { + "epoch": 0.67, + "learning_rate": 0.0002671374122367102, + "loss": 1.3181, + "theoretical_loss": 3.4307223784986647, + "tokens_seen": 2209808384 + }, + { + "epoch": 0.67, + "learning_rate": 0.0002671213640922769, + "loss": 1.301, + "theoretical_loss": 3.430714129893569, + "tokens_seen": 2209873920 + }, + { + "epoch": 0.67, + "learning_rate": 0.0002671053159478436, + "loss": 1.2462, + "theoretical_loss": 3.430705881601581, + "tokens_seen": 2209939456 + }, + { + "epoch": 0.67, + "learning_rate": 0.0002670892678034103, + "loss": 1.3581, + "theoretical_loss": 3.4306976336226804, + "tokens_seen": 2210004992 + }, + { + "epoch": 0.67, + "learning_rate": 0.000267073219658977, + "loss": 1.3585, + "theoretical_loss": 3.430689385956846, + "tokens_seen": 2210070528 + }, + { + "epoch": 0.67, + "learning_rate": 0.0002670571715145437, + "loss": 1.3309, + "theoretical_loss": 3.4306811386040565, + "tokens_seen": 2210136064 + }, + { + "epoch": 0.67, + "learning_rate": 0.00026704112337011033, + "loss": 1.3098, + "theoretical_loss": 3.430672891564291, + "tokens_seen": 2210201600 + }, + { + "epoch": 0.67, + "learning_rate": 0.00026702507522567703, + "loss": 1.3593, + "theoretical_loss": 3.4306646448375275, + "tokens_seen": 2210267136 + }, + { + "epoch": 0.67, + "learning_rate": 0.00026700902708124373, + "loss": 1.2462, + "theoretical_loss": 3.4306563984237455, + "tokens_seen": 2210332672 + }, + { + "epoch": 0.67, + "learning_rate": 0.00026699297893681043, + "loss": 1.2889, + "theoretical_loss": 3.4306481523229237, + "tokens_seen": 2210398208 + }, + { + "epoch": 0.67, + "learning_rate": 0.00026697693079237713, + "loss": 1.3513, + "theoretical_loss": 3.430639906535041, + "tokens_seen": 2210463744 + }, + { + "epoch": 0.67, + "learning_rate": 0.00026696088264794383, + "loss": 1.2445, + "theoretical_loss": 3.4306316610600764, + "tokens_seen": 2210529280 + }, + { + "epoch": 0.67, + "learning_rate": 0.00026694483450351054, + "loss": 1.3856, + "theoretical_loss": 3.4306234158980082, + "tokens_seen": 2210594816 + }, + { + "epoch": 0.67, + "learning_rate": 0.00026692878635907724, + "loss": 1.2798, + "theoretical_loss": 3.430615171048816, + "tokens_seen": 2210660352 + }, + { + "epoch": 0.67, + "learning_rate": 0.00026691273821464394, + "loss": 1.3828, + "theoretical_loss": 3.4306069265124783, + "tokens_seen": 2210725888 + }, + { + "epoch": 0.67, + "learning_rate": 0.00026689669007021064, + "loss": 1.2042, + "theoretical_loss": 3.4305986822889736, + "tokens_seen": 2210791424 + }, + { + "epoch": 0.67, + "learning_rate": 0.00026688064192577734, + "loss": 1.3897, + "theoretical_loss": 3.4305904383782813, + "tokens_seen": 2210856960 + }, + { + "epoch": 0.67, + "learning_rate": 0.00026686459378134404, + "loss": 1.3188, + "theoretical_loss": 3.4305821947803796, + "tokens_seen": 2210922496 + }, + { + "epoch": 0.67, + "learning_rate": 0.00026684854563691074, + "loss": 1.269, + "theoretical_loss": 3.4305739514952482, + "tokens_seen": 2210988032 + }, + { + "epoch": 0.67, + "learning_rate": 0.00026683249749247744, + "loss": 1.3645, + "theoretical_loss": 3.430565708522866, + "tokens_seen": 2211053568 + }, + { + "epoch": 0.67, + "learning_rate": 0.00026681644934804414, + "loss": 1.1663, + "theoretical_loss": 3.4305574658632105, + "tokens_seen": 2211119104 + }, + { + "epoch": 0.67, + "learning_rate": 0.00026680040120361084, + "loss": 1.2231, + "theoretical_loss": 3.4305492235162616, + "tokens_seen": 2211184640 + }, + { + "epoch": 0.67, + "learning_rate": 0.00026678435305917754, + "loss": 1.3632, + "theoretical_loss": 3.4305409814819985, + "tokens_seen": 2211250176 + }, + { + "epoch": 0.67, + "learning_rate": 0.00026676830491474424, + "loss": 1.364, + "theoretical_loss": 3.4305327397603995, + "tokens_seen": 2211315712 + }, + { + "epoch": 0.67, + "learning_rate": 0.00026675225677031094, + "loss": 1.3108, + "theoretical_loss": 3.430524498351444, + "tokens_seen": 2211381248 + }, + { + "epoch": 0.67, + "learning_rate": 0.00026673620862587764, + "loss": 1.2854, + "theoretical_loss": 3.4305162572551096, + "tokens_seen": 2211446784 + }, + { + "epoch": 0.67, + "learning_rate": 0.00026672016048144434, + "loss": 1.3453, + "theoretical_loss": 3.430508016471377, + "tokens_seen": 2211512320 + }, + { + "epoch": 0.67, + "learning_rate": 0.00026670411233701104, + "loss": 1.3058, + "theoretical_loss": 3.4304997760002234, + "tokens_seen": 2211577856 + }, + { + "epoch": 0.67, + "learning_rate": 0.00026668806419257775, + "loss": 1.2882, + "theoretical_loss": 3.430491535841629, + "tokens_seen": 2211643392 + }, + { + "epoch": 0.67, + "learning_rate": 0.00026667201604814445, + "loss": 1.2226, + "theoretical_loss": 3.4304832959955713, + "tokens_seen": 2211708928 + }, + { + "epoch": 0.67, + "learning_rate": 0.00026665596790371115, + "loss": 1.1664, + "theoretical_loss": 3.4304750564620305, + "tokens_seen": 2211774464 + }, + { + "epoch": 0.67, + "objective/train/docs_used": 805208, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.3710589408874512, + "objective/train/theoretical_loss": 3.4304668172409847, + "objective/train/tokens_used": 2232300000, + "theoretical_loss": 3.4304668172409847, + "tokens_seen": 2211840000 + }, + { + "epoch": 0.67, + "learning_rate": 0.00026663991975927785, + "loss": 1.3769, + "theoretical_loss": 3.4304668172409847, + "tokens_seen": 2211840000 + }, + { + "epoch": 0.67, + "learning_rate": 0.00026662387161484455, + "loss": 1.3229, + "theoretical_loss": 3.4304585783324133, + "tokens_seen": 2211905536 + }, + { + "epoch": 0.67, + "learning_rate": 0.00026660782347041125, + "loss": 1.218, + "theoretical_loss": 3.430450339736295, + "tokens_seen": 2211971072 + }, + { + "epoch": 0.67, + "learning_rate": 0.00026659177532597795, + "loss": 1.3734, + "theoretical_loss": 3.4304421014526083, + "tokens_seen": 2212036608 + }, + { + "epoch": 0.67, + "learning_rate": 0.00026657572718154465, + "loss": 1.3997, + "theoretical_loss": 3.430433863481333, + "tokens_seen": 2212102144 + }, + { + "epoch": 0.67, + "learning_rate": 0.00026655967903711135, + "loss": 1.3677, + "theoretical_loss": 3.430425625822447, + "tokens_seen": 2212167680 + }, + { + "epoch": 0.67, + "learning_rate": 0.00026654363089267805, + "loss": 1.2873, + "theoretical_loss": 3.43041738847593, + "tokens_seen": 2212233216 + }, + { + "epoch": 0.67, + "learning_rate": 0.00026652758274824475, + "loss": 1.3888, + "theoretical_loss": 3.4304091514417605, + "tokens_seen": 2212298752 + }, + { + "epoch": 0.67, + "learning_rate": 0.00026651153460381145, + "loss": 1.292, + "theoretical_loss": 3.4304009147199173, + "tokens_seen": 2212364288 + }, + { + "epoch": 0.67, + "learning_rate": 0.00026649548645937815, + "loss": 1.3509, + "theoretical_loss": 3.430392678310379, + "tokens_seen": 2212429824 + }, + { + "epoch": 0.67, + "learning_rate": 0.00026647943831494485, + "loss": 1.3642, + "theoretical_loss": 3.4303844422131258, + "tokens_seen": 2212495360 + }, + { + "epoch": 0.67, + "learning_rate": 0.00026646339017051155, + "loss": 1.4696, + "theoretical_loss": 3.430376206428135, + "tokens_seen": 2212560896 + }, + { + "epoch": 0.67, + "learning_rate": 0.00026644734202607825, + "loss": 1.3269, + "theoretical_loss": 3.4303679709553867, + "tokens_seen": 2212626432 + }, + { + "epoch": 0.67, + "learning_rate": 0.00026643129388164496, + "loss": 1.2469, + "theoretical_loss": 3.4303597357948594, + "tokens_seen": 2212691968 + }, + { + "epoch": 0.67, + "learning_rate": 0.00026641524573721166, + "loss": 1.2587, + "theoretical_loss": 3.430351500946532, + "tokens_seen": 2212757504 + }, + { + "epoch": 0.67, + "learning_rate": 0.00026639919759277836, + "loss": 1.3277, + "theoretical_loss": 3.4303432664103832, + "tokens_seen": 2212823040 + }, + { + "epoch": 0.67, + "learning_rate": 0.00026638314944834506, + "loss": 1.1852, + "theoretical_loss": 3.4303350321863926, + "tokens_seen": 2212888576 + }, + { + "epoch": 0.67, + "learning_rate": 0.00026636710130391176, + "loss": 1.2925, + "theoretical_loss": 3.4303267982745385, + "tokens_seen": 2212954112 + }, + { + "epoch": 0.67, + "learning_rate": 0.00026635105315947846, + "loss": 1.3647, + "theoretical_loss": 3.4303185646747996, + "tokens_seen": 2213019648 + }, + { + "epoch": 0.67, + "learning_rate": 0.00026633500501504516, + "loss": 1.3131, + "theoretical_loss": 3.4303103313871555, + "tokens_seen": 2213085184 + }, + { + "epoch": 0.67, + "learning_rate": 0.00026631895687061186, + "loss": 1.307, + "theoretical_loss": 3.430302098411585, + "tokens_seen": 2213150720 + }, + { + "epoch": 0.67, + "learning_rate": 0.00026630290872617856, + "loss": 1.2927, + "theoretical_loss": 3.430293865748067, + "tokens_seen": 2213216256 + }, + { + "epoch": 0.67, + "learning_rate": 0.00026628686058174526, + "loss": 1.349, + "theoretical_loss": 3.43028563339658, + "tokens_seen": 2213281792 + }, + { + "epoch": 0.67, + "learning_rate": 0.0002662708124373119, + "loss": 1.1841, + "theoretical_loss": 3.4302774013571034, + "tokens_seen": 2213347328 + }, + { + "epoch": 0.67, + "learning_rate": 0.0002662547642928786, + "loss": 1.3658, + "theoretical_loss": 3.430269169629616, + "tokens_seen": 2213412864 + }, + { + "epoch": 0.67, + "learning_rate": 0.0002662387161484453, + "loss": 1.2858, + "theoretical_loss": 3.4302609382140967, + "tokens_seen": 2213478400 + }, + { + "epoch": 0.67, + "learning_rate": 0.000266222668004012, + "loss": 1.2549, + "theoretical_loss": 3.430252707110524, + "tokens_seen": 2213543936 + }, + { + "epoch": 0.67, + "learning_rate": 0.0002662066198595787, + "loss": 1.3588, + "theoretical_loss": 3.430244476318878, + "tokens_seen": 2213609472 + }, + { + "epoch": 0.67, + "learning_rate": 0.0002661905717151454, + "loss": 1.4093, + "theoretical_loss": 3.4302362458391364, + "tokens_seen": 2213675008 + }, + { + "epoch": 0.67, + "learning_rate": 0.0002661745235707121, + "loss": 1.3327, + "theoretical_loss": 3.4302280156712794, + "tokens_seen": 2213740544 + }, + { + "epoch": 0.67, + "learning_rate": 0.0002661584754262788, + "loss": 1.1739, + "theoretical_loss": 3.430219785815285, + "tokens_seen": 2213806080 + }, + { + "epoch": 0.67, + "learning_rate": 0.0002661424272818455, + "loss": 1.2967, + "theoretical_loss": 3.430211556271132, + "tokens_seen": 2213871616 + }, + { + "epoch": 0.67, + "learning_rate": 0.0002661263791374122, + "loss": 1.3669, + "theoretical_loss": 3.4302033270388, + "tokens_seen": 2213937152 + }, + { + "epoch": 0.67, + "learning_rate": 0.0002661103309929789, + "loss": 1.2345, + "theoretical_loss": 3.4301950981182676, + "tokens_seen": 2214002688 + }, + { + "epoch": 0.67, + "learning_rate": 0.0002660942828485456, + "loss": 1.4488, + "theoretical_loss": 3.430186869509514, + "tokens_seen": 2214068224 + }, + { + "epoch": 0.67, + "learning_rate": 0.0002660782347041123, + "loss": 1.178, + "theoretical_loss": 3.4301786412125184, + "tokens_seen": 2214133760 + }, + { + "epoch": 0.67, + "learning_rate": 0.00026606218655967907, + "loss": 1.2246, + "theoretical_loss": 3.4301704132272586, + "tokens_seen": 2214199296 + }, + { + "epoch": 0.67, + "learning_rate": 0.00026604613841524577, + "loss": 1.3008, + "theoretical_loss": 3.4301621855537148, + "tokens_seen": 2214264832 + }, + { + "epoch": 0.67, + "learning_rate": 0.00026603009027081247, + "loss": 1.3395, + "theoretical_loss": 3.4301539581918656, + "tokens_seen": 2214330368 + }, + { + "epoch": 0.67, + "learning_rate": 0.00026601404212637917, + "loss": 1.2964, + "theoretical_loss": 3.43014573114169, + "tokens_seen": 2214395904 + }, + { + "epoch": 0.67, + "learning_rate": 0.00026599799398194587, + "loss": 1.4309, + "theoretical_loss": 3.4301375044031666, + "tokens_seen": 2214461440 + }, + { + "epoch": 0.67, + "learning_rate": 0.0002659819458375126, + "loss": 1.3682, + "theoretical_loss": 3.4301292779762744, + "tokens_seen": 2214526976 + }, + { + "epoch": 0.67, + "learning_rate": 0.0002659658976930793, + "loss": 1.3467, + "theoretical_loss": 3.430121051860993, + "tokens_seen": 2214592512 + }, + { + "epoch": 0.67, + "learning_rate": 0.000265949849548646, + "loss": 1.351, + "theoretical_loss": 3.4301128260573006, + "tokens_seen": 2214658048 + }, + { + "epoch": 0.67, + "learning_rate": 0.0002659338014042127, + "loss": 1.3814, + "theoretical_loss": 3.4301046005651767, + "tokens_seen": 2214723584 + }, + { + "epoch": 0.67, + "learning_rate": 0.0002659177532597794, + "loss": 1.247, + "theoretical_loss": 3.4300963753846005, + "tokens_seen": 2214789120 + }, + { + "epoch": 0.67, + "learning_rate": 0.0002659017051153461, + "loss": 1.2919, + "theoretical_loss": 3.4300881505155503, + "tokens_seen": 2214854656 + }, + { + "epoch": 0.67, + "learning_rate": 0.0002658856569709128, + "loss": 1.3489, + "theoretical_loss": 3.4300799259580055, + "tokens_seen": 2214920192 + }, + { + "epoch": 0.67, + "learning_rate": 0.0002658696088264795, + "loss": 1.3418, + "theoretical_loss": 3.4300717017119444, + "tokens_seen": 2214985728 + }, + { + "epoch": 0.67, + "learning_rate": 0.0002658535606820462, + "loss": 1.2341, + "theoretical_loss": 3.430063477777347, + "tokens_seen": 2215051264 + }, + { + "epoch": 0.67, + "objective/train/docs_used": 806281, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.2311538457870483, + "objective/train/theoretical_loss": 3.430055254154192, + "objective/train/tokens_used": 2235576800, + "theoretical_loss": 3.430055254154192, + "tokens_seen": 2215116800 + }, + { + "epoch": 0.67, + "learning_rate": 0.0002658375125376129, + "loss": 1.2659, + "theoretical_loss": 3.430055254154192, + "tokens_seen": 2215116800 + }, + { + "epoch": 0.67, + "learning_rate": 0.0002658214643931796, + "loss": 1.3338, + "theoretical_loss": 3.4300470308424584, + "tokens_seen": 2215182336 + }, + { + "epoch": 0.67, + "learning_rate": 0.0002658054162487463, + "loss": 1.3617, + "theoretical_loss": 3.430038807842125, + "tokens_seen": 2215247872 + }, + { + "epoch": 0.67, + "learning_rate": 0.000265789368104313, + "loss": 1.2346, + "theoretical_loss": 3.4300305851531703, + "tokens_seen": 2215313408 + }, + { + "epoch": 0.67, + "learning_rate": 0.0002657733199598797, + "loss": 1.3093, + "theoretical_loss": 3.4300223627755746, + "tokens_seen": 2215378944 + }, + { + "epoch": 0.67, + "learning_rate": 0.0002657572718154464, + "loss": 1.257, + "theoretical_loss": 3.4300141407093157, + "tokens_seen": 2215444480 + }, + { + "epoch": 0.67, + "learning_rate": 0.0002657412236710131, + "loss": 1.4735, + "theoretical_loss": 3.430005918954373, + "tokens_seen": 2215510016 + }, + { + "epoch": 0.67, + "learning_rate": 0.0002657251755265798, + "loss": 1.2623, + "theoretical_loss": 3.429997697510726, + "tokens_seen": 2215575552 + }, + { + "epoch": 0.67, + "learning_rate": 0.0002657091273821465, + "loss": 1.3391, + "theoretical_loss": 3.4299894763783527, + "tokens_seen": 2215641088 + }, + { + "epoch": 0.67, + "learning_rate": 0.0002656930792377132, + "loss": 1.2735, + "theoretical_loss": 3.429981255557233, + "tokens_seen": 2215706624 + }, + { + "epoch": 0.67, + "learning_rate": 0.0002656770310932799, + "loss": 1.2929, + "theoretical_loss": 3.429973035047345, + "tokens_seen": 2215772160 + }, + { + "epoch": 0.67, + "learning_rate": 0.0002656609829488466, + "loss": 1.3885, + "theoretical_loss": 3.4299648148486686, + "tokens_seen": 2215837696 + }, + { + "epoch": 0.67, + "learning_rate": 0.00026564493480441323, + "loss": 1.2842, + "theoretical_loss": 3.4299565949611828, + "tokens_seen": 2215903232 + }, + { + "epoch": 0.67, + "learning_rate": 0.00026562888665997993, + "loss": 1.3723, + "theoretical_loss": 3.429948375384866, + "tokens_seen": 2215968768 + }, + { + "epoch": 0.67, + "learning_rate": 0.00026561283851554663, + "loss": 1.241, + "theoretical_loss": 3.429940156119698, + "tokens_seen": 2216034304 + }, + { + "epoch": 0.67, + "learning_rate": 0.00026559679037111333, + "loss": 1.2456, + "theoretical_loss": 3.4299319371656565, + "tokens_seen": 2216099840 + }, + { + "epoch": 0.67, + "learning_rate": 0.00026558074222668003, + "loss": 1.2277, + "theoretical_loss": 3.429923718522722, + "tokens_seen": 2216165376 + }, + { + "epoch": 0.67, + "learning_rate": 0.00026556469408224674, + "loss": 1.2981, + "theoretical_loss": 3.429915500190873, + "tokens_seen": 2216230912 + }, + { + "epoch": 0.67, + "learning_rate": 0.00026554864593781344, + "loss": 1.3154, + "theoretical_loss": 3.429907282170088, + "tokens_seen": 2216296448 + }, + { + "epoch": 0.67, + "learning_rate": 0.00026553259779338014, + "loss": 1.3449, + "theoretical_loss": 3.4298990644603466, + "tokens_seen": 2216361984 + }, + { + "epoch": 0.67, + "learning_rate": 0.00026551654964894684, + "loss": 1.2707, + "theoretical_loss": 3.429890847061628, + "tokens_seen": 2216427520 + }, + { + "epoch": 0.67, + "learning_rate": 0.00026550050150451354, + "loss": 1.2761, + "theoretical_loss": 3.4298826299739105, + "tokens_seen": 2216493056 + }, + { + "epoch": 0.67, + "learning_rate": 0.00026548445336008024, + "loss": 1.2362, + "theoretical_loss": 3.4298744131971737, + "tokens_seen": 2216558592 + }, + { + "epoch": 0.67, + "learning_rate": 0.00026546840521564694, + "loss": 1.2833, + "theoretical_loss": 3.429866196731396, + "tokens_seen": 2216624128 + }, + { + "epoch": 0.67, + "learning_rate": 0.00026545235707121364, + "loss": 1.2193, + "theoretical_loss": 3.429857980576558, + "tokens_seen": 2216689664 + }, + { + "epoch": 0.67, + "learning_rate": 0.00026543630892678034, + "loss": 1.3187, + "theoretical_loss": 3.429849764732637, + "tokens_seen": 2216755200 + }, + { + "epoch": 0.67, + "learning_rate": 0.00026542026078234704, + "loss": 1.4318, + "theoretical_loss": 3.4298415491996126, + "tokens_seen": 2216820736 + }, + { + "epoch": 0.67, + "learning_rate": 0.00026540421263791374, + "loss": 1.2691, + "theoretical_loss": 3.4298333339774643, + "tokens_seen": 2216886272 + }, + { + "epoch": 0.67, + "learning_rate": 0.00026538816449348044, + "loss": 1.2556, + "theoretical_loss": 3.4298251190661704, + "tokens_seen": 2216951808 + }, + { + "epoch": 0.67, + "learning_rate": 0.00026537211634904714, + "loss": 1.2638, + "theoretical_loss": 3.429816904465711, + "tokens_seen": 2217017344 + }, + { + "epoch": 0.67, + "learning_rate": 0.00026535606820461384, + "loss": 1.3593, + "theoretical_loss": 3.429808690176064, + "tokens_seen": 2217082880 + }, + { + "epoch": 0.67, + "learning_rate": 0.00026534002006018054, + "loss": 1.2919, + "theoretical_loss": 3.429800476197209, + "tokens_seen": 2217148416 + }, + { + "epoch": 0.67, + "learning_rate": 0.00026532397191574724, + "loss": 1.358, + "theoretical_loss": 3.429792262529125, + "tokens_seen": 2217213952 + }, + { + "epoch": 0.67, + "learning_rate": 0.00026530792377131395, + "loss": 1.2874, + "theoretical_loss": 3.4297840491717912, + "tokens_seen": 2217279488 + }, + { + "epoch": 0.67, + "learning_rate": 0.00026529187562688065, + "loss": 1.262, + "theoretical_loss": 3.429775836125186, + "tokens_seen": 2217345024 + }, + { + "epoch": 0.67, + "learning_rate": 0.00026527582748244735, + "loss": 1.2836, + "theoretical_loss": 3.42976762338929, + "tokens_seen": 2217410560 + }, + { + "epoch": 0.67, + "learning_rate": 0.00026525977933801405, + "loss": 1.3446, + "theoretical_loss": 3.4297594109640803, + "tokens_seen": 2217476096 + }, + { + "epoch": 0.67, + "learning_rate": 0.00026524373119358075, + "loss": 1.2736, + "theoretical_loss": 3.4297511988495377, + "tokens_seen": 2217541632 + }, + { + "epoch": 0.67, + "learning_rate": 0.00026522768304914745, + "loss": 1.2476, + "theoretical_loss": 3.42974298704564, + "tokens_seen": 2217607168 + }, + { + "epoch": 0.67, + "learning_rate": 0.00026521163490471415, + "loss": 1.4326, + "theoretical_loss": 3.429734775552367, + "tokens_seen": 2217672704 + }, + { + "epoch": 0.67, + "learning_rate": 0.00026519558676028085, + "loss": 1.3233, + "theoretical_loss": 3.4297265643696973, + "tokens_seen": 2217738240 + }, + { + "epoch": 0.67, + "learning_rate": 0.00026517953861584755, + "loss": 1.355, + "theoretical_loss": 3.42971835349761, + "tokens_seen": 2217803776 + }, + { + "epoch": 0.67, + "learning_rate": 0.00026516349047141425, + "loss": 1.2443, + "theoretical_loss": 3.4297101429360852, + "tokens_seen": 2217869312 + }, + { + "epoch": 0.67, + "learning_rate": 0.00026514744232698095, + "loss": 1.2255, + "theoretical_loss": 3.4297019326851004, + "tokens_seen": 2217934848 + }, + { + "epoch": 0.67, + "learning_rate": 0.00026513139418254765, + "loss": 1.3645, + "theoretical_loss": 3.4296937227446356, + "tokens_seen": 2218000384 + }, + { + "epoch": 0.67, + "learning_rate": 0.00026511534603811435, + "loss": 1.3494, + "theoretical_loss": 3.4296855131146695, + "tokens_seen": 2218065920 + }, + { + "epoch": 0.67, + "learning_rate": 0.00026509929789368105, + "loss": 1.3167, + "theoretical_loss": 3.4296773037951818, + "tokens_seen": 2218131456 + }, + { + "epoch": 0.67, + "learning_rate": 0.00026508324974924775, + "loss": 1.3471, + "theoretical_loss": 3.429669094786151, + "tokens_seen": 2218196992 + }, + { + "epoch": 0.67, + "learning_rate": 0.00026506720160481446, + "loss": 1.2589, + "theoretical_loss": 3.4296608860875564, + "tokens_seen": 2218262528 + }, + { + "epoch": 0.67, + "learning_rate": 0.00026505115346038116, + "loss": 1.313, + "theoretical_loss": 3.429652677699377, + "tokens_seen": 2218328064 + }, + { + "epoch": 0.67, + "objective/train/docs_used": 807316, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.2874975204467773, + "objective/train/theoretical_loss": 3.429644469621592, + "objective/train/tokens_used": 2238853600, + "theoretical_loss": 3.429644469621592, + "tokens_seen": 2218393600 + }, + { + "epoch": 0.67, + "learning_rate": 0.00026503510531594786, + "loss": 1.2735, + "theoretical_loss": 3.429644469621592, + "tokens_seen": 2218393600 + }, + { + "epoch": 0.67, + "learning_rate": 0.00026501905717151456, + "loss": 1.2656, + "theoretical_loss": 3.4296362618541805, + "tokens_seen": 2218459136 + }, + { + "epoch": 0.67, + "learning_rate": 0.00026500300902708126, + "loss": 1.2119, + "theoretical_loss": 3.429628054397121, + "tokens_seen": 2218524672 + }, + { + "epoch": 0.67, + "learning_rate": 0.00026498696088264796, + "loss": 1.277, + "theoretical_loss": 3.429619847250394, + "tokens_seen": 2218590208 + }, + { + "epoch": 0.67, + "learning_rate": 0.00026497091273821466, + "loss": 1.3779, + "theoretical_loss": 3.4296116404139774, + "tokens_seen": 2218655744 + }, + { + "epoch": 0.67, + "learning_rate": 0.00026495486459378136, + "loss": 1.339, + "theoretical_loss": 3.4296034338878503, + "tokens_seen": 2218721280 + }, + { + "epoch": 0.67, + "learning_rate": 0.00026493881644934806, + "loss": 1.3713, + "theoretical_loss": 3.4295952276719923, + "tokens_seen": 2218786816 + }, + { + "epoch": 0.67, + "learning_rate": 0.00026492276830491476, + "loss": 1.234, + "theoretical_loss": 3.4295870217663826, + "tokens_seen": 2218852352 + }, + { + "epoch": 0.67, + "learning_rate": 0.00026490672016048146, + "loss": 1.2791, + "theoretical_loss": 3.4295788161709995, + "tokens_seen": 2218917888 + }, + { + "epoch": 0.67, + "learning_rate": 0.00026489067201604816, + "loss": 1.2339, + "theoretical_loss": 3.4295706108858233, + "tokens_seen": 2218983424 + }, + { + "epoch": 0.67, + "learning_rate": 0.0002648746238716148, + "loss": 1.3069, + "theoretical_loss": 3.4295624059108323, + "tokens_seen": 2219048960 + }, + { + "epoch": 0.67, + "learning_rate": 0.00026485857572718156, + "loss": 1.3025, + "theoretical_loss": 3.4295542012460056, + "tokens_seen": 2219114496 + }, + { + "epoch": 0.67, + "learning_rate": 0.00026484252758274826, + "loss": 1.2072, + "theoretical_loss": 3.4295459968913224, + "tokens_seen": 2219180032 + }, + { + "epoch": 0.67, + "learning_rate": 0.00026482647943831496, + "loss": 1.314, + "theoretical_loss": 3.4295377928467623, + "tokens_seen": 2219245568 + }, + { + "epoch": 0.67, + "learning_rate": 0.00026481043129388167, + "loss": 1.3669, + "theoretical_loss": 3.429529589112304, + "tokens_seen": 2219311104 + }, + { + "epoch": 0.67, + "learning_rate": 0.00026479438314944837, + "loss": 1.3348, + "theoretical_loss": 3.4295213856879263, + "tokens_seen": 2219376640 + }, + { + "epoch": 0.67, + "learning_rate": 0.00026477833500501507, + "loss": 1.3413, + "theoretical_loss": 3.4295131825736087, + "tokens_seen": 2219442176 + }, + { + "epoch": 0.67, + "learning_rate": 0.00026476228686058177, + "loss": 1.3693, + "theoretical_loss": 3.4295049797693307, + "tokens_seen": 2219507712 + }, + { + "epoch": 0.67, + "learning_rate": 0.00026474623871614847, + "loss": 1.2439, + "theoretical_loss": 3.429496777275071, + "tokens_seen": 2219573248 + }, + { + "epoch": 0.67, + "learning_rate": 0.00026473019057171517, + "loss": 1.2861, + "theoretical_loss": 3.4294885750908084, + "tokens_seen": 2219638784 + }, + { + "epoch": 0.67, + "learning_rate": 0.00026471414242728187, + "loss": 1.2114, + "theoretical_loss": 3.4294803732165224, + "tokens_seen": 2219704320 + }, + { + "epoch": 0.67, + "learning_rate": 0.00026469809428284857, + "loss": 1.2528, + "theoretical_loss": 3.4294721716521925, + "tokens_seen": 2219769856 + }, + { + "epoch": 0.67, + "learning_rate": 0.00026468204613841527, + "loss": 1.3568, + "theoretical_loss": 3.4294639703977974, + "tokens_seen": 2219835392 + }, + { + "epoch": 0.67, + "learning_rate": 0.00026466599799398197, + "loss": 1.2863, + "theoretical_loss": 3.429455769453316, + "tokens_seen": 2219900928 + }, + { + "epoch": 0.67, + "learning_rate": 0.00026464994984954867, + "loss": 1.2309, + "theoretical_loss": 3.4294475688187283, + "tokens_seen": 2219966464 + }, + { + "epoch": 0.67, + "learning_rate": 0.00026463390170511537, + "loss": 1.2306, + "theoretical_loss": 3.4294393684940125, + "tokens_seen": 2220032000 + }, + { + "epoch": 0.67, + "learning_rate": 0.0002646178535606821, + "loss": 1.2873, + "theoretical_loss": 3.429431168479148, + "tokens_seen": 2220097536 + }, + { + "epoch": 0.67, + "learning_rate": 0.0002646018054162488, + "loss": 1.2096, + "theoretical_loss": 3.4294229687741145, + "tokens_seen": 2220163072 + }, + { + "epoch": 0.67, + "learning_rate": 0.0002645857572718155, + "loss": 1.2743, + "theoretical_loss": 3.4294147693788903, + "tokens_seen": 2220228608 + }, + { + "epoch": 0.67, + "learning_rate": 0.0002645697091273822, + "loss": 1.3309, + "theoretical_loss": 3.4294065702934553, + "tokens_seen": 2220294144 + }, + { + "epoch": 0.67, + "learning_rate": 0.0002645536609829489, + "loss": 1.3714, + "theoretical_loss": 3.429398371517788, + "tokens_seen": 2220359680 + }, + { + "epoch": 0.67, + "learning_rate": 0.0002645376128385156, + "loss": 1.3718, + "theoretical_loss": 3.4293901730518686, + "tokens_seen": 2220425216 + }, + { + "epoch": 0.67, + "learning_rate": 0.0002645215646940823, + "loss": 1.2679, + "theoretical_loss": 3.4293819748956746, + "tokens_seen": 2220490752 + }, + { + "epoch": 0.67, + "learning_rate": 0.000264505516549649, + "loss": 1.2446, + "theoretical_loss": 3.4293737770491868, + "tokens_seen": 2220556288 + }, + { + "epoch": 0.67, + "learning_rate": 0.0002644894684052157, + "loss": 1.3725, + "theoretical_loss": 3.429365579512383, + "tokens_seen": 2220621824 + }, + { + "epoch": 0.67, + "learning_rate": 0.0002644734202607824, + "loss": 1.2618, + "theoretical_loss": 3.429357382285244, + "tokens_seen": 2220687360 + }, + { + "epoch": 0.67, + "learning_rate": 0.0002644573721163491, + "loss": 1.3051, + "theoretical_loss": 3.429349185367747, + "tokens_seen": 2220752896 + }, + { + "epoch": 0.67, + "learning_rate": 0.0002644413239719158, + "loss": 1.3796, + "theoretical_loss": 3.429340988759873, + "tokens_seen": 2220818432 + }, + { + "epoch": 0.67, + "learning_rate": 0.0002644252758274825, + "loss": 1.3, + "theoretical_loss": 3.4293327924615995, + "tokens_seen": 2220883968 + }, + { + "epoch": 0.67, + "learning_rate": 0.0002644092276830492, + "loss": 1.228, + "theoretical_loss": 3.429324596472907, + "tokens_seen": 2220949504 + }, + { + "epoch": 0.67, + "learning_rate": 0.0002643931795386159, + "loss": 1.3173, + "theoretical_loss": 3.429316400793774, + "tokens_seen": 2221015040 + }, + { + "epoch": 0.67, + "learning_rate": 0.0002643771313941826, + "loss": 1.2599, + "theoretical_loss": 3.4293082054241797, + "tokens_seen": 2221080576 + }, + { + "epoch": 0.67, + "learning_rate": 0.0002643610832497493, + "loss": 1.3087, + "theoretical_loss": 3.429300010364104, + "tokens_seen": 2221146112 + }, + { + "epoch": 0.67, + "learning_rate": 0.000264345035105316, + "loss": 1.2481, + "theoretical_loss": 3.429291815613525, + "tokens_seen": 2221211648 + }, + { + "epoch": 0.67, + "learning_rate": 0.0002643289869608827, + "loss": 1.2315, + "theoretical_loss": 3.4292836211724222, + "tokens_seen": 2221277184 + }, + { + "epoch": 0.67, + "learning_rate": 0.0002643129388164494, + "loss": 1.2883, + "theoretical_loss": 3.4292754270407753, + "tokens_seen": 2221342720 + }, + { + "epoch": 0.67, + "learning_rate": 0.0002642968906720161, + "loss": 1.2647, + "theoretical_loss": 3.429267233218563, + "tokens_seen": 2221408256 + }, + { + "epoch": 0.67, + "learning_rate": 0.0002642808425275828, + "loss": 1.2643, + "theoretical_loss": 3.4292590397057645, + "tokens_seen": 2221473792 + }, + { + "epoch": 0.67, + "learning_rate": 0.0002642647943831495, + "loss": 1.2515, + "theoretical_loss": 3.4292508465023595, + "tokens_seen": 2221539328 + }, + { + "epoch": 0.67, + "learning_rate": 0.0002642487462387162, + "loss": 1.377, + "theoretical_loss": 3.4292426536083265, + "tokens_seen": 2221604864 + }, + { + "epoch": 0.67, + "objective/train/docs_used": 808517, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.314875602722168, + "objective/train/theoretical_loss": 3.429234461023645, + "objective/train/tokens_used": 2242130400, + "theoretical_loss": 3.429234461023645, + "tokens_seen": 2221670400 + }, + { + "epoch": 0.67, + "learning_rate": 0.00026423269809428283, + "loss": 1.221, + "theoretical_loss": 3.429234461023645, + "tokens_seen": 2221670400 + }, + { + "epoch": 0.67, + "learning_rate": 0.00026421664994984953, + "loss": 1.2545, + "theoretical_loss": 3.429226268748294, + "tokens_seen": 2221735936 + }, + { + "epoch": 0.67, + "learning_rate": 0.00026420060180541624, + "loss": 1.337, + "theoretical_loss": 3.4292180767822535, + "tokens_seen": 2221801472 + }, + { + "epoch": 0.67, + "learning_rate": 0.00026418455366098294, + "loss": 1.3064, + "theoretical_loss": 3.429209885125502, + "tokens_seen": 2221867008 + }, + { + "epoch": 0.67, + "learning_rate": 0.00026416850551654964, + "loss": 1.3554, + "theoretical_loss": 3.429201693778018, + "tokens_seen": 2221932544 + }, + { + "epoch": 0.67, + "learning_rate": 0.00026415245737211634, + "loss": 1.3507, + "theoretical_loss": 3.429193502739782, + "tokens_seen": 2221998080 + }, + { + "epoch": 0.67, + "learning_rate": 0.00026413640922768304, + "loss": 1.4184, + "theoretical_loss": 3.4291853120107727, + "tokens_seen": 2222063616 + }, + { + "epoch": 0.67, + "learning_rate": 0.00026412036108324974, + "loss": 1.2474, + "theoretical_loss": 3.4291771215909694, + "tokens_seen": 2222129152 + }, + { + "epoch": 0.67, + "learning_rate": 0.00026410431293881644, + "loss": 1.3986, + "theoretical_loss": 3.4291689314803513, + "tokens_seen": 2222194688 + }, + { + "epoch": 0.67, + "learning_rate": 0.00026408826479438314, + "loss": 1.3568, + "theoretical_loss": 3.4291607416788974, + "tokens_seen": 2222260224 + }, + { + "epoch": 0.67, + "learning_rate": 0.00026407221664994984, + "loss": 1.3447, + "theoretical_loss": 3.429152552186587, + "tokens_seen": 2222325760 + }, + { + "epoch": 0.67, + "learning_rate": 0.00026405616850551654, + "loss": 1.3435, + "theoretical_loss": 3.4291443630033993, + "tokens_seen": 2222391296 + }, + { + "epoch": 0.67, + "learning_rate": 0.00026404012036108324, + "loss": 1.323, + "theoretical_loss": 3.4291361741293134, + "tokens_seen": 2222456832 + }, + { + "epoch": 0.67, + "learning_rate": 0.00026402407221664994, + "loss": 1.1774, + "theoretical_loss": 3.4291279855643086, + "tokens_seen": 2222522368 + }, + { + "epoch": 0.67, + "learning_rate": 0.00026400802407221664, + "loss": 1.4558, + "theoretical_loss": 3.4291197973083647, + "tokens_seen": 2222587904 + }, + { + "epoch": 0.67, + "learning_rate": 0.00026399197592778334, + "loss": 1.224, + "theoretical_loss": 3.42911160936146, + "tokens_seen": 2222653440 + }, + { + "epoch": 0.67, + "learning_rate": 0.00026397592778335004, + "loss": 1.3279, + "theoretical_loss": 3.4291034217235747, + "tokens_seen": 2222718976 + }, + { + "epoch": 0.67, + "learning_rate": 0.00026395987963891674, + "loss": 1.257, + "theoretical_loss": 3.429095234394687, + "tokens_seen": 2222784512 + }, + { + "epoch": 0.67, + "learning_rate": 0.00026394383149448345, + "loss": 1.3082, + "theoretical_loss": 3.4290870473747765, + "tokens_seen": 2222850048 + }, + { + "epoch": 0.67, + "learning_rate": 0.00026392778335005015, + "loss": 1.2988, + "theoretical_loss": 3.429078860663823, + "tokens_seen": 2222915584 + }, + { + "epoch": 0.67, + "learning_rate": 0.00026391173520561685, + "loss": 1.3099, + "theoretical_loss": 3.429070674261805, + "tokens_seen": 2222981120 + }, + { + "epoch": 0.67, + "learning_rate": 0.00026389568706118355, + "loss": 1.2105, + "theoretical_loss": 3.429062488168702, + "tokens_seen": 2223046656 + }, + { + "epoch": 0.67, + "learning_rate": 0.00026387963891675025, + "loss": 1.3071, + "theoretical_loss": 3.4290543023844933, + "tokens_seen": 2223112192 + }, + { + "epoch": 0.67, + "learning_rate": 0.00026386359077231695, + "loss": 1.361, + "theoretical_loss": 3.429046116909158, + "tokens_seen": 2223177728 + }, + { + "epoch": 0.67, + "learning_rate": 0.00026384754262788365, + "loss": 1.2616, + "theoretical_loss": 3.429037931742675, + "tokens_seen": 2223243264 + }, + { + "epoch": 0.67, + "learning_rate": 0.00026383149448345035, + "loss": 1.392, + "theoretical_loss": 3.429029746885025, + "tokens_seen": 2223308800 + }, + { + "epoch": 0.67, + "learning_rate": 0.00026381544633901705, + "loss": 1.3727, + "theoretical_loss": 3.4290215623361853, + "tokens_seen": 2223374336 + }, + { + "epoch": 0.67, + "learning_rate": 0.00026379939819458375, + "loss": 1.2876, + "theoretical_loss": 3.4290133780961365, + "tokens_seen": 2223439872 + }, + { + "epoch": 0.67, + "learning_rate": 0.00026378335005015045, + "loss": 1.3828, + "theoretical_loss": 3.429005194164857, + "tokens_seen": 2223505408 + }, + { + "epoch": 0.67, + "learning_rate": 0.00026376730190571715, + "loss": 1.2663, + "theoretical_loss": 3.428997010542327, + "tokens_seen": 2223570944 + }, + { + "epoch": 0.67, + "learning_rate": 0.00026375125376128385, + "loss": 1.4139, + "theoretical_loss": 3.4289888272285247, + "tokens_seen": 2223636480 + }, + { + "epoch": 0.67, + "learning_rate": 0.00026373520561685055, + "loss": 1.3521, + "theoretical_loss": 3.4289806442234303, + "tokens_seen": 2223702016 + }, + { + "epoch": 0.67, + "learning_rate": 0.00026371915747241725, + "loss": 1.2288, + "theoretical_loss": 3.4289724615270223, + "tokens_seen": 2223767552 + }, + { + "epoch": 0.67, + "learning_rate": 0.00026370310932798395, + "loss": 1.3772, + "theoretical_loss": 3.4289642791392803, + "tokens_seen": 2223833088 + }, + { + "epoch": 0.67, + "learning_rate": 0.00026368706118355066, + "loss": 1.4291, + "theoretical_loss": 3.428956097060184, + "tokens_seen": 2223898624 + }, + { + "epoch": 0.67, + "learning_rate": 0.00026367101303911736, + "loss": 1.3638, + "theoretical_loss": 3.4289479152897115, + "tokens_seen": 2223964160 + }, + { + "epoch": 0.67, + "learning_rate": 0.0002636549648946841, + "loss": 1.3226, + "theoretical_loss": 3.428939733827843, + "tokens_seen": 2224029696 + }, + { + "epoch": 0.67, + "learning_rate": 0.0002636389167502508, + "loss": 1.3292, + "theoretical_loss": 3.4289315526745576, + "tokens_seen": 2224095232 + }, + { + "epoch": 0.67, + "learning_rate": 0.0002636228686058175, + "loss": 1.3207, + "theoretical_loss": 3.4289233718298346, + "tokens_seen": 2224160768 + }, + { + "epoch": 0.67, + "learning_rate": 0.00026360682046138416, + "loss": 1.3262, + "theoretical_loss": 3.428915191293653, + "tokens_seen": 2224226304 + }, + { + "epoch": 0.67, + "learning_rate": 0.00026359077231695086, + "loss": 1.4267, + "theoretical_loss": 3.4289070110659923, + "tokens_seen": 2224291840 + }, + { + "epoch": 0.67, + "learning_rate": 0.00026357472417251756, + "loss": 1.3091, + "theoretical_loss": 3.428898831146832, + "tokens_seen": 2224357376 + }, + { + "epoch": 0.67, + "learning_rate": 0.00026355867602808426, + "loss": 1.3423, + "theoretical_loss": 3.428890651536151, + "tokens_seen": 2224422912 + }, + { + "epoch": 0.67, + "learning_rate": 0.00026354262788365096, + "loss": 1.337, + "theoretical_loss": 3.428882472233928, + "tokens_seen": 2224488448 + }, + { + "epoch": 0.67, + "learning_rate": 0.00026352657973921766, + "loss": 1.3569, + "theoretical_loss": 3.4288742932401437, + "tokens_seen": 2224553984 + }, + { + "epoch": 0.67, + "learning_rate": 0.00026351053159478436, + "loss": 1.3281, + "theoretical_loss": 3.4288661145547765, + "tokens_seen": 2224619520 + }, + { + "epoch": 0.67, + "learning_rate": 0.00026349448345035106, + "loss": 1.2776, + "theoretical_loss": 3.4288579361778053, + "tokens_seen": 2224685056 + }, + { + "epoch": 0.67, + "learning_rate": 0.00026347843530591776, + "loss": 1.3463, + "theoretical_loss": 3.4288497581092106, + "tokens_seen": 2224750592 + }, + { + "epoch": 0.67, + "learning_rate": 0.00026346238716148446, + "loss": 1.2549, + "theoretical_loss": 3.428841580348971, + "tokens_seen": 2224816128 + }, + { + "epoch": 0.67, + "learning_rate": 0.00026344633901705117, + "loss": 1.2465, + "theoretical_loss": 3.428833402897065, + "tokens_seen": 2224881664 + }, + { + "epoch": 0.67, + "objective/train/docs_used": 809850, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.4253478050231934, + "objective/train/theoretical_loss": 3.4288252257534735, + "objective/train/tokens_used": 2245407200, + "theoretical_loss": 3.4288252257534735, + "tokens_seen": 2224947200 + }, + { + "epoch": 0.67, + "learning_rate": 0.00026343029087261787, + "loss": 1.3274, + "theoretical_loss": 3.4288252257534735, + "tokens_seen": 2224947200 + }, + { + "epoch": 0.67, + "learning_rate": 0.00026341424272818457, + "loss": 1.239, + "theoretical_loss": 3.4288170489181744, + "tokens_seen": 2225012736 + }, + { + "epoch": 0.67, + "learning_rate": 0.00026339819458375127, + "loss": 1.2964, + "theoretical_loss": 3.428808872391148, + "tokens_seen": 2225078272 + }, + { + "epoch": 0.67, + "learning_rate": 0.00026338214643931797, + "loss": 1.2297, + "theoretical_loss": 3.428800696172373, + "tokens_seen": 2225143808 + }, + { + "epoch": 0.67, + "learning_rate": 0.00026336609829488467, + "loss": 1.2699, + "theoretical_loss": 3.428792520261829, + "tokens_seen": 2225209344 + }, + { + "epoch": 0.67, + "learning_rate": 0.00026335005015045137, + "loss": 1.2434, + "theoretical_loss": 3.428784344659495, + "tokens_seen": 2225274880 + }, + { + "epoch": 0.67, + "learning_rate": 0.00026333400200601807, + "loss": 1.195, + "theoretical_loss": 3.4287761693653502, + "tokens_seen": 2225340416 + }, + { + "epoch": 0.67, + "learning_rate": 0.00026331795386158477, + "loss": 1.424, + "theoretical_loss": 3.4287679943793745, + "tokens_seen": 2225405952 + }, + { + "epoch": 0.67, + "learning_rate": 0.00026330190571715147, + "loss": 1.3317, + "theoretical_loss": 3.428759819701547, + "tokens_seen": 2225471488 + }, + { + "epoch": 0.67, + "learning_rate": 0.00026328585757271817, + "loss": 1.3205, + "theoretical_loss": 3.4287516453318467, + "tokens_seen": 2225537024 + }, + { + "epoch": 0.67, + "learning_rate": 0.00026326980942828487, + "loss": 1.2546, + "theoretical_loss": 3.4287434712702534, + "tokens_seen": 2225602560 + }, + { + "epoch": 0.67, + "learning_rate": 0.00026325376128385157, + "loss": 1.3094, + "theoretical_loss": 3.428735297516746, + "tokens_seen": 2225668096 + }, + { + "epoch": 0.67, + "learning_rate": 0.0002632377131394183, + "loss": 1.3093, + "theoretical_loss": 3.4287271240713038, + "tokens_seen": 2225733632 + }, + { + "epoch": 0.67, + "learning_rate": 0.000263221664994985, + "loss": 1.2486, + "theoretical_loss": 3.4287189509339067, + "tokens_seen": 2225799168 + }, + { + "epoch": 0.67, + "learning_rate": 0.0002632056168505517, + "loss": 1.2986, + "theoretical_loss": 3.428710778104533, + "tokens_seen": 2225864704 + }, + { + "epoch": 0.67, + "learning_rate": 0.0002631895687061184, + "loss": 1.2423, + "theoretical_loss": 3.4287026055831626, + "tokens_seen": 2225930240 + }, + { + "epoch": 0.67, + "learning_rate": 0.0002631735205616851, + "loss": 1.336, + "theoretical_loss": 3.4286944333697753, + "tokens_seen": 2225995776 + }, + { + "epoch": 0.67, + "learning_rate": 0.0002631574724172518, + "loss": 1.241, + "theoretical_loss": 3.4286862614643496, + "tokens_seen": 2226061312 + }, + { + "epoch": 0.67, + "learning_rate": 0.0002631414242728185, + "loss": 1.2694, + "theoretical_loss": 3.4286780898668656, + "tokens_seen": 2226126848 + }, + { + "epoch": 0.67, + "learning_rate": 0.0002631253761283852, + "loss": 1.2614, + "theoretical_loss": 3.428669918577302, + "tokens_seen": 2226192384 + }, + { + "epoch": 0.67, + "learning_rate": 0.0002631093279839519, + "loss": 1.1958, + "theoretical_loss": 3.4286617475956382, + "tokens_seen": 2226257920 + }, + { + "epoch": 0.67, + "learning_rate": 0.0002630932798395186, + "loss": 1.2368, + "theoretical_loss": 3.428653576921854, + "tokens_seen": 2226323456 + }, + { + "epoch": 0.67, + "learning_rate": 0.0002630772316950853, + "loss": 1.3618, + "theoretical_loss": 3.428645406555928, + "tokens_seen": 2226388992 + }, + { + "epoch": 0.67, + "learning_rate": 0.000263061183550652, + "loss": 1.3576, + "theoretical_loss": 3.42863723649784, + "tokens_seen": 2226454528 + }, + { + "epoch": 0.67, + "learning_rate": 0.0002630451354062187, + "loss": 1.2695, + "theoretical_loss": 3.42862906674757, + "tokens_seen": 2226520064 + }, + { + "epoch": 0.67, + "learning_rate": 0.0002630290872617854, + "loss": 1.3773, + "theoretical_loss": 3.428620897305096, + "tokens_seen": 2226585600 + }, + { + "epoch": 0.67, + "learning_rate": 0.0002630130391173521, + "loss": 1.3825, + "theoretical_loss": 3.4286127281703984, + "tokens_seen": 2226651136 + }, + { + "epoch": 0.67, + "learning_rate": 0.0002629969909729188, + "loss": 1.336, + "theoretical_loss": 3.428604559343456, + "tokens_seen": 2226716672 + }, + { + "epoch": 0.67, + "learning_rate": 0.0002629809428284855, + "loss": 1.2909, + "theoretical_loss": 3.4285963908242483, + "tokens_seen": 2226782208 + }, + { + "epoch": 0.67, + "learning_rate": 0.0002629648946840522, + "loss": 1.3489, + "theoretical_loss": 3.4285882226127544, + "tokens_seen": 2226847744 + }, + { + "epoch": 0.67, + "learning_rate": 0.0002629488465396189, + "loss": 1.3535, + "theoretical_loss": 3.428580054708954, + "tokens_seen": 2226913280 + }, + { + "epoch": 0.67, + "learning_rate": 0.0002629327983951856, + "loss": 1.3449, + "theoretical_loss": 3.428571887112826, + "tokens_seen": 2226978816 + }, + { + "epoch": 0.67, + "learning_rate": 0.0002629167502507523, + "loss": 1.2463, + "theoretical_loss": 3.428563719824351, + "tokens_seen": 2227044352 + }, + { + "epoch": 0.67, + "learning_rate": 0.000262900702106319, + "loss": 1.3818, + "theoretical_loss": 3.4285555528435068, + "tokens_seen": 2227109888 + }, + { + "epoch": 0.67, + "learning_rate": 0.0002628846539618857, + "loss": 1.2785, + "theoretical_loss": 3.4285473861702735, + "tokens_seen": 2227175424 + }, + { + "epoch": 0.67, + "learning_rate": 0.0002628686058174524, + "loss": 1.3651, + "theoretical_loss": 3.4285392198046303, + "tokens_seen": 2227240960 + }, + { + "epoch": 0.67, + "learning_rate": 0.0002628525576730191, + "loss": 1.4074, + "theoretical_loss": 3.428531053746557, + "tokens_seen": 2227306496 + }, + { + "epoch": 0.67, + "learning_rate": 0.00026283650952858573, + "loss": 1.3044, + "theoretical_loss": 3.4285228879960323, + "tokens_seen": 2227372032 + }, + { + "epoch": 0.67, + "learning_rate": 0.00026282046138415244, + "loss": 1.3765, + "theoretical_loss": 3.428514722553036, + "tokens_seen": 2227437568 + }, + { + "epoch": 0.68, + "learning_rate": 0.00026280441323971914, + "loss": 1.3973, + "theoretical_loss": 3.4285065574175473, + "tokens_seen": 2227503104 + }, + { + "epoch": 0.68, + "learning_rate": 0.00026278836509528584, + "loss": 1.2112, + "theoretical_loss": 3.428498392589546, + "tokens_seen": 2227568640 + }, + { + "epoch": 0.68, + "learning_rate": 0.00026277231695085254, + "loss": 1.3185, + "theoretical_loss": 3.428490228069011, + "tokens_seen": 2227634176 + }, + { + "epoch": 0.68, + "learning_rate": 0.00026275626880641924, + "loss": 1.2937, + "theoretical_loss": 3.4284820638559212, + "tokens_seen": 2227699712 + }, + { + "epoch": 0.68, + "learning_rate": 0.00026274022066198594, + "loss": 1.3159, + "theoretical_loss": 3.428473899950257, + "tokens_seen": 2227765248 + }, + { + "epoch": 0.68, + "learning_rate": 0.00026272417251755264, + "loss": 1.3136, + "theoretical_loss": 3.4284657363519973, + "tokens_seen": 2227830784 + }, + { + "epoch": 0.68, + "learning_rate": 0.00026270812437311934, + "loss": 1.2525, + "theoretical_loss": 3.4284575730611215, + "tokens_seen": 2227896320 + }, + { + "epoch": 0.68, + "learning_rate": 0.00026269207622868604, + "loss": 1.2919, + "theoretical_loss": 3.428449410077609, + "tokens_seen": 2227961856 + }, + { + "epoch": 0.68, + "learning_rate": 0.00026267602808425274, + "loss": 1.2697, + "theoretical_loss": 3.428441247401439, + "tokens_seen": 2228027392 + }, + { + "epoch": 0.68, + "learning_rate": 0.00026265997993981944, + "loss": 1.4312, + "theoretical_loss": 3.4284330850325917, + "tokens_seen": 2228092928 + }, + { + "epoch": 0.68, + "learning_rate": 0.00026264393179538614, + "loss": 1.41, + "theoretical_loss": 3.428424922971045, + "tokens_seen": 2228158464 + }, + { + "epoch": 0.68, + "objective/train/docs_used": 811140, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.2091569900512695, + "objective/train/theoretical_loss": 3.4284167612167797, + "objective/train/tokens_used": 2248684000, + "theoretical_loss": 3.4284167612167797, + "tokens_seen": 2228224000 + }, + { + "epoch": 0.68, + "learning_rate": 0.00026262788365095284, + "loss": 1.1977, + "theoretical_loss": 3.4284167612167797, + "tokens_seen": 2228224000 + }, + { + "epoch": 0.68, + "learning_rate": 0.00026261183550651954, + "loss": 1.3358, + "theoretical_loss": 3.428408599769775, + "tokens_seen": 2228289536 + }, + { + "epoch": 0.68, + "learning_rate": 0.00026259578736208624, + "loss": 1.3448, + "theoretical_loss": 3.428400438630009, + "tokens_seen": 2228355072 + }, + { + "epoch": 0.68, + "learning_rate": 0.00026257973921765295, + "loss": 1.3821, + "theoretical_loss": 3.428392277797463, + "tokens_seen": 2228420608 + }, + { + "epoch": 0.68, + "learning_rate": 0.00026256369107321965, + "loss": 1.178, + "theoretical_loss": 3.428384117272115, + "tokens_seen": 2228486144 + }, + { + "epoch": 0.68, + "learning_rate": 0.00026254764292878635, + "loss": 1.3941, + "theoretical_loss": 3.428375957053945, + "tokens_seen": 2228551680 + }, + { + "epoch": 0.68, + "learning_rate": 0.00026253159478435305, + "loss": 1.2455, + "theoretical_loss": 3.428367797142932, + "tokens_seen": 2228617216 + }, + { + "epoch": 0.68, + "learning_rate": 0.00026251554663991975, + "loss": 1.2906, + "theoretical_loss": 3.4283596375390557, + "tokens_seen": 2228682752 + }, + { + "epoch": 0.68, + "learning_rate": 0.00026249949849548645, + "loss": 1.3624, + "theoretical_loss": 3.4283514782422957, + "tokens_seen": 2228748288 + }, + { + "epoch": 0.68, + "learning_rate": 0.00026248345035105315, + "loss": 1.4029, + "theoretical_loss": 3.4283433192526314, + "tokens_seen": 2228813824 + }, + { + "epoch": 0.68, + "learning_rate": 0.00026246740220661985, + "loss": 1.2905, + "theoretical_loss": 3.4283351605700414, + "tokens_seen": 2228879360 + }, + { + "epoch": 0.68, + "learning_rate": 0.0002624513540621866, + "loss": 1.4001, + "theoretical_loss": 3.4283270021945063, + "tokens_seen": 2228944896 + }, + { + "epoch": 0.68, + "learning_rate": 0.0002624353059177533, + "loss": 1.34, + "theoretical_loss": 3.4283188441260046, + "tokens_seen": 2229010432 + }, + { + "epoch": 0.68, + "learning_rate": 0.00026241925777332, + "loss": 1.3352, + "theoretical_loss": 3.428310686364516, + "tokens_seen": 2229075968 + }, + { + "epoch": 0.68, + "learning_rate": 0.0002624032096288867, + "loss": 1.3156, + "theoretical_loss": 3.4283025289100197, + "tokens_seen": 2229141504 + }, + { + "epoch": 0.68, + "learning_rate": 0.0002623871614844534, + "loss": 1.4266, + "theoretical_loss": 3.428294371762496, + "tokens_seen": 2229207040 + }, + { + "epoch": 0.68, + "learning_rate": 0.0002623711133400201, + "loss": 1.3161, + "theoretical_loss": 3.4282862149219233, + "tokens_seen": 2229272576 + }, + { + "epoch": 0.68, + "learning_rate": 0.0002623550651955868, + "loss": 1.2406, + "theoretical_loss": 3.4282780583882815, + "tokens_seen": 2229338112 + }, + { + "epoch": 0.68, + "learning_rate": 0.0002623390170511535, + "loss": 1.2919, + "theoretical_loss": 3.42826990216155, + "tokens_seen": 2229403648 + }, + { + "epoch": 0.68, + "learning_rate": 0.0002623229689067202, + "loss": 1.4146, + "theoretical_loss": 3.4282617462417084, + "tokens_seen": 2229469184 + }, + { + "epoch": 0.68, + "learning_rate": 0.0002623069207622869, + "loss": 1.3675, + "theoretical_loss": 3.428253590628736, + "tokens_seen": 2229534720 + }, + { + "epoch": 0.68, + "learning_rate": 0.0002622908726178536, + "loss": 1.3554, + "theoretical_loss": 3.428245435322612, + "tokens_seen": 2229600256 + }, + { + "epoch": 0.68, + "learning_rate": 0.0002622748244734203, + "loss": 1.3384, + "theoretical_loss": 3.428237280323316, + "tokens_seen": 2229665792 + }, + { + "epoch": 0.68, + "learning_rate": 0.000262258776328987, + "loss": 1.3905, + "theoretical_loss": 3.4282291256308275, + "tokens_seen": 2229731328 + }, + { + "epoch": 0.68, + "learning_rate": 0.0002622427281845537, + "loss": 1.3092, + "theoretical_loss": 3.4282209712451257, + "tokens_seen": 2229796864 + }, + { + "epoch": 0.68, + "learning_rate": 0.0002622266800401204, + "loss": 1.2396, + "theoretical_loss": 3.42821281716619, + "tokens_seen": 2229862400 + }, + { + "epoch": 0.68, + "learning_rate": 0.00026221063189568706, + "loss": 1.3795, + "theoretical_loss": 3.4282046633940007, + "tokens_seen": 2229927936 + }, + { + "epoch": 0.68, + "learning_rate": 0.00026219458375125376, + "loss": 1.3691, + "theoretical_loss": 3.428196509928536, + "tokens_seen": 2229993472 + }, + { + "epoch": 0.68, + "learning_rate": 0.00026217853560682046, + "loss": 1.2488, + "theoretical_loss": 3.4281883567697764, + "tokens_seen": 2230059008 + }, + { + "epoch": 0.68, + "learning_rate": 0.00026216248746238716, + "loss": 1.3613, + "theoretical_loss": 3.4281802039177007, + "tokens_seen": 2230124544 + }, + { + "epoch": 0.68, + "learning_rate": 0.00026214643931795386, + "loss": 1.2646, + "theoretical_loss": 3.4281720513722886, + "tokens_seen": 2230190080 + }, + { + "epoch": 0.68, + "learning_rate": 0.00026213039117352056, + "loss": 1.35, + "theoretical_loss": 3.4281638991335197, + "tokens_seen": 2230255616 + }, + { + "epoch": 0.68, + "learning_rate": 0.00026211434302908726, + "loss": 1.2296, + "theoretical_loss": 3.428155747201373, + "tokens_seen": 2230321152 + }, + { + "epoch": 0.68, + "learning_rate": 0.00026209829488465396, + "loss": 1.2889, + "theoretical_loss": 3.4281475955758287, + "tokens_seen": 2230386688 + }, + { + "epoch": 0.68, + "learning_rate": 0.00026208224674022066, + "loss": 1.2649, + "theoretical_loss": 3.4281394442568653, + "tokens_seen": 2230452224 + }, + { + "epoch": 0.68, + "learning_rate": 0.00026206619859578737, + "loss": 1.3402, + "theoretical_loss": 3.428131293244463, + "tokens_seen": 2230517760 + }, + { + "epoch": 0.68, + "learning_rate": 0.00026205015045135407, + "loss": 1.2776, + "theoretical_loss": 3.4281231425386007, + "tokens_seen": 2230583296 + }, + { + "epoch": 0.68, + "learning_rate": 0.00026203410230692077, + "loss": 1.2316, + "theoretical_loss": 3.4281149921392586, + "tokens_seen": 2230648832 + }, + { + "epoch": 0.68, + "learning_rate": 0.00026201805416248747, + "loss": 1.2185, + "theoretical_loss": 3.4281068420464154, + "tokens_seen": 2230714368 + }, + { + "epoch": 0.68, + "learning_rate": 0.00026200200601805417, + "loss": 1.3332, + "theoretical_loss": 3.428098692260051, + "tokens_seen": 2230779904 + }, + { + "epoch": 0.68, + "learning_rate": 0.00026198595787362087, + "loss": 1.2839, + "theoretical_loss": 3.428090542780145, + "tokens_seen": 2230845440 + }, + { + "epoch": 0.68, + "learning_rate": 0.00026196990972918757, + "loss": 1.3022, + "theoretical_loss": 3.4280823936066764, + "tokens_seen": 2230910976 + }, + { + "epoch": 0.68, + "learning_rate": 0.00026195386158475427, + "loss": 1.2906, + "theoretical_loss": 3.428074244739625, + "tokens_seen": 2230976512 + }, + { + "epoch": 0.68, + "learning_rate": 0.00026193781344032097, + "loss": 1.4521, + "theoretical_loss": 3.4280660961789704, + "tokens_seen": 2231042048 + }, + { + "epoch": 0.68, + "learning_rate": 0.00026192176529588767, + "loss": 1.4348, + "theoretical_loss": 3.4280579479246915, + "tokens_seen": 2231107584 + }, + { + "epoch": 0.68, + "learning_rate": 0.00026190571715145437, + "loss": 1.2486, + "theoretical_loss": 3.428049799976769, + "tokens_seen": 2231173120 + }, + { + "epoch": 0.68, + "learning_rate": 0.00026188966900702107, + "loss": 1.2901, + "theoretical_loss": 3.428041652335181, + "tokens_seen": 2231238656 + }, + { + "epoch": 0.68, + "learning_rate": 0.0002618736208625878, + "loss": 1.3906, + "theoretical_loss": 3.4280335049999073, + "tokens_seen": 2231304192 + }, + { + "epoch": 0.68, + "learning_rate": 0.0002618575727181545, + "loss": 1.373, + "theoretical_loss": 3.428025357970928, + "tokens_seen": 2231369728 + }, + { + "epoch": 0.68, + "learning_rate": 0.0002618415245737212, + "loss": 1.3323, + "theoretical_loss": 3.428017211248222, + "tokens_seen": 2231435264 + }, + { + "epoch": 0.68, + "objective/train/docs_used": 812344, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.1589101552963257, + "objective/train/theoretical_loss": 3.4280090648317696, + "objective/train/tokens_used": 2251960800, + "theoretical_loss": 3.4280090648317696, + "tokens_seen": 2231500800 + }, + { + "epoch": 0.68, + "learning_rate": 0.0002618254764292879, + "loss": 1.2756, + "theoretical_loss": 3.4280090648317696, + "tokens_seen": 2231500800 + }, + { + "epoch": 0.68, + "learning_rate": 0.0002618094282848546, + "loss": 1.3784, + "theoretical_loss": 3.428000918721549, + "tokens_seen": 2231566336 + }, + { + "epoch": 0.68, + "learning_rate": 0.0002617933801404213, + "loss": 1.3263, + "theoretical_loss": 3.427992772917541, + "tokens_seen": 2231631872 + }, + { + "epoch": 0.68, + "learning_rate": 0.000261777331995988, + "loss": 1.3841, + "theoretical_loss": 3.4279846274197237, + "tokens_seen": 2231697408 + }, + { + "epoch": 0.68, + "learning_rate": 0.0002617612838515547, + "loss": 1.34, + "theoretical_loss": 3.427976482228078, + "tokens_seen": 2231762944 + }, + { + "epoch": 0.68, + "learning_rate": 0.0002617452357071214, + "loss": 1.2601, + "theoretical_loss": 3.427968337342583, + "tokens_seen": 2231828480 + }, + { + "epoch": 0.68, + "learning_rate": 0.0002617291875626881, + "loss": 1.346, + "theoretical_loss": 3.4279601927632175, + "tokens_seen": 2231894016 + }, + { + "epoch": 0.68, + "learning_rate": 0.0002617131394182548, + "loss": 1.1783, + "theoretical_loss": 3.4279520484899617, + "tokens_seen": 2231959552 + }, + { + "epoch": 0.68, + "learning_rate": 0.0002616970912738215, + "loss": 1.3296, + "theoretical_loss": 3.4279439045227953, + "tokens_seen": 2232025088 + }, + { + "epoch": 0.68, + "learning_rate": 0.0002616810431293882, + "loss": 1.3297, + "theoretical_loss": 3.427935760861697, + "tokens_seen": 2232090624 + }, + { + "epoch": 0.68, + "learning_rate": 0.0002616649949849549, + "loss": 1.2859, + "theoretical_loss": 3.4279276175066467, + "tokens_seen": 2232156160 + }, + { + "epoch": 0.68, + "learning_rate": 0.0002616489468405216, + "loss": 1.2684, + "theoretical_loss": 3.427919474457624, + "tokens_seen": 2232221696 + }, + { + "epoch": 0.68, + "learning_rate": 0.0002616328986960883, + "loss": 1.3343, + "theoretical_loss": 3.4279113317146086, + "tokens_seen": 2232287232 + }, + { + "epoch": 0.68, + "learning_rate": 0.000261616850551655, + "loss": 1.2244, + "theoretical_loss": 3.4279031892775795, + "tokens_seen": 2232352768 + }, + { + "epoch": 0.68, + "learning_rate": 0.0002616008024072217, + "loss": 1.2874, + "theoretical_loss": 3.427895047146517, + "tokens_seen": 2232418304 + }, + { + "epoch": 0.68, + "learning_rate": 0.0002615847542627884, + "loss": 1.3759, + "theoretical_loss": 3.4278869053213996, + "tokens_seen": 2232483840 + }, + { + "epoch": 0.68, + "learning_rate": 0.0002615687061183551, + "loss": 1.2951, + "theoretical_loss": 3.4278787638022075, + "tokens_seen": 2232549376 + }, + { + "epoch": 0.68, + "learning_rate": 0.0002615526579739218, + "loss": 1.3441, + "theoretical_loss": 3.4278706225889204, + "tokens_seen": 2232614912 + }, + { + "epoch": 0.68, + "learning_rate": 0.0002615366098294885, + "loss": 1.286, + "theoretical_loss": 3.427862481681517, + "tokens_seen": 2232680448 + }, + { + "epoch": 0.68, + "learning_rate": 0.0002615205616850552, + "loss": 1.3664, + "theoretical_loss": 3.4278543410799776, + "tokens_seen": 2232745984 + }, + { + "epoch": 0.68, + "learning_rate": 0.0002615045135406219, + "loss": 1.3553, + "theoretical_loss": 3.4278462007842814, + "tokens_seen": 2232811520 + }, + { + "epoch": 0.68, + "learning_rate": 0.0002614884653961886, + "loss": 1.2947, + "theoretical_loss": 3.427838060794408, + "tokens_seen": 2232877056 + }, + { + "epoch": 0.68, + "learning_rate": 0.0002614724172517553, + "loss": 1.316, + "theoretical_loss": 3.427829921110337, + "tokens_seen": 2232942592 + }, + { + "epoch": 0.68, + "learning_rate": 0.000261456369107322, + "loss": 1.3531, + "theoretical_loss": 3.4278217817320478, + "tokens_seen": 2233008128 + }, + { + "epoch": 0.68, + "learning_rate": 0.00026144032096288864, + "loss": 1.242, + "theoretical_loss": 3.42781364265952, + "tokens_seen": 2233073664 + }, + { + "epoch": 0.68, + "learning_rate": 0.00026142427281845534, + "loss": 1.3894, + "theoretical_loss": 3.4278055038927335, + "tokens_seen": 2233139200 + }, + { + "epoch": 0.68, + "learning_rate": 0.00026140822467402204, + "loss": 1.3056, + "theoretical_loss": 3.427797365431667, + "tokens_seen": 2233204736 + }, + { + "epoch": 0.68, + "learning_rate": 0.00026139217652958874, + "loss": 1.3961, + "theoretical_loss": 3.4277892272763006, + "tokens_seen": 2233270272 + }, + { + "epoch": 0.68, + "learning_rate": 0.00026137612838515544, + "loss": 1.3227, + "theoretical_loss": 3.427781089426614, + "tokens_seen": 2233335808 + }, + { + "epoch": 0.68, + "learning_rate": 0.00026136008024072214, + "loss": 1.4128, + "theoretical_loss": 3.4277729518825866, + "tokens_seen": 2233401344 + }, + { + "epoch": 0.68, + "learning_rate": 0.00026134403209628884, + "loss": 1.3058, + "theoretical_loss": 3.4277648146441977, + "tokens_seen": 2233466880 + }, + { + "epoch": 0.68, + "learning_rate": 0.00026132798395185554, + "loss": 1.3006, + "theoretical_loss": 3.427756677711427, + "tokens_seen": 2233532416 + }, + { + "epoch": 0.68, + "learning_rate": 0.00026131193580742224, + "loss": 1.3346, + "theoretical_loss": 3.427748541084254, + "tokens_seen": 2233597952 + }, + { + "epoch": 0.68, + "learning_rate": 0.00026129588766298894, + "loss": 1.2614, + "theoretical_loss": 3.427740404762659, + "tokens_seen": 2233663488 + }, + { + "epoch": 0.68, + "learning_rate": 0.00026127983951855564, + "loss": 1.2889, + "theoretical_loss": 3.4277322687466203, + "tokens_seen": 2233729024 + }, + { + "epoch": 0.68, + "learning_rate": 0.00026126379137412234, + "loss": 1.3138, + "theoretical_loss": 3.427724133036118, + "tokens_seen": 2233794560 + }, + { + "epoch": 0.68, + "learning_rate": 0.0002612477432296891, + "loss": 1.324, + "theoretical_loss": 3.427715997631132, + "tokens_seen": 2233860096 + }, + { + "epoch": 0.68, + "learning_rate": 0.0002612316950852558, + "loss": 1.3283, + "theoretical_loss": 3.4277078625316415, + "tokens_seen": 2233925632 + }, + { + "epoch": 0.68, + "learning_rate": 0.0002612156469408225, + "loss": 1.3938, + "theoretical_loss": 3.427699727737626, + "tokens_seen": 2233991168 + }, + { + "epoch": 0.68, + "learning_rate": 0.0002611995987963892, + "loss": 1.349, + "theoretical_loss": 3.4276915932490657, + "tokens_seen": 2234056704 + }, + { + "epoch": 0.68, + "learning_rate": 0.0002611835506519559, + "loss": 1.3217, + "theoretical_loss": 3.4276834590659395, + "tokens_seen": 2234122240 + }, + { + "epoch": 0.68, + "learning_rate": 0.0002611675025075226, + "loss": 1.408, + "theoretical_loss": 3.4276753251882273, + "tokens_seen": 2234187776 + }, + { + "epoch": 0.68, + "learning_rate": 0.0002611514543630893, + "loss": 1.3045, + "theoretical_loss": 3.427667191615908, + "tokens_seen": 2234253312 + }, + { + "epoch": 0.68, + "learning_rate": 0.000261135406218656, + "loss": 1.3415, + "theoretical_loss": 3.4276590583489623, + "tokens_seen": 2234318848 + }, + { + "epoch": 0.68, + "learning_rate": 0.0002611193580742227, + "loss": 1.3878, + "theoretical_loss": 3.427650925387369, + "tokens_seen": 2234384384 + }, + { + "epoch": 0.68, + "learning_rate": 0.0002611033099297894, + "loss": 1.3714, + "theoretical_loss": 3.4276427927311075, + "tokens_seen": 2234449920 + }, + { + "epoch": 0.68, + "learning_rate": 0.0002610872617853561, + "loss": 1.2365, + "theoretical_loss": 3.427634660380158, + "tokens_seen": 2234515456 + }, + { + "epoch": 0.68, + "learning_rate": 0.0002610712136409228, + "loss": 1.2324, + "theoretical_loss": 3.4276265283345, + "tokens_seen": 2234580992 + }, + { + "epoch": 0.68, + "learning_rate": 0.0002610551654964895, + "loss": 1.4036, + "theoretical_loss": 3.427618396594113, + "tokens_seen": 2234646528 + }, + { + "epoch": 0.68, + "learning_rate": 0.0002610391173520562, + "loss": 1.3343, + "theoretical_loss": 3.4276102651589766, + "tokens_seen": 2234712064 + }, + { + "epoch": 0.68, + "objective/train/docs_used": 813661, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.326331377029419, + "objective/train/theoretical_loss": 3.42760213402907, + "objective/train/tokens_used": 2255237600, + "theoretical_loss": 3.42760213402907, + "tokens_seen": 2234777600 + }, + { + "epoch": 0.68, + "learning_rate": 0.0002610230692076229, + "loss": 1.3339, + "theoretical_loss": 3.42760213402907, + "tokens_seen": 2234777600 + }, + { + "epoch": 0.68, + "learning_rate": 0.0002610070210631896, + "loss": 1.323, + "theoretical_loss": 3.4275940032043732, + "tokens_seen": 2234843136 + }, + { + "epoch": 0.68, + "learning_rate": 0.0002609909729187563, + "loss": 1.3, + "theoretical_loss": 3.427585872684866, + "tokens_seen": 2234908672 + }, + { + "epoch": 0.68, + "learning_rate": 0.000260974924774323, + "loss": 1.2859, + "theoretical_loss": 3.427577742470527, + "tokens_seen": 2234974208 + }, + { + "epoch": 0.68, + "learning_rate": 0.0002609588766298897, + "loss": 1.1928, + "theoretical_loss": 3.4275696125613373, + "tokens_seen": 2235039744 + }, + { + "epoch": 0.68, + "learning_rate": 0.0002609428284854564, + "loss": 1.3088, + "theoretical_loss": 3.427561482957275, + "tokens_seen": 2235105280 + }, + { + "epoch": 0.68, + "learning_rate": 0.0002609267803410231, + "loss": 1.2016, + "theoretical_loss": 3.427553353658321, + "tokens_seen": 2235170816 + }, + { + "epoch": 0.68, + "learning_rate": 0.0002609107321965898, + "loss": 1.3665, + "theoretical_loss": 3.427545224664454, + "tokens_seen": 2235236352 + }, + { + "epoch": 0.68, + "learning_rate": 0.0002608946840521565, + "loss": 1.4135, + "theoretical_loss": 3.4275370959756537, + "tokens_seen": 2235301888 + }, + { + "epoch": 0.68, + "learning_rate": 0.0002608786359077232, + "loss": 1.3537, + "theoretical_loss": 3.4275289675919, + "tokens_seen": 2235367424 + }, + { + "epoch": 0.68, + "learning_rate": 0.0002608625877632899, + "loss": 1.307, + "theoretical_loss": 3.4275208395131727, + "tokens_seen": 2235432960 + }, + { + "epoch": 0.68, + "learning_rate": 0.0002608465396188566, + "loss": 1.2405, + "theoretical_loss": 3.427512711739451, + "tokens_seen": 2235498496 + }, + { + "epoch": 0.68, + "learning_rate": 0.0002608304914744233, + "loss": 1.3697, + "theoretical_loss": 3.427504584270715, + "tokens_seen": 2235564032 + }, + { + "epoch": 0.68, + "learning_rate": 0.00026081444332999, + "loss": 1.253, + "theoretical_loss": 3.427496457106943, + "tokens_seen": 2235629568 + }, + { + "epoch": 0.68, + "learning_rate": 0.00026079839518555666, + "loss": 1.3011, + "theoretical_loss": 3.4274883302481163, + "tokens_seen": 2235695104 + }, + { + "epoch": 0.68, + "learning_rate": 0.00026078234704112336, + "loss": 1.4236, + "theoretical_loss": 3.4274802036942136, + "tokens_seen": 2235760640 + }, + { + "epoch": 0.68, + "learning_rate": 0.00026076629889669006, + "loss": 1.2862, + "theoretical_loss": 3.427472077445215, + "tokens_seen": 2235826176 + }, + { + "epoch": 0.68, + "learning_rate": 0.00026075025075225676, + "loss": 1.3737, + "theoretical_loss": 3.4274639515010996, + "tokens_seen": 2235891712 + }, + { + "epoch": 0.68, + "learning_rate": 0.00026073420260782346, + "loss": 1.3641, + "theoretical_loss": 3.4274558258618475, + "tokens_seen": 2235957248 + }, + { + "epoch": 0.68, + "learning_rate": 0.00026071815446339016, + "loss": 1.3317, + "theoretical_loss": 3.4274477005274377, + "tokens_seen": 2236022784 + }, + { + "epoch": 0.68, + "learning_rate": 0.00026070210631895687, + "loss": 1.3303, + "theoretical_loss": 3.4274395754978504, + "tokens_seen": 2236088320 + }, + { + "epoch": 0.68, + "learning_rate": 0.00026068605817452357, + "loss": 1.3231, + "theoretical_loss": 3.427431450773065, + "tokens_seen": 2236153856 + }, + { + "epoch": 0.68, + "learning_rate": 0.00026067001003009027, + "loss": 1.3276, + "theoretical_loss": 3.4274233263530616, + "tokens_seen": 2236219392 + }, + { + "epoch": 0.68, + "learning_rate": 0.00026065396188565697, + "loss": 1.2292, + "theoretical_loss": 3.427415202237819, + "tokens_seen": 2236284928 + }, + { + "epoch": 0.68, + "learning_rate": 0.00026063791374122367, + "loss": 1.2984, + "theoretical_loss": 3.427407078427317, + "tokens_seen": 2236350464 + }, + { + "epoch": 0.68, + "learning_rate": 0.00026062186559679037, + "loss": 1.3191, + "theoretical_loss": 3.4273989549215362, + "tokens_seen": 2236416000 + }, + { + "epoch": 0.68, + "learning_rate": 0.00026060581745235707, + "loss": 1.2326, + "theoretical_loss": 3.427390831720455, + "tokens_seen": 2236481536 + }, + { + "epoch": 0.68, + "learning_rate": 0.00026058976930792377, + "loss": 1.2721, + "theoretical_loss": 3.4273827088240543, + "tokens_seen": 2236547072 + }, + { + "epoch": 0.68, + "learning_rate": 0.00026057372116349047, + "loss": 1.3251, + "theoretical_loss": 3.4273745862323124, + "tokens_seen": 2236612608 + }, + { + "epoch": 0.68, + "learning_rate": 0.00026055767301905717, + "loss": 1.2501, + "theoretical_loss": 3.4273664639452095, + "tokens_seen": 2236678144 + }, + { + "epoch": 0.68, + "learning_rate": 0.00026054162487462387, + "loss": 1.36, + "theoretical_loss": 3.4273583419627256, + "tokens_seen": 2236743680 + }, + { + "epoch": 0.68, + "learning_rate": 0.00026052557673019057, + "loss": 1.3224, + "theoretical_loss": 3.42735022028484, + "tokens_seen": 2236809216 + }, + { + "epoch": 0.68, + "learning_rate": 0.00026050952858575727, + "loss": 1.2746, + "theoretical_loss": 3.4273420989115326, + "tokens_seen": 2236874752 + }, + { + "epoch": 0.68, + "learning_rate": 0.000260493480441324, + "loss": 1.4039, + "theoretical_loss": 3.4273339778427827, + "tokens_seen": 2236940288 + }, + { + "epoch": 0.68, + "learning_rate": 0.0002604774322968907, + "loss": 1.342, + "theoretical_loss": 3.42732585707857, + "tokens_seen": 2237005824 + }, + { + "epoch": 0.68, + "learning_rate": 0.0002604613841524574, + "loss": 1.2859, + "theoretical_loss": 3.4273177366188747, + "tokens_seen": 2237071360 + }, + { + "epoch": 0.68, + "learning_rate": 0.0002604453360080241, + "loss": 1.2145, + "theoretical_loss": 3.427309616463676, + "tokens_seen": 2237136896 + }, + { + "epoch": 0.68, + "learning_rate": 0.0002604292878635908, + "loss": 1.4188, + "theoretical_loss": 3.4273014966129534, + "tokens_seen": 2237202432 + }, + { + "epoch": 0.68, + "learning_rate": 0.0002604132397191575, + "loss": 1.3189, + "theoretical_loss": 3.4272933770666865, + "tokens_seen": 2237267968 + }, + { + "epoch": 0.68, + "learning_rate": 0.0002603971915747242, + "loss": 1.3282, + "theoretical_loss": 3.4272852578248556, + "tokens_seen": 2237333504 + }, + { + "epoch": 0.68, + "learning_rate": 0.0002603811434302909, + "loss": 1.3575, + "theoretical_loss": 3.4272771388874403, + "tokens_seen": 2237399040 + }, + { + "epoch": 0.68, + "learning_rate": 0.0002603650952858576, + "loss": 1.4101, + "theoretical_loss": 3.4272690202544194, + "tokens_seen": 2237464576 + }, + { + "epoch": 0.68, + "learning_rate": 0.0002603490471414243, + "loss": 1.3417, + "theoretical_loss": 3.4272609019257736, + "tokens_seen": 2237530112 + }, + { + "epoch": 0.68, + "learning_rate": 0.000260332998996991, + "loss": 1.2023, + "theoretical_loss": 3.427252783901482, + "tokens_seen": 2237595648 + }, + { + "epoch": 0.68, + "learning_rate": 0.0002603169508525577, + "loss": 1.3495, + "theoretical_loss": 3.4272446661815246, + "tokens_seen": 2237661184 + }, + { + "epoch": 0.68, + "learning_rate": 0.0002603009027081244, + "loss": 1.3613, + "theoretical_loss": 3.4272365487658805, + "tokens_seen": 2237726720 + }, + { + "epoch": 0.68, + "learning_rate": 0.0002602848545636911, + "loss": 1.4429, + "theoretical_loss": 3.4272284316545303, + "tokens_seen": 2237792256 + }, + { + "epoch": 0.68, + "learning_rate": 0.0002602688064192578, + "loss": 1.4285, + "theoretical_loss": 3.4272203148474527, + "tokens_seen": 2237857792 + }, + { + "epoch": 0.68, + "learning_rate": 0.0002602527582748245, + "loss": 1.2296, + "theoretical_loss": 3.4272121983446278, + "tokens_seen": 2237923328 + }, + { + "epoch": 0.68, + "learning_rate": 0.0002602367101303912, + "loss": 1.3232, + "theoretical_loss": 3.427204082146036, + "tokens_seen": 2237988864 + }, + { + "epoch": 0.68, + "objective/train/docs_used": 814894, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.21431565284729, + "objective/train/theoretical_loss": 3.427195966251656, + "objective/train/tokens_used": 2258514400, + "theoretical_loss": 3.427195966251656, + "tokens_seen": 2238054400 + }, + { + "epoch": 0.68, + "learning_rate": 0.0002602206619859579, + "loss": 1.1878, + "theoretical_loss": 3.427195966251656, + "tokens_seen": 2238054400 + }, + { + "epoch": 0.68, + "learning_rate": 0.0002602046138415246, + "loss": 1.3369, + "theoretical_loss": 3.4271878506614675, + "tokens_seen": 2238119936 + }, + { + "epoch": 0.68, + "learning_rate": 0.0002601885656970913, + "loss": 1.3951, + "theoretical_loss": 3.4271797353754505, + "tokens_seen": 2238185472 + }, + { + "epoch": 0.68, + "learning_rate": 0.000260172517552658, + "loss": 1.387, + "theoretical_loss": 3.427171620393585, + "tokens_seen": 2238251008 + }, + { + "epoch": 0.68, + "learning_rate": 0.0002601564694082247, + "loss": 1.2848, + "theoretical_loss": 3.42716350571585, + "tokens_seen": 2238316544 + }, + { + "epoch": 0.68, + "learning_rate": 0.0002601404212637914, + "loss": 1.2283, + "theoretical_loss": 3.427155391342226, + "tokens_seen": 2238382080 + }, + { + "epoch": 0.68, + "learning_rate": 0.0002601243731193581, + "loss": 1.2658, + "theoretical_loss": 3.427147277272692, + "tokens_seen": 2238447616 + }, + { + "epoch": 0.68, + "learning_rate": 0.0002601083249749248, + "loss": 1.3056, + "theoretical_loss": 3.427139163507228, + "tokens_seen": 2238513152 + }, + { + "epoch": 0.68, + "learning_rate": 0.0002600922768304915, + "loss": 1.3855, + "theoretical_loss": 3.427131050045814, + "tokens_seen": 2238578688 + }, + { + "epoch": 0.68, + "learning_rate": 0.0002600762286860582, + "loss": 1.43, + "theoretical_loss": 3.427122936888429, + "tokens_seen": 2238644224 + }, + { + "epoch": 0.68, + "learning_rate": 0.0002600601805416249, + "loss": 1.3362, + "theoretical_loss": 3.427114824035053, + "tokens_seen": 2238709760 + }, + { + "epoch": 0.68, + "learning_rate": 0.0002600441323971916, + "loss": 1.2927, + "theoretical_loss": 3.427106711485666, + "tokens_seen": 2238775296 + }, + { + "epoch": 0.68, + "learning_rate": 0.0002600280842527583, + "loss": 1.4264, + "theoretical_loss": 3.4270985992402476, + "tokens_seen": 2238840832 + }, + { + "epoch": 0.68, + "learning_rate": 0.000260012036108325, + "loss": 1.4193, + "theoretical_loss": 3.427090487298778, + "tokens_seen": 2238906368 + }, + { + "epoch": 0.68, + "learning_rate": 0.0002599959879638917, + "loss": 1.2691, + "theoretical_loss": 3.4270823756612354, + "tokens_seen": 2238971904 + }, + { + "epoch": 0.68, + "learning_rate": 0.0002599799398194584, + "loss": 1.3372, + "theoretical_loss": 3.427074264327601, + "tokens_seen": 2239037440 + }, + { + "epoch": 0.68, + "learning_rate": 0.0002599638916750251, + "loss": 1.2551, + "theoretical_loss": 3.4270661532978535, + "tokens_seen": 2239102976 + }, + { + "epoch": 0.68, + "learning_rate": 0.0002599478435305918, + "loss": 1.3693, + "theoretical_loss": 3.4270580425719737, + "tokens_seen": 2239168512 + }, + { + "epoch": 0.68, + "learning_rate": 0.0002599317953861585, + "loss": 1.3006, + "theoretical_loss": 3.4270499321499406, + "tokens_seen": 2239234048 + }, + { + "epoch": 0.68, + "learning_rate": 0.0002599157472417252, + "loss": 1.3834, + "theoretical_loss": 3.427041822031734, + "tokens_seen": 2239299584 + }, + { + "epoch": 0.68, + "learning_rate": 0.0002598996990972919, + "loss": 1.229, + "theoretical_loss": 3.4270337122173333, + "tokens_seen": 2239365120 + }, + { + "epoch": 0.68, + "learning_rate": 0.0002598836509528586, + "loss": 1.3701, + "theoretical_loss": 3.427025602706719, + "tokens_seen": 2239430656 + }, + { + "epoch": 0.68, + "learning_rate": 0.0002598676028084253, + "loss": 1.3416, + "theoretical_loss": 3.4270174934998705, + "tokens_seen": 2239496192 + }, + { + "epoch": 0.68, + "learning_rate": 0.000259851554663992, + "loss": 1.371, + "theoretical_loss": 3.4270093845967673, + "tokens_seen": 2239561728 + }, + { + "epoch": 0.68, + "learning_rate": 0.0002598355065195587, + "loss": 1.2938, + "theoretical_loss": 3.427001275997389, + "tokens_seen": 2239627264 + }, + { + "epoch": 0.68, + "learning_rate": 0.0002598194583751254, + "loss": 1.3701, + "theoretical_loss": 3.4269931677017165, + "tokens_seen": 2239692800 + }, + { + "epoch": 0.68, + "learning_rate": 0.0002598034102306921, + "loss": 1.3885, + "theoretical_loss": 3.426985059709728, + "tokens_seen": 2239758336 + }, + { + "epoch": 0.68, + "learning_rate": 0.0002597873620862588, + "loss": 1.2876, + "theoretical_loss": 3.426976952021404, + "tokens_seen": 2239823872 + }, + { + "epoch": 0.68, + "learning_rate": 0.0002597713139418255, + "loss": 1.3125, + "theoretical_loss": 3.4269688446367246, + "tokens_seen": 2239889408 + }, + { + "epoch": 0.68, + "learning_rate": 0.0002597552657973922, + "loss": 1.3039, + "theoretical_loss": 3.4269607375556683, + "tokens_seen": 2239954944 + }, + { + "epoch": 0.68, + "learning_rate": 0.0002597392176529589, + "loss": 1.3426, + "theoretical_loss": 3.4269526307782163, + "tokens_seen": 2240020480 + }, + { + "epoch": 0.68, + "learning_rate": 0.0002597231695085256, + "loss": 1.3283, + "theoretical_loss": 3.4269445243043477, + "tokens_seen": 2240086016 + }, + { + "epoch": 0.68, + "learning_rate": 0.0002597071213640923, + "loss": 1.2118, + "theoretical_loss": 3.426936418134042, + "tokens_seen": 2240151552 + }, + { + "epoch": 0.68, + "learning_rate": 0.000259691073219659, + "loss": 1.4008, + "theoretical_loss": 3.4269283122672793, + "tokens_seen": 2240217088 + }, + { + "epoch": 0.68, + "learning_rate": 0.0002596750250752257, + "loss": 1.3181, + "theoretical_loss": 3.426920206704039, + "tokens_seen": 2240282624 + }, + { + "epoch": 0.68, + "learning_rate": 0.0002596589769307924, + "loss": 1.3543, + "theoretical_loss": 3.4269121014443016, + "tokens_seen": 2240348160 + }, + { + "epoch": 0.68, + "learning_rate": 0.0002596429287863591, + "loss": 1.2707, + "theoretical_loss": 3.426903996488046, + "tokens_seen": 2240413696 + }, + { + "epoch": 0.68, + "learning_rate": 0.0002596268806419258, + "loss": 1.3995, + "theoretical_loss": 3.426895891835253, + "tokens_seen": 2240479232 + }, + { + "epoch": 0.68, + "learning_rate": 0.0002596108324974925, + "loss": 1.3312, + "theoretical_loss": 3.426887787485901, + "tokens_seen": 2240544768 + }, + { + "epoch": 0.68, + "learning_rate": 0.0002595947843530592, + "loss": 1.1988, + "theoretical_loss": 3.426879683439971, + "tokens_seen": 2240610304 + }, + { + "epoch": 0.68, + "learning_rate": 0.0002595787362086259, + "loss": 1.3244, + "theoretical_loss": 3.4268715796974414, + "tokens_seen": 2240675840 + }, + { + "epoch": 0.68, + "learning_rate": 0.0002595626880641926, + "loss": 1.3009, + "theoretical_loss": 3.4268634762582937, + "tokens_seen": 2240741376 + }, + { + "epoch": 0.68, + "learning_rate": 0.0002595466399197593, + "loss": 1.2235, + "theoretical_loss": 3.426855373122506, + "tokens_seen": 2240806912 + }, + { + "epoch": 0.68, + "learning_rate": 0.000259530591775326, + "loss": 1.2504, + "theoretical_loss": 3.426847270290059, + "tokens_seen": 2240872448 + }, + { + "epoch": 0.68, + "learning_rate": 0.0002595145436308927, + "loss": 1.2795, + "theoretical_loss": 3.426839167760933, + "tokens_seen": 2240937984 + }, + { + "epoch": 0.68, + "learning_rate": 0.0002594984954864594, + "loss": 1.3657, + "theoretical_loss": 3.4268310655351066, + "tokens_seen": 2241003520 + }, + { + "epoch": 0.68, + "learning_rate": 0.0002594824473420261, + "loss": 1.3183, + "theoretical_loss": 3.42682296361256, + "tokens_seen": 2241069056 + }, + { + "epoch": 0.68, + "learning_rate": 0.0002594663991975928, + "loss": 1.3315, + "theoretical_loss": 3.426814861993273, + "tokens_seen": 2241134592 + }, + { + "epoch": 0.68, + "learning_rate": 0.0002594503510531595, + "loss": 1.2661, + "theoretical_loss": 3.4268067606772257, + "tokens_seen": 2241200128 + }, + { + "epoch": 0.68, + "learning_rate": 0.0002594343029087262, + "loss": 1.2645, + "theoretical_loss": 3.4267986596643976, + "tokens_seen": 2241265664 + }, + { + "epoch": 0.68, + "objective/train/docs_used": 816099, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.2895150184631348, + "objective/train/theoretical_loss": 3.4267905589547683, + "objective/train/tokens_used": 2261791200, + "theoretical_loss": 3.4267905589547683, + "tokens_seen": 2241331200 + }, + { + "epoch": 0.68, + "learning_rate": 0.0002594182547642929, + "loss": 1.2711, + "theoretical_loss": 3.4267905589547683, + "tokens_seen": 2241331200 + }, + { + "epoch": 0.68, + "learning_rate": 0.00025940220661985956, + "loss": 1.2638, + "theoretical_loss": 3.4267824585483178, + "tokens_seen": 2241396736 + }, + { + "epoch": 0.68, + "learning_rate": 0.00025938615847542626, + "loss": 1.3362, + "theoretical_loss": 3.4267743584450256, + "tokens_seen": 2241462272 + }, + { + "epoch": 0.68, + "learning_rate": 0.00025937011033099296, + "loss": 1.3325, + "theoretical_loss": 3.4267662586448724, + "tokens_seen": 2241527808 + }, + { + "epoch": 0.68, + "learning_rate": 0.00025935406218655966, + "loss": 1.2941, + "theoretical_loss": 3.4267581591478367, + "tokens_seen": 2241593344 + }, + { + "epoch": 0.68, + "learning_rate": 0.00025933801404212636, + "loss": 1.3212, + "theoretical_loss": 3.4267500599538994, + "tokens_seen": 2241658880 + }, + { + "epoch": 0.68, + "learning_rate": 0.00025932196589769307, + "loss": 1.2678, + "theoretical_loss": 3.4267419610630396, + "tokens_seen": 2241724416 + }, + { + "epoch": 0.68, + "learning_rate": 0.00025930591775325977, + "loss": 1.33, + "theoretical_loss": 3.4267338624752375, + "tokens_seen": 2241789952 + }, + { + "epoch": 0.68, + "learning_rate": 0.00025928986960882647, + "loss": 1.2175, + "theoretical_loss": 3.426725764190473, + "tokens_seen": 2241855488 + }, + { + "epoch": 0.68, + "learning_rate": 0.00025927382146439317, + "loss": 1.2487, + "theoretical_loss": 3.426717666208725, + "tokens_seen": 2241921024 + }, + { + "epoch": 0.68, + "learning_rate": 0.00025925777331995987, + "loss": 1.2625, + "theoretical_loss": 3.426709568529974, + "tokens_seen": 2241986560 + }, + { + "epoch": 0.68, + "learning_rate": 0.00025924172517552657, + "loss": 1.3342, + "theoretical_loss": 3.4267014711542005, + "tokens_seen": 2242052096 + }, + { + "epoch": 0.68, + "learning_rate": 0.00025922567703109327, + "loss": 1.2627, + "theoretical_loss": 3.426693374081383, + "tokens_seen": 2242117632 + }, + { + "epoch": 0.68, + "learning_rate": 0.00025920962888665997, + "loss": 1.376, + "theoretical_loss": 3.4266852773115017, + "tokens_seen": 2242183168 + }, + { + "epoch": 0.68, + "learning_rate": 0.00025919358074222667, + "loss": 1.2814, + "theoretical_loss": 3.4266771808445373, + "tokens_seen": 2242248704 + }, + { + "epoch": 0.68, + "learning_rate": 0.00025917753259779337, + "loss": 1.3207, + "theoretical_loss": 3.4266690846804684, + "tokens_seen": 2242314240 + }, + { + "epoch": 0.68, + "learning_rate": 0.00025916148445336007, + "loss": 1.3613, + "theoretical_loss": 3.4266609888192754, + "tokens_seen": 2242379776 + }, + { + "epoch": 0.68, + "learning_rate": 0.00025914543630892677, + "loss": 1.3218, + "theoretical_loss": 3.426652893260938, + "tokens_seen": 2242445312 + }, + { + "epoch": 0.68, + "learning_rate": 0.0002591293881644935, + "loss": 1.3444, + "theoretical_loss": 3.426644798005436, + "tokens_seen": 2242510848 + }, + { + "epoch": 0.68, + "learning_rate": 0.0002591133400200602, + "loss": 1.3095, + "theoretical_loss": 3.4266367030527496, + "tokens_seen": 2242576384 + }, + { + "epoch": 0.68, + "learning_rate": 0.0002590972918756269, + "loss": 1.3042, + "theoretical_loss": 3.426628608402858, + "tokens_seen": 2242641920 + }, + { + "epoch": 0.68, + "learning_rate": 0.0002590812437311936, + "loss": 1.2313, + "theoretical_loss": 3.4266205140557413, + "tokens_seen": 2242707456 + }, + { + "epoch": 0.68, + "learning_rate": 0.0002590651955867603, + "loss": 1.386, + "theoretical_loss": 3.4266124200113794, + "tokens_seen": 2242772992 + }, + { + "epoch": 0.68, + "learning_rate": 0.000259049147442327, + "loss": 1.2269, + "theoretical_loss": 3.4266043262697523, + "tokens_seen": 2242838528 + }, + { + "epoch": 0.68, + "learning_rate": 0.0002590330992978937, + "loss": 1.3378, + "theoretical_loss": 3.4265962328308395, + "tokens_seen": 2242904064 + }, + { + "epoch": 0.68, + "learning_rate": 0.0002590170511534604, + "loss": 1.3599, + "theoretical_loss": 3.426588139694621, + "tokens_seen": 2242969600 + }, + { + "epoch": 0.68, + "learning_rate": 0.0002590010030090271, + "loss": 1.2439, + "theoretical_loss": 3.4265800468610763, + "tokens_seen": 2243035136 + }, + { + "epoch": 0.68, + "learning_rate": 0.0002589849548645938, + "loss": 1.2538, + "theoretical_loss": 3.4265719543301856, + "tokens_seen": 2243100672 + }, + { + "epoch": 0.68, + "learning_rate": 0.0002589689067201605, + "loss": 1.3648, + "theoretical_loss": 3.4265638621019288, + "tokens_seen": 2243166208 + }, + { + "epoch": 0.68, + "learning_rate": 0.0002589528585757272, + "loss": 1.2842, + "theoretical_loss": 3.426555770176286, + "tokens_seen": 2243231744 + }, + { + "epoch": 0.68, + "learning_rate": 0.0002589368104312939, + "loss": 1.2049, + "theoretical_loss": 3.426547678553236, + "tokens_seen": 2243297280 + }, + { + "epoch": 0.68, + "learning_rate": 0.0002589207622868606, + "loss": 1.2278, + "theoretical_loss": 3.4265395872327593, + "tokens_seen": 2243362816 + }, + { + "epoch": 0.68, + "learning_rate": 0.0002589047141424273, + "loss": 1.3358, + "theoretical_loss": 3.426531496214836, + "tokens_seen": 2243428352 + }, + { + "epoch": 0.68, + "learning_rate": 0.000258888665997994, + "loss": 1.3269, + "theoretical_loss": 3.426523405499446, + "tokens_seen": 2243493888 + }, + { + "epoch": 0.68, + "learning_rate": 0.0002588726178535607, + "loss": 1.2868, + "theoretical_loss": 3.4265153150865686, + "tokens_seen": 2243559424 + }, + { + "epoch": 0.68, + "learning_rate": 0.00025885656970912744, + "loss": 1.2389, + "theoretical_loss": 3.4265072249761834, + "tokens_seen": 2243624960 + }, + { + "epoch": 0.68, + "learning_rate": 0.00025884052156469414, + "loss": 1.205, + "theoretical_loss": 3.4264991351682714, + "tokens_seen": 2243690496 + }, + { + "epoch": 0.68, + "learning_rate": 0.00025882447342026084, + "loss": 1.3666, + "theoretical_loss": 3.4264910456628117, + "tokens_seen": 2243756032 + }, + { + "epoch": 0.68, + "learning_rate": 0.00025880842527582754, + "loss": 1.2077, + "theoretical_loss": 3.4264829564597843, + "tokens_seen": 2243821568 + }, + { + "epoch": 0.68, + "learning_rate": 0.00025879237713139424, + "loss": 1.3477, + "theoretical_loss": 3.4264748675591687, + "tokens_seen": 2243887104 + }, + { + "epoch": 0.68, + "learning_rate": 0.00025877632898696094, + "loss": 1.3222, + "theoretical_loss": 3.426466778960945, + "tokens_seen": 2243952640 + }, + { + "epoch": 0.68, + "learning_rate": 0.0002587602808425276, + "loss": 1.3313, + "theoretical_loss": 3.4264586906650933, + "tokens_seen": 2244018176 + }, + { + "epoch": 0.68, + "learning_rate": 0.0002587442326980943, + "loss": 1.3359, + "theoretical_loss": 3.4264506026715935, + "tokens_seen": 2244083712 + }, + { + "epoch": 0.68, + "learning_rate": 0.000258728184553661, + "loss": 1.323, + "theoretical_loss": 3.426442514980425, + "tokens_seen": 2244149248 + }, + { + "epoch": 0.68, + "learning_rate": 0.0002587121364092277, + "loss": 1.3209, + "theoretical_loss": 3.4264344275915684, + "tokens_seen": 2244214784 + }, + { + "epoch": 0.68, + "learning_rate": 0.0002586960882647944, + "loss": 1.4169, + "theoretical_loss": 3.426426340505003, + "tokens_seen": 2244280320 + }, + { + "epoch": 0.68, + "learning_rate": 0.0002586800401203611, + "loss": 1.4554, + "theoretical_loss": 3.4264182537207084, + "tokens_seen": 2244345856 + }, + { + "epoch": 0.68, + "learning_rate": 0.0002586639919759278, + "loss": 1.262, + "theoretical_loss": 3.4264101672386653, + "tokens_seen": 2244411392 + }, + { + "epoch": 0.68, + "learning_rate": 0.0002586479438314945, + "loss": 1.3041, + "theoretical_loss": 3.4264020810588534, + "tokens_seen": 2244476928 + }, + { + "epoch": 0.68, + "learning_rate": 0.0002586318956870612, + "loss": 1.2099, + "theoretical_loss": 3.4263939951812517, + "tokens_seen": 2244542464 + }, + { + "epoch": 0.68, + "objective/train/docs_used": 817136, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.3404786586761475, + "objective/train/theoretical_loss": 3.426385909605841, + "objective/train/tokens_used": 2265068000, + "theoretical_loss": 3.426385909605841, + "tokens_seen": 2244608000 + }, + { + "epoch": 0.68, + "learning_rate": 0.0002586158475426279, + "loss": 1.3743, + "theoretical_loss": 3.426385909605841, + "tokens_seen": 2244608000 + }, + { + "epoch": 0.68, + "learning_rate": 0.0002585997993981946, + "loss": 1.2763, + "theoretical_loss": 3.426377824332601, + "tokens_seen": 2244673536 + }, + { + "epoch": 0.68, + "learning_rate": 0.0002585837512537613, + "loss": 1.2495, + "theoretical_loss": 3.4263697393615113, + "tokens_seen": 2244739072 + }, + { + "epoch": 0.68, + "learning_rate": 0.000258567703109328, + "loss": 1.282, + "theoretical_loss": 3.426361654692552, + "tokens_seen": 2244804608 + }, + { + "epoch": 0.68, + "learning_rate": 0.0002585516549648947, + "loss": 1.3272, + "theoretical_loss": 3.426353570325703, + "tokens_seen": 2244870144 + }, + { + "epoch": 0.68, + "learning_rate": 0.0002585356068204614, + "loss": 1.3583, + "theoretical_loss": 3.426345486260944, + "tokens_seen": 2244935680 + }, + { + "epoch": 0.68, + "learning_rate": 0.0002585195586760281, + "loss": 1.3734, + "theoretical_loss": 3.4263374024982554, + "tokens_seen": 2245001216 + }, + { + "epoch": 0.68, + "learning_rate": 0.0002585035105315948, + "loss": 1.2135, + "theoretical_loss": 3.4263293190376167, + "tokens_seen": 2245066752 + }, + { + "epoch": 0.68, + "learning_rate": 0.0002584874623871615, + "loss": 1.3674, + "theoretical_loss": 3.4263212358790076, + "tokens_seen": 2245132288 + }, + { + "epoch": 0.68, + "learning_rate": 0.0002584714142427282, + "loss": 1.3289, + "theoretical_loss": 3.4263131530224085, + "tokens_seen": 2245197824 + }, + { + "epoch": 0.68, + "learning_rate": 0.0002584553660982949, + "loss": 1.203, + "theoretical_loss": 3.426305070467799, + "tokens_seen": 2245263360 + }, + { + "epoch": 0.68, + "learning_rate": 0.0002584393179538616, + "loss": 1.3449, + "theoretical_loss": 3.426296988215159, + "tokens_seen": 2245328896 + }, + { + "epoch": 0.68, + "learning_rate": 0.0002584232698094283, + "loss": 1.4092, + "theoretical_loss": 3.4262889062644684, + "tokens_seen": 2245394432 + }, + { + "epoch": 0.68, + "learning_rate": 0.000258407221664995, + "loss": 1.3138, + "theoretical_loss": 3.4262808246157066, + "tokens_seen": 2245459968 + }, + { + "epoch": 0.68, + "learning_rate": 0.0002583911735205617, + "loss": 1.3241, + "theoretical_loss": 3.426272743268855, + "tokens_seen": 2245525504 + }, + { + "epoch": 0.68, + "learning_rate": 0.0002583751253761284, + "loss": 1.3532, + "theoretical_loss": 3.426264662223892, + "tokens_seen": 2245591040 + }, + { + "epoch": 0.68, + "learning_rate": 0.0002583590772316951, + "loss": 1.2111, + "theoretical_loss": 3.426256581480798, + "tokens_seen": 2245656576 + }, + { + "epoch": 0.68, + "learning_rate": 0.0002583430290872618, + "loss": 1.3461, + "theoretical_loss": 3.4262485010395536, + "tokens_seen": 2245722112 + }, + { + "epoch": 0.68, + "learning_rate": 0.0002583269809428285, + "loss": 1.2913, + "theoretical_loss": 3.426240420900138, + "tokens_seen": 2245787648 + }, + { + "epoch": 0.68, + "learning_rate": 0.0002583109327983952, + "loss": 1.3385, + "theoretical_loss": 3.426232341062531, + "tokens_seen": 2245853184 + }, + { + "epoch": 0.68, + "learning_rate": 0.0002582948846539619, + "loss": 1.3251, + "theoretical_loss": 3.4262242615267127, + "tokens_seen": 2245918720 + }, + { + "epoch": 0.68, + "learning_rate": 0.0002582788365095286, + "loss": 1.34, + "theoretical_loss": 3.4262161822926633, + "tokens_seen": 2245984256 + }, + { + "epoch": 0.68, + "learning_rate": 0.0002582627883650953, + "loss": 1.3342, + "theoretical_loss": 3.4262081033603624, + "tokens_seen": 2246049792 + }, + { + "epoch": 0.68, + "learning_rate": 0.000258246740220662, + "loss": 1.2687, + "theoretical_loss": 3.42620002472979, + "tokens_seen": 2246115328 + }, + { + "epoch": 0.68, + "learning_rate": 0.0002582306920762287, + "loss": 1.421, + "theoretical_loss": 3.426191946400926, + "tokens_seen": 2246180864 + }, + { + "epoch": 0.68, + "learning_rate": 0.0002582146439317954, + "loss": 1.3226, + "theoretical_loss": 3.4261838683737507, + "tokens_seen": 2246246400 + }, + { + "epoch": 0.68, + "learning_rate": 0.0002581985957873621, + "loss": 1.3183, + "theoretical_loss": 3.4261757906482435, + "tokens_seen": 2246311936 + }, + { + "epoch": 0.68, + "learning_rate": 0.0002581825476429288, + "loss": 1.215, + "theoretical_loss": 3.4261677132243844, + "tokens_seen": 2246377472 + }, + { + "epoch": 0.68, + "learning_rate": 0.0002581664994984955, + "loss": 1.3661, + "theoretical_loss": 3.4261596361021533, + "tokens_seen": 2246443008 + }, + { + "epoch": 0.68, + "learning_rate": 0.0002581504513540622, + "loss": 1.2453, + "theoretical_loss": 3.426151559281531, + "tokens_seen": 2246508544 + }, + { + "epoch": 0.68, + "learning_rate": 0.0002581344032096289, + "loss": 1.2501, + "theoretical_loss": 3.4261434827624964, + "tokens_seen": 2246574080 + }, + { + "epoch": 0.68, + "learning_rate": 0.0002581183550651956, + "loss": 1.391, + "theoretical_loss": 3.42613540654503, + "tokens_seen": 2246639616 + }, + { + "epoch": 0.68, + "learning_rate": 0.0002581023069207623, + "loss": 1.3038, + "theoretical_loss": 3.4261273306291113, + "tokens_seen": 2246705152 + }, + { + "epoch": 0.68, + "learning_rate": 0.000258086258776329, + "loss": 1.2313, + "theoretical_loss": 3.4261192550147204, + "tokens_seen": 2246770688 + }, + { + "epoch": 0.68, + "learning_rate": 0.0002580702106318957, + "loss": 1.3063, + "theoretical_loss": 3.4261111797018375, + "tokens_seen": 2246836224 + }, + { + "epoch": 0.68, + "learning_rate": 0.0002580541624874624, + "loss": 1.2983, + "theoretical_loss": 3.4261031046904424, + "tokens_seen": 2246901760 + }, + { + "epoch": 0.68, + "learning_rate": 0.0002580381143430291, + "loss": 1.2945, + "theoretical_loss": 3.426095029980515, + "tokens_seen": 2246967296 + }, + { + "epoch": 0.68, + "learning_rate": 0.0002580220661985958, + "loss": 1.4237, + "theoretical_loss": 3.4260869555720355, + "tokens_seen": 2247032832 + }, + { + "epoch": 0.68, + "learning_rate": 0.0002580060180541625, + "loss": 1.3557, + "theoretical_loss": 3.4260788814649836, + "tokens_seen": 2247098368 + }, + { + "epoch": 0.68, + "learning_rate": 0.00025798996990972916, + "loss": 1.3271, + "theoretical_loss": 3.4260708076593396, + "tokens_seen": 2247163904 + }, + { + "epoch": 0.68, + "learning_rate": 0.00025797392176529586, + "loss": 1.1932, + "theoretical_loss": 3.4260627341550824, + "tokens_seen": 2247229440 + }, + { + "epoch": 0.68, + "learning_rate": 0.00025795787362086257, + "loss": 1.3639, + "theoretical_loss": 3.426054660952193, + "tokens_seen": 2247294976 + }, + { + "epoch": 0.68, + "learning_rate": 0.00025794182547642927, + "loss": 1.4439, + "theoretical_loss": 3.4260465880506517, + "tokens_seen": 2247360512 + }, + { + "epoch": 0.68, + "learning_rate": 0.00025792577733199597, + "loss": 1.2776, + "theoretical_loss": 3.426038515450437, + "tokens_seen": 2247426048 + }, + { + "epoch": 0.68, + "learning_rate": 0.00025790972918756267, + "loss": 1.3912, + "theoretical_loss": 3.4260304431515305, + "tokens_seen": 2247491584 + }, + { + "epoch": 0.68, + "learning_rate": 0.00025789368104312937, + "loss": 1.3366, + "theoretical_loss": 3.426022371153911, + "tokens_seen": 2247557120 + }, + { + "epoch": 0.68, + "learning_rate": 0.00025787763289869607, + "loss": 1.3271, + "theoretical_loss": 3.4260142994575586, + "tokens_seen": 2247622656 + }, + { + "epoch": 0.68, + "learning_rate": 0.00025786158475426277, + "loss": 1.339, + "theoretical_loss": 3.426006228062454, + "tokens_seen": 2247688192 + }, + { + "epoch": 0.68, + "learning_rate": 0.00025784553660982947, + "loss": 1.2482, + "theoretical_loss": 3.4259981569685767, + "tokens_seen": 2247753728 + }, + { + "epoch": 0.68, + "learning_rate": 0.00025782948846539617, + "loss": 1.2852, + "theoretical_loss": 3.425990086175906, + "tokens_seen": 2247819264 + }, + { + "epoch": 0.68, + "objective/train/docs_used": 818241, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.202354073524475, + "objective/train/theoretical_loss": 3.4259820156844234, + "objective/train/tokens_used": 2268344800, + "theoretical_loss": 3.4259820156844234, + "tokens_seen": 2247884800 + }, + { + "epoch": 0.68, + "learning_rate": 0.00025781344032096287, + "loss": 1.2183, + "theoretical_loss": 3.4259820156844234, + "tokens_seen": 2247884800 + }, + { + "epoch": 0.68, + "learning_rate": 0.00025779739217652957, + "loss": 1.3129, + "theoretical_loss": 3.425973945494108, + "tokens_seen": 2247950336 + }, + { + "epoch": 0.68, + "learning_rate": 0.00025778134403209627, + "loss": 1.3235, + "theoretical_loss": 3.425965875604939, + "tokens_seen": 2248015872 + }, + { + "epoch": 0.68, + "learning_rate": 0.00025776529588766297, + "loss": 1.3437, + "theoretical_loss": 3.425957806016898, + "tokens_seen": 2248081408 + }, + { + "epoch": 0.68, + "learning_rate": 0.0002577492477432297, + "loss": 1.3071, + "theoretical_loss": 3.425949736729964, + "tokens_seen": 2248146944 + }, + { + "epoch": 0.68, + "learning_rate": 0.0002577331995987964, + "loss": 1.328, + "theoretical_loss": 3.425941667744117, + "tokens_seen": 2248212480 + }, + { + "epoch": 0.68, + "learning_rate": 0.0002577171514543631, + "loss": 1.2426, + "theoretical_loss": 3.425933599059337, + "tokens_seen": 2248278016 + }, + { + "epoch": 0.68, + "learning_rate": 0.0002577011033099298, + "loss": 1.2978, + "theoretical_loss": 3.4259255306756042, + "tokens_seen": 2248343552 + }, + { + "epoch": 0.68, + "learning_rate": 0.0002576850551654965, + "loss": 1.3831, + "theoretical_loss": 3.425917462592899, + "tokens_seen": 2248409088 + }, + { + "epoch": 0.68, + "learning_rate": 0.0002576690070210632, + "loss": 1.2205, + "theoretical_loss": 3.4259093948112005, + "tokens_seen": 2248474624 + }, + { + "epoch": 0.68, + "learning_rate": 0.00025765295887662993, + "loss": 1.2825, + "theoretical_loss": 3.4259013273304895, + "tokens_seen": 2248540160 + }, + { + "epoch": 0.68, + "learning_rate": 0.00025763691073219663, + "loss": 1.3626, + "theoretical_loss": 3.4258932601507457, + "tokens_seen": 2248605696 + }, + { + "epoch": 0.68, + "learning_rate": 0.00025762086258776333, + "loss": 1.3462, + "theoretical_loss": 3.4258851932719487, + "tokens_seen": 2248671232 + }, + { + "epoch": 0.68, + "learning_rate": 0.00025760481444333003, + "loss": 1.3253, + "theoretical_loss": 3.425877126694079, + "tokens_seen": 2248736768 + }, + { + "epoch": 0.68, + "learning_rate": 0.00025758876629889673, + "loss": 1.2174, + "theoretical_loss": 3.4258690604171163, + "tokens_seen": 2248802304 + }, + { + "epoch": 0.68, + "learning_rate": 0.00025757271815446343, + "loss": 1.373, + "theoretical_loss": 3.425860994441041, + "tokens_seen": 2248867840 + }, + { + "epoch": 0.68, + "learning_rate": 0.00025755667001003014, + "loss": 1.3022, + "theoretical_loss": 3.425852928765833, + "tokens_seen": 2248933376 + }, + { + "epoch": 0.68, + "learning_rate": 0.00025754062186559684, + "loss": 1.3198, + "theoretical_loss": 3.4258448633914718, + "tokens_seen": 2248998912 + }, + { + "epoch": 0.68, + "learning_rate": 0.00025752457372116354, + "loss": 1.3183, + "theoretical_loss": 3.425836798317938, + "tokens_seen": 2249064448 + }, + { + "epoch": 0.68, + "learning_rate": 0.00025750852557673024, + "loss": 1.3145, + "theoretical_loss": 3.4258287335452113, + "tokens_seen": 2249129984 + }, + { + "epoch": 0.68, + "learning_rate": 0.00025749247743229694, + "loss": 1.2994, + "theoretical_loss": 3.4258206690732718, + "tokens_seen": 2249195520 + }, + { + "epoch": 0.68, + "learning_rate": 0.00025747642928786364, + "loss": 1.3325, + "theoretical_loss": 3.4258126049020996, + "tokens_seen": 2249261056 + }, + { + "epoch": 0.68, + "learning_rate": 0.00025746038114343034, + "loss": 1.3017, + "theoretical_loss": 3.425804541031675, + "tokens_seen": 2249326592 + }, + { + "epoch": 0.68, + "learning_rate": 0.00025744433299899704, + "loss": 1.2217, + "theoretical_loss": 3.425796477461977, + "tokens_seen": 2249392128 + }, + { + "epoch": 0.68, + "learning_rate": 0.00025742828485456374, + "loss": 1.3043, + "theoretical_loss": 3.425788414192987, + "tokens_seen": 2249457664 + }, + { + "epoch": 0.68, + "learning_rate": 0.00025741223671013044, + "loss": 1.2726, + "theoretical_loss": 3.4257803512246836, + "tokens_seen": 2249523200 + }, + { + "epoch": 0.68, + "learning_rate": 0.00025739618856569714, + "loss": 1.3817, + "theoretical_loss": 3.425772288557048, + "tokens_seen": 2249588736 + }, + { + "epoch": 0.68, + "learning_rate": 0.00025738014042126384, + "loss": 1.2807, + "theoretical_loss": 3.4257642261900596, + "tokens_seen": 2249654272 + }, + { + "epoch": 0.68, + "learning_rate": 0.0002573640922768305, + "loss": 1.3434, + "theoretical_loss": 3.425756164123699, + "tokens_seen": 2249719808 + }, + { + "epoch": 0.68, + "learning_rate": 0.0002573480441323972, + "loss": 1.2529, + "theoretical_loss": 3.4257481023579457, + "tokens_seen": 2249785344 + }, + { + "epoch": 0.68, + "learning_rate": 0.0002573319959879639, + "loss": 1.1794, + "theoretical_loss": 3.425740040892779, + "tokens_seen": 2249850880 + }, + { + "epoch": 0.68, + "learning_rate": 0.0002573159478435306, + "loss": 1.2995, + "theoretical_loss": 3.425731979728181, + "tokens_seen": 2249916416 + }, + { + "epoch": 0.68, + "learning_rate": 0.0002572998996990973, + "loss": 1.3553, + "theoretical_loss": 3.42572391886413, + "tokens_seen": 2249981952 + }, + { + "epoch": 0.68, + "learning_rate": 0.000257283851554664, + "loss": 1.3595, + "theoretical_loss": 3.425715858300607, + "tokens_seen": 2250047488 + }, + { + "epoch": 0.68, + "learning_rate": 0.0002572678034102307, + "loss": 1.3711, + "theoretical_loss": 3.425707798037591, + "tokens_seen": 2250113024 + }, + { + "epoch": 0.68, + "learning_rate": 0.0002572517552657974, + "loss": 1.3801, + "theoretical_loss": 3.425699738075063, + "tokens_seen": 2250178560 + }, + { + "epoch": 0.68, + "learning_rate": 0.0002572357071213641, + "loss": 1.3258, + "theoretical_loss": 3.425691678413003, + "tokens_seen": 2250244096 + }, + { + "epoch": 0.68, + "learning_rate": 0.0002572196589769308, + "loss": 1.3129, + "theoretical_loss": 3.425683619051391, + "tokens_seen": 2250309632 + }, + { + "epoch": 0.68, + "learning_rate": 0.0002572036108324975, + "loss": 1.3357, + "theoretical_loss": 3.4256755599902062, + "tokens_seen": 2250375168 + }, + { + "epoch": 0.68, + "learning_rate": 0.0002571875626880642, + "loss": 1.2698, + "theoretical_loss": 3.4256675012294293, + "tokens_seen": 2250440704 + }, + { + "epoch": 0.68, + "learning_rate": 0.0002571715145436309, + "loss": 1.3162, + "theoretical_loss": 3.425659442769041, + "tokens_seen": 2250506240 + }, + { + "epoch": 0.68, + "learning_rate": 0.0002571554663991976, + "loss": 1.3556, + "theoretical_loss": 3.42565138460902, + "tokens_seen": 2250571776 + }, + { + "epoch": 0.68, + "learning_rate": 0.0002571394182547643, + "loss": 1.3319, + "theoretical_loss": 3.4256433267493476, + "tokens_seen": 2250637312 + }, + { + "epoch": 0.68, + "learning_rate": 0.000257123370110331, + "loss": 1.358, + "theoretical_loss": 3.425635269190003, + "tokens_seen": 2250702848 + }, + { + "epoch": 0.68, + "learning_rate": 0.0002571073219658977, + "loss": 1.2781, + "theoretical_loss": 3.4256272119309665, + "tokens_seen": 2250768384 + }, + { + "epoch": 0.68, + "learning_rate": 0.0002570912738214644, + "loss": 1.4097, + "theoretical_loss": 3.4256191549722184, + "tokens_seen": 2250833920 + }, + { + "epoch": 0.68, + "learning_rate": 0.0002570752256770311, + "loss": 1.3216, + "theoretical_loss": 3.4256110983137384, + "tokens_seen": 2250899456 + }, + { + "epoch": 0.68, + "learning_rate": 0.0002570591775325978, + "loss": 1.2983, + "theoretical_loss": 3.4256030419555072, + "tokens_seen": 2250964992 + }, + { + "epoch": 0.68, + "learning_rate": 0.0002570431293881645, + "loss": 1.2323, + "theoretical_loss": 3.4255949858975043, + "tokens_seen": 2251030528 + }, + { + "epoch": 0.68, + "learning_rate": 0.0002570270812437312, + "loss": 1.2909, + "theoretical_loss": 3.4255869301397097, + "tokens_seen": 2251096064 + }, + { + "epoch": 0.68, + "objective/train/docs_used": 819526, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.4209468364715576, + "objective/train/theoretical_loss": 3.4255788746821034, + "objective/train/tokens_used": 2271621600, + "theoretical_loss": 3.4255788746821034, + "tokens_seen": 2251161600 + }, + { + "epoch": 0.68, + "learning_rate": 0.0002570110330992979, + "loss": 1.4005, + "theoretical_loss": 3.4255788746821034, + "tokens_seen": 2251161600 + }, + { + "epoch": 0.68, + "learning_rate": 0.0002569949849548646, + "loss": 1.278, + "theoretical_loss": 3.4255708195246664, + "tokens_seen": 2251227136 + }, + { + "epoch": 0.68, + "learning_rate": 0.0002569789368104313, + "loss": 1.369, + "theoretical_loss": 3.4255627646673776, + "tokens_seen": 2251292672 + }, + { + "epoch": 0.68, + "learning_rate": 0.000256962888665998, + "loss": 1.3503, + "theoretical_loss": 3.425554710110218, + "tokens_seen": 2251358208 + }, + { + "epoch": 0.68, + "learning_rate": 0.0002569468405215647, + "loss": 1.3314, + "theoretical_loss": 3.4255466558531675, + "tokens_seen": 2251423744 + }, + { + "epoch": 0.68, + "learning_rate": 0.0002569307923771314, + "loss": 1.2392, + "theoretical_loss": 3.4255386018962057, + "tokens_seen": 2251489280 + }, + { + "epoch": 0.68, + "learning_rate": 0.0002569147442326981, + "loss": 1.3689, + "theoretical_loss": 3.425530548239313, + "tokens_seen": 2251554816 + }, + { + "epoch": 0.68, + "learning_rate": 0.0002568986960882648, + "loss": 1.3592, + "theoretical_loss": 3.425522494882469, + "tokens_seen": 2251620352 + }, + { + "epoch": 0.68, + "learning_rate": 0.0002568826479438315, + "loss": 1.3204, + "theoretical_loss": 3.425514441825655, + "tokens_seen": 2251685888 + }, + { + "epoch": 0.68, + "learning_rate": 0.0002568665997993982, + "loss": 1.3209, + "theoretical_loss": 3.42550638906885, + "tokens_seen": 2251751424 + }, + { + "epoch": 0.68, + "learning_rate": 0.0002568505516549649, + "loss": 1.3038, + "theoretical_loss": 3.4254983366120344, + "tokens_seen": 2251816960 + }, + { + "epoch": 0.68, + "learning_rate": 0.0002568345035105316, + "loss": 1.3527, + "theoretical_loss": 3.425490284455188, + "tokens_seen": 2251882496 + }, + { + "epoch": 0.68, + "learning_rate": 0.0002568184553660983, + "loss": 1.3349, + "theoretical_loss": 3.425482232598292, + "tokens_seen": 2251948032 + }, + { + "epoch": 0.68, + "learning_rate": 0.000256802407221665, + "loss": 1.2658, + "theoretical_loss": 3.4254741810413254, + "tokens_seen": 2252013568 + }, + { + "epoch": 0.68, + "learning_rate": 0.0002567863590772317, + "loss": 1.368, + "theoretical_loss": 3.4254661297842683, + "tokens_seen": 2252079104 + }, + { + "epoch": 0.68, + "learning_rate": 0.0002567703109327984, + "loss": 1.3001, + "theoretical_loss": 3.4254580788271016, + "tokens_seen": 2252144640 + }, + { + "epoch": 0.68, + "learning_rate": 0.0002567542627883651, + "loss": 1.3354, + "theoretical_loss": 3.425450028169805, + "tokens_seen": 2252210176 + }, + { + "epoch": 0.68, + "learning_rate": 0.0002567382146439318, + "loss": 1.2975, + "theoretical_loss": 3.4254419778123575, + "tokens_seen": 2252275712 + }, + { + "epoch": 0.68, + "learning_rate": 0.0002567221664994985, + "loss": 1.2531, + "theoretical_loss": 3.425433927754741, + "tokens_seen": 2252341248 + }, + { + "epoch": 0.68, + "learning_rate": 0.0002567061183550652, + "loss": 1.2914, + "theoretical_loss": 3.425425877996935, + "tokens_seen": 2252406784 + }, + { + "epoch": 0.68, + "learning_rate": 0.0002566900702106319, + "loss": 1.2359, + "theoretical_loss": 3.4254178285389187, + "tokens_seen": 2252472320 + }, + { + "epoch": 0.68, + "learning_rate": 0.0002566740220661986, + "loss": 1.2436, + "theoretical_loss": 3.425409779380674, + "tokens_seen": 2252537856 + }, + { + "epoch": 0.68, + "learning_rate": 0.0002566579739217653, + "loss": 1.2985, + "theoretical_loss": 3.425401730522179, + "tokens_seen": 2252603392 + }, + { + "epoch": 0.68, + "learning_rate": 0.000256641925777332, + "loss": 1.2337, + "theoretical_loss": 3.4253936819634156, + "tokens_seen": 2252668928 + }, + { + "epoch": 0.68, + "learning_rate": 0.0002566258776328987, + "loss": 1.2513, + "theoretical_loss": 3.4253856337043627, + "tokens_seen": 2252734464 + }, + { + "epoch": 0.68, + "learning_rate": 0.0002566098294884654, + "loss": 1.3133, + "theoretical_loss": 3.425377585745001, + "tokens_seen": 2252800000 + }, + { + "epoch": 0.68, + "learning_rate": 0.00025659378134403206, + "loss": 1.2594, + "theoretical_loss": 3.4253695380853104, + "tokens_seen": 2252865536 + }, + { + "epoch": 0.68, + "learning_rate": 0.00025657773319959877, + "loss": 1.337, + "theoretical_loss": 3.4253614907252707, + "tokens_seen": 2252931072 + }, + { + "epoch": 0.68, + "learning_rate": 0.00025656168505516547, + "loss": 1.2931, + "theoretical_loss": 3.425353443664863, + "tokens_seen": 2252996608 + }, + { + "epoch": 0.68, + "learning_rate": 0.00025654563691073217, + "loss": 1.3424, + "theoretical_loss": 3.4253453969040666, + "tokens_seen": 2253062144 + }, + { + "epoch": 0.68, + "learning_rate": 0.00025652958876629887, + "loss": 1.3585, + "theoretical_loss": 3.425337350442862, + "tokens_seen": 2253127680 + }, + { + "epoch": 0.68, + "learning_rate": 0.00025651354062186557, + "loss": 1.2804, + "theoretical_loss": 3.425329304281229, + "tokens_seen": 2253193216 + }, + { + "epoch": 0.68, + "learning_rate": 0.00025649749247743227, + "loss": 1.2577, + "theoretical_loss": 3.4253212584191477, + "tokens_seen": 2253258752 + }, + { + "epoch": 0.68, + "learning_rate": 0.00025648144433299897, + "loss": 1.3565, + "theoretical_loss": 3.425313212856599, + "tokens_seen": 2253324288 + }, + { + "epoch": 0.68, + "learning_rate": 0.00025646539618856567, + "loss": 1.3613, + "theoretical_loss": 3.4253051675935624, + "tokens_seen": 2253389824 + }, + { + "epoch": 0.68, + "learning_rate": 0.0002564493480441324, + "loss": 1.3031, + "theoretical_loss": 3.425297122630018, + "tokens_seen": 2253455360 + }, + { + "epoch": 0.68, + "learning_rate": 0.0002564332998996991, + "loss": 1.5101, + "theoretical_loss": 3.4252890779659464, + "tokens_seen": 2253520896 + }, + { + "epoch": 0.68, + "learning_rate": 0.0002564172517552658, + "loss": 1.3876, + "theoretical_loss": 3.425281033601327, + "tokens_seen": 2253586432 + }, + { + "epoch": 0.68, + "learning_rate": 0.0002564012036108325, + "loss": 1.3524, + "theoretical_loss": 3.4252729895361407, + "tokens_seen": 2253651968 + }, + { + "epoch": 0.68, + "learning_rate": 0.00025638515546639923, + "loss": 1.446, + "theoretical_loss": 3.425264945770367, + "tokens_seen": 2253717504 + }, + { + "epoch": 0.68, + "learning_rate": 0.00025636910732196593, + "loss": 1.3239, + "theoretical_loss": 3.4252569023039867, + "tokens_seen": 2253783040 + }, + { + "epoch": 0.68, + "learning_rate": 0.00025635305917753263, + "loss": 1.3897, + "theoretical_loss": 3.4252488591369796, + "tokens_seen": 2253848576 + }, + { + "epoch": 0.68, + "learning_rate": 0.00025633701103309933, + "loss": 1.2752, + "theoretical_loss": 3.4252408162693255, + "tokens_seen": 2253914112 + }, + { + "epoch": 0.68, + "learning_rate": 0.00025632096288866603, + "loss": 1.3009, + "theoretical_loss": 3.4252327737010053, + "tokens_seen": 2253979648 + }, + { + "epoch": 0.68, + "learning_rate": 0.00025630491474423273, + "loss": 1.3188, + "theoretical_loss": 3.425224731431999, + "tokens_seen": 2254045184 + }, + { + "epoch": 0.68, + "learning_rate": 0.00025628886659979943, + "loss": 1.3324, + "theoretical_loss": 3.425216689462286, + "tokens_seen": 2254110720 + }, + { + "epoch": 0.68, + "learning_rate": 0.00025627281845536613, + "loss": 1.3007, + "theoretical_loss": 3.4252086477918473, + "tokens_seen": 2254176256 + }, + { + "epoch": 0.68, + "learning_rate": 0.00025625677031093283, + "loss": 1.3374, + "theoretical_loss": 3.425200606420663, + "tokens_seen": 2254241792 + }, + { + "epoch": 0.68, + "learning_rate": 0.00025624072216649953, + "loss": 1.2321, + "theoretical_loss": 3.4251925653487127, + "tokens_seen": 2254307328 + }, + { + "epoch": 0.68, + "learning_rate": 0.00025622467402206623, + "loss": 1.2251, + "theoretical_loss": 3.425184524575977, + "tokens_seen": 2254372864 + }, + { + "epoch": 0.68, + "objective/train/docs_used": 820753, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.230039358139038, + "objective/train/theoretical_loss": 3.425176484102436, + "objective/train/tokens_used": 2274898400, + "theoretical_loss": 3.425176484102436, + "tokens_seen": 2254438400 + }, + { + "epoch": 0.68, + "learning_rate": 0.00025620862587763293, + "loss": 1.2906, + "theoretical_loss": 3.425176484102436, + "tokens_seen": 2254438400 + }, + { + "epoch": 0.68, + "learning_rate": 0.00025619257773319964, + "loss": 1.3374, + "theoretical_loss": 3.4251684439280696, + "tokens_seen": 2254503936 + }, + { + "epoch": 0.68, + "learning_rate": 0.00025617652958876634, + "loss": 1.2866, + "theoretical_loss": 3.4251604040528587, + "tokens_seen": 2254569472 + }, + { + "epoch": 0.68, + "learning_rate": 0.00025616048144433304, + "loss": 1.269, + "theoretical_loss": 3.4251523644767827, + "tokens_seen": 2254635008 + }, + { + "epoch": 0.68, + "learning_rate": 0.00025614443329989974, + "loss": 1.309, + "theoretical_loss": 3.425144325199822, + "tokens_seen": 2254700544 + }, + { + "epoch": 0.68, + "learning_rate": 0.00025612838515546644, + "loss": 1.2671, + "theoretical_loss": 3.425136286221957, + "tokens_seen": 2254766080 + }, + { + "epoch": 0.68, + "learning_rate": 0.00025611233701103314, + "loss": 1.3723, + "theoretical_loss": 3.4251282475431677, + "tokens_seen": 2254831616 + }, + { + "epoch": 0.68, + "learning_rate": 0.00025609628886659984, + "loss": 1.2762, + "theoretical_loss": 3.4251202091634343, + "tokens_seen": 2254897152 + }, + { + "epoch": 0.68, + "learning_rate": 0.00025608024072216654, + "loss": 1.2425, + "theoretical_loss": 3.4251121710827372, + "tokens_seen": 2254962688 + }, + { + "epoch": 0.68, + "learning_rate": 0.00025606419257773324, + "loss": 1.4286, + "theoretical_loss": 3.425104133301056, + "tokens_seen": 2255028224 + }, + { + "epoch": 0.68, + "learning_rate": 0.00025604814443329994, + "loss": 1.3691, + "theoretical_loss": 3.4250960958183714, + "tokens_seen": 2255093760 + }, + { + "epoch": 0.68, + "learning_rate": 0.00025603209628886664, + "loss": 1.3669, + "theoretical_loss": 3.4250880586346635, + "tokens_seen": 2255159296 + }, + { + "epoch": 0.68, + "learning_rate": 0.00025601604814443334, + "loss": 1.2991, + "theoretical_loss": 3.4250800217499124, + "tokens_seen": 2255224832 + }, + { + "epoch": 0.68, + "learning_rate": 0.00025600000000000004, + "loss": 1.3225, + "theoretical_loss": 3.4250719851640983, + "tokens_seen": 2255290368 + }, + { + "epoch": 0.68, + "learning_rate": 0.00025598395185556674, + "loss": 1.3508, + "theoretical_loss": 3.4250639488772014, + "tokens_seen": 2255355904 + }, + { + "epoch": 0.68, + "learning_rate": 0.0002559679037111334, + "loss": 1.4118, + "theoretical_loss": 3.425055912889202, + "tokens_seen": 2255421440 + }, + { + "epoch": 0.68, + "learning_rate": 0.0002559518555667001, + "loss": 1.2184, + "theoretical_loss": 3.4250478772000803, + "tokens_seen": 2255486976 + }, + { + "epoch": 0.68, + "learning_rate": 0.0002559358074222668, + "loss": 1.3514, + "theoretical_loss": 3.4250398418098165, + "tokens_seen": 2255552512 + }, + { + "epoch": 0.68, + "learning_rate": 0.0002559197592778335, + "loss": 1.3013, + "theoretical_loss": 3.4250318067183905, + "tokens_seen": 2255618048 + }, + { + "epoch": 0.68, + "learning_rate": 0.0002559037111334002, + "loss": 1.3485, + "theoretical_loss": 3.4250237719257823, + "tokens_seen": 2255683584 + }, + { + "epoch": 0.68, + "learning_rate": 0.0002558876629889669, + "loss": 1.2802, + "theoretical_loss": 3.4250157374319734, + "tokens_seen": 2255749120 + }, + { + "epoch": 0.68, + "learning_rate": 0.0002558716148445336, + "loss": 1.1964, + "theoretical_loss": 3.4250077032369424, + "tokens_seen": 2255814656 + }, + { + "epoch": 0.68, + "learning_rate": 0.0002558555667001003, + "loss": 1.2423, + "theoretical_loss": 3.4249996693406706, + "tokens_seen": 2255880192 + }, + { + "epoch": 0.68, + "learning_rate": 0.000255839518555667, + "loss": 1.2933, + "theoretical_loss": 3.424991635743138, + "tokens_seen": 2255945728 + }, + { + "epoch": 0.68, + "learning_rate": 0.0002558234704112337, + "loss": 1.2652, + "theoretical_loss": 3.4249836024443248, + "tokens_seen": 2256011264 + }, + { + "epoch": 0.68, + "learning_rate": 0.0002558074222668004, + "loss": 1.227, + "theoretical_loss": 3.4249755694442108, + "tokens_seen": 2256076800 + }, + { + "epoch": 0.68, + "learning_rate": 0.0002557913741223671, + "loss": 1.3136, + "theoretical_loss": 3.4249675367427765, + "tokens_seen": 2256142336 + }, + { + "epoch": 0.68, + "learning_rate": 0.0002557753259779338, + "loss": 1.2741, + "theoretical_loss": 3.424959504340002, + "tokens_seen": 2256207872 + }, + { + "epoch": 0.68, + "learning_rate": 0.0002557592778335005, + "loss": 1.2915, + "theoretical_loss": 3.424951472235868, + "tokens_seen": 2256273408 + }, + { + "epoch": 0.68, + "learning_rate": 0.0002557432296890672, + "loss": 1.2992, + "theoretical_loss": 3.4249434404303543, + "tokens_seen": 2256338944 + }, + { + "epoch": 0.68, + "learning_rate": 0.0002557271815446339, + "loss": 1.3651, + "theoretical_loss": 3.424935408923441, + "tokens_seen": 2256404480 + }, + { + "epoch": 0.68, + "learning_rate": 0.0002557111334002006, + "loss": 1.2891, + "theoretical_loss": 3.4249273777151084, + "tokens_seen": 2256470016 + }, + { + "epoch": 0.68, + "learning_rate": 0.0002556950852557673, + "loss": 1.2754, + "theoretical_loss": 3.4249193468053374, + "tokens_seen": 2256535552 + }, + { + "epoch": 0.68, + "learning_rate": 0.000255679037111334, + "loss": 1.3406, + "theoretical_loss": 3.424911316194107, + "tokens_seen": 2256601088 + }, + { + "epoch": 0.68, + "learning_rate": 0.0002556629889669007, + "loss": 1.2509, + "theoretical_loss": 3.4249032858813986, + "tokens_seen": 2256666624 + }, + { + "epoch": 0.68, + "learning_rate": 0.0002556469408224674, + "loss": 1.1912, + "theoretical_loss": 3.4248952558671917, + "tokens_seen": 2256732160 + }, + { + "epoch": 0.68, + "learning_rate": 0.0002556308926780341, + "loss": 1.2947, + "theoretical_loss": 3.424887226151467, + "tokens_seen": 2256797696 + }, + { + "epoch": 0.68, + "learning_rate": 0.0002556148445336008, + "loss": 1.3649, + "theoretical_loss": 3.4248791967342043, + "tokens_seen": 2256863232 + }, + { + "epoch": 0.68, + "learning_rate": 0.0002555987963891675, + "loss": 1.258, + "theoretical_loss": 3.4248711676153842, + "tokens_seen": 2256928768 + }, + { + "epoch": 0.68, + "learning_rate": 0.0002555827482447342, + "loss": 1.2996, + "theoretical_loss": 3.424863138794987, + "tokens_seen": 2256994304 + }, + { + "epoch": 0.68, + "learning_rate": 0.0002555667001003009, + "loss": 1.3821, + "theoretical_loss": 3.4248551102729925, + "tokens_seen": 2257059840 + }, + { + "epoch": 0.68, + "learning_rate": 0.0002555506519558676, + "loss": 1.3707, + "theoretical_loss": 3.4248470820493813, + "tokens_seen": 2257125376 + }, + { + "epoch": 0.68, + "learning_rate": 0.0002555346038114343, + "loss": 1.4314, + "theoretical_loss": 3.424839054124133, + "tokens_seen": 2257190912 + }, + { + "epoch": 0.68, + "learning_rate": 0.000255518555667001, + "loss": 1.3884, + "theoretical_loss": 3.424831026497229, + "tokens_seen": 2257256448 + }, + { + "epoch": 0.68, + "learning_rate": 0.0002555025075225677, + "loss": 1.3464, + "theoretical_loss": 3.424822999168649, + "tokens_seen": 2257321984 + }, + { + "epoch": 0.68, + "learning_rate": 0.0002554864593781344, + "loss": 1.3996, + "theoretical_loss": 3.424814972138373, + "tokens_seen": 2257387520 + }, + { + "epoch": 0.68, + "learning_rate": 0.0002554704112337011, + "loss": 1.3503, + "theoretical_loss": 3.4248069454063814, + "tokens_seen": 2257453056 + }, + { + "epoch": 0.68, + "learning_rate": 0.0002554543630892678, + "loss": 1.3081, + "theoretical_loss": 3.4247989189726544, + "tokens_seen": 2257518592 + }, + { + "epoch": 0.68, + "learning_rate": 0.0002554383149448345, + "loss": 1.2641, + "theoretical_loss": 3.4247908928371724, + "tokens_seen": 2257584128 + }, + { + "epoch": 0.68, + "learning_rate": 0.0002554222668004012, + "loss": 1.3792, + "theoretical_loss": 3.424782866999916, + "tokens_seen": 2257649664 + }, + { + "epoch": 0.68, + "objective/train/docs_used": 822060, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.2137627601623535, + "objective/train/theoretical_loss": 3.4247748414608647, + "objective/train/tokens_used": 2278175200, + "theoretical_loss": 3.4247748414608647, + "tokens_seen": 2257715200 + }, + { + "epoch": 0.68, + "learning_rate": 0.0002554062186559679, + "loss": 1.2598, + "theoretical_loss": 3.4247748414608647, + "tokens_seen": 2257715200 + }, + { + "epoch": 0.68, + "learning_rate": 0.0002553901705115346, + "loss": 1.2661, + "theoretical_loss": 3.424766816219999, + "tokens_seen": 2257780736 + }, + { + "epoch": 0.68, + "learning_rate": 0.0002553741223671013, + "loss": 1.2132, + "theoretical_loss": 3.4247587912773, + "tokens_seen": 2257846272 + }, + { + "epoch": 0.68, + "learning_rate": 0.000255358074222668, + "loss": 1.3233, + "theoretical_loss": 3.424750766632747, + "tokens_seen": 2257911808 + }, + { + "epoch": 0.68, + "learning_rate": 0.0002553420260782347, + "loss": 1.3258, + "theoretical_loss": 3.42474274228632, + "tokens_seen": 2257977344 + }, + { + "epoch": 0.68, + "learning_rate": 0.0002553259779338014, + "loss": 1.3087, + "theoretical_loss": 3.4247347182380006, + "tokens_seen": 2258042880 + }, + { + "epoch": 0.68, + "learning_rate": 0.0002553099297893681, + "loss": 1.3712, + "theoretical_loss": 3.424726694487768, + "tokens_seen": 2258108416 + }, + { + "epoch": 0.68, + "learning_rate": 0.0002552938816449348, + "loss": 1.3576, + "theoretical_loss": 3.4247186710356026, + "tokens_seen": 2258173952 + }, + { + "epoch": 0.68, + "learning_rate": 0.0002552778335005015, + "loss": 1.2594, + "theoretical_loss": 3.424710647881485, + "tokens_seen": 2258239488 + }, + { + "epoch": 0.68, + "learning_rate": 0.0002552617853560682, + "loss": 1.3209, + "theoretical_loss": 3.4247026250253954, + "tokens_seen": 2258305024 + }, + { + "epoch": 0.68, + "learning_rate": 0.0002552457372116349, + "loss": 1.2405, + "theoretical_loss": 3.424694602467314, + "tokens_seen": 2258370560 + }, + { + "epoch": 0.68, + "learning_rate": 0.0002552296890672016, + "loss": 1.364, + "theoretical_loss": 3.424686580207221, + "tokens_seen": 2258436096 + }, + { + "epoch": 0.68, + "learning_rate": 0.0002552136409227683, + "loss": 1.3557, + "theoretical_loss": 3.4246785582450965, + "tokens_seen": 2258501632 + }, + { + "epoch": 0.68, + "learning_rate": 0.000255197592778335, + "loss": 1.3748, + "theoretical_loss": 3.424670536580922, + "tokens_seen": 2258567168 + }, + { + "epoch": 0.68, + "learning_rate": 0.0002551815446339017, + "loss": 1.329, + "theoretical_loss": 3.424662515214676, + "tokens_seen": 2258632704 + }, + { + "epoch": 0.68, + "learning_rate": 0.0002551654964894684, + "loss": 1.3965, + "theoretical_loss": 3.42465449414634, + "tokens_seen": 2258698240 + }, + { + "epoch": 0.68, + "learning_rate": 0.0002551494483450351, + "loss": 1.3559, + "theoretical_loss": 3.4246464733758937, + "tokens_seen": 2258763776 + }, + { + "epoch": 0.68, + "learning_rate": 0.0002551334002006018, + "loss": 1.4279, + "theoretical_loss": 3.424638452903318, + "tokens_seen": 2258829312 + }, + { + "epoch": 0.68, + "learning_rate": 0.0002551173520561685, + "loss": 1.3059, + "theoretical_loss": 3.4246304327285926, + "tokens_seen": 2258894848 + }, + { + "epoch": 0.68, + "learning_rate": 0.0002551013039117352, + "loss": 1.3085, + "theoretical_loss": 3.424622412851698, + "tokens_seen": 2258960384 + }, + { + "epoch": 0.68, + "learning_rate": 0.0002550852557673019, + "loss": 1.4004, + "theoretical_loss": 3.4246143932726145, + "tokens_seen": 2259025920 + }, + { + "epoch": 0.68, + "learning_rate": 0.0002550692076228686, + "loss": 1.3283, + "theoretical_loss": 3.424606373991322, + "tokens_seen": 2259091456 + }, + { + "epoch": 0.68, + "learning_rate": 0.0002550531594784353, + "loss": 1.3992, + "theoretical_loss": 3.4245983550078023, + "tokens_seen": 2259156992 + }, + { + "epoch": 0.68, + "learning_rate": 0.000255037111334002, + "loss": 1.2718, + "theoretical_loss": 3.424590336322034, + "tokens_seen": 2259222528 + }, + { + "epoch": 0.68, + "learning_rate": 0.00025502106318956873, + "loss": 1.3159, + "theoretical_loss": 3.4245823179339983, + "tokens_seen": 2259288064 + }, + { + "epoch": 0.68, + "learning_rate": 0.00025500501504513543, + "loss": 1.3331, + "theoretical_loss": 3.424574299843675, + "tokens_seen": 2259353600 + }, + { + "epoch": 0.68, + "learning_rate": 0.00025498896690070213, + "loss": 1.2887, + "theoretical_loss": 3.4245662820510443, + "tokens_seen": 2259419136 + }, + { + "epoch": 0.68, + "learning_rate": 0.00025497291875626883, + "loss": 1.3537, + "theoretical_loss": 3.4245582645560875, + "tokens_seen": 2259484672 + }, + { + "epoch": 0.68, + "learning_rate": 0.00025495687061183553, + "loss": 1.4086, + "theoretical_loss": 3.424550247358784, + "tokens_seen": 2259550208 + }, + { + "epoch": 0.68, + "learning_rate": 0.00025494082246740223, + "loss": 1.3367, + "theoretical_loss": 3.4245422304591147, + "tokens_seen": 2259615744 + }, + { + "epoch": 0.68, + "learning_rate": 0.00025492477432296893, + "loss": 1.3667, + "theoretical_loss": 3.4245342138570596, + "tokens_seen": 2259681280 + }, + { + "epoch": 0.68, + "learning_rate": 0.00025490872617853563, + "loss": 1.2607, + "theoretical_loss": 3.424526197552599, + "tokens_seen": 2259746816 + }, + { + "epoch": 0.68, + "learning_rate": 0.00025489267803410233, + "loss": 1.2329, + "theoretical_loss": 3.424518181545713, + "tokens_seen": 2259812352 + }, + { + "epoch": 0.68, + "learning_rate": 0.00025487662988966903, + "loss": 1.4165, + "theoretical_loss": 3.4245101658363826, + "tokens_seen": 2259877888 + }, + { + "epoch": 0.68, + "learning_rate": 0.00025486058174523573, + "loss": 1.3948, + "theoretical_loss": 3.4245021504245874, + "tokens_seen": 2259943424 + }, + { + "epoch": 0.68, + "learning_rate": 0.00025484453360080243, + "loss": 1.2053, + "theoretical_loss": 3.424494135310308, + "tokens_seen": 2260008960 + }, + { + "epoch": 0.68, + "learning_rate": 0.00025482848545636913, + "loss": 1.2626, + "theoretical_loss": 3.4244861204935253, + "tokens_seen": 2260074496 + }, + { + "epoch": 0.68, + "learning_rate": 0.00025481243731193584, + "loss": 1.2792, + "theoretical_loss": 3.4244781059742184, + "tokens_seen": 2260140032 + }, + { + "epoch": 0.68, + "learning_rate": 0.00025479638916750254, + "loss": 1.2753, + "theoretical_loss": 3.424470091752369, + "tokens_seen": 2260205568 + }, + { + "epoch": 0.68, + "learning_rate": 0.00025478034102306924, + "loss": 1.1638, + "theoretical_loss": 3.4244620778279566, + "tokens_seen": 2260271104 + }, + { + "epoch": 0.68, + "learning_rate": 0.00025476429287863594, + "loss": 1.4794, + "theoretical_loss": 3.4244540642009613, + "tokens_seen": 2260336640 + }, + { + "epoch": 0.68, + "learning_rate": 0.00025474824473420264, + "loss": 1.31, + "theoretical_loss": 3.4244460508713646, + "tokens_seen": 2260402176 + }, + { + "epoch": 0.68, + "learning_rate": 0.00025473219658976934, + "loss": 1.3385, + "theoretical_loss": 3.4244380378391455, + "tokens_seen": 2260467712 + }, + { + "epoch": 0.69, + "learning_rate": 0.00025471614844533604, + "loss": 1.1809, + "theoretical_loss": 3.424430025104285, + "tokens_seen": 2260533248 + }, + { + "epoch": 0.69, + "learning_rate": 0.00025470010030090274, + "loss": 1.2838, + "theoretical_loss": 3.4244220126667635, + "tokens_seen": 2260598784 + }, + { + "epoch": 0.69, + "learning_rate": 0.00025468405215646944, + "loss": 1.2905, + "theoretical_loss": 3.4244140005265615, + "tokens_seen": 2260664320 + }, + { + "epoch": 0.69, + "learning_rate": 0.00025466800401203614, + "loss": 1.3742, + "theoretical_loss": 3.4244059886836586, + "tokens_seen": 2260729856 + }, + { + "epoch": 0.69, + "learning_rate": 0.00025465195586760284, + "loss": 1.3952, + "theoretical_loss": 3.424397977138036, + "tokens_seen": 2260795392 + }, + { + "epoch": 0.69, + "learning_rate": 0.00025463590772316954, + "loss": 1.3377, + "theoretical_loss": 3.4243899658896737, + "tokens_seen": 2260860928 + }, + { + "epoch": 0.69, + "learning_rate": 0.00025461985957873624, + "loss": 1.282, + "theoretical_loss": 3.4243819549385517, + "tokens_seen": 2260926464 + }, + { + "epoch": 0.69, + "objective/train/docs_used": 823330, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.2914601564407349, + "objective/train/theoretical_loss": 3.424373944284651, + "objective/train/tokens_used": 2281452000, + "theoretical_loss": 3.424373944284651, + "tokens_seen": 2260992000 + }, + { + "epoch": 0.69, + "learning_rate": 0.00025460381143430294, + "loss": 1.3142, + "theoretical_loss": 3.424373944284651, + "tokens_seen": 2260992000 + }, + { + "epoch": 0.69, + "learning_rate": 0.00025458776328986964, + "loss": 1.3131, + "theoretical_loss": 3.4243659339279517, + "tokens_seen": 2261057536 + }, + { + "epoch": 0.69, + "learning_rate": 0.00025457171514543634, + "loss": 1.3459, + "theoretical_loss": 3.424357923868434, + "tokens_seen": 2261123072 + }, + { + "epoch": 0.69, + "learning_rate": 0.000254555667001003, + "loss": 1.3474, + "theoretical_loss": 3.4243499141060783, + "tokens_seen": 2261188608 + }, + { + "epoch": 0.69, + "learning_rate": 0.0002545396188565697, + "loss": 1.378, + "theoretical_loss": 3.4243419046408654, + "tokens_seen": 2261254144 + }, + { + "epoch": 0.69, + "learning_rate": 0.0002545235707121364, + "loss": 1.2221, + "theoretical_loss": 3.424333895472775, + "tokens_seen": 2261319680 + }, + { + "epoch": 0.69, + "learning_rate": 0.0002545075225677031, + "loss": 1.4658, + "theoretical_loss": 3.4243258866017876, + "tokens_seen": 2261385216 + }, + { + "epoch": 0.69, + "learning_rate": 0.0002544914744232698, + "loss": 1.3773, + "theoretical_loss": 3.4243178780278836, + "tokens_seen": 2261450752 + }, + { + "epoch": 0.69, + "learning_rate": 0.0002544754262788365, + "loss": 1.43, + "theoretical_loss": 3.424309869751044, + "tokens_seen": 2261516288 + }, + { + "epoch": 0.69, + "learning_rate": 0.0002544593781344032, + "loss": 1.2931, + "theoretical_loss": 3.424301861771249, + "tokens_seen": 2261581824 + }, + { + "epoch": 0.69, + "learning_rate": 0.0002544433299899699, + "loss": 1.3403, + "theoretical_loss": 3.4242938540884778, + "tokens_seen": 2261647360 + }, + { + "epoch": 0.69, + "learning_rate": 0.0002544272818455366, + "loss": 1.2945, + "theoretical_loss": 3.424285846702712, + "tokens_seen": 2261712896 + }, + { + "epoch": 0.69, + "learning_rate": 0.0002544112337011033, + "loss": 1.3042, + "theoretical_loss": 3.4242778396139317, + "tokens_seen": 2261778432 + }, + { + "epoch": 0.69, + "learning_rate": 0.00025439518555667, + "loss": 1.4003, + "theoretical_loss": 3.4242698328221173, + "tokens_seen": 2261843968 + }, + { + "epoch": 0.69, + "learning_rate": 0.0002543791374122367, + "loss": 1.346, + "theoretical_loss": 3.424261826327249, + "tokens_seen": 2261909504 + }, + { + "epoch": 0.69, + "learning_rate": 0.0002543630892678034, + "loss": 1.3392, + "theoretical_loss": 3.424253820129307, + "tokens_seen": 2261975040 + }, + { + "epoch": 0.69, + "learning_rate": 0.0002543470411233701, + "loss": 1.4129, + "theoretical_loss": 3.4242458142282723, + "tokens_seen": 2262040576 + }, + { + "epoch": 0.69, + "learning_rate": 0.0002543309929789368, + "loss": 1.3446, + "theoretical_loss": 3.4242378086241247, + "tokens_seen": 2262106112 + }, + { + "epoch": 0.69, + "learning_rate": 0.0002543149448345035, + "loss": 1.3112, + "theoretical_loss": 3.424229803316845, + "tokens_seen": 2262171648 + }, + { + "epoch": 0.69, + "learning_rate": 0.0002542988966900702, + "loss": 1.4327, + "theoretical_loss": 3.424221798306413, + "tokens_seen": 2262237184 + }, + { + "epoch": 0.69, + "learning_rate": 0.0002542828485456369, + "loss": 1.3238, + "theoretical_loss": 3.42421379359281, + "tokens_seen": 2262302720 + }, + { + "epoch": 0.69, + "learning_rate": 0.0002542668004012036, + "loss": 1.4058, + "theoretical_loss": 3.424205789176016, + "tokens_seen": 2262368256 + }, + { + "epoch": 0.69, + "learning_rate": 0.0002542507522567703, + "loss": 1.357, + "theoretical_loss": 3.424197785056011, + "tokens_seen": 2262433792 + }, + { + "epoch": 0.69, + "learning_rate": 0.000254234704112337, + "loss": 1.2522, + "theoretical_loss": 3.4241897812327755, + "tokens_seen": 2262499328 + }, + { + "epoch": 0.69, + "learning_rate": 0.0002542186559679037, + "loss": 1.4475, + "theoretical_loss": 3.4241817777062904, + "tokens_seen": 2262564864 + }, + { + "epoch": 0.69, + "learning_rate": 0.0002542026078234704, + "loss": 1.3041, + "theoretical_loss": 3.424173774476536, + "tokens_seen": 2262630400 + }, + { + "epoch": 0.69, + "learning_rate": 0.0002541865596790371, + "loss": 1.412, + "theoretical_loss": 3.4241657715434917, + "tokens_seen": 2262695936 + }, + { + "epoch": 0.69, + "learning_rate": 0.0002541705115346038, + "loss": 1.2742, + "theoretical_loss": 3.4241577689071394, + "tokens_seen": 2262761472 + }, + { + "epoch": 0.69, + "learning_rate": 0.0002541544633901705, + "loss": 1.2901, + "theoretical_loss": 3.4241497665674587, + "tokens_seen": 2262827008 + }, + { + "epoch": 0.69, + "learning_rate": 0.0002541384152457372, + "loss": 1.4175, + "theoretical_loss": 3.42414176452443, + "tokens_seen": 2262892544 + }, + { + "epoch": 0.69, + "learning_rate": 0.0002541223671013039, + "loss": 1.3238, + "theoretical_loss": 3.424133762778034, + "tokens_seen": 2262958080 + }, + { + "epoch": 0.69, + "learning_rate": 0.0002541063189568706, + "loss": 1.4212, + "theoretical_loss": 3.424125761328251, + "tokens_seen": 2263023616 + }, + { + "epoch": 0.69, + "learning_rate": 0.0002540902708124373, + "loss": 1.2331, + "theoretical_loss": 3.4241177601750614, + "tokens_seen": 2263089152 + }, + { + "epoch": 0.69, + "learning_rate": 0.000254074222668004, + "loss": 1.3217, + "theoretical_loss": 3.4241097593184455, + "tokens_seen": 2263154688 + }, + { + "epoch": 0.69, + "learning_rate": 0.0002540581745235707, + "loss": 1.3589, + "theoretical_loss": 3.424101758758384, + "tokens_seen": 2263220224 + }, + { + "epoch": 0.69, + "learning_rate": 0.00025404212637913747, + "loss": 1.4166, + "theoretical_loss": 3.424093758494857, + "tokens_seen": 2263285760 + }, + { + "epoch": 0.69, + "learning_rate": 0.00025402607823470417, + "loss": 1.3557, + "theoretical_loss": 3.424085758527845, + "tokens_seen": 2263351296 + }, + { + "epoch": 0.69, + "learning_rate": 0.00025401003009027087, + "loss": 1.3701, + "theoretical_loss": 3.4240777588573286, + "tokens_seen": 2263416832 + }, + { + "epoch": 0.69, + "learning_rate": 0.00025399398194583757, + "loss": 1.2685, + "theoretical_loss": 3.424069759483288, + "tokens_seen": 2263482368 + }, + { + "epoch": 0.69, + "learning_rate": 0.00025397793380140427, + "loss": 1.2669, + "theoretical_loss": 3.4240617604057038, + "tokens_seen": 2263547904 + }, + { + "epoch": 0.69, + "learning_rate": 0.00025396188565697097, + "loss": 1.3196, + "theoretical_loss": 3.424053761624556, + "tokens_seen": 2263613440 + }, + { + "epoch": 0.69, + "learning_rate": 0.00025394583751253767, + "loss": 1.3601, + "theoretical_loss": 3.424045763139826, + "tokens_seen": 2263678976 + }, + { + "epoch": 0.69, + "learning_rate": 0.0002539297893681043, + "loss": 1.4019, + "theoretical_loss": 3.4240377649514935, + "tokens_seen": 2263744512 + }, + { + "epoch": 0.69, + "learning_rate": 0.000253913741223671, + "loss": 1.3313, + "theoretical_loss": 3.4240297670595385, + "tokens_seen": 2263810048 + }, + { + "epoch": 0.69, + "learning_rate": 0.0002538976930792377, + "loss": 1.2759, + "theoretical_loss": 3.4240217694639425, + "tokens_seen": 2263875584 + }, + { + "epoch": 0.69, + "learning_rate": 0.0002538816449348044, + "loss": 1.3217, + "theoretical_loss": 3.4240137721646855, + "tokens_seen": 2263941120 + }, + { + "epoch": 0.69, + "learning_rate": 0.0002538655967903711, + "loss": 1.3053, + "theoretical_loss": 3.424005775161748, + "tokens_seen": 2264006656 + }, + { + "epoch": 0.69, + "learning_rate": 0.0002538495486459378, + "loss": 1.2288, + "theoretical_loss": 3.42399777845511, + "tokens_seen": 2264072192 + }, + { + "epoch": 0.69, + "learning_rate": 0.0002538335005015045, + "loss": 1.2877, + "theoretical_loss": 3.4239897820447522, + "tokens_seen": 2264137728 + }, + { + "epoch": 0.69, + "learning_rate": 0.0002538174523570712, + "loss": 1.3183, + "theoretical_loss": 3.4239817859306556, + "tokens_seen": 2264203264 + }, + { + "epoch": 0.69, + "objective/train/docs_used": 824641, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.4494963884353638, + "objective/train/theoretical_loss": 3.4239737901128, + "objective/train/tokens_used": 2284728800, + "theoretical_loss": 3.4239737901128, + "tokens_seen": 2264268800 + }, + { + "epoch": 0.69, + "learning_rate": 0.0002538014042126379, + "loss": 1.3373, + "theoretical_loss": 3.4239737901128, + "tokens_seen": 2264268800 + }, + { + "epoch": 0.69, + "learning_rate": 0.0002537853560682046, + "loss": 1.3266, + "theoretical_loss": 3.423965794591166, + "tokens_seen": 2264334336 + }, + { + "epoch": 0.69, + "learning_rate": 0.0002537693079237713, + "loss": 1.477, + "theoretical_loss": 3.423957799365734, + "tokens_seen": 2264399872 + }, + { + "epoch": 0.69, + "learning_rate": 0.000253753259779338, + "loss": 1.2869, + "theoretical_loss": 3.4239498044364844, + "tokens_seen": 2264465408 + }, + { + "epoch": 0.69, + "learning_rate": 0.0002537372116349047, + "loss": 1.3853, + "theoretical_loss": 3.423941809803398, + "tokens_seen": 2264530944 + }, + { + "epoch": 0.69, + "learning_rate": 0.0002537211634904714, + "loss": 1.2494, + "theoretical_loss": 3.423933815466455, + "tokens_seen": 2264596480 + }, + { + "epoch": 0.69, + "learning_rate": 0.0002537051153460381, + "loss": 1.2697, + "theoretical_loss": 3.423925821425636, + "tokens_seen": 2264662016 + }, + { + "epoch": 0.69, + "learning_rate": 0.0002536890672016048, + "loss": 1.2417, + "theoretical_loss": 3.4239178276809215, + "tokens_seen": 2264727552 + }, + { + "epoch": 0.69, + "learning_rate": 0.0002536730190571715, + "loss": 1.2964, + "theoretical_loss": 3.4239098342322913, + "tokens_seen": 2264793088 + }, + { + "epoch": 0.69, + "learning_rate": 0.0002536569709127382, + "loss": 1.2983, + "theoretical_loss": 3.4239018410797266, + "tokens_seen": 2264858624 + }, + { + "epoch": 0.69, + "learning_rate": 0.00025364092276830493, + "loss": 1.3241, + "theoretical_loss": 3.4238938482232077, + "tokens_seen": 2264924160 + }, + { + "epoch": 0.69, + "learning_rate": 0.00025362487462387163, + "loss": 1.2135, + "theoretical_loss": 3.423885855662715, + "tokens_seen": 2264989696 + }, + { + "epoch": 0.69, + "learning_rate": 0.00025360882647943833, + "loss": 1.243, + "theoretical_loss": 3.4238778633982294, + "tokens_seen": 2265055232 + }, + { + "epoch": 0.69, + "learning_rate": 0.00025359277833500503, + "loss": 1.3194, + "theoretical_loss": 3.423869871429731, + "tokens_seen": 2265120768 + }, + { + "epoch": 0.69, + "learning_rate": 0.00025357673019057173, + "loss": 1.4564, + "theoretical_loss": 3.4238618797572, + "tokens_seen": 2265186304 + }, + { + "epoch": 0.69, + "learning_rate": 0.00025356068204613843, + "loss": 1.2349, + "theoretical_loss": 3.4238538883806173, + "tokens_seen": 2265251840 + }, + { + "epoch": 0.69, + "learning_rate": 0.00025354463390170513, + "loss": 1.2815, + "theoretical_loss": 3.423845897299963, + "tokens_seen": 2265317376 + }, + { + "epoch": 0.69, + "learning_rate": 0.00025352858575727183, + "loss": 1.3492, + "theoretical_loss": 3.423837906515218, + "tokens_seen": 2265382912 + }, + { + "epoch": 0.69, + "learning_rate": 0.00025351253761283853, + "loss": 1.3419, + "theoretical_loss": 3.423829916026362, + "tokens_seen": 2265448448 + }, + { + "epoch": 0.69, + "learning_rate": 0.00025349648946840523, + "loss": 1.3876, + "theoretical_loss": 3.423821925833377, + "tokens_seen": 2265513984 + }, + { + "epoch": 0.69, + "learning_rate": 0.00025348044132397193, + "loss": 1.2441, + "theoretical_loss": 3.423813935936242, + "tokens_seen": 2265579520 + }, + { + "epoch": 0.69, + "learning_rate": 0.00025346439317953863, + "loss": 1.3666, + "theoretical_loss": 3.4238059463349377, + "tokens_seen": 2265645056 + }, + { + "epoch": 0.69, + "learning_rate": 0.00025344834503510534, + "loss": 1.47, + "theoretical_loss": 3.4237979570294454, + "tokens_seen": 2265710592 + }, + { + "epoch": 0.69, + "learning_rate": 0.00025343229689067204, + "loss": 1.3611, + "theoretical_loss": 3.423789968019745, + "tokens_seen": 2265776128 + }, + { + "epoch": 0.69, + "learning_rate": 0.00025341624874623874, + "loss": 1.2741, + "theoretical_loss": 3.4237819793058173, + "tokens_seen": 2265841664 + }, + { + "epoch": 0.69, + "learning_rate": 0.00025340020060180544, + "loss": 1.3371, + "theoretical_loss": 3.4237739908876423, + "tokens_seen": 2265907200 + }, + { + "epoch": 0.69, + "learning_rate": 0.00025338415245737214, + "loss": 1.2922, + "theoretical_loss": 3.423766002765201, + "tokens_seen": 2265972736 + }, + { + "epoch": 0.69, + "learning_rate": 0.00025336810431293884, + "loss": 1.3062, + "theoretical_loss": 3.423758014938474, + "tokens_seen": 2266038272 + }, + { + "epoch": 0.69, + "learning_rate": 0.00025335205616850554, + "loss": 1.3478, + "theoretical_loss": 3.423750027407441, + "tokens_seen": 2266103808 + }, + { + "epoch": 0.69, + "learning_rate": 0.00025333600802407224, + "loss": 1.3284, + "theoretical_loss": 3.4237420401720833, + "tokens_seen": 2266169344 + }, + { + "epoch": 0.69, + "learning_rate": 0.00025331995987963894, + "loss": 1.3354, + "theoretical_loss": 3.423734053232381, + "tokens_seen": 2266234880 + }, + { + "epoch": 0.69, + "learning_rate": 0.00025330391173520564, + "loss": 1.3116, + "theoretical_loss": 3.4237260665883147, + "tokens_seen": 2266300416 + }, + { + "epoch": 0.69, + "learning_rate": 0.00025328786359077234, + "loss": 1.4215, + "theoretical_loss": 3.423718080239865, + "tokens_seen": 2266365952 + }, + { + "epoch": 0.69, + "learning_rate": 0.00025327181544633904, + "loss": 1.275, + "theoretical_loss": 3.423710094187012, + "tokens_seen": 2266431488 + }, + { + "epoch": 0.69, + "learning_rate": 0.00025325576730190574, + "loss": 1.3346, + "theoretical_loss": 3.423702108429737, + "tokens_seen": 2266497024 + }, + { + "epoch": 0.69, + "learning_rate": 0.00025323971915747244, + "loss": 1.31, + "theoretical_loss": 3.42369412296802, + "tokens_seen": 2266562560 + }, + { + "epoch": 0.69, + "learning_rate": 0.00025322367101303914, + "loss": 1.29, + "theoretical_loss": 3.4236861378018415, + "tokens_seen": 2266628096 + }, + { + "epoch": 0.69, + "learning_rate": 0.00025320762286860584, + "loss": 1.3433, + "theoretical_loss": 3.423678152931182, + "tokens_seen": 2266693632 + }, + { + "epoch": 0.69, + "learning_rate": 0.00025319157472417255, + "loss": 1.2529, + "theoretical_loss": 3.4236701683560224, + "tokens_seen": 2266759168 + }, + { + "epoch": 0.69, + "learning_rate": 0.00025317552657973925, + "loss": 1.4046, + "theoretical_loss": 3.423662184076343, + "tokens_seen": 2266824704 + }, + { + "epoch": 0.69, + "learning_rate": 0.0002531594784353059, + "loss": 1.3622, + "theoretical_loss": 3.4236542000921237, + "tokens_seen": 2266890240 + }, + { + "epoch": 0.69, + "learning_rate": 0.0002531434302908726, + "loss": 1.3152, + "theoretical_loss": 3.4236462164033457, + "tokens_seen": 2266955776 + }, + { + "epoch": 0.69, + "learning_rate": 0.0002531273821464393, + "loss": 1.285, + "theoretical_loss": 3.42363823300999, + "tokens_seen": 2267021312 + }, + { + "epoch": 0.69, + "learning_rate": 0.000253111334002006, + "loss": 1.3658, + "theoretical_loss": 3.423630249912036, + "tokens_seen": 2267086848 + }, + { + "epoch": 0.69, + "learning_rate": 0.0002530952858575727, + "loss": 1.3125, + "theoretical_loss": 3.423622267109465, + "tokens_seen": 2267152384 + }, + { + "epoch": 0.69, + "learning_rate": 0.0002530792377131394, + "loss": 1.3279, + "theoretical_loss": 3.423614284602257, + "tokens_seen": 2267217920 + }, + { + "epoch": 0.69, + "learning_rate": 0.0002530631895687061, + "loss": 1.2763, + "theoretical_loss": 3.423606302390393, + "tokens_seen": 2267283456 + }, + { + "epoch": 0.69, + "learning_rate": 0.0002530471414242728, + "loss": 1.4205, + "theoretical_loss": 3.423598320473854, + "tokens_seen": 2267348992 + }, + { + "epoch": 0.69, + "learning_rate": 0.0002530310932798395, + "loss": 1.2841, + "theoretical_loss": 3.423590338852619, + "tokens_seen": 2267414528 + }, + { + "epoch": 0.69, + "learning_rate": 0.0002530150451354062, + "loss": 1.2739, + "theoretical_loss": 3.42358235752667, + "tokens_seen": 2267480064 + }, + { + "epoch": 0.69, + "objective/train/docs_used": 825734, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.2516493797302246, + "objective/train/theoretical_loss": 3.4235743764959867, + "objective/train/tokens_used": 2288005600, + "theoretical_loss": 3.4235743764959867, + "tokens_seen": 2267545600 + }, + { + "epoch": 0.69, + "learning_rate": 0.0002529989969909729, + "loss": 1.245, + "theoretical_loss": 3.4235743764959867, + "tokens_seen": 2267545600 + }, + { + "epoch": 0.69, + "learning_rate": 0.0002529829488465396, + "loss": 1.2588, + "theoretical_loss": 3.42356639576055, + "tokens_seen": 2267611136 + }, + { + "epoch": 0.69, + "learning_rate": 0.0002529669007021063, + "loss": 1.4014, + "theoretical_loss": 3.4235584153203407, + "tokens_seen": 2267676672 + }, + { + "epoch": 0.69, + "learning_rate": 0.000252950852557673, + "loss": 1.2382, + "theoretical_loss": 3.423550435175339, + "tokens_seen": 2267742208 + }, + { + "epoch": 0.69, + "learning_rate": 0.0002529348044132397, + "loss": 1.3296, + "theoretical_loss": 3.423542455325525, + "tokens_seen": 2267807744 + }, + { + "epoch": 0.69, + "learning_rate": 0.0002529187562688064, + "loss": 1.3185, + "theoretical_loss": 3.42353447577088, + "tokens_seen": 2267873280 + }, + { + "epoch": 0.69, + "learning_rate": 0.0002529027081243731, + "loss": 1.2666, + "theoretical_loss": 3.4235264965113843, + "tokens_seen": 2267938816 + }, + { + "epoch": 0.69, + "learning_rate": 0.0002528866599799398, + "loss": 1.3907, + "theoretical_loss": 3.4235185175470186, + "tokens_seen": 2268004352 + }, + { + "epoch": 0.69, + "learning_rate": 0.0002528706118355065, + "loss": 1.3081, + "theoretical_loss": 3.423510538877763, + "tokens_seen": 2268069888 + }, + { + "epoch": 0.69, + "learning_rate": 0.0002528545636910732, + "loss": 1.3657, + "theoretical_loss": 3.4235025605035987, + "tokens_seen": 2268135424 + }, + { + "epoch": 0.69, + "learning_rate": 0.00025283851554663996, + "loss": 1.3389, + "theoretical_loss": 3.4234945824245058, + "tokens_seen": 2268200960 + }, + { + "epoch": 0.69, + "learning_rate": 0.00025282246740220666, + "loss": 1.297, + "theoretical_loss": 3.423486604640465, + "tokens_seen": 2268266496 + }, + { + "epoch": 0.69, + "learning_rate": 0.00025280641925777336, + "loss": 1.3367, + "theoretical_loss": 3.4234786271514563, + "tokens_seen": 2268332032 + }, + { + "epoch": 0.69, + "learning_rate": 0.00025279037111334006, + "loss": 1.2278, + "theoretical_loss": 3.423470649957461, + "tokens_seen": 2268397568 + }, + { + "epoch": 0.69, + "learning_rate": 0.00025277432296890676, + "loss": 1.3473, + "theoretical_loss": 3.4234626730584603, + "tokens_seen": 2268463104 + }, + { + "epoch": 0.69, + "learning_rate": 0.00025275827482447346, + "loss": 1.3112, + "theoretical_loss": 3.4234546964544332, + "tokens_seen": 2268528640 + }, + { + "epoch": 0.69, + "learning_rate": 0.00025274222668004016, + "loss": 1.3151, + "theoretical_loss": 3.4234467201453613, + "tokens_seen": 2268594176 + }, + { + "epoch": 0.69, + "learning_rate": 0.00025272617853560686, + "loss": 1.309, + "theoretical_loss": 3.4234387441312246, + "tokens_seen": 2268659712 + }, + { + "epoch": 0.69, + "learning_rate": 0.00025271013039117356, + "loss": 1.3282, + "theoretical_loss": 3.423430768412004, + "tokens_seen": 2268725248 + }, + { + "epoch": 0.69, + "learning_rate": 0.00025269408224674026, + "loss": 1.3326, + "theoretical_loss": 3.4234227929876804, + "tokens_seen": 2268790784 + }, + { + "epoch": 0.69, + "learning_rate": 0.00025267803410230697, + "loss": 1.2963, + "theoretical_loss": 3.4234148178582338, + "tokens_seen": 2268856320 + }, + { + "epoch": 0.69, + "learning_rate": 0.00025266198595787367, + "loss": 1.402, + "theoretical_loss": 3.4234068430236446, + "tokens_seen": 2268921856 + }, + { + "epoch": 0.69, + "learning_rate": 0.00025264593781344037, + "loss": 1.2789, + "theoretical_loss": 3.4233988684838943, + "tokens_seen": 2268987392 + }, + { + "epoch": 0.69, + "learning_rate": 0.00025262988966900707, + "loss": 1.2489, + "theoretical_loss": 3.423390894238963, + "tokens_seen": 2269052928 + }, + { + "epoch": 0.69, + "learning_rate": 0.00025261384152457377, + "loss": 1.325, + "theoretical_loss": 3.423382920288831, + "tokens_seen": 2269118464 + }, + { + "epoch": 0.69, + "learning_rate": 0.00025259779338014047, + "loss": 1.3865, + "theoretical_loss": 3.423374946633479, + "tokens_seen": 2269184000 + }, + { + "epoch": 0.69, + "learning_rate": 0.00025258174523570717, + "loss": 1.2332, + "theoretical_loss": 3.423366973272888, + "tokens_seen": 2269249536 + }, + { + "epoch": 0.69, + "learning_rate": 0.00025256569709127387, + "loss": 1.317, + "theoretical_loss": 3.4233590002070384, + "tokens_seen": 2269315072 + }, + { + "epoch": 0.69, + "learning_rate": 0.00025254964894684057, + "loss": 1.2659, + "theoretical_loss": 3.4233510274359107, + "tokens_seen": 2269380608 + }, + { + "epoch": 0.69, + "learning_rate": 0.00025253360080240727, + "loss": 1.2885, + "theoretical_loss": 3.4233430549594854, + "tokens_seen": 2269446144 + }, + { + "epoch": 0.69, + "learning_rate": 0.0002525175526579739, + "loss": 1.2726, + "theoretical_loss": 3.423335082777743, + "tokens_seen": 2269511680 + }, + { + "epoch": 0.69, + "learning_rate": 0.0002525015045135406, + "loss": 1.2583, + "theoretical_loss": 3.4233271108906647, + "tokens_seen": 2269577216 + }, + { + "epoch": 0.69, + "learning_rate": 0.0002524854563691073, + "loss": 1.2216, + "theoretical_loss": 3.4233191392982305, + "tokens_seen": 2269642752 + }, + { + "epoch": 0.69, + "learning_rate": 0.000252469408224674, + "loss": 1.2731, + "theoretical_loss": 3.423311168000421, + "tokens_seen": 2269708288 + }, + { + "epoch": 0.69, + "learning_rate": 0.0002524533600802407, + "loss": 1.3052, + "theoretical_loss": 3.4233031969972174, + "tokens_seen": 2269773824 + }, + { + "epoch": 0.69, + "learning_rate": 0.0002524373119358074, + "loss": 1.3351, + "theoretical_loss": 3.4232952262886, + "tokens_seen": 2269839360 + }, + { + "epoch": 0.69, + "learning_rate": 0.0002524212637913741, + "loss": 1.3011, + "theoretical_loss": 3.423287255874549, + "tokens_seen": 2269904896 + }, + { + "epoch": 0.69, + "learning_rate": 0.0002524052156469408, + "loss": 1.2697, + "theoretical_loss": 3.423279285755045, + "tokens_seen": 2269970432 + }, + { + "epoch": 0.69, + "learning_rate": 0.0002523891675025075, + "loss": 1.3803, + "theoretical_loss": 3.4232713159300694, + "tokens_seen": 2270035968 + }, + { + "epoch": 0.69, + "learning_rate": 0.0002523731193580742, + "loss": 1.3123, + "theoretical_loss": 3.4232633463996023, + "tokens_seen": 2270101504 + }, + { + "epoch": 0.69, + "learning_rate": 0.0002523570712136409, + "loss": 1.3368, + "theoretical_loss": 3.4232553771636245, + "tokens_seen": 2270167040 + }, + { + "epoch": 0.69, + "learning_rate": 0.0002523410230692076, + "loss": 1.3772, + "theoretical_loss": 3.4232474082221165, + "tokens_seen": 2270232576 + }, + { + "epoch": 0.69, + "learning_rate": 0.0002523249749247743, + "loss": 1.2692, + "theoretical_loss": 3.4232394395750587, + "tokens_seen": 2270298112 + }, + { + "epoch": 0.69, + "learning_rate": 0.000252308926780341, + "loss": 1.4307, + "theoretical_loss": 3.4232314712224317, + "tokens_seen": 2270363648 + }, + { + "epoch": 0.69, + "learning_rate": 0.0002522928786359077, + "loss": 1.2997, + "theoretical_loss": 3.423223503164217, + "tokens_seen": 2270429184 + }, + { + "epoch": 0.69, + "learning_rate": 0.00025227683049147443, + "loss": 1.3506, + "theoretical_loss": 3.423215535400394, + "tokens_seen": 2270494720 + }, + { + "epoch": 0.69, + "learning_rate": 0.00025226078234704113, + "loss": 1.2333, + "theoretical_loss": 3.423207567930944, + "tokens_seen": 2270560256 + }, + { + "epoch": 0.69, + "learning_rate": 0.00025224473420260783, + "loss": 1.2375, + "theoretical_loss": 3.423199600755848, + "tokens_seen": 2270625792 + }, + { + "epoch": 0.69, + "learning_rate": 0.00025222868605817453, + "loss": 1.2755, + "theoretical_loss": 3.423191633875086, + "tokens_seen": 2270691328 + }, + { + "epoch": 0.69, + "learning_rate": 0.00025221263791374123, + "loss": 1.4228, + "theoretical_loss": 3.4231836672886384, + "tokens_seen": 2270756864 + }, + { + "epoch": 0.69, + "objective/train/docs_used": 826907, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.2616167068481445, + "objective/train/theoretical_loss": 3.4231757009964863, + "objective/train/tokens_used": 2291282400, + "theoretical_loss": 3.4231757009964863, + "tokens_seen": 2270822400 + }, + { + "epoch": 0.69, + "learning_rate": 0.00025219658976930793, + "loss": 1.2981, + "theoretical_loss": 3.4231757009964863, + "tokens_seen": 2270822400 + }, + { + "epoch": 0.69, + "learning_rate": 0.00025218054162487463, + "loss": 1.2277, + "theoretical_loss": 3.4231677349986107, + "tokens_seen": 2270887936 + }, + { + "epoch": 0.69, + "learning_rate": 0.00025216449348044133, + "loss": 1.1543, + "theoretical_loss": 3.4231597692949913, + "tokens_seen": 2270953472 + }, + { + "epoch": 0.69, + "learning_rate": 0.00025214844533600803, + "loss": 1.2838, + "theoretical_loss": 3.4231518038856095, + "tokens_seen": 2271019008 + }, + { + "epoch": 0.69, + "learning_rate": 0.00025213239719157473, + "loss": 1.321, + "theoretical_loss": 3.4231438387704456, + "tokens_seen": 2271084544 + }, + { + "epoch": 0.69, + "learning_rate": 0.00025211634904714143, + "loss": 1.3198, + "theoretical_loss": 3.42313587394948, + "tokens_seen": 2271150080 + }, + { + "epoch": 0.69, + "learning_rate": 0.00025210030090270813, + "loss": 1.3673, + "theoretical_loss": 3.4231279094226945, + "tokens_seen": 2271215616 + }, + { + "epoch": 0.69, + "learning_rate": 0.00025208425275827483, + "loss": 1.2865, + "theoretical_loss": 3.4231199451900682, + "tokens_seen": 2271281152 + }, + { + "epoch": 0.69, + "learning_rate": 0.00025206820461384154, + "loss": 1.2723, + "theoretical_loss": 3.4231119812515827, + "tokens_seen": 2271346688 + }, + { + "epoch": 0.69, + "learning_rate": 0.00025205215646940824, + "loss": 1.3275, + "theoretical_loss": 3.4231040176072187, + "tokens_seen": 2271412224 + }, + { + "epoch": 0.69, + "learning_rate": 0.00025203610832497494, + "loss": 1.2169, + "theoretical_loss": 3.4230960542569564, + "tokens_seen": 2271477760 + }, + { + "epoch": 0.69, + "learning_rate": 0.00025202006018054164, + "loss": 1.3691, + "theoretical_loss": 3.4230880912007766, + "tokens_seen": 2271543296 + }, + { + "epoch": 0.69, + "learning_rate": 0.00025200401203610834, + "loss": 1.3628, + "theoretical_loss": 3.4230801284386594, + "tokens_seen": 2271608832 + }, + { + "epoch": 0.69, + "learning_rate": 0.00025198796389167504, + "loss": 1.4435, + "theoretical_loss": 3.423072165970587, + "tokens_seen": 2271674368 + }, + { + "epoch": 0.69, + "learning_rate": 0.00025197191574724174, + "loss": 1.3819, + "theoretical_loss": 3.4230642037965384, + "tokens_seen": 2271739904 + }, + { + "epoch": 0.69, + "learning_rate": 0.00025195586760280844, + "loss": 1.2443, + "theoretical_loss": 3.423056241916495, + "tokens_seen": 2271805440 + }, + { + "epoch": 0.69, + "learning_rate": 0.00025193981945837514, + "loss": 1.3218, + "theoretical_loss": 3.4230482803304376, + "tokens_seen": 2271870976 + }, + { + "epoch": 0.69, + "learning_rate": 0.00025192377131394184, + "loss": 1.2983, + "theoretical_loss": 3.423040319038347, + "tokens_seen": 2271936512 + }, + { + "epoch": 0.69, + "learning_rate": 0.00025190772316950854, + "loss": 1.5136, + "theoretical_loss": 3.423032358040203, + "tokens_seen": 2272002048 + }, + { + "epoch": 0.69, + "learning_rate": 0.00025189167502507524, + "loss": 1.3238, + "theoretical_loss": 3.423024397335987, + "tokens_seen": 2272067584 + }, + { + "epoch": 0.69, + "learning_rate": 0.00025187562688064194, + "loss": 1.2419, + "theoretical_loss": 3.4230164369256797, + "tokens_seen": 2272133120 + }, + { + "epoch": 0.69, + "learning_rate": 0.00025185957873620864, + "loss": 1.2862, + "theoretical_loss": 3.4230084768092612, + "tokens_seen": 2272198656 + }, + { + "epoch": 0.69, + "learning_rate": 0.00025184353059177534, + "loss": 1.2425, + "theoretical_loss": 3.423000516986713, + "tokens_seen": 2272264192 + }, + { + "epoch": 0.69, + "learning_rate": 0.00025182748244734204, + "loss": 1.2886, + "theoretical_loss": 3.4229925574580147, + "tokens_seen": 2272329728 + }, + { + "epoch": 0.69, + "learning_rate": 0.00025181143430290875, + "loss": 1.2216, + "theoretical_loss": 3.422984598223148, + "tokens_seen": 2272395264 + }, + { + "epoch": 0.69, + "learning_rate": 0.00025179538615847545, + "loss": 1.2886, + "theoretical_loss": 3.4229766392820933, + "tokens_seen": 2272460800 + }, + { + "epoch": 0.69, + "learning_rate": 0.00025177933801404215, + "loss": 1.1612, + "theoretical_loss": 3.4229686806348307, + "tokens_seen": 2272526336 + }, + { + "epoch": 0.69, + "learning_rate": 0.00025176328986960885, + "loss": 1.2273, + "theoretical_loss": 3.4229607222813416, + "tokens_seen": 2272591872 + }, + { + "epoch": 0.69, + "learning_rate": 0.0002517472417251755, + "loss": 1.3586, + "theoretical_loss": 3.4229527642216064, + "tokens_seen": 2272657408 + }, + { + "epoch": 0.69, + "learning_rate": 0.0002517311935807422, + "loss": 1.2493, + "theoretical_loss": 3.4229448064556056, + "tokens_seen": 2272722944 + }, + { + "epoch": 0.69, + "learning_rate": 0.0002517151454363089, + "loss": 1.4304, + "theoretical_loss": 3.42293684898332, + "tokens_seen": 2272788480 + }, + { + "epoch": 0.69, + "learning_rate": 0.0002516990972918756, + "loss": 1.3083, + "theoretical_loss": 3.422928891804731, + "tokens_seen": 2272854016 + }, + { + "epoch": 0.69, + "learning_rate": 0.0002516830491474423, + "loss": 1.3947, + "theoretical_loss": 3.4229209349198184, + "tokens_seen": 2272919552 + }, + { + "epoch": 0.69, + "learning_rate": 0.000251667001003009, + "loss": 1.279, + "theoretical_loss": 3.422912978328563, + "tokens_seen": 2272985088 + }, + { + "epoch": 0.69, + "learning_rate": 0.00025165095285857575, + "loss": 1.3244, + "theoretical_loss": 3.4229050220309456, + "tokens_seen": 2273050624 + }, + { + "epoch": 0.69, + "learning_rate": 0.00025163490471414245, + "loss": 1.2412, + "theoretical_loss": 3.4228970660269473, + "tokens_seen": 2273116160 + }, + { + "epoch": 0.69, + "learning_rate": 0.00025161885656970915, + "loss": 1.3223, + "theoretical_loss": 3.422889110316548, + "tokens_seen": 2273181696 + }, + { + "epoch": 0.69, + "learning_rate": 0.00025160280842527585, + "loss": 1.2684, + "theoretical_loss": 3.422881154899729, + "tokens_seen": 2273247232 + }, + { + "epoch": 0.69, + "learning_rate": 0.00025158676028084255, + "loss": 1.334, + "theoretical_loss": 3.422873199776471, + "tokens_seen": 2273312768 + }, + { + "epoch": 0.69, + "learning_rate": 0.00025157071213640926, + "loss": 1.3233, + "theoretical_loss": 3.4228652449467547, + "tokens_seen": 2273378304 + }, + { + "epoch": 0.69, + "learning_rate": 0.00025155466399197596, + "loss": 1.5333, + "theoretical_loss": 3.4228572904105605, + "tokens_seen": 2273443840 + }, + { + "epoch": 0.69, + "learning_rate": 0.00025153861584754266, + "loss": 1.3484, + "theoretical_loss": 3.4228493361678693, + "tokens_seen": 2273509376 + }, + { + "epoch": 0.69, + "learning_rate": 0.00025152256770310936, + "loss": 1.4386, + "theoretical_loss": 3.4228413822186616, + "tokens_seen": 2273574912 + }, + { + "epoch": 0.69, + "learning_rate": 0.00025150651955867606, + "loss": 1.2062, + "theoretical_loss": 3.4228334285629187, + "tokens_seen": 2273640448 + }, + { + "epoch": 0.69, + "learning_rate": 0.00025149047141424276, + "loss": 1.2534, + "theoretical_loss": 3.4228254752006206, + "tokens_seen": 2273705984 + }, + { + "epoch": 0.69, + "learning_rate": 0.00025147442326980946, + "loss": 1.2745, + "theoretical_loss": 3.4228175221317483, + "tokens_seen": 2273771520 + }, + { + "epoch": 0.69, + "learning_rate": 0.00025145837512537616, + "loss": 1.3425, + "theoretical_loss": 3.422809569356283, + "tokens_seen": 2273837056 + }, + { + "epoch": 0.69, + "learning_rate": 0.00025144232698094286, + "loss": 1.2663, + "theoretical_loss": 3.4228016168742044, + "tokens_seen": 2273902592 + }, + { + "epoch": 0.69, + "learning_rate": 0.00025142627883650956, + "loss": 1.2646, + "theoretical_loss": 3.422793664685494, + "tokens_seen": 2273968128 + }, + { + "epoch": 0.69, + "learning_rate": 0.00025141023069207626, + "loss": 1.3521, + "theoretical_loss": 3.4227857127901324, + "tokens_seen": 2274033664 + }, + { + "epoch": 0.69, + "objective/train/docs_used": 828122, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.2746100425720215, + "objective/train/theoretical_loss": 3.4227777611881, + "objective/train/tokens_used": 2294559200, + "theoretical_loss": 3.4227777611881, + "tokens_seen": 2274099200 + }, + { + "epoch": 0.69, + "learning_rate": 0.00025139418254764296, + "loss": 1.3117, + "theoretical_loss": 3.4227777611881, + "tokens_seen": 2274099200 + }, + { + "epoch": 0.69, + "learning_rate": 0.00025137813440320966, + "loss": 1.2747, + "theoretical_loss": 3.4227698098793784, + "tokens_seen": 2274164736 + }, + { + "epoch": 0.69, + "learning_rate": 0.00025136208625877636, + "loss": 1.3793, + "theoretical_loss": 3.422761858863947, + "tokens_seen": 2274230272 + }, + { + "epoch": 0.69, + "learning_rate": 0.00025134603811434306, + "loss": 1.4963, + "theoretical_loss": 3.4227539081417877, + "tokens_seen": 2274295808 + }, + { + "epoch": 0.69, + "learning_rate": 0.00025132998996990976, + "loss": 1.4352, + "theoretical_loss": 3.4227459577128805, + "tokens_seen": 2274361344 + }, + { + "epoch": 0.69, + "learning_rate": 0.00025131394182547647, + "loss": 1.4006, + "theoretical_loss": 3.4227380075772067, + "tokens_seen": 2274426880 + }, + { + "epoch": 0.69, + "learning_rate": 0.00025129789368104317, + "loss": 1.3524, + "theoretical_loss": 3.4227300577347464, + "tokens_seen": 2274492416 + }, + { + "epoch": 0.69, + "learning_rate": 0.00025128184553660987, + "loss": 1.3485, + "theoretical_loss": 3.422722108185481, + "tokens_seen": 2274557952 + }, + { + "epoch": 0.69, + "learning_rate": 0.00025126579739217657, + "loss": 1.3713, + "theoretical_loss": 3.4227141589293906, + "tokens_seen": 2274623488 + }, + { + "epoch": 0.69, + "learning_rate": 0.00025124974924774327, + "loss": 1.31, + "theoretical_loss": 3.4227062099664565, + "tokens_seen": 2274689024 + }, + { + "epoch": 0.69, + "learning_rate": 0.00025123370110330997, + "loss": 1.3204, + "theoretical_loss": 3.4226982612966586, + "tokens_seen": 2274754560 + }, + { + "epoch": 0.69, + "learning_rate": 0.00025121765295887667, + "loss": 1.2149, + "theoretical_loss": 3.4226903129199786, + "tokens_seen": 2274820096 + }, + { + "epoch": 0.69, + "learning_rate": 0.00025120160481444337, + "loss": 1.3205, + "theoretical_loss": 3.422682364836397, + "tokens_seen": 2274885632 + }, + { + "epoch": 0.69, + "learning_rate": 0.00025118555667001007, + "loss": 1.1799, + "theoretical_loss": 3.4226744170458945, + "tokens_seen": 2274951168 + }, + { + "epoch": 0.69, + "learning_rate": 0.00025116950852557677, + "loss": 1.3985, + "theoretical_loss": 3.4226664695484517, + "tokens_seen": 2275016704 + }, + { + "epoch": 0.69, + "learning_rate": 0.00025115346038114347, + "loss": 1.339, + "theoretical_loss": 3.4226585223440495, + "tokens_seen": 2275082240 + }, + { + "epoch": 0.69, + "learning_rate": 0.00025113741223671017, + "loss": 1.2626, + "theoretical_loss": 3.4226505754326686, + "tokens_seen": 2275147776 + }, + { + "epoch": 0.69, + "learning_rate": 0.0002511213640922768, + "loss": 1.4504, + "theoretical_loss": 3.4226426288142893, + "tokens_seen": 2275213312 + }, + { + "epoch": 0.69, + "learning_rate": 0.0002511053159478435, + "loss": 1.29, + "theoretical_loss": 3.4226346824888934, + "tokens_seen": 2275278848 + }, + { + "epoch": 0.69, + "learning_rate": 0.0002510892678034102, + "loss": 1.2596, + "theoretical_loss": 3.422626736456461, + "tokens_seen": 2275344384 + }, + { + "epoch": 0.69, + "learning_rate": 0.0002510732196589769, + "loss": 1.3514, + "theoretical_loss": 3.4226187907169723, + "tokens_seen": 2275409920 + }, + { + "epoch": 0.69, + "learning_rate": 0.0002510571715145436, + "loss": 1.2759, + "theoretical_loss": 3.4226108452704094, + "tokens_seen": 2275475456 + }, + { + "epoch": 0.69, + "learning_rate": 0.0002510411233701103, + "loss": 1.4978, + "theoretical_loss": 3.422602900116752, + "tokens_seen": 2275540992 + }, + { + "epoch": 0.69, + "learning_rate": 0.000251025075225677, + "loss": 1.3787, + "theoretical_loss": 3.422594955255981, + "tokens_seen": 2275606528 + }, + { + "epoch": 0.69, + "learning_rate": 0.0002510090270812437, + "loss": 1.3491, + "theoretical_loss": 3.422587010688078, + "tokens_seen": 2275672064 + }, + { + "epoch": 0.69, + "learning_rate": 0.0002509929789368104, + "loss": 1.2931, + "theoretical_loss": 3.422579066413023, + "tokens_seen": 2275737600 + }, + { + "epoch": 0.69, + "learning_rate": 0.0002509769307923771, + "loss": 1.3846, + "theoretical_loss": 3.4225711224307966, + "tokens_seen": 2275803136 + }, + { + "epoch": 0.69, + "learning_rate": 0.0002509608826479438, + "loss": 1.3379, + "theoretical_loss": 3.42256317874138, + "tokens_seen": 2275868672 + }, + { + "epoch": 0.69, + "learning_rate": 0.0002509448345035105, + "loss": 1.3083, + "theoretical_loss": 3.422555235344754, + "tokens_seen": 2275934208 + }, + { + "epoch": 0.69, + "learning_rate": 0.0002509287863590772, + "loss": 1.3937, + "theoretical_loss": 3.4225472922408993, + "tokens_seen": 2275999744 + }, + { + "epoch": 0.69, + "learning_rate": 0.0002509127382146439, + "loss": 1.2697, + "theoretical_loss": 3.4225393494297967, + "tokens_seen": 2276065280 + }, + { + "epoch": 0.69, + "learning_rate": 0.00025089669007021063, + "loss": 1.2518, + "theoretical_loss": 3.4225314069114265, + "tokens_seen": 2276130816 + }, + { + "epoch": 0.69, + "learning_rate": 0.00025088064192577733, + "loss": 1.3542, + "theoretical_loss": 3.4225234646857703, + "tokens_seen": 2276196352 + }, + { + "epoch": 0.69, + "learning_rate": 0.00025086459378134403, + "loss": 1.3201, + "theoretical_loss": 3.4225155227528083, + "tokens_seen": 2276261888 + }, + { + "epoch": 0.69, + "learning_rate": 0.00025084854563691073, + "loss": 1.1743, + "theoretical_loss": 3.4225075811125216, + "tokens_seen": 2276327424 + }, + { + "epoch": 0.69, + "learning_rate": 0.00025083249749247743, + "loss": 1.3261, + "theoretical_loss": 3.422499639764891, + "tokens_seen": 2276392960 + }, + { + "epoch": 0.69, + "learning_rate": 0.00025081644934804413, + "loss": 1.2629, + "theoretical_loss": 3.422491698709897, + "tokens_seen": 2276458496 + }, + { + "epoch": 0.69, + "learning_rate": 0.00025080040120361083, + "loss": 1.1965, + "theoretical_loss": 3.4224837579475205, + "tokens_seen": 2276524032 + }, + { + "epoch": 0.69, + "learning_rate": 0.00025078435305917753, + "loss": 1.3127, + "theoretical_loss": 3.422475817477742, + "tokens_seen": 2276589568 + }, + { + "epoch": 0.69, + "learning_rate": 0.00025076830491474423, + "loss": 1.4067, + "theoretical_loss": 3.422467877300543, + "tokens_seen": 2276655104 + }, + { + "epoch": 0.69, + "learning_rate": 0.00025075225677031093, + "loss": 1.3088, + "theoretical_loss": 3.422459937415904, + "tokens_seen": 2276720640 + }, + { + "epoch": 0.69, + "learning_rate": 0.00025073620862587763, + "loss": 1.3262, + "theoretical_loss": 3.4224519978238055, + "tokens_seen": 2276786176 + }, + { + "epoch": 0.69, + "learning_rate": 0.00025072016048144433, + "loss": 1.2414, + "theoretical_loss": 3.422444058524229, + "tokens_seen": 2276851712 + }, + { + "epoch": 0.69, + "learning_rate": 0.00025070411233701104, + "loss": 1.3159, + "theoretical_loss": 3.4224361195171547, + "tokens_seen": 2276917248 + }, + { + "epoch": 0.69, + "learning_rate": 0.00025068806419257774, + "loss": 1.2437, + "theoretical_loss": 3.422428180802563, + "tokens_seen": 2276982784 + }, + { + "epoch": 0.69, + "learning_rate": 0.00025067201604814444, + "loss": 1.2485, + "theoretical_loss": 3.422420242380436, + "tokens_seen": 2277048320 + }, + { + "epoch": 0.69, + "learning_rate": 0.00025065596790371114, + "loss": 1.3778, + "theoretical_loss": 3.422412304250753, + "tokens_seen": 2277113856 + }, + { + "epoch": 0.69, + "learning_rate": 0.00025063991975927784, + "loss": 1.341, + "theoretical_loss": 3.422404366413496, + "tokens_seen": 2277179392 + }, + { + "epoch": 0.69, + "learning_rate": 0.00025062387161484454, + "loss": 1.3093, + "theoretical_loss": 3.4223964288686455, + "tokens_seen": 2277244928 + }, + { + "epoch": 0.69, + "learning_rate": 0.00025060782347041124, + "loss": 1.3302, + "theoretical_loss": 3.422388491616182, + "tokens_seen": 2277310464 + }, + { + "epoch": 0.69, + "objective/train/docs_used": 829293, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.4799326658248901, + "objective/train/theoretical_loss": 3.422380554656087, + "objective/train/tokens_used": 2297836000, + "theoretical_loss": 3.422380554656087, + "tokens_seen": 2277376000 + }, + { + "epoch": 0.69, + "learning_rate": 0.00025059177532597794, + "loss": 1.4008, + "theoretical_loss": 3.422380554656087, + "tokens_seen": 2277376000 + }, + { + "epoch": 0.69, + "learning_rate": 0.00025057572718154464, + "loss": 1.2955, + "theoretical_loss": 3.42237261798834, + "tokens_seen": 2277441536 + }, + { + "epoch": 0.69, + "learning_rate": 0.00025055967903711134, + "loss": 1.3363, + "theoretical_loss": 3.422364681612923, + "tokens_seen": 2277507072 + }, + { + "epoch": 0.69, + "learning_rate": 0.00025054363089267804, + "loss": 1.3633, + "theoretical_loss": 3.4223567455298167, + "tokens_seen": 2277572608 + }, + { + "epoch": 0.69, + "learning_rate": 0.00025052758274824474, + "loss": 1.2674, + "theoretical_loss": 3.4223488097390016, + "tokens_seen": 2277638144 + }, + { + "epoch": 0.69, + "learning_rate": 0.00025051153460381144, + "loss": 1.2958, + "theoretical_loss": 3.422340874240458, + "tokens_seen": 2277703680 + }, + { + "epoch": 0.69, + "learning_rate": 0.00025049548645937814, + "loss": 1.3568, + "theoretical_loss": 3.422332939034168, + "tokens_seen": 2277769216 + }, + { + "epoch": 0.69, + "learning_rate": 0.00025047943831494484, + "loss": 1.2574, + "theoretical_loss": 3.4223250041201116, + "tokens_seen": 2277834752 + }, + { + "epoch": 0.69, + "learning_rate": 0.00025046339017051154, + "loss": 1.2663, + "theoretical_loss": 3.42231706949827, + "tokens_seen": 2277900288 + }, + { + "epoch": 0.69, + "learning_rate": 0.00025044734202607825, + "loss": 1.316, + "theoretical_loss": 3.4223091351686237, + "tokens_seen": 2277965824 + }, + { + "epoch": 0.69, + "learning_rate": 0.00025043129388164495, + "loss": 1.4141, + "theoretical_loss": 3.4223012011311535, + "tokens_seen": 2278031360 + }, + { + "epoch": 0.69, + "learning_rate": 0.00025041524573721165, + "loss": 1.3207, + "theoretical_loss": 3.4222932673858404, + "tokens_seen": 2278096896 + }, + { + "epoch": 0.69, + "learning_rate": 0.00025039919759277835, + "loss": 1.3003, + "theoretical_loss": 3.4222853339326655, + "tokens_seen": 2278162432 + }, + { + "epoch": 0.69, + "learning_rate": 0.00025038314944834505, + "loss": 1.2817, + "theoretical_loss": 3.4222774007716095, + "tokens_seen": 2278227968 + }, + { + "epoch": 0.69, + "learning_rate": 0.00025036710130391175, + "loss": 1.2795, + "theoretical_loss": 3.4222694679026526, + "tokens_seen": 2278293504 + }, + { + "epoch": 0.69, + "learning_rate": 0.00025035105315947845, + "loss": 1.2894, + "theoretical_loss": 3.4222615353257764, + "tokens_seen": 2278359040 + }, + { + "epoch": 0.69, + "learning_rate": 0.00025033500501504515, + "loss": 1.3085, + "theoretical_loss": 3.422253603040962, + "tokens_seen": 2278424576 + }, + { + "epoch": 0.69, + "learning_rate": 0.00025031895687061185, + "loss": 1.2696, + "theoretical_loss": 3.422245671048189, + "tokens_seen": 2278490112 + }, + { + "epoch": 0.69, + "learning_rate": 0.00025030290872617855, + "loss": 1.3817, + "theoretical_loss": 3.4222377393474392, + "tokens_seen": 2278555648 + }, + { + "epoch": 0.69, + "learning_rate": 0.00025028686058174525, + "loss": 1.3473, + "theoretical_loss": 3.422229807938694, + "tokens_seen": 2278621184 + }, + { + "epoch": 0.69, + "learning_rate": 0.00025027081243731195, + "loss": 1.2645, + "theoretical_loss": 3.422221876821933, + "tokens_seen": 2278686720 + }, + { + "epoch": 0.69, + "learning_rate": 0.00025025476429287865, + "loss": 1.2583, + "theoretical_loss": 3.422213945997137, + "tokens_seen": 2278752256 + }, + { + "epoch": 0.69, + "learning_rate": 0.00025023871614844535, + "loss": 1.2177, + "theoretical_loss": 3.422206015464288, + "tokens_seen": 2278817792 + }, + { + "epoch": 0.69, + "learning_rate": 0.00025022266800401205, + "loss": 1.2238, + "theoretical_loss": 3.422198085223366, + "tokens_seen": 2278883328 + }, + { + "epoch": 0.69, + "learning_rate": 0.00025020661985957875, + "loss": 1.3163, + "theoretical_loss": 3.422190155274352, + "tokens_seen": 2278948864 + }, + { + "epoch": 0.69, + "learning_rate": 0.00025019057171514546, + "loss": 1.2629, + "theoretical_loss": 3.422182225617228, + "tokens_seen": 2279014400 + }, + { + "epoch": 0.69, + "learning_rate": 0.00025017452357071216, + "loss": 1.2753, + "theoretical_loss": 3.422174296251973, + "tokens_seen": 2279079936 + }, + { + "epoch": 0.69, + "learning_rate": 0.00025015847542627886, + "loss": 1.26, + "theoretical_loss": 3.4221663671785687, + "tokens_seen": 2279145472 + }, + { + "epoch": 0.69, + "learning_rate": 0.00025014242728184556, + "loss": 1.318, + "theoretical_loss": 3.4221584383969965, + "tokens_seen": 2279211008 + }, + { + "epoch": 0.69, + "learning_rate": 0.00025012637913741226, + "loss": 1.3714, + "theoretical_loss": 3.422150509907236, + "tokens_seen": 2279276544 + }, + { + "epoch": 0.69, + "learning_rate": 0.00025011033099297896, + "loss": 1.2734, + "theoretical_loss": 3.422142581709269, + "tokens_seen": 2279342080 + }, + { + "epoch": 0.69, + "learning_rate": 0.00025009428284854566, + "loss": 1.3022, + "theoretical_loss": 3.422134653803077, + "tokens_seen": 2279407616 + }, + { + "epoch": 0.69, + "learning_rate": 0.00025007823470411236, + "loss": 1.3745, + "theoretical_loss": 3.4221267261886394, + "tokens_seen": 2279473152 + }, + { + "epoch": 0.69, + "learning_rate": 0.00025006218655967906, + "loss": 1.3466, + "theoretical_loss": 3.422118798865938, + "tokens_seen": 2279538688 + }, + { + "epoch": 0.69, + "learning_rate": 0.00025004613841524576, + "loss": 1.3491, + "theoretical_loss": 3.422110871834953, + "tokens_seen": 2279604224 + }, + { + "epoch": 0.69, + "learning_rate": 0.00025003009027081246, + "loss": 1.2657, + "theoretical_loss": 3.422102945095666, + "tokens_seen": 2279669760 + }, + { + "epoch": 0.69, + "learning_rate": 0.00025001404212637916, + "loss": 1.1792, + "theoretical_loss": 3.422095018648058, + "tokens_seen": 2279735296 + }, + { + "epoch": 0.69, + "learning_rate": 0.00024999799398194586, + "loss": 1.255, + "theoretical_loss": 3.4220870924921085, + "tokens_seen": 2279800832 + }, + { + "epoch": 0.69, + "learning_rate": 0.00024998194583751256, + "loss": 1.3807, + "theoretical_loss": 3.4220791666278, + "tokens_seen": 2279866368 + }, + { + "epoch": 0.69, + "learning_rate": 0.00024996589769307926, + "loss": 1.3524, + "theoretical_loss": 3.4220712410551126, + "tokens_seen": 2279931904 + }, + { + "epoch": 0.69, + "learning_rate": 0.00024994984954864597, + "loss": 1.3436, + "theoretical_loss": 3.4220633157740274, + "tokens_seen": 2279997440 + }, + { + "epoch": 0.69, + "learning_rate": 0.00024993380140421267, + "loss": 1.3747, + "theoretical_loss": 3.422055390784525, + "tokens_seen": 2280062976 + }, + { + "epoch": 0.69, + "learning_rate": 0.00024991775325977937, + "loss": 1.212, + "theoretical_loss": 3.422047466086587, + "tokens_seen": 2280128512 + }, + { + "epoch": 0.69, + "learning_rate": 0.00024990170511534607, + "loss": 1.4399, + "theoretical_loss": 3.4220395416801934, + "tokens_seen": 2280194048 + }, + { + "epoch": 0.69, + "learning_rate": 0.00024988565697091277, + "loss": 1.3241, + "theoretical_loss": 3.4220316175653256, + "tokens_seen": 2280259584 + }, + { + "epoch": 0.69, + "learning_rate": 0.00024986960882647947, + "loss": 1.334, + "theoretical_loss": 3.4220236937419646, + "tokens_seen": 2280325120 + }, + { + "epoch": 0.69, + "learning_rate": 0.00024985356068204617, + "loss": 1.3094, + "theoretical_loss": 3.422015770210091, + "tokens_seen": 2280390656 + }, + { + "epoch": 0.69, + "learning_rate": 0.00024983751253761287, + "loss": 1.2855, + "theoretical_loss": 3.4220078469696853, + "tokens_seen": 2280456192 + }, + { + "epoch": 0.69, + "learning_rate": 0.00024982146439317957, + "loss": 1.3867, + "theoretical_loss": 3.4219999240207297, + "tokens_seen": 2280521728 + }, + { + "epoch": 0.69, + "learning_rate": 0.00024980541624874627, + "loss": 1.3259, + "theoretical_loss": 3.4219920013632037, + "tokens_seen": 2280587264 + }, + { + "epoch": 0.69, + "objective/train/docs_used": 830439, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.2348090410232544, + "objective/train/theoretical_loss": 3.421984078997089, + "objective/train/tokens_used": 2301112800, + "theoretical_loss": 3.421984078997089, + "tokens_seen": 2280652800 + }, + { + "epoch": 0.69, + "learning_rate": 0.00024978936810431297, + "loss": 1.2336, + "theoretical_loss": 3.421984078997089, + "tokens_seen": 2280652800 + }, + { + "epoch": 0.69, + "learning_rate": 0.00024977331995987967, + "loss": 1.2477, + "theoretical_loss": 3.421976156922366, + "tokens_seen": 2280718336 + }, + { + "epoch": 0.69, + "learning_rate": 0.00024975727181544637, + "loss": 1.3143, + "theoretical_loss": 3.4219682351390164, + "tokens_seen": 2280783872 + }, + { + "epoch": 0.69, + "learning_rate": 0.0002497412236710131, + "loss": 1.4447, + "theoretical_loss": 3.4219603136470207, + "tokens_seen": 2280849408 + }, + { + "epoch": 0.69, + "learning_rate": 0.0002497251755265797, + "loss": 1.3195, + "theoretical_loss": 3.4219523924463595, + "tokens_seen": 2280914944 + }, + { + "epoch": 0.69, + "learning_rate": 0.0002497091273821464, + "loss": 1.2617, + "theoretical_loss": 3.4219444715370138, + "tokens_seen": 2280980480 + }, + { + "epoch": 0.69, + "learning_rate": 0.0002496930792377131, + "loss": 1.3254, + "theoretical_loss": 3.4219365509189648, + "tokens_seen": 2281046016 + }, + { + "epoch": 0.69, + "learning_rate": 0.0002496770310932798, + "loss": 1.3086, + "theoretical_loss": 3.4219286305921934, + "tokens_seen": 2281111552 + }, + { + "epoch": 0.69, + "learning_rate": 0.0002496609829488465, + "loss": 1.3004, + "theoretical_loss": 3.4219207105566802, + "tokens_seen": 2281177088 + }, + { + "epoch": 0.69, + "learning_rate": 0.0002496449348044132, + "loss": 1.1092, + "theoretical_loss": 3.421912790812407, + "tokens_seen": 2281242624 + }, + { + "epoch": 0.69, + "learning_rate": 0.0002496288866599799, + "loss": 1.4528, + "theoretical_loss": 3.421904871359353, + "tokens_seen": 2281308160 + }, + { + "epoch": 0.69, + "learning_rate": 0.0002496128385155466, + "loss": 1.2738, + "theoretical_loss": 3.421896952197501, + "tokens_seen": 2281373696 + }, + { + "epoch": 0.69, + "learning_rate": 0.0002495967903711133, + "loss": 1.2265, + "theoretical_loss": 3.4218890333268313, + "tokens_seen": 2281439232 + }, + { + "epoch": 0.69, + "learning_rate": 0.00024958074222668, + "loss": 1.343, + "theoretical_loss": 3.421881114747324, + "tokens_seen": 2281504768 + }, + { + "epoch": 0.69, + "learning_rate": 0.0002495646940822467, + "loss": 1.2465, + "theoretical_loss": 3.421873196458961, + "tokens_seen": 2281570304 + }, + { + "epoch": 0.69, + "learning_rate": 0.0002495486459378134, + "loss": 1.3759, + "theoretical_loss": 3.421865278461723, + "tokens_seen": 2281635840 + }, + { + "epoch": 0.69, + "learning_rate": 0.00024953259779338013, + "loss": 1.2689, + "theoretical_loss": 3.4218573607555904, + "tokens_seen": 2281701376 + }, + { + "epoch": 0.69, + "learning_rate": 0.00024951654964894683, + "loss": 1.3267, + "theoretical_loss": 3.421849443340545, + "tokens_seen": 2281766912 + }, + { + "epoch": 0.69, + "learning_rate": 0.00024950050150451353, + "loss": 1.2657, + "theoretical_loss": 3.4218415262165673, + "tokens_seen": 2281832448 + }, + { + "epoch": 0.69, + "learning_rate": 0.00024948445336008023, + "loss": 1.325, + "theoretical_loss": 3.421833609383638, + "tokens_seen": 2281897984 + }, + { + "epoch": 0.69, + "learning_rate": 0.00024946840521564693, + "loss": 1.1681, + "theoretical_loss": 3.4218256928417388, + "tokens_seen": 2281963520 + }, + { + "epoch": 0.69, + "learning_rate": 0.00024945235707121363, + "loss": 1.4412, + "theoretical_loss": 3.4218177765908497, + "tokens_seen": 2282029056 + }, + { + "epoch": 0.69, + "learning_rate": 0.00024943630892678033, + "loss": 1.2775, + "theoretical_loss": 3.421809860630952, + "tokens_seen": 2282094592 + }, + { + "epoch": 0.69, + "learning_rate": 0.00024942026078234703, + "loss": 1.358, + "theoretical_loss": 3.421801944962027, + "tokens_seen": 2282160128 + }, + { + "epoch": 0.69, + "learning_rate": 0.00024940421263791373, + "loss": 1.2644, + "theoretical_loss": 3.4217940295840554, + "tokens_seen": 2282225664 + }, + { + "epoch": 0.69, + "learning_rate": 0.00024938816449348043, + "loss": 1.329, + "theoretical_loss": 3.421786114497018, + "tokens_seen": 2282291200 + }, + { + "epoch": 0.69, + "learning_rate": 0.00024937211634904713, + "loss": 1.306, + "theoretical_loss": 3.4217781997008965, + "tokens_seen": 2282356736 + }, + { + "epoch": 0.69, + "learning_rate": 0.00024935606820461383, + "loss": 1.2568, + "theoretical_loss": 3.4217702851956706, + "tokens_seen": 2282422272 + }, + { + "epoch": 0.69, + "learning_rate": 0.00024934002006018053, + "loss": 1.3224, + "theoretical_loss": 3.421762370981322, + "tokens_seen": 2282487808 + }, + { + "epoch": 0.69, + "learning_rate": 0.00024932397191574724, + "loss": 1.3792, + "theoretical_loss": 3.4217544570578315, + "tokens_seen": 2282553344 + }, + { + "epoch": 0.69, + "learning_rate": 0.00024930792377131394, + "loss": 1.3865, + "theoretical_loss": 3.4217465434251806, + "tokens_seen": 2282618880 + }, + { + "epoch": 0.69, + "learning_rate": 0.00024929187562688064, + "loss": 1.1906, + "theoretical_loss": 3.4217386300833494, + "tokens_seen": 2282684416 + }, + { + "epoch": 0.69, + "learning_rate": 0.00024927582748244734, + "loss": 1.2775, + "theoretical_loss": 3.421730717032319, + "tokens_seen": 2282749952 + }, + { + "epoch": 0.69, + "learning_rate": 0.00024925977933801404, + "loss": 1.4173, + "theoretical_loss": 3.421722804272071, + "tokens_seen": 2282815488 + }, + { + "epoch": 0.69, + "learning_rate": 0.0002492437311935808, + "loss": 1.2247, + "theoretical_loss": 3.421714891802586, + "tokens_seen": 2282881024 + }, + { + "epoch": 0.69, + "learning_rate": 0.0002492276830491475, + "loss": 1.3099, + "theoretical_loss": 3.421706979623845, + "tokens_seen": 2282946560 + }, + { + "epoch": 0.69, + "learning_rate": 0.0002492116349047142, + "loss": 1.2436, + "theoretical_loss": 3.4216990677358288, + "tokens_seen": 2283012096 + }, + { + "epoch": 0.69, + "learning_rate": 0.0002491955867602809, + "loss": 1.3013, + "theoretical_loss": 3.421691156138518, + "tokens_seen": 2283077632 + }, + { + "epoch": 0.69, + "learning_rate": 0.0002491795386158476, + "loss": 1.2643, + "theoretical_loss": 3.421683244831895, + "tokens_seen": 2283143168 + }, + { + "epoch": 0.69, + "learning_rate": 0.0002491634904714143, + "loss": 1.2891, + "theoretical_loss": 3.4216753338159394, + "tokens_seen": 2283208704 + }, + { + "epoch": 0.69, + "learning_rate": 0.000249147442326981, + "loss": 1.3983, + "theoretical_loss": 3.4216674230906325, + "tokens_seen": 2283274240 + }, + { + "epoch": 0.69, + "learning_rate": 0.0002491313941825477, + "loss": 1.3301, + "theoretical_loss": 3.4216595126559555, + "tokens_seen": 2283339776 + }, + { + "epoch": 0.69, + "learning_rate": 0.0002491153460381144, + "loss": 1.199, + "theoretical_loss": 3.4216516025118895, + "tokens_seen": 2283405312 + }, + { + "epoch": 0.69, + "learning_rate": 0.0002490992978936811, + "loss": 1.2379, + "theoretical_loss": 3.421643692658415, + "tokens_seen": 2283470848 + }, + { + "epoch": 0.69, + "learning_rate": 0.00024908324974924775, + "loss": 1.3861, + "theoretical_loss": 3.421635783095513, + "tokens_seen": 2283536384 + }, + { + "epoch": 0.69, + "learning_rate": 0.00024906720160481445, + "loss": 1.2352, + "theoretical_loss": 3.4216278738231654, + "tokens_seen": 2283601920 + }, + { + "epoch": 0.69, + "learning_rate": 0.00024905115346038115, + "loss": 1.319, + "theoretical_loss": 3.4216199648413523, + "tokens_seen": 2283667456 + }, + { + "epoch": 0.69, + "learning_rate": 0.00024903510531594785, + "loss": 1.2361, + "theoretical_loss": 3.421612056150055, + "tokens_seen": 2283732992 + }, + { + "epoch": 0.69, + "learning_rate": 0.00024901905717151455, + "loss": 1.3978, + "theoretical_loss": 3.421604147749254, + "tokens_seen": 2283798528 + }, + { + "epoch": 0.69, + "learning_rate": 0.00024900300902708125, + "loss": 1.3316, + "theoretical_loss": 3.421596239638931, + "tokens_seen": 2283864064 + }, + { + "epoch": 0.69, + "objective/train/docs_used": 831582, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.3100173473358154, + "objective/train/theoretical_loss": 3.4215883318190667, + "objective/train/tokens_used": 2304389600, + "theoretical_loss": 3.4215883318190667, + "tokens_seen": 2283929600 + }, + { + "epoch": 0.69, + "learning_rate": 0.00024898696088264795, + "loss": 1.2673, + "theoretical_loss": 3.4215883318190667, + "tokens_seen": 2283929600 + }, + { + "epoch": 0.69, + "learning_rate": 0.00024897091273821465, + "loss": 1.3571, + "theoretical_loss": 3.421580424289642, + "tokens_seen": 2283995136 + }, + { + "epoch": 0.69, + "learning_rate": 0.00024895486459378135, + "loss": 1.2672, + "theoretical_loss": 3.421572517050638, + "tokens_seen": 2284060672 + }, + { + "epoch": 0.69, + "learning_rate": 0.00024893881644934805, + "loss": 1.3963, + "theoretical_loss": 3.4215646101020356, + "tokens_seen": 2284126208 + }, + { + "epoch": 0.69, + "learning_rate": 0.00024892276830491475, + "loss": 1.306, + "theoretical_loss": 3.4215567034438163, + "tokens_seen": 2284191744 + }, + { + "epoch": 0.69, + "learning_rate": 0.00024890672016048145, + "loss": 1.2844, + "theoretical_loss": 3.4215487970759604, + "tokens_seen": 2284257280 + }, + { + "epoch": 0.69, + "learning_rate": 0.00024889067201604815, + "loss": 1.2621, + "theoretical_loss": 3.4215408909984495, + "tokens_seen": 2284322816 + }, + { + "epoch": 0.69, + "learning_rate": 0.00024887462387161485, + "loss": 1.3776, + "theoretical_loss": 3.4215329852112637, + "tokens_seen": 2284388352 + }, + { + "epoch": 0.69, + "learning_rate": 0.00024885857572718155, + "loss": 1.2419, + "theoretical_loss": 3.4215250797143852, + "tokens_seen": 2284453888 + }, + { + "epoch": 0.69, + "learning_rate": 0.00024884252758274825, + "loss": 1.3801, + "theoretical_loss": 3.4215171745077946, + "tokens_seen": 2284519424 + }, + { + "epoch": 0.69, + "learning_rate": 0.00024882647943831496, + "loss": 1.332, + "theoretical_loss": 3.421509269591472, + "tokens_seen": 2284584960 + }, + { + "epoch": 0.69, + "learning_rate": 0.00024881043129388166, + "loss": 1.2832, + "theoretical_loss": 3.4215013649654, + "tokens_seen": 2284650496 + }, + { + "epoch": 0.69, + "learning_rate": 0.00024879438314944836, + "loss": 1.2883, + "theoretical_loss": 3.4214934606295584, + "tokens_seen": 2284716032 + }, + { + "epoch": 0.69, + "learning_rate": 0.00024877833500501506, + "loss": 1.2342, + "theoretical_loss": 3.4214855565839284, + "tokens_seen": 2284781568 + }, + { + "epoch": 0.69, + "learning_rate": 0.00024876228686058176, + "loss": 1.3083, + "theoretical_loss": 3.421477652828491, + "tokens_seen": 2284847104 + }, + { + "epoch": 0.69, + "learning_rate": 0.00024874623871614846, + "loss": 1.4924, + "theoretical_loss": 3.4214697493632285, + "tokens_seen": 2284912640 + }, + { + "epoch": 0.69, + "learning_rate": 0.00024873019057171516, + "loss": 1.2477, + "theoretical_loss": 3.4214618461881203, + "tokens_seen": 2284978176 + }, + { + "epoch": 0.69, + "learning_rate": 0.00024871414242728186, + "loss": 1.3226, + "theoretical_loss": 3.4214539433031477, + "tokens_seen": 2285043712 + }, + { + "epoch": 0.69, + "learning_rate": 0.00024869809428284856, + "loss": 1.3012, + "theoretical_loss": 3.4214460407082923, + "tokens_seen": 2285109248 + }, + { + "epoch": 0.69, + "learning_rate": 0.00024868204613841526, + "loss": 1.3204, + "theoretical_loss": 3.421438138403535, + "tokens_seen": 2285174784 + }, + { + "epoch": 0.69, + "learning_rate": 0.00024866599799398196, + "loss": 1.2178, + "theoretical_loss": 3.4214302363888565, + "tokens_seen": 2285240320 + }, + { + "epoch": 0.69, + "learning_rate": 0.00024864994984954866, + "loss": 1.2192, + "theoretical_loss": 3.4214223346642383, + "tokens_seen": 2285305856 + }, + { + "epoch": 0.69, + "learning_rate": 0.00024863390170511536, + "loss": 1.2205, + "theoretical_loss": 3.421414433229661, + "tokens_seen": 2285371392 + }, + { + "epoch": 0.69, + "learning_rate": 0.00024861785356068206, + "loss": 1.2522, + "theoretical_loss": 3.4214065320851055, + "tokens_seen": 2285436928 + }, + { + "epoch": 0.69, + "learning_rate": 0.00024860180541624876, + "loss": 1.2522, + "theoretical_loss": 3.4213986312305535, + "tokens_seen": 2285502464 + }, + { + "epoch": 0.69, + "learning_rate": 0.00024858575727181546, + "loss": 1.3298, + "theoretical_loss": 3.4213907306659856, + "tokens_seen": 2285568000 + }, + { + "epoch": 0.69, + "learning_rate": 0.00024856970912738217, + "loss": 1.297, + "theoretical_loss": 3.4213828303913827, + "tokens_seen": 2285633536 + }, + { + "epoch": 0.69, + "learning_rate": 0.00024855366098294887, + "loss": 1.1833, + "theoretical_loss": 3.4213749304067265, + "tokens_seen": 2285699072 + }, + { + "epoch": 0.69, + "learning_rate": 0.00024853761283851557, + "loss": 1.2912, + "theoretical_loss": 3.4213670307119974, + "tokens_seen": 2285764608 + }, + { + "epoch": 0.69, + "learning_rate": 0.00024852156469408227, + "loss": 1.3343, + "theoretical_loss": 3.4213591313071765, + "tokens_seen": 2285830144 + }, + { + "epoch": 0.69, + "learning_rate": 0.00024850551654964897, + "loss": 1.3109, + "theoretical_loss": 3.421351232192245, + "tokens_seen": 2285895680 + }, + { + "epoch": 0.69, + "learning_rate": 0.00024848946840521567, + "loss": 1.3715, + "theoretical_loss": 3.4213433333671843, + "tokens_seen": 2285961216 + }, + { + "epoch": 0.69, + "learning_rate": 0.00024847342026078237, + "loss": 1.3112, + "theoretical_loss": 3.421335434831975, + "tokens_seen": 2286026752 + }, + { + "epoch": 0.69, + "learning_rate": 0.00024845737211634907, + "loss": 1.2365, + "theoretical_loss": 3.421327536586598, + "tokens_seen": 2286092288 + }, + { + "epoch": 0.69, + "learning_rate": 0.00024844132397191577, + "loss": 1.2808, + "theoretical_loss": 3.4213196386310347, + "tokens_seen": 2286157824 + }, + { + "epoch": 0.69, + "learning_rate": 0.00024842527582748247, + "loss": 1.2562, + "theoretical_loss": 3.4213117409652662, + "tokens_seen": 2286223360 + }, + { + "epoch": 0.69, + "learning_rate": 0.00024840922768304917, + "loss": 1.2595, + "theoretical_loss": 3.4213038435892735, + "tokens_seen": 2286288896 + }, + { + "epoch": 0.69, + "learning_rate": 0.00024839317953861587, + "loss": 1.2529, + "theoretical_loss": 3.4212959465030375, + "tokens_seen": 2286354432 + }, + { + "epoch": 0.69, + "learning_rate": 0.0002483771313941826, + "loss": 1.2359, + "theoretical_loss": 3.4212880497065394, + "tokens_seen": 2286419968 + }, + { + "epoch": 0.69, + "learning_rate": 0.0002483610832497493, + "loss": 1.2908, + "theoretical_loss": 3.4212801531997603, + "tokens_seen": 2286485504 + }, + { + "epoch": 0.69, + "learning_rate": 0.000248345035105316, + "loss": 1.3504, + "theoretical_loss": 3.4212722569826814, + "tokens_seen": 2286551040 + }, + { + "epoch": 0.69, + "learning_rate": 0.0002483289869608827, + "loss": 1.247, + "theoretical_loss": 3.421264361055283, + "tokens_seen": 2286616576 + }, + { + "epoch": 0.69, + "learning_rate": 0.0002483129388164493, + "loss": 1.2872, + "theoretical_loss": 3.4212564654175472, + "tokens_seen": 2286682112 + }, + { + "epoch": 0.69, + "learning_rate": 0.000248296890672016, + "loss": 1.1681, + "theoretical_loss": 3.421248570069454, + "tokens_seen": 2286747648 + }, + { + "epoch": 0.69, + "learning_rate": 0.0002482808425275827, + "loss": 1.2818, + "theoretical_loss": 3.421240675010986, + "tokens_seen": 2286813184 + }, + { + "epoch": 0.69, + "learning_rate": 0.0002482647943831494, + "loss": 1.3198, + "theoretical_loss": 3.4212327802421227, + "tokens_seen": 2286878720 + }, + { + "epoch": 0.69, + "learning_rate": 0.0002482487462387161, + "loss": 1.3093, + "theoretical_loss": 3.421224885762846, + "tokens_seen": 2286944256 + }, + { + "epoch": 0.69, + "learning_rate": 0.0002482326980942828, + "loss": 1.2708, + "theoretical_loss": 3.4212169915731367, + "tokens_seen": 2287009792 + }, + { + "epoch": 0.69, + "learning_rate": 0.0002482166499498495, + "loss": 1.4226, + "theoretical_loss": 3.4212090976729765, + "tokens_seen": 2287075328 + }, + { + "epoch": 0.69, + "learning_rate": 0.0002482006018054162, + "loss": 1.2023, + "theoretical_loss": 3.4212012040623456, + "tokens_seen": 2287140864 + }, + { + "epoch": 0.69, + "objective/train/docs_used": 832763, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.312143087387085, + "objective/train/theoretical_loss": 3.4211933107412253, + "objective/train/tokens_used": 2307666400, + "theoretical_loss": 3.4211933107412253, + "tokens_seen": 2287206400 + }, + { + "epoch": 0.69, + "learning_rate": 0.0002481845536609829, + "loss": 1.2339, + "theoretical_loss": 3.4211933107412253, + "tokens_seen": 2287206400 + }, + { + "epoch": 0.69, + "learning_rate": 0.0002481685055165496, + "loss": 1.3965, + "theoretical_loss": 3.421185417709597, + "tokens_seen": 2287271936 + }, + { + "epoch": 0.69, + "learning_rate": 0.00024815245737211633, + "loss": 1.2346, + "theoretical_loss": 3.421177524967442, + "tokens_seen": 2287337472 + }, + { + "epoch": 0.69, + "learning_rate": 0.00024813640922768303, + "loss": 1.1788, + "theoretical_loss": 3.4211696325147405, + "tokens_seen": 2287403008 + }, + { + "epoch": 0.69, + "learning_rate": 0.00024812036108324973, + "loss": 1.3116, + "theoretical_loss": 3.4211617403514745, + "tokens_seen": 2287468544 + }, + { + "epoch": 0.69, + "learning_rate": 0.00024810431293881643, + "loss": 1.3194, + "theoretical_loss": 3.4211538484776245, + "tokens_seen": 2287534080 + }, + { + "epoch": 0.69, + "learning_rate": 0.00024808826479438313, + "loss": 1.2749, + "theoretical_loss": 3.421145956893172, + "tokens_seen": 2287599616 + }, + { + "epoch": 0.69, + "learning_rate": 0.00024807221664994983, + "loss": 1.3123, + "theoretical_loss": 3.4211380655980976, + "tokens_seen": 2287665152 + }, + { + "epoch": 0.69, + "learning_rate": 0.00024805616850551653, + "loss": 1.3489, + "theoretical_loss": 3.421130174592383, + "tokens_seen": 2287730688 + }, + { + "epoch": 0.69, + "learning_rate": 0.0002480401203610833, + "loss": 1.2924, + "theoretical_loss": 3.421122283876009, + "tokens_seen": 2287796224 + }, + { + "epoch": 0.69, + "learning_rate": 0.00024802407221665, + "loss": 1.3507, + "theoretical_loss": 3.4211143934489567, + "tokens_seen": 2287861760 + }, + { + "epoch": 0.69, + "learning_rate": 0.0002480080240722167, + "loss": 1.3059, + "theoretical_loss": 3.421106503311207, + "tokens_seen": 2287927296 + }, + { + "epoch": 0.69, + "learning_rate": 0.0002479919759277834, + "loss": 1.2538, + "theoretical_loss": 3.4210986134627417, + "tokens_seen": 2287992832 + }, + { + "epoch": 0.69, + "learning_rate": 0.0002479759277833501, + "loss": 1.2833, + "theoretical_loss": 3.421090723903541, + "tokens_seen": 2288058368 + }, + { + "epoch": 0.69, + "learning_rate": 0.0002479598796389168, + "loss": 1.2325, + "theoretical_loss": 3.4210828346335864, + "tokens_seen": 2288123904 + }, + { + "epoch": 0.69, + "learning_rate": 0.0002479438314944835, + "loss": 1.3909, + "theoretical_loss": 3.4210749456528595, + "tokens_seen": 2288189440 + }, + { + "epoch": 0.69, + "learning_rate": 0.0002479277833500502, + "loss": 1.2219, + "theoretical_loss": 3.421067056961341, + "tokens_seen": 2288254976 + }, + { + "epoch": 0.69, + "learning_rate": 0.0002479117352056169, + "loss": 1.3715, + "theoretical_loss": 3.4210591685590113, + "tokens_seen": 2288320512 + }, + { + "epoch": 0.69, + "learning_rate": 0.0002478956870611836, + "loss": 1.2834, + "theoretical_loss": 3.4210512804458526, + "tokens_seen": 2288386048 + }, + { + "epoch": 0.69, + "learning_rate": 0.0002478796389167503, + "loss": 1.3258, + "theoretical_loss": 3.4210433926218458, + "tokens_seen": 2288451584 + }, + { + "epoch": 0.69, + "learning_rate": 0.000247863590772317, + "loss": 1.2837, + "theoretical_loss": 3.4210355050869716, + "tokens_seen": 2288517120 + }, + { + "epoch": 0.69, + "learning_rate": 0.0002478475426278837, + "loss": 1.3009, + "theoretical_loss": 3.4210276178412116, + "tokens_seen": 2288582656 + }, + { + "epoch": 0.69, + "learning_rate": 0.0002478314944834504, + "loss": 1.3624, + "theoretical_loss": 3.421019730884546, + "tokens_seen": 2288648192 + }, + { + "epoch": 0.69, + "learning_rate": 0.0002478154463390171, + "loss": 1.2896, + "theoretical_loss": 3.4210118442169573, + "tokens_seen": 2288713728 + }, + { + "epoch": 0.69, + "learning_rate": 0.0002477993981945838, + "loss": 1.289, + "theoretical_loss": 3.4210039578384257, + "tokens_seen": 2288779264 + }, + { + "epoch": 0.69, + "learning_rate": 0.0002477833500501505, + "loss": 1.3408, + "theoretical_loss": 3.4209960717489327, + "tokens_seen": 2288844800 + }, + { + "epoch": 0.69, + "learning_rate": 0.0002477673019057172, + "loss": 1.3745, + "theoretical_loss": 3.420988185948459, + "tokens_seen": 2288910336 + }, + { + "epoch": 0.69, + "learning_rate": 0.0002477512537612839, + "loss": 1.2122, + "theoretical_loss": 3.4209803004369865, + "tokens_seen": 2288975872 + }, + { + "epoch": 0.69, + "learning_rate": 0.0002477352056168506, + "loss": 1.2994, + "theoretical_loss": 3.4209724152144956, + "tokens_seen": 2289041408 + }, + { + "epoch": 0.69, + "learning_rate": 0.0002477191574724173, + "loss": 1.3048, + "theoretical_loss": 3.420964530280968, + "tokens_seen": 2289106944 + }, + { + "epoch": 0.69, + "learning_rate": 0.000247703109327984, + "loss": 1.3295, + "theoretical_loss": 3.4209566456363842, + "tokens_seen": 2289172480 + }, + { + "epoch": 0.69, + "learning_rate": 0.00024768706118355065, + "loss": 1.3393, + "theoretical_loss": 3.4209487612807257, + "tokens_seen": 2289238016 + }, + { + "epoch": 0.69, + "learning_rate": 0.00024767101303911735, + "loss": 1.2835, + "theoretical_loss": 3.4209408772139738, + "tokens_seen": 2289303552 + }, + { + "epoch": 0.69, + "learning_rate": 0.00024765496489468405, + "loss": 1.3047, + "theoretical_loss": 3.4209329934361095, + "tokens_seen": 2289369088 + }, + { + "epoch": 0.69, + "learning_rate": 0.00024763891675025075, + "loss": 1.3509, + "theoretical_loss": 3.4209251099471136, + "tokens_seen": 2289434624 + }, + { + "epoch": 0.69, + "learning_rate": 0.00024762286860581745, + "loss": 1.3167, + "theoretical_loss": 3.4209172267469676, + "tokens_seen": 2289500160 + }, + { + "epoch": 0.69, + "learning_rate": 0.00024760682046138415, + "loss": 1.311, + "theoretical_loss": 3.420909343835653, + "tokens_seen": 2289565696 + }, + { + "epoch": 0.69, + "learning_rate": 0.00024759077231695085, + "loss": 1.4028, + "theoretical_loss": 3.42090146121315, + "tokens_seen": 2289631232 + }, + { + "epoch": 0.69, + "learning_rate": 0.00024757472417251755, + "loss": 1.3092, + "theoretical_loss": 3.420893578879441, + "tokens_seen": 2289696768 + }, + { + "epoch": 0.69, + "learning_rate": 0.00024755867602808425, + "loss": 1.2443, + "theoretical_loss": 3.420885696834506, + "tokens_seen": 2289762304 + }, + { + "epoch": 0.69, + "learning_rate": 0.00024754262788365095, + "loss": 1.3085, + "theoretical_loss": 3.4208778150783266, + "tokens_seen": 2289827840 + }, + { + "epoch": 0.69, + "learning_rate": 0.00024752657973921765, + "loss": 1.2836, + "theoretical_loss": 3.4208699336108843, + "tokens_seen": 2289893376 + }, + { + "epoch": 0.69, + "learning_rate": 0.00024751053159478435, + "loss": 1.1734, + "theoretical_loss": 3.4208620524321596, + "tokens_seen": 2289958912 + }, + { + "epoch": 0.69, + "learning_rate": 0.00024749448345035105, + "loss": 1.3495, + "theoretical_loss": 3.420854171542134, + "tokens_seen": 2290024448 + }, + { + "epoch": 0.69, + "learning_rate": 0.00024747843530591775, + "loss": 1.2976, + "theoretical_loss": 3.420846290940789, + "tokens_seen": 2290089984 + }, + { + "epoch": 0.69, + "learning_rate": 0.00024746238716148445, + "loss": 1.3782, + "theoretical_loss": 3.420838410628105, + "tokens_seen": 2290155520 + }, + { + "epoch": 0.69, + "learning_rate": 0.00024744633901705116, + "loss": 1.3154, + "theoretical_loss": 3.420830530604064, + "tokens_seen": 2290221056 + }, + { + "epoch": 0.69, + "learning_rate": 0.00024743029087261786, + "loss": 1.3162, + "theoretical_loss": 3.4208226508686463, + "tokens_seen": 2290286592 + }, + { + "epoch": 0.69, + "learning_rate": 0.00024741424272818456, + "loss": 1.3191, + "theoretical_loss": 3.420814771421834, + "tokens_seen": 2290352128 + }, + { + "epoch": 0.69, + "learning_rate": 0.00024739819458375126, + "loss": 1.2489, + "theoretical_loss": 3.4208068922636077, + "tokens_seen": 2290417664 + }, + { + "epoch": 0.69, + "objective/train/docs_used": 833966, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.3701281547546387, + "objective/train/theoretical_loss": 3.4207990133939488, + "objective/train/tokens_used": 2310943200, + "theoretical_loss": 3.4207990133939488, + "tokens_seen": 2290483200 + }, + { + "epoch": 0.69, + "learning_rate": 0.00024738214643931796, + "loss": 1.2947, + "theoretical_loss": 3.4207990133939488, + "tokens_seen": 2290483200 + }, + { + "epoch": 0.69, + "learning_rate": 0.00024736609829488466, + "loss": 1.3981, + "theoretical_loss": 3.4207911348128377, + "tokens_seen": 2290548736 + }, + { + "epoch": 0.69, + "learning_rate": 0.00024735005015045136, + "loss": 1.2142, + "theoretical_loss": 3.4207832565202567, + "tokens_seen": 2290614272 + }, + { + "epoch": 0.69, + "learning_rate": 0.00024733400200601806, + "loss": 1.2948, + "theoretical_loss": 3.4207753785161867, + "tokens_seen": 2290679808 + }, + { + "epoch": 0.69, + "learning_rate": 0.00024731795386158476, + "loss": 1.2968, + "theoretical_loss": 3.4207675008006087, + "tokens_seen": 2290745344 + }, + { + "epoch": 0.69, + "learning_rate": 0.00024730190571715146, + "loss": 1.2472, + "theoretical_loss": 3.4207596233735034, + "tokens_seen": 2290810880 + }, + { + "epoch": 0.69, + "learning_rate": 0.00024728585757271816, + "loss": 1.3085, + "theoretical_loss": 3.4207517462348527, + "tokens_seen": 2290876416 + }, + { + "epoch": 0.69, + "learning_rate": 0.00024726980942828486, + "loss": 1.2273, + "theoretical_loss": 3.420743869384638, + "tokens_seen": 2290941952 + }, + { + "epoch": 0.69, + "learning_rate": 0.00024725376128385156, + "loss": 1.2469, + "theoretical_loss": 3.4207359928228396, + "tokens_seen": 2291007488 + }, + { + "epoch": 0.69, + "learning_rate": 0.00024723771313941826, + "loss": 1.2479, + "theoretical_loss": 3.420728116549439, + "tokens_seen": 2291073024 + }, + { + "epoch": 0.69, + "learning_rate": 0.00024722166499498496, + "loss": 1.2617, + "theoretical_loss": 3.4207202405644175, + "tokens_seen": 2291138560 + }, + { + "epoch": 0.69, + "learning_rate": 0.00024720561685055167, + "loss": 1.2545, + "theoretical_loss": 3.4207123648677564, + "tokens_seen": 2291204096 + }, + { + "epoch": 0.69, + "learning_rate": 0.00024718956870611837, + "loss": 1.2864, + "theoretical_loss": 3.420704489459437, + "tokens_seen": 2291269632 + }, + { + "epoch": 0.69, + "learning_rate": 0.00024717352056168507, + "loss": 1.2089, + "theoretical_loss": 3.42069661433944, + "tokens_seen": 2291335168 + }, + { + "epoch": 0.69, + "learning_rate": 0.00024715747241725177, + "loss": 1.3163, + "theoretical_loss": 3.420688739507747, + "tokens_seen": 2291400704 + }, + { + "epoch": 0.69, + "learning_rate": 0.00024714142427281847, + "loss": 1.3361, + "theoretical_loss": 3.420680864964339, + "tokens_seen": 2291466240 + }, + { + "epoch": 0.69, + "learning_rate": 0.00024712537612838517, + "loss": 1.3086, + "theoretical_loss": 3.4206729907091975, + "tokens_seen": 2291531776 + }, + { + "epoch": 0.69, + "learning_rate": 0.00024710932798395187, + "loss": 1.2531, + "theoretical_loss": 3.4206651167423034, + "tokens_seen": 2291597312 + }, + { + "epoch": 0.69, + "learning_rate": 0.00024709327983951857, + "loss": 1.3243, + "theoretical_loss": 3.420657243063638, + "tokens_seen": 2291662848 + }, + { + "epoch": 0.69, + "learning_rate": 0.00024707723169508527, + "loss": 1.2781, + "theoretical_loss": 3.420649369673183, + "tokens_seen": 2291728384 + }, + { + "epoch": 0.69, + "learning_rate": 0.00024706118355065197, + "loss": 1.2046, + "theoretical_loss": 3.4206414965709184, + "tokens_seen": 2291793920 + }, + { + "epoch": 0.69, + "learning_rate": 0.00024704513540621867, + "loss": 1.2811, + "theoretical_loss": 3.4206336237568267, + "tokens_seen": 2291859456 + }, + { + "epoch": 0.69, + "learning_rate": 0.00024702908726178537, + "loss": 1.3783, + "theoretical_loss": 3.420625751230888, + "tokens_seen": 2291924992 + }, + { + "epoch": 0.69, + "learning_rate": 0.00024701303911735207, + "loss": 1.2139, + "theoretical_loss": 3.4206178789930846, + "tokens_seen": 2291990528 + }, + { + "epoch": 0.69, + "learning_rate": 0.0002469969909729188, + "loss": 1.2901, + "theoretical_loss": 3.4206100070433965, + "tokens_seen": 2292056064 + }, + { + "epoch": 0.69, + "learning_rate": 0.0002469809428284855, + "loss": 1.3763, + "theoretical_loss": 3.420602135381806, + "tokens_seen": 2292121600 + }, + { + "epoch": 0.69, + "learning_rate": 0.0002469648946840522, + "loss": 1.3262, + "theoretical_loss": 3.420594264008294, + "tokens_seen": 2292187136 + }, + { + "epoch": 0.69, + "learning_rate": 0.0002469488465396189, + "loss": 1.3195, + "theoretical_loss": 3.4205863929228415, + "tokens_seen": 2292252672 + }, + { + "epoch": 0.69, + "learning_rate": 0.0002469327983951856, + "loss": 1.2762, + "theoretical_loss": 3.42057852212543, + "tokens_seen": 2292318208 + }, + { + "epoch": 0.69, + "learning_rate": 0.0002469167502507522, + "loss": 1.3166, + "theoretical_loss": 3.4205706516160403, + "tokens_seen": 2292383744 + }, + { + "epoch": 0.69, + "learning_rate": 0.0002469007021063189, + "loss": 1.1814, + "theoretical_loss": 3.420562781394654, + "tokens_seen": 2292449280 + }, + { + "epoch": 0.69, + "learning_rate": 0.0002468846539618856, + "loss": 1.3249, + "theoretical_loss": 3.4205549114612523, + "tokens_seen": 2292514816 + }, + { + "epoch": 0.69, + "learning_rate": 0.0002468686058174523, + "loss": 1.3275, + "theoretical_loss": 3.4205470418158166, + "tokens_seen": 2292580352 + }, + { + "epoch": 0.69, + "learning_rate": 0.000246852557673019, + "loss": 1.2968, + "theoretical_loss": 3.4205391724583274, + "tokens_seen": 2292645888 + }, + { + "epoch": 0.69, + "learning_rate": 0.0002468365095285858, + "loss": 1.3591, + "theoretical_loss": 3.420531303388767, + "tokens_seen": 2292711424 + }, + { + "epoch": 0.69, + "learning_rate": 0.0002468204613841525, + "loss": 1.3086, + "theoretical_loss": 3.420523434607116, + "tokens_seen": 2292776960 + }, + { + "epoch": 0.69, + "learning_rate": 0.0002468044132397192, + "loss": 1.2351, + "theoretical_loss": 3.4205155661133553, + "tokens_seen": 2292842496 + }, + { + "epoch": 0.69, + "learning_rate": 0.0002467883650952859, + "loss": 1.4014, + "theoretical_loss": 3.4205076979074667, + "tokens_seen": 2292908032 + }, + { + "epoch": 0.69, + "learning_rate": 0.0002467723169508526, + "loss": 1.3429, + "theoretical_loss": 3.420499829989431, + "tokens_seen": 2292973568 + }, + { + "epoch": 0.69, + "learning_rate": 0.0002467562688064193, + "loss": 1.4206, + "theoretical_loss": 3.42049196235923, + "tokens_seen": 2293039104 + }, + { + "epoch": 0.69, + "learning_rate": 0.000246740220661986, + "loss": 1.2056, + "theoretical_loss": 3.4204840950168447, + "tokens_seen": 2293104640 + }, + { + "epoch": 0.69, + "learning_rate": 0.0002467241725175527, + "loss": 1.1648, + "theoretical_loss": 3.420476227962256, + "tokens_seen": 2293170176 + }, + { + "epoch": 0.69, + "learning_rate": 0.0002467081243731194, + "loss": 1.4013, + "theoretical_loss": 3.420468361195446, + "tokens_seen": 2293235712 + }, + { + "epoch": 0.69, + "learning_rate": 0.0002466920762286861, + "loss": 1.306, + "theoretical_loss": 3.4204604947163952, + "tokens_seen": 2293301248 + }, + { + "epoch": 0.69, + "learning_rate": 0.0002466760280842528, + "loss": 1.1943, + "theoretical_loss": 3.4204526285250854, + "tokens_seen": 2293366784 + }, + { + "epoch": 0.69, + "learning_rate": 0.0002466599799398195, + "loss": 1.2389, + "theoretical_loss": 3.420444762621497, + "tokens_seen": 2293432320 + }, + { + "epoch": 0.69, + "learning_rate": 0.0002466439317953862, + "loss": 1.2416, + "theoretical_loss": 3.4204368970056116, + "tokens_seen": 2293497856 + }, + { + "epoch": 0.7, + "learning_rate": 0.0002466278836509529, + "loss": 1.3428, + "theoretical_loss": 3.4204290316774113, + "tokens_seen": 2293563392 + }, + { + "epoch": 0.7, + "learning_rate": 0.0002466118355065196, + "loss": 1.2887, + "theoretical_loss": 3.420421166636876, + "tokens_seen": 2293628928 + }, + { + "epoch": 0.7, + "learning_rate": 0.0002465957873620863, + "loss": 1.3976, + "theoretical_loss": 3.420413301883988, + "tokens_seen": 2293694464 + }, + { + "debugging/Compilability": 1.0, + "debugging/distinct-1-grams": 0.7361949066418735, + "debugging/entropy-1-grams": 5.412433223050153, + "debugging/length": 455.6666666666667, + "debugging/num_segments": 15, + "debugging/score": 0.0066926720884253575, + "debugging/score_std": 0.006256546529562537, + "epoch": 0.7, + "objective/train/docs_used": 835219, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.2255016565322876, + "objective/train/theoretical_loss": 3.4204054374187285, + "objective/train/tokens_used": 2314220000, + "theoretical_loss": 3.4204054374187285, + "tokens_seen": 2293760000 + }, + { + "epoch": 0.7, + "learning_rate": 0.000246579739217653, + "loss": 1.2722, + "theoretical_loss": 3.4204054374187285, + "tokens_seen": 2293760000 + }, + { + "epoch": 0.7, + "learning_rate": 0.0002465636910732197, + "loss": 1.288, + "theoretical_loss": 3.420397573241078, + "tokens_seen": 2293825536 + }, + { + "epoch": 0.7, + "learning_rate": 0.0002465476429287864, + "loss": 1.3625, + "theoretical_loss": 3.4203897093510185, + "tokens_seen": 2293891072 + }, + { + "epoch": 0.7, + "learning_rate": 0.0002465315947843531, + "loss": 1.3222, + "theoretical_loss": 3.4203818457485307, + "tokens_seen": 2293956608 + }, + { + "epoch": 0.7, + "learning_rate": 0.0002465155466399198, + "loss": 1.3125, + "theoretical_loss": 3.4203739824335964, + "tokens_seen": 2294022144 + }, + { + "epoch": 0.7, + "learning_rate": 0.0002464994984954865, + "loss": 1.349, + "theoretical_loss": 3.4203661194061965, + "tokens_seen": 2294087680 + }, + { + "epoch": 0.7, + "learning_rate": 0.0002464834503510532, + "loss": 1.3505, + "theoretical_loss": 3.420358256666313, + "tokens_seen": 2294153216 + }, + { + "epoch": 0.7, + "learning_rate": 0.0002464674022066199, + "loss": 1.3231, + "theoretical_loss": 3.420350394213926, + "tokens_seen": 2294218752 + }, + { + "epoch": 0.7, + "learning_rate": 0.0002464513540621866, + "loss": 1.4435, + "theoretical_loss": 3.4203425320490175, + "tokens_seen": 2294284288 + }, + { + "epoch": 0.7, + "learning_rate": 0.0002464353059177533, + "loss": 1.3882, + "theoretical_loss": 3.4203346701715684, + "tokens_seen": 2294349824 + }, + { + "epoch": 0.7, + "learning_rate": 0.00024641925777332, + "loss": 1.4245, + "theoretical_loss": 3.4203268085815606, + "tokens_seen": 2294415360 + }, + { + "epoch": 0.7, + "learning_rate": 0.0002464032096288867, + "loss": 1.3831, + "theoretical_loss": 3.420318947278975, + "tokens_seen": 2294480896 + }, + { + "epoch": 0.7, + "learning_rate": 0.0002463871614844534, + "loss": 1.2933, + "theoretical_loss": 3.4203110862637924, + "tokens_seen": 2294546432 + }, + { + "epoch": 0.7, + "learning_rate": 0.0002463711133400201, + "loss": 1.3367, + "theoretical_loss": 3.420303225535995, + "tokens_seen": 2294611968 + }, + { + "epoch": 0.7, + "learning_rate": 0.0002463550651955868, + "loss": 1.301, + "theoretical_loss": 3.4202953650955634, + "tokens_seen": 2294677504 + }, + { + "epoch": 0.7, + "learning_rate": 0.0002463390170511535, + "loss": 1.2797, + "theoretical_loss": 3.4202875049424795, + "tokens_seen": 2294743040 + }, + { + "epoch": 0.7, + "learning_rate": 0.0002463229689067202, + "loss": 1.2184, + "theoretical_loss": 3.420279645076724, + "tokens_seen": 2294808576 + }, + { + "epoch": 0.7, + "learning_rate": 0.0002463069207622869, + "loss": 1.322, + "theoretical_loss": 3.4202717854982785, + "tokens_seen": 2294874112 + }, + { + "epoch": 0.7, + "learning_rate": 0.00024629087261785355, + "loss": 1.1975, + "theoretical_loss": 3.420263926207124, + "tokens_seen": 2294939648 + }, + { + "epoch": 0.7, + "learning_rate": 0.00024627482447342025, + "loss": 1.209, + "theoretical_loss": 3.4202560672032423, + "tokens_seen": 2295005184 + }, + { + "epoch": 0.7, + "learning_rate": 0.00024625877632898695, + "loss": 1.298, + "theoretical_loss": 3.420248208486614, + "tokens_seen": 2295070720 + }, + { + "epoch": 0.7, + "learning_rate": 0.00024624272818455365, + "loss": 1.265, + "theoretical_loss": 3.420240350057221, + "tokens_seen": 2295136256 + }, + { + "epoch": 0.7, + "learning_rate": 0.00024622668004012035, + "loss": 1.2555, + "theoretical_loss": 3.4202324919150446, + "tokens_seen": 2295201792 + }, + { + "epoch": 0.7, + "learning_rate": 0.00024621063189568705, + "loss": 1.2493, + "theoretical_loss": 3.420224634060066, + "tokens_seen": 2295267328 + }, + { + "epoch": 0.7, + "learning_rate": 0.00024619458375125375, + "loss": 1.287, + "theoretical_loss": 3.420216776492266, + "tokens_seen": 2295332864 + }, + { + "epoch": 0.7, + "learning_rate": 0.00024617853560682045, + "loss": 1.1929, + "theoretical_loss": 3.420208919211627, + "tokens_seen": 2295398400 + }, + { + "epoch": 0.7, + "learning_rate": 0.00024616248746238715, + "loss": 1.2653, + "theoretical_loss": 3.420201062218129, + "tokens_seen": 2295463936 + }, + { + "epoch": 0.7, + "learning_rate": 0.00024614643931795385, + "loss": 1.1853, + "theoretical_loss": 3.420193205511754, + "tokens_seen": 2295529472 + }, + { + "epoch": 0.7, + "learning_rate": 0.00024613039117352055, + "loss": 1.2881, + "theoretical_loss": 3.4201853490924834, + "tokens_seen": 2295595008 + }, + { + "epoch": 0.7, + "learning_rate": 0.00024611434302908725, + "loss": 1.3557, + "theoretical_loss": 3.420177492960298, + "tokens_seen": 2295660544 + }, + { + "epoch": 0.7, + "learning_rate": 0.00024609829488465395, + "loss": 1.3753, + "theoretical_loss": 3.42016963711518, + "tokens_seen": 2295726080 + }, + { + "epoch": 0.7, + "learning_rate": 0.00024608224674022066, + "loss": 1.354, + "theoretical_loss": 3.42016178155711, + "tokens_seen": 2295791616 + }, + { + "epoch": 0.7, + "learning_rate": 0.00024606619859578736, + "loss": 1.3079, + "theoretical_loss": 3.4201539262860696, + "tokens_seen": 2295857152 + }, + { + "epoch": 0.7, + "learning_rate": 0.00024605015045135406, + "loss": 1.2988, + "theoretical_loss": 3.4201460713020397, + "tokens_seen": 2295922688 + }, + { + "epoch": 0.7, + "learning_rate": 0.00024603410230692076, + "loss": 1.302, + "theoretical_loss": 3.420138216605002, + "tokens_seen": 2295988224 + }, + { + "epoch": 0.7, + "learning_rate": 0.00024601805416248746, + "loss": 1.3881, + "theoretical_loss": 3.4201303621949384, + "tokens_seen": 2296053760 + }, + { + "epoch": 0.7, + "learning_rate": 0.00024600200601805416, + "loss": 1.3836, + "theoretical_loss": 3.420122508071829, + "tokens_seen": 2296119296 + }, + { + "epoch": 0.7, + "learning_rate": 0.00024598595787362086, + "loss": 1.4148, + "theoretical_loss": 3.4201146542356557, + "tokens_seen": 2296184832 + }, + { + "epoch": 0.7, + "learning_rate": 0.00024596990972918756, + "loss": 1.3822, + "theoretical_loss": 3.4201068006864, + "tokens_seen": 2296250368 + }, + { + "epoch": 0.7, + "learning_rate": 0.00024595386158475426, + "loss": 1.1797, + "theoretical_loss": 3.420098947424043, + "tokens_seen": 2296315904 + }, + { + "epoch": 0.7, + "learning_rate": 0.00024593781344032096, + "loss": 1.3057, + "theoretical_loss": 3.4200910944485665, + "tokens_seen": 2296381440 + }, + { + "epoch": 0.7, + "learning_rate": 0.00024592176529588766, + "loss": 1.3338, + "theoretical_loss": 3.4200832417599507, + "tokens_seen": 2296446976 + }, + { + "epoch": 0.7, + "learning_rate": 0.00024590571715145436, + "loss": 1.3056, + "theoretical_loss": 3.4200753893581783, + "tokens_seen": 2296512512 + }, + { + "epoch": 0.7, + "learning_rate": 0.00024588966900702106, + "loss": 1.3831, + "theoretical_loss": 3.42006753724323, + "tokens_seen": 2296578048 + }, + { + "epoch": 0.7, + "learning_rate": 0.00024587362086258776, + "loss": 1.3956, + "theoretical_loss": 3.4200596854150866, + "tokens_seen": 2296643584 + }, + { + "epoch": 0.7, + "learning_rate": 0.00024585757271815446, + "loss": 1.2806, + "theoretical_loss": 3.4200518338737305, + "tokens_seen": 2296709120 + }, + { + "epoch": 0.7, + "learning_rate": 0.00024584152457372116, + "loss": 1.2392, + "theoretical_loss": 3.420043982619142, + "tokens_seen": 2296774656 + }, + { + "epoch": 0.7, + "learning_rate": 0.00024582547642928787, + "loss": 1.358, + "theoretical_loss": 3.4200361316513037, + "tokens_seen": 2296840192 + }, + { + "epoch": 0.7, + "learning_rate": 0.00024580942828485457, + "loss": 1.3815, + "theoretical_loss": 3.420028280970196, + "tokens_seen": 2296905728 + }, + { + "epoch": 0.7, + "learning_rate": 0.00024579338014042127, + "loss": 1.2617, + "theoretical_loss": 3.4200204305758004, + "tokens_seen": 2296971264 + }, + { + "epoch": 0.7, + "objective/train/docs_used": 836410, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.2487609386444092, + "objective/train/theoretical_loss": 3.4200125804680983, + "objective/train/tokens_used": 2317496800, + "theoretical_loss": 3.4200125804680983, + "tokens_seen": 2297036800 + }, + { + "epoch": 0.7, + "learning_rate": 0.00024577733199598797, + "loss": 1.2748, + "theoretical_loss": 3.4200125804680983, + "tokens_seen": 2297036800 + }, + { + "epoch": 0.7, + "learning_rate": 0.00024576128385155467, + "loss": 1.3065, + "theoretical_loss": 3.420004730647071, + "tokens_seen": 2297102336 + }, + { + "epoch": 0.7, + "learning_rate": 0.00024574523570712137, + "loss": 1.314, + "theoretical_loss": 3.4199968811127004, + "tokens_seen": 2297167872 + }, + { + "epoch": 0.7, + "learning_rate": 0.00024572918756268807, + "loss": 1.2682, + "theoretical_loss": 3.419989031864967, + "tokens_seen": 2297233408 + }, + { + "epoch": 0.7, + "learning_rate": 0.00024571313941825477, + "loss": 1.3134, + "theoretical_loss": 3.419981182903852, + "tokens_seen": 2297298944 + }, + { + "epoch": 0.7, + "learning_rate": 0.00024569709127382147, + "loss": 1.3196, + "theoretical_loss": 3.419973334229338, + "tokens_seen": 2297364480 + }, + { + "epoch": 0.7, + "learning_rate": 0.00024568104312938817, + "loss": 1.347, + "theoretical_loss": 3.4199654858414057, + "tokens_seen": 2297430016 + }, + { + "epoch": 0.7, + "learning_rate": 0.00024566499498495487, + "loss": 1.243, + "theoretical_loss": 3.419957637740036, + "tokens_seen": 2297495552 + }, + { + "epoch": 0.7, + "learning_rate": 0.00024564894684052157, + "loss": 1.3362, + "theoretical_loss": 3.419949789925211, + "tokens_seen": 2297561088 + }, + { + "epoch": 0.7, + "learning_rate": 0.0002456328986960883, + "loss": 1.2343, + "theoretical_loss": 3.4199419423969117, + "tokens_seen": 2297626624 + }, + { + "epoch": 0.7, + "learning_rate": 0.000245616850551655, + "loss": 1.278, + "theoretical_loss": 3.4199340951551194, + "tokens_seen": 2297692160 + }, + { + "epoch": 0.7, + "learning_rate": 0.0002456008024072217, + "loss": 1.241, + "theoretical_loss": 3.4199262481998156, + "tokens_seen": 2297757696 + }, + { + "epoch": 0.7, + "learning_rate": 0.0002455847542627884, + "loss": 1.313, + "theoretical_loss": 3.419918401530982, + "tokens_seen": 2297823232 + }, + { + "epoch": 0.7, + "learning_rate": 0.0002455687061183551, + "loss": 1.3288, + "theoretical_loss": 3.419910555148599, + "tokens_seen": 2297888768 + }, + { + "epoch": 0.7, + "learning_rate": 0.0002455526579739218, + "loss": 1.2945, + "theoretical_loss": 3.419902709052649, + "tokens_seen": 2297954304 + }, + { + "epoch": 0.7, + "learning_rate": 0.0002455366098294885, + "loss": 1.2501, + "theoretical_loss": 3.4198948632431128, + "tokens_seen": 2298019840 + }, + { + "epoch": 0.7, + "learning_rate": 0.0002455205616850552, + "loss": 1.373, + "theoretical_loss": 3.419887017719972, + "tokens_seen": 2298085376 + }, + { + "epoch": 0.7, + "learning_rate": 0.0002455045135406219, + "loss": 1.323, + "theoretical_loss": 3.419879172483208, + "tokens_seen": 2298150912 + }, + { + "epoch": 0.7, + "learning_rate": 0.0002454884653961886, + "loss": 1.2593, + "theoretical_loss": 3.419871327532802, + "tokens_seen": 2298216448 + }, + { + "epoch": 0.7, + "learning_rate": 0.0002454724172517553, + "loss": 1.1857, + "theoretical_loss": 3.4198634828687355, + "tokens_seen": 2298281984 + }, + { + "epoch": 0.7, + "learning_rate": 0.000245456369107322, + "loss": 1.3889, + "theoretical_loss": 3.41985563849099, + "tokens_seen": 2298347520 + }, + { + "epoch": 0.7, + "learning_rate": 0.0002454403209628887, + "loss": 1.3074, + "theoretical_loss": 3.4198477943995464, + "tokens_seen": 2298413056 + }, + { + "epoch": 0.7, + "learning_rate": 0.0002454242728184554, + "loss": 1.2779, + "theoretical_loss": 3.419839950594387, + "tokens_seen": 2298478592 + }, + { + "epoch": 0.7, + "learning_rate": 0.0002454082246740221, + "loss": 1.2841, + "theoretical_loss": 3.419832107075492, + "tokens_seen": 2298544128 + }, + { + "epoch": 0.7, + "learning_rate": 0.0002453921765295888, + "loss": 1.2956, + "theoretical_loss": 3.419824263842844, + "tokens_seen": 2298609664 + }, + { + "epoch": 0.7, + "learning_rate": 0.0002453761283851555, + "loss": 1.2129, + "theoretical_loss": 3.4198164208964235, + "tokens_seen": 2298675200 + }, + { + "epoch": 0.7, + "learning_rate": 0.0002453600802407222, + "loss": 1.1565, + "theoretical_loss": 3.4198085782362124, + "tokens_seen": 2298740736 + }, + { + "epoch": 0.7, + "learning_rate": 0.0002453440320962889, + "loss": 1.4285, + "theoretical_loss": 3.419800735862192, + "tokens_seen": 2298806272 + }, + { + "epoch": 0.7, + "learning_rate": 0.0002453279839518556, + "loss": 1.2801, + "theoretical_loss": 3.4197928937743436, + "tokens_seen": 2298871808 + }, + { + "epoch": 0.7, + "learning_rate": 0.0002453119358074223, + "loss": 1.214, + "theoretical_loss": 3.4197850519726485, + "tokens_seen": 2298937344 + }, + { + "epoch": 0.7, + "learning_rate": 0.000245295887662989, + "loss": 1.3541, + "theoretical_loss": 3.419777210457088, + "tokens_seen": 2299002880 + }, + { + "epoch": 0.7, + "learning_rate": 0.0002452798395185557, + "loss": 1.2681, + "theoretical_loss": 3.4197693692276436, + "tokens_seen": 2299068416 + }, + { + "epoch": 0.7, + "learning_rate": 0.0002452637913741224, + "loss": 1.2799, + "theoretical_loss": 3.4197615282842975, + "tokens_seen": 2299133952 + }, + { + "epoch": 0.7, + "learning_rate": 0.0002452477432296891, + "loss": 1.3173, + "theoretical_loss": 3.41975368762703, + "tokens_seen": 2299199488 + }, + { + "epoch": 0.7, + "learning_rate": 0.0002452316950852558, + "loss": 1.302, + "theoretical_loss": 3.4197458472558226, + "tokens_seen": 2299265024 + }, + { + "epoch": 0.7, + "learning_rate": 0.0002452156469408225, + "loss": 1.3201, + "theoretical_loss": 3.4197380071706576, + "tokens_seen": 2299330560 + }, + { + "epoch": 0.7, + "learning_rate": 0.0002451995987963892, + "loss": 1.3248, + "theoretical_loss": 3.4197301673715157, + "tokens_seen": 2299396096 + }, + { + "epoch": 0.7, + "learning_rate": 0.0002451835506519559, + "loss": 1.363, + "theoretical_loss": 3.419722327858379, + "tokens_seen": 2299461632 + }, + { + "epoch": 0.7, + "learning_rate": 0.0002451675025075226, + "loss": 1.1929, + "theoretical_loss": 3.4197144886312274, + "tokens_seen": 2299527168 + }, + { + "epoch": 0.7, + "learning_rate": 0.0002451514543630893, + "loss": 1.3281, + "theoretical_loss": 3.4197066496900437, + "tokens_seen": 2299592704 + }, + { + "epoch": 0.7, + "learning_rate": 0.000245135406218656, + "loss": 1.2938, + "theoretical_loss": 3.419698811034809, + "tokens_seen": 2299658240 + }, + { + "epoch": 0.7, + "learning_rate": 0.0002451193580742227, + "loss": 1.3939, + "theoretical_loss": 3.4196909726655047, + "tokens_seen": 2299723776 + }, + { + "epoch": 0.7, + "learning_rate": 0.0002451033099297894, + "loss": 1.3771, + "theoretical_loss": 3.4196831345821117, + "tokens_seen": 2299789312 + }, + { + "epoch": 0.7, + "learning_rate": 0.0002450872617853561, + "loss": 1.3236, + "theoretical_loss": 3.4196752967846127, + "tokens_seen": 2299854848 + }, + { + "epoch": 0.7, + "learning_rate": 0.0002450712136409228, + "loss": 1.2943, + "theoretical_loss": 3.4196674592729877, + "tokens_seen": 2299920384 + }, + { + "epoch": 0.7, + "learning_rate": 0.0002450551654964895, + "loss": 1.42, + "theoretical_loss": 3.419659622047219, + "tokens_seen": 2299985920 + }, + { + "epoch": 0.7, + "learning_rate": 0.0002450391173520562, + "loss": 1.3358, + "theoretical_loss": 3.4196517851072876, + "tokens_seen": 2300051456 + }, + { + "epoch": 0.7, + "learning_rate": 0.0002450230692076229, + "loss": 1.3811, + "theoretical_loss": 3.4196439484531753, + "tokens_seen": 2300116992 + }, + { + "epoch": 0.7, + "learning_rate": 0.0002450070210631896, + "loss": 1.3383, + "theoretical_loss": 3.419636112084863, + "tokens_seen": 2300182528 + }, + { + "epoch": 0.7, + "learning_rate": 0.0002449909729187563, + "loss": 1.2448, + "theoretical_loss": 3.4196282760023324, + "tokens_seen": 2300248064 + }, + { + "epoch": 0.7, + "objective/train/docs_used": 837515, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.3734774589538574, + "objective/train/theoretical_loss": 3.4196204402055654, + "objective/train/tokens_used": 2320773600, + "theoretical_loss": 3.4196204402055654, + "tokens_seen": 2300313600 + }, + { + "epoch": 0.7, + "learning_rate": 0.000244974924774323, + "loss": 1.3757, + "theoretical_loss": 3.4196204402055654, + "tokens_seen": 2300313600 + }, + { + "epoch": 0.7, + "learning_rate": 0.0002449588766298897, + "loss": 1.2402, + "theoretical_loss": 3.419612604694543, + "tokens_seen": 2300379136 + }, + { + "epoch": 0.7, + "learning_rate": 0.0002449428284854564, + "loss": 1.2445, + "theoretical_loss": 3.4196047694692466, + "tokens_seen": 2300444672 + }, + { + "epoch": 0.7, + "learning_rate": 0.0002449267803410231, + "loss": 1.3151, + "theoretical_loss": 3.419596934529658, + "tokens_seen": 2300510208 + }, + { + "epoch": 0.7, + "learning_rate": 0.0002449107321965898, + "loss": 1.3202, + "theoretical_loss": 3.419589099875758, + "tokens_seen": 2300575744 + }, + { + "epoch": 0.7, + "learning_rate": 0.0002448946840521565, + "loss": 1.3697, + "theoretical_loss": 3.4195812655075284, + "tokens_seen": 2300641280 + }, + { + "epoch": 0.7, + "learning_rate": 0.00024487863590772315, + "loss": 1.1543, + "theoretical_loss": 3.4195734314249506, + "tokens_seen": 2300706816 + }, + { + "epoch": 0.7, + "learning_rate": 0.00024486258776328985, + "loss": 1.3593, + "theoretical_loss": 3.4195655976280066, + "tokens_seen": 2300772352 + }, + { + "epoch": 0.7, + "learning_rate": 0.00024484653961885655, + "loss": 1.1741, + "theoretical_loss": 3.419557764116677, + "tokens_seen": 2300837888 + }, + { + "epoch": 0.7, + "learning_rate": 0.00024483049147442325, + "loss": 1.3673, + "theoretical_loss": 3.419549930890944, + "tokens_seen": 2300903424 + }, + { + "epoch": 0.7, + "learning_rate": 0.00024481444332998995, + "loss": 1.3151, + "theoretical_loss": 3.419542097950788, + "tokens_seen": 2300968960 + }, + { + "epoch": 0.7, + "learning_rate": 0.00024479839518555665, + "loss": 1.3354, + "theoretical_loss": 3.4195342652961918, + "tokens_seen": 2301034496 + }, + { + "epoch": 0.7, + "learning_rate": 0.00024478234704112335, + "loss": 1.2743, + "theoretical_loss": 3.4195264329271353, + "tokens_seen": 2301100032 + }, + { + "epoch": 0.7, + "learning_rate": 0.00024476629889669005, + "loss": 1.2859, + "theoretical_loss": 3.4195186008436016, + "tokens_seen": 2301165568 + }, + { + "epoch": 0.7, + "learning_rate": 0.00024475025075225675, + "loss": 1.2864, + "theoretical_loss": 3.419510769045571, + "tokens_seen": 2301231104 + }, + { + "epoch": 0.7, + "learning_rate": 0.00024473420260782345, + "loss": 1.2093, + "theoretical_loss": 3.4195029375330255, + "tokens_seen": 2301296640 + }, + { + "epoch": 0.7, + "learning_rate": 0.00024471815446339016, + "loss": 1.2735, + "theoretical_loss": 3.4194951063059467, + "tokens_seen": 2301362176 + }, + { + "epoch": 0.7, + "learning_rate": 0.00024470210631895686, + "loss": 1.3341, + "theoretical_loss": 3.419487275364315, + "tokens_seen": 2301427712 + }, + { + "epoch": 0.7, + "learning_rate": 0.00024468605817452356, + "loss": 1.2516, + "theoretical_loss": 3.4194794447081134, + "tokens_seen": 2301493248 + }, + { + "epoch": 0.7, + "learning_rate": 0.00024467001003009026, + "loss": 1.281, + "theoretical_loss": 3.4194716143373225, + "tokens_seen": 2301558784 + }, + { + "epoch": 0.7, + "learning_rate": 0.00024465396188565696, + "loss": 1.3177, + "theoretical_loss": 3.4194637842519238, + "tokens_seen": 2301624320 + }, + { + "epoch": 0.7, + "learning_rate": 0.00024463791374122366, + "loss": 1.369, + "theoretical_loss": 3.4194559544518985, + "tokens_seen": 2301689856 + }, + { + "epoch": 0.7, + "learning_rate": 0.00024462186559679036, + "loss": 1.3485, + "theoretical_loss": 3.419448124937229, + "tokens_seen": 2301755392 + }, + { + "epoch": 0.7, + "learning_rate": 0.00024460581745235706, + "loss": 1.2935, + "theoretical_loss": 3.419440295707896, + "tokens_seen": 2301820928 + }, + { + "epoch": 0.7, + "learning_rate": 0.00024458976930792376, + "loss": 1.3834, + "theoretical_loss": 3.4194324667638814, + "tokens_seen": 2301886464 + }, + { + "epoch": 0.7, + "learning_rate": 0.00024457372116349046, + "loss": 1.34, + "theoretical_loss": 3.4194246381051663, + "tokens_seen": 2301952000 + }, + { + "epoch": 0.7, + "learning_rate": 0.00024455767301905716, + "loss": 1.2741, + "theoretical_loss": 3.419416809731732, + "tokens_seen": 2302017536 + }, + { + "epoch": 0.7, + "learning_rate": 0.00024454162487462386, + "loss": 1.505, + "theoretical_loss": 3.419408981643561, + "tokens_seen": 2302083072 + }, + { + "epoch": 0.7, + "learning_rate": 0.00024452557673019056, + "loss": 1.3639, + "theoretical_loss": 3.4194011538406333, + "tokens_seen": 2302148608 + }, + { + "epoch": 0.7, + "learning_rate": 0.00024450952858575726, + "loss": 1.2386, + "theoretical_loss": 3.4193933263229317, + "tokens_seen": 2302214144 + }, + { + "epoch": 0.7, + "learning_rate": 0.00024449348044132396, + "loss": 1.3566, + "theoretical_loss": 3.4193854990904375, + "tokens_seen": 2302279680 + }, + { + "epoch": 0.7, + "learning_rate": 0.00024447743229689066, + "loss": 1.2433, + "theoretical_loss": 3.4193776721431313, + "tokens_seen": 2302345216 + }, + { + "epoch": 0.7, + "learning_rate": 0.00024446138415245737, + "loss": 1.2594, + "theoretical_loss": 3.4193698454809955, + "tokens_seen": 2302410752 + }, + { + "epoch": 0.7, + "learning_rate": 0.00024444533600802407, + "loss": 1.3052, + "theoretical_loss": 3.4193620191040113, + "tokens_seen": 2302476288 + }, + { + "epoch": 0.7, + "learning_rate": 0.0002444292878635908, + "loss": 1.2564, + "theoretical_loss": 3.4193541930121603, + "tokens_seen": 2302541824 + }, + { + "epoch": 0.7, + "learning_rate": 0.0002444132397191575, + "loss": 1.4136, + "theoretical_loss": 3.4193463672054234, + "tokens_seen": 2302607360 + }, + { + "epoch": 0.7, + "learning_rate": 0.0002443971915747242, + "loss": 1.2979, + "theoretical_loss": 3.419338541683783, + "tokens_seen": 2302672896 + }, + { + "epoch": 0.7, + "learning_rate": 0.0002443811434302909, + "loss": 1.3176, + "theoretical_loss": 3.41933071644722, + "tokens_seen": 2302738432 + }, + { + "epoch": 0.7, + "learning_rate": 0.0002443650952858576, + "loss": 1.3765, + "theoretical_loss": 3.4193228914957157, + "tokens_seen": 2302803968 + }, + { + "epoch": 0.7, + "learning_rate": 0.0002443490471414243, + "loss": 1.2851, + "theoretical_loss": 3.4193150668292525, + "tokens_seen": 2302869504 + }, + { + "epoch": 0.7, + "learning_rate": 0.000244332998996991, + "loss": 1.2615, + "theoretical_loss": 3.419307242447811, + "tokens_seen": 2302935040 + }, + { + "epoch": 0.7, + "learning_rate": 0.0002443169508525577, + "loss": 1.2962, + "theoretical_loss": 3.4192994183513736, + "tokens_seen": 2303000576 + }, + { + "epoch": 0.7, + "learning_rate": 0.0002443009027081244, + "loss": 1.3862, + "theoretical_loss": 3.4192915945399207, + "tokens_seen": 2303066112 + }, + { + "epoch": 0.7, + "learning_rate": 0.0002442848545636911, + "loss": 1.3626, + "theoretical_loss": 3.4192837710134345, + "tokens_seen": 2303131648 + }, + { + "epoch": 0.7, + "learning_rate": 0.0002442688064192578, + "loss": 1.3709, + "theoretical_loss": 3.419275947771897, + "tokens_seen": 2303197184 + }, + { + "epoch": 0.7, + "learning_rate": 0.0002442527582748245, + "loss": 1.2733, + "theoretical_loss": 3.4192681248152885, + "tokens_seen": 2303262720 + }, + { + "epoch": 0.7, + "learning_rate": 0.0002442367101303912, + "loss": 1.3811, + "theoretical_loss": 3.4192603021435914, + "tokens_seen": 2303328256 + }, + { + "epoch": 0.7, + "learning_rate": 0.0002442206619859579, + "loss": 1.3204, + "theoretical_loss": 3.419252479756787, + "tokens_seen": 2303393792 + }, + { + "epoch": 0.7, + "learning_rate": 0.0002442046138415246, + "loss": 1.3523, + "theoretical_loss": 3.4192446576548567, + "tokens_seen": 2303459328 + }, + { + "epoch": 0.7, + "learning_rate": 0.0002441885656970913, + "loss": 1.3478, + "theoretical_loss": 3.419236835837782, + "tokens_seen": 2303524864 + }, + { + "epoch": 0.7, + "objective/train/docs_used": 838679, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.2844657897949219, + "objective/train/theoretical_loss": 3.419229014305545, + "objective/train/tokens_used": 2324050400, + "theoretical_loss": 3.419229014305545, + "tokens_seen": 2303590400 + }, + { + "epoch": 0.7, + "learning_rate": 0.000244172517552658, + "loss": 1.3221, + "theoretical_loss": 3.419229014305545, + "tokens_seen": 2303590400 + }, + { + "epoch": 0.7, + "learning_rate": 0.0002441564694082247, + "loss": 1.3594, + "theoretical_loss": 3.419221193058126, + "tokens_seen": 2303655936 + }, + { + "epoch": 0.7, + "learning_rate": 0.0002441404212637914, + "loss": 1.4465, + "theoretical_loss": 3.4192133720955074, + "tokens_seen": 2303721472 + }, + { + "epoch": 0.7, + "learning_rate": 0.0002441243731193581, + "loss": 1.3313, + "theoretical_loss": 3.4192055514176714, + "tokens_seen": 2303787008 + }, + { + "epoch": 0.7, + "learning_rate": 0.0002441083249749248, + "loss": 1.327, + "theoretical_loss": 3.419197731024598, + "tokens_seen": 2303852544 + }, + { + "epoch": 0.7, + "learning_rate": 0.0002440922768304915, + "loss": 1.3415, + "theoretical_loss": 3.4191899109162693, + "tokens_seen": 2303918080 + }, + { + "epoch": 0.7, + "learning_rate": 0.0002440762286860582, + "loss": 1.3323, + "theoretical_loss": 3.4191820910926674, + "tokens_seen": 2303983616 + }, + { + "epoch": 0.7, + "learning_rate": 0.00024406018054162488, + "loss": 1.3236, + "theoretical_loss": 3.419174271553773, + "tokens_seen": 2304049152 + }, + { + "epoch": 0.7, + "learning_rate": 0.00024404413239719158, + "loss": 1.3372, + "theoretical_loss": 3.419166452299569, + "tokens_seen": 2304114688 + }, + { + "epoch": 0.7, + "learning_rate": 0.00024402808425275828, + "loss": 1.3005, + "theoretical_loss": 3.4191586333300354, + "tokens_seen": 2304180224 + }, + { + "epoch": 0.7, + "learning_rate": 0.00024401203610832498, + "loss": 1.3595, + "theoretical_loss": 3.419150814645154, + "tokens_seen": 2304245760 + }, + { + "epoch": 0.7, + "learning_rate": 0.00024399598796389168, + "loss": 1.2345, + "theoretical_loss": 3.4191429962449074, + "tokens_seen": 2304311296 + }, + { + "epoch": 0.7, + "learning_rate": 0.00024397993981945838, + "loss": 1.2913, + "theoretical_loss": 3.419135178129276, + "tokens_seen": 2304376832 + }, + { + "epoch": 0.7, + "learning_rate": 0.00024396389167502508, + "loss": 1.2593, + "theoretical_loss": 3.419127360298242, + "tokens_seen": 2304442368 + }, + { + "epoch": 0.7, + "learning_rate": 0.00024394784353059179, + "loss": 1.3035, + "theoretical_loss": 3.4191195427517864, + "tokens_seen": 2304507904 + }, + { + "epoch": 0.7, + "learning_rate": 0.00024393179538615849, + "loss": 1.3765, + "theoretical_loss": 3.4191117254898913, + "tokens_seen": 2304573440 + }, + { + "epoch": 0.7, + "learning_rate": 0.0002439157472417252, + "loss": 1.3105, + "theoretical_loss": 3.419103908512538, + "tokens_seen": 2304638976 + }, + { + "epoch": 0.7, + "learning_rate": 0.0002438996990972919, + "loss": 1.3006, + "theoretical_loss": 3.419096091819708, + "tokens_seen": 2304704512 + }, + { + "epoch": 0.7, + "learning_rate": 0.0002438836509528586, + "loss": 1.3386, + "theoretical_loss": 3.419088275411383, + "tokens_seen": 2304770048 + }, + { + "epoch": 0.7, + "learning_rate": 0.0002438676028084253, + "loss": 1.3137, + "theoretical_loss": 3.4190804592875446, + "tokens_seen": 2304835584 + }, + { + "epoch": 0.7, + "learning_rate": 0.000243851554663992, + "loss": 1.3227, + "theoretical_loss": 3.4190726434481737, + "tokens_seen": 2304901120 + }, + { + "epoch": 0.7, + "learning_rate": 0.0002438355065195587, + "loss": 1.3112, + "theoretical_loss": 3.419064827893253, + "tokens_seen": 2304966656 + }, + { + "epoch": 0.7, + "learning_rate": 0.0002438194583751254, + "loss": 1.3056, + "theoretical_loss": 3.4190570126227633, + "tokens_seen": 2305032192 + }, + { + "epoch": 0.7, + "learning_rate": 0.0002438034102306921, + "loss": 1.2249, + "theoretical_loss": 3.4190491976366864, + "tokens_seen": 2305097728 + }, + { + "epoch": 0.7, + "learning_rate": 0.0002437873620862588, + "loss": 1.2921, + "theoretical_loss": 3.4190413829350037, + "tokens_seen": 2305163264 + }, + { + "epoch": 0.7, + "learning_rate": 0.0002437713139418255, + "loss": 1.4142, + "theoretical_loss": 3.4190335685176967, + "tokens_seen": 2305228800 + }, + { + "epoch": 0.7, + "learning_rate": 0.0002437552657973922, + "loss": 1.3679, + "theoretical_loss": 3.419025754384747, + "tokens_seen": 2305294336 + }, + { + "epoch": 0.7, + "learning_rate": 0.0002437392176529589, + "loss": 1.1785, + "theoretical_loss": 3.4190179405361367, + "tokens_seen": 2305359872 + }, + { + "epoch": 0.7, + "learning_rate": 0.0002437231695085256, + "loss": 1.355, + "theoretical_loss": 3.4190101269718465, + "tokens_seen": 2305425408 + }, + { + "epoch": 0.7, + "learning_rate": 0.0002437071213640923, + "loss": 1.2813, + "theoretical_loss": 3.419002313691859, + "tokens_seen": 2305490944 + }, + { + "epoch": 0.7, + "learning_rate": 0.000243691073219659, + "loss": 1.3011, + "theoretical_loss": 3.418994500696155, + "tokens_seen": 2305556480 + }, + { + "epoch": 0.7, + "learning_rate": 0.00024367502507522567, + "loss": 1.3834, + "theoretical_loss": 3.4189866879847157, + "tokens_seen": 2305622016 + }, + { + "epoch": 0.7, + "learning_rate": 0.00024365897693079237, + "loss": 1.2456, + "theoretical_loss": 3.4189788755575243, + "tokens_seen": 2305687552 + }, + { + "epoch": 0.7, + "learning_rate": 0.00024364292878635907, + "loss": 1.3266, + "theoretical_loss": 3.4189710634145607, + "tokens_seen": 2305753088 + }, + { + "epoch": 0.7, + "learning_rate": 0.00024362688064192577, + "loss": 1.1867, + "theoretical_loss": 3.418963251555807, + "tokens_seen": 2305818624 + }, + { + "epoch": 0.7, + "learning_rate": 0.00024361083249749247, + "loss": 1.2893, + "theoretical_loss": 3.418955439981245, + "tokens_seen": 2305884160 + }, + { + "epoch": 0.7, + "learning_rate": 0.00024359478435305917, + "loss": 1.3657, + "theoretical_loss": 3.4189476286908564, + "tokens_seen": 2305949696 + }, + { + "epoch": 0.7, + "learning_rate": 0.00024357873620862587, + "loss": 1.3392, + "theoretical_loss": 3.418939817684622, + "tokens_seen": 2306015232 + }, + { + "epoch": 0.7, + "learning_rate": 0.00024356268806419257, + "loss": 1.2963, + "theoretical_loss": 3.4189320069625246, + "tokens_seen": 2306080768 + }, + { + "epoch": 0.7, + "learning_rate": 0.00024354663991975927, + "loss": 1.3692, + "theoretical_loss": 3.418924196524545, + "tokens_seen": 2306146304 + }, + { + "epoch": 0.7, + "learning_rate": 0.00024353059177532597, + "loss": 1.2825, + "theoretical_loss": 3.418916386370665, + "tokens_seen": 2306211840 + }, + { + "epoch": 0.7, + "learning_rate": 0.00024351454363089268, + "loss": 1.3854, + "theoretical_loss": 3.418908576500866, + "tokens_seen": 2306277376 + }, + { + "epoch": 0.7, + "learning_rate": 0.00024349849548645938, + "loss": 1.1989, + "theoretical_loss": 3.41890076691513, + "tokens_seen": 2306342912 + }, + { + "epoch": 0.7, + "learning_rate": 0.00024348244734202608, + "loss": 1.2574, + "theoretical_loss": 3.418892957613438, + "tokens_seen": 2306408448 + }, + { + "epoch": 0.7, + "learning_rate": 0.00024346639919759278, + "loss": 1.3201, + "theoretical_loss": 3.418885148595772, + "tokens_seen": 2306473984 + }, + { + "epoch": 0.7, + "learning_rate": 0.00024345035105315948, + "loss": 1.2779, + "theoretical_loss": 3.4188773398621137, + "tokens_seen": 2306539520 + }, + { + "epoch": 0.7, + "learning_rate": 0.00024343430290872618, + "loss": 1.3178, + "theoretical_loss": 3.4188695314124447, + "tokens_seen": 2306605056 + }, + { + "epoch": 0.7, + "learning_rate": 0.00024341825476429288, + "loss": 1.3601, + "theoretical_loss": 3.4188617232467458, + "tokens_seen": 2306670592 + }, + { + "epoch": 0.7, + "learning_rate": 0.00024340220661985958, + "loss": 1.232, + "theoretical_loss": 3.4188539153649993, + "tokens_seen": 2306736128 + }, + { + "epoch": 0.7, + "learning_rate": 0.00024338615847542628, + "loss": 1.3058, + "theoretical_loss": 3.4188461077671874, + "tokens_seen": 2306801664 + }, + { + "epoch": 0.7, + "objective/train/docs_used": 839893, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.3934584856033325, + "objective/train/theoretical_loss": 3.4188383004532903, + "objective/train/tokens_used": 2327327200, + "theoretical_loss": 3.4188383004532903, + "tokens_seen": 2306867200 + }, + { + "epoch": 0.7, + "learning_rate": 0.00024337011033099298, + "loss": 1.377, + "theoretical_loss": 3.4188383004532903, + "tokens_seen": 2306867200 + }, + { + "epoch": 0.7, + "learning_rate": 0.00024335406218655968, + "loss": 1.2019, + "theoretical_loss": 3.418830493423291, + "tokens_seen": 2306932736 + }, + { + "epoch": 0.7, + "learning_rate": 0.00024333801404212638, + "loss": 1.3896, + "theoretical_loss": 3.41882268667717, + "tokens_seen": 2306998272 + }, + { + "epoch": 0.7, + "learning_rate": 0.00024332196589769308, + "loss": 1.2308, + "theoretical_loss": 3.4188148802149096, + "tokens_seen": 2307063808 + }, + { + "epoch": 0.7, + "learning_rate": 0.00024330591775325978, + "loss": 1.3751, + "theoretical_loss": 3.4188070740364913, + "tokens_seen": 2307129344 + }, + { + "epoch": 0.7, + "learning_rate": 0.00024328986960882646, + "loss": 1.2663, + "theoretical_loss": 3.418799268141896, + "tokens_seen": 2307194880 + }, + { + "epoch": 0.7, + "learning_rate": 0.00024327382146439316, + "loss": 1.3274, + "theoretical_loss": 3.418791462531107, + "tokens_seen": 2307260416 + }, + { + "epoch": 0.7, + "learning_rate": 0.00024325777331995986, + "loss": 1.3066, + "theoretical_loss": 3.418783657204104, + "tokens_seen": 2307325952 + }, + { + "epoch": 0.7, + "learning_rate": 0.0002432417251755266, + "loss": 1.323, + "theoretical_loss": 3.41877585216087, + "tokens_seen": 2307391488 + }, + { + "epoch": 0.7, + "learning_rate": 0.00024322567703109331, + "loss": 1.2862, + "theoretical_loss": 3.418768047401386, + "tokens_seen": 2307457024 + }, + { + "epoch": 0.7, + "learning_rate": 0.00024320962888666001, + "loss": 1.1943, + "theoretical_loss": 3.418760242925633, + "tokens_seen": 2307522560 + }, + { + "epoch": 0.7, + "learning_rate": 0.00024319358074222672, + "loss": 1.3129, + "theoretical_loss": 3.4187524387335944, + "tokens_seen": 2307588096 + }, + { + "epoch": 0.7, + "learning_rate": 0.00024317753259779342, + "loss": 1.4011, + "theoretical_loss": 3.4187446348252504, + "tokens_seen": 2307653632 + }, + { + "epoch": 0.7, + "learning_rate": 0.00024316148445336012, + "loss": 1.3235, + "theoretical_loss": 3.418736831200583, + "tokens_seen": 2307719168 + }, + { + "epoch": 0.7, + "learning_rate": 0.00024314543630892682, + "loss": 1.2222, + "theoretical_loss": 3.4187290278595737, + "tokens_seen": 2307784704 + }, + { + "epoch": 0.7, + "learning_rate": 0.00024312938816449352, + "loss": 1.357, + "theoretical_loss": 3.4187212248022045, + "tokens_seen": 2307850240 + }, + { + "epoch": 0.7, + "learning_rate": 0.00024311334002006022, + "loss": 1.3951, + "theoretical_loss": 3.4187134220284565, + "tokens_seen": 2307915776 + }, + { + "epoch": 0.7, + "learning_rate": 0.00024309729187562692, + "loss": 1.1887, + "theoretical_loss": 3.418705619538312, + "tokens_seen": 2307981312 + }, + { + "epoch": 0.7, + "learning_rate": 0.00024308124373119362, + "loss": 1.266, + "theoretical_loss": 3.4186978173317524, + "tokens_seen": 2308046848 + }, + { + "epoch": 0.7, + "learning_rate": 0.00024306519558676032, + "loss": 1.3898, + "theoretical_loss": 3.418690015408759, + "tokens_seen": 2308112384 + }, + { + "epoch": 0.7, + "learning_rate": 0.00024304914744232702, + "loss": 1.3287, + "theoretical_loss": 3.418682213769314, + "tokens_seen": 2308177920 + }, + { + "epoch": 0.7, + "learning_rate": 0.0002430330992978937, + "loss": 1.2552, + "theoretical_loss": 3.418674412413398, + "tokens_seen": 2308243456 + }, + { + "epoch": 0.7, + "learning_rate": 0.0002430170511534604, + "loss": 1.3, + "theoretical_loss": 3.418666611340994, + "tokens_seen": 2308308992 + }, + { + "epoch": 0.7, + "learning_rate": 0.0002430010030090271, + "loss": 1.3115, + "theoretical_loss": 3.418658810552083, + "tokens_seen": 2308374528 + }, + { + "epoch": 0.7, + "learning_rate": 0.0002429849548645938, + "loss": 1.3695, + "theoretical_loss": 3.4186510100466463, + "tokens_seen": 2308440064 + }, + { + "epoch": 0.7, + "learning_rate": 0.0002429689067201605, + "loss": 1.3673, + "theoretical_loss": 3.418643209824666, + "tokens_seen": 2308505600 + }, + { + "epoch": 0.7, + "learning_rate": 0.0002429528585757272, + "loss": 1.3238, + "theoretical_loss": 3.418635409886124, + "tokens_seen": 2308571136 + }, + { + "epoch": 0.7, + "learning_rate": 0.0002429368104312939, + "loss": 1.4846, + "theoretical_loss": 3.4186276102310016, + "tokens_seen": 2308636672 + }, + { + "epoch": 0.7, + "learning_rate": 0.0002429207622868606, + "loss": 1.4058, + "theoretical_loss": 3.41861981085928, + "tokens_seen": 2308702208 + }, + { + "epoch": 0.7, + "learning_rate": 0.0002429047141424273, + "loss": 1.2713, + "theoretical_loss": 3.418612011770942, + "tokens_seen": 2308767744 + }, + { + "epoch": 0.7, + "learning_rate": 0.000242888665997994, + "loss": 1.3379, + "theoretical_loss": 3.418604212965968, + "tokens_seen": 2308833280 + }, + { + "epoch": 0.7, + "learning_rate": 0.0002428726178535607, + "loss": 1.3667, + "theoretical_loss": 3.4185964144443406, + "tokens_seen": 2308898816 + }, + { + "epoch": 0.7, + "learning_rate": 0.0002428565697091274, + "loss": 1.4033, + "theoretical_loss": 3.418588616206041, + "tokens_seen": 2308964352 + }, + { + "epoch": 0.7, + "learning_rate": 0.0002428405215646941, + "loss": 1.2655, + "theoretical_loss": 3.4185808182510513, + "tokens_seen": 2309029888 + }, + { + "epoch": 0.7, + "learning_rate": 0.0002428244734202608, + "loss": 1.3194, + "theoretical_loss": 3.4185730205793528, + "tokens_seen": 2309095424 + }, + { + "epoch": 0.7, + "learning_rate": 0.0002428084252758275, + "loss": 1.3931, + "theoretical_loss": 3.418565223190927, + "tokens_seen": 2309160960 + }, + { + "epoch": 0.7, + "learning_rate": 0.0002427923771313942, + "loss": 1.3868, + "theoretical_loss": 3.418557426085756, + "tokens_seen": 2309226496 + }, + { + "epoch": 0.7, + "learning_rate": 0.0002427763289869609, + "loss": 1.329, + "theoretical_loss": 3.418549629263821, + "tokens_seen": 2309292032 + }, + { + "epoch": 0.7, + "learning_rate": 0.0002427602808425276, + "loss": 1.343, + "theoretical_loss": 3.4185418327251043, + "tokens_seen": 2309357568 + }, + { + "epoch": 0.7, + "learning_rate": 0.0002427442326980943, + "loss": 1.4184, + "theoretical_loss": 3.4185340364695866, + "tokens_seen": 2309423104 + }, + { + "epoch": 0.7, + "learning_rate": 0.000242728184553661, + "loss": 1.2934, + "theoretical_loss": 3.418526240497251, + "tokens_seen": 2309488640 + }, + { + "epoch": 0.7, + "learning_rate": 0.0002427121364092277, + "loss": 1.2915, + "theoretical_loss": 3.418518444808078, + "tokens_seen": 2309554176 + }, + { + "epoch": 0.7, + "learning_rate": 0.0002426960882647944, + "loss": 1.3187, + "theoretical_loss": 3.4185106494020494, + "tokens_seen": 2309619712 + }, + { + "epoch": 0.7, + "learning_rate": 0.0002426800401203611, + "loss": 1.2876, + "theoretical_loss": 3.4185028542791476, + "tokens_seen": 2309685248 + }, + { + "epoch": 0.7, + "learning_rate": 0.0002426639919759278, + "loss": 1.25, + "theoretical_loss": 3.4184950594393535, + "tokens_seen": 2309750784 + }, + { + "epoch": 0.7, + "learning_rate": 0.00024264794383149448, + "loss": 1.3372, + "theoretical_loss": 3.4184872648826494, + "tokens_seen": 2309816320 + }, + { + "epoch": 0.7, + "learning_rate": 0.00024263189568706118, + "loss": 1.2749, + "theoretical_loss": 3.4184794706090162, + "tokens_seen": 2309881856 + }, + { + "epoch": 0.7, + "learning_rate": 0.00024261584754262788, + "loss": 1.3065, + "theoretical_loss": 3.4184716766184366, + "tokens_seen": 2309947392 + }, + { + "epoch": 0.7, + "learning_rate": 0.00024259979939819458, + "loss": 1.1973, + "theoretical_loss": 3.4184638829108915, + "tokens_seen": 2310012928 + }, + { + "epoch": 0.7, + "learning_rate": 0.00024258375125376129, + "loss": 1.3038, + "theoretical_loss": 3.418456089486363, + "tokens_seen": 2310078464 + }, + { + "epoch": 0.7, + "objective/train/docs_used": 841155, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.2097991704940796, + "objective/train/theoretical_loss": 3.418448296344833, + "objective/train/tokens_used": 2330604000, + "theoretical_loss": 3.418448296344833, + "tokens_seen": 2310144000 + }, + { + "epoch": 0.7, + "learning_rate": 0.00024256770310932799, + "loss": 1.3335, + "theoretical_loss": 3.418448296344833, + "tokens_seen": 2310144000 + }, + { + "epoch": 0.7, + "learning_rate": 0.0002425516549648947, + "loss": 1.2201, + "theoretical_loss": 3.4184405034862824, + "tokens_seen": 2310209536 + }, + { + "epoch": 0.7, + "learning_rate": 0.0002425356068204614, + "loss": 1.321, + "theoretical_loss": 3.418432710910693, + "tokens_seen": 2310275072 + }, + { + "epoch": 0.7, + "learning_rate": 0.0002425195586760281, + "loss": 1.3607, + "theoretical_loss": 3.4184249186180478, + "tokens_seen": 2310340608 + }, + { + "epoch": 0.7, + "learning_rate": 0.0002425035105315948, + "loss": 1.2741, + "theoretical_loss": 3.418417126608327, + "tokens_seen": 2310406144 + }, + { + "epoch": 0.7, + "learning_rate": 0.0002424874623871615, + "loss": 1.3487, + "theoretical_loss": 3.4184093348815128, + "tokens_seen": 2310471680 + }, + { + "epoch": 0.7, + "learning_rate": 0.0002424714142427282, + "loss": 1.4024, + "theoretical_loss": 3.418401543437587, + "tokens_seen": 2310537216 + }, + { + "epoch": 0.7, + "learning_rate": 0.0002424553660982949, + "loss": 1.3571, + "theoretical_loss": 3.4183937522765317, + "tokens_seen": 2310602752 + }, + { + "epoch": 0.7, + "learning_rate": 0.0002424393179538616, + "loss": 1.2229, + "theoretical_loss": 3.418385961398328, + "tokens_seen": 2310668288 + }, + { + "epoch": 0.7, + "learning_rate": 0.0002424232698094283, + "loss": 1.2403, + "theoretical_loss": 3.4183781708029572, + "tokens_seen": 2310733824 + }, + { + "epoch": 0.7, + "learning_rate": 0.000242407221664995, + "loss": 1.3016, + "theoretical_loss": 3.418370380490402, + "tokens_seen": 2310799360 + }, + { + "epoch": 0.7, + "learning_rate": 0.0002423911735205617, + "loss": 1.3163, + "theoretical_loss": 3.418362590460644, + "tokens_seen": 2310864896 + }, + { + "epoch": 0.7, + "learning_rate": 0.0002423751253761284, + "loss": 1.2233, + "theoretical_loss": 3.418354800713664, + "tokens_seen": 2310930432 + }, + { + "epoch": 0.7, + "learning_rate": 0.0002423590772316951, + "loss": 1.1992, + "theoretical_loss": 3.4183470112494447, + "tokens_seen": 2310995968 + }, + { + "epoch": 0.7, + "learning_rate": 0.0002423430290872618, + "loss": 1.4349, + "theoretical_loss": 3.4183392220679676, + "tokens_seen": 2311061504 + }, + { + "epoch": 0.7, + "learning_rate": 0.0002423269809428285, + "loss": 1.3406, + "theoretical_loss": 3.418331433169214, + "tokens_seen": 2311127040 + }, + { + "epoch": 0.7, + "learning_rate": 0.0002423109327983952, + "loss": 1.4591, + "theoretical_loss": 3.4183236445531664, + "tokens_seen": 2311192576 + }, + { + "epoch": 0.7, + "learning_rate": 0.0002422948846539619, + "loss": 1.1568, + "theoretical_loss": 3.4183158562198055, + "tokens_seen": 2311258112 + }, + { + "epoch": 0.7, + "learning_rate": 0.0002422788365095286, + "loss": 1.3233, + "theoretical_loss": 3.4183080681691136, + "tokens_seen": 2311323648 + }, + { + "epoch": 0.7, + "learning_rate": 0.00024226278836509527, + "loss": 1.3989, + "theoretical_loss": 3.4183002804010725, + "tokens_seen": 2311389184 + }, + { + "epoch": 0.7, + "learning_rate": 0.00024224674022066197, + "loss": 1.3407, + "theoretical_loss": 3.418292492915664, + "tokens_seen": 2311454720 + }, + { + "epoch": 0.7, + "learning_rate": 0.00024223069207622867, + "loss": 1.3844, + "theoretical_loss": 3.4182847057128694, + "tokens_seen": 2311520256 + }, + { + "epoch": 0.7, + "learning_rate": 0.00024221464393179537, + "loss": 1.2837, + "theoretical_loss": 3.418276918792671, + "tokens_seen": 2311585792 + }, + { + "epoch": 0.7, + "learning_rate": 0.00024219859578736207, + "loss": 1.4085, + "theoretical_loss": 3.4182691321550496, + "tokens_seen": 2311651328 + }, + { + "epoch": 0.7, + "learning_rate": 0.00024218254764292877, + "loss": 1.3707, + "theoretical_loss": 3.418261345799988, + "tokens_seen": 2311716864 + }, + { + "epoch": 0.7, + "learning_rate": 0.00024216649949849547, + "loss": 1.3868, + "theoretical_loss": 3.418253559727467, + "tokens_seen": 2311782400 + }, + { + "epoch": 0.7, + "learning_rate": 0.00024215045135406218, + "loss": 1.3248, + "theoretical_loss": 3.418245773937469, + "tokens_seen": 2311847936 + }, + { + "epoch": 0.7, + "learning_rate": 0.00024213440320962888, + "loss": 1.291, + "theoretical_loss": 3.4182379884299756, + "tokens_seen": 2311913472 + }, + { + "epoch": 0.7, + "learning_rate": 0.00024211835506519558, + "loss": 1.3015, + "theoretical_loss": 3.4182302032049687, + "tokens_seen": 2311979008 + }, + { + "epoch": 0.7, + "learning_rate": 0.00024210230692076228, + "loss": 1.3242, + "theoretical_loss": 3.4182224182624297, + "tokens_seen": 2312044544 + }, + { + "epoch": 0.7, + "learning_rate": 0.00024208625877632898, + "loss": 1.2399, + "theoretical_loss": 3.4182146336023402, + "tokens_seen": 2312110080 + }, + { + "epoch": 0.7, + "learning_rate": 0.00024207021063189568, + "loss": 1.3947, + "theoretical_loss": 3.4182068492246827, + "tokens_seen": 2312175616 + }, + { + "epoch": 0.7, + "learning_rate": 0.00024205416248746238, + "loss": 1.2738, + "theoretical_loss": 3.4181990651294383, + "tokens_seen": 2312241152 + }, + { + "epoch": 0.7, + "learning_rate": 0.00024203811434302913, + "loss": 1.2244, + "theoretical_loss": 3.418191281316589, + "tokens_seen": 2312306688 + }, + { + "epoch": 0.7, + "learning_rate": 0.0002420220661985958, + "loss": 1.3242, + "theoretical_loss": 3.4181834977861163, + "tokens_seen": 2312372224 + }, + { + "epoch": 0.7, + "learning_rate": 0.0002420060180541625, + "loss": 1.3162, + "theoretical_loss": 3.418175714538002, + "tokens_seen": 2312437760 + }, + { + "epoch": 0.7, + "learning_rate": 0.0002419899699097292, + "loss": 1.3338, + "theoretical_loss": 3.418167931572228, + "tokens_seen": 2312503296 + }, + { + "epoch": 0.7, + "learning_rate": 0.0002419739217652959, + "loss": 1.3602, + "theoretical_loss": 3.4181601488887763, + "tokens_seen": 2312568832 + }, + { + "epoch": 0.7, + "learning_rate": 0.0002419578736208626, + "loss": 1.266, + "theoretical_loss": 3.4181523664876283, + "tokens_seen": 2312634368 + }, + { + "epoch": 0.7, + "learning_rate": 0.0002419418254764293, + "loss": 1.2958, + "theoretical_loss": 3.4181445843687657, + "tokens_seen": 2312699904 + }, + { + "epoch": 0.7, + "learning_rate": 0.000241925777331996, + "loss": 1.3192, + "theoretical_loss": 3.4181368025321706, + "tokens_seen": 2312765440 + }, + { + "epoch": 0.7, + "learning_rate": 0.0002419097291875627, + "loss": 1.2861, + "theoretical_loss": 3.418129020977825, + "tokens_seen": 2312830976 + }, + { + "epoch": 0.7, + "learning_rate": 0.0002418936810431294, + "loss": 1.3635, + "theoretical_loss": 3.4181212397057097, + "tokens_seen": 2312896512 + }, + { + "epoch": 0.7, + "learning_rate": 0.0002418776328986961, + "loss": 1.2584, + "theoretical_loss": 3.4181134587158066, + "tokens_seen": 2312962048 + }, + { + "epoch": 0.7, + "learning_rate": 0.00024186158475426281, + "loss": 1.2855, + "theoretical_loss": 3.4181056780080987, + "tokens_seen": 2313027584 + }, + { + "epoch": 0.7, + "learning_rate": 0.00024184553660982951, + "loss": 1.3466, + "theoretical_loss": 3.418097897582567, + "tokens_seen": 2313093120 + }, + { + "epoch": 0.7, + "learning_rate": 0.00024182948846539622, + "loss": 1.3726, + "theoretical_loss": 3.418090117439193, + "tokens_seen": 2313158656 + }, + { + "epoch": 0.7, + "learning_rate": 0.00024181344032096292, + "loss": 1.3188, + "theoretical_loss": 3.418082337577958, + "tokens_seen": 2313224192 + }, + { + "epoch": 0.7, + "learning_rate": 0.00024179739217652962, + "loss": 1.3566, + "theoretical_loss": 3.418074557998845, + "tokens_seen": 2313289728 + }, + { + "epoch": 0.7, + "learning_rate": 0.00024178134403209632, + "loss": 1.368, + "theoretical_loss": 3.4180667787018355, + "tokens_seen": 2313355264 + }, + { + "epoch": 0.7, + "objective/train/docs_used": 842373, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.235954761505127, + "objective/train/theoretical_loss": 3.418058999686911, + "objective/train/tokens_used": 2333880800, + "theoretical_loss": 3.418058999686911, + "tokens_seen": 2313420800 + }, + { + "epoch": 0.7, + "learning_rate": 0.00024176529588766302, + "loss": 1.3014, + "theoretical_loss": 3.418058999686911, + "tokens_seen": 2313420800 + }, + { + "epoch": 0.7, + "learning_rate": 0.00024174924774322972, + "loss": 1.3512, + "theoretical_loss": 3.418051220954053, + "tokens_seen": 2313486336 + }, + { + "epoch": 0.7, + "learning_rate": 0.00024173319959879642, + "loss": 1.324, + "theoretical_loss": 3.4180434425032438, + "tokens_seen": 2313551872 + }, + { + "epoch": 0.7, + "learning_rate": 0.00024171715145436312, + "loss": 1.3106, + "theoretical_loss": 3.418035664334465, + "tokens_seen": 2313617408 + }, + { + "epoch": 0.7, + "learning_rate": 0.00024170110330992982, + "loss": 1.3389, + "theoretical_loss": 3.4180278864476983, + "tokens_seen": 2313682944 + }, + { + "epoch": 0.7, + "learning_rate": 0.00024168505516549652, + "loss": 1.3221, + "theoretical_loss": 3.4180201088429256, + "tokens_seen": 2313748480 + }, + { + "epoch": 0.7, + "learning_rate": 0.00024166900702106322, + "loss": 1.4107, + "theoretical_loss": 3.4180123315201287, + "tokens_seen": 2313814016 + }, + { + "epoch": 0.7, + "learning_rate": 0.00024165295887662992, + "loss": 1.384, + "theoretical_loss": 3.418004554479289, + "tokens_seen": 2313879552 + }, + { + "epoch": 0.7, + "learning_rate": 0.0002416369107321966, + "loss": 1.3344, + "theoretical_loss": 3.417996777720389, + "tokens_seen": 2313945088 + }, + { + "epoch": 0.7, + "learning_rate": 0.0002416208625877633, + "loss": 1.2142, + "theoretical_loss": 3.41798900124341, + "tokens_seen": 2314010624 + }, + { + "epoch": 0.7, + "learning_rate": 0.00024160481444333, + "loss": 1.2743, + "theoretical_loss": 3.417981225048334, + "tokens_seen": 2314076160 + }, + { + "epoch": 0.7, + "learning_rate": 0.0002415887662988967, + "loss": 1.3916, + "theoretical_loss": 3.417973449135143, + "tokens_seen": 2314141696 + }, + { + "epoch": 0.7, + "learning_rate": 0.0002415727181544634, + "loss": 1.3313, + "theoretical_loss": 3.4179656735038177, + "tokens_seen": 2314207232 + }, + { + "epoch": 0.7, + "learning_rate": 0.0002415566700100301, + "loss": 1.3259, + "theoretical_loss": 3.4179578981543415, + "tokens_seen": 2314272768 + }, + { + "epoch": 0.7, + "learning_rate": 0.0002415406218655968, + "loss": 1.3396, + "theoretical_loss": 3.417950123086695, + "tokens_seen": 2314338304 + }, + { + "epoch": 0.7, + "learning_rate": 0.0002415245737211635, + "loss": 1.2732, + "theoretical_loss": 3.41794234830086, + "tokens_seen": 2314403840 + }, + { + "epoch": 0.7, + "learning_rate": 0.0002415085255767302, + "loss": 1.3222, + "theoretical_loss": 3.4179345737968196, + "tokens_seen": 2314469376 + }, + { + "epoch": 0.7, + "learning_rate": 0.0002414924774322969, + "loss": 1.4172, + "theoretical_loss": 3.4179267995745546, + "tokens_seen": 2314534912 + }, + { + "epoch": 0.7, + "learning_rate": 0.0002414764292878636, + "loss": 1.2835, + "theoretical_loss": 3.4179190256340464, + "tokens_seen": 2314600448 + }, + { + "epoch": 0.7, + "learning_rate": 0.0002414603811434303, + "loss": 1.2967, + "theoretical_loss": 3.4179112519752777, + "tokens_seen": 2314665984 + }, + { + "epoch": 0.7, + "learning_rate": 0.000241444332998997, + "loss": 1.2254, + "theoretical_loss": 3.41790347859823, + "tokens_seen": 2314731520 + }, + { + "epoch": 0.7, + "learning_rate": 0.0002414282848545637, + "loss": 1.2779, + "theoretical_loss": 3.417895705502885, + "tokens_seen": 2314797056 + }, + { + "epoch": 0.7, + "learning_rate": 0.0002414122367101304, + "loss": 1.3517, + "theoretical_loss": 3.417887932689225, + "tokens_seen": 2314862592 + }, + { + "epoch": 0.7, + "learning_rate": 0.0002413961885656971, + "loss": 1.2705, + "theoretical_loss": 3.417880160157231, + "tokens_seen": 2314928128 + }, + { + "epoch": 0.7, + "learning_rate": 0.0002413801404212638, + "loss": 1.1523, + "theoretical_loss": 3.417872387906885, + "tokens_seen": 2314993664 + }, + { + "epoch": 0.7, + "learning_rate": 0.0002413640922768305, + "loss": 1.3041, + "theoretical_loss": 3.4178646159381696, + "tokens_seen": 2315059200 + }, + { + "epoch": 0.7, + "learning_rate": 0.0002413480441323972, + "loss": 1.3717, + "theoretical_loss": 3.4178568442510655, + "tokens_seen": 2315124736 + }, + { + "epoch": 0.7, + "learning_rate": 0.0002413319959879639, + "loss": 1.4448, + "theoretical_loss": 3.4178490728455557, + "tokens_seen": 2315190272 + }, + { + "epoch": 0.7, + "learning_rate": 0.0002413159478435306, + "loss": 1.2305, + "theoretical_loss": 3.417841301721621, + "tokens_seen": 2315255808 + }, + { + "epoch": 0.7, + "learning_rate": 0.0002412998996990973, + "loss": 1.3884, + "theoretical_loss": 3.4178335308792436, + "tokens_seen": 2315321344 + }, + { + "epoch": 0.7, + "learning_rate": 0.000241283851554664, + "loss": 1.2029, + "theoretical_loss": 3.417825760318406, + "tokens_seen": 2315386880 + }, + { + "epoch": 0.7, + "learning_rate": 0.0002412678034102307, + "loss": 1.2547, + "theoretical_loss": 3.417817990039089, + "tokens_seen": 2315452416 + }, + { + "epoch": 0.7, + "learning_rate": 0.00024125175526579738, + "loss": 1.3183, + "theoretical_loss": 3.4178102200412748, + "tokens_seen": 2315517952 + }, + { + "epoch": 0.7, + "learning_rate": 0.00024123570712136408, + "loss": 1.3083, + "theoretical_loss": 3.4178024503249453, + "tokens_seen": 2315583488 + }, + { + "epoch": 0.7, + "learning_rate": 0.00024121965897693078, + "loss": 1.3576, + "theoretical_loss": 3.4177946808900828, + "tokens_seen": 2315649024 + }, + { + "epoch": 0.7, + "learning_rate": 0.00024120361083249749, + "loss": 1.2594, + "theoretical_loss": 3.417786911736668, + "tokens_seen": 2315714560 + }, + { + "epoch": 0.7, + "learning_rate": 0.00024118756268806419, + "loss": 1.2555, + "theoretical_loss": 3.4177791428646835, + "tokens_seen": 2315780096 + }, + { + "epoch": 0.7, + "learning_rate": 0.0002411715145436309, + "loss": 1.3017, + "theoretical_loss": 3.4177713742741114, + "tokens_seen": 2315845632 + }, + { + "epoch": 0.7, + "learning_rate": 0.0002411554663991976, + "loss": 1.3643, + "theoretical_loss": 3.417763605964933, + "tokens_seen": 2315911168 + }, + { + "epoch": 0.7, + "learning_rate": 0.0002411394182547643, + "loss": 1.3849, + "theoretical_loss": 3.4177558379371304, + "tokens_seen": 2315976704 + }, + { + "epoch": 0.7, + "learning_rate": 0.000241123370110331, + "loss": 1.3013, + "theoretical_loss": 3.4177480701906853, + "tokens_seen": 2316042240 + }, + { + "epoch": 0.7, + "learning_rate": 0.0002411073219658977, + "loss": 1.3462, + "theoretical_loss": 3.4177403027255795, + "tokens_seen": 2316107776 + }, + { + "epoch": 0.7, + "learning_rate": 0.0002410912738214644, + "loss": 1.3358, + "theoretical_loss": 3.4177325355417953, + "tokens_seen": 2316173312 + }, + { + "epoch": 0.7, + "learning_rate": 0.0002410752256770311, + "loss": 1.327, + "theoretical_loss": 3.4177247686393137, + "tokens_seen": 2316238848 + }, + { + "epoch": 0.7, + "learning_rate": 0.0002410591775325978, + "loss": 1.2821, + "theoretical_loss": 3.417717002018118, + "tokens_seen": 2316304384 + }, + { + "epoch": 0.7, + "learning_rate": 0.0002410431293881645, + "loss": 1.3579, + "theoretical_loss": 3.4177092356781884, + "tokens_seen": 2316369920 + }, + { + "epoch": 0.7, + "learning_rate": 0.0002410270812437312, + "loss": 1.2679, + "theoretical_loss": 3.417701469619508, + "tokens_seen": 2316435456 + }, + { + "epoch": 0.7, + "learning_rate": 0.0002410110330992979, + "loss": 1.3674, + "theoretical_loss": 3.4176937038420574, + "tokens_seen": 2316500992 + }, + { + "epoch": 0.7, + "learning_rate": 0.0002409949849548646, + "loss": 1.2929, + "theoretical_loss": 3.41768593834582, + "tokens_seen": 2316566528 + }, + { + "epoch": 0.7, + "learning_rate": 0.0002409789368104313, + "loss": 1.3434, + "theoretical_loss": 3.417678173130776, + "tokens_seen": 2316632064 + }, + { + "epoch": 0.7, + "objective/train/docs_used": 843608, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.420849084854126, + "objective/train/theoretical_loss": 3.417670408196909, + "objective/train/tokens_used": 2337157600, + "theoretical_loss": 3.417670408196909, + "tokens_seen": 2316697600 + }, + { + "epoch": 0.7, + "learning_rate": 0.000240962888665998, + "loss": 1.4279, + "theoretical_loss": 3.417670408196909, + "tokens_seen": 2316697600 + }, + { + "epoch": 0.7, + "learning_rate": 0.0002409468405215647, + "loss": 1.2356, + "theoretical_loss": 3.4176626435441992, + "tokens_seen": 2316763136 + }, + { + "epoch": 0.7, + "learning_rate": 0.0002409307923771314, + "loss": 1.4069, + "theoretical_loss": 3.41765487917263, + "tokens_seen": 2316828672 + }, + { + "epoch": 0.7, + "learning_rate": 0.0002409147442326981, + "loss": 1.3422, + "theoretical_loss": 3.4176471150821826, + "tokens_seen": 2316894208 + }, + { + "epoch": 0.7, + "learning_rate": 0.0002408986960882648, + "loss": 1.4706, + "theoretical_loss": 3.4176393512728382, + "tokens_seen": 2316959744 + }, + { + "epoch": 0.7, + "learning_rate": 0.0002408826479438315, + "loss": 1.3786, + "theoretical_loss": 3.4176315877445793, + "tokens_seen": 2317025280 + }, + { + "epoch": 0.7, + "learning_rate": 0.00024086659979939817, + "loss": 1.3597, + "theoretical_loss": 3.4176238244973884, + "tokens_seen": 2317090816 + }, + { + "epoch": 0.7, + "learning_rate": 0.00024085055165496487, + "loss": 1.4262, + "theoretical_loss": 3.417616061531246, + "tokens_seen": 2317156352 + }, + { + "epoch": 0.7, + "learning_rate": 0.00024083450351053163, + "loss": 1.2515, + "theoretical_loss": 3.417608298846135, + "tokens_seen": 2317221888 + }, + { + "epoch": 0.7, + "learning_rate": 0.00024081845536609833, + "loss": 1.3399, + "theoretical_loss": 3.417600536442037, + "tokens_seen": 2317287424 + }, + { + "epoch": 0.7, + "learning_rate": 0.00024080240722166503, + "loss": 1.3625, + "theoretical_loss": 3.4175927743189343, + "tokens_seen": 2317352960 + }, + { + "epoch": 0.7, + "learning_rate": 0.00024078635907723173, + "loss": 1.3167, + "theoretical_loss": 3.417585012476808, + "tokens_seen": 2317418496 + }, + { + "epoch": 0.7, + "learning_rate": 0.00024077031093279843, + "loss": 1.2287, + "theoretical_loss": 3.41757725091564, + "tokens_seen": 2317484032 + }, + { + "epoch": 0.7, + "learning_rate": 0.00024075426278836513, + "loss": 1.4418, + "theoretical_loss": 3.4175694896354125, + "tokens_seen": 2317549568 + }, + { + "epoch": 0.7, + "learning_rate": 0.00024073821464393183, + "loss": 1.3966, + "theoretical_loss": 3.417561728636108, + "tokens_seen": 2317615104 + }, + { + "epoch": 0.7, + "learning_rate": 0.00024072216649949853, + "loss": 1.3166, + "theoretical_loss": 3.4175539679177076, + "tokens_seen": 2317680640 + }, + { + "epoch": 0.7, + "learning_rate": 0.00024070611835506523, + "loss": 1.4189, + "theoretical_loss": 3.4175462074801928, + "tokens_seen": 2317746176 + }, + { + "epoch": 0.7, + "learning_rate": 0.00024069007021063193, + "loss": 1.367, + "theoretical_loss": 3.4175384473235466, + "tokens_seen": 2317811712 + }, + { + "epoch": 0.7, + "learning_rate": 0.00024067402206619863, + "loss": 1.3138, + "theoretical_loss": 3.4175306874477505, + "tokens_seen": 2317877248 + }, + { + "epoch": 0.7, + "learning_rate": 0.00024065797392176533, + "loss": 1.3893, + "theoretical_loss": 3.417522927852786, + "tokens_seen": 2317942784 + }, + { + "epoch": 0.7, + "learning_rate": 0.00024064192577733203, + "loss": 1.2156, + "theoretical_loss": 3.4175151685386354, + "tokens_seen": 2318008320 + }, + { + "epoch": 0.7, + "learning_rate": 0.0002406258776328987, + "loss": 1.4198, + "theoretical_loss": 3.4175074095052804, + "tokens_seen": 2318073856 + }, + { + "epoch": 0.7, + "learning_rate": 0.0002406098294884654, + "loss": 1.3896, + "theoretical_loss": 3.4174996507527027, + "tokens_seen": 2318139392 + }, + { + "epoch": 0.7, + "learning_rate": 0.0002405937813440321, + "loss": 1.3228, + "theoretical_loss": 3.417491892280885, + "tokens_seen": 2318204928 + }, + { + "epoch": 0.7, + "learning_rate": 0.0002405777331995988, + "loss": 1.4257, + "theoretical_loss": 3.417484134089808, + "tokens_seen": 2318270464 + }, + { + "epoch": 0.7, + "learning_rate": 0.0002405616850551655, + "loss": 1.3105, + "theoretical_loss": 3.417476376179455, + "tokens_seen": 2318336000 + }, + { + "epoch": 0.7, + "learning_rate": 0.0002405456369107322, + "loss": 1.32, + "theoretical_loss": 3.4174686185498064, + "tokens_seen": 2318401536 + }, + { + "epoch": 0.7, + "learning_rate": 0.0002405295887662989, + "loss": 1.2909, + "theoretical_loss": 3.4174608612008455, + "tokens_seen": 2318467072 + }, + { + "epoch": 0.7, + "learning_rate": 0.0002405135406218656, + "loss": 1.2567, + "theoretical_loss": 3.4174531041325533, + "tokens_seen": 2318532608 + }, + { + "epoch": 0.7, + "learning_rate": 0.0002404974924774323, + "loss": 1.3075, + "theoretical_loss": 3.4174453473449122, + "tokens_seen": 2318598144 + }, + { + "epoch": 0.7, + "learning_rate": 0.00024048144433299901, + "loss": 1.396, + "theoretical_loss": 3.4174375908379035, + "tokens_seen": 2318663680 + }, + { + "epoch": 0.7, + "learning_rate": 0.00024046539618856571, + "loss": 1.3221, + "theoretical_loss": 3.41742983461151, + "tokens_seen": 2318729216 + }, + { + "epoch": 0.7, + "learning_rate": 0.00024044934804413242, + "loss": 1.3341, + "theoretical_loss": 3.4174220786657132, + "tokens_seen": 2318794752 + }, + { + "epoch": 0.7, + "learning_rate": 0.00024043329989969912, + "loss": 1.2689, + "theoretical_loss": 3.4174143230004947, + "tokens_seen": 2318860288 + }, + { + "epoch": 0.7, + "learning_rate": 0.00024041725175526582, + "loss": 1.3335, + "theoretical_loss": 3.4174065676158367, + "tokens_seen": 2318925824 + }, + { + "epoch": 0.7, + "learning_rate": 0.00024040120361083252, + "loss": 1.2904, + "theoretical_loss": 3.4173988125117214, + "tokens_seen": 2318991360 + }, + { + "epoch": 0.7, + "learning_rate": 0.00024038515546639922, + "loss": 1.3611, + "theoretical_loss": 3.41739105768813, + "tokens_seen": 2319056896 + }, + { + "epoch": 0.7, + "learning_rate": 0.00024036910732196592, + "loss": 1.3083, + "theoretical_loss": 3.417383303145045, + "tokens_seen": 2319122432 + }, + { + "epoch": 0.7, + "learning_rate": 0.00024035305917753262, + "loss": 1.341, + "theoretical_loss": 3.4173755488824487, + "tokens_seen": 2319187968 + }, + { + "epoch": 0.7, + "learning_rate": 0.00024033701103309932, + "loss": 1.4595, + "theoretical_loss": 3.417367794900322, + "tokens_seen": 2319253504 + }, + { + "epoch": 0.7, + "learning_rate": 0.00024032096288866602, + "loss": 1.3456, + "theoretical_loss": 3.417360041198647, + "tokens_seen": 2319319040 + }, + { + "epoch": 0.7, + "learning_rate": 0.00024030491474423272, + "loss": 1.2922, + "theoretical_loss": 3.4173522877774065, + "tokens_seen": 2319384576 + }, + { + "epoch": 0.7, + "learning_rate": 0.00024028886659979942, + "loss": 1.1931, + "theoretical_loss": 3.417344534636582, + "tokens_seen": 2319450112 + }, + { + "epoch": 0.7, + "learning_rate": 0.00024027281845536612, + "loss": 1.3097, + "theoretical_loss": 3.417336781776155, + "tokens_seen": 2319515648 + }, + { + "epoch": 0.7, + "learning_rate": 0.00024025677031093282, + "loss": 1.3421, + "theoretical_loss": 3.4173290291961083, + "tokens_seen": 2319581184 + }, + { + "epoch": 0.7, + "learning_rate": 0.0002402407221664995, + "loss": 1.3832, + "theoretical_loss": 3.417321276896423, + "tokens_seen": 2319646720 + }, + { + "epoch": 0.7, + "learning_rate": 0.0002402246740220662, + "loss": 1.4526, + "theoretical_loss": 3.4173135248770814, + "tokens_seen": 2319712256 + }, + { + "epoch": 0.7, + "learning_rate": 0.0002402086258776329, + "loss": 1.2291, + "theoretical_loss": 3.4173057731380654, + "tokens_seen": 2319777792 + }, + { + "epoch": 0.7, + "learning_rate": 0.0002401925777331996, + "loss": 1.2756, + "theoretical_loss": 3.4172980216793567, + "tokens_seen": 2319843328 + }, + { + "epoch": 0.7, + "learning_rate": 0.0002401765295887663, + "loss": 1.4218, + "theoretical_loss": 3.417290270500938, + "tokens_seen": 2319908864 + }, + { + "epoch": 0.7, + "objective/train/docs_used": 844817, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.1530518531799316, + "objective/train/theoretical_loss": 3.4172825196027903, + "objective/train/tokens_used": 2340434400, + "theoretical_loss": 3.4172825196027903, + "tokens_seen": 2319974400 + }, + { + "epoch": 0.7, + "learning_rate": 0.000240160481444333, + "loss": 1.1811, + "theoretical_loss": 3.4172825196027903, + "tokens_seen": 2319974400 + }, + { + "epoch": 0.7, + "learning_rate": 0.0002401444332998997, + "loss": 1.4717, + "theoretical_loss": 3.417274768984896, + "tokens_seen": 2320039936 + }, + { + "epoch": 0.7, + "learning_rate": 0.0002401283851554664, + "loss": 1.3319, + "theoretical_loss": 3.4172670186472374, + "tokens_seen": 2320105472 + }, + { + "epoch": 0.7, + "learning_rate": 0.0002401123370110331, + "loss": 1.2265, + "theoretical_loss": 3.4172592685897962, + "tokens_seen": 2320171008 + }, + { + "epoch": 0.7, + "learning_rate": 0.0002400962888665998, + "loss": 1.3208, + "theoretical_loss": 3.417251518812554, + "tokens_seen": 2320236544 + }, + { + "epoch": 0.7, + "learning_rate": 0.0002400802407221665, + "loss": 1.2247, + "theoretical_loss": 3.417243769315493, + "tokens_seen": 2320302080 + }, + { + "epoch": 0.7, + "learning_rate": 0.0002400641925777332, + "loss": 1.256, + "theoretical_loss": 3.417236020098595, + "tokens_seen": 2320367616 + }, + { + "epoch": 0.7, + "learning_rate": 0.0002400481444332999, + "loss": 1.3303, + "theoretical_loss": 3.4172282711618425, + "tokens_seen": 2320433152 + }, + { + "epoch": 0.7, + "learning_rate": 0.0002400320962888666, + "loss": 1.2805, + "theoretical_loss": 3.417220522505217, + "tokens_seen": 2320498688 + }, + { + "epoch": 0.7, + "learning_rate": 0.0002400160481444333, + "loss": 1.3258, + "theoretical_loss": 3.4172127741287004, + "tokens_seen": 2320564224 + }, + { + "epoch": 0.7, + "learning_rate": 0.00024, + "loss": 1.4894, + "theoretical_loss": 3.4172050260322746, + "tokens_seen": 2320629760 + }, + { + "epoch": 0.7, + "learning_rate": 0.0002399839518555667, + "loss": 1.3323, + "theoretical_loss": 3.417197278215922, + "tokens_seen": 2320695296 + }, + { + "epoch": 0.7, + "learning_rate": 0.0002399679037111334, + "loss": 1.2883, + "theoretical_loss": 3.4171895306796243, + "tokens_seen": 2320760832 + }, + { + "epoch": 0.7, + "learning_rate": 0.0002399518555667001, + "loss": 1.257, + "theoretical_loss": 3.4171817834233638, + "tokens_seen": 2320826368 + }, + { + "epoch": 0.7, + "learning_rate": 0.0002399358074222668, + "loss": 1.3695, + "theoretical_loss": 3.417174036447122, + "tokens_seen": 2320891904 + }, + { + "epoch": 0.7, + "learning_rate": 0.0002399197592778335, + "loss": 1.2939, + "theoretical_loss": 3.417166289750881, + "tokens_seen": 2320957440 + }, + { + "epoch": 0.7, + "learning_rate": 0.0002399037111334002, + "loss": 1.347, + "theoretical_loss": 3.417158543334623, + "tokens_seen": 2321022976 + }, + { + "epoch": 0.7, + "learning_rate": 0.0002398876629889669, + "loss": 1.3171, + "theoretical_loss": 3.4171507971983295, + "tokens_seen": 2321088512 + }, + { + "epoch": 0.7, + "learning_rate": 0.0002398716148445336, + "loss": 1.359, + "theoretical_loss": 3.4171430513419834, + "tokens_seen": 2321154048 + }, + { + "epoch": 0.7, + "learning_rate": 0.00023985556670010028, + "loss": 1.3632, + "theoretical_loss": 3.4171353057655653, + "tokens_seen": 2321219584 + }, + { + "epoch": 0.7, + "learning_rate": 0.00023983951855566699, + "loss": 1.2362, + "theoretical_loss": 3.4171275604690585, + "tokens_seen": 2321285120 + }, + { + "epoch": 0.7, + "learning_rate": 0.00023982347041123369, + "loss": 1.3256, + "theoretical_loss": 3.417119815452444, + "tokens_seen": 2321350656 + }, + { + "epoch": 0.7, + "learning_rate": 0.0002398074222668004, + "loss": 1.293, + "theoretical_loss": 3.4171120707157043, + "tokens_seen": 2321416192 + }, + { + "epoch": 0.7, + "learning_rate": 0.0002397913741223671, + "loss": 1.2886, + "theoretical_loss": 3.4171043262588214, + "tokens_seen": 2321481728 + }, + { + "epoch": 0.7, + "learning_rate": 0.0002397753259779338, + "loss": 1.4293, + "theoretical_loss": 3.4170965820817774, + "tokens_seen": 2321547264 + }, + { + "epoch": 0.7, + "learning_rate": 0.0002397592778335005, + "loss": 1.2952, + "theoretical_loss": 3.417088838184554, + "tokens_seen": 2321612800 + }, + { + "epoch": 0.7, + "learning_rate": 0.0002397432296890672, + "loss": 1.3502, + "theoretical_loss": 3.417081094567133, + "tokens_seen": 2321678336 + }, + { + "epoch": 0.7, + "learning_rate": 0.0002397271815446339, + "loss": 1.2475, + "theoretical_loss": 3.4170733512294964, + "tokens_seen": 2321743872 + }, + { + "epoch": 0.7, + "learning_rate": 0.0002397111334002006, + "loss": 1.3501, + "theoretical_loss": 3.4170656081716273, + "tokens_seen": 2321809408 + }, + { + "epoch": 0.7, + "learning_rate": 0.0002396950852557673, + "loss": 1.2414, + "theoretical_loss": 3.417057865393506, + "tokens_seen": 2321874944 + }, + { + "epoch": 0.7, + "learning_rate": 0.000239679037111334, + "loss": 1.304, + "theoretical_loss": 3.417050122895116, + "tokens_seen": 2321940480 + }, + { + "epoch": 0.7, + "learning_rate": 0.0002396629889669007, + "loss": 1.3994, + "theoretical_loss": 3.4170423806764383, + "tokens_seen": 2322006016 + }, + { + "epoch": 0.7, + "learning_rate": 0.0002396469408224674, + "loss": 1.3723, + "theoretical_loss": 3.417034638737455, + "tokens_seen": 2322071552 + }, + { + "epoch": 0.7, + "learning_rate": 0.00023963089267803415, + "loss": 1.2775, + "theoretical_loss": 3.417026897078149, + "tokens_seen": 2322137088 + }, + { + "epoch": 0.7, + "learning_rate": 0.00023961484453360085, + "loss": 1.1934, + "theoretical_loss": 3.4170191556985015, + "tokens_seen": 2322202624 + }, + { + "epoch": 0.7, + "learning_rate": 0.00023959879638916752, + "loss": 1.3366, + "theoretical_loss": 3.4170114145984942, + "tokens_seen": 2322268160 + }, + { + "epoch": 0.7, + "learning_rate": 0.00023958274824473422, + "loss": 1.3779, + "theoretical_loss": 3.41700367377811, + "tokens_seen": 2322333696 + }, + { + "epoch": 0.7, + "learning_rate": 0.00023956670010030092, + "loss": 1.3295, + "theoretical_loss": 3.41699593323733, + "tokens_seen": 2322399232 + }, + { + "epoch": 0.7, + "learning_rate": 0.00023955065195586762, + "loss": 1.2723, + "theoretical_loss": 3.416988192976137, + "tokens_seen": 2322464768 + }, + { + "epoch": 0.7, + "learning_rate": 0.00023953460381143432, + "loss": 1.2492, + "theoretical_loss": 3.4169804529945127, + "tokens_seen": 2322530304 + }, + { + "epoch": 0.7, + "learning_rate": 0.00023951855566700103, + "loss": 1.2326, + "theoretical_loss": 3.4169727132924392, + "tokens_seen": 2322595840 + }, + { + "epoch": 0.7, + "learning_rate": 0.00023950250752256773, + "loss": 1.4402, + "theoretical_loss": 3.416964973869898, + "tokens_seen": 2322661376 + }, + { + "epoch": 0.7, + "learning_rate": 0.00023948645937813443, + "loss": 1.2677, + "theoretical_loss": 3.416957234726872, + "tokens_seen": 2322726912 + }, + { + "epoch": 0.7, + "learning_rate": 0.00023947041123370113, + "loss": 1.296, + "theoretical_loss": 3.4169494958633426, + "tokens_seen": 2322792448 + }, + { + "epoch": 0.7, + "learning_rate": 0.00023945436308926783, + "loss": 1.3259, + "theoretical_loss": 3.416941757279292, + "tokens_seen": 2322857984 + }, + { + "epoch": 0.7, + "learning_rate": 0.00023943831494483453, + "loss": 1.4149, + "theoretical_loss": 3.416934018974702, + "tokens_seen": 2322923520 + }, + { + "epoch": 0.7, + "learning_rate": 0.00023942226680040123, + "loss": 1.3187, + "theoretical_loss": 3.4169262809495553, + "tokens_seen": 2322989056 + }, + { + "epoch": 0.7, + "learning_rate": 0.00023940621865596793, + "loss": 1.3364, + "theoretical_loss": 3.4169185432038334, + "tokens_seen": 2323054592 + }, + { + "epoch": 0.7, + "learning_rate": 0.00023939017051153463, + "loss": 1.2951, + "theoretical_loss": 3.416910805737518, + "tokens_seen": 2323120128 + }, + { + "epoch": 0.7, + "learning_rate": 0.00023937412236710133, + "loss": 1.2582, + "theoretical_loss": 3.4169030685505914, + "tokens_seen": 2323185664 + }, + { + "epoch": 0.7, + "objective/train/docs_used": 845757, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.2464518547058105, + "objective/train/theoretical_loss": 3.4168953316430364, + "objective/train/tokens_used": 2343711200, + "theoretical_loss": 3.4168953316430364, + "tokens_seen": 2323251200 + }, + { + "epoch": 0.7, + "learning_rate": 0.00023935807422266803, + "loss": 1.3344, + "theoretical_loss": 3.4168953316430364, + "tokens_seen": 2323251200 + }, + { + "epoch": 0.7, + "learning_rate": 0.00023934202607823473, + "loss": 1.2963, + "theoretical_loss": 3.416887595014834, + "tokens_seen": 2323316736 + }, + { + "epoch": 0.7, + "learning_rate": 0.00023932597793380143, + "loss": 1.4128, + "theoretical_loss": 3.4168798586659666, + "tokens_seen": 2323382272 + }, + { + "epoch": 0.7, + "learning_rate": 0.00023930992978936813, + "loss": 1.3446, + "theoretical_loss": 3.4168721225964163, + "tokens_seen": 2323447808 + }, + { + "epoch": 0.7, + "learning_rate": 0.00023929388164493483, + "loss": 1.3891, + "theoretical_loss": 3.4168643868061648, + "tokens_seen": 2323513344 + }, + { + "epoch": 0.7, + "learning_rate": 0.00023927783350050153, + "loss": 1.289, + "theoretical_loss": 3.416856651295195, + "tokens_seen": 2323578880 + }, + { + "epoch": 0.7, + "learning_rate": 0.00023926178535606824, + "loss": 1.3452, + "theoretical_loss": 3.4168489160634876, + "tokens_seen": 2323644416 + }, + { + "epoch": 0.7, + "learning_rate": 0.00023924573721163494, + "loss": 1.2797, + "theoretical_loss": 3.4168411811110255, + "tokens_seen": 2323709952 + }, + { + "epoch": 0.7, + "learning_rate": 0.00023922968906720164, + "loss": 1.3839, + "theoretical_loss": 3.416833446437791, + "tokens_seen": 2323775488 + }, + { + "epoch": 0.7, + "learning_rate": 0.0002392136409227683, + "loss": 1.2249, + "theoretical_loss": 3.416825712043766, + "tokens_seen": 2323841024 + }, + { + "epoch": 0.7, + "learning_rate": 0.000239197592778335, + "loss": 1.3966, + "theoretical_loss": 3.416817977928932, + "tokens_seen": 2323906560 + }, + { + "epoch": 0.7, + "learning_rate": 0.0002391815446339017, + "loss": 1.2851, + "theoretical_loss": 3.4168102440932717, + "tokens_seen": 2323972096 + }, + { + "epoch": 0.7, + "learning_rate": 0.0002391654964894684, + "loss": 1.2795, + "theoretical_loss": 3.4168025105367663, + "tokens_seen": 2324037632 + }, + { + "epoch": 0.7, + "learning_rate": 0.0002391494483450351, + "loss": 1.2999, + "theoretical_loss": 3.4167947772593985, + "tokens_seen": 2324103168 + }, + { + "epoch": 0.7, + "learning_rate": 0.0002391334002006018, + "loss": 1.3293, + "theoretical_loss": 3.4167870442611505, + "tokens_seen": 2324168704 + }, + { + "epoch": 0.7, + "learning_rate": 0.00023911735205616851, + "loss": 1.3342, + "theoretical_loss": 3.416779311542004, + "tokens_seen": 2324234240 + }, + { + "epoch": 0.7, + "learning_rate": 0.00023910130391173521, + "loss": 1.3059, + "theoretical_loss": 3.4167715791019413, + "tokens_seen": 2324299776 + }, + { + "epoch": 0.7, + "learning_rate": 0.00023908525576730192, + "loss": 1.3267, + "theoretical_loss": 3.4167638469409445, + "tokens_seen": 2324365312 + }, + { + "epoch": 0.7, + "learning_rate": 0.00023906920762286862, + "loss": 1.3715, + "theoretical_loss": 3.4167561150589947, + "tokens_seen": 2324430848 + }, + { + "epoch": 0.7, + "learning_rate": 0.00023905315947843532, + "loss": 1.3248, + "theoretical_loss": 3.4167483834560755, + "tokens_seen": 2324496384 + }, + { + "epoch": 0.7, + "learning_rate": 0.00023903711133400202, + "loss": 1.3301, + "theoretical_loss": 3.4167406521321677, + "tokens_seen": 2324561920 + }, + { + "epoch": 0.7, + "learning_rate": 0.00023902106318956872, + "loss": 1.2202, + "theoretical_loss": 3.416732921087254, + "tokens_seen": 2324627456 + }, + { + "epoch": 0.7, + "learning_rate": 0.00023900501504513542, + "loss": 1.3135, + "theoretical_loss": 3.416725190321317, + "tokens_seen": 2324692992 + }, + { + "epoch": 0.7, + "learning_rate": 0.00023898896690070212, + "loss": 1.2982, + "theoretical_loss": 3.416717459834337, + "tokens_seen": 2324758528 + }, + { + "epoch": 0.7, + "learning_rate": 0.00023897291875626882, + "loss": 1.3336, + "theoretical_loss": 3.416709729626298, + "tokens_seen": 2324824064 + }, + { + "epoch": 0.7, + "learning_rate": 0.00023895687061183552, + "loss": 1.247, + "theoretical_loss": 3.416701999697181, + "tokens_seen": 2324889600 + }, + { + "epoch": 0.7, + "learning_rate": 0.00023894082246740222, + "loss": 1.3397, + "theoretical_loss": 3.416694270046968, + "tokens_seen": 2324955136 + }, + { + "epoch": 0.7, + "learning_rate": 0.00023892477432296892, + "loss": 1.2224, + "theoretical_loss": 3.4166865406756415, + "tokens_seen": 2325020672 + }, + { + "epoch": 0.7, + "learning_rate": 0.00023890872617853562, + "loss": 1.3349, + "theoretical_loss": 3.416678811583184, + "tokens_seen": 2325086208 + }, + { + "epoch": 0.7, + "learning_rate": 0.00023889267803410232, + "loss": 1.1911, + "theoretical_loss": 3.4166710827695765, + "tokens_seen": 2325151744 + }, + { + "epoch": 0.7, + "learning_rate": 0.00023887662988966902, + "loss": 1.368, + "theoretical_loss": 3.416663354234802, + "tokens_seen": 2325217280 + }, + { + "epoch": 0.7, + "learning_rate": 0.00023886058174523572, + "loss": 1.2862, + "theoretical_loss": 3.416655625978842, + "tokens_seen": 2325282816 + }, + { + "epoch": 0.7, + "learning_rate": 0.00023884453360080242, + "loss": 1.3963, + "theoretical_loss": 3.4166478980016786, + "tokens_seen": 2325348352 + }, + { + "epoch": 0.7, + "learning_rate": 0.0002388284854563691, + "loss": 1.4229, + "theoretical_loss": 3.4166401703032943, + "tokens_seen": 2325413888 + }, + { + "epoch": 0.7, + "learning_rate": 0.0002388124373119358, + "loss": 1.345, + "theoretical_loss": 3.416632442883671, + "tokens_seen": 2325479424 + }, + { + "epoch": 0.7, + "learning_rate": 0.0002387963891675025, + "loss": 1.211, + "theoretical_loss": 3.416624715742791, + "tokens_seen": 2325544960 + }, + { + "epoch": 0.7, + "learning_rate": 0.0002387803410230692, + "loss": 1.4224, + "theoretical_loss": 3.416616988880636, + "tokens_seen": 2325610496 + }, + { + "epoch": 0.7, + "learning_rate": 0.0002387642928786359, + "loss": 1.2307, + "theoretical_loss": 3.416609262297188, + "tokens_seen": 2325676032 + }, + { + "epoch": 0.7, + "learning_rate": 0.0002387482447342026, + "loss": 1.2893, + "theoretical_loss": 3.4166015359924296, + "tokens_seen": 2325741568 + }, + { + "epoch": 0.7, + "learning_rate": 0.0002387321965897693, + "loss": 1.3013, + "theoretical_loss": 3.416593809966342, + "tokens_seen": 2325807104 + }, + { + "epoch": 0.7, + "learning_rate": 0.000238716148445336, + "loss": 1.2785, + "theoretical_loss": 3.416586084218909, + "tokens_seen": 2325872640 + }, + { + "epoch": 0.7, + "learning_rate": 0.0002387001003009027, + "loss": 1.3393, + "theoretical_loss": 3.416578358750111, + "tokens_seen": 2325938176 + }, + { + "epoch": 0.7, + "learning_rate": 0.0002386840521564694, + "loss": 1.3837, + "theoretical_loss": 3.416570633559931, + "tokens_seen": 2326003712 + }, + { + "epoch": 0.7, + "learning_rate": 0.0002386680040120361, + "loss": 1.346, + "theoretical_loss": 3.41656290864835, + "tokens_seen": 2326069248 + }, + { + "epoch": 0.7, + "learning_rate": 0.0002386519558676028, + "loss": 1.2875, + "theoretical_loss": 3.416555184015352, + "tokens_seen": 2326134784 + }, + { + "epoch": 0.7, + "learning_rate": 0.0002386359077231695, + "loss": 1.4024, + "theoretical_loss": 3.4165474596609173, + "tokens_seen": 2326200320 + }, + { + "epoch": 0.7, + "learning_rate": 0.0002386198595787362, + "loss": 1.3574, + "theoretical_loss": 3.416539735585029, + "tokens_seen": 2326265856 + }, + { + "epoch": 0.7, + "learning_rate": 0.0002386038114343029, + "loss": 1.3496, + "theoretical_loss": 3.416532011787669, + "tokens_seen": 2326331392 + }, + { + "epoch": 0.7, + "learning_rate": 0.0002385877632898696, + "loss": 1.2931, + "theoretical_loss": 3.4165242882688194, + "tokens_seen": 2326396928 + }, + { + "epoch": 0.7, + "learning_rate": 0.0002385717151454363, + "loss": 1.2964, + "theoretical_loss": 3.4165165650284623, + "tokens_seen": 2326462464 + }, + { + "epoch": 0.7, + "objective/train/docs_used": 847012, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.3350573778152466, + "objective/train/theoretical_loss": 3.4165088420665795, + "objective/train/tokens_used": 2346988000, + "theoretical_loss": 3.4165088420665795, + "tokens_seen": 2326528000 + }, + { + "epoch": 0.71, + "learning_rate": 0.000238555667001003, + "loss": 1.347, + "theoretical_loss": 3.4165088420665795, + "tokens_seen": 2326528000 + }, + { + "epoch": 0.71, + "learning_rate": 0.0002385396188565697, + "loss": 1.3201, + "theoretical_loss": 3.416501119383154, + "tokens_seen": 2326593536 + }, + { + "epoch": 0.71, + "learning_rate": 0.0002385235707121364, + "loss": 1.3141, + "theoretical_loss": 3.416493396978167, + "tokens_seen": 2326659072 + }, + { + "epoch": 0.71, + "learning_rate": 0.0002385075225677031, + "loss": 1.2615, + "theoretical_loss": 3.4164856748516006, + "tokens_seen": 2326724608 + }, + { + "epoch": 0.71, + "learning_rate": 0.0002384914744232698, + "loss": 1.3181, + "theoretical_loss": 3.4164779530034375, + "tokens_seen": 2326790144 + }, + { + "epoch": 0.71, + "learning_rate": 0.0002384754262788365, + "loss": 1.3772, + "theoretical_loss": 3.4164702314336597, + "tokens_seen": 2326855680 + }, + { + "epoch": 0.71, + "learning_rate": 0.0002384593781344032, + "loss": 1.2288, + "theoretical_loss": 3.416462510142249, + "tokens_seen": 2326921216 + }, + { + "epoch": 0.71, + "learning_rate": 0.00023844332998996989, + "loss": 1.2989, + "theoretical_loss": 3.4164547891291877, + "tokens_seen": 2326986752 + }, + { + "epoch": 0.71, + "learning_rate": 0.00023842728184553664, + "loss": 1.2809, + "theoretical_loss": 3.4164470683944583, + "tokens_seen": 2327052288 + }, + { + "epoch": 0.71, + "learning_rate": 0.00023841123370110334, + "loss": 1.2618, + "theoretical_loss": 3.4164393479380424, + "tokens_seen": 2327117824 + }, + { + "epoch": 0.71, + "learning_rate": 0.00023839518555667004, + "loss": 1.2148, + "theoretical_loss": 3.4164316277599225, + "tokens_seen": 2327183360 + }, + { + "epoch": 0.71, + "learning_rate": 0.00023837913741223674, + "loss": 1.3147, + "theoretical_loss": 3.4164239078600804, + "tokens_seen": 2327248896 + }, + { + "epoch": 0.71, + "learning_rate": 0.00023836308926780344, + "loss": 1.2639, + "theoretical_loss": 3.416416188238498, + "tokens_seen": 2327314432 + }, + { + "epoch": 0.71, + "learning_rate": 0.00023834704112337014, + "loss": 1.3573, + "theoretical_loss": 3.4164084688951584, + "tokens_seen": 2327379968 + }, + { + "epoch": 0.71, + "learning_rate": 0.00023833099297893684, + "loss": 1.2989, + "theoretical_loss": 3.416400749830043, + "tokens_seen": 2327445504 + }, + { + "epoch": 0.71, + "learning_rate": 0.00023831494483450355, + "loss": 1.347, + "theoretical_loss": 3.4163930310431336, + "tokens_seen": 2327511040 + }, + { + "epoch": 0.71, + "learning_rate": 0.00023829889669007025, + "loss": 1.3133, + "theoretical_loss": 3.4163853125344135, + "tokens_seen": 2327576576 + }, + { + "epoch": 0.71, + "learning_rate": 0.00023828284854563695, + "loss": 1.1517, + "theoretical_loss": 3.4163775943038637, + "tokens_seen": 2327642112 + }, + { + "epoch": 0.71, + "learning_rate": 0.00023826680040120365, + "loss": 1.2424, + "theoretical_loss": 3.4163698763514674, + "tokens_seen": 2327707648 + }, + { + "epoch": 0.71, + "learning_rate": 0.00023825075225677035, + "loss": 1.2678, + "theoretical_loss": 3.4163621586772055, + "tokens_seen": 2327773184 + }, + { + "epoch": 0.71, + "learning_rate": 0.00023823470411233705, + "loss": 1.1669, + "theoretical_loss": 3.416354441281061, + "tokens_seen": 2327838720 + }, + { + "epoch": 0.71, + "learning_rate": 0.00023821865596790375, + "loss": 1.3499, + "theoretical_loss": 3.416346724163016, + "tokens_seen": 2327904256 + }, + { + "epoch": 0.71, + "learning_rate": 0.00023820260782347042, + "loss": 1.1918, + "theoretical_loss": 3.416339007323052, + "tokens_seen": 2327969792 + }, + { + "epoch": 0.71, + "learning_rate": 0.00023818655967903712, + "loss": 1.425, + "theoretical_loss": 3.4163312907611525, + "tokens_seen": 2328035328 + }, + { + "epoch": 0.71, + "learning_rate": 0.00023817051153460382, + "loss": 1.3058, + "theoretical_loss": 3.4163235744772984, + "tokens_seen": 2328100864 + }, + { + "epoch": 0.71, + "learning_rate": 0.00023815446339017052, + "loss": 1.2675, + "theoretical_loss": 3.416315858471472, + "tokens_seen": 2328166400 + }, + { + "epoch": 0.71, + "learning_rate": 0.00023813841524573723, + "loss": 1.3337, + "theoretical_loss": 3.4163081427436564, + "tokens_seen": 2328231936 + }, + { + "epoch": 0.71, + "learning_rate": 0.00023812236710130393, + "loss": 1.325, + "theoretical_loss": 3.4163004272938324, + "tokens_seen": 2328297472 + }, + { + "epoch": 0.71, + "learning_rate": 0.00023810631895687063, + "loss": 1.2932, + "theoretical_loss": 3.416292712121983, + "tokens_seen": 2328363008 + }, + { + "epoch": 0.71, + "learning_rate": 0.00023809027081243733, + "loss": 1.3282, + "theoretical_loss": 3.4162849972280904, + "tokens_seen": 2328428544 + }, + { + "epoch": 0.71, + "learning_rate": 0.00023807422266800403, + "loss": 1.3528, + "theoretical_loss": 3.416277282612137, + "tokens_seen": 2328494080 + }, + { + "epoch": 0.71, + "learning_rate": 0.00023805817452357073, + "loss": 1.2584, + "theoretical_loss": 3.4162695682741036, + "tokens_seen": 2328559616 + }, + { + "epoch": 0.71, + "learning_rate": 0.00023804212637913743, + "loss": 1.3877, + "theoretical_loss": 3.4162618542139738, + "tokens_seen": 2328625152 + }, + { + "epoch": 0.71, + "learning_rate": 0.00023802607823470413, + "loss": 1.3038, + "theoretical_loss": 3.416254140431729, + "tokens_seen": 2328690688 + }, + { + "epoch": 0.71, + "learning_rate": 0.00023801003009027083, + "loss": 1.2928, + "theoretical_loss": 3.416246426927352, + "tokens_seen": 2328756224 + }, + { + "epoch": 0.71, + "learning_rate": 0.00023799398194583753, + "loss": 1.3311, + "theoretical_loss": 3.416238713700824, + "tokens_seen": 2328821760 + }, + { + "epoch": 0.71, + "learning_rate": 0.00023797793380140423, + "loss": 1.3552, + "theoretical_loss": 3.4162310007521284, + "tokens_seen": 2328887296 + }, + { + "epoch": 0.71, + "learning_rate": 0.00023796188565697093, + "loss": 1.2322, + "theoretical_loss": 3.4162232880812464, + "tokens_seen": 2328952832 + }, + { + "epoch": 0.71, + "learning_rate": 0.00023794583751253763, + "loss": 1.3149, + "theoretical_loss": 3.4162155756881605, + "tokens_seen": 2329018368 + }, + { + "epoch": 0.71, + "learning_rate": 0.00023792978936810433, + "loss": 1.473, + "theoretical_loss": 3.416207863572853, + "tokens_seen": 2329083904 + }, + { + "epoch": 0.71, + "learning_rate": 0.00023791374122367103, + "loss": 1.2149, + "theoretical_loss": 3.4162001517353064, + "tokens_seen": 2329149440 + }, + { + "epoch": 0.71, + "learning_rate": 0.00023789769307923773, + "loss": 1.2439, + "theoretical_loss": 3.416192440175502, + "tokens_seen": 2329214976 + }, + { + "epoch": 0.71, + "learning_rate": 0.00023788164493480444, + "loss": 1.2475, + "theoretical_loss": 3.4161847288934224, + "tokens_seen": 2329280512 + }, + { + "epoch": 0.71, + "learning_rate": 0.00023786559679037114, + "loss": 1.3562, + "theoretical_loss": 3.41617701788905, + "tokens_seen": 2329346048 + }, + { + "epoch": 0.71, + "learning_rate": 0.00023784954864593784, + "loss": 1.3296, + "theoretical_loss": 3.4161693071623667, + "tokens_seen": 2329411584 + }, + { + "epoch": 0.71, + "learning_rate": 0.00023783350050150454, + "loss": 1.3388, + "theoretical_loss": 3.416161596713355, + "tokens_seen": 2329477120 + }, + { + "epoch": 0.71, + "learning_rate": 0.0002378174523570712, + "loss": 1.3327, + "theoretical_loss": 3.4161538865419967, + "tokens_seen": 2329542656 + }, + { + "epoch": 0.71, + "learning_rate": 0.0002378014042126379, + "loss": 1.3532, + "theoretical_loss": 3.416146176648274, + "tokens_seen": 2329608192 + }, + { + "epoch": 0.71, + "learning_rate": 0.0002377853560682046, + "loss": 1.3534, + "theoretical_loss": 3.41613846703217, + "tokens_seen": 2329673728 + }, + { + "epoch": 0.71, + "learning_rate": 0.0002377693079237713, + "loss": 1.2692, + "theoretical_loss": 3.416130757693666, + "tokens_seen": 2329739264 + }, + { + "epoch": 0.71, + "objective/train/docs_used": 848291, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.2697076797485352, + "objective/train/theoretical_loss": 3.416123048632744, + "objective/train/tokens_used": 2350264800, + "theoretical_loss": 3.416123048632744, + "tokens_seen": 2329804800 + }, + { + "epoch": 0.71, + "learning_rate": 0.00023775325977933801, + "loss": 1.2959, + "theoretical_loss": 3.416123048632744, + "tokens_seen": 2329804800 + }, + { + "epoch": 0.71, + "learning_rate": 0.00023773721163490471, + "loss": 1.2385, + "theoretical_loss": 3.4161153398493864, + "tokens_seen": 2329870336 + }, + { + "epoch": 0.71, + "learning_rate": 0.00023772116349047141, + "loss": 1.326, + "theoretical_loss": 3.416107631343576, + "tokens_seen": 2329935872 + }, + { + "epoch": 0.71, + "learning_rate": 0.00023770511534603812, + "loss": 1.2868, + "theoretical_loss": 3.4160999231152944, + "tokens_seen": 2330001408 + }, + { + "epoch": 0.71, + "learning_rate": 0.00023768906720160482, + "loss": 1.3613, + "theoretical_loss": 3.416092215164524, + "tokens_seen": 2330066944 + }, + { + "epoch": 0.71, + "learning_rate": 0.00023767301905717152, + "loss": 1.2286, + "theoretical_loss": 3.416084507491247, + "tokens_seen": 2330132480 + }, + { + "epoch": 0.71, + "learning_rate": 0.00023765697091273822, + "loss": 1.3399, + "theoretical_loss": 3.4160768000954453, + "tokens_seen": 2330198016 + }, + { + "epoch": 0.71, + "learning_rate": 0.00023764092276830492, + "loss": 1.2526, + "theoretical_loss": 3.416069092977102, + "tokens_seen": 2330263552 + }, + { + "epoch": 0.71, + "learning_rate": 0.00023762487462387162, + "loss": 1.2665, + "theoretical_loss": 3.4160613861361977, + "tokens_seen": 2330329088 + }, + { + "epoch": 0.71, + "learning_rate": 0.00023760882647943832, + "loss": 1.2672, + "theoretical_loss": 3.416053679572716, + "tokens_seen": 2330394624 + }, + { + "epoch": 0.71, + "learning_rate": 0.00023759277833500502, + "loss": 1.39, + "theoretical_loss": 3.4160459732866393, + "tokens_seen": 2330460160 + }, + { + "epoch": 0.71, + "learning_rate": 0.00023757673019057172, + "loss": 1.407, + "theoretical_loss": 3.4160382672779486, + "tokens_seen": 2330525696 + }, + { + "epoch": 0.71, + "learning_rate": 0.00023756068204613842, + "loss": 1.2482, + "theoretical_loss": 3.416030561546627, + "tokens_seen": 2330591232 + }, + { + "epoch": 0.71, + "learning_rate": 0.00023754463390170512, + "loss": 1.2986, + "theoretical_loss": 3.4160228560926567, + "tokens_seen": 2330656768 + }, + { + "epoch": 0.71, + "learning_rate": 0.00023752858575727182, + "loss": 1.3404, + "theoretical_loss": 3.4160151509160195, + "tokens_seen": 2330722304 + }, + { + "epoch": 0.71, + "learning_rate": 0.00023751253761283852, + "loss": 1.3701, + "theoretical_loss": 3.416007446016698, + "tokens_seen": 2330787840 + }, + { + "epoch": 0.71, + "learning_rate": 0.00023749648946840522, + "loss": 1.3594, + "theoretical_loss": 3.415999741394674, + "tokens_seen": 2330853376 + }, + { + "epoch": 0.71, + "learning_rate": 0.00023748044132397192, + "loss": 1.2231, + "theoretical_loss": 3.4159920370499295, + "tokens_seen": 2330918912 + }, + { + "epoch": 0.71, + "learning_rate": 0.00023746439317953863, + "loss": 1.3195, + "theoretical_loss": 3.4159843329824477, + "tokens_seen": 2330984448 + }, + { + "epoch": 0.71, + "learning_rate": 0.00023744834503510533, + "loss": 1.3365, + "theoretical_loss": 3.41597662919221, + "tokens_seen": 2331049984 + }, + { + "epoch": 0.71, + "learning_rate": 0.000237432296890672, + "loss": 1.405, + "theoretical_loss": 3.4159689256791994, + "tokens_seen": 2331115520 + }, + { + "epoch": 0.71, + "learning_rate": 0.0002374162487462387, + "loss": 1.3038, + "theoretical_loss": 3.4159612224433973, + "tokens_seen": 2331181056 + }, + { + "epoch": 0.71, + "learning_rate": 0.0002374002006018054, + "loss": 1.3406, + "theoretical_loss": 3.4159535194847868, + "tokens_seen": 2331246592 + }, + { + "epoch": 0.71, + "learning_rate": 0.0002373841524573721, + "loss": 1.3649, + "theoretical_loss": 3.4159458168033487, + "tokens_seen": 2331312128 + }, + { + "epoch": 0.71, + "learning_rate": 0.0002373681043129388, + "loss": 1.3193, + "theoretical_loss": 3.415938114399067, + "tokens_seen": 2331377664 + }, + { + "epoch": 0.71, + "learning_rate": 0.0002373520561685055, + "loss": 1.2296, + "theoretical_loss": 3.4159304122719227, + "tokens_seen": 2331443200 + }, + { + "epoch": 0.71, + "learning_rate": 0.0002373360080240722, + "loss": 1.3167, + "theoretical_loss": 3.4159227104218983, + "tokens_seen": 2331508736 + }, + { + "epoch": 0.71, + "learning_rate": 0.0002373199598796389, + "loss": 1.2755, + "theoretical_loss": 3.4159150088489767, + "tokens_seen": 2331574272 + }, + { + "epoch": 0.71, + "learning_rate": 0.0002373039117352056, + "loss": 1.3593, + "theoretical_loss": 3.415907307553139, + "tokens_seen": 2331639808 + }, + { + "epoch": 0.71, + "learning_rate": 0.0002372878635907723, + "loss": 1.3453, + "theoretical_loss": 3.4158996065343685, + "tokens_seen": 2331705344 + }, + { + "epoch": 0.71, + "learning_rate": 0.000237271815446339, + "loss": 1.2756, + "theoretical_loss": 3.4158919057926473, + "tokens_seen": 2331770880 + }, + { + "epoch": 0.71, + "learning_rate": 0.0002372557673019057, + "loss": 1.3364, + "theoretical_loss": 3.4158842053279566, + "tokens_seen": 2331836416 + }, + { + "epoch": 0.71, + "learning_rate": 0.0002372397191574724, + "loss": 1.3322, + "theoretical_loss": 3.4158765051402797, + "tokens_seen": 2331901952 + }, + { + "epoch": 0.71, + "learning_rate": 0.00023722367101303916, + "loss": 1.3347, + "theoretical_loss": 3.4158688052295987, + "tokens_seen": 2331967488 + }, + { + "epoch": 0.71, + "learning_rate": 0.00023720762286860586, + "loss": 1.4509, + "theoretical_loss": 3.4158611055958956, + "tokens_seen": 2332033024 + }, + { + "epoch": 0.71, + "learning_rate": 0.00023719157472417256, + "loss": 1.3325, + "theoretical_loss": 3.415853406239153, + "tokens_seen": 2332098560 + }, + { + "epoch": 0.71, + "learning_rate": 0.00023717552657973924, + "loss": 1.2946, + "theoretical_loss": 3.4158457071593524, + "tokens_seen": 2332164096 + }, + { + "epoch": 0.71, + "learning_rate": 0.00023715947843530594, + "loss": 1.2452, + "theoretical_loss": 3.415838008356477, + "tokens_seen": 2332229632 + }, + { + "epoch": 0.71, + "learning_rate": 0.00023714343029087264, + "loss": 1.34, + "theoretical_loss": 3.4158303098305085, + "tokens_seen": 2332295168 + }, + { + "epoch": 0.71, + "learning_rate": 0.00023712738214643934, + "loss": 1.3613, + "theoretical_loss": 3.4158226115814294, + "tokens_seen": 2332360704 + }, + { + "epoch": 0.71, + "learning_rate": 0.00023711133400200604, + "loss": 1.3812, + "theoretical_loss": 3.4158149136092217, + "tokens_seen": 2332426240 + }, + { + "epoch": 0.71, + "learning_rate": 0.00023709528585757274, + "loss": 1.3293, + "theoretical_loss": 3.415807215913868, + "tokens_seen": 2332491776 + }, + { + "epoch": 0.71, + "learning_rate": 0.00023707923771313944, + "loss": 1.3387, + "theoretical_loss": 3.4157995184953505, + "tokens_seen": 2332557312 + }, + { + "epoch": 0.71, + "learning_rate": 0.00023706318956870614, + "loss": 1.2619, + "theoretical_loss": 3.415791821353651, + "tokens_seen": 2332622848 + }, + { + "epoch": 0.71, + "learning_rate": 0.00023704714142427284, + "loss": 1.1902, + "theoretical_loss": 3.415784124488752, + "tokens_seen": 2332688384 + }, + { + "epoch": 0.71, + "learning_rate": 0.00023703109327983954, + "loss": 1.2338, + "theoretical_loss": 3.415776427900636, + "tokens_seen": 2332753920 + }, + { + "epoch": 0.71, + "learning_rate": 0.00023701504513540624, + "loss": 1.2846, + "theoretical_loss": 3.415768731589285, + "tokens_seen": 2332819456 + }, + { + "epoch": 0.71, + "learning_rate": 0.00023699899699097294, + "loss": 1.2521, + "theoretical_loss": 3.4157610355546817, + "tokens_seen": 2332884992 + }, + { + "epoch": 0.71, + "learning_rate": 0.00023698294884653964, + "loss": 1.3011, + "theoretical_loss": 3.415753339796808, + "tokens_seen": 2332950528 + }, + { + "epoch": 0.71, + "learning_rate": 0.00023696690070210634, + "loss": 1.339, + "theoretical_loss": 3.4157456443156464, + "tokens_seen": 2333016064 + }, + { + "epoch": 0.71, + "objective/train/docs_used": 849464, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.3212370872497559, + "objective/train/theoretical_loss": 3.415737949111179, + "objective/train/tokens_used": 2353541600, + "theoretical_loss": 3.415737949111179, + "tokens_seen": 2333081600 + }, + { + "epoch": 0.71, + "learning_rate": 0.00023695085255767305, + "loss": 1.3552, + "theoretical_loss": 3.415737949111179, + "tokens_seen": 2333081600 + }, + { + "epoch": 0.71, + "learning_rate": 0.00023693480441323975, + "loss": 1.4106, + "theoretical_loss": 3.4157302541833885, + "tokens_seen": 2333147136 + }, + { + "epoch": 0.71, + "learning_rate": 0.00023691875626880645, + "loss": 1.3479, + "theoretical_loss": 3.4157225595322562, + "tokens_seen": 2333212672 + }, + { + "epoch": 0.71, + "learning_rate": 0.00023690270812437315, + "loss": 1.3003, + "theoretical_loss": 3.4157148651577653, + "tokens_seen": 2333278208 + }, + { + "epoch": 0.71, + "learning_rate": 0.00023688665997993985, + "loss": 1.2481, + "theoretical_loss": 3.4157071710598976, + "tokens_seen": 2333343744 + }, + { + "epoch": 0.71, + "learning_rate": 0.00023687061183550655, + "loss": 1.3576, + "theoretical_loss": 3.4156994772386358, + "tokens_seen": 2333409280 + }, + { + "epoch": 0.71, + "learning_rate": 0.00023685456369107325, + "loss": 1.3845, + "theoretical_loss": 3.415691783693962, + "tokens_seen": 2333474816 + }, + { + "epoch": 0.71, + "learning_rate": 0.00023683851554663995, + "loss": 1.3703, + "theoretical_loss": 3.4156840904258585, + "tokens_seen": 2333540352 + }, + { + "epoch": 0.71, + "learning_rate": 0.00023682246740220665, + "loss": 1.287, + "theoretical_loss": 3.415676397434307, + "tokens_seen": 2333605888 + }, + { + "epoch": 0.71, + "learning_rate": 0.00023680641925777335, + "loss": 1.3397, + "theoretical_loss": 3.4156687047192906, + "tokens_seen": 2333671424 + }, + { + "epoch": 0.71, + "learning_rate": 0.00023679037111334002, + "loss": 1.3474, + "theoretical_loss": 3.4156610122807916, + "tokens_seen": 2333736960 + }, + { + "epoch": 0.71, + "learning_rate": 0.00023677432296890673, + "loss": 1.3048, + "theoretical_loss": 3.4156533201187917, + "tokens_seen": 2333802496 + }, + { + "epoch": 0.71, + "learning_rate": 0.00023675827482447343, + "loss": 1.3217, + "theoretical_loss": 3.415645628233274, + "tokens_seen": 2333868032 + }, + { + "epoch": 0.71, + "learning_rate": 0.00023674222668004013, + "loss": 1.4027, + "theoretical_loss": 3.4156379366242193, + "tokens_seen": 2333933568 + }, + { + "epoch": 0.71, + "learning_rate": 0.00023672617853560683, + "loss": 1.3519, + "theoretical_loss": 3.415630245291612, + "tokens_seen": 2333999104 + }, + { + "epoch": 0.71, + "learning_rate": 0.00023671013039117353, + "loss": 1.2683, + "theoretical_loss": 3.4156225542354326, + "tokens_seen": 2334064640 + }, + { + "epoch": 0.71, + "learning_rate": 0.00023669408224674023, + "loss": 1.354, + "theoretical_loss": 3.4156148634556645, + "tokens_seen": 2334130176 + }, + { + "epoch": 0.71, + "learning_rate": 0.00023667803410230693, + "loss": 1.4097, + "theoretical_loss": 3.4156071729522894, + "tokens_seen": 2334195712 + }, + { + "epoch": 0.71, + "learning_rate": 0.00023666198595787363, + "loss": 1.237, + "theoretical_loss": 3.4155994827252902, + "tokens_seen": 2334261248 + }, + { + "epoch": 0.71, + "learning_rate": 0.00023664593781344033, + "loss": 1.3361, + "theoretical_loss": 3.4155917927746486, + "tokens_seen": 2334326784 + }, + { + "epoch": 0.71, + "learning_rate": 0.00023662988966900703, + "loss": 1.3592, + "theoretical_loss": 3.415584103100347, + "tokens_seen": 2334392320 + }, + { + "epoch": 0.71, + "learning_rate": 0.00023661384152457373, + "loss": 1.4076, + "theoretical_loss": 3.415576413702368, + "tokens_seen": 2334457856 + }, + { + "epoch": 0.71, + "learning_rate": 0.00023659779338014043, + "loss": 1.4186, + "theoretical_loss": 3.415568724580694, + "tokens_seen": 2334523392 + }, + { + "epoch": 0.71, + "learning_rate": 0.00023658174523570713, + "loss": 1.4905, + "theoretical_loss": 3.415561035735307, + "tokens_seen": 2334588928 + }, + { + "epoch": 0.71, + "learning_rate": 0.00023656569709127383, + "loss": 1.3226, + "theoretical_loss": 3.415553347166189, + "tokens_seen": 2334654464 + }, + { + "epoch": 0.71, + "learning_rate": 0.00023654964894684053, + "loss": 1.2816, + "theoretical_loss": 3.4155456588733233, + "tokens_seen": 2334720000 + }, + { + "epoch": 0.71, + "learning_rate": 0.00023653360080240723, + "loss": 1.3272, + "theoretical_loss": 3.4155379708566915, + "tokens_seen": 2334785536 + }, + { + "epoch": 0.71, + "learning_rate": 0.00023651755265797394, + "loss": 1.3775, + "theoretical_loss": 3.4155302831162757, + "tokens_seen": 2334851072 + }, + { + "epoch": 0.71, + "learning_rate": 0.00023650150451354064, + "loss": 1.3554, + "theoretical_loss": 3.415522595652059, + "tokens_seen": 2334916608 + }, + { + "epoch": 0.71, + "learning_rate": 0.00023648545636910734, + "loss": 1.3434, + "theoretical_loss": 3.415514908464023, + "tokens_seen": 2334982144 + }, + { + "epoch": 0.71, + "learning_rate": 0.00023646940822467404, + "loss": 1.1996, + "theoretical_loss": 3.4155072215521507, + "tokens_seen": 2335047680 + }, + { + "epoch": 0.71, + "learning_rate": 0.00023645336008024074, + "loss": 1.3417, + "theoretical_loss": 3.4154995349164237, + "tokens_seen": 2335113216 + }, + { + "epoch": 0.71, + "learning_rate": 0.00023643731193580744, + "loss": 1.3032, + "theoretical_loss": 3.415491848556825, + "tokens_seen": 2335178752 + }, + { + "epoch": 0.71, + "learning_rate": 0.00023642126379137414, + "loss": 1.3011, + "theoretical_loss": 3.4154841624733367, + "tokens_seen": 2335244288 + }, + { + "epoch": 0.71, + "learning_rate": 0.0002364052156469408, + "loss": 1.3456, + "theoretical_loss": 3.415476476665941, + "tokens_seen": 2335309824 + }, + { + "epoch": 0.71, + "learning_rate": 0.0002363891675025075, + "loss": 1.3082, + "theoretical_loss": 3.41546879113462, + "tokens_seen": 2335375360 + }, + { + "epoch": 0.71, + "learning_rate": 0.00023637311935807421, + "loss": 1.4445, + "theoretical_loss": 3.415461105879357, + "tokens_seen": 2335440896 + }, + { + "epoch": 0.71, + "learning_rate": 0.00023635707121364091, + "loss": 1.2645, + "theoretical_loss": 3.415453420900133, + "tokens_seen": 2335506432 + }, + { + "epoch": 0.71, + "learning_rate": 0.00023634102306920762, + "loss": 1.2033, + "theoretical_loss": 3.4154457361969315, + "tokens_seen": 2335571968 + }, + { + "epoch": 0.71, + "learning_rate": 0.00023632497492477432, + "loss": 1.3466, + "theoretical_loss": 3.4154380517697343, + "tokens_seen": 2335637504 + }, + { + "epoch": 0.71, + "learning_rate": 0.00023630892678034102, + "loss": 1.3629, + "theoretical_loss": 3.4154303676185234, + "tokens_seen": 2335703040 + }, + { + "epoch": 0.71, + "learning_rate": 0.00023629287863590772, + "loss": 1.3174, + "theoretical_loss": 3.415422683743282, + "tokens_seen": 2335768576 + }, + { + "epoch": 0.71, + "learning_rate": 0.00023627683049147442, + "loss": 1.2404, + "theoretical_loss": 3.415415000143992, + "tokens_seen": 2335834112 + }, + { + "epoch": 0.71, + "learning_rate": 0.00023626078234704112, + "loss": 1.3522, + "theoretical_loss": 3.415407316820635, + "tokens_seen": 2335899648 + }, + { + "epoch": 0.71, + "learning_rate": 0.00023624473420260782, + "loss": 1.1614, + "theoretical_loss": 3.4153996337731947, + "tokens_seen": 2335965184 + }, + { + "epoch": 0.71, + "learning_rate": 0.00023622868605817452, + "loss": 1.2884, + "theoretical_loss": 3.4153919510016526, + "tokens_seen": 2336030720 + }, + { + "epoch": 0.71, + "learning_rate": 0.00023621263791374122, + "loss": 1.3322, + "theoretical_loss": 3.4153842685059916, + "tokens_seen": 2336096256 + }, + { + "epoch": 0.71, + "learning_rate": 0.00023619658976930792, + "loss": 1.2638, + "theoretical_loss": 3.415376586286194, + "tokens_seen": 2336161792 + }, + { + "epoch": 0.71, + "learning_rate": 0.00023618054162487462, + "loss": 1.303, + "theoretical_loss": 3.415368904342241, + "tokens_seen": 2336227328 + }, + { + "epoch": 0.71, + "learning_rate": 0.00023616449348044132, + "loss": 1.4197, + "theoretical_loss": 3.415361222674117, + "tokens_seen": 2336292864 + }, + { + "epoch": 0.71, + "objective/train/docs_used": 850768, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.262451171875, + "objective/train/theoretical_loss": 3.415353541281802, + "objective/train/tokens_used": 2356818400, + "theoretical_loss": 3.415353541281802, + "tokens_seen": 2336358400 + }, + { + "epoch": 0.71, + "learning_rate": 0.00023614844533600802, + "loss": 1.3591, + "theoretical_loss": 3.415353541281802, + "tokens_seen": 2336358400 + }, + { + "epoch": 0.71, + "learning_rate": 0.00023613239719157472, + "loss": 1.3133, + "theoretical_loss": 3.4153458601652806, + "tokens_seen": 2336423936 + }, + { + "epoch": 0.71, + "learning_rate": 0.00023611634904714142, + "loss": 1.2762, + "theoretical_loss": 3.4153381793245337, + "tokens_seen": 2336489472 + }, + { + "epoch": 0.71, + "learning_rate": 0.00023610030090270812, + "loss": 1.2271, + "theoretical_loss": 3.415330498759544, + "tokens_seen": 2336555008 + }, + { + "epoch": 0.71, + "learning_rate": 0.00023608425275827483, + "loss": 1.3369, + "theoretical_loss": 3.415322818470294, + "tokens_seen": 2336620544 + }, + { + "epoch": 0.71, + "learning_rate": 0.00023606820461384153, + "loss": 1.2755, + "theoretical_loss": 3.4153151384567666, + "tokens_seen": 2336686080 + }, + { + "epoch": 0.71, + "learning_rate": 0.00023605215646940823, + "loss": 1.2996, + "theoretical_loss": 3.415307458718943, + "tokens_seen": 2336751616 + }, + { + "epoch": 0.71, + "learning_rate": 0.00023603610832497493, + "loss": 1.3601, + "theoretical_loss": 3.4152997792568063, + "tokens_seen": 2336817152 + }, + { + "epoch": 0.71, + "learning_rate": 0.00023602006018054166, + "loss": 1.3143, + "theoretical_loss": 3.415292100070339, + "tokens_seen": 2336882688 + }, + { + "epoch": 0.71, + "learning_rate": 0.00023600401203610836, + "loss": 1.3976, + "theoretical_loss": 3.415284421159523, + "tokens_seen": 2336948224 + }, + { + "epoch": 0.71, + "learning_rate": 0.00023598796389167506, + "loss": 1.3023, + "theoretical_loss": 3.415276742524341, + "tokens_seen": 2337013760 + }, + { + "epoch": 0.71, + "learning_rate": 0.00023597191574724176, + "loss": 1.3149, + "theoretical_loss": 3.415269064164775, + "tokens_seen": 2337079296 + }, + { + "epoch": 0.71, + "learning_rate": 0.00023595586760280846, + "loss": 1.3642, + "theoretical_loss": 3.4152613860808083, + "tokens_seen": 2337144832 + }, + { + "epoch": 0.71, + "learning_rate": 0.00023593981945837516, + "loss": 1.3821, + "theoretical_loss": 3.415253708272422, + "tokens_seen": 2337210368 + }, + { + "epoch": 0.71, + "learning_rate": 0.00023592377131394186, + "loss": 1.3282, + "theoretical_loss": 3.4152460307395995, + "tokens_seen": 2337275904 + }, + { + "epoch": 0.71, + "learning_rate": 0.00023590772316950856, + "loss": 1.3824, + "theoretical_loss": 3.415238353482323, + "tokens_seen": 2337341440 + }, + { + "epoch": 0.71, + "learning_rate": 0.00023589167502507526, + "loss": 1.3387, + "theoretical_loss": 3.415230676500574, + "tokens_seen": 2337406976 + }, + { + "epoch": 0.71, + "learning_rate": 0.00023587562688064196, + "loss": 1.2942, + "theoretical_loss": 3.415222999794336, + "tokens_seen": 2337472512 + }, + { + "epoch": 0.71, + "learning_rate": 0.00023585957873620866, + "loss": 1.4584, + "theoretical_loss": 3.415215323363591, + "tokens_seen": 2337538048 + }, + { + "epoch": 0.71, + "learning_rate": 0.00023584353059177536, + "loss": 1.2932, + "theoretical_loss": 3.4152076472083213, + "tokens_seen": 2337603584 + }, + { + "epoch": 0.71, + "learning_rate": 0.00023582748244734206, + "loss": 1.3436, + "theoretical_loss": 3.4151999713285095, + "tokens_seen": 2337669120 + }, + { + "epoch": 0.71, + "learning_rate": 0.00023581143430290876, + "loss": 1.325, + "theoretical_loss": 3.4151922957241374, + "tokens_seen": 2337734656 + }, + { + "epoch": 0.71, + "learning_rate": 0.00023579538615847546, + "loss": 1.2978, + "theoretical_loss": 3.415184620395188, + "tokens_seen": 2337800192 + }, + { + "epoch": 0.71, + "learning_rate": 0.00023577933801404214, + "loss": 1.2017, + "theoretical_loss": 3.415176945341644, + "tokens_seen": 2337865728 + }, + { + "epoch": 0.71, + "learning_rate": 0.00023576328986960884, + "loss": 1.4463, + "theoretical_loss": 3.415169270563487, + "tokens_seen": 2337931264 + }, + { + "epoch": 0.71, + "learning_rate": 0.00023574724172517554, + "loss": 1.3219, + "theoretical_loss": 3.4151615960607, + "tokens_seen": 2337996800 + }, + { + "epoch": 0.71, + "learning_rate": 0.00023573119358074224, + "loss": 1.2309, + "theoretical_loss": 3.4151539218332645, + "tokens_seen": 2338062336 + }, + { + "epoch": 0.71, + "learning_rate": 0.00023571514543630894, + "loss": 1.4685, + "theoretical_loss": 3.4151462478811636, + "tokens_seen": 2338127872 + }, + { + "epoch": 0.71, + "learning_rate": 0.00023569909729187564, + "loss": 1.4386, + "theoretical_loss": 3.4151385742043803, + "tokens_seen": 2338193408 + }, + { + "epoch": 0.71, + "learning_rate": 0.00023568304914744234, + "loss": 1.3008, + "theoretical_loss": 3.415130900802896, + "tokens_seen": 2338258944 + }, + { + "epoch": 0.71, + "learning_rate": 0.00023566700100300904, + "loss": 1.2828, + "theoretical_loss": 3.4151232276766934, + "tokens_seen": 2338324480 + }, + { + "epoch": 0.71, + "learning_rate": 0.00023565095285857574, + "loss": 1.4053, + "theoretical_loss": 3.415115554825755, + "tokens_seen": 2338390016 + }, + { + "epoch": 0.71, + "learning_rate": 0.00023563490471414244, + "loss": 1.2695, + "theoretical_loss": 3.415107882250063, + "tokens_seen": 2338455552 + }, + { + "epoch": 0.71, + "learning_rate": 0.00023561885656970914, + "loss": 1.27, + "theoretical_loss": 3.4151002099496, + "tokens_seen": 2338521088 + }, + { + "epoch": 0.71, + "learning_rate": 0.00023560280842527584, + "loss": 1.3455, + "theoretical_loss": 3.4150925379243486, + "tokens_seen": 2338586624 + }, + { + "epoch": 0.71, + "learning_rate": 0.00023558676028084255, + "loss": 1.2879, + "theoretical_loss": 3.415084866174291, + "tokens_seen": 2338652160 + }, + { + "epoch": 0.71, + "learning_rate": 0.00023557071213640925, + "loss": 1.3082, + "theoretical_loss": 3.4150771946994096, + "tokens_seen": 2338717696 + }, + { + "epoch": 0.71, + "learning_rate": 0.00023555466399197595, + "loss": 1.3354, + "theoretical_loss": 3.415069523499687, + "tokens_seen": 2338783232 + }, + { + "epoch": 0.71, + "learning_rate": 0.00023553861584754265, + "loss": 1.3903, + "theoretical_loss": 3.4150618525751053, + "tokens_seen": 2338848768 + }, + { + "epoch": 0.71, + "learning_rate": 0.00023552256770310935, + "loss": 1.3288, + "theoretical_loss": 3.4150541819256475, + "tokens_seen": 2338914304 + }, + { + "epoch": 0.71, + "learning_rate": 0.00023550651955867605, + "loss": 1.3141, + "theoretical_loss": 3.415046511551295, + "tokens_seen": 2338979840 + }, + { + "epoch": 0.71, + "learning_rate": 0.00023549047141424275, + "loss": 1.39, + "theoretical_loss": 3.415038841452031, + "tokens_seen": 2339045376 + }, + { + "epoch": 0.71, + "learning_rate": 0.00023547442326980945, + "loss": 1.3957, + "theoretical_loss": 3.415031171627838, + "tokens_seen": 2339110912 + }, + { + "epoch": 0.71, + "learning_rate": 0.00023545837512537615, + "loss": 1.3606, + "theoretical_loss": 3.4150235020786983, + "tokens_seen": 2339176448 + }, + { + "epoch": 0.71, + "learning_rate": 0.00023544232698094285, + "loss": 1.3032, + "theoretical_loss": 3.4150158328045936, + "tokens_seen": 2339241984 + }, + { + "epoch": 0.71, + "learning_rate": 0.00023542627883650955, + "loss": 1.3058, + "theoretical_loss": 3.4150081638055076, + "tokens_seen": 2339307520 + }, + { + "epoch": 0.71, + "learning_rate": 0.00023541023069207625, + "loss": 1.3102, + "theoretical_loss": 3.415000495081422, + "tokens_seen": 2339373056 + }, + { + "epoch": 0.71, + "learning_rate": 0.00023539418254764293, + "loss": 1.3182, + "theoretical_loss": 3.414992826632319, + "tokens_seen": 2339438592 + }, + { + "epoch": 0.71, + "learning_rate": 0.00023537813440320963, + "loss": 1.3293, + "theoretical_loss": 3.4149851584581814, + "tokens_seen": 2339504128 + }, + { + "epoch": 0.71, + "learning_rate": 0.00023536208625877633, + "loss": 1.332, + "theoretical_loss": 3.414977490558992, + "tokens_seen": 2339569664 + }, + { + "epoch": 0.71, + "objective/train/docs_used": 851815, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.395216941833496, + "objective/train/theoretical_loss": 3.414969822934733, + "objective/train/tokens_used": 2360095200, + "theoretical_loss": 3.414969822934733, + "tokens_seen": 2339635200 + }, + { + "epoch": 0.71, + "learning_rate": 0.00023534603811434303, + "loss": 1.3013, + "theoretical_loss": 3.414969822934733, + "tokens_seen": 2339635200 + }, + { + "epoch": 0.71, + "learning_rate": 0.00023532998996990973, + "loss": 1.3006, + "theoretical_loss": 3.414962155585386, + "tokens_seen": 2339700736 + }, + { + "epoch": 0.71, + "learning_rate": 0.00023531394182547643, + "loss": 1.2729, + "theoretical_loss": 3.4149544885109346, + "tokens_seen": 2339766272 + }, + { + "epoch": 0.71, + "learning_rate": 0.00023529789368104313, + "loss": 1.3162, + "theoretical_loss": 3.4149468217113608, + "tokens_seen": 2339831808 + }, + { + "epoch": 0.71, + "learning_rate": 0.00023528184553660983, + "loss": 1.307, + "theoretical_loss": 3.4149391551866466, + "tokens_seen": 2339897344 + }, + { + "epoch": 0.71, + "learning_rate": 0.00023526579739217653, + "loss": 1.4032, + "theoretical_loss": 3.4149314889367752, + "tokens_seen": 2339962880 + }, + { + "epoch": 0.71, + "learning_rate": 0.00023524974924774323, + "loss": 1.3517, + "theoretical_loss": 3.414923822961729, + "tokens_seen": 2340028416 + }, + { + "epoch": 0.71, + "learning_rate": 0.00023523370110330993, + "loss": 1.3118, + "theoretical_loss": 3.4149161572614894, + "tokens_seen": 2340093952 + }, + { + "epoch": 0.71, + "learning_rate": 0.00023521765295887663, + "loss": 1.2741, + "theoretical_loss": 3.4149084918360404, + "tokens_seen": 2340159488 + }, + { + "epoch": 0.71, + "learning_rate": 0.00023520160481444333, + "loss": 1.3466, + "theoretical_loss": 3.414900826685363, + "tokens_seen": 2340225024 + }, + { + "epoch": 0.71, + "learning_rate": 0.00023518555667001003, + "loss": 1.2869, + "theoretical_loss": 3.414893161809441, + "tokens_seen": 2340290560 + }, + { + "epoch": 0.71, + "learning_rate": 0.00023516950852557673, + "loss": 1.2912, + "theoretical_loss": 3.414885497208256, + "tokens_seen": 2340356096 + }, + { + "epoch": 0.71, + "learning_rate": 0.00023515346038114344, + "loss": 1.2959, + "theoretical_loss": 3.4148778328817904, + "tokens_seen": 2340421632 + }, + { + "epoch": 0.71, + "learning_rate": 0.00023513741223671014, + "loss": 1.2858, + "theoretical_loss": 3.414870168830027, + "tokens_seen": 2340487168 + }, + { + "epoch": 0.71, + "learning_rate": 0.00023512136409227684, + "loss": 1.3694, + "theoretical_loss": 3.414862505052948, + "tokens_seen": 2340552704 + }, + { + "epoch": 0.71, + "learning_rate": 0.00023510531594784354, + "loss": 1.2054, + "theoretical_loss": 3.414854841550537, + "tokens_seen": 2340618240 + }, + { + "epoch": 0.71, + "learning_rate": 0.00023508926780341024, + "loss": 1.3391, + "theoretical_loss": 3.4148471783227743, + "tokens_seen": 2340683776 + }, + { + "epoch": 0.71, + "learning_rate": 0.00023507321965897694, + "loss": 1.3317, + "theoretical_loss": 3.4148395153696445, + "tokens_seen": 2340749312 + }, + { + "epoch": 0.71, + "learning_rate": 0.00023505717151454364, + "loss": 1.3509, + "theoretical_loss": 3.4148318526911288, + "tokens_seen": 2340814848 + }, + { + "epoch": 0.71, + "learning_rate": 0.00023504112337011034, + "loss": 1.3662, + "theoretical_loss": 3.41482419028721, + "tokens_seen": 2340880384 + }, + { + "epoch": 0.71, + "learning_rate": 0.00023502507522567704, + "loss": 1.2322, + "theoretical_loss": 3.4148165281578704, + "tokens_seen": 2340945920 + }, + { + "epoch": 0.71, + "learning_rate": 0.00023500902708124371, + "loss": 1.312, + "theoretical_loss": 3.414808866303093, + "tokens_seen": 2341011456 + }, + { + "epoch": 0.71, + "learning_rate": 0.00023499297893681041, + "loss": 1.3336, + "theoretical_loss": 3.4148012047228598, + "tokens_seen": 2341076992 + }, + { + "epoch": 0.71, + "learning_rate": 0.00023497693079237711, + "loss": 1.2985, + "theoretical_loss": 3.414793543417153, + "tokens_seen": 2341142528 + }, + { + "epoch": 0.71, + "learning_rate": 0.00023496088264794382, + "loss": 1.3175, + "theoretical_loss": 3.414785882385956, + "tokens_seen": 2341208064 + }, + { + "epoch": 0.71, + "learning_rate": 0.00023494483450351052, + "loss": 1.3293, + "theoretical_loss": 3.414778221629251, + "tokens_seen": 2341273600 + }, + { + "epoch": 0.71, + "learning_rate": 0.00023492878635907722, + "loss": 1.3812, + "theoretical_loss": 3.41477056114702, + "tokens_seen": 2341339136 + }, + { + "epoch": 0.71, + "learning_rate": 0.00023491273821464392, + "loss": 1.3401, + "theoretical_loss": 3.4147629009392455, + "tokens_seen": 2341404672 + }, + { + "epoch": 0.71, + "learning_rate": 0.00023489669007021062, + "loss": 1.3597, + "theoretical_loss": 3.4147552410059103, + "tokens_seen": 2341470208 + }, + { + "epoch": 0.71, + "learning_rate": 0.00023488064192577732, + "loss": 1.3374, + "theoretical_loss": 3.414747581346997, + "tokens_seen": 2341535744 + }, + { + "epoch": 0.71, + "learning_rate": 0.00023486459378134402, + "loss": 1.3245, + "theoretical_loss": 3.4147399219624877, + "tokens_seen": 2341601280 + }, + { + "epoch": 0.71, + "learning_rate": 0.00023484854563691072, + "loss": 1.331, + "theoretical_loss": 3.4147322628523655, + "tokens_seen": 2341666816 + }, + { + "epoch": 0.71, + "learning_rate": 0.00023483249749247747, + "loss": 1.3021, + "theoretical_loss": 3.414724604016612, + "tokens_seen": 2341732352 + }, + { + "epoch": 0.71, + "learning_rate": 0.00023481644934804418, + "loss": 1.394, + "theoretical_loss": 3.4147169454552104, + "tokens_seen": 2341797888 + }, + { + "epoch": 0.71, + "learning_rate": 0.00023480040120361088, + "loss": 1.3529, + "theoretical_loss": 3.414709287168143, + "tokens_seen": 2341863424 + }, + { + "epoch": 0.71, + "learning_rate": 0.00023478435305917758, + "loss": 1.2834, + "theoretical_loss": 3.4147016291553918, + "tokens_seen": 2341928960 + }, + { + "epoch": 0.71, + "learning_rate": 0.00023476830491474425, + "loss": 1.43, + "theoretical_loss": 3.4146939714169404, + "tokens_seen": 2341994496 + }, + { + "epoch": 0.71, + "learning_rate": 0.00023475225677031095, + "loss": 1.4344, + "theoretical_loss": 3.4146863139527706, + "tokens_seen": 2342060032 + }, + { + "epoch": 0.71, + "learning_rate": 0.00023473620862587765, + "loss": 1.2955, + "theoretical_loss": 3.414678656762865, + "tokens_seen": 2342125568 + }, + { + "epoch": 0.71, + "learning_rate": 0.00023472016048144435, + "loss": 1.3216, + "theoretical_loss": 3.4146709998472056, + "tokens_seen": 2342191104 + }, + { + "epoch": 0.71, + "learning_rate": 0.00023470411233701105, + "loss": 1.345, + "theoretical_loss": 3.414663343205776, + "tokens_seen": 2342256640 + }, + { + "epoch": 0.71, + "learning_rate": 0.00023468806419257775, + "loss": 1.2667, + "theoretical_loss": 3.4146556868385574, + "tokens_seen": 2342322176 + }, + { + "epoch": 0.71, + "learning_rate": 0.00023467201604814445, + "loss": 1.3204, + "theoretical_loss": 3.4146480307455334, + "tokens_seen": 2342387712 + }, + { + "epoch": 0.71, + "learning_rate": 0.00023465596790371115, + "loss": 1.388, + "theoretical_loss": 3.414640374926686, + "tokens_seen": 2342453248 + }, + { + "epoch": 0.71, + "learning_rate": 0.00023463991975927786, + "loss": 1.3907, + "theoretical_loss": 3.4146327193819976, + "tokens_seen": 2342518784 + }, + { + "epoch": 0.71, + "learning_rate": 0.00023462387161484456, + "loss": 1.3002, + "theoretical_loss": 3.4146250641114513, + "tokens_seen": 2342584320 + }, + { + "epoch": 0.71, + "learning_rate": 0.00023460782347041126, + "loss": 1.3298, + "theoretical_loss": 3.414617409115029, + "tokens_seen": 2342649856 + }, + { + "epoch": 0.71, + "learning_rate": 0.00023459177532597796, + "loss": 1.3676, + "theoretical_loss": 3.414609754392714, + "tokens_seen": 2342715392 + }, + { + "epoch": 0.71, + "learning_rate": 0.00023457572718154466, + "loss": 1.3401, + "theoretical_loss": 3.4146020999444877, + "tokens_seen": 2342780928 + }, + { + "epoch": 0.71, + "learning_rate": 0.00023455967903711136, + "loss": 1.295, + "theoretical_loss": 3.414594445770333, + "tokens_seen": 2342846464 + }, + { + "epoch": 0.71, + "objective/train/docs_used": 852909, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.181896448135376, + "objective/train/theoretical_loss": 3.4145867918702333, + "objective/train/tokens_used": 2363372000, + "theoretical_loss": 3.4145867918702333, + "tokens_seen": 2342912000 + }, + { + "epoch": 0.71, + "learning_rate": 0.00023454363089267806, + "loss": 1.2618, + "theoretical_loss": 3.4145867918702333, + "tokens_seen": 2342912000 + }, + { + "epoch": 0.71, + "learning_rate": 0.00023452758274824476, + "loss": 1.3502, + "theoretical_loss": 3.41457913824417, + "tokens_seen": 2342977536 + }, + { + "epoch": 0.71, + "learning_rate": 0.00023451153460381146, + "loss": 1.4691, + "theoretical_loss": 3.414571484892126, + "tokens_seen": 2343043072 + }, + { + "epoch": 0.71, + "learning_rate": 0.00023449548645937816, + "loss": 1.3213, + "theoretical_loss": 3.414563831814084, + "tokens_seen": 2343108608 + }, + { + "epoch": 0.71, + "learning_rate": 0.00023447943831494486, + "loss": 1.302, + "theoretical_loss": 3.4145561790100265, + "tokens_seen": 2343174144 + }, + { + "epoch": 0.71, + "learning_rate": 0.00023446339017051156, + "loss": 1.3314, + "theoretical_loss": 3.414548526479936, + "tokens_seen": 2343239680 + }, + { + "epoch": 0.71, + "learning_rate": 0.00023444734202607826, + "loss": 1.3333, + "theoretical_loss": 3.414540874223795, + "tokens_seen": 2343305216 + }, + { + "epoch": 0.71, + "learning_rate": 0.00023443129388164496, + "loss": 1.3716, + "theoretical_loss": 3.4145332222415856, + "tokens_seen": 2343370752 + }, + { + "epoch": 0.71, + "learning_rate": 0.00023441524573721166, + "loss": 1.3741, + "theoretical_loss": 3.414525570533291, + "tokens_seen": 2343436288 + }, + { + "epoch": 0.71, + "learning_rate": 0.00023439919759277836, + "loss": 1.214, + "theoretical_loss": 3.4145179190988935, + "tokens_seen": 2343501824 + }, + { + "epoch": 0.71, + "learning_rate": 0.00023438314944834504, + "loss": 1.2407, + "theoretical_loss": 3.4145102679383754, + "tokens_seen": 2343567360 + }, + { + "epoch": 0.71, + "learning_rate": 0.00023436710130391174, + "loss": 1.3012, + "theoretical_loss": 3.4145026170517196, + "tokens_seen": 2343632896 + }, + { + "epoch": 0.71, + "learning_rate": 0.00023435105315947844, + "loss": 1.3284, + "theoretical_loss": 3.4144949664389084, + "tokens_seen": 2343698432 + }, + { + "epoch": 0.71, + "learning_rate": 0.00023433500501504514, + "loss": 1.3803, + "theoretical_loss": 3.4144873160999243, + "tokens_seen": 2343763968 + }, + { + "epoch": 0.71, + "learning_rate": 0.00023431895687061184, + "loss": 1.2979, + "theoretical_loss": 3.4144796660347505, + "tokens_seen": 2343829504 + }, + { + "epoch": 0.71, + "learning_rate": 0.00023430290872617854, + "loss": 1.2474, + "theoretical_loss": 3.4144720162433684, + "tokens_seen": 2343895040 + }, + { + "epoch": 0.71, + "learning_rate": 0.00023428686058174524, + "loss": 1.2757, + "theoretical_loss": 3.4144643667257615, + "tokens_seen": 2343960576 + }, + { + "epoch": 0.71, + "learning_rate": 0.00023427081243731194, + "loss": 1.2942, + "theoretical_loss": 3.414456717481912, + "tokens_seen": 2344026112 + }, + { + "epoch": 0.71, + "learning_rate": 0.00023425476429287864, + "loss": 1.2858, + "theoretical_loss": 3.4144490685118023, + "tokens_seen": 2344091648 + }, + { + "epoch": 0.71, + "learning_rate": 0.00023423871614844534, + "loss": 1.1721, + "theoretical_loss": 3.4144414198154154, + "tokens_seen": 2344157184 + }, + { + "epoch": 0.71, + "learning_rate": 0.00023422266800401204, + "loss": 1.3759, + "theoretical_loss": 3.414433771392733, + "tokens_seen": 2344222720 + }, + { + "epoch": 0.71, + "learning_rate": 0.00023420661985957875, + "loss": 1.336, + "theoretical_loss": 3.4144261232437385, + "tokens_seen": 2344288256 + }, + { + "epoch": 0.71, + "learning_rate": 0.00023419057171514545, + "loss": 1.3568, + "theoretical_loss": 3.4144184753684144, + "tokens_seen": 2344353792 + }, + { + "epoch": 0.71, + "learning_rate": 0.00023417452357071215, + "loss": 1.2581, + "theoretical_loss": 3.4144108277667424, + "tokens_seen": 2344419328 + }, + { + "epoch": 0.71, + "learning_rate": 0.00023415847542627885, + "loss": 1.3291, + "theoretical_loss": 3.4144031804387067, + "tokens_seen": 2344484864 + }, + { + "epoch": 0.71, + "learning_rate": 0.00023414242728184555, + "loss": 1.3358, + "theoretical_loss": 3.4143955333842877, + "tokens_seen": 2344550400 + }, + { + "epoch": 0.71, + "learning_rate": 0.00023412637913741225, + "loss": 1.403, + "theoretical_loss": 3.41438788660347, + "tokens_seen": 2344615936 + }, + { + "epoch": 0.71, + "learning_rate": 0.00023411033099297895, + "loss": 1.3728, + "theoretical_loss": 3.414380240096235, + "tokens_seen": 2344681472 + }, + { + "epoch": 0.71, + "learning_rate": 0.00023409428284854565, + "loss": 1.3746, + "theoretical_loss": 3.4143725938625655, + "tokens_seen": 2344747008 + }, + { + "epoch": 0.71, + "learning_rate": 0.00023407823470411235, + "loss": 1.3303, + "theoretical_loss": 3.414364947902444, + "tokens_seen": 2344812544 + }, + { + "epoch": 0.71, + "learning_rate": 0.00023406218655967905, + "loss": 1.3038, + "theoretical_loss": 3.414357302215853, + "tokens_seen": 2344878080 + }, + { + "epoch": 0.71, + "learning_rate": 0.00023404613841524575, + "loss": 1.4182, + "theoretical_loss": 3.4143496568027754, + "tokens_seen": 2344943616 + }, + { + "epoch": 0.71, + "learning_rate": 0.00023403009027081245, + "loss": 1.3909, + "theoretical_loss": 3.414342011663194, + "tokens_seen": 2345009152 + }, + { + "epoch": 0.71, + "learning_rate": 0.00023401404212637915, + "loss": 1.3291, + "theoretical_loss": 3.4143343667970907, + "tokens_seen": 2345074688 + }, + { + "epoch": 0.71, + "learning_rate": 0.00023399799398194583, + "loss": 1.3615, + "theoretical_loss": 3.414326722204448, + "tokens_seen": 2345140224 + }, + { + "epoch": 0.71, + "learning_rate": 0.00023398194583751253, + "loss": 1.366, + "theoretical_loss": 3.4143190778852492, + "tokens_seen": 2345205760 + }, + { + "epoch": 0.71, + "learning_rate": 0.00023396589769307923, + "loss": 1.3667, + "theoretical_loss": 3.414311433839477, + "tokens_seen": 2345271296 + }, + { + "epoch": 0.71, + "learning_rate": 0.00023394984954864593, + "loss": 1.2362, + "theoretical_loss": 3.4143037900671125, + "tokens_seen": 2345336832 + }, + { + "epoch": 0.71, + "learning_rate": 0.00023393380140421263, + "loss": 1.2011, + "theoretical_loss": 3.41429614656814, + "tokens_seen": 2345402368 + }, + { + "epoch": 0.71, + "learning_rate": 0.00023391775325977933, + "loss": 1.3498, + "theoretical_loss": 3.414288503342541, + "tokens_seen": 2345467904 + }, + { + "epoch": 0.71, + "learning_rate": 0.00023390170511534603, + "loss": 1.3046, + "theoretical_loss": 3.4142808603902988, + "tokens_seen": 2345533440 + }, + { + "epoch": 0.71, + "learning_rate": 0.00023388565697091273, + "loss": 1.3172, + "theoretical_loss": 3.4142732177113952, + "tokens_seen": 2345598976 + }, + { + "epoch": 0.71, + "learning_rate": 0.00023386960882647943, + "loss": 1.2248, + "theoretical_loss": 3.4142655753058135, + "tokens_seen": 2345664512 + }, + { + "epoch": 0.71, + "learning_rate": 0.00023385356068204613, + "loss": 1.2915, + "theoretical_loss": 3.414257933173536, + "tokens_seen": 2345730048 + }, + { + "epoch": 0.71, + "learning_rate": 0.00023383751253761283, + "loss": 1.2727, + "theoretical_loss": 3.4142502913145454, + "tokens_seen": 2345795584 + }, + { + "epoch": 0.71, + "learning_rate": 0.00023382146439317953, + "loss": 1.2591, + "theoretical_loss": 3.414242649728824, + "tokens_seen": 2345861120 + }, + { + "epoch": 0.71, + "learning_rate": 0.00023380541624874623, + "loss": 1.371, + "theoretical_loss": 3.4142350084163544, + "tokens_seen": 2345926656 + }, + { + "epoch": 0.71, + "learning_rate": 0.00023378936810431293, + "loss": 1.3286, + "theoretical_loss": 3.4142273673771197, + "tokens_seen": 2345992192 + }, + { + "epoch": 0.71, + "learning_rate": 0.00023377331995987964, + "loss": 1.3142, + "theoretical_loss": 3.414219726611102, + "tokens_seen": 2346057728 + }, + { + "epoch": 0.71, + "learning_rate": 0.00023375727181544634, + "loss": 1.3056, + "theoretical_loss": 3.414212086118284, + "tokens_seen": 2346123264 + }, + { + "epoch": 0.71, + "objective/train/docs_used": 854185, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.40086030960083, + "objective/train/theoretical_loss": 3.414204445898649, + "objective/train/tokens_used": 2366648800, + "theoretical_loss": 3.414204445898649, + "tokens_seen": 2346188800 + }, + { + "epoch": 0.71, + "learning_rate": 0.00023374122367101304, + "loss": 1.4091, + "theoretical_loss": 3.414204445898649, + "tokens_seen": 2346188800 + }, + { + "epoch": 0.71, + "learning_rate": 0.00023372517552657974, + "loss": 1.3472, + "theoretical_loss": 3.4141968059521783, + "tokens_seen": 2346254336 + }, + { + "epoch": 0.71, + "learning_rate": 0.00023370912738214644, + "loss": 1.3637, + "theoretical_loss": 3.4141891662788555, + "tokens_seen": 2346319872 + }, + { + "epoch": 0.71, + "learning_rate": 0.00023369307923771314, + "loss": 1.2607, + "theoretical_loss": 3.4141815268786626, + "tokens_seen": 2346385408 + }, + { + "epoch": 0.71, + "learning_rate": 0.00023367703109327984, + "loss": 1.304, + "theoretical_loss": 3.414173887751583, + "tokens_seen": 2346450944 + }, + { + "epoch": 0.71, + "learning_rate": 0.00023366098294884654, + "loss": 1.3008, + "theoretical_loss": 3.4141662488975983, + "tokens_seen": 2346516480 + }, + { + "epoch": 0.71, + "learning_rate": 0.00023364493480441324, + "loss": 1.2652, + "theoretical_loss": 3.414158610316692, + "tokens_seen": 2346582016 + }, + { + "epoch": 0.71, + "learning_rate": 0.00023362888665997997, + "loss": 1.4242, + "theoretical_loss": 3.414150972008846, + "tokens_seen": 2346647552 + }, + { + "epoch": 0.71, + "learning_rate": 0.00023361283851554667, + "loss": 1.3617, + "theoretical_loss": 3.4141433339740432, + "tokens_seen": 2346713088 + }, + { + "epoch": 0.71, + "learning_rate": 0.00023359679037111337, + "loss": 1.299, + "theoretical_loss": 3.4141356962122664, + "tokens_seen": 2346778624 + }, + { + "epoch": 0.71, + "learning_rate": 0.00023358074222668007, + "loss": 1.231, + "theoretical_loss": 3.4141280587234983, + "tokens_seen": 2346844160 + }, + { + "epoch": 0.71, + "learning_rate": 0.00023356469408224677, + "loss": 1.3782, + "theoretical_loss": 3.4141204215077208, + "tokens_seen": 2346909696 + }, + { + "epoch": 0.71, + "learning_rate": 0.00023354864593781347, + "loss": 1.3599, + "theoretical_loss": 3.4141127845649173, + "tokens_seen": 2346975232 + }, + { + "epoch": 0.71, + "learning_rate": 0.00023353259779338017, + "loss": 1.2717, + "theoretical_loss": 3.4141051478950706, + "tokens_seen": 2347040768 + }, + { + "epoch": 0.71, + "learning_rate": 0.00023351654964894687, + "loss": 1.2615, + "theoretical_loss": 3.414097511498162, + "tokens_seen": 2347106304 + }, + { + "epoch": 0.71, + "learning_rate": 0.00023350050150451357, + "loss": 1.3549, + "theoretical_loss": 3.4140898753741755, + "tokens_seen": 2347171840 + }, + { + "epoch": 0.71, + "learning_rate": 0.00023348445336008027, + "loss": 1.3455, + "theoretical_loss": 3.414082239523093, + "tokens_seen": 2347237376 + }, + { + "epoch": 0.71, + "learning_rate": 0.00023346840521564697, + "loss": 1.2284, + "theoretical_loss": 3.4140746039448975, + "tokens_seen": 2347302912 + }, + { + "epoch": 0.71, + "learning_rate": 0.00023345235707121368, + "loss": 1.2341, + "theoretical_loss": 3.4140669686395713, + "tokens_seen": 2347368448 + }, + { + "epoch": 0.71, + "learning_rate": 0.00023343630892678038, + "loss": 1.2828, + "theoretical_loss": 3.4140593336070975, + "tokens_seen": 2347433984 + }, + { + "epoch": 0.71, + "learning_rate": 0.00023342026078234708, + "loss": 1.3156, + "theoretical_loss": 3.414051698847458, + "tokens_seen": 2347499520 + }, + { + "epoch": 0.71, + "learning_rate": 0.00023340421263791378, + "loss": 1.3618, + "theoretical_loss": 3.414044064360636, + "tokens_seen": 2347565056 + }, + { + "epoch": 0.71, + "learning_rate": 0.00023338816449348048, + "loss": 1.3243, + "theoretical_loss": 3.4140364301466137, + "tokens_seen": 2347630592 + }, + { + "epoch": 0.71, + "learning_rate": 0.00023337211634904718, + "loss": 1.3216, + "theoretical_loss": 3.4140287962053746, + "tokens_seen": 2347696128 + }, + { + "epoch": 0.71, + "learning_rate": 0.00023335606820461385, + "loss": 1.2282, + "theoretical_loss": 3.4140211625369004, + "tokens_seen": 2347761664 + }, + { + "epoch": 0.71, + "learning_rate": 0.00023334002006018055, + "loss": 1.3006, + "theoretical_loss": 3.414013529141174, + "tokens_seen": 2347827200 + }, + { + "epoch": 0.71, + "learning_rate": 0.00023332397191574725, + "loss": 1.3405, + "theoretical_loss": 3.4140058960181787, + "tokens_seen": 2347892736 + }, + { + "epoch": 0.71, + "learning_rate": 0.00023330792377131395, + "loss": 1.2528, + "theoretical_loss": 3.4139982631678962, + "tokens_seen": 2347958272 + }, + { + "epoch": 0.71, + "learning_rate": 0.00023329187562688065, + "loss": 1.2926, + "theoretical_loss": 3.41399063059031, + "tokens_seen": 2348023808 + }, + { + "epoch": 0.71, + "learning_rate": 0.00023327582748244736, + "loss": 1.3103, + "theoretical_loss": 3.4139829982854017, + "tokens_seen": 2348089344 + }, + { + "epoch": 0.71, + "learning_rate": 0.00023325977933801406, + "loss": 1.3226, + "theoretical_loss": 3.4139753662531547, + "tokens_seen": 2348154880 + }, + { + "epoch": 0.71, + "learning_rate": 0.00023324373119358076, + "loss": 1.2683, + "theoretical_loss": 3.4139677344935517, + "tokens_seen": 2348220416 + }, + { + "epoch": 0.71, + "learning_rate": 0.00023322768304914746, + "loss": 1.2839, + "theoretical_loss": 3.4139601030065747, + "tokens_seen": 2348285952 + }, + { + "epoch": 0.71, + "learning_rate": 0.00023321163490471416, + "loss": 1.2533, + "theoretical_loss": 3.413952471792207, + "tokens_seen": 2348351488 + }, + { + "epoch": 0.71, + "learning_rate": 0.00023319558676028086, + "loss": 1.2814, + "theoretical_loss": 3.4139448408504314, + "tokens_seen": 2348417024 + }, + { + "epoch": 0.71, + "learning_rate": 0.00023317953861584756, + "loss": 1.2662, + "theoretical_loss": 3.4139372101812295, + "tokens_seen": 2348482560 + }, + { + "epoch": 0.71, + "learning_rate": 0.00023316349047141426, + "loss": 1.3521, + "theoretical_loss": 3.4139295797845852, + "tokens_seen": 2348548096 + }, + { + "epoch": 0.71, + "learning_rate": 0.00023314744232698096, + "loss": 1.2745, + "theoretical_loss": 3.4139219496604802, + "tokens_seen": 2348613632 + }, + { + "epoch": 0.71, + "learning_rate": 0.00023313139418254766, + "loss": 1.382, + "theoretical_loss": 3.413914319808898, + "tokens_seen": 2348679168 + }, + { + "epoch": 0.71, + "learning_rate": 0.00023311534603811436, + "loss": 1.3624, + "theoretical_loss": 3.4139066902298207, + "tokens_seen": 2348744704 + }, + { + "epoch": 0.71, + "learning_rate": 0.00023309929789368106, + "loss": 1.3268, + "theoretical_loss": 3.413899060923231, + "tokens_seen": 2348810240 + }, + { + "epoch": 0.71, + "learning_rate": 0.00023308324974924776, + "loss": 1.1776, + "theoretical_loss": 3.4138914318891116, + "tokens_seen": 2348875776 + }, + { + "epoch": 0.71, + "learning_rate": 0.00023306720160481446, + "loss": 1.2884, + "theoretical_loss": 3.4138838031274457, + "tokens_seen": 2348941312 + }, + { + "epoch": 0.71, + "learning_rate": 0.00023305115346038116, + "loss": 1.3831, + "theoretical_loss": 3.413876174638215, + "tokens_seen": 2349006848 + }, + { + "epoch": 0.71, + "learning_rate": 0.00023303510531594786, + "loss": 1.263, + "theoretical_loss": 3.413868546421403, + "tokens_seen": 2349072384 + }, + { + "epoch": 0.71, + "learning_rate": 0.00023301905717151457, + "loss": 1.3186, + "theoretical_loss": 3.413860918476992, + "tokens_seen": 2349137920 + }, + { + "epoch": 0.71, + "learning_rate": 0.00023300300902708127, + "loss": 1.2248, + "theoretical_loss": 3.4138532908049646, + "tokens_seen": 2349203456 + }, + { + "epoch": 0.71, + "learning_rate": 0.00023298696088264797, + "loss": 1.3719, + "theoretical_loss": 3.4138456634053034, + "tokens_seen": 2349268992 + }, + { + "epoch": 0.71, + "learning_rate": 0.00023297091273821464, + "loss": 1.4242, + "theoretical_loss": 3.413838036277992, + "tokens_seen": 2349334528 + }, + { + "epoch": 0.71, + "learning_rate": 0.00023295486459378134, + "loss": 1.2192, + "theoretical_loss": 3.4138304094230114, + "tokens_seen": 2349400064 + }, + { + "epoch": 0.71, + "objective/train/docs_used": 855374, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.4039301872253418, + "objective/train/theoretical_loss": 3.413822782840346, + "objective/train/tokens_used": 2369925600, + "theoretical_loss": 3.413822782840346, + "tokens_seen": 2349465600 + }, + { + "epoch": 0.71, + "learning_rate": 0.00023293881644934804, + "loss": 1.3742, + "theoretical_loss": 3.413822782840346, + "tokens_seen": 2349465600 + }, + { + "epoch": 0.71, + "learning_rate": 0.00023292276830491474, + "loss": 1.3249, + "theoretical_loss": 3.4138151565299775, + "tokens_seen": 2349531136 + }, + { + "epoch": 0.71, + "learning_rate": 0.00023290672016048144, + "loss": 1.3897, + "theoretical_loss": 3.4138075304918885, + "tokens_seen": 2349596672 + }, + { + "epoch": 0.71, + "learning_rate": 0.00023289067201604814, + "loss": 1.3323, + "theoretical_loss": 3.413799904726062, + "tokens_seen": 2349662208 + }, + { + "epoch": 0.71, + "learning_rate": 0.00023287462387161484, + "loss": 1.3961, + "theoretical_loss": 3.413792279232481, + "tokens_seen": 2349727744 + }, + { + "epoch": 0.71, + "learning_rate": 0.00023285857572718154, + "loss": 1.3669, + "theoretical_loss": 3.4137846540111276, + "tokens_seen": 2349793280 + }, + { + "epoch": 0.71, + "learning_rate": 0.00023284252758274825, + "loss": 1.4001, + "theoretical_loss": 3.413777029061985, + "tokens_seen": 2349858816 + }, + { + "epoch": 0.71, + "learning_rate": 0.00023282647943831495, + "loss": 1.3605, + "theoretical_loss": 3.4137694043850355, + "tokens_seen": 2349924352 + }, + { + "epoch": 0.71, + "learning_rate": 0.00023281043129388165, + "loss": 1.3401, + "theoretical_loss": 3.4137617799802618, + "tokens_seen": 2349989888 + }, + { + "epoch": 0.71, + "learning_rate": 0.00023279438314944835, + "loss": 1.3109, + "theoretical_loss": 3.413754155847647, + "tokens_seen": 2350055424 + }, + { + "epoch": 0.71, + "learning_rate": 0.00023277833500501505, + "loss": 1.2723, + "theoretical_loss": 3.413746531987173, + "tokens_seen": 2350120960 + }, + { + "epoch": 0.71, + "learning_rate": 0.00023276228686058175, + "loss": 1.3074, + "theoretical_loss": 3.4137389083988237, + "tokens_seen": 2350186496 + }, + { + "epoch": 0.71, + "learning_rate": 0.00023274623871614845, + "loss": 1.3216, + "theoretical_loss": 3.413731285082581, + "tokens_seen": 2350252032 + }, + { + "epoch": 0.71, + "learning_rate": 0.00023273019057171515, + "loss": 1.3202, + "theoretical_loss": 3.4137236620384277, + "tokens_seen": 2350317568 + }, + { + "epoch": 0.71, + "learning_rate": 0.00023271414242728185, + "loss": 1.1909, + "theoretical_loss": 3.413716039266346, + "tokens_seen": 2350383104 + }, + { + "epoch": 0.71, + "learning_rate": 0.00023269809428284855, + "loss": 1.3374, + "theoretical_loss": 3.4137084167663194, + "tokens_seen": 2350448640 + }, + { + "epoch": 0.71, + "learning_rate": 0.00023268204613841525, + "loss": 1.2467, + "theoretical_loss": 3.4137007945383306, + "tokens_seen": 2350514176 + }, + { + "epoch": 0.71, + "learning_rate": 0.00023266599799398195, + "loss": 1.366, + "theoretical_loss": 3.4136931725823616, + "tokens_seen": 2350579712 + }, + { + "epoch": 0.71, + "learning_rate": 0.00023264994984954865, + "loss": 1.3529, + "theoretical_loss": 3.413685550898396, + "tokens_seen": 2350645248 + }, + { + "epoch": 0.71, + "learning_rate": 0.00023263390170511535, + "loss": 1.2644, + "theoretical_loss": 3.4136779294864157, + "tokens_seen": 2350710784 + }, + { + "epoch": 0.71, + "learning_rate": 0.00023261785356068205, + "loss": 1.3415, + "theoretical_loss": 3.413670308346404, + "tokens_seen": 2350776320 + }, + { + "epoch": 0.71, + "learning_rate": 0.00023260180541624875, + "loss": 1.2974, + "theoretical_loss": 3.413662687478343, + "tokens_seen": 2350841856 + }, + { + "epoch": 0.71, + "learning_rate": 0.00023258575727181543, + "loss": 1.3784, + "theoretical_loss": 3.413655066882216, + "tokens_seen": 2350907392 + }, + { + "epoch": 0.71, + "learning_rate": 0.00023256970912738213, + "loss": 1.4158, + "theoretical_loss": 3.413647446558006, + "tokens_seen": 2350972928 + }, + { + "epoch": 0.71, + "learning_rate": 0.00023255366098294883, + "loss": 1.1836, + "theoretical_loss": 3.4136398265056944, + "tokens_seen": 2351038464 + }, + { + "epoch": 0.71, + "learning_rate": 0.00023253761283851553, + "loss": 1.3169, + "theoretical_loss": 3.413632206725265, + "tokens_seen": 2351104000 + }, + { + "epoch": 0.71, + "learning_rate": 0.00023252156469408223, + "loss": 1.2518, + "theoretical_loss": 3.4136245872167006, + "tokens_seen": 2351169536 + }, + { + "epoch": 0.71, + "learning_rate": 0.00023250551654964893, + "loss": 1.3413, + "theoretical_loss": 3.4136169679799835, + "tokens_seen": 2351235072 + }, + { + "epoch": 0.71, + "learning_rate": 0.00023248946840521563, + "loss": 1.3398, + "theoretical_loss": 3.413609349015096, + "tokens_seen": 2351300608 + }, + { + "epoch": 0.71, + "learning_rate": 0.00023247342026078233, + "loss": 1.263, + "theoretical_loss": 3.4136017303220223, + "tokens_seen": 2351366144 + }, + { + "epoch": 0.71, + "learning_rate": 0.00023245737211634903, + "loss": 1.326, + "theoretical_loss": 3.4135941119007436, + "tokens_seen": 2351431680 + }, + { + "epoch": 0.71, + "learning_rate": 0.00023244132397191573, + "loss": 1.4331, + "theoretical_loss": 3.413586493751243, + "tokens_seen": 2351497216 + }, + { + "epoch": 0.71, + "learning_rate": 0.0002324252758274825, + "loss": 1.3347, + "theoretical_loss": 3.4135788758735037, + "tokens_seen": 2351562752 + }, + { + "epoch": 0.71, + "learning_rate": 0.0002324092276830492, + "loss": 1.1963, + "theoretical_loss": 3.413571258267508, + "tokens_seen": 2351628288 + }, + { + "epoch": 0.71, + "learning_rate": 0.0002323931795386159, + "loss": 1.3044, + "theoretical_loss": 3.413563640933239, + "tokens_seen": 2351693824 + }, + { + "epoch": 0.71, + "learning_rate": 0.0002323771313941826, + "loss": 1.3707, + "theoretical_loss": 3.413556023870679, + "tokens_seen": 2351759360 + }, + { + "epoch": 0.71, + "learning_rate": 0.0002323610832497493, + "loss": 1.2927, + "theoretical_loss": 3.413548407079811, + "tokens_seen": 2351824896 + }, + { + "epoch": 0.71, + "learning_rate": 0.00023234503510531596, + "loss": 1.294, + "theoretical_loss": 3.413540790560618, + "tokens_seen": 2351890432 + }, + { + "epoch": 0.71, + "learning_rate": 0.00023232898696088267, + "loss": 1.2338, + "theoretical_loss": 3.4135331743130823, + "tokens_seen": 2351955968 + }, + { + "epoch": 0.71, + "learning_rate": 0.00023231293881644937, + "loss": 1.3287, + "theoretical_loss": 3.4135255583371866, + "tokens_seen": 2352021504 + }, + { + "epoch": 0.71, + "learning_rate": 0.00023229689067201607, + "loss": 1.3519, + "theoretical_loss": 3.413517942632914, + "tokens_seen": 2352087040 + }, + { + "epoch": 0.71, + "learning_rate": 0.00023228084252758277, + "loss": 1.3478, + "theoretical_loss": 3.413510327200247, + "tokens_seen": 2352152576 + }, + { + "epoch": 0.71, + "learning_rate": 0.00023226479438314947, + "loss": 1.2869, + "theoretical_loss": 3.4135027120391683, + "tokens_seen": 2352218112 + }, + { + "epoch": 0.71, + "learning_rate": 0.00023224874623871617, + "loss": 1.2315, + "theoretical_loss": 3.413495097149661, + "tokens_seen": 2352283648 + }, + { + "epoch": 0.71, + "learning_rate": 0.00023223269809428287, + "loss": 1.2905, + "theoretical_loss": 3.413487482531708, + "tokens_seen": 2352349184 + }, + { + "epoch": 0.71, + "learning_rate": 0.00023221664994984957, + "loss": 1.2671, + "theoretical_loss": 3.413479868185291, + "tokens_seen": 2352414720 + }, + { + "epoch": 0.71, + "learning_rate": 0.00023220060180541627, + "loss": 1.3795, + "theoretical_loss": 3.4134722541103937, + "tokens_seen": 2352480256 + }, + { + "epoch": 0.71, + "learning_rate": 0.00023218455366098297, + "loss": 1.3131, + "theoretical_loss": 3.4134646403069984, + "tokens_seen": 2352545792 + }, + { + "epoch": 0.71, + "learning_rate": 0.00023216850551654967, + "loss": 1.3505, + "theoretical_loss": 3.413457026775088, + "tokens_seen": 2352611328 + }, + { + "epoch": 0.71, + "learning_rate": 0.00023215245737211637, + "loss": 1.277, + "theoretical_loss": 3.4134494135146456, + "tokens_seen": 2352676864 + }, + { + "epoch": 0.71, + "objective/train/docs_used": 856640, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.4450786113739014, + "objective/train/theoretical_loss": 3.4134418005256535, + "objective/train/tokens_used": 2373202400, + "theoretical_loss": 3.4134418005256535, + "tokens_seen": 2352742400 + }, + { + "epoch": 0.71, + "learning_rate": 0.00023213640922768307, + "loss": 1.3485, + "theoretical_loss": 3.4134418005256535, + "tokens_seen": 2352742400 + }, + { + "epoch": 0.71, + "learning_rate": 0.00023212036108324977, + "loss": 1.3265, + "theoretical_loss": 3.4134341878080945, + "tokens_seen": 2352807936 + }, + { + "epoch": 0.71, + "learning_rate": 0.00023210431293881647, + "loss": 1.2999, + "theoretical_loss": 3.4134265753619513, + "tokens_seen": 2352873472 + }, + { + "epoch": 0.71, + "learning_rate": 0.00023208826479438317, + "loss": 1.2828, + "theoretical_loss": 3.4134189631872074, + "tokens_seen": 2352939008 + }, + { + "epoch": 0.71, + "learning_rate": 0.00023207221664994988, + "loss": 1.3419, + "theoretical_loss": 3.4134113512838447, + "tokens_seen": 2353004544 + }, + { + "epoch": 0.71, + "learning_rate": 0.00023205616850551658, + "loss": 1.3095, + "theoretical_loss": 3.413403739651846, + "tokens_seen": 2353070080 + }, + { + "epoch": 0.71, + "learning_rate": 0.00023204012036108328, + "loss": 1.3754, + "theoretical_loss": 3.4133961282911947, + "tokens_seen": 2353135616 + }, + { + "epoch": 0.71, + "learning_rate": 0.00023202407221664998, + "loss": 1.3778, + "theoretical_loss": 3.4133885172018728, + "tokens_seen": 2353201152 + }, + { + "epoch": 0.71, + "learning_rate": 0.00023200802407221668, + "loss": 1.3306, + "theoretical_loss": 3.413380906383864, + "tokens_seen": 2353266688 + }, + { + "epoch": 0.71, + "learning_rate": 0.00023199197592778338, + "loss": 1.2548, + "theoretical_loss": 3.4133732958371503, + "tokens_seen": 2353332224 + }, + { + "epoch": 0.71, + "learning_rate": 0.00023197592778335008, + "loss": 1.3431, + "theoretical_loss": 3.4133656855617147, + "tokens_seen": 2353397760 + }, + { + "epoch": 0.71, + "learning_rate": 0.00023195987963891675, + "loss": 1.2525, + "theoretical_loss": 3.4133580755575403, + "tokens_seen": 2353463296 + }, + { + "epoch": 0.71, + "learning_rate": 0.00023194383149448345, + "loss": 1.3087, + "theoretical_loss": 3.413350465824609, + "tokens_seen": 2353528832 + }, + { + "epoch": 0.71, + "learning_rate": 0.00023192778335005015, + "loss": 1.3218, + "theoretical_loss": 3.4133428563629047, + "tokens_seen": 2353594368 + }, + { + "epoch": 0.71, + "learning_rate": 0.00023191173520561685, + "loss": 1.2925, + "theoretical_loss": 3.4133352471724097, + "tokens_seen": 2353659904 + }, + { + "epoch": 0.71, + "learning_rate": 0.00023189568706118356, + "loss": 1.3466, + "theoretical_loss": 3.413327638253106, + "tokens_seen": 2353725440 + }, + { + "epoch": 0.71, + "learning_rate": 0.00023187963891675026, + "loss": 1.2923, + "theoretical_loss": 3.4133200296049777, + "tokens_seen": 2353790976 + }, + { + "epoch": 0.71, + "learning_rate": 0.00023186359077231696, + "loss": 1.4247, + "theoretical_loss": 3.4133124212280066, + "tokens_seen": 2353856512 + }, + { + "epoch": 0.71, + "learning_rate": 0.00023184754262788366, + "loss": 1.2563, + "theoretical_loss": 3.4133048131221764, + "tokens_seen": 2353922048 + }, + { + "epoch": 0.71, + "learning_rate": 0.00023183149448345036, + "loss": 1.222, + "theoretical_loss": 3.413297205287469, + "tokens_seen": 2353987584 + }, + { + "epoch": 0.71, + "learning_rate": 0.00023181544633901706, + "loss": 1.2101, + "theoretical_loss": 3.4132895977238675, + "tokens_seen": 2354053120 + }, + { + "epoch": 0.71, + "learning_rate": 0.00023179939819458376, + "loss": 1.3588, + "theoretical_loss": 3.413281990431355, + "tokens_seen": 2354118656 + }, + { + "epoch": 0.71, + "learning_rate": 0.00023178335005015046, + "loss": 1.2994, + "theoretical_loss": 3.413274383409914, + "tokens_seen": 2354184192 + }, + { + "epoch": 0.71, + "learning_rate": 0.00023176730190571716, + "loss": 1.2236, + "theoretical_loss": 3.413266776659527, + "tokens_seen": 2354249728 + }, + { + "epoch": 0.71, + "learning_rate": 0.00023175125376128386, + "loss": 1.3469, + "theoretical_loss": 3.4132591701801775, + "tokens_seen": 2354315264 + }, + { + "epoch": 0.71, + "learning_rate": 0.00023173520561685056, + "loss": 1.4036, + "theoretical_loss": 3.413251563971848, + "tokens_seen": 2354380800 + }, + { + "epoch": 0.71, + "learning_rate": 0.00023171915747241726, + "loss": 1.2612, + "theoretical_loss": 3.413243958034521, + "tokens_seen": 2354446336 + }, + { + "epoch": 0.71, + "learning_rate": 0.00023170310932798396, + "loss": 1.2, + "theoretical_loss": 3.4132363523681795, + "tokens_seen": 2354511872 + }, + { + "epoch": 0.71, + "learning_rate": 0.00023168706118355066, + "loss": 1.3324, + "theoretical_loss": 3.413228746972807, + "tokens_seen": 2354577408 + }, + { + "epoch": 0.71, + "learning_rate": 0.00023167101303911736, + "loss": 1.2549, + "theoretical_loss": 3.4132211418483847, + "tokens_seen": 2354642944 + }, + { + "epoch": 0.71, + "learning_rate": 0.00023165496489468407, + "loss": 1.2357, + "theoretical_loss": 3.413213536994897, + "tokens_seen": 2354708480 + }, + { + "epoch": 0.71, + "learning_rate": 0.00023163891675025077, + "loss": 1.3224, + "theoretical_loss": 3.4132059324123256, + "tokens_seen": 2354774016 + }, + { + "epoch": 0.71, + "learning_rate": 0.00023162286860581747, + "loss": 1.3206, + "theoretical_loss": 3.4131983281006537, + "tokens_seen": 2354839552 + }, + { + "epoch": 0.71, + "learning_rate": 0.00023160682046138417, + "loss": 1.2705, + "theoretical_loss": 3.4131907240598647, + "tokens_seen": 2354905088 + }, + { + "epoch": 0.71, + "learning_rate": 0.00023159077231695087, + "loss": 1.2844, + "theoretical_loss": 3.4131831202899408, + "tokens_seen": 2354970624 + }, + { + "epoch": 0.71, + "learning_rate": 0.00023157472417251754, + "loss": 1.1949, + "theoretical_loss": 3.4131755167908646, + "tokens_seen": 2355036160 + }, + { + "epoch": 0.71, + "learning_rate": 0.00023155867602808424, + "loss": 1.3097, + "theoretical_loss": 3.4131679135626194, + "tokens_seen": 2355101696 + }, + { + "epoch": 0.71, + "learning_rate": 0.00023154262788365094, + "loss": 1.3213, + "theoretical_loss": 3.4131603106051878, + "tokens_seen": 2355167232 + }, + { + "epoch": 0.71, + "learning_rate": 0.00023152657973921764, + "loss": 1.2793, + "theoretical_loss": 3.4131527079185524, + "tokens_seen": 2355232768 + }, + { + "epoch": 0.71, + "learning_rate": 0.00023151053159478434, + "loss": 1.3106, + "theoretical_loss": 3.413145105502697, + "tokens_seen": 2355298304 + }, + { + "epoch": 0.71, + "learning_rate": 0.00023149448345035104, + "loss": 1.3316, + "theoretical_loss": 3.413137503357603, + "tokens_seen": 2355363840 + }, + { + "epoch": 0.71, + "learning_rate": 0.00023147843530591774, + "loss": 1.3221, + "theoretical_loss": 3.4131299014832543, + "tokens_seen": 2355429376 + }, + { + "epoch": 0.71, + "learning_rate": 0.00023146238716148445, + "loss": 1.2656, + "theoretical_loss": 3.413122299879633, + "tokens_seen": 2355494912 + }, + { + "epoch": 0.71, + "learning_rate": 0.00023144633901705115, + "loss": 1.404, + "theoretical_loss": 3.4131146985467224, + "tokens_seen": 2355560448 + }, + { + "epoch": 0.71, + "learning_rate": 0.00023143029087261785, + "loss": 1.3327, + "theoretical_loss": 3.4131070974845055, + "tokens_seen": 2355625984 + }, + { + "epoch": 0.71, + "learning_rate": 0.00023141424272818455, + "loss": 1.2738, + "theoretical_loss": 3.4130994966929644, + "tokens_seen": 2355691520 + }, + { + "epoch": 0.71, + "learning_rate": 0.00023139819458375125, + "loss": 1.3231, + "theoretical_loss": 3.413091896172083, + "tokens_seen": 2355757056 + }, + { + "epoch": 0.71, + "learning_rate": 0.00023138214643931795, + "loss": 1.3787, + "theoretical_loss": 3.4130842959218426, + "tokens_seen": 2355822592 + }, + { + "epoch": 0.71, + "learning_rate": 0.00023136609829488465, + "loss": 1.3076, + "theoretical_loss": 3.4130766959422276, + "tokens_seen": 2355888128 + }, + { + "epoch": 0.71, + "learning_rate": 0.00023135005015045135, + "loss": 1.2883, + "theoretical_loss": 3.4130690962332197, + "tokens_seen": 2355953664 + }, + { + "epoch": 0.71, + "objective/train/docs_used": 857700, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.3081554174423218, + "objective/train/theoretical_loss": 3.4130614967948025, + "objective/train/tokens_used": 2376479200, + "theoretical_loss": 3.4130614967948025, + "tokens_seen": 2356019200 + }, + { + "epoch": 0.71, + "learning_rate": 0.00023133400200601805, + "loss": 1.3516, + "theoretical_loss": 3.4130614967948025, + "tokens_seen": 2356019200 + }, + { + "epoch": 0.71, + "learning_rate": 0.00023131795386158475, + "loss": 1.2917, + "theoretical_loss": 3.4130538976269587, + "tokens_seen": 2356084736 + }, + { + "epoch": 0.71, + "learning_rate": 0.00023130190571715145, + "loss": 1.3485, + "theoretical_loss": 3.413046298729671, + "tokens_seen": 2356150272 + }, + { + "epoch": 0.71, + "learning_rate": 0.00023128585757271815, + "loss": 1.2035, + "theoretical_loss": 3.4130387001029217, + "tokens_seen": 2356215808 + }, + { + "epoch": 0.71, + "learning_rate": 0.00023126980942828485, + "loss": 1.3269, + "theoretical_loss": 3.4130311017466943, + "tokens_seen": 2356281344 + }, + { + "epoch": 0.71, + "learning_rate": 0.00023125376128385155, + "loss": 1.3476, + "theoretical_loss": 3.4130235036609715, + "tokens_seen": 2356346880 + }, + { + "epoch": 0.71, + "learning_rate": 0.00023123771313941825, + "loss": 1.3361, + "theoretical_loss": 3.4130159058457368, + "tokens_seen": 2356412416 + }, + { + "epoch": 0.71, + "learning_rate": 0.00023122166499498498, + "loss": 1.2945, + "theoretical_loss": 3.413008308300972, + "tokens_seen": 2356477952 + }, + { + "epoch": 0.71, + "learning_rate": 0.00023120561685055168, + "loss": 1.3474, + "theoretical_loss": 3.41300071102666, + "tokens_seen": 2356543488 + }, + { + "epoch": 0.71, + "learning_rate": 0.00023118956870611838, + "loss": 1.2654, + "theoretical_loss": 3.4129931140227843, + "tokens_seen": 2356609024 + }, + { + "epoch": 0.71, + "learning_rate": 0.00023117352056168508, + "loss": 1.2966, + "theoretical_loss": 3.4129855172893278, + "tokens_seen": 2356674560 + }, + { + "epoch": 0.71, + "learning_rate": 0.00023115747241725178, + "loss": 1.275, + "theoretical_loss": 3.4129779208262723, + "tokens_seen": 2356740096 + }, + { + "epoch": 0.71, + "learning_rate": 0.00023114142427281849, + "loss": 1.3582, + "theoretical_loss": 3.412970324633602, + "tokens_seen": 2356805632 + }, + { + "epoch": 0.71, + "learning_rate": 0.00023112537612838519, + "loss": 1.3838, + "theoretical_loss": 3.4129627287112987, + "tokens_seen": 2356871168 + }, + { + "epoch": 0.71, + "learning_rate": 0.0002311093279839519, + "loss": 1.274, + "theoretical_loss": 3.412955133059346, + "tokens_seen": 2356936704 + }, + { + "epoch": 0.71, + "learning_rate": 0.0002310932798395186, + "loss": 1.2376, + "theoretical_loss": 3.4129475376777263, + "tokens_seen": 2357002240 + }, + { + "epoch": 0.71, + "learning_rate": 0.0002310772316950853, + "loss": 1.3781, + "theoretical_loss": 3.412939942566423, + "tokens_seen": 2357067776 + }, + { + "epoch": 0.71, + "learning_rate": 0.000231061183550652, + "loss": 1.2847, + "theoretical_loss": 3.412932347725418, + "tokens_seen": 2357133312 + }, + { + "epoch": 0.71, + "learning_rate": 0.0002310451354062187, + "loss": 1.3122, + "theoretical_loss": 3.412924753154695, + "tokens_seen": 2357198848 + }, + { + "epoch": 0.71, + "learning_rate": 0.0002310290872617854, + "loss": 1.3296, + "theoretical_loss": 3.4129171588542366, + "tokens_seen": 2357264384 + }, + { + "epoch": 0.71, + "learning_rate": 0.0002310130391173521, + "loss": 1.2437, + "theoretical_loss": 3.4129095648240257, + "tokens_seen": 2357329920 + }, + { + "epoch": 0.71, + "learning_rate": 0.0002309969909729188, + "loss": 1.2478, + "theoretical_loss": 3.412901971064045, + "tokens_seen": 2357395456 + }, + { + "epoch": 0.71, + "learning_rate": 0.0002309809428284855, + "loss": 1.3899, + "theoretical_loss": 3.4128943775742773, + "tokens_seen": 2357460992 + }, + { + "epoch": 0.71, + "learning_rate": 0.0002309648946840522, + "loss": 1.2714, + "theoretical_loss": 3.4128867843547064, + "tokens_seen": 2357526528 + }, + { + "epoch": 0.71, + "learning_rate": 0.0002309488465396189, + "loss": 1.2815, + "theoretical_loss": 3.412879191405314, + "tokens_seen": 2357592064 + }, + { + "epoch": 0.71, + "learning_rate": 0.00023093279839518557, + "loss": 1.2082, + "theoretical_loss": 3.412871598726084, + "tokens_seen": 2357657600 + }, + { + "epoch": 0.71, + "learning_rate": 0.00023091675025075227, + "loss": 1.2497, + "theoretical_loss": 3.4128640063169975, + "tokens_seen": 2357723136 + }, + { + "epoch": 0.71, + "learning_rate": 0.00023090070210631897, + "loss": 1.414, + "theoretical_loss": 3.41285641417804, + "tokens_seen": 2357788672 + }, + { + "epoch": 0.71, + "learning_rate": 0.00023088465396188567, + "loss": 1.2802, + "theoretical_loss": 3.412848822309192, + "tokens_seen": 2357854208 + }, + { + "epoch": 0.71, + "learning_rate": 0.00023086860581745237, + "loss": 1.329, + "theoretical_loss": 3.4128412307104377, + "tokens_seen": 2357919744 + }, + { + "epoch": 0.71, + "learning_rate": 0.00023085255767301907, + "loss": 1.2781, + "theoretical_loss": 3.41283363938176, + "tokens_seen": 2357985280 + }, + { + "epoch": 0.71, + "learning_rate": 0.00023083650952858577, + "loss": 1.2969, + "theoretical_loss": 3.412826048323141, + "tokens_seen": 2358050816 + }, + { + "epoch": 0.71, + "learning_rate": 0.00023082046138415247, + "loss": 1.2844, + "theoretical_loss": 3.412818457534564, + "tokens_seen": 2358116352 + }, + { + "epoch": 0.71, + "learning_rate": 0.00023080441323971917, + "loss": 1.4468, + "theoretical_loss": 3.412810867016012, + "tokens_seen": 2358181888 + }, + { + "epoch": 0.71, + "learning_rate": 0.00023078836509528587, + "loss": 1.2307, + "theoretical_loss": 3.4128032767674683, + "tokens_seen": 2358247424 + }, + { + "epoch": 0.71, + "learning_rate": 0.00023077231695085257, + "loss": 1.354, + "theoretical_loss": 3.4127956867889147, + "tokens_seen": 2358312960 + }, + { + "epoch": 0.71, + "learning_rate": 0.00023075626880641927, + "loss": 1.3544, + "theoretical_loss": 3.412788097080335, + "tokens_seen": 2358378496 + }, + { + "epoch": 0.71, + "learning_rate": 0.00023074022066198597, + "loss": 1.216, + "theoretical_loss": 3.4127805076417115, + "tokens_seen": 2358444032 + }, + { + "epoch": 0.71, + "learning_rate": 0.00023072417251755267, + "loss": 1.1907, + "theoretical_loss": 3.4127729184730278, + "tokens_seen": 2358509568 + }, + { + "epoch": 0.71, + "learning_rate": 0.00023070812437311938, + "loss": 1.3152, + "theoretical_loss": 3.412765329574266, + "tokens_seen": 2358575104 + }, + { + "epoch": 0.71, + "learning_rate": 0.00023069207622868608, + "loss": 1.3281, + "theoretical_loss": 3.4127577409454095, + "tokens_seen": 2358640640 + }, + { + "epoch": 0.71, + "learning_rate": 0.00023067602808425278, + "loss": 1.2611, + "theoretical_loss": 3.412750152586441, + "tokens_seen": 2358706176 + }, + { + "epoch": 0.71, + "learning_rate": 0.00023065997993981948, + "loss": 1.3258, + "theoretical_loss": 3.4127425644973437, + "tokens_seen": 2358771712 + }, + { + "epoch": 0.71, + "learning_rate": 0.00023064393179538618, + "loss": 1.3533, + "theoretical_loss": 3.4127349766781, + "tokens_seen": 2358837248 + }, + { + "epoch": 0.71, + "learning_rate": 0.00023062788365095288, + "loss": 1.373, + "theoretical_loss": 3.4127273891286936, + "tokens_seen": 2358902784 + }, + { + "epoch": 0.71, + "learning_rate": 0.00023061183550651958, + "loss": 1.3264, + "theoretical_loss": 3.4127198018491067, + "tokens_seen": 2358968320 + }, + { + "epoch": 0.71, + "learning_rate": 0.00023059578736208628, + "loss": 1.1957, + "theoretical_loss": 3.4127122148393223, + "tokens_seen": 2359033856 + }, + { + "epoch": 0.71, + "learning_rate": 0.00023057973921765298, + "loss": 1.2921, + "theoretical_loss": 3.4127046280993234, + "tokens_seen": 2359099392 + }, + { + "epoch": 0.71, + "learning_rate": 0.00023056369107321968, + "loss": 1.3917, + "theoretical_loss": 3.4126970416290927, + "tokens_seen": 2359164928 + }, + { + "epoch": 0.71, + "learning_rate": 0.00023054764292878635, + "loss": 1.1441, + "theoretical_loss": 3.412689455428614, + "tokens_seen": 2359230464 + }, + { + "epoch": 0.71, + "objective/train/docs_used": 858941, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.2477517127990723, + "objective/train/theoretical_loss": 3.412681869497869, + "objective/train/tokens_used": 2379756000, + "theoretical_loss": 3.412681869497869, + "tokens_seen": 2359296000 + }, + { + "epoch": 0.71, + "learning_rate": 0.00023053159478435306, + "loss": 1.2453, + "theoretical_loss": 3.412681869497869, + "tokens_seen": 2359296000 + }, + { + "epoch": 0.71, + "learning_rate": 0.00023051554663991976, + "loss": 1.3669, + "theoretical_loss": 3.4126742838368416, + "tokens_seen": 2359361536 + }, + { + "epoch": 0.71, + "learning_rate": 0.00023049949849548646, + "loss": 1.3846, + "theoretical_loss": 3.412666698445514, + "tokens_seen": 2359427072 + }, + { + "epoch": 0.71, + "learning_rate": 0.00023048345035105316, + "loss": 1.2307, + "theoretical_loss": 3.4126591133238695, + "tokens_seen": 2359492608 + }, + { + "epoch": 0.72, + "learning_rate": 0.00023046740220661986, + "loss": 1.353, + "theoretical_loss": 3.412651528471891, + "tokens_seen": 2359558144 + }, + { + "epoch": 0.72, + "learning_rate": 0.00023045135406218656, + "loss": 1.3608, + "theoretical_loss": 3.4126439438895613, + "tokens_seen": 2359623680 + }, + { + "epoch": 0.72, + "learning_rate": 0.00023043530591775326, + "loss": 1.2843, + "theoretical_loss": 3.4126363595768634, + "tokens_seen": 2359689216 + }, + { + "epoch": 0.72, + "learning_rate": 0.00023041925777331996, + "loss": 1.4481, + "theoretical_loss": 3.41262877553378, + "tokens_seen": 2359754752 + }, + { + "epoch": 0.72, + "learning_rate": 0.00023040320962888666, + "loss": 1.3621, + "theoretical_loss": 3.412621191760295, + "tokens_seen": 2359820288 + }, + { + "epoch": 0.72, + "learning_rate": 0.00023038716148445336, + "loss": 1.3567, + "theoretical_loss": 3.41261360825639, + "tokens_seen": 2359885824 + }, + { + "epoch": 0.72, + "learning_rate": 0.00023037111334002006, + "loss": 1.3479, + "theoretical_loss": 3.4126060250220487, + "tokens_seen": 2359951360 + }, + { + "epoch": 0.72, + "learning_rate": 0.00023035506519558676, + "loss": 1.2762, + "theoretical_loss": 3.412598442057254, + "tokens_seen": 2360016896 + }, + { + "epoch": 0.72, + "learning_rate": 0.00023033901705115346, + "loss": 1.3458, + "theoretical_loss": 3.412590859361988, + "tokens_seen": 2360082432 + }, + { + "epoch": 0.72, + "learning_rate": 0.00023032296890672016, + "loss": 1.3421, + "theoretical_loss": 3.4125832769362345, + "tokens_seen": 2360147968 + }, + { + "epoch": 0.72, + "learning_rate": 0.00023030692076228686, + "loss": 1.1682, + "theoretical_loss": 3.412575694779977, + "tokens_seen": 2360213504 + }, + { + "epoch": 0.72, + "learning_rate": 0.00023029087261785356, + "loss": 1.3886, + "theoretical_loss": 3.4125681128931973, + "tokens_seen": 2360279040 + }, + { + "epoch": 0.72, + "learning_rate": 0.00023027482447342027, + "loss": 1.3394, + "theoretical_loss": 3.4125605312758784, + "tokens_seen": 2360344576 + }, + { + "epoch": 0.72, + "learning_rate": 0.00023025877632898697, + "loss": 1.3063, + "theoretical_loss": 3.4125529499280036, + "tokens_seen": 2360410112 + }, + { + "epoch": 0.72, + "learning_rate": 0.00023024272818455367, + "loss": 1.2986, + "theoretical_loss": 3.412545368849556, + "tokens_seen": 2360475648 + }, + { + "epoch": 0.72, + "learning_rate": 0.00023022668004012037, + "loss": 1.2599, + "theoretical_loss": 3.4125377880405185, + "tokens_seen": 2360541184 + }, + { + "epoch": 0.72, + "learning_rate": 0.00023021063189568707, + "loss": 1.3584, + "theoretical_loss": 3.4125302075008737, + "tokens_seen": 2360606720 + }, + { + "epoch": 0.72, + "learning_rate": 0.00023019458375125377, + "loss": 1.3067, + "theoretical_loss": 3.412522627230605, + "tokens_seen": 2360672256 + }, + { + "epoch": 0.72, + "learning_rate": 0.00023017853560682047, + "loss": 1.3013, + "theoretical_loss": 3.412515047229695, + "tokens_seen": 2360737792 + }, + { + "epoch": 0.72, + "learning_rate": 0.00023016248746238714, + "loss": 1.2558, + "theoretical_loss": 3.4125074674981266, + "tokens_seen": 2360803328 + }, + { + "epoch": 0.72, + "learning_rate": 0.00023014643931795384, + "loss": 1.3893, + "theoretical_loss": 3.412499888035883, + "tokens_seen": 2360868864 + }, + { + "epoch": 0.72, + "learning_rate": 0.00023013039117352054, + "loss": 1.2986, + "theoretical_loss": 3.412492308842947, + "tokens_seen": 2360934400 + }, + { + "epoch": 0.72, + "learning_rate": 0.00023011434302908724, + "loss": 1.2571, + "theoretical_loss": 3.412484729919302, + "tokens_seen": 2360999936 + }, + { + "epoch": 0.72, + "learning_rate": 0.00023009829488465395, + "loss": 1.4111, + "theoretical_loss": 3.41247715126493, + "tokens_seen": 2361065472 + }, + { + "epoch": 0.72, + "learning_rate": 0.00023008224674022065, + "loss": 1.4081, + "theoretical_loss": 3.412469572879815, + "tokens_seen": 2361131008 + }, + { + "epoch": 0.72, + "learning_rate": 0.00023006619859578735, + "loss": 1.2945, + "theoretical_loss": 3.412461994763939, + "tokens_seen": 2361196544 + }, + { + "epoch": 0.72, + "learning_rate": 0.00023005015045135405, + "loss": 1.3243, + "theoretical_loss": 3.412454416917286, + "tokens_seen": 2361262080 + }, + { + "epoch": 0.72, + "learning_rate": 0.00023003410230692075, + "loss": 1.2495, + "theoretical_loss": 3.412446839339838, + "tokens_seen": 2361327616 + }, + { + "epoch": 0.72, + "learning_rate": 0.0002300180541624875, + "loss": 1.3254, + "theoretical_loss": 3.4124392620315787, + "tokens_seen": 2361393152 + }, + { + "epoch": 0.72, + "learning_rate": 0.0002300020060180542, + "loss": 1.375, + "theoretical_loss": 3.4124316849924905, + "tokens_seen": 2361458688 + }, + { + "epoch": 0.72, + "learning_rate": 0.0002299859578736209, + "loss": 1.289, + "theoretical_loss": 3.412424108222557, + "tokens_seen": 2361524224 + }, + { + "epoch": 0.72, + "learning_rate": 0.0002299699097291876, + "loss": 1.3636, + "theoretical_loss": 3.4124165317217603, + "tokens_seen": 2361589760 + }, + { + "epoch": 0.72, + "learning_rate": 0.0002299538615847543, + "loss": 1.2663, + "theoretical_loss": 3.4124089554900845, + "tokens_seen": 2361655296 + }, + { + "epoch": 0.72, + "learning_rate": 0.000229937813440321, + "loss": 1.3587, + "theoretical_loss": 3.4124013795275117, + "tokens_seen": 2361720832 + }, + { + "epoch": 0.72, + "learning_rate": 0.00022992176529588768, + "loss": 1.349, + "theoretical_loss": 3.412393803834025, + "tokens_seen": 2361786368 + }, + { + "epoch": 0.72, + "learning_rate": 0.00022990571715145438, + "loss": 1.2733, + "theoretical_loss": 3.4123862284096074, + "tokens_seen": 2361851904 + }, + { + "epoch": 0.72, + "learning_rate": 0.00022988966900702108, + "loss": 1.2715, + "theoretical_loss": 3.4123786532542417, + "tokens_seen": 2361917440 + }, + { + "epoch": 0.72, + "learning_rate": 0.00022987362086258778, + "loss": 1.3275, + "theoretical_loss": 3.4123710783679115, + "tokens_seen": 2361982976 + }, + { + "epoch": 0.72, + "learning_rate": 0.00022985757271815448, + "loss": 1.324, + "theoretical_loss": 3.4123635037505995, + "tokens_seen": 2362048512 + }, + { + "epoch": 0.72, + "learning_rate": 0.00022984152457372118, + "loss": 1.3199, + "theoretical_loss": 3.4123559294022883, + "tokens_seen": 2362114048 + }, + { + "epoch": 0.72, + "learning_rate": 0.00022982547642928788, + "loss": 1.3685, + "theoretical_loss": 3.412348355322962, + "tokens_seen": 2362179584 + }, + { + "epoch": 0.72, + "learning_rate": 0.00022980942828485458, + "loss": 1.3805, + "theoretical_loss": 3.4123407815126017, + "tokens_seen": 2362245120 + }, + { + "epoch": 0.72, + "learning_rate": 0.00022979338014042128, + "loss": 1.317, + "theoretical_loss": 3.412333207971192, + "tokens_seen": 2362310656 + }, + { + "epoch": 0.72, + "learning_rate": 0.00022977733199598799, + "loss": 1.3897, + "theoretical_loss": 3.4123256346987154, + "tokens_seen": 2362376192 + }, + { + "epoch": 0.72, + "learning_rate": 0.00022976128385155469, + "loss": 1.3065, + "theoretical_loss": 3.4123180616951547, + "tokens_seen": 2362441728 + }, + { + "epoch": 0.72, + "learning_rate": 0.00022974523570712139, + "loss": 1.2472, + "theoretical_loss": 3.412310488960493, + "tokens_seen": 2362507264 + }, + { + "epoch": 0.72, + "objective/train/docs_used": 860261, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.3941826820373535, + "objective/train/theoretical_loss": 3.4123029164947134, + "objective/train/tokens_used": 2383032800, + "theoretical_loss": 3.4123029164947134, + "tokens_seen": 2362572800 + }, + { + "epoch": 0.72, + "learning_rate": 0.0002297291875626881, + "loss": 1.3414, + "theoretical_loss": 3.4123029164947134, + "tokens_seen": 2362572800 + }, + { + "epoch": 0.72, + "learning_rate": 0.0002297131394182548, + "loss": 1.3287, + "theoretical_loss": 3.412295344297799, + "tokens_seen": 2362638336 + }, + { + "epoch": 0.72, + "learning_rate": 0.0002296970912738215, + "loss": 1.2731, + "theoretical_loss": 3.4122877723697327, + "tokens_seen": 2362703872 + }, + { + "epoch": 0.72, + "learning_rate": 0.0002296810431293882, + "loss": 1.241, + "theoretical_loss": 3.412280200710497, + "tokens_seen": 2362769408 + }, + { + "epoch": 0.72, + "learning_rate": 0.0002296649949849549, + "loss": 1.2955, + "theoretical_loss": 3.4122726293200754, + "tokens_seen": 2362834944 + }, + { + "epoch": 0.72, + "learning_rate": 0.0002296489468405216, + "loss": 1.281, + "theoretical_loss": 3.412265058198451, + "tokens_seen": 2362900480 + }, + { + "epoch": 0.72, + "learning_rate": 0.0002296328986960883, + "loss": 1.2886, + "theoretical_loss": 3.4122574873456064, + "tokens_seen": 2362966016 + }, + { + "epoch": 0.72, + "learning_rate": 0.000229616850551655, + "loss": 1.2487, + "theoretical_loss": 3.412249916761525, + "tokens_seen": 2363031552 + }, + { + "epoch": 0.72, + "learning_rate": 0.0002296008024072217, + "loss": 1.2677, + "theoretical_loss": 3.41224234644619, + "tokens_seen": 2363097088 + }, + { + "epoch": 0.72, + "learning_rate": 0.0002295847542627884, + "loss": 1.2847, + "theoretical_loss": 3.4122347763995835, + "tokens_seen": 2363162624 + }, + { + "epoch": 0.72, + "learning_rate": 0.0002295687061183551, + "loss": 1.4079, + "theoretical_loss": 3.412227206621689, + "tokens_seen": 2363228160 + }, + { + "epoch": 0.72, + "learning_rate": 0.0002295526579739218, + "loss": 1.3481, + "theoretical_loss": 3.4122196371124898, + "tokens_seen": 2363293696 + }, + { + "epoch": 0.72, + "learning_rate": 0.00022953660982948847, + "loss": 1.2979, + "theoretical_loss": 3.4122120678719687, + "tokens_seen": 2363359232 + }, + { + "epoch": 0.72, + "learning_rate": 0.00022952056168505517, + "loss": 1.2634, + "theoretical_loss": 3.4122044989001084, + "tokens_seen": 2363424768 + }, + { + "epoch": 0.72, + "learning_rate": 0.00022950451354062187, + "loss": 1.2772, + "theoretical_loss": 3.4121969301968926, + "tokens_seen": 2363490304 + }, + { + "epoch": 0.72, + "learning_rate": 0.00022948846539618857, + "loss": 1.2904, + "theoretical_loss": 3.4121893617623034, + "tokens_seen": 2363555840 + }, + { + "epoch": 0.72, + "learning_rate": 0.00022947241725175527, + "loss": 1.3767, + "theoretical_loss": 3.4121817935963246, + "tokens_seen": 2363621376 + }, + { + "epoch": 0.72, + "learning_rate": 0.00022945636910732197, + "loss": 1.3718, + "theoretical_loss": 3.412174225698939, + "tokens_seen": 2363686912 + }, + { + "epoch": 0.72, + "learning_rate": 0.00022944032096288867, + "loss": 1.3507, + "theoretical_loss": 3.41216665807013, + "tokens_seen": 2363752448 + }, + { + "epoch": 0.72, + "learning_rate": 0.00022942427281845537, + "loss": 1.2841, + "theoretical_loss": 3.4121590907098795, + "tokens_seen": 2363817984 + }, + { + "epoch": 0.72, + "learning_rate": 0.00022940822467402207, + "loss": 1.2845, + "theoretical_loss": 3.4121515236181716, + "tokens_seen": 2363883520 + }, + { + "epoch": 0.72, + "learning_rate": 0.00022939217652958877, + "loss": 1.317, + "theoretical_loss": 3.4121439567949885, + "tokens_seen": 2363949056 + }, + { + "epoch": 0.72, + "learning_rate": 0.00022937612838515547, + "loss": 1.231, + "theoretical_loss": 3.412136390240314, + "tokens_seen": 2364014592 + }, + { + "epoch": 0.72, + "learning_rate": 0.00022936008024072217, + "loss": 1.3478, + "theoretical_loss": 3.4121288239541308, + "tokens_seen": 2364080128 + }, + { + "epoch": 0.72, + "learning_rate": 0.00022934403209628888, + "loss": 1.1214, + "theoretical_loss": 3.412121257936422, + "tokens_seen": 2364145664 + }, + { + "epoch": 0.72, + "learning_rate": 0.00022932798395185558, + "loss": 1.2348, + "theoretical_loss": 3.4121136921871704, + "tokens_seen": 2364211200 + }, + { + "epoch": 0.72, + "learning_rate": 0.00022931193580742228, + "loss": 1.3523, + "theoretical_loss": 3.4121061267063593, + "tokens_seen": 2364276736 + }, + { + "epoch": 0.72, + "learning_rate": 0.00022929588766298898, + "loss": 1.4134, + "theoretical_loss": 3.4120985614939716, + "tokens_seen": 2364342272 + }, + { + "epoch": 0.72, + "learning_rate": 0.00022927983951855568, + "loss": 1.3303, + "theoretical_loss": 3.41209099654999, + "tokens_seen": 2364407808 + }, + { + "epoch": 0.72, + "learning_rate": 0.00022926379137412238, + "loss": 1.3392, + "theoretical_loss": 3.4120834318743984, + "tokens_seen": 2364473344 + }, + { + "epoch": 0.72, + "learning_rate": 0.00022924774322968908, + "loss": 1.3687, + "theoretical_loss": 3.4120758674671796, + "tokens_seen": 2364538880 + }, + { + "epoch": 0.72, + "learning_rate": 0.00022923169508525578, + "loss": 1.311, + "theoretical_loss": 3.412068303328316, + "tokens_seen": 2364604416 + }, + { + "epoch": 0.72, + "learning_rate": 0.00022921564694082248, + "loss": 1.2702, + "theoretical_loss": 3.412060739457791, + "tokens_seen": 2364669952 + }, + { + "epoch": 0.72, + "learning_rate": 0.00022919959879638918, + "loss": 1.2912, + "theoretical_loss": 3.4120531758555877, + "tokens_seen": 2364735488 + }, + { + "epoch": 0.72, + "learning_rate": 0.00022918355065195588, + "loss": 1.2949, + "theoretical_loss": 3.412045612521689, + "tokens_seen": 2364801024 + }, + { + "epoch": 0.72, + "learning_rate": 0.00022916750250752258, + "loss": 1.3089, + "theoretical_loss": 3.4120380494560782, + "tokens_seen": 2364866560 + }, + { + "epoch": 0.72, + "learning_rate": 0.00022915145436308926, + "loss": 1.3162, + "theoretical_loss": 3.412030486658738, + "tokens_seen": 2364932096 + }, + { + "epoch": 0.72, + "learning_rate": 0.00022913540621865596, + "loss": 1.2751, + "theoretical_loss": 3.412022924129652, + "tokens_seen": 2364997632 + }, + { + "epoch": 0.72, + "learning_rate": 0.00022911935807422266, + "loss": 1.2757, + "theoretical_loss": 3.4120153618688027, + "tokens_seen": 2365063168 + }, + { + "epoch": 0.72, + "learning_rate": 0.00022910330992978936, + "loss": 1.3282, + "theoretical_loss": 3.4120077998761733, + "tokens_seen": 2365128704 + }, + { + "epoch": 0.72, + "learning_rate": 0.00022908726178535606, + "loss": 1.3462, + "theoretical_loss": 3.412000238151747, + "tokens_seen": 2365194240 + }, + { + "epoch": 0.72, + "learning_rate": 0.00022907121364092276, + "loss": 1.3475, + "theoretical_loss": 3.411992676695507, + "tokens_seen": 2365259776 + }, + { + "epoch": 0.72, + "learning_rate": 0.00022905516549648946, + "loss": 1.2904, + "theoretical_loss": 3.411985115507436, + "tokens_seen": 2365325312 + }, + { + "epoch": 0.72, + "learning_rate": 0.00022903911735205616, + "loss": 1.2304, + "theoretical_loss": 3.411977554587517, + "tokens_seen": 2365390848 + }, + { + "epoch": 0.72, + "learning_rate": 0.00022902306920762286, + "loss": 1.2536, + "theoretical_loss": 3.4119699939357337, + "tokens_seen": 2365456384 + }, + { + "epoch": 0.72, + "learning_rate": 0.00022900702106318956, + "loss": 1.3375, + "theoretical_loss": 3.4119624335520684, + "tokens_seen": 2365521920 + }, + { + "epoch": 0.72, + "learning_rate": 0.00022899097291875626, + "loss": 1.3631, + "theoretical_loss": 3.4119548734365046, + "tokens_seen": 2365587456 + }, + { + "epoch": 0.72, + "learning_rate": 0.00022897492477432296, + "loss": 1.2939, + "theoretical_loss": 3.411947313589025, + "tokens_seen": 2365652992 + }, + { + "epoch": 0.72, + "learning_rate": 0.00022895887662988966, + "loss": 1.2983, + "theoretical_loss": 3.4119397540096132, + "tokens_seen": 2365718528 + }, + { + "epoch": 0.72, + "learning_rate": 0.00022894282848545636, + "loss": 1.3478, + "theoretical_loss": 3.411932194698252, + "tokens_seen": 2365784064 + }, + { + "epoch": 0.72, + "objective/train/docs_used": 861421, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.1135318279266357, + "objective/train/theoretical_loss": 3.411924635654924, + "objective/train/tokens_used": 2386309600, + "theoretical_loss": 3.411924635654924, + "tokens_seen": 2365849600 + }, + { + "epoch": 0.72, + "learning_rate": 0.00022892678034102306, + "loss": 1.1734, + "theoretical_loss": 3.411924635654924, + "tokens_seen": 2365849600 + }, + { + "epoch": 0.72, + "learning_rate": 0.00022891073219658977, + "loss": 1.3802, + "theoretical_loss": 3.411917076879613, + "tokens_seen": 2365915136 + }, + { + "epoch": 0.72, + "learning_rate": 0.00022889468405215647, + "loss": 1.3626, + "theoretical_loss": 3.411909518372302, + "tokens_seen": 2365980672 + }, + { + "epoch": 0.72, + "learning_rate": 0.00022887863590772317, + "loss": 1.326, + "theoretical_loss": 3.411901960132974, + "tokens_seen": 2366046208 + }, + { + "epoch": 0.72, + "learning_rate": 0.00022886258776328987, + "loss": 1.1494, + "theoretical_loss": 3.4118944021616118, + "tokens_seen": 2366111744 + }, + { + "epoch": 0.72, + "learning_rate": 0.00022884653961885657, + "loss": 1.3664, + "theoretical_loss": 3.4118868444581985, + "tokens_seen": 2366177280 + }, + { + "epoch": 0.72, + "learning_rate": 0.00022883049147442327, + "loss": 1.2103, + "theoretical_loss": 3.4118792870227175, + "tokens_seen": 2366242816 + }, + { + "epoch": 0.72, + "learning_rate": 0.00022881444332999, + "loss": 1.2964, + "theoretical_loss": 3.411871729855151, + "tokens_seen": 2366308352 + }, + { + "epoch": 0.72, + "learning_rate": 0.0002287983951855567, + "loss": 1.2777, + "theoretical_loss": 3.4118641729554837, + "tokens_seen": 2366373888 + }, + { + "epoch": 0.72, + "learning_rate": 0.0002287823470411234, + "loss": 1.2406, + "theoretical_loss": 3.411856616323697, + "tokens_seen": 2366439424 + }, + { + "epoch": 0.72, + "learning_rate": 0.0002287662988966901, + "loss": 1.3824, + "theoretical_loss": 3.4118490599597755, + "tokens_seen": 2366504960 + }, + { + "epoch": 0.72, + "learning_rate": 0.0002287502507522568, + "loss": 1.2956, + "theoretical_loss": 3.411841503863701, + "tokens_seen": 2366570496 + }, + { + "epoch": 0.72, + "learning_rate": 0.0002287342026078235, + "loss": 1.2016, + "theoretical_loss": 3.4118339480354574, + "tokens_seen": 2366636032 + }, + { + "epoch": 0.72, + "learning_rate": 0.0002287181544633902, + "loss": 1.275, + "theoretical_loss": 3.411826392475027, + "tokens_seen": 2366701568 + }, + { + "epoch": 0.72, + "learning_rate": 0.0002287021063189569, + "loss": 1.3545, + "theoretical_loss": 3.4118188371823943, + "tokens_seen": 2366767104 + }, + { + "epoch": 0.72, + "learning_rate": 0.0002286860581745236, + "loss": 1.2304, + "theoretical_loss": 3.411811282157541, + "tokens_seen": 2366832640 + }, + { + "epoch": 0.72, + "learning_rate": 0.0002286700100300903, + "loss": 1.3428, + "theoretical_loss": 3.4118037274004505, + "tokens_seen": 2366898176 + }, + { + "epoch": 0.72, + "learning_rate": 0.000228653961885657, + "loss": 1.459, + "theoretical_loss": 3.411796172911106, + "tokens_seen": 2366963712 + }, + { + "epoch": 0.72, + "learning_rate": 0.0002286379137412237, + "loss": 1.4386, + "theoretical_loss": 3.4117886186894912, + "tokens_seen": 2367029248 + }, + { + "epoch": 0.72, + "learning_rate": 0.0002286218655967904, + "loss": 1.2358, + "theoretical_loss": 3.4117810647355884, + "tokens_seen": 2367094784 + }, + { + "epoch": 0.72, + "learning_rate": 0.0002286058174523571, + "loss": 1.3178, + "theoretical_loss": 3.411773511049381, + "tokens_seen": 2367160320 + }, + { + "epoch": 0.72, + "learning_rate": 0.0002285897693079238, + "loss": 1.3034, + "theoretical_loss": 3.4117659576308523, + "tokens_seen": 2367225856 + }, + { + "epoch": 0.72, + "learning_rate": 0.0002285737211634905, + "loss": 1.2682, + "theoretical_loss": 3.411758404479985, + "tokens_seen": 2367291392 + }, + { + "epoch": 0.72, + "learning_rate": 0.0002285576730190572, + "loss": 1.3544, + "theoretical_loss": 3.4117508515967625, + "tokens_seen": 2367356928 + }, + { + "epoch": 0.72, + "learning_rate": 0.0002285416248746239, + "loss": 1.3375, + "theoretical_loss": 3.411743298981168, + "tokens_seen": 2367422464 + }, + { + "epoch": 0.72, + "learning_rate": 0.00022852557673019058, + "loss": 1.2593, + "theoretical_loss": 3.411735746633184, + "tokens_seen": 2367488000 + }, + { + "epoch": 0.72, + "learning_rate": 0.00022850952858575728, + "loss": 1.2854, + "theoretical_loss": 3.411728194552794, + "tokens_seen": 2367553536 + }, + { + "epoch": 0.72, + "learning_rate": 0.00022849348044132398, + "loss": 1.2884, + "theoretical_loss": 3.411720642739981, + "tokens_seen": 2367619072 + }, + { + "epoch": 0.72, + "learning_rate": 0.00022847743229689068, + "loss": 1.4285, + "theoretical_loss": 3.411713091194729, + "tokens_seen": 2367684608 + }, + { + "epoch": 0.72, + "learning_rate": 0.00022846138415245738, + "loss": 1.3023, + "theoretical_loss": 3.41170553991702, + "tokens_seen": 2367750144 + }, + { + "epoch": 0.72, + "learning_rate": 0.00022844533600802408, + "loss": 1.2974, + "theoretical_loss": 3.411697988906837, + "tokens_seen": 2367815680 + }, + { + "epoch": 0.72, + "learning_rate": 0.00022842928786359078, + "loss": 1.3482, + "theoretical_loss": 3.411690438164164, + "tokens_seen": 2367881216 + }, + { + "epoch": 0.72, + "learning_rate": 0.00022841323971915748, + "loss": 1.2352, + "theoretical_loss": 3.411682887688984, + "tokens_seen": 2367946752 + }, + { + "epoch": 0.72, + "learning_rate": 0.00022839719157472419, + "loss": 1.3502, + "theoretical_loss": 3.411675337481279, + "tokens_seen": 2368012288 + }, + { + "epoch": 0.72, + "learning_rate": 0.00022838114343029089, + "loss": 1.2517, + "theoretical_loss": 3.4116677875410337, + "tokens_seen": 2368077824 + }, + { + "epoch": 0.72, + "learning_rate": 0.0002283650952858576, + "loss": 1.2957, + "theoretical_loss": 3.4116602378682304, + "tokens_seen": 2368143360 + }, + { + "epoch": 0.72, + "learning_rate": 0.0002283490471414243, + "loss": 1.2015, + "theoretical_loss": 3.411652688462852, + "tokens_seen": 2368208896 + }, + { + "epoch": 0.72, + "learning_rate": 0.000228332998996991, + "loss": 1.3455, + "theoretical_loss": 3.4116451393248823, + "tokens_seen": 2368274432 + }, + { + "epoch": 0.72, + "learning_rate": 0.0002283169508525577, + "loss": 1.2137, + "theoretical_loss": 3.411637590454304, + "tokens_seen": 2368339968 + }, + { + "epoch": 0.72, + "learning_rate": 0.0002283009027081244, + "loss": 1.3476, + "theoretical_loss": 3.4116300418511, + "tokens_seen": 2368405504 + }, + { + "epoch": 0.72, + "learning_rate": 0.0002282848545636911, + "loss": 1.2789, + "theoretical_loss": 3.4116224935152544, + "tokens_seen": 2368471040 + }, + { + "epoch": 0.72, + "learning_rate": 0.0002282688064192578, + "loss": 1.3277, + "theoretical_loss": 3.411614945446749, + "tokens_seen": 2368536576 + }, + { + "epoch": 0.72, + "learning_rate": 0.0002282527582748245, + "loss": 1.2688, + "theoretical_loss": 3.411607397645568, + "tokens_seen": 2368602112 + }, + { + "epoch": 0.72, + "learning_rate": 0.0002282367101303912, + "loss": 1.3006, + "theoretical_loss": 3.4115998501116938, + "tokens_seen": 2368667648 + }, + { + "epoch": 0.72, + "learning_rate": 0.0002282206619859579, + "loss": 1.3536, + "theoretical_loss": 3.41159230284511, + "tokens_seen": 2368733184 + }, + { + "epoch": 0.72, + "learning_rate": 0.0002282046138415246, + "loss": 1.2475, + "theoretical_loss": 3.4115847558457997, + "tokens_seen": 2368798720 + }, + { + "epoch": 0.72, + "learning_rate": 0.0002281885656970913, + "loss": 1.3308, + "theoretical_loss": 3.4115772091137457, + "tokens_seen": 2368864256 + }, + { + "epoch": 0.72, + "learning_rate": 0.000228172517552658, + "loss": 1.4141, + "theoretical_loss": 3.411569662648932, + "tokens_seen": 2368929792 + }, + { + "epoch": 0.72, + "learning_rate": 0.0002281564694082247, + "loss": 1.3124, + "theoretical_loss": 3.4115621164513406, + "tokens_seen": 2368995328 + }, + { + "epoch": 0.72, + "learning_rate": 0.00022814042126379137, + "loss": 1.3501, + "theoretical_loss": 3.411554570520955, + "tokens_seen": 2369060864 + }, + { + "epoch": 0.72, + "objective/train/docs_used": 862685, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.2108964920043945, + "objective/train/theoretical_loss": 3.411547024857759, + "objective/train/tokens_used": 2389586400, + "theoretical_loss": 3.411547024857759, + "tokens_seen": 2369126400 + }, + { + "epoch": 0.72, + "learning_rate": 0.00022812437311935807, + "loss": 1.2576, + "theoretical_loss": 3.411547024857759, + "tokens_seen": 2369126400 + }, + { + "epoch": 0.72, + "learning_rate": 0.00022810832497492477, + "loss": 1.2143, + "theoretical_loss": 3.411539479461735, + "tokens_seen": 2369191936 + }, + { + "epoch": 0.72, + "learning_rate": 0.00022809227683049147, + "loss": 1.3081, + "theoretical_loss": 3.4115319343328667, + "tokens_seen": 2369257472 + }, + { + "epoch": 0.72, + "learning_rate": 0.00022807622868605817, + "loss": 1.3135, + "theoretical_loss": 3.4115243894711367, + "tokens_seen": 2369323008 + }, + { + "epoch": 0.72, + "learning_rate": 0.00022806018054162487, + "loss": 1.2638, + "theoretical_loss": 3.4115168448765285, + "tokens_seen": 2369388544 + }, + { + "epoch": 0.72, + "learning_rate": 0.00022804413239719157, + "loss": 1.391, + "theoretical_loss": 3.4115093005490253, + "tokens_seen": 2369454080 + }, + { + "epoch": 0.72, + "learning_rate": 0.00022802808425275827, + "loss": 1.2732, + "theoretical_loss": 3.41150175648861, + "tokens_seen": 2369519616 + }, + { + "epoch": 0.72, + "learning_rate": 0.00022801203610832497, + "loss": 1.3157, + "theoretical_loss": 3.411494212695266, + "tokens_seen": 2369585152 + }, + { + "epoch": 0.72, + "learning_rate": 0.00022799598796389167, + "loss": 1.3186, + "theoretical_loss": 3.4114866691689762, + "tokens_seen": 2369650688 + }, + { + "epoch": 0.72, + "learning_rate": 0.00022797993981945837, + "loss": 1.3567, + "theoretical_loss": 3.4114791259097244, + "tokens_seen": 2369716224 + }, + { + "epoch": 0.72, + "learning_rate": 0.00022796389167502508, + "loss": 1.2706, + "theoretical_loss": 3.4114715829174926, + "tokens_seen": 2369781760 + }, + { + "epoch": 0.72, + "learning_rate": 0.00022794784353059178, + "loss": 1.3051, + "theoretical_loss": 3.411464040192265, + "tokens_seen": 2369847296 + }, + { + "epoch": 0.72, + "learning_rate": 0.00022793179538615848, + "loss": 1.3396, + "theoretical_loss": 3.411456497734025, + "tokens_seen": 2369912832 + }, + { + "epoch": 0.72, + "learning_rate": 0.00022791574724172518, + "loss": 1.2775, + "theoretical_loss": 3.4114489555427543, + "tokens_seen": 2369978368 + }, + { + "epoch": 0.72, + "learning_rate": 0.00022789969909729188, + "loss": 1.2403, + "theoretical_loss": 3.4114414136184372, + "tokens_seen": 2370043904 + }, + { + "epoch": 0.72, + "learning_rate": 0.00022788365095285858, + "loss": 1.3347, + "theoretical_loss": 3.411433871961057, + "tokens_seen": 2370109440 + }, + { + "epoch": 0.72, + "learning_rate": 0.00022786760280842528, + "loss": 1.3043, + "theoretical_loss": 3.411426330570596, + "tokens_seen": 2370174976 + }, + { + "epoch": 0.72, + "learning_rate": 0.00022785155466399198, + "loss": 1.2978, + "theoretical_loss": 3.411418789447038, + "tokens_seen": 2370240512 + }, + { + "epoch": 0.72, + "learning_rate": 0.00022783550651955868, + "loss": 1.2957, + "theoretical_loss": 3.4114112485903663, + "tokens_seen": 2370306048 + }, + { + "epoch": 0.72, + "learning_rate": 0.00022781945837512538, + "loss": 1.3278, + "theoretical_loss": 3.4114037080005635, + "tokens_seen": 2370371584 + }, + { + "epoch": 0.72, + "learning_rate": 0.00022780341023069208, + "loss": 1.3436, + "theoretical_loss": 3.411396167677613, + "tokens_seen": 2370437120 + }, + { + "epoch": 0.72, + "learning_rate": 0.00022778736208625878, + "loss": 1.2364, + "theoretical_loss": 3.4113886276214984, + "tokens_seen": 2370502656 + }, + { + "epoch": 0.72, + "learning_rate": 0.00022777131394182548, + "loss": 1.372, + "theoretical_loss": 3.4113810878322024, + "tokens_seen": 2370568192 + }, + { + "epoch": 0.72, + "learning_rate": 0.00022775526579739216, + "loss": 1.3291, + "theoretical_loss": 3.411373548309708, + "tokens_seen": 2370633728 + }, + { + "epoch": 0.72, + "learning_rate": 0.00022773921765295886, + "loss": 1.369, + "theoretical_loss": 3.411366009053999, + "tokens_seen": 2370699264 + }, + { + "epoch": 0.72, + "learning_rate": 0.00022772316950852556, + "loss": 1.343, + "theoretical_loss": 3.4113584700650588, + "tokens_seen": 2370764800 + }, + { + "epoch": 0.72, + "learning_rate": 0.00022770712136409226, + "loss": 1.3196, + "theoretical_loss": 3.411350931342869, + "tokens_seen": 2370830336 + }, + { + "epoch": 0.72, + "learning_rate": 0.00022769107321965896, + "loss": 1.3162, + "theoretical_loss": 3.4113433928874146, + "tokens_seen": 2370895872 + }, + { + "epoch": 0.72, + "learning_rate": 0.00022767502507522566, + "loss": 1.3667, + "theoretical_loss": 3.4113358546986783, + "tokens_seen": 2370961408 + }, + { + "epoch": 0.72, + "learning_rate": 0.00022765897693079236, + "loss": 1.3634, + "theoretical_loss": 3.411328316776643, + "tokens_seen": 2371026944 + }, + { + "epoch": 0.72, + "learning_rate": 0.00022764292878635906, + "loss": 1.3968, + "theoretical_loss": 3.4113207791212914, + "tokens_seen": 2371092480 + }, + { + "epoch": 0.72, + "learning_rate": 0.00022762688064192582, + "loss": 1.3684, + "theoretical_loss": 3.4113132417326075, + "tokens_seen": 2371158016 + }, + { + "epoch": 0.72, + "learning_rate": 0.00022761083249749252, + "loss": 1.2406, + "theoretical_loss": 3.4113057046105744, + "tokens_seen": 2371223552 + }, + { + "epoch": 0.72, + "learning_rate": 0.00022759478435305922, + "loss": 1.3205, + "theoretical_loss": 3.411298167755175, + "tokens_seen": 2371289088 + }, + { + "epoch": 0.72, + "learning_rate": 0.00022757873620862592, + "loss": 1.4115, + "theoretical_loss": 3.4112906311663926, + "tokens_seen": 2371354624 + }, + { + "epoch": 0.72, + "learning_rate": 0.00022756268806419262, + "loss": 1.2644, + "theoretical_loss": 3.4112830948442108, + "tokens_seen": 2371420160 + }, + { + "epoch": 0.72, + "learning_rate": 0.00022754663991975932, + "loss": 1.3263, + "theoretical_loss": 3.4112755587886117, + "tokens_seen": 2371485696 + }, + { + "epoch": 0.72, + "learning_rate": 0.00022753059177532602, + "loss": 1.2697, + "theoretical_loss": 3.4112680229995798, + "tokens_seen": 2371551232 + }, + { + "epoch": 0.72, + "learning_rate": 0.00022751454363089272, + "loss": 1.3465, + "theoretical_loss": 3.4112604874770978, + "tokens_seen": 2371616768 + }, + { + "epoch": 0.72, + "learning_rate": 0.0002274984954864594, + "loss": 1.2717, + "theoretical_loss": 3.4112529522211483, + "tokens_seen": 2371682304 + }, + { + "epoch": 0.72, + "learning_rate": 0.0002274824473420261, + "loss": 1.3393, + "theoretical_loss": 3.4112454172317155, + "tokens_seen": 2371747840 + }, + { + "epoch": 0.72, + "learning_rate": 0.0002274663991975928, + "loss": 1.295, + "theoretical_loss": 3.4112378825087823, + "tokens_seen": 2371813376 + }, + { + "epoch": 0.72, + "learning_rate": 0.0002274503510531595, + "loss": 1.3434, + "theoretical_loss": 3.4112303480523316, + "tokens_seen": 2371878912 + }, + { + "epoch": 0.72, + "learning_rate": 0.0002274343029087262, + "loss": 1.2841, + "theoretical_loss": 3.411222813862347, + "tokens_seen": 2371944448 + }, + { + "epoch": 0.72, + "learning_rate": 0.0002274182547642929, + "loss": 1.2612, + "theoretical_loss": 3.411215279938811, + "tokens_seen": 2372009984 + }, + { + "epoch": 0.72, + "learning_rate": 0.0002274022066198596, + "loss": 1.2593, + "theoretical_loss": 3.4112077462817076, + "tokens_seen": 2372075520 + }, + { + "epoch": 0.72, + "learning_rate": 0.0002273861584754263, + "loss": 1.2908, + "theoretical_loss": 3.41120021289102, + "tokens_seen": 2372141056 + }, + { + "epoch": 0.72, + "learning_rate": 0.000227370110330993, + "loss": 1.3598, + "theoretical_loss": 3.4111926797667307, + "tokens_seen": 2372206592 + }, + { + "epoch": 0.72, + "learning_rate": 0.0002273540621865597, + "loss": 1.3027, + "theoretical_loss": 3.4111851469088235, + "tokens_seen": 2372272128 + }, + { + "epoch": 0.72, + "learning_rate": 0.0002273380140421264, + "loss": 1.3002, + "theoretical_loss": 3.411177614317282, + "tokens_seen": 2372337664 + }, + { + "epoch": 0.72, + "objective/train/docs_used": 863878, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.2855571508407593, + "objective/train/theoretical_loss": 3.4111700819920885, + "objective/train/tokens_used": 2392863200, + "theoretical_loss": 3.4111700819920885, + "tokens_seen": 2372403200 + }, + { + "epoch": 0.72, + "learning_rate": 0.0002273219658976931, + "loss": 1.2885, + "theoretical_loss": 3.4111700819920885, + "tokens_seen": 2372403200 + }, + { + "epoch": 0.72, + "learning_rate": 0.0002273059177532598, + "loss": 1.2357, + "theoretical_loss": 3.4111625499332265, + "tokens_seen": 2372468736 + }, + { + "epoch": 0.72, + "learning_rate": 0.0002272898696088265, + "loss": 1.1955, + "theoretical_loss": 3.4111550181406796, + "tokens_seen": 2372534272 + }, + { + "epoch": 0.72, + "learning_rate": 0.0002272738214643932, + "loss": 1.2195, + "theoretical_loss": 3.4111474866144307, + "tokens_seen": 2372599808 + }, + { + "epoch": 0.72, + "learning_rate": 0.0002272577733199599, + "loss": 1.2775, + "theoretical_loss": 3.4111399553544635, + "tokens_seen": 2372665344 + }, + { + "epoch": 0.72, + "learning_rate": 0.0002272417251755266, + "loss": 1.273, + "theoretical_loss": 3.4111324243607606, + "tokens_seen": 2372730880 + }, + { + "epoch": 0.72, + "learning_rate": 0.0002272256770310933, + "loss": 1.3046, + "theoretical_loss": 3.411124893633305, + "tokens_seen": 2372796416 + }, + { + "epoch": 0.72, + "learning_rate": 0.00022720962888666, + "loss": 1.2941, + "theoretical_loss": 3.4111173631720813, + "tokens_seen": 2372861952 + }, + { + "epoch": 0.72, + "learning_rate": 0.0002271935807422267, + "loss": 1.272, + "theoretical_loss": 3.4111098329770715, + "tokens_seen": 2372927488 + }, + { + "epoch": 0.72, + "learning_rate": 0.0002271775325977934, + "loss": 1.3574, + "theoretical_loss": 3.411102303048259, + "tokens_seen": 2372993024 + }, + { + "epoch": 0.72, + "learning_rate": 0.0002271614844533601, + "loss": 1.2453, + "theoretical_loss": 3.411094773385628, + "tokens_seen": 2373058560 + }, + { + "epoch": 0.72, + "learning_rate": 0.0002271454363089268, + "loss": 1.2948, + "theoretical_loss": 3.4110872439891597, + "tokens_seen": 2373124096 + }, + { + "epoch": 0.72, + "learning_rate": 0.0002271293881644935, + "loss": 1.381, + "theoretical_loss": 3.4110797148588397, + "tokens_seen": 2373189632 + }, + { + "epoch": 0.72, + "learning_rate": 0.00022711334002006018, + "loss": 1.3386, + "theoretical_loss": 3.41107218599465, + "tokens_seen": 2373255168 + }, + { + "epoch": 0.72, + "learning_rate": 0.00022709729187562688, + "loss": 1.2976, + "theoretical_loss": 3.4110646573965737, + "tokens_seen": 2373320704 + }, + { + "epoch": 0.72, + "learning_rate": 0.00022708124373119358, + "loss": 1.3497, + "theoretical_loss": 3.411057129064594, + "tokens_seen": 2373386240 + }, + { + "epoch": 0.72, + "learning_rate": 0.00022706519558676028, + "loss": 1.306, + "theoretical_loss": 3.4110496009986955, + "tokens_seen": 2373451776 + }, + { + "epoch": 0.72, + "learning_rate": 0.00022704914744232698, + "loss": 1.2232, + "theoretical_loss": 3.4110420731988595, + "tokens_seen": 2373517312 + }, + { + "epoch": 0.72, + "learning_rate": 0.00022703309929789369, + "loss": 1.3173, + "theoretical_loss": 3.411034545665071, + "tokens_seen": 2373582848 + }, + { + "epoch": 0.72, + "learning_rate": 0.00022701705115346039, + "loss": 1.3031, + "theoretical_loss": 3.411027018397312, + "tokens_seen": 2373648384 + }, + { + "epoch": 0.72, + "learning_rate": 0.00022700100300902709, + "loss": 1.2103, + "theoretical_loss": 3.4110194913955665, + "tokens_seen": 2373713920 + }, + { + "epoch": 0.72, + "learning_rate": 0.0002269849548645938, + "loss": 1.2958, + "theoretical_loss": 3.411011964659817, + "tokens_seen": 2373779456 + }, + { + "epoch": 0.72, + "learning_rate": 0.0002269689067201605, + "loss": 1.3099, + "theoretical_loss": 3.4110044381900475, + "tokens_seen": 2373844992 + }, + { + "epoch": 0.72, + "learning_rate": 0.0002269528585757272, + "loss": 1.4637, + "theoretical_loss": 3.410996911986241, + "tokens_seen": 2373910528 + }, + { + "epoch": 0.72, + "learning_rate": 0.0002269368104312939, + "loss": 1.3141, + "theoretical_loss": 3.4109893860483806, + "tokens_seen": 2373976064 + }, + { + "epoch": 0.72, + "learning_rate": 0.0002269207622868606, + "loss": 1.1912, + "theoretical_loss": 3.41098186037645, + "tokens_seen": 2374041600 + }, + { + "epoch": 0.72, + "learning_rate": 0.0002269047141424273, + "loss": 1.1695, + "theoretical_loss": 3.4109743349704322, + "tokens_seen": 2374107136 + }, + { + "epoch": 0.72, + "learning_rate": 0.000226888665997994, + "loss": 1.2573, + "theoretical_loss": 3.41096680983031, + "tokens_seen": 2374172672 + }, + { + "epoch": 0.72, + "learning_rate": 0.0002268726178535607, + "loss": 1.3062, + "theoretical_loss": 3.4109592849560673, + "tokens_seen": 2374238208 + }, + { + "epoch": 0.72, + "learning_rate": 0.0002268565697091274, + "loss": 1.3234, + "theoretical_loss": 3.4109517603476873, + "tokens_seen": 2374303744 + }, + { + "epoch": 0.72, + "learning_rate": 0.0002268405215646941, + "loss": 1.4371, + "theoretical_loss": 3.410944236005153, + "tokens_seen": 2374369280 + }, + { + "epoch": 0.72, + "learning_rate": 0.0002268244734202608, + "loss": 1.3294, + "theoretical_loss": 3.4109367119284477, + "tokens_seen": 2374434816 + }, + { + "epoch": 0.72, + "learning_rate": 0.0002268084252758275, + "loss": 1.2767, + "theoretical_loss": 3.4109291881175547, + "tokens_seen": 2374500352 + }, + { + "epoch": 0.72, + "learning_rate": 0.0002267923771313942, + "loss": 1.3006, + "theoretical_loss": 3.4109216645724576, + "tokens_seen": 2374565888 + }, + { + "epoch": 0.72, + "learning_rate": 0.0002267763289869609, + "loss": 1.2505, + "theoretical_loss": 3.4109141412931394, + "tokens_seen": 2374631424 + }, + { + "epoch": 0.72, + "learning_rate": 0.0002267602808425276, + "loss": 1.3086, + "theoretical_loss": 3.4109066182795833, + "tokens_seen": 2374696960 + }, + { + "epoch": 0.72, + "learning_rate": 0.0002267442326980943, + "loss": 1.2318, + "theoretical_loss": 3.410899095531773, + "tokens_seen": 2374762496 + }, + { + "epoch": 0.72, + "learning_rate": 0.00022672818455366097, + "loss": 1.2229, + "theoretical_loss": 3.410891573049691, + "tokens_seen": 2374828032 + }, + { + "epoch": 0.72, + "learning_rate": 0.00022671213640922767, + "loss": 1.2797, + "theoretical_loss": 3.410884050833321, + "tokens_seen": 2374893568 + }, + { + "epoch": 0.72, + "learning_rate": 0.00022669608826479437, + "loss": 1.3544, + "theoretical_loss": 3.4108765288826466, + "tokens_seen": 2374959104 + }, + { + "epoch": 0.72, + "learning_rate": 0.00022668004012036107, + "loss": 1.2275, + "theoretical_loss": 3.410869007197651, + "tokens_seen": 2375024640 + }, + { + "epoch": 0.72, + "learning_rate": 0.00022666399197592777, + "loss": 1.2702, + "theoretical_loss": 3.410861485778317, + "tokens_seen": 2375090176 + }, + { + "epoch": 0.72, + "learning_rate": 0.00022664794383149447, + "loss": 1.3447, + "theoretical_loss": 3.410853964624628, + "tokens_seen": 2375155712 + }, + { + "epoch": 0.72, + "learning_rate": 0.00022663189568706117, + "loss": 1.3402, + "theoretical_loss": 3.4108464437365678, + "tokens_seen": 2375221248 + }, + { + "epoch": 0.72, + "learning_rate": 0.00022661584754262787, + "loss": 1.2984, + "theoretical_loss": 3.410838923114119, + "tokens_seen": 2375286784 + }, + { + "epoch": 0.72, + "learning_rate": 0.00022659979939819458, + "loss": 1.3528, + "theoretical_loss": 3.4108314027572657, + "tokens_seen": 2375352320 + }, + { + "epoch": 0.72, + "learning_rate": 0.00022658375125376128, + "loss": 1.3474, + "theoretical_loss": 3.4108238826659902, + "tokens_seen": 2375417856 + }, + { + "epoch": 0.72, + "learning_rate": 0.00022656770310932798, + "loss": 1.3153, + "theoretical_loss": 3.4108163628402766, + "tokens_seen": 2375483392 + }, + { + "epoch": 0.72, + "learning_rate": 0.00022655165496489468, + "loss": 1.3268, + "theoretical_loss": 3.4108088432801082, + "tokens_seen": 2375548928 + }, + { + "epoch": 0.72, + "learning_rate": 0.00022653560682046138, + "loss": 1.2425, + "theoretical_loss": 3.4108013239854675, + "tokens_seen": 2375614464 + }, + { + "epoch": 0.72, + "objective/train/docs_used": 865150, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.344789743423462, + "objective/train/theoretical_loss": 3.410793804956339, + "objective/train/tokens_used": 2396140000, + "theoretical_loss": 3.410793804956339, + "tokens_seen": 2375680000 + }, + { + "epoch": 0.72, + "learning_rate": 0.00022651955867602808, + "loss": 1.2995, + "theoretical_loss": 3.410793804956339, + "tokens_seen": 2375680000 + }, + { + "epoch": 0.72, + "learning_rate": 0.00022650351053159478, + "loss": 1.1948, + "theoretical_loss": 3.4107862861927045, + "tokens_seen": 2375745536 + }, + { + "epoch": 0.72, + "learning_rate": 0.00022648746238716148, + "loss": 1.4291, + "theoretical_loss": 3.4107787676945485, + "tokens_seen": 2375811072 + }, + { + "epoch": 0.72, + "learning_rate": 0.00022647141424272818, + "loss": 1.2658, + "theoretical_loss": 3.4107712494618543, + "tokens_seen": 2375876608 + }, + { + "epoch": 0.72, + "learning_rate": 0.00022645536609829488, + "loss": 1.2412, + "theoretical_loss": 3.410763731494604, + "tokens_seen": 2375942144 + }, + { + "epoch": 0.72, + "learning_rate": 0.00022643931795386158, + "loss": 1.2874, + "theoretical_loss": 3.4107562137927827, + "tokens_seen": 2376007680 + }, + { + "epoch": 0.72, + "learning_rate": 0.0002264232698094283, + "loss": 1.3295, + "theoretical_loss": 3.410748696356372, + "tokens_seen": 2376073216 + }, + { + "epoch": 0.72, + "learning_rate": 0.000226407221664995, + "loss": 1.3976, + "theoretical_loss": 3.410741179185356, + "tokens_seen": 2376138752 + }, + { + "epoch": 0.72, + "learning_rate": 0.0002263911735205617, + "loss": 1.2736, + "theoretical_loss": 3.4107336622797186, + "tokens_seen": 2376204288 + }, + { + "epoch": 0.72, + "learning_rate": 0.0002263751253761284, + "loss": 1.2911, + "theoretical_loss": 3.410726145639442, + "tokens_seen": 2376269824 + }, + { + "epoch": 0.72, + "learning_rate": 0.0002263590772316951, + "loss": 1.3956, + "theoretical_loss": 3.41071862926451, + "tokens_seen": 2376335360 + }, + { + "epoch": 0.72, + "learning_rate": 0.0002263430290872618, + "loss": 1.3121, + "theoretical_loss": 3.410711113154906, + "tokens_seen": 2376400896 + }, + { + "epoch": 0.72, + "learning_rate": 0.0002263269809428285, + "loss": 1.233, + "theoretical_loss": 3.4107035973106132, + "tokens_seen": 2376466432 + }, + { + "epoch": 0.72, + "learning_rate": 0.00022631093279839521, + "loss": 1.2827, + "theoretical_loss": 3.410696081731615, + "tokens_seen": 2376531968 + }, + { + "epoch": 0.72, + "learning_rate": 0.00022629488465396191, + "loss": 1.2895, + "theoretical_loss": 3.4106885664178948, + "tokens_seen": 2376597504 + }, + { + "epoch": 0.72, + "learning_rate": 0.00022627883650952861, + "loss": 1.2708, + "theoretical_loss": 3.4106810513694357, + "tokens_seen": 2376663040 + }, + { + "epoch": 0.72, + "learning_rate": 0.00022626278836509532, + "loss": 1.2237, + "theoretical_loss": 3.410673536586221, + "tokens_seen": 2376728576 + }, + { + "epoch": 0.72, + "learning_rate": 0.00022624674022066202, + "loss": 1.4008, + "theoretical_loss": 3.4106660220682343, + "tokens_seen": 2376794112 + }, + { + "epoch": 0.72, + "learning_rate": 0.00022623069207622872, + "loss": 1.4197, + "theoretical_loss": 3.4106585078154583, + "tokens_seen": 2376859648 + }, + { + "epoch": 0.72, + "learning_rate": 0.00022621464393179542, + "loss": 1.2678, + "theoretical_loss": 3.4106509938278773, + "tokens_seen": 2376925184 + }, + { + "epoch": 0.72, + "learning_rate": 0.00022619859578736212, + "loss": 1.376, + "theoretical_loss": 3.410643480105474, + "tokens_seen": 2376990720 + }, + { + "epoch": 0.72, + "learning_rate": 0.00022618254764292882, + "loss": 1.4076, + "theoretical_loss": 3.4106359666482318, + "tokens_seen": 2377056256 + }, + { + "epoch": 0.72, + "learning_rate": 0.00022616649949849552, + "loss": 1.2433, + "theoretical_loss": 3.410628453456134, + "tokens_seen": 2377121792 + }, + { + "epoch": 0.72, + "learning_rate": 0.00022615045135406222, + "loss": 1.3566, + "theoretical_loss": 3.4106209405291645, + "tokens_seen": 2377187328 + }, + { + "epoch": 0.72, + "learning_rate": 0.00022613440320962892, + "loss": 1.2084, + "theoretical_loss": 3.4106134278673057, + "tokens_seen": 2377252864 + }, + { + "epoch": 0.72, + "learning_rate": 0.00022611835506519562, + "loss": 1.2769, + "theoretical_loss": 3.4106059154705415, + "tokens_seen": 2377318400 + }, + { + "epoch": 0.72, + "learning_rate": 0.0002261023069207623, + "loss": 1.2971, + "theoretical_loss": 3.410598403338855, + "tokens_seen": 2377383936 + }, + { + "epoch": 0.72, + "learning_rate": 0.000226086258776329, + "loss": 1.3674, + "theoretical_loss": 3.41059089147223, + "tokens_seen": 2377449472 + }, + { + "epoch": 0.72, + "learning_rate": 0.0002260702106318957, + "loss": 1.2527, + "theoretical_loss": 3.410583379870649, + "tokens_seen": 2377515008 + }, + { + "epoch": 0.72, + "learning_rate": 0.0002260541624874624, + "loss": 1.2413, + "theoretical_loss": 3.410575868534096, + "tokens_seen": 2377580544 + }, + { + "epoch": 0.72, + "learning_rate": 0.0002260381143430291, + "loss": 1.2511, + "theoretical_loss": 3.4105683574625543, + "tokens_seen": 2377646080 + }, + { + "epoch": 0.72, + "learning_rate": 0.0002260220661985958, + "loss": 1.3171, + "theoretical_loss": 3.4105608466560073, + "tokens_seen": 2377711616 + }, + { + "epoch": 0.72, + "learning_rate": 0.0002260060180541625, + "loss": 1.2734, + "theoretical_loss": 3.410553336114438, + "tokens_seen": 2377777152 + }, + { + "epoch": 0.72, + "learning_rate": 0.0002259899699097292, + "loss": 1.26, + "theoretical_loss": 3.41054582583783, + "tokens_seen": 2377842688 + }, + { + "epoch": 0.72, + "learning_rate": 0.0002259739217652959, + "loss": 1.2132, + "theoretical_loss": 3.4105383158261664, + "tokens_seen": 2377908224 + }, + { + "epoch": 0.72, + "learning_rate": 0.0002259578736208626, + "loss": 1.3539, + "theoretical_loss": 3.4105308060794313, + "tokens_seen": 2377973760 + }, + { + "epoch": 0.72, + "learning_rate": 0.0002259418254764293, + "loss": 1.3233, + "theoretical_loss": 3.410523296597607, + "tokens_seen": 2378039296 + }, + { + "epoch": 0.72, + "learning_rate": 0.000225925777331996, + "loss": 1.3024, + "theoretical_loss": 3.4105157873806773, + "tokens_seen": 2378104832 + }, + { + "epoch": 0.72, + "learning_rate": 0.0002259097291875627, + "loss": 1.327, + "theoretical_loss": 3.410508278428626, + "tokens_seen": 2378170368 + }, + { + "epoch": 0.72, + "learning_rate": 0.0002258936810431294, + "loss": 1.3759, + "theoretical_loss": 3.410500769741436, + "tokens_seen": 2378235904 + }, + { + "epoch": 0.72, + "learning_rate": 0.0002258776328986961, + "loss": 1.3035, + "theoretical_loss": 3.41049326131909, + "tokens_seen": 2378301440 + }, + { + "epoch": 0.72, + "learning_rate": 0.0002258615847542628, + "loss": 1.3189, + "theoretical_loss": 3.4104857531615727, + "tokens_seen": 2378366976 + }, + { + "epoch": 0.72, + "learning_rate": 0.0002258455366098295, + "loss": 1.2218, + "theoretical_loss": 3.410478245268867, + "tokens_seen": 2378432512 + }, + { + "epoch": 0.72, + "learning_rate": 0.0002258294884653962, + "loss": 1.3155, + "theoretical_loss": 3.410470737640956, + "tokens_seen": 2378498048 + }, + { + "epoch": 0.72, + "learning_rate": 0.0002258134403209629, + "loss": 1.2693, + "theoretical_loss": 3.410463230277823, + "tokens_seen": 2378563584 + }, + { + "epoch": 0.72, + "learning_rate": 0.0002257973921765296, + "loss": 1.3186, + "theoretical_loss": 3.4104557231794512, + "tokens_seen": 2378629120 + }, + { + "epoch": 0.72, + "learning_rate": 0.0002257813440320963, + "loss": 1.2207, + "theoretical_loss": 3.410448216345825, + "tokens_seen": 2378694656 + }, + { + "epoch": 0.72, + "learning_rate": 0.000225765295887663, + "loss": 1.2994, + "theoretical_loss": 3.4104407097769265, + "tokens_seen": 2378760192 + }, + { + "epoch": 0.72, + "learning_rate": 0.0002257492477432297, + "loss": 1.2854, + "theoretical_loss": 3.4104332034727403, + "tokens_seen": 2378825728 + }, + { + "epoch": 0.72, + "learning_rate": 0.0002257331995987964, + "loss": 1.3338, + "theoretical_loss": 3.4104256974332485, + "tokens_seen": 2378891264 + }, + { + "epoch": 0.72, + "objective/train/docs_used": 866449, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.110121488571167, + "objective/train/theoretical_loss": 3.4104181916584357, + "objective/train/tokens_used": 2399416800, + "theoretical_loss": 3.4104181916584357, + "tokens_seen": 2378956800 + }, + { + "epoch": 0.72, + "learning_rate": 0.00022571715145436308, + "loss": 1.2131, + "theoretical_loss": 3.4104181916584357, + "tokens_seen": 2378956800 + }, + { + "epoch": 0.72, + "learning_rate": 0.00022570110330992978, + "loss": 1.2734, + "theoretical_loss": 3.410410686148284, + "tokens_seen": 2379022336 + }, + { + "epoch": 0.72, + "learning_rate": 0.00022568505516549648, + "loss": 1.311, + "theoretical_loss": 3.4104031809027777, + "tokens_seen": 2379087872 + }, + { + "epoch": 0.72, + "learning_rate": 0.00022566900702106318, + "loss": 1.2727, + "theoretical_loss": 3.4103956759219, + "tokens_seen": 2379153408 + }, + { + "epoch": 0.72, + "learning_rate": 0.00022565295887662989, + "loss": 1.2503, + "theoretical_loss": 3.4103881712056343, + "tokens_seen": 2379218944 + }, + { + "epoch": 0.72, + "learning_rate": 0.00022563691073219659, + "loss": 1.2016, + "theoretical_loss": 3.410380666753964, + "tokens_seen": 2379284480 + }, + { + "epoch": 0.72, + "learning_rate": 0.0002256208625877633, + "loss": 1.4293, + "theoretical_loss": 3.4103731625668723, + "tokens_seen": 2379350016 + }, + { + "epoch": 0.72, + "learning_rate": 0.00022560481444333, + "loss": 1.3259, + "theoretical_loss": 3.4103656586443423, + "tokens_seen": 2379415552 + }, + { + "epoch": 0.72, + "learning_rate": 0.0002255887662988967, + "loss": 1.3, + "theoretical_loss": 3.410358154986358, + "tokens_seen": 2379481088 + }, + { + "epoch": 0.72, + "learning_rate": 0.0002255727181544634, + "loss": 1.3385, + "theoretical_loss": 3.4103506515929025, + "tokens_seen": 2379546624 + }, + { + "epoch": 0.72, + "learning_rate": 0.0002255566700100301, + "loss": 1.446, + "theoretical_loss": 3.4103431484639595, + "tokens_seen": 2379612160 + }, + { + "epoch": 0.72, + "learning_rate": 0.0002255406218655968, + "loss": 1.361, + "theoretical_loss": 3.4103356455995115, + "tokens_seen": 2379677696 + }, + { + "epoch": 0.72, + "learning_rate": 0.0002255245737211635, + "loss": 1.4015, + "theoretical_loss": 3.410328142999543, + "tokens_seen": 2379743232 + }, + { + "epoch": 0.72, + "learning_rate": 0.0002255085255767302, + "loss": 1.3128, + "theoretical_loss": 3.410320640664037, + "tokens_seen": 2379808768 + }, + { + "epoch": 0.72, + "learning_rate": 0.0002254924774322969, + "loss": 1.2242, + "theoretical_loss": 3.4103131385929766, + "tokens_seen": 2379874304 + }, + { + "epoch": 0.72, + "learning_rate": 0.0002254764292878636, + "loss": 1.3842, + "theoretical_loss": 3.410305636786345, + "tokens_seen": 2379939840 + }, + { + "epoch": 0.72, + "learning_rate": 0.0002254603811434303, + "loss": 1.3318, + "theoretical_loss": 3.4102981352441266, + "tokens_seen": 2380005376 + }, + { + "epoch": 0.72, + "learning_rate": 0.000225444332998997, + "loss": 1.3057, + "theoretical_loss": 3.410290633966304, + "tokens_seen": 2380070912 + }, + { + "epoch": 0.72, + "learning_rate": 0.0002254282848545637, + "loss": 1.4178, + "theoretical_loss": 3.410283132952861, + "tokens_seen": 2380136448 + }, + { + "epoch": 0.72, + "learning_rate": 0.0002254122367101304, + "loss": 1.2815, + "theoretical_loss": 3.4102756322037804, + "tokens_seen": 2380201984 + }, + { + "epoch": 0.72, + "learning_rate": 0.0002253961885656971, + "loss": 1.2334, + "theoretical_loss": 3.4102681317190457, + "tokens_seen": 2380267520 + }, + { + "epoch": 0.72, + "learning_rate": 0.0002253801404212638, + "loss": 1.228, + "theoretical_loss": 3.4102606314986414, + "tokens_seen": 2380333056 + }, + { + "epoch": 0.72, + "learning_rate": 0.0002253640922768305, + "loss": 1.2812, + "theoretical_loss": 3.4102531315425497, + "tokens_seen": 2380398592 + }, + { + "epoch": 0.72, + "learning_rate": 0.0002253480441323972, + "loss": 1.2169, + "theoretical_loss": 3.4102456318507546, + "tokens_seen": 2380464128 + }, + { + "epoch": 0.72, + "learning_rate": 0.00022533199598796387, + "loss": 1.3888, + "theoretical_loss": 3.410238132423239, + "tokens_seen": 2380529664 + }, + { + "epoch": 0.72, + "learning_rate": 0.00022531594784353057, + "loss": 1.3024, + "theoretical_loss": 3.410230633259987, + "tokens_seen": 2380595200 + }, + { + "epoch": 0.72, + "learning_rate": 0.00022529989969909727, + "loss": 1.3294, + "theoretical_loss": 3.4102231343609812, + "tokens_seen": 2380660736 + }, + { + "epoch": 0.72, + "learning_rate": 0.00022528385155466397, + "loss": 1.2247, + "theoretical_loss": 3.410215635726206, + "tokens_seen": 2380726272 + }, + { + "epoch": 0.72, + "learning_rate": 0.00022526780341023067, + "loss": 1.2872, + "theoretical_loss": 3.410208137355644, + "tokens_seen": 2380791808 + }, + { + "epoch": 0.72, + "learning_rate": 0.00022525175526579737, + "loss": 1.2126, + "theoretical_loss": 3.410200639249279, + "tokens_seen": 2380857344 + }, + { + "epoch": 0.72, + "learning_rate": 0.00022523570712136407, + "loss": 1.3035, + "theoretical_loss": 3.4101931414070945, + "tokens_seen": 2380922880 + }, + { + "epoch": 0.72, + "learning_rate": 0.00022521965897693083, + "loss": 1.3438, + "theoretical_loss": 3.4101856438290734, + "tokens_seen": 2380988416 + }, + { + "epoch": 0.72, + "learning_rate": 0.00022520361083249753, + "loss": 1.3313, + "theoretical_loss": 3.4101781465151997, + "tokens_seen": 2381053952 + }, + { + "epoch": 0.72, + "learning_rate": 0.00022518756268806423, + "loss": 1.2726, + "theoretical_loss": 3.4101706494654564, + "tokens_seen": 2381119488 + }, + { + "epoch": 0.72, + "learning_rate": 0.00022517151454363093, + "loss": 1.3835, + "theoretical_loss": 3.4101631526798273, + "tokens_seen": 2381185024 + }, + { + "epoch": 0.72, + "learning_rate": 0.00022515546639919763, + "loss": 1.3775, + "theoretical_loss": 3.4101556561582957, + "tokens_seen": 2381250560 + }, + { + "epoch": 0.72, + "learning_rate": 0.00022513941825476433, + "loss": 1.2757, + "theoretical_loss": 3.410148159900845, + "tokens_seen": 2381316096 + }, + { + "epoch": 0.72, + "learning_rate": 0.00022512337011033103, + "loss": 1.3823, + "theoretical_loss": 3.410140663907458, + "tokens_seen": 2381381632 + }, + { + "epoch": 0.72, + "learning_rate": 0.00022510732196589773, + "loss": 1.2304, + "theoretical_loss": 3.4101331681781195, + "tokens_seen": 2381447168 + }, + { + "epoch": 0.72, + "learning_rate": 0.0002250912738214644, + "loss": 1.2338, + "theoretical_loss": 3.4101256727128115, + "tokens_seen": 2381512704 + }, + { + "epoch": 0.72, + "learning_rate": 0.0002250752256770311, + "loss": 1.2854, + "theoretical_loss": 3.410118177511519, + "tokens_seen": 2381578240 + }, + { + "epoch": 0.72, + "learning_rate": 0.0002250591775325978, + "loss": 1.2499, + "theoretical_loss": 3.410110682574224, + "tokens_seen": 2381643776 + }, + { + "epoch": 0.72, + "learning_rate": 0.0002250431293881645, + "loss": 1.3823, + "theoretical_loss": 3.4101031879009103, + "tokens_seen": 2381709312 + }, + { + "epoch": 0.72, + "learning_rate": 0.0002250270812437312, + "loss": 1.2173, + "theoretical_loss": 3.4100956934915616, + "tokens_seen": 2381774848 + }, + { + "epoch": 0.72, + "learning_rate": 0.0002250110330992979, + "loss": 1.327, + "theoretical_loss": 3.4100881993461614, + "tokens_seen": 2381840384 + }, + { + "epoch": 0.72, + "learning_rate": 0.0002249949849548646, + "loss": 1.2594, + "theoretical_loss": 3.4100807054646927, + "tokens_seen": 2381905920 + }, + { + "epoch": 0.72, + "learning_rate": 0.0002249789368104313, + "loss": 1.3412, + "theoretical_loss": 3.4100732118471395, + "tokens_seen": 2381971456 + }, + { + "epoch": 0.72, + "learning_rate": 0.000224962888665998, + "loss": 1.2367, + "theoretical_loss": 3.410065718493485, + "tokens_seen": 2382036992 + }, + { + "epoch": 0.72, + "learning_rate": 0.0002249468405215647, + "loss": 1.2876, + "theoretical_loss": 3.410058225403713, + "tokens_seen": 2382102528 + }, + { + "epoch": 0.72, + "learning_rate": 0.00022493079237713141, + "loss": 1.2879, + "theoretical_loss": 3.410050732577806, + "tokens_seen": 2382168064 + }, + { + "epoch": 0.72, + "objective/train/docs_used": 867590, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.3140168190002441, + "objective/train/theoretical_loss": 3.410043240015748, + "objective/train/tokens_used": 2402693600, + "theoretical_loss": 3.410043240015748, + "tokens_seen": 2382233600 + }, + { + "epoch": 0.72, + "learning_rate": 0.00022491474423269811, + "loss": 1.2928, + "theoretical_loss": 3.410043240015748, + "tokens_seen": 2382233600 + }, + { + "epoch": 0.72, + "learning_rate": 0.00022489869608826482, + "loss": 1.2593, + "theoretical_loss": 3.4100357477175227, + "tokens_seen": 2382299136 + }, + { + "epoch": 0.72, + "learning_rate": 0.00022488264794383152, + "loss": 1.434, + "theoretical_loss": 3.4100282556831134, + "tokens_seen": 2382364672 + }, + { + "epoch": 0.72, + "learning_rate": 0.00022486659979939822, + "loss": 1.4258, + "theoretical_loss": 3.4100207639125033, + "tokens_seen": 2382430208 + }, + { + "epoch": 0.72, + "learning_rate": 0.00022485055165496492, + "loss": 1.3562, + "theoretical_loss": 3.4100132724056764, + "tokens_seen": 2382495744 + }, + { + "epoch": 0.72, + "learning_rate": 0.00022483450351053162, + "loss": 1.2049, + "theoretical_loss": 3.4100057811626154, + "tokens_seen": 2382561280 + }, + { + "epoch": 0.72, + "learning_rate": 0.00022481845536609832, + "loss": 1.2559, + "theoretical_loss": 3.4099982901833044, + "tokens_seen": 2382626816 + }, + { + "epoch": 0.72, + "learning_rate": 0.00022480240722166502, + "loss": 1.3164, + "theoretical_loss": 3.4099907994677263, + "tokens_seen": 2382692352 + }, + { + "epoch": 0.72, + "learning_rate": 0.00022478635907723172, + "loss": 1.3491, + "theoretical_loss": 3.409983309015865, + "tokens_seen": 2382757888 + }, + { + "epoch": 0.72, + "learning_rate": 0.00022477031093279842, + "loss": 1.2814, + "theoretical_loss": 3.409975818827704, + "tokens_seen": 2382823424 + }, + { + "epoch": 0.72, + "learning_rate": 0.00022475426278836512, + "loss": 1.2206, + "theoretical_loss": 3.4099683289032265, + "tokens_seen": 2382888960 + }, + { + "epoch": 0.72, + "learning_rate": 0.00022473821464393182, + "loss": 1.3115, + "theoretical_loss": 3.409960839242416, + "tokens_seen": 2382954496 + }, + { + "epoch": 0.72, + "learning_rate": 0.00022472216649949852, + "loss": 1.2878, + "theoretical_loss": 3.409953349845256, + "tokens_seen": 2383020032 + }, + { + "epoch": 0.72, + "learning_rate": 0.0002247061183550652, + "loss": 1.3018, + "theoretical_loss": 3.4099458607117303, + "tokens_seen": 2383085568 + }, + { + "epoch": 0.72, + "learning_rate": 0.0002246900702106319, + "loss": 1.286, + "theoretical_loss": 3.409938371841822, + "tokens_seen": 2383151104 + }, + { + "epoch": 0.72, + "learning_rate": 0.0002246740220661986, + "loss": 1.3758, + "theoretical_loss": 3.4099308832355146, + "tokens_seen": 2383216640 + }, + { + "epoch": 0.72, + "learning_rate": 0.0002246579739217653, + "loss": 1.3589, + "theoretical_loss": 3.4099233948927914, + "tokens_seen": 2383282176 + }, + { + "epoch": 0.72, + "learning_rate": 0.000224641925777332, + "loss": 1.2713, + "theoretical_loss": 3.4099159068136364, + "tokens_seen": 2383347712 + }, + { + "epoch": 0.72, + "learning_rate": 0.0002246258776328987, + "loss": 1.313, + "theoretical_loss": 3.4099084189980324, + "tokens_seen": 2383413248 + }, + { + "epoch": 0.72, + "learning_rate": 0.0002246098294884654, + "loss": 1.2139, + "theoretical_loss": 3.4099009314459634, + "tokens_seen": 2383478784 + }, + { + "epoch": 0.72, + "learning_rate": 0.0002245937813440321, + "loss": 1.2731, + "theoretical_loss": 3.409893444157413, + "tokens_seen": 2383544320 + }, + { + "epoch": 0.72, + "learning_rate": 0.0002245777331995988, + "loss": 1.3844, + "theoretical_loss": 3.409885957132364, + "tokens_seen": 2383609856 + }, + { + "epoch": 0.72, + "learning_rate": 0.0002245616850551655, + "loss": 1.3067, + "theoretical_loss": 3.409878470370801, + "tokens_seen": 2383675392 + }, + { + "epoch": 0.72, + "learning_rate": 0.0002245456369107322, + "loss": 1.2487, + "theoretical_loss": 3.4098709838727057, + "tokens_seen": 2383740928 + }, + { + "epoch": 0.72, + "learning_rate": 0.0002245295887662989, + "loss": 1.2846, + "theoretical_loss": 3.4098634976380633, + "tokens_seen": 2383806464 + }, + { + "epoch": 0.72, + "learning_rate": 0.0002245135406218656, + "loss": 1.3468, + "theoretical_loss": 3.409856011666857, + "tokens_seen": 2383872000 + }, + { + "epoch": 0.72, + "learning_rate": 0.0002244974924774323, + "loss": 1.3228, + "theoretical_loss": 3.4098485259590694, + "tokens_seen": 2383937536 + }, + { + "epoch": 0.72, + "learning_rate": 0.000224481444332999, + "loss": 1.2813, + "theoretical_loss": 3.409841040514684, + "tokens_seen": 2384003072 + }, + { + "epoch": 0.72, + "learning_rate": 0.0002244653961885657, + "loss": 1.358, + "theoretical_loss": 3.4098335553336856, + "tokens_seen": 2384068608 + }, + { + "epoch": 0.72, + "learning_rate": 0.0002244493480441324, + "loss": 1.283, + "theoretical_loss": 3.409826070416057, + "tokens_seen": 2384134144 + }, + { + "epoch": 0.72, + "learning_rate": 0.0002244332998996991, + "loss": 1.3199, + "theoretical_loss": 3.4098185857617813, + "tokens_seen": 2384199680 + }, + { + "epoch": 0.72, + "learning_rate": 0.0002244172517552658, + "loss": 1.347, + "theoretical_loss": 3.4098111013708423, + "tokens_seen": 2384265216 + }, + { + "epoch": 0.72, + "learning_rate": 0.0002244012036108325, + "loss": 1.229, + "theoretical_loss": 3.409803617243224, + "tokens_seen": 2384330752 + }, + { + "epoch": 0.72, + "learning_rate": 0.0002243851554663992, + "loss": 1.355, + "theoretical_loss": 3.4097961333789084, + "tokens_seen": 2384396288 + }, + { + "epoch": 0.72, + "learning_rate": 0.0002243691073219659, + "loss": 1.3163, + "theoretical_loss": 3.4097886497778807, + "tokens_seen": 2384461824 + }, + { + "epoch": 0.72, + "learning_rate": 0.0002243530591775326, + "loss": 1.2527, + "theoretical_loss": 3.4097811664401236, + "tokens_seen": 2384527360 + }, + { + "epoch": 0.72, + "learning_rate": 0.0002243370110330993, + "loss": 1.2437, + "theoretical_loss": 3.4097736833656205, + "tokens_seen": 2384592896 + }, + { + "epoch": 0.72, + "learning_rate": 0.000224320962888666, + "loss": 1.2825, + "theoretical_loss": 3.409766200554355, + "tokens_seen": 2384658432 + }, + { + "epoch": 0.72, + "learning_rate": 0.00022430491474423268, + "loss": 1.3608, + "theoretical_loss": 3.409758718006311, + "tokens_seen": 2384723968 + }, + { + "epoch": 0.72, + "learning_rate": 0.00022428886659979939, + "loss": 1.3214, + "theoretical_loss": 3.4097512357214717, + "tokens_seen": 2384789504 + }, + { + "epoch": 0.72, + "learning_rate": 0.00022427281845536609, + "loss": 1.3149, + "theoretical_loss": 3.4097437536998205, + "tokens_seen": 2384855040 + }, + { + "epoch": 0.72, + "learning_rate": 0.0002242567703109328, + "loss": 1.4516, + "theoretical_loss": 3.409736271941341, + "tokens_seen": 2384920576 + }, + { + "epoch": 0.72, + "learning_rate": 0.0002242407221664995, + "loss": 1.2888, + "theoretical_loss": 3.4097287904460165, + "tokens_seen": 2384986112 + }, + { + "epoch": 0.72, + "learning_rate": 0.0002242246740220662, + "loss": 1.3449, + "theoretical_loss": 3.409721309213831, + "tokens_seen": 2385051648 + }, + { + "epoch": 0.72, + "learning_rate": 0.0002242086258776329, + "loss": 1.2357, + "theoretical_loss": 3.409713828244768, + "tokens_seen": 2385117184 + }, + { + "epoch": 0.72, + "learning_rate": 0.0002241925777331996, + "loss": 1.2618, + "theoretical_loss": 3.4097063475388105, + "tokens_seen": 2385182720 + }, + { + "epoch": 0.72, + "learning_rate": 0.0002241765295887663, + "loss": 1.3028, + "theoretical_loss": 3.4096988670959427, + "tokens_seen": 2385248256 + }, + { + "epoch": 0.72, + "learning_rate": 0.000224160481444333, + "loss": 1.2891, + "theoretical_loss": 3.409691386916147, + "tokens_seen": 2385313792 + }, + { + "epoch": 0.72, + "learning_rate": 0.0002241444332998997, + "loss": 1.4534, + "theoretical_loss": 3.4096839069994083, + "tokens_seen": 2385379328 + }, + { + "epoch": 0.72, + "learning_rate": 0.0002241283851554664, + "loss": 1.3254, + "theoretical_loss": 3.409676427345709, + "tokens_seen": 2385444864 + }, + { + "epoch": 0.72, + "objective/train/docs_used": 868828, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.2849719524383545, + "objective/train/theoretical_loss": 3.4096689479550335, + "objective/train/tokens_used": 2405970400, + "theoretical_loss": 3.4096689479550335, + "tokens_seen": 2385510400 + }, + { + "epoch": 0.72, + "learning_rate": 0.0002241123370110331, + "loss": 1.3269, + "theoretical_loss": 3.4096689479550335, + "tokens_seen": 2385510400 + }, + { + "epoch": 0.72, + "learning_rate": 0.0002240962888665998, + "loss": 1.2642, + "theoretical_loss": 3.4096614688273648, + "tokens_seen": 2385575936 + }, + { + "epoch": 0.72, + "learning_rate": 0.0002240802407221665, + "loss": 1.3987, + "theoretical_loss": 3.4096539899626865, + "tokens_seen": 2385641472 + }, + { + "epoch": 0.72, + "learning_rate": 0.0002240641925777332, + "loss": 1.2947, + "theoretical_loss": 3.4096465113609824, + "tokens_seen": 2385707008 + }, + { + "epoch": 0.72, + "learning_rate": 0.0002240481444332999, + "loss": 1.2538, + "theoretical_loss": 3.4096390330222355, + "tokens_seen": 2385772544 + }, + { + "epoch": 0.72, + "learning_rate": 0.0002240320962888666, + "loss": 1.3376, + "theoretical_loss": 3.4096315549464298, + "tokens_seen": 2385838080 + }, + { + "epoch": 0.72, + "learning_rate": 0.00022401604814443332, + "loss": 1.2743, + "theoretical_loss": 3.4096240771335484, + "tokens_seen": 2385903616 + }, + { + "epoch": 0.72, + "learning_rate": 0.00022400000000000002, + "loss": 1.3306, + "theoretical_loss": 3.4096165995835754, + "tokens_seen": 2385969152 + }, + { + "epoch": 0.72, + "learning_rate": 0.00022398395185556672, + "loss": 1.3209, + "theoretical_loss": 3.409609122296494, + "tokens_seen": 2386034688 + }, + { + "epoch": 0.72, + "learning_rate": 0.00022396790371113343, + "loss": 1.3794, + "theoretical_loss": 3.4096016452722875, + "tokens_seen": 2386100224 + }, + { + "epoch": 0.72, + "learning_rate": 0.00022395185556670013, + "loss": 1.3375, + "theoretical_loss": 3.4095941685109397, + "tokens_seen": 2386165760 + }, + { + "epoch": 0.72, + "learning_rate": 0.00022393580742226683, + "loss": 1.4384, + "theoretical_loss": 3.4095866920124345, + "tokens_seen": 2386231296 + }, + { + "epoch": 0.72, + "learning_rate": 0.00022391975927783353, + "loss": 1.3528, + "theoretical_loss": 3.4095792157767546, + "tokens_seen": 2386296832 + }, + { + "epoch": 0.72, + "learning_rate": 0.00022390371113340023, + "loss": 1.2556, + "theoretical_loss": 3.4095717398038845, + "tokens_seen": 2386362368 + }, + { + "epoch": 0.72, + "learning_rate": 0.00022388766298896693, + "loss": 1.2648, + "theoretical_loss": 3.4095642640938073, + "tokens_seen": 2386427904 + }, + { + "epoch": 0.72, + "learning_rate": 0.00022387161484453363, + "loss": 1.3096, + "theoretical_loss": 3.409556788646506, + "tokens_seen": 2386493440 + }, + { + "epoch": 0.72, + "learning_rate": 0.00022385556670010033, + "loss": 1.3235, + "theoretical_loss": 3.4095493134619654, + "tokens_seen": 2386558976 + }, + { + "epoch": 0.72, + "learning_rate": 0.00022383951855566703, + "loss": 1.3176, + "theoretical_loss": 3.409541838540168, + "tokens_seen": 2386624512 + }, + { + "epoch": 0.72, + "learning_rate": 0.00022382347041123373, + "loss": 1.2828, + "theoretical_loss": 3.4095343638810975, + "tokens_seen": 2386690048 + }, + { + "epoch": 0.72, + "learning_rate": 0.00022380742226680043, + "loss": 1.341, + "theoretical_loss": 3.4095268894847375, + "tokens_seen": 2386755584 + }, + { + "epoch": 0.72, + "learning_rate": 0.00022379137412236713, + "loss": 1.3945, + "theoretical_loss": 3.409519415351072, + "tokens_seen": 2386821120 + }, + { + "epoch": 0.72, + "learning_rate": 0.00022377532597793383, + "loss": 1.3315, + "theoretical_loss": 3.4095119414800843, + "tokens_seen": 2386886656 + }, + { + "epoch": 0.72, + "learning_rate": 0.00022375927783350053, + "loss": 1.2866, + "theoretical_loss": 3.4095044678717574, + "tokens_seen": 2386952192 + }, + { + "epoch": 0.72, + "learning_rate": 0.00022374322968906723, + "loss": 1.3086, + "theoretical_loss": 3.409496994526076, + "tokens_seen": 2387017728 + }, + { + "epoch": 0.72, + "learning_rate": 0.00022372718154463393, + "loss": 1.3911, + "theoretical_loss": 3.4094895214430228, + "tokens_seen": 2387083264 + }, + { + "epoch": 0.72, + "learning_rate": 0.00022371113340020064, + "loss": 1.3235, + "theoretical_loss": 3.409482048622581, + "tokens_seen": 2387148800 + }, + { + "epoch": 0.72, + "learning_rate": 0.00022369508525576734, + "loss": 1.3337, + "theoretical_loss": 3.4094745760647354, + "tokens_seen": 2387214336 + }, + { + "epoch": 0.72, + "learning_rate": 0.000223679037111334, + "loss": 1.3576, + "theoretical_loss": 3.409467103769469, + "tokens_seen": 2387279872 + }, + { + "epoch": 0.72, + "learning_rate": 0.0002236629889669007, + "loss": 1.34, + "theoretical_loss": 3.4094596317367647, + "tokens_seen": 2387345408 + }, + { + "epoch": 0.72, + "learning_rate": 0.0002236469408224674, + "loss": 1.3322, + "theoretical_loss": 3.409452159966607, + "tokens_seen": 2387410944 + }, + { + "epoch": 0.72, + "learning_rate": 0.0002236308926780341, + "loss": 1.3757, + "theoretical_loss": 3.409444688458979, + "tokens_seen": 2387476480 + }, + { + "epoch": 0.72, + "learning_rate": 0.0002236148445336008, + "loss": 1.3266, + "theoretical_loss": 3.4094372172138643, + "tokens_seen": 2387542016 + }, + { + "epoch": 0.72, + "learning_rate": 0.0002235987963891675, + "loss": 1.32, + "theoretical_loss": 3.4094297462312464, + "tokens_seen": 2387607552 + }, + { + "epoch": 0.72, + "learning_rate": 0.0002235827482447342, + "loss": 1.2049, + "theoretical_loss": 3.409422275511109, + "tokens_seen": 2387673088 + }, + { + "epoch": 0.72, + "learning_rate": 0.00022356670010030091, + "loss": 1.2711, + "theoretical_loss": 3.409414805053436, + "tokens_seen": 2387738624 + }, + { + "epoch": 0.72, + "learning_rate": 0.00022355065195586761, + "loss": 1.2969, + "theoretical_loss": 3.4094073348582103, + "tokens_seen": 2387804160 + }, + { + "epoch": 0.72, + "learning_rate": 0.00022353460381143432, + "loss": 1.2287, + "theoretical_loss": 3.409399864925416, + "tokens_seen": 2387869696 + }, + { + "epoch": 0.72, + "learning_rate": 0.00022351855566700102, + "loss": 1.3584, + "theoretical_loss": 3.4093923952550367, + "tokens_seen": 2387935232 + }, + { + "epoch": 0.72, + "learning_rate": 0.00022350250752256772, + "loss": 1.1998, + "theoretical_loss": 3.409384925847055, + "tokens_seen": 2388000768 + }, + { + "epoch": 0.72, + "learning_rate": 0.00022348645937813442, + "loss": 1.3507, + "theoretical_loss": 3.4093774567014563, + "tokens_seen": 2388066304 + }, + { + "epoch": 0.72, + "learning_rate": 0.00022347041123370112, + "loss": 1.4326, + "theoretical_loss": 3.4093699878182226, + "tokens_seen": 2388131840 + }, + { + "epoch": 0.72, + "learning_rate": 0.00022345436308926782, + "loss": 1.249, + "theoretical_loss": 3.4093625191973382, + "tokens_seen": 2388197376 + }, + { + "epoch": 0.72, + "learning_rate": 0.00022343831494483452, + "loss": 1.2602, + "theoretical_loss": 3.4093550508387866, + "tokens_seen": 2388262912 + }, + { + "epoch": 0.72, + "learning_rate": 0.00022342226680040122, + "loss": 1.3401, + "theoretical_loss": 3.4093475827425515, + "tokens_seen": 2388328448 + }, + { + "epoch": 0.72, + "learning_rate": 0.00022340621865596792, + "loss": 1.3702, + "theoretical_loss": 3.4093401149086158, + "tokens_seen": 2388393984 + }, + { + "epoch": 0.72, + "learning_rate": 0.00022339017051153462, + "loss": 1.273, + "theoretical_loss": 3.4093326473369636, + "tokens_seen": 2388459520 + }, + { + "epoch": 0.72, + "learning_rate": 0.00022337412236710132, + "loss": 1.3199, + "theoretical_loss": 3.409325180027579, + "tokens_seen": 2388525056 + }, + { + "epoch": 0.72, + "learning_rate": 0.00022335807422266802, + "loss": 1.2529, + "theoretical_loss": 3.4093177129804446, + "tokens_seen": 2388590592 + }, + { + "epoch": 0.72, + "learning_rate": 0.00022334202607823472, + "loss": 1.3978, + "theoretical_loss": 3.4093102461955445, + "tokens_seen": 2388656128 + }, + { + "epoch": 0.72, + "learning_rate": 0.00022332597793380142, + "loss": 1.3655, + "theoretical_loss": 3.4093027796728625, + "tokens_seen": 2388721664 + }, + { + "epoch": 0.72, + "objective/train/docs_used": 870071, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.3472888469696045, + "objective/train/theoretical_loss": 3.409295313412382, + "objective/train/tokens_used": 2409247200, + "theoretical_loss": 3.409295313412382, + "tokens_seen": 2388787200 + }, + { + "epoch": 0.72, + "learning_rate": 0.00022330992978936812, + "loss": 1.4417, + "theoretical_loss": 3.409295313412382, + "tokens_seen": 2388787200 + }, + { + "epoch": 0.72, + "learning_rate": 0.0002232938816449348, + "loss": 1.2771, + "theoretical_loss": 3.4092878474140864, + "tokens_seen": 2388852736 + }, + { + "epoch": 0.72, + "learning_rate": 0.0002232778335005015, + "loss": 1.5629, + "theoretical_loss": 3.4092803816779593, + "tokens_seen": 2388918272 + }, + { + "epoch": 0.72, + "learning_rate": 0.0002232617853560682, + "loss": 1.2352, + "theoretical_loss": 3.4092729162039848, + "tokens_seen": 2388983808 + }, + { + "epoch": 0.72, + "learning_rate": 0.0002232457372116349, + "loss": 1.3163, + "theoretical_loss": 3.4092654509921463, + "tokens_seen": 2389049344 + }, + { + "epoch": 0.72, + "learning_rate": 0.0002232296890672016, + "loss": 1.3154, + "theoretical_loss": 3.4092579860424266, + "tokens_seen": 2389114880 + }, + { + "epoch": 0.72, + "learning_rate": 0.0002232136409227683, + "loss": 1.2422, + "theoretical_loss": 3.4092505213548105, + "tokens_seen": 2389180416 + }, + { + "epoch": 0.72, + "learning_rate": 0.000223197592778335, + "loss": 1.3469, + "theoretical_loss": 3.4092430569292813, + "tokens_seen": 2389245952 + }, + { + "epoch": 0.72, + "learning_rate": 0.0002231815446339017, + "loss": 1.2851, + "theoretical_loss": 3.409235592765822, + "tokens_seen": 2389311488 + }, + { + "epoch": 0.72, + "learning_rate": 0.0002231654964894684, + "loss": 1.2318, + "theoretical_loss": 3.4092281288644166, + "tokens_seen": 2389377024 + }, + { + "epoch": 0.72, + "learning_rate": 0.0002231494483450351, + "loss": 1.3901, + "theoretical_loss": 3.4092206652250487, + "tokens_seen": 2389442560 + }, + { + "epoch": 0.72, + "learning_rate": 0.0002231334002006018, + "loss": 1.2696, + "theoretical_loss": 3.4092132018477024, + "tokens_seen": 2389508096 + }, + { + "epoch": 0.72, + "learning_rate": 0.0002231173520561685, + "loss": 1.3431, + "theoretical_loss": 3.40920573873236, + "tokens_seen": 2389573632 + }, + { + "epoch": 0.72, + "learning_rate": 0.0002231013039117352, + "loss": 1.4116, + "theoretical_loss": 3.4091982758790067, + "tokens_seen": 2389639168 + }, + { + "epoch": 0.72, + "learning_rate": 0.0002230852557673019, + "loss": 1.3333, + "theoretical_loss": 3.4091908132876254, + "tokens_seen": 2389704704 + }, + { + "epoch": 0.72, + "learning_rate": 0.0002230692076228686, + "loss": 1.378, + "theoretical_loss": 3.4091833509581995, + "tokens_seen": 2389770240 + }, + { + "epoch": 0.72, + "learning_rate": 0.0002230531594784353, + "loss": 1.3128, + "theoretical_loss": 3.409175888890713, + "tokens_seen": 2389835776 + }, + { + "epoch": 0.72, + "learning_rate": 0.000223037111334002, + "loss": 1.3437, + "theoretical_loss": 3.409168427085149, + "tokens_seen": 2389901312 + }, + { + "epoch": 0.72, + "learning_rate": 0.0002230210631895687, + "loss": 1.315, + "theoretical_loss": 3.4091609655414916, + "tokens_seen": 2389966848 + }, + { + "epoch": 0.72, + "learning_rate": 0.0002230050150451354, + "loss": 1.2661, + "theoretical_loss": 3.4091535042597245, + "tokens_seen": 2390032384 + }, + { + "epoch": 0.72, + "learning_rate": 0.0002229889669007021, + "loss": 1.3443, + "theoretical_loss": 3.409146043239831, + "tokens_seen": 2390097920 + }, + { + "epoch": 0.72, + "learning_rate": 0.0002229729187562688, + "loss": 1.2834, + "theoretical_loss": 3.409138582481795, + "tokens_seen": 2390163456 + }, + { + "epoch": 0.72, + "learning_rate": 0.0002229568706118355, + "loss": 1.2613, + "theoretical_loss": 3.4091311219855998, + "tokens_seen": 2390228992 + }, + { + "epoch": 0.72, + "learning_rate": 0.0002229408224674022, + "loss": 1.3088, + "theoretical_loss": 3.409123661751229, + "tokens_seen": 2390294528 + }, + { + "epoch": 0.72, + "learning_rate": 0.0002229247743229689, + "loss": 1.2904, + "theoretical_loss": 3.4091162017786667, + "tokens_seen": 2390360064 + }, + { + "epoch": 0.72, + "learning_rate": 0.00022290872617853559, + "loss": 1.295, + "theoretical_loss": 3.4091087420678963, + "tokens_seen": 2390425600 + }, + { + "epoch": 0.72, + "learning_rate": 0.00022289267803410229, + "loss": 1.2269, + "theoretical_loss": 3.4091012826189013, + "tokens_seen": 2390491136 + }, + { + "epoch": 0.72, + "learning_rate": 0.000222876629889669, + "loss": 1.2484, + "theoretical_loss": 3.4090938234316654, + "tokens_seen": 2390556672 + }, + { + "epoch": 0.72, + "learning_rate": 0.0002228605817452357, + "loss": 1.2543, + "theoretical_loss": 3.4090863645061726, + "tokens_seen": 2390622208 + }, + { + "epoch": 0.72, + "learning_rate": 0.0002228445336008024, + "loss": 1.3359, + "theoretical_loss": 3.409078905842406, + "tokens_seen": 2390687744 + }, + { + "epoch": 0.72, + "learning_rate": 0.0002228284854563691, + "loss": 1.2795, + "theoretical_loss": 3.4090714474403496, + "tokens_seen": 2390753280 + }, + { + "epoch": 0.72, + "learning_rate": 0.00022281243731193584, + "loss": 1.2375, + "theoretical_loss": 3.409063989299987, + "tokens_seen": 2390818816 + }, + { + "epoch": 0.72, + "learning_rate": 0.00022279638916750254, + "loss": 1.3219, + "theoretical_loss": 3.409056531421302, + "tokens_seen": 2390884352 + }, + { + "epoch": 0.72, + "learning_rate": 0.00022278034102306924, + "loss": 1.3053, + "theoretical_loss": 3.409049073804278, + "tokens_seen": 2390949888 + }, + { + "epoch": 0.72, + "learning_rate": 0.00022276429287863595, + "loss": 1.3908, + "theoretical_loss": 3.409041616448898, + "tokens_seen": 2391015424 + }, + { + "epoch": 0.72, + "learning_rate": 0.00022274824473420265, + "loss": 1.449, + "theoretical_loss": 3.4090341593551465, + "tokens_seen": 2391080960 + }, + { + "epoch": 0.72, + "learning_rate": 0.00022273219658976935, + "loss": 1.365, + "theoretical_loss": 3.409026702523007, + "tokens_seen": 2391146496 + }, + { + "epoch": 0.72, + "learning_rate": 0.00022271614844533605, + "loss": 1.3221, + "theoretical_loss": 3.4090192459524635, + "tokens_seen": 2391212032 + }, + { + "epoch": 0.72, + "learning_rate": 0.00022270010030090275, + "loss": 1.2037, + "theoretical_loss": 3.4090117896434995, + "tokens_seen": 2391277568 + }, + { + "epoch": 0.72, + "learning_rate": 0.00022268405215646945, + "loss": 1.2741, + "theoretical_loss": 3.409004333596098, + "tokens_seen": 2391343104 + }, + { + "epoch": 0.72, + "learning_rate": 0.00022266800401203612, + "loss": 1.3723, + "theoretical_loss": 3.4089968778102424, + "tokens_seen": 2391408640 + }, + { + "epoch": 0.72, + "learning_rate": 0.00022265195586760282, + "loss": 1.3709, + "theoretical_loss": 3.4089894222859183, + "tokens_seen": 2391474176 + }, + { + "epoch": 0.72, + "learning_rate": 0.00022263590772316952, + "loss": 1.2789, + "theoretical_loss": 3.408981967023107, + "tokens_seen": 2391539712 + }, + { + "epoch": 0.72, + "learning_rate": 0.00022261985957873622, + "loss": 1.3722, + "theoretical_loss": 3.408974512021794, + "tokens_seen": 2391605248 + }, + { + "epoch": 0.72, + "learning_rate": 0.00022260381143430292, + "loss": 1.3901, + "theoretical_loss": 3.408967057281962, + "tokens_seen": 2391670784 + }, + { + "epoch": 0.72, + "learning_rate": 0.00022258776328986963, + "loss": 1.2558, + "theoretical_loss": 3.408959602803595, + "tokens_seen": 2391736320 + }, + { + "epoch": 0.72, + "learning_rate": 0.00022257171514543633, + "loss": 1.2719, + "theoretical_loss": 3.4089521485866765, + "tokens_seen": 2391801856 + }, + { + "epoch": 0.72, + "learning_rate": 0.00022255566700100303, + "loss": 1.3134, + "theoretical_loss": 3.40894469463119, + "tokens_seen": 2391867392 + }, + { + "epoch": 0.72, + "learning_rate": 0.00022253961885656973, + "loss": 1.3019, + "theoretical_loss": 3.4089372409371195, + "tokens_seen": 2391932928 + }, + { + "epoch": 0.72, + "learning_rate": 0.00022252357071213643, + "loss": 1.2841, + "theoretical_loss": 3.408929787504449, + "tokens_seen": 2391998464 + }, + { + "epoch": 0.72, + "objective/train/docs_used": 871244, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.4347987174987793, + "objective/train/theoretical_loss": 3.408922334333161, + "objective/train/tokens_used": 2412524000, + "theoretical_loss": 3.408922334333161, + "tokens_seen": 2392064000 + }, + { + "epoch": 0.72, + "learning_rate": 0.00022250752256770313, + "loss": 1.3866, + "theoretical_loss": 3.408922334333161, + "tokens_seen": 2392064000 + }, + { + "epoch": 0.72, + "learning_rate": 0.00022249147442326983, + "loss": 1.3114, + "theoretical_loss": 3.4089148814232404, + "tokens_seen": 2392129536 + }, + { + "epoch": 0.72, + "learning_rate": 0.00022247542627883653, + "loss": 1.3411, + "theoretical_loss": 3.40890742877467, + "tokens_seen": 2392195072 + }, + { + "epoch": 0.72, + "learning_rate": 0.00022245937813440323, + "loss": 1.3928, + "theoretical_loss": 3.408899976387434, + "tokens_seen": 2392260608 + }, + { + "epoch": 0.72, + "learning_rate": 0.00022244332998996993, + "loss": 1.2698, + "theoretical_loss": 3.4088925242615162, + "tokens_seen": 2392326144 + }, + { + "epoch": 0.72, + "learning_rate": 0.00022242728184553663, + "loss": 1.4063, + "theoretical_loss": 3.4088850723968998, + "tokens_seen": 2392391680 + }, + { + "epoch": 0.72, + "learning_rate": 0.00022241123370110333, + "loss": 1.3042, + "theoretical_loss": 3.408877620793569, + "tokens_seen": 2392457216 + }, + { + "epoch": 0.73, + "learning_rate": 0.00022239518555667003, + "loss": 1.2725, + "theoretical_loss": 3.4088701694515064, + "tokens_seen": 2392522752 + }, + { + "epoch": 0.73, + "learning_rate": 0.00022237913741223673, + "loss": 1.2894, + "theoretical_loss": 3.408862718370697, + "tokens_seen": 2392588288 + }, + { + "epoch": 0.73, + "learning_rate": 0.00022236308926780343, + "loss": 1.2973, + "theoretical_loss": 3.408855267551124, + "tokens_seen": 2392653824 + }, + { + "epoch": 0.73, + "learning_rate": 0.00022234704112337013, + "loss": 1.3005, + "theoretical_loss": 3.408847816992771, + "tokens_seen": 2392719360 + }, + { + "epoch": 0.73, + "learning_rate": 0.00022233099297893684, + "loss": 1.3479, + "theoretical_loss": 3.4088403666956215, + "tokens_seen": 2392784896 + }, + { + "epoch": 0.73, + "learning_rate": 0.00022231494483450354, + "loss": 1.2603, + "theoretical_loss": 3.4088329166596596, + "tokens_seen": 2392850432 + }, + { + "epoch": 0.73, + "learning_rate": 0.00022229889669007024, + "loss": 1.3294, + "theoretical_loss": 3.4088254668848683, + "tokens_seen": 2392915968 + }, + { + "epoch": 0.73, + "learning_rate": 0.0002222828485456369, + "loss": 1.3483, + "theoretical_loss": 3.4088180173712326, + "tokens_seen": 2392981504 + }, + { + "epoch": 0.73, + "learning_rate": 0.0002222668004012036, + "loss": 1.2286, + "theoretical_loss": 3.4088105681187346, + "tokens_seen": 2393047040 + }, + { + "epoch": 0.73, + "learning_rate": 0.0002222507522567703, + "loss": 1.3167, + "theoretical_loss": 3.4088031191273593, + "tokens_seen": 2393112576 + }, + { + "epoch": 0.73, + "learning_rate": 0.000222234704112337, + "loss": 1.3383, + "theoretical_loss": 3.40879567039709, + "tokens_seen": 2393178112 + }, + { + "epoch": 0.73, + "learning_rate": 0.0002222186559679037, + "loss": 1.2576, + "theoretical_loss": 3.4087882219279098, + "tokens_seen": 2393243648 + }, + { + "epoch": 0.73, + "learning_rate": 0.0002222026078234704, + "loss": 1.3972, + "theoretical_loss": 3.408780773719803, + "tokens_seen": 2393309184 + }, + { + "epoch": 0.73, + "learning_rate": 0.00022218655967903711, + "loss": 1.2837, + "theoretical_loss": 3.408773325772753, + "tokens_seen": 2393374720 + }, + { + "epoch": 0.73, + "learning_rate": 0.00022217051153460381, + "loss": 1.3373, + "theoretical_loss": 3.408765878086744, + "tokens_seen": 2393440256 + }, + { + "epoch": 0.73, + "learning_rate": 0.00022215446339017052, + "loss": 1.2692, + "theoretical_loss": 3.4087584306617593, + "tokens_seen": 2393505792 + }, + { + "epoch": 0.73, + "learning_rate": 0.00022213841524573722, + "loss": 1.3544, + "theoretical_loss": 3.4087509834977823, + "tokens_seen": 2393571328 + }, + { + "epoch": 0.73, + "learning_rate": 0.00022212236710130392, + "loss": 1.2877, + "theoretical_loss": 3.4087435365947973, + "tokens_seen": 2393636864 + }, + { + "epoch": 0.73, + "learning_rate": 0.00022210631895687062, + "loss": 1.3241, + "theoretical_loss": 3.408736089952788, + "tokens_seen": 2393702400 + }, + { + "epoch": 0.73, + "learning_rate": 0.00022209027081243732, + "loss": 1.397, + "theoretical_loss": 3.4087286435717377, + "tokens_seen": 2393767936 + }, + { + "epoch": 0.73, + "learning_rate": 0.00022207422266800402, + "loss": 1.3723, + "theoretical_loss": 3.4087211974516305, + "tokens_seen": 2393833472 + }, + { + "epoch": 0.73, + "learning_rate": 0.00022205817452357072, + "loss": 1.2433, + "theoretical_loss": 3.4087137515924497, + "tokens_seen": 2393899008 + }, + { + "epoch": 0.73, + "learning_rate": 0.00022204212637913742, + "loss": 1.4131, + "theoretical_loss": 3.4087063059941793, + "tokens_seen": 2393964544 + }, + { + "epoch": 0.73, + "learning_rate": 0.00022202607823470412, + "loss": 1.3494, + "theoretical_loss": 3.408698860656803, + "tokens_seen": 2394030080 + }, + { + "epoch": 0.73, + "learning_rate": 0.00022201003009027082, + "loss": 1.3055, + "theoretical_loss": 3.4086914155803045, + "tokens_seen": 2394095616 + }, + { + "epoch": 0.73, + "learning_rate": 0.00022199398194583752, + "loss": 1.2395, + "theoretical_loss": 3.4086839707646677, + "tokens_seen": 2394161152 + }, + { + "epoch": 0.73, + "learning_rate": 0.00022197793380140422, + "loss": 1.2209, + "theoretical_loss": 3.408676526209876, + "tokens_seen": 2394226688 + }, + { + "epoch": 0.73, + "learning_rate": 0.00022196188565697092, + "loss": 1.2971, + "theoretical_loss": 3.408669081915913, + "tokens_seen": 2394292224 + }, + { + "epoch": 0.73, + "learning_rate": 0.00022194583751253762, + "loss": 1.2345, + "theoretical_loss": 3.4086616378827626, + "tokens_seen": 2394357760 + }, + { + "epoch": 0.73, + "learning_rate": 0.00022192978936810432, + "loss": 1.3291, + "theoretical_loss": 3.4086541941104085, + "tokens_seen": 2394423296 + }, + { + "epoch": 0.73, + "learning_rate": 0.00022191374122367102, + "loss": 1.2319, + "theoretical_loss": 3.4086467505988347, + "tokens_seen": 2394488832 + }, + { + "epoch": 0.73, + "learning_rate": 0.0002218976930792377, + "loss": 1.2861, + "theoretical_loss": 3.408639307348025, + "tokens_seen": 2394554368 + }, + { + "epoch": 0.73, + "learning_rate": 0.0002218816449348044, + "loss": 1.2498, + "theoretical_loss": 3.4086318643579627, + "tokens_seen": 2394619904 + }, + { + "epoch": 0.73, + "learning_rate": 0.0002218655967903711, + "loss": 1.4333, + "theoretical_loss": 3.408624421628631, + "tokens_seen": 2394685440 + }, + { + "epoch": 0.73, + "learning_rate": 0.0002218495486459378, + "loss": 1.218, + "theoretical_loss": 3.408616979160015, + "tokens_seen": 2394750976 + }, + { + "epoch": 0.73, + "learning_rate": 0.0002218335005015045, + "loss": 1.3023, + "theoretical_loss": 3.408609536952098, + "tokens_seen": 2394816512 + }, + { + "epoch": 0.73, + "learning_rate": 0.0002218174523570712, + "loss": 1.2442, + "theoretical_loss": 3.408602095004863, + "tokens_seen": 2394882048 + }, + { + "epoch": 0.73, + "learning_rate": 0.0002218014042126379, + "loss": 1.3071, + "theoretical_loss": 3.4085946533182945, + "tokens_seen": 2394947584 + }, + { + "epoch": 0.73, + "learning_rate": 0.0002217853560682046, + "loss": 1.2698, + "theoretical_loss": 3.4085872118923755, + "tokens_seen": 2395013120 + }, + { + "epoch": 0.73, + "learning_rate": 0.0002217693079237713, + "loss": 1.3617, + "theoretical_loss": 3.4085797707270906, + "tokens_seen": 2395078656 + }, + { + "epoch": 0.73, + "learning_rate": 0.000221753259779338, + "loss": 1.3343, + "theoretical_loss": 3.408572329822423, + "tokens_seen": 2395144192 + }, + { + "epoch": 0.73, + "learning_rate": 0.0002217372116349047, + "loss": 1.3609, + "theoretical_loss": 3.408564889178357, + "tokens_seen": 2395209728 + }, + { + "epoch": 0.73, + "learning_rate": 0.0002217211634904714, + "loss": 1.34, + "theoretical_loss": 3.408557448794875, + "tokens_seen": 2395275264 + }, + { + "epoch": 0.73, + "objective/train/docs_used": 872384, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.238278865814209, + "objective/train/theoretical_loss": 3.4085500086719622, + "objective/train/tokens_used": 2415800800, + "theoretical_loss": 3.4085500086719622, + "tokens_seen": 2395340800 + }, + { + "epoch": 0.73, + "learning_rate": 0.0002217051153460381, + "loss": 1.2746, + "theoretical_loss": 3.4085500086719622, + "tokens_seen": 2395340800 + }, + { + "epoch": 0.73, + "learning_rate": 0.0002216890672016048, + "loss": 1.3743, + "theoretical_loss": 3.408542568809602, + "tokens_seen": 2395406336 + }, + { + "epoch": 0.73, + "learning_rate": 0.0002216730190571715, + "loss": 1.2318, + "theoretical_loss": 3.408535129207778, + "tokens_seen": 2395471872 + }, + { + "epoch": 0.73, + "learning_rate": 0.0002216569709127382, + "loss": 1.28, + "theoretical_loss": 3.4085276898664736, + "tokens_seen": 2395537408 + }, + { + "epoch": 0.73, + "learning_rate": 0.0002216409227683049, + "loss": 1.234, + "theoretical_loss": 3.4085202507856724, + "tokens_seen": 2395602944 + }, + { + "epoch": 0.73, + "learning_rate": 0.0002216248746238716, + "loss": 1.3272, + "theoretical_loss": 3.4085128119653594, + "tokens_seen": 2395668480 + }, + { + "epoch": 0.73, + "learning_rate": 0.00022160882647943834, + "loss": 1.3846, + "theoretical_loss": 3.408505373405517, + "tokens_seen": 2395734016 + }, + { + "epoch": 0.73, + "learning_rate": 0.00022159277833500504, + "loss": 1.3752, + "theoretical_loss": 3.40849793510613, + "tokens_seen": 2395799552 + }, + { + "epoch": 0.73, + "learning_rate": 0.00022157673019057174, + "loss": 1.237, + "theoretical_loss": 3.4084904970671817, + "tokens_seen": 2395865088 + }, + { + "epoch": 0.73, + "learning_rate": 0.00022156068204613844, + "loss": 1.3775, + "theoretical_loss": 3.4084830592886552, + "tokens_seen": 2395930624 + }, + { + "epoch": 0.73, + "learning_rate": 0.00022154463390170514, + "loss": 1.1854, + "theoretical_loss": 3.4084756217705356, + "tokens_seen": 2395996160 + }, + { + "epoch": 0.73, + "learning_rate": 0.00022152858575727184, + "loss": 1.3433, + "theoretical_loss": 3.4084681845128055, + "tokens_seen": 2396061696 + }, + { + "epoch": 0.73, + "learning_rate": 0.00022151253761283854, + "loss": 1.168, + "theoretical_loss": 3.4084607475154494, + "tokens_seen": 2396127232 + }, + { + "epoch": 0.73, + "learning_rate": 0.00022149648946840524, + "loss": 1.2147, + "theoretical_loss": 3.4084533107784507, + "tokens_seen": 2396192768 + }, + { + "epoch": 0.73, + "learning_rate": 0.00022148044132397194, + "loss": 1.3589, + "theoretical_loss": 3.408445874301793, + "tokens_seen": 2396258304 + }, + { + "epoch": 0.73, + "learning_rate": 0.00022146439317953864, + "loss": 1.3594, + "theoretical_loss": 3.4084384380854607, + "tokens_seen": 2396323840 + }, + { + "epoch": 0.73, + "learning_rate": 0.00022144834503510534, + "loss": 1.3621, + "theoretical_loss": 3.408431002129437, + "tokens_seen": 2396389376 + }, + { + "epoch": 0.73, + "learning_rate": 0.00022143229689067204, + "loss": 1.3597, + "theoretical_loss": 3.4084235664337057, + "tokens_seen": 2396454912 + }, + { + "epoch": 0.73, + "learning_rate": 0.00022141624874623874, + "loss": 1.2321, + "theoretical_loss": 3.4084161309982504, + "tokens_seen": 2396520448 + }, + { + "epoch": 0.73, + "learning_rate": 0.00022140020060180545, + "loss": 1.4375, + "theoretical_loss": 3.4084086958230557, + "tokens_seen": 2396585984 + }, + { + "epoch": 0.73, + "learning_rate": 0.00022138415245737215, + "loss": 1.2501, + "theoretical_loss": 3.408401260908105, + "tokens_seen": 2396651520 + }, + { + "epoch": 0.73, + "learning_rate": 0.00022136810431293885, + "loss": 1.2732, + "theoretical_loss": 3.4083938262533815, + "tokens_seen": 2396717056 + }, + { + "epoch": 0.73, + "learning_rate": 0.00022135205616850555, + "loss": 1.3198, + "theoretical_loss": 3.40838639185887, + "tokens_seen": 2396782592 + }, + { + "epoch": 0.73, + "learning_rate": 0.00022133600802407225, + "loss": 1.4033, + "theoretical_loss": 3.408378957724553, + "tokens_seen": 2396848128 + }, + { + "epoch": 0.73, + "learning_rate": 0.00022131995987963895, + "loss": 1.3241, + "theoretical_loss": 3.4083715238504153, + "tokens_seen": 2396913664 + }, + { + "epoch": 0.73, + "learning_rate": 0.00022130391173520565, + "loss": 1.3393, + "theoretical_loss": 3.4083640902364403, + "tokens_seen": 2396979200 + }, + { + "epoch": 0.73, + "learning_rate": 0.00022128786359077235, + "loss": 1.3078, + "theoretical_loss": 3.408356656882612, + "tokens_seen": 2397044736 + }, + { + "epoch": 0.73, + "learning_rate": 0.00022127181544633905, + "loss": 1.3116, + "theoretical_loss": 3.4083492237889135, + "tokens_seen": 2397110272 + }, + { + "epoch": 0.73, + "learning_rate": 0.00022125576730190572, + "loss": 1.345, + "theoretical_loss": 3.4083417909553297, + "tokens_seen": 2397175808 + }, + { + "epoch": 0.73, + "learning_rate": 0.00022123971915747242, + "loss": 1.2879, + "theoretical_loss": 3.4083343583818433, + "tokens_seen": 2397241344 + }, + { + "epoch": 0.73, + "learning_rate": 0.00022122367101303913, + "loss": 1.315, + "theoretical_loss": 3.408326926068439, + "tokens_seen": 2397306880 + }, + { + "epoch": 0.73, + "learning_rate": 0.00022120762286860583, + "loss": 1.2854, + "theoretical_loss": 3.4083194940150996, + "tokens_seen": 2397372416 + }, + { + "epoch": 0.73, + "learning_rate": 0.00022119157472417253, + "loss": 1.3292, + "theoretical_loss": 3.40831206222181, + "tokens_seen": 2397437952 + }, + { + "epoch": 0.73, + "learning_rate": 0.00022117552657973923, + "loss": 1.2997, + "theoretical_loss": 3.408304630688553, + "tokens_seen": 2397503488 + }, + { + "epoch": 0.73, + "learning_rate": 0.00022115947843530593, + "loss": 1.3317, + "theoretical_loss": 3.4082971994153133, + "tokens_seen": 2397569024 + }, + { + "epoch": 0.73, + "learning_rate": 0.00022114343029087263, + "loss": 1.2733, + "theoretical_loss": 3.4082897684020743, + "tokens_seen": 2397634560 + }, + { + "epoch": 0.73, + "learning_rate": 0.00022112738214643933, + "loss": 1.361, + "theoretical_loss": 3.4082823376488194, + "tokens_seen": 2397700096 + }, + { + "epoch": 0.73, + "learning_rate": 0.00022111133400200603, + "loss": 1.251, + "theoretical_loss": 3.4082749071555325, + "tokens_seen": 2397765632 + }, + { + "epoch": 0.73, + "learning_rate": 0.00022109528585757273, + "loss": 1.3184, + "theoretical_loss": 3.408267476922198, + "tokens_seen": 2397831168 + }, + { + "epoch": 0.73, + "learning_rate": 0.00022107923771313943, + "loss": 1.3545, + "theoretical_loss": 3.4082600469488, + "tokens_seen": 2397896704 + }, + { + "epoch": 0.73, + "learning_rate": 0.00022106318956870613, + "loss": 1.3537, + "theoretical_loss": 3.4082526172353207, + "tokens_seen": 2397962240 + }, + { + "epoch": 0.73, + "learning_rate": 0.00022104714142427283, + "loss": 1.3235, + "theoretical_loss": 3.408245187781745, + "tokens_seen": 2398027776 + }, + { + "epoch": 0.73, + "learning_rate": 0.00022103109327983953, + "loss": 1.378, + "theoretical_loss": 3.4082377585880566, + "tokens_seen": 2398093312 + }, + { + "epoch": 0.73, + "learning_rate": 0.00022101504513540623, + "loss": 1.2675, + "theoretical_loss": 3.408230329654239, + "tokens_seen": 2398158848 + }, + { + "epoch": 0.73, + "learning_rate": 0.00022099899699097293, + "loss": 1.3266, + "theoretical_loss": 3.4082229009802765, + "tokens_seen": 2398224384 + }, + { + "epoch": 0.73, + "learning_rate": 0.00022098294884653963, + "loss": 1.2491, + "theoretical_loss": 3.4082154725661526, + "tokens_seen": 2398289920 + }, + { + "epoch": 0.73, + "learning_rate": 0.00022096690070210634, + "loss": 1.3671, + "theoretical_loss": 3.4082080444118517, + "tokens_seen": 2398355456 + }, + { + "epoch": 0.73, + "learning_rate": 0.00022095085255767304, + "loss": 1.2717, + "theoretical_loss": 3.4082006165173566, + "tokens_seen": 2398420992 + }, + { + "epoch": 0.73, + "learning_rate": 0.00022093480441323974, + "loss": 1.2972, + "theoretical_loss": 3.4081931888826515, + "tokens_seen": 2398486528 + }, + { + "epoch": 0.73, + "learning_rate": 0.00022091875626880644, + "loss": 1.2738, + "theoretical_loss": 3.4081857615077205, + "tokens_seen": 2398552064 + }, + { + "epoch": 0.73, + "objective/train/docs_used": 873551, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.2822637557983398, + "objective/train/theoretical_loss": 3.4081783343925474, + "objective/train/tokens_used": 2419077600, + "theoretical_loss": 3.4081783343925474, + "tokens_seen": 2398617600 + }, + { + "epoch": 0.73, + "learning_rate": 0.00022090270812437314, + "loss": 1.2762, + "theoretical_loss": 3.4081783343925474, + "tokens_seen": 2398617600 + }, + { + "epoch": 0.73, + "learning_rate": 0.00022088665997993984, + "loss": 1.3134, + "theoretical_loss": 3.408170907537116, + "tokens_seen": 2398683136 + }, + { + "epoch": 0.73, + "learning_rate": 0.0002208706118355065, + "loss": 1.3132, + "theoretical_loss": 3.4081634809414094, + "tokens_seen": 2398748672 + }, + { + "epoch": 0.73, + "learning_rate": 0.0002208545636910732, + "loss": 1.2667, + "theoretical_loss": 3.4081560546054126, + "tokens_seen": 2398814208 + }, + { + "epoch": 0.73, + "learning_rate": 0.0002208385155466399, + "loss": 1.4221, + "theoretical_loss": 3.4081486285291085, + "tokens_seen": 2398879744 + }, + { + "epoch": 0.73, + "learning_rate": 0.00022082246740220661, + "loss": 1.294, + "theoretical_loss": 3.408141202712481, + "tokens_seen": 2398945280 + }, + { + "epoch": 0.73, + "learning_rate": 0.00022080641925777331, + "loss": 1.3131, + "theoretical_loss": 3.4081337771555145, + "tokens_seen": 2399010816 + }, + { + "epoch": 0.73, + "learning_rate": 0.00022079037111334002, + "loss": 1.2049, + "theoretical_loss": 3.4081263518581926, + "tokens_seen": 2399076352 + }, + { + "epoch": 0.73, + "learning_rate": 0.00022077432296890672, + "loss": 1.2233, + "theoretical_loss": 3.408118926820499, + "tokens_seen": 2399141888 + }, + { + "epoch": 0.73, + "learning_rate": 0.00022075827482447342, + "loss": 1.2557, + "theoretical_loss": 3.4081115020424173, + "tokens_seen": 2399207424 + }, + { + "epoch": 0.73, + "learning_rate": 0.00022074222668004012, + "loss": 1.2888, + "theoretical_loss": 3.408104077523932, + "tokens_seen": 2399272960 + }, + { + "epoch": 0.73, + "learning_rate": 0.00022072617853560682, + "loss": 1.2832, + "theoretical_loss": 3.408096653265026, + "tokens_seen": 2399338496 + }, + { + "epoch": 0.73, + "learning_rate": 0.00022071013039117352, + "loss": 1.277, + "theoretical_loss": 3.408089229265684, + "tokens_seen": 2399404032 + }, + { + "epoch": 0.73, + "learning_rate": 0.00022069408224674022, + "loss": 1.136, + "theoretical_loss": 3.4080818055258897, + "tokens_seen": 2399469568 + }, + { + "epoch": 0.73, + "learning_rate": 0.00022067803410230692, + "loss": 1.2952, + "theoretical_loss": 3.4080743820456263, + "tokens_seen": 2399535104 + }, + { + "epoch": 0.73, + "learning_rate": 0.00022066198595787362, + "loss": 1.3988, + "theoretical_loss": 3.4080669588248784, + "tokens_seen": 2399600640 + }, + { + "epoch": 0.73, + "learning_rate": 0.00022064593781344032, + "loss": 1.3454, + "theoretical_loss": 3.4080595358636296, + "tokens_seen": 2399666176 + }, + { + "epoch": 0.73, + "learning_rate": 0.00022062988966900702, + "loss": 1.4447, + "theoretical_loss": 3.4080521131618635, + "tokens_seen": 2399731712 + }, + { + "epoch": 0.73, + "learning_rate": 0.00022061384152457372, + "loss": 1.2868, + "theoretical_loss": 3.4080446907195636, + "tokens_seen": 2399797248 + }, + { + "epoch": 0.73, + "learning_rate": 0.00022059779338014042, + "loss": 1.3092, + "theoretical_loss": 3.408037268536715, + "tokens_seen": 2399862784 + }, + { + "epoch": 0.73, + "learning_rate": 0.00022058174523570712, + "loss": 1.4015, + "theoretical_loss": 3.4080298466133003, + "tokens_seen": 2399928320 + }, + { + "epoch": 0.73, + "learning_rate": 0.00022056569709127382, + "loss": 1.2406, + "theoretical_loss": 3.408022424949304, + "tokens_seen": 2399993856 + }, + { + "epoch": 0.73, + "learning_rate": 0.00022054964894684052, + "loss": 1.3197, + "theoretical_loss": 3.40801500354471, + "tokens_seen": 2400059392 + }, + { + "epoch": 0.73, + "learning_rate": 0.00022053360080240723, + "loss": 1.2719, + "theoretical_loss": 3.408007582399502, + "tokens_seen": 2400124928 + }, + { + "epoch": 0.73, + "learning_rate": 0.00022051755265797393, + "loss": 1.403, + "theoretical_loss": 3.4080001615136633, + "tokens_seen": 2400190464 + }, + { + "epoch": 0.73, + "learning_rate": 0.00022050150451354063, + "loss": 1.2642, + "theoretical_loss": 3.4079927408871784, + "tokens_seen": 2400256000 + }, + { + "epoch": 0.73, + "learning_rate": 0.0002204854563691073, + "loss": 1.2606, + "theoretical_loss": 3.407985320520031, + "tokens_seen": 2400321536 + }, + { + "epoch": 0.73, + "learning_rate": 0.000220469408224674, + "loss": 1.2484, + "theoretical_loss": 3.407977900412205, + "tokens_seen": 2400387072 + }, + { + "epoch": 0.73, + "learning_rate": 0.0002204533600802407, + "loss": 1.3282, + "theoretical_loss": 3.4079704805636846, + "tokens_seen": 2400452608 + }, + { + "epoch": 0.73, + "learning_rate": 0.0002204373119358074, + "loss": 1.3307, + "theoretical_loss": 3.4079630609744527, + "tokens_seen": 2400518144 + }, + { + "epoch": 0.73, + "learning_rate": 0.0002204212637913741, + "loss": 1.3325, + "theoretical_loss": 3.4079556416444943, + "tokens_seen": 2400583680 + }, + { + "epoch": 0.73, + "learning_rate": 0.00022040521564694086, + "loss": 1.3011, + "theoretical_loss": 3.407948222573792, + "tokens_seen": 2400649216 + }, + { + "epoch": 0.73, + "learning_rate": 0.00022038916750250756, + "loss": 1.3644, + "theoretical_loss": 3.407940803762331, + "tokens_seen": 2400714752 + }, + { + "epoch": 0.73, + "learning_rate": 0.00022037311935807426, + "loss": 1.3682, + "theoretical_loss": 3.4079333852100944, + "tokens_seen": 2400780288 + }, + { + "epoch": 0.73, + "learning_rate": 0.00022035707121364096, + "loss": 1.3839, + "theoretical_loss": 3.407925966917066, + "tokens_seen": 2400845824 + }, + { + "epoch": 0.73, + "learning_rate": 0.00022034102306920766, + "loss": 1.3519, + "theoretical_loss": 3.40791854888323, + "tokens_seen": 2400911360 + }, + { + "epoch": 0.73, + "learning_rate": 0.00022032497492477436, + "loss": 1.3119, + "theoretical_loss": 3.4079111311085697, + "tokens_seen": 2400976896 + }, + { + "epoch": 0.73, + "learning_rate": 0.00022030892678034106, + "loss": 1.2898, + "theoretical_loss": 3.4079037135930697, + "tokens_seen": 2401042432 + }, + { + "epoch": 0.73, + "learning_rate": 0.00022029287863590776, + "loss": 1.179, + "theoretical_loss": 3.4078962963367134, + "tokens_seen": 2401107968 + }, + { + "epoch": 0.73, + "learning_rate": 0.00022027683049147446, + "loss": 1.2627, + "theoretical_loss": 3.4078888793394855, + "tokens_seen": 2401173504 + }, + { + "epoch": 0.73, + "learning_rate": 0.00022026078234704116, + "loss": 1.362, + "theoretical_loss": 3.4078814626013685, + "tokens_seen": 2401239040 + }, + { + "epoch": 0.73, + "learning_rate": 0.00022024473420260784, + "loss": 1.2994, + "theoretical_loss": 3.4078740461223473, + "tokens_seen": 2401304576 + }, + { + "epoch": 0.73, + "learning_rate": 0.00022022868605817454, + "loss": 1.2954, + "theoretical_loss": 3.407866629902405, + "tokens_seen": 2401370112 + }, + { + "epoch": 0.73, + "learning_rate": 0.00022021263791374124, + "loss": 1.3352, + "theoretical_loss": 3.407859213941527, + "tokens_seen": 2401435648 + }, + { + "epoch": 0.73, + "learning_rate": 0.00022019658976930794, + "loss": 1.3551, + "theoretical_loss": 3.4078517982396956, + "tokens_seen": 2401501184 + }, + { + "epoch": 0.73, + "learning_rate": 0.00022018054162487464, + "loss": 1.333, + "theoretical_loss": 3.407844382796895, + "tokens_seen": 2401566720 + }, + { + "epoch": 0.73, + "learning_rate": 0.00022016449348044134, + "loss": 1.2235, + "theoretical_loss": 3.4078369676131093, + "tokens_seen": 2401632256 + }, + { + "epoch": 0.73, + "learning_rate": 0.00022014844533600804, + "loss": 1.2733, + "theoretical_loss": 3.4078295526883227, + "tokens_seen": 2401697792 + }, + { + "epoch": 0.73, + "learning_rate": 0.00022013239719157474, + "loss": 1.3241, + "theoretical_loss": 3.407822138022518, + "tokens_seen": 2401763328 + }, + { + "epoch": 0.73, + "learning_rate": 0.00022011634904714144, + "loss": 1.3004, + "theoretical_loss": 3.407814723615681, + "tokens_seen": 2401828864 + }, + { + "epoch": 0.73, + "objective/train/docs_used": 874334, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.2560763359069824, + "objective/train/theoretical_loss": 3.4078073094677936, + "objective/train/tokens_used": 2422354400, + "theoretical_loss": 3.4078073094677936, + "tokens_seen": 2401894400 + }, + { + "epoch": 0.73, + "learning_rate": 0.00022010030090270814, + "loss": 1.2551, + "theoretical_loss": 3.4078073094677936, + "tokens_seen": 2401894400 + }, + { + "epoch": 0.73, + "learning_rate": 0.00022008425275827484, + "loss": 1.3911, + "theoretical_loss": 3.407799895578841, + "tokens_seen": 2401959936 + }, + { + "epoch": 0.73, + "learning_rate": 0.00022006820461384154, + "loss": 1.3288, + "theoretical_loss": 3.4077924819488064, + "tokens_seen": 2402025472 + }, + { + "epoch": 0.73, + "learning_rate": 0.00022005215646940824, + "loss": 1.4141, + "theoretical_loss": 3.4077850685776743, + "tokens_seen": 2402091008 + }, + { + "epoch": 0.73, + "learning_rate": 0.00022003610832497494, + "loss": 1.3839, + "theoretical_loss": 3.407777655465428, + "tokens_seen": 2402156544 + }, + { + "epoch": 0.73, + "learning_rate": 0.00022002006018054165, + "loss": 1.2453, + "theoretical_loss": 3.4077702426120515, + "tokens_seen": 2402222080 + }, + { + "epoch": 0.73, + "learning_rate": 0.00022002006018054165, + "loss": 1.3454, + "theoretical_loss": 3.4077628300175284, + "tokens_seen": 2402287616 + }, + { + "epoch": 0.73, + "learning_rate": 0.00022000401203610835, + "loss": 1.3415, + "theoretical_loss": 3.4077554176818436, + "tokens_seen": 2402353152 + }, + { + "epoch": 0.73, + "learning_rate": 0.00021998796389167505, + "loss": 1.3516, + "theoretical_loss": 3.4077480056049803, + "tokens_seen": 2402418688 + }, + { + "epoch": 0.73, + "learning_rate": 0.00021997191574724175, + "loss": 1.3346, + "theoretical_loss": 3.4077405937869227, + "tokens_seen": 2402484224 + }, + { + "epoch": 0.73, + "learning_rate": 0.00021995586760280845, + "loss": 1.4602, + "theoretical_loss": 3.407733182227654, + "tokens_seen": 2402549760 + }, + { + "epoch": 0.73, + "learning_rate": 0.00021993981945837515, + "loss": 1.2745, + "theoretical_loss": 3.407725770927159, + "tokens_seen": 2402615296 + }, + { + "epoch": 0.73, + "learning_rate": 0.00021992377131394185, + "loss": 1.3744, + "theoretical_loss": 3.4077183598854215, + "tokens_seen": 2402680832 + }, + { + "epoch": 0.73, + "learning_rate": 0.00021990772316950855, + "loss": 1.3159, + "theoretical_loss": 3.407710949102425, + "tokens_seen": 2402746368 + }, + { + "epoch": 0.73, + "learning_rate": 0.00021989167502507525, + "loss": 1.2413, + "theoretical_loss": 3.4077035385781533, + "tokens_seen": 2402811904 + }, + { + "epoch": 0.73, + "learning_rate": 0.00021987562688064195, + "loss": 1.3712, + "theoretical_loss": 3.4076961283125904, + "tokens_seen": 2402877440 + }, + { + "epoch": 0.73, + "learning_rate": 0.00021985957873620862, + "loss": 1.5231, + "theoretical_loss": 3.407688718305721, + "tokens_seen": 2402942976 + }, + { + "epoch": 0.73, + "learning_rate": 0.00021984353059177533, + "loss": 1.3031, + "theoretical_loss": 3.4076813085575277, + "tokens_seen": 2403008512 + }, + { + "epoch": 0.73, + "learning_rate": 0.00021982748244734203, + "loss": 1.388, + "theoretical_loss": 3.4076738990679956, + "tokens_seen": 2403074048 + }, + { + "epoch": 0.73, + "learning_rate": 0.00021981143430290873, + "loss": 1.3097, + "theoretical_loss": 3.4076664898371076, + "tokens_seen": 2403139584 + }, + { + "epoch": 0.73, + "learning_rate": 0.00021979538615847543, + "loss": 1.4298, + "theoretical_loss": 3.407659080864849, + "tokens_seen": 2403205120 + }, + { + "epoch": 0.73, + "learning_rate": 0.00021977933801404213, + "loss": 1.2465, + "theoretical_loss": 3.4076516721512022, + "tokens_seen": 2403270656 + }, + { + "epoch": 0.73, + "learning_rate": 0.00021976328986960883, + "loss": 1.19, + "theoretical_loss": 3.407644263696152, + "tokens_seen": 2403336192 + }, + { + "epoch": 0.73, + "learning_rate": 0.00021974724172517553, + "loss": 1.3058, + "theoretical_loss": 3.407636855499682, + "tokens_seen": 2403401728 + }, + { + "epoch": 0.73, + "learning_rate": 0.00021973119358074223, + "loss": 1.3257, + "theoretical_loss": 3.407629447561776, + "tokens_seen": 2403467264 + }, + { + "epoch": 0.73, + "learning_rate": 0.00021971514543630893, + "loss": 1.4604, + "theoretical_loss": 3.4076220398824186, + "tokens_seen": 2403532800 + }, + { + "epoch": 0.73, + "learning_rate": 0.00021969909729187563, + "loss": 1.2098, + "theoretical_loss": 3.4076146324615935, + "tokens_seen": 2403598336 + }, + { + "epoch": 0.73, + "learning_rate": 0.00021968304914744233, + "loss": 1.2672, + "theoretical_loss": 3.4076072252992837, + "tokens_seen": 2403663872 + }, + { + "epoch": 0.73, + "learning_rate": 0.00021966700100300903, + "loss": 1.2878, + "theoretical_loss": 3.4075998183954743, + "tokens_seen": 2403729408 + }, + { + "epoch": 0.73, + "learning_rate": 0.00021965095285857573, + "loss": 1.4031, + "theoretical_loss": 3.407592411750149, + "tokens_seen": 2403794944 + }, + { + "epoch": 0.73, + "learning_rate": 0.00021963490471414243, + "loss": 1.2663, + "theoretical_loss": 3.407585005363291, + "tokens_seen": 2403860480 + }, + { + "epoch": 0.73, + "learning_rate": 0.00021961885656970913, + "loss": 1.4098, + "theoretical_loss": 3.4075775992348847, + "tokens_seen": 2403926016 + }, + { + "epoch": 0.73, + "learning_rate": 0.00021960280842527583, + "loss": 1.4649, + "theoretical_loss": 3.407570193364914, + "tokens_seen": 2403991552 + }, + { + "epoch": 0.73, + "learning_rate": 0.00021958676028084254, + "loss": 1.316, + "theoretical_loss": 3.4075627877533634, + "tokens_seen": 2404057088 + }, + { + "epoch": 0.73, + "learning_rate": 0.00021957071213640924, + "loss": 1.4309, + "theoretical_loss": 3.4075553824002163, + "tokens_seen": 2404122624 + }, + { + "epoch": 0.73, + "learning_rate": 0.00021955466399197594, + "loss": 1.3701, + "theoretical_loss": 3.407547977305456, + "tokens_seen": 2404188160 + }, + { + "epoch": 0.73, + "learning_rate": 0.00021953861584754264, + "loss": 1.2632, + "theoretical_loss": 3.407540572469068, + "tokens_seen": 2404253696 + }, + { + "epoch": 0.73, + "learning_rate": 0.00021952256770310934, + "loss": 1.4773, + "theoretical_loss": 3.407533167891035, + "tokens_seen": 2404319232 + }, + { + "epoch": 0.73, + "learning_rate": 0.00021950651955867604, + "loss": 1.4144, + "theoretical_loss": 3.4075257635713414, + "tokens_seen": 2404384768 + }, + { + "epoch": 0.73, + "learning_rate": 0.00021949047141424274, + "loss": 1.4436, + "theoretical_loss": 3.407518359509971, + "tokens_seen": 2404450304 + }, + { + "epoch": 0.73, + "learning_rate": 0.0002194744232698094, + "loss": 1.2792, + "theoretical_loss": 3.4075109557069077, + "tokens_seen": 2404515840 + }, + { + "epoch": 0.73, + "learning_rate": 0.00021945837512537611, + "loss": 1.2203, + "theoretical_loss": 3.4075035521621357, + "tokens_seen": 2404581376 + }, + { + "epoch": 0.73, + "learning_rate": 0.00021944232698094281, + "loss": 1.3065, + "theoretical_loss": 3.407496148875639, + "tokens_seen": 2404646912 + }, + { + "epoch": 0.73, + "learning_rate": 0.00021942627883650951, + "loss": 1.2858, + "theoretical_loss": 3.407488745847401, + "tokens_seen": 2404712448 + }, + { + "epoch": 0.73, + "learning_rate": 0.00021941023069207622, + "loss": 1.3239, + "theoretical_loss": 3.4074813430774062, + "tokens_seen": 2404777984 + }, + { + "epoch": 0.73, + "learning_rate": 0.00021939418254764292, + "loss": 1.2449, + "theoretical_loss": 3.4074739405656382, + "tokens_seen": 2404843520 + }, + { + "epoch": 0.73, + "learning_rate": 0.00021937813440320962, + "loss": 1.5424, + "theoretical_loss": 3.4074665383120815, + "tokens_seen": 2404909056 + }, + { + "epoch": 0.73, + "learning_rate": 0.00021936208625877632, + "loss": 1.1789, + "theoretical_loss": 3.407459136316719, + "tokens_seen": 2404974592 + }, + { + "epoch": 0.73, + "learning_rate": 0.00021934603811434302, + "loss": 1.2296, + "theoretical_loss": 3.407451734579536, + "tokens_seen": 2405040128 + }, + { + "epoch": 0.73, + "learning_rate": 0.00021932998996990972, + "loss": 1.3453, + "theoretical_loss": 3.4074443331005155, + "tokens_seen": 2405105664 + }, + { + "epoch": 0.73, + "objective/train/docs_used": 875512, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.243737816810608, + "objective/train/theoretical_loss": 3.407436931879642, + "objective/train/tokens_used": 2425631200, + "theoretical_loss": 3.407436931879642, + "tokens_seen": 2405171200 + }, + { + "epoch": 0.73, + "learning_rate": 0.00021931394182547642, + "loss": 1.3027, + "theoretical_loss": 3.407436931879642, + "tokens_seen": 2405171200 + }, + { + "epoch": 0.73, + "learning_rate": 0.00021929789368104312, + "loss": 1.3489, + "theoretical_loss": 3.407429530916899, + "tokens_seen": 2405236736 + }, + { + "epoch": 0.73, + "learning_rate": 0.00021928184553660982, + "loss": 1.3971, + "theoretical_loss": 3.4074221302122707, + "tokens_seen": 2405302272 + }, + { + "epoch": 0.73, + "learning_rate": 0.00021926579739217652, + "loss": 1.3716, + "theoretical_loss": 3.407414729765741, + "tokens_seen": 2405367808 + }, + { + "epoch": 0.73, + "learning_rate": 0.00021924974924774322, + "loss": 1.3719, + "theoretical_loss": 3.407407329577294, + "tokens_seen": 2405433344 + }, + { + "epoch": 0.73, + "learning_rate": 0.00021923370110330992, + "loss": 1.3311, + "theoretical_loss": 3.4073999296469135, + "tokens_seen": 2405498880 + }, + { + "epoch": 0.73, + "learning_rate": 0.00021921765295887665, + "loss": 1.4461, + "theoretical_loss": 3.407392529974584, + "tokens_seen": 2405564416 + }, + { + "epoch": 0.73, + "learning_rate": 0.00021920160481444335, + "loss": 1.2442, + "theoretical_loss": 3.4073851305602885, + "tokens_seen": 2405629952 + }, + { + "epoch": 0.73, + "learning_rate": 0.00021918555667001005, + "loss": 1.388, + "theoretical_loss": 3.4073777314040115, + "tokens_seen": 2405695488 + }, + { + "epoch": 0.73, + "learning_rate": 0.00021916950852557675, + "loss": 1.2981, + "theoretical_loss": 3.407370332505737, + "tokens_seen": 2405761024 + }, + { + "epoch": 0.73, + "learning_rate": 0.00021915346038114345, + "loss": 1.1927, + "theoretical_loss": 3.4073629338654494, + "tokens_seen": 2405826560 + }, + { + "epoch": 0.73, + "learning_rate": 0.00021913741223671015, + "loss": 1.4204, + "theoretical_loss": 3.407355535483132, + "tokens_seen": 2405892096 + }, + { + "epoch": 0.73, + "learning_rate": 0.00021912136409227685, + "loss": 1.3498, + "theoretical_loss": 3.4073481373587686, + "tokens_seen": 2405957632 + }, + { + "epoch": 0.73, + "learning_rate": 0.00021910531594784355, + "loss": 1.4071, + "theoretical_loss": 3.4073407394923443, + "tokens_seen": 2406023168 + }, + { + "epoch": 0.73, + "learning_rate": 0.00021908926780341026, + "loss": 1.3516, + "theoretical_loss": 3.4073333418838416, + "tokens_seen": 2406088704 + }, + { + "epoch": 0.73, + "learning_rate": 0.00021907321965897696, + "loss": 1.3315, + "theoretical_loss": 3.407325944533246, + "tokens_seen": 2406154240 + }, + { + "epoch": 0.73, + "learning_rate": 0.00021905717151454366, + "loss": 1.4013, + "theoretical_loss": 3.40731854744054, + "tokens_seen": 2406219776 + }, + { + "epoch": 0.73, + "learning_rate": 0.00021904112337011036, + "loss": 1.3666, + "theoretical_loss": 3.407311150605709, + "tokens_seen": 2406285312 + }, + { + "epoch": 0.73, + "learning_rate": 0.00021902507522567706, + "loss": 1.2995, + "theoretical_loss": 3.407303754028736, + "tokens_seen": 2406350848 + }, + { + "epoch": 0.73, + "learning_rate": 0.00021900902708124376, + "loss": 1.2984, + "theoretical_loss": 3.4072963577096056, + "tokens_seen": 2406416384 + }, + { + "epoch": 0.73, + "learning_rate": 0.00021899297893681046, + "loss": 1.2697, + "theoretical_loss": 3.407288961648301, + "tokens_seen": 2406481920 + }, + { + "epoch": 0.73, + "learning_rate": 0.00021897693079237716, + "loss": 1.3816, + "theoretical_loss": 3.407281565844807, + "tokens_seen": 2406547456 + }, + { + "epoch": 0.73, + "learning_rate": 0.00021896088264794386, + "loss": 1.2854, + "theoretical_loss": 3.407274170299107, + "tokens_seen": 2406612992 + }, + { + "epoch": 0.73, + "learning_rate": 0.00021894483450351056, + "loss": 1.3844, + "theoretical_loss": 3.4072667750111854, + "tokens_seen": 2406678528 + }, + { + "epoch": 0.73, + "learning_rate": 0.00021892878635907726, + "loss": 1.3598, + "theoretical_loss": 3.407259379981026, + "tokens_seen": 2406744064 + }, + { + "epoch": 0.73, + "learning_rate": 0.00021891273821464396, + "loss": 1.3226, + "theoretical_loss": 3.407251985208613, + "tokens_seen": 2406809600 + }, + { + "epoch": 0.73, + "learning_rate": 0.00021889669007021066, + "loss": 1.2864, + "theoretical_loss": 3.4072445906939306, + "tokens_seen": 2406875136 + }, + { + "epoch": 0.73, + "learning_rate": 0.00021888064192577736, + "loss": 1.3278, + "theoretical_loss": 3.407237196436962, + "tokens_seen": 2406940672 + }, + { + "epoch": 0.73, + "learning_rate": 0.00021886459378134406, + "loss": 1.2679, + "theoretical_loss": 3.4072298024376915, + "tokens_seen": 2407006208 + }, + { + "epoch": 0.73, + "learning_rate": 0.00021884854563691074, + "loss": 1.3429, + "theoretical_loss": 3.4072224086961036, + "tokens_seen": 2407071744 + }, + { + "epoch": 0.73, + "learning_rate": 0.00021883249749247744, + "loss": 1.3111, + "theoretical_loss": 3.4072150152121816, + "tokens_seen": 2407137280 + }, + { + "epoch": 0.73, + "learning_rate": 0.00021881644934804414, + "loss": 1.2952, + "theoretical_loss": 3.4072076219859104, + "tokens_seen": 2407202816 + }, + { + "epoch": 0.73, + "learning_rate": 0.00021880040120361084, + "loss": 1.4044, + "theoretical_loss": 3.407200229017273, + "tokens_seen": 2407268352 + }, + { + "epoch": 0.73, + "learning_rate": 0.00021878435305917754, + "loss": 1.4459, + "theoretical_loss": 3.4071928363062542, + "tokens_seen": 2407333888 + }, + { + "epoch": 0.73, + "learning_rate": 0.00021876830491474424, + "loss": 1.3265, + "theoretical_loss": 3.4071854438528377, + "tokens_seen": 2407399424 + }, + { + "epoch": 0.73, + "learning_rate": 0.00021875225677031094, + "loss": 1.2662, + "theoretical_loss": 3.407178051657007, + "tokens_seen": 2407464960 + }, + { + "epoch": 0.73, + "learning_rate": 0.00021873620862587764, + "loss": 1.3044, + "theoretical_loss": 3.4071706597187474, + "tokens_seen": 2407530496 + }, + { + "epoch": 0.73, + "learning_rate": 0.00021872016048144434, + "loss": 1.2792, + "theoretical_loss": 3.4071632680380417, + "tokens_seen": 2407596032 + }, + { + "epoch": 0.73, + "learning_rate": 0.00021870411233701104, + "loss": 1.3362, + "theoretical_loss": 3.4071558766148744, + "tokens_seen": 2407661568 + }, + { + "epoch": 0.73, + "learning_rate": 0.00021868806419257774, + "loss": 1.3307, + "theoretical_loss": 3.4071484854492295, + "tokens_seen": 2407727104 + }, + { + "epoch": 0.73, + "learning_rate": 0.00021867201604814444, + "loss": 1.4591, + "theoretical_loss": 3.407141094541091, + "tokens_seen": 2407792640 + }, + { + "epoch": 0.73, + "learning_rate": 0.00021865596790371115, + "loss": 1.4269, + "theoretical_loss": 3.407133703890443, + "tokens_seen": 2407858176 + }, + { + "epoch": 0.73, + "learning_rate": 0.00021863991975927785, + "loss": 1.3729, + "theoretical_loss": 3.4071263134972694, + "tokens_seen": 2407923712 + }, + { + "epoch": 0.73, + "learning_rate": 0.00021862387161484455, + "loss": 1.427, + "theoretical_loss": 3.407118923361554, + "tokens_seen": 2407989248 + }, + { + "epoch": 0.73, + "learning_rate": 0.00021860782347041125, + "loss": 1.3804, + "theoretical_loss": 3.407111533483281, + "tokens_seen": 2408054784 + }, + { + "epoch": 0.73, + "learning_rate": 0.00021859177532597795, + "loss": 1.4177, + "theoretical_loss": 3.407104143862435, + "tokens_seen": 2408120320 + }, + { + "epoch": 0.73, + "learning_rate": 0.00021857572718154465, + "loss": 1.3704, + "theoretical_loss": 3.4070967544989994, + "tokens_seen": 2408185856 + }, + { + "epoch": 0.73, + "learning_rate": 0.00021855967903711135, + "loss": 1.3304, + "theoretical_loss": 3.407089365392958, + "tokens_seen": 2408251392 + }, + { + "epoch": 0.73, + "learning_rate": 0.00021854363089267805, + "loss": 1.2386, + "theoretical_loss": 3.4070819765442955, + "tokens_seen": 2408316928 + }, + { + "epoch": 0.73, + "learning_rate": 0.00021852758274824475, + "loss": 1.3911, + "theoretical_loss": 3.407074587952996, + "tokens_seen": 2408382464 + }, + { + "epoch": 0.73, + "objective/train/docs_used": 876655, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.354355812072754, + "objective/train/theoretical_loss": 3.4070671996190423, + "objective/train/tokens_used": 2428908000, + "theoretical_loss": 3.4070671996190423, + "tokens_seen": 2408448000 + }, + { + "epoch": 0.73, + "learning_rate": 0.00021851153460381145, + "loss": 1.3428, + "theoretical_loss": 3.4070671996190423, + "tokens_seen": 2408448000 + }, + { + "epoch": 0.73, + "learning_rate": 0.00021849548645937815, + "loss": 1.3167, + "theoretical_loss": 3.4070598115424198, + "tokens_seen": 2408513536 + }, + { + "epoch": 0.73, + "learning_rate": 0.00021847943831494485, + "loss": 1.3469, + "theoretical_loss": 3.407052423723112, + "tokens_seen": 2408579072 + }, + { + "epoch": 0.73, + "learning_rate": 0.00021846339017051153, + "loss": 1.4258, + "theoretical_loss": 3.4070450361611027, + "tokens_seen": 2408644608 + }, + { + "epoch": 0.73, + "learning_rate": 0.00021844734202607823, + "loss": 1.2635, + "theoretical_loss": 3.4070376488563765, + "tokens_seen": 2408710144 + }, + { + "epoch": 0.73, + "learning_rate": 0.00021843129388164493, + "loss": 1.2892, + "theoretical_loss": 3.407030261808917, + "tokens_seen": 2408775680 + }, + { + "epoch": 0.73, + "learning_rate": 0.00021841524573721163, + "loss": 1.3059, + "theoretical_loss": 3.407022875018708, + "tokens_seen": 2408841216 + }, + { + "epoch": 0.73, + "learning_rate": 0.00021839919759277833, + "loss": 1.3519, + "theoretical_loss": 3.407015488485735, + "tokens_seen": 2408906752 + }, + { + "epoch": 0.73, + "learning_rate": 0.00021838314944834503, + "loss": 1.3127, + "theoretical_loss": 3.40700810220998, + "tokens_seen": 2408972288 + }, + { + "epoch": 0.73, + "learning_rate": 0.00021836710130391173, + "loss": 1.2546, + "theoretical_loss": 3.407000716191428, + "tokens_seen": 2409037824 + }, + { + "epoch": 0.73, + "learning_rate": 0.00021835105315947843, + "loss": 1.2706, + "theoretical_loss": 3.4069933304300637, + "tokens_seen": 2409103360 + }, + { + "epoch": 0.73, + "learning_rate": 0.00021833500501504513, + "loss": 1.3644, + "theoretical_loss": 3.40698594492587, + "tokens_seen": 2409168896 + }, + { + "epoch": 0.73, + "learning_rate": 0.00021831895687061183, + "loss": 1.2975, + "theoretical_loss": 3.406978559678832, + "tokens_seen": 2409234432 + }, + { + "epoch": 0.73, + "learning_rate": 0.00021830290872617853, + "loss": 1.2655, + "theoretical_loss": 3.4069711746889326, + "tokens_seen": 2409299968 + }, + { + "epoch": 0.73, + "learning_rate": 0.00021828686058174523, + "loss": 1.2828, + "theoretical_loss": 3.4069637899561567, + "tokens_seen": 2409365504 + }, + { + "epoch": 0.73, + "learning_rate": 0.00021827081243731193, + "loss": 1.3769, + "theoretical_loss": 3.406956405480488, + "tokens_seen": 2409431040 + }, + { + "epoch": 0.73, + "learning_rate": 0.00021825476429287863, + "loss": 1.4133, + "theoretical_loss": 3.406949021261911, + "tokens_seen": 2409496576 + }, + { + "epoch": 0.73, + "learning_rate": 0.00021823871614844533, + "loss": 1.3436, + "theoretical_loss": 3.406941637300409, + "tokens_seen": 2409562112 + }, + { + "epoch": 0.73, + "learning_rate": 0.00021822266800401204, + "loss": 1.3164, + "theoretical_loss": 3.406934253595967, + "tokens_seen": 2409627648 + }, + { + "epoch": 0.73, + "learning_rate": 0.00021820661985957874, + "loss": 1.42, + "theoretical_loss": 3.406926870148568, + "tokens_seen": 2409693184 + }, + { + "epoch": 0.73, + "learning_rate": 0.00021819057171514544, + "loss": 1.2004, + "theoretical_loss": 3.4069194869581967, + "tokens_seen": 2409758720 + }, + { + "epoch": 0.73, + "learning_rate": 0.00021817452357071214, + "loss": 1.3494, + "theoretical_loss": 3.4069121040248374, + "tokens_seen": 2409824256 + }, + { + "epoch": 0.73, + "learning_rate": 0.00021815847542627884, + "loss": 1.3266, + "theoretical_loss": 3.4069047213484738, + "tokens_seen": 2409889792 + }, + { + "epoch": 0.73, + "learning_rate": 0.00021814242728184554, + "loss": 1.2538, + "theoretical_loss": 3.40689733892909, + "tokens_seen": 2409955328 + }, + { + "epoch": 0.73, + "learning_rate": 0.00021812637913741224, + "loss": 1.4103, + "theoretical_loss": 3.40688995676667, + "tokens_seen": 2410020864 + }, + { + "epoch": 0.73, + "learning_rate": 0.00021811033099297894, + "loss": 1.3576, + "theoretical_loss": 3.4068825748611977, + "tokens_seen": 2410086400 + }, + { + "epoch": 0.73, + "learning_rate": 0.00021809428284854564, + "loss": 1.3186, + "theoretical_loss": 3.4068751932126577, + "tokens_seen": 2410151936 + }, + { + "epoch": 0.73, + "learning_rate": 0.00021807823470411231, + "loss": 1.2917, + "theoretical_loss": 3.4068678118210336, + "tokens_seen": 2410217472 + }, + { + "epoch": 0.73, + "learning_rate": 0.00021806218655967901, + "loss": 1.2691, + "theoretical_loss": 3.4068604306863097, + "tokens_seen": 2410283008 + }, + { + "epoch": 0.73, + "learning_rate": 0.00021804613841524572, + "loss": 1.3031, + "theoretical_loss": 3.40685304980847, + "tokens_seen": 2410348544 + }, + { + "epoch": 0.73, + "learning_rate": 0.00021803009027081242, + "loss": 1.3932, + "theoretical_loss": 3.406845669187499, + "tokens_seen": 2410414080 + }, + { + "epoch": 0.73, + "learning_rate": 0.00021801404212637917, + "loss": 1.3736, + "theoretical_loss": 3.40683828882338, + "tokens_seen": 2410479616 + }, + { + "epoch": 0.73, + "learning_rate": 0.00021799799398194587, + "loss": 1.326, + "theoretical_loss": 3.4068309087160973, + "tokens_seen": 2410545152 + }, + { + "epoch": 0.73, + "learning_rate": 0.00021798194583751257, + "loss": 1.323, + "theoretical_loss": 3.406823528865635, + "tokens_seen": 2410610688 + }, + { + "epoch": 0.73, + "learning_rate": 0.00021796589769307927, + "loss": 1.3813, + "theoretical_loss": 3.406816149271978, + "tokens_seen": 2410676224 + }, + { + "epoch": 0.73, + "learning_rate": 0.00021794984954864597, + "loss": 1.2787, + "theoretical_loss": 3.4068087699351093, + "tokens_seen": 2410741760 + }, + { + "epoch": 0.73, + "learning_rate": 0.00021793380140421267, + "loss": 1.3023, + "theoretical_loss": 3.406801390855013, + "tokens_seen": 2410807296 + }, + { + "epoch": 0.73, + "learning_rate": 0.00021791775325977937, + "loss": 1.2139, + "theoretical_loss": 3.406794012031674, + "tokens_seen": 2410872832 + }, + { + "epoch": 0.73, + "learning_rate": 0.00021790170511534608, + "loss": 1.3711, + "theoretical_loss": 3.4067866334650763, + "tokens_seen": 2410938368 + }, + { + "epoch": 0.73, + "learning_rate": 0.00021788565697091278, + "loss": 1.32, + "theoretical_loss": 3.406779255155203, + "tokens_seen": 2411003904 + }, + { + "epoch": 0.73, + "learning_rate": 0.00021786960882647948, + "loss": 1.3181, + "theoretical_loss": 3.406771877102039, + "tokens_seen": 2411069440 + }, + { + "epoch": 0.73, + "learning_rate": 0.00021785356068204618, + "loss": 1.2514, + "theoretical_loss": 3.406764499305569, + "tokens_seen": 2411134976 + }, + { + "epoch": 0.73, + "learning_rate": 0.00021783751253761288, + "loss": 1.4445, + "theoretical_loss": 3.4067571217657755, + "tokens_seen": 2411200512 + }, + { + "epoch": 0.73, + "learning_rate": 0.00021782146439317955, + "loss": 1.3113, + "theoretical_loss": 3.4067497444826436, + "tokens_seen": 2411266048 + }, + { + "epoch": 0.73, + "learning_rate": 0.00021780541624874625, + "loss": 1.2745, + "theoretical_loss": 3.406742367456157, + "tokens_seen": 2411331584 + }, + { + "epoch": 0.73, + "learning_rate": 0.00021778936810431295, + "loss": 1.3612, + "theoretical_loss": 3.4067349906863003, + "tokens_seen": 2411397120 + }, + { + "epoch": 0.73, + "learning_rate": 0.00021777331995987965, + "loss": 1.433, + "theoretical_loss": 3.406727614173057, + "tokens_seen": 2411462656 + }, + { + "epoch": 0.73, + "learning_rate": 0.00021775727181544635, + "loss": 1.3052, + "theoretical_loss": 3.406720237916412, + "tokens_seen": 2411528192 + }, + { + "epoch": 0.73, + "learning_rate": 0.00021774122367101305, + "loss": 1.3233, + "theoretical_loss": 3.4067128619163487, + "tokens_seen": 2411593728 + }, + { + "epoch": 0.73, + "learning_rate": 0.00021772517552657976, + "loss": 1.298, + "theoretical_loss": 3.406705486172851, + "tokens_seen": 2411659264 + }, + { + "epoch": 0.73, + "objective/train/docs_used": 877919, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.3863321542739868, + "objective/train/theoretical_loss": 3.4066981106859036, + "objective/train/tokens_used": 2432184800, + "theoretical_loss": 3.4066981106859036, + "tokens_seen": 2411724800 + }, + { + "epoch": 0.73, + "learning_rate": 0.00021770912738214646, + "loss": 1.3585, + "theoretical_loss": 3.4066981106859036, + "tokens_seen": 2411724800 + }, + { + "epoch": 0.73, + "learning_rate": 0.00021769307923771316, + "loss": 1.3217, + "theoretical_loss": 3.4066907354554905, + "tokens_seen": 2411790336 + }, + { + "epoch": 0.73, + "learning_rate": 0.00021767703109327986, + "loss": 1.3684, + "theoretical_loss": 3.406683360481596, + "tokens_seen": 2411855872 + }, + { + "epoch": 0.73, + "learning_rate": 0.00021766098294884656, + "loss": 1.2996, + "theoretical_loss": 3.4066759857642035, + "tokens_seen": 2411921408 + }, + { + "epoch": 0.73, + "learning_rate": 0.00021764493480441326, + "loss": 1.314, + "theoretical_loss": 3.4066686113032976, + "tokens_seen": 2411986944 + }, + { + "epoch": 0.73, + "learning_rate": 0.00021762888665997996, + "loss": 1.3046, + "theoretical_loss": 3.4066612370988625, + "tokens_seen": 2412052480 + }, + { + "epoch": 0.73, + "learning_rate": 0.00021761283851554666, + "loss": 1.3736, + "theoretical_loss": 3.4066538631508823, + "tokens_seen": 2412118016 + }, + { + "epoch": 0.73, + "learning_rate": 0.00021759679037111336, + "loss": 1.4033, + "theoretical_loss": 3.406646489459341, + "tokens_seen": 2412183552 + }, + { + "epoch": 0.73, + "learning_rate": 0.00021758074222668006, + "loss": 1.3399, + "theoretical_loss": 3.4066391160242224, + "tokens_seen": 2412249088 + }, + { + "epoch": 0.73, + "learning_rate": 0.00021756469408224676, + "loss": 1.3005, + "theoretical_loss": 3.406631742845511, + "tokens_seen": 2412314624 + }, + { + "epoch": 0.73, + "learning_rate": 0.00021754864593781346, + "loss": 1.4134, + "theoretical_loss": 3.4066243699231906, + "tokens_seen": 2412380160 + }, + { + "epoch": 0.73, + "learning_rate": 0.00021753259779338016, + "loss": 1.383, + "theoretical_loss": 3.4066169972572458, + "tokens_seen": 2412445696 + }, + { + "epoch": 0.73, + "learning_rate": 0.00021751654964894686, + "loss": 1.3464, + "theoretical_loss": 3.4066096248476607, + "tokens_seen": 2412511232 + }, + { + "epoch": 0.73, + "learning_rate": 0.00021750050150451356, + "loss": 1.331, + "theoretical_loss": 3.4066022526944186, + "tokens_seen": 2412576768 + }, + { + "epoch": 0.73, + "learning_rate": 0.00021748445336008026, + "loss": 1.281, + "theoretical_loss": 3.406594880797505, + "tokens_seen": 2412642304 + }, + { + "epoch": 0.73, + "learning_rate": 0.00021746840521564697, + "loss": 1.3023, + "theoretical_loss": 3.4065875091569024, + "tokens_seen": 2412707840 + }, + { + "epoch": 0.73, + "learning_rate": 0.00021745235707121367, + "loss": 1.3807, + "theoretical_loss": 3.4065801377725964, + "tokens_seen": 2412773376 + }, + { + "epoch": 0.73, + "learning_rate": 0.00021743630892678034, + "loss": 1.3304, + "theoretical_loss": 3.40657276664457, + "tokens_seen": 2412838912 + }, + { + "epoch": 0.73, + "learning_rate": 0.00021742026078234704, + "loss": 1.3498, + "theoretical_loss": 3.4065653957728084, + "tokens_seen": 2412904448 + }, + { + "epoch": 0.73, + "learning_rate": 0.00021740421263791374, + "loss": 1.3284, + "theoretical_loss": 3.4065580251572944, + "tokens_seen": 2412969984 + }, + { + "epoch": 0.73, + "learning_rate": 0.00021738816449348044, + "loss": 1.2696, + "theoretical_loss": 3.4065506547980138, + "tokens_seen": 2413035520 + }, + { + "epoch": 0.73, + "learning_rate": 0.00021737211634904714, + "loss": 1.4157, + "theoretical_loss": 3.406543284694949, + "tokens_seen": 2413101056 + }, + { + "epoch": 0.73, + "learning_rate": 0.00021735606820461384, + "loss": 1.2523, + "theoretical_loss": 3.406535914848085, + "tokens_seen": 2413166592 + }, + { + "epoch": 0.73, + "learning_rate": 0.00021734002006018054, + "loss": 1.359, + "theoretical_loss": 3.4065285452574066, + "tokens_seen": 2413232128 + }, + { + "epoch": 0.73, + "learning_rate": 0.00021732397191574724, + "loss": 1.3987, + "theoretical_loss": 3.406521175922897, + "tokens_seen": 2413297664 + }, + { + "epoch": 0.73, + "learning_rate": 0.00021730792377131394, + "loss": 1.3106, + "theoretical_loss": 3.40651380684454, + "tokens_seen": 2413363200 + }, + { + "epoch": 0.73, + "learning_rate": 0.00021729187562688065, + "loss": 1.3792, + "theoretical_loss": 3.406506438022321, + "tokens_seen": 2413428736 + }, + { + "epoch": 0.73, + "learning_rate": 0.00021727582748244735, + "loss": 1.3582, + "theoretical_loss": 3.406499069456223, + "tokens_seen": 2413494272 + }, + { + "epoch": 0.73, + "learning_rate": 0.00021725977933801405, + "loss": 1.3596, + "theoretical_loss": 3.406491701146231, + "tokens_seen": 2413559808 + }, + { + "epoch": 0.73, + "learning_rate": 0.00021724373119358075, + "loss": 1.3597, + "theoretical_loss": 3.4064843330923282, + "tokens_seen": 2413625344 + }, + { + "epoch": 0.73, + "learning_rate": 0.00021722768304914745, + "loss": 1.3927, + "theoretical_loss": 3.4064769652944995, + "tokens_seen": 2413690880 + }, + { + "epoch": 0.73, + "learning_rate": 0.00021721163490471415, + "loss": 1.2523, + "theoretical_loss": 3.4064695977527286, + "tokens_seen": 2413756416 + }, + { + "epoch": 0.73, + "learning_rate": 0.00021719558676028085, + "loss": 1.2482, + "theoretical_loss": 3.4064622304670005, + "tokens_seen": 2413821952 + }, + { + "epoch": 0.73, + "learning_rate": 0.00021717953861584755, + "loss": 1.3789, + "theoretical_loss": 3.406454863437298, + "tokens_seen": 2413887488 + }, + { + "epoch": 0.73, + "learning_rate": 0.00021716349047141425, + "loss": 1.4077, + "theoretical_loss": 3.4064474966636067, + "tokens_seen": 2413953024 + }, + { + "epoch": 0.73, + "learning_rate": 0.00021714744232698095, + "loss": 1.2185, + "theoretical_loss": 3.4064401301459095, + "tokens_seen": 2414018560 + }, + { + "epoch": 0.73, + "learning_rate": 0.00021713139418254765, + "loss": 1.3682, + "theoretical_loss": 3.4064327638841916, + "tokens_seen": 2414084096 + }, + { + "epoch": 0.73, + "learning_rate": 0.00021711534603811435, + "loss": 1.2993, + "theoretical_loss": 3.4064253978784365, + "tokens_seen": 2414149632 + }, + { + "epoch": 0.73, + "learning_rate": 0.00021709929789368105, + "loss": 1.1847, + "theoretical_loss": 3.406418032128628, + "tokens_seen": 2414215168 + }, + { + "epoch": 0.73, + "learning_rate": 0.00021708324974924775, + "loss": 1.2916, + "theoretical_loss": 3.406410666634751, + "tokens_seen": 2414280704 + }, + { + "epoch": 0.73, + "learning_rate": 0.00021706720160481445, + "loss": 1.2556, + "theoretical_loss": 3.40640330139679, + "tokens_seen": 2414346240 + }, + { + "epoch": 0.73, + "learning_rate": 0.00021705115346038113, + "loss": 1.3638, + "theoretical_loss": 3.406395936414728, + "tokens_seen": 2414411776 + }, + { + "epoch": 0.73, + "learning_rate": 0.00021703510531594783, + "loss": 1.3949, + "theoretical_loss": 3.4063885716885496, + "tokens_seen": 2414477312 + }, + { + "epoch": 0.73, + "learning_rate": 0.00021701905717151453, + "loss": 1.3677, + "theoretical_loss": 3.4063812072182396, + "tokens_seen": 2414542848 + }, + { + "epoch": 0.73, + "learning_rate": 0.00021700300902708123, + "loss": 1.3126, + "theoretical_loss": 3.4063738430037813, + "tokens_seen": 2414608384 + }, + { + "epoch": 0.73, + "learning_rate": 0.00021698696088264793, + "loss": 1.3557, + "theoretical_loss": 3.4063664790451598, + "tokens_seen": 2414673920 + }, + { + "epoch": 0.73, + "learning_rate": 0.00021697091273821463, + "loss": 1.3276, + "theoretical_loss": 3.406359115342358, + "tokens_seen": 2414739456 + }, + { + "epoch": 0.73, + "learning_rate": 0.00021695486459378133, + "loss": 1.3389, + "theoretical_loss": 3.4063517518953614, + "tokens_seen": 2414804992 + }, + { + "epoch": 0.73, + "learning_rate": 0.00021693881644934803, + "loss": 1.2947, + "theoretical_loss": 3.406344388704153, + "tokens_seen": 2414870528 + }, + { + "epoch": 0.73, + "learning_rate": 0.00021692276830491473, + "loss": 1.2037, + "theoretical_loss": 3.406337025768718, + "tokens_seen": 2414936064 + }, + { + "epoch": 0.73, + "objective/train/docs_used": 879191, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.2154524326324463, + "objective/train/theoretical_loss": 3.40632966308904, + "objective/train/tokens_used": 2435461600, + "theoretical_loss": 3.40632966308904, + "tokens_seen": 2415001600 + }, + { + "epoch": 0.73, + "learning_rate": 0.00021690672016048143, + "loss": 1.4167, + "theoretical_loss": 3.40632966308904, + "tokens_seen": 2415001600 + }, + { + "epoch": 0.73, + "learning_rate": 0.00021689067201604813, + "loss": 1.3479, + "theoretical_loss": 3.4063223006651033, + "tokens_seen": 2415067136 + }, + { + "epoch": 0.73, + "learning_rate": 0.00021687462387161483, + "loss": 1.3988, + "theoretical_loss": 3.406314938496892, + "tokens_seen": 2415132672 + }, + { + "epoch": 0.73, + "learning_rate": 0.00021685857572718154, + "loss": 1.3874, + "theoretical_loss": 3.4063075765843904, + "tokens_seen": 2415198208 + }, + { + "epoch": 0.73, + "learning_rate": 0.00021684252758274824, + "loss": 1.3433, + "theoretical_loss": 3.4063002149275823, + "tokens_seen": 2415263744 + }, + { + "epoch": 0.73, + "learning_rate": 0.00021682647943831494, + "loss": 1.2351, + "theoretical_loss": 3.4062928535264527, + "tokens_seen": 2415329280 + }, + { + "epoch": 0.73, + "learning_rate": 0.00021681043129388166, + "loss": 1.3732, + "theoretical_loss": 3.406285492380985, + "tokens_seen": 2415394816 + }, + { + "epoch": 0.73, + "learning_rate": 0.00021679438314944836, + "loss": 1.3351, + "theoretical_loss": 3.4062781314911637, + "tokens_seen": 2415460352 + }, + { + "epoch": 0.73, + "learning_rate": 0.00021677833500501507, + "loss": 1.2743, + "theoretical_loss": 3.406270770856973, + "tokens_seen": 2415525888 + }, + { + "epoch": 0.73, + "learning_rate": 0.00021676228686058177, + "loss": 1.288, + "theoretical_loss": 3.406263410478397, + "tokens_seen": 2415591424 + }, + { + "epoch": 0.73, + "learning_rate": 0.00021674623871614847, + "loss": 1.2881, + "theoretical_loss": 3.40625605035542, + "tokens_seen": 2415656960 + }, + { + "epoch": 0.73, + "learning_rate": 0.00021673019057171517, + "loss": 1.2969, + "theoretical_loss": 3.406248690488026, + "tokens_seen": 2415722496 + }, + { + "epoch": 0.73, + "learning_rate": 0.00021671414242728187, + "loss": 1.2739, + "theoretical_loss": 3.4062413308761994, + "tokens_seen": 2415788032 + }, + { + "epoch": 0.73, + "learning_rate": 0.00021669809428284857, + "loss": 1.3978, + "theoretical_loss": 3.4062339715199244, + "tokens_seen": 2415853568 + }, + { + "epoch": 0.73, + "learning_rate": 0.00021668204613841527, + "loss": 1.3964, + "theoretical_loss": 3.406226612419185, + "tokens_seen": 2415919104 + }, + { + "epoch": 0.73, + "learning_rate": 0.00021666599799398197, + "loss": 1.3536, + "theoretical_loss": 3.4062192535739655, + "tokens_seen": 2415984640 + }, + { + "epoch": 0.73, + "learning_rate": 0.00021664994984954867, + "loss": 1.433, + "theoretical_loss": 3.4062118949842497, + "tokens_seen": 2416050176 + }, + { + "epoch": 0.73, + "learning_rate": 0.00021663390170511537, + "loss": 1.2028, + "theoretical_loss": 3.406204536650023, + "tokens_seen": 2416115712 + }, + { + "epoch": 0.73, + "learning_rate": 0.00021661785356068207, + "loss": 1.3841, + "theoretical_loss": 3.406197178571268, + "tokens_seen": 2416181248 + }, + { + "epoch": 0.73, + "learning_rate": 0.00021660180541624877, + "loss": 1.3705, + "theoretical_loss": 3.40618982074797, + "tokens_seen": 2416246784 + }, + { + "epoch": 0.73, + "learning_rate": 0.00021658575727181547, + "loss": 1.3341, + "theoretical_loss": 3.406182463180113, + "tokens_seen": 2416312320 + }, + { + "epoch": 0.73, + "learning_rate": 0.00021656970912738217, + "loss": 1.3135, + "theoretical_loss": 3.406175105867681, + "tokens_seen": 2416377856 + }, + { + "epoch": 0.73, + "learning_rate": 0.00021655366098294887, + "loss": 1.2002, + "theoretical_loss": 3.4061677488106583, + "tokens_seen": 2416443392 + }, + { + "epoch": 0.73, + "learning_rate": 0.00021653761283851557, + "loss": 1.3545, + "theoretical_loss": 3.406160392009029, + "tokens_seen": 2416508928 + }, + { + "epoch": 0.73, + "learning_rate": 0.00021652156469408228, + "loss": 1.267, + "theoretical_loss": 3.406153035462778, + "tokens_seen": 2416574464 + }, + { + "epoch": 0.73, + "learning_rate": 0.00021650551654964898, + "loss": 1.2778, + "theoretical_loss": 3.406145679171888, + "tokens_seen": 2416640000 + }, + { + "epoch": 0.73, + "learning_rate": 0.00021648946840521568, + "loss": 1.3515, + "theoretical_loss": 3.4061383231363447, + "tokens_seen": 2416705536 + }, + { + "epoch": 0.73, + "learning_rate": 0.00021647342026078238, + "loss": 1.3048, + "theoretical_loss": 3.4061309673561317, + "tokens_seen": 2416771072 + }, + { + "epoch": 0.73, + "learning_rate": 0.00021645737211634908, + "loss": 1.4126, + "theoretical_loss": 3.4061236118312332, + "tokens_seen": 2416836608 + }, + { + "epoch": 0.73, + "learning_rate": 0.00021644132397191578, + "loss": 1.3679, + "theoretical_loss": 3.4061162565616336, + "tokens_seen": 2416902144 + }, + { + "epoch": 0.73, + "learning_rate": 0.00021642527582748245, + "loss": 1.3103, + "theoretical_loss": 3.406108901547317, + "tokens_seen": 2416967680 + }, + { + "epoch": 0.73, + "learning_rate": 0.00021640922768304915, + "loss": 1.2122, + "theoretical_loss": 3.4061015467882676, + "tokens_seen": 2417033216 + }, + { + "epoch": 0.73, + "learning_rate": 0.00021639317953861585, + "loss": 1.383, + "theoretical_loss": 3.4060941922844696, + "tokens_seen": 2417098752 + }, + { + "epoch": 0.73, + "learning_rate": 0.00021637713139418255, + "loss": 1.3051, + "theoretical_loss": 3.406086838035907, + "tokens_seen": 2417164288 + }, + { + "epoch": 0.73, + "learning_rate": 0.00021636108324974925, + "loss": 1.4415, + "theoretical_loss": 3.4060794840425648, + "tokens_seen": 2417229824 + }, + { + "epoch": 0.73, + "learning_rate": 0.00021634503510531596, + "loss": 1.2694, + "theoretical_loss": 3.4060721303044263, + "tokens_seen": 2417295360 + }, + { + "epoch": 0.73, + "learning_rate": 0.00021632898696088266, + "loss": 1.4545, + "theoretical_loss": 3.4060647768214762, + "tokens_seen": 2417360896 + }, + { + "epoch": 0.73, + "learning_rate": 0.00021631293881644936, + "loss": 1.3801, + "theoretical_loss": 3.406057423593699, + "tokens_seen": 2417426432 + }, + { + "epoch": 0.73, + "learning_rate": 0.00021629689067201606, + "loss": 1.3026, + "theoretical_loss": 3.4060500706210783, + "tokens_seen": 2417491968 + }, + { + "epoch": 0.73, + "learning_rate": 0.00021628084252758276, + "loss": 1.3949, + "theoretical_loss": 3.4060427179035986, + "tokens_seen": 2417557504 + }, + { + "epoch": 0.73, + "learning_rate": 0.00021626479438314946, + "loss": 1.2424, + "theoretical_loss": 3.406035365441244, + "tokens_seen": 2417623040 + }, + { + "epoch": 0.73, + "learning_rate": 0.00021624874623871616, + "loss": 1.3772, + "theoretical_loss": 3.406028013233999, + "tokens_seen": 2417688576 + }, + { + "epoch": 0.73, + "learning_rate": 0.00021623269809428286, + "loss": 1.4617, + "theoretical_loss": 3.4060206612818478, + "tokens_seen": 2417754112 + }, + { + "epoch": 0.73, + "learning_rate": 0.00021621664994984956, + "loss": 1.2147, + "theoretical_loss": 3.4060133095847744, + "tokens_seen": 2417819648 + }, + { + "epoch": 0.73, + "learning_rate": 0.00021620060180541626, + "loss": 1.3482, + "theoretical_loss": 3.4060059581427633, + "tokens_seen": 2417885184 + }, + { + "epoch": 0.73, + "learning_rate": 0.00021618455366098296, + "loss": 1.2967, + "theoretical_loss": 3.4059986069557984, + "tokens_seen": 2417950720 + }, + { + "epoch": 0.73, + "learning_rate": 0.00021616850551654966, + "loss": 1.3348, + "theoretical_loss": 3.4059912560238645, + "tokens_seen": 2418016256 + }, + { + "epoch": 0.73, + "learning_rate": 0.00021615245737211636, + "loss": 1.3627, + "theoretical_loss": 3.4059839053469454, + "tokens_seen": 2418081792 + }, + { + "epoch": 0.73, + "learning_rate": 0.00021613640922768306, + "loss": 1.3565, + "theoretical_loss": 3.4059765549250254, + "tokens_seen": 2418147328 + }, + { + "epoch": 0.73, + "learning_rate": 0.00021612036108324976, + "loss": 1.2441, + "theoretical_loss": 3.4059692047580885, + "tokens_seen": 2418212864 + }, + { + "epoch": 0.73, + "objective/train/docs_used": 880852, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.3205599784851074, + "objective/train/theoretical_loss": 3.4059618548461197, + "objective/train/tokens_used": 2438738400, + "theoretical_loss": 3.4059618548461197, + "tokens_seen": 2418278400 + }, + { + "epoch": 0.73, + "learning_rate": 0.00021610431293881646, + "loss": 1.3499, + "theoretical_loss": 3.4059618548461197, + "tokens_seen": 2418278400 + }, + { + "epoch": 0.73, + "learning_rate": 0.00021608826479438317, + "loss": 1.4183, + "theoretical_loss": 3.4059545051891025, + "tokens_seen": 2418343936 + }, + { + "epoch": 0.73, + "learning_rate": 0.00021607221664994987, + "loss": 1.2445, + "theoretical_loss": 3.4059471557870213, + "tokens_seen": 2418409472 + }, + { + "epoch": 0.73, + "learning_rate": 0.00021605616850551657, + "loss": 1.3016, + "theoretical_loss": 3.4059398066398607, + "tokens_seen": 2418475008 + }, + { + "epoch": 0.73, + "learning_rate": 0.00021604012036108324, + "loss": 1.2729, + "theoretical_loss": 3.4059324577476047, + "tokens_seen": 2418540544 + }, + { + "epoch": 0.73, + "learning_rate": 0.00021602407221664994, + "loss": 1.2553, + "theoretical_loss": 3.4059251091102376, + "tokens_seen": 2418606080 + }, + { + "epoch": 0.73, + "learning_rate": 0.00021600802407221664, + "loss": 1.3374, + "theoretical_loss": 3.4059177607277435, + "tokens_seen": 2418671616 + }, + { + "epoch": 0.73, + "learning_rate": 0.00021599197592778334, + "loss": 1.3109, + "theoretical_loss": 3.405910412600107, + "tokens_seen": 2418737152 + }, + { + "epoch": 0.73, + "learning_rate": 0.00021597592778335004, + "loss": 1.2895, + "theoretical_loss": 3.405903064727312, + "tokens_seen": 2418802688 + }, + { + "epoch": 0.73, + "learning_rate": 0.00021595987963891674, + "loss": 1.3592, + "theoretical_loss": 3.4058957171093427, + "tokens_seen": 2418868224 + }, + { + "epoch": 0.73, + "learning_rate": 0.00021594383149448344, + "loss": 1.3143, + "theoretical_loss": 3.405888369746184, + "tokens_seen": 2418933760 + }, + { + "epoch": 0.73, + "learning_rate": 0.00021592778335005014, + "loss": 1.2288, + "theoretical_loss": 3.405881022637819, + "tokens_seen": 2418999296 + }, + { + "epoch": 0.73, + "learning_rate": 0.00021591173520561685, + "loss": 1.2727, + "theoretical_loss": 3.405873675784233, + "tokens_seen": 2419064832 + }, + { + "epoch": 0.73, + "learning_rate": 0.00021589568706118355, + "loss": 1.3659, + "theoretical_loss": 3.4058663291854105, + "tokens_seen": 2419130368 + }, + { + "epoch": 0.73, + "learning_rate": 0.00021587963891675025, + "loss": 1.2406, + "theoretical_loss": 3.405858982841335, + "tokens_seen": 2419195904 + }, + { + "epoch": 0.73, + "learning_rate": 0.00021586359077231695, + "loss": 1.4389, + "theoretical_loss": 3.4058516367519904, + "tokens_seen": 2419261440 + }, + { + "epoch": 0.73, + "learning_rate": 0.00021584754262788365, + "loss": 1.445, + "theoretical_loss": 3.405844290917362, + "tokens_seen": 2419326976 + }, + { + "epoch": 0.73, + "learning_rate": 0.00021583149448345035, + "loss": 1.239, + "theoretical_loss": 3.405836945337433, + "tokens_seen": 2419392512 + }, + { + "epoch": 0.73, + "learning_rate": 0.00021581544633901705, + "loss": 1.2285, + "theoretical_loss": 3.405829600012189, + "tokens_seen": 2419458048 + }, + { + "epoch": 0.73, + "learning_rate": 0.00021579939819458375, + "loss": 1.2762, + "theoretical_loss": 3.4058222549416133, + "tokens_seen": 2419523584 + }, + { + "epoch": 0.73, + "learning_rate": 0.00021578335005015045, + "loss": 1.3935, + "theoretical_loss": 3.4058149101256903, + "tokens_seen": 2419589120 + }, + { + "epoch": 0.73, + "learning_rate": 0.00021576730190571715, + "loss": 1.25, + "theoretical_loss": 3.4058075655644044, + "tokens_seen": 2419654656 + }, + { + "epoch": 0.73, + "learning_rate": 0.00021575125376128385, + "loss": 1.2957, + "theoretical_loss": 3.4058002212577403, + "tokens_seen": 2419720192 + }, + { + "epoch": 0.73, + "learning_rate": 0.00021573520561685055, + "loss": 1.3261, + "theoretical_loss": 3.4057928772056814, + "tokens_seen": 2419785728 + }, + { + "epoch": 0.73, + "learning_rate": 0.00021571915747241725, + "loss": 1.2805, + "theoretical_loss": 3.4057855334082126, + "tokens_seen": 2419851264 + }, + { + "epoch": 0.73, + "learning_rate": 0.00021570310932798395, + "loss": 1.3301, + "theoretical_loss": 3.405778189865318, + "tokens_seen": 2419916800 + }, + { + "epoch": 0.73, + "learning_rate": 0.00021568706118355065, + "loss": 1.2439, + "theoretical_loss": 3.4057708465769814, + "tokens_seen": 2419982336 + }, + { + "epoch": 0.73, + "learning_rate": 0.00021567101303911735, + "loss": 1.3153, + "theoretical_loss": 3.405763503543188, + "tokens_seen": 2420047872 + }, + { + "epoch": 0.73, + "learning_rate": 0.00021565496489468403, + "loss": 1.4472, + "theoretical_loss": 3.4057561607639215, + "tokens_seen": 2420113408 + }, + { + "epoch": 0.73, + "learning_rate": 0.00021563891675025073, + "loss": 1.438, + "theoretical_loss": 3.4057488182391666, + "tokens_seen": 2420178944 + }, + { + "epoch": 0.73, + "learning_rate": 0.00021562286860581743, + "loss": 1.3901, + "theoretical_loss": 3.4057414759689073, + "tokens_seen": 2420244480 + }, + { + "epoch": 0.73, + "learning_rate": 0.00021560682046138418, + "loss": 1.3867, + "theoretical_loss": 3.405734133953128, + "tokens_seen": 2420310016 + }, + { + "epoch": 0.73, + "learning_rate": 0.00021559077231695089, + "loss": 1.2979, + "theoretical_loss": 3.4057267921918126, + "tokens_seen": 2420375552 + }, + { + "epoch": 0.73, + "learning_rate": 0.00021557472417251759, + "loss": 1.3023, + "theoretical_loss": 3.4057194506849457, + "tokens_seen": 2420441088 + }, + { + "epoch": 0.73, + "learning_rate": 0.0002155586760280843, + "loss": 1.1773, + "theoretical_loss": 3.4057121094325113, + "tokens_seen": 2420506624 + }, + { + "epoch": 0.73, + "learning_rate": 0.000215542627883651, + "loss": 1.3226, + "theoretical_loss": 3.4057047684344948, + "tokens_seen": 2420572160 + }, + { + "epoch": 0.73, + "learning_rate": 0.0002155265797392177, + "loss": 1.3123, + "theoretical_loss": 3.405697427690879, + "tokens_seen": 2420637696 + }, + { + "epoch": 0.73, + "learning_rate": 0.0002155105315947844, + "loss": 1.3, + "theoretical_loss": 3.4056900872016493, + "tokens_seen": 2420703232 + }, + { + "epoch": 0.73, + "learning_rate": 0.0002154944834503511, + "loss": 1.2766, + "theoretical_loss": 3.4056827469667894, + "tokens_seen": 2420768768 + }, + { + "epoch": 0.73, + "learning_rate": 0.0002154784353059178, + "loss": 1.3642, + "theoretical_loss": 3.4056754069862833, + "tokens_seen": 2420834304 + }, + { + "epoch": 0.73, + "learning_rate": 0.0002154623871614845, + "loss": 1.2592, + "theoretical_loss": 3.405668067260116, + "tokens_seen": 2420899840 + }, + { + "epoch": 0.73, + "learning_rate": 0.0002154463390170512, + "loss": 1.3133, + "theoretical_loss": 3.405660727788272, + "tokens_seen": 2420965376 + }, + { + "epoch": 0.73, + "learning_rate": 0.0002154302908726179, + "loss": 1.3477, + "theoretical_loss": 3.4056533885707347, + "tokens_seen": 2421030912 + }, + { + "epoch": 0.73, + "learning_rate": 0.0002154142427281846, + "loss": 1.384, + "theoretical_loss": 3.4056460496074896, + "tokens_seen": 2421096448 + }, + { + "epoch": 0.73, + "learning_rate": 0.00021539819458375127, + "loss": 1.3293, + "theoretical_loss": 3.4056387108985193, + "tokens_seen": 2421161984 + }, + { + "epoch": 0.73, + "learning_rate": 0.00021538214643931797, + "loss": 1.3254, + "theoretical_loss": 3.4056313724438096, + "tokens_seen": 2421227520 + }, + { + "epoch": 0.73, + "learning_rate": 0.00021536609829488467, + "loss": 1.3465, + "theoretical_loss": 3.4056240342433446, + "tokens_seen": 2421293056 + }, + { + "epoch": 0.73, + "learning_rate": 0.00021535005015045137, + "loss": 1.4372, + "theoretical_loss": 3.4056166962971077, + "tokens_seen": 2421358592 + }, + { + "epoch": 0.73, + "learning_rate": 0.00021533400200601807, + "loss": 1.2313, + "theoretical_loss": 3.4056093586050844, + "tokens_seen": 2421424128 + }, + { + "epoch": 0.73, + "learning_rate": 0.00021531795386158477, + "loss": 1.3731, + "theoretical_loss": 3.4056020211672577, + "tokens_seen": 2421489664 + }, + { + "epoch": 0.73, + "objective/train/docs_used": 881885, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.1910479068756104, + "objective/train/theoretical_loss": 3.4055946839836135, + "objective/train/tokens_used": 2442015200, + "theoretical_loss": 3.4055946839836135, + "tokens_seen": 2421555200 + }, + { + "epoch": 0.73, + "learning_rate": 0.00021530190571715147, + "loss": 1.2692, + "theoretical_loss": 3.4055946839836135, + "tokens_seen": 2421555200 + }, + { + "epoch": 0.73, + "learning_rate": 0.00021528585757271817, + "loss": 1.2204, + "theoretical_loss": 3.4055873470541345, + "tokens_seen": 2421620736 + }, + { + "epoch": 0.73, + "learning_rate": 0.00021526980942828487, + "loss": 1.3404, + "theoretical_loss": 3.4055800103788063, + "tokens_seen": 2421686272 + }, + { + "epoch": 0.73, + "learning_rate": 0.00021525376128385157, + "loss": 1.3611, + "theoretical_loss": 3.4055726739576127, + "tokens_seen": 2421751808 + }, + { + "epoch": 0.73, + "learning_rate": 0.00021523771313941827, + "loss": 1.309, + "theoretical_loss": 3.4055653377905375, + "tokens_seen": 2421817344 + }, + { + "epoch": 0.73, + "learning_rate": 0.00021522166499498497, + "loss": 1.4264, + "theoretical_loss": 3.4055580018775657, + "tokens_seen": 2421882880 + }, + { + "epoch": 0.73, + "learning_rate": 0.00021520561685055167, + "loss": 1.2461, + "theoretical_loss": 3.405550666218682, + "tokens_seen": 2421948416 + }, + { + "epoch": 0.73, + "learning_rate": 0.00021518956870611837, + "loss": 1.2615, + "theoretical_loss": 3.40554333081387, + "tokens_seen": 2422013952 + }, + { + "epoch": 0.73, + "learning_rate": 0.00021517352056168507, + "loss": 1.3873, + "theoretical_loss": 3.4055359956631137, + "tokens_seen": 2422079488 + }, + { + "epoch": 0.73, + "learning_rate": 0.00021515747241725178, + "loss": 1.1889, + "theoretical_loss": 3.4055286607663984, + "tokens_seen": 2422145024 + }, + { + "epoch": 0.73, + "learning_rate": 0.00021514142427281848, + "loss": 1.3994, + "theoretical_loss": 3.405521326123708, + "tokens_seen": 2422210560 + }, + { + "epoch": 0.73, + "learning_rate": 0.00021512537612838518, + "loss": 1.3263, + "theoretical_loss": 3.4055139917350266, + "tokens_seen": 2422276096 + }, + { + "epoch": 0.73, + "learning_rate": 0.00021510932798395188, + "loss": 1.4029, + "theoretical_loss": 3.4055066576003385, + "tokens_seen": 2422341632 + }, + { + "epoch": 0.73, + "learning_rate": 0.00021509327983951858, + "loss": 1.4657, + "theoretical_loss": 3.4054993237196287, + "tokens_seen": 2422407168 + }, + { + "epoch": 0.73, + "learning_rate": 0.00021507723169508528, + "loss": 1.4311, + "theoretical_loss": 3.405491990092881, + "tokens_seen": 2422472704 + }, + { + "epoch": 0.73, + "learning_rate": 0.00021506118355065198, + "loss": 1.3657, + "theoretical_loss": 3.4054846567200796, + "tokens_seen": 2422538240 + }, + { + "epoch": 0.73, + "learning_rate": 0.00021504513540621868, + "loss": 1.4389, + "theoretical_loss": 3.4054773236012097, + "tokens_seen": 2422603776 + }, + { + "epoch": 0.73, + "learning_rate": 0.00021502908726178538, + "loss": 1.4435, + "theoretical_loss": 3.4054699907362544, + "tokens_seen": 2422669312 + }, + { + "epoch": 0.73, + "learning_rate": 0.00021501303911735205, + "loss": 1.2968, + "theoretical_loss": 3.4054626581251988, + "tokens_seen": 2422734848 + }, + { + "epoch": 0.73, + "learning_rate": 0.00021499699097291875, + "loss": 1.2476, + "theoretical_loss": 3.405455325768027, + "tokens_seen": 2422800384 + }, + { + "epoch": 0.73, + "learning_rate": 0.00021498094282848546, + "loss": 1.3134, + "theoretical_loss": 3.4054479936647235, + "tokens_seen": 2422865920 + }, + { + "epoch": 0.73, + "learning_rate": 0.00021496489468405216, + "loss": 1.43, + "theoretical_loss": 3.405440661815273, + "tokens_seen": 2422931456 + }, + { + "epoch": 0.73, + "learning_rate": 0.00021494884653961886, + "loss": 1.387, + "theoretical_loss": 3.4054333302196587, + "tokens_seen": 2422996992 + }, + { + "epoch": 0.73, + "learning_rate": 0.00021493279839518556, + "loss": 1.2286, + "theoretical_loss": 3.4054259988778663, + "tokens_seen": 2423062528 + }, + { + "epoch": 0.73, + "learning_rate": 0.00021491675025075226, + "loss": 1.314, + "theoretical_loss": 3.405418667789879, + "tokens_seen": 2423128064 + }, + { + "epoch": 0.73, + "learning_rate": 0.00021490070210631896, + "loss": 1.4227, + "theoretical_loss": 3.405411336955682, + "tokens_seen": 2423193600 + }, + { + "epoch": 0.73, + "learning_rate": 0.00021488465396188566, + "loss": 1.2673, + "theoretical_loss": 3.4054040063752593, + "tokens_seen": 2423259136 + }, + { + "epoch": 0.73, + "learning_rate": 0.00021486860581745236, + "loss": 1.2907, + "theoretical_loss": 3.405396676048595, + "tokens_seen": 2423324672 + }, + { + "epoch": 0.73, + "learning_rate": 0.00021485255767301906, + "loss": 1.3644, + "theoretical_loss": 3.405389345975674, + "tokens_seen": 2423390208 + }, + { + "epoch": 0.73, + "learning_rate": 0.00021483650952858576, + "loss": 1.4496, + "theoretical_loss": 3.40538201615648, + "tokens_seen": 2423455744 + }, + { + "epoch": 0.73, + "learning_rate": 0.00021482046138415246, + "loss": 1.2686, + "theoretical_loss": 3.4053746865909984, + "tokens_seen": 2423521280 + }, + { + "epoch": 0.73, + "learning_rate": 0.00021480441323971916, + "loss": 1.3311, + "theoretical_loss": 3.4053673572792125, + "tokens_seen": 2423586816 + }, + { + "epoch": 0.73, + "learning_rate": 0.00021478836509528586, + "loss": 1.3685, + "theoretical_loss": 3.405360028221107, + "tokens_seen": 2423652352 + }, + { + "epoch": 0.73, + "learning_rate": 0.00021477231695085256, + "loss": 1.2446, + "theoretical_loss": 3.4053526994166665, + "tokens_seen": 2423717888 + }, + { + "epoch": 0.73, + "learning_rate": 0.00021475626880641926, + "loss": 1.275, + "theoretical_loss": 3.405345370865875, + "tokens_seen": 2423783424 + }, + { + "epoch": 0.73, + "learning_rate": 0.00021474022066198596, + "loss": 1.2022, + "theoretical_loss": 3.405338042568717, + "tokens_seen": 2423848960 + }, + { + "epoch": 0.73, + "learning_rate": 0.00021472417251755267, + "loss": 1.2482, + "theoretical_loss": 3.405330714525177, + "tokens_seen": 2423914496 + }, + { + "epoch": 0.73, + "learning_rate": 0.00021470812437311937, + "loss": 1.3954, + "theoretical_loss": 3.4053233867352395, + "tokens_seen": 2423980032 + }, + { + "epoch": 0.73, + "learning_rate": 0.00021469207622868607, + "loss": 1.2806, + "theoretical_loss": 3.405316059198888, + "tokens_seen": 2424045568 + }, + { + "epoch": 0.73, + "learning_rate": 0.00021467602808425277, + "loss": 1.2559, + "theoretical_loss": 3.405308731916108, + "tokens_seen": 2424111104 + }, + { + "epoch": 0.73, + "learning_rate": 0.00021465997993981947, + "loss": 1.2149, + "theoretical_loss": 3.4053014048868833, + "tokens_seen": 2424176640 + }, + { + "epoch": 0.73, + "learning_rate": 0.00021464393179538617, + "loss": 1.2974, + "theoretical_loss": 3.4052940781111984, + "tokens_seen": 2424242176 + }, + { + "epoch": 0.73, + "learning_rate": 0.00021462788365095284, + "loss": 1.3588, + "theoretical_loss": 3.4052867515890375, + "tokens_seen": 2424307712 + }, + { + "epoch": 0.73, + "learning_rate": 0.00021461183550651954, + "loss": 1.3102, + "theoretical_loss": 3.405279425320385, + "tokens_seen": 2424373248 + }, + { + "epoch": 0.73, + "learning_rate": 0.00021459578736208624, + "loss": 1.3635, + "theoretical_loss": 3.4052720993052255, + "tokens_seen": 2424438784 + }, + { + "epoch": 0.73, + "learning_rate": 0.00021457973921765294, + "loss": 1.2411, + "theoretical_loss": 3.4052647735435433, + "tokens_seen": 2424504320 + }, + { + "epoch": 0.73, + "learning_rate": 0.00021456369107321964, + "loss": 1.3512, + "theoretical_loss": 3.4052574480353224, + "tokens_seen": 2424569856 + }, + { + "epoch": 0.73, + "learning_rate": 0.00021454764292878635, + "loss": 1.3038, + "theoretical_loss": 3.405250122780548, + "tokens_seen": 2424635392 + }, + { + "epoch": 0.73, + "learning_rate": 0.00021453159478435305, + "loss": 1.2652, + "theoretical_loss": 3.405242797779204, + "tokens_seen": 2424700928 + }, + { + "epoch": 0.73, + "learning_rate": 0.00021451554663991975, + "loss": 1.3964, + "theoretical_loss": 3.405235473031274, + "tokens_seen": 2424766464 + }, + { + "epoch": 0.73, + "objective/train/docs_used": 882990, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.4437370300292969, + "objective/train/theoretical_loss": 3.405228148536744, + "objective/train/tokens_used": 2445292000, + "theoretical_loss": 3.405228148536744, + "tokens_seen": 2424832000 + }, + { + "epoch": 0.73, + "learning_rate": 0.00021449949849548645, + "loss": 1.3631, + "theoretical_loss": 3.405228148536744, + "tokens_seen": 2424832000 + }, + { + "epoch": 0.73, + "learning_rate": 0.00021448345035105315, + "loss": 1.2795, + "theoretical_loss": 3.405220824295597, + "tokens_seen": 2424897536 + }, + { + "epoch": 0.73, + "learning_rate": 0.00021446740220661985, + "loss": 1.4565, + "theoretical_loss": 3.405213500307818, + "tokens_seen": 2424963072 + }, + { + "epoch": 0.73, + "learning_rate": 0.00021445135406218655, + "loss": 1.3346, + "theoretical_loss": 3.4052061765733916, + "tokens_seen": 2425028608 + }, + { + "epoch": 0.73, + "learning_rate": 0.00021443530591775325, + "loss": 1.3093, + "theoretical_loss": 3.4051988530923016, + "tokens_seen": 2425094144 + }, + { + "epoch": 0.73, + "learning_rate": 0.00021441925777331995, + "loss": 1.4243, + "theoretical_loss": 3.405191529864533, + "tokens_seen": 2425159680 + }, + { + "epoch": 0.73, + "learning_rate": 0.0002144032096288867, + "loss": 1.3655, + "theoretical_loss": 3.4051842068900697, + "tokens_seen": 2425225216 + }, + { + "epoch": 0.73, + "learning_rate": 0.00021438716148445338, + "loss": 1.3207, + "theoretical_loss": 3.4051768841688963, + "tokens_seen": 2425290752 + }, + { + "epoch": 0.73, + "learning_rate": 0.00021437111334002008, + "loss": 1.367, + "theoretical_loss": 3.4051695617009967, + "tokens_seen": 2425356288 + }, + { + "epoch": 0.73, + "learning_rate": 0.00021435506519558678, + "loss": 1.3768, + "theoretical_loss": 3.4051622394863563, + "tokens_seen": 2425421824 + }, + { + "epoch": 0.73, + "learning_rate": 0.00021433901705115348, + "loss": 1.3079, + "theoretical_loss": 3.405154917524959, + "tokens_seen": 2425487360 + }, + { + "epoch": 0.74, + "learning_rate": 0.00021432296890672018, + "loss": 1.3622, + "theoretical_loss": 3.405147595816789, + "tokens_seen": 2425552896 + }, + { + "epoch": 0.74, + "learning_rate": 0.00021430692076228688, + "loss": 1.2667, + "theoretical_loss": 3.405140274361831, + "tokens_seen": 2425618432 + }, + { + "epoch": 0.74, + "learning_rate": 0.00021429087261785358, + "loss": 1.4353, + "theoretical_loss": 3.405132953160069, + "tokens_seen": 2425683968 + }, + { + "epoch": 0.74, + "learning_rate": 0.00021427482447342028, + "loss": 1.3855, + "theoretical_loss": 3.405125632211488, + "tokens_seen": 2425749504 + }, + { + "epoch": 0.74, + "learning_rate": 0.00021425877632898698, + "loss": 1.1921, + "theoretical_loss": 3.405118311516072, + "tokens_seen": 2425815040 + }, + { + "epoch": 0.74, + "learning_rate": 0.00021424272818455368, + "loss": 1.3163, + "theoretical_loss": 3.4051109910738053, + "tokens_seen": 2425880576 + }, + { + "epoch": 0.74, + "learning_rate": 0.00021422668004012038, + "loss": 1.2174, + "theoretical_loss": 3.4051036708846727, + "tokens_seen": 2425946112 + }, + { + "epoch": 0.74, + "learning_rate": 0.00021421063189568709, + "loss": 1.4354, + "theoretical_loss": 3.405096350948658, + "tokens_seen": 2426011648 + }, + { + "epoch": 0.74, + "learning_rate": 0.00021419458375125379, + "loss": 1.3762, + "theoretical_loss": 3.4050890312657467, + "tokens_seen": 2426077184 + }, + { + "epoch": 0.74, + "learning_rate": 0.0002141785356068205, + "loss": 1.2793, + "theoretical_loss": 3.4050817118359222, + "tokens_seen": 2426142720 + }, + { + "epoch": 0.74, + "learning_rate": 0.0002141624874623872, + "loss": 1.3298, + "theoretical_loss": 3.405074392659169, + "tokens_seen": 2426208256 + }, + { + "epoch": 0.74, + "learning_rate": 0.0002141464393179539, + "loss": 1.2999, + "theoretical_loss": 3.405067073735472, + "tokens_seen": 2426273792 + }, + { + "epoch": 0.74, + "learning_rate": 0.0002141303911735206, + "loss": 1.2886, + "theoretical_loss": 3.405059755064815, + "tokens_seen": 2426339328 + }, + { + "epoch": 0.74, + "learning_rate": 0.0002141143430290873, + "loss": 1.4336, + "theoretical_loss": 3.4050524366471833, + "tokens_seen": 2426404864 + }, + { + "epoch": 0.74, + "learning_rate": 0.000214098294884654, + "loss": 1.3675, + "theoretical_loss": 3.4050451184825605, + "tokens_seen": 2426470400 + }, + { + "epoch": 0.74, + "learning_rate": 0.0002140822467402207, + "loss": 1.3292, + "theoretical_loss": 3.4050378005709314, + "tokens_seen": 2426535936 + }, + { + "epoch": 0.74, + "learning_rate": 0.0002140661985957874, + "loss": 1.3737, + "theoretical_loss": 3.4050304829122804, + "tokens_seen": 2426601472 + }, + { + "epoch": 0.74, + "learning_rate": 0.0002140501504513541, + "loss": 1.3292, + "theoretical_loss": 3.405023165506592, + "tokens_seen": 2426667008 + }, + { + "epoch": 0.74, + "learning_rate": 0.0002140341023069208, + "loss": 1.3167, + "theoretical_loss": 3.40501584835385, + "tokens_seen": 2426732544 + }, + { + "epoch": 0.74, + "learning_rate": 0.0002140180541624875, + "loss": 1.3083, + "theoretical_loss": 3.4050085314540395, + "tokens_seen": 2426798080 + }, + { + "epoch": 0.74, + "learning_rate": 0.00021400200601805417, + "loss": 1.2865, + "theoretical_loss": 3.405001214807145, + "tokens_seen": 2426863616 + }, + { + "epoch": 0.74, + "learning_rate": 0.00021398595787362087, + "loss": 1.4217, + "theoretical_loss": 3.4049938984131507, + "tokens_seen": 2426929152 + }, + { + "epoch": 0.74, + "learning_rate": 0.00021396990972918757, + "loss": 1.188, + "theoretical_loss": 3.4049865822720404, + "tokens_seen": 2426994688 + }, + { + "epoch": 0.74, + "learning_rate": 0.00021395386158475427, + "loss": 1.2595, + "theoretical_loss": 3.4049792663838, + "tokens_seen": 2427060224 + }, + { + "epoch": 0.74, + "learning_rate": 0.00021393781344032097, + "loss": 1.358, + "theoretical_loss": 3.404971950748412, + "tokens_seen": 2427125760 + }, + { + "epoch": 0.74, + "learning_rate": 0.00021392176529588767, + "loss": 1.3232, + "theoretical_loss": 3.4049646353658627, + "tokens_seen": 2427191296 + }, + { + "epoch": 0.74, + "learning_rate": 0.00021390571715145437, + "loss": 1.2503, + "theoretical_loss": 3.404957320236136, + "tokens_seen": 2427256832 + }, + { + "epoch": 0.74, + "learning_rate": 0.00021388966900702107, + "loss": 1.2948, + "theoretical_loss": 3.404950005359215, + "tokens_seen": 2427322368 + }, + { + "epoch": 0.74, + "learning_rate": 0.00021387362086258777, + "loss": 1.3483, + "theoretical_loss": 3.404942690735086, + "tokens_seen": 2427387904 + }, + { + "epoch": 0.74, + "learning_rate": 0.00021385757271815447, + "loss": 1.2119, + "theoretical_loss": 3.4049353763637327, + "tokens_seen": 2427453440 + }, + { + "epoch": 0.74, + "learning_rate": 0.00021384152457372117, + "loss": 1.3709, + "theoretical_loss": 3.404928062245139, + "tokens_seen": 2427518976 + }, + { + "epoch": 0.74, + "learning_rate": 0.00021382547642928787, + "loss": 1.2733, + "theoretical_loss": 3.40492074837929, + "tokens_seen": 2427584512 + }, + { + "epoch": 0.74, + "learning_rate": 0.00021380942828485457, + "loss": 1.364, + "theoretical_loss": 3.4049134347661703, + "tokens_seen": 2427650048 + }, + { + "epoch": 0.74, + "learning_rate": 0.00021379338014042127, + "loss": 1.3036, + "theoretical_loss": 3.4049061214057637, + "tokens_seen": 2427715584 + }, + { + "epoch": 0.74, + "learning_rate": 0.00021377733199598798, + "loss": 1.2206, + "theoretical_loss": 3.404898808298055, + "tokens_seen": 2427781120 + }, + { + "epoch": 0.74, + "learning_rate": 0.00021376128385155468, + "loss": 1.3633, + "theoretical_loss": 3.4048914954430285, + "tokens_seen": 2427846656 + }, + { + "epoch": 0.74, + "learning_rate": 0.00021374523570712138, + "loss": 1.3317, + "theoretical_loss": 3.404884182840669, + "tokens_seen": 2427912192 + }, + { + "epoch": 0.74, + "learning_rate": 0.00021372918756268808, + "loss": 1.419, + "theoretical_loss": 3.4048768704909604, + "tokens_seen": 2427977728 + }, + { + "epoch": 0.74, + "learning_rate": 0.00021371313941825478, + "loss": 1.3819, + "theoretical_loss": 3.4048695583938877, + "tokens_seen": 2428043264 + }, + { + "epoch": 0.74, + "objective/train/docs_used": 884091, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.2091819047927856, + "objective/train/theoretical_loss": 3.404862246549435, + "objective/train/tokens_used": 2448568800, + "theoretical_loss": 3.404862246549435, + "tokens_seen": 2428108800 + }, + { + "epoch": 0.74, + "learning_rate": 0.00021369709127382148, + "loss": 1.3318, + "theoretical_loss": 3.404862246549435, + "tokens_seen": 2428108800 + }, + { + "epoch": 0.74, + "learning_rate": 0.00021368104312938818, + "loss": 1.2551, + "theoretical_loss": 3.404854934957587, + "tokens_seen": 2428174336 + }, + { + "epoch": 0.74, + "learning_rate": 0.00021366499498495488, + "loss": 1.3111, + "theoretical_loss": 3.404847623618328, + "tokens_seen": 2428239872 + }, + { + "epoch": 0.74, + "learning_rate": 0.00021364894684052158, + "loss": 1.3509, + "theoretical_loss": 3.4048403125316424, + "tokens_seen": 2428305408 + }, + { + "epoch": 0.74, + "learning_rate": 0.00021363289869608828, + "loss": 1.4741, + "theoretical_loss": 3.404833001697515, + "tokens_seen": 2428370944 + }, + { + "epoch": 0.74, + "learning_rate": 0.00021361685055165495, + "loss": 1.2467, + "theoretical_loss": 3.4048256911159296, + "tokens_seen": 2428436480 + }, + { + "epoch": 0.74, + "learning_rate": 0.00021360080240722166, + "loss": 1.2359, + "theoretical_loss": 3.404818380786871, + "tokens_seen": 2428502016 + }, + { + "epoch": 0.74, + "learning_rate": 0.00021358475426278836, + "loss": 1.3625, + "theoretical_loss": 3.4048110707103243, + "tokens_seen": 2428567552 + }, + { + "epoch": 0.74, + "learning_rate": 0.00021356870611835506, + "loss": 1.3115, + "theoretical_loss": 3.404803760886273, + "tokens_seen": 2428633088 + }, + { + "epoch": 0.74, + "learning_rate": 0.00021355265797392176, + "loss": 1.2971, + "theoretical_loss": 3.4047964513147018, + "tokens_seen": 2428698624 + }, + { + "epoch": 0.74, + "learning_rate": 0.00021353660982948846, + "loss": 1.3188, + "theoretical_loss": 3.404789141995596, + "tokens_seen": 2428764160 + }, + { + "epoch": 0.74, + "learning_rate": 0.00021352056168505516, + "loss": 1.2569, + "theoretical_loss": 3.4047818329289385, + "tokens_seen": 2428829696 + }, + { + "epoch": 0.74, + "learning_rate": 0.00021350451354062186, + "loss": 1.4316, + "theoretical_loss": 3.404774524114715, + "tokens_seen": 2428895232 + }, + { + "epoch": 0.74, + "learning_rate": 0.00021348846539618856, + "loss": 1.304, + "theoretical_loss": 3.40476721555291, + "tokens_seen": 2428960768 + }, + { + "epoch": 0.74, + "learning_rate": 0.00021347241725175526, + "loss": 1.3003, + "theoretical_loss": 3.404759907243507, + "tokens_seen": 2429026304 + }, + { + "epoch": 0.74, + "learning_rate": 0.00021345636910732196, + "loss": 1.3488, + "theoretical_loss": 3.4047525991864913, + "tokens_seen": 2429091840 + }, + { + "epoch": 0.74, + "learning_rate": 0.00021344032096288866, + "loss": 1.2703, + "theoretical_loss": 3.4047452913818472, + "tokens_seen": 2429157376 + }, + { + "epoch": 0.74, + "learning_rate": 0.00021342427281845536, + "loss": 1.3292, + "theoretical_loss": 3.404737983829559, + "tokens_seen": 2429222912 + }, + { + "epoch": 0.74, + "learning_rate": 0.00021340822467402206, + "loss": 1.3309, + "theoretical_loss": 3.4047306765296117, + "tokens_seen": 2429288448 + }, + { + "epoch": 0.74, + "learning_rate": 0.00021339217652958876, + "loss": 1.3244, + "theoretical_loss": 3.404723369481989, + "tokens_seen": 2429353984 + }, + { + "epoch": 0.74, + "learning_rate": 0.00021337612838515546, + "loss": 1.2876, + "theoretical_loss": 3.404716062686676, + "tokens_seen": 2429419520 + }, + { + "epoch": 0.74, + "learning_rate": 0.00021336008024072216, + "loss": 1.2646, + "theoretical_loss": 3.4047087561436564, + "tokens_seen": 2429485056 + }, + { + "epoch": 0.74, + "learning_rate": 0.00021334403209628887, + "loss": 1.2602, + "theoretical_loss": 3.404701449852916, + "tokens_seen": 2429550592 + }, + { + "epoch": 0.74, + "learning_rate": 0.00021332798395185557, + "loss": 1.4336, + "theoretical_loss": 3.4046941438144382, + "tokens_seen": 2429616128 + }, + { + "epoch": 0.74, + "learning_rate": 0.00021331193580742227, + "loss": 1.28, + "theoretical_loss": 3.4046868380282076, + "tokens_seen": 2429681664 + }, + { + "epoch": 0.74, + "learning_rate": 0.00021329588766298897, + "loss": 1.2968, + "theoretical_loss": 3.4046795324942094, + "tokens_seen": 2429747200 + }, + { + "epoch": 0.74, + "learning_rate": 0.00021327983951855567, + "loss": 1.352, + "theoretical_loss": 3.404672227212427, + "tokens_seen": 2429812736 + }, + { + "epoch": 0.74, + "learning_rate": 0.00021326379137412237, + "loss": 1.2532, + "theoretical_loss": 3.4046649221828456, + "tokens_seen": 2429878272 + }, + { + "epoch": 0.74, + "learning_rate": 0.00021324774322968907, + "loss": 1.2586, + "theoretical_loss": 3.4046576174054497, + "tokens_seen": 2429943808 + }, + { + "epoch": 0.74, + "learning_rate": 0.00021323169508525574, + "loss": 1.2512, + "theoretical_loss": 3.4046503128802237, + "tokens_seen": 2430009344 + }, + { + "epoch": 0.74, + "learning_rate": 0.00021321564694082244, + "loss": 1.4674, + "theoretical_loss": 3.404643008607152, + "tokens_seen": 2430074880 + }, + { + "epoch": 0.74, + "learning_rate": 0.0002131995987963892, + "loss": 1.3884, + "theoretical_loss": 3.404635704586219, + "tokens_seen": 2430140416 + }, + { + "epoch": 0.74, + "learning_rate": 0.0002131835506519559, + "loss": 1.4078, + "theoretical_loss": 3.4046284008174093, + "tokens_seen": 2430205952 + }, + { + "epoch": 0.74, + "learning_rate": 0.0002131675025075226, + "loss": 1.3285, + "theoretical_loss": 3.4046210973007076, + "tokens_seen": 2430271488 + }, + { + "epoch": 0.74, + "learning_rate": 0.0002131514543630893, + "loss": 1.3518, + "theoretical_loss": 3.404613794036098, + "tokens_seen": 2430337024 + }, + { + "epoch": 0.74, + "learning_rate": 0.000213135406218656, + "loss": 1.3483, + "theoretical_loss": 3.4046064910235656, + "tokens_seen": 2430402560 + }, + { + "epoch": 0.74, + "learning_rate": 0.0002131193580742227, + "loss": 1.3427, + "theoretical_loss": 3.404599188263094, + "tokens_seen": 2430468096 + }, + { + "epoch": 0.74, + "learning_rate": 0.0002131033099297894, + "loss": 1.381, + "theoretical_loss": 3.4045918857546686, + "tokens_seen": 2430533632 + }, + { + "epoch": 0.74, + "learning_rate": 0.0002130872617853561, + "loss": 1.3069, + "theoretical_loss": 3.404584583498274, + "tokens_seen": 2430599168 + }, + { + "epoch": 0.74, + "learning_rate": 0.0002130712136409228, + "loss": 1.2471, + "theoretical_loss": 3.4045772814938937, + "tokens_seen": 2430664704 + }, + { + "epoch": 0.74, + "learning_rate": 0.0002130551654964895, + "loss": 1.368, + "theoretical_loss": 3.4045699797415128, + "tokens_seen": 2430730240 + }, + { + "epoch": 0.74, + "learning_rate": 0.0002130391173520562, + "loss": 1.357, + "theoretical_loss": 3.404562678241116, + "tokens_seen": 2430795776 + }, + { + "epoch": 0.74, + "learning_rate": 0.0002130230692076229, + "loss": 1.2835, + "theoretical_loss": 3.404555376992687, + "tokens_seen": 2430861312 + }, + { + "epoch": 0.74, + "learning_rate": 0.0002130070210631896, + "loss": 1.3431, + "theoretical_loss": 3.4045480759962117, + "tokens_seen": 2430926848 + }, + { + "epoch": 0.74, + "learning_rate": 0.00021299097291875628, + "loss": 1.3748, + "theoretical_loss": 3.4045407752516734, + "tokens_seen": 2430992384 + }, + { + "epoch": 0.74, + "learning_rate": 0.00021297492477432298, + "loss": 1.3042, + "theoretical_loss": 3.4045334747590568, + "tokens_seen": 2431057920 + }, + { + "epoch": 0.74, + "learning_rate": 0.00021295887662988968, + "loss": 1.2677, + "theoretical_loss": 3.404526174518347, + "tokens_seen": 2431123456 + }, + { + "epoch": 0.74, + "learning_rate": 0.00021294282848545638, + "loss": 1.3013, + "theoretical_loss": 3.4045188745295283, + "tokens_seen": 2431188992 + }, + { + "epoch": 0.74, + "learning_rate": 0.00021292678034102308, + "loss": 1.3479, + "theoretical_loss": 3.4045115747925845, + "tokens_seen": 2431254528 + }, + { + "epoch": 0.74, + "learning_rate": 0.00021291073219658978, + "loss": 1.2834, + "theoretical_loss": 3.404504275307501, + "tokens_seen": 2431320064 + }, + { + "epoch": 0.74, + "objective/train/docs_used": 885315, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.3404672145843506, + "objective/train/theoretical_loss": 3.4044969760742623, + "objective/train/tokens_used": 2451845600, + "theoretical_loss": 3.4044969760742623, + "tokens_seen": 2431385600 + }, + { + "epoch": 0.74, + "learning_rate": 0.00021289468405215648, + "loss": 1.3608, + "theoretical_loss": 3.4044969760742623, + "tokens_seen": 2431385600 + }, + { + "epoch": 0.74, + "learning_rate": 0.00021287863590772318, + "loss": 1.3991, + "theoretical_loss": 3.404489677092852, + "tokens_seen": 2431451136 + }, + { + "epoch": 0.74, + "learning_rate": 0.00021286258776328988, + "loss": 1.2869, + "theoretical_loss": 3.404482378363256, + "tokens_seen": 2431516672 + }, + { + "epoch": 0.74, + "learning_rate": 0.00021284653961885659, + "loss": 1.3556, + "theoretical_loss": 3.4044750798854575, + "tokens_seen": 2431582208 + }, + { + "epoch": 0.74, + "learning_rate": 0.00021283049147442329, + "loss": 1.3715, + "theoretical_loss": 3.4044677816594415, + "tokens_seen": 2431647744 + }, + { + "epoch": 0.74, + "learning_rate": 0.00021281444332999, + "loss": 1.2883, + "theoretical_loss": 3.4044604836851935, + "tokens_seen": 2431713280 + }, + { + "epoch": 0.74, + "learning_rate": 0.0002127983951855567, + "loss": 1.3521, + "theoretical_loss": 3.4044531859626965, + "tokens_seen": 2431778816 + }, + { + "epoch": 0.74, + "learning_rate": 0.0002127823470411234, + "loss": 1.3541, + "theoretical_loss": 3.404445888491936, + "tokens_seen": 2431844352 + }, + { + "epoch": 0.74, + "learning_rate": 0.0002127662988966901, + "loss": 1.1613, + "theoretical_loss": 3.404438591272896, + "tokens_seen": 2431909888 + }, + { + "epoch": 0.74, + "learning_rate": 0.0002127502507522568, + "loss": 1.3443, + "theoretical_loss": 3.404431294305561, + "tokens_seen": 2431975424 + }, + { + "epoch": 0.74, + "learning_rate": 0.0002127342026078235, + "loss": 1.3408, + "theoretical_loss": 3.4044239975899164, + "tokens_seen": 2432040960 + }, + { + "epoch": 0.74, + "learning_rate": 0.0002127181544633902, + "loss": 1.3233, + "theoretical_loss": 3.404416701125946, + "tokens_seen": 2432106496 + }, + { + "epoch": 0.74, + "learning_rate": 0.0002127021063189569, + "loss": 1.2667, + "theoretical_loss": 3.4044094049136344, + "tokens_seen": 2432172032 + }, + { + "epoch": 0.74, + "learning_rate": 0.0002126860581745236, + "loss": 1.3072, + "theoretical_loss": 3.4044021089529664, + "tokens_seen": 2432237568 + }, + { + "epoch": 0.74, + "learning_rate": 0.0002126700100300903, + "loss": 1.2212, + "theoretical_loss": 3.404394813243926, + "tokens_seen": 2432303104 + }, + { + "epoch": 0.74, + "learning_rate": 0.000212653961885657, + "loss": 1.2607, + "theoretical_loss": 3.4043875177864984, + "tokens_seen": 2432368640 + }, + { + "epoch": 0.74, + "learning_rate": 0.0002126379137412237, + "loss": 1.3487, + "theoretical_loss": 3.4043802225806674, + "tokens_seen": 2432434176 + }, + { + "epoch": 0.74, + "learning_rate": 0.0002126218655967904, + "loss": 1.2747, + "theoretical_loss": 3.4043729276264187, + "tokens_seen": 2432499712 + }, + { + "epoch": 0.74, + "learning_rate": 0.00021260581745235707, + "loss": 1.3825, + "theoretical_loss": 3.4043656329237355, + "tokens_seen": 2432565248 + }, + { + "epoch": 0.74, + "learning_rate": 0.00021258976930792377, + "loss": 1.2592, + "theoretical_loss": 3.4043583384726035, + "tokens_seen": 2432630784 + }, + { + "epoch": 0.74, + "learning_rate": 0.00021257372116349047, + "loss": 1.3704, + "theoretical_loss": 3.4043510442730067, + "tokens_seen": 2432696320 + }, + { + "epoch": 0.74, + "learning_rate": 0.00021255767301905717, + "loss": 1.3466, + "theoretical_loss": 3.404343750324929, + "tokens_seen": 2432761856 + }, + { + "epoch": 0.74, + "learning_rate": 0.00021254162487462387, + "loss": 1.3121, + "theoretical_loss": 3.4043364566283563, + "tokens_seen": 2432827392 + }, + { + "epoch": 0.74, + "learning_rate": 0.00021252557673019057, + "loss": 1.3117, + "theoretical_loss": 3.4043291631832724, + "tokens_seen": 2432892928 + }, + { + "epoch": 0.74, + "learning_rate": 0.00021250952858575727, + "loss": 1.4199, + "theoretical_loss": 3.404321869989662, + "tokens_seen": 2432958464 + }, + { + "epoch": 0.74, + "learning_rate": 0.00021249348044132397, + "loss": 1.4296, + "theoretical_loss": 3.4043145770475096, + "tokens_seen": 2433024000 + }, + { + "epoch": 0.74, + "learning_rate": 0.00021247743229689067, + "loss": 1.3302, + "theoretical_loss": 3.4043072843567996, + "tokens_seen": 2433089536 + }, + { + "epoch": 0.74, + "learning_rate": 0.00021246138415245737, + "loss": 1.2372, + "theoretical_loss": 3.4042999919175165, + "tokens_seen": 2433155072 + }, + { + "epoch": 0.74, + "learning_rate": 0.00021244533600802407, + "loss": 1.3361, + "theoretical_loss": 3.4042926997296457, + "tokens_seen": 2433220608 + }, + { + "epoch": 0.74, + "learning_rate": 0.00021242928786359077, + "loss": 1.301, + "theoretical_loss": 3.4042854077931706, + "tokens_seen": 2433286144 + }, + { + "epoch": 0.74, + "learning_rate": 0.00021241323971915748, + "loss": 1.3899, + "theoretical_loss": 3.4042781161080766, + "tokens_seen": 2433351680 + }, + { + "epoch": 0.74, + "learning_rate": 0.00021239719157472418, + "loss": 1.3875, + "theoretical_loss": 3.404270824674348, + "tokens_seen": 2433417216 + }, + { + "epoch": 0.74, + "learning_rate": 0.00021238114343029088, + "loss": 1.2529, + "theoretical_loss": 3.404263533491969, + "tokens_seen": 2433482752 + }, + { + "epoch": 0.74, + "learning_rate": 0.00021236509528585758, + "loss": 1.2844, + "theoretical_loss": 3.4042562425609244, + "tokens_seen": 2433548288 + }, + { + "epoch": 0.74, + "learning_rate": 0.00021234904714142428, + "loss": 1.3088, + "theoretical_loss": 3.404248951881199, + "tokens_seen": 2433613824 + }, + { + "epoch": 0.74, + "learning_rate": 0.00021233299899699098, + "loss": 1.2009, + "theoretical_loss": 3.4042416614527777, + "tokens_seen": 2433679360 + }, + { + "epoch": 0.74, + "learning_rate": 0.00021231695085255768, + "loss": 1.3783, + "theoretical_loss": 3.404234371275644, + "tokens_seen": 2433744896 + }, + { + "epoch": 0.74, + "learning_rate": 0.00021230090270812438, + "loss": 1.4076, + "theoretical_loss": 3.4042270813497835, + "tokens_seen": 2433810432 + }, + { + "epoch": 0.74, + "learning_rate": 0.00021228485456369108, + "loss": 1.3102, + "theoretical_loss": 3.4042197916751804, + "tokens_seen": 2433875968 + }, + { + "epoch": 0.74, + "learning_rate": 0.00021226880641925778, + "loss": 1.3218, + "theoretical_loss": 3.404212502251819, + "tokens_seen": 2433941504 + }, + { + "epoch": 0.74, + "learning_rate": 0.00021225275827482448, + "loss": 1.2582, + "theoretical_loss": 3.404205213079684, + "tokens_seen": 2434007040 + }, + { + "epoch": 0.74, + "learning_rate": 0.00021223671013039118, + "loss": 1.2479, + "theoretical_loss": 3.40419792415876, + "tokens_seen": 2434072576 + }, + { + "epoch": 0.74, + "learning_rate": 0.00021222066198595786, + "loss": 1.2473, + "theoretical_loss": 3.404190635489032, + "tokens_seen": 2434138112 + }, + { + "epoch": 0.74, + "learning_rate": 0.00021220461384152456, + "loss": 1.3063, + "theoretical_loss": 3.404183347070484, + "tokens_seen": 2434203648 + }, + { + "epoch": 0.74, + "learning_rate": 0.00021218856569709126, + "loss": 1.3181, + "theoretical_loss": 3.404176058903101, + "tokens_seen": 2434269184 + }, + { + "epoch": 0.74, + "learning_rate": 0.00021217251755265796, + "loss": 1.4444, + "theoretical_loss": 3.4041687709868675, + "tokens_seen": 2434334720 + }, + { + "epoch": 0.74, + "learning_rate": 0.00021215646940822466, + "loss": 1.1947, + "theoretical_loss": 3.4041614833217677, + "tokens_seen": 2434400256 + }, + { + "epoch": 0.74, + "learning_rate": 0.00021214042126379136, + "loss": 1.2889, + "theoretical_loss": 3.404154195907787, + "tokens_seen": 2434465792 + }, + { + "epoch": 0.74, + "learning_rate": 0.00021212437311935806, + "loss": 1.3833, + "theoretical_loss": 3.404146908744909, + "tokens_seen": 2434531328 + }, + { + "epoch": 0.74, + "learning_rate": 0.00021210832497492476, + "loss": 1.4197, + "theoretical_loss": 3.404139621833119, + "tokens_seen": 2434596864 + }, + { + "epoch": 0.74, + "objective/train/docs_used": 886508, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.348459243774414, + "objective/train/theoretical_loss": 3.404132335172401, + "objective/train/tokens_used": 2455122400, + "theoretical_loss": 3.404132335172401, + "tokens_seen": 2434662400 + }, + { + "epoch": 0.74, + "learning_rate": 0.00021209227683049146, + "loss": 1.2916, + "theoretical_loss": 3.404132335172401, + "tokens_seen": 2434662400 + }, + { + "epoch": 0.74, + "learning_rate": 0.00021207622868605816, + "loss": 1.3497, + "theoretical_loss": 3.40412504876274, + "tokens_seen": 2434727936 + }, + { + "epoch": 0.74, + "learning_rate": 0.00021206018054162486, + "loss": 1.3221, + "theoretical_loss": 3.4041177626041206, + "tokens_seen": 2434793472 + }, + { + "epoch": 0.74, + "learning_rate": 0.00021204413239719156, + "loss": 1.3664, + "theoretical_loss": 3.4041104766965273, + "tokens_seen": 2434859008 + }, + { + "epoch": 0.74, + "learning_rate": 0.00021202808425275826, + "loss": 1.4023, + "theoretical_loss": 3.404103191039945, + "tokens_seen": 2434924544 + }, + { + "epoch": 0.74, + "learning_rate": 0.00021201203610832496, + "loss": 1.3006, + "theoretical_loss": 3.4040959056343576, + "tokens_seen": 2434990080 + }, + { + "epoch": 0.74, + "learning_rate": 0.00021199598796389172, + "loss": 1.164, + "theoretical_loss": 3.4040886204797505, + "tokens_seen": 2435055616 + }, + { + "epoch": 0.74, + "learning_rate": 0.00021197993981945842, + "loss": 1.4714, + "theoretical_loss": 3.4040813355761075, + "tokens_seen": 2435121152 + }, + { + "epoch": 0.74, + "learning_rate": 0.0002119638916750251, + "loss": 1.4071, + "theoretical_loss": 3.404074050923414, + "tokens_seen": 2435186688 + }, + { + "epoch": 0.74, + "learning_rate": 0.0002119478435305918, + "loss": 1.2243, + "theoretical_loss": 3.404066766521654, + "tokens_seen": 2435252224 + }, + { + "epoch": 0.74, + "learning_rate": 0.0002119317953861585, + "loss": 1.3003, + "theoretical_loss": 3.4040594823708124, + "tokens_seen": 2435317760 + }, + { + "epoch": 0.74, + "learning_rate": 0.0002119157472417252, + "loss": 1.1996, + "theoretical_loss": 3.4040521984708736, + "tokens_seen": 2435383296 + }, + { + "epoch": 0.74, + "learning_rate": 0.0002118996990972919, + "loss": 1.2808, + "theoretical_loss": 3.404044914821822, + "tokens_seen": 2435448832 + }, + { + "epoch": 0.74, + "learning_rate": 0.0002118836509528586, + "loss": 1.3046, + "theoretical_loss": 3.404037631423643, + "tokens_seen": 2435514368 + }, + { + "epoch": 0.74, + "learning_rate": 0.0002118676028084253, + "loss": 1.324, + "theoretical_loss": 3.404030348276321, + "tokens_seen": 2435579904 + }, + { + "epoch": 0.74, + "learning_rate": 0.000211851554663992, + "loss": 1.3804, + "theoretical_loss": 3.40402306537984, + "tokens_seen": 2435645440 + }, + { + "epoch": 0.74, + "learning_rate": 0.0002118355065195587, + "loss": 1.4519, + "theoretical_loss": 3.404015782734185, + "tokens_seen": 2435710976 + }, + { + "epoch": 0.74, + "learning_rate": 0.0002118194583751254, + "loss": 1.2662, + "theoretical_loss": 3.4040085003393403, + "tokens_seen": 2435776512 + }, + { + "epoch": 0.74, + "learning_rate": 0.0002118034102306921, + "loss": 1.2793, + "theoretical_loss": 3.4040012181952912, + "tokens_seen": 2435842048 + }, + { + "epoch": 0.74, + "learning_rate": 0.0002117873620862588, + "loss": 1.3787, + "theoretical_loss": 3.4039939363020215, + "tokens_seen": 2435907584 + }, + { + "epoch": 0.74, + "learning_rate": 0.0002117713139418255, + "loss": 1.297, + "theoretical_loss": 3.4039866546595166, + "tokens_seen": 2435973120 + }, + { + "epoch": 0.74, + "learning_rate": 0.0002117552657973922, + "loss": 1.2338, + "theoretical_loss": 3.4039793732677603, + "tokens_seen": 2436038656 + }, + { + "epoch": 0.74, + "learning_rate": 0.0002117392176529589, + "loss": 1.3261, + "theoretical_loss": 3.4039720921267382, + "tokens_seen": 2436104192 + }, + { + "epoch": 0.74, + "learning_rate": 0.0002117231695085256, + "loss": 1.2938, + "theoretical_loss": 3.4039648112364342, + "tokens_seen": 2436169728 + }, + { + "epoch": 0.74, + "learning_rate": 0.0002117071213640923, + "loss": 1.3958, + "theoretical_loss": 3.4039575305968333, + "tokens_seen": 2436235264 + }, + { + "epoch": 0.74, + "learning_rate": 0.000211691073219659, + "loss": 1.2048, + "theoretical_loss": 3.4039502502079193, + "tokens_seen": 2436300800 + }, + { + "epoch": 0.74, + "learning_rate": 0.0002116750250752257, + "loss": 1.2648, + "theoretical_loss": 3.403942970069678, + "tokens_seen": 2436366336 + }, + { + "epoch": 0.74, + "learning_rate": 0.0002116589769307924, + "loss": 1.2626, + "theoretical_loss": 3.4039356901820934, + "tokens_seen": 2436431872 + }, + { + "epoch": 0.74, + "learning_rate": 0.0002116429287863591, + "loss": 1.2515, + "theoretical_loss": 3.4039284105451504, + "tokens_seen": 2436497408 + }, + { + "epoch": 0.74, + "learning_rate": 0.0002116268806419258, + "loss": 1.2977, + "theoretical_loss": 3.4039211311588327, + "tokens_seen": 2436562944 + }, + { + "epoch": 0.74, + "learning_rate": 0.0002116108324974925, + "loss": 1.5202, + "theoretical_loss": 3.4039138520231265, + "tokens_seen": 2436628480 + }, + { + "epoch": 0.74, + "learning_rate": 0.0002115947843530592, + "loss": 1.28, + "theoretical_loss": 3.4039065731380154, + "tokens_seen": 2436694016 + }, + { + "epoch": 0.74, + "learning_rate": 0.00021157873620862588, + "loss": 1.3, + "theoretical_loss": 3.403899294503484, + "tokens_seen": 2436759552 + }, + { + "epoch": 0.74, + "learning_rate": 0.00021156268806419258, + "loss": 1.3304, + "theoretical_loss": 3.403892016119517, + "tokens_seen": 2436825088 + }, + { + "epoch": 0.74, + "learning_rate": 0.00021154663991975928, + "loss": 1.288, + "theoretical_loss": 3.4038847379860995, + "tokens_seen": 2436890624 + }, + { + "epoch": 0.74, + "learning_rate": 0.00021153059177532598, + "loss": 1.3005, + "theoretical_loss": 3.403877460103216, + "tokens_seen": 2436956160 + }, + { + "epoch": 0.74, + "learning_rate": 0.00021151454363089268, + "loss": 1.3633, + "theoretical_loss": 3.403870182470851, + "tokens_seen": 2437021696 + }, + { + "epoch": 0.74, + "learning_rate": 0.00021149849548645938, + "loss": 1.2813, + "theoretical_loss": 3.403862905088989, + "tokens_seen": 2437087232 + }, + { + "epoch": 0.74, + "learning_rate": 0.00021148244734202609, + "loss": 1.3544, + "theoretical_loss": 3.4038556279576144, + "tokens_seen": 2437152768 + }, + { + "epoch": 0.74, + "learning_rate": 0.00021146639919759279, + "loss": 1.3196, + "theoretical_loss": 3.4038483510767126, + "tokens_seen": 2437218304 + }, + { + "epoch": 0.74, + "learning_rate": 0.00021145035105315949, + "loss": 1.2987, + "theoretical_loss": 3.4038410744462677, + "tokens_seen": 2437283840 + }, + { + "epoch": 0.74, + "learning_rate": 0.0002114343029087262, + "loss": 1.2441, + "theoretical_loss": 3.4038337980662643, + "tokens_seen": 2437349376 + }, + { + "epoch": 0.74, + "learning_rate": 0.0002114182547642929, + "loss": 1.3683, + "theoretical_loss": 3.403826521936687, + "tokens_seen": 2437414912 + }, + { + "epoch": 0.74, + "learning_rate": 0.0002114022066198596, + "loss": 1.2852, + "theoretical_loss": 3.403819246057522, + "tokens_seen": 2437480448 + }, + { + "epoch": 0.74, + "learning_rate": 0.0002113861584754263, + "loss": 1.2819, + "theoretical_loss": 3.4038119704287517, + "tokens_seen": 2437545984 + }, + { + "epoch": 0.74, + "learning_rate": 0.000211370110330993, + "loss": 1.3873, + "theoretical_loss": 3.4038046950503618, + "tokens_seen": 2437611520 + }, + { + "epoch": 0.74, + "learning_rate": 0.0002113540621865597, + "loss": 1.341, + "theoretical_loss": 3.403797419922337, + "tokens_seen": 2437677056 + }, + { + "epoch": 0.74, + "learning_rate": 0.0002113380140421264, + "loss": 1.2994, + "theoretical_loss": 3.4037901450446615, + "tokens_seen": 2437742592 + }, + { + "epoch": 0.74, + "learning_rate": 0.0002113219658976931, + "loss": 1.33, + "theoretical_loss": 3.40378287041732, + "tokens_seen": 2437808128 + }, + { + "epoch": 0.74, + "learning_rate": 0.0002113059177532598, + "loss": 1.1713, + "theoretical_loss": 3.403775596040298, + "tokens_seen": 2437873664 + }, + { + "epoch": 0.74, + "objective/train/docs_used": 887686, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.3012793064117432, + "objective/train/theoretical_loss": 3.4037683219135793, + "objective/train/tokens_used": 2458399200, + "theoretical_loss": 3.4037683219135793, + "tokens_seen": 2437939200 + }, + { + "epoch": 0.74, + "learning_rate": 0.0002112898696088265, + "loss": 1.261, + "theoretical_loss": 3.4037683219135793, + "tokens_seen": 2437939200 + }, + { + "epoch": 0.74, + "learning_rate": 0.0002112738214643932, + "loss": 1.2619, + "theoretical_loss": 3.4037610480371487, + "tokens_seen": 2438004736 + }, + { + "epoch": 0.74, + "learning_rate": 0.0002112577733199599, + "loss": 1.2724, + "theoretical_loss": 3.4037537744109914, + "tokens_seen": 2438070272 + }, + { + "epoch": 0.74, + "learning_rate": 0.0002112417251755266, + "loss": 1.3173, + "theoretical_loss": 3.4037465010350916, + "tokens_seen": 2438135808 + }, + { + "epoch": 0.74, + "learning_rate": 0.0002112256770310933, + "loss": 1.2987, + "theoretical_loss": 3.403739227909434, + "tokens_seen": 2438201344 + }, + { + "epoch": 0.74, + "learning_rate": 0.00021120962888666, + "loss": 1.4404, + "theoretical_loss": 3.4037319550340026, + "tokens_seen": 2438266880 + }, + { + "epoch": 0.74, + "learning_rate": 0.00021119358074222667, + "loss": 1.2892, + "theoretical_loss": 3.4037246824087837, + "tokens_seen": 2438332416 + }, + { + "epoch": 0.74, + "learning_rate": 0.00021117753259779337, + "loss": 1.2331, + "theoretical_loss": 3.40371741003376, + "tokens_seen": 2438397952 + }, + { + "epoch": 0.74, + "learning_rate": 0.00021116148445336007, + "loss": 1.3511, + "theoretical_loss": 3.4037101379089183, + "tokens_seen": 2438463488 + }, + { + "epoch": 0.74, + "learning_rate": 0.00021114543630892677, + "loss": 1.338, + "theoretical_loss": 3.4037028660342417, + "tokens_seen": 2438529024 + }, + { + "epoch": 0.74, + "learning_rate": 0.00021112938816449347, + "loss": 1.2767, + "theoretical_loss": 3.403695594409715, + "tokens_seen": 2438594560 + }, + { + "epoch": 0.74, + "learning_rate": 0.00021111334002006017, + "loss": 1.1907, + "theoretical_loss": 3.403688323035323, + "tokens_seen": 2438660096 + }, + { + "epoch": 0.74, + "learning_rate": 0.00021109729187562687, + "loss": 1.2825, + "theoretical_loss": 3.403681051911051, + "tokens_seen": 2438725632 + }, + { + "epoch": 0.74, + "learning_rate": 0.00021108124373119357, + "loss": 1.3092, + "theoretical_loss": 3.4036737810368836, + "tokens_seen": 2438791168 + }, + { + "epoch": 0.74, + "learning_rate": 0.00021106519558676027, + "loss": 1.2319, + "theoretical_loss": 3.4036665104128048, + "tokens_seen": 2438856704 + }, + { + "epoch": 0.74, + "learning_rate": 0.00021104914744232698, + "loss": 1.3228, + "theoretical_loss": 3.4036592400387997, + "tokens_seen": 2438922240 + }, + { + "epoch": 0.74, + "learning_rate": 0.00021103309929789368, + "loss": 1.325, + "theoretical_loss": 3.4036519699148524, + "tokens_seen": 2438987776 + }, + { + "epoch": 0.74, + "learning_rate": 0.00021101705115346038, + "loss": 1.3043, + "theoretical_loss": 3.403644700040948, + "tokens_seen": 2439053312 + }, + { + "epoch": 0.74, + "learning_rate": 0.00021100100300902708, + "loss": 1.1508, + "theoretical_loss": 3.403637430417072, + "tokens_seen": 2439118848 + }, + { + "epoch": 0.74, + "learning_rate": 0.00021098495486459378, + "loss": 1.2752, + "theoretical_loss": 3.403630161043208, + "tokens_seen": 2439184384 + }, + { + "epoch": 0.74, + "learning_rate": 0.00021096890672016048, + "loss": 1.2524, + "theoretical_loss": 3.4036228919193405, + "tokens_seen": 2439249920 + }, + { + "epoch": 0.74, + "learning_rate": 0.00021095285857572718, + "loss": 1.2441, + "theoretical_loss": 3.4036156230454555, + "tokens_seen": 2439315456 + }, + { + "epoch": 0.74, + "learning_rate": 0.00021093681043129388, + "loss": 1.2284, + "theoretical_loss": 3.4036083544215363, + "tokens_seen": 2439380992 + }, + { + "epoch": 0.74, + "learning_rate": 0.00021092076228686058, + "loss": 1.3993, + "theoretical_loss": 3.4036010860475683, + "tokens_seen": 2439446528 + }, + { + "epoch": 0.74, + "learning_rate": 0.00021090471414242728, + "loss": 1.3001, + "theoretical_loss": 3.4035938179235363, + "tokens_seen": 2439512064 + }, + { + "epoch": 0.74, + "learning_rate": 0.00021088866599799398, + "loss": 1.405, + "theoretical_loss": 3.4035865500494245, + "tokens_seen": 2439577600 + }, + { + "epoch": 0.74, + "learning_rate": 0.00021087261785356068, + "loss": 1.3752, + "theoretical_loss": 3.403579282425218, + "tokens_seen": 2439643136 + }, + { + "epoch": 0.74, + "learning_rate": 0.00021085656970912738, + "loss": 1.3477, + "theoretical_loss": 3.4035720150509015, + "tokens_seen": 2439708672 + }, + { + "epoch": 0.74, + "learning_rate": 0.00021084052156469408, + "loss": 1.2697, + "theoretical_loss": 3.403564747926459, + "tokens_seen": 2439774208 + }, + { + "epoch": 0.74, + "learning_rate": 0.00021082447342026078, + "loss": 1.321, + "theoretical_loss": 3.403557481051876, + "tokens_seen": 2439839744 + }, + { + "epoch": 0.74, + "learning_rate": 0.0002108084252758275, + "loss": 1.4071, + "theoretical_loss": 3.403550214427137, + "tokens_seen": 2439905280 + }, + { + "epoch": 0.74, + "learning_rate": 0.0002107923771313942, + "loss": 1.3776, + "theoretical_loss": 3.4035429480522263, + "tokens_seen": 2439970816 + }, + { + "epoch": 0.74, + "learning_rate": 0.0002107763289869609, + "loss": 1.2383, + "theoretical_loss": 3.4035356819271296, + "tokens_seen": 2440036352 + }, + { + "epoch": 0.74, + "learning_rate": 0.00021076028084252761, + "loss": 1.3348, + "theoretical_loss": 3.4035284160518304, + "tokens_seen": 2440101888 + }, + { + "epoch": 0.74, + "learning_rate": 0.00021074423269809431, + "loss": 1.3473, + "theoretical_loss": 3.403521150426314, + "tokens_seen": 2440167424 + }, + { + "epoch": 0.74, + "learning_rate": 0.00021072818455366101, + "loss": 1.3934, + "theoretical_loss": 3.403513885050565, + "tokens_seen": 2440232960 + }, + { + "epoch": 0.74, + "learning_rate": 0.00021071213640922772, + "loss": 1.3506, + "theoretical_loss": 3.4035066199245683, + "tokens_seen": 2440298496 + }, + { + "epoch": 0.74, + "learning_rate": 0.00021069608826479442, + "loss": 1.2356, + "theoretical_loss": 3.403499355048308, + "tokens_seen": 2440364032 + }, + { + "epoch": 0.74, + "learning_rate": 0.00021068004012036112, + "loss": 1.3519, + "theoretical_loss": 3.40349209042177, + "tokens_seen": 2440429568 + }, + { + "epoch": 0.74, + "learning_rate": 0.00021066399197592782, + "loss": 1.3274, + "theoretical_loss": 3.403484826044938, + "tokens_seen": 2440495104 + }, + { + "epoch": 0.74, + "learning_rate": 0.00021064794383149452, + "loss": 1.3006, + "theoretical_loss": 3.403477561917797, + "tokens_seen": 2440560640 + }, + { + "epoch": 0.74, + "learning_rate": 0.00021063189568706122, + "loss": 1.253, + "theoretical_loss": 3.4034702980403315, + "tokens_seen": 2440626176 + }, + { + "epoch": 0.74, + "learning_rate": 0.00021061584754262792, + "loss": 1.3459, + "theoretical_loss": 3.4034630344125265, + "tokens_seen": 2440691712 + }, + { + "epoch": 0.74, + "learning_rate": 0.00021059979939819462, + "loss": 1.3888, + "theoretical_loss": 3.4034557710343667, + "tokens_seen": 2440757248 + }, + { + "epoch": 0.74, + "learning_rate": 0.00021058375125376132, + "loss": 1.2176, + "theoretical_loss": 3.4034485079058365, + "tokens_seen": 2440822784 + }, + { + "epoch": 0.74, + "learning_rate": 0.000210567703109328, + "loss": 1.3191, + "theoretical_loss": 3.4034412450269214, + "tokens_seen": 2440888320 + }, + { + "epoch": 0.74, + "learning_rate": 0.0002105516549648947, + "loss": 1.2982, + "theoretical_loss": 3.4034339823976048, + "tokens_seen": 2440953856 + }, + { + "epoch": 0.74, + "learning_rate": 0.0002105356068204614, + "loss": 1.2487, + "theoretical_loss": 3.403426720017873, + "tokens_seen": 2441019392 + }, + { + "epoch": 0.74, + "learning_rate": 0.0002105195586760281, + "loss": 1.3532, + "theoretical_loss": 3.4034194578877095, + "tokens_seen": 2441084928 + }, + { + "epoch": 0.74, + "learning_rate": 0.0002105035105315948, + "loss": 1.2418, + "theoretical_loss": 3.4034121960070993, + "tokens_seen": 2441150464 + }, + { + "epoch": 0.74, + "objective/train/docs_used": 888977, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.1388837099075317, + "objective/train/theoretical_loss": 3.4034049343760278, + "objective/train/tokens_used": 2461676000, + "theoretical_loss": 3.4034049343760278, + "tokens_seen": 2441216000 + }, + { + "epoch": 0.74, + "learning_rate": 0.0002104874623871615, + "loss": 1.203, + "theoretical_loss": 3.4034049343760278, + "tokens_seen": 2441216000 + }, + { + "epoch": 0.74, + "learning_rate": 0.0002104714142427282, + "loss": 1.2886, + "theoretical_loss": 3.403397672994479, + "tokens_seen": 2441281536 + }, + { + "epoch": 0.74, + "learning_rate": 0.0002104553660982949, + "loss": 1.3735, + "theoretical_loss": 3.4033904118624374, + "tokens_seen": 2441347072 + }, + { + "epoch": 0.74, + "learning_rate": 0.0002104393179538616, + "loss": 1.3171, + "theoretical_loss": 3.4033831509798884, + "tokens_seen": 2441412608 + }, + { + "epoch": 0.74, + "learning_rate": 0.0002104232698094283, + "loss": 1.248, + "theoretical_loss": 3.4033758903468168, + "tokens_seen": 2441478144 + }, + { + "epoch": 0.74, + "learning_rate": 0.000210407221664995, + "loss": 1.3455, + "theoretical_loss": 3.4033686299632064, + "tokens_seen": 2441543680 + }, + { + "epoch": 0.74, + "learning_rate": 0.0002103911735205617, + "loss": 1.3393, + "theoretical_loss": 3.4033613698290432, + "tokens_seen": 2441609216 + }, + { + "epoch": 0.74, + "learning_rate": 0.0002103751253761284, + "loss": 1.294, + "theoretical_loss": 3.4033541099443108, + "tokens_seen": 2441674752 + }, + { + "epoch": 0.74, + "learning_rate": 0.0002103590772316951, + "loss": 1.1911, + "theoretical_loss": 3.4033468503089948, + "tokens_seen": 2441740288 + }, + { + "epoch": 0.74, + "learning_rate": 0.0002103430290872618, + "loss": 1.2523, + "theoretical_loss": 3.4033395909230792, + "tokens_seen": 2441805824 + }, + { + "epoch": 0.74, + "learning_rate": 0.0002103269809428285, + "loss": 1.2278, + "theoretical_loss": 3.4033323317865496, + "tokens_seen": 2441871360 + }, + { + "epoch": 0.74, + "learning_rate": 0.0002103109327983952, + "loss": 1.3203, + "theoretical_loss": 3.40332507289939, + "tokens_seen": 2441936896 + }, + { + "epoch": 0.74, + "learning_rate": 0.0002102948846539619, + "loss": 1.3693, + "theoretical_loss": 3.403317814261585, + "tokens_seen": 2442002432 + }, + { + "epoch": 0.74, + "learning_rate": 0.0002102788365095286, + "loss": 1.392, + "theoretical_loss": 3.4033105558731203, + "tokens_seen": 2442067968 + }, + { + "epoch": 0.74, + "learning_rate": 0.0002102627883650953, + "loss": 1.2675, + "theoretical_loss": 3.40330329773398, + "tokens_seen": 2442133504 + }, + { + "epoch": 0.74, + "learning_rate": 0.000210246740220662, + "loss": 1.3148, + "theoretical_loss": 3.4032960398441485, + "tokens_seen": 2442199040 + }, + { + "epoch": 0.74, + "learning_rate": 0.0002102306920762287, + "loss": 1.2366, + "theoretical_loss": 3.403288782203611, + "tokens_seen": 2442264576 + }, + { + "epoch": 0.74, + "learning_rate": 0.0002102146439317954, + "loss": 1.3693, + "theoretical_loss": 3.4032815248123525, + "tokens_seen": 2442330112 + }, + { + "epoch": 0.74, + "learning_rate": 0.0002101985957873621, + "loss": 1.3233, + "theoretical_loss": 3.4032742676703576, + "tokens_seen": 2442395648 + }, + { + "epoch": 0.74, + "learning_rate": 0.00021018254764292878, + "loss": 1.362, + "theoretical_loss": 3.4032670107776104, + "tokens_seen": 2442461184 + }, + { + "epoch": 0.74, + "learning_rate": 0.00021016649949849548, + "loss": 1.3364, + "theoretical_loss": 3.4032597541340968, + "tokens_seen": 2442526720 + }, + { + "epoch": 0.74, + "learning_rate": 0.00021015045135406218, + "loss": 1.2877, + "theoretical_loss": 3.4032524977398, + "tokens_seen": 2442592256 + }, + { + "epoch": 0.74, + "learning_rate": 0.00021013440320962888, + "loss": 1.292, + "theoretical_loss": 3.4032452415947065, + "tokens_seen": 2442657792 + }, + { + "epoch": 0.74, + "learning_rate": 0.00021011835506519558, + "loss": 1.3618, + "theoretical_loss": 3.4032379856988, + "tokens_seen": 2442723328 + }, + { + "epoch": 0.74, + "learning_rate": 0.00021010230692076229, + "loss": 1.3109, + "theoretical_loss": 3.4032307300520657, + "tokens_seen": 2442788864 + }, + { + "epoch": 0.74, + "learning_rate": 0.00021008625877632899, + "loss": 1.4219, + "theoretical_loss": 3.403223474654488, + "tokens_seen": 2442854400 + }, + { + "epoch": 0.74, + "learning_rate": 0.0002100702106318957, + "loss": 1.3187, + "theoretical_loss": 3.403216219506051, + "tokens_seen": 2442919936 + }, + { + "epoch": 0.74, + "learning_rate": 0.0002100541624874624, + "loss": 1.2878, + "theoretical_loss": 3.4032089646067414, + "tokens_seen": 2442985472 + }, + { + "epoch": 0.74, + "learning_rate": 0.0002100381143430291, + "loss": 1.3508, + "theoretical_loss": 3.403201709956542, + "tokens_seen": 2443051008 + }, + { + "epoch": 0.74, + "learning_rate": 0.0002100220661985958, + "loss": 1.387, + "theoretical_loss": 3.4031944555554388, + "tokens_seen": 2443116544 + }, + { + "epoch": 0.74, + "learning_rate": 0.0002100060180541625, + "loss": 1.3159, + "theoretical_loss": 3.4031872014034166, + "tokens_seen": 2443182080 + }, + { + "epoch": 0.74, + "learning_rate": 0.0002099899699097292, + "loss": 1.3467, + "theoretical_loss": 3.403179947500459, + "tokens_seen": 2443247616 + }, + { + "epoch": 0.74, + "learning_rate": 0.0002099739217652959, + "loss": 1.3774, + "theoretical_loss": 3.4031726938465514, + "tokens_seen": 2443313152 + }, + { + "epoch": 0.74, + "learning_rate": 0.0002099578736208626, + "loss": 1.264, + "theoretical_loss": 3.403165440441679, + "tokens_seen": 2443378688 + }, + { + "epoch": 0.74, + "learning_rate": 0.0002099418254764293, + "loss": 1.3915, + "theoretical_loss": 3.4031581872858263, + "tokens_seen": 2443444224 + }, + { + "epoch": 0.74, + "learning_rate": 0.000209925777331996, + "loss": 1.293, + "theoretical_loss": 3.403150934378978, + "tokens_seen": 2443509760 + }, + { + "epoch": 0.74, + "learning_rate": 0.0002099097291875627, + "loss": 1.3318, + "theoretical_loss": 3.4031436817211187, + "tokens_seen": 2443575296 + }, + { + "epoch": 0.74, + "learning_rate": 0.0002098936810431294, + "loss": 1.2718, + "theoretical_loss": 3.4031364293122337, + "tokens_seen": 2443640832 + }, + { + "epoch": 0.74, + "learning_rate": 0.0002098776328986961, + "loss": 1.3516, + "theoretical_loss": 3.403129177152307, + "tokens_seen": 2443706368 + }, + { + "epoch": 0.74, + "learning_rate": 0.0002098615847542628, + "loss": 1.3212, + "theoretical_loss": 3.403121925241324, + "tokens_seen": 2443771904 + }, + { + "epoch": 0.74, + "learning_rate": 0.0002098455366098295, + "loss": 1.3728, + "theoretical_loss": 3.4031146735792692, + "tokens_seen": 2443837440 + }, + { + "epoch": 0.74, + "learning_rate": 0.0002098294884653962, + "loss": 1.3402, + "theoretical_loss": 3.4031074221661273, + "tokens_seen": 2443902976 + }, + { + "epoch": 0.74, + "learning_rate": 0.0002098134403209629, + "loss": 1.3207, + "theoretical_loss": 3.4031001710018836, + "tokens_seen": 2443968512 + }, + { + "epoch": 0.74, + "learning_rate": 0.00020979739217652957, + "loss": 1.3385, + "theoretical_loss": 3.4030929200865225, + "tokens_seen": 2444034048 + }, + { + "epoch": 0.74, + "learning_rate": 0.00020978134403209627, + "loss": 1.4254, + "theoretical_loss": 3.4030856694200287, + "tokens_seen": 2444099584 + }, + { + "epoch": 0.74, + "learning_rate": 0.00020976529588766297, + "loss": 1.3294, + "theoretical_loss": 3.4030784190023873, + "tokens_seen": 2444165120 + }, + { + "epoch": 0.74, + "learning_rate": 0.00020974924774322967, + "loss": 1.3248, + "theoretical_loss": 3.4030711688335824, + "tokens_seen": 2444230656 + }, + { + "epoch": 0.74, + "learning_rate": 0.00020973319959879637, + "loss": 1.3629, + "theoretical_loss": 3.4030639189135994, + "tokens_seen": 2444296192 + }, + { + "epoch": 0.74, + "learning_rate": 0.00020971715145436307, + "loss": 1.3647, + "theoretical_loss": 3.4030566692424236, + "tokens_seen": 2444361728 + }, + { + "epoch": 0.74, + "learning_rate": 0.00020970110330992977, + "loss": 1.2515, + "theoretical_loss": 3.4030494198200385, + "tokens_seen": 2444427264 + }, + { + "epoch": 0.74, + "objective/train/docs_used": 890203, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.4732046127319336, + "objective/train/theoretical_loss": 3.4030421706464296, + "objective/train/tokens_used": 2464952800, + "theoretical_loss": 3.4030421706464296, + "tokens_seen": 2444492800 + }, + { + "epoch": 0.74, + "learning_rate": 0.00020968505516549647, + "loss": 1.4211, + "theoretical_loss": 3.4030421706464296, + "tokens_seen": 2444492800 + }, + { + "epoch": 0.74, + "learning_rate": 0.00020966900702106318, + "loss": 1.3403, + "theoretical_loss": 3.4030349217215816, + "tokens_seen": 2444558336 + }, + { + "epoch": 0.74, + "learning_rate": 0.00020965295887662988, + "loss": 1.4068, + "theoretical_loss": 3.4030276730454796, + "tokens_seen": 2444623872 + }, + { + "epoch": 0.74, + "learning_rate": 0.00020963691073219658, + "loss": 1.3542, + "theoretical_loss": 3.403020424618108, + "tokens_seen": 2444689408 + }, + { + "epoch": 0.74, + "learning_rate": 0.00020962086258776328, + "loss": 1.3042, + "theoretical_loss": 3.403013176439452, + "tokens_seen": 2444754944 + }, + { + "epoch": 0.74, + "learning_rate": 0.00020960481444333003, + "loss": 1.3019, + "theoretical_loss": 3.4030059285094953, + "tokens_seen": 2444820480 + }, + { + "epoch": 0.74, + "learning_rate": 0.00020958876629889673, + "loss": 1.2486, + "theoretical_loss": 3.402998680828224, + "tokens_seen": 2444886016 + }, + { + "epoch": 0.74, + "learning_rate": 0.00020957271815446343, + "loss": 1.4569, + "theoretical_loss": 3.4029914333956226, + "tokens_seen": 2444951552 + }, + { + "epoch": 0.74, + "learning_rate": 0.00020955667001003013, + "loss": 1.353, + "theoretical_loss": 3.402984186211676, + "tokens_seen": 2445017088 + }, + { + "epoch": 0.74, + "learning_rate": 0.0002095406218655968, + "loss": 1.3815, + "theoretical_loss": 3.4029769392763685, + "tokens_seen": 2445082624 + }, + { + "epoch": 0.74, + "learning_rate": 0.0002095245737211635, + "loss": 1.3759, + "theoretical_loss": 3.402969692589685, + "tokens_seen": 2445148160 + }, + { + "epoch": 0.74, + "learning_rate": 0.0002095085255767302, + "loss": 1.3326, + "theoretical_loss": 3.4029624461516104, + "tokens_seen": 2445213696 + }, + { + "epoch": 0.74, + "learning_rate": 0.0002094924774322969, + "loss": 1.3144, + "theoretical_loss": 3.4029551999621295, + "tokens_seen": 2445279232 + }, + { + "epoch": 0.74, + "learning_rate": 0.0002094764292878636, + "loss": 1.2467, + "theoretical_loss": 3.4029479540212275, + "tokens_seen": 2445344768 + }, + { + "epoch": 0.74, + "learning_rate": 0.0002094603811434303, + "loss": 1.3632, + "theoretical_loss": 3.4029407083288885, + "tokens_seen": 2445410304 + }, + { + "epoch": 0.74, + "learning_rate": 0.000209444332998997, + "loss": 1.3564, + "theoretical_loss": 3.402933462885098, + "tokens_seen": 2445475840 + }, + { + "epoch": 0.74, + "learning_rate": 0.0002094282848545637, + "loss": 1.3677, + "theoretical_loss": 3.4029262176898403, + "tokens_seen": 2445541376 + }, + { + "epoch": 0.74, + "learning_rate": 0.0002094122367101304, + "loss": 1.4138, + "theoretical_loss": 3.4029189727431004, + "tokens_seen": 2445606912 + }, + { + "epoch": 0.74, + "learning_rate": 0.0002093961885656971, + "loss": 1.2434, + "theoretical_loss": 3.402911728044863, + "tokens_seen": 2445672448 + }, + { + "epoch": 0.74, + "learning_rate": 0.00020938014042126381, + "loss": 1.2343, + "theoretical_loss": 3.4029044835951137, + "tokens_seen": 2445737984 + }, + { + "epoch": 0.74, + "learning_rate": 0.00020936409227683051, + "loss": 1.2926, + "theoretical_loss": 3.402897239393836, + "tokens_seen": 2445803520 + }, + { + "epoch": 0.74, + "learning_rate": 0.00020934804413239722, + "loss": 1.276, + "theoretical_loss": 3.4028899954410154, + "tokens_seen": 2445869056 + }, + { + "epoch": 0.74, + "learning_rate": 0.00020933199598796392, + "loss": 1.2799, + "theoretical_loss": 3.4028827517366373, + "tokens_seen": 2445934592 + }, + { + "epoch": 0.74, + "learning_rate": 0.00020931594784353062, + "loss": 1.3754, + "theoretical_loss": 3.4028755082806854, + "tokens_seen": 2446000128 + }, + { + "epoch": 0.74, + "learning_rate": 0.00020929989969909732, + "loss": 1.1981, + "theoretical_loss": 3.4028682650731454, + "tokens_seen": 2446065664 + }, + { + "epoch": 0.74, + "learning_rate": 0.00020928385155466402, + "loss": 1.3229, + "theoretical_loss": 3.4028610221140014, + "tokens_seen": 2446131200 + }, + { + "epoch": 0.74, + "learning_rate": 0.00020926780341023072, + "loss": 1.4523, + "theoretical_loss": 3.402853779403239, + "tokens_seen": 2446196736 + }, + { + "epoch": 0.74, + "learning_rate": 0.00020925175526579742, + "loss": 1.2361, + "theoretical_loss": 3.4028465369408423, + "tokens_seen": 2446262272 + }, + { + "epoch": 0.74, + "learning_rate": 0.00020923570712136412, + "loss": 1.2764, + "theoretical_loss": 3.402839294726797, + "tokens_seen": 2446327808 + }, + { + "epoch": 0.74, + "learning_rate": 0.00020921965897693082, + "loss": 1.334, + "theoretical_loss": 3.4028320527610867, + "tokens_seen": 2446393344 + }, + { + "epoch": 0.74, + "learning_rate": 0.00020920361083249752, + "loss": 1.2801, + "theoretical_loss": 3.4028248110436974, + "tokens_seen": 2446458880 + }, + { + "epoch": 0.74, + "learning_rate": 0.00020918756268806422, + "loss": 1.2958, + "theoretical_loss": 3.402817569574613, + "tokens_seen": 2446524416 + }, + { + "epoch": 0.74, + "learning_rate": 0.00020917151454363092, + "loss": 1.3673, + "theoretical_loss": 3.4028103283538194, + "tokens_seen": 2446589952 + }, + { + "epoch": 0.74, + "learning_rate": 0.0002091554663991976, + "loss": 1.3391, + "theoretical_loss": 3.4028030873813004, + "tokens_seen": 2446655488 + }, + { + "epoch": 0.74, + "learning_rate": 0.0002091394182547643, + "loss": 1.3563, + "theoretical_loss": 3.4027958466570416, + "tokens_seen": 2446721024 + }, + { + "epoch": 0.74, + "learning_rate": 0.000209123370110331, + "loss": 1.3019, + "theoretical_loss": 3.4027886061810273, + "tokens_seen": 2446786560 + }, + { + "epoch": 0.74, + "learning_rate": 0.0002091073219658977, + "loss": 1.2967, + "theoretical_loss": 3.402781365953243, + "tokens_seen": 2446852096 + }, + { + "epoch": 0.74, + "learning_rate": 0.0002090912738214644, + "loss": 1.4179, + "theoretical_loss": 3.4027741259736723, + "tokens_seen": 2446917632 + }, + { + "epoch": 0.74, + "learning_rate": 0.0002090752256770311, + "loss": 1.3268, + "theoretical_loss": 3.4027668862423015, + "tokens_seen": 2446983168 + }, + { + "epoch": 0.74, + "learning_rate": 0.0002090591775325978, + "loss": 1.2515, + "theoretical_loss": 3.4027596467591144, + "tokens_seen": 2447048704 + }, + { + "epoch": 0.74, + "learning_rate": 0.0002090431293881645, + "loss": 1.4009, + "theoretical_loss": 3.4027524075240967, + "tokens_seen": 2447114240 + }, + { + "epoch": 0.74, + "learning_rate": 0.0002090270812437312, + "loss": 1.34, + "theoretical_loss": 3.402745168537232, + "tokens_seen": 2447179776 + }, + { + "epoch": 0.74, + "learning_rate": 0.0002090110330992979, + "loss": 1.3113, + "theoretical_loss": 3.4027379297985068, + "tokens_seen": 2447245312 + }, + { + "epoch": 0.74, + "learning_rate": 0.0002089949849548646, + "loss": 1.3241, + "theoretical_loss": 3.4027306913079043, + "tokens_seen": 2447310848 + }, + { + "epoch": 0.74, + "learning_rate": 0.0002089789368104313, + "loss": 1.2567, + "theoretical_loss": 3.4027234530654105, + "tokens_seen": 2447376384 + }, + { + "epoch": 0.74, + "learning_rate": 0.000208962888665998, + "loss": 1.3556, + "theoretical_loss": 3.4027162150710097, + "tokens_seen": 2447441920 + }, + { + "epoch": 0.74, + "learning_rate": 0.0002089468405215647, + "loss": 1.2648, + "theoretical_loss": 3.4027089773246875, + "tokens_seen": 2447507456 + }, + { + "epoch": 0.74, + "learning_rate": 0.0002089307923771314, + "loss": 1.2249, + "theoretical_loss": 3.4027017398264277, + "tokens_seen": 2447572992 + }, + { + "epoch": 0.74, + "learning_rate": 0.0002089147442326981, + "loss": 1.2794, + "theoretical_loss": 3.4026945025762156, + "tokens_seen": 2447638528 + }, + { + "epoch": 0.74, + "learning_rate": 0.0002088986960882648, + "loss": 1.2455, + "theoretical_loss": 3.4026872655740363, + "tokens_seen": 2447704064 + }, + { + "epoch": 0.74, + "objective/train/docs_used": 890833, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.1800854206085205, + "objective/train/theoretical_loss": 3.402680028819874, + "objective/train/tokens_used": 2468229600, + "theoretical_loss": 3.402680028819874, + "tokens_seen": 2447769600 + }, + { + "epoch": 0.74, + "learning_rate": 0.0002088826479438315, + "loss": 1.1892, + "theoretical_loss": 3.402680028819874, + "tokens_seen": 2447769600 + }, + { + "epoch": 0.74, + "learning_rate": 0.0002088665997993982, + "loss": 1.1033, + "theoretical_loss": 3.402672792313715, + "tokens_seen": 2447835136 + }, + { + "epoch": 0.74, + "learning_rate": 0.0002088505516549649, + "loss": 1.2456, + "theoretical_loss": 3.4026655560555423, + "tokens_seen": 2447900672 + }, + { + "epoch": 0.74, + "learning_rate": 0.0002088345035105316, + "loss": 1.2753, + "theoretical_loss": 3.402658320045342, + "tokens_seen": 2447966208 + }, + { + "epoch": 0.74, + "learning_rate": 0.0002088184553660983, + "loss": 1.2798, + "theoretical_loss": 3.4026510842830984, + "tokens_seen": 2448031744 + }, + { + "epoch": 0.74, + "learning_rate": 0.000208802407221665, + "loss": 1.357, + "theoretical_loss": 3.402643848768797, + "tokens_seen": 2448097280 + }, + { + "epoch": 0.74, + "learning_rate": 0.0002087863590772317, + "loss": 1.3018, + "theoretical_loss": 3.4026366135024215, + "tokens_seen": 2448162816 + }, + { + "epoch": 0.74, + "learning_rate": 0.00020877031093279838, + "loss": 1.3244, + "theoretical_loss": 3.402629378483958, + "tokens_seen": 2448228352 + }, + { + "epoch": 0.74, + "learning_rate": 0.00020875426278836508, + "loss": 1.2886, + "theoretical_loss": 3.402622143713391, + "tokens_seen": 2448293888 + }, + { + "epoch": 0.74, + "learning_rate": 0.00020873821464393179, + "loss": 1.3678, + "theoretical_loss": 3.4026149091907048, + "tokens_seen": 2448359424 + }, + { + "epoch": 0.74, + "learning_rate": 0.00020872216649949849, + "loss": 1.2754, + "theoretical_loss": 3.402607674915885, + "tokens_seen": 2448424960 + }, + { + "epoch": 0.74, + "learning_rate": 0.00020870611835506519, + "loss": 1.2207, + "theoretical_loss": 3.402600440888916, + "tokens_seen": 2448490496 + }, + { + "epoch": 0.74, + "learning_rate": 0.0002086900702106319, + "loss": 1.3246, + "theoretical_loss": 3.4025932071097835, + "tokens_seen": 2448556032 + }, + { + "epoch": 0.74, + "learning_rate": 0.0002086740220661986, + "loss": 1.3398, + "theoretical_loss": 3.4025859735784714, + "tokens_seen": 2448621568 + }, + { + "epoch": 0.74, + "learning_rate": 0.0002086579739217653, + "loss": 1.2924, + "theoretical_loss": 3.4025787402949645, + "tokens_seen": 2448687104 + }, + { + "epoch": 0.74, + "learning_rate": 0.000208641925777332, + "loss": 1.2784, + "theoretical_loss": 3.4025715072592484, + "tokens_seen": 2448752640 + }, + { + "epoch": 0.74, + "learning_rate": 0.0002086258776328987, + "loss": 1.3549, + "theoretical_loss": 3.402564274471308, + "tokens_seen": 2448818176 + }, + { + "epoch": 0.74, + "learning_rate": 0.0002086098294884654, + "loss": 1.2563, + "theoretical_loss": 3.4025570419311273, + "tokens_seen": 2448883712 + }, + { + "epoch": 0.74, + "learning_rate": 0.0002085937813440321, + "loss": 1.3831, + "theoretical_loss": 3.402549809638692, + "tokens_seen": 2448949248 + }, + { + "epoch": 0.74, + "learning_rate": 0.0002085777331995988, + "loss": 1.4269, + "theoretical_loss": 3.402542577593987, + "tokens_seen": 2449014784 + }, + { + "epoch": 0.74, + "learning_rate": 0.0002085616850551655, + "loss": 1.4859, + "theoretical_loss": 3.4025353457969967, + "tokens_seen": 2449080320 + }, + { + "epoch": 0.74, + "learning_rate": 0.0002085456369107322, + "loss": 1.2155, + "theoretical_loss": 3.4025281142477057, + "tokens_seen": 2449145856 + }, + { + "epoch": 0.74, + "learning_rate": 0.0002085295887662989, + "loss": 1.2906, + "theoretical_loss": 3.4025208829461002, + "tokens_seen": 2449211392 + }, + { + "epoch": 0.74, + "learning_rate": 0.0002085135406218656, + "loss": 1.2563, + "theoretical_loss": 3.4025136518921637, + "tokens_seen": 2449276928 + }, + { + "epoch": 0.74, + "learning_rate": 0.0002084974924774323, + "loss": 1.3436, + "theoretical_loss": 3.402506421085882, + "tokens_seen": 2449342464 + }, + { + "epoch": 0.74, + "learning_rate": 0.000208481444332999, + "loss": 1.2405, + "theoretical_loss": 3.40249919052724, + "tokens_seen": 2449408000 + }, + { + "epoch": 0.74, + "learning_rate": 0.0002084653961885657, + "loss": 1.2317, + "theoretical_loss": 3.4024919602162216, + "tokens_seen": 2449473536 + }, + { + "epoch": 0.74, + "learning_rate": 0.0002084493480441324, + "loss": 1.2684, + "theoretical_loss": 3.4024847301528123, + "tokens_seen": 2449539072 + }, + { + "epoch": 0.74, + "learning_rate": 0.0002084332998996991, + "loss": 1.2276, + "theoretical_loss": 3.4024775003369974, + "tokens_seen": 2449604608 + }, + { + "epoch": 0.74, + "learning_rate": 0.0002084172517552658, + "loss": 1.2966, + "theoretical_loss": 3.4024702707687613, + "tokens_seen": 2449670144 + }, + { + "epoch": 0.74, + "learning_rate": 0.00020840120361083253, + "loss": 1.2774, + "theoretical_loss": 3.4024630414480894, + "tokens_seen": 2449735680 + }, + { + "epoch": 0.74, + "learning_rate": 0.00020838515546639923, + "loss": 1.2464, + "theoretical_loss": 3.4024558123749657, + "tokens_seen": 2449801216 + }, + { + "epoch": 0.74, + "learning_rate": 0.00020836910732196593, + "loss": 1.2846, + "theoretical_loss": 3.402448583549376, + "tokens_seen": 2449866752 + }, + { + "epoch": 0.74, + "learning_rate": 0.00020835305917753263, + "loss": 1.2103, + "theoretical_loss": 3.4024413549713044, + "tokens_seen": 2449932288 + }, + { + "epoch": 0.74, + "learning_rate": 0.00020833701103309933, + "loss": 1.1136, + "theoretical_loss": 3.402434126640737, + "tokens_seen": 2449997824 + }, + { + "epoch": 0.74, + "learning_rate": 0.00020832096288866603, + "loss": 1.2701, + "theoretical_loss": 3.4024268985576573, + "tokens_seen": 2450063360 + }, + { + "epoch": 0.74, + "learning_rate": 0.00020830491474423273, + "loss": 1.2885, + "theoretical_loss": 3.4024196707220513, + "tokens_seen": 2450128896 + }, + { + "epoch": 0.74, + "learning_rate": 0.00020828886659979943, + "loss": 1.4428, + "theoretical_loss": 3.4024124431339033, + "tokens_seen": 2450194432 + }, + { + "epoch": 0.74, + "learning_rate": 0.00020827281845536613, + "loss": 1.2617, + "theoretical_loss": 3.4024052157931983, + "tokens_seen": 2450259968 + }, + { + "epoch": 0.74, + "learning_rate": 0.00020825677031093283, + "loss": 1.2474, + "theoretical_loss": 3.4023979886999216, + "tokens_seen": 2450325504 + }, + { + "epoch": 0.74, + "learning_rate": 0.00020824072216649953, + "loss": 1.0998, + "theoretical_loss": 3.402390761854057, + "tokens_seen": 2450391040 + }, + { + "epoch": 0.74, + "learning_rate": 0.00020822467402206623, + "loss": 1.4006, + "theoretical_loss": 3.402383535255591, + "tokens_seen": 2450456576 + }, + { + "epoch": 0.74, + "learning_rate": 0.00020820862587763293, + "loss": 1.3446, + "theoretical_loss": 3.4023763089045076, + "tokens_seen": 2450522112 + }, + { + "epoch": 0.74, + "learning_rate": 0.00020819257773319963, + "loss": 1.28, + "theoretical_loss": 3.4023690828007918, + "tokens_seen": 2450587648 + }, + { + "epoch": 0.74, + "learning_rate": 0.00020817652958876633, + "loss": 1.2773, + "theoretical_loss": 3.4023618569444283, + "tokens_seen": 2450653184 + }, + { + "epoch": 0.74, + "learning_rate": 0.00020816048144433304, + "loss": 1.2122, + "theoretical_loss": 3.4023546313354025, + "tokens_seen": 2450718720 + }, + { + "epoch": 0.74, + "learning_rate": 0.0002081444332998997, + "loss": 1.2563, + "theoretical_loss": 3.4023474059736993, + "tokens_seen": 2450784256 + }, + { + "epoch": 0.74, + "learning_rate": 0.0002081283851554664, + "loss": 1.3058, + "theoretical_loss": 3.4023401808593032, + "tokens_seen": 2450849792 + }, + { + "epoch": 0.74, + "learning_rate": 0.0002081123370110331, + "loss": 1.201, + "theoretical_loss": 3.4023329559921995, + "tokens_seen": 2450915328 + }, + { + "epoch": 0.74, + "learning_rate": 0.0002080962888665998, + "loss": 1.2729, + "theoretical_loss": 3.4023257313723727, + "tokens_seen": 2450980864 + }, + { + "epoch": 0.74, + "objective/train/docs_used": 891914, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.4016845226287842, + "objective/train/theoretical_loss": 3.402318506999808, + "objective/train/tokens_used": 2471506400, + "theoretical_loss": 3.402318506999808, + "tokens_seen": 2451046400 + }, + { + "epoch": 0.74, + "learning_rate": 0.0002080802407221665, + "loss": 1.3275, + "theoretical_loss": 3.402318506999808, + "tokens_seen": 2451046400 + }, + { + "epoch": 0.74, + "learning_rate": 0.0002080641925777332, + "loss": 1.1837, + "theoretical_loss": 3.402311282874491, + "tokens_seen": 2451111936 + }, + { + "epoch": 0.74, + "learning_rate": 0.0002080481444332999, + "loss": 1.3486, + "theoretical_loss": 3.402304058996405, + "tokens_seen": 2451177472 + }, + { + "epoch": 0.74, + "learning_rate": 0.0002080320962888666, + "loss": 1.199, + "theoretical_loss": 3.4022968353655365, + "tokens_seen": 2451243008 + }, + { + "epoch": 0.74, + "learning_rate": 0.00020801604814443331, + "loss": 1.1756, + "theoretical_loss": 3.4022896119818697, + "tokens_seen": 2451308544 + }, + { + "epoch": 0.74, + "learning_rate": 0.00020800000000000001, + "loss": 1.232, + "theoretical_loss": 3.4022823888453892, + "tokens_seen": 2451374080 + }, + { + "epoch": 0.74, + "learning_rate": 0.00020798395185556671, + "loss": 1.4353, + "theoretical_loss": 3.402275165956081, + "tokens_seen": 2451439616 + }, + { + "epoch": 0.74, + "learning_rate": 0.00020796790371113342, + "loss": 1.3097, + "theoretical_loss": 3.4022679433139293, + "tokens_seen": 2451505152 + }, + { + "epoch": 0.74, + "learning_rate": 0.00020795185556670012, + "loss": 1.2521, + "theoretical_loss": 3.402260720918919, + "tokens_seen": 2451570688 + }, + { + "epoch": 0.74, + "learning_rate": 0.00020793580742226682, + "loss": 1.304, + "theoretical_loss": 3.402253498771035, + "tokens_seen": 2451636224 + }, + { + "epoch": 0.74, + "learning_rate": 0.00020791975927783352, + "loss": 1.3213, + "theoretical_loss": 3.4022462768702626, + "tokens_seen": 2451701760 + }, + { + "epoch": 0.74, + "learning_rate": 0.00020790371113340022, + "loss": 1.1735, + "theoretical_loss": 3.4022390552165866, + "tokens_seen": 2451767296 + }, + { + "epoch": 0.74, + "learning_rate": 0.00020788766298896692, + "loss": 1.2621, + "theoretical_loss": 3.4022318338099917, + "tokens_seen": 2451832832 + }, + { + "epoch": 0.74, + "learning_rate": 0.00020787161484453362, + "loss": 1.304, + "theoretical_loss": 3.402224612650463, + "tokens_seen": 2451898368 + }, + { + "epoch": 0.74, + "learning_rate": 0.00020785556670010032, + "loss": 1.3764, + "theoretical_loss": 3.402217391737986, + "tokens_seen": 2451963904 + }, + { + "epoch": 0.74, + "learning_rate": 0.00020783951855566702, + "loss": 1.2432, + "theoretical_loss": 3.4022101710725448, + "tokens_seen": 2452029440 + }, + { + "epoch": 0.74, + "learning_rate": 0.00020782347041123372, + "loss": 1.2465, + "theoretical_loss": 3.4022029506541247, + "tokens_seen": 2452094976 + }, + { + "epoch": 0.74, + "learning_rate": 0.00020780742226680042, + "loss": 1.2417, + "theoretical_loss": 3.4021957304827106, + "tokens_seen": 2452160512 + }, + { + "epoch": 0.74, + "learning_rate": 0.00020779137412236712, + "loss": 1.2546, + "theoretical_loss": 3.4021885105582874, + "tokens_seen": 2452226048 + }, + { + "epoch": 0.74, + "learning_rate": 0.00020777532597793382, + "loss": 1.2276, + "theoretical_loss": 3.4021812908808404, + "tokens_seen": 2452291584 + }, + { + "epoch": 0.74, + "learning_rate": 0.0002077592778335005, + "loss": 1.288, + "theoretical_loss": 3.402174071450354, + "tokens_seen": 2452357120 + }, + { + "epoch": 0.74, + "learning_rate": 0.0002077432296890672, + "loss": 1.3494, + "theoretical_loss": 3.4021668522668134, + "tokens_seen": 2452422656 + }, + { + "epoch": 0.74, + "learning_rate": 0.0002077271815446339, + "loss": 1.2825, + "theoretical_loss": 3.4021596333302035, + "tokens_seen": 2452488192 + }, + { + "epoch": 0.74, + "learning_rate": 0.0002077111334002006, + "loss": 1.2194, + "theoretical_loss": 3.40215241464051, + "tokens_seen": 2452553728 + }, + { + "epoch": 0.74, + "learning_rate": 0.0002076950852557673, + "loss": 1.258, + "theoretical_loss": 3.4021451961977167, + "tokens_seen": 2452619264 + }, + { + "epoch": 0.74, + "learning_rate": 0.000207679037111334, + "loss": 1.2713, + "theoretical_loss": 3.402137978001809, + "tokens_seen": 2452684800 + }, + { + "epoch": 0.74, + "learning_rate": 0.0002076629889669007, + "loss": 1.2666, + "theoretical_loss": 3.402130760052772, + "tokens_seen": 2452750336 + }, + { + "epoch": 0.74, + "learning_rate": 0.0002076469408224674, + "loss": 1.2852, + "theoretical_loss": 3.4021235423505907, + "tokens_seen": 2452815872 + }, + { + "epoch": 0.74, + "learning_rate": 0.0002076308926780341, + "loss": 1.2327, + "theoretical_loss": 3.4021163248952497, + "tokens_seen": 2452881408 + }, + { + "epoch": 0.74, + "learning_rate": 0.0002076148445336008, + "loss": 1.326, + "theoretical_loss": 3.4021091076867345, + "tokens_seen": 2452946944 + }, + { + "epoch": 0.74, + "learning_rate": 0.0002075987963891675, + "loss": 1.1478, + "theoretical_loss": 3.4021018907250298, + "tokens_seen": 2453012480 + }, + { + "epoch": 0.74, + "learning_rate": 0.0002075827482447342, + "loss": 1.2211, + "theoretical_loss": 3.40209467401012, + "tokens_seen": 2453078016 + }, + { + "epoch": 0.74, + "learning_rate": 0.0002075667001003009, + "loss": 1.4449, + "theoretical_loss": 3.402087457541991, + "tokens_seen": 2453143552 + }, + { + "epoch": 0.74, + "learning_rate": 0.0002075506519558676, + "loss": 1.2054, + "theoretical_loss": 3.4020802413206273, + "tokens_seen": 2453209088 + }, + { + "epoch": 0.74, + "learning_rate": 0.0002075346038114343, + "loss": 1.36, + "theoretical_loss": 3.402073025346014, + "tokens_seen": 2453274624 + }, + { + "epoch": 0.74, + "learning_rate": 0.000207518555667001, + "loss": 1.3083, + "theoretical_loss": 3.402065809618136, + "tokens_seen": 2453340160 + }, + { + "epoch": 0.74, + "learning_rate": 0.0002075025075225677, + "loss": 1.2517, + "theoretical_loss": 3.402058594136978, + "tokens_seen": 2453405696 + }, + { + "epoch": 0.74, + "learning_rate": 0.0002074864593781344, + "loss": 1.3558, + "theoretical_loss": 3.4020513789025255, + "tokens_seen": 2453471232 + }, + { + "epoch": 0.74, + "learning_rate": 0.0002074704112337011, + "loss": 1.2009, + "theoretical_loss": 3.402044163914763, + "tokens_seen": 2453536768 + }, + { + "epoch": 0.74, + "learning_rate": 0.0002074543630892678, + "loss": 1.2331, + "theoretical_loss": 3.4020369491736764, + "tokens_seen": 2453602304 + }, + { + "epoch": 0.74, + "learning_rate": 0.0002074383149448345, + "loss": 1.2112, + "theoretical_loss": 3.4020297346792496, + "tokens_seen": 2453667840 + }, + { + "epoch": 0.74, + "learning_rate": 0.0002074222668004012, + "loss": 1.2859, + "theoretical_loss": 3.4020225204314674, + "tokens_seen": 2453733376 + }, + { + "epoch": 0.74, + "learning_rate": 0.0002074062186559679, + "loss": 1.3173, + "theoretical_loss": 3.4020153064303162, + "tokens_seen": 2453798912 + }, + { + "epoch": 0.74, + "learning_rate": 0.0002073901705115346, + "loss": 1.3198, + "theoretical_loss": 3.4020080926757794, + "tokens_seen": 2453864448 + }, + { + "epoch": 0.74, + "learning_rate": 0.00020737412236710128, + "loss": 1.2799, + "theoretical_loss": 3.402000879167843, + "tokens_seen": 2453929984 + }, + { + "epoch": 0.74, + "learning_rate": 0.00020735807422266799, + "loss": 1.1502, + "theoretical_loss": 3.401993665906492, + "tokens_seen": 2453995520 + }, + { + "epoch": 0.74, + "learning_rate": 0.00020734202607823469, + "loss": 1.0738, + "theoretical_loss": 3.4019864528917108, + "tokens_seen": 2454061056 + }, + { + "epoch": 0.74, + "learning_rate": 0.0002073259779338014, + "loss": 1.2219, + "theoretical_loss": 3.401979240123485, + "tokens_seen": 2454126592 + }, + { + "epoch": 0.74, + "learning_rate": 0.0002073099297893681, + "loss": 1.3038, + "theoretical_loss": 3.401972027601799, + "tokens_seen": 2454192128 + }, + { + "epoch": 0.74, + "learning_rate": 0.0002072938816449348, + "loss": 1.3507, + "theoretical_loss": 3.4019648153266377, + "tokens_seen": 2454257664 + }, + { + "epoch": 0.74, + "objective/train/docs_used": 893100, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.310814380645752, + "objective/train/theoretical_loss": 3.401957603297987, + "objective/train/tokens_used": 2474783200, + "theoretical_loss": 3.401957603297987, + "tokens_seen": 2454323200 + }, + { + "epoch": 0.74, + "learning_rate": 0.0002072778335005015, + "loss": 1.2398, + "theoretical_loss": 3.401957603297987, + "tokens_seen": 2454323200 + }, + { + "epoch": 0.74, + "learning_rate": 0.0002072617853560682, + "loss": 1.2425, + "theoretical_loss": 3.401950391515831, + "tokens_seen": 2454388736 + }, + { + "epoch": 0.74, + "learning_rate": 0.0002072457372116349, + "loss": 1.2609, + "theoretical_loss": 3.4019431799801554, + "tokens_seen": 2454454272 + }, + { + "epoch": 0.74, + "learning_rate": 0.0002072296890672016, + "loss": 1.3092, + "theoretical_loss": 3.4019359686909443, + "tokens_seen": 2454519808 + }, + { + "epoch": 0.74, + "learning_rate": 0.0002072136409227683, + "loss": 1.154, + "theoretical_loss": 3.4019287576481836, + "tokens_seen": 2454585344 + }, + { + "epoch": 0.74, + "learning_rate": 0.00020719759277833505, + "loss": 1.2213, + "theoretical_loss": 3.4019215468518578, + "tokens_seen": 2454650880 + }, + { + "epoch": 0.74, + "learning_rate": 0.00020718154463390175, + "loss": 1.2246, + "theoretical_loss": 3.4019143363019517, + "tokens_seen": 2454716416 + }, + { + "epoch": 0.74, + "learning_rate": 0.00020716549648946845, + "loss": 1.1046, + "theoretical_loss": 3.401907125998451, + "tokens_seen": 2454781952 + }, + { + "epoch": 0.74, + "learning_rate": 0.00020714944834503515, + "loss": 1.3142, + "theoretical_loss": 3.4018999159413403, + "tokens_seen": 2454847488 + }, + { + "epoch": 0.74, + "learning_rate": 0.00020713340020060182, + "loss": 1.2625, + "theoretical_loss": 3.4018927061306043, + "tokens_seen": 2454913024 + }, + { + "epoch": 0.74, + "learning_rate": 0.00020711735205616852, + "loss": 1.2667, + "theoretical_loss": 3.4018854965662286, + "tokens_seen": 2454978560 + }, + { + "epoch": 0.74, + "learning_rate": 0.00020710130391173522, + "loss": 1.3182, + "theoretical_loss": 3.401878287248198, + "tokens_seen": 2455044096 + }, + { + "epoch": 0.74, + "learning_rate": 0.00020708525576730192, + "loss": 1.1429, + "theoretical_loss": 3.401871078176497, + "tokens_seen": 2455109632 + }, + { + "epoch": 0.74, + "learning_rate": 0.00020706920762286862, + "loss": 1.3981, + "theoretical_loss": 3.4018638693511116, + "tokens_seen": 2455175168 + }, + { + "epoch": 0.74, + "learning_rate": 0.00020705315947843532, + "loss": 1.3852, + "theoretical_loss": 3.401856660772026, + "tokens_seen": 2455240704 + }, + { + "epoch": 0.74, + "learning_rate": 0.00020703711133400203, + "loss": 1.3391, + "theoretical_loss": 3.4018494524392255, + "tokens_seen": 2455306240 + }, + { + "epoch": 0.74, + "learning_rate": 0.00020702106318956873, + "loss": 1.3846, + "theoretical_loss": 3.401842244352695, + "tokens_seen": 2455371776 + }, + { + "epoch": 0.74, + "learning_rate": 0.00020700501504513543, + "loss": 1.2529, + "theoretical_loss": 3.40183503651242, + "tokens_seen": 2455437312 + }, + { + "epoch": 0.74, + "learning_rate": 0.00020698896690070213, + "loss": 1.1571, + "theoretical_loss": 3.4018278289183845, + "tokens_seen": 2455502848 + }, + { + "epoch": 0.74, + "learning_rate": 0.00020697291875626883, + "loss": 1.2197, + "theoretical_loss": 3.4018206215705744, + "tokens_seen": 2455568384 + }, + { + "epoch": 0.74, + "learning_rate": 0.00020695687061183553, + "loss": 1.204, + "theoretical_loss": 3.4018134144689745, + "tokens_seen": 2455633920 + }, + { + "epoch": 0.74, + "learning_rate": 0.00020694082246740223, + "loss": 1.3403, + "theoretical_loss": 3.4018062076135696, + "tokens_seen": 2455699456 + }, + { + "epoch": 0.74, + "learning_rate": 0.00020692477432296893, + "loss": 1.2289, + "theoretical_loss": 3.4017990010043446, + "tokens_seen": 2455764992 + }, + { + "epoch": 0.74, + "learning_rate": 0.00020690872617853563, + "loss": 1.1916, + "theoretical_loss": 3.4017917946412854, + "tokens_seen": 2455830528 + }, + { + "epoch": 0.74, + "learning_rate": 0.00020689267803410233, + "loss": 1.3797, + "theoretical_loss": 3.401784588524376, + "tokens_seen": 2455896064 + }, + { + "epoch": 0.74, + "learning_rate": 0.00020687662988966903, + "loss": 1.2606, + "theoretical_loss": 3.401777382653602, + "tokens_seen": 2455961600 + }, + { + "epoch": 0.74, + "learning_rate": 0.00020686058174523573, + "loss": 1.3437, + "theoretical_loss": 3.4017701770289483, + "tokens_seen": 2456027136 + }, + { + "epoch": 0.74, + "learning_rate": 0.00020684453360080243, + "loss": 1.3383, + "theoretical_loss": 3.4017629716504, + "tokens_seen": 2456092672 + }, + { + "epoch": 0.74, + "learning_rate": 0.00020682848545636913, + "loss": 1.2839, + "theoretical_loss": 3.4017557665179416, + "tokens_seen": 2456158208 + }, + { + "epoch": 0.74, + "learning_rate": 0.00020681243731193583, + "loss": 1.2751, + "theoretical_loss": 3.401748561631559, + "tokens_seen": 2456223744 + }, + { + "epoch": 0.74, + "learning_rate": 0.00020679638916750253, + "loss": 1.1687, + "theoretical_loss": 3.4017413569912365, + "tokens_seen": 2456289280 + }, + { + "epoch": 0.74, + "learning_rate": 0.00020678034102306924, + "loss": 1.3784, + "theoretical_loss": 3.4017341525969598, + "tokens_seen": 2456354816 + }, + { + "epoch": 0.74, + "learning_rate": 0.00020676429287863594, + "loss": 1.2013, + "theoretical_loss": 3.401726948448713, + "tokens_seen": 2456420352 + }, + { + "epoch": 0.74, + "learning_rate": 0.0002067482447342026, + "loss": 1.4293, + "theoretical_loss": 3.401719744546482, + "tokens_seen": 2456485888 + }, + { + "epoch": 0.74, + "learning_rate": 0.0002067321965897693, + "loss": 1.1534, + "theoretical_loss": 3.4017125408902515, + "tokens_seen": 2456551424 + }, + { + "epoch": 0.74, + "learning_rate": 0.000206716148445336, + "loss": 1.193, + "theoretical_loss": 3.4017053374800064, + "tokens_seen": 2456616960 + }, + { + "epoch": 0.74, + "learning_rate": 0.0002067001003009027, + "loss": 1.3001, + "theoretical_loss": 3.4016981343157324, + "tokens_seen": 2456682496 + }, + { + "epoch": 0.74, + "learning_rate": 0.0002066840521564694, + "loss": 1.1972, + "theoretical_loss": 3.4016909313974133, + "tokens_seen": 2456748032 + }, + { + "epoch": 0.74, + "learning_rate": 0.0002066680040120361, + "loss": 1.2789, + "theoretical_loss": 3.401683728725035, + "tokens_seen": 2456813568 + }, + { + "epoch": 0.74, + "learning_rate": 0.0002066519558676028, + "loss": 1.3162, + "theoretical_loss": 3.401676526298583, + "tokens_seen": 2456879104 + }, + { + "epoch": 0.74, + "learning_rate": 0.00020663590772316951, + "loss": 1.2207, + "theoretical_loss": 3.4016693241180413, + "tokens_seen": 2456944640 + }, + { + "epoch": 0.74, + "learning_rate": 0.00020661985957873621, + "loss": 1.3488, + "theoretical_loss": 3.401662122183396, + "tokens_seen": 2457010176 + }, + { + "epoch": 0.74, + "learning_rate": 0.00020660381143430292, + "loss": 1.2494, + "theoretical_loss": 3.4016549204946305, + "tokens_seen": 2457075712 + }, + { + "epoch": 0.74, + "learning_rate": 0.00020658776328986962, + "loss": 1.3143, + "theoretical_loss": 3.4016477190517316, + "tokens_seen": 2457141248 + }, + { + "epoch": 0.74, + "learning_rate": 0.00020657171514543632, + "loss": 1.3363, + "theoretical_loss": 3.4016405178546836, + "tokens_seen": 2457206784 + }, + { + "epoch": 0.74, + "learning_rate": 0.00020655566700100302, + "loss": 1.4204, + "theoretical_loss": 3.4016333169034714, + "tokens_seen": 2457272320 + }, + { + "epoch": 0.74, + "learning_rate": 0.00020653961885656972, + "loss": 1.2152, + "theoretical_loss": 3.4016261161980803, + "tokens_seen": 2457337856 + }, + { + "epoch": 0.74, + "learning_rate": 0.00020652357071213642, + "loss": 1.2715, + "theoretical_loss": 3.4016189157384953, + "tokens_seen": 2457403392 + }, + { + "epoch": 0.74, + "learning_rate": 0.00020650752256770312, + "loss": 1.2231, + "theoretical_loss": 3.4016117155247017, + "tokens_seen": 2457468928 + }, + { + "epoch": 0.74, + "learning_rate": 0.00020649147442326982, + "loss": 1.2239, + "theoretical_loss": 3.4016045155566843, + "tokens_seen": 2457534464 + }, + { + "epoch": 0.74, + "objective/train/docs_used": 894282, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.235457420349121, + "objective/train/theoretical_loss": 3.4015973158344277, + "objective/train/tokens_used": 2478060000, + "theoretical_loss": 3.4015973158344277, + "tokens_seen": 2457600000 + }, + { + "epoch": 0.74, + "learning_rate": 0.00020647542627883652, + "loss": 1.2467, + "theoretical_loss": 3.4015973158344277, + "tokens_seen": 2457600000 + }, + { + "epoch": 0.74, + "learning_rate": 0.00020645937813440322, + "loss": 1.2481, + "theoretical_loss": 3.4015901163579176, + "tokens_seen": 2457665536 + }, + { + "epoch": 0.74, + "learning_rate": 0.00020644332998996992, + "loss": 1.2489, + "theoretical_loss": 3.401582917127139, + "tokens_seen": 2457731072 + }, + { + "epoch": 0.74, + "learning_rate": 0.00020642728184553662, + "loss": 1.3301, + "theoretical_loss": 3.401575718142077, + "tokens_seen": 2457796608 + }, + { + "epoch": 0.74, + "learning_rate": 0.00020641123370110332, + "loss": 1.2332, + "theoretical_loss": 3.4015685194027165, + "tokens_seen": 2457862144 + }, + { + "epoch": 0.74, + "learning_rate": 0.00020639518555667002, + "loss": 1.2758, + "theoretical_loss": 3.4015613209090425, + "tokens_seen": 2457927680 + }, + { + "epoch": 0.74, + "learning_rate": 0.00020637913741223672, + "loss": 1.2245, + "theoretical_loss": 3.40155412266104, + "tokens_seen": 2457993216 + }, + { + "epoch": 0.74, + "learning_rate": 0.0002063630892678034, + "loss": 1.2991, + "theoretical_loss": 3.4015469246586942, + "tokens_seen": 2458058752 + }, + { + "epoch": 0.74, + "learning_rate": 0.0002063470411233701, + "loss": 1.1871, + "theoretical_loss": 3.4015397269019902, + "tokens_seen": 2458124288 + }, + { + "epoch": 0.74, + "learning_rate": 0.0002063309929789368, + "loss": 1.2335, + "theoretical_loss": 3.401532529390913, + "tokens_seen": 2458189824 + }, + { + "epoch": 0.74, + "learning_rate": 0.0002063149448345035, + "loss": 1.1928, + "theoretical_loss": 3.401525332125448, + "tokens_seen": 2458255360 + }, + { + "epoch": 0.74, + "learning_rate": 0.0002062988966900702, + "loss": 1.4711, + "theoretical_loss": 3.4015181351055794, + "tokens_seen": 2458320896 + }, + { + "epoch": 0.74, + "learning_rate": 0.0002062828485456369, + "loss": 1.3169, + "theoretical_loss": 3.4015109383312936, + "tokens_seen": 2458386432 + }, + { + "epoch": 0.74, + "learning_rate": 0.0002062668004012036, + "loss": 1.3619, + "theoretical_loss": 3.4015037418025744, + "tokens_seen": 2458451968 + }, + { + "epoch": 0.75, + "learning_rate": 0.0002062507522567703, + "loss": 1.3025, + "theoretical_loss": 3.4014965455194073, + "tokens_seen": 2458517504 + }, + { + "epoch": 0.75, + "learning_rate": 0.000206234704112337, + "loss": 1.3468, + "theoretical_loss": 3.401489349481778, + "tokens_seen": 2458583040 + }, + { + "epoch": 0.75, + "learning_rate": 0.0002062186559679037, + "loss": 1.2553, + "theoretical_loss": 3.4014821536896704, + "tokens_seen": 2458648576 + }, + { + "epoch": 0.75, + "learning_rate": 0.0002062026078234704, + "loss": 1.3377, + "theoretical_loss": 3.401474958143071, + "tokens_seen": 2458714112 + }, + { + "epoch": 0.75, + "learning_rate": 0.0002061865596790371, + "loss": 1.2635, + "theoretical_loss": 3.401467762841963, + "tokens_seen": 2458779648 + }, + { + "epoch": 0.75, + "learning_rate": 0.0002061705115346038, + "loss": 1.1818, + "theoretical_loss": 3.4014605677863337, + "tokens_seen": 2458845184 + }, + { + "epoch": 0.75, + "learning_rate": 0.0002061544633901705, + "loss": 1.2937, + "theoretical_loss": 3.4014533729761665, + "tokens_seen": 2458910720 + }, + { + "epoch": 0.75, + "learning_rate": 0.0002061384152457372, + "loss": 1.3191, + "theoretical_loss": 3.4014461784114474, + "tokens_seen": 2458976256 + }, + { + "epoch": 0.75, + "learning_rate": 0.0002061223671013039, + "loss": 1.2999, + "theoretical_loss": 3.401438984092161, + "tokens_seen": 2459041792 + }, + { + "epoch": 0.75, + "learning_rate": 0.0002061063189568706, + "loss": 1.1225, + "theoretical_loss": 3.401431790018292, + "tokens_seen": 2459107328 + }, + { + "epoch": 0.75, + "learning_rate": 0.0002060902708124373, + "loss": 1.1802, + "theoretical_loss": 3.401424596189827, + "tokens_seen": 2459172864 + }, + { + "epoch": 0.75, + "learning_rate": 0.000206074222668004, + "loss": 1.1822, + "theoretical_loss": 3.4014174026067496, + "tokens_seen": 2459238400 + }, + { + "epoch": 0.75, + "learning_rate": 0.0002060581745235707, + "loss": 1.2198, + "theoretical_loss": 3.401410209269045, + "tokens_seen": 2459303936 + }, + { + "epoch": 0.75, + "learning_rate": 0.0002060421263791374, + "loss": 1.2378, + "theoretical_loss": 3.4014030161766993, + "tokens_seen": 2459369472 + }, + { + "epoch": 0.75, + "learning_rate": 0.0002060260782347041, + "loss": 1.2186, + "theoretical_loss": 3.401395823329697, + "tokens_seen": 2459435008 + }, + { + "epoch": 0.75, + "learning_rate": 0.0002060100300902708, + "loss": 1.3243, + "theoretical_loss": 3.401388630728023, + "tokens_seen": 2459500544 + }, + { + "epoch": 0.75, + "learning_rate": 0.00020599398194583754, + "loss": 1.3261, + "theoretical_loss": 3.4013814383716623, + "tokens_seen": 2459566080 + }, + { + "epoch": 0.75, + "learning_rate": 0.00020597793380140424, + "loss": 1.1894, + "theoretical_loss": 3.401374246260601, + "tokens_seen": 2459631616 + }, + { + "epoch": 0.75, + "learning_rate": 0.00020596188565697094, + "loss": 1.3588, + "theoretical_loss": 3.4013670543948225, + "tokens_seen": 2459697152 + }, + { + "epoch": 0.75, + "learning_rate": 0.00020594583751253764, + "loss": 1.3236, + "theoretical_loss": 3.4013598627743136, + "tokens_seen": 2459762688 + }, + { + "epoch": 0.75, + "learning_rate": 0.00020592978936810434, + "loss": 1.1884, + "theoretical_loss": 3.4013526713990583, + "tokens_seen": 2459828224 + }, + { + "epoch": 0.75, + "learning_rate": 0.00020591374122367104, + "loss": 1.2856, + "theoretical_loss": 3.4013454802690424, + "tokens_seen": 2459893760 + }, + { + "epoch": 0.75, + "learning_rate": 0.00020589769307923774, + "loss": 1.347, + "theoretical_loss": 3.4013382893842503, + "tokens_seen": 2459959296 + }, + { + "epoch": 0.75, + "learning_rate": 0.00020588164493480444, + "loss": 1.3176, + "theoretical_loss": 3.401331098744668, + "tokens_seen": 2460024832 + }, + { + "epoch": 0.75, + "learning_rate": 0.00020586559679037114, + "loss": 1.2385, + "theoretical_loss": 3.4013239083502795, + "tokens_seen": 2460090368 + }, + { + "epoch": 0.75, + "learning_rate": 0.00020584954864593785, + "loss": 1.3374, + "theoretical_loss": 3.4013167182010706, + "tokens_seen": 2460155904 + }, + { + "epoch": 0.75, + "learning_rate": 0.00020583350050150455, + "loss": 1.2648, + "theoretical_loss": 3.4013095282970265, + "tokens_seen": 2460221440 + }, + { + "epoch": 0.75, + "learning_rate": 0.00020581745235707125, + "loss": 1.4165, + "theoretical_loss": 3.401302338638132, + "tokens_seen": 2460286976 + }, + { + "epoch": 0.75, + "learning_rate": 0.00020580140421263795, + "loss": 1.1938, + "theoretical_loss": 3.4012951492243726, + "tokens_seen": 2460352512 + }, + { + "epoch": 0.75, + "learning_rate": 0.00020578535606820465, + "loss": 1.3056, + "theoretical_loss": 3.4012879600557326, + "tokens_seen": 2460418048 + }, + { + "epoch": 0.75, + "learning_rate": 0.00020576930792377135, + "loss": 1.1958, + "theoretical_loss": 3.4012807711321984, + "tokens_seen": 2460483584 + }, + { + "epoch": 0.75, + "learning_rate": 0.00020575325977933805, + "loss": 1.2104, + "theoretical_loss": 3.4012735824537534, + "tokens_seen": 2460549120 + }, + { + "epoch": 0.75, + "learning_rate": 0.00020573721163490475, + "loss": 1.1543, + "theoretical_loss": 3.401266394020385, + "tokens_seen": 2460614656 + }, + { + "epoch": 0.75, + "learning_rate": 0.00020572116349047142, + "loss": 1.1701, + "theoretical_loss": 3.401259205832076, + "tokens_seen": 2460680192 + }, + { + "epoch": 0.75, + "learning_rate": 0.00020570511534603812, + "loss": 1.2684, + "theoretical_loss": 3.4012520178888126, + "tokens_seen": 2460745728 + }, + { + "epoch": 0.75, + "learning_rate": 0.00020568906720160482, + "loss": 1.1647, + "theoretical_loss": 3.40124483019058, + "tokens_seen": 2460811264 + }, + { + "epoch": 0.75, + "objective/train/docs_used": 895477, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.3390307426452637, + "objective/train/theoretical_loss": 3.401237642737363, + "objective/train/tokens_used": 2481336800, + "theoretical_loss": 3.401237642737363, + "tokens_seen": 2460876800 + }, + { + "epoch": 0.75, + "learning_rate": 0.00020567301905717153, + "loss": 1.3025, + "theoretical_loss": 3.401237642737363, + "tokens_seen": 2460876800 + }, + { + "epoch": 0.75, + "learning_rate": 0.00020565697091273823, + "loss": 1.3114, + "theoretical_loss": 3.4012304555291473, + "tokens_seen": 2460942336 + }, + { + "epoch": 0.75, + "learning_rate": 0.00020564092276830493, + "loss": 1.282, + "theoretical_loss": 3.4012232685659174, + "tokens_seen": 2461007872 + }, + { + "epoch": 0.75, + "learning_rate": 0.00020562487462387163, + "loss": 1.2504, + "theoretical_loss": 3.4012160818476587, + "tokens_seen": 2461073408 + }, + { + "epoch": 0.75, + "learning_rate": 0.00020560882647943833, + "loss": 1.2526, + "theoretical_loss": 3.401208895374356, + "tokens_seen": 2461138944 + }, + { + "epoch": 0.75, + "learning_rate": 0.00020559277833500503, + "loss": 1.218, + "theoretical_loss": 3.4012017091459947, + "tokens_seen": 2461204480 + }, + { + "epoch": 0.75, + "learning_rate": 0.00020557673019057173, + "loss": 1.3216, + "theoretical_loss": 3.40119452316256, + "tokens_seen": 2461270016 + }, + { + "epoch": 0.75, + "learning_rate": 0.00020556068204613843, + "loss": 1.3044, + "theoretical_loss": 3.4011873374240373, + "tokens_seen": 2461335552 + }, + { + "epoch": 0.75, + "learning_rate": 0.00020554463390170513, + "loss": 1.3011, + "theoretical_loss": 3.401180151930411, + "tokens_seen": 2461401088 + }, + { + "epoch": 0.75, + "learning_rate": 0.00020552858575727183, + "loss": 1.2848, + "theoretical_loss": 3.4011729666816666, + "tokens_seen": 2461466624 + }, + { + "epoch": 0.75, + "learning_rate": 0.00020551253761283853, + "loss": 1.2698, + "theoretical_loss": 3.4011657816777894, + "tokens_seen": 2461532160 + }, + { + "epoch": 0.75, + "learning_rate": 0.00020549648946840523, + "loss": 1.2159, + "theoretical_loss": 3.4011585969187648, + "tokens_seen": 2461597696 + }, + { + "epoch": 0.75, + "learning_rate": 0.00020548044132397193, + "loss": 1.2349, + "theoretical_loss": 3.4011514124045767, + "tokens_seen": 2461663232 + }, + { + "epoch": 0.75, + "learning_rate": 0.00020546439317953863, + "loss": 1.172, + "theoretical_loss": 3.4011442281352116, + "tokens_seen": 2461728768 + }, + { + "epoch": 0.75, + "learning_rate": 0.00020544834503510533, + "loss": 1.2403, + "theoretical_loss": 3.401137044110654, + "tokens_seen": 2461794304 + }, + { + "epoch": 0.75, + "learning_rate": 0.00020543229689067203, + "loss": 1.3121, + "theoretical_loss": 3.4011298603308893, + "tokens_seen": 2461859840 + }, + { + "epoch": 0.75, + "learning_rate": 0.00020541624874623874, + "loss": 1.3222, + "theoretical_loss": 3.401122676795902, + "tokens_seen": 2461925376 + }, + { + "epoch": 0.75, + "learning_rate": 0.00020540020060180544, + "loss": 1.2121, + "theoretical_loss": 3.401115493505678, + "tokens_seen": 2461990912 + }, + { + "epoch": 0.75, + "learning_rate": 0.00020538415245737214, + "loss": 1.247, + "theoretical_loss": 3.4011083104602022, + "tokens_seen": 2462056448 + }, + { + "epoch": 0.75, + "learning_rate": 0.00020536810431293884, + "loss": 1.1329, + "theoretical_loss": 3.40110112765946, + "tokens_seen": 2462121984 + }, + { + "epoch": 0.75, + "learning_rate": 0.00020535205616850554, + "loss": 1.2526, + "theoretical_loss": 3.401093945103436, + "tokens_seen": 2462187520 + }, + { + "epoch": 0.75, + "learning_rate": 0.0002053360080240722, + "loss": 1.3945, + "theoretical_loss": 3.4010867627921155, + "tokens_seen": 2462253056 + }, + { + "epoch": 0.75, + "learning_rate": 0.0002053199598796389, + "loss": 1.3481, + "theoretical_loss": 3.4010795807254843, + "tokens_seen": 2462318592 + }, + { + "epoch": 0.75, + "learning_rate": 0.0002053039117352056, + "loss": 1.2647, + "theoretical_loss": 3.4010723989035263, + "tokens_seen": 2462384128 + }, + { + "epoch": 0.75, + "learning_rate": 0.0002052878635907723, + "loss": 1.2874, + "theoretical_loss": 3.401065217326228, + "tokens_seen": 2462449664 + }, + { + "epoch": 0.75, + "learning_rate": 0.00020527181544633901, + "loss": 1.1828, + "theoretical_loss": 3.4010580359935734, + "tokens_seen": 2462515200 + }, + { + "epoch": 0.75, + "learning_rate": 0.00020525576730190571, + "loss": 1.3858, + "theoretical_loss": 3.4010508549055487, + "tokens_seen": 2462580736 + }, + { + "epoch": 0.75, + "learning_rate": 0.00020523971915747242, + "loss": 1.2704, + "theoretical_loss": 3.401043674062138, + "tokens_seen": 2462646272 + }, + { + "epoch": 0.75, + "learning_rate": 0.00020522367101303912, + "loss": 1.331, + "theoretical_loss": 3.4010364934633275, + "tokens_seen": 2462711808 + }, + { + "epoch": 0.75, + "learning_rate": 0.00020520762286860582, + "loss": 1.2294, + "theoretical_loss": 3.4010293131091016, + "tokens_seen": 2462777344 + }, + { + "epoch": 0.75, + "learning_rate": 0.00020519157472417252, + "loss": 1.4027, + "theoretical_loss": 3.401022132999446, + "tokens_seen": 2462842880 + }, + { + "epoch": 0.75, + "learning_rate": 0.00020517552657973922, + "loss": 1.4433, + "theoretical_loss": 3.401014953134345, + "tokens_seen": 2462908416 + }, + { + "epoch": 0.75, + "learning_rate": 0.00020515947843530592, + "loss": 1.1584, + "theoretical_loss": 3.401007773513785, + "tokens_seen": 2462973952 + }, + { + "epoch": 0.75, + "learning_rate": 0.00020514343029087262, + "loss": 1.316, + "theoretical_loss": 3.4010005941377504, + "tokens_seen": 2463039488 + }, + { + "epoch": 0.75, + "learning_rate": 0.00020512738214643932, + "loss": 1.266, + "theoretical_loss": 3.400993415006226, + "tokens_seen": 2463105024 + }, + { + "epoch": 0.75, + "learning_rate": 0.00020511133400200602, + "loss": 1.2138, + "theoretical_loss": 3.4009862361191976, + "tokens_seen": 2463170560 + }, + { + "epoch": 0.75, + "learning_rate": 0.00020509528585757272, + "loss": 1.3226, + "theoretical_loss": 3.4009790574766505, + "tokens_seen": 2463236096 + }, + { + "epoch": 0.75, + "learning_rate": 0.00020507923771313942, + "loss": 1.2614, + "theoretical_loss": 3.4009718790785692, + "tokens_seen": 2463301632 + }, + { + "epoch": 0.75, + "learning_rate": 0.00020506318956870612, + "loss": 1.3779, + "theoretical_loss": 3.40096470092494, + "tokens_seen": 2463367168 + }, + { + "epoch": 0.75, + "learning_rate": 0.00020504714142427282, + "loss": 1.2743, + "theoretical_loss": 3.4009575230157467, + "tokens_seen": 2463432704 + }, + { + "epoch": 0.75, + "learning_rate": 0.00020503109327983952, + "loss": 1.378, + "theoretical_loss": 3.4009503453509753, + "tokens_seen": 2463498240 + }, + { + "epoch": 0.75, + "learning_rate": 0.00020501504513540622, + "loss": 1.2254, + "theoretical_loss": 3.4009431679306106, + "tokens_seen": 2463563776 + }, + { + "epoch": 0.75, + "learning_rate": 0.00020499899699097292, + "loss": 1.1759, + "theoretical_loss": 3.400935990754638, + "tokens_seen": 2463629312 + }, + { + "epoch": 0.75, + "learning_rate": 0.00020498294884653963, + "loss": 1.2649, + "theoretical_loss": 3.4009288138230427, + "tokens_seen": 2463694848 + }, + { + "epoch": 0.75, + "learning_rate": 0.00020496690070210633, + "loss": 1.2877, + "theoretical_loss": 3.4009216371358097, + "tokens_seen": 2463760384 + }, + { + "epoch": 0.75, + "learning_rate": 0.000204950852557673, + "loss": 1.2216, + "theoretical_loss": 3.4009144606929245, + "tokens_seen": 2463825920 + }, + { + "epoch": 0.75, + "learning_rate": 0.0002049348044132397, + "loss": 1.1803, + "theoretical_loss": 3.400907284494372, + "tokens_seen": 2463891456 + }, + { + "epoch": 0.75, + "learning_rate": 0.0002049187562688064, + "loss": 1.2881, + "theoretical_loss": 3.4009001085401374, + "tokens_seen": 2463956992 + }, + { + "epoch": 0.75, + "learning_rate": 0.0002049027081243731, + "loss": 1.2546, + "theoretical_loss": 3.4008929328302058, + "tokens_seen": 2464022528 + }, + { + "epoch": 0.75, + "learning_rate": 0.0002048866599799398, + "loss": 1.278, + "theoretical_loss": 3.400885757364563, + "tokens_seen": 2464088064 + }, + { + "epoch": 0.75, + "objective/train/docs_used": 896700, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.115243911743164, + "objective/train/theoretical_loss": 3.4008785821431933, + "objective/train/tokens_used": 2484613600, + "theoretical_loss": 3.4008785821431933, + "tokens_seen": 2464153600 + }, + { + "epoch": 0.75, + "learning_rate": 0.0002048706118355065, + "loss": 1.2604, + "theoretical_loss": 3.4008785821431933, + "tokens_seen": 2464153600 + }, + { + "epoch": 0.75, + "learning_rate": 0.0002048545636910732, + "loss": 1.3184, + "theoretical_loss": 3.400871407166082, + "tokens_seen": 2464219136 + }, + { + "epoch": 0.75, + "learning_rate": 0.0002048385155466399, + "loss": 1.3373, + "theoretical_loss": 3.4008642324332152, + "tokens_seen": 2464284672 + }, + { + "epoch": 0.75, + "learning_rate": 0.0002048224674022066, + "loss": 1.2254, + "theoretical_loss": 3.4008570579445774, + "tokens_seen": 2464350208 + }, + { + "epoch": 0.75, + "learning_rate": 0.0002048064192577733, + "loss": 1.2948, + "theoretical_loss": 3.4008498837001535, + "tokens_seen": 2464415744 + }, + { + "epoch": 0.75, + "learning_rate": 0.00020479037111334006, + "loss": 1.219, + "theoretical_loss": 3.4008427096999294, + "tokens_seen": 2464481280 + }, + { + "epoch": 0.75, + "learning_rate": 0.00020477432296890676, + "loss": 1.2246, + "theoretical_loss": 3.40083553594389, + "tokens_seen": 2464546816 + }, + { + "epoch": 0.75, + "learning_rate": 0.00020475827482447346, + "loss": 1.2949, + "theoretical_loss": 3.40082836243202, + "tokens_seen": 2464612352 + }, + { + "epoch": 0.75, + "learning_rate": 0.00020474222668004016, + "loss": 1.2299, + "theoretical_loss": 3.400821189164305, + "tokens_seen": 2464677888 + }, + { + "epoch": 0.75, + "learning_rate": 0.00020472617853560686, + "loss": 1.1965, + "theoretical_loss": 3.400814016140731, + "tokens_seen": 2464743424 + }, + { + "epoch": 0.75, + "learning_rate": 0.00020471013039117354, + "loss": 1.1665, + "theoretical_loss": 3.4008068433612824, + "tokens_seen": 2464808960 + }, + { + "epoch": 0.75, + "learning_rate": 0.00020469408224674024, + "loss": 1.1299, + "theoretical_loss": 3.400799670825944, + "tokens_seen": 2464874496 + }, + { + "epoch": 0.75, + "learning_rate": 0.00020467803410230694, + "loss": 1.3652, + "theoretical_loss": 3.4007924985347016, + "tokens_seen": 2464940032 + }, + { + "epoch": 0.75, + "learning_rate": 0.00020466198595787364, + "loss": 1.2746, + "theoretical_loss": 3.40078532648754, + "tokens_seen": 2465005568 + }, + { + "epoch": 0.75, + "learning_rate": 0.00020464593781344034, + "loss": 1.3543, + "theoretical_loss": 3.4007781546844447, + "tokens_seen": 2465071104 + }, + { + "epoch": 0.75, + "learning_rate": 0.00020462988966900704, + "loss": 1.2545, + "theoretical_loss": 3.400770983125401, + "tokens_seen": 2465136640 + }, + { + "epoch": 0.75, + "learning_rate": 0.00020461384152457374, + "loss": 1.3363, + "theoretical_loss": 3.400763811810394, + "tokens_seen": 2465202176 + }, + { + "epoch": 0.75, + "learning_rate": 0.00020459779338014044, + "loss": 1.2186, + "theoretical_loss": 3.400756640739409, + "tokens_seen": 2465267712 + }, + { + "epoch": 0.75, + "learning_rate": 0.00020458174523570714, + "loss": 1.2686, + "theoretical_loss": 3.400749469912431, + "tokens_seen": 2465333248 + }, + { + "epoch": 0.75, + "learning_rate": 0.00020456569709127384, + "loss": 1.2148, + "theoretical_loss": 3.400742299329445, + "tokens_seen": 2465398784 + }, + { + "epoch": 0.75, + "learning_rate": 0.00020454964894684054, + "loss": 1.3341, + "theoretical_loss": 3.400735128990437, + "tokens_seen": 2465464320 + }, + { + "epoch": 0.75, + "learning_rate": 0.00020453360080240724, + "loss": 1.3208, + "theoretical_loss": 3.4007279588953914, + "tokens_seen": 2465529856 + }, + { + "epoch": 0.75, + "learning_rate": 0.00020451755265797394, + "loss": 1.3098, + "theoretical_loss": 3.4007207890442936, + "tokens_seen": 2465595392 + }, + { + "epoch": 0.75, + "learning_rate": 0.00020450150451354064, + "loss": 1.3101, + "theoretical_loss": 3.4007136194371292, + "tokens_seen": 2465660928 + }, + { + "epoch": 0.75, + "learning_rate": 0.00020448545636910734, + "loss": 1.2828, + "theoretical_loss": 3.4007064500738835, + "tokens_seen": 2465726464 + }, + { + "epoch": 0.75, + "learning_rate": 0.00020446940822467405, + "loss": 1.277, + "theoretical_loss": 3.400699280954541, + "tokens_seen": 2465792000 + }, + { + "epoch": 0.75, + "learning_rate": 0.00020445336008024075, + "loss": 1.3318, + "theoretical_loss": 3.400692112079087, + "tokens_seen": 2465857536 + }, + { + "epoch": 0.75, + "learning_rate": 0.00020443731193580745, + "loss": 1.0963, + "theoretical_loss": 3.4006849434475077, + "tokens_seen": 2465923072 + }, + { + "epoch": 0.75, + "learning_rate": 0.00020442126379137415, + "loss": 1.2578, + "theoretical_loss": 3.4006777750597874, + "tokens_seen": 2465988608 + }, + { + "epoch": 0.75, + "learning_rate": 0.00020440521564694085, + "loss": 1.2406, + "theoretical_loss": 3.4006706069159116, + "tokens_seen": 2466054144 + }, + { + "epoch": 0.75, + "learning_rate": 0.00020438916750250755, + "loss": 1.2691, + "theoretical_loss": 3.400663439015865, + "tokens_seen": 2466119680 + }, + { + "epoch": 0.75, + "learning_rate": 0.00020437311935807425, + "loss": 1.3551, + "theoretical_loss": 3.400656271359634, + "tokens_seen": 2466185216 + }, + { + "epoch": 0.75, + "learning_rate": 0.00020435707121364095, + "loss": 1.3658, + "theoretical_loss": 3.4006491039472033, + "tokens_seen": 2466250752 + }, + { + "epoch": 0.75, + "learning_rate": 0.00020434102306920765, + "loss": 1.2551, + "theoretical_loss": 3.4006419367785576, + "tokens_seen": 2466316288 + }, + { + "epoch": 0.75, + "learning_rate": 0.00020432497492477432, + "loss": 1.353, + "theoretical_loss": 3.4006347698536823, + "tokens_seen": 2466381824 + }, + { + "epoch": 0.75, + "learning_rate": 0.00020430892678034102, + "loss": 1.1684, + "theoretical_loss": 3.4006276031725635, + "tokens_seen": 2466447360 + }, + { + "epoch": 0.75, + "learning_rate": 0.00020429287863590773, + "loss": 1.4369, + "theoretical_loss": 3.4006204367351853, + "tokens_seen": 2466512896 + }, + { + "epoch": 0.75, + "learning_rate": 0.00020427683049147443, + "loss": 1.391, + "theoretical_loss": 3.4006132705415335, + "tokens_seen": 2466578432 + }, + { + "epoch": 0.75, + "learning_rate": 0.00020426078234704113, + "loss": 1.1919, + "theoretical_loss": 3.4006061045915934, + "tokens_seen": 2466643968 + }, + { + "epoch": 0.75, + "learning_rate": 0.00020424473420260783, + "loss": 1.2344, + "theoretical_loss": 3.40059893888535, + "tokens_seen": 2466709504 + }, + { + "epoch": 0.75, + "learning_rate": 0.00020422868605817453, + "loss": 1.347, + "theoretical_loss": 3.4005917734227884, + "tokens_seen": 2466775040 + }, + { + "epoch": 0.75, + "learning_rate": 0.00020421263791374123, + "loss": 1.295, + "theoretical_loss": 3.4005846082038946, + "tokens_seen": 2466840576 + }, + { + "epoch": 0.75, + "learning_rate": 0.00020419658976930793, + "loss": 1.2149, + "theoretical_loss": 3.400577443228653, + "tokens_seen": 2466906112 + }, + { + "epoch": 0.75, + "learning_rate": 0.00020418054162487463, + "loss": 1.26, + "theoretical_loss": 3.400570278497049, + "tokens_seen": 2466971648 + }, + { + "epoch": 0.75, + "learning_rate": 0.00020416449348044133, + "loss": 1.3761, + "theoretical_loss": 3.400563114009068, + "tokens_seen": 2467037184 + }, + { + "epoch": 0.75, + "learning_rate": 0.00020414844533600803, + "loss": 1.2435, + "theoretical_loss": 3.4005559497646956, + "tokens_seen": 2467102720 + }, + { + "epoch": 0.75, + "learning_rate": 0.00020413239719157473, + "loss": 1.4026, + "theoretical_loss": 3.4005487857639167, + "tokens_seen": 2467168256 + }, + { + "epoch": 0.75, + "learning_rate": 0.00020411634904714143, + "loss": 1.3484, + "theoretical_loss": 3.400541622006716, + "tokens_seen": 2467233792 + }, + { + "epoch": 0.75, + "learning_rate": 0.00020410030090270813, + "loss": 1.219, + "theoretical_loss": 3.40053445849308, + "tokens_seen": 2467299328 + }, + { + "epoch": 0.75, + "learning_rate": 0.00020408425275827483, + "loss": 1.3994, + "theoretical_loss": 3.4005272952229926, + "tokens_seen": 2467364864 + }, + { + "epoch": 0.75, + "objective/train/docs_used": 897881, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.262965202331543, + "objective/train/theoretical_loss": 3.40052013219644, + "objective/train/tokens_used": 2487890400, + "theoretical_loss": 3.40052013219644, + "tokens_seen": 2467430400 + }, + { + "epoch": 0.75, + "learning_rate": 0.00020406820461384153, + "loss": 1.2874, + "theoretical_loss": 3.40052013219644, + "tokens_seen": 2467430400 + }, + { + "epoch": 0.75, + "learning_rate": 0.00020405215646940823, + "loss": 1.334, + "theoretical_loss": 3.4005129694134073, + "tokens_seen": 2467495936 + }, + { + "epoch": 0.75, + "learning_rate": 0.00020403610832497494, + "loss": 1.223, + "theoretical_loss": 3.400505806873879, + "tokens_seen": 2467561472 + }, + { + "epoch": 0.75, + "learning_rate": 0.00020402006018054164, + "loss": 1.1891, + "theoretical_loss": 3.4004986445778416, + "tokens_seen": 2467627008 + }, + { + "epoch": 0.75, + "learning_rate": 0.00020400401203610834, + "loss": 1.3833, + "theoretical_loss": 3.400491482525279, + "tokens_seen": 2467692544 + }, + { + "epoch": 0.75, + "learning_rate": 0.00020398796389167504, + "loss": 1.2087, + "theoretical_loss": 3.4004843207161777, + "tokens_seen": 2467758080 + }, + { + "epoch": 0.75, + "learning_rate": 0.00020397191574724174, + "loss": 1.2763, + "theoretical_loss": 3.4004771591505225, + "tokens_seen": 2467823616 + }, + { + "epoch": 0.75, + "learning_rate": 0.00020395586760280844, + "loss": 1.3368, + "theoretical_loss": 3.4004699978282984, + "tokens_seen": 2467889152 + }, + { + "epoch": 0.75, + "learning_rate": 0.0002039398194583751, + "loss": 1.2626, + "theoretical_loss": 3.4004628367494907, + "tokens_seen": 2467954688 + }, + { + "epoch": 0.75, + "learning_rate": 0.0002039237713139418, + "loss": 1.3149, + "theoretical_loss": 3.400455675914085, + "tokens_seen": 2468020224 + }, + { + "epoch": 0.75, + "learning_rate": 0.0002039077231695085, + "loss": 1.1547, + "theoretical_loss": 3.4004485153220663, + "tokens_seen": 2468085760 + }, + { + "epoch": 0.75, + "learning_rate": 0.00020389167502507521, + "loss": 1.1689, + "theoretical_loss": 3.40044135497342, + "tokens_seen": 2468151296 + }, + { + "epoch": 0.75, + "learning_rate": 0.00020387562688064191, + "loss": 1.3652, + "theoretical_loss": 3.400434194868131, + "tokens_seen": 2468216832 + }, + { + "epoch": 0.75, + "learning_rate": 0.00020385957873620862, + "loss": 1.3427, + "theoretical_loss": 3.400427035006185, + "tokens_seen": 2468282368 + }, + { + "epoch": 0.75, + "learning_rate": 0.00020384353059177532, + "loss": 1.3013, + "theoretical_loss": 3.400419875387567, + "tokens_seen": 2468347904 + }, + { + "epoch": 0.75, + "learning_rate": 0.00020382748244734202, + "loss": 1.2682, + "theoretical_loss": 3.400412716012263, + "tokens_seen": 2468413440 + }, + { + "epoch": 0.75, + "learning_rate": 0.00020381143430290872, + "loss": 1.2661, + "theoretical_loss": 3.400405556880257, + "tokens_seen": 2468478976 + }, + { + "epoch": 0.75, + "learning_rate": 0.00020379538615847542, + "loss": 1.256, + "theoretical_loss": 3.4003983979915358, + "tokens_seen": 2468544512 + }, + { + "epoch": 0.75, + "learning_rate": 0.00020377933801404212, + "loss": 1.2299, + "theoretical_loss": 3.4003912393460833, + "tokens_seen": 2468610048 + }, + { + "epoch": 0.75, + "learning_rate": 0.00020376328986960882, + "loss": 1.2634, + "theoretical_loss": 3.4003840809438852, + "tokens_seen": 2468675584 + }, + { + "epoch": 0.75, + "learning_rate": 0.00020374724172517552, + "loss": 1.178, + "theoretical_loss": 3.4003769227849268, + "tokens_seen": 2468741120 + }, + { + "epoch": 0.75, + "learning_rate": 0.00020373119358074222, + "loss": 1.2464, + "theoretical_loss": 3.4003697648691937, + "tokens_seen": 2468806656 + }, + { + "epoch": 0.75, + "learning_rate": 0.00020371514543630892, + "loss": 1.3695, + "theoretical_loss": 3.4003626071966706, + "tokens_seen": 2468872192 + }, + { + "epoch": 0.75, + "learning_rate": 0.00020369909729187562, + "loss": 1.238, + "theoretical_loss": 3.4003554497673436, + "tokens_seen": 2468937728 + }, + { + "epoch": 0.75, + "learning_rate": 0.00020368304914744232, + "loss": 1.2986, + "theoretical_loss": 3.400348292581197, + "tokens_seen": 2469003264 + }, + { + "epoch": 0.75, + "learning_rate": 0.00020366700100300902, + "loss": 1.2602, + "theoretical_loss": 3.400341135638217, + "tokens_seen": 2469068800 + }, + { + "epoch": 0.75, + "learning_rate": 0.00020365095285857572, + "loss": 1.33, + "theoretical_loss": 3.400333978938388, + "tokens_seen": 2469134336 + }, + { + "epoch": 0.75, + "learning_rate": 0.00020363490471414242, + "loss": 1.3108, + "theoretical_loss": 3.4003268224816963, + "tokens_seen": 2469199872 + }, + { + "epoch": 0.75, + "learning_rate": 0.00020361885656970912, + "loss": 1.2461, + "theoretical_loss": 3.4003196662681265, + "tokens_seen": 2469265408 + }, + { + "epoch": 0.75, + "learning_rate": 0.00020360280842527585, + "loss": 1.2612, + "theoretical_loss": 3.4003125102976637, + "tokens_seen": 2469330944 + }, + { + "epoch": 0.75, + "learning_rate": 0.00020358676028084255, + "loss": 1.2792, + "theoretical_loss": 3.400305354570294, + "tokens_seen": 2469396480 + }, + { + "epoch": 0.75, + "learning_rate": 0.00020357071213640925, + "loss": 1.2202, + "theoretical_loss": 3.4002981990860017, + "tokens_seen": 2469462016 + }, + { + "epoch": 0.75, + "learning_rate": 0.00020355466399197595, + "loss": 1.2149, + "theoretical_loss": 3.400291043844773, + "tokens_seen": 2469527552 + }, + { + "epoch": 0.75, + "learning_rate": 0.00020353861584754266, + "loss": 1.3027, + "theoretical_loss": 3.4002838888465927, + "tokens_seen": 2469593088 + }, + { + "epoch": 0.75, + "learning_rate": 0.00020352256770310936, + "loss": 1.2222, + "theoretical_loss": 3.4002767340914457, + "tokens_seen": 2469658624 + }, + { + "epoch": 0.75, + "learning_rate": 0.00020350651955867606, + "loss": 1.3184, + "theoretical_loss": 3.4002695795793185, + "tokens_seen": 2469724160 + }, + { + "epoch": 0.75, + "learning_rate": 0.00020349047141424276, + "loss": 1.3249, + "theoretical_loss": 3.400262425310195, + "tokens_seen": 2469789696 + }, + { + "epoch": 0.75, + "learning_rate": 0.00020347442326980946, + "loss": 1.2253, + "theoretical_loss": 3.4002552712840615, + "tokens_seen": 2469855232 + }, + { + "epoch": 0.75, + "learning_rate": 0.00020345837512537616, + "loss": 1.1999, + "theoretical_loss": 3.400248117500903, + "tokens_seen": 2469920768 + }, + { + "epoch": 0.75, + "learning_rate": 0.00020344232698094286, + "loss": 1.253, + "theoretical_loss": 3.400240963960705, + "tokens_seen": 2469986304 + }, + { + "epoch": 0.75, + "learning_rate": 0.00020342627883650956, + "loss": 1.3581, + "theoretical_loss": 3.4002338106634524, + "tokens_seen": 2470051840 + }, + { + "epoch": 0.75, + "learning_rate": 0.00020341023069207626, + "loss": 1.4274, + "theoretical_loss": 3.4002266576091307, + "tokens_seen": 2470117376 + }, + { + "epoch": 0.75, + "learning_rate": 0.00020339418254764296, + "loss": 1.2888, + "theoretical_loss": 3.4002195047977253, + "tokens_seen": 2470182912 + }, + { + "epoch": 0.75, + "learning_rate": 0.00020337813440320966, + "loss": 1.2354, + "theoretical_loss": 3.400212352229221, + "tokens_seen": 2470248448 + }, + { + "epoch": 0.75, + "learning_rate": 0.00020336208625877636, + "loss": 1.3189, + "theoretical_loss": 3.4002051999036036, + "tokens_seen": 2470313984 + }, + { + "epoch": 0.75, + "learning_rate": 0.00020334603811434306, + "loss": 1.2411, + "theoretical_loss": 3.400198047820859, + "tokens_seen": 2470379520 + }, + { + "epoch": 0.75, + "learning_rate": 0.00020332998996990976, + "loss": 1.278, + "theoretical_loss": 3.400190895980971, + "tokens_seen": 2470445056 + }, + { + "epoch": 0.75, + "learning_rate": 0.00020331394182547646, + "loss": 1.1598, + "theoretical_loss": 3.4001837443839262, + "tokens_seen": 2470510592 + }, + { + "epoch": 0.75, + "learning_rate": 0.00020329789368104314, + "loss": 1.2773, + "theoretical_loss": 3.400176593029709, + "tokens_seen": 2470576128 + }, + { + "epoch": 0.75, + "learning_rate": 0.00020328184553660984, + "loss": 1.2857, + "theoretical_loss": 3.4001694419183055, + "tokens_seen": 2470641664 + }, + { + "epoch": 0.75, + "objective/train/docs_used": 899059, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.306131362915039, + "objective/train/theoretical_loss": 3.4001622910497007, + "objective/train/tokens_used": 2491167200, + "theoretical_loss": 3.4001622910497007, + "tokens_seen": 2470707200 + }, + { + "epoch": 0.75, + "learning_rate": 0.00020326579739217654, + "loss": 1.2722, + "theoretical_loss": 3.4001622910497007, + "tokens_seen": 2470707200 + }, + { + "epoch": 0.75, + "learning_rate": 0.00020324974924774324, + "loss": 1.1852, + "theoretical_loss": 3.4001551404238795, + "tokens_seen": 2470772736 + }, + { + "epoch": 0.75, + "learning_rate": 0.00020323370110330994, + "loss": 1.2477, + "theoretical_loss": 3.400147990040828, + "tokens_seen": 2470838272 + }, + { + "epoch": 0.75, + "learning_rate": 0.00020321765295887664, + "loss": 1.3114, + "theoretical_loss": 3.400140839900531, + "tokens_seen": 2470903808 + }, + { + "epoch": 0.75, + "learning_rate": 0.00020320160481444334, + "loss": 1.3323, + "theoretical_loss": 3.400133690002974, + "tokens_seen": 2470969344 + }, + { + "epoch": 0.75, + "learning_rate": 0.00020318555667001004, + "loss": 1.2129, + "theoretical_loss": 3.4001265403481424, + "tokens_seen": 2471034880 + }, + { + "epoch": 0.75, + "learning_rate": 0.00020316950852557674, + "loss": 1.2898, + "theoretical_loss": 3.400119390936021, + "tokens_seen": 2471100416 + }, + { + "epoch": 0.75, + "learning_rate": 0.00020315346038114344, + "loss": 1.1754, + "theoretical_loss": 3.4001122417665957, + "tokens_seen": 2471165952 + }, + { + "epoch": 0.75, + "learning_rate": 0.00020313741223671014, + "loss": 1.3026, + "theoretical_loss": 3.4001050928398517, + "tokens_seen": 2471231488 + }, + { + "epoch": 0.75, + "learning_rate": 0.00020312136409227684, + "loss": 1.3835, + "theoretical_loss": 3.400097944155774, + "tokens_seen": 2471297024 + }, + { + "epoch": 0.75, + "learning_rate": 0.00020310531594784355, + "loss": 1.3244, + "theoretical_loss": 3.4000907957143487, + "tokens_seen": 2471362560 + }, + { + "epoch": 0.75, + "learning_rate": 0.00020308926780341025, + "loss": 1.269, + "theoretical_loss": 3.4000836475155602, + "tokens_seen": 2471428096 + }, + { + "epoch": 0.75, + "learning_rate": 0.00020307321965897695, + "loss": 1.2348, + "theoretical_loss": 3.4000764995593946, + "tokens_seen": 2471493632 + }, + { + "epoch": 0.75, + "learning_rate": 0.00020305717151454365, + "loss": 1.2389, + "theoretical_loss": 3.4000693518458367, + "tokens_seen": 2471559168 + }, + { + "epoch": 0.75, + "learning_rate": 0.00020304112337011035, + "loss": 1.3483, + "theoretical_loss": 3.400062204374872, + "tokens_seen": 2471624704 + }, + { + "epoch": 0.75, + "learning_rate": 0.00020302507522567705, + "loss": 1.3879, + "theoretical_loss": 3.400055057146486, + "tokens_seen": 2471690240 + }, + { + "epoch": 0.75, + "learning_rate": 0.00020300902708124375, + "loss": 1.3128, + "theoretical_loss": 3.4000479101606635, + "tokens_seen": 2471755776 + }, + { + "epoch": 0.75, + "learning_rate": 0.00020299297893681045, + "loss": 1.2306, + "theoretical_loss": 3.4000407634173904, + "tokens_seen": 2471821312 + }, + { + "epoch": 0.75, + "learning_rate": 0.00020297693079237715, + "loss": 1.3258, + "theoretical_loss": 3.400033616916652, + "tokens_seen": 2471886848 + }, + { + "epoch": 0.75, + "learning_rate": 0.00020296088264794385, + "loss": 1.2855, + "theoretical_loss": 3.4000264706584336, + "tokens_seen": 2471952384 + }, + { + "epoch": 0.75, + "learning_rate": 0.00020294483450351055, + "loss": 1.3044, + "theoretical_loss": 3.40001932464272, + "tokens_seen": 2472017920 + }, + { + "epoch": 0.75, + "learning_rate": 0.00020292878635907725, + "loss": 1.3123, + "theoretical_loss": 3.4000121788694977, + "tokens_seen": 2472083456 + }, + { + "epoch": 0.75, + "learning_rate": 0.00020291273821464393, + "loss": 1.239, + "theoretical_loss": 3.4000050333387506, + "tokens_seen": 2472148992 + }, + { + "epoch": 0.75, + "learning_rate": 0.00020289669007021063, + "loss": 1.1867, + "theoretical_loss": 3.399997888050465, + "tokens_seen": 2472214528 + }, + { + "epoch": 0.75, + "learning_rate": 0.00020288064192577733, + "loss": 1.3448, + "theoretical_loss": 3.3999907430046266, + "tokens_seen": 2472280064 + }, + { + "epoch": 0.75, + "learning_rate": 0.00020286459378134403, + "loss": 1.2887, + "theoretical_loss": 3.3999835982012194, + "tokens_seen": 2472345600 + }, + { + "epoch": 0.75, + "learning_rate": 0.00020284854563691073, + "loss": 1.2758, + "theoretical_loss": 3.39997645364023, + "tokens_seen": 2472411136 + }, + { + "epoch": 0.75, + "learning_rate": 0.00020283249749247743, + "loss": 1.2381, + "theoretical_loss": 3.399969309321643, + "tokens_seen": 2472476672 + }, + { + "epoch": 0.75, + "learning_rate": 0.00020281644934804413, + "loss": 1.256, + "theoretical_loss": 3.399962165245444, + "tokens_seen": 2472542208 + }, + { + "epoch": 0.75, + "learning_rate": 0.00020280040120361083, + "loss": 1.321, + "theoretical_loss": 3.3999550214116185, + "tokens_seen": 2472607744 + }, + { + "epoch": 0.75, + "learning_rate": 0.00020278435305917753, + "loss": 1.3556, + "theoretical_loss": 3.3999478778201517, + "tokens_seen": 2472673280 + }, + { + "epoch": 0.75, + "learning_rate": 0.00020276830491474423, + "loss": 1.191, + "theoretical_loss": 3.399940734471029, + "tokens_seen": 2472738816 + }, + { + "epoch": 0.75, + "learning_rate": 0.00020275225677031093, + "loss": 1.2195, + "theoretical_loss": 3.3999335913642357, + "tokens_seen": 2472804352 + }, + { + "epoch": 0.75, + "learning_rate": 0.00020273620862587763, + "loss": 1.1669, + "theoretical_loss": 3.399926448499757, + "tokens_seen": 2472869888 + }, + { + "epoch": 0.75, + "learning_rate": 0.00020272016048144433, + "loss": 1.2576, + "theoretical_loss": 3.3999193058775785, + "tokens_seen": 2472935424 + }, + { + "epoch": 0.75, + "learning_rate": 0.00020270411233701103, + "loss": 1.1836, + "theoretical_loss": 3.3999121634976857, + "tokens_seen": 2473000960 + }, + { + "epoch": 0.75, + "learning_rate": 0.00020268806419257773, + "loss": 1.2671, + "theoretical_loss": 3.3999050213600635, + "tokens_seen": 2473066496 + }, + { + "epoch": 0.75, + "learning_rate": 0.00020267201604814444, + "loss": 1.3271, + "theoretical_loss": 3.3998978794646977, + "tokens_seen": 2473132032 + }, + { + "epoch": 0.75, + "learning_rate": 0.00020265596790371114, + "loss": 1.318, + "theoretical_loss": 3.3998907378115737, + "tokens_seen": 2473197568 + }, + { + "epoch": 0.75, + "learning_rate": 0.00020263991975927784, + "loss": 1.2727, + "theoretical_loss": 3.3998835964006764, + "tokens_seen": 2473263104 + }, + { + "epoch": 0.75, + "learning_rate": 0.00020262387161484454, + "loss": 1.3917, + "theoretical_loss": 3.3998764552319916, + "tokens_seen": 2473328640 + }, + { + "epoch": 0.75, + "learning_rate": 0.00020260782347041124, + "loss": 1.3008, + "theoretical_loss": 3.399869314305504, + "tokens_seen": 2473394176 + }, + { + "epoch": 0.75, + "learning_rate": 0.00020259177532597794, + "loss": 1.4052, + "theoretical_loss": 3.3998621736212002, + "tokens_seen": 2473459712 + }, + { + "epoch": 0.75, + "learning_rate": 0.00020257572718154464, + "loss": 1.2338, + "theoretical_loss": 3.3998550331790645, + "tokens_seen": 2473525248 + }, + { + "epoch": 0.75, + "learning_rate": 0.00020255967903711134, + "loss": 1.3577, + "theoretical_loss": 3.3998478929790825, + "tokens_seen": 2473590784 + }, + { + "epoch": 0.75, + "learning_rate": 0.00020254363089267804, + "loss": 1.241, + "theoretical_loss": 3.39984075302124, + "tokens_seen": 2473656320 + }, + { + "epoch": 0.75, + "learning_rate": 0.00020252758274824471, + "loss": 1.2951, + "theoretical_loss": 3.3998336133055216, + "tokens_seen": 2473721856 + }, + { + "epoch": 0.75, + "learning_rate": 0.00020251153460381141, + "loss": 1.1873, + "theoretical_loss": 3.3998264738319133, + "tokens_seen": 2473787392 + }, + { + "epoch": 0.75, + "learning_rate": 0.00020249548645937812, + "loss": 1.33, + "theoretical_loss": 3.3998193346004006, + "tokens_seen": 2473852928 + }, + { + "epoch": 0.75, + "learning_rate": 0.00020247943831494482, + "loss": 1.2692, + "theoretical_loss": 3.399812195610968, + "tokens_seen": 2473918464 + }, + { + "epoch": 0.75, + "objective/train/docs_used": 900350, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.2343543767929077, + "objective/train/theoretical_loss": 3.399805056863602, + "objective/train/tokens_used": 2494444000, + "theoretical_loss": 3.399805056863602, + "tokens_seen": 2473984000 + }, + { + "epoch": 0.75, + "learning_rate": 0.00020246339017051152, + "loss": 1.2828, + "theoretical_loss": 3.399805056863602, + "tokens_seen": 2473984000 + }, + { + "epoch": 0.75, + "learning_rate": 0.00020244734202607822, + "loss": 1.4353, + "theoretical_loss": 3.399797918358287, + "tokens_seen": 2474049536 + }, + { + "epoch": 0.75, + "learning_rate": 0.00020243129388164492, + "loss": 1.3358, + "theoretical_loss": 3.399790780095009, + "tokens_seen": 2474115072 + }, + { + "epoch": 0.75, + "learning_rate": 0.00020241524573721162, + "loss": 1.2052, + "theoretical_loss": 3.399783642073753, + "tokens_seen": 2474180608 + }, + { + "epoch": 0.75, + "learning_rate": 0.00020239919759277837, + "loss": 1.4282, + "theoretical_loss": 3.399776504294505, + "tokens_seen": 2474246144 + }, + { + "epoch": 0.75, + "learning_rate": 0.00020238314944834507, + "loss": 1.2415, + "theoretical_loss": 3.39976936675725, + "tokens_seen": 2474311680 + }, + { + "epoch": 0.75, + "learning_rate": 0.00020236710130391177, + "loss": 1.2827, + "theoretical_loss": 3.399762229461973, + "tokens_seen": 2474377216 + }, + { + "epoch": 0.75, + "learning_rate": 0.00020235105315947848, + "loss": 1.2485, + "theoretical_loss": 3.3997550924086597, + "tokens_seen": 2474442752 + }, + { + "epoch": 0.75, + "learning_rate": 0.00020233500501504518, + "loss": 1.2255, + "theoretical_loss": 3.3997479555972956, + "tokens_seen": 2474508288 + }, + { + "epoch": 0.75, + "learning_rate": 0.00020231895687061188, + "loss": 1.2794, + "theoretical_loss": 3.3997408190278664, + "tokens_seen": 2474573824 + }, + { + "epoch": 0.75, + "learning_rate": 0.00020230290872617858, + "loss": 1.3893, + "theoretical_loss": 3.399733682700357, + "tokens_seen": 2474639360 + }, + { + "epoch": 0.75, + "learning_rate": 0.00020228686058174525, + "loss": 1.3302, + "theoretical_loss": 3.3997265466147524, + "tokens_seen": 2474704896 + }, + { + "epoch": 0.75, + "learning_rate": 0.00020227081243731195, + "loss": 1.2667, + "theoretical_loss": 3.399719410771039, + "tokens_seen": 2474770432 + }, + { + "epoch": 0.75, + "learning_rate": 0.00020225476429287865, + "loss": 1.2401, + "theoretical_loss": 3.3997122751692013, + "tokens_seen": 2474835968 + }, + { + "epoch": 0.75, + "learning_rate": 0.00020223871614844535, + "loss": 1.2591, + "theoretical_loss": 3.3997051398092255, + "tokens_seen": 2474901504 + }, + { + "epoch": 0.75, + "learning_rate": 0.00020222266800401205, + "loss": 1.3404, + "theoretical_loss": 3.3996980046910963, + "tokens_seen": 2474967040 + }, + { + "epoch": 0.75, + "learning_rate": 0.00020220661985957875, + "loss": 1.3009, + "theoretical_loss": 3.3996908698147994, + "tokens_seen": 2475032576 + }, + { + "epoch": 0.75, + "learning_rate": 0.00020219057171514545, + "loss": 1.3788, + "theoretical_loss": 3.3996837351803206, + "tokens_seen": 2475098112 + }, + { + "epoch": 0.75, + "learning_rate": 0.00020217452357071215, + "loss": 1.297, + "theoretical_loss": 3.3996766007876444, + "tokens_seen": 2475163648 + }, + { + "epoch": 0.75, + "learning_rate": 0.00020215847542627886, + "loss": 1.369, + "theoretical_loss": 3.399669466636757, + "tokens_seen": 2475229184 + }, + { + "epoch": 0.75, + "learning_rate": 0.00020214242728184556, + "loss": 1.1349, + "theoretical_loss": 3.3996623327276434, + "tokens_seen": 2475294720 + }, + { + "epoch": 0.75, + "learning_rate": 0.00020212637913741226, + "loss": 1.3828, + "theoretical_loss": 3.3996551990602892, + "tokens_seen": 2475360256 + }, + { + "epoch": 0.75, + "learning_rate": 0.00020211033099297896, + "loss": 1.3745, + "theoretical_loss": 3.3996480656346795, + "tokens_seen": 2475425792 + }, + { + "epoch": 0.75, + "learning_rate": 0.00020209428284854566, + "loss": 1.2699, + "theoretical_loss": 3.3996409324508, + "tokens_seen": 2475491328 + }, + { + "epoch": 0.75, + "learning_rate": 0.00020207823470411236, + "loss": 1.2276, + "theoretical_loss": 3.399633799508636, + "tokens_seen": 2475556864 + }, + { + "epoch": 0.75, + "learning_rate": 0.00020206218655967906, + "loss": 1.3369, + "theoretical_loss": 3.399626666808173, + "tokens_seen": 2475622400 + }, + { + "epoch": 0.75, + "learning_rate": 0.00020204613841524576, + "loss": 1.3476, + "theoretical_loss": 3.3996195343493962, + "tokens_seen": 2475687936 + }, + { + "epoch": 0.75, + "learning_rate": 0.00020203009027081246, + "loss": 1.1892, + "theoretical_loss": 3.3996124021322918, + "tokens_seen": 2475753472 + }, + { + "epoch": 0.75, + "learning_rate": 0.00020201404212637916, + "loss": 1.1953, + "theoretical_loss": 3.3996052701568438, + "tokens_seen": 2475819008 + }, + { + "epoch": 0.75, + "learning_rate": 0.00020199799398194586, + "loss": 1.356, + "theoretical_loss": 3.399598138423039, + "tokens_seen": 2475884544 + }, + { + "epoch": 0.75, + "learning_rate": 0.00020198194583751256, + "loss": 1.2395, + "theoretical_loss": 3.3995910069308617, + "tokens_seen": 2475950080 + }, + { + "epoch": 0.75, + "learning_rate": 0.00020196589769307926, + "loss": 1.3086, + "theoretical_loss": 3.399583875680298, + "tokens_seen": 2476015616 + }, + { + "epoch": 0.75, + "learning_rate": 0.00020194984954864596, + "loss": 1.1928, + "theoretical_loss": 3.399576744671333, + "tokens_seen": 2476081152 + }, + { + "epoch": 0.75, + "learning_rate": 0.00020193380140421266, + "loss": 1.3425, + "theoretical_loss": 3.3995696139039526, + "tokens_seen": 2476146688 + }, + { + "epoch": 0.75, + "learning_rate": 0.00020191775325977937, + "loss": 1.269, + "theoretical_loss": 3.3995624833781415, + "tokens_seen": 2476212224 + }, + { + "epoch": 0.75, + "learning_rate": 0.00020190170511534604, + "loss": 1.3469, + "theoretical_loss": 3.399555353093886, + "tokens_seen": 2476277760 + }, + { + "epoch": 0.75, + "learning_rate": 0.00020188565697091274, + "loss": 1.2706, + "theoretical_loss": 3.3995482230511707, + "tokens_seen": 2476343296 + }, + { + "epoch": 0.75, + "learning_rate": 0.00020186960882647944, + "loss": 1.2989, + "theoretical_loss": 3.399541093249981, + "tokens_seen": 2476408832 + }, + { + "epoch": 0.75, + "learning_rate": 0.00020185356068204614, + "loss": 1.2669, + "theoretical_loss": 3.3995339636903035, + "tokens_seen": 2476474368 + }, + { + "epoch": 0.75, + "learning_rate": 0.00020183751253761284, + "loss": 1.2971, + "theoretical_loss": 3.3995268343721223, + "tokens_seen": 2476539904 + }, + { + "epoch": 0.75, + "learning_rate": 0.00020182146439317954, + "loss": 1.122, + "theoretical_loss": 3.3995197052954236, + "tokens_seen": 2476605440 + }, + { + "epoch": 0.75, + "learning_rate": 0.00020180541624874624, + "loss": 1.323, + "theoretical_loss": 3.3995125764601926, + "tokens_seen": 2476670976 + }, + { + "epoch": 0.75, + "learning_rate": 0.00020178936810431294, + "loss": 1.293, + "theoretical_loss": 3.3995054478664146, + "tokens_seen": 2476736512 + }, + { + "epoch": 0.75, + "learning_rate": 0.00020177331995987964, + "loss": 1.2445, + "theoretical_loss": 3.399498319514075, + "tokens_seen": 2476802048 + }, + { + "epoch": 0.75, + "learning_rate": 0.00020175727181544634, + "loss": 1.2024, + "theoretical_loss": 3.39949119140316, + "tokens_seen": 2476867584 + }, + { + "epoch": 0.75, + "learning_rate": 0.00020174122367101304, + "loss": 1.3097, + "theoretical_loss": 3.3994840635336536, + "tokens_seen": 2476933120 + }, + { + "epoch": 0.75, + "learning_rate": 0.00020172517552657975, + "loss": 1.2776, + "theoretical_loss": 3.3994769359055423, + "tokens_seen": 2476998656 + }, + { + "epoch": 0.75, + "learning_rate": 0.00020170912738214645, + "loss": 1.1961, + "theoretical_loss": 3.3994698085188118, + "tokens_seen": 2477064192 + }, + { + "epoch": 0.75, + "learning_rate": 0.00020169307923771315, + "loss": 1.3089, + "theoretical_loss": 3.3994626813734463, + "tokens_seen": 2477129728 + }, + { + "epoch": 0.75, + "learning_rate": 0.00020167703109327985, + "loss": 1.2995, + "theoretical_loss": 3.3994555544694323, + "tokens_seen": 2477195264 + }, + { + "epoch": 0.75, + "objective/train/docs_used": 901501, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.2307312488555908, + "objective/train/theoretical_loss": 3.399448427806755, + "objective/train/tokens_used": 2497720800, + "theoretical_loss": 3.399448427806755, + "tokens_seen": 2477260800 + }, + { + "epoch": 0.75, + "learning_rate": 0.00020166098294884655, + "loss": 1.2348, + "theoretical_loss": 3.399448427806755, + "tokens_seen": 2477260800 + }, + { + "epoch": 0.75, + "learning_rate": 0.00020164493480441325, + "loss": 1.2022, + "theoretical_loss": 3.3994413013853997, + "tokens_seen": 2477326336 + }, + { + "epoch": 0.75, + "learning_rate": 0.00020162888665997995, + "loss": 1.3412, + "theoretical_loss": 3.399434175205352, + "tokens_seen": 2477391872 + }, + { + "epoch": 0.75, + "learning_rate": 0.00020161283851554665, + "loss": 1.3402, + "theoretical_loss": 3.3994270492665972, + "tokens_seen": 2477457408 + }, + { + "epoch": 0.75, + "learning_rate": 0.00020159679037111335, + "loss": 1.3091, + "theoretical_loss": 3.3994199235691203, + "tokens_seen": 2477522944 + }, + { + "epoch": 0.75, + "learning_rate": 0.00020158074222668005, + "loss": 1.2695, + "theoretical_loss": 3.399412798112908, + "tokens_seen": 2477588480 + }, + { + "epoch": 0.75, + "learning_rate": 0.00020156469408224675, + "loss": 1.168, + "theoretical_loss": 3.3994056728979443, + "tokens_seen": 2477654016 + }, + { + "epoch": 0.75, + "learning_rate": 0.00020154864593781345, + "loss": 1.2499, + "theoretical_loss": 3.399398547924216, + "tokens_seen": 2477719552 + }, + { + "epoch": 0.75, + "learning_rate": 0.00020153259779338015, + "loss": 1.2913, + "theoretical_loss": 3.3993914231917075, + "tokens_seen": 2477785088 + }, + { + "epoch": 0.75, + "learning_rate": 0.00020151654964894683, + "loss": 1.3822, + "theoretical_loss": 3.399384298700405, + "tokens_seen": 2477850624 + }, + { + "epoch": 0.75, + "learning_rate": 0.00020150050150451353, + "loss": 1.2492, + "theoretical_loss": 3.3993771744502927, + "tokens_seen": 2477916160 + }, + { + "epoch": 0.75, + "learning_rate": 0.00020148445336008023, + "loss": 1.3584, + "theoretical_loss": 3.399370050441358, + "tokens_seen": 2477981696 + }, + { + "epoch": 0.75, + "learning_rate": 0.00020146840521564693, + "loss": 1.2198, + "theoretical_loss": 3.399362926673585, + "tokens_seen": 2478047232 + }, + { + "epoch": 0.75, + "learning_rate": 0.00020145235707121363, + "loss": 1.3276, + "theoretical_loss": 3.3993558031469595, + "tokens_seen": 2478112768 + }, + { + "epoch": 0.75, + "learning_rate": 0.00020143630892678033, + "loss": 1.3714, + "theoretical_loss": 3.3993486798614665, + "tokens_seen": 2478178304 + }, + { + "epoch": 0.75, + "learning_rate": 0.00020142026078234703, + "loss": 1.1954, + "theoretical_loss": 3.3993415568170926, + "tokens_seen": 2478243840 + }, + { + "epoch": 0.75, + "learning_rate": 0.00020140421263791373, + "loss": 1.2465, + "theoretical_loss": 3.399334434013822, + "tokens_seen": 2478309376 + }, + { + "epoch": 0.75, + "learning_rate": 0.00020138816449348043, + "loss": 1.3448, + "theoretical_loss": 3.399327311451641, + "tokens_seen": 2478374912 + }, + { + "epoch": 0.75, + "learning_rate": 0.00020137211634904713, + "loss": 1.3528, + "theoretical_loss": 3.3993201891305347, + "tokens_seen": 2478440448 + }, + { + "epoch": 0.75, + "learning_rate": 0.00020135606820461383, + "loss": 1.294, + "theoretical_loss": 3.3993130670504885, + "tokens_seen": 2478505984 + }, + { + "epoch": 0.75, + "learning_rate": 0.00020134002006018053, + "loss": 1.4132, + "theoretical_loss": 3.399305945211488, + "tokens_seen": 2478571520 + }, + { + "epoch": 0.75, + "learning_rate": 0.00020132397191574723, + "loss": 1.2868, + "theoretical_loss": 3.399298823613519, + "tokens_seen": 2478637056 + }, + { + "epoch": 0.75, + "learning_rate": 0.00020130792377131393, + "loss": 1.3484, + "theoretical_loss": 3.3992917022565665, + "tokens_seen": 2478702592 + }, + { + "epoch": 0.75, + "learning_rate": 0.00020129187562688064, + "loss": 1.257, + "theoretical_loss": 3.3992845811406163, + "tokens_seen": 2478768128 + }, + { + "epoch": 0.75, + "learning_rate": 0.00020127582748244734, + "loss": 1.2783, + "theoretical_loss": 3.3992774602656537, + "tokens_seen": 2478833664 + }, + { + "epoch": 0.75, + "learning_rate": 0.00020125977933801404, + "loss": 1.3047, + "theoretical_loss": 3.399270339631664, + "tokens_seen": 2478899200 + }, + { + "epoch": 0.75, + "learning_rate": 0.00020124373119358074, + "loss": 1.4229, + "theoretical_loss": 3.399263219238633, + "tokens_seen": 2478964736 + }, + { + "epoch": 0.75, + "learning_rate": 0.00020122768304914744, + "loss": 1.3243, + "theoretical_loss": 3.399256099086546, + "tokens_seen": 2479030272 + }, + { + "epoch": 0.75, + "learning_rate": 0.00020121163490471414, + "loss": 1.1807, + "theoretical_loss": 3.3992489791753884, + "tokens_seen": 2479095808 + }, + { + "epoch": 0.75, + "learning_rate": 0.00020119558676028087, + "loss": 1.2958, + "theoretical_loss": 3.3992418595051457, + "tokens_seen": 2479161344 + }, + { + "epoch": 0.75, + "learning_rate": 0.00020117953861584757, + "loss": 1.2835, + "theoretical_loss": 3.3992347400758036, + "tokens_seen": 2479226880 + }, + { + "epoch": 0.75, + "learning_rate": 0.00020116349047141427, + "loss": 1.2478, + "theoretical_loss": 3.3992276208873475, + "tokens_seen": 2479292416 + }, + { + "epoch": 0.75, + "learning_rate": 0.00020114744232698097, + "loss": 1.4144, + "theoretical_loss": 3.3992205019397628, + "tokens_seen": 2479357952 + }, + { + "epoch": 0.75, + "learning_rate": 0.00020113139418254767, + "loss": 1.2279, + "theoretical_loss": 3.399213383233035, + "tokens_seen": 2479423488 + }, + { + "epoch": 0.75, + "learning_rate": 0.00020111534603811437, + "loss": 1.2697, + "theoretical_loss": 3.3992062647671495, + "tokens_seen": 2479489024 + }, + { + "epoch": 0.75, + "learning_rate": 0.00020109929789368107, + "loss": 1.2708, + "theoretical_loss": 3.399199146542092, + "tokens_seen": 2479554560 + }, + { + "epoch": 0.75, + "learning_rate": 0.00020108324974924777, + "loss": 1.4197, + "theoretical_loss": 3.399192028557848, + "tokens_seen": 2479620096 + }, + { + "epoch": 0.75, + "learning_rate": 0.00020106720160481447, + "loss": 1.2248, + "theoretical_loss": 3.3991849108144025, + "tokens_seen": 2479685632 + }, + { + "epoch": 0.75, + "learning_rate": 0.00020105115346038117, + "loss": 1.417, + "theoretical_loss": 3.3991777933117415, + "tokens_seen": 2479751168 + }, + { + "epoch": 0.75, + "learning_rate": 0.00020103510531594787, + "loss": 1.2983, + "theoretical_loss": 3.3991706760498506, + "tokens_seen": 2479816704 + }, + { + "epoch": 0.75, + "learning_rate": 0.00020101905717151457, + "loss": 1.3002, + "theoretical_loss": 3.3991635590287146, + "tokens_seen": 2479882240 + }, + { + "epoch": 0.75, + "learning_rate": 0.00020100300902708127, + "loss": 1.2797, + "theoretical_loss": 3.3991564422483203, + "tokens_seen": 2479947776 + }, + { + "epoch": 0.75, + "learning_rate": 0.00020098696088264797, + "loss": 1.2795, + "theoretical_loss": 3.399149325708651, + "tokens_seen": 2480013312 + }, + { + "epoch": 0.75, + "learning_rate": 0.00020097091273821468, + "loss": 1.3747, + "theoretical_loss": 3.3991422094096944, + "tokens_seen": 2480078848 + }, + { + "epoch": 0.75, + "learning_rate": 0.00020095486459378138, + "loss": 1.3744, + "theoretical_loss": 3.3991350933514353, + "tokens_seen": 2480144384 + }, + { + "epoch": 0.75, + "learning_rate": 0.00020093881644934808, + "loss": 1.3869, + "theoretical_loss": 3.3991279775338583, + "tokens_seen": 2480209920 + }, + { + "epoch": 0.75, + "learning_rate": 0.00020092276830491478, + "loss": 1.3496, + "theoretical_loss": 3.39912086195695, + "tokens_seen": 2480275456 + }, + { + "epoch": 0.75, + "learning_rate": 0.00020090672016048148, + "loss": 1.3007, + "theoretical_loss": 3.3991137466206958, + "tokens_seen": 2480340992 + }, + { + "epoch": 0.75, + "learning_rate": 0.00020089067201604815, + "loss": 1.3979, + "theoretical_loss": 3.3991066315250804, + "tokens_seen": 2480406528 + }, + { + "epoch": 0.75, + "learning_rate": 0.00020087462387161485, + "loss": 1.3008, + "theoretical_loss": 3.3990995166700904, + "tokens_seen": 2480472064 + }, + { + "epoch": 0.75, + "objective/train/docs_used": 902720, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.3832130432128906, + "objective/train/theoretical_loss": 3.39909240205571, + "objective/train/tokens_used": 2500997600, + "theoretical_loss": 3.39909240205571, + "tokens_seen": 2480537600 + }, + { + "epoch": 0.75, + "learning_rate": 0.00020085857572718155, + "loss": 1.3229, + "theoretical_loss": 3.39909240205571, + "tokens_seen": 2480537600 + }, + { + "epoch": 0.75, + "learning_rate": 0.00020084252758274825, + "loss": 1.3183, + "theoretical_loss": 3.3990852876819257, + "tokens_seen": 2480603136 + }, + { + "epoch": 0.75, + "learning_rate": 0.00020082647943831495, + "loss": 1.268, + "theoretical_loss": 3.399078173548723, + "tokens_seen": 2480668672 + }, + { + "epoch": 0.75, + "learning_rate": 0.00020081043129388165, + "loss": 1.2841, + "theoretical_loss": 3.399071059656087, + "tokens_seen": 2480734208 + }, + { + "epoch": 0.75, + "learning_rate": 0.00020079438314944836, + "loss": 1.2874, + "theoretical_loss": 3.3990639460040035, + "tokens_seen": 2480799744 + }, + { + "epoch": 0.75, + "learning_rate": 0.00020077833500501506, + "loss": 1.3061, + "theoretical_loss": 3.3990568325924575, + "tokens_seen": 2480865280 + }, + { + "epoch": 0.75, + "learning_rate": 0.00020076228686058176, + "loss": 1.3916, + "theoretical_loss": 3.399049719421435, + "tokens_seen": 2480930816 + }, + { + "epoch": 0.75, + "learning_rate": 0.00020074623871614846, + "loss": 1.222, + "theoretical_loss": 3.399042606490922, + "tokens_seen": 2480996352 + }, + { + "epoch": 0.75, + "learning_rate": 0.00020073019057171516, + "loss": 1.1664, + "theoretical_loss": 3.399035493800903, + "tokens_seen": 2481061888 + }, + { + "epoch": 0.75, + "learning_rate": 0.00020071414242728186, + "loss": 1.3716, + "theoretical_loss": 3.3990283813513633, + "tokens_seen": 2481127424 + }, + { + "epoch": 0.75, + "learning_rate": 0.00020069809428284856, + "loss": 1.3786, + "theoretical_loss": 3.39902126914229, + "tokens_seen": 2481192960 + }, + { + "epoch": 0.75, + "learning_rate": 0.00020068204613841526, + "loss": 1.2563, + "theoretical_loss": 3.399014157173667, + "tokens_seen": 2481258496 + }, + { + "epoch": 0.75, + "learning_rate": 0.00020066599799398196, + "loss": 1.3077, + "theoretical_loss": 3.3990070454454813, + "tokens_seen": 2481324032 + }, + { + "epoch": 0.75, + "learning_rate": 0.00020064994984954866, + "loss": 1.273, + "theoretical_loss": 3.398999933957717, + "tokens_seen": 2481389568 + }, + { + "epoch": 0.75, + "learning_rate": 0.00020063390170511536, + "loss": 1.2568, + "theoretical_loss": 3.39899282271036, + "tokens_seen": 2481455104 + }, + { + "epoch": 0.75, + "learning_rate": 0.00020061785356068206, + "loss": 1.2382, + "theoretical_loss": 3.3989857117033964, + "tokens_seen": 2481520640 + }, + { + "epoch": 0.75, + "learning_rate": 0.00020060180541624876, + "loss": 1.4405, + "theoretical_loss": 3.3989786009368115, + "tokens_seen": 2481586176 + }, + { + "epoch": 0.75, + "learning_rate": 0.00020058575727181546, + "loss": 1.2866, + "theoretical_loss": 3.3989714904105908, + "tokens_seen": 2481651712 + }, + { + "epoch": 0.75, + "learning_rate": 0.00020056970912738216, + "loss": 1.3369, + "theoretical_loss": 3.3989643801247196, + "tokens_seen": 2481717248 + }, + { + "epoch": 0.75, + "learning_rate": 0.00020055366098294886, + "loss": 1.3716, + "theoretical_loss": 3.3989572700791832, + "tokens_seen": 2481782784 + }, + { + "epoch": 0.75, + "learning_rate": 0.00020053761283851557, + "loss": 1.1997, + "theoretical_loss": 3.3989501602739676, + "tokens_seen": 2481848320 + }, + { + "epoch": 0.75, + "learning_rate": 0.00020052156469408227, + "loss": 1.2807, + "theoretical_loss": 3.3989430507090588, + "tokens_seen": 2481913856 + }, + { + "epoch": 0.75, + "learning_rate": 0.00020050551654964894, + "loss": 1.3062, + "theoretical_loss": 3.3989359413844413, + "tokens_seen": 2481979392 + }, + { + "epoch": 0.75, + "learning_rate": 0.00020048946840521564, + "loss": 1.1731, + "theoretical_loss": 3.398928832300101, + "tokens_seen": 2482044928 + }, + { + "epoch": 0.75, + "learning_rate": 0.00020047342026078234, + "loss": 1.3288, + "theoretical_loss": 3.398921723456024, + "tokens_seen": 2482110464 + }, + { + "epoch": 0.75, + "learning_rate": 0.00020045737211634904, + "loss": 1.3876, + "theoretical_loss": 3.398914614852195, + "tokens_seen": 2482176000 + }, + { + "epoch": 0.75, + "learning_rate": 0.00020044132397191574, + "loss": 1.1657, + "theoretical_loss": 3.3989075064886, + "tokens_seen": 2482241536 + }, + { + "epoch": 0.75, + "learning_rate": 0.00020042527582748244, + "loss": 1.2907, + "theoretical_loss": 3.3989003983652246, + "tokens_seen": 2482307072 + }, + { + "epoch": 0.75, + "learning_rate": 0.00020040922768304914, + "loss": 1.2175, + "theoretical_loss": 3.3988932904820537, + "tokens_seen": 2482372608 + }, + { + "epoch": 0.75, + "learning_rate": 0.00020039317953861584, + "loss": 1.2934, + "theoretical_loss": 3.3988861828390737, + "tokens_seen": 2482438144 + }, + { + "epoch": 0.75, + "learning_rate": 0.00020037713139418254, + "loss": 1.3594, + "theoretical_loss": 3.3988790754362697, + "tokens_seen": 2482503680 + }, + { + "epoch": 0.75, + "learning_rate": 0.00020036108324974925, + "loss": 1.3541, + "theoretical_loss": 3.3988719682736273, + "tokens_seen": 2482569216 + }, + { + "epoch": 0.75, + "learning_rate": 0.00020034503510531595, + "loss": 1.3021, + "theoretical_loss": 3.3988648613511323, + "tokens_seen": 2482634752 + }, + { + "epoch": 0.75, + "learning_rate": 0.00020032898696088265, + "loss": 1.294, + "theoretical_loss": 3.3988577546687697, + "tokens_seen": 2482700288 + }, + { + "epoch": 0.75, + "learning_rate": 0.00020031293881644935, + "loss": 1.2989, + "theoretical_loss": 3.3988506482265253, + "tokens_seen": 2482765824 + }, + { + "epoch": 0.75, + "learning_rate": 0.00020029689067201605, + "loss": 1.3742, + "theoretical_loss": 3.3988435420243848, + "tokens_seen": 2482831360 + }, + { + "epoch": 0.75, + "learning_rate": 0.00020028084252758275, + "loss": 1.3126, + "theoretical_loss": 3.3988364360623335, + "tokens_seen": 2482896896 + }, + { + "epoch": 0.75, + "learning_rate": 0.00020026479438314945, + "loss": 1.2655, + "theoretical_loss": 3.3988293303403574, + "tokens_seen": 2482962432 + }, + { + "epoch": 0.75, + "learning_rate": 0.00020024874623871615, + "loss": 1.1799, + "theoretical_loss": 3.3988222248584417, + "tokens_seen": 2483027968 + }, + { + "epoch": 0.75, + "learning_rate": 0.00020023269809428285, + "loss": 1.2942, + "theoretical_loss": 3.398815119616572, + "tokens_seen": 2483093504 + }, + { + "epoch": 0.75, + "learning_rate": 0.00020021664994984955, + "loss": 1.3239, + "theoretical_loss": 3.3988080146147337, + "tokens_seen": 2483159040 + }, + { + "epoch": 0.75, + "learning_rate": 0.00020020060180541625, + "loss": 1.1865, + "theoretical_loss": 3.3988009098529126, + "tokens_seen": 2483224576 + }, + { + "epoch": 0.75, + "learning_rate": 0.00020018455366098295, + "loss": 1.2899, + "theoretical_loss": 3.398793805331094, + "tokens_seen": 2483290112 + }, + { + "epoch": 0.75, + "learning_rate": 0.00020016850551654965, + "loss": 1.3579, + "theoretical_loss": 3.3987867010492643, + "tokens_seen": 2483355648 + }, + { + "epoch": 0.75, + "learning_rate": 0.00020015245737211635, + "loss": 1.2967, + "theoretical_loss": 3.3987795970074077, + "tokens_seen": 2483421184 + }, + { + "epoch": 0.75, + "learning_rate": 0.00020013640922768305, + "loss": 1.3322, + "theoretical_loss": 3.398772493205511, + "tokens_seen": 2483486720 + }, + { + "epoch": 0.75, + "learning_rate": 0.00020012036108324973, + "loss": 1.2746, + "theoretical_loss": 3.398765389643559, + "tokens_seen": 2483552256 + }, + { + "epoch": 0.75, + "learning_rate": 0.00020010431293881643, + "loss": 1.3561, + "theoretical_loss": 3.398758286321537, + "tokens_seen": 2483617792 + }, + { + "epoch": 0.75, + "learning_rate": 0.00020008826479438313, + "loss": 1.2711, + "theoretical_loss": 3.398751183239432, + "tokens_seen": 2483683328 + }, + { + "epoch": 0.75, + "learning_rate": 0.00020007221664994983, + "loss": 1.3108, + "theoretical_loss": 3.3987440803972278, + "tokens_seen": 2483748864 + }, + { + "epoch": 0.75, + "objective/train/docs_used": 903880, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.3537970781326294, + "objective/train/theoretical_loss": 3.3987369777949112, + "objective/train/tokens_used": 2504274400, + "theoretical_loss": 3.3987369777949112, + "tokens_seen": 2483814400 + }, + { + "epoch": 0.75, + "learning_rate": 0.00020005616850551653, + "loss": 1.2819, + "theoretical_loss": 3.3987369777949112, + "tokens_seen": 2483814400 + }, + { + "epoch": 0.75, + "learning_rate": 0.00020004012036108323, + "loss": 1.3311, + "theoretical_loss": 3.3987298754324673, + "tokens_seen": 2483879936 + }, + { + "epoch": 0.75, + "learning_rate": 0.00020002407221664993, + "loss": 1.2919, + "theoretical_loss": 3.3987227733098817, + "tokens_seen": 2483945472 + }, + { + "epoch": 0.75, + "learning_rate": 0.00020000802407221663, + "loss": 1.2842, + "theoretical_loss": 3.39871567142714, + "tokens_seen": 2484011008 + }, + { + "epoch": 0.75, + "learning_rate": 0.00019999197592778336, + "loss": 1.3399, + "theoretical_loss": 3.3987085697842274, + "tokens_seen": 2484076544 + }, + { + "epoch": 0.75, + "learning_rate": 0.00019997592778335006, + "loss": 1.3438, + "theoretical_loss": 3.3987014683811303, + "tokens_seen": 2484142080 + }, + { + "epoch": 0.75, + "learning_rate": 0.00019995987963891676, + "loss": 1.2276, + "theoretical_loss": 3.398694367217834, + "tokens_seen": 2484207616 + }, + { + "epoch": 0.75, + "learning_rate": 0.00019994383149448346, + "loss": 1.2797, + "theoretical_loss": 3.3986872662943233, + "tokens_seen": 2484273152 + }, + { + "epoch": 0.75, + "learning_rate": 0.00019992778335005016, + "loss": 1.146, + "theoretical_loss": 3.3986801656105845, + "tokens_seen": 2484338688 + }, + { + "epoch": 0.75, + "learning_rate": 0.00019991173520561686, + "loss": 1.2499, + "theoretical_loss": 3.3986730651666033, + "tokens_seen": 2484404224 + }, + { + "epoch": 0.75, + "learning_rate": 0.00019989568706118356, + "loss": 1.2463, + "theoretical_loss": 3.398665964962365, + "tokens_seen": 2484469760 + }, + { + "epoch": 0.75, + "learning_rate": 0.00019987963891675026, + "loss": 1.1659, + "theoretical_loss": 3.3986588649978553, + "tokens_seen": 2484535296 + }, + { + "epoch": 0.75, + "learning_rate": 0.00019986359077231696, + "loss": 1.2568, + "theoretical_loss": 3.3986517652730597, + "tokens_seen": 2484600832 + }, + { + "epoch": 0.75, + "learning_rate": 0.00019984754262788367, + "loss": 1.3118, + "theoretical_loss": 3.3986446657879634, + "tokens_seen": 2484666368 + }, + { + "epoch": 0.75, + "learning_rate": 0.00019983149448345037, + "loss": 1.2293, + "theoretical_loss": 3.398637566542553, + "tokens_seen": 2484731904 + }, + { + "epoch": 0.75, + "learning_rate": 0.00019981544633901707, + "loss": 1.2445, + "theoretical_loss": 3.3986304675368126, + "tokens_seen": 2484797440 + }, + { + "epoch": 0.75, + "learning_rate": 0.00019979939819458374, + "loss": 1.2622, + "theoretical_loss": 3.3986233687707292, + "tokens_seen": 2484862976 + }, + { + "epoch": 0.75, + "learning_rate": 0.00019978335005015044, + "loss": 1.2834, + "theoretical_loss": 3.398616270244288, + "tokens_seen": 2484928512 + }, + { + "epoch": 0.75, + "learning_rate": 0.00019976730190571714, + "loss": 1.2931, + "theoretical_loss": 3.398609171957474, + "tokens_seen": 2484994048 + }, + { + "epoch": 0.75, + "learning_rate": 0.00019975125376128384, + "loss": 1.2603, + "theoretical_loss": 3.3986020739102734, + "tokens_seen": 2485059584 + }, + { + "epoch": 0.75, + "learning_rate": 0.00019973520561685054, + "loss": 1.4083, + "theoretical_loss": 3.3985949761026717, + "tokens_seen": 2485125120 + }, + { + "epoch": 0.75, + "learning_rate": 0.00019971915747241724, + "loss": 1.333, + "theoretical_loss": 3.398587878534654, + "tokens_seen": 2485190656 + }, + { + "epoch": 0.75, + "learning_rate": 0.00019970310932798394, + "loss": 1.2778, + "theoretical_loss": 3.3985807812062068, + "tokens_seen": 2485256192 + }, + { + "epoch": 0.75, + "learning_rate": 0.00019968706118355067, + "loss": 1.2412, + "theoretical_loss": 3.3985736841173146, + "tokens_seen": 2485321728 + }, + { + "epoch": 0.75, + "learning_rate": 0.00019967101303911737, + "loss": 1.2894, + "theoretical_loss": 3.3985665872679642, + "tokens_seen": 2485387264 + }, + { + "epoch": 0.75, + "learning_rate": 0.00019965496489468407, + "loss": 1.3484, + "theoretical_loss": 3.3985594906581404, + "tokens_seen": 2485452800 + }, + { + "epoch": 0.75, + "learning_rate": 0.00019963891675025077, + "loss": 1.2335, + "theoretical_loss": 3.3985523942878286, + "tokens_seen": 2485518336 + }, + { + "epoch": 0.75, + "learning_rate": 0.00019962286860581747, + "loss": 1.2191, + "theoretical_loss": 3.3985452981570154, + "tokens_seen": 2485583872 + }, + { + "epoch": 0.75, + "learning_rate": 0.00019960682046138418, + "loss": 1.2887, + "theoretical_loss": 3.3985382022656854, + "tokens_seen": 2485649408 + }, + { + "epoch": 0.75, + "learning_rate": 0.00019959077231695088, + "loss": 1.3345, + "theoretical_loss": 3.3985311066138246, + "tokens_seen": 2485714944 + }, + { + "epoch": 0.75, + "learning_rate": 0.00019957472417251758, + "loss": 1.2605, + "theoretical_loss": 3.398524011201419, + "tokens_seen": 2485780480 + }, + { + "epoch": 0.75, + "learning_rate": 0.00019955867602808428, + "loss": 1.3003, + "theoretical_loss": 3.3985169160284534, + "tokens_seen": 2485846016 + }, + { + "epoch": 0.75, + "learning_rate": 0.00019954262788365098, + "loss": 1.2435, + "theoretical_loss": 3.398509821094914, + "tokens_seen": 2485911552 + }, + { + "epoch": 0.75, + "learning_rate": 0.00019952657973921768, + "loss": 1.1804, + "theoretical_loss": 3.3985027264007863, + "tokens_seen": 2485977088 + }, + { + "epoch": 0.75, + "learning_rate": 0.00019951053159478438, + "loss": 1.3348, + "theoretical_loss": 3.3984956319460555, + "tokens_seen": 2486042624 + }, + { + "epoch": 0.75, + "learning_rate": 0.00019949448345035108, + "loss": 1.2432, + "theoretical_loss": 3.3984885377307075, + "tokens_seen": 2486108160 + }, + { + "epoch": 0.75, + "learning_rate": 0.00019947843530591775, + "loss": 1.4473, + "theoretical_loss": 3.3984814437547284, + "tokens_seen": 2486173696 + }, + { + "epoch": 0.75, + "learning_rate": 0.00019946238716148445, + "loss": 1.2471, + "theoretical_loss": 3.398474350018103, + "tokens_seen": 2486239232 + }, + { + "epoch": 0.75, + "learning_rate": 0.00019944633901705115, + "loss": 1.3172, + "theoretical_loss": 3.3984672565208176, + "tokens_seen": 2486304768 + }, + { + "epoch": 0.75, + "learning_rate": 0.00019943029087261786, + "loss": 1.3015, + "theoretical_loss": 3.3984601632628575, + "tokens_seen": 2486370304 + }, + { + "epoch": 0.75, + "learning_rate": 0.00019941424272818456, + "loss": 1.2881, + "theoretical_loss": 3.3984530702442077, + "tokens_seen": 2486435840 + }, + { + "epoch": 0.75, + "learning_rate": 0.00019939819458375126, + "loss": 1.3871, + "theoretical_loss": 3.3984459774648554, + "tokens_seen": 2486501376 + }, + { + "epoch": 0.75, + "learning_rate": 0.00019938214643931796, + "loss": 1.3545, + "theoretical_loss": 3.3984388849247846, + "tokens_seen": 2486566912 + }, + { + "epoch": 0.75, + "learning_rate": 0.00019936609829488466, + "loss": 1.2834, + "theoretical_loss": 3.398431792623982, + "tokens_seen": 2486632448 + }, + { + "epoch": 0.75, + "learning_rate": 0.00019935005015045136, + "loss": 1.3067, + "theoretical_loss": 3.3984247005624324, + "tokens_seen": 2486697984 + }, + { + "epoch": 0.75, + "learning_rate": 0.00019933400200601806, + "loss": 1.3116, + "theoretical_loss": 3.3984176087401217, + "tokens_seen": 2486763520 + }, + { + "epoch": 0.75, + "learning_rate": 0.00019931795386158476, + "loss": 1.2525, + "theoretical_loss": 3.398410517157036, + "tokens_seen": 2486829056 + }, + { + "epoch": 0.75, + "learning_rate": 0.00019930190571715146, + "loss": 1.3022, + "theoretical_loss": 3.398403425813161, + "tokens_seen": 2486894592 + }, + { + "epoch": 0.75, + "learning_rate": 0.00019928585757271816, + "loss": 1.2772, + "theoretical_loss": 3.3983963347084813, + "tokens_seen": 2486960128 + }, + { + "epoch": 0.75, + "learning_rate": 0.00019926980942828486, + "loss": 1.2404, + "theoretical_loss": 3.398389243842983, + "tokens_seen": 2487025664 + }, + { + "epoch": 0.75, + "objective/train/docs_used": 905256, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.322628378868103, + "objective/train/theoretical_loss": 3.398382153216652, + "objective/train/tokens_used": 2507551200, + "theoretical_loss": 3.398382153216652, + "tokens_seen": 2487091200 + }, + { + "epoch": 0.75, + "learning_rate": 0.00019925376128385156, + "loss": 1.2982, + "theoretical_loss": 3.398382153216652, + "tokens_seen": 2487091200 + }, + { + "epoch": 0.75, + "learning_rate": 0.00019923771313941826, + "loss": 1.3048, + "theoretical_loss": 3.398375062829474, + "tokens_seen": 2487156736 + }, + { + "epoch": 0.75, + "learning_rate": 0.00019922166499498496, + "loss": 1.3576, + "theoretical_loss": 3.3983679726814344, + "tokens_seen": 2487222272 + }, + { + "epoch": 0.75, + "learning_rate": 0.00019920561685055166, + "loss": 1.3336, + "theoretical_loss": 3.398360882772519, + "tokens_seen": 2487287808 + }, + { + "epoch": 0.75, + "learning_rate": 0.00019918956870611836, + "loss": 1.2991, + "theoretical_loss": 3.398353793102713, + "tokens_seen": 2487353344 + }, + { + "epoch": 0.75, + "learning_rate": 0.00019917352056168507, + "loss": 1.1459, + "theoretical_loss": 3.3983467036720025, + "tokens_seen": 2487418880 + }, + { + "epoch": 0.75, + "learning_rate": 0.00019915747241725177, + "loss": 1.4048, + "theoretical_loss": 3.398339614480373, + "tokens_seen": 2487484416 + }, + { + "epoch": 0.75, + "learning_rate": 0.00019914142427281847, + "loss": 1.3639, + "theoretical_loss": 3.3983325255278096, + "tokens_seen": 2487549952 + }, + { + "epoch": 0.75, + "learning_rate": 0.00019912537612838517, + "loss": 1.1959, + "theoretical_loss": 3.398325436814299, + "tokens_seen": 2487615488 + }, + { + "epoch": 0.75, + "learning_rate": 0.00019910932798395187, + "loss": 1.3635, + "theoretical_loss": 3.398318348339826, + "tokens_seen": 2487681024 + }, + { + "epoch": 0.75, + "learning_rate": 0.00019909327983951857, + "loss": 1.2934, + "theoretical_loss": 3.398311260104377, + "tokens_seen": 2487746560 + }, + { + "epoch": 0.75, + "learning_rate": 0.00019907723169508527, + "loss": 1.3309, + "theoretical_loss": 3.398304172107937, + "tokens_seen": 2487812096 + }, + { + "epoch": 0.75, + "learning_rate": 0.00019906118355065197, + "loss": 1.2685, + "theoretical_loss": 3.3982970843504914, + "tokens_seen": 2487877632 + }, + { + "epoch": 0.75, + "learning_rate": 0.00019904513540621867, + "loss": 1.2631, + "theoretical_loss": 3.398289996832027, + "tokens_seen": 2487943168 + }, + { + "epoch": 0.75, + "learning_rate": 0.00019902908726178537, + "loss": 1.356, + "theoretical_loss": 3.398282909552528, + "tokens_seen": 2488008704 + }, + { + "epoch": 0.75, + "learning_rate": 0.00019901303911735207, + "loss": 1.4465, + "theoretical_loss": 3.398275822511981, + "tokens_seen": 2488074240 + }, + { + "epoch": 0.75, + "learning_rate": 0.00019899699097291877, + "loss": 1.3734, + "theoretical_loss": 3.3982687357103716, + "tokens_seen": 2488139776 + }, + { + "epoch": 0.75, + "learning_rate": 0.00019898094282848547, + "loss": 1.3595, + "theoretical_loss": 3.398261649147685, + "tokens_seen": 2488205312 + }, + { + "epoch": 0.75, + "learning_rate": 0.00019896489468405217, + "loss": 1.2955, + "theoretical_loss": 3.398254562823907, + "tokens_seen": 2488270848 + }, + { + "epoch": 0.75, + "learning_rate": 0.00019894884653961887, + "loss": 1.3564, + "theoretical_loss": 3.398247476739024, + "tokens_seen": 2488336384 + }, + { + "epoch": 0.75, + "learning_rate": 0.00019893279839518557, + "loss": 1.2614, + "theoretical_loss": 3.3982403908930205, + "tokens_seen": 2488401920 + }, + { + "epoch": 0.75, + "learning_rate": 0.00019891675025075228, + "loss": 1.3027, + "theoretical_loss": 3.398233305285883, + "tokens_seen": 2488467456 + }, + { + "epoch": 0.75, + "learning_rate": 0.00019890070210631898, + "loss": 1.2914, + "theoretical_loss": 3.3982262199175968, + "tokens_seen": 2488532992 + }, + { + "epoch": 0.75, + "learning_rate": 0.00019888465396188568, + "loss": 1.3559, + "theoretical_loss": 3.3982191347881474, + "tokens_seen": 2488598528 + }, + { + "epoch": 0.75, + "learning_rate": 0.00019886860581745238, + "loss": 1.3528, + "theoretical_loss": 3.398212049897521, + "tokens_seen": 2488664064 + }, + { + "epoch": 0.75, + "learning_rate": 0.00019885255767301908, + "loss": 1.3344, + "theoretical_loss": 3.3982049652457023, + "tokens_seen": 2488729600 + }, + { + "epoch": 0.75, + "learning_rate": 0.00019883650952858578, + "loss": 1.3083, + "theoretical_loss": 3.398197880832678, + "tokens_seen": 2488795136 + }, + { + "epoch": 0.75, + "learning_rate": 0.00019882046138415248, + "loss": 1.3463, + "theoretical_loss": 3.398190796658433, + "tokens_seen": 2488860672 + }, + { + "epoch": 0.75, + "learning_rate": 0.00019880441323971918, + "loss": 1.2971, + "theoretical_loss": 3.398183712722954, + "tokens_seen": 2488926208 + }, + { + "epoch": 0.75, + "learning_rate": 0.00019878836509528585, + "loss": 1.3623, + "theoretical_loss": 3.3981766290262256, + "tokens_seen": 2488991744 + }, + { + "epoch": 0.75, + "learning_rate": 0.00019877231695085255, + "loss": 1.2885, + "theoretical_loss": 3.3981695455682335, + "tokens_seen": 2489057280 + }, + { + "epoch": 0.75, + "learning_rate": 0.00019875626880641925, + "loss": 1.3163, + "theoretical_loss": 3.398162462348964, + "tokens_seen": 2489122816 + }, + { + "epoch": 0.75, + "learning_rate": 0.00019874022066198596, + "loss": 1.2446, + "theoretical_loss": 3.398155379368403, + "tokens_seen": 2489188352 + }, + { + "epoch": 0.75, + "learning_rate": 0.00019872417251755266, + "loss": 1.3227, + "theoretical_loss": 3.398148296626535, + "tokens_seen": 2489253888 + }, + { + "epoch": 0.75, + "learning_rate": 0.00019870812437311936, + "loss": 1.2266, + "theoretical_loss": 3.3981412141233465, + "tokens_seen": 2489319424 + }, + { + "epoch": 0.75, + "learning_rate": 0.00019869207622868606, + "loss": 1.3006, + "theoretical_loss": 3.398134131858823, + "tokens_seen": 2489384960 + }, + { + "epoch": 0.75, + "learning_rate": 0.00019867602808425276, + "loss": 1.2127, + "theoretical_loss": 3.39812704983295, + "tokens_seen": 2489450496 + }, + { + "epoch": 0.75, + "learning_rate": 0.00019865997993981946, + "loss": 1.4291, + "theoretical_loss": 3.3981199680457137, + "tokens_seen": 2489516032 + }, + { + "epoch": 0.75, + "learning_rate": 0.00019864393179538616, + "loss": 1.4189, + "theoretical_loss": 3.398112886497099, + "tokens_seen": 2489581568 + }, + { + "epoch": 0.75, + "learning_rate": 0.00019862788365095286, + "loss": 1.3057, + "theoretical_loss": 3.3981058051870923, + "tokens_seen": 2489647104 + }, + { + "epoch": 0.75, + "learning_rate": 0.00019861183550651956, + "loss": 1.3369, + "theoretical_loss": 3.398098724115679, + "tokens_seen": 2489712640 + }, + { + "epoch": 0.75, + "learning_rate": 0.00019859578736208626, + "loss": 1.2235, + "theoretical_loss": 3.3980916432828447, + "tokens_seen": 2489778176 + }, + { + "epoch": 0.75, + "learning_rate": 0.00019857973921765296, + "loss": 1.2563, + "theoretical_loss": 3.3980845626885756, + "tokens_seen": 2489843712 + }, + { + "epoch": 0.75, + "learning_rate": 0.00019856369107321966, + "loss": 1.1996, + "theoretical_loss": 3.398077482332856, + "tokens_seen": 2489909248 + }, + { + "epoch": 0.75, + "learning_rate": 0.00019854764292878636, + "loss": 1.2977, + "theoretical_loss": 3.398070402215673, + "tokens_seen": 2489974784 + }, + { + "epoch": 0.75, + "learning_rate": 0.00019853159478435306, + "loss": 1.2773, + "theoretical_loss": 3.398063322337012, + "tokens_seen": 2490040320 + }, + { + "epoch": 0.75, + "learning_rate": 0.00019851554663991976, + "loss": 1.2333, + "theoretical_loss": 3.3980562426968586, + "tokens_seen": 2490105856 + }, + { + "epoch": 0.75, + "learning_rate": 0.00019849949849548646, + "loss": 1.343, + "theoretical_loss": 3.3980491632951977, + "tokens_seen": 2490171392 + }, + { + "epoch": 0.75, + "learning_rate": 0.0001984834503510532, + "loss": 1.1853, + "theoretical_loss": 3.3980420841320162, + "tokens_seen": 2490236928 + }, + { + "epoch": 0.75, + "learning_rate": 0.00019846740220661987, + "loss": 1.3387, + "theoretical_loss": 3.398035005207299, + "tokens_seen": 2490302464 + }, + { + "epoch": 0.75, + "objective/train/docs_used": 906478, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.212127923965454, + "objective/train/theoretical_loss": 3.3980279265210322, + "objective/train/tokens_used": 2510828000, + "theoretical_loss": 3.3980279265210322, + "tokens_seen": 2490368000 + }, + { + "epoch": 0.75, + "learning_rate": 0.00019845135406218657, + "loss": 1.2881, + "theoretical_loss": 3.3980279265210322, + "tokens_seen": 2490368000 + }, + { + "epoch": 0.75, + "learning_rate": 0.00019843530591775327, + "loss": 1.2653, + "theoretical_loss": 3.3980208480732013, + "tokens_seen": 2490433536 + }, + { + "epoch": 0.75, + "learning_rate": 0.00019841925777331997, + "loss": 1.3093, + "theoretical_loss": 3.398013769863792, + "tokens_seen": 2490499072 + }, + { + "epoch": 0.75, + "learning_rate": 0.00019840320962888667, + "loss": 1.1857, + "theoretical_loss": 3.39800669189279, + "tokens_seen": 2490564608 + }, + { + "epoch": 0.75, + "learning_rate": 0.00019838716148445337, + "loss": 1.2273, + "theoretical_loss": 3.397999614160181, + "tokens_seen": 2490630144 + }, + { + "epoch": 0.75, + "learning_rate": 0.00019837111334002007, + "loss": 1.3489, + "theoretical_loss": 3.397992536665951, + "tokens_seen": 2490695680 + }, + { + "epoch": 0.75, + "learning_rate": 0.00019835506519558677, + "loss": 1.2327, + "theoretical_loss": 3.3979854594100853, + "tokens_seen": 2490761216 + }, + { + "epoch": 0.75, + "learning_rate": 0.00019833901705115347, + "loss": 1.3457, + "theoretical_loss": 3.3979783823925693, + "tokens_seen": 2490826752 + }, + { + "epoch": 0.75, + "learning_rate": 0.00019832296890672017, + "loss": 1.2969, + "theoretical_loss": 3.39797130561339, + "tokens_seen": 2490892288 + }, + { + "epoch": 0.75, + "learning_rate": 0.00019830692076228687, + "loss": 1.3473, + "theoretical_loss": 3.3979642290725316, + "tokens_seen": 2490957824 + }, + { + "epoch": 0.75, + "learning_rate": 0.00019829087261785357, + "loss": 1.3477, + "theoretical_loss": 3.3979571527699806, + "tokens_seen": 2491023360 + }, + { + "epoch": 0.75, + "learning_rate": 0.00019827482447342027, + "loss": 1.2895, + "theoretical_loss": 3.3979500767057225, + "tokens_seen": 2491088896 + }, + { + "epoch": 0.75, + "learning_rate": 0.00019825877632898697, + "loss": 1.2602, + "theoretical_loss": 3.397943000879743, + "tokens_seen": 2491154432 + }, + { + "epoch": 0.75, + "learning_rate": 0.00019824272818455367, + "loss": 1.3861, + "theoretical_loss": 3.397935925292028, + "tokens_seen": 2491219968 + }, + { + "epoch": 0.75, + "learning_rate": 0.00019822668004012038, + "loss": 1.3663, + "theoretical_loss": 3.397928849942563, + "tokens_seen": 2491285504 + }, + { + "epoch": 0.75, + "learning_rate": 0.00019821063189568708, + "loss": 1.3975, + "theoretical_loss": 3.3979217748313335, + "tokens_seen": 2491351040 + }, + { + "epoch": 0.75, + "learning_rate": 0.00019819458375125378, + "loss": 1.3293, + "theoretical_loss": 3.397914699958326, + "tokens_seen": 2491416576 + }, + { + "epoch": 0.75, + "learning_rate": 0.00019817853560682048, + "loss": 1.3197, + "theoretical_loss": 3.3979076253235254, + "tokens_seen": 2491482112 + }, + { + "epoch": 0.76, + "learning_rate": 0.00019816248746238718, + "loss": 1.2522, + "theoretical_loss": 3.397900550926918, + "tokens_seen": 2491547648 + }, + { + "epoch": 0.76, + "learning_rate": 0.00019814643931795388, + "loss": 1.2949, + "theoretical_loss": 3.3978934767684885, + "tokens_seen": 2491613184 + }, + { + "epoch": 0.76, + "learning_rate": 0.00019813039117352058, + "loss": 1.4729, + "theoretical_loss": 3.3978864028482243, + "tokens_seen": 2491678720 + }, + { + "epoch": 0.76, + "learning_rate": 0.00019811434302908728, + "loss": 1.2717, + "theoretical_loss": 3.3978793291661096, + "tokens_seen": 2491744256 + }, + { + "epoch": 0.76, + "learning_rate": 0.00019809829488465398, + "loss": 1.3425, + "theoretical_loss": 3.397872255722131, + "tokens_seen": 2491809792 + }, + { + "epoch": 0.76, + "learning_rate": 0.00019808224674022065, + "loss": 1.3189, + "theoretical_loss": 3.3978651825162736, + "tokens_seen": 2491875328 + }, + { + "epoch": 0.76, + "learning_rate": 0.00019806619859578735, + "loss": 1.322, + "theoretical_loss": 3.3978581095485234, + "tokens_seen": 2491940864 + }, + { + "epoch": 0.76, + "learning_rate": 0.00019805015045135406, + "loss": 1.2924, + "theoretical_loss": 3.3978510368188664, + "tokens_seen": 2492006400 + }, + { + "epoch": 0.76, + "learning_rate": 0.00019803410230692076, + "loss": 1.3454, + "theoretical_loss": 3.3978439643272877, + "tokens_seen": 2492071936 + }, + { + "epoch": 0.76, + "learning_rate": 0.00019801805416248746, + "loss": 1.3087, + "theoretical_loss": 3.397836892073774, + "tokens_seen": 2492137472 + }, + { + "epoch": 0.76, + "learning_rate": 0.00019800200601805416, + "loss": 1.3611, + "theoretical_loss": 3.39782982005831, + "tokens_seen": 2492203008 + }, + { + "epoch": 0.76, + "learning_rate": 0.00019798595787362086, + "loss": 1.2619, + "theoretical_loss": 3.397822748280882, + "tokens_seen": 2492268544 + }, + { + "epoch": 0.76, + "learning_rate": 0.00019796990972918756, + "loss": 1.3714, + "theoretical_loss": 3.3978156767414753, + "tokens_seen": 2492334080 + }, + { + "epoch": 0.76, + "learning_rate": 0.00019795386158475426, + "loss": 1.1924, + "theoretical_loss": 3.3978086054400762, + "tokens_seen": 2492399616 + }, + { + "epoch": 0.76, + "learning_rate": 0.00019793781344032096, + "loss": 1.1791, + "theoretical_loss": 3.3978015343766703, + "tokens_seen": 2492465152 + }, + { + "epoch": 0.76, + "learning_rate": 0.00019792176529588766, + "loss": 1.4027, + "theoretical_loss": 3.397794463551243, + "tokens_seen": 2492530688 + }, + { + "epoch": 0.76, + "learning_rate": 0.00019790571715145436, + "loss": 1.3121, + "theoretical_loss": 3.39778739296378, + "tokens_seen": 2492596224 + }, + { + "epoch": 0.76, + "learning_rate": 0.0001978896690070211, + "loss": 1.3396, + "theoretical_loss": 3.3977803226142673, + "tokens_seen": 2492661760 + }, + { + "epoch": 0.76, + "learning_rate": 0.0001978736208625878, + "loss": 1.3694, + "theoretical_loss": 3.3977732525026907, + "tokens_seen": 2492727296 + }, + { + "epoch": 0.76, + "learning_rate": 0.0001978575727181545, + "loss": 1.276, + "theoretical_loss": 3.397766182629036, + "tokens_seen": 2492792832 + }, + { + "epoch": 0.76, + "learning_rate": 0.0001978415245737212, + "loss": 1.273, + "theoretical_loss": 3.3977591129932883, + "tokens_seen": 2492858368 + }, + { + "epoch": 0.76, + "learning_rate": 0.0001978254764292879, + "loss": 1.3485, + "theoretical_loss": 3.3977520435954345, + "tokens_seen": 2492923904 + }, + { + "epoch": 0.76, + "learning_rate": 0.0001978094282848546, + "loss": 1.2269, + "theoretical_loss": 3.397744974435459, + "tokens_seen": 2492989440 + }, + { + "epoch": 0.76, + "learning_rate": 0.0001977933801404213, + "loss": 1.2669, + "theoretical_loss": 3.3977379055133485, + "tokens_seen": 2493054976 + }, + { + "epoch": 0.76, + "learning_rate": 0.000197777331995988, + "loss": 1.3765, + "theoretical_loss": 3.397730836829088, + "tokens_seen": 2493120512 + }, + { + "epoch": 0.76, + "learning_rate": 0.00019776128385155467, + "loss": 1.2759, + "theoretical_loss": 3.397723768382664, + "tokens_seen": 2493186048 + }, + { + "epoch": 0.76, + "learning_rate": 0.00019774523570712137, + "loss": 1.3393, + "theoretical_loss": 3.3977167001740622, + "tokens_seen": 2493251584 + }, + { + "epoch": 0.76, + "learning_rate": 0.00019772918756268807, + "loss": 1.3531, + "theoretical_loss": 3.397709632203268, + "tokens_seen": 2493317120 + }, + { + "epoch": 0.76, + "learning_rate": 0.00019771313941825477, + "loss": 1.2896, + "theoretical_loss": 3.3977025644702668, + "tokens_seen": 2493382656 + }, + { + "epoch": 0.76, + "learning_rate": 0.00019769709127382147, + "loss": 1.4047, + "theoretical_loss": 3.397695496975045, + "tokens_seen": 2493448192 + }, + { + "epoch": 0.76, + "learning_rate": 0.00019768104312938817, + "loss": 1.3381, + "theoretical_loss": 3.397688429717588, + "tokens_seen": 2493513728 + }, + { + "epoch": 0.76, + "learning_rate": 0.00019766499498495487, + "loss": 1.2167, + "theoretical_loss": 3.397681362697882, + "tokens_seen": 2493579264 + }, + { + "epoch": 0.76, + "objective/train/docs_used": 907814, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.2638299465179443, + "objective/train/theoretical_loss": 3.3976742959159116, + "objective/train/tokens_used": 2514104800, + "theoretical_loss": 3.3976742959159116, + "tokens_seen": 2493644800 + }, + { + "epoch": 0.76, + "learning_rate": 0.00019764894684052157, + "loss": 1.3043, + "theoretical_loss": 3.3976742959159116, + "tokens_seen": 2493644800 + }, + { + "epoch": 0.76, + "learning_rate": 0.00019763289869608827, + "loss": 1.3346, + "theoretical_loss": 3.397667229371664, + "tokens_seen": 2493710336 + }, + { + "epoch": 0.76, + "learning_rate": 0.00019761685055165497, + "loss": 1.313, + "theoretical_loss": 3.3976601630651246, + "tokens_seen": 2493775872 + }, + { + "epoch": 0.76, + "learning_rate": 0.00019760080240722167, + "loss": 1.1832, + "theoretical_loss": 3.3976530969962786, + "tokens_seen": 2493841408 + }, + { + "epoch": 0.76, + "learning_rate": 0.00019758475426278837, + "loss": 1.2897, + "theoretical_loss": 3.3976460311651118, + "tokens_seen": 2493906944 + }, + { + "epoch": 0.76, + "learning_rate": 0.00019756870611835507, + "loss": 1.2813, + "theoretical_loss": 3.3976389655716104, + "tokens_seen": 2493972480 + }, + { + "epoch": 0.76, + "learning_rate": 0.00019755265797392178, + "loss": 1.3118, + "theoretical_loss": 3.3976319002157602, + "tokens_seen": 2494038016 + }, + { + "epoch": 0.76, + "learning_rate": 0.00019753660982948848, + "loss": 1.3091, + "theoretical_loss": 3.397624835097546, + "tokens_seen": 2494103552 + }, + { + "epoch": 0.76, + "learning_rate": 0.00019752056168505518, + "loss": 1.3295, + "theoretical_loss": 3.3976177702169554, + "tokens_seen": 2494169088 + }, + { + "epoch": 0.76, + "learning_rate": 0.00019750451354062188, + "loss": 1.2173, + "theoretical_loss": 3.3976107055739724, + "tokens_seen": 2494234624 + }, + { + "epoch": 0.76, + "learning_rate": 0.00019748846539618858, + "loss": 1.3136, + "theoretical_loss": 3.3976036411685833, + "tokens_seen": 2494300160 + }, + { + "epoch": 0.76, + "learning_rate": 0.00019747241725175528, + "loss": 1.2791, + "theoretical_loss": 3.3975965770007743, + "tokens_seen": 2494365696 + }, + { + "epoch": 0.76, + "learning_rate": 0.00019745636910732198, + "loss": 1.3801, + "theoretical_loss": 3.397589513070531, + "tokens_seen": 2494431232 + }, + { + "epoch": 0.76, + "learning_rate": 0.00019744032096288868, + "loss": 1.2785, + "theoretical_loss": 3.3975824493778384, + "tokens_seen": 2494496768 + }, + { + "epoch": 0.76, + "learning_rate": 0.00019742427281845538, + "loss": 1.264, + "theoretical_loss": 3.3975753859226834, + "tokens_seen": 2494562304 + }, + { + "epoch": 0.76, + "learning_rate": 0.00019740822467402208, + "loss": 1.2314, + "theoretical_loss": 3.397568322705051, + "tokens_seen": 2494627840 + }, + { + "epoch": 0.76, + "learning_rate": 0.00019739217652958878, + "loss": 1.1547, + "theoretical_loss": 3.3975612597249274, + "tokens_seen": 2494693376 + }, + { + "epoch": 0.76, + "learning_rate": 0.00019737612838515545, + "loss": 1.3302, + "theoretical_loss": 3.3975541969822984, + "tokens_seen": 2494758912 + }, + { + "epoch": 0.76, + "learning_rate": 0.00019736008024072216, + "loss": 1.2439, + "theoretical_loss": 3.3975471344771493, + "tokens_seen": 2494824448 + }, + { + "epoch": 0.76, + "learning_rate": 0.00019734403209628886, + "loss": 1.2942, + "theoretical_loss": 3.397540072209466, + "tokens_seen": 2494889984 + }, + { + "epoch": 0.76, + "learning_rate": 0.00019732798395185556, + "loss": 1.2808, + "theoretical_loss": 3.397533010179235, + "tokens_seen": 2494955520 + }, + { + "epoch": 0.76, + "learning_rate": 0.00019731193580742226, + "loss": 1.2811, + "theoretical_loss": 3.3975259483864413, + "tokens_seen": 2495021056 + }, + { + "epoch": 0.76, + "learning_rate": 0.00019729588766298899, + "loss": 1.2653, + "theoretical_loss": 3.3975188868310706, + "tokens_seen": 2495086592 + }, + { + "epoch": 0.76, + "learning_rate": 0.00019727983951855569, + "loss": 1.2984, + "theoretical_loss": 3.397511825513109, + "tokens_seen": 2495152128 + }, + { + "epoch": 0.76, + "learning_rate": 0.0001972637913741224, + "loss": 1.3048, + "theoretical_loss": 3.397504764432543, + "tokens_seen": 2495217664 + }, + { + "epoch": 0.76, + "learning_rate": 0.0001972477432296891, + "loss": 1.2114, + "theoretical_loss": 3.397497703589357, + "tokens_seen": 2495283200 + }, + { + "epoch": 0.76, + "learning_rate": 0.0001972316950852558, + "loss": 1.2627, + "theoretical_loss": 3.397490642983538, + "tokens_seen": 2495348736 + }, + { + "epoch": 0.76, + "learning_rate": 0.0001972156469408225, + "loss": 1.2785, + "theoretical_loss": 3.3974835826150707, + "tokens_seen": 2495414272 + }, + { + "epoch": 0.76, + "learning_rate": 0.0001971995987963892, + "loss": 1.3055, + "theoretical_loss": 3.3974765224839416, + "tokens_seen": 2495479808 + }, + { + "epoch": 0.76, + "learning_rate": 0.0001971835506519559, + "loss": 1.3293, + "theoretical_loss": 3.397469462590136, + "tokens_seen": 2495545344 + }, + { + "epoch": 0.76, + "learning_rate": 0.0001971675025075226, + "loss": 1.2556, + "theoretical_loss": 3.3974624029336407, + "tokens_seen": 2495610880 + }, + { + "epoch": 0.76, + "learning_rate": 0.0001971514543630893, + "loss": 1.3397, + "theoretical_loss": 3.39745534351444, + "tokens_seen": 2495676416 + }, + { + "epoch": 0.76, + "learning_rate": 0.000197135406218656, + "loss": 1.2465, + "theoretical_loss": 3.397448284332521, + "tokens_seen": 2495741952 + }, + { + "epoch": 0.76, + "learning_rate": 0.0001971193580742227, + "loss": 1.2935, + "theoretical_loss": 3.397441225387869, + "tokens_seen": 2495807488 + }, + { + "epoch": 0.76, + "learning_rate": 0.0001971033099297894, + "loss": 1.2736, + "theoretical_loss": 3.3974341666804695, + "tokens_seen": 2495873024 + }, + { + "epoch": 0.76, + "learning_rate": 0.0001970872617853561, + "loss": 1.3396, + "theoretical_loss": 3.3974271082103087, + "tokens_seen": 2495938560 + }, + { + "epoch": 0.76, + "learning_rate": 0.00019707121364092277, + "loss": 1.2471, + "theoretical_loss": 3.397420049977372, + "tokens_seen": 2496004096 + }, + { + "epoch": 0.76, + "learning_rate": 0.00019705516549648947, + "loss": 1.2276, + "theoretical_loss": 3.397412991981646, + "tokens_seen": 2496069632 + }, + { + "epoch": 0.76, + "learning_rate": 0.00019703911735205617, + "loss": 1.3975, + "theoretical_loss": 3.3974059342231158, + "tokens_seen": 2496135168 + }, + { + "epoch": 0.76, + "learning_rate": 0.00019702306920762287, + "loss": 1.3382, + "theoretical_loss": 3.397398876701767, + "tokens_seen": 2496200704 + }, + { + "epoch": 0.76, + "learning_rate": 0.00019700702106318957, + "loss": 1.2905, + "theoretical_loss": 3.397391819417586, + "tokens_seen": 2496266240 + }, + { + "epoch": 0.76, + "learning_rate": 0.00019699097291875627, + "loss": 1.2842, + "theoretical_loss": 3.3973847623705584, + "tokens_seen": 2496331776 + }, + { + "epoch": 0.76, + "learning_rate": 0.00019697492477432297, + "loss": 1.2624, + "theoretical_loss": 3.39737770556067, + "tokens_seen": 2496397312 + }, + { + "epoch": 0.76, + "learning_rate": 0.00019695887662988967, + "loss": 1.2437, + "theoretical_loss": 3.3973706489879065, + "tokens_seen": 2496462848 + }, + { + "epoch": 0.76, + "learning_rate": 0.00019694282848545637, + "loss": 1.3412, + "theoretical_loss": 3.397363592652254, + "tokens_seen": 2496528384 + }, + { + "epoch": 0.76, + "learning_rate": 0.00019692678034102307, + "loss": 1.2065, + "theoretical_loss": 3.3973565365536977, + "tokens_seen": 2496593920 + }, + { + "epoch": 0.76, + "learning_rate": 0.00019691073219658977, + "loss": 1.3034, + "theoretical_loss": 3.397349480692224, + "tokens_seen": 2496659456 + }, + { + "epoch": 0.76, + "learning_rate": 0.00019689468405215647, + "loss": 1.3431, + "theoretical_loss": 3.3973424250678184, + "tokens_seen": 2496724992 + }, + { + "epoch": 0.76, + "learning_rate": 0.00019687863590772317, + "loss": 1.248, + "theoretical_loss": 3.397335369680467, + "tokens_seen": 2496790528 + }, + { + "epoch": 0.76, + "learning_rate": 0.00019686258776328988, + "loss": 1.3723, + "theoretical_loss": 3.3973283145301556, + "tokens_seen": 2496856064 + }, + { + "epoch": 0.76, + "objective/train/docs_used": 908925, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.295708179473877, + "objective/train/theoretical_loss": 3.39732125961687, + "objective/train/tokens_used": 2517381600, + "theoretical_loss": 3.39732125961687, + "tokens_seen": 2496921600 + }, + { + "epoch": 0.76, + "learning_rate": 0.00019684653961885658, + "loss": 1.2662, + "theoretical_loss": 3.39732125961687, + "tokens_seen": 2496921600 + }, + { + "epoch": 0.76, + "learning_rate": 0.00019683049147442328, + "loss": 1.2245, + "theoretical_loss": 3.3973142049405953, + "tokens_seen": 2496987136 + }, + { + "epoch": 0.76, + "learning_rate": 0.00019681444332998998, + "loss": 1.2852, + "theoretical_loss": 3.397307150501318, + "tokens_seen": 2497052672 + }, + { + "epoch": 0.76, + "learning_rate": 0.00019679839518555668, + "loss": 1.2905, + "theoretical_loss": 3.3973000962990243, + "tokens_seen": 2497118208 + }, + { + "epoch": 0.76, + "learning_rate": 0.00019678234704112338, + "loss": 1.2852, + "theoretical_loss": 3.397293042333699, + "tokens_seen": 2497183744 + }, + { + "epoch": 0.76, + "learning_rate": 0.00019676629889669008, + "loss": 1.3428, + "theoretical_loss": 3.397285988605329, + "tokens_seen": 2497249280 + }, + { + "epoch": 0.76, + "learning_rate": 0.00019675025075225678, + "loss": 1.2144, + "theoretical_loss": 3.3972789351138992, + "tokens_seen": 2497314816 + }, + { + "epoch": 0.76, + "learning_rate": 0.00019673420260782348, + "loss": 1.3486, + "theoretical_loss": 3.397271881859396, + "tokens_seen": 2497380352 + }, + { + "epoch": 0.76, + "learning_rate": 0.00019671815446339018, + "loss": 1.2319, + "theoretical_loss": 3.397264828841805, + "tokens_seen": 2497445888 + }, + { + "epoch": 0.76, + "learning_rate": 0.00019670210631895688, + "loss": 1.3342, + "theoretical_loss": 3.3972577760611116, + "tokens_seen": 2497511424 + }, + { + "epoch": 0.76, + "learning_rate": 0.00019668605817452358, + "loss": 1.2607, + "theoretical_loss": 3.3972507235173026, + "tokens_seen": 2497576960 + }, + { + "epoch": 0.76, + "learning_rate": 0.00019667001003009028, + "loss": 1.2988, + "theoretical_loss": 3.397243671210363, + "tokens_seen": 2497642496 + }, + { + "epoch": 0.76, + "learning_rate": 0.00019665396188565698, + "loss": 1.3036, + "theoretical_loss": 3.397236619140279, + "tokens_seen": 2497708032 + }, + { + "epoch": 0.76, + "learning_rate": 0.00019663791374122368, + "loss": 1.3188, + "theoretical_loss": 3.3972295673070363, + "tokens_seen": 2497773568 + }, + { + "epoch": 0.76, + "learning_rate": 0.00019662186559679038, + "loss": 1.2471, + "theoretical_loss": 3.397222515710621, + "tokens_seen": 2497839104 + }, + { + "epoch": 0.76, + "learning_rate": 0.00019660581745235709, + "loss": 1.3538, + "theoretical_loss": 3.3972154643510186, + "tokens_seen": 2497904640 + }, + { + "epoch": 0.76, + "learning_rate": 0.00019658976930792379, + "loss": 1.2749, + "theoretical_loss": 3.397208413228215, + "tokens_seen": 2497970176 + }, + { + "epoch": 0.76, + "learning_rate": 0.0001965737211634905, + "loss": 1.3032, + "theoretical_loss": 3.397201362342196, + "tokens_seen": 2498035712 + }, + { + "epoch": 0.76, + "learning_rate": 0.0001965576730190572, + "loss": 1.3333, + "theoretical_loss": 3.397194311692948, + "tokens_seen": 2498101248 + }, + { + "epoch": 0.76, + "learning_rate": 0.0001965416248746239, + "loss": 1.1894, + "theoretical_loss": 3.3971872612804566, + "tokens_seen": 2498166784 + }, + { + "epoch": 0.76, + "learning_rate": 0.0001965255767301906, + "loss": 1.1939, + "theoretical_loss": 3.3971802111047067, + "tokens_seen": 2498232320 + }, + { + "epoch": 0.76, + "learning_rate": 0.0001965095285857573, + "loss": 1.2864, + "theoretical_loss": 3.397173161165685, + "tokens_seen": 2498297856 + }, + { + "epoch": 0.76, + "learning_rate": 0.000196493480441324, + "loss": 1.3338, + "theoretical_loss": 3.3971661114633775, + "tokens_seen": 2498363392 + }, + { + "epoch": 0.76, + "learning_rate": 0.0001964774322968907, + "loss": 1.1979, + "theoretical_loss": 3.3971590619977694, + "tokens_seen": 2498428928 + }, + { + "epoch": 0.76, + "learning_rate": 0.0001964613841524574, + "loss": 1.2064, + "theoretical_loss": 3.397152012768847, + "tokens_seen": 2498494464 + }, + { + "epoch": 0.76, + "learning_rate": 0.0001964453360080241, + "loss": 1.3759, + "theoretical_loss": 3.3971449637765962, + "tokens_seen": 2498560000 + }, + { + "epoch": 0.76, + "learning_rate": 0.0001964292878635908, + "loss": 1.3807, + "theoretical_loss": 3.3971379150210024, + "tokens_seen": 2498625536 + }, + { + "epoch": 0.76, + "learning_rate": 0.0001964132397191575, + "loss": 1.1916, + "theoretical_loss": 3.397130866502052, + "tokens_seen": 2498691072 + }, + { + "epoch": 0.76, + "learning_rate": 0.0001963971915747242, + "loss": 1.3273, + "theoretical_loss": 3.3971238182197303, + "tokens_seen": 2498756608 + }, + { + "epoch": 0.76, + "learning_rate": 0.0001963811434302909, + "loss": 1.3383, + "theoretical_loss": 3.3971167701740237, + "tokens_seen": 2498822144 + }, + { + "epoch": 0.76, + "learning_rate": 0.00019636509528585757, + "loss": 1.2983, + "theoretical_loss": 3.3971097223649176, + "tokens_seen": 2498887680 + }, + { + "epoch": 0.76, + "learning_rate": 0.00019634904714142427, + "loss": 1.3649, + "theoretical_loss": 3.397102674792398, + "tokens_seen": 2498953216 + }, + { + "epoch": 0.76, + "learning_rate": 0.00019633299899699097, + "loss": 1.255, + "theoretical_loss": 3.3970956274564506, + "tokens_seen": 2499018752 + }, + { + "epoch": 0.76, + "learning_rate": 0.00019631695085255767, + "loss": 1.2776, + "theoretical_loss": 3.3970885803570616, + "tokens_seen": 2499084288 + }, + { + "epoch": 0.76, + "learning_rate": 0.00019630090270812437, + "loss": 1.1373, + "theoretical_loss": 3.397081533494217, + "tokens_seen": 2499149824 + }, + { + "epoch": 0.76, + "learning_rate": 0.00019628485456369107, + "loss": 1.3337, + "theoretical_loss": 3.397074486867902, + "tokens_seen": 2499215360 + }, + { + "epoch": 0.76, + "learning_rate": 0.00019626880641925777, + "loss": 1.2495, + "theoretical_loss": 3.3970674404781027, + "tokens_seen": 2499280896 + }, + { + "epoch": 0.76, + "learning_rate": 0.00019625275827482447, + "loss": 1.2499, + "theoretical_loss": 3.3970603943248054, + "tokens_seen": 2499346432 + }, + { + "epoch": 0.76, + "learning_rate": 0.00019623671013039117, + "loss": 1.2569, + "theoretical_loss": 3.3970533484079954, + "tokens_seen": 2499411968 + }, + { + "epoch": 0.76, + "learning_rate": 0.00019622066198595787, + "loss": 1.2934, + "theoretical_loss": 3.397046302727659, + "tokens_seen": 2499477504 + }, + { + "epoch": 0.76, + "learning_rate": 0.00019620461384152457, + "loss": 1.3122, + "theoretical_loss": 3.3970392572837818, + "tokens_seen": 2499543040 + }, + { + "epoch": 0.76, + "learning_rate": 0.00019618856569709127, + "loss": 1.2001, + "theoretical_loss": 3.3970322120763496, + "tokens_seen": 2499608576 + }, + { + "epoch": 0.76, + "learning_rate": 0.00019617251755265798, + "loss": 1.3474, + "theoretical_loss": 3.3970251671053484, + "tokens_seen": 2499674112 + }, + { + "epoch": 0.76, + "learning_rate": 0.00019615646940822468, + "loss": 1.3324, + "theoretical_loss": 3.3970181223707643, + "tokens_seen": 2499739648 + }, + { + "epoch": 0.76, + "learning_rate": 0.00019614042126379138, + "loss": 1.3453, + "theoretical_loss": 3.3970110778725826, + "tokens_seen": 2499805184 + }, + { + "epoch": 0.76, + "learning_rate": 0.00019612437311935808, + "loss": 1.3264, + "theoretical_loss": 3.3970040336107896, + "tokens_seen": 2499870720 + }, + { + "epoch": 0.76, + "learning_rate": 0.00019610832497492478, + "loss": 1.3729, + "theoretical_loss": 3.396996989585371, + "tokens_seen": 2499936256 + }, + { + "epoch": 0.76, + "learning_rate": 0.0001960922768304915, + "loss": 1.2541, + "theoretical_loss": 3.396989945796313, + "tokens_seen": 2500001792 + }, + { + "epoch": 0.76, + "learning_rate": 0.0001960762286860582, + "loss": 1.3176, + "theoretical_loss": 3.3969829022436007, + "tokens_seen": 2500067328 + }, + { + "epoch": 0.76, + "learning_rate": 0.0001960601805416249, + "loss": 1.2705, + "theoretical_loss": 3.3969758589272208, + "tokens_seen": 2500132864 + }, + { + "epoch": 0.76, + "objective/train/docs_used": 910166, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.2162121534347534, + "objective/train/theoretical_loss": 3.396968815847159, + "objective/train/tokens_used": 2520658400, + "theoretical_loss": 3.396968815847159, + "tokens_seen": 2500198400 + }, + { + "epoch": 0.76, + "learning_rate": 0.00019604413239719158, + "loss": 1.2516, + "theoretical_loss": 3.396968815847159, + "tokens_seen": 2500198400 + }, + { + "epoch": 0.76, + "learning_rate": 0.00019602808425275828, + "loss": 1.1513, + "theoretical_loss": 3.396961773003401, + "tokens_seen": 2500263936 + }, + { + "epoch": 0.76, + "learning_rate": 0.00019601203610832498, + "loss": 1.2599, + "theoretical_loss": 3.396954730395932, + "tokens_seen": 2500329472 + }, + { + "epoch": 0.76, + "learning_rate": 0.00019599598796389168, + "loss": 1.333, + "theoretical_loss": 3.3969476880247393, + "tokens_seen": 2500395008 + }, + { + "epoch": 0.76, + "learning_rate": 0.00019597993981945838, + "loss": 1.2786, + "theoretical_loss": 3.396940645889808, + "tokens_seen": 2500460544 + }, + { + "epoch": 0.76, + "learning_rate": 0.00019596389167502508, + "loss": 1.1597, + "theoretical_loss": 3.3969336039911244, + "tokens_seen": 2500526080 + }, + { + "epoch": 0.76, + "learning_rate": 0.00019594784353059178, + "loss": 1.3877, + "theoretical_loss": 3.3969265623286735, + "tokens_seen": 2500591616 + }, + { + "epoch": 0.76, + "learning_rate": 0.00019593179538615848, + "loss": 1.2458, + "theoretical_loss": 3.3969195209024416, + "tokens_seen": 2500657152 + }, + { + "epoch": 0.76, + "learning_rate": 0.00019591574724172519, + "loss": 1.2402, + "theoretical_loss": 3.396912479712415, + "tokens_seen": 2500722688 + }, + { + "epoch": 0.76, + "learning_rate": 0.00019589969909729189, + "loss": 1.2234, + "theoretical_loss": 3.3969054387585795, + "tokens_seen": 2500788224 + }, + { + "epoch": 0.76, + "learning_rate": 0.0001958836509528586, + "loss": 1.342, + "theoretical_loss": 3.3968983980409204, + "tokens_seen": 2500853760 + }, + { + "epoch": 0.76, + "learning_rate": 0.0001958676028084253, + "loss": 1.2941, + "theoretical_loss": 3.396891357559424, + "tokens_seen": 2500919296 + }, + { + "epoch": 0.76, + "learning_rate": 0.000195851554663992, + "loss": 1.2563, + "theoretical_loss": 3.3968843173140764, + "tokens_seen": 2500984832 + }, + { + "epoch": 0.76, + "learning_rate": 0.0001958355065195587, + "loss": 1.2964, + "theoretical_loss": 3.3968772773048634, + "tokens_seen": 2501050368 + }, + { + "epoch": 0.76, + "learning_rate": 0.0001958194583751254, + "loss": 1.2612, + "theoretical_loss": 3.39687023753177, + "tokens_seen": 2501115904 + }, + { + "epoch": 0.76, + "learning_rate": 0.0001958034102306921, + "loss": 1.3465, + "theoretical_loss": 3.396863197994784, + "tokens_seen": 2501181440 + }, + { + "epoch": 0.76, + "learning_rate": 0.0001957873620862588, + "loss": 1.2147, + "theoretical_loss": 3.396856158693889, + "tokens_seen": 2501246976 + }, + { + "epoch": 0.76, + "learning_rate": 0.0001957713139418255, + "loss": 1.2048, + "theoretical_loss": 3.3968491196290724, + "tokens_seen": 2501312512 + }, + { + "epoch": 0.76, + "learning_rate": 0.0001957552657973922, + "loss": 1.1962, + "theoretical_loss": 3.3968420808003197, + "tokens_seen": 2501378048 + }, + { + "epoch": 0.76, + "learning_rate": 0.0001957392176529589, + "loss": 1.22, + "theoretical_loss": 3.3968350422076172, + "tokens_seen": 2501443584 + }, + { + "epoch": 0.76, + "learning_rate": 0.0001957231695085256, + "loss": 1.1592, + "theoretical_loss": 3.39682800385095, + "tokens_seen": 2501509120 + }, + { + "epoch": 0.76, + "learning_rate": 0.0001957071213640923, + "loss": 1.1877, + "theoretical_loss": 3.3968209657303046, + "tokens_seen": 2501574656 + }, + { + "epoch": 0.76, + "learning_rate": 0.000195691073219659, + "loss": 1.2555, + "theoretical_loss": 3.396813927845667, + "tokens_seen": 2501640192 + }, + { + "epoch": 0.76, + "learning_rate": 0.0001956750250752257, + "loss": 1.3102, + "theoretical_loss": 3.396806890197022, + "tokens_seen": 2501705728 + }, + { + "epoch": 0.76, + "learning_rate": 0.00019565897693079237, + "loss": 1.3071, + "theoretical_loss": 3.396799852784357, + "tokens_seen": 2501771264 + }, + { + "epoch": 0.76, + "learning_rate": 0.00019564292878635907, + "loss": 1.1765, + "theoretical_loss": 3.396792815607657, + "tokens_seen": 2501836800 + }, + { + "epoch": 0.76, + "learning_rate": 0.00019562688064192577, + "loss": 1.3022, + "theoretical_loss": 3.3967857786669082, + "tokens_seen": 2501902336 + }, + { + "epoch": 0.76, + "learning_rate": 0.00019561083249749247, + "loss": 1.2663, + "theoretical_loss": 3.3967787419620965, + "tokens_seen": 2501967872 + }, + { + "epoch": 0.76, + "learning_rate": 0.00019559478435305917, + "loss": 1.3761, + "theoretical_loss": 3.396771705493208, + "tokens_seen": 2502033408 + }, + { + "epoch": 0.76, + "learning_rate": 0.00019557873620862587, + "loss": 1.2449, + "theoretical_loss": 3.3967646692602282, + "tokens_seen": 2502098944 + }, + { + "epoch": 0.76, + "learning_rate": 0.00019556268806419257, + "loss": 1.3142, + "theoretical_loss": 3.3967576332631433, + "tokens_seen": 2502164480 + }, + { + "epoch": 0.76, + "learning_rate": 0.00019554663991975927, + "loss": 1.2375, + "theoretical_loss": 3.3967505975019385, + "tokens_seen": 2502230016 + }, + { + "epoch": 0.76, + "learning_rate": 0.00019553059177532597, + "loss": 1.3658, + "theoretical_loss": 3.396743561976601, + "tokens_seen": 2502295552 + }, + { + "epoch": 0.76, + "learning_rate": 0.00019551454363089267, + "loss": 1.386, + "theoretical_loss": 3.396736526687116, + "tokens_seen": 2502361088 + }, + { + "epoch": 0.76, + "learning_rate": 0.00019549849548645937, + "loss": 1.3783, + "theoretical_loss": 3.396729491633469, + "tokens_seen": 2502426624 + }, + { + "epoch": 0.76, + "learning_rate": 0.0001954824473420261, + "loss": 1.3231, + "theoretical_loss": 3.396722456815647, + "tokens_seen": 2502492160 + }, + { + "epoch": 0.76, + "learning_rate": 0.0001954663991975928, + "loss": 1.2424, + "theoretical_loss": 3.3967154222336347, + "tokens_seen": 2502557696 + }, + { + "epoch": 0.76, + "learning_rate": 0.0001954503510531595, + "loss": 1.2525, + "theoretical_loss": 3.396708387887419, + "tokens_seen": 2502623232 + }, + { + "epoch": 0.76, + "learning_rate": 0.0001954343029087262, + "loss": 1.3138, + "theoretical_loss": 3.3967013537769852, + "tokens_seen": 2502688768 + }, + { + "epoch": 0.76, + "learning_rate": 0.0001954182547642929, + "loss": 1.3006, + "theoretical_loss": 3.3966943199023194, + "tokens_seen": 2502754304 + }, + { + "epoch": 0.76, + "learning_rate": 0.0001954022066198596, + "loss": 1.3031, + "theoretical_loss": 3.3966872862634077, + "tokens_seen": 2502819840 + }, + { + "epoch": 0.76, + "learning_rate": 0.0001953861584754263, + "loss": 1.3285, + "theoretical_loss": 3.396680252860236, + "tokens_seen": 2502885376 + }, + { + "epoch": 0.76, + "learning_rate": 0.000195370110330993, + "loss": 1.3152, + "theoretical_loss": 3.39667321969279, + "tokens_seen": 2502950912 + }, + { + "epoch": 0.76, + "learning_rate": 0.0001953540621865597, + "loss": 1.2323, + "theoretical_loss": 3.396666186761056, + "tokens_seen": 2503016448 + }, + { + "epoch": 0.76, + "learning_rate": 0.00019533801404212638, + "loss": 1.2752, + "theoretical_loss": 3.3966591540650195, + "tokens_seen": 2503081984 + }, + { + "epoch": 0.76, + "learning_rate": 0.00019532196589769308, + "loss": 1.3087, + "theoretical_loss": 3.3966521216046663, + "tokens_seen": 2503147520 + }, + { + "epoch": 0.76, + "learning_rate": 0.00019530591775325978, + "loss": 1.3116, + "theoretical_loss": 3.396645089379983, + "tokens_seen": 2503213056 + }, + { + "epoch": 0.76, + "learning_rate": 0.00019528986960882648, + "loss": 1.3633, + "theoretical_loss": 3.3966380573909554, + "tokens_seen": 2503278592 + }, + { + "epoch": 0.76, + "learning_rate": 0.00019527382146439318, + "loss": 1.3012, + "theoretical_loss": 3.3966310256375687, + "tokens_seen": 2503344128 + }, + { + "epoch": 0.76, + "learning_rate": 0.00019525777331995988, + "loss": 1.2388, + "theoretical_loss": 3.3966239941198095, + "tokens_seen": 2503409664 + }, + { + "epoch": 0.76, + "objective/train/docs_used": 911421, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.3977479934692383, + "objective/train/theoretical_loss": 3.3966169628376637, + "objective/train/tokens_used": 2523935200, + "theoretical_loss": 3.3966169628376637, + "tokens_seen": 2503475200 + }, + { + "epoch": 0.76, + "learning_rate": 0.00019524172517552659, + "loss": 1.3339, + "theoretical_loss": 3.3966169628376637, + "tokens_seen": 2503475200 + }, + { + "epoch": 0.76, + "learning_rate": 0.00019522567703109329, + "loss": 1.2799, + "theoretical_loss": 3.3966099317911174, + "tokens_seen": 2503540736 + }, + { + "epoch": 0.76, + "learning_rate": 0.00019520962888665999, + "loss": 1.2085, + "theoretical_loss": 3.396602900980156, + "tokens_seen": 2503606272 + }, + { + "epoch": 0.76, + "learning_rate": 0.0001951935807422267, + "loss": 1.2138, + "theoretical_loss": 3.3965958704047656, + "tokens_seen": 2503671808 + }, + { + "epoch": 0.76, + "learning_rate": 0.0001951775325977934, + "loss": 1.238, + "theoretical_loss": 3.3965888400649322, + "tokens_seen": 2503737344 + }, + { + "epoch": 0.76, + "learning_rate": 0.0001951614844533601, + "loss": 1.3157, + "theoretical_loss": 3.396581809960642, + "tokens_seen": 2503802880 + }, + { + "epoch": 0.76, + "learning_rate": 0.0001951454363089268, + "loss": 1.2708, + "theoretical_loss": 3.3965747800918806, + "tokens_seen": 2503868416 + }, + { + "epoch": 0.76, + "learning_rate": 0.0001951293881644935, + "loss": 1.3089, + "theoretical_loss": 3.396567750458634, + "tokens_seen": 2503933952 + }, + { + "epoch": 0.76, + "learning_rate": 0.0001951133400200602, + "loss": 1.3848, + "theoretical_loss": 3.3965607210608884, + "tokens_seen": 2503999488 + }, + { + "epoch": 0.76, + "learning_rate": 0.0001950972918756269, + "loss": 1.2439, + "theoretical_loss": 3.3965536918986294, + "tokens_seen": 2504065024 + }, + { + "epoch": 0.76, + "learning_rate": 0.0001950812437311936, + "loss": 1.2322, + "theoretical_loss": 3.3965466629718435, + "tokens_seen": 2504130560 + }, + { + "epoch": 0.76, + "learning_rate": 0.0001950651955867603, + "loss": 1.1831, + "theoretical_loss": 3.396539634280516, + "tokens_seen": 2504196096 + }, + { + "epoch": 0.76, + "learning_rate": 0.000195049147442327, + "loss": 1.2533, + "theoretical_loss": 3.396532605824633, + "tokens_seen": 2504261632 + }, + { + "epoch": 0.76, + "learning_rate": 0.0001950330992978937, + "loss": 1.2453, + "theoretical_loss": 3.3965255776041805, + "tokens_seen": 2504327168 + }, + { + "epoch": 0.76, + "learning_rate": 0.0001950170511534604, + "loss": 1.2718, + "theoretical_loss": 3.3965185496191443, + "tokens_seen": 2504392704 + }, + { + "epoch": 0.76, + "learning_rate": 0.0001950010030090271, + "loss": 1.3493, + "theoretical_loss": 3.3965115218695114, + "tokens_seen": 2504458240 + }, + { + "epoch": 0.76, + "learning_rate": 0.0001949849548645938, + "loss": 1.2758, + "theoretical_loss": 3.3965044943552662, + "tokens_seen": 2504523776 + }, + { + "epoch": 0.76, + "learning_rate": 0.0001949689067201605, + "loss": 1.2994, + "theoretical_loss": 3.396497467076396, + "tokens_seen": 2504589312 + }, + { + "epoch": 0.76, + "learning_rate": 0.00019495285857572717, + "loss": 1.1906, + "theoretical_loss": 3.3964904400328857, + "tokens_seen": 2504654848 + }, + { + "epoch": 0.76, + "learning_rate": 0.00019493681043129387, + "loss": 1.3032, + "theoretical_loss": 3.396483413224722, + "tokens_seen": 2504720384 + }, + { + "epoch": 0.76, + "learning_rate": 0.00019492076228686057, + "loss": 1.3369, + "theoretical_loss": 3.3964763866518903, + "tokens_seen": 2504785920 + }, + { + "epoch": 0.76, + "learning_rate": 0.00019490471414242727, + "loss": 1.347, + "theoretical_loss": 3.3964693603143767, + "tokens_seen": 2504851456 + }, + { + "epoch": 0.76, + "learning_rate": 0.000194888665997994, + "loss": 1.2728, + "theoretical_loss": 3.396462334212168, + "tokens_seen": 2504916992 + }, + { + "epoch": 0.76, + "learning_rate": 0.0001948726178535607, + "loss": 1.2096, + "theoretical_loss": 3.3964553083452484, + "tokens_seen": 2504982528 + }, + { + "epoch": 0.76, + "learning_rate": 0.0001948565697091274, + "loss": 1.134, + "theoretical_loss": 3.3964482827136058, + "tokens_seen": 2505048064 + }, + { + "epoch": 0.76, + "learning_rate": 0.0001948405215646941, + "loss": 1.3895, + "theoretical_loss": 3.3964412573172247, + "tokens_seen": 2505113600 + }, + { + "epoch": 0.76, + "learning_rate": 0.0001948244734202608, + "loss": 1.2798, + "theoretical_loss": 3.396434232156092, + "tokens_seen": 2505179136 + }, + { + "epoch": 0.76, + "learning_rate": 0.0001948084252758275, + "loss": 1.242, + "theoretical_loss": 3.396427207230193, + "tokens_seen": 2505244672 + }, + { + "epoch": 0.76, + "learning_rate": 0.0001947923771313942, + "loss": 1.1861, + "theoretical_loss": 3.3964201825395146, + "tokens_seen": 2505310208 + }, + { + "epoch": 0.76, + "learning_rate": 0.0001947763289869609, + "loss": 1.3213, + "theoretical_loss": 3.3964131580840418, + "tokens_seen": 2505375744 + }, + { + "epoch": 0.76, + "learning_rate": 0.0001947602808425276, + "loss": 1.2654, + "theoretical_loss": 3.396406133863761, + "tokens_seen": 2505441280 + }, + { + "epoch": 0.76, + "learning_rate": 0.0001947442326980943, + "loss": 1.3825, + "theoretical_loss": 3.396399109878658, + "tokens_seen": 2505506816 + }, + { + "epoch": 0.76, + "learning_rate": 0.000194728184553661, + "loss": 1.3324, + "theoretical_loss": 3.396392086128719, + "tokens_seen": 2505572352 + }, + { + "epoch": 0.76, + "learning_rate": 0.0001947121364092277, + "loss": 1.3497, + "theoretical_loss": 3.3963850626139296, + "tokens_seen": 2505637888 + }, + { + "epoch": 0.76, + "learning_rate": 0.0001946960882647944, + "loss": 1.2001, + "theoretical_loss": 3.396378039334276, + "tokens_seen": 2505703424 + }, + { + "epoch": 0.76, + "learning_rate": 0.0001946800401203611, + "loss": 1.3412, + "theoretical_loss": 3.3963710162897445, + "tokens_seen": 2505768960 + }, + { + "epoch": 0.76, + "learning_rate": 0.0001946639919759278, + "loss": 1.2589, + "theoretical_loss": 3.3963639934803207, + "tokens_seen": 2505834496 + }, + { + "epoch": 0.76, + "learning_rate": 0.00019464794383149448, + "loss": 1.2577, + "theoretical_loss": 3.3963569709059906, + "tokens_seen": 2505900032 + }, + { + "epoch": 0.76, + "learning_rate": 0.00019463189568706118, + "loss": 1.3083, + "theoretical_loss": 3.3963499485667406, + "tokens_seen": 2505965568 + }, + { + "epoch": 0.76, + "learning_rate": 0.00019461584754262788, + "loss": 1.3036, + "theoretical_loss": 3.396342926462556, + "tokens_seen": 2506031104 + }, + { + "epoch": 0.76, + "learning_rate": 0.00019459979939819458, + "loss": 1.2669, + "theoretical_loss": 3.396335904593423, + "tokens_seen": 2506096640 + }, + { + "epoch": 0.76, + "learning_rate": 0.00019458375125376128, + "loss": 1.3913, + "theoretical_loss": 3.3963288829593283, + "tokens_seen": 2506162176 + }, + { + "epoch": 0.76, + "learning_rate": 0.00019456770310932798, + "loss": 1.3311, + "theoretical_loss": 3.396321861560257, + "tokens_seen": 2506227712 + }, + { + "epoch": 0.76, + "learning_rate": 0.00019455165496489469, + "loss": 1.1209, + "theoretical_loss": 3.396314840396195, + "tokens_seen": 2506293248 + }, + { + "epoch": 0.76, + "learning_rate": 0.00019453560682046139, + "loss": 1.1422, + "theoretical_loss": 3.396307819467129, + "tokens_seen": 2506358784 + }, + { + "epoch": 0.76, + "learning_rate": 0.0001945195586760281, + "loss": 1.4025, + "theoretical_loss": 3.3963007987730447, + "tokens_seen": 2506424320 + }, + { + "epoch": 0.76, + "learning_rate": 0.0001945035105315948, + "loss": 1.2874, + "theoretical_loss": 3.396293778313928, + "tokens_seen": 2506489856 + }, + { + "epoch": 0.76, + "learning_rate": 0.0001944874623871615, + "loss": 1.3252, + "theoretical_loss": 3.396286758089765, + "tokens_seen": 2506555392 + }, + { + "epoch": 0.76, + "learning_rate": 0.0001944714142427282, + "loss": 1.2709, + "theoretical_loss": 3.3962797381005414, + "tokens_seen": 2506620928 + }, + { + "epoch": 0.76, + "learning_rate": 0.0001944553660982949, + "loss": 1.2932, + "theoretical_loss": 3.396272718346244, + "tokens_seen": 2506686464 + }, + { + "epoch": 0.76, + "objective/train/docs_used": 912553, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.3604350090026855, + "objective/train/theoretical_loss": 3.396265698826858, + "objective/train/tokens_used": 2527212000, + "theoretical_loss": 3.396265698826858, + "tokens_seen": 2506752000 + }, + { + "epoch": 0.76, + "learning_rate": 0.0001944393179538616, + "loss": 1.3474, + "theoretical_loss": 3.396265698826858, + "tokens_seen": 2506752000 + }, + { + "epoch": 0.76, + "learning_rate": 0.0001944232698094283, + "loss": 1.2067, + "theoretical_loss": 3.3962586795423695, + "tokens_seen": 2506817536 + }, + { + "epoch": 0.76, + "learning_rate": 0.000194407221664995, + "loss": 1.3999, + "theoretical_loss": 3.3962516604927644, + "tokens_seen": 2506883072 + }, + { + "epoch": 0.76, + "learning_rate": 0.0001943911735205617, + "loss": 1.1351, + "theoretical_loss": 3.396244641678029, + "tokens_seen": 2506948608 + }, + { + "epoch": 0.76, + "learning_rate": 0.0001943751253761284, + "loss": 1.2921, + "theoretical_loss": 3.3962376230981497, + "tokens_seen": 2507014144 + }, + { + "epoch": 0.76, + "learning_rate": 0.0001943590772316951, + "loss": 1.2327, + "theoretical_loss": 3.3962306047531117, + "tokens_seen": 2507079680 + }, + { + "epoch": 0.76, + "learning_rate": 0.0001943430290872618, + "loss": 1.352, + "theoretical_loss": 3.3962235866429014, + "tokens_seen": 2507145216 + }, + { + "epoch": 0.76, + "learning_rate": 0.0001943269809428285, + "loss": 1.1947, + "theoretical_loss": 3.3962165687675046, + "tokens_seen": 2507210752 + }, + { + "epoch": 0.76, + "learning_rate": 0.0001943109327983952, + "loss": 1.2196, + "theoretical_loss": 3.3962095511269075, + "tokens_seen": 2507276288 + }, + { + "epoch": 0.76, + "learning_rate": 0.0001942948846539619, + "loss": 1.2662, + "theoretical_loss": 3.3962025337210964, + "tokens_seen": 2507341824 + }, + { + "epoch": 0.76, + "learning_rate": 0.0001942788365095286, + "loss": 1.2232, + "theoretical_loss": 3.3961955165500566, + "tokens_seen": 2507407360 + }, + { + "epoch": 0.76, + "learning_rate": 0.0001942627883650953, + "loss": 1.256, + "theoretical_loss": 3.396188499613775, + "tokens_seen": 2507472896 + }, + { + "epoch": 0.76, + "learning_rate": 0.000194246740220662, + "loss": 1.3052, + "theoretical_loss": 3.3961814829122363, + "tokens_seen": 2507538432 + }, + { + "epoch": 0.76, + "learning_rate": 0.0001942306920762287, + "loss": 1.2469, + "theoretical_loss": 3.3961744664454274, + "tokens_seen": 2507603968 + }, + { + "epoch": 0.76, + "learning_rate": 0.0001942146439317954, + "loss": 1.3443, + "theoretical_loss": 3.3961674502133348, + "tokens_seen": 2507669504 + }, + { + "epoch": 0.76, + "learning_rate": 0.0001941985957873621, + "loss": 1.3393, + "theoretical_loss": 3.3961604342159437, + "tokens_seen": 2507735040 + }, + { + "epoch": 0.76, + "learning_rate": 0.0001941825476429288, + "loss": 1.2783, + "theoretical_loss": 3.39615341845324, + "tokens_seen": 2507800576 + }, + { + "epoch": 0.76, + "learning_rate": 0.0001941664994984955, + "loss": 1.3107, + "theoretical_loss": 3.3961464029252104, + "tokens_seen": 2507866112 + }, + { + "epoch": 0.76, + "learning_rate": 0.0001941504513540622, + "loss": 1.2881, + "theoretical_loss": 3.3961393876318406, + "tokens_seen": 2507931648 + }, + { + "epoch": 0.76, + "learning_rate": 0.0001941344032096289, + "loss": 1.2241, + "theoretical_loss": 3.3961323725731165, + "tokens_seen": 2507997184 + }, + { + "epoch": 0.76, + "learning_rate": 0.0001941183550651956, + "loss": 1.4035, + "theoretical_loss": 3.3961253577490242, + "tokens_seen": 2508062720 + }, + { + "epoch": 0.76, + "learning_rate": 0.0001941023069207623, + "loss": 1.2745, + "theoretical_loss": 3.3961183431595496, + "tokens_seen": 2508128256 + }, + { + "epoch": 0.76, + "learning_rate": 0.000194086258776329, + "loss": 1.314, + "theoretical_loss": 3.396111328804679, + "tokens_seen": 2508193792 + }, + { + "epoch": 0.76, + "learning_rate": 0.0001940702106318957, + "loss": 1.3438, + "theoretical_loss": 3.3961043146843983, + "tokens_seen": 2508259328 + }, + { + "epoch": 0.76, + "learning_rate": 0.0001940541624874624, + "loss": 1.3005, + "theoretical_loss": 3.3960973007986937, + "tokens_seen": 2508324864 + }, + { + "epoch": 0.76, + "learning_rate": 0.0001940381143430291, + "loss": 1.2044, + "theoretical_loss": 3.3960902871475507, + "tokens_seen": 2508390400 + }, + { + "epoch": 0.76, + "learning_rate": 0.0001940220661985958, + "loss": 1.2942, + "theoretical_loss": 3.3960832737309556, + "tokens_seen": 2508455936 + }, + { + "epoch": 0.76, + "learning_rate": 0.0001940060180541625, + "loss": 1.2882, + "theoretical_loss": 3.3960762605488948, + "tokens_seen": 2508521472 + }, + { + "epoch": 0.76, + "learning_rate": 0.0001939899699097292, + "loss": 1.2566, + "theoretical_loss": 3.396069247601354, + "tokens_seen": 2508587008 + }, + { + "epoch": 0.76, + "learning_rate": 0.0001939739217652959, + "loss": 1.3058, + "theoretical_loss": 3.396062234888319, + "tokens_seen": 2508652544 + }, + { + "epoch": 0.76, + "learning_rate": 0.0001939578736208626, + "loss": 1.2734, + "theoretical_loss": 3.3960552224097764, + "tokens_seen": 2508718080 + }, + { + "epoch": 0.76, + "learning_rate": 0.00019394182547642928, + "loss": 1.3401, + "theoretical_loss": 3.396048210165712, + "tokens_seen": 2508783616 + }, + { + "epoch": 0.76, + "learning_rate": 0.00019392577733199598, + "loss": 1.3141, + "theoretical_loss": 3.3960411981561114, + "tokens_seen": 2508849152 + }, + { + "epoch": 0.76, + "learning_rate": 0.00019390972918756268, + "loss": 1.3573, + "theoretical_loss": 3.396034186380961, + "tokens_seen": 2508914688 + }, + { + "epoch": 0.76, + "learning_rate": 0.00019389368104312938, + "loss": 1.2568, + "theoretical_loss": 3.3960271748402473, + "tokens_seen": 2508980224 + }, + { + "epoch": 0.76, + "learning_rate": 0.00019387763289869608, + "loss": 1.2235, + "theoretical_loss": 3.3960201635339553, + "tokens_seen": 2509045760 + }, + { + "epoch": 0.76, + "learning_rate": 0.00019386158475426279, + "loss": 1.2594, + "theoretical_loss": 3.396013152462072, + "tokens_seen": 2509111296 + }, + { + "epoch": 0.76, + "learning_rate": 0.00019384553660982949, + "loss": 1.2709, + "theoretical_loss": 3.3960061416245826, + "tokens_seen": 2509176832 + }, + { + "epoch": 0.76, + "learning_rate": 0.0001938294884653962, + "loss": 1.1423, + "theoretical_loss": 3.395999131021474, + "tokens_seen": 2509242368 + }, + { + "epoch": 0.76, + "learning_rate": 0.0001938134403209629, + "loss": 1.2876, + "theoretical_loss": 3.3959921206527315, + "tokens_seen": 2509307904 + }, + { + "epoch": 0.76, + "learning_rate": 0.0001937973921765296, + "loss": 1.3704, + "theoretical_loss": 3.3959851105183416, + "tokens_seen": 2509373440 + }, + { + "epoch": 0.76, + "learning_rate": 0.0001937813440320963, + "loss": 1.344, + "theoretical_loss": 3.3959781006182905, + "tokens_seen": 2509438976 + }, + { + "epoch": 0.76, + "learning_rate": 0.000193765295887663, + "loss": 1.3882, + "theoretical_loss": 3.395971090952564, + "tokens_seen": 2509504512 + }, + { + "epoch": 0.76, + "learning_rate": 0.0001937492477432297, + "loss": 1.2659, + "theoretical_loss": 3.395964081521148, + "tokens_seen": 2509570048 + }, + { + "epoch": 0.76, + "learning_rate": 0.0001937331995987964, + "loss": 1.2203, + "theoretical_loss": 3.395957072324028, + "tokens_seen": 2509635584 + }, + { + "epoch": 0.76, + "learning_rate": 0.0001937171514543631, + "loss": 1.272, + "theoretical_loss": 3.3959500633611914, + "tokens_seen": 2509701120 + }, + { + "epoch": 0.76, + "learning_rate": 0.0001937011033099298, + "loss": 1.3558, + "theoretical_loss": 3.3959430546326237, + "tokens_seen": 2509766656 + }, + { + "epoch": 0.76, + "learning_rate": 0.00019368505516549652, + "loss": 1.3067, + "theoretical_loss": 3.3959360461383103, + "tokens_seen": 2509832192 + }, + { + "epoch": 0.76, + "learning_rate": 0.00019366900702106322, + "loss": 1.162, + "theoretical_loss": 3.395929037878238, + "tokens_seen": 2509897728 + }, + { + "epoch": 0.76, + "learning_rate": 0.00019365295887662992, + "loss": 1.3494, + "theoretical_loss": 3.395922029852393, + "tokens_seen": 2509963264 + }, + { + "epoch": 0.76, + "objective/train/docs_used": 913305, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.405246615409851, + "objective/train/theoretical_loss": 3.395915022060761, + "objective/train/tokens_used": 2530488800, + "theoretical_loss": 3.395915022060761, + "tokens_seen": 2510028800 + }, + { + "epoch": 0.76, + "learning_rate": 0.00019363691073219662, + "loss": 1.3546, + "theoretical_loss": 3.395915022060761, + "tokens_seen": 2510028800 + }, + { + "epoch": 0.76, + "learning_rate": 0.0001936208625877633, + "loss": 1.3685, + "theoretical_loss": 3.3959080145033274, + "tokens_seen": 2510094336 + }, + { + "epoch": 0.76, + "learning_rate": 0.00019360481444333, + "loss": 1.2814, + "theoretical_loss": 3.395901007180079, + "tokens_seen": 2510159872 + }, + { + "epoch": 0.76, + "learning_rate": 0.0001935887662988967, + "loss": 1.2967, + "theoretical_loss": 3.395894000091002, + "tokens_seen": 2510225408 + }, + { + "epoch": 0.76, + "learning_rate": 0.0001935727181544634, + "loss": 1.3062, + "theoretical_loss": 3.3958869932360822, + "tokens_seen": 2510290944 + }, + { + "epoch": 0.76, + "learning_rate": 0.0001935566700100301, + "loss": 1.431, + "theoretical_loss": 3.3958799866153058, + "tokens_seen": 2510356480 + }, + { + "epoch": 0.76, + "learning_rate": 0.0001935406218655968, + "loss": 1.1913, + "theoretical_loss": 3.3958729802286585, + "tokens_seen": 2510422016 + }, + { + "epoch": 0.76, + "learning_rate": 0.0001935245737211635, + "loss": 1.3654, + "theoretical_loss": 3.3958659740761266, + "tokens_seen": 2510487552 + }, + { + "epoch": 0.76, + "learning_rate": 0.0001935085255767302, + "loss": 1.2994, + "theoretical_loss": 3.3958589681576963, + "tokens_seen": 2510553088 + }, + { + "epoch": 0.76, + "learning_rate": 0.0001934924774322969, + "loss": 1.2345, + "theoretical_loss": 3.395851962473354, + "tokens_seen": 2510618624 + }, + { + "epoch": 0.76, + "learning_rate": 0.0001934764292878636, + "loss": 1.2337, + "theoretical_loss": 3.3958449570230846, + "tokens_seen": 2510684160 + }, + { + "epoch": 0.76, + "learning_rate": 0.0001934603811434303, + "loss": 1.1794, + "theoretical_loss": 3.3958379518068753, + "tokens_seen": 2510749696 + }, + { + "epoch": 0.76, + "learning_rate": 0.000193444332998997, + "loss": 1.2775, + "theoretical_loss": 3.3958309468247116, + "tokens_seen": 2510815232 + }, + { + "epoch": 0.76, + "learning_rate": 0.0001934282848545637, + "loss": 1.3644, + "theoretical_loss": 3.3958239420765794, + "tokens_seen": 2510880768 + }, + { + "epoch": 0.76, + "learning_rate": 0.0001934122367101304, + "loss": 1.3474, + "theoretical_loss": 3.395816937562466, + "tokens_seen": 2510946304 + }, + { + "epoch": 0.76, + "learning_rate": 0.0001933961885656971, + "loss": 1.2072, + "theoretical_loss": 3.3958099332823557, + "tokens_seen": 2511011840 + }, + { + "epoch": 0.76, + "learning_rate": 0.0001933801404212638, + "loss": 1.2245, + "theoretical_loss": 3.3958029292362357, + "tokens_seen": 2511077376 + }, + { + "epoch": 0.76, + "learning_rate": 0.0001933640922768305, + "loss": 1.3635, + "theoretical_loss": 3.395795925424092, + "tokens_seen": 2511142912 + }, + { + "epoch": 0.76, + "learning_rate": 0.0001933480441323972, + "loss": 1.2057, + "theoretical_loss": 3.3957889218459103, + "tokens_seen": 2511208448 + }, + { + "epoch": 0.76, + "learning_rate": 0.0001933319959879639, + "loss": 1.3711, + "theoretical_loss": 3.3957819185016773, + "tokens_seen": 2511273984 + }, + { + "epoch": 0.76, + "learning_rate": 0.0001933159478435306, + "loss": 1.3193, + "theoretical_loss": 3.395774915391378, + "tokens_seen": 2511339520 + }, + { + "epoch": 0.76, + "learning_rate": 0.0001932998996990973, + "loss": 1.2945, + "theoretical_loss": 3.3957679125149998, + "tokens_seen": 2511405056 + }, + { + "epoch": 0.76, + "learning_rate": 0.000193283851554664, + "loss": 1.4351, + "theoretical_loss": 3.3957609098725277, + "tokens_seen": 2511470592 + }, + { + "epoch": 0.76, + "learning_rate": 0.0001932678034102307, + "loss": 1.3473, + "theoretical_loss": 3.3957539074639485, + "tokens_seen": 2511536128 + }, + { + "epoch": 0.76, + "learning_rate": 0.0001932517552657974, + "loss": 1.3485, + "theoretical_loss": 3.395746905289248, + "tokens_seen": 2511601664 + }, + { + "epoch": 0.76, + "learning_rate": 0.00019323570712136408, + "loss": 1.2447, + "theoretical_loss": 3.395739903348412, + "tokens_seen": 2511667200 + }, + { + "epoch": 0.76, + "learning_rate": 0.00019321965897693078, + "loss": 1.4376, + "theoretical_loss": 3.395732901641427, + "tokens_seen": 2511732736 + }, + { + "epoch": 0.76, + "learning_rate": 0.00019320361083249748, + "loss": 1.2322, + "theoretical_loss": 3.395725900168279, + "tokens_seen": 2511798272 + }, + { + "epoch": 0.76, + "learning_rate": 0.00019318756268806419, + "loss": 1.2909, + "theoretical_loss": 3.395718898928954, + "tokens_seen": 2511863808 + }, + { + "epoch": 0.76, + "learning_rate": 0.00019317151454363089, + "loss": 1.3107, + "theoretical_loss": 3.395711897923438, + "tokens_seen": 2511929344 + }, + { + "epoch": 0.76, + "learning_rate": 0.00019315546639919759, + "loss": 1.2808, + "theoretical_loss": 3.3957048971517176, + "tokens_seen": 2511994880 + }, + { + "epoch": 0.76, + "learning_rate": 0.0001931394182547643, + "loss": 1.2461, + "theoretical_loss": 3.395697896613778, + "tokens_seen": 2512060416 + }, + { + "epoch": 0.76, + "learning_rate": 0.000193123370110331, + "loss": 1.4177, + "theoretical_loss": 3.395690896309606, + "tokens_seen": 2512125952 + }, + { + "epoch": 0.76, + "learning_rate": 0.0001931073219658977, + "loss": 1.3566, + "theoretical_loss": 3.3956838962391878, + "tokens_seen": 2512191488 + }, + { + "epoch": 0.76, + "learning_rate": 0.00019309127382146442, + "loss": 1.2646, + "theoretical_loss": 3.395676896402509, + "tokens_seen": 2512257024 + }, + { + "epoch": 0.76, + "learning_rate": 0.00019307522567703112, + "loss": 1.2376, + "theoretical_loss": 3.395669896799556, + "tokens_seen": 2512322560 + }, + { + "epoch": 0.76, + "learning_rate": 0.00019305917753259782, + "loss": 1.2971, + "theoretical_loss": 3.3956628974303147, + "tokens_seen": 2512388096 + }, + { + "epoch": 0.76, + "learning_rate": 0.00019304312938816452, + "loss": 1.1956, + "theoretical_loss": 3.3956558982947715, + "tokens_seen": 2512453632 + }, + { + "epoch": 0.76, + "learning_rate": 0.00019302708124373122, + "loss": 1.3478, + "theoretical_loss": 3.395648899392912, + "tokens_seen": 2512519168 + }, + { + "epoch": 0.76, + "learning_rate": 0.00019301103309929792, + "loss": 1.2708, + "theoretical_loss": 3.395641900724723, + "tokens_seen": 2512584704 + }, + { + "epoch": 0.76, + "learning_rate": 0.00019299498495486462, + "loss": 1.2196, + "theoretical_loss": 3.39563490229019, + "tokens_seen": 2512650240 + }, + { + "epoch": 0.76, + "learning_rate": 0.00019297893681043132, + "loss": 1.3271, + "theoretical_loss": 3.3956279040892996, + "tokens_seen": 2512715776 + }, + { + "epoch": 0.76, + "learning_rate": 0.00019296288866599802, + "loss": 1.1941, + "theoretical_loss": 3.395620906122037, + "tokens_seen": 2512781312 + }, + { + "epoch": 0.76, + "learning_rate": 0.00019294684052156472, + "loss": 1.3656, + "theoretical_loss": 3.3956139083883894, + "tokens_seen": 2512846848 + }, + { + "epoch": 0.76, + "learning_rate": 0.0001929307923771314, + "loss": 1.2512, + "theoretical_loss": 3.3956069108883424, + "tokens_seen": 2512912384 + }, + { + "epoch": 0.76, + "learning_rate": 0.0001929147442326981, + "loss": 1.3096, + "theoretical_loss": 3.395599913621882, + "tokens_seen": 2512977920 + }, + { + "epoch": 0.76, + "learning_rate": 0.0001928986960882648, + "loss": 1.3023, + "theoretical_loss": 3.3955929165889946, + "tokens_seen": 2513043456 + }, + { + "epoch": 0.76, + "learning_rate": 0.0001928826479438315, + "loss": 1.3104, + "theoretical_loss": 3.3955859197896663, + "tokens_seen": 2513108992 + }, + { + "epoch": 0.76, + "learning_rate": 0.0001928665997993982, + "loss": 1.2478, + "theoretical_loss": 3.3955789232238827, + "tokens_seen": 2513174528 + }, + { + "epoch": 0.76, + "learning_rate": 0.0001928505516549649, + "loss": 1.1813, + "theoretical_loss": 3.395571926891631, + "tokens_seen": 2513240064 + }, + { + "epoch": 0.76, + "objective/train/docs_used": 914593, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.444047451019287, + "objective/train/theoretical_loss": 3.395564930792896, + "objective/train/tokens_used": 2533765600, + "theoretical_loss": 3.395564930792896, + "tokens_seen": 2513305600 + }, + { + "epoch": 0.76, + "learning_rate": 0.0001928345035105316, + "loss": 1.3548, + "theoretical_loss": 3.395564930792896, + "tokens_seen": 2513305600 + }, + { + "epoch": 0.76, + "learning_rate": 0.0001928184553660983, + "loss": 1.4225, + "theoretical_loss": 3.395557934927665, + "tokens_seen": 2513371136 + }, + { + "epoch": 0.76, + "learning_rate": 0.000192802407221665, + "loss": 1.2981, + "theoretical_loss": 3.395550939295923, + "tokens_seen": 2513436672 + }, + { + "epoch": 0.76, + "learning_rate": 0.0001927863590772317, + "loss": 1.3322, + "theoretical_loss": 3.395543943897657, + "tokens_seen": 2513502208 + }, + { + "epoch": 0.76, + "learning_rate": 0.0001927703109327984, + "loss": 1.4303, + "theoretical_loss": 3.395536948732852, + "tokens_seen": 2513567744 + }, + { + "epoch": 0.76, + "learning_rate": 0.0001927542627883651, + "loss": 1.2383, + "theoretical_loss": 3.395529953801496, + "tokens_seen": 2513633280 + }, + { + "epoch": 0.76, + "learning_rate": 0.0001927382146439318, + "loss": 1.3815, + "theoretical_loss": 3.3955229591035736, + "tokens_seen": 2513698816 + }, + { + "epoch": 0.76, + "learning_rate": 0.0001927221664994985, + "loss": 1.2649, + "theoretical_loss": 3.3955159646390713, + "tokens_seen": 2513764352 + }, + { + "epoch": 0.76, + "learning_rate": 0.0001927061183550652, + "loss": 1.293, + "theoretical_loss": 3.3955089704079753, + "tokens_seen": 2513829888 + }, + { + "epoch": 0.76, + "learning_rate": 0.0001926900702106319, + "loss": 1.3139, + "theoretical_loss": 3.395501976410272, + "tokens_seen": 2513895424 + }, + { + "epoch": 0.76, + "learning_rate": 0.0001926740220661986, + "loss": 1.2832, + "theoretical_loss": 3.395494982645947, + "tokens_seen": 2513960960 + }, + { + "epoch": 0.76, + "learning_rate": 0.0001926579739217653, + "loss": 1.2367, + "theoretical_loss": 3.3954879891149865, + "tokens_seen": 2514026496 + }, + { + "epoch": 0.76, + "learning_rate": 0.000192641925777332, + "loss": 1.1928, + "theoretical_loss": 3.395480995817377, + "tokens_seen": 2514092032 + }, + { + "epoch": 0.76, + "learning_rate": 0.0001926258776328987, + "loss": 1.1703, + "theoretical_loss": 3.3954740027531045, + "tokens_seen": 2514157568 + }, + { + "epoch": 0.76, + "learning_rate": 0.0001926098294884654, + "loss": 1.3555, + "theoretical_loss": 3.395467009922155, + "tokens_seen": 2514223104 + }, + { + "epoch": 0.76, + "learning_rate": 0.0001925937813440321, + "loss": 1.3264, + "theoretical_loss": 3.395460017324515, + "tokens_seen": 2514288640 + }, + { + "epoch": 0.76, + "learning_rate": 0.0001925777331995988, + "loss": 1.2204, + "theoretical_loss": 3.39545302496017, + "tokens_seen": 2514354176 + }, + { + "epoch": 0.76, + "learning_rate": 0.0001925616850551655, + "loss": 1.1747, + "theoretical_loss": 3.3954460328291067, + "tokens_seen": 2514419712 + }, + { + "epoch": 0.76, + "learning_rate": 0.00019254563691073218, + "loss": 1.2936, + "theoretical_loss": 3.395439040931311, + "tokens_seen": 2514485248 + }, + { + "epoch": 0.76, + "learning_rate": 0.00019252958876629888, + "loss": 1.2585, + "theoretical_loss": 3.395432049266769, + "tokens_seen": 2514550784 + }, + { + "epoch": 0.76, + "learning_rate": 0.00019251354062186558, + "loss": 1.2128, + "theoretical_loss": 3.3954250578354666, + "tokens_seen": 2514616320 + }, + { + "epoch": 0.76, + "learning_rate": 0.00019249749247743229, + "loss": 1.2273, + "theoretical_loss": 3.3954180666373905, + "tokens_seen": 2514681856 + }, + { + "epoch": 0.76, + "learning_rate": 0.000192481444332999, + "loss": 1.2607, + "theoretical_loss": 3.395411075672526, + "tokens_seen": 2514747392 + }, + { + "epoch": 0.76, + "learning_rate": 0.00019246539618856571, + "loss": 1.3557, + "theoretical_loss": 3.3954040849408607, + "tokens_seen": 2514812928 + }, + { + "epoch": 0.76, + "learning_rate": 0.00019244934804413241, + "loss": 1.2963, + "theoretical_loss": 3.3953970944423797, + "tokens_seen": 2514878464 + }, + { + "epoch": 0.76, + "learning_rate": 0.00019243329989969911, + "loss": 1.2741, + "theoretical_loss": 3.395390104177069, + "tokens_seen": 2514944000 + }, + { + "epoch": 0.76, + "learning_rate": 0.00019241725175526582, + "loss": 1.2708, + "theoretical_loss": 3.395383114144915, + "tokens_seen": 2515009536 + }, + { + "epoch": 0.76, + "learning_rate": 0.00019240120361083252, + "loss": 1.3851, + "theoretical_loss": 3.395376124345904, + "tokens_seen": 2515075072 + }, + { + "epoch": 0.76, + "learning_rate": 0.00019238515546639922, + "loss": 1.3615, + "theoretical_loss": 3.3953691347800223, + "tokens_seen": 2515140608 + }, + { + "epoch": 0.76, + "learning_rate": 0.00019236910732196592, + "loss": 1.3108, + "theoretical_loss": 3.3953621454472556, + "tokens_seen": 2515206144 + }, + { + "epoch": 0.76, + "learning_rate": 0.00019235305917753262, + "loss": 1.253, + "theoretical_loss": 3.3953551563475903, + "tokens_seen": 2515271680 + }, + { + "epoch": 0.76, + "learning_rate": 0.00019233701103309932, + "loss": 1.2589, + "theoretical_loss": 3.3953481674810124, + "tokens_seen": 2515337216 + }, + { + "epoch": 0.76, + "learning_rate": 0.00019232096288866602, + "loss": 1.2516, + "theoretical_loss": 3.395341178847508, + "tokens_seen": 2515402752 + }, + { + "epoch": 0.76, + "learning_rate": 0.00019230491474423272, + "loss": 1.3733, + "theoretical_loss": 3.3953341904470635, + "tokens_seen": 2515468288 + }, + { + "epoch": 0.76, + "learning_rate": 0.00019228886659979942, + "loss": 1.3193, + "theoretical_loss": 3.3953272022796654, + "tokens_seen": 2515533824 + }, + { + "epoch": 0.76, + "learning_rate": 0.00019227281845536612, + "loss": 1.3066, + "theoretical_loss": 3.395320214345299, + "tokens_seen": 2515599360 + }, + { + "epoch": 0.76, + "learning_rate": 0.00019225677031093282, + "loss": 1.3521, + "theoretical_loss": 3.395313226643951, + "tokens_seen": 2515664896 + }, + { + "epoch": 0.76, + "learning_rate": 0.00019224072216649952, + "loss": 1.3327, + "theoretical_loss": 3.3953062391756075, + "tokens_seen": 2515730432 + }, + { + "epoch": 0.76, + "learning_rate": 0.0001922246740220662, + "loss": 1.3064, + "theoretical_loss": 3.3952992519402545, + "tokens_seen": 2515795968 + }, + { + "epoch": 0.76, + "learning_rate": 0.0001922086258776329, + "loss": 1.305, + "theoretical_loss": 3.3952922649378783, + "tokens_seen": 2515861504 + }, + { + "epoch": 0.76, + "learning_rate": 0.0001921925777331996, + "loss": 1.2095, + "theoretical_loss": 3.395285278168465, + "tokens_seen": 2515927040 + }, + { + "epoch": 0.76, + "learning_rate": 0.0001921765295887663, + "loss": 1.3063, + "theoretical_loss": 3.395278291632001, + "tokens_seen": 2515992576 + }, + { + "epoch": 0.76, + "learning_rate": 0.000192160481444333, + "loss": 1.2377, + "theoretical_loss": 3.395271305328472, + "tokens_seen": 2516058112 + }, + { + "epoch": 0.76, + "learning_rate": 0.0001921444332998997, + "loss": 1.2018, + "theoretical_loss": 3.3952643192578646, + "tokens_seen": 2516123648 + }, + { + "epoch": 0.76, + "learning_rate": 0.0001921283851554664, + "loss": 1.2034, + "theoretical_loss": 3.395257333420165, + "tokens_seen": 2516189184 + }, + { + "epoch": 0.76, + "learning_rate": 0.0001921123370110331, + "loss": 1.2653, + "theoretical_loss": 3.3952503478153586, + "tokens_seen": 2516254720 + }, + { + "epoch": 0.76, + "learning_rate": 0.0001920962888665998, + "loss": 1.4107, + "theoretical_loss": 3.3952433624434324, + "tokens_seen": 2516320256 + }, + { + "epoch": 0.76, + "learning_rate": 0.0001920802407221665, + "loss": 1.3868, + "theoretical_loss": 3.3952363773043723, + "tokens_seen": 2516385792 + }, + { + "epoch": 0.76, + "learning_rate": 0.0001920641925777332, + "loss": 1.3208, + "theoretical_loss": 3.3952293923981647, + "tokens_seen": 2516451328 + }, + { + "epoch": 0.76, + "learning_rate": 0.0001920481444332999, + "loss": 1.3414, + "theoretical_loss": 3.3952224077247957, + "tokens_seen": 2516516864 + }, + { + "epoch": 0.76, + "objective/train/docs_used": 916228, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.4000651836395264, + "objective/train/theoretical_loss": 3.3952154232842506, + "objective/train/tokens_used": 2537042400, + "theoretical_loss": 3.3952154232842506, + "tokens_seen": 2516582400 + }, + { + "epoch": 0.76, + "learning_rate": 0.0001920320962888666, + "loss": 1.3702, + "theoretical_loss": 3.3952154232842506, + "tokens_seen": 2516582400 + }, + { + "epoch": 0.76, + "learning_rate": 0.0001920160481444333, + "loss": 1.3114, + "theoretical_loss": 3.3952084390765167, + "tokens_seen": 2516647936 + }, + { + "epoch": 0.76, + "learning_rate": 0.000192, + "loss": 1.3131, + "theoretical_loss": 3.3952014551015797, + "tokens_seen": 2516713472 + }, + { + "epoch": 0.76, + "learning_rate": 0.0001919839518555667, + "loss": 1.2453, + "theoretical_loss": 3.395194471359426, + "tokens_seen": 2516779008 + }, + { + "epoch": 0.76, + "learning_rate": 0.0001919679037111334, + "loss": 1.2483, + "theoretical_loss": 3.3951874878500417, + "tokens_seen": 2516844544 + }, + { + "epoch": 0.76, + "learning_rate": 0.0001919518555667001, + "loss": 1.2093, + "theoretical_loss": 3.395180504573413, + "tokens_seen": 2516910080 + }, + { + "epoch": 0.76, + "learning_rate": 0.0001919358074222668, + "loss": 1.2748, + "theoretical_loss": 3.3951735215295256, + "tokens_seen": 2516975616 + }, + { + "epoch": 0.76, + "learning_rate": 0.0001919197592778335, + "loss": 1.2769, + "theoretical_loss": 3.3951665387183665, + "tokens_seen": 2517041152 + }, + { + "epoch": 0.76, + "learning_rate": 0.0001919037111334002, + "loss": 1.3429, + "theoretical_loss": 3.3951595561399213, + "tokens_seen": 2517106688 + }, + { + "epoch": 0.76, + "learning_rate": 0.0001918876629889669, + "loss": 1.2886, + "theoretical_loss": 3.395152573794176, + "tokens_seen": 2517172224 + }, + { + "epoch": 0.76, + "learning_rate": 0.0001918716148445336, + "loss": 1.2091, + "theoretical_loss": 3.3951455916811177, + "tokens_seen": 2517237760 + }, + { + "epoch": 0.76, + "learning_rate": 0.0001918555667001003, + "loss": 1.3211, + "theoretical_loss": 3.395138609800732, + "tokens_seen": 2517303296 + }, + { + "epoch": 0.76, + "learning_rate": 0.000191839518555667, + "loss": 1.345, + "theoretical_loss": 3.395131628153005, + "tokens_seen": 2517368832 + }, + { + "epoch": 0.76, + "learning_rate": 0.0001918234704112337, + "loss": 1.3021, + "theoretical_loss": 3.395124646737923, + "tokens_seen": 2517434368 + }, + { + "epoch": 0.76, + "learning_rate": 0.0001918074222668004, + "loss": 1.2864, + "theoretical_loss": 3.3951176655554716, + "tokens_seen": 2517499904 + }, + { + "epoch": 0.76, + "learning_rate": 0.0001917913741223671, + "loss": 1.2779, + "theoretical_loss": 3.3951106846056383, + "tokens_seen": 2517565440 + }, + { + "epoch": 0.76, + "learning_rate": 0.00019177532597793381, + "loss": 1.3647, + "theoretical_loss": 3.3951037038884087, + "tokens_seen": 2517630976 + }, + { + "epoch": 0.76, + "learning_rate": 0.00019175927783350051, + "loss": 1.3041, + "theoretical_loss": 3.395096723403768, + "tokens_seen": 2517696512 + }, + { + "epoch": 0.76, + "learning_rate": 0.00019174322968906722, + "loss": 1.2122, + "theoretical_loss": 3.395089743151704, + "tokens_seen": 2517762048 + }, + { + "epoch": 0.76, + "learning_rate": 0.00019172718154463392, + "loss": 1.2695, + "theoretical_loss": 3.3950827631322023, + "tokens_seen": 2517827584 + }, + { + "epoch": 0.76, + "learning_rate": 0.00019171113340020062, + "loss": 1.2206, + "theoretical_loss": 3.3950757833452485, + "tokens_seen": 2517893120 + }, + { + "epoch": 0.76, + "learning_rate": 0.00019169508525576732, + "loss": 1.3152, + "theoretical_loss": 3.3950688037908296, + "tokens_seen": 2517958656 + }, + { + "epoch": 0.76, + "learning_rate": 0.00019167903711133402, + "loss": 1.2112, + "theoretical_loss": 3.3950618244689315, + "tokens_seen": 2518024192 + }, + { + "epoch": 0.76, + "learning_rate": 0.00019166298896690072, + "loss": 1.212, + "theoretical_loss": 3.3950548453795406, + "tokens_seen": 2518089728 + }, + { + "epoch": 0.76, + "learning_rate": 0.00019164694082246742, + "loss": 1.3705, + "theoretical_loss": 3.3950478665226425, + "tokens_seen": 2518155264 + }, + { + "epoch": 0.76, + "learning_rate": 0.00019163089267803412, + "loss": 1.227, + "theoretical_loss": 3.3950408878982237, + "tokens_seen": 2518220800 + }, + { + "epoch": 0.76, + "learning_rate": 0.00019161484453360082, + "loss": 1.2483, + "theoretical_loss": 3.395033909506271, + "tokens_seen": 2518286336 + }, + { + "epoch": 0.76, + "learning_rate": 0.00019159879638916752, + "loss": 1.3835, + "theoretical_loss": 3.395026931346769, + "tokens_seen": 2518351872 + }, + { + "epoch": 0.76, + "learning_rate": 0.00019158274824473422, + "loss": 1.3421, + "theoretical_loss": 3.3950199534197063, + "tokens_seen": 2518417408 + }, + { + "epoch": 0.76, + "learning_rate": 0.00019156670010030092, + "loss": 1.3018, + "theoretical_loss": 3.3950129757250673, + "tokens_seen": 2518482944 + }, + { + "epoch": 0.76, + "learning_rate": 0.00019155065195586762, + "loss": 1.2212, + "theoretical_loss": 3.3950059982628384, + "tokens_seen": 2518548480 + }, + { + "epoch": 0.76, + "learning_rate": 0.00019153460381143432, + "loss": 1.4309, + "theoretical_loss": 3.3949990210330068, + "tokens_seen": 2518614016 + }, + { + "epoch": 0.76, + "learning_rate": 0.000191518555667001, + "loss": 1.1993, + "theoretical_loss": 3.3949920440355577, + "tokens_seen": 2518679552 + }, + { + "epoch": 0.76, + "learning_rate": 0.0001915025075225677, + "loss": 1.2358, + "theoretical_loss": 3.3949850672704773, + "tokens_seen": 2518745088 + }, + { + "epoch": 0.76, + "learning_rate": 0.0001914864593781344, + "loss": 1.3645, + "theoretical_loss": 3.394978090737753, + "tokens_seen": 2518810624 + }, + { + "epoch": 0.76, + "learning_rate": 0.0001914704112337011, + "loss": 1.3232, + "theoretical_loss": 3.3949711144373698, + "tokens_seen": 2518876160 + }, + { + "epoch": 0.76, + "learning_rate": 0.0001914543630892678, + "loss": 1.2143, + "theoretical_loss": 3.394964138369314, + "tokens_seen": 2518941696 + }, + { + "epoch": 0.76, + "learning_rate": 0.0001914383149448345, + "loss": 1.4161, + "theoretical_loss": 3.3949571625335726, + "tokens_seen": 2519007232 + }, + { + "epoch": 0.76, + "learning_rate": 0.0001914222668004012, + "loss": 1.2554, + "theoretical_loss": 3.394950186930131, + "tokens_seen": 2519072768 + }, + { + "epoch": 0.76, + "learning_rate": 0.0001914062186559679, + "loss": 1.2464, + "theoretical_loss": 3.394943211558976, + "tokens_seen": 2519138304 + }, + { + "epoch": 0.76, + "learning_rate": 0.0001913901705115346, + "loss": 1.2862, + "theoretical_loss": 3.3949362364200937, + "tokens_seen": 2519203840 + }, + { + "epoch": 0.76, + "learning_rate": 0.0001913741223671013, + "loss": 1.3148, + "theoretical_loss": 3.39492926151347, + "tokens_seen": 2519269376 + }, + { + "epoch": 0.76, + "learning_rate": 0.000191358074222668, + "loss": 1.2449, + "theoretical_loss": 3.3949222868390914, + "tokens_seen": 2519334912 + }, + { + "epoch": 0.76, + "learning_rate": 0.0001913420260782347, + "loss": 1.3305, + "theoretical_loss": 3.394915312396944, + "tokens_seen": 2519400448 + }, + { + "epoch": 0.76, + "learning_rate": 0.0001913259779338014, + "loss": 1.3463, + "theoretical_loss": 3.3949083381870144, + "tokens_seen": 2519465984 + }, + { + "epoch": 0.76, + "learning_rate": 0.0001913099297893681, + "loss": 1.1947, + "theoretical_loss": 3.394901364209288, + "tokens_seen": 2519531520 + }, + { + "epoch": 0.76, + "learning_rate": 0.0001912938816449348, + "loss": 1.2883, + "theoretical_loss": 3.394894390463752, + "tokens_seen": 2519597056 + }, + { + "epoch": 0.76, + "learning_rate": 0.00019127783350050153, + "loss": 1.2074, + "theoretical_loss": 3.394887416950392, + "tokens_seen": 2519662592 + }, + { + "epoch": 0.76, + "learning_rate": 0.00019126178535606823, + "loss": 1.2789, + "theoretical_loss": 3.3948804436691944, + "tokens_seen": 2519728128 + }, + { + "epoch": 0.76, + "learning_rate": 0.00019124573721163493, + "loss": 1.2577, + "theoretical_loss": 3.3948734706201456, + "tokens_seen": 2519793664 + }, + { + "epoch": 0.76, + "objective/train/docs_used": 917453, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.2612054347991943, + "objective/train/theoretical_loss": 3.394866497803232, + "objective/train/tokens_used": 2540319200, + "theoretical_loss": 3.394866497803232, + "tokens_seen": 2519859200 + }, + { + "epoch": 0.76, + "learning_rate": 0.00019122968906720164, + "loss": 1.1734, + "theoretical_loss": 3.394866497803232, + "tokens_seen": 2519859200 + }, + { + "epoch": 0.76, + "learning_rate": 0.0001912136409227683, + "loss": 1.238, + "theoretical_loss": 3.394859525218439, + "tokens_seen": 2519924736 + }, + { + "epoch": 0.76, + "learning_rate": 0.000191197592778335, + "loss": 1.4183, + "theoretical_loss": 3.3948525528657534, + "tokens_seen": 2519990272 + }, + { + "epoch": 0.76, + "learning_rate": 0.0001911815446339017, + "loss": 1.2035, + "theoretical_loss": 3.3948455807451614, + "tokens_seen": 2520055808 + }, + { + "epoch": 0.76, + "learning_rate": 0.0001911654964894684, + "loss": 1.2458, + "theoretical_loss": 3.3948386088566496, + "tokens_seen": 2520121344 + }, + { + "epoch": 0.76, + "learning_rate": 0.0001911494483450351, + "loss": 1.2168, + "theoretical_loss": 3.3948316372002036, + "tokens_seen": 2520186880 + }, + { + "epoch": 0.76, + "learning_rate": 0.0001911334002006018, + "loss": 1.324, + "theoretical_loss": 3.3948246657758094, + "tokens_seen": 2520252416 + }, + { + "epoch": 0.76, + "learning_rate": 0.0001911173520561685, + "loss": 1.3172, + "theoretical_loss": 3.3948176945834545, + "tokens_seen": 2520317952 + }, + { + "epoch": 0.76, + "learning_rate": 0.0001911013039117352, + "loss": 1.281, + "theoretical_loss": 3.394810723623124, + "tokens_seen": 2520383488 + }, + { + "epoch": 0.76, + "learning_rate": 0.00019108525576730191, + "loss": 1.3013, + "theoretical_loss": 3.3948037528948047, + "tokens_seen": 2520449024 + }, + { + "epoch": 0.76, + "learning_rate": 0.00019106920762286861, + "loss": 1.2888, + "theoretical_loss": 3.394796782398483, + "tokens_seen": 2520514560 + }, + { + "epoch": 0.76, + "learning_rate": 0.00019105315947843532, + "loss": 1.2909, + "theoretical_loss": 3.3947898121341447, + "tokens_seen": 2520580096 + }, + { + "epoch": 0.76, + "learning_rate": 0.00019103711133400202, + "loss": 1.3664, + "theoretical_loss": 3.3947828421017756, + "tokens_seen": 2520645632 + }, + { + "epoch": 0.76, + "learning_rate": 0.00019102106318956872, + "loss": 1.3043, + "theoretical_loss": 3.3947758723013632, + "tokens_seen": 2520711168 + }, + { + "epoch": 0.76, + "learning_rate": 0.00019100501504513542, + "loss": 1.2718, + "theoretical_loss": 3.394768902732893, + "tokens_seen": 2520776704 + }, + { + "epoch": 0.76, + "learning_rate": 0.00019098896690070212, + "loss": 1.3149, + "theoretical_loss": 3.394761933396351, + "tokens_seen": 2520842240 + }, + { + "epoch": 0.76, + "learning_rate": 0.00019097291875626882, + "loss": 1.3528, + "theoretical_loss": 3.394754964291724, + "tokens_seen": 2520907776 + }, + { + "epoch": 0.76, + "learning_rate": 0.00019095687061183552, + "loss": 1.4124, + "theoretical_loss": 3.3947479954189976, + "tokens_seen": 2520973312 + }, + { + "epoch": 0.76, + "learning_rate": 0.00019094082246740222, + "loss": 1.3654, + "theoretical_loss": 3.394741026778159, + "tokens_seen": 2521038848 + }, + { + "epoch": 0.76, + "learning_rate": 0.00019092477432296892, + "loss": 1.2917, + "theoretical_loss": 3.394734058369194, + "tokens_seen": 2521104384 + }, + { + "epoch": 0.76, + "learning_rate": 0.00019090872617853562, + "loss": 1.2168, + "theoretical_loss": 3.394727090192089, + "tokens_seen": 2521169920 + }, + { + "epoch": 0.76, + "learning_rate": 0.00019089267803410232, + "loss": 1.3038, + "theoretical_loss": 3.3947201222468295, + "tokens_seen": 2521235456 + }, + { + "epoch": 0.76, + "learning_rate": 0.00019087662988966902, + "loss": 1.2468, + "theoretical_loss": 3.3947131545334024, + "tokens_seen": 2521300992 + }, + { + "epoch": 0.76, + "learning_rate": 0.00019086058174523572, + "loss": 1.3202, + "theoretical_loss": 3.3947061870517943, + "tokens_seen": 2521366528 + }, + { + "epoch": 0.76, + "learning_rate": 0.00019084453360080242, + "loss": 1.3172, + "theoretical_loss": 3.394699219801991, + "tokens_seen": 2521432064 + }, + { + "epoch": 0.76, + "learning_rate": 0.0001908284854563691, + "loss": 1.3664, + "theoretical_loss": 3.3946922527839782, + "tokens_seen": 2521497600 + }, + { + "epoch": 0.76, + "learning_rate": 0.0001908124373119358, + "loss": 1.2859, + "theoretical_loss": 3.3946852859977437, + "tokens_seen": 2521563136 + }, + { + "epoch": 0.76, + "learning_rate": 0.0001907963891675025, + "loss": 1.2367, + "theoretical_loss": 3.394678319443272, + "tokens_seen": 2521628672 + }, + { + "epoch": 0.76, + "learning_rate": 0.0001907803410230692, + "loss": 1.2256, + "theoretical_loss": 3.3946713531205504, + "tokens_seen": 2521694208 + }, + { + "epoch": 0.76, + "learning_rate": 0.0001907642928786359, + "loss": 1.3756, + "theoretical_loss": 3.3946643870295654, + "tokens_seen": 2521759744 + }, + { + "epoch": 0.76, + "learning_rate": 0.0001907482447342026, + "loss": 1.3711, + "theoretical_loss": 3.3946574211703027, + "tokens_seen": 2521825280 + }, + { + "epoch": 0.76, + "learning_rate": 0.0001907321965897693, + "loss": 1.2626, + "theoretical_loss": 3.3946504555427484, + "tokens_seen": 2521890816 + }, + { + "epoch": 0.76, + "learning_rate": 0.000190716148445336, + "loss": 1.295, + "theoretical_loss": 3.3946434901468896, + "tokens_seen": 2521956352 + }, + { + "epoch": 0.76, + "learning_rate": 0.0001907001003009027, + "loss": 1.3649, + "theoretical_loss": 3.3946365249827117, + "tokens_seen": 2522021888 + }, + { + "epoch": 0.76, + "learning_rate": 0.00019068405215646943, + "loss": 1.3447, + "theoretical_loss": 3.3946295600502014, + "tokens_seen": 2522087424 + }, + { + "epoch": 0.76, + "learning_rate": 0.00019066800401203613, + "loss": 1.2115, + "theoretical_loss": 3.394622595349345, + "tokens_seen": 2522152960 + }, + { + "epoch": 0.76, + "learning_rate": 0.00019065195586760283, + "loss": 1.2423, + "theoretical_loss": 3.3946156308801285, + "tokens_seen": 2522218496 + }, + { + "epoch": 0.76, + "learning_rate": 0.00019063590772316953, + "loss": 1.2514, + "theoretical_loss": 3.3946086666425384, + "tokens_seen": 2522284032 + }, + { + "epoch": 0.76, + "learning_rate": 0.00019061985957873623, + "loss": 1.3424, + "theoretical_loss": 3.394601702636561, + "tokens_seen": 2522349568 + }, + { + "epoch": 0.76, + "learning_rate": 0.00019060381143430293, + "loss": 1.2756, + "theoretical_loss": 3.3945947388621827, + "tokens_seen": 2522415104 + }, + { + "epoch": 0.76, + "learning_rate": 0.00019058776328986963, + "loss": 1.2588, + "theoretical_loss": 3.394587775319389, + "tokens_seen": 2522480640 + }, + { + "epoch": 0.76, + "learning_rate": 0.00019057171514543633, + "loss": 1.3006, + "theoretical_loss": 3.3945808120081677, + "tokens_seen": 2522546176 + }, + { + "epoch": 0.76, + "learning_rate": 0.00019055566700100303, + "loss": 1.3057, + "theoretical_loss": 3.394573848928504, + "tokens_seen": 2522611712 + }, + { + "epoch": 0.76, + "learning_rate": 0.00019053961885656974, + "loss": 1.2927, + "theoretical_loss": 3.3945668860803835, + "tokens_seen": 2522677248 + }, + { + "epoch": 0.76, + "learning_rate": 0.00019052357071213644, + "loss": 1.2122, + "theoretical_loss": 3.394559923463794, + "tokens_seen": 2522742784 + }, + { + "epoch": 0.76, + "learning_rate": 0.0001905075225677031, + "loss": 1.3192, + "theoretical_loss": 3.3945529610787215, + "tokens_seen": 2522808320 + }, + { + "epoch": 0.76, + "learning_rate": 0.0001904914744232698, + "loss": 1.3101, + "theoretical_loss": 3.3945459989251514, + "tokens_seen": 2522873856 + }, + { + "epoch": 0.76, + "learning_rate": 0.0001904754262788365, + "loss": 1.2696, + "theoretical_loss": 3.3945390370030704, + "tokens_seen": 2522939392 + }, + { + "epoch": 0.76, + "learning_rate": 0.0001904593781344032, + "loss": 1.318, + "theoretical_loss": 3.394532075312465, + "tokens_seen": 2523004928 + }, + { + "epoch": 0.76, + "learning_rate": 0.0001904433299899699, + "loss": 1.3385, + "theoretical_loss": 3.3945251138533212, + "tokens_seen": 2523070464 + }, + { + "epoch": 0.76, + "objective/train/docs_used": 918647, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.365435004234314, + "objective/train/theoretical_loss": 3.394518152625626, + "objective/train/tokens_used": 2543596000, + "theoretical_loss": 3.394518152625626, + "tokens_seen": 2523136000 + }, + { + "epoch": 0.76, + "learning_rate": 0.0001904272818455366, + "loss": 1.3688, + "theoretical_loss": 3.394518152625626, + "tokens_seen": 2523136000 + }, + { + "epoch": 0.76, + "learning_rate": 0.0001904112337011033, + "loss": 1.2405, + "theoretical_loss": 3.394511191629365, + "tokens_seen": 2523201536 + }, + { + "epoch": 0.76, + "learning_rate": 0.00019039518555667001, + "loss": 1.2033, + "theoretical_loss": 3.3945042308645244, + "tokens_seen": 2523267072 + }, + { + "epoch": 0.76, + "learning_rate": 0.00019037913741223671, + "loss": 1.3526, + "theoretical_loss": 3.394497270331091, + "tokens_seen": 2523332608 + }, + { + "epoch": 0.76, + "learning_rate": 0.00019036308926780342, + "loss": 1.2731, + "theoretical_loss": 3.3944903100290507, + "tokens_seen": 2523398144 + }, + { + "epoch": 0.76, + "learning_rate": 0.00019034704112337012, + "loss": 1.2801, + "theoretical_loss": 3.39448334995839, + "tokens_seen": 2523463680 + }, + { + "epoch": 0.76, + "learning_rate": 0.00019033099297893682, + "loss": 1.4452, + "theoretical_loss": 3.3944763901190953, + "tokens_seen": 2523529216 + }, + { + "epoch": 0.76, + "learning_rate": 0.00019031494483450352, + "loss": 1.3248, + "theoretical_loss": 3.394469430511153, + "tokens_seen": 2523594752 + }, + { + "epoch": 0.76, + "learning_rate": 0.00019029889669007022, + "loss": 1.3078, + "theoretical_loss": 3.3944624711345486, + "tokens_seen": 2523660288 + }, + { + "epoch": 0.76, + "learning_rate": 0.00019028284854563692, + "loss": 1.2548, + "theoretical_loss": 3.394455511989269, + "tokens_seen": 2523725824 + }, + { + "epoch": 0.76, + "learning_rate": 0.00019026680040120362, + "loss": 1.3819, + "theoretical_loss": 3.394448553075301, + "tokens_seen": 2523791360 + }, + { + "epoch": 0.76, + "learning_rate": 0.00019025075225677032, + "loss": 1.4185, + "theoretical_loss": 3.39444159439263, + "tokens_seen": 2523856896 + }, + { + "epoch": 0.76, + "learning_rate": 0.00019023470411233702, + "loss": 1.2655, + "theoretical_loss": 3.3944346359412423, + "tokens_seen": 2523922432 + }, + { + "epoch": 0.76, + "learning_rate": 0.00019021865596790372, + "loss": 1.3234, + "theoretical_loss": 3.3944276777211253, + "tokens_seen": 2523987968 + }, + { + "epoch": 0.76, + "learning_rate": 0.00019020260782347042, + "loss": 1.1625, + "theoretical_loss": 3.3944207197322642, + "tokens_seen": 2524053504 + }, + { + "epoch": 0.76, + "learning_rate": 0.00019018655967903712, + "loss": 1.3263, + "theoretical_loss": 3.3944137619746457, + "tokens_seen": 2524119040 + }, + { + "epoch": 0.76, + "learning_rate": 0.00019017051153460382, + "loss": 1.2146, + "theoretical_loss": 3.3944068044482565, + "tokens_seen": 2524184576 + }, + { + "epoch": 0.76, + "learning_rate": 0.00019015446339017052, + "loss": 1.2419, + "theoretical_loss": 3.394399847153082, + "tokens_seen": 2524250112 + }, + { + "epoch": 0.76, + "learning_rate": 0.00019013841524573722, + "loss": 1.3019, + "theoretical_loss": 3.3943928900891094, + "tokens_seen": 2524315648 + }, + { + "epoch": 0.76, + "learning_rate": 0.0001901223671013039, + "loss": 1.2448, + "theoretical_loss": 3.3943859332563244, + "tokens_seen": 2524381184 + }, + { + "epoch": 0.76, + "learning_rate": 0.0001901063189568706, + "loss": 1.3471, + "theoretical_loss": 3.394378976654714, + "tokens_seen": 2524446720 + }, + { + "epoch": 0.77, + "learning_rate": 0.0001900902708124373, + "loss": 1.1825, + "theoretical_loss": 3.3943720202842638, + "tokens_seen": 2524512256 + }, + { + "epoch": 0.77, + "learning_rate": 0.00019007422266800403, + "loss": 1.2324, + "theoretical_loss": 3.39436506414496, + "tokens_seen": 2524577792 + }, + { + "epoch": 0.77, + "learning_rate": 0.00019005817452357073, + "loss": 1.2995, + "theoretical_loss": 3.3943581082367897, + "tokens_seen": 2524643328 + }, + { + "epoch": 0.77, + "learning_rate": 0.00019004212637913743, + "loss": 1.2257, + "theoretical_loss": 3.3943511525597394, + "tokens_seen": 2524708864 + }, + { + "epoch": 0.77, + "learning_rate": 0.00019002607823470413, + "loss": 1.3648, + "theoretical_loss": 3.394344197113794, + "tokens_seen": 2524774400 + }, + { + "epoch": 0.77, + "learning_rate": 0.00019001003009027083, + "loss": 1.3337, + "theoretical_loss": 3.394337241898941, + "tokens_seen": 2524839936 + }, + { + "epoch": 0.77, + "learning_rate": 0.00018999398194583753, + "loss": 1.2921, + "theoretical_loss": 3.394330286915167, + "tokens_seen": 2524905472 + }, + { + "epoch": 0.77, + "learning_rate": 0.00018997793380140423, + "loss": 1.3429, + "theoretical_loss": 3.394323332162457, + "tokens_seen": 2524971008 + }, + { + "epoch": 0.77, + "learning_rate": 0.00018996188565697093, + "loss": 1.2813, + "theoretical_loss": 3.3943163776407985, + "tokens_seen": 2525036544 + }, + { + "epoch": 0.77, + "learning_rate": 0.00018994583751253763, + "loss": 1.3897, + "theoretical_loss": 3.394309423350177, + "tokens_seen": 2525102080 + }, + { + "epoch": 0.77, + "learning_rate": 0.00018992978936810433, + "loss": 1.2975, + "theoretical_loss": 3.3943024692905794, + "tokens_seen": 2525167616 + }, + { + "epoch": 0.77, + "learning_rate": 0.00018991374122367103, + "loss": 1.1842, + "theoretical_loss": 3.394295515461992, + "tokens_seen": 2525233152 + }, + { + "epoch": 0.77, + "learning_rate": 0.00018989769307923773, + "loss": 1.3838, + "theoretical_loss": 3.3942885618644008, + "tokens_seen": 2525298688 + }, + { + "epoch": 0.77, + "learning_rate": 0.00018988164493480443, + "loss": 1.3035, + "theoretical_loss": 3.3942816084977925, + "tokens_seen": 2525364224 + }, + { + "epoch": 0.77, + "learning_rate": 0.00018986559679037114, + "loss": 1.2796, + "theoretical_loss": 3.394274655362153, + "tokens_seen": 2525429760 + }, + { + "epoch": 0.77, + "learning_rate": 0.00018984954864593784, + "loss": 1.4029, + "theoretical_loss": 3.3942677024574692, + "tokens_seen": 2525495296 + }, + { + "epoch": 0.77, + "learning_rate": 0.00018983350050150454, + "loss": 1.1614, + "theoretical_loss": 3.3942607497837267, + "tokens_seen": 2525560832 + }, + { + "epoch": 0.77, + "learning_rate": 0.00018981745235707124, + "loss": 1.2289, + "theoretical_loss": 3.3942537973409124, + "tokens_seen": 2525626368 + }, + { + "epoch": 0.77, + "learning_rate": 0.0001898014042126379, + "loss": 1.2881, + "theoretical_loss": 3.394246845129013, + "tokens_seen": 2525691904 + }, + { + "epoch": 0.77, + "learning_rate": 0.0001897853560682046, + "loss": 1.2927, + "theoretical_loss": 3.3942398931480136, + "tokens_seen": 2525757440 + }, + { + "epoch": 0.77, + "learning_rate": 0.0001897693079237713, + "loss": 1.2706, + "theoretical_loss": 3.3942329413979015, + "tokens_seen": 2525822976 + }, + { + "epoch": 0.77, + "learning_rate": 0.000189753259779338, + "loss": 1.3826, + "theoretical_loss": 3.394225989878663, + "tokens_seen": 2525888512 + }, + { + "epoch": 0.77, + "learning_rate": 0.0001897372116349047, + "loss": 1.3522, + "theoretical_loss": 3.3942190385902844, + "tokens_seen": 2525954048 + }, + { + "epoch": 0.77, + "learning_rate": 0.00018972116349047141, + "loss": 1.2731, + "theoretical_loss": 3.3942120875327513, + "tokens_seen": 2526019584 + }, + { + "epoch": 0.77, + "learning_rate": 0.00018970511534603811, + "loss": 1.2251, + "theoretical_loss": 3.3942051367060513, + "tokens_seen": 2526085120 + }, + { + "epoch": 0.77, + "learning_rate": 0.00018968906720160481, + "loss": 1.1927, + "theoretical_loss": 3.3941981861101698, + "tokens_seen": 2526150656 + }, + { + "epoch": 0.77, + "learning_rate": 0.00018967301905717152, + "loss": 1.2513, + "theoretical_loss": 3.3941912357450934, + "tokens_seen": 2526216192 + }, + { + "epoch": 0.77, + "learning_rate": 0.00018965697091273822, + "loss": 1.3564, + "theoretical_loss": 3.394184285610809, + "tokens_seen": 2526281728 + }, + { + "epoch": 0.77, + "learning_rate": 0.00018964092276830492, + "loss": 1.2944, + "theoretical_loss": 3.3941773357073015, + "tokens_seen": 2526347264 + }, + { + "epoch": 0.77, + "objective/train/docs_used": 919806, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.1638953685760498, + "objective/train/theoretical_loss": 3.394170386034559, + "objective/train/tokens_used": 2546872800, + "theoretical_loss": 3.394170386034559, + "tokens_seen": 2526412800 + }, + { + "epoch": 0.77, + "learning_rate": 0.00018962487462387162, + "loss": 1.3401, + "theoretical_loss": 3.394170386034559, + "tokens_seen": 2526412800 + }, + { + "epoch": 0.77, + "learning_rate": 0.00018960882647943832, + "loss": 1.294, + "theoretical_loss": 3.3941634365925664, + "tokens_seen": 2526478336 + }, + { + "epoch": 0.77, + "learning_rate": 0.00018959277833500502, + "loss": 1.3893, + "theoretical_loss": 3.394156487381311, + "tokens_seen": 2526543872 + }, + { + "epoch": 0.77, + "learning_rate": 0.00018957673019057172, + "loss": 1.2161, + "theoretical_loss": 3.394149538400779, + "tokens_seen": 2526609408 + }, + { + "epoch": 0.77, + "learning_rate": 0.00018956068204613842, + "loss": 1.3058, + "theoretical_loss": 3.3941425896509565, + "tokens_seen": 2526674944 + }, + { + "epoch": 0.77, + "learning_rate": 0.00018954463390170512, + "loss": 1.3054, + "theoretical_loss": 3.3941356411318298, + "tokens_seen": 2526740480 + }, + { + "epoch": 0.77, + "learning_rate": 0.00018952858575727182, + "loss": 1.2543, + "theoretical_loss": 3.3941286928433856, + "tokens_seen": 2526806016 + }, + { + "epoch": 0.77, + "learning_rate": 0.00018951253761283852, + "loss": 1.4197, + "theoretical_loss": 3.3941217447856102, + "tokens_seen": 2526871552 + }, + { + "epoch": 0.77, + "learning_rate": 0.00018949648946840522, + "loss": 1.2096, + "theoretical_loss": 3.39411479695849, + "tokens_seen": 2526937088 + }, + { + "epoch": 0.77, + "learning_rate": 0.00018948044132397192, + "loss": 1.2944, + "theoretical_loss": 3.3941078493620105, + "tokens_seen": 2527002624 + }, + { + "epoch": 0.77, + "learning_rate": 0.00018946439317953862, + "loss": 1.2935, + "theoretical_loss": 3.3941009019961594, + "tokens_seen": 2527068160 + }, + { + "epoch": 0.77, + "learning_rate": 0.00018944834503510532, + "loss": 1.3587, + "theoretical_loss": 3.394093954860922, + "tokens_seen": 2527133696 + }, + { + "epoch": 0.77, + "learning_rate": 0.00018943229689067203, + "loss": 1.3329, + "theoretical_loss": 3.3940870079562853, + "tokens_seen": 2527199232 + }, + { + "epoch": 0.77, + "learning_rate": 0.00018941624874623873, + "loss": 1.3355, + "theoretical_loss": 3.3940800612822355, + "tokens_seen": 2527264768 + }, + { + "epoch": 0.77, + "learning_rate": 0.00018940020060180543, + "loss": 1.3484, + "theoretical_loss": 3.3940731148387586, + "tokens_seen": 2527330304 + }, + { + "epoch": 0.77, + "learning_rate": 0.00018938415245737213, + "loss": 1.2625, + "theoretical_loss": 3.3940661686258418, + "tokens_seen": 2527395840 + }, + { + "epoch": 0.77, + "learning_rate": 0.00018936810431293883, + "loss": 1.3072, + "theoretical_loss": 3.3940592226434707, + "tokens_seen": 2527461376 + }, + { + "epoch": 0.77, + "learning_rate": 0.00018935205616850553, + "loss": 1.283, + "theoretical_loss": 3.3940522768916317, + "tokens_seen": 2527526912 + }, + { + "epoch": 0.77, + "learning_rate": 0.00018933600802407223, + "loss": 1.2385, + "theoretical_loss": 3.394045331370312, + "tokens_seen": 2527592448 + }, + { + "epoch": 0.77, + "learning_rate": 0.00018931995987963893, + "loss": 1.2712, + "theoretical_loss": 3.394038386079497, + "tokens_seen": 2527657984 + }, + { + "epoch": 0.77, + "learning_rate": 0.00018930391173520563, + "loss": 1.3263, + "theoretical_loss": 3.394031441019173, + "tokens_seen": 2527723520 + }, + { + "epoch": 0.77, + "learning_rate": 0.00018928786359077233, + "loss": 1.2539, + "theoretical_loss": 3.3940244961893278, + "tokens_seen": 2527789056 + }, + { + "epoch": 0.77, + "learning_rate": 0.00018927181544633903, + "loss": 1.2942, + "theoretical_loss": 3.394017551589946, + "tokens_seen": 2527854592 + }, + { + "epoch": 0.77, + "learning_rate": 0.00018925576730190573, + "loss": 1.2545, + "theoretical_loss": 3.3940106072210154, + "tokens_seen": 2527920128 + }, + { + "epoch": 0.77, + "learning_rate": 0.00018923971915747243, + "loss": 1.4011, + "theoretical_loss": 3.394003663082521, + "tokens_seen": 2527985664 + }, + { + "epoch": 0.77, + "learning_rate": 0.00018922367101303913, + "loss": 1.2719, + "theoretical_loss": 3.3939967191744507, + "tokens_seen": 2528051200 + }, + { + "epoch": 0.77, + "learning_rate": 0.00018920762286860583, + "loss": 1.2308, + "theoretical_loss": 3.39398977549679, + "tokens_seen": 2528116736 + }, + { + "epoch": 0.77, + "learning_rate": 0.00018919157472417253, + "loss": 1.3513, + "theoretical_loss": 3.3939828320495247, + "tokens_seen": 2528182272 + }, + { + "epoch": 0.77, + "learning_rate": 0.00018917552657973924, + "loss": 1.1963, + "theoretical_loss": 3.3939758888326423, + "tokens_seen": 2528247808 + }, + { + "epoch": 0.77, + "learning_rate": 0.00018915947843530594, + "loss": 1.4076, + "theoretical_loss": 3.393968945846129, + "tokens_seen": 2528313344 + }, + { + "epoch": 0.77, + "learning_rate": 0.00018914343029087264, + "loss": 1.2816, + "theoretical_loss": 3.3939620030899706, + "tokens_seen": 2528378880 + }, + { + "epoch": 0.77, + "learning_rate": 0.00018912738214643934, + "loss": 1.321, + "theoretical_loss": 3.393955060564154, + "tokens_seen": 2528444416 + }, + { + "epoch": 0.77, + "learning_rate": 0.00018911133400200604, + "loss": 1.236, + "theoretical_loss": 3.393948118268666, + "tokens_seen": 2528509952 + }, + { + "epoch": 0.77, + "learning_rate": 0.0001890952858575727, + "loss": 1.3222, + "theoretical_loss": 3.3939411762034917, + "tokens_seen": 2528575488 + }, + { + "epoch": 0.77, + "learning_rate": 0.0001890792377131394, + "loss": 1.2101, + "theoretical_loss": 3.393934234368618, + "tokens_seen": 2528641024 + }, + { + "epoch": 0.77, + "learning_rate": 0.0001890631895687061, + "loss": 1.3689, + "theoretical_loss": 3.3939272927640323, + "tokens_seen": 2528706560 + }, + { + "epoch": 0.77, + "learning_rate": 0.0001890471414242728, + "loss": 1.2485, + "theoretical_loss": 3.3939203513897196, + "tokens_seen": 2528772096 + }, + { + "epoch": 0.77, + "learning_rate": 0.00018903109327983951, + "loss": 1.3424, + "theoretical_loss": 3.393913410245667, + "tokens_seen": 2528837632 + }, + { + "epoch": 0.77, + "learning_rate": 0.00018901504513540621, + "loss": 1.2234, + "theoretical_loss": 3.393906469331861, + "tokens_seen": 2528903168 + }, + { + "epoch": 0.77, + "learning_rate": 0.00018899899699097292, + "loss": 1.2267, + "theoretical_loss": 3.393899528648287, + "tokens_seen": 2528968704 + }, + { + "epoch": 0.77, + "learning_rate": 0.00018898294884653962, + "loss": 1.3198, + "theoretical_loss": 3.3938925881949333, + "tokens_seen": 2529034240 + }, + { + "epoch": 0.77, + "learning_rate": 0.00018896690070210632, + "loss": 1.3373, + "theoretical_loss": 3.3938856479717847, + "tokens_seen": 2529099776 + }, + { + "epoch": 0.77, + "learning_rate": 0.00018895085255767302, + "loss": 1.3047, + "theoretical_loss": 3.393878707978828, + "tokens_seen": 2529165312 + }, + { + "epoch": 0.77, + "learning_rate": 0.00018893480441323972, + "loss": 1.3448, + "theoretical_loss": 3.3938717682160497, + "tokens_seen": 2529230848 + }, + { + "epoch": 0.77, + "learning_rate": 0.00018891875626880642, + "loss": 1.3114, + "theoretical_loss": 3.393864828683436, + "tokens_seen": 2529296384 + }, + { + "epoch": 0.77, + "learning_rate": 0.00018890270812437312, + "loss": 1.3121, + "theoretical_loss": 3.393857889380974, + "tokens_seen": 2529361920 + }, + { + "epoch": 0.77, + "learning_rate": 0.00018888665997993985, + "loss": 1.2167, + "theoretical_loss": 3.393850950308649, + "tokens_seen": 2529427456 + }, + { + "epoch": 0.77, + "learning_rate": 0.00018887061183550655, + "loss": 1.3148, + "theoretical_loss": 3.393844011466448, + "tokens_seen": 2529492992 + }, + { + "epoch": 0.77, + "learning_rate": 0.00018885456369107325, + "loss": 1.318, + "theoretical_loss": 3.3938370728543577, + "tokens_seen": 2529558528 + }, + { + "epoch": 0.77, + "learning_rate": 0.00018883851554663995, + "loss": 1.3731, + "theoretical_loss": 3.3938301344723643, + "tokens_seen": 2529624064 + }, + { + "epoch": 0.77, + "objective/train/docs_used": 920541, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.250909447669983, + "objective/train/theoretical_loss": 3.3938231963204535, + "objective/train/tokens_used": 2550149600, + "theoretical_loss": 3.3938231963204535, + "tokens_seen": 2529689600 + }, + { + "epoch": 0.77, + "learning_rate": 0.00018882246740220665, + "loss": 1.3115, + "theoretical_loss": 3.3938231963204535, + "tokens_seen": 2529689600 + }, + { + "epoch": 0.77, + "learning_rate": 0.00018880641925777335, + "loss": 1.3118, + "theoretical_loss": 3.393816258398613, + "tokens_seen": 2529755136 + }, + { + "epoch": 0.77, + "learning_rate": 0.00018879037111334002, + "loss": 1.4095, + "theoretical_loss": 3.3938093207068283, + "tokens_seen": 2529820672 + }, + { + "epoch": 0.77, + "learning_rate": 0.00018877432296890672, + "loss": 1.186, + "theoretical_loss": 3.3938023832450854, + "tokens_seen": 2529886208 + }, + { + "epoch": 0.77, + "learning_rate": 0.00018875827482447342, + "loss": 1.3127, + "theoretical_loss": 3.3937954460133724, + "tokens_seen": 2529951744 + }, + { + "epoch": 0.77, + "learning_rate": 0.00018874222668004013, + "loss": 1.3832, + "theoretical_loss": 3.393788509011674, + "tokens_seen": 2530017280 + }, + { + "epoch": 0.77, + "learning_rate": 0.00018872617853560683, + "loss": 1.3121, + "theoretical_loss": 3.393781572239978, + "tokens_seen": 2530082816 + }, + { + "epoch": 0.77, + "learning_rate": 0.00018871013039117353, + "loss": 1.2766, + "theoretical_loss": 3.393774635698269, + "tokens_seen": 2530148352 + }, + { + "epoch": 0.77, + "learning_rate": 0.00018869408224674023, + "loss": 1.3326, + "theoretical_loss": 3.3937676993865353, + "tokens_seen": 2530213888 + }, + { + "epoch": 0.77, + "learning_rate": 0.00018867803410230693, + "loss": 1.3134, + "theoretical_loss": 3.3937607633047624, + "tokens_seen": 2530279424 + }, + { + "epoch": 0.77, + "learning_rate": 0.00018866198595787363, + "loss": 1.2621, + "theoretical_loss": 3.3937538274529366, + "tokens_seen": 2530344960 + }, + { + "epoch": 0.77, + "learning_rate": 0.00018864593781344033, + "loss": 1.3943, + "theoretical_loss": 3.3937468918310447, + "tokens_seen": 2530410496 + }, + { + "epoch": 0.77, + "learning_rate": 0.00018862988966900703, + "loss": 1.2376, + "theoretical_loss": 3.3937399564390733, + "tokens_seen": 2530476032 + }, + { + "epoch": 0.77, + "learning_rate": 0.00018861384152457373, + "loss": 1.2255, + "theoretical_loss": 3.3937330212770083, + "tokens_seen": 2530541568 + }, + { + "epoch": 0.77, + "learning_rate": 0.00018859779338014043, + "loss": 1.3222, + "theoretical_loss": 3.393726086344836, + "tokens_seen": 2530607104 + }, + { + "epoch": 0.77, + "learning_rate": 0.00018858174523570713, + "loss": 1.2999, + "theoretical_loss": 3.393719151642544, + "tokens_seen": 2530672640 + }, + { + "epoch": 0.77, + "learning_rate": 0.00018856569709127383, + "loss": 1.1533, + "theoretical_loss": 3.3937122171701173, + "tokens_seen": 2530738176 + }, + { + "epoch": 0.77, + "learning_rate": 0.00018854964894684053, + "loss": 1.3002, + "theoretical_loss": 3.393705282927543, + "tokens_seen": 2530803712 + }, + { + "epoch": 0.77, + "learning_rate": 0.00018853360080240723, + "loss": 1.3209, + "theoretical_loss": 3.3936983489148074, + "tokens_seen": 2530869248 + }, + { + "epoch": 0.77, + "learning_rate": 0.00018851755265797393, + "loss": 1.2544, + "theoretical_loss": 3.3936914151318973, + "tokens_seen": 2530934784 + }, + { + "epoch": 0.77, + "learning_rate": 0.00018850150451354063, + "loss": 1.2988, + "theoretical_loss": 3.3936844815787985, + "tokens_seen": 2531000320 + }, + { + "epoch": 0.77, + "learning_rate": 0.00018848545636910734, + "loss": 1.2903, + "theoretical_loss": 3.3936775482554977, + "tokens_seen": 2531065856 + }, + { + "epoch": 0.77, + "learning_rate": 0.00018846940822467404, + "loss": 1.3101, + "theoretical_loss": 3.3936706151619815, + "tokens_seen": 2531131392 + }, + { + "epoch": 0.77, + "learning_rate": 0.00018845336008024074, + "loss": 1.3453, + "theoretical_loss": 3.3936636822982367, + "tokens_seen": 2531196928 + }, + { + "epoch": 0.77, + "learning_rate": 0.00018843731193580744, + "loss": 1.2776, + "theoretical_loss": 3.3936567496642485, + "tokens_seen": 2531262464 + }, + { + "epoch": 0.77, + "learning_rate": 0.00018842126379137414, + "loss": 1.3793, + "theoretical_loss": 3.3936498172600045, + "tokens_seen": 2531328000 + }, + { + "epoch": 0.77, + "learning_rate": 0.0001884052156469408, + "loss": 1.3275, + "theoretical_loss": 3.393642885085491, + "tokens_seen": 2531393536 + }, + { + "epoch": 0.77, + "learning_rate": 0.0001883891675025075, + "loss": 1.3884, + "theoretical_loss": 3.3936359531406937, + "tokens_seen": 2531459072 + }, + { + "epoch": 0.77, + "learning_rate": 0.0001883731193580742, + "loss": 1.282, + "theoretical_loss": 3.3936290214255997, + "tokens_seen": 2531524608 + }, + { + "epoch": 0.77, + "learning_rate": 0.0001883570712136409, + "loss": 1.2768, + "theoretical_loss": 3.3936220899401954, + "tokens_seen": 2531590144 + }, + { + "epoch": 0.77, + "learning_rate": 0.00018834102306920761, + "loss": 1.3318, + "theoretical_loss": 3.393615158684467, + "tokens_seen": 2531655680 + }, + { + "epoch": 0.77, + "learning_rate": 0.00018832497492477431, + "loss": 1.3474, + "theoretical_loss": 3.393608227658401, + "tokens_seen": 2531721216 + }, + { + "epoch": 0.77, + "learning_rate": 0.00018830892678034102, + "loss": 1.256, + "theoretical_loss": 3.3936012968619838, + "tokens_seen": 2531786752 + }, + { + "epoch": 0.77, + "learning_rate": 0.00018829287863590772, + "loss": 1.2553, + "theoretical_loss": 3.393594366295202, + "tokens_seen": 2531852288 + }, + { + "epoch": 0.77, + "learning_rate": 0.00018827683049147444, + "loss": 1.2287, + "theoretical_loss": 3.393587435958042, + "tokens_seen": 2531917824 + }, + { + "epoch": 0.77, + "learning_rate": 0.00018826078234704114, + "loss": 1.2137, + "theoretical_loss": 3.39358050585049, + "tokens_seen": 2531983360 + }, + { + "epoch": 0.77, + "learning_rate": 0.00018824473420260784, + "loss": 1.2717, + "theoretical_loss": 3.393573575972533, + "tokens_seen": 2532048896 + }, + { + "epoch": 0.77, + "learning_rate": 0.00018822868605817455, + "loss": 1.3334, + "theoretical_loss": 3.393566646324157, + "tokens_seen": 2532114432 + }, + { + "epoch": 0.77, + "learning_rate": 0.00018821263791374125, + "loss": 1.3239, + "theoretical_loss": 3.393559716905349, + "tokens_seen": 2532179968 + }, + { + "epoch": 0.77, + "learning_rate": 0.00018819658976930795, + "loss": 1.3059, + "theoretical_loss": 3.3935527877160947, + "tokens_seen": 2532245504 + }, + { + "epoch": 0.77, + "learning_rate": 0.00018818054162487465, + "loss": 1.3277, + "theoretical_loss": 3.3935458587563807, + "tokens_seen": 2532311040 + }, + { + "epoch": 0.77, + "learning_rate": 0.00018816449348044135, + "loss": 1.2305, + "theoretical_loss": 3.3935389300261942, + "tokens_seen": 2532376576 + }, + { + "epoch": 0.77, + "learning_rate": 0.00018814844533600805, + "loss": 1.2592, + "theoretical_loss": 3.3935320015255206, + "tokens_seen": 2532442112 + }, + { + "epoch": 0.77, + "learning_rate": 0.00018813239719157475, + "loss": 1.3706, + "theoretical_loss": 3.393525073254347, + "tokens_seen": 2532507648 + }, + { + "epoch": 0.77, + "learning_rate": 0.00018811634904714145, + "loss": 1.2858, + "theoretical_loss": 3.39351814521266, + "tokens_seen": 2532573184 + }, + { + "epoch": 0.77, + "learning_rate": 0.00018810030090270815, + "loss": 1.3124, + "theoretical_loss": 3.3935112174004454, + "tokens_seen": 2532638720 + }, + { + "epoch": 0.77, + "learning_rate": 0.00018808425275827482, + "loss": 1.2477, + "theoretical_loss": 3.39350428981769, + "tokens_seen": 2532704256 + }, + { + "epoch": 0.77, + "learning_rate": 0.00018806820461384152, + "loss": 1.3303, + "theoretical_loss": 3.3934973624643807, + "tokens_seen": 2532769792 + }, + { + "epoch": 0.77, + "learning_rate": 0.00018805215646940823, + "loss": 1.2599, + "theoretical_loss": 3.393490435340503, + "tokens_seen": 2532835328 + }, + { + "epoch": 0.77, + "learning_rate": 0.00018803610832497493, + "loss": 1.257, + "theoretical_loss": 3.3934835084460446, + "tokens_seen": 2532900864 + }, + { + "epoch": 0.77, + "objective/train/docs_used": 921695, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.308255672454834, + "objective/train/theoretical_loss": 3.393476581780991, + "objective/train/tokens_used": 2553426400, + "theoretical_loss": 3.393476581780991, + "tokens_seen": 2532966400 + }, + { + "epoch": 0.77, + "learning_rate": 0.00018802006018054163, + "loss": 1.2924, + "theoretical_loss": 3.393476581780991, + "tokens_seen": 2532966400 + }, + { + "epoch": 0.77, + "learning_rate": 0.00018800401203610833, + "loss": 1.3327, + "theoretical_loss": 3.393469655345329, + "tokens_seen": 2533031936 + }, + { + "epoch": 0.77, + "learning_rate": 0.00018798796389167503, + "loss": 1.276, + "theoretical_loss": 3.3934627291390456, + "tokens_seen": 2533097472 + }, + { + "epoch": 0.77, + "learning_rate": 0.00018797191574724173, + "loss": 1.2451, + "theoretical_loss": 3.393455803162126, + "tokens_seen": 2533163008 + }, + { + "epoch": 0.77, + "learning_rate": 0.00018795586760280843, + "loss": 1.2708, + "theoretical_loss": 3.3934488774145577, + "tokens_seen": 2533228544 + }, + { + "epoch": 0.77, + "learning_rate": 0.00018793981945837513, + "loss": 1.247, + "theoretical_loss": 3.393441951896327, + "tokens_seen": 2533294080 + }, + { + "epoch": 0.77, + "learning_rate": 0.00018792377131394183, + "loss": 1.2633, + "theoretical_loss": 3.3934350266074196, + "tokens_seen": 2533359616 + }, + { + "epoch": 0.77, + "learning_rate": 0.00018790772316950853, + "loss": 1.3472, + "theoretical_loss": 3.3934281015478227, + "tokens_seen": 2533425152 + }, + { + "epoch": 0.77, + "learning_rate": 0.00018789167502507523, + "loss": 1.2695, + "theoretical_loss": 3.3934211767175233, + "tokens_seen": 2533490688 + }, + { + "epoch": 0.77, + "learning_rate": 0.00018787562688064193, + "loss": 1.299, + "theoretical_loss": 3.3934142521165067, + "tokens_seen": 2533556224 + }, + { + "epoch": 0.77, + "learning_rate": 0.00018785957873620863, + "loss": 1.3168, + "theoretical_loss": 3.3934073277447605, + "tokens_seen": 2533621760 + }, + { + "epoch": 0.77, + "learning_rate": 0.00018784353059177533, + "loss": 1.2196, + "theoretical_loss": 3.39340040360227, + "tokens_seen": 2533687296 + }, + { + "epoch": 0.77, + "learning_rate": 0.00018782748244734203, + "loss": 1.345, + "theoretical_loss": 3.393393479689023, + "tokens_seen": 2533752832 + }, + { + "epoch": 0.77, + "learning_rate": 0.00018781143430290873, + "loss": 1.232, + "theoretical_loss": 3.3933865560050043, + "tokens_seen": 2533818368 + }, + { + "epoch": 0.77, + "learning_rate": 0.00018779538615847544, + "loss": 1.4252, + "theoretical_loss": 3.393379632550202, + "tokens_seen": 2533883904 + }, + { + "epoch": 0.77, + "learning_rate": 0.00018777933801404214, + "loss": 1.2242, + "theoretical_loss": 3.393372709324602, + "tokens_seen": 2533949440 + }, + { + "epoch": 0.77, + "learning_rate": 0.00018776328986960884, + "loss": 1.335, + "theoretical_loss": 3.3933657863281903, + "tokens_seen": 2534014976 + }, + { + "epoch": 0.77, + "learning_rate": 0.00018774724172517554, + "loss": 1.2723, + "theoretical_loss": 3.3933588635609544, + "tokens_seen": 2534080512 + }, + { + "epoch": 0.77, + "learning_rate": 0.00018773119358074224, + "loss": 1.3565, + "theoretical_loss": 3.39335194102288, + "tokens_seen": 2534146048 + }, + { + "epoch": 0.77, + "learning_rate": 0.00018771514543630894, + "loss": 1.3189, + "theoretical_loss": 3.3933450187139536, + "tokens_seen": 2534211584 + }, + { + "epoch": 0.77, + "learning_rate": 0.0001876990972918756, + "loss": 1.2171, + "theoretical_loss": 3.393338096634162, + "tokens_seen": 2534277120 + }, + { + "epoch": 0.77, + "learning_rate": 0.00018768304914744234, + "loss": 1.3095, + "theoretical_loss": 3.3933311747834916, + "tokens_seen": 2534342656 + }, + { + "epoch": 0.77, + "learning_rate": 0.00018766700100300904, + "loss": 1.2212, + "theoretical_loss": 3.3933242531619285, + "tokens_seen": 2534408192 + }, + { + "epoch": 0.77, + "learning_rate": 0.00018765095285857574, + "loss": 1.3527, + "theoretical_loss": 3.3933173317694596, + "tokens_seen": 2534473728 + }, + { + "epoch": 0.77, + "learning_rate": 0.00018763490471414244, + "loss": 1.2248, + "theoretical_loss": 3.393310410606072, + "tokens_seen": 2534539264 + }, + { + "epoch": 0.77, + "learning_rate": 0.00018761885656970914, + "loss": 1.2855, + "theoretical_loss": 3.3933034896717507, + "tokens_seen": 2534604800 + }, + { + "epoch": 0.77, + "learning_rate": 0.00018760280842527584, + "loss": 1.285, + "theoretical_loss": 3.3932965689664836, + "tokens_seen": 2534670336 + }, + { + "epoch": 0.77, + "learning_rate": 0.00018758676028084254, + "loss": 1.308, + "theoretical_loss": 3.3932896484902564, + "tokens_seen": 2534735872 + }, + { + "epoch": 0.77, + "learning_rate": 0.00018757071213640924, + "loss": 1.3227, + "theoretical_loss": 3.393282728243056, + "tokens_seen": 2534801408 + }, + { + "epoch": 0.77, + "learning_rate": 0.00018755466399197595, + "loss": 1.232, + "theoretical_loss": 3.3932758082248684, + "tokens_seen": 2534866944 + }, + { + "epoch": 0.77, + "learning_rate": 0.00018753861584754265, + "loss": 1.3761, + "theoretical_loss": 3.393268888435681, + "tokens_seen": 2534932480 + }, + { + "epoch": 0.77, + "learning_rate": 0.00018752256770310935, + "loss": 1.1838, + "theoretical_loss": 3.3932619688754793, + "tokens_seen": 2534998016 + }, + { + "epoch": 0.77, + "learning_rate": 0.00018750651955867605, + "loss": 1.338, + "theoretical_loss": 3.3932550495442504, + "tokens_seen": 2535063552 + }, + { + "epoch": 0.77, + "learning_rate": 0.00018749047141424275, + "loss": 1.3706, + "theoretical_loss": 3.3932481304419806, + "tokens_seen": 2535129088 + }, + { + "epoch": 0.77, + "learning_rate": 0.00018747442326980945, + "loss": 1.3636, + "theoretical_loss": 3.393241211568656, + "tokens_seen": 2535194624 + }, + { + "epoch": 0.77, + "learning_rate": 0.00018745837512537615, + "loss": 1.3136, + "theoretical_loss": 3.393234292924264, + "tokens_seen": 2535260160 + }, + { + "epoch": 0.77, + "learning_rate": 0.00018744232698094285, + "loss": 1.3188, + "theoretical_loss": 3.3932273745087906, + "tokens_seen": 2535325696 + }, + { + "epoch": 0.77, + "learning_rate": 0.00018742627883650955, + "loss": 1.3691, + "theoretical_loss": 3.3932204563222226, + "tokens_seen": 2535391232 + }, + { + "epoch": 0.77, + "learning_rate": 0.00018741023069207625, + "loss": 1.2683, + "theoretical_loss": 3.393213538364546, + "tokens_seen": 2535456768 + }, + { + "epoch": 0.77, + "learning_rate": 0.00018739418254764295, + "loss": 1.2871, + "theoretical_loss": 3.393206620635748, + "tokens_seen": 2535522304 + }, + { + "epoch": 0.77, + "learning_rate": 0.00018737813440320962, + "loss": 1.2956, + "theoretical_loss": 3.393199703135814, + "tokens_seen": 2535587840 + }, + { + "epoch": 0.77, + "learning_rate": 0.00018736208625877633, + "loss": 1.2797, + "theoretical_loss": 3.3931927858647315, + "tokens_seen": 2535653376 + }, + { + "epoch": 0.77, + "learning_rate": 0.00018734603811434303, + "loss": 1.2047, + "theoretical_loss": 3.3931858688224867, + "tokens_seen": 2535718912 + }, + { + "epoch": 0.77, + "learning_rate": 0.00018732998996990973, + "loss": 1.211, + "theoretical_loss": 3.3931789520090665, + "tokens_seen": 2535784448 + }, + { + "epoch": 0.77, + "learning_rate": 0.00018731394182547643, + "loss": 1.2651, + "theoretical_loss": 3.393172035424457, + "tokens_seen": 2535849984 + }, + { + "epoch": 0.77, + "learning_rate": 0.00018729789368104313, + "loss": 1.3113, + "theoretical_loss": 3.3931651190686445, + "tokens_seen": 2535915520 + }, + { + "epoch": 0.77, + "learning_rate": 0.00018728184553660983, + "loss": 1.2982, + "theoretical_loss": 3.393158202941616, + "tokens_seen": 2535981056 + }, + { + "epoch": 0.77, + "learning_rate": 0.00018726579739217653, + "loss": 1.2388, + "theoretical_loss": 3.3931512870433576, + "tokens_seen": 2536046592 + }, + { + "epoch": 0.77, + "learning_rate": 0.00018724974924774323, + "loss": 1.3371, + "theoretical_loss": 3.393144371373856, + "tokens_seen": 2536112128 + }, + { + "epoch": 0.77, + "learning_rate": 0.00018723370110330993, + "loss": 1.3167, + "theoretical_loss": 3.393137455933098, + "tokens_seen": 2536177664 + }, + { + "epoch": 0.77, + "objective/train/docs_used": 922895, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.3763655424118042, + "objective/train/theoretical_loss": 3.39313054072107, + "objective/train/tokens_used": 2556703200, + "theoretical_loss": 3.39313054072107, + "tokens_seen": 2536243200 + }, + { + "epoch": 0.77, + "learning_rate": 0.00018721765295887663, + "loss": 1.3774, + "theoretical_loss": 3.39313054072107, + "tokens_seen": 2536243200 + }, + { + "epoch": 0.77, + "learning_rate": 0.00018720160481444333, + "loss": 1.3008, + "theoretical_loss": 3.3931236257377577, + "tokens_seen": 2536308736 + }, + { + "epoch": 0.77, + "learning_rate": 0.00018718555667001003, + "loss": 1.3774, + "theoretical_loss": 3.393116710983149, + "tokens_seen": 2536374272 + }, + { + "epoch": 0.77, + "learning_rate": 0.00018716950852557673, + "loss": 1.2684, + "theoretical_loss": 3.3931097964572294, + "tokens_seen": 2536439808 + }, + { + "epoch": 0.77, + "learning_rate": 0.00018715346038114343, + "loss": 1.3797, + "theoretical_loss": 3.393102882159986, + "tokens_seen": 2536505344 + }, + { + "epoch": 0.77, + "learning_rate": 0.00018713741223671013, + "loss": 1.2744, + "theoretical_loss": 3.393095968091405, + "tokens_seen": 2536570880 + }, + { + "epoch": 0.77, + "learning_rate": 0.00018712136409227684, + "loss": 1.2575, + "theoretical_loss": 3.3930890542514733, + "tokens_seen": 2536636416 + }, + { + "epoch": 0.77, + "learning_rate": 0.00018710531594784354, + "loss": 1.2412, + "theoretical_loss": 3.393082140640177, + "tokens_seen": 2536701952 + }, + { + "epoch": 0.77, + "learning_rate": 0.00018708926780341024, + "loss": 1.2806, + "theoretical_loss": 3.393075227257503, + "tokens_seen": 2536767488 + }, + { + "epoch": 0.77, + "learning_rate": 0.00018707321965897694, + "loss": 1.3245, + "theoretical_loss": 3.3930683141034375, + "tokens_seen": 2536833024 + }, + { + "epoch": 0.77, + "learning_rate": 0.00018705717151454364, + "loss": 1.3418, + "theoretical_loss": 3.3930614011779667, + "tokens_seen": 2536898560 + }, + { + "epoch": 0.77, + "learning_rate": 0.00018704112337011034, + "loss": 1.2735, + "theoretical_loss": 3.3930544884810785, + "tokens_seen": 2536964096 + }, + { + "epoch": 0.77, + "learning_rate": 0.00018702507522567704, + "loss": 1.188, + "theoretical_loss": 3.393047576012758, + "tokens_seen": 2537029632 + }, + { + "epoch": 0.77, + "learning_rate": 0.00018700902708124374, + "loss": 1.2707, + "theoretical_loss": 3.393040663772992, + "tokens_seen": 2537095168 + }, + { + "epoch": 0.77, + "learning_rate": 0.00018699297893681044, + "loss": 1.2649, + "theoretical_loss": 3.3930337517617684, + "tokens_seen": 2537160704 + }, + { + "epoch": 0.77, + "learning_rate": 0.00018697693079237714, + "loss": 1.3029, + "theoretical_loss": 3.393026839979072, + "tokens_seen": 2537226240 + }, + { + "epoch": 0.77, + "learning_rate": 0.00018696088264794384, + "loss": 1.3839, + "theoretical_loss": 3.39301992842489, + "tokens_seen": 2537291776 + }, + { + "epoch": 0.77, + "learning_rate": 0.00018694483450351054, + "loss": 1.343, + "theoretical_loss": 3.393013017099209, + "tokens_seen": 2537357312 + }, + { + "epoch": 0.77, + "learning_rate": 0.00018692878635907724, + "loss": 1.3001, + "theoretical_loss": 3.3930061060020154, + "tokens_seen": 2537422848 + }, + { + "epoch": 0.77, + "learning_rate": 0.00018691273821464394, + "loss": 1.2547, + "theoretical_loss": 3.392999195133296, + "tokens_seen": 2537488384 + }, + { + "epoch": 0.77, + "learning_rate": 0.00018689669007021064, + "loss": 1.3757, + "theoretical_loss": 3.392992284493037, + "tokens_seen": 2537553920 + }, + { + "epoch": 0.77, + "learning_rate": 0.00018688064192577734, + "loss": 1.326, + "theoretical_loss": 3.3929853740812255, + "tokens_seen": 2537619456 + }, + { + "epoch": 0.77, + "learning_rate": 0.00018686459378134405, + "loss": 1.3661, + "theoretical_loss": 3.3929784638978475, + "tokens_seen": 2537684992 + }, + { + "epoch": 0.77, + "learning_rate": 0.00018684854563691075, + "loss": 1.1634, + "theoretical_loss": 3.39297155394289, + "tokens_seen": 2537750528 + }, + { + "epoch": 0.77, + "learning_rate": 0.00018683249749247745, + "loss": 1.3219, + "theoretical_loss": 3.3929646442163386, + "tokens_seen": 2537816064 + }, + { + "epoch": 0.77, + "learning_rate": 0.00018681644934804415, + "loss": 1.3043, + "theoretical_loss": 3.392957734718181, + "tokens_seen": 2537881600 + }, + { + "epoch": 0.77, + "learning_rate": 0.00018680040120361085, + "loss": 1.1716, + "theoretical_loss": 3.3929508254484033, + "tokens_seen": 2537947136 + }, + { + "epoch": 0.77, + "learning_rate": 0.00018678435305917755, + "loss": 1.4173, + "theoretical_loss": 3.392943916406992, + "tokens_seen": 2538012672 + }, + { + "epoch": 0.77, + "learning_rate": 0.00018676830491474425, + "loss": 1.3026, + "theoretical_loss": 3.392937007593934, + "tokens_seen": 2538078208 + }, + { + "epoch": 0.77, + "learning_rate": 0.00018675225677031095, + "loss": 1.3872, + "theoretical_loss": 3.392930099009215, + "tokens_seen": 2538143744 + }, + { + "epoch": 0.77, + "learning_rate": 0.00018673620862587765, + "loss": 1.2928, + "theoretical_loss": 3.392923190652822, + "tokens_seen": 2538209280 + }, + { + "epoch": 0.77, + "learning_rate": 0.00018672016048144435, + "loss": 1.3327, + "theoretical_loss": 3.3929162825247423, + "tokens_seen": 2538274816 + }, + { + "epoch": 0.77, + "learning_rate": 0.00018670411233701105, + "loss": 1.2769, + "theoretical_loss": 3.3929093746249617, + "tokens_seen": 2538340352 + }, + { + "epoch": 0.77, + "learning_rate": 0.00018668806419257773, + "loss": 1.245, + "theoretical_loss": 3.392902466953467, + "tokens_seen": 2538405888 + }, + { + "epoch": 0.77, + "learning_rate": 0.00018667201604814443, + "loss": 1.2458, + "theoretical_loss": 3.3928955595102446, + "tokens_seen": 2538471424 + }, + { + "epoch": 0.77, + "learning_rate": 0.00018665596790371113, + "loss": 1.2326, + "theoretical_loss": 3.392888652295281, + "tokens_seen": 2538536960 + }, + { + "epoch": 0.77, + "learning_rate": 0.00018663991975927783, + "loss": 1.2661, + "theoretical_loss": 3.3928817453085625, + "tokens_seen": 2538602496 + }, + { + "epoch": 0.77, + "learning_rate": 0.00018662387161484453, + "loss": 1.2776, + "theoretical_loss": 3.3928748385500764, + "tokens_seen": 2538668032 + }, + { + "epoch": 0.77, + "learning_rate": 0.00018660782347041123, + "loss": 1.236, + "theoretical_loss": 3.392867932019809, + "tokens_seen": 2538733568 + }, + { + "epoch": 0.77, + "learning_rate": 0.00018659177532597793, + "loss": 1.3422, + "theoretical_loss": 3.392861025717747, + "tokens_seen": 2538799104 + }, + { + "epoch": 0.77, + "learning_rate": 0.00018657572718154463, + "loss": 1.2618, + "theoretical_loss": 3.3928541196438764, + "tokens_seen": 2538864640 + }, + { + "epoch": 0.77, + "learning_rate": 0.00018655967903711133, + "loss": 1.2243, + "theoretical_loss": 3.392847213798184, + "tokens_seen": 2538930176 + }, + { + "epoch": 0.77, + "learning_rate": 0.00018654363089267803, + "loss": 1.2823, + "theoretical_loss": 3.392840308180657, + "tokens_seen": 2538995712 + }, + { + "epoch": 0.77, + "learning_rate": 0.00018652758274824473, + "loss": 1.2031, + "theoretical_loss": 3.3928334027912816, + "tokens_seen": 2539061248 + }, + { + "epoch": 0.77, + "learning_rate": 0.00018651153460381143, + "loss": 1.3025, + "theoretical_loss": 3.3928264976300437, + "tokens_seen": 2539126784 + }, + { + "epoch": 0.77, + "learning_rate": 0.00018649548645937813, + "loss": 1.3248, + "theoretical_loss": 3.3928195926969305, + "tokens_seen": 2539192320 + }, + { + "epoch": 0.77, + "learning_rate": 0.00018647943831494486, + "loss": 1.1955, + "theoretical_loss": 3.3928126879919285, + "tokens_seen": 2539257856 + }, + { + "epoch": 0.77, + "learning_rate": 0.00018646339017051156, + "loss": 1.2234, + "theoretical_loss": 3.3928057835150245, + "tokens_seen": 2539323392 + }, + { + "epoch": 0.77, + "learning_rate": 0.00018644734202607826, + "loss": 1.1946, + "theoretical_loss": 3.3927988792662047, + "tokens_seen": 2539388928 + }, + { + "epoch": 0.77, + "learning_rate": 0.00018643129388164496, + "loss": 1.2815, + "theoretical_loss": 3.392791975245456, + "tokens_seen": 2539454464 + }, + { + "epoch": 0.77, + "objective/train/docs_used": 924078, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.268160104751587, + "objective/train/theoretical_loss": 3.3927850714527645, + "objective/train/tokens_used": 2559980000, + "theoretical_loss": 3.3927850714527645, + "tokens_seen": 2539520000 + }, + { + "epoch": 0.77, + "learning_rate": 0.00018641524573721166, + "loss": 1.3444, + "theoretical_loss": 3.3927850714527645, + "tokens_seen": 2539520000 + }, + { + "epoch": 0.77, + "learning_rate": 0.00018639919759277836, + "loss": 1.264, + "theoretical_loss": 3.3927781678881175, + "tokens_seen": 2539585536 + }, + { + "epoch": 0.77, + "learning_rate": 0.00018638314944834506, + "loss": 1.2807, + "theoretical_loss": 3.3927712645515005, + "tokens_seen": 2539651072 + }, + { + "epoch": 0.77, + "learning_rate": 0.00018636710130391174, + "loss": 1.3698, + "theoretical_loss": 3.3927643614429015, + "tokens_seen": 2539716608 + }, + { + "epoch": 0.77, + "learning_rate": 0.00018635105315947844, + "loss": 1.2557, + "theoretical_loss": 3.392757458562306, + "tokens_seen": 2539782144 + }, + { + "epoch": 0.77, + "learning_rate": 0.00018633500501504514, + "loss": 1.1581, + "theoretical_loss": 3.392750555909701, + "tokens_seen": 2539847680 + }, + { + "epoch": 0.77, + "learning_rate": 0.00018631895687061184, + "loss": 1.2969, + "theoretical_loss": 3.392743653485073, + "tokens_seen": 2539913216 + }, + { + "epoch": 0.77, + "learning_rate": 0.00018630290872617854, + "loss": 1.2418, + "theoretical_loss": 3.3927367512884086, + "tokens_seen": 2539978752 + }, + { + "epoch": 0.77, + "learning_rate": 0.00018628686058174524, + "loss": 1.3214, + "theoretical_loss": 3.3927298493196947, + "tokens_seen": 2540044288 + }, + { + "epoch": 0.77, + "learning_rate": 0.00018627081243731194, + "loss": 1.11, + "theoretical_loss": 3.392722947578917, + "tokens_seen": 2540109824 + }, + { + "epoch": 0.77, + "learning_rate": 0.00018625476429287864, + "loss": 1.2677, + "theoretical_loss": 3.392716046066063, + "tokens_seen": 2540175360 + }, + { + "epoch": 0.77, + "learning_rate": 0.00018623871614844534, + "loss": 1.2759, + "theoretical_loss": 3.392709144781119, + "tokens_seen": 2540240896 + }, + { + "epoch": 0.77, + "learning_rate": 0.00018622266800401204, + "loss": 1.2133, + "theoretical_loss": 3.3927022437240715, + "tokens_seen": 2540306432 + }, + { + "epoch": 0.77, + "learning_rate": 0.00018620661985957874, + "loss": 1.2949, + "theoretical_loss": 3.3926953428949074, + "tokens_seen": 2540371968 + }, + { + "epoch": 0.77, + "learning_rate": 0.00018619057171514544, + "loss": 1.3028, + "theoretical_loss": 3.3926884422936125, + "tokens_seen": 2540437504 + }, + { + "epoch": 0.77, + "learning_rate": 0.00018617452357071215, + "loss": 1.2671, + "theoretical_loss": 3.3926815419201746, + "tokens_seen": 2540503040 + }, + { + "epoch": 0.77, + "learning_rate": 0.00018615847542627885, + "loss": 1.2221, + "theoretical_loss": 3.3926746417745792, + "tokens_seen": 2540568576 + }, + { + "epoch": 0.77, + "learning_rate": 0.00018614242728184555, + "loss": 1.2606, + "theoretical_loss": 3.3926677418568136, + "tokens_seen": 2540634112 + }, + { + "epoch": 0.77, + "learning_rate": 0.00018612637913741225, + "loss": 1.3013, + "theoretical_loss": 3.392660842166864, + "tokens_seen": 2540699648 + }, + { + "epoch": 0.77, + "learning_rate": 0.00018611033099297895, + "loss": 1.2431, + "theoretical_loss": 3.3926539427047175, + "tokens_seen": 2540765184 + }, + { + "epoch": 0.77, + "learning_rate": 0.00018609428284854565, + "loss": 1.2316, + "theoretical_loss": 3.39264704347036, + "tokens_seen": 2540830720 + }, + { + "epoch": 0.77, + "learning_rate": 0.00018607823470411235, + "loss": 1.2154, + "theoretical_loss": 3.3926401444637784, + "tokens_seen": 2540896256 + }, + { + "epoch": 0.77, + "learning_rate": 0.00018606218655967905, + "loss": 1.284, + "theoretical_loss": 3.3926332456849595, + "tokens_seen": 2540961792 + }, + { + "epoch": 0.77, + "learning_rate": 0.00018604613841524575, + "loss": 1.1708, + "theoretical_loss": 3.39262634713389, + "tokens_seen": 2541027328 + }, + { + "epoch": 0.77, + "learning_rate": 0.00018603009027081245, + "loss": 1.3323, + "theoretical_loss": 3.392619448810556, + "tokens_seen": 2541092864 + }, + { + "epoch": 0.77, + "learning_rate": 0.00018601404212637915, + "loss": 1.3273, + "theoretical_loss": 3.3926125507149445, + "tokens_seen": 2541158400 + }, + { + "epoch": 0.77, + "learning_rate": 0.00018599799398194585, + "loss": 1.2215, + "theoretical_loss": 3.3926056528470423, + "tokens_seen": 2541223936 + }, + { + "epoch": 0.77, + "learning_rate": 0.00018598194583751253, + "loss": 1.3077, + "theoretical_loss": 3.3925987552068353, + "tokens_seen": 2541289472 + }, + { + "epoch": 0.77, + "learning_rate": 0.00018596589769307923, + "loss": 1.2927, + "theoretical_loss": 3.3925918577943106, + "tokens_seen": 2541355008 + }, + { + "epoch": 0.77, + "learning_rate": 0.00018594984954864593, + "loss": 1.1974, + "theoretical_loss": 3.392584960609455, + "tokens_seen": 2541420544 + }, + { + "epoch": 0.77, + "learning_rate": 0.00018593380140421263, + "loss": 1.2917, + "theoretical_loss": 3.3925780636522544, + "tokens_seen": 2541486080 + }, + { + "epoch": 0.77, + "learning_rate": 0.00018591775325977933, + "loss": 1.2742, + "theoretical_loss": 3.3925711669226963, + "tokens_seen": 2541551616 + }, + { + "epoch": 0.77, + "learning_rate": 0.00018590170511534603, + "loss": 1.3318, + "theoretical_loss": 3.3925642704207664, + "tokens_seen": 2541617152 + }, + { + "epoch": 0.77, + "learning_rate": 0.00018588565697091276, + "loss": 1.2057, + "theoretical_loss": 3.392557374146452, + "tokens_seen": 2541682688 + }, + { + "epoch": 0.77, + "learning_rate": 0.00018586960882647946, + "loss": 1.2773, + "theoretical_loss": 3.3925504780997398, + "tokens_seen": 2541748224 + }, + { + "epoch": 0.77, + "learning_rate": 0.00018585356068204616, + "loss": 1.3355, + "theoretical_loss": 3.392543582280616, + "tokens_seen": 2541813760 + }, + { + "epoch": 0.77, + "learning_rate": 0.00018583751253761286, + "loss": 1.3112, + "theoretical_loss": 3.392536686689067, + "tokens_seen": 2541879296 + }, + { + "epoch": 0.77, + "learning_rate": 0.00018582146439317956, + "loss": 1.3274, + "theoretical_loss": 3.3925297913250807, + "tokens_seen": 2541944832 + }, + { + "epoch": 0.77, + "learning_rate": 0.00018580541624874626, + "loss": 1.309, + "theoretical_loss": 3.392522896188642, + "tokens_seen": 2542010368 + }, + { + "epoch": 0.77, + "learning_rate": 0.00018578936810431296, + "loss": 1.2835, + "theoretical_loss": 3.3925160012797386, + "tokens_seen": 2542075904 + }, + { + "epoch": 0.77, + "learning_rate": 0.00018577331995987966, + "loss": 1.2379, + "theoretical_loss": 3.392509106598357, + "tokens_seen": 2542141440 + }, + { + "epoch": 0.77, + "learning_rate": 0.00018575727181544636, + "loss": 1.2805, + "theoretical_loss": 3.392502212144483, + "tokens_seen": 2542206976 + }, + { + "epoch": 0.77, + "learning_rate": 0.00018574122367101306, + "loss": 1.309, + "theoretical_loss": 3.392495317918105, + "tokens_seen": 2542272512 + }, + { + "epoch": 0.77, + "learning_rate": 0.00018572517552657976, + "loss": 1.4298, + "theoretical_loss": 3.392488423919208, + "tokens_seen": 2542338048 + }, + { + "epoch": 0.77, + "learning_rate": 0.00018570912738214646, + "loss": 1.1893, + "theoretical_loss": 3.392481530147779, + "tokens_seen": 2542403584 + }, + { + "epoch": 0.77, + "learning_rate": 0.00018569307923771316, + "loss": 1.3159, + "theoretical_loss": 3.392474636603805, + "tokens_seen": 2542469120 + }, + { + "epoch": 0.77, + "learning_rate": 0.00018567703109327987, + "loss": 1.3245, + "theoretical_loss": 3.3924677432872725, + "tokens_seen": 2542534656 + }, + { + "epoch": 0.77, + "learning_rate": 0.00018566098294884654, + "loss": 1.2865, + "theoretical_loss": 3.392460850198168, + "tokens_seen": 2542600192 + }, + { + "epoch": 0.77, + "learning_rate": 0.00018564493480441324, + "loss": 1.3896, + "theoretical_loss": 3.3924539573364783, + "tokens_seen": 2542665728 + }, + { + "epoch": 0.77, + "learning_rate": 0.00018562888665997994, + "loss": 1.2312, + "theoretical_loss": 3.3924470647021896, + "tokens_seen": 2542731264 + }, + { + "epoch": 0.77, + "objective/train/docs_used": 925229, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.0784614086151123, + "objective/train/theoretical_loss": 3.3924401722952893, + "objective/train/tokens_used": 2563256800, + "theoretical_loss": 3.3924401722952893, + "tokens_seen": 2542796800 + }, + { + "epoch": 0.77, + "learning_rate": 0.00018561283851554664, + "loss": 1.1881, + "theoretical_loss": 3.3924401722952893, + "tokens_seen": 2542796800 + }, + { + "epoch": 0.77, + "learning_rate": 0.00018559679037111334, + "loss": 1.3486, + "theoretical_loss": 3.3924332801157635, + "tokens_seen": 2542862336 + }, + { + "epoch": 0.77, + "learning_rate": 0.00018558074222668004, + "loss": 1.3321, + "theoretical_loss": 3.392426388163599, + "tokens_seen": 2542927872 + }, + { + "epoch": 0.77, + "learning_rate": 0.00018556469408224674, + "loss": 1.3033, + "theoretical_loss": 3.392419496438782, + "tokens_seen": 2542993408 + }, + { + "epoch": 0.77, + "learning_rate": 0.00018554864593781344, + "loss": 1.4195, + "theoretical_loss": 3.3924126049413, + "tokens_seen": 2543058944 + }, + { + "epoch": 0.77, + "learning_rate": 0.00018553259779338014, + "loss": 1.3068, + "theoretical_loss": 3.392405713671139, + "tokens_seen": 2543124480 + }, + { + "epoch": 0.77, + "learning_rate": 0.00018551654964894684, + "loss": 1.3099, + "theoretical_loss": 3.392398822628286, + "tokens_seen": 2543190016 + }, + { + "epoch": 0.77, + "learning_rate": 0.00018550050150451355, + "loss": 1.2963, + "theoretical_loss": 3.3923919318127274, + "tokens_seen": 2543255552 + }, + { + "epoch": 0.77, + "learning_rate": 0.00018548445336008025, + "loss": 1.288, + "theoretical_loss": 3.3923850412244496, + "tokens_seen": 2543321088 + }, + { + "epoch": 0.77, + "learning_rate": 0.00018546840521564695, + "loss": 1.2582, + "theoretical_loss": 3.3923781508634403, + "tokens_seen": 2543386624 + }, + { + "epoch": 0.77, + "learning_rate": 0.00018545235707121365, + "loss": 1.3272, + "theoretical_loss": 3.3923712607296848, + "tokens_seen": 2543452160 + }, + { + "epoch": 0.77, + "learning_rate": 0.00018543630892678035, + "loss": 1.386, + "theoretical_loss": 3.3923643708231705, + "tokens_seen": 2543517696 + }, + { + "epoch": 0.77, + "learning_rate": 0.00018542026078234705, + "loss": 1.3199, + "theoretical_loss": 3.392357481143884, + "tokens_seen": 2543583232 + }, + { + "epoch": 0.77, + "learning_rate": 0.00018540421263791375, + "loss": 1.2944, + "theoretical_loss": 3.3923505916918115, + "tokens_seen": 2543648768 + }, + { + "epoch": 0.77, + "learning_rate": 0.00018538816449348045, + "loss": 1.2257, + "theoretical_loss": 3.3923437024669405, + "tokens_seen": 2543714304 + }, + { + "epoch": 0.77, + "learning_rate": 0.00018537211634904715, + "loss": 1.32, + "theoretical_loss": 3.392336813469257, + "tokens_seen": 2543779840 + }, + { + "epoch": 0.77, + "learning_rate": 0.00018535606820461385, + "loss": 1.3675, + "theoretical_loss": 3.392329924698748, + "tokens_seen": 2543845376 + }, + { + "epoch": 0.77, + "learning_rate": 0.00018534002006018055, + "loss": 1.3197, + "theoretical_loss": 3.3923230361553998, + "tokens_seen": 2543910912 + }, + { + "epoch": 0.77, + "learning_rate": 0.00018532397191574725, + "loss": 1.3057, + "theoretical_loss": 3.392316147839199, + "tokens_seen": 2543976448 + }, + { + "epoch": 0.77, + "learning_rate": 0.00018530792377131395, + "loss": 1.2665, + "theoretical_loss": 3.392309259750133, + "tokens_seen": 2544041984 + }, + { + "epoch": 0.77, + "learning_rate": 0.00018529187562688065, + "loss": 1.2384, + "theoretical_loss": 3.3923023718881877, + "tokens_seen": 2544107520 + }, + { + "epoch": 0.77, + "learning_rate": 0.00018527582748244735, + "loss": 1.2687, + "theoretical_loss": 3.39229548425335, + "tokens_seen": 2544173056 + }, + { + "epoch": 0.77, + "learning_rate": 0.00018525977933801405, + "loss": 1.341, + "theoretical_loss": 3.392288596845607, + "tokens_seen": 2544238592 + }, + { + "epoch": 0.77, + "learning_rate": 0.00018524373119358076, + "loss": 1.2405, + "theoretical_loss": 3.3922817096649442, + "tokens_seen": 2544304128 + }, + { + "epoch": 0.77, + "learning_rate": 0.00018522768304914746, + "loss": 1.2002, + "theoretical_loss": 3.3922748227113497, + "tokens_seen": 2544369664 + }, + { + "epoch": 0.77, + "learning_rate": 0.00018521163490471416, + "loss": 1.2739, + "theoretical_loss": 3.392267935984809, + "tokens_seen": 2544435200 + }, + { + "epoch": 0.77, + "learning_rate": 0.00018519558676028086, + "loss": 1.1508, + "theoretical_loss": 3.3922610494853096, + "tokens_seen": 2544500736 + }, + { + "epoch": 0.77, + "learning_rate": 0.00018517953861584756, + "loss": 1.2928, + "theoretical_loss": 3.3922541632128373, + "tokens_seen": 2544566272 + }, + { + "epoch": 0.77, + "learning_rate": 0.00018516349047141426, + "loss": 1.2531, + "theoretical_loss": 3.39224727716738, + "tokens_seen": 2544631808 + }, + { + "epoch": 0.77, + "learning_rate": 0.00018514744232698096, + "loss": 1.3135, + "theoretical_loss": 3.392240391348923, + "tokens_seen": 2544697344 + }, + { + "epoch": 0.77, + "learning_rate": 0.00018513139418254766, + "loss": 1.2081, + "theoretical_loss": 3.392233505757454, + "tokens_seen": 2544762880 + }, + { + "epoch": 0.77, + "learning_rate": 0.00018511534603811436, + "loss": 1.2702, + "theoretical_loss": 3.392226620392959, + "tokens_seen": 2544828416 + }, + { + "epoch": 0.77, + "learning_rate": 0.00018509929789368106, + "loss": 1.272, + "theoretical_loss": 3.3922197352554253, + "tokens_seen": 2544893952 + }, + { + "epoch": 0.77, + "learning_rate": 0.00018508324974924776, + "loss": 1.2145, + "theoretical_loss": 3.392212850344839, + "tokens_seen": 2544959488 + }, + { + "epoch": 0.77, + "learning_rate": 0.00018506720160481446, + "loss": 1.4398, + "theoretical_loss": 3.392205965661187, + "tokens_seen": 2545025024 + }, + { + "epoch": 0.77, + "learning_rate": 0.00018505115346038116, + "loss": 1.2978, + "theoretical_loss": 3.3921990812044562, + "tokens_seen": 2545090560 + }, + { + "epoch": 0.77, + "learning_rate": 0.00018503510531594786, + "loss": 1.2344, + "theoretical_loss": 3.392192196974633, + "tokens_seen": 2545156096 + }, + { + "epoch": 0.77, + "learning_rate": 0.00018501905717151456, + "loss": 1.3244, + "theoretical_loss": 3.392185312971704, + "tokens_seen": 2545221632 + }, + { + "epoch": 0.77, + "learning_rate": 0.00018500300902708126, + "loss": 1.3058, + "theoretical_loss": 3.3921784291956563, + "tokens_seen": 2545287168 + }, + { + "epoch": 0.77, + "learning_rate": 0.00018498696088264797, + "loss": 1.4048, + "theoretical_loss": 3.3921715456464763, + "tokens_seen": 2545352704 + }, + { + "epoch": 0.77, + "learning_rate": 0.00018497091273821464, + "loss": 1.204, + "theoretical_loss": 3.392164662324151, + "tokens_seen": 2545418240 + }, + { + "epoch": 0.77, + "learning_rate": 0.00018495486459378134, + "loss": 1.2987, + "theoretical_loss": 3.392157779228666, + "tokens_seen": 2545483776 + }, + { + "epoch": 0.77, + "learning_rate": 0.00018493881644934804, + "loss": 1.2815, + "theoretical_loss": 3.3921508963600093, + "tokens_seen": 2545549312 + }, + { + "epoch": 0.77, + "learning_rate": 0.00018492276830491474, + "loss": 1.2003, + "theoretical_loss": 3.3921440137181667, + "tokens_seen": 2545614848 + }, + { + "epoch": 0.77, + "learning_rate": 0.00018490672016048144, + "loss": 1.2715, + "theoretical_loss": 3.392137131303125, + "tokens_seen": 2545680384 + }, + { + "epoch": 0.77, + "learning_rate": 0.00018489067201604814, + "loss": 1.2622, + "theoretical_loss": 3.3921302491148717, + "tokens_seen": 2545745920 + }, + { + "epoch": 0.77, + "learning_rate": 0.00018487462387161484, + "loss": 1.3345, + "theoretical_loss": 3.392123367153393, + "tokens_seen": 2545811456 + }, + { + "epoch": 0.77, + "learning_rate": 0.00018485857572718154, + "loss": 1.2008, + "theoretical_loss": 3.392116485418675, + "tokens_seen": 2545876992 + }, + { + "epoch": 0.77, + "learning_rate": 0.00018484252758274824, + "loss": 1.2015, + "theoretical_loss": 3.392109603910705, + "tokens_seen": 2545942528 + }, + { + "epoch": 0.77, + "learning_rate": 0.00018482647943831494, + "loss": 1.194, + "theoretical_loss": 3.3921027226294695, + "tokens_seen": 2546008064 + }, + { + "epoch": 0.77, + "objective/train/docs_used": 926243, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.2009003162384033, + "objective/train/theoretical_loss": 3.392095841574956, + "objective/train/tokens_used": 2566533600, + "theoretical_loss": 3.392095841574956, + "tokens_seen": 2546073600 + }, + { + "epoch": 0.77, + "learning_rate": 0.00018481043129388165, + "loss": 1.2059, + "theoretical_loss": 3.392095841574956, + "tokens_seen": 2546073600 + }, + { + "epoch": 0.77, + "learning_rate": 0.00018479438314944835, + "loss": 1.2662, + "theoretical_loss": 3.3920889607471496, + "tokens_seen": 2546139136 + }, + { + "epoch": 0.77, + "learning_rate": 0.00018477833500501505, + "loss": 1.3379, + "theoretical_loss": 3.3920820801460385, + "tokens_seen": 2546204672 + }, + { + "epoch": 0.77, + "learning_rate": 0.00018476228686058175, + "loss": 1.221, + "theoretical_loss": 3.3920751997716083, + "tokens_seen": 2546270208 + }, + { + "epoch": 0.77, + "learning_rate": 0.00018474623871614845, + "loss": 1.3131, + "theoretical_loss": 3.392068319623846, + "tokens_seen": 2546335744 + }, + { + "epoch": 0.77, + "learning_rate": 0.00018473019057171515, + "loss": 1.2162, + "theoretical_loss": 3.392061439702739, + "tokens_seen": 2546401280 + }, + { + "epoch": 0.77, + "learning_rate": 0.00018471414242728185, + "loss": 1.2508, + "theoretical_loss": 3.3920545600082734, + "tokens_seen": 2546466816 + }, + { + "epoch": 0.77, + "learning_rate": 0.00018469809428284855, + "loss": 1.1626, + "theoretical_loss": 3.392047680540436, + "tokens_seen": 2546532352 + }, + { + "epoch": 0.77, + "learning_rate": 0.00018468204613841528, + "loss": 1.2375, + "theoretical_loss": 3.392040801299213, + "tokens_seen": 2546597888 + }, + { + "epoch": 0.77, + "learning_rate": 0.00018466599799398198, + "loss": 1.2083, + "theoretical_loss": 3.392033922284592, + "tokens_seen": 2546663424 + }, + { + "epoch": 0.77, + "learning_rate": 0.00018464994984954865, + "loss": 1.2391, + "theoretical_loss": 3.392027043496559, + "tokens_seen": 2546728960 + }, + { + "epoch": 0.77, + "learning_rate": 0.00018463390170511535, + "loss": 1.2371, + "theoretical_loss": 3.392020164935101, + "tokens_seen": 2546794496 + }, + { + "epoch": 0.77, + "learning_rate": 0.00018461785356068205, + "loss": 1.2876, + "theoretical_loss": 3.392013286600205, + "tokens_seen": 2546860032 + }, + { + "epoch": 0.77, + "learning_rate": 0.00018460180541624875, + "loss": 1.2016, + "theoretical_loss": 3.392006408491857, + "tokens_seen": 2546925568 + }, + { + "epoch": 0.77, + "learning_rate": 0.00018458575727181545, + "loss": 1.2383, + "theoretical_loss": 3.391999530610044, + "tokens_seen": 2546991104 + }, + { + "epoch": 0.77, + "learning_rate": 0.00018456970912738215, + "loss": 1.1877, + "theoretical_loss": 3.3919926529547535, + "tokens_seen": 2547056640 + }, + { + "epoch": 0.77, + "learning_rate": 0.00018455366098294886, + "loss": 1.251, + "theoretical_loss": 3.391985775525971, + "tokens_seen": 2547122176 + }, + { + "epoch": 0.77, + "learning_rate": 0.00018453761283851556, + "loss": 1.2544, + "theoretical_loss": 3.391978898323684, + "tokens_seen": 2547187712 + }, + { + "epoch": 0.77, + "learning_rate": 0.00018452156469408226, + "loss": 1.195, + "theoretical_loss": 3.3919720213478786, + "tokens_seen": 2547253248 + }, + { + "epoch": 0.77, + "learning_rate": 0.00018450551654964896, + "loss": 1.1119, + "theoretical_loss": 3.391965144598542, + "tokens_seen": 2547318784 + }, + { + "epoch": 0.77, + "learning_rate": 0.00018448946840521566, + "loss": 1.3465, + "theoretical_loss": 3.391958268075661, + "tokens_seen": 2547384320 + }, + { + "epoch": 0.77, + "learning_rate": 0.00018447342026078236, + "loss": 1.2354, + "theoretical_loss": 3.391951391779222, + "tokens_seen": 2547449856 + }, + { + "epoch": 0.77, + "learning_rate": 0.00018445737211634906, + "loss": 1.171, + "theoretical_loss": 3.391944515709212, + "tokens_seen": 2547515392 + }, + { + "epoch": 0.77, + "learning_rate": 0.00018444132397191576, + "loss": 1.2756, + "theoretical_loss": 3.391937639865617, + "tokens_seen": 2547580928 + }, + { + "epoch": 0.77, + "learning_rate": 0.00018442527582748246, + "loss": 1.2694, + "theoretical_loss": 3.3919307642484244, + "tokens_seen": 2547646464 + }, + { + "epoch": 0.77, + "learning_rate": 0.00018440922768304916, + "loss": 1.1912, + "theoretical_loss": 3.391923888857621, + "tokens_seen": 2547712000 + }, + { + "epoch": 0.77, + "learning_rate": 0.00018439317953861586, + "loss": 1.3636, + "theoretical_loss": 3.3919170136931935, + "tokens_seen": 2547777536 + }, + { + "epoch": 0.77, + "learning_rate": 0.00018437713139418256, + "loss": 1.3353, + "theoretical_loss": 3.391910138755128, + "tokens_seen": 2547843072 + }, + { + "epoch": 0.77, + "learning_rate": 0.00018436108324974926, + "loss": 1.2687, + "theoretical_loss": 3.3919032640434117, + "tokens_seen": 2547908608 + }, + { + "epoch": 0.77, + "learning_rate": 0.00018434503510531596, + "loss": 1.1786, + "theoretical_loss": 3.391896389558031, + "tokens_seen": 2547974144 + }, + { + "epoch": 0.77, + "learning_rate": 0.00018432898696088266, + "loss": 1.2306, + "theoretical_loss": 3.3918895152989736, + "tokens_seen": 2548039680 + }, + { + "epoch": 0.77, + "learning_rate": 0.00018431293881644936, + "loss": 1.3165, + "theoretical_loss": 3.3918826412662253, + "tokens_seen": 2548105216 + }, + { + "epoch": 0.77, + "learning_rate": 0.00018429689067201607, + "loss": 1.2994, + "theoretical_loss": 3.391875767459773, + "tokens_seen": 2548170752 + }, + { + "epoch": 0.77, + "learning_rate": 0.00018428084252758277, + "loss": 1.2351, + "theoretical_loss": 3.3918688938796033, + "tokens_seen": 2548236288 + }, + { + "epoch": 0.77, + "learning_rate": 0.00018426479438314944, + "loss": 1.3347, + "theoretical_loss": 3.3918620205257035, + "tokens_seen": 2548301824 + }, + { + "epoch": 0.77, + "learning_rate": 0.00018424874623871614, + "loss": 1.29, + "theoretical_loss": 3.3918551473980596, + "tokens_seen": 2548367360 + }, + { + "epoch": 0.77, + "learning_rate": 0.00018423269809428284, + "loss": 1.3071, + "theoretical_loss": 3.3918482744966587, + "tokens_seen": 2548432896 + }, + { + "epoch": 0.77, + "learning_rate": 0.00018421664994984954, + "loss": 1.197, + "theoretical_loss": 3.391841401821488, + "tokens_seen": 2548498432 + }, + { + "epoch": 0.77, + "learning_rate": 0.00018420060180541624, + "loss": 1.295, + "theoretical_loss": 3.391834529372533, + "tokens_seen": 2548563968 + }, + { + "epoch": 0.77, + "learning_rate": 0.00018418455366098294, + "loss": 1.2865, + "theoretical_loss": 3.391827657149782, + "tokens_seen": 2548629504 + }, + { + "epoch": 0.77, + "learning_rate": 0.00018416850551654964, + "loss": 1.2333, + "theoretical_loss": 3.3918207851532203, + "tokens_seen": 2548695040 + }, + { + "epoch": 0.77, + "learning_rate": 0.00018415245737211634, + "loss": 1.2064, + "theoretical_loss": 3.3918139133828356, + "tokens_seen": 2548760576 + }, + { + "epoch": 0.77, + "learning_rate": 0.00018413640922768304, + "loss": 1.2593, + "theoretical_loss": 3.3918070418386144, + "tokens_seen": 2548826112 + }, + { + "epoch": 0.77, + "learning_rate": 0.00018412036108324975, + "loss": 1.3201, + "theoretical_loss": 3.391800170520543, + "tokens_seen": 2548891648 + }, + { + "epoch": 0.77, + "learning_rate": 0.00018410431293881645, + "loss": 1.1926, + "theoretical_loss": 3.391793299428609, + "tokens_seen": 2548957184 + }, + { + "epoch": 0.77, + "learning_rate": 0.00018408826479438315, + "loss": 1.2802, + "theoretical_loss": 3.391786428562798, + "tokens_seen": 2549022720 + }, + { + "epoch": 0.77, + "learning_rate": 0.00018407221664994987, + "loss": 1.2521, + "theoretical_loss": 3.3917795579230976, + "tokens_seen": 2549088256 + }, + { + "epoch": 0.77, + "learning_rate": 0.00018405616850551658, + "loss": 1.2818, + "theoretical_loss": 3.3917726875094947, + "tokens_seen": 2549153792 + }, + { + "epoch": 0.77, + "learning_rate": 0.00018404012036108328, + "loss": 1.3522, + "theoretical_loss": 3.3917658173219754, + "tokens_seen": 2549219328 + }, + { + "epoch": 0.77, + "learning_rate": 0.00018402407221664998, + "loss": 1.261, + "theoretical_loss": 3.3917589473605267, + "tokens_seen": 2549284864 + }, + { + "epoch": 0.77, + "objective/train/docs_used": 927515, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.273855209350586, + "objective/train/theoretical_loss": 3.3917520776251355, + "objective/train/tokens_used": 2569810400, + "theoretical_loss": 3.3917520776251355, + "tokens_seen": 2549350400 + }, + { + "epoch": 0.77, + "learning_rate": 0.00018400802407221668, + "loss": 1.3114, + "theoretical_loss": 3.3917520776251355, + "tokens_seen": 2549350400 + }, + { + "epoch": 0.77, + "learning_rate": 0.00018399197592778338, + "loss": 1.2768, + "theoretical_loss": 3.3917452081157884, + "tokens_seen": 2549415936 + }, + { + "epoch": 0.77, + "learning_rate": 0.00018397592778335008, + "loss": 1.1356, + "theoretical_loss": 3.391738338832472, + "tokens_seen": 2549481472 + }, + { + "epoch": 0.77, + "learning_rate": 0.00018395987963891678, + "loss": 1.4229, + "theoretical_loss": 3.3917314697751735, + "tokens_seen": 2549547008 + }, + { + "epoch": 0.77, + "learning_rate": 0.00018394383149448345, + "loss": 1.254, + "theoretical_loss": 3.391724600943879, + "tokens_seen": 2549612544 + }, + { + "epoch": 0.77, + "learning_rate": 0.00018392778335005015, + "loss": 1.2513, + "theoretical_loss": 3.391717732338576, + "tokens_seen": 2549678080 + }, + { + "epoch": 0.77, + "learning_rate": 0.00018391173520561685, + "loss": 1.1621, + "theoretical_loss": 3.3917108639592506, + "tokens_seen": 2549743616 + }, + { + "epoch": 0.77, + "learning_rate": 0.00018389568706118355, + "loss": 1.2797, + "theoretical_loss": 3.3917039958058903, + "tokens_seen": 2549809152 + }, + { + "epoch": 0.77, + "learning_rate": 0.00018387963891675025, + "loss": 1.237, + "theoretical_loss": 3.3916971278784813, + "tokens_seen": 2549874688 + }, + { + "epoch": 0.77, + "learning_rate": 0.00018386359077231696, + "loss": 1.249, + "theoretical_loss": 3.3916902601770103, + "tokens_seen": 2549940224 + }, + { + "epoch": 0.77, + "learning_rate": 0.00018384754262788366, + "loss": 1.2985, + "theoretical_loss": 3.3916833927014642, + "tokens_seen": 2550005760 + }, + { + "epoch": 0.77, + "learning_rate": 0.00018383149448345036, + "loss": 1.2805, + "theoretical_loss": 3.39167652545183, + "tokens_seen": 2550071296 + }, + { + "epoch": 0.77, + "learning_rate": 0.00018381544633901706, + "loss": 1.1962, + "theoretical_loss": 3.391669658428094, + "tokens_seen": 2550136832 + }, + { + "epoch": 0.77, + "learning_rate": 0.00018379939819458376, + "loss": 1.225, + "theoretical_loss": 3.3916627916302433, + "tokens_seen": 2550202368 + }, + { + "epoch": 0.77, + "learning_rate": 0.00018378335005015046, + "loss": 1.363, + "theoretical_loss": 3.3916559250582647, + "tokens_seen": 2550267904 + }, + { + "epoch": 0.77, + "learning_rate": 0.00018376730190571716, + "loss": 1.2816, + "theoretical_loss": 3.391649058712145, + "tokens_seen": 2550333440 + }, + { + "epoch": 0.77, + "learning_rate": 0.00018375125376128386, + "loss": 1.2336, + "theoretical_loss": 3.3916421925918705, + "tokens_seen": 2550398976 + }, + { + "epoch": 0.77, + "learning_rate": 0.00018373520561685056, + "loss": 1.3121, + "theoretical_loss": 3.3916353266974286, + "tokens_seen": 2550464512 + }, + { + "epoch": 0.77, + "learning_rate": 0.00018371915747241726, + "loss": 1.2009, + "theoretical_loss": 3.391628461028805, + "tokens_seen": 2550530048 + }, + { + "epoch": 0.77, + "learning_rate": 0.00018370310932798396, + "loss": 1.2228, + "theoretical_loss": 3.3916215955859883, + "tokens_seen": 2550595584 + }, + { + "epoch": 0.77, + "learning_rate": 0.00018368706118355066, + "loss": 1.3009, + "theoretical_loss": 3.3916147303689637, + "tokens_seen": 2550661120 + }, + { + "epoch": 0.77, + "learning_rate": 0.00018367101303911736, + "loss": 1.2676, + "theoretical_loss": 3.3916078653777184, + "tokens_seen": 2550726656 + }, + { + "epoch": 0.77, + "learning_rate": 0.00018365496489468406, + "loss": 1.3322, + "theoretical_loss": 3.3916010006122397, + "tokens_seen": 2550792192 + }, + { + "epoch": 0.77, + "learning_rate": 0.00018363891675025076, + "loss": 1.4029, + "theoretical_loss": 3.3915941360725133, + "tokens_seen": 2550857728 + }, + { + "epoch": 0.77, + "learning_rate": 0.00018362286860581747, + "loss": 1.1409, + "theoretical_loss": 3.3915872717585267, + "tokens_seen": 2550923264 + }, + { + "epoch": 0.77, + "learning_rate": 0.00018360682046138417, + "loss": 1.1849, + "theoretical_loss": 3.3915804076702667, + "tokens_seen": 2550988800 + }, + { + "epoch": 0.77, + "learning_rate": 0.00018359077231695087, + "loss": 1.1264, + "theoretical_loss": 3.39157354380772, + "tokens_seen": 2551054336 + }, + { + "epoch": 0.77, + "learning_rate": 0.00018357472417251757, + "loss": 1.2392, + "theoretical_loss": 3.3915666801708735, + "tokens_seen": 2551119872 + }, + { + "epoch": 0.77, + "learning_rate": 0.00018355867602808424, + "loss": 1.3188, + "theoretical_loss": 3.3915598167597136, + "tokens_seen": 2551185408 + }, + { + "epoch": 0.77, + "learning_rate": 0.00018354262788365094, + "loss": 1.2347, + "theoretical_loss": 3.3915529535742275, + "tokens_seen": 2551250944 + }, + { + "epoch": 0.77, + "learning_rate": 0.00018352657973921764, + "loss": 1.2423, + "theoretical_loss": 3.3915460906144017, + "tokens_seen": 2551316480 + }, + { + "epoch": 0.77, + "learning_rate": 0.00018351053159478434, + "loss": 1.1663, + "theoretical_loss": 3.391539227880223, + "tokens_seen": 2551382016 + }, + { + "epoch": 0.77, + "learning_rate": 0.00018349448345035104, + "loss": 1.2541, + "theoretical_loss": 3.391532365371678, + "tokens_seen": 2551447552 + }, + { + "epoch": 0.77, + "learning_rate": 0.00018347843530591777, + "loss": 1.1326, + "theoretical_loss": 3.3915255030887543, + "tokens_seen": 2551513088 + }, + { + "epoch": 0.77, + "learning_rate": 0.00018346238716148447, + "loss": 1.3373, + "theoretical_loss": 3.3915186410314377, + "tokens_seen": 2551578624 + }, + { + "epoch": 0.77, + "learning_rate": 0.00018344633901705117, + "loss": 1.3043, + "theoretical_loss": 3.3915117791997154, + "tokens_seen": 2551644160 + }, + { + "epoch": 0.77, + "learning_rate": 0.00018343029087261787, + "loss": 1.1934, + "theoretical_loss": 3.3915049175935743, + "tokens_seen": 2551709696 + }, + { + "epoch": 0.77, + "learning_rate": 0.00018341424272818457, + "loss": 1.2372, + "theoretical_loss": 3.3914980562130013, + "tokens_seen": 2551775232 + }, + { + "epoch": 0.77, + "learning_rate": 0.00018339819458375127, + "loss": 1.1636, + "theoretical_loss": 3.3914911950579825, + "tokens_seen": 2551840768 + }, + { + "epoch": 0.77, + "learning_rate": 0.00018338214643931797, + "loss": 1.3073, + "theoretical_loss": 3.3914843341285055, + "tokens_seen": 2551906304 + }, + { + "epoch": 0.77, + "learning_rate": 0.00018336609829488468, + "loss": 1.274, + "theoretical_loss": 3.391477473424557, + "tokens_seen": 2551971840 + }, + { + "epoch": 0.77, + "learning_rate": 0.00018335005015045138, + "loss": 1.2546, + "theoretical_loss": 3.391470612946123, + "tokens_seen": 2552037376 + }, + { + "epoch": 0.77, + "learning_rate": 0.00018333400200601808, + "loss": 1.2901, + "theoretical_loss": 3.3914637526931917, + "tokens_seen": 2552102912 + }, + { + "epoch": 0.77, + "learning_rate": 0.00018331795386158478, + "loss": 1.178, + "theoretical_loss": 3.3914568926657482, + "tokens_seen": 2552168448 + }, + { + "epoch": 0.77, + "learning_rate": 0.00018330190571715148, + "loss": 1.1686, + "theoretical_loss": 3.391450032863781, + "tokens_seen": 2552233984 + }, + { + "epoch": 0.77, + "learning_rate": 0.00018328585757271818, + "loss": 1.2933, + "theoretical_loss": 3.3914431732872754, + "tokens_seen": 2552299520 + }, + { + "epoch": 0.77, + "learning_rate": 0.00018326980942828488, + "loss": 1.324, + "theoretical_loss": 3.391436313936219, + "tokens_seen": 2552365056 + }, + { + "epoch": 0.77, + "learning_rate": 0.00018325376128385155, + "loss": 1.3764, + "theoretical_loss": 3.3914294548105985, + "tokens_seen": 2552430592 + }, + { + "epoch": 0.77, + "learning_rate": 0.00018323771313941825, + "loss": 1.1758, + "theoretical_loss": 3.3914225959104005, + "tokens_seen": 2552496128 + }, + { + "epoch": 0.77, + "learning_rate": 0.00018322166499498495, + "loss": 1.3142, + "theoretical_loss": 3.3914157372356124, + "tokens_seen": 2552561664 + }, + { + "epoch": 0.77, + "objective/train/docs_used": 928605, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.2177066802978516, + "objective/train/theoretical_loss": 3.39140887878622, + "objective/train/tokens_used": 2573087200, + "theoretical_loss": 3.39140887878622, + "tokens_seen": 2552627200 + }, + { + "epoch": 0.77, + "learning_rate": 0.00018320561685055165, + "loss": 1.1874, + "theoretical_loss": 3.39140887878622, + "tokens_seen": 2552627200 + }, + { + "epoch": 0.77, + "learning_rate": 0.00018318956870611836, + "loss": 1.3565, + "theoretical_loss": 3.391402020562211, + "tokens_seen": 2552692736 + }, + { + "epoch": 0.77, + "learning_rate": 0.00018317352056168506, + "loss": 1.2866, + "theoretical_loss": 3.3913951625635717, + "tokens_seen": 2552758272 + }, + { + "epoch": 0.77, + "learning_rate": 0.00018315747241725176, + "loss": 1.2432, + "theoretical_loss": 3.3913883047902895, + "tokens_seen": 2552823808 + }, + { + "epoch": 0.77, + "learning_rate": 0.00018314142427281846, + "loss": 1.2749, + "theoretical_loss": 3.3913814472423507, + "tokens_seen": 2552889344 + }, + { + "epoch": 0.77, + "learning_rate": 0.00018312537612838516, + "loss": 1.2217, + "theoretical_loss": 3.3913745899197423, + "tokens_seen": 2552954880 + }, + { + "epoch": 0.77, + "learning_rate": 0.00018310932798395186, + "loss": 1.1702, + "theoretical_loss": 3.3913677328224505, + "tokens_seen": 2553020416 + }, + { + "epoch": 0.77, + "learning_rate": 0.00018309327983951856, + "loss": 1.3052, + "theoretical_loss": 3.391360875950463, + "tokens_seen": 2553085952 + }, + { + "epoch": 0.77, + "learning_rate": 0.00018307723169508526, + "loss": 1.3394, + "theoretical_loss": 3.3913540193037663, + "tokens_seen": 2553151488 + }, + { + "epoch": 0.77, + "learning_rate": 0.00018306118355065196, + "loss": 1.2569, + "theoretical_loss": 3.391347162882347, + "tokens_seen": 2553217024 + }, + { + "epoch": 0.77, + "learning_rate": 0.00018304513540621866, + "loss": 1.2724, + "theoretical_loss": 3.391340306686192, + "tokens_seen": 2553282560 + }, + { + "epoch": 0.77, + "learning_rate": 0.00018302908726178536, + "loss": 1.2412, + "theoretical_loss": 3.3913334507152886, + "tokens_seen": 2553348096 + }, + { + "epoch": 0.77, + "learning_rate": 0.00018301303911735206, + "loss": 1.2652, + "theoretical_loss": 3.3913265949696227, + "tokens_seen": 2553413632 + }, + { + "epoch": 0.77, + "learning_rate": 0.00018299699097291876, + "loss": 1.1818, + "theoretical_loss": 3.391319739449182, + "tokens_seen": 2553479168 + }, + { + "epoch": 0.77, + "learning_rate": 0.00018298094282848546, + "loss": 1.3742, + "theoretical_loss": 3.391312884153953, + "tokens_seen": 2553544704 + }, + { + "epoch": 0.77, + "learning_rate": 0.00018296489468405216, + "loss": 1.1992, + "theoretical_loss": 3.3913060290839225, + "tokens_seen": 2553610240 + }, + { + "epoch": 0.77, + "learning_rate": 0.00018294884653961886, + "loss": 1.1929, + "theoretical_loss": 3.391299174239077, + "tokens_seen": 2553675776 + }, + { + "epoch": 0.77, + "learning_rate": 0.00018293279839518557, + "loss": 1.2376, + "theoretical_loss": 3.391292319619404, + "tokens_seen": 2553741312 + }, + { + "epoch": 0.77, + "learning_rate": 0.00018291675025075227, + "loss": 1.2732, + "theoretical_loss": 3.3912854652248896, + "tokens_seen": 2553806848 + }, + { + "epoch": 0.77, + "learning_rate": 0.00018290070210631897, + "loss": 1.3132, + "theoretical_loss": 3.391278611055521, + "tokens_seen": 2553872384 + }, + { + "epoch": 0.77, + "learning_rate": 0.00018288465396188567, + "loss": 1.2788, + "theoretical_loss": 3.3912717571112854, + "tokens_seen": 2553937920 + }, + { + "epoch": 0.77, + "learning_rate": 0.00018286860581745237, + "loss": 1.1946, + "theoretical_loss": 3.391264903392169, + "tokens_seen": 2554003456 + }, + { + "epoch": 0.77, + "learning_rate": 0.00018285255767301907, + "loss": 1.3055, + "theoretical_loss": 3.391258049898159, + "tokens_seen": 2554068992 + }, + { + "epoch": 0.77, + "learning_rate": 0.00018283650952858577, + "loss": 1.278, + "theoretical_loss": 3.3912511966292422, + "tokens_seen": 2554134528 + }, + { + "epoch": 0.77, + "learning_rate": 0.00018282046138415247, + "loss": 1.1984, + "theoretical_loss": 3.391244343585405, + "tokens_seen": 2554200064 + }, + { + "epoch": 0.77, + "learning_rate": 0.00018280441323971917, + "loss": 1.2666, + "theoretical_loss": 3.3912374907666347, + "tokens_seen": 2554265600 + }, + { + "epoch": 0.77, + "learning_rate": 0.00018278836509528587, + "loss": 1.2897, + "theoretical_loss": 3.391230638172918, + "tokens_seen": 2554331136 + }, + { + "epoch": 0.77, + "learning_rate": 0.00018277231695085257, + "loss": 1.239, + "theoretical_loss": 3.3912237858042413, + "tokens_seen": 2554396672 + }, + { + "epoch": 0.77, + "learning_rate": 0.00018275626880641927, + "loss": 1.2369, + "theoretical_loss": 3.3912169336605924, + "tokens_seen": 2554462208 + }, + { + "epoch": 0.77, + "learning_rate": 0.00018274022066198597, + "loss": 1.1656, + "theoretical_loss": 3.3912100817419577, + "tokens_seen": 2554527744 + }, + { + "epoch": 0.77, + "learning_rate": 0.00018272417251755267, + "loss": 1.2596, + "theoretical_loss": 3.3912032300483235, + "tokens_seen": 2554593280 + }, + { + "epoch": 0.77, + "learning_rate": 0.00018270812437311937, + "loss": 1.3025, + "theoretical_loss": 3.3911963785796777, + "tokens_seen": 2554658816 + }, + { + "epoch": 0.77, + "learning_rate": 0.00018269207622868607, + "loss": 1.219, + "theoretical_loss": 3.3911895273360058, + "tokens_seen": 2554724352 + }, + { + "epoch": 0.77, + "learning_rate": 0.00018267602808425278, + "loss": 1.3502, + "theoretical_loss": 3.391182676317296, + "tokens_seen": 2554789888 + }, + { + "epoch": 0.77, + "learning_rate": 0.00018265997993981948, + "loss": 1.3147, + "theoretical_loss": 3.391175825523534, + "tokens_seen": 2554855424 + }, + { + "epoch": 0.77, + "learning_rate": 0.00018264393179538618, + "loss": 1.2953, + "theoretical_loss": 3.3911689749547076, + "tokens_seen": 2554920960 + }, + { + "epoch": 0.77, + "learning_rate": 0.00018262788365095288, + "loss": 1.3329, + "theoretical_loss": 3.391162124610803, + "tokens_seen": 2554986496 + }, + { + "epoch": 0.77, + "learning_rate": 0.00018261183550651958, + "loss": 1.2191, + "theoretical_loss": 3.391155274491807, + "tokens_seen": 2555052032 + }, + { + "epoch": 0.77, + "learning_rate": 0.00018259578736208628, + "loss": 1.3812, + "theoretical_loss": 3.391148424597707, + "tokens_seen": 2555117568 + }, + { + "epoch": 0.77, + "learning_rate": 0.00018257973921765298, + "loss": 1.2573, + "theoretical_loss": 3.3911415749284894, + "tokens_seen": 2555183104 + }, + { + "epoch": 0.77, + "learning_rate": 0.00018256369107321968, + "loss": 1.2867, + "theoretical_loss": 3.3911347254841413, + "tokens_seen": 2555248640 + }, + { + "epoch": 0.77, + "learning_rate": 0.00018254764292878635, + "loss": 1.3218, + "theoretical_loss": 3.3911278762646493, + "tokens_seen": 2555314176 + }, + { + "epoch": 0.77, + "learning_rate": 0.00018253159478435305, + "loss": 1.36, + "theoretical_loss": 3.3911210272700005, + "tokens_seen": 2555379712 + }, + { + "epoch": 0.77, + "learning_rate": 0.00018251554663991975, + "loss": 1.2654, + "theoretical_loss": 3.3911141785001817, + "tokens_seen": 2555445248 + }, + { + "epoch": 0.77, + "learning_rate": 0.00018249949849548646, + "loss": 1.3503, + "theoretical_loss": 3.3911073299551795, + "tokens_seen": 2555510784 + }, + { + "epoch": 0.77, + "learning_rate": 0.00018248345035105316, + "loss": 1.1886, + "theoretical_loss": 3.391100481634981, + "tokens_seen": 2555576320 + }, + { + "epoch": 0.77, + "learning_rate": 0.00018246740220661986, + "loss": 1.2818, + "theoretical_loss": 3.3910936335395734, + "tokens_seen": 2555641856 + }, + { + "epoch": 0.77, + "learning_rate": 0.00018245135406218656, + "loss": 1.3196, + "theoretical_loss": 3.391086785668943, + "tokens_seen": 2555707392 + }, + { + "epoch": 0.77, + "learning_rate": 0.00018243530591775326, + "loss": 1.2272, + "theoretical_loss": 3.3910799380230765, + "tokens_seen": 2555772928 + }, + { + "epoch": 0.77, + "learning_rate": 0.00018241925777331996, + "loss": 1.2418, + "theoretical_loss": 3.391073090601961, + "tokens_seen": 2555838464 + }, + { + "epoch": 0.77, + "objective/train/docs_used": 929757, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.2973575592041016, + "objective/train/theoretical_loss": 3.391066243405584, + "objective/train/tokens_used": 2576364000, + "theoretical_loss": 3.391066243405584, + "tokens_seen": 2555904000 + }, + { + "epoch": 0.77, + "learning_rate": 0.00018240320962888666, + "loss": 1.2753, + "theoretical_loss": 3.391066243405584, + "tokens_seen": 2555904000 + }, + { + "epoch": 0.77, + "learning_rate": 0.00018238716148445336, + "loss": 1.3677, + "theoretical_loss": 3.3910593964339313, + "tokens_seen": 2555969536 + }, + { + "epoch": 0.77, + "learning_rate": 0.00018237111334002006, + "loss": 1.2575, + "theoretical_loss": 3.3910525496869908, + "tokens_seen": 2556035072 + }, + { + "epoch": 0.77, + "learning_rate": 0.00018235506519558676, + "loss": 1.3512, + "theoretical_loss": 3.3910457031647483, + "tokens_seen": 2556100608 + }, + { + "epoch": 0.77, + "learning_rate": 0.00018233901705115346, + "loss": 1.2712, + "theoretical_loss": 3.3910388568671914, + "tokens_seen": 2556166144 + }, + { + "epoch": 0.77, + "learning_rate": 0.00018232296890672016, + "loss": 1.325, + "theoretical_loss": 3.391032010794307, + "tokens_seen": 2556231680 + }, + { + "epoch": 0.77, + "learning_rate": 0.00018230692076228686, + "loss": 1.3903, + "theoretical_loss": 3.3910251649460816, + "tokens_seen": 2556297216 + }, + { + "epoch": 0.77, + "learning_rate": 0.00018229087261785356, + "loss": 1.2275, + "theoretical_loss": 3.391018319322502, + "tokens_seen": 2556362752 + }, + { + "epoch": 0.77, + "learning_rate": 0.0001822748244734203, + "loss": 1.302, + "theoretical_loss": 3.3910114739235553, + "tokens_seen": 2556428288 + }, + { + "epoch": 0.77, + "learning_rate": 0.000182258776328987, + "loss": 1.3291, + "theoretical_loss": 3.391004628749229, + "tokens_seen": 2556493824 + }, + { + "epoch": 0.77, + "learning_rate": 0.0001822427281845537, + "loss": 1.1271, + "theoretical_loss": 3.3909977837995084, + "tokens_seen": 2556559360 + }, + { + "epoch": 0.77, + "learning_rate": 0.00018222668004012037, + "loss": 1.3206, + "theoretical_loss": 3.3909909390743818, + "tokens_seen": 2556624896 + }, + { + "epoch": 0.77, + "learning_rate": 0.00018221063189568707, + "loss": 1.2122, + "theoretical_loss": 3.3909840945738354, + "tokens_seen": 2556690432 + }, + { + "epoch": 0.77, + "learning_rate": 0.00018219458375125377, + "loss": 1.2957, + "theoretical_loss": 3.390977250297856, + "tokens_seen": 2556755968 + }, + { + "epoch": 0.77, + "learning_rate": 0.00018217853560682047, + "loss": 1.3421, + "theoretical_loss": 3.390970406246431, + "tokens_seen": 2556821504 + }, + { + "epoch": 0.77, + "learning_rate": 0.00018216248746238717, + "loss": 1.2855, + "theoretical_loss": 3.390963562419547, + "tokens_seen": 2556887040 + }, + { + "epoch": 0.77, + "learning_rate": 0.00018214643931795387, + "loss": 1.1438, + "theoretical_loss": 3.3909567188171907, + "tokens_seen": 2556952576 + }, + { + "epoch": 0.77, + "learning_rate": 0.00018213039117352057, + "loss": 1.2025, + "theoretical_loss": 3.3909498754393494, + "tokens_seen": 2557018112 + }, + { + "epoch": 0.77, + "learning_rate": 0.00018211434302908727, + "loss": 1.2715, + "theoretical_loss": 3.3909430322860095, + "tokens_seen": 2557083648 + }, + { + "epoch": 0.77, + "learning_rate": 0.00018209829488465397, + "loss": 1.2951, + "theoretical_loss": 3.3909361893571583, + "tokens_seen": 2557149184 + }, + { + "epoch": 0.77, + "learning_rate": 0.00018208224674022067, + "loss": 1.3009, + "theoretical_loss": 3.3909293466527823, + "tokens_seen": 2557214720 + }, + { + "epoch": 0.77, + "learning_rate": 0.00018206619859578737, + "loss": 1.2983, + "theoretical_loss": 3.3909225041728686, + "tokens_seen": 2557280256 + }, + { + "epoch": 0.77, + "learning_rate": 0.00018205015045135407, + "loss": 1.329, + "theoretical_loss": 3.390915661917404, + "tokens_seen": 2557345792 + }, + { + "epoch": 0.77, + "learning_rate": 0.00018203410230692077, + "loss": 1.2332, + "theoretical_loss": 3.3909088198863757, + "tokens_seen": 2557411328 + }, + { + "epoch": 0.77, + "learning_rate": 0.00018201805416248747, + "loss": 1.2779, + "theoretical_loss": 3.3909019780797705, + "tokens_seen": 2557476864 + }, + { + "epoch": 0.78, + "learning_rate": 0.00018200200601805417, + "loss": 1.3243, + "theoretical_loss": 3.3908951364975746, + "tokens_seen": 2557542400 + }, + { + "epoch": 0.78, + "learning_rate": 0.00018198595787362088, + "loss": 1.3476, + "theoretical_loss": 3.3908882951397756, + "tokens_seen": 2557607936 + }, + { + "epoch": 0.78, + "learning_rate": 0.00018196990972918758, + "loss": 1.2089, + "theoretical_loss": 3.3908814540063603, + "tokens_seen": 2557673472 + }, + { + "epoch": 0.78, + "learning_rate": 0.00018195386158475428, + "loss": 1.2749, + "theoretical_loss": 3.3908746130973153, + "tokens_seen": 2557739008 + }, + { + "epoch": 0.78, + "learning_rate": 0.00018193781344032098, + "loss": 1.2698, + "theoretical_loss": 3.390867772412628, + "tokens_seen": 2557804544 + }, + { + "epoch": 0.78, + "learning_rate": 0.00018192176529588768, + "loss": 1.3149, + "theoretical_loss": 3.3908609319522847, + "tokens_seen": 2557870080 + }, + { + "epoch": 0.78, + "learning_rate": 0.00018190571715145438, + "loss": 1.3508, + "theoretical_loss": 3.390854091716273, + "tokens_seen": 2557935616 + }, + { + "epoch": 0.78, + "learning_rate": 0.00018188966900702108, + "loss": 1.2107, + "theoretical_loss": 3.390847251704579, + "tokens_seen": 2558001152 + }, + { + "epoch": 0.78, + "learning_rate": 0.00018187362086258778, + "loss": 1.3098, + "theoretical_loss": 3.3908404119171895, + "tokens_seen": 2558066688 + }, + { + "epoch": 0.78, + "learning_rate": 0.00018185757271815448, + "loss": 1.3543, + "theoretical_loss": 3.3908335723540928, + "tokens_seen": 2558132224 + }, + { + "epoch": 0.78, + "learning_rate": 0.00018184152457372115, + "loss": 1.261, + "theoretical_loss": 3.3908267330152744, + "tokens_seen": 2558197760 + }, + { + "epoch": 0.78, + "learning_rate": 0.00018182547642928785, + "loss": 1.3581, + "theoretical_loss": 3.3908198939007215, + "tokens_seen": 2558263296 + }, + { + "epoch": 0.78, + "learning_rate": 0.00018180942828485456, + "loss": 1.1934, + "theoretical_loss": 3.390813055010421, + "tokens_seen": 2558328832 + }, + { + "epoch": 0.78, + "learning_rate": 0.00018179338014042126, + "loss": 1.2679, + "theoretical_loss": 3.3908062163443606, + "tokens_seen": 2558394368 + }, + { + "epoch": 0.78, + "learning_rate": 0.00018177733199598796, + "loss": 1.3008, + "theoretical_loss": 3.3907993779025265, + "tokens_seen": 2558459904 + }, + { + "epoch": 0.78, + "learning_rate": 0.00018176128385155466, + "loss": 1.2764, + "theoretical_loss": 3.390792539684905, + "tokens_seen": 2558525440 + }, + { + "epoch": 0.78, + "learning_rate": 0.00018174523570712136, + "loss": 1.3961, + "theoretical_loss": 3.3907857016914846, + "tokens_seen": 2558590976 + }, + { + "epoch": 0.78, + "learning_rate": 0.00018172918756268806, + "loss": 1.2277, + "theoretical_loss": 3.3907788639222507, + "tokens_seen": 2558656512 + }, + { + "epoch": 0.78, + "learning_rate": 0.00018171313941825476, + "loss": 1.2642, + "theoretical_loss": 3.390772026377191, + "tokens_seen": 2558722048 + }, + { + "epoch": 0.78, + "learning_rate": 0.00018169709127382146, + "loss": 1.3382, + "theoretical_loss": 3.390765189056292, + "tokens_seen": 2558787584 + }, + { + "epoch": 0.78, + "learning_rate": 0.0001816810431293882, + "loss": 1.2519, + "theoretical_loss": 3.390758351959541, + "tokens_seen": 2558853120 + }, + { + "epoch": 0.78, + "learning_rate": 0.0001816649949849549, + "loss": 1.3054, + "theoretical_loss": 3.3907515150869245, + "tokens_seen": 2558918656 + }, + { + "epoch": 0.78, + "learning_rate": 0.0001816489468405216, + "loss": 1.2721, + "theoretical_loss": 3.39074467843843, + "tokens_seen": 2558984192 + }, + { + "epoch": 0.78, + "learning_rate": 0.0001816328986960883, + "loss": 1.2129, + "theoretical_loss": 3.390737842014044, + "tokens_seen": 2559049728 + }, + { + "epoch": 0.78, + "learning_rate": 0.000181616850551655, + "loss": 1.218, + "theoretical_loss": 3.390731005813753, + "tokens_seen": 2559115264 + }, + { + "epoch": 0.78, + "objective/train/docs_used": 930868, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.116724967956543, + "objective/train/theoretical_loss": 3.390724169837545, + "objective/train/tokens_used": 2579640800, + "theoretical_loss": 3.390724169837545, + "tokens_seen": 2559180800 + }, + { + "epoch": 0.78, + "learning_rate": 0.0001816008024072217, + "loss": 1.1549, + "theoretical_loss": 3.390724169837545, + "tokens_seen": 2559180800 + }, + { + "epoch": 0.78, + "learning_rate": 0.0001815847542627884, + "loss": 1.2003, + "theoretical_loss": 3.390717334085406, + "tokens_seen": 2559246336 + }, + { + "epoch": 0.78, + "learning_rate": 0.0001815687061183551, + "loss": 1.2045, + "theoretical_loss": 3.390710498557323, + "tokens_seen": 2559311872 + }, + { + "epoch": 0.78, + "learning_rate": 0.0001815526579739218, + "loss": 1.3158, + "theoretical_loss": 3.3907036632532837, + "tokens_seen": 2559377408 + }, + { + "epoch": 0.78, + "learning_rate": 0.0001815366098294885, + "loss": 1.3326, + "theoretical_loss": 3.390696828173274, + "tokens_seen": 2559442944 + }, + { + "epoch": 0.78, + "learning_rate": 0.00018152056168505517, + "loss": 1.3027, + "theoretical_loss": 3.3906899933172814, + "tokens_seen": 2559508480 + }, + { + "epoch": 0.78, + "learning_rate": 0.00018150451354062187, + "loss": 1.2675, + "theoretical_loss": 3.3906831586852926, + "tokens_seen": 2559574016 + }, + { + "epoch": 0.78, + "learning_rate": 0.00018148846539618857, + "loss": 1.2622, + "theoretical_loss": 3.390676324277295, + "tokens_seen": 2559639552 + }, + { + "epoch": 0.78, + "learning_rate": 0.00018147241725175527, + "loss": 1.2171, + "theoretical_loss": 3.3906694900932743, + "tokens_seen": 2559705088 + }, + { + "epoch": 0.78, + "learning_rate": 0.00018145636910732197, + "loss": 1.3365, + "theoretical_loss": 3.390662656133219, + "tokens_seen": 2559770624 + }, + { + "epoch": 0.78, + "learning_rate": 0.00018144032096288867, + "loss": 1.2029, + "theoretical_loss": 3.390655822397115, + "tokens_seen": 2559836160 + }, + { + "epoch": 0.78, + "learning_rate": 0.00018142427281845537, + "loss": 1.1802, + "theoretical_loss": 3.3906489888849496, + "tokens_seen": 2559901696 + }, + { + "epoch": 0.78, + "learning_rate": 0.00018140822467402207, + "loss": 1.1868, + "theoretical_loss": 3.39064215559671, + "tokens_seen": 2559967232 + }, + { + "epoch": 0.78, + "learning_rate": 0.00018139217652958877, + "loss": 1.1732, + "theoretical_loss": 3.3906353225323826, + "tokens_seen": 2560032768 + }, + { + "epoch": 0.78, + "learning_rate": 0.00018137612838515547, + "loss": 1.1832, + "theoretical_loss": 3.3906284896919545, + "tokens_seen": 2560098304 + }, + { + "epoch": 0.78, + "learning_rate": 0.00018136008024072217, + "loss": 1.1565, + "theoretical_loss": 3.390621657075412, + "tokens_seen": 2560163840 + }, + { + "epoch": 0.78, + "learning_rate": 0.00018134403209628887, + "loss": 1.3836, + "theoretical_loss": 3.3906148246827437, + "tokens_seen": 2560229376 + }, + { + "epoch": 0.78, + "learning_rate": 0.00018132798395185557, + "loss": 1.2315, + "theoretical_loss": 3.390607992513935, + "tokens_seen": 2560294912 + }, + { + "epoch": 0.78, + "learning_rate": 0.00018131193580742228, + "loss": 1.2629, + "theoretical_loss": 3.3906011605689734, + "tokens_seen": 2560360448 + }, + { + "epoch": 0.78, + "learning_rate": 0.00018129588766298898, + "loss": 1.268, + "theoretical_loss": 3.390594328847846, + "tokens_seen": 2560425984 + }, + { + "epoch": 0.78, + "learning_rate": 0.00018127983951855568, + "loss": 1.1836, + "theoretical_loss": 3.3905874973505394, + "tokens_seen": 2560491520 + }, + { + "epoch": 0.78, + "learning_rate": 0.00018126379137412238, + "loss": 1.3001, + "theoretical_loss": 3.3905806660770406, + "tokens_seen": 2560557056 + }, + { + "epoch": 0.78, + "learning_rate": 0.00018124774322968908, + "loss": 1.3092, + "theoretical_loss": 3.3905738350273364, + "tokens_seen": 2560622592 + }, + { + "epoch": 0.78, + "learning_rate": 0.00018123169508525578, + "loss": 1.1852, + "theoretical_loss": 3.390567004201414, + "tokens_seen": 2560688128 + }, + { + "epoch": 0.78, + "learning_rate": 0.00018121564694082248, + "loss": 1.2009, + "theoretical_loss": 3.3905601735992605, + "tokens_seen": 2560753664 + }, + { + "epoch": 0.78, + "learning_rate": 0.00018119959879638918, + "loss": 1.2015, + "theoretical_loss": 3.390553343220863, + "tokens_seen": 2560819200 + }, + { + "epoch": 0.78, + "learning_rate": 0.00018118355065195588, + "loss": 1.2549, + "theoretical_loss": 3.3905465130662074, + "tokens_seen": 2560884736 + }, + { + "epoch": 0.78, + "learning_rate": 0.00018116750250752258, + "loss": 1.2068, + "theoretical_loss": 3.3905396831352816, + "tokens_seen": 2560950272 + }, + { + "epoch": 0.78, + "learning_rate": 0.00018115145436308928, + "loss": 1.3101, + "theoretical_loss": 3.3905328534280725, + "tokens_seen": 2561015808 + }, + { + "epoch": 0.78, + "learning_rate": 0.00018113540621865595, + "loss": 1.2716, + "theoretical_loss": 3.3905260239445663, + "tokens_seen": 2561081344 + }, + { + "epoch": 0.78, + "learning_rate": 0.00018111935807422266, + "loss": 1.2035, + "theoretical_loss": 3.390519194684751, + "tokens_seen": 2561146880 + }, + { + "epoch": 0.78, + "learning_rate": 0.00018110330992978936, + "loss": 1.2205, + "theoretical_loss": 3.3905123656486125, + "tokens_seen": 2561212416 + }, + { + "epoch": 0.78, + "learning_rate": 0.00018108726178535606, + "loss": 1.261, + "theoretical_loss": 3.3905055368361383, + "tokens_seen": 2561277952 + }, + { + "epoch": 0.78, + "learning_rate": 0.00018107121364092278, + "loss": 1.2881, + "theoretical_loss": 3.390498708247316, + "tokens_seen": 2561343488 + }, + { + "epoch": 0.78, + "learning_rate": 0.00018105516549648949, + "loss": 1.2541, + "theoretical_loss": 3.390491879882131, + "tokens_seen": 2561409024 + }, + { + "epoch": 0.78, + "learning_rate": 0.00018103911735205619, + "loss": 1.1685, + "theoretical_loss": 3.390485051740572, + "tokens_seen": 2561474560 + }, + { + "epoch": 0.78, + "learning_rate": 0.0001810230692076229, + "loss": 1.1907, + "theoretical_loss": 3.3904782238226243, + "tokens_seen": 2561540096 + }, + { + "epoch": 0.78, + "learning_rate": 0.0001810070210631896, + "loss": 1.2376, + "theoretical_loss": 3.3904713961282757, + "tokens_seen": 2561605632 + }, + { + "epoch": 0.78, + "learning_rate": 0.0001809909729187563, + "loss": 1.3225, + "theoretical_loss": 3.3904645686575137, + "tokens_seen": 2561671168 + }, + { + "epoch": 0.78, + "learning_rate": 0.000180974924774323, + "loss": 1.3192, + "theoretical_loss": 3.390457741410324, + "tokens_seen": 2561736704 + }, + { + "epoch": 0.78, + "learning_rate": 0.0001809588766298897, + "loss": 1.3263, + "theoretical_loss": 3.3904509143866948, + "tokens_seen": 2561802240 + }, + { + "epoch": 0.78, + "learning_rate": 0.0001809428284854564, + "loss": 1.2375, + "theoretical_loss": 3.390444087586612, + "tokens_seen": 2561867776 + }, + { + "epoch": 0.78, + "learning_rate": 0.0001809267803410231, + "loss": 1.3026, + "theoretical_loss": 3.3904372610100633, + "tokens_seen": 2561933312 + }, + { + "epoch": 0.78, + "learning_rate": 0.0001809107321965898, + "loss": 1.2144, + "theoretical_loss": 3.390430434657035, + "tokens_seen": 2561998848 + }, + { + "epoch": 0.78, + "learning_rate": 0.0001808946840521565, + "loss": 1.2583, + "theoretical_loss": 3.390423608527515, + "tokens_seen": 2562064384 + }, + { + "epoch": 0.78, + "learning_rate": 0.0001808786359077232, + "loss": 1.3306, + "theoretical_loss": 3.3904167826214895, + "tokens_seen": 2562129920 + }, + { + "epoch": 0.78, + "learning_rate": 0.0001808625877632899, + "loss": 1.2515, + "theoretical_loss": 3.3904099569389454, + "tokens_seen": 2562195456 + }, + { + "epoch": 0.78, + "learning_rate": 0.0001808465396188566, + "loss": 1.1817, + "theoretical_loss": 3.3904031314798706, + "tokens_seen": 2562260992 + }, + { + "epoch": 0.78, + "learning_rate": 0.00018083049147442327, + "loss": 1.2798, + "theoretical_loss": 3.390396306244251, + "tokens_seen": 2562326528 + }, + { + "epoch": 0.78, + "learning_rate": 0.00018081444332998997, + "loss": 1.2195, + "theoretical_loss": 3.3903894812320736, + "tokens_seen": 2562392064 + }, + { + "epoch": 0.78, + "objective/train/docs_used": 931999, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.2126555442810059, + "objective/train/theoretical_loss": 3.390382656443326, + "objective/train/tokens_used": 2582917600, + "theoretical_loss": 3.390382656443326, + "tokens_seen": 2562457600 + }, + { + "epoch": 0.78, + "learning_rate": 0.00018079839518555667, + "loss": 1.2852, + "theoretical_loss": 3.390382656443326, + "tokens_seen": 2562457600 + }, + { + "epoch": 0.78, + "learning_rate": 0.00018078234704112337, + "loss": 1.2902, + "theoretical_loss": 3.3903758318779955, + "tokens_seen": 2562523136 + }, + { + "epoch": 0.78, + "learning_rate": 0.00018076629889669007, + "loss": 1.2806, + "theoretical_loss": 3.3903690075360684, + "tokens_seen": 2562588672 + }, + { + "epoch": 0.78, + "learning_rate": 0.00018075025075225677, + "loss": 1.3921, + "theoretical_loss": 3.3903621834175315, + "tokens_seen": 2562654208 + }, + { + "epoch": 0.78, + "learning_rate": 0.00018073420260782347, + "loss": 1.2695, + "theoretical_loss": 3.390355359522372, + "tokens_seen": 2562719744 + }, + { + "epoch": 0.78, + "learning_rate": 0.00018071815446339017, + "loss": 1.3087, + "theoretical_loss": 3.3903485358505767, + "tokens_seen": 2562785280 + }, + { + "epoch": 0.78, + "learning_rate": 0.00018070210631895687, + "loss": 1.2549, + "theoretical_loss": 3.390341712402133, + "tokens_seen": 2562850816 + }, + { + "epoch": 0.78, + "learning_rate": 0.00018068605817452357, + "loss": 1.2922, + "theoretical_loss": 3.390334889177028, + "tokens_seen": 2562916352 + }, + { + "epoch": 0.78, + "learning_rate": 0.00018067001003009027, + "loss": 1.1999, + "theoretical_loss": 3.3903280661752486, + "tokens_seen": 2562981888 + }, + { + "epoch": 0.78, + "learning_rate": 0.00018065396188565697, + "loss": 1.2343, + "theoretical_loss": 3.390321243396781, + "tokens_seen": 2563047424 + }, + { + "epoch": 0.78, + "learning_rate": 0.00018063791374122367, + "loss": 1.2732, + "theoretical_loss": 3.3903144208416127, + "tokens_seen": 2563112960 + }, + { + "epoch": 0.78, + "learning_rate": 0.00018062186559679038, + "loss": 1.2524, + "theoretical_loss": 3.390307598509731, + "tokens_seen": 2563178496 + }, + { + "epoch": 0.78, + "learning_rate": 0.00018060581745235708, + "loss": 1.2289, + "theoretical_loss": 3.390300776401123, + "tokens_seen": 2563244032 + }, + { + "epoch": 0.78, + "learning_rate": 0.00018058976930792378, + "loss": 1.2721, + "theoretical_loss": 3.390293954515774, + "tokens_seen": 2563309568 + }, + { + "epoch": 0.78, + "learning_rate": 0.00018057372116349048, + "loss": 1.2983, + "theoretical_loss": 3.3902871328536732, + "tokens_seen": 2563375104 + }, + { + "epoch": 0.78, + "learning_rate": 0.00018055767301905718, + "loss": 1.2179, + "theoretical_loss": 3.3902803114148066, + "tokens_seen": 2563440640 + }, + { + "epoch": 0.78, + "learning_rate": 0.00018054162487462388, + "loss": 1.2106, + "theoretical_loss": 3.390273490199161, + "tokens_seen": 2563506176 + }, + { + "epoch": 0.78, + "learning_rate": 0.00018052557673019058, + "loss": 1.2303, + "theoretical_loss": 3.390266669206724, + "tokens_seen": 2563571712 + }, + { + "epoch": 0.78, + "learning_rate": 0.00018050952858575728, + "loss": 1.3786, + "theoretical_loss": 3.390259848437482, + "tokens_seen": 2563637248 + }, + { + "epoch": 0.78, + "learning_rate": 0.00018049348044132398, + "loss": 1.2654, + "theoretical_loss": 3.390253027891422, + "tokens_seen": 2563702784 + }, + { + "epoch": 0.78, + "learning_rate": 0.00018047743229689068, + "loss": 1.3094, + "theoretical_loss": 3.3902462075685316, + "tokens_seen": 2563768320 + }, + { + "epoch": 0.78, + "learning_rate": 0.00018046138415245738, + "loss": 1.2734, + "theoretical_loss": 3.390239387468797, + "tokens_seen": 2563833856 + }, + { + "epoch": 0.78, + "learning_rate": 0.00018044533600802408, + "loss": 1.2376, + "theoretical_loss": 3.390232567592206, + "tokens_seen": 2563899392 + }, + { + "epoch": 0.78, + "learning_rate": 0.00018042928786359078, + "loss": 1.3137, + "theoretical_loss": 3.3902257479387448, + "tokens_seen": 2563964928 + }, + { + "epoch": 0.78, + "learning_rate": 0.00018041323971915748, + "loss": 1.2882, + "theoretical_loss": 3.390218928508401, + "tokens_seen": 2564030464 + }, + { + "epoch": 0.78, + "learning_rate": 0.00018039719157472418, + "loss": 1.2776, + "theoretical_loss": 3.3902121093011615, + "tokens_seen": 2564096000 + }, + { + "epoch": 0.78, + "learning_rate": 0.00018038114343029088, + "loss": 1.1066, + "theoretical_loss": 3.390205290317013, + "tokens_seen": 2564161536 + }, + { + "epoch": 0.78, + "learning_rate": 0.00018036509528585759, + "loss": 1.2203, + "theoretical_loss": 3.3901984715559426, + "tokens_seen": 2564227072 + }, + { + "epoch": 0.78, + "learning_rate": 0.00018034904714142429, + "loss": 1.322, + "theoretical_loss": 3.3901916530179372, + "tokens_seen": 2564292608 + }, + { + "epoch": 0.78, + "learning_rate": 0.000180332998996991, + "loss": 1.142, + "theoretical_loss": 3.3901848347029846, + "tokens_seen": 2564358144 + }, + { + "epoch": 0.78, + "learning_rate": 0.0001803169508525577, + "loss": 1.2686, + "theoretical_loss": 3.390178016611071, + "tokens_seen": 2564423680 + }, + { + "epoch": 0.78, + "learning_rate": 0.0001803009027081244, + "loss": 1.258, + "theoretical_loss": 3.390171198742183, + "tokens_seen": 2564489216 + }, + { + "epoch": 0.78, + "learning_rate": 0.0001802848545636911, + "loss": 1.3258, + "theoretical_loss": 3.390164381096309, + "tokens_seen": 2564554752 + }, + { + "epoch": 0.78, + "learning_rate": 0.0001802688064192578, + "loss": 1.3221, + "theoretical_loss": 3.390157563673435, + "tokens_seen": 2564620288 + }, + { + "epoch": 0.78, + "learning_rate": 0.0001802527582748245, + "loss": 1.2682, + "theoretical_loss": 3.390150746473548, + "tokens_seen": 2564685824 + }, + { + "epoch": 0.78, + "learning_rate": 0.0001802367101303912, + "loss": 1.2724, + "theoretical_loss": 3.3901439294966353, + "tokens_seen": 2564751360 + }, + { + "epoch": 0.78, + "learning_rate": 0.0001802206619859579, + "loss": 1.2726, + "theoretical_loss": 3.390137112742684, + "tokens_seen": 2564816896 + }, + { + "epoch": 0.78, + "learning_rate": 0.0001802046138415246, + "loss": 1.2927, + "theoretical_loss": 3.3901302962116806, + "tokens_seen": 2564882432 + }, + { + "epoch": 0.78, + "learning_rate": 0.0001801885656970913, + "loss": 1.3567, + "theoretical_loss": 3.3901234799036124, + "tokens_seen": 2564947968 + }, + { + "epoch": 0.78, + "learning_rate": 0.000180172517552658, + "loss": 1.2397, + "theoretical_loss": 3.3901166638184668, + "tokens_seen": 2565013504 + }, + { + "epoch": 0.78, + "learning_rate": 0.0001801564694082247, + "loss": 1.1702, + "theoretical_loss": 3.3901098479562304, + "tokens_seen": 2565079040 + }, + { + "epoch": 0.78, + "learning_rate": 0.0001801404212637914, + "loss": 1.199, + "theoretical_loss": 3.3901030323168904, + "tokens_seen": 2565144576 + }, + { + "epoch": 0.78, + "learning_rate": 0.00018012437311935807, + "loss": 1.358, + "theoretical_loss": 3.3900962169004334, + "tokens_seen": 2565210112 + }, + { + "epoch": 0.78, + "learning_rate": 0.00018010832497492477, + "loss": 1.1871, + "theoretical_loss": 3.390089401706847, + "tokens_seen": 2565275648 + }, + { + "epoch": 0.78, + "learning_rate": 0.00018009227683049147, + "loss": 1.2688, + "theoretical_loss": 3.390082586736118, + "tokens_seen": 2565341184 + }, + { + "epoch": 0.78, + "learning_rate": 0.00018007622868605817, + "loss": 1.3602, + "theoretical_loss": 3.390075771988233, + "tokens_seen": 2565406720 + }, + { + "epoch": 0.78, + "learning_rate": 0.00018006018054162487, + "loss": 1.2136, + "theoretical_loss": 3.3900689574631797, + "tokens_seen": 2565472256 + }, + { + "epoch": 0.78, + "learning_rate": 0.00018004413239719157, + "loss": 1.213, + "theoretical_loss": 3.390062143160945, + "tokens_seen": 2565537792 + }, + { + "epoch": 0.78, + "learning_rate": 0.00018002808425275827, + "loss": 1.2728, + "theoretical_loss": 3.390055329081515, + "tokens_seen": 2565603328 + }, + { + "epoch": 0.78, + "learning_rate": 0.00018001203610832497, + "loss": 1.282, + "theoretical_loss": 3.390048515224878, + "tokens_seen": 2565668864 + }, + { + "epoch": 0.78, + "objective/train/docs_used": 933221, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.4744858741760254, + "objective/train/theoretical_loss": 3.3900417015910205, + "objective/train/tokens_used": 2586194400, + "theoretical_loss": 3.3900417015910205, + "tokens_seen": 2565734400 + }, + { + "epoch": 0.78, + "learning_rate": 0.00017999598796389167, + "loss": 1.3517, + "theoretical_loss": 3.3900417015910205, + "tokens_seen": 2565734400 + }, + { + "epoch": 0.78, + "learning_rate": 0.00017997993981945837, + "loss": 1.2484, + "theoretical_loss": 3.390034888179929, + "tokens_seen": 2565799936 + }, + { + "epoch": 0.78, + "learning_rate": 0.00017996389167502507, + "loss": 1.2673, + "theoretical_loss": 3.3900280749915916, + "tokens_seen": 2565865472 + }, + { + "epoch": 0.78, + "learning_rate": 0.00017994784353059177, + "loss": 1.2536, + "theoretical_loss": 3.3900212620259946, + "tokens_seen": 2565931008 + }, + { + "epoch": 0.78, + "learning_rate": 0.00017993179538615848, + "loss": 1.255, + "theoretical_loss": 3.3900144492831252, + "tokens_seen": 2565996544 + }, + { + "epoch": 0.78, + "learning_rate": 0.00017991574724172518, + "loss": 1.2321, + "theoretical_loss": 3.3900076367629706, + "tokens_seen": 2566062080 + }, + { + "epoch": 0.78, + "learning_rate": 0.00017989969909729188, + "loss": 1.2605, + "theoretical_loss": 3.3900008244655173, + "tokens_seen": 2566127616 + }, + { + "epoch": 0.78, + "learning_rate": 0.00017988365095285858, + "loss": 1.2972, + "theoretical_loss": 3.389994012390753, + "tokens_seen": 2566193152 + }, + { + "epoch": 0.78, + "learning_rate": 0.0001798676028084253, + "loss": 1.2728, + "theoretical_loss": 3.3899872005386644, + "tokens_seen": 2566258688 + }, + { + "epoch": 0.78, + "learning_rate": 0.000179851554663992, + "loss": 1.3035, + "theoretical_loss": 3.389980388909238, + "tokens_seen": 2566324224 + }, + { + "epoch": 0.78, + "learning_rate": 0.0001798355065195587, + "loss": 1.3186, + "theoretical_loss": 3.389973577502462, + "tokens_seen": 2566389760 + }, + { + "epoch": 0.78, + "learning_rate": 0.0001798194583751254, + "loss": 1.2745, + "theoretical_loss": 3.389966766318323, + "tokens_seen": 2566455296 + }, + { + "epoch": 0.78, + "learning_rate": 0.00017980341023069208, + "loss": 1.2172, + "theoretical_loss": 3.389959955356807, + "tokens_seen": 2566520832 + }, + { + "epoch": 0.78, + "learning_rate": 0.00017978736208625878, + "loss": 1.3696, + "theoretical_loss": 3.3899531446179028, + "tokens_seen": 2566586368 + }, + { + "epoch": 0.78, + "learning_rate": 0.00017977131394182548, + "loss": 1.3423, + "theoretical_loss": 3.389946334101596, + "tokens_seen": 2566651904 + }, + { + "epoch": 0.78, + "learning_rate": 0.00017975526579739218, + "loss": 1.2535, + "theoretical_loss": 3.3899395238078744, + "tokens_seen": 2566717440 + }, + { + "epoch": 0.78, + "learning_rate": 0.00017973921765295888, + "loss": 1.1948, + "theoretical_loss": 3.389932713736725, + "tokens_seen": 2566782976 + }, + { + "epoch": 0.78, + "learning_rate": 0.00017972316950852558, + "loss": 1.2385, + "theoretical_loss": 3.389925903888134, + "tokens_seen": 2566848512 + }, + { + "epoch": 0.78, + "learning_rate": 0.00017970712136409228, + "loss": 1.2272, + "theoretical_loss": 3.3899190942620896, + "tokens_seen": 2566914048 + }, + { + "epoch": 0.78, + "learning_rate": 0.00017969107321965898, + "loss": 1.2272, + "theoretical_loss": 3.3899122848585783, + "tokens_seen": 2566979584 + }, + { + "epoch": 0.78, + "learning_rate": 0.00017967502507522569, + "loss": 1.4213, + "theoretical_loss": 3.3899054756775873, + "tokens_seen": 2567045120 + }, + { + "epoch": 0.78, + "learning_rate": 0.00017965897693079239, + "loss": 1.1787, + "theoretical_loss": 3.3898986667191036, + "tokens_seen": 2567110656 + }, + { + "epoch": 0.78, + "learning_rate": 0.0001796429287863591, + "loss": 1.279, + "theoretical_loss": 3.389891857983114, + "tokens_seen": 2567176192 + }, + { + "epoch": 0.78, + "learning_rate": 0.0001796268806419258, + "loss": 1.3594, + "theoretical_loss": 3.389885049469606, + "tokens_seen": 2567241728 + }, + { + "epoch": 0.78, + "learning_rate": 0.0001796108324974925, + "loss": 1.2197, + "theoretical_loss": 3.3898782411785664, + "tokens_seen": 2567307264 + }, + { + "epoch": 0.78, + "learning_rate": 0.0001795947843530592, + "loss": 1.2951, + "theoretical_loss": 3.389871433109982, + "tokens_seen": 2567372800 + }, + { + "epoch": 0.78, + "learning_rate": 0.0001795787362086259, + "loss": 1.3362, + "theoretical_loss": 3.38986462526384, + "tokens_seen": 2567438336 + }, + { + "epoch": 0.78, + "learning_rate": 0.0001795626880641926, + "loss": 1.2632, + "theoretical_loss": 3.389857817640128, + "tokens_seen": 2567503872 + }, + { + "epoch": 0.78, + "learning_rate": 0.0001795466399197593, + "loss": 1.2285, + "theoretical_loss": 3.389851010238832, + "tokens_seen": 2567569408 + }, + { + "epoch": 0.78, + "learning_rate": 0.000179530591775326, + "loss": 1.2557, + "theoretical_loss": 3.3898442030599405, + "tokens_seen": 2567634944 + }, + { + "epoch": 0.78, + "learning_rate": 0.0001795145436308927, + "loss": 1.1929, + "theoretical_loss": 3.3898373961034394, + "tokens_seen": 2567700480 + }, + { + "epoch": 0.78, + "learning_rate": 0.0001794984954864594, + "loss": 1.2865, + "theoretical_loss": 3.389830589369316, + "tokens_seen": 2567766016 + }, + { + "epoch": 0.78, + "learning_rate": 0.0001794824473420261, + "loss": 1.3523, + "theoretical_loss": 3.3898237828575573, + "tokens_seen": 2567831552 + }, + { + "epoch": 0.78, + "learning_rate": 0.0001794663991975928, + "loss": 1.2693, + "theoretical_loss": 3.389816976568151, + "tokens_seen": 2567897088 + }, + { + "epoch": 0.78, + "learning_rate": 0.0001794503510531595, + "loss": 1.2852, + "theoretical_loss": 3.3898101705010832, + "tokens_seen": 2567962624 + }, + { + "epoch": 0.78, + "learning_rate": 0.0001794343029087262, + "loss": 1.2339, + "theoretical_loss": 3.3898033646563417, + "tokens_seen": 2568028160 + }, + { + "epoch": 0.78, + "learning_rate": 0.00017941825476429287, + "loss": 1.2587, + "theoretical_loss": 3.389796559033913, + "tokens_seen": 2568093696 + }, + { + "epoch": 0.78, + "learning_rate": 0.00017940220661985957, + "loss": 1.3, + "theoretical_loss": 3.389789753633785, + "tokens_seen": 2568159232 + }, + { + "epoch": 0.78, + "learning_rate": 0.00017938615847542627, + "loss": 1.2822, + "theoretical_loss": 3.389782948455944, + "tokens_seen": 2568224768 + }, + { + "epoch": 0.78, + "learning_rate": 0.00017937011033099297, + "loss": 1.3347, + "theoretical_loss": 3.3897761435003773, + "tokens_seen": 2568290304 + }, + { + "epoch": 0.78, + "learning_rate": 0.00017935406218655967, + "loss": 1.3882, + "theoretical_loss": 3.3897693387670715, + "tokens_seen": 2568355840 + }, + { + "epoch": 0.78, + "learning_rate": 0.00017933801404212637, + "loss": 1.2261, + "theoretical_loss": 3.3897625342560147, + "tokens_seen": 2568421376 + }, + { + "epoch": 0.78, + "learning_rate": 0.00017932196589769307, + "loss": 1.1746, + "theoretical_loss": 3.3897557299671934, + "tokens_seen": 2568486912 + }, + { + "epoch": 0.78, + "learning_rate": 0.00017930591775325977, + "loss": 1.2989, + "theoretical_loss": 3.3897489259005944, + "tokens_seen": 2568552448 + }, + { + "epoch": 0.78, + "learning_rate": 0.00017928986960882647, + "loss": 1.2763, + "theoretical_loss": 3.389742122056205, + "tokens_seen": 2568617984 + }, + { + "epoch": 0.78, + "learning_rate": 0.0001792738214643932, + "loss": 1.2417, + "theoretical_loss": 3.389735318434013, + "tokens_seen": 2568683520 + }, + { + "epoch": 0.78, + "learning_rate": 0.0001792577733199599, + "loss": 1.3541, + "theoretical_loss": 3.389728515034004, + "tokens_seen": 2568749056 + }, + { + "epoch": 0.78, + "learning_rate": 0.0001792417251755266, + "loss": 1.2969, + "theoretical_loss": 3.389721711856166, + "tokens_seen": 2568814592 + }, + { + "epoch": 0.78, + "learning_rate": 0.0001792256770310933, + "loss": 1.2913, + "theoretical_loss": 3.3897149089004865, + "tokens_seen": 2568880128 + }, + { + "epoch": 0.78, + "learning_rate": 0.00017920962888666, + "loss": 1.1932, + "theoretical_loss": 3.3897081061669514, + "tokens_seen": 2568945664 + }, + { + "epoch": 0.78, + "objective/train/docs_used": 934338, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.1593440771102905, + "objective/train/theoretical_loss": 3.389701303655549, + "objective/train/tokens_used": 2589471200, + "theoretical_loss": 3.389701303655549, + "tokens_seen": 2569011200 + }, + { + "epoch": 0.78, + "learning_rate": 0.0001791935807422267, + "loss": 1.1977, + "theoretical_loss": 3.389701303655549, + "tokens_seen": 2569011200 + }, + { + "epoch": 0.78, + "learning_rate": 0.0001791775325977934, + "loss": 1.3418, + "theoretical_loss": 3.3896945013662654, + "tokens_seen": 2569076736 + }, + { + "epoch": 0.78, + "learning_rate": 0.0001791614844533601, + "loss": 1.2704, + "theoretical_loss": 3.3896876992990883, + "tokens_seen": 2569142272 + }, + { + "epoch": 0.78, + "learning_rate": 0.0001791454363089268, + "loss": 1.45, + "theoretical_loss": 3.3896808974540042, + "tokens_seen": 2569207808 + }, + { + "epoch": 0.78, + "learning_rate": 0.0001791293881644935, + "loss": 1.3707, + "theoretical_loss": 3.389674095831001, + "tokens_seen": 2569273344 + }, + { + "epoch": 0.78, + "learning_rate": 0.00017911334002006018, + "loss": 1.2292, + "theoretical_loss": 3.389667294430065, + "tokens_seen": 2569338880 + }, + { + "epoch": 0.78, + "learning_rate": 0.00017909729187562688, + "loss": 1.1758, + "theoretical_loss": 3.389660493251184, + "tokens_seen": 2569404416 + }, + { + "epoch": 0.78, + "learning_rate": 0.00017908124373119358, + "loss": 1.2549, + "theoretical_loss": 3.389653692294344, + "tokens_seen": 2569469952 + }, + { + "epoch": 0.78, + "learning_rate": 0.00017906519558676028, + "loss": 1.3251, + "theoretical_loss": 3.3896468915595337, + "tokens_seen": 2569535488 + }, + { + "epoch": 0.78, + "learning_rate": 0.00017904914744232698, + "loss": 1.2503, + "theoretical_loss": 3.3896400910467386, + "tokens_seen": 2569601024 + }, + { + "epoch": 0.78, + "learning_rate": 0.00017903309929789368, + "loss": 1.3576, + "theoretical_loss": 3.3896332907559468, + "tokens_seen": 2569666560 + }, + { + "epoch": 0.78, + "learning_rate": 0.00017901705115346038, + "loss": 1.1894, + "theoretical_loss": 3.389626490687145, + "tokens_seen": 2569732096 + }, + { + "epoch": 0.78, + "learning_rate": 0.00017900100300902709, + "loss": 1.286, + "theoretical_loss": 3.3896196908403198, + "tokens_seen": 2569797632 + }, + { + "epoch": 0.78, + "learning_rate": 0.00017898495486459379, + "loss": 1.4363, + "theoretical_loss": 3.3896128912154593, + "tokens_seen": 2569863168 + }, + { + "epoch": 0.78, + "learning_rate": 0.0001789689067201605, + "loss": 1.2982, + "theoretical_loss": 3.38960609181255, + "tokens_seen": 2569928704 + }, + { + "epoch": 0.78, + "learning_rate": 0.0001789528585757272, + "loss": 1.2165, + "theoretical_loss": 3.38959929263158, + "tokens_seen": 2569994240 + }, + { + "epoch": 0.78, + "learning_rate": 0.0001789368104312939, + "loss": 1.2782, + "theoretical_loss": 3.3895924936725343, + "tokens_seen": 2570059776 + }, + { + "epoch": 0.78, + "learning_rate": 0.0001789207622868606, + "loss": 1.2345, + "theoretical_loss": 3.389585694935402, + "tokens_seen": 2570125312 + }, + { + "epoch": 0.78, + "learning_rate": 0.0001789047141424273, + "loss": 1.2578, + "theoretical_loss": 3.389578896420169, + "tokens_seen": 2570190848 + }, + { + "epoch": 0.78, + "learning_rate": 0.000178888665997994, + "loss": 1.291, + "theoretical_loss": 3.389572098126823, + "tokens_seen": 2570256384 + }, + { + "epoch": 0.78, + "learning_rate": 0.0001788726178535607, + "loss": 1.337, + "theoretical_loss": 3.389565300055351, + "tokens_seen": 2570321920 + }, + { + "epoch": 0.78, + "learning_rate": 0.0001788565697091274, + "loss": 1.2799, + "theoretical_loss": 3.3895585022057397, + "tokens_seen": 2570387456 + }, + { + "epoch": 0.78, + "learning_rate": 0.0001788405215646941, + "loss": 1.295, + "theoretical_loss": 3.389551704577977, + "tokens_seen": 2570452992 + }, + { + "epoch": 0.78, + "learning_rate": 0.0001788244734202608, + "loss": 1.2578, + "theoretical_loss": 3.3895449071720494, + "tokens_seen": 2570518528 + }, + { + "epoch": 0.78, + "learning_rate": 0.0001788084252758275, + "loss": 1.3938, + "theoretical_loss": 3.389538109987944, + "tokens_seen": 2570584064 + }, + { + "epoch": 0.78, + "learning_rate": 0.0001787923771313942, + "loss": 1.2794, + "theoretical_loss": 3.389531313025648, + "tokens_seen": 2570649600 + }, + { + "epoch": 0.78, + "learning_rate": 0.0001787763289869609, + "loss": 1.3461, + "theoretical_loss": 3.3895245162851486, + "tokens_seen": 2570715136 + }, + { + "epoch": 0.78, + "learning_rate": 0.0001787602808425276, + "loss": 1.3632, + "theoretical_loss": 3.389517719766433, + "tokens_seen": 2570780672 + }, + { + "epoch": 0.78, + "learning_rate": 0.0001787442326980943, + "loss": 1.3954, + "theoretical_loss": 3.389510923469488, + "tokens_seen": 2570846208 + }, + { + "epoch": 0.78, + "learning_rate": 0.00017872818455366097, + "loss": 1.2495, + "theoretical_loss": 3.389504127394301, + "tokens_seen": 2570911744 + }, + { + "epoch": 0.78, + "learning_rate": 0.00017871213640922767, + "loss": 1.2545, + "theoretical_loss": 3.3894973315408587, + "tokens_seen": 2570977280 + }, + { + "epoch": 0.78, + "learning_rate": 0.00017869608826479437, + "loss": 1.3637, + "theoretical_loss": 3.389490535909149, + "tokens_seen": 2571042816 + }, + { + "epoch": 0.78, + "learning_rate": 0.00017868004012036107, + "loss": 1.3396, + "theoretical_loss": 3.3894837404991582, + "tokens_seen": 2571108352 + }, + { + "epoch": 0.78, + "learning_rate": 0.0001786639919759278, + "loss": 1.2839, + "theoretical_loss": 3.3894769453108737, + "tokens_seen": 2571173888 + }, + { + "epoch": 0.78, + "learning_rate": 0.0001786479438314945, + "loss": 1.4913, + "theoretical_loss": 3.3894701503442826, + "tokens_seen": 2571239424 + }, + { + "epoch": 0.78, + "learning_rate": 0.0001786318956870612, + "loss": 1.2469, + "theoretical_loss": 3.3894633555993723, + "tokens_seen": 2571304960 + }, + { + "epoch": 0.78, + "learning_rate": 0.0001786158475426279, + "loss": 1.2321, + "theoretical_loss": 3.3894565610761296, + "tokens_seen": 2571370496 + }, + { + "epoch": 0.78, + "learning_rate": 0.0001785997993981946, + "loss": 1.3231, + "theoretical_loss": 3.3894497667745416, + "tokens_seen": 2571436032 + }, + { + "epoch": 0.78, + "learning_rate": 0.0001785837512537613, + "loss": 1.2475, + "theoretical_loss": 3.389442972694596, + "tokens_seen": 2571501568 + }, + { + "epoch": 0.78, + "learning_rate": 0.000178567703109328, + "loss": 1.1777, + "theoretical_loss": 3.389436178836279, + "tokens_seen": 2571567104 + }, + { + "epoch": 0.78, + "learning_rate": 0.0001785516549648947, + "loss": 1.3074, + "theoretical_loss": 3.3894293851995783, + "tokens_seen": 2571632640 + }, + { + "epoch": 0.78, + "learning_rate": 0.0001785356068204614, + "loss": 1.2198, + "theoretical_loss": 3.3894225917844807, + "tokens_seen": 2571698176 + }, + { + "epoch": 0.78, + "learning_rate": 0.0001785195586760281, + "loss": 1.1862, + "theoretical_loss": 3.3894157985909734, + "tokens_seen": 2571763712 + }, + { + "epoch": 0.78, + "learning_rate": 0.0001785035105315948, + "loss": 1.3304, + "theoretical_loss": 3.389409005619044, + "tokens_seen": 2571829248 + }, + { + "epoch": 0.78, + "learning_rate": 0.0001784874623871615, + "loss": 1.2921, + "theoretical_loss": 3.389402212868679, + "tokens_seen": 2571894784 + }, + { + "epoch": 0.78, + "learning_rate": 0.0001784714142427282, + "loss": 1.3207, + "theoretical_loss": 3.3893954203398664, + "tokens_seen": 2571960320 + }, + { + "epoch": 0.78, + "learning_rate": 0.0001784553660982949, + "loss": 1.3182, + "theoretical_loss": 3.3893886280325924, + "tokens_seen": 2572025856 + }, + { + "epoch": 0.78, + "learning_rate": 0.0001784393179538616, + "loss": 1.3483, + "theoretical_loss": 3.3893818359468444, + "tokens_seen": 2572091392 + }, + { + "epoch": 0.78, + "learning_rate": 0.0001784232698094283, + "loss": 1.2698, + "theoretical_loss": 3.389375044082609, + "tokens_seen": 2572156928 + }, + { + "epoch": 0.78, + "learning_rate": 0.00017840722166499498, + "loss": 1.5021, + "theoretical_loss": 3.3893682524398745, + "tokens_seen": 2572222464 + }, + { + "epoch": 0.78, + "objective/train/docs_used": 935480, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.2225403785705566, + "objective/train/theoretical_loss": 3.3893614610186273, + "objective/train/tokens_used": 2592748000, + "theoretical_loss": 3.3893614610186273, + "tokens_seen": 2572288000 + }, + { + "epoch": 0.78, + "learning_rate": 0.00017839117352056168, + "loss": 1.3223, + "theoretical_loss": 3.3893614610186273, + "tokens_seen": 2572288000 + }, + { + "epoch": 0.78, + "learning_rate": 0.00017837512537612838, + "loss": 1.3524, + "theoretical_loss": 3.389354669818855, + "tokens_seen": 2572353536 + }, + { + "epoch": 0.78, + "learning_rate": 0.00017835907723169508, + "loss": 1.3877, + "theoretical_loss": 3.3893478788405442, + "tokens_seen": 2572419072 + }, + { + "epoch": 0.78, + "learning_rate": 0.00017834302908726178, + "loss": 1.3394, + "theoretical_loss": 3.3893410880836825, + "tokens_seen": 2572484608 + }, + { + "epoch": 0.78, + "learning_rate": 0.00017832698094282848, + "loss": 1.3454, + "theoretical_loss": 3.389334297548256, + "tokens_seen": 2572550144 + }, + { + "epoch": 0.78, + "learning_rate": 0.00017831093279839519, + "loss": 1.355, + "theoretical_loss": 3.3893275072342535, + "tokens_seen": 2572615680 + }, + { + "epoch": 0.78, + "learning_rate": 0.00017829488465396189, + "loss": 1.3053, + "theoretical_loss": 3.389320717141661, + "tokens_seen": 2572681216 + }, + { + "epoch": 0.78, + "learning_rate": 0.0001782788365095286, + "loss": 1.4246, + "theoretical_loss": 3.3893139272704658, + "tokens_seen": 2572746752 + }, + { + "epoch": 0.78, + "learning_rate": 0.0001782627883650953, + "loss": 1.1965, + "theoretical_loss": 3.389307137620655, + "tokens_seen": 2572812288 + }, + { + "epoch": 0.78, + "learning_rate": 0.000178246740220662, + "loss": 1.2382, + "theoretical_loss": 3.3893003481922164, + "tokens_seen": 2572877824 + }, + { + "epoch": 0.78, + "learning_rate": 0.0001782306920762287, + "loss": 1.5456, + "theoretical_loss": 3.3892935589851363, + "tokens_seen": 2572943360 + }, + { + "epoch": 0.78, + "learning_rate": 0.0001782146439317954, + "loss": 1.2984, + "theoretical_loss": 3.3892867699994023, + "tokens_seen": 2573008896 + }, + { + "epoch": 0.78, + "learning_rate": 0.0001781985957873621, + "loss": 1.2731, + "theoretical_loss": 3.389279981235001, + "tokens_seen": 2573074432 + }, + { + "epoch": 0.78, + "learning_rate": 0.0001781825476429288, + "loss": 1.1955, + "theoretical_loss": 3.3892731926919204, + "tokens_seen": 2573139968 + }, + { + "epoch": 0.78, + "learning_rate": 0.0001781664994984955, + "loss": 1.2744, + "theoretical_loss": 3.3892664043701473, + "tokens_seen": 2573205504 + }, + { + "epoch": 0.78, + "learning_rate": 0.0001781504513540622, + "loss": 1.1933, + "theoretical_loss": 3.3892596162696687, + "tokens_seen": 2573271040 + }, + { + "epoch": 0.78, + "learning_rate": 0.0001781344032096289, + "loss": 1.3113, + "theoretical_loss": 3.3892528283904717, + "tokens_seen": 2573336576 + }, + { + "epoch": 0.78, + "learning_rate": 0.0001781183550651956, + "loss": 1.2724, + "theoretical_loss": 3.3892460407325435, + "tokens_seen": 2573402112 + }, + { + "epoch": 0.78, + "learning_rate": 0.0001781023069207623, + "loss": 1.2062, + "theoretical_loss": 3.3892392532958717, + "tokens_seen": 2573467648 + }, + { + "epoch": 0.78, + "learning_rate": 0.000178086258776329, + "loss": 1.3913, + "theoretical_loss": 3.389232466080443, + "tokens_seen": 2573533184 + }, + { + "epoch": 0.78, + "learning_rate": 0.0001780702106318957, + "loss": 1.3803, + "theoretical_loss": 3.3892256790862447, + "tokens_seen": 2573598720 + }, + { + "epoch": 0.78, + "learning_rate": 0.0001780541624874624, + "loss": 1.3175, + "theoretical_loss": 3.3892188923132633, + "tokens_seen": 2573664256 + }, + { + "epoch": 0.78, + "learning_rate": 0.0001780381143430291, + "loss": 1.419, + "theoretical_loss": 3.3892121057614872, + "tokens_seen": 2573729792 + }, + { + "epoch": 0.78, + "learning_rate": 0.0001780220661985958, + "loss": 1.2065, + "theoretical_loss": 3.3892053194309026, + "tokens_seen": 2573795328 + }, + { + "epoch": 0.78, + "learning_rate": 0.0001780060180541625, + "loss": 1.3427, + "theoretical_loss": 3.389198533321497, + "tokens_seen": 2573860864 + }, + { + "epoch": 0.78, + "learning_rate": 0.0001779899699097292, + "loss": 1.1974, + "theoretical_loss": 3.389191747433258, + "tokens_seen": 2573926400 + }, + { + "epoch": 0.78, + "learning_rate": 0.0001779739217652959, + "loss": 1.3347, + "theoretical_loss": 3.3891849617661713, + "tokens_seen": 2573991936 + }, + { + "epoch": 0.78, + "learning_rate": 0.0001779578736208626, + "loss": 1.0908, + "theoretical_loss": 3.3891781763202258, + "tokens_seen": 2574057472 + }, + { + "epoch": 0.78, + "learning_rate": 0.0001779418254764293, + "loss": 1.2123, + "theoretical_loss": 3.3891713910954078, + "tokens_seen": 2574123008 + }, + { + "epoch": 0.78, + "learning_rate": 0.000177925777331996, + "loss": 1.3218, + "theoretical_loss": 3.3891646060917044, + "tokens_seen": 2574188544 + }, + { + "epoch": 0.78, + "learning_rate": 0.0001779097291875627, + "loss": 1.3302, + "theoretical_loss": 3.389157821309103, + "tokens_seen": 2574254080 + }, + { + "epoch": 0.78, + "learning_rate": 0.0001778936810431294, + "loss": 1.3151, + "theoretical_loss": 3.3891510367475908, + "tokens_seen": 2574319616 + }, + { + "epoch": 0.78, + "learning_rate": 0.0001778776328986961, + "loss": 1.3132, + "theoretical_loss": 3.3891442524071547, + "tokens_seen": 2574385152 + }, + { + "epoch": 0.78, + "learning_rate": 0.0001778615847542628, + "loss": 1.2618, + "theoretical_loss": 3.389137468287782, + "tokens_seen": 2574450688 + }, + { + "epoch": 0.78, + "learning_rate": 0.0001778455366098295, + "loss": 1.2161, + "theoretical_loss": 3.3891306843894604, + "tokens_seen": 2574516224 + }, + { + "epoch": 0.78, + "learning_rate": 0.0001778294884653962, + "loss": 1.3203, + "theoretical_loss": 3.389123900712176, + "tokens_seen": 2574581760 + }, + { + "epoch": 0.78, + "learning_rate": 0.0001778134403209629, + "loss": 1.3036, + "theoretical_loss": 3.3891171172559167, + "tokens_seen": 2574647296 + }, + { + "epoch": 0.78, + "learning_rate": 0.0001777973921765296, + "loss": 1.3054, + "theoretical_loss": 3.38911033402067, + "tokens_seen": 2574712832 + }, + { + "epoch": 0.78, + "learning_rate": 0.0001777813440320963, + "loss": 1.2603, + "theoretical_loss": 3.3891035510064222, + "tokens_seen": 2574778368 + }, + { + "epoch": 0.78, + "learning_rate": 0.000177765295887663, + "loss": 1.2947, + "theoretical_loss": 3.389096768213161, + "tokens_seen": 2574843904 + }, + { + "epoch": 0.78, + "learning_rate": 0.0001777492477432297, + "loss": 1.2906, + "theoretical_loss": 3.389089985640873, + "tokens_seen": 2574909440 + }, + { + "epoch": 0.78, + "learning_rate": 0.0001777331995987964, + "loss": 1.2938, + "theoretical_loss": 3.3890832032895464, + "tokens_seen": 2574974976 + }, + { + "epoch": 0.78, + "learning_rate": 0.0001777171514543631, + "loss": 1.3974, + "theoretical_loss": 3.3890764211591677, + "tokens_seen": 2575040512 + }, + { + "epoch": 0.78, + "learning_rate": 0.00017770110330992978, + "loss": 1.1357, + "theoretical_loss": 3.3890696392497244, + "tokens_seen": 2575106048 + }, + { + "epoch": 0.78, + "learning_rate": 0.00017768505516549648, + "loss": 1.1018, + "theoretical_loss": 3.389062857561203, + "tokens_seen": 2575171584 + }, + { + "epoch": 0.78, + "learning_rate": 0.00017766900702106318, + "loss": 1.3432, + "theoretical_loss": 3.3890560760935915, + "tokens_seen": 2575237120 + }, + { + "epoch": 0.78, + "learning_rate": 0.00017765295887662988, + "loss": 1.2299, + "theoretical_loss": 3.3890492948468767, + "tokens_seen": 2575302656 + }, + { + "epoch": 0.78, + "learning_rate": 0.00017763691073219658, + "loss": 1.4149, + "theoretical_loss": 3.3890425138210456, + "tokens_seen": 2575368192 + }, + { + "epoch": 0.78, + "learning_rate": 0.00017762086258776329, + "loss": 1.2211, + "theoretical_loss": 3.3890357330160854, + "tokens_seen": 2575433728 + }, + { + "epoch": 0.78, + "learning_rate": 0.00017760481444332999, + "loss": 1.3349, + "theoretical_loss": 3.3890289524319837, + "tokens_seen": 2575499264 + }, + { + "epoch": 0.78, + "objective/train/docs_used": 936685, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.3407886028289795, + "objective/train/theoretical_loss": 3.389022172068728, + "objective/train/tokens_used": 2596024800, + "theoretical_loss": 3.389022172068728, + "tokens_seen": 2575564800 + }, + { + "epoch": 0.78, + "learning_rate": 0.0001775887662988967, + "loss": 1.2596, + "theoretical_loss": 3.389022172068728, + "tokens_seen": 2575564800 + }, + { + "epoch": 0.78, + "learning_rate": 0.0001775727181544634, + "loss": 1.1666, + "theoretical_loss": 3.3890153919263044, + "tokens_seen": 2575630336 + }, + { + "epoch": 0.78, + "learning_rate": 0.0001775566700100301, + "loss": 1.3373, + "theoretical_loss": 3.3890086120047007, + "tokens_seen": 2575695872 + }, + { + "epoch": 0.78, + "learning_rate": 0.0001775406218655968, + "loss": 1.3342, + "theoretical_loss": 3.389001832303904, + "tokens_seen": 2575761408 + }, + { + "epoch": 0.78, + "learning_rate": 0.0001775245737211635, + "loss": 1.3527, + "theoretical_loss": 3.3889950528239017, + "tokens_seen": 2575826944 + }, + { + "epoch": 0.78, + "learning_rate": 0.0001775085255767302, + "loss": 1.2402, + "theoretical_loss": 3.3889882735646806, + "tokens_seen": 2575892480 + }, + { + "epoch": 0.78, + "learning_rate": 0.0001774924774322969, + "loss": 1.2668, + "theoretical_loss": 3.3889814945262287, + "tokens_seen": 2575958016 + }, + { + "epoch": 0.78, + "learning_rate": 0.00017747642928786362, + "loss": 1.1415, + "theoretical_loss": 3.388974715708532, + "tokens_seen": 2576023552 + }, + { + "epoch": 0.78, + "learning_rate": 0.00017746038114343032, + "loss": 1.2977, + "theoretical_loss": 3.3889679371115786, + "tokens_seen": 2576089088 + }, + { + "epoch": 0.78, + "learning_rate": 0.00017744433299899702, + "loss": 1.3252, + "theoretical_loss": 3.3889611587353556, + "tokens_seen": 2576154624 + }, + { + "epoch": 0.78, + "learning_rate": 0.00017742828485456372, + "loss": 1.3043, + "theoretical_loss": 3.3889543805798494, + "tokens_seen": 2576220160 + }, + { + "epoch": 0.78, + "learning_rate": 0.00017741223671013042, + "loss": 1.2743, + "theoretical_loss": 3.388947602645048, + "tokens_seen": 2576285696 + }, + { + "epoch": 0.78, + "learning_rate": 0.0001773961885656971, + "loss": 1.4329, + "theoretical_loss": 3.3889408249309385, + "tokens_seen": 2576351232 + }, + { + "epoch": 0.78, + "learning_rate": 0.0001773801404212638, + "loss": 1.3492, + "theoretical_loss": 3.388934047437508, + "tokens_seen": 2576416768 + }, + { + "epoch": 0.78, + "learning_rate": 0.0001773640922768305, + "loss": 1.2856, + "theoretical_loss": 3.3889272701647437, + "tokens_seen": 2576482304 + }, + { + "epoch": 0.78, + "learning_rate": 0.0001773480441323972, + "loss": 1.3626, + "theoretical_loss": 3.388920493112633, + "tokens_seen": 2576547840 + }, + { + "epoch": 0.78, + "learning_rate": 0.0001773319959879639, + "loss": 1.3498, + "theoretical_loss": 3.388913716281163, + "tokens_seen": 2576613376 + }, + { + "epoch": 0.78, + "learning_rate": 0.0001773159478435306, + "loss": 1.1391, + "theoretical_loss": 3.3889069396703206, + "tokens_seen": 2576678912 + }, + { + "epoch": 0.78, + "learning_rate": 0.0001772998996990973, + "loss": 1.2692, + "theoretical_loss": 3.388900163280093, + "tokens_seen": 2576744448 + }, + { + "epoch": 0.78, + "learning_rate": 0.000177283851554664, + "loss": 1.2453, + "theoretical_loss": 3.3888933871104676, + "tokens_seen": 2576809984 + }, + { + "epoch": 0.78, + "learning_rate": 0.0001772678034102307, + "loss": 1.1876, + "theoretical_loss": 3.388886611161432, + "tokens_seen": 2576875520 + }, + { + "epoch": 0.78, + "learning_rate": 0.0001772517552657974, + "loss": 1.3643, + "theoretical_loss": 3.3888798354329728, + "tokens_seen": 2576941056 + }, + { + "epoch": 0.78, + "learning_rate": 0.0001772357071213641, + "loss": 1.3903, + "theoretical_loss": 3.3888730599250776, + "tokens_seen": 2577006592 + }, + { + "epoch": 0.78, + "learning_rate": 0.0001772196589769308, + "loss": 1.3803, + "theoretical_loss": 3.388866284637733, + "tokens_seen": 2577072128 + }, + { + "epoch": 0.78, + "learning_rate": 0.0001772036108324975, + "loss": 1.2882, + "theoretical_loss": 3.3888595095709273, + "tokens_seen": 2577137664 + }, + { + "epoch": 0.78, + "learning_rate": 0.0001771875626880642, + "loss": 1.2516, + "theoretical_loss": 3.388852734724647, + "tokens_seen": 2577203200 + }, + { + "epoch": 0.78, + "learning_rate": 0.0001771715145436309, + "loss": 1.1731, + "theoretical_loss": 3.388845960098879, + "tokens_seen": 2577268736 + }, + { + "epoch": 0.78, + "learning_rate": 0.0001771554663991976, + "loss": 1.2516, + "theoretical_loss": 3.3888391856936115, + "tokens_seen": 2577334272 + }, + { + "epoch": 0.78, + "learning_rate": 0.0001771394182547643, + "loss": 1.1634, + "theoretical_loss": 3.3888324115088304, + "tokens_seen": 2577399808 + }, + { + "epoch": 0.78, + "learning_rate": 0.000177123370110331, + "loss": 1.1089, + "theoretical_loss": 3.3888256375445245, + "tokens_seen": 2577465344 + }, + { + "epoch": 0.78, + "learning_rate": 0.0001771073219658977, + "loss": 1.4291, + "theoretical_loss": 3.3888188638006795, + "tokens_seen": 2577530880 + }, + { + "epoch": 0.78, + "learning_rate": 0.0001770912738214644, + "loss": 1.3211, + "theoretical_loss": 3.3888120902772836, + "tokens_seen": 2577596416 + }, + { + "epoch": 0.78, + "learning_rate": 0.0001770752256770311, + "loss": 1.2987, + "theoretical_loss": 3.3888053169743237, + "tokens_seen": 2577661952 + }, + { + "epoch": 0.78, + "learning_rate": 0.0001770591775325978, + "loss": 1.3843, + "theoretical_loss": 3.388798543891787, + "tokens_seen": 2577727488 + }, + { + "epoch": 0.78, + "learning_rate": 0.0001770431293881645, + "loss": 1.3056, + "theoretical_loss": 3.3887917710296604, + "tokens_seen": 2577793024 + }, + { + "epoch": 0.78, + "learning_rate": 0.0001770270812437312, + "loss": 1.2161, + "theoretical_loss": 3.388784998387932, + "tokens_seen": 2577858560 + }, + { + "epoch": 0.78, + "learning_rate": 0.00017701103309929788, + "loss": 1.216, + "theoretical_loss": 3.3887782259665884, + "tokens_seen": 2577924096 + }, + { + "epoch": 0.78, + "learning_rate": 0.00017699498495486458, + "loss": 1.355, + "theoretical_loss": 3.388771453765617, + "tokens_seen": 2577989632 + }, + { + "epoch": 0.78, + "learning_rate": 0.00017697893681043128, + "loss": 1.3035, + "theoretical_loss": 3.3887646817850046, + "tokens_seen": 2578055168 + }, + { + "epoch": 0.78, + "learning_rate": 0.00017696288866599798, + "loss": 1.1634, + "theoretical_loss": 3.388757910024739, + "tokens_seen": 2578120704 + }, + { + "epoch": 0.78, + "learning_rate": 0.00017694684052156469, + "loss": 1.329, + "theoretical_loss": 3.3887511384848072, + "tokens_seen": 2578186240 + }, + { + "epoch": 0.78, + "learning_rate": 0.00017693079237713139, + "loss": 1.4337, + "theoretical_loss": 3.388744367165197, + "tokens_seen": 2578251776 + }, + { + "epoch": 0.78, + "learning_rate": 0.00017691474423269809, + "loss": 1.2992, + "theoretical_loss": 3.3887375960658943, + "tokens_seen": 2578317312 + }, + { + "epoch": 0.78, + "learning_rate": 0.0001768986960882648, + "loss": 1.3457, + "theoretical_loss": 3.388730825186887, + "tokens_seen": 2578382848 + }, + { + "epoch": 0.78, + "learning_rate": 0.0001768826479438315, + "loss": 1.3758, + "theoretical_loss": 3.388724054528163, + "tokens_seen": 2578448384 + }, + { + "epoch": 0.78, + "learning_rate": 0.00017686659979939822, + "loss": 1.4555, + "theoretical_loss": 3.388717284089709, + "tokens_seen": 2578513920 + }, + { + "epoch": 0.78, + "learning_rate": 0.00017685055165496492, + "loss": 1.1951, + "theoretical_loss": 3.388710513871512, + "tokens_seen": 2578579456 + }, + { + "epoch": 0.78, + "learning_rate": 0.00017683450351053162, + "loss": 1.2231, + "theoretical_loss": 3.3887037438735597, + "tokens_seen": 2578644992 + }, + { + "epoch": 0.78, + "learning_rate": 0.00017681845536609832, + "loss": 1.4228, + "theoretical_loss": 3.3886969740958386, + "tokens_seen": 2578710528 + }, + { + "epoch": 0.78, + "learning_rate": 0.00017680240722166502, + "loss": 1.2852, + "theoretical_loss": 3.388690204538337, + "tokens_seen": 2578776064 + }, + { + "epoch": 0.78, + "objective/train/docs_used": 937774, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.1851321458816528, + "objective/train/theoretical_loss": 3.388683435201041, + "objective/train/tokens_used": 2599301600, + "theoretical_loss": 3.388683435201041, + "tokens_seen": 2578841600 + }, + { + "epoch": 0.78, + "learning_rate": 0.00017678635907723172, + "loss": 1.3241, + "theoretical_loss": 3.388683435201041, + "tokens_seen": 2578841600 + }, + { + "epoch": 0.78, + "learning_rate": 0.00017677031093279842, + "loss": 1.3753, + "theoretical_loss": 3.3886766660839385, + "tokens_seen": 2578907136 + }, + { + "epoch": 0.78, + "learning_rate": 0.00017675426278836512, + "loss": 1.1349, + "theoretical_loss": 3.3886698971870173, + "tokens_seen": 2578972672 + }, + { + "epoch": 0.78, + "learning_rate": 0.00017673821464393182, + "loss": 1.2579, + "theoretical_loss": 3.388663128510263, + "tokens_seen": 2579038208 + }, + { + "epoch": 0.78, + "learning_rate": 0.00017672216649949852, + "loss": 1.2743, + "theoretical_loss": 3.3886563600536648, + "tokens_seen": 2579103744 + }, + { + "epoch": 0.78, + "learning_rate": 0.00017670611835506522, + "loss": 1.4244, + "theoretical_loss": 3.3886495918172086, + "tokens_seen": 2579169280 + }, + { + "epoch": 0.78, + "learning_rate": 0.0001766900702106319, + "loss": 1.3607, + "theoretical_loss": 3.3886428238008817, + "tokens_seen": 2579234816 + }, + { + "epoch": 0.78, + "learning_rate": 0.0001766740220661986, + "loss": 1.2469, + "theoretical_loss": 3.3886360560046724, + "tokens_seen": 2579300352 + }, + { + "epoch": 0.78, + "learning_rate": 0.0001766579739217653, + "loss": 1.2307, + "theoretical_loss": 3.3886292884285667, + "tokens_seen": 2579365888 + }, + { + "epoch": 0.78, + "learning_rate": 0.000176641925777332, + "loss": 1.4538, + "theoretical_loss": 3.3886225210725525, + "tokens_seen": 2579431424 + }, + { + "epoch": 0.78, + "learning_rate": 0.0001766258776328987, + "loss": 1.4206, + "theoretical_loss": 3.388615753936617, + "tokens_seen": 2579496960 + }, + { + "epoch": 0.78, + "learning_rate": 0.0001766098294884654, + "loss": 1.3943, + "theoretical_loss": 3.3886089870207474, + "tokens_seen": 2579562496 + }, + { + "epoch": 0.78, + "learning_rate": 0.0001765937813440321, + "loss": 1.3807, + "theoretical_loss": 3.3886022203249304, + "tokens_seen": 2579628032 + }, + { + "epoch": 0.78, + "learning_rate": 0.0001765777331995988, + "loss": 1.2678, + "theoretical_loss": 3.3885954538491543, + "tokens_seen": 2579693568 + }, + { + "epoch": 0.78, + "learning_rate": 0.0001765616850551655, + "loss": 1.2618, + "theoretical_loss": 3.388588687593406, + "tokens_seen": 2579759104 + }, + { + "epoch": 0.78, + "learning_rate": 0.0001765456369107322, + "loss": 1.3133, + "theoretical_loss": 3.3885819215576722, + "tokens_seen": 2579824640 + }, + { + "epoch": 0.78, + "learning_rate": 0.0001765295887662989, + "loss": 1.2499, + "theoretical_loss": 3.388575155741941, + "tokens_seen": 2579890176 + }, + { + "epoch": 0.78, + "learning_rate": 0.0001765135406218656, + "loss": 1.2749, + "theoretical_loss": 3.388568390146199, + "tokens_seen": 2579955712 + }, + { + "epoch": 0.78, + "learning_rate": 0.0001764974924774323, + "loss": 1.2141, + "theoretical_loss": 3.3885616247704338, + "tokens_seen": 2580021248 + }, + { + "epoch": 0.78, + "learning_rate": 0.000176481444332999, + "loss": 1.369, + "theoretical_loss": 3.388554859614632, + "tokens_seen": 2580086784 + }, + { + "epoch": 0.78, + "learning_rate": 0.0001764653961885657, + "loss": 1.2204, + "theoretical_loss": 3.3885480946787823, + "tokens_seen": 2580152320 + }, + { + "epoch": 0.78, + "learning_rate": 0.0001764493480441324, + "loss": 1.0259, + "theoretical_loss": 3.3885413299628704, + "tokens_seen": 2580217856 + }, + { + "epoch": 0.78, + "learning_rate": 0.0001764332998996991, + "loss": 1.2564, + "theoretical_loss": 3.3885345654668844, + "tokens_seen": 2580283392 + }, + { + "epoch": 0.78, + "learning_rate": 0.0001764172517552658, + "loss": 1.3094, + "theoretical_loss": 3.3885278011908113, + "tokens_seen": 2580348928 + }, + { + "epoch": 0.78, + "learning_rate": 0.0001764012036108325, + "loss": 1.2371, + "theoretical_loss": 3.3885210371346384, + "tokens_seen": 2580414464 + }, + { + "epoch": 0.78, + "learning_rate": 0.0001763851554663992, + "loss": 1.3163, + "theoretical_loss": 3.388514273298353, + "tokens_seen": 2580480000 + }, + { + "epoch": 0.78, + "learning_rate": 0.0001763691073219659, + "loss": 1.2326, + "theoretical_loss": 3.3885075096819426, + "tokens_seen": 2580545536 + }, + { + "epoch": 0.78, + "learning_rate": 0.0001763530591775326, + "loss": 1.2598, + "theoretical_loss": 3.388500746285394, + "tokens_seen": 2580611072 + }, + { + "epoch": 0.78, + "learning_rate": 0.0001763370110330993, + "loss": 1.3535, + "theoretical_loss": 3.388493983108695, + "tokens_seen": 2580676608 + }, + { + "epoch": 0.78, + "learning_rate": 0.000176320962888666, + "loss": 1.4234, + "theoretical_loss": 3.3884872201518323, + "tokens_seen": 2580742144 + }, + { + "epoch": 0.78, + "learning_rate": 0.00017630491474423268, + "loss": 1.2621, + "theoretical_loss": 3.388480457414794, + "tokens_seen": 2580807680 + }, + { + "epoch": 0.78, + "learning_rate": 0.00017628886659979938, + "loss": 1.2475, + "theoretical_loss": 3.3884736948975664, + "tokens_seen": 2580873216 + }, + { + "epoch": 0.78, + "learning_rate": 0.0001762728184553661, + "loss": 1.2626, + "theoretical_loss": 3.3884669326001373, + "tokens_seen": 2580938752 + }, + { + "epoch": 0.78, + "learning_rate": 0.0001762567703109328, + "loss": 1.2327, + "theoretical_loss": 3.3884601705224937, + "tokens_seen": 2581004288 + }, + { + "epoch": 0.78, + "learning_rate": 0.0001762407221664995, + "loss": 1.2757, + "theoretical_loss": 3.388453408664623, + "tokens_seen": 2581069824 + }, + { + "epoch": 0.78, + "learning_rate": 0.00017622467402206621, + "loss": 1.3643, + "theoretical_loss": 3.3884466470265133, + "tokens_seen": 2581135360 + }, + { + "epoch": 0.78, + "learning_rate": 0.00017620862587763291, + "loss": 1.2722, + "theoretical_loss": 3.3884398856081503, + "tokens_seen": 2581200896 + }, + { + "epoch": 0.78, + "learning_rate": 0.00017619257773319961, + "loss": 1.3193, + "theoretical_loss": 3.3884331244095227, + "tokens_seen": 2581266432 + }, + { + "epoch": 0.78, + "learning_rate": 0.00017617652958876632, + "loss": 1.297, + "theoretical_loss": 3.388426363430617, + "tokens_seen": 2581331968 + }, + { + "epoch": 0.78, + "learning_rate": 0.00017616048144433302, + "loss": 1.234, + "theoretical_loss": 3.38841960267142, + "tokens_seen": 2581397504 + }, + { + "epoch": 0.78, + "learning_rate": 0.00017614443329989972, + "loss": 1.3594, + "theoretical_loss": 3.388412842131921, + "tokens_seen": 2581463040 + }, + { + "epoch": 0.78, + "learning_rate": 0.00017612838515546642, + "loss": 1.2866, + "theoretical_loss": 3.388406081812105, + "tokens_seen": 2581528576 + }, + { + "epoch": 0.78, + "learning_rate": 0.00017611233701103312, + "loss": 1.2651, + "theoretical_loss": 3.3883993217119603, + "tokens_seen": 2581594112 + }, + { + "epoch": 0.78, + "learning_rate": 0.00017609628886659982, + "loss": 1.1847, + "theoretical_loss": 3.3883925618314743, + "tokens_seen": 2581659648 + }, + { + "epoch": 0.78, + "learning_rate": 0.00017608024072216652, + "loss": 1.3599, + "theoretical_loss": 3.3883858021706335, + "tokens_seen": 2581725184 + }, + { + "epoch": 0.78, + "learning_rate": 0.00017606419257773322, + "loss": 1.2415, + "theoretical_loss": 3.3883790427294267, + "tokens_seen": 2581790720 + }, + { + "epoch": 0.78, + "learning_rate": 0.00017604814443329992, + "loss": 1.2482, + "theoretical_loss": 3.38837228350784, + "tokens_seen": 2581856256 + }, + { + "epoch": 0.78, + "learning_rate": 0.00017603209628886662, + "loss": 1.2792, + "theoretical_loss": 3.3883655245058604, + "tokens_seen": 2581921792 + }, + { + "epoch": 0.78, + "learning_rate": 0.00017601604814443332, + "loss": 1.2856, + "theoretical_loss": 3.388358765723476, + "tokens_seen": 2581987328 + }, + { + "epoch": 0.78, + "learning_rate": 0.00017600000000000002, + "loss": 1.2119, + "theoretical_loss": 3.388352007160674, + "tokens_seen": 2582052864 + }, + { + "epoch": 0.78, + "objective/train/docs_used": 938961, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.3607200384140015, + "objective/train/theoretical_loss": 3.3883452488174415, + "objective/train/tokens_used": 2602578400, + "theoretical_loss": 3.3883452488174415, + "tokens_seen": 2582118400 + }, + { + "epoch": 0.78, + "learning_rate": 0.0001759839518555667, + "loss": 1.3567, + "theoretical_loss": 3.3883452488174415, + "tokens_seen": 2582118400 + }, + { + "epoch": 0.78, + "learning_rate": 0.0001759679037111334, + "loss": 1.3152, + "theoretical_loss": 3.3883384906937657, + "tokens_seen": 2582183936 + }, + { + "epoch": 0.78, + "learning_rate": 0.0001759518555667001, + "loss": 1.3645, + "theoretical_loss": 3.388331732789634, + "tokens_seen": 2582249472 + }, + { + "epoch": 0.78, + "learning_rate": 0.0001759358074222668, + "loss": 1.3013, + "theoretical_loss": 3.388324975105034, + "tokens_seen": 2582315008 + }, + { + "epoch": 0.78, + "learning_rate": 0.0001759197592778335, + "loss": 1.3242, + "theoretical_loss": 3.3883182176399522, + "tokens_seen": 2582380544 + }, + { + "epoch": 0.78, + "learning_rate": 0.0001759037111334002, + "loss": 1.2171, + "theoretical_loss": 3.3883114603943767, + "tokens_seen": 2582446080 + }, + { + "epoch": 0.78, + "learning_rate": 0.0001758876629889669, + "loss": 1.244, + "theoretical_loss": 3.3883047033682945, + "tokens_seen": 2582511616 + }, + { + "epoch": 0.78, + "learning_rate": 0.0001758716148445336, + "loss": 1.3406, + "theoretical_loss": 3.388297946561693, + "tokens_seen": 2582577152 + }, + { + "epoch": 0.78, + "learning_rate": 0.0001758555667001003, + "loss": 1.2124, + "theoretical_loss": 3.3882911899745594, + "tokens_seen": 2582642688 + }, + { + "epoch": 0.78, + "learning_rate": 0.000175839518555667, + "loss": 1.2326, + "theoretical_loss": 3.388284433606881, + "tokens_seen": 2582708224 + }, + { + "epoch": 0.78, + "learning_rate": 0.0001758234704112337, + "loss": 1.2581, + "theoretical_loss": 3.388277677458645, + "tokens_seen": 2582773760 + }, + { + "epoch": 0.78, + "learning_rate": 0.0001758074222668004, + "loss": 1.3818, + "theoretical_loss": 3.388270921529839, + "tokens_seen": 2582839296 + }, + { + "epoch": 0.78, + "learning_rate": 0.0001757913741223671, + "loss": 1.2632, + "theoretical_loss": 3.38826416582045, + "tokens_seen": 2582904832 + }, + { + "epoch": 0.78, + "learning_rate": 0.0001757753259779338, + "loss": 1.2278, + "theoretical_loss": 3.3882574103304655, + "tokens_seen": 2582970368 + }, + { + "epoch": 0.78, + "learning_rate": 0.0001757592778335005, + "loss": 1.3713, + "theoretical_loss": 3.3882506550598728, + "tokens_seen": 2583035904 + }, + { + "epoch": 0.78, + "learning_rate": 0.0001757432296890672, + "loss": 1.393, + "theoretical_loss": 3.388243900008659, + "tokens_seen": 2583101440 + }, + { + "epoch": 0.78, + "learning_rate": 0.0001757271815446339, + "loss": 1.2988, + "theoretical_loss": 3.3882371451768116, + "tokens_seen": 2583166976 + }, + { + "epoch": 0.78, + "learning_rate": 0.0001757111334002006, + "loss": 1.3435, + "theoretical_loss": 3.388230390564318, + "tokens_seen": 2583232512 + }, + { + "epoch": 0.78, + "learning_rate": 0.0001756950852557673, + "loss": 1.3127, + "theoretical_loss": 3.388223636171165, + "tokens_seen": 2583298048 + }, + { + "epoch": 0.78, + "learning_rate": 0.000175679037111334, + "loss": 1.291, + "theoretical_loss": 3.388216881997341, + "tokens_seen": 2583363584 + }, + { + "epoch": 0.78, + "learning_rate": 0.0001756629889669007, + "loss": 1.2187, + "theoretical_loss": 3.388210128042832, + "tokens_seen": 2583429120 + }, + { + "epoch": 0.78, + "learning_rate": 0.0001756469408224674, + "loss": 1.2685, + "theoretical_loss": 3.388203374307626, + "tokens_seen": 2583494656 + }, + { + "epoch": 0.78, + "learning_rate": 0.0001756308926780341, + "loss": 1.3233, + "theoretical_loss": 3.3881966207917102, + "tokens_seen": 2583560192 + }, + { + "epoch": 0.78, + "learning_rate": 0.0001756148445336008, + "loss": 1.2498, + "theoretical_loss": 3.3881898674950723, + "tokens_seen": 2583625728 + }, + { + "epoch": 0.78, + "learning_rate": 0.0001755987963891675, + "loss": 1.2698, + "theoretical_loss": 3.388183114417699, + "tokens_seen": 2583691264 + }, + { + "epoch": 0.78, + "learning_rate": 0.0001755827482447342, + "loss": 1.2342, + "theoretical_loss": 3.3881763615595784, + "tokens_seen": 2583756800 + }, + { + "epoch": 0.78, + "learning_rate": 0.0001755667001003009, + "loss": 1.2085, + "theoretical_loss": 3.3881696089206965, + "tokens_seen": 2583822336 + }, + { + "epoch": 0.78, + "learning_rate": 0.0001755506519558676, + "loss": 1.2592, + "theoretical_loss": 3.388162856501042, + "tokens_seen": 2583887872 + }, + { + "epoch": 0.78, + "learning_rate": 0.00017553460381143431, + "loss": 1.3348, + "theoretical_loss": 3.3881561043006014, + "tokens_seen": 2583953408 + }, + { + "epoch": 0.78, + "learning_rate": 0.00017551855566700101, + "loss": 1.3629, + "theoretical_loss": 3.3881493523193624, + "tokens_seen": 2584018944 + }, + { + "epoch": 0.78, + "learning_rate": 0.00017550250752256772, + "loss": 1.148, + "theoretical_loss": 3.3881426005573125, + "tokens_seen": 2584084480 + }, + { + "epoch": 0.78, + "learning_rate": 0.00017548645937813442, + "loss": 1.2917, + "theoretical_loss": 3.3881358490144384, + "tokens_seen": 2584150016 + }, + { + "epoch": 0.78, + "learning_rate": 0.00017547041123370112, + "loss": 1.2643, + "theoretical_loss": 3.3881290976907277, + "tokens_seen": 2584215552 + }, + { + "epoch": 0.78, + "learning_rate": 0.00017545436308926782, + "loss": 1.3383, + "theoretical_loss": 3.3881223465861683, + "tokens_seen": 2584281088 + }, + { + "epoch": 0.78, + "learning_rate": 0.00017543831494483452, + "loss": 1.3191, + "theoretical_loss": 3.3881155957007465, + "tokens_seen": 2584346624 + }, + { + "epoch": 0.78, + "learning_rate": 0.00017542226680040122, + "loss": 1.2613, + "theoretical_loss": 3.3881088450344503, + "tokens_seen": 2584412160 + }, + { + "epoch": 0.78, + "learning_rate": 0.00017540621865596792, + "loss": 1.252, + "theoretical_loss": 3.388102094587267, + "tokens_seen": 2584477696 + }, + { + "epoch": 0.78, + "learning_rate": 0.00017539017051153462, + "loss": 1.216, + "theoretical_loss": 3.3880953443591837, + "tokens_seen": 2584543232 + }, + { + "epoch": 0.78, + "learning_rate": 0.00017537412236710132, + "loss": 1.2882, + "theoretical_loss": 3.3880885943501875, + "tokens_seen": 2584608768 + }, + { + "epoch": 0.78, + "learning_rate": 0.00017535807422266802, + "loss": 1.359, + "theoretical_loss": 3.3880818445602667, + "tokens_seen": 2584674304 + }, + { + "epoch": 0.78, + "learning_rate": 0.00017534202607823472, + "loss": 1.2515, + "theoretical_loss": 3.388075094989408, + "tokens_seen": 2584739840 + }, + { + "epoch": 0.78, + "learning_rate": 0.00017532597793380142, + "loss": 1.1875, + "theoretical_loss": 3.388068345637598, + "tokens_seen": 2584805376 + }, + { + "epoch": 0.78, + "learning_rate": 0.00017530992978936812, + "loss": 1.3637, + "theoretical_loss": 3.3880615965048255, + "tokens_seen": 2584870912 + }, + { + "epoch": 0.78, + "learning_rate": 0.00017529388164493482, + "loss": 1.225, + "theoretical_loss": 3.388054847591077, + "tokens_seen": 2584936448 + }, + { + "epoch": 0.78, + "learning_rate": 0.0001752778335005015, + "loss": 1.333, + "theoretical_loss": 3.38804809889634, + "tokens_seen": 2585001984 + }, + { + "epoch": 0.78, + "learning_rate": 0.0001752617853560682, + "loss": 1.2791, + "theoretical_loss": 3.3880413504206017, + "tokens_seen": 2585067520 + }, + { + "epoch": 0.78, + "learning_rate": 0.0001752457372116349, + "loss": 1.1539, + "theoretical_loss": 3.3880346021638497, + "tokens_seen": 2585133056 + }, + { + "epoch": 0.78, + "learning_rate": 0.0001752296890672016, + "loss": 1.339, + "theoretical_loss": 3.388027854126071, + "tokens_seen": 2585198592 + }, + { + "epoch": 0.78, + "learning_rate": 0.0001752136409227683, + "loss": 1.3633, + "theoretical_loss": 3.388021106307253, + "tokens_seen": 2585264128 + }, + { + "epoch": 0.78, + "learning_rate": 0.000175197592778335, + "loss": 1.2076, + "theoretical_loss": 3.3880143587073834, + "tokens_seen": 2585329664 + }, + { + "epoch": 0.78, + "objective/train/docs_used": 939988, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.217339277267456, + "objective/train/theoretical_loss": 3.3880076113264495, + "objective/train/tokens_used": 2605855200, + "theoretical_loss": 3.3880076113264495, + "tokens_seen": 2585395200 + }, + { + "epoch": 0.78, + "learning_rate": 0.0001751815446339017, + "loss": 1.2116, + "theoretical_loss": 3.3880076113264495, + "tokens_seen": 2585395200 + }, + { + "epoch": 0.78, + "learning_rate": 0.0001751654964894684, + "loss": 1.2189, + "theoretical_loss": 3.3880008641644386, + "tokens_seen": 2585460736 + }, + { + "epoch": 0.78, + "learning_rate": 0.0001751494483450351, + "loss": 1.2687, + "theoretical_loss": 3.3879941172213375, + "tokens_seen": 2585526272 + }, + { + "epoch": 0.78, + "learning_rate": 0.0001751334002006018, + "loss": 1.3577, + "theoretical_loss": 3.3879873704971337, + "tokens_seen": 2585591808 + }, + { + "epoch": 0.78, + "learning_rate": 0.0001751173520561685, + "loss": 1.2791, + "theoretical_loss": 3.3879806239918153, + "tokens_seen": 2585657344 + }, + { + "epoch": 0.78, + "learning_rate": 0.0001751013039117352, + "loss": 1.3112, + "theoretical_loss": 3.3879738777053694, + "tokens_seen": 2585722880 + }, + { + "epoch": 0.78, + "learning_rate": 0.0001750852557673019, + "loss": 1.084, + "theoretical_loss": 3.3879671316377826, + "tokens_seen": 2585788416 + }, + { + "epoch": 0.78, + "learning_rate": 0.00017506920762286863, + "loss": 1.198, + "theoretical_loss": 3.387960385789043, + "tokens_seen": 2585853952 + }, + { + "epoch": 0.78, + "learning_rate": 0.00017505315947843533, + "loss": 1.2481, + "theoretical_loss": 3.3879536401591377, + "tokens_seen": 2585919488 + }, + { + "epoch": 0.78, + "learning_rate": 0.00017503711133400203, + "loss": 1.2424, + "theoretical_loss": 3.387946894748054, + "tokens_seen": 2585985024 + }, + { + "epoch": 0.78, + "learning_rate": 0.00017502106318956873, + "loss": 1.223, + "theoretical_loss": 3.3879401495557797, + "tokens_seen": 2586050560 + }, + { + "epoch": 0.78, + "learning_rate": 0.00017500501504513543, + "loss": 1.3903, + "theoretical_loss": 3.3879334045823017, + "tokens_seen": 2586116096 + }, + { + "epoch": 0.78, + "learning_rate": 0.00017498896690070214, + "loss": 1.4158, + "theoretical_loss": 3.387926659827607, + "tokens_seen": 2586181632 + }, + { + "epoch": 0.78, + "learning_rate": 0.0001749729187562688, + "loss": 1.2558, + "theoretical_loss": 3.387919915291684, + "tokens_seen": 2586247168 + }, + { + "epoch": 0.78, + "learning_rate": 0.0001749568706118355, + "loss": 1.4108, + "theoretical_loss": 3.387913170974519, + "tokens_seen": 2586312704 + }, + { + "epoch": 0.78, + "learning_rate": 0.0001749408224674022, + "loss": 1.3322, + "theoretical_loss": 3.3879064268761003, + "tokens_seen": 2586378240 + }, + { + "epoch": 0.78, + "learning_rate": 0.0001749247743229689, + "loss": 1.199, + "theoretical_loss": 3.3878996829964145, + "tokens_seen": 2586443776 + }, + { + "epoch": 0.78, + "learning_rate": 0.0001749087261785356, + "loss": 1.2781, + "theoretical_loss": 3.3878929393354493, + "tokens_seen": 2586509312 + }, + { + "epoch": 0.78, + "learning_rate": 0.0001748926780341023, + "loss": 1.3168, + "theoretical_loss": 3.3878861958931923, + "tokens_seen": 2586574848 + }, + { + "epoch": 0.78, + "learning_rate": 0.000174876629889669, + "loss": 1.3734, + "theoretical_loss": 3.3878794526696305, + "tokens_seen": 2586640384 + }, + { + "epoch": 0.78, + "learning_rate": 0.0001748605817452357, + "loss": 1.2962, + "theoretical_loss": 3.387872709664751, + "tokens_seen": 2586705920 + }, + { + "epoch": 0.78, + "learning_rate": 0.00017484453360080241, + "loss": 1.2907, + "theoretical_loss": 3.387865966878542, + "tokens_seen": 2586771456 + }, + { + "epoch": 0.78, + "learning_rate": 0.00017482848545636911, + "loss": 1.3554, + "theoretical_loss": 3.38785922431099, + "tokens_seen": 2586836992 + }, + { + "epoch": 0.78, + "learning_rate": 0.00017481243731193582, + "loss": 1.4376, + "theoretical_loss": 3.387852481962083, + "tokens_seen": 2586902528 + }, + { + "epoch": 0.78, + "learning_rate": 0.00017479638916750252, + "loss": 1.4244, + "theoretical_loss": 3.3878457398318083, + "tokens_seen": 2586968064 + }, + { + "epoch": 0.78, + "learning_rate": 0.00017478034102306922, + "loss": 1.2056, + "theoretical_loss": 3.387838997920153, + "tokens_seen": 2587033600 + }, + { + "epoch": 0.78, + "learning_rate": 0.00017476429287863592, + "loss": 1.2522, + "theoretical_loss": 3.3878322562271044, + "tokens_seen": 2587099136 + }, + { + "epoch": 0.78, + "learning_rate": 0.00017474824473420262, + "loss": 1.3694, + "theoretical_loss": 3.38782551475265, + "tokens_seen": 2587164672 + }, + { + "epoch": 0.78, + "learning_rate": 0.00017473219658976932, + "loss": 1.3163, + "theoretical_loss": 3.3878187734967775, + "tokens_seen": 2587230208 + }, + { + "epoch": 0.78, + "learning_rate": 0.00017471614844533602, + "loss": 1.2358, + "theoretical_loss": 3.387812032459474, + "tokens_seen": 2587295744 + }, + { + "epoch": 0.78, + "learning_rate": 0.00017470010030090272, + "loss": 1.158, + "theoretical_loss": 3.387805291640727, + "tokens_seen": 2587361280 + }, + { + "epoch": 0.78, + "learning_rate": 0.00017468405215646942, + "loss": 1.3932, + "theoretical_loss": 3.3877985510405235, + "tokens_seen": 2587426816 + }, + { + "epoch": 0.78, + "learning_rate": 0.00017466800401203612, + "loss": 1.2403, + "theoretical_loss": 3.387791810658851, + "tokens_seen": 2587492352 + }, + { + "epoch": 0.78, + "learning_rate": 0.00017465195586760282, + "loss": 1.3663, + "theoretical_loss": 3.3877850704956978, + "tokens_seen": 2587557888 + }, + { + "epoch": 0.78, + "learning_rate": 0.00017463590772316952, + "loss": 1.3781, + "theoretical_loss": 3.3877783305510496, + "tokens_seen": 2587623424 + }, + { + "epoch": 0.78, + "learning_rate": 0.00017461985957873622, + "loss": 1.1304, + "theoretical_loss": 3.3877715908248955, + "tokens_seen": 2587688960 + }, + { + "epoch": 0.78, + "learning_rate": 0.00017460381143430292, + "loss": 1.3467, + "theoretical_loss": 3.387764851317222, + "tokens_seen": 2587754496 + }, + { + "epoch": 0.78, + "learning_rate": 0.0001745877632898696, + "loss": 1.3963, + "theoretical_loss": 3.387758112028016, + "tokens_seen": 2587820032 + }, + { + "epoch": 0.78, + "learning_rate": 0.0001745717151454363, + "loss": 1.2952, + "theoretical_loss": 3.3877513729572657, + "tokens_seen": 2587885568 + }, + { + "epoch": 0.78, + "learning_rate": 0.000174555667001003, + "loss": 1.3139, + "theoretical_loss": 3.3877446341049584, + "tokens_seen": 2587951104 + }, + { + "epoch": 0.78, + "learning_rate": 0.0001745396188565697, + "loss": 1.2417, + "theoretical_loss": 3.387737895471081, + "tokens_seen": 2588016640 + }, + { + "epoch": 0.78, + "learning_rate": 0.0001745235707121364, + "loss": 1.246, + "theoretical_loss": 3.3877311570556214, + "tokens_seen": 2588082176 + }, + { + "epoch": 0.78, + "learning_rate": 0.0001745075225677031, + "loss": 1.2256, + "theoretical_loss": 3.387724418858567, + "tokens_seen": 2588147712 + }, + { + "epoch": 0.78, + "learning_rate": 0.0001744914744232698, + "loss": 1.2849, + "theoretical_loss": 3.3877176808799043, + "tokens_seen": 2588213248 + }, + { + "epoch": 0.78, + "learning_rate": 0.0001744754262788365, + "loss": 1.4002, + "theoretical_loss": 3.3877109431196217, + "tokens_seen": 2588278784 + }, + { + "epoch": 0.78, + "learning_rate": 0.00017445937813440323, + "loss": 1.2665, + "theoretical_loss": 3.3877042055777067, + "tokens_seen": 2588344320 + }, + { + "epoch": 0.78, + "learning_rate": 0.00017444332998996993, + "loss": 1.241, + "theoretical_loss": 3.387697468254146, + "tokens_seen": 2588409856 + }, + { + "epoch": 0.78, + "learning_rate": 0.00017442728184553663, + "loss": 1.3881, + "theoretical_loss": 3.3876907311489273, + "tokens_seen": 2588475392 + }, + { + "epoch": 0.78, + "learning_rate": 0.00017441123370110333, + "loss": 1.3052, + "theoretical_loss": 3.387683994262038, + "tokens_seen": 2588540928 + }, + { + "epoch": 0.78, + "learning_rate": 0.00017439518555667003, + "loss": 1.4066, + "theoretical_loss": 3.3876772575934657, + "tokens_seen": 2588606464 + }, + { + "epoch": 0.78, + "objective/train/docs_used": 941211, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.2885043621063232, + "objective/train/theoretical_loss": 3.387670521143197, + "objective/train/tokens_used": 2609132000, + "theoretical_loss": 3.387670521143197, + "tokens_seen": 2588672000 + }, + { + "epoch": 0.78, + "learning_rate": 0.00017437913741223673, + "loss": 1.264, + "theoretical_loss": 3.387670521143197, + "tokens_seen": 2588672000 + }, + { + "epoch": 0.78, + "learning_rate": 0.00017436308926780343, + "loss": 1.3612, + "theoretical_loss": 3.3876637849112203, + "tokens_seen": 2588737536 + }, + { + "epoch": 0.78, + "learning_rate": 0.00017434704112337013, + "loss": 1.3029, + "theoretical_loss": 3.3876570488975224, + "tokens_seen": 2588803072 + }, + { + "epoch": 0.78, + "learning_rate": 0.00017433099297893683, + "loss": 1.3583, + "theoretical_loss": 3.3876503131020907, + "tokens_seen": 2588868608 + }, + { + "epoch": 0.78, + "learning_rate": 0.00017431494483450353, + "loss": 1.1725, + "theoretical_loss": 3.387643577524913, + "tokens_seen": 2588934144 + }, + { + "epoch": 0.78, + "learning_rate": 0.00017429889669007024, + "loss": 1.2993, + "theoretical_loss": 3.3876368421659766, + "tokens_seen": 2588999680 + }, + { + "epoch": 0.78, + "learning_rate": 0.00017428284854563694, + "loss": 1.2913, + "theoretical_loss": 3.3876301070252683, + "tokens_seen": 2589065216 + }, + { + "epoch": 0.78, + "learning_rate": 0.0001742668004012036, + "loss": 1.316, + "theoretical_loss": 3.3876233721027766, + "tokens_seen": 2589130752 + }, + { + "epoch": 0.78, + "learning_rate": 0.0001742507522567703, + "loss": 1.4425, + "theoretical_loss": 3.3876166373984877, + "tokens_seen": 2589196288 + }, + { + "epoch": 0.78, + "learning_rate": 0.000174234704112337, + "loss": 1.2896, + "theoretical_loss": 3.3876099029123896, + "tokens_seen": 2589261824 + }, + { + "epoch": 0.78, + "learning_rate": 0.0001742186559679037, + "loss": 1.2676, + "theoretical_loss": 3.3876031686444703, + "tokens_seen": 2589327360 + }, + { + "epoch": 0.78, + "learning_rate": 0.0001742026078234704, + "loss": 1.2579, + "theoretical_loss": 3.387596434594716, + "tokens_seen": 2589392896 + }, + { + "epoch": 0.78, + "learning_rate": 0.0001741865596790371, + "loss": 1.2145, + "theoretical_loss": 3.3875897007631153, + "tokens_seen": 2589458432 + }, + { + "epoch": 0.78, + "learning_rate": 0.00017417051153460381, + "loss": 1.2381, + "theoretical_loss": 3.3875829671496547, + "tokens_seen": 2589523968 + }, + { + "epoch": 0.78, + "learning_rate": 0.00017415446339017051, + "loss": 1.3577, + "theoretical_loss": 3.387576233754322, + "tokens_seen": 2589589504 + }, + { + "epoch": 0.78, + "learning_rate": 0.00017413841524573721, + "loss": 1.2373, + "theoretical_loss": 3.3875695005771043, + "tokens_seen": 2589655040 + }, + { + "epoch": 0.78, + "learning_rate": 0.00017412236710130392, + "loss": 1.3087, + "theoretical_loss": 3.3875627676179896, + "tokens_seen": 2589720576 + }, + { + "epoch": 0.78, + "learning_rate": 0.00017410631895687062, + "loss": 1.2146, + "theoretical_loss": 3.387556034876965, + "tokens_seen": 2589786112 + }, + { + "epoch": 0.78, + "learning_rate": 0.00017409027081243732, + "loss": 1.276, + "theoretical_loss": 3.3875493023540173, + "tokens_seen": 2589851648 + }, + { + "epoch": 0.78, + "learning_rate": 0.00017407422266800402, + "loss": 1.3401, + "theoretical_loss": 3.3875425700491353, + "tokens_seen": 2589917184 + }, + { + "epoch": 0.78, + "learning_rate": 0.00017405817452357072, + "loss": 1.3546, + "theoretical_loss": 3.3875358379623055, + "tokens_seen": 2589982720 + }, + { + "epoch": 0.78, + "learning_rate": 0.00017404212637913742, + "loss": 1.2809, + "theoretical_loss": 3.387529106093515, + "tokens_seen": 2590048256 + }, + { + "epoch": 0.78, + "learning_rate": 0.00017402607823470412, + "loss": 1.1958, + "theoretical_loss": 3.387522374442752, + "tokens_seen": 2590113792 + }, + { + "epoch": 0.78, + "learning_rate": 0.00017401003009027082, + "loss": 1.2024, + "theoretical_loss": 3.387515643010004, + "tokens_seen": 2590179328 + }, + { + "epoch": 0.78, + "learning_rate": 0.00017399398194583752, + "loss": 1.1347, + "theoretical_loss": 3.3875089117952575, + "tokens_seen": 2590244864 + }, + { + "epoch": 0.78, + "learning_rate": 0.00017397793380140422, + "loss": 1.1809, + "theoretical_loss": 3.3875021807985006, + "tokens_seen": 2590310400 + }, + { + "epoch": 0.78, + "learning_rate": 0.00017396188565697092, + "loss": 1.2307, + "theoretical_loss": 3.3874954500197205, + "tokens_seen": 2590375936 + }, + { + "epoch": 0.78, + "learning_rate": 0.00017394583751253762, + "loss": 1.3295, + "theoretical_loss": 3.387488719458905, + "tokens_seen": 2590441472 + }, + { + "epoch": 0.79, + "learning_rate": 0.00017392978936810432, + "loss": 1.1401, + "theoretical_loss": 3.387481989116041, + "tokens_seen": 2590507008 + }, + { + "epoch": 0.79, + "learning_rate": 0.00017391374122367102, + "loss": 1.3182, + "theoretical_loss": 3.387475258991116, + "tokens_seen": 2590572544 + }, + { + "epoch": 0.79, + "learning_rate": 0.00017389769307923772, + "loss": 1.2472, + "theoretical_loss": 3.387468529084118, + "tokens_seen": 2590638080 + }, + { + "epoch": 0.79, + "learning_rate": 0.0001738816449348044, + "loss": 1.1979, + "theoretical_loss": 3.387461799395034, + "tokens_seen": 2590703616 + }, + { + "epoch": 0.79, + "learning_rate": 0.00017386559679037113, + "loss": 1.2779, + "theoretical_loss": 3.3874550699238513, + "tokens_seen": 2590769152 + }, + { + "epoch": 0.79, + "learning_rate": 0.00017384954864593783, + "loss": 1.3523, + "theoretical_loss": 3.3874483406705576, + "tokens_seen": 2590834688 + }, + { + "epoch": 0.79, + "learning_rate": 0.00017383350050150453, + "loss": 1.354, + "theoretical_loss": 3.38744161163514, + "tokens_seen": 2590900224 + }, + { + "epoch": 0.79, + "learning_rate": 0.00017381745235707123, + "loss": 1.2611, + "theoretical_loss": 3.387434882817586, + "tokens_seen": 2590965760 + }, + { + "epoch": 0.79, + "learning_rate": 0.00017380140421263793, + "loss": 1.2918, + "theoretical_loss": 3.3874281542178837, + "tokens_seen": 2591031296 + }, + { + "epoch": 0.79, + "learning_rate": 0.00017378535606820463, + "loss": 1.1931, + "theoretical_loss": 3.3874214258360196, + "tokens_seen": 2591096832 + }, + { + "epoch": 0.79, + "learning_rate": 0.00017376930792377133, + "loss": 1.1398, + "theoretical_loss": 3.387414697671982, + "tokens_seen": 2591162368 + }, + { + "epoch": 0.79, + "learning_rate": 0.00017375325977933803, + "loss": 1.3485, + "theoretical_loss": 3.3874079697257575, + "tokens_seen": 2591227904 + }, + { + "epoch": 0.79, + "learning_rate": 0.00017373721163490473, + "loss": 1.3063, + "theoretical_loss": 3.3874012419973343, + "tokens_seen": 2591293440 + }, + { + "epoch": 0.79, + "learning_rate": 0.00017372116349047143, + "loss": 1.1765, + "theoretical_loss": 3.387394514486699, + "tokens_seen": 2591358976 + }, + { + "epoch": 0.79, + "learning_rate": 0.00017370511534603813, + "loss": 1.4098, + "theoretical_loss": 3.3873877871938403, + "tokens_seen": 2591424512 + }, + { + "epoch": 0.79, + "learning_rate": 0.00017368906720160483, + "loss": 1.2485, + "theoretical_loss": 3.387381060118744, + "tokens_seen": 2591490048 + }, + { + "epoch": 0.79, + "learning_rate": 0.00017367301905717153, + "loss": 1.3784, + "theoretical_loss": 3.3873743332613993, + "tokens_seen": 2591555584 + }, + { + "epoch": 0.79, + "learning_rate": 0.00017365697091273823, + "loss": 1.2193, + "theoretical_loss": 3.387367606621792, + "tokens_seen": 2591621120 + }, + { + "epoch": 0.79, + "learning_rate": 0.00017364092276830493, + "loss": 1.193, + "theoretical_loss": 3.387360880199911, + "tokens_seen": 2591686656 + }, + { + "epoch": 0.79, + "learning_rate": 0.00017362487462387164, + "loss": 1.2862, + "theoretical_loss": 3.3873541539957426, + "tokens_seen": 2591752192 + }, + { + "epoch": 0.79, + "learning_rate": 0.00017360882647943834, + "loss": 1.3119, + "theoretical_loss": 3.387347428009275, + "tokens_seen": 2591817728 + }, + { + "epoch": 0.79, + "learning_rate": 0.00017359277833500504, + "loss": 1.3049, + "theoretical_loss": 3.387340702240495, + "tokens_seen": 2591883264 + }, + { + "epoch": 0.79, + "objective/train/docs_used": 942495, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.2394979000091553, + "objective/train/theoretical_loss": 3.3873339766893906, + "objective/train/tokens_used": 2612408800, + "theoretical_loss": 3.3873339766893906, + "tokens_seen": 2591948800 + }, + { + "epoch": 0.79, + "learning_rate": 0.00017357673019057174, + "loss": 1.2356, + "theoretical_loss": 3.3873339766893906, + "tokens_seen": 2591948800 + }, + { + "epoch": 0.79, + "learning_rate": 0.0001735606820461384, + "loss": 1.2617, + "theoretical_loss": 3.387327251355949, + "tokens_seen": 2592014336 + }, + { + "epoch": 0.79, + "learning_rate": 0.0001735446339017051, + "loss": 1.3657, + "theoretical_loss": 3.387320526240158, + "tokens_seen": 2592079872 + }, + { + "epoch": 0.79, + "learning_rate": 0.0001735285857572718, + "loss": 1.2744, + "theoretical_loss": 3.387313801342004, + "tokens_seen": 2592145408 + }, + { + "epoch": 0.79, + "learning_rate": 0.0001735125376128385, + "loss": 1.3447, + "theoretical_loss": 3.387307076661476, + "tokens_seen": 2592210944 + }, + { + "epoch": 0.79, + "learning_rate": 0.0001734964894684052, + "loss": 1.2155, + "theoretical_loss": 3.3873003521985603, + "tokens_seen": 2592276480 + }, + { + "epoch": 0.79, + "learning_rate": 0.00017348044132397191, + "loss": 1.2949, + "theoretical_loss": 3.387293627953245, + "tokens_seen": 2592342016 + }, + { + "epoch": 0.79, + "learning_rate": 0.00017346439317953861, + "loss": 1.3903, + "theoretical_loss": 3.387286903925517, + "tokens_seen": 2592407552 + }, + { + "epoch": 0.79, + "learning_rate": 0.00017344834503510532, + "loss": 1.3329, + "theoretical_loss": 3.3872801801153645, + "tokens_seen": 2592473088 + }, + { + "epoch": 0.79, + "learning_rate": 0.00017343229689067202, + "loss": 1.2932, + "theoretical_loss": 3.387273456522774, + "tokens_seen": 2592538624 + }, + { + "epoch": 0.79, + "learning_rate": 0.00017341624874623872, + "loss": 1.2981, + "theoretical_loss": 3.3872667331477335, + "tokens_seen": 2592604160 + }, + { + "epoch": 0.79, + "learning_rate": 0.00017340020060180542, + "loss": 1.338, + "theoretical_loss": 3.387260009990231, + "tokens_seen": 2592669696 + }, + { + "epoch": 0.79, + "learning_rate": 0.00017338415245737212, + "loss": 1.2664, + "theoretical_loss": 3.3872532870502527, + "tokens_seen": 2592735232 + }, + { + "epoch": 0.79, + "learning_rate": 0.00017336810431293882, + "loss": 1.3186, + "theoretical_loss": 3.387246564327787, + "tokens_seen": 2592800768 + }, + { + "epoch": 0.79, + "learning_rate": 0.00017335205616850552, + "loss": 1.2901, + "theoretical_loss": 3.3872398418228213, + "tokens_seen": 2592866304 + }, + { + "epoch": 0.79, + "learning_rate": 0.00017333600802407222, + "loss": 1.1972, + "theoretical_loss": 3.3872331195353427, + "tokens_seen": 2592931840 + }, + { + "epoch": 0.79, + "learning_rate": 0.00017331995987963892, + "loss": 1.2447, + "theoretical_loss": 3.387226397465339, + "tokens_seen": 2592997376 + }, + { + "epoch": 0.79, + "learning_rate": 0.00017330391173520562, + "loss": 1.2243, + "theoretical_loss": 3.3872196756127972, + "tokens_seen": 2593062912 + }, + { + "epoch": 0.79, + "learning_rate": 0.00017328786359077232, + "loss": 1.1928, + "theoretical_loss": 3.3872129539777056, + "tokens_seen": 2593128448 + }, + { + "epoch": 0.79, + "learning_rate": 0.00017327181544633905, + "loss": 1.2129, + "theoretical_loss": 3.387206232560051, + "tokens_seen": 2593193984 + }, + { + "epoch": 0.79, + "learning_rate": 0.00017325576730190572, + "loss": 1.3008, + "theoretical_loss": 3.387199511359821, + "tokens_seen": 2593259520 + }, + { + "epoch": 0.79, + "learning_rate": 0.00017323971915747242, + "loss": 1.1558, + "theoretical_loss": 3.3871927903770027, + "tokens_seen": 2593325056 + }, + { + "epoch": 0.79, + "learning_rate": 0.00017322367101303912, + "loss": 1.3046, + "theoretical_loss": 3.3871860696115847, + "tokens_seen": 2593390592 + }, + { + "epoch": 0.79, + "learning_rate": 0.00017320762286860582, + "loss": 1.2805, + "theoretical_loss": 3.3871793490635533, + "tokens_seen": 2593456128 + }, + { + "epoch": 0.79, + "learning_rate": 0.00017319157472417253, + "loss": 1.2455, + "theoretical_loss": 3.3871726287328965, + "tokens_seen": 2593521664 + }, + { + "epoch": 0.79, + "learning_rate": 0.00017317552657973923, + "loss": 1.3751, + "theoretical_loss": 3.3871659086196018, + "tokens_seen": 2593587200 + }, + { + "epoch": 0.79, + "learning_rate": 0.00017315947843530593, + "loss": 1.2448, + "theoretical_loss": 3.3871591887236567, + "tokens_seen": 2593652736 + }, + { + "epoch": 0.79, + "learning_rate": 0.00017314343029087263, + "loss": 1.2525, + "theoretical_loss": 3.387152469045048, + "tokens_seen": 2593718272 + }, + { + "epoch": 0.79, + "learning_rate": 0.00017312738214643933, + "loss": 1.2657, + "theoretical_loss": 3.3871457495837642, + "tokens_seen": 2593783808 + }, + { + "epoch": 0.79, + "learning_rate": 0.00017311133400200603, + "loss": 1.2667, + "theoretical_loss": 3.3871390303397924, + "tokens_seen": 2593849344 + }, + { + "epoch": 0.79, + "learning_rate": 0.00017309528585757273, + "loss": 1.3496, + "theoretical_loss": 3.38713231131312, + "tokens_seen": 2593914880 + }, + { + "epoch": 0.79, + "learning_rate": 0.00017307923771313943, + "loss": 1.1493, + "theoretical_loss": 3.3871255925037342, + "tokens_seen": 2593980416 + }, + { + "epoch": 0.79, + "learning_rate": 0.00017306318956870613, + "loss": 1.2932, + "theoretical_loss": 3.387118873911623, + "tokens_seen": 2594045952 + }, + { + "epoch": 0.79, + "learning_rate": 0.00017304714142427283, + "loss": 1.4031, + "theoretical_loss": 3.3871121555367734, + "tokens_seen": 2594111488 + }, + { + "epoch": 0.79, + "learning_rate": 0.00017303109327983953, + "loss": 1.3315, + "theoretical_loss": 3.3871054373791734, + "tokens_seen": 2594177024 + }, + { + "epoch": 0.79, + "learning_rate": 0.00017301504513540623, + "loss": 1.3192, + "theoretical_loss": 3.3870987194388102, + "tokens_seen": 2594242560 + }, + { + "epoch": 0.79, + "learning_rate": 0.00017299899699097293, + "loss": 1.3283, + "theoretical_loss": 3.387092001715671, + "tokens_seen": 2594308096 + }, + { + "epoch": 0.79, + "learning_rate": 0.00017298294884653963, + "loss": 1.2183, + "theoretical_loss": 3.387085284209744, + "tokens_seen": 2594373632 + }, + { + "epoch": 0.79, + "learning_rate": 0.00017296690070210633, + "loss": 1.3299, + "theoretical_loss": 3.3870785669210166, + "tokens_seen": 2594439168 + }, + { + "epoch": 0.79, + "learning_rate": 0.00017295085255767303, + "loss": 1.2127, + "theoretical_loss": 3.3870718498494754, + "tokens_seen": 2594504704 + }, + { + "epoch": 0.79, + "learning_rate": 0.00017293480441323974, + "loss": 1.3219, + "theoretical_loss": 3.387065132995109, + "tokens_seen": 2594570240 + }, + { + "epoch": 0.79, + "learning_rate": 0.00017291875626880644, + "loss": 1.3322, + "theoretical_loss": 3.3870584163579034, + "tokens_seen": 2594635776 + }, + { + "epoch": 0.79, + "learning_rate": 0.00017290270812437314, + "loss": 1.351, + "theoretical_loss": 3.3870516999378477, + "tokens_seen": 2594701312 + }, + { + "epoch": 0.79, + "learning_rate": 0.00017288665997993984, + "loss": 1.2931, + "theoretical_loss": 3.387044983734929, + "tokens_seen": 2594766848 + }, + { + "epoch": 0.79, + "learning_rate": 0.0001728706118355065, + "loss": 1.3104, + "theoretical_loss": 3.3870382677491344, + "tokens_seen": 2594832384 + }, + { + "epoch": 0.79, + "learning_rate": 0.0001728545636910732, + "loss": 1.2658, + "theoretical_loss": 3.3870315519804515, + "tokens_seen": 2594897920 + }, + { + "epoch": 0.79, + "learning_rate": 0.0001728385155466399, + "loss": 1.314, + "theoretical_loss": 3.3870248364288678, + "tokens_seen": 2594963456 + }, + { + "epoch": 0.79, + "learning_rate": 0.0001728224674022066, + "loss": 1.3385, + "theoretical_loss": 3.387018121094371, + "tokens_seen": 2595028992 + }, + { + "epoch": 0.79, + "learning_rate": 0.0001728064192577733, + "loss": 1.3187, + "theoretical_loss": 3.3870114059769483, + "tokens_seen": 2595094528 + }, + { + "epoch": 0.79, + "learning_rate": 0.00017279037111334001, + "loss": 1.3042, + "theoretical_loss": 3.387004691076587, + "tokens_seen": 2595160064 + }, + { + "epoch": 0.79, + "objective/train/docs_used": 943650, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.309882640838623, + "objective/train/theoretical_loss": 3.3869979763932756, + "objective/train/tokens_used": 2615685600, + "theoretical_loss": 3.3869979763932756, + "tokens_seen": 2595225600 + }, + { + "epoch": 0.79, + "learning_rate": 0.00017277432296890671, + "loss": 1.4219, + "theoretical_loss": 3.3869979763932756, + "tokens_seen": 2595225600 + }, + { + "epoch": 0.79, + "learning_rate": 0.00017275827482447342, + "loss": 1.3572, + "theoretical_loss": 3.3869912619270006, + "tokens_seen": 2595291136 + }, + { + "epoch": 0.79, + "learning_rate": 0.00017274222668004012, + "loss": 1.286, + "theoretical_loss": 3.3869845476777503, + "tokens_seen": 2595356672 + }, + { + "epoch": 0.79, + "learning_rate": 0.00017272617853560682, + "loss": 1.4289, + "theoretical_loss": 3.386977833645511, + "tokens_seen": 2595422208 + }, + { + "epoch": 0.79, + "learning_rate": 0.00017271013039117352, + "loss": 1.405, + "theoretical_loss": 3.386971119830272, + "tokens_seen": 2595487744 + }, + { + "epoch": 0.79, + "learning_rate": 0.00017269408224674022, + "loss": 1.3791, + "theoretical_loss": 3.386964406232019, + "tokens_seen": 2595553280 + }, + { + "epoch": 0.79, + "learning_rate": 0.00017267803410230692, + "loss": 1.2932, + "theoretical_loss": 3.38695769285074, + "tokens_seen": 2595618816 + }, + { + "epoch": 0.79, + "learning_rate": 0.00017266198595787365, + "loss": 1.2641, + "theoretical_loss": 3.3869509796864237, + "tokens_seen": 2595684352 + }, + { + "epoch": 0.79, + "learning_rate": 0.00017264593781344035, + "loss": 1.3393, + "theoretical_loss": 3.3869442667390564, + "tokens_seen": 2595749888 + }, + { + "epoch": 0.79, + "learning_rate": 0.00017262988966900705, + "loss": 1.3478, + "theoretical_loss": 3.3869375540086257, + "tokens_seen": 2595815424 + }, + { + "epoch": 0.79, + "learning_rate": 0.00017261384152457375, + "loss": 1.3182, + "theoretical_loss": 3.3869308414951194, + "tokens_seen": 2595880960 + }, + { + "epoch": 0.79, + "learning_rate": 0.00017259779338014045, + "loss": 1.2127, + "theoretical_loss": 3.3869241291985253, + "tokens_seen": 2595946496 + }, + { + "epoch": 0.79, + "learning_rate": 0.00017258174523570715, + "loss": 1.4147, + "theoretical_loss": 3.38691741711883, + "tokens_seen": 2596012032 + }, + { + "epoch": 0.79, + "learning_rate": 0.00017256569709127385, + "loss": 1.2653, + "theoretical_loss": 3.386910705256022, + "tokens_seen": 2596077568 + }, + { + "epoch": 0.79, + "learning_rate": 0.00017254964894684052, + "loss": 1.2702, + "theoretical_loss": 3.386903993610088, + "tokens_seen": 2596143104 + }, + { + "epoch": 0.79, + "learning_rate": 0.00017253360080240722, + "loss": 1.3175, + "theoretical_loss": 3.3868972821810166, + "tokens_seen": 2596208640 + }, + { + "epoch": 0.79, + "learning_rate": 0.00017251755265797392, + "loss": 1.4883, + "theoretical_loss": 3.386890570968794, + "tokens_seen": 2596274176 + }, + { + "epoch": 0.79, + "learning_rate": 0.00017250150451354063, + "loss": 1.326, + "theoretical_loss": 3.3868838599734086, + "tokens_seen": 2596339712 + }, + { + "epoch": 0.79, + "learning_rate": 0.00017248545636910733, + "loss": 1.2891, + "theoretical_loss": 3.3868771491948477, + "tokens_seen": 2596405248 + }, + { + "epoch": 0.79, + "learning_rate": 0.00017246940822467403, + "loss": 1.2568, + "theoretical_loss": 3.386870438633099, + "tokens_seen": 2596470784 + }, + { + "epoch": 0.79, + "learning_rate": 0.00017245336008024073, + "loss": 1.1802, + "theoretical_loss": 3.3868637282881493, + "tokens_seen": 2596536320 + }, + { + "epoch": 0.79, + "learning_rate": 0.00017243731193580743, + "loss": 1.2773, + "theoretical_loss": 3.3868570181599873, + "tokens_seen": 2596601856 + }, + { + "epoch": 0.79, + "learning_rate": 0.00017242126379137413, + "loss": 1.2707, + "theoretical_loss": 3.3868503082485995, + "tokens_seen": 2596667392 + }, + { + "epoch": 0.79, + "learning_rate": 0.00017240521564694083, + "loss": 1.2599, + "theoretical_loss": 3.3868435985539738, + "tokens_seen": 2596732928 + }, + { + "epoch": 0.79, + "learning_rate": 0.00017238916750250753, + "loss": 1.2439, + "theoretical_loss": 3.386836889076098, + "tokens_seen": 2596798464 + }, + { + "epoch": 0.79, + "learning_rate": 0.00017237311935807423, + "loss": 1.203, + "theoretical_loss": 3.3868301798149587, + "tokens_seen": 2596864000 + }, + { + "epoch": 0.79, + "learning_rate": 0.00017235707121364093, + "loss": 1.2408, + "theoretical_loss": 3.3868234707705445, + "tokens_seen": 2596929536 + }, + { + "epoch": 0.79, + "learning_rate": 0.00017234102306920763, + "loss": 1.3189, + "theoretical_loss": 3.3868167619428426, + "tokens_seen": 2596995072 + }, + { + "epoch": 0.79, + "learning_rate": 0.00017232497492477433, + "loss": 1.3395, + "theoretical_loss": 3.38681005333184, + "tokens_seen": 2597060608 + }, + { + "epoch": 0.79, + "learning_rate": 0.00017230892678034103, + "loss": 1.156, + "theoretical_loss": 3.386803344937525, + "tokens_seen": 2597126144 + }, + { + "epoch": 0.79, + "learning_rate": 0.00017229287863590773, + "loss": 1.3825, + "theoretical_loss": 3.386796636759885, + "tokens_seen": 2597191680 + }, + { + "epoch": 0.79, + "learning_rate": 0.00017227683049147443, + "loss": 1.3337, + "theoretical_loss": 3.3867899287989074, + "tokens_seen": 2597257216 + }, + { + "epoch": 0.79, + "learning_rate": 0.00017226078234704113, + "loss": 1.2693, + "theoretical_loss": 3.386783221054579, + "tokens_seen": 2597322752 + }, + { + "epoch": 0.79, + "learning_rate": 0.00017224473420260784, + "loss": 1.3113, + "theoretical_loss": 3.386776513526889, + "tokens_seen": 2597388288 + }, + { + "epoch": 0.79, + "learning_rate": 0.00017222868605817454, + "loss": 1.2974, + "theoretical_loss": 3.386769806215823, + "tokens_seen": 2597453824 + }, + { + "epoch": 0.79, + "learning_rate": 0.00017221263791374124, + "loss": 1.3067, + "theoretical_loss": 3.38676309912137, + "tokens_seen": 2597519360 + }, + { + "epoch": 0.79, + "learning_rate": 0.00017219658976930794, + "loss": 1.2196, + "theoretical_loss": 3.386756392243517, + "tokens_seen": 2597584896 + }, + { + "epoch": 0.79, + "learning_rate": 0.00017218054162487464, + "loss": 1.2784, + "theoretical_loss": 3.3867496855822514, + "tokens_seen": 2597650432 + }, + { + "epoch": 0.79, + "learning_rate": 0.0001721644934804413, + "loss": 1.1797, + "theoretical_loss": 3.386742979137561, + "tokens_seen": 2597715968 + }, + { + "epoch": 0.79, + "learning_rate": 0.000172148445336008, + "loss": 1.2736, + "theoretical_loss": 3.3867362729094332, + "tokens_seen": 2597781504 + }, + { + "epoch": 0.79, + "learning_rate": 0.0001721323971915747, + "loss": 1.3173, + "theoretical_loss": 3.3867295668978556, + "tokens_seen": 2597847040 + }, + { + "epoch": 0.79, + "learning_rate": 0.0001721163490471414, + "loss": 1.299, + "theoretical_loss": 3.386722861102816, + "tokens_seen": 2597912576 + }, + { + "epoch": 0.79, + "learning_rate": 0.00017210030090270811, + "loss": 1.1846, + "theoretical_loss": 3.3867161555243017, + "tokens_seen": 2597978112 + }, + { + "epoch": 0.79, + "learning_rate": 0.00017208425275827481, + "loss": 1.3464, + "theoretical_loss": 3.3867094501623, + "tokens_seen": 2598043648 + }, + { + "epoch": 0.79, + "learning_rate": 0.00017206820461384154, + "loss": 1.3639, + "theoretical_loss": 3.3867027450167986, + "tokens_seen": 2598109184 + }, + { + "epoch": 0.79, + "learning_rate": 0.00017205215646940824, + "loss": 1.4186, + "theoretical_loss": 3.3866960400877852, + "tokens_seen": 2598174720 + }, + { + "epoch": 0.79, + "learning_rate": 0.00017203610832497494, + "loss": 1.3133, + "theoretical_loss": 3.3866893353752476, + "tokens_seen": 2598240256 + }, + { + "epoch": 0.79, + "learning_rate": 0.00017202006018054164, + "loss": 1.3051, + "theoretical_loss": 3.3866826308791724, + "tokens_seen": 2598305792 + }, + { + "epoch": 0.79, + "learning_rate": 0.00017200401203610835, + "loss": 1.3466, + "theoretical_loss": 3.3866759265995485, + "tokens_seen": 2598371328 + }, + { + "epoch": 0.79, + "learning_rate": 0.00017198796389167505, + "loss": 1.363, + "theoretical_loss": 3.386669222536362, + "tokens_seen": 2598436864 + }, + { + "epoch": 0.79, + "objective/train/docs_used": 944930, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.213080644607544, + "objective/train/theoretical_loss": 3.3866625186896018, + "objective/train/tokens_used": 2618962400, + "theoretical_loss": 3.3866625186896018, + "tokens_seen": 2598502400 + }, + { + "epoch": 0.79, + "learning_rate": 0.00017197191574724175, + "loss": 1.2099, + "theoretical_loss": 3.3866625186896018, + "tokens_seen": 2598502400 + }, + { + "epoch": 0.79, + "learning_rate": 0.00017195586760280845, + "loss": 1.2772, + "theoretical_loss": 3.386655815059255, + "tokens_seen": 2598567936 + }, + { + "epoch": 0.79, + "learning_rate": 0.00017193981945837515, + "loss": 1.2256, + "theoretical_loss": 3.3866491116453084, + "tokens_seen": 2598633472 + }, + { + "epoch": 0.79, + "learning_rate": 0.00017192377131394185, + "loss": 1.3216, + "theoretical_loss": 3.3866424084477504, + "tokens_seen": 2598699008 + }, + { + "epoch": 0.79, + "learning_rate": 0.00017190772316950855, + "loss": 1.4172, + "theoretical_loss": 3.386635705466569, + "tokens_seen": 2598764544 + }, + { + "epoch": 0.79, + "learning_rate": 0.00017189167502507525, + "loss": 1.2989, + "theoretical_loss": 3.38662900270175, + "tokens_seen": 2598830080 + }, + { + "epoch": 0.79, + "learning_rate": 0.00017187562688064195, + "loss": 1.3409, + "theoretical_loss": 3.3866223001532827, + "tokens_seen": 2598895616 + }, + { + "epoch": 0.79, + "learning_rate": 0.00017185957873620865, + "loss": 1.3145, + "theoretical_loss": 3.3866155978211534, + "tokens_seen": 2598961152 + }, + { + "epoch": 0.79, + "learning_rate": 0.00017184353059177532, + "loss": 1.4401, + "theoretical_loss": 3.3866088957053506, + "tokens_seen": 2599026688 + }, + { + "epoch": 0.79, + "learning_rate": 0.00017182748244734202, + "loss": 1.2536, + "theoretical_loss": 3.386602193805862, + "tokens_seen": 2599092224 + }, + { + "epoch": 0.79, + "learning_rate": 0.00017181143430290873, + "loss": 1.3091, + "theoretical_loss": 3.386595492122674, + "tokens_seen": 2599157760 + }, + { + "epoch": 0.79, + "learning_rate": 0.00017179538615847543, + "loss": 1.2744, + "theoretical_loss": 3.3865887906557752, + "tokens_seen": 2599223296 + }, + { + "epoch": 0.79, + "learning_rate": 0.00017177933801404213, + "loss": 1.3316, + "theoretical_loss": 3.386582089405153, + "tokens_seen": 2599288832 + }, + { + "epoch": 0.79, + "learning_rate": 0.00017176328986960883, + "loss": 1.4454, + "theoretical_loss": 3.3865753883707943, + "tokens_seen": 2599354368 + }, + { + "epoch": 0.79, + "learning_rate": 0.00017174724172517553, + "loss": 1.264, + "theoretical_loss": 3.3865686875526873, + "tokens_seen": 2599419904 + }, + { + "epoch": 0.79, + "learning_rate": 0.00017173119358074223, + "loss": 1.2304, + "theoretical_loss": 3.3865619869508197, + "tokens_seen": 2599485440 + }, + { + "epoch": 0.79, + "learning_rate": 0.00017171514543630893, + "loss": 1.1993, + "theoretical_loss": 3.3865552865651782, + "tokens_seen": 2599550976 + }, + { + "epoch": 0.79, + "learning_rate": 0.00017169909729187563, + "loss": 1.2815, + "theoretical_loss": 3.386548586395752, + "tokens_seen": 2599616512 + }, + { + "epoch": 0.79, + "learning_rate": 0.00017168304914744233, + "loss": 1.1929, + "theoretical_loss": 3.3865418864425267, + "tokens_seen": 2599682048 + }, + { + "epoch": 0.79, + "learning_rate": 0.00017166700100300903, + "loss": 1.2672, + "theoretical_loss": 3.3865351867054914, + "tokens_seen": 2599747584 + }, + { + "epoch": 0.79, + "learning_rate": 0.00017165095285857573, + "loss": 1.2974, + "theoretical_loss": 3.3865284871846324, + "tokens_seen": 2599813120 + }, + { + "epoch": 0.79, + "learning_rate": 0.00017163490471414243, + "loss": 1.346, + "theoretical_loss": 3.3865217878799387, + "tokens_seen": 2599878656 + }, + { + "epoch": 0.79, + "learning_rate": 0.00017161885656970913, + "loss": 1.2935, + "theoretical_loss": 3.3865150887913966, + "tokens_seen": 2599944192 + }, + { + "epoch": 0.79, + "learning_rate": 0.00017160280842527583, + "loss": 1.2663, + "theoretical_loss": 3.386508389918995, + "tokens_seen": 2600009728 + }, + { + "epoch": 0.79, + "learning_rate": 0.00017158676028084253, + "loss": 1.3888, + "theoretical_loss": 3.38650169126272, + "tokens_seen": 2600075264 + }, + { + "epoch": 0.79, + "learning_rate": 0.00017157071213640924, + "loss": 1.3601, + "theoretical_loss": 3.38649499282256, + "tokens_seen": 2600140800 + }, + { + "epoch": 0.79, + "learning_rate": 0.00017155466399197594, + "loss": 1.2843, + "theoretical_loss": 3.3864882945985024, + "tokens_seen": 2600206336 + }, + { + "epoch": 0.79, + "learning_rate": 0.00017153861584754264, + "loss": 1.2749, + "theoretical_loss": 3.386481596590535, + "tokens_seen": 2600271872 + }, + { + "epoch": 0.79, + "learning_rate": 0.00017152256770310934, + "loss": 1.1818, + "theoretical_loss": 3.386474898798645, + "tokens_seen": 2600337408 + }, + { + "epoch": 0.79, + "learning_rate": 0.00017150651955867604, + "loss": 1.2617, + "theoretical_loss": 3.3864682012228204, + "tokens_seen": 2600402944 + }, + { + "epoch": 0.79, + "learning_rate": 0.00017149047141424274, + "loss": 1.2054, + "theoretical_loss": 3.386461503863049, + "tokens_seen": 2600468480 + }, + { + "epoch": 0.79, + "learning_rate": 0.00017147442326980944, + "loss": 1.3484, + "theoretical_loss": 3.386454806719317, + "tokens_seen": 2600534016 + }, + { + "epoch": 0.79, + "learning_rate": 0.00017145837512537614, + "loss": 1.268, + "theoretical_loss": 3.3864481097916137, + "tokens_seen": 2600599552 + }, + { + "epoch": 0.79, + "learning_rate": 0.00017144232698094284, + "loss": 1.4223, + "theoretical_loss": 3.386441413079926, + "tokens_seen": 2600665088 + }, + { + "epoch": 0.79, + "learning_rate": 0.00017142627883650954, + "loss": 1.2708, + "theoretical_loss": 3.3864347165842412, + "tokens_seen": 2600730624 + }, + { + "epoch": 0.79, + "learning_rate": 0.00017141023069207624, + "loss": 1.3364, + "theoretical_loss": 3.386428020304547, + "tokens_seen": 2600796160 + }, + { + "epoch": 0.79, + "learning_rate": 0.00017139418254764294, + "loss": 1.3323, + "theoretical_loss": 3.386421324240831, + "tokens_seen": 2600861696 + }, + { + "epoch": 0.79, + "learning_rate": 0.00017137813440320964, + "loss": 1.4174, + "theoretical_loss": 3.3864146283930814, + "tokens_seen": 2600927232 + }, + { + "epoch": 0.79, + "learning_rate": 0.00017136208625877634, + "loss": 1.3786, + "theoretical_loss": 3.3864079327612853, + "tokens_seen": 2600992768 + }, + { + "epoch": 0.79, + "learning_rate": 0.00017134603811434304, + "loss": 1.3702, + "theoretical_loss": 3.38640123734543, + "tokens_seen": 2601058304 + }, + { + "epoch": 0.79, + "learning_rate": 0.00017132998996990974, + "loss": 1.4408, + "theoretical_loss": 3.3863945421455037, + "tokens_seen": 2601123840 + }, + { + "epoch": 0.79, + "learning_rate": 0.00017131394182547645, + "loss": 1.3365, + "theoretical_loss": 3.3863878471614934, + "tokens_seen": 2601189376 + }, + { + "epoch": 0.79, + "learning_rate": 0.00017129789368104315, + "loss": 1.3121, + "theoretical_loss": 3.386381152393387, + "tokens_seen": 2601254912 + }, + { + "epoch": 0.79, + "learning_rate": 0.00017128184553660985, + "loss": 1.2484, + "theoretical_loss": 3.386374457841172, + "tokens_seen": 2601320448 + }, + { + "epoch": 0.79, + "learning_rate": 0.00017126579739217655, + "loss": 1.282, + "theoretical_loss": 3.3863677635048366, + "tokens_seen": 2601385984 + }, + { + "epoch": 0.79, + "learning_rate": 0.00017124974924774325, + "loss": 1.2558, + "theoretical_loss": 3.3863610693843675, + "tokens_seen": 2601451520 + }, + { + "epoch": 0.79, + "learning_rate": 0.00017123370110330995, + "loss": 1.3171, + "theoretical_loss": 3.3863543754797525, + "tokens_seen": 2601517056 + }, + { + "epoch": 0.79, + "learning_rate": 0.00017121765295887665, + "loss": 1.4135, + "theoretical_loss": 3.38634768179098, + "tokens_seen": 2601582592 + }, + { + "epoch": 0.79, + "learning_rate": 0.00017120160481444335, + "loss": 1.2543, + "theoretical_loss": 3.3863409883180364, + "tokens_seen": 2601648128 + }, + { + "epoch": 0.79, + "learning_rate": 0.00017118555667001005, + "loss": 1.1778, + "theoretical_loss": 3.38633429506091, + "tokens_seen": 2601713664 + }, + { + "epoch": 0.79, + "objective/train/docs_used": 946148, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.4879472255706787, + "objective/train/theoretical_loss": 3.3863276020195885, + "objective/train/tokens_used": 2622239200, + "theoretical_loss": 3.3863276020195885, + "tokens_seen": 2601779200 + }, + { + "epoch": 0.79, + "learning_rate": 0.00017116950852557675, + "loss": 1.4025, + "theoretical_loss": 3.3863276020195885, + "tokens_seen": 2601779200 + }, + { + "epoch": 0.79, + "learning_rate": 0.00017115346038114342, + "loss": 1.2229, + "theoretical_loss": 3.386320909194059, + "tokens_seen": 2601844736 + }, + { + "epoch": 0.79, + "learning_rate": 0.00017113741223671013, + "loss": 1.262, + "theoretical_loss": 3.38631421658431, + "tokens_seen": 2601910272 + }, + { + "epoch": 0.79, + "learning_rate": 0.00017112136409227683, + "loss": 1.4643, + "theoretical_loss": 3.386307524190328, + "tokens_seen": 2601975808 + }, + { + "epoch": 0.79, + "learning_rate": 0.00017110531594784353, + "loss": 1.3731, + "theoretical_loss": 3.3863008320121013, + "tokens_seen": 2602041344 + }, + { + "epoch": 0.79, + "learning_rate": 0.00017108926780341023, + "loss": 1.2427, + "theoretical_loss": 3.3862941400496176, + "tokens_seen": 2602106880 + }, + { + "epoch": 0.79, + "learning_rate": 0.00017107321965897693, + "loss": 1.2077, + "theoretical_loss": 3.386287448302864, + "tokens_seen": 2602172416 + }, + { + "epoch": 0.79, + "learning_rate": 0.00017105717151454363, + "loss": 1.3849, + "theoretical_loss": 3.3862807567718285, + "tokens_seen": 2602237952 + }, + { + "epoch": 0.79, + "learning_rate": 0.00017104112337011033, + "loss": 1.3093, + "theoretical_loss": 3.3862740654564987, + "tokens_seen": 2602303488 + }, + { + "epoch": 0.79, + "learning_rate": 0.00017102507522567703, + "loss": 1.4041, + "theoretical_loss": 3.3862673743568616, + "tokens_seen": 2602369024 + }, + { + "epoch": 0.79, + "learning_rate": 0.00017100902708124373, + "loss": 1.3326, + "theoretical_loss": 3.386260683472906, + "tokens_seen": 2602434560 + }, + { + "epoch": 0.79, + "learning_rate": 0.00017099297893681043, + "loss": 1.428, + "theoretical_loss": 3.386253992804618, + "tokens_seen": 2602500096 + }, + { + "epoch": 0.79, + "learning_rate": 0.00017097693079237713, + "loss": 1.3738, + "theoretical_loss": 3.3862473023519866, + "tokens_seen": 2602565632 + }, + { + "epoch": 0.79, + "learning_rate": 0.00017096088264794383, + "loss": 1.223, + "theoretical_loss": 3.3862406121149986, + "tokens_seen": 2602631168 + }, + { + "epoch": 0.79, + "learning_rate": 0.00017094483450351053, + "loss": 1.2492, + "theoretical_loss": 3.386233922093642, + "tokens_seen": 2602696704 + }, + { + "epoch": 0.79, + "learning_rate": 0.00017092878635907723, + "loss": 1.3589, + "theoretical_loss": 3.386227232287905, + "tokens_seen": 2602762240 + }, + { + "epoch": 0.79, + "learning_rate": 0.00017091273821464393, + "loss": 1.5347, + "theoretical_loss": 3.3862205426977736, + "tokens_seen": 2602827776 + }, + { + "epoch": 0.79, + "learning_rate": 0.00017089669007021063, + "loss": 1.1736, + "theoretical_loss": 3.3862138533232367, + "tokens_seen": 2602893312 + }, + { + "epoch": 0.79, + "learning_rate": 0.00017088064192577734, + "loss": 1.2965, + "theoretical_loss": 3.3862071641642815, + "tokens_seen": 2602958848 + }, + { + "epoch": 0.79, + "learning_rate": 0.00017086459378134406, + "loss": 1.3133, + "theoretical_loss": 3.3862004752208956, + "tokens_seen": 2603024384 + }, + { + "epoch": 0.79, + "learning_rate": 0.00017084854563691076, + "loss": 1.3521, + "theoretical_loss": 3.3861937864930667, + "tokens_seen": 2603089920 + }, + { + "epoch": 0.79, + "learning_rate": 0.00017083249749247744, + "loss": 1.2594, + "theoretical_loss": 3.386187097980783, + "tokens_seen": 2603155456 + }, + { + "epoch": 0.79, + "learning_rate": 0.00017081644934804414, + "loss": 1.2591, + "theoretical_loss": 3.386180409684031, + "tokens_seen": 2603220992 + }, + { + "epoch": 0.79, + "learning_rate": 0.00017080040120361084, + "loss": 1.3803, + "theoretical_loss": 3.386173721602799, + "tokens_seen": 2603286528 + }, + { + "epoch": 0.79, + "learning_rate": 0.00017078435305917754, + "loss": 1.2897, + "theoretical_loss": 3.386167033737075, + "tokens_seen": 2603352064 + }, + { + "epoch": 0.79, + "learning_rate": 0.00017076830491474424, + "loss": 1.3265, + "theoretical_loss": 3.3861603460868457, + "tokens_seen": 2603417600 + }, + { + "epoch": 0.79, + "learning_rate": 0.00017075225677031094, + "loss": 1.2515, + "theoretical_loss": 3.3861536586520993, + "tokens_seen": 2603483136 + }, + { + "epoch": 0.79, + "learning_rate": 0.00017073620862587764, + "loss": 1.2251, + "theoretical_loss": 3.386146971432823, + "tokens_seen": 2603548672 + }, + { + "epoch": 0.79, + "learning_rate": 0.00017072016048144434, + "loss": 1.3184, + "theoretical_loss": 3.3861402844290054, + "tokens_seen": 2603614208 + }, + { + "epoch": 0.79, + "learning_rate": 0.00017070411233701104, + "loss": 1.1775, + "theoretical_loss": 3.386133597640633, + "tokens_seen": 2603679744 + }, + { + "epoch": 0.79, + "learning_rate": 0.00017068806419257774, + "loss": 1.3861, + "theoretical_loss": 3.386126911067694, + "tokens_seen": 2603745280 + }, + { + "epoch": 0.79, + "learning_rate": 0.00017067201604814444, + "loss": 1.3275, + "theoretical_loss": 3.3861202247101767, + "tokens_seen": 2603810816 + }, + { + "epoch": 0.79, + "learning_rate": 0.00017065596790371114, + "loss": 1.2965, + "theoretical_loss": 3.3861135385680674, + "tokens_seen": 2603876352 + }, + { + "epoch": 0.79, + "learning_rate": 0.00017063991975927784, + "loss": 1.3382, + "theoretical_loss": 3.3861068526413542, + "tokens_seen": 2603941888 + }, + { + "epoch": 0.79, + "learning_rate": 0.00017062387161484455, + "loss": 1.368, + "theoretical_loss": 3.3861001669300252, + "tokens_seen": 2604007424 + }, + { + "epoch": 0.79, + "learning_rate": 0.00017060782347041125, + "loss": 1.2287, + "theoretical_loss": 3.3860934814340675, + "tokens_seen": 2604072960 + }, + { + "epoch": 0.79, + "learning_rate": 0.00017059177532597795, + "loss": 1.2844, + "theoretical_loss": 3.3860867961534695, + "tokens_seen": 2604138496 + }, + { + "epoch": 0.79, + "learning_rate": 0.00017057572718154465, + "loss": 1.3473, + "theoretical_loss": 3.386080111088218, + "tokens_seen": 2604204032 + }, + { + "epoch": 0.79, + "learning_rate": 0.00017055967903711135, + "loss": 1.4039, + "theoretical_loss": 3.3860734262383008, + "tokens_seen": 2604269568 + }, + { + "epoch": 0.79, + "learning_rate": 0.00017054363089267805, + "loss": 1.289, + "theoretical_loss": 3.386066741603706, + "tokens_seen": 2604335104 + }, + { + "epoch": 0.79, + "learning_rate": 0.00017052758274824475, + "loss": 1.2693, + "theoretical_loss": 3.3860600571844204, + "tokens_seen": 2604400640 + }, + { + "epoch": 0.79, + "learning_rate": 0.00017051153460381145, + "loss": 1.168, + "theoretical_loss": 3.3860533729804327, + "tokens_seen": 2604466176 + }, + { + "epoch": 0.79, + "learning_rate": 0.00017049548645937815, + "loss": 1.3032, + "theoretical_loss": 3.38604668899173, + "tokens_seen": 2604531712 + }, + { + "epoch": 0.79, + "learning_rate": 0.00017047943831494485, + "loss": 1.3496, + "theoretical_loss": 3.3860400052183, + "tokens_seen": 2604597248 + }, + { + "epoch": 0.79, + "learning_rate": 0.00017046339017051155, + "loss": 1.4211, + "theoretical_loss": 3.3860333216601304, + "tokens_seen": 2604662784 + }, + { + "epoch": 0.79, + "learning_rate": 0.00017044734202607823, + "loss": 1.275, + "theoretical_loss": 3.3860266383172086, + "tokens_seen": 2604728320 + }, + { + "epoch": 0.79, + "learning_rate": 0.00017043129388164493, + "loss": 1.3792, + "theoretical_loss": 3.3860199551895227, + "tokens_seen": 2604793856 + }, + { + "epoch": 0.79, + "learning_rate": 0.00017041524573721163, + "loss": 1.3755, + "theoretical_loss": 3.3860132722770597, + "tokens_seen": 2604859392 + }, + { + "epoch": 0.79, + "learning_rate": 0.00017039919759277833, + "loss": 1.2242, + "theoretical_loss": 3.386006589579808, + "tokens_seen": 2604924928 + }, + { + "epoch": 0.79, + "learning_rate": 0.00017038314944834503, + "loss": 1.2096, + "theoretical_loss": 3.385999907097755, + "tokens_seen": 2604990464 + }, + { + "epoch": 0.79, + "objective/train/docs_used": 947237, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.1654672622680664, + "objective/train/theoretical_loss": 3.385993224830888, + "objective/train/tokens_used": 2625516000, + "theoretical_loss": 3.385993224830888, + "tokens_seen": 2605056000 + }, + { + "epoch": 0.79, + "learning_rate": 0.00017036710130391173, + "loss": 1.2808, + "theoretical_loss": 3.385993224830888, + "tokens_seen": 2605056000 + }, + { + "epoch": 0.79, + "learning_rate": 0.00017035105315947843, + "loss": 1.2918, + "theoretical_loss": 3.3859865427791953, + "tokens_seen": 2605121536 + }, + { + "epoch": 0.79, + "learning_rate": 0.00017033500501504513, + "loss": 1.2418, + "theoretical_loss": 3.385979860942664, + "tokens_seen": 2605187072 + }, + { + "epoch": 0.79, + "learning_rate": 0.00017031895687061183, + "loss": 1.2474, + "theoretical_loss": 3.385973179321282, + "tokens_seen": 2605252608 + }, + { + "epoch": 0.79, + "learning_rate": 0.00017030290872617853, + "loss": 1.3574, + "theoretical_loss": 3.3859664979150366, + "tokens_seen": 2605318144 + }, + { + "epoch": 0.79, + "learning_rate": 0.00017028686058174523, + "loss": 1.2846, + "theoretical_loss": 3.385959816723916, + "tokens_seen": 2605383680 + }, + { + "epoch": 0.79, + "learning_rate": 0.00017027081243731193, + "loss": 1.1961, + "theoretical_loss": 3.385953135747908, + "tokens_seen": 2605449216 + }, + { + "epoch": 0.79, + "learning_rate": 0.00017025476429287866, + "loss": 1.3685, + "theoretical_loss": 3.3859464549869993, + "tokens_seen": 2605514752 + }, + { + "epoch": 0.79, + "learning_rate": 0.00017023871614844536, + "loss": 1.2894, + "theoretical_loss": 3.3859397744411783, + "tokens_seen": 2605580288 + }, + { + "epoch": 0.79, + "learning_rate": 0.00017022266800401206, + "loss": 1.2165, + "theoretical_loss": 3.385933094110433, + "tokens_seen": 2605645824 + }, + { + "epoch": 0.79, + "learning_rate": 0.00017020661985957876, + "loss": 1.2378, + "theoretical_loss": 3.38592641399475, + "tokens_seen": 2605711360 + }, + { + "epoch": 0.79, + "learning_rate": 0.00017019057171514546, + "loss": 1.2971, + "theoretical_loss": 3.385919734094118, + "tokens_seen": 2605776896 + }, + { + "epoch": 0.79, + "learning_rate": 0.00017017452357071216, + "loss": 1.3509, + "theoretical_loss": 3.385913054408524, + "tokens_seen": 2605842432 + }, + { + "epoch": 0.79, + "learning_rate": 0.00017015847542627886, + "loss": 1.2726, + "theoretical_loss": 3.3859063749379557, + "tokens_seen": 2605907968 + }, + { + "epoch": 0.79, + "learning_rate": 0.00017014242728184556, + "loss": 1.2554, + "theoretical_loss": 3.3858996956824012, + "tokens_seen": 2605973504 + }, + { + "epoch": 0.79, + "learning_rate": 0.00017012637913741224, + "loss": 1.3199, + "theoretical_loss": 3.385893016641848, + "tokens_seen": 2606039040 + }, + { + "epoch": 0.79, + "learning_rate": 0.00017011033099297894, + "loss": 1.3708, + "theoretical_loss": 3.3858863378162836, + "tokens_seen": 2606104576 + }, + { + "epoch": 0.79, + "learning_rate": 0.00017009428284854564, + "loss": 1.2996, + "theoretical_loss": 3.385879659205696, + "tokens_seen": 2606170112 + }, + { + "epoch": 0.79, + "learning_rate": 0.00017007823470411234, + "loss": 1.2699, + "theoretical_loss": 3.3858729808100727, + "tokens_seen": 2606235648 + }, + { + "epoch": 0.79, + "learning_rate": 0.00017006218655967904, + "loss": 1.3018, + "theoretical_loss": 3.385866302629401, + "tokens_seen": 2606301184 + }, + { + "epoch": 0.79, + "learning_rate": 0.00017004613841524574, + "loss": 1.2533, + "theoretical_loss": 3.385859624663669, + "tokens_seen": 2606366720 + }, + { + "epoch": 0.79, + "learning_rate": 0.00017003009027081244, + "loss": 1.2222, + "theoretical_loss": 3.3858529469128644, + "tokens_seen": 2606432256 + }, + { + "epoch": 0.79, + "learning_rate": 0.00017001404212637914, + "loss": 1.1953, + "theoretical_loss": 3.385846269376975, + "tokens_seen": 2606497792 + }, + { + "epoch": 0.79, + "learning_rate": 0.00016999799398194584, + "loss": 1.2881, + "theoretical_loss": 3.385839592055988, + "tokens_seen": 2606563328 + }, + { + "epoch": 0.79, + "learning_rate": 0.00016998194583751254, + "loss": 1.5236, + "theoretical_loss": 3.385832914949891, + "tokens_seen": 2606628864 + }, + { + "epoch": 0.79, + "learning_rate": 0.00016996589769307924, + "loss": 1.3541, + "theoretical_loss": 3.3858262380586726, + "tokens_seen": 2606694400 + }, + { + "epoch": 0.79, + "learning_rate": 0.00016994984954864594, + "loss": 1.2822, + "theoretical_loss": 3.38581956138232, + "tokens_seen": 2606759936 + }, + { + "epoch": 0.79, + "learning_rate": 0.00016993380140421265, + "loss": 1.2674, + "theoretical_loss": 3.38581288492082, + "tokens_seen": 2606825472 + }, + { + "epoch": 0.79, + "learning_rate": 0.00016991775325977935, + "loss": 1.3176, + "theoretical_loss": 3.3858062086741616, + "tokens_seen": 2606891008 + }, + { + "epoch": 0.79, + "learning_rate": 0.00016990170511534605, + "loss": 1.2106, + "theoretical_loss": 3.385799532642332, + "tokens_seen": 2606956544 + }, + { + "epoch": 0.79, + "learning_rate": 0.00016988565697091275, + "loss": 1.3183, + "theoretical_loss": 3.3857928568253186, + "tokens_seen": 2607022080 + }, + { + "epoch": 0.79, + "learning_rate": 0.00016986960882647945, + "loss": 1.2768, + "theoretical_loss": 3.38578618122311, + "tokens_seen": 2607087616 + }, + { + "epoch": 0.79, + "learning_rate": 0.00016985356068204615, + "loss": 1.3488, + "theoretical_loss": 3.3857795058356923, + "tokens_seen": 2607153152 + }, + { + "epoch": 0.79, + "learning_rate": 0.00016983751253761285, + "loss": 1.2651, + "theoretical_loss": 3.3857728306630546, + "tokens_seen": 2607218688 + }, + { + "epoch": 0.79, + "learning_rate": 0.00016982146439317955, + "loss": 1.259, + "theoretical_loss": 3.385766155705184, + "tokens_seen": 2607284224 + }, + { + "epoch": 0.79, + "learning_rate": 0.00016980541624874625, + "loss": 1.2395, + "theoretical_loss": 3.385759480962068, + "tokens_seen": 2607349760 + }, + { + "epoch": 0.79, + "learning_rate": 0.00016978936810431295, + "loss": 1.3675, + "theoretical_loss": 3.385752806433695, + "tokens_seen": 2607415296 + }, + { + "epoch": 0.79, + "learning_rate": 0.00016977331995987965, + "loss": 1.2657, + "theoretical_loss": 3.385746132120052, + "tokens_seen": 2607480832 + }, + { + "epoch": 0.79, + "learning_rate": 0.00016975727181544635, + "loss": 1.3162, + "theoretical_loss": 3.3857394580211273, + "tokens_seen": 2607546368 + }, + { + "epoch": 0.79, + "learning_rate": 0.00016974122367101303, + "loss": 1.3025, + "theoretical_loss": 3.3857327841369083, + "tokens_seen": 2607611904 + }, + { + "epoch": 0.79, + "learning_rate": 0.00016972517552657973, + "loss": 1.2368, + "theoretical_loss": 3.3857261104673824, + "tokens_seen": 2607677440 + }, + { + "epoch": 0.79, + "learning_rate": 0.00016970912738214643, + "loss": 1.1824, + "theoretical_loss": 3.3857194370125376, + "tokens_seen": 2607742976 + }, + { + "epoch": 0.79, + "learning_rate": 0.00016969307923771313, + "loss": 1.2611, + "theoretical_loss": 3.3857127637723616, + "tokens_seen": 2607808512 + }, + { + "epoch": 0.79, + "learning_rate": 0.00016967703109327983, + "loss": 1.2412, + "theoretical_loss": 3.385706090746842, + "tokens_seen": 2607874048 + }, + { + "epoch": 0.79, + "learning_rate": 0.00016966098294884656, + "loss": 1.2691, + "theoretical_loss": 3.3856994179359665, + "tokens_seen": 2607939584 + }, + { + "epoch": 0.79, + "learning_rate": 0.00016964493480441326, + "loss": 1.3295, + "theoretical_loss": 3.385692745339723, + "tokens_seen": 2608005120 + }, + { + "epoch": 0.79, + "learning_rate": 0.00016962888665997996, + "loss": 1.3558, + "theoretical_loss": 3.3856860729580993, + "tokens_seen": 2608070656 + }, + { + "epoch": 0.79, + "learning_rate": 0.00016961283851554666, + "loss": 1.3585, + "theoretical_loss": 3.3856794007910826, + "tokens_seen": 2608136192 + }, + { + "epoch": 0.79, + "learning_rate": 0.00016959679037111336, + "loss": 1.1458, + "theoretical_loss": 3.385672728838661, + "tokens_seen": 2608201728 + }, + { + "epoch": 0.79, + "learning_rate": 0.00016958074222668006, + "loss": 1.1908, + "theoretical_loss": 3.385666057100822, + "tokens_seen": 2608267264 + }, + { + "epoch": 0.79, + "objective/train/docs_used": 948381, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.2919771671295166, + "objective/train/theoretical_loss": 3.3856593855775534, + "objective/train/tokens_used": 2628792800, + "theoretical_loss": 3.3856593855775534, + "tokens_seen": 2608332800 + }, + { + "epoch": 0.79, + "learning_rate": 0.00016956469408224676, + "loss": 1.2957, + "theoretical_loss": 3.3856593855775534, + "tokens_seen": 2608332800 + }, + { + "epoch": 0.79, + "learning_rate": 0.00016954864593781346, + "loss": 1.3747, + "theoretical_loss": 3.3856527142688426, + "tokens_seen": 2608398336 + }, + { + "epoch": 0.79, + "learning_rate": 0.00016953259779338016, + "loss": 1.2954, + "theoretical_loss": 3.385646043174678, + "tokens_seen": 2608463872 + }, + { + "epoch": 0.79, + "learning_rate": 0.00016951654964894686, + "loss": 1.2494, + "theoretical_loss": 3.3856393722950466, + "tokens_seen": 2608529408 + }, + { + "epoch": 0.79, + "learning_rate": 0.00016950050150451356, + "loss": 1.3098, + "theoretical_loss": 3.385632701629937, + "tokens_seen": 2608594944 + }, + { + "epoch": 0.79, + "learning_rate": 0.00016948445336008026, + "loss": 1.3078, + "theoretical_loss": 3.385626031179336, + "tokens_seen": 2608660480 + }, + { + "epoch": 0.79, + "learning_rate": 0.00016946840521564696, + "loss": 1.3066, + "theoretical_loss": 3.3856193609432315, + "tokens_seen": 2608726016 + }, + { + "epoch": 0.79, + "learning_rate": 0.00016945235707121366, + "loss": 1.2044, + "theoretical_loss": 3.3856126909216115, + "tokens_seen": 2608791552 + }, + { + "epoch": 0.79, + "learning_rate": 0.00016943630892678034, + "loss": 1.251, + "theoretical_loss": 3.3856060211144636, + "tokens_seen": 2608857088 + }, + { + "epoch": 0.79, + "learning_rate": 0.00016942026078234704, + "loss": 1.4089, + "theoretical_loss": 3.3855993515217757, + "tokens_seen": 2608922624 + }, + { + "epoch": 0.79, + "learning_rate": 0.00016940421263791374, + "loss": 1.2077, + "theoretical_loss": 3.385592682143535, + "tokens_seen": 2608988160 + }, + { + "epoch": 0.79, + "learning_rate": 0.00016938816449348044, + "loss": 1.2128, + "theoretical_loss": 3.3855860129797297, + "tokens_seen": 2609053696 + }, + { + "epoch": 0.79, + "learning_rate": 0.00016937211634904714, + "loss": 1.3581, + "theoretical_loss": 3.385579344030347, + "tokens_seen": 2609119232 + }, + { + "epoch": 0.79, + "learning_rate": 0.00016935606820461384, + "loss": 1.3483, + "theoretical_loss": 3.385572675295376, + "tokens_seen": 2609184768 + }, + { + "epoch": 0.79, + "learning_rate": 0.00016934002006018054, + "loss": 1.2837, + "theoretical_loss": 3.385566006774803, + "tokens_seen": 2609250304 + }, + { + "epoch": 0.79, + "learning_rate": 0.00016932397191574724, + "loss": 1.3737, + "theoretical_loss": 3.3855593384686156, + "tokens_seen": 2609315840 + }, + { + "epoch": 0.79, + "learning_rate": 0.00016930792377131394, + "loss": 1.1929, + "theoretical_loss": 3.385552670376802, + "tokens_seen": 2609381376 + }, + { + "epoch": 0.79, + "learning_rate": 0.00016929187562688064, + "loss": 1.3658, + "theoretical_loss": 3.3855460024993502, + "tokens_seen": 2609446912 + }, + { + "epoch": 0.79, + "learning_rate": 0.00016927582748244734, + "loss": 1.3376, + "theoretical_loss": 3.385539334836248, + "tokens_seen": 2609512448 + }, + { + "epoch": 0.79, + "learning_rate": 0.00016925977933801405, + "loss": 1.2672, + "theoretical_loss": 3.3855326673874826, + "tokens_seen": 2609577984 + }, + { + "epoch": 0.79, + "learning_rate": 0.00016924373119358075, + "loss": 1.3076, + "theoretical_loss": 3.385526000153042, + "tokens_seen": 2609643520 + }, + { + "epoch": 0.79, + "learning_rate": 0.00016922768304914745, + "loss": 1.1423, + "theoretical_loss": 3.3855193331329136, + "tokens_seen": 2609709056 + }, + { + "epoch": 0.79, + "learning_rate": 0.00016921163490471415, + "loss": 1.2381, + "theoretical_loss": 3.385512666327086, + "tokens_seen": 2609774592 + }, + { + "epoch": 0.79, + "learning_rate": 0.00016919558676028085, + "loss": 1.2995, + "theoretical_loss": 3.3855059997355457, + "tokens_seen": 2609840128 + }, + { + "epoch": 0.79, + "learning_rate": 0.00016917953861584755, + "loss": 1.3201, + "theoretical_loss": 3.3854993333582817, + "tokens_seen": 2609905664 + }, + { + "epoch": 0.79, + "learning_rate": 0.00016916349047141425, + "loss": 1.3657, + "theoretical_loss": 3.3854926671952805, + "tokens_seen": 2609971200 + }, + { + "epoch": 0.79, + "learning_rate": 0.00016914744232698095, + "loss": 1.4269, + "theoretical_loss": 3.3854860012465307, + "tokens_seen": 2610036736 + }, + { + "epoch": 0.79, + "learning_rate": 0.00016913139418254765, + "loss": 1.221, + "theoretical_loss": 3.3854793355120196, + "tokens_seen": 2610102272 + }, + { + "epoch": 0.79, + "learning_rate": 0.00016911534603811435, + "loss": 1.2614, + "theoretical_loss": 3.3854726699917355, + "tokens_seen": 2610167808 + }, + { + "epoch": 0.79, + "learning_rate": 0.00016909929789368105, + "loss": 1.3273, + "theoretical_loss": 3.3854660046856653, + "tokens_seen": 2610233344 + }, + { + "epoch": 0.79, + "learning_rate": 0.00016908324974924775, + "loss": 1.2723, + "theoretical_loss": 3.3854593395937975, + "tokens_seen": 2610298880 + }, + { + "epoch": 0.79, + "learning_rate": 0.00016906720160481445, + "loss": 1.3231, + "theoretical_loss": 3.3854526747161193, + "tokens_seen": 2610364416 + }, + { + "epoch": 0.79, + "learning_rate": 0.00016905115346038115, + "loss": 1.2046, + "theoretical_loss": 3.3854460100526187, + "tokens_seen": 2610429952 + }, + { + "epoch": 0.79, + "learning_rate": 0.00016903510531594785, + "loss": 1.334, + "theoretical_loss": 3.3854393456032836, + "tokens_seen": 2610495488 + }, + { + "epoch": 0.79, + "learning_rate": 0.00016901905717151455, + "loss": 1.38, + "theoretical_loss": 3.3854326813681013, + "tokens_seen": 2610561024 + }, + { + "epoch": 0.79, + "learning_rate": 0.00016900300902708126, + "loss": 1.2069, + "theoretical_loss": 3.38542601734706, + "tokens_seen": 2610626560 + }, + { + "epoch": 0.79, + "learning_rate": 0.00016898696088264796, + "loss": 1.215, + "theoretical_loss": 3.385419353540147, + "tokens_seen": 2610692096 + }, + { + "epoch": 0.79, + "learning_rate": 0.00016897091273821466, + "loss": 1.2952, + "theoretical_loss": 3.3854126899473505, + "tokens_seen": 2610757632 + }, + { + "epoch": 0.79, + "learning_rate": 0.00016895486459378136, + "loss": 1.3365, + "theoretical_loss": 3.385406026568658, + "tokens_seen": 2610823168 + }, + { + "epoch": 0.79, + "learning_rate": 0.00016893881644934806, + "loss": 1.3193, + "theoretical_loss": 3.385399363404057, + "tokens_seen": 2610888704 + }, + { + "epoch": 0.79, + "learning_rate": 0.00016892276830491476, + "loss": 1.3099, + "theoretical_loss": 3.3853927004535356, + "tokens_seen": 2610954240 + }, + { + "epoch": 0.79, + "learning_rate": 0.00016890672016048146, + "loss": 1.3358, + "theoretical_loss": 3.3853860377170815, + "tokens_seen": 2611019776 + }, + { + "epoch": 0.79, + "learning_rate": 0.00016889067201604816, + "loss": 1.3818, + "theoretical_loss": 3.385379375194683, + "tokens_seen": 2611085312 + }, + { + "epoch": 0.79, + "learning_rate": 0.00016887462387161486, + "loss": 1.3207, + "theoretical_loss": 3.3853727128863262, + "tokens_seen": 2611150848 + }, + { + "epoch": 0.79, + "learning_rate": 0.00016885857572718156, + "loss": 1.2804, + "theoretical_loss": 3.3853660507920007, + "tokens_seen": 2611216384 + }, + { + "epoch": 0.79, + "learning_rate": 0.00016884252758274826, + "loss": 1.3123, + "theoretical_loss": 3.3853593889116933, + "tokens_seen": 2611281920 + }, + { + "epoch": 0.79, + "learning_rate": 0.00016882647943831496, + "loss": 1.3476, + "theoretical_loss": 3.3853527272453916, + "tokens_seen": 2611347456 + }, + { + "epoch": 0.79, + "learning_rate": 0.00016881043129388166, + "loss": 1.2813, + "theoretical_loss": 3.385346065793084, + "tokens_seen": 2611412992 + }, + { + "epoch": 0.79, + "learning_rate": 0.00016879438314944836, + "loss": 1.2905, + "theoretical_loss": 3.385339404554758, + "tokens_seen": 2611478528 + }, + { + "epoch": 0.79, + "learning_rate": 0.00016877833500501506, + "loss": 1.3575, + "theoretical_loss": 3.385332743530401, + "tokens_seen": 2611544064 + }, + { + "epoch": 0.79, + "objective/train/docs_used": 949550, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.202101707458496, + "objective/train/theoretical_loss": 3.385326082720001, + "objective/train/tokens_used": 2632069600, + "theoretical_loss": 3.385326082720001, + "tokens_seen": 2611609600 + }, + { + "epoch": 0.79, + "learning_rate": 0.00016876228686058176, + "loss": 1.3427, + "theoretical_loss": 3.385326082720001, + "tokens_seen": 2611609600 + }, + { + "epoch": 0.79, + "learning_rate": 0.00016874623871614847, + "loss": 1.4682, + "theoretical_loss": 3.385319422123546, + "tokens_seen": 2611675136 + }, + { + "epoch": 0.79, + "learning_rate": 0.00016873019057171514, + "loss": 1.3255, + "theoretical_loss": 3.3853127617410235, + "tokens_seen": 2611740672 + }, + { + "epoch": 0.79, + "learning_rate": 0.00016871414242728184, + "loss": 1.2897, + "theoretical_loss": 3.385306101572421, + "tokens_seen": 2611806208 + }, + { + "epoch": 0.79, + "learning_rate": 0.00016869809428284854, + "loss": 1.4562, + "theoretical_loss": 3.385299441617727, + "tokens_seen": 2611871744 + }, + { + "epoch": 0.79, + "learning_rate": 0.00016868204613841524, + "loss": 1.3091, + "theoretical_loss": 3.385292781876929, + "tokens_seen": 2611937280 + }, + { + "epoch": 0.79, + "learning_rate": 0.00016866599799398194, + "loss": 1.2865, + "theoretical_loss": 3.3852861223500144, + "tokens_seen": 2612002816 + }, + { + "epoch": 0.79, + "learning_rate": 0.00016864994984954864, + "loss": 1.3772, + "theoretical_loss": 3.385279463036971, + "tokens_seen": 2612068352 + }, + { + "epoch": 0.79, + "learning_rate": 0.00016863390170511534, + "loss": 1.2847, + "theoretical_loss": 3.385272803937787, + "tokens_seen": 2612133888 + }, + { + "epoch": 0.79, + "learning_rate": 0.00016861785356068204, + "loss": 1.3095, + "theoretical_loss": 3.3852661450524497, + "tokens_seen": 2612199424 + }, + { + "epoch": 0.79, + "learning_rate": 0.00016860180541624874, + "loss": 1.4657, + "theoretical_loss": 3.3852594863809475, + "tokens_seen": 2612264960 + }, + { + "epoch": 0.79, + "learning_rate": 0.00016858575727181544, + "loss": 1.32, + "theoretical_loss": 3.385252827923267, + "tokens_seen": 2612330496 + }, + { + "epoch": 0.79, + "learning_rate": 0.00016856970912738215, + "loss": 1.3531, + "theoretical_loss": 3.385246169679397, + "tokens_seen": 2612396032 + }, + { + "epoch": 0.79, + "learning_rate": 0.00016855366098294885, + "loss": 1.317, + "theoretical_loss": 3.3852395116493255, + "tokens_seen": 2612461568 + }, + { + "epoch": 0.79, + "learning_rate": 0.00016853761283851555, + "loss": 1.3531, + "theoretical_loss": 3.385232853833039, + "tokens_seen": 2612527104 + }, + { + "epoch": 0.79, + "learning_rate": 0.00016852156469408225, + "loss": 1.107, + "theoretical_loss": 3.385226196230527, + "tokens_seen": 2612592640 + }, + { + "epoch": 0.79, + "learning_rate": 0.00016850551654964895, + "loss": 1.2768, + "theoretical_loss": 3.385219538841776, + "tokens_seen": 2612658176 + }, + { + "epoch": 0.79, + "learning_rate": 0.00016848946840521565, + "loss": 1.1592, + "theoretical_loss": 3.3852128816667735, + "tokens_seen": 2612723712 + }, + { + "epoch": 0.79, + "learning_rate": 0.00016847342026078235, + "loss": 1.3463, + "theoretical_loss": 3.385206224705508, + "tokens_seen": 2612789248 + }, + { + "epoch": 0.79, + "learning_rate": 0.00016845737211634908, + "loss": 1.3948, + "theoretical_loss": 3.3851995679579674, + "tokens_seen": 2612854784 + }, + { + "epoch": 0.79, + "learning_rate": 0.00016844132397191578, + "loss": 1.2912, + "theoretical_loss": 3.385192911424139, + "tokens_seen": 2612920320 + }, + { + "epoch": 0.79, + "learning_rate": 0.00016842527582748248, + "loss": 1.2237, + "theoretical_loss": 3.385186255104011, + "tokens_seen": 2612985856 + }, + { + "epoch": 0.79, + "learning_rate": 0.00016840922768304915, + "loss": 1.2947, + "theoretical_loss": 3.3851795989975706, + "tokens_seen": 2613051392 + }, + { + "epoch": 0.79, + "learning_rate": 0.00016839317953861585, + "loss": 1.2717, + "theoretical_loss": 3.3851729431048065, + "tokens_seen": 2613116928 + }, + { + "epoch": 0.79, + "learning_rate": 0.00016837713139418255, + "loss": 1.3135, + "theoretical_loss": 3.3851662874257054, + "tokens_seen": 2613182464 + }, + { + "epoch": 0.79, + "learning_rate": 0.00016836108324974925, + "loss": 1.3382, + "theoretical_loss": 3.385159631960256, + "tokens_seen": 2613248000 + }, + { + "epoch": 0.79, + "learning_rate": 0.00016834503510531595, + "loss": 1.3647, + "theoretical_loss": 3.3851529767084454, + "tokens_seen": 2613313536 + }, + { + "epoch": 0.79, + "learning_rate": 0.00016832898696088265, + "loss": 1.2652, + "theoretical_loss": 3.3851463216702617, + "tokens_seen": 2613379072 + }, + { + "epoch": 0.79, + "learning_rate": 0.00016831293881644936, + "loss": 1.1868, + "theoretical_loss": 3.385139666845693, + "tokens_seen": 2613444608 + }, + { + "epoch": 0.79, + "learning_rate": 0.00016829689067201606, + "loss": 1.4313, + "theoretical_loss": 3.3851330122347267, + "tokens_seen": 2613510144 + }, + { + "epoch": 0.79, + "learning_rate": 0.00016828084252758276, + "loss": 1.3123, + "theoretical_loss": 3.3851263578373505, + "tokens_seen": 2613575680 + }, + { + "epoch": 0.79, + "learning_rate": 0.00016826479438314946, + "loss": 1.2696, + "theoretical_loss": 3.3851197036535523, + "tokens_seen": 2613641216 + }, + { + "epoch": 0.79, + "learning_rate": 0.00016824874623871616, + "loss": 1.2946, + "theoretical_loss": 3.38511304968332, + "tokens_seen": 2613706752 + }, + { + "epoch": 0.79, + "learning_rate": 0.00016823269809428286, + "loss": 1.3285, + "theoretical_loss": 3.385106395926641, + "tokens_seen": 2613772288 + }, + { + "epoch": 0.79, + "learning_rate": 0.00016821664994984956, + "loss": 1.3815, + "theoretical_loss": 3.385099742383504, + "tokens_seen": 2613837824 + }, + { + "epoch": 0.79, + "learning_rate": 0.00016820060180541626, + "loss": 1.2545, + "theoretical_loss": 3.3850930890538957, + "tokens_seen": 2613903360 + }, + { + "epoch": 0.79, + "learning_rate": 0.00016818455366098296, + "loss": 1.3386, + "theoretical_loss": 3.385086435937805, + "tokens_seen": 2613968896 + }, + { + "epoch": 0.79, + "learning_rate": 0.00016816850551654966, + "loss": 1.3043, + "theoretical_loss": 3.385079783035218, + "tokens_seen": 2614034432 + }, + { + "epoch": 0.79, + "learning_rate": 0.00016815245737211636, + "loss": 1.268, + "theoretical_loss": 3.3850731303461243, + "tokens_seen": 2614099968 + }, + { + "epoch": 0.79, + "learning_rate": 0.00016813640922768306, + "loss": 1.3731, + "theoretical_loss": 3.3850664778705113, + "tokens_seen": 2614165504 + }, + { + "epoch": 0.79, + "learning_rate": 0.00016812036108324976, + "loss": 1.1731, + "theoretical_loss": 3.3850598256083657, + "tokens_seen": 2614231040 + }, + { + "epoch": 0.79, + "learning_rate": 0.00016810431293881646, + "loss": 1.2718, + "theoretical_loss": 3.385053173559677, + "tokens_seen": 2614296576 + }, + { + "epoch": 0.79, + "learning_rate": 0.00016808826479438316, + "loss": 1.2819, + "theoretical_loss": 3.385046521724431, + "tokens_seen": 2614362112 + }, + { + "epoch": 0.79, + "learning_rate": 0.00016807221664994986, + "loss": 1.2157, + "theoretical_loss": 3.3850398701026174, + "tokens_seen": 2614427648 + }, + { + "epoch": 0.79, + "learning_rate": 0.00016805616850551657, + "loss": 1.3407, + "theoretical_loss": 3.385033218694223, + "tokens_seen": 2614493184 + }, + { + "epoch": 0.79, + "learning_rate": 0.00016804012036108327, + "loss": 1.2694, + "theoretical_loss": 3.385026567499236, + "tokens_seen": 2614558720 + }, + { + "epoch": 0.79, + "learning_rate": 0.00016802407221664994, + "loss": 1.2868, + "theoretical_loss": 3.3850199165176433, + "tokens_seen": 2614624256 + }, + { + "epoch": 0.79, + "learning_rate": 0.00016800802407221664, + "loss": 1.246, + "theoretical_loss": 3.385013265749434, + "tokens_seen": 2614689792 + }, + { + "epoch": 0.79, + "learning_rate": 0.00016799197592778334, + "loss": 1.1863, + "theoretical_loss": 3.385006615194595, + "tokens_seen": 2614755328 + }, + { + "epoch": 0.79, + "learning_rate": 0.00016797592778335004, + "loss": 1.1953, + "theoretical_loss": 3.3849999648531144, + "tokens_seen": 2614820864 + }, + { + "epoch": 0.79, + "objective/train/docs_used": 950739, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.3132990598678589, + "objective/train/theoretical_loss": 3.38499331472498, + "objective/train/tokens_used": 2635346400, + "theoretical_loss": 3.38499331472498, + "tokens_seen": 2614886400 + }, + { + "epoch": 0.79, + "learning_rate": 0.00016795987963891674, + "loss": 1.3447, + "theoretical_loss": 3.38499331472498, + "tokens_seen": 2614886400 + }, + { + "epoch": 0.79, + "learning_rate": 0.00016794383149448344, + "loss": 1.1927, + "theoretical_loss": 3.38498666481018, + "tokens_seen": 2614951936 + }, + { + "epoch": 0.79, + "learning_rate": 0.00016792778335005014, + "loss": 1.342, + "theoretical_loss": 3.3849800151087015, + "tokens_seen": 2615017472 + }, + { + "epoch": 0.79, + "learning_rate": 0.00016791173520561684, + "loss": 1.3587, + "theoretical_loss": 3.3849733656205325, + "tokens_seen": 2615083008 + }, + { + "epoch": 0.79, + "learning_rate": 0.00016789568706118354, + "loss": 1.3348, + "theoretical_loss": 3.3849667163456614, + "tokens_seen": 2615148544 + }, + { + "epoch": 0.79, + "learning_rate": 0.00016787963891675025, + "loss": 1.3266, + "theoretical_loss": 3.384960067284075, + "tokens_seen": 2615214080 + }, + { + "epoch": 0.79, + "learning_rate": 0.00016786359077231697, + "loss": 1.3131, + "theoretical_loss": 3.384953418435762, + "tokens_seen": 2615279616 + }, + { + "epoch": 0.79, + "learning_rate": 0.00016784754262788367, + "loss": 1.2865, + "theoretical_loss": 3.38494676980071, + "tokens_seen": 2615345152 + }, + { + "epoch": 0.79, + "learning_rate": 0.00016783149448345037, + "loss": 1.2327, + "theoretical_loss": 3.3849401213789068, + "tokens_seen": 2615410688 + }, + { + "epoch": 0.79, + "learning_rate": 0.00016781544633901708, + "loss": 1.2629, + "theoretical_loss": 3.3849334731703395, + "tokens_seen": 2615476224 + }, + { + "epoch": 0.79, + "learning_rate": 0.00016779939819458378, + "loss": 1.3276, + "theoretical_loss": 3.3849268251749973, + "tokens_seen": 2615541760 + }, + { + "epoch": 0.79, + "learning_rate": 0.00016778335005015048, + "loss": 1.268, + "theoretical_loss": 3.3849201773928668, + "tokens_seen": 2615607296 + }, + { + "epoch": 0.79, + "learning_rate": 0.00016776730190571718, + "loss": 1.2336, + "theoretical_loss": 3.384913529823937, + "tokens_seen": 2615672832 + }, + { + "epoch": 0.79, + "learning_rate": 0.00016775125376128388, + "loss": 1.3113, + "theoretical_loss": 3.384906882468194, + "tokens_seen": 2615738368 + }, + { + "epoch": 0.79, + "learning_rate": 0.00016773520561685058, + "loss": 1.3562, + "theoretical_loss": 3.384900235325627, + "tokens_seen": 2615803904 + }, + { + "epoch": 0.79, + "learning_rate": 0.00016771915747241728, + "loss": 1.2545, + "theoretical_loss": 3.3848935883962237, + "tokens_seen": 2615869440 + }, + { + "epoch": 0.79, + "learning_rate": 0.00016770310932798395, + "loss": 1.3409, + "theoretical_loss": 3.384886941679971, + "tokens_seen": 2615934976 + }, + { + "epoch": 0.79, + "learning_rate": 0.00016768706118355065, + "loss": 1.3327, + "theoretical_loss": 3.3848802951768584, + "tokens_seen": 2616000512 + }, + { + "epoch": 0.79, + "learning_rate": 0.00016767101303911735, + "loss": 1.2063, + "theoretical_loss": 3.384873648886872, + "tokens_seen": 2616066048 + }, + { + "epoch": 0.79, + "learning_rate": 0.00016765496489468405, + "loss": 1.2916, + "theoretical_loss": 3.384867002810001, + "tokens_seen": 2616131584 + }, + { + "epoch": 0.79, + "learning_rate": 0.00016763891675025075, + "loss": 1.295, + "theoretical_loss": 3.384860356946232, + "tokens_seen": 2616197120 + }, + { + "epoch": 0.79, + "learning_rate": 0.00016762286860581746, + "loss": 1.2788, + "theoretical_loss": 3.3848537112955537, + "tokens_seen": 2616262656 + }, + { + "epoch": 0.79, + "learning_rate": 0.00016760682046138416, + "loss": 1.3336, + "theoretical_loss": 3.3848470658579535, + "tokens_seen": 2616328192 + }, + { + "epoch": 0.79, + "learning_rate": 0.00016759077231695086, + "loss": 1.2944, + "theoretical_loss": 3.3848404206334193, + "tokens_seen": 2616393728 + }, + { + "epoch": 0.79, + "learning_rate": 0.00016757472417251756, + "loss": 1.2493, + "theoretical_loss": 3.384833775621939, + "tokens_seen": 2616459264 + }, + { + "epoch": 0.79, + "learning_rate": 0.00016755867602808426, + "loss": 1.3624, + "theoretical_loss": 3.3848271308235005, + "tokens_seen": 2616524800 + }, + { + "epoch": 0.79, + "learning_rate": 0.00016754262788365096, + "loss": 1.2802, + "theoretical_loss": 3.3848204862380915, + "tokens_seen": 2616590336 + }, + { + "epoch": 0.79, + "learning_rate": 0.00016752657973921766, + "loss": 1.1507, + "theoretical_loss": 3.3848138418657, + "tokens_seen": 2616655872 + }, + { + "epoch": 0.79, + "learning_rate": 0.00016751053159478436, + "loss": 1.3259, + "theoretical_loss": 3.384807197706314, + "tokens_seen": 2616721408 + }, + { + "epoch": 0.79, + "learning_rate": 0.00016749448345035106, + "loss": 1.402, + "theoretical_loss": 3.384800553759921, + "tokens_seen": 2616786944 + }, + { + "epoch": 0.79, + "learning_rate": 0.00016747843530591776, + "loss": 1.3367, + "theoretical_loss": 3.3847939100265085, + "tokens_seen": 2616852480 + }, + { + "epoch": 0.79, + "learning_rate": 0.00016746238716148446, + "loss": 1.2727, + "theoretical_loss": 3.384787266506065, + "tokens_seen": 2616918016 + }, + { + "epoch": 0.79, + "learning_rate": 0.00016744633901705116, + "loss": 1.1444, + "theoretical_loss": 3.384780623198578, + "tokens_seen": 2616983552 + }, + { + "epoch": 0.79, + "learning_rate": 0.00016743029087261786, + "loss": 1.3372, + "theoretical_loss": 3.3847739801040357, + "tokens_seen": 2617049088 + }, + { + "epoch": 0.79, + "learning_rate": 0.00016741424272818456, + "loss": 1.3329, + "theoretical_loss": 3.3847673372224256, + "tokens_seen": 2617114624 + }, + { + "epoch": 0.79, + "learning_rate": 0.00016739819458375126, + "loss": 1.231, + "theoretical_loss": 3.3847606945537354, + "tokens_seen": 2617180160 + }, + { + "epoch": 0.79, + "learning_rate": 0.00016738214643931797, + "loss": 1.2998, + "theoretical_loss": 3.3847540520979535, + "tokens_seen": 2617245696 + }, + { + "epoch": 0.79, + "learning_rate": 0.00016736609829488467, + "loss": 1.2773, + "theoretical_loss": 3.384747409855067, + "tokens_seen": 2617311232 + }, + { + "epoch": 0.79, + "learning_rate": 0.00016735005015045137, + "loss": 1.2412, + "theoretical_loss": 3.3847407678250647, + "tokens_seen": 2617376768 + }, + { + "epoch": 0.79, + "learning_rate": 0.00016733400200601807, + "loss": 1.2982, + "theoretical_loss": 3.3847341260079338, + "tokens_seen": 2617442304 + }, + { + "epoch": 0.79, + "learning_rate": 0.00016731795386158474, + "loss": 1.2507, + "theoretical_loss": 3.384727484403662, + "tokens_seen": 2617507840 + }, + { + "epoch": 0.79, + "learning_rate": 0.00016730190571715144, + "loss": 1.2962, + "theoretical_loss": 3.3847208430122375, + "tokens_seen": 2617573376 + }, + { + "epoch": 0.79, + "learning_rate": 0.00016728585757271814, + "loss": 1.3201, + "theoretical_loss": 3.384714201833648, + "tokens_seen": 2617638912 + }, + { + "epoch": 0.79, + "learning_rate": 0.00016726980942828484, + "loss": 1.1935, + "theoretical_loss": 3.384707560867881, + "tokens_seen": 2617704448 + }, + { + "epoch": 0.79, + "learning_rate": 0.00016725376128385157, + "loss": 1.2857, + "theoretical_loss": 3.384700920114925, + "tokens_seen": 2617769984 + }, + { + "epoch": 0.79, + "learning_rate": 0.00016723771313941827, + "loss": 1.3281, + "theoretical_loss": 3.3846942795747683, + "tokens_seen": 2617835520 + }, + { + "epoch": 0.79, + "learning_rate": 0.00016722166499498497, + "loss": 1.3127, + "theoretical_loss": 3.3846876392473972, + "tokens_seen": 2617901056 + }, + { + "epoch": 0.79, + "learning_rate": 0.00016720561685055167, + "loss": 1.4061, + "theoretical_loss": 3.384680999132801, + "tokens_seen": 2617966592 + }, + { + "epoch": 0.79, + "learning_rate": 0.00016718956870611837, + "loss": 1.2267, + "theoretical_loss": 3.3846743592309667, + "tokens_seen": 2618032128 + }, + { + "epoch": 0.79, + "learning_rate": 0.00016717352056168507, + "loss": 1.2172, + "theoretical_loss": 3.3846677195418824, + "tokens_seen": 2618097664 + }, + { + "epoch": 0.79, + "objective/train/docs_used": 951960, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.3764095306396484, + "objective/train/theoretical_loss": 3.384661080065536, + "objective/train/tokens_used": 2638623200, + "theoretical_loss": 3.384661080065536, + "tokens_seen": 2618163200 + }, + { + "epoch": 0.79, + "learning_rate": 0.00016715747241725177, + "loss": 1.3741, + "theoretical_loss": 3.384661080065536, + "tokens_seen": 2618163200 + }, + { + "epoch": 0.79, + "learning_rate": 0.00016714142427281847, + "loss": 1.1848, + "theoretical_loss": 3.3846544408019152, + "tokens_seen": 2618228736 + }, + { + "epoch": 0.79, + "learning_rate": 0.00016712537612838518, + "loss": 1.2311, + "theoretical_loss": 3.3846478017510084, + "tokens_seen": 2618294272 + }, + { + "epoch": 0.79, + "learning_rate": 0.00016710932798395188, + "loss": 1.2898, + "theoretical_loss": 3.384641162912803, + "tokens_seen": 2618359808 + }, + { + "epoch": 0.79, + "learning_rate": 0.00016709327983951858, + "loss": 1.2247, + "theoretical_loss": 3.3846345242872866, + "tokens_seen": 2618425344 + }, + { + "epoch": 0.79, + "learning_rate": 0.00016707723169508528, + "loss": 1.2286, + "theoretical_loss": 3.3846278858744476, + "tokens_seen": 2618490880 + }, + { + "epoch": 0.79, + "learning_rate": 0.00016706118355065198, + "loss": 1.244, + "theoretical_loss": 3.3846212476742736, + "tokens_seen": 2618556416 + }, + { + "epoch": 0.79, + "learning_rate": 0.00016704513540621868, + "loss": 1.2797, + "theoretical_loss": 3.3846146096867527, + "tokens_seen": 2618621952 + }, + { + "epoch": 0.79, + "learning_rate": 0.00016702908726178538, + "loss": 1.4112, + "theoretical_loss": 3.3846079719118727, + "tokens_seen": 2618687488 + }, + { + "epoch": 0.79, + "learning_rate": 0.00016701303911735205, + "loss": 1.2824, + "theoretical_loss": 3.384601334349621, + "tokens_seen": 2618753024 + }, + { + "epoch": 0.79, + "learning_rate": 0.00016699699097291875, + "loss": 1.4076, + "theoretical_loss": 3.3845946969999856, + "tokens_seen": 2618818560 + }, + { + "epoch": 0.79, + "learning_rate": 0.00016698094282848545, + "loss": 1.2713, + "theoretical_loss": 3.384588059862955, + "tokens_seen": 2618884096 + }, + { + "epoch": 0.79, + "learning_rate": 0.00016696489468405215, + "loss": 1.3584, + "theoretical_loss": 3.3845814229385165, + "tokens_seen": 2618949632 + }, + { + "epoch": 0.79, + "learning_rate": 0.00016694884653961886, + "loss": 1.3907, + "theoretical_loss": 3.384574786226658, + "tokens_seen": 2619015168 + }, + { + "epoch": 0.79, + "learning_rate": 0.00016693279839518556, + "loss": 1.2662, + "theoretical_loss": 3.384568149727368, + "tokens_seen": 2619080704 + }, + { + "epoch": 0.79, + "learning_rate": 0.00016691675025075226, + "loss": 1.3268, + "theoretical_loss": 3.3845615134406337, + "tokens_seen": 2619146240 + }, + { + "epoch": 0.79, + "learning_rate": 0.00016690070210631896, + "loss": 1.4256, + "theoretical_loss": 3.384554877366443, + "tokens_seen": 2619211776 + }, + { + "epoch": 0.79, + "learning_rate": 0.00016688465396188566, + "loss": 1.3146, + "theoretical_loss": 3.384548241504784, + "tokens_seen": 2619277312 + }, + { + "epoch": 0.79, + "learning_rate": 0.00016686860581745236, + "loss": 1.2402, + "theoretical_loss": 3.3845416058556443, + "tokens_seen": 2619342848 + }, + { + "epoch": 0.79, + "learning_rate": 0.00016685255767301906, + "loss": 1.2726, + "theoretical_loss": 3.384534970419012, + "tokens_seen": 2619408384 + }, + { + "epoch": 0.79, + "learning_rate": 0.00016683650952858576, + "loss": 1.3369, + "theoretical_loss": 3.3845283351948754, + "tokens_seen": 2619473920 + }, + { + "epoch": 0.79, + "learning_rate": 0.00016682046138415246, + "loss": 1.1858, + "theoretical_loss": 3.3845217001832215, + "tokens_seen": 2619539456 + }, + { + "epoch": 0.79, + "learning_rate": 0.00016680441323971916, + "loss": 1.2118, + "theoretical_loss": 3.384515065384039, + "tokens_seen": 2619604992 + }, + { + "epoch": 0.79, + "learning_rate": 0.00016678836509528586, + "loss": 1.2984, + "theoretical_loss": 3.384508430797315, + "tokens_seen": 2619670528 + }, + { + "epoch": 0.79, + "learning_rate": 0.00016677231695085256, + "loss": 1.2479, + "theoretical_loss": 3.384501796423038, + "tokens_seen": 2619736064 + }, + { + "epoch": 0.79, + "learning_rate": 0.00016675626880641926, + "loss": 1.3303, + "theoretical_loss": 3.3844951622611954, + "tokens_seen": 2619801600 + }, + { + "epoch": 0.79, + "learning_rate": 0.00016674022066198596, + "loss": 1.2184, + "theoretical_loss": 3.384488528311776, + "tokens_seen": 2619867136 + }, + { + "epoch": 0.79, + "learning_rate": 0.00016672417251755266, + "loss": 1.2713, + "theoretical_loss": 3.3844818945747663, + "tokens_seen": 2619932672 + }, + { + "epoch": 0.79, + "learning_rate": 0.00016670812437311936, + "loss": 1.2734, + "theoretical_loss": 3.384475261050155, + "tokens_seen": 2619998208 + }, + { + "epoch": 0.79, + "learning_rate": 0.00016669207622868607, + "loss": 1.3567, + "theoretical_loss": 3.38446862773793, + "tokens_seen": 2620063744 + }, + { + "epoch": 0.79, + "learning_rate": 0.00016667602808425277, + "loss": 1.2725, + "theoretical_loss": 3.3844619946380794, + "tokens_seen": 2620129280 + }, + { + "epoch": 0.79, + "learning_rate": 0.00016665997993981947, + "loss": 1.312, + "theoretical_loss": 3.3844553617505904, + "tokens_seen": 2620194816 + }, + { + "epoch": 0.79, + "learning_rate": 0.00016664393179538617, + "loss": 1.349, + "theoretical_loss": 3.3844487290754515, + "tokens_seen": 2620260352 + }, + { + "epoch": 0.79, + "learning_rate": 0.00016662788365095287, + "loss": 1.3201, + "theoretical_loss": 3.3844420966126503, + "tokens_seen": 2620325888 + }, + { + "epoch": 0.79, + "learning_rate": 0.00016661183550651957, + "loss": 1.2592, + "theoretical_loss": 3.384435464362175, + "tokens_seen": 2620391424 + }, + { + "epoch": 0.79, + "learning_rate": 0.00016659578736208627, + "loss": 1.4203, + "theoretical_loss": 3.3844288323240126, + "tokens_seen": 2620456960 + }, + { + "epoch": 0.79, + "learning_rate": 0.00016657973921765297, + "loss": 1.4246, + "theoretical_loss": 3.384422200498152, + "tokens_seen": 2620522496 + }, + { + "epoch": 0.79, + "learning_rate": 0.00016656369107321967, + "loss": 1.2132, + "theoretical_loss": 3.384415568884581, + "tokens_seen": 2620588032 + }, + { + "epoch": 0.79, + "learning_rate": 0.00016654764292878637, + "loss": 1.3157, + "theoretical_loss": 3.384408937483287, + "tokens_seen": 2620653568 + }, + { + "epoch": 0.79, + "learning_rate": 0.00016653159478435307, + "loss": 1.2838, + "theoretical_loss": 3.384402306294258, + "tokens_seen": 2620719104 + }, + { + "epoch": 0.79, + "learning_rate": 0.00016651554663991977, + "loss": 1.2514, + "theoretical_loss": 3.384395675317482, + "tokens_seen": 2620784640 + }, + { + "epoch": 0.79, + "learning_rate": 0.00016649949849548647, + "loss": 1.2134, + "theoretical_loss": 3.384389044552947, + "tokens_seen": 2620850176 + }, + { + "epoch": 0.79, + "learning_rate": 0.00016648345035105317, + "loss": 1.2329, + "theoretical_loss": 3.384382414000641, + "tokens_seen": 2620915712 + }, + { + "epoch": 0.79, + "learning_rate": 0.00016646740220661987, + "loss": 1.3262, + "theoretical_loss": 3.3843757836605515, + "tokens_seen": 2620981248 + }, + { + "epoch": 0.79, + "learning_rate": 0.00016645135406218657, + "loss": 1.2016, + "theoretical_loss": 3.3843691535326665, + "tokens_seen": 2621046784 + }, + { + "epoch": 0.79, + "learning_rate": 0.00016643530591775328, + "loss": 1.21, + "theoretical_loss": 3.384362523616974, + "tokens_seen": 2621112320 + }, + { + "epoch": 0.79, + "learning_rate": 0.00016641925777331998, + "loss": 1.2085, + "theoretical_loss": 3.3843558939134626, + "tokens_seen": 2621177856 + }, + { + "epoch": 0.79, + "learning_rate": 0.00016640320962888668, + "loss": 1.2946, + "theoretical_loss": 3.384349264422119, + "tokens_seen": 2621243392 + }, + { + "epoch": 0.79, + "learning_rate": 0.00016638716148445338, + "loss": 1.2625, + "theoretical_loss": 3.3843426351429318, + "tokens_seen": 2621308928 + }, + { + "epoch": 0.79, + "learning_rate": 0.00016637111334002008, + "loss": 1.327, + "theoretical_loss": 3.3843360060758885, + "tokens_seen": 2621374464 + }, + { + "debugging/Compilability": 1.0, + "debugging/distinct-1-grams": 0.7183788677974512, + "debugging/entropy-1-grams": 5.311092611595205, + "debugging/length": 475.35714285714283, + "debugging/num_segments": 14, + "debugging/score": 0.005309415865714915, + "debugging/score_std": 0.0067471201632161425, + "epoch": 0.79, + "objective/train/docs_used": 953088, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.3135061264038086, + "objective/train/theoretical_loss": 3.3843293772209777, + "objective/train/tokens_used": 2641900000, + "theoretical_loss": 3.3843293772209777, + "tokens_seen": 2621440000 + }, + { + "epoch": 0.79, + "learning_rate": 0.00016635506519558678, + "loss": 1.3332, + "theoretical_loss": 3.3843293772209777, + "tokens_seen": 2621440000 + }, + { + "epoch": 0.79, + "learning_rate": 0.00016633901705115348, + "loss": 1.237, + "theoretical_loss": 3.3843227485781866, + "tokens_seen": 2621505536 + }, + { + "epoch": 0.79, + "learning_rate": 0.00016632296890672018, + "loss": 1.1916, + "theoretical_loss": 3.384316120147503, + "tokens_seen": 2621571072 + }, + { + "epoch": 0.79, + "learning_rate": 0.00016630692076228685, + "loss": 1.2871, + "theoretical_loss": 3.384309491928916, + "tokens_seen": 2621636608 + }, + { + "epoch": 0.79, + "learning_rate": 0.00016629087261785355, + "loss": 1.2802, + "theoretical_loss": 3.3843028639224118, + "tokens_seen": 2621702144 + }, + { + "epoch": 0.79, + "learning_rate": 0.00016627482447342025, + "loss": 1.2315, + "theoretical_loss": 3.38429623612798, + "tokens_seen": 2621767680 + }, + { + "epoch": 0.79, + "learning_rate": 0.00016625877632898696, + "loss": 1.2773, + "theoretical_loss": 3.384289608545607, + "tokens_seen": 2621833216 + }, + { + "epoch": 0.79, + "learning_rate": 0.00016624272818455366, + "loss": 1.2723, + "theoretical_loss": 3.384282981175282, + "tokens_seen": 2621898752 + }, + { + "epoch": 0.79, + "learning_rate": 0.00016622668004012036, + "loss": 1.3759, + "theoretical_loss": 3.3842763540169924, + "tokens_seen": 2621964288 + }, + { + "epoch": 0.79, + "learning_rate": 0.00016621063189568706, + "loss": 1.2888, + "theoretical_loss": 3.3842697270707256, + "tokens_seen": 2622029824 + }, + { + "epoch": 0.79, + "learning_rate": 0.00016619458375125376, + "loss": 1.2775, + "theoretical_loss": 3.3842631003364705, + "tokens_seen": 2622095360 + }, + { + "epoch": 0.79, + "learning_rate": 0.00016617853560682046, + "loss": 1.3289, + "theoretical_loss": 3.3842564738142142, + "tokens_seen": 2622160896 + }, + { + "epoch": 0.79, + "learning_rate": 0.00016616248746238716, + "loss": 1.2332, + "theoretical_loss": 3.384249847503945, + "tokens_seen": 2622226432 + }, + { + "epoch": 0.79, + "learning_rate": 0.00016614643931795386, + "loss": 1.1734, + "theoretical_loss": 3.3842432214056504, + "tokens_seen": 2622291968 + }, + { + "epoch": 0.79, + "learning_rate": 0.00016613039117352056, + "loss": 1.2143, + "theoretical_loss": 3.3842365955193188, + "tokens_seen": 2622357504 + }, + { + "epoch": 0.79, + "learning_rate": 0.00016611434302908726, + "loss": 1.2821, + "theoretical_loss": 3.3842299698449385, + "tokens_seen": 2622423040 + }, + { + "epoch": 0.79, + "learning_rate": 0.00016609829488465396, + "loss": 1.2613, + "theoretical_loss": 3.3842233443824963, + "tokens_seen": 2622488576 + }, + { + "epoch": 0.79, + "learning_rate": 0.00016608224674022066, + "loss": 1.3117, + "theoretical_loss": 3.384216719131981, + "tokens_seen": 2622554112 + }, + { + "epoch": 0.79, + "learning_rate": 0.00016606619859578736, + "loss": 1.2412, + "theoretical_loss": 3.3842100940933806, + "tokens_seen": 2622619648 + }, + { + "epoch": 0.79, + "learning_rate": 0.0001660501504513541, + "loss": 1.2526, + "theoretical_loss": 3.3842034692666823, + "tokens_seen": 2622685184 + }, + { + "epoch": 0.79, + "learning_rate": 0.0001660341023069208, + "loss": 1.2275, + "theoretical_loss": 3.384196844651874, + "tokens_seen": 2622750720 + }, + { + "epoch": 0.79, + "learning_rate": 0.0001660180541624875, + "loss": 1.3046, + "theoretical_loss": 3.3841902202489447, + "tokens_seen": 2622816256 + }, + { + "epoch": 0.79, + "learning_rate": 0.0001660020060180542, + "loss": 1.2029, + "theoretical_loss": 3.3841835960578814, + "tokens_seen": 2622881792 + }, + { + "epoch": 0.79, + "learning_rate": 0.00016598595787362087, + "loss": 1.2347, + "theoretical_loss": 3.3841769720786727, + "tokens_seen": 2622947328 + }, + { + "epoch": 0.79, + "learning_rate": 0.00016596990972918757, + "loss": 1.3529, + "theoretical_loss": 3.3841703483113057, + "tokens_seen": 2623012864 + }, + { + "epoch": 0.79, + "learning_rate": 0.00016595386158475427, + "loss": 1.3799, + "theoretical_loss": 3.3841637247557688, + "tokens_seen": 2623078400 + }, + { + "epoch": 0.79, + "learning_rate": 0.00016593781344032097, + "loss": 1.3489, + "theoretical_loss": 3.38415710141205, + "tokens_seen": 2623143936 + }, + { + "epoch": 0.79, + "learning_rate": 0.00016592176529588767, + "loss": 1.4414, + "theoretical_loss": 3.3841504782801373, + "tokens_seen": 2623209472 + }, + { + "epoch": 0.79, + "learning_rate": 0.00016590571715145437, + "loss": 1.2284, + "theoretical_loss": 3.3841438553600183, + "tokens_seen": 2623275008 + }, + { + "epoch": 0.79, + "learning_rate": 0.00016588966900702107, + "loss": 1.4314, + "theoretical_loss": 3.384137232651681, + "tokens_seen": 2623340544 + }, + { + "epoch": 0.79, + "learning_rate": 0.00016587362086258777, + "loss": 1.3236, + "theoretical_loss": 3.3841306101551134, + "tokens_seen": 2623406080 + }, + { + "epoch": 0.79, + "learning_rate": 0.00016585757271815447, + "loss": 1.3366, + "theoretical_loss": 3.3841239878703036, + "tokens_seen": 2623471616 + }, + { + "epoch": 0.8, + "learning_rate": 0.00016584152457372117, + "loss": 1.2956, + "theoretical_loss": 3.3841173657972394, + "tokens_seen": 2623537152 + }, + { + "epoch": 0.8, + "learning_rate": 0.00016582547642928787, + "loss": 1.3487, + "theoretical_loss": 3.384110743935909, + "tokens_seen": 2623602688 + }, + { + "epoch": 0.8, + "learning_rate": 0.00016580942828485457, + "loss": 1.2778, + "theoretical_loss": 3.3841041222863, + "tokens_seen": 2623668224 + }, + { + "epoch": 0.8, + "learning_rate": 0.00016579338014042127, + "loss": 1.3981, + "theoretical_loss": 3.3840975008484, + "tokens_seen": 2623733760 + }, + { + "epoch": 0.8, + "learning_rate": 0.00016577733199598797, + "loss": 1.2557, + "theoretical_loss": 3.3840908796221982, + "tokens_seen": 2623799296 + }, + { + "epoch": 0.8, + "learning_rate": 0.00016576128385155468, + "loss": 1.2805, + "theoretical_loss": 3.384084258607681, + "tokens_seen": 2623864832 + }, + { + "epoch": 0.8, + "learning_rate": 0.00016574523570712138, + "loss": 1.2733, + "theoretical_loss": 3.384077637804838, + "tokens_seen": 2623930368 + }, + { + "epoch": 0.8, + "learning_rate": 0.00016572918756268808, + "loss": 1.3065, + "theoretical_loss": 3.384071017213655, + "tokens_seen": 2623995904 + }, + { + "epoch": 0.8, + "learning_rate": 0.00016571313941825478, + "loss": 1.3168, + "theoretical_loss": 3.384064396834122, + "tokens_seen": 2624061440 + }, + { + "epoch": 0.8, + "learning_rate": 0.00016569709127382148, + "loss": 1.3238, + "theoretical_loss": 3.384057776666226, + "tokens_seen": 2624126976 + }, + { + "epoch": 0.8, + "learning_rate": 0.00016568104312938818, + "loss": 1.3756, + "theoretical_loss": 3.3840511567099547, + "tokens_seen": 2624192512 + }, + { + "epoch": 0.8, + "learning_rate": 0.00016566499498495488, + "loss": 1.2495, + "theoretical_loss": 3.384044536965297, + "tokens_seen": 2624258048 + }, + { + "epoch": 0.8, + "learning_rate": 0.00016564894684052158, + "loss": 1.2792, + "theoretical_loss": 3.3840379174322397, + "tokens_seen": 2624323584 + }, + { + "epoch": 0.8, + "learning_rate": 0.00016563289869608828, + "loss": 1.1951, + "theoretical_loss": 3.384031298110772, + "tokens_seen": 2624389120 + }, + { + "epoch": 0.8, + "learning_rate": 0.00016561685055165498, + "loss": 1.3822, + "theoretical_loss": 3.3840246790008806, + "tokens_seen": 2624454656 + }, + { + "epoch": 0.8, + "learning_rate": 0.00016560080240722165, + "loss": 1.2518, + "theoretical_loss": 3.3840180601025542, + "tokens_seen": 2624520192 + }, + { + "epoch": 0.8, + "learning_rate": 0.00016558475426278835, + "loss": 1.3989, + "theoretical_loss": 3.3840114414157805, + "tokens_seen": 2624585728 + }, + { + "epoch": 0.8, + "learning_rate": 0.00016556870611835506, + "loss": 1.4212, + "theoretical_loss": 3.3840048229405477, + "tokens_seen": 2624651264 + }, + { + "epoch": 0.8, + "objective/train/docs_used": 954280, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.2140306234359741, + "objective/train/theoretical_loss": 3.3839982046768435, + "objective/train/tokens_used": 2645176800, + "theoretical_loss": 3.3839982046768435, + "tokens_seen": 2624716800 + }, + { + "epoch": 0.8, + "learning_rate": 0.00016555265797392176, + "loss": 1.2731, + "theoretical_loss": 3.3839982046768435, + "tokens_seen": 2624716800 + }, + { + "epoch": 0.8, + "learning_rate": 0.00016553660982948846, + "loss": 1.2795, + "theoretical_loss": 3.383991586624656, + "tokens_seen": 2624782336 + }, + { + "epoch": 0.8, + "learning_rate": 0.00016552056168505516, + "loss": 1.2844, + "theoretical_loss": 3.383984968783973, + "tokens_seen": 2624847872 + }, + { + "epoch": 0.8, + "learning_rate": 0.00016550451354062186, + "loss": 1.3859, + "theoretical_loss": 3.3839783511547825, + "tokens_seen": 2624913408 + }, + { + "epoch": 0.8, + "learning_rate": 0.00016548846539618856, + "loss": 1.2856, + "theoretical_loss": 3.3839717337370727, + "tokens_seen": 2624978944 + }, + { + "epoch": 0.8, + "learning_rate": 0.00016547241725175526, + "loss": 1.3739, + "theoretical_loss": 3.3839651165308315, + "tokens_seen": 2625044480 + }, + { + "epoch": 0.8, + "learning_rate": 0.000165456369107322, + "loss": 1.3586, + "theoretical_loss": 3.383958499536047, + "tokens_seen": 2625110016 + }, + { + "epoch": 0.8, + "learning_rate": 0.0001654403209628887, + "loss": 1.3656, + "theoretical_loss": 3.383951882752706, + "tokens_seen": 2625175552 + }, + { + "epoch": 0.8, + "learning_rate": 0.0001654242728184554, + "loss": 1.2645, + "theoretical_loss": 3.3839452661807985, + "tokens_seen": 2625241088 + }, + { + "epoch": 0.8, + "learning_rate": 0.0001654082246740221, + "loss": 1.2355, + "theoretical_loss": 3.383938649820311, + "tokens_seen": 2625306624 + }, + { + "epoch": 0.8, + "learning_rate": 0.0001653921765295888, + "loss": 1.3693, + "theoretical_loss": 3.3839320336712317, + "tokens_seen": 2625372160 + }, + { + "epoch": 0.8, + "learning_rate": 0.0001653761283851555, + "loss": 1.2476, + "theoretical_loss": 3.3839254177335487, + "tokens_seen": 2625437696 + }, + { + "epoch": 0.8, + "learning_rate": 0.0001653600802407222, + "loss": 1.4079, + "theoretical_loss": 3.3839188020072495, + "tokens_seen": 2625503232 + }, + { + "epoch": 0.8, + "learning_rate": 0.0001653440320962889, + "loss": 1.2561, + "theoretical_loss": 3.3839121864923234, + "tokens_seen": 2625568768 + }, + { + "epoch": 0.8, + "learning_rate": 0.0001653279839518556, + "loss": 1.391, + "theoretical_loss": 3.383905571188757, + "tokens_seen": 2625634304 + }, + { + "epoch": 0.8, + "learning_rate": 0.0001653119358074223, + "loss": 1.2892, + "theoretical_loss": 3.383898956096539, + "tokens_seen": 2625699840 + }, + { + "epoch": 0.8, + "learning_rate": 0.00016529588766298897, + "loss": 1.185, + "theoretical_loss": 3.383892341215657, + "tokens_seen": 2625765376 + }, + { + "epoch": 0.8, + "learning_rate": 0.00016527983951855567, + "loss": 1.2319, + "theoretical_loss": 3.383885726546099, + "tokens_seen": 2625830912 + }, + { + "epoch": 0.8, + "learning_rate": 0.00016526379137412237, + "loss": 1.3002, + "theoretical_loss": 3.383879112087853, + "tokens_seen": 2625896448 + }, + { + "epoch": 0.8, + "learning_rate": 0.00016524774322968907, + "loss": 1.2946, + "theoretical_loss": 3.3838724978409074, + "tokens_seen": 2625961984 + }, + { + "epoch": 0.8, + "learning_rate": 0.00016523169508525577, + "loss": 1.313, + "theoretical_loss": 3.38386588380525, + "tokens_seen": 2626027520 + }, + { + "epoch": 0.8, + "learning_rate": 0.00016521564694082247, + "loss": 1.374, + "theoretical_loss": 3.383859269980868, + "tokens_seen": 2626093056 + }, + { + "epoch": 0.8, + "learning_rate": 0.00016519959879638917, + "loss": 1.3464, + "theoretical_loss": 3.3838526563677505, + "tokens_seen": 2626158592 + }, + { + "epoch": 0.8, + "learning_rate": 0.00016518355065195587, + "loss": 1.367, + "theoretical_loss": 3.383846042965885, + "tokens_seen": 2626224128 + }, + { + "epoch": 0.8, + "learning_rate": 0.00016516750250752257, + "loss": 1.3171, + "theoretical_loss": 3.3838394297752594, + "tokens_seen": 2626289664 + }, + { + "epoch": 0.8, + "learning_rate": 0.00016515145436308927, + "loss": 1.2671, + "theoretical_loss": 3.3838328167958616, + "tokens_seen": 2626355200 + }, + { + "epoch": 0.8, + "learning_rate": 0.00016513540621865597, + "loss": 1.2643, + "theoretical_loss": 3.38382620402768, + "tokens_seen": 2626420736 + }, + { + "epoch": 0.8, + "learning_rate": 0.00016511935807422267, + "loss": 1.1816, + "theoretical_loss": 3.383819591470702, + "tokens_seen": 2626486272 + }, + { + "epoch": 0.8, + "learning_rate": 0.00016510330992978937, + "loss": 1.3282, + "theoretical_loss": 3.383812979124916, + "tokens_seen": 2626551808 + }, + { + "epoch": 0.8, + "learning_rate": 0.00016508726178535607, + "loss": 1.2932, + "theoretical_loss": 3.38380636699031, + "tokens_seen": 2626617344 + }, + { + "epoch": 0.8, + "learning_rate": 0.00016507121364092278, + "loss": 1.2923, + "theoretical_loss": 3.383799755066872, + "tokens_seen": 2626682880 + }, + { + "epoch": 0.8, + "learning_rate": 0.00016505516549648948, + "loss": 1.3068, + "theoretical_loss": 3.3837931433545894, + "tokens_seen": 2626748416 + }, + { + "epoch": 0.8, + "learning_rate": 0.00016503911735205618, + "loss": 1.263, + "theoretical_loss": 3.383786531853451, + "tokens_seen": 2626813952 + }, + { + "epoch": 0.8, + "learning_rate": 0.00016502306920762288, + "loss": 1.3382, + "theoretical_loss": 3.3837799205634442, + "tokens_seen": 2626879488 + }, + { + "epoch": 0.8, + "learning_rate": 0.00016500702106318958, + "loss": 1.3098, + "theoretical_loss": 3.3837733094845572, + "tokens_seen": 2626945024 + }, + { + "epoch": 0.8, + "learning_rate": 0.00016499097291875628, + "loss": 1.4109, + "theoretical_loss": 3.3837666986167783, + "tokens_seen": 2627010560 + }, + { + "epoch": 0.8, + "learning_rate": 0.00016497492477432298, + "loss": 1.2712, + "theoretical_loss": 3.3837600879600953, + "tokens_seen": 2627076096 + }, + { + "epoch": 0.8, + "learning_rate": 0.00016495887662988968, + "loss": 1.346, + "theoretical_loss": 3.383753477514496, + "tokens_seen": 2627141632 + }, + { + "epoch": 0.8, + "learning_rate": 0.00016494282848545638, + "loss": 1.2576, + "theoretical_loss": 3.383746867279968, + "tokens_seen": 2627207168 + }, + { + "epoch": 0.8, + "learning_rate": 0.00016492678034102308, + "loss": 1.2372, + "theoretical_loss": 3.3837402572565005, + "tokens_seen": 2627272704 + }, + { + "epoch": 0.8, + "learning_rate": 0.00016491073219658975, + "loss": 1.2871, + "theoretical_loss": 3.3837336474440805, + "tokens_seen": 2627338240 + }, + { + "epoch": 0.8, + "learning_rate": 0.00016489468405215646, + "loss": 1.3564, + "theoretical_loss": 3.383727037842696, + "tokens_seen": 2627403776 + }, + { + "epoch": 0.8, + "learning_rate": 0.00016487863590772316, + "loss": 1.2777, + "theoretical_loss": 3.3837204284523357, + "tokens_seen": 2627469312 + }, + { + "epoch": 0.8, + "learning_rate": 0.00016486258776328988, + "loss": 1.3472, + "theoretical_loss": 3.383713819272987, + "tokens_seen": 2627534848 + }, + { + "epoch": 0.8, + "learning_rate": 0.00016484653961885658, + "loss": 1.2142, + "theoretical_loss": 3.3837072103046384, + "tokens_seen": 2627600384 + }, + { + "epoch": 0.8, + "learning_rate": 0.00016483049147442328, + "loss": 1.3289, + "theoretical_loss": 3.383700601547277, + "tokens_seen": 2627665920 + }, + { + "epoch": 0.8, + "learning_rate": 0.00016481444332998999, + "loss": 1.2533, + "theoretical_loss": 3.383693993000892, + "tokens_seen": 2627731456 + }, + { + "epoch": 0.8, + "learning_rate": 0.00016479839518555669, + "loss": 1.2929, + "theoretical_loss": 3.3836873846654703, + "tokens_seen": 2627796992 + }, + { + "epoch": 0.8, + "learning_rate": 0.0001647823470411234, + "loss": 1.2851, + "theoretical_loss": 3.383680776541001, + "tokens_seen": 2627862528 + }, + { + "epoch": 0.8, + "learning_rate": 0.0001647662988966901, + "loss": 1.2829, + "theoretical_loss": 3.383674168627471, + "tokens_seen": 2627928064 + }, + { + "epoch": 0.8, + "objective/train/docs_used": 955375, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.3425976037979126, + "objective/train/theoretical_loss": 3.3836675609248688, + "objective/train/tokens_used": 2648453600, + "theoretical_loss": 3.3836675609248688, + "tokens_seen": 2627993600 + }, + { + "epoch": 0.8, + "learning_rate": 0.0001647502507522568, + "loss": 1.3211, + "theoretical_loss": 3.3836675609248688, + "tokens_seen": 2627993600 + }, + { + "epoch": 0.8, + "learning_rate": 0.0001647342026078235, + "loss": 1.3542, + "theoretical_loss": 3.3836609534331825, + "tokens_seen": 2628059136 + }, + { + "epoch": 0.8, + "learning_rate": 0.0001647181544633902, + "loss": 1.3235, + "theoretical_loss": 3.3836543461524, + "tokens_seen": 2628124672 + }, + { + "epoch": 0.8, + "learning_rate": 0.0001647021063189569, + "loss": 1.2602, + "theoretical_loss": 3.3836477390825097, + "tokens_seen": 2628190208 + }, + { + "epoch": 0.8, + "learning_rate": 0.0001646860581745236, + "loss": 1.2235, + "theoretical_loss": 3.3836411322234987, + "tokens_seen": 2628255744 + }, + { + "epoch": 0.8, + "learning_rate": 0.0001646700100300903, + "loss": 1.2826, + "theoretical_loss": 3.383634525575356, + "tokens_seen": 2628321280 + }, + { + "epoch": 0.8, + "learning_rate": 0.000164653961885657, + "loss": 1.2722, + "theoretical_loss": 3.3836279191380694, + "tokens_seen": 2628386816 + }, + { + "epoch": 0.8, + "learning_rate": 0.0001646379137412237, + "loss": 1.3292, + "theoretical_loss": 3.383621312911626, + "tokens_seen": 2628452352 + }, + { + "epoch": 0.8, + "learning_rate": 0.0001646218655967904, + "loss": 1.2959, + "theoretical_loss": 3.3836147068960147, + "tokens_seen": 2628517888 + }, + { + "epoch": 0.8, + "learning_rate": 0.0001646058174523571, + "loss": 1.2974, + "theoretical_loss": 3.3836081010912236, + "tokens_seen": 2628583424 + }, + { + "epoch": 0.8, + "learning_rate": 0.00016458976930792377, + "loss": 1.2143, + "theoretical_loss": 3.38360149549724, + "tokens_seen": 2628648960 + }, + { + "epoch": 0.8, + "learning_rate": 0.00016457372116349047, + "loss": 1.2627, + "theoretical_loss": 3.3835948901140527, + "tokens_seen": 2628714496 + }, + { + "epoch": 0.8, + "learning_rate": 0.00016455767301905717, + "loss": 1.3309, + "theoretical_loss": 3.3835882849416494, + "tokens_seen": 2628780032 + }, + { + "epoch": 0.8, + "learning_rate": 0.00016454162487462387, + "loss": 1.2882, + "theoretical_loss": 3.3835816799800176, + "tokens_seen": 2628845568 + }, + { + "epoch": 0.8, + "learning_rate": 0.00016452557673019057, + "loss": 1.2369, + "theoretical_loss": 3.383575075229146, + "tokens_seen": 2628911104 + }, + { + "epoch": 0.8, + "learning_rate": 0.00016450952858575727, + "loss": 1.2767, + "theoretical_loss": 3.383568470689023, + "tokens_seen": 2628976640 + }, + { + "epoch": 0.8, + "learning_rate": 0.00016449348044132397, + "loss": 1.305, + "theoretical_loss": 3.3835618663596354, + "tokens_seen": 2629042176 + }, + { + "epoch": 0.8, + "learning_rate": 0.00016447743229689067, + "loss": 1.2108, + "theoretical_loss": 3.383555262240972, + "tokens_seen": 2629107712 + }, + { + "epoch": 0.8, + "learning_rate": 0.00016446138415245737, + "loss": 1.3259, + "theoretical_loss": 3.3835486583330208, + "tokens_seen": 2629173248 + }, + { + "epoch": 0.8, + "learning_rate": 0.00016444533600802407, + "loss": 1.1386, + "theoretical_loss": 3.38354205463577, + "tokens_seen": 2629238784 + }, + { + "epoch": 0.8, + "learning_rate": 0.00016442928786359077, + "loss": 1.3025, + "theoretical_loss": 3.3835354511492066, + "tokens_seen": 2629304320 + }, + { + "epoch": 0.8, + "learning_rate": 0.00016441323971915747, + "loss": 1.2286, + "theoretical_loss": 3.38352884787332, + "tokens_seen": 2629369856 + }, + { + "epoch": 0.8, + "learning_rate": 0.00016439719157472417, + "loss": 1.3307, + "theoretical_loss": 3.3835222448080975, + "tokens_seen": 2629435392 + }, + { + "epoch": 0.8, + "learning_rate": 0.00016438114343029088, + "loss": 1.3411, + "theoretical_loss": 3.383515641953527, + "tokens_seen": 2629500928 + }, + { + "epoch": 0.8, + "learning_rate": 0.00016436509528585758, + "loss": 1.339, + "theoretical_loss": 3.3835090393095966, + "tokens_seen": 2629566464 + }, + { + "epoch": 0.8, + "learning_rate": 0.00016434904714142428, + "loss": 1.2292, + "theoretical_loss": 3.383502436876295, + "tokens_seen": 2629632000 + }, + { + "epoch": 0.8, + "learning_rate": 0.00016433299899699098, + "loss": 1.2441, + "theoretical_loss": 3.3834958346536093, + "tokens_seen": 2629697536 + }, + { + "epoch": 0.8, + "learning_rate": 0.00016431695085255768, + "loss": 1.3157, + "theoretical_loss": 3.3834892326415282, + "tokens_seen": 2629763072 + }, + { + "epoch": 0.8, + "learning_rate": 0.00016430090270812438, + "loss": 1.2577, + "theoretical_loss": 3.3834826308400396, + "tokens_seen": 2629828608 + }, + { + "epoch": 0.8, + "learning_rate": 0.00016428485456369108, + "loss": 1.3097, + "theoretical_loss": 3.383476029249131, + "tokens_seen": 2629894144 + }, + { + "epoch": 0.8, + "learning_rate": 0.00016426880641925778, + "loss": 1.2597, + "theoretical_loss": 3.3834694278687913, + "tokens_seen": 2629959680 + }, + { + "epoch": 0.8, + "learning_rate": 0.00016425275827482448, + "loss": 1.315, + "theoretical_loss": 3.3834628266990077, + "tokens_seen": 2630025216 + }, + { + "epoch": 0.8, + "learning_rate": 0.00016423671013039118, + "loss": 1.2739, + "theoretical_loss": 3.383456225739769, + "tokens_seen": 2630090752 + }, + { + "epoch": 0.8, + "learning_rate": 0.00016422066198595788, + "loss": 1.266, + "theoretical_loss": 3.3834496249910626, + "tokens_seen": 2630156288 + }, + { + "epoch": 0.8, + "learning_rate": 0.00016420461384152458, + "loss": 1.3437, + "theoretical_loss": 3.383443024452877, + "tokens_seen": 2630221824 + }, + { + "epoch": 0.8, + "learning_rate": 0.00016418856569709128, + "loss": 1.244, + "theoretical_loss": 3.3834364241252, + "tokens_seen": 2630287360 + }, + { + "epoch": 0.8, + "learning_rate": 0.00016417251755265798, + "loss": 1.2145, + "theoretical_loss": 3.3834298240080196, + "tokens_seen": 2630352896 + }, + { + "epoch": 0.8, + "learning_rate": 0.00016415646940822468, + "loss": 1.2275, + "theoretical_loss": 3.383423224101324, + "tokens_seen": 2630418432 + }, + { + "epoch": 0.8, + "learning_rate": 0.00016414042126379138, + "loss": 1.2147, + "theoretical_loss": 3.383416624405101, + "tokens_seen": 2630483968 + }, + { + "epoch": 0.8, + "learning_rate": 0.00016412437311935809, + "loss": 1.1903, + "theoretical_loss": 3.383410024919339, + "tokens_seen": 2630549504 + }, + { + "epoch": 0.8, + "learning_rate": 0.00016410832497492479, + "loss": 1.2925, + "theoretical_loss": 3.3834034256440257, + "tokens_seen": 2630615040 + }, + { + "epoch": 0.8, + "learning_rate": 0.0001640922768304915, + "loss": 1.3234, + "theoretical_loss": 3.3833968265791494, + "tokens_seen": 2630680576 + }, + { + "epoch": 0.8, + "learning_rate": 0.0001640762286860582, + "loss": 1.4092, + "theoretical_loss": 3.383390227724698, + "tokens_seen": 2630746112 + }, + { + "epoch": 0.8, + "learning_rate": 0.0001640601805416249, + "loss": 1.2649, + "theoretical_loss": 3.38338362908066, + "tokens_seen": 2630811648 + }, + { + "epoch": 0.8, + "learning_rate": 0.0001640441323971916, + "loss": 1.3779, + "theoretical_loss": 3.383377030647023, + "tokens_seen": 2630877184 + }, + { + "epoch": 0.8, + "learning_rate": 0.0001640280842527583, + "loss": 1.3294, + "theoretical_loss": 3.3833704324237748, + "tokens_seen": 2630942720 + }, + { + "epoch": 0.8, + "learning_rate": 0.000164012036108325, + "loss": 1.2222, + "theoretical_loss": 3.3833638344109036, + "tokens_seen": 2631008256 + }, + { + "epoch": 0.8, + "learning_rate": 0.0001639959879638917, + "loss": 1.3047, + "theoretical_loss": 3.383357236608398, + "tokens_seen": 2631073792 + }, + { + "epoch": 0.8, + "learning_rate": 0.0001639799398194584, + "loss": 1.279, + "theoretical_loss": 3.3833506390162453, + "tokens_seen": 2631139328 + }, + { + "epoch": 0.8, + "learning_rate": 0.0001639638916750251, + "loss": 1.317, + "theoretical_loss": 3.3833440416344347, + "tokens_seen": 2631204864 + }, + { + "epoch": 0.8, + "objective/train/docs_used": 956570, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.192093849182129, + "objective/train/theoretical_loss": 3.3833374444629527, + "objective/train/tokens_used": 2651730400, + "theoretical_loss": 3.3833374444629527, + "tokens_seen": 2631270400 + }, + { + "epoch": 0.8, + "learning_rate": 0.0001639478435305918, + "loss": 1.3054, + "theoretical_loss": 3.3833374444629527, + "tokens_seen": 2631270400 + }, + { + "epoch": 0.8, + "learning_rate": 0.0001639317953861585, + "loss": 1.4321, + "theoretical_loss": 3.3833308475017883, + "tokens_seen": 2631335936 + }, + { + "epoch": 0.8, + "learning_rate": 0.0001639157472417252, + "loss": 1.2817, + "theoretical_loss": 3.3833242507509294, + "tokens_seen": 2631401472 + }, + { + "epoch": 0.8, + "learning_rate": 0.0001638996990972919, + "loss": 1.1497, + "theoretical_loss": 3.383317654210364, + "tokens_seen": 2631467008 + }, + { + "epoch": 0.8, + "learning_rate": 0.00016388365095285857, + "loss": 1.2592, + "theoretical_loss": 3.3833110578800807, + "tokens_seen": 2631532544 + }, + { + "epoch": 0.8, + "learning_rate": 0.00016386760280842527, + "loss": 1.3741, + "theoretical_loss": 3.3833044617600665, + "tokens_seen": 2631598080 + }, + { + "epoch": 0.8, + "learning_rate": 0.00016385155466399197, + "loss": 1.379, + "theoretical_loss": 3.3832978658503103, + "tokens_seen": 2631663616 + }, + { + "epoch": 0.8, + "learning_rate": 0.00016383550651955867, + "loss": 1.1912, + "theoretical_loss": 3.3832912701508, + "tokens_seen": 2631729152 + }, + { + "epoch": 0.8, + "learning_rate": 0.00016381945837512537, + "loss": 1.3312, + "theoretical_loss": 3.383284674661523, + "tokens_seen": 2631794688 + }, + { + "epoch": 0.8, + "learning_rate": 0.00016380341023069207, + "loss": 1.2543, + "theoretical_loss": 3.3832780793824684, + "tokens_seen": 2631860224 + }, + { + "epoch": 0.8, + "learning_rate": 0.00016378736208625877, + "loss": 1.3037, + "theoretical_loss": 3.3832714843136236, + "tokens_seen": 2631925760 + }, + { + "epoch": 0.8, + "learning_rate": 0.00016377131394182547, + "loss": 1.326, + "theoretical_loss": 3.383264889454977, + "tokens_seen": 2631991296 + }, + { + "epoch": 0.8, + "learning_rate": 0.00016375526579739217, + "loss": 1.2093, + "theoretical_loss": 3.383258294806516, + "tokens_seen": 2632056832 + }, + { + "epoch": 0.8, + "learning_rate": 0.00016373921765295887, + "loss": 1.3377, + "theoretical_loss": 3.3832517003682296, + "tokens_seen": 2632122368 + }, + { + "epoch": 0.8, + "learning_rate": 0.00016372316950852557, + "loss": 1.3315, + "theoretical_loss": 3.3832451061401057, + "tokens_seen": 2632187904 + }, + { + "epoch": 0.8, + "learning_rate": 0.00016370712136409227, + "loss": 1.3364, + "theoretical_loss": 3.383238512122132, + "tokens_seen": 2632253440 + }, + { + "epoch": 0.8, + "learning_rate": 0.00016369107321965898, + "loss": 1.3005, + "theoretical_loss": 3.3832319183142965, + "tokens_seen": 2632318976 + }, + { + "epoch": 0.8, + "learning_rate": 0.00016367502507522568, + "loss": 1.3905, + "theoretical_loss": 3.3832253247165873, + "tokens_seen": 2632384512 + }, + { + "epoch": 0.8, + "learning_rate": 0.0001636589769307924, + "loss": 1.2399, + "theoretical_loss": 3.383218731328993, + "tokens_seen": 2632450048 + }, + { + "epoch": 0.8, + "learning_rate": 0.0001636429287863591, + "loss": 1.3007, + "theoretical_loss": 3.3832121381515012, + "tokens_seen": 2632515584 + }, + { + "epoch": 0.8, + "learning_rate": 0.0001636268806419258, + "loss": 1.2611, + "theoretical_loss": 3.3832055451841, + "tokens_seen": 2632581120 + }, + { + "epoch": 0.8, + "learning_rate": 0.0001636108324974925, + "loss": 1.2232, + "theoretical_loss": 3.3831989524267776, + "tokens_seen": 2632646656 + }, + { + "epoch": 0.8, + "learning_rate": 0.0001635947843530592, + "loss": 1.2842, + "theoretical_loss": 3.383192359879522, + "tokens_seen": 2632712192 + }, + { + "epoch": 0.8, + "learning_rate": 0.00016357873620862588, + "loss": 1.4081, + "theoretical_loss": 3.3831857675423214, + "tokens_seen": 2632777728 + }, + { + "epoch": 0.8, + "learning_rate": 0.00016356268806419258, + "loss": 1.2538, + "theoretical_loss": 3.383179175415164, + "tokens_seen": 2632843264 + }, + { + "epoch": 0.8, + "learning_rate": 0.00016354663991975928, + "loss": 1.1213, + "theoretical_loss": 3.383172583498037, + "tokens_seen": 2632908800 + }, + { + "epoch": 0.8, + "learning_rate": 0.00016353059177532598, + "loss": 1.2503, + "theoretical_loss": 3.38316599179093, + "tokens_seen": 2632974336 + }, + { + "epoch": 0.8, + "learning_rate": 0.00016351454363089268, + "loss": 1.2473, + "theoretical_loss": 3.3831594002938297, + "tokens_seen": 2633039872 + }, + { + "epoch": 0.8, + "learning_rate": 0.00016349849548645938, + "loss": 1.2206, + "theoretical_loss": 3.3831528090067247, + "tokens_seen": 2633105408 + }, + { + "epoch": 0.8, + "learning_rate": 0.00016348244734202608, + "loss": 1.3126, + "theoretical_loss": 3.3831462179296032, + "tokens_seen": 2633170944 + }, + { + "epoch": 0.8, + "learning_rate": 0.00016346639919759278, + "loss": 1.1947, + "theoretical_loss": 3.3831396270624534, + "tokens_seen": 2633236480 + }, + { + "epoch": 0.8, + "learning_rate": 0.00016345035105315949, + "loss": 1.2649, + "theoretical_loss": 3.3831330364052628, + "tokens_seen": 2633302016 + }, + { + "epoch": 0.8, + "learning_rate": 0.00016343430290872619, + "loss": 1.3497, + "theoretical_loss": 3.38312644595802, + "tokens_seen": 2633367552 + }, + { + "epoch": 0.8, + "learning_rate": 0.00016341825476429289, + "loss": 1.2008, + "theoretical_loss": 3.383119855720713, + "tokens_seen": 2633433088 + }, + { + "epoch": 0.8, + "learning_rate": 0.0001634022066198596, + "loss": 1.2026, + "theoretical_loss": 3.38311326569333, + "tokens_seen": 2633498624 + }, + { + "epoch": 0.8, + "learning_rate": 0.0001633861584754263, + "loss": 1.1262, + "theoretical_loss": 3.3831066758758586, + "tokens_seen": 2633564160 + }, + { + "epoch": 0.8, + "learning_rate": 0.000163370110330993, + "loss": 1.1544, + "theoretical_loss": 3.3831000862682874, + "tokens_seen": 2633629696 + }, + { + "epoch": 0.8, + "learning_rate": 0.0001633540621865597, + "loss": 1.3047, + "theoretical_loss": 3.3830934968706043, + "tokens_seen": 2633695232 + }, + { + "epoch": 0.8, + "learning_rate": 0.0001633380140421264, + "loss": 1.2995, + "theoretical_loss": 3.3830869076827974, + "tokens_seen": 2633760768 + }, + { + "epoch": 0.8, + "learning_rate": 0.0001633219658976931, + "loss": 1.3201, + "theoretical_loss": 3.383080318704855, + "tokens_seen": 2633826304 + }, + { + "epoch": 0.8, + "learning_rate": 0.0001633059177532598, + "loss": 1.3009, + "theoretical_loss": 3.383073729936765, + "tokens_seen": 2633891840 + }, + { + "epoch": 0.8, + "learning_rate": 0.0001632898696088265, + "loss": 1.2187, + "theoretical_loss": 3.3830671413785147, + "tokens_seen": 2633957376 + }, + { + "epoch": 0.8, + "learning_rate": 0.0001632738214643932, + "loss": 1.2455, + "theoretical_loss": 3.3830605530300932, + "tokens_seen": 2634022912 + }, + { + "epoch": 0.8, + "learning_rate": 0.0001632577733199599, + "loss": 1.2316, + "theoretical_loss": 3.383053964891489, + "tokens_seen": 2634088448 + }, + { + "epoch": 0.8, + "learning_rate": 0.0001632417251755266, + "loss": 1.3783, + "theoretical_loss": 3.383047376962689, + "tokens_seen": 2634153984 + }, + { + "epoch": 0.8, + "learning_rate": 0.0001632256770310933, + "loss": 1.3137, + "theoretical_loss": 3.3830407892436822, + "tokens_seen": 2634219520 + }, + { + "epoch": 0.8, + "learning_rate": 0.00016320962888666, + "loss": 1.25, + "theoretical_loss": 3.3830342017344566, + "tokens_seen": 2634285056 + }, + { + "epoch": 0.8, + "learning_rate": 0.00016319358074222667, + "loss": 1.3049, + "theoretical_loss": 3.3830276144349996, + "tokens_seen": 2634350592 + }, + { + "epoch": 0.8, + "learning_rate": 0.00016317753259779337, + "loss": 1.2099, + "theoretical_loss": 3.3830210273452996, + "tokens_seen": 2634416128 + }, + { + "epoch": 0.8, + "learning_rate": 0.00016316148445336007, + "loss": 1.3167, + "theoretical_loss": 3.3830144404653453, + "tokens_seen": 2634481664 + }, + { + "epoch": 0.8, + "objective/train/docs_used": 957617, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.2366855144500732, + "objective/train/theoretical_loss": 3.3830078537951245, + "objective/train/tokens_used": 2655007200, + "theoretical_loss": 3.3830078537951245, + "tokens_seen": 2634547200 + }, + { + "epoch": 0.8, + "learning_rate": 0.00016314543630892677, + "loss": 1.1988, + "theoretical_loss": 3.3830078537951245, + "tokens_seen": 2634547200 + }, + { + "epoch": 0.8, + "learning_rate": 0.00016312938816449347, + "loss": 1.248, + "theoretical_loss": 3.383001267334625, + "tokens_seen": 2634612736 + }, + { + "epoch": 0.8, + "learning_rate": 0.00016311334002006017, + "loss": 1.4596, + "theoretical_loss": 3.3829946810838347, + "tokens_seen": 2634678272 + }, + { + "epoch": 0.8, + "learning_rate": 0.00016309729187562687, + "loss": 1.1803, + "theoretical_loss": 3.3829880950427427, + "tokens_seen": 2634743808 + }, + { + "epoch": 0.8, + "learning_rate": 0.00016308124373119357, + "loss": 1.4015, + "theoretical_loss": 3.3829815092113362, + "tokens_seen": 2634809344 + }, + { + "epoch": 0.8, + "learning_rate": 0.00016306519558676027, + "loss": 1.3384, + "theoretical_loss": 3.382974923589604, + "tokens_seen": 2634874880 + }, + { + "epoch": 0.8, + "learning_rate": 0.000163049147442327, + "loss": 1.2027, + "theoretical_loss": 3.382968338177533, + "tokens_seen": 2634940416 + }, + { + "epoch": 0.8, + "learning_rate": 0.0001630330992978937, + "loss": 1.2851, + "theoretical_loss": 3.382961752975113, + "tokens_seen": 2635005952 + }, + { + "epoch": 0.8, + "learning_rate": 0.0001630170511534604, + "loss": 1.3337, + "theoretical_loss": 3.3829551679823306, + "tokens_seen": 2635071488 + }, + { + "epoch": 0.8, + "learning_rate": 0.0001630010030090271, + "loss": 1.2549, + "theoretical_loss": 3.382948583199175, + "tokens_seen": 2635137024 + }, + { + "epoch": 0.8, + "learning_rate": 0.0001629849548645938, + "loss": 1.2559, + "theoretical_loss": 3.3829419986256335, + "tokens_seen": 2635202560 + }, + { + "epoch": 0.8, + "learning_rate": 0.0001629689067201605, + "loss": 1.2766, + "theoretical_loss": 3.382935414261695, + "tokens_seen": 2635268096 + }, + { + "epoch": 0.8, + "learning_rate": 0.0001629528585757272, + "loss": 1.2075, + "theoretical_loss": 3.382928830107347, + "tokens_seen": 2635333632 + }, + { + "epoch": 0.8, + "learning_rate": 0.0001629368104312939, + "loss": 1.2898, + "theoretical_loss": 3.3829222461625776, + "tokens_seen": 2635399168 + }, + { + "epoch": 0.8, + "learning_rate": 0.0001629207622868606, + "loss": 1.1724, + "theoretical_loss": 3.3829156624273753, + "tokens_seen": 2635464704 + }, + { + "epoch": 0.8, + "learning_rate": 0.0001629047141424273, + "loss": 1.3588, + "theoretical_loss": 3.3829090789017284, + "tokens_seen": 2635530240 + }, + { + "epoch": 0.8, + "learning_rate": 0.000162888665997994, + "loss": 1.3252, + "theoretical_loss": 3.382902495585624, + "tokens_seen": 2635595776 + }, + { + "epoch": 0.8, + "learning_rate": 0.00016287261785356068, + "loss": 1.2831, + "theoretical_loss": 3.3828959124790514, + "tokens_seen": 2635661312 + }, + { + "epoch": 0.8, + "learning_rate": 0.00016285656970912738, + "loss": 1.3255, + "theoretical_loss": 3.382889329581998, + "tokens_seen": 2635726848 + }, + { + "epoch": 0.8, + "learning_rate": 0.00016284052156469408, + "loss": 1.2941, + "theoretical_loss": 3.3828827468944525, + "tokens_seen": 2635792384 + }, + { + "epoch": 0.8, + "learning_rate": 0.00016282447342026078, + "loss": 1.3362, + "theoretical_loss": 3.3828761644164023, + "tokens_seen": 2635857920 + }, + { + "epoch": 0.8, + "learning_rate": 0.00016280842527582748, + "loss": 1.286, + "theoretical_loss": 3.382869582147836, + "tokens_seen": 2635923456 + }, + { + "epoch": 0.8, + "learning_rate": 0.00016279237713139418, + "loss": 1.2594, + "theoretical_loss": 3.3828630000887414, + "tokens_seen": 2635988992 + }, + { + "epoch": 0.8, + "learning_rate": 0.00016277632898696088, + "loss": 1.2137, + "theoretical_loss": 3.382856418239107, + "tokens_seen": 2636054528 + }, + { + "epoch": 0.8, + "learning_rate": 0.00016276028084252759, + "loss": 1.254, + "theoretical_loss": 3.382849836598921, + "tokens_seen": 2636120064 + }, + { + "epoch": 0.8, + "learning_rate": 0.00016274423269809429, + "loss": 1.2019, + "theoretical_loss": 3.382843255168171, + "tokens_seen": 2636185600 + }, + { + "epoch": 0.8, + "learning_rate": 0.000162728184553661, + "loss": 1.1784, + "theoretical_loss": 3.382836673946845, + "tokens_seen": 2636251136 + }, + { + "epoch": 0.8, + "learning_rate": 0.0001627121364092277, + "loss": 1.2539, + "theoretical_loss": 3.3828300929349324, + "tokens_seen": 2636316672 + }, + { + "epoch": 0.8, + "learning_rate": 0.0001626960882647944, + "loss": 1.2941, + "theoretical_loss": 3.38282351213242, + "tokens_seen": 2636382208 + }, + { + "epoch": 0.8, + "learning_rate": 0.0001626800401203611, + "loss": 1.2394, + "theoretical_loss": 3.3828169315392964, + "tokens_seen": 2636447744 + }, + { + "epoch": 0.8, + "learning_rate": 0.0001626639919759278, + "loss": 1.2609, + "theoretical_loss": 3.38281035115555, + "tokens_seen": 2636513280 + }, + { + "epoch": 0.8, + "learning_rate": 0.0001626479438314945, + "loss": 1.2612, + "theoretical_loss": 3.3828037709811687, + "tokens_seen": 2636578816 + }, + { + "epoch": 0.8, + "learning_rate": 0.0001626318956870612, + "loss": 1.4179, + "theoretical_loss": 3.3827971910161407, + "tokens_seen": 2636644352 + }, + { + "epoch": 0.8, + "learning_rate": 0.0001626158475426279, + "loss": 1.3497, + "theoretical_loss": 3.3827906112604538, + "tokens_seen": 2636709888 + }, + { + "epoch": 0.8, + "learning_rate": 0.0001625997993981946, + "loss": 1.2843, + "theoretical_loss": 3.3827840317140963, + "tokens_seen": 2636775424 + }, + { + "epoch": 0.8, + "learning_rate": 0.0001625837512537613, + "loss": 1.2148, + "theoretical_loss": 3.3827774523770566, + "tokens_seen": 2636840960 + }, + { + "epoch": 0.8, + "learning_rate": 0.000162567703109328, + "loss": 1.3406, + "theoretical_loss": 3.3827708732493225, + "tokens_seen": 2636906496 + }, + { + "epoch": 0.8, + "learning_rate": 0.0001625516549648947, + "loss": 1.2976, + "theoretical_loss": 3.382764294330882, + "tokens_seen": 2636972032 + }, + { + "epoch": 0.8, + "learning_rate": 0.0001625356068204614, + "loss": 1.2686, + "theoretical_loss": 3.3827577156217243, + "tokens_seen": 2637037568 + }, + { + "epoch": 0.8, + "learning_rate": 0.0001625195586760281, + "loss": 1.314, + "theoretical_loss": 3.3827511371218364, + "tokens_seen": 2637103104 + }, + { + "epoch": 0.8, + "learning_rate": 0.0001625035105315948, + "loss": 1.2278, + "theoretical_loss": 3.382744558831207, + "tokens_seen": 2637168640 + }, + { + "epoch": 0.8, + "learning_rate": 0.00016248746238716147, + "loss": 1.3049, + "theoretical_loss": 3.382737980749824, + "tokens_seen": 2637234176 + }, + { + "epoch": 0.8, + "learning_rate": 0.00016247141424272817, + "loss": 1.2999, + "theoretical_loss": 3.3827314028776754, + "tokens_seen": 2637299712 + }, + { + "epoch": 0.8, + "learning_rate": 0.0001624553660982949, + "loss": 1.1636, + "theoretical_loss": 3.38272482521475, + "tokens_seen": 2637365248 + }, + { + "epoch": 0.8, + "learning_rate": 0.0001624393179538616, + "loss": 1.2705, + "theoretical_loss": 3.382718247761035, + "tokens_seen": 2637430784 + }, + { + "epoch": 0.8, + "learning_rate": 0.0001624232698094283, + "loss": 1.2898, + "theoretical_loss": 3.382711670516519, + "tokens_seen": 2637496320 + }, + { + "epoch": 0.8, + "learning_rate": 0.000162407221664995, + "loss": 1.3398, + "theoretical_loss": 3.382705093481191, + "tokens_seen": 2637561856 + }, + { + "epoch": 0.8, + "learning_rate": 0.0001623911735205617, + "loss": 1.2404, + "theoretical_loss": 3.3826985166550374, + "tokens_seen": 2637627392 + }, + { + "epoch": 0.8, + "learning_rate": 0.0001623751253761284, + "loss": 1.287, + "theoretical_loss": 3.3826919400380477, + "tokens_seen": 2637692928 + }, + { + "epoch": 0.8, + "learning_rate": 0.0001623590772316951, + "loss": 1.1866, + "theoretical_loss": 3.38268536363021, + "tokens_seen": 2637758464 + }, + { + "epoch": 0.8, + "objective/train/docs_used": 958785, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.2947386503219604, + "objective/train/theoretical_loss": 3.382678787431512, + "objective/train/tokens_used": 2658284000, + "theoretical_loss": 3.382678787431512, + "tokens_seen": 2637824000 + }, + { + "epoch": 0.8, + "learning_rate": 0.0001623430290872618, + "loss": 1.3, + "theoretical_loss": 3.382678787431512, + "tokens_seen": 2637824000 + }, + { + "epoch": 0.8, + "learning_rate": 0.0001623269809428285, + "loss": 1.1431, + "theoretical_loss": 3.3826722114419416, + "tokens_seen": 2637889536 + }, + { + "epoch": 0.8, + "learning_rate": 0.0001623109327983952, + "loss": 1.206, + "theoretical_loss": 3.3826656356614873, + "tokens_seen": 2637955072 + }, + { + "epoch": 0.8, + "learning_rate": 0.0001622948846539619, + "loss": 1.2709, + "theoretical_loss": 3.3826590600901376, + "tokens_seen": 2638020608 + }, + { + "epoch": 0.8, + "learning_rate": 0.0001622788365095286, + "loss": 1.3034, + "theoretical_loss": 3.3826524847278803, + "tokens_seen": 2638086144 + }, + { + "epoch": 0.8, + "learning_rate": 0.0001622627883650953, + "loss": 1.3748, + "theoretical_loss": 3.3826459095747037, + "tokens_seen": 2638151680 + }, + { + "epoch": 0.8, + "learning_rate": 0.000162246740220662, + "loss": 1.2883, + "theoretical_loss": 3.3826393346305954, + "tokens_seen": 2638217216 + }, + { + "epoch": 0.8, + "learning_rate": 0.0001622306920762287, + "loss": 1.2727, + "theoretical_loss": 3.3826327598955443, + "tokens_seen": 2638282752 + }, + { + "epoch": 0.8, + "learning_rate": 0.0001622146439317954, + "loss": 1.2152, + "theoretical_loss": 3.382626185369538, + "tokens_seen": 2638348288 + }, + { + "epoch": 0.8, + "learning_rate": 0.0001621985957873621, + "loss": 1.3463, + "theoretical_loss": 3.382619611052565, + "tokens_seen": 2638413824 + }, + { + "epoch": 0.8, + "learning_rate": 0.0001621825476429288, + "loss": 1.2021, + "theoretical_loss": 3.3826130369446137, + "tokens_seen": 2638479360 + }, + { + "epoch": 0.8, + "learning_rate": 0.00016216649949849548, + "loss": 1.2773, + "theoretical_loss": 3.382606463045672, + "tokens_seen": 2638544896 + }, + { + "epoch": 0.8, + "learning_rate": 0.00016215045135406218, + "loss": 1.229, + "theoretical_loss": 3.3825998893557276, + "tokens_seen": 2638610432 + }, + { + "epoch": 0.8, + "learning_rate": 0.00016213440320962888, + "loss": 1.2841, + "theoretical_loss": 3.382593315874769, + "tokens_seen": 2638675968 + }, + { + "epoch": 0.8, + "learning_rate": 0.00016211835506519558, + "loss": 1.2815, + "theoretical_loss": 3.382586742602785, + "tokens_seen": 2638741504 + }, + { + "epoch": 0.8, + "learning_rate": 0.00016210230692076228, + "loss": 1.2268, + "theoretical_loss": 3.3825801695397626, + "tokens_seen": 2638807040 + }, + { + "epoch": 0.8, + "learning_rate": 0.00016208625877632898, + "loss": 1.2322, + "theoretical_loss": 3.382573596685691, + "tokens_seen": 2638872576 + }, + { + "epoch": 0.8, + "learning_rate": 0.00016207021063189569, + "loss": 1.3466, + "theoretical_loss": 3.382567024040558, + "tokens_seen": 2638938112 + }, + { + "epoch": 0.8, + "learning_rate": 0.00016207021063189569, + "loss": 1.3195, + "theoretical_loss": 3.3825604516043515, + "tokens_seen": 2639003648 + }, + { + "epoch": 0.8, + "learning_rate": 0.00016205416248746239, + "loss": 1.2393, + "theoretical_loss": 3.3825538793770598, + "tokens_seen": 2639069184 + }, + { + "epoch": 0.8, + "learning_rate": 0.0001620381143430291, + "loss": 1.3339, + "theoretical_loss": 3.3825473073586716, + "tokens_seen": 2639134720 + }, + { + "epoch": 0.8, + "learning_rate": 0.0001620220661985958, + "loss": 1.322, + "theoretical_loss": 3.382540735549174, + "tokens_seen": 2639200256 + }, + { + "epoch": 0.8, + "learning_rate": 0.0001620060180541625, + "loss": 1.3117, + "theoretical_loss": 3.382534163948556, + "tokens_seen": 2639265792 + }, + { + "epoch": 0.8, + "learning_rate": 0.0001619899699097292, + "loss": 1.271, + "theoretical_loss": 3.382527592556806, + "tokens_seen": 2639331328 + }, + { + "epoch": 0.8, + "learning_rate": 0.0001619739217652959, + "loss": 1.2827, + "theoretical_loss": 3.382521021373911, + "tokens_seen": 2639396864 + }, + { + "epoch": 0.8, + "learning_rate": 0.0001619578736208626, + "loss": 1.2756, + "theoretical_loss": 3.382514450399861, + "tokens_seen": 2639462400 + }, + { + "epoch": 0.8, + "learning_rate": 0.0001619418254764293, + "loss": 1.247, + "theoretical_loss": 3.3825078796346424, + "tokens_seen": 2639527936 + }, + { + "epoch": 0.8, + "learning_rate": 0.000161925777331996, + "loss": 1.2934, + "theoretical_loss": 3.382501309078244, + "tokens_seen": 2639593472 + }, + { + "epoch": 0.8, + "learning_rate": 0.0001619097291875627, + "loss": 1.2337, + "theoretical_loss": 3.382494738730654, + "tokens_seen": 2639659008 + }, + { + "epoch": 0.8, + "learning_rate": 0.0001618936810431294, + "loss": 1.3042, + "theoretical_loss": 3.382488168591861, + "tokens_seen": 2639724544 + }, + { + "epoch": 0.8, + "learning_rate": 0.0001618776328986961, + "loss": 1.1736, + "theoretical_loss": 3.382481598661853, + "tokens_seen": 2639790080 + }, + { + "epoch": 0.8, + "learning_rate": 0.0001618615847542628, + "loss": 1.3069, + "theoretical_loss": 3.3824750289406174, + "tokens_seen": 2639855616 + }, + { + "epoch": 0.8, + "learning_rate": 0.0001618455366098295, + "loss": 1.1745, + "theoretical_loss": 3.3824684594281433, + "tokens_seen": 2639921152 + }, + { + "epoch": 0.8, + "learning_rate": 0.0001618294884653962, + "loss": 1.2284, + "theoretical_loss": 3.382461890124419, + "tokens_seen": 2639986688 + }, + { + "epoch": 0.8, + "learning_rate": 0.0001618134403209629, + "loss": 1.2609, + "theoretical_loss": 3.3824553210294317, + "tokens_seen": 2640052224 + }, + { + "epoch": 0.8, + "learning_rate": 0.0001617973921765296, + "loss": 1.196, + "theoretical_loss": 3.38244875214317, + "tokens_seen": 2640117760 + }, + { + "epoch": 0.8, + "learning_rate": 0.0001617813440320963, + "loss": 1.2588, + "theoretical_loss": 3.382442183465623, + "tokens_seen": 2640183296 + }, + { + "epoch": 0.8, + "learning_rate": 0.000161765295887663, + "loss": 1.2749, + "theoretical_loss": 3.382435614996777, + "tokens_seen": 2640248832 + }, + { + "epoch": 0.8, + "learning_rate": 0.0001617492477432297, + "loss": 1.3015, + "theoretical_loss": 3.3824290467366223, + "tokens_seen": 2640314368 + }, + { + "epoch": 0.8, + "learning_rate": 0.0001617331995987964, + "loss": 1.3046, + "theoretical_loss": 3.382422478685146, + "tokens_seen": 2640379904 + }, + { + "epoch": 0.8, + "learning_rate": 0.0001617171514543631, + "loss": 1.2672, + "theoretical_loss": 3.3824159108423357, + "tokens_seen": 2640445440 + }, + { + "epoch": 0.8, + "learning_rate": 0.0001617011033099298, + "loss": 1.3173, + "theoretical_loss": 3.382409343208181, + "tokens_seen": 2640510976 + }, + { + "epoch": 0.8, + "learning_rate": 0.0001616850551654965, + "loss": 1.3002, + "theoretical_loss": 3.382402775782669, + "tokens_seen": 2640576512 + }, + { + "epoch": 0.8, + "learning_rate": 0.0001616690070210632, + "loss": 1.1976, + "theoretical_loss": 3.3823962085657886, + "tokens_seen": 2640642048 + }, + { + "epoch": 0.8, + "learning_rate": 0.0001616529588766299, + "loss": 1.2962, + "theoretical_loss": 3.382389641557527, + "tokens_seen": 2640707584 + }, + { + "epoch": 0.8, + "learning_rate": 0.0001616369107321966, + "loss": 1.2598, + "theoretical_loss": 3.3823830747578736, + "tokens_seen": 2640773120 + }, + { + "epoch": 0.8, + "learning_rate": 0.0001616208625877633, + "loss": 1.2492, + "theoretical_loss": 3.382376508166816, + "tokens_seen": 2640838656 + }, + { + "epoch": 0.8, + "learning_rate": 0.00016160481444333, + "loss": 1.177, + "theoretical_loss": 3.3823699417843427, + "tokens_seen": 2640904192 + }, + { + "epoch": 0.8, + "learning_rate": 0.0001615887662988967, + "loss": 1.1501, + "theoretical_loss": 3.382363375610441, + "tokens_seen": 2640969728 + }, + { + "epoch": 0.8, + "learning_rate": 0.0001615727181544634, + "loss": 1.2519, + "theoretical_loss": 3.3823568096451004, + "tokens_seen": 2641035264 + }, + { + "epoch": 0.8, + "objective/train/docs_used": 959716, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.329784631729126, + "objective/train/theoretical_loss": 3.3823502438883084, + "objective/train/tokens_used": 2661560800, + "theoretical_loss": 3.3823502438883084, + "tokens_seen": 2641100800 + }, + { + "epoch": 0.8, + "learning_rate": 0.0001615566700100301, + "loss": 1.2451, + "theoretical_loss": 3.3823502438883084, + "tokens_seen": 2641100800 + }, + { + "epoch": 0.8, + "learning_rate": 0.0001615406218655968, + "loss": 1.311, + "theoretical_loss": 3.3823436783400527, + "tokens_seen": 2641166336 + }, + { + "epoch": 0.8, + "learning_rate": 0.0001615245737211635, + "loss": 1.3067, + "theoretical_loss": 3.3823371130003226, + "tokens_seen": 2641231872 + }, + { + "epoch": 0.8, + "learning_rate": 0.0001615085255767302, + "loss": 1.3046, + "theoretical_loss": 3.3823305478691053, + "tokens_seen": 2641297408 + }, + { + "epoch": 0.8, + "learning_rate": 0.0001614924774322969, + "loss": 1.2815, + "theoretical_loss": 3.3823239829463896, + "tokens_seen": 2641362944 + }, + { + "epoch": 0.8, + "learning_rate": 0.00016147642928786358, + "loss": 1.2369, + "theoretical_loss": 3.382317418232164, + "tokens_seen": 2641428480 + }, + { + "epoch": 0.8, + "learning_rate": 0.00016146038114343028, + "loss": 1.3609, + "theoretical_loss": 3.3823108537264157, + "tokens_seen": 2641494016 + }, + { + "epoch": 0.8, + "learning_rate": 0.00016144433299899698, + "loss": 1.1543, + "theoretical_loss": 3.382304289429134, + "tokens_seen": 2641559552 + }, + { + "epoch": 0.8, + "learning_rate": 0.00016142828485456368, + "loss": 1.2593, + "theoretical_loss": 3.3822977253403064, + "tokens_seen": 2641625088 + }, + { + "epoch": 0.8, + "learning_rate": 0.00016141223671013038, + "loss": 1.2439, + "theoretical_loss": 3.382291161459921, + "tokens_seen": 2641690624 + }, + { + "epoch": 0.8, + "learning_rate": 0.00016139618856569708, + "loss": 1.2534, + "theoretical_loss": 3.3822845977879665, + "tokens_seen": 2641756160 + }, + { + "epoch": 0.8, + "learning_rate": 0.00016138014042126379, + "loss": 1.2308, + "theoretical_loss": 3.382278034324431, + "tokens_seen": 2641821696 + }, + { + "epoch": 0.8, + "learning_rate": 0.00016136409227683049, + "loss": 1.2, + "theoretical_loss": 3.3822714710693025, + "tokens_seen": 2641887232 + }, + { + "epoch": 0.8, + "learning_rate": 0.0001613480441323972, + "loss": 1.3312, + "theoretical_loss": 3.3822649080225693, + "tokens_seen": 2641952768 + }, + { + "epoch": 0.8, + "learning_rate": 0.0001613319959879639, + "loss": 1.2788, + "theoretical_loss": 3.38225834518422, + "tokens_seen": 2642018304 + }, + { + "epoch": 0.8, + "learning_rate": 0.0001613159478435306, + "loss": 1.3265, + "theoretical_loss": 3.3822517825542424, + "tokens_seen": 2642083840 + }, + { + "epoch": 0.8, + "learning_rate": 0.0001612998996990973, + "loss": 1.3157, + "theoretical_loss": 3.3822452201326247, + "tokens_seen": 2642149376 + }, + { + "epoch": 0.8, + "learning_rate": 0.000161283851554664, + "loss": 1.3106, + "theoretical_loss": 3.382238657919355, + "tokens_seen": 2642214912 + }, + { + "epoch": 0.8, + "learning_rate": 0.0001612678034102307, + "loss": 1.2191, + "theoretical_loss": 3.3822320959144223, + "tokens_seen": 2642280448 + }, + { + "epoch": 0.8, + "learning_rate": 0.00016125175526579742, + "loss": 1.2654, + "theoretical_loss": 3.3822255341178136, + "tokens_seen": 2642345984 + }, + { + "epoch": 0.8, + "learning_rate": 0.00016123570712136412, + "loss": 1.2189, + "theoretical_loss": 3.382218972529518, + "tokens_seen": 2642411520 + }, + { + "epoch": 0.8, + "learning_rate": 0.00016121965897693082, + "loss": 1.2813, + "theoretical_loss": 3.3822124111495233, + "tokens_seen": 2642477056 + }, + { + "epoch": 0.8, + "learning_rate": 0.00016120361083249752, + "loss": 1.2092, + "theoretical_loss": 3.3822058499778183, + "tokens_seen": 2642542592 + }, + { + "epoch": 0.8, + "learning_rate": 0.00016118756268806422, + "loss": 1.0883, + "theoretical_loss": 3.3821992890143906, + "tokens_seen": 2642608128 + }, + { + "epoch": 0.8, + "learning_rate": 0.00016117151454363092, + "loss": 1.2549, + "theoretical_loss": 3.382192728259229, + "tokens_seen": 2642673664 + }, + { + "epoch": 0.8, + "learning_rate": 0.0001611554663991976, + "loss": 1.2193, + "theoretical_loss": 3.3821861677123204, + "tokens_seen": 2642739200 + }, + { + "epoch": 0.8, + "learning_rate": 0.0001611394182547643, + "loss": 1.063, + "theoretical_loss": 3.382179607373655, + "tokens_seen": 2642804736 + }, + { + "epoch": 0.8, + "learning_rate": 0.000161123370110331, + "loss": 1.3265, + "theoretical_loss": 3.38217304724322, + "tokens_seen": 2642870272 + }, + { + "epoch": 0.8, + "learning_rate": 0.0001611073219658977, + "loss": 1.1789, + "theoretical_loss": 3.382166487321003, + "tokens_seen": 2642935808 + }, + { + "epoch": 0.8, + "learning_rate": 0.0001610912738214644, + "loss": 1.2838, + "theoretical_loss": 3.382159927606993, + "tokens_seen": 2643001344 + }, + { + "epoch": 0.8, + "learning_rate": 0.0001610752256770311, + "loss": 1.3445, + "theoretical_loss": 3.3821533681011786, + "tokens_seen": 2643066880 + }, + { + "epoch": 0.8, + "learning_rate": 0.0001610591775325978, + "loss": 1.201, + "theoretical_loss": 3.3821468088035473, + "tokens_seen": 2643132416 + }, + { + "epoch": 0.8, + "learning_rate": 0.0001610431293881645, + "loss": 1.357, + "theoretical_loss": 3.3821402497140873, + "tokens_seen": 2643197952 + }, + { + "epoch": 0.8, + "learning_rate": 0.0001610270812437312, + "loss": 1.1884, + "theoretical_loss": 3.3821336908327875, + "tokens_seen": 2643263488 + }, + { + "epoch": 0.8, + "learning_rate": 0.0001610110330992979, + "loss": 1.3134, + "theoretical_loss": 3.382127132159636, + "tokens_seen": 2643329024 + }, + { + "epoch": 0.8, + "learning_rate": 0.0001609949849548646, + "loss": 1.2456, + "theoretical_loss": 3.38212057369462, + "tokens_seen": 2643394560 + }, + { + "epoch": 0.8, + "learning_rate": 0.0001609789368104313, + "loss": 1.2225, + "theoretical_loss": 3.382114015437729, + "tokens_seen": 2643460096 + }, + { + "epoch": 0.8, + "learning_rate": 0.000160962888665998, + "loss": 1.2504, + "theoretical_loss": 3.3821074573889502, + "tokens_seen": 2643525632 + }, + { + "epoch": 0.8, + "learning_rate": 0.0001609468405215647, + "loss": 1.2084, + "theoretical_loss": 3.382100899548273, + "tokens_seen": 2643591168 + }, + { + "epoch": 0.8, + "learning_rate": 0.0001609307923771314, + "loss": 1.1765, + "theoretical_loss": 3.382094341915685, + "tokens_seen": 2643656704 + }, + { + "epoch": 0.8, + "learning_rate": 0.0001609147442326981, + "loss": 1.2074, + "theoretical_loss": 3.382087784491174, + "tokens_seen": 2643722240 + }, + { + "epoch": 0.8, + "learning_rate": 0.0001608986960882648, + "loss": 1.3468, + "theoretical_loss": 3.382081227274729, + "tokens_seen": 2643787776 + }, + { + "epoch": 0.8, + "learning_rate": 0.0001608826479438315, + "loss": 1.2568, + "theoretical_loss": 3.382074670266338, + "tokens_seen": 2643853312 + }, + { + "epoch": 0.8, + "learning_rate": 0.0001608665997993982, + "loss": 1.1722, + "theoretical_loss": 3.3820681134659893, + "tokens_seen": 2643918848 + }, + { + "epoch": 0.8, + "learning_rate": 0.0001608505516549649, + "loss": 1.2877, + "theoretical_loss": 3.3820615568736705, + "tokens_seen": 2643984384 + }, + { + "epoch": 0.8, + "learning_rate": 0.0001608345035105316, + "loss": 1.25, + "theoretical_loss": 3.382055000489371, + "tokens_seen": 2644049920 + }, + { + "epoch": 0.8, + "learning_rate": 0.0001608184553660983, + "loss": 1.3764, + "theoretical_loss": 3.382048444313078, + "tokens_seen": 2644115456 + }, + { + "epoch": 0.8, + "learning_rate": 0.000160802407221665, + "loss": 1.2343, + "theoretical_loss": 3.38204188834478, + "tokens_seen": 2644180992 + }, + { + "epoch": 0.8, + "learning_rate": 0.0001607863590772317, + "loss": 1.3084, + "theoretical_loss": 3.382035332584466, + "tokens_seen": 2644246528 + }, + { + "epoch": 0.8, + "learning_rate": 0.00016077031093279838, + "loss": 1.2358, + "theoretical_loss": 3.382028777032123, + "tokens_seen": 2644312064 + }, + { + "epoch": 0.8, + "objective/train/docs_used": 960592, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.2820556163787842, + "objective/train/theoretical_loss": 3.3820222216877402, + "objective/train/tokens_used": 2664837600, + "theoretical_loss": 3.3820222216877402, + "tokens_seen": 2644377600 + }, + { + "epoch": 0.8, + "learning_rate": 0.00016075426278836508, + "loss": 1.2188, + "theoretical_loss": 3.3820222216877402, + "tokens_seen": 2644377600 + }, + { + "epoch": 0.8, + "learning_rate": 0.00016073821464393178, + "loss": 1.281, + "theoretical_loss": 3.3820156665513057, + "tokens_seen": 2644443136 + }, + { + "epoch": 0.8, + "learning_rate": 0.00016072216649949848, + "loss": 1.2469, + "theoretical_loss": 3.3820091116228075, + "tokens_seen": 2644508672 + }, + { + "epoch": 0.8, + "learning_rate": 0.00016070611835506519, + "loss": 1.2399, + "theoretical_loss": 3.3820025569022336, + "tokens_seen": 2644574208 + }, + { + "epoch": 0.8, + "learning_rate": 0.00016069007021063189, + "loss": 1.2828, + "theoretical_loss": 3.381996002389573, + "tokens_seen": 2644639744 + }, + { + "epoch": 0.8, + "learning_rate": 0.00016067402206619859, + "loss": 1.326, + "theoretical_loss": 3.3819894480848136, + "tokens_seen": 2644705280 + }, + { + "epoch": 0.8, + "learning_rate": 0.00016065797392176531, + "loss": 1.2774, + "theoretical_loss": 3.381982893987943, + "tokens_seen": 2644770816 + }, + { + "epoch": 0.8, + "learning_rate": 0.00016064192577733201, + "loss": 1.3158, + "theoretical_loss": 3.3819763400989507, + "tokens_seen": 2644836352 + }, + { + "epoch": 0.8, + "learning_rate": 0.00016062587763289872, + "loss": 1.3021, + "theoretical_loss": 3.381969786417824, + "tokens_seen": 2644901888 + }, + { + "epoch": 0.8, + "learning_rate": 0.00016060982948846542, + "loss": 1.2253, + "theoretical_loss": 3.381963232944552, + "tokens_seen": 2644967424 + }, + { + "epoch": 0.8, + "learning_rate": 0.00016059378134403212, + "loss": 1.3424, + "theoretical_loss": 3.3819566796791216, + "tokens_seen": 2645032960 + }, + { + "epoch": 0.8, + "learning_rate": 0.00016057773319959882, + "loss": 1.3586, + "theoretical_loss": 3.381950126621523, + "tokens_seen": 2645098496 + }, + { + "epoch": 0.8, + "learning_rate": 0.00016056168505516552, + "loss": 1.2128, + "theoretical_loss": 3.3819435737717423, + "tokens_seen": 2645164032 + }, + { + "epoch": 0.8, + "learning_rate": 0.00016054563691073222, + "loss": 1.3494, + "theoretical_loss": 3.3819370211297692, + "tokens_seen": 2645229568 + }, + { + "epoch": 0.8, + "learning_rate": 0.00016052958876629892, + "loss": 1.2544, + "theoretical_loss": 3.3819304686955913, + "tokens_seen": 2645295104 + }, + { + "epoch": 0.8, + "learning_rate": 0.00016051354062186562, + "loss": 1.3197, + "theoretical_loss": 3.381923916469198, + "tokens_seen": 2645360640 + }, + { + "epoch": 0.8, + "learning_rate": 0.00016049749247743232, + "loss": 1.2413, + "theoretical_loss": 3.381917364450576, + "tokens_seen": 2645426176 + }, + { + "epoch": 0.8, + "learning_rate": 0.00016048144433299902, + "loss": 1.2929, + "theoretical_loss": 3.381910812639714, + "tokens_seen": 2645491712 + }, + { + "epoch": 0.8, + "learning_rate": 0.00016046539618856572, + "loss": 1.2878, + "theoretical_loss": 3.381904261036601, + "tokens_seen": 2645557248 + }, + { + "epoch": 0.8, + "learning_rate": 0.0001604493480441324, + "loss": 1.207, + "theoretical_loss": 3.3818977096412244, + "tokens_seen": 2645622784 + }, + { + "epoch": 0.8, + "learning_rate": 0.0001604332998996991, + "loss": 1.3429, + "theoretical_loss": 3.3818911584535734, + "tokens_seen": 2645688320 + }, + { + "epoch": 0.8, + "learning_rate": 0.0001604172517552658, + "loss": 1.2748, + "theoretical_loss": 3.3818846074736357, + "tokens_seen": 2645753856 + }, + { + "epoch": 0.8, + "learning_rate": 0.0001604012036108325, + "loss": 1.2678, + "theoretical_loss": 3.381878056701399, + "tokens_seen": 2645819392 + }, + { + "epoch": 0.8, + "learning_rate": 0.0001603851554663992, + "loss": 1.2796, + "theoretical_loss": 3.3818715061368527, + "tokens_seen": 2645884928 + }, + { + "epoch": 0.8, + "learning_rate": 0.0001603691073219659, + "loss": 1.246, + "theoretical_loss": 3.3818649557799847, + "tokens_seen": 2645950464 + }, + { + "epoch": 0.8, + "learning_rate": 0.0001603530591775326, + "loss": 1.2259, + "theoretical_loss": 3.3818584056307825, + "tokens_seen": 2646016000 + }, + { + "epoch": 0.8, + "learning_rate": 0.0001603370110330993, + "loss": 1.173, + "theoretical_loss": 3.3818518556892356, + "tokens_seen": 2646081536 + }, + { + "epoch": 0.8, + "learning_rate": 0.000160320962888666, + "loss": 1.2957, + "theoretical_loss": 3.3818453059553315, + "tokens_seen": 2646147072 + }, + { + "epoch": 0.8, + "learning_rate": 0.0001603049147442327, + "loss": 1.0962, + "theoretical_loss": 3.381838756429058, + "tokens_seen": 2646212608 + }, + { + "epoch": 0.8, + "learning_rate": 0.0001602888665997994, + "loss": 1.2199, + "theoretical_loss": 3.3818322071104046, + "tokens_seen": 2646278144 + }, + { + "epoch": 0.8, + "learning_rate": 0.0001602728184553661, + "loss": 1.2457, + "theoretical_loss": 3.381825657999359, + "tokens_seen": 2646343680 + }, + { + "epoch": 0.8, + "learning_rate": 0.0001602567703109328, + "loss": 1.23, + "theoretical_loss": 3.3818191090959093, + "tokens_seen": 2646409216 + }, + { + "epoch": 0.8, + "learning_rate": 0.0001602407221664995, + "loss": 1.2239, + "theoretical_loss": 3.381812560400044, + "tokens_seen": 2646474752 + }, + { + "epoch": 0.8, + "learning_rate": 0.0001602246740220662, + "loss": 1.268, + "theoretical_loss": 3.3818060119117517, + "tokens_seen": 2646540288 + }, + { + "epoch": 0.8, + "learning_rate": 0.0001602086258776329, + "loss": 1.1094, + "theoretical_loss": 3.38179946363102, + "tokens_seen": 2646605824 + }, + { + "epoch": 0.8, + "learning_rate": 0.0001601925777331996, + "loss": 1.2918, + "theoretical_loss": 3.3817929155578375, + "tokens_seen": 2646671360 + }, + { + "epoch": 0.8, + "learning_rate": 0.0001601765295887663, + "loss": 1.221, + "theoretical_loss": 3.381786367692192, + "tokens_seen": 2646736896 + }, + { + "epoch": 0.8, + "learning_rate": 0.000160160481444333, + "loss": 1.2751, + "theoretical_loss": 3.381779820034073, + "tokens_seen": 2646802432 + }, + { + "epoch": 0.8, + "learning_rate": 0.0001601444332998997, + "loss": 1.2673, + "theoretical_loss": 3.3817732725834677, + "tokens_seen": 2646867968 + }, + { + "epoch": 0.8, + "learning_rate": 0.0001601283851554664, + "loss": 1.2629, + "theoretical_loss": 3.381766725340365, + "tokens_seen": 2646933504 + }, + { + "epoch": 0.8, + "learning_rate": 0.0001601123370110331, + "loss": 1.2415, + "theoretical_loss": 3.3817601783047526, + "tokens_seen": 2646999040 + }, + { + "epoch": 0.8, + "learning_rate": 0.0001600962888665998, + "loss": 1.3127, + "theoretical_loss": 3.3817536314766192, + "tokens_seen": 2647064576 + }, + { + "epoch": 0.8, + "learning_rate": 0.0001600802407221665, + "loss": 1.2054, + "theoretical_loss": 3.3817470848559528, + "tokens_seen": 2647130112 + }, + { + "epoch": 0.8, + "learning_rate": 0.00016006419257773318, + "loss": 1.2509, + "theoretical_loss": 3.381740538442742, + "tokens_seen": 2647195648 + }, + { + "epoch": 0.8, + "learning_rate": 0.0001600481444332999, + "loss": 1.1677, + "theoretical_loss": 3.381733992236975, + "tokens_seen": 2647261184 + }, + { + "epoch": 0.8, + "learning_rate": 0.0001600320962888666, + "loss": 1.2942, + "theoretical_loss": 3.3817274462386404, + "tokens_seen": 2647326720 + }, + { + "epoch": 0.8, + "learning_rate": 0.0001600160481444333, + "loss": 1.2634, + "theoretical_loss": 3.3817209004477258, + "tokens_seen": 2647392256 + }, + { + "epoch": 0.8, + "learning_rate": 0.00016, + "loss": 1.154, + "theoretical_loss": 3.3817143548642195, + "tokens_seen": 2647457792 + }, + { + "epoch": 0.8, + "learning_rate": 0.00015998395185556671, + "loss": 1.2146, + "theoretical_loss": 3.381707809488111, + "tokens_seen": 2647523328 + }, + { + "epoch": 0.8, + "learning_rate": 0.00015996790371113341, + "loss": 1.2114, + "theoretical_loss": 3.381701264319387, + "tokens_seen": 2647588864 + }, + { + "epoch": 0.8, + "objective/train/docs_used": 961712, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.3633036613464355, + "objective/train/theoretical_loss": 3.3816947193580367, + "objective/train/tokens_used": 2668114400, + "theoretical_loss": 3.3816947193580367, + "tokens_seen": 2647654400 + }, + { + "epoch": 0.8, + "learning_rate": 0.00015995185556670011, + "loss": 1.458, + "theoretical_loss": 3.3816947193580367, + "tokens_seen": 2647654400 + }, + { + "epoch": 0.8, + "learning_rate": 0.00015993580742226682, + "loss": 1.2503, + "theoretical_loss": 3.381688174604048, + "tokens_seen": 2647719936 + }, + { + "epoch": 0.8, + "learning_rate": 0.00015991975927783352, + "loss": 1.2357, + "theoretical_loss": 3.38168163005741, + "tokens_seen": 2647785472 + }, + { + "epoch": 0.8, + "learning_rate": 0.00015990371113340022, + "loss": 1.1758, + "theoretical_loss": 3.3816750857181104, + "tokens_seen": 2647851008 + }, + { + "epoch": 0.8, + "learning_rate": 0.00015988766298896692, + "loss": 1.2907, + "theoretical_loss": 3.381668541586137, + "tokens_seen": 2647916544 + }, + { + "epoch": 0.8, + "learning_rate": 0.00015987161484453362, + "loss": 1.1819, + "theoretical_loss": 3.381661997661479, + "tokens_seen": 2647982080 + }, + { + "epoch": 0.8, + "learning_rate": 0.00015985556670010032, + "loss": 1.2836, + "theoretical_loss": 3.3816554539441244, + "tokens_seen": 2648047616 + }, + { + "epoch": 0.8, + "learning_rate": 0.00015983951855566702, + "loss": 1.2342, + "theoretical_loss": 3.3816489104340612, + "tokens_seen": 2648113152 + }, + { + "epoch": 0.8, + "learning_rate": 0.00015982347041123372, + "loss": 1.2241, + "theoretical_loss": 3.381642367131278, + "tokens_seen": 2648178688 + }, + { + "epoch": 0.8, + "learning_rate": 0.00015980742226680042, + "loss": 1.1401, + "theoretical_loss": 3.381635824035763, + "tokens_seen": 2648244224 + }, + { + "epoch": 0.8, + "learning_rate": 0.00015979137412236712, + "loss": 1.2445, + "theoretical_loss": 3.3816292811475046, + "tokens_seen": 2648309760 + }, + { + "epoch": 0.8, + "learning_rate": 0.00015977532597793382, + "loss": 1.2567, + "theoretical_loss": 3.3816227384664908, + "tokens_seen": 2648375296 + }, + { + "epoch": 0.8, + "learning_rate": 0.00015975927783350052, + "loss": 1.2112, + "theoretical_loss": 3.3816161959927107, + "tokens_seen": 2648440832 + }, + { + "epoch": 0.8, + "learning_rate": 0.0001597432296890672, + "loss": 1.2585, + "theoretical_loss": 3.3816096537261515, + "tokens_seen": 2648506368 + }, + { + "epoch": 0.8, + "learning_rate": 0.0001597271815446339, + "loss": 1.2949, + "theoretical_loss": 3.381603111666802, + "tokens_seen": 2648571904 + }, + { + "epoch": 0.8, + "learning_rate": 0.0001597111334002006, + "loss": 1.2674, + "theoretical_loss": 3.381596569814651, + "tokens_seen": 2648637440 + }, + { + "epoch": 0.8, + "learning_rate": 0.0001596950852557673, + "loss": 1.2645, + "theoretical_loss": 3.3815900281696867, + "tokens_seen": 2648702976 + }, + { + "epoch": 0.8, + "learning_rate": 0.000159679037111334, + "loss": 1.3473, + "theoretical_loss": 3.3815834867318966, + "tokens_seen": 2648768512 + }, + { + "epoch": 0.8, + "learning_rate": 0.0001596629889669007, + "loss": 1.3613, + "theoretical_loss": 3.3815769455012696, + "tokens_seen": 2648834048 + }, + { + "epoch": 0.8, + "learning_rate": 0.0001596469408224674, + "loss": 1.2884, + "theoretical_loss": 3.381570404477794, + "tokens_seen": 2648899584 + }, + { + "epoch": 0.8, + "learning_rate": 0.0001596308926780341, + "loss": 1.2467, + "theoretical_loss": 3.381563863661458, + "tokens_seen": 2648965120 + }, + { + "epoch": 0.8, + "learning_rate": 0.0001596148445336008, + "loss": 1.2517, + "theoretical_loss": 3.3815573230522498, + "tokens_seen": 2649030656 + }, + { + "epoch": 0.8, + "learning_rate": 0.0001595987963891675, + "loss": 1.1821, + "theoretical_loss": 3.3815507826501583, + "tokens_seen": 2649096192 + }, + { + "epoch": 0.8, + "learning_rate": 0.0001595827482447342, + "loss": 1.2012, + "theoretical_loss": 3.381544242455171, + "tokens_seen": 2649161728 + }, + { + "epoch": 0.8, + "learning_rate": 0.0001595667001003009, + "loss": 1.2828, + "theoretical_loss": 3.381537702467277, + "tokens_seen": 2649227264 + }, + { + "epoch": 0.8, + "learning_rate": 0.0001595506519558676, + "loss": 1.3688, + "theoretical_loss": 3.3815311626864637, + "tokens_seen": 2649292800 + }, + { + "epoch": 0.8, + "learning_rate": 0.0001595346038114343, + "loss": 1.2935, + "theoretical_loss": 3.3815246231127203, + "tokens_seen": 2649358336 + }, + { + "epoch": 0.8, + "learning_rate": 0.000159518555667001, + "loss": 1.2966, + "theoretical_loss": 3.381518083746035, + "tokens_seen": 2649423872 + }, + { + "epoch": 0.8, + "learning_rate": 0.0001595025075225677, + "loss": 1.2907, + "theoretical_loss": 3.3815115445863952, + "tokens_seen": 2649489408 + }, + { + "epoch": 0.8, + "learning_rate": 0.0001594864593781344, + "loss": 1.2361, + "theoretical_loss": 3.3815050056337905, + "tokens_seen": 2649554944 + }, + { + "epoch": 0.8, + "learning_rate": 0.0001594704112337011, + "loss": 1.3284, + "theoretical_loss": 3.3814984668882087, + "tokens_seen": 2649620480 + }, + { + "epoch": 0.8, + "learning_rate": 0.00015945436308926783, + "loss": 1.2712, + "theoretical_loss": 3.3814919283496376, + "tokens_seen": 2649686016 + }, + { + "epoch": 0.8, + "learning_rate": 0.0001594383149448345, + "loss": 1.1895, + "theoretical_loss": 3.3814853900180664, + "tokens_seen": 2649751552 + }, + { + "epoch": 0.8, + "learning_rate": 0.0001594222668004012, + "loss": 1.3524, + "theoretical_loss": 3.3814788518934833, + "tokens_seen": 2649817088 + }, + { + "epoch": 0.8, + "learning_rate": 0.0001594062186559679, + "loss": 1.2691, + "theoretical_loss": 3.3814723139758756, + "tokens_seen": 2649882624 + }, + { + "epoch": 0.8, + "learning_rate": 0.0001593901705115346, + "loss": 1.264, + "theoretical_loss": 3.381465776265233, + "tokens_seen": 2649948160 + }, + { + "epoch": 0.8, + "learning_rate": 0.0001593741223671013, + "loss": 1.2707, + "theoretical_loss": 3.381459238761543, + "tokens_seen": 2650013696 + }, + { + "epoch": 0.8, + "learning_rate": 0.000159358074222668, + "loss": 1.3342, + "theoretical_loss": 3.381452701464794, + "tokens_seen": 2650079232 + }, + { + "epoch": 0.8, + "learning_rate": 0.0001593420260782347, + "loss": 1.1849, + "theoretical_loss": 3.3814461643749745, + "tokens_seen": 2650144768 + }, + { + "epoch": 0.8, + "learning_rate": 0.0001593259779338014, + "loss": 1.1631, + "theoretical_loss": 3.381439627492073, + "tokens_seen": 2650210304 + }, + { + "epoch": 0.8, + "learning_rate": 0.0001593099297893681, + "loss": 1.3175, + "theoretical_loss": 3.3814330908160777, + "tokens_seen": 2650275840 + }, + { + "epoch": 0.8, + "learning_rate": 0.00015929388164493481, + "loss": 1.0833, + "theoretical_loss": 3.3814265543469766, + "tokens_seen": 2650341376 + }, + { + "epoch": 0.8, + "learning_rate": 0.00015927783350050151, + "loss": 1.2265, + "theoretical_loss": 3.381420018084759, + "tokens_seen": 2650406912 + }, + { + "epoch": 0.8, + "learning_rate": 0.00015926178535606822, + "loss": 1.2497, + "theoretical_loss": 3.3814134820294117, + "tokens_seen": 2650472448 + }, + { + "epoch": 0.8, + "learning_rate": 0.00015924573721163492, + "loss": 1.2301, + "theoretical_loss": 3.381406946180924, + "tokens_seen": 2650537984 + }, + { + "epoch": 0.8, + "learning_rate": 0.00015922968906720162, + "loss": 1.3477, + "theoretical_loss": 3.3814004105392845, + "tokens_seen": 2650603520 + }, + { + "epoch": 0.8, + "learning_rate": 0.00015921364092276832, + "loss": 1.2914, + "theoretical_loss": 3.3813938751044814, + "tokens_seen": 2650669056 + }, + { + "epoch": 0.8, + "learning_rate": 0.00015919759277833502, + "loss": 1.3216, + "theoretical_loss": 3.381387339876502, + "tokens_seen": 2650734592 + }, + { + "epoch": 0.8, + "learning_rate": 0.00015918154463390172, + "loss": 1.1583, + "theoretical_loss": 3.381380804855336, + "tokens_seen": 2650800128 + }, + { + "epoch": 0.8, + "learning_rate": 0.00015916549648946842, + "loss": 1.2823, + "theoretical_loss": 3.381374270040971, + "tokens_seen": 2650865664 + }, + { + "epoch": 0.8, + "objective/train/docs_used": 962920, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.0632336139678955, + "objective/train/theoretical_loss": 3.381367735433396, + "objective/train/tokens_used": 2671391200, + "theoretical_loss": 3.381367735433396, + "tokens_seen": 2650931200 + }, + { + "epoch": 0.8, + "learning_rate": 0.00015914944834503512, + "loss": 1.1577, + "theoretical_loss": 3.381367735433396, + "tokens_seen": 2650931200 + }, + { + "epoch": 0.8, + "learning_rate": 0.00015913340020060182, + "loss": 1.3197, + "theoretical_loss": 3.381361201032598, + "tokens_seen": 2650996736 + }, + { + "epoch": 0.8, + "learning_rate": 0.00015911735205616852, + "loss": 1.238, + "theoretical_loss": 3.381354666838567, + "tokens_seen": 2651062272 + }, + { + "epoch": 0.8, + "learning_rate": 0.00015910130391173522, + "loss": 1.3169, + "theoretical_loss": 3.3813481328512904, + "tokens_seen": 2651127808 + }, + { + "epoch": 0.8, + "learning_rate": 0.00015908525576730192, + "loss": 1.206, + "theoretical_loss": 3.3813415990707565, + "tokens_seen": 2651193344 + }, + { + "epoch": 0.8, + "learning_rate": 0.00015906920762286862, + "loss": 1.1528, + "theoretical_loss": 3.381335065496954, + "tokens_seen": 2651258880 + }, + { + "epoch": 0.8, + "learning_rate": 0.0001590531594784353, + "loss": 1.1646, + "theoretical_loss": 3.381328532129871, + "tokens_seen": 2651324416 + }, + { + "epoch": 0.8, + "learning_rate": 0.000159037111334002, + "loss": 1.3123, + "theoretical_loss": 3.3813219989694963, + "tokens_seen": 2651389952 + }, + { + "epoch": 0.8, + "learning_rate": 0.0001590210631895687, + "loss": 1.4271, + "theoretical_loss": 3.381315466015818, + "tokens_seen": 2651455488 + }, + { + "epoch": 0.8, + "learning_rate": 0.0001590050150451354, + "loss": 1.2983, + "theoretical_loss": 3.3813089332688238, + "tokens_seen": 2651521024 + }, + { + "epoch": 0.8, + "learning_rate": 0.0001589889669007021, + "loss": 1.2867, + "theoretical_loss": 3.381302400728503, + "tokens_seen": 2651586560 + }, + { + "epoch": 0.8, + "learning_rate": 0.0001589729187562688, + "loss": 1.2946, + "theoretical_loss": 3.3812958683948438, + "tokens_seen": 2651652096 + }, + { + "epoch": 0.8, + "learning_rate": 0.0001589568706118355, + "loss": 1.2222, + "theoretical_loss": 3.381289336267834, + "tokens_seen": 2651717632 + }, + { + "epoch": 0.8, + "learning_rate": 0.0001589408224674022, + "loss": 1.2569, + "theoretical_loss": 3.381282804347462, + "tokens_seen": 2651783168 + }, + { + "epoch": 0.8, + "learning_rate": 0.0001589247743229689, + "loss": 1.2855, + "theoretical_loss": 3.3812762726337167, + "tokens_seen": 2651848704 + }, + { + "epoch": 0.8, + "learning_rate": 0.0001589087261785356, + "loss": 1.2406, + "theoretical_loss": 3.3812697411265864, + "tokens_seen": 2651914240 + }, + { + "epoch": 0.8, + "learning_rate": 0.0001588926780341023, + "loss": 1.2332, + "theoretical_loss": 3.381263209826059, + "tokens_seen": 2651979776 + }, + { + "epoch": 0.8, + "learning_rate": 0.000158876629889669, + "loss": 1.3101, + "theoretical_loss": 3.3812566787321234, + "tokens_seen": 2652045312 + }, + { + "epoch": 0.8, + "learning_rate": 0.0001588605817452357, + "loss": 1.2007, + "theoretical_loss": 3.381250147844767, + "tokens_seen": 2652110848 + }, + { + "epoch": 0.8, + "learning_rate": 0.00015884453360080243, + "loss": 1.3539, + "theoretical_loss": 3.38124361716398, + "tokens_seen": 2652176384 + }, + { + "epoch": 0.8, + "learning_rate": 0.00015882848545636913, + "loss": 1.2403, + "theoretical_loss": 3.3812370866897488, + "tokens_seen": 2652241920 + }, + { + "epoch": 0.8, + "learning_rate": 0.00015881243731193583, + "loss": 1.28, + "theoretical_loss": 3.3812305564220626, + "tokens_seen": 2652307456 + }, + { + "epoch": 0.8, + "learning_rate": 0.00015879638916750253, + "loss": 1.1919, + "theoretical_loss": 3.38122402636091, + "tokens_seen": 2652372992 + }, + { + "epoch": 0.8, + "learning_rate": 0.00015878034102306923, + "loss": 1.2553, + "theoretical_loss": 3.3812174965062787, + "tokens_seen": 2652438528 + }, + { + "epoch": 0.8, + "learning_rate": 0.00015876429287863593, + "loss": 1.2488, + "theoretical_loss": 3.381210966858158, + "tokens_seen": 2652504064 + }, + { + "epoch": 0.8, + "learning_rate": 0.00015874824473420264, + "loss": 1.386, + "theoretical_loss": 3.3812044374165353, + "tokens_seen": 2652569600 + }, + { + "epoch": 0.8, + "learning_rate": 0.0001587321965897693, + "loss": 1.2553, + "theoretical_loss": 3.3811979081813996, + "tokens_seen": 2652635136 + }, + { + "epoch": 0.8, + "learning_rate": 0.000158716148445336, + "loss": 1.1707, + "theoretical_loss": 3.381191379152739, + "tokens_seen": 2652700672 + }, + { + "epoch": 0.8, + "learning_rate": 0.0001587001003009027, + "loss": 1.1859, + "theoretical_loss": 3.381184850330542, + "tokens_seen": 2652766208 + }, + { + "epoch": 0.8, + "learning_rate": 0.0001586840521564694, + "loss": 1.3311, + "theoretical_loss": 3.381178321714797, + "tokens_seen": 2652831744 + }, + { + "epoch": 0.8, + "learning_rate": 0.0001586680040120361, + "loss": 1.2694, + "theoretical_loss": 3.3811717933054917, + "tokens_seen": 2652897280 + }, + { + "epoch": 0.8, + "learning_rate": 0.0001586519558676028, + "loss": 1.0917, + "theoretical_loss": 3.3811652651026156, + "tokens_seen": 2652962816 + }, + { + "epoch": 0.8, + "learning_rate": 0.0001586359077231695, + "loss": 1.2036, + "theoretical_loss": 3.3811587371061567, + "tokens_seen": 2653028352 + }, + { + "epoch": 0.8, + "learning_rate": 0.0001586198595787362, + "loss": 1.3026, + "theoretical_loss": 3.381152209316103, + "tokens_seen": 2653093888 + }, + { + "epoch": 0.8, + "learning_rate": 0.00015860381143430291, + "loss": 1.1439, + "theoretical_loss": 3.3811456817324426, + "tokens_seen": 2653159424 + }, + { + "epoch": 0.8, + "learning_rate": 0.00015858776328986961, + "loss": 1.2285, + "theoretical_loss": 3.3811391543551648, + "tokens_seen": 2653224960 + }, + { + "epoch": 0.8, + "learning_rate": 0.00015857171514543632, + "loss": 1.1923, + "theoretical_loss": 3.3811326271842574, + "tokens_seen": 2653290496 + }, + { + "epoch": 0.8, + "learning_rate": 0.00015855566700100302, + "loss": 1.2664, + "theoretical_loss": 3.381126100219709, + "tokens_seen": 2653356032 + }, + { + "epoch": 0.8, + "learning_rate": 0.00015853961885656972, + "loss": 1.2215, + "theoretical_loss": 3.381119573461508, + "tokens_seen": 2653421568 + }, + { + "epoch": 0.8, + "learning_rate": 0.00015852357071213642, + "loss": 1.2162, + "theoretical_loss": 3.3811130469096424, + "tokens_seen": 2653487104 + }, + { + "epoch": 0.8, + "learning_rate": 0.00015850752256770312, + "loss": 1.2472, + "theoretical_loss": 3.381106520564101, + "tokens_seen": 2653552640 + }, + { + "epoch": 0.8, + "learning_rate": 0.00015849147442326982, + "loss": 1.2643, + "theoretical_loss": 3.381099994424872, + "tokens_seen": 2653618176 + }, + { + "epoch": 0.8, + "learning_rate": 0.00015847542627883652, + "loss": 1.3101, + "theoretical_loss": 3.381093468491944, + "tokens_seen": 2653683712 + }, + { + "epoch": 0.8, + "learning_rate": 0.00015845937813440322, + "loss": 1.1893, + "theoretical_loss": 3.3810869427653047, + "tokens_seen": 2653749248 + }, + { + "epoch": 0.8, + "learning_rate": 0.00015844332998996992, + "loss": 1.187, + "theoretical_loss": 3.3810804172449433, + "tokens_seen": 2653814784 + }, + { + "epoch": 0.8, + "learning_rate": 0.00015842728184553662, + "loss": 1.3057, + "theoretical_loss": 3.381073891930848, + "tokens_seen": 2653880320 + }, + { + "epoch": 0.8, + "learning_rate": 0.00015841123370110332, + "loss": 1.2493, + "theoretical_loss": 3.381067366823007, + "tokens_seen": 2653945856 + }, + { + "epoch": 0.8, + "learning_rate": 0.00015839518555667002, + "loss": 1.2065, + "theoretical_loss": 3.3810608419214088, + "tokens_seen": 2654011392 + }, + { + "epoch": 0.8, + "learning_rate": 0.00015837913741223672, + "loss": 1.3398, + "theoretical_loss": 3.381054317226041, + "tokens_seen": 2654076928 + }, + { + "epoch": 0.8, + "learning_rate": 0.00015836308926780342, + "loss": 1.337, + "theoretical_loss": 3.3810477927368936, + "tokens_seen": 2654142464 + }, + { + "epoch": 0.8, + "objective/train/docs_used": 964093, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.2230384349822998, + "objective/train/theoretical_loss": 3.381041268453954, + "objective/train/tokens_used": 2674668000, + "theoretical_loss": 3.381041268453954, + "tokens_seen": 2654208000 + }, + { + "epoch": 0.8, + "learning_rate": 0.0001583470411233701, + "loss": 1.1625, + "theoretical_loss": 3.381041268453954, + "tokens_seen": 2654208000 + }, + { + "epoch": 0.8, + "learning_rate": 0.0001583309929789368, + "loss": 1.2416, + "theoretical_loss": 3.38103474437721, + "tokens_seen": 2654273536 + }, + { + "epoch": 0.8, + "learning_rate": 0.0001583149448345035, + "loss": 1.3163, + "theoretical_loss": 3.3810282205066517, + "tokens_seen": 2654339072 + }, + { + "epoch": 0.8, + "learning_rate": 0.0001582988966900702, + "loss": 1.2917, + "theoretical_loss": 3.381021696842266, + "tokens_seen": 2654404608 + }, + { + "epoch": 0.8, + "learning_rate": 0.0001582828485456369, + "loss": 1.1827, + "theoretical_loss": 3.3810151733840415, + "tokens_seen": 2654470144 + }, + { + "epoch": 0.8, + "learning_rate": 0.0001582668004012036, + "loss": 1.2018, + "theoretical_loss": 3.3810086501319674, + "tokens_seen": 2654535680 + }, + { + "epoch": 0.8, + "learning_rate": 0.00015825075225677033, + "loss": 1.1715, + "theoretical_loss": 3.381002127086031, + "tokens_seen": 2654601216 + }, + { + "epoch": 0.8, + "learning_rate": 0.00015823470411233703, + "loss": 1.2329, + "theoretical_loss": 3.380995604246222, + "tokens_seen": 2654666752 + }, + { + "epoch": 0.8, + "learning_rate": 0.00015821865596790373, + "loss": 1.2106, + "theoretical_loss": 3.3809890816125274, + "tokens_seen": 2654732288 + }, + { + "epoch": 0.8, + "learning_rate": 0.00015820260782347043, + "loss": 1.2724, + "theoretical_loss": 3.3809825591849365, + "tokens_seen": 2654797824 + }, + { + "epoch": 0.8, + "learning_rate": 0.00015818655967903713, + "loss": 1.3037, + "theoretical_loss": 3.3809760369634376, + "tokens_seen": 2654863360 + }, + { + "epoch": 0.8, + "learning_rate": 0.00015817051153460383, + "loss": 1.2947, + "theoretical_loss": 3.380969514948019, + "tokens_seen": 2654928896 + }, + { + "epoch": 0.8, + "learning_rate": 0.00015815446339017053, + "loss": 1.2899, + "theoretical_loss": 3.380962993138669, + "tokens_seen": 2654994432 + }, + { + "epoch": 0.8, + "learning_rate": 0.00015813841524573723, + "loss": 1.1055, + "theoretical_loss": 3.380956471535376, + "tokens_seen": 2655059968 + }, + { + "epoch": 0.8, + "learning_rate": 0.00015812236710130393, + "loss": 1.3903, + "theoretical_loss": 3.3809499501381284, + "tokens_seen": 2655125504 + }, + { + "epoch": 0.8, + "learning_rate": 0.00015810631895687063, + "loss": 1.1126, + "theoretical_loss": 3.3809434289469147, + "tokens_seen": 2655191040 + }, + { + "epoch": 0.8, + "learning_rate": 0.00015809027081243733, + "loss": 1.2758, + "theoretical_loss": 3.3809369079617233, + "tokens_seen": 2655256576 + }, + { + "epoch": 0.8, + "learning_rate": 0.00015807422266800404, + "loss": 1.1852, + "theoretical_loss": 3.3809303871825427, + "tokens_seen": 2655322112 + }, + { + "epoch": 0.8, + "learning_rate": 0.00015805817452357074, + "loss": 1.2178, + "theoretical_loss": 3.380923866609361, + "tokens_seen": 2655387648 + }, + { + "epoch": 0.8, + "learning_rate": 0.00015804212637913744, + "loss": 1.3833, + "theoretical_loss": 3.3809173462421667, + "tokens_seen": 2655453184 + }, + { + "epoch": 0.8, + "learning_rate": 0.0001580260782347041, + "loss": 1.3001, + "theoretical_loss": 3.3809108260809486, + "tokens_seen": 2655518720 + }, + { + "epoch": 0.8, + "learning_rate": 0.0001580100300902708, + "loss": 1.3019, + "theoretical_loss": 3.380904306125695, + "tokens_seen": 2655584256 + }, + { + "epoch": 0.8, + "learning_rate": 0.0001579939819458375, + "loss": 1.2794, + "theoretical_loss": 3.380897786376394, + "tokens_seen": 2655649792 + }, + { + "epoch": 0.8, + "learning_rate": 0.0001579779338014042, + "loss": 1.2759, + "theoretical_loss": 3.380891266833034, + "tokens_seen": 2655715328 + }, + { + "epoch": 0.8, + "learning_rate": 0.0001579618856569709, + "loss": 1.3034, + "theoretical_loss": 3.3808847474956036, + "tokens_seen": 2655780864 + }, + { + "epoch": 0.8, + "learning_rate": 0.0001579458375125376, + "loss": 1.3205, + "theoretical_loss": 3.3808782283640912, + "tokens_seen": 2655846400 + }, + { + "epoch": 0.8, + "learning_rate": 0.00015792978936810431, + "loss": 1.2404, + "theoretical_loss": 3.3808717094384853, + "tokens_seen": 2655911936 + }, + { + "epoch": 0.8, + "learning_rate": 0.00015791374122367101, + "loss": 1.3, + "theoretical_loss": 3.380865190718774, + "tokens_seen": 2655977472 + }, + { + "epoch": 0.8, + "learning_rate": 0.00015789769307923771, + "loss": 1.3204, + "theoretical_loss": 3.380858672204946, + "tokens_seen": 2656043008 + }, + { + "epoch": 0.8, + "learning_rate": 0.00015788164493480442, + "loss": 1.3415, + "theoretical_loss": 3.38085215389699, + "tokens_seen": 2656108544 + }, + { + "epoch": 0.8, + "learning_rate": 0.00015786559679037112, + "loss": 1.262, + "theoretical_loss": 3.3808456357948935, + "tokens_seen": 2656174080 + }, + { + "epoch": 0.8, + "learning_rate": 0.00015784954864593782, + "loss": 1.348, + "theoretical_loss": 3.380839117898646, + "tokens_seen": 2656239616 + }, + { + "epoch": 0.8, + "learning_rate": 0.00015783350050150452, + "loss": 1.2674, + "theoretical_loss": 3.380832600208235, + "tokens_seen": 2656305152 + }, + { + "epoch": 0.8, + "learning_rate": 0.00015781745235707122, + "loss": 1.3162, + "theoretical_loss": 3.380826082723649, + "tokens_seen": 2656370688 + }, + { + "epoch": 0.8, + "learning_rate": 0.00015780140421263792, + "loss": 1.2124, + "theoretical_loss": 3.3808195654448774, + "tokens_seen": 2656436224 + }, + { + "epoch": 0.81, + "learning_rate": 0.00015778535606820462, + "loss": 1.2259, + "theoretical_loss": 3.3808130483719077, + "tokens_seen": 2656501760 + }, + { + "epoch": 0.81, + "learning_rate": 0.00015776930792377132, + "loss": 1.2921, + "theoretical_loss": 3.380806531504729, + "tokens_seen": 2656567296 + }, + { + "epoch": 0.81, + "learning_rate": 0.00015775325977933802, + "loss": 1.3229, + "theoretical_loss": 3.380800014843329, + "tokens_seen": 2656632832 + }, + { + "epoch": 0.81, + "learning_rate": 0.00015773721163490472, + "loss": 1.4156, + "theoretical_loss": 3.380793498387696, + "tokens_seen": 2656698368 + }, + { + "epoch": 0.81, + "learning_rate": 0.00015772116349047142, + "loss": 1.2346, + "theoretical_loss": 3.3807869821378196, + "tokens_seen": 2656763904 + }, + { + "epoch": 0.81, + "learning_rate": 0.00015770511534603812, + "loss": 1.3159, + "theoretical_loss": 3.380780466093687, + "tokens_seen": 2656829440 + }, + { + "epoch": 0.81, + "learning_rate": 0.00015768906720160482, + "loss": 1.3293, + "theoretical_loss": 3.3807739502552874, + "tokens_seen": 2656894976 + }, + { + "epoch": 0.81, + "learning_rate": 0.00015767301905717152, + "loss": 1.2281, + "theoretical_loss": 3.380767434622609, + "tokens_seen": 2656960512 + }, + { + "epoch": 0.81, + "learning_rate": 0.00015765697091273822, + "loss": 1.2328, + "theoretical_loss": 3.38076091919564, + "tokens_seen": 2657026048 + }, + { + "epoch": 0.81, + "learning_rate": 0.00015764092276830493, + "loss": 1.2637, + "theoretical_loss": 3.380754403974369, + "tokens_seen": 2657091584 + }, + { + "epoch": 0.81, + "learning_rate": 0.00015762487462387163, + "loss": 1.3106, + "theoretical_loss": 3.3807478889587848, + "tokens_seen": 2657157120 + }, + { + "epoch": 0.81, + "learning_rate": 0.00015760882647943833, + "loss": 1.4626, + "theoretical_loss": 3.380741374148875, + "tokens_seen": 2657222656 + }, + { + "epoch": 0.81, + "learning_rate": 0.00015759277833500503, + "loss": 1.2894, + "theoretical_loss": 3.3807348595446287, + "tokens_seen": 2657288192 + }, + { + "epoch": 0.81, + "learning_rate": 0.00015757673019057173, + "loss": 1.2429, + "theoretical_loss": 3.3807283451460344, + "tokens_seen": 2657353728 + }, + { + "epoch": 0.81, + "learning_rate": 0.00015756068204613843, + "loss": 1.293, + "theoretical_loss": 3.38072183095308, + "tokens_seen": 2657419264 + }, + { + "epoch": 0.81, + "objective/train/docs_used": 965310, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.3755323886871338, + "objective/train/theoretical_loss": 3.3807153169657544, + "objective/train/tokens_used": 2677944800, + "theoretical_loss": 3.3807153169657544, + "tokens_seen": 2657484800 + }, + { + "epoch": 0.81, + "learning_rate": 0.00015754463390170513, + "loss": 1.1976, + "theoretical_loss": 3.3807153169657544, + "tokens_seen": 2657484800 + }, + { + "epoch": 0.81, + "learning_rate": 0.00015752858575727183, + "loss": 1.216, + "theoretical_loss": 3.3807088031840458, + "tokens_seen": 2657550336 + }, + { + "epoch": 0.81, + "learning_rate": 0.00015751253761283853, + "loss": 1.3128, + "theoretical_loss": 3.380702289607943, + "tokens_seen": 2657615872 + }, + { + "epoch": 0.81, + "learning_rate": 0.00015749648946840523, + "loss": 1.202, + "theoretical_loss": 3.3806957762374337, + "tokens_seen": 2657681408 + }, + { + "epoch": 0.81, + "learning_rate": 0.00015748044132397193, + "loss": 1.2169, + "theoretical_loss": 3.3806892630725067, + "tokens_seen": 2657746944 + }, + { + "epoch": 0.81, + "learning_rate": 0.00015746439317953863, + "loss": 1.2193, + "theoretical_loss": 3.380682750113151, + "tokens_seen": 2657812480 + }, + { + "epoch": 0.81, + "learning_rate": 0.00015744834503510533, + "loss": 1.2572, + "theoretical_loss": 3.380676237359354, + "tokens_seen": 2657878016 + }, + { + "epoch": 0.81, + "learning_rate": 0.00015743229689067203, + "loss": 1.2984, + "theoretical_loss": 3.3806697248111055, + "tokens_seen": 2657943552 + }, + { + "epoch": 0.81, + "learning_rate": 0.00015741624874623873, + "loss": 1.2003, + "theoretical_loss": 3.3806632124683924, + "tokens_seen": 2658009088 + }, + { + "epoch": 0.81, + "learning_rate": 0.00015740020060180543, + "loss": 1.3062, + "theoretical_loss": 3.3806567003312047, + "tokens_seen": 2658074624 + }, + { + "epoch": 0.81, + "learning_rate": 0.00015738415245737214, + "loss": 1.3077, + "theoretical_loss": 3.3806501883995295, + "tokens_seen": 2658140160 + }, + { + "epoch": 0.81, + "learning_rate": 0.00015736810431293884, + "loss": 1.3473, + "theoretical_loss": 3.3806436766733556, + "tokens_seen": 2658205696 + }, + { + "epoch": 0.81, + "learning_rate": 0.00015735205616850554, + "loss": 1.2361, + "theoretical_loss": 3.380637165152672, + "tokens_seen": 2658271232 + }, + { + "epoch": 0.81, + "learning_rate": 0.0001573360080240722, + "loss": 1.3263, + "theoretical_loss": 3.380630653837467, + "tokens_seen": 2658336768 + }, + { + "epoch": 0.81, + "learning_rate": 0.0001573199598796389, + "loss": 1.2617, + "theoretical_loss": 3.3806241427277284, + "tokens_seen": 2658402304 + }, + { + "epoch": 0.81, + "learning_rate": 0.0001573039117352056, + "loss": 1.117, + "theoretical_loss": 3.3806176318234455, + "tokens_seen": 2658467840 + }, + { + "epoch": 0.81, + "learning_rate": 0.0001572878635907723, + "loss": 1.351, + "theoretical_loss": 3.3806111211246064, + "tokens_seen": 2658533376 + }, + { + "epoch": 0.81, + "learning_rate": 0.000157271815446339, + "loss": 1.4361, + "theoretical_loss": 3.380604610631199, + "tokens_seen": 2658598912 + }, + { + "epoch": 0.81, + "learning_rate": 0.0001572557673019057, + "loss": 1.2533, + "theoretical_loss": 3.380598100343213, + "tokens_seen": 2658664448 + }, + { + "epoch": 0.81, + "learning_rate": 0.00015723971915747241, + "loss": 1.1932, + "theoretical_loss": 3.3805915902606354, + "tokens_seen": 2658729984 + }, + { + "epoch": 0.81, + "learning_rate": 0.00015722367101303911, + "loss": 1.3644, + "theoretical_loss": 3.3805850803834554, + "tokens_seen": 2658795520 + }, + { + "epoch": 0.81, + "learning_rate": 0.00015720762286860582, + "loss": 1.4258, + "theoretical_loss": 3.380578570711662, + "tokens_seen": 2658861056 + }, + { + "epoch": 0.81, + "learning_rate": 0.00015719157472417252, + "loss": 1.3153, + "theoretical_loss": 3.3805720612452426, + "tokens_seen": 2658926592 + }, + { + "epoch": 0.81, + "learning_rate": 0.00015717552657973922, + "loss": 1.3301, + "theoretical_loss": 3.380565551984186, + "tokens_seen": 2658992128 + }, + { + "epoch": 0.81, + "learning_rate": 0.00015715947843530592, + "loss": 1.1992, + "theoretical_loss": 3.3805590429284815, + "tokens_seen": 2659057664 + }, + { + "epoch": 0.81, + "learning_rate": 0.00015714343029087262, + "loss": 1.1347, + "theoretical_loss": 3.3805525340781166, + "tokens_seen": 2659123200 + }, + { + "epoch": 0.81, + "learning_rate": 0.00015712738214643932, + "loss": 1.3746, + "theoretical_loss": 3.38054602543308, + "tokens_seen": 2659188736 + }, + { + "epoch": 0.81, + "learning_rate": 0.00015711133400200602, + "loss": 1.1828, + "theoretical_loss": 3.38053951699336, + "tokens_seen": 2659254272 + }, + { + "epoch": 0.81, + "learning_rate": 0.00015709528585757272, + "loss": 1.3043, + "theoretical_loss": 3.3805330087589454, + "tokens_seen": 2659319808 + }, + { + "epoch": 0.81, + "learning_rate": 0.00015707923771313942, + "loss": 1.3169, + "theoretical_loss": 3.3805265007298244, + "tokens_seen": 2659385344 + }, + { + "epoch": 0.81, + "learning_rate": 0.00015706318956870612, + "loss": 1.1475, + "theoretical_loss": 3.380519992905986, + "tokens_seen": 2659450880 + }, + { + "epoch": 0.81, + "learning_rate": 0.00015704714142427285, + "loss": 1.2939, + "theoretical_loss": 3.380513485287418, + "tokens_seen": 2659516416 + }, + { + "epoch": 0.81, + "learning_rate": 0.00015703109327983955, + "loss": 1.271, + "theoretical_loss": 3.380506977874109, + "tokens_seen": 2659581952 + }, + { + "epoch": 0.81, + "learning_rate": 0.00015701504513540622, + "loss": 1.3307, + "theoretical_loss": 3.3805004706660475, + "tokens_seen": 2659647488 + }, + { + "epoch": 0.81, + "learning_rate": 0.00015699899699097292, + "loss": 1.3129, + "theoretical_loss": 3.3804939636632225, + "tokens_seen": 2659713024 + }, + { + "epoch": 0.81, + "learning_rate": 0.00015698294884653962, + "loss": 1.2773, + "theoretical_loss": 3.3804874568656214, + "tokens_seen": 2659778560 + }, + { + "epoch": 0.81, + "learning_rate": 0.00015696690070210632, + "loss": 1.154, + "theoretical_loss": 3.380480950273234, + "tokens_seen": 2659844096 + }, + { + "epoch": 0.81, + "learning_rate": 0.00015695085255767303, + "loss": 1.2227, + "theoretical_loss": 3.3804744438860475, + "tokens_seen": 2659909632 + }, + { + "epoch": 0.81, + "learning_rate": 0.00015693480441323973, + "loss": 1.2909, + "theoretical_loss": 3.380467937704051, + "tokens_seen": 2659975168 + }, + { + "epoch": 0.81, + "learning_rate": 0.00015691875626880643, + "loss": 1.2536, + "theoretical_loss": 3.3804614317272335, + "tokens_seen": 2660040704 + }, + { + "epoch": 0.81, + "learning_rate": 0.00015690270812437313, + "loss": 1.3534, + "theoretical_loss": 3.3804549259555823, + "tokens_seen": 2660106240 + }, + { + "epoch": 0.81, + "learning_rate": 0.00015688665997993983, + "loss": 1.2687, + "theoretical_loss": 3.3804484203890866, + "tokens_seen": 2660171776 + }, + { + "epoch": 0.81, + "learning_rate": 0.00015687061183550653, + "loss": 1.2384, + "theoretical_loss": 3.3804419150277347, + "tokens_seen": 2660237312 + }, + { + "epoch": 0.81, + "learning_rate": 0.00015685456369107323, + "loss": 1.4067, + "theoretical_loss": 3.3804354098715153, + "tokens_seen": 2660302848 + }, + { + "epoch": 0.81, + "learning_rate": 0.00015683851554663993, + "loss": 1.3167, + "theoretical_loss": 3.3804289049204166, + "tokens_seen": 2660368384 + }, + { + "epoch": 0.81, + "learning_rate": 0.00015682246740220663, + "loss": 1.2123, + "theoretical_loss": 3.380422400174427, + "tokens_seen": 2660433920 + }, + { + "epoch": 0.81, + "learning_rate": 0.00015680641925777333, + "loss": 1.2378, + "theoretical_loss": 3.3804158956335355, + "tokens_seen": 2660499456 + }, + { + "epoch": 0.81, + "learning_rate": 0.00015679037111334003, + "loss": 1.1672, + "theoretical_loss": 3.3804093912977304, + "tokens_seen": 2660564992 + }, + { + "epoch": 0.81, + "learning_rate": 0.00015677432296890673, + "loss": 1.29, + "theoretical_loss": 3.3804028871669995, + "tokens_seen": 2660630528 + }, + { + "epoch": 0.81, + "learning_rate": 0.00015675827482447343, + "loss": 1.2928, + "theoretical_loss": 3.380396383241332, + "tokens_seen": 2660696064 + }, + { + "epoch": 0.81, + "objective/train/docs_used": 966512, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.3489511013031006, + "objective/train/theoretical_loss": 3.3803898795207163, + "objective/train/tokens_used": 2681221600, + "theoretical_loss": 3.3803898795207163, + "tokens_seen": 2660761600 + }, + { + "epoch": 0.81, + "learning_rate": 0.00015674222668004013, + "loss": 1.2663, + "theoretical_loss": 3.3803898795207163, + "tokens_seen": 2660761600 + }, + { + "epoch": 0.81, + "learning_rate": 0.00015672617853560683, + "loss": 1.2784, + "theoretical_loss": 3.380383376005141, + "tokens_seen": 2660827136 + }, + { + "epoch": 0.81, + "learning_rate": 0.00015671013039117353, + "loss": 1.2329, + "theoretical_loss": 3.380376872694594, + "tokens_seen": 2660892672 + }, + { + "epoch": 0.81, + "learning_rate": 0.00015669408224674024, + "loss": 1.3024, + "theoretical_loss": 3.380370369589064, + "tokens_seen": 2660958208 + }, + { + "epoch": 0.81, + "learning_rate": 0.00015667803410230694, + "loss": 1.1757, + "theoretical_loss": 3.3803638666885396, + "tokens_seen": 2661023744 + }, + { + "epoch": 0.81, + "learning_rate": 0.00015666198595787364, + "loss": 1.3464, + "theoretical_loss": 3.38035736399301, + "tokens_seen": 2661089280 + }, + { + "epoch": 0.81, + "learning_rate": 0.00015664593781344034, + "loss": 1.3007, + "theoretical_loss": 3.3803508615024627, + "tokens_seen": 2661154816 + }, + { + "epoch": 0.81, + "learning_rate": 0.000156629889669007, + "loss": 1.3134, + "theoretical_loss": 3.380344359216886, + "tokens_seen": 2661220352 + }, + { + "epoch": 0.81, + "learning_rate": 0.0001566138415245737, + "loss": 1.2494, + "theoretical_loss": 3.3803378571362694, + "tokens_seen": 2661285888 + }, + { + "epoch": 0.81, + "learning_rate": 0.0001565977933801404, + "loss": 1.2919, + "theoretical_loss": 3.380331355260601, + "tokens_seen": 2661351424 + }, + { + "epoch": 0.81, + "learning_rate": 0.0001565817452357071, + "loss": 1.3589, + "theoretical_loss": 3.380324853589869, + "tokens_seen": 2661416960 + }, + { + "epoch": 0.81, + "learning_rate": 0.0001565656970912738, + "loss": 1.3197, + "theoretical_loss": 3.380318352124062, + "tokens_seen": 2661482496 + }, + { + "epoch": 0.81, + "learning_rate": 0.00015654964894684051, + "loss": 1.2714, + "theoretical_loss": 3.3803118508631687, + "tokens_seen": 2661548032 + }, + { + "epoch": 0.81, + "learning_rate": 0.00015653360080240721, + "loss": 1.2242, + "theoretical_loss": 3.3803053498071773, + "tokens_seen": 2661613568 + }, + { + "epoch": 0.81, + "learning_rate": 0.00015651755265797392, + "loss": 1.3062, + "theoretical_loss": 3.3802988489560764, + "tokens_seen": 2661679104 + }, + { + "epoch": 0.81, + "learning_rate": 0.00015650150451354062, + "loss": 1.2556, + "theoretical_loss": 3.3802923483098546, + "tokens_seen": 2661744640 + }, + { + "epoch": 0.81, + "learning_rate": 0.00015648545636910732, + "loss": 1.3162, + "theoretical_loss": 3.3802858478685005, + "tokens_seen": 2661810176 + }, + { + "epoch": 0.81, + "learning_rate": 0.00015646940822467402, + "loss": 1.2893, + "theoretical_loss": 3.3802793476320026, + "tokens_seen": 2661875712 + }, + { + "epoch": 0.81, + "learning_rate": 0.00015645336008024074, + "loss": 1.31, + "theoretical_loss": 3.380272847600349, + "tokens_seen": 2661941248 + }, + { + "epoch": 0.81, + "learning_rate": 0.00015643731193580745, + "loss": 1.306, + "theoretical_loss": 3.380266347773529, + "tokens_seen": 2662006784 + }, + { + "epoch": 0.81, + "learning_rate": 0.00015642126379137415, + "loss": 1.3312, + "theoretical_loss": 3.38025984815153, + "tokens_seen": 2662072320 + }, + { + "epoch": 0.81, + "learning_rate": 0.00015640521564694085, + "loss": 1.2793, + "theoretical_loss": 3.3802533487343407, + "tokens_seen": 2662137856 + }, + { + "epoch": 0.81, + "learning_rate": 0.00015638916750250755, + "loss": 1.3532, + "theoretical_loss": 3.380246849521951, + "tokens_seen": 2662203392 + }, + { + "epoch": 0.81, + "learning_rate": 0.00015637311935807425, + "loss": 1.3397, + "theoretical_loss": 3.380240350514348, + "tokens_seen": 2662268928 + }, + { + "epoch": 0.81, + "learning_rate": 0.00015635707121364095, + "loss": 1.2509, + "theoretical_loss": 3.38023385171152, + "tokens_seen": 2662334464 + }, + { + "epoch": 0.81, + "learning_rate": 0.00015634102306920765, + "loss": 1.1801, + "theoretical_loss": 3.3802273531134563, + "tokens_seen": 2662400000 + }, + { + "epoch": 0.81, + "learning_rate": 0.00015632497492477435, + "loss": 1.2441, + "theoretical_loss": 3.3802208547201458, + "tokens_seen": 2662465536 + }, + { + "epoch": 0.81, + "learning_rate": 0.00015630892678034102, + "loss": 1.3483, + "theoretical_loss": 3.3802143565315763, + "tokens_seen": 2662531072 + }, + { + "epoch": 0.81, + "learning_rate": 0.00015629287863590772, + "loss": 1.3334, + "theoretical_loss": 3.380207858547736, + "tokens_seen": 2662596608 + }, + { + "epoch": 0.81, + "learning_rate": 0.00015627683049147442, + "loss": 1.404, + "theoretical_loss": 3.380201360768614, + "tokens_seen": 2662662144 + }, + { + "epoch": 0.81, + "learning_rate": 0.00015626078234704113, + "loss": 1.3347, + "theoretical_loss": 3.3801948631941987, + "tokens_seen": 2662727680 + }, + { + "epoch": 0.81, + "learning_rate": 0.00015624473420260783, + "loss": 1.2458, + "theoretical_loss": 3.3801883658244782, + "tokens_seen": 2662793216 + }, + { + "epoch": 0.81, + "learning_rate": 0.00015622868605817453, + "loss": 1.3996, + "theoretical_loss": 3.380181868659442, + "tokens_seen": 2662858752 + }, + { + "epoch": 0.81, + "learning_rate": 0.00015621263791374123, + "loss": 1.3642, + "theoretical_loss": 3.3801753716990777, + "tokens_seen": 2662924288 + }, + { + "epoch": 0.81, + "learning_rate": 0.00015619658976930793, + "loss": 1.2947, + "theoretical_loss": 3.380168874943374, + "tokens_seen": 2662989824 + }, + { + "epoch": 0.81, + "learning_rate": 0.00015618054162487463, + "loss": 1.2159, + "theoretical_loss": 3.3801623783923196, + "tokens_seen": 2663055360 + }, + { + "epoch": 0.81, + "learning_rate": 0.00015616449348044133, + "loss": 1.2507, + "theoretical_loss": 3.380155882045903, + "tokens_seen": 2663120896 + }, + { + "epoch": 0.81, + "learning_rate": 0.00015614844533600803, + "loss": 1.2178, + "theoretical_loss": 3.3801493859041125, + "tokens_seen": 2663186432 + }, + { + "epoch": 0.81, + "learning_rate": 0.00015613239719157473, + "loss": 1.3427, + "theoretical_loss": 3.380142889966937, + "tokens_seen": 2663251968 + }, + { + "epoch": 0.81, + "learning_rate": 0.00015611634904714143, + "loss": 1.4077, + "theoretical_loss": 3.3801363942343645, + "tokens_seen": 2663317504 + }, + { + "epoch": 0.81, + "learning_rate": 0.00015610030090270813, + "loss": 1.2323, + "theoretical_loss": 3.380129898706384, + "tokens_seen": 2663383040 + }, + { + "epoch": 0.81, + "learning_rate": 0.00015608425275827483, + "loss": 1.2579, + "theoretical_loss": 3.3801234033829837, + "tokens_seen": 2663448576 + }, + { + "epoch": 0.81, + "learning_rate": 0.00015606820461384153, + "loss": 1.2518, + "theoretical_loss": 3.3801169082641525, + "tokens_seen": 2663514112 + }, + { + "epoch": 0.81, + "learning_rate": 0.00015605215646940823, + "loss": 1.2155, + "theoretical_loss": 3.3801104133498785, + "tokens_seen": 2663579648 + }, + { + "epoch": 0.81, + "learning_rate": 0.00015603610832497493, + "loss": 1.2214, + "theoretical_loss": 3.3801039186401507, + "tokens_seen": 2663645184 + }, + { + "epoch": 0.81, + "learning_rate": 0.00015602006018054163, + "loss": 1.2739, + "theoretical_loss": 3.3800974241349566, + "tokens_seen": 2663710720 + }, + { + "epoch": 0.81, + "learning_rate": 0.00015600401203610834, + "loss": 1.1823, + "theoretical_loss": 3.380090929834286, + "tokens_seen": 2663776256 + }, + { + "epoch": 0.81, + "learning_rate": 0.00015598796389167504, + "loss": 1.3489, + "theoretical_loss": 3.380084435738127, + "tokens_seen": 2663841792 + }, + { + "epoch": 0.81, + "learning_rate": 0.00015597191574724174, + "loss": 1.3711, + "theoretical_loss": 3.380077941846468, + "tokens_seen": 2663907328 + }, + { + "epoch": 0.81, + "learning_rate": 0.00015595586760280844, + "loss": 1.2144, + "theoretical_loss": 3.3800714481592973, + "tokens_seen": 2663972864 + }, + { + "epoch": 0.81, + "objective/train/docs_used": 967741, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.2659364938735962, + "objective/train/theoretical_loss": 3.3800649546766035, + "objective/train/tokens_used": 2684498400, + "theoretical_loss": 3.3800649546766035, + "tokens_seen": 2664038400 + }, + { + "epoch": 0.81, + "learning_rate": 0.00015593981945837514, + "loss": 1.3171, + "theoretical_loss": 3.3800649546766035, + "tokens_seen": 2664038400 + }, + { + "epoch": 0.81, + "learning_rate": 0.0001559237713139418, + "loss": 1.3185, + "theoretical_loss": 3.380058461398376, + "tokens_seen": 2664103936 + }, + { + "epoch": 0.81, + "learning_rate": 0.0001559077231695085, + "loss": 1.2841, + "theoretical_loss": 3.380051968324602, + "tokens_seen": 2664169472 + }, + { + "epoch": 0.81, + "learning_rate": 0.0001558916750250752, + "loss": 1.3555, + "theoretical_loss": 3.380045475455271, + "tokens_seen": 2664235008 + }, + { + "epoch": 0.81, + "learning_rate": 0.0001558756268806419, + "loss": 1.2735, + "theoretical_loss": 3.380038982790371, + "tokens_seen": 2664300544 + }, + { + "epoch": 0.81, + "learning_rate": 0.00015585957873620861, + "loss": 1.2653, + "theoretical_loss": 3.380032490329891, + "tokens_seen": 2664366080 + }, + { + "epoch": 0.81, + "learning_rate": 0.00015584353059177534, + "loss": 1.1399, + "theoretical_loss": 3.380025998073819, + "tokens_seen": 2664431616 + }, + { + "epoch": 0.81, + "learning_rate": 0.00015582748244734204, + "loss": 1.1196, + "theoretical_loss": 3.380019506022144, + "tokens_seen": 2664497152 + }, + { + "epoch": 0.81, + "learning_rate": 0.00015581143430290874, + "loss": 1.2988, + "theoretical_loss": 3.380013014174854, + "tokens_seen": 2664562688 + }, + { + "epoch": 0.81, + "learning_rate": 0.00015579538615847544, + "loss": 1.3049, + "theoretical_loss": 3.3800065225319385, + "tokens_seen": 2664628224 + }, + { + "epoch": 0.81, + "learning_rate": 0.00015577933801404214, + "loss": 1.2165, + "theoretical_loss": 3.380000031093385, + "tokens_seen": 2664693760 + }, + { + "epoch": 0.81, + "learning_rate": 0.00015576328986960885, + "loss": 1.3004, + "theoretical_loss": 3.379993539859183, + "tokens_seen": 2664759296 + }, + { + "epoch": 0.81, + "learning_rate": 0.00015574724172517555, + "loss": 1.2289, + "theoretical_loss": 3.37998704882932, + "tokens_seen": 2664824832 + }, + { + "epoch": 0.81, + "learning_rate": 0.00015573119358074225, + "loss": 1.3461, + "theoretical_loss": 3.379980558003785, + "tokens_seen": 2664890368 + }, + { + "epoch": 0.81, + "learning_rate": 0.00015571514543630895, + "loss": 1.3203, + "theoretical_loss": 3.379974067382567, + "tokens_seen": 2664955904 + }, + { + "epoch": 0.81, + "learning_rate": 0.00015569909729187565, + "loss": 1.3113, + "theoretical_loss": 3.3799675769656536, + "tokens_seen": 2665021440 + }, + { + "epoch": 0.81, + "learning_rate": 0.00015568304914744235, + "loss": 1.242, + "theoretical_loss": 3.3799610867530343, + "tokens_seen": 2665086976 + }, + { + "epoch": 0.81, + "learning_rate": 0.00015566700100300905, + "loss": 1.2684, + "theoretical_loss": 3.379954596744697, + "tokens_seen": 2665152512 + }, + { + "epoch": 0.81, + "learning_rate": 0.00015565095285857575, + "loss": 1.2444, + "theoretical_loss": 3.3799481069406307, + "tokens_seen": 2665218048 + }, + { + "epoch": 0.81, + "learning_rate": 0.00015563490471414245, + "loss": 1.2134, + "theoretical_loss": 3.3799416173408234, + "tokens_seen": 2665283584 + }, + { + "epoch": 0.81, + "learning_rate": 0.00015561885656970912, + "loss": 1.1869, + "theoretical_loss": 3.3799351279452647, + "tokens_seen": 2665349120 + }, + { + "epoch": 0.81, + "learning_rate": 0.00015560280842527582, + "loss": 1.2511, + "theoretical_loss": 3.379928638753942, + "tokens_seen": 2665414656 + }, + { + "epoch": 0.81, + "learning_rate": 0.00015558676028084252, + "loss": 1.356, + "theoretical_loss": 3.379922149766844, + "tokens_seen": 2665480192 + }, + { + "epoch": 0.81, + "learning_rate": 0.00015557071213640923, + "loss": 1.2038, + "theoretical_loss": 3.37991566098396, + "tokens_seen": 2665545728 + }, + { + "epoch": 0.81, + "learning_rate": 0.00015555466399197593, + "loss": 1.2682, + "theoretical_loss": 3.3799091724052777, + "tokens_seen": 2665611264 + }, + { + "epoch": 0.81, + "learning_rate": 0.00015553861584754263, + "loss": 1.2778, + "theoretical_loss": 3.379902684030786, + "tokens_seen": 2665676800 + }, + { + "epoch": 0.81, + "learning_rate": 0.00015552256770310933, + "loss": 1.2772, + "theoretical_loss": 3.379896195860474, + "tokens_seen": 2665742336 + }, + { + "epoch": 0.81, + "learning_rate": 0.00015550651955867603, + "loss": 1.2765, + "theoretical_loss": 3.3798897078943293, + "tokens_seen": 2665807872 + }, + { + "epoch": 0.81, + "learning_rate": 0.00015549047141424273, + "loss": 1.2334, + "theoretical_loss": 3.379883220132341, + "tokens_seen": 2665873408 + }, + { + "epoch": 0.81, + "learning_rate": 0.00015547442326980943, + "loss": 1.308, + "theoretical_loss": 3.379876732574498, + "tokens_seen": 2665938944 + }, + { + "epoch": 0.81, + "learning_rate": 0.00015545837512537613, + "loss": 1.3074, + "theoretical_loss": 3.3798702452207876, + "tokens_seen": 2666004480 + }, + { + "epoch": 0.81, + "learning_rate": 0.00015544232698094283, + "loss": 1.3205, + "theoretical_loss": 3.3798637580712, + "tokens_seen": 2666070016 + }, + { + "epoch": 0.81, + "learning_rate": 0.00015542627883650953, + "loss": 1.3512, + "theoretical_loss": 3.379857271125722, + "tokens_seen": 2666135552 + }, + { + "epoch": 0.81, + "learning_rate": 0.00015541023069207623, + "loss": 1.2858, + "theoretical_loss": 3.379850784384344, + "tokens_seen": 2666201088 + }, + { + "epoch": 0.81, + "learning_rate": 0.00015539418254764293, + "loss": 1.2292, + "theoretical_loss": 3.379844297847053, + "tokens_seen": 2666266624 + }, + { + "epoch": 0.81, + "learning_rate": 0.00015537813440320963, + "loss": 1.2143, + "theoretical_loss": 3.379837811513839, + "tokens_seen": 2666332160 + }, + { + "epoch": 0.81, + "learning_rate": 0.00015536208625877633, + "loss": 1.2774, + "theoretical_loss": 3.3798313253846892, + "tokens_seen": 2666397696 + }, + { + "epoch": 0.81, + "learning_rate": 0.00015534603811434303, + "loss": 1.1395, + "theoretical_loss": 3.3798248394595927, + "tokens_seen": 2666463232 + }, + { + "epoch": 0.81, + "learning_rate": 0.00015532998996990974, + "loss": 1.2557, + "theoretical_loss": 3.3798183537385382, + "tokens_seen": 2666528768 + }, + { + "epoch": 0.81, + "learning_rate": 0.00015531394182547644, + "loss": 1.2515, + "theoretical_loss": 3.3798118682215144, + "tokens_seen": 2666594304 + }, + { + "epoch": 0.81, + "learning_rate": 0.00015529789368104314, + "loss": 1.1755, + "theoretical_loss": 3.3798053829085095, + "tokens_seen": 2666659840 + }, + { + "epoch": 0.81, + "learning_rate": 0.00015528184553660984, + "loss": 1.2712, + "theoretical_loss": 3.3797988977995126, + "tokens_seen": 2666725376 + }, + { + "epoch": 0.81, + "learning_rate": 0.00015526579739217654, + "loss": 1.3303, + "theoretical_loss": 3.3797924128945116, + "tokens_seen": 2666790912 + }, + { + "epoch": 0.81, + "learning_rate": 0.00015524974924774324, + "loss": 1.222, + "theoretical_loss": 3.379785928193495, + "tokens_seen": 2666856448 + }, + { + "epoch": 0.81, + "learning_rate": 0.00015523370110330994, + "loss": 1.1875, + "theoretical_loss": 3.379779443696452, + "tokens_seen": 2666921984 + }, + { + "epoch": 0.81, + "learning_rate": 0.00015521765295887664, + "loss": 1.1373, + "theoretical_loss": 3.379772959403371, + "tokens_seen": 2666987520 + }, + { + "epoch": 0.81, + "learning_rate": 0.00015520160481444334, + "loss": 1.2093, + "theoretical_loss": 3.3797664753142405, + "tokens_seen": 2667053056 + }, + { + "epoch": 0.81, + "learning_rate": 0.00015518555667001004, + "loss": 1.2829, + "theoretical_loss": 3.379759991429049, + "tokens_seen": 2667118592 + }, + { + "epoch": 0.81, + "learning_rate": 0.00015516950852557674, + "loss": 1.3059, + "theoretical_loss": 3.379753507747785, + "tokens_seen": 2667184128 + }, + { + "epoch": 0.81, + "learning_rate": 0.00015515346038114344, + "loss": 1.304, + "theoretical_loss": 3.3797470242704373, + "tokens_seen": 2667249664 + }, + { + "epoch": 0.81, + "objective/train/docs_used": 968905, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.2724499702453613, + "objective/train/theoretical_loss": 3.3797405409969947, + "objective/train/tokens_used": 2687775200, + "theoretical_loss": 3.3797405409969947, + "tokens_seen": 2667315200 + }, + { + "epoch": 0.81, + "learning_rate": 0.00015513741223671014, + "loss": 1.2438, + "theoretical_loss": 3.3797405409969947, + "tokens_seen": 2667315200 + }, + { + "epoch": 0.81, + "learning_rate": 0.00015512136409227684, + "loss": 1.2664, + "theoretical_loss": 3.379734057927445, + "tokens_seen": 2667380736 + }, + { + "epoch": 0.81, + "learning_rate": 0.00015510531594784354, + "loss": 1.2522, + "theoretical_loss": 3.3797275750617772, + "tokens_seen": 2667446272 + }, + { + "epoch": 0.81, + "learning_rate": 0.00015508926780341024, + "loss": 1.3338, + "theoretical_loss": 3.37972109239998, + "tokens_seen": 2667511808 + }, + { + "epoch": 0.81, + "learning_rate": 0.00015507321965897695, + "loss": 1.2607, + "theoretical_loss": 3.379714609942042, + "tokens_seen": 2667577344 + }, + { + "epoch": 0.81, + "learning_rate": 0.00015505717151454365, + "loss": 1.2729, + "theoretical_loss": 3.3797081276879517, + "tokens_seen": 2667642880 + }, + { + "epoch": 0.81, + "learning_rate": 0.00015504112337011035, + "loss": 1.2006, + "theoretical_loss": 3.379701645637698, + "tokens_seen": 2667708416 + }, + { + "epoch": 0.81, + "learning_rate": 0.00015502507522567705, + "loss": 1.2128, + "theoretical_loss": 3.3796951637912684, + "tokens_seen": 2667773952 + }, + { + "epoch": 0.81, + "learning_rate": 0.00015500902708124375, + "loss": 1.3796, + "theoretical_loss": 3.3796886821486525, + "tokens_seen": 2667839488 + }, + { + "epoch": 0.81, + "learning_rate": 0.00015499297893681045, + "loss": 1.267, + "theoretical_loss": 3.379682200709839, + "tokens_seen": 2667905024 + }, + { + "epoch": 0.81, + "learning_rate": 0.00015497693079237715, + "loss": 1.297, + "theoretical_loss": 3.3796757194748155, + "tokens_seen": 2667970560 + }, + { + "epoch": 0.81, + "learning_rate": 0.00015496088264794385, + "loss": 1.2461, + "theoretical_loss": 3.3796692384435714, + "tokens_seen": 2668036096 + }, + { + "epoch": 0.81, + "learning_rate": 0.00015494483450351055, + "loss": 1.2976, + "theoretical_loss": 3.379662757616095, + "tokens_seen": 2668101632 + }, + { + "epoch": 0.81, + "learning_rate": 0.00015492878635907725, + "loss": 1.111, + "theoretical_loss": 3.379656276992375, + "tokens_seen": 2668167168 + }, + { + "epoch": 0.81, + "learning_rate": 0.00015491273821464392, + "loss": 1.2378, + "theoretical_loss": 3.3796497965724, + "tokens_seen": 2668232704 + }, + { + "epoch": 0.81, + "learning_rate": 0.00015489669007021063, + "loss": 1.2698, + "theoretical_loss": 3.3796433163561588, + "tokens_seen": 2668298240 + }, + { + "epoch": 0.81, + "learning_rate": 0.00015488064192577733, + "loss": 1.2206, + "theoretical_loss": 3.3796368363436393, + "tokens_seen": 2668363776 + }, + { + "epoch": 0.81, + "learning_rate": 0.00015486459378134403, + "loss": 1.2174, + "theoretical_loss": 3.3796303565348307, + "tokens_seen": 2668429312 + }, + { + "epoch": 0.81, + "learning_rate": 0.00015484854563691073, + "loss": 1.2899, + "theoretical_loss": 3.3796238769297213, + "tokens_seen": 2668494848 + }, + { + "epoch": 0.81, + "learning_rate": 0.00015483249749247743, + "loss": 1.2567, + "theoretical_loss": 3.3796173975282997, + "tokens_seen": 2668560384 + }, + { + "epoch": 0.81, + "learning_rate": 0.00015481644934804413, + "loss": 1.2314, + "theoretical_loss": 3.379610918330555, + "tokens_seen": 2668625920 + }, + { + "epoch": 0.81, + "learning_rate": 0.00015480040120361083, + "loss": 1.3083, + "theoretical_loss": 3.3796044393364753, + "tokens_seen": 2668691456 + }, + { + "epoch": 0.81, + "learning_rate": 0.00015478435305917753, + "loss": 1.3771, + "theoretical_loss": 3.3795979605460493, + "tokens_seen": 2668756992 + }, + { + "epoch": 0.81, + "learning_rate": 0.00015476830491474423, + "loss": 1.2314, + "theoretical_loss": 3.3795914819592654, + "tokens_seen": 2668822528 + }, + { + "epoch": 0.81, + "learning_rate": 0.00015475225677031093, + "loss": 1.3718, + "theoretical_loss": 3.3795850035761124, + "tokens_seen": 2668888064 + }, + { + "epoch": 0.81, + "learning_rate": 0.00015473620862587763, + "loss": 1.3078, + "theoretical_loss": 3.379578525396579, + "tokens_seen": 2668953600 + }, + { + "epoch": 0.81, + "learning_rate": 0.00015472016048144433, + "loss": 1.2137, + "theoretical_loss": 3.3795720474206536, + "tokens_seen": 2669019136 + }, + { + "epoch": 0.81, + "learning_rate": 0.00015470411233701103, + "loss": 1.1813, + "theoretical_loss": 3.379565569648325, + "tokens_seen": 2669084672 + }, + { + "epoch": 0.81, + "learning_rate": 0.00015468806419257773, + "loss": 1.2492, + "theoretical_loss": 3.3795590920795817, + "tokens_seen": 2669150208 + }, + { + "epoch": 0.81, + "learning_rate": 0.00015467201604814443, + "loss": 1.2191, + "theoretical_loss": 3.379552614714412, + "tokens_seen": 2669215744 + }, + { + "epoch": 0.81, + "learning_rate": 0.00015465596790371113, + "loss": 1.3108, + "theoretical_loss": 3.379546137552805, + "tokens_seen": 2669281280 + }, + { + "epoch": 0.81, + "learning_rate": 0.00015463991975927786, + "loss": 1.2177, + "theoretical_loss": 3.3795396605947494, + "tokens_seen": 2669346816 + }, + { + "epoch": 0.81, + "learning_rate": 0.00015462387161484456, + "loss": 1.2674, + "theoretical_loss": 3.379533183840233, + "tokens_seen": 2669412352 + }, + { + "epoch": 0.81, + "learning_rate": 0.00015460782347041126, + "loss": 1.2441, + "theoretical_loss": 3.379526707289245, + "tokens_seen": 2669477888 + }, + { + "epoch": 0.81, + "learning_rate": 0.00015459177532597794, + "loss": 1.2956, + "theoretical_loss": 3.3795202309417745, + "tokens_seen": 2669543424 + }, + { + "epoch": 0.81, + "learning_rate": 0.00015457572718154464, + "loss": 1.2351, + "theoretical_loss": 3.3795137547978094, + "tokens_seen": 2669608960 + }, + { + "epoch": 0.81, + "learning_rate": 0.00015455967903711134, + "loss": 1.2702, + "theoretical_loss": 3.379507278857338, + "tokens_seen": 2669674496 + }, + { + "epoch": 0.81, + "learning_rate": 0.00015454363089267804, + "loss": 1.2748, + "theoretical_loss": 3.3795008031203495, + "tokens_seen": 2669740032 + }, + { + "epoch": 0.81, + "learning_rate": 0.00015452758274824474, + "loss": 1.2325, + "theoretical_loss": 3.3794943275868325, + "tokens_seen": 2669805568 + }, + { + "epoch": 0.81, + "learning_rate": 0.00015451153460381144, + "loss": 1.2271, + "theoretical_loss": 3.3794878522567755, + "tokens_seen": 2669871104 + }, + { + "epoch": 0.81, + "learning_rate": 0.00015449548645937814, + "loss": 1.198, + "theoretical_loss": 3.379481377130167, + "tokens_seen": 2669936640 + }, + { + "epoch": 0.81, + "learning_rate": 0.00015447943831494484, + "loss": 1.24, + "theoretical_loss": 3.3794749022069954, + "tokens_seen": 2670002176 + }, + { + "epoch": 0.81, + "learning_rate": 0.00015446339017051154, + "loss": 1.2617, + "theoretical_loss": 3.37946842748725, + "tokens_seen": 2670067712 + }, + { + "epoch": 0.81, + "learning_rate": 0.00015444734202607824, + "loss": 1.2706, + "theoretical_loss": 3.379461952970919, + "tokens_seen": 2670133248 + }, + { + "epoch": 0.81, + "learning_rate": 0.00015443129388164494, + "loss": 1.1822, + "theoretical_loss": 3.3794554786579907, + "tokens_seen": 2670198784 + }, + { + "epoch": 0.81, + "learning_rate": 0.00015441524573721164, + "loss": 1.3009, + "theoretical_loss": 3.3794490045484546, + "tokens_seen": 2670264320 + }, + { + "epoch": 0.81, + "learning_rate": 0.00015439919759277834, + "loss": 1.1933, + "theoretical_loss": 3.3794425306422986, + "tokens_seen": 2670329856 + }, + { + "epoch": 0.81, + "learning_rate": 0.00015438314944834505, + "loss": 1.2459, + "theoretical_loss": 3.3794360569395114, + "tokens_seen": 2670395392 + }, + { + "epoch": 0.81, + "learning_rate": 0.00015436710130391175, + "loss": 1.2014, + "theoretical_loss": 3.379429583440082, + "tokens_seen": 2670460928 + }, + { + "epoch": 0.81, + "learning_rate": 0.00015435105315947845, + "loss": 1.4017, + "theoretical_loss": 3.3794231101439984, + "tokens_seen": 2670526464 + }, + { + "epoch": 0.81, + "objective/train/docs_used": 970113, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.248903751373291, + "objective/train/theoretical_loss": 3.37941663705125, + "objective/train/tokens_used": 2691052000, + "theoretical_loss": 3.37941663705125, + "tokens_seen": 2670592000 + }, + { + "epoch": 0.81, + "learning_rate": 0.00015433500501504515, + "loss": 1.2051, + "theoretical_loss": 3.37941663705125, + "tokens_seen": 2670592000 + }, + { + "epoch": 0.81, + "learning_rate": 0.00015431895687061185, + "loss": 1.2633, + "theoretical_loss": 3.3794101641618246, + "tokens_seen": 2670657536 + }, + { + "epoch": 0.81, + "learning_rate": 0.00015430290872617855, + "loss": 1.3264, + "theoretical_loss": 3.3794036914757113, + "tokens_seen": 2670723072 + }, + { + "epoch": 0.81, + "learning_rate": 0.00015428686058174525, + "loss": 1.2547, + "theoretical_loss": 3.379397218992899, + "tokens_seen": 2670788608 + }, + { + "epoch": 0.81, + "learning_rate": 0.00015427081243731195, + "loss": 1.2379, + "theoretical_loss": 3.3793907467133755, + "tokens_seen": 2670854144 + }, + { + "epoch": 0.81, + "learning_rate": 0.00015425476429287865, + "loss": 1.3355, + "theoretical_loss": 3.3793842746371303, + "tokens_seen": 2670919680 + }, + { + "epoch": 0.81, + "learning_rate": 0.00015423871614844535, + "loss": 1.2526, + "theoretical_loss": 3.3793778027641515, + "tokens_seen": 2670985216 + }, + { + "epoch": 0.81, + "learning_rate": 0.00015422266800401205, + "loss": 1.2491, + "theoretical_loss": 3.3793713310944278, + "tokens_seen": 2671050752 + }, + { + "epoch": 0.81, + "learning_rate": 0.00015420661985957873, + "loss": 1.3514, + "theoretical_loss": 3.3793648596279477, + "tokens_seen": 2671116288 + }, + { + "epoch": 0.81, + "learning_rate": 0.00015419057171514543, + "loss": 1.2565, + "theoretical_loss": 3.3793583883647003, + "tokens_seen": 2671181824 + }, + { + "epoch": 0.81, + "learning_rate": 0.00015417452357071213, + "loss": 1.3285, + "theoretical_loss": 3.3793519173046738, + "tokens_seen": 2671247360 + }, + { + "epoch": 0.81, + "learning_rate": 0.00015415847542627883, + "loss": 1.2016, + "theoretical_loss": 3.379345446447857, + "tokens_seen": 2671312896 + }, + { + "epoch": 0.81, + "learning_rate": 0.00015414242728184553, + "loss": 1.2677, + "theoretical_loss": 3.379338975794239, + "tokens_seen": 2671378432 + }, + { + "epoch": 0.81, + "learning_rate": 0.00015412637913741223, + "loss": 1.1481, + "theoretical_loss": 3.3793325053438075, + "tokens_seen": 2671443968 + }, + { + "epoch": 0.81, + "learning_rate": 0.00015411033099297893, + "loss": 1.2092, + "theoretical_loss": 3.3793260350965513, + "tokens_seen": 2671509504 + }, + { + "epoch": 0.81, + "learning_rate": 0.00015409428284854563, + "loss": 1.2415, + "theoretical_loss": 3.3793195650524597, + "tokens_seen": 2671575040 + }, + { + "epoch": 0.81, + "learning_rate": 0.00015407823470411233, + "loss": 1.3111, + "theoretical_loss": 3.379313095211521, + "tokens_seen": 2671640576 + }, + { + "epoch": 0.81, + "learning_rate": 0.00015406218655967903, + "loss": 1.1844, + "theoretical_loss": 3.3793066255737236, + "tokens_seen": 2671706112 + }, + { + "epoch": 0.81, + "learning_rate": 0.00015404613841524576, + "loss": 1.2128, + "theoretical_loss": 3.3793001561390565, + "tokens_seen": 2671771648 + }, + { + "epoch": 0.81, + "learning_rate": 0.00015403009027081246, + "loss": 1.2606, + "theoretical_loss": 3.3792936869075083, + "tokens_seen": 2671837184 + }, + { + "epoch": 0.81, + "learning_rate": 0.00015401404212637916, + "loss": 1.2503, + "theoretical_loss": 3.3792872178790674, + "tokens_seen": 2671902720 + }, + { + "epoch": 0.81, + "learning_rate": 0.00015399799398194586, + "loss": 1.3312, + "theoretical_loss": 3.379280749053722, + "tokens_seen": 2671968256 + }, + { + "epoch": 0.81, + "learning_rate": 0.00015398194583751256, + "loss": 1.2743, + "theoretical_loss": 3.379274280431462, + "tokens_seen": 2672033792 + }, + { + "epoch": 0.81, + "learning_rate": 0.00015396589769307926, + "loss": 1.3109, + "theoretical_loss": 3.379267812012275, + "tokens_seen": 2672099328 + }, + { + "epoch": 0.81, + "learning_rate": 0.00015394984954864596, + "loss": 1.1276, + "theoretical_loss": 3.3792613437961503, + "tokens_seen": 2672164864 + }, + { + "epoch": 0.81, + "learning_rate": 0.00015393380140421266, + "loss": 1.347, + "theoretical_loss": 3.3792548757830763, + "tokens_seen": 2672230400 + }, + { + "epoch": 0.81, + "learning_rate": 0.00015391775325977936, + "loss": 1.2711, + "theoretical_loss": 3.3792484079730416, + "tokens_seen": 2672295936 + }, + { + "epoch": 0.81, + "learning_rate": 0.00015390170511534606, + "loss": 1.2378, + "theoretical_loss": 3.3792419403660343, + "tokens_seen": 2672361472 + }, + { + "epoch": 0.81, + "learning_rate": 0.00015388565697091274, + "loss": 1.2251, + "theoretical_loss": 3.379235472962044, + "tokens_seen": 2672427008 + }, + { + "epoch": 0.81, + "learning_rate": 0.00015386960882647944, + "loss": 1.3133, + "theoretical_loss": 3.379229005761059, + "tokens_seen": 2672492544 + }, + { + "epoch": 0.81, + "learning_rate": 0.00015385356068204614, + "loss": 1.2078, + "theoretical_loss": 3.3792225387630674, + "tokens_seen": 2672558080 + }, + { + "epoch": 0.81, + "learning_rate": 0.00015383751253761284, + "loss": 1.2469, + "theoretical_loss": 3.3792160719680586, + "tokens_seen": 2672623616 + }, + { + "epoch": 0.81, + "learning_rate": 0.00015382146439317954, + "loss": 1.2863, + "theoretical_loss": 3.379209605376021, + "tokens_seen": 2672689152 + }, + { + "epoch": 0.81, + "learning_rate": 0.00015380541624874624, + "loss": 1.2621, + "theoretical_loss": 3.3792031389869432, + "tokens_seen": 2672754688 + }, + { + "epoch": 0.81, + "learning_rate": 0.00015378936810431294, + "loss": 1.1869, + "theoretical_loss": 3.379196672800814, + "tokens_seen": 2672820224 + }, + { + "epoch": 0.81, + "learning_rate": 0.00015377331995987964, + "loss": 1.2698, + "theoretical_loss": 3.379190206817622, + "tokens_seen": 2672885760 + }, + { + "epoch": 0.81, + "learning_rate": 0.00015375727181544634, + "loss": 1.3084, + "theoretical_loss": 3.3791837410373553, + "tokens_seen": 2672951296 + }, + { + "epoch": 0.81, + "learning_rate": 0.00015374122367101304, + "loss": 1.2782, + "theoretical_loss": 3.3791772754600036, + "tokens_seen": 2673016832 + }, + { + "epoch": 0.81, + "learning_rate": 0.00015372517552657974, + "loss": 1.1864, + "theoretical_loss": 3.379170810085555, + "tokens_seen": 2673082368 + }, + { + "epoch": 0.81, + "learning_rate": 0.00015370912738214644, + "loss": 1.329, + "theoretical_loss": 3.379164344913998, + "tokens_seen": 2673147904 + }, + { + "epoch": 0.81, + "learning_rate": 0.00015369307923771315, + "loss": 1.3756, + "theoretical_loss": 3.3791578799453212, + "tokens_seen": 2673213440 + }, + { + "epoch": 0.81, + "learning_rate": 0.00015367703109327985, + "loss": 1.2936, + "theoretical_loss": 3.3791514151795137, + "tokens_seen": 2673278976 + }, + { + "epoch": 0.81, + "learning_rate": 0.00015366098294884655, + "loss": 1.236, + "theoretical_loss": 3.3791449506165643, + "tokens_seen": 2673344512 + }, + { + "epoch": 0.81, + "learning_rate": 0.00015364493480441325, + "loss": 1.1397, + "theoretical_loss": 3.3791384862564606, + "tokens_seen": 2673410048 + }, + { + "epoch": 0.81, + "learning_rate": 0.00015362888665997995, + "loss": 1.2968, + "theoretical_loss": 3.3791320220991925, + "tokens_seen": 2673475584 + }, + { + "epoch": 0.81, + "learning_rate": 0.00015361283851554665, + "loss": 1.2832, + "theoretical_loss": 3.379125558144748, + "tokens_seen": 2673541120 + }, + { + "epoch": 0.81, + "learning_rate": 0.00015359679037111335, + "loss": 1.2845, + "theoretical_loss": 3.3791190943931158, + "tokens_seen": 2673606656 + }, + { + "epoch": 0.81, + "learning_rate": 0.00015358074222668005, + "loss": 1.2524, + "theoretical_loss": 3.3791126308442845, + "tokens_seen": 2673672192 + }, + { + "epoch": 0.81, + "learning_rate": 0.00015356469408224675, + "loss": 1.1688, + "theoretical_loss": 3.3791061674982434, + "tokens_seen": 2673737728 + }, + { + "epoch": 0.81, + "learning_rate": 0.00015354864593781345, + "loss": 1.2215, + "theoretical_loss": 3.3790997043549806, + "tokens_seen": 2673803264 + }, + { + "epoch": 0.81, + "objective/train/docs_used": 971344, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.409189224243164, + "objective/train/theoretical_loss": 3.3790932414144845, + "objective/train/tokens_used": 2694328800, + "theoretical_loss": 3.3790932414144845, + "tokens_seen": 2673868800 + }, + { + "epoch": 0.81, + "learning_rate": 0.00015353259779338015, + "loss": 1.4158, + "theoretical_loss": 3.3790932414144845, + "tokens_seen": 2673868800 + }, + { + "epoch": 0.81, + "learning_rate": 0.00015351654964894685, + "loss": 1.352, + "theoretical_loss": 3.3790867786767445, + "tokens_seen": 2673934336 + }, + { + "epoch": 0.81, + "learning_rate": 0.00015350050150451353, + "loss": 1.2976, + "theoretical_loss": 3.379080316141749, + "tokens_seen": 2673999872 + }, + { + "epoch": 0.81, + "learning_rate": 0.00015348445336008023, + "loss": 1.2042, + "theoretical_loss": 3.379073853809486, + "tokens_seen": 2674065408 + }, + { + "epoch": 0.81, + "learning_rate": 0.00015346840521564693, + "loss": 1.3002, + "theoretical_loss": 3.3790673916799454, + "tokens_seen": 2674130944 + }, + { + "epoch": 0.81, + "learning_rate": 0.00015345235707121366, + "loss": 1.2249, + "theoretical_loss": 3.379060929753115, + "tokens_seen": 2674196480 + }, + { + "epoch": 0.81, + "learning_rate": 0.00015343630892678036, + "loss": 1.351, + "theoretical_loss": 3.3790544680289836, + "tokens_seen": 2674262016 + }, + { + "epoch": 0.81, + "learning_rate": 0.00015342026078234706, + "loss": 1.1621, + "theoretical_loss": 3.37904800650754, + "tokens_seen": 2674327552 + }, + { + "epoch": 0.81, + "learning_rate": 0.00015340421263791376, + "loss": 1.1948, + "theoretical_loss": 3.3790415451887728, + "tokens_seen": 2674393088 + }, + { + "epoch": 0.81, + "learning_rate": 0.00015338816449348046, + "loss": 1.2602, + "theoretical_loss": 3.3790350840726706, + "tokens_seen": 2674458624 + }, + { + "epoch": 0.81, + "learning_rate": 0.00015337211634904716, + "loss": 1.2095, + "theoretical_loss": 3.3790286231592224, + "tokens_seen": 2674524160 + }, + { + "epoch": 0.81, + "learning_rate": 0.00015335606820461386, + "loss": 1.2782, + "theoretical_loss": 3.3790221624484165, + "tokens_seen": 2674589696 + }, + { + "epoch": 0.81, + "learning_rate": 0.00015334002006018056, + "loss": 1.1592, + "theoretical_loss": 3.379015701940242, + "tokens_seen": 2674655232 + }, + { + "epoch": 0.81, + "learning_rate": 0.00015332397191574726, + "loss": 1.3172, + "theoretical_loss": 3.3790092416346873, + "tokens_seen": 2674720768 + }, + { + "epoch": 0.81, + "learning_rate": 0.00015330792377131396, + "loss": 1.2888, + "theoretical_loss": 3.379002781531741, + "tokens_seen": 2674786304 + }, + { + "epoch": 0.81, + "learning_rate": 0.00015329187562688066, + "loss": 1.2379, + "theoretical_loss": 3.378996321631392, + "tokens_seen": 2674851840 + }, + { + "epoch": 0.81, + "learning_rate": 0.00015327582748244736, + "loss": 1.3855, + "theoretical_loss": 3.3789898619336287, + "tokens_seen": 2674917376 + }, + { + "epoch": 0.81, + "learning_rate": 0.00015325977933801406, + "loss": 1.2152, + "theoretical_loss": 3.37898340243844, + "tokens_seen": 2674982912 + }, + { + "epoch": 0.81, + "learning_rate": 0.00015324373119358076, + "loss": 1.2844, + "theoretical_loss": 3.378976943145815, + "tokens_seen": 2675048448 + }, + { + "epoch": 0.81, + "learning_rate": 0.00015322768304914746, + "loss": 1.3748, + "theoretical_loss": 3.3789704840557415, + "tokens_seen": 2675113984 + }, + { + "epoch": 0.81, + "learning_rate": 0.00015321163490471416, + "loss": 1.2004, + "theoretical_loss": 3.3789640251682087, + "tokens_seen": 2675179520 + }, + { + "epoch": 0.81, + "learning_rate": 0.00015319558676028084, + "loss": 1.2889, + "theoretical_loss": 3.378957566483205, + "tokens_seen": 2675245056 + }, + { + "epoch": 0.81, + "learning_rate": 0.00015317953861584754, + "loss": 1.2809, + "theoretical_loss": 3.3789511080007197, + "tokens_seen": 2675310592 + }, + { + "epoch": 0.81, + "learning_rate": 0.00015316349047141424, + "loss": 1.2766, + "theoretical_loss": 3.378944649720741, + "tokens_seen": 2675376128 + }, + { + "epoch": 0.81, + "learning_rate": 0.00015314744232698094, + "loss": 1.2145, + "theoretical_loss": 3.3789381916432575, + "tokens_seen": 2675441664 + }, + { + "epoch": 0.81, + "learning_rate": 0.00015313139418254764, + "loss": 1.2232, + "theoretical_loss": 3.3789317337682583, + "tokens_seen": 2675507200 + }, + { + "epoch": 0.81, + "learning_rate": 0.00015311534603811434, + "loss": 1.2672, + "theoretical_loss": 3.378925276095732, + "tokens_seen": 2675572736 + }, + { + "epoch": 0.81, + "learning_rate": 0.00015309929789368104, + "loss": 1.3833, + "theoretical_loss": 3.3789188186256665, + "tokens_seen": 2675638272 + }, + { + "epoch": 0.81, + "learning_rate": 0.00015308324974924774, + "loss": 1.1975, + "theoretical_loss": 3.3789123613580516, + "tokens_seen": 2675703808 + }, + { + "epoch": 0.81, + "learning_rate": 0.00015306720160481444, + "loss": 1.1562, + "theoretical_loss": 3.3789059042928753, + "tokens_seen": 2675769344 + }, + { + "epoch": 0.81, + "learning_rate": 0.00015305115346038114, + "loss": 1.152, + "theoretical_loss": 3.378899447430127, + "tokens_seen": 2675834880 + }, + { + "epoch": 0.81, + "learning_rate": 0.00015303510531594784, + "loss": 1.2895, + "theoretical_loss": 3.378892990769794, + "tokens_seen": 2675900416 + }, + { + "epoch": 0.81, + "learning_rate": 0.00015301905717151455, + "loss": 1.2466, + "theoretical_loss": 3.378886534311867, + "tokens_seen": 2675965952 + }, + { + "epoch": 0.81, + "learning_rate": 0.00015300300902708125, + "loss": 1.2274, + "theoretical_loss": 3.378880078056333, + "tokens_seen": 2676031488 + }, + { + "epoch": 0.81, + "learning_rate": 0.00015298696088264795, + "loss": 1.2217, + "theoretical_loss": 3.378873622003182, + "tokens_seen": 2676097024 + }, + { + "epoch": 0.81, + "learning_rate": 0.00015297091273821465, + "loss": 1.2923, + "theoretical_loss": 3.3788671661524012, + "tokens_seen": 2676162560 + }, + { + "epoch": 0.81, + "learning_rate": 0.00015295486459378135, + "loss": 1.2469, + "theoretical_loss": 3.3788607105039805, + "tokens_seen": 2676228096 + }, + { + "epoch": 0.81, + "learning_rate": 0.00015293881644934805, + "loss": 1.3415, + "theoretical_loss": 3.378854255057908, + "tokens_seen": 2676293632 + }, + { + "epoch": 0.81, + "learning_rate": 0.00015292276830491475, + "loss": 1.326, + "theoretical_loss": 3.378847799814173, + "tokens_seen": 2676359168 + }, + { + "epoch": 0.81, + "learning_rate": 0.00015290672016048145, + "loss": 1.3517, + "theoretical_loss": 3.3788413447727637, + "tokens_seen": 2676424704 + }, + { + "epoch": 0.81, + "learning_rate": 0.00015289067201604815, + "loss": 1.2461, + "theoretical_loss": 3.378834889933669, + "tokens_seen": 2676490240 + }, + { + "epoch": 0.81, + "learning_rate": 0.00015287462387161485, + "loss": 1.3106, + "theoretical_loss": 3.378828435296877, + "tokens_seen": 2676555776 + }, + { + "epoch": 0.81, + "learning_rate": 0.00015285857572718155, + "loss": 1.2635, + "theoretical_loss": 3.378821980862378, + "tokens_seen": 2676621312 + }, + { + "epoch": 0.81, + "learning_rate": 0.00015284252758274825, + "loss": 1.3235, + "theoretical_loss": 3.3788155266301585, + "tokens_seen": 2676686848 + }, + { + "epoch": 0.81, + "learning_rate": 0.00015282647943831495, + "loss": 1.2464, + "theoretical_loss": 3.378809072600209, + "tokens_seen": 2676752384 + }, + { + "epoch": 0.81, + "learning_rate": 0.00015281043129388165, + "loss": 1.3748, + "theoretical_loss": 3.3788026187725175, + "tokens_seen": 2676817920 + }, + { + "epoch": 0.81, + "learning_rate": 0.00015279438314944835, + "loss": 1.3286, + "theoretical_loss": 3.378796165147073, + "tokens_seen": 2676883456 + }, + { + "epoch": 0.81, + "learning_rate": 0.00015277833500501505, + "loss": 1.3018, + "theoretical_loss": 3.3787897117238637, + "tokens_seen": 2676948992 + }, + { + "epoch": 0.81, + "learning_rate": 0.00015276228686058176, + "loss": 1.3154, + "theoretical_loss": 3.378783258502879, + "tokens_seen": 2677014528 + }, + { + "epoch": 0.81, + "learning_rate": 0.00015274623871614846, + "loss": 1.371, + "theoretical_loss": 3.3787768054841067, + "tokens_seen": 2677080064 + }, + { + "epoch": 0.81, + "objective/train/docs_used": 972612, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.336362361907959, + "objective/train/theoretical_loss": 3.3787703526675363, + "objective/train/tokens_used": 2697605600, + "theoretical_loss": 3.3787703526675363, + "tokens_seen": 2677145600 + }, + { + "epoch": 0.81, + "learning_rate": 0.00015273019057171516, + "loss": 1.2885, + "theoretical_loss": 3.3787703526675363, + "tokens_seen": 2677145600 + }, + { + "epoch": 0.81, + "learning_rate": 0.00015271414242728186, + "loss": 1.235, + "theoretical_loss": 3.378763900053156, + "tokens_seen": 2677211136 + }, + { + "epoch": 0.81, + "learning_rate": 0.00015269809428284856, + "loss": 1.303, + "theoretical_loss": 3.3787574476409548, + "tokens_seen": 2677276672 + }, + { + "epoch": 0.81, + "learning_rate": 0.00015268204613841526, + "loss": 1.2443, + "theoretical_loss": 3.378750995430922, + "tokens_seen": 2677342208 + }, + { + "epoch": 0.81, + "learning_rate": 0.00015266599799398196, + "loss": 1.2493, + "theoretical_loss": 3.378744543423045, + "tokens_seen": 2677407744 + }, + { + "epoch": 0.81, + "learning_rate": 0.00015264994984954866, + "loss": 1.3231, + "theoretical_loss": 3.3787380916173135, + "tokens_seen": 2677473280 + }, + { + "epoch": 0.81, + "learning_rate": 0.00015263390170511536, + "loss": 1.2107, + "theoretical_loss": 3.378731640013716, + "tokens_seen": 2677538816 + }, + { + "epoch": 0.81, + "learning_rate": 0.00015261785356068206, + "loss": 1.2936, + "theoretical_loss": 3.3787251886122407, + "tokens_seen": 2677604352 + }, + { + "epoch": 0.81, + "learning_rate": 0.00015260180541624876, + "loss": 1.2846, + "theoretical_loss": 3.3787187374128775, + "tokens_seen": 2677669888 + }, + { + "epoch": 0.81, + "learning_rate": 0.00015258575727181546, + "loss": 1.2789, + "theoretical_loss": 3.378712286415614, + "tokens_seen": 2677735424 + }, + { + "epoch": 0.81, + "learning_rate": 0.00015256970912738216, + "loss": 1.2473, + "theoretical_loss": 3.3787058356204396, + "tokens_seen": 2677800960 + }, + { + "epoch": 0.81, + "learning_rate": 0.00015255366098294886, + "loss": 1.1953, + "theoretical_loss": 3.3786993850273426, + "tokens_seen": 2677866496 + }, + { + "epoch": 0.81, + "learning_rate": 0.00015253761283851556, + "loss": 1.3923, + "theoretical_loss": 3.3786929346363115, + "tokens_seen": 2677932032 + }, + { + "epoch": 0.81, + "learning_rate": 0.00015252156469408226, + "loss": 1.2242, + "theoretical_loss": 3.3786864844473357, + "tokens_seen": 2677997568 + }, + { + "epoch": 0.81, + "learning_rate": 0.00015250551654964897, + "loss": 1.2472, + "theoretical_loss": 3.3786800344604035, + "tokens_seen": 2678063104 + }, + { + "epoch": 0.81, + "learning_rate": 0.00015248946840521564, + "loss": 1.2654, + "theoretical_loss": 3.3786735846755036, + "tokens_seen": 2678128640 + }, + { + "epoch": 0.81, + "learning_rate": 0.00015247342026078234, + "loss": 1.3105, + "theoretical_loss": 3.378667135092625, + "tokens_seen": 2678194176 + }, + { + "epoch": 0.81, + "learning_rate": 0.00015245737211634904, + "loss": 1.2718, + "theoretical_loss": 3.3786606857117567, + "tokens_seen": 2678259712 + }, + { + "epoch": 0.81, + "learning_rate": 0.00015244132397191574, + "loss": 1.2612, + "theoretical_loss": 3.3786542365328867, + "tokens_seen": 2678325248 + }, + { + "epoch": 0.81, + "learning_rate": 0.00015242527582748244, + "loss": 1.3937, + "theoretical_loss": 3.3786477875560044, + "tokens_seen": 2678390784 + }, + { + "epoch": 0.81, + "learning_rate": 0.00015240922768304914, + "loss": 1.2652, + "theoretical_loss": 3.378641338781098, + "tokens_seen": 2678456320 + }, + { + "epoch": 0.81, + "learning_rate": 0.00015239317953861584, + "loss": 1.213, + "theoretical_loss": 3.378634890208156, + "tokens_seen": 2678521856 + }, + { + "epoch": 0.81, + "learning_rate": 0.00015237713139418254, + "loss": 1.3151, + "theoretical_loss": 3.378628441837168, + "tokens_seen": 2678587392 + }, + { + "epoch": 0.81, + "learning_rate": 0.00015236108324974924, + "loss": 1.4296, + "theoretical_loss": 3.3786219936681223, + "tokens_seen": 2678652928 + }, + { + "epoch": 0.81, + "learning_rate": 0.00015234503510531594, + "loss": 1.2619, + "theoretical_loss": 3.3786155457010074, + "tokens_seen": 2678718464 + }, + { + "epoch": 0.81, + "learning_rate": 0.00015232898696088265, + "loss": 1.2615, + "theoretical_loss": 3.3786090979358123, + "tokens_seen": 2678784000 + }, + { + "epoch": 0.81, + "learning_rate": 0.00015231293881644935, + "loss": 1.3919, + "theoretical_loss": 3.378602650372526, + "tokens_seen": 2678849536 + }, + { + "epoch": 0.81, + "learning_rate": 0.00015229689067201605, + "loss": 1.288, + "theoretical_loss": 3.3785962030111367, + "tokens_seen": 2678915072 + }, + { + "epoch": 0.81, + "learning_rate": 0.00015228084252758275, + "loss": 1.3437, + "theoretical_loss": 3.3785897558516336, + "tokens_seen": 2678980608 + }, + { + "epoch": 0.81, + "learning_rate": 0.00015226479438314945, + "loss": 1.2584, + "theoretical_loss": 3.378583308894005, + "tokens_seen": 2679046144 + }, + { + "epoch": 0.81, + "learning_rate": 0.00015224874623871618, + "loss": 1.3952, + "theoretical_loss": 3.37857686213824, + "tokens_seen": 2679111680 + }, + { + "epoch": 0.81, + "learning_rate": 0.00015223269809428288, + "loss": 1.3478, + "theoretical_loss": 3.3785704155843272, + "tokens_seen": 2679177216 + }, + { + "epoch": 0.81, + "learning_rate": 0.00015221664994984958, + "loss": 1.2001, + "theoretical_loss": 3.378563969232255, + "tokens_seen": 2679242752 + }, + { + "epoch": 0.81, + "learning_rate": 0.00015220060180541628, + "loss": 1.2502, + "theoretical_loss": 3.378557523082013, + "tokens_seen": 2679308288 + }, + { + "epoch": 0.81, + "learning_rate": 0.00015218455366098298, + "loss": 1.2559, + "theoretical_loss": 3.378551077133589, + "tokens_seen": 2679373824 + }, + { + "epoch": 0.81, + "learning_rate": 0.00015216850551654965, + "loss": 1.227, + "theoretical_loss": 3.3785446313869727, + "tokens_seen": 2679439360 + }, + { + "epoch": 0.81, + "learning_rate": 0.00015215245737211635, + "loss": 1.2952, + "theoretical_loss": 3.378538185842152, + "tokens_seen": 2679504896 + }, + { + "epoch": 0.81, + "learning_rate": 0.00015213640922768305, + "loss": 1.3015, + "theoretical_loss": 3.378531740499116, + "tokens_seen": 2679570432 + }, + { + "epoch": 0.81, + "learning_rate": 0.00015212036108324975, + "loss": 1.4263, + "theoretical_loss": 3.3785252953578535, + "tokens_seen": 2679635968 + }, + { + "epoch": 0.81, + "learning_rate": 0.00015210431293881645, + "loss": 1.2738, + "theoretical_loss": 3.378518850418353, + "tokens_seen": 2679701504 + }, + { + "epoch": 0.81, + "learning_rate": 0.00015208826479438315, + "loss": 1.1938, + "theoretical_loss": 3.3785124056806035, + "tokens_seen": 2679767040 + }, + { + "epoch": 0.81, + "learning_rate": 0.00015207221664994986, + "loss": 1.253, + "theoretical_loss": 3.378505961144594, + "tokens_seen": 2679832576 + }, + { + "epoch": 0.81, + "learning_rate": 0.00015205616850551656, + "loss": 1.3208, + "theoretical_loss": 3.3784995168103125, + "tokens_seen": 2679898112 + }, + { + "epoch": 0.81, + "learning_rate": 0.00015204012036108326, + "loss": 1.1541, + "theoretical_loss": 3.378493072677748, + "tokens_seen": 2679963648 + }, + { + "epoch": 0.81, + "learning_rate": 0.00015202407221664996, + "loss": 1.2732, + "theoretical_loss": 3.37848662874689, + "tokens_seen": 2680029184 + }, + { + "epoch": 0.81, + "learning_rate": 0.00015200802407221666, + "loss": 1.2956, + "theoretical_loss": 3.378480185017726, + "tokens_seen": 2680094720 + }, + { + "epoch": 0.81, + "learning_rate": 0.00015199197592778336, + "loss": 1.3834, + "theoretical_loss": 3.378473741490246, + "tokens_seen": 2680160256 + }, + { + "epoch": 0.81, + "learning_rate": 0.00015197592778335006, + "loss": 1.2867, + "theoretical_loss": 3.3784672981644386, + "tokens_seen": 2680225792 + }, + { + "epoch": 0.81, + "learning_rate": 0.00015195987963891676, + "loss": 1.2555, + "theoretical_loss": 3.3784608550402915, + "tokens_seen": 2680291328 + }, + { + "epoch": 0.81, + "learning_rate": 0.00015194383149448346, + "loss": 1.2775, + "theoretical_loss": 3.378454412117794, + "tokens_seen": 2680356864 + }, + { + "epoch": 0.81, + "objective/train/docs_used": 973756, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.4277327060699463, + "objective/train/theoretical_loss": 3.378447969396935, + "objective/train/tokens_used": 2700882400, + "theoretical_loss": 3.378447969396935, + "tokens_seen": 2680422400 + }, + { + "epoch": 0.81, + "learning_rate": 0.00015192778335005016, + "loss": 1.3366, + "theoretical_loss": 3.378447969396935, + "tokens_seen": 2680422400 + }, + { + "epoch": 0.81, + "learning_rate": 0.00015191173520561686, + "loss": 1.2338, + "theoretical_loss": 3.3784415268777037, + "tokens_seen": 2680487936 + }, + { + "epoch": 0.81, + "learning_rate": 0.00015189568706118356, + "loss": 1.1666, + "theoretical_loss": 3.3784350845600883, + "tokens_seen": 2680553472 + }, + { + "epoch": 0.81, + "learning_rate": 0.00015187963891675026, + "loss": 1.2758, + "theoretical_loss": 3.378428642444077, + "tokens_seen": 2680619008 + }, + { + "epoch": 0.81, + "learning_rate": 0.00015186359077231696, + "loss": 1.3045, + "theoretical_loss": 3.37842220052966, + "tokens_seen": 2680684544 + }, + { + "epoch": 0.81, + "learning_rate": 0.00015184754262788366, + "loss": 1.2282, + "theoretical_loss": 3.378415758816825, + "tokens_seen": 2680750080 + }, + { + "epoch": 0.81, + "learning_rate": 0.00015183149448345037, + "loss": 1.2278, + "theoretical_loss": 3.378409317305561, + "tokens_seen": 2680815616 + }, + { + "epoch": 0.81, + "learning_rate": 0.00015181544633901707, + "loss": 1.2638, + "theoretical_loss": 3.3784028759958566, + "tokens_seen": 2680881152 + }, + { + "epoch": 0.81, + "learning_rate": 0.00015179939819458377, + "loss": 1.2778, + "theoretical_loss": 3.378396434887701, + "tokens_seen": 2680946688 + }, + { + "epoch": 0.81, + "learning_rate": 0.00015178335005015044, + "loss": 1.3264, + "theoretical_loss": 3.3783899939810826, + "tokens_seen": 2681012224 + }, + { + "epoch": 0.81, + "learning_rate": 0.00015176730190571714, + "loss": 1.2848, + "theoretical_loss": 3.3783835532759903, + "tokens_seen": 2681077760 + }, + { + "epoch": 0.81, + "learning_rate": 0.00015175125376128384, + "loss": 1.3202, + "theoretical_loss": 3.378377112772413, + "tokens_seen": 2681143296 + }, + { + "epoch": 0.81, + "learning_rate": 0.00015173520561685054, + "loss": 1.4095, + "theoretical_loss": 3.3783706724703393, + "tokens_seen": 2681208832 + }, + { + "epoch": 0.81, + "learning_rate": 0.00015171915747241724, + "loss": 1.2317, + "theoretical_loss": 3.3783642323697585, + "tokens_seen": 2681274368 + }, + { + "epoch": 0.81, + "learning_rate": 0.00015170310932798394, + "loss": 1.2377, + "theoretical_loss": 3.3783577924706583, + "tokens_seen": 2681339904 + }, + { + "epoch": 0.81, + "learning_rate": 0.00015168706118355064, + "loss": 1.2688, + "theoretical_loss": 3.378351352773028, + "tokens_seen": 2681405440 + }, + { + "epoch": 0.81, + "learning_rate": 0.00015167101303911734, + "loss": 1.3109, + "theoretical_loss": 3.3783449132768566, + "tokens_seen": 2681470976 + }, + { + "epoch": 0.81, + "learning_rate": 0.00015165496489468404, + "loss": 1.2506, + "theoretical_loss": 3.378338473982133, + "tokens_seen": 2681536512 + }, + { + "epoch": 0.81, + "learning_rate": 0.00015163891675025077, + "loss": 1.3105, + "theoretical_loss": 3.378332034888845, + "tokens_seen": 2681602048 + }, + { + "epoch": 0.81, + "learning_rate": 0.00015162286860581747, + "loss": 1.3348, + "theoretical_loss": 3.3783255959969827, + "tokens_seen": 2681667584 + }, + { + "epoch": 0.81, + "learning_rate": 0.00015160682046138417, + "loss": 1.2146, + "theoretical_loss": 3.378319157306534, + "tokens_seen": 2681733120 + }, + { + "epoch": 0.81, + "learning_rate": 0.00015159077231695087, + "loss": 1.2358, + "theoretical_loss": 3.378312718817488, + "tokens_seen": 2681798656 + }, + { + "epoch": 0.81, + "learning_rate": 0.00015157472417251758, + "loss": 1.2451, + "theoretical_loss": 3.3783062805298334, + "tokens_seen": 2681864192 + }, + { + "epoch": 0.81, + "learning_rate": 0.00015155867602808428, + "loss": 1.2481, + "theoretical_loss": 3.3782998424435586, + "tokens_seen": 2681929728 + }, + { + "epoch": 0.81, + "learning_rate": 0.00015154262788365098, + "loss": 1.2241, + "theoretical_loss": 3.378293404558653, + "tokens_seen": 2681995264 + }, + { + "epoch": 0.81, + "learning_rate": 0.00015152657973921768, + "loss": 1.3082, + "theoretical_loss": 3.3782869668751054, + "tokens_seen": 2682060800 + }, + { + "epoch": 0.81, + "learning_rate": 0.00015151053159478438, + "loss": 1.3734, + "theoretical_loss": 3.378280529392904, + "tokens_seen": 2682126336 + }, + { + "epoch": 0.81, + "learning_rate": 0.00015149448345035108, + "loss": 1.2354, + "theoretical_loss": 3.3782740921120378, + "tokens_seen": 2682191872 + }, + { + "epoch": 0.81, + "learning_rate": 0.00015147843530591775, + "loss": 1.3373, + "theoretical_loss": 3.378267655032496, + "tokens_seen": 2682257408 + }, + { + "epoch": 0.81, + "learning_rate": 0.00015146238716148445, + "loss": 1.3612, + "theoretical_loss": 3.378261218154267, + "tokens_seen": 2682322944 + }, + { + "epoch": 0.81, + "learning_rate": 0.00015144633901705115, + "loss": 1.2859, + "theoretical_loss": 3.3782547814773394, + "tokens_seen": 2682388480 + }, + { + "epoch": 0.81, + "learning_rate": 0.00015143029087261785, + "loss": 1.269, + "theoretical_loss": 3.3782483450017025, + "tokens_seen": 2682454016 + }, + { + "epoch": 0.81, + "learning_rate": 0.00015141424272818455, + "loss": 1.2445, + "theoretical_loss": 3.3782419087273445, + "tokens_seen": 2682519552 + }, + { + "epoch": 0.81, + "learning_rate": 0.00015139819458375126, + "loss": 1.1861, + "theoretical_loss": 3.378235472654255, + "tokens_seen": 2682585088 + }, + { + "epoch": 0.81, + "learning_rate": 0.00015138214643931796, + "loss": 1.3595, + "theoretical_loss": 3.3782290367824217, + "tokens_seen": 2682650624 + }, + { + "epoch": 0.81, + "learning_rate": 0.00015136609829488466, + "loss": 1.2708, + "theoretical_loss": 3.3782226011118346, + "tokens_seen": 2682716160 + }, + { + "epoch": 0.81, + "learning_rate": 0.00015135005015045136, + "loss": 1.1814, + "theoretical_loss": 3.3782161656424816, + "tokens_seen": 2682781696 + }, + { + "epoch": 0.81, + "learning_rate": 0.00015133400200601806, + "loss": 1.2956, + "theoretical_loss": 3.3782097303743517, + "tokens_seen": 2682847232 + }, + { + "epoch": 0.81, + "learning_rate": 0.00015131795386158476, + "loss": 1.2735, + "theoretical_loss": 3.3782032953074337, + "tokens_seen": 2682912768 + }, + { + "epoch": 0.81, + "learning_rate": 0.00015130190571715146, + "loss": 1.2473, + "theoretical_loss": 3.378196860441717, + "tokens_seen": 2682978304 + }, + { + "epoch": 0.81, + "learning_rate": 0.00015128585757271816, + "loss": 1.317, + "theoretical_loss": 3.3781904257771895, + "tokens_seen": 2683043840 + }, + { + "epoch": 0.81, + "learning_rate": 0.00015126980942828486, + "loss": 1.3085, + "theoretical_loss": 3.3781839913138403, + "tokens_seen": 2683109376 + }, + { + "epoch": 0.81, + "learning_rate": 0.00015125376128385156, + "loss": 1.3237, + "theoretical_loss": 3.378177557051658, + "tokens_seen": 2683174912 + }, + { + "epoch": 0.81, + "learning_rate": 0.00015123771313941826, + "loss": 1.2314, + "theoretical_loss": 3.378171122990632, + "tokens_seen": 2683240448 + }, + { + "epoch": 0.81, + "learning_rate": 0.00015122166499498496, + "loss": 1.2374, + "theoretical_loss": 3.3781646891307506, + "tokens_seen": 2683305984 + }, + { + "epoch": 0.81, + "learning_rate": 0.00015120561685055166, + "loss": 1.2623, + "theoretical_loss": 3.378158255472003, + "tokens_seen": 2683371520 + }, + { + "epoch": 0.81, + "learning_rate": 0.00015118956870611836, + "loss": 1.2971, + "theoretical_loss": 3.378151822014378, + "tokens_seen": 2683437056 + }, + { + "epoch": 0.81, + "learning_rate": 0.00015117352056168506, + "loss": 1.2212, + "theoretical_loss": 3.378145388757863, + "tokens_seen": 2683502592 + }, + { + "epoch": 0.81, + "learning_rate": 0.00015115747241725176, + "loss": 1.2356, + "theoretical_loss": 3.378138955702449, + "tokens_seen": 2683568128 + }, + { + "epoch": 0.81, + "learning_rate": 0.00015114142427281847, + "loss": 1.1751, + "theoretical_loss": 3.378132522848124, + "tokens_seen": 2683633664 + }, + { + "epoch": 0.81, + "objective/train/docs_used": 974889, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.277236819267273, + "objective/train/theoretical_loss": 3.3781260901948755, + "objective/train/tokens_used": 2704159200, + "theoretical_loss": 3.3781260901948755, + "tokens_seen": 2683699200 + }, + { + "epoch": 0.81, + "learning_rate": 0.00015112537612838517, + "loss": 1.2626, + "theoretical_loss": 3.3781260901948755, + "tokens_seen": 2683699200 + }, + { + "epoch": 0.81, + "learning_rate": 0.00015110932798395187, + "loss": 1.3654, + "theoretical_loss": 3.378119657742694, + "tokens_seen": 2683764736 + }, + { + "epoch": 0.81, + "learning_rate": 0.00015109327983951854, + "loss": 1.2632, + "theoretical_loss": 3.3781132254915676, + "tokens_seen": 2683830272 + }, + { + "epoch": 0.81, + "learning_rate": 0.00015107723169508524, + "loss": 1.2705, + "theoretical_loss": 3.378106793441485, + "tokens_seen": 2683895808 + }, + { + "epoch": 0.81, + "learning_rate": 0.00015106118355065194, + "loss": 1.3839, + "theoretical_loss": 3.3781003615924354, + "tokens_seen": 2683961344 + }, + { + "epoch": 0.81, + "learning_rate": 0.00015104513540621867, + "loss": 1.3141, + "theoretical_loss": 3.378093929944407, + "tokens_seen": 2684026880 + }, + { + "epoch": 0.81, + "learning_rate": 0.00015102908726178537, + "loss": 1.3106, + "theoretical_loss": 3.3780874984973894, + "tokens_seen": 2684092416 + }, + { + "epoch": 0.81, + "learning_rate": 0.00015101303911735207, + "loss": 1.305, + "theoretical_loss": 3.378081067251371, + "tokens_seen": 2684157952 + }, + { + "epoch": 0.81, + "learning_rate": 0.00015099699097291877, + "loss": 1.3178, + "theoretical_loss": 3.3780746362063403, + "tokens_seen": 2684223488 + }, + { + "epoch": 0.81, + "learning_rate": 0.00015098094282848547, + "loss": 1.2836, + "theoretical_loss": 3.3780682053622866, + "tokens_seen": 2684289024 + }, + { + "epoch": 0.81, + "learning_rate": 0.00015096489468405217, + "loss": 1.3242, + "theoretical_loss": 3.3780617747191988, + "tokens_seen": 2684354560 + }, + { + "epoch": 0.81, + "learning_rate": 0.00015094884653961887, + "loss": 1.2763, + "theoretical_loss": 3.378055344277065, + "tokens_seen": 2684420096 + }, + { + "epoch": 0.81, + "learning_rate": 0.00015093279839518557, + "loss": 1.2055, + "theoretical_loss": 3.3780489140358747, + "tokens_seen": 2684485632 + }, + { + "epoch": 0.81, + "learning_rate": 0.00015091675025075227, + "loss": 1.3316, + "theoretical_loss": 3.3780424839956167, + "tokens_seen": 2684551168 + }, + { + "epoch": 0.81, + "learning_rate": 0.00015090070210631897, + "loss": 1.2717, + "theoretical_loss": 3.3780360541562793, + "tokens_seen": 2684616704 + }, + { + "epoch": 0.81, + "learning_rate": 0.00015088465396188568, + "loss": 1.3561, + "theoretical_loss": 3.3780296245178514, + "tokens_seen": 2684682240 + }, + { + "epoch": 0.81, + "learning_rate": 0.00015086860581745238, + "loss": 1.2396, + "theoretical_loss": 3.3780231950803223, + "tokens_seen": 2684747776 + }, + { + "epoch": 0.81, + "learning_rate": 0.00015085255767301908, + "loss": 1.2659, + "theoretical_loss": 3.3780167658436806, + "tokens_seen": 2684813312 + }, + { + "epoch": 0.81, + "learning_rate": 0.00015083650952858578, + "loss": 1.2394, + "theoretical_loss": 3.3780103368079155, + "tokens_seen": 2684878848 + }, + { + "epoch": 0.81, + "learning_rate": 0.00015082046138415248, + "loss": 1.2765, + "theoretical_loss": 3.3780039079730146, + "tokens_seen": 2684944384 + }, + { + "epoch": 0.81, + "learning_rate": 0.00015080441323971918, + "loss": 1.3064, + "theoretical_loss": 3.377997479338968, + "tokens_seen": 2685009920 + }, + { + "epoch": 0.81, + "learning_rate": 0.00015078836509528588, + "loss": 1.3377, + "theoretical_loss": 3.377991050905764, + "tokens_seen": 2685075456 + }, + { + "epoch": 0.81, + "learning_rate": 0.00015077231695085255, + "loss": 1.25, + "theoretical_loss": 3.3779846226733916, + "tokens_seen": 2685140992 + }, + { + "epoch": 0.81, + "learning_rate": 0.00015075626880641925, + "loss": 1.2899, + "theoretical_loss": 3.3779781946418392, + "tokens_seen": 2685206528 + }, + { + "epoch": 0.81, + "learning_rate": 0.00015074022066198595, + "loss": 1.2902, + "theoretical_loss": 3.377971766811096, + "tokens_seen": 2685272064 + }, + { + "epoch": 0.81, + "learning_rate": 0.00015072417251755265, + "loss": 1.2516, + "theoretical_loss": 3.3779653391811504, + "tokens_seen": 2685337600 + }, + { + "epoch": 0.81, + "learning_rate": 0.00015070812437311936, + "loss": 1.296, + "theoretical_loss": 3.377958911751992, + "tokens_seen": 2685403136 + }, + { + "epoch": 0.81, + "learning_rate": 0.00015069207622868606, + "loss": 1.2474, + "theoretical_loss": 3.3779524845236093, + "tokens_seen": 2685468672 + }, + { + "epoch": 0.81, + "learning_rate": 0.00015067602808425276, + "loss": 1.2079, + "theoretical_loss": 3.3779460574959908, + "tokens_seen": 2685534208 + }, + { + "epoch": 0.81, + "learning_rate": 0.00015065997993981946, + "loss": 1.1825, + "theoretical_loss": 3.3779396306691254, + "tokens_seen": 2685599744 + }, + { + "epoch": 0.81, + "learning_rate": 0.00015064393179538616, + "loss": 1.2846, + "theoretical_loss": 3.3779332040430026, + "tokens_seen": 2685665280 + }, + { + "epoch": 0.81, + "learning_rate": 0.00015062788365095286, + "loss": 1.2925, + "theoretical_loss": 3.3779267776176103, + "tokens_seen": 2685730816 + }, + { + "epoch": 0.81, + "learning_rate": 0.00015061183550651956, + "loss": 1.3688, + "theoretical_loss": 3.377920351392938, + "tokens_seen": 2685796352 + }, + { + "epoch": 0.81, + "learning_rate": 0.00015059578736208626, + "loss": 1.2541, + "theoretical_loss": 3.377913925368974, + "tokens_seen": 2685861888 + }, + { + "epoch": 0.81, + "learning_rate": 0.00015057973921765296, + "loss": 1.2209, + "theoretical_loss": 3.3779074995457075, + "tokens_seen": 2685927424 + }, + { + "epoch": 0.81, + "learning_rate": 0.00015056369107321966, + "loss": 1.2726, + "theoretical_loss": 3.377901073923127, + "tokens_seen": 2685992960 + }, + { + "epoch": 0.81, + "learning_rate": 0.00015054764292878636, + "loss": 1.2276, + "theoretical_loss": 3.377894648501222, + "tokens_seen": 2686058496 + }, + { + "epoch": 0.81, + "learning_rate": 0.00015053159478435306, + "loss": 1.2552, + "theoretical_loss": 3.3778882232799807, + "tokens_seen": 2686124032 + }, + { + "epoch": 0.81, + "learning_rate": 0.00015051554663991976, + "loss": 1.2462, + "theoretical_loss": 3.3778817982593923, + "tokens_seen": 2686189568 + }, + { + "epoch": 0.81, + "learning_rate": 0.00015049949849548646, + "loss": 1.2854, + "theoretical_loss": 3.3778753734394455, + "tokens_seen": 2686255104 + }, + { + "epoch": 0.81, + "learning_rate": 0.00015048345035105316, + "loss": 1.285, + "theoretical_loss": 3.3778689488201294, + "tokens_seen": 2686320640 + }, + { + "epoch": 0.81, + "learning_rate": 0.00015046740220661986, + "loss": 1.2891, + "theoretical_loss": 3.377862524401432, + "tokens_seen": 2686386176 + }, + { + "epoch": 0.81, + "learning_rate": 0.00015045135406218657, + "loss": 1.3551, + "theoretical_loss": 3.377856100183343, + "tokens_seen": 2686451712 + }, + { + "epoch": 0.81, + "learning_rate": 0.00015043530591775327, + "loss": 1.1536, + "theoretical_loss": 3.377849676165851, + "tokens_seen": 2686517248 + }, + { + "epoch": 0.81, + "learning_rate": 0.00015041925777331997, + "loss": 1.2287, + "theoretical_loss": 3.377843252348945, + "tokens_seen": 2686582784 + }, + { + "epoch": 0.81, + "learning_rate": 0.00015040320962888667, + "loss": 1.2752, + "theoretical_loss": 3.3778368287326135, + "tokens_seen": 2686648320 + }, + { + "epoch": 0.81, + "learning_rate": 0.00015038716148445337, + "loss": 1.3162, + "theoretical_loss": 3.3778304053168453, + "tokens_seen": 2686713856 + }, + { + "epoch": 0.81, + "learning_rate": 0.00015037111334002007, + "loss": 1.2761, + "theoretical_loss": 3.3778239821016296, + "tokens_seen": 2686779392 + }, + { + "epoch": 0.81, + "learning_rate": 0.00015035506519558677, + "loss": 1.3127, + "theoretical_loss": 3.377817559086955, + "tokens_seen": 2686844928 + }, + { + "epoch": 0.81, + "learning_rate": 0.00015033901705115347, + "loss": 1.2737, + "theoretical_loss": 3.3778111362728103, + "tokens_seen": 2686910464 + }, + { + "epoch": 0.81, + "objective/train/docs_used": 976171, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.2426645755767822, + "objective/train/theoretical_loss": 3.377804713659185, + "objective/train/tokens_used": 2707436000, + "theoretical_loss": 3.377804713659185, + "tokens_seen": 2686976000 + }, + { + "epoch": 0.81, + "learning_rate": 0.00015032296890672017, + "loss": 1.2473, + "theoretical_loss": 3.377804713659185, + "tokens_seen": 2686976000 + }, + { + "epoch": 0.81, + "learning_rate": 0.00015030692076228687, + "loss": 1.3457, + "theoretical_loss": 3.377798291246067, + "tokens_seen": 2687041536 + }, + { + "epoch": 0.81, + "learning_rate": 0.00015029087261785357, + "loss": 1.2732, + "theoretical_loss": 3.3777918690334454, + "tokens_seen": 2687107072 + }, + { + "epoch": 0.81, + "learning_rate": 0.00015027482447342027, + "loss": 1.2041, + "theoretical_loss": 3.37778544702131, + "tokens_seen": 2687172608 + }, + { + "epoch": 0.81, + "learning_rate": 0.00015025877632898697, + "loss": 1.2018, + "theoretical_loss": 3.377779025209648, + "tokens_seen": 2687238144 + }, + { + "epoch": 0.81, + "learning_rate": 0.00015024272818455367, + "loss": 1.3648, + "theoretical_loss": 3.3777726035984497, + "tokens_seen": 2687303680 + }, + { + "epoch": 0.81, + "learning_rate": 0.00015022668004012037, + "loss": 1.2206, + "theoretical_loss": 3.3777661821877034, + "tokens_seen": 2687369216 + }, + { + "epoch": 0.81, + "learning_rate": 0.00015021063189568707, + "loss": 1.2279, + "theoretical_loss": 3.3777597609773977, + "tokens_seen": 2687434752 + }, + { + "epoch": 0.81, + "learning_rate": 0.00015019458375125378, + "loss": 1.2424, + "theoretical_loss": 3.377753339967522, + "tokens_seen": 2687500288 + }, + { + "epoch": 0.81, + "learning_rate": 0.00015017853560682048, + "loss": 1.2286, + "theoretical_loss": 3.3777469191580645, + "tokens_seen": 2687565824 + }, + { + "epoch": 0.81, + "learning_rate": 0.00015016248746238718, + "loss": 1.2611, + "theoretical_loss": 3.3777404985490147, + "tokens_seen": 2687631360 + }, + { + "epoch": 0.81, + "learning_rate": 0.00015014643931795388, + "loss": 1.2755, + "theoretical_loss": 3.377734078140361, + "tokens_seen": 2687696896 + }, + { + "epoch": 0.81, + "learning_rate": 0.00015013039117352058, + "loss": 1.3011, + "theoretical_loss": 3.3777276579320925, + "tokens_seen": 2687762432 + }, + { + "epoch": 0.81, + "learning_rate": 0.00015011434302908728, + "loss": 1.2809, + "theoretical_loss": 3.377721237924198, + "tokens_seen": 2687827968 + }, + { + "epoch": 0.81, + "learning_rate": 0.00015009829488465398, + "loss": 1.341, + "theoretical_loss": 3.3777148181166665, + "tokens_seen": 2687893504 + }, + { + "epoch": 0.81, + "learning_rate": 0.00015008224674022068, + "loss": 1.3256, + "theoretical_loss": 3.3777083985094865, + "tokens_seen": 2687959040 + }, + { + "epoch": 0.81, + "learning_rate": 0.00015006619859578735, + "loss": 1.312, + "theoretical_loss": 3.377701979102647, + "tokens_seen": 2688024576 + }, + { + "epoch": 0.81, + "learning_rate": 0.00015005015045135405, + "loss": 1.2729, + "theoretical_loss": 3.3776955598961376, + "tokens_seen": 2688090112 + }, + { + "epoch": 0.81, + "learning_rate": 0.00015003410230692075, + "loss": 1.1222, + "theoretical_loss": 3.3776891408899457, + "tokens_seen": 2688155648 + }, + { + "epoch": 0.81, + "learning_rate": 0.00015001805416248746, + "loss": 1.3472, + "theoretical_loss": 3.377682722084061, + "tokens_seen": 2688221184 + }, + { + "epoch": 0.81, + "learning_rate": 0.00015000200601805416, + "loss": 1.3291, + "theoretical_loss": 3.3776763034784727, + "tokens_seen": 2688286720 + }, + { + "epoch": 0.81, + "learning_rate": 0.00014998595787362086, + "loss": 1.2949, + "theoretical_loss": 3.3776698850731695, + "tokens_seen": 2688352256 + }, + { + "epoch": 0.81, + "learning_rate": 0.00014996990972918756, + "loss": 1.2898, + "theoretical_loss": 3.37766346686814, + "tokens_seen": 2688417792 + }, + { + "epoch": 0.81, + "learning_rate": 0.00014995386158475426, + "loss": 1.3179, + "theoretical_loss": 3.377657048863373, + "tokens_seen": 2688483328 + }, + { + "epoch": 0.81, + "learning_rate": 0.00014993781344032096, + "loss": 1.3038, + "theoretical_loss": 3.377650631058857, + "tokens_seen": 2688548864 + }, + { + "epoch": 0.81, + "learning_rate": 0.00014992176529588766, + "loss": 1.2686, + "theoretical_loss": 3.377644213454582, + "tokens_seen": 2688614400 + }, + { + "epoch": 0.81, + "learning_rate": 0.00014990571715145436, + "loss": 1.255, + "theoretical_loss": 3.377637796050536, + "tokens_seen": 2688679936 + }, + { + "epoch": 0.81, + "learning_rate": 0.00014988966900702106, + "loss": 1.291, + "theoretical_loss": 3.3776313788467087, + "tokens_seen": 2688745472 + }, + { + "epoch": 0.81, + "learning_rate": 0.00014987362086258776, + "loss": 1.2784, + "theoretical_loss": 3.377624961843088, + "tokens_seen": 2688811008 + }, + { + "epoch": 0.81, + "learning_rate": 0.00014985757271815446, + "loss": 1.2139, + "theoretical_loss": 3.377618545039663, + "tokens_seen": 2688876544 + }, + { + "epoch": 0.81, + "learning_rate": 0.0001498415245737212, + "loss": 1.3085, + "theoretical_loss": 3.377612128436423, + "tokens_seen": 2688942080 + }, + { + "epoch": 0.81, + "learning_rate": 0.0001498254764292879, + "loss": 1.2421, + "theoretical_loss": 3.3776057120333567, + "tokens_seen": 2689007616 + }, + { + "epoch": 0.81, + "learning_rate": 0.0001498094282848546, + "loss": 1.4425, + "theoretical_loss": 3.3775992958304526, + "tokens_seen": 2689073152 + }, + { + "epoch": 0.81, + "learning_rate": 0.0001497933801404213, + "loss": 1.1911, + "theoretical_loss": 3.3775928798277, + "tokens_seen": 2689138688 + }, + { + "epoch": 0.81, + "learning_rate": 0.000149777331995988, + "loss": 1.3562, + "theoretical_loss": 3.3775864640250877, + "tokens_seen": 2689204224 + }, + { + "epoch": 0.81, + "learning_rate": 0.00014976128385155467, + "loss": 1.1203, + "theoretical_loss": 3.377580048422604, + "tokens_seen": 2689269760 + }, + { + "epoch": 0.81, + "learning_rate": 0.00014974523570712137, + "loss": 1.2538, + "theoretical_loss": 3.3775736330202393, + "tokens_seen": 2689335296 + }, + { + "epoch": 0.81, + "learning_rate": 0.00014972918756268807, + "loss": 1.1957, + "theoretical_loss": 3.377567217817981, + "tokens_seen": 2689400832 + }, + { + "epoch": 0.81, + "learning_rate": 0.00014971313941825477, + "loss": 1.2955, + "theoretical_loss": 3.3775608028158186, + "tokens_seen": 2689466368 + }, + { + "epoch": 0.82, + "learning_rate": 0.00014969709127382147, + "loss": 1.3258, + "theoretical_loss": 3.3775543880137406, + "tokens_seen": 2689531904 + }, + { + "epoch": 0.82, + "learning_rate": 0.00014968104312938817, + "loss": 1.3705, + "theoretical_loss": 3.3775479734117364, + "tokens_seen": 2689597440 + }, + { + "epoch": 0.82, + "learning_rate": 0.00014966499498495487, + "loss": 1.171, + "theoretical_loss": 3.377541559009795, + "tokens_seen": 2689662976 + }, + { + "epoch": 0.82, + "learning_rate": 0.00014964894684052157, + "loss": 1.3139, + "theoretical_loss": 3.3775351448079043, + "tokens_seen": 2689728512 + }, + { + "epoch": 0.82, + "learning_rate": 0.00014963289869608827, + "loss": 1.2792, + "theoretical_loss": 3.3775287308060538, + "tokens_seen": 2689794048 + }, + { + "epoch": 0.82, + "learning_rate": 0.00014961685055165497, + "loss": 1.3405, + "theoretical_loss": 3.3775223170042326, + "tokens_seen": 2689859584 + }, + { + "epoch": 0.82, + "learning_rate": 0.00014960080240722167, + "loss": 1.2927, + "theoretical_loss": 3.3775159034024296, + "tokens_seen": 2689925120 + }, + { + "epoch": 0.82, + "learning_rate": 0.00014958475426278837, + "loss": 1.2523, + "theoretical_loss": 3.3775094900006333, + "tokens_seen": 2689990656 + }, + { + "epoch": 0.82, + "learning_rate": 0.00014956870611835507, + "loss": 1.2863, + "theoretical_loss": 3.377503076798833, + "tokens_seen": 2690056192 + }, + { + "epoch": 0.82, + "learning_rate": 0.00014955265797392177, + "loss": 1.2291, + "theoretical_loss": 3.377496663797017, + "tokens_seen": 2690121728 + }, + { + "epoch": 0.82, + "learning_rate": 0.00014953660982948847, + "loss": 1.2228, + "theoretical_loss": 3.3774902509951747, + "tokens_seen": 2690187264 + }, + { + "epoch": 0.82, + "objective/train/docs_used": 977388, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.346511960029602, + "objective/train/theoretical_loss": 3.377483838393295, + "objective/train/tokens_used": 2710712800, + "theoretical_loss": 3.377483838393295, + "tokens_seen": 2690252800 + }, + { + "epoch": 0.82, + "learning_rate": 0.00014952056168505518, + "loss": 1.2953, + "theoretical_loss": 3.377483838393295, + "tokens_seen": 2690252800 + }, + { + "epoch": 0.82, + "learning_rate": 0.00014950451354062188, + "loss": 1.2875, + "theoretical_loss": 3.3774774259913665, + "tokens_seen": 2690318336 + }, + { + "epoch": 0.82, + "learning_rate": 0.00014948846539618858, + "loss": 1.2392, + "theoretical_loss": 3.377471013789378, + "tokens_seen": 2690383872 + }, + { + "epoch": 0.82, + "learning_rate": 0.00014947241725175528, + "loss": 1.2177, + "theoretical_loss": 3.377464601787319, + "tokens_seen": 2690449408 + }, + { + "epoch": 0.82, + "learning_rate": 0.00014945636910732198, + "loss": 1.2413, + "theoretical_loss": 3.3774581899851777, + "tokens_seen": 2690514944 + }, + { + "epoch": 0.82, + "learning_rate": 0.00014944032096288868, + "loss": 1.3979, + "theoretical_loss": 3.3774517783829436, + "tokens_seen": 2690580480 + }, + { + "epoch": 0.82, + "learning_rate": 0.00014942427281845538, + "loss": 1.2233, + "theoretical_loss": 3.3774453669806053, + "tokens_seen": 2690646016 + }, + { + "epoch": 0.82, + "learning_rate": 0.00014940822467402208, + "loss": 1.2244, + "theoretical_loss": 3.3774389557781515, + "tokens_seen": 2690711552 + }, + { + "epoch": 0.82, + "learning_rate": 0.00014939217652958878, + "loss": 1.2532, + "theoretical_loss": 3.3774325447755715, + "tokens_seen": 2690777088 + }, + { + "epoch": 0.82, + "learning_rate": 0.00014937612838515545, + "loss": 1.3156, + "theoretical_loss": 3.377426133972854, + "tokens_seen": 2690842624 + }, + { + "epoch": 0.82, + "learning_rate": 0.00014936008024072215, + "loss": 1.2433, + "theoretical_loss": 3.3774197233699876, + "tokens_seen": 2690908160 + }, + { + "epoch": 0.82, + "learning_rate": 0.00014934403209628885, + "loss": 1.2793, + "theoretical_loss": 3.377413312966962, + "tokens_seen": 2690973696 + }, + { + "epoch": 0.82, + "learning_rate": 0.00014932798395185556, + "loss": 1.2674, + "theoretical_loss": 3.3774069027637657, + "tokens_seen": 2691039232 + }, + { + "epoch": 0.82, + "learning_rate": 0.00014931193580742226, + "loss": 1.198, + "theoretical_loss": 3.377400492760387, + "tokens_seen": 2691104768 + }, + { + "epoch": 0.82, + "learning_rate": 0.00014929588766298896, + "loss": 1.322, + "theoretical_loss": 3.3773940829568154, + "tokens_seen": 2691170304 + }, + { + "epoch": 0.82, + "learning_rate": 0.00014927983951855566, + "loss": 1.2045, + "theoretical_loss": 3.37738767335304, + "tokens_seen": 2691235840 + }, + { + "epoch": 0.82, + "learning_rate": 0.00014926379137412236, + "loss": 1.23, + "theoretical_loss": 3.3773812639490495, + "tokens_seen": 2691301376 + }, + { + "epoch": 0.82, + "learning_rate": 0.00014924774322968909, + "loss": 1.2638, + "theoretical_loss": 3.377374854744833, + "tokens_seen": 2691366912 + }, + { + "epoch": 0.82, + "learning_rate": 0.0001492316950852558, + "loss": 1.2405, + "theoretical_loss": 3.3773684457403785, + "tokens_seen": 2691432448 + }, + { + "epoch": 0.82, + "learning_rate": 0.0001492156469408225, + "loss": 1.2825, + "theoretical_loss": 3.377362036935676, + "tokens_seen": 2691497984 + }, + { + "epoch": 0.82, + "learning_rate": 0.0001491995987963892, + "loss": 1.2542, + "theoretical_loss": 3.3773556283307142, + "tokens_seen": 2691563520 + }, + { + "epoch": 0.82, + "learning_rate": 0.0001491835506519559, + "loss": 1.3, + "theoretical_loss": 3.377349219925481, + "tokens_seen": 2691629056 + }, + { + "epoch": 0.82, + "learning_rate": 0.0001491675025075226, + "loss": 1.1568, + "theoretical_loss": 3.377342811719967, + "tokens_seen": 2691694592 + }, + { + "epoch": 0.82, + "learning_rate": 0.0001491514543630893, + "loss": 1.2766, + "theoretical_loss": 3.3773364037141596, + "tokens_seen": 2691760128 + }, + { + "epoch": 0.82, + "learning_rate": 0.000149135406218656, + "loss": 1.2754, + "theoretical_loss": 3.377329995908049, + "tokens_seen": 2691825664 + }, + { + "epoch": 0.82, + "learning_rate": 0.0001491193580742227, + "loss": 1.2425, + "theoretical_loss": 3.377323588301623, + "tokens_seen": 2691891200 + }, + { + "epoch": 0.82, + "learning_rate": 0.0001491033099297894, + "loss": 1.3646, + "theoretical_loss": 3.377317180894871, + "tokens_seen": 2691956736 + }, + { + "epoch": 0.82, + "learning_rate": 0.0001490872617853561, + "loss": 1.3498, + "theoretical_loss": 3.377310773687782, + "tokens_seen": 2692022272 + }, + { + "epoch": 0.82, + "learning_rate": 0.0001490712136409228, + "loss": 1.2768, + "theoretical_loss": 3.3773043666803444, + "tokens_seen": 2692087808 + }, + { + "epoch": 0.82, + "learning_rate": 0.00014905516549648947, + "loss": 1.2841, + "theoretical_loss": 3.377297959872548, + "tokens_seen": 2692153344 + }, + { + "epoch": 0.82, + "learning_rate": 0.00014903911735205617, + "loss": 1.3308, + "theoretical_loss": 3.377291553264381, + "tokens_seen": 2692218880 + }, + { + "epoch": 0.82, + "learning_rate": 0.00014902306920762287, + "loss": 1.1289, + "theoretical_loss": 3.377285146855833, + "tokens_seen": 2692284416 + }, + { + "epoch": 0.82, + "learning_rate": 0.00014900702106318957, + "loss": 1.2543, + "theoretical_loss": 3.377278740646892, + "tokens_seen": 2692349952 + }, + { + "epoch": 0.82, + "learning_rate": 0.00014899097291875627, + "loss": 1.2128, + "theoretical_loss": 3.3772723346375475, + "tokens_seen": 2692415488 + }, + { + "epoch": 0.82, + "learning_rate": 0.00014897492477432297, + "loss": 1.2192, + "theoretical_loss": 3.3772659288277884, + "tokens_seen": 2692481024 + }, + { + "epoch": 0.82, + "learning_rate": 0.00014895887662988967, + "loss": 1.2332, + "theoretical_loss": 3.3772595232176035, + "tokens_seen": 2692546560 + }, + { + "epoch": 0.82, + "learning_rate": 0.00014894282848545637, + "loss": 1.2971, + "theoretical_loss": 3.377253117806982, + "tokens_seen": 2692612096 + }, + { + "epoch": 0.82, + "learning_rate": 0.00014892678034102307, + "loss": 1.2019, + "theoretical_loss": 3.3772467125959125, + "tokens_seen": 2692677632 + }, + { + "epoch": 0.82, + "learning_rate": 0.00014891073219658977, + "loss": 1.2699, + "theoretical_loss": 3.3772403075843838, + "tokens_seen": 2692743168 + }, + { + "epoch": 0.82, + "learning_rate": 0.00014889468405215647, + "loss": 1.2539, + "theoretical_loss": 3.377233902772385, + "tokens_seen": 2692808704 + }, + { + "epoch": 0.82, + "learning_rate": 0.00014887863590772317, + "loss": 1.2284, + "theoretical_loss": 3.3772274981599057, + "tokens_seen": 2692874240 + }, + { + "epoch": 0.82, + "learning_rate": 0.00014886258776328987, + "loss": 1.3178, + "theoretical_loss": 3.377221093746934, + "tokens_seen": 2692939776 + }, + { + "epoch": 0.82, + "learning_rate": 0.00014884653961885657, + "loss": 1.2258, + "theoretical_loss": 3.3772146895334587, + "tokens_seen": 2693005312 + }, + { + "epoch": 0.82, + "learning_rate": 0.00014883049147442328, + "loss": 1.2158, + "theoretical_loss": 3.377208285519469, + "tokens_seen": 2693070848 + }, + { + "epoch": 0.82, + "learning_rate": 0.00014881444332998998, + "loss": 1.2455, + "theoretical_loss": 3.3772018817049543, + "tokens_seen": 2693136384 + }, + { + "epoch": 0.82, + "learning_rate": 0.00014879839518555668, + "loss": 1.347, + "theoretical_loss": 3.377195478089903, + "tokens_seen": 2693201920 + }, + { + "epoch": 0.82, + "learning_rate": 0.00014878234704112338, + "loss": 1.2609, + "theoretical_loss": 3.3771890746743045, + "tokens_seen": 2693267456 + }, + { + "epoch": 0.82, + "learning_rate": 0.00014876629889669008, + "loss": 1.2136, + "theoretical_loss": 3.377182671458147, + "tokens_seen": 2693332992 + }, + { + "epoch": 0.82, + "learning_rate": 0.00014875025075225678, + "loss": 1.3592, + "theoretical_loss": 3.3771762684414197, + "tokens_seen": 2693398528 + }, + { + "epoch": 0.82, + "learning_rate": 0.00014873420260782348, + "loss": 1.2874, + "theoretical_loss": 3.3771698656241123, + "tokens_seen": 2693464064 + }, + { + "epoch": 0.82, + "objective/train/docs_used": 978638, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.297573447227478, + "objective/train/theoretical_loss": 3.377163463006213, + "objective/train/tokens_used": 2713989600, + "theoretical_loss": 3.377163463006213, + "tokens_seen": 2693529600 + }, + { + "epoch": 0.82, + "learning_rate": 0.00014871815446339018, + "loss": 1.3047, + "theoretical_loss": 3.377163463006213, + "tokens_seen": 2693529600 + }, + { + "epoch": 0.82, + "learning_rate": 0.00014870210631895688, + "loss": 1.2317, + "theoretical_loss": 3.3771570605877104, + "tokens_seen": 2693595136 + }, + { + "epoch": 0.82, + "learning_rate": 0.00014868605817452358, + "loss": 1.2376, + "theoretical_loss": 3.377150658368594, + "tokens_seen": 2693660672 + }, + { + "epoch": 0.82, + "learning_rate": 0.00014867001003009025, + "loss": 1.415, + "theoretical_loss": 3.3771442563488527, + "tokens_seen": 2693726208 + }, + { + "epoch": 0.82, + "learning_rate": 0.00014865396188565696, + "loss": 1.3057, + "theoretical_loss": 3.377137854528476, + "tokens_seen": 2693791744 + }, + { + "epoch": 0.82, + "learning_rate": 0.00014863791374122368, + "loss": 1.3325, + "theoretical_loss": 3.3771314529074514, + "tokens_seen": 2693857280 + }, + { + "epoch": 0.82, + "learning_rate": 0.00014862186559679038, + "loss": 1.1907, + "theoretical_loss": 3.377125051485769, + "tokens_seen": 2693922816 + }, + { + "epoch": 0.82, + "learning_rate": 0.00014860581745235708, + "loss": 1.3396, + "theoretical_loss": 3.377118650263417, + "tokens_seen": 2693988352 + }, + { + "epoch": 0.82, + "learning_rate": 0.00014858976930792378, + "loss": 1.2694, + "theoretical_loss": 3.377112249240385, + "tokens_seen": 2694053888 + }, + { + "epoch": 0.82, + "learning_rate": 0.00014857372116349049, + "loss": 1.248, + "theoretical_loss": 3.377105848416662, + "tokens_seen": 2694119424 + }, + { + "epoch": 0.82, + "learning_rate": 0.00014855767301905719, + "loss": 1.3064, + "theoretical_loss": 3.3770994477922365, + "tokens_seen": 2694184960 + }, + { + "epoch": 0.82, + "learning_rate": 0.0001485416248746239, + "loss": 1.1602, + "theoretical_loss": 3.3770930473670977, + "tokens_seen": 2694250496 + }, + { + "epoch": 0.82, + "learning_rate": 0.0001485255767301906, + "loss": 1.3179, + "theoretical_loss": 3.377086647141234, + "tokens_seen": 2694316032 + }, + { + "epoch": 0.82, + "learning_rate": 0.0001485095285857573, + "loss": 1.2771, + "theoretical_loss": 3.3770802471146353, + "tokens_seen": 2694381568 + }, + { + "epoch": 0.82, + "learning_rate": 0.000148493480441324, + "loss": 1.3177, + "theoretical_loss": 3.37707384728729, + "tokens_seen": 2694447104 + }, + { + "epoch": 0.82, + "learning_rate": 0.0001484774322968907, + "loss": 1.2721, + "theoretical_loss": 3.3770674476591864, + "tokens_seen": 2694512640 + }, + { + "epoch": 0.82, + "learning_rate": 0.0001484613841524574, + "loss": 1.1691, + "theoretical_loss": 3.3770610482303147, + "tokens_seen": 2694578176 + }, + { + "epoch": 0.82, + "learning_rate": 0.0001484453360080241, + "loss": 1.3544, + "theoretical_loss": 3.377054649000663, + "tokens_seen": 2694643712 + }, + { + "epoch": 0.82, + "learning_rate": 0.0001484292878635908, + "loss": 1.1779, + "theoretical_loss": 3.3770482499702212, + "tokens_seen": 2694709248 + }, + { + "epoch": 0.82, + "learning_rate": 0.0001484132397191575, + "loss": 1.3658, + "theoretical_loss": 3.3770418511389773, + "tokens_seen": 2694774784 + }, + { + "epoch": 0.82, + "learning_rate": 0.0001483971915747242, + "loss": 1.3326, + "theoretical_loss": 3.37703545250692, + "tokens_seen": 2694840320 + }, + { + "epoch": 0.82, + "learning_rate": 0.0001483811434302909, + "loss": 1.253, + "theoretical_loss": 3.3770290540740393, + "tokens_seen": 2694905856 + }, + { + "epoch": 0.82, + "learning_rate": 0.0001483650952858576, + "loss": 1.3461, + "theoretical_loss": 3.3770226558403236, + "tokens_seen": 2694971392 + }, + { + "epoch": 0.82, + "learning_rate": 0.00014834904714142427, + "loss": 1.1673, + "theoretical_loss": 3.377016257805762, + "tokens_seen": 2695036928 + }, + { + "epoch": 0.82, + "learning_rate": 0.00014833299899699097, + "loss": 1.2563, + "theoretical_loss": 3.377009859970343, + "tokens_seen": 2695102464 + }, + { + "epoch": 0.82, + "learning_rate": 0.00014831695085255767, + "loss": 1.2967, + "theoretical_loss": 3.3770034623340566, + "tokens_seen": 2695168000 + }, + { + "epoch": 0.82, + "learning_rate": 0.00014830090270812437, + "loss": 1.1918, + "theoretical_loss": 3.3769970648968908, + "tokens_seen": 2695233536 + }, + { + "epoch": 0.82, + "learning_rate": 0.00014828485456369107, + "loss": 1.354, + "theoretical_loss": 3.3769906676588346, + "tokens_seen": 2695299072 + }, + { + "epoch": 0.82, + "learning_rate": 0.00014826880641925777, + "loss": 1.4159, + "theoretical_loss": 3.3769842706198774, + "tokens_seen": 2695364608 + }, + { + "epoch": 0.82, + "learning_rate": 0.00014825275827482447, + "loss": 1.288, + "theoretical_loss": 3.376977873780008, + "tokens_seen": 2695430144 + }, + { + "epoch": 0.82, + "learning_rate": 0.00014823671013039117, + "loss": 1.3125, + "theoretical_loss": 3.3769714771392154, + "tokens_seen": 2695495680 + }, + { + "epoch": 0.82, + "learning_rate": 0.00014822066198595787, + "loss": 1.2939, + "theoretical_loss": 3.3769650806974885, + "tokens_seen": 2695561216 + }, + { + "epoch": 0.82, + "learning_rate": 0.00014820461384152457, + "loss": 1.2306, + "theoretical_loss": 3.376958684454816, + "tokens_seen": 2695626752 + }, + { + "epoch": 0.82, + "learning_rate": 0.00014818856569709127, + "loss": 1.2845, + "theoretical_loss": 3.3769522884111876, + "tokens_seen": 2695692288 + }, + { + "epoch": 0.82, + "learning_rate": 0.00014817251755265797, + "loss": 1.3431, + "theoretical_loss": 3.3769458925665914, + "tokens_seen": 2695757824 + }, + { + "epoch": 0.82, + "learning_rate": 0.00014815646940822467, + "loss": 1.2766, + "theoretical_loss": 3.3769394969210174, + "tokens_seen": 2695823360 + }, + { + "epoch": 0.82, + "learning_rate": 0.00014814042126379138, + "loss": 1.2741, + "theoretical_loss": 3.376933101474453, + "tokens_seen": 2695888896 + }, + { + "epoch": 0.82, + "learning_rate": 0.00014812437311935808, + "loss": 1.2007, + "theoretical_loss": 3.3769267062268886, + "tokens_seen": 2695954432 + }, + { + "epoch": 0.82, + "learning_rate": 0.00014810832497492478, + "loss": 1.2606, + "theoretical_loss": 3.376920311178313, + "tokens_seen": 2696019968 + }, + { + "epoch": 0.82, + "learning_rate": 0.00014809227683049148, + "loss": 1.3202, + "theoretical_loss": 3.3769139163287147, + "tokens_seen": 2696085504 + }, + { + "epoch": 0.82, + "learning_rate": 0.00014807622868605818, + "loss": 1.282, + "theoretical_loss": 3.3769075216780826, + "tokens_seen": 2696151040 + }, + { + "epoch": 0.82, + "learning_rate": 0.00014806018054162488, + "loss": 1.2565, + "theoretical_loss": 3.376901127226406, + "tokens_seen": 2696216576 + }, + { + "epoch": 0.82, + "learning_rate": 0.00014804413239719158, + "loss": 1.2328, + "theoretical_loss": 3.376894732973674, + "tokens_seen": 2696282112 + }, + { + "epoch": 0.82, + "learning_rate": 0.00014802808425275828, + "loss": 1.2682, + "theoretical_loss": 3.376888338919875, + "tokens_seen": 2696347648 + }, + { + "epoch": 0.82, + "learning_rate": 0.00014801203610832498, + "loss": 1.3096, + "theoretical_loss": 3.3768819450649987, + "tokens_seen": 2696413184 + }, + { + "epoch": 0.82, + "learning_rate": 0.00014799598796389168, + "loss": 1.2077, + "theoretical_loss": 3.376875551409033, + "tokens_seen": 2696478720 + }, + { + "epoch": 0.82, + "learning_rate": 0.00014797993981945838, + "loss": 1.2233, + "theoretical_loss": 3.3768691579519685, + "tokens_seen": 2696544256 + }, + { + "epoch": 0.82, + "learning_rate": 0.00014796389167502508, + "loss": 1.2824, + "theoretical_loss": 3.3768627646937928, + "tokens_seen": 2696609792 + }, + { + "epoch": 0.82, + "learning_rate": 0.00014794784353059178, + "loss": 1.317, + "theoretical_loss": 3.3768563716344957, + "tokens_seen": 2696675328 + }, + { + "epoch": 0.82, + "learning_rate": 0.00014793179538615848, + "loss": 1.2929, + "theoretical_loss": 3.3768499787740653, + "tokens_seen": 2696740864 + }, + { + "epoch": 0.82, + "objective/train/docs_used": 979818, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.4779577255249023, + "objective/train/theoretical_loss": 3.3768435861124915, + "objective/train/tokens_used": 2717266400, + "theoretical_loss": 3.3768435861124915, + "tokens_seen": 2696806400 + }, + { + "epoch": 0.82, + "learning_rate": 0.00014791574724172518, + "loss": 1.3747, + "theoretical_loss": 3.3768435861124915, + "tokens_seen": 2696806400 + }, + { + "epoch": 0.82, + "learning_rate": 0.00014789969909729188, + "loss": 1.3856, + "theoretical_loss": 3.3768371936497625, + "tokens_seen": 2696871936 + }, + { + "epoch": 0.82, + "learning_rate": 0.00014788365095285859, + "loss": 1.2643, + "theoretical_loss": 3.376830801385868, + "tokens_seen": 2696937472 + }, + { + "epoch": 0.82, + "learning_rate": 0.00014786760280842529, + "loss": 1.2009, + "theoretical_loss": 3.376824409320797, + "tokens_seen": 2697003008 + }, + { + "epoch": 0.82, + "learning_rate": 0.000147851554663992, + "loss": 1.3177, + "theoretical_loss": 3.3768180174545375, + "tokens_seen": 2697068544 + }, + { + "epoch": 0.82, + "learning_rate": 0.0001478355065195587, + "loss": 1.2486, + "theoretical_loss": 3.376811625787079, + "tokens_seen": 2697134080 + }, + { + "epoch": 0.82, + "learning_rate": 0.0001478194583751254, + "loss": 1.2509, + "theoretical_loss": 3.376805234318411, + "tokens_seen": 2697199616 + }, + { + "epoch": 0.82, + "learning_rate": 0.0001478034102306921, + "loss": 1.3666, + "theoretical_loss": 3.3767988430485225, + "tokens_seen": 2697265152 + }, + { + "epoch": 0.82, + "learning_rate": 0.0001477873620862588, + "loss": 1.3874, + "theoretical_loss": 3.3767924519774013, + "tokens_seen": 2697330688 + }, + { + "epoch": 0.82, + "learning_rate": 0.0001477713139418255, + "loss": 1.3367, + "theoretical_loss": 3.3767860611050375, + "tokens_seen": 2697396224 + }, + { + "epoch": 0.82, + "learning_rate": 0.0001477552657973922, + "loss": 1.2471, + "theoretical_loss": 3.3767796704314197, + "tokens_seen": 2697461760 + }, + { + "epoch": 0.82, + "learning_rate": 0.0001477392176529589, + "loss": 1.2588, + "theoretical_loss": 3.376773279956537, + "tokens_seen": 2697527296 + }, + { + "epoch": 0.82, + "learning_rate": 0.0001477231695085256, + "loss": 1.3066, + "theoretical_loss": 3.3767668896803786, + "tokens_seen": 2697592832 + }, + { + "epoch": 0.82, + "learning_rate": 0.0001477071213640923, + "loss": 1.3455, + "theoretical_loss": 3.3767604996029332, + "tokens_seen": 2697658368 + }, + { + "epoch": 0.82, + "learning_rate": 0.000147691073219659, + "loss": 1.3142, + "theoretical_loss": 3.3767541097241898, + "tokens_seen": 2697723904 + }, + { + "epoch": 0.82, + "learning_rate": 0.0001476750250752257, + "loss": 1.278, + "theoretical_loss": 3.376747720044137, + "tokens_seen": 2697789440 + }, + { + "epoch": 0.82, + "learning_rate": 0.00014765897693079237, + "loss": 1.2237, + "theoretical_loss": 3.3767413305627647, + "tokens_seen": 2697854976 + }, + { + "epoch": 0.82, + "learning_rate": 0.00014764292878635907, + "loss": 1.2496, + "theoretical_loss": 3.3767349412800614, + "tokens_seen": 2697920512 + }, + { + "epoch": 0.82, + "learning_rate": 0.00014762688064192577, + "loss": 1.2658, + "theoretical_loss": 3.376728552196016, + "tokens_seen": 2697986048 + }, + { + "epoch": 0.82, + "learning_rate": 0.00014761083249749247, + "loss": 1.2441, + "theoretical_loss": 3.3767221633106175, + "tokens_seen": 2698051584 + }, + { + "epoch": 0.82, + "learning_rate": 0.00014759478435305917, + "loss": 1.1654, + "theoretical_loss": 3.3767157746238547, + "tokens_seen": 2698117120 + }, + { + "epoch": 0.82, + "learning_rate": 0.00014757873620862587, + "loss": 1.2401, + "theoretical_loss": 3.3767093861357176, + "tokens_seen": 2698182656 + }, + { + "epoch": 0.82, + "learning_rate": 0.00014756268806419257, + "loss": 1.2272, + "theoretical_loss": 3.3767029978461944, + "tokens_seen": 2698248192 + }, + { + "epoch": 0.82, + "learning_rate": 0.00014754663991975927, + "loss": 1.2851, + "theoretical_loss": 3.3766966097552737, + "tokens_seen": 2698313728 + }, + { + "epoch": 0.82, + "learning_rate": 0.00014753059177532597, + "loss": 1.2394, + "theoretical_loss": 3.3766902218629458, + "tokens_seen": 2698379264 + }, + { + "epoch": 0.82, + "learning_rate": 0.00014751454363089267, + "loss": 1.2314, + "theoretical_loss": 3.3766838341691985, + "tokens_seen": 2698444800 + }, + { + "epoch": 0.82, + "learning_rate": 0.00014749849548645937, + "loss": 1.2342, + "theoretical_loss": 3.3766774466740213, + "tokens_seen": 2698510336 + }, + { + "epoch": 0.82, + "learning_rate": 0.00014748244734202607, + "loss": 1.2347, + "theoretical_loss": 3.376671059377403, + "tokens_seen": 2698575872 + }, + { + "epoch": 0.82, + "learning_rate": 0.00014746639919759278, + "loss": 1.1656, + "theoretical_loss": 3.376664672279333, + "tokens_seen": 2698641408 + }, + { + "epoch": 0.82, + "learning_rate": 0.00014745035105315948, + "loss": 1.3153, + "theoretical_loss": 3.3766582853797997, + "tokens_seen": 2698706944 + }, + { + "epoch": 0.82, + "learning_rate": 0.0001474343029087262, + "loss": 1.3235, + "theoretical_loss": 3.3766518986787926, + "tokens_seen": 2698772480 + }, + { + "epoch": 0.82, + "learning_rate": 0.0001474182547642929, + "loss": 1.3227, + "theoretical_loss": 3.3766455121763004, + "tokens_seen": 2698838016 + }, + { + "epoch": 0.82, + "learning_rate": 0.0001474022066198596, + "loss": 1.294, + "theoretical_loss": 3.3766391258723125, + "tokens_seen": 2698903552 + }, + { + "epoch": 0.82, + "learning_rate": 0.0001473861584754263, + "loss": 1.3, + "theoretical_loss": 3.376632739766818, + "tokens_seen": 2698969088 + }, + { + "epoch": 0.82, + "learning_rate": 0.000147370110330993, + "loss": 1.3003, + "theoretical_loss": 3.376626353859805, + "tokens_seen": 2699034624 + }, + { + "epoch": 0.82, + "learning_rate": 0.0001473540621865597, + "loss": 1.2822, + "theoretical_loss": 3.376619968151263, + "tokens_seen": 2699100160 + }, + { + "epoch": 0.82, + "learning_rate": 0.00014733801404212638, + "loss": 1.2739, + "theoretical_loss": 3.376613582641182, + "tokens_seen": 2699165696 + }, + { + "epoch": 0.82, + "learning_rate": 0.00014732196589769308, + "loss": 1.1701, + "theoretical_loss": 3.376607197329549, + "tokens_seen": 2699231232 + }, + { + "epoch": 0.82, + "learning_rate": 0.00014730591775325978, + "loss": 1.186, + "theoretical_loss": 3.376600812216355, + "tokens_seen": 2699296768 + }, + { + "epoch": 0.82, + "learning_rate": 0.00014728986960882648, + "loss": 1.1827, + "theoretical_loss": 3.3765944273015878, + "tokens_seen": 2699362304 + }, + { + "epoch": 0.82, + "learning_rate": 0.00014727382146439318, + "loss": 1.2555, + "theoretical_loss": 3.376588042585237, + "tokens_seen": 2699427840 + }, + { + "epoch": 0.82, + "learning_rate": 0.00014725777331995988, + "loss": 1.2694, + "theoretical_loss": 3.376581658067291, + "tokens_seen": 2699493376 + }, + { + "epoch": 0.82, + "learning_rate": 0.00014724172517552658, + "loss": 1.2093, + "theoretical_loss": 3.3765752737477395, + "tokens_seen": 2699558912 + }, + { + "epoch": 0.82, + "learning_rate": 0.00014722567703109328, + "loss": 1.1753, + "theoretical_loss": 3.3765688896265713, + "tokens_seen": 2699624448 + }, + { + "epoch": 0.82, + "learning_rate": 0.00014720962888665999, + "loss": 1.3301, + "theoretical_loss": 3.3765625057037756, + "tokens_seen": 2699689984 + }, + { + "epoch": 0.82, + "learning_rate": 0.00014719358074222669, + "loss": 1.3613, + "theoretical_loss": 3.3765561219793403, + "tokens_seen": 2699755520 + }, + { + "epoch": 0.82, + "learning_rate": 0.00014717753259779339, + "loss": 1.384, + "theoretical_loss": 3.376549738453256, + "tokens_seen": 2699821056 + }, + { + "epoch": 0.82, + "learning_rate": 0.0001471614844533601, + "loss": 1.2238, + "theoretical_loss": 3.376543355125511, + "tokens_seen": 2699886592 + }, + { + "epoch": 0.82, + "learning_rate": 0.0001471454363089268, + "loss": 1.2402, + "theoretical_loss": 3.376536971996094, + "tokens_seen": 2699952128 + }, + { + "epoch": 0.82, + "learning_rate": 0.0001471293881644935, + "loss": 1.3116, + "theoretical_loss": 3.376530589064995, + "tokens_seen": 2700017664 + }, + { + "epoch": 0.82, + "objective/train/docs_used": 981079, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.40153169631958, + "objective/train/theoretical_loss": 3.3765242063322014, + "objective/train/tokens_used": 2720543200, + "theoretical_loss": 3.3765242063322014, + "tokens_seen": 2700083200 + }, + { + "epoch": 0.82, + "learning_rate": 0.0001471133400200602, + "loss": 1.3543, + "theoretical_loss": 3.3765242063322014, + "tokens_seen": 2700083200 + }, + { + "epoch": 0.82, + "learning_rate": 0.0001470972918756269, + "loss": 1.2075, + "theoretical_loss": 3.376517823797704, + "tokens_seen": 2700148736 + }, + { + "epoch": 0.82, + "learning_rate": 0.0001470812437311936, + "loss": 1.26, + "theoretical_loss": 3.3765114414614907, + "tokens_seen": 2700214272 + }, + { + "epoch": 0.82, + "learning_rate": 0.0001470651955867603, + "loss": 1.1999, + "theoretical_loss": 3.376505059323551, + "tokens_seen": 2700279808 + }, + { + "epoch": 0.82, + "learning_rate": 0.000147049147442327, + "loss": 1.2824, + "theoretical_loss": 3.3764986773838737, + "tokens_seen": 2700345344 + }, + { + "epoch": 0.82, + "learning_rate": 0.0001470330992978937, + "loss": 1.3018, + "theoretical_loss": 3.3764922956424477, + "tokens_seen": 2700410880 + }, + { + "epoch": 0.82, + "learning_rate": 0.0001470170511534604, + "loss": 1.2583, + "theoretical_loss": 3.3764859140992627, + "tokens_seen": 2700476416 + }, + { + "epoch": 0.82, + "learning_rate": 0.0001470010030090271, + "loss": 1.3466, + "theoretical_loss": 3.376479532754307, + "tokens_seen": 2700541952 + }, + { + "epoch": 0.82, + "learning_rate": 0.0001469849548645938, + "loss": 1.2865, + "theoretical_loss": 3.37647315160757, + "tokens_seen": 2700607488 + }, + { + "epoch": 0.82, + "learning_rate": 0.0001469689067201605, + "loss": 1.3301, + "theoretical_loss": 3.3764667706590408, + "tokens_seen": 2700673024 + }, + { + "epoch": 0.82, + "learning_rate": 0.00014695285857572717, + "loss": 1.2878, + "theoretical_loss": 3.3764603899087082, + "tokens_seen": 2700738560 + }, + { + "epoch": 0.82, + "learning_rate": 0.00014693681043129387, + "loss": 1.257, + "theoretical_loss": 3.376454009356561, + "tokens_seen": 2700804096 + }, + { + "epoch": 0.82, + "learning_rate": 0.00014692076228686057, + "loss": 1.2587, + "theoretical_loss": 3.3764476290025893, + "tokens_seen": 2700869632 + }, + { + "epoch": 0.82, + "learning_rate": 0.00014690471414242727, + "loss": 1.2299, + "theoretical_loss": 3.3764412488467808, + "tokens_seen": 2700935168 + }, + { + "epoch": 0.82, + "learning_rate": 0.00014688866599799397, + "loss": 1.2588, + "theoretical_loss": 3.3764348688891253, + "tokens_seen": 2701000704 + }, + { + "epoch": 0.82, + "learning_rate": 0.00014687261785356067, + "loss": 1.3055, + "theoretical_loss": 3.3764284891296117, + "tokens_seen": 2701066240 + }, + { + "epoch": 0.82, + "learning_rate": 0.00014685656970912737, + "loss": 1.2627, + "theoretical_loss": 3.376422109568229, + "tokens_seen": 2701131776 + }, + { + "epoch": 0.82, + "learning_rate": 0.0001468405215646941, + "loss": 1.3466, + "theoretical_loss": 3.376415730204966, + "tokens_seen": 2701197312 + }, + { + "epoch": 0.82, + "learning_rate": 0.0001468244734202608, + "loss": 1.3037, + "theoretical_loss": 3.376409351039812, + "tokens_seen": 2701262848 + }, + { + "epoch": 0.82, + "learning_rate": 0.0001468084252758275, + "loss": 1.2878, + "theoretical_loss": 3.3764029720727566, + "tokens_seen": 2701328384 + }, + { + "epoch": 0.82, + "learning_rate": 0.0001467923771313942, + "loss": 1.2375, + "theoretical_loss": 3.376396593303788, + "tokens_seen": 2701393920 + }, + { + "epoch": 0.82, + "learning_rate": 0.0001467763289869609, + "loss": 1.2369, + "theoretical_loss": 3.3763902147328952, + "tokens_seen": 2701459456 + }, + { + "epoch": 0.82, + "learning_rate": 0.0001467602808425276, + "loss": 1.3689, + "theoretical_loss": 3.376383836360068, + "tokens_seen": 2701524992 + }, + { + "epoch": 0.82, + "learning_rate": 0.0001467442326980943, + "loss": 1.3292, + "theoretical_loss": 3.3763774581852948, + "tokens_seen": 2701590528 + }, + { + "epoch": 0.82, + "learning_rate": 0.000146728184553661, + "loss": 1.2386, + "theoretical_loss": 3.3763710802085645, + "tokens_seen": 2701656064 + }, + { + "epoch": 0.82, + "learning_rate": 0.0001467121364092277, + "loss": 1.243, + "theoretical_loss": 3.376364702429867, + "tokens_seen": 2701721600 + }, + { + "epoch": 0.82, + "learning_rate": 0.0001466960882647944, + "loss": 1.2878, + "theoretical_loss": 3.376358324849191, + "tokens_seen": 2701787136 + }, + { + "epoch": 0.82, + "learning_rate": 0.0001466800401203611, + "loss": 1.2519, + "theoretical_loss": 3.3763519474665245, + "tokens_seen": 2701852672 + }, + { + "epoch": 0.82, + "learning_rate": 0.0001466639919759278, + "loss": 1.3079, + "theoretical_loss": 3.3763455702818583, + "tokens_seen": 2701918208 + }, + { + "epoch": 0.82, + "learning_rate": 0.0001466479438314945, + "loss": 1.2239, + "theoretical_loss": 3.37633919329518, + "tokens_seen": 2701983744 + }, + { + "epoch": 0.82, + "learning_rate": 0.00014663189568706118, + "loss": 1.3739, + "theoretical_loss": 3.37633281650648, + "tokens_seen": 2702049280 + }, + { + "epoch": 0.82, + "learning_rate": 0.00014661584754262788, + "loss": 1.1938, + "theoretical_loss": 3.3763264399157458, + "tokens_seen": 2702114816 + }, + { + "epoch": 0.82, + "learning_rate": 0.00014659979939819458, + "loss": 1.2008, + "theoretical_loss": 3.3763200635229675, + "tokens_seen": 2702180352 + }, + { + "epoch": 0.82, + "learning_rate": 0.00014658375125376128, + "loss": 1.2998, + "theoretical_loss": 3.376313687328134, + "tokens_seen": 2702245888 + }, + { + "epoch": 0.82, + "learning_rate": 0.00014656770310932798, + "loss": 1.2609, + "theoretical_loss": 3.3763073113312343, + "tokens_seen": 2702311424 + }, + { + "epoch": 0.82, + "learning_rate": 0.00014655165496489468, + "loss": 1.2373, + "theoretical_loss": 3.3763009355322575, + "tokens_seen": 2702376960 + }, + { + "epoch": 0.82, + "learning_rate": 0.00014653560682046138, + "loss": 1.1796, + "theoretical_loss": 3.3762945599311927, + "tokens_seen": 2702442496 + }, + { + "epoch": 0.82, + "learning_rate": 0.00014651955867602809, + "loss": 1.3552, + "theoretical_loss": 3.376288184528028, + "tokens_seen": 2702508032 + }, + { + "epoch": 0.82, + "learning_rate": 0.00014650351053159479, + "loss": 1.2805, + "theoretical_loss": 3.3762818093227542, + "tokens_seen": 2702573568 + }, + { + "epoch": 0.82, + "learning_rate": 0.0001464874623871615, + "loss": 1.2485, + "theoretical_loss": 3.3762754343153594, + "tokens_seen": 2702639104 + }, + { + "epoch": 0.82, + "learning_rate": 0.0001464714142427282, + "loss": 1.2088, + "theoretical_loss": 3.376269059505832, + "tokens_seen": 2702704640 + }, + { + "epoch": 0.82, + "learning_rate": 0.0001464553660982949, + "loss": 1.307, + "theoretical_loss": 3.3762626848941624, + "tokens_seen": 2702770176 + }, + { + "epoch": 0.82, + "learning_rate": 0.0001464393179538616, + "loss": 1.2819, + "theoretical_loss": 3.376256310480339, + "tokens_seen": 2702835712 + }, + { + "epoch": 0.82, + "learning_rate": 0.0001464232698094283, + "loss": 1.2105, + "theoretical_loss": 3.3762499362643505, + "tokens_seen": 2702901248 + }, + { + "epoch": 0.82, + "learning_rate": 0.000146407221664995, + "loss": 1.3775, + "theoretical_loss": 3.3762435622461866, + "tokens_seen": 2702966784 + }, + { + "epoch": 0.82, + "learning_rate": 0.0001463911735205617, + "loss": 1.354, + "theoretical_loss": 3.376237188425836, + "tokens_seen": 2703032320 + }, + { + "epoch": 0.82, + "learning_rate": 0.0001463751253761284, + "loss": 1.268, + "theoretical_loss": 3.376230814803288, + "tokens_seen": 2703097856 + }, + { + "epoch": 0.82, + "learning_rate": 0.0001463590772316951, + "loss": 1.3144, + "theoretical_loss": 3.3762244413785316, + "tokens_seen": 2703163392 + }, + { + "epoch": 0.82, + "learning_rate": 0.0001463430290872618, + "loss": 1.2997, + "theoretical_loss": 3.3762180681515557, + "tokens_seen": 2703228928 + }, + { + "epoch": 0.82, + "learning_rate": 0.0001463269809428285, + "loss": 1.2682, + "theoretical_loss": 3.3762116951223495, + "tokens_seen": 2703294464 + }, + { + "epoch": 0.82, + "objective/train/docs_used": 982308, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.2470022439956665, + "objective/train/theoretical_loss": 3.376205322290902, + "objective/train/tokens_used": 2723820000, + "theoretical_loss": 3.376205322290902, + "tokens_seen": 2703360000 + }, + { + "epoch": 0.82, + "learning_rate": 0.0001463109327983952, + "loss": 1.2963, + "theoretical_loss": 3.376205322290902, + "tokens_seen": 2703360000 + }, + { + "epoch": 0.82, + "learning_rate": 0.0001462948846539619, + "loss": 1.1549, + "theoretical_loss": 3.3761989496572022, + "tokens_seen": 2703425536 + }, + { + "epoch": 0.82, + "learning_rate": 0.0001462788365095286, + "loss": 1.2914, + "theoretical_loss": 3.3761925772212398, + "tokens_seen": 2703491072 + }, + { + "epoch": 0.82, + "learning_rate": 0.0001462627883650953, + "loss": 1.2249, + "theoretical_loss": 3.376186204983003, + "tokens_seen": 2703556608 + }, + { + "epoch": 0.82, + "learning_rate": 0.00014624674022066197, + "loss": 1.2862, + "theoretical_loss": 3.3761798329424813, + "tokens_seen": 2703622144 + }, + { + "epoch": 0.82, + "learning_rate": 0.0001462306920762287, + "loss": 1.2504, + "theoretical_loss": 3.376173461099664, + "tokens_seen": 2703687680 + }, + { + "epoch": 0.82, + "learning_rate": 0.0001462146439317954, + "loss": 1.261, + "theoretical_loss": 3.3761670894545395, + "tokens_seen": 2703753216 + }, + { + "epoch": 0.82, + "learning_rate": 0.0001461985957873621, + "loss": 1.2598, + "theoretical_loss": 3.376160718007097, + "tokens_seen": 2703818752 + }, + { + "epoch": 0.82, + "learning_rate": 0.0001461825476429288, + "loss": 1.3083, + "theoretical_loss": 3.3761543467573265, + "tokens_seen": 2703884288 + }, + { + "epoch": 0.82, + "learning_rate": 0.0001461664994984955, + "loss": 1.3634, + "theoretical_loss": 3.3761479757052157, + "tokens_seen": 2703949824 + }, + { + "epoch": 0.82, + "learning_rate": 0.0001461504513540622, + "loss": 1.2682, + "theoretical_loss": 3.376141604850755, + "tokens_seen": 2704015360 + }, + { + "epoch": 0.82, + "learning_rate": 0.0001461344032096289, + "loss": 1.3853, + "theoretical_loss": 3.3761352341939324, + "tokens_seen": 2704080896 + }, + { + "epoch": 0.82, + "learning_rate": 0.0001461183550651956, + "loss": 1.2687, + "theoretical_loss": 3.3761288637347375, + "tokens_seen": 2704146432 + }, + { + "epoch": 0.82, + "learning_rate": 0.0001461023069207623, + "loss": 1.2904, + "theoretical_loss": 3.3761224934731597, + "tokens_seen": 2704211968 + }, + { + "epoch": 0.82, + "learning_rate": 0.000146086258776329, + "loss": 1.3419, + "theoretical_loss": 3.3761161234091874, + "tokens_seen": 2704277504 + }, + { + "epoch": 0.82, + "learning_rate": 0.0001460702106318957, + "loss": 1.2249, + "theoretical_loss": 3.3761097535428104, + "tokens_seen": 2704343040 + }, + { + "epoch": 0.82, + "learning_rate": 0.0001460541624874624, + "loss": 1.2483, + "theoretical_loss": 3.376103383874017, + "tokens_seen": 2704408576 + }, + { + "epoch": 0.82, + "learning_rate": 0.0001460381143430291, + "loss": 1.2928, + "theoretical_loss": 3.376097014402797, + "tokens_seen": 2704474112 + }, + { + "epoch": 0.82, + "learning_rate": 0.0001460220661985958, + "loss": 1.2259, + "theoretical_loss": 3.3760906451291386, + "tokens_seen": 2704539648 + }, + { + "epoch": 0.82, + "learning_rate": 0.0001460060180541625, + "loss": 1.1781, + "theoretical_loss": 3.3760842760530316, + "tokens_seen": 2704605184 + }, + { + "epoch": 0.82, + "learning_rate": 0.0001459899699097292, + "loss": 1.1441, + "theoretical_loss": 3.376077907174465, + "tokens_seen": 2704670720 + }, + { + "epoch": 0.82, + "learning_rate": 0.0001459739217652959, + "loss": 1.2881, + "theoretical_loss": 3.376071538493428, + "tokens_seen": 2704736256 + }, + { + "epoch": 0.82, + "learning_rate": 0.0001459578736208626, + "loss": 1.1784, + "theoretical_loss": 3.376065170009909, + "tokens_seen": 2704801792 + }, + { + "epoch": 0.82, + "learning_rate": 0.0001459418254764293, + "loss": 1.299, + "theoretical_loss": 3.376058801723898, + "tokens_seen": 2704867328 + }, + { + "epoch": 0.82, + "learning_rate": 0.00014592577733199598, + "loss": 1.3057, + "theoretical_loss": 3.376052433635383, + "tokens_seen": 2704932864 + }, + { + "epoch": 0.82, + "learning_rate": 0.00014590972918756268, + "loss": 1.2769, + "theoretical_loss": 3.3760460657443545, + "tokens_seen": 2704998400 + }, + { + "epoch": 0.82, + "learning_rate": 0.00014589368104312938, + "loss": 1.2534, + "theoretical_loss": 3.3760396980508007, + "tokens_seen": 2705063936 + }, + { + "epoch": 0.82, + "learning_rate": 0.00014587763289869608, + "loss": 1.2283, + "theoretical_loss": 3.3760333305547108, + "tokens_seen": 2705129472 + }, + { + "epoch": 0.82, + "learning_rate": 0.00014586158475426278, + "loss": 1.2881, + "theoretical_loss": 3.376026963256074, + "tokens_seen": 2705195008 + }, + { + "epoch": 0.82, + "learning_rate": 0.00014584553660982948, + "loss": 1.3326, + "theoretical_loss": 3.3760205961548793, + "tokens_seen": 2705260544 + }, + { + "epoch": 0.82, + "learning_rate": 0.00014582948846539619, + "loss": 1.2526, + "theoretical_loss": 3.3760142292511155, + "tokens_seen": 2705326080 + }, + { + "epoch": 0.82, + "learning_rate": 0.00014581344032096289, + "loss": 1.3681, + "theoretical_loss": 3.376007862544772, + "tokens_seen": 2705391616 + }, + { + "epoch": 0.82, + "learning_rate": 0.0001457973921765296, + "loss": 1.2566, + "theoretical_loss": 3.376001496035838, + "tokens_seen": 2705457152 + }, + { + "epoch": 0.82, + "learning_rate": 0.0001457813440320963, + "loss": 1.3097, + "theoretical_loss": 3.375995129724303, + "tokens_seen": 2705522688 + }, + { + "epoch": 0.82, + "learning_rate": 0.000145765295887663, + "loss": 1.2514, + "theoretical_loss": 3.375988763610155, + "tokens_seen": 2705588224 + }, + { + "epoch": 0.82, + "learning_rate": 0.0001457492477432297, + "loss": 1.3017, + "theoretical_loss": 3.375982397693384, + "tokens_seen": 2705653760 + }, + { + "epoch": 0.82, + "learning_rate": 0.0001457331995987964, + "loss": 1.273, + "theoretical_loss": 3.3759760319739787, + "tokens_seen": 2705719296 + }, + { + "epoch": 0.82, + "learning_rate": 0.0001457171514543631, + "loss": 1.2562, + "theoretical_loss": 3.3759696664519288, + "tokens_seen": 2705784832 + }, + { + "epoch": 0.82, + "learning_rate": 0.0001457011033099298, + "loss": 1.4455, + "theoretical_loss": 3.3759633011272223, + "tokens_seen": 2705850368 + }, + { + "epoch": 0.82, + "learning_rate": 0.0001456850551654965, + "loss": 1.2452, + "theoretical_loss": 3.3759569359998487, + "tokens_seen": 2705915904 + }, + { + "epoch": 0.82, + "learning_rate": 0.0001456690070210632, + "loss": 1.232, + "theoretical_loss": 3.3759505710697977, + "tokens_seen": 2705981440 + }, + { + "epoch": 0.82, + "learning_rate": 0.0001456529588766299, + "loss": 1.3147, + "theoretical_loss": 3.375944206337058, + "tokens_seen": 2706046976 + }, + { + "epoch": 0.82, + "learning_rate": 0.00014563691073219662, + "loss": 1.2602, + "theoretical_loss": 3.3759378418016186, + "tokens_seen": 2706112512 + }, + { + "epoch": 0.82, + "learning_rate": 0.0001456208625877633, + "loss": 1.1888, + "theoretical_loss": 3.375931477463469, + "tokens_seen": 2706178048 + }, + { + "epoch": 0.82, + "learning_rate": 0.00014560481444333, + "loss": 1.2785, + "theoretical_loss": 3.375925113322598, + "tokens_seen": 2706243584 + }, + { + "epoch": 0.82, + "learning_rate": 0.0001455887662988967, + "loss": 1.2693, + "theoretical_loss": 3.3759187493789944, + "tokens_seen": 2706309120 + }, + { + "epoch": 0.82, + "learning_rate": 0.0001455727181544634, + "loss": 1.2234, + "theoretical_loss": 3.375912385632648, + "tokens_seen": 2706374656 + }, + { + "epoch": 0.82, + "learning_rate": 0.0001455566700100301, + "loss": 1.3043, + "theoretical_loss": 3.375906022083547, + "tokens_seen": 2706440192 + }, + { + "epoch": 0.82, + "learning_rate": 0.0001455406218655968, + "loss": 1.3047, + "theoretical_loss": 3.3758996587316816, + "tokens_seen": 2706505728 + }, + { + "epoch": 0.82, + "learning_rate": 0.0001455245737211635, + "loss": 1.3065, + "theoretical_loss": 3.37589329557704, + "tokens_seen": 2706571264 + }, + { + "epoch": 0.82, + "objective/train/docs_used": 983478, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.2508280277252197, + "objective/train/theoretical_loss": 3.375886932619612, + "objective/train/tokens_used": 2727096800, + "theoretical_loss": 3.375886932619612, + "tokens_seen": 2706636800 + }, + { + "epoch": 0.82, + "learning_rate": 0.0001455085255767302, + "loss": 1.2496, + "theoretical_loss": 3.375886932619612, + "tokens_seen": 2706636800 + }, + { + "epoch": 0.82, + "learning_rate": 0.0001454924774322969, + "loss": 1.36, + "theoretical_loss": 3.375880569859387, + "tokens_seen": 2706702336 + }, + { + "epoch": 0.82, + "learning_rate": 0.0001454764292878636, + "loss": 1.2592, + "theoretical_loss": 3.3758742072963526, + "tokens_seen": 2706767872 + }, + { + "epoch": 0.82, + "learning_rate": 0.0001454603811434303, + "loss": 1.3416, + "theoretical_loss": 3.3758678449304993, + "tokens_seen": 2706833408 + }, + { + "epoch": 0.82, + "learning_rate": 0.000145444332998997, + "loss": 1.3082, + "theoretical_loss": 3.375861482761815, + "tokens_seen": 2706898944 + }, + { + "epoch": 0.82, + "learning_rate": 0.0001454282848545637, + "loss": 1.3462, + "theoretical_loss": 3.3758551207902903, + "tokens_seen": 2706964480 + }, + { + "epoch": 0.82, + "learning_rate": 0.0001454122367101304, + "loss": 1.2726, + "theoretical_loss": 3.3758487590159136, + "tokens_seen": 2707030016 + }, + { + "epoch": 0.82, + "learning_rate": 0.0001453961885656971, + "loss": 1.3316, + "theoretical_loss": 3.3758423974386735, + "tokens_seen": 2707095552 + }, + { + "epoch": 0.82, + "learning_rate": 0.0001453801404212638, + "loss": 1.2324, + "theoretical_loss": 3.37583603605856, + "tokens_seen": 2707161088 + }, + { + "epoch": 0.82, + "learning_rate": 0.0001453640922768305, + "loss": 1.2289, + "theoretical_loss": 3.3758296748755616, + "tokens_seen": 2707226624 + }, + { + "epoch": 0.82, + "learning_rate": 0.0001453480441323972, + "loss": 1.3156, + "theoretical_loss": 3.375823313889668, + "tokens_seen": 2707292160 + }, + { + "epoch": 0.82, + "learning_rate": 0.0001453319959879639, + "loss": 1.288, + "theoretical_loss": 3.375816953100868, + "tokens_seen": 2707357696 + }, + { + "epoch": 0.82, + "learning_rate": 0.0001453159478435306, + "loss": 1.3363, + "theoretical_loss": 3.37581059250915, + "tokens_seen": 2707423232 + }, + { + "epoch": 0.82, + "learning_rate": 0.0001452998996990973, + "loss": 1.1892, + "theoretical_loss": 3.3758042321145045, + "tokens_seen": 2707488768 + }, + { + "epoch": 0.82, + "learning_rate": 0.000145283851554664, + "loss": 1.2801, + "theoretical_loss": 3.3757978719169195, + "tokens_seen": 2707554304 + }, + { + "epoch": 0.82, + "learning_rate": 0.0001452678034102307, + "loss": 1.3305, + "theoretical_loss": 3.3757915119163853, + "tokens_seen": 2707619840 + }, + { + "epoch": 0.82, + "learning_rate": 0.0001452517552657974, + "loss": 1.2569, + "theoretical_loss": 3.37578515211289, + "tokens_seen": 2707685376 + }, + { + "epoch": 0.82, + "learning_rate": 0.00014523570712136408, + "loss": 1.2712, + "theoretical_loss": 3.3757787925064227, + "tokens_seen": 2707750912 + }, + { + "epoch": 0.82, + "learning_rate": 0.00014521965897693078, + "loss": 1.4439, + "theoretical_loss": 3.375772433096973, + "tokens_seen": 2707816448 + }, + { + "epoch": 0.82, + "learning_rate": 0.00014520361083249748, + "loss": 1.3621, + "theoretical_loss": 3.3757660738845296, + "tokens_seen": 2707881984 + }, + { + "epoch": 0.82, + "learning_rate": 0.00014518756268806418, + "loss": 1.3924, + "theoretical_loss": 3.375759714869082, + "tokens_seen": 2707947520 + }, + { + "epoch": 0.82, + "learning_rate": 0.00014517151454363088, + "loss": 1.4671, + "theoretical_loss": 3.3757533560506197, + "tokens_seen": 2708013056 + }, + { + "epoch": 0.82, + "learning_rate": 0.00014515546639919759, + "loss": 1.326, + "theoretical_loss": 3.375746997429131, + "tokens_seen": 2708078592 + }, + { + "epoch": 0.82, + "learning_rate": 0.00014513941825476429, + "loss": 1.3693, + "theoretical_loss": 3.375740639004606, + "tokens_seen": 2708144128 + }, + { + "epoch": 0.82, + "learning_rate": 0.00014512337011033099, + "loss": 1.4516, + "theoretical_loss": 3.3757342807770323, + "tokens_seen": 2708209664 + }, + { + "epoch": 0.82, + "learning_rate": 0.0001451073219658977, + "loss": 1.3968, + "theoretical_loss": 3.3757279227464005, + "tokens_seen": 2708275200 + }, + { + "epoch": 0.82, + "learning_rate": 0.0001450912738214644, + "loss": 1.3103, + "theoretical_loss": 3.3757215649126993, + "tokens_seen": 2708340736 + }, + { + "epoch": 0.82, + "learning_rate": 0.0001450752256770311, + "loss": 1.4291, + "theoretical_loss": 3.375715207275918, + "tokens_seen": 2708406272 + }, + { + "epoch": 0.82, + "learning_rate": 0.0001450591775325978, + "loss": 1.269, + "theoretical_loss": 3.3757088498360446, + "tokens_seen": 2708471808 + }, + { + "epoch": 0.82, + "learning_rate": 0.00014504312938816452, + "loss": 1.3913, + "theoretical_loss": 3.37570249259307, + "tokens_seen": 2708537344 + }, + { + "epoch": 0.82, + "learning_rate": 0.00014502708124373122, + "loss": 1.3039, + "theoretical_loss": 3.3756961355469817, + "tokens_seen": 2708602880 + }, + { + "epoch": 0.82, + "learning_rate": 0.00014501103309929792, + "loss": 1.2911, + "theoretical_loss": 3.3756897786977698, + "tokens_seen": 2708668416 + }, + { + "epoch": 0.82, + "learning_rate": 0.00014499498495486462, + "loss": 1.4271, + "theoretical_loss": 3.3756834220454235, + "tokens_seen": 2708733952 + }, + { + "epoch": 0.82, + "learning_rate": 0.00014497893681043132, + "loss": 1.5736, + "theoretical_loss": 3.3756770655899313, + "tokens_seen": 2708799488 + }, + { + "epoch": 0.82, + "learning_rate": 0.00014496288866599802, + "loss": 1.3655, + "theoretical_loss": 3.375670709331283, + "tokens_seen": 2708865024 + }, + { + "epoch": 0.82, + "learning_rate": 0.00014494684052156472, + "loss": 1.2796, + "theoretical_loss": 3.3756643532694675, + "tokens_seen": 2708930560 + }, + { + "epoch": 0.82, + "learning_rate": 0.00014493079237713142, + "loss": 1.2905, + "theoretical_loss": 3.375657997404474, + "tokens_seen": 2708996096 + }, + { + "epoch": 0.82, + "learning_rate": 0.0001449147442326981, + "loss": 1.2248, + "theoretical_loss": 3.375651641736291, + "tokens_seen": 2709061632 + }, + { + "epoch": 0.82, + "learning_rate": 0.0001448986960882648, + "loss": 1.3859, + "theoretical_loss": 3.3756452862649087, + "tokens_seen": 2709127168 + }, + { + "epoch": 0.82, + "learning_rate": 0.0001448826479438315, + "loss": 1.2437, + "theoretical_loss": 3.3756389309903154, + "tokens_seen": 2709192704 + }, + { + "epoch": 0.82, + "learning_rate": 0.0001448665997993982, + "loss": 1.325, + "theoretical_loss": 3.375632575912501, + "tokens_seen": 2709258240 + }, + { + "epoch": 0.82, + "learning_rate": 0.0001448505516549649, + "loss": 1.4055, + "theoretical_loss": 3.375626221031454, + "tokens_seen": 2709323776 + }, + { + "epoch": 0.82, + "learning_rate": 0.0001448345035105316, + "loss": 1.3356, + "theoretical_loss": 3.3756198663471637, + "tokens_seen": 2709389312 + }, + { + "epoch": 0.82, + "learning_rate": 0.0001448184553660983, + "loss": 1.3314, + "theoretical_loss": 3.375613511859619, + "tokens_seen": 2709454848 + }, + { + "epoch": 0.82, + "learning_rate": 0.000144802407221665, + "loss": 1.2572, + "theoretical_loss": 3.37560715756881, + "tokens_seen": 2709520384 + }, + { + "epoch": 0.82, + "learning_rate": 0.0001447863590772317, + "loss": 1.2522, + "theoretical_loss": 3.375600803474725, + "tokens_seen": 2709585920 + }, + { + "epoch": 0.82, + "learning_rate": 0.0001447703109327984, + "loss": 1.2535, + "theoretical_loss": 3.3755944495773536, + "tokens_seen": 2709651456 + }, + { + "epoch": 0.82, + "learning_rate": 0.0001447542627883651, + "loss": 1.3727, + "theoretical_loss": 3.3755880958766844, + "tokens_seen": 2709716992 + }, + { + "epoch": 0.82, + "learning_rate": 0.0001447382146439318, + "loss": 1.3815, + "theoretical_loss": 3.375581742372707, + "tokens_seen": 2709782528 + }, + { + "epoch": 0.82, + "learning_rate": 0.0001447221664994985, + "loss": 1.1801, + "theoretical_loss": 3.37557538906541, + "tokens_seen": 2709848064 + }, + { + "epoch": 0.82, + "objective/train/docs_used": 984630, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.3604110479354858, + "objective/train/theoretical_loss": 3.3755690359547836, + "objective/train/tokens_used": 2730373600, + "theoretical_loss": 3.3755690359547836, + "tokens_seen": 2709913600 + }, + { + "epoch": 0.82, + "learning_rate": 0.0001447061183550652, + "loss": 1.2232, + "theoretical_loss": 3.3755690359547836, + "tokens_seen": 2709913600 + }, + { + "epoch": 0.82, + "learning_rate": 0.0001446900702106319, + "loss": 1.3413, + "theoretical_loss": 3.375562683040816, + "tokens_seen": 2709979136 + }, + { + "epoch": 0.82, + "learning_rate": 0.0001446740220661986, + "loss": 1.3037, + "theoretical_loss": 3.375556330323497, + "tokens_seen": 2710044672 + }, + { + "epoch": 0.82, + "learning_rate": 0.0001446579739217653, + "loss": 1.4428, + "theoretical_loss": 3.375549977802815, + "tokens_seen": 2710110208 + }, + { + "epoch": 0.82, + "learning_rate": 0.000144641925777332, + "loss": 1.3476, + "theoretical_loss": 3.37554362547876, + "tokens_seen": 2710175744 + }, + { + "epoch": 0.82, + "learning_rate": 0.0001446258776328987, + "loss": 1.2586, + "theoretical_loss": 3.3755372733513207, + "tokens_seen": 2710241280 + }, + { + "epoch": 0.82, + "learning_rate": 0.0001446098294884654, + "loss": 1.2444, + "theoretical_loss": 3.3755309214204865, + "tokens_seen": 2710306816 + }, + { + "epoch": 0.82, + "learning_rate": 0.0001445937813440321, + "loss": 1.3693, + "theoretical_loss": 3.375524569686246, + "tokens_seen": 2710372352 + }, + { + "epoch": 0.82, + "learning_rate": 0.0001445777331995988, + "loss": 1.3085, + "theoretical_loss": 3.375518218148589, + "tokens_seen": 2710437888 + }, + { + "epoch": 0.82, + "learning_rate": 0.0001445616850551655, + "loss": 1.4039, + "theoretical_loss": 3.3755118668075044, + "tokens_seen": 2710503424 + }, + { + "epoch": 0.82, + "learning_rate": 0.0001445456369107322, + "loss": 1.2319, + "theoretical_loss": 3.3755055156629816, + "tokens_seen": 2710568960 + }, + { + "epoch": 0.82, + "learning_rate": 0.00014452958876629888, + "loss": 1.2783, + "theoretical_loss": 3.375499164715009, + "tokens_seen": 2710634496 + }, + { + "epoch": 0.82, + "learning_rate": 0.00014451354062186558, + "loss": 1.3755, + "theoretical_loss": 3.375492813963577, + "tokens_seen": 2710700032 + }, + { + "epoch": 0.82, + "learning_rate": 0.00014449749247743228, + "loss": 1.4086, + "theoretical_loss": 3.3754864634086736, + "tokens_seen": 2710765568 + }, + { + "epoch": 0.82, + "learning_rate": 0.00014448144433299898, + "loss": 1.3186, + "theoretical_loss": 3.3754801130502887, + "tokens_seen": 2710831104 + }, + { + "epoch": 0.82, + "learning_rate": 0.00014446539618856569, + "loss": 1.3672, + "theoretical_loss": 3.3754737628884115, + "tokens_seen": 2710896640 + }, + { + "epoch": 0.82, + "learning_rate": 0.00014444934804413239, + "loss": 1.2968, + "theoretical_loss": 3.3754674129230304, + "tokens_seen": 2710962176 + }, + { + "epoch": 0.82, + "learning_rate": 0.00014443329989969911, + "loss": 1.289, + "theoretical_loss": 3.375461063154135, + "tokens_seen": 2711027712 + }, + { + "epoch": 0.82, + "learning_rate": 0.00014441725175526581, + "loss": 1.1758, + "theoretical_loss": 3.375454713581715, + "tokens_seen": 2711093248 + }, + { + "epoch": 0.82, + "learning_rate": 0.00014440120361083251, + "loss": 1.3257, + "theoretical_loss": 3.3754483642057584, + "tokens_seen": 2711158784 + }, + { + "epoch": 0.82, + "learning_rate": 0.00014438515546639922, + "loss": 1.2944, + "theoretical_loss": 3.375442015026256, + "tokens_seen": 2711224320 + }, + { + "epoch": 0.82, + "learning_rate": 0.00014436910732196592, + "loss": 1.3309, + "theoretical_loss": 3.3754356660431952, + "tokens_seen": 2711289856 + }, + { + "epoch": 0.82, + "learning_rate": 0.00014435305917753262, + "loss": 1.4577, + "theoretical_loss": 3.3754293172565664, + "tokens_seen": 2711355392 + }, + { + "epoch": 0.82, + "learning_rate": 0.00014433701103309932, + "loss": 1.2628, + "theoretical_loss": 3.375422968666358, + "tokens_seen": 2711420928 + }, + { + "epoch": 0.82, + "learning_rate": 0.00014432096288866602, + "loss": 1.364, + "theoretical_loss": 3.3754166202725604, + "tokens_seen": 2711486464 + }, + { + "epoch": 0.82, + "learning_rate": 0.00014430491474423272, + "loss": 1.3445, + "theoretical_loss": 3.3754102720751615, + "tokens_seen": 2711552000 + }, + { + "epoch": 0.82, + "learning_rate": 0.00014428886659979942, + "loss": 1.2073, + "theoretical_loss": 3.375403924074151, + "tokens_seen": 2711617536 + }, + { + "epoch": 0.82, + "learning_rate": 0.00014427281845536612, + "loss": 1.2136, + "theoretical_loss": 3.375397576269518, + "tokens_seen": 2711683072 + }, + { + "epoch": 0.82, + "learning_rate": 0.00014425677031093282, + "loss": 1.3789, + "theoretical_loss": 3.3753912286612513, + "tokens_seen": 2711748608 + }, + { + "epoch": 0.82, + "learning_rate": 0.00014424072216649952, + "loss": 1.2877, + "theoretical_loss": 3.3753848812493406, + "tokens_seen": 2711814144 + }, + { + "epoch": 0.82, + "learning_rate": 0.00014422467402206622, + "loss": 1.1636, + "theoretical_loss": 3.3753785340337754, + "tokens_seen": 2711879680 + }, + { + "epoch": 0.82, + "learning_rate": 0.0001442086258776329, + "loss": 1.2716, + "theoretical_loss": 3.375372187014544, + "tokens_seen": 2711945216 + }, + { + "epoch": 0.82, + "learning_rate": 0.0001441925777331996, + "loss": 1.4173, + "theoretical_loss": 3.375365840191636, + "tokens_seen": 2712010752 + }, + { + "epoch": 0.82, + "learning_rate": 0.0001441765295887663, + "loss": 1.3225, + "theoretical_loss": 3.375359493565041, + "tokens_seen": 2712076288 + }, + { + "epoch": 0.82, + "learning_rate": 0.000144160481444333, + "loss": 1.3273, + "theoretical_loss": 3.3753531471347475, + "tokens_seen": 2712141824 + }, + { + "epoch": 0.82, + "learning_rate": 0.0001441444332998997, + "loss": 1.2177, + "theoretical_loss": 3.375346800900745, + "tokens_seen": 2712207360 + }, + { + "epoch": 0.82, + "learning_rate": 0.0001441283851554664, + "loss": 1.3159, + "theoretical_loss": 3.3753404548630224, + "tokens_seen": 2712272896 + }, + { + "epoch": 0.82, + "learning_rate": 0.0001441123370110331, + "loss": 1.2587, + "theoretical_loss": 3.375334109021569, + "tokens_seen": 2712338432 + }, + { + "epoch": 0.82, + "learning_rate": 0.0001440962888665998, + "loss": 1.2898, + "theoretical_loss": 3.375327763376375, + "tokens_seen": 2712403968 + }, + { + "epoch": 0.82, + "learning_rate": 0.0001440802407221665, + "loss": 1.3801, + "theoretical_loss": 3.3753214179274282, + "tokens_seen": 2712469504 + }, + { + "epoch": 0.82, + "learning_rate": 0.0001440641925777332, + "loss": 1.3549, + "theoretical_loss": 3.3753150726747183, + "tokens_seen": 2712535040 + }, + { + "epoch": 0.82, + "learning_rate": 0.0001440481444332999, + "loss": 1.3715, + "theoretical_loss": 3.3753087276182345, + "tokens_seen": 2712600576 + }, + { + "epoch": 0.82, + "learning_rate": 0.0001440320962888666, + "loss": 1.4693, + "theoretical_loss": 3.375302382757966, + "tokens_seen": 2712666112 + }, + { + "epoch": 0.82, + "learning_rate": 0.0001440160481444333, + "loss": 1.1917, + "theoretical_loss": 3.375296038093902, + "tokens_seen": 2712731648 + }, + { + "epoch": 0.82, + "learning_rate": 0.000144, + "loss": 1.3927, + "theoretical_loss": 3.3752896936260313, + "tokens_seen": 2712797184 + }, + { + "epoch": 0.82, + "learning_rate": 0.0001439839518555667, + "loss": 1.3487, + "theoretical_loss": 3.3752833493543437, + "tokens_seen": 2712862720 + }, + { + "epoch": 0.82, + "learning_rate": 0.0001439679037111334, + "loss": 1.2601, + "theoretical_loss": 3.3752770052788286, + "tokens_seen": 2712928256 + }, + { + "epoch": 0.82, + "learning_rate": 0.0001439518555667001, + "loss": 1.3455, + "theoretical_loss": 3.3752706613994743, + "tokens_seen": 2712993792 + }, + { + "epoch": 0.82, + "learning_rate": 0.0001439358074222668, + "loss": 1.4117, + "theoretical_loss": 3.3752643177162702, + "tokens_seen": 2713059328 + }, + { + "epoch": 0.82, + "learning_rate": 0.0001439197592778335, + "loss": 1.4043, + "theoretical_loss": 3.375257974229206, + "tokens_seen": 2713124864 + }, + { + "epoch": 0.82, + "objective/train/docs_used": 985832, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.2908244132995605, + "objective/train/theoretical_loss": 3.375251630938271, + "objective/train/tokens_used": 2733650400, + "theoretical_loss": 3.375251630938271, + "tokens_seen": 2713190400 + }, + { + "epoch": 0.82, + "learning_rate": 0.0001439037111334002, + "loss": 1.3655, + "theoretical_loss": 3.375251630938271, + "tokens_seen": 2713190400 + }, + { + "epoch": 0.82, + "learning_rate": 0.0001438876629889669, + "loss": 1.2478, + "theoretical_loss": 3.3752452878434536, + "tokens_seen": 2713255936 + }, + { + "epoch": 0.82, + "learning_rate": 0.0001438716148445336, + "loss": 1.2707, + "theoretical_loss": 3.3752389449447433, + "tokens_seen": 2713321472 + }, + { + "epoch": 0.82, + "learning_rate": 0.0001438555667001003, + "loss": 1.2592, + "theoretical_loss": 3.37523260224213, + "tokens_seen": 2713387008 + }, + { + "epoch": 0.82, + "learning_rate": 0.000143839518555667, + "loss": 1.3228, + "theoretical_loss": 3.375226259735602, + "tokens_seen": 2713452544 + }, + { + "epoch": 0.82, + "learning_rate": 0.0001438234704112337, + "loss": 1.3238, + "theoretical_loss": 3.3752199174251487, + "tokens_seen": 2713518080 + }, + { + "epoch": 0.82, + "learning_rate": 0.0001438074222668004, + "loss": 1.3931, + "theoretical_loss": 3.37521357531076, + "tokens_seen": 2713583616 + }, + { + "epoch": 0.82, + "learning_rate": 0.0001437913741223671, + "loss": 1.221, + "theoretical_loss": 3.375207233392424, + "tokens_seen": 2713649152 + }, + { + "epoch": 0.82, + "learning_rate": 0.0001437753259779338, + "loss": 1.3016, + "theoretical_loss": 3.3752008916701306, + "tokens_seen": 2713714688 + }, + { + "epoch": 0.82, + "learning_rate": 0.0001437592778335005, + "loss": 1.2819, + "theoretical_loss": 3.375194550143869, + "tokens_seen": 2713780224 + }, + { + "epoch": 0.82, + "learning_rate": 0.00014374322968906721, + "loss": 1.28, + "theoretical_loss": 3.3751882088136282, + "tokens_seen": 2713845760 + }, + { + "epoch": 0.82, + "learning_rate": 0.00014372718154463391, + "loss": 1.2103, + "theoretical_loss": 3.375181867679397, + "tokens_seen": 2713911296 + }, + { + "epoch": 0.82, + "learning_rate": 0.00014371113340020062, + "loss": 1.3805, + "theoretical_loss": 3.375175526741166, + "tokens_seen": 2713976832 + }, + { + "epoch": 0.82, + "learning_rate": 0.00014369508525576732, + "loss": 1.2013, + "theoretical_loss": 3.375169185998923, + "tokens_seen": 2714042368 + }, + { + "epoch": 0.82, + "learning_rate": 0.00014367903711133402, + "loss": 1.2153, + "theoretical_loss": 3.3751628454526577, + "tokens_seen": 2714107904 + }, + { + "epoch": 0.82, + "learning_rate": 0.00014366298896690072, + "loss": 1.233, + "theoretical_loss": 3.3751565051023595, + "tokens_seen": 2714173440 + }, + { + "epoch": 0.82, + "learning_rate": 0.00014364694082246742, + "loss": 1.2703, + "theoretical_loss": 3.3751501649480176, + "tokens_seen": 2714238976 + }, + { + "epoch": 0.82, + "learning_rate": 0.00014363089267803412, + "loss": 1.2829, + "theoretical_loss": 3.3751438249896206, + "tokens_seen": 2714304512 + }, + { + "epoch": 0.82, + "learning_rate": 0.00014361484453360082, + "loss": 1.3395, + "theoretical_loss": 3.375137485227158, + "tokens_seen": 2714370048 + }, + { + "epoch": 0.82, + "learning_rate": 0.00014359879638916752, + "loss": 1.2722, + "theoretical_loss": 3.3751311456606197, + "tokens_seen": 2714435584 + }, + { + "epoch": 0.82, + "learning_rate": 0.00014358274824473422, + "loss": 1.4232, + "theoretical_loss": 3.375124806289994, + "tokens_seen": 2714501120 + }, + { + "epoch": 0.82, + "learning_rate": 0.00014356670010030092, + "loss": 1.3389, + "theoretical_loss": 3.3751184671152705, + "tokens_seen": 2714566656 + }, + { + "epoch": 0.82, + "learning_rate": 0.00014355065195586762, + "loss": 1.2487, + "theoretical_loss": 3.375112128136439, + "tokens_seen": 2714632192 + }, + { + "epoch": 0.82, + "learning_rate": 0.00014353460381143432, + "loss": 1.2304, + "theoretical_loss": 3.3751057893534875, + "tokens_seen": 2714697728 + }, + { + "epoch": 0.82, + "learning_rate": 0.000143518555667001, + "loss": 1.2388, + "theoretical_loss": 3.375099450766406, + "tokens_seen": 2714763264 + }, + { + "epoch": 0.82, + "learning_rate": 0.0001435025075225677, + "loss": 1.2283, + "theoretical_loss": 3.3750931123751835, + "tokens_seen": 2714828800 + }, + { + "epoch": 0.82, + "learning_rate": 0.0001434864593781344, + "loss": 1.2954, + "theoretical_loss": 3.375086774179809, + "tokens_seen": 2714894336 + }, + { + "epoch": 0.82, + "learning_rate": 0.0001434704112337011, + "loss": 1.2643, + "theoretical_loss": 3.375080436180273, + "tokens_seen": 2714959872 + }, + { + "epoch": 0.82, + "learning_rate": 0.0001434543630892678, + "loss": 1.3615, + "theoretical_loss": 3.3750740983765626, + "tokens_seen": 2715025408 + }, + { + "epoch": 0.82, + "learning_rate": 0.0001434383149448345, + "loss": 1.1761, + "theoretical_loss": 3.375067760768669, + "tokens_seen": 2715090944 + }, + { + "epoch": 0.82, + "learning_rate": 0.0001434222668004012, + "loss": 1.2406, + "theoretical_loss": 3.37506142335658, + "tokens_seen": 2715156480 + }, + { + "epoch": 0.82, + "learning_rate": 0.0001434062186559679, + "loss": 1.2152, + "theoretical_loss": 3.3750550861402857, + "tokens_seen": 2715222016 + }, + { + "epoch": 0.82, + "learning_rate": 0.0001433901705115346, + "loss": 1.3308, + "theoretical_loss": 3.375048749119775, + "tokens_seen": 2715287552 + }, + { + "epoch": 0.82, + "learning_rate": 0.0001433741223671013, + "loss": 1.3341, + "theoretical_loss": 3.3750424122950373, + "tokens_seen": 2715353088 + }, + { + "epoch": 0.82, + "learning_rate": 0.000143358074222668, + "loss": 1.3326, + "theoretical_loss": 3.3750360756660616, + "tokens_seen": 2715418624 + }, + { + "epoch": 0.82, + "learning_rate": 0.0001433420260782347, + "loss": 1.4106, + "theoretical_loss": 3.3750297392328372, + "tokens_seen": 2715484160 + }, + { + "epoch": 0.82, + "learning_rate": 0.0001433259779338014, + "loss": 1.2463, + "theoretical_loss": 3.375023402995353, + "tokens_seen": 2715549696 + }, + { + "epoch": 0.82, + "learning_rate": 0.0001433099297893681, + "loss": 1.1798, + "theoretical_loss": 3.3750170669535993, + "tokens_seen": 2715615232 + }, + { + "epoch": 0.82, + "learning_rate": 0.0001432938816449348, + "loss": 1.2496, + "theoretical_loss": 3.375010731107564, + "tokens_seen": 2715680768 + }, + { + "epoch": 0.82, + "learning_rate": 0.0001432778335005015, + "loss": 1.4291, + "theoretical_loss": 3.375004395457237, + "tokens_seen": 2715746304 + }, + { + "epoch": 0.82, + "learning_rate": 0.0001432617853560682, + "loss": 1.4489, + "theoretical_loss": 3.3749980600026075, + "tokens_seen": 2715811840 + }, + { + "epoch": 0.82, + "learning_rate": 0.0001432457372116349, + "loss": 1.2532, + "theoretical_loss": 3.3749917247436647, + "tokens_seen": 2715877376 + }, + { + "epoch": 0.82, + "learning_rate": 0.00014322968906720163, + "loss": 1.3015, + "theoretical_loss": 3.374985389680398, + "tokens_seen": 2715942912 + }, + { + "epoch": 0.82, + "learning_rate": 0.00014321364092276833, + "loss": 1.3027, + "theoretical_loss": 3.3749790548127963, + "tokens_seen": 2716008448 + }, + { + "epoch": 0.82, + "learning_rate": 0.000143197592778335, + "loss": 1.2828, + "theoretical_loss": 3.3749727201408493, + "tokens_seen": 2716073984 + }, + { + "epoch": 0.82, + "learning_rate": 0.0001431815446339017, + "loss": 1.2649, + "theoretical_loss": 3.3749663856645453, + "tokens_seen": 2716139520 + }, + { + "epoch": 0.82, + "learning_rate": 0.0001431654964894684, + "loss": 1.3109, + "theoretical_loss": 3.3749600513838747, + "tokens_seen": 2716205056 + }, + { + "epoch": 0.82, + "learning_rate": 0.0001431494483450351, + "loss": 1.4376, + "theoretical_loss": 3.374953717298826, + "tokens_seen": 2716270592 + }, + { + "epoch": 0.82, + "learning_rate": 0.0001431334002006018, + "loss": 1.2493, + "theoretical_loss": 3.374947383409389, + "tokens_seen": 2716336128 + }, + { + "epoch": 0.82, + "learning_rate": 0.0001431173520561685, + "loss": 1.2635, + "theoretical_loss": 3.3749410497155523, + "tokens_seen": 2716401664 + }, + { + "epoch": 0.82, + "objective/train/docs_used": 987010, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.1263468265533447, + "objective/train/theoretical_loss": 3.374934716217305, + "objective/train/tokens_used": 2736927200, + "theoretical_loss": 3.374934716217305, + "tokens_seen": 2716467200 + }, + { + "epoch": 0.82, + "learning_rate": 0.0001431013039117352, + "loss": 1.2836, + "theoretical_loss": 3.374934716217305, + "tokens_seen": 2716467200 + }, + { + "epoch": 0.82, + "learning_rate": 0.0001430852557673019, + "loss": 1.2482, + "theoretical_loss": 3.3749283829146375, + "tokens_seen": 2716532736 + }, + { + "epoch": 0.82, + "learning_rate": 0.0001430692076228686, + "loss": 1.3245, + "theoretical_loss": 3.374922049807538, + "tokens_seen": 2716598272 + }, + { + "epoch": 0.82, + "learning_rate": 0.00014305315947843531, + "loss": 1.3829, + "theoretical_loss": 3.374915716895996, + "tokens_seen": 2716663808 + }, + { + "epoch": 0.82, + "learning_rate": 0.00014303711133400201, + "loss": 1.2599, + "theoretical_loss": 3.374909384180001, + "tokens_seen": 2716729344 + }, + { + "epoch": 0.82, + "learning_rate": 0.00014302106318956872, + "loss": 1.2548, + "theoretical_loss": 3.374903051659542, + "tokens_seen": 2716794880 + }, + { + "epoch": 0.82, + "learning_rate": 0.00014300501504513542, + "loss": 1.2727, + "theoretical_loss": 3.374896719334608, + "tokens_seen": 2716860416 + }, + { + "epoch": 0.82, + "learning_rate": 0.00014298896690070212, + "loss": 1.323, + "theoretical_loss": 3.374890387205189, + "tokens_seen": 2716925952 + }, + { + "epoch": 0.82, + "learning_rate": 0.00014297291875626882, + "loss": 1.3213, + "theoretical_loss": 3.3748840552712736, + "tokens_seen": 2716991488 + }, + { + "epoch": 0.82, + "learning_rate": 0.00014295687061183552, + "loss": 1.4319, + "theoretical_loss": 3.374877723532851, + "tokens_seen": 2717057024 + }, + { + "epoch": 0.82, + "learning_rate": 0.00014294082246740222, + "loss": 1.4686, + "theoretical_loss": 3.374871391989911, + "tokens_seen": 2717122560 + }, + { + "epoch": 0.82, + "learning_rate": 0.00014292477432296892, + "loss": 1.3479, + "theoretical_loss": 3.3748650606424424, + "tokens_seen": 2717188096 + }, + { + "epoch": 0.82, + "learning_rate": 0.00014290872617853562, + "loss": 1.3356, + "theoretical_loss": 3.3748587294904344, + "tokens_seen": 2717253632 + }, + { + "epoch": 0.82, + "learning_rate": 0.00014289267803410232, + "loss": 1.4347, + "theoretical_loss": 3.3748523985338768, + "tokens_seen": 2717319168 + }, + { + "epoch": 0.82, + "learning_rate": 0.00014287662988966902, + "loss": 1.433, + "theoretical_loss": 3.3748460677727583, + "tokens_seen": 2717384704 + }, + { + "epoch": 0.82, + "learning_rate": 0.00014286058174523572, + "loss": 1.3592, + "theoretical_loss": 3.3748397372070684, + "tokens_seen": 2717450240 + }, + { + "epoch": 0.82, + "learning_rate": 0.00014284453360080242, + "loss": 1.3769, + "theoretical_loss": 3.374833406836796, + "tokens_seen": 2717515776 + }, + { + "epoch": 0.82, + "learning_rate": 0.00014282848545636912, + "loss": 1.215, + "theoretical_loss": 3.3748270766619313, + "tokens_seen": 2717581312 + }, + { + "epoch": 0.82, + "learning_rate": 0.0001428124373119358, + "loss": 1.3014, + "theoretical_loss": 3.3748207466824622, + "tokens_seen": 2717646848 + }, + { + "epoch": 0.82, + "learning_rate": 0.0001427963891675025, + "loss": 1.2775, + "theoretical_loss": 3.374814416898379, + "tokens_seen": 2717712384 + }, + { + "epoch": 0.82, + "learning_rate": 0.0001427803410230692, + "loss": 1.3082, + "theoretical_loss": 3.374808087309671, + "tokens_seen": 2717777920 + }, + { + "epoch": 0.82, + "learning_rate": 0.0001427642928786359, + "loss": 1.2146, + "theoretical_loss": 3.3748017579163267, + "tokens_seen": 2717843456 + }, + { + "epoch": 0.82, + "learning_rate": 0.0001427482447342026, + "loss": 1.2775, + "theoretical_loss": 3.3747954287183357, + "tokens_seen": 2717908992 + }, + { + "epoch": 0.82, + "learning_rate": 0.0001427321965897693, + "loss": 1.3543, + "theoretical_loss": 3.374789099715687, + "tokens_seen": 2717974528 + }, + { + "epoch": 0.82, + "learning_rate": 0.000142716148445336, + "loss": 1.2373, + "theoretical_loss": 3.3747827709083706, + "tokens_seen": 2718040064 + }, + { + "epoch": 0.82, + "learning_rate": 0.0001427001003009027, + "loss": 1.2624, + "theoretical_loss": 3.3747764422963753, + "tokens_seen": 2718105600 + }, + { + "epoch": 0.82, + "learning_rate": 0.0001426840521564694, + "loss": 1.2413, + "theoretical_loss": 3.3747701138796904, + "tokens_seen": 2718171136 + }, + { + "epoch": 0.82, + "learning_rate": 0.0001426680040120361, + "loss": 1.3149, + "theoretical_loss": 3.3747637856583053, + "tokens_seen": 2718236672 + }, + { + "epoch": 0.82, + "learning_rate": 0.0001426519558676028, + "loss": 1.3548, + "theoretical_loss": 3.374757457632209, + "tokens_seen": 2718302208 + }, + { + "epoch": 0.82, + "learning_rate": 0.00014263590772316953, + "loss": 1.3627, + "theoretical_loss": 3.3747511298013904, + "tokens_seen": 2718367744 + }, + { + "epoch": 0.82, + "learning_rate": 0.00014261985957873623, + "loss": 1.302, + "theoretical_loss": 3.3747448021658397, + "tokens_seen": 2718433280 + }, + { + "epoch": 0.82, + "learning_rate": 0.00014260381143430293, + "loss": 1.3339, + "theoretical_loss": 3.374738474725546, + "tokens_seen": 2718498816 + }, + { + "epoch": 0.82, + "learning_rate": 0.00014258776328986963, + "loss": 1.2135, + "theoretical_loss": 3.3747321474804974, + "tokens_seen": 2718564352 + }, + { + "epoch": 0.82, + "learning_rate": 0.00014257171514543633, + "loss": 1.2435, + "theoretical_loss": 3.374725820430685, + "tokens_seen": 2718629888 + }, + { + "epoch": 0.82, + "learning_rate": 0.00014255566700100303, + "loss": 1.291, + "theoretical_loss": 3.3747194935760962, + "tokens_seen": 2718695424 + }, + { + "epoch": 0.82, + "learning_rate": 0.00014253961885656973, + "loss": 1.189, + "theoretical_loss": 3.374713166916722, + "tokens_seen": 2718760960 + }, + { + "epoch": 0.82, + "learning_rate": 0.00014252357071213643, + "loss": 1.239, + "theoretical_loss": 3.3747068404525504, + "tokens_seen": 2718826496 + }, + { + "epoch": 0.82, + "learning_rate": 0.00014250752256770314, + "loss": 1.2507, + "theoretical_loss": 3.3747005141835715, + "tokens_seen": 2718892032 + }, + { + "epoch": 0.82, + "learning_rate": 0.0001424914744232698, + "loss": 1.2014, + "theoretical_loss": 3.374694188109774, + "tokens_seen": 2718957568 + }, + { + "epoch": 0.82, + "learning_rate": 0.0001424754262788365, + "loss": 1.3475, + "theoretical_loss": 3.3746878622311476, + "tokens_seen": 2719023104 + }, + { + "epoch": 0.82, + "learning_rate": 0.0001424593781344032, + "loss": 1.3492, + "theoretical_loss": 3.3746815365476808, + "tokens_seen": 2719088640 + }, + { + "epoch": 0.82, + "learning_rate": 0.0001424433299899699, + "loss": 1.2499, + "theoretical_loss": 3.3746752110593636, + "tokens_seen": 2719154176 + }, + { + "epoch": 0.82, + "learning_rate": 0.0001424272818455366, + "loss": 1.373, + "theoretical_loss": 3.3746688857661855, + "tokens_seen": 2719219712 + }, + { + "epoch": 0.82, + "learning_rate": 0.0001424112337011033, + "loss": 1.3055, + "theoretical_loss": 3.3746625606681353, + "tokens_seen": 2719285248 + }, + { + "epoch": 0.82, + "learning_rate": 0.00014239518555667, + "loss": 1.4027, + "theoretical_loss": 3.374656235765202, + "tokens_seen": 2719350784 + }, + { + "epoch": 0.82, + "learning_rate": 0.0001423791374122367, + "loss": 1.2729, + "theoretical_loss": 3.3746499110573756, + "tokens_seen": 2719416320 + }, + { + "epoch": 0.82, + "learning_rate": 0.00014236308926780341, + "loss": 1.3417, + "theoretical_loss": 3.3746435865446447, + "tokens_seen": 2719481856 + }, + { + "epoch": 0.82, + "learning_rate": 0.00014234704112337011, + "loss": 1.2096, + "theoretical_loss": 3.374637262226999, + "tokens_seen": 2719547392 + }, + { + "epoch": 0.82, + "learning_rate": 0.00014233099297893682, + "loss": 1.3076, + "theoretical_loss": 3.374630938104428, + "tokens_seen": 2719612928 + }, + { + "epoch": 0.82, + "learning_rate": 0.00014231494483450352, + "loss": 1.3231, + "theoretical_loss": 3.3746246141769207, + "tokens_seen": 2719678464 + }, + { + "epoch": 0.82, + "objective/train/docs_used": 988161, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.332650899887085, + "objective/train/theoretical_loss": 3.3746182904444657, + "objective/train/tokens_used": 2740204000, + "theoretical_loss": 3.3746182904444657, + "tokens_seen": 2719744000 + }, + { + "epoch": 0.82, + "learning_rate": 0.00014229889669007022, + "loss": 1.2397, + "theoretical_loss": 3.3746182904444657, + "tokens_seen": 2719744000 + }, + { + "epoch": 0.82, + "learning_rate": 0.00014228284854563692, + "loss": 1.2864, + "theoretical_loss": 3.374611966907054, + "tokens_seen": 2719809536 + }, + { + "epoch": 0.82, + "learning_rate": 0.00014226680040120362, + "loss": 1.2732, + "theoretical_loss": 3.3746056435646725, + "tokens_seen": 2719875072 + }, + { + "epoch": 0.82, + "learning_rate": 0.00014225075225677032, + "loss": 1.3867, + "theoretical_loss": 3.3745993204173126, + "tokens_seen": 2719940608 + }, + { + "epoch": 0.82, + "learning_rate": 0.00014223470411233702, + "loss": 1.2075, + "theoretical_loss": 3.374592997464963, + "tokens_seen": 2720006144 + }, + { + "epoch": 0.82, + "learning_rate": 0.00014221865596790372, + "loss": 1.4007, + "theoretical_loss": 3.3745866747076123, + "tokens_seen": 2720071680 + }, + { + "epoch": 0.82, + "learning_rate": 0.00014220260782347042, + "loss": 1.3244, + "theoretical_loss": 3.37458035214525, + "tokens_seen": 2720137216 + }, + { + "epoch": 0.82, + "learning_rate": 0.00014218655967903712, + "loss": 1.3339, + "theoretical_loss": 3.3745740297778664, + "tokens_seen": 2720202752 + }, + { + "epoch": 0.82, + "learning_rate": 0.00014217051153460382, + "loss": 1.3976, + "theoretical_loss": 3.3745677076054497, + "tokens_seen": 2720268288 + }, + { + "epoch": 0.82, + "learning_rate": 0.00014215446339017052, + "loss": 1.3621, + "theoretical_loss": 3.3745613856279895, + "tokens_seen": 2720333824 + }, + { + "epoch": 0.82, + "learning_rate": 0.00014213841524573722, + "loss": 1.3777, + "theoretical_loss": 3.374555063845475, + "tokens_seen": 2720399360 + }, + { + "epoch": 0.82, + "learning_rate": 0.00014212236710130392, + "loss": 1.2924, + "theoretical_loss": 3.374548742257896, + "tokens_seen": 2720464896 + }, + { + "epoch": 0.82, + "learning_rate": 0.0001421063189568706, + "loss": 1.3031, + "theoretical_loss": 3.3745424208652413, + "tokens_seen": 2720530432 + }, + { + "epoch": 0.82, + "learning_rate": 0.0001420902708124373, + "loss": 1.4075, + "theoretical_loss": 3.3745360996675, + "tokens_seen": 2720595968 + }, + { + "epoch": 0.82, + "learning_rate": 0.000142074222668004, + "loss": 1.3902, + "theoretical_loss": 3.374529778664662, + "tokens_seen": 2720661504 + }, + { + "epoch": 0.82, + "learning_rate": 0.0001420581745235707, + "loss": 1.2421, + "theoretical_loss": 3.374523457856716, + "tokens_seen": 2720727040 + }, + { + "epoch": 0.82, + "learning_rate": 0.0001420421263791374, + "loss": 1.3323, + "theoretical_loss": 3.374517137243652, + "tokens_seen": 2720792576 + }, + { + "epoch": 0.82, + "learning_rate": 0.00014202607823470413, + "loss": 1.242, + "theoretical_loss": 3.3745108168254587, + "tokens_seen": 2720858112 + }, + { + "epoch": 0.82, + "learning_rate": 0.00014201003009027083, + "loss": 1.1833, + "theoretical_loss": 3.3745044966021256, + "tokens_seen": 2720923648 + }, + { + "epoch": 0.82, + "learning_rate": 0.00014199398194583753, + "loss": 1.3003, + "theoretical_loss": 3.374498176573642, + "tokens_seen": 2720989184 + }, + { + "epoch": 0.82, + "learning_rate": 0.00014197793380140423, + "loss": 1.2719, + "theoretical_loss": 3.374491856739997, + "tokens_seen": 2721054720 + }, + { + "epoch": 0.82, + "learning_rate": 0.00014196188565697093, + "loss": 1.2259, + "theoretical_loss": 3.3744855371011804, + "tokens_seen": 2721120256 + }, + { + "epoch": 0.82, + "learning_rate": 0.00014194583751253763, + "loss": 1.3197, + "theoretical_loss": 3.3744792176571807, + "tokens_seen": 2721185792 + }, + { + "epoch": 0.82, + "learning_rate": 0.00014192978936810433, + "loss": 1.2422, + "theoretical_loss": 3.3744728984079884, + "tokens_seen": 2721251328 + }, + { + "epoch": 0.82, + "learning_rate": 0.00014191374122367103, + "loss": 1.3486, + "theoretical_loss": 3.3744665793535917, + "tokens_seen": 2721316864 + }, + { + "epoch": 0.82, + "learning_rate": 0.00014189769307923773, + "loss": 1.3285, + "theoretical_loss": 3.37446026049398, + "tokens_seen": 2721382400 + }, + { + "epoch": 0.82, + "learning_rate": 0.00014188164493480443, + "loss": 1.2326, + "theoretical_loss": 3.3744539418291435, + "tokens_seen": 2721447936 + }, + { + "epoch": 0.82, + "learning_rate": 0.00014186559679037113, + "loss": 1.2777, + "theoretical_loss": 3.3744476233590706, + "tokens_seen": 2721513472 + }, + { + "epoch": 0.82, + "learning_rate": 0.00014184954864593783, + "loss": 1.3888, + "theoretical_loss": 3.374441305083751, + "tokens_seen": 2721579008 + }, + { + "epoch": 0.82, + "learning_rate": 0.00014183350050150454, + "loss": 1.3248, + "theoretical_loss": 3.3744349870031742, + "tokens_seen": 2721644544 + }, + { + "epoch": 0.82, + "learning_rate": 0.00014181745235707124, + "loss": 1.2625, + "theoretical_loss": 3.3744286691173286, + "tokens_seen": 2721710080 + }, + { + "epoch": 0.82, + "learning_rate": 0.0001418014042126379, + "loss": 1.1543, + "theoretical_loss": 3.3744223514262046, + "tokens_seen": 2721775616 + }, + { + "epoch": 0.82, + "learning_rate": 0.0001417853560682046, + "loss": 1.2849, + "theoretical_loss": 3.374416033929791, + "tokens_seen": 2721841152 + }, + { + "epoch": 0.82, + "learning_rate": 0.0001417693079237713, + "loss": 1.3331, + "theoretical_loss": 3.3744097166280773, + "tokens_seen": 2721906688 + }, + { + "epoch": 0.82, + "learning_rate": 0.000141753259779338, + "loss": 1.217, + "theoretical_loss": 3.3744033995210523, + "tokens_seen": 2721972224 + }, + { + "epoch": 0.82, + "learning_rate": 0.0001417372116349047, + "loss": 1.4018, + "theoretical_loss": 3.374397082608706, + "tokens_seen": 2722037760 + }, + { + "epoch": 0.82, + "learning_rate": 0.0001417211634904714, + "loss": 1.3006, + "theoretical_loss": 3.3743907658910275, + "tokens_seen": 2722103296 + }, + { + "epoch": 0.82, + "learning_rate": 0.0001417051153460381, + "loss": 1.1856, + "theoretical_loss": 3.374384449368006, + "tokens_seen": 2722168832 + }, + { + "epoch": 0.82, + "learning_rate": 0.00014168906720160481, + "loss": 1.1717, + "theoretical_loss": 3.3743781330396305, + "tokens_seen": 2722234368 + }, + { + "epoch": 0.82, + "learning_rate": 0.00014167301905717151, + "loss": 1.2894, + "theoretical_loss": 3.374371816905891, + "tokens_seen": 2722299904 + }, + { + "epoch": 0.82, + "learning_rate": 0.00014165697091273821, + "loss": 1.1986, + "theoretical_loss": 3.374365500966776, + "tokens_seen": 2722365440 + }, + { + "epoch": 0.82, + "learning_rate": 0.00014164092276830492, + "loss": 1.3613, + "theoretical_loss": 3.374359185222276, + "tokens_seen": 2722430976 + }, + { + "epoch": 0.82, + "learning_rate": 0.00014162487462387162, + "loss": 1.2946, + "theoretical_loss": 3.374352869672379, + "tokens_seen": 2722496512 + }, + { + "epoch": 0.83, + "learning_rate": 0.00014160882647943832, + "loss": 1.2597, + "theoretical_loss": 3.374346554317075, + "tokens_seen": 2722562048 + }, + { + "epoch": 0.83, + "learning_rate": 0.00014159277833500502, + "loss": 1.2628, + "theoretical_loss": 3.3743402391563535, + "tokens_seen": 2722627584 + }, + { + "epoch": 0.83, + "learning_rate": 0.00014157673019057172, + "loss": 1.2942, + "theoretical_loss": 3.3743339241902035, + "tokens_seen": 2722693120 + }, + { + "epoch": 0.83, + "learning_rate": 0.00014156068204613842, + "loss": 1.1675, + "theoretical_loss": 3.3743276094186143, + "tokens_seen": 2722758656 + }, + { + "epoch": 0.83, + "learning_rate": 0.00014154463390170512, + "loss": 1.1936, + "theoretical_loss": 3.3743212948415753, + "tokens_seen": 2722824192 + }, + { + "epoch": 0.83, + "learning_rate": 0.00014152858575727182, + "loss": 1.3074, + "theoretical_loss": 3.3743149804590757, + "tokens_seen": 2722889728 + }, + { + "epoch": 0.83, + "learning_rate": 0.00014151253761283852, + "loss": 1.2952, + "theoretical_loss": 3.374308666271105, + "tokens_seen": 2722955264 + }, + { + "epoch": 0.83, + "objective/train/docs_used": 989325, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.2467668056488037, + "objective/train/theoretical_loss": 3.374302352277653, + "objective/train/tokens_used": 2743480800, + "theoretical_loss": 3.374302352277653, + "tokens_seen": 2723020800 + }, + { + "epoch": 0.83, + "learning_rate": 0.00014149648946840522, + "loss": 1.2612, + "theoretical_loss": 3.374302352277653, + "tokens_seen": 2723020800 + }, + { + "epoch": 0.83, + "learning_rate": 0.00014148044132397192, + "loss": 1.2429, + "theoretical_loss": 3.374296038478708, + "tokens_seen": 2723086336 + }, + { + "epoch": 0.83, + "learning_rate": 0.00014146439317953862, + "loss": 1.2458, + "theoretical_loss": 3.37428972487426, + "tokens_seen": 2723151872 + }, + { + "epoch": 0.83, + "learning_rate": 0.00014144834503510532, + "loss": 1.2509, + "theoretical_loss": 3.3742834114642983, + "tokens_seen": 2723217408 + }, + { + "epoch": 0.83, + "learning_rate": 0.00014143229689067202, + "loss": 1.2779, + "theoretical_loss": 3.374277098248812, + "tokens_seen": 2723282944 + }, + { + "epoch": 0.83, + "learning_rate": 0.00014141624874623872, + "loss": 1.1955, + "theoretical_loss": 3.3742707852277904, + "tokens_seen": 2723348480 + }, + { + "epoch": 0.83, + "learning_rate": 0.00014140020060180543, + "loss": 1.2666, + "theoretical_loss": 3.3742644724012227, + "tokens_seen": 2723414016 + }, + { + "epoch": 0.83, + "learning_rate": 0.00014138415245737213, + "loss": 1.2341, + "theoretical_loss": 3.3742581597690986, + "tokens_seen": 2723479552 + }, + { + "epoch": 0.83, + "learning_rate": 0.00014136810431293883, + "loss": 1.2568, + "theoretical_loss": 3.374251847331408, + "tokens_seen": 2723545088 + }, + { + "epoch": 0.83, + "learning_rate": 0.00014135205616850553, + "loss": 1.378, + "theoretical_loss": 3.374245535088139, + "tokens_seen": 2723610624 + }, + { + "epoch": 0.83, + "learning_rate": 0.00014133600802407223, + "loss": 1.3236, + "theoretical_loss": 3.3742392230392815, + "tokens_seen": 2723676160 + }, + { + "epoch": 0.83, + "learning_rate": 0.00014131995987963893, + "loss": 1.1481, + "theoretical_loss": 3.3742329111848246, + "tokens_seen": 2723741696 + }, + { + "epoch": 0.83, + "learning_rate": 0.00014130391173520563, + "loss": 1.2816, + "theoretical_loss": 3.3742265995247585, + "tokens_seen": 2723807232 + }, + { + "epoch": 0.83, + "learning_rate": 0.00014128786359077233, + "loss": 1.3184, + "theoretical_loss": 3.3742202880590715, + "tokens_seen": 2723872768 + }, + { + "epoch": 0.83, + "learning_rate": 0.00014127181544633903, + "loss": 1.3727, + "theoretical_loss": 3.3742139767877535, + "tokens_seen": 2723938304 + }, + { + "epoch": 0.83, + "learning_rate": 0.00014125576730190573, + "loss": 1.2273, + "theoretical_loss": 3.3742076657107933, + "tokens_seen": 2724003840 + }, + { + "epoch": 0.83, + "learning_rate": 0.00014123971915747243, + "loss": 1.1423, + "theoretical_loss": 3.374201354828181, + "tokens_seen": 2724069376 + }, + { + "epoch": 0.83, + "learning_rate": 0.00014122367101303913, + "loss": 1.3334, + "theoretical_loss": 3.3741950441399053, + "tokens_seen": 2724134912 + }, + { + "epoch": 0.83, + "learning_rate": 0.00014120762286860583, + "loss": 1.354, + "theoretical_loss": 3.3741887336459557, + "tokens_seen": 2724200448 + }, + { + "epoch": 0.83, + "learning_rate": 0.00014119157472417253, + "loss": 1.2431, + "theoretical_loss": 3.374182423346322, + "tokens_seen": 2724265984 + }, + { + "epoch": 0.83, + "learning_rate": 0.00014117552657973923, + "loss": 1.2265, + "theoretical_loss": 3.374176113240993, + "tokens_seen": 2724331520 + }, + { + "epoch": 0.83, + "learning_rate": 0.00014115947843530593, + "loss": 1.2482, + "theoretical_loss": 3.3741698033299583, + "tokens_seen": 2724397056 + }, + { + "epoch": 0.83, + "learning_rate": 0.00014114343029087264, + "loss": 1.2832, + "theoretical_loss": 3.374163493613207, + "tokens_seen": 2724462592 + }, + { + "epoch": 0.83, + "learning_rate": 0.00014112738214643934, + "loss": 1.2265, + "theoretical_loss": 3.3741571840907287, + "tokens_seen": 2724528128 + }, + { + "epoch": 0.83, + "learning_rate": 0.00014111133400200604, + "loss": 1.2591, + "theoretical_loss": 3.3741508747625124, + "tokens_seen": 2724593664 + }, + { + "epoch": 0.83, + "learning_rate": 0.0001410952858575727, + "loss": 1.2844, + "theoretical_loss": 3.374144565628548, + "tokens_seen": 2724659200 + }, + { + "epoch": 0.83, + "learning_rate": 0.0001410792377131394, + "loss": 1.2747, + "theoretical_loss": 3.3741382566888243, + "tokens_seen": 2724724736 + }, + { + "epoch": 0.83, + "learning_rate": 0.0001410631895687061, + "loss": 1.2467, + "theoretical_loss": 3.374131947943331, + "tokens_seen": 2724790272 + }, + { + "epoch": 0.83, + "learning_rate": 0.0001410471414242728, + "loss": 1.3275, + "theoretical_loss": 3.3741256393920573, + "tokens_seen": 2724855808 + }, + { + "epoch": 0.83, + "learning_rate": 0.0001410310932798395, + "loss": 1.2941, + "theoretical_loss": 3.3741193310349926, + "tokens_seen": 2724921344 + }, + { + "epoch": 0.83, + "learning_rate": 0.0001410150451354062, + "loss": 1.375, + "theoretical_loss": 3.3741130228721263, + "tokens_seen": 2724986880 + }, + { + "epoch": 0.83, + "learning_rate": 0.00014099899699097291, + "loss": 1.1512, + "theoretical_loss": 3.374106714903448, + "tokens_seen": 2725052416 + }, + { + "epoch": 0.83, + "learning_rate": 0.00014098294884653961, + "loss": 1.2467, + "theoretical_loss": 3.3741004071289464, + "tokens_seen": 2725117952 + }, + { + "epoch": 0.83, + "learning_rate": 0.00014096690070210632, + "loss": 1.3467, + "theoretical_loss": 3.374094099548611, + "tokens_seen": 2725183488 + }, + { + "epoch": 0.83, + "learning_rate": 0.00014095085255767302, + "loss": 1.254, + "theoretical_loss": 3.374087792162432, + "tokens_seen": 2725249024 + }, + { + "epoch": 0.83, + "learning_rate": 0.00014093480441323972, + "loss": 1.3684, + "theoretical_loss": 3.3740814849703975, + "tokens_seen": 2725314560 + }, + { + "epoch": 0.83, + "learning_rate": 0.00014091875626880642, + "loss": 1.2026, + "theoretical_loss": 3.3740751779724976, + "tokens_seen": 2725380096 + }, + { + "epoch": 0.83, + "learning_rate": 0.00014090270812437312, + "loss": 1.2689, + "theoretical_loss": 3.3740688711687215, + "tokens_seen": 2725445632 + }, + { + "epoch": 0.83, + "learning_rate": 0.00014088665997993982, + "loss": 1.2648, + "theoretical_loss": 3.3740625645590585, + "tokens_seen": 2725511168 + }, + { + "epoch": 0.83, + "learning_rate": 0.00014087061183550652, + "loss": 1.2572, + "theoretical_loss": 3.374056258143498, + "tokens_seen": 2725576704 + }, + { + "epoch": 0.83, + "learning_rate": 0.00014085456369107322, + "loss": 1.2778, + "theoretical_loss": 3.37404995192203, + "tokens_seen": 2725642240 + }, + { + "epoch": 0.83, + "learning_rate": 0.00014083851554663995, + "loss": 1.2966, + "theoretical_loss": 3.3740436458946426, + "tokens_seen": 2725707776 + }, + { + "epoch": 0.83, + "learning_rate": 0.00014082246740220665, + "loss": 1.2759, + "theoretical_loss": 3.374037340061326, + "tokens_seen": 2725773312 + }, + { + "epoch": 0.83, + "learning_rate": 0.00014080641925777335, + "loss": 1.2038, + "theoretical_loss": 3.3740310344220696, + "tokens_seen": 2725838848 + }, + { + "epoch": 0.83, + "learning_rate": 0.00014079037111334005, + "loss": 1.245, + "theoretical_loss": 3.3740247289768623, + "tokens_seen": 2725904384 + }, + { + "epoch": 0.83, + "learning_rate": 0.00014077432296890672, + "loss": 1.2961, + "theoretical_loss": 3.3740184237256936, + "tokens_seen": 2725969920 + }, + { + "epoch": 0.83, + "learning_rate": 0.00014075827482447342, + "loss": 1.2177, + "theoretical_loss": 3.374012118668553, + "tokens_seen": 2726035456 + }, + { + "epoch": 0.83, + "learning_rate": 0.00014074222668004012, + "loss": 1.2903, + "theoretical_loss": 3.37400581380543, + "tokens_seen": 2726100992 + }, + { + "epoch": 0.83, + "learning_rate": 0.00014072617853560682, + "loss": 1.3274, + "theoretical_loss": 3.373999509136314, + "tokens_seen": 2726166528 + }, + { + "epoch": 0.83, + "learning_rate": 0.00014071013039117353, + "loss": 1.2058, + "theoretical_loss": 3.373993204661194, + "tokens_seen": 2726232064 + }, + { + "epoch": 0.83, + "objective/train/docs_used": 990556, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.2092616558074951, + "objective/train/theoretical_loss": 3.373986900380059, + "objective/train/tokens_used": 2746757600, + "theoretical_loss": 3.373986900380059, + "tokens_seen": 2726297600 + }, + { + "epoch": 0.83, + "learning_rate": 0.00014069408224674023, + "loss": 1.2636, + "theoretical_loss": 3.373986900380059, + "tokens_seen": 2726297600 + }, + { + "epoch": 0.83, + "learning_rate": 0.00014067803410230693, + "loss": 1.293, + "theoretical_loss": 3.3739805962928995, + "tokens_seen": 2726363136 + }, + { + "epoch": 0.83, + "learning_rate": 0.00014066198595787363, + "loss": 1.3997, + "theoretical_loss": 3.373974292399704, + "tokens_seen": 2726428672 + }, + { + "epoch": 0.83, + "learning_rate": 0.00014064593781344033, + "loss": 1.2591, + "theoretical_loss": 3.3739679887004623, + "tokens_seen": 2726494208 + }, + { + "epoch": 0.83, + "learning_rate": 0.00014062988966900703, + "loss": 1.2423, + "theoretical_loss": 3.3739616851951633, + "tokens_seen": 2726559744 + }, + { + "epoch": 0.83, + "learning_rate": 0.00014061384152457373, + "loss": 1.2832, + "theoretical_loss": 3.373955381883797, + "tokens_seen": 2726625280 + }, + { + "epoch": 0.83, + "learning_rate": 0.00014059779338014043, + "loss": 1.3324, + "theoretical_loss": 3.373949078766352, + "tokens_seen": 2726690816 + }, + { + "epoch": 0.83, + "learning_rate": 0.00014058174523570713, + "loss": 1.2602, + "theoretical_loss": 3.373942775842819, + "tokens_seen": 2726756352 + }, + { + "epoch": 0.83, + "learning_rate": 0.00014056569709127383, + "loss": 1.2282, + "theoretical_loss": 3.3739364731131856, + "tokens_seen": 2726821888 + }, + { + "epoch": 0.83, + "learning_rate": 0.00014054964894684053, + "loss": 1.2572, + "theoretical_loss": 3.3739301705774425, + "tokens_seen": 2726887424 + }, + { + "epoch": 0.83, + "learning_rate": 0.00014053360080240723, + "loss": 1.2567, + "theoretical_loss": 3.373923868235579, + "tokens_seen": 2726952960 + }, + { + "epoch": 0.83, + "learning_rate": 0.00014051755265797393, + "loss": 1.2163, + "theoretical_loss": 3.3739175660875835, + "tokens_seen": 2727018496 + }, + { + "epoch": 0.83, + "learning_rate": 0.00014050150451354063, + "loss": 1.2937, + "theoretical_loss": 3.3739112641334463, + "tokens_seen": 2727084032 + }, + { + "epoch": 0.83, + "learning_rate": 0.00014048545636910733, + "loss": 1.346, + "theoretical_loss": 3.3739049623731563, + "tokens_seen": 2727149568 + }, + { + "epoch": 0.83, + "learning_rate": 0.00014046940822467403, + "loss": 1.2363, + "theoretical_loss": 3.373898660806703, + "tokens_seen": 2727215104 + }, + { + "epoch": 0.83, + "learning_rate": 0.00014045336008024074, + "loss": 1.2318, + "theoretical_loss": 3.373892359434076, + "tokens_seen": 2727280640 + }, + { + "epoch": 0.83, + "learning_rate": 0.00014043731193580744, + "loss": 1.3143, + "theoretical_loss": 3.3738860582552648, + "tokens_seen": 2727346176 + }, + { + "epoch": 0.83, + "learning_rate": 0.00014042126379137414, + "loss": 1.2435, + "theoretical_loss": 3.373879757270258, + "tokens_seen": 2727411712 + }, + { + "epoch": 0.83, + "learning_rate": 0.00014040521564694084, + "loss": 1.3243, + "theoretical_loss": 3.373873456479046, + "tokens_seen": 2727477248 + }, + { + "epoch": 0.83, + "learning_rate": 0.0001403891675025075, + "loss": 1.3316, + "theoretical_loss": 3.373867155881617, + "tokens_seen": 2727542784 + }, + { + "epoch": 0.83, + "learning_rate": 0.0001403731193580742, + "loss": 1.296, + "theoretical_loss": 3.3738608554779614, + "tokens_seen": 2727608320 + }, + { + "epoch": 0.83, + "learning_rate": 0.0001403570712136409, + "loss": 1.2042, + "theoretical_loss": 3.3738545552680685, + "tokens_seen": 2727673856 + }, + { + "epoch": 0.83, + "learning_rate": 0.0001403410230692076, + "loss": 1.3066, + "theoretical_loss": 3.373848255251927, + "tokens_seen": 2727739392 + }, + { + "epoch": 0.83, + "learning_rate": 0.0001403249749247743, + "loss": 1.2827, + "theoretical_loss": 3.373841955429527, + "tokens_seen": 2727804928 + }, + { + "epoch": 0.83, + "learning_rate": 0.00014030892678034101, + "loss": 1.3339, + "theoretical_loss": 3.3738356558008578, + "tokens_seen": 2727870464 + }, + { + "epoch": 0.83, + "learning_rate": 0.00014029287863590771, + "loss": 1.4269, + "theoretical_loss": 3.373829356365908, + "tokens_seen": 2727936000 + }, + { + "epoch": 0.83, + "learning_rate": 0.00014027683049147442, + "loss": 1.2947, + "theoretical_loss": 3.373823057124668, + "tokens_seen": 2728001536 + }, + { + "epoch": 0.83, + "learning_rate": 0.00014026078234704112, + "loss": 1.3572, + "theoretical_loss": 3.3738167580771266, + "tokens_seen": 2728067072 + }, + { + "epoch": 0.83, + "learning_rate": 0.00014024473420260782, + "loss": 1.3136, + "theoretical_loss": 3.3738104592232734, + "tokens_seen": 2728132608 + }, + { + "epoch": 0.83, + "learning_rate": 0.00014022868605817454, + "loss": 1.2398, + "theoretical_loss": 3.3738041605630977, + "tokens_seen": 2728198144 + }, + { + "epoch": 0.83, + "learning_rate": 0.00014021263791374124, + "loss": 1.3996, + "theoretical_loss": 3.373797862096589, + "tokens_seen": 2728263680 + }, + { + "epoch": 0.83, + "learning_rate": 0.00014019658976930795, + "loss": 1.2245, + "theoretical_loss": 3.3737915638237364, + "tokens_seen": 2728329216 + }, + { + "epoch": 0.83, + "learning_rate": 0.00014018054162487465, + "loss": 1.2908, + "theoretical_loss": 3.3737852657445297, + "tokens_seen": 2728394752 + }, + { + "epoch": 0.83, + "learning_rate": 0.00014016449348044135, + "loss": 1.312, + "theoretical_loss": 3.373778967858958, + "tokens_seen": 2728460288 + }, + { + "epoch": 0.83, + "learning_rate": 0.00014014844533600805, + "loss": 1.3437, + "theoretical_loss": 3.3737726701670114, + "tokens_seen": 2728525824 + }, + { + "epoch": 0.83, + "learning_rate": 0.00014013239719157475, + "loss": 1.2522, + "theoretical_loss": 3.373766372668678, + "tokens_seen": 2728591360 + }, + { + "epoch": 0.83, + "learning_rate": 0.00014011634904714145, + "loss": 1.2066, + "theoretical_loss": 3.3737600753639483, + "tokens_seen": 2728656896 + }, + { + "epoch": 0.83, + "learning_rate": 0.00014010030090270815, + "loss": 1.1844, + "theoretical_loss": 3.3737537782528113, + "tokens_seen": 2728722432 + }, + { + "epoch": 0.83, + "learning_rate": 0.00014008425275827485, + "loss": 1.2469, + "theoretical_loss": 3.3737474813352564, + "tokens_seen": 2728787968 + }, + { + "epoch": 0.83, + "learning_rate": 0.00014006820461384152, + "loss": 1.34, + "theoretical_loss": 3.373741184611273, + "tokens_seen": 2728853504 + }, + { + "epoch": 0.83, + "learning_rate": 0.00014005215646940822, + "loss": 1.3521, + "theoretical_loss": 3.3737348880808504, + "tokens_seen": 2728919040 + }, + { + "epoch": 0.83, + "learning_rate": 0.00014003610832497492, + "loss": 1.2228, + "theoretical_loss": 3.373728591743978, + "tokens_seen": 2728984576 + }, + { + "epoch": 0.83, + "learning_rate": 0.00014002006018054163, + "loss": 1.2416, + "theoretical_loss": 3.3737222956006456, + "tokens_seen": 2729050112 + }, + { + "epoch": 0.83, + "learning_rate": 0.00014000401203610833, + "loss": 1.2826, + "theoretical_loss": 3.373715999650842, + "tokens_seen": 2729115648 + }, + { + "epoch": 0.83, + "learning_rate": 0.00013998796389167503, + "loss": 1.3561, + "theoretical_loss": 3.3737097038945576, + "tokens_seen": 2729181184 + }, + { + "epoch": 0.83, + "learning_rate": 0.00013997191574724173, + "loss": 1.2299, + "theoretical_loss": 3.3737034083317803, + "tokens_seen": 2729246720 + }, + { + "epoch": 0.83, + "learning_rate": 0.00013995586760280843, + "loss": 1.3905, + "theoretical_loss": 3.3736971129625006, + "tokens_seen": 2729312256 + }, + { + "epoch": 0.83, + "learning_rate": 0.00013993981945837513, + "loss": 1.1864, + "theoretical_loss": 3.373690817786708, + "tokens_seen": 2729377792 + }, + { + "epoch": 0.83, + "learning_rate": 0.00013992377131394183, + "loss": 1.342, + "theoretical_loss": 3.3736845228043912, + "tokens_seen": 2729443328 + }, + { + "epoch": 0.83, + "learning_rate": 0.00013990772316950853, + "loss": 1.221, + "theoretical_loss": 3.37367822801554, + "tokens_seen": 2729508864 + }, + { + "epoch": 0.83, + "objective/train/docs_used": 991698, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.3475942611694336, + "objective/train/theoretical_loss": 3.3736719334201437, + "objective/train/tokens_used": 2750034400, + "theoretical_loss": 3.3736719334201437, + "tokens_seen": 2729574400 + }, + { + "epoch": 0.83, + "learning_rate": 0.00013989167502507523, + "loss": 1.2389, + "theoretical_loss": 3.3736719334201437, + "tokens_seen": 2729574400 + }, + { + "epoch": 0.83, + "learning_rate": 0.00013987562688064193, + "loss": 1.2995, + "theoretical_loss": 3.373665639018192, + "tokens_seen": 2729639936 + }, + { + "epoch": 0.83, + "learning_rate": 0.00013985957873620863, + "loss": 1.2222, + "theoretical_loss": 3.373659344809674, + "tokens_seen": 2729705472 + }, + { + "epoch": 0.83, + "learning_rate": 0.00013984353059177533, + "loss": 1.4088, + "theoretical_loss": 3.3736530507945792, + "tokens_seen": 2729771008 + }, + { + "epoch": 0.83, + "learning_rate": 0.00013982748244734203, + "loss": 1.3056, + "theoretical_loss": 3.373646756972897, + "tokens_seen": 2729836544 + }, + { + "epoch": 0.83, + "learning_rate": 0.00013981143430290873, + "loss": 1.4127, + "theoretical_loss": 3.373640463344617, + "tokens_seen": 2729902080 + }, + { + "epoch": 0.83, + "learning_rate": 0.00013979538615847543, + "loss": 1.2148, + "theoretical_loss": 3.373634169909728, + "tokens_seen": 2729967616 + }, + { + "epoch": 0.83, + "learning_rate": 0.00013977933801404214, + "loss": 1.2866, + "theoretical_loss": 3.37362787666822, + "tokens_seen": 2730033152 + }, + { + "epoch": 0.83, + "learning_rate": 0.00013976328986960884, + "loss": 1.2663, + "theoretical_loss": 3.3736215836200825, + "tokens_seen": 2730098688 + }, + { + "epoch": 0.83, + "learning_rate": 0.00013974724172517554, + "loss": 1.364, + "theoretical_loss": 3.3736152907653043, + "tokens_seen": 2730164224 + }, + { + "epoch": 0.83, + "learning_rate": 0.00013973119358074224, + "loss": 1.286, + "theoretical_loss": 3.373608998103876, + "tokens_seen": 2730229760 + }, + { + "epoch": 0.83, + "learning_rate": 0.00013971514543630894, + "loss": 1.2593, + "theoretical_loss": 3.3736027056357853, + "tokens_seen": 2730295296 + }, + { + "epoch": 0.83, + "learning_rate": 0.00013969909729187564, + "loss": 1.2625, + "theoretical_loss": 3.3735964133610232, + "tokens_seen": 2730360832 + }, + { + "epoch": 0.83, + "learning_rate": 0.0001396830491474423, + "loss": 1.2535, + "theoretical_loss": 3.373590121279578, + "tokens_seen": 2730426368 + }, + { + "epoch": 0.83, + "learning_rate": 0.000139667001003009, + "loss": 1.2286, + "theoretical_loss": 3.37358382939144, + "tokens_seen": 2730491904 + }, + { + "epoch": 0.83, + "learning_rate": 0.0001396509528585757, + "loss": 1.4738, + "theoretical_loss": 3.373577537696598, + "tokens_seen": 2730557440 + }, + { + "epoch": 0.83, + "learning_rate": 0.00013963490471414244, + "loss": 1.3201, + "theoretical_loss": 3.3735712461950413, + "tokens_seen": 2730622976 + }, + { + "epoch": 0.83, + "learning_rate": 0.00013961885656970914, + "loss": 1.3415, + "theoretical_loss": 3.37356495488676, + "tokens_seen": 2730688512 + }, + { + "epoch": 0.83, + "learning_rate": 0.00013960280842527584, + "loss": 1.2548, + "theoretical_loss": 3.3735586637717434, + "tokens_seen": 2730754048 + }, + { + "epoch": 0.83, + "learning_rate": 0.00013958676028084254, + "loss": 1.3287, + "theoretical_loss": 3.3735523728499803, + "tokens_seen": 2730819584 + }, + { + "epoch": 0.83, + "learning_rate": 0.00013957071213640924, + "loss": 1.31, + "theoretical_loss": 3.373546082121461, + "tokens_seen": 2730885120 + }, + { + "epoch": 0.83, + "learning_rate": 0.00013955466399197594, + "loss": 1.3067, + "theoretical_loss": 3.373539791586174, + "tokens_seen": 2730950656 + }, + { + "epoch": 0.83, + "learning_rate": 0.00013953861584754264, + "loss": 1.2725, + "theoretical_loss": 3.373533501244109, + "tokens_seen": 2731016192 + }, + { + "epoch": 0.83, + "learning_rate": 0.00013952256770310935, + "loss": 1.2342, + "theoretical_loss": 3.373527211095256, + "tokens_seen": 2731081728 + }, + { + "epoch": 0.83, + "learning_rate": 0.00013950651955867605, + "loss": 1.243, + "theoretical_loss": 3.373520921139604, + "tokens_seen": 2731147264 + }, + { + "epoch": 0.83, + "learning_rate": 0.00013949047141424275, + "loss": 1.1916, + "theoretical_loss": 3.3735146313771427, + "tokens_seen": 2731212800 + }, + { + "epoch": 0.83, + "learning_rate": 0.00013947442326980945, + "loss": 1.3404, + "theoretical_loss": 3.373508341807861, + "tokens_seen": 2731278336 + }, + { + "epoch": 0.83, + "learning_rate": 0.00013945837512537615, + "loss": 1.3559, + "theoretical_loss": 3.3735020524317485, + "tokens_seen": 2731343872 + }, + { + "epoch": 0.83, + "learning_rate": 0.00013944232698094285, + "loss": 1.3613, + "theoretical_loss": 3.373495763248795, + "tokens_seen": 2731409408 + }, + { + "epoch": 0.83, + "learning_rate": 0.00013942627883650955, + "loss": 1.2635, + "theoretical_loss": 3.3734894742589896, + "tokens_seen": 2731474944 + }, + { + "epoch": 0.83, + "learning_rate": 0.00013941023069207625, + "loss": 1.1921, + "theoretical_loss": 3.3734831854623217, + "tokens_seen": 2731540480 + }, + { + "epoch": 0.83, + "learning_rate": 0.00013939418254764295, + "loss": 1.1988, + "theoretical_loss": 3.373476896858781, + "tokens_seen": 2731606016 + }, + { + "epoch": 0.83, + "learning_rate": 0.00013937813440320962, + "loss": 1.29, + "theoretical_loss": 3.3734706084483568, + "tokens_seen": 2731671552 + }, + { + "epoch": 0.83, + "learning_rate": 0.00013936208625877632, + "loss": 1.3172, + "theoretical_loss": 3.3734643202310384, + "tokens_seen": 2731737088 + }, + { + "epoch": 0.83, + "learning_rate": 0.00013934603811434303, + "loss": 1.3015, + "theoretical_loss": 3.3734580322068157, + "tokens_seen": 2731802624 + }, + { + "epoch": 0.83, + "learning_rate": 0.00013932998996990973, + "loss": 1.405, + "theoretical_loss": 3.3734517443756777, + "tokens_seen": 2731868160 + }, + { + "epoch": 0.83, + "learning_rate": 0.00013931394182547643, + "loss": 1.2836, + "theoretical_loss": 3.3734454567376138, + "tokens_seen": 2731933696 + }, + { + "epoch": 0.83, + "learning_rate": 0.00013929789368104313, + "loss": 1.256, + "theoretical_loss": 3.373439169292614, + "tokens_seen": 2731999232 + }, + { + "epoch": 0.83, + "learning_rate": 0.00013928184553660983, + "loss": 1.2926, + "theoretical_loss": 3.373432882040667, + "tokens_seen": 2732064768 + }, + { + "epoch": 0.83, + "learning_rate": 0.00013926579739217653, + "loss": 1.2346, + "theoretical_loss": 3.3734265949817623, + "tokens_seen": 2732130304 + }, + { + "epoch": 0.83, + "learning_rate": 0.00013924974924774323, + "loss": 1.2655, + "theoretical_loss": 3.37342030811589, + "tokens_seen": 2732195840 + }, + { + "epoch": 0.83, + "learning_rate": 0.00013923370110330993, + "loss": 1.3021, + "theoretical_loss": 3.3734140214430393, + "tokens_seen": 2732261376 + }, + { + "epoch": 0.83, + "learning_rate": 0.00013921765295887663, + "loss": 1.3419, + "theoretical_loss": 3.373407734963199, + "tokens_seen": 2732326912 + }, + { + "epoch": 0.83, + "learning_rate": 0.00013920160481444333, + "loss": 1.3655, + "theoretical_loss": 3.3734014486763595, + "tokens_seen": 2732392448 + }, + { + "epoch": 0.83, + "learning_rate": 0.00013918555667001003, + "loss": 1.3212, + "theoretical_loss": 3.3733951625825096, + "tokens_seen": 2732457984 + }, + { + "epoch": 0.83, + "learning_rate": 0.00013916950852557673, + "loss": 1.3427, + "theoretical_loss": 3.373388876681639, + "tokens_seen": 2732523520 + }, + { + "epoch": 0.83, + "learning_rate": 0.00013915346038114343, + "loss": 1.256, + "theoretical_loss": 3.373382590973737, + "tokens_seen": 2732589056 + }, + { + "epoch": 0.83, + "learning_rate": 0.00013913741223671013, + "loss": 1.3348, + "theoretical_loss": 3.3733763054587933, + "tokens_seen": 2732654592 + }, + { + "epoch": 0.83, + "learning_rate": 0.00013912136409227683, + "loss": 1.3249, + "theoretical_loss": 3.373370020136797, + "tokens_seen": 2732720128 + }, + { + "epoch": 0.83, + "learning_rate": 0.00013910531594784353, + "loss": 1.2516, + "theoretical_loss": 3.373363735007738, + "tokens_seen": 2732785664 + }, + { + "epoch": 0.83, + "objective/train/docs_used": 992760, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.2704886198043823, + "objective/train/theoretical_loss": 3.3733574500716053, + "objective/train/tokens_used": 2753311200, + "theoretical_loss": 3.3733574500716053, + "tokens_seen": 2732851200 + }, + { + "epoch": 0.83, + "learning_rate": 0.00013908926780341024, + "loss": 1.4378, + "theoretical_loss": 3.3733574500716053, + "tokens_seen": 2732851200 + }, + { + "epoch": 0.83, + "learning_rate": 0.00013907321965897694, + "loss": 1.3001, + "theoretical_loss": 3.3733511653283887, + "tokens_seen": 2732916736 + }, + { + "epoch": 0.83, + "learning_rate": 0.00013905717151454364, + "loss": 1.2968, + "theoretical_loss": 3.3733448807780775, + "tokens_seen": 2732982272 + }, + { + "epoch": 0.83, + "learning_rate": 0.00013904112337011034, + "loss": 1.2773, + "theoretical_loss": 3.373338596420661, + "tokens_seen": 2733047808 + }, + { + "epoch": 0.83, + "learning_rate": 0.00013902507522567704, + "loss": 1.2646, + "theoretical_loss": 3.373332312256129, + "tokens_seen": 2733113344 + }, + { + "epoch": 0.83, + "learning_rate": 0.00013900902708124374, + "loss": 1.3468, + "theoretical_loss": 3.3733260282844704, + "tokens_seen": 2733178880 + }, + { + "epoch": 0.83, + "learning_rate": 0.00013899297893681044, + "loss": 1.1946, + "theoretical_loss": 3.3733197445056753, + "tokens_seen": 2733244416 + }, + { + "epoch": 0.83, + "learning_rate": 0.00013897693079237714, + "loss": 1.1848, + "theoretical_loss": 3.3733134609197326, + "tokens_seen": 2733309952 + }, + { + "epoch": 0.83, + "learning_rate": 0.00013896088264794384, + "loss": 1.1925, + "theoretical_loss": 3.373307177526632, + "tokens_seen": 2733375488 + }, + { + "epoch": 0.83, + "learning_rate": 0.00013894483450351054, + "loss": 1.2489, + "theoretical_loss": 3.3733008943263636, + "tokens_seen": 2733441024 + }, + { + "epoch": 0.83, + "learning_rate": 0.00013892878635907724, + "loss": 1.2916, + "theoretical_loss": 3.3732946113189155, + "tokens_seen": 2733506560 + }, + { + "epoch": 0.83, + "learning_rate": 0.00013891273821464394, + "loss": 1.3579, + "theoretical_loss": 3.3732883285042785, + "tokens_seen": 2733572096 + }, + { + "epoch": 0.83, + "learning_rate": 0.00013889669007021064, + "loss": 1.3849, + "theoretical_loss": 3.373282045882441, + "tokens_seen": 2733637632 + }, + { + "epoch": 0.83, + "learning_rate": 0.00013888064192577734, + "loss": 1.2863, + "theoretical_loss": 3.3732757634533934, + "tokens_seen": 2733703168 + }, + { + "epoch": 0.83, + "learning_rate": 0.00013886459378134404, + "loss": 1.2257, + "theoretical_loss": 3.373269481217124, + "tokens_seen": 2733768704 + }, + { + "epoch": 0.83, + "learning_rate": 0.00013884854563691074, + "loss": 1.2796, + "theoretical_loss": 3.3732631991736235, + "tokens_seen": 2733834240 + }, + { + "epoch": 0.83, + "learning_rate": 0.00013883249749247745, + "loss": 1.2685, + "theoretical_loss": 3.373256917322881, + "tokens_seen": 2733899776 + }, + { + "epoch": 0.83, + "learning_rate": 0.00013881644934804415, + "loss": 1.3587, + "theoretical_loss": 3.3732506356648853, + "tokens_seen": 2733965312 + }, + { + "epoch": 0.83, + "learning_rate": 0.00013880040120361085, + "loss": 1.2633, + "theoretical_loss": 3.3732443541996266, + "tokens_seen": 2734030848 + }, + { + "epoch": 0.83, + "learning_rate": 0.00013878435305917755, + "loss": 1.2936, + "theoretical_loss": 3.373238072927094, + "tokens_seen": 2734096384 + }, + { + "epoch": 0.83, + "learning_rate": 0.00013876830491474425, + "loss": 1.253, + "theoretical_loss": 3.3732317918472767, + "tokens_seen": 2734161920 + }, + { + "epoch": 0.83, + "learning_rate": 0.00013875225677031095, + "loss": 1.3435, + "theoretical_loss": 3.3732255109601654, + "tokens_seen": 2734227456 + }, + { + "epoch": 0.83, + "learning_rate": 0.00013873620862587765, + "loss": 1.2897, + "theoretical_loss": 3.373219230265748, + "tokens_seen": 2734292992 + }, + { + "epoch": 0.83, + "learning_rate": 0.00013872016048144435, + "loss": 1.1441, + "theoretical_loss": 3.373212949764015, + "tokens_seen": 2734358528 + }, + { + "epoch": 0.83, + "learning_rate": 0.00013870411233701105, + "loss": 1.2364, + "theoretical_loss": 3.3732066694549556, + "tokens_seen": 2734424064 + }, + { + "epoch": 0.83, + "learning_rate": 0.00013868806419257775, + "loss": 1.3021, + "theoretical_loss": 3.373200389338559, + "tokens_seen": 2734489600 + }, + { + "epoch": 0.83, + "learning_rate": 0.00013867201604814442, + "loss": 1.3799, + "theoretical_loss": 3.3731941094148152, + "tokens_seen": 2734555136 + }, + { + "epoch": 0.83, + "learning_rate": 0.00013865596790371113, + "loss": 1.258, + "theoretical_loss": 3.373187829683713, + "tokens_seen": 2734620672 + }, + { + "epoch": 0.83, + "learning_rate": 0.00013863991975927783, + "loss": 1.2421, + "theoretical_loss": 3.373181550145243, + "tokens_seen": 2734686208 + }, + { + "epoch": 0.83, + "learning_rate": 0.00013862387161484453, + "loss": 1.2162, + "theoretical_loss": 3.373175270799393, + "tokens_seen": 2734751744 + }, + { + "epoch": 0.83, + "learning_rate": 0.00013860782347041123, + "loss": 1.394, + "theoretical_loss": 3.373168991646154, + "tokens_seen": 2734817280 + }, + { + "epoch": 0.83, + "learning_rate": 0.00013859177532597793, + "loss": 1.2826, + "theoretical_loss": 3.3731627126855144, + "tokens_seen": 2734882816 + }, + { + "epoch": 0.83, + "learning_rate": 0.00013857572718154463, + "loss": 1.2224, + "theoretical_loss": 3.3731564339174644, + "tokens_seen": 2734948352 + }, + { + "epoch": 0.83, + "learning_rate": 0.00013855967903711133, + "loss": 1.3175, + "theoretical_loss": 3.3731501553419934, + "tokens_seen": 2735013888 + }, + { + "epoch": 0.83, + "learning_rate": 0.00013854363089267803, + "loss": 1.2754, + "theoretical_loss": 3.3731438769590905, + "tokens_seen": 2735079424 + }, + { + "epoch": 0.83, + "learning_rate": 0.00013852758274824473, + "loss": 1.2376, + "theoretical_loss": 3.3731375987687455, + "tokens_seen": 2735144960 + }, + { + "epoch": 0.83, + "learning_rate": 0.00013851153460381143, + "loss": 1.0933, + "theoretical_loss": 3.373131320770948, + "tokens_seen": 2735210496 + }, + { + "epoch": 0.83, + "learning_rate": 0.00013849548645937813, + "loss": 1.3283, + "theoretical_loss": 3.373125042965687, + "tokens_seen": 2735276032 + }, + { + "epoch": 0.83, + "learning_rate": 0.00013847943831494483, + "loss": 1.339, + "theoretical_loss": 3.373118765352952, + "tokens_seen": 2735341568 + }, + { + "epoch": 0.83, + "learning_rate": 0.00013846339017051153, + "loss": 1.2556, + "theoretical_loss": 3.373112487932733, + "tokens_seen": 2735407104 + }, + { + "epoch": 0.83, + "learning_rate": 0.00013844734202607823, + "loss": 1.2549, + "theoretical_loss": 3.373106210705019, + "tokens_seen": 2735472640 + }, + { + "epoch": 0.83, + "learning_rate": 0.00013843129388164496, + "loss": 1.1301, + "theoretical_loss": 3.3730999336698, + "tokens_seen": 2735538176 + }, + { + "epoch": 0.83, + "learning_rate": 0.00013841524573721166, + "loss": 1.3577, + "theoretical_loss": 3.373093656827065, + "tokens_seen": 2735603712 + }, + { + "epoch": 0.83, + "learning_rate": 0.00013839919759277836, + "loss": 1.294, + "theoretical_loss": 3.3730873801768038, + "tokens_seen": 2735669248 + }, + { + "epoch": 0.83, + "learning_rate": 0.00013838314944834506, + "loss": 1.2261, + "theoretical_loss": 3.3730811037190054, + "tokens_seen": 2735734784 + }, + { + "epoch": 0.83, + "learning_rate": 0.00013836710130391176, + "loss": 1.3197, + "theoretical_loss": 3.37307482745366, + "tokens_seen": 2735800320 + }, + { + "epoch": 0.83, + "learning_rate": 0.00013835105315947844, + "loss": 1.3224, + "theoretical_loss": 3.3730685513807566, + "tokens_seen": 2735865856 + }, + { + "epoch": 0.83, + "learning_rate": 0.00013833500501504514, + "loss": 1.374, + "theoretical_loss": 3.373062275500285, + "tokens_seen": 2735931392 + }, + { + "epoch": 0.83, + "learning_rate": 0.00013831895687061184, + "loss": 1.1398, + "theoretical_loss": 3.3730559998122343, + "tokens_seen": 2735996928 + }, + { + "epoch": 0.83, + "learning_rate": 0.00013830290872617854, + "loss": 1.2539, + "theoretical_loss": 3.3730497243165942, + "tokens_seen": 2736062464 + }, + { + "epoch": 0.83, + "objective/train/docs_used": 993861, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.322221040725708, + "objective/train/theoretical_loss": 3.373043449013354, + "objective/train/tokens_used": 2756588000, + "theoretical_loss": 3.373043449013354, + "tokens_seen": 2736128000 + }, + { + "epoch": 0.83, + "learning_rate": 0.00013828686058174524, + "loss": 1.2668, + "theoretical_loss": 3.373043449013354, + "tokens_seen": 2736128000 + }, + { + "epoch": 0.83, + "learning_rate": 0.00013827081243731194, + "loss": 1.2498, + "theoretical_loss": 3.3730371739025036, + "tokens_seen": 2736193536 + }, + { + "epoch": 0.83, + "learning_rate": 0.00013825476429287864, + "loss": 1.307, + "theoretical_loss": 3.373030898984032, + "tokens_seen": 2736259072 + }, + { + "epoch": 0.83, + "learning_rate": 0.00013823871614844534, + "loss": 1.2741, + "theoretical_loss": 3.3730246242579294, + "tokens_seen": 2736324608 + }, + { + "epoch": 0.83, + "learning_rate": 0.00013822266800401204, + "loss": 1.2482, + "theoretical_loss": 3.373018349724185, + "tokens_seen": 2736390144 + }, + { + "epoch": 0.83, + "learning_rate": 0.00013820661985957874, + "loss": 1.1641, + "theoretical_loss": 3.3730120753827877, + "tokens_seen": 2736455680 + }, + { + "epoch": 0.83, + "learning_rate": 0.00013819057171514544, + "loss": 1.2031, + "theoretical_loss": 3.3730058012337274, + "tokens_seen": 2736521216 + }, + { + "epoch": 0.83, + "learning_rate": 0.00013817452357071214, + "loss": 1.2101, + "theoretical_loss": 3.3729995272769937, + "tokens_seen": 2736586752 + }, + { + "epoch": 0.83, + "learning_rate": 0.00013815847542627884, + "loss": 1.3462, + "theoretical_loss": 3.3729932535125764, + "tokens_seen": 2736652288 + }, + { + "epoch": 0.83, + "learning_rate": 0.00013814242728184555, + "loss": 1.2116, + "theoretical_loss": 3.3729869799404644, + "tokens_seen": 2736717824 + }, + { + "epoch": 0.83, + "learning_rate": 0.00013812637913741225, + "loss": 1.3176, + "theoretical_loss": 3.372980706560648, + "tokens_seen": 2736783360 + }, + { + "epoch": 0.83, + "learning_rate": 0.00013811033099297895, + "loss": 1.1996, + "theoretical_loss": 3.3729744333731153, + "tokens_seen": 2736848896 + }, + { + "epoch": 0.83, + "learning_rate": 0.00013809428284854565, + "loss": 1.2602, + "theoretical_loss": 3.372968160377857, + "tokens_seen": 2736914432 + }, + { + "epoch": 0.83, + "learning_rate": 0.00013807823470411235, + "loss": 1.2886, + "theoretical_loss": 3.3729618875748626, + "tokens_seen": 2736979968 + }, + { + "epoch": 0.83, + "learning_rate": 0.00013806218655967905, + "loss": 1.2049, + "theoretical_loss": 3.3729556149641207, + "tokens_seen": 2737045504 + }, + { + "epoch": 0.83, + "learning_rate": 0.00013804613841524575, + "loss": 1.3034, + "theoretical_loss": 3.372949342545622, + "tokens_seen": 2737111040 + }, + { + "epoch": 0.83, + "learning_rate": 0.00013803009027081245, + "loss": 1.236, + "theoretical_loss": 3.372943070319355, + "tokens_seen": 2737176576 + }, + { + "epoch": 0.83, + "learning_rate": 0.00013801404212637915, + "loss": 1.2623, + "theoretical_loss": 3.3729367982853096, + "tokens_seen": 2737242112 + }, + { + "epoch": 0.83, + "learning_rate": 0.00013799799398194585, + "loss": 1.2235, + "theoretical_loss": 3.3729305264434752, + "tokens_seen": 2737307648 + }, + { + "epoch": 0.83, + "learning_rate": 0.00013798194583751255, + "loss": 1.2469, + "theoretical_loss": 3.3729242547938414, + "tokens_seen": 2737373184 + }, + { + "epoch": 0.83, + "learning_rate": 0.00013796589769307923, + "loss": 1.1663, + "theoretical_loss": 3.372917983336398, + "tokens_seen": 2737438720 + }, + { + "epoch": 0.83, + "learning_rate": 0.00013794984954864593, + "loss": 1.3949, + "theoretical_loss": 3.372911712071134, + "tokens_seen": 2737504256 + }, + { + "epoch": 0.83, + "learning_rate": 0.00013793380140421263, + "loss": 1.2376, + "theoretical_loss": 3.372905440998039, + "tokens_seen": 2737569792 + }, + { + "epoch": 0.83, + "learning_rate": 0.00013791775325977933, + "loss": 1.3585, + "theoretical_loss": 3.3728991701171025, + "tokens_seen": 2737635328 + }, + { + "epoch": 0.83, + "learning_rate": 0.00013790170511534603, + "loss": 1.3255, + "theoretical_loss": 3.3728928994283147, + "tokens_seen": 2737700864 + }, + { + "epoch": 0.83, + "learning_rate": 0.00013788565697091273, + "loss": 1.1976, + "theoretical_loss": 3.3728866289316644, + "tokens_seen": 2737766400 + }, + { + "epoch": 0.83, + "learning_rate": 0.00013786960882647943, + "loss": 1.454, + "theoretical_loss": 3.3728803586271408, + "tokens_seen": 2737831936 + }, + { + "epoch": 0.83, + "learning_rate": 0.00013785356068204613, + "loss": 1.2864, + "theoretical_loss": 3.3728740885147346, + "tokens_seen": 2737897472 + }, + { + "epoch": 0.83, + "learning_rate": 0.00013783751253761286, + "loss": 1.3346, + "theoretical_loss": 3.3728678185944343, + "tokens_seen": 2737963008 + }, + { + "epoch": 0.83, + "learning_rate": 0.00013782146439317956, + "loss": 1.3025, + "theoretical_loss": 3.3728615488662292, + "tokens_seen": 2738028544 + }, + { + "epoch": 0.83, + "learning_rate": 0.00013780541624874626, + "loss": 1.2521, + "theoretical_loss": 3.37285527933011, + "tokens_seen": 2738094080 + }, + { + "epoch": 0.83, + "learning_rate": 0.00013778936810431296, + "loss": 1.3147, + "theoretical_loss": 3.3728490099860653, + "tokens_seen": 2738159616 + }, + { + "epoch": 0.83, + "learning_rate": 0.00013777331995987966, + "loss": 1.251, + "theoretical_loss": 3.372842740834085, + "tokens_seen": 2738225152 + }, + { + "epoch": 0.83, + "learning_rate": 0.00013775727181544636, + "loss": 1.2333, + "theoretical_loss": 3.3728364718741584, + "tokens_seen": 2738290688 + }, + { + "epoch": 0.83, + "learning_rate": 0.00013774122367101306, + "loss": 1.2994, + "theoretical_loss": 3.3728302031062753, + "tokens_seen": 2738356224 + }, + { + "epoch": 0.83, + "learning_rate": 0.00013772517552657976, + "loss": 1.2202, + "theoretical_loss": 3.372823934530425, + "tokens_seen": 2738421760 + }, + { + "epoch": 0.83, + "learning_rate": 0.00013770912738214646, + "loss": 1.1551, + "theoretical_loss": 3.372817666146597, + "tokens_seen": 2738487296 + }, + { + "epoch": 0.83, + "learning_rate": 0.00013769307923771316, + "loss": 1.3169, + "theoretical_loss": 3.3728113979547807, + "tokens_seen": 2738552832 + }, + { + "epoch": 0.83, + "learning_rate": 0.00013767703109327986, + "loss": 1.2055, + "theoretical_loss": 3.3728051299549655, + "tokens_seen": 2738618368 + }, + { + "epoch": 0.83, + "learning_rate": 0.00013766098294884654, + "loss": 1.2932, + "theoretical_loss": 3.3727988621471416, + "tokens_seen": 2738683904 + }, + { + "epoch": 0.83, + "learning_rate": 0.00013764493480441324, + "loss": 1.2296, + "theoretical_loss": 3.3727925945312984, + "tokens_seen": 2738749440 + }, + { + "epoch": 0.83, + "learning_rate": 0.00013762888665997994, + "loss": 1.2998, + "theoretical_loss": 3.372786327107425, + "tokens_seen": 2738814976 + }, + { + "epoch": 0.83, + "learning_rate": 0.00013761283851554664, + "loss": 1.2044, + "theoretical_loss": 3.3727800598755113, + "tokens_seen": 2738880512 + }, + { + "epoch": 0.83, + "learning_rate": 0.00013759679037111334, + "loss": 1.2442, + "theoretical_loss": 3.372773792835546, + "tokens_seen": 2738946048 + }, + { + "epoch": 0.83, + "learning_rate": 0.00013758074222668004, + "loss": 1.3456, + "theoretical_loss": 3.37276752598752, + "tokens_seen": 2739011584 + }, + { + "epoch": 0.83, + "learning_rate": 0.00013756469408224674, + "loss": 1.3325, + "theoretical_loss": 3.372761259331422, + "tokens_seen": 2739077120 + }, + { + "epoch": 0.83, + "learning_rate": 0.00013754864593781344, + "loss": 1.2699, + "theoretical_loss": 3.3727549928672413, + "tokens_seen": 2739142656 + }, + { + "epoch": 0.83, + "learning_rate": 0.00013753259779338014, + "loss": 1.2783, + "theoretical_loss": 3.372748726594968, + "tokens_seen": 2739208192 + }, + { + "epoch": 0.83, + "learning_rate": 0.00013751654964894684, + "loss": 1.2173, + "theoretical_loss": 3.372742460514591, + "tokens_seen": 2739273728 + }, + { + "epoch": 0.83, + "learning_rate": 0.00013750050150451354, + "loss": 1.2622, + "theoretical_loss": 3.3727361946261003, + "tokens_seen": 2739339264 + }, + { + "epoch": 0.83, + "objective/train/docs_used": 995217, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.3427839279174805, + "objective/train/theoretical_loss": 3.372729928929486, + "objective/train/tokens_used": 2759864800, + "theoretical_loss": 3.372729928929486, + "tokens_seen": 2739404800 + }, + { + "epoch": 0.83, + "learning_rate": 0.00013748445336008024, + "loss": 1.2928, + "theoretical_loss": 3.372729928929486, + "tokens_seen": 2739404800 + }, + { + "epoch": 0.83, + "learning_rate": 0.00013746840521564695, + "loss": 1.3476, + "theoretical_loss": 3.372723663424736, + "tokens_seen": 2739470336 + }, + { + "epoch": 0.83, + "learning_rate": 0.00013745235707121365, + "loss": 1.281, + "theoretical_loss": 3.3727173981118415, + "tokens_seen": 2739535872 + }, + { + "epoch": 0.83, + "learning_rate": 0.00013743630892678035, + "loss": 1.3873, + "theoretical_loss": 3.372711132990791, + "tokens_seen": 2739601408 + }, + { + "epoch": 0.83, + "learning_rate": 0.00013742026078234705, + "loss": 1.3091, + "theoretical_loss": 3.3727048680615743, + "tokens_seen": 2739666944 + }, + { + "epoch": 0.83, + "learning_rate": 0.00013740421263791375, + "loss": 1.2495, + "theoretical_loss": 3.3726986033241815, + "tokens_seen": 2739732480 + }, + { + "epoch": 0.83, + "learning_rate": 0.00013738816449348045, + "loss": 1.1341, + "theoretical_loss": 3.3726923387786014, + "tokens_seen": 2739798016 + }, + { + "epoch": 0.83, + "learning_rate": 0.00013737211634904715, + "loss": 1.1808, + "theoretical_loss": 3.372686074424824, + "tokens_seen": 2739863552 + }, + { + "epoch": 0.83, + "learning_rate": 0.00013735606820461385, + "loss": 1.3175, + "theoretical_loss": 3.372679810262839, + "tokens_seen": 2739929088 + }, + { + "epoch": 0.83, + "learning_rate": 0.00013734002006018055, + "loss": 1.2473, + "theoretical_loss": 3.3726735462926345, + "tokens_seen": 2739994624 + }, + { + "epoch": 0.83, + "learning_rate": 0.00013732397191574725, + "loss": 1.2938, + "theoretical_loss": 3.3726672825142017, + "tokens_seen": 2740060160 + }, + { + "epoch": 0.83, + "learning_rate": 0.00013730792377131395, + "loss": 1.1963, + "theoretical_loss": 3.3726610189275297, + "tokens_seen": 2740125696 + }, + { + "epoch": 0.83, + "learning_rate": 0.00013729187562688065, + "loss": 1.1833, + "theoretical_loss": 3.372654755532608, + "tokens_seen": 2740191232 + }, + { + "epoch": 0.83, + "learning_rate": 0.00013727582748244733, + "loss": 1.2453, + "theoretical_loss": 3.3726484923294255, + "tokens_seen": 2740256768 + }, + { + "epoch": 0.83, + "learning_rate": 0.00013725977933801403, + "loss": 1.2275, + "theoretical_loss": 3.3726422293179725, + "tokens_seen": 2740322304 + }, + { + "epoch": 0.83, + "learning_rate": 0.00013724373119358073, + "loss": 1.3639, + "theoretical_loss": 3.3726359664982386, + "tokens_seen": 2740387840 + }, + { + "epoch": 0.83, + "learning_rate": 0.00013722768304914745, + "loss": 1.2188, + "theoretical_loss": 3.3726297038702135, + "tokens_seen": 2740453376 + }, + { + "epoch": 0.83, + "learning_rate": 0.00013721163490471416, + "loss": 1.2332, + "theoretical_loss": 3.372623441433886, + "tokens_seen": 2740518912 + }, + { + "epoch": 0.83, + "learning_rate": 0.00013719558676028086, + "loss": 1.3933, + "theoretical_loss": 3.3726171791892456, + "tokens_seen": 2740584448 + }, + { + "epoch": 0.83, + "learning_rate": 0.00013717953861584756, + "loss": 1.2314, + "theoretical_loss": 3.3726109171362824, + "tokens_seen": 2740649984 + }, + { + "epoch": 0.83, + "learning_rate": 0.00013716349047141426, + "loss": 1.3304, + "theoretical_loss": 3.3726046552749858, + "tokens_seen": 2740715520 + }, + { + "epoch": 0.83, + "learning_rate": 0.00013714744232698096, + "loss": 1.2201, + "theoretical_loss": 3.3725983936053456, + "tokens_seen": 2740781056 + }, + { + "epoch": 0.83, + "learning_rate": 0.00013713139418254766, + "loss": 1.2508, + "theoretical_loss": 3.372592132127351, + "tokens_seen": 2740846592 + }, + { + "epoch": 0.83, + "learning_rate": 0.00013711534603811436, + "loss": 1.2172, + "theoretical_loss": 3.3725858708409913, + "tokens_seen": 2740912128 + }, + { + "epoch": 0.83, + "learning_rate": 0.00013709929789368106, + "loss": 1.3837, + "theoretical_loss": 3.3725796097462566, + "tokens_seen": 2740977664 + }, + { + "epoch": 0.83, + "learning_rate": 0.00013708324974924776, + "loss": 1.2678, + "theoretical_loss": 3.3725733488431366, + "tokens_seen": 2741043200 + }, + { + "epoch": 0.83, + "learning_rate": 0.00013706720160481446, + "loss": 1.2121, + "theoretical_loss": 3.37256708813162, + "tokens_seen": 2741108736 + }, + { + "epoch": 0.83, + "learning_rate": 0.00013705115346038116, + "loss": 1.2854, + "theoretical_loss": 3.372560827611697, + "tokens_seen": 2741174272 + }, + { + "epoch": 0.83, + "learning_rate": 0.00013703510531594786, + "loss": 1.2212, + "theoretical_loss": 3.3725545672833572, + "tokens_seen": 2741239808 + }, + { + "epoch": 0.83, + "learning_rate": 0.00013701905717151456, + "loss": 1.2874, + "theoretical_loss": 3.37254830714659, + "tokens_seen": 2741305344 + }, + { + "epoch": 0.83, + "learning_rate": 0.00013700300902708126, + "loss": 1.4009, + "theoretical_loss": 3.3725420472013843, + "tokens_seen": 2741370880 + }, + { + "epoch": 0.83, + "learning_rate": 0.00013698696088264796, + "loss": 1.2422, + "theoretical_loss": 3.372535787447731, + "tokens_seen": 2741436416 + }, + { + "epoch": 0.83, + "learning_rate": 0.00013697091273821466, + "loss": 1.2133, + "theoretical_loss": 3.372529527885619, + "tokens_seen": 2741501952 + }, + { + "epoch": 0.83, + "learning_rate": 0.00013695486459378134, + "loss": 1.2952, + "theoretical_loss": 3.372523268515037, + "tokens_seen": 2741567488 + }, + { + "epoch": 0.83, + "learning_rate": 0.00013693881644934804, + "loss": 1.2327, + "theoretical_loss": 3.372517009335976, + "tokens_seen": 2741633024 + }, + { + "epoch": 0.83, + "learning_rate": 0.00013692276830491474, + "loss": 1.256, + "theoretical_loss": 3.3725107503484253, + "tokens_seen": 2741698560 + }, + { + "epoch": 0.83, + "learning_rate": 0.00013690672016048144, + "loss": 1.3559, + "theoretical_loss": 3.372504491552373, + "tokens_seen": 2741764096 + }, + { + "epoch": 0.83, + "learning_rate": 0.00013689067201604814, + "loss": 1.2503, + "theoretical_loss": 3.372498232947811, + "tokens_seen": 2741829632 + }, + { + "epoch": 0.83, + "learning_rate": 0.00013687462387161484, + "loss": 1.2775, + "theoretical_loss": 3.372491974534727, + "tokens_seen": 2741895168 + }, + { + "epoch": 0.83, + "learning_rate": 0.00013685857572718154, + "loss": 1.2687, + "theoretical_loss": 3.3724857163131112, + "tokens_seen": 2741960704 + }, + { + "epoch": 0.83, + "learning_rate": 0.00013684252758274824, + "loss": 1.2877, + "theoretical_loss": 3.372479458282953, + "tokens_seen": 2742026240 + }, + { + "epoch": 0.83, + "learning_rate": 0.00013682647943831494, + "loss": 1.2305, + "theoretical_loss": 3.3724732004442424, + "tokens_seen": 2742091776 + }, + { + "epoch": 0.83, + "learning_rate": 0.00013681043129388164, + "loss": 1.3118, + "theoretical_loss": 3.3724669427969687, + "tokens_seen": 2742157312 + }, + { + "epoch": 0.83, + "learning_rate": 0.00013679438314944834, + "loss": 1.3255, + "theoretical_loss": 3.3724606853411214, + "tokens_seen": 2742222848 + }, + { + "epoch": 0.83, + "learning_rate": 0.00013677833500501505, + "loss": 1.1827, + "theoretical_loss": 3.37245442807669, + "tokens_seen": 2742288384 + }, + { + "epoch": 0.83, + "learning_rate": 0.00013676228686058175, + "loss": 1.2043, + "theoretical_loss": 3.3724481710036645, + "tokens_seen": 2742353920 + }, + { + "epoch": 0.83, + "learning_rate": 0.00013674623871614845, + "loss": 1.1253, + "theoretical_loss": 3.372441914122034, + "tokens_seen": 2742419456 + }, + { + "epoch": 0.83, + "learning_rate": 0.00013673019057171515, + "loss": 1.263, + "theoretical_loss": 3.372435657431788, + "tokens_seen": 2742484992 + }, + { + "epoch": 0.83, + "learning_rate": 0.00013671414242728185, + "loss": 1.2583, + "theoretical_loss": 3.3724294009329165, + "tokens_seen": 2742550528 + }, + { + "epoch": 0.83, + "learning_rate": 0.00013669809428284855, + "loss": 1.2659, + "theoretical_loss": 3.372423144625409, + "tokens_seen": 2742616064 + }, + { + "epoch": 0.83, + "objective/train/docs_used": 996586, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.2646888494491577, + "objective/train/theoretical_loss": 3.372416888509255, + "objective/train/tokens_used": 2763141600, + "theoretical_loss": 3.372416888509255, + "tokens_seen": 2742681600 + }, + { + "epoch": 0.83, + "learning_rate": 0.00013668204613841525, + "loss": 1.2206, + "theoretical_loss": 3.372416888509255, + "tokens_seen": 2742681600 + }, + { + "epoch": 0.83, + "learning_rate": 0.00013666599799398195, + "loss": 1.3085, + "theoretical_loss": 3.372410632584444, + "tokens_seen": 2742747136 + }, + { + "epoch": 0.83, + "learning_rate": 0.00013664994984954865, + "loss": 1.3077, + "theoretical_loss": 3.3724043768509655, + "tokens_seen": 2742812672 + }, + { + "epoch": 0.83, + "learning_rate": 0.00013663390170511535, + "loss": 1.2897, + "theoretical_loss": 3.3723981213088092, + "tokens_seen": 2742878208 + }, + { + "epoch": 0.83, + "learning_rate": 0.00013661785356068205, + "loss": 1.2696, + "theoretical_loss": 3.372391865957965, + "tokens_seen": 2742943744 + }, + { + "epoch": 0.83, + "learning_rate": 0.00013660180541624875, + "loss": 1.315, + "theoretical_loss": 3.3723856107984216, + "tokens_seen": 2743009280 + }, + { + "epoch": 0.83, + "learning_rate": 0.00013658575727181545, + "loss": 1.2724, + "theoretical_loss": 3.3723793558301693, + "tokens_seen": 2743074816 + }, + { + "epoch": 0.83, + "learning_rate": 0.00013656970912738215, + "loss": 1.2464, + "theoretical_loss": 3.3723731010531974, + "tokens_seen": 2743140352 + }, + { + "epoch": 0.83, + "learning_rate": 0.00013655366098294885, + "loss": 1.3315, + "theoretical_loss": 3.372366846467496, + "tokens_seen": 2743205888 + }, + { + "epoch": 0.83, + "learning_rate": 0.00013653761283851555, + "loss": 1.2265, + "theoretical_loss": 3.372360592073054, + "tokens_seen": 2743271424 + }, + { + "epoch": 0.83, + "learning_rate": 0.00013652156469408226, + "loss": 1.3671, + "theoretical_loss": 3.3723543378698615, + "tokens_seen": 2743336960 + }, + { + "epoch": 0.83, + "learning_rate": 0.00013650551654964896, + "loss": 1.28, + "theoretical_loss": 3.3723480838579074, + "tokens_seen": 2743402496 + }, + { + "epoch": 0.83, + "learning_rate": 0.00013648946840521566, + "loss": 1.3183, + "theoretical_loss": 3.372341830037182, + "tokens_seen": 2743468032 + }, + { + "epoch": 0.83, + "learning_rate": 0.00013647342026078236, + "loss": 1.2617, + "theoretical_loss": 3.3723355764076746, + "tokens_seen": 2743533568 + }, + { + "epoch": 0.83, + "learning_rate": 0.00013645737211634906, + "loss": 1.1766, + "theoretical_loss": 3.372329322969375, + "tokens_seen": 2743599104 + }, + { + "epoch": 0.83, + "learning_rate": 0.00013644132397191576, + "loss": 1.2473, + "theoretical_loss": 3.372323069722272, + "tokens_seen": 2743664640 + }, + { + "epoch": 0.83, + "learning_rate": 0.00013642527582748246, + "loss": 1.1869, + "theoretical_loss": 3.3723168166663564, + "tokens_seen": 2743730176 + }, + { + "epoch": 0.83, + "learning_rate": 0.00013640922768304916, + "loss": 1.276, + "theoretical_loss": 3.3723105638016166, + "tokens_seen": 2743795712 + }, + { + "epoch": 0.83, + "learning_rate": 0.00013639317953861586, + "loss": 1.2189, + "theoretical_loss": 3.372304311128043, + "tokens_seen": 2743861248 + }, + { + "epoch": 0.83, + "learning_rate": 0.00013637713139418256, + "loss": 1.2136, + "theoretical_loss": 3.372298058645625, + "tokens_seen": 2743926784 + }, + { + "epoch": 0.83, + "learning_rate": 0.00013636108324974926, + "loss": 1.2709, + "theoretical_loss": 3.372291806354352, + "tokens_seen": 2743992320 + }, + { + "epoch": 0.83, + "learning_rate": 0.00013634503510531596, + "loss": 1.2985, + "theoretical_loss": 3.372285554254214, + "tokens_seen": 2744057856 + }, + { + "epoch": 0.83, + "learning_rate": 0.00013632898696088266, + "loss": 1.2089, + "theoretical_loss": 3.3722793023452, + "tokens_seen": 2744123392 + }, + { + "epoch": 0.83, + "learning_rate": 0.00013631293881644936, + "loss": 1.2352, + "theoretical_loss": 3.3722730506273, + "tokens_seen": 2744188928 + }, + { + "epoch": 0.83, + "learning_rate": 0.00013629689067201606, + "loss": 1.2529, + "theoretical_loss": 3.3722667991005033, + "tokens_seen": 2744254464 + }, + { + "epoch": 0.83, + "learning_rate": 0.00013628084252758276, + "loss": 1.2882, + "theoretical_loss": 3.3722605477648, + "tokens_seen": 2744320000 + }, + { + "epoch": 0.83, + "learning_rate": 0.00013626479438314947, + "loss": 1.2799, + "theoretical_loss": 3.372254296620179, + "tokens_seen": 2744385536 + }, + { + "epoch": 0.83, + "learning_rate": 0.00013624874623871614, + "loss": 1.2893, + "theoretical_loss": 3.372248045666631, + "tokens_seen": 2744451072 + }, + { + "epoch": 0.83, + "learning_rate": 0.00013623269809428284, + "loss": 1.2953, + "theoretical_loss": 3.372241794904144, + "tokens_seen": 2744516608 + }, + { + "epoch": 0.83, + "learning_rate": 0.00013621664994984954, + "loss": 1.2955, + "theoretical_loss": 3.372235544332709, + "tokens_seen": 2744582144 + }, + { + "epoch": 0.83, + "learning_rate": 0.00013620060180541624, + "loss": 1.256, + "theoretical_loss": 3.372229293952315, + "tokens_seen": 2744647680 + }, + { + "epoch": 0.83, + "learning_rate": 0.00013618455366098294, + "loss": 1.2819, + "theoretical_loss": 3.3722230437629515, + "tokens_seen": 2744713216 + }, + { + "epoch": 0.83, + "learning_rate": 0.00013616850551654964, + "loss": 1.356, + "theoretical_loss": 3.3722167937646086, + "tokens_seen": 2744778752 + }, + { + "epoch": 0.83, + "learning_rate": 0.00013615245737211634, + "loss": 1.2256, + "theoretical_loss": 3.3722105439572756, + "tokens_seen": 2744844288 + }, + { + "epoch": 0.83, + "learning_rate": 0.00013613640922768304, + "loss": 1.2956, + "theoretical_loss": 3.372204294340942, + "tokens_seen": 2744909824 + }, + { + "epoch": 0.83, + "learning_rate": 0.00013612036108324974, + "loss": 1.2074, + "theoretical_loss": 3.3721980449155966, + "tokens_seen": 2744975360 + }, + { + "epoch": 0.83, + "learning_rate": 0.00013610431293881644, + "loss": 1.2437, + "theoretical_loss": 3.372191795681231, + "tokens_seen": 2745040896 + }, + { + "epoch": 0.83, + "learning_rate": 0.00013608826479438315, + "loss": 1.1305, + "theoretical_loss": 3.372185546637833, + "tokens_seen": 2745106432 + }, + { + "epoch": 0.83, + "learning_rate": 0.00013607221664994985, + "loss": 1.1651, + "theoretical_loss": 3.3721792977853933, + "tokens_seen": 2745171968 + }, + { + "epoch": 0.83, + "learning_rate": 0.00013605616850551655, + "loss": 1.3066, + "theoretical_loss": 3.372173049123901, + "tokens_seen": 2745237504 + }, + { + "epoch": 0.83, + "learning_rate": 0.00013604012036108325, + "loss": 1.2043, + "theoretical_loss": 3.372166800653346, + "tokens_seen": 2745303040 + }, + { + "epoch": 0.83, + "learning_rate": 0.00013602407221664998, + "loss": 1.2837, + "theoretical_loss": 3.3721605523737175, + "tokens_seen": 2745368576 + }, + { + "epoch": 0.83, + "learning_rate": 0.00013600802407221668, + "loss": 1.1996, + "theoretical_loss": 3.3721543042850053, + "tokens_seen": 2745434112 + }, + { + "epoch": 0.83, + "learning_rate": 0.00013599197592778338, + "loss": 1.2194, + "theoretical_loss": 3.372148056387199, + "tokens_seen": 2745499648 + }, + { + "epoch": 0.83, + "learning_rate": 0.00013597592778335008, + "loss": 1.2298, + "theoretical_loss": 3.3721418086802886, + "tokens_seen": 2745565184 + }, + { + "epoch": 0.83, + "learning_rate": 0.00013595987963891678, + "loss": 1.3406, + "theoretical_loss": 3.3721355611642627, + "tokens_seen": 2745630720 + }, + { + "epoch": 0.83, + "learning_rate": 0.00013594383149448345, + "loss": 1.3527, + "theoretical_loss": 3.372129313839112, + "tokens_seen": 2745696256 + }, + { + "epoch": 0.83, + "learning_rate": 0.00013592778335005015, + "loss": 1.1598, + "theoretical_loss": 3.372123066704826, + "tokens_seen": 2745761792 + }, + { + "epoch": 0.83, + "learning_rate": 0.00013591173520561685, + "loss": 1.3388, + "theoretical_loss": 3.3721168197613935, + "tokens_seen": 2745827328 + }, + { + "epoch": 0.83, + "learning_rate": 0.00013589568706118355, + "loss": 1.2304, + "theoretical_loss": 3.372110573008805, + "tokens_seen": 2745892864 + }, + { + "epoch": 0.83, + "objective/train/docs_used": 997819, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.286963701248169, + "objective/train/theoretical_loss": 3.372104326447049, + "objective/train/tokens_used": 2766418400, + "theoretical_loss": 3.372104326447049, + "tokens_seen": 2745958400 + }, + { + "epoch": 0.83, + "learning_rate": 0.00013587963891675025, + "loss": 1.251, + "theoretical_loss": 3.372104326447049, + "tokens_seen": 2745958400 + }, + { + "epoch": 0.83, + "learning_rate": 0.00013586359077231695, + "loss": 1.2713, + "theoretical_loss": 3.3720980800761167, + "tokens_seen": 2746023936 + }, + { + "epoch": 0.83, + "learning_rate": 0.00013584754262788365, + "loss": 1.2629, + "theoretical_loss": 3.372091833895996, + "tokens_seen": 2746089472 + }, + { + "epoch": 0.83, + "learning_rate": 0.00013583149448345036, + "loss": 1.3127, + "theoretical_loss": 3.3720855879066782, + "tokens_seen": 2746155008 + }, + { + "epoch": 0.83, + "learning_rate": 0.00013581544633901706, + "loss": 1.3296, + "theoretical_loss": 3.372079342108152, + "tokens_seen": 2746220544 + }, + { + "epoch": 0.83, + "learning_rate": 0.00013579939819458376, + "loss": 1.2706, + "theoretical_loss": 3.372073096500406, + "tokens_seen": 2746286080 + }, + { + "epoch": 0.83, + "learning_rate": 0.00013578335005015046, + "loss": 1.1566, + "theoretical_loss": 3.3720668510834324, + "tokens_seen": 2746351616 + }, + { + "epoch": 0.83, + "learning_rate": 0.00013576730190571716, + "loss": 1.2693, + "theoretical_loss": 3.3720606058572185, + "tokens_seen": 2746417152 + }, + { + "epoch": 0.83, + "learning_rate": 0.00013575125376128386, + "loss": 1.1938, + "theoretical_loss": 3.372054360821755, + "tokens_seen": 2746482688 + }, + { + "epoch": 0.83, + "learning_rate": 0.00013573520561685056, + "loss": 1.2543, + "theoretical_loss": 3.372048115977031, + "tokens_seen": 2746548224 + }, + { + "epoch": 0.83, + "learning_rate": 0.00013571915747241726, + "loss": 1.3405, + "theoretical_loss": 3.372041871323037, + "tokens_seen": 2746613760 + }, + { + "epoch": 0.83, + "learning_rate": 0.00013570310932798396, + "loss": 1.3435, + "theoretical_loss": 3.3720356268597618, + "tokens_seen": 2746679296 + }, + { + "epoch": 0.83, + "learning_rate": 0.00013568706118355066, + "loss": 1.284, + "theoretical_loss": 3.372029382587195, + "tokens_seen": 2746744832 + }, + { + "epoch": 0.83, + "learning_rate": 0.00013567101303911736, + "loss": 1.1418, + "theoretical_loss": 3.3720231385053268, + "tokens_seen": 2746810368 + }, + { + "epoch": 0.83, + "learning_rate": 0.00013565496489468406, + "loss": 1.2412, + "theoretical_loss": 3.3720168946141467, + "tokens_seen": 2746875904 + }, + { + "epoch": 0.83, + "learning_rate": 0.00013563891675025076, + "loss": 1.2518, + "theoretical_loss": 3.3720106509136434, + "tokens_seen": 2746941440 + }, + { + "epoch": 0.83, + "learning_rate": 0.00013562286860581746, + "loss": 1.2688, + "theoretical_loss": 3.372004407403808, + "tokens_seen": 2747006976 + }, + { + "epoch": 0.83, + "learning_rate": 0.00013560682046138416, + "loss": 1.2656, + "theoretical_loss": 3.371998164084629, + "tokens_seen": 2747072512 + }, + { + "epoch": 0.83, + "learning_rate": 0.00013559077231695087, + "loss": 1.201, + "theoretical_loss": 3.3719919209560967, + "tokens_seen": 2747138048 + }, + { + "epoch": 0.83, + "learning_rate": 0.00013557472417251757, + "loss": 1.1671, + "theoretical_loss": 3.3719856780182003, + "tokens_seen": 2747203584 + }, + { + "epoch": 0.83, + "learning_rate": 0.00013555867602808424, + "loss": 1.2905, + "theoretical_loss": 3.3719794352709296, + "tokens_seen": 2747269120 + }, + { + "epoch": 0.83, + "learning_rate": 0.00013554262788365094, + "loss": 1.2906, + "theoretical_loss": 3.3719731927142744, + "tokens_seen": 2747334656 + }, + { + "epoch": 0.83, + "learning_rate": 0.00013552657973921764, + "loss": 1.3555, + "theoretical_loss": 3.3719669503482237, + "tokens_seen": 2747400192 + }, + { + "epoch": 0.83, + "learning_rate": 0.00013551053159478434, + "loss": 1.2646, + "theoretical_loss": 3.371960708172768, + "tokens_seen": 2747465728 + }, + { + "epoch": 0.83, + "learning_rate": 0.00013549448345035104, + "loss": 1.1509, + "theoretical_loss": 3.3719544661878964, + "tokens_seen": 2747531264 + }, + { + "epoch": 0.83, + "learning_rate": 0.00013547843530591774, + "loss": 1.3063, + "theoretical_loss": 3.3719482243935985, + "tokens_seen": 2747596800 + }, + { + "epoch": 0.83, + "learning_rate": 0.00013546238716148444, + "loss": 1.3518, + "theoretical_loss": 3.3719419827898647, + "tokens_seen": 2747662336 + }, + { + "epoch": 0.83, + "learning_rate": 0.00013544633901705114, + "loss": 1.2119, + "theoretical_loss": 3.3719357413766833, + "tokens_seen": 2747727872 + }, + { + "epoch": 0.83, + "learning_rate": 0.00013543029087261787, + "loss": 1.2769, + "theoretical_loss": 3.371929500154045, + "tokens_seen": 2747793408 + }, + { + "epoch": 0.83, + "learning_rate": 0.00013541424272818457, + "loss": 1.2041, + "theoretical_loss": 3.3719232591219392, + "tokens_seen": 2747858944 + }, + { + "epoch": 0.83, + "learning_rate": 0.00013539819458375127, + "loss": 1.2611, + "theoretical_loss": 3.3719170182803557, + "tokens_seen": 2747924480 + }, + { + "epoch": 0.83, + "learning_rate": 0.00013538214643931797, + "loss": 1.2324, + "theoretical_loss": 3.3719107776292834, + "tokens_seen": 2747990016 + }, + { + "epoch": 0.83, + "learning_rate": 0.00013536609829488467, + "loss": 1.27, + "theoretical_loss": 3.371904537168713, + "tokens_seen": 2748055552 + }, + { + "epoch": 0.83, + "learning_rate": 0.00013535005015045137, + "loss": 1.2801, + "theoretical_loss": 3.3718982968986326, + "tokens_seen": 2748121088 + }, + { + "epoch": 0.83, + "learning_rate": 0.00013533400200601808, + "loss": 1.2766, + "theoretical_loss": 3.371892056819034, + "tokens_seen": 2748186624 + }, + { + "epoch": 0.83, + "learning_rate": 0.00013531795386158478, + "loss": 1.3393, + "theoretical_loss": 3.371885816929905, + "tokens_seen": 2748252160 + }, + { + "epoch": 0.83, + "learning_rate": 0.00013530190571715148, + "loss": 1.3052, + "theoretical_loss": 3.3718795772312355, + "tokens_seen": 2748317696 + }, + { + "epoch": 0.83, + "learning_rate": 0.00013528585757271818, + "loss": 1.3065, + "theoretical_loss": 3.371873337723016, + "tokens_seen": 2748383232 + }, + { + "epoch": 0.83, + "learning_rate": 0.00013526980942828488, + "loss": 1.3314, + "theoretical_loss": 3.371867098405236, + "tokens_seen": 2748448768 + }, + { + "epoch": 0.83, + "learning_rate": 0.00013525376128385158, + "loss": 1.2096, + "theoretical_loss": 3.3718608592778847, + "tokens_seen": 2748514304 + }, + { + "epoch": 0.83, + "learning_rate": 0.00013523771313941825, + "loss": 1.2199, + "theoretical_loss": 3.371854620340952, + "tokens_seen": 2748579840 + }, + { + "epoch": 0.83, + "learning_rate": 0.00013522166499498495, + "loss": 1.3059, + "theoretical_loss": 3.3718483815944267, + "tokens_seen": 2748645376 + }, + { + "epoch": 0.83, + "learning_rate": 0.00013520561685055165, + "loss": 1.2104, + "theoretical_loss": 3.3718421430382994, + "tokens_seen": 2748710912 + }, + { + "epoch": 0.83, + "learning_rate": 0.00013518956870611835, + "loss": 1.3225, + "theoretical_loss": 3.3718359046725603, + "tokens_seen": 2748776448 + }, + { + "epoch": 0.83, + "learning_rate": 0.00013517352056168505, + "loss": 1.2906, + "theoretical_loss": 3.371829666497198, + "tokens_seen": 2748841984 + }, + { + "epoch": 0.83, + "learning_rate": 0.00013515747241725176, + "loss": 1.2591, + "theoretical_loss": 3.3718234285122017, + "tokens_seen": 2748907520 + }, + { + "epoch": 0.83, + "learning_rate": 0.00013514142427281846, + "loss": 1.3165, + "theoretical_loss": 3.3718171907175627, + "tokens_seen": 2748973056 + }, + { + "epoch": 0.83, + "learning_rate": 0.00013512537612838516, + "loss": 1.224, + "theoretical_loss": 3.371810953113269, + "tokens_seen": 2749038592 + }, + { + "epoch": 0.83, + "learning_rate": 0.00013510932798395186, + "loss": 1.2833, + "theoretical_loss": 3.371804715699312, + "tokens_seen": 2749104128 + }, + { + "epoch": 0.83, + "learning_rate": 0.00013509327983951856, + "loss": 1.2572, + "theoretical_loss": 3.3717984784756796, + "tokens_seen": 2749169664 + }, + { + "epoch": 0.83, + "objective/train/docs_used": 998936, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.3776664733886719, + "objective/train/theoretical_loss": 3.3717922414423622, + "objective/train/tokens_used": 2769695200, + "theoretical_loss": 3.3717922414423622, + "tokens_seen": 2749235200 + }, + { + "epoch": 0.83, + "learning_rate": 0.00013507723169508526, + "loss": 1.366, + "theoretical_loss": 3.3717922414423622, + "tokens_seen": 2749235200 + }, + { + "epoch": 0.83, + "learning_rate": 0.00013506118355065196, + "loss": 1.2957, + "theoretical_loss": 3.3717860045993495, + "tokens_seen": 2749300736 + }, + { + "epoch": 0.83, + "learning_rate": 0.00013504513540621866, + "loss": 1.2195, + "theoretical_loss": 3.3717797679466317, + "tokens_seen": 2749366272 + }, + { + "epoch": 0.83, + "learning_rate": 0.00013502908726178536, + "loss": 1.2588, + "theoretical_loss": 3.3717735314841972, + "tokens_seen": 2749431808 + }, + { + "epoch": 0.83, + "learning_rate": 0.00013501303911735206, + "loss": 1.3405, + "theoretical_loss": 3.3717672952120368, + "tokens_seen": 2749497344 + }, + { + "epoch": 0.83, + "learning_rate": 0.00013499699097291876, + "loss": 1.2717, + "theoretical_loss": 3.3717610591301392, + "tokens_seen": 2749562880 + }, + { + "epoch": 0.83, + "learning_rate": 0.00013498094282848546, + "loss": 1.3148, + "theoretical_loss": 3.371754823238495, + "tokens_seen": 2749628416 + }, + { + "epoch": 0.83, + "learning_rate": 0.00013496489468405216, + "loss": 1.1877, + "theoretical_loss": 3.3717485875370934, + "tokens_seen": 2749693952 + }, + { + "epoch": 0.83, + "learning_rate": 0.00013494884653961886, + "loss": 1.2351, + "theoretical_loss": 3.3717423520259238, + "tokens_seen": 2749759488 + }, + { + "epoch": 0.83, + "learning_rate": 0.00013493279839518556, + "loss": 1.2495, + "theoretical_loss": 3.371736116704976, + "tokens_seen": 2749825024 + }, + { + "epoch": 0.83, + "learning_rate": 0.00013491675025075226, + "loss": 1.2406, + "theoretical_loss": 3.3717298815742405, + "tokens_seen": 2749890560 + }, + { + "epoch": 0.83, + "learning_rate": 0.00013490070210631897, + "loss": 1.1898, + "theoretical_loss": 3.371723646633706, + "tokens_seen": 2749956096 + }, + { + "epoch": 0.83, + "learning_rate": 0.00013488465396188567, + "loss": 1.1774, + "theoretical_loss": 3.3717174118833624, + "tokens_seen": 2750021632 + }, + { + "epoch": 0.83, + "learning_rate": 0.00013486860581745237, + "loss": 1.2823, + "theoretical_loss": 3.371711177323199, + "tokens_seen": 2750087168 + }, + { + "epoch": 0.83, + "learning_rate": 0.00013485255767301904, + "loss": 1.2941, + "theoretical_loss": 3.3717049429532064, + "tokens_seen": 2750152704 + }, + { + "epoch": 0.83, + "learning_rate": 0.00013483650952858574, + "loss": 1.3094, + "theoretical_loss": 3.3716987087733736, + "tokens_seen": 2750218240 + }, + { + "epoch": 0.83, + "learning_rate": 0.00013482046138415247, + "loss": 1.3215, + "theoretical_loss": 3.3716924747836905, + "tokens_seen": 2750283776 + }, + { + "epoch": 0.83, + "learning_rate": 0.00013480441323971917, + "loss": 1.2559, + "theoretical_loss": 3.371686240984147, + "tokens_seen": 2750349312 + }, + { + "epoch": 0.83, + "learning_rate": 0.00013478836509528587, + "loss": 1.1969, + "theoretical_loss": 3.3716800073747315, + "tokens_seen": 2750414848 + }, + { + "epoch": 0.83, + "learning_rate": 0.00013477231695085257, + "loss": 1.3476, + "theoretical_loss": 3.371673773955435, + "tokens_seen": 2750480384 + }, + { + "epoch": 0.83, + "learning_rate": 0.00013475626880641927, + "loss": 1.2049, + "theoretical_loss": 3.3716675407262473, + "tokens_seen": 2750545920 + }, + { + "epoch": 0.83, + "learning_rate": 0.00013474022066198597, + "loss": 1.2816, + "theoretical_loss": 3.371661307687157, + "tokens_seen": 2750611456 + }, + { + "epoch": 0.83, + "learning_rate": 0.00013472417251755267, + "loss": 1.2405, + "theoretical_loss": 3.371655074838155, + "tokens_seen": 2750676992 + }, + { + "epoch": 0.83, + "learning_rate": 0.00013470812437311937, + "loss": 1.2287, + "theoretical_loss": 3.3716488421792294, + "tokens_seen": 2750742528 + }, + { + "epoch": 0.83, + "learning_rate": 0.00013469207622868607, + "loss": 1.4408, + "theoretical_loss": 3.3716426097103716, + "tokens_seen": 2750808064 + }, + { + "epoch": 0.83, + "learning_rate": 0.00013467602808425277, + "loss": 1.3198, + "theoretical_loss": 3.37163637743157, + "tokens_seen": 2750873600 + }, + { + "epoch": 0.83, + "learning_rate": 0.00013465997993981947, + "loss": 1.1469, + "theoretical_loss": 3.371630145342815, + "tokens_seen": 2750939136 + }, + { + "epoch": 0.83, + "learning_rate": 0.00013464393179538618, + "loss": 1.263, + "theoretical_loss": 3.3716239134440955, + "tokens_seen": 2751004672 + }, + { + "epoch": 0.83, + "learning_rate": 0.00013462788365095288, + "loss": 1.1982, + "theoretical_loss": 3.3716176817354024, + "tokens_seen": 2751070208 + }, + { + "epoch": 0.83, + "learning_rate": 0.00013461183550651958, + "loss": 1.2603, + "theoretical_loss": 3.3716114502167245, + "tokens_seen": 2751135744 + }, + { + "epoch": 0.83, + "learning_rate": 0.00013459578736208628, + "loss": 1.1871, + "theoretical_loss": 3.3716052188880514, + "tokens_seen": 2751201280 + }, + { + "epoch": 0.83, + "learning_rate": 0.00013457973921765298, + "loss": 1.241, + "theoretical_loss": 3.371598987749373, + "tokens_seen": 2751266816 + }, + { + "epoch": 0.83, + "learning_rate": 0.00013456369107321968, + "loss": 1.3575, + "theoretical_loss": 3.371592756800679, + "tokens_seen": 2751332352 + }, + { + "epoch": 0.83, + "learning_rate": 0.00013454764292878638, + "loss": 1.3168, + "theoretical_loss": 3.3715865260419595, + "tokens_seen": 2751397888 + }, + { + "epoch": 0.83, + "learning_rate": 0.00013453159478435305, + "loss": 1.2713, + "theoretical_loss": 3.3715802954732035, + "tokens_seen": 2751463424 + }, + { + "epoch": 0.83, + "learning_rate": 0.00013451554663991975, + "loss": 1.2683, + "theoretical_loss": 3.3715740650944013, + "tokens_seen": 2751528960 + }, + { + "epoch": 0.83, + "learning_rate": 0.00013449949849548645, + "loss": 1.1977, + "theoretical_loss": 3.371567834905542, + "tokens_seen": 2751594496 + }, + { + "epoch": 0.83, + "learning_rate": 0.00013448345035105315, + "loss": 1.2325, + "theoretical_loss": 3.3715616049066153, + "tokens_seen": 2751660032 + }, + { + "epoch": 0.83, + "learning_rate": 0.00013446740220661986, + "loss": 1.2172, + "theoretical_loss": 3.3715553750976115, + "tokens_seen": 2751725568 + }, + { + "epoch": 0.83, + "learning_rate": 0.00013445135406218656, + "loss": 1.2158, + "theoretical_loss": 3.37154914547852, + "tokens_seen": 2751791104 + }, + { + "epoch": 0.83, + "learning_rate": 0.00013443530591775326, + "loss": 1.2827, + "theoretical_loss": 3.37154291604933, + "tokens_seen": 2751856640 + }, + { + "epoch": 0.83, + "learning_rate": 0.00013441925777331996, + "loss": 1.1626, + "theoretical_loss": 3.371536686810032, + "tokens_seen": 2751922176 + }, + { + "epoch": 0.83, + "learning_rate": 0.00013440320962888666, + "loss": 1.2481, + "theoretical_loss": 3.3715304577606147, + "tokens_seen": 2751987712 + }, + { + "epoch": 0.83, + "learning_rate": 0.00013438716148445336, + "loss": 1.1919, + "theoretical_loss": 3.3715242289010687, + "tokens_seen": 2752053248 + }, + { + "epoch": 0.83, + "learning_rate": 0.00013437111334002006, + "loss": 1.2081, + "theoretical_loss": 3.3715180002313834, + "tokens_seen": 2752118784 + }, + { + "epoch": 0.83, + "learning_rate": 0.00013435506519558676, + "loss": 1.3711, + "theoretical_loss": 3.3715117717515484, + "tokens_seen": 2752184320 + }, + { + "epoch": 0.83, + "learning_rate": 0.00013433901705115346, + "loss": 1.2445, + "theoretical_loss": 3.3715055434615535, + "tokens_seen": 2752249856 + }, + { + "epoch": 0.83, + "learning_rate": 0.00013432296890672016, + "loss": 1.2365, + "theoretical_loss": 3.371499315361388, + "tokens_seen": 2752315392 + }, + { + "epoch": 0.83, + "learning_rate": 0.00013430692076228686, + "loss": 1.2521, + "theoretical_loss": 3.3714930874510425, + "tokens_seen": 2752380928 + }, + { + "epoch": 0.83, + "learning_rate": 0.00013429087261785356, + "loss": 1.2621, + "theoretical_loss": 3.371486859730506, + "tokens_seen": 2752446464 + }, + { + "epoch": 0.83, + "objective/train/docs_used": 1000102, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.2884702682495117, + "objective/train/theoretical_loss": 3.371480632199768, + "objective/train/tokens_used": 2772972000, + "theoretical_loss": 3.371480632199768, + "tokens_seen": 2752512000 + }, + { + "epoch": 0.83, + "learning_rate": 0.00013427482447342026, + "loss": 1.2891, + "theoretical_loss": 3.371480632199768, + "tokens_seen": 2752512000 + }, + { + "epoch": 0.83, + "learning_rate": 0.00013425877632898696, + "loss": 1.1781, + "theoretical_loss": 3.3714744048588186, + "tokens_seen": 2752577536 + }, + { + "epoch": 0.83, + "learning_rate": 0.00013424272818455366, + "loss": 1.357, + "theoretical_loss": 3.371468177707648, + "tokens_seen": 2752643072 + }, + { + "epoch": 0.83, + "learning_rate": 0.00013422668004012036, + "loss": 1.2128, + "theoretical_loss": 3.3714619507462444, + "tokens_seen": 2752708608 + }, + { + "epoch": 0.83, + "learning_rate": 0.00013421063189568707, + "loss": 1.3133, + "theoretical_loss": 3.371455723974599, + "tokens_seen": 2752774144 + }, + { + "epoch": 0.83, + "learning_rate": 0.00013419458375125377, + "loss": 1.2225, + "theoretical_loss": 3.371449497392701, + "tokens_seen": 2752839680 + }, + { + "epoch": 0.83, + "learning_rate": 0.00013417853560682047, + "loss": 1.1937, + "theoretical_loss": 3.3714432710005395, + "tokens_seen": 2752905216 + }, + { + "epoch": 0.83, + "learning_rate": 0.00013416248746238717, + "loss": 1.3363, + "theoretical_loss": 3.371437044798105, + "tokens_seen": 2752970752 + }, + { + "epoch": 0.83, + "learning_rate": 0.00013414643931795387, + "loss": 1.3056, + "theoretical_loss": 3.371430818785387, + "tokens_seen": 2753036288 + }, + { + "epoch": 0.83, + "learning_rate": 0.00013413039117352057, + "loss": 1.4636, + "theoretical_loss": 3.371424592962375, + "tokens_seen": 2753101824 + }, + { + "epoch": 0.83, + "learning_rate": 0.00013411434302908727, + "loss": 1.2602, + "theoretical_loss": 3.371418367329059, + "tokens_seen": 2753167360 + }, + { + "epoch": 0.83, + "learning_rate": 0.00013409829488465397, + "loss": 1.3562, + "theoretical_loss": 3.371412141885428, + "tokens_seen": 2753232896 + }, + { + "epoch": 0.83, + "learning_rate": 0.00013408224674022067, + "loss": 1.2508, + "theoretical_loss": 3.371405916631473, + "tokens_seen": 2753298432 + }, + { + "epoch": 0.83, + "learning_rate": 0.00013406619859578737, + "loss": 1.2354, + "theoretical_loss": 3.3713996915671824, + "tokens_seen": 2753363968 + }, + { + "epoch": 0.83, + "learning_rate": 0.00013405015045135407, + "loss": 1.3372, + "theoretical_loss": 3.371393466692547, + "tokens_seen": 2753429504 + }, + { + "epoch": 0.83, + "learning_rate": 0.00013403410230692077, + "loss": 1.2885, + "theoretical_loss": 3.3713872420075557, + "tokens_seen": 2753495040 + }, + { + "epoch": 0.83, + "learning_rate": 0.00013401805416248747, + "loss": 1.2127, + "theoretical_loss": 3.3713810175121983, + "tokens_seen": 2753560576 + }, + { + "epoch": 0.83, + "learning_rate": 0.00013400200601805417, + "loss": 1.3175, + "theoretical_loss": 3.371374793206465, + "tokens_seen": 2753626112 + }, + { + "epoch": 0.83, + "learning_rate": 0.00013398595787362087, + "loss": 1.1469, + "theoretical_loss": 3.3713685690903445, + "tokens_seen": 2753691648 + }, + { + "epoch": 0.83, + "learning_rate": 0.00013396990972918757, + "loss": 1.2803, + "theoretical_loss": 3.3713623451638277, + "tokens_seen": 2753757184 + }, + { + "epoch": 0.83, + "learning_rate": 0.00013395386158475428, + "loss": 1.2817, + "theoretical_loss": 3.3713561214269037, + "tokens_seen": 2753822720 + }, + { + "epoch": 0.83, + "learning_rate": 0.00013393781344032098, + "loss": 1.3363, + "theoretical_loss": 3.3713498978795626, + "tokens_seen": 2753888256 + }, + { + "epoch": 0.83, + "learning_rate": 0.00013392176529588768, + "loss": 1.2465, + "theoretical_loss": 3.3713436745217935, + "tokens_seen": 2753953792 + }, + { + "epoch": 0.83, + "learning_rate": 0.00013390571715145438, + "loss": 1.1557, + "theoretical_loss": 3.371337451353587, + "tokens_seen": 2754019328 + }, + { + "epoch": 0.83, + "learning_rate": 0.00013388966900702108, + "loss": 1.2556, + "theoretical_loss": 3.3713312283749315, + "tokens_seen": 2754084864 + }, + { + "epoch": 0.83, + "learning_rate": 0.00013387362086258778, + "loss": 1.3721, + "theoretical_loss": 3.3713250055858177, + "tokens_seen": 2754150400 + }, + { + "epoch": 0.83, + "learning_rate": 0.00013385757271815448, + "loss": 1.1655, + "theoretical_loss": 3.3713187829862354, + "tokens_seen": 2754215936 + }, + { + "epoch": 0.83, + "learning_rate": 0.00013384152457372115, + "loss": 1.2158, + "theoretical_loss": 3.371312560576174, + "tokens_seen": 2754281472 + }, + { + "epoch": 0.83, + "learning_rate": 0.00013382547642928785, + "loss": 1.29, + "theoretical_loss": 3.371306338355623, + "tokens_seen": 2754347008 + }, + { + "epoch": 0.83, + "learning_rate": 0.00013380942828485455, + "loss": 1.1762, + "theoretical_loss": 3.3713001163245724, + "tokens_seen": 2754412544 + }, + { + "epoch": 0.83, + "learning_rate": 0.00013379338014042125, + "loss": 1.266, + "theoretical_loss": 3.3712938944830118, + "tokens_seen": 2754478080 + }, + { + "epoch": 0.83, + "learning_rate": 0.00013377733199598796, + "loss": 1.249, + "theoretical_loss": 3.3712876728309316, + "tokens_seen": 2754543616 + }, + { + "epoch": 0.83, + "learning_rate": 0.00013376128385155466, + "loss": 1.2995, + "theoretical_loss": 3.3712814513683202, + "tokens_seen": 2754609152 + }, + { + "epoch": 0.83, + "learning_rate": 0.00013374523570712136, + "loss": 1.1982, + "theoretical_loss": 3.3712752300951685, + "tokens_seen": 2754674688 + }, + { + "epoch": 0.83, + "learning_rate": 0.00013372918756268806, + "loss": 1.2562, + "theoretical_loss": 3.371269009011465, + "tokens_seen": 2754740224 + }, + { + "epoch": 0.83, + "learning_rate": 0.00013371313941825476, + "loss": 1.2885, + "theoretical_loss": 3.371262788117201, + "tokens_seen": 2754805760 + }, + { + "epoch": 0.83, + "learning_rate": 0.00013369709127382146, + "loss": 1.3262, + "theoretical_loss": 3.3712565674123653, + "tokens_seen": 2754871296 + }, + { + "epoch": 0.83, + "learning_rate": 0.00013368104312938816, + "loss": 1.2037, + "theoretical_loss": 3.3712503468969475, + "tokens_seen": 2754936832 + }, + { + "epoch": 0.83, + "learning_rate": 0.00013366499498495486, + "loss": 1.2787, + "theoretical_loss": 3.3712441265709376, + "tokens_seen": 2755002368 + }, + { + "epoch": 0.83, + "learning_rate": 0.00013364894684052156, + "loss": 1.2365, + "theoretical_loss": 3.3712379064343256, + "tokens_seen": 2755067904 + }, + { + "epoch": 0.83, + "learning_rate": 0.0001336328986960883, + "loss": 1.2872, + "theoretical_loss": 3.371231686487101, + "tokens_seen": 2755133440 + }, + { + "epoch": 0.83, + "learning_rate": 0.000133616850551655, + "loss": 1.3268, + "theoretical_loss": 3.371225466729253, + "tokens_seen": 2755198976 + }, + { + "epoch": 0.83, + "learning_rate": 0.0001336008024072217, + "loss": 1.3127, + "theoretical_loss": 3.371219247160772, + "tokens_seen": 2755264512 + }, + { + "epoch": 0.83, + "learning_rate": 0.0001335847542627884, + "loss": 1.3624, + "theoretical_loss": 3.371213027781647, + "tokens_seen": 2755330048 + }, + { + "epoch": 0.83, + "learning_rate": 0.0001335687061183551, + "loss": 1.3258, + "theoretical_loss": 3.371206808591869, + "tokens_seen": 2755395584 + }, + { + "epoch": 0.83, + "learning_rate": 0.0001335526579739218, + "loss": 1.2731, + "theoretical_loss": 3.3712005895914263, + "tokens_seen": 2755461120 + }, + { + "epoch": 0.84, + "learning_rate": 0.0001335366098294885, + "loss": 1.3114, + "theoretical_loss": 3.3711943707803096, + "tokens_seen": 2755526656 + }, + { + "epoch": 0.84, + "learning_rate": 0.00013352056168505517, + "loss": 1.2304, + "theoretical_loss": 3.371188152158509, + "tokens_seen": 2755592192 + }, + { + "epoch": 0.84, + "learning_rate": 0.00013350451354062187, + "loss": 1.3198, + "theoretical_loss": 3.371181933726013, + "tokens_seen": 2755657728 + }, + { + "epoch": 0.84, + "learning_rate": 0.00013348846539618857, + "loss": 1.4059, + "theoretical_loss": 3.3711757154828117, + "tokens_seen": 2755723264 + }, + { + "epoch": 0.84, + "objective/train/docs_used": 1001278, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.0827292203903198, + "objective/train/theoretical_loss": 3.371169497428895, + "objective/train/tokens_used": 2776248800, + "theoretical_loss": 3.371169497428895, + "tokens_seen": 2755788800 + }, + { + "epoch": 0.84, + "learning_rate": 0.00013347241725175527, + "loss": 1.1825, + "theoretical_loss": 3.371169497428895, + "tokens_seen": 2755788800 + }, + { + "epoch": 0.84, + "learning_rate": 0.00013345636910732197, + "loss": 1.2913, + "theoretical_loss": 3.371163279564253, + "tokens_seen": 2755854336 + }, + { + "epoch": 0.84, + "learning_rate": 0.00013344032096288867, + "loss": 1.3779, + "theoretical_loss": 3.371157061888875, + "tokens_seen": 2755919872 + }, + { + "epoch": 0.84, + "learning_rate": 0.00013342427281845537, + "loss": 1.2279, + "theoretical_loss": 3.371150844402751, + "tokens_seen": 2755985408 + }, + { + "epoch": 0.84, + "learning_rate": 0.00013340822467402207, + "loss": 1.2796, + "theoretical_loss": 3.3711446271058705, + "tokens_seen": 2756050944 + }, + { + "epoch": 0.84, + "learning_rate": 0.00013339217652958877, + "loss": 1.3124, + "theoretical_loss": 3.3711384099982236, + "tokens_seen": 2756116480 + }, + { + "epoch": 0.84, + "learning_rate": 0.00013337612838515547, + "loss": 1.3066, + "theoretical_loss": 3.3711321930797995, + "tokens_seen": 2756182016 + }, + { + "epoch": 0.84, + "learning_rate": 0.00013336008024072217, + "loss": 1.2009, + "theoretical_loss": 3.371125976350588, + "tokens_seen": 2756247552 + }, + { + "epoch": 0.84, + "learning_rate": 0.00013334403209628887, + "loss": 1.2006, + "theoretical_loss": 3.371119759810579, + "tokens_seen": 2756313088 + }, + { + "epoch": 0.84, + "learning_rate": 0.00013332798395185557, + "loss": 1.3068, + "theoretical_loss": 3.3711135434597628, + "tokens_seen": 2756378624 + }, + { + "epoch": 0.84, + "learning_rate": 0.00013331193580742227, + "loss": 1.3115, + "theoretical_loss": 3.371107327298129, + "tokens_seen": 2756444160 + }, + { + "epoch": 0.84, + "learning_rate": 0.00013329588766298897, + "loss": 1.1899, + "theoretical_loss": 3.3711011113256664, + "tokens_seen": 2756509696 + }, + { + "epoch": 0.84, + "learning_rate": 0.00013327983951855568, + "loss": 1.393, + "theoretical_loss": 3.371094895542365, + "tokens_seen": 2756575232 + }, + { + "epoch": 0.84, + "learning_rate": 0.00013326379137412238, + "loss": 1.2216, + "theoretical_loss": 3.3710886799482154, + "tokens_seen": 2756640768 + }, + { + "epoch": 0.84, + "learning_rate": 0.00013324774322968908, + "loss": 1.2124, + "theoretical_loss": 3.371082464543207, + "tokens_seen": 2756706304 + }, + { + "epoch": 0.84, + "learning_rate": 0.00013323169508525578, + "loss": 1.2435, + "theoretical_loss": 3.371076249327329, + "tokens_seen": 2756771840 + }, + { + "epoch": 0.84, + "learning_rate": 0.00013321564694082248, + "loss": 1.2859, + "theoretical_loss": 3.3710700343005717, + "tokens_seen": 2756837376 + }, + { + "epoch": 0.84, + "learning_rate": 0.00013319959879638918, + "loss": 1.2905, + "theoretical_loss": 3.3710638194629245, + "tokens_seen": 2756902912 + }, + { + "epoch": 0.84, + "learning_rate": 0.00013318355065195588, + "loss": 1.167, + "theoretical_loss": 3.3710576048143777, + "tokens_seen": 2756968448 + }, + { + "epoch": 0.84, + "learning_rate": 0.00013316750250752258, + "loss": 1.3693, + "theoretical_loss": 3.3710513903549204, + "tokens_seen": 2757033984 + }, + { + "epoch": 0.84, + "learning_rate": 0.00013315145436308928, + "loss": 1.2992, + "theoretical_loss": 3.371045176084543, + "tokens_seen": 2757099520 + }, + { + "epoch": 0.84, + "learning_rate": 0.00013313540621865595, + "loss": 1.3652, + "theoretical_loss": 3.3710389620032344, + "tokens_seen": 2757165056 + }, + { + "epoch": 0.84, + "learning_rate": 0.00013311935807422265, + "loss": 1.3087, + "theoretical_loss": 3.371032748110985, + "tokens_seen": 2757230592 + }, + { + "epoch": 0.84, + "learning_rate": 0.00013310330992978936, + "loss": 1.2346, + "theoretical_loss": 3.3710265344077843, + "tokens_seen": 2757296128 + }, + { + "epoch": 0.84, + "learning_rate": 0.00013308726178535606, + "loss": 1.3586, + "theoretical_loss": 3.3710203208936225, + "tokens_seen": 2757361664 + }, + { + "epoch": 0.84, + "learning_rate": 0.00013307121364092276, + "loss": 1.3127, + "theoretical_loss": 3.371014107568489, + "tokens_seen": 2757427200 + }, + { + "epoch": 0.84, + "learning_rate": 0.00013305516549648946, + "loss": 1.3005, + "theoretical_loss": 3.371007894432373, + "tokens_seen": 2757492736 + }, + { + "epoch": 0.84, + "learning_rate": 0.00013303911735205616, + "loss": 1.3491, + "theoretical_loss": 3.3710016814852652, + "tokens_seen": 2757558272 + }, + { + "epoch": 0.84, + "learning_rate": 0.00013302306920762289, + "loss": 1.2886, + "theoretical_loss": 3.3709954687271555, + "tokens_seen": 2757623808 + }, + { + "epoch": 0.84, + "learning_rate": 0.00013300702106318959, + "loss": 1.3413, + "theoretical_loss": 3.3709892561580324, + "tokens_seen": 2757689344 + }, + { + "epoch": 0.84, + "learning_rate": 0.0001329909729187563, + "loss": 1.2128, + "theoretical_loss": 3.3709830437778865, + "tokens_seen": 2757754880 + }, + { + "epoch": 0.84, + "learning_rate": 0.000132974924774323, + "loss": 1.2731, + "theoretical_loss": 3.3709768315867077, + "tokens_seen": 2757820416 + }, + { + "epoch": 0.84, + "learning_rate": 0.0001329588766298897, + "loss": 1.3376, + "theoretical_loss": 3.3709706195844857, + "tokens_seen": 2757885952 + }, + { + "epoch": 0.84, + "learning_rate": 0.0001329428284854564, + "loss": 1.342, + "theoretical_loss": 3.3709644077712095, + "tokens_seen": 2757951488 + }, + { + "epoch": 0.84, + "learning_rate": 0.0001329267803410231, + "loss": 1.3747, + "theoretical_loss": 3.3709581961468698, + "tokens_seen": 2758017024 + }, + { + "epoch": 0.84, + "learning_rate": 0.0001329107321965898, + "loss": 1.1841, + "theoretical_loss": 3.370951984711456, + "tokens_seen": 2758082560 + }, + { + "epoch": 0.84, + "learning_rate": 0.0001328946840521565, + "loss": 1.3352, + "theoretical_loss": 3.3709457734649577, + "tokens_seen": 2758148096 + }, + { + "epoch": 0.84, + "learning_rate": 0.0001328786359077232, + "loss": 1.2641, + "theoretical_loss": 3.370939562407365, + "tokens_seen": 2758213632 + }, + { + "epoch": 0.84, + "learning_rate": 0.0001328625877632899, + "loss": 1.3387, + "theoretical_loss": 3.370933351538668, + "tokens_seen": 2758279168 + }, + { + "epoch": 0.84, + "learning_rate": 0.0001328465396188566, + "loss": 1.2586, + "theoretical_loss": 3.3709271408588553, + "tokens_seen": 2758344704 + }, + { + "epoch": 0.84, + "learning_rate": 0.0001328304914744233, + "loss": 1.2919, + "theoretical_loss": 3.3709209303679177, + "tokens_seen": 2758410240 + }, + { + "epoch": 0.84, + "learning_rate": 0.00013281444332998997, + "loss": 1.3599, + "theoretical_loss": 3.3709147200658443, + "tokens_seen": 2758475776 + }, + { + "epoch": 0.84, + "learning_rate": 0.00013279839518555667, + "loss": 1.2086, + "theoretical_loss": 3.3709085099526255, + "tokens_seen": 2758541312 + }, + { + "epoch": 0.84, + "learning_rate": 0.00013278234704112337, + "loss": 1.2582, + "theoretical_loss": 3.3709023000282503, + "tokens_seen": 2758606848 + }, + { + "epoch": 0.84, + "learning_rate": 0.00013276629889669007, + "loss": 1.3354, + "theoretical_loss": 3.3708960902927094, + "tokens_seen": 2758672384 + }, + { + "epoch": 0.84, + "learning_rate": 0.00013275025075225677, + "loss": 1.2909, + "theoretical_loss": 3.370889880745992, + "tokens_seen": 2758737920 + }, + { + "epoch": 0.84, + "learning_rate": 0.00013273420260782347, + "loss": 1.3207, + "theoretical_loss": 3.370883671388088, + "tokens_seen": 2758803456 + }, + { + "epoch": 0.84, + "learning_rate": 0.00013271815446339017, + "loss": 1.2837, + "theoretical_loss": 3.3708774622189868, + "tokens_seen": 2758868992 + }, + { + "epoch": 0.84, + "learning_rate": 0.00013270210631895687, + "loss": 1.3275, + "theoretical_loss": 3.370871253238679, + "tokens_seen": 2758934528 + }, + { + "epoch": 0.84, + "learning_rate": 0.00013268605817452357, + "loss": 1.3047, + "theoretical_loss": 3.3708650444471537, + "tokens_seen": 2759000064 + }, + { + "epoch": 0.84, + "objective/train/docs_used": 1002432, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.4378294944763184, + "objective/train/theoretical_loss": 3.370858835844401, + "objective/train/tokens_used": 2779525600, + "theoretical_loss": 3.370858835844401, + "tokens_seen": 2759065600 + }, + { + "epoch": 0.84, + "learning_rate": 0.00013267001003009027, + "loss": 1.3561, + "theoretical_loss": 3.370858835844401, + "tokens_seen": 2759065600 + }, + { + "epoch": 0.84, + "learning_rate": 0.00013265396188565697, + "loss": 1.2212, + "theoretical_loss": 3.37085262743041, + "tokens_seen": 2759131136 + }, + { + "epoch": 0.84, + "learning_rate": 0.00013263791374122367, + "loss": 1.2665, + "theoretical_loss": 3.3708464192051717, + "tokens_seen": 2759196672 + }, + { + "epoch": 0.84, + "learning_rate": 0.00013262186559679037, + "loss": 1.329, + "theoretical_loss": 3.370840211168675, + "tokens_seen": 2759262208 + }, + { + "epoch": 0.84, + "learning_rate": 0.00013260581745235707, + "loss": 1.2193, + "theoretical_loss": 3.3708340033209097, + "tokens_seen": 2759327744 + }, + { + "epoch": 0.84, + "learning_rate": 0.00013258976930792378, + "loss": 1.1737, + "theoretical_loss": 3.370827795661866, + "tokens_seen": 2759393280 + }, + { + "epoch": 0.84, + "learning_rate": 0.00013257372116349048, + "loss": 1.2773, + "theoretical_loss": 3.370821588191533, + "tokens_seen": 2759458816 + }, + { + "epoch": 0.84, + "learning_rate": 0.00013255767301905718, + "loss": 1.3733, + "theoretical_loss": 3.3708153809099013, + "tokens_seen": 2759524352 + }, + { + "epoch": 0.84, + "learning_rate": 0.00013254162487462388, + "loss": 1.2902, + "theoretical_loss": 3.3708091738169603, + "tokens_seen": 2759589888 + }, + { + "epoch": 0.84, + "learning_rate": 0.00013252557673019058, + "loss": 1.2051, + "theoretical_loss": 3.3708029669127, + "tokens_seen": 2759655424 + }, + { + "epoch": 0.84, + "learning_rate": 0.00013250952858575728, + "loss": 1.2335, + "theoretical_loss": 3.37079676019711, + "tokens_seen": 2759720960 + }, + { + "epoch": 0.84, + "learning_rate": 0.00013249348044132398, + "loss": 1.181, + "theoretical_loss": 3.3707905536701794, + "tokens_seen": 2759786496 + }, + { + "epoch": 0.84, + "learning_rate": 0.00013247743229689068, + "loss": 1.1947, + "theoretical_loss": 3.3707843473318992, + "tokens_seen": 2759852032 + }, + { + "epoch": 0.84, + "learning_rate": 0.00013246138415245738, + "loss": 1.2823, + "theoretical_loss": 3.3707781411822584, + "tokens_seen": 2759917568 + }, + { + "epoch": 0.84, + "learning_rate": 0.00013244533600802408, + "loss": 1.3314, + "theoretical_loss": 3.3707719352212475, + "tokens_seen": 2759983104 + }, + { + "epoch": 0.84, + "learning_rate": 0.00013242928786359078, + "loss": 1.2544, + "theoretical_loss": 3.370765729448855, + "tokens_seen": 2760048640 + }, + { + "epoch": 0.84, + "learning_rate": 0.00013241323971915748, + "loss": 1.3043, + "theoretical_loss": 3.3707595238650723, + "tokens_seen": 2760114176 + }, + { + "epoch": 0.84, + "learning_rate": 0.00013239719157472418, + "loss": 1.2502, + "theoretical_loss": 3.370753318469888, + "tokens_seen": 2760179712 + }, + { + "epoch": 0.84, + "learning_rate": 0.00013238114343029088, + "loss": 1.2475, + "theoretical_loss": 3.3707471132632922, + "tokens_seen": 2760245248 + }, + { + "epoch": 0.84, + "learning_rate": 0.00013236509528585758, + "loss": 1.266, + "theoretical_loss": 3.3707409082452746, + "tokens_seen": 2760310784 + }, + { + "epoch": 0.84, + "learning_rate": 0.00013234904714142428, + "loss": 1.2637, + "theoretical_loss": 3.3707347034158257, + "tokens_seen": 2760376320 + }, + { + "epoch": 0.84, + "learning_rate": 0.00013233299899699099, + "loss": 1.4272, + "theoretical_loss": 3.3707284987749344, + "tokens_seen": 2760441856 + }, + { + "epoch": 0.84, + "learning_rate": 0.00013231695085255769, + "loss": 1.2996, + "theoretical_loss": 3.370722294322591, + "tokens_seen": 2760507392 + }, + { + "epoch": 0.84, + "learning_rate": 0.0001323009027081244, + "loss": 1.2877, + "theoretical_loss": 3.370716090058785, + "tokens_seen": 2760572928 + }, + { + "epoch": 0.84, + "learning_rate": 0.0001322848545636911, + "loss": 1.2835, + "theoretical_loss": 3.3707098859835063, + "tokens_seen": 2760638464 + }, + { + "epoch": 0.84, + "learning_rate": 0.0001322688064192578, + "loss": 1.3397, + "theoretical_loss": 3.370703682096745, + "tokens_seen": 2760704000 + }, + { + "epoch": 0.84, + "learning_rate": 0.0001322527582748245, + "loss": 1.3372, + "theoretical_loss": 3.3706974783984904, + "tokens_seen": 2760769536 + }, + { + "epoch": 0.84, + "learning_rate": 0.0001322367101303912, + "loss": 1.2636, + "theoretical_loss": 3.3706912748887325, + "tokens_seen": 2760835072 + }, + { + "epoch": 0.84, + "learning_rate": 0.0001322206619859579, + "loss": 1.2254, + "theoretical_loss": 3.370685071567461, + "tokens_seen": 2760900608 + }, + { + "epoch": 0.84, + "learning_rate": 0.0001322046138415246, + "loss": 1.2634, + "theoretical_loss": 3.3706788684346662, + "tokens_seen": 2760966144 + }, + { + "epoch": 0.84, + "learning_rate": 0.0001321885656970913, + "loss": 1.2937, + "theoretical_loss": 3.3706726654903374, + "tokens_seen": 2761031680 + }, + { + "epoch": 0.84, + "learning_rate": 0.000132172517552658, + "loss": 1.2234, + "theoretical_loss": 3.3706664627344645, + "tokens_seen": 2761097216 + }, + { + "epoch": 0.84, + "learning_rate": 0.0001321564694082247, + "loss": 1.2048, + "theoretical_loss": 3.3706602601670372, + "tokens_seen": 2761162752 + }, + { + "epoch": 0.84, + "learning_rate": 0.0001321404212637914, + "loss": 1.231, + "theoretical_loss": 3.3706540577880455, + "tokens_seen": 2761228288 + }, + { + "epoch": 0.84, + "learning_rate": 0.0001321243731193581, + "loss": 1.2843, + "theoretical_loss": 3.3706478555974795, + "tokens_seen": 2761293824 + }, + { + "epoch": 0.84, + "learning_rate": 0.00013210832497492477, + "loss": 1.3202, + "theoretical_loss": 3.370641653595328, + "tokens_seen": 2761359360 + }, + { + "epoch": 0.84, + "learning_rate": 0.00013209227683049147, + "loss": 1.3572, + "theoretical_loss": 3.370635451781582, + "tokens_seen": 2761424896 + }, + { + "epoch": 0.84, + "learning_rate": 0.00013207622868605817, + "loss": 1.341, + "theoretical_loss": 3.3706292501562305, + "tokens_seen": 2761490432 + }, + { + "epoch": 0.84, + "learning_rate": 0.00013206018054162487, + "loss": 1.3257, + "theoretical_loss": 3.3706230487192634, + "tokens_seen": 2761555968 + }, + { + "epoch": 0.84, + "learning_rate": 0.00013204413239719157, + "loss": 1.2736, + "theoretical_loss": 3.3706168474706706, + "tokens_seen": 2761621504 + }, + { + "epoch": 0.84, + "learning_rate": 0.00013202808425275827, + "loss": 1.348, + "theoretical_loss": 3.3706106464104426, + "tokens_seen": 2761687040 + }, + { + "epoch": 0.84, + "learning_rate": 0.00013201203610832497, + "loss": 1.1921, + "theoretical_loss": 3.370604445538568, + "tokens_seen": 2761752576 + }, + { + "epoch": 0.84, + "learning_rate": 0.00013199598796389167, + "loss": 1.3635, + "theoretical_loss": 3.3705982448550373, + "tokens_seen": 2761818112 + }, + { + "epoch": 0.84, + "learning_rate": 0.00013197993981945837, + "loss": 1.2242, + "theoretical_loss": 3.3705920443598405, + "tokens_seen": 2761883648 + }, + { + "epoch": 0.84, + "learning_rate": 0.00013196389167502507, + "loss": 1.2091, + "theoretical_loss": 3.3705858440529664, + "tokens_seen": 2761949184 + }, + { + "epoch": 0.84, + "learning_rate": 0.00013194784353059177, + "loss": 1.3046, + "theoretical_loss": 3.370579643934406, + "tokens_seen": 2762014720 + }, + { + "epoch": 0.84, + "learning_rate": 0.00013193179538615847, + "loss": 1.159, + "theoretical_loss": 3.370573444004149, + "tokens_seen": 2762080256 + }, + { + "epoch": 0.84, + "learning_rate": 0.00013191574724172517, + "loss": 1.2475, + "theoretical_loss": 3.370567244262184, + "tokens_seen": 2762145792 + }, + { + "epoch": 0.84, + "learning_rate": 0.00013189969909729188, + "loss": 1.1683, + "theoretical_loss": 3.370561044708502, + "tokens_seen": 2762211328 + }, + { + "epoch": 0.84, + "learning_rate": 0.00013188365095285858, + "loss": 1.2086, + "theoretical_loss": 3.3705548453430927, + "tokens_seen": 2762276864 + }, + { + "epoch": 0.84, + "objective/train/docs_used": 1003480, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.2719088792800903, + "objective/train/theoretical_loss": 3.370548646165946, + "objective/train/tokens_used": 2782802400, + "theoretical_loss": 3.370548646165946, + "tokens_seen": 2762342400 + }, + { + "epoch": 0.84, + "learning_rate": 0.00013186760280842528, + "loss": 1.2671, + "theoretical_loss": 3.370548646165946, + "tokens_seen": 2762342400 + }, + { + "epoch": 0.84, + "learning_rate": 0.00013185155466399198, + "loss": 1.249, + "theoretical_loss": 3.3705424471770504, + "tokens_seen": 2762407936 + }, + { + "epoch": 0.84, + "learning_rate": 0.00013183550651955868, + "loss": 1.3505, + "theoretical_loss": 3.3705362483763976, + "tokens_seen": 2762473472 + }, + { + "epoch": 0.84, + "learning_rate": 0.0001318194583751254, + "loss": 1.342, + "theoretical_loss": 3.370530049763976, + "tokens_seen": 2762539008 + }, + { + "epoch": 0.84, + "learning_rate": 0.00013180341023069208, + "loss": 1.1785, + "theoretical_loss": 3.3705238513397764, + "tokens_seen": 2762604544 + }, + { + "epoch": 0.84, + "learning_rate": 0.00013178736208625878, + "loss": 1.3156, + "theoretical_loss": 3.370517653103788, + "tokens_seen": 2762670080 + }, + { + "epoch": 0.84, + "learning_rate": 0.00013177131394182548, + "loss": 1.3413, + "theoretical_loss": 3.3705114550560005, + "tokens_seen": 2762735616 + }, + { + "epoch": 0.84, + "learning_rate": 0.00013175526579739218, + "loss": 1.2614, + "theoretical_loss": 3.3705052571964043, + "tokens_seen": 2762801152 + }, + { + "epoch": 0.84, + "learning_rate": 0.00013173921765295888, + "loss": 1.2974, + "theoretical_loss": 3.370499059524989, + "tokens_seen": 2762866688 + }, + { + "epoch": 0.84, + "learning_rate": 0.00013172316950852558, + "loss": 1.3543, + "theoretical_loss": 3.3704928620417443, + "tokens_seen": 2762932224 + }, + { + "epoch": 0.84, + "learning_rate": 0.00013170712136409228, + "loss": 1.3827, + "theoretical_loss": 3.37048666474666, + "tokens_seen": 2762997760 + }, + { + "epoch": 0.84, + "learning_rate": 0.00013169107321965898, + "loss": 1.2788, + "theoretical_loss": 3.370480467639726, + "tokens_seen": 2763063296 + }, + { + "epoch": 0.84, + "learning_rate": 0.00013167502507522568, + "loss": 1.1742, + "theoretical_loss": 3.3704742707209325, + "tokens_seen": 2763128832 + }, + { + "epoch": 0.84, + "learning_rate": 0.00013165897693079239, + "loss": 1.2217, + "theoretical_loss": 3.370468073990269, + "tokens_seen": 2763194368 + }, + { + "epoch": 0.84, + "learning_rate": 0.00013164292878635909, + "loss": 1.1285, + "theoretical_loss": 3.370461877447725, + "tokens_seen": 2763259904 + }, + { + "epoch": 0.84, + "learning_rate": 0.00013162688064192579, + "loss": 1.252, + "theoretical_loss": 3.3704556810932904, + "tokens_seen": 2763325440 + }, + { + "epoch": 0.84, + "learning_rate": 0.0001316108324974925, + "loss": 1.2404, + "theoretical_loss": 3.3704494849269553, + "tokens_seen": 2763390976 + }, + { + "epoch": 0.84, + "learning_rate": 0.0001315947843530592, + "loss": 1.1956, + "theoretical_loss": 3.3704432889487097, + "tokens_seen": 2763456512 + }, + { + "epoch": 0.84, + "learning_rate": 0.0001315787362086259, + "loss": 1.3245, + "theoretical_loss": 3.3704370931585435, + "tokens_seen": 2763522048 + }, + { + "epoch": 0.84, + "learning_rate": 0.0001315626880641926, + "loss": 1.2514, + "theoretical_loss": 3.370430897556446, + "tokens_seen": 2763587584 + }, + { + "epoch": 0.84, + "learning_rate": 0.0001315466399197593, + "loss": 1.2692, + "theoretical_loss": 3.370424702142407, + "tokens_seen": 2763653120 + }, + { + "epoch": 0.84, + "learning_rate": 0.000131530591775326, + "loss": 1.2779, + "theoretical_loss": 3.3704185069164168, + "tokens_seen": 2763718656 + }, + { + "epoch": 0.84, + "learning_rate": 0.0001315145436308927, + "loss": 1.2643, + "theoretical_loss": 3.370412311878465, + "tokens_seen": 2763784192 + }, + { + "epoch": 0.84, + "learning_rate": 0.0001314984954864594, + "loss": 1.268, + "theoretical_loss": 3.3704061170285415, + "tokens_seen": 2763849728 + }, + { + "epoch": 0.84, + "learning_rate": 0.0001314824473420261, + "loss": 1.287, + "theoretical_loss": 3.370399922366636, + "tokens_seen": 2763915264 + }, + { + "epoch": 0.84, + "learning_rate": 0.0001314663991975928, + "loss": 1.2663, + "theoretical_loss": 3.3703937278927385, + "tokens_seen": 2763980800 + }, + { + "epoch": 0.84, + "learning_rate": 0.0001314503510531595, + "loss": 1.3377, + "theoretical_loss": 3.3703875336068387, + "tokens_seen": 2764046336 + }, + { + "epoch": 0.84, + "learning_rate": 0.0001314343029087262, + "loss": 1.2689, + "theoretical_loss": 3.370381339508927, + "tokens_seen": 2764111872 + }, + { + "epoch": 0.84, + "learning_rate": 0.00013141825476429287, + "loss": 1.1881, + "theoretical_loss": 3.370375145598992, + "tokens_seen": 2764177408 + }, + { + "epoch": 0.84, + "learning_rate": 0.00013140220661985957, + "loss": 1.4534, + "theoretical_loss": 3.370368951877025, + "tokens_seen": 2764242944 + }, + { + "epoch": 0.84, + "learning_rate": 0.00013138615847542627, + "loss": 1.2849, + "theoretical_loss": 3.3703627583430147, + "tokens_seen": 2764308480 + }, + { + "epoch": 0.84, + "learning_rate": 0.00013137011033099297, + "loss": 1.2899, + "theoretical_loss": 3.3703565649969516, + "tokens_seen": 2764374016 + }, + { + "epoch": 0.84, + "learning_rate": 0.00013135406218655967, + "loss": 1.3463, + "theoretical_loss": 3.370350371838825, + "tokens_seen": 2764439552 + }, + { + "epoch": 0.84, + "learning_rate": 0.00013133801404212637, + "loss": 1.2345, + "theoretical_loss": 3.370344178868625, + "tokens_seen": 2764505088 + }, + { + "epoch": 0.84, + "learning_rate": 0.00013132196589769307, + "loss": 1.3014, + "theoretical_loss": 3.370337986086342, + "tokens_seen": 2764570624 + }, + { + "epoch": 0.84, + "learning_rate": 0.00013130591775325977, + "loss": 1.2894, + "theoretical_loss": 3.3703317934919648, + "tokens_seen": 2764636160 + }, + { + "epoch": 0.84, + "learning_rate": 0.00013128986960882647, + "loss": 1.2318, + "theoretical_loss": 3.3703256010854843, + "tokens_seen": 2764701696 + }, + { + "epoch": 0.84, + "learning_rate": 0.00013127382146439317, + "loss": 1.2566, + "theoretical_loss": 3.3703194088668895, + "tokens_seen": 2764767232 + }, + { + "epoch": 0.84, + "learning_rate": 0.00013125777331995987, + "loss": 1.2395, + "theoretical_loss": 3.3703132168361707, + "tokens_seen": 2764832768 + }, + { + "epoch": 0.84, + "learning_rate": 0.00013124172517552657, + "loss": 1.2732, + "theoretical_loss": 3.3703070249933176, + "tokens_seen": 2764898304 + }, + { + "epoch": 0.84, + "learning_rate": 0.0001312256770310933, + "loss": 1.169, + "theoretical_loss": 3.3703008333383204, + "tokens_seen": 2764963840 + }, + { + "epoch": 0.84, + "learning_rate": 0.00013120962888666, + "loss": 1.1358, + "theoretical_loss": 3.370294641871168, + "tokens_seen": 2765029376 + }, + { + "epoch": 0.84, + "learning_rate": 0.0001311935807422267, + "loss": 1.2192, + "theoretical_loss": 3.3702884505918513, + "tokens_seen": 2765094912 + }, + { + "epoch": 0.84, + "learning_rate": 0.0001311775325977934, + "loss": 1.1979, + "theoretical_loss": 3.3702822595003594, + "tokens_seen": 2765160448 + }, + { + "epoch": 0.84, + "learning_rate": 0.0001311614844533601, + "loss": 1.2709, + "theoretical_loss": 3.3702760685966826, + "tokens_seen": 2765225984 + }, + { + "epoch": 0.84, + "learning_rate": 0.0001311454363089268, + "loss": 1.2883, + "theoretical_loss": 3.3702698778808107, + "tokens_seen": 2765291520 + }, + { + "epoch": 0.84, + "learning_rate": 0.0001311293881644935, + "loss": 1.3234, + "theoretical_loss": 3.3702636873527334, + "tokens_seen": 2765357056 + }, + { + "epoch": 0.84, + "learning_rate": 0.0001311133400200602, + "loss": 1.2119, + "theoretical_loss": 3.3702574970124406, + "tokens_seen": 2765422592 + }, + { + "epoch": 0.84, + "learning_rate": 0.00013109729187562688, + "loss": 1.3498, + "theoretical_loss": 3.370251306859922, + "tokens_seen": 2765488128 + }, + { + "epoch": 0.84, + "learning_rate": 0.00013108124373119358, + "loss": 1.2086, + "theoretical_loss": 3.370245116895168, + "tokens_seen": 2765553664 + }, + { + "epoch": 0.84, + "objective/train/docs_used": 1004597, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.151955246925354, + "objective/train/theoretical_loss": 3.370238927118168, + "objective/train/tokens_used": 2786079200, + "theoretical_loss": 3.370238927118168, + "tokens_seen": 2765619200 + }, + { + "epoch": 0.84, + "learning_rate": 0.00013106519558676028, + "loss": 1.2045, + "theoretical_loss": 3.370238927118168, + "tokens_seen": 2765619200 + }, + { + "epoch": 0.84, + "learning_rate": 0.00013104914744232698, + "loss": 1.2515, + "theoretical_loss": 3.370232737528912, + "tokens_seen": 2765684736 + }, + { + "epoch": 0.84, + "learning_rate": 0.00013103309929789368, + "loss": 1.2404, + "theoretical_loss": 3.3702265481273894, + "tokens_seen": 2765750272 + }, + { + "epoch": 0.84, + "learning_rate": 0.00013101705115346038, + "loss": 1.1925, + "theoretical_loss": 3.370220358913591, + "tokens_seen": 2765815808 + }, + { + "epoch": 0.84, + "learning_rate": 0.00013100100300902708, + "loss": 1.321, + "theoretical_loss": 3.3702141698875057, + "tokens_seen": 2765881344 + }, + { + "epoch": 0.84, + "learning_rate": 0.00013098495486459378, + "loss": 1.2838, + "theoretical_loss": 3.3702079810491243, + "tokens_seen": 2765946880 + }, + { + "epoch": 0.84, + "learning_rate": 0.00013096890672016049, + "loss": 1.3467, + "theoretical_loss": 3.3702017923984355, + "tokens_seen": 2766012416 + }, + { + "epoch": 0.84, + "learning_rate": 0.00013095285857572719, + "loss": 1.2865, + "theoretical_loss": 3.3701956039354304, + "tokens_seen": 2766077952 + }, + { + "epoch": 0.84, + "learning_rate": 0.0001309368104312939, + "loss": 1.2681, + "theoretical_loss": 3.3701894156600978, + "tokens_seen": 2766143488 + }, + { + "epoch": 0.84, + "learning_rate": 0.0001309207622868606, + "loss": 1.3504, + "theoretical_loss": 3.370183227572428, + "tokens_seen": 2766209024 + }, + { + "epoch": 0.84, + "learning_rate": 0.0001309047141424273, + "loss": 1.1991, + "theoretical_loss": 3.370177039672411, + "tokens_seen": 2766274560 + }, + { + "epoch": 0.84, + "learning_rate": 0.000130888665997994, + "loss": 1.3333, + "theoretical_loss": 3.3701708519600366, + "tokens_seen": 2766340096 + }, + { + "epoch": 0.84, + "learning_rate": 0.0001308726178535607, + "loss": 1.2761, + "theoretical_loss": 3.3701646644352943, + "tokens_seen": 2766405632 + }, + { + "epoch": 0.84, + "learning_rate": 0.0001308565697091274, + "loss": 1.222, + "theoretical_loss": 3.3701584770981747, + "tokens_seen": 2766471168 + }, + { + "epoch": 0.84, + "learning_rate": 0.0001308405215646941, + "loss": 1.3138, + "theoretical_loss": 3.370152289948667, + "tokens_seen": 2766536704 + }, + { + "epoch": 0.84, + "learning_rate": 0.0001308244734202608, + "loss": 1.2892, + "theoretical_loss": 3.3701461029867614, + "tokens_seen": 2766602240 + }, + { + "epoch": 0.84, + "learning_rate": 0.0001308084252758275, + "loss": 1.3717, + "theoretical_loss": 3.370139916212448, + "tokens_seen": 2766667776 + }, + { + "epoch": 0.84, + "learning_rate": 0.0001307923771313942, + "loss": 1.2786, + "theoretical_loss": 3.370133729625716, + "tokens_seen": 2766733312 + }, + { + "epoch": 0.84, + "learning_rate": 0.0001307763289869609, + "loss": 1.2663, + "theoretical_loss": 3.3701275432265554, + "tokens_seen": 2766798848 + }, + { + "epoch": 0.84, + "learning_rate": 0.0001307602808425276, + "loss": 1.2106, + "theoretical_loss": 3.370121357014957, + "tokens_seen": 2766864384 + }, + { + "epoch": 0.84, + "learning_rate": 0.0001307442326980943, + "loss": 1.254, + "theoretical_loss": 3.370115170990909, + "tokens_seen": 2766929920 + }, + { + "epoch": 0.84, + "learning_rate": 0.000130728184553661, + "loss": 1.2791, + "theoretical_loss": 3.370108985154403, + "tokens_seen": 2766995456 + }, + { + "epoch": 0.84, + "learning_rate": 0.00013071213640922767, + "loss": 1.3455, + "theoretical_loss": 3.370102799505428, + "tokens_seen": 2767060992 + }, + { + "epoch": 0.84, + "learning_rate": 0.00013069608826479437, + "loss": 1.2038, + "theoretical_loss": 3.3700966140439736, + "tokens_seen": 2767126528 + }, + { + "epoch": 0.84, + "learning_rate": 0.00013068004012036107, + "loss": 1.2122, + "theoretical_loss": 3.3700904287700304, + "tokens_seen": 2767192064 + }, + { + "epoch": 0.84, + "learning_rate": 0.00013066399197592777, + "loss": 1.3054, + "theoretical_loss": 3.370084243683588, + "tokens_seen": 2767257600 + }, + { + "epoch": 0.84, + "learning_rate": 0.00013064794383149447, + "loss": 1.1834, + "theoretical_loss": 3.3700780587846357, + "tokens_seen": 2767323136 + }, + { + "epoch": 0.84, + "learning_rate": 0.00013063189568706117, + "loss": 1.3278, + "theoretical_loss": 3.3700718740731643, + "tokens_seen": 2767388672 + }, + { + "epoch": 0.84, + "learning_rate": 0.0001306158475426279, + "loss": 1.2913, + "theoretical_loss": 3.370065689549163, + "tokens_seen": 2767454208 + }, + { + "epoch": 0.84, + "learning_rate": 0.0001305997993981946, + "loss": 1.2365, + "theoretical_loss": 3.3700595052126223, + "tokens_seen": 2767519744 + }, + { + "epoch": 0.84, + "learning_rate": 0.0001305837512537613, + "loss": 1.2603, + "theoretical_loss": 3.3700533210635313, + "tokens_seen": 2767585280 + }, + { + "epoch": 0.84, + "learning_rate": 0.000130567703109328, + "loss": 1.1379, + "theoretical_loss": 3.3700471371018805, + "tokens_seen": 2767650816 + }, + { + "epoch": 0.84, + "learning_rate": 0.0001305516549648947, + "loss": 1.2706, + "theoretical_loss": 3.3700409533276594, + "tokens_seen": 2767716352 + }, + { + "epoch": 0.84, + "learning_rate": 0.0001305356068204614, + "loss": 1.2771, + "theoretical_loss": 3.370034769740858, + "tokens_seen": 2767781888 + }, + { + "epoch": 0.84, + "learning_rate": 0.0001305195586760281, + "loss": 1.2439, + "theoretical_loss": 3.3700285863414665, + "tokens_seen": 2767847424 + }, + { + "epoch": 0.84, + "learning_rate": 0.0001305035105315948, + "loss": 1.2561, + "theoretical_loss": 3.3700224031294748, + "tokens_seen": 2767912960 + }, + { + "epoch": 0.84, + "learning_rate": 0.0001304874623871615, + "loss": 1.3953, + "theoretical_loss": 3.370016220104872, + "tokens_seen": 2767978496 + }, + { + "epoch": 0.84, + "learning_rate": 0.0001304714142427282, + "loss": 1.2788, + "theoretical_loss": 3.3700100372676487, + "tokens_seen": 2768044032 + }, + { + "epoch": 0.84, + "learning_rate": 0.0001304553660982949, + "loss": 1.2759, + "theoretical_loss": 3.3700038546177944, + "tokens_seen": 2768109568 + }, + { + "epoch": 0.84, + "learning_rate": 0.0001304393179538616, + "loss": 1.1884, + "theoretical_loss": 3.369997672155299, + "tokens_seen": 2768175104 + }, + { + "epoch": 0.84, + "learning_rate": 0.0001304232698094283, + "loss": 1.262, + "theoretical_loss": 3.369991489880153, + "tokens_seen": 2768240640 + }, + { + "epoch": 0.84, + "learning_rate": 0.000130407221664995, + "loss": 1.276, + "theoretical_loss": 3.369985307792346, + "tokens_seen": 2768306176 + }, + { + "epoch": 0.84, + "learning_rate": 0.00013039117352056168, + "loss": 1.1823, + "theoretical_loss": 3.3699791258918674, + "tokens_seen": 2768371712 + }, + { + "epoch": 0.84, + "learning_rate": 0.00013037512537612838, + "loss": 1.2956, + "theoretical_loss": 3.3699729441787074, + "tokens_seen": 2768437248 + }, + { + "epoch": 0.84, + "learning_rate": 0.00013035907723169508, + "loss": 1.2285, + "theoretical_loss": 3.3699667626528558, + "tokens_seen": 2768502784 + }, + { + "epoch": 0.84, + "learning_rate": 0.00013034302908726178, + "loss": 1.3183, + "theoretical_loss": 3.3699605813143023, + "tokens_seen": 2768568320 + }, + { + "epoch": 0.84, + "learning_rate": 0.00013032698094282848, + "loss": 1.31, + "theoretical_loss": 3.3699544001630377, + "tokens_seen": 2768633856 + }, + { + "epoch": 0.84, + "learning_rate": 0.00013031093279839518, + "loss": 1.1522, + "theoretical_loss": 3.369948219199051, + "tokens_seen": 2768699392 + }, + { + "epoch": 0.84, + "learning_rate": 0.00013029488465396188, + "loss": 1.2473, + "theoretical_loss": 3.3699420384223324, + "tokens_seen": 2768764928 + }, + { + "epoch": 0.84, + "learning_rate": 0.00013027883650952859, + "loss": 1.371, + "theoretical_loss": 3.3699358578328718, + "tokens_seen": 2768830464 + }, + { + "epoch": 0.84, + "objective/train/docs_used": 1005833, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.352116346359253, + "objective/train/theoretical_loss": 3.369929677430659, + "objective/train/tokens_used": 2789356000, + "theoretical_loss": 3.369929677430659, + "tokens_seen": 2768896000 + }, + { + "epoch": 0.84, + "learning_rate": 0.00013026278836509529, + "loss": 1.3049, + "theoretical_loss": 3.369929677430659, + "tokens_seen": 2768896000 + }, + { + "epoch": 0.84, + "learning_rate": 0.000130246740220662, + "loss": 1.3495, + "theoretical_loss": 3.369923497215684, + "tokens_seen": 2768961536 + }, + { + "epoch": 0.84, + "learning_rate": 0.0001302306920762287, + "loss": 1.2902, + "theoretical_loss": 3.3699173171879364, + "tokens_seen": 2769027072 + }, + { + "epoch": 0.84, + "learning_rate": 0.0001302146439317954, + "loss": 1.3715, + "theoretical_loss": 3.369911137347407, + "tokens_seen": 2769092608 + }, + { + "epoch": 0.84, + "learning_rate": 0.0001301985957873621, + "loss": 1.1908, + "theoretical_loss": 3.3699049576940845, + "tokens_seen": 2769158144 + }, + { + "epoch": 0.84, + "learning_rate": 0.0001301825476429288, + "loss": 1.3346, + "theoretical_loss": 3.3698987782279595, + "tokens_seen": 2769223680 + }, + { + "epoch": 0.84, + "learning_rate": 0.0001301664994984955, + "loss": 1.28, + "theoretical_loss": 3.369892598949021, + "tokens_seen": 2769289216 + }, + { + "epoch": 0.84, + "learning_rate": 0.0001301504513540622, + "loss": 1.2823, + "theoretical_loss": 3.3698864198572607, + "tokens_seen": 2769354752 + }, + { + "epoch": 0.84, + "learning_rate": 0.0001301344032096289, + "loss": 1.3077, + "theoretical_loss": 3.369880240952667, + "tokens_seen": 2769420288 + }, + { + "epoch": 0.84, + "learning_rate": 0.0001301183550651956, + "loss": 1.2901, + "theoretical_loss": 3.3698740622352306, + "tokens_seen": 2769485824 + }, + { + "epoch": 0.84, + "learning_rate": 0.0001301023069207623, + "loss": 1.4455, + "theoretical_loss": 3.3698678837049405, + "tokens_seen": 2769551360 + }, + { + "epoch": 0.84, + "learning_rate": 0.000130086258776329, + "loss": 1.3786, + "theoretical_loss": 3.3698617053617874, + "tokens_seen": 2769616896 + }, + { + "epoch": 0.84, + "learning_rate": 0.0001300702106318957, + "loss": 1.2559, + "theoretical_loss": 3.369855527205761, + "tokens_seen": 2769682432 + }, + { + "epoch": 0.84, + "learning_rate": 0.0001300541624874624, + "loss": 1.3035, + "theoretical_loss": 3.3698493492368513, + "tokens_seen": 2769747968 + }, + { + "epoch": 0.84, + "learning_rate": 0.0001300381143430291, + "loss": 1.3269, + "theoretical_loss": 3.369843171455048, + "tokens_seen": 2769813504 + }, + { + "epoch": 0.84, + "learning_rate": 0.0001300220661985958, + "loss": 1.1924, + "theoretical_loss": 3.369836993860341, + "tokens_seen": 2769879040 + }, + { + "epoch": 0.84, + "learning_rate": 0.0001300060180541625, + "loss": 1.2742, + "theoretical_loss": 3.36983081645272, + "tokens_seen": 2769944576 + }, + { + "epoch": 0.84, + "learning_rate": 0.0001299899699097292, + "loss": 1.2625, + "theoretical_loss": 3.369824639232175, + "tokens_seen": 2770010112 + }, + { + "epoch": 0.84, + "learning_rate": 0.0001299739217652959, + "loss": 1.3228, + "theoretical_loss": 3.3698184621986966, + "tokens_seen": 2770075648 + }, + { + "epoch": 0.84, + "learning_rate": 0.0001299578736208626, + "loss": 1.3639, + "theoretical_loss": 3.3698122853522743, + "tokens_seen": 2770141184 + }, + { + "epoch": 0.84, + "learning_rate": 0.0001299418254764293, + "loss": 1.3834, + "theoretical_loss": 3.369806108692898, + "tokens_seen": 2770206720 + }, + { + "epoch": 0.84, + "learning_rate": 0.000129925777331996, + "loss": 1.226, + "theoretical_loss": 3.369799932220557, + "tokens_seen": 2770272256 + }, + { + "epoch": 0.84, + "learning_rate": 0.0001299097291875627, + "loss": 1.2356, + "theoretical_loss": 3.369793755935242, + "tokens_seen": 2770337792 + }, + { + "epoch": 0.84, + "learning_rate": 0.0001298936810431294, + "loss": 1.3115, + "theoretical_loss": 3.3697875798369425, + "tokens_seen": 2770403328 + }, + { + "epoch": 0.84, + "learning_rate": 0.0001298776328986961, + "loss": 1.1893, + "theoretical_loss": 3.3697814039256486, + "tokens_seen": 2770468864 + }, + { + "epoch": 0.84, + "learning_rate": 0.0001298615847542628, + "loss": 1.2489, + "theoretical_loss": 3.3697752282013504, + "tokens_seen": 2770534400 + }, + { + "epoch": 0.84, + "learning_rate": 0.0001298455366098295, + "loss": 1.3426, + "theoretical_loss": 3.3697690526640374, + "tokens_seen": 2770599936 + }, + { + "epoch": 0.84, + "learning_rate": 0.0001298294884653962, + "loss": 1.4225, + "theoretical_loss": 3.3697628773136996, + "tokens_seen": 2770665472 + }, + { + "epoch": 0.84, + "learning_rate": 0.0001298134403209629, + "loss": 1.1908, + "theoretical_loss": 3.3697567021503274, + "tokens_seen": 2770731008 + }, + { + "epoch": 0.84, + "learning_rate": 0.0001297973921765296, + "loss": 1.3128, + "theoretical_loss": 3.36975052717391, + "tokens_seen": 2770796544 + }, + { + "epoch": 0.84, + "learning_rate": 0.0001297813440320963, + "loss": 1.2274, + "theoretical_loss": 3.369744352384438, + "tokens_seen": 2770862080 + }, + { + "epoch": 0.84, + "learning_rate": 0.000129765295887663, + "loss": 1.3205, + "theoretical_loss": 3.3697381777819007, + "tokens_seen": 2770927616 + }, + { + "epoch": 0.84, + "learning_rate": 0.0001297492477432297, + "loss": 1.2132, + "theoretical_loss": 3.3697320033662885, + "tokens_seen": 2770993152 + }, + { + "epoch": 0.84, + "learning_rate": 0.0001297331995987964, + "loss": 1.1954, + "theoretical_loss": 3.369725829137591, + "tokens_seen": 2771058688 + }, + { + "epoch": 0.84, + "learning_rate": 0.0001297171514543631, + "loss": 1.2088, + "theoretical_loss": 3.369719655095798, + "tokens_seen": 2771124224 + }, + { + "epoch": 0.84, + "learning_rate": 0.00012970110330992978, + "loss": 1.3133, + "theoretical_loss": 3.3697134812409, + "tokens_seen": 2771189760 + }, + { + "epoch": 0.84, + "learning_rate": 0.00012968505516549648, + "loss": 1.2519, + "theoretical_loss": 3.3697073075728863, + "tokens_seen": 2771255296 + }, + { + "epoch": 0.84, + "learning_rate": 0.00012966900702106318, + "loss": 1.209, + "theoretical_loss": 3.3697011340917475, + "tokens_seen": 2771320832 + }, + { + "epoch": 0.84, + "learning_rate": 0.00012965295887662988, + "loss": 1.2861, + "theoretical_loss": 3.3696949607974727, + "tokens_seen": 2771386368 + }, + { + "epoch": 0.84, + "learning_rate": 0.00012963691073219658, + "loss": 1.2963, + "theoretical_loss": 3.3696887876900528, + "tokens_seen": 2771451904 + }, + { + "epoch": 0.84, + "learning_rate": 0.00012962086258776328, + "loss": 1.3311, + "theoretical_loss": 3.3696826147694767, + "tokens_seen": 2771517440 + }, + { + "epoch": 0.84, + "learning_rate": 0.00012960481444332998, + "loss": 1.349, + "theoretical_loss": 3.369676442035735, + "tokens_seen": 2771582976 + }, + { + "epoch": 0.84, + "learning_rate": 0.00012958876629889669, + "loss": 1.1747, + "theoretical_loss": 3.369670269488817, + "tokens_seen": 2771648512 + }, + { + "epoch": 0.84, + "learning_rate": 0.00012957271815446339, + "loss": 1.2622, + "theoretical_loss": 3.3696640971287137, + "tokens_seen": 2771714048 + }, + { + "epoch": 0.84, + "learning_rate": 0.0001295566700100301, + "loss": 1.2589, + "theoretical_loss": 3.369657924955414, + "tokens_seen": 2771779584 + }, + { + "epoch": 0.84, + "learning_rate": 0.0001295406218655968, + "loss": 1.2172, + "theoretical_loss": 3.3696517529689087, + "tokens_seen": 2771845120 + }, + { + "epoch": 0.84, + "learning_rate": 0.0001295245737211635, + "loss": 1.2953, + "theoretical_loss": 3.3696455811691868, + "tokens_seen": 2771910656 + }, + { + "epoch": 0.84, + "learning_rate": 0.0001295085255767302, + "loss": 1.4688, + "theoretical_loss": 3.369639409556239, + "tokens_seen": 2771976192 + }, + { + "epoch": 0.84, + "learning_rate": 0.0001294924774322969, + "loss": 1.246, + "theoretical_loss": 3.3696332381300547, + "tokens_seen": 2772041728 + }, + { + "epoch": 0.84, + "learning_rate": 0.0001294764292878636, + "loss": 1.3557, + "theoretical_loss": 3.369627066890624, + "tokens_seen": 2772107264 + }, + { + "epoch": 0.84, + "objective/train/docs_used": 1006627, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.2924408912658691, + "objective/train/theoretical_loss": 3.369620895837937, + "objective/train/tokens_used": 2792632800, + "theoretical_loss": 3.369620895837937, + "tokens_seen": 2772172800 + }, + { + "epoch": 0.84, + "learning_rate": 0.0001294603811434303, + "loss": 1.2842, + "theoretical_loss": 3.369620895837937, + "tokens_seen": 2772172800 + }, + { + "epoch": 0.84, + "learning_rate": 0.000129444332998997, + "loss": 1.2392, + "theoretical_loss": 3.3696147249719837, + "tokens_seen": 2772238336 + }, + { + "epoch": 0.84, + "learning_rate": 0.00012942828485456372, + "loss": 1.234, + "theoretical_loss": 3.3696085542927534, + "tokens_seen": 2772303872 + }, + { + "epoch": 0.84, + "learning_rate": 0.00012941223671013042, + "loss": 1.2616, + "theoretical_loss": 3.369602383800237, + "tokens_seen": 2772369408 + }, + { + "epoch": 0.84, + "learning_rate": 0.00012939618856569712, + "loss": 1.2713, + "theoretical_loss": 3.3695962134944235, + "tokens_seen": 2772434944 + }, + { + "epoch": 0.84, + "learning_rate": 0.0001293801404212638, + "loss": 1.2304, + "theoretical_loss": 3.3695900433753034, + "tokens_seen": 2772500480 + }, + { + "epoch": 0.84, + "learning_rate": 0.0001293640922768305, + "loss": 1.2204, + "theoretical_loss": 3.369583873442867, + "tokens_seen": 2772566016 + }, + { + "epoch": 0.84, + "learning_rate": 0.0001293480441323972, + "loss": 1.235, + "theoretical_loss": 3.369577703697103, + "tokens_seen": 2772631552 + }, + { + "epoch": 0.84, + "learning_rate": 0.0001293319959879639, + "loss": 1.1814, + "theoretical_loss": 3.3695715341380024, + "tokens_seen": 2772697088 + }, + { + "epoch": 0.84, + "learning_rate": 0.0001293159478435306, + "loss": 1.2112, + "theoretical_loss": 3.369565364765555, + "tokens_seen": 2772762624 + }, + { + "epoch": 0.84, + "learning_rate": 0.0001292998996990973, + "loss": 1.2058, + "theoretical_loss": 3.3695591955797504, + "tokens_seen": 2772828160 + }, + { + "epoch": 0.84, + "learning_rate": 0.000129283851554664, + "loss": 1.3612, + "theoretical_loss": 3.3695530265805784, + "tokens_seen": 2772893696 + }, + { + "epoch": 0.84, + "learning_rate": 0.0001292678034102307, + "loss": 1.2855, + "theoretical_loss": 3.3695468577680296, + "tokens_seen": 2772959232 + }, + { + "epoch": 0.84, + "learning_rate": 0.0001292517552657974, + "loss": 1.2547, + "theoretical_loss": 3.3695406891420934, + "tokens_seen": 2773024768 + }, + { + "epoch": 0.84, + "learning_rate": 0.0001292357071213641, + "loss": 1.2451, + "theoretical_loss": 3.3695345207027603, + "tokens_seen": 2773090304 + }, + { + "epoch": 0.84, + "learning_rate": 0.0001292196589769308, + "loss": 1.309, + "theoretical_loss": 3.3695283524500197, + "tokens_seen": 2773155840 + }, + { + "epoch": 0.84, + "learning_rate": 0.0001292036108324975, + "loss": 1.2977, + "theoretical_loss": 3.369522184383862, + "tokens_seen": 2773221376 + }, + { + "epoch": 0.84, + "learning_rate": 0.0001291875626880642, + "loss": 1.2888, + "theoretical_loss": 3.369516016504276, + "tokens_seen": 2773286912 + }, + { + "epoch": 0.84, + "learning_rate": 0.0001291715145436309, + "loss": 1.3169, + "theoretical_loss": 3.369509848811253, + "tokens_seen": 2773352448 + }, + { + "epoch": 0.84, + "learning_rate": 0.0001291554663991976, + "loss": 1.2538, + "theoretical_loss": 3.3695036813047827, + "tokens_seen": 2773417984 + }, + { + "epoch": 0.84, + "learning_rate": 0.0001291394182547643, + "loss": 1.3029, + "theoretical_loss": 3.3694975139848546, + "tokens_seen": 2773483520 + }, + { + "epoch": 0.84, + "learning_rate": 0.000129123370110331, + "loss": 1.2326, + "theoretical_loss": 3.3694913468514596, + "tokens_seen": 2773549056 + }, + { + "epoch": 0.84, + "learning_rate": 0.0001291073219658977, + "loss": 1.2675, + "theoretical_loss": 3.369485179904586, + "tokens_seen": 2773614592 + }, + { + "epoch": 0.84, + "learning_rate": 0.0001290912738214644, + "loss": 1.3218, + "theoretical_loss": 3.369479013144225, + "tokens_seen": 2773680128 + }, + { + "epoch": 0.84, + "learning_rate": 0.0001290752256770311, + "loss": 1.1243, + "theoretical_loss": 3.3694728465703663, + "tokens_seen": 2773745664 + }, + { + "epoch": 0.84, + "learning_rate": 0.0001290591775325978, + "loss": 1.3207, + "theoretical_loss": 3.3694666801829998, + "tokens_seen": 2773811200 + }, + { + "epoch": 0.84, + "learning_rate": 0.0001290431293881645, + "loss": 1.3116, + "theoretical_loss": 3.369460513982115, + "tokens_seen": 2773876736 + }, + { + "epoch": 0.84, + "learning_rate": 0.0001290270812437312, + "loss": 1.3448, + "theoretical_loss": 3.369454347967703, + "tokens_seen": 2773942272 + }, + { + "epoch": 0.84, + "learning_rate": 0.0001290110330992979, + "loss": 1.2983, + "theoretical_loss": 3.3694481821397524, + "tokens_seen": 2774007808 + }, + { + "epoch": 0.84, + "learning_rate": 0.00012899498495486458, + "loss": 1.2145, + "theoretical_loss": 3.369442016498254, + "tokens_seen": 2774073344 + }, + { + "epoch": 0.84, + "learning_rate": 0.00012897893681043128, + "loss": 1.3046, + "theoretical_loss": 3.369435851043198, + "tokens_seen": 2774138880 + }, + { + "epoch": 0.84, + "learning_rate": 0.00012896288866599798, + "loss": 1.2396, + "theoretical_loss": 3.3694296857745734, + "tokens_seen": 2774204416 + }, + { + "epoch": 0.84, + "learning_rate": 0.00012894684052156468, + "loss": 1.3094, + "theoretical_loss": 3.369423520692371, + "tokens_seen": 2774269952 + }, + { + "epoch": 0.84, + "learning_rate": 0.00012893079237713138, + "loss": 1.1917, + "theoretical_loss": 3.36941735579658, + "tokens_seen": 2774335488 + }, + { + "epoch": 0.84, + "learning_rate": 0.00012891474423269809, + "loss": 1.213, + "theoretical_loss": 3.369411191087191, + "tokens_seen": 2774401024 + }, + { + "epoch": 0.84, + "learning_rate": 0.00012889869608826479, + "loss": 1.2676, + "theoretical_loss": 3.369405026564194, + "tokens_seen": 2774466560 + }, + { + "epoch": 0.84, + "learning_rate": 0.00012888264794383149, + "loss": 1.2463, + "theoretical_loss": 3.3693988622275786, + "tokens_seen": 2774532096 + }, + { + "epoch": 0.84, + "learning_rate": 0.0001288665997993982, + "loss": 1.2333, + "theoretical_loss": 3.369392698077335, + "tokens_seen": 2774597632 + }, + { + "epoch": 0.84, + "learning_rate": 0.0001288505516549649, + "loss": 1.2326, + "theoretical_loss": 3.3693865341134526, + "tokens_seen": 2774663168 + }, + { + "epoch": 0.84, + "learning_rate": 0.0001288345035105316, + "loss": 1.2592, + "theoretical_loss": 3.3693803703359224, + "tokens_seen": 2774728704 + }, + { + "epoch": 0.84, + "learning_rate": 0.00012881845536609832, + "loss": 1.1817, + "theoretical_loss": 3.369374206744733, + "tokens_seen": 2774794240 + }, + { + "epoch": 0.84, + "learning_rate": 0.00012880240722166502, + "loss": 1.3334, + "theoretical_loss": 3.369368043339876, + "tokens_seen": 2774859776 + }, + { + "epoch": 0.84, + "learning_rate": 0.00012878635907723172, + "loss": 1.1755, + "theoretical_loss": 3.36936188012134, + "tokens_seen": 2774925312 + }, + { + "epoch": 0.84, + "learning_rate": 0.00012877031093279842, + "loss": 1.2464, + "theoretical_loss": 3.3693557170891157, + "tokens_seen": 2774990848 + }, + { + "epoch": 0.84, + "learning_rate": 0.00012875426278836512, + "loss": 1.2448, + "theoretical_loss": 3.3693495542431924, + "tokens_seen": 2775056384 + }, + { + "epoch": 0.84, + "learning_rate": 0.00012873821464393182, + "loss": 1.2432, + "theoretical_loss": 3.369343391583561, + "tokens_seen": 2775121920 + }, + { + "epoch": 0.84, + "learning_rate": 0.00012872216649949852, + "loss": 1.2231, + "theoretical_loss": 3.369337229110211, + "tokens_seen": 2775187456 + }, + { + "epoch": 0.84, + "learning_rate": 0.00012870611835506522, + "loss": 1.2347, + "theoretical_loss": 3.369331066823132, + "tokens_seen": 2775252992 + }, + { + "epoch": 0.84, + "learning_rate": 0.00012869007021063192, + "loss": 1.2901, + "theoretical_loss": 3.3693249047223146, + "tokens_seen": 2775318528 + }, + { + "epoch": 0.84, + "learning_rate": 0.0001286740220661986, + "loss": 1.4282, + "theoretical_loss": 3.3693187428077485, + "tokens_seen": 2775384064 + }, + { + "epoch": 0.84, + "objective/train/docs_used": 1007859, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.239830493927002, + "objective/train/theoretical_loss": 3.3693125810794236, + "objective/train/tokens_used": 2795909600, + "theoretical_loss": 3.3693125810794236, + "tokens_seen": 2775449600 + }, + { + "epoch": 0.84, + "learning_rate": 0.0001286579739217653, + "loss": 1.2442, + "theoretical_loss": 3.3693125810794236, + "tokens_seen": 2775449600 + }, + { + "epoch": 0.84, + "learning_rate": 0.000128641925777332, + "loss": 1.3016, + "theoretical_loss": 3.3693064195373297, + "tokens_seen": 2775515136 + }, + { + "epoch": 0.84, + "learning_rate": 0.0001286258776328987, + "loss": 1.2011, + "theoretical_loss": 3.3693002581814575, + "tokens_seen": 2775580672 + }, + { + "epoch": 0.84, + "learning_rate": 0.0001286098294884654, + "loss": 1.2327, + "theoretical_loss": 3.369294097011796, + "tokens_seen": 2775646208 + }, + { + "epoch": 0.84, + "learning_rate": 0.0001285937813440321, + "loss": 1.275, + "theoretical_loss": 3.3692879360283357, + "tokens_seen": 2775711744 + }, + { + "epoch": 0.84, + "learning_rate": 0.0001285777331995988, + "loss": 1.2067, + "theoretical_loss": 3.369281775231067, + "tokens_seen": 2775777280 + }, + { + "epoch": 0.84, + "learning_rate": 0.0001285616850551655, + "loss": 1.331, + "theoretical_loss": 3.369275614619979, + "tokens_seen": 2775842816 + }, + { + "epoch": 0.84, + "learning_rate": 0.0001285456369107322, + "loss": 1.2526, + "theoretical_loss": 3.3692694541950625, + "tokens_seen": 2775908352 + }, + { + "epoch": 0.84, + "learning_rate": 0.0001285295887662989, + "loss": 1.3159, + "theoretical_loss": 3.3692632939563065, + "tokens_seen": 2775973888 + }, + { + "epoch": 0.84, + "learning_rate": 0.0001285135406218656, + "loss": 1.1552, + "theoretical_loss": 3.369257133903702, + "tokens_seen": 2776039424 + }, + { + "epoch": 0.84, + "learning_rate": 0.0001284974924774323, + "loss": 1.3997, + "theoretical_loss": 3.3692509740372385, + "tokens_seen": 2776104960 + }, + { + "epoch": 0.84, + "learning_rate": 0.000128481444332999, + "loss": 1.2198, + "theoretical_loss": 3.369244814356906, + "tokens_seen": 2776170496 + }, + { + "epoch": 0.84, + "learning_rate": 0.0001284653961885657, + "loss": 1.2736, + "theoretical_loss": 3.3692386548626945, + "tokens_seen": 2776236032 + }, + { + "epoch": 0.84, + "learning_rate": 0.0001284493480441324, + "loss": 1.2764, + "theoretical_loss": 3.369232495554594, + "tokens_seen": 2776301568 + }, + { + "epoch": 0.84, + "learning_rate": 0.0001284332998996991, + "loss": 1.245, + "theoretical_loss": 3.369226336432594, + "tokens_seen": 2776367104 + }, + { + "epoch": 0.84, + "learning_rate": 0.0001284172517552658, + "loss": 1.1906, + "theoretical_loss": 3.3692201774966857, + "tokens_seen": 2776432640 + }, + { + "epoch": 0.84, + "learning_rate": 0.0001284012036108325, + "loss": 1.1993, + "theoretical_loss": 3.3692140187468578, + "tokens_seen": 2776498176 + }, + { + "epoch": 0.84, + "learning_rate": 0.0001283851554663992, + "loss": 1.2985, + "theoretical_loss": 3.369207860183101, + "tokens_seen": 2776563712 + }, + { + "epoch": 0.84, + "learning_rate": 0.0001283691073219659, + "loss": 1.2725, + "theoretical_loss": 3.3692017018054052, + "tokens_seen": 2776629248 + }, + { + "epoch": 0.84, + "learning_rate": 0.0001283530591775326, + "loss": 1.2681, + "theoretical_loss": 3.36919554361376, + "tokens_seen": 2776694784 + }, + { + "epoch": 0.84, + "learning_rate": 0.0001283370110330993, + "loss": 1.2623, + "theoretical_loss": 3.369189385608156, + "tokens_seen": 2776760320 + }, + { + "epoch": 0.84, + "learning_rate": 0.000128320962888666, + "loss": 1.2896, + "theoretical_loss": 3.369183227788583, + "tokens_seen": 2776825856 + }, + { + "epoch": 0.84, + "learning_rate": 0.0001283049147442327, + "loss": 1.2034, + "theoretical_loss": 3.3691770701550303, + "tokens_seen": 2776891392 + }, + { + "epoch": 0.84, + "learning_rate": 0.00012828886659979938, + "loss": 1.3406, + "theoretical_loss": 3.369170912707489, + "tokens_seen": 2776956928 + }, + { + "epoch": 0.84, + "learning_rate": 0.00012827281845536608, + "loss": 1.3398, + "theoretical_loss": 3.369164755445948, + "tokens_seen": 2777022464 + }, + { + "epoch": 0.84, + "learning_rate": 0.00012825677031093278, + "loss": 1.2634, + "theoretical_loss": 3.369158598370398, + "tokens_seen": 2777088000 + }, + { + "epoch": 0.84, + "learning_rate": 0.00012824072216649948, + "loss": 1.3537, + "theoretical_loss": 3.3691524414808294, + "tokens_seen": 2777153536 + }, + { + "epoch": 0.84, + "learning_rate": 0.0001282246740220662, + "loss": 1.3276, + "theoretical_loss": 3.3691462847772313, + "tokens_seen": 2777219072 + }, + { + "epoch": 0.84, + "learning_rate": 0.0001282086258776329, + "loss": 1.3532, + "theoretical_loss": 3.369140128259594, + "tokens_seen": 2777284608 + }, + { + "epoch": 0.84, + "learning_rate": 0.00012819257773319961, + "loss": 1.2199, + "theoretical_loss": 3.3691339719279076, + "tokens_seen": 2777350144 + }, + { + "epoch": 0.84, + "learning_rate": 0.00012817652958876631, + "loss": 1.2285, + "theoretical_loss": 3.3691278157821616, + "tokens_seen": 2777415680 + }, + { + "epoch": 0.84, + "learning_rate": 0.00012816048144433301, + "loss": 1.2113, + "theoretical_loss": 3.369121659822347, + "tokens_seen": 2777481216 + }, + { + "epoch": 0.84, + "learning_rate": 0.00012814443329989972, + "loss": 1.2324, + "theoretical_loss": 3.3691155040484526, + "tokens_seen": 2777546752 + }, + { + "epoch": 0.84, + "learning_rate": 0.00012812838515546642, + "loss": 1.3, + "theoretical_loss": 3.3691093484604693, + "tokens_seen": 2777612288 + }, + { + "epoch": 0.84, + "learning_rate": 0.00012811233701103312, + "loss": 1.2604, + "theoretical_loss": 3.3691031930583866, + "tokens_seen": 2777677824 + }, + { + "epoch": 0.84, + "learning_rate": 0.00012809628886659982, + "loss": 1.2398, + "theoretical_loss": 3.369097037842195, + "tokens_seen": 2777743360 + }, + { + "epoch": 0.84, + "learning_rate": 0.00012808024072216652, + "loss": 1.2783, + "theoretical_loss": 3.369090882811884, + "tokens_seen": 2777808896 + }, + { + "epoch": 0.84, + "learning_rate": 0.00012806419257773322, + "loss": 1.1899, + "theoretical_loss": 3.369084727967444, + "tokens_seen": 2777874432 + }, + { + "epoch": 0.84, + "learning_rate": 0.00012804814443329992, + "loss": 1.2958, + "theoretical_loss": 3.3690785733088644, + "tokens_seen": 2777939968 + }, + { + "epoch": 0.84, + "learning_rate": 0.00012803209628886662, + "loss": 1.3087, + "theoretical_loss": 3.369072418836136, + "tokens_seen": 2778005504 + }, + { + "epoch": 0.84, + "learning_rate": 0.00012801604814443332, + "loss": 1.2598, + "theoretical_loss": 3.3690662645492484, + "tokens_seen": 2778071040 + }, + { + "epoch": 0.84, + "learning_rate": 0.00012800000000000002, + "loss": 1.2689, + "theoretical_loss": 3.3690601104481916, + "tokens_seen": 2778136576 + }, + { + "epoch": 0.84, + "learning_rate": 0.0001279839518555667, + "loss": 1.222, + "theoretical_loss": 3.3690539565329556, + "tokens_seen": 2778202112 + }, + { + "epoch": 0.84, + "learning_rate": 0.0001279679037111334, + "loss": 1.2193, + "theoretical_loss": 3.3690478028035304, + "tokens_seen": 2778267648 + }, + { + "epoch": 0.84, + "learning_rate": 0.0001279518555667001, + "loss": 1.3453, + "theoretical_loss": 3.3690416492599056, + "tokens_seen": 2778333184 + }, + { + "epoch": 0.84, + "learning_rate": 0.0001279358074222668, + "loss": 1.2062, + "theoretical_loss": 3.369035495902072, + "tokens_seen": 2778398720 + }, + { + "epoch": 0.84, + "learning_rate": 0.0001279197592778335, + "loss": 1.2526, + "theoretical_loss": 3.3690293427300197, + "tokens_seen": 2778464256 + }, + { + "epoch": 0.84, + "learning_rate": 0.0001279037111334002, + "loss": 1.3052, + "theoretical_loss": 3.3690231897437375, + "tokens_seen": 2778529792 + }, + { + "epoch": 0.84, + "learning_rate": 0.0001278876629889669, + "loss": 1.3816, + "theoretical_loss": 3.369017036943217, + "tokens_seen": 2778595328 + }, + { + "epoch": 0.84, + "learning_rate": 0.0001278716148445336, + "loss": 1.3476, + "theoretical_loss": 3.3690108843284468, + "tokens_seen": 2778660864 + }, + { + "epoch": 0.84, + "objective/train/docs_used": 1009547, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.2582721710205078, + "objective/train/theoretical_loss": 3.3690047318994174, + "objective/train/tokens_used": 2799186400, + "theoretical_loss": 3.3690047318994174, + "tokens_seen": 2778726400 + }, + { + "epoch": 0.84, + "learning_rate": 0.0001278555667001003, + "loss": 1.2267, + "theoretical_loss": 3.3690047318994174, + "tokens_seen": 2778726400 + }, + { + "epoch": 0.84, + "learning_rate": 0.000127839518555667, + "loss": 1.2957, + "theoretical_loss": 3.368998579656119, + "tokens_seen": 2778791936 + }, + { + "epoch": 0.84, + "learning_rate": 0.0001278234704112337, + "loss": 1.2369, + "theoretical_loss": 3.368992427598542, + "tokens_seen": 2778857472 + }, + { + "epoch": 0.84, + "learning_rate": 0.0001278074222668004, + "loss": 1.3388, + "theoretical_loss": 3.3689862757266753, + "tokens_seen": 2778923008 + }, + { + "epoch": 0.84, + "learning_rate": 0.0001277913741223671, + "loss": 1.3416, + "theoretical_loss": 3.36898012404051, + "tokens_seen": 2778988544 + }, + { + "epoch": 0.84, + "learning_rate": 0.0001277753259779338, + "loss": 1.2189, + "theoretical_loss": 3.368973972540035, + "tokens_seen": 2779054080 + }, + { + "epoch": 0.84, + "learning_rate": 0.0001277592778335005, + "loss": 1.23, + "theoretical_loss": 3.3689678212252416, + "tokens_seen": 2779119616 + }, + { + "epoch": 0.84, + "learning_rate": 0.0001277432296890672, + "loss": 1.223, + "theoretical_loss": 3.368961670096119, + "tokens_seen": 2779185152 + }, + { + "epoch": 0.84, + "learning_rate": 0.0001277271815446339, + "loss": 1.115, + "theoretical_loss": 3.3689555191526575, + "tokens_seen": 2779250688 + }, + { + "epoch": 0.84, + "learning_rate": 0.0001277111334002006, + "loss": 1.2655, + "theoretical_loss": 3.368949368394847, + "tokens_seen": 2779316224 + }, + { + "epoch": 0.84, + "learning_rate": 0.0001276950852557673, + "loss": 1.3202, + "theoretical_loss": 3.3689432178226775, + "tokens_seen": 2779381760 + }, + { + "epoch": 0.84, + "learning_rate": 0.000127679037111334, + "loss": 1.2334, + "theoretical_loss": 3.368937067436139, + "tokens_seen": 2779447296 + }, + { + "epoch": 0.84, + "learning_rate": 0.0001276629889669007, + "loss": 1.3604, + "theoretical_loss": 3.3689309172352218, + "tokens_seen": 2779512832 + }, + { + "epoch": 0.84, + "learning_rate": 0.0001276469408224674, + "loss": 1.2956, + "theoretical_loss": 3.3689247672199154, + "tokens_seen": 2779578368 + }, + { + "epoch": 0.84, + "learning_rate": 0.0001276308926780341, + "loss": 1.2986, + "theoretical_loss": 3.3689186173902104, + "tokens_seen": 2779643904 + }, + { + "epoch": 0.84, + "learning_rate": 0.0001276148445336008, + "loss": 1.1591, + "theoretical_loss": 3.3689124677460964, + "tokens_seen": 2779709440 + }, + { + "epoch": 0.84, + "learning_rate": 0.0001275987963891675, + "loss": 1.2265, + "theoretical_loss": 3.3689063182875634, + "tokens_seen": 2779774976 + }, + { + "epoch": 0.84, + "learning_rate": 0.0001275827482447342, + "loss": 1.1818, + "theoretical_loss": 3.368900169014602, + "tokens_seen": 2779840512 + }, + { + "epoch": 0.84, + "learning_rate": 0.0001275667001003009, + "loss": 1.2815, + "theoretical_loss": 3.3688940199272017, + "tokens_seen": 2779906048 + }, + { + "epoch": 0.84, + "learning_rate": 0.0001275506519558676, + "loss": 1.1837, + "theoretical_loss": 3.3688878710253523, + "tokens_seen": 2779971584 + }, + { + "epoch": 0.84, + "learning_rate": 0.0001275346038114343, + "loss": 1.2152, + "theoretical_loss": 3.3688817223090446, + "tokens_seen": 2780037120 + }, + { + "epoch": 0.84, + "learning_rate": 0.000127518555667001, + "loss": 1.2841, + "theoretical_loss": 3.3688755737782685, + "tokens_seen": 2780102656 + }, + { + "epoch": 0.84, + "learning_rate": 0.00012750250752256771, + "loss": 1.2535, + "theoretical_loss": 3.368869425433013, + "tokens_seen": 2780168192 + }, + { + "epoch": 0.84, + "learning_rate": 0.00012748645937813441, + "loss": 1.3277, + "theoretical_loss": 3.3688632772732694, + "tokens_seen": 2780233728 + }, + { + "epoch": 0.84, + "learning_rate": 0.00012747041123370112, + "loss": 1.274, + "theoretical_loss": 3.368857129299027, + "tokens_seen": 2780299264 + }, + { + "epoch": 0.84, + "learning_rate": 0.00012745436308926782, + "loss": 1.2503, + "theoretical_loss": 3.368850981510276, + "tokens_seen": 2780364800 + }, + { + "epoch": 0.84, + "learning_rate": 0.00012743831494483452, + "loss": 1.326, + "theoretical_loss": 3.3688448339070063, + "tokens_seen": 2780430336 + }, + { + "epoch": 0.84, + "learning_rate": 0.00012742226680040122, + "loss": 1.3548, + "theoretical_loss": 3.3688386864892084, + "tokens_seen": 2780495872 + }, + { + "epoch": 0.84, + "learning_rate": 0.00012740621865596792, + "loss": 1.2249, + "theoretical_loss": 3.368832539256872, + "tokens_seen": 2780561408 + }, + { + "epoch": 0.84, + "learning_rate": 0.00012739017051153462, + "loss": 1.2, + "theoretical_loss": 3.368826392209987, + "tokens_seen": 2780626944 + }, + { + "epoch": 0.84, + "learning_rate": 0.00012737412236710132, + "loss": 1.2614, + "theoretical_loss": 3.3688202453485436, + "tokens_seen": 2780692480 + }, + { + "epoch": 0.84, + "learning_rate": 0.00012735807422266802, + "loss": 1.3112, + "theoretical_loss": 3.368814098672532, + "tokens_seen": 2780758016 + }, + { + "epoch": 0.84, + "learning_rate": 0.00012734202607823472, + "loss": 1.2893, + "theoretical_loss": 3.3688079521819416, + "tokens_seen": 2780823552 + }, + { + "epoch": 0.84, + "learning_rate": 0.00012732597793380142, + "loss": 1.3872, + "theoretical_loss": 3.3688018058767635, + "tokens_seen": 2780889088 + }, + { + "epoch": 0.84, + "learning_rate": 0.00012730992978936812, + "loss": 1.2317, + "theoretical_loss": 3.368795659756987, + "tokens_seen": 2780954624 + }, + { + "epoch": 0.84, + "learning_rate": 0.00012729388164493482, + "loss": 1.2041, + "theoretical_loss": 3.368789513822602, + "tokens_seen": 2781020160 + }, + { + "epoch": 0.84, + "learning_rate": 0.0001272778335005015, + "loss": 1.2653, + "theoretical_loss": 3.368783368073599, + "tokens_seen": 2781085696 + }, + { + "epoch": 0.84, + "learning_rate": 0.0001272617853560682, + "loss": 1.2726, + "theoretical_loss": 3.368777222509968, + "tokens_seen": 2781151232 + }, + { + "epoch": 0.84, + "learning_rate": 0.0001272457372116349, + "loss": 1.2713, + "theoretical_loss": 3.3687710771316985, + "tokens_seen": 2781216768 + }, + { + "epoch": 0.84, + "learning_rate": 0.0001272296890672016, + "loss": 1.3413, + "theoretical_loss": 3.3687649319387813, + "tokens_seen": 2781282304 + }, + { + "epoch": 0.84, + "learning_rate": 0.0001272136409227683, + "loss": 1.4153, + "theoretical_loss": 3.368758786931206, + "tokens_seen": 2781347840 + }, + { + "epoch": 0.84, + "learning_rate": 0.000127197592778335, + "loss": 1.3301, + "theoretical_loss": 3.3687526421089626, + "tokens_seen": 2781413376 + }, + { + "epoch": 0.84, + "learning_rate": 0.0001271815446339017, + "loss": 1.3117, + "theoretical_loss": 3.3687464974720416, + "tokens_seen": 2781478912 + }, + { + "epoch": 0.84, + "learning_rate": 0.0001271654964894684, + "loss": 1.264, + "theoretical_loss": 3.3687403530204323, + "tokens_seen": 2781544448 + }, + { + "epoch": 0.84, + "learning_rate": 0.0001271494483450351, + "loss": 1.2547, + "theoretical_loss": 3.3687342087541254, + "tokens_seen": 2781609984 + }, + { + "epoch": 0.84, + "learning_rate": 0.0001271334002006018, + "loss": 1.3298, + "theoretical_loss": 3.3687280646731104, + "tokens_seen": 2781675520 + }, + { + "epoch": 0.84, + "learning_rate": 0.0001271173520561685, + "loss": 1.2419, + "theoretical_loss": 3.3687219207773778, + "tokens_seen": 2781741056 + }, + { + "epoch": 0.84, + "learning_rate": 0.0001271013039117352, + "loss": 1.2185, + "theoretical_loss": 3.368715777066918, + "tokens_seen": 2781806592 + }, + { + "epoch": 0.84, + "learning_rate": 0.0001270852557673019, + "loss": 1.3617, + "theoretical_loss": 3.36870963354172, + "tokens_seen": 2781872128 + }, + { + "epoch": 0.84, + "learning_rate": 0.0001270692076228686, + "loss": 1.2186, + "theoretical_loss": 3.3687034902017743, + "tokens_seen": 2781937664 + }, + { + "epoch": 0.84, + "objective/train/docs_used": 1010370, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.3678696155548096, + "objective/train/theoretical_loss": 3.368697347047071, + "objective/train/tokens_used": 2802463200, + "theoretical_loss": 3.368697347047071, + "tokens_seen": 2782003200 + }, + { + "epoch": 0.84, + "learning_rate": 0.0001270531594784353, + "loss": 1.3448, + "theoretical_loss": 3.368697347047071, + "tokens_seen": 2782003200 + }, + { + "epoch": 0.84, + "learning_rate": 0.000127037111334002, + "loss": 1.3094, + "theoretical_loss": 3.3686912040776007, + "tokens_seen": 2782068736 + }, + { + "epoch": 0.84, + "learning_rate": 0.00012702106318956873, + "loss": 1.2447, + "theoretical_loss": 3.3686850612933528, + "tokens_seen": 2782134272 + }, + { + "epoch": 0.84, + "learning_rate": 0.00012700501504513543, + "loss": 1.2781, + "theoretical_loss": 3.368678918694317, + "tokens_seen": 2782199808 + }, + { + "epoch": 0.84, + "learning_rate": 0.00012698896690070213, + "loss": 1.3925, + "theoretical_loss": 3.3686727762804844, + "tokens_seen": 2782265344 + }, + { + "epoch": 0.84, + "learning_rate": 0.00012697291875626883, + "loss": 1.1738, + "theoretical_loss": 3.3686666340518445, + "tokens_seen": 2782330880 + }, + { + "epoch": 0.84, + "learning_rate": 0.0001269568706118355, + "loss": 1.2281, + "theoretical_loss": 3.368660492008387, + "tokens_seen": 2782396416 + }, + { + "epoch": 0.84, + "learning_rate": 0.0001269408224674022, + "loss": 1.1979, + "theoretical_loss": 3.368654350150103, + "tokens_seen": 2782461952 + }, + { + "epoch": 0.84, + "learning_rate": 0.0001269247743229689, + "loss": 1.1922, + "theoretical_loss": 3.3686482084769813, + "tokens_seen": 2782527488 + }, + { + "epoch": 0.84, + "learning_rate": 0.0001269087261785356, + "loss": 1.2882, + "theoretical_loss": 3.3686420669890125, + "tokens_seen": 2782593024 + }, + { + "epoch": 0.84, + "learning_rate": 0.0001268926780341023, + "loss": 1.3333, + "theoretical_loss": 3.3686359256861866, + "tokens_seen": 2782658560 + }, + { + "epoch": 0.84, + "learning_rate": 0.000126876629889669, + "loss": 1.3015, + "theoretical_loss": 3.368629784568494, + "tokens_seen": 2782724096 + }, + { + "epoch": 0.84, + "learning_rate": 0.0001268605817452357, + "loss": 1.2904, + "theoretical_loss": 3.3686236436359245, + "tokens_seen": 2782789632 + }, + { + "epoch": 0.84, + "learning_rate": 0.0001268445336008024, + "loss": 1.3199, + "theoretical_loss": 3.368617502888468, + "tokens_seen": 2782855168 + }, + { + "epoch": 0.84, + "learning_rate": 0.0001268284854563691, + "loss": 1.2938, + "theoretical_loss": 3.3686113623261154, + "tokens_seen": 2782920704 + }, + { + "epoch": 0.84, + "learning_rate": 0.00012681243731193581, + "loss": 1.2225, + "theoretical_loss": 3.3686052219488554, + "tokens_seen": 2782986240 + }, + { + "epoch": 0.84, + "learning_rate": 0.00012679638916750251, + "loss": 1.276, + "theoretical_loss": 3.368599081756679, + "tokens_seen": 2783051776 + }, + { + "epoch": 0.84, + "learning_rate": 0.00012678034102306922, + "loss": 1.2963, + "theoretical_loss": 3.3685929417495757, + "tokens_seen": 2783117312 + }, + { + "epoch": 0.84, + "learning_rate": 0.00012676429287863592, + "loss": 1.301, + "theoretical_loss": 3.3685868019275365, + "tokens_seen": 2783182848 + }, + { + "epoch": 0.84, + "learning_rate": 0.00012674824473420262, + "loss": 1.3349, + "theoretical_loss": 3.36858066229055, + "tokens_seen": 2783248384 + }, + { + "epoch": 0.84, + "learning_rate": 0.00012673219658976932, + "loss": 1.2421, + "theoretical_loss": 3.3685745228386077, + "tokens_seen": 2783313920 + }, + { + "epoch": 0.84, + "learning_rate": 0.00012671614844533602, + "loss": 1.2634, + "theoretical_loss": 3.3685683835716986, + "tokens_seen": 2783379456 + }, + { + "epoch": 0.84, + "learning_rate": 0.00012670010030090272, + "loss": 1.3043, + "theoretical_loss": 3.368562244489814, + "tokens_seen": 2783444992 + }, + { + "epoch": 0.84, + "learning_rate": 0.00012668405215646942, + "loss": 1.3611, + "theoretical_loss": 3.3685561055929427, + "tokens_seen": 2783510528 + }, + { + "epoch": 0.84, + "learning_rate": 0.00012666800401203612, + "loss": 1.2736, + "theoretical_loss": 3.3685499668810754, + "tokens_seen": 2783576064 + }, + { + "epoch": 0.84, + "learning_rate": 0.00012665195586760282, + "loss": 1.3015, + "theoretical_loss": 3.368543828354202, + "tokens_seen": 2783641600 + }, + { + "epoch": 0.84, + "learning_rate": 0.00012663590772316952, + "loss": 1.3171, + "theoretical_loss": 3.3685376900123125, + "tokens_seen": 2783707136 + }, + { + "epoch": 0.84, + "learning_rate": 0.00012661985957873622, + "loss": 1.1706, + "theoretical_loss": 3.3685315518553973, + "tokens_seen": 2783772672 + }, + { + "epoch": 0.84, + "learning_rate": 0.00012660381143430292, + "loss": 1.2774, + "theoretical_loss": 3.3685254138834457, + "tokens_seen": 2783838208 + }, + { + "epoch": 0.84, + "learning_rate": 0.00012658776328986962, + "loss": 1.1755, + "theoretical_loss": 3.368519276096449, + "tokens_seen": 2783903744 + }, + { + "epoch": 0.84, + "learning_rate": 0.0001265717151454363, + "loss": 1.3024, + "theoretical_loss": 3.3685131384943965, + "tokens_seen": 2783969280 + }, + { + "epoch": 0.84, + "learning_rate": 0.000126555667001003, + "loss": 1.1247, + "theoretical_loss": 3.368507001077278, + "tokens_seen": 2784034816 + }, + { + "epoch": 0.84, + "learning_rate": 0.0001265396188565697, + "loss": 1.2504, + "theoretical_loss": 3.3685008638450844, + "tokens_seen": 2784100352 + }, + { + "epoch": 0.84, + "learning_rate": 0.0001265235707121364, + "loss": 1.2553, + "theoretical_loss": 3.368494726797805, + "tokens_seen": 2784165888 + }, + { + "epoch": 0.84, + "learning_rate": 0.0001265075225677031, + "loss": 1.3085, + "theoretical_loss": 3.3684885899354304, + "tokens_seen": 2784231424 + }, + { + "epoch": 0.84, + "learning_rate": 0.0001264914744232698, + "loss": 1.4123, + "theoretical_loss": 3.3684824532579505, + "tokens_seen": 2784296960 + }, + { + "epoch": 0.84, + "learning_rate": 0.0001264754262788365, + "loss": 1.2639, + "theoretical_loss": 3.368476316765355, + "tokens_seen": 2784362496 + }, + { + "epoch": 0.84, + "learning_rate": 0.0001264593781344032, + "loss": 1.2012, + "theoretical_loss": 3.368470180457635, + "tokens_seen": 2784428032 + }, + { + "epoch": 0.84, + "learning_rate": 0.0001264433299899699, + "loss": 1.2632, + "theoretical_loss": 3.3684640443347793, + "tokens_seen": 2784493568 + }, + { + "epoch": 0.84, + "learning_rate": 0.0001264272818455366, + "loss": 1.2701, + "theoretical_loss": 3.3684579083967785, + "tokens_seen": 2784559104 + }, + { + "epoch": 0.84, + "learning_rate": 0.00012641123370110333, + "loss": 1.2268, + "theoretical_loss": 3.368451772643623, + "tokens_seen": 2784624640 + }, + { + "epoch": 0.84, + "learning_rate": 0.00012639518555667003, + "loss": 1.3266, + "theoretical_loss": 3.368445637075303, + "tokens_seen": 2784690176 + }, + { + "epoch": 0.84, + "learning_rate": 0.00012637913741223673, + "loss": 1.3219, + "theoretical_loss": 3.368439501691808, + "tokens_seen": 2784755712 + }, + { + "epoch": 0.84, + "learning_rate": 0.00012636308926780343, + "loss": 1.3804, + "theoretical_loss": 3.3684333664931283, + "tokens_seen": 2784821248 + }, + { + "epoch": 0.84, + "learning_rate": 0.00012634704112337013, + "loss": 1.2918, + "theoretical_loss": 3.368427231479254, + "tokens_seen": 2784886784 + }, + { + "epoch": 0.84, + "learning_rate": 0.00012633099297893683, + "loss": 1.3788, + "theoretical_loss": 3.368421096650175, + "tokens_seen": 2784952320 + }, + { + "epoch": 0.84, + "learning_rate": 0.00012631494483450353, + "loss": 1.1975, + "theoretical_loss": 3.3684149620058816, + "tokens_seen": 2785017856 + }, + { + "epoch": 0.84, + "learning_rate": 0.00012629889669007023, + "loss": 1.261, + "theoretical_loss": 3.3684088275463644, + "tokens_seen": 2785083392 + }, + { + "epoch": 0.84, + "learning_rate": 0.00012628284854563694, + "loss": 1.2602, + "theoretical_loss": 3.3684026932716122, + "tokens_seen": 2785148928 + }, + { + "epoch": 0.84, + "learning_rate": 0.00012626680040120364, + "loss": 1.2706, + "theoretical_loss": 3.3683965591816163, + "tokens_seen": 2785214464 + }, + { + "epoch": 0.84, + "objective/train/docs_used": 1011592, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.277856469154358, + "objective/train/theoretical_loss": 3.3683904252763663, + "objective/train/tokens_used": 2805740000, + "theoretical_loss": 3.3683904252763663, + "tokens_seen": 2785280000 + }, + { + "epoch": 0.84, + "learning_rate": 0.0001262507522567703, + "loss": 1.2914, + "theoretical_loss": 3.3683904252763663, + "tokens_seen": 2785280000 + }, + { + "epoch": 0.84, + "learning_rate": 0.000126234704112337, + "loss": 1.2424, + "theoretical_loss": 3.368384291555852, + "tokens_seen": 2785345536 + }, + { + "epoch": 0.84, + "learning_rate": 0.0001262186559679037, + "loss": 1.2628, + "theoretical_loss": 3.3683781580200645, + "tokens_seen": 2785411072 + }, + { + "epoch": 0.84, + "learning_rate": 0.0001262026078234704, + "loss": 1.4201, + "theoretical_loss": 3.3683720246689925, + "tokens_seen": 2785476608 + }, + { + "epoch": 0.84, + "learning_rate": 0.0001261865596790371, + "loss": 1.2641, + "theoretical_loss": 3.368365891502627, + "tokens_seen": 2785542144 + }, + { + "epoch": 0.84, + "learning_rate": 0.0001261705115346038, + "loss": 1.2858, + "theoretical_loss": 3.368359758520958, + "tokens_seen": 2785607680 + }, + { + "epoch": 0.84, + "learning_rate": 0.0001261544633901705, + "loss": 1.2497, + "theoretical_loss": 3.3683536257239752, + "tokens_seen": 2785673216 + }, + { + "epoch": 0.84, + "learning_rate": 0.00012613841524573721, + "loss": 1.2378, + "theoretical_loss": 3.368347493111669, + "tokens_seen": 2785738752 + }, + { + "epoch": 0.84, + "learning_rate": 0.00012612236710130391, + "loss": 1.2629, + "theoretical_loss": 3.36834136068403, + "tokens_seen": 2785804288 + }, + { + "epoch": 0.84, + "learning_rate": 0.00012610631895687061, + "loss": 1.2835, + "theoretical_loss": 3.368335228441047, + "tokens_seen": 2785869824 + }, + { + "epoch": 0.84, + "learning_rate": 0.00012609027081243732, + "loss": 1.3808, + "theoretical_loss": 3.3683290963827113, + "tokens_seen": 2785935360 + }, + { + "epoch": 0.84, + "learning_rate": 0.00012607422266800402, + "loss": 1.3404, + "theoretical_loss": 3.3683229645090123, + "tokens_seen": 2786000896 + }, + { + "epoch": 0.84, + "learning_rate": 0.00012605817452357072, + "loss": 1.2091, + "theoretical_loss": 3.36831683281994, + "tokens_seen": 2786066432 + }, + { + "epoch": 0.84, + "learning_rate": 0.00012604212637913742, + "loss": 1.2064, + "theoretical_loss": 3.368310701315486, + "tokens_seen": 2786131968 + }, + { + "epoch": 0.84, + "learning_rate": 0.00012602607823470412, + "loss": 1.292, + "theoretical_loss": 3.368304569995638, + "tokens_seen": 2786197504 + }, + { + "epoch": 0.84, + "learning_rate": 0.00012601003009027082, + "loss": 1.3162, + "theoretical_loss": 3.3682984388603883, + "tokens_seen": 2786263040 + }, + { + "epoch": 0.84, + "learning_rate": 0.00012599398194583752, + "loss": 1.1958, + "theoretical_loss": 3.3682923079097256, + "tokens_seen": 2786328576 + }, + { + "epoch": 0.84, + "learning_rate": 0.00012597793380140422, + "loss": 1.2422, + "theoretical_loss": 3.36828617714364, + "tokens_seen": 2786394112 + }, + { + "epoch": 0.84, + "learning_rate": 0.00012596188565697092, + "loss": 1.1959, + "theoretical_loss": 3.3682800465621225, + "tokens_seen": 2786459648 + }, + { + "epoch": 0.84, + "learning_rate": 0.00012594583751253762, + "loss": 1.3266, + "theoretical_loss": 3.3682739161651627, + "tokens_seen": 2786525184 + }, + { + "epoch": 0.84, + "learning_rate": 0.00012592978936810432, + "loss": 1.381, + "theoretical_loss": 3.3682677859527512, + "tokens_seen": 2786590720 + }, + { + "epoch": 0.84, + "learning_rate": 0.00012591374122367102, + "loss": 1.2593, + "theoretical_loss": 3.368261655924877, + "tokens_seen": 2786656256 + }, + { + "epoch": 0.84, + "learning_rate": 0.00012589769307923772, + "loss": 1.2079, + "theoretical_loss": 3.3682555260815312, + "tokens_seen": 2786721792 + }, + { + "epoch": 0.84, + "learning_rate": 0.00012588164493480442, + "loss": 1.3704, + "theoretical_loss": 3.368249396422703, + "tokens_seen": 2786787328 + }, + { + "epoch": 0.84, + "learning_rate": 0.0001258655967903711, + "loss": 1.2848, + "theoretical_loss": 3.368243266948384, + "tokens_seen": 2786852864 + }, + { + "epoch": 0.84, + "learning_rate": 0.0001258495486459378, + "loss": 1.2394, + "theoretical_loss": 3.368237137658563, + "tokens_seen": 2786918400 + }, + { + "epoch": 0.84, + "learning_rate": 0.0001258335005015045, + "loss": 1.2216, + "theoretical_loss": 3.3682310085532303, + "tokens_seen": 2786983936 + }, + { + "epoch": 0.84, + "learning_rate": 0.00012581745235707123, + "loss": 1.2819, + "theoretical_loss": 3.3682248796323764, + "tokens_seen": 2787049472 + }, + { + "epoch": 0.84, + "learning_rate": 0.00012580140421263793, + "loss": 1.2497, + "theoretical_loss": 3.368218750895991, + "tokens_seen": 2787115008 + }, + { + "epoch": 0.84, + "learning_rate": 0.00012578535606820463, + "loss": 1.3052, + "theoretical_loss": 3.3682126223440645, + "tokens_seen": 2787180544 + }, + { + "epoch": 0.84, + "learning_rate": 0.00012576930792377133, + "loss": 1.3572, + "theoretical_loss": 3.368206493976587, + "tokens_seen": 2787246080 + }, + { + "epoch": 0.84, + "learning_rate": 0.00012575325977933803, + "loss": 1.3056, + "theoretical_loss": 3.368200365793548, + "tokens_seen": 2787311616 + }, + { + "epoch": 0.84, + "learning_rate": 0.00012573721163490473, + "loss": 1.2531, + "theoretical_loss": 3.368194237794939, + "tokens_seen": 2787377152 + }, + { + "epoch": 0.84, + "learning_rate": 0.00012572116349047143, + "loss": 1.326, + "theoretical_loss": 3.368188109980749, + "tokens_seen": 2787442688 + }, + { + "epoch": 0.84, + "learning_rate": 0.00012570511534603813, + "loss": 1.2876, + "theoretical_loss": 3.368181982350968, + "tokens_seen": 2787508224 + }, + { + "epoch": 0.84, + "learning_rate": 0.00012568906720160483, + "loss": 1.3378, + "theoretical_loss": 3.368175854905587, + "tokens_seen": 2787573760 + }, + { + "epoch": 0.84, + "learning_rate": 0.00012567301905717153, + "loss": 1.2777, + "theoretical_loss": 3.368169727644595, + "tokens_seen": 2787639296 + }, + { + "epoch": 0.84, + "learning_rate": 0.00012565697091273823, + "loss": 1.3214, + "theoretical_loss": 3.3681636005679834, + "tokens_seen": 2787704832 + }, + { + "epoch": 0.84, + "learning_rate": 0.00012564092276830493, + "loss": 1.2818, + "theoretical_loss": 3.3681574736757414, + "tokens_seen": 2787770368 + }, + { + "epoch": 0.84, + "learning_rate": 0.00012562487462387163, + "loss": 1.2398, + "theoretical_loss": 3.368151346967859, + "tokens_seen": 2787835904 + }, + { + "epoch": 0.84, + "learning_rate": 0.00012560882647943833, + "loss": 1.2015, + "theoretical_loss": 3.368145220444327, + "tokens_seen": 2787901440 + }, + { + "epoch": 0.84, + "learning_rate": 0.00012559277833500504, + "loss": 1.2049, + "theoretical_loss": 3.3681390941051355, + "tokens_seen": 2787966976 + }, + { + "epoch": 0.84, + "learning_rate": 0.00012557673019057174, + "loss": 1.2715, + "theoretical_loss": 3.368132967950274, + "tokens_seen": 2788032512 + }, + { + "epoch": 0.84, + "learning_rate": 0.0001255606820461384, + "loss": 1.1439, + "theoretical_loss": 3.3681268419797328, + "tokens_seen": 2788098048 + }, + { + "epoch": 0.84, + "learning_rate": 0.0001255446339017051, + "loss": 1.2318, + "theoretical_loss": 3.3681207161935025, + "tokens_seen": 2788163584 + }, + { + "epoch": 0.84, + "learning_rate": 0.0001255285857572718, + "loss": 1.243, + "theoretical_loss": 3.368114590591573, + "tokens_seen": 2788229120 + }, + { + "epoch": 0.84, + "learning_rate": 0.0001255125376128385, + "loss": 1.3531, + "theoretical_loss": 3.3681084651739335, + "tokens_seen": 2788294656 + }, + { + "epoch": 0.84, + "learning_rate": 0.0001254964894684052, + "loss": 1.2682, + "theoretical_loss": 3.3681023399405756, + "tokens_seen": 2788360192 + }, + { + "epoch": 0.84, + "learning_rate": 0.0001254804413239719, + "loss": 1.2864, + "theoretical_loss": 3.3680962148914886, + "tokens_seen": 2788425728 + }, + { + "epoch": 0.84, + "learning_rate": 0.0001254643931795386, + "loss": 1.2983, + "theoretical_loss": 3.368090090026663, + "tokens_seen": 2788491264 + }, + { + "epoch": 0.84, + "objective/train/docs_used": 1012721, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.2645659446716309, + "objective/train/theoretical_loss": 3.3680839653460883, + "objective/train/tokens_used": 2809016800, + "theoretical_loss": 3.3680839653460883, + "tokens_seen": 2788556800 + }, + { + "epoch": 0.85, + "learning_rate": 0.00012544834503510531, + "loss": 1.2204, + "theoretical_loss": 3.3680839653460883, + "tokens_seen": 2788556800 + }, + { + "epoch": 0.85, + "learning_rate": 0.00012543229689067201, + "loss": 1.2145, + "theoretical_loss": 3.368077840849755, + "tokens_seen": 2788622336 + }, + { + "epoch": 0.85, + "learning_rate": 0.00012541624874623872, + "loss": 1.2902, + "theoretical_loss": 3.3680717165376537, + "tokens_seen": 2788687872 + }, + { + "epoch": 0.85, + "learning_rate": 0.00012540020060180542, + "loss": 1.1869, + "theoretical_loss": 3.368065592409774, + "tokens_seen": 2788753408 + }, + { + "epoch": 0.85, + "learning_rate": 0.00012538415245737212, + "loss": 1.2562, + "theoretical_loss": 3.368059468466106, + "tokens_seen": 2788818944 + }, + { + "epoch": 0.85, + "learning_rate": 0.00012536810431293882, + "loss": 1.2284, + "theoretical_loss": 3.36805334470664, + "tokens_seen": 2788884480 + }, + { + "epoch": 0.85, + "learning_rate": 0.00012535205616850552, + "loss": 1.2752, + "theoretical_loss": 3.3680472211313663, + "tokens_seen": 2788950016 + }, + { + "epoch": 0.85, + "learning_rate": 0.00012533600802407222, + "loss": 1.2413, + "theoretical_loss": 3.3680410977402744, + "tokens_seen": 2789015552 + }, + { + "epoch": 0.85, + "learning_rate": 0.00012531995987963892, + "loss": 1.3037, + "theoretical_loss": 3.368034974533355, + "tokens_seen": 2789081088 + }, + { + "epoch": 0.85, + "learning_rate": 0.00012530391173520562, + "loss": 1.3154, + "theoretical_loss": 3.3680288515105983, + "tokens_seen": 2789146624 + }, + { + "epoch": 0.85, + "learning_rate": 0.00012528786359077232, + "loss": 1.41, + "theoretical_loss": 3.368022728671994, + "tokens_seen": 2789212160 + }, + { + "epoch": 0.85, + "learning_rate": 0.00012527181544633902, + "loss": 1.1415, + "theoretical_loss": 3.3680166060175325, + "tokens_seen": 2789277696 + }, + { + "epoch": 0.85, + "learning_rate": 0.00012525576730190572, + "loss": 1.2515, + "theoretical_loss": 3.368010483547204, + "tokens_seen": 2789343232 + }, + { + "epoch": 0.85, + "learning_rate": 0.00012523971915747242, + "loss": 1.2951, + "theoretical_loss": 3.368004361260998, + "tokens_seen": 2789408768 + }, + { + "epoch": 0.85, + "learning_rate": 0.00012522367101303912, + "loss": 1.298, + "theoretical_loss": 3.367998239158906, + "tokens_seen": 2789474304 + }, + { + "epoch": 0.85, + "learning_rate": 0.00012520762286860582, + "loss": 1.3629, + "theoretical_loss": 3.3679921172409166, + "tokens_seen": 2789539840 + }, + { + "epoch": 0.85, + "learning_rate": 0.00012519157472417252, + "loss": 1.3064, + "theoretical_loss": 3.367985995507021, + "tokens_seen": 2789605376 + }, + { + "epoch": 0.85, + "learning_rate": 0.00012517552657973922, + "loss": 1.355, + "theoretical_loss": 3.367979873957209, + "tokens_seen": 2789670912 + }, + { + "epoch": 0.85, + "learning_rate": 0.00012515947843530593, + "loss": 1.3039, + "theoretical_loss": 3.3679737525914706, + "tokens_seen": 2789736448 + }, + { + "epoch": 0.85, + "learning_rate": 0.00012514343029087263, + "loss": 1.3311, + "theoretical_loss": 3.3679676314097957, + "tokens_seen": 2789801984 + }, + { + "epoch": 0.85, + "learning_rate": 0.00012512738214643933, + "loss": 1.3049, + "theoretical_loss": 3.3679615104121754, + "tokens_seen": 2789867520 + }, + { + "epoch": 0.85, + "learning_rate": 0.00012511133400200603, + "loss": 1.3038, + "theoretical_loss": 3.367955389598599, + "tokens_seen": 2789933056 + }, + { + "epoch": 0.85, + "learning_rate": 0.00012509528585757273, + "loss": 1.3356, + "theoretical_loss": 3.367949268969057, + "tokens_seen": 2789998592 + }, + { + "epoch": 0.85, + "learning_rate": 0.00012507923771313943, + "loss": 1.2964, + "theoretical_loss": 3.3679431485235396, + "tokens_seen": 2790064128 + }, + { + "epoch": 0.85, + "learning_rate": 0.00012506318956870613, + "loss": 1.3301, + "theoretical_loss": 3.3679370282620362, + "tokens_seen": 2790129664 + }, + { + "epoch": 0.85, + "learning_rate": 0.00012504714142427283, + "loss": 1.1915, + "theoretical_loss": 3.3679309081845377, + "tokens_seen": 2790195200 + }, + { + "epoch": 0.85, + "learning_rate": 0.00012503109327983953, + "loss": 1.3202, + "theoretical_loss": 3.3679247882910346, + "tokens_seen": 2790260736 + }, + { + "epoch": 0.85, + "learning_rate": 0.00012501504513540623, + "loss": 1.2479, + "theoretical_loss": 3.367918668581516, + "tokens_seen": 2790326272 + }, + { + "epoch": 0.85, + "learning_rate": 0.00012499899699097293, + "loss": 1.2925, + "theoretical_loss": 3.3679125490559727, + "tokens_seen": 2790391808 + }, + { + "epoch": 0.85, + "learning_rate": 0.00012498294884653963, + "loss": 1.2767, + "theoretical_loss": 3.3679064297143944, + "tokens_seen": 2790457344 + }, + { + "epoch": 0.85, + "learning_rate": 0.00012496690070210633, + "loss": 1.279, + "theoretical_loss": 3.367900310556772, + "tokens_seen": 2790522880 + }, + { + "epoch": 0.85, + "learning_rate": 0.00012495085255767303, + "loss": 1.3146, + "theoretical_loss": 3.367894191583095, + "tokens_seen": 2790588416 + }, + { + "epoch": 0.85, + "learning_rate": 0.00012493480441323973, + "loss": 1.2168, + "theoretical_loss": 3.367888072793354, + "tokens_seen": 2790653952 + }, + { + "epoch": 0.85, + "learning_rate": 0.00012491875626880643, + "loss": 1.2245, + "theoretical_loss": 3.3678819541875384, + "tokens_seen": 2790719488 + }, + { + "epoch": 0.85, + "learning_rate": 0.00012490270812437314, + "loss": 1.3241, + "theoretical_loss": 3.3678758357656395, + "tokens_seen": 2790785024 + }, + { + "epoch": 0.85, + "learning_rate": 0.00012488665997993984, + "loss": 1.3224, + "theoretical_loss": 3.3678697175276464, + "tokens_seen": 2790850560 + }, + { + "epoch": 0.85, + "learning_rate": 0.00012487061183550654, + "loss": 1.2416, + "theoretical_loss": 3.3678635994735497, + "tokens_seen": 2790916096 + }, + { + "epoch": 0.85, + "learning_rate": 0.0001248545636910732, + "loss": 1.1923, + "theoretical_loss": 3.3678574816033393, + "tokens_seen": 2790981632 + }, + { + "epoch": 0.85, + "learning_rate": 0.0001248385155466399, + "loss": 1.3288, + "theoretical_loss": 3.3678513639170062, + "tokens_seen": 2791047168 + }, + { + "epoch": 0.85, + "learning_rate": 0.0001248224674022066, + "loss": 1.2291, + "theoretical_loss": 3.3678452464145394, + "tokens_seen": 2791112704 + }, + { + "epoch": 0.85, + "learning_rate": 0.0001248064192577733, + "loss": 1.222, + "theoretical_loss": 3.36783912909593, + "tokens_seen": 2791178240 + }, + { + "epoch": 0.85, + "learning_rate": 0.00012479037111334, + "loss": 1.3414, + "theoretical_loss": 3.3678330119611672, + "tokens_seen": 2791243776 + }, + { + "epoch": 0.85, + "learning_rate": 0.0001247743229689067, + "loss": 1.3269, + "theoretical_loss": 3.3678268950102423, + "tokens_seen": 2791309312 + }, + { + "epoch": 0.85, + "learning_rate": 0.00012475827482447341, + "loss": 1.3641, + "theoretical_loss": 3.3678207782431446, + "tokens_seen": 2791374848 + }, + { + "epoch": 0.85, + "learning_rate": 0.00012474222668004011, + "loss": 1.2414, + "theoretical_loss": 3.3678146616598648, + "tokens_seen": 2791440384 + }, + { + "epoch": 0.85, + "learning_rate": 0.00012472617853560682, + "loss": 1.3748, + "theoretical_loss": 3.3678085452603925, + "tokens_seen": 2791505920 + }, + { + "epoch": 0.85, + "learning_rate": 0.00012471013039117352, + "loss": 1.2667, + "theoretical_loss": 3.367802429044718, + "tokens_seen": 2791571456 + }, + { + "epoch": 0.85, + "learning_rate": 0.00012469408224674022, + "loss": 1.3126, + "theoretical_loss": 3.367796313012832, + "tokens_seen": 2791636992 + }, + { + "epoch": 0.85, + "learning_rate": 0.00012467803410230692, + "loss": 1.3666, + "theoretical_loss": 3.367790197164724, + "tokens_seen": 2791702528 + }, + { + "epoch": 0.85, + "learning_rate": 0.00012466198595787362, + "loss": 1.2616, + "theoretical_loss": 3.3677840815003846, + "tokens_seen": 2791768064 + }, + { + "epoch": 0.85, + "objective/train/docs_used": 1014349, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.406116247177124, + "objective/train/theoretical_loss": 3.3677779660198035, + "objective/train/tokens_used": 2812293600, + "theoretical_loss": 3.3677779660198035, + "tokens_seen": 2791833600 + }, + { + "epoch": 0.85, + "learning_rate": 0.00012464593781344032, + "loss": 1.345, + "theoretical_loss": 3.3677779660198035, + "tokens_seen": 2791833600 + }, + { + "epoch": 0.85, + "learning_rate": 0.00012462988966900702, + "loss": 1.2252, + "theoretical_loss": 3.3677718507229715, + "tokens_seen": 2791899136 + }, + { + "epoch": 0.85, + "learning_rate": 0.00012461384152457375, + "loss": 1.268, + "theoretical_loss": 3.3677657356098782, + "tokens_seen": 2791964672 + }, + { + "epoch": 0.85, + "learning_rate": 0.00012459779338014045, + "loss": 1.2293, + "theoretical_loss": 3.3677596206805145, + "tokens_seen": 2792030208 + }, + { + "epoch": 0.85, + "learning_rate": 0.00012458174523570715, + "loss": 1.2, + "theoretical_loss": 3.3677535059348696, + "tokens_seen": 2792095744 + }, + { + "epoch": 0.85, + "learning_rate": 0.00012456569709127385, + "loss": 1.308, + "theoretical_loss": 3.367747391372934, + "tokens_seen": 2792161280 + }, + { + "epoch": 0.85, + "learning_rate": 0.00012454964894684055, + "loss": 1.3216, + "theoretical_loss": 3.3677412769946984, + "tokens_seen": 2792226816 + }, + { + "epoch": 0.85, + "learning_rate": 0.00012453360080240722, + "loss": 1.3122, + "theoretical_loss": 3.3677351628001526, + "tokens_seen": 2792292352 + }, + { + "epoch": 0.85, + "learning_rate": 0.00012451755265797392, + "loss": 1.2771, + "theoretical_loss": 3.367729048789286, + "tokens_seen": 2792357888 + }, + { + "epoch": 0.85, + "learning_rate": 0.00012450150451354062, + "loss": 1.3513, + "theoretical_loss": 3.3677229349620905, + "tokens_seen": 2792423424 + }, + { + "epoch": 0.85, + "learning_rate": 0.00012448545636910732, + "loss": 1.3193, + "theoretical_loss": 3.367716821318555, + "tokens_seen": 2792488960 + }, + { + "epoch": 0.85, + "learning_rate": 0.00012446940822467403, + "loss": 1.2714, + "theoretical_loss": 3.3677107078586697, + "tokens_seen": 2792554496 + }, + { + "epoch": 0.85, + "learning_rate": 0.00012445336008024073, + "loss": 1.2915, + "theoretical_loss": 3.367704594582425, + "tokens_seen": 2792620032 + }, + { + "epoch": 0.85, + "learning_rate": 0.00012443731193580743, + "loss": 1.3219, + "theoretical_loss": 3.3676984814898114, + "tokens_seen": 2792685568 + }, + { + "epoch": 0.85, + "learning_rate": 0.00012442126379137413, + "loss": 1.1718, + "theoretical_loss": 3.3676923685808187, + "tokens_seen": 2792751104 + }, + { + "epoch": 0.85, + "learning_rate": 0.00012440521564694083, + "loss": 1.3046, + "theoretical_loss": 3.3676862558554372, + "tokens_seen": 2792816640 + }, + { + "epoch": 0.85, + "learning_rate": 0.00012438916750250753, + "loss": 1.3248, + "theoretical_loss": 3.367680143313657, + "tokens_seen": 2792882176 + }, + { + "epoch": 0.85, + "learning_rate": 0.00012437311935807423, + "loss": 1.2364, + "theoretical_loss": 3.367674030955468, + "tokens_seen": 2792947712 + }, + { + "epoch": 0.85, + "learning_rate": 0.00012435707121364093, + "loss": 1.2874, + "theoretical_loss": 3.367667918780861, + "tokens_seen": 2793013248 + }, + { + "epoch": 0.85, + "learning_rate": 0.00012434102306920763, + "loss": 1.2829, + "theoretical_loss": 3.367661806789826, + "tokens_seen": 2793078784 + }, + { + "epoch": 0.85, + "learning_rate": 0.00012432497492477433, + "loss": 1.2396, + "theoretical_loss": 3.3676556949823526, + "tokens_seen": 2793144320 + }, + { + "epoch": 0.85, + "learning_rate": 0.00012430892678034103, + "loss": 1.2487, + "theoretical_loss": 3.367649583358432, + "tokens_seen": 2793209856 + }, + { + "epoch": 0.85, + "learning_rate": 0.00012429287863590773, + "loss": 1.2646, + "theoretical_loss": 3.3676434719180532, + "tokens_seen": 2793275392 + }, + { + "epoch": 0.85, + "learning_rate": 0.00012427683049147443, + "loss": 1.259, + "theoretical_loss": 3.367637360661208, + "tokens_seen": 2793340928 + }, + { + "epoch": 0.85, + "learning_rate": 0.00012426078234704113, + "loss": 1.2543, + "theoretical_loss": 3.367631249587885, + "tokens_seen": 2793406464 + }, + { + "epoch": 0.85, + "learning_rate": 0.00012424473420260783, + "loss": 1.1996, + "theoretical_loss": 3.3676251386980747, + "tokens_seen": 2793472000 + }, + { + "epoch": 0.85, + "learning_rate": 0.00012422868605817453, + "loss": 1.2667, + "theoretical_loss": 3.3676190279917675, + "tokens_seen": 2793537536 + }, + { + "epoch": 0.85, + "learning_rate": 0.00012421263791374124, + "loss": 1.2761, + "theoretical_loss": 3.3676129174689535, + "tokens_seen": 2793603072 + }, + { + "epoch": 0.85, + "learning_rate": 0.00012419658976930794, + "loss": 1.3458, + "theoretical_loss": 3.3676068071296235, + "tokens_seen": 2793668608 + }, + { + "epoch": 0.85, + "learning_rate": 0.00012418054162487464, + "loss": 1.2828, + "theoretical_loss": 3.367600696973767, + "tokens_seen": 2793734144 + }, + { + "epoch": 0.85, + "learning_rate": 0.00012416449348044134, + "loss": 1.3295, + "theoretical_loss": 3.3675945870013746, + "tokens_seen": 2793799680 + }, + { + "epoch": 0.85, + "learning_rate": 0.000124148445336008, + "loss": 1.2444, + "theoretical_loss": 3.3675884772124363, + "tokens_seen": 2793865216 + }, + { + "epoch": 0.85, + "learning_rate": 0.0001241323971915747, + "loss": 1.1818, + "theoretical_loss": 3.367582367606942, + "tokens_seen": 2793930752 + }, + { + "epoch": 0.85, + "learning_rate": 0.0001241163490471414, + "loss": 1.3806, + "theoretical_loss": 3.367576258184882, + "tokens_seen": 2793996288 + }, + { + "epoch": 0.85, + "learning_rate": 0.0001241003009027081, + "loss": 1.1816, + "theoretical_loss": 3.367570148946247, + "tokens_seen": 2794061824 + }, + { + "epoch": 0.85, + "learning_rate": 0.0001240842527582748, + "loss": 1.2991, + "theoretical_loss": 3.3675640398910267, + "tokens_seen": 2794127360 + }, + { + "epoch": 0.85, + "learning_rate": 0.00012406820461384151, + "loss": 1.2095, + "theoretical_loss": 3.3675579310192116, + "tokens_seen": 2794192896 + }, + { + "epoch": 0.85, + "learning_rate": 0.00012405215646940821, + "loss": 1.3446, + "theoretical_loss": 3.3675518223307916, + "tokens_seen": 2794258432 + }, + { + "epoch": 0.85, + "learning_rate": 0.00012403610832497492, + "loss": 1.2517, + "theoretical_loss": 3.367545713825757, + "tokens_seen": 2794323968 + }, + { + "epoch": 0.85, + "learning_rate": 0.00012402006018054164, + "loss": 1.2641, + "theoretical_loss": 3.367539605504098, + "tokens_seen": 2794389504 + }, + { + "epoch": 0.85, + "learning_rate": 0.00012400401203610834, + "loss": 1.2822, + "theoretical_loss": 3.3675334973658044, + "tokens_seen": 2794455040 + }, + { + "epoch": 0.85, + "learning_rate": 0.00012398796389167504, + "loss": 1.3267, + "theoretical_loss": 3.3675273894108675, + "tokens_seen": 2794520576 + }, + { + "epoch": 0.85, + "learning_rate": 0.00012397191574724175, + "loss": 1.2951, + "theoretical_loss": 3.3675212816392763, + "tokens_seen": 2794586112 + }, + { + "epoch": 0.85, + "learning_rate": 0.00012395586760280845, + "loss": 1.2603, + "theoretical_loss": 3.3675151740510216, + "tokens_seen": 2794651648 + }, + { + "epoch": 0.85, + "learning_rate": 0.00012393981945837515, + "loss": 1.327, + "theoretical_loss": 3.3675090666460936, + "tokens_seen": 2794717184 + }, + { + "epoch": 0.85, + "learning_rate": 0.00012392377131394185, + "loss": 1.3115, + "theoretical_loss": 3.3675029594244825, + "tokens_seen": 2794782720 + }, + { + "epoch": 0.85, + "learning_rate": 0.00012390772316950855, + "loss": 1.3762, + "theoretical_loss": 3.367496852386178, + "tokens_seen": 2794848256 + }, + { + "epoch": 0.85, + "learning_rate": 0.00012389167502507525, + "loss": 1.2083, + "theoretical_loss": 3.367490745531171, + "tokens_seen": 2794913792 + }, + { + "epoch": 0.85, + "learning_rate": 0.00012387562688064195, + "loss": 1.2787, + "theoretical_loss": 3.3674846388594513, + "tokens_seen": 2794979328 + }, + { + "epoch": 0.85, + "learning_rate": 0.00012385957873620865, + "loss": 1.2068, + "theoretical_loss": 3.367478532371009, + "tokens_seen": 2795044864 + }, + { + "epoch": 0.85, + "objective/train/docs_used": 1015624, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.3005752563476562, + "objective/train/theoretical_loss": 3.367472426065835, + "objective/train/tokens_used": 2815570400, + "theoretical_loss": 3.367472426065835, + "tokens_seen": 2795110400 + }, + { + "epoch": 0.85, + "learning_rate": 0.00012384353059177532, + "loss": 1.2471, + "theoretical_loss": 3.367472426065835, + "tokens_seen": 2795110400 + }, + { + "epoch": 0.85, + "learning_rate": 0.00012382748244734202, + "loss": 1.1962, + "theoretical_loss": 3.3674663199439183, + "tokens_seen": 2795175936 + }, + { + "epoch": 0.85, + "learning_rate": 0.00012381143430290872, + "loss": 1.2016, + "theoretical_loss": 3.3674602140052503, + "tokens_seen": 2795241472 + }, + { + "epoch": 0.85, + "learning_rate": 0.00012379538615847542, + "loss": 1.3212, + "theoretical_loss": 3.3674541082498206, + "tokens_seen": 2795307008 + }, + { + "epoch": 0.85, + "learning_rate": 0.00012377933801404213, + "loss": 1.3067, + "theoretical_loss": 3.3674480026776195, + "tokens_seen": 2795372544 + }, + { + "epoch": 0.85, + "learning_rate": 0.00012376328986960883, + "loss": 1.2804, + "theoretical_loss": 3.367441897288637, + "tokens_seen": 2795438080 + }, + { + "epoch": 0.85, + "learning_rate": 0.00012374724172517553, + "loss": 1.3126, + "theoretical_loss": 3.3674357920828637, + "tokens_seen": 2795503616 + }, + { + "epoch": 0.85, + "learning_rate": 0.00012373119358074223, + "loss": 1.2222, + "theoretical_loss": 3.3674296870602896, + "tokens_seen": 2795569152 + }, + { + "epoch": 0.85, + "learning_rate": 0.00012371514543630893, + "loss": 1.3684, + "theoretical_loss": 3.367423582220905, + "tokens_seen": 2795634688 + }, + { + "epoch": 0.85, + "learning_rate": 0.00012369909729187563, + "loss": 1.2474, + "theoretical_loss": 3.3674174775647, + "tokens_seen": 2795700224 + }, + { + "epoch": 0.85, + "learning_rate": 0.00012368304914744233, + "loss": 1.3008, + "theoretical_loss": 3.367411373091665, + "tokens_seen": 2795765760 + }, + { + "epoch": 0.85, + "learning_rate": 0.00012366700100300903, + "loss": 1.3194, + "theoretical_loss": 3.3674052688017895, + "tokens_seen": 2795831296 + }, + { + "epoch": 0.85, + "learning_rate": 0.00012365095285857573, + "loss": 1.2043, + "theoretical_loss": 3.367399164695065, + "tokens_seen": 2795896832 + }, + { + "epoch": 0.85, + "learning_rate": 0.00012363490471414243, + "loss": 1.2665, + "theoretical_loss": 3.36739306077148, + "tokens_seen": 2795962368 + }, + { + "epoch": 0.85, + "learning_rate": 0.00012361885656970913, + "loss": 1.278, + "theoretical_loss": 3.3673869570310266, + "tokens_seen": 2796027904 + }, + { + "epoch": 0.85, + "learning_rate": 0.00012360280842527583, + "loss": 1.2781, + "theoretical_loss": 3.367380853473694, + "tokens_seen": 2796093440 + }, + { + "epoch": 0.85, + "learning_rate": 0.00012358676028084253, + "loss": 1.237, + "theoretical_loss": 3.367374750099472, + "tokens_seen": 2796158976 + }, + { + "epoch": 0.85, + "learning_rate": 0.00012357071213640923, + "loss": 1.2744, + "theoretical_loss": 3.367368646908352, + "tokens_seen": 2796224512 + }, + { + "epoch": 0.85, + "learning_rate": 0.00012355466399197593, + "loss": 1.2805, + "theoretical_loss": 3.3673625439003234, + "tokens_seen": 2796290048 + }, + { + "epoch": 0.85, + "learning_rate": 0.00012353861584754264, + "loss": 1.2134, + "theoretical_loss": 3.3673564410753767, + "tokens_seen": 2796355584 + }, + { + "epoch": 0.85, + "learning_rate": 0.00012352256770310934, + "loss": 1.3168, + "theoretical_loss": 3.3673503384335017, + "tokens_seen": 2796421120 + }, + { + "epoch": 0.85, + "learning_rate": 0.00012350651955867604, + "loss": 1.2422, + "theoretical_loss": 3.367344235974689, + "tokens_seen": 2796486656 + }, + { + "epoch": 0.85, + "learning_rate": 0.00012349047141424274, + "loss": 1.2321, + "theoretical_loss": 3.3673381336989285, + "tokens_seen": 2796552192 + }, + { + "epoch": 0.85, + "learning_rate": 0.00012347442326980944, + "loss": 1.2688, + "theoretical_loss": 3.3673320316062108, + "tokens_seen": 2796617728 + }, + { + "epoch": 0.85, + "learning_rate": 0.0001234583751253761, + "loss": 1.3539, + "theoretical_loss": 3.367325929696526, + "tokens_seen": 2796683264 + }, + { + "epoch": 0.85, + "learning_rate": 0.0001234423269809428, + "loss": 1.303, + "theoretical_loss": 3.3673198279698644, + "tokens_seen": 2796748800 + }, + { + "epoch": 0.85, + "learning_rate": 0.0001234262788365095, + "loss": 1.3111, + "theoretical_loss": 3.367313726426216, + "tokens_seen": 2796814336 + }, + { + "epoch": 0.85, + "learning_rate": 0.00012341023069207624, + "loss": 1.3276, + "theoretical_loss": 3.367307625065571, + "tokens_seen": 2796879872 + }, + { + "epoch": 0.85, + "learning_rate": 0.00012339418254764294, + "loss": 1.1791, + "theoretical_loss": 3.36730152388792, + "tokens_seen": 2796945408 + }, + { + "epoch": 0.85, + "learning_rate": 0.00012337813440320964, + "loss": 1.309, + "theoretical_loss": 3.367295422893253, + "tokens_seen": 2797010944 + }, + { + "epoch": 0.85, + "learning_rate": 0.00012336208625877634, + "loss": 1.3539, + "theoretical_loss": 3.36728932208156, + "tokens_seen": 2797076480 + }, + { + "epoch": 0.85, + "learning_rate": 0.00012334603811434304, + "loss": 1.3152, + "theoretical_loss": 3.3672832214528317, + "tokens_seen": 2797142016 + }, + { + "epoch": 0.85, + "learning_rate": 0.00012332998996990974, + "loss": 1.1983, + "theoretical_loss": 3.3672771210070573, + "tokens_seen": 2797207552 + }, + { + "epoch": 0.85, + "learning_rate": 0.00012331394182547644, + "loss": 1.2655, + "theoretical_loss": 3.3672710207442287, + "tokens_seen": 2797273088 + }, + { + "epoch": 0.85, + "learning_rate": 0.00012329789368104314, + "loss": 1.4418, + "theoretical_loss": 3.3672649206643346, + "tokens_seen": 2797338624 + }, + { + "epoch": 0.85, + "learning_rate": 0.00012328184553660985, + "loss": 1.3032, + "theoretical_loss": 3.367258820767366, + "tokens_seen": 2797404160 + }, + { + "epoch": 0.85, + "learning_rate": 0.00012326579739217655, + "loss": 1.2707, + "theoretical_loss": 3.367252721053313, + "tokens_seen": 2797469696 + }, + { + "epoch": 0.85, + "learning_rate": 0.00012324974924774325, + "loss": 1.3275, + "theoretical_loss": 3.367246621522166, + "tokens_seen": 2797535232 + }, + { + "epoch": 0.85, + "learning_rate": 0.00012323370110330995, + "loss": 1.3121, + "theoretical_loss": 3.367240522173915, + "tokens_seen": 2797600768 + }, + { + "epoch": 0.85, + "learning_rate": 0.00012321765295887665, + "loss": 1.291, + "theoretical_loss": 3.3672344230085502, + "tokens_seen": 2797666304 + }, + { + "epoch": 0.85, + "learning_rate": 0.00012320160481444335, + "loss": 1.36, + "theoretical_loss": 3.367228324026062, + "tokens_seen": 2797731840 + }, + { + "epoch": 0.85, + "learning_rate": 0.00012318555667001005, + "loss": 1.2225, + "theoretical_loss": 3.36722222522644, + "tokens_seen": 2797797376 + }, + { + "epoch": 0.85, + "learning_rate": 0.00012316950852557675, + "loss": 1.3296, + "theoretical_loss": 3.3672161266096756, + "tokens_seen": 2797862912 + }, + { + "epoch": 0.85, + "learning_rate": 0.00012315346038114345, + "loss": 1.2965, + "theoretical_loss": 3.3672100281757578, + "tokens_seen": 2797928448 + }, + { + "epoch": 0.85, + "learning_rate": 0.00012313741223671012, + "loss": 1.2413, + "theoretical_loss": 3.367203929924678, + "tokens_seen": 2797993984 + }, + { + "epoch": 0.85, + "learning_rate": 0.00012312136409227682, + "loss": 1.362, + "theoretical_loss": 3.3671978318564255, + "tokens_seen": 2798059520 + }, + { + "epoch": 0.85, + "learning_rate": 0.00012310531594784353, + "loss": 1.2769, + "theoretical_loss": 3.367191733970991, + "tokens_seen": 2798125056 + }, + { + "epoch": 0.85, + "learning_rate": 0.00012308926780341023, + "loss": 1.2574, + "theoretical_loss": 3.3671856362683648, + "tokens_seen": 2798190592 + }, + { + "epoch": 0.85, + "learning_rate": 0.00012307321965897693, + "loss": 1.2157, + "theoretical_loss": 3.3671795387485366, + "tokens_seen": 2798256128 + }, + { + "epoch": 0.85, + "learning_rate": 0.00012305717151454363, + "loss": 1.2738, + "theoretical_loss": 3.3671734414114978, + "tokens_seen": 2798321664 + }, + { + "epoch": 0.85, + "objective/train/docs_used": 1016369, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.327946662902832, + "objective/train/theoretical_loss": 3.367167344257237, + "objective/train/tokens_used": 2818847200, + "theoretical_loss": 3.367167344257237, + "tokens_seen": 2798387200 + }, + { + "epoch": 0.85, + "learning_rate": 0.00012304112337011033, + "loss": 1.2658, + "theoretical_loss": 3.367167344257237, + "tokens_seen": 2798387200 + }, + { + "epoch": 0.85, + "learning_rate": 0.00012302507522567703, + "loss": 1.3214, + "theoretical_loss": 3.3671612472857455, + "tokens_seen": 2798452736 + }, + { + "epoch": 0.85, + "learning_rate": 0.00012300902708124373, + "loss": 1.2157, + "theoretical_loss": 3.3671551504970134, + "tokens_seen": 2798518272 + }, + { + "epoch": 0.85, + "learning_rate": 0.00012299297893681043, + "loss": 1.3117, + "theoretical_loss": 3.367149053891031, + "tokens_seen": 2798583808 + }, + { + "epoch": 0.85, + "learning_rate": 0.00012297693079237713, + "loss": 1.2759, + "theoretical_loss": 3.3671429574677885, + "tokens_seen": 2798649344 + }, + { + "epoch": 0.85, + "learning_rate": 0.00012296088264794383, + "loss": 1.1914, + "theoretical_loss": 3.367136861227276, + "tokens_seen": 2798714880 + }, + { + "epoch": 0.85, + "learning_rate": 0.00012294483450351053, + "loss": 1.2565, + "theoretical_loss": 3.3671307651694837, + "tokens_seen": 2798780416 + }, + { + "epoch": 0.85, + "learning_rate": 0.00012292878635907723, + "loss": 1.2192, + "theoretical_loss": 3.3671246692944017, + "tokens_seen": 2798845952 + }, + { + "epoch": 0.85, + "learning_rate": 0.00012291273821464393, + "loss": 1.1981, + "theoretical_loss": 3.367118573602021, + "tokens_seen": 2798911488 + }, + { + "epoch": 0.85, + "learning_rate": 0.00012289669007021063, + "loss": 1.2656, + "theoretical_loss": 3.367112478092331, + "tokens_seen": 2798977024 + }, + { + "epoch": 0.85, + "learning_rate": 0.00012288064192577733, + "loss": 1.2957, + "theoretical_loss": 3.3671063827653227, + "tokens_seen": 2799042560 + }, + { + "epoch": 0.85, + "learning_rate": 0.00012286459378134403, + "loss": 1.3521, + "theoretical_loss": 3.3671002876209855, + "tokens_seen": 2799108096 + }, + { + "epoch": 0.85, + "learning_rate": 0.00012284854563691074, + "loss": 1.23, + "theoretical_loss": 3.3670941926593105, + "tokens_seen": 2799173632 + }, + { + "epoch": 0.85, + "learning_rate": 0.00012283249749247744, + "loss": 1.32, + "theoretical_loss": 3.3670880978802873, + "tokens_seen": 2799239168 + }, + { + "epoch": 0.85, + "learning_rate": 0.00012281644934804414, + "loss": 1.3249, + "theoretical_loss": 3.3670820032839064, + "tokens_seen": 2799304704 + }, + { + "epoch": 0.85, + "learning_rate": 0.00012280040120361084, + "loss": 1.2186, + "theoretical_loss": 3.367075908870158, + "tokens_seen": 2799370240 + }, + { + "epoch": 0.85, + "learning_rate": 0.00012278435305917754, + "loss": 1.2669, + "theoretical_loss": 3.3670698146390325, + "tokens_seen": 2799435776 + }, + { + "epoch": 0.85, + "learning_rate": 0.00012276830491474424, + "loss": 1.1812, + "theoretical_loss": 3.36706372059052, + "tokens_seen": 2799501312 + }, + { + "epoch": 0.85, + "learning_rate": 0.00012275225677031094, + "loss": 1.3019, + "theoretical_loss": 3.367057626724611, + "tokens_seen": 2799566848 + }, + { + "epoch": 0.85, + "learning_rate": 0.00012273620862587764, + "loss": 1.1781, + "theoretical_loss": 3.367051533041295, + "tokens_seen": 2799632384 + }, + { + "epoch": 0.85, + "learning_rate": 0.00012272016048144434, + "loss": 1.2696, + "theoretical_loss": 3.367045439540563, + "tokens_seen": 2799697920 + }, + { + "epoch": 0.85, + "learning_rate": 0.00012270411233701104, + "loss": 1.3449, + "theoretical_loss": 3.3670393462224055, + "tokens_seen": 2799763456 + }, + { + "epoch": 0.85, + "learning_rate": 0.00012268806419257774, + "loss": 1.3249, + "theoretical_loss": 3.367033253086812, + "tokens_seen": 2799828992 + }, + { + "epoch": 0.85, + "learning_rate": 0.00012267201604814444, + "loss": 1.3165, + "theoretical_loss": 3.367027160133773, + "tokens_seen": 2799894528 + }, + { + "epoch": 0.85, + "learning_rate": 0.00012265596790371114, + "loss": 1.3748, + "theoretical_loss": 3.367021067363279, + "tokens_seen": 2799960064 + }, + { + "epoch": 0.85, + "learning_rate": 0.00012263991975927784, + "loss": 1.3508, + "theoretical_loss": 3.36701497477532, + "tokens_seen": 2800025600 + }, + { + "epoch": 0.85, + "learning_rate": 0.00012262387161484454, + "loss": 1.3231, + "theoretical_loss": 3.3670088823698867, + "tokens_seen": 2800091136 + }, + { + "epoch": 0.85, + "learning_rate": 0.00012260782347041124, + "loss": 1.2752, + "theoretical_loss": 3.3670027901469686, + "tokens_seen": 2800156672 + }, + { + "epoch": 0.85, + "learning_rate": 0.00012259177532597795, + "loss": 1.2587, + "theoretical_loss": 3.3669966981065564, + "tokens_seen": 2800222208 + }, + { + "epoch": 0.85, + "learning_rate": 0.00012257572718154465, + "loss": 1.2818, + "theoretical_loss": 3.3669906062486405, + "tokens_seen": 2800287744 + }, + { + "epoch": 0.85, + "learning_rate": 0.00012255967903711135, + "loss": 1.2085, + "theoretical_loss": 3.366984514573211, + "tokens_seen": 2800353280 + }, + { + "epoch": 0.85, + "learning_rate": 0.00012254363089267805, + "loss": 1.3104, + "theoretical_loss": 3.3669784230802584, + "tokens_seen": 2800418816 + }, + { + "epoch": 0.85, + "learning_rate": 0.00012252758274824475, + "loss": 1.3072, + "theoretical_loss": 3.366972331769772, + "tokens_seen": 2800484352 + }, + { + "epoch": 0.85, + "learning_rate": 0.00012251153460381145, + "loss": 1.2195, + "theoretical_loss": 3.3669662406417435, + "tokens_seen": 2800549888 + }, + { + "epoch": 0.85, + "learning_rate": 0.00012249548645937815, + "loss": 1.3051, + "theoretical_loss": 3.3669601496961623, + "tokens_seen": 2800615424 + }, + { + "epoch": 0.85, + "learning_rate": 0.00012247943831494485, + "loss": 1.1696, + "theoretical_loss": 3.3669540589330187, + "tokens_seen": 2800680960 + }, + { + "epoch": 0.85, + "learning_rate": 0.00012246339017051155, + "loss": 1.2148, + "theoretical_loss": 3.366947968352303, + "tokens_seen": 2800746496 + }, + { + "epoch": 0.85, + "learning_rate": 0.00012244734202607825, + "loss": 1.2838, + "theoretical_loss": 3.3669418779540057, + "tokens_seen": 2800812032 + }, + { + "epoch": 0.85, + "learning_rate": 0.00012243129388164492, + "loss": 1.2495, + "theoretical_loss": 3.3669357877381167, + "tokens_seen": 2800877568 + }, + { + "epoch": 0.85, + "learning_rate": 0.00012241524573721163, + "loss": 1.3409, + "theoretical_loss": 3.366929697704627, + "tokens_seen": 2800943104 + }, + { + "epoch": 0.85, + "learning_rate": 0.00012239919759277833, + "loss": 1.2724, + "theoretical_loss": 3.3669236078535256, + "tokens_seen": 2801008640 + }, + { + "epoch": 0.85, + "learning_rate": 0.00012238314944834503, + "loss": 1.3101, + "theoretical_loss": 3.3669175181848043, + "tokens_seen": 2801074176 + }, + { + "epoch": 0.85, + "learning_rate": 0.00012236710130391173, + "loss": 1.3412, + "theoretical_loss": 3.3669114286984523, + "tokens_seen": 2801139712 + }, + { + "epoch": 0.85, + "learning_rate": 0.00012235105315947843, + "loss": 1.1372, + "theoretical_loss": 3.36690533939446, + "tokens_seen": 2801205248 + }, + { + "epoch": 0.85, + "learning_rate": 0.00012233500501504513, + "loss": 1.2561, + "theoretical_loss": 3.3668992502728177, + "tokens_seen": 2801270784 + }, + { + "epoch": 0.85, + "learning_rate": 0.00012231895687061183, + "loss": 1.3791, + "theoretical_loss": 3.3668931613335165, + "tokens_seen": 2801336320 + }, + { + "epoch": 0.85, + "learning_rate": 0.00012230290872617853, + "loss": 1.2783, + "theoretical_loss": 3.3668870725765454, + "tokens_seen": 2801401856 + }, + { + "epoch": 0.85, + "learning_rate": 0.00012228686058174523, + "loss": 1.3727, + "theoretical_loss": 3.3668809840018956, + "tokens_seen": 2801467392 + }, + { + "epoch": 0.85, + "learning_rate": 0.00012227081243731193, + "loss": 1.3555, + "theoretical_loss": 3.366874895609557, + "tokens_seen": 2801532928 + }, + { + "epoch": 0.85, + "learning_rate": 0.00012225476429287863, + "loss": 1.2073, + "theoretical_loss": 3.36686880739952, + "tokens_seen": 2801598464 + }, + { + "epoch": 0.85, + "objective/train/docs_used": 1017536, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.2849243879318237, + "objective/train/theoretical_loss": 3.3668627193717744, + "objective/train/tokens_used": 2822124000, + "theoretical_loss": 3.3668627193717744, + "tokens_seen": 2801664000 + }, + { + "epoch": 0.85, + "learning_rate": 0.00012223871614844533, + "loss": 1.3022, + "theoretical_loss": 3.3668627193717744, + "tokens_seen": 2801664000 + }, + { + "epoch": 0.85, + "learning_rate": 0.00012222266800401203, + "loss": 1.3831, + "theoretical_loss": 3.366856631526311, + "tokens_seen": 2801729536 + }, + { + "epoch": 0.85, + "learning_rate": 0.00012220661985957876, + "loss": 1.4226, + "theoretical_loss": 3.36685054386312, + "tokens_seen": 2801795072 + }, + { + "epoch": 0.85, + "learning_rate": 0.00012219057171514546, + "loss": 1.3254, + "theoretical_loss": 3.366844456382192, + "tokens_seen": 2801860608 + }, + { + "epoch": 0.85, + "learning_rate": 0.00012217452357071216, + "loss": 1.2632, + "theoretical_loss": 3.3668383690835166, + "tokens_seen": 2801926144 + }, + { + "epoch": 0.85, + "learning_rate": 0.00012215847542627886, + "loss": 1.2836, + "theoretical_loss": 3.3668322819670844, + "tokens_seen": 2801991680 + }, + { + "epoch": 0.85, + "learning_rate": 0.00012214242728184556, + "loss": 1.3957, + "theoretical_loss": 3.3668261950328855, + "tokens_seen": 2802057216 + }, + { + "epoch": 0.85, + "learning_rate": 0.00012212637913741224, + "loss": 1.307, + "theoretical_loss": 3.3668201082809106, + "tokens_seen": 2802122752 + }, + { + "epoch": 0.85, + "learning_rate": 0.00012211033099297894, + "loss": 1.2906, + "theoretical_loss": 3.36681402171115, + "tokens_seen": 2802188288 + }, + { + "epoch": 0.85, + "learning_rate": 0.00012209428284854564, + "loss": 1.3096, + "theoretical_loss": 3.3668079353235933, + "tokens_seen": 2802253824 + }, + { + "epoch": 0.85, + "learning_rate": 0.00012207823470411234, + "loss": 1.3094, + "theoretical_loss": 3.3668018491182314, + "tokens_seen": 2802319360 + }, + { + "epoch": 0.85, + "learning_rate": 0.00012206218655967905, + "loss": 1.2677, + "theoretical_loss": 3.3667957630950545, + "tokens_seen": 2802384896 + }, + { + "epoch": 0.85, + "learning_rate": 0.00012204613841524575, + "loss": 1.2435, + "theoretical_loss": 3.3667896772540526, + "tokens_seen": 2802450432 + }, + { + "epoch": 0.85, + "learning_rate": 0.00012203009027081244, + "loss": 1.1685, + "theoretical_loss": 3.3667835915952162, + "tokens_seen": 2802515968 + }, + { + "epoch": 0.85, + "learning_rate": 0.00012201404212637914, + "loss": 1.2227, + "theoretical_loss": 3.3667775061185354, + "tokens_seen": 2802581504 + }, + { + "epoch": 0.85, + "learning_rate": 0.00012199799398194584, + "loss": 1.1957, + "theoretical_loss": 3.366771420824001, + "tokens_seen": 2802647040 + }, + { + "epoch": 0.85, + "learning_rate": 0.00012198194583751254, + "loss": 1.3025, + "theoretical_loss": 3.3667653357116025, + "tokens_seen": 2802712576 + }, + { + "epoch": 0.85, + "learning_rate": 0.00012196589769307924, + "loss": 1.1989, + "theoretical_loss": 3.366759250781331, + "tokens_seen": 2802778112 + }, + { + "epoch": 0.85, + "learning_rate": 0.00012194984954864594, + "loss": 1.3296, + "theoretical_loss": 3.3667531660331766, + "tokens_seen": 2802843648 + }, + { + "epoch": 0.85, + "learning_rate": 0.00012193380140421264, + "loss": 1.2866, + "theoretical_loss": 3.3667470814671288, + "tokens_seen": 2802909184 + }, + { + "epoch": 0.85, + "learning_rate": 0.00012191775325977934, + "loss": 1.239, + "theoretical_loss": 3.3667409970831788, + "tokens_seen": 2802974720 + }, + { + "epoch": 0.85, + "learning_rate": 0.00012190170511534605, + "loss": 1.3395, + "theoretical_loss": 3.3667349128813164, + "tokens_seen": 2803040256 + }, + { + "epoch": 0.85, + "learning_rate": 0.00012188565697091275, + "loss": 1.3651, + "theoretical_loss": 3.3667288288615325, + "tokens_seen": 2803105792 + }, + { + "epoch": 0.85, + "learning_rate": 0.00012186960882647945, + "loss": 1.2869, + "theoretical_loss": 3.3667227450238166, + "tokens_seen": 2803171328 + }, + { + "epoch": 0.85, + "learning_rate": 0.00012185356068204615, + "loss": 1.2856, + "theoretical_loss": 3.3667166613681596, + "tokens_seen": 2803236864 + }, + { + "epoch": 0.85, + "learning_rate": 0.00012183751253761283, + "loss": 1.3007, + "theoretical_loss": 3.3667105778945516, + "tokens_seen": 2803302400 + }, + { + "epoch": 0.85, + "learning_rate": 0.00012182146439317954, + "loss": 1.3119, + "theoretical_loss": 3.366704494602983, + "tokens_seen": 2803367936 + }, + { + "epoch": 0.85, + "learning_rate": 0.00012180541624874624, + "loss": 1.3193, + "theoretical_loss": 3.3666984114934437, + "tokens_seen": 2803433472 + }, + { + "epoch": 0.85, + "learning_rate": 0.00012178936810431294, + "loss": 1.3247, + "theoretical_loss": 3.3666923285659243, + "tokens_seen": 2803499008 + }, + { + "epoch": 0.85, + "learning_rate": 0.00012177331995987964, + "loss": 1.282, + "theoretical_loss": 3.366686245820415, + "tokens_seen": 2803564544 + }, + { + "epoch": 0.85, + "learning_rate": 0.00012175727181544634, + "loss": 1.2644, + "theoretical_loss": 3.3666801632569063, + "tokens_seen": 2803630080 + }, + { + "epoch": 0.85, + "learning_rate": 0.00012174122367101304, + "loss": 1.2536, + "theoretical_loss": 3.3666740808753883, + "tokens_seen": 2803695616 + }, + { + "epoch": 0.85, + "learning_rate": 0.00012172517552657974, + "loss": 1.199, + "theoretical_loss": 3.3666679986758514, + "tokens_seen": 2803761152 + }, + { + "epoch": 0.85, + "learning_rate": 0.00012170912738214644, + "loss": 1.2434, + "theoretical_loss": 3.366661916658286, + "tokens_seen": 2803826688 + }, + { + "epoch": 0.85, + "learning_rate": 0.00012169307923771314, + "loss": 1.2575, + "theoretical_loss": 3.366655834822682, + "tokens_seen": 2803892224 + }, + { + "epoch": 0.85, + "learning_rate": 0.00012167703109327984, + "loss": 1.3892, + "theoretical_loss": 3.36664975316903, + "tokens_seen": 2803957760 + }, + { + "epoch": 0.85, + "learning_rate": 0.00012166098294884654, + "loss": 1.3105, + "theoretical_loss": 3.3666436716973203, + "tokens_seen": 2804023296 + }, + { + "epoch": 0.85, + "learning_rate": 0.00012164493480441323, + "loss": 1.3097, + "theoretical_loss": 3.366637590407543, + "tokens_seen": 2804088832 + }, + { + "epoch": 0.85, + "learning_rate": 0.00012162888665997993, + "loss": 1.2485, + "theoretical_loss": 3.366631509299689, + "tokens_seen": 2804154368 + }, + { + "epoch": 0.85, + "learning_rate": 0.00012161283851554666, + "loss": 1.2494, + "theoretical_loss": 3.366625428373748, + "tokens_seen": 2804219904 + }, + { + "epoch": 0.85, + "learning_rate": 0.00012159679037111336, + "loss": 1.2329, + "theoretical_loss": 3.366619347629711, + "tokens_seen": 2804285440 + }, + { + "epoch": 0.85, + "learning_rate": 0.00012158074222668006, + "loss": 1.2409, + "theoretical_loss": 3.366613267067567, + "tokens_seen": 2804350976 + }, + { + "epoch": 0.85, + "learning_rate": 0.00012156469408224676, + "loss": 1.3575, + "theoretical_loss": 3.3666071866873075, + "tokens_seen": 2804416512 + }, + { + "epoch": 0.85, + "learning_rate": 0.00012154864593781346, + "loss": 1.3262, + "theoretical_loss": 3.3666011064889223, + "tokens_seen": 2804482048 + }, + { + "epoch": 0.85, + "learning_rate": 0.00012153259779338016, + "loss": 1.2425, + "theoretical_loss": 3.366595026472402, + "tokens_seen": 2804547584 + }, + { + "epoch": 0.85, + "learning_rate": 0.00012151654964894685, + "loss": 1.3281, + "theoretical_loss": 3.3665889466377363, + "tokens_seen": 2804613120 + }, + { + "epoch": 0.85, + "learning_rate": 0.00012150050150451355, + "loss": 1.3397, + "theoretical_loss": 3.3665828669849165, + "tokens_seen": 2804678656 + }, + { + "epoch": 0.85, + "learning_rate": 0.00012148445336008025, + "loss": 1.2651, + "theoretical_loss": 3.366576787513932, + "tokens_seen": 2804744192 + }, + { + "epoch": 0.85, + "learning_rate": 0.00012146840521564695, + "loss": 1.3111, + "theoretical_loss": 3.366570708224774, + "tokens_seen": 2804809728 + }, + { + "epoch": 0.85, + "learning_rate": 0.00012145235707121365, + "loss": 1.168, + "theoretical_loss": 3.366564629117432, + "tokens_seen": 2804875264 + }, + { + "epoch": 0.85, + "objective/train/docs_used": 1018816, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.3423362970352173, + "objective/train/theoretical_loss": 3.366558550191896, + "objective/train/tokens_used": 2825400800, + "theoretical_loss": 3.366558550191896, + "tokens_seen": 2804940800 + }, + { + "epoch": 0.85, + "learning_rate": 0.00012143630892678035, + "loss": 1.3321, + "theoretical_loss": 3.366558550191896, + "tokens_seen": 2804940800 + }, + { + "epoch": 0.85, + "learning_rate": 0.00012142026078234705, + "loss": 1.2093, + "theoretical_loss": 3.3665524714481574, + "tokens_seen": 2805006336 + }, + { + "epoch": 0.85, + "learning_rate": 0.00012140421263791375, + "loss": 1.3162, + "theoretical_loss": 3.3665463928862063, + "tokens_seen": 2805071872 + }, + { + "epoch": 0.85, + "learning_rate": 0.00012138816449348045, + "loss": 1.313, + "theoretical_loss": 3.3665403145060324, + "tokens_seen": 2805137408 + }, + { + "epoch": 0.85, + "learning_rate": 0.00012137211634904715, + "loss": 1.3484, + "theoretical_loss": 3.3665342363076265, + "tokens_seen": 2805202944 + }, + { + "epoch": 0.85, + "learning_rate": 0.00012135606820461385, + "loss": 1.2154, + "theoretical_loss": 3.366528158290979, + "tokens_seen": 2805268480 + }, + { + "epoch": 0.85, + "learning_rate": 0.00012134002006018055, + "loss": 1.2116, + "theoretical_loss": 3.36652208045608, + "tokens_seen": 2805334016 + }, + { + "epoch": 0.85, + "learning_rate": 0.00012132397191574724, + "loss": 1.3645, + "theoretical_loss": 3.3665160028029195, + "tokens_seen": 2805399552 + }, + { + "epoch": 0.85, + "learning_rate": 0.00012130792377131394, + "loss": 1.3238, + "theoretical_loss": 3.3665099253314885, + "tokens_seen": 2805465088 + }, + { + "epoch": 0.85, + "learning_rate": 0.00012129187562688064, + "loss": 1.3171, + "theoretical_loss": 3.3665038480417766, + "tokens_seen": 2805530624 + }, + { + "epoch": 0.85, + "learning_rate": 0.00012127582748244734, + "loss": 1.3217, + "theoretical_loss": 3.3664977709337744, + "tokens_seen": 2805596160 + }, + { + "epoch": 0.85, + "learning_rate": 0.00012125977933801404, + "loss": 1.2154, + "theoretical_loss": 3.3664916940074727, + "tokens_seen": 2805661696 + }, + { + "epoch": 0.85, + "learning_rate": 0.00012124373119358074, + "loss": 1.3757, + "theoretical_loss": 3.3664856172628617, + "tokens_seen": 2805727232 + }, + { + "epoch": 0.85, + "learning_rate": 0.00012122768304914745, + "loss": 1.2852, + "theoretical_loss": 3.366479540699931, + "tokens_seen": 2805792768 + }, + { + "epoch": 0.85, + "learning_rate": 0.00012121163490471415, + "loss": 1.3032, + "theoretical_loss": 3.3664734643186716, + "tokens_seen": 2805858304 + }, + { + "epoch": 0.85, + "learning_rate": 0.00012119558676028085, + "loss": 1.327, + "theoretical_loss": 3.3664673881190734, + "tokens_seen": 2805923840 + }, + { + "epoch": 0.85, + "learning_rate": 0.00012117953861584755, + "loss": 1.2782, + "theoretical_loss": 3.366461312101127, + "tokens_seen": 2805989376 + }, + { + "epoch": 0.85, + "learning_rate": 0.00012116349047141425, + "loss": 1.2824, + "theoretical_loss": 3.3664552362648226, + "tokens_seen": 2806054912 + }, + { + "epoch": 0.85, + "learning_rate": 0.00012114744232698095, + "loss": 1.2512, + "theoretical_loss": 3.3664491606101508, + "tokens_seen": 2806120448 + }, + { + "epoch": 0.85, + "learning_rate": 0.00012113139418254764, + "loss": 1.2214, + "theoretical_loss": 3.366443085137101, + "tokens_seen": 2806185984 + }, + { + "epoch": 0.85, + "learning_rate": 0.00012111534603811434, + "loss": 1.284, + "theoretical_loss": 3.3664370098456655, + "tokens_seen": 2806251520 + }, + { + "epoch": 0.85, + "learning_rate": 0.00012109929789368104, + "loss": 1.325, + "theoretical_loss": 3.3664309347358325, + "tokens_seen": 2806317056 + }, + { + "epoch": 0.85, + "learning_rate": 0.00012108324974924774, + "loss": 1.1823, + "theoretical_loss": 3.3664248598075934, + "tokens_seen": 2806382592 + }, + { + "epoch": 0.85, + "learning_rate": 0.00012106720160481444, + "loss": 1.283, + "theoretical_loss": 3.3664187850609384, + "tokens_seen": 2806448128 + }, + { + "epoch": 0.85, + "learning_rate": 0.00012105115346038114, + "loss": 1.2814, + "theoretical_loss": 3.366412710495858, + "tokens_seen": 2806513664 + }, + { + "epoch": 0.85, + "learning_rate": 0.00012103510531594784, + "loss": 1.3038, + "theoretical_loss": 3.366406636112342, + "tokens_seen": 2806579200 + }, + { + "epoch": 0.85, + "learning_rate": 0.00012101905717151457, + "loss": 1.1901, + "theoretical_loss": 3.366400561910381, + "tokens_seen": 2806644736 + }, + { + "epoch": 0.85, + "learning_rate": 0.00012100300902708125, + "loss": 1.2608, + "theoretical_loss": 3.366394487889966, + "tokens_seen": 2806710272 + }, + { + "epoch": 0.85, + "learning_rate": 0.00012098696088264795, + "loss": 1.2613, + "theoretical_loss": 3.366388414051086, + "tokens_seen": 2806775808 + }, + { + "epoch": 0.85, + "learning_rate": 0.00012097091273821466, + "loss": 1.3481, + "theoretical_loss": 3.366382340393732, + "tokens_seen": 2806841344 + }, + { + "epoch": 0.85, + "learning_rate": 0.00012095486459378136, + "loss": 1.3225, + "theoretical_loss": 3.3663762669178947, + "tokens_seen": 2806906880 + }, + { + "epoch": 0.85, + "learning_rate": 0.00012093881644934806, + "loss": 1.2792, + "theoretical_loss": 3.366370193623564, + "tokens_seen": 2806972416 + }, + { + "epoch": 0.85, + "learning_rate": 0.00012092276830491476, + "loss": 1.2551, + "theoretical_loss": 3.3663641205107306, + "tokens_seen": 2807037952 + }, + { + "epoch": 0.85, + "learning_rate": 0.00012090672016048146, + "loss": 1.2432, + "theoretical_loss": 3.3663580475793844, + "tokens_seen": 2807103488 + }, + { + "epoch": 0.85, + "learning_rate": 0.00012089067201604816, + "loss": 1.1862, + "theoretical_loss": 3.366351974829516, + "tokens_seen": 2807169024 + }, + { + "epoch": 0.85, + "learning_rate": 0.00012087462387161486, + "loss": 1.1732, + "theoretical_loss": 3.3663459022611155, + "tokens_seen": 2807234560 + }, + { + "epoch": 0.85, + "learning_rate": 0.00012085857572718156, + "loss": 1.1989, + "theoretical_loss": 3.366339829874174, + "tokens_seen": 2807300096 + }, + { + "epoch": 0.85, + "learning_rate": 0.00012084252758274826, + "loss": 1.3102, + "theoretical_loss": 3.3663337576686807, + "tokens_seen": 2807365632 + }, + { + "epoch": 0.85, + "learning_rate": 0.00012082647943831496, + "loss": 1.3649, + "theoretical_loss": 3.3663276856446265, + "tokens_seen": 2807431168 + }, + { + "epoch": 0.85, + "learning_rate": 0.00012081043129388165, + "loss": 1.3414, + "theoretical_loss": 3.3663216138020022, + "tokens_seen": 2807496704 + }, + { + "epoch": 0.85, + "learning_rate": 0.00012079438314944835, + "loss": 1.2328, + "theoretical_loss": 3.3663155421407973, + "tokens_seen": 2807562240 + }, + { + "epoch": 0.85, + "learning_rate": 0.00012077833500501505, + "loss": 1.3311, + "theoretical_loss": 3.3663094706610024, + "tokens_seen": 2807627776 + }, + { + "epoch": 0.85, + "learning_rate": 0.00012076228686058175, + "loss": 1.2205, + "theoretical_loss": 3.366303399362608, + "tokens_seen": 2807693312 + }, + { + "epoch": 0.85, + "learning_rate": 0.00012074623871614845, + "loss": 1.2619, + "theoretical_loss": 3.366297328245605, + "tokens_seen": 2807758848 + }, + { + "epoch": 0.85, + "learning_rate": 0.00012073019057171515, + "loss": 1.2639, + "theoretical_loss": 3.3662912573099826, + "tokens_seen": 2807824384 + }, + { + "epoch": 0.85, + "learning_rate": 0.00012071414242728185, + "loss": 1.3126, + "theoretical_loss": 3.366285186555732, + "tokens_seen": 2807889920 + }, + { + "epoch": 0.85, + "learning_rate": 0.00012069809428284855, + "loss": 1.3055, + "theoretical_loss": 3.3662791159828434, + "tokens_seen": 2807955456 + }, + { + "epoch": 0.85, + "learning_rate": 0.00012068204613841525, + "loss": 1.3538, + "theoretical_loss": 3.366273045591307, + "tokens_seen": 2808020992 + }, + { + "epoch": 0.85, + "learning_rate": 0.00012066599799398195, + "loss": 1.2483, + "theoretical_loss": 3.3662669753811127, + "tokens_seen": 2808086528 + }, + { + "epoch": 0.85, + "learning_rate": 0.00012064994984954865, + "loss": 1.3144, + "theoretical_loss": 3.3662609053522514, + "tokens_seen": 2808152064 + }, + { + "epoch": 0.85, + "objective/train/docs_used": 1020099, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.4596006870269775, + "objective/train/theoretical_loss": 3.366254835504714, + "objective/train/tokens_used": 2828677600, + "theoretical_loss": 3.366254835504714, + "tokens_seen": 2808217600 + }, + { + "epoch": 0.85, + "learning_rate": 0.00012063390170511536, + "loss": 1.3545, + "theoretical_loss": 3.366254835504714, + "tokens_seen": 2808217600 + }, + { + "epoch": 0.85, + "learning_rate": 0.00012061785356068204, + "loss": 1.2633, + "theoretical_loss": 3.3662487658384896, + "tokens_seen": 2808283136 + }, + { + "epoch": 0.85, + "learning_rate": 0.00012060180541624874, + "loss": 1.3424, + "theoretical_loss": 3.366242696353569, + "tokens_seen": 2808348672 + }, + { + "epoch": 0.85, + "learning_rate": 0.00012058575727181544, + "loss": 1.2575, + "theoretical_loss": 3.3662366270499433, + "tokens_seen": 2808414208 + }, + { + "epoch": 0.85, + "learning_rate": 0.00012056970912738214, + "loss": 1.1983, + "theoretical_loss": 3.366230557927602, + "tokens_seen": 2808479744 + }, + { + "epoch": 0.85, + "learning_rate": 0.00012055366098294884, + "loss": 1.2787, + "theoretical_loss": 3.366224488986536, + "tokens_seen": 2808545280 + }, + { + "epoch": 0.85, + "learning_rate": 0.00012053761283851555, + "loss": 1.3757, + "theoretical_loss": 3.366218420226735, + "tokens_seen": 2808610816 + }, + { + "epoch": 0.85, + "learning_rate": 0.00012052156469408225, + "loss": 1.3364, + "theoretical_loss": 3.3662123516481905, + "tokens_seen": 2808676352 + }, + { + "epoch": 0.85, + "learning_rate": 0.00012050551654964895, + "loss": 1.2608, + "theoretical_loss": 3.366206283250891, + "tokens_seen": 2808741888 + }, + { + "epoch": 0.85, + "learning_rate": 0.00012048946840521565, + "loss": 1.2505, + "theoretical_loss": 3.366200215034829, + "tokens_seen": 2808807424 + }, + { + "epoch": 0.85, + "learning_rate": 0.00012047342026078235, + "loss": 1.2067, + "theoretical_loss": 3.3661941469999936, + "tokens_seen": 2808872960 + }, + { + "epoch": 0.85, + "learning_rate": 0.00012045737211634905, + "loss": 1.2474, + "theoretical_loss": 3.366188079146375, + "tokens_seen": 2808938496 + }, + { + "epoch": 0.85, + "learning_rate": 0.00012044132397191575, + "loss": 1.249, + "theoretical_loss": 3.3661820114739642, + "tokens_seen": 2809004032 + }, + { + "epoch": 0.85, + "learning_rate": 0.00012042527582748244, + "loss": 1.2938, + "theoretical_loss": 3.3661759439827508, + "tokens_seen": 2809069568 + }, + { + "epoch": 0.85, + "learning_rate": 0.00012040922768304916, + "loss": 1.401, + "theoretical_loss": 3.366169876672726, + "tokens_seen": 2809135104 + }, + { + "epoch": 0.85, + "learning_rate": 0.00012039317953861586, + "loss": 1.2434, + "theoretical_loss": 3.36616380954388, + "tokens_seen": 2809200640 + }, + { + "epoch": 0.85, + "learning_rate": 0.00012037713139418257, + "loss": 1.1432, + "theoretical_loss": 3.366157742596203, + "tokens_seen": 2809266176 + }, + { + "epoch": 0.85, + "learning_rate": 0.00012036108324974927, + "loss": 1.2841, + "theoretical_loss": 3.3661516758296854, + "tokens_seen": 2809331712 + }, + { + "epoch": 0.85, + "learning_rate": 0.00012034503510531597, + "loss": 1.3358, + "theoretical_loss": 3.3661456092443176, + "tokens_seen": 2809397248 + }, + { + "epoch": 0.85, + "learning_rate": 0.00012032898696088267, + "loss": 1.3201, + "theoretical_loss": 3.3661395428400898, + "tokens_seen": 2809462784 + }, + { + "epoch": 0.85, + "learning_rate": 0.00012031293881644935, + "loss": 1.2966, + "theoretical_loss": 3.366133476616992, + "tokens_seen": 2809528320 + }, + { + "epoch": 0.85, + "learning_rate": 0.00012029689067201605, + "loss": 1.2934, + "theoretical_loss": 3.3661274105750154, + "tokens_seen": 2809593856 + }, + { + "epoch": 0.85, + "learning_rate": 0.00012028084252758276, + "loss": 1.3187, + "theoretical_loss": 3.3661213447141503, + "tokens_seen": 2809659392 + }, + { + "epoch": 0.85, + "learning_rate": 0.00012026479438314946, + "loss": 1.2451, + "theoretical_loss": 3.3661152790343865, + "tokens_seen": 2809724928 + }, + { + "epoch": 0.85, + "learning_rate": 0.00012024874623871616, + "loss": 1.1519, + "theoretical_loss": 3.3661092135357142, + "tokens_seen": 2809790464 + }, + { + "epoch": 0.85, + "learning_rate": 0.00012023269809428286, + "loss": 1.3609, + "theoretical_loss": 3.3661031482181247, + "tokens_seen": 2809856000 + }, + { + "epoch": 0.85, + "learning_rate": 0.00012021664994984956, + "loss": 1.1798, + "theoretical_loss": 3.3660970830816077, + "tokens_seen": 2809921536 + }, + { + "epoch": 0.85, + "learning_rate": 0.00012020060180541626, + "loss": 1.2513, + "theoretical_loss": 3.366091018126154, + "tokens_seen": 2809987072 + }, + { + "epoch": 0.85, + "learning_rate": 0.00012018455366098296, + "loss": 1.1642, + "theoretical_loss": 3.366084953351753, + "tokens_seen": 2810052608 + }, + { + "epoch": 0.85, + "learning_rate": 0.00012016850551654966, + "loss": 1.2193, + "theoretical_loss": 3.3660788887583966, + "tokens_seen": 2810118144 + }, + { + "epoch": 0.85, + "learning_rate": 0.00012015245737211636, + "loss": 1.3264, + "theoretical_loss": 3.366072824346074, + "tokens_seen": 2810183680 + }, + { + "epoch": 0.85, + "learning_rate": 0.00012013640922768306, + "loss": 1.2406, + "theoretical_loss": 3.3660667601147756, + "tokens_seen": 2810249216 + }, + { + "epoch": 0.85, + "learning_rate": 0.00012012036108324975, + "loss": 1.2081, + "theoretical_loss": 3.3660606960644928, + "tokens_seen": 2810314752 + }, + { + "epoch": 0.85, + "learning_rate": 0.00012010431293881645, + "loss": 1.2452, + "theoretical_loss": 3.366054632195215, + "tokens_seen": 2810380288 + }, + { + "epoch": 0.85, + "learning_rate": 0.00012008826479438315, + "loss": 1.2412, + "theoretical_loss": 3.3660485685069323, + "tokens_seen": 2810445824 + }, + { + "epoch": 0.85, + "learning_rate": 0.00012007221664994985, + "loss": 1.1158, + "theoretical_loss": 3.3660425049996365, + "tokens_seen": 2810511360 + }, + { + "epoch": 0.85, + "learning_rate": 0.00012005616850551655, + "loss": 1.2818, + "theoretical_loss": 3.366036441673317, + "tokens_seen": 2810576896 + }, + { + "epoch": 0.85, + "learning_rate": 0.00012004012036108325, + "loss": 1.1825, + "theoretical_loss": 3.3660303785279635, + "tokens_seen": 2810642432 + }, + { + "epoch": 0.85, + "learning_rate": 0.00012002407221664995, + "loss": 1.2804, + "theoretical_loss": 3.3660243155635676, + "tokens_seen": 2810707968 + }, + { + "epoch": 0.85, + "learning_rate": 0.00012000802407221665, + "loss": 1.2671, + "theoretical_loss": 3.3660182527801195, + "tokens_seen": 2810773504 + }, + { + "epoch": 0.85, + "learning_rate": 0.00011999197592778335, + "loss": 1.2696, + "theoretical_loss": 3.366012190177609, + "tokens_seen": 2810839040 + }, + { + "epoch": 0.85, + "learning_rate": 0.00011997592778335005, + "loss": 1.2992, + "theoretical_loss": 3.3660061277560276, + "tokens_seen": 2810904576 + }, + { + "epoch": 0.85, + "learning_rate": 0.00011995987963891675, + "loss": 1.3073, + "theoretical_loss": 3.366000065515364, + "tokens_seen": 2810970112 + }, + { + "epoch": 0.85, + "learning_rate": 0.00011994383149448346, + "loss": 1.2881, + "theoretical_loss": 3.3659940034556097, + "tokens_seen": 2811035648 + }, + { + "epoch": 0.85, + "learning_rate": 0.00011992778335005014, + "loss": 1.2989, + "theoretical_loss": 3.365987941576755, + "tokens_seen": 2811101184 + }, + { + "epoch": 0.85, + "learning_rate": 0.00011991173520561684, + "loss": 1.3757, + "theoretical_loss": 3.36598187987879, + "tokens_seen": 2811166720 + }, + { + "epoch": 0.85, + "learning_rate": 0.00011989568706118354, + "loss": 1.3794, + "theoretical_loss": 3.3659758183617057, + "tokens_seen": 2811232256 + }, + { + "epoch": 0.85, + "learning_rate": 0.00011987963891675024, + "loss": 1.357, + "theoretical_loss": 3.3659697570254914, + "tokens_seen": 2811297792 + }, + { + "epoch": 0.85, + "learning_rate": 0.00011986359077231694, + "loss": 1.3748, + "theoretical_loss": 3.365963695870138, + "tokens_seen": 2811363328 + }, + { + "epoch": 0.85, + "learning_rate": 0.00011984754262788365, + "loss": 1.2799, + "theoretical_loss": 3.365957634895637, + "tokens_seen": 2811428864 + }, + { + "epoch": 0.85, + "objective/train/docs_used": 1021297, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.2894008159637451, + "objective/train/theoretical_loss": 3.365951574101977, + "objective/train/tokens_used": 2831954400, + "theoretical_loss": 3.365951574101977, + "tokens_seen": 2811494400 + }, + { + "epoch": 0.85, + "learning_rate": 0.00011983149448345035, + "loss": 1.2424, + "theoretical_loss": 3.365951574101977, + "tokens_seen": 2811494400 + }, + { + "epoch": 0.85, + "learning_rate": 0.00011981544633901707, + "loss": 1.3147, + "theoretical_loss": 3.365945513489149, + "tokens_seen": 2811559936 + }, + { + "epoch": 0.85, + "learning_rate": 0.00011979939819458376, + "loss": 1.2631, + "theoretical_loss": 3.365939453057144, + "tokens_seen": 2811625472 + }, + { + "epoch": 0.85, + "learning_rate": 0.00011978335005015046, + "loss": 1.3173, + "theoretical_loss": 3.365933392805952, + "tokens_seen": 2811691008 + }, + { + "epoch": 0.85, + "learning_rate": 0.00011976730190571716, + "loss": 1.2653, + "theoretical_loss": 3.3659273327355628, + "tokens_seen": 2811756544 + }, + { + "epoch": 0.85, + "learning_rate": 0.00011975125376128386, + "loss": 1.2237, + "theoretical_loss": 3.3659212728459678, + "tokens_seen": 2811822080 + }, + { + "epoch": 0.85, + "learning_rate": 0.00011973520561685056, + "loss": 1.2297, + "theoretical_loss": 3.365915213137157, + "tokens_seen": 2811887616 + }, + { + "epoch": 0.85, + "learning_rate": 0.00011971915747241726, + "loss": 1.239, + "theoretical_loss": 3.3659091536091204, + "tokens_seen": 2811953152 + }, + { + "epoch": 0.85, + "learning_rate": 0.00011970310932798396, + "loss": 1.2792, + "theoretical_loss": 3.365903094261849, + "tokens_seen": 2812018688 + }, + { + "epoch": 0.85, + "learning_rate": 0.00011968706118355067, + "loss": 1.3125, + "theoretical_loss": 3.365897035095333, + "tokens_seen": 2812084224 + }, + { + "epoch": 0.85, + "learning_rate": 0.00011967101303911737, + "loss": 1.3347, + "theoretical_loss": 3.3658909761095623, + "tokens_seen": 2812149760 + }, + { + "epoch": 0.85, + "learning_rate": 0.00011965496489468407, + "loss": 1.237, + "theoretical_loss": 3.3658849173045278, + "tokens_seen": 2812215296 + }, + { + "epoch": 0.85, + "learning_rate": 0.00011963891675025077, + "loss": 1.2231, + "theoretical_loss": 3.3658788586802197, + "tokens_seen": 2812280832 + }, + { + "epoch": 0.85, + "learning_rate": 0.00011962286860581747, + "loss": 1.3116, + "theoretical_loss": 3.3658728002366285, + "tokens_seen": 2812346368 + }, + { + "epoch": 0.85, + "learning_rate": 0.00011960682046138416, + "loss": 1.3629, + "theoretical_loss": 3.365866741973745, + "tokens_seen": 2812411904 + }, + { + "epoch": 0.85, + "learning_rate": 0.00011959077231695086, + "loss": 1.1881, + "theoretical_loss": 3.365860683891559, + "tokens_seen": 2812477440 + }, + { + "epoch": 0.85, + "learning_rate": 0.00011957472417251756, + "loss": 1.287, + "theoretical_loss": 3.3658546259900612, + "tokens_seen": 2812542976 + }, + { + "epoch": 0.85, + "learning_rate": 0.00011955867602808426, + "loss": 1.2439, + "theoretical_loss": 3.3658485682692416, + "tokens_seen": 2812608512 + }, + { + "epoch": 0.85, + "learning_rate": 0.00011954262788365096, + "loss": 1.2746, + "theoretical_loss": 3.365842510729091, + "tokens_seen": 2812674048 + }, + { + "epoch": 0.85, + "learning_rate": 0.00011952657973921766, + "loss": 1.335, + "theoretical_loss": 3.3658364533696, + "tokens_seen": 2812739584 + }, + { + "epoch": 0.85, + "learning_rate": 0.00011951053159478436, + "loss": 1.3106, + "theoretical_loss": 3.3658303961907583, + "tokens_seen": 2812805120 + }, + { + "epoch": 0.85, + "learning_rate": 0.00011949448345035106, + "loss": 1.2902, + "theoretical_loss": 3.3658243391925566, + "tokens_seen": 2812870656 + }, + { + "epoch": 0.85, + "learning_rate": 0.00011947843530591776, + "loss": 1.26, + "theoretical_loss": 3.3658182823749856, + "tokens_seen": 2812936192 + }, + { + "epoch": 0.85, + "learning_rate": 0.00011946238716148446, + "loss": 1.3077, + "theoretical_loss": 3.365812225738036, + "tokens_seen": 2813001728 + }, + { + "epoch": 0.85, + "learning_rate": 0.00011944633901705116, + "loss": 1.2848, + "theoretical_loss": 3.365806169281697, + "tokens_seen": 2813067264 + }, + { + "epoch": 0.85, + "learning_rate": 0.00011943029087261786, + "loss": 1.3428, + "theoretical_loss": 3.3658001130059603, + "tokens_seen": 2813132800 + }, + { + "epoch": 0.85, + "learning_rate": 0.00011941424272818455, + "loss": 1.2634, + "theoretical_loss": 3.3657940569108153, + "tokens_seen": 2813198336 + }, + { + "epoch": 0.85, + "learning_rate": 0.00011939819458375125, + "loss": 1.2924, + "theoretical_loss": 3.365788000996253, + "tokens_seen": 2813263872 + }, + { + "epoch": 0.85, + "learning_rate": 0.00011938214643931795, + "loss": 1.3022, + "theoretical_loss": 3.3657819452622637, + "tokens_seen": 2813329408 + }, + { + "epoch": 0.85, + "learning_rate": 0.00011936609829488465, + "loss": 1.2729, + "theoretical_loss": 3.3657758897088375, + "tokens_seen": 2813394944 + }, + { + "epoch": 0.85, + "learning_rate": 0.00011935005015045135, + "loss": 1.3285, + "theoretical_loss": 3.3657698343359654, + "tokens_seen": 2813460480 + }, + { + "epoch": 0.85, + "learning_rate": 0.00011933400200601805, + "loss": 1.257, + "theoretical_loss": 3.3657637791436374, + "tokens_seen": 2813526016 + }, + { + "epoch": 0.85, + "learning_rate": 0.00011931795386158475, + "loss": 1.2119, + "theoretical_loss": 3.3657577241318437, + "tokens_seen": 2813591552 + }, + { + "epoch": 0.85, + "learning_rate": 0.00011930190571715145, + "loss": 1.3101, + "theoretical_loss": 3.3657516693005753, + "tokens_seen": 2813657088 + }, + { + "epoch": 0.85, + "learning_rate": 0.00011928585757271815, + "loss": 1.211, + "theoretical_loss": 3.365745614649822, + "tokens_seen": 2813722624 + }, + { + "epoch": 0.85, + "learning_rate": 0.00011926980942828485, + "loss": 1.2313, + "theoretical_loss": 3.365739560179575, + "tokens_seen": 2813788160 + }, + { + "epoch": 0.85, + "learning_rate": 0.00011925376128385156, + "loss": 1.2656, + "theoretical_loss": 3.3657335058898243, + "tokens_seen": 2813853696 + }, + { + "epoch": 0.85, + "learning_rate": 0.00011923771313941826, + "loss": 1.3203, + "theoretical_loss": 3.36572745178056, + "tokens_seen": 2813919232 + }, + { + "epoch": 0.85, + "learning_rate": 0.00011922166499498494, + "loss": 1.1893, + "theoretical_loss": 3.365721397851773, + "tokens_seen": 2813984768 + }, + { + "epoch": 0.85, + "learning_rate": 0.00011920561685055167, + "loss": 1.2379, + "theoretical_loss": 3.365715344103453, + "tokens_seen": 2814050304 + }, + { + "epoch": 0.85, + "learning_rate": 0.00011918956870611837, + "loss": 1.2774, + "theoretical_loss": 3.3657092905355914, + "tokens_seen": 2814115840 + }, + { + "epoch": 0.85, + "learning_rate": 0.00011917352056168507, + "loss": 1.3355, + "theoretical_loss": 3.365703237148178, + "tokens_seen": 2814181376 + }, + { + "epoch": 0.85, + "learning_rate": 0.00011915747241725177, + "loss": 1.1875, + "theoretical_loss": 3.365697183941203, + "tokens_seen": 2814246912 + }, + { + "epoch": 0.85, + "learning_rate": 0.00011914142427281847, + "loss": 1.3117, + "theoretical_loss": 3.3656911309146578, + "tokens_seen": 2814312448 + }, + { + "epoch": 0.85, + "learning_rate": 0.00011912537612838517, + "loss": 1.2703, + "theoretical_loss": 3.3656850780685317, + "tokens_seen": 2814377984 + }, + { + "epoch": 0.85, + "learning_rate": 0.00011910932798395187, + "loss": 1.2804, + "theoretical_loss": 3.365679025402816, + "tokens_seen": 2814443520 + }, + { + "epoch": 0.85, + "learning_rate": 0.00011909327983951856, + "loss": 1.4396, + "theoretical_loss": 3.3656729729175003, + "tokens_seen": 2814509056 + }, + { + "epoch": 0.85, + "learning_rate": 0.00011907723169508526, + "loss": 1.2204, + "theoretical_loss": 3.3656669206125756, + "tokens_seen": 2814574592 + }, + { + "epoch": 0.85, + "learning_rate": 0.00011906118355065196, + "loss": 1.3122, + "theoretical_loss": 3.3656608684880323, + "tokens_seen": 2814640128 + }, + { + "epoch": 0.85, + "learning_rate": 0.00011904513540621866, + "loss": 1.2954, + "theoretical_loss": 3.3656548165438607, + "tokens_seen": 2814705664 + }, + { + "epoch": 0.85, + "objective/train/docs_used": 1022508, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.217341661453247, + "objective/train/theoretical_loss": 3.365648764780051, + "objective/train/tokens_used": 2835231200, + "theoretical_loss": 3.365648764780051, + "tokens_seen": 2814771200 + }, + { + "epoch": 0.85, + "learning_rate": 0.00011902908726178536, + "loss": 1.3389, + "theoretical_loss": 3.365648764780051, + "tokens_seen": 2814771200 + }, + { + "epoch": 0.85, + "learning_rate": 0.00011901303911735206, + "loss": 1.3301, + "theoretical_loss": 3.3656427131965945, + "tokens_seen": 2814836736 + }, + { + "epoch": 0.85, + "learning_rate": 0.00011899699097291877, + "loss": 1.2235, + "theoretical_loss": 3.3656366617934803, + "tokens_seen": 2814902272 + }, + { + "epoch": 0.85, + "learning_rate": 0.00011898094282848547, + "loss": 1.2989, + "theoretical_loss": 3.3656306105706997, + "tokens_seen": 2814967808 + }, + { + "epoch": 0.85, + "learning_rate": 0.00011896489468405217, + "loss": 1.2548, + "theoretical_loss": 3.365624559528243, + "tokens_seen": 2815033344 + }, + { + "epoch": 0.85, + "learning_rate": 0.00011894884653961887, + "loss": 1.1981, + "theoretical_loss": 3.3656185086661004, + "tokens_seen": 2815098880 + }, + { + "epoch": 0.85, + "learning_rate": 0.00011893279839518557, + "loss": 1.2736, + "theoretical_loss": 3.3656124579842626, + "tokens_seen": 2815164416 + }, + { + "epoch": 0.85, + "learning_rate": 0.00011891675025075227, + "loss": 1.3346, + "theoretical_loss": 3.36560640748272, + "tokens_seen": 2815229952 + }, + { + "epoch": 0.85, + "learning_rate": 0.00011890070210631896, + "loss": 1.2332, + "theoretical_loss": 3.3656003571614628, + "tokens_seen": 2815295488 + }, + { + "epoch": 0.85, + "learning_rate": 0.00011888465396188566, + "loss": 1.2381, + "theoretical_loss": 3.3655943070204817, + "tokens_seen": 2815361024 + }, + { + "epoch": 0.85, + "learning_rate": 0.00011886860581745236, + "loss": 1.2636, + "theoretical_loss": 3.365588257059767, + "tokens_seen": 2815426560 + }, + { + "epoch": 0.85, + "learning_rate": 0.00011885255767301906, + "loss": 1.277, + "theoretical_loss": 3.365582207279309, + "tokens_seen": 2815492096 + }, + { + "epoch": 0.85, + "learning_rate": 0.00011883650952858576, + "loss": 1.3499, + "theoretical_loss": 3.3655761576790986, + "tokens_seen": 2815557632 + }, + { + "epoch": 0.85, + "learning_rate": 0.00011882046138415246, + "loss": 1.3557, + "theoretical_loss": 3.365570108259126, + "tokens_seen": 2815623168 + }, + { + "epoch": 0.85, + "learning_rate": 0.00011880441323971916, + "loss": 1.3556, + "theoretical_loss": 3.365564059019381, + "tokens_seen": 2815688704 + }, + { + "epoch": 0.85, + "learning_rate": 0.00011878836509528586, + "loss": 1.2586, + "theoretical_loss": 3.365558009959855, + "tokens_seen": 2815754240 + }, + { + "epoch": 0.85, + "learning_rate": 0.00011877231695085256, + "loss": 1.2581, + "theoretical_loss": 3.365551961080538, + "tokens_seen": 2815819776 + }, + { + "epoch": 0.85, + "learning_rate": 0.00011875626880641926, + "loss": 1.2459, + "theoretical_loss": 3.36554591238142, + "tokens_seen": 2815885312 + }, + { + "epoch": 0.85, + "learning_rate": 0.00011874022066198596, + "loss": 1.2715, + "theoretical_loss": 3.3655398638624927, + "tokens_seen": 2815950848 + }, + { + "epoch": 0.85, + "learning_rate": 0.00011872417251755266, + "loss": 1.3858, + "theoretical_loss": 3.3655338155237455, + "tokens_seen": 2816016384 + }, + { + "epoch": 0.85, + "learning_rate": 0.00011870812437311935, + "loss": 1.3521, + "theoretical_loss": 3.3655277673651685, + "tokens_seen": 2816081920 + }, + { + "epoch": 0.85, + "learning_rate": 0.00011869207622868605, + "loss": 1.3802, + "theoretical_loss": 3.3655217193867535, + "tokens_seen": 2816147456 + }, + { + "epoch": 0.85, + "learning_rate": 0.00011867602808425275, + "loss": 1.2079, + "theoretical_loss": 3.3655156715884895, + "tokens_seen": 2816212992 + }, + { + "epoch": 0.85, + "learning_rate": 0.00011865997993981945, + "loss": 1.3315, + "theoretical_loss": 3.365509623970368, + "tokens_seen": 2816278528 + }, + { + "epoch": 0.85, + "learning_rate": 0.00011864393179538615, + "loss": 1.3501, + "theoretical_loss": 3.365503576532379, + "tokens_seen": 2816344064 + }, + { + "epoch": 0.85, + "learning_rate": 0.00011862788365095285, + "loss": 1.2866, + "theoretical_loss": 3.365497529274513, + "tokens_seen": 2816409600 + }, + { + "epoch": 0.85, + "learning_rate": 0.00011861183550651958, + "loss": 1.328, + "theoretical_loss": 3.3654914821967603, + "tokens_seen": 2816475136 + }, + { + "epoch": 0.85, + "learning_rate": 0.00011859578736208628, + "loss": 1.2546, + "theoretical_loss": 3.3654854352991115, + "tokens_seen": 2816540672 + }, + { + "epoch": 0.85, + "learning_rate": 0.00011857973921765297, + "loss": 1.2884, + "theoretical_loss": 3.3654793885815573, + "tokens_seen": 2816606208 + }, + { + "epoch": 0.85, + "learning_rate": 0.00011856369107321967, + "loss": 1.2319, + "theoretical_loss": 3.3654733420440874, + "tokens_seen": 2816671744 + }, + { + "epoch": 0.85, + "learning_rate": 0.00011854764292878637, + "loss": 1.2688, + "theoretical_loss": 3.365467295686693, + "tokens_seen": 2816737280 + }, + { + "epoch": 0.85, + "learning_rate": 0.00011853159478435307, + "loss": 1.3307, + "theoretical_loss": 3.365461249509364, + "tokens_seen": 2816802816 + }, + { + "epoch": 0.85, + "learning_rate": 0.00011851554663991977, + "loss": 1.3013, + "theoretical_loss": 3.3654552035120915, + "tokens_seen": 2816868352 + }, + { + "epoch": 0.85, + "learning_rate": 0.00011849949849548647, + "loss": 1.3381, + "theoretical_loss": 3.3654491576948655, + "tokens_seen": 2816933888 + }, + { + "epoch": 0.85, + "learning_rate": 0.00011848345035105317, + "loss": 1.334, + "theoretical_loss": 3.3654431120576764, + "tokens_seen": 2816999424 + }, + { + "epoch": 0.85, + "learning_rate": 0.00011846740220661987, + "loss": 1.2798, + "theoretical_loss": 3.365437066600515, + "tokens_seen": 2817064960 + }, + { + "epoch": 0.85, + "learning_rate": 0.00011845135406218657, + "loss": 1.3214, + "theoretical_loss": 3.365431021323371, + "tokens_seen": 2817130496 + }, + { + "epoch": 0.85, + "learning_rate": 0.00011843530591775327, + "loss": 1.3562, + "theoretical_loss": 3.365424976226236, + "tokens_seen": 2817196032 + }, + { + "epoch": 0.85, + "learning_rate": 0.00011841925777331997, + "loss": 1.2937, + "theoretical_loss": 3.365418931309099, + "tokens_seen": 2817261568 + }, + { + "epoch": 0.85, + "learning_rate": 0.00011840320962888668, + "loss": 1.2696, + "theoretical_loss": 3.365412886571952, + "tokens_seen": 2817327104 + }, + { + "epoch": 0.85, + "learning_rate": 0.00011838716148445336, + "loss": 1.3714, + "theoretical_loss": 3.3654068420147842, + "tokens_seen": 2817392640 + }, + { + "epoch": 0.85, + "learning_rate": 0.00011837111334002006, + "loss": 1.2604, + "theoretical_loss": 3.3654007976375873, + "tokens_seen": 2817458176 + }, + { + "epoch": 0.85, + "learning_rate": 0.00011835506519558676, + "loss": 1.3604, + "theoretical_loss": 3.3653947534403503, + "tokens_seen": 2817523712 + }, + { + "epoch": 0.85, + "learning_rate": 0.00011833901705115346, + "loss": 1.3108, + "theoretical_loss": 3.3653887094230646, + "tokens_seen": 2817589248 + }, + { + "epoch": 0.85, + "learning_rate": 0.00011832296890672017, + "loss": 1.3431, + "theoretical_loss": 3.3653826655857206, + "tokens_seen": 2817654784 + }, + { + "epoch": 0.85, + "learning_rate": 0.00011830692076228687, + "loss": 1.2724, + "theoretical_loss": 3.3653766219283083, + "tokens_seen": 2817720320 + }, + { + "epoch": 0.85, + "learning_rate": 0.00011829087261785357, + "loss": 1.3281, + "theoretical_loss": 3.3653705784508188, + "tokens_seen": 2817785856 + }, + { + "epoch": 0.85, + "learning_rate": 0.00011827482447342027, + "loss": 1.3178, + "theoretical_loss": 3.365364535153242, + "tokens_seen": 2817851392 + }, + { + "epoch": 0.85, + "learning_rate": 0.00011825877632898697, + "loss": 1.3267, + "theoretical_loss": 3.365358492035569, + "tokens_seen": 2817916928 + }, + { + "epoch": 0.85, + "learning_rate": 0.00011824272818455367, + "loss": 1.249, + "theoretical_loss": 3.3653524490977893, + "tokens_seen": 2817982464 + }, + { + "epoch": 0.85, + "objective/train/docs_used": 1024155, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.2416443824768066, + "objective/train/theoretical_loss": 3.3653464063398943, + "objective/train/tokens_used": 2838508000, + "theoretical_loss": 3.3653464063398943, + "tokens_seen": 2818048000 + }, + { + "epoch": 0.85, + "learning_rate": 0.00011822668004012037, + "loss": 1.2666, + "theoretical_loss": 3.3653464063398943, + "tokens_seen": 2818048000 + }, + { + "epoch": 0.85, + "learning_rate": 0.00011821063189568707, + "loss": 1.2705, + "theoretical_loss": 3.3653403637618737, + "tokens_seen": 2818113536 + }, + { + "epoch": 0.85, + "learning_rate": 0.00011819458375125376, + "loss": 1.2423, + "theoretical_loss": 3.3653343213637186, + "tokens_seen": 2818179072 + }, + { + "epoch": 0.85, + "learning_rate": 0.00011817853560682046, + "loss": 1.3835, + "theoretical_loss": 3.365328279145419, + "tokens_seen": 2818244608 + }, + { + "epoch": 0.85, + "learning_rate": 0.00011816248746238716, + "loss": 1.3529, + "theoretical_loss": 3.365322237106966, + "tokens_seen": 2818310144 + }, + { + "epoch": 0.85, + "learning_rate": 0.00011814643931795386, + "loss": 1.2732, + "theoretical_loss": 3.365316195248349, + "tokens_seen": 2818375680 + }, + { + "epoch": 0.85, + "learning_rate": 0.00011813039117352056, + "loss": 1.3692, + "theoretical_loss": 3.3653101535695598, + "tokens_seen": 2818441216 + }, + { + "epoch": 0.85, + "learning_rate": 0.00011811434302908726, + "loss": 1.2224, + "theoretical_loss": 3.3653041120705875, + "tokens_seen": 2818506752 + }, + { + "epoch": 0.85, + "learning_rate": 0.00011809829488465396, + "loss": 1.2742, + "theoretical_loss": 3.3652980707514235, + "tokens_seen": 2818572288 + }, + { + "epoch": 0.85, + "learning_rate": 0.00011808224674022066, + "loss": 1.2413, + "theoretical_loss": 3.365292029612058, + "tokens_seen": 2818637824 + }, + { + "epoch": 0.85, + "learning_rate": 0.00011806619859578736, + "loss": 1.4333, + "theoretical_loss": 3.3652859886524813, + "tokens_seen": 2818703360 + }, + { + "epoch": 0.85, + "learning_rate": 0.00011805015045135406, + "loss": 1.2508, + "theoretical_loss": 3.3652799478726845, + "tokens_seen": 2818768896 + }, + { + "epoch": 0.85, + "learning_rate": 0.00011803410230692076, + "loss": 1.2992, + "theoretical_loss": 3.365273907272657, + "tokens_seen": 2818834432 + }, + { + "epoch": 0.85, + "learning_rate": 0.00011801805416248746, + "loss": 1.3765, + "theoretical_loss": 3.36526786685239, + "tokens_seen": 2818899968 + }, + { + "epoch": 0.85, + "learning_rate": 0.00011800200601805418, + "loss": 1.2511, + "theoretical_loss": 3.3652618266118743, + "tokens_seen": 2818965504 + }, + { + "epoch": 0.85, + "learning_rate": 0.00011798595787362088, + "loss": 1.4421, + "theoretical_loss": 3.3652557865510992, + "tokens_seen": 2819031040 + }, + { + "epoch": 0.85, + "learning_rate": 0.00011796990972918758, + "loss": 1.3067, + "theoretical_loss": 3.3652497466700564, + "tokens_seen": 2819096576 + }, + { + "epoch": 0.85, + "learning_rate": 0.00011795386158475428, + "loss": 1.1785, + "theoretical_loss": 3.3652437069687355, + "tokens_seen": 2819162112 + }, + { + "epoch": 0.85, + "learning_rate": 0.00011793781344032098, + "loss": 1.1767, + "theoretical_loss": 3.365237667447128, + "tokens_seen": 2819227648 + }, + { + "epoch": 0.85, + "learning_rate": 0.00011792176529588768, + "loss": 1.2997, + "theoretical_loss": 3.365231628105223, + "tokens_seen": 2819293184 + }, + { + "epoch": 0.85, + "learning_rate": 0.00011790571715145438, + "loss": 1.3196, + "theoretical_loss": 3.3652255889430123, + "tokens_seen": 2819358720 + }, + { + "epoch": 0.85, + "learning_rate": 0.00011788966900702107, + "loss": 1.3565, + "theoretical_loss": 3.365219549960485, + "tokens_seen": 2819424256 + }, + { + "epoch": 0.85, + "learning_rate": 0.00011787362086258777, + "loss": 1.2979, + "theoretical_loss": 3.365213511157633, + "tokens_seen": 2819489792 + }, + { + "epoch": 0.85, + "learning_rate": 0.00011785757271815447, + "loss": 1.3104, + "theoretical_loss": 3.3652074725344456, + "tokens_seen": 2819555328 + }, + { + "epoch": 0.85, + "learning_rate": 0.00011784152457372117, + "loss": 1.3163, + "theoretical_loss": 3.365201434090914, + "tokens_seen": 2819620864 + }, + { + "epoch": 0.85, + "learning_rate": 0.00011782547642928787, + "loss": 1.2243, + "theoretical_loss": 3.365195395827029, + "tokens_seen": 2819686400 + }, + { + "epoch": 0.85, + "learning_rate": 0.00011780942828485457, + "loss": 1.3443, + "theoretical_loss": 3.36518935774278, + "tokens_seen": 2819751936 + }, + { + "epoch": 0.85, + "learning_rate": 0.00011779338014042127, + "loss": 1.3781, + "theoretical_loss": 3.3651833198381578, + "tokens_seen": 2819817472 + }, + { + "epoch": 0.85, + "learning_rate": 0.00011777733199598797, + "loss": 1.3164, + "theoretical_loss": 3.3651772821131534, + "tokens_seen": 2819883008 + }, + { + "epoch": 0.85, + "learning_rate": 0.00011776128385155467, + "loss": 1.292, + "theoretical_loss": 3.3651712445677573, + "tokens_seen": 2819948544 + }, + { + "epoch": 0.85, + "learning_rate": 0.00011774523570712137, + "loss": 1.1603, + "theoretical_loss": 3.3651652072019593, + "tokens_seen": 2820014080 + }, + { + "epoch": 0.85, + "learning_rate": 0.00011772918756268808, + "loss": 1.2377, + "theoretical_loss": 3.3651591700157506, + "tokens_seen": 2820079616 + }, + { + "epoch": 0.85, + "learning_rate": 0.00011771313941825478, + "loss": 1.2379, + "theoretical_loss": 3.365153133009121, + "tokens_seen": 2820145152 + }, + { + "epoch": 0.85, + "learning_rate": 0.00011769709127382146, + "loss": 1.3475, + "theoretical_loss": 3.3651470961820618, + "tokens_seen": 2820210688 + }, + { + "epoch": 0.85, + "learning_rate": 0.00011768104312938816, + "loss": 1.3091, + "theoretical_loss": 3.3651410595345626, + "tokens_seen": 2820276224 + }, + { + "epoch": 0.85, + "learning_rate": 0.00011766499498495486, + "loss": 1.3098, + "theoretical_loss": 3.365135023066614, + "tokens_seen": 2820341760 + }, + { + "epoch": 0.85, + "learning_rate": 0.00011764894684052156, + "loss": 1.3174, + "theoretical_loss": 3.3651289867782075, + "tokens_seen": 2820407296 + }, + { + "epoch": 0.85, + "learning_rate": 0.00011763289869608827, + "loss": 1.1955, + "theoretical_loss": 3.3651229506693325, + "tokens_seen": 2820472832 + }, + { + "epoch": 0.85, + "learning_rate": 0.00011761685055165497, + "loss": 1.2628, + "theoretical_loss": 3.36511691473998, + "tokens_seen": 2820538368 + }, + { + "epoch": 0.85, + "learning_rate": 0.00011760080240722167, + "loss": 1.3174, + "theoretical_loss": 3.3651108789901403, + "tokens_seen": 2820603904 + }, + { + "epoch": 0.85, + "learning_rate": 0.00011758475426278837, + "loss": 1.2446, + "theoretical_loss": 3.365104843419804, + "tokens_seen": 2820669440 + }, + { + "epoch": 0.85, + "learning_rate": 0.00011756870611835507, + "loss": 1.3404, + "theoretical_loss": 3.3650988080289617, + "tokens_seen": 2820734976 + }, + { + "epoch": 0.85, + "learning_rate": 0.00011755265797392177, + "loss": 1.3105, + "theoretical_loss": 3.3650927728176034, + "tokens_seen": 2820800512 + }, + { + "epoch": 0.85, + "learning_rate": 0.00011753660982948847, + "loss": 1.2903, + "theoretical_loss": 3.3650867377857203, + "tokens_seen": 2820866048 + }, + { + "epoch": 0.85, + "learning_rate": 0.00011752056168505517, + "loss": 1.1645, + "theoretical_loss": 3.365080702933302, + "tokens_seen": 2820931584 + }, + { + "epoch": 0.85, + "learning_rate": 0.00011750451354062186, + "loss": 1.2106, + "theoretical_loss": 3.3650746682603403, + "tokens_seen": 2820997120 + }, + { + "epoch": 0.85, + "learning_rate": 0.00011748846539618856, + "loss": 1.2401, + "theoretical_loss": 3.3650686337668243, + "tokens_seen": 2821062656 + }, + { + "epoch": 0.85, + "learning_rate": 0.00011747241725175526, + "loss": 1.2719, + "theoretical_loss": 3.365062599452745, + "tokens_seen": 2821128192 + }, + { + "epoch": 0.85, + "learning_rate": 0.00011745636910732196, + "loss": 1.3484, + "theoretical_loss": 3.3650565653180937, + "tokens_seen": 2821193728 + }, + { + "epoch": 0.85, + "learning_rate": 0.00011744032096288866, + "loss": 1.3102, + "theoretical_loss": 3.3650505313628596, + "tokens_seen": 2821259264 + }, + { + "epoch": 0.85, + "objective/train/docs_used": 1025368, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.3580979108810425, + "objective/train/theoretical_loss": 3.365044497587034, + "objective/train/tokens_used": 2841784800, + "theoretical_loss": 3.365044497587034, + "tokens_seen": 2821324800 + }, + { + "epoch": 0.85, + "learning_rate": 0.00011742427281845536, + "loss": 1.3546, + "theoretical_loss": 3.365044497587034, + "tokens_seen": 2821324800 + }, + { + "epoch": 0.85, + "learning_rate": 0.00011740822467402209, + "loss": 1.3223, + "theoretical_loss": 3.365038463990607, + "tokens_seen": 2821390336 + }, + { + "epoch": 0.85, + "learning_rate": 0.00011739217652958879, + "loss": 1.3635, + "theoretical_loss": 3.36503243057357, + "tokens_seen": 2821455872 + }, + { + "epoch": 0.86, + "learning_rate": 0.00011737612838515548, + "loss": 1.1944, + "theoretical_loss": 3.365026397335912, + "tokens_seen": 2821521408 + }, + { + "epoch": 0.86, + "learning_rate": 0.00011736008024072218, + "loss": 1.3455, + "theoretical_loss": 3.3650203642776244, + "tokens_seen": 2821586944 + }, + { + "epoch": 0.86, + "learning_rate": 0.00011734403209628888, + "loss": 1.3309, + "theoretical_loss": 3.3650143313986978, + "tokens_seen": 2821652480 + }, + { + "epoch": 0.86, + "learning_rate": 0.00011732798395185558, + "loss": 1.2402, + "theoretical_loss": 3.3650082986991228, + "tokens_seen": 2821718016 + }, + { + "epoch": 0.86, + "learning_rate": 0.00011731193580742228, + "loss": 1.2818, + "theoretical_loss": 3.365002266178889, + "tokens_seen": 2821783552 + }, + { + "epoch": 0.86, + "learning_rate": 0.00011729588766298898, + "loss": 1.3116, + "theoretical_loss": 3.3649962338379877, + "tokens_seen": 2821849088 + }, + { + "epoch": 0.86, + "learning_rate": 0.00011727983951855568, + "loss": 1.2523, + "theoretical_loss": 3.3649902016764095, + "tokens_seen": 2821914624 + }, + { + "epoch": 0.86, + "learning_rate": 0.00011726379137412238, + "loss": 1.2181, + "theoretical_loss": 3.3649841696941443, + "tokens_seen": 2821980160 + }, + { + "epoch": 0.86, + "learning_rate": 0.00011724774322968908, + "loss": 1.3518, + "theoretical_loss": 3.364978137891183, + "tokens_seen": 2822045696 + }, + { + "epoch": 0.86, + "learning_rate": 0.00011723169508525578, + "loss": 1.2759, + "theoretical_loss": 3.3649721062675164, + "tokens_seen": 2822111232 + }, + { + "epoch": 0.86, + "learning_rate": 0.00011721564694082248, + "loss": 1.2991, + "theoretical_loss": 3.3649660748231343, + "tokens_seen": 2822176768 + }, + { + "epoch": 0.86, + "learning_rate": 0.00011719959879638918, + "loss": 1.2886, + "theoretical_loss": 3.3649600435580274, + "tokens_seen": 2822242304 + }, + { + "epoch": 0.86, + "learning_rate": 0.00011718355065195587, + "loss": 1.3032, + "theoretical_loss": 3.3649540124721864, + "tokens_seen": 2822307840 + }, + { + "epoch": 0.86, + "learning_rate": 0.00011716750250752257, + "loss": 1.2897, + "theoretical_loss": 3.364947981565602, + "tokens_seen": 2822373376 + }, + { + "epoch": 0.86, + "learning_rate": 0.00011715145436308927, + "loss": 1.2772, + "theoretical_loss": 3.364941950838264, + "tokens_seen": 2822438912 + }, + { + "epoch": 0.86, + "learning_rate": 0.00011713540621865597, + "loss": 1.1978, + "theoretical_loss": 3.364935920290164, + "tokens_seen": 2822504448 + }, + { + "epoch": 0.86, + "learning_rate": 0.00011711935807422267, + "loss": 1.3472, + "theoretical_loss": 3.3649298899212914, + "tokens_seen": 2822569984 + }, + { + "epoch": 0.86, + "learning_rate": 0.00011710330992978937, + "loss": 1.2954, + "theoretical_loss": 3.3649238597316375, + "tokens_seen": 2822635520 + }, + { + "epoch": 0.86, + "learning_rate": 0.00011708726178535607, + "loss": 1.3085, + "theoretical_loss": 3.3649178297211924, + "tokens_seen": 2822701056 + }, + { + "epoch": 0.86, + "learning_rate": 0.00011707121364092277, + "loss": 1.2616, + "theoretical_loss": 3.3649117998899465, + "tokens_seen": 2822766592 + }, + { + "epoch": 0.86, + "learning_rate": 0.00011705516549648947, + "loss": 1.2963, + "theoretical_loss": 3.364905770237891, + "tokens_seen": 2822832128 + }, + { + "epoch": 0.86, + "learning_rate": 0.00011703911735205618, + "loss": 1.3034, + "theoretical_loss": 3.364899740765016, + "tokens_seen": 2822897664 + }, + { + "epoch": 0.86, + "learning_rate": 0.00011702306920762288, + "loss": 1.1933, + "theoretical_loss": 3.3648937114713116, + "tokens_seen": 2822963200 + }, + { + "epoch": 0.86, + "learning_rate": 0.00011700702106318958, + "loss": 1.2893, + "theoretical_loss": 3.364887682356769, + "tokens_seen": 2823028736 + }, + { + "epoch": 0.86, + "learning_rate": 0.00011699097291875626, + "loss": 1.227, + "theoretical_loss": 3.364881653421378, + "tokens_seen": 2823094272 + }, + { + "epoch": 0.86, + "learning_rate": 0.00011697492477432296, + "loss": 1.337, + "theoretical_loss": 3.36487562466513, + "tokens_seen": 2823159808 + }, + { + "epoch": 0.86, + "learning_rate": 0.00011695887662988966, + "loss": 1.206, + "theoretical_loss": 3.364869596088014, + "tokens_seen": 2823225344 + }, + { + "epoch": 0.86, + "learning_rate": 0.00011694282848545637, + "loss": 1.2576, + "theoretical_loss": 3.3648635676900227, + "tokens_seen": 2823290880 + }, + { + "epoch": 0.86, + "learning_rate": 0.00011692678034102307, + "loss": 1.289, + "theoretical_loss": 3.3648575394711453, + "tokens_seen": 2823356416 + }, + { + "epoch": 0.86, + "learning_rate": 0.00011691073219658977, + "loss": 1.2517, + "theoretical_loss": 3.364851511431372, + "tokens_seen": 2823421952 + }, + { + "epoch": 0.86, + "learning_rate": 0.00011689468405215647, + "loss": 1.2753, + "theoretical_loss": 3.3648454835706945, + "tokens_seen": 2823487488 + }, + { + "epoch": 0.86, + "learning_rate": 0.00011687863590772317, + "loss": 1.3182, + "theoretical_loss": 3.3648394558891024, + "tokens_seen": 2823553024 + }, + { + "epoch": 0.86, + "learning_rate": 0.00011686258776328987, + "loss": 1.3132, + "theoretical_loss": 3.3648334283865866, + "tokens_seen": 2823618560 + }, + { + "epoch": 0.86, + "learning_rate": 0.00011684653961885657, + "loss": 1.3077, + "theoretical_loss": 3.364827401063137, + "tokens_seen": 2823684096 + }, + { + "epoch": 0.86, + "learning_rate": 0.00011683049147442327, + "loss": 1.2642, + "theoretical_loss": 3.364821373918745, + "tokens_seen": 2823749632 + }, + { + "epoch": 0.86, + "learning_rate": 0.00011681444332998998, + "loss": 1.2654, + "theoretical_loss": 3.364815346953401, + "tokens_seen": 2823815168 + }, + { + "epoch": 0.86, + "learning_rate": 0.00011679839518555668, + "loss": 1.2342, + "theoretical_loss": 3.3648093201670948, + "tokens_seen": 2823880704 + }, + { + "epoch": 0.86, + "learning_rate": 0.00011678234704112339, + "loss": 1.2402, + "theoretical_loss": 3.3648032935598176, + "tokens_seen": 2823946240 + }, + { + "epoch": 0.86, + "learning_rate": 0.00011676629889669009, + "loss": 1.1267, + "theoretical_loss": 3.3647972671315594, + "tokens_seen": 2824011776 + }, + { + "epoch": 0.86, + "learning_rate": 0.00011675025075225679, + "loss": 1.3682, + "theoretical_loss": 3.3647912408823113, + "tokens_seen": 2824077312 + }, + { + "epoch": 0.86, + "learning_rate": 0.00011673420260782349, + "loss": 1.3238, + "theoretical_loss": 3.3647852148120636, + "tokens_seen": 2824142848 + }, + { + "epoch": 0.86, + "learning_rate": 0.00011671815446339019, + "loss": 1.3296, + "theoretical_loss": 3.3647791889208065, + "tokens_seen": 2824208384 + }, + { + "epoch": 0.86, + "learning_rate": 0.00011670210631895689, + "loss": 1.2885, + "theoretical_loss": 3.3647731632085316, + "tokens_seen": 2824273920 + }, + { + "epoch": 0.86, + "learning_rate": 0.00011668605817452359, + "loss": 1.2675, + "theoretical_loss": 3.364767137675228, + "tokens_seen": 2824339456 + }, + { + "epoch": 0.86, + "learning_rate": 0.00011667001003009028, + "loss": 1.1918, + "theoretical_loss": 3.3647611123208874, + "tokens_seen": 2824404992 + }, + { + "epoch": 0.86, + "learning_rate": 0.00011665396188565698, + "loss": 1.2566, + "theoretical_loss": 3.3647550871454994, + "tokens_seen": 2824470528 + }, + { + "epoch": 0.86, + "learning_rate": 0.00011663791374122368, + "loss": 1.2423, + "theoretical_loss": 3.364749062149055, + "tokens_seen": 2824536064 + }, + { + "epoch": 0.86, + "objective/train/docs_used": 1026596, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.267777919769287, + "objective/train/theoretical_loss": 3.364743037331545, + "objective/train/tokens_used": 2845061600, + "theoretical_loss": 3.364743037331545, + "tokens_seen": 2824601600 + }, + { + "epoch": 0.86, + "learning_rate": 0.00011662186559679038, + "loss": 1.2718, + "theoretical_loss": 3.364743037331545, + "tokens_seen": 2824601600 + }, + { + "epoch": 0.86, + "learning_rate": 0.00011660581745235708, + "loss": 1.2086, + "theoretical_loss": 3.3647370126929594, + "tokens_seen": 2824667136 + }, + { + "epoch": 0.86, + "learning_rate": 0.00011658976930792378, + "loss": 1.2455, + "theoretical_loss": 3.364730988233289, + "tokens_seen": 2824732672 + }, + { + "epoch": 0.86, + "learning_rate": 0.00011657372116349048, + "loss": 1.2805, + "theoretical_loss": 3.3647249639525243, + "tokens_seen": 2824798208 + }, + { + "epoch": 0.86, + "learning_rate": 0.00011655767301905718, + "loss": 1.193, + "theoretical_loss": 3.364718939850656, + "tokens_seen": 2824863744 + }, + { + "epoch": 0.86, + "learning_rate": 0.00011654162487462388, + "loss": 1.2463, + "theoretical_loss": 3.3647129159276745, + "tokens_seen": 2824929280 + }, + { + "epoch": 0.86, + "learning_rate": 0.00011652557673019058, + "loss": 1.2976, + "theoretical_loss": 3.36470689218357, + "tokens_seen": 2824994816 + }, + { + "epoch": 0.86, + "learning_rate": 0.00011650952858575728, + "loss": 1.2816, + "theoretical_loss": 3.3647008686183337, + "tokens_seen": 2825060352 + }, + { + "epoch": 0.86, + "learning_rate": 0.00011649348044132398, + "loss": 1.2147, + "theoretical_loss": 3.3646948452319556, + "tokens_seen": 2825125888 + }, + { + "epoch": 0.86, + "learning_rate": 0.00011647743229689067, + "loss": 1.2861, + "theoretical_loss": 3.364688822024427, + "tokens_seen": 2825191424 + }, + { + "epoch": 0.86, + "learning_rate": 0.00011646138415245737, + "loss": 1.2534, + "theoretical_loss": 3.3646827989957373, + "tokens_seen": 2825256960 + }, + { + "epoch": 0.86, + "learning_rate": 0.00011644533600802407, + "loss": 1.2933, + "theoretical_loss": 3.3646767761458776, + "tokens_seen": 2825322496 + }, + { + "epoch": 0.86, + "learning_rate": 0.00011642928786359077, + "loss": 1.1862, + "theoretical_loss": 3.3646707534748383, + "tokens_seen": 2825388032 + }, + { + "epoch": 0.86, + "learning_rate": 0.00011641323971915747, + "loss": 1.4433, + "theoretical_loss": 3.3646647309826108, + "tokens_seen": 2825453568 + }, + { + "epoch": 0.86, + "learning_rate": 0.00011639719157472417, + "loss": 1.3567, + "theoretical_loss": 3.364658708669184, + "tokens_seen": 2825519104 + }, + { + "epoch": 0.86, + "learning_rate": 0.00011638114343029087, + "loss": 1.154, + "theoretical_loss": 3.3646526865345505, + "tokens_seen": 2825584640 + }, + { + "epoch": 0.86, + "learning_rate": 0.00011636509528585757, + "loss": 1.2565, + "theoretical_loss": 3.3646466645786988, + "tokens_seen": 2825650176 + }, + { + "epoch": 0.86, + "learning_rate": 0.00011634904714142428, + "loss": 1.1771, + "theoretical_loss": 3.364640642801621, + "tokens_seen": 2825715712 + }, + { + "epoch": 0.86, + "learning_rate": 0.00011633299899699098, + "loss": 1.2044, + "theoretical_loss": 3.3646346212033067, + "tokens_seen": 2825781248 + }, + { + "epoch": 0.86, + "learning_rate": 0.00011631695085255768, + "loss": 1.3016, + "theoretical_loss": 3.364628599783747, + "tokens_seen": 2825846784 + }, + { + "epoch": 0.86, + "learning_rate": 0.00011630090270812438, + "loss": 1.2814, + "theoretical_loss": 3.364622578542932, + "tokens_seen": 2825912320 + }, + { + "epoch": 0.86, + "learning_rate": 0.00011628485456369106, + "loss": 1.3839, + "theoretical_loss": 3.3646165574808524, + "tokens_seen": 2825977856 + }, + { + "epoch": 0.86, + "learning_rate": 0.00011626880641925777, + "loss": 1.1841, + "theoretical_loss": 3.364610536597499, + "tokens_seen": 2826043392 + }, + { + "epoch": 0.86, + "learning_rate": 0.00011625275827482447, + "loss": 1.1205, + "theoretical_loss": 3.3646045158928626, + "tokens_seen": 2826108928 + }, + { + "epoch": 0.86, + "learning_rate": 0.00011623671013039117, + "loss": 1.3004, + "theoretical_loss": 3.364598495366933, + "tokens_seen": 2826174464 + }, + { + "epoch": 0.86, + "learning_rate": 0.00011622066198595787, + "loss": 1.2788, + "theoretical_loss": 3.364592475019701, + "tokens_seen": 2826240000 + }, + { + "epoch": 0.86, + "learning_rate": 0.0001162046138415246, + "loss": 1.3649, + "theoretical_loss": 3.364586454851157, + "tokens_seen": 2826305536 + }, + { + "epoch": 0.86, + "learning_rate": 0.0001161885656970913, + "loss": 1.2706, + "theoretical_loss": 3.364580434861292, + "tokens_seen": 2826371072 + }, + { + "epoch": 0.86, + "learning_rate": 0.00011617251755265798, + "loss": 1.4161, + "theoretical_loss": 3.3645744150500967, + "tokens_seen": 2826436608 + }, + { + "epoch": 0.86, + "learning_rate": 0.00011615646940822468, + "loss": 1.2856, + "theoretical_loss": 3.3645683954175607, + "tokens_seen": 2826502144 + }, + { + "epoch": 0.86, + "learning_rate": 0.00011614042126379138, + "loss": 1.2266, + "theoretical_loss": 3.3645623759636756, + "tokens_seen": 2826567680 + }, + { + "epoch": 0.86, + "learning_rate": 0.00011612437311935808, + "loss": 1.273, + "theoretical_loss": 3.364556356688431, + "tokens_seen": 2826633216 + }, + { + "epoch": 0.86, + "learning_rate": 0.00011610832497492478, + "loss": 1.308, + "theoretical_loss": 3.3645503375918184, + "tokens_seen": 2826698752 + }, + { + "epoch": 0.86, + "learning_rate": 0.00011609227683049149, + "loss": 1.3249, + "theoretical_loss": 3.3645443186738277, + "tokens_seen": 2826764288 + }, + { + "epoch": 0.86, + "learning_rate": 0.00011607622868605819, + "loss": 1.2254, + "theoretical_loss": 3.36453829993445, + "tokens_seen": 2826829824 + }, + { + "epoch": 0.86, + "learning_rate": 0.00011606018054162489, + "loss": 1.1769, + "theoretical_loss": 3.364532281373675, + "tokens_seen": 2826895360 + }, + { + "epoch": 0.86, + "learning_rate": 0.00011604413239719159, + "loss": 1.2791, + "theoretical_loss": 3.3645262629914945, + "tokens_seen": 2826960896 + }, + { + "epoch": 0.86, + "learning_rate": 0.00011602808425275829, + "loss": 1.2322, + "theoretical_loss": 3.364520244787898, + "tokens_seen": 2827026432 + }, + { + "epoch": 0.86, + "learning_rate": 0.00011601203610832499, + "loss": 1.2964, + "theoretical_loss": 3.3645142267628763, + "tokens_seen": 2827091968 + }, + { + "epoch": 0.86, + "learning_rate": 0.00011599598796389169, + "loss": 1.2878, + "theoretical_loss": 3.36450820891642, + "tokens_seen": 2827157504 + }, + { + "epoch": 0.86, + "learning_rate": 0.00011597993981945838, + "loss": 1.239, + "theoretical_loss": 3.36450219124852, + "tokens_seen": 2827223040 + }, + { + "epoch": 0.86, + "learning_rate": 0.00011596389167502508, + "loss": 1.265, + "theoretical_loss": 3.364496173759167, + "tokens_seen": 2827288576 + }, + { + "epoch": 0.86, + "learning_rate": 0.00011594784353059178, + "loss": 1.2164, + "theoretical_loss": 3.3644901564483503, + "tokens_seen": 2827354112 + }, + { + "epoch": 0.86, + "learning_rate": 0.00011593179538615848, + "loss": 1.3012, + "theoretical_loss": 3.364484139316062, + "tokens_seen": 2827419648 + }, + { + "epoch": 0.86, + "learning_rate": 0.00011591574724172518, + "loss": 1.2357, + "theoretical_loss": 3.3644781223622915, + "tokens_seen": 2827485184 + }, + { + "epoch": 0.86, + "learning_rate": 0.00011589969909729188, + "loss": 1.2713, + "theoretical_loss": 3.3644721055870304, + "tokens_seen": 2827550720 + }, + { + "epoch": 0.86, + "learning_rate": 0.00011588365095285858, + "loss": 1.3618, + "theoretical_loss": 3.3644660889902682, + "tokens_seen": 2827616256 + }, + { + "epoch": 0.86, + "learning_rate": 0.00011586760280842528, + "loss": 1.258, + "theoretical_loss": 3.364460072571996, + "tokens_seen": 2827681792 + }, + { + "epoch": 0.86, + "learning_rate": 0.00011585155466399198, + "loss": 1.2594, + "theoretical_loss": 3.364454056332205, + "tokens_seen": 2827747328 + }, + { + "epoch": 0.86, + "learning_rate": 0.00011583550651955868, + "loss": 1.3414, + "theoretical_loss": 3.3644480402708847, + "tokens_seen": 2827812864 + }, + { + "epoch": 0.86, + "objective/train/docs_used": 1027853, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.3197367191314697, + "objective/train/theoretical_loss": 3.364442024388026, + "objective/train/tokens_used": 2848338400, + "theoretical_loss": 3.364442024388026, + "tokens_seen": 2827878400 + }, + { + "epoch": 0.86, + "learning_rate": 0.00011581945837512538, + "loss": 1.2837, + "theoretical_loss": 3.364442024388026, + "tokens_seen": 2827878400 + }, + { + "epoch": 0.86, + "learning_rate": 0.00011580341023069208, + "loss": 1.3262, + "theoretical_loss": 3.36443600868362, + "tokens_seen": 2827943936 + }, + { + "epoch": 0.86, + "learning_rate": 0.00011578736208625877, + "loss": 1.292, + "theoretical_loss": 3.3644299931576565, + "tokens_seen": 2828009472 + }, + { + "epoch": 0.86, + "learning_rate": 0.00011577131394182547, + "loss": 1.3119, + "theoretical_loss": 3.3644239778101266, + "tokens_seen": 2828075008 + }, + { + "epoch": 0.86, + "learning_rate": 0.00011575526579739217, + "loss": 1.3201, + "theoretical_loss": 3.3644179626410207, + "tokens_seen": 2828140544 + }, + { + "epoch": 0.86, + "learning_rate": 0.00011573921765295887, + "loss": 1.1842, + "theoretical_loss": 3.3644119476503294, + "tokens_seen": 2828206080 + }, + { + "epoch": 0.86, + "learning_rate": 0.00011572316950852557, + "loss": 1.2394, + "theoretical_loss": 3.364405932838043, + "tokens_seen": 2828271616 + }, + { + "epoch": 0.86, + "learning_rate": 0.00011570712136409227, + "loss": 1.2068, + "theoretical_loss": 3.3643999182041524, + "tokens_seen": 2828337152 + }, + { + "epoch": 0.86, + "learning_rate": 0.00011569107321965897, + "loss": 1.2492, + "theoretical_loss": 3.364393903748648, + "tokens_seen": 2828402688 + }, + { + "epoch": 0.86, + "learning_rate": 0.00011567502507522567, + "loss": 1.3215, + "theoretical_loss": 3.3643878894715207, + "tokens_seen": 2828468224 + }, + { + "epoch": 0.86, + "learning_rate": 0.00011565897693079238, + "loss": 1.3191, + "theoretical_loss": 3.364381875372761, + "tokens_seen": 2828533760 + }, + { + "epoch": 0.86, + "learning_rate": 0.00011564292878635908, + "loss": 1.2765, + "theoretical_loss": 3.364375861452359, + "tokens_seen": 2828599296 + }, + { + "epoch": 0.86, + "learning_rate": 0.00011562688064192578, + "loss": 1.1901, + "theoretical_loss": 3.3643698477103055, + "tokens_seen": 2828664832 + }, + { + "epoch": 0.86, + "learning_rate": 0.00011561083249749249, + "loss": 1.3188, + "theoretical_loss": 3.3643638341465913, + "tokens_seen": 2828730368 + }, + { + "epoch": 0.86, + "learning_rate": 0.00011559478435305919, + "loss": 1.2422, + "theoretical_loss": 3.364357820761207, + "tokens_seen": 2828795904 + }, + { + "epoch": 0.86, + "learning_rate": 0.00011557873620862589, + "loss": 1.2939, + "theoretical_loss": 3.3643518075541428, + "tokens_seen": 2828861440 + }, + { + "epoch": 0.86, + "learning_rate": 0.00011556268806419259, + "loss": 1.3924, + "theoretical_loss": 3.3643457945253896, + "tokens_seen": 2828926976 + }, + { + "epoch": 0.86, + "learning_rate": 0.0001155466399197593, + "loss": 1.2866, + "theoretical_loss": 3.3643397816749383, + "tokens_seen": 2828992512 + }, + { + "epoch": 0.86, + "learning_rate": 0.000115530591775326, + "loss": 1.3352, + "theoretical_loss": 3.3643337690027786, + "tokens_seen": 2829058048 + }, + { + "epoch": 0.86, + "learning_rate": 0.0001155145436308927, + "loss": 1.2876, + "theoretical_loss": 3.364327756508902, + "tokens_seen": 2829123584 + }, + { + "epoch": 0.86, + "learning_rate": 0.0001154984954864594, + "loss": 1.2693, + "theoretical_loss": 3.3643217441932984, + "tokens_seen": 2829189120 + }, + { + "epoch": 0.86, + "learning_rate": 0.0001154824473420261, + "loss": 1.4122, + "theoretical_loss": 3.3643157320559585, + "tokens_seen": 2829254656 + }, + { + "epoch": 0.86, + "learning_rate": 0.00011546639919759278, + "loss": 1.2789, + "theoretical_loss": 3.3643097200968732, + "tokens_seen": 2829320192 + }, + { + "epoch": 0.86, + "learning_rate": 0.00011545035105315948, + "loss": 1.2726, + "theoretical_loss": 3.364303708316033, + "tokens_seen": 2829385728 + }, + { + "epoch": 0.86, + "learning_rate": 0.00011543430290872618, + "loss": 1.3177, + "theoretical_loss": 3.364297696713428, + "tokens_seen": 2829451264 + }, + { + "epoch": 0.86, + "learning_rate": 0.00011541825476429289, + "loss": 1.2473, + "theoretical_loss": 3.3642916852890496, + "tokens_seen": 2829516800 + }, + { + "epoch": 0.86, + "learning_rate": 0.00011540220661985959, + "loss": 1.2781, + "theoretical_loss": 3.364285674042888, + "tokens_seen": 2829582336 + }, + { + "epoch": 0.86, + "learning_rate": 0.00011538615847542629, + "loss": 1.2133, + "theoretical_loss": 3.3642796629749334, + "tokens_seen": 2829647872 + }, + { + "epoch": 0.86, + "learning_rate": 0.00011537011033099299, + "loss": 1.2656, + "theoretical_loss": 3.364273652085177, + "tokens_seen": 2829713408 + }, + { + "epoch": 0.86, + "learning_rate": 0.00011535406218655969, + "loss": 1.2296, + "theoretical_loss": 3.364267641373609, + "tokens_seen": 2829778944 + }, + { + "epoch": 0.86, + "learning_rate": 0.00011533801404212639, + "loss": 1.2782, + "theoretical_loss": 3.36426163084022, + "tokens_seen": 2829844480 + }, + { + "epoch": 0.86, + "learning_rate": 0.00011532196589769309, + "loss": 1.2225, + "theoretical_loss": 3.364255620485001, + "tokens_seen": 2829910016 + }, + { + "epoch": 0.86, + "learning_rate": 0.00011530591775325979, + "loss": 1.3407, + "theoretical_loss": 3.364249610307942, + "tokens_seen": 2829975552 + }, + { + "epoch": 0.86, + "learning_rate": 0.00011528986960882649, + "loss": 1.233, + "theoretical_loss": 3.3642436003090346, + "tokens_seen": 2830041088 + }, + { + "epoch": 0.86, + "learning_rate": 0.00011527382146439318, + "loss": 1.2845, + "theoretical_loss": 3.364237590488268, + "tokens_seen": 2830106624 + }, + { + "epoch": 0.86, + "learning_rate": 0.00011525777331995988, + "loss": 1.2768, + "theoretical_loss": 3.364231580845634, + "tokens_seen": 2830172160 + }, + { + "epoch": 0.86, + "learning_rate": 0.00011524172517552658, + "loss": 1.2844, + "theoretical_loss": 3.364225571381122, + "tokens_seen": 2830237696 + }, + { + "epoch": 0.86, + "learning_rate": 0.00011522567703109328, + "loss": 1.2184, + "theoretical_loss": 3.364219562094724, + "tokens_seen": 2830303232 + }, + { + "epoch": 0.86, + "learning_rate": 0.00011520962888665998, + "loss": 1.2897, + "theoretical_loss": 3.3642135529864294, + "tokens_seen": 2830368768 + }, + { + "epoch": 0.86, + "learning_rate": 0.00011519358074222668, + "loss": 1.2799, + "theoretical_loss": 3.36420754405623, + "tokens_seen": 2830434304 + }, + { + "epoch": 0.86, + "learning_rate": 0.00011517753259779338, + "loss": 1.3149, + "theoretical_loss": 3.364201535304115, + "tokens_seen": 2830499840 + }, + { + "epoch": 0.86, + "learning_rate": 0.00011516148445336008, + "loss": 1.1135, + "theoretical_loss": 3.364195526730076, + "tokens_seen": 2830565376 + }, + { + "epoch": 0.86, + "learning_rate": 0.00011514543630892678, + "loss": 1.3112, + "theoretical_loss": 3.364189518334103, + "tokens_seen": 2830630912 + }, + { + "epoch": 0.86, + "learning_rate": 0.00011512938816449348, + "loss": 1.1607, + "theoretical_loss": 3.364183510116187, + "tokens_seen": 2830696448 + }, + { + "epoch": 0.86, + "learning_rate": 0.00011511334002006018, + "loss": 1.2447, + "theoretical_loss": 3.3641775020763185, + "tokens_seen": 2830761984 + }, + { + "epoch": 0.86, + "learning_rate": 0.00011509729187562688, + "loss": 1.2866, + "theoretical_loss": 3.3641714942144882, + "tokens_seen": 2830827520 + }, + { + "epoch": 0.86, + "learning_rate": 0.00011508124373119357, + "loss": 1.2973, + "theoretical_loss": 3.3641654865306867, + "tokens_seen": 2830893056 + }, + { + "epoch": 0.86, + "learning_rate": 0.00011506519558676027, + "loss": 1.3125, + "theoretical_loss": 3.364159479024904, + "tokens_seen": 2830958592 + }, + { + "epoch": 0.86, + "learning_rate": 0.00011504914744232697, + "loss": 1.2946, + "theoretical_loss": 3.3641534716971315, + "tokens_seen": 2831024128 + }, + { + "epoch": 0.86, + "learning_rate": 0.00011503309929789367, + "loss": 1.239, + "theoretical_loss": 3.3641474645473597, + "tokens_seen": 2831089664 + }, + { + "epoch": 0.86, + "objective/train/docs_used": 1029215, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.2607598304748535, + "objective/train/theoretical_loss": 3.364141457575579, + "objective/train/tokens_used": 2851615200, + "theoretical_loss": 3.364141457575579, + "tokens_seen": 2831155200 + }, + { + "epoch": 0.86, + "learning_rate": 0.00011501705115346037, + "loss": 1.2498, + "theoretical_loss": 3.364141457575579, + "tokens_seen": 2831155200 + }, + { + "epoch": 0.86, + "learning_rate": 0.0001150010030090271, + "loss": 1.3335, + "theoretical_loss": 3.3641354507817796, + "tokens_seen": 2831220736 + }, + { + "epoch": 0.86, + "learning_rate": 0.0001149849548645938, + "loss": 1.3273, + "theoretical_loss": 3.3641294441659526, + "tokens_seen": 2831286272 + }, + { + "epoch": 0.86, + "learning_rate": 0.0001149689067201605, + "loss": 1.2286, + "theoretical_loss": 3.3641234377280886, + "tokens_seen": 2831351808 + }, + { + "epoch": 0.86, + "learning_rate": 0.00011495285857572719, + "loss": 1.216, + "theoretical_loss": 3.3641174314681783, + "tokens_seen": 2831417344 + }, + { + "epoch": 0.86, + "learning_rate": 0.00011493681043129389, + "loss": 1.2654, + "theoretical_loss": 3.364111425386212, + "tokens_seen": 2831482880 + }, + { + "epoch": 0.86, + "learning_rate": 0.00011492076228686059, + "loss": 1.3515, + "theoretical_loss": 3.3641054194821804, + "tokens_seen": 2831548416 + }, + { + "epoch": 0.86, + "learning_rate": 0.00011490471414242729, + "loss": 1.2995, + "theoretical_loss": 3.3640994137560742, + "tokens_seen": 2831613952 + }, + { + "epoch": 0.86, + "learning_rate": 0.00011488866599799399, + "loss": 1.277, + "theoretical_loss": 3.364093408207884, + "tokens_seen": 2831679488 + }, + { + "epoch": 0.86, + "learning_rate": 0.00011487261785356069, + "loss": 1.2687, + "theoretical_loss": 3.3640874028376007, + "tokens_seen": 2831745024 + }, + { + "epoch": 0.86, + "learning_rate": 0.0001148565697091274, + "loss": 1.202, + "theoretical_loss": 3.364081397645214, + "tokens_seen": 2831810560 + }, + { + "epoch": 0.86, + "learning_rate": 0.0001148405215646941, + "loss": 1.3545, + "theoretical_loss": 3.3640753926307156, + "tokens_seen": 2831876096 + }, + { + "epoch": 0.86, + "learning_rate": 0.0001148244734202608, + "loss": 1.2258, + "theoretical_loss": 3.3640693877940957, + "tokens_seen": 2831941632 + }, + { + "epoch": 0.86, + "learning_rate": 0.0001148084252758275, + "loss": 1.3018, + "theoretical_loss": 3.3640633831353446, + "tokens_seen": 2832007168 + }, + { + "epoch": 0.86, + "learning_rate": 0.0001147923771313942, + "loss": 1.3522, + "theoretical_loss": 3.364057378654453, + "tokens_seen": 2832072704 + }, + { + "epoch": 0.86, + "learning_rate": 0.0001147763289869609, + "loss": 1.2888, + "theoretical_loss": 3.364051374351412, + "tokens_seen": 2832138240 + }, + { + "epoch": 0.86, + "learning_rate": 0.00011476028084252758, + "loss": 1.2576, + "theoretical_loss": 3.3640453702262114, + "tokens_seen": 2832203776 + }, + { + "epoch": 0.86, + "learning_rate": 0.00011474423269809428, + "loss": 1.2952, + "theoretical_loss": 3.364039366278843, + "tokens_seen": 2832269312 + }, + { + "epoch": 0.86, + "learning_rate": 0.00011472818455366099, + "loss": 1.3929, + "theoretical_loss": 3.3640333625092964, + "tokens_seen": 2832334848 + }, + { + "epoch": 0.86, + "learning_rate": 0.00011471213640922769, + "loss": 1.1872, + "theoretical_loss": 3.3640273589175624, + "tokens_seen": 2832400384 + }, + { + "epoch": 0.86, + "learning_rate": 0.00011469608826479439, + "loss": 1.3063, + "theoretical_loss": 3.364021355503632, + "tokens_seen": 2832465920 + }, + { + "epoch": 0.86, + "learning_rate": 0.00011468004012036109, + "loss": 1.2297, + "theoretical_loss": 3.3640153522674954, + "tokens_seen": 2832531456 + }, + { + "epoch": 0.86, + "learning_rate": 0.00011466399197592779, + "loss": 1.284, + "theoretical_loss": 3.3640093492091436, + "tokens_seen": 2832596992 + }, + { + "epoch": 0.86, + "learning_rate": 0.00011464794383149449, + "loss": 1.2234, + "theoretical_loss": 3.3640033463285666, + "tokens_seen": 2832662528 + }, + { + "epoch": 0.86, + "learning_rate": 0.00011463189568706119, + "loss": 1.2904, + "theoretical_loss": 3.3639973436257558, + "tokens_seen": 2832728064 + }, + { + "epoch": 0.86, + "learning_rate": 0.00011461584754262789, + "loss": 1.3581, + "theoretical_loss": 3.3639913411007014, + "tokens_seen": 2832793600 + }, + { + "epoch": 0.86, + "learning_rate": 0.00011459979939819459, + "loss": 1.313, + "theoretical_loss": 3.3639853387533942, + "tokens_seen": 2832859136 + }, + { + "epoch": 0.86, + "learning_rate": 0.00011458375125376129, + "loss": 1.3482, + "theoretical_loss": 3.3639793365838244, + "tokens_seen": 2832924672 + }, + { + "epoch": 0.86, + "learning_rate": 0.00011456770310932798, + "loss": 1.2892, + "theoretical_loss": 3.363973334591983, + "tokens_seen": 2832990208 + }, + { + "epoch": 0.86, + "learning_rate": 0.00011455165496489468, + "loss": 1.2746, + "theoretical_loss": 3.363967332777861, + "tokens_seen": 2833055744 + }, + { + "epoch": 0.86, + "learning_rate": 0.00011453560682046138, + "loss": 1.3196, + "theoretical_loss": 3.3639613311414482, + "tokens_seen": 2833121280 + }, + { + "epoch": 0.86, + "learning_rate": 0.00011451955867602808, + "loss": 1.2532, + "theoretical_loss": 3.363955329682736, + "tokens_seen": 2833186816 + }, + { + "epoch": 0.86, + "learning_rate": 0.00011450351053159478, + "loss": 1.371, + "theoretical_loss": 3.3639493284017146, + "tokens_seen": 2833252352 + }, + { + "epoch": 0.86, + "learning_rate": 0.00011448746238716148, + "loss": 1.2755, + "theoretical_loss": 3.363943327298374, + "tokens_seen": 2833317888 + }, + { + "epoch": 0.86, + "learning_rate": 0.00011447141424272818, + "loss": 1.2503, + "theoretical_loss": 3.3639373263727066, + "tokens_seen": 2833383424 + }, + { + "epoch": 0.86, + "learning_rate": 0.00011445536609829488, + "loss": 1.3148, + "theoretical_loss": 3.363931325624701, + "tokens_seen": 2833448960 + }, + { + "epoch": 0.86, + "learning_rate": 0.00011443931795386158, + "loss": 1.2816, + "theoretical_loss": 3.3639253250543493, + "tokens_seen": 2833514496 + }, + { + "epoch": 0.86, + "learning_rate": 0.00011442326980942828, + "loss": 1.2952, + "theoretical_loss": 3.3639193246616417, + "tokens_seen": 2833580032 + }, + { + "epoch": 0.86, + "learning_rate": 0.000114407221664995, + "loss": 1.2955, + "theoretical_loss": 3.3639133244465684, + "tokens_seen": 2833645568 + }, + { + "epoch": 0.86, + "learning_rate": 0.0001143911735205617, + "loss": 1.3364, + "theoretical_loss": 3.36390732440912, + "tokens_seen": 2833711104 + }, + { + "epoch": 0.86, + "learning_rate": 0.0001143751253761284, + "loss": 1.2943, + "theoretical_loss": 3.3639013245492886, + "tokens_seen": 2833776640 + }, + { + "epoch": 0.86, + "learning_rate": 0.0001143590772316951, + "loss": 1.3598, + "theoretical_loss": 3.363895324867063, + "tokens_seen": 2833842176 + }, + { + "epoch": 0.86, + "learning_rate": 0.0001143430290872618, + "loss": 1.2606, + "theoretical_loss": 3.3638893253624347, + "tokens_seen": 2833907712 + }, + { + "epoch": 0.86, + "learning_rate": 0.0001143269809428285, + "loss": 1.1991, + "theoretical_loss": 3.363883326035394, + "tokens_seen": 2833973248 + }, + { + "epoch": 0.86, + "learning_rate": 0.0001143109327983952, + "loss": 1.2625, + "theoretical_loss": 3.363877326885932, + "tokens_seen": 2834038784 + }, + { + "epoch": 0.86, + "learning_rate": 0.0001142948846539619, + "loss": 1.2946, + "theoretical_loss": 3.3638713279140395, + "tokens_seen": 2834104320 + }, + { + "epoch": 0.86, + "learning_rate": 0.0001142788365095286, + "loss": 1.3984, + "theoretical_loss": 3.3638653291197063, + "tokens_seen": 2834169856 + }, + { + "epoch": 0.86, + "learning_rate": 0.00011426278836509529, + "loss": 1.2551, + "theoretical_loss": 3.3638593305029234, + "tokens_seen": 2834235392 + }, + { + "epoch": 0.86, + "learning_rate": 0.00011424674022066199, + "loss": 1.1248, + "theoretical_loss": 3.363853332063681, + "tokens_seen": 2834300928 + }, + { + "epoch": 0.86, + "learning_rate": 0.00011423069207622869, + "loss": 1.1469, + "theoretical_loss": 3.363847333801971, + "tokens_seen": 2834366464 + }, + { + "epoch": 0.86, + "objective/train/docs_used": 1030428, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.3357394933700562, + "objective/train/theoretical_loss": 3.363841335717783, + "objective/train/tokens_used": 2854892000, + "theoretical_loss": 3.363841335717783, + "tokens_seen": 2834432000 + }, + { + "epoch": 0.86, + "learning_rate": 0.00011421464393179539, + "loss": 1.2429, + "theoretical_loss": 3.363841335717783, + "tokens_seen": 2834432000 + }, + { + "epoch": 0.86, + "learning_rate": 0.00011419859578736209, + "loss": 1.2935, + "theoretical_loss": 3.3638353378111083, + "tokens_seen": 2834497536 + }, + { + "epoch": 0.86, + "learning_rate": 0.0001141825476429288, + "loss": 1.3048, + "theoretical_loss": 3.363829340081937, + "tokens_seen": 2834563072 + }, + { + "epoch": 0.86, + "learning_rate": 0.0001141664994984955, + "loss": 1.2964, + "theoretical_loss": 3.3638233425302597, + "tokens_seen": 2834628608 + }, + { + "epoch": 0.86, + "learning_rate": 0.0001141504513540622, + "loss": 1.3335, + "theoretical_loss": 3.3638173451560673, + "tokens_seen": 2834694144 + }, + { + "epoch": 0.86, + "learning_rate": 0.0001141344032096289, + "loss": 1.1739, + "theoretical_loss": 3.3638113479593508, + "tokens_seen": 2834759680 + }, + { + "epoch": 0.86, + "learning_rate": 0.0001141183550651956, + "loss": 1.2981, + "theoretical_loss": 3.3638053509401002, + "tokens_seen": 2834825216 + }, + { + "epoch": 0.86, + "learning_rate": 0.0001141023069207623, + "loss": 1.3068, + "theoretical_loss": 3.363799354098306, + "tokens_seen": 2834890752 + }, + { + "epoch": 0.86, + "learning_rate": 0.000114086258776329, + "loss": 1.2779, + "theoretical_loss": 3.3637933574339596, + "tokens_seen": 2834956288 + }, + { + "epoch": 0.86, + "learning_rate": 0.00011407021063189568, + "loss": 1.317, + "theoretical_loss": 3.363787360947051, + "tokens_seen": 2835021824 + }, + { + "epoch": 0.86, + "learning_rate": 0.00011405416248746238, + "loss": 1.2621, + "theoretical_loss": 3.3637813646375716, + "tokens_seen": 2835087360 + }, + { + "epoch": 0.86, + "learning_rate": 0.00011403811434302909, + "loss": 1.2841, + "theoretical_loss": 3.3637753685055114, + "tokens_seen": 2835152896 + }, + { + "epoch": 0.86, + "learning_rate": 0.00011402206619859579, + "loss": 1.3031, + "theoretical_loss": 3.3637693725508613, + "tokens_seen": 2835218432 + }, + { + "epoch": 0.86, + "learning_rate": 0.00011400601805416249, + "loss": 1.2961, + "theoretical_loss": 3.363763376773612, + "tokens_seen": 2835283968 + }, + { + "epoch": 0.86, + "learning_rate": 0.00011398996990972919, + "loss": 1.2712, + "theoretical_loss": 3.3637573811737536, + "tokens_seen": 2835349504 + }, + { + "epoch": 0.86, + "learning_rate": 0.00011397392176529589, + "loss": 1.352, + "theoretical_loss": 3.3637513857512773, + "tokens_seen": 2835415040 + }, + { + "epoch": 0.86, + "learning_rate": 0.00011395787362086259, + "loss": 1.2104, + "theoretical_loss": 3.363745390506174, + "tokens_seen": 2835480576 + }, + { + "epoch": 0.86, + "learning_rate": 0.00011394182547642929, + "loss": 1.3316, + "theoretical_loss": 3.363739395438434, + "tokens_seen": 2835546112 + }, + { + "epoch": 0.86, + "learning_rate": 0.00011392577733199599, + "loss": 1.2167, + "theoretical_loss": 3.363733400548047, + "tokens_seen": 2835611648 + }, + { + "epoch": 0.86, + "learning_rate": 0.00011390972918756269, + "loss": 1.3149, + "theoretical_loss": 3.3637274058350055, + "tokens_seen": 2835677184 + }, + { + "epoch": 0.86, + "learning_rate": 0.00011389368104312939, + "loss": 1.3158, + "theoretical_loss": 3.363721411299299, + "tokens_seen": 2835742720 + }, + { + "epoch": 0.86, + "learning_rate": 0.00011387763289869608, + "loss": 1.296, + "theoretical_loss": 3.3637154169409187, + "tokens_seen": 2835808256 + }, + { + "epoch": 0.86, + "learning_rate": 0.00011386158475426278, + "loss": 1.3914, + "theoretical_loss": 3.3637094227598547, + "tokens_seen": 2835873792 + }, + { + "epoch": 0.86, + "learning_rate": 0.00011384553660982948, + "loss": 1.2949, + "theoretical_loss": 3.363703428756098, + "tokens_seen": 2835939328 + }, + { + "epoch": 0.86, + "learning_rate": 0.00011382948846539618, + "loss": 1.2779, + "theoretical_loss": 3.363697434929639, + "tokens_seen": 2836004864 + }, + { + "epoch": 0.86, + "learning_rate": 0.00011381344032096291, + "loss": 1.3426, + "theoretical_loss": 3.3636914412804684, + "tokens_seen": 2836070400 + }, + { + "epoch": 0.86, + "learning_rate": 0.00011379739217652961, + "loss": 1.3725, + "theoretical_loss": 3.3636854478085776, + "tokens_seen": 2836135936 + }, + { + "epoch": 0.86, + "learning_rate": 0.00011378134403209631, + "loss": 1.2587, + "theoretical_loss": 3.3636794545139566, + "tokens_seen": 2836201472 + }, + { + "epoch": 0.86, + "learning_rate": 0.00011376529588766301, + "loss": 1.3146, + "theoretical_loss": 3.3636734613965955, + "tokens_seen": 2836267008 + }, + { + "epoch": 0.86, + "learning_rate": 0.0001137492477432297, + "loss": 1.3261, + "theoretical_loss": 3.363667468456486, + "tokens_seen": 2836332544 + }, + { + "epoch": 0.86, + "learning_rate": 0.0001137331995987964, + "loss": 1.2854, + "theoretical_loss": 3.3636614756936183, + "tokens_seen": 2836398080 + }, + { + "epoch": 0.86, + "learning_rate": 0.0001137171514543631, + "loss": 1.2829, + "theoretical_loss": 3.363655483107983, + "tokens_seen": 2836463616 + }, + { + "epoch": 0.86, + "learning_rate": 0.0001137011033099298, + "loss": 1.3544, + "theoretical_loss": 3.3636494906995713, + "tokens_seen": 2836529152 + }, + { + "epoch": 0.86, + "learning_rate": 0.0001136850551654965, + "loss": 1.2815, + "theoretical_loss": 3.363643498468373, + "tokens_seen": 2836594688 + }, + { + "epoch": 0.86, + "learning_rate": 0.0001136690070210632, + "loss": 1.2734, + "theoretical_loss": 3.363637506414379, + "tokens_seen": 2836660224 + }, + { + "epoch": 0.86, + "learning_rate": 0.0001136529588766299, + "loss": 1.2872, + "theoretical_loss": 3.363631514537581, + "tokens_seen": 2836725760 + }, + { + "epoch": 0.86, + "learning_rate": 0.0001136369107321966, + "loss": 1.2738, + "theoretical_loss": 3.363625522837968, + "tokens_seen": 2836791296 + }, + { + "epoch": 0.86, + "learning_rate": 0.0001136208625877633, + "loss": 1.3835, + "theoretical_loss": 3.363619531315532, + "tokens_seen": 2836856832 + }, + { + "epoch": 0.86, + "learning_rate": 0.00011360481444333, + "loss": 1.3959, + "theoretical_loss": 3.363613539970263, + "tokens_seen": 2836922368 + }, + { + "epoch": 0.86, + "learning_rate": 0.0001135887662988967, + "loss": 1.2628, + "theoretical_loss": 3.3636075488021517, + "tokens_seen": 2836987904 + }, + { + "epoch": 0.86, + "learning_rate": 0.0001135727181544634, + "loss": 1.2059, + "theoretical_loss": 3.3636015578111893, + "tokens_seen": 2837053440 + }, + { + "epoch": 0.86, + "learning_rate": 0.00011355667001003009, + "loss": 1.2868, + "theoretical_loss": 3.3635955669973656, + "tokens_seen": 2837118976 + }, + { + "epoch": 0.86, + "learning_rate": 0.00011354062186559679, + "loss": 1.3421, + "theoretical_loss": 3.3635895763606722, + "tokens_seen": 2837184512 + }, + { + "epoch": 0.86, + "learning_rate": 0.00011352457372116349, + "loss": 1.3117, + "theoretical_loss": 3.363583585901099, + "tokens_seen": 2837250048 + }, + { + "epoch": 0.86, + "learning_rate": 0.00011350852557673019, + "loss": 1.3435, + "theoretical_loss": 3.363577595618637, + "tokens_seen": 2837315584 + }, + { + "epoch": 0.86, + "learning_rate": 0.0001134924774322969, + "loss": 1.2738, + "theoretical_loss": 3.363571605513277, + "tokens_seen": 2837381120 + }, + { + "epoch": 0.86, + "learning_rate": 0.0001134764292878636, + "loss": 1.2594, + "theoretical_loss": 3.3635656155850095, + "tokens_seen": 2837446656 + }, + { + "epoch": 0.86, + "learning_rate": 0.0001134603811434303, + "loss": 1.2602, + "theoretical_loss": 3.363559625833825, + "tokens_seen": 2837512192 + }, + { + "epoch": 0.86, + "learning_rate": 0.000113444332998997, + "loss": 1.3409, + "theoretical_loss": 3.3635536362597147, + "tokens_seen": 2837577728 + }, + { + "epoch": 0.86, + "learning_rate": 0.0001134282848545637, + "loss": 1.2786, + "theoretical_loss": 3.363547646862669, + "tokens_seen": 2837643264 + }, + { + "epoch": 0.86, + "objective/train/docs_used": 1031594, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.1104670763015747, + "objective/train/theoretical_loss": 3.3635416576426787, + "objective/train/tokens_used": 2858168800, + "theoretical_loss": 3.3635416576426787, + "tokens_seen": 2837708800 + }, + { + "epoch": 0.86, + "learning_rate": 0.0001134122367101304, + "loss": 1.2354, + "theoretical_loss": 3.3635416576426787, + "tokens_seen": 2837708800 + }, + { + "epoch": 0.86, + "learning_rate": 0.0001133961885656971, + "loss": 1.3378, + "theoretical_loss": 3.3635356685997335, + "tokens_seen": 2837774336 + }, + { + "epoch": 0.86, + "learning_rate": 0.0001133801404212638, + "loss": 1.3473, + "theoretical_loss": 3.3635296797338254, + "tokens_seen": 2837839872 + }, + { + "epoch": 0.86, + "learning_rate": 0.00011336409227683049, + "loss": 1.2532, + "theoretical_loss": 3.3635236910449446, + "tokens_seen": 2837905408 + }, + { + "epoch": 0.86, + "learning_rate": 0.00011334804413239719, + "loss": 1.3111, + "theoretical_loss": 3.3635177025330822, + "tokens_seen": 2837970944 + }, + { + "epoch": 0.86, + "learning_rate": 0.00011333199598796389, + "loss": 1.2378, + "theoretical_loss": 3.3635117141982276, + "tokens_seen": 2838036480 + }, + { + "epoch": 0.86, + "learning_rate": 0.00011331594784353059, + "loss": 1.2892, + "theoretical_loss": 3.3635057260403727, + "tokens_seen": 2838102016 + }, + { + "epoch": 0.86, + "learning_rate": 0.00011329989969909729, + "loss": 1.3123, + "theoretical_loss": 3.363499738059508, + "tokens_seen": 2838167552 + }, + { + "epoch": 0.86, + "learning_rate": 0.00011328385155466399, + "loss": 1.232, + "theoretical_loss": 3.363493750255624, + "tokens_seen": 2838233088 + }, + { + "epoch": 0.86, + "learning_rate": 0.00011326780341023069, + "loss": 1.2226, + "theoretical_loss": 3.363487762628711, + "tokens_seen": 2838298624 + }, + { + "epoch": 0.86, + "learning_rate": 0.00011325175526579739, + "loss": 1.2742, + "theoretical_loss": 3.36348177517876, + "tokens_seen": 2838364160 + }, + { + "epoch": 0.86, + "learning_rate": 0.00011323570712136409, + "loss": 1.2361, + "theoretical_loss": 3.363475787905762, + "tokens_seen": 2838429696 + }, + { + "epoch": 0.86, + "learning_rate": 0.00011321965897693079, + "loss": 1.3207, + "theoretical_loss": 3.3634698008097077, + "tokens_seen": 2838495232 + }, + { + "epoch": 0.86, + "learning_rate": 0.0001132036108324975, + "loss": 1.2094, + "theoretical_loss": 3.363463813890587, + "tokens_seen": 2838560768 + }, + { + "epoch": 0.86, + "learning_rate": 0.0001131875626880642, + "loss": 1.197, + "theoretical_loss": 3.363457827148391, + "tokens_seen": 2838626304 + }, + { + "epoch": 0.86, + "learning_rate": 0.0001131715145436309, + "loss": 1.2152, + "theoretical_loss": 3.3634518405831106, + "tokens_seen": 2838691840 + }, + { + "epoch": 0.86, + "learning_rate": 0.00011315546639919761, + "loss": 1.177, + "theoretical_loss": 3.3634458541947367, + "tokens_seen": 2838757376 + }, + { + "epoch": 0.86, + "learning_rate": 0.00011313941825476431, + "loss": 1.2601, + "theoretical_loss": 3.363439867983259, + "tokens_seen": 2838822912 + }, + { + "epoch": 0.86, + "learning_rate": 0.00011312337011033101, + "loss": 1.2621, + "theoretical_loss": 3.3634338819486693, + "tokens_seen": 2838888448 + }, + { + "epoch": 0.86, + "learning_rate": 0.00011310732196589771, + "loss": 1.1958, + "theoretical_loss": 3.363427896090958, + "tokens_seen": 2838953984 + }, + { + "epoch": 0.86, + "learning_rate": 0.00011309127382146441, + "loss": 1.3224, + "theoretical_loss": 3.3634219104101155, + "tokens_seen": 2839019520 + }, + { + "epoch": 0.86, + "learning_rate": 0.00011307522567703111, + "loss": 1.4233, + "theoretical_loss": 3.3634159249061324, + "tokens_seen": 2839085056 + }, + { + "epoch": 0.86, + "learning_rate": 0.00011305917753259781, + "loss": 1.2852, + "theoretical_loss": 3.363409939579, + "tokens_seen": 2839150592 + }, + { + "epoch": 0.86, + "learning_rate": 0.0001130431293881645, + "loss": 1.278, + "theoretical_loss": 3.363403954428708, + "tokens_seen": 2839216128 + }, + { + "epoch": 0.86, + "learning_rate": 0.0001130270812437312, + "loss": 1.2476, + "theoretical_loss": 3.363397969455248, + "tokens_seen": 2839281664 + }, + { + "epoch": 0.86, + "learning_rate": 0.0001130110330992979, + "loss": 1.2271, + "theoretical_loss": 3.3633919846586102, + "tokens_seen": 2839347200 + }, + { + "epoch": 0.86, + "learning_rate": 0.0001129949849548646, + "loss": 1.2537, + "theoretical_loss": 3.3633860000387856, + "tokens_seen": 2839412736 + }, + { + "epoch": 0.86, + "learning_rate": 0.0001129789368104313, + "loss": 1.3056, + "theoretical_loss": 3.363380015595765, + "tokens_seen": 2839478272 + }, + { + "epoch": 0.86, + "learning_rate": 0.000112962888665998, + "loss": 1.3498, + "theoretical_loss": 3.3633740313295384, + "tokens_seen": 2839543808 + }, + { + "epoch": 0.86, + "learning_rate": 0.0001129468405215647, + "loss": 1.2306, + "theoretical_loss": 3.3633680472400975, + "tokens_seen": 2839609344 + }, + { + "epoch": 0.86, + "learning_rate": 0.0001129307923771314, + "loss": 1.2735, + "theoretical_loss": 3.363362063327432, + "tokens_seen": 2839674880 + }, + { + "epoch": 0.86, + "learning_rate": 0.0001129147442326981, + "loss": 1.3666, + "theoretical_loss": 3.3633560795915334, + "tokens_seen": 2839740416 + }, + { + "epoch": 0.86, + "learning_rate": 0.0001128986960882648, + "loss": 1.2682, + "theoretical_loss": 3.363350096032392, + "tokens_seen": 2839805952 + }, + { + "epoch": 0.86, + "learning_rate": 0.0001128826479438315, + "loss": 1.2753, + "theoretical_loss": 3.3633441126499983, + "tokens_seen": 2839871488 + }, + { + "epoch": 0.86, + "learning_rate": 0.0001128665997993982, + "loss": 1.2782, + "theoretical_loss": 3.3633381294443434, + "tokens_seen": 2839937024 + }, + { + "epoch": 0.86, + "learning_rate": 0.00011285055165496489, + "loss": 1.2339, + "theoretical_loss": 3.3633321464154182, + "tokens_seen": 2840002560 + }, + { + "epoch": 0.86, + "learning_rate": 0.00011283450351053159, + "loss": 1.3654, + "theoretical_loss": 3.3633261635632126, + "tokens_seen": 2840068096 + }, + { + "epoch": 0.86, + "learning_rate": 0.00011281845536609829, + "loss": 1.3256, + "theoretical_loss": 3.363320180887718, + "tokens_seen": 2840133632 + }, + { + "epoch": 0.86, + "learning_rate": 0.000112802407221665, + "loss": 1.281, + "theoretical_loss": 3.3633141983889248, + "tokens_seen": 2840199168 + }, + { + "epoch": 0.86, + "learning_rate": 0.0001127863590772317, + "loss": 1.2533, + "theoretical_loss": 3.363308216066824, + "tokens_seen": 2840264704 + }, + { + "epoch": 0.86, + "learning_rate": 0.0001127703109327984, + "loss": 1.2272, + "theoretical_loss": 3.363302233921406, + "tokens_seen": 2840330240 + }, + { + "epoch": 0.86, + "learning_rate": 0.0001127542627883651, + "loss": 1.2428, + "theoretical_loss": 3.363296251952661, + "tokens_seen": 2840395776 + }, + { + "epoch": 0.86, + "learning_rate": 0.0001127382146439318, + "loss": 1.3163, + "theoretical_loss": 3.3632902701605807, + "tokens_seen": 2840461312 + }, + { + "epoch": 0.86, + "learning_rate": 0.0001127221664994985, + "loss": 1.2619, + "theoretical_loss": 3.3632842885451555, + "tokens_seen": 2840526848 + }, + { + "epoch": 0.86, + "learning_rate": 0.0001127061183550652, + "loss": 1.2466, + "theoretical_loss": 3.363278307106376, + "tokens_seen": 2840592384 + }, + { + "epoch": 0.86, + "learning_rate": 0.0001126900702106319, + "loss": 1.2343, + "theoretical_loss": 3.3632723258442327, + "tokens_seen": 2840657920 + }, + { + "epoch": 0.86, + "learning_rate": 0.0001126740220661986, + "loss": 1.2314, + "theoretical_loss": 3.3632663447587166, + "tokens_seen": 2840723456 + }, + { + "epoch": 0.86, + "learning_rate": 0.00011265797392176529, + "loss": 1.2923, + "theoretical_loss": 3.363260363849818, + "tokens_seen": 2840788992 + }, + { + "epoch": 0.86, + "learning_rate": 0.00011264192577733199, + "loss": 1.3241, + "theoretical_loss": 3.363254383117529, + "tokens_seen": 2840854528 + }, + { + "epoch": 0.86, + "learning_rate": 0.00011262587763289869, + "loss": 1.3221, + "theoretical_loss": 3.3632484025618385, + "tokens_seen": 2840920064 + }, + { + "epoch": 0.86, + "objective/train/docs_used": 1032706, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.3313937187194824, + "objective/train/theoretical_loss": 3.3632424221827377, + "objective/train/tokens_used": 2861445600, + "theoretical_loss": 3.3632424221827377, + "tokens_seen": 2840985600 + }, + { + "epoch": 0.86, + "learning_rate": 0.00011260982948846541, + "loss": 1.2683, + "theoretical_loss": 3.3632424221827377, + "tokens_seen": 2840985600 + }, + { + "epoch": 0.86, + "learning_rate": 0.00011259378134403212, + "loss": 1.3492, + "theoretical_loss": 3.363236441980218, + "tokens_seen": 2841051136 + }, + { + "epoch": 0.86, + "learning_rate": 0.00011257773319959882, + "loss": 1.1853, + "theoretical_loss": 3.3632304619542697, + "tokens_seen": 2841116672 + }, + { + "epoch": 0.86, + "learning_rate": 0.00011256168505516552, + "loss": 1.4147, + "theoretical_loss": 3.3632244821048833, + "tokens_seen": 2841182208 + }, + { + "epoch": 0.86, + "learning_rate": 0.0001125456369107322, + "loss": 1.3337, + "theoretical_loss": 3.36321850243205, + "tokens_seen": 2841247744 + }, + { + "epoch": 0.86, + "learning_rate": 0.0001125295887662989, + "loss": 1.1567, + "theoretical_loss": 3.36321252293576, + "tokens_seen": 2841313280 + }, + { + "epoch": 0.86, + "learning_rate": 0.0001125135406218656, + "loss": 1.2913, + "theoretical_loss": 3.363206543616004, + "tokens_seen": 2841378816 + }, + { + "epoch": 0.86, + "learning_rate": 0.0001124974924774323, + "loss": 1.2903, + "theoretical_loss": 3.3632005644727734, + "tokens_seen": 2841444352 + }, + { + "epoch": 0.86, + "learning_rate": 0.000112481444332999, + "loss": 1.3493, + "theoretical_loss": 3.3631945855060583, + "tokens_seen": 2841509888 + }, + { + "epoch": 0.86, + "learning_rate": 0.00011246539618856571, + "loss": 1.3251, + "theoretical_loss": 3.3631886067158496, + "tokens_seen": 2841575424 + }, + { + "epoch": 0.86, + "learning_rate": 0.00011244934804413241, + "loss": 1.3143, + "theoretical_loss": 3.363182628102138, + "tokens_seen": 2841640960 + }, + { + "epoch": 0.86, + "learning_rate": 0.00011243329989969911, + "loss": 1.3202, + "theoretical_loss": 3.3631766496649145, + "tokens_seen": 2841706496 + }, + { + "epoch": 0.86, + "learning_rate": 0.00011241725175526581, + "loss": 1.2405, + "theoretical_loss": 3.3631706714041694, + "tokens_seen": 2841772032 + }, + { + "epoch": 0.86, + "learning_rate": 0.00011240120361083251, + "loss": 1.3292, + "theoretical_loss": 3.3631646933198933, + "tokens_seen": 2841837568 + }, + { + "epoch": 0.86, + "learning_rate": 0.00011238515546639921, + "loss": 1.2733, + "theoretical_loss": 3.3631587154120774, + "tokens_seen": 2841903104 + }, + { + "epoch": 0.86, + "learning_rate": 0.00011236910732196591, + "loss": 1.3471, + "theoretical_loss": 3.3631527376807124, + "tokens_seen": 2841968640 + }, + { + "epoch": 0.86, + "learning_rate": 0.0001123530591775326, + "loss": 1.2626, + "theoretical_loss": 3.3631467601257885, + "tokens_seen": 2842034176 + }, + { + "epoch": 0.86, + "learning_rate": 0.0001123370110330993, + "loss": 1.2182, + "theoretical_loss": 3.3631407827472968, + "tokens_seen": 2842099712 + }, + { + "epoch": 0.86, + "learning_rate": 0.000112320962888666, + "loss": 1.3532, + "theoretical_loss": 3.3631348055452284, + "tokens_seen": 2842165248 + }, + { + "epoch": 0.86, + "learning_rate": 0.0001123049147442327, + "loss": 1.264, + "theoretical_loss": 3.3631288285195735, + "tokens_seen": 2842230784 + }, + { + "epoch": 0.86, + "learning_rate": 0.0001122888665997994, + "loss": 1.3141, + "theoretical_loss": 3.3631228516703224, + "tokens_seen": 2842296320 + }, + { + "epoch": 0.86, + "learning_rate": 0.0001122728184553661, + "loss": 1.2539, + "theoretical_loss": 3.363116874997467, + "tokens_seen": 2842361856 + }, + { + "epoch": 0.86, + "learning_rate": 0.0001122567703109328, + "loss": 1.2729, + "theoretical_loss": 3.363110898500997, + "tokens_seen": 2842427392 + }, + { + "epoch": 0.86, + "learning_rate": 0.0001122407221664995, + "loss": 1.3308, + "theoretical_loss": 3.363104922180904, + "tokens_seen": 2842492928 + }, + { + "epoch": 0.86, + "learning_rate": 0.0001122246740220662, + "loss": 1.283, + "theoretical_loss": 3.363098946037178, + "tokens_seen": 2842558464 + }, + { + "epoch": 0.86, + "learning_rate": 0.0001122086258776329, + "loss": 1.2706, + "theoretical_loss": 3.36309297006981, + "tokens_seen": 2842624000 + }, + { + "epoch": 0.86, + "learning_rate": 0.0001121925777331996, + "loss": 1.3284, + "theoretical_loss": 3.3630869942787904, + "tokens_seen": 2842689536 + }, + { + "epoch": 0.86, + "learning_rate": 0.0001121765295887663, + "loss": 1.1577, + "theoretical_loss": 3.3630810186641105, + "tokens_seen": 2842755072 + }, + { + "epoch": 0.86, + "learning_rate": 0.000112160481444333, + "loss": 1.3942, + "theoretical_loss": 3.363075043225761, + "tokens_seen": 2842820608 + }, + { + "epoch": 0.86, + "learning_rate": 0.00011214443329989969, + "loss": 1.3182, + "theoretical_loss": 3.363069067963732, + "tokens_seen": 2842886144 + }, + { + "epoch": 0.86, + "learning_rate": 0.0001121283851554664, + "loss": 1.26, + "theoretical_loss": 3.363063092878015, + "tokens_seen": 2842951680 + }, + { + "epoch": 0.86, + "learning_rate": 0.0001121123370110331, + "loss": 1.2991, + "theoretical_loss": 3.3630571179685997, + "tokens_seen": 2843017216 + }, + { + "epoch": 0.86, + "learning_rate": 0.0001120962888665998, + "loss": 1.2701, + "theoretical_loss": 3.363051143235478, + "tokens_seen": 2843082752 + }, + { + "epoch": 0.86, + "learning_rate": 0.0001120802407221665, + "loss": 1.2173, + "theoretical_loss": 3.3630451686786405, + "tokens_seen": 2843148288 + }, + { + "epoch": 0.86, + "learning_rate": 0.0001120641925777332, + "loss": 1.2908, + "theoretical_loss": 3.363039194298077, + "tokens_seen": 2843213824 + }, + { + "epoch": 0.86, + "learning_rate": 0.0001120481444332999, + "loss": 1.2544, + "theoretical_loss": 3.363033220093779, + "tokens_seen": 2843279360 + }, + { + "epoch": 0.86, + "learning_rate": 0.0001120320962888666, + "loss": 1.2613, + "theoretical_loss": 3.363027246065737, + "tokens_seen": 2843344896 + }, + { + "epoch": 0.86, + "learning_rate": 0.0001120160481444333, + "loss": 1.2689, + "theoretical_loss": 3.363021272213942, + "tokens_seen": 2843410432 + }, + { + "epoch": 0.86, + "learning_rate": 0.00011200000000000001, + "loss": 1.2815, + "theoretical_loss": 3.363015298538384, + "tokens_seen": 2843475968 + }, + { + "epoch": 0.86, + "learning_rate": 0.00011198395185556671, + "loss": 1.2663, + "theoretical_loss": 3.3630093250390547, + "tokens_seen": 2843541504 + }, + { + "epoch": 0.86, + "learning_rate": 0.00011196790371113341, + "loss": 1.3371, + "theoretical_loss": 3.3630033517159443, + "tokens_seen": 2843607040 + }, + { + "epoch": 0.86, + "learning_rate": 0.00011195185556670011, + "loss": 1.3064, + "theoretical_loss": 3.3629973785690437, + "tokens_seen": 2843672576 + }, + { + "epoch": 0.86, + "learning_rate": 0.00011193580742226681, + "loss": 1.3467, + "theoretical_loss": 3.362991405598344, + "tokens_seen": 2843738112 + }, + { + "epoch": 0.86, + "learning_rate": 0.00011191975927783352, + "loss": 1.2212, + "theoretical_loss": 3.3629854328038347, + "tokens_seen": 2843803648 + }, + { + "epoch": 0.86, + "learning_rate": 0.00011190371113340022, + "loss": 1.4037, + "theoretical_loss": 3.3629794601855076, + "tokens_seen": 2843869184 + }, + { + "epoch": 0.86, + "learning_rate": 0.00011188766298896692, + "loss": 1.22, + "theoretical_loss": 3.3629734877433535, + "tokens_seen": 2843934720 + }, + { + "epoch": 0.86, + "learning_rate": 0.00011187161484453362, + "loss": 1.3305, + "theoretical_loss": 3.362967515477363, + "tokens_seen": 2844000256 + }, + { + "epoch": 0.86, + "learning_rate": 0.00011185556670010032, + "loss": 1.2655, + "theoretical_loss": 3.3629615433875264, + "tokens_seen": 2844065792 + }, + { + "epoch": 0.86, + "learning_rate": 0.000111839518555667, + "loss": 1.3126, + "theoretical_loss": 3.362955571473835, + "tokens_seen": 2844131328 + }, + { + "epoch": 0.86, + "learning_rate": 0.0001118234704112337, + "loss": 1.3533, + "theoretical_loss": 3.362949599736279, + "tokens_seen": 2844196864 + }, + { + "epoch": 0.86, + "objective/train/docs_used": 1033870, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.3258209228515625, + "objective/train/theoretical_loss": 3.3629436281748495, + "objective/train/tokens_used": 2864722400, + "theoretical_loss": 3.3629436281748495, + "tokens_seen": 2844262400 + }, + { + "epoch": 0.86, + "learning_rate": 0.0001118074222668004, + "loss": 1.287, + "theoretical_loss": 3.3629436281748495, + "tokens_seen": 2844262400 + }, + { + "epoch": 0.86, + "learning_rate": 0.0001117913741223671, + "loss": 1.1689, + "theoretical_loss": 3.362937656789537, + "tokens_seen": 2844327936 + }, + { + "epoch": 0.86, + "learning_rate": 0.00011177532597793381, + "loss": 1.3727, + "theoretical_loss": 3.3629316855803326, + "tokens_seen": 2844393472 + }, + { + "epoch": 0.86, + "learning_rate": 0.00011175927783350051, + "loss": 1.294, + "theoretical_loss": 3.362925714547227, + "tokens_seen": 2844459008 + }, + { + "epoch": 0.86, + "learning_rate": 0.00011174322968906721, + "loss": 1.3157, + "theoretical_loss": 3.3629197436902105, + "tokens_seen": 2844524544 + }, + { + "epoch": 0.86, + "learning_rate": 0.00011172718154463391, + "loss": 1.3702, + "theoretical_loss": 3.3629137730092746, + "tokens_seen": 2844590080 + }, + { + "epoch": 0.86, + "learning_rate": 0.00011171113340020061, + "loss": 1.2285, + "theoretical_loss": 3.3629078025044095, + "tokens_seen": 2844655616 + }, + { + "epoch": 0.86, + "learning_rate": 0.00011169508525576731, + "loss": 1.3037, + "theoretical_loss": 3.362901832175606, + "tokens_seen": 2844721152 + }, + { + "epoch": 0.86, + "learning_rate": 0.00011167903711133401, + "loss": 1.192, + "theoretical_loss": 3.362895862022855, + "tokens_seen": 2844786688 + }, + { + "epoch": 0.86, + "learning_rate": 0.00011166298896690071, + "loss": 1.1953, + "theoretical_loss": 3.362889892046147, + "tokens_seen": 2844852224 + }, + { + "epoch": 0.86, + "learning_rate": 0.0001116469408224674, + "loss": 1.2447, + "theoretical_loss": 3.3628839222454734, + "tokens_seen": 2844917760 + }, + { + "epoch": 0.86, + "learning_rate": 0.0001116308926780341, + "loss": 1.2945, + "theoretical_loss": 3.362877952620824, + "tokens_seen": 2844983296 + }, + { + "epoch": 0.86, + "learning_rate": 0.0001116148445336008, + "loss": 1.2941, + "theoretical_loss": 3.3628719831721905, + "tokens_seen": 2845048832 + }, + { + "epoch": 0.86, + "learning_rate": 0.0001115987963891675, + "loss": 1.2001, + "theoretical_loss": 3.3628660138995627, + "tokens_seen": 2845114368 + }, + { + "epoch": 0.86, + "learning_rate": 0.0001115827482447342, + "loss": 1.3198, + "theoretical_loss": 3.3628600448029324, + "tokens_seen": 2845179904 + }, + { + "epoch": 0.86, + "learning_rate": 0.0001115667001003009, + "loss": 1.1656, + "theoretical_loss": 3.3628540758822894, + "tokens_seen": 2845245440 + }, + { + "epoch": 0.86, + "learning_rate": 0.0001115506519558676, + "loss": 1.3579, + "theoretical_loss": 3.3628481071376255, + "tokens_seen": 2845310976 + }, + { + "epoch": 0.86, + "learning_rate": 0.0001115346038114343, + "loss": 1.3797, + "theoretical_loss": 3.36284213856893, + "tokens_seen": 2845376512 + }, + { + "epoch": 0.86, + "learning_rate": 0.000111518555667001, + "loss": 1.2437, + "theoretical_loss": 3.362836170176195, + "tokens_seen": 2845442048 + }, + { + "epoch": 0.86, + "learning_rate": 0.0001115025075225677, + "loss": 1.3044, + "theoretical_loss": 3.3628302019594107, + "tokens_seen": 2845507584 + }, + { + "epoch": 0.86, + "learning_rate": 0.0001114864593781344, + "loss": 1.278, + "theoretical_loss": 3.362824233918568, + "tokens_seen": 2845573120 + }, + { + "epoch": 0.86, + "learning_rate": 0.0001114704112337011, + "loss": 1.3005, + "theoretical_loss": 3.362818266053657, + "tokens_seen": 2845638656 + }, + { + "epoch": 0.86, + "learning_rate": 0.00011145436308926779, + "loss": 1.281, + "theoretical_loss": 3.3628122983646698, + "tokens_seen": 2845704192 + }, + { + "epoch": 0.86, + "learning_rate": 0.0001114383149448345, + "loss": 1.3193, + "theoretical_loss": 3.3628063308515963, + "tokens_seen": 2845769728 + }, + { + "epoch": 0.86, + "learning_rate": 0.0001114222668004012, + "loss": 1.2751, + "theoretical_loss": 3.362800363514427, + "tokens_seen": 2845835264 + }, + { + "epoch": 0.86, + "learning_rate": 0.00011140621865596792, + "loss": 1.3126, + "theoretical_loss": 3.3627943963531535, + "tokens_seen": 2845900800 + }, + { + "epoch": 0.86, + "learning_rate": 0.00011139017051153462, + "loss": 1.2199, + "theoretical_loss": 3.3627884293677655, + "tokens_seen": 2845966336 + }, + { + "epoch": 0.86, + "learning_rate": 0.00011137412236710132, + "loss": 1.2591, + "theoretical_loss": 3.362782462558255, + "tokens_seen": 2846031872 + }, + { + "epoch": 0.86, + "learning_rate": 0.00011135807422266802, + "loss": 1.3494, + "theoretical_loss": 3.362776495924612, + "tokens_seen": 2846097408 + }, + { + "epoch": 0.86, + "learning_rate": 0.00011134202607823472, + "loss": 1.303, + "theoretical_loss": 3.362770529466827, + "tokens_seen": 2846162944 + }, + { + "epoch": 0.86, + "learning_rate": 0.00011132597793380141, + "loss": 1.2279, + "theoretical_loss": 3.362764563184891, + "tokens_seen": 2846228480 + }, + { + "epoch": 0.86, + "learning_rate": 0.00011130992978936811, + "loss": 1.2575, + "theoretical_loss": 3.3627585970787957, + "tokens_seen": 2846294016 + }, + { + "epoch": 0.86, + "learning_rate": 0.00011129388164493481, + "loss": 1.3275, + "theoretical_loss": 3.3627526311485307, + "tokens_seen": 2846359552 + }, + { + "epoch": 0.86, + "learning_rate": 0.00011127783350050151, + "loss": 1.244, + "theoretical_loss": 3.362746665394087, + "tokens_seen": 2846425088 + }, + { + "epoch": 0.86, + "learning_rate": 0.00011126178535606821, + "loss": 1.3322, + "theoretical_loss": 3.362740699815456, + "tokens_seen": 2846490624 + }, + { + "epoch": 0.86, + "learning_rate": 0.00011124573721163491, + "loss": 1.2472, + "theoretical_loss": 3.3627347344126277, + "tokens_seen": 2846556160 + }, + { + "epoch": 0.86, + "learning_rate": 0.00011122968906720162, + "loss": 1.2418, + "theoretical_loss": 3.362728769185593, + "tokens_seen": 2846621696 + }, + { + "epoch": 0.86, + "learning_rate": 0.00011121364092276832, + "loss": 1.1858, + "theoretical_loss": 3.3627228041343433, + "tokens_seen": 2846687232 + }, + { + "epoch": 0.86, + "learning_rate": 0.00011119759277833502, + "loss": 1.3183, + "theoretical_loss": 3.3627168392588684, + "tokens_seen": 2846752768 + }, + { + "epoch": 0.86, + "learning_rate": 0.00011118154463390172, + "loss": 1.2594, + "theoretical_loss": 3.36271087455916, + "tokens_seen": 2846818304 + }, + { + "epoch": 0.86, + "learning_rate": 0.00011116549648946842, + "loss": 1.2022, + "theoretical_loss": 3.362704910035209, + "tokens_seen": 2846883840 + }, + { + "epoch": 0.86, + "learning_rate": 0.00011114944834503512, + "loss": 1.2039, + "theoretical_loss": 3.3626989456870047, + "tokens_seen": 2846949376 + }, + { + "epoch": 0.86, + "learning_rate": 0.0001111334002006018, + "loss": 1.2521, + "theoretical_loss": 3.362692981514539, + "tokens_seen": 2847014912 + }, + { + "epoch": 0.86, + "learning_rate": 0.0001111173520561685, + "loss": 1.2913, + "theoretical_loss": 3.362687017517803, + "tokens_seen": 2847080448 + }, + { + "epoch": 0.86, + "learning_rate": 0.0001111013039117352, + "loss": 1.2589, + "theoretical_loss": 3.362681053696787, + "tokens_seen": 2847145984 + }, + { + "epoch": 0.86, + "learning_rate": 0.00011108525576730191, + "loss": 1.2139, + "theoretical_loss": 3.3626750900514812, + "tokens_seen": 2847211520 + }, + { + "epoch": 0.86, + "learning_rate": 0.00011106920762286861, + "loss": 1.2509, + "theoretical_loss": 3.362669126581877, + "tokens_seen": 2847277056 + }, + { + "epoch": 0.86, + "learning_rate": 0.00011105315947843531, + "loss": 1.2758, + "theoretical_loss": 3.3626631632879653, + "tokens_seen": 2847342592 + }, + { + "epoch": 0.86, + "learning_rate": 0.00011103711133400201, + "loss": 1.274, + "theoretical_loss": 3.362657200169737, + "tokens_seen": 2847408128 + }, + { + "epoch": 0.86, + "learning_rate": 0.00011102106318956871, + "loss": 1.324, + "theoretical_loss": 3.3626512372271824, + "tokens_seen": 2847473664 + }, + { + "epoch": 0.86, + "objective/train/docs_used": 1035131, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.470363736152649, + "objective/train/theoretical_loss": 3.3626452744602924, + "objective/train/tokens_used": 2867999200, + "theoretical_loss": 3.3626452744602924, + "tokens_seen": 2847539200 + }, + { + "epoch": 0.86, + "learning_rate": 0.00011100501504513541, + "loss": 1.4166, + "theoretical_loss": 3.3626452744602924, + "tokens_seen": 2847539200 + }, + { + "epoch": 0.86, + "learning_rate": 0.00011098896690070211, + "loss": 1.2331, + "theoretical_loss": 3.3626393118690574, + "tokens_seen": 2847604736 + }, + { + "epoch": 0.86, + "learning_rate": 0.00011097291875626881, + "loss": 1.2685, + "theoretical_loss": 3.3626333494534695, + "tokens_seen": 2847670272 + }, + { + "epoch": 0.86, + "learning_rate": 0.00011095687061183551, + "loss": 1.1455, + "theoretical_loss": 3.362627387213518, + "tokens_seen": 2847735808 + }, + { + "epoch": 0.86, + "learning_rate": 0.0001109408224674022, + "loss": 1.2143, + "theoretical_loss": 3.3626214251491944, + "tokens_seen": 2847801344 + }, + { + "epoch": 0.86, + "learning_rate": 0.0001109247743229689, + "loss": 1.3138, + "theoretical_loss": 3.3626154632604894, + "tokens_seen": 2847866880 + }, + { + "epoch": 0.86, + "learning_rate": 0.0001109087261785356, + "loss": 1.3051, + "theoretical_loss": 3.362609501547394, + "tokens_seen": 2847932416 + }, + { + "epoch": 0.86, + "learning_rate": 0.0001108926780341023, + "loss": 1.155, + "theoretical_loss": 3.3626035400098986, + "tokens_seen": 2847997952 + }, + { + "epoch": 0.86, + "learning_rate": 0.000110876629889669, + "loss": 1.318, + "theoretical_loss": 3.362597578647994, + "tokens_seen": 2848063488 + }, + { + "epoch": 0.86, + "learning_rate": 0.0001108605817452357, + "loss": 1.3072, + "theoretical_loss": 3.3625916174616712, + "tokens_seen": 2848129024 + }, + { + "epoch": 0.86, + "learning_rate": 0.0001108445336008024, + "loss": 1.3643, + "theoretical_loss": 3.362585656450921, + "tokens_seen": 2848194560 + }, + { + "epoch": 0.86, + "learning_rate": 0.0001108284854563691, + "loss": 1.2639, + "theoretical_loss": 3.362579695615734, + "tokens_seen": 2848260096 + }, + { + "epoch": 0.86, + "learning_rate": 0.0001108124373119358, + "loss": 1.2752, + "theoretical_loss": 3.3625737349561016, + "tokens_seen": 2848325632 + }, + { + "epoch": 0.86, + "learning_rate": 0.00011079638916750252, + "loss": 1.2523, + "theoretical_loss": 3.3625677744720135, + "tokens_seen": 2848391168 + }, + { + "epoch": 0.86, + "learning_rate": 0.00011078034102306922, + "loss": 1.2469, + "theoretical_loss": 3.362561814163461, + "tokens_seen": 2848456704 + }, + { + "epoch": 0.86, + "learning_rate": 0.00011076429287863592, + "loss": 1.2754, + "theoretical_loss": 3.362555854030435, + "tokens_seen": 2848522240 + }, + { + "epoch": 0.86, + "learning_rate": 0.00011074824473420262, + "loss": 1.2388, + "theoretical_loss": 3.362549894072927, + "tokens_seen": 2848587776 + }, + { + "epoch": 0.86, + "learning_rate": 0.00011073219658976932, + "loss": 1.2514, + "theoretical_loss": 3.3625439342909265, + "tokens_seen": 2848653312 + }, + { + "epoch": 0.86, + "learning_rate": 0.00011071614844533602, + "loss": 1.2097, + "theoretical_loss": 3.362537974684425, + "tokens_seen": 2848718848 + }, + { + "epoch": 0.86, + "learning_rate": 0.00011070010030090272, + "loss": 1.2767, + "theoretical_loss": 3.3625320152534126, + "tokens_seen": 2848784384 + }, + { + "epoch": 0.86, + "learning_rate": 0.00011068405215646942, + "loss": 1.3115, + "theoretical_loss": 3.3625260559978813, + "tokens_seen": 2848849920 + }, + { + "epoch": 0.86, + "learning_rate": 0.00011066800401203612, + "loss": 1.312, + "theoretical_loss": 3.3625200969178213, + "tokens_seen": 2848915456 + }, + { + "epoch": 0.86, + "learning_rate": 0.00011065195586760282, + "loss": 1.2243, + "theoretical_loss": 3.362514138013223, + "tokens_seen": 2848980992 + }, + { + "epoch": 0.86, + "learning_rate": 0.00011063590772316953, + "loss": 1.3267, + "theoretical_loss": 3.3625081792840774, + "tokens_seen": 2849046528 + }, + { + "epoch": 0.86, + "learning_rate": 0.00011061985957873621, + "loss": 1.2611, + "theoretical_loss": 3.362502220730376, + "tokens_seen": 2849112064 + }, + { + "epoch": 0.86, + "learning_rate": 0.00011060381143430291, + "loss": 1.2859, + "theoretical_loss": 3.362496262352108, + "tokens_seen": 2849177600 + }, + { + "epoch": 0.86, + "learning_rate": 0.00011058776328986961, + "loss": 1.2844, + "theoretical_loss": 3.3624903041492664, + "tokens_seen": 2849243136 + }, + { + "epoch": 0.86, + "learning_rate": 0.00011057171514543631, + "loss": 1.2792, + "theoretical_loss": 3.3624843461218403, + "tokens_seen": 2849308672 + }, + { + "epoch": 0.86, + "learning_rate": 0.00011055566700100301, + "loss": 1.3227, + "theoretical_loss": 3.362478388269821, + "tokens_seen": 2849374208 + }, + { + "epoch": 0.86, + "learning_rate": 0.00011053961885656972, + "loss": 1.2543, + "theoretical_loss": 3.3624724305931992, + "tokens_seen": 2849439744 + }, + { + "epoch": 0.86, + "learning_rate": 0.00011052357071213642, + "loss": 1.3289, + "theoretical_loss": 3.362466473091966, + "tokens_seen": 2849505280 + }, + { + "epoch": 0.86, + "learning_rate": 0.00011050752256770312, + "loss": 1.2951, + "theoretical_loss": 3.3624605157661125, + "tokens_seen": 2849570816 + }, + { + "epoch": 0.86, + "learning_rate": 0.00011049147442326982, + "loss": 1.2398, + "theoretical_loss": 3.3624545586156285, + "tokens_seen": 2849636352 + }, + { + "epoch": 0.86, + "learning_rate": 0.00011047542627883652, + "loss": 1.2866, + "theoretical_loss": 3.3624486016405055, + "tokens_seen": 2849701888 + }, + { + "epoch": 0.86, + "learning_rate": 0.00011045937813440322, + "loss": 1.2716, + "theoretical_loss": 3.362442644840734, + "tokens_seen": 2849767424 + }, + { + "epoch": 0.86, + "learning_rate": 0.00011044332998996992, + "loss": 1.1824, + "theoretical_loss": 3.362436688216305, + "tokens_seen": 2849832960 + }, + { + "epoch": 0.86, + "learning_rate": 0.0001104272818455366, + "loss": 1.2802, + "theoretical_loss": 3.3624307317672093, + "tokens_seen": 2849898496 + }, + { + "epoch": 0.86, + "learning_rate": 0.00011041123370110331, + "loss": 1.3613, + "theoretical_loss": 3.3624247754934378, + "tokens_seen": 2849964032 + }, + { + "epoch": 0.86, + "learning_rate": 0.00011039518555667001, + "loss": 1.3485, + "theoretical_loss": 3.362418819394981, + "tokens_seen": 2850029568 + }, + { + "epoch": 0.86, + "learning_rate": 0.00011037913741223671, + "loss": 1.3424, + "theoretical_loss": 3.36241286347183, + "tokens_seen": 2850095104 + }, + { + "epoch": 0.86, + "learning_rate": 0.00011036308926780341, + "loss": 1.3333, + "theoretical_loss": 3.3624069077239755, + "tokens_seen": 2850160640 + }, + { + "epoch": 0.86, + "learning_rate": 0.00011034704112337011, + "loss": 1.2939, + "theoretical_loss": 3.3624009521514084, + "tokens_seen": 2850226176 + }, + { + "epoch": 0.86, + "learning_rate": 0.00011033099297893681, + "loss": 1.2198, + "theoretical_loss": 3.3623949967541193, + "tokens_seen": 2850291712 + }, + { + "epoch": 0.86, + "learning_rate": 0.00011031494483450351, + "loss": 1.3078, + "theoretical_loss": 3.362389041532099, + "tokens_seen": 2850357248 + }, + { + "epoch": 0.86, + "learning_rate": 0.00011029889669007021, + "loss": 1.2174, + "theoretical_loss": 3.362383086485339, + "tokens_seen": 2850422784 + }, + { + "epoch": 0.86, + "learning_rate": 0.00011028284854563691, + "loss": 1.2429, + "theoretical_loss": 3.3623771316138287, + "tokens_seen": 2850488320 + }, + { + "epoch": 0.86, + "learning_rate": 0.00011026680040120361, + "loss": 1.3425, + "theoretical_loss": 3.36237117691756, + "tokens_seen": 2850553856 + }, + { + "epoch": 0.86, + "learning_rate": 0.00011025075225677031, + "loss": 1.3018, + "theoretical_loss": 3.362365222396524, + "tokens_seen": 2850619392 + }, + { + "epoch": 0.86, + "learning_rate": 0.000110234704112337, + "loss": 1.253, + "theoretical_loss": 3.3623592680507106, + "tokens_seen": 2850684928 + }, + { + "epoch": 0.86, + "learning_rate": 0.0001102186559679037, + "loss": 1.3571, + "theoretical_loss": 3.3623533138801114, + "tokens_seen": 2850750464 + }, + { + "epoch": 0.86, + "objective/train/docs_used": 1036421, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.2483084201812744, + "objective/train/theoretical_loss": 3.3623473598847164, + "objective/train/tokens_used": 2871276000, + "theoretical_loss": 3.3623473598847164, + "tokens_seen": 2850816000 + }, + { + "epoch": 0.86, + "learning_rate": 0.00011020260782347043, + "loss": 1.2398, + "theoretical_loss": 3.3623473598847164, + "tokens_seen": 2850816000 + }, + { + "epoch": 0.86, + "learning_rate": 0.00011018655967903713, + "loss": 1.2034, + "theoretical_loss": 3.362341406064517, + "tokens_seen": 2850881536 + }, + { + "epoch": 0.86, + "learning_rate": 0.00011017051153460383, + "loss": 1.2063, + "theoretical_loss": 3.362335452419504, + "tokens_seen": 2850947072 + }, + { + "epoch": 0.86, + "learning_rate": 0.00011015446339017053, + "loss": 1.2886, + "theoretical_loss": 3.3623294989496677, + "tokens_seen": 2851012608 + }, + { + "epoch": 0.86, + "learning_rate": 0.00011013841524573723, + "loss": 1.3212, + "theoretical_loss": 3.362323545655, + "tokens_seen": 2851078144 + }, + { + "epoch": 0.86, + "learning_rate": 0.00011012236710130392, + "loss": 1.3438, + "theoretical_loss": 3.36231759253549, + "tokens_seen": 2851143680 + }, + { + "epoch": 0.86, + "learning_rate": 0.00011010631895687062, + "loss": 1.0911, + "theoretical_loss": 3.3623116395911303, + "tokens_seen": 2851209216 + }, + { + "epoch": 0.86, + "learning_rate": 0.00011009027081243732, + "loss": 1.2617, + "theoretical_loss": 3.362305686821911, + "tokens_seen": 2851274752 + }, + { + "epoch": 0.86, + "learning_rate": 0.00011007422266800402, + "loss": 1.2953, + "theoretical_loss": 3.3622997342278227, + "tokens_seen": 2851340288 + }, + { + "epoch": 0.86, + "learning_rate": 0.00011005817452357072, + "loss": 1.2519, + "theoretical_loss": 3.362293781808856, + "tokens_seen": 2851405824 + }, + { + "epoch": 0.86, + "learning_rate": 0.00011004212637913742, + "loss": 1.2562, + "theoretical_loss": 3.3622878295650027, + "tokens_seen": 2851471360 + }, + { + "epoch": 0.86, + "learning_rate": 0.00011002607823470412, + "loss": 1.2862, + "theoretical_loss": 3.362281877496253, + "tokens_seen": 2851536896 + }, + { + "epoch": 0.86, + "learning_rate": 0.00011001003009027082, + "loss": 1.1743, + "theoretical_loss": 3.3622759256025976, + "tokens_seen": 2851602432 + }, + { + "epoch": 0.86, + "learning_rate": 0.00010999398194583752, + "loss": 1.2558, + "theoretical_loss": 3.362269973884027, + "tokens_seen": 2851667968 + }, + { + "epoch": 0.86, + "learning_rate": 0.00010997793380140422, + "loss": 1.2501, + "theoretical_loss": 3.3622640223405336, + "tokens_seen": 2851733504 + }, + { + "epoch": 0.86, + "learning_rate": 0.00010996188565697092, + "loss": 1.2289, + "theoretical_loss": 3.3622580709721066, + "tokens_seen": 2851799040 + }, + { + "epoch": 0.86, + "learning_rate": 0.00010994583751253763, + "loss": 1.1914, + "theoretical_loss": 3.3622521197787374, + "tokens_seen": 2851864576 + }, + { + "epoch": 0.86, + "learning_rate": 0.00010992978936810431, + "loss": 1.2115, + "theoretical_loss": 3.3622461687604166, + "tokens_seen": 2851930112 + }, + { + "epoch": 0.86, + "learning_rate": 0.00010991374122367101, + "loss": 1.3355, + "theoretical_loss": 3.362240217917136, + "tokens_seen": 2851995648 + }, + { + "epoch": 0.86, + "learning_rate": 0.00010989769307923771, + "loss": 1.2522, + "theoretical_loss": 3.3622342672488847, + "tokens_seen": 2852061184 + }, + { + "epoch": 0.86, + "learning_rate": 0.00010988164493480441, + "loss": 1.2083, + "theoretical_loss": 3.362228316755655, + "tokens_seen": 2852126720 + }, + { + "epoch": 0.86, + "learning_rate": 0.00010986559679037111, + "loss": 1.3143, + "theoretical_loss": 3.3622223664374373, + "tokens_seen": 2852192256 + }, + { + "epoch": 0.86, + "learning_rate": 0.00010984954864593782, + "loss": 1.241, + "theoretical_loss": 3.3622164162942223, + "tokens_seen": 2852257792 + }, + { + "epoch": 0.86, + "learning_rate": 0.00010983350050150452, + "loss": 1.185, + "theoretical_loss": 3.3622104663260006, + "tokens_seen": 2852323328 + }, + { + "epoch": 0.86, + "learning_rate": 0.00010981745235707122, + "loss": 1.3075, + "theoretical_loss": 3.3622045165327634, + "tokens_seen": 2852388864 + }, + { + "epoch": 0.86, + "learning_rate": 0.00010980140421263792, + "loss": 1.3158, + "theoretical_loss": 3.362198566914502, + "tokens_seen": 2852454400 + }, + { + "epoch": 0.86, + "learning_rate": 0.00010978535606820462, + "loss": 1.3527, + "theoretical_loss": 3.362192617471206, + "tokens_seen": 2852519936 + }, + { + "epoch": 0.86, + "learning_rate": 0.00010976930792377132, + "loss": 1.325, + "theoretical_loss": 3.3621866682028676, + "tokens_seen": 2852585472 + }, + { + "epoch": 0.86, + "learning_rate": 0.00010975325977933802, + "loss": 1.2238, + "theoretical_loss": 3.3621807191094764, + "tokens_seen": 2852651008 + }, + { + "epoch": 0.86, + "learning_rate": 0.0001097372116349047, + "loss": 1.1684, + "theoretical_loss": 3.3621747701910243, + "tokens_seen": 2852716544 + }, + { + "epoch": 0.86, + "learning_rate": 0.00010972116349047141, + "loss": 1.4172, + "theoretical_loss": 3.362168821447501, + "tokens_seen": 2852782080 + }, + { + "epoch": 0.86, + "learning_rate": 0.00010970511534603811, + "loss": 1.2698, + "theoretical_loss": 3.3621628728788986, + "tokens_seen": 2852847616 + }, + { + "epoch": 0.86, + "learning_rate": 0.00010968906720160481, + "loss": 1.2999, + "theoretical_loss": 3.362156924485207, + "tokens_seen": 2852913152 + }, + { + "epoch": 0.86, + "learning_rate": 0.00010967301905717151, + "loss": 1.2813, + "theoretical_loss": 3.3621509762664177, + "tokens_seen": 2852978688 + }, + { + "epoch": 0.86, + "learning_rate": 0.00010965697091273821, + "loss": 1.2346, + "theoretical_loss": 3.3621450282225207, + "tokens_seen": 2853044224 + }, + { + "epoch": 0.86, + "learning_rate": 0.00010964092276830491, + "loss": 1.2603, + "theoretical_loss": 3.362139080353508, + "tokens_seen": 2853109760 + }, + { + "epoch": 0.86, + "learning_rate": 0.00010962487462387161, + "loss": 1.2566, + "theoretical_loss": 3.362133132659369, + "tokens_seen": 2853175296 + }, + { + "epoch": 0.86, + "learning_rate": 0.00010960882647943833, + "loss": 1.2748, + "theoretical_loss": 3.362127185140096, + "tokens_seen": 2853240832 + }, + { + "epoch": 0.86, + "learning_rate": 0.00010959277833500503, + "loss": 1.2854, + "theoretical_loss": 3.3621212377956784, + "tokens_seen": 2853306368 + }, + { + "epoch": 0.86, + "learning_rate": 0.00010957673019057173, + "loss": 1.2472, + "theoretical_loss": 3.3621152906261083, + "tokens_seen": 2853371904 + }, + { + "epoch": 0.86, + "learning_rate": 0.00010956068204613843, + "loss": 1.2494, + "theoretical_loss": 3.362109343631376, + "tokens_seen": 2853437440 + }, + { + "epoch": 0.86, + "learning_rate": 0.00010954463390170513, + "loss": 1.2887, + "theoretical_loss": 3.3621033968114724, + "tokens_seen": 2853502976 + }, + { + "epoch": 0.86, + "learning_rate": 0.00010952858575727183, + "loss": 1.226, + "theoretical_loss": 3.362097450166388, + "tokens_seen": 2853568512 + }, + { + "epoch": 0.86, + "learning_rate": 0.00010951253761283853, + "loss": 1.2609, + "theoretical_loss": 3.3620915036961145, + "tokens_seen": 2853634048 + }, + { + "epoch": 0.86, + "learning_rate": 0.00010949648946840523, + "loss": 1.2937, + "theoretical_loss": 3.362085557400642, + "tokens_seen": 2853699584 + }, + { + "epoch": 0.86, + "learning_rate": 0.00010948044132397193, + "loss": 1.2226, + "theoretical_loss": 3.362079611279962, + "tokens_seen": 2853765120 + }, + { + "epoch": 0.86, + "learning_rate": 0.00010946439317953863, + "loss": 1.2936, + "theoretical_loss": 3.3620736653340644, + "tokens_seen": 2853830656 + }, + { + "epoch": 0.86, + "learning_rate": 0.00010944834503510533, + "loss": 1.3376, + "theoretical_loss": 3.3620677195629405, + "tokens_seen": 2853896192 + }, + { + "epoch": 0.86, + "learning_rate": 0.00010943229689067203, + "loss": 1.2254, + "theoretical_loss": 3.3620617739665817, + "tokens_seen": 2853961728 + }, + { + "epoch": 0.86, + "learning_rate": 0.00010941624874623872, + "loss": 1.3463, + "theoretical_loss": 3.362055828544978, + "tokens_seen": 2854027264 + }, + { + "epoch": 0.86, + "objective/train/docs_used": 1037150, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.2597136497497559, + "objective/train/theoretical_loss": 3.3620498832981207, + "objective/train/tokens_used": 2874552800, + "theoretical_loss": 3.3620498832981207, + "tokens_seen": 2854092800 + }, + { + "epoch": 0.86, + "learning_rate": 0.00010940020060180542, + "loss": 1.256, + "theoretical_loss": 3.3620498832981207, + "tokens_seen": 2854092800 + }, + { + "epoch": 0.86, + "learning_rate": 0.00010938415245737212, + "loss": 1.3051, + "theoretical_loss": 3.362043938226001, + "tokens_seen": 2854158336 + }, + { + "epoch": 0.86, + "learning_rate": 0.00010936810431293882, + "loss": 1.281, + "theoretical_loss": 3.3620379933286086, + "tokens_seen": 2854223872 + }, + { + "epoch": 0.86, + "learning_rate": 0.00010935205616850552, + "loss": 1.2766, + "theoretical_loss": 3.3620320486059354, + "tokens_seen": 2854289408 + }, + { + "epoch": 0.86, + "learning_rate": 0.00010933600802407222, + "loss": 1.372, + "theoretical_loss": 3.362026104057972, + "tokens_seen": 2854354944 + }, + { + "epoch": 0.86, + "learning_rate": 0.00010931995987963892, + "loss": 1.2618, + "theoretical_loss": 3.3620201596847092, + "tokens_seen": 2854420480 + }, + { + "epoch": 0.86, + "learning_rate": 0.00010930391173520562, + "loss": 1.2625, + "theoretical_loss": 3.362014215486138, + "tokens_seen": 2854486016 + }, + { + "epoch": 0.87, + "learning_rate": 0.00010928786359077232, + "loss": 1.3141, + "theoretical_loss": 3.362008271462249, + "tokens_seen": 2854551552 + }, + { + "epoch": 0.87, + "learning_rate": 0.00010927181544633902, + "loss": 1.3062, + "theoretical_loss": 3.362002327613033, + "tokens_seen": 2854617088 + }, + { + "epoch": 0.87, + "learning_rate": 0.00010925576730190573, + "loss": 1.2942, + "theoretical_loss": 3.361996383938481, + "tokens_seen": 2854682624 + }, + { + "epoch": 0.87, + "learning_rate": 0.00010923971915747243, + "loss": 1.2927, + "theoretical_loss": 3.361990440438584, + "tokens_seen": 2854748160 + }, + { + "epoch": 0.87, + "learning_rate": 0.00010922367101303911, + "loss": 1.2773, + "theoretical_loss": 3.361984497113333, + "tokens_seen": 2854813696 + }, + { + "epoch": 0.87, + "learning_rate": 0.00010920762286860581, + "loss": 1.3383, + "theoretical_loss": 3.361978553962718, + "tokens_seen": 2854879232 + }, + { + "epoch": 0.87, + "learning_rate": 0.00010919157472417251, + "loss": 1.2445, + "theoretical_loss": 3.3619726109867307, + "tokens_seen": 2854944768 + }, + { + "epoch": 0.87, + "learning_rate": 0.00010917552657973922, + "loss": 1.2598, + "theoretical_loss": 3.361966668185362, + "tokens_seen": 2855010304 + }, + { + "epoch": 0.87, + "learning_rate": 0.00010915947843530592, + "loss": 1.3074, + "theoretical_loss": 3.3619607255586024, + "tokens_seen": 2855075840 + }, + { + "epoch": 0.87, + "learning_rate": 0.00010914343029087262, + "loss": 1.3601, + "theoretical_loss": 3.3619547831064427, + "tokens_seen": 2855141376 + }, + { + "epoch": 0.87, + "learning_rate": 0.00010912738214643932, + "loss": 1.2195, + "theoretical_loss": 3.3619488408288736, + "tokens_seen": 2855206912 + }, + { + "epoch": 0.87, + "learning_rate": 0.00010911133400200602, + "loss": 1.2179, + "theoretical_loss": 3.3619428987258866, + "tokens_seen": 2855272448 + }, + { + "epoch": 0.87, + "learning_rate": 0.00010909528585757272, + "loss": 1.2941, + "theoretical_loss": 3.361936956797472, + "tokens_seen": 2855337984 + }, + { + "epoch": 0.87, + "learning_rate": 0.00010907923771313942, + "loss": 1.269, + "theoretical_loss": 3.3619310150436212, + "tokens_seen": 2855403520 + }, + { + "epoch": 0.87, + "learning_rate": 0.00010906318956870612, + "loss": 1.3588, + "theoretical_loss": 3.3619250734643247, + "tokens_seen": 2855469056 + }, + { + "epoch": 0.87, + "learning_rate": 0.00010904714142427282, + "loss": 1.2522, + "theoretical_loss": 3.3619191320595734, + "tokens_seen": 2855534592 + }, + { + "epoch": 0.87, + "learning_rate": 0.00010903109327983951, + "loss": 1.2625, + "theoretical_loss": 3.361913190829358, + "tokens_seen": 2855600128 + }, + { + "epoch": 0.87, + "learning_rate": 0.00010901504513540621, + "loss": 1.3557, + "theoretical_loss": 3.3619072497736697, + "tokens_seen": 2855665664 + }, + { + "epoch": 0.87, + "learning_rate": 0.00010899899699097294, + "loss": 1.2086, + "theoretical_loss": 3.361901308892499, + "tokens_seen": 2855731200 + }, + { + "epoch": 0.87, + "learning_rate": 0.00010898294884653964, + "loss": 1.2985, + "theoretical_loss": 3.361895368185837, + "tokens_seen": 2855796736 + }, + { + "epoch": 0.87, + "learning_rate": 0.00010896690070210634, + "loss": 1.2981, + "theoretical_loss": 3.361889427653675, + "tokens_seen": 2855862272 + }, + { + "epoch": 0.87, + "learning_rate": 0.00010895085255767304, + "loss": 1.3287, + "theoretical_loss": 3.3618834872960033, + "tokens_seen": 2855927808 + }, + { + "epoch": 0.87, + "learning_rate": 0.00010893480441323974, + "loss": 1.3936, + "theoretical_loss": 3.3618775471128126, + "tokens_seen": 2855993344 + }, + { + "epoch": 0.87, + "learning_rate": 0.00010891875626880644, + "loss": 1.345, + "theoretical_loss": 3.3618716071040944, + "tokens_seen": 2856058880 + }, + { + "epoch": 0.87, + "learning_rate": 0.00010890270812437313, + "loss": 1.1979, + "theoretical_loss": 3.361865667269839, + "tokens_seen": 2856124416 + }, + { + "epoch": 0.87, + "learning_rate": 0.00010888665997993983, + "loss": 1.2439, + "theoretical_loss": 3.3618597276100375, + "tokens_seen": 2856189952 + }, + { + "epoch": 0.87, + "learning_rate": 0.00010887061183550653, + "loss": 1.1864, + "theoretical_loss": 3.361853788124681, + "tokens_seen": 2856255488 + }, + { + "epoch": 0.87, + "learning_rate": 0.00010885456369107323, + "loss": 1.3109, + "theoretical_loss": 3.3618478488137598, + "tokens_seen": 2856321024 + }, + { + "epoch": 0.87, + "learning_rate": 0.00010883851554663993, + "loss": 1.2338, + "theoretical_loss": 3.3618419096772656, + "tokens_seen": 2856386560 + }, + { + "epoch": 0.87, + "learning_rate": 0.00010882246740220663, + "loss": 1.3316, + "theoretical_loss": 3.3618359707151884, + "tokens_seen": 2856452096 + }, + { + "epoch": 0.87, + "learning_rate": 0.00010880641925777333, + "loss": 1.4107, + "theoretical_loss": 3.3618300319275196, + "tokens_seen": 2856517632 + }, + { + "epoch": 0.87, + "learning_rate": 0.00010879037111334003, + "loss": 1.3035, + "theoretical_loss": 3.36182409331425, + "tokens_seen": 2856583168 + }, + { + "epoch": 0.87, + "learning_rate": 0.00010877432296890673, + "loss": 1.1985, + "theoretical_loss": 3.3618181548753707, + "tokens_seen": 2856648704 + }, + { + "epoch": 0.87, + "learning_rate": 0.00010875827482447343, + "loss": 1.3495, + "theoretical_loss": 3.3618122166108724, + "tokens_seen": 2856714240 + }, + { + "epoch": 0.87, + "learning_rate": 0.00010874222668004013, + "loss": 1.3054, + "theoretical_loss": 3.3618062785207456, + "tokens_seen": 2856779776 + }, + { + "epoch": 0.87, + "learning_rate": 0.00010872617853560683, + "loss": 1.2611, + "theoretical_loss": 3.3618003406049817, + "tokens_seen": 2856845312 + }, + { + "epoch": 0.87, + "learning_rate": 0.00010871013039117352, + "loss": 1.245, + "theoretical_loss": 3.361794402863571, + "tokens_seen": 2856910848 + }, + { + "epoch": 0.87, + "learning_rate": 0.00010869408224674022, + "loss": 1.3253, + "theoretical_loss": 3.361788465296505, + "tokens_seen": 2856976384 + }, + { + "epoch": 0.87, + "learning_rate": 0.00010867803410230692, + "loss": 1.2213, + "theoretical_loss": 3.3617825279037743, + "tokens_seen": 2857041920 + }, + { + "epoch": 0.87, + "learning_rate": 0.00010866198595787362, + "loss": 1.2255, + "theoretical_loss": 3.36177659068537, + "tokens_seen": 2857107456 + }, + { + "epoch": 0.87, + "learning_rate": 0.00010864593781344032, + "loss": 1.2867, + "theoretical_loss": 3.3617706536412824, + "tokens_seen": 2857172992 + }, + { + "epoch": 0.87, + "learning_rate": 0.00010862988966900702, + "loss": 1.2175, + "theoretical_loss": 3.361764716771503, + "tokens_seen": 2857238528 + }, + { + "epoch": 0.87, + "learning_rate": 0.00010861384152457372, + "loss": 1.2619, + "theoretical_loss": 3.3617587800760225, + "tokens_seen": 2857304064 + }, + { + "epoch": 0.87, + "objective/train/docs_used": 1038225, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.3500590324401855, + "objective/train/theoretical_loss": 3.3617528435548314, + "objective/train/tokens_used": 2877829600, + "theoretical_loss": 3.3617528435548314, + "tokens_seen": 2857369600 + }, + { + "epoch": 0.87, + "learning_rate": 0.00010859779338014042, + "loss": 1.3546, + "theoretical_loss": 3.3617528435548314, + "tokens_seen": 2857369600 + }, + { + "epoch": 0.87, + "learning_rate": 0.00010858174523570713, + "loss": 1.2929, + "theoretical_loss": 3.3617469072079214, + "tokens_seen": 2857435136 + }, + { + "epoch": 0.87, + "learning_rate": 0.00010856569709127383, + "loss": 1.2899, + "theoretical_loss": 3.3617409710352826, + "tokens_seen": 2857500672 + }, + { + "epoch": 0.87, + "learning_rate": 0.00010854964894684053, + "loss": 1.3321, + "theoretical_loss": 3.361735035036906, + "tokens_seen": 2857566208 + }, + { + "epoch": 0.87, + "learning_rate": 0.00010853360080240723, + "loss": 1.193, + "theoretical_loss": 3.3617290992127833, + "tokens_seen": 2857631744 + }, + { + "epoch": 0.87, + "learning_rate": 0.00010851755265797391, + "loss": 1.3808, + "theoretical_loss": 3.3617231635629046, + "tokens_seen": 2857697280 + }, + { + "epoch": 0.87, + "learning_rate": 0.00010850150451354061, + "loss": 1.2491, + "theoretical_loss": 3.3617172280872607, + "tokens_seen": 2857762816 + }, + { + "epoch": 0.87, + "learning_rate": 0.00010848545636910732, + "loss": 1.2817, + "theoretical_loss": 3.3617112927858432, + "tokens_seen": 2857828352 + }, + { + "epoch": 0.87, + "learning_rate": 0.00010846940822467402, + "loss": 1.1914, + "theoretical_loss": 3.361705357658642, + "tokens_seen": 2857893888 + }, + { + "epoch": 0.87, + "learning_rate": 0.00010845336008024072, + "loss": 1.33, + "theoretical_loss": 3.361699422705649, + "tokens_seen": 2857959424 + }, + { + "epoch": 0.87, + "learning_rate": 0.00010843731193580742, + "loss": 1.3427, + "theoretical_loss": 3.3616934879268543, + "tokens_seen": 2858024960 + }, + { + "epoch": 0.87, + "learning_rate": 0.00010842126379137412, + "loss": 1.3218, + "theoretical_loss": 3.3616875533222492, + "tokens_seen": 2858090496 + }, + { + "epoch": 0.87, + "learning_rate": 0.00010840521564694083, + "loss": 1.3467, + "theoretical_loss": 3.361681618891825, + "tokens_seen": 2858156032 + }, + { + "epoch": 0.87, + "learning_rate": 0.00010838916750250753, + "loss": 1.2379, + "theoretical_loss": 3.3616756846355718, + "tokens_seen": 2858221568 + }, + { + "epoch": 0.87, + "learning_rate": 0.00010837311935807423, + "loss": 1.1906, + "theoretical_loss": 3.3616697505534807, + "tokens_seen": 2858287104 + }, + { + "epoch": 0.87, + "learning_rate": 0.00010835707121364093, + "loss": 1.245, + "theoretical_loss": 3.3616638166455424, + "tokens_seen": 2858352640 + }, + { + "epoch": 0.87, + "learning_rate": 0.00010834102306920763, + "loss": 1.3522, + "theoretical_loss": 3.3616578829117487, + "tokens_seen": 2858418176 + }, + { + "epoch": 0.87, + "learning_rate": 0.00010832497492477434, + "loss": 1.3697, + "theoretical_loss": 3.3616519493520896, + "tokens_seen": 2858483712 + }, + { + "epoch": 0.87, + "learning_rate": 0.00010830892678034104, + "loss": 1.3726, + "theoretical_loss": 3.3616460159665564, + "tokens_seen": 2858549248 + }, + { + "epoch": 0.87, + "learning_rate": 0.00010829287863590774, + "loss": 1.2337, + "theoretical_loss": 3.36164008275514, + "tokens_seen": 2858614784 + }, + { + "epoch": 0.87, + "learning_rate": 0.00010827683049147444, + "loss": 1.3405, + "theoretical_loss": 3.361634149717831, + "tokens_seen": 2858680320 + }, + { + "epoch": 0.87, + "learning_rate": 0.00010826078234704114, + "loss": 1.3728, + "theoretical_loss": 3.3616282168546205, + "tokens_seen": 2858745856 + }, + { + "epoch": 0.87, + "learning_rate": 0.00010824473420260784, + "loss": 1.3708, + "theoretical_loss": 3.3616222841654992, + "tokens_seen": 2858811392 + }, + { + "epoch": 0.87, + "learning_rate": 0.00010822868605817454, + "loss": 1.2917, + "theoretical_loss": 3.3616163516504587, + "tokens_seen": 2858876928 + }, + { + "epoch": 0.87, + "learning_rate": 0.00010821263791374123, + "loss": 1.2298, + "theoretical_loss": 3.361610419309489, + "tokens_seen": 2858942464 + }, + { + "epoch": 0.87, + "learning_rate": 0.00010819658976930793, + "loss": 1.3007, + "theoretical_loss": 3.3616044871425816, + "tokens_seen": 2859008000 + }, + { + "epoch": 0.87, + "learning_rate": 0.00010818054162487463, + "loss": 1.3276, + "theoretical_loss": 3.361598555149727, + "tokens_seen": 2859073536 + }, + { + "epoch": 0.87, + "learning_rate": 0.00010816449348044133, + "loss": 1.3207, + "theoretical_loss": 3.3615926233309166, + "tokens_seen": 2859139072 + }, + { + "epoch": 0.87, + "learning_rate": 0.00010814844533600803, + "loss": 1.307, + "theoretical_loss": 3.3615866916861408, + "tokens_seen": 2859204608 + }, + { + "epoch": 0.87, + "learning_rate": 0.00010813239719157473, + "loss": 1.3081, + "theoretical_loss": 3.361580760215391, + "tokens_seen": 2859270144 + }, + { + "epoch": 0.87, + "learning_rate": 0.00010811634904714143, + "loss": 1.215, + "theoretical_loss": 3.3615748289186573, + "tokens_seen": 2859335680 + }, + { + "epoch": 0.87, + "learning_rate": 0.00010810030090270813, + "loss": 1.2521, + "theoretical_loss": 3.3615688977959315, + "tokens_seen": 2859401216 + }, + { + "epoch": 0.87, + "learning_rate": 0.00010808425275827483, + "loss": 1.2993, + "theoretical_loss": 3.3615629668472042, + "tokens_seen": 2859466752 + }, + { + "epoch": 0.87, + "learning_rate": 0.00010806820461384153, + "loss": 1.3682, + "theoretical_loss": 3.361557036072466, + "tokens_seen": 2859532288 + }, + { + "epoch": 0.87, + "learning_rate": 0.00010805215646940823, + "loss": 1.3406, + "theoretical_loss": 3.361551105471708, + "tokens_seen": 2859597824 + }, + { + "epoch": 0.87, + "learning_rate": 0.00010803610832497493, + "loss": 1.2529, + "theoretical_loss": 3.3615451750449212, + "tokens_seen": 2859663360 + }, + { + "epoch": 0.87, + "learning_rate": 0.00010802006018054162, + "loss": 1.1449, + "theoretical_loss": 3.361539244792097, + "tokens_seen": 2859728896 + }, + { + "epoch": 0.87, + "learning_rate": 0.00010800401203610832, + "loss": 1.2876, + "theoretical_loss": 3.361533314713225, + "tokens_seen": 2859794432 + }, + { + "epoch": 0.87, + "learning_rate": 0.00010798796389167502, + "loss": 1.3277, + "theoretical_loss": 3.3615273848082974, + "tokens_seen": 2859859968 + }, + { + "epoch": 0.87, + "learning_rate": 0.00010797191574724172, + "loss": 1.1652, + "theoretical_loss": 3.3615214550773045, + "tokens_seen": 2859925504 + }, + { + "epoch": 0.87, + "learning_rate": 0.00010795586760280842, + "loss": 1.3796, + "theoretical_loss": 3.3615155255202374, + "tokens_seen": 2859991040 + }, + { + "epoch": 0.87, + "learning_rate": 0.00010793981945837512, + "loss": 1.2444, + "theoretical_loss": 3.3615095961370867, + "tokens_seen": 2860056576 + }, + { + "epoch": 0.87, + "learning_rate": 0.00010792377131394182, + "loss": 1.2765, + "theoretical_loss": 3.3615036669278435, + "tokens_seen": 2860122112 + }, + { + "epoch": 0.87, + "learning_rate": 0.00010790772316950852, + "loss": 1.2832, + "theoretical_loss": 3.361497737892499, + "tokens_seen": 2860187648 + }, + { + "epoch": 0.87, + "learning_rate": 0.00010789167502507523, + "loss": 1.3709, + "theoretical_loss": 3.361491809031044, + "tokens_seen": 2860253184 + }, + { + "epoch": 0.87, + "learning_rate": 0.00010787562688064193, + "loss": 1.3088, + "theoretical_loss": 3.361485880343469, + "tokens_seen": 2860318720 + }, + { + "epoch": 0.87, + "learning_rate": 0.00010785957873620863, + "loss": 1.246, + "theoretical_loss": 3.3614799518297653, + "tokens_seen": 2860384256 + }, + { + "epoch": 0.87, + "learning_rate": 0.00010784353059177533, + "loss": 1.235, + "theoretical_loss": 3.3614740234899236, + "tokens_seen": 2860449792 + }, + { + "epoch": 0.87, + "learning_rate": 0.00010782748244734201, + "loss": 1.3342, + "theoretical_loss": 3.361468095323935, + "tokens_seen": 2860515328 + }, + { + "epoch": 0.87, + "learning_rate": 0.00010781143430290871, + "loss": 1.3866, + "theoretical_loss": 3.3614621673317906, + "tokens_seen": 2860580864 + }, + { + "epoch": 0.87, + "objective/train/docs_used": 1039526, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.180164098739624, + "objective/train/theoretical_loss": 3.3614562395134806, + "objective/train/tokens_used": 2881106400, + "theoretical_loss": 3.3614562395134806, + "tokens_seen": 2860646400 + }, + { + "epoch": 0.87, + "learning_rate": 0.00010779538615847544, + "loss": 1.2683, + "theoretical_loss": 3.3614562395134806, + "tokens_seen": 2860646400 + }, + { + "epoch": 0.87, + "learning_rate": 0.00010777933801404214, + "loss": 1.2557, + "theoretical_loss": 3.361450311868997, + "tokens_seen": 2860711936 + }, + { + "epoch": 0.87, + "learning_rate": 0.00010776328986960884, + "loss": 1.2651, + "theoretical_loss": 3.3614443843983297, + "tokens_seen": 2860777472 + }, + { + "epoch": 0.87, + "learning_rate": 0.00010774724172517554, + "loss": 1.3353, + "theoretical_loss": 3.36143845710147, + "tokens_seen": 2860843008 + }, + { + "epoch": 0.87, + "learning_rate": 0.00010773119358074225, + "loss": 1.3216, + "theoretical_loss": 3.3614325299784094, + "tokens_seen": 2860908544 + }, + { + "epoch": 0.87, + "learning_rate": 0.00010771514543630895, + "loss": 1.3133, + "theoretical_loss": 3.3614266030291375, + "tokens_seen": 2860974080 + }, + { + "epoch": 0.87, + "learning_rate": 0.00010769909729187563, + "loss": 1.2268, + "theoretical_loss": 3.3614206762536467, + "tokens_seen": 2861039616 + }, + { + "epoch": 0.87, + "learning_rate": 0.00010768304914744233, + "loss": 1.2081, + "theoretical_loss": 3.361414749651927, + "tokens_seen": 2861105152 + }, + { + "epoch": 0.87, + "learning_rate": 0.00010766700100300903, + "loss": 1.2942, + "theoretical_loss": 3.3614088232239694, + "tokens_seen": 2861170688 + }, + { + "epoch": 0.87, + "learning_rate": 0.00010765095285857573, + "loss": 1.305, + "theoretical_loss": 3.361402896969765, + "tokens_seen": 2861236224 + }, + { + "epoch": 0.87, + "learning_rate": 0.00010763490471414244, + "loss": 1.2954, + "theoretical_loss": 3.361396970889305, + "tokens_seen": 2861301760 + }, + { + "epoch": 0.87, + "learning_rate": 0.00010761885656970914, + "loss": 1.2055, + "theoretical_loss": 3.3613910449825797, + "tokens_seen": 2861367296 + }, + { + "epoch": 0.87, + "learning_rate": 0.00010760280842527584, + "loss": 1.3584, + "theoretical_loss": 3.3613851192495803, + "tokens_seen": 2861432832 + }, + { + "epoch": 0.87, + "learning_rate": 0.00010758676028084254, + "loss": 1.1733, + "theoretical_loss": 3.361379193690298, + "tokens_seen": 2861498368 + }, + { + "epoch": 0.87, + "learning_rate": 0.00010757071213640924, + "loss": 1.3247, + "theoretical_loss": 3.3613732683047237, + "tokens_seen": 2861563904 + }, + { + "epoch": 0.87, + "learning_rate": 0.00010755466399197594, + "loss": 1.2066, + "theoretical_loss": 3.361367343092848, + "tokens_seen": 2861629440 + }, + { + "epoch": 0.87, + "learning_rate": 0.00010753861584754264, + "loss": 1.2753, + "theoretical_loss": 3.361361418054662, + "tokens_seen": 2861694976 + }, + { + "epoch": 0.87, + "learning_rate": 0.00010752256770310934, + "loss": 1.3446, + "theoretical_loss": 3.361355493190157, + "tokens_seen": 2861760512 + }, + { + "epoch": 0.87, + "learning_rate": 0.00010750651955867603, + "loss": 1.3441, + "theoretical_loss": 3.361349568499323, + "tokens_seen": 2861826048 + }, + { + "epoch": 0.87, + "learning_rate": 0.00010749047141424273, + "loss": 1.2007, + "theoretical_loss": 3.3613436439821514, + "tokens_seen": 2861891584 + }, + { + "epoch": 0.87, + "learning_rate": 0.00010747442326980943, + "loss": 1.2032, + "theoretical_loss": 3.3613377196386334, + "tokens_seen": 2861957120 + }, + { + "epoch": 0.87, + "learning_rate": 0.00010745837512537613, + "loss": 1.2275, + "theoretical_loss": 3.3613317954687596, + "tokens_seen": 2862022656 + }, + { + "epoch": 0.87, + "learning_rate": 0.00010744232698094283, + "loss": 1.2932, + "theoretical_loss": 3.361325871472521, + "tokens_seen": 2862088192 + }, + { + "epoch": 0.87, + "learning_rate": 0.00010742627883650953, + "loss": 1.3237, + "theoretical_loss": 3.361319947649909, + "tokens_seen": 2862153728 + }, + { + "epoch": 0.87, + "learning_rate": 0.00010741023069207623, + "loss": 1.3452, + "theoretical_loss": 3.3613140240009143, + "tokens_seen": 2862219264 + }, + { + "epoch": 0.87, + "learning_rate": 0.00010739418254764293, + "loss": 1.1996, + "theoretical_loss": 3.361308100525527, + "tokens_seen": 2862284800 + }, + { + "epoch": 0.87, + "learning_rate": 0.00010737813440320963, + "loss": 1.2777, + "theoretical_loss": 3.3613021772237395, + "tokens_seen": 2862350336 + }, + { + "epoch": 0.87, + "learning_rate": 0.00010736208625877633, + "loss": 1.2178, + "theoretical_loss": 3.3612962540955413, + "tokens_seen": 2862415872 + }, + { + "epoch": 0.87, + "learning_rate": 0.00010734603811434303, + "loss": 1.2691, + "theoretical_loss": 3.3612903311409243, + "tokens_seen": 2862481408 + }, + { + "epoch": 0.87, + "learning_rate": 0.00010732998996990973, + "loss": 1.3217, + "theoretical_loss": 3.3612844083598796, + "tokens_seen": 2862546944 + }, + { + "epoch": 0.87, + "learning_rate": 0.00010731394182547642, + "loss": 1.2421, + "theoretical_loss": 3.361278485752397, + "tokens_seen": 2862612480 + }, + { + "epoch": 0.87, + "learning_rate": 0.00010729789368104312, + "loss": 1.2845, + "theoretical_loss": 3.3612725633184684, + "tokens_seen": 2862678016 + }, + { + "epoch": 0.87, + "learning_rate": 0.00010728184553660982, + "loss": 1.351, + "theoretical_loss": 3.3612666410580845, + "tokens_seen": 2862743552 + }, + { + "epoch": 0.87, + "learning_rate": 0.00010726579739217652, + "loss": 1.2688, + "theoretical_loss": 3.361260718971236, + "tokens_seen": 2862809088 + }, + { + "epoch": 0.87, + "learning_rate": 0.00010724974924774322, + "loss": 1.2382, + "theoretical_loss": 3.361254797057914, + "tokens_seen": 2862874624 + }, + { + "epoch": 0.87, + "learning_rate": 0.00010723370110330992, + "loss": 1.2666, + "theoretical_loss": 3.36124887531811, + "tokens_seen": 2862940160 + }, + { + "epoch": 0.87, + "learning_rate": 0.00010721765295887662, + "loss": 1.2937, + "theoretical_loss": 3.361242953751814, + "tokens_seen": 2863005696 + }, + { + "epoch": 0.87, + "learning_rate": 0.00010720160481444335, + "loss": 1.255, + "theoretical_loss": 3.3612370323590173, + "tokens_seen": 2863071232 + }, + { + "epoch": 0.87, + "learning_rate": 0.00010718555667001004, + "loss": 1.2096, + "theoretical_loss": 3.361231111139711, + "tokens_seen": 2863136768 + }, + { + "epoch": 0.87, + "learning_rate": 0.00010716950852557674, + "loss": 1.3798, + "theoretical_loss": 3.3612251900938857, + "tokens_seen": 2863202304 + }, + { + "epoch": 0.87, + "learning_rate": 0.00010715346038114344, + "loss": 1.2645, + "theoretical_loss": 3.3612192692215332, + "tokens_seen": 2863267840 + }, + { + "epoch": 0.87, + "learning_rate": 0.00010713741223671014, + "loss": 1.2767, + "theoretical_loss": 3.3612133485226434, + "tokens_seen": 2863333376 + }, + { + "epoch": 0.87, + "learning_rate": 0.00010712136409227684, + "loss": 1.2861, + "theoretical_loss": 3.361207427997208, + "tokens_seen": 2863398912 + }, + { + "epoch": 0.87, + "learning_rate": 0.00010710531594784354, + "loss": 1.2385, + "theoretical_loss": 3.3612015076452177, + "tokens_seen": 2863464448 + }, + { + "epoch": 0.87, + "learning_rate": 0.00010708926780341024, + "loss": 1.2436, + "theoretical_loss": 3.3611955874666632, + "tokens_seen": 2863529984 + }, + { + "epoch": 0.87, + "learning_rate": 0.00010707321965897694, + "loss": 1.306, + "theoretical_loss": 3.361189667461536, + "tokens_seen": 2863595520 + }, + { + "epoch": 0.87, + "learning_rate": 0.00010705717151454364, + "loss": 1.3136, + "theoretical_loss": 3.361183747629826, + "tokens_seen": 2863661056 + }, + { + "epoch": 0.87, + "learning_rate": 0.00010704112337011035, + "loss": 1.1225, + "theoretical_loss": 3.3611778279715256, + "tokens_seen": 2863726592 + }, + { + "epoch": 0.87, + "learning_rate": 0.00010702507522567705, + "loss": 1.1788, + "theoretical_loss": 3.3611719084866247, + "tokens_seen": 2863792128 + }, + { + "epoch": 0.87, + "learning_rate": 0.00010700902708124375, + "loss": 1.2505, + "theoretical_loss": 3.361165989175115, + "tokens_seen": 2863857664 + }, + { + "epoch": 0.87, + "objective/train/docs_used": 1040737, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.2356295585632324, + "objective/train/theoretical_loss": 3.3611600700369864, + "objective/train/tokens_used": 2884383200, + "theoretical_loss": 3.3611600700369864, + "tokens_seen": 2863923200 + }, + { + "epoch": 0.87, + "learning_rate": 0.00010699297893681043, + "loss": 1.1992, + "theoretical_loss": 3.3611600700369864, + "tokens_seen": 2863923200 + }, + { + "epoch": 0.87, + "learning_rate": 0.00010697693079237713, + "loss": 1.278, + "theoretical_loss": 3.3611541510722307, + "tokens_seen": 2863988736 + }, + { + "epoch": 0.87, + "learning_rate": 0.00010696088264794383, + "loss": 1.2823, + "theoretical_loss": 3.3611482322808386, + "tokens_seen": 2864054272 + }, + { + "epoch": 0.87, + "learning_rate": 0.00010694483450351054, + "loss": 1.2331, + "theoretical_loss": 3.361142313662801, + "tokens_seen": 2864119808 + }, + { + "epoch": 0.87, + "learning_rate": 0.00010692878635907724, + "loss": 1.3873, + "theoretical_loss": 3.361136395218109, + "tokens_seen": 2864185344 + }, + { + "epoch": 0.87, + "learning_rate": 0.00010691273821464394, + "loss": 1.2817, + "theoretical_loss": 3.3611304769467534, + "tokens_seen": 2864250880 + }, + { + "epoch": 0.87, + "learning_rate": 0.00010689669007021064, + "loss": 1.2181, + "theoretical_loss": 3.3611245588487257, + "tokens_seen": 2864316416 + }, + { + "epoch": 0.87, + "learning_rate": 0.00010688064192577734, + "loss": 1.2704, + "theoretical_loss": 3.361118640924016, + "tokens_seen": 2864381952 + }, + { + "epoch": 0.87, + "learning_rate": 0.00010686459378134404, + "loss": 1.2433, + "theoretical_loss": 3.361112723172616, + "tokens_seen": 2864447488 + }, + { + "epoch": 0.87, + "learning_rate": 0.00010684854563691074, + "loss": 1.3648, + "theoretical_loss": 3.361106805594516, + "tokens_seen": 2864513024 + }, + { + "epoch": 0.87, + "learning_rate": 0.00010683249749247744, + "loss": 1.3297, + "theoretical_loss": 3.361100888189707, + "tokens_seen": 2864578560 + }, + { + "epoch": 0.87, + "learning_rate": 0.00010681644934804414, + "loss": 1.2348, + "theoretical_loss": 3.361094970958181, + "tokens_seen": 2864644096 + }, + { + "epoch": 0.87, + "learning_rate": 0.00010680040120361083, + "loss": 1.2629, + "theoretical_loss": 3.361089053899928, + "tokens_seen": 2864709632 + }, + { + "epoch": 0.87, + "learning_rate": 0.00010678435305917753, + "loss": 1.2372, + "theoretical_loss": 3.3610831370149388, + "tokens_seen": 2864775168 + }, + { + "epoch": 0.87, + "learning_rate": 0.00010676830491474423, + "loss": 1.3078, + "theoretical_loss": 3.361077220303205, + "tokens_seen": 2864840704 + }, + { + "epoch": 0.87, + "learning_rate": 0.00010675225677031093, + "loss": 1.2971, + "theoretical_loss": 3.3610713037647173, + "tokens_seen": 2864906240 + }, + { + "epoch": 0.87, + "learning_rate": 0.00010673620862587763, + "loss": 1.2987, + "theoretical_loss": 3.3610653873994667, + "tokens_seen": 2864971776 + }, + { + "epoch": 0.87, + "learning_rate": 0.00010672016048144433, + "loss": 1.2692, + "theoretical_loss": 3.361059471207444, + "tokens_seen": 2865037312 + }, + { + "epoch": 0.87, + "learning_rate": 0.00010670411233701103, + "loss": 1.2261, + "theoretical_loss": 3.3610535551886405, + "tokens_seen": 2865102848 + }, + { + "epoch": 0.87, + "learning_rate": 0.00010668806419257773, + "loss": 1.2897, + "theoretical_loss": 3.361047639343047, + "tokens_seen": 2865168384 + }, + { + "epoch": 0.87, + "learning_rate": 0.00010667201604814443, + "loss": 1.3114, + "theoretical_loss": 3.3610417236706542, + "tokens_seen": 2865233920 + }, + { + "epoch": 0.87, + "learning_rate": 0.00010665596790371113, + "loss": 1.232, + "theoretical_loss": 3.3610358081714535, + "tokens_seen": 2865299456 + }, + { + "epoch": 0.87, + "learning_rate": 0.00010663991975927783, + "loss": 1.2877, + "theoretical_loss": 3.3610298928454356, + "tokens_seen": 2865364992 + }, + { + "epoch": 0.87, + "learning_rate": 0.00010662387161484453, + "loss": 1.3243, + "theoretical_loss": 3.361023977692591, + "tokens_seen": 2865430528 + }, + { + "epoch": 0.87, + "learning_rate": 0.00010660782347041122, + "loss": 1.306, + "theoretical_loss": 3.361018062712912, + "tokens_seen": 2865496064 + }, + { + "epoch": 0.87, + "learning_rate": 0.00010659177532597795, + "loss": 1.356, + "theoretical_loss": 3.3610121479063886, + "tokens_seen": 2865561600 + }, + { + "epoch": 0.87, + "learning_rate": 0.00010657572718154465, + "loss": 1.3453, + "theoretical_loss": 3.3610062332730117, + "tokens_seen": 2865627136 + }, + { + "epoch": 0.87, + "learning_rate": 0.00010655967903711135, + "loss": 1.2422, + "theoretical_loss": 3.3610003188127724, + "tokens_seen": 2865692672 + }, + { + "epoch": 0.87, + "learning_rate": 0.00010654363089267805, + "loss": 1.2895, + "theoretical_loss": 3.360994404525662, + "tokens_seen": 2865758208 + }, + { + "epoch": 0.87, + "learning_rate": 0.00010652758274824475, + "loss": 1.2656, + "theoretical_loss": 3.3609884904116716, + "tokens_seen": 2865823744 + }, + { + "epoch": 0.87, + "learning_rate": 0.00010651153460381145, + "loss": 1.2981, + "theoretical_loss": 3.3609825764707915, + "tokens_seen": 2865889280 + }, + { + "epoch": 0.87, + "learning_rate": 0.00010649548645937814, + "loss": 1.2807, + "theoretical_loss": 3.3609766627030133, + "tokens_seen": 2865954816 + }, + { + "epoch": 0.87, + "learning_rate": 0.00010647943831494484, + "loss": 1.3114, + "theoretical_loss": 3.3609707491083274, + "tokens_seen": 2866020352 + }, + { + "epoch": 0.87, + "learning_rate": 0.00010646339017051154, + "loss": 1.1844, + "theoretical_loss": 3.3609648356867248, + "tokens_seen": 2866085888 + }, + { + "epoch": 0.87, + "learning_rate": 0.00010644734202607824, + "loss": 1.2135, + "theoretical_loss": 3.360958922438197, + "tokens_seen": 2866151424 + }, + { + "epoch": 0.87, + "learning_rate": 0.00010643129388164494, + "loss": 1.2785, + "theoretical_loss": 3.360953009362735, + "tokens_seen": 2866216960 + }, + { + "epoch": 0.87, + "learning_rate": 0.00010641524573721164, + "loss": 1.2807, + "theoretical_loss": 3.360947096460329, + "tokens_seen": 2866282496 + }, + { + "epoch": 0.87, + "learning_rate": 0.00010639919759277834, + "loss": 1.2148, + "theoretical_loss": 3.360941183730971, + "tokens_seen": 2866348032 + }, + { + "epoch": 0.87, + "learning_rate": 0.00010638314944834504, + "loss": 1.3552, + "theoretical_loss": 3.360935271174651, + "tokens_seen": 2866413568 + }, + { + "epoch": 0.87, + "learning_rate": 0.00010636710130391174, + "loss": 1.3352, + "theoretical_loss": 3.3609293587913607, + "tokens_seen": 2866479104 + }, + { + "epoch": 0.87, + "learning_rate": 0.00010635105315947845, + "loss": 1.3264, + "theoretical_loss": 3.360923446581091, + "tokens_seen": 2866544640 + }, + { + "epoch": 0.87, + "learning_rate": 0.00010633500501504515, + "loss": 1.235, + "theoretical_loss": 3.360917534543832, + "tokens_seen": 2866610176 + }, + { + "epoch": 0.87, + "learning_rate": 0.00010631895687061185, + "loss": 1.268, + "theoretical_loss": 3.360911622679576, + "tokens_seen": 2866675712 + }, + { + "epoch": 0.87, + "learning_rate": 0.00010630290872617853, + "loss": 1.2671, + "theoretical_loss": 3.3609057109883134, + "tokens_seen": 2866741248 + }, + { + "epoch": 0.87, + "learning_rate": 0.00010628686058174523, + "loss": 1.2533, + "theoretical_loss": 3.3608997994700345, + "tokens_seen": 2866806784 + }, + { + "epoch": 0.87, + "learning_rate": 0.00010627081243731194, + "loss": 1.2162, + "theoretical_loss": 3.3608938881247314, + "tokens_seen": 2866872320 + }, + { + "epoch": 0.87, + "learning_rate": 0.00010625476429287864, + "loss": 1.2481, + "theoretical_loss": 3.3608879769523945, + "tokens_seen": 2866937856 + }, + { + "epoch": 0.87, + "learning_rate": 0.00010623871614844534, + "loss": 1.3062, + "theoretical_loss": 3.3608820659530148, + "tokens_seen": 2867003392 + }, + { + "epoch": 0.87, + "learning_rate": 0.00010622266800401204, + "loss": 1.2867, + "theoretical_loss": 3.3608761551265838, + "tokens_seen": 2867068928 + }, + { + "epoch": 0.87, + "learning_rate": 0.00010620661985957874, + "loss": 1.3298, + "theoretical_loss": 3.360870244473092, + "tokens_seen": 2867134464 + }, + { + "epoch": 0.87, + "objective/train/docs_used": 1042051, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.2878918647766113, + "objective/train/theoretical_loss": 3.36086433399253, + "objective/train/tokens_used": 2887660000, + "theoretical_loss": 3.36086433399253, + "tokens_seen": 2867200000 + }, + { + "epoch": 0.87, + "learning_rate": 0.00010619057171514544, + "loss": 1.3374, + "theoretical_loss": 3.36086433399253, + "tokens_seen": 2867200000 + }, + { + "epoch": 0.87, + "learning_rate": 0.00010617452357071214, + "loss": 1.168, + "theoretical_loss": 3.3608584236848893, + "tokens_seen": 2867265536 + }, + { + "epoch": 0.87, + "learning_rate": 0.00010615847542627884, + "loss": 1.2872, + "theoretical_loss": 3.360852513550161, + "tokens_seen": 2867331072 + }, + { + "epoch": 0.87, + "learning_rate": 0.00010614242728184554, + "loss": 1.3046, + "theoretical_loss": 3.360846603588336, + "tokens_seen": 2867396608 + }, + { + "epoch": 0.87, + "learning_rate": 0.00010612637913741224, + "loss": 1.2187, + "theoretical_loss": 3.360840693799405, + "tokens_seen": 2867462144 + }, + { + "epoch": 0.87, + "learning_rate": 0.00010611033099297893, + "loss": 1.2866, + "theoretical_loss": 3.3608347841833597, + "tokens_seen": 2867527680 + }, + { + "epoch": 0.87, + "learning_rate": 0.00010609428284854563, + "loss": 1.2567, + "theoretical_loss": 3.36082887474019, + "tokens_seen": 2867593216 + }, + { + "epoch": 0.87, + "learning_rate": 0.00010607823470411233, + "loss": 1.2752, + "theoretical_loss": 3.360822965469888, + "tokens_seen": 2867658752 + }, + { + "epoch": 0.87, + "learning_rate": 0.00010606218655967903, + "loss": 1.2751, + "theoretical_loss": 3.3608170563724435, + "tokens_seen": 2867724288 + }, + { + "epoch": 0.87, + "learning_rate": 0.00010604613841524573, + "loss": 1.2531, + "theoretical_loss": 3.3608111474478486, + "tokens_seen": 2867789824 + }, + { + "epoch": 0.87, + "learning_rate": 0.00010603009027081243, + "loss": 1.2408, + "theoretical_loss": 3.360805238696094, + "tokens_seen": 2867855360 + }, + { + "epoch": 0.87, + "learning_rate": 0.00010601404212637913, + "loss": 1.1905, + "theoretical_loss": 3.3607993301171706, + "tokens_seen": 2867920896 + }, + { + "epoch": 0.87, + "learning_rate": 0.00010599799398194586, + "loss": 1.2877, + "theoretical_loss": 3.360793421711069, + "tokens_seen": 2867986432 + }, + { + "epoch": 0.87, + "learning_rate": 0.00010598194583751255, + "loss": 1.3264, + "theoretical_loss": 3.360787513477781, + "tokens_seen": 2868051968 + }, + { + "epoch": 0.87, + "learning_rate": 0.00010596589769307925, + "loss": 1.2238, + "theoretical_loss": 3.360781605417297, + "tokens_seen": 2868117504 + }, + { + "epoch": 0.87, + "learning_rate": 0.00010594984954864595, + "loss": 1.2305, + "theoretical_loss": 3.3607756975296077, + "tokens_seen": 2868183040 + }, + { + "epoch": 0.87, + "learning_rate": 0.00010593380140421265, + "loss": 1.2144, + "theoretical_loss": 3.360769789814705, + "tokens_seen": 2868248576 + }, + { + "epoch": 0.87, + "learning_rate": 0.00010591775325977935, + "loss": 1.258, + "theoretical_loss": 3.3607638822725794, + "tokens_seen": 2868314112 + }, + { + "epoch": 0.87, + "learning_rate": 0.00010590170511534605, + "loss": 1.2672, + "theoretical_loss": 3.3607579749032217, + "tokens_seen": 2868379648 + }, + { + "epoch": 0.87, + "learning_rate": 0.00010588565697091275, + "loss": 1.2922, + "theoretical_loss": 3.360752067706623, + "tokens_seen": 2868445184 + }, + { + "epoch": 0.87, + "learning_rate": 0.00010586960882647945, + "loss": 1.4352, + "theoretical_loss": 3.3607461606827753, + "tokens_seen": 2868510720 + }, + { + "epoch": 0.87, + "learning_rate": 0.00010585356068204615, + "loss": 1.1924, + "theoretical_loss": 3.360740253831668, + "tokens_seen": 2868576256 + }, + { + "epoch": 0.87, + "learning_rate": 0.00010583751253761285, + "loss": 1.2001, + "theoretical_loss": 3.360734347153293, + "tokens_seen": 2868641792 + }, + { + "epoch": 0.87, + "learning_rate": 0.00010582146439317955, + "loss": 1.3442, + "theoretical_loss": 3.360728440647641, + "tokens_seen": 2868707328 + }, + { + "epoch": 0.87, + "learning_rate": 0.00010580541624874625, + "loss": 1.3413, + "theoretical_loss": 3.3607225343147036, + "tokens_seen": 2868772864 + }, + { + "epoch": 0.87, + "learning_rate": 0.00010578936810431294, + "loss": 1.3599, + "theoretical_loss": 3.360716628154471, + "tokens_seen": 2868838400 + }, + { + "epoch": 0.87, + "learning_rate": 0.00010577331995987964, + "loss": 1.2737, + "theoretical_loss": 3.360710722166935, + "tokens_seen": 2868903936 + }, + { + "epoch": 0.87, + "learning_rate": 0.00010575727181544634, + "loss": 1.1928, + "theoretical_loss": 3.3607048163520856, + "tokens_seen": 2868969472 + }, + { + "epoch": 0.87, + "learning_rate": 0.00010574122367101304, + "loss": 1.4087, + "theoretical_loss": 3.3606989107099148, + "tokens_seen": 2869035008 + }, + { + "epoch": 0.87, + "learning_rate": 0.00010572517552657974, + "loss": 1.2031, + "theoretical_loss": 3.360693005240413, + "tokens_seen": 2869100544 + }, + { + "epoch": 0.87, + "learning_rate": 0.00010570912738214644, + "loss": 1.2605, + "theoretical_loss": 3.360687099943571, + "tokens_seen": 2869166080 + }, + { + "epoch": 0.87, + "learning_rate": 0.00010569307923771314, + "loss": 1.2857, + "theoretical_loss": 3.3606811948193807, + "tokens_seen": 2869231616 + }, + { + "epoch": 0.87, + "learning_rate": 0.00010567703109327985, + "loss": 1.2579, + "theoretical_loss": 3.3606752898678325, + "tokens_seen": 2869297152 + }, + { + "epoch": 0.87, + "learning_rate": 0.00010566098294884655, + "loss": 1.3382, + "theoretical_loss": 3.3606693850889178, + "tokens_seen": 2869362688 + }, + { + "epoch": 0.87, + "learning_rate": 0.00010564493480441325, + "loss": 1.2566, + "theoretical_loss": 3.3606634804826268, + "tokens_seen": 2869428224 + }, + { + "epoch": 0.87, + "learning_rate": 0.00010562888665997995, + "loss": 1.3313, + "theoretical_loss": 3.3606575760489514, + "tokens_seen": 2869493760 + }, + { + "epoch": 0.87, + "learning_rate": 0.00010561283851554665, + "loss": 1.3257, + "theoretical_loss": 3.360651671787882, + "tokens_seen": 2869559296 + }, + { + "epoch": 0.87, + "learning_rate": 0.00010559679037111333, + "loss": 1.2915, + "theoretical_loss": 3.36064576769941, + "tokens_seen": 2869624832 + }, + { + "epoch": 0.87, + "learning_rate": 0.00010558074222668004, + "loss": 1.287, + "theoretical_loss": 3.3606398637835264, + "tokens_seen": 2869690368 + }, + { + "epoch": 0.87, + "learning_rate": 0.00010556469408224674, + "loss": 1.3591, + "theoretical_loss": 3.360633960040222, + "tokens_seen": 2869755904 + }, + { + "epoch": 0.87, + "learning_rate": 0.00010554864593781344, + "loss": 1.2143, + "theoretical_loss": 3.360628056469488, + "tokens_seen": 2869821440 + }, + { + "epoch": 0.87, + "learning_rate": 0.00010553259779338014, + "loss": 1.1978, + "theoretical_loss": 3.360622153071315, + "tokens_seen": 2869886976 + }, + { + "epoch": 0.87, + "learning_rate": 0.00010551654964894684, + "loss": 1.1928, + "theoretical_loss": 3.3606162498456946, + "tokens_seen": 2869952512 + }, + { + "epoch": 0.87, + "learning_rate": 0.00010550050150451354, + "loss": 1.1688, + "theoretical_loss": 3.360610346792617, + "tokens_seen": 2870018048 + }, + { + "epoch": 0.87, + "learning_rate": 0.00010548445336008024, + "loss": 1.223, + "theoretical_loss": 3.3606044439120746, + "tokens_seen": 2870083584 + }, + { + "epoch": 0.87, + "learning_rate": 0.00010546840521564694, + "loss": 1.3601, + "theoretical_loss": 3.360598541204057, + "tokens_seen": 2870149120 + }, + { + "epoch": 0.87, + "learning_rate": 0.00010545235707121364, + "loss": 1.2282, + "theoretical_loss": 3.3605926386685563, + "tokens_seen": 2870214656 + }, + { + "epoch": 0.87, + "learning_rate": 0.00010543630892678034, + "loss": 1.2886, + "theoretical_loss": 3.360586736305563, + "tokens_seen": 2870280192 + }, + { + "epoch": 0.87, + "learning_rate": 0.00010542026078234704, + "loss": 1.2181, + "theoretical_loss": 3.3605808341150674, + "tokens_seen": 2870345728 + }, + { + "epoch": 0.87, + "learning_rate": 0.00010540421263791376, + "loss": 1.1513, + "theoretical_loss": 3.360574932097062, + "tokens_seen": 2870411264 + }, + { + "epoch": 0.87, + "objective/train/docs_used": 1043211, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.1554994583129883, + "objective/train/theoretical_loss": 3.3605690302515367, + "objective/train/tokens_used": 2890936800, + "theoretical_loss": 3.3605690302515367, + "tokens_seen": 2870476800 + }, + { + "epoch": 0.87, + "learning_rate": 0.00010538816449348046, + "loss": 1.262, + "theoretical_loss": 3.3605690302515367, + "tokens_seen": 2870476800 + }, + { + "epoch": 0.87, + "learning_rate": 0.00010537211634904716, + "loss": 1.3388, + "theoretical_loss": 3.360563128578483, + "tokens_seen": 2870542336 + }, + { + "epoch": 0.87, + "learning_rate": 0.00010535606820461386, + "loss": 1.2061, + "theoretical_loss": 3.3605572270778916, + "tokens_seen": 2870607872 + }, + { + "epoch": 0.87, + "learning_rate": 0.00010534002006018056, + "loss": 1.2094, + "theoretical_loss": 3.3605513257497543, + "tokens_seen": 2870673408 + }, + { + "epoch": 0.87, + "learning_rate": 0.00010532397191574726, + "loss": 1.329, + "theoretical_loss": 3.3605454245940614, + "tokens_seen": 2870738944 + }, + { + "epoch": 0.87, + "learning_rate": 0.00010530792377131396, + "loss": 1.2597, + "theoretical_loss": 3.3605395236108038, + "tokens_seen": 2870804480 + }, + { + "epoch": 0.87, + "learning_rate": 0.00010529187562688066, + "loss": 1.2297, + "theoretical_loss": 3.3605336227999727, + "tokens_seen": 2870870016 + }, + { + "epoch": 0.87, + "learning_rate": 0.00010527582748244735, + "loss": 1.2431, + "theoretical_loss": 3.3605277221615597, + "tokens_seen": 2870935552 + }, + { + "epoch": 0.87, + "learning_rate": 0.00010525977933801405, + "loss": 1.2058, + "theoretical_loss": 3.3605218216955555, + "tokens_seen": 2871001088 + }, + { + "epoch": 0.87, + "learning_rate": 0.00010524373119358075, + "loss": 1.2476, + "theoretical_loss": 3.3605159214019507, + "tokens_seen": 2871066624 + }, + { + "epoch": 0.87, + "learning_rate": 0.00010522768304914745, + "loss": 1.29, + "theoretical_loss": 3.3605100212807364, + "tokens_seen": 2871132160 + }, + { + "epoch": 0.87, + "learning_rate": 0.00010521163490471415, + "loss": 1.2238, + "theoretical_loss": 3.3605041213319042, + "tokens_seen": 2871197696 + }, + { + "epoch": 0.87, + "learning_rate": 0.00010519558676028085, + "loss": 1.4107, + "theoretical_loss": 3.360498221555445, + "tokens_seen": 2871263232 + }, + { + "epoch": 0.87, + "learning_rate": 0.00010517953861584755, + "loss": 1.3012, + "theoretical_loss": 3.360492321951349, + "tokens_seen": 2871328768 + }, + { + "epoch": 0.87, + "learning_rate": 0.00010516349047141425, + "loss": 1.3941, + "theoretical_loss": 3.360486422519608, + "tokens_seen": 2871394304 + }, + { + "epoch": 0.87, + "learning_rate": 0.00010514744232698095, + "loss": 1.3128, + "theoretical_loss": 3.3604805232602133, + "tokens_seen": 2871459840 + }, + { + "epoch": 0.87, + "learning_rate": 0.00010513139418254765, + "loss": 1.3028, + "theoretical_loss": 3.360474624173155, + "tokens_seen": 2871525376 + }, + { + "epoch": 0.87, + "learning_rate": 0.00010511534603811435, + "loss": 1.2669, + "theoretical_loss": 3.360468725258425, + "tokens_seen": 2871590912 + }, + { + "epoch": 0.87, + "learning_rate": 0.00010509929789368105, + "loss": 1.358, + "theoretical_loss": 3.360462826516014, + "tokens_seen": 2871656448 + }, + { + "epoch": 0.87, + "learning_rate": 0.00010508324974924774, + "loss": 1.3131, + "theoretical_loss": 3.360456927945913, + "tokens_seen": 2871721984 + }, + { + "epoch": 0.87, + "learning_rate": 0.00010506720160481444, + "loss": 1.3161, + "theoretical_loss": 3.3604510295481127, + "tokens_seen": 2871787520 + }, + { + "epoch": 0.87, + "learning_rate": 0.00010505115346038114, + "loss": 1.4121, + "theoretical_loss": 3.3604451313226047, + "tokens_seen": 2871853056 + }, + { + "epoch": 0.87, + "learning_rate": 0.00010503510531594784, + "loss": 1.3467, + "theoretical_loss": 3.3604392332693798, + "tokens_seen": 2871918592 + }, + { + "epoch": 0.87, + "learning_rate": 0.00010501905717151454, + "loss": 1.3505, + "theoretical_loss": 3.360433335388429, + "tokens_seen": 2871984128 + }, + { + "epoch": 0.87, + "learning_rate": 0.00010500300902708124, + "loss": 1.2416, + "theoretical_loss": 3.3604274376797436, + "tokens_seen": 2872049664 + }, + { + "epoch": 0.87, + "learning_rate": 0.00010498696088264795, + "loss": 1.3255, + "theoretical_loss": 3.360421540143314, + "tokens_seen": 2872115200 + }, + { + "epoch": 0.87, + "learning_rate": 0.00010497091273821465, + "loss": 1.2834, + "theoretical_loss": 3.3604156427791323, + "tokens_seen": 2872180736 + }, + { + "epoch": 0.87, + "learning_rate": 0.00010495486459378135, + "loss": 1.2555, + "theoretical_loss": 3.360409745587188, + "tokens_seen": 2872246272 + }, + { + "epoch": 0.87, + "learning_rate": 0.00010493881644934805, + "loss": 1.2426, + "theoretical_loss": 3.360403848567474, + "tokens_seen": 2872311808 + }, + { + "epoch": 0.87, + "learning_rate": 0.00010492276830491475, + "loss": 1.2436, + "theoretical_loss": 3.36039795171998, + "tokens_seen": 2872377344 + }, + { + "epoch": 0.87, + "learning_rate": 0.00010490672016048145, + "loss": 1.3301, + "theoretical_loss": 3.3603920550446977, + "tokens_seen": 2872442880 + }, + { + "epoch": 0.87, + "learning_rate": 0.00010489067201604814, + "loss": 1.2738, + "theoretical_loss": 3.3603861585416173, + "tokens_seen": 2872508416 + }, + { + "epoch": 0.87, + "learning_rate": 0.00010487462387161484, + "loss": 1.1879, + "theoretical_loss": 3.3603802622107306, + "tokens_seen": 2872573952 + }, + { + "epoch": 0.87, + "learning_rate": 0.00010485857572718154, + "loss": 1.1762, + "theoretical_loss": 3.360374366052029, + "tokens_seen": 2872639488 + }, + { + "epoch": 0.87, + "learning_rate": 0.00010484252758274824, + "loss": 1.2331, + "theoretical_loss": 3.3603684700655023, + "tokens_seen": 2872705024 + }, + { + "epoch": 0.87, + "learning_rate": 0.00010482647943831494, + "loss": 1.182, + "theoretical_loss": 3.3603625742511425, + "tokens_seen": 2872770560 + }, + { + "epoch": 0.87, + "learning_rate": 0.00010481043129388164, + "loss": 1.2301, + "theoretical_loss": 3.36035667860894, + "tokens_seen": 2872836096 + }, + { + "epoch": 0.87, + "learning_rate": 0.00010479438314944837, + "loss": 1.407, + "theoretical_loss": 3.3603507831388866, + "tokens_seen": 2872901632 + }, + { + "epoch": 0.87, + "learning_rate": 0.00010477833500501507, + "loss": 1.2686, + "theoretical_loss": 3.360344887840973, + "tokens_seen": 2872967168 + }, + { + "epoch": 0.87, + "learning_rate": 0.00010476228686058175, + "loss": 1.2915, + "theoretical_loss": 3.3603389927151905, + "tokens_seen": 2873032704 + }, + { + "epoch": 0.87, + "learning_rate": 0.00010474623871614845, + "loss": 1.2673, + "theoretical_loss": 3.3603330977615293, + "tokens_seen": 2873098240 + }, + { + "epoch": 0.87, + "learning_rate": 0.00010473019057171516, + "loss": 1.4093, + "theoretical_loss": 3.3603272029799816, + "tokens_seen": 2873163776 + }, + { + "epoch": 0.87, + "learning_rate": 0.00010471414242728186, + "loss": 1.3589, + "theoretical_loss": 3.3603213083705374, + "tokens_seen": 2873229312 + }, + { + "epoch": 0.87, + "learning_rate": 0.00010469809428284856, + "loss": 1.2739, + "theoretical_loss": 3.3603154139331886, + "tokens_seen": 2873294848 + }, + { + "epoch": 0.87, + "learning_rate": 0.00010468204613841526, + "loss": 1.3009, + "theoretical_loss": 3.3603095196679256, + "tokens_seen": 2873360384 + }, + { + "epoch": 0.87, + "learning_rate": 0.00010466599799398196, + "loss": 1.2238, + "theoretical_loss": 3.3603036255747396, + "tokens_seen": 2873425920 + }, + { + "epoch": 0.87, + "learning_rate": 0.00010464994984954866, + "loss": 1.2636, + "theoretical_loss": 3.360297731653622, + "tokens_seen": 2873491456 + }, + { + "epoch": 0.87, + "learning_rate": 0.00010463390170511536, + "loss": 1.3392, + "theoretical_loss": 3.3602918379045636, + "tokens_seen": 2873556992 + }, + { + "epoch": 0.87, + "learning_rate": 0.00010461785356068206, + "loss": 1.316, + "theoretical_loss": 3.3602859443275555, + "tokens_seen": 2873622528 + }, + { + "epoch": 0.87, + "learning_rate": 0.00010460180541624876, + "loss": 1.2064, + "theoretical_loss": 3.3602800509225887, + "tokens_seen": 2873688064 + }, + { + "epoch": 0.87, + "objective/train/docs_used": 1044447, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.252935767173767, + "objective/train/theoretical_loss": 3.360274157689654, + "objective/train/tokens_used": 2894213600, + "theoretical_loss": 3.360274157689654, + "tokens_seen": 2873753600 + }, + { + "epoch": 0.87, + "learning_rate": 0.00010458575727181546, + "loss": 1.2474, + "theoretical_loss": 3.360274157689654, + "tokens_seen": 2873753600 + }, + { + "epoch": 0.87, + "learning_rate": 0.00010456970912738215, + "loss": 1.2155, + "theoretical_loss": 3.3602682646287434, + "tokens_seen": 2873819136 + }, + { + "epoch": 0.87, + "learning_rate": 0.00010455366098294885, + "loss": 1.159, + "theoretical_loss": 3.360262371739847, + "tokens_seen": 2873884672 + }, + { + "epoch": 0.87, + "learning_rate": 0.00010453761283851555, + "loss": 1.2122, + "theoretical_loss": 3.360256479022956, + "tokens_seen": 2873950208 + }, + { + "epoch": 0.87, + "learning_rate": 0.00010452156469408225, + "loss": 1.2391, + "theoretical_loss": 3.360250586478062, + "tokens_seen": 2874015744 + }, + { + "epoch": 0.87, + "learning_rate": 0.00010450551654964895, + "loss": 1.2839, + "theoretical_loss": 3.3602446941051554, + "tokens_seen": 2874081280 + }, + { + "epoch": 0.87, + "learning_rate": 0.00010448946840521565, + "loss": 1.3748, + "theoretical_loss": 3.3602388019042273, + "tokens_seen": 2874146816 + }, + { + "epoch": 0.87, + "learning_rate": 0.00010447342026078235, + "loss": 1.37, + "theoretical_loss": 3.3602329098752692, + "tokens_seen": 2874212352 + }, + { + "epoch": 0.87, + "learning_rate": 0.00010445737211634905, + "loss": 1.2567, + "theoretical_loss": 3.360227018018272, + "tokens_seen": 2874277888 + }, + { + "epoch": 0.87, + "learning_rate": 0.00010444132397191575, + "loss": 1.2798, + "theoretical_loss": 3.3602211263332267, + "tokens_seen": 2874343424 + }, + { + "epoch": 0.87, + "learning_rate": 0.00010442527582748245, + "loss": 1.2674, + "theoretical_loss": 3.3602152348201244, + "tokens_seen": 2874408960 + }, + { + "epoch": 0.87, + "learning_rate": 0.00010440922768304915, + "loss": 1.1884, + "theoretical_loss": 3.3602093434789557, + "tokens_seen": 2874474496 + }, + { + "epoch": 0.87, + "learning_rate": 0.00010439317953861586, + "loss": 1.195, + "theoretical_loss": 3.3602034523097126, + "tokens_seen": 2874540032 + }, + { + "epoch": 0.87, + "learning_rate": 0.00010437713139418254, + "loss": 1.2526, + "theoretical_loss": 3.3601975613123853, + "tokens_seen": 2874605568 + }, + { + "epoch": 0.87, + "learning_rate": 0.00010436108324974924, + "loss": 1.4221, + "theoretical_loss": 3.360191670486966, + "tokens_seen": 2874671104 + }, + { + "epoch": 0.87, + "learning_rate": 0.00010434503510531594, + "loss": 1.2906, + "theoretical_loss": 3.3601857798334445, + "tokens_seen": 2874736640 + }, + { + "epoch": 0.87, + "learning_rate": 0.00010432898696088264, + "loss": 1.28, + "theoretical_loss": 3.360179889351812, + "tokens_seen": 2874802176 + }, + { + "epoch": 0.87, + "learning_rate": 0.00010431293881644934, + "loss": 1.2285, + "theoretical_loss": 3.3601739990420603, + "tokens_seen": 2874867712 + }, + { + "epoch": 0.87, + "learning_rate": 0.00010429689067201605, + "loss": 1.3533, + "theoretical_loss": 3.36016810890418, + "tokens_seen": 2874933248 + }, + { + "epoch": 0.87, + "learning_rate": 0.00010428084252758275, + "loss": 1.2493, + "theoretical_loss": 3.360162218938162, + "tokens_seen": 2874998784 + }, + { + "epoch": 0.87, + "learning_rate": 0.00010426479438314945, + "loss": 1.3805, + "theoretical_loss": 3.360156329143998, + "tokens_seen": 2875064320 + }, + { + "epoch": 0.87, + "learning_rate": 0.00010424874623871615, + "loss": 1.4076, + "theoretical_loss": 3.3601504395216786, + "tokens_seen": 2875129856 + }, + { + "epoch": 0.87, + "learning_rate": 0.00010423269809428285, + "loss": 1.2346, + "theoretical_loss": 3.360144550071195, + "tokens_seen": 2875195392 + }, + { + "epoch": 0.87, + "learning_rate": 0.00010421664994984955, + "loss": 1.3154, + "theoretical_loss": 3.360138660792538, + "tokens_seen": 2875260928 + }, + { + "epoch": 0.87, + "learning_rate": 0.00010420060180541626, + "loss": 1.2273, + "theoretical_loss": 3.3601327716856995, + "tokens_seen": 2875326464 + }, + { + "epoch": 0.87, + "learning_rate": 0.00010418455366098296, + "loss": 1.4188, + "theoretical_loss": 3.360126882750669, + "tokens_seen": 2875392000 + }, + { + "epoch": 0.87, + "learning_rate": 0.00010416850551654966, + "loss": 1.3084, + "theoretical_loss": 3.3601209939874392, + "tokens_seen": 2875457536 + }, + { + "epoch": 0.87, + "learning_rate": 0.00010415245737211636, + "loss": 1.2352, + "theoretical_loss": 3.360115105396001, + "tokens_seen": 2875523072 + }, + { + "epoch": 0.87, + "learning_rate": 0.00010413640922768307, + "loss": 1.3065, + "theoretical_loss": 3.360109216976344, + "tokens_seen": 2875588608 + }, + { + "epoch": 0.87, + "learning_rate": 0.00010412036108324977, + "loss": 1.248, + "theoretical_loss": 3.3601033287284605, + "tokens_seen": 2875654144 + }, + { + "epoch": 0.87, + "learning_rate": 0.00010410431293881647, + "loss": 1.3018, + "theoretical_loss": 3.360097440652342, + "tokens_seen": 2875719680 + }, + { + "epoch": 0.87, + "learning_rate": 0.00010408826479438317, + "loss": 1.2592, + "theoretical_loss": 3.3600915527479778, + "tokens_seen": 2875785216 + }, + { + "epoch": 0.87, + "learning_rate": 0.00010407221664994985, + "loss": 1.3598, + "theoretical_loss": 3.3600856650153608, + "tokens_seen": 2875850752 + }, + { + "epoch": 0.87, + "learning_rate": 0.00010405616850551655, + "loss": 1.2321, + "theoretical_loss": 3.360079777454481, + "tokens_seen": 2875916288 + }, + { + "epoch": 0.87, + "learning_rate": 0.00010404012036108326, + "loss": 1.241, + "theoretical_loss": 3.3600738900653306, + "tokens_seen": 2875981824 + }, + { + "epoch": 0.87, + "learning_rate": 0.00010402407221664996, + "loss": 1.3545, + "theoretical_loss": 3.3600680028478993, + "tokens_seen": 2876047360 + }, + { + "epoch": 0.87, + "learning_rate": 0.00010400802407221666, + "loss": 1.275, + "theoretical_loss": 3.3600621158021786, + "tokens_seen": 2876112896 + }, + { + "epoch": 0.87, + "learning_rate": 0.00010399197592778336, + "loss": 1.2768, + "theoretical_loss": 3.36005622892816, + "tokens_seen": 2876178432 + }, + { + "epoch": 0.87, + "learning_rate": 0.00010397592778335006, + "loss": 1.2807, + "theoretical_loss": 3.3600503422258345, + "tokens_seen": 2876243968 + }, + { + "epoch": 0.87, + "learning_rate": 0.00010395987963891676, + "loss": 1.4232, + "theoretical_loss": 3.360044455695193, + "tokens_seen": 2876309504 + }, + { + "epoch": 0.87, + "learning_rate": 0.00010394383149448346, + "loss": 1.2043, + "theoretical_loss": 3.360038569336227, + "tokens_seen": 2876375040 + }, + { + "epoch": 0.87, + "learning_rate": 0.00010392778335005016, + "loss": 1.2924, + "theoretical_loss": 3.3600326831489267, + "tokens_seen": 2876440576 + }, + { + "epoch": 0.87, + "learning_rate": 0.00010391173520561686, + "loss": 1.289, + "theoretical_loss": 3.3600267971332833, + "tokens_seen": 2876506112 + }, + { + "epoch": 0.87, + "learning_rate": 0.00010389568706118356, + "loss": 1.2593, + "theoretical_loss": 3.360020911289289, + "tokens_seen": 2876571648 + }, + { + "epoch": 0.87, + "learning_rate": 0.00010387963891675025, + "loss": 1.1952, + "theoretical_loss": 3.360015025616934, + "tokens_seen": 2876637184 + }, + { + "epoch": 0.87, + "learning_rate": 0.00010386359077231695, + "loss": 1.305, + "theoretical_loss": 3.3600091401162095, + "tokens_seen": 2876702720 + }, + { + "epoch": 0.87, + "learning_rate": 0.00010384754262788365, + "loss": 1.2966, + "theoretical_loss": 3.360003254787107, + "tokens_seen": 2876768256 + }, + { + "epoch": 0.87, + "learning_rate": 0.00010383149448345035, + "loss": 1.3519, + "theoretical_loss": 3.3599973696296166, + "tokens_seen": 2876833792 + }, + { + "epoch": 0.87, + "learning_rate": 0.00010381544633901705, + "loss": 1.3583, + "theoretical_loss": 3.35999148464373, + "tokens_seen": 2876899328 + }, + { + "epoch": 0.87, + "learning_rate": 0.00010379939819458375, + "loss": 1.2504, + "theoretical_loss": 3.3599855998294386, + "tokens_seen": 2876964864 + }, + { + "epoch": 0.87, + "objective/train/docs_used": 1045577, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.447434902191162, + "objective/train/theoretical_loss": 3.359979715186733, + "objective/train/tokens_used": 2897490400, + "theoretical_loss": 3.359979715186733, + "tokens_seen": 2877030400 + }, + { + "epoch": 0.87, + "learning_rate": 0.00010378335005015045, + "loss": 1.3369, + "theoretical_loss": 3.359979715186733, + "tokens_seen": 2877030400 + }, + { + "epoch": 0.87, + "learning_rate": 0.00010376730190571715, + "loss": 1.4093, + "theoretical_loss": 3.3599738307156044, + "tokens_seen": 2877095936 + }, + { + "epoch": 0.87, + "learning_rate": 0.00010375125376128385, + "loss": 1.3144, + "theoretical_loss": 3.359967946416044, + "tokens_seen": 2877161472 + }, + { + "epoch": 0.87, + "learning_rate": 0.00010373520561685055, + "loss": 1.2077, + "theoretical_loss": 3.3599620622880426, + "tokens_seen": 2877227008 + }, + { + "epoch": 0.87, + "learning_rate": 0.00010371915747241725, + "loss": 1.2963, + "theoretical_loss": 3.3599561783315917, + "tokens_seen": 2877292544 + }, + { + "epoch": 0.87, + "learning_rate": 0.00010370310932798396, + "loss": 1.2685, + "theoretical_loss": 3.3599502945466826, + "tokens_seen": 2877358080 + }, + { + "epoch": 0.87, + "learning_rate": 0.00010368706118355064, + "loss": 1.3069, + "theoretical_loss": 3.3599444109333056, + "tokens_seen": 2877423616 + }, + { + "epoch": 0.87, + "learning_rate": 0.00010367101303911734, + "loss": 1.1896, + "theoretical_loss": 3.3599385274914524, + "tokens_seen": 2877489152 + }, + { + "epoch": 0.87, + "learning_rate": 0.00010365496489468404, + "loss": 1.2645, + "theoretical_loss": 3.359932644221114, + "tokens_seen": 2877554688 + }, + { + "epoch": 0.87, + "learning_rate": 0.00010363891675025074, + "loss": 1.2324, + "theoretical_loss": 3.359926761122281, + "tokens_seen": 2877620224 + }, + { + "epoch": 0.87, + "learning_rate": 0.00010362286860581744, + "loss": 1.2496, + "theoretical_loss": 3.359920878194945, + "tokens_seen": 2877685760 + }, + { + "epoch": 0.87, + "learning_rate": 0.00010360682046138415, + "loss": 1.3083, + "theoretical_loss": 3.359914995439097, + "tokens_seen": 2877751296 + }, + { + "epoch": 0.87, + "learning_rate": 0.00010359077231695087, + "loss": 1.2952, + "theoretical_loss": 3.3599091128547283, + "tokens_seen": 2877816832 + }, + { + "epoch": 0.87, + "learning_rate": 0.00010357472417251757, + "loss": 1.1962, + "theoretical_loss": 3.3599032304418297, + "tokens_seen": 2877882368 + }, + { + "epoch": 0.87, + "learning_rate": 0.00010355867602808426, + "loss": 1.3454, + "theoretical_loss": 3.359897348200392, + "tokens_seen": 2877947904 + }, + { + "epoch": 0.87, + "learning_rate": 0.00010354262788365096, + "loss": 1.3726, + "theoretical_loss": 3.3598914661304073, + "tokens_seen": 2878013440 + }, + { + "epoch": 0.87, + "learning_rate": 0.00010352657973921766, + "loss": 1.331, + "theoretical_loss": 3.359885584231866, + "tokens_seen": 2878078976 + }, + { + "epoch": 0.87, + "learning_rate": 0.00010351053159478436, + "loss": 1.444, + "theoretical_loss": 3.3598797025047586, + "tokens_seen": 2878144512 + }, + { + "epoch": 0.87, + "learning_rate": 0.00010349448345035106, + "loss": 1.2944, + "theoretical_loss": 3.3598738209490775, + "tokens_seen": 2878210048 + }, + { + "epoch": 0.87, + "learning_rate": 0.00010347843530591776, + "loss": 1.3111, + "theoretical_loss": 3.359867939564813, + "tokens_seen": 2878275584 + }, + { + "epoch": 0.87, + "learning_rate": 0.00010346238716148446, + "loss": 1.2731, + "theoretical_loss": 3.359862058351956, + "tokens_seen": 2878341120 + }, + { + "epoch": 0.87, + "learning_rate": 0.00010344633901705117, + "loss": 1.271, + "theoretical_loss": 3.3598561773104985, + "tokens_seen": 2878406656 + }, + { + "epoch": 0.87, + "learning_rate": 0.00010343029087261787, + "loss": 1.2406, + "theoretical_loss": 3.3598502964404307, + "tokens_seen": 2878472192 + }, + { + "epoch": 0.87, + "learning_rate": 0.00010341424272818457, + "loss": 1.1956, + "theoretical_loss": 3.3598444157417444, + "tokens_seen": 2878537728 + }, + { + "epoch": 0.87, + "learning_rate": 0.00010339819458375127, + "loss": 1.1771, + "theoretical_loss": 3.35983853521443, + "tokens_seen": 2878603264 + }, + { + "epoch": 0.87, + "learning_rate": 0.00010338214643931797, + "loss": 1.3452, + "theoretical_loss": 3.3598326548584794, + "tokens_seen": 2878668800 + }, + { + "epoch": 0.87, + "learning_rate": 0.00010336609829488466, + "loss": 1.3031, + "theoretical_loss": 3.359826774673883, + "tokens_seen": 2878734336 + }, + { + "epoch": 0.87, + "learning_rate": 0.00010335005015045136, + "loss": 1.3601, + "theoretical_loss": 3.3598208946606323, + "tokens_seen": 2878799872 + }, + { + "epoch": 0.87, + "learning_rate": 0.00010333400200601806, + "loss": 1.3741, + "theoretical_loss": 3.359815014818718, + "tokens_seen": 2878865408 + }, + { + "epoch": 0.87, + "learning_rate": 0.00010331795386158476, + "loss": 1.3678, + "theoretical_loss": 3.359809135148132, + "tokens_seen": 2878930944 + }, + { + "epoch": 0.87, + "learning_rate": 0.00010330190571715146, + "loss": 1.2564, + "theoretical_loss": 3.3598032556488646, + "tokens_seen": 2878996480 + }, + { + "epoch": 0.87, + "learning_rate": 0.00010328585757271816, + "loss": 1.32, + "theoretical_loss": 3.3597973763209072, + "tokens_seen": 2879062016 + }, + { + "epoch": 0.87, + "learning_rate": 0.00010326980942828486, + "loss": 1.1805, + "theoretical_loss": 3.3597914971642515, + "tokens_seen": 2879127552 + }, + { + "epoch": 0.87, + "learning_rate": 0.00010325376128385156, + "loss": 1.3596, + "theoretical_loss": 3.359785618178887, + "tokens_seen": 2879193088 + }, + { + "epoch": 0.87, + "learning_rate": 0.00010323771313941826, + "loss": 1.2919, + "theoretical_loss": 3.359779739364807, + "tokens_seen": 2879258624 + }, + { + "epoch": 0.87, + "learning_rate": 0.00010322166499498496, + "loss": 1.2147, + "theoretical_loss": 3.359773860722001, + "tokens_seen": 2879324160 + }, + { + "epoch": 0.87, + "learning_rate": 0.00010320561685055166, + "loss": 1.2107, + "theoretical_loss": 3.3597679822504602, + "tokens_seen": 2879389696 + }, + { + "epoch": 0.87, + "learning_rate": 0.00010318956870611836, + "loss": 1.3022, + "theoretical_loss": 3.3597621039501764, + "tokens_seen": 2879455232 + }, + { + "epoch": 0.87, + "learning_rate": 0.00010317352056168505, + "loss": 1.3316, + "theoretical_loss": 3.3597562258211404, + "tokens_seen": 2879520768 + }, + { + "epoch": 0.87, + "learning_rate": 0.00010315747241725175, + "loss": 1.3298, + "theoretical_loss": 3.3597503478633435, + "tokens_seen": 2879586304 + }, + { + "epoch": 0.87, + "learning_rate": 0.00010314142427281845, + "loss": 1.1938, + "theoretical_loss": 3.3597444700767762, + "tokens_seen": 2879651840 + }, + { + "epoch": 0.87, + "learning_rate": 0.00010312537612838515, + "loss": 1.2176, + "theoretical_loss": 3.3597385924614303, + "tokens_seen": 2879717376 + }, + { + "epoch": 0.87, + "learning_rate": 0.00010310932798395185, + "loss": 1.2639, + "theoretical_loss": 3.3597327150172966, + "tokens_seen": 2879782912 + }, + { + "epoch": 0.87, + "learning_rate": 0.00010309327983951855, + "loss": 1.3225, + "theoretical_loss": 3.3597268377443665, + "tokens_seen": 2879848448 + }, + { + "epoch": 0.87, + "learning_rate": 0.00010307723169508525, + "loss": 1.3763, + "theoretical_loss": 3.3597209606426306, + "tokens_seen": 2879913984 + }, + { + "epoch": 0.87, + "learning_rate": 0.00010306118355065195, + "loss": 1.2792, + "theoretical_loss": 3.3597150837120804, + "tokens_seen": 2879979520 + }, + { + "epoch": 0.87, + "learning_rate": 0.00010304513540621865, + "loss": 1.3297, + "theoretical_loss": 3.359709206952707, + "tokens_seen": 2880045056 + }, + { + "epoch": 0.87, + "learning_rate": 0.00010302908726178535, + "loss": 1.2721, + "theoretical_loss": 3.3597033303645016, + "tokens_seen": 2880110592 + }, + { + "epoch": 0.87, + "learning_rate": 0.00010301303911735206, + "loss": 1.3001, + "theoretical_loss": 3.3596974539474553, + "tokens_seen": 2880176128 + }, + { + "epoch": 0.87, + "learning_rate": 0.00010299699097291877, + "loss": 1.3161, + "theoretical_loss": 3.3596915777015584, + "tokens_seen": 2880241664 + }, + { + "epoch": 0.87, + "objective/train/docs_used": 1046730, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.2681565284729004, + "objective/train/theoretical_loss": 3.3596857016268036, + "objective/train/tokens_used": 2900767200, + "theoretical_loss": 3.3596857016268036, + "tokens_seen": 2880307200 + }, + { + "epoch": 0.87, + "learning_rate": 0.00010298094282848547, + "loss": 1.3998, + "theoretical_loss": 3.3596857016268036, + "tokens_seen": 2880307200 + }, + { + "epoch": 0.87, + "learning_rate": 0.00010296489468405217, + "loss": 1.273, + "theoretical_loss": 3.3596798257231804, + "tokens_seen": 2880372736 + }, + { + "epoch": 0.87, + "learning_rate": 0.00010294884653961887, + "loss": 1.3143, + "theoretical_loss": 3.3596739499906807, + "tokens_seen": 2880438272 + }, + { + "epoch": 0.87, + "learning_rate": 0.00010293279839518557, + "loss": 1.3738, + "theoretical_loss": 3.3596680744292957, + "tokens_seen": 2880503808 + }, + { + "epoch": 0.87, + "learning_rate": 0.00010291675025075227, + "loss": 1.2456, + "theoretical_loss": 3.3596621990390165, + "tokens_seen": 2880569344 + }, + { + "epoch": 0.87, + "learning_rate": 0.00010290070210631897, + "loss": 1.1536, + "theoretical_loss": 3.3596563238198343, + "tokens_seen": 2880634880 + }, + { + "epoch": 0.87, + "learning_rate": 0.00010288465396188567, + "loss": 1.2221, + "theoretical_loss": 3.35965044877174, + "tokens_seen": 2880700416 + }, + { + "epoch": 0.87, + "learning_rate": 0.00010286860581745237, + "loss": 1.2871, + "theoretical_loss": 3.359644573894725, + "tokens_seen": 2880765952 + }, + { + "epoch": 0.87, + "learning_rate": 0.00010285255767301906, + "loss": 1.2677, + "theoretical_loss": 3.3596386991887797, + "tokens_seen": 2880831488 + }, + { + "epoch": 0.87, + "learning_rate": 0.00010283650952858576, + "loss": 1.3193, + "theoretical_loss": 3.359632824653896, + "tokens_seen": 2880897024 + }, + { + "epoch": 0.87, + "learning_rate": 0.00010282046138415246, + "loss": 1.2198, + "theoretical_loss": 3.3596269502900644, + "tokens_seen": 2880962560 + }, + { + "epoch": 0.87, + "learning_rate": 0.00010280441323971916, + "loss": 1.2842, + "theoretical_loss": 3.3596210760972767, + "tokens_seen": 2881028096 + }, + { + "epoch": 0.87, + "learning_rate": 0.00010278836509528586, + "loss": 1.2675, + "theoretical_loss": 3.3596152020755237, + "tokens_seen": 2881093632 + }, + { + "epoch": 0.87, + "learning_rate": 0.00010277231695085257, + "loss": 1.2295, + "theoretical_loss": 3.3596093282247965, + "tokens_seen": 2881159168 + }, + { + "epoch": 0.87, + "learning_rate": 0.00010275626880641927, + "loss": 1.2619, + "theoretical_loss": 3.3596034545450864, + "tokens_seen": 2881224704 + }, + { + "epoch": 0.87, + "learning_rate": 0.00010274022066198597, + "loss": 1.2667, + "theoretical_loss": 3.3595975810363843, + "tokens_seen": 2881290240 + }, + { + "epoch": 0.87, + "learning_rate": 0.00010272417251755267, + "loss": 1.253, + "theoretical_loss": 3.3595917076986814, + "tokens_seen": 2881355776 + }, + { + "epoch": 0.87, + "learning_rate": 0.00010270812437311937, + "loss": 1.1921, + "theoretical_loss": 3.359585834531969, + "tokens_seen": 2881421312 + }, + { + "epoch": 0.87, + "learning_rate": 0.00010269207622868607, + "loss": 1.3543, + "theoretical_loss": 3.359579961536238, + "tokens_seen": 2881486848 + }, + { + "epoch": 0.87, + "learning_rate": 0.00010267602808425277, + "loss": 1.2843, + "theoretical_loss": 3.3595740887114798, + "tokens_seen": 2881552384 + }, + { + "epoch": 0.87, + "learning_rate": 0.00010265997993981946, + "loss": 1.2895, + "theoretical_loss": 3.359568216057685, + "tokens_seen": 2881617920 + }, + { + "epoch": 0.87, + "learning_rate": 0.00010264393179538616, + "loss": 1.3377, + "theoretical_loss": 3.3595623435748454, + "tokens_seen": 2881683456 + }, + { + "epoch": 0.87, + "learning_rate": 0.00010262788365095286, + "loss": 1.2272, + "theoretical_loss": 3.3595564712629518, + "tokens_seen": 2881748992 + }, + { + "epoch": 0.87, + "learning_rate": 0.00010261183550651956, + "loss": 1.2898, + "theoretical_loss": 3.3595505991219956, + "tokens_seen": 2881814528 + }, + { + "epoch": 0.87, + "learning_rate": 0.00010259578736208626, + "loss": 1.4724, + "theoretical_loss": 3.3595447271519676, + "tokens_seen": 2881880064 + }, + { + "epoch": 0.87, + "learning_rate": 0.00010257973921765296, + "loss": 1.3213, + "theoretical_loss": 3.359538855352859, + "tokens_seen": 2881945600 + }, + { + "epoch": 0.87, + "learning_rate": 0.00010256369107321966, + "loss": 1.1873, + "theoretical_loss": 3.3595329837246606, + "tokens_seen": 2882011136 + }, + { + "epoch": 0.87, + "learning_rate": 0.00010254764292878636, + "loss": 1.3503, + "theoretical_loss": 3.3595271122673642, + "tokens_seen": 2882076672 + }, + { + "epoch": 0.87, + "learning_rate": 0.00010253159478435306, + "loss": 1.341, + "theoretical_loss": 3.359521240980961, + "tokens_seen": 2882142208 + }, + { + "epoch": 0.87, + "learning_rate": 0.00010251554663991976, + "loss": 1.3276, + "theoretical_loss": 3.359515369865442, + "tokens_seen": 2882207744 + }, + { + "epoch": 0.87, + "learning_rate": 0.00010249949849548646, + "loss": 1.3209, + "theoretical_loss": 3.359509498920797, + "tokens_seen": 2882273280 + }, + { + "epoch": 0.87, + "learning_rate": 0.00010248345035105316, + "loss": 1.2028, + "theoretical_loss": 3.3595036281470194, + "tokens_seen": 2882338816 + }, + { + "epoch": 0.87, + "learning_rate": 0.00010246740220661985, + "loss": 1.1909, + "theoretical_loss": 3.3594977575440987, + "tokens_seen": 2882404352 + }, + { + "epoch": 0.87, + "learning_rate": 0.00010245135406218655, + "loss": 1.2434, + "theoretical_loss": 3.359491887112027, + "tokens_seen": 2882469888 + }, + { + "epoch": 0.87, + "learning_rate": 0.00010243530591775325, + "loss": 1.4019, + "theoretical_loss": 3.359486016850795, + "tokens_seen": 2882535424 + }, + { + "epoch": 0.87, + "learning_rate": 0.00010241925777331995, + "loss": 1.2587, + "theoretical_loss": 3.3594801467603936, + "tokens_seen": 2882600960 + }, + { + "epoch": 0.87, + "learning_rate": 0.00010240320962888665, + "loss": 1.302, + "theoretical_loss": 3.3594742768408143, + "tokens_seen": 2882666496 + }, + { + "epoch": 0.87, + "learning_rate": 0.00010238716148445338, + "loss": 1.2842, + "theoretical_loss": 3.359468407092048, + "tokens_seen": 2882732032 + }, + { + "epoch": 0.87, + "learning_rate": 0.00010237111334002008, + "loss": 1.237, + "theoretical_loss": 3.3594625375140863, + "tokens_seen": 2882797568 + }, + { + "epoch": 0.87, + "learning_rate": 0.00010235506519558677, + "loss": 1.2313, + "theoretical_loss": 3.35945666810692, + "tokens_seen": 2882863104 + }, + { + "epoch": 0.87, + "learning_rate": 0.00010233901705115347, + "loss": 1.2643, + "theoretical_loss": 3.35945079887054, + "tokens_seen": 2882928640 + }, + { + "epoch": 0.87, + "learning_rate": 0.00010232296890672017, + "loss": 1.3659, + "theoretical_loss": 3.359444929804938, + "tokens_seen": 2882994176 + }, + { + "epoch": 0.87, + "learning_rate": 0.00010230692076228687, + "loss": 1.3003, + "theoretical_loss": 3.359439060910105, + "tokens_seen": 2883059712 + }, + { + "epoch": 0.87, + "learning_rate": 0.00010229087261785357, + "loss": 1.2503, + "theoretical_loss": 3.359433192186032, + "tokens_seen": 2883125248 + }, + { + "epoch": 0.87, + "learning_rate": 0.00010227482447342027, + "loss": 1.2576, + "theoretical_loss": 3.35942732363271, + "tokens_seen": 2883190784 + }, + { + "epoch": 0.87, + "learning_rate": 0.00010225877632898697, + "loss": 1.1682, + "theoretical_loss": 3.3594214552501307, + "tokens_seen": 2883256320 + }, + { + "epoch": 0.87, + "learning_rate": 0.00010224272818455367, + "loss": 1.3277, + "theoretical_loss": 3.3594155870382845, + "tokens_seen": 2883321856 + }, + { + "epoch": 0.87, + "learning_rate": 0.00010222668004012037, + "loss": 1.2746, + "theoretical_loss": 3.359409718997163, + "tokens_seen": 2883387392 + }, + { + "epoch": 0.87, + "learning_rate": 0.00010221063189568707, + "loss": 1.2468, + "theoretical_loss": 3.359403851126758, + "tokens_seen": 2883452928 + }, + { + "epoch": 0.87, + "learning_rate": 0.00010219458375125377, + "loss": 1.3248, + "theoretical_loss": 3.359397983427059, + "tokens_seen": 2883518464 + }, + { + "epoch": 0.87, + "objective/train/docs_used": 1048090, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.2763785123825073, + "objective/train/theoretical_loss": 3.3593921158980584, + "objective/train/tokens_used": 2904044000, + "theoretical_loss": 3.3593921158980584, + "tokens_seen": 2883584000 + }, + { + "epoch": 0.87, + "learning_rate": 0.00010217853560682047, + "loss": 1.2691, + "theoretical_loss": 3.3593921158980584, + "tokens_seen": 2883584000 + }, + { + "epoch": 0.87, + "learning_rate": 0.00010216248746238716, + "loss": 1.2344, + "theoretical_loss": 3.359386248539747, + "tokens_seen": 2883649536 + }, + { + "epoch": 0.87, + "learning_rate": 0.00010214643931795386, + "loss": 1.2656, + "theoretical_loss": 3.3593803813521164, + "tokens_seen": 2883715072 + }, + { + "epoch": 0.87, + "learning_rate": 0.00010213039117352056, + "loss": 1.3291, + "theoretical_loss": 3.3593745143351574, + "tokens_seen": 2883780608 + }, + { + "epoch": 0.87, + "learning_rate": 0.00010211434302908726, + "loss": 1.203, + "theoretical_loss": 3.3593686474888607, + "tokens_seen": 2883846144 + }, + { + "epoch": 0.87, + "learning_rate": 0.00010209829488465396, + "loss": 1.3914, + "theoretical_loss": 3.359362780813218, + "tokens_seen": 2883911680 + }, + { + "epoch": 0.87, + "learning_rate": 0.00010208224674022067, + "loss": 1.1815, + "theoretical_loss": 3.3593569143082207, + "tokens_seen": 2883977216 + }, + { + "epoch": 0.87, + "learning_rate": 0.00010206619859578737, + "loss": 1.2277, + "theoretical_loss": 3.3593510479738597, + "tokens_seen": 2884042752 + }, + { + "epoch": 0.87, + "learning_rate": 0.00010205015045135407, + "loss": 1.2523, + "theoretical_loss": 3.3593451818101254, + "tokens_seen": 2884108288 + }, + { + "epoch": 0.87, + "learning_rate": 0.00010203410230692077, + "loss": 1.2378, + "theoretical_loss": 3.35933931581701, + "tokens_seen": 2884173824 + }, + { + "epoch": 0.87, + "learning_rate": 0.00010201805416248747, + "loss": 1.3188, + "theoretical_loss": 3.3593334499945042, + "tokens_seen": 2884239360 + }, + { + "epoch": 0.87, + "learning_rate": 0.00010200200601805417, + "loss": 1.3632, + "theoretical_loss": 3.3593275843425996, + "tokens_seen": 2884304896 + }, + { + "epoch": 0.87, + "learning_rate": 0.00010198595787362087, + "loss": 1.237, + "theoretical_loss": 3.3593217188612865, + "tokens_seen": 2884370432 + }, + { + "epoch": 0.87, + "learning_rate": 0.00010196990972918756, + "loss": 1.2703, + "theoretical_loss": 3.359315853550557, + "tokens_seen": 2884435968 + }, + { + "epoch": 0.87, + "learning_rate": 0.00010195386158475426, + "loss": 1.2627, + "theoretical_loss": 3.3593099884104016, + "tokens_seen": 2884501504 + }, + { + "epoch": 0.87, + "learning_rate": 0.00010193781344032096, + "loss": 1.2743, + "theoretical_loss": 3.359304123440812, + "tokens_seen": 2884567040 + }, + { + "epoch": 0.87, + "learning_rate": 0.00010192176529588766, + "loss": 1.274, + "theoretical_loss": 3.359298258641779, + "tokens_seen": 2884632576 + }, + { + "epoch": 0.87, + "learning_rate": 0.00010190571715145436, + "loss": 1.2062, + "theoretical_loss": 3.3592923940132935, + "tokens_seen": 2884698112 + }, + { + "epoch": 0.87, + "learning_rate": 0.00010188966900702106, + "loss": 1.3058, + "theoretical_loss": 3.359286529555347, + "tokens_seen": 2884763648 + }, + { + "epoch": 0.87, + "learning_rate": 0.00010187362086258776, + "loss": 1.2353, + "theoretical_loss": 3.359280665267931, + "tokens_seen": 2884829184 + }, + { + "epoch": 0.87, + "learning_rate": 0.00010185757271815446, + "loss": 1.2785, + "theoretical_loss": 3.3592748011510363, + "tokens_seen": 2884894720 + }, + { + "epoch": 0.87, + "learning_rate": 0.00010184152457372116, + "loss": 1.3301, + "theoretical_loss": 3.359268937204654, + "tokens_seen": 2884960256 + }, + { + "epoch": 0.87, + "learning_rate": 0.00010182547642928786, + "loss": 1.2638, + "theoretical_loss": 3.3592630734287754, + "tokens_seen": 2885025792 + }, + { + "epoch": 0.87, + "learning_rate": 0.00010180942828485456, + "loss": 1.248, + "theoretical_loss": 3.359257209823392, + "tokens_seen": 2885091328 + }, + { + "epoch": 0.87, + "learning_rate": 0.00010179338014042128, + "loss": 1.2835, + "theoretical_loss": 3.359251346388494, + "tokens_seen": 2885156864 + }, + { + "epoch": 0.87, + "learning_rate": 0.00010177733199598798, + "loss": 1.2695, + "theoretical_loss": 3.3592454831240732, + "tokens_seen": 2885222400 + }, + { + "epoch": 0.87, + "learning_rate": 0.00010176128385155468, + "loss": 1.2984, + "theoretical_loss": 3.359239620030121, + "tokens_seen": 2885287936 + }, + { + "epoch": 0.87, + "learning_rate": 0.00010174523570712138, + "loss": 1.427, + "theoretical_loss": 3.3592337571066286, + "tokens_seen": 2885353472 + }, + { + "epoch": 0.87, + "learning_rate": 0.00010172918756268808, + "loss": 1.2114, + "theoretical_loss": 3.3592278943535865, + "tokens_seen": 2885419008 + }, + { + "epoch": 0.87, + "learning_rate": 0.00010171313941825478, + "loss": 1.2981, + "theoretical_loss": 3.3592220317709867, + "tokens_seen": 2885484544 + }, + { + "epoch": 0.87, + "learning_rate": 0.00010169709127382148, + "loss": 1.2649, + "theoretical_loss": 3.3592161693588194, + "tokens_seen": 2885550080 + }, + { + "epoch": 0.87, + "learning_rate": 0.00010168104312938818, + "loss": 1.2476, + "theoretical_loss": 3.3592103071170767, + "tokens_seen": 2885615616 + }, + { + "epoch": 0.87, + "learning_rate": 0.00010166499498495488, + "loss": 1.2463, + "theoretical_loss": 3.359204445045749, + "tokens_seen": 2885681152 + }, + { + "epoch": 0.87, + "learning_rate": 0.00010164894684052157, + "loss": 1.3109, + "theoretical_loss": 3.359198583144828, + "tokens_seen": 2885746688 + }, + { + "epoch": 0.87, + "learning_rate": 0.00010163289869608827, + "loss": 1.3236, + "theoretical_loss": 3.3591927214143054, + "tokens_seen": 2885812224 + }, + { + "epoch": 0.87, + "learning_rate": 0.00010161685055165497, + "loss": 1.4288, + "theoretical_loss": 3.359186859854171, + "tokens_seen": 2885877760 + }, + { + "epoch": 0.87, + "learning_rate": 0.00010160080240722167, + "loss": 1.3037, + "theoretical_loss": 3.3591809984644168, + "tokens_seen": 2885943296 + }, + { + "epoch": 0.87, + "learning_rate": 0.00010158475426278837, + "loss": 1.2518, + "theoretical_loss": 3.359175137245034, + "tokens_seen": 2886008832 + }, + { + "epoch": 0.87, + "learning_rate": 0.00010156870611835507, + "loss": 1.3199, + "theoretical_loss": 3.3591692761960137, + "tokens_seen": 2886074368 + }, + { + "epoch": 0.87, + "learning_rate": 0.00010155265797392177, + "loss": 1.223, + "theoretical_loss": 3.359163415317347, + "tokens_seen": 2886139904 + }, + { + "epoch": 0.87, + "learning_rate": 0.00010153660982948847, + "loss": 1.3033, + "theoretical_loss": 3.359157554609025, + "tokens_seen": 2886205440 + }, + { + "epoch": 0.87, + "learning_rate": 0.00010152056168505517, + "loss": 1.3491, + "theoretical_loss": 3.3591516940710395, + "tokens_seen": 2886270976 + }, + { + "epoch": 0.87, + "learning_rate": 0.00010150451354062187, + "loss": 1.3845, + "theoretical_loss": 3.359145833703381, + "tokens_seen": 2886336512 + }, + { + "epoch": 0.87, + "learning_rate": 0.00010148846539618858, + "loss": 1.0724, + "theoretical_loss": 3.35913997350604, + "tokens_seen": 2886402048 + }, + { + "epoch": 0.87, + "learning_rate": 0.00010147241725175528, + "loss": 1.3541, + "theoretical_loss": 3.3591341134790094, + "tokens_seen": 2886467584 + }, + { + "epoch": 0.87, + "learning_rate": 0.00010145636910732196, + "loss": 1.2512, + "theoretical_loss": 3.3591282536222797, + "tokens_seen": 2886533120 + }, + { + "epoch": 0.87, + "learning_rate": 0.00010144032096288866, + "loss": 1.2811, + "theoretical_loss": 3.3591223939358414, + "tokens_seen": 2886598656 + }, + { + "epoch": 0.87, + "learning_rate": 0.00010142427281845536, + "loss": 1.2295, + "theoretical_loss": 3.3591165344196865, + "tokens_seen": 2886664192 + }, + { + "epoch": 0.87, + "learning_rate": 0.00010140822467402206, + "loss": 1.3127, + "theoretical_loss": 3.3591106750738056, + "tokens_seen": 2886729728 + }, + { + "epoch": 0.87, + "learning_rate": 0.00010139217652958877, + "loss": 1.2583, + "theoretical_loss": 3.3591048158981907, + "tokens_seen": 2886795264 + }, + { + "epoch": 0.87, + "objective/train/docs_used": 1049289, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.1765341758728027, + "objective/train/theoretical_loss": 3.359098956892832, + "objective/train/tokens_used": 2907320800, + "theoretical_loss": 3.359098956892832, + "tokens_seen": 2886860800 + }, + { + "epoch": 0.87, + "learning_rate": 0.00010137612838515547, + "loss": 1.181, + "theoretical_loss": 3.359098956892832, + "tokens_seen": 2886860800 + }, + { + "epoch": 0.87, + "learning_rate": 0.00010136008024072217, + "loss": 1.2814, + "theoretical_loss": 3.359093098057721, + "tokens_seen": 2886926336 + }, + { + "epoch": 0.87, + "learning_rate": 0.00010134403209628887, + "loss": 1.2342, + "theoretical_loss": 3.3590872393928493, + "tokens_seen": 2886991872 + }, + { + "epoch": 0.87, + "learning_rate": 0.00010132798395185557, + "loss": 1.263, + "theoretical_loss": 3.359081380898208, + "tokens_seen": 2887057408 + }, + { + "epoch": 0.87, + "learning_rate": 0.00010131193580742227, + "loss": 1.3435, + "theoretical_loss": 3.359075522573788, + "tokens_seen": 2887122944 + }, + { + "epoch": 0.87, + "learning_rate": 0.00010129588766298897, + "loss": 1.2129, + "theoretical_loss": 3.3590696644195805, + "tokens_seen": 2887188480 + }, + { + "epoch": 0.87, + "learning_rate": 0.00010127983951855567, + "loss": 1.2322, + "theoretical_loss": 3.359063806435577, + "tokens_seen": 2887254016 + }, + { + "epoch": 0.87, + "learning_rate": 0.00010126379137412236, + "loss": 1.2416, + "theoretical_loss": 3.3590579486217687, + "tokens_seen": 2887319552 + }, + { + "epoch": 0.87, + "learning_rate": 0.00010124774322968906, + "loss": 1.3387, + "theoretical_loss": 3.359052090978146, + "tokens_seen": 2887385088 + }, + { + "epoch": 0.87, + "learning_rate": 0.00010123169508525576, + "loss": 1.2454, + "theoretical_loss": 3.359046233504701, + "tokens_seen": 2887450624 + }, + { + "epoch": 0.88, + "learning_rate": 0.00010121564694082246, + "loss": 1.3222, + "theoretical_loss": 3.3590403762014245, + "tokens_seen": 2887516160 + }, + { + "epoch": 0.88, + "learning_rate": 0.00010119959879638919, + "loss": 1.3488, + "theoretical_loss": 3.3590345190683077, + "tokens_seen": 2887581696 + }, + { + "epoch": 0.88, + "learning_rate": 0.00010118355065195589, + "loss": 1.2169, + "theoretical_loss": 3.359028662105342, + "tokens_seen": 2887647232 + }, + { + "epoch": 0.88, + "learning_rate": 0.00010116750250752259, + "loss": 1.3245, + "theoretical_loss": 3.3590228053125184, + "tokens_seen": 2887712768 + }, + { + "epoch": 0.88, + "learning_rate": 0.00010115145436308929, + "loss": 1.2776, + "theoretical_loss": 3.3590169486898285, + "tokens_seen": 2887778304 + }, + { + "epoch": 0.88, + "learning_rate": 0.00010113540621865598, + "loss": 1.5158, + "theoretical_loss": 3.3590110922372625, + "tokens_seen": 2887843840 + }, + { + "epoch": 0.88, + "learning_rate": 0.00010111935807422268, + "loss": 1.3324, + "theoretical_loss": 3.359005235954813, + "tokens_seen": 2887909376 + }, + { + "epoch": 0.88, + "learning_rate": 0.00010110330992978938, + "loss": 1.242, + "theoretical_loss": 3.3589993798424693, + "tokens_seen": 2887974912 + }, + { + "epoch": 0.88, + "learning_rate": 0.00010108726178535608, + "loss": 1.2891, + "theoretical_loss": 3.3589935239002244, + "tokens_seen": 2888040448 + }, + { + "epoch": 0.88, + "learning_rate": 0.00010107121364092278, + "loss": 1.1788, + "theoretical_loss": 3.3589876681280693, + "tokens_seen": 2888105984 + }, + { + "epoch": 0.88, + "learning_rate": 0.00010105516549648948, + "loss": 1.263, + "theoretical_loss": 3.358981812525994, + "tokens_seen": 2888171520 + }, + { + "epoch": 0.88, + "learning_rate": 0.00010103911735205618, + "loss": 1.3644, + "theoretical_loss": 3.358975957093991, + "tokens_seen": 2888237056 + }, + { + "epoch": 0.88, + "learning_rate": 0.00010102306920762288, + "loss": 1.2102, + "theoretical_loss": 3.3589701018320506, + "tokens_seen": 2888302592 + }, + { + "epoch": 0.88, + "learning_rate": 0.00010100702106318958, + "loss": 1.2682, + "theoretical_loss": 3.3589642467401646, + "tokens_seen": 2888368128 + }, + { + "epoch": 0.88, + "learning_rate": 0.00010099097291875628, + "loss": 1.2772, + "theoretical_loss": 3.358958391818324, + "tokens_seen": 2888433664 + }, + { + "epoch": 0.88, + "learning_rate": 0.00010097492477432298, + "loss": 1.2952, + "theoretical_loss": 3.3589525370665196, + "tokens_seen": 2888499200 + }, + { + "epoch": 0.88, + "learning_rate": 0.00010095887662988968, + "loss": 1.2451, + "theoretical_loss": 3.358946682484743, + "tokens_seen": 2888564736 + }, + { + "epoch": 0.88, + "learning_rate": 0.00010094282848545637, + "loss": 1.2315, + "theoretical_loss": 3.3589408280729858, + "tokens_seen": 2888630272 + }, + { + "epoch": 0.88, + "learning_rate": 0.00010092678034102307, + "loss": 1.3059, + "theoretical_loss": 3.3589349738312384, + "tokens_seen": 2888695808 + }, + { + "epoch": 0.88, + "learning_rate": 0.00010091073219658977, + "loss": 1.2487, + "theoretical_loss": 3.3589291197594924, + "tokens_seen": 2888761344 + }, + { + "epoch": 0.88, + "learning_rate": 0.00010089468405215647, + "loss": 1.3718, + "theoretical_loss": 3.3589232658577393, + "tokens_seen": 2888826880 + }, + { + "epoch": 0.88, + "learning_rate": 0.00010087863590772317, + "loss": 1.2812, + "theoretical_loss": 3.3589174121259693, + "tokens_seen": 2888892416 + }, + { + "epoch": 0.88, + "learning_rate": 0.00010086258776328987, + "loss": 1.3134, + "theoretical_loss": 3.358911558564175, + "tokens_seen": 2888957952 + }, + { + "epoch": 0.88, + "learning_rate": 0.00010084653961885657, + "loss": 1.1515, + "theoretical_loss": 3.3589057051723463, + "tokens_seen": 2889023488 + }, + { + "epoch": 0.88, + "learning_rate": 0.00010083049147442327, + "loss": 1.2574, + "theoretical_loss": 3.3588998519504756, + "tokens_seen": 2889089024 + }, + { + "epoch": 0.88, + "learning_rate": 0.00010081444332998997, + "loss": 1.3626, + "theoretical_loss": 3.3588939988985533, + "tokens_seen": 2889154560 + }, + { + "epoch": 0.88, + "learning_rate": 0.00010079839518555668, + "loss": 1.2962, + "theoretical_loss": 3.3588881460165707, + "tokens_seen": 2889220096 + }, + { + "epoch": 0.88, + "learning_rate": 0.00010078234704112338, + "loss": 1.2997, + "theoretical_loss": 3.3588822933045193, + "tokens_seen": 2889285632 + }, + { + "epoch": 0.88, + "learning_rate": 0.00010076629889669008, + "loss": 1.1974, + "theoretical_loss": 3.35887644076239, + "tokens_seen": 2889351168 + }, + { + "epoch": 0.88, + "learning_rate": 0.00010075025075225676, + "loss": 1.2177, + "theoretical_loss": 3.3588705883901744, + "tokens_seen": 2889416704 + }, + { + "epoch": 0.88, + "learning_rate": 0.00010073420260782346, + "loss": 1.2164, + "theoretical_loss": 3.358864736187863, + "tokens_seen": 2889482240 + }, + { + "epoch": 0.88, + "learning_rate": 0.00010071815446339016, + "loss": 1.357, + "theoretical_loss": 3.358858884155448, + "tokens_seen": 2889547776 + }, + { + "epoch": 0.88, + "learning_rate": 0.00010070210631895687, + "loss": 1.3335, + "theoretical_loss": 3.35885303229292, + "tokens_seen": 2889613312 + }, + { + "epoch": 0.88, + "learning_rate": 0.00010068605817452357, + "loss": 1.1862, + "theoretical_loss": 3.3588471806002707, + "tokens_seen": 2889678848 + }, + { + "epoch": 0.88, + "learning_rate": 0.00010067001003009027, + "loss": 1.358, + "theoretical_loss": 3.3588413290774906, + "tokens_seen": 2889744384 + }, + { + "epoch": 0.88, + "learning_rate": 0.00010065396188565697, + "loss": 1.2523, + "theoretical_loss": 3.358835477724571, + "tokens_seen": 2889809920 + }, + { + "epoch": 0.88, + "learning_rate": 0.00010063791374122367, + "loss": 1.3258, + "theoretical_loss": 3.3588296265415036, + "tokens_seen": 2889875456 + }, + { + "epoch": 0.88, + "learning_rate": 0.00010062186559679037, + "loss": 1.2581, + "theoretical_loss": 3.3588237755282795, + "tokens_seen": 2889940992 + }, + { + "epoch": 0.88, + "learning_rate": 0.00010060581745235707, + "loss": 1.3088, + "theoretical_loss": 3.3588179246848897, + "tokens_seen": 2890006528 + }, + { + "epoch": 0.88, + "learning_rate": 0.00010058976930792378, + "loss": 1.2632, + "theoretical_loss": 3.3588120740113254, + "tokens_seen": 2890072064 + }, + { + "epoch": 0.88, + "objective/train/docs_used": 1050561, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.297946810722351, + "objective/train/theoretical_loss": 3.358806223507578, + "objective/train/tokens_used": 2910597600, + "theoretical_loss": 3.358806223507578, + "tokens_seen": 2890137600 + }, + { + "epoch": 0.88, + "learning_rate": 0.00010057372116349048, + "loss": 1.2809, + "theoretical_loss": 3.358806223507578, + "tokens_seen": 2890137600 + }, + { + "epoch": 0.88, + "learning_rate": 0.00010055767301905718, + "loss": 1.2541, + "theoretical_loss": 3.358800373173639, + "tokens_seen": 2890203136 + }, + { + "epoch": 0.88, + "learning_rate": 0.00010054162487462389, + "loss": 1.1936, + "theoretical_loss": 3.358794523009499, + "tokens_seen": 2890268672 + }, + { + "epoch": 0.88, + "learning_rate": 0.00010052557673019059, + "loss": 1.2404, + "theoretical_loss": 3.35878867301515, + "tokens_seen": 2890334208 + }, + { + "epoch": 0.88, + "learning_rate": 0.00010050952858575729, + "loss": 1.3158, + "theoretical_loss": 3.358782823190582, + "tokens_seen": 2890399744 + }, + { + "epoch": 0.88, + "learning_rate": 0.00010049348044132399, + "loss": 1.3215, + "theoretical_loss": 3.3587769735357873, + "tokens_seen": 2890465280 + }, + { + "epoch": 0.88, + "learning_rate": 0.00010047743229689069, + "loss": 1.2738, + "theoretical_loss": 3.3587711240507567, + "tokens_seen": 2890530816 + }, + { + "epoch": 0.88, + "learning_rate": 0.00010046138415245739, + "loss": 1.3179, + "theoretical_loss": 3.3587652747354815, + "tokens_seen": 2890596352 + }, + { + "epoch": 0.88, + "learning_rate": 0.00010044533600802408, + "loss": 1.2858, + "theoretical_loss": 3.358759425589953, + "tokens_seen": 2890661888 + }, + { + "epoch": 0.88, + "learning_rate": 0.00010042928786359078, + "loss": 1.2976, + "theoretical_loss": 3.3587535766141627, + "tokens_seen": 2890727424 + }, + { + "epoch": 0.88, + "learning_rate": 0.00010041323971915748, + "loss": 1.2993, + "theoretical_loss": 3.358747727808101, + "tokens_seen": 2890792960 + }, + { + "epoch": 0.88, + "learning_rate": 0.00010039719157472418, + "loss": 1.1924, + "theoretical_loss": 3.35874187917176, + "tokens_seen": 2890858496 + }, + { + "epoch": 0.88, + "learning_rate": 0.00010038114343029088, + "loss": 1.311, + "theoretical_loss": 3.3587360307051304, + "tokens_seen": 2890924032 + }, + { + "epoch": 0.88, + "learning_rate": 0.00010036509528585758, + "loss": 1.2632, + "theoretical_loss": 3.3587301824082036, + "tokens_seen": 2890989568 + }, + { + "epoch": 0.88, + "learning_rate": 0.00010034904714142428, + "loss": 1.2615, + "theoretical_loss": 3.3587243342809705, + "tokens_seen": 2891055104 + }, + { + "epoch": 0.88, + "learning_rate": 0.00010033299899699098, + "loss": 1.2544, + "theoretical_loss": 3.358718486323423, + "tokens_seen": 2891120640 + }, + { + "epoch": 0.88, + "learning_rate": 0.00010031695085255768, + "loss": 1.2961, + "theoretical_loss": 3.3587126385355517, + "tokens_seen": 2891186176 + }, + { + "epoch": 0.88, + "learning_rate": 0.00010030090270812438, + "loss": 1.2553, + "theoretical_loss": 3.3587067909173483, + "tokens_seen": 2891251712 + }, + { + "epoch": 0.88, + "learning_rate": 0.00010028485456369108, + "loss": 1.206, + "theoretical_loss": 3.3587009434688033, + "tokens_seen": 2891317248 + }, + { + "epoch": 0.88, + "learning_rate": 0.00010026880641925778, + "loss": 1.2086, + "theoretical_loss": 3.3586950961899094, + "tokens_seen": 2891382784 + }, + { + "epoch": 0.88, + "learning_rate": 0.00010025275827482447, + "loss": 1.3142, + "theoretical_loss": 3.358689249080656, + "tokens_seen": 2891448320 + }, + { + "epoch": 0.88, + "learning_rate": 0.00010023671013039117, + "loss": 1.3411, + "theoretical_loss": 3.3586834021410357, + "tokens_seen": 2891513856 + }, + { + "epoch": 0.88, + "learning_rate": 0.00010022066198595787, + "loss": 1.2273, + "theoretical_loss": 3.358677555371039, + "tokens_seen": 2891579392 + }, + { + "epoch": 0.88, + "learning_rate": 0.00010020461384152457, + "loss": 1.2704, + "theoretical_loss": 3.3586717087706575, + "tokens_seen": 2891644928 + }, + { + "epoch": 0.88, + "learning_rate": 0.00010018856569709127, + "loss": 1.2395, + "theoretical_loss": 3.3586658623398824, + "tokens_seen": 2891710464 + }, + { + "epoch": 0.88, + "learning_rate": 0.00010017251755265797, + "loss": 1.3067, + "theoretical_loss": 3.358660016078705, + "tokens_seen": 2891776000 + }, + { + "epoch": 0.88, + "learning_rate": 0.00010015646940822467, + "loss": 1.3168, + "theoretical_loss": 3.3586541699871164, + "tokens_seen": 2891841536 + }, + { + "epoch": 0.88, + "learning_rate": 0.00010014042126379137, + "loss": 1.2923, + "theoretical_loss": 3.3586483240651077, + "tokens_seen": 2891907072 + }, + { + "epoch": 0.88, + "learning_rate": 0.00010012437311935807, + "loss": 1.3059, + "theoretical_loss": 3.35864247831267, + "tokens_seen": 2891972608 + }, + { + "epoch": 0.88, + "learning_rate": 0.00010010832497492478, + "loss": 1.3389, + "theoretical_loss": 3.3586366327297954, + "tokens_seen": 2892038144 + }, + { + "epoch": 0.88, + "learning_rate": 0.00010009227683049148, + "loss": 1.3045, + "theoretical_loss": 3.358630787316474, + "tokens_seen": 2892103680 + }, + { + "epoch": 0.88, + "learning_rate": 0.00010007622868605818, + "loss": 1.2661, + "theoretical_loss": 3.358624942072698, + "tokens_seen": 2892169216 + }, + { + "epoch": 0.88, + "learning_rate": 0.00010006018054162486, + "loss": 1.2427, + "theoretical_loss": 3.3586190969984577, + "tokens_seen": 2892234752 + }, + { + "epoch": 0.88, + "learning_rate": 0.00010004413239719156, + "loss": 1.2342, + "theoretical_loss": 3.358613252093745, + "tokens_seen": 2892300288 + }, + { + "epoch": 0.88, + "learning_rate": 0.00010002808425275827, + "loss": 1.1906, + "theoretical_loss": 3.3586074073585515, + "tokens_seen": 2892365824 + }, + { + "epoch": 0.88, + "learning_rate": 0.00010001203610832497, + "loss": 1.2596, + "theoretical_loss": 3.3586015627928676, + "tokens_seen": 2892431360 + }, + { + "epoch": 0.88, + "learning_rate": 9.999598796389168e-05, + "loss": 1.2577, + "theoretical_loss": 3.358595718396685, + "tokens_seen": 2892496896 + }, + { + "epoch": 0.88, + "learning_rate": 9.997993981945838e-05, + "loss": 1.3105, + "theoretical_loss": 3.358589874169995, + "tokens_seen": 2892562432 + }, + { + "epoch": 0.88, + "learning_rate": 9.996389167502508e-05, + "loss": 1.2539, + "theoretical_loss": 3.3585840301127883, + "tokens_seen": 2892627968 + }, + { + "epoch": 0.88, + "learning_rate": 9.994784353059178e-05, + "loss": 1.2466, + "theoretical_loss": 3.3585781862250568, + "tokens_seen": 2892693504 + }, + { + "epoch": 0.88, + "learning_rate": 9.993179538615848e-05, + "loss": 1.2894, + "theoretical_loss": 3.3585723425067915, + "tokens_seen": 2892759040 + }, + { + "epoch": 0.88, + "learning_rate": 9.991574724172518e-05, + "loss": 1.3065, + "theoretical_loss": 3.3585664989579835, + "tokens_seen": 2892824576 + }, + { + "epoch": 0.88, + "learning_rate": 9.989969909729187e-05, + "loss": 1.3518, + "theoretical_loss": 3.358560655578624, + "tokens_seen": 2892890112 + }, + { + "epoch": 0.88, + "learning_rate": 9.988365095285857e-05, + "loss": 1.2725, + "theoretical_loss": 3.358554812368705, + "tokens_seen": 2892955648 + }, + { + "epoch": 0.88, + "learning_rate": 9.986760280842527e-05, + "loss": 1.2185, + "theoretical_loss": 3.3585489693282167, + "tokens_seen": 2893021184 + }, + { + "epoch": 0.88, + "learning_rate": 9.985155466399197e-05, + "loss": 1.3179, + "theoretical_loss": 3.358543126457151, + "tokens_seen": 2893086720 + }, + { + "epoch": 0.88, + "learning_rate": 9.983550651955869e-05, + "loss": 1.2242, + "theoretical_loss": 3.358537283755499, + "tokens_seen": 2893152256 + }, + { + "epoch": 0.88, + "learning_rate": 9.981945837512539e-05, + "loss": 1.3035, + "theoretical_loss": 3.358531441223252, + "tokens_seen": 2893217792 + }, + { + "epoch": 0.88, + "learning_rate": 9.980341023069209e-05, + "loss": 1.1952, + "theoretical_loss": 3.358525598860401, + "tokens_seen": 2893283328 + }, + { + "epoch": 0.88, + "learning_rate": 9.978736208625879e-05, + "loss": 1.2689, + "theoretical_loss": 3.3585197566669374, + "tokens_seen": 2893348864 + }, + { + "epoch": 0.88, + "objective/train/docs_used": 1051818, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.2313382625579834, + "objective/train/theoretical_loss": 3.3585139146428524, + "objective/train/tokens_used": 2913874400, + "theoretical_loss": 3.3585139146428524, + "tokens_seen": 2893414400 + }, + { + "epoch": 0.88, + "learning_rate": 9.977131394182549e-05, + "loss": 1.2538, + "theoretical_loss": 3.3585139146428524, + "tokens_seen": 2893414400 + }, + { + "epoch": 0.88, + "learning_rate": 9.975526579739219e-05, + "loss": 1.3429, + "theoretical_loss": 3.358508072788138, + "tokens_seen": 2893479936 + }, + { + "epoch": 0.88, + "learning_rate": 9.973921765295888e-05, + "loss": 1.1886, + "theoretical_loss": 3.3585022311027837, + "tokens_seen": 2893545472 + }, + { + "epoch": 0.88, + "learning_rate": 9.972316950852558e-05, + "loss": 1.2089, + "theoretical_loss": 3.3584963895867825, + "tokens_seen": 2893611008 + }, + { + "epoch": 0.88, + "learning_rate": 9.970712136409228e-05, + "loss": 1.256, + "theoretical_loss": 3.358490548240125, + "tokens_seen": 2893676544 + }, + { + "epoch": 0.88, + "learning_rate": 9.969107321965898e-05, + "loss": 1.2958, + "theoretical_loss": 3.3584847070628023, + "tokens_seen": 2893742080 + }, + { + "epoch": 0.88, + "learning_rate": 9.967502507522568e-05, + "loss": 1.3218, + "theoretical_loss": 3.358478866054806, + "tokens_seen": 2893807616 + }, + { + "epoch": 0.88, + "learning_rate": 9.965897693079238e-05, + "loss": 1.2762, + "theoretical_loss": 3.358473025216127, + "tokens_seen": 2893873152 + }, + { + "epoch": 0.88, + "learning_rate": 9.964292878635908e-05, + "loss": 1.2142, + "theoretical_loss": 3.3584671845467566, + "tokens_seen": 2893938688 + }, + { + "epoch": 0.88, + "learning_rate": 9.962688064192578e-05, + "loss": 1.4181, + "theoretical_loss": 3.3584613440466864, + "tokens_seen": 2894004224 + }, + { + "epoch": 0.88, + "learning_rate": 9.961083249749248e-05, + "loss": 1.3357, + "theoretical_loss": 3.3584555037159074, + "tokens_seen": 2894069760 + }, + { + "epoch": 0.88, + "learning_rate": 9.959478435305918e-05, + "loss": 1.2218, + "theoretical_loss": 3.3584496635544108, + "tokens_seen": 2894135296 + }, + { + "epoch": 0.88, + "learning_rate": 9.957873620862588e-05, + "loss": 1.3412, + "theoretical_loss": 3.3584438235621876, + "tokens_seen": 2894200832 + }, + { + "epoch": 0.88, + "learning_rate": 9.956268806419258e-05, + "loss": 1.2112, + "theoretical_loss": 3.35843798373923, + "tokens_seen": 2894266368 + }, + { + "epoch": 0.88, + "learning_rate": 9.954663991975928e-05, + "loss": 1.3495, + "theoretical_loss": 3.3584321440855285, + "tokens_seen": 2894331904 + }, + { + "epoch": 0.88, + "learning_rate": 9.953059177532598e-05, + "loss": 1.3349, + "theoretical_loss": 3.3584263046010747, + "tokens_seen": 2894397440 + }, + { + "epoch": 0.88, + "learning_rate": 9.951454363089269e-05, + "loss": 1.3296, + "theoretical_loss": 3.3584204652858594, + "tokens_seen": 2894462976 + }, + { + "epoch": 0.88, + "learning_rate": 9.949849548645939e-05, + "loss": 1.4325, + "theoretical_loss": 3.3584146261398744, + "tokens_seen": 2894528512 + }, + { + "epoch": 0.88, + "learning_rate": 9.948244734202609e-05, + "loss": 1.2561, + "theoretical_loss": 3.3584087871631105, + "tokens_seen": 2894594048 + }, + { + "epoch": 0.88, + "learning_rate": 9.946639919759279e-05, + "loss": 1.3132, + "theoretical_loss": 3.3584029483555593, + "tokens_seen": 2894659584 + }, + { + "epoch": 0.88, + "learning_rate": 9.945035105315949e-05, + "loss": 1.2208, + "theoretical_loss": 3.3583971097172123, + "tokens_seen": 2894725120 + }, + { + "epoch": 0.88, + "learning_rate": 9.943430290872619e-05, + "loss": 1.2105, + "theoretical_loss": 3.3583912712480597, + "tokens_seen": 2894790656 + }, + { + "epoch": 0.88, + "learning_rate": 9.941825476429289e-05, + "loss": 1.2763, + "theoretical_loss": 3.358385432948094, + "tokens_seen": 2894856192 + }, + { + "epoch": 0.88, + "learning_rate": 9.940220661985959e-05, + "loss": 1.3873, + "theoretical_loss": 3.358379594817306, + "tokens_seen": 2894921728 + }, + { + "epoch": 0.88, + "learning_rate": 9.938615847542628e-05, + "loss": 1.2439, + "theoretical_loss": 3.358373756855687, + "tokens_seen": 2894987264 + }, + { + "epoch": 0.88, + "learning_rate": 9.937011033099298e-05, + "loss": 1.1727, + "theoretical_loss": 3.3583679190632276, + "tokens_seen": 2895052800 + }, + { + "epoch": 0.88, + "learning_rate": 9.935406218655968e-05, + "loss": 1.3418, + "theoretical_loss": 3.35836208143992, + "tokens_seen": 2895118336 + }, + { + "epoch": 0.88, + "learning_rate": 9.933801404212638e-05, + "loss": 1.2638, + "theoretical_loss": 3.3583562439857557, + "tokens_seen": 2895183872 + }, + { + "epoch": 0.88, + "learning_rate": 9.932196589769308e-05, + "loss": 1.2965, + "theoretical_loss": 3.358350406700725, + "tokens_seen": 2895249408 + }, + { + "epoch": 0.88, + "learning_rate": 9.930591775325978e-05, + "loss": 1.2289, + "theoretical_loss": 3.358344569584819, + "tokens_seen": 2895314944 + }, + { + "epoch": 0.88, + "learning_rate": 9.928986960882648e-05, + "loss": 1.2311, + "theoretical_loss": 3.35833873263803, + "tokens_seen": 2895380480 + }, + { + "epoch": 0.88, + "learning_rate": 9.927382146439318e-05, + "loss": 1.1961, + "theoretical_loss": 3.3583328958603493, + "tokens_seen": 2895446016 + }, + { + "epoch": 0.88, + "learning_rate": 9.925777331995988e-05, + "loss": 1.2195, + "theoretical_loss": 3.358327059251767, + "tokens_seen": 2895511552 + }, + { + "epoch": 0.88, + "learning_rate": 9.92417251755266e-05, + "loss": 1.1938, + "theoretical_loss": 3.3583212228122754, + "tokens_seen": 2895577088 + }, + { + "epoch": 0.88, + "learning_rate": 9.922567703109328e-05, + "loss": 1.2428, + "theoretical_loss": 3.3583153865418653, + "tokens_seen": 2895642624 + }, + { + "epoch": 0.88, + "learning_rate": 9.920962888665998e-05, + "loss": 1.25, + "theoretical_loss": 3.3583095504405285, + "tokens_seen": 2895708160 + }, + { + "epoch": 0.88, + "learning_rate": 9.919358074222668e-05, + "loss": 1.298, + "theoretical_loss": 3.3583037145082555, + "tokens_seen": 2895773696 + }, + { + "epoch": 0.88, + "learning_rate": 9.917753259779339e-05, + "loss": 1.274, + "theoretical_loss": 3.358297878745038, + "tokens_seen": 2895839232 + }, + { + "epoch": 0.88, + "learning_rate": 9.916148445336009e-05, + "loss": 1.2429, + "theoretical_loss": 3.3582920431508674, + "tokens_seen": 2895904768 + }, + { + "epoch": 0.88, + "learning_rate": 9.914543630892679e-05, + "loss": 1.2785, + "theoretical_loss": 3.358286207725735, + "tokens_seen": 2895970304 + }, + { + "epoch": 0.88, + "learning_rate": 9.912938816449349e-05, + "loss": 1.323, + "theoretical_loss": 3.3582803724696313, + "tokens_seen": 2896035840 + }, + { + "epoch": 0.88, + "learning_rate": 9.911334002006019e-05, + "loss": 1.2594, + "theoretical_loss": 3.358274537382549, + "tokens_seen": 2896101376 + }, + { + "epoch": 0.88, + "learning_rate": 9.909729187562689e-05, + "loss": 1.2544, + "theoretical_loss": 3.3582687024644775, + "tokens_seen": 2896166912 + }, + { + "epoch": 0.88, + "learning_rate": 9.908124373119359e-05, + "loss": 1.2102, + "theoretical_loss": 3.35826286771541, + "tokens_seen": 2896232448 + }, + { + "epoch": 0.88, + "learning_rate": 9.906519558676029e-05, + "loss": 1.2056, + "theoretical_loss": 3.3582570331353367, + "tokens_seen": 2896297984 + }, + { + "epoch": 0.88, + "learning_rate": 9.904914744232699e-05, + "loss": 1.2856, + "theoretical_loss": 3.358251198724249, + "tokens_seen": 2896363520 + }, + { + "epoch": 0.88, + "learning_rate": 9.903309929789368e-05, + "loss": 1.2858, + "theoretical_loss": 3.3582453644821384, + "tokens_seen": 2896429056 + }, + { + "epoch": 0.88, + "learning_rate": 9.901705115346038e-05, + "loss": 1.3396, + "theoretical_loss": 3.358239530408996, + "tokens_seen": 2896494592 + }, + { + "epoch": 0.88, + "learning_rate": 9.900100300902708e-05, + "loss": 1.2154, + "theoretical_loss": 3.358233696504813, + "tokens_seen": 2896560128 + }, + { + "epoch": 0.88, + "learning_rate": 9.898495486459378e-05, + "loss": 1.2305, + "theoretical_loss": 3.3582278627695814, + "tokens_seen": 2896625664 + }, + { + "epoch": 0.88, + "objective/train/docs_used": 1053062, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.2450370788574219, + "objective/train/theoretical_loss": 3.358222029203292, + "objective/train/tokens_used": 2917151200, + "theoretical_loss": 3.358222029203292, + "tokens_seen": 2896691200 + }, + { + "epoch": 0.88, + "learning_rate": 9.896890672016048e-05, + "loss": 1.2502, + "theoretical_loss": 3.358222029203292, + "tokens_seen": 2896691200 + }, + { + "epoch": 0.88, + "learning_rate": 9.895285857572718e-05, + "loss": 1.2722, + "theoretical_loss": 3.358216195805935, + "tokens_seen": 2896756736 + }, + { + "epoch": 0.88, + "learning_rate": 9.89368104312939e-05, + "loss": 1.2275, + "theoretical_loss": 3.3582103625775037, + "tokens_seen": 2896822272 + }, + { + "epoch": 0.88, + "learning_rate": 9.89207622868606e-05, + "loss": 1.2283, + "theoretical_loss": 3.358204529517988, + "tokens_seen": 2896887808 + }, + { + "epoch": 0.88, + "learning_rate": 9.89047141424273e-05, + "loss": 1.3359, + "theoretical_loss": 3.358198696627379, + "tokens_seen": 2896953344 + }, + { + "epoch": 0.88, + "learning_rate": 9.8888665997994e-05, + "loss": 1.2883, + "theoretical_loss": 3.3581928639056695, + "tokens_seen": 2897018880 + }, + { + "epoch": 0.88, + "learning_rate": 9.887261785356068e-05, + "loss": 1.3566, + "theoretical_loss": 3.3581870313528492, + "tokens_seen": 2897084416 + }, + { + "epoch": 0.88, + "learning_rate": 9.885656970912738e-05, + "loss": 1.3437, + "theoretical_loss": 3.3581811989689103, + "tokens_seen": 2897149952 + }, + { + "epoch": 0.88, + "learning_rate": 9.884052156469408e-05, + "loss": 1.3514, + "theoretical_loss": 3.358175366753844, + "tokens_seen": 2897215488 + }, + { + "epoch": 0.88, + "learning_rate": 9.882447342026079e-05, + "loss": 1.3934, + "theoretical_loss": 3.358169534707641, + "tokens_seen": 2897281024 + }, + { + "epoch": 0.88, + "learning_rate": 9.880842527582749e-05, + "loss": 1.297, + "theoretical_loss": 3.3581637028302933, + "tokens_seen": 2897346560 + }, + { + "epoch": 0.88, + "learning_rate": 9.879237713139419e-05, + "loss": 1.3828, + "theoretical_loss": 3.358157871121792, + "tokens_seen": 2897412096 + }, + { + "epoch": 0.88, + "learning_rate": 9.877632898696089e-05, + "loss": 1.2786, + "theoretical_loss": 3.358152039582128, + "tokens_seen": 2897477632 + }, + { + "epoch": 0.88, + "learning_rate": 9.876028084252759e-05, + "loss": 1.2868, + "theoretical_loss": 3.3581462082112927, + "tokens_seen": 2897543168 + }, + { + "epoch": 0.88, + "learning_rate": 9.874423269809429e-05, + "loss": 1.2475, + "theoretical_loss": 3.358140377009278, + "tokens_seen": 2897608704 + }, + { + "epoch": 0.88, + "learning_rate": 9.872818455366099e-05, + "loss": 1.2374, + "theoretical_loss": 3.3581345459760747, + "tokens_seen": 2897674240 + }, + { + "epoch": 0.88, + "learning_rate": 9.871213640922769e-05, + "loss": 1.2971, + "theoretical_loss": 3.358128715111674, + "tokens_seen": 2897739776 + }, + { + "epoch": 0.88, + "learning_rate": 9.869608826479439e-05, + "loss": 1.2838, + "theoretical_loss": 3.3581228844160678, + "tokens_seen": 2897805312 + }, + { + "epoch": 0.88, + "learning_rate": 9.868004012036108e-05, + "loss": 1.3046, + "theoretical_loss": 3.3581170538892464, + "tokens_seen": 2897870848 + }, + { + "epoch": 0.88, + "learning_rate": 9.866399197592778e-05, + "loss": 1.2896, + "theoretical_loss": 3.358111223531202, + "tokens_seen": 2897936384 + }, + { + "epoch": 0.88, + "learning_rate": 9.864794383149449e-05, + "loss": 1.2241, + "theoretical_loss": 3.3581053933419254, + "tokens_seen": 2898001920 + }, + { + "epoch": 0.88, + "learning_rate": 9.86318956870612e-05, + "loss": 1.2214, + "theoretical_loss": 3.3580995633214084, + "tokens_seen": 2898067456 + }, + { + "epoch": 0.88, + "learning_rate": 9.86158475426279e-05, + "loss": 1.2969, + "theoretical_loss": 3.3580937334696417, + "tokens_seen": 2898132992 + }, + { + "epoch": 0.88, + "learning_rate": 9.85997993981946e-05, + "loss": 1.2061, + "theoretical_loss": 3.3580879037866165, + "tokens_seen": 2898198528 + }, + { + "epoch": 0.88, + "learning_rate": 9.85837512537613e-05, + "loss": 1.3278, + "theoretical_loss": 3.3580820742723247, + "tokens_seen": 2898264064 + }, + { + "epoch": 0.88, + "learning_rate": 9.8567703109328e-05, + "loss": 1.2198, + "theoretical_loss": 3.3580762449267576, + "tokens_seen": 2898329600 + }, + { + "epoch": 0.88, + "learning_rate": 9.85516549648947e-05, + "loss": 1.1829, + "theoretical_loss": 3.358070415749906, + "tokens_seen": 2898395136 + }, + { + "epoch": 0.88, + "learning_rate": 9.853560682046138e-05, + "loss": 1.2937, + "theoretical_loss": 3.3580645867417616, + "tokens_seen": 2898460672 + }, + { + "epoch": 0.88, + "learning_rate": 9.851955867602808e-05, + "loss": 1.3188, + "theoretical_loss": 3.358058757902316, + "tokens_seen": 2898526208 + }, + { + "epoch": 0.88, + "learning_rate": 9.850351053159478e-05, + "loss": 1.1774, + "theoretical_loss": 3.358052929231559, + "tokens_seen": 2898591744 + }, + { + "epoch": 0.88, + "learning_rate": 9.848746238716149e-05, + "loss": 1.2107, + "theoretical_loss": 3.358047100729484, + "tokens_seen": 2898657280 + }, + { + "epoch": 0.88, + "learning_rate": 9.847141424272819e-05, + "loss": 1.2899, + "theoretical_loss": 3.358041272396081, + "tokens_seen": 2898722816 + }, + { + "epoch": 0.88, + "learning_rate": 9.845536609829489e-05, + "loss": 1.287, + "theoretical_loss": 3.358035444231341, + "tokens_seen": 2898788352 + }, + { + "epoch": 0.88, + "learning_rate": 9.843931795386159e-05, + "loss": 1.3224, + "theoretical_loss": 3.3580296162352563, + "tokens_seen": 2898853888 + }, + { + "epoch": 0.88, + "learning_rate": 9.842326980942829e-05, + "loss": 1.2146, + "theoretical_loss": 3.358023788407818, + "tokens_seen": 2898919424 + }, + { + "epoch": 0.88, + "learning_rate": 9.840722166499499e-05, + "loss": 1.1484, + "theoretical_loss": 3.3580179607490166, + "tokens_seen": 2898984960 + }, + { + "epoch": 0.88, + "learning_rate": 9.839117352056169e-05, + "loss": 1.2111, + "theoretical_loss": 3.358012133258845, + "tokens_seen": 2899050496 + }, + { + "epoch": 0.88, + "learning_rate": 9.837512537612839e-05, + "loss": 1.1896, + "theoretical_loss": 3.358006305937293, + "tokens_seen": 2899116032 + }, + { + "epoch": 0.88, + "learning_rate": 9.835907723169509e-05, + "loss": 1.3683, + "theoretical_loss": 3.358000478784352, + "tokens_seen": 2899181568 + }, + { + "epoch": 0.88, + "learning_rate": 9.834302908726179e-05, + "loss": 1.3166, + "theoretical_loss": 3.357994651800014, + "tokens_seen": 2899247104 + }, + { + "epoch": 0.88, + "learning_rate": 9.832698094282849e-05, + "loss": 1.3062, + "theoretical_loss": 3.35798882498427, + "tokens_seen": 2899312640 + }, + { + "epoch": 0.88, + "learning_rate": 9.831093279839519e-05, + "loss": 1.2973, + "theoretical_loss": 3.3579829983371114, + "tokens_seen": 2899378176 + }, + { + "epoch": 0.88, + "learning_rate": 9.829488465396189e-05, + "loss": 1.3074, + "theoretical_loss": 3.35797717185853, + "tokens_seen": 2899443712 + }, + { + "epoch": 0.88, + "learning_rate": 9.82788365095286e-05, + "loss": 1.2688, + "theoretical_loss": 3.357971345548516, + "tokens_seen": 2899509248 + }, + { + "epoch": 0.88, + "learning_rate": 9.82627883650953e-05, + "loss": 1.2396, + "theoretical_loss": 3.3579655194070614, + "tokens_seen": 2899574784 + }, + { + "epoch": 0.88, + "learning_rate": 9.8246740220662e-05, + "loss": 1.1859, + "theoretical_loss": 3.3579596934341573, + "tokens_seen": 2899640320 + }, + { + "epoch": 0.88, + "learning_rate": 9.82306920762287e-05, + "loss": 1.3892, + "theoretical_loss": 3.3579538676297953, + "tokens_seen": 2899705856 + }, + { + "epoch": 0.88, + "learning_rate": 9.82146439317954e-05, + "loss": 1.2972, + "theoretical_loss": 3.357948041993967, + "tokens_seen": 2899771392 + }, + { + "epoch": 0.88, + "learning_rate": 9.81985957873621e-05, + "loss": 1.2099, + "theoretical_loss": 3.3579422165266624, + "tokens_seen": 2899836928 + }, + { + "epoch": 0.88, + "learning_rate": 9.818254764292878e-05, + "loss": 1.2814, + "theoretical_loss": 3.357936391227874, + "tokens_seen": 2899902464 + }, + { + "epoch": 0.88, + "objective/train/docs_used": 1054339, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.3116545677185059, + "objective/train/theoretical_loss": 3.357930566097593, + "objective/train/tokens_used": 2920428000, + "theoretical_loss": 3.357930566097593, + "tokens_seen": 2899968000 + }, + { + "epoch": 0.88, + "learning_rate": 9.816649949849548e-05, + "loss": 1.3388, + "theoretical_loss": 3.357930566097593, + "tokens_seen": 2899968000 + }, + { + "epoch": 0.88, + "learning_rate": 9.815045135406219e-05, + "loss": 1.3, + "theoretical_loss": 3.35792474113581, + "tokens_seen": 2900033536 + }, + { + "epoch": 0.88, + "learning_rate": 9.813440320962889e-05, + "loss": 1.3511, + "theoretical_loss": 3.357918916342517, + "tokens_seen": 2900099072 + }, + { + "epoch": 0.88, + "learning_rate": 9.811835506519559e-05, + "loss": 1.3322, + "theoretical_loss": 3.357913091717706, + "tokens_seen": 2900164608 + }, + { + "epoch": 0.88, + "learning_rate": 9.810230692076229e-05, + "loss": 1.2839, + "theoretical_loss": 3.3579072672613663, + "tokens_seen": 2900230144 + }, + { + "epoch": 0.88, + "learning_rate": 9.808625877632899e-05, + "loss": 1.2969, + "theoretical_loss": 3.357901442973491, + "tokens_seen": 2900295680 + }, + { + "epoch": 0.88, + "learning_rate": 9.807021063189569e-05, + "loss": 1.3181, + "theoretical_loss": 3.3578956188540707, + "tokens_seen": 2900361216 + }, + { + "epoch": 0.88, + "learning_rate": 9.805416248746239e-05, + "loss": 1.1815, + "theoretical_loss": 3.357889794903097, + "tokens_seen": 2900426752 + }, + { + "epoch": 0.88, + "learning_rate": 9.80381143430291e-05, + "loss": 1.2131, + "theoretical_loss": 3.3578839711205606, + "tokens_seen": 2900492288 + }, + { + "epoch": 0.88, + "learning_rate": 9.802206619859579e-05, + "loss": 1.2337, + "theoretical_loss": 3.3578781475064536, + "tokens_seen": 2900557824 + }, + { + "epoch": 0.88, + "learning_rate": 9.800601805416249e-05, + "loss": 1.2614, + "theoretical_loss": 3.3578723240607666, + "tokens_seen": 2900623360 + }, + { + "epoch": 0.88, + "learning_rate": 9.798996990972919e-05, + "loss": 1.3233, + "theoretical_loss": 3.3578665007834916, + "tokens_seen": 2900688896 + }, + { + "epoch": 0.88, + "learning_rate": 9.797392176529589e-05, + "loss": 1.2151, + "theoretical_loss": 3.35786067767462, + "tokens_seen": 2900754432 + }, + { + "epoch": 0.88, + "learning_rate": 9.795787362086259e-05, + "loss": 1.231, + "theoretical_loss": 3.3578548547341422, + "tokens_seen": 2900819968 + }, + { + "epoch": 0.88, + "learning_rate": 9.79418254764293e-05, + "loss": 1.1843, + "theoretical_loss": 3.3578490319620506, + "tokens_seen": 2900885504 + }, + { + "epoch": 0.88, + "learning_rate": 9.7925777331996e-05, + "loss": 1.1744, + "theoretical_loss": 3.3578432093583355, + "tokens_seen": 2900951040 + }, + { + "epoch": 0.88, + "learning_rate": 9.79097291875627e-05, + "loss": 1.2098, + "theoretical_loss": 3.357837386922989, + "tokens_seen": 2901016576 + }, + { + "epoch": 0.88, + "learning_rate": 9.78936810431294e-05, + "loss": 1.3477, + "theoretical_loss": 3.3578315646560024, + "tokens_seen": 2901082112 + }, + { + "epoch": 0.88, + "learning_rate": 9.78776328986961e-05, + "loss": 1.2764, + "theoretical_loss": 3.3578257425573663, + "tokens_seen": 2901147648 + }, + { + "epoch": 0.88, + "learning_rate": 9.78615847542628e-05, + "loss": 1.294, + "theoretical_loss": 3.357819920627073, + "tokens_seen": 2901213184 + }, + { + "epoch": 0.88, + "learning_rate": 9.78455366098295e-05, + "loss": 1.1801, + "theoretical_loss": 3.3578140988651133, + "tokens_seen": 2901278720 + }, + { + "epoch": 0.88, + "learning_rate": 9.782948846539618e-05, + "loss": 1.3532, + "theoretical_loss": 3.357808277271478, + "tokens_seen": 2901344256 + }, + { + "epoch": 0.88, + "learning_rate": 9.781344032096288e-05, + "loss": 1.2042, + "theoretical_loss": 3.35780245584616, + "tokens_seen": 2901409792 + }, + { + "epoch": 0.88, + "learning_rate": 9.779739217652959e-05, + "loss": 1.3, + "theoretical_loss": 3.357796634589149, + "tokens_seen": 2901475328 + }, + { + "epoch": 0.88, + "learning_rate": 9.778134403209629e-05, + "loss": 1.2607, + "theoretical_loss": 3.3577908135004373, + "tokens_seen": 2901540864 + }, + { + "epoch": 0.88, + "learning_rate": 9.776529588766299e-05, + "loss": 1.1933, + "theoretical_loss": 3.3577849925800156, + "tokens_seen": 2901606400 + }, + { + "epoch": 0.88, + "learning_rate": 9.774924774322969e-05, + "loss": 1.2897, + "theoretical_loss": 3.3577791718278758, + "tokens_seen": 2901671936 + }, + { + "epoch": 0.88, + "learning_rate": 9.77331995987964e-05, + "loss": 1.2217, + "theoretical_loss": 3.357773351244009, + "tokens_seen": 2901737472 + }, + { + "epoch": 0.88, + "learning_rate": 9.77171514543631e-05, + "loss": 1.3042, + "theoretical_loss": 3.3577675308284065, + "tokens_seen": 2901803008 + }, + { + "epoch": 0.88, + "learning_rate": 9.77011033099298e-05, + "loss": 1.1787, + "theoretical_loss": 3.3577617105810598, + "tokens_seen": 2901868544 + }, + { + "epoch": 0.88, + "learning_rate": 9.76850551654965e-05, + "loss": 1.348, + "theoretical_loss": 3.3577558905019598, + "tokens_seen": 2901934080 + }, + { + "epoch": 0.88, + "learning_rate": 9.766900702106319e-05, + "loss": 1.2755, + "theoretical_loss": 3.357750070591098, + "tokens_seen": 2901999616 + }, + { + "epoch": 0.88, + "learning_rate": 9.765295887662989e-05, + "loss": 1.2177, + "theoretical_loss": 3.3577442508484663, + "tokens_seen": 2902065152 + }, + { + "epoch": 0.88, + "learning_rate": 9.763691073219659e-05, + "loss": 1.3112, + "theoretical_loss": 3.3577384312740555, + "tokens_seen": 2902130688 + }, + { + "epoch": 0.88, + "learning_rate": 9.762086258776329e-05, + "loss": 1.2833, + "theoretical_loss": 3.357732611867857, + "tokens_seen": 2902196224 + }, + { + "epoch": 0.88, + "learning_rate": 9.760481444332999e-05, + "loss": 1.2091, + "theoretical_loss": 3.357726792629862, + "tokens_seen": 2902261760 + }, + { + "epoch": 0.88, + "learning_rate": 9.75887662988967e-05, + "loss": 1.2421, + "theoretical_loss": 3.3577209735600624, + "tokens_seen": 2902327296 + }, + { + "epoch": 0.88, + "learning_rate": 9.75727181544634e-05, + "loss": 1.2064, + "theoretical_loss": 3.357715154658449, + "tokens_seen": 2902392832 + }, + { + "epoch": 0.88, + "learning_rate": 9.75566700100301e-05, + "loss": 1.2681, + "theoretical_loss": 3.3577093359250134, + "tokens_seen": 2902458368 + }, + { + "epoch": 0.88, + "learning_rate": 9.75406218655968e-05, + "loss": 1.3169, + "theoretical_loss": 3.357703517359747, + "tokens_seen": 2902523904 + }, + { + "epoch": 0.88, + "learning_rate": 9.75245737211635e-05, + "loss": 1.3129, + "theoretical_loss": 3.3576976989626406, + "tokens_seen": 2902589440 + }, + { + "epoch": 0.88, + "learning_rate": 9.75085255767302e-05, + "loss": 1.2612, + "theoretical_loss": 3.357691880733686, + "tokens_seen": 2902654976 + }, + { + "epoch": 0.88, + "learning_rate": 9.74924774322969e-05, + "loss": 1.2463, + "theoretical_loss": 3.3576860626728746, + "tokens_seen": 2902720512 + }, + { + "epoch": 0.88, + "learning_rate": 9.747642928786358e-05, + "loss": 1.3103, + "theoretical_loss": 3.3576802447801977, + "tokens_seen": 2902786048 + }, + { + "epoch": 0.88, + "learning_rate": 9.746038114343029e-05, + "loss": 1.2749, + "theoretical_loss": 3.3576744270556462, + "tokens_seen": 2902851584 + }, + { + "epoch": 0.88, + "learning_rate": 9.7444332998997e-05, + "loss": 1.1659, + "theoretical_loss": 3.3576686094992123, + "tokens_seen": 2902917120 + }, + { + "epoch": 0.88, + "learning_rate": 9.74282848545637e-05, + "loss": 1.257, + "theoretical_loss": 3.3576627921108866, + "tokens_seen": 2902982656 + }, + { + "epoch": 0.88, + "learning_rate": 9.74122367101304e-05, + "loss": 1.1893, + "theoretical_loss": 3.357656974890661, + "tokens_seen": 2903048192 + }, + { + "epoch": 0.88, + "learning_rate": 9.73961885656971e-05, + "loss": 1.2811, + "theoretical_loss": 3.3576511578385264, + "tokens_seen": 2903113728 + }, + { + "epoch": 0.88, + "learning_rate": 9.73801404212638e-05, + "loss": 1.3112, + "theoretical_loss": 3.357645340954474, + "tokens_seen": 2903179264 + }, + { + "epoch": 0.88, + "objective/train/docs_used": 1055596, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.3417952060699463, + "objective/train/theoretical_loss": 3.357639524238496, + "objective/train/tokens_used": 2923704800, + "theoretical_loss": 3.357639524238496, + "tokens_seen": 2903244800 + }, + { + "epoch": 0.88, + "learning_rate": 9.73640922768305e-05, + "loss": 1.3109, + "theoretical_loss": 3.357639524238496, + "tokens_seen": 2903244800 + }, + { + "epoch": 0.88, + "learning_rate": 9.73480441323972e-05, + "loss": 1.3789, + "theoretical_loss": 3.357633707690583, + "tokens_seen": 2903310336 + }, + { + "epoch": 0.88, + "learning_rate": 9.73319959879639e-05, + "loss": 1.2207, + "theoretical_loss": 3.3576278913107265, + "tokens_seen": 2903375872 + }, + { + "epoch": 0.88, + "learning_rate": 9.731594784353059e-05, + "loss": 1.292, + "theoretical_loss": 3.357622075098918, + "tokens_seen": 2903441408 + }, + { + "epoch": 0.88, + "learning_rate": 9.729989969909729e-05, + "loss": 1.3202, + "theoretical_loss": 3.3576162590551486, + "tokens_seen": 2903506944 + }, + { + "epoch": 0.88, + "learning_rate": 9.728385155466399e-05, + "loss": 1.2926, + "theoretical_loss": 3.35761044317941, + "tokens_seen": 2903572480 + }, + { + "epoch": 0.88, + "learning_rate": 9.726780341023069e-05, + "loss": 1.1129, + "theoretical_loss": 3.3576046274716935, + "tokens_seen": 2903638016 + }, + { + "epoch": 0.88, + "learning_rate": 9.72517552657974e-05, + "loss": 1.2925, + "theoretical_loss": 3.35759881193199, + "tokens_seen": 2903703552 + }, + { + "epoch": 0.88, + "learning_rate": 9.72357071213641e-05, + "loss": 1.1689, + "theoretical_loss": 3.357592996560291, + "tokens_seen": 2903769088 + }, + { + "epoch": 0.88, + "learning_rate": 9.72196589769308e-05, + "loss": 1.257, + "theoretical_loss": 3.3575871813565885, + "tokens_seen": 2903834624 + }, + { + "epoch": 0.88, + "learning_rate": 9.72036108324975e-05, + "loss": 1.2556, + "theoretical_loss": 3.3575813663208733, + "tokens_seen": 2903900160 + }, + { + "epoch": 0.88, + "learning_rate": 9.71875626880642e-05, + "loss": 1.3278, + "theoretical_loss": 3.3575755514531367, + "tokens_seen": 2903965696 + }, + { + "epoch": 0.88, + "learning_rate": 9.71715145436309e-05, + "loss": 1.3159, + "theoretical_loss": 3.35756973675337, + "tokens_seen": 2904031232 + }, + { + "epoch": 0.88, + "learning_rate": 9.71554663991976e-05, + "loss": 1.2257, + "theoretical_loss": 3.3575639222215656, + "tokens_seen": 2904096768 + }, + { + "epoch": 0.88, + "learning_rate": 9.71394182547643e-05, + "loss": 1.225, + "theoretical_loss": 3.357558107857713, + "tokens_seen": 2904162304 + }, + { + "epoch": 0.88, + "learning_rate": 9.7123370110331e-05, + "loss": 1.3704, + "theoretical_loss": 3.357552293661805, + "tokens_seen": 2904227840 + }, + { + "epoch": 0.88, + "learning_rate": 9.71073219658977e-05, + "loss": 1.3099, + "theoretical_loss": 3.357546479633833, + "tokens_seen": 2904293376 + }, + { + "epoch": 0.88, + "learning_rate": 9.70912738214644e-05, + "loss": 1.1834, + "theoretical_loss": 3.3575406657737874, + "tokens_seen": 2904358912 + }, + { + "epoch": 0.88, + "learning_rate": 9.70752256770311e-05, + "loss": 1.3154, + "theoretical_loss": 3.3575348520816597, + "tokens_seen": 2904424448 + }, + { + "epoch": 0.88, + "learning_rate": 9.70591775325978e-05, + "loss": 1.1742, + "theoretical_loss": 3.357529038557442, + "tokens_seen": 2904489984 + }, + { + "epoch": 0.88, + "learning_rate": 9.70431293881645e-05, + "loss": 1.2789, + "theoretical_loss": 3.357523225201126, + "tokens_seen": 2904555520 + }, + { + "epoch": 0.88, + "learning_rate": 9.70270812437312e-05, + "loss": 1.2677, + "theoretical_loss": 3.3575174120127014, + "tokens_seen": 2904621056 + }, + { + "epoch": 0.88, + "learning_rate": 9.70110330992979e-05, + "loss": 1.3091, + "theoretical_loss": 3.3575115989921605, + "tokens_seen": 2904686592 + }, + { + "epoch": 0.88, + "learning_rate": 9.69949849548646e-05, + "loss": 1.2459, + "theoretical_loss": 3.357505786139495, + "tokens_seen": 2904752128 + }, + { + "epoch": 0.88, + "learning_rate": 9.69789368104313e-05, + "loss": 1.299, + "theoretical_loss": 3.357499973454696, + "tokens_seen": 2904817664 + }, + { + "epoch": 0.88, + "learning_rate": 9.696288866599799e-05, + "loss": 1.1092, + "theoretical_loss": 3.3574941609377547, + "tokens_seen": 2904883200 + }, + { + "epoch": 0.88, + "learning_rate": 9.694684052156469e-05, + "loss": 1.135, + "theoretical_loss": 3.3574883485886624, + "tokens_seen": 2904948736 + }, + { + "epoch": 0.88, + "learning_rate": 9.693079237713139e-05, + "loss": 1.2017, + "theoretical_loss": 3.357482536407411, + "tokens_seen": 2905014272 + }, + { + "epoch": 0.88, + "learning_rate": 9.69147442326981e-05, + "loss": 1.2658, + "theoretical_loss": 3.357476724393991, + "tokens_seen": 2905079808 + }, + { + "epoch": 0.88, + "learning_rate": 9.68986960882648e-05, + "loss": 1.2592, + "theoretical_loss": 3.357470912548395, + "tokens_seen": 2905145344 + }, + { + "epoch": 0.88, + "learning_rate": 9.68826479438315e-05, + "loss": 1.2765, + "theoretical_loss": 3.3574651008706127, + "tokens_seen": 2905210880 + }, + { + "epoch": 0.88, + "learning_rate": 9.68665997993982e-05, + "loss": 1.3678, + "theoretical_loss": 3.357459289360637, + "tokens_seen": 2905276416 + }, + { + "epoch": 0.88, + "learning_rate": 9.68505516549649e-05, + "loss": 1.3191, + "theoretical_loss": 3.3574534780184586, + "tokens_seen": 2905341952 + }, + { + "epoch": 0.88, + "learning_rate": 9.683450351053161e-05, + "loss": 1.1725, + "theoretical_loss": 3.357447666844069, + "tokens_seen": 2905407488 + }, + { + "epoch": 0.88, + "learning_rate": 9.681845536609831e-05, + "loss": 1.1734, + "theoretical_loss": 3.3574418558374592, + "tokens_seen": 2905473024 + }, + { + "epoch": 0.88, + "learning_rate": 9.6802407221665e-05, + "loss": 1.1515, + "theoretical_loss": 3.357436044998621, + "tokens_seen": 2905538560 + }, + { + "epoch": 0.88, + "learning_rate": 9.67863590772317e-05, + "loss": 1.2814, + "theoretical_loss": 3.3574302343275457, + "tokens_seen": 2905604096 + }, + { + "epoch": 0.88, + "learning_rate": 9.67703109327984e-05, + "loss": 1.2783, + "theoretical_loss": 3.357424423824225, + "tokens_seen": 2905669632 + }, + { + "epoch": 0.88, + "learning_rate": 9.67542627883651e-05, + "loss": 1.3006, + "theoretical_loss": 3.3574186134886492, + "tokens_seen": 2905735168 + }, + { + "epoch": 0.88, + "learning_rate": 9.67382146439318e-05, + "loss": 1.3367, + "theoretical_loss": 3.357412803320811, + "tokens_seen": 2905800704 + }, + { + "epoch": 0.88, + "learning_rate": 9.67221664994985e-05, + "loss": 1.2783, + "theoretical_loss": 3.357406993320701, + "tokens_seen": 2905866240 + }, + { + "epoch": 0.88, + "learning_rate": 9.67061183550652e-05, + "loss": 1.2853, + "theoretical_loss": 3.3574011834883106, + "tokens_seen": 2905931776 + }, + { + "epoch": 0.88, + "learning_rate": 9.66900702106319e-05, + "loss": 1.2407, + "theoretical_loss": 3.3573953738236315, + "tokens_seen": 2905997312 + }, + { + "epoch": 0.88, + "learning_rate": 9.66740220661986e-05, + "loss": 1.3255, + "theoretical_loss": 3.3573895643266547, + "tokens_seen": 2906062848 + }, + { + "epoch": 0.88, + "learning_rate": 9.66579739217653e-05, + "loss": 1.2245, + "theoretical_loss": 3.3573837549973717, + "tokens_seen": 2906128384 + }, + { + "epoch": 0.88, + "learning_rate": 9.6641925777332e-05, + "loss": 1.3209, + "theoretical_loss": 3.357377945835774, + "tokens_seen": 2906193920 + }, + { + "epoch": 0.88, + "learning_rate": 9.66258776328987e-05, + "loss": 1.212, + "theoretical_loss": 3.357372136841853, + "tokens_seen": 2906259456 + }, + { + "epoch": 0.88, + "learning_rate": 9.660982948846539e-05, + "loss": 1.2173, + "theoretical_loss": 3.3573663280156, + "tokens_seen": 2906324992 + }, + { + "epoch": 0.88, + "learning_rate": 9.659378134403209e-05, + "loss": 1.3477, + "theoretical_loss": 3.357360519357006, + "tokens_seen": 2906390528 + }, + { + "epoch": 0.88, + "learning_rate": 9.657773319959879e-05, + "loss": 1.334, + "theoretical_loss": 3.3573547108660633, + "tokens_seen": 2906456064 + }, + { + "epoch": 0.88, + "objective/train/docs_used": 1056686, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.1998437643051147, + "objective/train/theoretical_loss": 3.3573489025427623, + "objective/train/tokens_used": 2926981600, + "theoretical_loss": 3.3573489025427623, + "tokens_seen": 2906521600 + }, + { + "epoch": 0.88, + "learning_rate": 9.65616850551655e-05, + "loss": 1.226, + "theoretical_loss": 3.3573489025427623, + "tokens_seen": 2906521600 + }, + { + "epoch": 0.88, + "learning_rate": 9.654563691073221e-05, + "loss": 1.2055, + "theoretical_loss": 3.357343094387095, + "tokens_seen": 2906587136 + }, + { + "epoch": 0.88, + "learning_rate": 9.652958876629891e-05, + "loss": 1.2069, + "theoretical_loss": 3.357337286399053, + "tokens_seen": 2906652672 + }, + { + "epoch": 0.88, + "learning_rate": 9.651354062186561e-05, + "loss": 1.3292, + "theoretical_loss": 3.357331478578627, + "tokens_seen": 2906718208 + }, + { + "epoch": 0.88, + "learning_rate": 9.649749247743231e-05, + "loss": 1.2654, + "theoretical_loss": 3.3573256709258086, + "tokens_seen": 2906783744 + }, + { + "epoch": 0.88, + "learning_rate": 9.648144433299901e-05, + "loss": 1.3614, + "theoretical_loss": 3.3573198634405887, + "tokens_seen": 2906849280 + }, + { + "epoch": 0.88, + "learning_rate": 9.64653961885657e-05, + "loss": 1.3582, + "theoretical_loss": 3.35731405612296, + "tokens_seen": 2906914816 + }, + { + "epoch": 0.88, + "learning_rate": 9.64493480441324e-05, + "loss": 1.2849, + "theoretical_loss": 3.357308248972913, + "tokens_seen": 2906980352 + }, + { + "epoch": 0.88, + "learning_rate": 9.64332998996991e-05, + "loss": 1.1861, + "theoretical_loss": 3.357302441990439, + "tokens_seen": 2907045888 + }, + { + "epoch": 0.88, + "learning_rate": 9.64172517552658e-05, + "loss": 1.2598, + "theoretical_loss": 3.35729663517553, + "tokens_seen": 2907111424 + }, + { + "epoch": 0.88, + "learning_rate": 9.64012036108325e-05, + "loss": 1.2256, + "theoretical_loss": 3.3572908285281766, + "tokens_seen": 2907176960 + }, + { + "epoch": 0.88, + "learning_rate": 9.63851554663992e-05, + "loss": 1.3822, + "theoretical_loss": 3.3572850220483703, + "tokens_seen": 2907242496 + }, + { + "epoch": 0.88, + "learning_rate": 9.63691073219659e-05, + "loss": 1.2011, + "theoretical_loss": 3.3572792157361033, + "tokens_seen": 2907308032 + }, + { + "epoch": 0.88, + "learning_rate": 9.63530591775326e-05, + "loss": 1.2271, + "theoretical_loss": 3.3572734095913663, + "tokens_seen": 2907373568 + }, + { + "epoch": 0.88, + "learning_rate": 9.63370110330993e-05, + "loss": 1.2802, + "theoretical_loss": 3.357267603614151, + "tokens_seen": 2907439104 + }, + { + "epoch": 0.88, + "learning_rate": 9.6320962888666e-05, + "loss": 1.2184, + "theoretical_loss": 3.357261797804448, + "tokens_seen": 2907504640 + }, + { + "epoch": 0.88, + "learning_rate": 9.63049147442327e-05, + "loss": 1.3153, + "theoretical_loss": 3.35725599216225, + "tokens_seen": 2907570176 + }, + { + "epoch": 0.88, + "learning_rate": 9.62888665997994e-05, + "loss": 1.3236, + "theoretical_loss": 3.3572501866875477, + "tokens_seen": 2907635712 + }, + { + "epoch": 0.88, + "learning_rate": 9.627281845536609e-05, + "loss": 1.2461, + "theoretical_loss": 3.357244381380332, + "tokens_seen": 2907701248 + }, + { + "epoch": 0.88, + "learning_rate": 9.625677031093279e-05, + "loss": 1.2015, + "theoretical_loss": 3.357238576240595, + "tokens_seen": 2907766784 + }, + { + "epoch": 0.88, + "learning_rate": 9.62407221664995e-05, + "loss": 1.305, + "theoretical_loss": 3.357232771268328, + "tokens_seen": 2907832320 + }, + { + "epoch": 0.88, + "learning_rate": 9.622467402206621e-05, + "loss": 1.3022, + "theoretical_loss": 3.3572269664635224, + "tokens_seen": 2907897856 + }, + { + "epoch": 0.88, + "learning_rate": 9.620862587763291e-05, + "loss": 1.1662, + "theoretical_loss": 3.3572211618261694, + "tokens_seen": 2907963392 + }, + { + "epoch": 0.88, + "learning_rate": 9.619257773319961e-05, + "loss": 1.163, + "theoretical_loss": 3.35721535735626, + "tokens_seen": 2908028928 + }, + { + "epoch": 0.88, + "learning_rate": 9.617652958876631e-05, + "loss": 1.2889, + "theoretical_loss": 3.357209553053787, + "tokens_seen": 2908094464 + }, + { + "epoch": 0.88, + "learning_rate": 9.616048144433301e-05, + "loss": 1.3064, + "theoretical_loss": 3.35720374891874, + "tokens_seen": 2908160000 + }, + { + "epoch": 0.88, + "learning_rate": 9.614443329989971e-05, + "loss": 1.313, + "theoretical_loss": 3.357197944951112, + "tokens_seen": 2908225536 + }, + { + "epoch": 0.88, + "learning_rate": 9.612838515546641e-05, + "loss": 1.33, + "theoretical_loss": 3.357192141150893, + "tokens_seen": 2908291072 + }, + { + "epoch": 0.88, + "learning_rate": 9.61123370110331e-05, + "loss": 1.1845, + "theoretical_loss": 3.3571863375180753, + "tokens_seen": 2908356608 + }, + { + "epoch": 0.88, + "learning_rate": 9.60962888665998e-05, + "loss": 1.2297, + "theoretical_loss": 3.35718053405265, + "tokens_seen": 2908422144 + }, + { + "epoch": 0.88, + "learning_rate": 9.60802407221665e-05, + "loss": 1.2205, + "theoretical_loss": 3.357174730754609, + "tokens_seen": 2908487680 + }, + { + "epoch": 0.88, + "learning_rate": 9.60641925777332e-05, + "loss": 1.2874, + "theoretical_loss": 3.3571689276239427, + "tokens_seen": 2908553216 + }, + { + "epoch": 0.88, + "learning_rate": 9.60481444332999e-05, + "loss": 1.3253, + "theoretical_loss": 3.3571631246606435, + "tokens_seen": 2908618752 + }, + { + "epoch": 0.88, + "learning_rate": 9.60320962888666e-05, + "loss": 1.224, + "theoretical_loss": 3.357157321864702, + "tokens_seen": 2908684288 + }, + { + "epoch": 0.88, + "learning_rate": 9.60160481444333e-05, + "loss": 1.1571, + "theoretical_loss": 3.35715151923611, + "tokens_seen": 2908749824 + }, + { + "epoch": 0.88, + "learning_rate": 9.6e-05, + "loss": 1.2487, + "theoretical_loss": 3.3571457167748595, + "tokens_seen": 2908815360 + }, + { + "epoch": 0.88, + "learning_rate": 9.59839518555667e-05, + "loss": 1.1665, + "theoretical_loss": 3.357139914480941, + "tokens_seen": 2908880896 + }, + { + "epoch": 0.88, + "learning_rate": 9.59679037111334e-05, + "loss": 1.3301, + "theoretical_loss": 3.3571341123543457, + "tokens_seen": 2908946432 + }, + { + "epoch": 0.88, + "learning_rate": 9.59518555667001e-05, + "loss": 1.3468, + "theoretical_loss": 3.357128310395066, + "tokens_seen": 2909011968 + }, + { + "epoch": 0.88, + "learning_rate": 9.59358074222668e-05, + "loss": 1.2733, + "theoretical_loss": 3.3571225086030925, + "tokens_seen": 2909077504 + }, + { + "epoch": 0.88, + "learning_rate": 9.59197592778335e-05, + "loss": 1.351, + "theoretical_loss": 3.357116706978417, + "tokens_seen": 2909143040 + }, + { + "epoch": 0.88, + "learning_rate": 9.59037111334002e-05, + "loss": 1.2337, + "theoretical_loss": 3.357110905521031, + "tokens_seen": 2909208576 + }, + { + "epoch": 0.88, + "learning_rate": 9.588766298896691e-05, + "loss": 1.3086, + "theoretical_loss": 3.3571051042309255, + "tokens_seen": 2909274112 + }, + { + "epoch": 0.88, + "learning_rate": 9.587161484453361e-05, + "loss": 1.3166, + "theoretical_loss": 3.3570993031080922, + "tokens_seen": 2909339648 + }, + { + "epoch": 0.88, + "learning_rate": 9.585556670010031e-05, + "loss": 1.3181, + "theoretical_loss": 3.357093502152522, + "tokens_seen": 2909405184 + }, + { + "epoch": 0.88, + "learning_rate": 9.583951855566701e-05, + "loss": 1.2664, + "theoretical_loss": 3.3570877013642075, + "tokens_seen": 2909470720 + }, + { + "epoch": 0.88, + "learning_rate": 9.582347041123371e-05, + "loss": 1.3277, + "theoretical_loss": 3.3570819007431387, + "tokens_seen": 2909536256 + }, + { + "epoch": 0.88, + "learning_rate": 9.580742226680041e-05, + "loss": 1.16, + "theoretical_loss": 3.357076100289308, + "tokens_seen": 2909601792 + }, + { + "epoch": 0.88, + "learning_rate": 9.579137412236711e-05, + "loss": 1.2845, + "theoretical_loss": 3.3570703000027065, + "tokens_seen": 2909667328 + }, + { + "epoch": 0.88, + "learning_rate": 9.577532597793381e-05, + "loss": 1.2365, + "theoretical_loss": 3.3570644998833257, + "tokens_seen": 2909732864 + }, + { + "epoch": 0.88, + "objective/train/docs_used": 1057798, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.2415955066680908, + "objective/train/theoretical_loss": 3.357058699931157, + "objective/train/tokens_used": 2930258400, + "theoretical_loss": 3.357058699931157, + "tokens_seen": 2909798400 + }, + { + "epoch": 0.88, + "learning_rate": 9.57592778335005e-05, + "loss": 1.2497, + "theoretical_loss": 3.357058699931157, + "tokens_seen": 2909798400 + }, + { + "epoch": 0.88, + "learning_rate": 9.57432296890672e-05, + "loss": 1.1567, + "theoretical_loss": 3.3570529001461913, + "tokens_seen": 2909863936 + }, + { + "epoch": 0.88, + "learning_rate": 9.57271815446339e-05, + "loss": 1.3183, + "theoretical_loss": 3.357047100528421, + "tokens_seen": 2909929472 + }, + { + "epoch": 0.88, + "learning_rate": 9.57111334002006e-05, + "loss": 1.3059, + "theoretical_loss": 3.357041301077836, + "tokens_seen": 2909995008 + }, + { + "epoch": 0.88, + "learning_rate": 9.56950852557673e-05, + "loss": 1.2643, + "theoretical_loss": 3.3570355017944293, + "tokens_seen": 2910060544 + }, + { + "epoch": 0.88, + "learning_rate": 9.5679037111334e-05, + "loss": 1.3115, + "theoretical_loss": 3.357029702678192, + "tokens_seen": 2910126080 + }, + { + "epoch": 0.88, + "learning_rate": 9.56629889669007e-05, + "loss": 1.2729, + "theoretical_loss": 3.3570239037291145, + "tokens_seen": 2910191616 + }, + { + "epoch": 0.88, + "learning_rate": 9.56469408224674e-05, + "loss": 1.3131, + "theoretical_loss": 3.357018104947189, + "tokens_seen": 2910257152 + }, + { + "epoch": 0.88, + "learning_rate": 9.563089267803412e-05, + "loss": 1.3354, + "theoretical_loss": 3.357012306332407, + "tokens_seen": 2910322688 + }, + { + "epoch": 0.88, + "learning_rate": 9.561484453360082e-05, + "loss": 1.2695, + "theoretical_loss": 3.35700650788476, + "tokens_seen": 2910388224 + }, + { + "epoch": 0.88, + "learning_rate": 9.55987963891675e-05, + "loss": 1.3302, + "theoretical_loss": 3.3570007096042387, + "tokens_seen": 2910453760 + }, + { + "epoch": 0.88, + "learning_rate": 9.55827482447342e-05, + "loss": 1.29, + "theoretical_loss": 3.3569949114908355, + "tokens_seen": 2910519296 + }, + { + "epoch": 0.88, + "learning_rate": 9.55667001003009e-05, + "loss": 1.3522, + "theoretical_loss": 3.3569891135445413, + "tokens_seen": 2910584832 + }, + { + "epoch": 0.88, + "learning_rate": 9.55506519558676e-05, + "loss": 1.3088, + "theoretical_loss": 3.356983315765347, + "tokens_seen": 2910650368 + }, + { + "epoch": 0.88, + "learning_rate": 9.553460381143431e-05, + "loss": 1.2267, + "theoretical_loss": 3.356977518153245, + "tokens_seen": 2910715904 + }, + { + "epoch": 0.88, + "learning_rate": 9.551855566700101e-05, + "loss": 1.3381, + "theoretical_loss": 3.356971720708226, + "tokens_seen": 2910781440 + }, + { + "epoch": 0.88, + "learning_rate": 9.550250752256771e-05, + "loss": 1.296, + "theoretical_loss": 3.3569659234302818, + "tokens_seen": 2910846976 + }, + { + "epoch": 0.88, + "learning_rate": 9.548645937813441e-05, + "loss": 1.2008, + "theoretical_loss": 3.3569601263194038, + "tokens_seen": 2910912512 + }, + { + "epoch": 0.88, + "learning_rate": 9.547041123370111e-05, + "loss": 1.2352, + "theoretical_loss": 3.3569543293755832, + "tokens_seen": 2910978048 + }, + { + "epoch": 0.88, + "learning_rate": 9.545436308926781e-05, + "loss": 1.2622, + "theoretical_loss": 3.3569485325988113, + "tokens_seen": 2911043584 + }, + { + "epoch": 0.88, + "learning_rate": 9.543831494483451e-05, + "loss": 1.3388, + "theoretical_loss": 3.3569427359890804, + "tokens_seen": 2911109120 + }, + { + "epoch": 0.88, + "learning_rate": 9.542226680040121e-05, + "loss": 1.2843, + "theoretical_loss": 3.3569369395463813, + "tokens_seen": 2911174656 + }, + { + "epoch": 0.88, + "learning_rate": 9.54062186559679e-05, + "loss": 1.3463, + "theoretical_loss": 3.356931143270705, + "tokens_seen": 2911240192 + }, + { + "epoch": 0.88, + "learning_rate": 9.53901705115346e-05, + "loss": 1.1839, + "theoretical_loss": 3.3569253471620435, + "tokens_seen": 2911305728 + }, + { + "epoch": 0.88, + "learning_rate": 9.53741223671013e-05, + "loss": 1.2734, + "theoretical_loss": 3.356919551220388, + "tokens_seen": 2911371264 + }, + { + "epoch": 0.88, + "learning_rate": 9.5358074222668e-05, + "loss": 1.266, + "theoretical_loss": 3.3569137554457305, + "tokens_seen": 2911436800 + }, + { + "epoch": 0.88, + "learning_rate": 9.534202607823471e-05, + "loss": 1.2652, + "theoretical_loss": 3.3569079598380616, + "tokens_seen": 2911502336 + }, + { + "epoch": 0.88, + "learning_rate": 9.532597793380142e-05, + "loss": 1.3392, + "theoretical_loss": 3.3569021643973733, + "tokens_seen": 2911567872 + }, + { + "epoch": 0.88, + "learning_rate": 9.530992978936812e-05, + "loss": 1.2763, + "theoretical_loss": 3.356896369123657, + "tokens_seen": 2911633408 + }, + { + "epoch": 0.88, + "learning_rate": 9.529388164493482e-05, + "loss": 1.2653, + "theoretical_loss": 3.3568905740169033, + "tokens_seen": 2911698944 + }, + { + "epoch": 0.88, + "learning_rate": 9.527783350050152e-05, + "loss": 1.2697, + "theoretical_loss": 3.3568847790771046, + "tokens_seen": 2911764480 + }, + { + "epoch": 0.88, + "learning_rate": 9.526178535606822e-05, + "loss": 1.3198, + "theoretical_loss": 3.3568789843042524, + "tokens_seen": 2911830016 + }, + { + "epoch": 0.88, + "learning_rate": 9.52457372116349e-05, + "loss": 1.2222, + "theoretical_loss": 3.3568731896983373, + "tokens_seen": 2911895552 + }, + { + "epoch": 0.88, + "learning_rate": 9.52296890672016e-05, + "loss": 1.2673, + "theoretical_loss": 3.3568673952593513, + "tokens_seen": 2911961088 + }, + { + "epoch": 0.88, + "learning_rate": 9.52136409227683e-05, + "loss": 1.2962, + "theoretical_loss": 3.3568616009872856, + "tokens_seen": 2912026624 + }, + { + "epoch": 0.88, + "learning_rate": 9.519759277833501e-05, + "loss": 1.191, + "theoretical_loss": 3.356855806882132, + "tokens_seen": 2912092160 + }, + { + "epoch": 0.88, + "learning_rate": 9.518154463390171e-05, + "loss": 1.1733, + "theoretical_loss": 3.3568500129438816, + "tokens_seen": 2912157696 + }, + { + "epoch": 0.88, + "learning_rate": 9.516549648946841e-05, + "loss": 1.1986, + "theoretical_loss": 3.3568442191725256, + "tokens_seen": 2912223232 + }, + { + "epoch": 0.88, + "learning_rate": 9.514944834503511e-05, + "loss": 1.269, + "theoretical_loss": 3.356838425568056, + "tokens_seen": 2912288768 + }, + { + "epoch": 0.88, + "learning_rate": 9.513340020060181e-05, + "loss": 1.2687, + "theoretical_loss": 3.3568326321304642, + "tokens_seen": 2912354304 + }, + { + "epoch": 0.88, + "learning_rate": 9.511735205616851e-05, + "loss": 1.2724, + "theoretical_loss": 3.3568268388597415, + "tokens_seen": 2912419840 + }, + { + "epoch": 0.88, + "learning_rate": 9.510130391173521e-05, + "loss": 1.3627, + "theoretical_loss": 3.356821045755879, + "tokens_seen": 2912485376 + }, + { + "epoch": 0.88, + "learning_rate": 9.508525576730191e-05, + "loss": 1.3515, + "theoretical_loss": 3.3568152528188686, + "tokens_seen": 2912550912 + }, + { + "epoch": 0.88, + "learning_rate": 9.506920762286861e-05, + "loss": 1.3533, + "theoretical_loss": 3.3568094600487015, + "tokens_seen": 2912616448 + }, + { + "epoch": 0.88, + "learning_rate": 9.50531594784353e-05, + "loss": 1.1917, + "theoretical_loss": 3.356803667445369, + "tokens_seen": 2912681984 + }, + { + "epoch": 0.88, + "learning_rate": 9.503711133400201e-05, + "loss": 1.2608, + "theoretical_loss": 3.356797875008863, + "tokens_seen": 2912747520 + }, + { + "epoch": 0.88, + "learning_rate": 9.502106318956871e-05, + "loss": 1.2229, + "theoretical_loss": 3.3567920827391746, + "tokens_seen": 2912813056 + }, + { + "epoch": 0.88, + "learning_rate": 9.500501504513541e-05, + "loss": 1.2762, + "theoretical_loss": 3.3567862906362955, + "tokens_seen": 2912878592 + }, + { + "epoch": 0.88, + "learning_rate": 9.498896690070212e-05, + "loss": 1.2693, + "theoretical_loss": 3.3567804987002168, + "tokens_seen": 2912944128 + }, + { + "epoch": 0.88, + "learning_rate": 9.497291875626882e-05, + "loss": 1.2361, + "theoretical_loss": 3.35677470693093, + "tokens_seen": 2913009664 + }, + { + "epoch": 0.88, + "objective/train/docs_used": 1058971, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.2594022750854492, + "objective/train/theoretical_loss": 3.3567689153284266, + "objective/train/tokens_used": 2933535200, + "theoretical_loss": 3.3567689153284266, + "tokens_seen": 2913075200 + }, + { + "epoch": 0.88, + "learning_rate": 9.495687061183552e-05, + "loss": 1.2503, + "theoretical_loss": 3.3567689153284266, + "tokens_seen": 2913075200 + }, + { + "epoch": 0.88, + "learning_rate": 9.494082246740222e-05, + "loss": 1.3288, + "theoretical_loss": 3.3567631238926983, + "tokens_seen": 2913140736 + }, + { + "epoch": 0.88, + "learning_rate": 9.492477432296892e-05, + "loss": 1.2672, + "theoretical_loss": 3.3567573326237365, + "tokens_seen": 2913206272 + }, + { + "epoch": 0.88, + "learning_rate": 9.490872617853562e-05, + "loss": 1.2684, + "theoretical_loss": 3.3567515415215325, + "tokens_seen": 2913271808 + }, + { + "epoch": 0.88, + "learning_rate": 9.48926780341023e-05, + "loss": 1.2597, + "theoretical_loss": 3.3567457505860774, + "tokens_seen": 2913337344 + }, + { + "epoch": 0.88, + "learning_rate": 9.4876629889669e-05, + "loss": 1.3587, + "theoretical_loss": 3.356739959817363, + "tokens_seen": 2913402880 + }, + { + "epoch": 0.88, + "learning_rate": 9.486058174523571e-05, + "loss": 1.3188, + "theoretical_loss": 3.356734169215381, + "tokens_seen": 2913468416 + }, + { + "epoch": 0.88, + "learning_rate": 9.484453360080241e-05, + "loss": 1.3605, + "theoretical_loss": 3.3567283787801223, + "tokens_seen": 2913533952 + }, + { + "epoch": 0.88, + "learning_rate": 9.484453360080241e-05, + "loss": 1.3105, + "theoretical_loss": 3.356722588511579, + "tokens_seen": 2913599488 + }, + { + "epoch": 0.88, + "learning_rate": 9.482848545636911e-05, + "loss": 1.2503, + "theoretical_loss": 3.356716798409742, + "tokens_seen": 2913665024 + }, + { + "epoch": 0.88, + "learning_rate": 9.481243731193581e-05, + "loss": 1.2741, + "theoretical_loss": 3.356711008474603, + "tokens_seen": 2913730560 + }, + { + "epoch": 0.88, + "learning_rate": 9.479638916750251e-05, + "loss": 1.2684, + "theoretical_loss": 3.3567052187061535, + "tokens_seen": 2913796096 + }, + { + "epoch": 0.88, + "learning_rate": 9.478034102306921e-05, + "loss": 1.2271, + "theoretical_loss": 3.3566994291043843, + "tokens_seen": 2913861632 + }, + { + "epoch": 0.88, + "learning_rate": 9.476429287863591e-05, + "loss": 1.1974, + "theoretical_loss": 3.356693639669288, + "tokens_seen": 2913927168 + }, + { + "epoch": 0.88, + "learning_rate": 9.474824473420261e-05, + "loss": 1.2876, + "theoretical_loss": 3.356687850400855, + "tokens_seen": 2913992704 + }, + { + "epoch": 0.88, + "learning_rate": 9.473219658976931e-05, + "loss": 1.2394, + "theoretical_loss": 3.3566820612990775, + "tokens_seen": 2914058240 + }, + { + "epoch": 0.88, + "learning_rate": 9.471614844533601e-05, + "loss": 1.2634, + "theoretical_loss": 3.3566762723639467, + "tokens_seen": 2914123776 + }, + { + "epoch": 0.88, + "learning_rate": 9.470010030090271e-05, + "loss": 1.2304, + "theoretical_loss": 3.3566704835954537, + "tokens_seen": 2914189312 + }, + { + "epoch": 0.88, + "learning_rate": 9.468405215646941e-05, + "loss": 1.2811, + "theoretical_loss": 3.3566646949935905, + "tokens_seen": 2914254848 + }, + { + "epoch": 0.88, + "learning_rate": 9.466800401203611e-05, + "loss": 1.2477, + "theoretical_loss": 3.3566589065583483, + "tokens_seen": 2914320384 + }, + { + "epoch": 0.88, + "learning_rate": 9.465195586760282e-05, + "loss": 1.2774, + "theoretical_loss": 3.3566531182897186, + "tokens_seen": 2914385920 + }, + { + "epoch": 0.88, + "learning_rate": 9.463590772316952e-05, + "loss": 1.4371, + "theoretical_loss": 3.356647330187693, + "tokens_seen": 2914451456 + }, + { + "epoch": 0.88, + "learning_rate": 9.461985957873622e-05, + "loss": 1.2979, + "theoretical_loss": 3.3566415422522624, + "tokens_seen": 2914516992 + }, + { + "epoch": 0.88, + "learning_rate": 9.460381143430292e-05, + "loss": 1.3089, + "theoretical_loss": 3.356635754483419, + "tokens_seen": 2914582528 + }, + { + "epoch": 0.88, + "learning_rate": 9.458776328986962e-05, + "loss": 1.3662, + "theoretical_loss": 3.3566299668811537, + "tokens_seen": 2914648064 + }, + { + "epoch": 0.88, + "learning_rate": 9.457171514543632e-05, + "loss": 1.3502, + "theoretical_loss": 3.3566241794454585, + "tokens_seen": 2914713600 + }, + { + "epoch": 0.88, + "learning_rate": 9.455566700100302e-05, + "loss": 1.1923, + "theoretical_loss": 3.356618392176324, + "tokens_seen": 2914779136 + }, + { + "epoch": 0.88, + "learning_rate": 9.45396188565697e-05, + "loss": 1.2169, + "theoretical_loss": 3.3566126050737424, + "tokens_seen": 2914844672 + }, + { + "epoch": 0.88, + "learning_rate": 9.45235707121364e-05, + "loss": 1.2149, + "theoretical_loss": 3.3566068181377053, + "tokens_seen": 2914910208 + }, + { + "epoch": 0.88, + "learning_rate": 9.450752256770311e-05, + "loss": 1.3174, + "theoretical_loss": 3.3566010313682035, + "tokens_seen": 2914975744 + }, + { + "epoch": 0.88, + "learning_rate": 9.449147442326981e-05, + "loss": 1.3646, + "theoretical_loss": 3.356595244765229, + "tokens_seen": 2915041280 + }, + { + "epoch": 0.88, + "learning_rate": 9.447542627883651e-05, + "loss": 1.1875, + "theoretical_loss": 3.3565894583287728, + "tokens_seen": 2915106816 + }, + { + "epoch": 0.88, + "learning_rate": 9.445937813440321e-05, + "loss": 1.276, + "theoretical_loss": 3.356583672058827, + "tokens_seen": 2915172352 + }, + { + "epoch": 0.88, + "learning_rate": 9.444332998996992e-05, + "loss": 1.29, + "theoretical_loss": 3.356577885955382, + "tokens_seen": 2915237888 + }, + { + "epoch": 0.88, + "learning_rate": 9.442728184553662e-05, + "loss": 1.271, + "theoretical_loss": 3.3565721000184308, + "tokens_seen": 2915303424 + }, + { + "epoch": 0.88, + "learning_rate": 9.441123370110332e-05, + "loss": 1.3226, + "theoretical_loss": 3.3565663142479636, + "tokens_seen": 2915368960 + }, + { + "epoch": 0.88, + "learning_rate": 9.439518555667001e-05, + "loss": 1.3196, + "theoretical_loss": 3.356560528643972, + "tokens_seen": 2915434496 + }, + { + "epoch": 0.88, + "learning_rate": 9.437913741223671e-05, + "loss": 1.1826, + "theoretical_loss": 3.3565547432064484, + "tokens_seen": 2915500032 + }, + { + "epoch": 0.88, + "learning_rate": 9.436308926780341e-05, + "loss": 1.358, + "theoretical_loss": 3.356548957935383, + "tokens_seen": 2915565568 + }, + { + "epoch": 0.88, + "learning_rate": 9.434704112337011e-05, + "loss": 1.2994, + "theoretical_loss": 3.356543172830768, + "tokens_seen": 2915631104 + }, + { + "epoch": 0.88, + "learning_rate": 9.433099297893681e-05, + "loss": 1.2612, + "theoretical_loss": 3.356537387892595, + "tokens_seen": 2915696640 + }, + { + "epoch": 0.88, + "learning_rate": 9.431494483450351e-05, + "loss": 1.3114, + "theoretical_loss": 3.356531603120855, + "tokens_seen": 2915762176 + }, + { + "epoch": 0.88, + "learning_rate": 9.429889669007022e-05, + "loss": 1.3135, + "theoretical_loss": 3.35652581851554, + "tokens_seen": 2915827712 + }, + { + "epoch": 0.88, + "learning_rate": 9.428284854563692e-05, + "loss": 1.249, + "theoretical_loss": 3.356520034076641, + "tokens_seen": 2915893248 + }, + { + "epoch": 0.88, + "learning_rate": 9.426680040120362e-05, + "loss": 1.3455, + "theoretical_loss": 3.3565142498041496, + "tokens_seen": 2915958784 + }, + { + "epoch": 0.88, + "learning_rate": 9.425075225677032e-05, + "loss": 1.262, + "theoretical_loss": 3.3565084656980573, + "tokens_seen": 2916024320 + }, + { + "epoch": 0.88, + "learning_rate": 9.423470411233702e-05, + "loss": 1.1678, + "theoretical_loss": 3.356502681758356, + "tokens_seen": 2916089856 + }, + { + "epoch": 0.88, + "learning_rate": 9.421865596790372e-05, + "loss": 1.3436, + "theoretical_loss": 3.356496897985036, + "tokens_seen": 2916155392 + }, + { + "epoch": 0.88, + "learning_rate": 9.42026078234704e-05, + "loss": 1.298, + "theoretical_loss": 3.3564911143780902, + "tokens_seen": 2916220928 + }, + { + "epoch": 0.88, + "learning_rate": 9.41865596790371e-05, + "loss": 1.3395, + "theoretical_loss": 3.356485330937509, + "tokens_seen": 2916286464 + }, + { + "epoch": 0.88, + "objective/train/docs_used": 1060149, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.193345546722412, + "objective/train/theoretical_loss": 3.3564795476632847, + "objective/train/tokens_used": 2936812000, + "theoretical_loss": 3.3564795476632847, + "tokens_seen": 2916352000 + }, + { + "epoch": 0.88, + "learning_rate": 9.417051153460381e-05, + "loss": 1.2015, + "theoretical_loss": 3.3564795476632847, + "tokens_seen": 2916352000 + }, + { + "epoch": 0.88, + "learning_rate": 9.415446339017051e-05, + "loss": 1.2279, + "theoretical_loss": 3.356473764555408, + "tokens_seen": 2916417536 + }, + { + "epoch": 0.88, + "learning_rate": 9.413841524573722e-05, + "loss": 1.3295, + "theoretical_loss": 3.356467981613871, + "tokens_seen": 2916483072 + }, + { + "epoch": 0.88, + "learning_rate": 9.412236710130392e-05, + "loss": 1.2345, + "theoretical_loss": 3.3564621988386647, + "tokens_seen": 2916548608 + }, + { + "epoch": 0.88, + "learning_rate": 9.410631895687062e-05, + "loss": 1.3469, + "theoretical_loss": 3.356456416229781, + "tokens_seen": 2916614144 + }, + { + "epoch": 0.88, + "learning_rate": 9.409027081243732e-05, + "loss": 1.2507, + "theoretical_loss": 3.356450633787211, + "tokens_seen": 2916679680 + }, + { + "epoch": 0.88, + "learning_rate": 9.407422266800402e-05, + "loss": 1.217, + "theoretical_loss": 3.3564448515109464, + "tokens_seen": 2916745216 + }, + { + "epoch": 0.88, + "learning_rate": 9.405817452357073e-05, + "loss": 1.3007, + "theoretical_loss": 3.3564390694009787, + "tokens_seen": 2916810752 + }, + { + "epoch": 0.88, + "learning_rate": 9.404212637913741e-05, + "loss": 1.2169, + "theoretical_loss": 3.3564332874572993, + "tokens_seen": 2916876288 + }, + { + "epoch": 0.88, + "learning_rate": 9.402607823470411e-05, + "loss": 1.2855, + "theoretical_loss": 3.3564275056799, + "tokens_seen": 2916941824 + }, + { + "epoch": 0.88, + "learning_rate": 9.401003009027081e-05, + "loss": 1.2437, + "theoretical_loss": 3.3564217240687713, + "tokens_seen": 2917007360 + }, + { + "epoch": 0.88, + "learning_rate": 9.399398194583751e-05, + "loss": 1.2651, + "theoretical_loss": 3.3564159426239057, + "tokens_seen": 2917072896 + }, + { + "epoch": 0.88, + "learning_rate": 9.397793380140421e-05, + "loss": 1.3257, + "theoretical_loss": 3.3564101613452944, + "tokens_seen": 2917138432 + }, + { + "epoch": 0.88, + "learning_rate": 9.396188565697092e-05, + "loss": 1.3107, + "theoretical_loss": 3.356404380232929, + "tokens_seen": 2917203968 + }, + { + "epoch": 0.88, + "learning_rate": 9.394583751253762e-05, + "loss": 1.3173, + "theoretical_loss": 3.3563985992868006, + "tokens_seen": 2917269504 + }, + { + "epoch": 0.88, + "learning_rate": 9.392978936810432e-05, + "loss": 1.2511, + "theoretical_loss": 3.356392818506901, + "tokens_seen": 2917335040 + }, + { + "epoch": 0.88, + "learning_rate": 9.391374122367102e-05, + "loss": 1.2544, + "theoretical_loss": 3.3563870378932217, + "tokens_seen": 2917400576 + }, + { + "epoch": 0.88, + "learning_rate": 9.391374122367102e-05, + "loss": 1.2221, + "theoretical_loss": 3.3563812574457534, + "tokens_seen": 2917466112 + }, + { + "epoch": 0.88, + "learning_rate": 9.389769307923772e-05, + "loss": 1.2915, + "theoretical_loss": 3.356375477164489, + "tokens_seen": 2917531648 + }, + { + "epoch": 0.88, + "learning_rate": 9.388164493480442e-05, + "loss": 1.2416, + "theoretical_loss": 3.356369697049419, + "tokens_seen": 2917597184 + }, + { + "epoch": 0.88, + "learning_rate": 9.386559679037112e-05, + "loss": 1.2198, + "theoretical_loss": 3.3563639171005355, + "tokens_seen": 2917662720 + }, + { + "epoch": 0.88, + "learning_rate": 9.38495486459378e-05, + "loss": 1.2654, + "theoretical_loss": 3.3563581373178293, + "tokens_seen": 2917728256 + }, + { + "epoch": 0.88, + "learning_rate": 9.383350050150452e-05, + "loss": 1.3393, + "theoretical_loss": 3.3563523577012924, + "tokens_seen": 2917793792 + }, + { + "epoch": 0.88, + "learning_rate": 9.381745235707122e-05, + "loss": 1.2092, + "theoretical_loss": 3.356346578250916, + "tokens_seen": 2917859328 + }, + { + "epoch": 0.88, + "learning_rate": 9.380140421263792e-05, + "loss": 1.2482, + "theoretical_loss": 3.3563407989666914, + "tokens_seen": 2917924864 + }, + { + "epoch": 0.88, + "learning_rate": 9.378535606820462e-05, + "loss": 1.2653, + "theoretical_loss": 3.356335019848611, + "tokens_seen": 2917990400 + }, + { + "epoch": 0.88, + "learning_rate": 9.376930792377132e-05, + "loss": 1.2569, + "theoretical_loss": 3.356329240896665, + "tokens_seen": 2918055936 + }, + { + "epoch": 0.88, + "learning_rate": 9.375325977933802e-05, + "loss": 1.2989, + "theoretical_loss": 3.356323462110846, + "tokens_seen": 2918121472 + }, + { + "epoch": 0.88, + "learning_rate": 9.373721163490472e-05, + "loss": 1.3931, + "theoretical_loss": 3.3563176834911452, + "tokens_seen": 2918187008 + }, + { + "epoch": 0.88, + "learning_rate": 9.372116349047142e-05, + "loss": 1.2127, + "theoretical_loss": 3.3563119050375536, + "tokens_seen": 2918252544 + }, + { + "epoch": 0.88, + "learning_rate": 9.370511534603813e-05, + "loss": 1.3078, + "theoretical_loss": 3.3563061267500633, + "tokens_seen": 2918318080 + }, + { + "epoch": 0.88, + "learning_rate": 9.368906720160481e-05, + "loss": 1.2923, + "theoretical_loss": 3.3563003486286656, + "tokens_seen": 2918383616 + }, + { + "epoch": 0.88, + "learning_rate": 9.367301905717151e-05, + "loss": 1.3292, + "theoretical_loss": 3.356294570673352, + "tokens_seen": 2918449152 + }, + { + "epoch": 0.88, + "learning_rate": 9.365697091273821e-05, + "loss": 1.2475, + "theoretical_loss": 3.3562887928841136, + "tokens_seen": 2918514688 + }, + { + "epoch": 0.88, + "learning_rate": 9.364092276830491e-05, + "loss": 1.2825, + "theoretical_loss": 3.356283015260942, + "tokens_seen": 2918580224 + }, + { + "epoch": 0.88, + "learning_rate": 9.362487462387162e-05, + "loss": 1.1836, + "theoretical_loss": 3.35627723780383, + "tokens_seen": 2918645760 + }, + { + "epoch": 0.88, + "learning_rate": 9.360882647943832e-05, + "loss": 1.2448, + "theoretical_loss": 3.3562714605127675, + "tokens_seen": 2918711296 + }, + { + "epoch": 0.88, + "learning_rate": 9.359277833500502e-05, + "loss": 1.2668, + "theoretical_loss": 3.3562656833877464, + "tokens_seen": 2918776832 + }, + { + "epoch": 0.88, + "learning_rate": 9.357673019057172e-05, + "loss": 1.185, + "theoretical_loss": 3.3562599064287584, + "tokens_seen": 2918842368 + }, + { + "epoch": 0.88, + "learning_rate": 9.356068204613842e-05, + "loss": 1.3097, + "theoretical_loss": 3.3562541296357953, + "tokens_seen": 2918907904 + }, + { + "epoch": 0.88, + "learning_rate": 9.354463390170512e-05, + "loss": 1.2877, + "theoretical_loss": 3.356248353008848, + "tokens_seen": 2918973440 + }, + { + "epoch": 0.88, + "learning_rate": 9.352858575727182e-05, + "loss": 1.3781, + "theoretical_loss": 3.356242576547908, + "tokens_seen": 2919038976 + }, + { + "epoch": 0.88, + "learning_rate": 9.351253761283852e-05, + "loss": 1.2783, + "theoretical_loss": 3.356236800252967, + "tokens_seen": 2919104512 + }, + { + "epoch": 0.88, + "learning_rate": 9.349648946840522e-05, + "loss": 1.3251, + "theoretical_loss": 3.3562310241240167, + "tokens_seen": 2919170048 + }, + { + "epoch": 0.88, + "learning_rate": 9.348044132397192e-05, + "loss": 1.3622, + "theoretical_loss": 3.356225248161049, + "tokens_seen": 2919235584 + }, + { + "epoch": 0.88, + "learning_rate": 9.346439317953862e-05, + "loss": 1.1828, + "theoretical_loss": 3.3562194723640544, + "tokens_seen": 2919301120 + }, + { + "epoch": 0.88, + "learning_rate": 9.344834503510532e-05, + "loss": 1.3105, + "theoretical_loss": 3.356213696733025, + "tokens_seen": 2919366656 + }, + { + "epoch": 0.88, + "learning_rate": 9.343229689067202e-05, + "loss": 1.2039, + "theoretical_loss": 3.356207921267952, + "tokens_seen": 2919432192 + }, + { + "epoch": 0.88, + "learning_rate": 9.341624874623872e-05, + "loss": 1.3645, + "theoretical_loss": 3.356202145968827, + "tokens_seen": 2919497728 + }, + { + "epoch": 0.88, + "learning_rate": 9.340020060180542e-05, + "loss": 1.3516, + "theoretical_loss": 3.356196370835642, + "tokens_seen": 2919563264 + }, + { + "epoch": 0.88, + "objective/train/docs_used": 1061216, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.2517013549804688, + "objective/train/theoretical_loss": 3.3561905958683877, + "objective/train/tokens_used": 2940088800, + "theoretical_loss": 3.3561905958683877, + "tokens_seen": 2919628800 + }, + { + "epoch": 0.88, + "learning_rate": 9.338415245737212e-05, + "loss": 1.2291, + "theoretical_loss": 3.3561905958683877, + "tokens_seen": 2919628800 + }, + { + "epoch": 0.88, + "learning_rate": 9.336810431293883e-05, + "loss": 1.2987, + "theoretical_loss": 3.356184821067056, + "tokens_seen": 2919694336 + }, + { + "epoch": 0.88, + "learning_rate": 9.335205616850553e-05, + "loss": 1.3585, + "theoretical_loss": 3.3561790464316386, + "tokens_seen": 2919759872 + }, + { + "epoch": 0.88, + "learning_rate": 9.333600802407221e-05, + "loss": 1.1874, + "theoretical_loss": 3.3561732719621267, + "tokens_seen": 2919825408 + }, + { + "epoch": 0.88, + "learning_rate": 9.331995987963891e-05, + "loss": 1.3614, + "theoretical_loss": 3.356167497658512, + "tokens_seen": 2919890944 + }, + { + "epoch": 0.88, + "learning_rate": 9.330391173520561e-05, + "loss": 1.2813, + "theoretical_loss": 3.356161723520786, + "tokens_seen": 2919956480 + }, + { + "epoch": 0.88, + "learning_rate": 9.328786359077231e-05, + "loss": 1.1652, + "theoretical_loss": 3.35615594954894, + "tokens_seen": 2920022016 + }, + { + "epoch": 0.88, + "learning_rate": 9.327181544633902e-05, + "loss": 1.2496, + "theoretical_loss": 3.356150175742966, + "tokens_seen": 2920087552 + }, + { + "epoch": 0.88, + "learning_rate": 9.325576730190572e-05, + "loss": 1.296, + "theoretical_loss": 3.3561444021028546, + "tokens_seen": 2920153088 + }, + { + "epoch": 0.88, + "learning_rate": 9.323971915747243e-05, + "loss": 1.3335, + "theoretical_loss": 3.3561386286285986, + "tokens_seen": 2920218624 + }, + { + "epoch": 0.88, + "learning_rate": 9.322367101303913e-05, + "loss": 1.2916, + "theoretical_loss": 3.356132855320188, + "tokens_seen": 2920284160 + }, + { + "epoch": 0.88, + "learning_rate": 9.320762286860583e-05, + "loss": 1.3398, + "theoretical_loss": 3.3561270821776157, + "tokens_seen": 2920349696 + }, + { + "epoch": 0.88, + "learning_rate": 9.319157472417253e-05, + "loss": 1.3521, + "theoretical_loss": 3.3561213092008724, + "tokens_seen": 2920415232 + }, + { + "epoch": 0.88, + "learning_rate": 9.317552657973922e-05, + "loss": 1.2411, + "theoretical_loss": 3.35611553638995, + "tokens_seen": 2920480768 + }, + { + "epoch": 0.89, + "learning_rate": 9.315947843530592e-05, + "loss": 1.2685, + "theoretical_loss": 3.35610976374484, + "tokens_seen": 2920546304 + }, + { + "epoch": 0.89, + "learning_rate": 9.314343029087262e-05, + "loss": 1.3056, + "theoretical_loss": 3.3561039912655333, + "tokens_seen": 2920611840 + }, + { + "epoch": 0.89, + "learning_rate": 9.312738214643932e-05, + "loss": 1.3475, + "theoretical_loss": 3.356098218952022, + "tokens_seen": 2920677376 + }, + { + "epoch": 0.89, + "learning_rate": 9.311133400200602e-05, + "loss": 1.427, + "theoretical_loss": 3.356092446804298, + "tokens_seen": 2920742912 + }, + { + "epoch": 0.89, + "learning_rate": 9.309528585757272e-05, + "loss": 1.3201, + "theoretical_loss": 3.3560866748223517, + "tokens_seen": 2920808448 + }, + { + "epoch": 0.89, + "learning_rate": 9.307923771313942e-05, + "loss": 1.3955, + "theoretical_loss": 3.3560809030061756, + "tokens_seen": 2920873984 + }, + { + "epoch": 0.89, + "learning_rate": 9.306318956870612e-05, + "loss": 1.1877, + "theoretical_loss": 3.3560751313557606, + "tokens_seen": 2920939520 + }, + { + "epoch": 0.89, + "learning_rate": 9.304714142427282e-05, + "loss": 1.245, + "theoretical_loss": 3.356069359871099, + "tokens_seen": 2921005056 + }, + { + "epoch": 0.89, + "learning_rate": 9.303109327983952e-05, + "loss": 1.2514, + "theoretical_loss": 3.3560635885521815, + "tokens_seen": 2921070592 + }, + { + "epoch": 0.89, + "learning_rate": 9.301504513540623e-05, + "loss": 1.2525, + "theoretical_loss": 3.356057817399, + "tokens_seen": 2921136128 + }, + { + "epoch": 0.89, + "learning_rate": 9.299899699097293e-05, + "loss": 1.2478, + "theoretical_loss": 3.3560520464115458, + "tokens_seen": 2921201664 + }, + { + "epoch": 0.89, + "learning_rate": 9.298294884653961e-05, + "loss": 1.1946, + "theoretical_loss": 3.3560462755898106, + "tokens_seen": 2921267200 + }, + { + "epoch": 0.89, + "learning_rate": 9.296690070210631e-05, + "loss": 1.3142, + "theoretical_loss": 3.3560405049337865, + "tokens_seen": 2921332736 + }, + { + "epoch": 0.89, + "learning_rate": 9.295085255767301e-05, + "loss": 1.256, + "theoretical_loss": 3.3560347344434636, + "tokens_seen": 2921398272 + }, + { + "epoch": 0.89, + "learning_rate": 9.293480441323973e-05, + "loss": 1.2738, + "theoretical_loss": 3.356028964118835, + "tokens_seen": 2921463808 + }, + { + "epoch": 0.89, + "learning_rate": 9.291875626880643e-05, + "loss": 1.3827, + "theoretical_loss": 3.356023193959891, + "tokens_seen": 2921529344 + }, + { + "epoch": 0.89, + "learning_rate": 9.290270812437313e-05, + "loss": 1.3621, + "theoretical_loss": 3.356017423966623, + "tokens_seen": 2921594880 + }, + { + "epoch": 0.89, + "learning_rate": 9.288665997993983e-05, + "loss": 1.3152, + "theoretical_loss": 3.3560116541390244, + "tokens_seen": 2921660416 + }, + { + "epoch": 0.89, + "learning_rate": 9.287061183550653e-05, + "loss": 1.2209, + "theoretical_loss": 3.3560058844770846, + "tokens_seen": 2921725952 + }, + { + "epoch": 0.89, + "learning_rate": 9.285456369107323e-05, + "loss": 1.2417, + "theoretical_loss": 3.3560001149807963, + "tokens_seen": 2921791488 + }, + { + "epoch": 0.89, + "learning_rate": 9.283851554663993e-05, + "loss": 1.3359, + "theoretical_loss": 3.3559943456501506, + "tokens_seen": 2921857024 + }, + { + "epoch": 0.89, + "learning_rate": 9.282246740220662e-05, + "loss": 1.2516, + "theoretical_loss": 3.3559885764851396, + "tokens_seen": 2921922560 + }, + { + "epoch": 0.89, + "learning_rate": 9.280641925777332e-05, + "loss": 1.331, + "theoretical_loss": 3.355982807485754, + "tokens_seen": 2921988096 + }, + { + "epoch": 0.89, + "learning_rate": 9.279037111334002e-05, + "loss": 1.3594, + "theoretical_loss": 3.3559770386519854, + "tokens_seen": 2922053632 + }, + { + "epoch": 0.89, + "learning_rate": 9.277432296890672e-05, + "loss": 1.1798, + "theoretical_loss": 3.3559712699838258, + "tokens_seen": 2922119168 + }, + { + "epoch": 0.89, + "learning_rate": 9.275827482447342e-05, + "loss": 1.3078, + "theoretical_loss": 3.3559655014812666, + "tokens_seen": 2922184704 + }, + { + "epoch": 0.89, + "learning_rate": 9.274222668004012e-05, + "loss": 1.2673, + "theoretical_loss": 3.3559597331443, + "tokens_seen": 2922250240 + }, + { + "epoch": 0.89, + "learning_rate": 9.272617853560682e-05, + "loss": 1.1646, + "theoretical_loss": 3.355953964972916, + "tokens_seen": 2922315776 + }, + { + "epoch": 0.89, + "learning_rate": 9.271013039117352e-05, + "loss": 1.3013, + "theoretical_loss": 3.3559481969671072, + "tokens_seen": 2922381312 + }, + { + "epoch": 0.89, + "learning_rate": 9.269408224674022e-05, + "loss": 1.1942, + "theoretical_loss": 3.355942429126865, + "tokens_seen": 2922446848 + }, + { + "epoch": 0.89, + "learning_rate": 9.267803410230693e-05, + "loss": 1.366, + "theoretical_loss": 3.3559366614521804, + "tokens_seen": 2922512384 + }, + { + "epoch": 0.89, + "learning_rate": 9.266198595787363e-05, + "loss": 1.3277, + "theoretical_loss": 3.355930893943046, + "tokens_seen": 2922577920 + }, + { + "epoch": 0.89, + "learning_rate": 9.264593781344033e-05, + "loss": 1.199, + "theoretical_loss": 3.3559251265994523, + "tokens_seen": 2922643456 + }, + { + "epoch": 0.89, + "learning_rate": 9.262988966900703e-05, + "loss": 1.3142, + "theoretical_loss": 3.3559193594213914, + "tokens_seen": 2922708992 + }, + { + "epoch": 0.89, + "learning_rate": 9.261384152457373e-05, + "loss": 1.1894, + "theoretical_loss": 3.3559135924088546, + "tokens_seen": 2922774528 + }, + { + "epoch": 0.89, + "learning_rate": 9.259779338014043e-05, + "loss": 1.325, + "theoretical_loss": 3.3559078255618338, + "tokens_seen": 2922840064 + }, + { + "epoch": 0.89, + "objective/train/docs_used": 1062456, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.2637553215026855, + "objective/train/theoretical_loss": 3.3559020588803197, + "objective/train/tokens_used": 2943365600, + "theoretical_loss": 3.3559020588803197, + "tokens_seen": 2922905600 + }, + { + "epoch": 0.89, + "learning_rate": 9.258174523570713e-05, + "loss": 1.2235, + "theoretical_loss": 3.3559020588803197, + "tokens_seen": 2922905600 + }, + { + "epoch": 0.89, + "learning_rate": 9.256569709127383e-05, + "loss": 1.2351, + "theoretical_loss": 3.3558962923643048, + "tokens_seen": 2922971136 + }, + { + "epoch": 0.89, + "learning_rate": 9.254964894684053e-05, + "loss": 1.2314, + "theoretical_loss": 3.3558905260137806, + "tokens_seen": 2923036672 + }, + { + "epoch": 0.89, + "learning_rate": 9.253360080240723e-05, + "loss": 1.3321, + "theoretical_loss": 3.3558847598287374, + "tokens_seen": 2923102208 + }, + { + "epoch": 0.89, + "learning_rate": 9.251755265797393e-05, + "loss": 1.2103, + "theoretical_loss": 3.355878993809168, + "tokens_seen": 2923167744 + }, + { + "epoch": 0.89, + "learning_rate": 9.250150451354063e-05, + "loss": 1.3779, + "theoretical_loss": 3.355873227955064, + "tokens_seen": 2923233280 + }, + { + "epoch": 0.89, + "learning_rate": 9.248545636910732e-05, + "loss": 1.1829, + "theoretical_loss": 3.355867462266416, + "tokens_seen": 2923298816 + }, + { + "epoch": 0.89, + "learning_rate": 9.246940822467402e-05, + "loss": 1.3183, + "theoretical_loss": 3.355861696743216, + "tokens_seen": 2923364352 + }, + { + "epoch": 0.89, + "learning_rate": 9.245336008024072e-05, + "loss": 1.1915, + "theoretical_loss": 3.355855931385456, + "tokens_seen": 2923429888 + }, + { + "epoch": 0.89, + "learning_rate": 9.243731193580742e-05, + "loss": 1.2764, + "theoretical_loss": 3.355850166193127, + "tokens_seen": 2923495424 + }, + { + "epoch": 0.89, + "learning_rate": 9.242126379137412e-05, + "loss": 1.4207, + "theoretical_loss": 3.3558444011662205, + "tokens_seen": 2923560960 + }, + { + "epoch": 0.89, + "learning_rate": 9.240521564694082e-05, + "loss": 1.2367, + "theoretical_loss": 3.3558386363047283, + "tokens_seen": 2923626496 + }, + { + "epoch": 0.89, + "learning_rate": 9.238916750250752e-05, + "loss": 1.3243, + "theoretical_loss": 3.3558328716086416, + "tokens_seen": 2923692032 + }, + { + "epoch": 0.89, + "learning_rate": 9.237311935807422e-05, + "loss": 1.2179, + "theoretical_loss": 3.355827107077953, + "tokens_seen": 2923757568 + }, + { + "epoch": 0.89, + "learning_rate": 9.235707121364092e-05, + "loss": 1.2418, + "theoretical_loss": 3.3558213427126526, + "tokens_seen": 2923823104 + }, + { + "epoch": 0.89, + "learning_rate": 9.234102306920764e-05, + "loss": 1.3651, + "theoretical_loss": 3.3558155785127326, + "tokens_seen": 2923888640 + }, + { + "epoch": 0.89, + "learning_rate": 9.232497492477433e-05, + "loss": 1.2588, + "theoretical_loss": 3.355809814478185, + "tokens_seen": 2923954176 + }, + { + "epoch": 0.89, + "learning_rate": 9.230892678034103e-05, + "loss": 1.384, + "theoretical_loss": 3.3558040506090006, + "tokens_seen": 2924019712 + }, + { + "epoch": 0.89, + "learning_rate": 9.229287863590773e-05, + "loss": 1.2497, + "theoretical_loss": 3.3557982869051712, + "tokens_seen": 2924085248 + }, + { + "epoch": 0.89, + "learning_rate": 9.227683049147443e-05, + "loss": 1.3604, + "theoretical_loss": 3.3557925233666888, + "tokens_seen": 2924150784 + }, + { + "epoch": 0.89, + "learning_rate": 9.226078234704113e-05, + "loss": 1.318, + "theoretical_loss": 3.3557867599935443, + "tokens_seen": 2924216320 + }, + { + "epoch": 0.89, + "learning_rate": 9.224473420260783e-05, + "loss": 1.3048, + "theoretical_loss": 3.3557809967857293, + "tokens_seen": 2924281856 + }, + { + "epoch": 0.89, + "learning_rate": 9.222868605817453e-05, + "loss": 1.1773, + "theoretical_loss": 3.355775233743236, + "tokens_seen": 2924347392 + }, + { + "epoch": 0.89, + "learning_rate": 9.221263791374123e-05, + "loss": 1.3464, + "theoretical_loss": 3.355769470866055, + "tokens_seen": 2924412928 + }, + { + "epoch": 0.89, + "learning_rate": 9.219658976930793e-05, + "loss": 1.3466, + "theoretical_loss": 3.355763708154179, + "tokens_seen": 2924478464 + }, + { + "epoch": 0.89, + "learning_rate": 9.218054162487463e-05, + "loss": 1.3984, + "theoretical_loss": 3.355757945607598, + "tokens_seen": 2924544000 + }, + { + "epoch": 0.89, + "learning_rate": 9.216449348044133e-05, + "loss": 1.3674, + "theoretical_loss": 3.3557521832263055, + "tokens_seen": 2924609536 + }, + { + "epoch": 0.89, + "learning_rate": 9.214844533600803e-05, + "loss": 1.4135, + "theoretical_loss": 3.3557464210102914, + "tokens_seen": 2924675072 + }, + { + "epoch": 0.89, + "learning_rate": 9.213239719157472e-05, + "loss": 1.2696, + "theoretical_loss": 3.355740658959548, + "tokens_seen": 2924740608 + }, + { + "epoch": 0.89, + "learning_rate": 9.211634904714142e-05, + "loss": 1.2608, + "theoretical_loss": 3.355734897074067, + "tokens_seen": 2924806144 + }, + { + "epoch": 0.89, + "learning_rate": 9.210030090270812e-05, + "loss": 1.2244, + "theoretical_loss": 3.3557291353538394, + "tokens_seen": 2924871680 + }, + { + "epoch": 0.89, + "learning_rate": 9.208425275827482e-05, + "loss": 1.2274, + "theoretical_loss": 3.3557233737988574, + "tokens_seen": 2924937216 + }, + { + "epoch": 0.89, + "learning_rate": 9.206820461384152e-05, + "loss": 1.331, + "theoretical_loss": 3.355717612409112, + "tokens_seen": 2925002752 + }, + { + "epoch": 0.89, + "learning_rate": 9.205215646940822e-05, + "loss": 1.3723, + "theoretical_loss": 3.355711851184595, + "tokens_seen": 2925068288 + }, + { + "epoch": 0.89, + "learning_rate": 9.203610832497494e-05, + "loss": 1.2483, + "theoretical_loss": 3.355706090125298, + "tokens_seen": 2925133824 + }, + { + "epoch": 0.89, + "learning_rate": 9.202006018054164e-05, + "loss": 1.218, + "theoretical_loss": 3.3557003292312126, + "tokens_seen": 2925199360 + }, + { + "epoch": 0.89, + "learning_rate": 9.200401203610834e-05, + "loss": 1.3552, + "theoretical_loss": 3.3556945685023303, + "tokens_seen": 2925264896 + }, + { + "epoch": 0.89, + "learning_rate": 9.198796389167504e-05, + "loss": 1.2908, + "theoretical_loss": 3.355688807938642, + "tokens_seen": 2925330432 + }, + { + "epoch": 0.89, + "learning_rate": 9.197191574724173e-05, + "loss": 1.1897, + "theoretical_loss": 3.3556830475401407, + "tokens_seen": 2925395968 + }, + { + "epoch": 0.89, + "learning_rate": 9.195586760280843e-05, + "loss": 1.2428, + "theoretical_loss": 3.3556772873068166, + "tokens_seen": 2925461504 + }, + { + "epoch": 0.89, + "learning_rate": 9.193981945837513e-05, + "loss": 1.2687, + "theoretical_loss": 3.355671527238662, + "tokens_seen": 2925527040 + }, + { + "epoch": 0.89, + "learning_rate": 9.192377131394183e-05, + "loss": 1.3419, + "theoretical_loss": 3.3556657673356685, + "tokens_seen": 2925592576 + }, + { + "epoch": 0.89, + "learning_rate": 9.190772316950853e-05, + "loss": 1.2028, + "theoretical_loss": 3.355660007597827, + "tokens_seen": 2925658112 + }, + { + "epoch": 0.89, + "learning_rate": 9.189167502507523e-05, + "loss": 1.2334, + "theoretical_loss": 3.3556542480251297, + "tokens_seen": 2925723648 + }, + { + "epoch": 0.89, + "learning_rate": 9.187562688064193e-05, + "loss": 1.2918, + "theoretical_loss": 3.355648488617568, + "tokens_seen": 2925789184 + }, + { + "epoch": 0.89, + "learning_rate": 9.185957873620863e-05, + "loss": 1.2882, + "theoretical_loss": 3.3556427293751336, + "tokens_seen": 2925854720 + }, + { + "epoch": 0.89, + "learning_rate": 9.184353059177533e-05, + "loss": 1.2196, + "theoretical_loss": 3.3556369702978177, + "tokens_seen": 2925920256 + }, + { + "epoch": 0.89, + "learning_rate": 9.182748244734203e-05, + "loss": 1.3689, + "theoretical_loss": 3.355631211385612, + "tokens_seen": 2925985792 + }, + { + "epoch": 0.89, + "learning_rate": 9.181143430290873e-05, + "loss": 1.3499, + "theoretical_loss": 3.3556254526385083, + "tokens_seen": 2926051328 + }, + { + "epoch": 0.89, + "learning_rate": 9.179538615847543e-05, + "loss": 1.206, + "theoretical_loss": 3.355619694056498, + "tokens_seen": 2926116864 + }, + { + "epoch": 0.89, + "objective/train/docs_used": 1063661, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.2596585750579834, + "objective/train/theoretical_loss": 3.355613935639573, + "objective/train/tokens_used": 2946642400, + "theoretical_loss": 3.355613935639573, + "tokens_seen": 2926182400 + }, + { + "epoch": 0.89, + "learning_rate": 9.177933801404212e-05, + "loss": 1.278, + "theoretical_loss": 3.355613935639573, + "tokens_seen": 2926182400 + }, + { + "epoch": 0.89, + "learning_rate": 9.176328986960882e-05, + "loss": 1.1758, + "theoretical_loss": 3.3556081773877238, + "tokens_seen": 2926247936 + }, + { + "epoch": 0.89, + "learning_rate": 9.174724172517552e-05, + "loss": 1.2049, + "theoretical_loss": 3.355602419300943, + "tokens_seen": 2926313472 + }, + { + "epoch": 0.89, + "learning_rate": 9.173119358074224e-05, + "loss": 1.2668, + "theoretical_loss": 3.3555966613792223, + "tokens_seen": 2926379008 + }, + { + "epoch": 0.89, + "learning_rate": 9.171514543630894e-05, + "loss": 1.2221, + "theoretical_loss": 3.3555909036225526, + "tokens_seen": 2926444544 + }, + { + "epoch": 0.89, + "learning_rate": 9.169909729187564e-05, + "loss": 1.3081, + "theoretical_loss": 3.3555851460309256, + "tokens_seen": 2926510080 + }, + { + "epoch": 0.89, + "learning_rate": 9.168304914744234e-05, + "loss": 1.2305, + "theoretical_loss": 3.3555793886043332, + "tokens_seen": 2926575616 + }, + { + "epoch": 0.89, + "learning_rate": 9.166700100300904e-05, + "loss": 1.2578, + "theoretical_loss": 3.3555736313427666, + "tokens_seen": 2926641152 + }, + { + "epoch": 0.89, + "learning_rate": 9.165095285857574e-05, + "loss": 1.3085, + "theoretical_loss": 3.355567874246218, + "tokens_seen": 2926706688 + }, + { + "epoch": 0.89, + "learning_rate": 9.163490471414244e-05, + "loss": 1.2023, + "theoretical_loss": 3.355562117314678, + "tokens_seen": 2926772224 + }, + { + "epoch": 0.89, + "learning_rate": 9.161885656970913e-05, + "loss": 1.2388, + "theoretical_loss": 3.355556360548139, + "tokens_seen": 2926837760 + }, + { + "epoch": 0.89, + "learning_rate": 9.160280842527583e-05, + "loss": 1.208, + "theoretical_loss": 3.3555506039465923, + "tokens_seen": 2926903296 + }, + { + "epoch": 0.89, + "learning_rate": 9.158676028084253e-05, + "loss": 1.2554, + "theoretical_loss": 3.3555448475100293, + "tokens_seen": 2926968832 + }, + { + "epoch": 0.89, + "learning_rate": 9.157071213640923e-05, + "loss": 1.2249, + "theoretical_loss": 3.355539091238442, + "tokens_seen": 2927034368 + }, + { + "epoch": 0.89, + "learning_rate": 9.155466399197593e-05, + "loss": 1.3503, + "theoretical_loss": 3.3555333351318213, + "tokens_seen": 2927099904 + }, + { + "epoch": 0.89, + "learning_rate": 9.153861584754263e-05, + "loss": 1.332, + "theoretical_loss": 3.3555275791901593, + "tokens_seen": 2927165440 + }, + { + "epoch": 0.89, + "learning_rate": 9.152256770310933e-05, + "loss": 1.3037, + "theoretical_loss": 3.3555218234134476, + "tokens_seen": 2927230976 + }, + { + "epoch": 0.89, + "learning_rate": 9.150651955867603e-05, + "loss": 1.27, + "theoretical_loss": 3.3555160678016778, + "tokens_seen": 2927296512 + }, + { + "epoch": 0.89, + "learning_rate": 9.149047141424273e-05, + "loss": 1.3252, + "theoretical_loss": 3.3555103123548413, + "tokens_seen": 2927362048 + }, + { + "epoch": 0.89, + "learning_rate": 9.147442326980943e-05, + "loss": 1.3632, + "theoretical_loss": 3.35550455707293, + "tokens_seen": 2927427584 + }, + { + "epoch": 0.89, + "learning_rate": 9.145837512537613e-05, + "loss": 1.247, + "theoretical_loss": 3.3554988019559344, + "tokens_seen": 2927493120 + }, + { + "epoch": 0.89, + "learning_rate": 9.144232698094283e-05, + "loss": 1.3109, + "theoretical_loss": 3.3554930470038475, + "tokens_seen": 2927558656 + }, + { + "epoch": 0.89, + "learning_rate": 9.142627883650953e-05, + "loss": 1.3262, + "theoretical_loss": 3.35548729221666, + "tokens_seen": 2927624192 + }, + { + "epoch": 0.89, + "learning_rate": 9.141023069207623e-05, + "loss": 1.2698, + "theoretical_loss": 3.355481537594364, + "tokens_seen": 2927689728 + }, + { + "epoch": 0.89, + "learning_rate": 9.139418254764294e-05, + "loss": 1.2776, + "theoretical_loss": 3.3554757831369506, + "tokens_seen": 2927755264 + }, + { + "epoch": 0.89, + "learning_rate": 9.137813440320964e-05, + "loss": 1.2181, + "theoretical_loss": 3.3554700288444117, + "tokens_seen": 2927820800 + }, + { + "epoch": 0.89, + "learning_rate": 9.136208625877634e-05, + "loss": 1.2587, + "theoretical_loss": 3.3554642747167387, + "tokens_seen": 2927886336 + }, + { + "epoch": 0.89, + "learning_rate": 9.134603811434304e-05, + "loss": 1.3371, + "theoretical_loss": 3.3554585207539236, + "tokens_seen": 2927951872 + }, + { + "epoch": 0.89, + "learning_rate": 9.132998996990974e-05, + "loss": 1.3348, + "theoretical_loss": 3.3554527669559575, + "tokens_seen": 2928017408 + }, + { + "epoch": 0.89, + "learning_rate": 9.131394182547644e-05, + "loss": 1.269, + "theoretical_loss": 3.355447013322832, + "tokens_seen": 2928082944 + }, + { + "epoch": 0.89, + "learning_rate": 9.129789368104314e-05, + "loss": 1.1958, + "theoretical_loss": 3.355441259854539, + "tokens_seen": 2928148480 + }, + { + "epoch": 0.89, + "learning_rate": 9.128184553660984e-05, + "loss": 1.2228, + "theoretical_loss": 3.3554355065510704, + "tokens_seen": 2928214016 + }, + { + "epoch": 0.89, + "learning_rate": 9.126579739217653e-05, + "loss": 1.2389, + "theoretical_loss": 3.355429753412417, + "tokens_seen": 2928279552 + }, + { + "epoch": 0.89, + "learning_rate": 9.124974924774323e-05, + "loss": 1.2772, + "theoretical_loss": 3.3554240004385703, + "tokens_seen": 2928345088 + }, + { + "epoch": 0.89, + "learning_rate": 9.123370110330993e-05, + "loss": 1.234, + "theoretical_loss": 3.3554182476295225, + "tokens_seen": 2928410624 + }, + { + "epoch": 0.89, + "learning_rate": 9.121765295887663e-05, + "loss": 1.237, + "theoretical_loss": 3.3554124949852655, + "tokens_seen": 2928476160 + }, + { + "epoch": 0.89, + "learning_rate": 9.120160481444333e-05, + "loss": 1.2794, + "theoretical_loss": 3.3554067425057896, + "tokens_seen": 2928541696 + }, + { + "epoch": 0.89, + "learning_rate": 9.118555667001003e-05, + "loss": 1.3175, + "theoretical_loss": 3.3554009901910877, + "tokens_seen": 2928607232 + }, + { + "epoch": 0.89, + "learning_rate": 9.116950852557673e-05, + "loss": 1.3394, + "theoretical_loss": 3.355395238041151, + "tokens_seen": 2928672768 + }, + { + "epoch": 0.89, + "learning_rate": 9.115346038114343e-05, + "loss": 1.2797, + "theoretical_loss": 3.3553894860559708, + "tokens_seen": 2928738304 + }, + { + "epoch": 0.89, + "learning_rate": 9.113741223671015e-05, + "loss": 1.3291, + "theoretical_loss": 3.355383734235539, + "tokens_seen": 2928803840 + }, + { + "epoch": 0.89, + "learning_rate": 9.112136409227685e-05, + "loss": 1.25, + "theoretical_loss": 3.355377982579847, + "tokens_seen": 2928869376 + }, + { + "epoch": 0.89, + "learning_rate": 9.110531594784353e-05, + "loss": 1.287, + "theoretical_loss": 3.3553722310888863, + "tokens_seen": 2928934912 + }, + { + "epoch": 0.89, + "learning_rate": 9.108926780341023e-05, + "loss": 1.2634, + "theoretical_loss": 3.355366479762649, + "tokens_seen": 2929000448 + }, + { + "epoch": 0.89, + "learning_rate": 9.107321965897693e-05, + "loss": 1.2242, + "theoretical_loss": 3.355360728601126, + "tokens_seen": 2929065984 + }, + { + "epoch": 0.89, + "learning_rate": 9.105717151454364e-05, + "loss": 1.2225, + "theoretical_loss": 3.3553549776043097, + "tokens_seen": 2929131520 + }, + { + "epoch": 0.89, + "learning_rate": 9.104112337011034e-05, + "loss": 1.2401, + "theoretical_loss": 3.3553492267721907, + "tokens_seen": 2929197056 + }, + { + "epoch": 0.89, + "learning_rate": 9.102507522567704e-05, + "loss": 1.3404, + "theoretical_loss": 3.3553434761047614, + "tokens_seen": 2929262592 + }, + { + "epoch": 0.89, + "learning_rate": 9.100902708124374e-05, + "loss": 1.2466, + "theoretical_loss": 3.355337725602013, + "tokens_seen": 2929328128 + }, + { + "epoch": 0.89, + "learning_rate": 9.099297893681044e-05, + "loss": 1.2563, + "theoretical_loss": 3.3553319752639377, + "tokens_seen": 2929393664 + }, + { + "epoch": 0.89, + "objective/train/docs_used": 1064613, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.3261871337890625, + "objective/train/theoretical_loss": 3.355326225090526, + "objective/train/tokens_used": 2949919200, + "theoretical_loss": 3.355326225090526, + "tokens_seen": 2929459200 + }, + { + "epoch": 0.89, + "learning_rate": 9.097693079237714e-05, + "loss": 1.2965, + "theoretical_loss": 3.355326225090526, + "tokens_seen": 2929459200 + }, + { + "epoch": 0.89, + "learning_rate": 9.096088264794384e-05, + "loss": 1.3383, + "theoretical_loss": 3.3553204750817707, + "tokens_seen": 2929524736 + }, + { + "epoch": 0.89, + "learning_rate": 9.094483450351054e-05, + "loss": 1.3288, + "theoretical_loss": 3.3553147252376627, + "tokens_seen": 2929590272 + }, + { + "epoch": 0.89, + "learning_rate": 9.092878635907724e-05, + "loss": 1.2587, + "theoretical_loss": 3.3553089755581937, + "tokens_seen": 2929655808 + }, + { + "epoch": 0.89, + "learning_rate": 9.091273821464393e-05, + "loss": 1.2466, + "theoretical_loss": 3.3553032260433553, + "tokens_seen": 2929721344 + }, + { + "epoch": 0.89, + "learning_rate": 9.089669007021063e-05, + "loss": 1.3068, + "theoretical_loss": 3.3552974766931394, + "tokens_seen": 2929786880 + }, + { + "epoch": 0.89, + "learning_rate": 9.088064192577733e-05, + "loss": 1.3585, + "theoretical_loss": 3.355291727507537, + "tokens_seen": 2929852416 + }, + { + "epoch": 0.89, + "learning_rate": 9.086459378134403e-05, + "loss": 1.3381, + "theoretical_loss": 3.35528597848654, + "tokens_seen": 2929917952 + }, + { + "epoch": 0.89, + "learning_rate": 9.084854563691073e-05, + "loss": 1.2856, + "theoretical_loss": 3.3552802296301407, + "tokens_seen": 2929983488 + }, + { + "epoch": 0.89, + "learning_rate": 9.083249749247744e-05, + "loss": 1.2949, + "theoretical_loss": 3.3552744809383297, + "tokens_seen": 2930049024 + }, + { + "epoch": 0.89, + "learning_rate": 9.081644934804414e-05, + "loss": 1.1985, + "theoretical_loss": 3.355268732411099, + "tokens_seen": 2930114560 + }, + { + "epoch": 0.89, + "learning_rate": 9.080040120361085e-05, + "loss": 1.2678, + "theoretical_loss": 3.3552629840484403, + "tokens_seen": 2930180096 + }, + { + "epoch": 0.89, + "learning_rate": 9.078435305917755e-05, + "loss": 1.1391, + "theoretical_loss": 3.355257235850345, + "tokens_seen": 2930245632 + }, + { + "epoch": 0.89, + "learning_rate": 9.076830491474425e-05, + "loss": 1.3852, + "theoretical_loss": 3.3552514878168047, + "tokens_seen": 2930311168 + }, + { + "epoch": 0.89, + "learning_rate": 9.075225677031093e-05, + "loss": 1.2644, + "theoretical_loss": 3.3552457399478115, + "tokens_seen": 2930376704 + }, + { + "epoch": 0.89, + "learning_rate": 9.073620862587763e-05, + "loss": 1.3027, + "theoretical_loss": 3.355239992243356, + "tokens_seen": 2930442240 + }, + { + "epoch": 0.89, + "learning_rate": 9.072016048144434e-05, + "loss": 1.1245, + "theoretical_loss": 3.355234244703431, + "tokens_seen": 2930507776 + }, + { + "epoch": 0.89, + "learning_rate": 9.070411233701104e-05, + "loss": 1.3316, + "theoretical_loss": 3.3552284973280275, + "tokens_seen": 2930573312 + }, + { + "epoch": 0.89, + "learning_rate": 9.068806419257774e-05, + "loss": 1.3446, + "theoretical_loss": 3.355222750117137, + "tokens_seen": 2930638848 + }, + { + "epoch": 0.89, + "learning_rate": 9.067201604814444e-05, + "loss": 1.3447, + "theoretical_loss": 3.3552170030707513, + "tokens_seen": 2930704384 + }, + { + "epoch": 0.89, + "learning_rate": 9.065596790371114e-05, + "loss": 1.2988, + "theoretical_loss": 3.3552112561888623, + "tokens_seen": 2930769920 + }, + { + "epoch": 0.89, + "learning_rate": 9.063991975927784e-05, + "loss": 1.2689, + "theoretical_loss": 3.355205509471461, + "tokens_seen": 2930835456 + }, + { + "epoch": 0.89, + "learning_rate": 9.062387161484454e-05, + "loss": 1.3848, + "theoretical_loss": 3.3551997629185393, + "tokens_seen": 2930900992 + }, + { + "epoch": 0.89, + "learning_rate": 9.060782347041124e-05, + "loss": 1.3423, + "theoretical_loss": 3.355194016530089, + "tokens_seen": 2930966528 + }, + { + "epoch": 0.89, + "learning_rate": 9.059177532597794e-05, + "loss": 1.2408, + "theoretical_loss": 3.355188270306101, + "tokens_seen": 2931032064 + }, + { + "epoch": 0.89, + "learning_rate": 9.057572718154464e-05, + "loss": 1.364, + "theoretical_loss": 3.355182524246568, + "tokens_seen": 2931097600 + }, + { + "epoch": 0.89, + "learning_rate": 9.055967903711133e-05, + "loss": 1.2415, + "theoretical_loss": 3.355176778351481, + "tokens_seen": 2931163136 + }, + { + "epoch": 0.89, + "learning_rate": 9.054363089267803e-05, + "loss": 1.3166, + "theoretical_loss": 3.3551710326208317, + "tokens_seen": 2931228672 + }, + { + "epoch": 0.89, + "learning_rate": 9.052758274824474e-05, + "loss": 1.2252, + "theoretical_loss": 3.3551652870546116, + "tokens_seen": 2931294208 + }, + { + "epoch": 0.89, + "learning_rate": 9.051153460381144e-05, + "loss": 1.4405, + "theoretical_loss": 3.355159541652813, + "tokens_seen": 2931359744 + }, + { + "epoch": 0.89, + "learning_rate": 9.049548645937814e-05, + "loss": 1.2811, + "theoretical_loss": 3.355153796415426, + "tokens_seen": 2931425280 + }, + { + "epoch": 0.89, + "learning_rate": 9.047943831494484e-05, + "loss": 1.2575, + "theoretical_loss": 3.355148051342444, + "tokens_seen": 2931490816 + }, + { + "epoch": 0.89, + "learning_rate": 9.046339017051155e-05, + "loss": 1.3141, + "theoretical_loss": 3.3551423064338572, + "tokens_seen": 2931556352 + }, + { + "epoch": 0.89, + "learning_rate": 9.044734202607825e-05, + "loss": 1.1534, + "theoretical_loss": 3.3551365616896582, + "tokens_seen": 2931621888 + }, + { + "epoch": 0.89, + "learning_rate": 9.043129388164495e-05, + "loss": 1.207, + "theoretical_loss": 3.355130817109838, + "tokens_seen": 2931687424 + }, + { + "epoch": 0.89, + "learning_rate": 9.041524573721163e-05, + "loss": 1.3106, + "theoretical_loss": 3.3551250726943884, + "tokens_seen": 2931752960 + }, + { + "epoch": 0.89, + "learning_rate": 9.039919759277833e-05, + "loss": 1.2172, + "theoretical_loss": 3.355119328443301, + "tokens_seen": 2931818496 + }, + { + "epoch": 0.89, + "learning_rate": 9.038314944834503e-05, + "loss": 1.3323, + "theoretical_loss": 3.3551135843565683, + "tokens_seen": 2931884032 + }, + { + "epoch": 0.89, + "learning_rate": 9.036710130391174e-05, + "loss": 1.1327, + "theoretical_loss": 3.3551078404341803, + "tokens_seen": 2931949568 + }, + { + "epoch": 0.89, + "learning_rate": 9.035105315947844e-05, + "loss": 1.1399, + "theoretical_loss": 3.35510209667613, + "tokens_seen": 2932015104 + }, + { + "epoch": 0.89, + "learning_rate": 9.033500501504514e-05, + "loss": 1.3409, + "theoretical_loss": 3.3550963530824083, + "tokens_seen": 2932080640 + }, + { + "epoch": 0.89, + "learning_rate": 9.031895687061184e-05, + "loss": 1.2216, + "theoretical_loss": 3.355090609653007, + "tokens_seen": 2932146176 + }, + { + "epoch": 0.89, + "learning_rate": 9.030290872617854e-05, + "loss": 1.2184, + "theoretical_loss": 3.355084866387917, + "tokens_seen": 2932211712 + }, + { + "epoch": 0.89, + "learning_rate": 9.028686058174524e-05, + "loss": 1.2053, + "theoretical_loss": 3.3550791232871315, + "tokens_seen": 2932277248 + }, + { + "epoch": 0.89, + "learning_rate": 9.027081243731194e-05, + "loss": 1.3297, + "theoretical_loss": 3.355073380350641, + "tokens_seen": 2932342784 + }, + { + "epoch": 0.89, + "learning_rate": 9.025476429287864e-05, + "loss": 1.2495, + "theoretical_loss": 3.3550676375784376, + "tokens_seen": 2932408320 + }, + { + "epoch": 0.89, + "learning_rate": 9.023871614844534e-05, + "loss": 1.253, + "theoretical_loss": 3.3550618949705124, + "tokens_seen": 2932473856 + }, + { + "epoch": 0.89, + "learning_rate": 9.022266800401204e-05, + "loss": 1.3218, + "theoretical_loss": 3.3550561525268576, + "tokens_seen": 2932539392 + }, + { + "epoch": 0.89, + "learning_rate": 9.020661985957874e-05, + "loss": 1.4226, + "theoretical_loss": 3.3550504102474648, + "tokens_seen": 2932604928 + }, + { + "epoch": 0.89, + "learning_rate": 9.019057171514544e-05, + "loss": 1.312, + "theoretical_loss": 3.355044668132325, + "tokens_seen": 2932670464 + }, + { + "epoch": 0.89, + "objective/train/docs_used": 1065874, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.1663317680358887, + "objective/train/theoretical_loss": 3.3550389261814306, + "objective/train/tokens_used": 2953196000, + "theoretical_loss": 3.3550389261814306, + "tokens_seen": 2932736000 + }, + { + "epoch": 0.89, + "learning_rate": 9.017452357071214e-05, + "loss": 1.2789, + "theoretical_loss": 3.3550389261814306, + "tokens_seen": 2932736000 + }, + { + "epoch": 0.89, + "learning_rate": 9.015847542627884e-05, + "loss": 1.2782, + "theoretical_loss": 3.355033184394773, + "tokens_seen": 2932801536 + }, + { + "epoch": 0.89, + "learning_rate": 9.014242728184554e-05, + "loss": 1.3466, + "theoretical_loss": 3.355027442772343, + "tokens_seen": 2932867072 + }, + { + "epoch": 0.89, + "learning_rate": 9.012637913741224e-05, + "loss": 1.2438, + "theoretical_loss": 3.3550217013141337, + "tokens_seen": 2932932608 + }, + { + "epoch": 0.89, + "learning_rate": 9.011033099297895e-05, + "loss": 1.2062, + "theoretical_loss": 3.3550159600201357, + "tokens_seen": 2932998144 + }, + { + "epoch": 0.89, + "learning_rate": 9.009428284854565e-05, + "loss": 1.2869, + "theoretical_loss": 3.355010218890341, + "tokens_seen": 2933063680 + }, + { + "epoch": 0.89, + "learning_rate": 9.007823470411235e-05, + "loss": 1.2786, + "theoretical_loss": 3.355004477924741, + "tokens_seen": 2933129216 + }, + { + "epoch": 0.89, + "learning_rate": 9.006218655967903e-05, + "loss": 1.225, + "theoretical_loss": 3.354998737123328, + "tokens_seen": 2933194752 + }, + { + "epoch": 0.89, + "learning_rate": 9.004613841524573e-05, + "loss": 1.1781, + "theoretical_loss": 3.3549929964860925, + "tokens_seen": 2933260288 + }, + { + "epoch": 0.89, + "learning_rate": 9.003009027081244e-05, + "loss": 1.2236, + "theoretical_loss": 3.354987256013027, + "tokens_seen": 2933325824 + }, + { + "epoch": 0.89, + "learning_rate": 9.001404212637914e-05, + "loss": 1.3399, + "theoretical_loss": 3.354981515704123, + "tokens_seen": 2933391360 + }, + { + "epoch": 0.89, + "learning_rate": 8.999799398194584e-05, + "loss": 1.1408, + "theoretical_loss": 3.354975775559372, + "tokens_seen": 2933456896 + }, + { + "epoch": 0.89, + "learning_rate": 8.998194583751254e-05, + "loss": 1.3463, + "theoretical_loss": 3.3549700355787655, + "tokens_seen": 2933522432 + }, + { + "epoch": 0.89, + "learning_rate": 8.996589769307924e-05, + "loss": 1.1952, + "theoretical_loss": 3.3549642957622954, + "tokens_seen": 2933587968 + }, + { + "epoch": 0.89, + "learning_rate": 8.994984954864594e-05, + "loss": 1.3042, + "theoretical_loss": 3.3549585561099535, + "tokens_seen": 2933653504 + }, + { + "epoch": 0.89, + "learning_rate": 8.993380140421265e-05, + "loss": 1.2449, + "theoretical_loss": 3.354952816621731, + "tokens_seen": 2933719040 + }, + { + "epoch": 0.89, + "learning_rate": 8.991775325977935e-05, + "loss": 1.3177, + "theoretical_loss": 3.35494707729762, + "tokens_seen": 2933784576 + }, + { + "epoch": 0.89, + "learning_rate": 8.990170511534604e-05, + "loss": 1.2534, + "theoretical_loss": 3.3549413381376114, + "tokens_seen": 2933850112 + }, + { + "epoch": 0.89, + "learning_rate": 8.988565697091274e-05, + "loss": 1.292, + "theoretical_loss": 3.354935599141698, + "tokens_seen": 2933915648 + }, + { + "epoch": 0.89, + "learning_rate": 8.986960882647944e-05, + "loss": 1.289, + "theoretical_loss": 3.3549298603098703, + "tokens_seen": 2933981184 + }, + { + "epoch": 0.89, + "learning_rate": 8.985356068204614e-05, + "loss": 1.2968, + "theoretical_loss": 3.3549241216421204, + "tokens_seen": 2934046720 + }, + { + "epoch": 0.89, + "learning_rate": 8.983751253761284e-05, + "loss": 1.2405, + "theoretical_loss": 3.35491838313844, + "tokens_seen": 2934112256 + }, + { + "epoch": 0.89, + "learning_rate": 8.982146439317954e-05, + "loss": 1.3096, + "theoretical_loss": 3.354912644798821, + "tokens_seen": 2934177792 + }, + { + "epoch": 0.89, + "learning_rate": 8.980541624874624e-05, + "loss": 1.1268, + "theoretical_loss": 3.3549069066232544, + "tokens_seen": 2934243328 + }, + { + "epoch": 0.89, + "learning_rate": 8.978936810431294e-05, + "loss": 1.3151, + "theoretical_loss": 3.3549011686117325, + "tokens_seen": 2934308864 + }, + { + "epoch": 0.89, + "learning_rate": 8.977331995987965e-05, + "loss": 1.3125, + "theoretical_loss": 3.3548954307642465, + "tokens_seen": 2934374400 + }, + { + "epoch": 0.89, + "learning_rate": 8.975727181544635e-05, + "loss": 1.3723, + "theoretical_loss": 3.3548896930807883, + "tokens_seen": 2934439936 + }, + { + "epoch": 0.89, + "learning_rate": 8.974122367101305e-05, + "loss": 1.2444, + "theoretical_loss": 3.354883955561349, + "tokens_seen": 2934505472 + }, + { + "epoch": 0.89, + "learning_rate": 8.972517552657975e-05, + "loss": 1.2539, + "theoretical_loss": 3.354878218205921, + "tokens_seen": 2934571008 + }, + { + "epoch": 0.89, + "learning_rate": 8.970912738214643e-05, + "loss": 1.293, + "theoretical_loss": 3.3548724810144956, + "tokens_seen": 2934636544 + }, + { + "epoch": 0.89, + "learning_rate": 8.969307923771313e-05, + "loss": 1.3117, + "theoretical_loss": 3.3548667439870643, + "tokens_seen": 2934702080 + }, + { + "epoch": 0.89, + "learning_rate": 8.967703109327984e-05, + "loss": 1.2256, + "theoretical_loss": 3.3548610071236196, + "tokens_seen": 2934767616 + }, + { + "epoch": 0.89, + "learning_rate": 8.966098294884654e-05, + "loss": 1.2418, + "theoretical_loss": 3.3548552704241517, + "tokens_seen": 2934833152 + }, + { + "epoch": 0.89, + "learning_rate": 8.964493480441324e-05, + "loss": 1.3503, + "theoretical_loss": 3.3548495338886535, + "tokens_seen": 2934898688 + }, + { + "epoch": 0.89, + "learning_rate": 8.962888665997995e-05, + "loss": 1.3575, + "theoretical_loss": 3.354843797517116, + "tokens_seen": 2934964224 + }, + { + "epoch": 0.89, + "learning_rate": 8.961283851554665e-05, + "loss": 1.3225, + "theoretical_loss": 3.354838061309531, + "tokens_seen": 2935029760 + }, + { + "epoch": 0.89, + "learning_rate": 8.959679037111335e-05, + "loss": 1.3064, + "theoretical_loss": 3.3548323252658903, + "tokens_seen": 2935095296 + }, + { + "epoch": 0.89, + "learning_rate": 8.958074222668005e-05, + "loss": 1.2351, + "theoretical_loss": 3.3548265893861853, + "tokens_seen": 2935160832 + }, + { + "epoch": 0.89, + "learning_rate": 8.956469408224675e-05, + "loss": 1.2916, + "theoretical_loss": 3.3548208536704083, + "tokens_seen": 2935226368 + }, + { + "epoch": 0.89, + "learning_rate": 8.954864593781344e-05, + "loss": 1.2491, + "theoretical_loss": 3.35481511811855, + "tokens_seen": 2935291904 + }, + { + "epoch": 0.89, + "learning_rate": 8.953259779338014e-05, + "loss": 1.216, + "theoretical_loss": 3.3548093827306027, + "tokens_seen": 2935357440 + }, + { + "epoch": 0.89, + "learning_rate": 8.951654964894684e-05, + "loss": 1.2453, + "theoretical_loss": 3.3548036475065577, + "tokens_seen": 2935422976 + }, + { + "epoch": 0.89, + "learning_rate": 8.950050150451354e-05, + "loss": 1.246, + "theoretical_loss": 3.354797912446407, + "tokens_seen": 2935488512 + }, + { + "epoch": 0.89, + "learning_rate": 8.948445336008024e-05, + "loss": 1.2719, + "theoretical_loss": 3.354792177550142, + "tokens_seen": 2935554048 + }, + { + "epoch": 0.89, + "learning_rate": 8.946840521564694e-05, + "loss": 1.3203, + "theoretical_loss": 3.3547864428177543, + "tokens_seen": 2935619584 + }, + { + "epoch": 0.89, + "learning_rate": 8.945235707121364e-05, + "loss": 1.2657, + "theoretical_loss": 3.354780708249236, + "tokens_seen": 2935685120 + }, + { + "epoch": 0.89, + "learning_rate": 8.943630892678035e-05, + "loss": 1.1855, + "theoretical_loss": 3.3547749738445782, + "tokens_seen": 2935750656 + }, + { + "epoch": 0.89, + "learning_rate": 8.942026078234705e-05, + "loss": 1.249, + "theoretical_loss": 3.354769239603773, + "tokens_seen": 2935816192 + }, + { + "epoch": 0.89, + "learning_rate": 8.940421263791375e-05, + "loss": 1.2286, + "theoretical_loss": 3.3547635055268117, + "tokens_seen": 2935881728 + }, + { + "epoch": 0.89, + "learning_rate": 8.938816449348045e-05, + "loss": 1.3078, + "theoretical_loss": 3.354757771613686, + "tokens_seen": 2935947264 + }, + { + "epoch": 0.89, + "objective/train/docs_used": 1067139, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.237708330154419, + "objective/train/theoretical_loss": 3.354752037864388, + "objective/train/tokens_used": 2956472800, + "theoretical_loss": 3.354752037864388, + "tokens_seen": 2936012800 + }, + { + "epoch": 0.89, + "learning_rate": 8.937211634904715e-05, + "loss": 1.2705, + "theoretical_loss": 3.354752037864388, + "tokens_seen": 2936012800 + }, + { + "epoch": 0.89, + "learning_rate": 8.935606820461383e-05, + "loss": 1.2819, + "theoretical_loss": 3.354746304278909, + "tokens_seen": 2936078336 + }, + { + "epoch": 0.89, + "learning_rate": 8.934002006018054e-05, + "loss": 1.2612, + "theoretical_loss": 3.354740570857241, + "tokens_seen": 2936143872 + }, + { + "epoch": 0.89, + "learning_rate": 8.932397191574725e-05, + "loss": 1.3152, + "theoretical_loss": 3.354734837599375, + "tokens_seen": 2936209408 + }, + { + "epoch": 0.89, + "learning_rate": 8.930792377131395e-05, + "loss": 1.222, + "theoretical_loss": 3.3547291045053034, + "tokens_seen": 2936274944 + }, + { + "epoch": 0.89, + "learning_rate": 8.929187562688065e-05, + "loss": 1.1889, + "theoretical_loss": 3.3547233715750173, + "tokens_seen": 2936340480 + }, + { + "epoch": 0.89, + "learning_rate": 8.927582748244735e-05, + "loss": 1.227, + "theoretical_loss": 3.3547176388085083, + "tokens_seen": 2936406016 + }, + { + "epoch": 0.89, + "learning_rate": 8.925977933801405e-05, + "loss": 1.286, + "theoretical_loss": 3.354711906205769, + "tokens_seen": 2936471552 + }, + { + "epoch": 0.89, + "learning_rate": 8.924373119358075e-05, + "loss": 1.1777, + "theoretical_loss": 3.35470617376679, + "tokens_seen": 2936537088 + }, + { + "epoch": 0.89, + "learning_rate": 8.922768304914745e-05, + "loss": 1.2663, + "theoretical_loss": 3.3547004414915635, + "tokens_seen": 2936602624 + }, + { + "epoch": 0.89, + "learning_rate": 8.921163490471415e-05, + "loss": 1.2483, + "theoretical_loss": 3.354694709380081, + "tokens_seen": 2936668160 + }, + { + "epoch": 0.89, + "learning_rate": 8.919558676028084e-05, + "loss": 1.185, + "theoretical_loss": 3.3546889774323345, + "tokens_seen": 2936733696 + }, + { + "epoch": 0.89, + "learning_rate": 8.917953861584754e-05, + "loss": 1.2736, + "theoretical_loss": 3.3546832456483147, + "tokens_seen": 2936799232 + }, + { + "epoch": 0.89, + "learning_rate": 8.916349047141424e-05, + "loss": 1.3045, + "theoretical_loss": 3.3546775140280145, + "tokens_seen": 2936864768 + }, + { + "epoch": 0.89, + "learning_rate": 8.914744232698094e-05, + "loss": 1.3475, + "theoretical_loss": 3.354671782571425, + "tokens_seen": 2936930304 + }, + { + "epoch": 0.89, + "learning_rate": 8.913139418254764e-05, + "loss": 1.2923, + "theoretical_loss": 3.354666051278538, + "tokens_seen": 2936995840 + }, + { + "epoch": 0.89, + "learning_rate": 8.911534603811434e-05, + "loss": 1.2604, + "theoretical_loss": 3.3546603201493452, + "tokens_seen": 2937061376 + }, + { + "epoch": 0.89, + "learning_rate": 8.909929789368104e-05, + "loss": 1.1865, + "theoretical_loss": 3.3546545891838377, + "tokens_seen": 2937126912 + }, + { + "epoch": 0.89, + "learning_rate": 8.908324974924775e-05, + "loss": 1.2442, + "theoretical_loss": 3.3546488583820078, + "tokens_seen": 2937192448 + }, + { + "epoch": 0.89, + "learning_rate": 8.906720160481445e-05, + "loss": 1.362, + "theoretical_loss": 3.354643127743847, + "tokens_seen": 2937257984 + }, + { + "epoch": 0.89, + "learning_rate": 8.905115346038115e-05, + "loss": 1.2852, + "theoretical_loss": 3.354637397269347, + "tokens_seen": 2937323520 + }, + { + "epoch": 0.89, + "learning_rate": 8.903510531594785e-05, + "loss": 1.2197, + "theoretical_loss": 3.3546316669584995, + "tokens_seen": 2937389056 + }, + { + "epoch": 0.89, + "learning_rate": 8.901905717151455e-05, + "loss": 1.3458, + "theoretical_loss": 3.354625936811296, + "tokens_seen": 2937454592 + }, + { + "epoch": 0.89, + "learning_rate": 8.900300902708125e-05, + "loss": 1.2935, + "theoretical_loss": 3.3546202068277284, + "tokens_seen": 2937520128 + }, + { + "epoch": 0.89, + "learning_rate": 8.898696088264795e-05, + "loss": 1.2737, + "theoretical_loss": 3.3546144770077886, + "tokens_seen": 2937585664 + }, + { + "epoch": 0.89, + "learning_rate": 8.897091273821465e-05, + "loss": 1.3216, + "theoretical_loss": 3.3546087473514676, + "tokens_seen": 2937651200 + }, + { + "epoch": 0.89, + "learning_rate": 8.895486459378135e-05, + "loss": 1.317, + "theoretical_loss": 3.354603017858757, + "tokens_seen": 2937716736 + }, + { + "epoch": 0.89, + "learning_rate": 8.893881644934805e-05, + "loss": 1.3135, + "theoretical_loss": 3.3545972885296496, + "tokens_seen": 2937782272 + }, + { + "epoch": 0.89, + "learning_rate": 8.892276830491475e-05, + "loss": 1.2554, + "theoretical_loss": 3.354591559364136, + "tokens_seen": 2937847808 + }, + { + "epoch": 0.89, + "learning_rate": 8.890672016048145e-05, + "loss": 1.1943, + "theoretical_loss": 3.3545858303622085, + "tokens_seen": 2937913344 + }, + { + "epoch": 0.89, + "learning_rate": 8.889067201604815e-05, + "loss": 1.2776, + "theoretical_loss": 3.354580101523858, + "tokens_seen": 2937978880 + }, + { + "epoch": 0.89, + "learning_rate": 8.887462387161485e-05, + "loss": 1.297, + "theoretical_loss": 3.3545743728490773, + "tokens_seen": 2938044416 + }, + { + "epoch": 0.89, + "learning_rate": 8.885857572718155e-05, + "loss": 1.2618, + "theoretical_loss": 3.3545686443378573, + "tokens_seen": 2938109952 + }, + { + "epoch": 0.89, + "learning_rate": 8.884252758274824e-05, + "loss": 1.2036, + "theoretical_loss": 3.35456291599019, + "tokens_seen": 2938175488 + }, + { + "epoch": 0.89, + "learning_rate": 8.882647943831494e-05, + "loss": 1.2982, + "theoretical_loss": 3.354557187806067, + "tokens_seen": 2938241024 + }, + { + "epoch": 0.89, + "learning_rate": 8.881043129388164e-05, + "loss": 1.3302, + "theoretical_loss": 3.3545514597854798, + "tokens_seen": 2938306560 + }, + { + "epoch": 0.89, + "learning_rate": 8.879438314944834e-05, + "loss": 1.3277, + "theoretical_loss": 3.3545457319284204, + "tokens_seen": 2938372096 + }, + { + "epoch": 0.89, + "learning_rate": 8.877833500501504e-05, + "loss": 1.2863, + "theoretical_loss": 3.35454000423488, + "tokens_seen": 2938437632 + }, + { + "epoch": 0.89, + "learning_rate": 8.876228686058174e-05, + "loss": 1.133, + "theoretical_loss": 3.354534276704851, + "tokens_seen": 2938503168 + }, + { + "epoch": 0.89, + "learning_rate": 8.874623871614845e-05, + "loss": 1.1632, + "theoretical_loss": 3.354528549338325, + "tokens_seen": 2938568704 + }, + { + "epoch": 0.89, + "learning_rate": 8.873019057171516e-05, + "loss": 1.1828, + "theoretical_loss": 3.3545228221352925, + "tokens_seen": 2938634240 + }, + { + "epoch": 0.89, + "learning_rate": 8.871414242728186e-05, + "loss": 1.27, + "theoretical_loss": 3.354517095095747, + "tokens_seen": 2938699776 + }, + { + "epoch": 0.89, + "learning_rate": 8.869809428284855e-05, + "loss": 1.2351, + "theoretical_loss": 3.3545113682196783, + "tokens_seen": 2938765312 + }, + { + "epoch": 0.89, + "learning_rate": 8.868204613841525e-05, + "loss": 1.2765, + "theoretical_loss": 3.3545056415070795, + "tokens_seen": 2938830848 + }, + { + "epoch": 0.89, + "learning_rate": 8.866599799398195e-05, + "loss": 1.2615, + "theoretical_loss": 3.354499914957942, + "tokens_seen": 2938896384 + }, + { + "epoch": 0.89, + "learning_rate": 8.864994984954865e-05, + "loss": 1.2573, + "theoretical_loss": 3.354494188572257, + "tokens_seen": 2938961920 + }, + { + "epoch": 0.89, + "learning_rate": 8.863390170511535e-05, + "loss": 1.2844, + "theoretical_loss": 3.354488462350017, + "tokens_seen": 2939027456 + }, + { + "epoch": 0.89, + "learning_rate": 8.861785356068205e-05, + "loss": 1.292, + "theoretical_loss": 3.3544827362912124, + "tokens_seen": 2939092992 + }, + { + "epoch": 0.89, + "learning_rate": 8.860180541624875e-05, + "loss": 1.2801, + "theoretical_loss": 3.3544770103958363, + "tokens_seen": 2939158528 + }, + { + "epoch": 0.89, + "learning_rate": 8.858575727181545e-05, + "loss": 1.22, + "theoretical_loss": 3.3544712846638793, + "tokens_seen": 2939224064 + }, + { + "epoch": 0.89, + "objective/train/docs_used": 1068503, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.3247673511505127, + "objective/train/theoretical_loss": 3.354465559095334, + "objective/train/tokens_used": 2959749600, + "theoretical_loss": 3.354465559095334, + "tokens_seen": 2939289600 + }, + { + "epoch": 0.89, + "learning_rate": 8.856970912738215e-05, + "loss": 1.3337, + "theoretical_loss": 3.354465559095334, + "tokens_seen": 2939289600 + }, + { + "epoch": 0.89, + "learning_rate": 8.855366098294885e-05, + "loss": 1.2796, + "theoretical_loss": 3.354459833690192, + "tokens_seen": 2939355136 + }, + { + "epoch": 0.89, + "learning_rate": 8.853761283851555e-05, + "loss": 1.2981, + "theoretical_loss": 3.3544541084484436, + "tokens_seen": 2939420672 + }, + { + "epoch": 0.89, + "learning_rate": 8.852156469408225e-05, + "loss": 1.3208, + "theoretical_loss": 3.3544483833700824, + "tokens_seen": 2939486208 + }, + { + "epoch": 0.89, + "learning_rate": 8.850551654964894e-05, + "loss": 1.2619, + "theoretical_loss": 3.354442658455099, + "tokens_seen": 2939551744 + }, + { + "epoch": 0.89, + "learning_rate": 8.848946840521564e-05, + "loss": 1.2495, + "theoretical_loss": 3.3544369337034854, + "tokens_seen": 2939617280 + }, + { + "epoch": 0.89, + "learning_rate": 8.847342026078234e-05, + "loss": 1.3487, + "theoretical_loss": 3.354431209115233, + "tokens_seen": 2939682816 + }, + { + "epoch": 0.89, + "learning_rate": 8.845737211634904e-05, + "loss": 1.2998, + "theoretical_loss": 3.354425484690334, + "tokens_seen": 2939748352 + }, + { + "epoch": 0.89, + "learning_rate": 8.844132397191574e-05, + "loss": 1.298, + "theoretical_loss": 3.35441976042878, + "tokens_seen": 2939813888 + }, + { + "epoch": 0.89, + "learning_rate": 8.842527582748246e-05, + "loss": 1.3071, + "theoretical_loss": 3.3544140363305623, + "tokens_seen": 2939879424 + }, + { + "epoch": 0.89, + "learning_rate": 8.840922768304916e-05, + "loss": 1.3461, + "theoretical_loss": 3.354408312395673, + "tokens_seen": 2939944960 + }, + { + "epoch": 0.89, + "learning_rate": 8.839317953861586e-05, + "loss": 1.2145, + "theoretical_loss": 3.3544025886241036, + "tokens_seen": 2940010496 + }, + { + "epoch": 0.89, + "learning_rate": 8.837713139418256e-05, + "loss": 1.2613, + "theoretical_loss": 3.3543968650158456, + "tokens_seen": 2940076032 + }, + { + "epoch": 0.89, + "learning_rate": 8.836108324974926e-05, + "loss": 1.3034, + "theoretical_loss": 3.354391141570891, + "tokens_seen": 2940141568 + }, + { + "epoch": 0.89, + "learning_rate": 8.834503510531595e-05, + "loss": 1.2252, + "theoretical_loss": 3.3543854182892314, + "tokens_seen": 2940207104 + }, + { + "epoch": 0.89, + "learning_rate": 8.832898696088265e-05, + "loss": 1.3394, + "theoretical_loss": 3.354379695170859, + "tokens_seen": 2940272640 + }, + { + "epoch": 0.89, + "learning_rate": 8.831293881644935e-05, + "loss": 1.2507, + "theoretical_loss": 3.3543739722157646, + "tokens_seen": 2940338176 + }, + { + "epoch": 0.89, + "learning_rate": 8.829689067201605e-05, + "loss": 1.2313, + "theoretical_loss": 3.35436824942394, + "tokens_seen": 2940403712 + }, + { + "epoch": 0.89, + "learning_rate": 8.828084252758275e-05, + "loss": 1.3157, + "theoretical_loss": 3.354362526795378, + "tokens_seen": 2940469248 + }, + { + "epoch": 0.89, + "learning_rate": 8.826479438314945e-05, + "loss": 1.2238, + "theoretical_loss": 3.3543568043300693, + "tokens_seen": 2940534784 + }, + { + "epoch": 0.89, + "learning_rate": 8.824874623871615e-05, + "loss": 1.2783, + "theoretical_loss": 3.3543510820280056, + "tokens_seen": 2940600320 + }, + { + "epoch": 0.89, + "learning_rate": 8.823269809428285e-05, + "loss": 1.2975, + "theoretical_loss": 3.354345359889179, + "tokens_seen": 2940665856 + }, + { + "epoch": 0.89, + "learning_rate": 8.821664994984955e-05, + "loss": 1.3384, + "theoretical_loss": 3.354339637913581, + "tokens_seen": 2940731392 + }, + { + "epoch": 0.89, + "learning_rate": 8.820060180541625e-05, + "loss": 1.257, + "theoretical_loss": 3.3543339161012034, + "tokens_seen": 2940796928 + }, + { + "epoch": 0.89, + "learning_rate": 8.818455366098295e-05, + "loss": 1.3753, + "theoretical_loss": 3.354328194452038, + "tokens_seen": 2940862464 + }, + { + "epoch": 0.89, + "learning_rate": 8.816850551654965e-05, + "loss": 1.3815, + "theoretical_loss": 3.354322472966076, + "tokens_seen": 2940928000 + }, + { + "epoch": 0.89, + "learning_rate": 8.815245737211634e-05, + "loss": 1.2186, + "theoretical_loss": 3.35431675164331, + "tokens_seen": 2940993536 + }, + { + "epoch": 0.89, + "learning_rate": 8.813640922768306e-05, + "loss": 1.2595, + "theoretical_loss": 3.354311030483731, + "tokens_seen": 2941059072 + }, + { + "epoch": 0.89, + "learning_rate": 8.812036108324976e-05, + "loss": 1.3201, + "theoretical_loss": 3.354305309487331, + "tokens_seen": 2941124608 + }, + { + "epoch": 0.89, + "learning_rate": 8.810431293881646e-05, + "loss": 1.3163, + "theoretical_loss": 3.354299588654101, + "tokens_seen": 2941190144 + }, + { + "epoch": 0.89, + "learning_rate": 8.808826479438316e-05, + "loss": 1.1411, + "theoretical_loss": 3.354293867984034, + "tokens_seen": 2941255680 + }, + { + "epoch": 0.89, + "learning_rate": 8.807221664994986e-05, + "loss": 1.2941, + "theoretical_loss": 3.3542881474771207, + "tokens_seen": 2941321216 + }, + { + "epoch": 0.89, + "learning_rate": 8.805616850551656e-05, + "loss": 1.2667, + "theoretical_loss": 3.3542824271333536, + "tokens_seen": 2941386752 + }, + { + "epoch": 0.89, + "learning_rate": 8.804012036108326e-05, + "loss": 1.2963, + "theoretical_loss": 3.3542767069527235, + "tokens_seen": 2941452288 + }, + { + "epoch": 0.89, + "learning_rate": 8.802407221664996e-05, + "loss": 1.1983, + "theoretical_loss": 3.354270986935223, + "tokens_seen": 2941517824 + }, + { + "epoch": 0.89, + "learning_rate": 8.800802407221666e-05, + "loss": 1.3284, + "theoretical_loss": 3.354265267080843, + "tokens_seen": 2941583360 + }, + { + "epoch": 0.89, + "learning_rate": 8.799197592778335e-05, + "loss": 1.3289, + "theoretical_loss": 3.3542595473895758, + "tokens_seen": 2941648896 + }, + { + "epoch": 0.89, + "learning_rate": 8.797592778335005e-05, + "loss": 1.2142, + "theoretical_loss": 3.3542538278614127, + "tokens_seen": 2941714432 + }, + { + "epoch": 0.89, + "learning_rate": 8.795987963891675e-05, + "loss": 1.322, + "theoretical_loss": 3.354248108496346, + "tokens_seen": 2941779968 + }, + { + "epoch": 0.89, + "learning_rate": 8.794383149448345e-05, + "loss": 1.2057, + "theoretical_loss": 3.3542423892943667, + "tokens_seen": 2941845504 + }, + { + "epoch": 0.89, + "learning_rate": 8.792778335005015e-05, + "loss": 1.1375, + "theoretical_loss": 3.354236670255467, + "tokens_seen": 2941911040 + }, + { + "epoch": 0.89, + "learning_rate": 8.791173520561685e-05, + "loss": 1.2582, + "theoretical_loss": 3.3542309513796384, + "tokens_seen": 2941976576 + }, + { + "epoch": 0.89, + "learning_rate": 8.789568706118355e-05, + "loss": 1.3229, + "theoretical_loss": 3.354225232666873, + "tokens_seen": 2942042112 + }, + { + "epoch": 0.89, + "learning_rate": 8.787963891675025e-05, + "loss": 1.2369, + "theoretical_loss": 3.354219514117162, + "tokens_seen": 2942107648 + }, + { + "epoch": 0.89, + "learning_rate": 8.786359077231695e-05, + "loss": 1.2304, + "theoretical_loss": 3.354213795730497, + "tokens_seen": 2942173184 + }, + { + "epoch": 0.89, + "learning_rate": 8.784754262788365e-05, + "loss": 1.2388, + "theoretical_loss": 3.3542080775068706, + "tokens_seen": 2942238720 + }, + { + "epoch": 0.89, + "learning_rate": 8.783149448345035e-05, + "loss": 1.3692, + "theoretical_loss": 3.354202359446274, + "tokens_seen": 2942304256 + }, + { + "epoch": 0.89, + "learning_rate": 8.781544633901706e-05, + "loss": 1.3983, + "theoretical_loss": 3.3541966415486986, + "tokens_seen": 2942369792 + }, + { + "epoch": 0.89, + "learning_rate": 8.779939819458376e-05, + "loss": 1.2964, + "theoretical_loss": 3.3541909238141363, + "tokens_seen": 2942435328 + }, + { + "epoch": 0.89, + "learning_rate": 8.778335005015046e-05, + "loss": 1.2772, + "theoretical_loss": 3.354185206242579, + "tokens_seen": 2942500864 + }, + { + "epoch": 0.89, + "objective/train/docs_used": 1069745, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.4973185062408447, + "objective/train/theoretical_loss": 3.354179488834019, + "objective/train/tokens_used": 2963026400, + "theoretical_loss": 3.354179488834019, + "tokens_seen": 2942566400 + }, + { + "epoch": 0.89, + "learning_rate": 8.776730190571716e-05, + "loss": 1.4179, + "theoretical_loss": 3.354179488834019, + "tokens_seen": 2942566400 + }, + { + "epoch": 0.89, + "learning_rate": 8.775125376128386e-05, + "loss": 1.3111, + "theoretical_loss": 3.3541737715884468, + "tokens_seen": 2942631936 + }, + { + "epoch": 0.89, + "learning_rate": 8.773520561685056e-05, + "loss": 1.2672, + "theoretical_loss": 3.3541680545058545, + "tokens_seen": 2942697472 + }, + { + "epoch": 0.89, + "learning_rate": 8.771915747241726e-05, + "loss": 1.2722, + "theoretical_loss": 3.3541623375862346, + "tokens_seen": 2942763008 + }, + { + "epoch": 0.89, + "learning_rate": 8.770310932798396e-05, + "loss": 1.2917, + "theoretical_loss": 3.354156620829578, + "tokens_seen": 2942828544 + }, + { + "epoch": 0.89, + "learning_rate": 8.768706118355066e-05, + "loss": 1.3385, + "theoretical_loss": 3.3541509042358766, + "tokens_seen": 2942894080 + }, + { + "epoch": 0.89, + "learning_rate": 8.767101303911736e-05, + "loss": 1.2937, + "theoretical_loss": 3.3541451878051225, + "tokens_seen": 2942959616 + }, + { + "epoch": 0.89, + "learning_rate": 8.765496489468406e-05, + "loss": 1.3181, + "theoretical_loss": 3.354139471537307, + "tokens_seen": 2943025152 + }, + { + "epoch": 0.89, + "learning_rate": 8.763891675025075e-05, + "loss": 1.2637, + "theoretical_loss": 3.354133755432422, + "tokens_seen": 2943090688 + }, + { + "epoch": 0.89, + "learning_rate": 8.762286860581745e-05, + "loss": 1.3168, + "theoretical_loss": 3.354128039490459, + "tokens_seen": 2943156224 + }, + { + "epoch": 0.89, + "learning_rate": 8.760682046138415e-05, + "loss": 1.2311, + "theoretical_loss": 3.35412232371141, + "tokens_seen": 2943221760 + }, + { + "epoch": 0.89, + "learning_rate": 8.759077231695085e-05, + "loss": 1.2715, + "theoretical_loss": 3.354116608095267, + "tokens_seen": 2943287296 + }, + { + "epoch": 0.89, + "learning_rate": 8.757472417251755e-05, + "loss": 1.2445, + "theoretical_loss": 3.3541108926420207, + "tokens_seen": 2943352832 + }, + { + "epoch": 0.89, + "learning_rate": 8.755867602808425e-05, + "loss": 1.2647, + "theoretical_loss": 3.354105177351664, + "tokens_seen": 2943418368 + }, + { + "epoch": 0.89, + "learning_rate": 8.754262788365095e-05, + "loss": 1.2864, + "theoretical_loss": 3.354099462224188, + "tokens_seen": 2943483904 + }, + { + "epoch": 0.89, + "learning_rate": 8.752657973921767e-05, + "loss": 1.2245, + "theoretical_loss": 3.3540937472595846, + "tokens_seen": 2943549440 + }, + { + "epoch": 0.89, + "learning_rate": 8.751053159478437e-05, + "loss": 1.3268, + "theoretical_loss": 3.3540880324578453, + "tokens_seen": 2943614976 + }, + { + "epoch": 0.89, + "learning_rate": 8.749448345035107e-05, + "loss": 1.3912, + "theoretical_loss": 3.3540823178189623, + "tokens_seen": 2943680512 + }, + { + "epoch": 0.89, + "learning_rate": 8.747843530591775e-05, + "loss": 1.2849, + "theoretical_loss": 3.3540766033429272, + "tokens_seen": 2943746048 + }, + { + "epoch": 0.89, + "learning_rate": 8.746238716148446e-05, + "loss": 1.221, + "theoretical_loss": 3.3540708890297317, + "tokens_seen": 2943811584 + }, + { + "epoch": 0.89, + "learning_rate": 8.744633901705116e-05, + "loss": 1.2806, + "theoretical_loss": 3.3540651748793673, + "tokens_seen": 2943877120 + }, + { + "epoch": 0.89, + "learning_rate": 8.743029087261786e-05, + "loss": 1.4496, + "theoretical_loss": 3.3540594608918255, + "tokens_seen": 2943942656 + }, + { + "epoch": 0.89, + "learning_rate": 8.741424272818456e-05, + "loss": 1.2381, + "theoretical_loss": 3.354053747067099, + "tokens_seen": 2944008192 + }, + { + "epoch": 0.89, + "learning_rate": 8.739819458375126e-05, + "loss": 1.2845, + "theoretical_loss": 3.3540480334051783, + "tokens_seen": 2944073728 + }, + { + "epoch": 0.89, + "learning_rate": 8.738214643931796e-05, + "loss": 1.2763, + "theoretical_loss": 3.354042319906056, + "tokens_seen": 2944139264 + }, + { + "epoch": 0.89, + "learning_rate": 8.736609829488466e-05, + "loss": 1.3156, + "theoretical_loss": 3.354036606569724, + "tokens_seen": 2944204800 + }, + { + "epoch": 0.89, + "learning_rate": 8.735005015045136e-05, + "loss": 1.2796, + "theoretical_loss": 3.3540308933961733, + "tokens_seen": 2944270336 + }, + { + "epoch": 0.89, + "learning_rate": 8.733400200601806e-05, + "loss": 1.3248, + "theoretical_loss": 3.3540251803853964, + "tokens_seen": 2944335872 + }, + { + "epoch": 0.89, + "learning_rate": 8.731795386158476e-05, + "loss": 1.4216, + "theoretical_loss": 3.354019467537384, + "tokens_seen": 2944401408 + }, + { + "epoch": 0.89, + "learning_rate": 8.730190571715146e-05, + "loss": 1.2812, + "theoretical_loss": 3.354013754852129, + "tokens_seen": 2944466944 + }, + { + "epoch": 0.89, + "learning_rate": 8.728585757271815e-05, + "loss": 1.2883, + "theoretical_loss": 3.354008042329623, + "tokens_seen": 2944532480 + }, + { + "epoch": 0.89, + "learning_rate": 8.726980942828485e-05, + "loss": 1.2628, + "theoretical_loss": 3.3540023299698563, + "tokens_seen": 2944598016 + }, + { + "epoch": 0.89, + "learning_rate": 8.725376128385155e-05, + "loss": 1.2122, + "theoretical_loss": 3.353996617772822, + "tokens_seen": 2944663552 + }, + { + "epoch": 0.89, + "learning_rate": 8.723771313941825e-05, + "loss": 1.1831, + "theoretical_loss": 3.353990905738512, + "tokens_seen": 2944729088 + }, + { + "epoch": 0.89, + "learning_rate": 8.722166499498496e-05, + "loss": 1.272, + "theoretical_loss": 3.353985193866918, + "tokens_seen": 2944794624 + }, + { + "epoch": 0.89, + "learning_rate": 8.720561685055167e-05, + "loss": 1.4644, + "theoretical_loss": 3.3539794821580307, + "tokens_seen": 2944860160 + }, + { + "epoch": 0.89, + "learning_rate": 8.718956870611837e-05, + "loss": 1.3197, + "theoretical_loss": 3.353973770611842, + "tokens_seen": 2944925696 + }, + { + "epoch": 0.89, + "learning_rate": 8.717352056168507e-05, + "loss": 1.2369, + "theoretical_loss": 3.353968059228345, + "tokens_seen": 2944991232 + }, + { + "epoch": 0.89, + "learning_rate": 8.715747241725177e-05, + "loss": 1.2974, + "theoretical_loss": 3.35396234800753, + "tokens_seen": 2945056768 + }, + { + "epoch": 0.89, + "learning_rate": 8.714142427281847e-05, + "loss": 1.3558, + "theoretical_loss": 3.3539566369493894, + "tokens_seen": 2945122304 + }, + { + "epoch": 0.89, + "learning_rate": 8.712537612838516e-05, + "loss": 1.2452, + "theoretical_loss": 3.3539509260539155, + "tokens_seen": 2945187840 + }, + { + "epoch": 0.89, + "learning_rate": 8.710932798395186e-05, + "loss": 1.325, + "theoretical_loss": 3.3539452153210987, + "tokens_seen": 2945253376 + }, + { + "epoch": 0.89, + "learning_rate": 8.709327983951856e-05, + "loss": 1.2388, + "theoretical_loss": 3.3539395047509317, + "tokens_seen": 2945318912 + }, + { + "epoch": 0.89, + "learning_rate": 8.707723169508526e-05, + "loss": 1.2638, + "theoretical_loss": 3.3539337943434058, + "tokens_seen": 2945384448 + }, + { + "epoch": 0.89, + "learning_rate": 8.706118355065196e-05, + "loss": 1.3434, + "theoretical_loss": 3.3539280840985133, + "tokens_seen": 2945449984 + }, + { + "epoch": 0.89, + "learning_rate": 8.704513540621866e-05, + "loss": 1.3742, + "theoretical_loss": 3.3539223740162454, + "tokens_seen": 2945515520 + }, + { + "epoch": 0.89, + "learning_rate": 8.702908726178536e-05, + "loss": 1.2994, + "theoretical_loss": 3.353916664096594, + "tokens_seen": 2945581056 + }, + { + "epoch": 0.89, + "learning_rate": 8.701303911735206e-05, + "loss": 1.2039, + "theoretical_loss": 3.3539109543395513, + "tokens_seen": 2945646592 + }, + { + "epoch": 0.89, + "learning_rate": 8.699699097291876e-05, + "loss": 1.3656, + "theoretical_loss": 3.3539052447451083, + "tokens_seen": 2945712128 + }, + { + "epoch": 0.89, + "learning_rate": 8.698094282848546e-05, + "loss": 1.3505, + "theoretical_loss": 3.353899535313257, + "tokens_seen": 2945777664 + }, + { + "epoch": 0.89, + "objective/train/docs_used": 1070956, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.227071762084961, + "objective/train/theoretical_loss": 3.35389382604399, + "objective/train/tokens_used": 2966303200, + "theoretical_loss": 3.35389382604399, + "tokens_seen": 2945843200 + }, + { + "epoch": 0.89, + "learning_rate": 8.696489468405216e-05, + "loss": 1.2119, + "theoretical_loss": 3.35389382604399, + "tokens_seen": 2945843200 + }, + { + "epoch": 0.89, + "learning_rate": 8.694884653961886e-05, + "loss": 1.2959, + "theoretical_loss": 3.3538881169372976, + "tokens_seen": 2945908736 + }, + { + "epoch": 0.89, + "learning_rate": 8.693279839518556e-05, + "loss": 1.2791, + "theoretical_loss": 3.353882407993172, + "tokens_seen": 2945974272 + }, + { + "epoch": 0.89, + "learning_rate": 8.691675025075226e-05, + "loss": 1.3061, + "theoretical_loss": 3.353876699211606, + "tokens_seen": 2946039808 + }, + { + "epoch": 0.89, + "learning_rate": 8.690070210631896e-05, + "loss": 1.3124, + "theoretical_loss": 3.3538709905925903, + "tokens_seen": 2946105344 + }, + { + "epoch": 0.89, + "learning_rate": 8.688465396188566e-05, + "loss": 1.1453, + "theoretical_loss": 3.353865282136117, + "tokens_seen": 2946170880 + }, + { + "epoch": 0.89, + "learning_rate": 8.686860581745237e-05, + "loss": 1.3295, + "theoretical_loss": 3.3538595738421777, + "tokens_seen": 2946236416 + }, + { + "epoch": 0.89, + "learning_rate": 8.685255767301907e-05, + "loss": 1.3069, + "theoretical_loss": 3.3538538657107644, + "tokens_seen": 2946301952 + }, + { + "epoch": 0.89, + "learning_rate": 8.683650952858577e-05, + "loss": 1.3948, + "theoretical_loss": 3.3538481577418686, + "tokens_seen": 2946367488 + }, + { + "epoch": 0.89, + "learning_rate": 8.682046138415247e-05, + "loss": 1.3305, + "theoretical_loss": 3.353842449935482, + "tokens_seen": 2946433024 + }, + { + "epoch": 0.89, + "learning_rate": 8.680441323971917e-05, + "loss": 1.1795, + "theoretical_loss": 3.3538367422915965, + "tokens_seen": 2946498560 + }, + { + "epoch": 0.89, + "learning_rate": 8.678836509528587e-05, + "loss": 1.2946, + "theoretical_loss": 3.353831034810204, + "tokens_seen": 2946564096 + }, + { + "epoch": 0.89, + "learning_rate": 8.677231695085256e-05, + "loss": 1.3453, + "theoretical_loss": 3.3538253274912964, + "tokens_seen": 2946629632 + }, + { + "epoch": 0.89, + "learning_rate": 8.675626880641926e-05, + "loss": 1.2359, + "theoretical_loss": 3.353819620334865, + "tokens_seen": 2946695168 + }, + { + "epoch": 0.89, + "learning_rate": 8.674022066198596e-05, + "loss": 1.26, + "theoretical_loss": 3.3538139133409017, + "tokens_seen": 2946760704 + }, + { + "epoch": 0.89, + "learning_rate": 8.672417251755266e-05, + "loss": 1.229, + "theoretical_loss": 3.3538082065093984, + "tokens_seen": 2946826240 + }, + { + "epoch": 0.89, + "learning_rate": 8.670812437311936e-05, + "loss": 1.2907, + "theoretical_loss": 3.353802499840347, + "tokens_seen": 2946891776 + }, + { + "epoch": 0.89, + "learning_rate": 8.669207622868606e-05, + "loss": 1.4433, + "theoretical_loss": 3.3537967933337387, + "tokens_seen": 2946957312 + }, + { + "epoch": 0.89, + "learning_rate": 8.667602808425276e-05, + "loss": 1.2972, + "theoretical_loss": 3.353791086989566, + "tokens_seen": 2947022848 + }, + { + "epoch": 0.89, + "learning_rate": 8.665997993981946e-05, + "loss": 1.2953, + "theoretical_loss": 3.3537853808078197, + "tokens_seen": 2947088384 + }, + { + "epoch": 0.89, + "learning_rate": 8.664393179538616e-05, + "loss": 1.2445, + "theoretical_loss": 3.353779674788493, + "tokens_seen": 2947153920 + }, + { + "epoch": 0.89, + "learning_rate": 8.662788365095286e-05, + "loss": 1.1914, + "theoretical_loss": 3.353773968931576, + "tokens_seen": 2947219456 + }, + { + "epoch": 0.89, + "learning_rate": 8.661183550651956e-05, + "loss": 1.4041, + "theoretical_loss": 3.3537682632370616, + "tokens_seen": 2947284992 + }, + { + "epoch": 0.89, + "learning_rate": 8.659578736208626e-05, + "loss": 1.2353, + "theoretical_loss": 3.3537625577049415, + "tokens_seen": 2947350528 + }, + { + "epoch": 0.89, + "learning_rate": 8.657973921765296e-05, + "loss": 1.3047, + "theoretical_loss": 3.3537568523352066, + "tokens_seen": 2947416064 + }, + { + "epoch": 0.89, + "learning_rate": 8.656369107321966e-05, + "loss": 1.2009, + "theoretical_loss": 3.35375114712785, + "tokens_seen": 2947481600 + }, + { + "epoch": 0.89, + "learning_rate": 8.654764292878636e-05, + "loss": 1.322, + "theoretical_loss": 3.353745442082862, + "tokens_seen": 2947547136 + }, + { + "epoch": 0.89, + "learning_rate": 8.653159478435307e-05, + "loss": 1.2949, + "theoretical_loss": 3.3537397372002355, + "tokens_seen": 2947612672 + }, + { + "epoch": 0.89, + "learning_rate": 8.651554663991977e-05, + "loss": 1.3527, + "theoretical_loss": 3.353734032479962, + "tokens_seen": 2947678208 + }, + { + "epoch": 0.89, + "learning_rate": 8.649949849548647e-05, + "loss": 1.271, + "theoretical_loss": 3.353728327922033, + "tokens_seen": 2947743744 + }, + { + "epoch": 0.89, + "learning_rate": 8.648345035105317e-05, + "loss": 1.1752, + "theoretical_loss": 3.3537226235264406, + "tokens_seen": 2947809280 + }, + { + "epoch": 0.89, + "learning_rate": 8.646740220661987e-05, + "loss": 1.2545, + "theoretical_loss": 3.353716919293176, + "tokens_seen": 2947874816 + }, + { + "epoch": 0.89, + "learning_rate": 8.645135406218657e-05, + "loss": 1.3169, + "theoretical_loss": 3.3537112152222317, + "tokens_seen": 2947940352 + }, + { + "epoch": 0.89, + "learning_rate": 8.643530591775326e-05, + "loss": 1.2415, + "theoretical_loss": 3.353705511313599, + "tokens_seen": 2948005888 + }, + { + "epoch": 0.89, + "learning_rate": 8.641925777331996e-05, + "loss": 1.3008, + "theoretical_loss": 3.35369980756727, + "tokens_seen": 2948071424 + }, + { + "epoch": 0.89, + "learning_rate": 8.640320962888666e-05, + "loss": 1.1987, + "theoretical_loss": 3.3536941039832366, + "tokens_seen": 2948136960 + }, + { + "epoch": 0.89, + "learning_rate": 8.638716148445336e-05, + "loss": 1.1927, + "theoretical_loss": 3.3536884005614898, + "tokens_seen": 2948202496 + }, + { + "epoch": 0.89, + "learning_rate": 8.637111334002006e-05, + "loss": 1.2083, + "theoretical_loss": 3.353682697302022, + "tokens_seen": 2948268032 + }, + { + "epoch": 0.89, + "learning_rate": 8.635506519558676e-05, + "loss": 1.3412, + "theoretical_loss": 3.3536769942048243, + "tokens_seen": 2948333568 + }, + { + "epoch": 0.89, + "learning_rate": 8.633901705115346e-05, + "loss": 1.1583, + "theoretical_loss": 3.35367129126989, + "tokens_seen": 2948399104 + }, + { + "epoch": 0.89, + "learning_rate": 8.632296890672017e-05, + "loss": 1.2359, + "theoretical_loss": 3.353665588497209, + "tokens_seen": 2948464640 + }, + { + "epoch": 0.89, + "learning_rate": 8.630692076228687e-05, + "loss": 1.249, + "theoretical_loss": 3.353659885886774, + "tokens_seen": 2948530176 + }, + { + "epoch": 0.89, + "learning_rate": 8.629087261785357e-05, + "loss": 1.304, + "theoretical_loss": 3.3536541834385774, + "tokens_seen": 2948595712 + }, + { + "epoch": 0.89, + "learning_rate": 8.627482447342026e-05, + "loss": 1.3189, + "theoretical_loss": 3.35364848115261, + "tokens_seen": 2948661248 + }, + { + "epoch": 0.89, + "learning_rate": 8.625877632898696e-05, + "loss": 1.2478, + "theoretical_loss": 3.3536427790288634, + "tokens_seen": 2948726784 + }, + { + "epoch": 0.89, + "learning_rate": 8.624272818455366e-05, + "loss": 1.2541, + "theoretical_loss": 3.3536370770673303, + "tokens_seen": 2948792320 + }, + { + "epoch": 0.89, + "learning_rate": 8.622668004012036e-05, + "loss": 1.2938, + "theoretical_loss": 3.353631375268002, + "tokens_seen": 2948857856 + }, + { + "epoch": 0.89, + "learning_rate": 8.621063189568706e-05, + "loss": 1.1918, + "theoretical_loss": 3.3536256736308703, + "tokens_seen": 2948923392 + }, + { + "epoch": 0.89, + "learning_rate": 8.619458375125376e-05, + "loss": 1.3198, + "theoretical_loss": 3.353619972155927, + "tokens_seen": 2948988928 + }, + { + "epoch": 0.89, + "learning_rate": 8.617853560682047e-05, + "loss": 1.3742, + "theoretical_loss": 3.3536142708431638, + "tokens_seen": 2949054464 + }, + { + "debugging/Compilability": 1.0, + "debugging/distinct-1-grams": 0.7681721806430071, + "debugging/entropy-1-grams": 4.8521974401642165, + "debugging/length": 515.7, + "debugging/num_segments": 10, + "debugging/score": 0.004562939445799932, + "debugging/score_std": 0.004018777249710917, + "epoch": 0.89, + "objective/train/docs_used": 1072071, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.1408500671386719, + "objective/train/theoretical_loss": 3.3536085696925726, + "objective/train/tokens_used": 2969580000, + "theoretical_loss": 3.3536085696925726, + "tokens_seen": 2949120000 + }, + { + "epoch": 0.89, + "learning_rate": 8.616248746238717e-05, + "loss": 1.1661, + "theoretical_loss": 3.3536085696925726, + "tokens_seen": 2949120000 + }, + { + "epoch": 0.89, + "learning_rate": 8.614643931795387e-05, + "loss": 1.2206, + "theoretical_loss": 3.3536028687041455, + "tokens_seen": 2949185536 + }, + { + "epoch": 0.89, + "learning_rate": 8.613039117352057e-05, + "loss": 1.1856, + "theoretical_loss": 3.3535971678778735, + "tokens_seen": 2949251072 + }, + { + "epoch": 0.89, + "learning_rate": 8.611434302908727e-05, + "loss": 1.1799, + "theoretical_loss": 3.3535914672137492, + "tokens_seen": 2949316608 + }, + { + "epoch": 0.89, + "learning_rate": 8.609829488465397e-05, + "loss": 1.241, + "theoretical_loss": 3.3535857667117637, + "tokens_seen": 2949382144 + }, + { + "epoch": 0.89, + "learning_rate": 8.608224674022066e-05, + "loss": 1.251, + "theoretical_loss": 3.353580066371909, + "tokens_seen": 2949447680 + }, + { + "epoch": 0.89, + "learning_rate": 8.606619859578736e-05, + "loss": 1.3013, + "theoretical_loss": 3.3535743661941773, + "tokens_seen": 2949513216 + }, + { + "epoch": 0.89, + "learning_rate": 8.605015045135406e-05, + "loss": 1.3319, + "theoretical_loss": 3.35356866617856, + "tokens_seen": 2949578752 + }, + { + "epoch": 0.89, + "learning_rate": 8.603410230692077e-05, + "loss": 1.2944, + "theoretical_loss": 3.3535629663250486, + "tokens_seen": 2949644288 + }, + { + "epoch": 0.89, + "learning_rate": 8.601805416248747e-05, + "loss": 1.2012, + "theoretical_loss": 3.3535572666336355, + "tokens_seen": 2949709824 + }, + { + "epoch": 0.89, + "learning_rate": 8.600200601805417e-05, + "loss": 1.216, + "theoretical_loss": 3.3535515671043123, + "tokens_seen": 2949775360 + }, + { + "epoch": 0.89, + "learning_rate": 8.598595787362087e-05, + "loss": 1.236, + "theoretical_loss": 3.353545867737071, + "tokens_seen": 2949840896 + }, + { + "epoch": 0.89, + "learning_rate": 8.596990972918757e-05, + "loss": 1.1853, + "theoretical_loss": 3.353540168531903, + "tokens_seen": 2949906432 + }, + { + "epoch": 0.89, + "learning_rate": 8.595386158475427e-05, + "loss": 1.2698, + "theoretical_loss": 3.3535344694887996, + "tokens_seen": 2949971968 + }, + { + "epoch": 0.89, + "learning_rate": 8.593781344032098e-05, + "loss": 1.2932, + "theoretical_loss": 3.353528770607754, + "tokens_seen": 2950037504 + }, + { + "epoch": 0.89, + "learning_rate": 8.592176529588766e-05, + "loss": 1.3257, + "theoretical_loss": 3.3535230718887563, + "tokens_seen": 2950103040 + }, + { + "epoch": 0.89, + "learning_rate": 8.590571715145436e-05, + "loss": 1.1668, + "theoretical_loss": 3.3535173733318, + "tokens_seen": 2950168576 + }, + { + "epoch": 0.89, + "learning_rate": 8.588966900702106e-05, + "loss": 1.2431, + "theoretical_loss": 3.3535116749368754, + "tokens_seen": 2950234112 + }, + { + "epoch": 0.89, + "learning_rate": 8.587362086258776e-05, + "loss": 1.274, + "theoretical_loss": 3.3535059767039757, + "tokens_seen": 2950299648 + }, + { + "epoch": 0.89, + "learning_rate": 8.585757271815446e-05, + "loss": 1.2933, + "theoretical_loss": 3.353500278633091, + "tokens_seen": 2950365184 + }, + { + "epoch": 0.89, + "learning_rate": 8.584152457372117e-05, + "loss": 1.2634, + "theoretical_loss": 3.353494580724215, + "tokens_seen": 2950430720 + }, + { + "epoch": 0.89, + "learning_rate": 8.582547642928787e-05, + "loss": 1.3332, + "theoretical_loss": 3.353488882977338, + "tokens_seen": 2950496256 + }, + { + "epoch": 0.89, + "learning_rate": 8.580942828485457e-05, + "loss": 1.1069, + "theoretical_loss": 3.3534831853924527, + "tokens_seen": 2950561792 + }, + { + "epoch": 0.89, + "learning_rate": 8.579338014042127e-05, + "loss": 1.2131, + "theoretical_loss": 3.3534774879695504, + "tokens_seen": 2950627328 + }, + { + "epoch": 0.89, + "learning_rate": 8.577733199598797e-05, + "loss": 1.2663, + "theoretical_loss": 3.353471790708623, + "tokens_seen": 2950692864 + }, + { + "epoch": 0.89, + "learning_rate": 8.576128385155467e-05, + "loss": 1.3387, + "theoretical_loss": 3.3534660936096623, + "tokens_seen": 2950758400 + }, + { + "epoch": 0.89, + "learning_rate": 8.574523570712137e-05, + "loss": 1.3384, + "theoretical_loss": 3.35346039667266, + "tokens_seen": 2950823936 + }, + { + "epoch": 0.89, + "learning_rate": 8.572918756268807e-05, + "loss": 1.2886, + "theoretical_loss": 3.3534546998976085, + "tokens_seen": 2950889472 + }, + { + "epoch": 0.89, + "learning_rate": 8.571313941825477e-05, + "loss": 1.2055, + "theoretical_loss": 3.3534490032844984, + "tokens_seen": 2950955008 + }, + { + "epoch": 0.89, + "learning_rate": 8.569709127382147e-05, + "loss": 1.1926, + "theoretical_loss": 3.3534433068333227, + "tokens_seen": 2951020544 + }, + { + "epoch": 0.89, + "learning_rate": 8.568104312938817e-05, + "loss": 1.26, + "theoretical_loss": 3.3534376105440726, + "tokens_seen": 2951086080 + }, + { + "epoch": 0.89, + "learning_rate": 8.566499498495487e-05, + "loss": 1.1785, + "theoretical_loss": 3.35343191441674, + "tokens_seen": 2951151616 + }, + { + "epoch": 0.89, + "learning_rate": 8.564894684052157e-05, + "loss": 1.2488, + "theoretical_loss": 3.3534262184513173, + "tokens_seen": 2951217152 + }, + { + "epoch": 0.89, + "learning_rate": 8.563289869608827e-05, + "loss": 1.2188, + "theoretical_loss": 3.353420522647795, + "tokens_seen": 2951282688 + }, + { + "epoch": 0.89, + "learning_rate": 8.561685055165497e-05, + "loss": 1.2848, + "theoretical_loss": 3.353414827006166, + "tokens_seen": 2951348224 + }, + { + "epoch": 0.89, + "learning_rate": 8.560080240722167e-05, + "loss": 1.3005, + "theoretical_loss": 3.353409131526422, + "tokens_seen": 2951413760 + }, + { + "epoch": 0.89, + "learning_rate": 8.558475426278838e-05, + "loss": 1.2086, + "theoretical_loss": 3.353403436208554, + "tokens_seen": 2951479296 + }, + { + "epoch": 0.89, + "learning_rate": 8.556870611835506e-05, + "loss": 1.3196, + "theoretical_loss": 3.3533977410525546, + "tokens_seen": 2951544832 + }, + { + "epoch": 0.89, + "learning_rate": 8.555265797392176e-05, + "loss": 1.1741, + "theoretical_loss": 3.3533920460584152, + "tokens_seen": 2951610368 + }, + { + "epoch": 0.89, + "learning_rate": 8.553660982948846e-05, + "loss": 1.2204, + "theoretical_loss": 3.353386351226128, + "tokens_seen": 2951675904 + }, + { + "epoch": 0.89, + "learning_rate": 8.552056168505516e-05, + "loss": 1.2621, + "theoretical_loss": 3.3533806565556845, + "tokens_seen": 2951741440 + }, + { + "epoch": 0.89, + "learning_rate": 8.550451354062187e-05, + "loss": 1.3438, + "theoretical_loss": 3.3533749620470767, + "tokens_seen": 2951806976 + }, + { + "epoch": 0.89, + "learning_rate": 8.548846539618857e-05, + "loss": 1.2377, + "theoretical_loss": 3.3533692677002964, + "tokens_seen": 2951872512 + }, + { + "epoch": 0.89, + "learning_rate": 8.547241725175527e-05, + "loss": 1.3554, + "theoretical_loss": 3.3533635735153347, + "tokens_seen": 2951938048 + }, + { + "epoch": 0.89, + "learning_rate": 8.545636910732197e-05, + "loss": 1.3652, + "theoretical_loss": 3.3533578794921843, + "tokens_seen": 2952003584 + }, + { + "epoch": 0.89, + "learning_rate": 8.544032096288867e-05, + "loss": 1.1665, + "theoretical_loss": 3.353352185630837, + "tokens_seen": 2952069120 + }, + { + "epoch": 0.89, + "learning_rate": 8.542427281845538e-05, + "loss": 1.3419, + "theoretical_loss": 3.3533464919312843, + "tokens_seen": 2952134656 + }, + { + "epoch": 0.89, + "learning_rate": 8.540822467402207e-05, + "loss": 1.3298, + "theoretical_loss": 3.3533407983935177, + "tokens_seen": 2952200192 + }, + { + "epoch": 0.89, + "learning_rate": 8.539217652958877e-05, + "loss": 1.1383, + "theoretical_loss": 3.3533351050175293, + "tokens_seen": 2952265728 + }, + { + "epoch": 0.89, + "learning_rate": 8.537612838515547e-05, + "loss": 1.4098, + "theoretical_loss": 3.3533294118033115, + "tokens_seen": 2952331264 + }, + { + "epoch": 0.89, + "objective/train/docs_used": 1073247, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.3161585330963135, + "objective/train/theoretical_loss": 3.353323718750855, + "objective/train/tokens_used": 2972856800, + "theoretical_loss": 3.353323718750855, + "tokens_seen": 2952396800 + }, + { + "epoch": 0.89, + "learning_rate": 8.536008024072217e-05, + "loss": 1.2165, + "theoretical_loss": 3.353323718750855, + "tokens_seen": 2952396800 + }, + { + "epoch": 0.89, + "learning_rate": 8.534403209628887e-05, + "loss": 1.2549, + "theoretical_loss": 3.3533180258601525, + "tokens_seen": 2952462336 + }, + { + "epoch": 0.89, + "learning_rate": 8.532798395185557e-05, + "loss": 1.2803, + "theoretical_loss": 3.3533123331311954, + "tokens_seen": 2952527872 + }, + { + "epoch": 0.89, + "learning_rate": 8.531193580742227e-05, + "loss": 1.1739, + "theoretical_loss": 3.3533066405639755, + "tokens_seen": 2952593408 + }, + { + "epoch": 0.89, + "learning_rate": 8.529588766298897e-05, + "loss": 1.3443, + "theoretical_loss": 3.353300948158485, + "tokens_seen": 2952658944 + }, + { + "epoch": 0.89, + "learning_rate": 8.527983951855567e-05, + "loss": 1.2311, + "theoretical_loss": 3.3532952559147153, + "tokens_seen": 2952724480 + }, + { + "epoch": 0.89, + "learning_rate": 8.526379137412237e-05, + "loss": 1.1786, + "theoretical_loss": 3.3532895638326585, + "tokens_seen": 2952790016 + }, + { + "epoch": 0.89, + "learning_rate": 8.524774322968908e-05, + "loss": 1.314, + "theoretical_loss": 3.353283871912306, + "tokens_seen": 2952855552 + }, + { + "epoch": 0.89, + "learning_rate": 8.523169508525578e-05, + "loss": 1.2623, + "theoretical_loss": 3.35327818015365, + "tokens_seen": 2952921088 + }, + { + "epoch": 0.89, + "learning_rate": 8.521564694082246e-05, + "loss": 1.2439, + "theoretical_loss": 3.3532724885566827, + "tokens_seen": 2952986624 + }, + { + "epoch": 0.89, + "learning_rate": 8.519959879638916e-05, + "loss": 1.2946, + "theoretical_loss": 3.3532667971213947, + "tokens_seen": 2953052160 + }, + { + "epoch": 0.89, + "learning_rate": 8.518355065195586e-05, + "loss": 1.234, + "theoretical_loss": 3.353261105847779, + "tokens_seen": 2953117696 + }, + { + "epoch": 0.89, + "learning_rate": 8.516750250752256e-05, + "loss": 1.2343, + "theoretical_loss": 3.353255414735827, + "tokens_seen": 2953183232 + }, + { + "epoch": 0.89, + "learning_rate": 8.515145436308927e-05, + "loss": 1.3433, + "theoretical_loss": 3.35324972378553, + "tokens_seen": 2953248768 + }, + { + "epoch": 0.89, + "learning_rate": 8.513540621865597e-05, + "loss": 1.3145, + "theoretical_loss": 3.353244032996881, + "tokens_seen": 2953314304 + }, + { + "epoch": 0.89, + "learning_rate": 8.511935807422268e-05, + "loss": 1.1355, + "theoretical_loss": 3.3532383423698704, + "tokens_seen": 2953379840 + }, + { + "epoch": 0.89, + "learning_rate": 8.510330992978938e-05, + "loss": 1.2328, + "theoretical_loss": 3.3532326519044915, + "tokens_seen": 2953445376 + }, + { + "epoch": 0.9, + "learning_rate": 8.508726178535608e-05, + "loss": 1.1932, + "theoretical_loss": 3.353226961600735, + "tokens_seen": 2953510912 + }, + { + "epoch": 0.9, + "learning_rate": 8.507121364092278e-05, + "loss": 1.2634, + "theoretical_loss": 3.353221271458593, + "tokens_seen": 2953576448 + }, + { + "epoch": 0.9, + "learning_rate": 8.505516549648947e-05, + "loss": 1.2996, + "theoretical_loss": 3.3532155814780573, + "tokens_seen": 2953641984 + }, + { + "epoch": 0.9, + "learning_rate": 8.503911735205617e-05, + "loss": 1.2342, + "theoretical_loss": 3.3532098916591204, + "tokens_seen": 2953707520 + }, + { + "epoch": 0.9, + "learning_rate": 8.502306920762287e-05, + "loss": 1.2245, + "theoretical_loss": 3.353204202001773, + "tokens_seen": 2953773056 + }, + { + "epoch": 0.9, + "learning_rate": 8.500702106318957e-05, + "loss": 1.2785, + "theoretical_loss": 3.3531985125060078, + "tokens_seen": 2953838592 + }, + { + "epoch": 0.9, + "learning_rate": 8.499097291875627e-05, + "loss": 1.2147, + "theoretical_loss": 3.3531928231718164, + "tokens_seen": 2953904128 + }, + { + "epoch": 0.9, + "learning_rate": 8.497492477432297e-05, + "loss": 1.3305, + "theoretical_loss": 3.3531871339991906, + "tokens_seen": 2953969664 + }, + { + "epoch": 0.9, + "learning_rate": 8.495887662988967e-05, + "loss": 1.2266, + "theoretical_loss": 3.353181444988122, + "tokens_seen": 2954035200 + }, + { + "epoch": 0.9, + "learning_rate": 8.494282848545637e-05, + "loss": 1.313, + "theoretical_loss": 3.3531757561386026, + "tokens_seen": 2954100736 + }, + { + "epoch": 0.9, + "learning_rate": 8.492678034102307e-05, + "loss": 1.2729, + "theoretical_loss": 3.353170067450624, + "tokens_seen": 2954166272 + }, + { + "epoch": 0.9, + "learning_rate": 8.491073219658977e-05, + "loss": 1.1707, + "theoretical_loss": 3.3531643789241787, + "tokens_seen": 2954231808 + }, + { + "epoch": 0.9, + "learning_rate": 8.489468405215648e-05, + "loss": 1.2214, + "theoretical_loss": 3.3531586905592583, + "tokens_seen": 2954297344 + }, + { + "epoch": 0.9, + "learning_rate": 8.487863590772318e-05, + "loss": 1.2812, + "theoretical_loss": 3.353153002355854, + "tokens_seen": 2954362880 + }, + { + "epoch": 0.9, + "learning_rate": 8.486258776328986e-05, + "loss": 1.2691, + "theoretical_loss": 3.3531473143139583, + "tokens_seen": 2954428416 + }, + { + "epoch": 0.9, + "learning_rate": 8.484653961885656e-05, + "loss": 1.1923, + "theoretical_loss": 3.3531416264335627, + "tokens_seen": 2954493952 + }, + { + "epoch": 0.9, + "learning_rate": 8.483049147442328e-05, + "loss": 1.2669, + "theoretical_loss": 3.353135938714659, + "tokens_seen": 2954559488 + }, + { + "epoch": 0.9, + "learning_rate": 8.481444332998998e-05, + "loss": 1.1763, + "theoretical_loss": 3.353130251157239, + "tokens_seen": 2954625024 + }, + { + "epoch": 0.9, + "learning_rate": 8.479839518555668e-05, + "loss": 1.2983, + "theoretical_loss": 3.353124563761295, + "tokens_seen": 2954690560 + }, + { + "epoch": 0.9, + "learning_rate": 8.478234704112338e-05, + "loss": 1.1533, + "theoretical_loss": 3.3531188765268185, + "tokens_seen": 2954756096 + }, + { + "epoch": 0.9, + "learning_rate": 8.476629889669008e-05, + "loss": 1.2564, + "theoretical_loss": 3.3531131894538015, + "tokens_seen": 2954821632 + }, + { + "epoch": 0.9, + "learning_rate": 8.475025075225678e-05, + "loss": 1.1834, + "theoretical_loss": 3.353107502542236, + "tokens_seen": 2954887168 + }, + { + "epoch": 0.9, + "learning_rate": 8.473420260782348e-05, + "loss": 1.2032, + "theoretical_loss": 3.3531018157921126, + "tokens_seen": 2954952704 + }, + { + "epoch": 0.9, + "learning_rate": 8.471815446339017e-05, + "loss": 1.296, + "theoretical_loss": 3.3530961292034247, + "tokens_seen": 2955018240 + }, + { + "epoch": 0.9, + "learning_rate": 8.470210631895687e-05, + "loss": 1.2677, + "theoretical_loss": 3.353090442776163, + "tokens_seen": 2955083776 + }, + { + "epoch": 0.9, + "learning_rate": 8.468605817452357e-05, + "loss": 1.2488, + "theoretical_loss": 3.3530847565103206, + "tokens_seen": 2955149312 + }, + { + "epoch": 0.9, + "learning_rate": 8.467001003009027e-05, + "loss": 1.2274, + "theoretical_loss": 3.3530790704058884, + "tokens_seen": 2955214848 + }, + { + "epoch": 0.9, + "learning_rate": 8.465396188565697e-05, + "loss": 1.273, + "theoretical_loss": 3.3530733844628577, + "tokens_seen": 2955280384 + }, + { + "epoch": 0.9, + "learning_rate": 8.463791374122367e-05, + "loss": 1.2776, + "theoretical_loss": 3.3530676986812216, + "tokens_seen": 2955345920 + }, + { + "epoch": 0.9, + "learning_rate": 8.462186559679037e-05, + "loss": 1.2223, + "theoretical_loss": 3.3530620130609714, + "tokens_seen": 2955411456 + }, + { + "epoch": 0.9, + "learning_rate": 8.460581745235707e-05, + "loss": 1.2017, + "theoretical_loss": 3.353056327602099, + "tokens_seen": 2955476992 + }, + { + "epoch": 0.9, + "learning_rate": 8.458976930792377e-05, + "loss": 1.3134, + "theoretical_loss": 3.3530506423045963, + "tokens_seen": 2955542528 + }, + { + "epoch": 0.9, + "learning_rate": 8.457372116349047e-05, + "loss": 1.3187, + "theoretical_loss": 3.3530449571684544, + "tokens_seen": 2955608064 + }, + { + "epoch": 0.9, + "objective/train/docs_used": 1074453, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.3133764266967773, + "objective/train/theoretical_loss": 3.3530392721936666, + "objective/train/tokens_used": 2976133600, + "theoretical_loss": 3.3530392721936666, + "tokens_seen": 2955673600 + }, + { + "epoch": 0.9, + "learning_rate": 8.455767301905718e-05, + "loss": 1.3577, + "theoretical_loss": 3.3530392721936666, + "tokens_seen": 2955673600 + }, + { + "epoch": 0.9, + "learning_rate": 8.454162487462388e-05, + "loss": 1.3055, + "theoretical_loss": 3.3530335873802235, + "tokens_seen": 2955739136 + }, + { + "epoch": 0.9, + "learning_rate": 8.452557673019058e-05, + "loss": 1.1997, + "theoretical_loss": 3.353027902728117, + "tokens_seen": 2955804672 + }, + { + "epoch": 0.9, + "learning_rate": 8.450952858575728e-05, + "loss": 1.2771, + "theoretical_loss": 3.3530222182373395, + "tokens_seen": 2955870208 + }, + { + "epoch": 0.9, + "learning_rate": 8.449348044132398e-05, + "loss": 1.2927, + "theoretical_loss": 3.353016533907883, + "tokens_seen": 2955935744 + }, + { + "epoch": 0.9, + "learning_rate": 8.447743229689068e-05, + "loss": 1.2486, + "theoretical_loss": 3.353010849739739, + "tokens_seen": 2956001280 + }, + { + "epoch": 0.9, + "learning_rate": 8.446138415245738e-05, + "loss": 1.2491, + "theoretical_loss": 3.353005165732899, + "tokens_seen": 2956066816 + }, + { + "epoch": 0.9, + "learning_rate": 8.444533600802408e-05, + "loss": 1.3611, + "theoretical_loss": 3.3529994818873554, + "tokens_seen": 2956132352 + }, + { + "epoch": 0.9, + "learning_rate": 8.442928786359078e-05, + "loss": 1.2384, + "theoretical_loss": 3.3529937982030997, + "tokens_seen": 2956197888 + }, + { + "epoch": 0.9, + "learning_rate": 8.441323971915748e-05, + "loss": 1.2875, + "theoretical_loss": 3.352988114680124, + "tokens_seen": 2956263424 + }, + { + "epoch": 0.9, + "learning_rate": 8.439719157472418e-05, + "loss": 1.193, + "theoretical_loss": 3.3529824313184196, + "tokens_seen": 2956328960 + }, + { + "epoch": 0.9, + "learning_rate": 8.438114343029088e-05, + "loss": 1.2595, + "theoretical_loss": 3.352976748117979, + "tokens_seen": 2956394496 + }, + { + "epoch": 0.9, + "learning_rate": 8.436509528585757e-05, + "loss": 1.2648, + "theoretical_loss": 3.352971065078794, + "tokens_seen": 2956460032 + }, + { + "epoch": 0.9, + "learning_rate": 8.434904714142427e-05, + "loss": 1.2514, + "theoretical_loss": 3.352965382200856, + "tokens_seen": 2956525568 + }, + { + "epoch": 0.9, + "learning_rate": 8.433299899699097e-05, + "loss": 1.2639, + "theoretical_loss": 3.3529596994841575, + "tokens_seen": 2956591104 + }, + { + "epoch": 0.9, + "learning_rate": 8.431695085255767e-05, + "loss": 1.3583, + "theoretical_loss": 3.35295401692869, + "tokens_seen": 2956656640 + }, + { + "epoch": 0.9, + "learning_rate": 8.430090270812437e-05, + "loss": 1.2257, + "theoretical_loss": 3.352948334534445, + "tokens_seen": 2956722176 + }, + { + "epoch": 0.9, + "learning_rate": 8.428485456369107e-05, + "loss": 1.235, + "theoretical_loss": 3.3529426523014147, + "tokens_seen": 2956787712 + }, + { + "epoch": 0.9, + "learning_rate": 8.426880641925777e-05, + "loss": 1.1805, + "theoretical_loss": 3.352936970229591, + "tokens_seen": 2956853248 + }, + { + "epoch": 0.9, + "learning_rate": 8.425275827482447e-05, + "loss": 1.3097, + "theoretical_loss": 3.3529312883189655, + "tokens_seen": 2956918784 + }, + { + "epoch": 0.9, + "learning_rate": 8.423671013039117e-05, + "loss": 1.3483, + "theoretical_loss": 3.3529256065695305, + "tokens_seen": 2956984320 + }, + { + "epoch": 0.9, + "learning_rate": 8.422066198595789e-05, + "loss": 1.3079, + "theoretical_loss": 3.3529199249812778, + "tokens_seen": 2957049856 + }, + { + "epoch": 0.9, + "learning_rate": 8.420461384152458e-05, + "loss": 1.2531, + "theoretical_loss": 3.3529142435541988, + "tokens_seen": 2957115392 + }, + { + "epoch": 0.9, + "learning_rate": 8.418856569709128e-05, + "loss": 1.3402, + "theoretical_loss": 3.3529085622882855, + "tokens_seen": 2957180928 + }, + { + "epoch": 0.9, + "learning_rate": 8.417251755265798e-05, + "loss": 1.3935, + "theoretical_loss": 3.35290288118353, + "tokens_seen": 2957246464 + }, + { + "epoch": 0.9, + "learning_rate": 8.415646940822468e-05, + "loss": 1.3164, + "theoretical_loss": 3.3528972002399238, + "tokens_seen": 2957312000 + }, + { + "epoch": 0.9, + "learning_rate": 8.414042126379138e-05, + "loss": 1.2582, + "theoretical_loss": 3.3528915194574593, + "tokens_seen": 2957377536 + }, + { + "epoch": 0.9, + "learning_rate": 8.412437311935808e-05, + "loss": 1.279, + "theoretical_loss": 3.352885838836128, + "tokens_seen": 2957443072 + }, + { + "epoch": 0.9, + "learning_rate": 8.410832497492478e-05, + "loss": 1.3373, + "theoretical_loss": 3.3528801583759216, + "tokens_seen": 2957508608 + }, + { + "epoch": 0.9, + "learning_rate": 8.409227683049148e-05, + "loss": 1.2182, + "theoretical_loss": 3.3528744780768323, + "tokens_seen": 2957574144 + }, + { + "epoch": 0.9, + "learning_rate": 8.407622868605818e-05, + "loss": 1.1381, + "theoretical_loss": 3.352868797938852, + "tokens_seen": 2957639680 + }, + { + "epoch": 0.9, + "learning_rate": 8.406018054162488e-05, + "loss": 1.2795, + "theoretical_loss": 3.352863117961972, + "tokens_seen": 2957705216 + }, + { + "epoch": 0.9, + "learning_rate": 8.404413239719158e-05, + "loss": 1.1572, + "theoretical_loss": 3.3528574381461844, + "tokens_seen": 2957770752 + }, + { + "epoch": 0.9, + "learning_rate": 8.402808425275828e-05, + "loss": 1.2994, + "theoretical_loss": 3.3528517584914814, + "tokens_seen": 2957836288 + }, + { + "epoch": 0.9, + "learning_rate": 8.401203610832497e-05, + "loss": 1.1544, + "theoretical_loss": 3.352846078997855, + "tokens_seen": 2957901824 + }, + { + "epoch": 0.9, + "learning_rate": 8.399598796389167e-05, + "loss": 1.2604, + "theoretical_loss": 3.352840399665296, + "tokens_seen": 2957967360 + }, + { + "epoch": 0.9, + "learning_rate": 8.397993981945837e-05, + "loss": 1.2414, + "theoretical_loss": 3.3528347204937976, + "tokens_seen": 2958032896 + }, + { + "epoch": 0.9, + "learning_rate": 8.396389167502507e-05, + "loss": 1.2053, + "theoretical_loss": 3.352829041483351, + "tokens_seen": 2958098432 + }, + { + "epoch": 0.9, + "learning_rate": 8.394784353059177e-05, + "loss": 1.3024, + "theoretical_loss": 3.352823362633948, + "tokens_seen": 2958163968 + }, + { + "epoch": 0.9, + "learning_rate": 8.393179538615849e-05, + "loss": 1.2489, + "theoretical_loss": 3.3528176839455806, + "tokens_seen": 2958229504 + }, + { + "epoch": 0.9, + "learning_rate": 8.391574724172519e-05, + "loss": 1.3656, + "theoretical_loss": 3.3528120054182406, + "tokens_seen": 2958295040 + }, + { + "epoch": 0.9, + "learning_rate": 8.389969909729189e-05, + "loss": 1.2388, + "theoretical_loss": 3.35280632705192, + "tokens_seen": 2958360576 + }, + { + "epoch": 0.9, + "learning_rate": 8.388365095285859e-05, + "loss": 1.2726, + "theoretical_loss": 3.3528006488466104, + "tokens_seen": 2958426112 + }, + { + "epoch": 0.9, + "learning_rate": 8.386760280842529e-05, + "loss": 1.3172, + "theoretical_loss": 3.352794970802304, + "tokens_seen": 2958491648 + }, + { + "epoch": 0.9, + "learning_rate": 8.385155466399198e-05, + "loss": 1.3601, + "theoretical_loss": 3.3527892929189926, + "tokens_seen": 2958557184 + }, + { + "epoch": 0.9, + "learning_rate": 8.383550651955868e-05, + "loss": 1.2858, + "theoretical_loss": 3.352783615196668, + "tokens_seen": 2958622720 + }, + { + "epoch": 0.9, + "learning_rate": 8.381945837512538e-05, + "loss": 1.3119, + "theoretical_loss": 3.352777937635322, + "tokens_seen": 2958688256 + }, + { + "epoch": 0.9, + "learning_rate": 8.380341023069208e-05, + "loss": 1.2496, + "theoretical_loss": 3.3527722602349463, + "tokens_seen": 2958753792 + }, + { + "epoch": 0.9, + "learning_rate": 8.378736208625878e-05, + "loss": 1.2528, + "theoretical_loss": 3.3527665829955335, + "tokens_seen": 2958819328 + }, + { + "epoch": 0.9, + "learning_rate": 8.377131394182548e-05, + "loss": 1.1666, + "theoretical_loss": 3.3527609059170747, + "tokens_seen": 2958884864 + }, + { + "epoch": 0.9, + "objective/train/docs_used": 1075681, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.3514275550842285, + "objective/train/theoretical_loss": 3.3527552289995617, + "objective/train/tokens_used": 2979410400, + "theoretical_loss": 3.3527552289995617, + "tokens_seen": 2958950400 + }, + { + "epoch": 0.9, + "learning_rate": 8.375526579739218e-05, + "loss": 1.2992, + "theoretical_loss": 3.3527552289995617, + "tokens_seen": 2958950400 + }, + { + "epoch": 0.9, + "learning_rate": 8.373921765295888e-05, + "loss": 1.3126, + "theoretical_loss": 3.3527495522429875, + "tokens_seen": 2959015936 + }, + { + "epoch": 0.9, + "learning_rate": 8.372316950852558e-05, + "loss": 1.225, + "theoretical_loss": 3.3527438756473424, + "tokens_seen": 2959081472 + }, + { + "epoch": 0.9, + "learning_rate": 8.370712136409228e-05, + "loss": 1.3172, + "theoretical_loss": 3.3527381992126193, + "tokens_seen": 2959147008 + }, + { + "epoch": 0.9, + "learning_rate": 8.369107321965898e-05, + "loss": 1.3608, + "theoretical_loss": 3.35273252293881, + "tokens_seen": 2959212544 + }, + { + "epoch": 0.9, + "learning_rate": 8.367502507522568e-05, + "loss": 1.2072, + "theoretical_loss": 3.3527268468259064, + "tokens_seen": 2959278080 + }, + { + "epoch": 0.9, + "learning_rate": 8.365897693079237e-05, + "loss": 1.2771, + "theoretical_loss": 3.3527211708738998, + "tokens_seen": 2959343616 + }, + { + "epoch": 0.9, + "learning_rate": 8.364292878635907e-05, + "loss": 1.2577, + "theoretical_loss": 3.3527154950827827, + "tokens_seen": 2959409152 + }, + { + "epoch": 0.9, + "learning_rate": 8.362688064192579e-05, + "loss": 1.2292, + "theoretical_loss": 3.3527098194525466, + "tokens_seen": 2959474688 + }, + { + "epoch": 0.9, + "learning_rate": 8.361083249749249e-05, + "loss": 1.2455, + "theoretical_loss": 3.3527041439831833, + "tokens_seen": 2959540224 + }, + { + "epoch": 0.9, + "learning_rate": 8.359478435305919e-05, + "loss": 1.3586, + "theoretical_loss": 3.352698468674685, + "tokens_seen": 2959605760 + }, + { + "epoch": 0.9, + "learning_rate": 8.357873620862589e-05, + "loss": 1.2899, + "theoretical_loss": 3.352692793527044, + "tokens_seen": 2959671296 + }, + { + "epoch": 0.9, + "learning_rate": 8.356268806419259e-05, + "loss": 1.2992, + "theoretical_loss": 3.352687118540251, + "tokens_seen": 2959736832 + }, + { + "epoch": 0.9, + "learning_rate": 8.354663991975929e-05, + "loss": 1.3093, + "theoretical_loss": 3.352681443714299, + "tokens_seen": 2959802368 + }, + { + "epoch": 0.9, + "learning_rate": 8.353059177532599e-05, + "loss": 1.2353, + "theoretical_loss": 3.352675769049179, + "tokens_seen": 2959867904 + }, + { + "epoch": 0.9, + "learning_rate": 8.351454363089269e-05, + "loss": 1.3041, + "theoretical_loss": 3.3526700945448837, + "tokens_seen": 2959933440 + }, + { + "epoch": 0.9, + "learning_rate": 8.349849548645938e-05, + "loss": 1.2543, + "theoretical_loss": 3.352664420201404, + "tokens_seen": 2959998976 + }, + { + "epoch": 0.9, + "learning_rate": 8.348244734202608e-05, + "loss": 1.2421, + "theoretical_loss": 3.352658746018733, + "tokens_seen": 2960064512 + }, + { + "epoch": 0.9, + "learning_rate": 8.346639919759278e-05, + "loss": 1.2525, + "theoretical_loss": 3.352653071996862, + "tokens_seen": 2960130048 + }, + { + "epoch": 0.9, + "learning_rate": 8.345035105315948e-05, + "loss": 1.2216, + "theoretical_loss": 3.3526473981357823, + "tokens_seen": 2960195584 + }, + { + "epoch": 0.9, + "learning_rate": 8.343430290872618e-05, + "loss": 1.2635, + "theoretical_loss": 3.3526417244354865, + "tokens_seen": 2960261120 + }, + { + "epoch": 0.9, + "learning_rate": 8.341825476429288e-05, + "loss": 1.3045, + "theoretical_loss": 3.352636050895966, + "tokens_seen": 2960326656 + }, + { + "epoch": 0.9, + "learning_rate": 8.340220661985958e-05, + "loss": 1.3541, + "theoretical_loss": 3.3526303775172135, + "tokens_seen": 2960392192 + }, + { + "epoch": 0.9, + "learning_rate": 8.338615847542628e-05, + "loss": 1.2008, + "theoretical_loss": 3.3526247042992203, + "tokens_seen": 2960457728 + }, + { + "epoch": 0.9, + "learning_rate": 8.337011033099298e-05, + "loss": 1.2247, + "theoretical_loss": 3.352619031241978, + "tokens_seen": 2960523264 + }, + { + "epoch": 0.9, + "learning_rate": 8.335406218655968e-05, + "loss": 1.2839, + "theoretical_loss": 3.3526133583454794, + "tokens_seen": 2960588800 + }, + { + "epoch": 0.9, + "learning_rate": 8.333801404212638e-05, + "loss": 1.2247, + "theoretical_loss": 3.352607685609715, + "tokens_seen": 2960654336 + }, + { + "epoch": 0.9, + "learning_rate": 8.332196589769308e-05, + "loss": 1.2674, + "theoretical_loss": 3.3526020130346783, + "tokens_seen": 2960719872 + }, + { + "epoch": 0.9, + "learning_rate": 8.330591775325978e-05, + "loss": 1.3171, + "theoretical_loss": 3.3525963406203596, + "tokens_seen": 2960785408 + }, + { + "epoch": 0.9, + "learning_rate": 8.328986960882648e-05, + "loss": 1.1915, + "theoretical_loss": 3.3525906683667523, + "tokens_seen": 2960850944 + }, + { + "epoch": 0.9, + "learning_rate": 8.327382146439319e-05, + "loss": 1.293, + "theoretical_loss": 3.352584996273847, + "tokens_seen": 2960916480 + }, + { + "epoch": 0.9, + "learning_rate": 8.325777331995989e-05, + "loss": 1.244, + "theoretical_loss": 3.3525793243416366, + "tokens_seen": 2960982016 + }, + { + "epoch": 0.9, + "learning_rate": 8.324172517552659e-05, + "loss": 1.2034, + "theoretical_loss": 3.3525736525701126, + "tokens_seen": 2961047552 + }, + { + "epoch": 0.9, + "learning_rate": 8.322567703109329e-05, + "loss": 1.2337, + "theoretical_loss": 3.352567980959267, + "tokens_seen": 2961113088 + }, + { + "epoch": 0.9, + "learning_rate": 8.320962888665999e-05, + "loss": 1.2813, + "theoretical_loss": 3.352562309509091, + "tokens_seen": 2961178624 + }, + { + "epoch": 0.9, + "learning_rate": 8.319358074222669e-05, + "loss": 1.2177, + "theoretical_loss": 3.3525566382195775, + "tokens_seen": 2961244160 + }, + { + "epoch": 0.9, + "learning_rate": 8.317753259779339e-05, + "loss": 1.211, + "theoretical_loss": 3.3525509670907176, + "tokens_seen": 2961309696 + }, + { + "epoch": 0.9, + "learning_rate": 8.316148445336009e-05, + "loss": 1.2568, + "theoretical_loss": 3.3525452961225035, + "tokens_seen": 2961375232 + }, + { + "epoch": 0.9, + "learning_rate": 8.314543630892678e-05, + "loss": 1.178, + "theoretical_loss": 3.352539625314927, + "tokens_seen": 2961440768 + }, + { + "epoch": 0.9, + "learning_rate": 8.312938816449348e-05, + "loss": 1.1692, + "theoretical_loss": 3.352533954667981, + "tokens_seen": 2961506304 + }, + { + "epoch": 0.9, + "learning_rate": 8.311334002006018e-05, + "loss": 1.3702, + "theoretical_loss": 3.3525282841816555, + "tokens_seen": 2961571840 + }, + { + "epoch": 0.9, + "learning_rate": 8.309729187562688e-05, + "loss": 1.2461, + "theoretical_loss": 3.352522613855944, + "tokens_seen": 2961637376 + }, + { + "epoch": 0.9, + "learning_rate": 8.308124373119358e-05, + "loss": 1.1887, + "theoretical_loss": 3.3525169436908375, + "tokens_seen": 2961702912 + }, + { + "epoch": 0.9, + "learning_rate": 8.306519558676028e-05, + "loss": 1.2303, + "theoretical_loss": 3.352511273686328, + "tokens_seen": 2961768448 + }, + { + "epoch": 0.9, + "learning_rate": 8.304914744232698e-05, + "loss": 1.2475, + "theoretical_loss": 3.352505603842408, + "tokens_seen": 2961833984 + }, + { + "epoch": 0.9, + "learning_rate": 8.303309929789368e-05, + "loss": 1.2753, + "theoretical_loss": 3.352499934159069, + "tokens_seen": 2961899520 + }, + { + "epoch": 0.9, + "learning_rate": 8.30170511534604e-05, + "loss": 1.3651, + "theoretical_loss": 3.3524942646363027, + "tokens_seen": 2961965056 + }, + { + "epoch": 0.9, + "learning_rate": 8.30010030090271e-05, + "loss": 1.2275, + "theoretical_loss": 3.3524885952741013, + "tokens_seen": 2962030592 + }, + { + "epoch": 0.9, + "learning_rate": 8.298495486459378e-05, + "loss": 1.3042, + "theoretical_loss": 3.3524829260724562, + "tokens_seen": 2962096128 + }, + { + "epoch": 0.9, + "learning_rate": 8.296890672016048e-05, + "loss": 1.2474, + "theoretical_loss": 3.35247725703136, + "tokens_seen": 2962161664 + }, + { + "epoch": 0.9, + "objective/train/docs_used": 1076860, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.5284337997436523, + "objective/train/theoretical_loss": 3.3524715881508045, + "objective/train/tokens_used": 2982687200, + "theoretical_loss": 3.3524715881508045, + "tokens_seen": 2962227200 + }, + { + "epoch": 0.9, + "learning_rate": 8.295285857572718e-05, + "loss": 1.3271, + "theoretical_loss": 3.3524715881508045, + "tokens_seen": 2962227200 + }, + { + "epoch": 0.9, + "learning_rate": 8.293681043129389e-05, + "loss": 1.1875, + "theoretical_loss": 3.3524659194307813, + "tokens_seen": 2962292736 + }, + { + "epoch": 0.9, + "learning_rate": 8.292076228686059e-05, + "loss": 1.3699, + "theoretical_loss": 3.352460250871282, + "tokens_seen": 2962358272 + }, + { + "epoch": 0.9, + "learning_rate": 8.290471414242729e-05, + "loss": 1.3667, + "theoretical_loss": 3.3524545824722995, + "tokens_seen": 2962423808 + }, + { + "epoch": 0.9, + "learning_rate": 8.288866599799399e-05, + "loss": 1.2673, + "theoretical_loss": 3.352448914233825, + "tokens_seen": 2962489344 + }, + { + "epoch": 0.9, + "learning_rate": 8.287261785356069e-05, + "loss": 1.3019, + "theoretical_loss": 3.3524432461558504, + "tokens_seen": 2962554880 + }, + { + "epoch": 0.9, + "learning_rate": 8.285656970912739e-05, + "loss": 1.2957, + "theoretical_loss": 3.3524375782383675, + "tokens_seen": 2962620416 + }, + { + "epoch": 0.9, + "learning_rate": 8.284052156469409e-05, + "loss": 1.2424, + "theoretical_loss": 3.3524319104813687, + "tokens_seen": 2962685952 + }, + { + "epoch": 0.9, + "learning_rate": 8.282447342026079e-05, + "loss": 1.2059, + "theoretical_loss": 3.3524262428848455, + "tokens_seen": 2962751488 + }, + { + "epoch": 0.9, + "learning_rate": 8.280842527582749e-05, + "loss": 1.3259, + "theoretical_loss": 3.35242057544879, + "tokens_seen": 2962817024 + }, + { + "epoch": 0.9, + "learning_rate": 8.279237713139418e-05, + "loss": 1.2835, + "theoretical_loss": 3.352414908173194, + "tokens_seen": 2962882560 + }, + { + "epoch": 0.9, + "learning_rate": 8.277632898696088e-05, + "loss": 1.2151, + "theoretical_loss": 3.3524092410580493, + "tokens_seen": 2962948096 + }, + { + "epoch": 0.9, + "learning_rate": 8.276028084252758e-05, + "loss": 1.3218, + "theoretical_loss": 3.352403574103348, + "tokens_seen": 2963013632 + }, + { + "epoch": 0.9, + "learning_rate": 8.274423269809428e-05, + "loss": 1.3685, + "theoretical_loss": 3.352397907309082, + "tokens_seen": 2963079168 + }, + { + "epoch": 0.9, + "learning_rate": 8.2728184553661e-05, + "loss": 1.2371, + "theoretical_loss": 3.3523922406752433, + "tokens_seen": 2963144704 + }, + { + "epoch": 0.9, + "learning_rate": 8.27121364092277e-05, + "loss": 1.1582, + "theoretical_loss": 3.352386574201824, + "tokens_seen": 2963210240 + }, + { + "epoch": 0.9, + "learning_rate": 8.26960882647944e-05, + "loss": 1.2997, + "theoretical_loss": 3.352380907888815, + "tokens_seen": 2963275776 + }, + { + "epoch": 0.9, + "learning_rate": 8.26800401203611e-05, + "loss": 1.2565, + "theoretical_loss": 3.352375241736209, + "tokens_seen": 2963341312 + }, + { + "epoch": 0.9, + "learning_rate": 8.26639919759278e-05, + "loss": 1.286, + "theoretical_loss": 3.3523695757439986, + "tokens_seen": 2963406848 + }, + { + "epoch": 0.9, + "learning_rate": 8.264794383149448e-05, + "loss": 1.3078, + "theoretical_loss": 3.3523639099121745, + "tokens_seen": 2963472384 + }, + { + "epoch": 0.9, + "learning_rate": 8.263189568706118e-05, + "loss": 1.2012, + "theoretical_loss": 3.3523582442407287, + "tokens_seen": 2963537920 + }, + { + "epoch": 0.9, + "learning_rate": 8.261584754262788e-05, + "loss": 1.3115, + "theoretical_loss": 3.3523525787296538, + "tokens_seen": 2963603456 + }, + { + "epoch": 0.9, + "learning_rate": 8.259979939819459e-05, + "loss": 1.203, + "theoretical_loss": 3.352346913378941, + "tokens_seen": 2963668992 + }, + { + "epoch": 0.9, + "learning_rate": 8.258375125376129e-05, + "loss": 1.1933, + "theoretical_loss": 3.352341248188583, + "tokens_seen": 2963734528 + }, + { + "epoch": 0.9, + "learning_rate": 8.256770310932799e-05, + "loss": 1.3332, + "theoretical_loss": 3.3523355831585713, + "tokens_seen": 2963800064 + }, + { + "epoch": 0.9, + "learning_rate": 8.255165496489469e-05, + "loss": 1.2338, + "theoretical_loss": 3.3523299182888975, + "tokens_seen": 2963865600 + }, + { + "epoch": 0.9, + "learning_rate": 8.253560682046139e-05, + "loss": 1.1554, + "theoretical_loss": 3.352324253579554, + "tokens_seen": 2963931136 + }, + { + "epoch": 0.9, + "learning_rate": 8.251955867602809e-05, + "loss": 1.2463, + "theoretical_loss": 3.3523185890305327, + "tokens_seen": 2963996672 + }, + { + "epoch": 0.9, + "learning_rate": 8.250351053159479e-05, + "loss": 1.3688, + "theoretical_loss": 3.3523129246418253, + "tokens_seen": 2964062208 + }, + { + "epoch": 0.9, + "learning_rate": 8.248746238716149e-05, + "loss": 1.2179, + "theoretical_loss": 3.352307260413424, + "tokens_seen": 2964127744 + }, + { + "epoch": 0.9, + "learning_rate": 8.247141424272819e-05, + "loss": 1.183, + "theoretical_loss": 3.3523015963453204, + "tokens_seen": 2964193280 + }, + { + "epoch": 0.9, + "learning_rate": 8.245536609829488e-05, + "loss": 1.3128, + "theoretical_loss": 3.352295932437506, + "tokens_seen": 2964258816 + }, + { + "epoch": 0.9, + "learning_rate": 8.243931795386158e-05, + "loss": 1.2531, + "theoretical_loss": 3.352290268689974, + "tokens_seen": 2964324352 + }, + { + "epoch": 0.9, + "learning_rate": 8.242326980942829e-05, + "loss": 1.2538, + "theoretical_loss": 3.352284605102715, + "tokens_seen": 2964389888 + }, + { + "epoch": 0.9, + "learning_rate": 8.240722166499499e-05, + "loss": 1.2212, + "theoretical_loss": 3.352278941675722, + "tokens_seen": 2964455424 + }, + { + "epoch": 0.9, + "learning_rate": 8.23911735205617e-05, + "loss": 1.2764, + "theoretical_loss": 3.352273278408986, + "tokens_seen": 2964520960 + }, + { + "epoch": 0.9, + "learning_rate": 8.23751253761284e-05, + "loss": 1.3919, + "theoretical_loss": 3.3522676153024995, + "tokens_seen": 2964586496 + }, + { + "epoch": 0.9, + "learning_rate": 8.23590772316951e-05, + "loss": 1.29, + "theoretical_loss": 3.3522619523562542, + "tokens_seen": 2964652032 + }, + { + "epoch": 0.9, + "learning_rate": 8.23430290872618e-05, + "loss": 1.2629, + "theoretical_loss": 3.352256289570242, + "tokens_seen": 2964717568 + }, + { + "epoch": 0.9, + "learning_rate": 8.23269809428285e-05, + "loss": 1.302, + "theoretical_loss": 3.352250626944455, + "tokens_seen": 2964783104 + }, + { + "epoch": 0.9, + "learning_rate": 8.23109327983952e-05, + "loss": 1.2644, + "theoretical_loss": 3.3522449644788854, + "tokens_seen": 2964848640 + }, + { + "epoch": 0.9, + "learning_rate": 8.229488465396188e-05, + "loss": 1.2659, + "theoretical_loss": 3.3522393021735244, + "tokens_seen": 2964914176 + }, + { + "epoch": 0.9, + "learning_rate": 8.227883650952858e-05, + "loss": 1.3175, + "theoretical_loss": 3.352233640028364, + "tokens_seen": 2964979712 + }, + { + "epoch": 0.9, + "learning_rate": 8.226278836509528e-05, + "loss": 1.21, + "theoretical_loss": 3.352227978043397, + "tokens_seen": 2965045248 + }, + { + "epoch": 0.9, + "learning_rate": 8.224674022066199e-05, + "loss": 1.233, + "theoretical_loss": 3.3522223162186147, + "tokens_seen": 2965110784 + }, + { + "epoch": 0.9, + "learning_rate": 8.223069207622869e-05, + "loss": 1.279, + "theoretical_loss": 3.3522166545540086, + "tokens_seen": 2965176320 + }, + { + "epoch": 0.9, + "learning_rate": 8.221464393179539e-05, + "loss": 1.1813, + "theoretical_loss": 3.3522109930495714, + "tokens_seen": 2965241856 + }, + { + "epoch": 0.9, + "learning_rate": 8.219859578736209e-05, + "loss": 1.3063, + "theoretical_loss": 3.352205331705295, + "tokens_seen": 2965307392 + }, + { + "epoch": 0.9, + "learning_rate": 8.218254764292879e-05, + "loss": 1.3304, + "theoretical_loss": 3.3521996705211707, + "tokens_seen": 2965372928 + }, + { + "epoch": 0.9, + "learning_rate": 8.216649949849549e-05, + "loss": 1.3056, + "theoretical_loss": 3.3521940094971905, + "tokens_seen": 2965438464 + }, + { + "epoch": 0.9, + "objective/train/docs_used": 1078080, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.3496520519256592, + "objective/train/theoretical_loss": 3.352188348633347, + "objective/train/tokens_used": 2985964000, + "theoretical_loss": 3.352188348633347, + "tokens_seen": 2965504000 + }, + { + "epoch": 0.9, + "learning_rate": 8.215045135406219e-05, + "loss": 1.2967, + "theoretical_loss": 3.352188348633347, + "tokens_seen": 2965504000 + }, + { + "epoch": 0.9, + "learning_rate": 8.213440320962889e-05, + "loss": 1.2783, + "theoretical_loss": 3.3521826879296315, + "tokens_seen": 2965569536 + }, + { + "epoch": 0.9, + "learning_rate": 8.211835506519559e-05, + "loss": 1.2305, + "theoretical_loss": 3.3521770273860367, + "tokens_seen": 2965635072 + }, + { + "epoch": 0.9, + "learning_rate": 8.210230692076229e-05, + "loss": 1.188, + "theoretical_loss": 3.3521713670025535, + "tokens_seen": 2965700608 + }, + { + "epoch": 0.9, + "learning_rate": 8.208625877632899e-05, + "loss": 1.3006, + "theoretical_loss": 3.3521657067791746, + "tokens_seen": 2965766144 + }, + { + "epoch": 0.9, + "learning_rate": 8.207021063189569e-05, + "loss": 1.2915, + "theoretical_loss": 3.3521600467158916, + "tokens_seen": 2965831680 + }, + { + "epoch": 0.9, + "learning_rate": 8.205416248746239e-05, + "loss": 1.2149, + "theoretical_loss": 3.3521543868126966, + "tokens_seen": 2965897216 + }, + { + "epoch": 0.9, + "learning_rate": 8.20381143430291e-05, + "loss": 1.2566, + "theoretical_loss": 3.3521487270695816, + "tokens_seen": 2965962752 + }, + { + "epoch": 0.9, + "learning_rate": 8.20220661985958e-05, + "loss": 1.2549, + "theoretical_loss": 3.352143067486538, + "tokens_seen": 2966028288 + }, + { + "epoch": 0.9, + "learning_rate": 8.20060180541625e-05, + "loss": 1.3082, + "theoretical_loss": 3.3521374080635584, + "tokens_seen": 2966093824 + }, + { + "epoch": 0.9, + "learning_rate": 8.19899699097292e-05, + "loss": 1.1911, + "theoretical_loss": 3.352131748800634, + "tokens_seen": 2966159360 + }, + { + "epoch": 0.9, + "learning_rate": 8.19739217652959e-05, + "loss": 1.2397, + "theoretical_loss": 3.352126089697758, + "tokens_seen": 2966224896 + }, + { + "epoch": 0.9, + "learning_rate": 8.19578736208626e-05, + "loss": 1.2718, + "theoretical_loss": 3.352120430754921, + "tokens_seen": 2966290432 + }, + { + "epoch": 0.9, + "learning_rate": 8.194182547642928e-05, + "loss": 1.2069, + "theoretical_loss": 3.3521147719721154, + "tokens_seen": 2966355968 + }, + { + "epoch": 0.9, + "learning_rate": 8.192577733199598e-05, + "loss": 1.3137, + "theoretical_loss": 3.3521091133493335, + "tokens_seen": 2966421504 + }, + { + "epoch": 0.9, + "learning_rate": 8.190972918756269e-05, + "loss": 1.2715, + "theoretical_loss": 3.3521034548865667, + "tokens_seen": 2966487040 + }, + { + "epoch": 0.9, + "learning_rate": 8.189368104312939e-05, + "loss": 1.201, + "theoretical_loss": 3.352097796583807, + "tokens_seen": 2966552576 + }, + { + "epoch": 0.9, + "learning_rate": 8.187763289869609e-05, + "loss": 1.115, + "theoretical_loss": 3.352092138441047, + "tokens_seen": 2966618112 + }, + { + "epoch": 0.9, + "learning_rate": 8.186158475426279e-05, + "loss": 1.2555, + "theoretical_loss": 3.352086480458278, + "tokens_seen": 2966683648 + }, + { + "epoch": 0.9, + "learning_rate": 8.184553660982949e-05, + "loss": 1.2514, + "theoretical_loss": 3.352080822635492, + "tokens_seen": 2966749184 + }, + { + "epoch": 0.9, + "learning_rate": 8.18294884653962e-05, + "loss": 1.2382, + "theoretical_loss": 3.352075164972681, + "tokens_seen": 2966814720 + }, + { + "epoch": 0.9, + "learning_rate": 8.18134403209629e-05, + "loss": 1.2546, + "theoretical_loss": 3.3520695074698375, + "tokens_seen": 2966880256 + }, + { + "epoch": 0.9, + "learning_rate": 8.17973921765296e-05, + "loss": 1.1979, + "theoretical_loss": 3.352063850126952, + "tokens_seen": 2966945792 + }, + { + "epoch": 0.9, + "learning_rate": 8.178134403209629e-05, + "loss": 1.1339, + "theoretical_loss": 3.3520581929440185, + "tokens_seen": 2967011328 + }, + { + "epoch": 0.9, + "learning_rate": 8.176529588766299e-05, + "loss": 1.1589, + "theoretical_loss": 3.352052535921027, + "tokens_seen": 2967076864 + }, + { + "epoch": 0.9, + "learning_rate": 8.174924774322969e-05, + "loss": 1.3156, + "theoretical_loss": 3.352046879057971, + "tokens_seen": 2967142400 + }, + { + "epoch": 0.9, + "learning_rate": 8.173319959879639e-05, + "loss": 1.2038, + "theoretical_loss": 3.352041222354841, + "tokens_seen": 2967207936 + }, + { + "epoch": 0.9, + "learning_rate": 8.171715145436309e-05, + "loss": 1.3422, + "theoretical_loss": 3.3520355658116303, + "tokens_seen": 2967273472 + }, + { + "epoch": 0.9, + "learning_rate": 8.17011033099298e-05, + "loss": 1.213, + "theoretical_loss": 3.35202990942833, + "tokens_seen": 2967339008 + }, + { + "epoch": 0.9, + "learning_rate": 8.16850551654965e-05, + "loss": 1.248, + "theoretical_loss": 3.352024253204932, + "tokens_seen": 2967404544 + }, + { + "epoch": 0.9, + "learning_rate": 8.16690070210632e-05, + "loss": 1.3384, + "theoretical_loss": 3.3520185971414285, + "tokens_seen": 2967470080 + }, + { + "epoch": 0.9, + "learning_rate": 8.16529588766299e-05, + "loss": 1.2348, + "theoretical_loss": 3.3520129412378115, + "tokens_seen": 2967535616 + }, + { + "epoch": 0.9, + "learning_rate": 8.16369107321966e-05, + "loss": 1.2562, + "theoretical_loss": 3.3520072854940732, + "tokens_seen": 2967601152 + }, + { + "epoch": 0.9, + "learning_rate": 8.16208625877633e-05, + "loss": 1.27, + "theoretical_loss": 3.352001629910205, + "tokens_seen": 2967666688 + }, + { + "epoch": 0.9, + "learning_rate": 8.160481444333e-05, + "loss": 1.2306, + "theoretical_loss": 3.3519959744861993, + "tokens_seen": 2967732224 + }, + { + "epoch": 0.9, + "learning_rate": 8.158876629889668e-05, + "loss": 1.3902, + "theoretical_loss": 3.3519903192220477, + "tokens_seen": 2967797760 + }, + { + "epoch": 0.9, + "learning_rate": 8.157271815446339e-05, + "loss": 1.2629, + "theoretical_loss": 3.3519846641177424, + "tokens_seen": 2967863296 + }, + { + "epoch": 0.9, + "learning_rate": 8.155667001003009e-05, + "loss": 1.3211, + "theoretical_loss": 3.3519790091732755, + "tokens_seen": 2967928832 + }, + { + "epoch": 0.9, + "learning_rate": 8.154062186559679e-05, + "loss": 1.2683, + "theoretical_loss": 3.3519733543886385, + "tokens_seen": 2967994368 + }, + { + "epoch": 0.9, + "learning_rate": 8.15245737211635e-05, + "loss": 1.2554, + "theoretical_loss": 3.351967699763823, + "tokens_seen": 2968059904 + }, + { + "epoch": 0.9, + "learning_rate": 8.15085255767302e-05, + "loss": 1.2883, + "theoretical_loss": 3.3519620452988224, + "tokens_seen": 2968125440 + }, + { + "epoch": 0.9, + "learning_rate": 8.14924774322969e-05, + "loss": 1.2732, + "theoretical_loss": 3.3519563909936276, + "tokens_seen": 2968190976 + }, + { + "epoch": 0.9, + "learning_rate": 8.14764292878636e-05, + "loss": 1.2273, + "theoretical_loss": 3.3519507368482304, + "tokens_seen": 2968256512 + }, + { + "epoch": 0.9, + "learning_rate": 8.14603811434303e-05, + "loss": 1.2566, + "theoretical_loss": 3.351945082862623, + "tokens_seen": 2968322048 + }, + { + "epoch": 0.9, + "learning_rate": 8.1444332998997e-05, + "loss": 1.2901, + "theoretical_loss": 3.351939429036798, + "tokens_seen": 2968387584 + }, + { + "epoch": 0.9, + "learning_rate": 8.142828485456369e-05, + "loss": 1.3297, + "theoretical_loss": 3.3519337753707465, + "tokens_seen": 2968453120 + }, + { + "epoch": 0.9, + "learning_rate": 8.141223671013039e-05, + "loss": 1.2255, + "theoretical_loss": 3.3519281218644608, + "tokens_seen": 2968518656 + }, + { + "epoch": 0.9, + "learning_rate": 8.139618856569709e-05, + "loss": 1.2534, + "theoretical_loss": 3.351922468517933, + "tokens_seen": 2968584192 + }, + { + "epoch": 0.9, + "learning_rate": 8.138014042126379e-05, + "loss": 1.2413, + "theoretical_loss": 3.3519168153311547, + "tokens_seen": 2968649728 + }, + { + "epoch": 0.9, + "learning_rate": 8.13640922768305e-05, + "loss": 1.2165, + "theoretical_loss": 3.351911162304118, + "tokens_seen": 2968715264 + }, + { + "epoch": 0.9, + "objective/train/docs_used": 1079221, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.271983027458191, + "objective/train/theoretical_loss": 3.351905509436815, + "objective/train/tokens_used": 2989240800, + "theoretical_loss": 3.351905509436815, + "tokens_seen": 2968780800 + }, + { + "epoch": 0.9, + "learning_rate": 8.13480441323972e-05, + "loss": 1.2301, + "theoretical_loss": 3.351905509436815, + "tokens_seen": 2968780800 + }, + { + "epoch": 0.9, + "learning_rate": 8.13319959879639e-05, + "loss": 1.3476, + "theoretical_loss": 3.3518998567292373, + "tokens_seen": 2968846336 + }, + { + "epoch": 0.9, + "learning_rate": 8.13159478435306e-05, + "loss": 1.3109, + "theoretical_loss": 3.3518942041813777, + "tokens_seen": 2968911872 + }, + { + "epoch": 0.9, + "learning_rate": 8.12998996990973e-05, + "loss": 1.1614, + "theoretical_loss": 3.351888551793227, + "tokens_seen": 2968977408 + }, + { + "epoch": 0.9, + "learning_rate": 8.1283851554664e-05, + "loss": 1.2045, + "theoretical_loss": 3.351882899564778, + "tokens_seen": 2969042944 + }, + { + "epoch": 0.9, + "learning_rate": 8.12678034102307e-05, + "loss": 1.2001, + "theoretical_loss": 3.3518772474960223, + "tokens_seen": 2969108480 + }, + { + "epoch": 0.9, + "learning_rate": 8.12517552657974e-05, + "loss": 1.2414, + "theoretical_loss": 3.3518715955869522, + "tokens_seen": 2969174016 + }, + { + "epoch": 0.9, + "learning_rate": 8.123570712136408e-05, + "loss": 1.3345, + "theoretical_loss": 3.3518659438375593, + "tokens_seen": 2969239552 + }, + { + "epoch": 0.9, + "learning_rate": 8.12196589769308e-05, + "loss": 1.1671, + "theoretical_loss": 3.3518602922478355, + "tokens_seen": 2969305088 + }, + { + "epoch": 0.9, + "learning_rate": 8.12036108324975e-05, + "loss": 1.238, + "theoretical_loss": 3.3518546408177734, + "tokens_seen": 2969370624 + }, + { + "epoch": 0.9, + "learning_rate": 8.11875626880642e-05, + "loss": 1.1856, + "theoretical_loss": 3.3518489895473644, + "tokens_seen": 2969436160 + }, + { + "epoch": 0.9, + "learning_rate": 8.11715145436309e-05, + "loss": 1.3333, + "theoretical_loss": 3.3518433384366, + "tokens_seen": 2969501696 + }, + { + "epoch": 0.9, + "learning_rate": 8.11554663991976e-05, + "loss": 1.1453, + "theoretical_loss": 3.3518376874854736, + "tokens_seen": 2969567232 + }, + { + "epoch": 0.9, + "learning_rate": 8.11394182547643e-05, + "loss": 1.2279, + "theoretical_loss": 3.351832036693976, + "tokens_seen": 2969632768 + }, + { + "epoch": 0.9, + "learning_rate": 8.1123370110331e-05, + "loss": 1.21, + "theoretical_loss": 3.3518263860620996, + "tokens_seen": 2969698304 + }, + { + "epoch": 0.9, + "learning_rate": 8.11073219658977e-05, + "loss": 1.2744, + "theoretical_loss": 3.351820735589836, + "tokens_seen": 2969763840 + }, + { + "epoch": 0.9, + "learning_rate": 8.10912738214644e-05, + "loss": 1.1781, + "theoretical_loss": 3.3518150852771775, + "tokens_seen": 2969829376 + }, + { + "epoch": 0.9, + "learning_rate": 8.107522567703109e-05, + "loss": 1.2025, + "theoretical_loss": 3.351809435124116, + "tokens_seen": 2969894912 + }, + { + "epoch": 0.9, + "learning_rate": 8.105917753259779e-05, + "loss": 1.3258, + "theoretical_loss": 3.351803785130644, + "tokens_seen": 2969960448 + }, + { + "epoch": 0.9, + "learning_rate": 8.104312938816449e-05, + "loss": 1.2932, + "theoretical_loss": 3.351798135296752, + "tokens_seen": 2970025984 + }, + { + "epoch": 0.9, + "learning_rate": 8.102708124373119e-05, + "loss": 1.2929, + "theoretical_loss": 3.3517924856224335, + "tokens_seen": 2970091520 + }, + { + "epoch": 0.9, + "learning_rate": 8.10110330992979e-05, + "loss": 1.2521, + "theoretical_loss": 3.35178683610768, + "tokens_seen": 2970157056 + }, + { + "epoch": 0.9, + "learning_rate": 8.09949849548646e-05, + "loss": 1.3154, + "theoretical_loss": 3.351781186752483, + "tokens_seen": 2970222592 + }, + { + "epoch": 0.9, + "learning_rate": 8.09789368104313e-05, + "loss": 1.2056, + "theoretical_loss": 3.3517755375568354, + "tokens_seen": 2970288128 + }, + { + "epoch": 0.9, + "learning_rate": 8.0962888665998e-05, + "loss": 1.2555, + "theoretical_loss": 3.351769888520728, + "tokens_seen": 2970353664 + }, + { + "epoch": 0.9, + "learning_rate": 8.09468405215647e-05, + "loss": 1.2825, + "theoretical_loss": 3.351764239644154, + "tokens_seen": 2970419200 + }, + { + "epoch": 0.9, + "learning_rate": 8.09307923771314e-05, + "loss": 1.2171, + "theoretical_loss": 3.3517585909271044, + "tokens_seen": 2970484736 + }, + { + "epoch": 0.9, + "learning_rate": 8.09147442326981e-05, + "loss": 1.2695, + "theoretical_loss": 3.3517529423695716, + "tokens_seen": 2970550272 + }, + { + "epoch": 0.9, + "learning_rate": 8.08986960882648e-05, + "loss": 1.286, + "theoretical_loss": 3.351747293971547, + "tokens_seen": 2970615808 + }, + { + "epoch": 0.9, + "learning_rate": 8.08826479438315e-05, + "loss": 1.3257, + "theoretical_loss": 3.351741645733024, + "tokens_seen": 2970681344 + }, + { + "epoch": 0.9, + "learning_rate": 8.08665997993982e-05, + "loss": 1.1846, + "theoretical_loss": 3.3517359976539933, + "tokens_seen": 2970746880 + }, + { + "epoch": 0.9, + "learning_rate": 8.08505516549649e-05, + "loss": 1.2826, + "theoretical_loss": 3.351730349734447, + "tokens_seen": 2970812416 + }, + { + "epoch": 0.9, + "learning_rate": 8.08345035105316e-05, + "loss": 1.2207, + "theoretical_loss": 3.3517247019743777, + "tokens_seen": 2970877952 + }, + { + "epoch": 0.9, + "learning_rate": 8.08184553660983e-05, + "loss": 1.1563, + "theoretical_loss": 3.3517190543737767, + "tokens_seen": 2970943488 + }, + { + "epoch": 0.9, + "learning_rate": 8.0802407221665e-05, + "loss": 1.299, + "theoretical_loss": 3.351713406932636, + "tokens_seen": 2971009024 + }, + { + "epoch": 0.9, + "learning_rate": 8.07863590772317e-05, + "loss": 1.229, + "theoretical_loss": 3.3517077596509486, + "tokens_seen": 2971074560 + }, + { + "epoch": 0.9, + "learning_rate": 8.07703109327984e-05, + "loss": 1.2704, + "theoretical_loss": 3.3517021125287054, + "tokens_seen": 2971140096 + }, + { + "epoch": 0.9, + "learning_rate": 8.07542627883651e-05, + "loss": 1.3785, + "theoretical_loss": 3.3516964655658987, + "tokens_seen": 2971205632 + }, + { + "epoch": 0.9, + "learning_rate": 8.073821464393179e-05, + "loss": 1.2087, + "theoretical_loss": 3.3516908187625205, + "tokens_seen": 2971271168 + }, + { + "epoch": 0.9, + "learning_rate": 8.072216649949849e-05, + "loss": 1.2616, + "theoretical_loss": 3.351685172118563, + "tokens_seen": 2971336704 + }, + { + "epoch": 0.9, + "learning_rate": 8.070611835506519e-05, + "loss": 1.2688, + "theoretical_loss": 3.351679525634018, + "tokens_seen": 2971402240 + }, + { + "epoch": 0.9, + "learning_rate": 8.069007021063189e-05, + "loss": 1.2858, + "theoretical_loss": 3.351673879308877, + "tokens_seen": 2971467776 + }, + { + "epoch": 0.9, + "learning_rate": 8.06740220661986e-05, + "loss": 1.1854, + "theoretical_loss": 3.351668233143133, + "tokens_seen": 2971533312 + }, + { + "epoch": 0.9, + "learning_rate": 8.06579739217653e-05, + "loss": 1.2998, + "theoretical_loss": 3.351662587136777, + "tokens_seen": 2971598848 + }, + { + "epoch": 0.9, + "learning_rate": 8.0641925777332e-05, + "loss": 1.1864, + "theoretical_loss": 3.351656941289802, + "tokens_seen": 2971664384 + }, + { + "epoch": 0.9, + "learning_rate": 8.062587763289871e-05, + "loss": 1.2181, + "theoretical_loss": 3.3516512956021987, + "tokens_seen": 2971729920 + }, + { + "epoch": 0.9, + "learning_rate": 8.060982948846541e-05, + "loss": 1.2321, + "theoretical_loss": 3.35164565007396, + "tokens_seen": 2971795456 + }, + { + "epoch": 0.9, + "learning_rate": 8.059378134403211e-05, + "loss": 1.2641, + "theoretical_loss": 3.351640004705078, + "tokens_seen": 2971860992 + }, + { + "epoch": 0.9, + "learning_rate": 8.05777331995988e-05, + "loss": 1.2938, + "theoretical_loss": 3.351634359495544, + "tokens_seen": 2971926528 + }, + { + "epoch": 0.9, + "learning_rate": 8.05616850551655e-05, + "loss": 1.1684, + "theoretical_loss": 3.3516287144453503, + "tokens_seen": 2971992064 + }, + { + "epoch": 0.9, + "objective/train/docs_used": 1080506, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.239708423614502, + "objective/train/theoretical_loss": 3.3516230695544893, + "objective/train/tokens_used": 2992517600, + "theoretical_loss": 3.3516230695544893, + "tokens_seen": 2972057600 + }, + { + "epoch": 0.9, + "learning_rate": 8.05456369107322e-05, + "loss": 1.1949, + "theoretical_loss": 3.3516230695544893, + "tokens_seen": 2972057600 + }, + { + "epoch": 0.9, + "learning_rate": 8.05295887662989e-05, + "loss": 1.2353, + "theoretical_loss": 3.351617424822952, + "tokens_seen": 2972123136 + }, + { + "epoch": 0.9, + "learning_rate": 8.05135406218656e-05, + "loss": 1.2472, + "theoretical_loss": 3.3516117802507317, + "tokens_seen": 2972188672 + }, + { + "epoch": 0.9, + "learning_rate": 8.04974924774323e-05, + "loss": 1.1929, + "theoretical_loss": 3.3516061358378195, + "tokens_seen": 2972254208 + }, + { + "epoch": 0.9, + "learning_rate": 8.0481444332999e-05, + "loss": 1.1836, + "theoretical_loss": 3.3516004915842075, + "tokens_seen": 2972319744 + }, + { + "epoch": 0.9, + "learning_rate": 8.04653961885657e-05, + "loss": 1.2261, + "theoretical_loss": 3.3515948474898876, + "tokens_seen": 2972385280 + }, + { + "epoch": 0.9, + "learning_rate": 8.04493480441324e-05, + "loss": 1.2456, + "theoretical_loss": 3.3515892035548522, + "tokens_seen": 2972450816 + }, + { + "epoch": 0.9, + "learning_rate": 8.04332998996991e-05, + "loss": 1.3475, + "theoretical_loss": 3.351583559779093, + "tokens_seen": 2972516352 + }, + { + "epoch": 0.9, + "learning_rate": 8.04172517552658e-05, + "loss": 1.3898, + "theoretical_loss": 3.3515779161626025, + "tokens_seen": 2972581888 + }, + { + "epoch": 0.9, + "learning_rate": 8.04012036108325e-05, + "loss": 1.1455, + "theoretical_loss": 3.351572272705372, + "tokens_seen": 2972647424 + }, + { + "epoch": 0.9, + "learning_rate": 8.038515546639919e-05, + "loss": 1.2233, + "theoretical_loss": 3.351566629407394, + "tokens_seen": 2972712960 + }, + { + "epoch": 0.9, + "learning_rate": 8.036910732196589e-05, + "loss": 1.1881, + "theoretical_loss": 3.3515609862686597, + "tokens_seen": 2972778496 + }, + { + "epoch": 0.9, + "learning_rate": 8.035305917753259e-05, + "loss": 1.2751, + "theoretical_loss": 3.351555343289162, + "tokens_seen": 2972844032 + }, + { + "epoch": 0.9, + "learning_rate": 8.033701103309929e-05, + "loss": 1.1759, + "theoretical_loss": 3.351549700468892, + "tokens_seen": 2972909568 + }, + { + "epoch": 0.9, + "learning_rate": 8.032096288866601e-05, + "loss": 1.2488, + "theoretical_loss": 3.351544057807843, + "tokens_seen": 2972975104 + }, + { + "epoch": 0.9, + "learning_rate": 8.030491474423271e-05, + "loss": 1.1923, + "theoretical_loss": 3.351538415306006, + "tokens_seen": 2973040640 + }, + { + "epoch": 0.9, + "learning_rate": 8.028886659979941e-05, + "loss": 1.3052, + "theoretical_loss": 3.351532772963373, + "tokens_seen": 2973106176 + }, + { + "epoch": 0.9, + "learning_rate": 8.027281845536611e-05, + "loss": 1.1982, + "theoretical_loss": 3.3515271307799366, + "tokens_seen": 2973171712 + }, + { + "epoch": 0.9, + "learning_rate": 8.025677031093281e-05, + "loss": 1.3461, + "theoretical_loss": 3.3515214887556883, + "tokens_seen": 2973237248 + }, + { + "epoch": 0.9, + "learning_rate": 8.024072216649951e-05, + "loss": 1.2366, + "theoretical_loss": 3.3515158468906203, + "tokens_seen": 2973302784 + }, + { + "epoch": 0.9, + "learning_rate": 8.02246740220662e-05, + "loss": 1.2906, + "theoretical_loss": 3.3515102051847245, + "tokens_seen": 2973368320 + }, + { + "epoch": 0.9, + "learning_rate": 8.02086258776329e-05, + "loss": 1.3416, + "theoretical_loss": 3.3515045636379925, + "tokens_seen": 2973433856 + }, + { + "epoch": 0.9, + "learning_rate": 8.01925777331996e-05, + "loss": 1.2318, + "theoretical_loss": 3.3514989222504177, + "tokens_seen": 2973499392 + }, + { + "epoch": 0.9, + "learning_rate": 8.01765295887663e-05, + "loss": 1.2804, + "theoretical_loss": 3.35149328102199, + "tokens_seen": 2973564928 + }, + { + "epoch": 0.9, + "learning_rate": 8.0160481444333e-05, + "loss": 1.2257, + "theoretical_loss": 3.3514876399527034, + "tokens_seen": 2973630464 + }, + { + "epoch": 0.9, + "learning_rate": 8.01444332998997e-05, + "loss": 1.2335, + "theoretical_loss": 3.351481999042549, + "tokens_seen": 2973696000 + }, + { + "epoch": 0.9, + "learning_rate": 8.01283851554664e-05, + "loss": 1.3526, + "theoretical_loss": 3.3514763582915186, + "tokens_seen": 2973761536 + }, + { + "epoch": 0.9, + "learning_rate": 8.01123370110331e-05, + "loss": 1.2765, + "theoretical_loss": 3.3514707176996046, + "tokens_seen": 2973827072 + }, + { + "epoch": 0.9, + "learning_rate": 8.00962888665998e-05, + "loss": 1.3921, + "theoretical_loss": 3.3514650772667984, + "tokens_seen": 2973892608 + }, + { + "epoch": 0.9, + "learning_rate": 8.00802407221665e-05, + "loss": 1.23, + "theoretical_loss": 3.351459436993093, + "tokens_seen": 2973958144 + }, + { + "epoch": 0.9, + "learning_rate": 8.00641925777332e-05, + "loss": 1.2436, + "theoretical_loss": 3.35145379687848, + "tokens_seen": 2974023680 + }, + { + "epoch": 0.9, + "learning_rate": 8.00481444332999e-05, + "loss": 1.2434, + "theoretical_loss": 3.3514481569229506, + "tokens_seen": 2974089216 + }, + { + "epoch": 0.9, + "learning_rate": 8.003209628886659e-05, + "loss": 1.3843, + "theoretical_loss": 3.351442517126498, + "tokens_seen": 2974154752 + }, + { + "epoch": 0.9, + "learning_rate": 8.00160481444333e-05, + "loss": 1.2202, + "theoretical_loss": 3.351436877489114, + "tokens_seen": 2974220288 + }, + { + "epoch": 0.9, + "learning_rate": 8e-05, + "loss": 1.2528, + "theoretical_loss": 3.3514312380107896, + "tokens_seen": 2974285824 + }, + { + "epoch": 0.9, + "learning_rate": 7.998395185556671e-05, + "loss": 1.3049, + "theoretical_loss": 3.351425598691518, + "tokens_seen": 2974351360 + }, + { + "epoch": 0.9, + "learning_rate": 7.996790371113341e-05, + "loss": 1.2499, + "theoretical_loss": 3.3514199595312903, + "tokens_seen": 2974416896 + }, + { + "epoch": 0.9, + "learning_rate": 7.995185556670011e-05, + "loss": 1.2962, + "theoretical_loss": 3.3514143205300995, + "tokens_seen": 2974482432 + }, + { + "epoch": 0.9, + "learning_rate": 7.993580742226681e-05, + "loss": 1.3082, + "theoretical_loss": 3.3514086816879365, + "tokens_seen": 2974547968 + }, + { + "epoch": 0.9, + "learning_rate": 7.991975927783351e-05, + "loss": 1.3052, + "theoretical_loss": 3.3514030430047943, + "tokens_seen": 2974613504 + }, + { + "epoch": 0.9, + "learning_rate": 7.990371113340021e-05, + "loss": 1.1902, + "theoretical_loss": 3.3513974044806645, + "tokens_seen": 2974679040 + }, + { + "epoch": 0.9, + "learning_rate": 7.988766298896691e-05, + "loss": 1.2199, + "theoretical_loss": 3.3513917661155386, + "tokens_seen": 2974744576 + }, + { + "epoch": 0.9, + "learning_rate": 7.98716148445336e-05, + "loss": 1.2375, + "theoretical_loss": 3.3513861279094095, + "tokens_seen": 2974810112 + }, + { + "epoch": 0.9, + "learning_rate": 7.98555667001003e-05, + "loss": 1.3153, + "theoretical_loss": 3.3513804898622688, + "tokens_seen": 2974875648 + }, + { + "epoch": 0.9, + "learning_rate": 7.9839518555667e-05, + "loss": 1.2388, + "theoretical_loss": 3.3513748519741084, + "tokens_seen": 2974941184 + }, + { + "epoch": 0.9, + "learning_rate": 7.98234704112337e-05, + "loss": 1.2324, + "theoretical_loss": 3.3513692142449205, + "tokens_seen": 2975006720 + }, + { + "epoch": 0.9, + "learning_rate": 7.98074222668004e-05, + "loss": 1.2353, + "theoretical_loss": 3.351363576674697, + "tokens_seen": 2975072256 + }, + { + "epoch": 0.9, + "learning_rate": 7.97913741223671e-05, + "loss": 1.2552, + "theoretical_loss": 3.35135793926343, + "tokens_seen": 2975137792 + }, + { + "epoch": 0.9, + "learning_rate": 7.97753259779338e-05, + "loss": 1.2812, + "theoretical_loss": 3.3513523020111116, + "tokens_seen": 2975203328 + }, + { + "epoch": 0.9, + "learning_rate": 7.97592778335005e-05, + "loss": 1.1794, + "theoretical_loss": 3.3513466649177337, + "tokens_seen": 2975268864 + }, + { + "epoch": 0.9, + "objective/train/docs_used": 1081640, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.2726154327392578, + "objective/train/theoretical_loss": 3.3513410279832883, + "objective/train/tokens_used": 2995794400, + "theoretical_loss": 3.3513410279832883, + "tokens_seen": 2975334400 + }, + { + "epoch": 0.9, + "learning_rate": 7.97432296890672e-05, + "loss": 1.231, + "theoretical_loss": 3.3513410279832883, + "tokens_seen": 2975334400 + }, + { + "epoch": 0.9, + "learning_rate": 7.972718154463392e-05, + "loss": 1.276, + "theoretical_loss": 3.3513353912077672, + "tokens_seen": 2975399936 + }, + { + "epoch": 0.9, + "learning_rate": 7.97111334002006e-05, + "loss": 1.2618, + "theoretical_loss": 3.3513297545911627, + "tokens_seen": 2975465472 + }, + { + "epoch": 0.9, + "learning_rate": 7.96950852557673e-05, + "loss": 1.2492, + "theoretical_loss": 3.351324118133467, + "tokens_seen": 2975531008 + }, + { + "epoch": 0.9, + "learning_rate": 7.9679037111334e-05, + "loss": 1.2737, + "theoretical_loss": 3.3513184818346717, + "tokens_seen": 2975596544 + }, + { + "epoch": 0.9, + "learning_rate": 7.96629889669007e-05, + "loss": 1.2354, + "theoretical_loss": 3.3513128456947694, + "tokens_seen": 2975662080 + }, + { + "epoch": 0.9, + "learning_rate": 7.964694082246741e-05, + "loss": 1.1993, + "theoretical_loss": 3.3513072097137515, + "tokens_seen": 2975727616 + }, + { + "epoch": 0.9, + "learning_rate": 7.963089267803411e-05, + "loss": 1.2048, + "theoretical_loss": 3.35130157389161, + "tokens_seen": 2975793152 + }, + { + "epoch": 0.9, + "learning_rate": 7.961484453360081e-05, + "loss": 1.2498, + "theoretical_loss": 3.3512959382283376, + "tokens_seen": 2975858688 + }, + { + "epoch": 0.9, + "learning_rate": 7.959879638916751e-05, + "loss": 1.2996, + "theoretical_loss": 3.351290302723926, + "tokens_seen": 2975924224 + }, + { + "epoch": 0.9, + "learning_rate": 7.958274824473421e-05, + "loss": 1.2332, + "theoretical_loss": 3.351284667378367, + "tokens_seen": 2975989760 + }, + { + "epoch": 0.9, + "learning_rate": 7.956670010030091e-05, + "loss": 1.3661, + "theoretical_loss": 3.3512790321916524, + "tokens_seen": 2976055296 + }, + { + "epoch": 0.9, + "learning_rate": 7.955065195586761e-05, + "loss": 1.2577, + "theoretical_loss": 3.3512733971637747, + "tokens_seen": 2976120832 + }, + { + "epoch": 0.9, + "learning_rate": 7.953460381143431e-05, + "loss": 1.1997, + "theoretical_loss": 3.3512677622947264, + "tokens_seen": 2976186368 + }, + { + "epoch": 0.9, + "learning_rate": 7.9518555667001e-05, + "loss": 1.2863, + "theoretical_loss": 3.3512621275844983, + "tokens_seen": 2976251904 + }, + { + "epoch": 0.9, + "learning_rate": 7.95025075225677e-05, + "loss": 1.2656, + "theoretical_loss": 3.3512564930330835, + "tokens_seen": 2976317440 + }, + { + "epoch": 0.9, + "learning_rate": 7.94864593781344e-05, + "loss": 1.2125, + "theoretical_loss": 3.3512508586404732, + "tokens_seen": 2976382976 + }, + { + "epoch": 0.9, + "learning_rate": 7.94704112337011e-05, + "loss": 1.3417, + "theoretical_loss": 3.35124522440666, + "tokens_seen": 2976448512 + }, + { + "epoch": 0.9, + "learning_rate": 7.94543630892678e-05, + "loss": 1.3535, + "theoretical_loss": 3.3512395903316357, + "tokens_seen": 2976514048 + }, + { + "epoch": 0.9, + "learning_rate": 7.94383149448345e-05, + "loss": 1.3327, + "theoretical_loss": 3.3512339564153923, + "tokens_seen": 2976579584 + }, + { + "epoch": 0.9, + "learning_rate": 7.942226680040122e-05, + "loss": 1.3035, + "theoretical_loss": 3.3512283226579225, + "tokens_seen": 2976645120 + }, + { + "epoch": 0.9, + "learning_rate": 7.940621865596792e-05, + "loss": 1.3872, + "theoretical_loss": 3.3512226890592167, + "tokens_seen": 2976710656 + }, + { + "epoch": 0.9, + "learning_rate": 7.939017051153462e-05, + "loss": 1.2256, + "theoretical_loss": 3.351217055619269, + "tokens_seen": 2976776192 + }, + { + "epoch": 0.9, + "learning_rate": 7.937412236710132e-05, + "loss": 1.2494, + "theoretical_loss": 3.3512114223380696, + "tokens_seen": 2976841728 + }, + { + "epoch": 0.9, + "learning_rate": 7.9358074222668e-05, + "loss": 1.1762, + "theoretical_loss": 3.3512057892156117, + "tokens_seen": 2976907264 + }, + { + "epoch": 0.9, + "learning_rate": 7.93420260782347e-05, + "loss": 1.2789, + "theoretical_loss": 3.3512001562518874, + "tokens_seen": 2976972800 + }, + { + "epoch": 0.9, + "learning_rate": 7.93259779338014e-05, + "loss": 1.1814, + "theoretical_loss": 3.3511945234468876, + "tokens_seen": 2977038336 + }, + { + "epoch": 0.9, + "learning_rate": 7.93099297893681e-05, + "loss": 1.336, + "theoretical_loss": 3.3511888908006053, + "tokens_seen": 2977103872 + }, + { + "epoch": 0.9, + "learning_rate": 7.929388164493481e-05, + "loss": 1.2589, + "theoretical_loss": 3.351183258313032, + "tokens_seen": 2977169408 + }, + { + "epoch": 0.9, + "learning_rate": 7.927783350050151e-05, + "loss": 1.1557, + "theoretical_loss": 3.3511776259841604, + "tokens_seen": 2977234944 + }, + { + "epoch": 0.9, + "learning_rate": 7.926178535606821e-05, + "loss": 1.2589, + "theoretical_loss": 3.3511719938139817, + "tokens_seen": 2977300480 + }, + { + "epoch": 0.9, + "learning_rate": 7.924573721163491e-05, + "loss": 1.2041, + "theoretical_loss": 3.3511663618024885, + "tokens_seen": 2977366016 + }, + { + "epoch": 0.9, + "learning_rate": 7.922968906720161e-05, + "loss": 1.2823, + "theoretical_loss": 3.351160729949673, + "tokens_seen": 2977431552 + }, + { + "epoch": 0.9, + "learning_rate": 7.921364092276831e-05, + "loss": 1.3992, + "theoretical_loss": 3.3511550982555267, + "tokens_seen": 2977497088 + }, + { + "epoch": 0.9, + "learning_rate": 7.919759277833501e-05, + "loss": 1.3102, + "theoretical_loss": 3.351149466720042, + "tokens_seen": 2977562624 + }, + { + "epoch": 0.9, + "learning_rate": 7.918154463390171e-05, + "loss": 1.3057, + "theoretical_loss": 3.3511438353432106, + "tokens_seen": 2977628160 + }, + { + "epoch": 0.9, + "learning_rate": 7.91654964894684e-05, + "loss": 1.3674, + "theoretical_loss": 3.351138204125025, + "tokens_seen": 2977693696 + }, + { + "epoch": 0.9, + "learning_rate": 7.91494483450351e-05, + "loss": 1.3001, + "theoretical_loss": 3.351132573065477, + "tokens_seen": 2977759232 + }, + { + "epoch": 0.9, + "learning_rate": 7.91334002006018e-05, + "loss": 1.2774, + "theoretical_loss": 3.3511269421645586, + "tokens_seen": 2977824768 + }, + { + "epoch": 0.9, + "learning_rate": 7.911735205616851e-05, + "loss": 1.3671, + "theoretical_loss": 3.3511213114222618, + "tokens_seen": 2977890304 + }, + { + "epoch": 0.9, + "learning_rate": 7.910130391173521e-05, + "loss": 1.3214, + "theoretical_loss": 3.3511156808385785, + "tokens_seen": 2977955840 + }, + { + "epoch": 0.9, + "learning_rate": 7.908525576730192e-05, + "loss": 1.2642, + "theoretical_loss": 3.351110050413501, + "tokens_seen": 2978021376 + }, + { + "epoch": 0.9, + "learning_rate": 7.906920762286862e-05, + "loss": 1.2162, + "theoretical_loss": 3.351104420147022, + "tokens_seen": 2978086912 + }, + { + "epoch": 0.9, + "learning_rate": 7.905315947843532e-05, + "loss": 1.3355, + "theoretical_loss": 3.3510987900391322, + "tokens_seen": 2978152448 + }, + { + "epoch": 0.9, + "learning_rate": 7.903711133400202e-05, + "loss": 1.1725, + "theoretical_loss": 3.3510931600898246, + "tokens_seen": 2978217984 + }, + { + "epoch": 0.9, + "learning_rate": 7.902106318956872e-05, + "loss": 1.2762, + "theoretical_loss": 3.3510875302990906, + "tokens_seen": 2978283520 + }, + { + "epoch": 0.9, + "learning_rate": 7.90050150451354e-05, + "loss": 1.2357, + "theoretical_loss": 3.3510819006669226, + "tokens_seen": 2978349056 + }, + { + "epoch": 0.9, + "learning_rate": 7.89889669007021e-05, + "loss": 1.2282, + "theoretical_loss": 3.3510762711933126, + "tokens_seen": 2978414592 + }, + { + "epoch": 0.9, + "learning_rate": 7.89729187562688e-05, + "loss": 1.3008, + "theoretical_loss": 3.3510706418782528, + "tokens_seen": 2978480128 + }, + { + "epoch": 0.9, + "learning_rate": 7.895687061183551e-05, + "loss": 1.4011, + "theoretical_loss": 3.3510650127217354, + "tokens_seen": 2978545664 + }, + { + "epoch": 0.9, + "objective/train/docs_used": 1082921, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.2551854848861694, + "objective/train/theoretical_loss": 3.3510593837237517, + "objective/train/tokens_used": 2999071200, + "theoretical_loss": 3.3510593837237517, + "tokens_seen": 2978611200 + }, + { + "epoch": 0.9, + "learning_rate": 7.894082246740221e-05, + "loss": 1.2279, + "theoretical_loss": 3.3510593837237517, + "tokens_seen": 2978611200 + }, + { + "epoch": 0.9, + "learning_rate": 7.892477432296891e-05, + "loss": 1.2024, + "theoretical_loss": 3.3510537548842945, + "tokens_seen": 2978676736 + }, + { + "epoch": 0.9, + "learning_rate": 7.890872617853561e-05, + "loss": 1.4152, + "theoretical_loss": 3.3510481262033553, + "tokens_seen": 2978742272 + }, + { + "epoch": 0.9, + "learning_rate": 7.889267803410231e-05, + "loss": 1.1846, + "theoretical_loss": 3.3510424976809263, + "tokens_seen": 2978807808 + }, + { + "epoch": 0.9, + "learning_rate": 7.887662988966901e-05, + "loss": 1.3266, + "theoretical_loss": 3.3510368693170003, + "tokens_seen": 2978873344 + }, + { + "epoch": 0.9, + "learning_rate": 7.886058174523571e-05, + "loss": 1.2221, + "theoretical_loss": 3.351031241111568, + "tokens_seen": 2978938880 + }, + { + "epoch": 0.9, + "learning_rate": 7.884453360080241e-05, + "loss": 1.3296, + "theoretical_loss": 3.3510256130646225, + "tokens_seen": 2979004416 + }, + { + "epoch": 0.9, + "learning_rate": 7.882848545636911e-05, + "loss": 1.1914, + "theoretical_loss": 3.3510199851761553, + "tokens_seen": 2979069952 + }, + { + "epoch": 0.9, + "learning_rate": 7.881243731193581e-05, + "loss": 1.1759, + "theoretical_loss": 3.351014357446159, + "tokens_seen": 2979135488 + }, + { + "epoch": 0.9, + "learning_rate": 7.879638916750251e-05, + "loss": 1.254, + "theoretical_loss": 3.351008729874625, + "tokens_seen": 2979201024 + }, + { + "epoch": 0.9, + "learning_rate": 7.878034102306921e-05, + "loss": 1.2644, + "theoretical_loss": 3.3510031024615454, + "tokens_seen": 2979266560 + }, + { + "epoch": 0.9, + "learning_rate": 7.876429287863591e-05, + "loss": 1.3612, + "theoretical_loss": 3.350997475206913, + "tokens_seen": 2979332096 + }, + { + "epoch": 0.9, + "learning_rate": 7.874824473420262e-05, + "loss": 1.2476, + "theoretical_loss": 3.350991848110719, + "tokens_seen": 2979397632 + }, + { + "epoch": 0.9, + "learning_rate": 7.873219658976932e-05, + "loss": 1.169, + "theoretical_loss": 3.350986221172956, + "tokens_seen": 2979463168 + }, + { + "epoch": 0.9, + "learning_rate": 7.871614844533602e-05, + "loss": 1.2864, + "theoretical_loss": 3.350980594393616, + "tokens_seen": 2979528704 + }, + { + "epoch": 0.9, + "learning_rate": 7.870010030090272e-05, + "loss": 1.3659, + "theoretical_loss": 3.3509749677726908, + "tokens_seen": 2979594240 + }, + { + "epoch": 0.9, + "learning_rate": 7.868405215646942e-05, + "loss": 1.2824, + "theoretical_loss": 3.3509693413101727, + "tokens_seen": 2979659776 + }, + { + "epoch": 0.9, + "learning_rate": 7.86680040120361e-05, + "loss": 1.2591, + "theoretical_loss": 3.3509637150060536, + "tokens_seen": 2979725312 + }, + { + "epoch": 0.9, + "learning_rate": 7.86519558676028e-05, + "loss": 1.1977, + "theoretical_loss": 3.3509580888603256, + "tokens_seen": 2979790848 + }, + { + "epoch": 0.9, + "learning_rate": 7.86359077231695e-05, + "loss": 1.2795, + "theoretical_loss": 3.3509524628729803, + "tokens_seen": 2979856384 + }, + { + "epoch": 0.9, + "learning_rate": 7.861985957873621e-05, + "loss": 1.216, + "theoretical_loss": 3.3509468370440105, + "tokens_seen": 2979921920 + }, + { + "epoch": 0.9, + "learning_rate": 7.860381143430291e-05, + "loss": 1.191, + "theoretical_loss": 3.3509412113734083, + "tokens_seen": 2979987456 + }, + { + "epoch": 0.9, + "learning_rate": 7.858776328986961e-05, + "loss": 1.2445, + "theoretical_loss": 3.350935585861165, + "tokens_seen": 2980052992 + }, + { + "epoch": 0.9, + "learning_rate": 7.857171514543631e-05, + "loss": 1.3202, + "theoretical_loss": 3.3509299605072735, + "tokens_seen": 2980118528 + }, + { + "epoch": 0.9, + "learning_rate": 7.855566700100301e-05, + "loss": 1.3512, + "theoretical_loss": 3.350924335311725, + "tokens_seen": 2980184064 + }, + { + "epoch": 0.9, + "learning_rate": 7.853961885656971e-05, + "loss": 1.2062, + "theoretical_loss": 3.350918710274512, + "tokens_seen": 2980249600 + }, + { + "epoch": 0.9, + "learning_rate": 7.852357071213642e-05, + "loss": 1.368, + "theoretical_loss": 3.3509130853956273, + "tokens_seen": 2980315136 + }, + { + "epoch": 0.9, + "learning_rate": 7.850752256770311e-05, + "loss": 1.3547, + "theoretical_loss": 3.3509074606750615, + "tokens_seen": 2980380672 + }, + { + "epoch": 0.9, + "learning_rate": 7.849147442326981e-05, + "loss": 1.2842, + "theoretical_loss": 3.350901836112808, + "tokens_seen": 2980446208 + }, + { + "epoch": 0.9, + "learning_rate": 7.847542627883651e-05, + "loss": 1.2748, + "theoretical_loss": 3.3508962117088577, + "tokens_seen": 2980511744 + }, + { + "epoch": 0.9, + "learning_rate": 7.845937813440321e-05, + "loss": 1.2435, + "theoretical_loss": 3.350890587463203, + "tokens_seen": 2980577280 + }, + { + "epoch": 0.9, + "learning_rate": 7.844332998996991e-05, + "loss": 1.1956, + "theoretical_loss": 3.350884963375837, + "tokens_seen": 2980642816 + }, + { + "epoch": 0.9, + "learning_rate": 7.842728184553661e-05, + "loss": 1.1514, + "theoretical_loss": 3.3508793394467506, + "tokens_seen": 2980708352 + }, + { + "epoch": 0.9, + "learning_rate": 7.841123370110332e-05, + "loss": 1.2909, + "theoretical_loss": 3.3508737156759363, + "tokens_seen": 2980773888 + }, + { + "epoch": 0.9, + "learning_rate": 7.839518555667002e-05, + "loss": 1.2362, + "theoretical_loss": 3.3508680920633855, + "tokens_seen": 2980839424 + }, + { + "epoch": 0.9, + "learning_rate": 7.837913741223672e-05, + "loss": 1.3846, + "theoretical_loss": 3.3508624686090913, + "tokens_seen": 2980904960 + }, + { + "epoch": 0.9, + "learning_rate": 7.836308926780342e-05, + "loss": 1.3159, + "theoretical_loss": 3.3508568453130456, + "tokens_seen": 2980970496 + }, + { + "epoch": 0.9, + "learning_rate": 7.834704112337012e-05, + "loss": 1.3769, + "theoretical_loss": 3.3508512221752396, + "tokens_seen": 2981036032 + }, + { + "epoch": 0.9, + "learning_rate": 7.833099297893682e-05, + "loss": 1.2319, + "theoretical_loss": 3.3508455991956665, + "tokens_seen": 2981101568 + }, + { + "epoch": 0.9, + "learning_rate": 7.83149448345035e-05, + "loss": 1.2902, + "theoretical_loss": 3.3508399763743175, + "tokens_seen": 2981167104 + }, + { + "epoch": 0.9, + "learning_rate": 7.82988966900702e-05, + "loss": 1.2379, + "theoretical_loss": 3.350834353711185, + "tokens_seen": 2981232640 + }, + { + "epoch": 0.9, + "learning_rate": 7.82828485456369e-05, + "loss": 1.2655, + "theoretical_loss": 3.3508287312062612, + "tokens_seen": 2981298176 + }, + { + "epoch": 0.9, + "learning_rate": 7.826680040120361e-05, + "loss": 1.2645, + "theoretical_loss": 3.350823108859538, + "tokens_seen": 2981363712 + }, + { + "epoch": 0.9, + "learning_rate": 7.825075225677031e-05, + "loss": 1.1769, + "theoretical_loss": 3.3508174866710077, + "tokens_seen": 2981429248 + }, + { + "epoch": 0.9, + "learning_rate": 7.823470411233701e-05, + "loss": 1.2329, + "theoretical_loss": 3.3508118646406615, + "tokens_seen": 2981494784 + }, + { + "epoch": 0.9, + "learning_rate": 7.821865596790372e-05, + "loss": 1.2756, + "theoretical_loss": 3.350806242768493, + "tokens_seen": 2981560320 + }, + { + "epoch": 0.9, + "learning_rate": 7.820260782347042e-05, + "loss": 1.2474, + "theoretical_loss": 3.3508006210544927, + "tokens_seen": 2981625856 + }, + { + "epoch": 0.9, + "learning_rate": 7.818655967903712e-05, + "loss": 1.2415, + "theoretical_loss": 3.350794999498654, + "tokens_seen": 2981691392 + }, + { + "epoch": 0.9, + "learning_rate": 7.817051153460382e-05, + "loss": 1.2507, + "theoretical_loss": 3.350789378100968, + "tokens_seen": 2981756928 + }, + { + "epoch": 0.9, + "learning_rate": 7.815446339017051e-05, + "loss": 1.2702, + "theoretical_loss": 3.350783756861427, + "tokens_seen": 2981822464 + }, + { + "epoch": 0.9, + "objective/train/docs_used": 1084154, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.2337791919708252, + "objective/train/theoretical_loss": 3.3507781357800237, + "objective/train/tokens_used": 3002348000, + "theoretical_loss": 3.3507781357800237, + "tokens_seen": 2981888000 + }, + { + "epoch": 0.9, + "learning_rate": 7.813841524573721e-05, + "loss": 1.22, + "theoretical_loss": 3.3507781357800237, + "tokens_seen": 2981888000 + }, + { + "epoch": 0.9, + "learning_rate": 7.812236710130391e-05, + "loss": 1.2785, + "theoretical_loss": 3.3507725148567493, + "tokens_seen": 2981953536 + }, + { + "epoch": 0.9, + "learning_rate": 7.810631895687061e-05, + "loss": 1.1812, + "theoretical_loss": 3.3507668940915964, + "tokens_seen": 2982019072 + }, + { + "epoch": 0.9, + "learning_rate": 7.809027081243731e-05, + "loss": 1.1568, + "theoretical_loss": 3.350761273484557, + "tokens_seen": 2982084608 + }, + { + "epoch": 0.9, + "learning_rate": 7.807422266800401e-05, + "loss": 1.3081, + "theoretical_loss": 3.3507556530356233, + "tokens_seen": 2982150144 + }, + { + "epoch": 0.9, + "learning_rate": 7.805817452357072e-05, + "loss": 1.2451, + "theoretical_loss": 3.350750032744787, + "tokens_seen": 2982215680 + }, + { + "epoch": 0.9, + "learning_rate": 7.804212637913742e-05, + "loss": 1.1929, + "theoretical_loss": 3.3507444126120403, + "tokens_seen": 2982281216 + }, + { + "epoch": 0.9, + "learning_rate": 7.802607823470412e-05, + "loss": 1.3022, + "theoretical_loss": 3.3507387926373755, + "tokens_seen": 2982346752 + }, + { + "epoch": 0.9, + "learning_rate": 7.801003009027082e-05, + "loss": 1.319, + "theoretical_loss": 3.3507331728207843, + "tokens_seen": 2982412288 + }, + { + "epoch": 0.9, + "learning_rate": 7.799398194583752e-05, + "loss": 1.3098, + "theoretical_loss": 3.3507275531622596, + "tokens_seen": 2982477824 + }, + { + "epoch": 0.9, + "learning_rate": 7.797793380140422e-05, + "loss": 1.2978, + "theoretical_loss": 3.3507219336617924, + "tokens_seen": 2982543360 + }, + { + "epoch": 0.9, + "learning_rate": 7.79618856569709e-05, + "loss": 1.2136, + "theoretical_loss": 3.350716314319375, + "tokens_seen": 2982608896 + }, + { + "epoch": 0.9, + "learning_rate": 7.79458375125376e-05, + "loss": 1.3087, + "theoretical_loss": 3.3507106951350005, + "tokens_seen": 2982674432 + }, + { + "epoch": 0.9, + "learning_rate": 7.792978936810431e-05, + "loss": 1.2577, + "theoretical_loss": 3.3507050761086603, + "tokens_seen": 2982739968 + }, + { + "epoch": 0.9, + "learning_rate": 7.791374122367102e-05, + "loss": 1.2268, + "theoretical_loss": 3.3506994572403457, + "tokens_seen": 2982805504 + }, + { + "epoch": 0.9, + "learning_rate": 7.789769307923772e-05, + "loss": 1.3058, + "theoretical_loss": 3.35069383853005, + "tokens_seen": 2982871040 + }, + { + "epoch": 0.9, + "learning_rate": 7.788164493480442e-05, + "loss": 1.3389, + "theoretical_loss": 3.350688219977765, + "tokens_seen": 2982936576 + }, + { + "epoch": 0.9, + "learning_rate": 7.786559679037112e-05, + "loss": 1.3176, + "theoretical_loss": 3.350682601583482, + "tokens_seen": 2983002112 + }, + { + "epoch": 0.9, + "learning_rate": 7.784954864593782e-05, + "loss": 1.1859, + "theoretical_loss": 3.350676983347194, + "tokens_seen": 2983067648 + }, + { + "epoch": 0.9, + "learning_rate": 7.783350050150452e-05, + "loss": 1.1936, + "theoretical_loss": 3.3506713652688926, + "tokens_seen": 2983133184 + }, + { + "epoch": 0.9, + "learning_rate": 7.781745235707123e-05, + "loss": 1.1854, + "theoretical_loss": 3.3506657473485704, + "tokens_seen": 2983198720 + }, + { + "epoch": 0.9, + "learning_rate": 7.780140421263791e-05, + "loss": 1.2712, + "theoretical_loss": 3.3506601295862186, + "tokens_seen": 2983264256 + }, + { + "epoch": 0.9, + "learning_rate": 7.778535606820461e-05, + "loss": 1.1953, + "theoretical_loss": 3.3506545119818303, + "tokens_seen": 2983329792 + }, + { + "epoch": 0.9, + "learning_rate": 7.776930792377131e-05, + "loss": 1.2934, + "theoretical_loss": 3.350648894535397, + "tokens_seen": 2983395328 + }, + { + "epoch": 0.9, + "learning_rate": 7.775325977933801e-05, + "loss": 1.2168, + "theoretical_loss": 3.3506432772469106, + "tokens_seen": 2983460864 + }, + { + "epoch": 0.9, + "learning_rate": 7.773721163490471e-05, + "loss": 1.2161, + "theoretical_loss": 3.350637660116364, + "tokens_seen": 2983526400 + }, + { + "epoch": 0.9, + "learning_rate": 7.772116349047142e-05, + "loss": 1.28, + "theoretical_loss": 3.350632043143748, + "tokens_seen": 2983591936 + }, + { + "epoch": 0.9, + "learning_rate": 7.770511534603812e-05, + "loss": 1.1725, + "theoretical_loss": 3.3506264263290566, + "tokens_seen": 2983657472 + }, + { + "epoch": 0.9, + "learning_rate": 7.768906720160482e-05, + "loss": 1.2261, + "theoretical_loss": 3.35062080967228, + "tokens_seen": 2983723008 + }, + { + "epoch": 0.9, + "learning_rate": 7.767301905717152e-05, + "loss": 1.2784, + "theoretical_loss": 3.350615193173411, + "tokens_seen": 2983788544 + }, + { + "epoch": 0.9, + "learning_rate": 7.765697091273822e-05, + "loss": 1.1661, + "theoretical_loss": 3.3506095768324418, + "tokens_seen": 2983854080 + }, + { + "epoch": 0.9, + "learning_rate": 7.764092276830492e-05, + "loss": 1.2756, + "theoretical_loss": 3.350603960649365, + "tokens_seen": 2983919616 + }, + { + "epoch": 0.9, + "learning_rate": 7.762487462387162e-05, + "loss": 1.2177, + "theoretical_loss": 3.3505983446241716, + "tokens_seen": 2983985152 + }, + { + "epoch": 0.9, + "learning_rate": 7.760882647943832e-05, + "loss": 1.2278, + "theoretical_loss": 3.350592728756854, + "tokens_seen": 2984050688 + }, + { + "epoch": 0.9, + "learning_rate": 7.759277833500502e-05, + "loss": 1.2126, + "theoretical_loss": 3.350587113047405, + "tokens_seen": 2984116224 + }, + { + "epoch": 0.9, + "learning_rate": 7.757673019057172e-05, + "loss": 1.1992, + "theoretical_loss": 3.350581497495816, + "tokens_seen": 2984181760 + }, + { + "epoch": 0.9, + "learning_rate": 7.756068204613842e-05, + "loss": 1.1597, + "theoretical_loss": 3.3505758821020795, + "tokens_seen": 2984247296 + }, + { + "epoch": 0.9, + "learning_rate": 7.754463390170512e-05, + "loss": 1.1129, + "theoretical_loss": 3.350570266866187, + "tokens_seen": 2984312832 + }, + { + "epoch": 0.9, + "learning_rate": 7.752858575727182e-05, + "loss": 1.2476, + "theoretical_loss": 3.3505646517881313, + "tokens_seen": 2984378368 + }, + { + "epoch": 0.9, + "learning_rate": 7.751253761283852e-05, + "loss": 1.256, + "theoretical_loss": 3.3505590368679044, + "tokens_seen": 2984443904 + }, + { + "epoch": 0.9, + "learning_rate": 7.749648946840522e-05, + "loss": 1.263, + "theoretical_loss": 3.3505534221054982, + "tokens_seen": 2984509440 + }, + { + "epoch": 0.9, + "learning_rate": 7.748044132397192e-05, + "loss": 1.3046, + "theoretical_loss": 3.3505478075009045, + "tokens_seen": 2984574976 + }, + { + "epoch": 0.9, + "learning_rate": 7.746439317953863e-05, + "loss": 1.1567, + "theoretical_loss": 3.3505421930541157, + "tokens_seen": 2984640512 + }, + { + "epoch": 0.9, + "learning_rate": 7.744834503510531e-05, + "loss": 1.243, + "theoretical_loss": 3.350536578765124, + "tokens_seen": 2984706048 + }, + { + "epoch": 0.9, + "learning_rate": 7.743229689067201e-05, + "loss": 1.2893, + "theoretical_loss": 3.350530964633921, + "tokens_seen": 2984771584 + }, + { + "epoch": 0.9, + "learning_rate": 7.741624874623871e-05, + "loss": 1.3087, + "theoretical_loss": 3.3505253506605, + "tokens_seen": 2984837120 + }, + { + "epoch": 0.9, + "learning_rate": 7.740020060180541e-05, + "loss": 1.3399, + "theoretical_loss": 3.350519736844852, + "tokens_seen": 2984902656 + }, + { + "epoch": 0.9, + "learning_rate": 7.738415245737212e-05, + "loss": 1.1872, + "theoretical_loss": 3.350514123186969, + "tokens_seen": 2984968192 + }, + { + "epoch": 0.9, + "learning_rate": 7.736810431293882e-05, + "loss": 1.1944, + "theoretical_loss": 3.350508509686844, + "tokens_seen": 2985033728 + }, + { + "epoch": 0.9, + "learning_rate": 7.735205616850552e-05, + "loss": 1.1847, + "theoretical_loss": 3.3505028963444685, + "tokens_seen": 2985099264 + }, + { + "epoch": 0.9, + "objective/train/docs_used": 1085234, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.2422221899032593, + "objective/train/theoretical_loss": 3.3504972831598345, + "objective/train/tokens_used": 3005624800, + "theoretical_loss": 3.3504972831598345, + "tokens_seen": 2985164800 + }, + { + "epoch": 0.9, + "learning_rate": 7.733600802407222e-05, + "loss": 1.1936, + "theoretical_loss": 3.3504972831598345, + "tokens_seen": 2985164800 + }, + { + "epoch": 0.9, + "learning_rate": 7.731995987963893e-05, + "loss": 1.2617, + "theoretical_loss": 3.3504916701329344, + "tokens_seen": 2985230336 + }, + { + "epoch": 0.9, + "learning_rate": 7.730391173520563e-05, + "loss": 1.2394, + "theoretical_loss": 3.3504860572637605, + "tokens_seen": 2985295872 + }, + { + "epoch": 0.9, + "learning_rate": 7.728786359077232e-05, + "loss": 1.273, + "theoretical_loss": 3.350480444552304, + "tokens_seen": 2985361408 + }, + { + "epoch": 0.9, + "learning_rate": 7.727181544633902e-05, + "loss": 1.2218, + "theoretical_loss": 3.3504748319985582, + "tokens_seen": 2985426944 + }, + { + "epoch": 0.9, + "learning_rate": 7.725576730190572e-05, + "loss": 1.2733, + "theoretical_loss": 3.3504692196025143, + "tokens_seen": 2985492480 + }, + { + "epoch": 0.9, + "learning_rate": 7.723971915747242e-05, + "loss": 1.2916, + "theoretical_loss": 3.350463607364165, + "tokens_seen": 2985558016 + }, + { + "epoch": 0.9, + "learning_rate": 7.722367101303912e-05, + "loss": 1.3614, + "theoretical_loss": 3.3504579952835023, + "tokens_seen": 2985623552 + }, + { + "epoch": 0.9, + "learning_rate": 7.720762286860582e-05, + "loss": 1.2419, + "theoretical_loss": 3.3504523833605178, + "tokens_seen": 2985689088 + }, + { + "epoch": 0.9, + "learning_rate": 7.719157472417252e-05, + "loss": 1.2982, + "theoretical_loss": 3.350446771595204, + "tokens_seen": 2985754624 + }, + { + "epoch": 0.9, + "learning_rate": 7.717552657973922e-05, + "loss": 1.2575, + "theoretical_loss": 3.350441159987553, + "tokens_seen": 2985820160 + }, + { + "epoch": 0.9, + "learning_rate": 7.715947843530592e-05, + "loss": 1.2234, + "theoretical_loss": 3.350435548537557, + "tokens_seen": 2985885696 + }, + { + "epoch": 0.9, + "learning_rate": 7.714343029087262e-05, + "loss": 1.2486, + "theoretical_loss": 3.350429937245208, + "tokens_seen": 2985951232 + }, + { + "epoch": 0.9, + "learning_rate": 7.712738214643933e-05, + "loss": 1.2584, + "theoretical_loss": 3.350424326110498, + "tokens_seen": 2986016768 + }, + { + "epoch": 0.9, + "learning_rate": 7.711133400200603e-05, + "loss": 1.2159, + "theoretical_loss": 3.3504187151334195, + "tokens_seen": 2986082304 + }, + { + "epoch": 0.9, + "learning_rate": 7.709528585757271e-05, + "loss": 1.2888, + "theoretical_loss": 3.350413104313964, + "tokens_seen": 2986147840 + }, + { + "epoch": 0.9, + "learning_rate": 7.707923771313941e-05, + "loss": 1.2746, + "theoretical_loss": 3.350407493652124, + "tokens_seen": 2986213376 + }, + { + "epoch": 0.9, + "learning_rate": 7.706318956870611e-05, + "loss": 1.2269, + "theoretical_loss": 3.350401883147892, + "tokens_seen": 2986278912 + }, + { + "epoch": 0.9, + "learning_rate": 7.704714142427281e-05, + "loss": 1.2801, + "theoretical_loss": 3.3503962728012593, + "tokens_seen": 2986344448 + }, + { + "epoch": 0.9, + "learning_rate": 7.703109327983952e-05, + "loss": 1.2849, + "theoretical_loss": 3.3503906626122184, + "tokens_seen": 2986409984 + }, + { + "epoch": 0.9, + "learning_rate": 7.701504513540623e-05, + "loss": 1.284, + "theoretical_loss": 3.3503850525807612, + "tokens_seen": 2986475520 + }, + { + "epoch": 0.91, + "learning_rate": 7.699899699097293e-05, + "loss": 1.2396, + "theoretical_loss": 3.3503794427068803, + "tokens_seen": 2986541056 + }, + { + "epoch": 0.91, + "learning_rate": 7.698294884653963e-05, + "loss": 1.2227, + "theoretical_loss": 3.350373832990568, + "tokens_seen": 2986606592 + }, + { + "epoch": 0.91, + "learning_rate": 7.696690070210633e-05, + "loss": 1.2374, + "theoretical_loss": 3.3503682234318153, + "tokens_seen": 2986672128 + }, + { + "epoch": 0.91, + "learning_rate": 7.695085255767303e-05, + "loss": 1.3227, + "theoretical_loss": 3.3503626140306153, + "tokens_seen": 2986737664 + }, + { + "epoch": 0.91, + "learning_rate": 7.693480441323972e-05, + "loss": 1.1296, + "theoretical_loss": 3.3503570047869595, + "tokens_seen": 2986803200 + }, + { + "epoch": 0.91, + "learning_rate": 7.691875626880642e-05, + "loss": 1.2462, + "theoretical_loss": 3.3503513957008404, + "tokens_seen": 2986868736 + }, + { + "epoch": 0.91, + "learning_rate": 7.690270812437312e-05, + "loss": 1.2799, + "theoretical_loss": 3.35034578677225, + "tokens_seen": 2986934272 + }, + { + "epoch": 0.91, + "learning_rate": 7.688665997993982e-05, + "loss": 1.2417, + "theoretical_loss": 3.3503401780011806, + "tokens_seen": 2986999808 + }, + { + "epoch": 0.91, + "learning_rate": 7.687061183550652e-05, + "loss": 1.14, + "theoretical_loss": 3.3503345693876243, + "tokens_seen": 2987065344 + }, + { + "epoch": 0.91, + "learning_rate": 7.685456369107322e-05, + "loss": 1.2677, + "theoretical_loss": 3.3503289609315727, + "tokens_seen": 2987130880 + }, + { + "epoch": 0.91, + "learning_rate": 7.683851554663992e-05, + "loss": 1.3044, + "theoretical_loss": 3.3503233526330183, + "tokens_seen": 2987196416 + }, + { + "epoch": 0.91, + "learning_rate": 7.682246740220662e-05, + "loss": 1.2316, + "theoretical_loss": 3.350317744491954, + "tokens_seen": 2987261952 + }, + { + "epoch": 0.91, + "learning_rate": 7.680641925777332e-05, + "loss": 1.1335, + "theoretical_loss": 3.35031213650837, + "tokens_seen": 2987327488 + }, + { + "epoch": 0.91, + "learning_rate": 7.679037111334003e-05, + "loss": 1.2705, + "theoretical_loss": 3.3503065286822604, + "tokens_seen": 2987393024 + }, + { + "epoch": 0.91, + "learning_rate": 7.677432296890673e-05, + "loss": 1.2032, + "theoretical_loss": 3.3503009210136163, + "tokens_seen": 2987458560 + }, + { + "epoch": 0.91, + "learning_rate": 7.675827482447343e-05, + "loss": 1.2007, + "theoretical_loss": 3.35029531350243, + "tokens_seen": 2987524096 + }, + { + "epoch": 0.91, + "learning_rate": 7.674222668004011e-05, + "loss": 1.1586, + "theoretical_loss": 3.3502897061486934, + "tokens_seen": 2987589632 + }, + { + "epoch": 0.91, + "learning_rate": 7.672617853560683e-05, + "loss": 1.2678, + "theoretical_loss": 3.3502840989523994, + "tokens_seen": 2987655168 + }, + { + "epoch": 0.91, + "learning_rate": 7.671013039117353e-05, + "loss": 1.2864, + "theoretical_loss": 3.3502784919135395, + "tokens_seen": 2987720704 + }, + { + "epoch": 0.91, + "learning_rate": 7.669408224674023e-05, + "loss": 1.2829, + "theoretical_loss": 3.3502728850321057, + "tokens_seen": 2987786240 + }, + { + "epoch": 0.91, + "learning_rate": 7.667803410230693e-05, + "loss": 1.3407, + "theoretical_loss": 3.3502672783080905, + "tokens_seen": 2987851776 + }, + { + "epoch": 0.91, + "learning_rate": 7.666198595787363e-05, + "loss": 1.2723, + "theoretical_loss": 3.3502616717414857, + "tokens_seen": 2987917312 + }, + { + "epoch": 0.91, + "learning_rate": 7.664593781344033e-05, + "loss": 1.2701, + "theoretical_loss": 3.3502560653322835, + "tokens_seen": 2987982848 + }, + { + "epoch": 0.91, + "learning_rate": 7.662988966900703e-05, + "loss": 1.374, + "theoretical_loss": 3.3502504590804767, + "tokens_seen": 2988048384 + }, + { + "epoch": 0.91, + "learning_rate": 7.661384152457373e-05, + "loss": 1.2295, + "theoretical_loss": 3.3502448529860565, + "tokens_seen": 2988113920 + }, + { + "epoch": 0.91, + "learning_rate": 7.659779338014042e-05, + "loss": 1.2708, + "theoretical_loss": 3.350239247049015, + "tokens_seen": 2988179456 + }, + { + "epoch": 0.91, + "learning_rate": 7.658174523570712e-05, + "loss": 1.2188, + "theoretical_loss": 3.3502336412693454, + "tokens_seen": 2988244992 + }, + { + "epoch": 0.91, + "learning_rate": 7.656569709127382e-05, + "loss": 1.3242, + "theoretical_loss": 3.350228035647039, + "tokens_seen": 2988310528 + }, + { + "epoch": 0.91, + "learning_rate": 7.654964894684052e-05, + "loss": 1.2922, + "theoretical_loss": 3.350222430182088, + "tokens_seen": 2988376064 + }, + { + "epoch": 0.91, + "objective/train/docs_used": 1086412, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.1151775121688843, + "objective/train/theoretical_loss": 3.350216824874485, + "objective/train/tokens_used": 3008901600, + "theoretical_loss": 3.350216824874485, + "tokens_seen": 2988441600 + }, + { + "epoch": 0.91, + "learning_rate": 7.653360080240722e-05, + "loss": 1.2368, + "theoretical_loss": 3.350216824874485, + "tokens_seen": 2988441600 + }, + { + "epoch": 0.91, + "learning_rate": 7.651755265797392e-05, + "loss": 1.3434, + "theoretical_loss": 3.3502112197242213, + "tokens_seen": 2988507136 + }, + { + "epoch": 0.91, + "learning_rate": 7.650150451354062e-05, + "loss": 1.2166, + "theoretical_loss": 3.3502056147312897, + "tokens_seen": 2988572672 + }, + { + "epoch": 0.91, + "learning_rate": 7.648545636910732e-05, + "loss": 1.1825, + "theoretical_loss": 3.3502000098956817, + "tokens_seen": 2988638208 + }, + { + "epoch": 0.91, + "learning_rate": 7.646940822467402e-05, + "loss": 1.1968, + "theoretical_loss": 3.3501944052173904, + "tokens_seen": 2988703744 + }, + { + "epoch": 0.91, + "learning_rate": 7.645336008024072e-05, + "loss": 1.2481, + "theoretical_loss": 3.350188800696407, + "tokens_seen": 2988769280 + }, + { + "epoch": 0.91, + "learning_rate": 7.643731193580743e-05, + "loss": 1.2591, + "theoretical_loss": 3.350183196332724, + "tokens_seen": 2988834816 + }, + { + "epoch": 0.91, + "learning_rate": 7.642126379137413e-05, + "loss": 1.3002, + "theoretical_loss": 3.350177592126334, + "tokens_seen": 2988900352 + }, + { + "epoch": 0.91, + "learning_rate": 7.640521564694083e-05, + "loss": 1.1387, + "theoretical_loss": 3.3501719880772285, + "tokens_seen": 2988965888 + }, + { + "epoch": 0.91, + "learning_rate": 7.638916750250753e-05, + "loss": 1.2516, + "theoretical_loss": 3.3501663841854, + "tokens_seen": 2989031424 + }, + { + "epoch": 0.91, + "learning_rate": 7.637311935807423e-05, + "loss": 1.2978, + "theoretical_loss": 3.3501607804508406, + "tokens_seen": 2989096960 + }, + { + "epoch": 0.91, + "learning_rate": 7.635707121364093e-05, + "loss": 1.2653, + "theoretical_loss": 3.3501551768735416, + "tokens_seen": 2989162496 + }, + { + "epoch": 0.91, + "learning_rate": 7.634102306920763e-05, + "loss": 1.1745, + "theoretical_loss": 3.3501495734534963, + "tokens_seen": 2989228032 + }, + { + "epoch": 0.91, + "learning_rate": 7.632497492477433e-05, + "loss": 1.2107, + "theoretical_loss": 3.3501439701906968, + "tokens_seen": 2989293568 + }, + { + "epoch": 0.91, + "learning_rate": 7.630892678034103e-05, + "loss": 1.3375, + "theoretical_loss": 3.350138367085134, + "tokens_seen": 2989359104 + }, + { + "epoch": 0.91, + "learning_rate": 7.629287863590773e-05, + "loss": 1.2601, + "theoretical_loss": 3.3501327641368013, + "tokens_seen": 2989424640 + }, + { + "epoch": 0.91, + "learning_rate": 7.627683049147443e-05, + "loss": 1.2121, + "theoretical_loss": 3.3501271613456907, + "tokens_seen": 2989490176 + }, + { + "epoch": 0.91, + "learning_rate": 7.626078234704113e-05, + "loss": 1.3942, + "theoretical_loss": 3.350121558711794, + "tokens_seen": 2989555712 + }, + { + "epoch": 0.91, + "learning_rate": 7.624473420260782e-05, + "loss": 1.1356, + "theoretical_loss": 3.350115956235103, + "tokens_seen": 2989621248 + }, + { + "epoch": 0.91, + "learning_rate": 7.622868605817452e-05, + "loss": 1.3095, + "theoretical_loss": 3.3501103539156105, + "tokens_seen": 2989686784 + }, + { + "epoch": 0.91, + "learning_rate": 7.621263791374122e-05, + "loss": 1.329, + "theoretical_loss": 3.350104751753308, + "tokens_seen": 2989752320 + }, + { + "epoch": 0.91, + "learning_rate": 7.619658976930792e-05, + "loss": 1.3002, + "theoretical_loss": 3.3500991497481887, + "tokens_seen": 2989817856 + }, + { + "epoch": 0.91, + "learning_rate": 7.618054162487462e-05, + "loss": 1.3007, + "theoretical_loss": 3.3500935479002436, + "tokens_seen": 2989883392 + }, + { + "epoch": 0.91, + "learning_rate": 7.616449348044132e-05, + "loss": 1.2903, + "theoretical_loss": 3.3500879462094657, + "tokens_seen": 2989948928 + }, + { + "epoch": 0.91, + "learning_rate": 7.614844533600802e-05, + "loss": 1.3217, + "theoretical_loss": 3.3500823446758465, + "tokens_seen": 2990014464 + }, + { + "epoch": 0.91, + "learning_rate": 7.613239719157472e-05, + "loss": 1.1517, + "theoretical_loss": 3.350076743299378, + "tokens_seen": 2990080000 + }, + { + "epoch": 0.91, + "learning_rate": 7.611634904714144e-05, + "loss": 1.1086, + "theoretical_loss": 3.3500711420800533, + "tokens_seen": 2990145536 + }, + { + "epoch": 0.91, + "learning_rate": 7.610030090270814e-05, + "loss": 1.1566, + "theoretical_loss": 3.350065541017864, + "tokens_seen": 2990211072 + }, + { + "epoch": 0.91, + "learning_rate": 7.608425275827483e-05, + "loss": 1.2712, + "theoretical_loss": 3.350059940112802, + "tokens_seen": 2990276608 + }, + { + "epoch": 0.91, + "learning_rate": 7.606820461384153e-05, + "loss": 1.248, + "theoretical_loss": 3.35005433936486, + "tokens_seen": 2990342144 + }, + { + "epoch": 0.91, + "learning_rate": 7.605215646940823e-05, + "loss": 1.2363, + "theoretical_loss": 3.3500487387740296, + "tokens_seen": 2990407680 + }, + { + "epoch": 0.91, + "learning_rate": 7.603610832497493e-05, + "loss": 1.2169, + "theoretical_loss": 3.3500431383403035, + "tokens_seen": 2990473216 + }, + { + "epoch": 0.91, + "learning_rate": 7.602006018054163e-05, + "loss": 1.2391, + "theoretical_loss": 3.3500375380636735, + "tokens_seen": 2990538752 + }, + { + "epoch": 0.91, + "learning_rate": 7.600401203610833e-05, + "loss": 1.1735, + "theoretical_loss": 3.3500319379441317, + "tokens_seen": 2990604288 + }, + { + "epoch": 0.91, + "learning_rate": 7.598796389167503e-05, + "loss": 1.2244, + "theoretical_loss": 3.3500263379816704, + "tokens_seen": 2990669824 + }, + { + "epoch": 0.91, + "learning_rate": 7.597191574724173e-05, + "loss": 1.3174, + "theoretical_loss": 3.3500207381762817, + "tokens_seen": 2990735360 + }, + { + "epoch": 0.91, + "learning_rate": 7.595586760280843e-05, + "loss": 1.2643, + "theoretical_loss": 3.3500151385279575, + "tokens_seen": 2990800896 + }, + { + "epoch": 0.91, + "learning_rate": 7.593981945837513e-05, + "loss": 1.2478, + "theoretical_loss": 3.3500095390366904, + "tokens_seen": 2990866432 + }, + { + "epoch": 0.91, + "learning_rate": 7.592377131394183e-05, + "loss": 1.3274, + "theoretical_loss": 3.3500039397024723, + "tokens_seen": 2990931968 + }, + { + "epoch": 0.91, + "learning_rate": 7.590772316950853e-05, + "loss": 1.1827, + "theoretical_loss": 3.3499983405252958, + "tokens_seen": 2990997504 + }, + { + "epoch": 0.91, + "learning_rate": 7.589167502507522e-05, + "loss": 1.2293, + "theoretical_loss": 3.3499927415051522, + "tokens_seen": 2991063040 + }, + { + "epoch": 0.91, + "learning_rate": 7.587562688064192e-05, + "loss": 1.3008, + "theoretical_loss": 3.3499871426420347, + "tokens_seen": 2991128576 + }, + { + "epoch": 0.91, + "learning_rate": 7.585957873620862e-05, + "loss": 1.2374, + "theoretical_loss": 3.349981543935934, + "tokens_seen": 2991194112 + }, + { + "epoch": 0.91, + "learning_rate": 7.584353059177532e-05, + "loss": 1.3953, + "theoretical_loss": 3.349975945386844, + "tokens_seen": 2991259648 + }, + { + "epoch": 0.91, + "learning_rate": 7.582748244734202e-05, + "loss": 1.2749, + "theoretical_loss": 3.3499703469947555, + "tokens_seen": 2991325184 + }, + { + "epoch": 0.91, + "learning_rate": 7.581143430290874e-05, + "loss": 1.1703, + "theoretical_loss": 3.3499647487596613, + "tokens_seen": 2991390720 + }, + { + "epoch": 0.91, + "learning_rate": 7.579538615847544e-05, + "loss": 1.292, + "theoretical_loss": 3.3499591506815536, + "tokens_seen": 2991456256 + }, + { + "epoch": 0.91, + "learning_rate": 7.577933801404214e-05, + "loss": 1.271, + "theoretical_loss": 3.3499535527604243, + "tokens_seen": 2991521792 + }, + { + "epoch": 0.91, + "learning_rate": 7.576328986960884e-05, + "loss": 1.202, + "theoretical_loss": 3.3499479549962654, + "tokens_seen": 2991587328 + }, + { + "epoch": 0.91, + "learning_rate": 7.574724172517554e-05, + "loss": 1.2558, + "theoretical_loss": 3.3499423573890694, + "tokens_seen": 2991652864 + }, + { + "epoch": 0.91, + "objective/train/docs_used": 1087632, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.2733005285263062, + "objective/train/theoretical_loss": 3.349936759938829, + "objective/train/tokens_used": 3012178400, + "theoretical_loss": 3.349936759938829, + "tokens_seen": 2991718400 + }, + { + "epoch": 0.91, + "learning_rate": 7.573119358074223e-05, + "loss": 1.2065, + "theoretical_loss": 3.349936759938829, + "tokens_seen": 2991718400 + }, + { + "epoch": 0.91, + "learning_rate": 7.571514543630893e-05, + "loss": 1.267, + "theoretical_loss": 3.3499311626455346, + "tokens_seen": 2991783936 + }, + { + "epoch": 0.91, + "learning_rate": 7.569909729187563e-05, + "loss": 1.2685, + "theoretical_loss": 3.34992556550918, + "tokens_seen": 2991849472 + }, + { + "epoch": 0.91, + "learning_rate": 7.568304914744233e-05, + "loss": 1.3047, + "theoretical_loss": 3.3499199685297567, + "tokens_seen": 2991915008 + }, + { + "epoch": 0.91, + "learning_rate": 7.566700100300903e-05, + "loss": 1.3375, + "theoretical_loss": 3.349914371707257, + "tokens_seen": 2991980544 + }, + { + "epoch": 0.91, + "learning_rate": 7.565095285857573e-05, + "loss": 1.2835, + "theoretical_loss": 3.3499087750416736, + "tokens_seen": 2992046080 + }, + { + "epoch": 0.91, + "learning_rate": 7.563490471414243e-05, + "loss": 1.2805, + "theoretical_loss": 3.3499031785329976, + "tokens_seen": 2992111616 + }, + { + "epoch": 0.91, + "learning_rate": 7.561885656970913e-05, + "loss": 1.231, + "theoretical_loss": 3.349897582181222, + "tokens_seen": 2992177152 + }, + { + "epoch": 0.91, + "learning_rate": 7.560280842527583e-05, + "loss": 1.3009, + "theoretical_loss": 3.3498919859863383, + "tokens_seen": 2992242688 + }, + { + "epoch": 0.91, + "learning_rate": 7.558676028084253e-05, + "loss": 1.2401, + "theoretical_loss": 3.349886389948339, + "tokens_seen": 2992308224 + }, + { + "epoch": 0.91, + "learning_rate": 7.557071213640923e-05, + "loss": 1.1938, + "theoretical_loss": 3.349880794067217, + "tokens_seen": 2992373760 + }, + { + "epoch": 0.91, + "learning_rate": 7.555466399197593e-05, + "loss": 1.1926, + "theoretical_loss": 3.3498751983429633, + "tokens_seen": 2992439296 + }, + { + "epoch": 0.91, + "learning_rate": 7.553861584754262e-05, + "loss": 1.2155, + "theoretical_loss": 3.3498696027755703, + "tokens_seen": 2992504832 + }, + { + "epoch": 0.91, + "learning_rate": 7.552256770310933e-05, + "loss": 1.1988, + "theoretical_loss": 3.3498640073650305, + "tokens_seen": 2992570368 + }, + { + "epoch": 0.91, + "learning_rate": 7.550651955867604e-05, + "loss": 1.2642, + "theoretical_loss": 3.349858412111336, + "tokens_seen": 2992635904 + }, + { + "epoch": 0.91, + "learning_rate": 7.549047141424274e-05, + "loss": 1.2607, + "theoretical_loss": 3.349852817014479, + "tokens_seen": 2992701440 + }, + { + "epoch": 0.91, + "learning_rate": 7.547442326980944e-05, + "loss": 1.187, + "theoretical_loss": 3.3498472220744517, + "tokens_seen": 2992766976 + }, + { + "epoch": 0.91, + "learning_rate": 7.545837512537614e-05, + "loss": 1.0994, + "theoretical_loss": 3.349841627291246, + "tokens_seen": 2992832512 + }, + { + "epoch": 0.91, + "learning_rate": 7.544232698094284e-05, + "loss": 1.3215, + "theoretical_loss": 3.349836032664854, + "tokens_seen": 2992898048 + }, + { + "epoch": 0.91, + "learning_rate": 7.542627883650954e-05, + "loss": 1.1893, + "theoretical_loss": 3.3498304381952684, + "tokens_seen": 2992963584 + }, + { + "epoch": 0.91, + "learning_rate": 7.541023069207624e-05, + "loss": 1.2739, + "theoretical_loss": 3.349824843882481, + "tokens_seen": 2993029120 + }, + { + "epoch": 0.91, + "learning_rate": 7.539418254764294e-05, + "loss": 1.2604, + "theoretical_loss": 3.3498192497264845, + "tokens_seen": 2993094656 + }, + { + "epoch": 0.91, + "learning_rate": 7.537813440320963e-05, + "loss": 1.1352, + "theoretical_loss": 3.34981365572727, + "tokens_seen": 2993160192 + }, + { + "epoch": 0.91, + "learning_rate": 7.536208625877633e-05, + "loss": 1.2251, + "theoretical_loss": 3.3498080618848305, + "tokens_seen": 2993225728 + }, + { + "epoch": 0.91, + "learning_rate": 7.534603811434303e-05, + "loss": 1.2078, + "theoretical_loss": 3.349802468199158, + "tokens_seen": 2993291264 + }, + { + "epoch": 0.91, + "learning_rate": 7.532998996990973e-05, + "loss": 1.2731, + "theoretical_loss": 3.349796874670245, + "tokens_seen": 2993356800 + }, + { + "epoch": 0.91, + "learning_rate": 7.531394182547643e-05, + "loss": 1.1076, + "theoretical_loss": 3.349791281298083, + "tokens_seen": 2993422336 + }, + { + "epoch": 0.91, + "learning_rate": 7.529789368104313e-05, + "loss": 1.2594, + "theoretical_loss": 3.3497856880826644, + "tokens_seen": 2993487872 + }, + { + "epoch": 0.91, + "learning_rate": 7.528184553660983e-05, + "loss": 1.1731, + "theoretical_loss": 3.3497800950239816, + "tokens_seen": 2993553408 + }, + { + "epoch": 0.91, + "learning_rate": 7.526579739217653e-05, + "loss": 1.2123, + "theoretical_loss": 3.3497745021220267, + "tokens_seen": 2993618944 + }, + { + "epoch": 0.91, + "learning_rate": 7.524974924774323e-05, + "loss": 1.2292, + "theoretical_loss": 3.3497689093767917, + "tokens_seen": 2993684480 + }, + { + "epoch": 0.91, + "learning_rate": 7.523370110330993e-05, + "loss": 1.226, + "theoretical_loss": 3.3497633167882688, + "tokens_seen": 2993750016 + }, + { + "epoch": 0.91, + "learning_rate": 7.521765295887663e-05, + "loss": 1.3266, + "theoretical_loss": 3.3497577243564507, + "tokens_seen": 2993815552 + }, + { + "epoch": 0.91, + "learning_rate": 7.520160481444333e-05, + "loss": 1.2699, + "theoretical_loss": 3.3497521320813286, + "tokens_seen": 2993881088 + }, + { + "epoch": 0.91, + "learning_rate": 7.518555667001003e-05, + "loss": 1.2778, + "theoretical_loss": 3.3497465399628954, + "tokens_seen": 2993946624 + }, + { + "epoch": 0.91, + "learning_rate": 7.516950852557673e-05, + "loss": 1.1866, + "theoretical_loss": 3.3497409480011435, + "tokens_seen": 2994012160 + }, + { + "epoch": 0.91, + "learning_rate": 7.515346038114344e-05, + "loss": 1.1695, + "theoretical_loss": 3.3497353561960646, + "tokens_seen": 2994077696 + }, + { + "epoch": 0.91, + "learning_rate": 7.513741223671014e-05, + "loss": 1.3113, + "theoretical_loss": 3.3497297645476505, + "tokens_seen": 2994143232 + }, + { + "epoch": 0.91, + "learning_rate": 7.512136409227684e-05, + "loss": 1.266, + "theoretical_loss": 3.349724173055894, + "tokens_seen": 2994208768 + }, + { + "epoch": 0.91, + "learning_rate": 7.510531594784354e-05, + "loss": 1.2035, + "theoretical_loss": 3.3497185817207877, + "tokens_seen": 2994274304 + }, + { + "epoch": 0.91, + "learning_rate": 7.508926780341024e-05, + "loss": 1.2646, + "theoretical_loss": 3.349712990542323, + "tokens_seen": 2994339840 + }, + { + "epoch": 0.91, + "learning_rate": 7.507321965897694e-05, + "loss": 1.2661, + "theoretical_loss": 3.3497073995204922, + "tokens_seen": 2994405376 + }, + { + "epoch": 0.91, + "learning_rate": 7.505717151454364e-05, + "loss": 1.3037, + "theoretical_loss": 3.3497018086552877, + "tokens_seen": 2994470912 + }, + { + "epoch": 0.91, + "learning_rate": 7.504112337011034e-05, + "loss": 1.2332, + "theoretical_loss": 3.3496962179467014, + "tokens_seen": 2994536448 + }, + { + "epoch": 0.91, + "learning_rate": 7.502507522567703e-05, + "loss": 1.2532, + "theoretical_loss": 3.349690627394726, + "tokens_seen": 2994601984 + }, + { + "epoch": 0.91, + "learning_rate": 7.500902708124373e-05, + "loss": 1.348, + "theoretical_loss": 3.349685036999353, + "tokens_seen": 2994667520 + }, + { + "epoch": 0.91, + "learning_rate": 7.499297893681043e-05, + "loss": 1.1414, + "theoretical_loss": 3.349679446760575, + "tokens_seen": 2994733056 + }, + { + "epoch": 0.91, + "learning_rate": 7.497693079237713e-05, + "loss": 1.2685, + "theoretical_loss": 3.3496738566783844, + "tokens_seen": 2994798592 + }, + { + "epoch": 0.91, + "learning_rate": 7.496088264794383e-05, + "loss": 1.1519, + "theoretical_loss": 3.3496682667527726, + "tokens_seen": 2994864128 + }, + { + "epoch": 0.91, + "learning_rate": 7.494483450351053e-05, + "loss": 1.2535, + "theoretical_loss": 3.3496626769837325, + "tokens_seen": 2994929664 + }, + { + "epoch": 0.91, + "objective/train/docs_used": 1088743, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.2331516742706299, + "objective/train/theoretical_loss": 3.3496570873712566, + "objective/train/tokens_used": 3015455200, + "theoretical_loss": 3.3496570873712566, + "tokens_seen": 2994995200 + }, + { + "epoch": 0.91, + "learning_rate": 7.492878635907723e-05, + "loss": 1.2019, + "theoretical_loss": 3.3496570873712566, + "tokens_seen": 2994995200 + }, + { + "epoch": 0.91, + "learning_rate": 7.491273821464395e-05, + "loss": 1.2368, + "theoretical_loss": 3.3496514979153362, + "tokens_seen": 2995060736 + }, + { + "epoch": 0.91, + "learning_rate": 7.489669007021065e-05, + "loss": 1.2744, + "theoretical_loss": 3.3496459086159636, + "tokens_seen": 2995126272 + }, + { + "epoch": 0.91, + "learning_rate": 7.488064192577733e-05, + "loss": 1.2241, + "theoretical_loss": 3.3496403194731315, + "tokens_seen": 2995191808 + }, + { + "epoch": 0.91, + "learning_rate": 7.486459378134403e-05, + "loss": 1.3113, + "theoretical_loss": 3.349634730486832, + "tokens_seen": 2995257344 + }, + { + "epoch": 0.91, + "learning_rate": 7.484854563691073e-05, + "loss": 1.1958, + "theoretical_loss": 3.349629141657057, + "tokens_seen": 2995322880 + }, + { + "epoch": 0.91, + "learning_rate": 7.483249749247743e-05, + "loss": 1.1324, + "theoretical_loss": 3.349623552983799, + "tokens_seen": 2995388416 + }, + { + "epoch": 0.91, + "learning_rate": 7.481644934804414e-05, + "loss": 1.2021, + "theoretical_loss": 3.3496179644670496, + "tokens_seen": 2995453952 + }, + { + "epoch": 0.91, + "learning_rate": 7.480040120361084e-05, + "loss": 1.1859, + "theoretical_loss": 3.3496123761068017, + "tokens_seen": 2995519488 + }, + { + "epoch": 0.91, + "learning_rate": 7.478435305917754e-05, + "loss": 1.2147, + "theoretical_loss": 3.3496067879030473, + "tokens_seen": 2995585024 + }, + { + "epoch": 0.91, + "learning_rate": 7.476830491474424e-05, + "loss": 1.1702, + "theoretical_loss": 3.3496011998557784, + "tokens_seen": 2995650560 + }, + { + "epoch": 0.91, + "learning_rate": 7.475225677031094e-05, + "loss": 1.2547, + "theoretical_loss": 3.3495956119649875, + "tokens_seen": 2995716096 + }, + { + "epoch": 0.91, + "learning_rate": 7.473620862587764e-05, + "loss": 1.2686, + "theoretical_loss": 3.3495900242306664, + "tokens_seen": 2995781632 + }, + { + "epoch": 0.91, + "learning_rate": 7.472016048144434e-05, + "loss": 1.1579, + "theoretical_loss": 3.3495844366528074, + "tokens_seen": 2995847168 + }, + { + "epoch": 0.91, + "learning_rate": 7.470411233701104e-05, + "loss": 1.2165, + "theoretical_loss": 3.349578849231403, + "tokens_seen": 2995912704 + }, + { + "epoch": 0.91, + "learning_rate": 7.468806419257773e-05, + "loss": 1.3353, + "theoretical_loss": 3.3495732619664453, + "tokens_seen": 2995978240 + }, + { + "epoch": 0.91, + "learning_rate": 7.467201604814443e-05, + "loss": 1.2684, + "theoretical_loss": 3.349567674857926, + "tokens_seen": 2996043776 + }, + { + "epoch": 0.91, + "learning_rate": 7.465596790371113e-05, + "loss": 1.3325, + "theoretical_loss": 3.349562087905838, + "tokens_seen": 2996109312 + }, + { + "epoch": 0.91, + "learning_rate": 7.463991975927783e-05, + "loss": 1.2368, + "theoretical_loss": 3.3495565011101727, + "tokens_seen": 2996174848 + }, + { + "epoch": 0.91, + "learning_rate": 7.462387161484454e-05, + "loss": 1.3069, + "theoretical_loss": 3.3495509144709232, + "tokens_seen": 2996240384 + }, + { + "epoch": 0.91, + "learning_rate": 7.460782347041124e-05, + "loss": 1.2407, + "theoretical_loss": 3.349545327988081, + "tokens_seen": 2996305920 + }, + { + "epoch": 0.91, + "learning_rate": 7.459177532597794e-05, + "loss": 1.2595, + "theoretical_loss": 3.3495397416616393, + "tokens_seen": 2996371456 + }, + { + "epoch": 0.91, + "learning_rate": 7.457572718154464e-05, + "loss": 1.253, + "theoretical_loss": 3.3495341554915887, + "tokens_seen": 2996436992 + }, + { + "epoch": 0.91, + "learning_rate": 7.455967903711135e-05, + "loss": 1.1907, + "theoretical_loss": 3.349528569477923, + "tokens_seen": 2996502528 + }, + { + "epoch": 0.91, + "learning_rate": 7.454363089267805e-05, + "loss": 1.2856, + "theoretical_loss": 3.3495229836206333, + "tokens_seen": 2996568064 + }, + { + "epoch": 0.91, + "learning_rate": 7.452758274824473e-05, + "loss": 1.2797, + "theoretical_loss": 3.349517397919712, + "tokens_seen": 2996633600 + }, + { + "epoch": 0.91, + "learning_rate": 7.451153460381143e-05, + "loss": 1.2352, + "theoretical_loss": 3.3495118123751517, + "tokens_seen": 2996699136 + }, + { + "epoch": 0.91, + "learning_rate": 7.449548645937813e-05, + "loss": 1.2392, + "theoretical_loss": 3.3495062269869447, + "tokens_seen": 2996764672 + }, + { + "epoch": 0.91, + "learning_rate": 7.447943831494484e-05, + "loss": 1.181, + "theoretical_loss": 3.3495006417550823, + "tokens_seen": 2996830208 + }, + { + "epoch": 0.91, + "learning_rate": 7.446339017051154e-05, + "loss": 1.2595, + "theoretical_loss": 3.3494950566795576, + "tokens_seen": 2996895744 + }, + { + "epoch": 0.91, + "learning_rate": 7.444734202607824e-05, + "loss": 1.2601, + "theoretical_loss": 3.3494894717603625, + "tokens_seen": 2996961280 + }, + { + "epoch": 0.91, + "learning_rate": 7.443129388164494e-05, + "loss": 1.1759, + "theoretical_loss": 3.3494838869974894, + "tokens_seen": 2997026816 + }, + { + "epoch": 0.91, + "learning_rate": 7.441524573721164e-05, + "loss": 1.2938, + "theoretical_loss": 3.34947830239093, + "tokens_seen": 2997092352 + }, + { + "epoch": 0.91, + "learning_rate": 7.439919759277834e-05, + "loss": 1.1696, + "theoretical_loss": 3.349472717940677, + "tokens_seen": 2997157888 + }, + { + "epoch": 0.91, + "learning_rate": 7.438314944834504e-05, + "loss": 1.3291, + "theoretical_loss": 3.3494671336467223, + "tokens_seen": 2997223424 + }, + { + "epoch": 0.91, + "learning_rate": 7.436710130391174e-05, + "loss": 1.1988, + "theoretical_loss": 3.3494615495090585, + "tokens_seen": 2997288960 + }, + { + "epoch": 0.91, + "learning_rate": 7.435105315947844e-05, + "loss": 1.2356, + "theoretical_loss": 3.349455965527677, + "tokens_seen": 2997354496 + }, + { + "epoch": 0.91, + "learning_rate": 7.433500501504513e-05, + "loss": 1.2433, + "theoretical_loss": 3.3494503817025714, + "tokens_seen": 2997420032 + }, + { + "epoch": 0.91, + "learning_rate": 7.431895687061184e-05, + "loss": 1.264, + "theoretical_loss": 3.3494447980337325, + "tokens_seen": 2997485568 + }, + { + "epoch": 0.91, + "learning_rate": 7.430290872617854e-05, + "loss": 1.2975, + "theoretical_loss": 3.349439214521153, + "tokens_seen": 2997551104 + }, + { + "epoch": 0.91, + "learning_rate": 7.428686058174524e-05, + "loss": 1.2147, + "theoretical_loss": 3.3494336311648256, + "tokens_seen": 2997616640 + }, + { + "epoch": 0.91, + "learning_rate": 7.427081243731194e-05, + "loss": 1.1231, + "theoretical_loss": 3.349428047964742, + "tokens_seen": 2997682176 + }, + { + "epoch": 0.91, + "learning_rate": 7.425476429287864e-05, + "loss": 1.2541, + "theoretical_loss": 3.3494224649208943, + "tokens_seen": 2997747712 + }, + { + "epoch": 0.91, + "learning_rate": 7.423871614844534e-05, + "loss": 1.3012, + "theoretical_loss": 3.349416882033275, + "tokens_seen": 2997813248 + }, + { + "epoch": 0.91, + "learning_rate": 7.422266800401205e-05, + "loss": 1.2196, + "theoretical_loss": 3.3494112993018765, + "tokens_seen": 2997878784 + }, + { + "epoch": 0.91, + "learning_rate": 7.420661985957875e-05, + "loss": 1.1875, + "theoretical_loss": 3.3494057167266904, + "tokens_seen": 2997944320 + }, + { + "epoch": 0.91, + "learning_rate": 7.419057171514545e-05, + "loss": 1.2417, + "theoretical_loss": 3.3494001343077096, + "tokens_seen": 2998009856 + }, + { + "epoch": 0.91, + "learning_rate": 7.417452357071213e-05, + "loss": 1.2321, + "theoretical_loss": 3.3493945520449255, + "tokens_seen": 2998075392 + }, + { + "epoch": 0.91, + "learning_rate": 7.415847542627883e-05, + "loss": 1.2634, + "theoretical_loss": 3.3493889699383312, + "tokens_seen": 2998140928 + }, + { + "epoch": 0.91, + "learning_rate": 7.414242728184553e-05, + "loss": 1.1658, + "theoretical_loss": 3.3493833879879182, + "tokens_seen": 2998206464 + }, + { + "epoch": 0.91, + "objective/train/docs_used": 1089857, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.3672313690185547, + "objective/train/theoretical_loss": 3.3493778061936794, + "objective/train/tokens_used": 3018732000, + "theoretical_loss": 3.3493778061936794, + "tokens_seen": 2998272000 + }, + { + "epoch": 0.91, + "learning_rate": 7.412637913741224e-05, + "loss": 1.3233, + "theoretical_loss": 3.3493778061936794, + "tokens_seen": 2998272000 + }, + { + "epoch": 0.91, + "learning_rate": 7.411033099297894e-05, + "loss": 1.2652, + "theoretical_loss": 3.3493722245556063, + "tokens_seen": 2998337536 + }, + { + "epoch": 0.91, + "learning_rate": 7.409428284854564e-05, + "loss": 1.1995, + "theoretical_loss": 3.349366643073692, + "tokens_seen": 2998403072 + }, + { + "epoch": 0.91, + "learning_rate": 7.407823470411234e-05, + "loss": 1.1533, + "theoretical_loss": 3.3493610617479277, + "tokens_seen": 2998468608 + }, + { + "epoch": 0.91, + "learning_rate": 7.406218655967904e-05, + "loss": 1.2351, + "theoretical_loss": 3.349355480578306, + "tokens_seen": 2998534144 + }, + { + "epoch": 0.91, + "learning_rate": 7.404613841524574e-05, + "loss": 1.2455, + "theoretical_loss": 3.3493498995648197, + "tokens_seen": 2998599680 + }, + { + "epoch": 0.91, + "learning_rate": 7.403009027081244e-05, + "loss": 1.3114, + "theoretical_loss": 3.3493443187074603, + "tokens_seen": 2998665216 + }, + { + "epoch": 0.91, + "learning_rate": 7.401404212637914e-05, + "loss": 1.2146, + "theoretical_loss": 3.34933873800622, + "tokens_seen": 2998730752 + }, + { + "epoch": 0.91, + "learning_rate": 7.399799398194584e-05, + "loss": 1.3537, + "theoretical_loss": 3.349333157461092, + "tokens_seen": 2998796288 + }, + { + "epoch": 0.91, + "learning_rate": 7.398194583751254e-05, + "loss": 1.2597, + "theoretical_loss": 3.349327577072067, + "tokens_seen": 2998861824 + }, + { + "epoch": 0.91, + "learning_rate": 7.396589769307924e-05, + "loss": 1.3069, + "theoretical_loss": 3.349321996839138, + "tokens_seen": 2998927360 + }, + { + "epoch": 0.91, + "learning_rate": 7.394984954864594e-05, + "loss": 1.2126, + "theoretical_loss": 3.349316416762298, + "tokens_seen": 2998992896 + }, + { + "epoch": 0.91, + "learning_rate": 7.393380140421264e-05, + "loss": 1.1137, + "theoretical_loss": 3.3493108368415383, + "tokens_seen": 2999058432 + }, + { + "epoch": 0.91, + "learning_rate": 7.391775325977934e-05, + "loss": 1.2246, + "theoretical_loss": 3.3493052570768507, + "tokens_seen": 2999123968 + }, + { + "epoch": 0.91, + "learning_rate": 7.390170511534604e-05, + "loss": 1.3834, + "theoretical_loss": 3.3492996774682284, + "tokens_seen": 2999189504 + }, + { + "epoch": 0.91, + "learning_rate": 7.388565697091275e-05, + "loss": 1.2694, + "theoretical_loss": 3.3492940980156636, + "tokens_seen": 2999255040 + }, + { + "epoch": 0.91, + "learning_rate": 7.386960882647945e-05, + "loss": 1.1943, + "theoretical_loss": 3.3492885187191477, + "tokens_seen": 2999320576 + }, + { + "epoch": 0.91, + "learning_rate": 7.385356068204615e-05, + "loss": 1.2894, + "theoretical_loss": 3.3492829395786736, + "tokens_seen": 2999386112 + }, + { + "epoch": 0.91, + "learning_rate": 7.383751253761285e-05, + "loss": 1.1792, + "theoretical_loss": 3.349277360594233, + "tokens_seen": 2999451648 + }, + { + "epoch": 0.91, + "learning_rate": 7.382146439317953e-05, + "loss": 1.3144, + "theoretical_loss": 3.3492717817658186, + "tokens_seen": 2999517184 + }, + { + "epoch": 0.91, + "learning_rate": 7.380541624874623e-05, + "loss": 1.2996, + "theoretical_loss": 3.3492662030934226, + "tokens_seen": 2999582720 + }, + { + "epoch": 0.91, + "learning_rate": 7.378936810431294e-05, + "loss": 1.229, + "theoretical_loss": 3.3492606245770373, + "tokens_seen": 2999648256 + }, + { + "epoch": 0.91, + "learning_rate": 7.377331995987964e-05, + "loss": 1.1703, + "theoretical_loss": 3.3492550462166544, + "tokens_seen": 2999713792 + }, + { + "epoch": 0.91, + "learning_rate": 7.375727181544634e-05, + "loss": 1.3254, + "theoretical_loss": 3.3492494680122666, + "tokens_seen": 2999779328 + }, + { + "epoch": 0.91, + "learning_rate": 7.374122367101304e-05, + "loss": 1.3101, + "theoretical_loss": 3.3492438899638657, + "tokens_seen": 2999844864 + }, + { + "epoch": 0.91, + "learning_rate": 7.372517552657974e-05, + "loss": 1.1743, + "theoretical_loss": 3.349238312071445, + "tokens_seen": 2999910400 + }, + { + "epoch": 0.91, + "learning_rate": 7.370912738214645e-05, + "loss": 1.2148, + "theoretical_loss": 3.3492327343349952, + "tokens_seen": 2999975936 + }, + { + "epoch": 0.91, + "learning_rate": 7.369307923771315e-05, + "loss": 1.2634, + "theoretical_loss": 3.3492271567545093, + "tokens_seen": 3000041472 + }, + { + "epoch": 0.91, + "learning_rate": 7.367703109327985e-05, + "loss": 1.254, + "theoretical_loss": 3.34922157932998, + "tokens_seen": 3000107008 + }, + { + "epoch": 0.91, + "learning_rate": 7.366098294884654e-05, + "loss": 1.2636, + "theoretical_loss": 3.3492160020613984, + "tokens_seen": 3000172544 + }, + { + "epoch": 0.91, + "learning_rate": 7.364493480441324e-05, + "loss": 1.2236, + "theoretical_loss": 3.349210424948758, + "tokens_seen": 3000238080 + }, + { + "epoch": 0.91, + "learning_rate": 7.362888665997994e-05, + "loss": 1.2673, + "theoretical_loss": 3.34920484799205, + "tokens_seen": 3000303616 + }, + { + "epoch": 0.91, + "learning_rate": 7.361283851554664e-05, + "loss": 1.1992, + "theoretical_loss": 3.3491992711912673, + "tokens_seen": 3000369152 + }, + { + "epoch": 0.91, + "learning_rate": 7.359679037111334e-05, + "loss": 1.1993, + "theoretical_loss": 3.3491936945464023, + "tokens_seen": 3000434688 + }, + { + "epoch": 0.91, + "learning_rate": 7.358074222668004e-05, + "loss": 1.215, + "theoretical_loss": 3.349188118057446, + "tokens_seen": 3000500224 + }, + { + "epoch": 0.91, + "learning_rate": 7.356469408224674e-05, + "loss": 1.2058, + "theoretical_loss": 3.349182541724392, + "tokens_seen": 3000565760 + }, + { + "epoch": 0.91, + "learning_rate": 7.354864593781344e-05, + "loss": 1.1874, + "theoretical_loss": 3.3491769655472314, + "tokens_seen": 3000631296 + }, + { + "epoch": 0.91, + "learning_rate": 7.353259779338015e-05, + "loss": 1.287, + "theoretical_loss": 3.3491713895259574, + "tokens_seen": 3000696832 + }, + { + "epoch": 0.91, + "learning_rate": 7.351654964894685e-05, + "loss": 1.2269, + "theoretical_loss": 3.349165813660562, + "tokens_seen": 3000762368 + }, + { + "epoch": 0.91, + "learning_rate": 7.350050150451355e-05, + "loss": 1.2518, + "theoretical_loss": 3.3491602379510375, + "tokens_seen": 3000827904 + }, + { + "epoch": 0.91, + "learning_rate": 7.348445336008025e-05, + "loss": 1.2508, + "theoretical_loss": 3.3491546623973756, + "tokens_seen": 3000893440 + }, + { + "epoch": 0.91, + "learning_rate": 7.346840521564693e-05, + "loss": 1.3281, + "theoretical_loss": 3.349149086999569, + "tokens_seen": 3000958976 + }, + { + "epoch": 0.91, + "learning_rate": 7.345235707121364e-05, + "loss": 1.2823, + "theoretical_loss": 3.3491435117576094, + "tokens_seen": 3001024512 + }, + { + "epoch": 0.91, + "learning_rate": 7.343630892678034e-05, + "loss": 1.2471, + "theoretical_loss": 3.34913793667149, + "tokens_seen": 3001090048 + }, + { + "epoch": 0.91, + "learning_rate": 7.342026078234705e-05, + "loss": 1.191, + "theoretical_loss": 3.3491323617412023, + "tokens_seen": 3001155584 + }, + { + "epoch": 0.91, + "learning_rate": 7.340421263791375e-05, + "loss": 1.1478, + "theoretical_loss": 3.349126786966739, + "tokens_seen": 3001221120 + }, + { + "epoch": 0.91, + "learning_rate": 7.338816449348045e-05, + "loss": 1.3031, + "theoretical_loss": 3.349121212348092, + "tokens_seen": 3001286656 + }, + { + "epoch": 0.91, + "learning_rate": 7.337211634904715e-05, + "loss": 1.3436, + "theoretical_loss": 3.3491156378852533, + "tokens_seen": 3001352192 + }, + { + "epoch": 0.91, + "learning_rate": 7.335606820461385e-05, + "loss": 1.2283, + "theoretical_loss": 3.3491100635782156, + "tokens_seen": 3001417728 + }, + { + "epoch": 0.91, + "learning_rate": 7.334002006018055e-05, + "loss": 1.2473, + "theoretical_loss": 3.349104489426971, + "tokens_seen": 3001483264 + }, + { + "epoch": 0.91, + "objective/train/docs_used": 1091037, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.1632734537124634, + "objective/train/theoretical_loss": 3.349098915431512, + "objective/train/tokens_used": 3022008800, + "theoretical_loss": 3.349098915431512, + "tokens_seen": 3001548800 + }, + { + "epoch": 0.91, + "learning_rate": 7.332397191574725e-05, + "loss": 1.184, + "theoretical_loss": 3.349098915431512, + "tokens_seen": 3001548800 + }, + { + "epoch": 0.91, + "learning_rate": 7.330792377131394e-05, + "loss": 1.3047, + "theoretical_loss": 3.3490933415918303, + "tokens_seen": 3001614336 + }, + { + "epoch": 0.91, + "learning_rate": 7.329187562688064e-05, + "loss": 1.2944, + "theoretical_loss": 3.349087767907919, + "tokens_seen": 3001679872 + }, + { + "epoch": 0.91, + "learning_rate": 7.327582748244734e-05, + "loss": 1.2589, + "theoretical_loss": 3.3490821943797693, + "tokens_seen": 3001745408 + }, + { + "epoch": 0.91, + "learning_rate": 7.325977933801404e-05, + "loss": 1.2869, + "theoretical_loss": 3.349076621007374, + "tokens_seen": 3001810944 + }, + { + "epoch": 0.91, + "learning_rate": 7.324373119358074e-05, + "loss": 1.1896, + "theoretical_loss": 3.3490710477907255, + "tokens_seen": 3001876480 + }, + { + "epoch": 0.91, + "learning_rate": 7.322768304914744e-05, + "loss": 1.2045, + "theoretical_loss": 3.3490654747298154, + "tokens_seen": 3001942016 + }, + { + "epoch": 0.91, + "learning_rate": 7.321163490471414e-05, + "loss": 1.2381, + "theoretical_loss": 3.3490599018246368, + "tokens_seen": 3002007552 + }, + { + "epoch": 0.91, + "learning_rate": 7.319558676028085e-05, + "loss": 1.3123, + "theoretical_loss": 3.3490543290751815, + "tokens_seen": 3002073088 + }, + { + "epoch": 0.91, + "learning_rate": 7.317953861584755e-05, + "loss": 1.2709, + "theoretical_loss": 3.3490487564814417, + "tokens_seen": 3002138624 + }, + { + "epoch": 0.91, + "learning_rate": 7.316349047141425e-05, + "loss": 1.342, + "theoretical_loss": 3.34904318404341, + "tokens_seen": 3002204160 + }, + { + "epoch": 0.91, + "learning_rate": 7.314744232698095e-05, + "loss": 1.2459, + "theoretical_loss": 3.3490376117610783, + "tokens_seen": 3002269696 + }, + { + "epoch": 0.91, + "learning_rate": 7.313139418254765e-05, + "loss": 1.131, + "theoretical_loss": 3.3490320396344386, + "tokens_seen": 3002335232 + }, + { + "epoch": 0.91, + "learning_rate": 7.311534603811435e-05, + "loss": 1.2198, + "theoretical_loss": 3.3490264676634833, + "tokens_seen": 3002400768 + }, + { + "epoch": 0.91, + "learning_rate": 7.309929789368105e-05, + "loss": 1.2418, + "theoretical_loss": 3.3490208958482057, + "tokens_seen": 3002466304 + }, + { + "epoch": 0.91, + "learning_rate": 7.308324974924775e-05, + "loss": 1.3456, + "theoretical_loss": 3.3490153241885965, + "tokens_seen": 3002531840 + }, + { + "epoch": 0.91, + "learning_rate": 7.306720160481445e-05, + "loss": 1.2233, + "theoretical_loss": 3.349009752684649, + "tokens_seen": 3002597376 + }, + { + "epoch": 0.91, + "learning_rate": 7.305115346038115e-05, + "loss": 1.1381, + "theoretical_loss": 3.3490041813363547, + "tokens_seen": 3002662912 + }, + { + "epoch": 0.91, + "learning_rate": 7.303510531594785e-05, + "loss": 1.2001, + "theoretical_loss": 3.3489986101437066, + "tokens_seen": 3002728448 + }, + { + "epoch": 0.91, + "learning_rate": 7.301905717151455e-05, + "loss": 1.3012, + "theoretical_loss": 3.3489930391066967, + "tokens_seen": 3002793984 + }, + { + "epoch": 0.91, + "learning_rate": 7.300300902708125e-05, + "loss": 1.3064, + "theoretical_loss": 3.348987468225317, + "tokens_seen": 3002859520 + }, + { + "epoch": 0.91, + "learning_rate": 7.298696088264795e-05, + "loss": 1.2672, + "theoretical_loss": 3.34898189749956, + "tokens_seen": 3002925056 + }, + { + "epoch": 0.91, + "learning_rate": 7.297091273821465e-05, + "loss": 1.2589, + "theoretical_loss": 3.3489763269294177, + "tokens_seen": 3002990592 + }, + { + "epoch": 0.91, + "learning_rate": 7.295486459378134e-05, + "loss": 1.2908, + "theoretical_loss": 3.3489707565148827, + "tokens_seen": 3003056128 + }, + { + "epoch": 0.91, + "learning_rate": 7.293881644934804e-05, + "loss": 1.2607, + "theoretical_loss": 3.348965186255947, + "tokens_seen": 3003121664 + }, + { + "epoch": 0.91, + "learning_rate": 7.292276830491474e-05, + "loss": 1.2278, + "theoretical_loss": 3.348959616152603, + "tokens_seen": 3003187200 + }, + { + "epoch": 0.91, + "learning_rate": 7.290672016048144e-05, + "loss": 1.2672, + "theoretical_loss": 3.348954046204843, + "tokens_seen": 3003252736 + }, + { + "epoch": 0.91, + "learning_rate": 7.289067201604814e-05, + "loss": 1.1934, + "theoretical_loss": 3.3489484764126587, + "tokens_seen": 3003318272 + }, + { + "epoch": 0.91, + "learning_rate": 7.287462387161484e-05, + "loss": 1.2613, + "theoretical_loss": 3.3489429067760437, + "tokens_seen": 3003383808 + }, + { + "epoch": 0.91, + "learning_rate": 7.285857572718154e-05, + "loss": 1.2349, + "theoretical_loss": 3.3489373372949887, + "tokens_seen": 3003449344 + }, + { + "epoch": 0.91, + "learning_rate": 7.284252758274825e-05, + "loss": 1.1904, + "theoretical_loss": 3.3489317679694866, + "tokens_seen": 3003514880 + }, + { + "epoch": 0.91, + "learning_rate": 7.282647943831495e-05, + "loss": 1.2689, + "theoretical_loss": 3.3489261987995302, + "tokens_seen": 3003580416 + }, + { + "epoch": 0.91, + "learning_rate": 7.281043129388165e-05, + "loss": 1.2208, + "theoretical_loss": 3.348920629785111, + "tokens_seen": 3003645952 + }, + { + "epoch": 0.91, + "learning_rate": 7.279438314944835e-05, + "loss": 1.1826, + "theoretical_loss": 3.3489150609262217, + "tokens_seen": 3003711488 + }, + { + "epoch": 0.91, + "learning_rate": 7.277833500501505e-05, + "loss": 1.2493, + "theoretical_loss": 3.348909492222854, + "tokens_seen": 3003777024 + }, + { + "epoch": 0.91, + "learning_rate": 7.276228686058175e-05, + "loss": 1.2563, + "theoretical_loss": 3.348903923675001, + "tokens_seen": 3003842560 + }, + { + "epoch": 0.91, + "learning_rate": 7.274623871614845e-05, + "loss": 1.2569, + "theoretical_loss": 3.348898355282654, + "tokens_seen": 3003908096 + }, + { + "epoch": 0.91, + "learning_rate": 7.273019057171515e-05, + "loss": 1.3457, + "theoretical_loss": 3.3488927870458065, + "tokens_seen": 3003973632 + }, + { + "epoch": 0.91, + "learning_rate": 7.271414242728185e-05, + "loss": 1.1275, + "theoretical_loss": 3.3488872189644496, + "tokens_seen": 3004039168 + }, + { + "epoch": 0.91, + "learning_rate": 7.269809428284855e-05, + "loss": 1.2885, + "theoretical_loss": 3.3488816510385764, + "tokens_seen": 3004104704 + }, + { + "epoch": 0.91, + "learning_rate": 7.268204613841525e-05, + "loss": 1.0936, + "theoretical_loss": 3.3488760832681783, + "tokens_seen": 3004170240 + }, + { + "epoch": 0.91, + "learning_rate": 7.266599799398195e-05, + "loss": 1.254, + "theoretical_loss": 3.3488705156532483, + "tokens_seen": 3004235776 + }, + { + "epoch": 0.91, + "learning_rate": 7.264994984954865e-05, + "loss": 1.2868, + "theoretical_loss": 3.3488649481937784, + "tokens_seen": 3004301312 + }, + { + "epoch": 0.91, + "learning_rate": 7.263390170511535e-05, + "loss": 1.1698, + "theoretical_loss": 3.348859380889761, + "tokens_seen": 3004366848 + }, + { + "epoch": 0.91, + "learning_rate": 7.261785356068204e-05, + "loss": 1.3666, + "theoretical_loss": 3.3488538137411883, + "tokens_seen": 3004432384 + }, + { + "epoch": 0.91, + "learning_rate": 7.260180541624874e-05, + "loss": 1.2434, + "theoretical_loss": 3.3488482467480525, + "tokens_seen": 3004497920 + }, + { + "epoch": 0.91, + "learning_rate": 7.258575727181544e-05, + "loss": 1.1873, + "theoretical_loss": 3.3488426799103457, + "tokens_seen": 3004563456 + }, + { + "epoch": 0.91, + "learning_rate": 7.256970912738214e-05, + "loss": 1.3348, + "theoretical_loss": 3.3488371132280603, + "tokens_seen": 3004628992 + }, + { + "epoch": 0.91, + "learning_rate": 7.255366098294884e-05, + "loss": 1.2347, + "theoretical_loss": 3.348831546701189, + "tokens_seen": 3004694528 + }, + { + "epoch": 0.91, + "learning_rate": 7.253761283851554e-05, + "loss": 1.1955, + "theoretical_loss": 3.3488259803297233, + "tokens_seen": 3004760064 + }, + { + "epoch": 0.91, + "objective/train/docs_used": 1092332, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.353142261505127, + "objective/train/theoretical_loss": 3.348820414113656, + "objective/train/tokens_used": 3025285600, + "theoretical_loss": 3.348820414113656, + "tokens_seen": 3004825600 + }, + { + "epoch": 0.91, + "learning_rate": 7.252156469408226e-05, + "loss": 1.2866, + "theoretical_loss": 3.348820414113656, + "tokens_seen": 3004825600 + }, + { + "epoch": 0.91, + "learning_rate": 7.250551654964896e-05, + "loss": 1.2181, + "theoretical_loss": 3.34881484805298, + "tokens_seen": 3004891136 + }, + { + "epoch": 0.91, + "learning_rate": 7.248946840521566e-05, + "loss": 1.2473, + "theoretical_loss": 3.348809282147686, + "tokens_seen": 3004956672 + }, + { + "epoch": 0.91, + "learning_rate": 7.247342026078236e-05, + "loss": 1.1949, + "theoretical_loss": 3.348803716397767, + "tokens_seen": 3005022208 + }, + { + "epoch": 0.91, + "learning_rate": 7.245737211634905e-05, + "loss": 1.1628, + "theoretical_loss": 3.3487981508032156, + "tokens_seen": 3005087744 + }, + { + "epoch": 0.91, + "learning_rate": 7.244132397191575e-05, + "loss": 1.2851, + "theoretical_loss": 3.3487925853640244, + "tokens_seen": 3005153280 + }, + { + "epoch": 0.91, + "learning_rate": 7.242527582748245e-05, + "loss": 1.2647, + "theoretical_loss": 3.3487870200801844, + "tokens_seen": 3005218816 + }, + { + "epoch": 0.91, + "learning_rate": 7.240922768304915e-05, + "loss": 1.1322, + "theoretical_loss": 3.348781454951689, + "tokens_seen": 3005284352 + }, + { + "epoch": 0.91, + "learning_rate": 7.239317953861585e-05, + "loss": 1.2226, + "theoretical_loss": 3.34877588997853, + "tokens_seen": 3005349888 + }, + { + "epoch": 0.91, + "learning_rate": 7.237713139418255e-05, + "loss": 1.2243, + "theoretical_loss": 3.3487703251606993, + "tokens_seen": 3005415424 + }, + { + "epoch": 0.91, + "learning_rate": 7.236108324974925e-05, + "loss": 1.2194, + "theoretical_loss": 3.3487647604981903, + "tokens_seen": 3005480960 + }, + { + "epoch": 0.91, + "learning_rate": 7.234503510531595e-05, + "loss": 1.1961, + "theoretical_loss": 3.3487591959909944, + "tokens_seen": 3005546496 + }, + { + "epoch": 0.91, + "learning_rate": 7.232898696088265e-05, + "loss": 1.2382, + "theoretical_loss": 3.348753631639104, + "tokens_seen": 3005612032 + }, + { + "epoch": 0.91, + "learning_rate": 7.231293881644935e-05, + "loss": 1.2486, + "theoretical_loss": 3.3487480674425116, + "tokens_seen": 3005677568 + }, + { + "epoch": 0.91, + "learning_rate": 7.229689067201605e-05, + "loss": 1.166, + "theoretical_loss": 3.3487425034012093, + "tokens_seen": 3005743104 + }, + { + "epoch": 0.91, + "learning_rate": 7.228084252758275e-05, + "loss": 1.2063, + "theoretical_loss": 3.3487369395151894, + "tokens_seen": 3005808640 + }, + { + "epoch": 0.91, + "learning_rate": 7.226479438314944e-05, + "loss": 1.2805, + "theoretical_loss": 3.3487313757844444, + "tokens_seen": 3005874176 + }, + { + "epoch": 0.91, + "learning_rate": 7.224874623871614e-05, + "loss": 1.237, + "theoretical_loss": 3.348725812208966, + "tokens_seen": 3005939712 + }, + { + "epoch": 0.91, + "learning_rate": 7.223269809428284e-05, + "loss": 1.2296, + "theoretical_loss": 3.348720248788747, + "tokens_seen": 3006005248 + }, + { + "epoch": 0.91, + "learning_rate": 7.221664994984956e-05, + "loss": 1.1437, + "theoretical_loss": 3.3487146855237797, + "tokens_seen": 3006070784 + }, + { + "epoch": 0.91, + "learning_rate": 7.220060180541626e-05, + "loss": 1.3761, + "theoretical_loss": 3.348709122414056, + "tokens_seen": 3006136320 + }, + { + "epoch": 0.91, + "learning_rate": 7.218455366098296e-05, + "loss": 1.3239, + "theoretical_loss": 3.348703559459569, + "tokens_seen": 3006201856 + }, + { + "epoch": 0.91, + "learning_rate": 7.216850551654966e-05, + "loss": 1.287, + "theoretical_loss": 3.34869799666031, + "tokens_seen": 3006267392 + }, + { + "epoch": 0.91, + "learning_rate": 7.215245737211636e-05, + "loss": 1.2203, + "theoretical_loss": 3.3486924340162716, + "tokens_seen": 3006332928 + }, + { + "epoch": 0.91, + "learning_rate": 7.213640922768306e-05, + "loss": 1.2023, + "theoretical_loss": 3.3486868715274465, + "tokens_seen": 3006398464 + }, + { + "epoch": 0.91, + "learning_rate": 7.212036108324976e-05, + "loss": 1.307, + "theoretical_loss": 3.3486813091938266, + "tokens_seen": 3006464000 + }, + { + "epoch": 0.91, + "learning_rate": 7.210431293881645e-05, + "loss": 1.2028, + "theoretical_loss": 3.3486757470154043, + "tokens_seen": 3006529536 + }, + { + "epoch": 0.91, + "learning_rate": 7.208826479438315e-05, + "loss": 1.2336, + "theoretical_loss": 3.348670184992171, + "tokens_seen": 3006595072 + }, + { + "epoch": 0.91, + "learning_rate": 7.207221664994985e-05, + "loss": 1.3434, + "theoretical_loss": 3.348664623124121, + "tokens_seen": 3006660608 + }, + { + "epoch": 0.91, + "learning_rate": 7.205616850551655e-05, + "loss": 1.2982, + "theoretical_loss": 3.3486590614112446, + "tokens_seen": 3006726144 + }, + { + "epoch": 0.91, + "learning_rate": 7.204012036108325e-05, + "loss": 1.3113, + "theoretical_loss": 3.3486534998535356, + "tokens_seen": 3006791680 + }, + { + "epoch": 0.91, + "learning_rate": 7.202407221664995e-05, + "loss": 1.2812, + "theoretical_loss": 3.348647938450985, + "tokens_seen": 3006857216 + }, + { + "epoch": 0.91, + "learning_rate": 7.200802407221665e-05, + "loss": 1.3411, + "theoretical_loss": 3.348642377203586, + "tokens_seen": 3006922752 + }, + { + "epoch": 0.91, + "learning_rate": 7.199197592778335e-05, + "loss": 1.2343, + "theoretical_loss": 3.3486368161113305, + "tokens_seen": 3006988288 + }, + { + "epoch": 0.91, + "learning_rate": 7.197592778335005e-05, + "loss": 1.2589, + "theoretical_loss": 3.348631255174211, + "tokens_seen": 3007053824 + }, + { + "epoch": 0.91, + "learning_rate": 7.195987963891675e-05, + "loss": 1.27, + "theoretical_loss": 3.3486256943922195, + "tokens_seen": 3007119360 + }, + { + "epoch": 0.91, + "learning_rate": 7.194383149448345e-05, + "loss": 1.2162, + "theoretical_loss": 3.348620133765348, + "tokens_seen": 3007184896 + }, + { + "epoch": 0.91, + "learning_rate": 7.192778335005015e-05, + "loss": 1.1761, + "theoretical_loss": 3.34861457329359, + "tokens_seen": 3007250432 + }, + { + "epoch": 0.91, + "learning_rate": 7.191173520561686e-05, + "loss": 1.2861, + "theoretical_loss": 3.3486090129769366, + "tokens_seen": 3007315968 + }, + { + "epoch": 0.91, + "learning_rate": 7.189568706118356e-05, + "loss": 1.2452, + "theoretical_loss": 3.3486034528153805, + "tokens_seen": 3007381504 + }, + { + "epoch": 0.91, + "learning_rate": 7.187963891675026e-05, + "loss": 1.3726, + "theoretical_loss": 3.348597892808914, + "tokens_seen": 3007447040 + }, + { + "epoch": 0.91, + "learning_rate": 7.186359077231696e-05, + "loss": 1.3179, + "theoretical_loss": 3.34859233295753, + "tokens_seen": 3007512576 + }, + { + "epoch": 0.91, + "learning_rate": 7.184754262788366e-05, + "loss": 1.194, + "theoretical_loss": 3.3485867732612196, + "tokens_seen": 3007578112 + }, + { + "epoch": 0.91, + "learning_rate": 7.183149448345036e-05, + "loss": 1.141, + "theoretical_loss": 3.3485812137199757, + "tokens_seen": 3007643648 + }, + { + "epoch": 0.91, + "learning_rate": 7.181544633901706e-05, + "loss": 1.302, + "theoretical_loss": 3.348575654333791, + "tokens_seen": 3007709184 + }, + { + "epoch": 0.91, + "learning_rate": 7.179939819458376e-05, + "loss": 1.1967, + "theoretical_loss": 3.348570095102657, + "tokens_seen": 3007774720 + }, + { + "epoch": 0.91, + "learning_rate": 7.178335005015046e-05, + "loss": 1.2498, + "theoretical_loss": 3.348564536026567, + "tokens_seen": 3007840256 + }, + { + "epoch": 0.91, + "learning_rate": 7.176730190571716e-05, + "loss": 1.2801, + "theoretical_loss": 3.3485589771055118, + "tokens_seen": 3007905792 + }, + { + "epoch": 0.91, + "learning_rate": 7.175125376128385e-05, + "loss": 1.1126, + "theoretical_loss": 3.348553418339485, + "tokens_seen": 3007971328 + }, + { + "epoch": 0.91, + "learning_rate": 7.173520561685055e-05, + "loss": 1.2581, + "theoretical_loss": 3.3485478597284786, + "tokens_seen": 3008036864 + }, + { + "epoch": 0.91, + "objective/train/docs_used": 1093533, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.109999179840088, + "objective/train/theoretical_loss": 3.348542301272485, + "objective/train/tokens_used": 3028562400, + "theoretical_loss": 3.348542301272485, + "tokens_seen": 3008102400 + }, + { + "epoch": 0.91, + "learning_rate": 7.171915747241725e-05, + "loss": 1.1624, + "theoretical_loss": 3.348542301272485, + "tokens_seen": 3008102400 + }, + { + "epoch": 0.91, + "learning_rate": 7.170310932798395e-05, + "loss": 1.1765, + "theoretical_loss": 3.348536742971496, + "tokens_seen": 3008167936 + }, + { + "epoch": 0.91, + "learning_rate": 7.168706118355065e-05, + "loss": 1.3037, + "theoretical_loss": 3.348531184825504, + "tokens_seen": 3008233472 + }, + { + "epoch": 0.91, + "learning_rate": 7.167101303911735e-05, + "loss": 1.298, + "theoretical_loss": 3.348525626834502, + "tokens_seen": 3008299008 + }, + { + "epoch": 0.91, + "learning_rate": 7.165496489468405e-05, + "loss": 1.1823, + "theoretical_loss": 3.3485200689984813, + "tokens_seen": 3008364544 + }, + { + "epoch": 0.91, + "learning_rate": 7.163891675025075e-05, + "loss": 1.2602, + "theoretical_loss": 3.348514511317435, + "tokens_seen": 3008430080 + }, + { + "epoch": 0.91, + "learning_rate": 7.162286860581745e-05, + "loss": 1.2416, + "theoretical_loss": 3.348508953791355, + "tokens_seen": 3008495616 + }, + { + "epoch": 0.91, + "learning_rate": 7.160682046138417e-05, + "loss": 1.0895, + "theoretical_loss": 3.3485033964202335, + "tokens_seen": 3008561152 + }, + { + "epoch": 0.91, + "learning_rate": 7.159077231695085e-05, + "loss": 1.2362, + "theoretical_loss": 3.348497839204063, + "tokens_seen": 3008626688 + }, + { + "epoch": 0.91, + "learning_rate": 7.157472417251756e-05, + "loss": 1.3142, + "theoretical_loss": 3.3484922821428365, + "tokens_seen": 3008692224 + }, + { + "epoch": 0.91, + "learning_rate": 7.155867602808426e-05, + "loss": 1.3868, + "theoretical_loss": 3.348486725236545, + "tokens_seen": 3008757760 + }, + { + "epoch": 0.91, + "learning_rate": 7.154262788365096e-05, + "loss": 1.2416, + "theoretical_loss": 3.3484811684851814, + "tokens_seen": 3008823296 + }, + { + "epoch": 0.91, + "learning_rate": 7.152657973921766e-05, + "loss": 1.2445, + "theoretical_loss": 3.3484756118887384, + "tokens_seen": 3008888832 + }, + { + "epoch": 0.91, + "learning_rate": 7.151053159478436e-05, + "loss": 1.3171, + "theoretical_loss": 3.3484700554472075, + "tokens_seen": 3008954368 + }, + { + "epoch": 0.91, + "learning_rate": 7.149448345035106e-05, + "loss": 1.2157, + "theoretical_loss": 3.348464499160582, + "tokens_seen": 3009019904 + }, + { + "epoch": 0.91, + "learning_rate": 7.147843530591776e-05, + "loss": 1.2843, + "theoretical_loss": 3.348458943028853, + "tokens_seen": 3009085440 + }, + { + "epoch": 0.91, + "learning_rate": 7.146238716148446e-05, + "loss": 1.1829, + "theoretical_loss": 3.3484533870520137, + "tokens_seen": 3009150976 + }, + { + "epoch": 0.91, + "learning_rate": 7.144633901705116e-05, + "loss": 1.3504, + "theoretical_loss": 3.3484478312300565, + "tokens_seen": 3009216512 + }, + { + "epoch": 0.91, + "learning_rate": 7.143029087261786e-05, + "loss": 1.2515, + "theoretical_loss": 3.348442275562973, + "tokens_seen": 3009282048 + }, + { + "epoch": 0.91, + "learning_rate": 7.141424272818456e-05, + "loss": 1.3267, + "theoretical_loss": 3.3484367200507563, + "tokens_seen": 3009347584 + }, + { + "epoch": 0.91, + "learning_rate": 7.139819458375125e-05, + "loss": 1.2418, + "theoretical_loss": 3.3484311646933977, + "tokens_seen": 3009413120 + }, + { + "epoch": 0.91, + "learning_rate": 7.138214643931795e-05, + "loss": 1.3107, + "theoretical_loss": 3.3484256094908904, + "tokens_seen": 3009478656 + }, + { + "epoch": 0.91, + "learning_rate": 7.136609829488465e-05, + "loss": 1.209, + "theoretical_loss": 3.3484200544432263, + "tokens_seen": 3009544192 + }, + { + "epoch": 0.91, + "learning_rate": 7.135005015045135e-05, + "loss": 1.1295, + "theoretical_loss": 3.3484144995503984, + "tokens_seen": 3009609728 + }, + { + "epoch": 0.91, + "learning_rate": 7.133400200601805e-05, + "loss": 1.2487, + "theoretical_loss": 3.3484089448123977, + "tokens_seen": 3009675264 + }, + { + "epoch": 0.91, + "learning_rate": 7.131795386158477e-05, + "loss": 1.2749, + "theoretical_loss": 3.3484033902292176, + "tokens_seen": 3009740800 + }, + { + "epoch": 0.91, + "learning_rate": 7.130190571715147e-05, + "loss": 1.3128, + "theoretical_loss": 3.34839783580085, + "tokens_seen": 3009806336 + }, + { + "epoch": 0.91, + "learning_rate": 7.128585757271817e-05, + "loss": 1.1952, + "theoretical_loss": 3.3483922815272873, + "tokens_seen": 3009871872 + }, + { + "epoch": 0.91, + "learning_rate": 7.126980942828487e-05, + "loss": 1.2829, + "theoretical_loss": 3.348386727408522, + "tokens_seen": 3009937408 + }, + { + "epoch": 0.91, + "learning_rate": 7.125376128385157e-05, + "loss": 1.1267, + "theoretical_loss": 3.3483811734445457, + "tokens_seen": 3010002944 + }, + { + "epoch": 0.91, + "learning_rate": 7.123771313941825e-05, + "loss": 1.2748, + "theoretical_loss": 3.348375619635352, + "tokens_seen": 3010068480 + }, + { + "epoch": 0.91, + "learning_rate": 7.122166499498496e-05, + "loss": 1.2612, + "theoretical_loss": 3.348370065980932, + "tokens_seen": 3010134016 + }, + { + "epoch": 0.91, + "learning_rate": 7.120561685055166e-05, + "loss": 1.2137, + "theoretical_loss": 3.3483645124812784, + "tokens_seen": 3010199552 + }, + { + "epoch": 0.91, + "learning_rate": 7.118956870611836e-05, + "loss": 1.3108, + "theoretical_loss": 3.348358959136384, + "tokens_seen": 3010265088 + }, + { + "epoch": 0.91, + "learning_rate": 7.117352056168506e-05, + "loss": 1.2365, + "theoretical_loss": 3.34835340594624, + "tokens_seen": 3010330624 + }, + { + "epoch": 0.91, + "learning_rate": 7.115747241725176e-05, + "loss": 1.3457, + "theoretical_loss": 3.3483478529108397, + "tokens_seen": 3010396160 + }, + { + "epoch": 0.91, + "learning_rate": 7.114142427281846e-05, + "loss": 1.226, + "theoretical_loss": 3.3483423000301755, + "tokens_seen": 3010461696 + }, + { + "epoch": 0.91, + "learning_rate": 7.112537612838516e-05, + "loss": 1.2558, + "theoretical_loss": 3.348336747304239, + "tokens_seen": 3010527232 + }, + { + "epoch": 0.91, + "learning_rate": 7.110932798395186e-05, + "loss": 1.1902, + "theoretical_loss": 3.348331194733023, + "tokens_seen": 3010592768 + }, + { + "epoch": 0.91, + "learning_rate": 7.109327983951856e-05, + "loss": 1.1982, + "theoretical_loss": 3.3483256423165195, + "tokens_seen": 3010658304 + }, + { + "epoch": 0.91, + "learning_rate": 7.107723169508526e-05, + "loss": 1.2676, + "theoretical_loss": 3.348320090054721, + "tokens_seen": 3010723840 + }, + { + "epoch": 0.91, + "learning_rate": 7.106118355065196e-05, + "loss": 1.24, + "theoretical_loss": 3.3483145379476205, + "tokens_seen": 3010789376 + }, + { + "epoch": 0.91, + "learning_rate": 7.104513540621865e-05, + "loss": 1.2254, + "theoretical_loss": 3.3483089859952093, + "tokens_seen": 3010854912 + }, + { + "epoch": 0.91, + "learning_rate": 7.102908726178535e-05, + "loss": 1.249, + "theoretical_loss": 3.3483034341974798, + "tokens_seen": 3010920448 + }, + { + "epoch": 0.91, + "learning_rate": 7.101303911735206e-05, + "loss": 1.2055, + "theoretical_loss": 3.3482978825544247, + "tokens_seen": 3010985984 + }, + { + "epoch": 0.91, + "learning_rate": 7.099699097291876e-05, + "loss": 1.2653, + "theoretical_loss": 3.3482923310660366, + "tokens_seen": 3011051520 + }, + { + "epoch": 0.91, + "learning_rate": 7.098094282848547e-05, + "loss": 1.1087, + "theoretical_loss": 3.348286779732307, + "tokens_seen": 3011117056 + }, + { + "epoch": 0.91, + "learning_rate": 7.096489468405217e-05, + "loss": 1.2635, + "theoretical_loss": 3.3482812285532293, + "tokens_seen": 3011182592 + }, + { + "epoch": 0.91, + "learning_rate": 7.094884653961887e-05, + "loss": 1.2897, + "theoretical_loss": 3.348275677528795, + "tokens_seen": 3011248128 + }, + { + "epoch": 0.91, + "learning_rate": 7.093279839518557e-05, + "loss": 1.2793, + "theoretical_loss": 3.3482701266589965, + "tokens_seen": 3011313664 + }, + { + "epoch": 0.91, + "objective/train/docs_used": 1094710, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.3427112102508545, + "objective/train/theoretical_loss": 3.348264575943826, + "objective/train/tokens_used": 3031839200, + "theoretical_loss": 3.348264575943826, + "tokens_seen": 3011379200 + }, + { + "epoch": 0.91, + "learning_rate": 7.091675025075227e-05, + "loss": 1.3202, + "theoretical_loss": 3.348264575943826, + "tokens_seen": 3011379200 + }, + { + "epoch": 0.91, + "learning_rate": 7.090070210631895e-05, + "loss": 1.2908, + "theoretical_loss": 3.3482590253832765, + "tokens_seen": 3011444736 + }, + { + "epoch": 0.91, + "learning_rate": 7.088465396188566e-05, + "loss": 1.2874, + "theoretical_loss": 3.34825347497734, + "tokens_seen": 3011510272 + }, + { + "epoch": 0.91, + "learning_rate": 7.086860581745236e-05, + "loss": 1.3743, + "theoretical_loss": 3.3482479247260084, + "tokens_seen": 3011575808 + }, + { + "epoch": 0.91, + "learning_rate": 7.085255767301906e-05, + "loss": 1.2483, + "theoretical_loss": 3.348242374629274, + "tokens_seen": 3011641344 + }, + { + "epoch": 0.91, + "learning_rate": 7.083650952858576e-05, + "loss": 1.2019, + "theoretical_loss": 3.34823682468713, + "tokens_seen": 3011706880 + }, + { + "epoch": 0.91, + "learning_rate": 7.082046138415246e-05, + "loss": 1.1963, + "theoretical_loss": 3.3482312748995686, + "tokens_seen": 3011772416 + }, + { + "epoch": 0.91, + "learning_rate": 7.080441323971916e-05, + "loss": 1.1934, + "theoretical_loss": 3.3482257252665817, + "tokens_seen": 3011837952 + }, + { + "epoch": 0.91, + "learning_rate": 7.078836509528586e-05, + "loss": 1.2893, + "theoretical_loss": 3.3482201757881613, + "tokens_seen": 3011903488 + }, + { + "epoch": 0.91, + "learning_rate": 7.077231695085256e-05, + "loss": 1.347, + "theoretical_loss": 3.3482146264643005, + "tokens_seen": 3011969024 + }, + { + "epoch": 0.91, + "learning_rate": 7.075626880641926e-05, + "loss": 1.4531, + "theoretical_loss": 3.3482090772949906, + "tokens_seen": 3012034560 + }, + { + "epoch": 0.91, + "learning_rate": 7.074022066198596e-05, + "loss": 1.2179, + "theoretical_loss": 3.348203528280225, + "tokens_seen": 3012100096 + }, + { + "epoch": 0.91, + "learning_rate": 7.072417251755266e-05, + "loss": 1.2232, + "theoretical_loss": 3.348197979419996, + "tokens_seen": 3012165632 + }, + { + "epoch": 0.91, + "learning_rate": 7.070812437311936e-05, + "loss": 1.2038, + "theoretical_loss": 3.3481924307142954, + "tokens_seen": 3012231168 + }, + { + "epoch": 0.91, + "learning_rate": 7.069207622868606e-05, + "loss": 1.2119, + "theoretical_loss": 3.3481868821631156, + "tokens_seen": 3012296704 + }, + { + "epoch": 0.91, + "learning_rate": 7.067602808425276e-05, + "loss": 1.3062, + "theoretical_loss": 3.348181333766449, + "tokens_seen": 3012362240 + }, + { + "epoch": 0.91, + "learning_rate": 7.065997993981946e-05, + "loss": 1.2699, + "theoretical_loss": 3.348175785524288, + "tokens_seen": 3012427776 + }, + { + "epoch": 0.91, + "learning_rate": 7.064393179538616e-05, + "loss": 1.2553, + "theoretical_loss": 3.3481702374366247, + "tokens_seen": 3012493312 + }, + { + "epoch": 0.91, + "learning_rate": 7.062788365095287e-05, + "loss": 1.2029, + "theoretical_loss": 3.3481646895034523, + "tokens_seen": 3012558848 + }, + { + "epoch": 0.91, + "learning_rate": 7.061183550651957e-05, + "loss": 1.2853, + "theoretical_loss": 3.3481591417247616, + "tokens_seen": 3012624384 + }, + { + "epoch": 0.91, + "learning_rate": 7.059578736208627e-05, + "loss": 1.3624, + "theoretical_loss": 3.3481535941005465, + "tokens_seen": 3012689920 + }, + { + "epoch": 0.91, + "learning_rate": 7.057973921765297e-05, + "loss": 1.2333, + "theoretical_loss": 3.3481480466307985, + "tokens_seen": 3012755456 + }, + { + "epoch": 0.91, + "learning_rate": 7.056369107321967e-05, + "loss": 1.1755, + "theoretical_loss": 3.34814249931551, + "tokens_seen": 3012820992 + }, + { + "epoch": 0.91, + "learning_rate": 7.054764292878636e-05, + "loss": 1.274, + "theoretical_loss": 3.3481369521546736, + "tokens_seen": 3012886528 + }, + { + "epoch": 0.91, + "learning_rate": 7.053159478435306e-05, + "loss": 1.3315, + "theoretical_loss": 3.3481314051482816, + "tokens_seen": 3012952064 + }, + { + "epoch": 0.91, + "learning_rate": 7.051554663991976e-05, + "loss": 1.2046, + "theoretical_loss": 3.3481258582963256, + "tokens_seen": 3013017600 + }, + { + "epoch": 0.91, + "learning_rate": 7.049949849548646e-05, + "loss": 1.2862, + "theoretical_loss": 3.348120311598799, + "tokens_seen": 3013083136 + }, + { + "epoch": 0.91, + "learning_rate": 7.048345035105316e-05, + "loss": 1.2746, + "theoretical_loss": 3.3481147650556937, + "tokens_seen": 3013148672 + }, + { + "epoch": 0.91, + "learning_rate": 7.046740220661986e-05, + "loss": 1.237, + "theoretical_loss": 3.348109218667002, + "tokens_seen": 3013214208 + }, + { + "epoch": 0.91, + "learning_rate": 7.045135406218656e-05, + "loss": 1.3765, + "theoretical_loss": 3.3481036724327167, + "tokens_seen": 3013279744 + }, + { + "epoch": 0.91, + "learning_rate": 7.043530591775326e-05, + "loss": 1.2997, + "theoretical_loss": 3.348098126352829, + "tokens_seen": 3013345280 + }, + { + "epoch": 0.91, + "learning_rate": 7.041925777331997e-05, + "loss": 1.2688, + "theoretical_loss": 3.3480925804273323, + "tokens_seen": 3013410816 + }, + { + "epoch": 0.91, + "learning_rate": 7.040320962888667e-05, + "loss": 1.2293, + "theoretical_loss": 3.348087034656219, + "tokens_seen": 3013476352 + }, + { + "epoch": 0.91, + "learning_rate": 7.038716148445336e-05, + "loss": 1.2401, + "theoretical_loss": 3.3480814890394806, + "tokens_seen": 3013541888 + }, + { + "epoch": 0.91, + "learning_rate": 7.037111334002006e-05, + "loss": 1.2811, + "theoretical_loss": 3.34807594357711, + "tokens_seen": 3013607424 + }, + { + "epoch": 0.91, + "learning_rate": 7.035506519558676e-05, + "loss": 1.222, + "theoretical_loss": 3.3480703982690994, + "tokens_seen": 3013672960 + }, + { + "epoch": 0.91, + "learning_rate": 7.033901705115346e-05, + "loss": 1.3061, + "theoretical_loss": 3.3480648531154413, + "tokens_seen": 3013738496 + }, + { + "epoch": 0.91, + "learning_rate": 7.032296890672016e-05, + "loss": 1.2882, + "theoretical_loss": 3.348059308116128, + "tokens_seen": 3013804032 + }, + { + "epoch": 0.91, + "learning_rate": 7.030692076228686e-05, + "loss": 1.2276, + "theoretical_loss": 3.348053763271152, + "tokens_seen": 3013869568 + }, + { + "epoch": 0.91, + "learning_rate": 7.029087261785357e-05, + "loss": 1.3363, + "theoretical_loss": 3.348048218580505, + "tokens_seen": 3013935104 + }, + { + "epoch": 0.91, + "learning_rate": 7.027482447342027e-05, + "loss": 1.1744, + "theoretical_loss": 3.34804267404418, + "tokens_seen": 3014000640 + }, + { + "epoch": 0.91, + "learning_rate": 7.025877632898697e-05, + "loss": 1.169, + "theoretical_loss": 3.348037129662169, + "tokens_seen": 3014066176 + }, + { + "epoch": 0.91, + "learning_rate": 7.024272818455367e-05, + "loss": 1.248, + "theoretical_loss": 3.348031585434465, + "tokens_seen": 3014131712 + }, + { + "epoch": 0.91, + "learning_rate": 7.022668004012037e-05, + "loss": 1.2241, + "theoretical_loss": 3.3480260413610594, + "tokens_seen": 3014197248 + }, + { + "epoch": 0.91, + "learning_rate": 7.021063189568707e-05, + "loss": 1.2612, + "theoretical_loss": 3.348020497441945, + "tokens_seen": 3014262784 + }, + { + "epoch": 0.91, + "learning_rate": 7.019458375125376e-05, + "loss": 1.2876, + "theoretical_loss": 3.348014953677114, + "tokens_seen": 3014328320 + }, + { + "epoch": 0.91, + "learning_rate": 7.017853560682046e-05, + "loss": 1.2608, + "theoretical_loss": 3.3480094100665596, + "tokens_seen": 3014393856 + }, + { + "epoch": 0.91, + "learning_rate": 7.016248746238716e-05, + "loss": 1.2408, + "theoretical_loss": 3.3480038666102727, + "tokens_seen": 3014459392 + }, + { + "epoch": 0.91, + "learning_rate": 7.014643931795386e-05, + "loss": 1.1569, + "theoretical_loss": 3.347998323308247, + "tokens_seen": 3014524928 + }, + { + "epoch": 0.91, + "learning_rate": 7.013039117352056e-05, + "loss": 1.2748, + "theoretical_loss": 3.347992780160474, + "tokens_seen": 3014590464 + }, + { + "epoch": 0.91, + "objective/train/docs_used": 1095882, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.292837381362915, + "objective/train/theoretical_loss": 3.3479872371669464, + "objective/train/tokens_used": 3035116000, + "theoretical_loss": 3.3479872371669464, + "tokens_seen": 3014656000 + }, + { + "epoch": 0.91, + "learning_rate": 7.011434302908727e-05, + "loss": 1.3175, + "theoretical_loss": 3.3479872371669464, + "tokens_seen": 3014656000 + }, + { + "epoch": 0.91, + "learning_rate": 7.009829488465397e-05, + "loss": 1.2443, + "theoretical_loss": 3.3479816943276566, + "tokens_seen": 3014721536 + }, + { + "epoch": 0.91, + "learning_rate": 7.008224674022067e-05, + "loss": 1.2716, + "theoretical_loss": 3.3479761516425963, + "tokens_seen": 3014787072 + }, + { + "epoch": 0.91, + "learning_rate": 7.006619859578737e-05, + "loss": 1.1998, + "theoretical_loss": 3.347970609111759, + "tokens_seen": 3014852608 + }, + { + "epoch": 0.91, + "learning_rate": 7.005015045135407e-05, + "loss": 1.2376, + "theoretical_loss": 3.347965066735136, + "tokens_seen": 3014918144 + }, + { + "epoch": 0.91, + "learning_rate": 7.003410230692076e-05, + "loss": 1.2803, + "theoretical_loss": 3.3479595245127203, + "tokens_seen": 3014983680 + }, + { + "epoch": 0.91, + "learning_rate": 7.001805416248746e-05, + "loss": 1.2372, + "theoretical_loss": 3.347953982444504, + "tokens_seen": 3015049216 + }, + { + "epoch": 0.91, + "learning_rate": 7.000200601805416e-05, + "loss": 1.2953, + "theoretical_loss": 3.3479484405304794, + "tokens_seen": 3015114752 + }, + { + "epoch": 0.91, + "learning_rate": 6.998595787362086e-05, + "loss": 1.2277, + "theoretical_loss": 3.347942898770639, + "tokens_seen": 3015180288 + }, + { + "epoch": 0.91, + "learning_rate": 6.996990972918756e-05, + "loss": 1.2558, + "theoretical_loss": 3.3479373571649753, + "tokens_seen": 3015245824 + }, + { + "epoch": 0.91, + "learning_rate": 6.995386158475426e-05, + "loss": 1.2888, + "theoretical_loss": 3.3479318157134808, + "tokens_seen": 3015311360 + }, + { + "epoch": 0.91, + "learning_rate": 6.993781344032097e-05, + "loss": 1.2295, + "theoretical_loss": 3.347926274416147, + "tokens_seen": 3015376896 + }, + { + "epoch": 0.91, + "learning_rate": 6.992176529588767e-05, + "loss": 1.282, + "theoretical_loss": 3.347920733272967, + "tokens_seen": 3015442432 + }, + { + "epoch": 0.91, + "learning_rate": 6.990571715145437e-05, + "loss": 1.2006, + "theoretical_loss": 3.3479151922839328, + "tokens_seen": 3015507968 + }, + { + "epoch": 0.91, + "learning_rate": 6.988966900702107e-05, + "loss": 1.2967, + "theoretical_loss": 3.3479096514490374, + "tokens_seen": 3015573504 + }, + { + "epoch": 0.91, + "learning_rate": 6.987362086258777e-05, + "loss": 1.157, + "theoretical_loss": 3.347904110768272, + "tokens_seen": 3015639040 + }, + { + "epoch": 0.91, + "learning_rate": 6.985757271815447e-05, + "loss": 1.2308, + "theoretical_loss": 3.34789857024163, + "tokens_seen": 3015704576 + }, + { + "epoch": 0.91, + "learning_rate": 6.984152457372116e-05, + "loss": 1.2921, + "theoretical_loss": 3.3478930298691036, + "tokens_seen": 3015770112 + }, + { + "epoch": 0.91, + "learning_rate": 6.982547642928786e-05, + "loss": 1.2192, + "theoretical_loss": 3.347887489650685, + "tokens_seen": 3015835648 + }, + { + "epoch": 0.91, + "learning_rate": 6.980942828485457e-05, + "loss": 1.3349, + "theoretical_loss": 3.3478819495863665, + "tokens_seen": 3015901184 + }, + { + "epoch": 0.91, + "learning_rate": 6.979338014042127e-05, + "loss": 1.237, + "theoretical_loss": 3.3478764096761404, + "tokens_seen": 3015966720 + }, + { + "epoch": 0.91, + "learning_rate": 6.977733199598797e-05, + "loss": 1.088, + "theoretical_loss": 3.3478708699199995, + "tokens_seen": 3016032256 + }, + { + "epoch": 0.91, + "learning_rate": 6.976128385155467e-05, + "loss": 1.2756, + "theoretical_loss": 3.3478653303179353, + "tokens_seen": 3016097792 + }, + { + "epoch": 0.91, + "learning_rate": 6.974523570712137e-05, + "loss": 1.1992, + "theoretical_loss": 3.3478597908699412, + "tokens_seen": 3016163328 + }, + { + "epoch": 0.91, + "learning_rate": 6.972918756268807e-05, + "loss": 1.1917, + "theoretical_loss": 3.347854251576009, + "tokens_seen": 3016228864 + }, + { + "epoch": 0.91, + "learning_rate": 6.971313941825477e-05, + "loss": 1.2561, + "theoretical_loss": 3.347848712436131, + "tokens_seen": 3016294400 + }, + { + "epoch": 0.91, + "learning_rate": 6.969709127382148e-05, + "loss": 1.2464, + "theoretical_loss": 3.3478431734502996, + "tokens_seen": 3016359936 + }, + { + "epoch": 0.91, + "learning_rate": 6.968104312938816e-05, + "loss": 1.1848, + "theoretical_loss": 3.3478376346185077, + "tokens_seen": 3016425472 + }, + { + "epoch": 0.91, + "learning_rate": 6.966499498495486e-05, + "loss": 1.2285, + "theoretical_loss": 3.347832095940747, + "tokens_seen": 3016491008 + }, + { + "epoch": 0.91, + "learning_rate": 6.964894684052156e-05, + "loss": 1.2037, + "theoretical_loss": 3.3478265574170103, + "tokens_seen": 3016556544 + }, + { + "epoch": 0.91, + "learning_rate": 6.963289869608826e-05, + "loss": 1.2896, + "theoretical_loss": 3.3478210190472897, + "tokens_seen": 3016622080 + }, + { + "epoch": 0.91, + "learning_rate": 6.961685055165496e-05, + "loss": 1.1954, + "theoretical_loss": 3.3478154808315774, + "tokens_seen": 3016687616 + }, + { + "epoch": 0.91, + "learning_rate": 6.960080240722167e-05, + "loss": 1.3591, + "theoretical_loss": 3.3478099427698664, + "tokens_seen": 3016753152 + }, + { + "epoch": 0.91, + "learning_rate": 6.958475426278837e-05, + "loss": 1.3055, + "theoretical_loss": 3.3478044048621487, + "tokens_seen": 3016818688 + }, + { + "epoch": 0.91, + "learning_rate": 6.956870611835507e-05, + "loss": 1.4339, + "theoretical_loss": 3.3477988671084167, + "tokens_seen": 3016884224 + }, + { + "epoch": 0.91, + "learning_rate": 6.955265797392177e-05, + "loss": 1.186, + "theoretical_loss": 3.3477933295086624, + "tokens_seen": 3016949760 + }, + { + "epoch": 0.91, + "learning_rate": 6.953660982948847e-05, + "loss": 1.2536, + "theoretical_loss": 3.3477877920628787, + "tokens_seen": 3017015296 + }, + { + "epoch": 0.91, + "learning_rate": 6.952056168505517e-05, + "loss": 1.2354, + "theoretical_loss": 3.347782254771058, + "tokens_seen": 3017080832 + }, + { + "epoch": 0.91, + "learning_rate": 6.950451354062187e-05, + "loss": 1.2133, + "theoretical_loss": 3.3477767176331925, + "tokens_seen": 3017146368 + }, + { + "epoch": 0.91, + "learning_rate": 6.948846539618857e-05, + "loss": 1.3466, + "theoretical_loss": 3.3477711806492745, + "tokens_seen": 3017211904 + }, + { + "epoch": 0.91, + "learning_rate": 6.947241725175527e-05, + "loss": 1.2901, + "theoretical_loss": 3.3477656438192964, + "tokens_seen": 3017277440 + }, + { + "epoch": 0.91, + "learning_rate": 6.945636910732197e-05, + "loss": 1.2812, + "theoretical_loss": 3.347760107143251, + "tokens_seen": 3017342976 + }, + { + "epoch": 0.91, + "learning_rate": 6.944032096288867e-05, + "loss": 1.2504, + "theoretical_loss": 3.3477545706211296, + "tokens_seen": 3017408512 + }, + { + "epoch": 0.91, + "learning_rate": 6.942427281845537e-05, + "loss": 1.2538, + "theoretical_loss": 3.3477490342529257, + "tokens_seen": 3017474048 + }, + { + "epoch": 0.91, + "learning_rate": 6.940822467402207e-05, + "loss": 1.2683, + "theoretical_loss": 3.347743498038631, + "tokens_seen": 3017539584 + }, + { + "epoch": 0.91, + "learning_rate": 6.939217652958877e-05, + "loss": 1.2099, + "theoretical_loss": 3.3477379619782384, + "tokens_seen": 3017605120 + }, + { + "epoch": 0.91, + "learning_rate": 6.937612838515547e-05, + "loss": 1.2306, + "theoretical_loss": 3.34773242607174, + "tokens_seen": 3017670656 + }, + { + "epoch": 0.91, + "learning_rate": 6.936008024072217e-05, + "loss": 1.2764, + "theoretical_loss": 3.347726890319128, + "tokens_seen": 3017736192 + }, + { + "epoch": 0.91, + "learning_rate": 6.934403209628888e-05, + "loss": 1.1927, + "theoretical_loss": 3.3477213547203952, + "tokens_seen": 3017801728 + }, + { + "epoch": 0.91, + "learning_rate": 6.932798395185556e-05, + "loss": 1.2214, + "theoretical_loss": 3.3477158192755336, + "tokens_seen": 3017867264 + }, + { + "epoch": 0.91, + "objective/train/docs_used": 1096954, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.186356544494629, + "objective/train/theoretical_loss": 3.3477102839845356, + "objective/train/tokens_used": 3038392800, + "theoretical_loss": 3.3477102839845356, + "tokens_seen": 3017932800 + }, + { + "epoch": 0.91, + "learning_rate": 6.931193580742226e-05, + "loss": 1.1685, + "theoretical_loss": 3.3477102839845356, + "tokens_seen": 3017932800 + }, + { + "epoch": 0.91, + "learning_rate": 6.929588766298896e-05, + "loss": 1.2402, + "theoretical_loss": 3.3477047488473937, + "tokens_seen": 3017998336 + }, + { + "epoch": 0.91, + "learning_rate": 6.927983951855566e-05, + "loss": 1.1728, + "theoretical_loss": 3.3476992138641006, + "tokens_seen": 3018063872 + }, + { + "epoch": 0.91, + "learning_rate": 6.926379137412237e-05, + "loss": 1.1176, + "theoretical_loss": 3.347693679034648, + "tokens_seen": 3018129408 + }, + { + "epoch": 0.91, + "learning_rate": 6.924774322968907e-05, + "loss": 1.2724, + "theoretical_loss": 3.347688144359029, + "tokens_seen": 3018194944 + }, + { + "epoch": 0.91, + "learning_rate": 6.923169508525577e-05, + "loss": 1.2158, + "theoretical_loss": 3.3476826098372356, + "tokens_seen": 3018260480 + }, + { + "epoch": 0.91, + "learning_rate": 6.921564694082248e-05, + "loss": 1.2178, + "theoretical_loss": 3.34767707546926, + "tokens_seen": 3018326016 + }, + { + "epoch": 0.91, + "learning_rate": 6.919959879638918e-05, + "loss": 1.3536, + "theoretical_loss": 3.347671541255095, + "tokens_seen": 3018391552 + }, + { + "epoch": 0.91, + "learning_rate": 6.918355065195588e-05, + "loss": 1.3741, + "theoretical_loss": 3.3476660071947326, + "tokens_seen": 3018457088 + }, + { + "epoch": 0.91, + "learning_rate": 6.916750250752257e-05, + "loss": 1.19, + "theoretical_loss": 3.347660473288166, + "tokens_seen": 3018522624 + }, + { + "epoch": 0.91, + "learning_rate": 6.915145436308927e-05, + "loss": 1.2982, + "theoretical_loss": 3.3476549395353863, + "tokens_seen": 3018588160 + }, + { + "epoch": 0.91, + "learning_rate": 6.913540621865597e-05, + "loss": 1.2683, + "theoretical_loss": 3.347649405936387, + "tokens_seen": 3018653696 + }, + { + "epoch": 0.91, + "learning_rate": 6.911935807422267e-05, + "loss": 1.3379, + "theoretical_loss": 3.34764387249116, + "tokens_seen": 3018719232 + }, + { + "epoch": 0.91, + "learning_rate": 6.910330992978937e-05, + "loss": 1.1926, + "theoretical_loss": 3.347638339199697, + "tokens_seen": 3018784768 + }, + { + "epoch": 0.91, + "learning_rate": 6.908726178535607e-05, + "loss": 1.3649, + "theoretical_loss": 3.347632806061992, + "tokens_seen": 3018850304 + }, + { + "epoch": 0.91, + "learning_rate": 6.907121364092277e-05, + "loss": 1.2524, + "theoretical_loss": 3.3476272730780363, + "tokens_seen": 3018915840 + }, + { + "epoch": 0.91, + "learning_rate": 6.905516549648947e-05, + "loss": 1.2898, + "theoretical_loss": 3.3476217402478223, + "tokens_seen": 3018981376 + }, + { + "epoch": 0.91, + "learning_rate": 6.903911735205617e-05, + "loss": 1.1994, + "theoretical_loss": 3.3476162075713427, + "tokens_seen": 3019046912 + }, + { + "epoch": 0.91, + "learning_rate": 6.902306920762287e-05, + "loss": 1.2153, + "theoretical_loss": 3.34761067504859, + "tokens_seen": 3019112448 + }, + { + "epoch": 0.91, + "learning_rate": 6.900702106318958e-05, + "loss": 1.2846, + "theoretical_loss": 3.3476051426795563, + "tokens_seen": 3019177984 + }, + { + "epoch": 0.91, + "learning_rate": 6.899097291875628e-05, + "loss": 1.336, + "theoretical_loss": 3.347599610464234, + "tokens_seen": 3019243520 + }, + { + "epoch": 0.91, + "learning_rate": 6.897492477432296e-05, + "loss": 1.3436, + "theoretical_loss": 3.3475940784026155, + "tokens_seen": 3019309056 + }, + { + "epoch": 0.91, + "learning_rate": 6.895887662988966e-05, + "loss": 1.1113, + "theoretical_loss": 3.347588546494693, + "tokens_seen": 3019374592 + }, + { + "epoch": 0.91, + "learning_rate": 6.894282848545636e-05, + "loss": 1.196, + "theoretical_loss": 3.3475830147404597, + "tokens_seen": 3019440128 + }, + { + "epoch": 0.92, + "learning_rate": 6.892678034102306e-05, + "loss": 1.1521, + "theoretical_loss": 3.3475774831399074, + "tokens_seen": 3019505664 + }, + { + "epoch": 0.92, + "learning_rate": 6.891073219658978e-05, + "loss": 1.2288, + "theoretical_loss": 3.3475719516930287, + "tokens_seen": 3019571200 + }, + { + "epoch": 0.92, + "learning_rate": 6.889468405215648e-05, + "loss": 1.218, + "theoretical_loss": 3.347566420399815, + "tokens_seen": 3019636736 + }, + { + "epoch": 0.92, + "learning_rate": 6.887863590772318e-05, + "loss": 1.3063, + "theoretical_loss": 3.34756088926026, + "tokens_seen": 3019702272 + }, + { + "epoch": 0.92, + "learning_rate": 6.886258776328988e-05, + "loss": 1.1875, + "theoretical_loss": 3.347555358274356, + "tokens_seen": 3019767808 + }, + { + "epoch": 0.92, + "learning_rate": 6.884653961885658e-05, + "loss": 1.1583, + "theoretical_loss": 3.3475498274420947, + "tokens_seen": 3019833344 + }, + { + "epoch": 0.92, + "learning_rate": 6.883049147442327e-05, + "loss": 1.2356, + "theoretical_loss": 3.347544296763469, + "tokens_seen": 3019898880 + }, + { + "epoch": 0.92, + "learning_rate": 6.881444332998997e-05, + "loss": 1.2308, + "theoretical_loss": 3.347538766238471, + "tokens_seen": 3019964416 + }, + { + "epoch": 0.92, + "learning_rate": 6.879839518555667e-05, + "loss": 1.276, + "theoretical_loss": 3.3475332358670933, + "tokens_seen": 3020029952 + }, + { + "epoch": 0.92, + "learning_rate": 6.878234704112337e-05, + "loss": 1.3302, + "theoretical_loss": 3.347527705649328, + "tokens_seen": 3020095488 + }, + { + "epoch": 0.92, + "learning_rate": 6.876629889669007e-05, + "loss": 1.2707, + "theoretical_loss": 3.347522175585168, + "tokens_seen": 3020161024 + }, + { + "epoch": 0.92, + "learning_rate": 6.875025075225677e-05, + "loss": 1.2238, + "theoretical_loss": 3.347516645674605, + "tokens_seen": 3020226560 + }, + { + "epoch": 0.92, + "learning_rate": 6.873420260782347e-05, + "loss": 1.259, + "theoretical_loss": 3.3475111159176323, + "tokens_seen": 3020292096 + }, + { + "epoch": 0.92, + "learning_rate": 6.871815446339017e-05, + "loss": 1.2771, + "theoretical_loss": 3.3475055863142416, + "tokens_seen": 3020357632 + }, + { + "epoch": 0.92, + "learning_rate": 6.870210631895687e-05, + "loss": 1.2426, + "theoretical_loss": 3.3475000568644258, + "tokens_seen": 3020423168 + }, + { + "epoch": 0.92, + "learning_rate": 6.868605817452357e-05, + "loss": 1.2171, + "theoretical_loss": 3.3474945275681764, + "tokens_seen": 3020488704 + }, + { + "epoch": 0.92, + "learning_rate": 6.867001003009028e-05, + "loss": 1.3192, + "theoretical_loss": 3.347488998425487, + "tokens_seen": 3020554240 + }, + { + "epoch": 0.92, + "learning_rate": 6.865396188565698e-05, + "loss": 1.1288, + "theoretical_loss": 3.3474834694363493, + "tokens_seen": 3020619776 + }, + { + "epoch": 0.92, + "learning_rate": 6.863791374122366e-05, + "loss": 1.2234, + "theoretical_loss": 3.347477940600756, + "tokens_seen": 3020685312 + }, + { + "epoch": 0.92, + "learning_rate": 6.862186559679036e-05, + "loss": 1.2938, + "theoretical_loss": 3.3474724119186994, + "tokens_seen": 3020750848 + }, + { + "epoch": 0.92, + "learning_rate": 6.860581745235708e-05, + "loss": 1.1798, + "theoretical_loss": 3.3474668833901715, + "tokens_seen": 3020816384 + }, + { + "epoch": 0.92, + "learning_rate": 6.858976930792378e-05, + "loss": 1.265, + "theoretical_loss": 3.3474613550151653, + "tokens_seen": 3020881920 + }, + { + "epoch": 0.92, + "learning_rate": 6.857372116349048e-05, + "loss": 1.2204, + "theoretical_loss": 3.3474558267936727, + "tokens_seen": 3020947456 + }, + { + "epoch": 0.92, + "learning_rate": 6.855767301905718e-05, + "loss": 1.2497, + "theoretical_loss": 3.3474502987256867, + "tokens_seen": 3021012992 + }, + { + "epoch": 0.92, + "learning_rate": 6.854162487462388e-05, + "loss": 1.241, + "theoretical_loss": 3.347444770811199, + "tokens_seen": 3021078528 + }, + { + "epoch": 0.92, + "learning_rate": 6.852557673019058e-05, + "loss": 1.1709, + "theoretical_loss": 3.3474392430502027, + "tokens_seen": 3021144064 + }, + { + "epoch": 0.92, + "objective/train/docs_used": 1098075, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.1954524517059326, + "objective/train/theoretical_loss": 3.34743371544269, + "objective/train/tokens_used": 3041669600, + "theoretical_loss": 3.34743371544269, + "tokens_seen": 3021209600 + }, + { + "epoch": 0.92, + "learning_rate": 6.850952858575728e-05, + "loss": 1.1997, + "theoretical_loss": 3.34743371544269, + "tokens_seen": 3021209600 + }, + { + "epoch": 0.92, + "learning_rate": 6.849348044132398e-05, + "loss": 1.2318, + "theoretical_loss": 3.347428187988653, + "tokens_seen": 3021275136 + }, + { + "epoch": 0.92, + "learning_rate": 6.847743229689067e-05, + "loss": 1.2867, + "theoretical_loss": 3.3474226606880846, + "tokens_seen": 3021340672 + }, + { + "epoch": 0.92, + "learning_rate": 6.846138415245737e-05, + "loss": 1.2833, + "theoretical_loss": 3.3474171335409766, + "tokens_seen": 3021406208 + }, + { + "epoch": 0.92, + "learning_rate": 6.844533600802407e-05, + "loss": 1.1818, + "theoretical_loss": 3.347411606547322, + "tokens_seen": 3021471744 + }, + { + "epoch": 0.92, + "learning_rate": 6.842928786359077e-05, + "loss": 1.2172, + "theoretical_loss": 3.3474060797071123, + "tokens_seen": 3021537280 + }, + { + "epoch": 0.92, + "learning_rate": 6.841323971915747e-05, + "loss": 1.3012, + "theoretical_loss": 3.347400553020341, + "tokens_seen": 3021602816 + }, + { + "epoch": 0.92, + "learning_rate": 6.839719157472417e-05, + "loss": 1.1602, + "theoretical_loss": 3.3473950264870003, + "tokens_seen": 3021668352 + }, + { + "epoch": 0.92, + "learning_rate": 6.838114343029087e-05, + "loss": 1.3639, + "theoretical_loss": 3.347389500107082, + "tokens_seen": 3021733888 + }, + { + "epoch": 0.92, + "learning_rate": 6.836509528585757e-05, + "loss": 1.2866, + "theoretical_loss": 3.3473839738805795, + "tokens_seen": 3021799424 + }, + { + "epoch": 0.92, + "learning_rate": 6.834904714142427e-05, + "loss": 1.2851, + "theoretical_loss": 3.3473784478074844, + "tokens_seen": 3021864960 + }, + { + "epoch": 0.92, + "learning_rate": 6.833299899699097e-05, + "loss": 1.253, + "theoretical_loss": 3.347372921887789, + "tokens_seen": 3021930496 + }, + { + "epoch": 0.92, + "learning_rate": 6.831695085255768e-05, + "loss": 1.266, + "theoretical_loss": 3.347367396121486, + "tokens_seen": 3021996032 + }, + { + "epoch": 0.92, + "learning_rate": 6.830090270812438e-05, + "loss": 1.185, + "theoretical_loss": 3.347361870508568, + "tokens_seen": 3022061568 + }, + { + "epoch": 0.92, + "learning_rate": 6.828485456369108e-05, + "loss": 1.2817, + "theoretical_loss": 3.3473563450490276, + "tokens_seen": 3022127104 + }, + { + "epoch": 0.92, + "learning_rate": 6.826880641925778e-05, + "loss": 1.2922, + "theoretical_loss": 3.3473508197428563, + "tokens_seen": 3022192640 + }, + { + "epoch": 0.92, + "learning_rate": 6.825275827482448e-05, + "loss": 1.1785, + "theoretical_loss": 3.3473452945900477, + "tokens_seen": 3022258176 + }, + { + "epoch": 0.92, + "learning_rate": 6.823671013039118e-05, + "loss": 1.3159, + "theoretical_loss": 3.3473397695905933, + "tokens_seen": 3022323712 + }, + { + "epoch": 0.92, + "learning_rate": 6.822066198595788e-05, + "loss": 1.2511, + "theoretical_loss": 3.347334244744486, + "tokens_seen": 3022389248 + }, + { + "epoch": 0.92, + "learning_rate": 6.820461384152458e-05, + "loss": 1.3097, + "theoretical_loss": 3.347328720051718, + "tokens_seen": 3022454784 + }, + { + "epoch": 0.92, + "learning_rate": 6.818856569709128e-05, + "loss": 1.2184, + "theoretical_loss": 3.3473231955122817, + "tokens_seen": 3022520320 + }, + { + "epoch": 0.92, + "learning_rate": 6.817251755265798e-05, + "loss": 1.2726, + "theoretical_loss": 3.3473176711261696, + "tokens_seen": 3022585856 + }, + { + "epoch": 0.92, + "learning_rate": 6.815646940822468e-05, + "loss": 1.0677, + "theoretical_loss": 3.3473121468933744, + "tokens_seen": 3022651392 + }, + { + "epoch": 0.92, + "learning_rate": 6.814042126379138e-05, + "loss": 1.2076, + "theoretical_loss": 3.347306622813888, + "tokens_seen": 3022716928 + }, + { + "epoch": 0.92, + "learning_rate": 6.812437311935807e-05, + "loss": 1.2501, + "theoretical_loss": 3.347301098887703, + "tokens_seen": 3022782464 + }, + { + "epoch": 0.92, + "learning_rate": 6.810832497492477e-05, + "loss": 1.1359, + "theoretical_loss": 3.347295575114812, + "tokens_seen": 3022848000 + }, + { + "epoch": 0.92, + "learning_rate": 6.809227683049147e-05, + "loss": 1.2412, + "theoretical_loss": 3.3472900514952073, + "tokens_seen": 3022913536 + }, + { + "epoch": 0.92, + "learning_rate": 6.807622868605817e-05, + "loss": 1.1198, + "theoretical_loss": 3.347284528028881, + "tokens_seen": 3022979072 + }, + { + "epoch": 0.92, + "learning_rate": 6.806018054162487e-05, + "loss": 1.3086, + "theoretical_loss": 3.3472790047158267, + "tokens_seen": 3023044608 + }, + { + "epoch": 0.92, + "learning_rate": 6.804413239719157e-05, + "loss": 1.288, + "theoretical_loss": 3.3472734815560354, + "tokens_seen": 3023110144 + }, + { + "epoch": 0.92, + "learning_rate": 6.802808425275827e-05, + "loss": 1.1786, + "theoretical_loss": 3.3472679585495, + "tokens_seen": 3023175680 + }, + { + "epoch": 0.92, + "learning_rate": 6.801203610832499e-05, + "loss": 1.2206, + "theoretical_loss": 3.3472624356962135, + "tokens_seen": 3023241216 + }, + { + "epoch": 0.92, + "learning_rate": 6.799598796389169e-05, + "loss": 1.3083, + "theoretical_loss": 3.3472569129961673, + "tokens_seen": 3023306752 + }, + { + "epoch": 0.92, + "learning_rate": 6.797993981945839e-05, + "loss": 1.3554, + "theoretical_loss": 3.347251390449355, + "tokens_seen": 3023372288 + }, + { + "epoch": 0.92, + "learning_rate": 6.796389167502508e-05, + "loss": 1.2925, + "theoretical_loss": 3.347245868055768, + "tokens_seen": 3023437824 + }, + { + "epoch": 0.92, + "learning_rate": 6.794784353059178e-05, + "loss": 1.2172, + "theoretical_loss": 3.347240345815399, + "tokens_seen": 3023503360 + }, + { + "epoch": 0.92, + "learning_rate": 6.793179538615848e-05, + "loss": 1.2888, + "theoretical_loss": 3.347234823728241, + "tokens_seen": 3023568896 + }, + { + "epoch": 0.92, + "learning_rate": 6.791574724172518e-05, + "loss": 1.2591, + "theoretical_loss": 3.347229301794286, + "tokens_seen": 3023634432 + }, + { + "epoch": 0.92, + "learning_rate": 6.789969909729188e-05, + "loss": 1.3025, + "theoretical_loss": 3.3472237800135263, + "tokens_seen": 3023699968 + }, + { + "epoch": 0.92, + "learning_rate": 6.788365095285858e-05, + "loss": 1.2366, + "theoretical_loss": 3.347218258385954, + "tokens_seen": 3023765504 + }, + { + "epoch": 0.92, + "learning_rate": 6.786760280842528e-05, + "loss": 1.1684, + "theoretical_loss": 3.3472127369115627, + "tokens_seen": 3023831040 + }, + { + "epoch": 0.92, + "learning_rate": 6.785155466399198e-05, + "loss": 1.2193, + "theoretical_loss": 3.347207215590344, + "tokens_seen": 3023896576 + }, + { + "epoch": 0.92, + "learning_rate": 6.783550651955868e-05, + "loss": 1.3262, + "theoretical_loss": 3.34720169442229, + "tokens_seen": 3023962112 + }, + { + "epoch": 0.92, + "learning_rate": 6.781945837512538e-05, + "loss": 1.2915, + "theoretical_loss": 3.347196173407394, + "tokens_seen": 3024027648 + }, + { + "epoch": 0.92, + "learning_rate": 6.780341023069208e-05, + "loss": 1.2554, + "theoretical_loss": 3.347190652545648, + "tokens_seen": 3024093184 + }, + { + "epoch": 0.92, + "learning_rate": 6.778736208625878e-05, + "loss": 1.3236, + "theoretical_loss": 3.347185131837044, + "tokens_seen": 3024158720 + }, + { + "epoch": 0.92, + "learning_rate": 6.777131394182547e-05, + "loss": 1.307, + "theoretical_loss": 3.3471796112815753, + "tokens_seen": 3024224256 + }, + { + "epoch": 0.92, + "learning_rate": 6.775526579739217e-05, + "loss": 1.1484, + "theoretical_loss": 3.347174090879234, + "tokens_seen": 3024289792 + }, + { + "epoch": 0.92, + "learning_rate": 6.773921765295887e-05, + "loss": 1.2927, + "theoretical_loss": 3.347168570630012, + "tokens_seen": 3024355328 + }, + { + "epoch": 0.92, + "learning_rate": 6.772316950852557e-05, + "loss": 1.2287, + "theoretical_loss": 3.3471630505339025, + "tokens_seen": 3024420864 + }, + { + "epoch": 0.92, + "objective/train/docs_used": 1099208, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.2432944774627686, + "objective/train/theoretical_loss": 3.3471575305908976, + "objective/train/tokens_used": 3044946400, + "theoretical_loss": 3.3471575305908976, + "tokens_seen": 3024486400 + }, + { + "epoch": 0.92, + "learning_rate": 6.770712136409229e-05, + "loss": 1.2947, + "theoretical_loss": 3.3471575305908976, + "tokens_seen": 3024486400 + }, + { + "epoch": 0.92, + "learning_rate": 6.769107321965899e-05, + "loss": 1.3467, + "theoretical_loss": 3.3471520108009893, + "tokens_seen": 3024551936 + }, + { + "epoch": 0.92, + "learning_rate": 6.767502507522569e-05, + "loss": 1.2219, + "theoretical_loss": 3.347146491164171, + "tokens_seen": 3024617472 + }, + { + "epoch": 0.92, + "learning_rate": 6.765897693079239e-05, + "loss": 1.1885, + "theoretical_loss": 3.3471409716804343, + "tokens_seen": 3024683008 + }, + { + "epoch": 0.92, + "learning_rate": 6.764292878635909e-05, + "loss": 1.2456, + "theoretical_loss": 3.3471354523497725, + "tokens_seen": 3024748544 + }, + { + "epoch": 0.92, + "learning_rate": 6.762688064192579e-05, + "loss": 1.3624, + "theoretical_loss": 3.347129933172177, + "tokens_seen": 3024814080 + }, + { + "epoch": 0.92, + "learning_rate": 6.761083249749248e-05, + "loss": 1.1691, + "theoretical_loss": 3.347124414147641, + "tokens_seen": 3024879616 + }, + { + "epoch": 0.92, + "learning_rate": 6.759478435305918e-05, + "loss": 1.2126, + "theoretical_loss": 3.347118895276157, + "tokens_seen": 3024945152 + }, + { + "epoch": 0.92, + "learning_rate": 6.757873620862588e-05, + "loss": 1.2118, + "theoretical_loss": 3.3471133765577163, + "tokens_seen": 3025010688 + }, + { + "epoch": 0.92, + "learning_rate": 6.756268806419258e-05, + "loss": 1.1141, + "theoretical_loss": 3.3471078579923126, + "tokens_seen": 3025076224 + }, + { + "epoch": 0.92, + "learning_rate": 6.754663991975928e-05, + "loss": 1.3837, + "theoretical_loss": 3.347102339579938, + "tokens_seen": 3025141760 + }, + { + "epoch": 0.92, + "learning_rate": 6.753059177532598e-05, + "loss": 1.2332, + "theoretical_loss": 3.3470968213205845, + "tokens_seen": 3025207296 + }, + { + "epoch": 0.92, + "learning_rate": 6.751454363089268e-05, + "loss": 1.2845, + "theoretical_loss": 3.3470913032142455, + "tokens_seen": 3025272832 + }, + { + "epoch": 0.92, + "learning_rate": 6.749849548645938e-05, + "loss": 1.2248, + "theoretical_loss": 3.347085785260912, + "tokens_seen": 3025338368 + }, + { + "epoch": 0.92, + "learning_rate": 6.748244734202608e-05, + "loss": 1.3004, + "theoretical_loss": 3.347080267460578, + "tokens_seen": 3025403904 + }, + { + "epoch": 0.92, + "learning_rate": 6.746639919759278e-05, + "loss": 1.3083, + "theoretical_loss": 3.347074749813235, + "tokens_seen": 3025469440 + }, + { + "epoch": 0.92, + "learning_rate": 6.745035105315948e-05, + "loss": 1.2161, + "theoretical_loss": 3.3470692323188755, + "tokens_seen": 3025534976 + }, + { + "epoch": 0.92, + "learning_rate": 6.743430290872618e-05, + "loss": 1.2247, + "theoretical_loss": 3.347063714977492, + "tokens_seen": 3025600512 + }, + { + "epoch": 0.92, + "learning_rate": 6.741825476429287e-05, + "loss": 1.2727, + "theoretical_loss": 3.3470581977890776, + "tokens_seen": 3025666048 + }, + { + "epoch": 0.92, + "learning_rate": 6.740220661985958e-05, + "loss": 1.3228, + "theoretical_loss": 3.347052680753624, + "tokens_seen": 3025731584 + }, + { + "epoch": 0.92, + "learning_rate": 6.738615847542629e-05, + "loss": 1.2661, + "theoretical_loss": 3.3470471638711237, + "tokens_seen": 3025797120 + }, + { + "epoch": 0.92, + "learning_rate": 6.737011033099299e-05, + "loss": 1.3551, + "theoretical_loss": 3.347041647141569, + "tokens_seen": 3025862656 + }, + { + "epoch": 0.92, + "learning_rate": 6.735406218655969e-05, + "loss": 1.2621, + "theoretical_loss": 3.3470361305649527, + "tokens_seen": 3025928192 + }, + { + "epoch": 0.92, + "learning_rate": 6.733801404212639e-05, + "loss": 1.2825, + "theoretical_loss": 3.347030614141268, + "tokens_seen": 3025993728 + }, + { + "epoch": 0.92, + "learning_rate": 6.732196589769309e-05, + "loss": 1.3126, + "theoretical_loss": 3.347025097870506, + "tokens_seen": 3026059264 + }, + { + "epoch": 0.92, + "learning_rate": 6.730591775325979e-05, + "loss": 1.329, + "theoretical_loss": 3.3470195817526593, + "tokens_seen": 3026124800 + }, + { + "epoch": 0.92, + "learning_rate": 6.728986960882649e-05, + "loss": 1.2413, + "theoretical_loss": 3.347014065787721, + "tokens_seen": 3026190336 + }, + { + "epoch": 0.92, + "learning_rate": 6.727382146439319e-05, + "loss": 1.1772, + "theoretical_loss": 3.3470085499756834, + "tokens_seen": 3026255872 + }, + { + "epoch": 0.92, + "learning_rate": 6.725777331995988e-05, + "loss": 1.2379, + "theoretical_loss": 3.347003034316539, + "tokens_seen": 3026321408 + }, + { + "epoch": 0.92, + "learning_rate": 6.724172517552658e-05, + "loss": 1.269, + "theoretical_loss": 3.3469975188102796, + "tokens_seen": 3026386944 + }, + { + "epoch": 0.92, + "learning_rate": 6.722567703109328e-05, + "loss": 1.1982, + "theoretical_loss": 3.3469920034568985, + "tokens_seen": 3026452480 + }, + { + "epoch": 0.92, + "learning_rate": 6.720962888665998e-05, + "loss": 1.3185, + "theoretical_loss": 3.3469864882563876, + "tokens_seen": 3026518016 + }, + { + "epoch": 0.92, + "learning_rate": 6.719358074222668e-05, + "loss": 1.2854, + "theoretical_loss": 3.3469809732087397, + "tokens_seen": 3026583552 + }, + { + "epoch": 0.92, + "learning_rate": 6.717753259779338e-05, + "loss": 1.2779, + "theoretical_loss": 3.346975458313947, + "tokens_seen": 3026649088 + }, + { + "epoch": 0.92, + "learning_rate": 6.716148445336008e-05, + "loss": 1.243, + "theoretical_loss": 3.346969943572002, + "tokens_seen": 3026714624 + }, + { + "epoch": 0.92, + "learning_rate": 6.714543630892678e-05, + "loss": 1.3007, + "theoretical_loss": 3.346964428982897, + "tokens_seen": 3026780160 + }, + { + "epoch": 0.92, + "learning_rate": 6.712938816449348e-05, + "loss": 1.2487, + "theoretical_loss": 3.346958914546625, + "tokens_seen": 3026845696 + }, + { + "epoch": 0.92, + "learning_rate": 6.711334002006018e-05, + "loss": 1.1695, + "theoretical_loss": 3.3469534002631782, + "tokens_seen": 3026911232 + }, + { + "epoch": 0.92, + "learning_rate": 6.709729187562688e-05, + "loss": 1.3187, + "theoretical_loss": 3.346947886132549, + "tokens_seen": 3026976768 + }, + { + "epoch": 0.92, + "learning_rate": 6.708124373119358e-05, + "loss": 1.2188, + "theoretical_loss": 3.3469423721547296, + "tokens_seen": 3027042304 + }, + { + "epoch": 0.92, + "learning_rate": 6.706519558676028e-05, + "loss": 1.3352, + "theoretical_loss": 3.3469368583297125, + "tokens_seen": 3027107840 + }, + { + "epoch": 0.92, + "learning_rate": 6.704914744232698e-05, + "loss": 1.3642, + "theoretical_loss": 3.3469313446574906, + "tokens_seen": 3027173376 + }, + { + "epoch": 0.92, + "learning_rate": 6.703309929789369e-05, + "loss": 1.2736, + "theoretical_loss": 3.346925831138056, + "tokens_seen": 3027238912 + }, + { + "epoch": 0.92, + "learning_rate": 6.701705115346039e-05, + "loss": 1.249, + "theoretical_loss": 3.346920317771401, + "tokens_seen": 3027304448 + }, + { + "epoch": 0.92, + "learning_rate": 6.700100300902709e-05, + "loss": 1.2992, + "theoretical_loss": 3.3469148045575188, + "tokens_seen": 3027369984 + }, + { + "epoch": 0.92, + "learning_rate": 6.698495486459379e-05, + "loss": 1.3085, + "theoretical_loss": 3.3469092914964014, + "tokens_seen": 3027435520 + }, + { + "epoch": 0.92, + "learning_rate": 6.696890672016049e-05, + "loss": 1.22, + "theoretical_loss": 3.346903778588041, + "tokens_seen": 3027501056 + }, + { + "epoch": 0.92, + "learning_rate": 6.695285857572719e-05, + "loss": 1.1631, + "theoretical_loss": 3.34689826583243, + "tokens_seen": 3027566592 + }, + { + "epoch": 0.92, + "learning_rate": 6.693681043129389e-05, + "loss": 1.2842, + "theoretical_loss": 3.346892753229562, + "tokens_seen": 3027632128 + }, + { + "epoch": 0.92, + "learning_rate": 6.692076228686058e-05, + "loss": 1.3863, + "theoretical_loss": 3.3468872407794277, + "tokens_seen": 3027697664 + }, + { + "epoch": 0.92, + "objective/train/docs_used": 1100401, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.3156346082687378, + "objective/train/theoretical_loss": 3.346881728482021, + "objective/train/tokens_used": 3048223200, + "theoretical_loss": 3.346881728482021, + "tokens_seen": 3027763200 + }, + { + "epoch": 0.92, + "learning_rate": 6.690471414242728e-05, + "loss": 1.2409, + "theoretical_loss": 3.346881728482021, + "tokens_seen": 3027763200 + }, + { + "epoch": 0.92, + "learning_rate": 6.688866599799398e-05, + "loss": 1.224, + "theoretical_loss": 3.346876216337334, + "tokens_seen": 3027828736 + }, + { + "epoch": 0.92, + "learning_rate": 6.687261785356068e-05, + "loss": 1.1926, + "theoretical_loss": 3.346870704345358, + "tokens_seen": 3027894272 + }, + { + "epoch": 0.92, + "learning_rate": 6.685656970912738e-05, + "loss": 1.2797, + "theoretical_loss": 3.3468651925060877, + "tokens_seen": 3027959808 + }, + { + "epoch": 0.92, + "learning_rate": 6.684052156469408e-05, + "loss": 1.0795, + "theoretical_loss": 3.346859680819514, + "tokens_seen": 3028025344 + }, + { + "epoch": 0.92, + "learning_rate": 6.682447342026078e-05, + "loss": 1.2474, + "theoretical_loss": 3.3468541692856295, + "tokens_seen": 3028090880 + }, + { + "epoch": 0.92, + "learning_rate": 6.68084252758275e-05, + "loss": 1.2637, + "theoretical_loss": 3.3468486579044265, + "tokens_seen": 3028156416 + }, + { + "epoch": 0.92, + "learning_rate": 6.67923771313942e-05, + "loss": 1.2976, + "theoretical_loss": 3.3468431466758983, + "tokens_seen": 3028221952 + }, + { + "epoch": 0.92, + "learning_rate": 6.67763289869609e-05, + "loss": 1.2695, + "theoretical_loss": 3.346837635600037, + "tokens_seen": 3028287488 + }, + { + "epoch": 0.92, + "learning_rate": 6.676028084252758e-05, + "loss": 1.2912, + "theoretical_loss": 3.3468321246768347, + "tokens_seen": 3028353024 + }, + { + "epoch": 0.92, + "learning_rate": 6.674423269809428e-05, + "loss": 1.3228, + "theoretical_loss": 3.346826613906284, + "tokens_seen": 3028418560 + }, + { + "epoch": 0.92, + "learning_rate": 6.672818455366098e-05, + "loss": 1.3568, + "theoretical_loss": 3.346821103288378, + "tokens_seen": 3028484096 + }, + { + "epoch": 0.92, + "learning_rate": 6.671213640922768e-05, + "loss": 1.1464, + "theoretical_loss": 3.3468155928231083, + "tokens_seen": 3028549632 + }, + { + "epoch": 0.92, + "learning_rate": 6.669608826479439e-05, + "loss": 1.1917, + "theoretical_loss": 3.346810082510468, + "tokens_seen": 3028615168 + }, + { + "epoch": 0.92, + "learning_rate": 6.668004012036109e-05, + "loss": 1.2228, + "theoretical_loss": 3.346804572350449, + "tokens_seen": 3028680704 + }, + { + "epoch": 0.92, + "learning_rate": 6.666399197592779e-05, + "loss": 1.2672, + "theoretical_loss": 3.3467990623430444, + "tokens_seen": 3028746240 + }, + { + "epoch": 0.92, + "learning_rate": 6.664794383149449e-05, + "loss": 1.267, + "theoretical_loss": 3.346793552488246, + "tokens_seen": 3028811776 + }, + { + "epoch": 0.92, + "learning_rate": 6.663189568706119e-05, + "loss": 1.265, + "theoretical_loss": 3.346788042786047, + "tokens_seen": 3028877312 + }, + { + "epoch": 0.92, + "learning_rate": 6.661584754262789e-05, + "loss": 1.2473, + "theoretical_loss": 3.346782533236439, + "tokens_seen": 3028942848 + }, + { + "epoch": 0.92, + "learning_rate": 6.659979939819459e-05, + "loss": 1.253, + "theoretical_loss": 3.3467770238394157, + "tokens_seen": 3029008384 + }, + { + "epoch": 0.92, + "learning_rate": 6.658375125376129e-05, + "loss": 1.3098, + "theoretical_loss": 3.3467715145949684, + "tokens_seen": 3029073920 + }, + { + "epoch": 0.92, + "learning_rate": 6.656770310932798e-05, + "loss": 1.3289, + "theoretical_loss": 3.34676600550309, + "tokens_seen": 3029139456 + }, + { + "epoch": 0.92, + "learning_rate": 6.655165496489468e-05, + "loss": 1.3453, + "theoretical_loss": 3.346760496563773, + "tokens_seen": 3029204992 + }, + { + "epoch": 0.92, + "learning_rate": 6.653560682046138e-05, + "loss": 1.202, + "theoretical_loss": 3.3467549877770097, + "tokens_seen": 3029270528 + }, + { + "epoch": 0.92, + "learning_rate": 6.651955867602808e-05, + "loss": 1.3858, + "theoretical_loss": 3.346749479142793, + "tokens_seen": 3029336064 + }, + { + "epoch": 0.92, + "learning_rate": 6.650351053159479e-05, + "loss": 1.2822, + "theoretical_loss": 3.346743970661115, + "tokens_seen": 3029401600 + }, + { + "epoch": 0.92, + "learning_rate": 6.64874623871615e-05, + "loss": 1.1512, + "theoretical_loss": 3.3467384623319685, + "tokens_seen": 3029467136 + }, + { + "epoch": 0.92, + "learning_rate": 6.64714142427282e-05, + "loss": 1.2368, + "theoretical_loss": 3.3467329541553457, + "tokens_seen": 3029532672 + }, + { + "epoch": 0.92, + "learning_rate": 6.64553660982949e-05, + "loss": 1.3303, + "theoretical_loss": 3.346727446131239, + "tokens_seen": 3029598208 + }, + { + "epoch": 0.92, + "learning_rate": 6.64393179538616e-05, + "loss": 1.3293, + "theoretical_loss": 3.346721938259641, + "tokens_seen": 3029663744 + }, + { + "epoch": 0.92, + "learning_rate": 6.64232698094283e-05, + "loss": 1.1509, + "theoretical_loss": 3.3467164305405444, + "tokens_seen": 3029729280 + }, + { + "epoch": 0.92, + "learning_rate": 6.640722166499498e-05, + "loss": 1.1878, + "theoretical_loss": 3.346710922973941, + "tokens_seen": 3029794816 + }, + { + "epoch": 0.92, + "learning_rate": 6.639117352056168e-05, + "loss": 1.2794, + "theoretical_loss": 3.3467054155598244, + "tokens_seen": 3029860352 + }, + { + "epoch": 0.92, + "learning_rate": 6.637512537612838e-05, + "loss": 1.2099, + "theoretical_loss": 3.346699908298186, + "tokens_seen": 3029925888 + }, + { + "epoch": 0.92, + "learning_rate": 6.635907723169509e-05, + "loss": 1.2572, + "theoretical_loss": 3.346694401189019, + "tokens_seen": 3029991424 + }, + { + "epoch": 0.92, + "learning_rate": 6.634302908726179e-05, + "loss": 1.2134, + "theoretical_loss": 3.346688894232315, + "tokens_seen": 3030056960 + }, + { + "epoch": 0.92, + "learning_rate": 6.632698094282849e-05, + "loss": 1.2947, + "theoretical_loss": 3.346683387428068, + "tokens_seen": 3030122496 + }, + { + "epoch": 0.92, + "learning_rate": 6.631093279839519e-05, + "loss": 1.2447, + "theoretical_loss": 3.3466778807762694, + "tokens_seen": 3030188032 + }, + { + "epoch": 0.92, + "learning_rate": 6.629488465396189e-05, + "loss": 1.2314, + "theoretical_loss": 3.346672374276911, + "tokens_seen": 3030253568 + }, + { + "epoch": 0.92, + "learning_rate": 6.627883650952859e-05, + "loss": 1.2022, + "theoretical_loss": 3.346666867929987, + "tokens_seen": 3030319104 + }, + { + "epoch": 0.92, + "learning_rate": 6.626278836509529e-05, + "loss": 1.2761, + "theoretical_loss": 3.346661361735489, + "tokens_seen": 3030384640 + }, + { + "epoch": 0.92, + "learning_rate": 6.624674022066199e-05, + "loss": 1.3291, + "theoretical_loss": 3.3466558556934087, + "tokens_seen": 3030450176 + }, + { + "epoch": 0.92, + "learning_rate": 6.623069207622869e-05, + "loss": 1.2623, + "theoretical_loss": 3.34665034980374, + "tokens_seen": 3030515712 + }, + { + "epoch": 0.92, + "learning_rate": 6.621464393179539e-05, + "loss": 1.1917, + "theoretical_loss": 3.346644844066475, + "tokens_seen": 3030581248 + }, + { + "epoch": 0.92, + "learning_rate": 6.619859578736209e-05, + "loss": 1.1881, + "theoretical_loss": 3.346639338481605, + "tokens_seen": 3030646784 + }, + { + "epoch": 0.92, + "learning_rate": 6.618254764292879e-05, + "loss": 1.2367, + "theoretical_loss": 3.3466338330491245, + "tokens_seen": 3030712320 + }, + { + "epoch": 0.92, + "learning_rate": 6.616649949849549e-05, + "loss": 1.3523, + "theoretical_loss": 3.3466283277690243, + "tokens_seen": 3030777856 + }, + { + "epoch": 0.92, + "learning_rate": 6.61504513540622e-05, + "loss": 1.3193, + "theoretical_loss": 3.3466228226412977, + "tokens_seen": 3030843392 + }, + { + "epoch": 0.92, + "learning_rate": 6.61344032096289e-05, + "loss": 1.2567, + "theoretical_loss": 3.3466173176659373, + "tokens_seen": 3030908928 + }, + { + "epoch": 0.92, + "learning_rate": 6.61183550651956e-05, + "loss": 1.2691, + "theoretical_loss": 3.346611812842935, + "tokens_seen": 3030974464 + }, + { + "epoch": 0.92, + "objective/train/docs_used": 1101695, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.1832518577575684, + "objective/train/theoretical_loss": 3.3466063081722837, + "objective/train/tokens_used": 3051500000, + "theoretical_loss": 3.3466063081722837, + "tokens_seen": 3031040000 + }, + { + "epoch": 0.92, + "learning_rate": 6.61023069207623e-05, + "loss": 1.2438, + "theoretical_loss": 3.3466063081722837, + "tokens_seen": 3031040000 + }, + { + "epoch": 0.92, + "learning_rate": 6.6086258776329e-05, + "loss": 1.2548, + "theoretical_loss": 3.3466008036539754, + "tokens_seen": 3031105536 + }, + { + "epoch": 0.92, + "learning_rate": 6.60702106318957e-05, + "loss": 1.3746, + "theoretical_loss": 3.3465952992880035, + "tokens_seen": 3031171072 + }, + { + "epoch": 0.92, + "learning_rate": 6.605416248746238e-05, + "loss": 1.2305, + "theoretical_loss": 3.3465897950743595, + "tokens_seen": 3031236608 + }, + { + "epoch": 0.92, + "learning_rate": 6.603811434302908e-05, + "loss": 1.1819, + "theoretical_loss": 3.3465842910130363, + "tokens_seen": 3031302144 + }, + { + "epoch": 0.92, + "learning_rate": 6.602206619859578e-05, + "loss": 1.4145, + "theoretical_loss": 3.346578787104027, + "tokens_seen": 3031367680 + }, + { + "epoch": 0.92, + "learning_rate": 6.600601805416249e-05, + "loss": 1.2828, + "theoretical_loss": 3.3465732833473227, + "tokens_seen": 3031433216 + }, + { + "epoch": 0.92, + "learning_rate": 6.598996990972919e-05, + "loss": 1.2775, + "theoretical_loss": 3.3465677797429176, + "tokens_seen": 3031498752 + }, + { + "epoch": 0.92, + "learning_rate": 6.597392176529589e-05, + "loss": 1.2568, + "theoretical_loss": 3.346562276290803, + "tokens_seen": 3031564288 + }, + { + "epoch": 0.92, + "learning_rate": 6.595787362086259e-05, + "loss": 1.2265, + "theoretical_loss": 3.3465567729909713, + "tokens_seen": 3031629824 + }, + { + "epoch": 0.92, + "learning_rate": 6.594182547642929e-05, + "loss": 1.2492, + "theoretical_loss": 3.346551269843416, + "tokens_seen": 3031695360 + }, + { + "epoch": 0.92, + "learning_rate": 6.592577733199599e-05, + "loss": 1.2552, + "theoretical_loss": 3.346545766848129, + "tokens_seen": 3031760896 + }, + { + "epoch": 0.92, + "learning_rate": 6.59097291875627e-05, + "loss": 1.3424, + "theoretical_loss": 3.3465402640051023, + "tokens_seen": 3031826432 + }, + { + "epoch": 0.92, + "learning_rate": 6.589368104312939e-05, + "loss": 1.2135, + "theoretical_loss": 3.346534761314329, + "tokens_seen": 3031891968 + }, + { + "epoch": 0.92, + "learning_rate": 6.587763289869609e-05, + "loss": 1.2511, + "theoretical_loss": 3.3465292587758015, + "tokens_seen": 3031957504 + }, + { + "epoch": 0.92, + "learning_rate": 6.586158475426279e-05, + "loss": 1.2081, + "theoretical_loss": 3.3465237563895123, + "tokens_seen": 3032023040 + }, + { + "epoch": 0.92, + "learning_rate": 6.584553660982949e-05, + "loss": 1.2604, + "theoretical_loss": 3.346518254155454, + "tokens_seen": 3032088576 + }, + { + "epoch": 0.92, + "learning_rate": 6.582948846539619e-05, + "loss": 1.3277, + "theoretical_loss": 3.346512752073619, + "tokens_seen": 3032154112 + }, + { + "epoch": 0.92, + "learning_rate": 6.581344032096289e-05, + "loss": 1.2666, + "theoretical_loss": 3.3465072501439996, + "tokens_seen": 3032219648 + }, + { + "epoch": 0.92, + "learning_rate": 6.57973921765296e-05, + "loss": 1.2579, + "theoretical_loss": 3.3465017483665886, + "tokens_seen": 3032285184 + }, + { + "epoch": 0.92, + "learning_rate": 6.57813440320963e-05, + "loss": 1.2142, + "theoretical_loss": 3.3464962467413786, + "tokens_seen": 3032350720 + }, + { + "epoch": 0.92, + "learning_rate": 6.5765295887663e-05, + "loss": 1.1665, + "theoretical_loss": 3.3464907452683614, + "tokens_seen": 3032416256 + }, + { + "epoch": 0.92, + "learning_rate": 6.57492477432297e-05, + "loss": 1.3539, + "theoretical_loss": 3.34648524394753, + "tokens_seen": 3032481792 + }, + { + "epoch": 0.92, + "learning_rate": 6.57331995987964e-05, + "loss": 1.2661, + "theoretical_loss": 3.3464797427788775, + "tokens_seen": 3032547328 + }, + { + "epoch": 0.92, + "learning_rate": 6.57171514543631e-05, + "loss": 1.2727, + "theoretical_loss": 3.346474241762395, + "tokens_seen": 3032612864 + }, + { + "epoch": 0.92, + "learning_rate": 6.570110330992978e-05, + "loss": 1.3246, + "theoretical_loss": 3.3464687408980764, + "tokens_seen": 3032678400 + }, + { + "epoch": 0.92, + "learning_rate": 6.568505516549648e-05, + "loss": 1.3004, + "theoretical_loss": 3.3464632401859133, + "tokens_seen": 3032743936 + }, + { + "epoch": 0.92, + "learning_rate": 6.566900702106319e-05, + "loss": 1.2496, + "theoretical_loss": 3.3464577396258983, + "tokens_seen": 3032809472 + }, + { + "epoch": 0.92, + "learning_rate": 6.565295887662989e-05, + "loss": 1.3282, + "theoretical_loss": 3.3464522392180247, + "tokens_seen": 3032875008 + }, + { + "epoch": 0.92, + "learning_rate": 6.563691073219659e-05, + "loss": 1.3088, + "theoretical_loss": 3.346446738962284, + "tokens_seen": 3032940544 + }, + { + "epoch": 0.92, + "learning_rate": 6.562086258776329e-05, + "loss": 1.313, + "theoretical_loss": 3.3464412388586693, + "tokens_seen": 3033006080 + }, + { + "epoch": 0.92, + "learning_rate": 6.560481444333e-05, + "loss": 1.2506, + "theoretical_loss": 3.3464357389071724, + "tokens_seen": 3033071616 + }, + { + "epoch": 0.92, + "learning_rate": 6.55887662988967e-05, + "loss": 1.2654, + "theoretical_loss": 3.3464302391077867, + "tokens_seen": 3033137152 + }, + { + "epoch": 0.92, + "learning_rate": 6.55727181544634e-05, + "loss": 1.2806, + "theoretical_loss": 3.3464247394605042, + "tokens_seen": 3033202688 + }, + { + "epoch": 0.92, + "learning_rate": 6.55566700100301e-05, + "loss": 1.2467, + "theoretical_loss": 3.3464192399653174, + "tokens_seen": 3033268224 + }, + { + "epoch": 0.92, + "learning_rate": 6.554062186559679e-05, + "loss": 1.2725, + "theoretical_loss": 3.346413740622219, + "tokens_seen": 3033333760 + }, + { + "epoch": 0.92, + "learning_rate": 6.552457372116349e-05, + "loss": 1.2193, + "theoretical_loss": 3.346408241431202, + "tokens_seen": 3033399296 + }, + { + "epoch": 0.92, + "learning_rate": 6.550852557673019e-05, + "loss": 1.3534, + "theoretical_loss": 3.3464027423922578, + "tokens_seen": 3033464832 + }, + { + "epoch": 0.92, + "learning_rate": 6.549247743229689e-05, + "loss": 1.2775, + "theoretical_loss": 3.3463972435053795, + "tokens_seen": 3033530368 + }, + { + "epoch": 0.92, + "learning_rate": 6.547642928786359e-05, + "loss": 1.2424, + "theoretical_loss": 3.3463917447705596, + "tokens_seen": 3033595904 + }, + { + "epoch": 0.92, + "learning_rate": 6.54603811434303e-05, + "loss": 1.3802, + "theoretical_loss": 3.346386246187791, + "tokens_seen": 3033661440 + }, + { + "epoch": 0.92, + "learning_rate": 6.5444332998997e-05, + "loss": 1.2266, + "theoretical_loss": 3.346380747757065, + "tokens_seen": 3033726976 + }, + { + "epoch": 0.92, + "learning_rate": 6.54282848545637e-05, + "loss": 1.1834, + "theoretical_loss": 3.346375249478376, + "tokens_seen": 3033792512 + }, + { + "epoch": 0.92, + "learning_rate": 6.54122367101304e-05, + "loss": 1.2616, + "theoretical_loss": 3.3463697513517143, + "tokens_seen": 3033858048 + }, + { + "epoch": 0.92, + "learning_rate": 6.53961885656971e-05, + "loss": 1.2537, + "theoretical_loss": 3.346364253377074, + "tokens_seen": 3033923584 + }, + { + "epoch": 0.92, + "learning_rate": 6.53801404212638e-05, + "loss": 1.3396, + "theoretical_loss": 3.346358755554447, + "tokens_seen": 3033989120 + }, + { + "epoch": 0.92, + "learning_rate": 6.53640922768305e-05, + "loss": 1.4084, + "theoretical_loss": 3.346353257883826, + "tokens_seen": 3034054656 + }, + { + "epoch": 0.92, + "learning_rate": 6.534804413239718e-05, + "loss": 1.3244, + "theoretical_loss": 3.3463477603652034, + "tokens_seen": 3034120192 + }, + { + "epoch": 0.92, + "learning_rate": 6.533199598796389e-05, + "loss": 1.2764, + "theoretical_loss": 3.346342262998572, + "tokens_seen": 3034185728 + }, + { + "epoch": 0.92, + "learning_rate": 6.531594784353059e-05, + "loss": 1.2703, + "theoretical_loss": 3.3463367657839242, + "tokens_seen": 3034251264 + }, + { + "epoch": 0.92, + "objective/train/docs_used": 1102938, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.3146312236785889, + "objective/train/theoretical_loss": 3.3463312687212525, + "objective/train/tokens_used": 3054776800, + "theoretical_loss": 3.3463312687212525, + "tokens_seen": 3034316800 + }, + { + "epoch": 0.92, + "learning_rate": 6.52998996990973e-05, + "loss": 1.2607, + "theoretical_loss": 3.3463312687212525, + "tokens_seen": 3034316800 + }, + { + "epoch": 0.92, + "learning_rate": 6.5283851554664e-05, + "loss": 1.2596, + "theoretical_loss": 3.3463257718105486, + "tokens_seen": 3034382336 + }, + { + "epoch": 0.92, + "learning_rate": 6.52678034102307e-05, + "loss": 1.3821, + "theoretical_loss": 3.3463202750518066, + "tokens_seen": 3034447872 + }, + { + "epoch": 0.92, + "learning_rate": 6.52517552657974e-05, + "loss": 1.1723, + "theoretical_loss": 3.3463147784450173, + "tokens_seen": 3034513408 + }, + { + "epoch": 0.92, + "learning_rate": 6.52357071213641e-05, + "loss": 1.2756, + "theoretical_loss": 3.346309281990175, + "tokens_seen": 3034578944 + }, + { + "epoch": 0.92, + "learning_rate": 6.52196589769308e-05, + "loss": 1.2271, + "theoretical_loss": 3.3463037856872706, + "tokens_seen": 3034644480 + }, + { + "epoch": 0.92, + "learning_rate": 6.52036108324975e-05, + "loss": 1.3038, + "theoretical_loss": 3.3462982895362976, + "tokens_seen": 3034710016 + }, + { + "epoch": 0.92, + "learning_rate": 6.518756268806419e-05, + "loss": 1.2213, + "theoretical_loss": 3.3462927935372484, + "tokens_seen": 3034775552 + }, + { + "epoch": 0.92, + "learning_rate": 6.517151454363089e-05, + "loss": 1.2842, + "theoretical_loss": 3.346287297690115, + "tokens_seen": 3034841088 + }, + { + "epoch": 0.92, + "learning_rate": 6.515546639919759e-05, + "loss": 1.1574, + "theoretical_loss": 3.3462818019948903, + "tokens_seen": 3034906624 + }, + { + "epoch": 0.92, + "learning_rate": 6.513941825476429e-05, + "loss": 1.2618, + "theoretical_loss": 3.3462763064515673, + "tokens_seen": 3034972160 + }, + { + "epoch": 0.92, + "learning_rate": 6.5123370110331e-05, + "loss": 1.1583, + "theoretical_loss": 3.3462708110601374, + "tokens_seen": 3035037696 + }, + { + "epoch": 0.92, + "learning_rate": 6.51073219658977e-05, + "loss": 1.231, + "theoretical_loss": 3.3462653158205944, + "tokens_seen": 3035103232 + }, + { + "epoch": 0.92, + "learning_rate": 6.50912738214644e-05, + "loss": 1.1941, + "theoretical_loss": 3.3462598207329295, + "tokens_seen": 3035168768 + }, + { + "epoch": 0.92, + "learning_rate": 6.50752256770311e-05, + "loss": 1.2815, + "theoretical_loss": 3.346254325797136, + "tokens_seen": 3035234304 + }, + { + "epoch": 0.92, + "learning_rate": 6.50591775325978e-05, + "loss": 1.33, + "theoretical_loss": 3.346248831013207, + "tokens_seen": 3035299840 + }, + { + "epoch": 0.92, + "learning_rate": 6.50431293881645e-05, + "loss": 1.2433, + "theoretical_loss": 3.3462433363811335, + "tokens_seen": 3035365376 + }, + { + "epoch": 0.92, + "learning_rate": 6.50270812437312e-05, + "loss": 1.2301, + "theoretical_loss": 3.3462378419009093, + "tokens_seen": 3035430912 + }, + { + "epoch": 0.92, + "learning_rate": 6.50110330992979e-05, + "loss": 1.2743, + "theoretical_loss": 3.346232347572526, + "tokens_seen": 3035496448 + }, + { + "epoch": 0.92, + "learning_rate": 6.49949849548646e-05, + "loss": 1.3058, + "theoretical_loss": 3.346226853395977, + "tokens_seen": 3035561984 + }, + { + "epoch": 0.92, + "learning_rate": 6.49789368104313e-05, + "loss": 1.301, + "theoretical_loss": 3.3462213593712544, + "tokens_seen": 3035627520 + }, + { + "epoch": 0.92, + "learning_rate": 6.4962888665998e-05, + "loss": 1.2796, + "theoretical_loss": 3.346215865498351, + "tokens_seen": 3035693056 + }, + { + "epoch": 0.92, + "learning_rate": 6.49468405215647e-05, + "loss": 1.1701, + "theoretical_loss": 3.346210371777259, + "tokens_seen": 3035758592 + }, + { + "epoch": 0.92, + "learning_rate": 6.49307923771314e-05, + "loss": 1.3359, + "theoretical_loss": 3.346204878207971, + "tokens_seen": 3035824128 + }, + { + "epoch": 0.92, + "learning_rate": 6.49147442326981e-05, + "loss": 1.297, + "theoretical_loss": 3.3461993847904794, + "tokens_seen": 3035889664 + }, + { + "epoch": 0.92, + "learning_rate": 6.48986960882648e-05, + "loss": 1.2255, + "theoretical_loss": 3.346193891524777, + "tokens_seen": 3035955200 + }, + { + "epoch": 0.92, + "learning_rate": 6.48826479438315e-05, + "loss": 1.2244, + "theoretical_loss": 3.346188398410856, + "tokens_seen": 3036020736 + }, + { + "epoch": 0.92, + "learning_rate": 6.48665997993982e-05, + "loss": 1.2654, + "theoretical_loss": 3.346182905448709, + "tokens_seen": 3036086272 + }, + { + "epoch": 0.92, + "learning_rate": 6.485055165496489e-05, + "loss": 1.1518, + "theoretical_loss": 3.3461774126383292, + "tokens_seen": 3036151808 + }, + { + "epoch": 0.92, + "learning_rate": 6.483450351053159e-05, + "loss": 1.1662, + "theoretical_loss": 3.3461719199797084, + "tokens_seen": 3036217344 + }, + { + "epoch": 0.92, + "learning_rate": 6.481845536609829e-05, + "loss": 1.302, + "theoretical_loss": 3.346166427472839, + "tokens_seen": 3036282880 + }, + { + "epoch": 0.92, + "learning_rate": 6.480240722166499e-05, + "loss": 1.2098, + "theoretical_loss": 3.346160935117714, + "tokens_seen": 3036348416 + }, + { + "epoch": 0.92, + "learning_rate": 6.478635907723169e-05, + "loss": 1.3332, + "theoretical_loss": 3.346155442914326, + "tokens_seen": 3036413952 + }, + { + "epoch": 0.92, + "learning_rate": 6.47703109327984e-05, + "loss": 1.3026, + "theoretical_loss": 3.3461499508626673, + "tokens_seen": 3036479488 + }, + { + "epoch": 0.92, + "learning_rate": 6.47542627883651e-05, + "loss": 1.1941, + "theoretical_loss": 3.3461444589627307, + "tokens_seen": 3036545024 + }, + { + "epoch": 0.92, + "learning_rate": 6.47382146439318e-05, + "loss": 1.3374, + "theoretical_loss": 3.3461389672145083, + "tokens_seen": 3036610560 + }, + { + "epoch": 0.92, + "learning_rate": 6.47221664994985e-05, + "loss": 1.1673, + "theoretical_loss": 3.3461334756179926, + "tokens_seen": 3036676096 + }, + { + "epoch": 0.92, + "learning_rate": 6.470611835506521e-05, + "loss": 1.168, + "theoretical_loss": 3.3461279841731764, + "tokens_seen": 3036741632 + }, + { + "epoch": 0.92, + "learning_rate": 6.46900702106319e-05, + "loss": 1.3635, + "theoretical_loss": 3.346122492880052, + "tokens_seen": 3036807168 + }, + { + "epoch": 0.92, + "learning_rate": 6.46740220661986e-05, + "loss": 1.2992, + "theoretical_loss": 3.346117001738613, + "tokens_seen": 3036872704 + }, + { + "epoch": 0.92, + "learning_rate": 6.46579739217653e-05, + "loss": 1.3541, + "theoretical_loss": 3.3461115107488504, + "tokens_seen": 3036938240 + }, + { + "epoch": 0.92, + "learning_rate": 6.4641925777332e-05, + "loss": 1.2896, + "theoretical_loss": 3.3461060199107573, + "tokens_seen": 3037003776 + }, + { + "epoch": 0.92, + "learning_rate": 6.46258776328987e-05, + "loss": 1.1689, + "theoretical_loss": 3.3461005292243264, + "tokens_seen": 3037069312 + }, + { + "epoch": 0.92, + "learning_rate": 6.46098294884654e-05, + "loss": 1.3247, + "theoretical_loss": 3.3460950386895507, + "tokens_seen": 3037134848 + }, + { + "epoch": 0.92, + "learning_rate": 6.45937813440321e-05, + "loss": 1.247, + "theoretical_loss": 3.3460895483064217, + "tokens_seen": 3037200384 + }, + { + "epoch": 0.92, + "learning_rate": 6.45777331995988e-05, + "loss": 1.2491, + "theoretical_loss": 3.3460840580749327, + "tokens_seen": 3037265920 + }, + { + "epoch": 0.92, + "learning_rate": 6.45616850551655e-05, + "loss": 1.2472, + "theoretical_loss": 3.346078567995076, + "tokens_seen": 3037331456 + }, + { + "epoch": 0.92, + "learning_rate": 6.45456369107322e-05, + "loss": 1.0795, + "theoretical_loss": 3.346073078066844, + "tokens_seen": 3037396992 + }, + { + "epoch": 0.92, + "learning_rate": 6.45295887662989e-05, + "loss": 1.3289, + "theoretical_loss": 3.3460675882902295, + "tokens_seen": 3037462528 + }, + { + "epoch": 0.92, + "learning_rate": 6.45135406218656e-05, + "loss": 1.2375, + "theoretical_loss": 3.346062098665225, + "tokens_seen": 3037528064 + }, + { + "epoch": 0.92, + "objective/train/docs_used": 1103994, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.3579102754592896, + "objective/train/theoretical_loss": 3.346056609191823, + "objective/train/tokens_used": 3058053600, + "theoretical_loss": 3.346056609191823, + "tokens_seen": 3037593600 + }, + { + "epoch": 0.92, + "learning_rate": 6.449749247743229e-05, + "loss": 1.3542, + "theoretical_loss": 3.346056609191823, + "tokens_seen": 3037593600 + }, + { + "epoch": 0.92, + "learning_rate": 6.448144433299899e-05, + "loss": 1.3032, + "theoretical_loss": 3.3460511198700162, + "tokens_seen": 3037659136 + }, + { + "epoch": 0.92, + "learning_rate": 6.446539618856569e-05, + "loss": 1.1146, + "theoretical_loss": 3.3460456306997965, + "tokens_seen": 3037724672 + }, + { + "epoch": 0.92, + "learning_rate": 6.444934804413239e-05, + "loss": 1.2348, + "theoretical_loss": 3.346040141681157, + "tokens_seen": 3037790208 + }, + { + "epoch": 0.92, + "learning_rate": 6.44332998996991e-05, + "loss": 1.336, + "theoretical_loss": 3.3460346528140903, + "tokens_seen": 3037855744 + }, + { + "epoch": 0.92, + "learning_rate": 6.44172517552658e-05, + "loss": 1.2107, + "theoretical_loss": 3.3460291640985886, + "tokens_seen": 3037921280 + }, + { + "epoch": 0.92, + "learning_rate": 6.440120361083251e-05, + "loss": 1.3563, + "theoretical_loss": 3.346023675534645, + "tokens_seen": 3037986816 + }, + { + "epoch": 0.92, + "learning_rate": 6.438515546639921e-05, + "loss": 1.2089, + "theoretical_loss": 3.3460181871222514, + "tokens_seen": 3038052352 + }, + { + "epoch": 0.92, + "learning_rate": 6.436910732196591e-05, + "loss": 1.1877, + "theoretical_loss": 3.3460126988614007, + "tokens_seen": 3038117888 + }, + { + "epoch": 0.92, + "learning_rate": 6.435305917753261e-05, + "loss": 1.3723, + "theoretical_loss": 3.3460072107520853, + "tokens_seen": 3038183424 + }, + { + "epoch": 0.92, + "learning_rate": 6.43370110330993e-05, + "loss": 1.2512, + "theoretical_loss": 3.346001722794298, + "tokens_seen": 3038248960 + }, + { + "epoch": 0.92, + "learning_rate": 6.4320962888666e-05, + "loss": 1.3389, + "theoretical_loss": 3.345996234988031, + "tokens_seen": 3038314496 + }, + { + "epoch": 0.92, + "learning_rate": 6.43049147442327e-05, + "loss": 1.2039, + "theoretical_loss": 3.345990747333277, + "tokens_seen": 3038380032 + }, + { + "epoch": 0.92, + "learning_rate": 6.42888665997994e-05, + "loss": 1.1994, + "theoretical_loss": 3.3459852598300284, + "tokens_seen": 3038445568 + }, + { + "epoch": 0.92, + "learning_rate": 6.42728184553661e-05, + "loss": 1.2968, + "theoretical_loss": 3.3459797724782785, + "tokens_seen": 3038511104 + }, + { + "epoch": 0.92, + "learning_rate": 6.42567703109328e-05, + "loss": 1.0839, + "theoretical_loss": 3.3459742852780185, + "tokens_seen": 3038576640 + }, + { + "epoch": 0.92, + "learning_rate": 6.42407221664995e-05, + "loss": 1.2259, + "theoretical_loss": 3.3459687982292423, + "tokens_seen": 3038642176 + }, + { + "epoch": 0.92, + "learning_rate": 6.42246740220662e-05, + "loss": 1.1842, + "theoretical_loss": 3.345963311331942, + "tokens_seen": 3038707712 + }, + { + "epoch": 0.92, + "learning_rate": 6.42086258776329e-05, + "loss": 1.2323, + "theoretical_loss": 3.3459578245861095, + "tokens_seen": 3038773248 + }, + { + "epoch": 0.92, + "learning_rate": 6.41925777331996e-05, + "loss": 1.1866, + "theoretical_loss": 3.345952337991738, + "tokens_seen": 3038838784 + }, + { + "epoch": 0.92, + "learning_rate": 6.41765295887663e-05, + "loss": 1.3029, + "theoretical_loss": 3.34594685154882, + "tokens_seen": 3038904320 + }, + { + "epoch": 0.92, + "learning_rate": 6.4160481444333e-05, + "loss": 1.2011, + "theoretical_loss": 3.345941365257348, + "tokens_seen": 3038969856 + }, + { + "epoch": 0.92, + "learning_rate": 6.414443329989969e-05, + "loss": 1.1733, + "theoretical_loss": 3.3459358791173144, + "tokens_seen": 3039035392 + }, + { + "epoch": 0.92, + "learning_rate": 6.412838515546639e-05, + "loss": 1.2011, + "theoretical_loss": 3.345930393128712, + "tokens_seen": 3039100928 + }, + { + "epoch": 0.92, + "learning_rate": 6.41123370110331e-05, + "loss": 1.3023, + "theoretical_loss": 3.3459249072915336, + "tokens_seen": 3039166464 + }, + { + "epoch": 0.92, + "learning_rate": 6.409628886659981e-05, + "loss": 1.1705, + "theoretical_loss": 3.345919421605771, + "tokens_seen": 3039232000 + }, + { + "epoch": 0.92, + "learning_rate": 6.408024072216651e-05, + "loss": 1.2689, + "theoretical_loss": 3.345913936071417, + "tokens_seen": 3039297536 + }, + { + "epoch": 0.92, + "learning_rate": 6.406419257773321e-05, + "loss": 1.2749, + "theoretical_loss": 3.3459084506884644, + "tokens_seen": 3039363072 + }, + { + "epoch": 0.92, + "learning_rate": 6.404814443329991e-05, + "loss": 1.3327, + "theoretical_loss": 3.345902965456906, + "tokens_seen": 3039428608 + }, + { + "epoch": 0.92, + "learning_rate": 6.403209628886661e-05, + "loss": 1.2641, + "theoretical_loss": 3.345897480376734, + "tokens_seen": 3039494144 + }, + { + "epoch": 0.92, + "learning_rate": 6.401604814443331e-05, + "loss": 1.2493, + "theoretical_loss": 3.3458919954479405, + "tokens_seen": 3039559680 + }, + { + "epoch": 0.92, + "learning_rate": 6.400000000000001e-05, + "loss": 1.2659, + "theoretical_loss": 3.345886510670519, + "tokens_seen": 3039625216 + }, + { + "epoch": 0.92, + "learning_rate": 6.39839518555667e-05, + "loss": 1.275, + "theoretical_loss": 3.345881026044461, + "tokens_seen": 3039690752 + }, + { + "epoch": 0.92, + "learning_rate": 6.39679037111334e-05, + "loss": 1.2451, + "theoretical_loss": 3.34587554156976, + "tokens_seen": 3039756288 + }, + { + "epoch": 0.92, + "learning_rate": 6.39518555667001e-05, + "loss": 1.287, + "theoretical_loss": 3.3458700572464086, + "tokens_seen": 3039821824 + }, + { + "epoch": 0.92, + "learning_rate": 6.39358074222668e-05, + "loss": 1.1806, + "theoretical_loss": 3.3458645730743988, + "tokens_seen": 3039887360 + }, + { + "epoch": 0.92, + "learning_rate": 6.39197592778335e-05, + "loss": 1.2406, + "theoretical_loss": 3.345859089053723, + "tokens_seen": 3039952896 + }, + { + "epoch": 0.92, + "learning_rate": 6.39037111334002e-05, + "loss": 1.3014, + "theoretical_loss": 3.345853605184374, + "tokens_seen": 3040018432 + }, + { + "epoch": 0.92, + "learning_rate": 6.38876629889669e-05, + "loss": 1.2353, + "theoretical_loss": 3.345848121466345, + "tokens_seen": 3040083968 + }, + { + "epoch": 0.92, + "learning_rate": 6.38716148445336e-05, + "loss": 1.2803, + "theoretical_loss": 3.345842637899628, + "tokens_seen": 3040149504 + }, + { + "epoch": 0.92, + "learning_rate": 6.38555667001003e-05, + "loss": 1.1852, + "theoretical_loss": 3.345837154484215, + "tokens_seen": 3040215040 + }, + { + "epoch": 0.92, + "learning_rate": 6.3839518555667e-05, + "loss": 1.1795, + "theoretical_loss": 3.3458316712201, + "tokens_seen": 3040280576 + }, + { + "epoch": 0.92, + "learning_rate": 6.38234704112337e-05, + "loss": 1.2853, + "theoretical_loss": 3.345826188107274, + "tokens_seen": 3040346112 + }, + { + "epoch": 0.92, + "learning_rate": 6.38074222668004e-05, + "loss": 1.3242, + "theoretical_loss": 3.3458207051457305, + "tokens_seen": 3040411648 + }, + { + "epoch": 0.92, + "learning_rate": 6.37913741223671e-05, + "loss": 1.2392, + "theoretical_loss": 3.345815222335462, + "tokens_seen": 3040477184 + }, + { + "epoch": 0.92, + "learning_rate": 6.37753259779338e-05, + "loss": 1.2206, + "theoretical_loss": 3.3458097396764606, + "tokens_seen": 3040542720 + }, + { + "epoch": 0.92, + "learning_rate": 6.37592778335005e-05, + "loss": 1.3108, + "theoretical_loss": 3.3458042571687194, + "tokens_seen": 3040608256 + }, + { + "epoch": 0.92, + "learning_rate": 6.374322968906721e-05, + "loss": 1.2097, + "theoretical_loss": 3.3457987748122306, + "tokens_seen": 3040673792 + }, + { + "epoch": 0.92, + "learning_rate": 6.372718154463391e-05, + "loss": 1.2256, + "theoretical_loss": 3.345793292606987, + "tokens_seen": 3040739328 + }, + { + "epoch": 0.92, + "learning_rate": 6.371113340020061e-05, + "loss": 1.3164, + "theoretical_loss": 3.3457878105529812, + "tokens_seen": 3040804864 + }, + { + "epoch": 0.92, + "objective/train/docs_used": 1105174, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.4151077270507812, + "objective/train/theoretical_loss": 3.3457823286502055, + "objective/train/tokens_used": 3061330400, + "theoretical_loss": 3.3457823286502055, + "tokens_seen": 3040870400 + }, + { + "epoch": 0.92, + "learning_rate": 6.369508525576731e-05, + "loss": 1.3605, + "theoretical_loss": 3.3457823286502055, + "tokens_seen": 3040870400 + }, + { + "epoch": 0.92, + "learning_rate": 6.367903711133401e-05, + "loss": 1.2359, + "theoretical_loss": 3.345776846898653, + "tokens_seen": 3040935936 + }, + { + "epoch": 0.92, + "learning_rate": 6.366298896690071e-05, + "loss": 1.196, + "theoretical_loss": 3.345771365298315, + "tokens_seen": 3041001472 + }, + { + "epoch": 0.92, + "learning_rate": 6.364694082246741e-05, + "loss": 1.2335, + "theoretical_loss": 3.345765883849186, + "tokens_seen": 3041067008 + }, + { + "epoch": 0.92, + "learning_rate": 6.36308926780341e-05, + "loss": 1.2754, + "theoretical_loss": 3.345760402551257, + "tokens_seen": 3041132544 + }, + { + "epoch": 0.92, + "learning_rate": 6.36148445336008e-05, + "loss": 1.211, + "theoretical_loss": 3.345754921404521, + "tokens_seen": 3041198080 + }, + { + "epoch": 0.92, + "learning_rate": 6.35987963891675e-05, + "loss": 1.1969, + "theoretical_loss": 3.345749440408971, + "tokens_seen": 3041263616 + }, + { + "epoch": 0.92, + "learning_rate": 6.35827482447342e-05, + "loss": 1.1696, + "theoretical_loss": 3.345743959564599, + "tokens_seen": 3041329152 + }, + { + "epoch": 0.92, + "learning_rate": 6.35667001003009e-05, + "loss": 1.2188, + "theoretical_loss": 3.345738478871398, + "tokens_seen": 3041394688 + }, + { + "epoch": 0.92, + "learning_rate": 6.35506519558676e-05, + "loss": 1.1718, + "theoretical_loss": 3.34573299832936, + "tokens_seen": 3041460224 + }, + { + "epoch": 0.92, + "learning_rate": 6.35346038114343e-05, + "loss": 1.267, + "theoretical_loss": 3.3457275179384784, + "tokens_seen": 3041525760 + }, + { + "epoch": 0.92, + "learning_rate": 6.3518555667001e-05, + "loss": 1.1845, + "theoretical_loss": 3.345722037698745, + "tokens_seen": 3041591296 + }, + { + "epoch": 0.92, + "learning_rate": 6.350250752256772e-05, + "loss": 1.203, + "theoretical_loss": 3.345716557610153, + "tokens_seen": 3041656832 + }, + { + "epoch": 0.92, + "learning_rate": 6.348645937813442e-05, + "loss": 1.229, + "theoretical_loss": 3.3457110776726946, + "tokens_seen": 3041722368 + }, + { + "epoch": 0.92, + "learning_rate": 6.34704112337011e-05, + "loss": 1.25, + "theoretical_loss": 3.3457055978863623, + "tokens_seen": 3041787904 + }, + { + "epoch": 0.92, + "learning_rate": 6.34543630892678e-05, + "loss": 1.1943, + "theoretical_loss": 3.345700118251149, + "tokens_seen": 3041853440 + }, + { + "epoch": 0.92, + "learning_rate": 6.34383149448345e-05, + "loss": 1.3544, + "theoretical_loss": 3.345694638767047, + "tokens_seen": 3041918976 + }, + { + "epoch": 0.92, + "learning_rate": 6.34222668004012e-05, + "loss": 1.2354, + "theoretical_loss": 3.345689159434049, + "tokens_seen": 3041984512 + }, + { + "epoch": 0.92, + "learning_rate": 6.340621865596791e-05, + "loss": 1.2864, + "theoretical_loss": 3.345683680252148, + "tokens_seen": 3042050048 + }, + { + "epoch": 0.92, + "learning_rate": 6.339017051153461e-05, + "loss": 1.2806, + "theoretical_loss": 3.345678201221336, + "tokens_seen": 3042115584 + }, + { + "epoch": 0.92, + "learning_rate": 6.337412236710131e-05, + "loss": 1.2153, + "theoretical_loss": 3.345672722341605, + "tokens_seen": 3042181120 + }, + { + "epoch": 0.92, + "learning_rate": 6.335807422266801e-05, + "loss": 1.1576, + "theoretical_loss": 3.345667243612949, + "tokens_seen": 3042246656 + }, + { + "epoch": 0.92, + "learning_rate": 6.334202607823471e-05, + "loss": 1.2176, + "theoretical_loss": 3.3456617650353597, + "tokens_seen": 3042312192 + }, + { + "epoch": 0.92, + "learning_rate": 6.332597793380141e-05, + "loss": 1.294, + "theoretical_loss": 3.34565628660883, + "tokens_seen": 3042377728 + }, + { + "epoch": 0.92, + "learning_rate": 6.330992978936811e-05, + "loss": 1.2696, + "theoretical_loss": 3.3456508083333523, + "tokens_seen": 3042443264 + }, + { + "epoch": 0.92, + "learning_rate": 6.329388164493481e-05, + "loss": 1.1848, + "theoretical_loss": 3.345645330208919, + "tokens_seen": 3042508800 + }, + { + "epoch": 0.92, + "learning_rate": 6.32778335005015e-05, + "loss": 1.3222, + "theoretical_loss": 3.3456398522355233, + "tokens_seen": 3042574336 + }, + { + "epoch": 0.92, + "learning_rate": 6.32617853560682e-05, + "loss": 1.3326, + "theoretical_loss": 3.345634374413157, + "tokens_seen": 3042639872 + }, + { + "epoch": 0.92, + "learning_rate": 6.32457372116349e-05, + "loss": 1.2322, + "theoretical_loss": 3.3456288967418133, + "tokens_seen": 3042705408 + }, + { + "epoch": 0.92, + "learning_rate": 6.32296890672016e-05, + "loss": 1.2472, + "theoretical_loss": 3.3456234192214844, + "tokens_seen": 3042770944 + }, + { + "epoch": 0.92, + "learning_rate": 6.32136409227683e-05, + "loss": 1.2782, + "theoretical_loss": 3.345617941852163, + "tokens_seen": 3042836480 + }, + { + "epoch": 0.92, + "learning_rate": 6.319759277833502e-05, + "loss": 1.2546, + "theoretical_loss": 3.345612464633842, + "tokens_seen": 3042902016 + }, + { + "epoch": 0.92, + "learning_rate": 6.318154463390172e-05, + "loss": 1.3079, + "theoretical_loss": 3.345606987566514, + "tokens_seen": 3042967552 + }, + { + "epoch": 0.92, + "learning_rate": 6.316549648946842e-05, + "loss": 1.2457, + "theoretical_loss": 3.345601510650171, + "tokens_seen": 3043033088 + }, + { + "epoch": 0.92, + "learning_rate": 6.314944834503512e-05, + "loss": 1.2982, + "theoretical_loss": 3.3455960338848056, + "tokens_seen": 3043098624 + }, + { + "epoch": 0.92, + "learning_rate": 6.313340020060182e-05, + "loss": 1.2231, + "theoretical_loss": 3.3455905572704108, + "tokens_seen": 3043164160 + }, + { + "epoch": 0.92, + "learning_rate": 6.31173520561685e-05, + "loss": 1.2753, + "theoretical_loss": 3.345585080806979, + "tokens_seen": 3043229696 + }, + { + "epoch": 0.92, + "learning_rate": 6.31013039117352e-05, + "loss": 1.3561, + "theoretical_loss": 3.345579604494503, + "tokens_seen": 3043295232 + }, + { + "epoch": 0.92, + "learning_rate": 6.30852557673019e-05, + "loss": 1.2081, + "theoretical_loss": 3.3455741283329754, + "tokens_seen": 3043360768 + }, + { + "epoch": 0.92, + "learning_rate": 6.306920762286861e-05, + "loss": 1.2073, + "theoretical_loss": 3.345568652322388, + "tokens_seen": 3043426304 + }, + { + "epoch": 0.92, + "learning_rate": 6.305315947843531e-05, + "loss": 1.2626, + "theoretical_loss": 3.345563176462735, + "tokens_seen": 3043491840 + }, + { + "epoch": 0.92, + "learning_rate": 6.303711133400201e-05, + "loss": 1.2273, + "theoretical_loss": 3.345557700754007, + "tokens_seen": 3043557376 + }, + { + "epoch": 0.92, + "learning_rate": 6.302106318956871e-05, + "loss": 1.1392, + "theoretical_loss": 3.345552225196198, + "tokens_seen": 3043622912 + }, + { + "epoch": 0.92, + "learning_rate": 6.300501504513541e-05, + "loss": 1.1971, + "theoretical_loss": 3.3455467497893006, + "tokens_seen": 3043688448 + }, + { + "epoch": 0.92, + "learning_rate": 6.298896690070211e-05, + "loss": 1.1833, + "theoretical_loss": 3.3455412745333066, + "tokens_seen": 3043753984 + }, + { + "epoch": 0.92, + "learning_rate": 6.297291875626881e-05, + "loss": 1.299, + "theoretical_loss": 3.345535799428209, + "tokens_seen": 3043819520 + }, + { + "epoch": 0.92, + "learning_rate": 6.295687061183551e-05, + "loss": 1.185, + "theoretical_loss": 3.3455303244740007, + "tokens_seen": 3043885056 + }, + { + "epoch": 0.92, + "learning_rate": 6.294082246740221e-05, + "loss": 1.236, + "theoretical_loss": 3.3455248496706735, + "tokens_seen": 3043950592 + }, + { + "epoch": 0.92, + "learning_rate": 6.29247743229689e-05, + "loss": 1.1814, + "theoretical_loss": 3.3455193750182204, + "tokens_seen": 3044016128 + }, + { + "epoch": 0.92, + "learning_rate": 6.290872617853561e-05, + "loss": 1.3075, + "theoretical_loss": 3.345513900516634, + "tokens_seen": 3044081664 + }, + { + "epoch": 0.92, + "objective/train/docs_used": 1106378, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.3095901012420654, + "objective/train/theoretical_loss": 3.3455084261659076, + "objective/train/tokens_used": 3064607200, + "theoretical_loss": 3.3455084261659076, + "tokens_seen": 3044147200 + }, + { + "epoch": 0.92, + "learning_rate": 6.289267803410231e-05, + "loss": 1.2583, + "theoretical_loss": 3.3455084261659076, + "tokens_seen": 3044147200 + }, + { + "epoch": 0.92, + "learning_rate": 6.287662988966901e-05, + "loss": 1.1895, + "theoretical_loss": 3.345502951966033, + "tokens_seen": 3044212736 + }, + { + "epoch": 0.92, + "learning_rate": 6.286058174523572e-05, + "loss": 1.23, + "theoretical_loss": 3.3454974779170024, + "tokens_seen": 3044278272 + }, + { + "epoch": 0.92, + "learning_rate": 6.284453360080242e-05, + "loss": 1.37, + "theoretical_loss": 3.345492004018809, + "tokens_seen": 3044343808 + }, + { + "epoch": 0.92, + "learning_rate": 6.282848545636912e-05, + "loss": 1.2227, + "theoretical_loss": 3.3454865302714456, + "tokens_seen": 3044409344 + }, + { + "epoch": 0.92, + "learning_rate": 6.281243731193582e-05, + "loss": 1.2908, + "theoretical_loss": 3.3454810566749043, + "tokens_seen": 3044474880 + }, + { + "epoch": 0.92, + "learning_rate": 6.279638916750252e-05, + "loss": 1.2073, + "theoretical_loss": 3.3454755832291783, + "tokens_seen": 3044540416 + }, + { + "epoch": 0.92, + "learning_rate": 6.27803410230692e-05, + "loss": 1.2914, + "theoretical_loss": 3.3454701099342596, + "tokens_seen": 3044605952 + }, + { + "epoch": 0.92, + "learning_rate": 6.27642928786359e-05, + "loss": 1.2722, + "theoretical_loss": 3.345464636790141, + "tokens_seen": 3044671488 + }, + { + "epoch": 0.92, + "learning_rate": 6.27482447342026e-05, + "loss": 1.3432, + "theoretical_loss": 3.3454591637968147, + "tokens_seen": 3044737024 + }, + { + "epoch": 0.92, + "learning_rate": 6.27321965897693e-05, + "loss": 1.21, + "theoretical_loss": 3.3454536909542743, + "tokens_seen": 3044802560 + }, + { + "epoch": 0.92, + "learning_rate": 6.271614844533601e-05, + "loss": 1.2961, + "theoretical_loss": 3.3454482182625114, + "tokens_seen": 3044868096 + }, + { + "epoch": 0.92, + "learning_rate": 6.270010030090271e-05, + "loss": 1.2332, + "theoretical_loss": 3.3454427457215195, + "tokens_seen": 3044933632 + }, + { + "epoch": 0.92, + "learning_rate": 6.268405215646941e-05, + "loss": 1.1965, + "theoretical_loss": 3.3454372733312905, + "tokens_seen": 3044999168 + }, + { + "epoch": 0.92, + "learning_rate": 6.266800401203611e-05, + "loss": 1.3138, + "theoretical_loss": 3.345431801091817, + "tokens_seen": 3045064704 + }, + { + "epoch": 0.92, + "learning_rate": 6.265195586760281e-05, + "loss": 1.2995, + "theoretical_loss": 3.345426329003092, + "tokens_seen": 3045130240 + }, + { + "epoch": 0.92, + "learning_rate": 6.263590772316951e-05, + "loss": 1.1987, + "theoretical_loss": 3.3454208570651076, + "tokens_seen": 3045195776 + }, + { + "epoch": 0.92, + "learning_rate": 6.261985957873621e-05, + "loss": 1.1446, + "theoretical_loss": 3.345415385277857, + "tokens_seen": 3045261312 + }, + { + "epoch": 0.92, + "learning_rate": 6.260381143430291e-05, + "loss": 1.4006, + "theoretical_loss": 3.345409913641333, + "tokens_seen": 3045326848 + }, + { + "epoch": 0.92, + "learning_rate": 6.258776328986961e-05, + "loss": 1.2007, + "theoretical_loss": 3.345404442155527, + "tokens_seen": 3045392384 + }, + { + "epoch": 0.92, + "learning_rate": 6.257171514543631e-05, + "loss": 1.3032, + "theoretical_loss": 3.345398970820433, + "tokens_seen": 3045457920 + }, + { + "epoch": 0.92, + "learning_rate": 6.255566700100301e-05, + "loss": 1.2556, + "theoretical_loss": 3.3453934996360424, + "tokens_seen": 3045523456 + }, + { + "epoch": 0.92, + "learning_rate": 6.253961885656971e-05, + "loss": 1.2797, + "theoretical_loss": 3.3453880286023487, + "tokens_seen": 3045588992 + }, + { + "epoch": 0.92, + "learning_rate": 6.252357071213641e-05, + "loss": 1.3373, + "theoretical_loss": 3.345382557719344, + "tokens_seen": 3045654528 + }, + { + "epoch": 0.92, + "learning_rate": 6.250752256770312e-05, + "loss": 1.35, + "theoretical_loss": 3.345377086987021, + "tokens_seen": 3045720064 + }, + { + "epoch": 0.92, + "learning_rate": 6.249147442326982e-05, + "loss": 1.2419, + "theoretical_loss": 3.345371616405372, + "tokens_seen": 3045785600 + }, + { + "epoch": 0.92, + "learning_rate": 6.247542627883652e-05, + "loss": 1.1885, + "theoretical_loss": 3.345366145974391, + "tokens_seen": 3045851136 + }, + { + "epoch": 0.92, + "learning_rate": 6.245937813440322e-05, + "loss": 1.1871, + "theoretical_loss": 3.345360675694069, + "tokens_seen": 3045916672 + }, + { + "epoch": 0.92, + "learning_rate": 6.244332998996992e-05, + "loss": 1.2338, + "theoretical_loss": 3.345355205564399, + "tokens_seen": 3045982208 + }, + { + "epoch": 0.92, + "learning_rate": 6.24272818455366e-05, + "loss": 1.2335, + "theoretical_loss": 3.3453497355853745, + "tokens_seen": 3046047744 + }, + { + "epoch": 0.92, + "learning_rate": 6.24112337011033e-05, + "loss": 1.2422, + "theoretical_loss": 3.345344265756987, + "tokens_seen": 3046113280 + }, + { + "epoch": 0.92, + "learning_rate": 6.239518555667e-05, + "loss": 1.2698, + "theoretical_loss": 3.345338796079229, + "tokens_seen": 3046178816 + }, + { + "epoch": 0.92, + "learning_rate": 6.237913741223671e-05, + "loss": 1.1923, + "theoretical_loss": 3.3453333265520944, + "tokens_seen": 3046244352 + }, + { + "epoch": 0.92, + "learning_rate": 6.236308926780341e-05, + "loss": 1.2612, + "theoretical_loss": 3.345327857175575, + "tokens_seen": 3046309888 + }, + { + "epoch": 0.92, + "learning_rate": 6.234704112337011e-05, + "loss": 1.2538, + "theoretical_loss": 3.345322387949663, + "tokens_seen": 3046375424 + }, + { + "epoch": 0.92, + "learning_rate": 6.233099297893681e-05, + "loss": 1.2329, + "theoretical_loss": 3.345316918874352, + "tokens_seen": 3046440960 + }, + { + "epoch": 0.92, + "learning_rate": 6.231494483450351e-05, + "loss": 1.2341, + "theoretical_loss": 3.3453114499496337, + "tokens_seen": 3046506496 + }, + { + "epoch": 0.92, + "learning_rate": 6.229889669007022e-05, + "loss": 1.2208, + "theoretical_loss": 3.3453059811755015, + "tokens_seen": 3046572032 + }, + { + "epoch": 0.92, + "learning_rate": 6.228284854563692e-05, + "loss": 1.1412, + "theoretical_loss": 3.345300512551947, + "tokens_seen": 3046637568 + }, + { + "epoch": 0.92, + "learning_rate": 6.226680040120361e-05, + "loss": 1.2907, + "theoretical_loss": 3.3452950440789637, + "tokens_seen": 3046703104 + }, + { + "epoch": 0.92, + "learning_rate": 6.225075225677031e-05, + "loss": 1.2379, + "theoretical_loss": 3.3452895757565444, + "tokens_seen": 3046768640 + }, + { + "epoch": 0.92, + "learning_rate": 6.223470411233701e-05, + "loss": 1.1266, + "theoretical_loss": 3.345284107584681, + "tokens_seen": 3046834176 + }, + { + "epoch": 0.92, + "learning_rate": 6.221865596790371e-05, + "loss": 1.2796, + "theoretical_loss": 3.3452786395633662, + "tokens_seen": 3046899712 + }, + { + "epoch": 0.92, + "learning_rate": 6.220260782347041e-05, + "loss": 1.1365, + "theoretical_loss": 3.3452731716925928, + "tokens_seen": 3046965248 + }, + { + "epoch": 0.92, + "learning_rate": 6.218655967903711e-05, + "loss": 1.2192, + "theoretical_loss": 3.345267703972354, + "tokens_seen": 3047030784 + }, + { + "epoch": 0.92, + "learning_rate": 6.217051153460382e-05, + "loss": 1.3499, + "theoretical_loss": 3.345262236402641, + "tokens_seen": 3047096320 + }, + { + "epoch": 0.92, + "learning_rate": 6.215446339017052e-05, + "loss": 1.3405, + "theoretical_loss": 3.3452567689834476, + "tokens_seen": 3047161856 + }, + { + "epoch": 0.92, + "learning_rate": 6.213841524573722e-05, + "loss": 1.2714, + "theoretical_loss": 3.3452513017147663, + "tokens_seen": 3047227392 + }, + { + "epoch": 0.92, + "learning_rate": 6.212236710130392e-05, + "loss": 1.2353, + "theoretical_loss": 3.3452458345965894, + "tokens_seen": 3047292928 + }, + { + "epoch": 0.92, + "learning_rate": 6.210631895687062e-05, + "loss": 1.3199, + "theoretical_loss": 3.3452403676289095, + "tokens_seen": 3047358464 + }, + { + "epoch": 0.92, + "objective/train/docs_used": 1107536, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.3020427227020264, + "objective/train/theoretical_loss": 3.3452349008117195, + "objective/train/tokens_used": 3067884000, + "theoretical_loss": 3.3452349008117195, + "tokens_seen": 3047424000 + }, + { + "epoch": 0.92, + "learning_rate": 6.209027081243732e-05, + "loss": 1.3645, + "theoretical_loss": 3.3452349008117195, + "tokens_seen": 3047424000 + }, + { + "epoch": 0.92, + "learning_rate": 6.2074222668004e-05, + "loss": 1.098, + "theoretical_loss": 3.3452294341450117, + "tokens_seen": 3047489536 + }, + { + "epoch": 0.92, + "learning_rate": 6.20581745235707e-05, + "loss": 1.2508, + "theoretical_loss": 3.3452239676287787, + "tokens_seen": 3047555072 + }, + { + "epoch": 0.92, + "learning_rate": 6.20421263791374e-05, + "loss": 1.3073, + "theoretical_loss": 3.345218501263014, + "tokens_seen": 3047620608 + }, + { + "epoch": 0.92, + "learning_rate": 6.202607823470411e-05, + "loss": 1.2683, + "theoretical_loss": 3.345213035047709, + "tokens_seen": 3047686144 + }, + { + "epoch": 0.92, + "learning_rate": 6.201003009027082e-05, + "loss": 1.1617, + "theoretical_loss": 3.3452075689828566, + "tokens_seen": 3047751680 + }, + { + "epoch": 0.92, + "learning_rate": 6.199398194583752e-05, + "loss": 1.2792, + "theoretical_loss": 3.3452021030684502, + "tokens_seen": 3047817216 + }, + { + "epoch": 0.92, + "learning_rate": 6.197793380140422e-05, + "loss": 1.2344, + "theoretical_loss": 3.3451966373044817, + "tokens_seen": 3047882752 + }, + { + "epoch": 0.92, + "learning_rate": 6.196188565697092e-05, + "loss": 1.4103, + "theoretical_loss": 3.345191171690944, + "tokens_seen": 3047948288 + }, + { + "epoch": 0.92, + "learning_rate": 6.194583751253762e-05, + "loss": 1.2985, + "theoretical_loss": 3.345185706227829, + "tokens_seen": 3048013824 + }, + { + "epoch": 0.92, + "learning_rate": 6.192978936810432e-05, + "loss": 1.2873, + "theoretical_loss": 3.3451802409151306, + "tokens_seen": 3048079360 + }, + { + "epoch": 0.92, + "learning_rate": 6.191374122367101e-05, + "loss": 1.312, + "theoretical_loss": 3.3451747757528407, + "tokens_seen": 3048144896 + }, + { + "epoch": 0.92, + "learning_rate": 6.189769307923771e-05, + "loss": 1.2164, + "theoretical_loss": 3.3451693107409524, + "tokens_seen": 3048210432 + }, + { + "epoch": 0.92, + "learning_rate": 6.188164493480441e-05, + "loss": 1.249, + "theoretical_loss": 3.3451638458794575, + "tokens_seen": 3048275968 + }, + { + "epoch": 0.92, + "learning_rate": 6.186559679037111e-05, + "loss": 1.228, + "theoretical_loss": 3.345158381168349, + "tokens_seen": 3048341504 + }, + { + "epoch": 0.92, + "learning_rate": 6.184954864593781e-05, + "loss": 1.2391, + "theoretical_loss": 3.3451529166076197, + "tokens_seen": 3048407040 + }, + { + "epoch": 0.92, + "learning_rate": 6.183350050150452e-05, + "loss": 1.2873, + "theoretical_loss": 3.3451474521972626, + "tokens_seen": 3048472576 + }, + { + "epoch": 0.92, + "learning_rate": 6.181745235707122e-05, + "loss": 1.2782, + "theoretical_loss": 3.3451419879372692, + "tokens_seen": 3048538112 + }, + { + "epoch": 0.92, + "learning_rate": 6.180140421263792e-05, + "loss": 1.33, + "theoretical_loss": 3.3451365238276334, + "tokens_seen": 3048603648 + }, + { + "epoch": 0.92, + "learning_rate": 6.178535606820462e-05, + "loss": 1.225, + "theoretical_loss": 3.345131059868347, + "tokens_seen": 3048669184 + }, + { + "epoch": 0.92, + "learning_rate": 6.176930792377132e-05, + "loss": 1.228, + "theoretical_loss": 3.3451255960594026, + "tokens_seen": 3048734720 + }, + { + "epoch": 0.92, + "learning_rate": 6.175325977933802e-05, + "loss": 1.281, + "theoretical_loss": 3.345120132400793, + "tokens_seen": 3048800256 + }, + { + "epoch": 0.92, + "learning_rate": 6.173721163490472e-05, + "loss": 1.1666, + "theoretical_loss": 3.3451146688925117, + "tokens_seen": 3048865792 + }, + { + "epoch": 0.92, + "learning_rate": 6.17211634904714e-05, + "loss": 1.2518, + "theoretical_loss": 3.3451092055345497, + "tokens_seen": 3048931328 + }, + { + "epoch": 0.92, + "learning_rate": 6.170511534603812e-05, + "loss": 1.2379, + "theoretical_loss": 3.345103742326901, + "tokens_seen": 3048996864 + }, + { + "epoch": 0.92, + "learning_rate": 6.168906720160482e-05, + "loss": 1.2003, + "theoretical_loss": 3.345098279269558, + "tokens_seen": 3049062400 + }, + { + "epoch": 0.92, + "learning_rate": 6.167301905717152e-05, + "loss": 1.2597, + "theoretical_loss": 3.345092816362512, + "tokens_seen": 3049127936 + }, + { + "epoch": 0.92, + "learning_rate": 6.165697091273822e-05, + "loss": 1.2172, + "theoretical_loss": 3.3450873536057575, + "tokens_seen": 3049193472 + }, + { + "epoch": 0.92, + "learning_rate": 6.164092276830492e-05, + "loss": 1.2758, + "theoretical_loss": 3.3450818909992863, + "tokens_seen": 3049259008 + }, + { + "epoch": 0.92, + "learning_rate": 6.162487462387162e-05, + "loss": 1.2474, + "theoretical_loss": 3.345076428543091, + "tokens_seen": 3049324544 + }, + { + "epoch": 0.92, + "learning_rate": 6.160882647943832e-05, + "loss": 1.0953, + "theoretical_loss": 3.3450709662371643, + "tokens_seen": 3049390080 + }, + { + "epoch": 0.92, + "learning_rate": 6.159277833500502e-05, + "loss": 1.2292, + "theoretical_loss": 3.3450655040814987, + "tokens_seen": 3049455616 + }, + { + "epoch": 0.92, + "learning_rate": 6.157673019057173e-05, + "loss": 1.2799, + "theoretical_loss": 3.3450600420760868, + "tokens_seen": 3049521152 + }, + { + "epoch": 0.92, + "learning_rate": 6.156068204613841e-05, + "loss": 1.3439, + "theoretical_loss": 3.3450545802209217, + "tokens_seen": 3049586688 + }, + { + "epoch": 0.92, + "learning_rate": 6.154463390170511e-05, + "loss": 1.2427, + "theoretical_loss": 3.3450491185159956, + "tokens_seen": 3049652224 + }, + { + "epoch": 0.92, + "learning_rate": 6.152858575727181e-05, + "loss": 1.2595, + "theoretical_loss": 3.345043656961302, + "tokens_seen": 3049717760 + }, + { + "epoch": 0.92, + "learning_rate": 6.151253761283851e-05, + "loss": 1.2305, + "theoretical_loss": 3.345038195556832, + "tokens_seen": 3049783296 + }, + { + "epoch": 0.92, + "learning_rate": 6.149648946840521e-05, + "loss": 1.2252, + "theoretical_loss": 3.3450327343025794, + "tokens_seen": 3049848832 + }, + { + "epoch": 0.92, + "learning_rate": 6.148044132397192e-05, + "loss": 1.2198, + "theoretical_loss": 3.345027273198536, + "tokens_seen": 3049914368 + }, + { + "epoch": 0.92, + "learning_rate": 6.146439317953862e-05, + "loss": 1.2291, + "theoretical_loss": 3.3450218122446955, + "tokens_seen": 3049979904 + }, + { + "epoch": 0.92, + "learning_rate": 6.144834503510532e-05, + "loss": 1.23, + "theoretical_loss": 3.34501635144105, + "tokens_seen": 3050045440 + }, + { + "epoch": 0.92, + "learning_rate": 6.143229689067202e-05, + "loss": 1.2297, + "theoretical_loss": 3.345010890787592, + "tokens_seen": 3050110976 + }, + { + "epoch": 0.92, + "learning_rate": 6.141624874623872e-05, + "loss": 1.2843, + "theoretical_loss": 3.3450054302843144, + "tokens_seen": 3050176512 + }, + { + "epoch": 0.92, + "learning_rate": 6.140020060180542e-05, + "loss": 1.3304, + "theoretical_loss": 3.3449999699312096, + "tokens_seen": 3050242048 + }, + { + "epoch": 0.92, + "learning_rate": 6.138415245737212e-05, + "loss": 1.2886, + "theoretical_loss": 3.34499450972827, + "tokens_seen": 3050307584 + }, + { + "epoch": 0.92, + "learning_rate": 6.136810431293882e-05, + "loss": 1.27, + "theoretical_loss": 3.3449890496754895, + "tokens_seen": 3050373120 + }, + { + "epoch": 0.92, + "learning_rate": 6.135205616850552e-05, + "loss": 1.107, + "theoretical_loss": 3.344983589772859, + "tokens_seen": 3050438656 + }, + { + "epoch": 0.92, + "learning_rate": 6.133600802407222e-05, + "loss": 1.289, + "theoretical_loss": 3.3449781300203725, + "tokens_seen": 3050504192 + }, + { + "epoch": 0.92, + "learning_rate": 6.131995987963892e-05, + "loss": 1.2717, + "theoretical_loss": 3.344972670418022, + "tokens_seen": 3050569728 + }, + { + "epoch": 0.92, + "learning_rate": 6.130391173520562e-05, + "loss": 1.2972, + "theoretical_loss": 3.3449672109658, + "tokens_seen": 3050635264 + }, + { + "epoch": 0.92, + "objective/train/docs_used": 1108683, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.315598964691162, + "objective/train/theoretical_loss": 3.3449617516637, + "objective/train/tokens_used": 3071160800, + "theoretical_loss": 3.3449617516637, + "tokens_seen": 3050700800 + }, + { + "epoch": 0.92, + "learning_rate": 6.128786359077232e-05, + "loss": 1.2395, + "theoretical_loss": 3.3449617516637, + "tokens_seen": 3050700800 + }, + { + "epoch": 0.92, + "learning_rate": 6.127181544633902e-05, + "loss": 1.2621, + "theoretical_loss": 3.3449562925117133, + "tokens_seen": 3050766336 + }, + { + "epoch": 0.92, + "learning_rate": 6.125576730190572e-05, + "loss": 1.256, + "theoretical_loss": 3.344950833509834, + "tokens_seen": 3050831872 + }, + { + "epoch": 0.92, + "learning_rate": 6.123971915747242e-05, + "loss": 1.299, + "theoretical_loss": 3.344945374658054, + "tokens_seen": 3050897408 + }, + { + "epoch": 0.92, + "learning_rate": 6.122367101303913e-05, + "loss": 1.1358, + "theoretical_loss": 3.344939915956366, + "tokens_seen": 3050962944 + }, + { + "epoch": 0.92, + "learning_rate": 6.120762286860581e-05, + "loss": 1.2577, + "theoretical_loss": 3.3449344574047624, + "tokens_seen": 3051028480 + }, + { + "epoch": 0.92, + "learning_rate": 6.119157472417251e-05, + "loss": 1.2546, + "theoretical_loss": 3.3449289990032365, + "tokens_seen": 3051094016 + }, + { + "epoch": 0.92, + "learning_rate": 6.117552657973921e-05, + "loss": 1.2161, + "theoretical_loss": 3.3449235407517803, + "tokens_seen": 3051159552 + }, + { + "epoch": 0.92, + "learning_rate": 6.115947843530591e-05, + "loss": 1.3434, + "theoretical_loss": 3.344918082650387, + "tokens_seen": 3051225088 + }, + { + "epoch": 0.92, + "learning_rate": 6.114343029087262e-05, + "loss": 1.3208, + "theoretical_loss": 3.344912624699049, + "tokens_seen": 3051290624 + }, + { + "epoch": 0.92, + "learning_rate": 6.112738214643932e-05, + "loss": 1.3675, + "theoretical_loss": 3.3449071668977584, + "tokens_seen": 3051356160 + }, + { + "epoch": 0.92, + "learning_rate": 6.111133400200602e-05, + "loss": 1.2192, + "theoretical_loss": 3.3449017092465088, + "tokens_seen": 3051421696 + }, + { + "epoch": 0.92, + "learning_rate": 6.109528585757273e-05, + "loss": 1.1879, + "theoretical_loss": 3.3448962517452925, + "tokens_seen": 3051487232 + }, + { + "epoch": 0.92, + "learning_rate": 6.107923771313943e-05, + "loss": 1.2489, + "theoretical_loss": 3.344890794394102, + "tokens_seen": 3051552768 + }, + { + "epoch": 0.92, + "learning_rate": 6.106318956870612e-05, + "loss": 1.2134, + "theoretical_loss": 3.3448853371929297, + "tokens_seen": 3051618304 + }, + { + "epoch": 0.92, + "learning_rate": 6.104714142427282e-05, + "loss": 1.155, + "theoretical_loss": 3.344879880141769, + "tokens_seen": 3051683840 + }, + { + "epoch": 0.92, + "learning_rate": 6.1031093279839526e-05, + "loss": 1.1756, + "theoretical_loss": 3.3448744232406122, + "tokens_seen": 3051749376 + }, + { + "epoch": 0.92, + "learning_rate": 6.101504513540622e-05, + "loss": 1.2679, + "theoretical_loss": 3.3448689664894515, + "tokens_seen": 3051814912 + }, + { + "epoch": 0.92, + "learning_rate": 6.099899699097292e-05, + "loss": 1.2962, + "theoretical_loss": 3.34486350988828, + "tokens_seen": 3051880448 + }, + { + "epoch": 0.92, + "learning_rate": 6.098294884653962e-05, + "loss": 1.2011, + "theoretical_loss": 3.3448580534370906, + "tokens_seen": 3051945984 + }, + { + "epoch": 0.92, + "learning_rate": 6.096690070210632e-05, + "loss": 1.3041, + "theoretical_loss": 3.344852597135876, + "tokens_seen": 3052011520 + }, + { + "epoch": 0.92, + "learning_rate": 6.095085255767302e-05, + "loss": 1.2982, + "theoretical_loss": 3.344847140984628, + "tokens_seen": 3052077056 + }, + { + "epoch": 0.92, + "learning_rate": 6.0934804413239723e-05, + "loss": 1.2762, + "theoretical_loss": 3.34484168498334, + "tokens_seen": 3052142592 + }, + { + "epoch": 0.92, + "learning_rate": 6.091875626880642e-05, + "loss": 1.2067, + "theoretical_loss": 3.3448362291320044, + "tokens_seen": 3052208128 + }, + { + "epoch": 0.92, + "learning_rate": 6.090270812437312e-05, + "loss": 1.3168, + "theoretical_loss": 3.344830773430614, + "tokens_seen": 3052273664 + }, + { + "epoch": 0.92, + "learning_rate": 6.088665997993982e-05, + "loss": 1.2873, + "theoretical_loss": 3.3448253178791614, + "tokens_seen": 3052339200 + }, + { + "epoch": 0.92, + "learning_rate": 6.087061183550652e-05, + "loss": 1.1792, + "theoretical_loss": 3.3448198624776393, + "tokens_seen": 3052404736 + }, + { + "epoch": 0.92, + "learning_rate": 6.085456369107322e-05, + "loss": 1.2838, + "theoretical_loss": 3.34481440722604, + "tokens_seen": 3052470272 + }, + { + "epoch": 0.93, + "learning_rate": 6.083851554663992e-05, + "loss": 1.2885, + "theoretical_loss": 3.3448089521243567, + "tokens_seen": 3052535808 + }, + { + "epoch": 0.93, + "learning_rate": 6.0822467402206614e-05, + "loss": 1.2677, + "theoretical_loss": 3.3448034971725815, + "tokens_seen": 3052601344 + }, + { + "epoch": 0.93, + "learning_rate": 6.080641925777333e-05, + "loss": 1.3468, + "theoretical_loss": 3.3447980423707078, + "tokens_seen": 3052666880 + }, + { + "epoch": 0.93, + "learning_rate": 6.079037111334003e-05, + "loss": 1.1897, + "theoretical_loss": 3.3447925877187275, + "tokens_seen": 3052732416 + }, + { + "epoch": 0.93, + "learning_rate": 6.077432296890673e-05, + "loss": 1.2988, + "theoretical_loss": 3.3447871332166335, + "tokens_seen": 3052797952 + }, + { + "epoch": 0.93, + "learning_rate": 6.0758274824473424e-05, + "loss": 1.2161, + "theoretical_loss": 3.344781678864419, + "tokens_seen": 3052863488 + }, + { + "epoch": 0.93, + "learning_rate": 6.0742226680040124e-05, + "loss": 1.2446, + "theoretical_loss": 3.344776224662076, + "tokens_seen": 3052929024 + }, + { + "epoch": 0.93, + "learning_rate": 6.0726178535606825e-05, + "loss": 1.3479, + "theoretical_loss": 3.3447707706095975, + "tokens_seen": 3052994560 + }, + { + "epoch": 0.93, + "learning_rate": 6.0710130391173526e-05, + "loss": 1.2586, + "theoretical_loss": 3.344765316706976, + "tokens_seen": 3053060096 + }, + { + "epoch": 0.93, + "learning_rate": 6.0694082246740226e-05, + "loss": 1.1711, + "theoretical_loss": 3.344759862954204, + "tokens_seen": 3053125632 + }, + { + "epoch": 0.93, + "learning_rate": 6.067803410230693e-05, + "loss": 1.3186, + "theoretical_loss": 3.3447544093512747, + "tokens_seen": 3053191168 + }, + { + "epoch": 0.93, + "learning_rate": 6.066198595787362e-05, + "loss": 1.1687, + "theoretical_loss": 3.34474895589818, + "tokens_seen": 3053256704 + }, + { + "epoch": 0.93, + "learning_rate": 6.064593781344032e-05, + "loss": 1.2542, + "theoretical_loss": 3.344743502594914, + "tokens_seen": 3053322240 + }, + { + "epoch": 0.93, + "learning_rate": 6.062988966900702e-05, + "loss": 1.1937, + "theoretical_loss": 3.3447380494414674, + "tokens_seen": 3053387776 + }, + { + "epoch": 0.93, + "learning_rate": 6.061384152457372e-05, + "loss": 1.3495, + "theoretical_loss": 3.3447325964378347, + "tokens_seen": 3053453312 + }, + { + "epoch": 0.93, + "learning_rate": 6.059779338014042e-05, + "loss": 1.3711, + "theoretical_loss": 3.344727143584007, + "tokens_seen": 3053518848 + }, + { + "epoch": 0.93, + "learning_rate": 6.0581745235707124e-05, + "loss": 1.2468, + "theoretical_loss": 3.344721690879978, + "tokens_seen": 3053584384 + }, + { + "epoch": 0.93, + "learning_rate": 6.056569709127382e-05, + "loss": 1.1824, + "theoretical_loss": 3.3447162383257405, + "tokens_seen": 3053649920 + }, + { + "epoch": 0.93, + "learning_rate": 6.054964894684052e-05, + "loss": 1.145, + "theoretical_loss": 3.3447107859212863, + "tokens_seen": 3053715456 + }, + { + "epoch": 0.93, + "learning_rate": 6.053360080240722e-05, + "loss": 1.2364, + "theoretical_loss": 3.3447053336666084, + "tokens_seen": 3053780992 + }, + { + "epoch": 0.93, + "learning_rate": 6.051755265797392e-05, + "loss": 1.2488, + "theoretical_loss": 3.3446998815616995, + "tokens_seen": 3053846528 + }, + { + "epoch": 0.93, + "learning_rate": 6.050150451354063e-05, + "loss": 1.2518, + "theoretical_loss": 3.3446944296065526, + "tokens_seen": 3053912064 + }, + { + "epoch": 0.93, + "objective/train/docs_used": 1109702, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.1798789501190186, + "objective/train/theoretical_loss": 3.34468897780116, + "objective/train/tokens_used": 3074437600, + "theoretical_loss": 3.34468897780116, + "tokens_seen": 3053977600 + }, + { + "epoch": 0.93, + "learning_rate": 6.048545636910733e-05, + "loss": 1.2063, + "theoretical_loss": 3.34468897780116, + "tokens_seen": 3053977600 + }, + { + "epoch": 0.93, + "learning_rate": 6.046940822467403e-05, + "loss": 1.2511, + "theoretical_loss": 3.344683526145515, + "tokens_seen": 3054043136 + }, + { + "epoch": 0.93, + "learning_rate": 6.045336008024073e-05, + "loss": 1.1705, + "theoretical_loss": 3.344678074639609, + "tokens_seen": 3054108672 + }, + { + "epoch": 0.93, + "learning_rate": 6.043731193580743e-05, + "loss": 1.308, + "theoretical_loss": 3.344672623283436, + "tokens_seen": 3054174208 + }, + { + "epoch": 0.93, + "learning_rate": 6.042126379137413e-05, + "loss": 1.2896, + "theoretical_loss": 3.344667172076988, + "tokens_seen": 3054239744 + }, + { + "epoch": 0.93, + "learning_rate": 6.0405215646940824e-05, + "loss": 1.2719, + "theoretical_loss": 3.344661721020258, + "tokens_seen": 3054305280 + }, + { + "epoch": 0.93, + "learning_rate": 6.0389167502507525e-05, + "loss": 1.1319, + "theoretical_loss": 3.344656270113238, + "tokens_seen": 3054370816 + }, + { + "epoch": 0.93, + "learning_rate": 6.0373119358074225e-05, + "loss": 1.3082, + "theoretical_loss": 3.3446508193559215, + "tokens_seen": 3054436352 + }, + { + "epoch": 0.93, + "learning_rate": 6.0357071213640926e-05, + "loss": 1.2999, + "theoretical_loss": 3.3446453687483007, + "tokens_seen": 3054501888 + }, + { + "epoch": 0.93, + "learning_rate": 6.0341023069207627e-05, + "loss": 1.3405, + "theoretical_loss": 3.3446399182903686, + "tokens_seen": 3054567424 + }, + { + "epoch": 0.93, + "learning_rate": 6.032497492477433e-05, + "loss": 1.2381, + "theoretical_loss": 3.344634467982117, + "tokens_seen": 3054632960 + }, + { + "epoch": 0.93, + "learning_rate": 6.030892678034102e-05, + "loss": 1.1781, + "theoretical_loss": 3.3446290178235403, + "tokens_seen": 3054698496 + }, + { + "epoch": 0.93, + "learning_rate": 6.029287863590772e-05, + "loss": 1.2336, + "theoretical_loss": 3.3446235678146294, + "tokens_seen": 3054764032 + }, + { + "epoch": 0.93, + "learning_rate": 6.027683049147442e-05, + "loss": 1.2889, + "theoretical_loss": 3.344618117955378, + "tokens_seen": 3054829568 + }, + { + "epoch": 0.93, + "learning_rate": 6.026078234704112e-05, + "loss": 1.2476, + "theoretical_loss": 3.3446126682457784, + "tokens_seen": 3054895104 + }, + { + "epoch": 0.93, + "learning_rate": 6.0244734202607824e-05, + "loss": 1.2998, + "theoretical_loss": 3.3446072186858236, + "tokens_seen": 3054960640 + }, + { + "epoch": 0.93, + "learning_rate": 6.0228686058174524e-05, + "loss": 1.2747, + "theoretical_loss": 3.344601769275506, + "tokens_seen": 3055026176 + }, + { + "epoch": 0.93, + "learning_rate": 6.021263791374122e-05, + "loss": 1.3088, + "theoretical_loss": 3.3445963200148183, + "tokens_seen": 3055091712 + }, + { + "epoch": 0.93, + "learning_rate": 6.019658976930793e-05, + "loss": 1.2504, + "theoretical_loss": 3.344590870903753, + "tokens_seen": 3055157248 + }, + { + "epoch": 0.93, + "learning_rate": 6.018054162487463e-05, + "loss": 1.3107, + "theoretical_loss": 3.3445854219423032, + "tokens_seen": 3055222784 + }, + { + "epoch": 0.93, + "learning_rate": 6.0164493480441334e-05, + "loss": 1.3602, + "theoretical_loss": 3.3445799731304615, + "tokens_seen": 3055288320 + }, + { + "epoch": 0.93, + "learning_rate": 6.014844533600803e-05, + "loss": 1.2389, + "theoretical_loss": 3.34457452446822, + "tokens_seen": 3055353856 + }, + { + "epoch": 0.93, + "learning_rate": 6.013239719157473e-05, + "loss": 1.295, + "theoretical_loss": 3.3445690759555724, + "tokens_seen": 3055419392 + }, + { + "epoch": 0.93, + "learning_rate": 6.011634904714143e-05, + "loss": 1.2564, + "theoretical_loss": 3.344563627592511, + "tokens_seen": 3055484928 + }, + { + "epoch": 0.93, + "learning_rate": 6.010030090270813e-05, + "loss": 1.1837, + "theoretical_loss": 3.3445581793790278, + "tokens_seen": 3055550464 + }, + { + "epoch": 0.93, + "learning_rate": 6.008425275827483e-05, + "loss": 1.2296, + "theoretical_loss": 3.344552731315116, + "tokens_seen": 3055616000 + }, + { + "epoch": 0.93, + "learning_rate": 6.006820461384153e-05, + "loss": 1.3274, + "theoretical_loss": 3.3445472834007686, + "tokens_seen": 3055681536 + }, + { + "epoch": 0.93, + "learning_rate": 6.0052156469408224e-05, + "loss": 1.2418, + "theoretical_loss": 3.3445418356359777, + "tokens_seen": 3055747072 + }, + { + "epoch": 0.93, + "learning_rate": 6.0036108324974925e-05, + "loss": 1.2353, + "theoretical_loss": 3.3445363880207366, + "tokens_seen": 3055812608 + }, + { + "epoch": 0.93, + "learning_rate": 6.0020060180541626e-05, + "loss": 1.2769, + "theoretical_loss": 3.344530940555037, + "tokens_seen": 3055878144 + }, + { + "epoch": 0.93, + "learning_rate": 6.0004012036108326e-05, + "loss": 1.2822, + "theoretical_loss": 3.3445254932388733, + "tokens_seen": 3055943680 + }, + { + "epoch": 0.93, + "learning_rate": 5.998796389167503e-05, + "loss": 1.2392, + "theoretical_loss": 3.3445200460722364, + "tokens_seen": 3056009216 + }, + { + "epoch": 0.93, + "learning_rate": 5.997191574724173e-05, + "loss": 1.2797, + "theoretical_loss": 3.34451459905512, + "tokens_seen": 3056074752 + }, + { + "epoch": 0.93, + "learning_rate": 5.995586760280842e-05, + "loss": 1.2028, + "theoretical_loss": 3.3445091521875163, + "tokens_seen": 3056140288 + }, + { + "epoch": 0.93, + "learning_rate": 5.993981945837512e-05, + "loss": 1.0834, + "theoretical_loss": 3.3445037054694184, + "tokens_seen": 3056205824 + }, + { + "epoch": 0.93, + "learning_rate": 5.992377131394182e-05, + "loss": 1.2392, + "theoretical_loss": 3.3444982589008188, + "tokens_seen": 3056271360 + }, + { + "epoch": 0.93, + "learning_rate": 5.990772316950854e-05, + "loss": 1.2803, + "theoretical_loss": 3.34449281248171, + "tokens_seen": 3056336896 + }, + { + "epoch": 0.93, + "learning_rate": 5.989167502507523e-05, + "loss": 1.2883, + "theoretical_loss": 3.344487366212085, + "tokens_seen": 3056402432 + }, + { + "epoch": 0.93, + "learning_rate": 5.987562688064193e-05, + "loss": 1.2458, + "theoretical_loss": 3.3444819200919365, + "tokens_seen": 3056467968 + }, + { + "epoch": 0.93, + "learning_rate": 5.985957873620863e-05, + "loss": 1.2304, + "theoretical_loss": 3.3444764741212567, + "tokens_seen": 3056533504 + }, + { + "epoch": 0.93, + "learning_rate": 5.984353059177533e-05, + "loss": 1.2401, + "theoretical_loss": 3.3444710283000387, + "tokens_seen": 3056599040 + }, + { + "epoch": 0.93, + "learning_rate": 5.982748244734203e-05, + "loss": 1.2662, + "theoretical_loss": 3.3444655826282754, + "tokens_seen": 3056664576 + }, + { + "epoch": 0.93, + "learning_rate": 5.9811434302908734e-05, + "loss": 1.3067, + "theoretical_loss": 3.3444601371059592, + "tokens_seen": 3056730112 + }, + { + "epoch": 0.93, + "learning_rate": 5.979538615847543e-05, + "loss": 1.2317, + "theoretical_loss": 3.344454691733083, + "tokens_seen": 3056795648 + }, + { + "epoch": 0.93, + "learning_rate": 5.977933801404213e-05, + "loss": 1.1691, + "theoretical_loss": 3.3444492465096394, + "tokens_seen": 3056861184 + }, + { + "epoch": 0.93, + "learning_rate": 5.976328986960883e-05, + "loss": 1.244, + "theoretical_loss": 3.3444438014356206, + "tokens_seen": 3056926720 + }, + { + "epoch": 0.93, + "learning_rate": 5.974724172517553e-05, + "loss": 1.1944, + "theoretical_loss": 3.3444383565110196, + "tokens_seen": 3056992256 + }, + { + "epoch": 0.93, + "learning_rate": 5.973119358074223e-05, + "loss": 1.194, + "theoretical_loss": 3.3444329117358294, + "tokens_seen": 3057057792 + }, + { + "epoch": 0.93, + "learning_rate": 5.971514543630893e-05, + "loss": 1.2543, + "theoretical_loss": 3.3444274671100427, + "tokens_seen": 3057123328 + }, + { + "epoch": 0.93, + "learning_rate": 5.9699097291875625e-05, + "loss": 1.3271, + "theoretical_loss": 3.344422022633652, + "tokens_seen": 3057188864 + }, + { + "epoch": 0.93, + "objective/train/docs_used": 1110949, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.3445825576782227, + "objective/train/theoretical_loss": 3.34441657830665, + "objective/train/tokens_used": 3077714400, + "theoretical_loss": 3.34441657830665, + "tokens_seen": 3057254400 + }, + { + "epoch": 0.93, + "learning_rate": 5.9683049147442326e-05, + "loss": 1.2545, + "theoretical_loss": 3.34441657830665, + "tokens_seen": 3057254400 + }, + { + "epoch": 0.93, + "learning_rate": 5.9667001003009026e-05, + "loss": 1.219, + "theoretical_loss": 3.3444111341290297, + "tokens_seen": 3057319936 + }, + { + "epoch": 0.93, + "learning_rate": 5.965095285857573e-05, + "loss": 1.2693, + "theoretical_loss": 3.344405690100783, + "tokens_seen": 3057385472 + }, + { + "epoch": 0.93, + "learning_rate": 5.963490471414243e-05, + "loss": 1.2236, + "theoretical_loss": 3.344400246221903, + "tokens_seen": 3057451008 + }, + { + "epoch": 0.93, + "learning_rate": 5.961885656970913e-05, + "loss": 1.3034, + "theoretical_loss": 3.344394802492383, + "tokens_seen": 3057516544 + }, + { + "epoch": 0.93, + "learning_rate": 5.9602808425275835e-05, + "loss": 1.1868, + "theoretical_loss": 3.344389358912215, + "tokens_seen": 3057582080 + }, + { + "epoch": 0.93, + "learning_rate": 5.9586760280842536e-05, + "loss": 1.274, + "theoretical_loss": 3.344383915481392, + "tokens_seen": 3057647616 + }, + { + "epoch": 0.93, + "learning_rate": 5.957071213640924e-05, + "loss": 1.2231, + "theoretical_loss": 3.3443784721999066, + "tokens_seen": 3057713152 + }, + { + "epoch": 0.93, + "learning_rate": 5.955466399197594e-05, + "loss": 1.2519, + "theoretical_loss": 3.3443730290677514, + "tokens_seen": 3057778688 + }, + { + "epoch": 0.93, + "learning_rate": 5.953861584754263e-05, + "loss": 1.3338, + "theoretical_loss": 3.3443675860849194, + "tokens_seen": 3057844224 + }, + { + "epoch": 0.93, + "learning_rate": 5.952256770310933e-05, + "loss": 1.211, + "theoretical_loss": 3.344362143251403, + "tokens_seen": 3057909760 + }, + { + "epoch": 0.93, + "learning_rate": 5.950651955867603e-05, + "loss": 1.2981, + "theoretical_loss": 3.344356700567195, + "tokens_seen": 3057975296 + }, + { + "epoch": 0.93, + "learning_rate": 5.949047141424273e-05, + "loss": 1.3283, + "theoretical_loss": 3.3443512580322885, + "tokens_seen": 3058040832 + }, + { + "epoch": 0.93, + "learning_rate": 5.9474423269809434e-05, + "loss": 1.2003, + "theoretical_loss": 3.344345815646675, + "tokens_seen": 3058106368 + }, + { + "epoch": 0.93, + "learning_rate": 5.9458375125376134e-05, + "loss": 1.2498, + "theoretical_loss": 3.3443403734103487, + "tokens_seen": 3058171904 + }, + { + "epoch": 0.93, + "learning_rate": 5.944232698094283e-05, + "loss": 1.2263, + "theoretical_loss": 3.344334931323302, + "tokens_seen": 3058237440 + }, + { + "epoch": 0.93, + "learning_rate": 5.942627883650953e-05, + "loss": 1.3255, + "theoretical_loss": 3.3443294893855264, + "tokens_seen": 3058302976 + }, + { + "epoch": 0.93, + "learning_rate": 5.941023069207623e-05, + "loss": 1.292, + "theoretical_loss": 3.344324047597016, + "tokens_seen": 3058368512 + }, + { + "epoch": 0.93, + "learning_rate": 5.939418254764293e-05, + "loss": 1.1769, + "theoretical_loss": 3.344318605957763, + "tokens_seen": 3058434048 + }, + { + "epoch": 0.93, + "learning_rate": 5.937813440320963e-05, + "loss": 1.2905, + "theoretical_loss": 3.34431316446776, + "tokens_seen": 3058499584 + }, + { + "epoch": 0.93, + "learning_rate": 5.936208625877633e-05, + "loss": 1.256, + "theoretical_loss": 3.3443077231269993, + "tokens_seen": 3058565120 + }, + { + "epoch": 0.93, + "learning_rate": 5.9346038114343025e-05, + "loss": 1.1681, + "theoretical_loss": 3.3443022819354744, + "tokens_seen": 3058630656 + }, + { + "epoch": 0.93, + "learning_rate": 5.9329989969909726e-05, + "loss": 1.2484, + "theoretical_loss": 3.344296840893178, + "tokens_seen": 3058696192 + }, + { + "epoch": 0.93, + "learning_rate": 5.9313941825476427e-05, + "loss": 1.2773, + "theoretical_loss": 3.3442914000001025, + "tokens_seen": 3058761728 + }, + { + "epoch": 0.93, + "learning_rate": 5.929789368104314e-05, + "loss": 1.2823, + "theoretical_loss": 3.34428595925624, + "tokens_seen": 3058827264 + }, + { + "epoch": 0.93, + "learning_rate": 5.9281845536609835e-05, + "loss": 1.28, + "theoretical_loss": 3.3442805186615843, + "tokens_seen": 3058892800 + }, + { + "epoch": 0.93, + "learning_rate": 5.9265797392176535e-05, + "loss": 1.2827, + "theoretical_loss": 3.344275078216128, + "tokens_seen": 3058958336 + }, + { + "epoch": 0.93, + "learning_rate": 5.9249749247743236e-05, + "loss": 1.1891, + "theoretical_loss": 3.344269637919863, + "tokens_seen": 3059023872 + }, + { + "epoch": 0.93, + "learning_rate": 5.9233701103309936e-05, + "loss": 1.1957, + "theoretical_loss": 3.3442641977727825, + "tokens_seen": 3059089408 + }, + { + "epoch": 0.93, + "learning_rate": 5.921765295887664e-05, + "loss": 1.2417, + "theoretical_loss": 3.3442587577748792, + "tokens_seen": 3059154944 + }, + { + "epoch": 0.93, + "learning_rate": 5.920160481444334e-05, + "loss": 1.1348, + "theoretical_loss": 3.344253317926146, + "tokens_seen": 3059220480 + }, + { + "epoch": 0.93, + "learning_rate": 5.918555667001003e-05, + "loss": 1.2426, + "theoretical_loss": 3.344247878226575, + "tokens_seen": 3059286016 + }, + { + "epoch": 0.93, + "learning_rate": 5.916950852557673e-05, + "loss": 1.2585, + "theoretical_loss": 3.34424243867616, + "tokens_seen": 3059351552 + }, + { + "epoch": 0.93, + "learning_rate": 5.915346038114343e-05, + "loss": 1.2771, + "theoretical_loss": 3.3442369992748926, + "tokens_seen": 3059417088 + }, + { + "epoch": 0.93, + "learning_rate": 5.9137412236710134e-05, + "loss": 1.2628, + "theoretical_loss": 3.3442315600227657, + "tokens_seen": 3059482624 + }, + { + "epoch": 0.93, + "learning_rate": 5.9121364092276834e-05, + "loss": 1.284, + "theoretical_loss": 3.344226120919773, + "tokens_seen": 3059548160 + }, + { + "epoch": 0.93, + "learning_rate": 5.9105315947843535e-05, + "loss": 1.2205, + "theoretical_loss": 3.344220681965906, + "tokens_seen": 3059613696 + }, + { + "epoch": 0.93, + "learning_rate": 5.908926780341023e-05, + "loss": 1.264, + "theoretical_loss": 3.344215243161158, + "tokens_seen": 3059679232 + }, + { + "epoch": 0.93, + "learning_rate": 5.907321965897693e-05, + "loss": 1.2531, + "theoretical_loss": 3.3442098045055215, + "tokens_seen": 3059744768 + }, + { + "epoch": 0.93, + "learning_rate": 5.905717151454363e-05, + "loss": 1.2489, + "theoretical_loss": 3.3442043659989893, + "tokens_seen": 3059810304 + }, + { + "epoch": 0.93, + "learning_rate": 5.904112337011033e-05, + "loss": 1.2054, + "theoretical_loss": 3.3441989276415542, + "tokens_seen": 3059875840 + }, + { + "epoch": 0.93, + "learning_rate": 5.902507522567703e-05, + "loss": 1.2732, + "theoretical_loss": 3.3441934894332093, + "tokens_seen": 3059941376 + }, + { + "epoch": 0.93, + "learning_rate": 5.900902708124373e-05, + "loss": 1.2896, + "theoretical_loss": 3.3441880513739464, + "tokens_seen": 3060006912 + }, + { + "epoch": 0.93, + "learning_rate": 5.899297893681044e-05, + "loss": 1.2736, + "theoretical_loss": 3.344182613463759, + "tokens_seen": 3060072448 + }, + { + "epoch": 0.93, + "learning_rate": 5.897693079237714e-05, + "loss": 1.1838, + "theoretical_loss": 3.3441771757026397, + "tokens_seen": 3060137984 + }, + { + "epoch": 0.93, + "learning_rate": 5.896088264794384e-05, + "loss": 1.2174, + "theoretical_loss": 3.344171738090581, + "tokens_seen": 3060203520 + }, + { + "epoch": 0.93, + "learning_rate": 5.8944834503510534e-05, + "loss": 1.135, + "theoretical_loss": 3.3441663006275753, + "tokens_seen": 3060269056 + }, + { + "epoch": 0.93, + "learning_rate": 5.8928786359077235e-05, + "loss": 1.1602, + "theoretical_loss": 3.344160863313616, + "tokens_seen": 3060334592 + }, + { + "epoch": 0.93, + "learning_rate": 5.8912738214643936e-05, + "loss": 1.2315, + "theoretical_loss": 3.3441554261486957, + "tokens_seen": 3060400128 + }, + { + "epoch": 0.93, + "learning_rate": 5.8896690070210636e-05, + "loss": 1.2532, + "theoretical_loss": 3.3441499891328066, + "tokens_seen": 3060465664 + }, + { + "epoch": 0.93, + "objective/train/docs_used": 1112043, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.1829795837402344, + "objective/train/theoretical_loss": 3.344144552265942, + "objective/train/tokens_used": 3080991200, + "theoretical_loss": 3.344144552265942, + "tokens_seen": 3060531200 + }, + { + "epoch": 0.93, + "learning_rate": 5.888064192577734e-05, + "loss": 1.2151, + "theoretical_loss": 3.344144552265942, + "tokens_seen": 3060531200 + }, + { + "epoch": 0.93, + "learning_rate": 5.886459378134404e-05, + "loss": 1.2029, + "theoretical_loss": 3.3441391155480944, + "tokens_seen": 3060596736 + }, + { + "epoch": 0.93, + "learning_rate": 5.884854563691073e-05, + "loss": 1.2221, + "theoretical_loss": 3.344133678979256, + "tokens_seen": 3060662272 + }, + { + "epoch": 0.93, + "learning_rate": 5.883249749247743e-05, + "loss": 1.2707, + "theoretical_loss": 3.344128242559421, + "tokens_seen": 3060727808 + }, + { + "epoch": 0.93, + "learning_rate": 5.881644934804413e-05, + "loss": 1.2558, + "theoretical_loss": 3.344122806288581, + "tokens_seen": 3060793344 + }, + { + "epoch": 0.93, + "learning_rate": 5.880040120361083e-05, + "loss": 1.1639, + "theoretical_loss": 3.344117370166728, + "tokens_seen": 3060858880 + }, + { + "epoch": 0.93, + "learning_rate": 5.8784353059177534e-05, + "loss": 1.2147, + "theoretical_loss": 3.3441119341938568, + "tokens_seen": 3060924416 + }, + { + "epoch": 0.93, + "learning_rate": 5.8768304914744235e-05, + "loss": 1.2325, + "theoretical_loss": 3.344106498369958, + "tokens_seen": 3060989952 + }, + { + "epoch": 0.93, + "learning_rate": 5.875225677031093e-05, + "loss": 1.2346, + "theoretical_loss": 3.344101062695026, + "tokens_seen": 3061055488 + }, + { + "epoch": 0.93, + "learning_rate": 5.873620862587763e-05, + "loss": 1.1929, + "theoretical_loss": 3.3440956271690525, + "tokens_seen": 3061121024 + }, + { + "epoch": 0.93, + "learning_rate": 5.872016048144433e-05, + "loss": 1.2064, + "theoretical_loss": 3.3440901917920307, + "tokens_seen": 3061186560 + }, + { + "epoch": 0.93, + "learning_rate": 5.8704112337011044e-05, + "loss": 1.1779, + "theoretical_loss": 3.344084756563953, + "tokens_seen": 3061252096 + }, + { + "epoch": 0.93, + "learning_rate": 5.868806419257774e-05, + "loss": 1.2741, + "theoretical_loss": 3.3440793214848124, + "tokens_seen": 3061317632 + }, + { + "epoch": 0.93, + "learning_rate": 5.867201604814444e-05, + "loss": 1.2426, + "theoretical_loss": 3.3440738865546016, + "tokens_seen": 3061383168 + }, + { + "epoch": 0.93, + "learning_rate": 5.865596790371114e-05, + "loss": 1.2578, + "theoretical_loss": 3.344068451773313, + "tokens_seen": 3061448704 + }, + { + "epoch": 0.93, + "learning_rate": 5.863991975927784e-05, + "loss": 1.1615, + "theoretical_loss": 3.34406301714094, + "tokens_seen": 3061514240 + }, + { + "epoch": 0.93, + "learning_rate": 5.862387161484454e-05, + "loss": 1.2677, + "theoretical_loss": 3.3440575826574745, + "tokens_seen": 3061579776 + }, + { + "epoch": 0.93, + "learning_rate": 5.860782347041124e-05, + "loss": 1.2576, + "theoretical_loss": 3.34405214832291, + "tokens_seen": 3061645312 + }, + { + "epoch": 0.93, + "learning_rate": 5.8591775325977935e-05, + "loss": 1.2352, + "theoretical_loss": 3.3440467141372388, + "tokens_seen": 3061710848 + }, + { + "epoch": 0.93, + "learning_rate": 5.8575727181544635e-05, + "loss": 1.3234, + "theoretical_loss": 3.3440412801004538, + "tokens_seen": 3061776384 + }, + { + "epoch": 0.93, + "learning_rate": 5.8559679037111336e-05, + "loss": 1.2182, + "theoretical_loss": 3.3440358462125475, + "tokens_seen": 3061841920 + }, + { + "epoch": 0.93, + "learning_rate": 5.854363089267804e-05, + "loss": 1.2838, + "theoretical_loss": 3.344030412473513, + "tokens_seen": 3061907456 + }, + { + "epoch": 0.93, + "learning_rate": 5.852758274824474e-05, + "loss": 1.2737, + "theoretical_loss": 3.3440249788833425, + "tokens_seen": 3061972992 + }, + { + "epoch": 0.93, + "learning_rate": 5.851153460381144e-05, + "loss": 1.2465, + "theoretical_loss": 3.3440195454420296, + "tokens_seen": 3062038528 + }, + { + "epoch": 0.93, + "learning_rate": 5.849548645937813e-05, + "loss": 1.3137, + "theoretical_loss": 3.344014112149566, + "tokens_seen": 3062104064 + }, + { + "epoch": 0.93, + "learning_rate": 5.847943831494483e-05, + "loss": 1.1726, + "theoretical_loss": 3.3440086790059453, + "tokens_seen": 3062169600 + }, + { + "epoch": 0.93, + "learning_rate": 5.846339017051153e-05, + "loss": 1.2398, + "theoretical_loss": 3.3440032460111597, + "tokens_seen": 3062235136 + }, + { + "epoch": 0.93, + "learning_rate": 5.8447342026078234e-05, + "loss": 1.2716, + "theoretical_loss": 3.343997813165202, + "tokens_seen": 3062300672 + }, + { + "epoch": 0.93, + "learning_rate": 5.8431293881644934e-05, + "loss": 1.2385, + "theoretical_loss": 3.3439923804680656, + "tokens_seen": 3062366208 + }, + { + "epoch": 0.93, + "learning_rate": 5.8415245737211635e-05, + "loss": 1.2716, + "theoretical_loss": 3.343986947919742, + "tokens_seen": 3062431744 + }, + { + "epoch": 0.93, + "learning_rate": 5.839919759277834e-05, + "loss": 1.2364, + "theoretical_loss": 3.343981515520225, + "tokens_seen": 3062497280 + }, + { + "epoch": 0.93, + "learning_rate": 5.838314944834504e-05, + "loss": 1.3039, + "theoretical_loss": 3.343976083269507, + "tokens_seen": 3062562816 + }, + { + "epoch": 0.93, + "learning_rate": 5.8367101303911744e-05, + "loss": 1.24, + "theoretical_loss": 3.343970651167581, + "tokens_seen": 3062628352 + }, + { + "epoch": 0.93, + "learning_rate": 5.8351053159478444e-05, + "loss": 1.2541, + "theoretical_loss": 3.343965219214439, + "tokens_seen": 3062693888 + }, + { + "epoch": 0.93, + "learning_rate": 5.833500501504514e-05, + "loss": 1.2077, + "theoretical_loss": 3.343959787410074, + "tokens_seen": 3062759424 + }, + { + "epoch": 0.93, + "learning_rate": 5.831895687061184e-05, + "loss": 1.346, + "theoretical_loss": 3.3439543557544793, + "tokens_seen": 3062824960 + }, + { + "epoch": 0.93, + "learning_rate": 5.830290872617854e-05, + "loss": 1.227, + "theoretical_loss": 3.3439489242476474, + "tokens_seen": 3062890496 + }, + { + "epoch": 0.93, + "learning_rate": 5.828686058174524e-05, + "loss": 1.2258, + "theoretical_loss": 3.343943492889571, + "tokens_seen": 3062956032 + }, + { + "epoch": 0.93, + "learning_rate": 5.827081243731194e-05, + "loss": 1.2528, + "theoretical_loss": 3.3439380616802423, + "tokens_seen": 3063021568 + }, + { + "epoch": 0.93, + "learning_rate": 5.825476429287864e-05, + "loss": 1.2671, + "theoretical_loss": 3.343932630619655, + "tokens_seen": 3063087104 + }, + { + "epoch": 0.93, + "learning_rate": 5.8238716148445335e-05, + "loss": 1.334, + "theoretical_loss": 3.343927199707801, + "tokens_seen": 3063152640 + }, + { + "epoch": 0.93, + "learning_rate": 5.8222668004012036e-05, + "loss": 1.2445, + "theoretical_loss": 3.343921768944673, + "tokens_seen": 3063218176 + }, + { + "epoch": 0.93, + "learning_rate": 5.8206619859578736e-05, + "loss": 1.2304, + "theoretical_loss": 3.343916338330265, + "tokens_seen": 3063283712 + }, + { + "epoch": 0.93, + "learning_rate": 5.819057171514544e-05, + "loss": 1.2681, + "theoretical_loss": 3.3439109078645686, + "tokens_seen": 3063349248 + }, + { + "epoch": 0.93, + "learning_rate": 5.817452357071214e-05, + "loss": 1.2397, + "theoretical_loss": 3.3439054775475765, + "tokens_seen": 3063414784 + }, + { + "epoch": 0.93, + "learning_rate": 5.815847542627884e-05, + "loss": 1.2476, + "theoretical_loss": 3.3439000473792824, + "tokens_seen": 3063480320 + }, + { + "epoch": 0.93, + "learning_rate": 5.814242728184553e-05, + "loss": 1.2986, + "theoretical_loss": 3.3438946173596777, + "tokens_seen": 3063545856 + }, + { + "epoch": 0.93, + "learning_rate": 5.812637913741223e-05, + "loss": 1.2747, + "theoretical_loss": 3.3438891874887564, + "tokens_seen": 3063611392 + }, + { + "epoch": 0.93, + "learning_rate": 5.8110330992978933e-05, + "loss": 1.2565, + "theoretical_loss": 3.3438837577665104, + "tokens_seen": 3063676928 + }, + { + "epoch": 0.93, + "learning_rate": 5.809428284854565e-05, + "loss": 1.1755, + "theoretical_loss": 3.3438783281929334, + "tokens_seen": 3063742464 + }, + { + "epoch": 0.93, + "objective/train/docs_used": 1113272, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.3536903858184814, + "objective/train/theoretical_loss": 3.3438728987680166, + "objective/train/tokens_used": 3084268000, + "theoretical_loss": 3.3438728987680166, + "tokens_seen": 3063808000 + }, + { + "epoch": 0.93, + "learning_rate": 5.807823470411234e-05, + "loss": 1.307, + "theoretical_loss": 3.3438728987680166, + "tokens_seen": 3063808000 + }, + { + "epoch": 0.93, + "learning_rate": 5.806218655967904e-05, + "loss": 1.2471, + "theoretical_loss": 3.3438674694917543, + "tokens_seen": 3063873536 + }, + { + "epoch": 0.93, + "learning_rate": 5.804613841524574e-05, + "loss": 1.3082, + "theoretical_loss": 3.343862040364139, + "tokens_seen": 3063939072 + }, + { + "epoch": 0.93, + "learning_rate": 5.8030090270812443e-05, + "loss": 1.1457, + "theoretical_loss": 3.343856611385162, + "tokens_seen": 3064004608 + }, + { + "epoch": 0.93, + "learning_rate": 5.8014042126379144e-05, + "loss": 1.1986, + "theoretical_loss": 3.343851182554818, + "tokens_seen": 3064070144 + }, + { + "epoch": 0.93, + "learning_rate": 5.7997993981945845e-05, + "loss": 1.1917, + "theoretical_loss": 3.3438457538730986, + "tokens_seen": 3064135680 + }, + { + "epoch": 0.93, + "learning_rate": 5.798194583751254e-05, + "loss": 1.2684, + "theoretical_loss": 3.343840325339997, + "tokens_seen": 3064201216 + }, + { + "epoch": 0.93, + "learning_rate": 5.796589769307924e-05, + "loss": 1.3217, + "theoretical_loss": 3.3438348969555056, + "tokens_seen": 3064266752 + }, + { + "epoch": 0.93, + "learning_rate": 5.794984954864594e-05, + "loss": 1.1574, + "theoretical_loss": 3.3438294687196173, + "tokens_seen": 3064332288 + }, + { + "epoch": 0.93, + "learning_rate": 5.793380140421264e-05, + "loss": 1.1468, + "theoretical_loss": 3.3438240406323247, + "tokens_seen": 3064397824 + }, + { + "epoch": 0.93, + "learning_rate": 5.791775325977934e-05, + "loss": 1.2133, + "theoretical_loss": 3.343818612693621, + "tokens_seen": 3064463360 + }, + { + "epoch": 0.93, + "learning_rate": 5.790170511534604e-05, + "loss": 1.2816, + "theoretical_loss": 3.3438131849034987, + "tokens_seen": 3064528896 + }, + { + "epoch": 0.93, + "learning_rate": 5.7885656970912736e-05, + "loss": 1.2433, + "theoretical_loss": 3.343807757261951, + "tokens_seen": 3064594432 + }, + { + "epoch": 0.93, + "learning_rate": 5.7869608826479436e-05, + "loss": 1.3061, + "theoretical_loss": 3.3438023297689696, + "tokens_seen": 3064659968 + }, + { + "epoch": 0.93, + "learning_rate": 5.785356068204614e-05, + "loss": 1.3317, + "theoretical_loss": 3.3437969024245477, + "tokens_seen": 3064725504 + }, + { + "epoch": 0.93, + "learning_rate": 5.783751253761284e-05, + "loss": 1.2579, + "theoretical_loss": 3.343791475228679, + "tokens_seen": 3064791040 + }, + { + "epoch": 0.93, + "learning_rate": 5.782146439317954e-05, + "loss": 1.1889, + "theoretical_loss": 3.3437860481813546, + "tokens_seen": 3064856576 + }, + { + "epoch": 0.93, + "learning_rate": 5.7805416248746246e-05, + "loss": 1.3825, + "theoretical_loss": 3.3437806212825687, + "tokens_seen": 3064922112 + }, + { + "epoch": 0.93, + "learning_rate": 5.7789368104312946e-05, + "loss": 1.2457, + "theoretical_loss": 3.3437751945323138, + "tokens_seen": 3064987648 + }, + { + "epoch": 0.93, + "learning_rate": 5.777331995987965e-05, + "loss": 1.2649, + "theoretical_loss": 3.343769767930582, + "tokens_seen": 3065053184 + }, + { + "epoch": 0.93, + "learning_rate": 5.775727181544635e-05, + "loss": 1.2642, + "theoretical_loss": 3.343764341477366, + "tokens_seen": 3065118720 + }, + { + "epoch": 0.93, + "learning_rate": 5.774122367101305e-05, + "loss": 1.3108, + "theoretical_loss": 3.3437589151726597, + "tokens_seen": 3065184256 + }, + { + "epoch": 0.93, + "learning_rate": 5.772517552657974e-05, + "loss": 1.2549, + "theoretical_loss": 3.3437534890164544, + "tokens_seen": 3065249792 + }, + { + "epoch": 0.93, + "learning_rate": 5.770912738214644e-05, + "loss": 1.2041, + "theoretical_loss": 3.343748063008744, + "tokens_seen": 3065315328 + }, + { + "epoch": 0.93, + "learning_rate": 5.769307923771314e-05, + "loss": 1.2171, + "theoretical_loss": 3.3437426371495214, + "tokens_seen": 3065380864 + }, + { + "epoch": 0.93, + "learning_rate": 5.7677031093279844e-05, + "loss": 1.2287, + "theoretical_loss": 3.343737211438778, + "tokens_seen": 3065446400 + }, + { + "epoch": 0.93, + "learning_rate": 5.7660982948846544e-05, + "loss": 1.2498, + "theoretical_loss": 3.343731785876508, + "tokens_seen": 3065511936 + }, + { + "epoch": 0.93, + "learning_rate": 5.7644934804413245e-05, + "loss": 1.258, + "theoretical_loss": 3.343726360462703, + "tokens_seen": 3065577472 + }, + { + "epoch": 0.93, + "learning_rate": 5.762888665997994e-05, + "loss": 1.1634, + "theoretical_loss": 3.3437209351973567, + "tokens_seen": 3065643008 + }, + { + "epoch": 0.93, + "learning_rate": 5.761283851554664e-05, + "loss": 1.3075, + "theoretical_loss": 3.3437155100804614, + "tokens_seen": 3065708544 + }, + { + "epoch": 0.93, + "learning_rate": 5.759679037111334e-05, + "loss": 1.2264, + "theoretical_loss": 3.3437100851120096, + "tokens_seen": 3065774080 + }, + { + "epoch": 0.93, + "learning_rate": 5.758074222668004e-05, + "loss": 1.296, + "theoretical_loss": 3.343704660291995, + "tokens_seen": 3065839616 + }, + { + "epoch": 0.93, + "learning_rate": 5.756469408224674e-05, + "loss": 1.2249, + "theoretical_loss": 3.3436992356204094, + "tokens_seen": 3065905152 + }, + { + "epoch": 0.93, + "learning_rate": 5.754864593781344e-05, + "loss": 1.182, + "theoretical_loss": 3.3436938110972463, + "tokens_seen": 3065970688 + }, + { + "epoch": 0.93, + "learning_rate": 5.7532597793380136e-05, + "loss": 1.1766, + "theoretical_loss": 3.343688386722498, + "tokens_seen": 3066036224 + }, + { + "epoch": 0.93, + "learning_rate": 5.751654964894684e-05, + "loss": 1.1461, + "theoretical_loss": 3.343682962496157, + "tokens_seen": 3066101760 + }, + { + "epoch": 0.93, + "learning_rate": 5.750050150451355e-05, + "loss": 1.2565, + "theoretical_loss": 3.3436775384182162, + "tokens_seen": 3066167296 + }, + { + "epoch": 0.93, + "learning_rate": 5.748445336008025e-05, + "loss": 1.2267, + "theoretical_loss": 3.3436721144886694, + "tokens_seen": 3066232832 + }, + { + "epoch": 0.93, + "learning_rate": 5.7468405215646945e-05, + "loss": 1.2533, + "theoretical_loss": 3.3436666907075083, + "tokens_seen": 3066298368 + }, + { + "epoch": 0.93, + "learning_rate": 5.7452357071213646e-05, + "loss": 1.3727, + "theoretical_loss": 3.343661267074726, + "tokens_seen": 3066363904 + }, + { + "epoch": 0.93, + "learning_rate": 5.7436308926780347e-05, + "loss": 1.4442, + "theoretical_loss": 3.343655843590315, + "tokens_seen": 3066429440 + }, + { + "epoch": 0.93, + "learning_rate": 5.742026078234705e-05, + "loss": 1.1951, + "theoretical_loss": 3.3436504202542685, + "tokens_seen": 3066494976 + }, + { + "epoch": 0.93, + "learning_rate": 5.740421263791375e-05, + "loss": 1.274, + "theoretical_loss": 3.343644997066579, + "tokens_seen": 3066560512 + }, + { + "epoch": 0.93, + "learning_rate": 5.738816449348045e-05, + "loss": 1.1933, + "theoretical_loss": 3.343639574027239, + "tokens_seen": 3066626048 + }, + { + "epoch": 0.93, + "learning_rate": 5.737211634904714e-05, + "loss": 1.1933, + "theoretical_loss": 3.343634151136242, + "tokens_seen": 3066691584 + }, + { + "epoch": 0.93, + "learning_rate": 5.735606820461384e-05, + "loss": 1.1666, + "theoretical_loss": 3.3436287283935804, + "tokens_seen": 3066757120 + }, + { + "epoch": 0.93, + "learning_rate": 5.7340020060180544e-05, + "loss": 1.2509, + "theoretical_loss": 3.343623305799247, + "tokens_seen": 3066822656 + }, + { + "epoch": 0.93, + "learning_rate": 5.7323971915747244e-05, + "loss": 1.2366, + "theoretical_loss": 3.343617883353234, + "tokens_seen": 3066888192 + }, + { + "epoch": 0.93, + "learning_rate": 5.7307923771313945e-05, + "loss": 1.2455, + "theoretical_loss": 3.3436124610555353, + "tokens_seen": 3066953728 + }, + { + "epoch": 0.93, + "learning_rate": 5.7291875626880646e-05, + "loss": 1.273, + "theoretical_loss": 3.3436070389061427, + "tokens_seen": 3067019264 + }, + { + "epoch": 0.93, + "objective/train/docs_used": 1114524, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.3610692024230957, + "objective/train/theoretical_loss": 3.3436016169050493, + "objective/train/tokens_used": 3087544800, + "theoretical_loss": 3.3436016169050493, + "tokens_seen": 3067084800 + }, + { + "epoch": 0.93, + "learning_rate": 5.727582748244734e-05, + "loss": 1.3328, + "theoretical_loss": 3.3436016169050493, + "tokens_seen": 3067084800 + }, + { + "epoch": 0.93, + "learning_rate": 5.725977933801404e-05, + "loss": 1.2731, + "theoretical_loss": 3.343596195052248, + "tokens_seen": 3067150336 + }, + { + "epoch": 0.93, + "learning_rate": 5.724373119358074e-05, + "loss": 1.0798, + "theoretical_loss": 3.3435907733477315, + "tokens_seen": 3067215872 + }, + { + "epoch": 0.93, + "learning_rate": 5.722768304914744e-05, + "loss": 1.1783, + "theoretical_loss": 3.343585351791493, + "tokens_seen": 3067281408 + }, + { + "epoch": 0.93, + "learning_rate": 5.721163490471414e-05, + "loss": 1.247, + "theoretical_loss": 3.343579930383524, + "tokens_seen": 3067346944 + }, + { + "epoch": 0.93, + "learning_rate": 5.719558676028085e-05, + "loss": 1.204, + "theoretical_loss": 3.343574509123819, + "tokens_seen": 3067412480 + }, + { + "epoch": 0.93, + "learning_rate": 5.717953861584755e-05, + "loss": 1.2291, + "theoretical_loss": 3.3435690880123694, + "tokens_seen": 3067478016 + }, + { + "epoch": 0.93, + "learning_rate": 5.716349047141425e-05, + "loss": 1.2992, + "theoretical_loss": 3.3435636670491684, + "tokens_seen": 3067543552 + }, + { + "epoch": 0.93, + "learning_rate": 5.714744232698095e-05, + "loss": 1.3133, + "theoretical_loss": 3.3435582462342093, + "tokens_seen": 3067609088 + }, + { + "epoch": 0.93, + "learning_rate": 5.7131394182547645e-05, + "loss": 1.2227, + "theoretical_loss": 3.3435528255674845, + "tokens_seen": 3067674624 + }, + { + "epoch": 0.93, + "learning_rate": 5.7115346038114346e-05, + "loss": 1.1642, + "theoretical_loss": 3.343547405048986, + "tokens_seen": 3067740160 + }, + { + "epoch": 0.93, + "learning_rate": 5.7099297893681046e-05, + "loss": 1.3132, + "theoretical_loss": 3.3435419846787076, + "tokens_seen": 3067805696 + }, + { + "epoch": 0.93, + "learning_rate": 5.708324974924775e-05, + "loss": 1.1773, + "theoretical_loss": 3.343536564456642, + "tokens_seen": 3067871232 + }, + { + "epoch": 0.93, + "learning_rate": 5.706720160481445e-05, + "loss": 1.3163, + "theoretical_loss": 3.3435311443827818, + "tokens_seen": 3067936768 + }, + { + "epoch": 0.93, + "learning_rate": 5.705115346038115e-05, + "loss": 1.3102, + "theoretical_loss": 3.3435257244571197, + "tokens_seen": 3068002304 + }, + { + "epoch": 0.93, + "learning_rate": 5.703510531594784e-05, + "loss": 1.1775, + "theoretical_loss": 3.3435203046796484, + "tokens_seen": 3068067840 + }, + { + "epoch": 0.93, + "learning_rate": 5.701905717151454e-05, + "loss": 1.3348, + "theoretical_loss": 3.343514885050361, + "tokens_seen": 3068133376 + }, + { + "epoch": 0.93, + "learning_rate": 5.700300902708124e-05, + "loss": 1.2768, + "theoretical_loss": 3.34350946556925, + "tokens_seen": 3068198912 + }, + { + "epoch": 0.93, + "learning_rate": 5.6986960882647944e-05, + "loss": 1.2233, + "theoretical_loss": 3.343504046236308, + "tokens_seen": 3068264448 + }, + { + "epoch": 0.93, + "learning_rate": 5.6970912738214645e-05, + "loss": 1.3016, + "theoretical_loss": 3.343498627051528, + "tokens_seen": 3068329984 + }, + { + "epoch": 0.93, + "learning_rate": 5.6954864593781345e-05, + "loss": 1.2667, + "theoretical_loss": 3.343493208014903, + "tokens_seen": 3068395520 + }, + { + "epoch": 0.93, + "learning_rate": 5.693881644934804e-05, + "loss": 1.3132, + "theoretical_loss": 3.343487789126426, + "tokens_seen": 3068461056 + }, + { + "epoch": 0.93, + "learning_rate": 5.692276830491474e-05, + "loss": 1.2357, + "theoretical_loss": 3.3434823703860888, + "tokens_seen": 3068526592 + }, + { + "epoch": 0.93, + "learning_rate": 5.6906720160481454e-05, + "loss": 1.2241, + "theoretical_loss": 3.3434769517938854, + "tokens_seen": 3068592128 + }, + { + "epoch": 0.93, + "learning_rate": 5.6890672016048155e-05, + "loss": 1.3198, + "theoretical_loss": 3.343471533349808, + "tokens_seen": 3068657664 + }, + { + "epoch": 0.93, + "learning_rate": 5.687462387161485e-05, + "loss": 1.1854, + "theoretical_loss": 3.343466115053849, + "tokens_seen": 3068723200 + }, + { + "epoch": 0.93, + "learning_rate": 5.685857572718155e-05, + "loss": 1.2254, + "theoretical_loss": 3.3434606969060017, + "tokens_seen": 3068788736 + }, + { + "epoch": 0.93, + "learning_rate": 5.684252758274825e-05, + "loss": 1.2515, + "theoretical_loss": 3.3434552789062586, + "tokens_seen": 3068854272 + }, + { + "epoch": 0.93, + "learning_rate": 5.682647943831495e-05, + "loss": 1.2652, + "theoretical_loss": 3.3434498610546126, + "tokens_seen": 3068919808 + }, + { + "epoch": 0.93, + "learning_rate": 5.681043129388165e-05, + "loss": 1.2041, + "theoretical_loss": 3.343444443351057, + "tokens_seen": 3068985344 + }, + { + "epoch": 0.93, + "learning_rate": 5.679438314944835e-05, + "loss": 1.1892, + "theoretical_loss": 3.343439025795584, + "tokens_seen": 3069050880 + }, + { + "epoch": 0.93, + "learning_rate": 5.6778335005015045e-05, + "loss": 1.2757, + "theoretical_loss": 3.343433608388186, + "tokens_seen": 3069116416 + }, + { + "epoch": 0.93, + "learning_rate": 5.6762286860581746e-05, + "loss": 1.2597, + "theoretical_loss": 3.343428191128857, + "tokens_seen": 3069181952 + }, + { + "epoch": 0.93, + "learning_rate": 5.674623871614845e-05, + "loss": 1.2441, + "theoretical_loss": 3.3434227740175886, + "tokens_seen": 3069247488 + }, + { + "epoch": 0.93, + "learning_rate": 5.673019057171515e-05, + "loss": 1.1106, + "theoretical_loss": 3.343417357054374, + "tokens_seen": 3069313024 + }, + { + "epoch": 0.93, + "learning_rate": 5.671414242728185e-05, + "loss": 1.2334, + "theoretical_loss": 3.3434119402392066, + "tokens_seen": 3069378560 + }, + { + "epoch": 0.93, + "learning_rate": 5.669809428284855e-05, + "loss": 1.218, + "theoretical_loss": 3.343406523572078, + "tokens_seen": 3069444096 + }, + { + "epoch": 0.93, + "learning_rate": 5.668204613841524e-05, + "loss": 1.2115, + "theoretical_loss": 3.3434011070529825, + "tokens_seen": 3069509632 + }, + { + "epoch": 0.93, + "learning_rate": 5.666599799398194e-05, + "loss": 1.1529, + "theoretical_loss": 3.3433956906819113, + "tokens_seen": 3069575168 + }, + { + "epoch": 0.93, + "learning_rate": 5.6649949849548644e-05, + "loss": 1.2366, + "theoretical_loss": 3.3433902744588586, + "tokens_seen": 3069640704 + }, + { + "epoch": 0.93, + "learning_rate": 5.6633901705115344e-05, + "loss": 1.3754, + "theoretical_loss": 3.343384858383816, + "tokens_seen": 3069706240 + }, + { + "epoch": 0.93, + "learning_rate": 5.6617853560682045e-05, + "loss": 1.2391, + "theoretical_loss": 3.343379442456777, + "tokens_seen": 3069771776 + }, + { + "epoch": 0.93, + "learning_rate": 5.660180541624875e-05, + "loss": 1.2004, + "theoretical_loss": 3.3433740266777345, + "tokens_seen": 3069837312 + }, + { + "epoch": 0.93, + "learning_rate": 5.658575727181545e-05, + "loss": 1.1933, + "theoretical_loss": 3.3433686110466807, + "tokens_seen": 3069902848 + }, + { + "epoch": 0.93, + "learning_rate": 5.6569709127382154e-05, + "loss": 1.3053, + "theoretical_loss": 3.343363195563609, + "tokens_seen": 3069968384 + }, + { + "epoch": 0.93, + "learning_rate": 5.6553660982948854e-05, + "loss": 1.2839, + "theoretical_loss": 3.3433577802285117, + "tokens_seen": 3070033920 + }, + { + "epoch": 0.93, + "learning_rate": 5.6537612838515555e-05, + "loss": 1.3117, + "theoretical_loss": 3.343352365041382, + "tokens_seen": 3070099456 + }, + { + "epoch": 0.93, + "learning_rate": 5.652156469408225e-05, + "loss": 1.2361, + "theoretical_loss": 3.3433469500022124, + "tokens_seen": 3070164992 + }, + { + "epoch": 0.93, + "learning_rate": 5.650551654964895e-05, + "loss": 1.2837, + "theoretical_loss": 3.343341535110996, + "tokens_seen": 3070230528 + }, + { + "epoch": 0.93, + "learning_rate": 5.648946840521565e-05, + "loss": 1.3112, + "theoretical_loss": 3.343336120367725, + "tokens_seen": 3070296064 + }, + { + "epoch": 0.93, + "objective/train/docs_used": 1115728, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.2020659446716309, + "objective/train/theoretical_loss": 3.3433307057723933, + "objective/train/tokens_used": 3090821600, + "theoretical_loss": 3.3433307057723933, + "tokens_seen": 3070361600 + }, + { + "epoch": 0.93, + "learning_rate": 5.647342026078235e-05, + "loss": 1.2389, + "theoretical_loss": 3.3433307057723933, + "tokens_seen": 3070361600 + }, + { + "epoch": 0.93, + "learning_rate": 5.645737211634905e-05, + "loss": 1.2893, + "theoretical_loss": 3.3433252913249927, + "tokens_seen": 3070427136 + }, + { + "epoch": 0.93, + "learning_rate": 5.644132397191575e-05, + "loss": 1.2447, + "theoretical_loss": 3.343319877025516, + "tokens_seen": 3070492672 + }, + { + "epoch": 0.93, + "learning_rate": 5.6425275827482446e-05, + "loss": 1.2836, + "theoretical_loss": 3.3433144628739564, + "tokens_seen": 3070558208 + }, + { + "epoch": 0.93, + "learning_rate": 5.6409227683049147e-05, + "loss": 1.3324, + "theoretical_loss": 3.343309048870307, + "tokens_seen": 3070623744 + }, + { + "epoch": 0.93, + "learning_rate": 5.639317953861585e-05, + "loss": 1.2443, + "theoretical_loss": 3.34330363501456, + "tokens_seen": 3070689280 + }, + { + "epoch": 0.93, + "learning_rate": 5.637713139418255e-05, + "loss": 1.3187, + "theoretical_loss": 3.343298221306709, + "tokens_seen": 3070754816 + }, + { + "epoch": 0.93, + "learning_rate": 5.636108324974925e-05, + "loss": 1.253, + "theoretical_loss": 3.3432928077467454, + "tokens_seen": 3070820352 + }, + { + "epoch": 0.93, + "learning_rate": 5.634503510531595e-05, + "loss": 1.3575, + "theoretical_loss": 3.3432873943346633, + "tokens_seen": 3070885888 + }, + { + "epoch": 0.93, + "learning_rate": 5.632898696088264e-05, + "loss": 1.1492, + "theoretical_loss": 3.343281981070455, + "tokens_seen": 3070951424 + }, + { + "epoch": 0.93, + "learning_rate": 5.6312938816449344e-05, + "loss": 1.296, + "theoretical_loss": 3.3432765679541134, + "tokens_seen": 3071016960 + }, + { + "epoch": 0.93, + "learning_rate": 5.629689067201606e-05, + "loss": 1.2826, + "theoretical_loss": 3.343271154985631, + "tokens_seen": 3071082496 + }, + { + "epoch": 0.93, + "learning_rate": 5.628084252758276e-05, + "loss": 1.2115, + "theoretical_loss": 3.343265742165001, + "tokens_seen": 3071148032 + }, + { + "epoch": 0.93, + "learning_rate": 5.626479438314945e-05, + "loss": 1.0861, + "theoretical_loss": 3.343260329492216, + "tokens_seen": 3071213568 + }, + { + "epoch": 0.93, + "learning_rate": 5.624874623871615e-05, + "loss": 1.224, + "theoretical_loss": 3.343254916967269, + "tokens_seen": 3071279104 + }, + { + "epoch": 0.93, + "learning_rate": 5.6232698094282854e-05, + "loss": 1.1984, + "theoretical_loss": 3.343249504590153, + "tokens_seen": 3071344640 + }, + { + "epoch": 0.93, + "learning_rate": 5.6216649949849554e-05, + "loss": 1.3385, + "theoretical_loss": 3.34324409236086, + "tokens_seen": 3071410176 + }, + { + "epoch": 0.93, + "learning_rate": 5.6200601805416255e-05, + "loss": 1.16, + "theoretical_loss": 3.3432386802793834, + "tokens_seen": 3071475712 + }, + { + "epoch": 0.93, + "learning_rate": 5.6184553660982955e-05, + "loss": 1.3382, + "theoretical_loss": 3.343233268345716, + "tokens_seen": 3071541248 + }, + { + "epoch": 0.93, + "learning_rate": 5.616850551654965e-05, + "loss": 1.2589, + "theoretical_loss": 3.3432278565598508, + "tokens_seen": 3071606784 + }, + { + "epoch": 0.93, + "learning_rate": 5.615245737211635e-05, + "loss": 1.2662, + "theoretical_loss": 3.34322244492178, + "tokens_seen": 3071672320 + }, + { + "epoch": 0.93, + "learning_rate": 5.613640922768305e-05, + "loss": 1.2212, + "theoretical_loss": 3.343217033431497, + "tokens_seen": 3071737856 + }, + { + "epoch": 0.93, + "learning_rate": 5.612036108324975e-05, + "loss": 1.2367, + "theoretical_loss": 3.343211622088994, + "tokens_seen": 3071803392 + }, + { + "epoch": 0.93, + "learning_rate": 5.610431293881645e-05, + "loss": 1.2136, + "theoretical_loss": 3.343206210894264, + "tokens_seen": 3071868928 + }, + { + "epoch": 0.93, + "learning_rate": 5.608826479438315e-05, + "loss": 1.3248, + "theoretical_loss": 3.3432007998473003, + "tokens_seen": 3071934464 + }, + { + "epoch": 0.93, + "learning_rate": 5.6072216649949846e-05, + "loss": 1.2964, + "theoretical_loss": 3.3431953889480956, + "tokens_seen": 3072000000 + }, + { + "epoch": 0.93, + "learning_rate": 5.605616850551655e-05, + "loss": 1.3291, + "theoretical_loss": 3.343189978196642, + "tokens_seen": 3072065536 + }, + { + "epoch": 0.93, + "learning_rate": 5.604012036108325e-05, + "loss": 1.237, + "theoretical_loss": 3.3431845675929335, + "tokens_seen": 3072131072 + }, + { + "epoch": 0.93, + "learning_rate": 5.602407221664995e-05, + "loss": 1.2697, + "theoretical_loss": 3.3431791571369613, + "tokens_seen": 3072196608 + }, + { + "epoch": 0.93, + "learning_rate": 5.600802407221665e-05, + "loss": 1.2274, + "theoretical_loss": 3.34317374682872, + "tokens_seen": 3072262144 + }, + { + "epoch": 0.93, + "learning_rate": 5.5991975927783356e-05, + "loss": 1.2194, + "theoretical_loss": 3.3431683366682012, + "tokens_seen": 3072327680 + }, + { + "epoch": 0.93, + "learning_rate": 5.597592778335006e-05, + "loss": 1.2656, + "theoretical_loss": 3.343162926655398, + "tokens_seen": 3072393216 + }, + { + "epoch": 0.93, + "learning_rate": 5.595987963891676e-05, + "loss": 1.3104, + "theoretical_loss": 3.3431575167903036, + "tokens_seen": 3072458752 + }, + { + "epoch": 0.93, + "learning_rate": 5.594383149448346e-05, + "loss": 1.1897, + "theoretical_loss": 3.3431521070729104, + "tokens_seen": 3072524288 + }, + { + "epoch": 0.93, + "learning_rate": 5.592778335005016e-05, + "loss": 1.2582, + "theoretical_loss": 3.343146697503211, + "tokens_seen": 3072589824 + }, + { + "epoch": 0.93, + "learning_rate": 5.591173520561685e-05, + "loss": 1.2868, + "theoretical_loss": 3.343141288081199, + "tokens_seen": 3072655360 + }, + { + "epoch": 0.93, + "learning_rate": 5.589568706118355e-05, + "loss": 1.2834, + "theoretical_loss": 3.3431358788068666, + "tokens_seen": 3072720896 + }, + { + "epoch": 0.93, + "learning_rate": 5.5879638916750254e-05, + "loss": 1.2455, + "theoretical_loss": 3.3431304696802067, + "tokens_seen": 3072786432 + }, + { + "epoch": 0.93, + "learning_rate": 5.5863590772316955e-05, + "loss": 1.328, + "theoretical_loss": 3.343125060701212, + "tokens_seen": 3072851968 + }, + { + "epoch": 0.93, + "learning_rate": 5.5847542627883655e-05, + "loss": 1.319, + "theoretical_loss": 3.3431196518698756, + "tokens_seen": 3072917504 + }, + { + "epoch": 0.93, + "learning_rate": 5.5831494483450356e-05, + "loss": 1.2617, + "theoretical_loss": 3.3431142431861907, + "tokens_seen": 3072983040 + }, + { + "epoch": 0.93, + "learning_rate": 5.581544633901705e-05, + "loss": 1.266, + "theoretical_loss": 3.3431088346501494, + "tokens_seen": 3073048576 + }, + { + "epoch": 0.93, + "learning_rate": 5.579939819458375e-05, + "loss": 1.2974, + "theoretical_loss": 3.3431034262617443, + "tokens_seen": 3073114112 + }, + { + "epoch": 0.93, + "learning_rate": 5.578335005015045e-05, + "loss": 1.2818, + "theoretical_loss": 3.3430980180209695, + "tokens_seen": 3073179648 + }, + { + "epoch": 0.93, + "learning_rate": 5.576730190571715e-05, + "loss": 1.2284, + "theoretical_loss": 3.3430926099278166, + "tokens_seen": 3073245184 + }, + { + "epoch": 0.93, + "learning_rate": 5.575125376128385e-05, + "loss": 1.2242, + "theoretical_loss": 3.343087201982279, + "tokens_seen": 3073310720 + }, + { + "epoch": 0.93, + "learning_rate": 5.573520561685055e-05, + "loss": 1.3212, + "theoretical_loss": 3.343081794184349, + "tokens_seen": 3073376256 + }, + { + "epoch": 0.93, + "learning_rate": 5.571915747241725e-05, + "loss": 1.1875, + "theoretical_loss": 3.3430763865340203, + "tokens_seen": 3073441792 + }, + { + "epoch": 0.93, + "learning_rate": 5.570310932798396e-05, + "loss": 1.2322, + "theoretical_loss": 3.343070979031285, + "tokens_seen": 3073507328 + }, + { + "epoch": 0.93, + "learning_rate": 5.568706118355066e-05, + "loss": 1.251, + "theoretical_loss": 3.343065571676136, + "tokens_seen": 3073572864 + }, + { + "epoch": 0.93, + "objective/train/docs_used": 1116766, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.0349520444869995, + "objective/train/theoretical_loss": 3.3430601644685662, + "objective/train/tokens_used": 3094098400, + "theoretical_loss": 3.3430601644685662, + "tokens_seen": 3073638400 + }, + { + "epoch": 0.93, + "learning_rate": 5.567101303911736e-05, + "loss": 1.0819, + "theoretical_loss": 3.3430601644685662, + "tokens_seen": 3073638400 + }, + { + "epoch": 0.93, + "learning_rate": 5.5654964894684056e-05, + "loss": 1.1439, + "theoretical_loss": 3.343054757408569, + "tokens_seen": 3073703936 + }, + { + "epoch": 0.93, + "learning_rate": 5.563891675025076e-05, + "loss": 1.2359, + "theoretical_loss": 3.3430493504961363, + "tokens_seen": 3073769472 + }, + { + "epoch": 0.93, + "learning_rate": 5.562286860581746e-05, + "loss": 1.2736, + "theoretical_loss": 3.343043943731262, + "tokens_seen": 3073835008 + }, + { + "epoch": 0.93, + "learning_rate": 5.560682046138416e-05, + "loss": 1.2136, + "theoretical_loss": 3.3430385371139373, + "tokens_seen": 3073900544 + }, + { + "epoch": 0.93, + "learning_rate": 5.559077231695086e-05, + "loss": 1.2215, + "theoretical_loss": 3.3430331306441565, + "tokens_seen": 3073966080 + }, + { + "epoch": 0.93, + "learning_rate": 5.557472417251756e-05, + "loss": 1.2021, + "theoretical_loss": 3.343027724321912, + "tokens_seen": 3074031616 + }, + { + "epoch": 0.93, + "learning_rate": 5.555867602808425e-05, + "loss": 1.2449, + "theoretical_loss": 3.343022318147196, + "tokens_seen": 3074097152 + }, + { + "epoch": 0.93, + "learning_rate": 5.5542627883650954e-05, + "loss": 1.2209, + "theoretical_loss": 3.3430169121200026, + "tokens_seen": 3074162688 + }, + { + "epoch": 0.93, + "learning_rate": 5.5526579739217654e-05, + "loss": 1.1697, + "theoretical_loss": 3.3430115062403236, + "tokens_seen": 3074228224 + }, + { + "epoch": 0.93, + "learning_rate": 5.5510531594784355e-05, + "loss": 1.2574, + "theoretical_loss": 3.3430061005081524, + "tokens_seen": 3074293760 + }, + { + "epoch": 0.93, + "learning_rate": 5.5494483450351056e-05, + "loss": 1.2526, + "theoretical_loss": 3.3430006949234814, + "tokens_seen": 3074359296 + }, + { + "epoch": 0.93, + "learning_rate": 5.5478435305917756e-05, + "loss": 1.4754, + "theoretical_loss": 3.3429952894863035, + "tokens_seen": 3074424832 + }, + { + "epoch": 0.93, + "learning_rate": 5.546238716148445e-05, + "loss": 1.1629, + "theoretical_loss": 3.3429898841966117, + "tokens_seen": 3074490368 + }, + { + "epoch": 0.93, + "learning_rate": 5.544633901705115e-05, + "loss": 1.199, + "theoretical_loss": 3.3429844790543988, + "tokens_seen": 3074555904 + }, + { + "epoch": 0.93, + "learning_rate": 5.543029087261785e-05, + "loss": 1.228, + "theoretical_loss": 3.3429790740596577, + "tokens_seen": 3074621440 + }, + { + "epoch": 0.93, + "learning_rate": 5.541424272818455e-05, + "loss": 1.1928, + "theoretical_loss": 3.342973669212381, + "tokens_seen": 3074686976 + }, + { + "epoch": 0.93, + "learning_rate": 5.539819458375126e-05, + "loss": 1.2641, + "theoretical_loss": 3.3429682645125616, + "tokens_seen": 3074752512 + }, + { + "epoch": 0.93, + "learning_rate": 5.538214643931796e-05, + "loss": 1.2061, + "theoretical_loss": 3.3429628599601924, + "tokens_seen": 3074818048 + }, + { + "epoch": 0.93, + "learning_rate": 5.536609829488466e-05, + "loss": 1.2166, + "theoretical_loss": 3.3429574555552666, + "tokens_seen": 3074883584 + }, + { + "epoch": 0.93, + "learning_rate": 5.535005015045136e-05, + "loss": 1.2413, + "theoretical_loss": 3.342952051297776, + "tokens_seen": 3074949120 + }, + { + "epoch": 0.93, + "learning_rate": 5.533400200601806e-05, + "loss": 1.229, + "theoretical_loss": 3.342946647187715, + "tokens_seen": 3075014656 + }, + { + "epoch": 0.93, + "learning_rate": 5.531795386158476e-05, + "loss": 1.2402, + "theoretical_loss": 3.342941243225075, + "tokens_seen": 3075080192 + }, + { + "epoch": 0.93, + "learning_rate": 5.5301905717151456e-05, + "loss": 1.205, + "theoretical_loss": 3.3429358394098494, + "tokens_seen": 3075145728 + }, + { + "epoch": 0.93, + "learning_rate": 5.528585757271816e-05, + "loss": 1.236, + "theoretical_loss": 3.3429304357420313, + "tokens_seen": 3075211264 + }, + { + "epoch": 0.93, + "learning_rate": 5.526980942828486e-05, + "loss": 1.2231, + "theoretical_loss": 3.342925032221613, + "tokens_seen": 3075276800 + }, + { + "epoch": 0.93, + "learning_rate": 5.525376128385156e-05, + "loss": 1.2909, + "theoretical_loss": 3.3429196288485876, + "tokens_seen": 3075342336 + }, + { + "epoch": 0.93, + "learning_rate": 5.523771313941826e-05, + "loss": 1.1955, + "theoretical_loss": 3.3429142256229483, + "tokens_seen": 3075407872 + }, + { + "epoch": 0.93, + "learning_rate": 5.522166499498496e-05, + "loss": 1.1771, + "theoretical_loss": 3.342908822544687, + "tokens_seen": 3075473408 + }, + { + "epoch": 0.93, + "learning_rate": 5.5205616850551653e-05, + "loss": 1.3007, + "theoretical_loss": 3.3429034196137977, + "tokens_seen": 3075538944 + }, + { + "epoch": 0.93, + "learning_rate": 5.5189568706118354e-05, + "loss": 1.3175, + "theoretical_loss": 3.342898016830272, + "tokens_seen": 3075604480 + }, + { + "epoch": 0.93, + "learning_rate": 5.5173520561685055e-05, + "loss": 1.3372, + "theoretical_loss": 3.342892614194104, + "tokens_seen": 3075670016 + }, + { + "epoch": 0.93, + "learning_rate": 5.5157472417251755e-05, + "loss": 1.2164, + "theoretical_loss": 3.3428872117052855, + "tokens_seen": 3075735552 + }, + { + "epoch": 0.93, + "learning_rate": 5.5141424272818456e-05, + "loss": 1.2159, + "theoretical_loss": 3.34288180936381, + "tokens_seen": 3075801088 + }, + { + "epoch": 0.93, + "learning_rate": 5.512537612838516e-05, + "loss": 1.1083, + "theoretical_loss": 3.3428764071696704, + "tokens_seen": 3075866624 + }, + { + "epoch": 0.93, + "learning_rate": 5.510932798395185e-05, + "loss": 1.1822, + "theoretical_loss": 3.3428710051228587, + "tokens_seen": 3075932160 + }, + { + "epoch": 0.93, + "learning_rate": 5.5093279839518565e-05, + "loss": 1.2519, + "theoretical_loss": 3.342865603223369, + "tokens_seen": 3075997696 + }, + { + "epoch": 0.93, + "learning_rate": 5.5077231695085265e-05, + "loss": 1.2434, + "theoretical_loss": 3.3428602014711926, + "tokens_seen": 3076063232 + }, + { + "epoch": 0.93, + "learning_rate": 5.506118355065196e-05, + "loss": 1.2346, + "theoretical_loss": 3.342854799866324, + "tokens_seen": 3076128768 + }, + { + "epoch": 0.93, + "learning_rate": 5.504513540621866e-05, + "loss": 1.2509, + "theoretical_loss": 3.3428493984087546, + "tokens_seen": 3076194304 + }, + { + "epoch": 0.93, + "learning_rate": 5.502908726178536e-05, + "loss": 1.2029, + "theoretical_loss": 3.3428439970984782, + "tokens_seen": 3076259840 + }, + { + "epoch": 0.93, + "learning_rate": 5.501303911735206e-05, + "loss": 1.071, + "theoretical_loss": 3.342838595935487, + "tokens_seen": 3076325376 + }, + { + "epoch": 0.93, + "learning_rate": 5.499699097291876e-05, + "loss": 1.2244, + "theoretical_loss": 3.3428331949197747, + "tokens_seen": 3076390912 + }, + { + "epoch": 0.93, + "learning_rate": 5.498094282848546e-05, + "loss": 1.2065, + "theoretical_loss": 3.3428277940513333, + "tokens_seen": 3076456448 + }, + { + "epoch": 0.93, + "learning_rate": 5.4964894684052156e-05, + "loss": 1.2645, + "theoretical_loss": 3.342822393330156, + "tokens_seen": 3076521984 + }, + { + "epoch": 0.93, + "learning_rate": 5.494884653961886e-05, + "loss": 1.3102, + "theoretical_loss": 3.342816992756236, + "tokens_seen": 3076587520 + }, + { + "epoch": 0.93, + "learning_rate": 5.493279839518556e-05, + "loss": 1.2092, + "theoretical_loss": 3.342811592329565, + "tokens_seen": 3076653056 + }, + { + "epoch": 0.93, + "learning_rate": 5.491675025075226e-05, + "loss": 1.1664, + "theoretical_loss": 3.3428061920501375, + "tokens_seen": 3076718592 + }, + { + "epoch": 0.93, + "learning_rate": 5.490070210631896e-05, + "loss": 1.2799, + "theoretical_loss": 3.342800791917945, + "tokens_seen": 3076784128 + }, + { + "epoch": 0.93, + "learning_rate": 5.488465396188566e-05, + "loss": 1.3335, + "theoretical_loss": 3.342795391932981, + "tokens_seen": 3076849664 + }, + { + "epoch": 0.93, + "objective/train/docs_used": 1117976, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.2596763372421265, + "objective/train/theoretical_loss": 3.3427899920952377, + "objective/train/tokens_used": 3097375200, + "theoretical_loss": 3.3427899920952377, + "tokens_seen": 3076915200 + }, + { + "epoch": 0.93, + "learning_rate": 5.486860581745235e-05, + "loss": 1.2389, + "theoretical_loss": 3.3427899920952377, + "tokens_seen": 3076915200 + }, + { + "epoch": 0.93, + "learning_rate": 5.4852557673019054e-05, + "loss": 1.2962, + "theoretical_loss": 3.342784592404709, + "tokens_seen": 3076980736 + }, + { + "epoch": 0.93, + "learning_rate": 5.4836509528585755e-05, + "loss": 1.171, + "theoretical_loss": 3.3427791928613866, + "tokens_seen": 3077046272 + }, + { + "epoch": 0.93, + "learning_rate": 5.4820461384152455e-05, + "loss": 1.2159, + "theoretical_loss": 3.3427737934652644, + "tokens_seen": 3077111808 + }, + { + "epoch": 0.93, + "learning_rate": 5.480441323971916e-05, + "loss": 1.2411, + "theoretical_loss": 3.3427683942163346, + "tokens_seen": 3077177344 + }, + { + "epoch": 0.93, + "learning_rate": 5.478836509528586e-05, + "loss": 1.2199, + "theoretical_loss": 3.3427629951145903, + "tokens_seen": 3077242880 + }, + { + "epoch": 0.93, + "learning_rate": 5.4772316950852564e-05, + "loss": 1.298, + "theoretical_loss": 3.342757596160024, + "tokens_seen": 3077308416 + }, + { + "epoch": 0.93, + "learning_rate": 5.4756268806419264e-05, + "loss": 1.2866, + "theoretical_loss": 3.3427521973526293, + "tokens_seen": 3077373952 + }, + { + "epoch": 0.93, + "learning_rate": 5.4740220661985965e-05, + "loss": 1.3633, + "theoretical_loss": 3.3427467986923984, + "tokens_seen": 3077439488 + }, + { + "epoch": 0.93, + "learning_rate": 5.4724172517552666e-05, + "loss": 1.2408, + "theoretical_loss": 3.342741400179324, + "tokens_seen": 3077505024 + }, + { + "epoch": 0.93, + "learning_rate": 5.470812437311936e-05, + "loss": 1.2413, + "theoretical_loss": 3.3427360018133996, + "tokens_seen": 3077570560 + }, + { + "epoch": 0.93, + "learning_rate": 5.469207622868606e-05, + "loss": 1.3743, + "theoretical_loss": 3.342730603594618, + "tokens_seen": 3077636096 + }, + { + "epoch": 0.93, + "learning_rate": 5.467602808425276e-05, + "loss": 1.2595, + "theoretical_loss": 3.3427252055229713, + "tokens_seen": 3077701632 + }, + { + "epoch": 0.93, + "learning_rate": 5.465997993981946e-05, + "loss": 1.1626, + "theoretical_loss": 3.342719807598453, + "tokens_seen": 3077767168 + }, + { + "epoch": 0.93, + "learning_rate": 5.464393179538616e-05, + "loss": 1.3563, + "theoretical_loss": 3.342714409821056, + "tokens_seen": 3077832704 + }, + { + "epoch": 0.93, + "learning_rate": 5.462788365095286e-05, + "loss": 1.3126, + "theoretical_loss": 3.342709012190773, + "tokens_seen": 3077898240 + }, + { + "epoch": 0.93, + "learning_rate": 5.461183550651956e-05, + "loss": 1.3067, + "theoretical_loss": 3.3427036147075966, + "tokens_seen": 3077963776 + }, + { + "epoch": 0.93, + "learning_rate": 5.459578736208626e-05, + "loss": 1.2282, + "theoretical_loss": 3.3426982173715203, + "tokens_seen": 3078029312 + }, + { + "epoch": 0.93, + "learning_rate": 5.457973921765296e-05, + "loss": 1.2346, + "theoretical_loss": 3.342692820182536, + "tokens_seen": 3078094848 + }, + { + "epoch": 0.93, + "learning_rate": 5.456369107321966e-05, + "loss": 1.3703, + "theoretical_loss": 3.3426874231406374, + "tokens_seen": 3078160384 + }, + { + "epoch": 0.93, + "learning_rate": 5.454764292878636e-05, + "loss": 1.3036, + "theoretical_loss": 3.342682026245817, + "tokens_seen": 3078225920 + }, + { + "epoch": 0.93, + "learning_rate": 5.453159478435306e-05, + "loss": 1.2735, + "theoretical_loss": 3.342676629498068, + "tokens_seen": 3078291456 + }, + { + "epoch": 0.93, + "learning_rate": 5.4515546639919754e-05, + "loss": 1.2106, + "theoretical_loss": 3.3426712328973824, + "tokens_seen": 3078356992 + }, + { + "epoch": 0.93, + "learning_rate": 5.449949849548647e-05, + "loss": 1.2961, + "theoretical_loss": 3.3426658364437545, + "tokens_seen": 3078422528 + }, + { + "epoch": 0.93, + "learning_rate": 5.448345035105317e-05, + "loss": 1.1368, + "theoretical_loss": 3.3426604401371756, + "tokens_seen": 3078488064 + }, + { + "epoch": 0.93, + "learning_rate": 5.446740220661987e-05, + "loss": 1.1763, + "theoretical_loss": 3.34265504397764, + "tokens_seen": 3078553600 + }, + { + "epoch": 0.93, + "learning_rate": 5.445135406218656e-05, + "loss": 1.2172, + "theoretical_loss": 3.342649647965139, + "tokens_seen": 3078619136 + }, + { + "epoch": 0.93, + "learning_rate": 5.4435305917753264e-05, + "loss": 1.3657, + "theoretical_loss": 3.342644252099667, + "tokens_seen": 3078684672 + }, + { + "epoch": 0.93, + "learning_rate": 5.4419257773319964e-05, + "loss": 1.1946, + "theoretical_loss": 3.3426388563812157, + "tokens_seen": 3078750208 + }, + { + "epoch": 0.93, + "learning_rate": 5.4403209628886665e-05, + "loss": 1.2663, + "theoretical_loss": 3.3426334608097785, + "tokens_seen": 3078815744 + }, + { + "epoch": 0.93, + "learning_rate": 5.4387161484453365e-05, + "loss": 1.2806, + "theoretical_loss": 3.342628065385348, + "tokens_seen": 3078881280 + }, + { + "epoch": 0.93, + "learning_rate": 5.4371113340020066e-05, + "loss": 1.2086, + "theoretical_loss": 3.342622670107918, + "tokens_seen": 3078946816 + }, + { + "epoch": 0.93, + "learning_rate": 5.435506519558676e-05, + "loss": 1.3598, + "theoretical_loss": 3.34261727497748, + "tokens_seen": 3079012352 + }, + { + "epoch": 0.93, + "learning_rate": 5.433901705115346e-05, + "loss": 1.1957, + "theoretical_loss": 3.342611879994028, + "tokens_seen": 3079077888 + }, + { + "epoch": 0.93, + "learning_rate": 5.432296890672016e-05, + "loss": 1.2397, + "theoretical_loss": 3.342606485157554, + "tokens_seen": 3079143424 + }, + { + "epoch": 0.93, + "learning_rate": 5.430692076228686e-05, + "loss": 1.205, + "theoretical_loss": 3.3426010904680514, + "tokens_seen": 3079208960 + }, + { + "epoch": 0.93, + "learning_rate": 5.429087261785356e-05, + "loss": 1.1684, + "theoretical_loss": 3.342595695925513, + "tokens_seen": 3079274496 + }, + { + "epoch": 0.93, + "learning_rate": 5.427482447342026e-05, + "loss": 1.2303, + "theoretical_loss": 3.342590301529931, + "tokens_seen": 3079340032 + }, + { + "epoch": 0.93, + "learning_rate": 5.425877632898696e-05, + "loss": 1.2086, + "theoretical_loss": 3.3425849072812994, + "tokens_seen": 3079405568 + }, + { + "epoch": 0.93, + "learning_rate": 5.424272818455366e-05, + "loss": 1.2574, + "theoretical_loss": 3.3425795131796106, + "tokens_seen": 3079471104 + }, + { + "epoch": 0.93, + "learning_rate": 5.422668004012036e-05, + "loss": 1.212, + "theoretical_loss": 3.342574119224857, + "tokens_seen": 3079536640 + }, + { + "epoch": 0.93, + "learning_rate": 5.421063189568706e-05, + "loss": 1.1632, + "theoretical_loss": 3.342568725417032, + "tokens_seen": 3079602176 + }, + { + "epoch": 0.93, + "learning_rate": 5.4194583751253766e-05, + "loss": 1.3005, + "theoretical_loss": 3.3425633317561285, + "tokens_seen": 3079667712 + }, + { + "epoch": 0.93, + "learning_rate": 5.417853560682047e-05, + "loss": 1.2468, + "theoretical_loss": 3.3425579382421393, + "tokens_seen": 3079733248 + }, + { + "epoch": 0.93, + "learning_rate": 5.416248746238717e-05, + "loss": 1.2825, + "theoretical_loss": 3.3425525448750566, + "tokens_seen": 3079798784 + }, + { + "epoch": 0.93, + "learning_rate": 5.414643931795387e-05, + "loss": 1.1873, + "theoretical_loss": 3.342547151654874, + "tokens_seen": 3079864320 + }, + { + "epoch": 0.93, + "learning_rate": 5.413039117352057e-05, + "loss": 1.2333, + "theoretical_loss": 3.342541758581585, + "tokens_seen": 3079929856 + }, + { + "epoch": 0.93, + "learning_rate": 5.411434302908727e-05, + "loss": 1.2151, + "theoretical_loss": 3.3425363656551808, + "tokens_seen": 3079995392 + }, + { + "epoch": 0.93, + "learning_rate": 5.409829488465396e-05, + "loss": 1.2567, + "theoretical_loss": 3.342530972875655, + "tokens_seen": 3080060928 + }, + { + "epoch": 0.93, + "learning_rate": 5.4082246740220664e-05, + "loss": 1.1513, + "theoretical_loss": 3.3425255802430014, + "tokens_seen": 3080126464 + }, + { + "epoch": 0.93, + "objective/train/docs_used": 1119127, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.0942118167877197, + "objective/train/theoretical_loss": 3.342520187757212, + "objective/train/tokens_used": 3100652000, + "theoretical_loss": 3.342520187757212, + "tokens_seen": 3080192000 + }, + { + "epoch": 0.93, + "learning_rate": 5.4066198595787365e-05, + "loss": 1.1572, + "theoretical_loss": 3.342520187757212, + "tokens_seen": 3080192000 + }, + { + "epoch": 0.93, + "learning_rate": 5.4050150451354065e-05, + "loss": 1.295, + "theoretical_loss": 3.3425147954182792, + "tokens_seen": 3080257536 + }, + { + "epoch": 0.93, + "learning_rate": 5.4034102306920766e-05, + "loss": 1.155, + "theoretical_loss": 3.342509403226197, + "tokens_seen": 3080323072 + }, + { + "epoch": 0.93, + "learning_rate": 5.4018054162487467e-05, + "loss": 1.2672, + "theoretical_loss": 3.3425040111809574, + "tokens_seen": 3080388608 + }, + { + "epoch": 0.93, + "learning_rate": 5.400200601805416e-05, + "loss": 1.2934, + "theoretical_loss": 3.342498619282554, + "tokens_seen": 3080454144 + }, + { + "epoch": 0.93, + "learning_rate": 5.398595787362086e-05, + "loss": 1.3242, + "theoretical_loss": 3.342493227530979, + "tokens_seen": 3080519680 + }, + { + "epoch": 0.93, + "learning_rate": 5.396990972918756e-05, + "loss": 1.3055, + "theoretical_loss": 3.3424878359262253, + "tokens_seen": 3080585216 + }, + { + "epoch": 0.93, + "learning_rate": 5.395386158475426e-05, + "loss": 1.2744, + "theoretical_loss": 3.3424824444682866, + "tokens_seen": 3080650752 + }, + { + "epoch": 0.93, + "learning_rate": 5.393781344032096e-05, + "loss": 1.2328, + "theoretical_loss": 3.342477053157155, + "tokens_seen": 3080716288 + }, + { + "epoch": 0.93, + "learning_rate": 5.3921765295887664e-05, + "loss": 1.0937, + "theoretical_loss": 3.342471661992824, + "tokens_seen": 3080781824 + }, + { + "epoch": 0.93, + "learning_rate": 5.390571715145436e-05, + "loss": 1.2558, + "theoretical_loss": 3.3424662709752857, + "tokens_seen": 3080847360 + }, + { + "epoch": 0.93, + "learning_rate": 5.388966900702107e-05, + "loss": 1.1707, + "theoretical_loss": 3.3424608801045332, + "tokens_seen": 3080912896 + }, + { + "epoch": 0.93, + "learning_rate": 5.387362086258777e-05, + "loss": 1.2351, + "theoretical_loss": 3.34245548938056, + "tokens_seen": 3080978432 + }, + { + "epoch": 0.93, + "learning_rate": 5.385757271815447e-05, + "loss": 1.2464, + "theoretical_loss": 3.342450098803358, + "tokens_seen": 3081043968 + }, + { + "epoch": 0.93, + "learning_rate": 5.384152457372117e-05, + "loss": 1.2949, + "theoretical_loss": 3.3424447083729207, + "tokens_seen": 3081109504 + }, + { + "epoch": 0.93, + "learning_rate": 5.382547642928787e-05, + "loss": 1.2389, + "theoretical_loss": 3.3424393180892413, + "tokens_seen": 3081175040 + }, + { + "epoch": 0.93, + "learning_rate": 5.380942828485457e-05, + "loss": 1.2593, + "theoretical_loss": 3.342433927952312, + "tokens_seen": 3081240576 + }, + { + "epoch": 0.93, + "learning_rate": 5.379338014042127e-05, + "loss": 1.1848, + "theoretical_loss": 3.342428537962126, + "tokens_seen": 3081306112 + }, + { + "epoch": 0.93, + "learning_rate": 5.377733199598797e-05, + "loss": 1.182, + "theoretical_loss": 3.3424231481186766, + "tokens_seen": 3081371648 + }, + { + "epoch": 0.93, + "learning_rate": 5.376128385155467e-05, + "loss": 1.1951, + "theoretical_loss": 3.3424177584219557, + "tokens_seen": 3081437184 + }, + { + "epoch": 0.93, + "learning_rate": 5.3745235707121364e-05, + "loss": 1.1305, + "theoretical_loss": 3.3424123688719565, + "tokens_seen": 3081502720 + }, + { + "epoch": 0.93, + "learning_rate": 5.3729187562688064e-05, + "loss": 1.395, + "theoretical_loss": 3.342406979468673, + "tokens_seen": 3081568256 + }, + { + "epoch": 0.93, + "learning_rate": 5.3713139418254765e-05, + "loss": 1.2405, + "theoretical_loss": 3.3424015902120963, + "tokens_seen": 3081633792 + }, + { + "epoch": 0.93, + "learning_rate": 5.3697091273821466e-05, + "loss": 1.2396, + "theoretical_loss": 3.3423962011022206, + "tokens_seen": 3081699328 + }, + { + "epoch": 0.93, + "learning_rate": 5.3681043129388166e-05, + "loss": 1.2527, + "theoretical_loss": 3.342390812139038, + "tokens_seen": 3081764864 + }, + { + "epoch": 0.93, + "learning_rate": 5.366499498495487e-05, + "loss": 1.1605, + "theoretical_loss": 3.342385423322542, + "tokens_seen": 3081830400 + }, + { + "epoch": 0.93, + "learning_rate": 5.364894684052156e-05, + "loss": 1.2051, + "theoretical_loss": 3.3423800346527255, + "tokens_seen": 3081895936 + }, + { + "epoch": 0.93, + "learning_rate": 5.363289869608826e-05, + "loss": 1.1647, + "theoretical_loss": 3.342374646129581, + "tokens_seen": 3081961472 + }, + { + "epoch": 0.93, + "learning_rate": 5.361685055165496e-05, + "loss": 1.2098, + "theoretical_loss": 3.3423692577531012, + "tokens_seen": 3082027008 + }, + { + "epoch": 0.93, + "learning_rate": 5.3600802407221676e-05, + "loss": 1.2798, + "theoretical_loss": 3.34236386952328, + "tokens_seen": 3082092544 + }, + { + "epoch": 0.93, + "learning_rate": 5.358475426278837e-05, + "loss": 1.2618, + "theoretical_loss": 3.3423584814401086, + "tokens_seen": 3082158080 + }, + { + "epoch": 0.93, + "learning_rate": 5.356870611835507e-05, + "loss": 1.2452, + "theoretical_loss": 3.3423530935035815, + "tokens_seen": 3082223616 + }, + { + "epoch": 0.93, + "learning_rate": 5.355265797392177e-05, + "loss": 1.1805, + "theoretical_loss": 3.342347705713691, + "tokens_seen": 3082289152 + }, + { + "epoch": 0.93, + "learning_rate": 5.353660982948847e-05, + "loss": 1.0814, + "theoretical_loss": 3.34234231807043, + "tokens_seen": 3082354688 + }, + { + "epoch": 0.93, + "learning_rate": 5.352056168505517e-05, + "loss": 1.1357, + "theoretical_loss": 3.342336930573791, + "tokens_seen": 3082420224 + }, + { + "epoch": 0.93, + "learning_rate": 5.350451354062187e-05, + "loss": 1.1563, + "theoretical_loss": 3.3423315432237675, + "tokens_seen": 3082485760 + }, + { + "epoch": 0.93, + "learning_rate": 5.348846539618857e-05, + "loss": 1.4016, + "theoretical_loss": 3.342326156020352, + "tokens_seen": 3082551296 + }, + { + "epoch": 0.93, + "learning_rate": 5.347241725175527e-05, + "loss": 1.2054, + "theoretical_loss": 3.342320768963538, + "tokens_seen": 3082616832 + }, + { + "epoch": 0.93, + "learning_rate": 5.345636910732197e-05, + "loss": 1.1861, + "theoretical_loss": 3.342315382053317, + "tokens_seen": 3082682368 + }, + { + "epoch": 0.93, + "learning_rate": 5.344032096288867e-05, + "loss": 1.2381, + "theoretical_loss": 3.3423099952896838, + "tokens_seen": 3082747904 + }, + { + "epoch": 0.93, + "learning_rate": 5.342427281845537e-05, + "loss": 1.191, + "theoretical_loss": 3.34230460867263, + "tokens_seen": 3082813440 + }, + { + "epoch": 0.93, + "learning_rate": 5.340822467402207e-05, + "loss": 1.1872, + "theoretical_loss": 3.342299222202149, + "tokens_seen": 3082878976 + }, + { + "epoch": 0.93, + "learning_rate": 5.3392176529588764e-05, + "loss": 1.2735, + "theoretical_loss": 3.342293835878233, + "tokens_seen": 3082944512 + }, + { + "epoch": 0.93, + "learning_rate": 5.3376128385155465e-05, + "loss": 1.2502, + "theoretical_loss": 3.3422884497008756, + "tokens_seen": 3083010048 + }, + { + "epoch": 0.93, + "learning_rate": 5.3360080240722165e-05, + "loss": 1.2726, + "theoretical_loss": 3.3422830636700698, + "tokens_seen": 3083075584 + }, + { + "epoch": 0.93, + "learning_rate": 5.3344032096288866e-05, + "loss": 1.3682, + "theoretical_loss": 3.342277677785808, + "tokens_seen": 3083141120 + }, + { + "epoch": 0.93, + "learning_rate": 5.332798395185557e-05, + "loss": 1.2748, + "theoretical_loss": 3.3422722920480834, + "tokens_seen": 3083206656 + }, + { + "epoch": 0.93, + "learning_rate": 5.331193580742227e-05, + "loss": 1.2978, + "theoretical_loss": 3.3422669064568886, + "tokens_seen": 3083272192 + }, + { + "epoch": 0.93, + "learning_rate": 5.3295887662988975e-05, + "loss": 1.2091, + "theoretical_loss": 3.342261521012217, + "tokens_seen": 3083337728 + }, + { + "epoch": 0.93, + "learning_rate": 5.3279839518555675e-05, + "loss": 1.184, + "theoretical_loss": 3.3422561357140608, + "tokens_seen": 3083403264 + }, + { + "epoch": 0.93, + "objective/train/docs_used": 1120337, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.2360115051269531, + "objective/train/theoretical_loss": 3.3422507505624135, + "objective/train/tokens_used": 3103928800, + "theoretical_loss": 3.3422507505624135, + "tokens_seen": 3083468800 + }, + { + "epoch": 0.93, + "learning_rate": 5.3263791374122376e-05, + "loss": 1.2252, + "theoretical_loss": 3.3422507505624135, + "tokens_seen": 3083468800 + }, + { + "epoch": 0.93, + "learning_rate": 5.324774322968907e-05, + "loss": 1.2391, + "theoretical_loss": 3.342245365557268, + "tokens_seen": 3083534336 + }, + { + "epoch": 0.93, + "learning_rate": 5.323169508525577e-05, + "loss": 1.2372, + "theoretical_loss": 3.3422399806986167, + "tokens_seen": 3083599872 + }, + { + "epoch": 0.93, + "learning_rate": 5.321564694082247e-05, + "loss": 1.2393, + "theoretical_loss": 3.342234595986453, + "tokens_seen": 3083665408 + }, + { + "epoch": 0.93, + "learning_rate": 5.319959879638917e-05, + "loss": 1.1463, + "theoretical_loss": 3.3422292114207695, + "tokens_seen": 3083730944 + }, + { + "epoch": 0.93, + "learning_rate": 5.318355065195587e-05, + "loss": 1.087, + "theoretical_loss": 3.3422238270015594, + "tokens_seen": 3083796480 + }, + { + "epoch": 0.93, + "learning_rate": 5.316750250752257e-05, + "loss": 1.3054, + "theoretical_loss": 3.3422184427288153, + "tokens_seen": 3083862016 + }, + { + "epoch": 0.93, + "learning_rate": 5.315145436308927e-05, + "loss": 1.227, + "theoretical_loss": 3.34221305860253, + "tokens_seen": 3083927552 + }, + { + "epoch": 0.93, + "learning_rate": 5.313540621865597e-05, + "loss": 1.2059, + "theoretical_loss": 3.3422076746226965, + "tokens_seen": 3083993088 + }, + { + "epoch": 0.93, + "learning_rate": 5.311935807422267e-05, + "loss": 1.2839, + "theoretical_loss": 3.3422022907893085, + "tokens_seen": 3084058624 + }, + { + "epoch": 0.93, + "learning_rate": 5.310330992978937e-05, + "loss": 1.2429, + "theoretical_loss": 3.3421969071023576, + "tokens_seen": 3084124160 + }, + { + "epoch": 0.93, + "learning_rate": 5.308726178535607e-05, + "loss": 1.2149, + "theoretical_loss": 3.3421915235618376, + "tokens_seen": 3084189696 + }, + { + "epoch": 0.93, + "learning_rate": 5.307121364092277e-05, + "loss": 1.2992, + "theoretical_loss": 3.3421861401677413, + "tokens_seen": 3084255232 + }, + { + "epoch": 0.93, + "learning_rate": 5.3055165496489464e-05, + "loss": 1.3225, + "theoretical_loss": 3.3421807569200612, + "tokens_seen": 3084320768 + }, + { + "epoch": 0.93, + "learning_rate": 5.3039117352056165e-05, + "loss": 1.1988, + "theoretical_loss": 3.3421753738187903, + "tokens_seen": 3084386304 + }, + { + "epoch": 0.93, + "learning_rate": 5.3023069207622865e-05, + "loss": 1.2154, + "theoretical_loss": 3.342169990863922, + "tokens_seen": 3084451840 + }, + { + "epoch": 0.93, + "learning_rate": 5.3007021063189566e-05, + "loss": 1.2293, + "theoretical_loss": 3.342164608055449, + "tokens_seen": 3084517376 + }, + { + "epoch": 0.93, + "learning_rate": 5.299097291875627e-05, + "loss": 1.224, + "theoretical_loss": 3.3421592253933636, + "tokens_seen": 3084582912 + }, + { + "epoch": 0.93, + "learning_rate": 5.2974924774322974e-05, + "loss": 1.2596, + "theoretical_loss": 3.3421538428776594, + "tokens_seen": 3084648448 + }, + { + "epoch": 0.93, + "learning_rate": 5.2958876629889675e-05, + "loss": 1.2025, + "theoretical_loss": 3.342148460508329, + "tokens_seen": 3084713984 + }, + { + "epoch": 0.93, + "learning_rate": 5.2942828485456375e-05, + "loss": 1.2221, + "theoretical_loss": 3.342143078285366, + "tokens_seen": 3084779520 + }, + { + "epoch": 0.93, + "learning_rate": 5.2926780341023076e-05, + "loss": 1.2228, + "theoretical_loss": 3.342137696208762, + "tokens_seen": 3084845056 + }, + { + "epoch": 0.93, + "learning_rate": 5.2910732196589776e-05, + "loss": 1.2296, + "theoretical_loss": 3.3421323142785107, + "tokens_seen": 3084910592 + }, + { + "epoch": 0.93, + "learning_rate": 5.289468405215647e-05, + "loss": 1.3025, + "theoretical_loss": 3.342126932494605, + "tokens_seen": 3084976128 + }, + { + "epoch": 0.93, + "learning_rate": 5.287863590772317e-05, + "loss": 1.1482, + "theoretical_loss": 3.342121550857038, + "tokens_seen": 3085041664 + }, + { + "epoch": 0.93, + "learning_rate": 5.286258776328987e-05, + "loss": 1.2575, + "theoretical_loss": 3.342116169365802, + "tokens_seen": 3085107200 + }, + { + "epoch": 0.93, + "learning_rate": 5.284653961885657e-05, + "loss": 1.1557, + "theoretical_loss": 3.3421107880208902, + "tokens_seen": 3085172736 + }, + { + "epoch": 0.93, + "learning_rate": 5.283049147442327e-05, + "loss": 1.2077, + "theoretical_loss": 3.342105406822296, + "tokens_seen": 3085238272 + }, + { + "epoch": 0.93, + "learning_rate": 5.2814443329989973e-05, + "loss": 1.2312, + "theoretical_loss": 3.3421000257700118, + "tokens_seen": 3085303808 + }, + { + "epoch": 0.93, + "learning_rate": 5.279839518555667e-05, + "loss": 1.3544, + "theoretical_loss": 3.3420946448640305, + "tokens_seen": 3085369344 + }, + { + "epoch": 0.93, + "learning_rate": 5.278234704112337e-05, + "loss": 1.2195, + "theoretical_loss": 3.342089264104345, + "tokens_seen": 3085434880 + }, + { + "epoch": 0.94, + "learning_rate": 5.276629889669007e-05, + "loss": 1.1385, + "theoretical_loss": 3.3420838834909485, + "tokens_seen": 3085500416 + }, + { + "epoch": 0.94, + "learning_rate": 5.275025075225677e-05, + "loss": 1.1858, + "theoretical_loss": 3.342078503023834, + "tokens_seen": 3085565952 + }, + { + "epoch": 0.94, + "learning_rate": 5.273420260782347e-05, + "loss": 1.3256, + "theoretical_loss": 3.342073122702994, + "tokens_seen": 3085631488 + }, + { + "epoch": 0.94, + "learning_rate": 5.271815446339017e-05, + "loss": 1.3791, + "theoretical_loss": 3.3420677425284215, + "tokens_seen": 3085697024 + }, + { + "epoch": 0.94, + "learning_rate": 5.270210631895688e-05, + "loss": 1.2344, + "theoretical_loss": 3.3420623625001094, + "tokens_seen": 3085762560 + }, + { + "epoch": 0.94, + "learning_rate": 5.268605817452358e-05, + "loss": 1.2414, + "theoretical_loss": 3.342056982618051, + "tokens_seen": 3085828096 + }, + { + "epoch": 0.94, + "learning_rate": 5.267001003009028e-05, + "loss": 1.3237, + "theoretical_loss": 3.342051602882239, + "tokens_seen": 3085893632 + }, + { + "epoch": 0.94, + "learning_rate": 5.265396188565698e-05, + "loss": 1.2111, + "theoretical_loss": 3.342046223292666, + "tokens_seen": 3085959168 + }, + { + "epoch": 0.94, + "learning_rate": 5.2637913741223674e-05, + "loss": 1.2627, + "theoretical_loss": 3.3420408438493254, + "tokens_seen": 3086024704 + }, + { + "epoch": 0.94, + "learning_rate": 5.2621865596790374e-05, + "loss": 1.3427, + "theoretical_loss": 3.3420354645522097, + "tokens_seen": 3086090240 + }, + { + "epoch": 0.94, + "learning_rate": 5.2605817452357075e-05, + "loss": 1.1844, + "theoretical_loss": 3.342030085401312, + "tokens_seen": 3086155776 + }, + { + "epoch": 0.94, + "learning_rate": 5.2589769307923776e-05, + "loss": 1.2584, + "theoretical_loss": 3.3420247063966255, + "tokens_seen": 3086221312 + }, + { + "epoch": 0.94, + "learning_rate": 5.2573721163490476e-05, + "loss": 1.3122, + "theoretical_loss": 3.3420193275381425, + "tokens_seen": 3086286848 + }, + { + "epoch": 0.94, + "learning_rate": 5.255767301905718e-05, + "loss": 1.2076, + "theoretical_loss": 3.3420139488258567, + "tokens_seen": 3086352384 + }, + { + "epoch": 0.94, + "learning_rate": 5.254162487462387e-05, + "loss": 1.3052, + "theoretical_loss": 3.34200857025976, + "tokens_seen": 3086417920 + }, + { + "epoch": 0.94, + "learning_rate": 5.252557673019057e-05, + "loss": 1.3316, + "theoretical_loss": 3.3420031918398467, + "tokens_seen": 3086483456 + }, + { + "epoch": 0.94, + "learning_rate": 5.250952858575727e-05, + "loss": 1.2134, + "theoretical_loss": 3.3419978135661084, + "tokens_seen": 3086548992 + }, + { + "epoch": 0.94, + "learning_rate": 5.249348044132397e-05, + "loss": 1.2517, + "theoretical_loss": 3.341992435438539, + "tokens_seen": 3086614528 + }, + { + "epoch": 0.94, + "learning_rate": 5.247743229689067e-05, + "loss": 1.2392, + "theoretical_loss": 3.3419870574571306, + "tokens_seen": 3086680064 + }, + { + "epoch": 0.94, + "objective/train/docs_used": 1121586, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.2318816184997559, + "objective/train/theoretical_loss": 3.3419816796218766, + "objective/train/tokens_used": 3107205600, + "theoretical_loss": 3.3419816796218766, + "tokens_seen": 3086745600 + }, + { + "epoch": 0.94, + "learning_rate": 5.2461384152457374e-05, + "loss": 1.1981, + "theoretical_loss": 3.3419816796218766, + "tokens_seen": 3086745600 + }, + { + "epoch": 0.94, + "learning_rate": 5.244533600802407e-05, + "loss": 1.2847, + "theoretical_loss": 3.34197630193277, + "tokens_seen": 3086811136 + }, + { + "epoch": 0.94, + "learning_rate": 5.242928786359077e-05, + "loss": 1.2524, + "theoretical_loss": 3.3419709243898037, + "tokens_seen": 3086876672 + }, + { + "epoch": 0.94, + "learning_rate": 5.241323971915747e-05, + "loss": 1.2384, + "theoretical_loss": 3.34196554699297, + "tokens_seen": 3086942208 + }, + { + "epoch": 0.94, + "learning_rate": 5.239719157472418e-05, + "loss": 1.2236, + "theoretical_loss": 3.341960169742263, + "tokens_seen": 3087007744 + }, + { + "epoch": 0.94, + "learning_rate": 5.238114343029088e-05, + "loss": 1.2194, + "theoretical_loss": 3.341954792637675, + "tokens_seen": 3087073280 + }, + { + "epoch": 0.94, + "learning_rate": 5.236509528585758e-05, + "loss": 1.1916, + "theoretical_loss": 3.3419494156791982, + "tokens_seen": 3087138816 + }, + { + "epoch": 0.94, + "learning_rate": 5.234904714142428e-05, + "loss": 1.2486, + "theoretical_loss": 3.3419440388668264, + "tokens_seen": 3087204352 + }, + { + "epoch": 0.94, + "learning_rate": 5.233299899699098e-05, + "loss": 1.2533, + "theoretical_loss": 3.3419386622005525, + "tokens_seen": 3087269888 + }, + { + "epoch": 0.94, + "learning_rate": 5.231695085255768e-05, + "loss": 1.2674, + "theoretical_loss": 3.341933285680369, + "tokens_seen": 3087335424 + }, + { + "epoch": 0.94, + "learning_rate": 5.230090270812438e-05, + "loss": 1.3633, + "theoretical_loss": 3.3419279093062695, + "tokens_seen": 3087400960 + }, + { + "epoch": 0.94, + "learning_rate": 5.2284854563691074e-05, + "loss": 1.1186, + "theoretical_loss": 3.3419225330782467, + "tokens_seen": 3087466496 + }, + { + "epoch": 0.94, + "learning_rate": 5.2268806419257775e-05, + "loss": 1.4149, + "theoretical_loss": 3.341917156996293, + "tokens_seen": 3087532032 + }, + { + "epoch": 0.94, + "learning_rate": 5.2252758274824475e-05, + "loss": 1.2237, + "theoretical_loss": 3.3419117810604018, + "tokens_seen": 3087597568 + }, + { + "epoch": 0.94, + "learning_rate": 5.2236710130391176e-05, + "loss": 1.2363, + "theoretical_loss": 3.3419064052705654, + "tokens_seen": 3087663104 + }, + { + "epoch": 0.94, + "learning_rate": 5.222066198595788e-05, + "loss": 1.2988, + "theoretical_loss": 3.341901029626778, + "tokens_seen": 3087728640 + }, + { + "epoch": 0.94, + "learning_rate": 5.220461384152458e-05, + "loss": 1.2859, + "theoretical_loss": 3.3418956541290314, + "tokens_seen": 3087794176 + }, + { + "epoch": 0.94, + "learning_rate": 5.218856569709127e-05, + "loss": 1.2209, + "theoretical_loss": 3.341890278777319, + "tokens_seen": 3087859712 + }, + { + "epoch": 0.94, + "learning_rate": 5.217251755265797e-05, + "loss": 1.2381, + "theoretical_loss": 3.341884903571634, + "tokens_seen": 3087925248 + }, + { + "epoch": 0.94, + "learning_rate": 5.215646940822467e-05, + "loss": 1.3257, + "theoretical_loss": 3.3418795285119685, + "tokens_seen": 3087990784 + }, + { + "epoch": 0.94, + "learning_rate": 5.214042126379137e-05, + "loss": 1.2507, + "theoretical_loss": 3.341874153598316, + "tokens_seen": 3088056320 + }, + { + "epoch": 0.94, + "learning_rate": 5.2124373119358074e-05, + "loss": 1.2546, + "theoretical_loss": 3.341868778830669, + "tokens_seen": 3088121856 + }, + { + "epoch": 0.94, + "learning_rate": 5.2108324974924774e-05, + "loss": 1.2384, + "theoretical_loss": 3.3418634042090214, + "tokens_seen": 3088187392 + }, + { + "epoch": 0.94, + "learning_rate": 5.209227683049148e-05, + "loss": 1.2876, + "theoretical_loss": 3.3418580297333653, + "tokens_seen": 3088252928 + }, + { + "epoch": 0.94, + "learning_rate": 5.207622868605818e-05, + "loss": 1.1622, + "theoretical_loss": 3.3418526554036942, + "tokens_seen": 3088318464 + }, + { + "epoch": 0.94, + "learning_rate": 5.206018054162488e-05, + "loss": 1.2891, + "theoretical_loss": 3.34184728122, + "tokens_seen": 3088384000 + }, + { + "epoch": 0.94, + "learning_rate": 5.2044132397191584e-05, + "loss": 1.2461, + "theoretical_loss": 3.3418419071822774, + "tokens_seen": 3088449536 + }, + { + "epoch": 0.94, + "learning_rate": 5.202808425275828e-05, + "loss": 1.126, + "theoretical_loss": 3.3418365332905173, + "tokens_seen": 3088515072 + }, + { + "epoch": 0.94, + "learning_rate": 5.201203610832498e-05, + "loss": 1.3401, + "theoretical_loss": 3.3418311595447143, + "tokens_seen": 3088580608 + }, + { + "epoch": 0.94, + "learning_rate": 5.199598796389168e-05, + "loss": 1.2074, + "theoretical_loss": 3.34182578594486, + "tokens_seen": 3088646144 + }, + { + "epoch": 0.94, + "learning_rate": 5.197993981945838e-05, + "loss": 1.2088, + "theoretical_loss": 3.3418204124909483, + "tokens_seen": 3088711680 + }, + { + "epoch": 0.94, + "learning_rate": 5.196389167502508e-05, + "loss": 1.1442, + "theoretical_loss": 3.341815039182972, + "tokens_seen": 3088777216 + }, + { + "epoch": 0.94, + "learning_rate": 5.194784353059178e-05, + "loss": 1.2253, + "theoretical_loss": 3.3418096660209238, + "tokens_seen": 3088842752 + }, + { + "epoch": 0.94, + "learning_rate": 5.1931795386158474e-05, + "loss": 1.2887, + "theoretical_loss": 3.3418042930047966, + "tokens_seen": 3088908288 + }, + { + "epoch": 0.94, + "learning_rate": 5.1915747241725175e-05, + "loss": 1.1003, + "theoretical_loss": 3.3417989201345835, + "tokens_seen": 3088973824 + }, + { + "epoch": 0.94, + "learning_rate": 5.1899699097291876e-05, + "loss": 1.3208, + "theoretical_loss": 3.341793547410277, + "tokens_seen": 3089039360 + }, + { + "epoch": 0.94, + "learning_rate": 5.1883650952858576e-05, + "loss": 1.2734, + "theoretical_loss": 3.341788174831871, + "tokens_seen": 3089104896 + }, + { + "epoch": 0.94, + "learning_rate": 5.186760280842528e-05, + "loss": 1.3054, + "theoretical_loss": 3.341782802399358, + "tokens_seen": 3089170432 + }, + { + "epoch": 0.94, + "learning_rate": 5.185155466399198e-05, + "loss": 1.2164, + "theoretical_loss": 3.3417774301127303, + "tokens_seen": 3089235968 + }, + { + "epoch": 0.94, + "learning_rate": 5.183550651955867e-05, + "loss": 1.2533, + "theoretical_loss": 3.3417720579719816, + "tokens_seen": 3089301504 + }, + { + "epoch": 0.94, + "learning_rate": 5.181945837512537e-05, + "loss": 1.322, + "theoretical_loss": 3.3417666859771047, + "tokens_seen": 3089367040 + }, + { + "epoch": 0.94, + "learning_rate": 5.180341023069207e-05, + "loss": 1.173, + "theoretical_loss": 3.3417613141280924, + "tokens_seen": 3089432576 + }, + { + "epoch": 0.94, + "learning_rate": 5.178736208625879e-05, + "loss": 1.2308, + "theoretical_loss": 3.3417559424249377, + "tokens_seen": 3089498112 + }, + { + "epoch": 0.94, + "learning_rate": 5.177131394182548e-05, + "loss": 1.1787, + "theoretical_loss": 3.3417505708676334, + "tokens_seen": 3089563648 + }, + { + "epoch": 0.94, + "learning_rate": 5.175526579739218e-05, + "loss": 1.167, + "theoretical_loss": 3.341745199456173, + "tokens_seen": 3089629184 + }, + { + "epoch": 0.94, + "learning_rate": 5.173921765295888e-05, + "loss": 1.2347, + "theoretical_loss": 3.3417398281905486, + "tokens_seen": 3089694720 + }, + { + "epoch": 0.94, + "learning_rate": 5.172316950852558e-05, + "loss": 1.1745, + "theoretical_loss": 3.341734457070754, + "tokens_seen": 3089760256 + }, + { + "epoch": 0.94, + "learning_rate": 5.170712136409228e-05, + "loss": 1.2324, + "theoretical_loss": 3.341729086096781, + "tokens_seen": 3089825792 + }, + { + "epoch": 0.94, + "learning_rate": 5.1691073219658984e-05, + "loss": 1.3319, + "theoretical_loss": 3.3417237152686243, + "tokens_seen": 3089891328 + }, + { + "epoch": 0.94, + "learning_rate": 5.167502507522568e-05, + "loss": 1.174, + "theoretical_loss": 3.3417183445862753, + "tokens_seen": 3089956864 + }, + { + "epoch": 0.94, + "objective/train/docs_used": 1122843, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.1462211608886719, + "objective/train/theoretical_loss": 3.3417129740497273, + "objective/train/tokens_used": 3110482400, + "theoretical_loss": 3.3417129740497273, + "tokens_seen": 3090022400 + }, + { + "epoch": 0.94, + "learning_rate": 5.165897693079238e-05, + "loss": 1.1598, + "theoretical_loss": 3.3417129740497273, + "tokens_seen": 3090022400 + }, + { + "epoch": 0.94, + "learning_rate": 5.164292878635908e-05, + "loss": 1.3312, + "theoretical_loss": 3.341707603658974, + "tokens_seen": 3090087936 + }, + { + "epoch": 0.94, + "learning_rate": 5.162688064192578e-05, + "loss": 1.3709, + "theoretical_loss": 3.3417022334140074, + "tokens_seen": 3090153472 + }, + { + "epoch": 0.94, + "learning_rate": 5.161083249749248e-05, + "loss": 1.2772, + "theoretical_loss": 3.341696863314821, + "tokens_seen": 3090219008 + }, + { + "epoch": 0.94, + "learning_rate": 5.159478435305918e-05, + "loss": 1.2984, + "theoretical_loss": 3.3416914933614073, + "tokens_seen": 3090284544 + }, + { + "epoch": 0.94, + "learning_rate": 5.1578736208625875e-05, + "loss": 1.1276, + "theoretical_loss": 3.34168612355376, + "tokens_seen": 3090350080 + }, + { + "epoch": 0.94, + "learning_rate": 5.1562688064192576e-05, + "loss": 1.2475, + "theoretical_loss": 3.341680753891871, + "tokens_seen": 3090415616 + }, + { + "epoch": 0.94, + "learning_rate": 5.1546639919759276e-05, + "loss": 1.2387, + "theoretical_loss": 3.3416753843757343, + "tokens_seen": 3090481152 + }, + { + "epoch": 0.94, + "learning_rate": 5.153059177532598e-05, + "loss": 1.274, + "theoretical_loss": 3.3416700150053424, + "tokens_seen": 3090546688 + }, + { + "epoch": 0.94, + "learning_rate": 5.151454363089268e-05, + "loss": 1.2362, + "theoretical_loss": 3.341664645780688, + "tokens_seen": 3090612224 + }, + { + "epoch": 0.94, + "learning_rate": 5.1498495486459385e-05, + "loss": 1.1806, + "theoretical_loss": 3.3416592767017645, + "tokens_seen": 3090677760 + }, + { + "epoch": 0.94, + "learning_rate": 5.1482447342026085e-05, + "loss": 1.3052, + "theoretical_loss": 3.3416539077685643, + "tokens_seen": 3090743296 + }, + { + "epoch": 0.94, + "learning_rate": 5.1466399197592786e-05, + "loss": 1.2741, + "theoretical_loss": 3.3416485389810813, + "tokens_seen": 3090808832 + }, + { + "epoch": 0.94, + "learning_rate": 5.145035105315949e-05, + "loss": 1.2743, + "theoretical_loss": 3.3416431703393075, + "tokens_seen": 3090874368 + }, + { + "epoch": 0.94, + "learning_rate": 5.143430290872619e-05, + "loss": 1.2203, + "theoretical_loss": 3.341637801843236, + "tokens_seen": 3090939904 + }, + { + "epoch": 0.94, + "learning_rate": 5.141825476429288e-05, + "loss": 1.2146, + "theoretical_loss": 3.3416324334928604, + "tokens_seen": 3091005440 + }, + { + "epoch": 0.94, + "learning_rate": 5.140220661985958e-05, + "loss": 1.2228, + "theoretical_loss": 3.341627065288173, + "tokens_seen": 3091070976 + }, + { + "epoch": 0.94, + "learning_rate": 5.138615847542628e-05, + "loss": 1.3378, + "theoretical_loss": 3.3416216972291672, + "tokens_seen": 3091136512 + }, + { + "epoch": 0.94, + "learning_rate": 5.137011033099298e-05, + "loss": 1.2046, + "theoretical_loss": 3.3416163293158356, + "tokens_seen": 3091202048 + }, + { + "epoch": 0.94, + "learning_rate": 5.1354062186559684e-05, + "loss": 1.2473, + "theoretical_loss": 3.3416109615481715, + "tokens_seen": 3091267584 + }, + { + "epoch": 0.94, + "learning_rate": 5.1338014042126384e-05, + "loss": 1.3834, + "theoretical_loss": 3.3416055939261673, + "tokens_seen": 3091333120 + }, + { + "epoch": 0.94, + "learning_rate": 5.132196589769308e-05, + "loss": 1.3522, + "theoretical_loss": 3.341600226449817, + "tokens_seen": 3091398656 + }, + { + "epoch": 0.94, + "learning_rate": 5.130591775325978e-05, + "loss": 1.2491, + "theoretical_loss": 3.341594859119112, + "tokens_seen": 3091464192 + }, + { + "epoch": 0.94, + "learning_rate": 5.128986960882648e-05, + "loss": 1.2669, + "theoretical_loss": 3.3415894919340468, + "tokens_seen": 3091529728 + }, + { + "epoch": 0.94, + "learning_rate": 5.127382146439318e-05, + "loss": 1.2342, + "theoretical_loss": 3.3415841248946134, + "tokens_seen": 3091595264 + }, + { + "epoch": 0.94, + "learning_rate": 5.125777331995988e-05, + "loss": 1.3281, + "theoretical_loss": 3.3415787580008054, + "tokens_seen": 3091660800 + }, + { + "epoch": 0.94, + "learning_rate": 5.124172517552658e-05, + "loss": 1.3499, + "theoretical_loss": 3.341573391252615, + "tokens_seen": 3091726336 + }, + { + "epoch": 0.94, + "learning_rate": 5.1225677031093275e-05, + "loss": 1.3288, + "theoretical_loss": 3.341568024650036, + "tokens_seen": 3091791872 + }, + { + "epoch": 0.94, + "learning_rate": 5.1209628886659976e-05, + "loss": 1.3286, + "theoretical_loss": 3.3415626581930606, + "tokens_seen": 3091857408 + }, + { + "epoch": 0.94, + "learning_rate": 5.119358074222669e-05, + "loss": 1.3119, + "theoretical_loss": 3.3415572918816823, + "tokens_seen": 3091922944 + }, + { + "epoch": 0.94, + "learning_rate": 5.1177532597793384e-05, + "loss": 1.3053, + "theoretical_loss": 3.341551925715894, + "tokens_seen": 3091988480 + }, + { + "epoch": 0.94, + "learning_rate": 5.1161484453360085e-05, + "loss": 1.1828, + "theoretical_loss": 3.3415465596956886, + "tokens_seen": 3092054016 + }, + { + "epoch": 0.94, + "learning_rate": 5.1145436308926785e-05, + "loss": 1.1936, + "theoretical_loss": 3.3415411938210586, + "tokens_seen": 3092119552 + }, + { + "epoch": 0.94, + "learning_rate": 5.1129388164493486e-05, + "loss": 1.2513, + "theoretical_loss": 3.341535828091998, + "tokens_seen": 3092185088 + }, + { + "epoch": 0.94, + "learning_rate": 5.1113340020060187e-05, + "loss": 1.2919, + "theoretical_loss": 3.3415304625084987, + "tokens_seen": 3092250624 + }, + { + "epoch": 0.94, + "learning_rate": 5.109729187562689e-05, + "loss": 1.1988, + "theoretical_loss": 3.341525097070554, + "tokens_seen": 3092316160 + }, + { + "epoch": 0.94, + "learning_rate": 5.108124373119358e-05, + "loss": 1.2931, + "theoretical_loss": 3.3415197317781575, + "tokens_seen": 3092381696 + }, + { + "epoch": 0.94, + "learning_rate": 5.106519558676028e-05, + "loss": 1.306, + "theoretical_loss": 3.3415143666313014, + "tokens_seen": 3092447232 + }, + { + "epoch": 0.94, + "learning_rate": 5.104914744232698e-05, + "loss": 1.2326, + "theoretical_loss": 3.3415090016299787, + "tokens_seen": 3092512768 + }, + { + "epoch": 0.94, + "learning_rate": 5.103309929789368e-05, + "loss": 1.3427, + "theoretical_loss": 3.341503636774183, + "tokens_seen": 3092578304 + }, + { + "epoch": 0.94, + "learning_rate": 5.1017051153460384e-05, + "loss": 1.2525, + "theoretical_loss": 3.3414982720639066, + "tokens_seen": 3092643840 + }, + { + "epoch": 0.94, + "learning_rate": 5.1001003009027084e-05, + "loss": 1.2721, + "theoretical_loss": 3.3414929074991426, + "tokens_seen": 3092709376 + }, + { + "epoch": 0.94, + "learning_rate": 5.098495486459378e-05, + "loss": 1.2465, + "theoretical_loss": 3.341487543079884, + "tokens_seen": 3092774912 + }, + { + "epoch": 0.94, + "learning_rate": 5.096890672016048e-05, + "loss": 1.2482, + "theoretical_loss": 3.341482178806124, + "tokens_seen": 3092840448 + }, + { + "epoch": 0.94, + "learning_rate": 5.095285857572718e-05, + "loss": 1.2125, + "theoretical_loss": 3.341476814677856, + "tokens_seen": 3092905984 + }, + { + "epoch": 0.94, + "learning_rate": 5.093681043129388e-05, + "loss": 1.3272, + "theoretical_loss": 3.341471450695072, + "tokens_seen": 3092971520 + }, + { + "epoch": 0.94, + "learning_rate": 5.092076228686058e-05, + "loss": 1.1855, + "theoretical_loss": 3.341466086857765, + "tokens_seen": 3093037056 + }, + { + "epoch": 0.94, + "learning_rate": 5.090471414242728e-05, + "loss": 1.2555, + "theoretical_loss": 3.3414607231659286, + "tokens_seen": 3093102592 + }, + { + "epoch": 0.94, + "learning_rate": 5.088866599799399e-05, + "loss": 1.2437, + "theoretical_loss": 3.3414553596195558, + "tokens_seen": 3093168128 + }, + { + "epoch": 0.94, + "learning_rate": 5.087261785356069e-05, + "loss": 1.2333, + "theoretical_loss": 3.341449996218639, + "tokens_seen": 3093233664 + }, + { + "epoch": 0.94, + "objective/train/docs_used": 1123676, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.2998976707458496, + "objective/train/theoretical_loss": 3.3414446329631717, + "objective/train/tokens_used": 3113759200, + "theoretical_loss": 3.3414446329631717, + "tokens_seen": 3093299200 + }, + { + "epoch": 0.94, + "learning_rate": 5.085656970912739e-05, + "loss": 1.2756, + "theoretical_loss": 3.3414446329631717, + "tokens_seen": 3093299200 + }, + { + "epoch": 0.94, + "learning_rate": 5.084052156469409e-05, + "loss": 1.2264, + "theoretical_loss": 3.3414392698531463, + "tokens_seen": 3093364736 + }, + { + "epoch": 0.94, + "learning_rate": 5.0824473420260784e-05, + "loss": 1.2981, + "theoretical_loss": 3.3414339068885566, + "tokens_seen": 3093430272 + }, + { + "epoch": 0.94, + "learning_rate": 5.0808425275827485e-05, + "loss": 1.3474, + "theoretical_loss": 3.3414285440693945, + "tokens_seen": 3093495808 + }, + { + "epoch": 0.94, + "learning_rate": 5.0792377131394186e-05, + "loss": 1.3344, + "theoretical_loss": 3.341423181395654, + "tokens_seen": 3093561344 + }, + { + "epoch": 0.94, + "learning_rate": 5.0776328986960886e-05, + "loss": 1.2805, + "theoretical_loss": 3.3414178188673276, + "tokens_seen": 3093626880 + }, + { + "epoch": 0.94, + "learning_rate": 5.076028084252759e-05, + "loss": 1.297, + "theoretical_loss": 3.3414124564844085, + "tokens_seen": 3093692416 + }, + { + "epoch": 0.94, + "learning_rate": 5.074423269809429e-05, + "loss": 1.3212, + "theoretical_loss": 3.341407094246889, + "tokens_seen": 3093757952 + }, + { + "epoch": 0.94, + "learning_rate": 5.072818455366098e-05, + "loss": 1.2998, + "theoretical_loss": 3.341401732154763, + "tokens_seen": 3093823488 + }, + { + "epoch": 0.94, + "learning_rate": 5.071213640922768e-05, + "loss": 1.19, + "theoretical_loss": 3.341396370208023, + "tokens_seen": 3093889024 + }, + { + "epoch": 0.94, + "learning_rate": 5.069608826479438e-05, + "loss": 1.249, + "theoretical_loss": 3.341391008406662, + "tokens_seen": 3093954560 + }, + { + "epoch": 0.94, + "learning_rate": 5.068004012036108e-05, + "loss": 1.3212, + "theoretical_loss": 3.3413856467506724, + "tokens_seen": 3094020096 + }, + { + "epoch": 0.94, + "learning_rate": 5.0663991975927784e-05, + "loss": 1.2828, + "theoretical_loss": 3.3413802852400485, + "tokens_seen": 3094085632 + }, + { + "epoch": 0.94, + "learning_rate": 5.0647943831494485e-05, + "loss": 1.3143, + "theoretical_loss": 3.3413749238747825, + "tokens_seen": 3094151168 + }, + { + "epoch": 0.94, + "learning_rate": 5.063189568706118e-05, + "loss": 1.1708, + "theoretical_loss": 3.3413695626548674, + "tokens_seen": 3094216704 + }, + { + "epoch": 0.94, + "learning_rate": 5.061584754262788e-05, + "loss": 1.2026, + "theoretical_loss": 3.341364201580296, + "tokens_seen": 3094282240 + }, + { + "epoch": 0.94, + "learning_rate": 5.059979939819459e-05, + "loss": 1.2449, + "theoretical_loss": 3.341358840651062, + "tokens_seen": 3094347776 + }, + { + "epoch": 0.94, + "learning_rate": 5.0583751253761294e-05, + "loss": 1.3255, + "theoretical_loss": 3.3413534798671574, + "tokens_seen": 3094413312 + }, + { + "epoch": 0.94, + "learning_rate": 5.056770310932799e-05, + "loss": 1.313, + "theoretical_loss": 3.341348119228576, + "tokens_seen": 3094478848 + }, + { + "epoch": 0.94, + "learning_rate": 5.055165496489469e-05, + "loss": 1.2211, + "theoretical_loss": 3.34134275873531, + "tokens_seen": 3094544384 + }, + { + "epoch": 0.94, + "learning_rate": 5.053560682046139e-05, + "loss": 1.157, + "theoretical_loss": 3.3413373983873536, + "tokens_seen": 3094609920 + }, + { + "epoch": 0.94, + "learning_rate": 5.051955867602809e-05, + "loss": 1.4143, + "theoretical_loss": 3.3413320381846985, + "tokens_seen": 3094675456 + }, + { + "epoch": 0.94, + "learning_rate": 5.050351053159479e-05, + "loss": 1.413, + "theoretical_loss": 3.3413266781273383, + "tokens_seen": 3094740992 + }, + { + "epoch": 0.94, + "learning_rate": 5.048746238716149e-05, + "loss": 1.3073, + "theoretical_loss": 3.341321318215266, + "tokens_seen": 3094806528 + }, + { + "epoch": 0.94, + "learning_rate": 5.0471414242728185e-05, + "loss": 1.2535, + "theoretical_loss": 3.341315958448474, + "tokens_seen": 3094872064 + }, + { + "epoch": 0.94, + "learning_rate": 5.0455366098294885e-05, + "loss": 1.328, + "theoretical_loss": 3.3413105988269565, + "tokens_seen": 3094937600 + }, + { + "epoch": 0.94, + "learning_rate": 5.0439317953861586e-05, + "loss": 1.3128, + "theoretical_loss": 3.3413052393507057, + "tokens_seen": 3095003136 + }, + { + "epoch": 0.94, + "learning_rate": 5.042326980942829e-05, + "loss": 1.2016, + "theoretical_loss": 3.3412998800197142, + "tokens_seen": 3095068672 + }, + { + "epoch": 0.94, + "learning_rate": 5.040722166499499e-05, + "loss": 1.2193, + "theoretical_loss": 3.3412945208339755, + "tokens_seen": 3095134208 + }, + { + "epoch": 0.94, + "learning_rate": 5.039117352056169e-05, + "loss": 1.2862, + "theoretical_loss": 3.341289161793483, + "tokens_seen": 3095199744 + }, + { + "epoch": 0.94, + "learning_rate": 5.037512537612838e-05, + "loss": 1.0708, + "theoretical_loss": 3.3412838028982286, + "tokens_seen": 3095265280 + }, + { + "epoch": 0.94, + "learning_rate": 5.035907723169508e-05, + "loss": 1.2467, + "theoretical_loss": 3.341278444148206, + "tokens_seen": 3095330816 + }, + { + "epoch": 0.94, + "learning_rate": 5.034302908726178e-05, + "loss": 1.2963, + "theoretical_loss": 3.341273085543408, + "tokens_seen": 3095396352 + }, + { + "epoch": 0.94, + "learning_rate": 5.0326980942828484e-05, + "loss": 1.2295, + "theoretical_loss": 3.341267727083828, + "tokens_seen": 3095461888 + }, + { + "epoch": 0.94, + "learning_rate": 5.0310932798395184e-05, + "loss": 1.3089, + "theoretical_loss": 3.341262368769459, + "tokens_seen": 3095527424 + }, + { + "epoch": 0.94, + "learning_rate": 5.029488465396189e-05, + "loss": 1.1851, + "theoretical_loss": 3.341257010600293, + "tokens_seen": 3095592960 + }, + { + "epoch": 0.94, + "learning_rate": 5.027883650952859e-05, + "loss": 1.2321, + "theoretical_loss": 3.341251652576324, + "tokens_seen": 3095658496 + }, + { + "epoch": 0.94, + "learning_rate": 5.026278836509529e-05, + "loss": 1.3148, + "theoretical_loss": 3.3412462946975445, + "tokens_seen": 3095724032 + }, + { + "epoch": 0.94, + "learning_rate": 5.0246740220661994e-05, + "loss": 1.2638, + "theoretical_loss": 3.3412409369639473, + "tokens_seen": 3095789568 + }, + { + "epoch": 0.94, + "learning_rate": 5.0230692076228694e-05, + "loss": 1.2178, + "theoretical_loss": 3.341235579375526, + "tokens_seen": 3095855104 + }, + { + "epoch": 0.94, + "learning_rate": 5.021464393179539e-05, + "loss": 1.1672, + "theoretical_loss": 3.3412302219322734, + "tokens_seen": 3095920640 + }, + { + "epoch": 0.94, + "learning_rate": 5.019859578736209e-05, + "loss": 1.2292, + "theoretical_loss": 3.3412248646341824, + "tokens_seen": 3095986176 + }, + { + "epoch": 0.94, + "learning_rate": 5.018254764292879e-05, + "loss": 1.2289, + "theoretical_loss": 3.341219507481246, + "tokens_seen": 3096051712 + }, + { + "epoch": 0.94, + "learning_rate": 5.016649949849549e-05, + "loss": 1.2229, + "theoretical_loss": 3.341214150473457, + "tokens_seen": 3096117248 + }, + { + "epoch": 0.94, + "learning_rate": 5.015045135406219e-05, + "loss": 1.2241, + "theoretical_loss": 3.341208793610809, + "tokens_seen": 3096182784 + }, + { + "epoch": 0.94, + "learning_rate": 5.013440320962889e-05, + "loss": 1.276, + "theoretical_loss": 3.341203436893294, + "tokens_seen": 3096248320 + }, + { + "epoch": 0.94, + "learning_rate": 5.0118355065195585e-05, + "loss": 1.2698, + "theoretical_loss": 3.3411980803209063, + "tokens_seen": 3096313856 + }, + { + "epoch": 0.94, + "learning_rate": 5.0102306920762286e-05, + "loss": 1.1506, + "theoretical_loss": 3.341192723893638, + "tokens_seen": 3096379392 + }, + { + "epoch": 0.94, + "learning_rate": 5.0086258776328986e-05, + "loss": 1.2683, + "theoretical_loss": 3.341187367611482, + "tokens_seen": 3096444928 + }, + { + "epoch": 0.94, + "learning_rate": 5.007021063189569e-05, + "loss": 1.1683, + "theoretical_loss": 3.3411820114744315, + "tokens_seen": 3096510464 + }, + { + "epoch": 0.94, + "objective/train/docs_used": 1124894, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.2610440254211426, + "objective/train/theoretical_loss": 3.3411766554824798, + "objective/train/tokens_used": 3117036000, + "theoretical_loss": 3.3411766554824798, + "tokens_seen": 3096576000 + }, + { + "epoch": 0.94, + "learning_rate": 5.005416248746239e-05, + "loss": 1.254, + "theoretical_loss": 3.3411766554824798, + "tokens_seen": 3096576000 + }, + { + "epoch": 0.94, + "learning_rate": 5.003811434302909e-05, + "loss": 1.1473, + "theoretical_loss": 3.3411712996356195, + "tokens_seen": 3096641536 + }, + { + "epoch": 0.94, + "learning_rate": 5.002206619859578e-05, + "loss": 1.3875, + "theoretical_loss": 3.341165943933844, + "tokens_seen": 3096707072 + }, + { + "epoch": 0.94, + "learning_rate": 5.000601805416248e-05, + "loss": 1.2247, + "theoretical_loss": 3.341160588377146, + "tokens_seen": 3096772608 + }, + { + "epoch": 0.94, + "learning_rate": 4.998996990972919e-05, + "loss": 1.2826, + "theoretical_loss": 3.3411552329655185, + "tokens_seen": 3096838144 + }, + { + "epoch": 0.94, + "learning_rate": 4.997392176529589e-05, + "loss": 1.2748, + "theoretical_loss": 3.341149877698955, + "tokens_seen": 3096903680 + }, + { + "epoch": 0.94, + "learning_rate": 4.995787362086259e-05, + "loss": 1.239, + "theoretical_loss": 3.3411445225774474, + "tokens_seen": 3096969216 + }, + { + "epoch": 0.94, + "learning_rate": 4.9941825476429285e-05, + "loss": 1.3149, + "theoretical_loss": 3.34113916760099, + "tokens_seen": 3097034752 + }, + { + "epoch": 0.94, + "learning_rate": 4.9925777331995986e-05, + "loss": 1.3526, + "theoretical_loss": 3.3411338127695744, + "tokens_seen": 3097100288 + }, + { + "epoch": 0.94, + "learning_rate": 4.9909729187562693e-05, + "loss": 1.2389, + "theoretical_loss": 3.341128458083195, + "tokens_seen": 3097165824 + }, + { + "epoch": 0.94, + "learning_rate": 4.9893681043129394e-05, + "loss": 1.2467, + "theoretical_loss": 3.341123103541844, + "tokens_seen": 3097231360 + }, + { + "epoch": 0.94, + "learning_rate": 4.9877632898696095e-05, + "loss": 1.2702, + "theoretical_loss": 3.3411177491455146, + "tokens_seen": 3097296896 + }, + { + "epoch": 0.94, + "learning_rate": 4.986158475426279e-05, + "loss": 1.3124, + "theoretical_loss": 3.3411123948942, + "tokens_seen": 3097362432 + }, + { + "epoch": 0.94, + "learning_rate": 4.984553660982949e-05, + "loss": 1.2424, + "theoretical_loss": 3.3411070407878927, + "tokens_seen": 3097427968 + }, + { + "epoch": 0.94, + "learning_rate": 4.982948846539619e-05, + "loss": 1.2551, + "theoretical_loss": 3.341101686826586, + "tokens_seen": 3097493504 + }, + { + "epoch": 0.94, + "learning_rate": 4.981344032096289e-05, + "loss": 1.2429, + "theoretical_loss": 3.341096333010273, + "tokens_seen": 3097559040 + }, + { + "epoch": 0.94, + "learning_rate": 4.979739217652959e-05, + "loss": 1.2845, + "theoretical_loss": 3.3410909793389463, + "tokens_seen": 3097624576 + }, + { + "epoch": 0.94, + "learning_rate": 4.978134403209629e-05, + "loss": 1.2942, + "theoretical_loss": 3.3410856258125996, + "tokens_seen": 3097690112 + }, + { + "epoch": 0.94, + "learning_rate": 4.976529588766299e-05, + "loss": 1.1456, + "theoretical_loss": 3.3410802724312254, + "tokens_seen": 3097755648 + }, + { + "epoch": 0.94, + "learning_rate": 4.974924774322969e-05, + "loss": 1.2243, + "theoretical_loss": 3.341074919194817, + "tokens_seen": 3097821184 + }, + { + "epoch": 0.94, + "learning_rate": 4.9733199598796394e-05, + "loss": 1.2947, + "theoretical_loss": 3.341069566103367, + "tokens_seen": 3097886720 + }, + { + "epoch": 0.94, + "learning_rate": 4.9717151454363094e-05, + "loss": 1.2647, + "theoretical_loss": 3.341064213156869, + "tokens_seen": 3097952256 + }, + { + "epoch": 0.94, + "learning_rate": 4.9701103309929795e-05, + "loss": 1.202, + "theoretical_loss": 3.3410588603553153, + "tokens_seen": 3098017792 + }, + { + "epoch": 0.94, + "learning_rate": 4.968505516549649e-05, + "loss": 1.3245, + "theoretical_loss": 3.3410535076986996, + "tokens_seen": 3098083328 + }, + { + "epoch": 0.94, + "learning_rate": 4.966900702106319e-05, + "loss": 1.2252, + "theoretical_loss": 3.341048155187014, + "tokens_seen": 3098148864 + }, + { + "epoch": 0.94, + "learning_rate": 4.965295887662989e-05, + "loss": 1.2706, + "theoretical_loss": 3.3410428028202523, + "tokens_seen": 3098214400 + }, + { + "epoch": 0.94, + "learning_rate": 4.963691073219659e-05, + "loss": 1.1831, + "theoretical_loss": 3.3410374505984075, + "tokens_seen": 3098279936 + }, + { + "epoch": 0.94, + "learning_rate": 4.96208625877633e-05, + "loss": 1.2059, + "theoretical_loss": 3.3410320985214725, + "tokens_seen": 3098345472 + }, + { + "epoch": 0.94, + "learning_rate": 4.960481444332999e-05, + "loss": 1.3145, + "theoretical_loss": 3.34102674658944, + "tokens_seen": 3098411008 + }, + { + "epoch": 0.94, + "learning_rate": 4.958876629889669e-05, + "loss": 1.267, + "theoretical_loss": 3.3410213948023033, + "tokens_seen": 3098476544 + }, + { + "epoch": 0.94, + "learning_rate": 4.957271815446339e-05, + "loss": 1.1881, + "theoretical_loss": 3.3410160431600553, + "tokens_seen": 3098542080 + }, + { + "epoch": 0.94, + "learning_rate": 4.9556670010030094e-05, + "loss": 1.254, + "theoretical_loss": 3.3410106916626887, + "tokens_seen": 3098607616 + }, + { + "epoch": 0.94, + "learning_rate": 4.9540621865596795e-05, + "loss": 1.2455, + "theoretical_loss": 3.3410053403101974, + "tokens_seen": 3098673152 + }, + { + "epoch": 0.94, + "learning_rate": 4.9524573721163495e-05, + "loss": 1.3426, + "theoretical_loss": 3.3409999891025737, + "tokens_seen": 3098738688 + }, + { + "epoch": 0.94, + "learning_rate": 4.950852557673019e-05, + "loss": 1.2022, + "theoretical_loss": 3.3409946380398106, + "tokens_seen": 3098804224 + }, + { + "epoch": 0.94, + "learning_rate": 4.949247743229689e-05, + "loss": 1.3227, + "theoretical_loss": 3.340989287121902, + "tokens_seen": 3098869760 + }, + { + "epoch": 0.94, + "learning_rate": 4.947642928786359e-05, + "loss": 1.2027, + "theoretical_loss": 3.3409839363488394, + "tokens_seen": 3098935296 + }, + { + "epoch": 0.94, + "learning_rate": 4.94603811434303e-05, + "loss": 1.3085, + "theoretical_loss": 3.340978585720617, + "tokens_seen": 3099000832 + }, + { + "epoch": 0.94, + "learning_rate": 4.9444332998997e-05, + "loss": 1.2692, + "theoretical_loss": 3.3409732352372274, + "tokens_seen": 3099066368 + }, + { + "epoch": 0.94, + "learning_rate": 4.942828485456369e-05, + "loss": 1.2856, + "theoretical_loss": 3.3409678848986637, + "tokens_seen": 3099131904 + }, + { + "epoch": 0.94, + "learning_rate": 4.941223671013039e-05, + "loss": 1.1272, + "theoretical_loss": 3.340962534704919, + "tokens_seen": 3099197440 + }, + { + "epoch": 0.94, + "learning_rate": 4.9396188565697093e-05, + "loss": 1.2263, + "theoretical_loss": 3.340957184655986, + "tokens_seen": 3099262976 + }, + { + "epoch": 0.94, + "learning_rate": 4.9380140421263794e-05, + "loss": 1.1608, + "theoretical_loss": 3.3409518347518583, + "tokens_seen": 3099328512 + }, + { + "epoch": 0.94, + "learning_rate": 4.9364092276830495e-05, + "loss": 1.2013, + "theoretical_loss": 3.340946484992528, + "tokens_seen": 3099394048 + }, + { + "epoch": 0.94, + "learning_rate": 4.9348044132397195e-05, + "loss": 1.2433, + "theoretical_loss": 3.340941135377989, + "tokens_seen": 3099459584 + }, + { + "epoch": 0.94, + "learning_rate": 4.933199598796389e-05, + "loss": 1.2344, + "theoretical_loss": 3.340935785908234, + "tokens_seen": 3099525120 + }, + { + "epoch": 0.94, + "learning_rate": 4.93159478435306e-05, + "loss": 1.22, + "theoretical_loss": 3.340930436583256, + "tokens_seen": 3099590656 + }, + { + "epoch": 0.94, + "learning_rate": 4.92998996990973e-05, + "loss": 1.2106, + "theoretical_loss": 3.340925087403048, + "tokens_seen": 3099656192 + }, + { + "epoch": 0.94, + "learning_rate": 4.9283851554664e-05, + "loss": 1.2133, + "theoretical_loss": 3.340919738367603, + "tokens_seen": 3099721728 + }, + { + "epoch": 0.94, + "learning_rate": 4.926780341023069e-05, + "loss": 1.3033, + "theoretical_loss": 3.3409143894769144, + "tokens_seen": 3099787264 + }, + { + "epoch": 0.94, + "objective/train/docs_used": 1126162, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.2644875049591064, + "objective/train/theoretical_loss": 3.3409090407309745, + "objective/train/tokens_used": 3120312800, + "theoretical_loss": 3.3409090407309745, + "tokens_seen": 3099852800 + }, + { + "epoch": 0.94, + "learning_rate": 4.925175526579739e-05, + "loss": 1.2539, + "theoretical_loss": 3.3409090407309745, + "tokens_seen": 3099852800 + }, + { + "epoch": 0.94, + "learning_rate": 4.923570712136409e-05, + "loss": 1.269, + "theoretical_loss": 3.3409036921297766, + "tokens_seen": 3099918336 + }, + { + "epoch": 0.94, + "learning_rate": 4.9219658976930794e-05, + "loss": 1.3286, + "theoretical_loss": 3.3408983436733144, + "tokens_seen": 3099983872 + }, + { + "epoch": 0.94, + "learning_rate": 4.9203610832497494e-05, + "loss": 1.2774, + "theoretical_loss": 3.34089299536158, + "tokens_seen": 3100049408 + }, + { + "epoch": 0.94, + "learning_rate": 4.9187562688064195e-05, + "loss": 1.2862, + "theoretical_loss": 3.3408876471945668, + "tokens_seen": 3100114944 + }, + { + "epoch": 0.94, + "learning_rate": 4.9171514543630896e-05, + "loss": 1.2673, + "theoretical_loss": 3.340882299172268, + "tokens_seen": 3100180480 + }, + { + "epoch": 0.94, + "learning_rate": 4.9155466399197596e-05, + "loss": 1.3124, + "theoretical_loss": 3.340876951294676, + "tokens_seen": 3100246016 + }, + { + "epoch": 0.94, + "learning_rate": 4.91394182547643e-05, + "loss": 1.2884, + "theoretical_loss": 3.3408716035617845, + "tokens_seen": 3100311552 + }, + { + "epoch": 0.94, + "learning_rate": 4.9123370110331e-05, + "loss": 1.2166, + "theoretical_loss": 3.3408662559735864, + "tokens_seen": 3100377088 + }, + { + "epoch": 0.94, + "learning_rate": 4.91073219658977e-05, + "loss": 1.2516, + "theoretical_loss": 3.3408609085300744, + "tokens_seen": 3100442624 + }, + { + "epoch": 0.94, + "learning_rate": 4.909127382146439e-05, + "loss": 1.2799, + "theoretical_loss": 3.3408555612312423, + "tokens_seen": 3100508160 + }, + { + "epoch": 0.94, + "learning_rate": 4.907522567703109e-05, + "loss": 1.204, + "theoretical_loss": 3.340850214077082, + "tokens_seen": 3100573696 + }, + { + "epoch": 0.94, + "learning_rate": 4.905917753259779e-05, + "loss": 1.2195, + "theoretical_loss": 3.340844867067587, + "tokens_seen": 3100639232 + }, + { + "epoch": 0.94, + "learning_rate": 4.9043129388164494e-05, + "loss": 1.1733, + "theoretical_loss": 3.340839520202751, + "tokens_seen": 3100704768 + }, + { + "epoch": 0.94, + "learning_rate": 4.9027081243731194e-05, + "loss": 1.2451, + "theoretical_loss": 3.340834173482566, + "tokens_seen": 3100770304 + }, + { + "epoch": 0.94, + "learning_rate": 4.9011033099297895e-05, + "loss": 1.2919, + "theoretical_loss": 3.3408288269070257, + "tokens_seen": 3100835840 + }, + { + "epoch": 0.94, + "learning_rate": 4.8994984954864596e-05, + "loss": 1.2659, + "theoretical_loss": 3.340823480476123, + "tokens_seen": 3100901376 + }, + { + "epoch": 0.94, + "learning_rate": 4.8978936810431296e-05, + "loss": 1.2577, + "theoretical_loss": 3.340818134189851, + "tokens_seen": 3100966912 + }, + { + "epoch": 0.94, + "learning_rate": 4.8962888665998e-05, + "loss": 1.2613, + "theoretical_loss": 3.340812788048202, + "tokens_seen": 3101032448 + }, + { + "epoch": 0.94, + "learning_rate": 4.89468405215647e-05, + "loss": 1.1869, + "theoretical_loss": 3.3408074420511698, + "tokens_seen": 3101097984 + }, + { + "epoch": 0.94, + "learning_rate": 4.89307923771314e-05, + "loss": 1.2836, + "theoretical_loss": 3.3408020961987472, + "tokens_seen": 3101163520 + }, + { + "epoch": 0.94, + "learning_rate": 4.891474423269809e-05, + "loss": 1.2056, + "theoretical_loss": 3.3407967504909277, + "tokens_seen": 3101229056 + }, + { + "epoch": 0.94, + "learning_rate": 4.889869608826479e-05, + "loss": 1.2262, + "theoretical_loss": 3.3407914049277037, + "tokens_seen": 3101294592 + }, + { + "epoch": 0.94, + "learning_rate": 4.8882647943831493e-05, + "loss": 1.2885, + "theoretical_loss": 3.3407860595090684, + "tokens_seen": 3101360128 + }, + { + "epoch": 0.94, + "learning_rate": 4.88665997993982e-05, + "loss": 1.2204, + "theoretical_loss": 3.3407807142350148, + "tokens_seen": 3101425664 + }, + { + "epoch": 0.94, + "learning_rate": 4.88505516549649e-05, + "loss": 1.3008, + "theoretical_loss": 3.3407753691055357, + "tokens_seen": 3101491200 + }, + { + "epoch": 0.94, + "learning_rate": 4.8834503510531595e-05, + "loss": 1.2968, + "theoretical_loss": 3.340770024120625, + "tokens_seen": 3101556736 + }, + { + "epoch": 0.94, + "learning_rate": 4.8818455366098296e-05, + "loss": 1.2359, + "theoretical_loss": 3.3407646792802748, + "tokens_seen": 3101622272 + }, + { + "epoch": 0.94, + "learning_rate": 4.8802407221664997e-05, + "loss": 1.2712, + "theoretical_loss": 3.3407593345844786, + "tokens_seen": 3101687808 + }, + { + "epoch": 0.94, + "learning_rate": 4.87863590772317e-05, + "loss": 1.199, + "theoretical_loss": 3.3407539900332295, + "tokens_seen": 3101753344 + }, + { + "epoch": 0.94, + "learning_rate": 4.87703109327984e-05, + "loss": 1.3587, + "theoretical_loss": 3.3407486456265203, + "tokens_seen": 3101818880 + }, + { + "epoch": 0.94, + "learning_rate": 4.87542627883651e-05, + "loss": 1.2741, + "theoretical_loss": 3.3407433013643444, + "tokens_seen": 3101884416 + }, + { + "epoch": 0.94, + "learning_rate": 4.873821464393179e-05, + "loss": 1.3524, + "theoretical_loss": 3.340737957246694, + "tokens_seen": 3101949952 + }, + { + "epoch": 0.94, + "learning_rate": 4.87221664994985e-05, + "loss": 1.1556, + "theoretical_loss": 3.3407326132735635, + "tokens_seen": 3102015488 + }, + { + "epoch": 0.94, + "learning_rate": 4.87061183550652e-05, + "loss": 1.2264, + "theoretical_loss": 3.3407272694449444, + "tokens_seen": 3102081024 + }, + { + "epoch": 0.94, + "learning_rate": 4.86900702106319e-05, + "loss": 1.2734, + "theoretical_loss": 3.340721925760831, + "tokens_seen": 3102146560 + }, + { + "epoch": 0.94, + "learning_rate": 4.86740220661986e-05, + "loss": 1.2365, + "theoretical_loss": 3.3407165822212153, + "tokens_seen": 3102212096 + }, + { + "epoch": 0.94, + "learning_rate": 4.8657973921765296e-05, + "loss": 1.2149, + "theoretical_loss": 3.340711238826091, + "tokens_seen": 3102277632 + }, + { + "epoch": 0.94, + "learning_rate": 4.8641925777331996e-05, + "loss": 1.2707, + "theoretical_loss": 3.340705895575451, + "tokens_seen": 3102343168 + }, + { + "epoch": 0.94, + "learning_rate": 4.86258776328987e-05, + "loss": 1.4105, + "theoretical_loss": 3.3407005524692885, + "tokens_seen": 3102408704 + }, + { + "epoch": 0.94, + "learning_rate": 4.86098294884654e-05, + "loss": 1.229, + "theoretical_loss": 3.3406952095075964, + "tokens_seen": 3102474240 + }, + { + "epoch": 0.94, + "learning_rate": 4.85937813440321e-05, + "loss": 1.3211, + "theoretical_loss": 3.3406898666903677, + "tokens_seen": 3102539776 + }, + { + "epoch": 0.94, + "learning_rate": 4.85777331995988e-05, + "loss": 1.3018, + "theoretical_loss": 3.3406845240175955, + "tokens_seen": 3102605312 + }, + { + "epoch": 0.94, + "learning_rate": 4.85616850551655e-05, + "loss": 1.2975, + "theoretical_loss": 3.3406791814892727, + "tokens_seen": 3102670848 + }, + { + "epoch": 0.94, + "learning_rate": 4.85456369107322e-05, + "loss": 1.3249, + "theoretical_loss": 3.3406738391053925, + "tokens_seen": 3102736384 + }, + { + "epoch": 0.94, + "learning_rate": 4.85295887662989e-05, + "loss": 1.1205, + "theoretical_loss": 3.340668496865948, + "tokens_seen": 3102801920 + }, + { + "epoch": 0.94, + "learning_rate": 4.85135406218656e-05, + "loss": 1.2432, + "theoretical_loss": 3.340663154770932, + "tokens_seen": 3102867456 + }, + { + "epoch": 0.94, + "learning_rate": 4.84974924774323e-05, + "loss": 1.2309, + "theoretical_loss": 3.3406578128203375, + "tokens_seen": 3102932992 + }, + { + "epoch": 0.94, + "learning_rate": 4.8481444332998996e-05, + "loss": 1.2949, + "theoretical_loss": 3.340652471014158, + "tokens_seen": 3102998528 + }, + { + "epoch": 0.94, + "learning_rate": 4.8465396188565696e-05, + "loss": 1.336, + "theoretical_loss": 3.3406471293523863, + "tokens_seen": 3103064064 + }, + { + "epoch": 0.94, + "objective/train/docs_used": 1127317, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.1397931575775146, + "objective/train/theoretical_loss": 3.340641787835015, + "objective/train/tokens_used": 3123589600, + "theoretical_loss": 3.340641787835015, + "tokens_seen": 3103129600 + }, + { + "epoch": 0.94, + "learning_rate": 4.84493480441324e-05, + "loss": 1.2264, + "theoretical_loss": 3.340641787835015, + "tokens_seen": 3103129600 + }, + { + "epoch": 0.94, + "learning_rate": 4.84332998996991e-05, + "loss": 1.2192, + "theoretical_loss": 3.340636446462038, + "tokens_seen": 3103195136 + }, + { + "epoch": 0.94, + "learning_rate": 4.8417251755265805e-05, + "loss": 1.2362, + "theoretical_loss": 3.340631105233448, + "tokens_seen": 3103260672 + }, + { + "epoch": 0.94, + "learning_rate": 4.84012036108325e-05, + "loss": 1.2475, + "theoretical_loss": 3.340625764149238, + "tokens_seen": 3103326208 + }, + { + "epoch": 0.94, + "learning_rate": 4.83851554663992e-05, + "loss": 1.2592, + "theoretical_loss": 3.340620423209401, + "tokens_seen": 3103391744 + }, + { + "epoch": 0.94, + "learning_rate": 4.83691073219659e-05, + "loss": 1.2557, + "theoretical_loss": 3.3406150824139296, + "tokens_seen": 3103457280 + }, + { + "epoch": 0.94, + "learning_rate": 4.83530591775326e-05, + "loss": 1.2163, + "theoretical_loss": 3.3406097417628176, + "tokens_seen": 3103522816 + }, + { + "epoch": 0.94, + "learning_rate": 4.83370110330993e-05, + "loss": 1.3101, + "theoretical_loss": 3.3406044012560576, + "tokens_seen": 3103588352 + }, + { + "epoch": 0.94, + "learning_rate": 4.8320962888666e-05, + "loss": 1.1398, + "theoretical_loss": 3.340599060893643, + "tokens_seen": 3103653888 + }, + { + "epoch": 0.94, + "learning_rate": 4.8304914744232696e-05, + "loss": 1.2611, + "theoretical_loss": 3.3405937206755665, + "tokens_seen": 3103719424 + }, + { + "epoch": 0.94, + "learning_rate": 4.8288866599799397e-05, + "loss": 1.327, + "theoretical_loss": 3.3405883806018215, + "tokens_seen": 3103784960 + }, + { + "epoch": 0.94, + "learning_rate": 4.8272818455366104e-05, + "loss": 1.3094, + "theoretical_loss": 3.340583040672401, + "tokens_seen": 3103850496 + }, + { + "epoch": 0.94, + "learning_rate": 4.8256770310932805e-05, + "loss": 1.3262, + "theoretical_loss": 3.3405777008872977, + "tokens_seen": 3103916032 + }, + { + "epoch": 0.94, + "learning_rate": 4.8240722166499505e-05, + "loss": 1.2121, + "theoretical_loss": 3.340572361246505, + "tokens_seen": 3103981568 + }, + { + "epoch": 0.94, + "learning_rate": 4.82246740220662e-05, + "loss": 1.2658, + "theoretical_loss": 3.3405670217500156, + "tokens_seen": 3104047104 + }, + { + "epoch": 0.94, + "learning_rate": 4.82086258776329e-05, + "loss": 1.291, + "theoretical_loss": 3.340561682397823, + "tokens_seen": 3104112640 + }, + { + "epoch": 0.94, + "learning_rate": 4.81925777331996e-05, + "loss": 1.2735, + "theoretical_loss": 3.34055634318992, + "tokens_seen": 3104178176 + }, + { + "epoch": 0.94, + "learning_rate": 4.81765295887663e-05, + "loss": 1.2674, + "theoretical_loss": 3.3405510041263, + "tokens_seen": 3104243712 + }, + { + "epoch": 0.94, + "learning_rate": 4.8160481444333e-05, + "loss": 1.2016, + "theoretical_loss": 3.340545665206955, + "tokens_seen": 3104309248 + }, + { + "epoch": 0.94, + "learning_rate": 4.81444332998997e-05, + "loss": 1.1415, + "theoretical_loss": 3.340540326431879, + "tokens_seen": 3104374784 + }, + { + "epoch": 0.94, + "learning_rate": 4.8128385155466396e-05, + "loss": 1.376, + "theoretical_loss": 3.3405349878010653, + "tokens_seen": 3104440320 + }, + { + "epoch": 0.94, + "learning_rate": 4.8112337011033104e-05, + "loss": 1.4126, + "theoretical_loss": 3.3405296493145062, + "tokens_seen": 3104505856 + }, + { + "epoch": 0.94, + "learning_rate": 4.8096288866599804e-05, + "loss": 1.1673, + "theoretical_loss": 3.340524310972195, + "tokens_seen": 3104571392 + }, + { + "epoch": 0.94, + "learning_rate": 4.8080240722166505e-05, + "loss": 1.2235, + "theoretical_loss": 3.340518972774125, + "tokens_seen": 3104636928 + }, + { + "epoch": 0.94, + "learning_rate": 4.8064192577733205e-05, + "loss": 1.2754, + "theoretical_loss": 3.340513634720289, + "tokens_seen": 3104702464 + }, + { + "epoch": 0.94, + "learning_rate": 4.80481444332999e-05, + "loss": 1.1885, + "theoretical_loss": 3.3405082968106803, + "tokens_seen": 3104768000 + }, + { + "epoch": 0.94, + "learning_rate": 4.80320962888666e-05, + "loss": 1.2569, + "theoretical_loss": 3.3405029590452915, + "tokens_seen": 3104833536 + }, + { + "epoch": 0.94, + "learning_rate": 4.80160481444333e-05, + "loss": 1.151, + "theoretical_loss": 3.340497621424116, + "tokens_seen": 3104899072 + }, + { + "epoch": 0.94, + "learning_rate": 4.8e-05, + "loss": 1.2499, + "theoretical_loss": 3.340492283947147, + "tokens_seen": 3104964608 + }, + { + "epoch": 0.94, + "learning_rate": 4.79839518555667e-05, + "loss": 1.2564, + "theoretical_loss": 3.3404869466143774, + "tokens_seen": 3105030144 + }, + { + "epoch": 0.94, + "learning_rate": 4.79679037111334e-05, + "loss": 1.2727, + "theoretical_loss": 3.3404816094258, + "tokens_seen": 3105095680 + }, + { + "epoch": 0.94, + "learning_rate": 4.79518555667001e-05, + "loss": 1.3261, + "theoretical_loss": 3.3404762723814083, + "tokens_seen": 3105161216 + }, + { + "epoch": 0.94, + "learning_rate": 4.7935807422266804e-05, + "loss": 1.2228, + "theoretical_loss": 3.3404709354811954, + "tokens_seen": 3105226752 + }, + { + "epoch": 0.94, + "learning_rate": 4.7919759277833504e-05, + "loss": 1.2313, + "theoretical_loss": 3.340465598725154, + "tokens_seen": 3105292288 + }, + { + "epoch": 0.94, + "learning_rate": 4.7903711133400205e-05, + "loss": 1.2611, + "theoretical_loss": 3.340460262113277, + "tokens_seen": 3105357824 + }, + { + "epoch": 0.94, + "learning_rate": 4.7887662988966906e-05, + "loss": 1.2824, + "theoretical_loss": 3.340454925645558, + "tokens_seen": 3105423360 + }, + { + "epoch": 0.94, + "learning_rate": 4.78716148445336e-05, + "loss": 1.2578, + "theoretical_loss": 3.3404495893219894, + "tokens_seen": 3105488896 + }, + { + "epoch": 0.94, + "learning_rate": 4.78555667001003e-05, + "loss": 1.2201, + "theoretical_loss": 3.3404442531425653, + "tokens_seen": 3105554432 + }, + { + "epoch": 0.94, + "learning_rate": 4.7839518555667e-05, + "loss": 1.2408, + "theoretical_loss": 3.3404389171072775, + "tokens_seen": 3105619968 + }, + { + "epoch": 0.94, + "learning_rate": 4.78234704112337e-05, + "loss": 1.1861, + "theoretical_loss": 3.34043358121612, + "tokens_seen": 3105685504 + }, + { + "epoch": 0.94, + "learning_rate": 4.780742226680041e-05, + "loss": 1.293, + "theoretical_loss": 3.3404282454690857, + "tokens_seen": 3105751040 + }, + { + "epoch": 0.94, + "learning_rate": 4.77913741223671e-05, + "loss": 1.2574, + "theoretical_loss": 3.3404229098661675, + "tokens_seen": 3105816576 + }, + { + "epoch": 0.94, + "learning_rate": 4.77753259779338e-05, + "loss": 1.2976, + "theoretical_loss": 3.3404175744073585, + "tokens_seen": 3105882112 + }, + { + "epoch": 0.94, + "learning_rate": 4.7759277833500504e-05, + "loss": 1.2865, + "theoretical_loss": 3.3404122390926516, + "tokens_seen": 3105947648 + }, + { + "epoch": 0.94, + "learning_rate": 4.7743229689067205e-05, + "loss": 1.3868, + "theoretical_loss": 3.3404069039220405, + "tokens_seen": 3106013184 + }, + { + "epoch": 0.94, + "learning_rate": 4.7727181544633905e-05, + "loss": 1.3894, + "theoretical_loss": 3.3404015688955173, + "tokens_seen": 3106078720 + }, + { + "epoch": 0.94, + "learning_rate": 4.7711133400200606e-05, + "loss": 1.2965, + "theoretical_loss": 3.3403962340130757, + "tokens_seen": 3106144256 + }, + { + "epoch": 0.94, + "learning_rate": 4.76950852557673e-05, + "loss": 1.2451, + "theoretical_loss": 3.340390899274709, + "tokens_seen": 3106209792 + }, + { + "epoch": 0.94, + "learning_rate": 4.7679037111334e-05, + "loss": 1.3395, + "theoretical_loss": 3.3403855646804095, + "tokens_seen": 3106275328 + }, + { + "epoch": 0.94, + "learning_rate": 4.766298896690071e-05, + "loss": 1.253, + "theoretical_loss": 3.340380230230171, + "tokens_seen": 3106340864 + }, + { + "epoch": 0.94, + "objective/train/docs_used": 1128625, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.2088756561279297, + "objective/train/theoretical_loss": 3.340374895923986, + "objective/train/tokens_used": 3126866400, + "theoretical_loss": 3.340374895923986, + "tokens_seen": 3106406400 + }, + { + "epoch": 0.94, + "learning_rate": 4.764694082246741e-05, + "loss": 1.207, + "theoretical_loss": 3.340374895923986, + "tokens_seen": 3106406400 + }, + { + "epoch": 0.94, + "learning_rate": 4.763089267803411e-05, + "loss": 1.2642, + "theoretical_loss": 3.340369561761848, + "tokens_seen": 3106471936 + }, + { + "epoch": 0.94, + "learning_rate": 4.76148445336008e-05, + "loss": 1.2936, + "theoretical_loss": 3.3403642277437497, + "tokens_seen": 3106537472 + }, + { + "epoch": 0.94, + "learning_rate": 4.7598796389167504e-05, + "loss": 1.1667, + "theoretical_loss": 3.3403588938696847, + "tokens_seen": 3106603008 + }, + { + "epoch": 0.94, + "learning_rate": 4.7582748244734204e-05, + "loss": 1.2542, + "theoretical_loss": 3.3403535601396452, + "tokens_seen": 3106668544 + }, + { + "epoch": 0.94, + "learning_rate": 4.7566700100300905e-05, + "loss": 1.2562, + "theoretical_loss": 3.3403482265536253, + "tokens_seen": 3106734080 + }, + { + "epoch": 0.94, + "learning_rate": 4.7550651955867605e-05, + "loss": 1.189, + "theoretical_loss": 3.3403428931116177, + "tokens_seen": 3106799616 + }, + { + "epoch": 0.94, + "learning_rate": 4.7534603811434306e-05, + "loss": 1.3568, + "theoretical_loss": 3.340337559813615, + "tokens_seen": 3106865152 + }, + { + "epoch": 0.94, + "learning_rate": 4.751855566700101e-05, + "loss": 1.1701, + "theoretical_loss": 3.3403322266596107, + "tokens_seen": 3106930688 + }, + { + "epoch": 0.94, + "learning_rate": 4.750250752256771e-05, + "loss": 1.1838, + "theoretical_loss": 3.340326893649598, + "tokens_seen": 3106996224 + }, + { + "epoch": 0.94, + "learning_rate": 4.748645937813441e-05, + "loss": 1.2767, + "theoretical_loss": 3.3403215607835697, + "tokens_seen": 3107061760 + }, + { + "epoch": 0.94, + "learning_rate": 4.747041123370111e-05, + "loss": 1.2005, + "theoretical_loss": 3.3403162280615186, + "tokens_seen": 3107127296 + }, + { + "epoch": 0.94, + "learning_rate": 4.745436308926781e-05, + "loss": 1.2257, + "theoretical_loss": 3.3403108954834386, + "tokens_seen": 3107192832 + }, + { + "epoch": 0.94, + "learning_rate": 4.74383149448345e-05, + "loss": 1.2654, + "theoretical_loss": 3.340305563049322, + "tokens_seen": 3107258368 + }, + { + "epoch": 0.94, + "learning_rate": 4.7422266800401204e-05, + "loss": 1.3098, + "theoretical_loss": 3.3403002307591625, + "tokens_seen": 3107323904 + }, + { + "epoch": 0.94, + "learning_rate": 4.7406218655967904e-05, + "loss": 1.2839, + "theoretical_loss": 3.340294898612953, + "tokens_seen": 3107389440 + }, + { + "epoch": 0.94, + "learning_rate": 4.7390170511534605e-05, + "loss": 1.3043, + "theoretical_loss": 3.340289566610686, + "tokens_seen": 3107454976 + }, + { + "epoch": 0.94, + "learning_rate": 4.7374122367101306e-05, + "loss": 1.3638, + "theoretical_loss": 3.3402842347523554, + "tokens_seen": 3107520512 + }, + { + "epoch": 0.94, + "learning_rate": 4.7358074222668006e-05, + "loss": 1.2368, + "theoretical_loss": 3.3402789030379534, + "tokens_seen": 3107586048 + }, + { + "epoch": 0.94, + "learning_rate": 4.734202607823471e-05, + "loss": 1.2313, + "theoretical_loss": 3.340273571467474, + "tokens_seen": 3107651584 + }, + { + "epoch": 0.94, + "learning_rate": 4.732597793380141e-05, + "loss": 1.2769, + "theoretical_loss": 3.34026824004091, + "tokens_seen": 3107717120 + }, + { + "epoch": 0.94, + "learning_rate": 4.730992978936811e-05, + "loss": 1.21, + "theoretical_loss": 3.3402629087582545, + "tokens_seen": 3107782656 + }, + { + "epoch": 0.94, + "learning_rate": 4.729388164493481e-05, + "loss": 1.2211, + "theoretical_loss": 3.3402575776195, + "tokens_seen": 3107848192 + }, + { + "epoch": 0.94, + "learning_rate": 4.727783350050151e-05, + "loss": 1.2513, + "theoretical_loss": 3.34025224662464, + "tokens_seen": 3107913728 + }, + { + "epoch": 0.94, + "learning_rate": 4.72617853560682e-05, + "loss": 1.3022, + "theoretical_loss": 3.3402469157736676, + "tokens_seen": 3107979264 + }, + { + "epoch": 0.94, + "learning_rate": 4.7245737211634904e-05, + "loss": 1.2144, + "theoretical_loss": 3.340241585066576, + "tokens_seen": 3108044800 + }, + { + "epoch": 0.94, + "learning_rate": 4.7229689067201605e-05, + "loss": 1.3155, + "theoretical_loss": 3.340236254503358, + "tokens_seen": 3108110336 + }, + { + "epoch": 0.94, + "learning_rate": 4.721364092276831e-05, + "loss": 1.26, + "theoretical_loss": 3.340230924084007, + "tokens_seen": 3108175872 + }, + { + "epoch": 0.94, + "learning_rate": 4.7197592778335006e-05, + "loss": 1.2451, + "theoretical_loss": 3.340225593808516, + "tokens_seen": 3108241408 + }, + { + "epoch": 0.94, + "learning_rate": 4.7181544633901706e-05, + "loss": 1.2078, + "theoretical_loss": 3.3402202636768776, + "tokens_seen": 3108306944 + }, + { + "epoch": 0.94, + "learning_rate": 4.716549648946841e-05, + "loss": 1.2546, + "theoretical_loss": 3.3402149336890856, + "tokens_seen": 3108372480 + }, + { + "epoch": 0.94, + "learning_rate": 4.714944834503511e-05, + "loss": 1.2367, + "theoretical_loss": 3.3402096038451328, + "tokens_seen": 3108438016 + }, + { + "epoch": 0.94, + "learning_rate": 4.713340020060181e-05, + "loss": 1.2214, + "theoretical_loss": 3.3402042741450124, + "tokens_seen": 3108503552 + }, + { + "epoch": 0.94, + "learning_rate": 4.711735205616851e-05, + "loss": 1.3032, + "theoretical_loss": 3.3401989445887175, + "tokens_seen": 3108569088 + }, + { + "epoch": 0.94, + "learning_rate": 4.71013039117352e-05, + "loss": 1.2632, + "theoretical_loss": 3.3401936151762404, + "tokens_seen": 3108634624 + }, + { + "epoch": 0.94, + "learning_rate": 4.7085255767301904e-05, + "loss": 1.0963, + "theoretical_loss": 3.3401882859075753, + "tokens_seen": 3108700160 + }, + { + "epoch": 0.94, + "learning_rate": 4.706920762286861e-05, + "loss": 1.1853, + "theoretical_loss": 3.340182956782715, + "tokens_seen": 3108765696 + }, + { + "epoch": 0.94, + "learning_rate": 4.705315947843531e-05, + "loss": 1.2698, + "theoretical_loss": 3.3401776278016517, + "tokens_seen": 3108831232 + }, + { + "epoch": 0.94, + "learning_rate": 4.703711133400201e-05, + "loss": 1.279, + "theoretical_loss": 3.3401722989643794, + "tokens_seen": 3108896768 + }, + { + "epoch": 0.94, + "learning_rate": 4.7021063189568706e-05, + "loss": 1.2049, + "theoretical_loss": 3.340166970270891, + "tokens_seen": 3108962304 + }, + { + "epoch": 0.94, + "learning_rate": 4.700501504513541e-05, + "loss": 1.2463, + "theoretical_loss": 3.34016164172118, + "tokens_seen": 3109027840 + }, + { + "epoch": 0.94, + "learning_rate": 4.698896690070211e-05, + "loss": 1.241, + "theoretical_loss": 3.340156313315239, + "tokens_seen": 3109093376 + }, + { + "epoch": 0.94, + "learning_rate": 4.697291875626881e-05, + "loss": 1.2469, + "theoretical_loss": 3.340150985053061, + "tokens_seen": 3109158912 + }, + { + "epoch": 0.94, + "learning_rate": 4.695687061183551e-05, + "loss": 1.35, + "theoretical_loss": 3.340145656934639, + "tokens_seen": 3109224448 + }, + { + "epoch": 0.94, + "learning_rate": 4.694082246740221e-05, + "loss": 1.2863, + "theoretical_loss": 3.3401403289599663, + "tokens_seen": 3109289984 + }, + { + "epoch": 0.94, + "learning_rate": 4.69247743229689e-05, + "loss": 1.2183, + "theoretical_loss": 3.3401350011290365, + "tokens_seen": 3109355520 + }, + { + "epoch": 0.94, + "learning_rate": 4.690872617853561e-05, + "loss": 1.2816, + "theoretical_loss": 3.340129673441842, + "tokens_seen": 3109421056 + }, + { + "epoch": 0.94, + "learning_rate": 4.689267803410231e-05, + "loss": 1.2184, + "theoretical_loss": 3.340124345898376, + "tokens_seen": 3109486592 + }, + { + "epoch": 0.94, + "learning_rate": 4.687662988966901e-05, + "loss": 1.2485, + "theoretical_loss": 3.3401190184986316, + "tokens_seen": 3109552128 + }, + { + "epoch": 0.94, + "learning_rate": 4.686058174523571e-05, + "loss": 1.2758, + "theoretical_loss": 3.340113691242602, + "tokens_seen": 3109617664 + }, + { + "epoch": 0.94, + "objective/train/docs_used": 1129708, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.3501499891281128, + "objective/train/theoretical_loss": 3.3401083641302804, + "objective/train/tokens_used": 3130143200, + "theoretical_loss": 3.3401083641302804, + "tokens_seen": 3109683200 + }, + { + "epoch": 0.94, + "learning_rate": 4.6844533600802406e-05, + "loss": 1.3316, + "theoretical_loss": 3.3401083641302804, + "tokens_seen": 3109683200 + }, + { + "epoch": 0.94, + "learning_rate": 4.682848545636911e-05, + "loss": 1.3131, + "theoretical_loss": 3.34010303716166, + "tokens_seen": 3109748736 + }, + { + "epoch": 0.94, + "learning_rate": 4.681243731193581e-05, + "loss": 1.416, + "theoretical_loss": 3.3400977103367335, + "tokens_seen": 3109814272 + }, + { + "epoch": 0.94, + "learning_rate": 4.679638916750251e-05, + "loss": 1.3304, + "theoretical_loss": 3.3400923836554943, + "tokens_seen": 3109879808 + }, + { + "epoch": 0.94, + "learning_rate": 4.678034102306921e-05, + "loss": 1.2865, + "theoretical_loss": 3.340087057117935, + "tokens_seen": 3109945344 + }, + { + "epoch": 0.94, + "learning_rate": 4.676429287863591e-05, + "loss": 1.2242, + "theoretical_loss": 3.3400817307240493, + "tokens_seen": 3110010880 + }, + { + "epoch": 0.94, + "learning_rate": 4.674824473420261e-05, + "loss": 1.2364, + "theoretical_loss": 3.3400764044738303, + "tokens_seen": 3110076416 + }, + { + "epoch": 0.94, + "learning_rate": 4.673219658976931e-05, + "loss": 1.2681, + "theoretical_loss": 3.3400710783672705, + "tokens_seen": 3110141952 + }, + { + "epoch": 0.94, + "learning_rate": 4.671614844533601e-05, + "loss": 1.3103, + "theoretical_loss": 3.3400657524043638, + "tokens_seen": 3110207488 + }, + { + "epoch": 0.94, + "learning_rate": 4.670010030090271e-05, + "loss": 1.2561, + "theoretical_loss": 3.3400604265851026, + "tokens_seen": 3110273024 + }, + { + "epoch": 0.94, + "learning_rate": 4.668405215646941e-05, + "loss": 1.3005, + "theoretical_loss": 3.34005510090948, + "tokens_seen": 3110338560 + }, + { + "epoch": 0.94, + "learning_rate": 4.6668004012036106e-05, + "loss": 1.2763, + "theoretical_loss": 3.34004977537749, + "tokens_seen": 3110404096 + }, + { + "epoch": 0.94, + "learning_rate": 4.665195586760281e-05, + "loss": 1.2817, + "theoretical_loss": 3.3400444499891244, + "tokens_seen": 3110469632 + }, + { + "epoch": 0.94, + "learning_rate": 4.663590772316951e-05, + "loss": 1.2696, + "theoretical_loss": 3.3400391247443775, + "tokens_seen": 3110535168 + }, + { + "epoch": 0.94, + "learning_rate": 4.6619859578736215e-05, + "loss": 1.2113, + "theoretical_loss": 3.3400337996432414, + "tokens_seen": 3110600704 + }, + { + "epoch": 0.94, + "learning_rate": 4.6603811434302916e-05, + "loss": 1.2831, + "theoretical_loss": 3.3400284746857096, + "tokens_seen": 3110666240 + }, + { + "epoch": 0.94, + "learning_rate": 4.658776328986961e-05, + "loss": 1.289, + "theoretical_loss": 3.340023149871776, + "tokens_seen": 3110731776 + }, + { + "epoch": 0.94, + "learning_rate": 4.657171514543631e-05, + "loss": 1.2477, + "theoretical_loss": 3.340017825201432, + "tokens_seen": 3110797312 + }, + { + "epoch": 0.94, + "learning_rate": 4.655566700100301e-05, + "loss": 1.235, + "theoretical_loss": 3.340012500674672, + "tokens_seen": 3110862848 + }, + { + "epoch": 0.94, + "learning_rate": 4.653961885656971e-05, + "loss": 1.1551, + "theoretical_loss": 3.3400071762914894, + "tokens_seen": 3110928384 + }, + { + "epoch": 0.94, + "learning_rate": 4.652357071213641e-05, + "loss": 1.3279, + "theoretical_loss": 3.340001852051876, + "tokens_seen": 3110993920 + }, + { + "epoch": 0.94, + "learning_rate": 4.650752256770311e-05, + "loss": 1.2198, + "theoretical_loss": 3.339996527955826, + "tokens_seen": 3111059456 + }, + { + "epoch": 0.94, + "learning_rate": 4.649147442326981e-05, + "loss": 1.4153, + "theoretical_loss": 3.339991204003332, + "tokens_seen": 3111124992 + }, + { + "epoch": 0.94, + "learning_rate": 4.647542627883651e-05, + "loss": 1.2746, + "theoretical_loss": 3.339985880194387, + "tokens_seen": 3111190528 + }, + { + "epoch": 0.94, + "learning_rate": 4.6459378134403215e-05, + "loss": 1.2251, + "theoretical_loss": 3.3399805565289844, + "tokens_seen": 3111256064 + }, + { + "epoch": 0.94, + "learning_rate": 4.6443329989969915e-05, + "loss": 1.2616, + "theoretical_loss": 3.339975233007117, + "tokens_seen": 3111321600 + }, + { + "epoch": 0.94, + "learning_rate": 4.6427281845536616e-05, + "loss": 1.3477, + "theoretical_loss": 3.3399699096287785, + "tokens_seen": 3111387136 + }, + { + "epoch": 0.94, + "learning_rate": 4.641123370110331e-05, + "loss": 1.2066, + "theoretical_loss": 3.339964586393961, + "tokens_seen": 3111452672 + }, + { + "epoch": 0.94, + "learning_rate": 4.639518555667001e-05, + "loss": 1.3022, + "theoretical_loss": 3.339959263302659, + "tokens_seen": 3111518208 + }, + { + "epoch": 0.94, + "learning_rate": 4.637913741223671e-05, + "loss": 1.2938, + "theoretical_loss": 3.3399539403548646, + "tokens_seen": 3111583744 + }, + { + "epoch": 0.94, + "learning_rate": 4.636308926780341e-05, + "loss": 1.353, + "theoretical_loss": 3.339948617550571, + "tokens_seen": 3111649280 + }, + { + "epoch": 0.94, + "learning_rate": 4.634704112337011e-05, + "loss": 1.3666, + "theoretical_loss": 3.3399432948897716, + "tokens_seen": 3111714816 + }, + { + "epoch": 0.94, + "learning_rate": 4.633099297893681e-05, + "loss": 1.3597, + "theoretical_loss": 3.339937972372459, + "tokens_seen": 3111780352 + }, + { + "epoch": 0.94, + "learning_rate": 4.6314944834503514e-05, + "loss": 1.2499, + "theoretical_loss": 3.339932649998627, + "tokens_seen": 3111845888 + }, + { + "epoch": 0.94, + "learning_rate": 4.6298896690070214e-05, + "loss": 1.3575, + "theoretical_loss": 3.3399273277682684, + "tokens_seen": 3111911424 + }, + { + "epoch": 0.94, + "learning_rate": 4.6282848545636915e-05, + "loss": 1.2242, + "theoretical_loss": 3.339922005681376, + "tokens_seen": 3111976960 + }, + { + "epoch": 0.94, + "learning_rate": 4.6266800401203616e-05, + "loss": 1.4241, + "theoretical_loss": 3.3399166837379433, + "tokens_seen": 3112042496 + }, + { + "epoch": 0.94, + "learning_rate": 4.6250752256770316e-05, + "loss": 1.3636, + "theoretical_loss": 3.339911361937964, + "tokens_seen": 3112108032 + }, + { + "epoch": 0.94, + "learning_rate": 4.623470411233701e-05, + "loss": 1.2766, + "theoretical_loss": 3.33990604028143, + "tokens_seen": 3112173568 + }, + { + "epoch": 0.94, + "learning_rate": 4.621865596790371e-05, + "loss": 1.3336, + "theoretical_loss": 3.3399007187683347, + "tokens_seen": 3112239104 + }, + { + "epoch": 0.94, + "learning_rate": 4.620260782347041e-05, + "loss": 1.2446, + "theoretical_loss": 3.3398953973986716, + "tokens_seen": 3112304640 + }, + { + "epoch": 0.94, + "learning_rate": 4.618655967903711e-05, + "loss": 1.1761, + "theoretical_loss": 3.3398900761724333, + "tokens_seen": 3112370176 + }, + { + "epoch": 0.94, + "learning_rate": 4.617051153460382e-05, + "loss": 1.3296, + "theoretical_loss": 3.339884755089614, + "tokens_seen": 3112435712 + }, + { + "epoch": 0.94, + "learning_rate": 4.615446339017051e-05, + "loss": 1.2353, + "theoretical_loss": 3.339879434150206, + "tokens_seen": 3112501248 + }, + { + "epoch": 0.94, + "learning_rate": 4.6138415245737214e-05, + "loss": 1.2579, + "theoretical_loss": 3.339874113354202, + "tokens_seen": 3112566784 + }, + { + "epoch": 0.94, + "learning_rate": 4.6122367101303914e-05, + "loss": 1.3243, + "theoretical_loss": 3.339868792701596, + "tokens_seen": 3112632320 + }, + { + "epoch": 0.94, + "learning_rate": 4.6106318956870615e-05, + "loss": 1.2678, + "theoretical_loss": 3.3398634721923806, + "tokens_seen": 3112697856 + }, + { + "epoch": 0.94, + "learning_rate": 4.6090270812437316e-05, + "loss": 1.2896, + "theoretical_loss": 3.339858151826549, + "tokens_seen": 3112763392 + }, + { + "epoch": 0.94, + "learning_rate": 4.6074222668004016e-05, + "loss": 1.2861, + "theoretical_loss": 3.339852831604095, + "tokens_seen": 3112828928 + }, + { + "epoch": 0.94, + "learning_rate": 4.605817452357071e-05, + "loss": 1.2761, + "theoretical_loss": 3.3398475115250106, + "tokens_seen": 3112894464 + }, + { + "epoch": 0.94, + "objective/train/docs_used": 1130876, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.3423080444335938, + "objective/train/theoretical_loss": 3.3398421915892893, + "objective/train/tokens_used": 3133420000, + "theoretical_loss": 3.3398421915892893, + "tokens_seen": 3112960000 + }, + { + "epoch": 0.94, + "learning_rate": 4.604212637913741e-05, + "loss": 1.2354, + "theoretical_loss": 3.3398421915892893, + "tokens_seen": 3112960000 + }, + { + "epoch": 0.94, + "learning_rate": 4.602607823470411e-05, + "loss": 1.2763, + "theoretical_loss": 3.3398368717969245, + "tokens_seen": 3113025536 + }, + { + "epoch": 0.94, + "learning_rate": 4.601003009027082e-05, + "loss": 1.1979, + "theoretical_loss": 3.3398315521479094, + "tokens_seen": 3113091072 + }, + { + "epoch": 0.94, + "learning_rate": 4.599398194583752e-05, + "loss": 1.312, + "theoretical_loss": 3.3398262326422365, + "tokens_seen": 3113156608 + }, + { + "epoch": 0.94, + "learning_rate": 4.5977933801404213e-05, + "loss": 1.2765, + "theoretical_loss": 3.3398209132798993, + "tokens_seen": 3113222144 + }, + { + "epoch": 0.94, + "learning_rate": 4.5961885656970914e-05, + "loss": 1.1337, + "theoretical_loss": 3.3398155940608913, + "tokens_seen": 3113287680 + }, + { + "epoch": 0.94, + "learning_rate": 4.5945837512537615e-05, + "loss": 1.2998, + "theoretical_loss": 3.339810274985205, + "tokens_seen": 3113353216 + }, + { + "epoch": 0.94, + "learning_rate": 4.5929789368104315e-05, + "loss": 1.3654, + "theoretical_loss": 3.339804956052834, + "tokens_seen": 3113418752 + }, + { + "epoch": 0.94, + "learning_rate": 4.5913741223671016e-05, + "loss": 1.1395, + "theoretical_loss": 3.339799637263771, + "tokens_seen": 3113484288 + }, + { + "epoch": 0.94, + "learning_rate": 4.5897693079237717e-05, + "loss": 1.3149, + "theoretical_loss": 3.339794318618009, + "tokens_seen": 3113549824 + }, + { + "epoch": 0.94, + "learning_rate": 4.588164493480441e-05, + "loss": 1.1392, + "theoretical_loss": 3.3397890001155415, + "tokens_seen": 3113615360 + }, + { + "epoch": 0.94, + "learning_rate": 4.586559679037112e-05, + "loss": 1.2459, + "theoretical_loss": 3.3397836817563618, + "tokens_seen": 3113680896 + }, + { + "epoch": 0.94, + "learning_rate": 4.584954864593782e-05, + "loss": 1.172, + "theoretical_loss": 3.3397783635404625, + "tokens_seen": 3113746432 + }, + { + "epoch": 0.94, + "learning_rate": 4.583350050150452e-05, + "loss": 1.1434, + "theoretical_loss": 3.339773045467837, + "tokens_seen": 3113811968 + }, + { + "epoch": 0.94, + "learning_rate": 4.581745235707122e-05, + "loss": 1.3277, + "theoretical_loss": 3.339767727538479, + "tokens_seen": 3113877504 + }, + { + "epoch": 0.94, + "learning_rate": 4.5801404212637914e-05, + "loss": 1.3428, + "theoretical_loss": 3.339762409752381, + "tokens_seen": 3113943040 + }, + { + "epoch": 0.94, + "learning_rate": 4.5785356068204614e-05, + "loss": 1.3041, + "theoretical_loss": 3.3397570921095356, + "tokens_seen": 3114008576 + }, + { + "epoch": 0.94, + "learning_rate": 4.5769307923771315e-05, + "loss": 1.3063, + "theoretical_loss": 3.3397517746099363, + "tokens_seen": 3114074112 + }, + { + "epoch": 0.94, + "learning_rate": 4.5753259779338016e-05, + "loss": 1.1853, + "theoretical_loss": 3.339746457253577, + "tokens_seen": 3114139648 + }, + { + "epoch": 0.94, + "learning_rate": 4.5737211634904716e-05, + "loss": 1.1547, + "theoretical_loss": 3.33974114004045, + "tokens_seen": 3114205184 + }, + { + "epoch": 0.94, + "learning_rate": 4.572116349047142e-05, + "loss": 1.1816, + "theoretical_loss": 3.3397358229705487, + "tokens_seen": 3114270720 + }, + { + "epoch": 0.94, + "learning_rate": 4.570511534603812e-05, + "loss": 1.2975, + "theoretical_loss": 3.3397305060438667, + "tokens_seen": 3114336256 + }, + { + "epoch": 0.94, + "learning_rate": 4.568906720160482e-05, + "loss": 1.2943, + "theoretical_loss": 3.339725189260396, + "tokens_seen": 3114401792 + }, + { + "epoch": 0.94, + "learning_rate": 4.567301905717152e-05, + "loss": 1.1182, + "theoretical_loss": 3.3397198726201305, + "tokens_seen": 3114467328 + }, + { + "epoch": 0.94, + "learning_rate": 4.565697091273822e-05, + "loss": 1.1409, + "theoretical_loss": 3.3397145561230634, + "tokens_seen": 3114532864 + }, + { + "epoch": 0.94, + "learning_rate": 4.564092276830492e-05, + "loss": 1.3395, + "theoretical_loss": 3.3397092397691877, + "tokens_seen": 3114598400 + }, + { + "epoch": 0.94, + "learning_rate": 4.5624874623871614e-05, + "loss": 1.2417, + "theoretical_loss": 3.339703923558496, + "tokens_seen": 3114663936 + }, + { + "epoch": 0.94, + "learning_rate": 4.5608826479438314e-05, + "loss": 1.1724, + "theoretical_loss": 3.3396986074909822, + "tokens_seen": 3114729472 + }, + { + "epoch": 0.94, + "learning_rate": 4.5592778335005015e-05, + "loss": 1.3422, + "theoretical_loss": 3.339693291566639, + "tokens_seen": 3114795008 + }, + { + "epoch": 0.94, + "learning_rate": 4.5576730190571716e-05, + "loss": 1.3071, + "theoretical_loss": 3.33968797578546, + "tokens_seen": 3114860544 + }, + { + "epoch": 0.94, + "learning_rate": 4.556068204613842e-05, + "loss": 1.2265, + "theoretical_loss": 3.3396826601474374, + "tokens_seen": 3114926080 + }, + { + "epoch": 0.94, + "learning_rate": 4.554463390170512e-05, + "loss": 1.2336, + "theoretical_loss": 3.339677344652565, + "tokens_seen": 3114991616 + }, + { + "epoch": 0.94, + "learning_rate": 4.552858575727182e-05, + "loss": 1.2746, + "theoretical_loss": 3.3396720293008366, + "tokens_seen": 3115057152 + }, + { + "epoch": 0.94, + "learning_rate": 4.551253761283852e-05, + "loss": 1.3499, + "theoretical_loss": 3.339666714092244, + "tokens_seen": 3115122688 + }, + { + "epoch": 0.94, + "learning_rate": 4.549648946840522e-05, + "loss": 1.2704, + "theoretical_loss": 3.339661399026781, + "tokens_seen": 3115188224 + }, + { + "epoch": 0.94, + "learning_rate": 4.548044132397192e-05, + "loss": 1.3211, + "theoretical_loss": 3.3396560841044405, + "tokens_seen": 3115253760 + }, + { + "epoch": 0.94, + "learning_rate": 4.546439317953862e-05, + "loss": 1.2404, + "theoretical_loss": 3.3396507693252158, + "tokens_seen": 3115319296 + }, + { + "epoch": 0.94, + "learning_rate": 4.5448345035105314e-05, + "loss": 1.2177, + "theoretical_loss": 3.3396454546891006, + "tokens_seen": 3115384832 + }, + { + "epoch": 0.94, + "learning_rate": 4.5432296890672015e-05, + "loss": 1.258, + "theoretical_loss": 3.339640140196087, + "tokens_seen": 3115450368 + }, + { + "epoch": 0.94, + "learning_rate": 4.541624874623872e-05, + "loss": 1.3244, + "theoretical_loss": 3.3396348258461686, + "tokens_seen": 3115515904 + }, + { + "epoch": 0.94, + "learning_rate": 4.540020060180542e-05, + "loss": 1.2418, + "theoretical_loss": 3.3396295116393384, + "tokens_seen": 3115581440 + }, + { + "epoch": 0.94, + "learning_rate": 4.538415245737212e-05, + "loss": 1.3424, + "theoretical_loss": 3.33962419757559, + "tokens_seen": 3115646976 + }, + { + "epoch": 0.94, + "learning_rate": 4.536810431293882e-05, + "loss": 1.2617, + "theoretical_loss": 3.3396188836549157, + "tokens_seen": 3115712512 + }, + { + "epoch": 0.94, + "learning_rate": 4.535205616850552e-05, + "loss": 1.1997, + "theoretical_loss": 3.3396135698773097, + "tokens_seen": 3115778048 + }, + { + "epoch": 0.94, + "learning_rate": 4.533600802407222e-05, + "loss": 1.2334, + "theoretical_loss": 3.339608256242764, + "tokens_seen": 3115843584 + }, + { + "epoch": 0.94, + "learning_rate": 4.531995987963892e-05, + "loss": 1.3007, + "theoretical_loss": 3.339602942751273, + "tokens_seen": 3115909120 + }, + { + "epoch": 0.94, + "learning_rate": 4.530391173520562e-05, + "loss": 1.2365, + "theoretical_loss": 3.339597629402829, + "tokens_seen": 3115974656 + }, + { + "epoch": 0.94, + "learning_rate": 4.528786359077232e-05, + "loss": 1.2084, + "theoretical_loss": 3.339592316197425, + "tokens_seen": 3116040192 + }, + { + "epoch": 0.94, + "learning_rate": 4.5271815446339014e-05, + "loss": 1.2711, + "theoretical_loss": 3.3395870031350547, + "tokens_seen": 3116105728 + }, + { + "epoch": 0.94, + "learning_rate": 4.525576730190572e-05, + "loss": 1.1976, + "theoretical_loss": 3.3395816902157107, + "tokens_seen": 3116171264 + }, + { + "epoch": 0.94, + "objective/train/docs_used": 1132106, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.2136402130126953, + "objective/train/theoretical_loss": 3.3395763774393865, + "objective/train/tokens_used": 3136696800, + "theoretical_loss": 3.3395763774393865, + "tokens_seen": 3116236800 + }, + { + "epoch": 0.94, + "learning_rate": 4.523971915747242e-05, + "loss": 1.2698, + "theoretical_loss": 3.3395763774393865, + "tokens_seen": 3116236800 + }, + { + "epoch": 0.94, + "learning_rate": 4.522367101303912e-05, + "loss": 1.261, + "theoretical_loss": 3.3395710648060755, + "tokens_seen": 3116302336 + }, + { + "epoch": 0.94, + "learning_rate": 4.520762286860582e-05, + "loss": 1.3295, + "theoretical_loss": 3.33956575231577, + "tokens_seen": 3116367872 + }, + { + "epoch": 0.94, + "learning_rate": 4.519157472417252e-05, + "loss": 1.2686, + "theoretical_loss": 3.339560439968464, + "tokens_seen": 3116433408 + }, + { + "epoch": 0.94, + "learning_rate": 4.517552657973922e-05, + "loss": 1.3177, + "theoretical_loss": 3.3395551277641506, + "tokens_seen": 3116498944 + }, + { + "epoch": 0.94, + "learning_rate": 4.515947843530592e-05, + "loss": 1.2855, + "theoretical_loss": 3.3395498157028225, + "tokens_seen": 3116564480 + }, + { + "epoch": 0.94, + "learning_rate": 4.514343029087262e-05, + "loss": 1.236, + "theoretical_loss": 3.339544503784473, + "tokens_seen": 3116630016 + }, + { + "epoch": 0.94, + "learning_rate": 4.512738214643932e-05, + "loss": 1.307, + "theoretical_loss": 3.3395391920090947, + "tokens_seen": 3116695552 + }, + { + "epoch": 0.94, + "learning_rate": 4.511133400200602e-05, + "loss": 1.3305, + "theoretical_loss": 3.3395338803766816, + "tokens_seen": 3116761088 + }, + { + "epoch": 0.94, + "learning_rate": 4.509528585757272e-05, + "loss": 1.3073, + "theoretical_loss": 3.3395285688872267, + "tokens_seen": 3116826624 + }, + { + "epoch": 0.94, + "learning_rate": 4.507923771313942e-05, + "loss": 1.3297, + "theoretical_loss": 3.3395232575407228, + "tokens_seen": 3116892160 + }, + { + "epoch": 0.94, + "learning_rate": 4.506318956870612e-05, + "loss": 1.2666, + "theoretical_loss": 3.339517946337163, + "tokens_seen": 3116957696 + }, + { + "epoch": 0.94, + "learning_rate": 4.504714142427282e-05, + "loss": 1.2931, + "theoretical_loss": 3.339512635276541, + "tokens_seen": 3117023232 + }, + { + "epoch": 0.94, + "learning_rate": 4.503109327983952e-05, + "loss": 1.386, + "theoretical_loss": 3.33950732435885, + "tokens_seen": 3117088768 + }, + { + "epoch": 0.94, + "learning_rate": 4.501504513540622e-05, + "loss": 1.3281, + "theoretical_loss": 3.339502013584082, + "tokens_seen": 3117154304 + }, + { + "epoch": 0.94, + "learning_rate": 4.499899699097292e-05, + "loss": 1.3134, + "theoretical_loss": 3.339496702952231, + "tokens_seen": 3117219840 + }, + { + "epoch": 0.94, + "learning_rate": 4.498294884653962e-05, + "loss": 1.3853, + "theoretical_loss": 3.3394913924632905, + "tokens_seen": 3117285376 + }, + { + "epoch": 0.94, + "learning_rate": 4.4966900702106326e-05, + "loss": 1.4082, + "theoretical_loss": 3.3394860821172534, + "tokens_seen": 3117350912 + }, + { + "epoch": 0.94, + "learning_rate": 4.495085255767302e-05, + "loss": 1.4015, + "theoretical_loss": 3.339480771914112, + "tokens_seen": 3117416448 + }, + { + "epoch": 0.94, + "learning_rate": 4.493480441323972e-05, + "loss": 1.3104, + "theoretical_loss": 3.3394754618538607, + "tokens_seen": 3117481984 + }, + { + "epoch": 0.94, + "learning_rate": 4.491875626880642e-05, + "loss": 1.2753, + "theoretical_loss": 3.3394701519364918, + "tokens_seen": 3117547520 + }, + { + "epoch": 0.94, + "learning_rate": 4.490270812437312e-05, + "loss": 1.336, + "theoretical_loss": 3.3394648421619983, + "tokens_seen": 3117613056 + }, + { + "epoch": 0.94, + "learning_rate": 4.488665997993982e-05, + "loss": 1.2712, + "theoretical_loss": 3.3394595325303746, + "tokens_seen": 3117678592 + }, + { + "epoch": 0.94, + "learning_rate": 4.487061183550652e-05, + "loss": 1.1729, + "theoretical_loss": 3.3394542230416127, + "tokens_seen": 3117744128 + }, + { + "epoch": 0.94, + "learning_rate": 4.485456369107322e-05, + "loss": 1.3853, + "theoretical_loss": 3.3394489136957057, + "tokens_seen": 3117809664 + }, + { + "epoch": 0.94, + "learning_rate": 4.483851554663992e-05, + "loss": 1.3973, + "theoretical_loss": 3.3394436044926477, + "tokens_seen": 3117875200 + }, + { + "epoch": 0.94, + "learning_rate": 4.482246740220662e-05, + "loss": 1.4622, + "theoretical_loss": 3.339438295432431, + "tokens_seen": 3117940736 + }, + { + "epoch": 0.94, + "learning_rate": 4.4806419257773326e-05, + "loss": 1.3661, + "theoretical_loss": 3.3394329865150487, + "tokens_seen": 3118006272 + }, + { + "epoch": 0.94, + "learning_rate": 4.4790371113340026e-05, + "loss": 1.3583, + "theoretical_loss": 3.3394276777404945, + "tokens_seen": 3118071808 + }, + { + "epoch": 0.94, + "learning_rate": 4.477432296890672e-05, + "loss": 1.2439, + "theoretical_loss": 3.3394223691087617, + "tokens_seen": 3118137344 + }, + { + "epoch": 0.94, + "learning_rate": 4.475827482447342e-05, + "loss": 1.2939, + "theoretical_loss": 3.339417060619843, + "tokens_seen": 3118202880 + }, + { + "epoch": 0.94, + "learning_rate": 4.474222668004012e-05, + "loss": 1.3273, + "theoretical_loss": 3.3394117522737314, + "tokens_seen": 3118268416 + }, + { + "epoch": 0.94, + "learning_rate": 4.472617853560682e-05, + "loss": 1.4027, + "theoretical_loss": 3.3394064440704208, + "tokens_seen": 3118333952 + }, + { + "epoch": 0.94, + "learning_rate": 4.471013039117352e-05, + "loss": 1.313, + "theoretical_loss": 3.3394011360099034, + "tokens_seen": 3118399488 + }, + { + "epoch": 0.94, + "learning_rate": 4.4694082246740224e-05, + "loss": 1.3595, + "theoretical_loss": 3.339395828092173, + "tokens_seen": 3118465024 + }, + { + "epoch": 0.95, + "learning_rate": 4.467803410230692e-05, + "loss": 1.368, + "theoretical_loss": 3.339390520317223, + "tokens_seen": 3118530560 + }, + { + "epoch": 0.95, + "learning_rate": 4.4661985957873625e-05, + "loss": 1.2875, + "theoretical_loss": 3.3393852126850456, + "tokens_seen": 3118596096 + }, + { + "epoch": 0.95, + "learning_rate": 4.4645937813440325e-05, + "loss": 1.1879, + "theoretical_loss": 3.3393799051956345, + "tokens_seen": 3118661632 + }, + { + "epoch": 0.95, + "learning_rate": 4.4629889669007026e-05, + "loss": 1.2656, + "theoretical_loss": 3.339374597848983, + "tokens_seen": 3118727168 + }, + { + "epoch": 0.95, + "learning_rate": 4.461384152457373e-05, + "loss": 1.3017, + "theoretical_loss": 3.3393692906450845, + "tokens_seen": 3118792704 + }, + { + "epoch": 0.95, + "learning_rate": 4.459779338014042e-05, + "loss": 1.3132, + "theoretical_loss": 3.3393639835839313, + "tokens_seen": 3118858240 + }, + { + "epoch": 0.95, + "learning_rate": 4.458174523570712e-05, + "loss": 1.1369, + "theoretical_loss": 3.3393586766655172, + "tokens_seen": 3118923776 + }, + { + "epoch": 0.95, + "learning_rate": 4.456569709127382e-05, + "loss": 1.4226, + "theoretical_loss": 3.339353369889835, + "tokens_seen": 3118989312 + }, + { + "epoch": 0.95, + "learning_rate": 4.454964894684052e-05, + "loss": 1.2543, + "theoretical_loss": 3.3393480632568786, + "tokens_seen": 3119054848 + }, + { + "epoch": 0.95, + "learning_rate": 4.453360080240722e-05, + "loss": 1.2874, + "theoretical_loss": 3.3393427567666403, + "tokens_seen": 3119120384 + }, + { + "epoch": 0.95, + "learning_rate": 4.4517552657973924e-05, + "loss": 1.2255, + "theoretical_loss": 3.339337450419114, + "tokens_seen": 3119185920 + }, + { + "epoch": 0.95, + "learning_rate": 4.4501504513540624e-05, + "loss": 1.2263, + "theoretical_loss": 3.339332144214292, + "tokens_seen": 3119251456 + }, + { + "epoch": 0.95, + "learning_rate": 4.4485456369107325e-05, + "loss": 1.316, + "theoretical_loss": 3.339326838152168, + "tokens_seen": 3119316992 + }, + { + "epoch": 0.95, + "learning_rate": 4.4469408224674026e-05, + "loss": 1.3506, + "theoretical_loss": 3.339321532232735, + "tokens_seen": 3119382528 + }, + { + "epoch": 0.95, + "learning_rate": 4.4453360080240726e-05, + "loss": 1.1759, + "theoretical_loss": 3.3393162264559866, + "tokens_seen": 3119448064 + }, + { + "epoch": 0.95, + "objective/train/docs_used": 1133291, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.493255376815796, + "objective/train/theoretical_loss": 3.3393109208219154, + "objective/train/tokens_used": 3139973600, + "theoretical_loss": 3.3393109208219154, + "tokens_seen": 3119513600 + }, + { + "epoch": 0.95, + "learning_rate": 4.443731193580743e-05, + "loss": 1.4145, + "theoretical_loss": 3.3393109208219154, + "tokens_seen": 3119513600 + }, + { + "epoch": 0.95, + "learning_rate": 4.442126379137412e-05, + "loss": 1.2506, + "theoretical_loss": 3.339305615330515, + "tokens_seen": 3119579136 + }, + { + "epoch": 0.95, + "learning_rate": 4.440521564694082e-05, + "loss": 1.2503, + "theoretical_loss": 3.339300309981778, + "tokens_seen": 3119644672 + }, + { + "epoch": 0.95, + "learning_rate": 4.438916750250752e-05, + "loss": 1.1358, + "theoretical_loss": 3.3392950047756984, + "tokens_seen": 3119710208 + }, + { + "epoch": 0.95, + "learning_rate": 4.437311935807422e-05, + "loss": 1.2988, + "theoretical_loss": 3.3392896997122685, + "tokens_seen": 3119775744 + }, + { + "epoch": 0.95, + "learning_rate": 4.435707121364093e-05, + "loss": 1.3806, + "theoretical_loss": 3.3392843947914823, + "tokens_seen": 3119841280 + }, + { + "epoch": 0.95, + "learning_rate": 4.4341023069207624e-05, + "loss": 1.4227, + "theoretical_loss": 3.3392790900133322, + "tokens_seen": 3119906816 + }, + { + "epoch": 0.95, + "learning_rate": 4.4324974924774325e-05, + "loss": 1.4391, + "theoretical_loss": 3.3392737853778116, + "tokens_seen": 3119972352 + }, + { + "epoch": 0.95, + "learning_rate": 4.4308926780341025e-05, + "loss": 1.2179, + "theoretical_loss": 3.339268480884914, + "tokens_seen": 3120037888 + }, + { + "epoch": 0.95, + "learning_rate": 4.4292878635907726e-05, + "loss": 1.3778, + "theoretical_loss": 3.339263176534632, + "tokens_seen": 3120103424 + }, + { + "epoch": 0.95, + "learning_rate": 4.4276830491474426e-05, + "loss": 1.2132, + "theoretical_loss": 3.3392578723269595, + "tokens_seen": 3120168960 + }, + { + "epoch": 0.95, + "learning_rate": 4.426078234704113e-05, + "loss": 1.2153, + "theoretical_loss": 3.339252568261889, + "tokens_seen": 3120234496 + }, + { + "epoch": 0.95, + "learning_rate": 4.424473420260782e-05, + "loss": 1.2875, + "theoretical_loss": 3.339247264339414, + "tokens_seen": 3120300032 + }, + { + "epoch": 0.95, + "learning_rate": 4.422868605817452e-05, + "loss": 1.2719, + "theoretical_loss": 3.339241960559528, + "tokens_seen": 3120365568 + }, + { + "epoch": 0.95, + "learning_rate": 4.421263791374123e-05, + "loss": 1.5051, + "theoretical_loss": 3.339236656922224, + "tokens_seen": 3120431104 + }, + { + "epoch": 0.95, + "learning_rate": 4.419658976930793e-05, + "loss": 1.3236, + "theoretical_loss": 3.339231353427494, + "tokens_seen": 3120496640 + }, + { + "epoch": 0.95, + "learning_rate": 4.418054162487463e-05, + "loss": 1.2577, + "theoretical_loss": 3.339226050075333, + "tokens_seen": 3120562176 + }, + { + "epoch": 0.95, + "learning_rate": 4.4164493480441324e-05, + "loss": 1.3424, + "theoretical_loss": 3.3392207468657324, + "tokens_seen": 3120627712 + }, + { + "epoch": 0.95, + "learning_rate": 4.4148445336008025e-05, + "loss": 1.2343, + "theoretical_loss": 3.339215443798687, + "tokens_seen": 3120693248 + }, + { + "epoch": 0.95, + "learning_rate": 4.4132397191574725e-05, + "loss": 1.3538, + "theoretical_loss": 3.339210140874189, + "tokens_seen": 3120758784 + }, + { + "epoch": 0.95, + "learning_rate": 4.4116349047141426e-05, + "loss": 1.3527, + "theoretical_loss": 3.339204838092232, + "tokens_seen": 3120824320 + }, + { + "epoch": 0.95, + "learning_rate": 4.410030090270813e-05, + "loss": 1.264, + "theoretical_loss": 3.339199535452809, + "tokens_seen": 3120889856 + }, + { + "epoch": 0.95, + "learning_rate": 4.408425275827483e-05, + "loss": 1.2999, + "theoretical_loss": 3.339194232955913, + "tokens_seen": 3120955392 + }, + { + "epoch": 0.95, + "learning_rate": 4.406820461384153e-05, + "loss": 1.3112, + "theoretical_loss": 3.3391889306015377, + "tokens_seen": 3121020928 + }, + { + "epoch": 0.95, + "learning_rate": 4.405215646940823e-05, + "loss": 1.1791, + "theoretical_loss": 3.3391836283896756, + "tokens_seen": 3121086464 + }, + { + "epoch": 0.95, + "learning_rate": 4.403610832497493e-05, + "loss": 1.2857, + "theoretical_loss": 3.3391783263203205, + "tokens_seen": 3121152000 + }, + { + "epoch": 0.95, + "learning_rate": 4.402006018054163e-05, + "loss": 1.2839, + "theoretical_loss": 3.3391730243934648, + "tokens_seen": 3121217536 + }, + { + "epoch": 0.95, + "learning_rate": 4.400401203610833e-05, + "loss": 1.1957, + "theoretical_loss": 3.339167722609103, + "tokens_seen": 3121283072 + }, + { + "epoch": 0.95, + "learning_rate": 4.3987963891675024e-05, + "loss": 1.3617, + "theoretical_loss": 3.339162420967227, + "tokens_seen": 3121348608 + }, + { + "epoch": 0.95, + "learning_rate": 4.3971915747241725e-05, + "loss": 1.2077, + "theoretical_loss": 3.33915711946783, + "tokens_seen": 3121414144 + }, + { + "epoch": 0.95, + "learning_rate": 4.3955867602808426e-05, + "loss": 1.3355, + "theoretical_loss": 3.3391518181109063, + "tokens_seen": 3121479680 + }, + { + "epoch": 0.95, + "learning_rate": 4.3939819458375126e-05, + "loss": 1.1544, + "theoretical_loss": 3.339146516896448, + "tokens_seen": 3121545216 + }, + { + "epoch": 0.95, + "learning_rate": 4.392377131394183e-05, + "loss": 1.2481, + "theoretical_loss": 3.3391412158244487, + "tokens_seen": 3121610752 + }, + { + "epoch": 0.95, + "learning_rate": 4.390772316950853e-05, + "loss": 1.1877, + "theoretical_loss": 3.339135914894902, + "tokens_seen": 3121676288 + }, + { + "epoch": 0.95, + "learning_rate": 4.389167502507523e-05, + "loss": 1.2463, + "theoretical_loss": 3.3391306141078, + "tokens_seen": 3121741824 + }, + { + "epoch": 0.95, + "learning_rate": 4.387562688064193e-05, + "loss": 1.3924, + "theoretical_loss": 3.339125313463137, + "tokens_seen": 3121807360 + }, + { + "epoch": 0.95, + "learning_rate": 4.385957873620863e-05, + "loss": 1.3418, + "theoretical_loss": 3.3391200129609055, + "tokens_seen": 3121872896 + }, + { + "epoch": 0.95, + "learning_rate": 4.384353059177533e-05, + "loss": 1.3839, + "theoretical_loss": 3.3391147126010985, + "tokens_seen": 3121938432 + }, + { + "epoch": 0.95, + "learning_rate": 4.382748244734203e-05, + "loss": 1.2121, + "theoretical_loss": 3.33910941238371, + "tokens_seen": 3122003968 + }, + { + "epoch": 0.95, + "learning_rate": 4.3811434302908725e-05, + "loss": 1.3215, + "theoretical_loss": 3.339104112308733, + "tokens_seen": 3122069504 + }, + { + "epoch": 0.95, + "learning_rate": 4.3795386158475425e-05, + "loss": 1.1681, + "theoretical_loss": 3.33909881237616, + "tokens_seen": 3122135040 + }, + { + "epoch": 0.95, + "learning_rate": 4.3779338014042126e-05, + "loss": 1.4165, + "theoretical_loss": 3.3390935125859844, + "tokens_seen": 3122200576 + }, + { + "epoch": 0.95, + "learning_rate": 4.376328986960883e-05, + "loss": 1.3045, + "theoretical_loss": 3.3390882129382002, + "tokens_seen": 3122266112 + }, + { + "epoch": 0.95, + "learning_rate": 4.3747241725175534e-05, + "loss": 1.2547, + "theoretical_loss": 3.3390829134328, + "tokens_seen": 3122331648 + }, + { + "epoch": 0.95, + "learning_rate": 4.373119358074223e-05, + "loss": 1.2922, + "theoretical_loss": 3.3390776140697764, + "tokens_seen": 3122397184 + }, + { + "epoch": 0.95, + "learning_rate": 4.371514543630893e-05, + "loss": 1.1832, + "theoretical_loss": 3.3390723148491235, + "tokens_seen": 3122462720 + }, + { + "epoch": 0.95, + "learning_rate": 4.369909729187563e-05, + "loss": 1.2678, + "theoretical_loss": 3.3390670157708344, + "tokens_seen": 3122528256 + }, + { + "epoch": 0.95, + "learning_rate": 4.368304914744233e-05, + "loss": 1.309, + "theoretical_loss": 3.3390617168349017, + "tokens_seen": 3122593792 + }, + { + "epoch": 0.95, + "learning_rate": 4.366700100300903e-05, + "loss": 1.2855, + "theoretical_loss": 3.3390564180413187, + "tokens_seen": 3122659328 + }, + { + "epoch": 0.95, + "learning_rate": 4.365095285857573e-05, + "loss": 1.3074, + "theoretical_loss": 3.339051119390079, + "tokens_seen": 3122724864 + }, + { + "epoch": 0.95, + "objective/train/docs_used": 1134384, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.3156681060791016, + "objective/train/theoretical_loss": 3.339045820881176, + "objective/train/tokens_used": 3143250400, + "theoretical_loss": 3.339045820881176, + "tokens_seen": 3122790400 + }, + { + "epoch": 0.95, + "learning_rate": 4.3634904714142425e-05, + "loss": 1.3445, + "theoretical_loss": 3.339045820881176, + "tokens_seen": 3122790400 + }, + { + "epoch": 0.95, + "learning_rate": 4.3618856569709125e-05, + "loss": 1.3336, + "theoretical_loss": 3.339040522514602, + "tokens_seen": 3122855936 + }, + { + "epoch": 0.95, + "learning_rate": 4.360280842527583e-05, + "loss": 1.2666, + "theoretical_loss": 3.339035224290351, + "tokens_seen": 3122921472 + }, + { + "epoch": 0.95, + "learning_rate": 4.3586760280842533e-05, + "loss": 1.2462, + "theoretical_loss": 3.3390299262084158, + "tokens_seen": 3122987008 + }, + { + "epoch": 0.95, + "learning_rate": 4.3570712136409234e-05, + "loss": 1.3206, + "theoretical_loss": 3.3390246282687897, + "tokens_seen": 3123052544 + }, + { + "epoch": 0.95, + "learning_rate": 4.355466399197593e-05, + "loss": 1.2269, + "theoretical_loss": 3.3390193304714657, + "tokens_seen": 3123118080 + }, + { + "epoch": 0.95, + "learning_rate": 4.353861584754263e-05, + "loss": 1.2685, + "theoretical_loss": 3.3390140328164373, + "tokens_seen": 3123183616 + }, + { + "epoch": 0.95, + "learning_rate": 4.352256770310933e-05, + "loss": 1.2097, + "theoretical_loss": 3.3390087353036972, + "tokens_seen": 3123249152 + }, + { + "epoch": 0.95, + "learning_rate": 4.350651955867603e-05, + "loss": 1.2815, + "theoretical_loss": 3.3390034379332394, + "tokens_seen": 3123314688 + }, + { + "epoch": 0.95, + "learning_rate": 4.349047141424273e-05, + "loss": 1.1575, + "theoretical_loss": 3.338998140705056, + "tokens_seen": 3123380224 + }, + { + "epoch": 0.95, + "learning_rate": 4.347442326980943e-05, + "loss": 1.3429, + "theoretical_loss": 3.338992843619142, + "tokens_seen": 3123445760 + }, + { + "epoch": 0.95, + "learning_rate": 4.345837512537613e-05, + "loss": 1.1642, + "theoretical_loss": 3.3389875466754884, + "tokens_seen": 3123511296 + }, + { + "epoch": 0.95, + "learning_rate": 4.344232698094283e-05, + "loss": 1.3765, + "theoretical_loss": 3.3389822498740895, + "tokens_seen": 3123576832 + }, + { + "epoch": 0.95, + "learning_rate": 4.342627883650953e-05, + "loss": 1.1775, + "theoretical_loss": 3.3389769532149387, + "tokens_seen": 3123642368 + }, + { + "epoch": 0.95, + "learning_rate": 4.3410230692076234e-05, + "loss": 1.3653, + "theoretical_loss": 3.3389716566980283, + "tokens_seen": 3123707904 + }, + { + "epoch": 0.95, + "learning_rate": 4.3394182547642934e-05, + "loss": 1.3277, + "theoretical_loss": 3.3389663603233526, + "tokens_seen": 3123773440 + }, + { + "epoch": 0.95, + "learning_rate": 4.337813440320963e-05, + "loss": 1.2593, + "theoretical_loss": 3.3389610640909044, + "tokens_seen": 3123838976 + }, + { + "epoch": 0.95, + "learning_rate": 4.336208625877633e-05, + "loss": 1.4248, + "theoretical_loss": 3.3389557680006767, + "tokens_seen": 3123904512 + }, + { + "epoch": 0.95, + "learning_rate": 4.334603811434303e-05, + "loss": 1.286, + "theoretical_loss": 3.338950472052663, + "tokens_seen": 3123970048 + }, + { + "epoch": 0.95, + "learning_rate": 4.332998996990973e-05, + "loss": 1.1727, + "theoretical_loss": 3.3389451762468556, + "tokens_seen": 3124035584 + }, + { + "epoch": 0.95, + "learning_rate": 4.331394182547643e-05, + "loss": 1.3043, + "theoretical_loss": 3.338939880583249, + "tokens_seen": 3124101120 + }, + { + "epoch": 0.95, + "learning_rate": 4.329789368104313e-05, + "loss": 1.2969, + "theoretical_loss": 3.3389345850618355, + "tokens_seen": 3124166656 + }, + { + "epoch": 0.95, + "learning_rate": 4.328184553660983e-05, + "loss": 1.3782, + "theoretical_loss": 3.3389292896826084, + "tokens_seen": 3124232192 + }, + { + "epoch": 0.95, + "learning_rate": 4.326579739217653e-05, + "loss": 1.2782, + "theoretical_loss": 3.3389239944455613, + "tokens_seen": 3124297728 + }, + { + "epoch": 0.95, + "learning_rate": 4.324974924774323e-05, + "loss": 1.3192, + "theoretical_loss": 3.338918699350687, + "tokens_seen": 3124363264 + }, + { + "epoch": 0.95, + "learning_rate": 4.3233701103309934e-05, + "loss": 1.38, + "theoretical_loss": 3.3389134043979793, + "tokens_seen": 3124428800 + }, + { + "epoch": 0.95, + "learning_rate": 4.321765295887663e-05, + "loss": 1.1776, + "theoretical_loss": 3.3389081095874307, + "tokens_seen": 3124494336 + }, + { + "epoch": 0.95, + "learning_rate": 4.320160481444333e-05, + "loss": 1.431, + "theoretical_loss": 3.338902814919035, + "tokens_seen": 3124559872 + }, + { + "epoch": 0.95, + "learning_rate": 4.318555667001003e-05, + "loss": 1.359, + "theoretical_loss": 3.3388975203927846, + "tokens_seen": 3124625408 + }, + { + "epoch": 0.95, + "learning_rate": 4.316950852557673e-05, + "loss": 1.2799, + "theoretical_loss": 3.3388922260086735, + "tokens_seen": 3124690944 + }, + { + "epoch": 0.95, + "learning_rate": 4.315346038114344e-05, + "loss": 1.2282, + "theoretical_loss": 3.3388869317666945, + "tokens_seen": 3124756480 + }, + { + "epoch": 0.95, + "learning_rate": 4.313741223671013e-05, + "loss": 1.2429, + "theoretical_loss": 3.338881637666841, + "tokens_seen": 3124822016 + }, + { + "epoch": 0.95, + "learning_rate": 4.312136409227683e-05, + "loss": 1.1937, + "theoretical_loss": 3.338876343709106, + "tokens_seen": 3124887552 + }, + { + "epoch": 0.95, + "learning_rate": 4.310531594784353e-05, + "loss": 1.2214, + "theoretical_loss": 3.3388710498934824, + "tokens_seen": 3124953088 + }, + { + "epoch": 0.95, + "learning_rate": 4.308926780341023e-05, + "loss": 1.3248, + "theoretical_loss": 3.338865756219964, + "tokens_seen": 3125018624 + }, + { + "epoch": 0.95, + "learning_rate": 4.3073219658976933e-05, + "loss": 1.3471, + "theoretical_loss": 3.3388604626885448, + "tokens_seen": 3125084160 + }, + { + "epoch": 0.95, + "learning_rate": 4.3057171514543634e-05, + "loss": 1.3076, + "theoretical_loss": 3.3388551692992157, + "tokens_seen": 3125149696 + }, + { + "epoch": 0.95, + "learning_rate": 4.304112337011033e-05, + "loss": 1.2767, + "theoretical_loss": 3.338849876051972, + "tokens_seen": 3125215232 + }, + { + "epoch": 0.95, + "learning_rate": 4.302507522567703e-05, + "loss": 1.2993, + "theoretical_loss": 3.338844582946806, + "tokens_seen": 3125280768 + }, + { + "epoch": 0.95, + "learning_rate": 4.3009027081243736e-05, + "loss": 1.2029, + "theoretical_loss": 3.338839289983711, + "tokens_seen": 3125346304 + }, + { + "epoch": 0.95, + "learning_rate": 4.2992978936810437e-05, + "loss": 1.3223, + "theoretical_loss": 3.3388339971626797, + "tokens_seen": 3125411840 + }, + { + "epoch": 0.95, + "learning_rate": 4.297693079237714e-05, + "loss": 1.2574, + "theoretical_loss": 3.3388287044837064, + "tokens_seen": 3125477376 + }, + { + "epoch": 0.95, + "learning_rate": 4.296088264794383e-05, + "loss": 1.3172, + "theoretical_loss": 3.338823411946784, + "tokens_seen": 3125542912 + }, + { + "epoch": 0.95, + "learning_rate": 4.294483450351053e-05, + "loss": 1.2353, + "theoretical_loss": 3.338818119551905, + "tokens_seen": 3125608448 + }, + { + "epoch": 0.95, + "learning_rate": 4.292878635907723e-05, + "loss": 1.2718, + "theoretical_loss": 3.338812827299063, + "tokens_seen": 3125673984 + }, + { + "epoch": 0.95, + "learning_rate": 4.291273821464393e-05, + "loss": 1.3108, + "theoretical_loss": 3.338807535188252, + "tokens_seen": 3125739520 + }, + { + "epoch": 0.95, + "learning_rate": 4.2896690070210634e-05, + "loss": 1.3326, + "theoretical_loss": 3.3388022432194635, + "tokens_seen": 3125805056 + }, + { + "epoch": 0.95, + "learning_rate": 4.2880641925777334e-05, + "loss": 1.2234, + "theoretical_loss": 3.3387969513926925, + "tokens_seen": 3125870592 + }, + { + "epoch": 0.95, + "learning_rate": 4.2864593781344035e-05, + "loss": 1.3046, + "theoretical_loss": 3.338791659707931, + "tokens_seen": 3125936128 + }, + { + "epoch": 0.95, + "learning_rate": 4.2848545636910736e-05, + "loss": 1.2185, + "theoretical_loss": 3.338786368165173, + "tokens_seen": 3126001664 + }, + { + "epoch": 0.95, + "objective/train/docs_used": 1135598, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.3123224973678589, + "objective/train/theoretical_loss": 3.338781076764411, + "objective/train/tokens_used": 3146527200, + "theoretical_loss": 3.338781076764411, + "tokens_seen": 3126067200 + }, + { + "epoch": 0.95, + "learning_rate": 4.2832497492477436e-05, + "loss": 1.2991, + "theoretical_loss": 3.338781076764411, + "tokens_seen": 3126067200 + }, + { + "epoch": 0.95, + "learning_rate": 4.281644934804414e-05, + "loss": 1.4235, + "theoretical_loss": 3.338775785505639, + "tokens_seen": 3126132736 + }, + { + "epoch": 0.95, + "learning_rate": 4.280040120361084e-05, + "loss": 1.3703, + "theoretical_loss": 3.338770494388849, + "tokens_seen": 3126198272 + }, + { + "epoch": 0.95, + "learning_rate": 4.278435305917753e-05, + "loss": 1.2896, + "theoretical_loss": 3.338765203414036, + "tokens_seen": 3126263808 + }, + { + "epoch": 0.95, + "learning_rate": 4.276830491474423e-05, + "loss": 1.1763, + "theoretical_loss": 3.3387599125811915, + "tokens_seen": 3126329344 + }, + { + "epoch": 0.95, + "learning_rate": 4.275225677031093e-05, + "loss": 1.1684, + "theoretical_loss": 3.3387546218903097, + "tokens_seen": 3126394880 + }, + { + "epoch": 0.95, + "learning_rate": 4.273620862587763e-05, + "loss": 1.1968, + "theoretical_loss": 3.338749331341383, + "tokens_seen": 3126460416 + }, + { + "epoch": 0.95, + "learning_rate": 4.2720160481444334e-05, + "loss": 1.4358, + "theoretical_loss": 3.338744040934406, + "tokens_seen": 3126525952 + }, + { + "epoch": 0.95, + "learning_rate": 4.2704112337011034e-05, + "loss": 1.2484, + "theoretical_loss": 3.3387387506693704, + "tokens_seen": 3126591488 + }, + { + "epoch": 0.95, + "learning_rate": 4.2688064192577735e-05, + "loss": 1.241, + "theoretical_loss": 3.3387334605462704, + "tokens_seen": 3126657024 + }, + { + "epoch": 0.95, + "learning_rate": 4.2672016048144436e-05, + "loss": 1.3693, + "theoretical_loss": 3.3387281705650986, + "tokens_seen": 3126722560 + }, + { + "epoch": 0.95, + "learning_rate": 4.2655967903711136e-05, + "loss": 1.3199, + "theoretical_loss": 3.3387228807258493, + "tokens_seen": 3126788096 + }, + { + "epoch": 0.95, + "learning_rate": 4.263991975927784e-05, + "loss": 1.3256, + "theoretical_loss": 3.338717591028514, + "tokens_seen": 3126853632 + }, + { + "epoch": 0.95, + "learning_rate": 4.262387161484454e-05, + "loss": 1.1788, + "theoretical_loss": 3.3387123014730875, + "tokens_seen": 3126919168 + }, + { + "epoch": 0.95, + "learning_rate": 4.260782347041123e-05, + "loss": 1.3227, + "theoretical_loss": 3.3387070120595617, + "tokens_seen": 3126984704 + }, + { + "epoch": 0.95, + "learning_rate": 4.259177532597793e-05, + "loss": 1.2763, + "theoretical_loss": 3.338701722787931, + "tokens_seen": 3127050240 + }, + { + "epoch": 0.95, + "learning_rate": 4.257572718154463e-05, + "loss": 1.2156, + "theoretical_loss": 3.338696433658188, + "tokens_seen": 3127115776 + }, + { + "epoch": 0.95, + "learning_rate": 4.255967903711134e-05, + "loss": 1.3614, + "theoretical_loss": 3.3386911446703262, + "tokens_seen": 3127181312 + }, + { + "epoch": 0.95, + "learning_rate": 4.254363089267804e-05, + "loss": 1.3018, + "theoretical_loss": 3.3386858558243384, + "tokens_seen": 3127246848 + }, + { + "epoch": 0.95, + "learning_rate": 4.2527582748244735e-05, + "loss": 1.3413, + "theoretical_loss": 3.3386805671202184, + "tokens_seen": 3127312384 + }, + { + "epoch": 0.95, + "learning_rate": 4.2511534603811435e-05, + "loss": 1.2943, + "theoretical_loss": 3.3386752785579588, + "tokens_seen": 3127377920 + }, + { + "epoch": 0.95, + "learning_rate": 4.2495486459378136e-05, + "loss": 1.2296, + "theoretical_loss": 3.338669990137553, + "tokens_seen": 3127443456 + }, + { + "epoch": 0.95, + "learning_rate": 4.2479438314944837e-05, + "loss": 1.3369, + "theoretical_loss": 3.3386647018589946, + "tokens_seen": 3127508992 + }, + { + "epoch": 0.95, + "learning_rate": 4.246339017051154e-05, + "loss": 1.2734, + "theoretical_loss": 3.3386594137222767, + "tokens_seen": 3127574528 + }, + { + "epoch": 0.95, + "learning_rate": 4.244734202607824e-05, + "loss": 1.2699, + "theoretical_loss": 3.338654125727392, + "tokens_seen": 3127640064 + }, + { + "epoch": 0.95, + "learning_rate": 4.243129388164493e-05, + "loss": 1.3795, + "theoretical_loss": 3.338648837874334, + "tokens_seen": 3127705600 + }, + { + "epoch": 0.95, + "learning_rate": 4.241524573721164e-05, + "loss": 1.222, + "theoretical_loss": 3.3386435501630967, + "tokens_seen": 3127771136 + }, + { + "epoch": 0.95, + "learning_rate": 4.239919759277834e-05, + "loss": 1.2625, + "theoretical_loss": 3.338638262593672, + "tokens_seen": 3127836672 + }, + { + "epoch": 0.95, + "learning_rate": 4.238314944834504e-05, + "loss": 1.4041, + "theoretical_loss": 3.338632975166054, + "tokens_seen": 3127902208 + }, + { + "epoch": 0.95, + "learning_rate": 4.236710130391174e-05, + "loss": 1.1604, + "theoretical_loss": 3.338627687880236, + "tokens_seen": 3127967744 + }, + { + "epoch": 0.95, + "learning_rate": 4.2351053159478435e-05, + "loss": 1.3072, + "theoretical_loss": 3.3386224007362104, + "tokens_seen": 3128033280 + }, + { + "epoch": 0.95, + "learning_rate": 4.2335005015045135e-05, + "loss": 1.2121, + "theoretical_loss": 3.3386171137339713, + "tokens_seen": 3128098816 + }, + { + "epoch": 0.95, + "learning_rate": 4.2318956870611836e-05, + "loss": 1.1755, + "theoretical_loss": 3.3386118268735117, + "tokens_seen": 3128164352 + }, + { + "epoch": 0.95, + "learning_rate": 4.230290872617854e-05, + "loss": 1.1322, + "theoretical_loss": 3.3386065401548244, + "tokens_seen": 3128229888 + }, + { + "epoch": 0.95, + "learning_rate": 4.228686058174524e-05, + "loss": 1.224, + "theoretical_loss": 3.3386012535779033, + "tokens_seen": 3128295424 + }, + { + "epoch": 0.95, + "learning_rate": 4.227081243731194e-05, + "loss": 1.2487, + "theoretical_loss": 3.338595967142741, + "tokens_seen": 3128360960 + }, + { + "epoch": 0.95, + "learning_rate": 4.225476429287864e-05, + "loss": 1.2566, + "theoretical_loss": 3.338590680849331, + "tokens_seen": 3128426496 + }, + { + "epoch": 0.95, + "learning_rate": 4.223871614844534e-05, + "loss": 1.3184, + "theoretical_loss": 3.3385853946976667, + "tokens_seen": 3128492032 + }, + { + "epoch": 0.95, + "learning_rate": 4.222266800401204e-05, + "loss": 1.2875, + "theoretical_loss": 3.3385801086877414, + "tokens_seen": 3128557568 + }, + { + "epoch": 0.95, + "learning_rate": 4.220661985957874e-05, + "loss": 1.2224, + "theoretical_loss": 3.3385748228195475, + "tokens_seen": 3128623104 + }, + { + "epoch": 0.95, + "learning_rate": 4.219057171514544e-05, + "loss": 1.2392, + "theoretical_loss": 3.338569537093079, + "tokens_seen": 3128688640 + }, + { + "epoch": 0.95, + "learning_rate": 4.2174523570712135e-05, + "loss": 1.2963, + "theoretical_loss": 3.338564251508329, + "tokens_seen": 3128754176 + }, + { + "epoch": 0.95, + "learning_rate": 4.2158475426278836e-05, + "loss": 1.2157, + "theoretical_loss": 3.338558966065291, + "tokens_seen": 3128819712 + }, + { + "epoch": 0.95, + "learning_rate": 4.2142427281845536e-05, + "loss": 1.3131, + "theoretical_loss": 3.3385536807639578, + "tokens_seen": 3128885248 + }, + { + "epoch": 0.95, + "learning_rate": 4.212637913741224e-05, + "loss": 1.2596, + "theoretical_loss": 3.3385483956043225, + "tokens_seen": 3128950784 + }, + { + "epoch": 0.95, + "learning_rate": 4.2110330992978944e-05, + "loss": 1.3149, + "theoretical_loss": 3.3385431105863788, + "tokens_seen": 3129016320 + }, + { + "epoch": 0.95, + "learning_rate": 4.209428284854564e-05, + "loss": 1.1507, + "theoretical_loss": 3.3385378257101195, + "tokens_seen": 3129081856 + }, + { + "epoch": 0.95, + "learning_rate": 4.207823470411234e-05, + "loss": 1.2683, + "theoretical_loss": 3.3385325409755384, + "tokens_seen": 3129147392 + }, + { + "epoch": 0.95, + "learning_rate": 4.206218655967904e-05, + "loss": 1.313, + "theoretical_loss": 3.338527256382628, + "tokens_seen": 3129212928 + }, + { + "epoch": 0.95, + "learning_rate": 4.204613841524574e-05, + "loss": 1.1708, + "theoretical_loss": 3.338521971931382, + "tokens_seen": 3129278464 + }, + { + "epoch": 0.95, + "objective/train/docs_used": 1136781, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.3042328357696533, + "objective/train/theoretical_loss": 3.338516687621794, + "objective/train/tokens_used": 3149804000, + "theoretical_loss": 3.338516687621794, + "tokens_seen": 3129344000 + }, + { + "epoch": 0.95, + "learning_rate": 4.203009027081244e-05, + "loss": 1.2256, + "theoretical_loss": 3.338516687621794, + "tokens_seen": 3129344000 + }, + { + "epoch": 0.95, + "learning_rate": 4.201404212637914e-05, + "loss": 1.26, + "theoretical_loss": 3.3385114034538566, + "tokens_seen": 3129409536 + }, + { + "epoch": 0.95, + "learning_rate": 4.1997993981945835e-05, + "loss": 1.3078, + "theoretical_loss": 3.338506119427563, + "tokens_seen": 3129475072 + }, + { + "epoch": 0.95, + "learning_rate": 4.1981945837512536e-05, + "loss": 1.2684, + "theoretical_loss": 3.3385008355429067, + "tokens_seen": 3129540608 + }, + { + "epoch": 0.95, + "learning_rate": 4.196589769307924e-05, + "loss": 1.2338, + "theoretical_loss": 3.3384955517998813, + "tokens_seen": 3129606144 + }, + { + "epoch": 0.95, + "learning_rate": 4.1949849548645944e-05, + "loss": 1.184, + "theoretical_loss": 3.338490268198479, + "tokens_seen": 3129671680 + }, + { + "epoch": 0.95, + "learning_rate": 4.1933801404212645e-05, + "loss": 1.2991, + "theoretical_loss": 3.3384849847386944, + "tokens_seen": 3129737216 + }, + { + "epoch": 0.95, + "learning_rate": 4.191775325977934e-05, + "loss": 1.2945, + "theoretical_loss": 3.3384797014205194, + "tokens_seen": 3129802752 + }, + { + "epoch": 0.95, + "learning_rate": 4.190170511534604e-05, + "loss": 1.3402, + "theoretical_loss": 3.338474418243948, + "tokens_seen": 3129868288 + }, + { + "epoch": 0.95, + "learning_rate": 4.188565697091274e-05, + "loss": 1.2891, + "theoretical_loss": 3.3384691352089737, + "tokens_seen": 3129933824 + }, + { + "epoch": 0.95, + "learning_rate": 4.186960882647944e-05, + "loss": 1.4029, + "theoretical_loss": 3.338463852315589, + "tokens_seen": 3129999360 + }, + { + "epoch": 0.95, + "learning_rate": 4.185356068204614e-05, + "loss": 1.3555, + "theoretical_loss": 3.3384585695637874, + "tokens_seen": 3130064896 + }, + { + "epoch": 0.95, + "learning_rate": 4.183751253761284e-05, + "loss": 1.2503, + "theoretical_loss": 3.3384532869535626, + "tokens_seen": 3130130432 + }, + { + "epoch": 0.95, + "learning_rate": 4.1821464393179535e-05, + "loss": 1.2494, + "theoretical_loss": 3.338448004484907, + "tokens_seen": 3130195968 + }, + { + "epoch": 0.95, + "learning_rate": 4.180541624874624e-05, + "loss": 1.2971, + "theoretical_loss": 3.338442722157815, + "tokens_seen": 3130261504 + }, + { + "epoch": 0.95, + "learning_rate": 4.1789368104312944e-05, + "loss": 1.3106, + "theoretical_loss": 3.3384374399722785, + "tokens_seen": 3130327040 + }, + { + "epoch": 0.95, + "learning_rate": 4.1773319959879644e-05, + "loss": 1.4477, + "theoretical_loss": 3.338432157928292, + "tokens_seen": 3130392576 + }, + { + "epoch": 0.95, + "learning_rate": 4.1757271815446345e-05, + "loss": 1.2036, + "theoretical_loss": 3.3384268760258475, + "tokens_seen": 3130458112 + }, + { + "epoch": 0.95, + "learning_rate": 4.174122367101304e-05, + "loss": 1.2813, + "theoretical_loss": 3.338421594264939, + "tokens_seen": 3130523648 + }, + { + "epoch": 0.95, + "learning_rate": 4.172517552657974e-05, + "loss": 1.2473, + "theoretical_loss": 3.3384163126455597, + "tokens_seen": 3130589184 + }, + { + "epoch": 0.95, + "learning_rate": 4.170912738214644e-05, + "loss": 1.3193, + "theoretical_loss": 3.338411031167703, + "tokens_seen": 3130654720 + }, + { + "epoch": 0.95, + "learning_rate": 4.169307923771314e-05, + "loss": 1.2861, + "theoretical_loss": 3.3384057498313617, + "tokens_seen": 3130720256 + }, + { + "epoch": 0.95, + "learning_rate": 4.167703109327984e-05, + "loss": 1.2404, + "theoretical_loss": 3.338400468636529, + "tokens_seen": 3130785792 + }, + { + "epoch": 0.95, + "learning_rate": 4.166098294884654e-05, + "loss": 1.2418, + "theoretical_loss": 3.338395187583199, + "tokens_seen": 3130851328 + }, + { + "epoch": 0.95, + "learning_rate": 4.164493480441324e-05, + "loss": 1.1932, + "theoretical_loss": 3.338389906671364, + "tokens_seen": 3130916864 + }, + { + "epoch": 0.95, + "learning_rate": 4.162888665997994e-05, + "loss": 1.206, + "theoretical_loss": 3.3383846259010177, + "tokens_seen": 3130982400 + }, + { + "epoch": 0.95, + "learning_rate": 4.1612838515546644e-05, + "loss": 1.289, + "theoretical_loss": 3.3383793452721533, + "tokens_seen": 3131047936 + }, + { + "epoch": 0.95, + "learning_rate": 4.1596790371113344e-05, + "loss": 1.2717, + "theoretical_loss": 3.338374064784764, + "tokens_seen": 3131113472 + }, + { + "epoch": 0.95, + "learning_rate": 4.1580742226680045e-05, + "loss": 1.3277, + "theoretical_loss": 3.3383687844388428, + "tokens_seen": 3131179008 + }, + { + "epoch": 0.95, + "learning_rate": 4.156469408224674e-05, + "loss": 1.345, + "theoretical_loss": 3.3383635042343833, + "tokens_seen": 3131244544 + }, + { + "epoch": 0.95, + "learning_rate": 4.154864593781344e-05, + "loss": 1.3083, + "theoretical_loss": 3.3383582241713787, + "tokens_seen": 3131310080 + }, + { + "epoch": 0.95, + "learning_rate": 4.153259779338014e-05, + "loss": 1.2177, + "theoretical_loss": 3.3383529442498228, + "tokens_seen": 3131375616 + }, + { + "epoch": 0.95, + "learning_rate": 4.151654964894684e-05, + "loss": 1.1849, + "theoretical_loss": 3.3383476644697074, + "tokens_seen": 3131441152 + }, + { + "epoch": 0.95, + "learning_rate": 4.150050150451355e-05, + "loss": 1.2296, + "theoretical_loss": 3.338342384831027, + "tokens_seen": 3131506688 + }, + { + "epoch": 0.95, + "learning_rate": 4.148445336008024e-05, + "loss": 1.2236, + "theoretical_loss": 3.3383371053337743, + "tokens_seen": 3131572224 + }, + { + "epoch": 0.95, + "learning_rate": 4.146840521564694e-05, + "loss": 1.2424, + "theoretical_loss": 3.3383318259779426, + "tokens_seen": 3131637760 + }, + { + "epoch": 0.95, + "learning_rate": 4.145235707121364e-05, + "loss": 1.3231, + "theoretical_loss": 3.3383265467635255, + "tokens_seen": 3131703296 + }, + { + "epoch": 0.95, + "learning_rate": 4.1436308926780344e-05, + "loss": 1.3306, + "theoretical_loss": 3.3383212676905156, + "tokens_seen": 3131768832 + }, + { + "epoch": 0.95, + "learning_rate": 4.1420260782347045e-05, + "loss": 1.1872, + "theoretical_loss": 3.3383159887589073, + "tokens_seen": 3131834368 + }, + { + "epoch": 0.95, + "learning_rate": 4.1404212637913745e-05, + "loss": 1.1881, + "theoretical_loss": 3.3383107099686926, + "tokens_seen": 3131899904 + }, + { + "epoch": 0.95, + "learning_rate": 4.138816449348044e-05, + "loss": 1.2567, + "theoretical_loss": 3.3383054313198652, + "tokens_seen": 3131965440 + }, + { + "epoch": 0.95, + "learning_rate": 4.137211634904714e-05, + "loss": 1.3367, + "theoretical_loss": 3.3383001528124185, + "tokens_seen": 3132030976 + }, + { + "epoch": 0.95, + "learning_rate": 4.135606820461385e-05, + "loss": 1.2245, + "theoretical_loss": 3.338294874446346, + "tokens_seen": 3132096512 + }, + { + "epoch": 0.95, + "learning_rate": 4.134002006018055e-05, + "loss": 1.2597, + "theoretical_loss": 3.3382895962216406, + "tokens_seen": 3132162048 + }, + { + "epoch": 0.95, + "learning_rate": 4.132397191574724e-05, + "loss": 1.2721, + "theoretical_loss": 3.338284318138295, + "tokens_seen": 3132227584 + }, + { + "epoch": 0.95, + "learning_rate": 4.130792377131394e-05, + "loss": 1.4066, + "theoretical_loss": 3.3382790401963036, + "tokens_seen": 3132293120 + }, + { + "epoch": 0.95, + "learning_rate": 4.129187562688064e-05, + "loss": 1.2649, + "theoretical_loss": 3.3382737623956587, + "tokens_seen": 3132358656 + }, + { + "epoch": 0.95, + "learning_rate": 4.1275827482447343e-05, + "loss": 1.1458, + "theoretical_loss": 3.3382684847363544, + "tokens_seen": 3132424192 + }, + { + "epoch": 0.95, + "learning_rate": 4.1259779338014044e-05, + "loss": 1.2942, + "theoretical_loss": 3.3382632072183833, + "tokens_seen": 3132489728 + }, + { + "epoch": 0.95, + "learning_rate": 4.1243731193580745e-05, + "loss": 1.3092, + "theoretical_loss": 3.3382579298417387, + "tokens_seen": 3132555264 + }, + { + "epoch": 0.95, + "objective/train/docs_used": 1137994, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.3896937370300293, + "objective/train/theoretical_loss": 3.3382526526064145, + "objective/train/tokens_used": 3153080800, + "theoretical_loss": 3.3382526526064145, + "tokens_seen": 3132620800 + }, + { + "epoch": 0.95, + "learning_rate": 4.122768304914744e-05, + "loss": 1.3328, + "theoretical_loss": 3.3382526526064145, + "tokens_seen": 3132620800 + }, + { + "epoch": 0.95, + "learning_rate": 4.1211634904714146e-05, + "loss": 1.281, + "theoretical_loss": 3.338247375512403, + "tokens_seen": 3132686336 + }, + { + "epoch": 0.95, + "learning_rate": 4.119558676028085e-05, + "loss": 1.1411, + "theoretical_loss": 3.3382420985596983, + "tokens_seen": 3132751872 + }, + { + "epoch": 0.95, + "learning_rate": 4.117953861584755e-05, + "loss": 1.299, + "theoretical_loss": 3.338236821748293, + "tokens_seen": 3132817408 + }, + { + "epoch": 0.95, + "learning_rate": 4.116349047141425e-05, + "loss": 1.2851, + "theoretical_loss": 3.338231545078181, + "tokens_seen": 3132882944 + }, + { + "epoch": 0.95, + "learning_rate": 4.114744232698094e-05, + "loss": 1.2278, + "theoretical_loss": 3.338226268549355, + "tokens_seen": 3132948480 + }, + { + "epoch": 0.95, + "learning_rate": 4.113139418254764e-05, + "loss": 1.3345, + "theoretical_loss": 3.338220992161809, + "tokens_seen": 3133014016 + }, + { + "epoch": 0.95, + "learning_rate": 4.111534603811434e-05, + "loss": 1.2181, + "theoretical_loss": 3.3382157159155352, + "tokens_seen": 3133079552 + }, + { + "epoch": 0.95, + "learning_rate": 4.1099297893681044e-05, + "loss": 1.2021, + "theoretical_loss": 3.3382104398105277, + "tokens_seen": 3133145088 + }, + { + "epoch": 0.95, + "learning_rate": 4.1083249749247744e-05, + "loss": 1.2578, + "theoretical_loss": 3.338205163846779, + "tokens_seen": 3133210624 + }, + { + "epoch": 0.95, + "learning_rate": 4.1067201604814445e-05, + "loss": 1.1411, + "theoretical_loss": 3.3381998880242834, + "tokens_seen": 3133276160 + }, + { + "epoch": 0.95, + "learning_rate": 4.1051153460381146e-05, + "loss": 1.244, + "theoretical_loss": 3.3381946123430337, + "tokens_seen": 3133341696 + }, + { + "epoch": 0.95, + "learning_rate": 4.1035105315947846e-05, + "loss": 1.2805, + "theoretical_loss": 3.3381893368030227, + "tokens_seen": 3133407232 + }, + { + "epoch": 0.95, + "learning_rate": 4.101905717151455e-05, + "loss": 1.196, + "theoretical_loss": 3.338184061404244, + "tokens_seen": 3133472768 + }, + { + "epoch": 0.95, + "learning_rate": 4.100300902708125e-05, + "loss": 1.2302, + "theoretical_loss": 3.338178786146691, + "tokens_seen": 3133538304 + }, + { + "epoch": 0.95, + "learning_rate": 4.098696088264795e-05, + "loss": 1.1155, + "theoretical_loss": 3.338173511030357, + "tokens_seen": 3133603840 + }, + { + "epoch": 0.95, + "learning_rate": 4.097091273821464e-05, + "loss": 1.2484, + "theoretical_loss": 3.3381682360552354, + "tokens_seen": 3133669376 + }, + { + "epoch": 0.95, + "learning_rate": 4.095486459378134e-05, + "loss": 1.3924, + "theoretical_loss": 3.3381629612213186, + "tokens_seen": 3133734912 + }, + { + "epoch": 0.95, + "learning_rate": 4.093881644934804e-05, + "loss": 1.2852, + "theoretical_loss": 3.338157686528601, + "tokens_seen": 3133800448 + }, + { + "epoch": 0.95, + "learning_rate": 4.0922768304914744e-05, + "loss": 1.1819, + "theoretical_loss": 3.338152411977075, + "tokens_seen": 3133865984 + }, + { + "epoch": 0.95, + "learning_rate": 4.090672016048145e-05, + "loss": 1.2091, + "theoretical_loss": 3.338147137566734, + "tokens_seen": 3133931520 + }, + { + "epoch": 0.95, + "learning_rate": 4.0890672016048145e-05, + "loss": 1.203, + "theoretical_loss": 3.3381418632975723, + "tokens_seen": 3133997056 + }, + { + "epoch": 0.95, + "learning_rate": 4.0874623871614846e-05, + "loss": 1.2183, + "theoretical_loss": 3.3381365891695816, + "tokens_seen": 3134062592 + }, + { + "epoch": 0.95, + "learning_rate": 4.0858575727181546e-05, + "loss": 1.2249, + "theoretical_loss": 3.338131315182756, + "tokens_seen": 3134128128 + }, + { + "epoch": 0.95, + "learning_rate": 4.084252758274825e-05, + "loss": 1.3625, + "theoretical_loss": 3.3381260413370892, + "tokens_seen": 3134193664 + }, + { + "epoch": 0.95, + "learning_rate": 4.082647943831495e-05, + "loss": 1.2603, + "theoretical_loss": 3.3381207676325735, + "tokens_seen": 3134259200 + }, + { + "epoch": 0.95, + "learning_rate": 4.081043129388165e-05, + "loss": 1.3234, + "theoretical_loss": 3.3381154940692026, + "tokens_seen": 3134324736 + }, + { + "epoch": 0.95, + "learning_rate": 4.079438314944834e-05, + "loss": 1.2973, + "theoretical_loss": 3.3381102206469704, + "tokens_seen": 3134390272 + }, + { + "epoch": 0.95, + "learning_rate": 4.077833500501504e-05, + "loss": 1.3012, + "theoretical_loss": 3.338104947365869, + "tokens_seen": 3134455808 + }, + { + "epoch": 0.95, + "learning_rate": 4.076228686058175e-05, + "loss": 1.2847, + "theoretical_loss": 3.338099674225892, + "tokens_seen": 3134521344 + }, + { + "epoch": 0.95, + "learning_rate": 4.074623871614845e-05, + "loss": 1.2413, + "theoretical_loss": 3.3380944012270337, + "tokens_seen": 3134586880 + }, + { + "epoch": 0.95, + "learning_rate": 4.073019057171515e-05, + "loss": 1.1975, + "theoretical_loss": 3.338089128369286, + "tokens_seen": 3134652416 + }, + { + "epoch": 0.95, + "learning_rate": 4.0714142427281845e-05, + "loss": 1.2911, + "theoretical_loss": 3.338083855652643, + "tokens_seen": 3134717952 + }, + { + "epoch": 0.95, + "learning_rate": 4.0698094282848546e-05, + "loss": 1.1887, + "theoretical_loss": 3.3380785830770976, + "tokens_seen": 3134783488 + }, + { + "epoch": 0.95, + "learning_rate": 4.068204613841525e-05, + "loss": 1.3269, + "theoretical_loss": 3.3380733106426432, + "tokens_seen": 3134849024 + }, + { + "epoch": 0.95, + "learning_rate": 4.066599799398195e-05, + "loss": 1.3265, + "theoretical_loss": 3.3380680383492733, + "tokens_seen": 3134914560 + }, + { + "epoch": 0.95, + "learning_rate": 4.064994984954865e-05, + "loss": 1.3078, + "theoretical_loss": 3.338062766196981, + "tokens_seen": 3134980096 + }, + { + "epoch": 0.95, + "learning_rate": 4.063390170511535e-05, + "loss": 1.2358, + "theoretical_loss": 3.338057494185759, + "tokens_seen": 3135045632 + }, + { + "epoch": 0.95, + "learning_rate": 4.061785356068204e-05, + "loss": 1.235, + "theoretical_loss": 3.3380522223156017, + "tokens_seen": 3135111168 + }, + { + "epoch": 0.95, + "learning_rate": 4.060180541624875e-05, + "loss": 1.2654, + "theoretical_loss": 3.338046950586502, + "tokens_seen": 3135176704 + }, + { + "epoch": 0.95, + "learning_rate": 4.058575727181545e-05, + "loss": 1.248, + "theoretical_loss": 3.338041678998452, + "tokens_seen": 3135242240 + }, + { + "epoch": 0.95, + "learning_rate": 4.056970912738215e-05, + "loss": 1.2749, + "theoretical_loss": 3.3380364075514466, + "tokens_seen": 3135307776 + }, + { + "epoch": 0.95, + "learning_rate": 4.055366098294885e-05, + "loss": 1.226, + "theoretical_loss": 3.3380311362454784, + "tokens_seen": 3135373312 + }, + { + "epoch": 0.95, + "learning_rate": 4.0537612838515546e-05, + "loss": 1.1356, + "theoretical_loss": 3.3380258650805406, + "tokens_seen": 3135438848 + }, + { + "epoch": 0.95, + "learning_rate": 4.0521564694082246e-05, + "loss": 1.3433, + "theoretical_loss": 3.3380205940566268, + "tokens_seen": 3135504384 + }, + { + "epoch": 0.95, + "learning_rate": 4.050551654964895e-05, + "loss": 1.2869, + "theoretical_loss": 3.33801532317373, + "tokens_seen": 3135569920 + }, + { + "epoch": 0.95, + "learning_rate": 4.048946840521565e-05, + "loss": 1.4233, + "theoretical_loss": 3.3380100524318435, + "tokens_seen": 3135635456 + }, + { + "epoch": 0.95, + "learning_rate": 4.047342026078235e-05, + "loss": 1.2805, + "theoretical_loss": 3.33800478183096, + "tokens_seen": 3135700992 + }, + { + "epoch": 0.95, + "learning_rate": 4.045737211634905e-05, + "loss": 1.2742, + "theoretical_loss": 3.3379995113710743, + "tokens_seen": 3135766528 + }, + { + "epoch": 0.95, + "learning_rate": 4.044132397191575e-05, + "loss": 1.3846, + "theoretical_loss": 3.3379942410521783, + "tokens_seen": 3135832064 + }, + { + "epoch": 0.95, + "objective/train/docs_used": 1139200, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.490768551826477, + "objective/train/theoretical_loss": 3.337988970874266, + "objective/train/tokens_used": 3156357600, + "theoretical_loss": 3.337988970874266, + "tokens_seen": 3135897600 + }, + { + "epoch": 0.95, + "learning_rate": 4.042527582748245e-05, + "loss": 1.4031, + "theoretical_loss": 3.337988970874266, + "tokens_seen": 3135897600 + }, + { + "epoch": 0.95, + "learning_rate": 4.040922768304915e-05, + "loss": 1.3374, + "theoretical_loss": 3.3379837008373303, + "tokens_seen": 3135963136 + }, + { + "epoch": 0.95, + "learning_rate": 4.039317953861585e-05, + "loss": 1.285, + "theoretical_loss": 3.3379784309413645, + "tokens_seen": 3136028672 + }, + { + "epoch": 0.95, + "learning_rate": 4.037713139418255e-05, + "loss": 1.1863, + "theoretical_loss": 3.3379731611863623, + "tokens_seen": 3136094208 + }, + { + "epoch": 0.95, + "learning_rate": 4.0361083249749246e-05, + "loss": 1.1974, + "theoretical_loss": 3.3379678915723168, + "tokens_seen": 3136159744 + }, + { + "epoch": 0.95, + "learning_rate": 4.0345035105315946e-05, + "loss": 1.2576, + "theoretical_loss": 3.337962622099221, + "tokens_seen": 3136225280 + }, + { + "epoch": 0.95, + "learning_rate": 4.032898696088265e-05, + "loss": 1.2145, + "theoretical_loss": 3.3379573527670683, + "tokens_seen": 3136290816 + }, + { + "epoch": 0.95, + "learning_rate": 4.0312938816449354e-05, + "loss": 1.2812, + "theoretical_loss": 3.337952083575852, + "tokens_seen": 3136356352 + }, + { + "epoch": 0.95, + "learning_rate": 4.0296890672016055e-05, + "loss": 1.2533, + "theoretical_loss": 3.3379468145255657, + "tokens_seen": 3136421888 + }, + { + "epoch": 0.95, + "learning_rate": 4.028084252758275e-05, + "loss": 1.294, + "theoretical_loss": 3.3379415456162023, + "tokens_seen": 3136487424 + }, + { + "epoch": 0.95, + "learning_rate": 4.026479438314945e-05, + "loss": 1.2913, + "theoretical_loss": 3.337936276847755, + "tokens_seen": 3136552960 + }, + { + "epoch": 0.95, + "learning_rate": 4.024874623871615e-05, + "loss": 1.2297, + "theoretical_loss": 3.3379310082202176, + "tokens_seen": 3136618496 + }, + { + "epoch": 0.95, + "learning_rate": 4.023269809428285e-05, + "loss": 1.1987, + "theoretical_loss": 3.337925739733583, + "tokens_seen": 3136684032 + }, + { + "epoch": 0.95, + "learning_rate": 4.021664994984955e-05, + "loss": 1.3196, + "theoretical_loss": 3.3379204713878448, + "tokens_seen": 3136749568 + }, + { + "epoch": 0.95, + "learning_rate": 4.020060180541625e-05, + "loss": 1.3623, + "theoretical_loss": 3.3379152031829955, + "tokens_seen": 3136815104 + }, + { + "epoch": 0.95, + "learning_rate": 4.0184553660982946e-05, + "loss": 1.1519, + "theoretical_loss": 3.337909935119029, + "tokens_seen": 3136880640 + }, + { + "epoch": 0.95, + "learning_rate": 4.0168505516549647e-05, + "loss": 1.3128, + "theoretical_loss": 3.3379046671959394, + "tokens_seen": 3136946176 + }, + { + "epoch": 0.95, + "learning_rate": 4.0152457372116354e-05, + "loss": 1.243, + "theoretical_loss": 3.3378993994137183, + "tokens_seen": 3137011712 + }, + { + "epoch": 0.95, + "learning_rate": 4.0136409227683055e-05, + "loss": 1.17, + "theoretical_loss": 3.33789413177236, + "tokens_seen": 3137077248 + }, + { + "epoch": 0.95, + "learning_rate": 4.0120361083249755e-05, + "loss": 1.2559, + "theoretical_loss": 3.3378888642718576, + "tokens_seen": 3137142784 + }, + { + "epoch": 0.95, + "learning_rate": 4.010431293881645e-05, + "loss": 1.3856, + "theoretical_loss": 3.3378835969122047, + "tokens_seen": 3137208320 + }, + { + "epoch": 0.95, + "learning_rate": 4.008826479438315e-05, + "loss": 1.2283, + "theoretical_loss": 3.337878329693394, + "tokens_seen": 3137273856 + }, + { + "epoch": 0.95, + "learning_rate": 4.007221664994985e-05, + "loss": 1.3759, + "theoretical_loss": 3.337873062615419, + "tokens_seen": 3137339392 + }, + { + "epoch": 0.95, + "learning_rate": 4.005616850551655e-05, + "loss": 1.29, + "theoretical_loss": 3.337867795678273, + "tokens_seen": 3137404928 + }, + { + "epoch": 0.95, + "learning_rate": 4.004012036108325e-05, + "loss": 1.343, + "theoretical_loss": 3.33786252888195, + "tokens_seen": 3137470464 + }, + { + "epoch": 0.95, + "learning_rate": 4.002407221664995e-05, + "loss": 1.4682, + "theoretical_loss": 3.337857262226442, + "tokens_seen": 3137536000 + }, + { + "epoch": 0.95, + "learning_rate": 4.000802407221665e-05, + "loss": 1.2764, + "theoretical_loss": 3.3378519957117434, + "tokens_seen": 3137601536 + }, + { + "epoch": 0.95, + "learning_rate": 3.9991975927783354e-05, + "loss": 1.2791, + "theoretical_loss": 3.337846729337847, + "tokens_seen": 3137667072 + }, + { + "epoch": 0.95, + "learning_rate": 3.9975927783350054e-05, + "loss": 1.3089, + "theoretical_loss": 3.3378414631047457, + "tokens_seen": 3137732608 + }, + { + "epoch": 0.95, + "learning_rate": 3.9959879638916755e-05, + "loss": 1.1695, + "theoretical_loss": 3.3378361970124337, + "tokens_seen": 3137798144 + }, + { + "epoch": 0.95, + "learning_rate": 3.9943831494483455e-05, + "loss": 1.2105, + "theoretical_loss": 3.337830931060904, + "tokens_seen": 3137863680 + }, + { + "epoch": 0.95, + "learning_rate": 3.992778335005015e-05, + "loss": 1.2916, + "theoretical_loss": 3.3378256652501492, + "tokens_seen": 3137929216 + }, + { + "epoch": 0.95, + "learning_rate": 3.991173520561685e-05, + "loss": 1.215, + "theoretical_loss": 3.3378203995801634, + "tokens_seen": 3137994752 + }, + { + "epoch": 0.95, + "learning_rate": 3.989568706118355e-05, + "loss": 1.2645, + "theoretical_loss": 3.3378151340509397, + "tokens_seen": 3138060288 + }, + { + "epoch": 0.95, + "learning_rate": 3.987963891675025e-05, + "loss": 1.3261, + "theoretical_loss": 3.3378098686624713, + "tokens_seen": 3138125824 + }, + { + "epoch": 0.95, + "learning_rate": 3.986359077231696e-05, + "loss": 1.1917, + "theoretical_loss": 3.3378046034147513, + "tokens_seen": 3138191360 + }, + { + "epoch": 0.95, + "learning_rate": 3.984754262788365e-05, + "loss": 1.2062, + "theoretical_loss": 3.3377993383077733, + "tokens_seen": 3138256896 + }, + { + "epoch": 0.95, + "learning_rate": 3.983149448345035e-05, + "loss": 1.2283, + "theoretical_loss": 3.3377940733415308, + "tokens_seen": 3138322432 + }, + { + "epoch": 0.95, + "learning_rate": 3.9815446339017054e-05, + "loss": 1.3541, + "theoretical_loss": 3.3377888085160166, + "tokens_seen": 3138387968 + }, + { + "epoch": 0.95, + "learning_rate": 3.9799398194583754e-05, + "loss": 1.2724, + "theoretical_loss": 3.3377835438312244, + "tokens_seen": 3138453504 + }, + { + "epoch": 0.95, + "learning_rate": 3.9783350050150455e-05, + "loss": 1.3574, + "theoretical_loss": 3.3377782792871473, + "tokens_seen": 3138519040 + }, + { + "epoch": 0.95, + "learning_rate": 3.9767301905717156e-05, + "loss": 1.2133, + "theoretical_loss": 3.337773014883778, + "tokens_seen": 3138584576 + }, + { + "epoch": 0.95, + "learning_rate": 3.975125376128385e-05, + "loss": 1.2803, + "theoretical_loss": 3.3377677506211114, + "tokens_seen": 3138650112 + }, + { + "epoch": 0.95, + "learning_rate": 3.973520561685055e-05, + "loss": 1.2118, + "theoretical_loss": 3.3377624864991393, + "tokens_seen": 3138715648 + }, + { + "epoch": 0.95, + "learning_rate": 3.971915747241725e-05, + "loss": 1.2192, + "theoretical_loss": 3.3377572225178556, + "tokens_seen": 3138781184 + }, + { + "epoch": 0.95, + "learning_rate": 3.970310932798396e-05, + "loss": 1.3196, + "theoretical_loss": 3.3377519586772535, + "tokens_seen": 3138846720 + }, + { + "epoch": 0.95, + "learning_rate": 3.968706118355066e-05, + "loss": 1.2323, + "theoretical_loss": 3.3377466949773265, + "tokens_seen": 3138912256 + }, + { + "epoch": 0.95, + "learning_rate": 3.967101303911735e-05, + "loss": 1.2444, + "theoretical_loss": 3.3377414314180673, + "tokens_seen": 3138977792 + }, + { + "epoch": 0.95, + "learning_rate": 3.965496489468405e-05, + "loss": 1.3364, + "theoretical_loss": 3.33773616799947, + "tokens_seen": 3139043328 + }, + { + "epoch": 0.95, + "learning_rate": 3.9638916750250754e-05, + "loss": 1.1635, + "theoretical_loss": 3.3377309047215276, + "tokens_seen": 3139108864 + }, + { + "epoch": 0.95, + "objective/train/docs_used": 1140387, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.2010536193847656, + "objective/train/theoretical_loss": 3.337725641584233, + "objective/train/tokens_used": 3159634400, + "theoretical_loss": 3.337725641584233, + "tokens_seen": 3139174400 + }, + { + "epoch": 0.95, + "learning_rate": 3.9622868605817455e-05, + "loss": 1.2263, + "theoretical_loss": 3.337725641584233, + "tokens_seen": 3139174400 + }, + { + "epoch": 0.95, + "learning_rate": 3.9606820461384155e-05, + "loss": 1.2171, + "theoretical_loss": 3.33772037858758, + "tokens_seen": 3139239936 + }, + { + "epoch": 0.95, + "learning_rate": 3.9590772316950856e-05, + "loss": 1.2995, + "theoretical_loss": 3.337715115731562, + "tokens_seen": 3139305472 + }, + { + "epoch": 0.95, + "learning_rate": 3.957472417251755e-05, + "loss": 1.2786, + "theoretical_loss": 3.3377098530161717, + "tokens_seen": 3139371008 + }, + { + "epoch": 0.95, + "learning_rate": 3.955867602808426e-05, + "loss": 1.2814, + "theoretical_loss": 3.337704590441403, + "tokens_seen": 3139436544 + }, + { + "epoch": 0.95, + "learning_rate": 3.954262788365096e-05, + "loss": 1.1235, + "theoretical_loss": 3.337699328007249, + "tokens_seen": 3139502080 + }, + { + "epoch": 0.95, + "learning_rate": 3.952657973921766e-05, + "loss": 1.351, + "theoretical_loss": 3.337694065713703, + "tokens_seen": 3139567616 + }, + { + "epoch": 0.95, + "learning_rate": 3.951053159478436e-05, + "loss": 1.285, + "theoretical_loss": 3.3376888035607584, + "tokens_seen": 3139633152 + }, + { + "epoch": 0.95, + "learning_rate": 3.949448345035105e-05, + "loss": 1.1406, + "theoretical_loss": 3.337683541548408, + "tokens_seen": 3139698688 + }, + { + "epoch": 0.95, + "learning_rate": 3.9478435305917754e-05, + "loss": 1.3385, + "theoretical_loss": 3.3376782796766458, + "tokens_seen": 3139764224 + }, + { + "epoch": 0.95, + "learning_rate": 3.9462387161484454e-05, + "loss": 1.3136, + "theoretical_loss": 3.337673017945465, + "tokens_seen": 3139829760 + }, + { + "epoch": 0.95, + "learning_rate": 3.9446339017051155e-05, + "loss": 1.3033, + "theoretical_loss": 3.3376677563548585, + "tokens_seen": 3139895296 + }, + { + "epoch": 0.95, + "learning_rate": 3.9430290872617855e-05, + "loss": 1.1968, + "theoretical_loss": 3.3376624949048197, + "tokens_seen": 3139960832 + }, + { + "epoch": 0.95, + "learning_rate": 3.9414242728184556e-05, + "loss": 1.2345, + "theoretical_loss": 3.3376572335953423, + "tokens_seen": 3140026368 + }, + { + "epoch": 0.95, + "learning_rate": 3.939819458375126e-05, + "loss": 1.241, + "theoretical_loss": 3.337651972426419, + "tokens_seen": 3140091904 + }, + { + "epoch": 0.95, + "learning_rate": 3.938214643931796e-05, + "loss": 1.2188, + "theoretical_loss": 3.337646711398044, + "tokens_seen": 3140157440 + }, + { + "epoch": 0.95, + "learning_rate": 3.936609829488466e-05, + "loss": 1.3248, + "theoretical_loss": 3.33764145051021, + "tokens_seen": 3140222976 + }, + { + "epoch": 0.95, + "learning_rate": 3.935005015045136e-05, + "loss": 1.313, + "theoretical_loss": 3.33763618976291, + "tokens_seen": 3140288512 + }, + { + "epoch": 0.95, + "learning_rate": 3.933400200601805e-05, + "loss": 1.3812, + "theoretical_loss": 3.337630929156138, + "tokens_seen": 3140354048 + }, + { + "epoch": 0.95, + "learning_rate": 3.931795386158475e-05, + "loss": 1.3068, + "theoretical_loss": 3.337625668689887, + "tokens_seen": 3140419584 + }, + { + "epoch": 0.95, + "learning_rate": 3.9301905717151454e-05, + "loss": 1.2552, + "theoretical_loss": 3.3376204083641507, + "tokens_seen": 3140485120 + }, + { + "epoch": 0.95, + "learning_rate": 3.9285857572718154e-05, + "loss": 1.297, + "theoretical_loss": 3.3376151481789216, + "tokens_seen": 3140550656 + }, + { + "epoch": 0.95, + "learning_rate": 3.9269809428284855e-05, + "loss": 1.191, + "theoretical_loss": 3.3376098881341933, + "tokens_seen": 3140616192 + }, + { + "epoch": 0.95, + "learning_rate": 3.9253761283851556e-05, + "loss": 1.2877, + "theoretical_loss": 3.33760462822996, + "tokens_seen": 3140681728 + }, + { + "epoch": 0.95, + "learning_rate": 3.9237713139418256e-05, + "loss": 1.3693, + "theoretical_loss": 3.3375993684662135, + "tokens_seen": 3140747264 + }, + { + "epoch": 0.95, + "learning_rate": 3.922166499498496e-05, + "loss": 1.1655, + "theoretical_loss": 3.3375941088429486, + "tokens_seen": 3140812800 + }, + { + "epoch": 0.95, + "learning_rate": 3.920561685055166e-05, + "loss": 1.1848, + "theoretical_loss": 3.3375888493601575, + "tokens_seen": 3140878336 + }, + { + "epoch": 0.95, + "learning_rate": 3.918956870611836e-05, + "loss": 1.372, + "theoretical_loss": 3.337583590017834, + "tokens_seen": 3140943872 + }, + { + "epoch": 0.95, + "learning_rate": 3.917352056168506e-05, + "loss": 1.3775, + "theoretical_loss": 3.3375783308159717, + "tokens_seen": 3141009408 + }, + { + "epoch": 0.95, + "learning_rate": 3.915747241725175e-05, + "loss": 1.2918, + "theoretical_loss": 3.337573071754563, + "tokens_seen": 3141074944 + }, + { + "epoch": 0.95, + "learning_rate": 3.914142427281845e-05, + "loss": 1.1882, + "theoretical_loss": 3.3375678128336026, + "tokens_seen": 3141140480 + }, + { + "epoch": 0.95, + "learning_rate": 3.9125376128385154e-05, + "loss": 1.3523, + "theoretical_loss": 3.3375625540530827, + "tokens_seen": 3141206016 + }, + { + "epoch": 0.95, + "learning_rate": 3.910932798395186e-05, + "loss": 1.2963, + "theoretical_loss": 3.3375572954129966, + "tokens_seen": 3141271552 + }, + { + "epoch": 0.95, + "learning_rate": 3.909327983951856e-05, + "loss": 1.1566, + "theoretical_loss": 3.3375520369133382, + "tokens_seen": 3141337088 + }, + { + "epoch": 0.95, + "learning_rate": 3.9077231695085256e-05, + "loss": 1.3237, + "theoretical_loss": 3.337546778554101, + "tokens_seen": 3141402624 + }, + { + "epoch": 0.95, + "learning_rate": 3.9061183550651957e-05, + "loss": 1.3337, + "theoretical_loss": 3.337541520335277, + "tokens_seen": 3141468160 + }, + { + "epoch": 0.95, + "learning_rate": 3.904513540621866e-05, + "loss": 1.133, + "theoretical_loss": 3.337536262256861, + "tokens_seen": 3141533696 + }, + { + "epoch": 0.95, + "learning_rate": 3.902908726178536e-05, + "loss": 1.2928, + "theoretical_loss": 3.337531004318846, + "tokens_seen": 3141599232 + }, + { + "epoch": 0.95, + "learning_rate": 3.901303911735206e-05, + "loss": 1.2989, + "theoretical_loss": 3.3375257465212247, + "tokens_seen": 3141664768 + }, + { + "epoch": 0.95, + "learning_rate": 3.899699097291876e-05, + "loss": 1.2191, + "theoretical_loss": 3.3375204888639907, + "tokens_seen": 3141730304 + }, + { + "epoch": 0.95, + "learning_rate": 3.898094282848545e-05, + "loss": 1.1198, + "theoretical_loss": 3.3375152313471377, + "tokens_seen": 3141795840 + }, + { + "epoch": 0.95, + "learning_rate": 3.8964894684052154e-05, + "loss": 1.2907, + "theoretical_loss": 3.3375099739706586, + "tokens_seen": 3141861376 + }, + { + "epoch": 0.95, + "learning_rate": 3.894884653961886e-05, + "loss": 1.2016, + "theoretical_loss": 3.337504716734547, + "tokens_seen": 3141926912 + }, + { + "epoch": 0.95, + "learning_rate": 3.893279839518556e-05, + "loss": 1.2855, + "theoretical_loss": 3.337499459638796, + "tokens_seen": 3141992448 + }, + { + "epoch": 0.95, + "learning_rate": 3.891675025075226e-05, + "loss": 1.2818, + "theoretical_loss": 3.337494202683399, + "tokens_seen": 3142057984 + }, + { + "epoch": 0.95, + "learning_rate": 3.8900702106318956e-05, + "loss": 1.2294, + "theoretical_loss": 3.3374889458683494, + "tokens_seen": 3142123520 + }, + { + "epoch": 0.95, + "learning_rate": 3.888465396188566e-05, + "loss": 1.3564, + "theoretical_loss": 3.3374836891936406, + "tokens_seen": 3142189056 + }, + { + "epoch": 0.95, + "learning_rate": 3.886860581745236e-05, + "loss": 1.3205, + "theoretical_loss": 3.3374784326592657, + "tokens_seen": 3142254592 + }, + { + "epoch": 0.95, + "learning_rate": 3.885255767301906e-05, + "loss": 1.3073, + "theoretical_loss": 3.3374731762652177, + "tokens_seen": 3142320128 + }, + { + "epoch": 0.95, + "learning_rate": 3.883650952858576e-05, + "loss": 1.2953, + "theoretical_loss": 3.337467920011491, + "tokens_seen": 3142385664 + }, + { + "epoch": 0.95, + "objective/train/docs_used": 1141495, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.2814786434173584, + "objective/train/theoretical_loss": 3.337462663898078, + "objective/train/tokens_used": 3162911200, + "theoretical_loss": 3.337462663898078, + "tokens_seen": 3142451200 + }, + { + "epoch": 0.95, + "learning_rate": 3.882046138415246e-05, + "loss": 1.3222, + "theoretical_loss": 3.337462663898078, + "tokens_seen": 3142451200 + }, + { + "epoch": 0.95, + "learning_rate": 3.880441323971916e-05, + "loss": 1.2506, + "theoretical_loss": 3.337457407924972, + "tokens_seen": 3142516736 + }, + { + "epoch": 0.95, + "learning_rate": 3.878836509528586e-05, + "loss": 1.3518, + "theoretical_loss": 3.337452152092167, + "tokens_seen": 3142582272 + }, + { + "epoch": 0.95, + "learning_rate": 3.877231695085256e-05, + "loss": 1.3472, + "theoretical_loss": 3.337446896399656, + "tokens_seen": 3142647808 + }, + { + "epoch": 0.95, + "learning_rate": 3.875626880641926e-05, + "loss": 1.2618, + "theoretical_loss": 3.337441640847432, + "tokens_seen": 3142713344 + }, + { + "epoch": 0.95, + "learning_rate": 3.874022066198596e-05, + "loss": 1.3095, + "theoretical_loss": 3.337436385435489, + "tokens_seen": 3142778880 + }, + { + "epoch": 0.95, + "learning_rate": 3.8724172517552656e-05, + "loss": 1.3582, + "theoretical_loss": 3.3374311301638198, + "tokens_seen": 3142844416 + }, + { + "epoch": 0.95, + "learning_rate": 3.870812437311936e-05, + "loss": 1.3125, + "theoretical_loss": 3.337425875032418, + "tokens_seen": 3142909952 + }, + { + "epoch": 0.95, + "learning_rate": 3.869207622868606e-05, + "loss": 1.1707, + "theoretical_loss": 3.337420620041277, + "tokens_seen": 3142975488 + }, + { + "epoch": 0.95, + "learning_rate": 3.867602808425276e-05, + "loss": 1.1693, + "theoretical_loss": 3.337415365190389, + "tokens_seen": 3143041024 + }, + { + "epoch": 0.95, + "learning_rate": 3.8659979939819466e-05, + "loss": 1.2187, + "theoretical_loss": 3.337410110479749, + "tokens_seen": 3143106560 + }, + { + "epoch": 0.95, + "learning_rate": 3.864393179538616e-05, + "loss": 1.2808, + "theoretical_loss": 3.33740485590935, + "tokens_seen": 3143172096 + }, + { + "epoch": 0.95, + "learning_rate": 3.862788365095286e-05, + "loss": 1.2479, + "theoretical_loss": 3.3373996014791842, + "tokens_seen": 3143237632 + }, + { + "epoch": 0.95, + "learning_rate": 3.861183550651956e-05, + "loss": 1.2998, + "theoretical_loss": 3.337394347189246, + "tokens_seen": 3143303168 + }, + { + "epoch": 0.95, + "learning_rate": 3.859578736208626e-05, + "loss": 1.2963, + "theoretical_loss": 3.3373890930395285, + "tokens_seen": 3143368704 + }, + { + "epoch": 0.95, + "learning_rate": 3.857973921765296e-05, + "loss": 1.1862, + "theoretical_loss": 3.3373838390300246, + "tokens_seen": 3143434240 + }, + { + "epoch": 0.95, + "learning_rate": 3.856369107321966e-05, + "loss": 1.2981, + "theoretical_loss": 3.337378585160728, + "tokens_seen": 3143499776 + }, + { + "epoch": 0.95, + "learning_rate": 3.8547642928786356e-05, + "loss": 1.2126, + "theoretical_loss": 3.3373733314316323, + "tokens_seen": 3143565312 + }, + { + "epoch": 0.95, + "learning_rate": 3.853159478435306e-05, + "loss": 1.3041, + "theoretical_loss": 3.3373680778427306, + "tokens_seen": 3143630848 + }, + { + "epoch": 0.95, + "learning_rate": 3.851554663991976e-05, + "loss": 1.3354, + "theoretical_loss": 3.3373628243940163, + "tokens_seen": 3143696384 + }, + { + "epoch": 0.95, + "learning_rate": 3.8499498495486465e-05, + "loss": 1.2245, + "theoretical_loss": 3.337357571085482, + "tokens_seen": 3143761920 + }, + { + "epoch": 0.95, + "learning_rate": 3.8483450351053166e-05, + "loss": 1.3158, + "theoretical_loss": 3.3373523179171225, + "tokens_seen": 3143827456 + }, + { + "epoch": 0.95, + "learning_rate": 3.846740220661986e-05, + "loss": 1.3109, + "theoretical_loss": 3.3373470648889296, + "tokens_seen": 3143892992 + }, + { + "epoch": 0.95, + "learning_rate": 3.845135406218656e-05, + "loss": 1.2734, + "theoretical_loss": 3.3373418120008975, + "tokens_seen": 3143958528 + }, + { + "epoch": 0.95, + "learning_rate": 3.843530591775326e-05, + "loss": 1.3053, + "theoretical_loss": 3.3373365592530195, + "tokens_seen": 3144024064 + }, + { + "epoch": 0.95, + "learning_rate": 3.841925777331996e-05, + "loss": 1.239, + "theoretical_loss": 3.337331306645289, + "tokens_seen": 3144089600 + }, + { + "epoch": 0.95, + "learning_rate": 3.840320962888666e-05, + "loss": 1.3037, + "theoretical_loss": 3.3373260541776992, + "tokens_seen": 3144155136 + }, + { + "epoch": 0.95, + "learning_rate": 3.838716148445336e-05, + "loss": 1.3167, + "theoretical_loss": 3.337320801850243, + "tokens_seen": 3144220672 + }, + { + "epoch": 0.95, + "learning_rate": 3.837111334002006e-05, + "loss": 1.2784, + "theoretical_loss": 3.3373155496629145, + "tokens_seen": 3144286208 + }, + { + "epoch": 0.95, + "learning_rate": 3.8355065195586764e-05, + "loss": 1.3205, + "theoretical_loss": 3.3373102976157063, + "tokens_seen": 3144351744 + }, + { + "epoch": 0.95, + "learning_rate": 3.8339017051153465e-05, + "loss": 1.1554, + "theoretical_loss": 3.3373050457086126, + "tokens_seen": 3144417280 + }, + { + "epoch": 0.95, + "learning_rate": 3.8322968906720165e-05, + "loss": 1.3157, + "theoretical_loss": 3.337299793941626, + "tokens_seen": 3144482816 + }, + { + "epoch": 0.95, + "learning_rate": 3.8306920762286866e-05, + "loss": 1.2108, + "theoretical_loss": 3.3372945423147398, + "tokens_seen": 3144548352 + }, + { + "epoch": 0.95, + "learning_rate": 3.829087261785356e-05, + "loss": 1.2099, + "theoretical_loss": 3.337289290827948, + "tokens_seen": 3144613888 + }, + { + "epoch": 0.95, + "learning_rate": 3.827482447342026e-05, + "loss": 1.3118, + "theoretical_loss": 3.3372840394812435, + "tokens_seen": 3144679424 + }, + { + "epoch": 0.95, + "learning_rate": 3.825877632898696e-05, + "loss": 1.3007, + "theoretical_loss": 3.33727878827462, + "tokens_seen": 3144744960 + }, + { + "epoch": 0.95, + "learning_rate": 3.824272818455366e-05, + "loss": 1.3057, + "theoretical_loss": 3.3372735372080706, + "tokens_seen": 3144810496 + }, + { + "epoch": 0.95, + "learning_rate": 3.822668004012036e-05, + "loss": 1.2408, + "theoretical_loss": 3.337268286281588, + "tokens_seen": 3144876032 + }, + { + "epoch": 0.95, + "learning_rate": 3.821063189568706e-05, + "loss": 1.1717, + "theoretical_loss": 3.337263035495167, + "tokens_seen": 3144941568 + }, + { + "epoch": 0.95, + "learning_rate": 3.8194583751253764e-05, + "loss": 1.3294, + "theoretical_loss": 3.3372577848487994, + "tokens_seen": 3145007104 + }, + { + "epoch": 0.95, + "learning_rate": 3.8178535606820464e-05, + "loss": 1.207, + "theoretical_loss": 3.3372525343424795, + "tokens_seen": 3145072640 + }, + { + "epoch": 0.95, + "learning_rate": 3.8162487462387165e-05, + "loss": 1.2969, + "theoretical_loss": 3.3372472839762004, + "tokens_seen": 3145138176 + }, + { + "epoch": 0.95, + "learning_rate": 3.8146439317953866e-05, + "loss": 1.1986, + "theoretical_loss": 3.3372420337499555, + "tokens_seen": 3145203712 + }, + { + "epoch": 0.95, + "learning_rate": 3.8130391173520566e-05, + "loss": 1.2388, + "theoretical_loss": 3.337236783663738, + "tokens_seen": 3145269248 + }, + { + "epoch": 0.95, + "learning_rate": 3.811434302908726e-05, + "loss": 1.323, + "theoretical_loss": 3.3372315337175418, + "tokens_seen": 3145334784 + }, + { + "epoch": 0.95, + "learning_rate": 3.809829488465396e-05, + "loss": 1.3596, + "theoretical_loss": 3.3372262839113596, + "tokens_seen": 3145400320 + }, + { + "epoch": 0.95, + "learning_rate": 3.808224674022066e-05, + "loss": 1.2121, + "theoretical_loss": 3.337221034245185, + "tokens_seen": 3145465856 + }, + { + "epoch": 0.95, + "learning_rate": 3.806619859578736e-05, + "loss": 1.1674, + "theoretical_loss": 3.3372157847190107, + "tokens_seen": 3145531392 + }, + { + "epoch": 0.95, + "learning_rate": 3.805015045135407e-05, + "loss": 1.2535, + "theoretical_loss": 3.337210535332831, + "tokens_seen": 3145596928 + }, + { + "epoch": 0.95, + "learning_rate": 3.803410230692076e-05, + "loss": 1.1827, + "theoretical_loss": 3.3372052860866392, + "tokens_seen": 3145662464 + }, + { + "epoch": 0.95, + "objective/train/docs_used": 1142692, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.2966246604919434, + "objective/train/theoretical_loss": 3.3372000369804278, + "objective/train/tokens_used": 3166188000, + "theoretical_loss": 3.3372000369804278, + "tokens_seen": 3145728000 + }, + { + "epoch": 0.95, + "learning_rate": 3.8018054162487464e-05, + "loss": 1.2361, + "theoretical_loss": 3.3372000369804278, + "tokens_seen": 3145728000 + }, + { + "epoch": 0.95, + "learning_rate": 3.8002006018054165e-05, + "loss": 1.2251, + "theoretical_loss": 3.337194788014191, + "tokens_seen": 3145793536 + }, + { + "epoch": 0.95, + "learning_rate": 3.7985957873620865e-05, + "loss": 1.209, + "theoretical_loss": 3.3371895391879223, + "tokens_seen": 3145859072 + }, + { + "epoch": 0.95, + "learning_rate": 3.7969909729187566e-05, + "loss": 1.2448, + "theoretical_loss": 3.337184290501614, + "tokens_seen": 3145924608 + }, + { + "epoch": 0.95, + "learning_rate": 3.7953861584754266e-05, + "loss": 1.3435, + "theoretical_loss": 3.3371790419552605, + "tokens_seen": 3145990144 + }, + { + "epoch": 0.95, + "learning_rate": 3.793781344032096e-05, + "loss": 1.279, + "theoretical_loss": 3.337173793548854, + "tokens_seen": 3146055680 + }, + { + "epoch": 0.95, + "learning_rate": 3.792176529588766e-05, + "loss": 1.2664, + "theoretical_loss": 3.337168545282389, + "tokens_seen": 3146121216 + }, + { + "epoch": 0.95, + "learning_rate": 3.790571715145437e-05, + "loss": 1.161, + "theoretical_loss": 3.3371632971558585, + "tokens_seen": 3146186752 + }, + { + "epoch": 0.95, + "learning_rate": 3.788966900702107e-05, + "loss": 1.3619, + "theoretical_loss": 3.337158049169256, + "tokens_seen": 3146252288 + }, + { + "epoch": 0.95, + "learning_rate": 3.787362086258777e-05, + "loss": 1.2841, + "theoretical_loss": 3.337152801322574, + "tokens_seen": 3146317824 + }, + { + "epoch": 0.95, + "learning_rate": 3.7857572718154463e-05, + "loss": 1.3084, + "theoretical_loss": 3.337147553615807, + "tokens_seen": 3146383360 + }, + { + "epoch": 0.95, + "learning_rate": 3.7841524573721164e-05, + "loss": 1.3092, + "theoretical_loss": 3.3371423060489476, + "tokens_seen": 3146448896 + }, + { + "epoch": 0.95, + "learning_rate": 3.7825476429287865e-05, + "loss": 1.223, + "theoretical_loss": 3.3371370586219893, + "tokens_seen": 3146514432 + }, + { + "epoch": 0.95, + "learning_rate": 3.7809428284854565e-05, + "loss": 1.2023, + "theoretical_loss": 3.3371318113349258, + "tokens_seen": 3146579968 + }, + { + "epoch": 0.95, + "learning_rate": 3.7793380140421266e-05, + "loss": 1.1859, + "theoretical_loss": 3.33712656418775, + "tokens_seen": 3146645504 + }, + { + "epoch": 0.95, + "learning_rate": 3.777733199598797e-05, + "loss": 1.2225, + "theoretical_loss": 3.3371213171804555, + "tokens_seen": 3146711040 + }, + { + "epoch": 0.95, + "learning_rate": 3.776128385155467e-05, + "loss": 1.2429, + "theoretical_loss": 3.3371160703130354, + "tokens_seen": 3146776576 + }, + { + "epoch": 0.95, + "learning_rate": 3.774523570712137e-05, + "loss": 1.2936, + "theoretical_loss": 3.3371108235854834, + "tokens_seen": 3146842112 + }, + { + "epoch": 0.95, + "learning_rate": 3.772918756268807e-05, + "loss": 1.2618, + "theoretical_loss": 3.337105576997793, + "tokens_seen": 3146907648 + }, + { + "epoch": 0.95, + "learning_rate": 3.771313941825477e-05, + "loss": 1.2937, + "theoretical_loss": 3.337100330549957, + "tokens_seen": 3146973184 + }, + { + "epoch": 0.95, + "learning_rate": 3.769709127382147e-05, + "loss": 1.3069, + "theoretical_loss": 3.3370950842419695, + "tokens_seen": 3147038720 + }, + { + "epoch": 0.95, + "learning_rate": 3.7681043129388164e-05, + "loss": 1.2355, + "theoretical_loss": 3.337089838073823, + "tokens_seen": 3147104256 + }, + { + "epoch": 0.95, + "learning_rate": 3.7664994984954864e-05, + "loss": 1.3914, + "theoretical_loss": 3.3370845920455112, + "tokens_seen": 3147169792 + }, + { + "epoch": 0.95, + "learning_rate": 3.7648946840521565e-05, + "loss": 1.2056, + "theoretical_loss": 3.337079346157028, + "tokens_seen": 3147235328 + }, + { + "epoch": 0.95, + "learning_rate": 3.7632898696088266e-05, + "loss": 1.2643, + "theoretical_loss": 3.337074100408366, + "tokens_seen": 3147300864 + }, + { + "epoch": 0.95, + "learning_rate": 3.7616850551654966e-05, + "loss": 1.2532, + "theoretical_loss": 3.3370688547995186, + "tokens_seen": 3147366400 + }, + { + "epoch": 0.95, + "learning_rate": 3.760080240722167e-05, + "loss": 1.2458, + "theoretical_loss": 3.33706360933048, + "tokens_seen": 3147431936 + }, + { + "epoch": 0.95, + "learning_rate": 3.758475426278837e-05, + "loss": 1.2655, + "theoretical_loss": 3.337058364001243, + "tokens_seen": 3147497472 + }, + { + "epoch": 0.95, + "learning_rate": 3.756870611835507e-05, + "loss": 1.262, + "theoretical_loss": 3.3370531188118004, + "tokens_seen": 3147563008 + }, + { + "epoch": 0.95, + "learning_rate": 3.755265797392177e-05, + "loss": 1.2731, + "theoretical_loss": 3.3370478737621463, + "tokens_seen": 3147628544 + }, + { + "epoch": 0.95, + "learning_rate": 3.753660982948847e-05, + "loss": 1.2138, + "theoretical_loss": 3.3370426288522737, + "tokens_seen": 3147694080 + }, + { + "epoch": 0.95, + "learning_rate": 3.752056168505517e-05, + "loss": 1.2216, + "theoretical_loss": 3.3370373840821763, + "tokens_seen": 3147759616 + }, + { + "epoch": 0.95, + "learning_rate": 3.7504513540621864e-05, + "loss": 1.3144, + "theoretical_loss": 3.3370321394518476, + "tokens_seen": 3147825152 + }, + { + "epoch": 0.95, + "learning_rate": 3.7488465396188564e-05, + "loss": 1.2211, + "theoretical_loss": 3.3370268949612805, + "tokens_seen": 3147890688 + }, + { + "epoch": 0.95, + "learning_rate": 3.7472417251755265e-05, + "loss": 1.2999, + "theoretical_loss": 3.3370216506104686, + "tokens_seen": 3147956224 + }, + { + "epoch": 0.95, + "learning_rate": 3.745636910732197e-05, + "loss": 1.3181, + "theoretical_loss": 3.337016406399405, + "tokens_seen": 3148021760 + }, + { + "epoch": 0.95, + "learning_rate": 3.7440320962888666e-05, + "loss": 1.1872, + "theoretical_loss": 3.337011162328083, + "tokens_seen": 3148087296 + }, + { + "epoch": 0.95, + "learning_rate": 3.742427281845537e-05, + "loss": 1.2626, + "theoretical_loss": 3.337005918396497, + "tokens_seen": 3148152832 + }, + { + "epoch": 0.95, + "learning_rate": 3.740822467402207e-05, + "loss": 1.1952, + "theoretical_loss": 3.337000674604639, + "tokens_seen": 3148218368 + }, + { + "epoch": 0.95, + "learning_rate": 3.739217652958877e-05, + "loss": 1.2715, + "theoretical_loss": 3.336995430952503, + "tokens_seen": 3148283904 + }, + { + "epoch": 0.95, + "learning_rate": 3.737612838515547e-05, + "loss": 1.256, + "theoretical_loss": 3.3369901874400827, + "tokens_seen": 3148349440 + }, + { + "epoch": 0.95, + "learning_rate": 3.736008024072217e-05, + "loss": 1.3493, + "theoretical_loss": 3.336984944067371, + "tokens_seen": 3148414976 + }, + { + "epoch": 0.95, + "learning_rate": 3.7344032096288863e-05, + "loss": 1.2844, + "theoretical_loss": 3.336979700834361, + "tokens_seen": 3148480512 + }, + { + "epoch": 0.95, + "learning_rate": 3.7327983951855564e-05, + "loss": 1.3293, + "theoretical_loss": 3.336974457741047, + "tokens_seen": 3148546048 + }, + { + "epoch": 0.95, + "learning_rate": 3.731193580742227e-05, + "loss": 1.2945, + "theoretical_loss": 3.3369692147874215, + "tokens_seen": 3148611584 + }, + { + "epoch": 0.95, + "learning_rate": 3.729588766298897e-05, + "loss": 1.3121, + "theoretical_loss": 3.336963971973478, + "tokens_seen": 3148677120 + }, + { + "epoch": 0.95, + "learning_rate": 3.727983951855567e-05, + "loss": 1.2195, + "theoretical_loss": 3.3369587292992104, + "tokens_seen": 3148742656 + }, + { + "epoch": 0.95, + "learning_rate": 3.7263791374122367e-05, + "loss": 1.3305, + "theoretical_loss": 3.336953486764612, + "tokens_seen": 3148808192 + }, + { + "epoch": 0.95, + "learning_rate": 3.724774322968907e-05, + "loss": 1.2068, + "theoretical_loss": 3.336948244369675, + "tokens_seen": 3148873728 + }, + { + "epoch": 0.95, + "learning_rate": 3.723169508525577e-05, + "loss": 1.1847, + "theoretical_loss": 3.3369430021143947, + "tokens_seen": 3148939264 + }, + { + "epoch": 0.95, + "objective/train/docs_used": 1143957, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.1387308835983276, + "objective/train/theoretical_loss": 3.336937759998763, + "objective/train/tokens_used": 3169464800, + "theoretical_loss": 3.336937759998763, + "tokens_seen": 3149004800 + }, + { + "epoch": 0.95, + "learning_rate": 3.721564694082247e-05, + "loss": 1.16, + "theoretical_loss": 3.336937759998763, + "tokens_seen": 3149004800 + }, + { + "epoch": 0.95, + "learning_rate": 3.719959879638917e-05, + "loss": 1.2231, + "theoretical_loss": 3.3369325180227736, + "tokens_seen": 3149070336 + }, + { + "epoch": 0.95, + "learning_rate": 3.718355065195587e-05, + "loss": 1.174, + "theoretical_loss": 3.33692727618642, + "tokens_seen": 3149135872 + }, + { + "epoch": 0.95, + "learning_rate": 3.7167502507522564e-05, + "loss": 1.2362, + "theoretical_loss": 3.336922034489696, + "tokens_seen": 3149201408 + }, + { + "epoch": 0.95, + "learning_rate": 3.715145436308927e-05, + "loss": 1.3024, + "theoretical_loss": 3.3369167929325942, + "tokens_seen": 3149266944 + }, + { + "epoch": 0.95, + "learning_rate": 3.713540621865597e-05, + "loss": 1.2496, + "theoretical_loss": 3.3369115515151084, + "tokens_seen": 3149332480 + }, + { + "epoch": 0.95, + "learning_rate": 3.711935807422267e-05, + "loss": 1.3293, + "theoretical_loss": 3.336906310237232, + "tokens_seen": 3149398016 + }, + { + "epoch": 0.95, + "learning_rate": 3.710330992978937e-05, + "loss": 1.1048, + "theoretical_loss": 3.3369010690989582, + "tokens_seen": 3149463552 + }, + { + "epoch": 0.95, + "learning_rate": 3.708726178535607e-05, + "loss": 1.28, + "theoretical_loss": 3.33689582810028, + "tokens_seen": 3149529088 + }, + { + "epoch": 0.95, + "learning_rate": 3.707121364092277e-05, + "loss": 1.2824, + "theoretical_loss": 3.336890587241192, + "tokens_seen": 3149594624 + }, + { + "epoch": 0.95, + "learning_rate": 3.705516549648947e-05, + "loss": 1.3085, + "theoretical_loss": 3.3368853465216866, + "tokens_seen": 3149660160 + }, + { + "epoch": 0.95, + "learning_rate": 3.703911735205617e-05, + "loss": 1.1808, + "theoretical_loss": 3.336880105941757, + "tokens_seen": 3149725696 + }, + { + "epoch": 0.95, + "learning_rate": 3.702306920762287e-05, + "loss": 1.2795, + "theoretical_loss": 3.3368748655013976, + "tokens_seen": 3149791232 + }, + { + "epoch": 0.95, + "learning_rate": 3.700702106318957e-05, + "loss": 1.254, + "theoretical_loss": 3.3368696252006003, + "tokens_seen": 3149856768 + }, + { + "epoch": 0.95, + "learning_rate": 3.699097291875627e-05, + "loss": 1.2521, + "theoretical_loss": 3.33686438503936, + "tokens_seen": 3149922304 + }, + { + "epoch": 0.95, + "learning_rate": 3.697492477432297e-05, + "loss": 1.3052, + "theoretical_loss": 3.336859145017669, + "tokens_seen": 3149987840 + }, + { + "epoch": 0.95, + "learning_rate": 3.695887662988967e-05, + "loss": 1.2738, + "theoretical_loss": 3.3368539051355213, + "tokens_seen": 3150053376 + }, + { + "epoch": 0.95, + "learning_rate": 3.694282848545637e-05, + "loss": 1.2664, + "theoretical_loss": 3.3368486653929104, + "tokens_seen": 3150118912 + }, + { + "epoch": 0.95, + "learning_rate": 3.692678034102307e-05, + "loss": 1.2441, + "theoretical_loss": 3.3368434257898287, + "tokens_seen": 3150184448 + }, + { + "epoch": 0.95, + "learning_rate": 3.691073219658977e-05, + "loss": 1.3006, + "theoretical_loss": 3.3368381863262706, + "tokens_seen": 3150249984 + }, + { + "epoch": 0.95, + "learning_rate": 3.689468405215647e-05, + "loss": 1.2328, + "theoretical_loss": 3.336832947002229, + "tokens_seen": 3150315520 + }, + { + "epoch": 0.95, + "learning_rate": 3.687863590772317e-05, + "loss": 1.2822, + "theoretical_loss": 3.3368277078176973, + "tokens_seen": 3150381056 + }, + { + "epoch": 0.95, + "learning_rate": 3.686258776328987e-05, + "loss": 1.2907, + "theoretical_loss": 3.3368224687726693, + "tokens_seen": 3150446592 + }, + { + "epoch": 0.95, + "learning_rate": 3.6846539618856576e-05, + "loss": 1.2179, + "theoretical_loss": 3.336817229867138, + "tokens_seen": 3150512128 + }, + { + "epoch": 0.95, + "learning_rate": 3.683049147442327e-05, + "loss": 1.2047, + "theoretical_loss": 3.3368119911010967, + "tokens_seen": 3150577664 + }, + { + "epoch": 0.95, + "learning_rate": 3.681444332998997e-05, + "loss": 1.2216, + "theoretical_loss": 3.336806752474539, + "tokens_seen": 3150643200 + }, + { + "epoch": 0.95, + "learning_rate": 3.679839518555667e-05, + "loss": 1.3329, + "theoretical_loss": 3.3368015139874583, + "tokens_seen": 3150708736 + }, + { + "epoch": 0.95, + "learning_rate": 3.678234704112337e-05, + "loss": 1.3078, + "theoretical_loss": 3.3367962756398475, + "tokens_seen": 3150774272 + }, + { + "epoch": 0.95, + "learning_rate": 3.676629889669007e-05, + "loss": 1.3341, + "theoretical_loss": 3.3367910374317007, + "tokens_seen": 3150839808 + }, + { + "epoch": 0.95, + "learning_rate": 3.675025075225677e-05, + "loss": 1.1974, + "theoretical_loss": 3.3367857993630112, + "tokens_seen": 3150905344 + }, + { + "epoch": 0.95, + "learning_rate": 3.673420260782347e-05, + "loss": 1.2817, + "theoretical_loss": 3.3367805614337716, + "tokens_seen": 3150970880 + }, + { + "epoch": 0.95, + "learning_rate": 3.671815446339017e-05, + "loss": 1.2165, + "theoretical_loss": 3.3367753236439763, + "tokens_seen": 3151036416 + }, + { + "epoch": 0.95, + "learning_rate": 3.6702106318956875e-05, + "loss": 1.2376, + "theoretical_loss": 3.336770085993618, + "tokens_seen": 3151101952 + }, + { + "epoch": 0.95, + "learning_rate": 3.6686058174523576e-05, + "loss": 1.242, + "theoretical_loss": 3.3367648484826904, + "tokens_seen": 3151167488 + }, + { + "epoch": 0.95, + "learning_rate": 3.6670010030090277e-05, + "loss": 1.2087, + "theoretical_loss": 3.336759611111187, + "tokens_seen": 3151233024 + }, + { + "epoch": 0.95, + "learning_rate": 3.665396188565697e-05, + "loss": 1.2463, + "theoretical_loss": 3.3367543738791006, + "tokens_seen": 3151298560 + }, + { + "epoch": 0.95, + "learning_rate": 3.663791374122367e-05, + "loss": 1.2893, + "theoretical_loss": 3.336749136786425, + "tokens_seen": 3151364096 + }, + { + "epoch": 0.95, + "learning_rate": 3.662186559679037e-05, + "loss": 1.2491, + "theoretical_loss": 3.336743899833154, + "tokens_seen": 3151429632 + }, + { + "epoch": 0.95, + "learning_rate": 3.660581745235707e-05, + "loss": 1.3682, + "theoretical_loss": 3.3367386630192803, + "tokens_seen": 3151495168 + }, + { + "epoch": 0.96, + "learning_rate": 3.658976930792377e-05, + "loss": 1.2601, + "theoretical_loss": 3.336733426344798, + "tokens_seen": 3151560704 + }, + { + "epoch": 0.96, + "learning_rate": 3.6573721163490474e-05, + "loss": 1.2018, + "theoretical_loss": 3.3367281898096994, + "tokens_seen": 3151626240 + }, + { + "epoch": 0.96, + "learning_rate": 3.6557673019057174e-05, + "loss": 1.3361, + "theoretical_loss": 3.336722953413979, + "tokens_seen": 3151691776 + }, + { + "epoch": 0.96, + "learning_rate": 3.6541624874623875e-05, + "loss": 1.1951, + "theoretical_loss": 3.3367177171576294, + "tokens_seen": 3151757312 + }, + { + "epoch": 0.96, + "learning_rate": 3.6525576730190575e-05, + "loss": 1.2959, + "theoretical_loss": 3.3367124810406446, + "tokens_seen": 3151822848 + }, + { + "epoch": 0.96, + "learning_rate": 3.6509528585757276e-05, + "loss": 1.1893, + "theoretical_loss": 3.3367072450630175, + "tokens_seen": 3151888384 + }, + { + "epoch": 0.96, + "learning_rate": 3.649348044132398e-05, + "loss": 1.2368, + "theoretical_loss": 3.3367020092247417, + "tokens_seen": 3151953920 + }, + { + "epoch": 0.96, + "learning_rate": 3.647743229689067e-05, + "loss": 1.2085, + "theoretical_loss": 3.336696773525811, + "tokens_seen": 3152019456 + }, + { + "epoch": 0.96, + "learning_rate": 3.646138415245737e-05, + "loss": 1.2659, + "theoretical_loss": 3.336691537966218, + "tokens_seen": 3152084992 + }, + { + "epoch": 0.96, + "learning_rate": 3.644533600802407e-05, + "loss": 1.3987, + "theoretical_loss": 3.3366863025459566, + "tokens_seen": 3152150528 + }, + { + "epoch": 0.96, + "learning_rate": 3.642928786359077e-05, + "loss": 1.165, + "theoretical_loss": 3.3366810672650202, + "tokens_seen": 3152216064 + }, + { + "epoch": 0.96, + "objective/train/docs_used": 1145159, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.283656358718872, + "objective/train/theoretical_loss": 3.336675832123402, + "objective/train/tokens_used": 3172741600, + "theoretical_loss": 3.336675832123402, + "tokens_seen": 3152281600 + }, + { + "epoch": 0.96, + "learning_rate": 3.641323971915747e-05, + "loss": 1.2499, + "theoretical_loss": 3.336675832123402, + "tokens_seen": 3152281600 + }, + { + "epoch": 0.96, + "learning_rate": 3.6397191574724174e-05, + "loss": 1.2568, + "theoretical_loss": 3.3366705971210955, + "tokens_seen": 3152347136 + }, + { + "epoch": 0.96, + "learning_rate": 3.6381143430290874e-05, + "loss": 1.169, + "theoretical_loss": 3.3366653622580937, + "tokens_seen": 3152412672 + }, + { + "epoch": 0.96, + "learning_rate": 3.6365095285857575e-05, + "loss": 1.2516, + "theoretical_loss": 3.3366601275343912, + "tokens_seen": 3152478208 + }, + { + "epoch": 0.96, + "learning_rate": 3.6349047141424276e-05, + "loss": 1.1709, + "theoretical_loss": 3.33665489294998, + "tokens_seen": 3152543744 + }, + { + "epoch": 0.96, + "learning_rate": 3.6332998996990976e-05, + "loss": 1.1861, + "theoretical_loss": 3.336649658504854, + "tokens_seen": 3152609280 + }, + { + "epoch": 0.96, + "learning_rate": 3.631695085255768e-05, + "loss": 1.3212, + "theoretical_loss": 3.336644424199007, + "tokens_seen": 3152674816 + }, + { + "epoch": 0.96, + "learning_rate": 3.630090270812437e-05, + "loss": 1.2998, + "theoretical_loss": 3.3366391900324324, + "tokens_seen": 3152740352 + }, + { + "epoch": 0.96, + "learning_rate": 3.628485456369107e-05, + "loss": 1.3291, + "theoretical_loss": 3.3366339560051226, + "tokens_seen": 3152805888 + }, + { + "epoch": 0.96, + "learning_rate": 3.626880641925777e-05, + "loss": 1.1376, + "theoretical_loss": 3.336628722117072, + "tokens_seen": 3152871424 + }, + { + "epoch": 0.96, + "learning_rate": 3.625275827482448e-05, + "loss": 1.1544, + "theoretical_loss": 3.3366234883682733, + "tokens_seen": 3152936960 + }, + { + "epoch": 0.96, + "learning_rate": 3.623671013039118e-05, + "loss": 1.2685, + "theoretical_loss": 3.336618254758721, + "tokens_seen": 3153002496 + }, + { + "epoch": 0.96, + "learning_rate": 3.6220661985957874e-05, + "loss": 1.3045, + "theoretical_loss": 3.3366130212884073, + "tokens_seen": 3153068032 + }, + { + "epoch": 0.96, + "learning_rate": 3.6204613841524575e-05, + "loss": 1.2543, + "theoretical_loss": 3.336607787957326, + "tokens_seen": 3153133568 + }, + { + "epoch": 0.96, + "learning_rate": 3.6188565697091275e-05, + "loss": 1.2564, + "theoretical_loss": 3.3366025547654705, + "tokens_seen": 3153199104 + }, + { + "epoch": 0.96, + "learning_rate": 3.6172517552657976e-05, + "loss": 1.3487, + "theoretical_loss": 3.3365973217128344, + "tokens_seen": 3153264640 + }, + { + "epoch": 0.96, + "learning_rate": 3.6156469408224677e-05, + "loss": 1.2698, + "theoretical_loss": 3.3365920887994114, + "tokens_seen": 3153330176 + }, + { + "epoch": 0.96, + "learning_rate": 3.614042126379138e-05, + "loss": 1.2922, + "theoretical_loss": 3.336586856025194, + "tokens_seen": 3153395712 + }, + { + "epoch": 0.96, + "learning_rate": 3.612437311935807e-05, + "loss": 1.1502, + "theoretical_loss": 3.336581623390176, + "tokens_seen": 3153461248 + }, + { + "epoch": 0.96, + "learning_rate": 3.610832497492478e-05, + "loss": 1.2352, + "theoretical_loss": 3.3365763908943515, + "tokens_seen": 3153526784 + }, + { + "epoch": 0.96, + "learning_rate": 3.609227683049148e-05, + "loss": 1.2168, + "theoretical_loss": 3.336571158537713, + "tokens_seen": 3153592320 + }, + { + "epoch": 0.96, + "learning_rate": 3.607622868605818e-05, + "loss": 1.1726, + "theoretical_loss": 3.3365659263202536, + "tokens_seen": 3153657856 + }, + { + "epoch": 0.96, + "learning_rate": 3.606018054162488e-05, + "loss": 1.2621, + "theoretical_loss": 3.3365606942419683, + "tokens_seen": 3153723392 + }, + { + "epoch": 0.96, + "learning_rate": 3.6044132397191574e-05, + "loss": 1.1871, + "theoretical_loss": 3.336555462302849, + "tokens_seen": 3153788928 + }, + { + "epoch": 0.96, + "learning_rate": 3.6028084252758275e-05, + "loss": 1.1143, + "theoretical_loss": 3.3365502305028896, + "tokens_seen": 3153854464 + }, + { + "epoch": 0.96, + "learning_rate": 3.6012036108324975e-05, + "loss": 1.2841, + "theoretical_loss": 3.3365449988420837, + "tokens_seen": 3153920000 + }, + { + "epoch": 0.96, + "learning_rate": 3.5995987963891676e-05, + "loss": 1.2538, + "theoretical_loss": 3.3365397673204242, + "tokens_seen": 3153985536 + }, + { + "epoch": 0.96, + "learning_rate": 3.597993981945838e-05, + "loss": 1.2027, + "theoretical_loss": 3.336534535937905, + "tokens_seen": 3154051072 + }, + { + "epoch": 0.96, + "learning_rate": 3.596389167502508e-05, + "loss": 1.1731, + "theoretical_loss": 3.3365293046945195, + "tokens_seen": 3154116608 + }, + { + "epoch": 0.96, + "learning_rate": 3.594784353059178e-05, + "loss": 1.1566, + "theoretical_loss": 3.3365240735902613, + "tokens_seen": 3154182144 + }, + { + "epoch": 0.96, + "learning_rate": 3.593179538615848e-05, + "loss": 1.1711, + "theoretical_loss": 3.336518842625123, + "tokens_seen": 3154247680 + }, + { + "epoch": 0.96, + "learning_rate": 3.591574724172518e-05, + "loss": 1.2078, + "theoretical_loss": 3.3365136117990986, + "tokens_seen": 3154313216 + }, + { + "epoch": 0.96, + "learning_rate": 3.589969909729188e-05, + "loss": 1.2714, + "theoretical_loss": 3.336508381112181, + "tokens_seen": 3154378752 + }, + { + "epoch": 0.96, + "learning_rate": 3.588365095285858e-05, + "loss": 1.22, + "theoretical_loss": 3.3365031505643645, + "tokens_seen": 3154444288 + }, + { + "epoch": 0.96, + "learning_rate": 3.5867602808425274e-05, + "loss": 1.2145, + "theoretical_loss": 3.336497920155642, + "tokens_seen": 3154509824 + }, + { + "epoch": 0.96, + "learning_rate": 3.5851554663991975e-05, + "loss": 1.2446, + "theoretical_loss": 3.336492689886007, + "tokens_seen": 3154575360 + }, + { + "epoch": 0.96, + "learning_rate": 3.5835506519558676e-05, + "loss": 1.2738, + "theoretical_loss": 3.3364874597554524, + "tokens_seen": 3154640896 + }, + { + "epoch": 0.96, + "learning_rate": 3.5819458375125376e-05, + "loss": 1.1879, + "theoretical_loss": 3.3364822297639725, + "tokens_seen": 3154706432 + }, + { + "epoch": 0.96, + "learning_rate": 3.5803410230692084e-05, + "loss": 1.217, + "theoretical_loss": 3.33647699991156, + "tokens_seen": 3154771968 + }, + { + "epoch": 0.96, + "learning_rate": 3.578736208625878e-05, + "loss": 1.3302, + "theoretical_loss": 3.3364717701982083, + "tokens_seen": 3154837504 + }, + { + "epoch": 0.96, + "learning_rate": 3.577131394182548e-05, + "loss": 1.3711, + "theoretical_loss": 3.3364665406239116, + "tokens_seen": 3154903040 + }, + { + "epoch": 0.96, + "learning_rate": 3.575526579739218e-05, + "loss": 1.2109, + "theoretical_loss": 3.3364613111886627, + "tokens_seen": 3154968576 + }, + { + "epoch": 0.96, + "learning_rate": 3.573921765295888e-05, + "loss": 1.3277, + "theoretical_loss": 3.3364560818924547, + "tokens_seen": 3155034112 + }, + { + "epoch": 0.96, + "learning_rate": 3.572316950852558e-05, + "loss": 1.1777, + "theoretical_loss": 3.336450852735282, + "tokens_seen": 3155099648 + }, + { + "epoch": 0.96, + "learning_rate": 3.570712136409228e-05, + "loss": 1.425, + "theoretical_loss": 3.3364456237171374, + "tokens_seen": 3155165184 + }, + { + "epoch": 0.96, + "learning_rate": 3.5691073219658975e-05, + "loss": 1.247, + "theoretical_loss": 3.3364403948380144, + "tokens_seen": 3155230720 + }, + { + "epoch": 0.96, + "learning_rate": 3.5675025075225675e-05, + "loss": 1.2291, + "theoretical_loss": 3.336435166097906, + "tokens_seen": 3155296256 + }, + { + "epoch": 0.96, + "learning_rate": 3.565897693079238e-05, + "loss": 1.3211, + "theoretical_loss": 3.3364299374968063, + "tokens_seen": 3155361792 + }, + { + "epoch": 0.96, + "learning_rate": 3.564292878635908e-05, + "loss": 1.2084, + "theoretical_loss": 3.3364247090347083, + "tokens_seen": 3155427328 + }, + { + "epoch": 0.96, + "learning_rate": 3.5626880641925784e-05, + "loss": 1.2749, + "theoretical_loss": 3.336419480711605, + "tokens_seen": 3155492864 + }, + { + "epoch": 0.96, + "objective/train/docs_used": 1146462, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.28457772731781, + "objective/train/theoretical_loss": 3.336414252527491, + "objective/train/tokens_used": 3176018400, + "theoretical_loss": 3.336414252527491, + "tokens_seen": 3155558400 + }, + { + "epoch": 0.96, + "learning_rate": 3.561083249749248e-05, + "loss": 1.265, + "theoretical_loss": 3.336414252527491, + "tokens_seen": 3155558400 + }, + { + "epoch": 0.96, + "learning_rate": 3.559478435305918e-05, + "loss": 1.2521, + "theoretical_loss": 3.336409024482359, + "tokens_seen": 3155623936 + }, + { + "epoch": 0.96, + "learning_rate": 3.557873620862588e-05, + "loss": 1.2591, + "theoretical_loss": 3.3364037965762026, + "tokens_seen": 3155689472 + }, + { + "epoch": 0.96, + "learning_rate": 3.556268806419258e-05, + "loss": 1.2673, + "theoretical_loss": 3.336398568809015, + "tokens_seen": 3155755008 + }, + { + "epoch": 0.96, + "learning_rate": 3.554663991975928e-05, + "loss": 1.28, + "theoretical_loss": 3.3363933411807896, + "tokens_seen": 3155820544 + }, + { + "epoch": 0.96, + "learning_rate": 3.553059177532598e-05, + "loss": 1.339, + "theoretical_loss": 3.33638811369152, + "tokens_seen": 3155886080 + }, + { + "epoch": 0.96, + "learning_rate": 3.5514543630892675e-05, + "loss": 1.2551, + "theoretical_loss": 3.3363828863411995, + "tokens_seen": 3155951616 + }, + { + "epoch": 0.96, + "learning_rate": 3.549849548645938e-05, + "loss": 1.2819, + "theoretical_loss": 3.336377659129822, + "tokens_seen": 3156017152 + }, + { + "epoch": 0.96, + "learning_rate": 3.548244734202608e-05, + "loss": 1.1823, + "theoretical_loss": 3.33637243205738, + "tokens_seen": 3156082688 + }, + { + "epoch": 0.96, + "learning_rate": 3.5466399197592783e-05, + "loss": 1.0897, + "theoretical_loss": 3.3363672051238678, + "tokens_seen": 3156148224 + }, + { + "epoch": 0.96, + "learning_rate": 3.545035105315948e-05, + "loss": 1.2756, + "theoretical_loss": 3.336361978329278, + "tokens_seen": 3156213760 + }, + { + "epoch": 0.96, + "learning_rate": 3.543430290872618e-05, + "loss": 1.3019, + "theoretical_loss": 3.336356751673605, + "tokens_seen": 3156279296 + }, + { + "epoch": 0.96, + "learning_rate": 3.541825476429288e-05, + "loss": 1.1798, + "theoretical_loss": 3.3363515251568416, + "tokens_seen": 3156344832 + }, + { + "epoch": 0.96, + "learning_rate": 3.540220661985958e-05, + "loss": 1.2258, + "theoretical_loss": 3.336346298778981, + "tokens_seen": 3156410368 + }, + { + "epoch": 0.96, + "learning_rate": 3.538615847542628e-05, + "loss": 1.3224, + "theoretical_loss": 3.3363410725400175, + "tokens_seen": 3156475904 + }, + { + "epoch": 0.96, + "learning_rate": 3.537011033099298e-05, + "loss": 1.242, + "theoretical_loss": 3.3363358464399435, + "tokens_seen": 3156541440 + }, + { + "epoch": 0.96, + "learning_rate": 3.535406218655968e-05, + "loss": 1.2177, + "theoretical_loss": 3.336330620478753, + "tokens_seen": 3156606976 + }, + { + "epoch": 0.96, + "learning_rate": 3.533801404212638e-05, + "loss": 1.2514, + "theoretical_loss": 3.3363253946564395, + "tokens_seen": 3156672512 + }, + { + "epoch": 0.96, + "learning_rate": 3.532196589769308e-05, + "loss": 1.3361, + "theoretical_loss": 3.336320168972996, + "tokens_seen": 3156738048 + }, + { + "epoch": 0.96, + "learning_rate": 3.530591775325978e-05, + "loss": 1.2685, + "theoretical_loss": 3.3363149434284165, + "tokens_seen": 3156803584 + }, + { + "epoch": 0.96, + "learning_rate": 3.5289869608826484e-05, + "loss": 1.2489, + "theoretical_loss": 3.3363097180226937, + "tokens_seen": 3156869120 + }, + { + "epoch": 0.96, + "learning_rate": 3.527382146439318e-05, + "loss": 1.3327, + "theoretical_loss": 3.336304492755822, + "tokens_seen": 3156934656 + }, + { + "epoch": 0.96, + "learning_rate": 3.525777331995988e-05, + "loss": 1.2949, + "theoretical_loss": 3.336299267627794, + "tokens_seen": 3157000192 + }, + { + "epoch": 0.96, + "learning_rate": 3.524172517552658e-05, + "loss": 1.2619, + "theoretical_loss": 3.3362940426386034, + "tokens_seen": 3157065728 + }, + { + "epoch": 0.96, + "learning_rate": 3.522567703109328e-05, + "loss": 1.1955, + "theoretical_loss": 3.336288817788243, + "tokens_seen": 3157131264 + }, + { + "epoch": 0.96, + "learning_rate": 3.520962888665999e-05, + "loss": 1.2748, + "theoretical_loss": 3.3362835930767076, + "tokens_seen": 3157196800 + }, + { + "epoch": 0.96, + "learning_rate": 3.519358074222668e-05, + "loss": 1.2414, + "theoretical_loss": 3.33627836850399, + "tokens_seen": 3157262336 + }, + { + "epoch": 0.96, + "learning_rate": 3.517753259779338e-05, + "loss": 1.3593, + "theoretical_loss": 3.3362731440700832, + "tokens_seen": 3157327872 + }, + { + "epoch": 0.96, + "learning_rate": 3.516148445336008e-05, + "loss": 1.2738, + "theoretical_loss": 3.336267919774981, + "tokens_seen": 3157393408 + }, + { + "epoch": 0.96, + "learning_rate": 3.514543630892678e-05, + "loss": 1.2709, + "theoretical_loss": 3.3362626956186765, + "tokens_seen": 3157458944 + }, + { + "epoch": 0.96, + "learning_rate": 3.512938816449348e-05, + "loss": 1.2303, + "theoretical_loss": 3.3362574716011637, + "tokens_seen": 3157524480 + }, + { + "epoch": 0.96, + "learning_rate": 3.5113340020060184e-05, + "loss": 1.2572, + "theoretical_loss": 3.336252247722436, + "tokens_seen": 3157590016 + }, + { + "epoch": 0.96, + "learning_rate": 3.509729187562688e-05, + "loss": 1.2302, + "theoretical_loss": 3.336247023982486, + "tokens_seen": 3157655552 + }, + { + "epoch": 0.96, + "learning_rate": 3.508124373119358e-05, + "loss": 1.2245, + "theoretical_loss": 3.336241800381308, + "tokens_seen": 3157721088 + }, + { + "epoch": 0.96, + "learning_rate": 3.506519558676028e-05, + "loss": 1.206, + "theoretical_loss": 3.3362365769188953, + "tokens_seen": 3157786624 + }, + { + "epoch": 0.96, + "learning_rate": 3.5049147442326986e-05, + "loss": 1.304, + "theoretical_loss": 3.336231353595241, + "tokens_seen": 3157852160 + }, + { + "epoch": 0.96, + "learning_rate": 3.503309929789369e-05, + "loss": 1.2804, + "theoretical_loss": 3.3362261304103384, + "tokens_seen": 3157917696 + }, + { + "epoch": 0.96, + "learning_rate": 3.501705115346038e-05, + "loss": 1.2059, + "theoretical_loss": 3.3362209073641815, + "tokens_seen": 3157983232 + }, + { + "epoch": 0.96, + "learning_rate": 3.500100300902708e-05, + "loss": 1.0346, + "theoretical_loss": 3.3362156844567634, + "tokens_seen": 3158048768 + }, + { + "epoch": 0.96, + "learning_rate": 3.498495486459378e-05, + "loss": 1.2598, + "theoretical_loss": 3.336210461688078, + "tokens_seen": 3158114304 + }, + { + "epoch": 0.96, + "learning_rate": 3.496890672016048e-05, + "loss": 1.2772, + "theoretical_loss": 3.336205239058118, + "tokens_seen": 3158179840 + }, + { + "epoch": 0.96, + "learning_rate": 3.4952858575727183e-05, + "loss": 1.3959, + "theoretical_loss": 3.3362000165668775, + "tokens_seen": 3158245376 + }, + { + "epoch": 0.96, + "learning_rate": 3.4936810431293884e-05, + "loss": 1.2796, + "theoretical_loss": 3.3361947942143493, + "tokens_seen": 3158310912 + }, + { + "epoch": 0.96, + "learning_rate": 3.492076228686058e-05, + "loss": 1.2258, + "theoretical_loss": 3.336189572000527, + "tokens_seen": 3158376448 + }, + { + "epoch": 0.96, + "learning_rate": 3.4904714142427285e-05, + "loss": 1.2742, + "theoretical_loss": 3.3361843499254045, + "tokens_seen": 3158441984 + }, + { + "epoch": 0.96, + "learning_rate": 3.4888665997993986e-05, + "loss": 1.242, + "theoretical_loss": 3.336179127988975, + "tokens_seen": 3158507520 + }, + { + "epoch": 0.96, + "learning_rate": 3.4872617853560687e-05, + "loss": 1.1957, + "theoretical_loss": 3.336173906191232, + "tokens_seen": 3158573056 + }, + { + "epoch": 0.96, + "learning_rate": 3.485656970912739e-05, + "loss": 1.4, + "theoretical_loss": 3.3361686845321685, + "tokens_seen": 3158638592 + }, + { + "epoch": 0.96, + "learning_rate": 3.484052156469408e-05, + "loss": 1.3321, + "theoretical_loss": 3.3361634630117782, + "tokens_seen": 3158704128 + }, + { + "epoch": 0.96, + "learning_rate": 3.482447342026078e-05, + "loss": 1.2491, + "theoretical_loss": 3.336158241630055, + "tokens_seen": 3158769664 + }, + { + "epoch": 0.96, + "objective/train/docs_used": 1147634, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.1722795963287354, + "objective/train/theoretical_loss": 3.336153020386991, + "objective/train/tokens_used": 3179295200, + "theoretical_loss": 3.336153020386991, + "tokens_seen": 3158835200 + }, + { + "epoch": 0.96, + "learning_rate": 3.480842527582748e-05, + "loss": 1.1999, + "theoretical_loss": 3.336153020386991, + "tokens_seen": 3158835200 + }, + { + "epoch": 0.96, + "learning_rate": 3.479237713139418e-05, + "loss": 1.2143, + "theoretical_loss": 3.3361477992825814, + "tokens_seen": 3158900736 + }, + { + "epoch": 0.96, + "learning_rate": 3.4776328986960884e-05, + "loss": 1.3086, + "theoretical_loss": 3.3361425783168186, + "tokens_seen": 3158966272 + }, + { + "epoch": 0.96, + "learning_rate": 3.4760280842527584e-05, + "loss": 1.2695, + "theoretical_loss": 3.3361373574896964, + "tokens_seen": 3159031808 + }, + { + "epoch": 0.96, + "learning_rate": 3.4744232698094285e-05, + "loss": 1.2457, + "theoretical_loss": 3.3361321368012082, + "tokens_seen": 3159097344 + }, + { + "epoch": 0.96, + "learning_rate": 3.4728184553660986e-05, + "loss": 1.2003, + "theoretical_loss": 3.336126916251347, + "tokens_seen": 3159162880 + }, + { + "epoch": 0.96, + "learning_rate": 3.4712136409227686e-05, + "loss": 1.2767, + "theoretical_loss": 3.3361216958401068, + "tokens_seen": 3159228416 + }, + { + "epoch": 0.96, + "learning_rate": 3.469608826479439e-05, + "loss": 1.2148, + "theoretical_loss": 3.336116475567481, + "tokens_seen": 3159293952 + }, + { + "epoch": 0.96, + "learning_rate": 3.468004012036109e-05, + "loss": 1.204, + "theoretical_loss": 3.3361112554334627, + "tokens_seen": 3159359488 + }, + { + "epoch": 0.96, + "learning_rate": 3.466399197592778e-05, + "loss": 1.346, + "theoretical_loss": 3.3361060354380454, + "tokens_seen": 3159425024 + }, + { + "epoch": 0.96, + "learning_rate": 3.464794383149448e-05, + "loss": 1.3878, + "theoretical_loss": 3.3361008155812226, + "tokens_seen": 3159490560 + }, + { + "epoch": 0.96, + "learning_rate": 3.463189568706118e-05, + "loss": 1.1611, + "theoretical_loss": 3.3360955958629877, + "tokens_seen": 3159556096 + }, + { + "epoch": 0.96, + "learning_rate": 3.461584754262788e-05, + "loss": 1.2826, + "theoretical_loss": 3.3360903762833347, + "tokens_seen": 3159621632 + }, + { + "epoch": 0.96, + "learning_rate": 3.459979939819459e-05, + "loss": 1.3199, + "theoretical_loss": 3.3360851568422563, + "tokens_seen": 3159687168 + }, + { + "epoch": 0.96, + "learning_rate": 3.4583751253761284e-05, + "loss": 1.2411, + "theoretical_loss": 3.3360799375397465, + "tokens_seen": 3159752704 + }, + { + "epoch": 0.96, + "learning_rate": 3.4567703109327985e-05, + "loss": 1.2497, + "theoretical_loss": 3.3360747183757984, + "tokens_seen": 3159818240 + }, + { + "epoch": 0.96, + "learning_rate": 3.4551654964894686e-05, + "loss": 1.2515, + "theoretical_loss": 3.3360694993504056, + "tokens_seen": 3159883776 + }, + { + "epoch": 0.96, + "learning_rate": 3.4535606820461386e-05, + "loss": 1.3504, + "theoretical_loss": 3.336064280463561, + "tokens_seen": 3159949312 + }, + { + "epoch": 0.96, + "learning_rate": 3.451955867602809e-05, + "loss": 1.3159, + "theoretical_loss": 3.3360590617152592, + "tokens_seen": 3160014848 + }, + { + "epoch": 0.96, + "learning_rate": 3.450351053159479e-05, + "loss": 1.3625, + "theoretical_loss": 3.3360538431054927, + "tokens_seen": 3160080384 + }, + { + "epoch": 0.96, + "learning_rate": 3.448746238716148e-05, + "loss": 1.2992, + "theoretical_loss": 3.3360486246342553, + "tokens_seen": 3160145920 + }, + { + "epoch": 0.96, + "learning_rate": 3.447141424272818e-05, + "loss": 1.296, + "theoretical_loss": 3.3360434063015405, + "tokens_seen": 3160211456 + }, + { + "epoch": 0.96, + "learning_rate": 3.445536609829489e-05, + "loss": 1.2217, + "theoretical_loss": 3.3360381881073415, + "tokens_seen": 3160276992 + }, + { + "epoch": 0.96, + "learning_rate": 3.443931795386159e-05, + "loss": 1.1795, + "theoretical_loss": 3.336032970051652, + "tokens_seen": 3160342528 + }, + { + "epoch": 0.96, + "learning_rate": 3.442326980942829e-05, + "loss": 1.2408, + "theoretical_loss": 3.336027752134465, + "tokens_seen": 3160408064 + }, + { + "epoch": 0.96, + "learning_rate": 3.4407221664994985e-05, + "loss": 1.2954, + "theoretical_loss": 3.3360225343557746, + "tokens_seen": 3160473600 + }, + { + "epoch": 0.96, + "learning_rate": 3.4391173520561685e-05, + "loss": 1.2355, + "theoretical_loss": 3.3360173167155738, + "tokens_seen": 3160539136 + }, + { + "epoch": 0.96, + "learning_rate": 3.4375125376128386e-05, + "loss": 1.2298, + "theoretical_loss": 3.3360120992138564, + "tokens_seen": 3160604672 + }, + { + "epoch": 0.96, + "learning_rate": 3.4359077231695087e-05, + "loss": 1.2003, + "theoretical_loss": 3.3360068818506154, + "tokens_seen": 3160670208 + }, + { + "epoch": 0.96, + "learning_rate": 3.434302908726179e-05, + "loss": 1.2537, + "theoretical_loss": 3.3360016646258446, + "tokens_seen": 3160735744 + }, + { + "epoch": 0.96, + "learning_rate": 3.432698094282849e-05, + "loss": 1.3082, + "theoretical_loss": 3.3359964475395376, + "tokens_seen": 3160801280 + }, + { + "epoch": 0.96, + "learning_rate": 3.431093279839518e-05, + "loss": 1.3236, + "theoretical_loss": 3.3359912305916875, + "tokens_seen": 3160866816 + }, + { + "epoch": 0.96, + "learning_rate": 3.429488465396189e-05, + "loss": 1.1909, + "theoretical_loss": 3.3359860137822874, + "tokens_seen": 3160932352 + }, + { + "epoch": 0.96, + "learning_rate": 3.427883650952859e-05, + "loss": 1.152, + "theoretical_loss": 3.3359807971113318, + "tokens_seen": 3160997888 + }, + { + "epoch": 0.96, + "learning_rate": 3.426278836509529e-05, + "loss": 1.1769, + "theoretical_loss": 3.335975580578814, + "tokens_seen": 3161063424 + }, + { + "epoch": 0.96, + "learning_rate": 3.424674022066199e-05, + "loss": 1.2458, + "theoretical_loss": 3.335970364184726, + "tokens_seen": 3161128960 + }, + { + "epoch": 0.96, + "learning_rate": 3.4230692076228685e-05, + "loss": 1.2665, + "theoretical_loss": 3.3359651479290626, + "tokens_seen": 3161194496 + }, + { + "epoch": 0.96, + "learning_rate": 3.4214643931795386e-05, + "loss": 1.3019, + "theoretical_loss": 3.3359599318118174, + "tokens_seen": 3161260032 + }, + { + "epoch": 0.96, + "learning_rate": 3.4198595787362086e-05, + "loss": 1.1935, + "theoretical_loss": 3.3359547158329828, + "tokens_seen": 3161325568 + }, + { + "epoch": 0.96, + "learning_rate": 3.418254764292879e-05, + "loss": 1.3155, + "theoretical_loss": 3.3359494999925534, + "tokens_seen": 3161391104 + }, + { + "epoch": 0.96, + "learning_rate": 3.416649949849549e-05, + "loss": 1.1475, + "theoretical_loss": 3.335944284290522, + "tokens_seen": 3161456640 + }, + { + "epoch": 0.96, + "learning_rate": 3.415045135406219e-05, + "loss": 1.182, + "theoretical_loss": 3.3359390687268817, + "tokens_seen": 3161522176 + }, + { + "epoch": 0.96, + "learning_rate": 3.413440320962889e-05, + "loss": 1.1604, + "theoretical_loss": 3.335933853301627, + "tokens_seen": 3161587712 + }, + { + "epoch": 0.96, + "learning_rate": 3.411835506519559e-05, + "loss": 1.2685, + "theoretical_loss": 3.3359286380147504, + "tokens_seen": 3161653248 + }, + { + "epoch": 0.96, + "learning_rate": 3.410230692076229e-05, + "loss": 1.2044, + "theoretical_loss": 3.335923422866246, + "tokens_seen": 3161718784 + }, + { + "epoch": 0.96, + "learning_rate": 3.408625877632899e-05, + "loss": 1.2476, + "theoretical_loss": 3.335918207856107, + "tokens_seen": 3161784320 + }, + { + "epoch": 0.96, + "learning_rate": 3.407021063189569e-05, + "loss": 1.3347, + "theoretical_loss": 3.335912992984327, + "tokens_seen": 3161849856 + }, + { + "epoch": 0.96, + "learning_rate": 3.4054162487462385e-05, + "loss": 1.255, + "theoretical_loss": 3.3359077782508995, + "tokens_seen": 3161915392 + }, + { + "epoch": 0.96, + "learning_rate": 3.4038114343029086e-05, + "loss": 1.2647, + "theoretical_loss": 3.3359025636558175, + "tokens_seen": 3161980928 + }, + { + "epoch": 0.96, + "learning_rate": 3.4022066198595786e-05, + "loss": 1.0978, + "theoretical_loss": 3.3358973491990747, + "tokens_seen": 3162046464 + }, + { + "epoch": 0.96, + "objective/train/docs_used": 1148896, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.23533296585083, + "objective/train/theoretical_loss": 3.3358921348806647, + "objective/train/tokens_used": 3182572000, + "theoretical_loss": 3.3358921348806647, + "tokens_seen": 3162112000 + }, + { + "epoch": 0.96, + "learning_rate": 3.4006018054162494e-05, + "loss": 1.2441, + "theoretical_loss": 3.3358921348806647, + "tokens_seen": 3162112000 + }, + { + "epoch": 0.96, + "learning_rate": 3.3989969909729194e-05, + "loss": 1.2836, + "theoretical_loss": 3.335886920700581, + "tokens_seen": 3162177536 + }, + { + "epoch": 0.96, + "learning_rate": 3.397392176529589e-05, + "loss": 1.2241, + "theoretical_loss": 3.335881706658817, + "tokens_seen": 3162243072 + }, + { + "epoch": 0.96, + "learning_rate": 3.395787362086259e-05, + "loss": 1.2291, + "theoretical_loss": 3.335876492755366, + "tokens_seen": 3162308608 + }, + { + "epoch": 0.96, + "learning_rate": 3.394182547642929e-05, + "loss": 1.2818, + "theoretical_loss": 3.3358712789902216, + "tokens_seen": 3162374144 + }, + { + "epoch": 0.96, + "learning_rate": 3.392577733199599e-05, + "loss": 1.3177, + "theoretical_loss": 3.335866065363377, + "tokens_seen": 3162439680 + }, + { + "epoch": 0.96, + "learning_rate": 3.390972918756269e-05, + "loss": 1.2453, + "theoretical_loss": 3.3358608518748265, + "tokens_seen": 3162505216 + }, + { + "epoch": 0.96, + "learning_rate": 3.389368104312939e-05, + "loss": 1.2498, + "theoretical_loss": 3.3358556385245626, + "tokens_seen": 3162570752 + }, + { + "epoch": 0.96, + "learning_rate": 3.3877632898696085e-05, + "loss": 1.2677, + "theoretical_loss": 3.3358504253125796, + "tokens_seen": 3162636288 + }, + { + "epoch": 0.96, + "learning_rate": 3.3861584754262786e-05, + "loss": 1.2537, + "theoretical_loss": 3.33584521223887, + "tokens_seen": 3162701824 + }, + { + "epoch": 0.96, + "learning_rate": 3.384553660982949e-05, + "loss": 1.2514, + "theoretical_loss": 3.335839999303428, + "tokens_seen": 3162767360 + }, + { + "epoch": 0.96, + "learning_rate": 3.3829488465396194e-05, + "loss": 1.1768, + "theoretical_loss": 3.335834786506247, + "tokens_seen": 3162832896 + }, + { + "epoch": 0.96, + "learning_rate": 3.3813440320962895e-05, + "loss": 1.3151, + "theoretical_loss": 3.33582957384732, + "tokens_seen": 3162898432 + }, + { + "epoch": 0.96, + "learning_rate": 3.379739217652959e-05, + "loss": 1.2007, + "theoretical_loss": 3.3358243613266407, + "tokens_seen": 3162963968 + }, + { + "epoch": 0.96, + "learning_rate": 3.378134403209629e-05, + "loss": 1.3041, + "theoretical_loss": 3.335819148944203, + "tokens_seen": 3163029504 + }, + { + "epoch": 0.96, + "learning_rate": 3.376529588766299e-05, + "loss": 1.2387, + "theoretical_loss": 3.3358139366999997, + "tokens_seen": 3163095040 + }, + { + "epoch": 0.96, + "learning_rate": 3.374924774322969e-05, + "loss": 1.2057, + "theoretical_loss": 3.335808724594025, + "tokens_seen": 3163160576 + }, + { + "epoch": 0.96, + "learning_rate": 3.373319959879639e-05, + "loss": 1.217, + "theoretical_loss": 3.3358035126262715, + "tokens_seen": 3163226112 + }, + { + "epoch": 0.96, + "learning_rate": 3.371715145436309e-05, + "loss": 1.1371, + "theoretical_loss": 3.335798300796734, + "tokens_seen": 3163291648 + }, + { + "epoch": 0.96, + "learning_rate": 3.370110330992979e-05, + "loss": 1.3386, + "theoretical_loss": 3.335793089105404, + "tokens_seen": 3163357184 + }, + { + "epoch": 0.96, + "learning_rate": 3.368505516549649e-05, + "loss": 1.3489, + "theoretical_loss": 3.335787877552277, + "tokens_seen": 3163422720 + }, + { + "epoch": 0.96, + "learning_rate": 3.3669007021063194e-05, + "loss": 1.167, + "theoretical_loss": 3.3357826661373453, + "tokens_seen": 3163488256 + }, + { + "epoch": 0.96, + "learning_rate": 3.3652958876629894e-05, + "loss": 1.2221, + "theoretical_loss": 3.3357774548606023, + "tokens_seen": 3163553792 + }, + { + "epoch": 0.96, + "learning_rate": 3.3636910732196595e-05, + "loss": 1.2175, + "theoretical_loss": 3.335772243722042, + "tokens_seen": 3163619328 + }, + { + "epoch": 0.96, + "learning_rate": 3.362086258776329e-05, + "loss": 1.3546, + "theoretical_loss": 3.335767032721658, + "tokens_seen": 3163684864 + }, + { + "epoch": 0.96, + "learning_rate": 3.360481444332999e-05, + "loss": 1.2981, + "theoretical_loss": 3.335761821859443, + "tokens_seen": 3163750400 + }, + { + "epoch": 0.96, + "learning_rate": 3.358876629889669e-05, + "loss": 1.2975, + "theoretical_loss": 3.335756611135391, + "tokens_seen": 3163815936 + }, + { + "epoch": 0.96, + "learning_rate": 3.357271815446339e-05, + "loss": 1.3049, + "theoretical_loss": 3.3357514005494955, + "tokens_seen": 3163881472 + }, + { + "epoch": 0.96, + "learning_rate": 3.355667001003009e-05, + "loss": 1.2809, + "theoretical_loss": 3.3357461901017498, + "tokens_seen": 3163947008 + }, + { + "epoch": 0.96, + "learning_rate": 3.354062186559679e-05, + "loss": 1.3018, + "theoretical_loss": 3.3357409797921473, + "tokens_seen": 3164012544 + }, + { + "epoch": 0.96, + "learning_rate": 3.352457372116349e-05, + "loss": 1.1999, + "theoretical_loss": 3.335735769620682, + "tokens_seen": 3164078080 + }, + { + "epoch": 0.96, + "learning_rate": 3.350852557673019e-05, + "loss": 1.321, + "theoretical_loss": 3.335730559587347, + "tokens_seen": 3164143616 + }, + { + "epoch": 0.96, + "learning_rate": 3.3492477432296894e-05, + "loss": 1.3242, + "theoretical_loss": 3.3357253496921357, + "tokens_seen": 3164209152 + }, + { + "epoch": 0.96, + "learning_rate": 3.3476429287863594e-05, + "loss": 1.1768, + "theoretical_loss": 3.3357201399350416, + "tokens_seen": 3164274688 + }, + { + "epoch": 0.96, + "learning_rate": 3.346038114343029e-05, + "loss": 1.3498, + "theoretical_loss": 3.3357149303160583, + "tokens_seen": 3164340224 + }, + { + "epoch": 0.96, + "learning_rate": 3.344433299899699e-05, + "loss": 1.2496, + "theoretical_loss": 3.3357097208351787, + "tokens_seen": 3164405760 + }, + { + "epoch": 0.96, + "learning_rate": 3.342828485456369e-05, + "loss": 1.1902, + "theoretical_loss": 3.3357045114923976, + "tokens_seen": 3164471296 + }, + { + "epoch": 0.96, + "learning_rate": 3.341223671013039e-05, + "loss": 1.2793, + "theoretical_loss": 3.335699302287707, + "tokens_seen": 3164536832 + }, + { + "epoch": 0.96, + "learning_rate": 3.33961885656971e-05, + "loss": 1.2692, + "theoretical_loss": 3.335694093221101, + "tokens_seen": 3164602368 + }, + { + "epoch": 0.96, + "learning_rate": 3.338014042126379e-05, + "loss": 1.2634, + "theoretical_loss": 3.335688884292574, + "tokens_seen": 3164667904 + }, + { + "epoch": 0.96, + "learning_rate": 3.336409227683049e-05, + "loss": 1.2781, + "theoretical_loss": 3.3356836755021178, + "tokens_seen": 3164733440 + }, + { + "epoch": 0.96, + "learning_rate": 3.334804413239719e-05, + "loss": 1.3321, + "theoretical_loss": 3.335678466849727, + "tokens_seen": 3164798976 + }, + { + "epoch": 0.96, + "learning_rate": 3.333199598796389e-05, + "loss": 1.3127, + "theoretical_loss": 3.3356732583353947, + "tokens_seen": 3164864512 + }, + { + "epoch": 0.96, + "learning_rate": 3.3315947843530594e-05, + "loss": 1.3314, + "theoretical_loss": 3.3356680499591147, + "tokens_seen": 3164930048 + }, + { + "epoch": 0.96, + "learning_rate": 3.3299899699097295e-05, + "loss": 1.2789, + "theoretical_loss": 3.33566284172088, + "tokens_seen": 3164995584 + }, + { + "epoch": 0.96, + "learning_rate": 3.328385155466399e-05, + "loss": 1.2107, + "theoretical_loss": 3.3356576336206842, + "tokens_seen": 3165061120 + }, + { + "epoch": 0.96, + "learning_rate": 3.326780341023069e-05, + "loss": 1.2656, + "theoretical_loss": 3.3356524256585214, + "tokens_seen": 3165126656 + }, + { + "epoch": 0.96, + "learning_rate": 3.3251755265797396e-05, + "loss": 1.335, + "theoretical_loss": 3.335647217834384, + "tokens_seen": 3165192192 + }, + { + "epoch": 0.96, + "learning_rate": 3.32357071213641e-05, + "loss": 1.3084, + "theoretical_loss": 3.3356420101482667, + "tokens_seen": 3165257728 + }, + { + "epoch": 0.96, + "learning_rate": 3.32196589769308e-05, + "loss": 1.2714, + "theoretical_loss": 3.335636802600162, + "tokens_seen": 3165323264 + }, + { + "epoch": 0.96, + "objective/train/docs_used": 1150157, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.2747828960418701, + "objective/train/theoretical_loss": 3.3356315951900637, + "objective/train/tokens_used": 3185848800, + "theoretical_loss": 3.3356315951900637, + "tokens_seen": 3165388800 + }, + { + "epoch": 0.96, + "learning_rate": 3.320361083249749e-05, + "loss": 1.2414, + "theoretical_loss": 3.3356315951900637, + "tokens_seen": 3165388800 + }, + { + "epoch": 0.96, + "learning_rate": 3.318756268806419e-05, + "loss": 1.2204, + "theoretical_loss": 3.3356263879179653, + "tokens_seen": 3165454336 + }, + { + "epoch": 0.96, + "learning_rate": 3.317151454363089e-05, + "loss": 1.2162, + "theoretical_loss": 3.3356211807838605, + "tokens_seen": 3165519872 + }, + { + "epoch": 0.96, + "learning_rate": 3.3155466399197594e-05, + "loss": 1.3166, + "theoretical_loss": 3.3356159737877427, + "tokens_seen": 3165585408 + }, + { + "epoch": 0.96, + "learning_rate": 3.3139418254764294e-05, + "loss": 1.163, + "theoretical_loss": 3.335610766929605, + "tokens_seen": 3165650944 + }, + { + "epoch": 0.96, + "learning_rate": 3.3123370110330995e-05, + "loss": 1.2813, + "theoretical_loss": 3.335605560209441, + "tokens_seen": 3165716480 + }, + { + "epoch": 0.96, + "learning_rate": 3.3107321965897695e-05, + "loss": 1.243, + "theoretical_loss": 3.3356003536272447, + "tokens_seen": 3165782016 + }, + { + "epoch": 0.96, + "learning_rate": 3.3091273821464396e-05, + "loss": 1.0884, + "theoretical_loss": 3.3355951471830094, + "tokens_seen": 3165847552 + }, + { + "epoch": 0.96, + "learning_rate": 3.30752256770311e-05, + "loss": 1.2517, + "theoretical_loss": 3.335589940876728, + "tokens_seen": 3165913088 + }, + { + "epoch": 0.96, + "learning_rate": 3.30591775325978e-05, + "loss": 1.2014, + "theoretical_loss": 3.335584734708395, + "tokens_seen": 3165978624 + }, + { + "epoch": 0.96, + "learning_rate": 3.30431293881645e-05, + "loss": 1.1194, + "theoretical_loss": 3.335579528678003, + "tokens_seen": 3166044160 + }, + { + "epoch": 0.96, + "learning_rate": 3.302708124373119e-05, + "loss": 1.2542, + "theoretical_loss": 3.3355743227855457, + "tokens_seen": 3166109696 + }, + { + "epoch": 0.96, + "learning_rate": 3.301103309929789e-05, + "loss": 1.289, + "theoretical_loss": 3.3355691170310164, + "tokens_seen": 3166175232 + }, + { + "epoch": 0.96, + "learning_rate": 3.299498495486459e-05, + "loss": 1.2885, + "theoretical_loss": 3.3355639114144093, + "tokens_seen": 3166240768 + }, + { + "epoch": 0.96, + "learning_rate": 3.2978936810431294e-05, + "loss": 1.1642, + "theoretical_loss": 3.3355587059357177, + "tokens_seen": 3166306304 + }, + { + "epoch": 0.96, + "learning_rate": 3.2962888665997994e-05, + "loss": 1.2442, + "theoretical_loss": 3.3355535005949344, + "tokens_seen": 3166371840 + }, + { + "epoch": 0.96, + "learning_rate": 3.2946840521564695e-05, + "loss": 1.1167, + "theoretical_loss": 3.3355482953920537, + "tokens_seen": 3166437376 + }, + { + "epoch": 0.96, + "learning_rate": 3.2930792377131396e-05, + "loss": 1.2847, + "theoretical_loss": 3.335543090327069, + "tokens_seen": 3166502912 + }, + { + "epoch": 0.96, + "learning_rate": 3.2914744232698096e-05, + "loss": 1.2817, + "theoretical_loss": 3.335537885399973, + "tokens_seen": 3166568448 + }, + { + "epoch": 0.96, + "learning_rate": 3.28986960882648e-05, + "loss": 1.3447, + "theoretical_loss": 3.33553268061076, + "tokens_seen": 3166633984 + }, + { + "epoch": 0.96, + "learning_rate": 3.28826479438315e-05, + "loss": 1.256, + "theoretical_loss": 3.335527475959423, + "tokens_seen": 3166699520 + }, + { + "epoch": 0.96, + "learning_rate": 3.28665997993982e-05, + "loss": 1.2308, + "theoretical_loss": 3.335522271445956, + "tokens_seen": 3166765056 + }, + { + "epoch": 0.96, + "learning_rate": 3.285055165496489e-05, + "loss": 1.2719, + "theoretical_loss": 3.3355170670703522, + "tokens_seen": 3166830592 + }, + { + "epoch": 0.96, + "learning_rate": 3.283450351053159e-05, + "loss": 1.2247, + "theoretical_loss": 3.335511862832605, + "tokens_seen": 3166896128 + }, + { + "epoch": 0.96, + "learning_rate": 3.281845536609829e-05, + "loss": 1.3491, + "theoretical_loss": 3.335506658732708, + "tokens_seen": 3166961664 + }, + { + "epoch": 0.96, + "learning_rate": 3.2802407221665e-05, + "loss": 1.2058, + "theoretical_loss": 3.3355014547706547, + "tokens_seen": 3167027200 + }, + { + "epoch": 0.96, + "learning_rate": 3.27863590772317e-05, + "loss": 1.2732, + "theoretical_loss": 3.3354962509464388, + "tokens_seen": 3167092736 + }, + { + "epoch": 0.96, + "learning_rate": 3.2770310932798395e-05, + "loss": 1.3729, + "theoretical_loss": 3.335491047260054, + "tokens_seen": 3167158272 + }, + { + "epoch": 0.96, + "learning_rate": 3.2754262788365096e-05, + "loss": 1.2805, + "theoretical_loss": 3.335485843711493, + "tokens_seen": 3167223808 + }, + { + "epoch": 0.96, + "learning_rate": 3.2738214643931796e-05, + "loss": 1.2143, + "theoretical_loss": 3.3354806403007493, + "tokens_seen": 3167289344 + }, + { + "epoch": 0.96, + "learning_rate": 3.27221664994985e-05, + "loss": 1.201, + "theoretical_loss": 3.3354754370278172, + "tokens_seen": 3167354880 + }, + { + "epoch": 0.96, + "learning_rate": 3.27061183550652e-05, + "loss": 1.2245, + "theoretical_loss": 3.3354702338926896, + "tokens_seen": 3167420416 + }, + { + "epoch": 0.96, + "learning_rate": 3.26900702106319e-05, + "loss": 1.197, + "theoretical_loss": 3.33546503089536, + "tokens_seen": 3167485952 + }, + { + "epoch": 0.96, + "learning_rate": 3.267402206619859e-05, + "loss": 1.2632, + "theoretical_loss": 3.3354598280358227, + "tokens_seen": 3167551488 + }, + { + "epoch": 0.96, + "learning_rate": 3.265797392176529e-05, + "loss": 1.3214, + "theoretical_loss": 3.33545462531407, + "tokens_seen": 3167617024 + }, + { + "epoch": 0.96, + "learning_rate": 3.2641925777332e-05, + "loss": 1.266, + "theoretical_loss": 3.335449422730097, + "tokens_seen": 3167682560 + }, + { + "epoch": 0.96, + "learning_rate": 3.26258776328987e-05, + "loss": 1.2562, + "theoretical_loss": 3.335444220283895, + "tokens_seen": 3167748096 + }, + { + "epoch": 0.96, + "learning_rate": 3.26098294884654e-05, + "loss": 1.201, + "theoretical_loss": 3.3354390179754594, + "tokens_seen": 3167813632 + }, + { + "epoch": 0.96, + "learning_rate": 3.2593781344032095e-05, + "loss": 1.2407, + "theoretical_loss": 3.335433815804783, + "tokens_seen": 3167879168 + }, + { + "epoch": 0.96, + "learning_rate": 3.2577733199598796e-05, + "loss": 1.2421, + "theoretical_loss": 3.335428613771859, + "tokens_seen": 3167944704 + }, + { + "epoch": 0.96, + "learning_rate": 3.25616850551655e-05, + "loss": 1.2386, + "theoretical_loss": 3.3354234118766817, + "tokens_seen": 3168010240 + }, + { + "epoch": 0.96, + "learning_rate": 3.25456369107322e-05, + "loss": 1.2734, + "theoretical_loss": 3.3354182101192436, + "tokens_seen": 3168075776 + }, + { + "epoch": 0.96, + "learning_rate": 3.25295887662989e-05, + "loss": 1.3103, + "theoretical_loss": 3.335413008499539, + "tokens_seen": 3168141312 + }, + { + "epoch": 0.96, + "learning_rate": 3.25135406218656e-05, + "loss": 1.3677, + "theoretical_loss": 3.335407807017561, + "tokens_seen": 3168206848 + }, + { + "epoch": 0.96, + "learning_rate": 3.24974924774323e-05, + "loss": 1.2973, + "theoretical_loss": 3.335402605673303, + "tokens_seen": 3168272384 + }, + { + "epoch": 0.96, + "learning_rate": 3.2481444332999e-05, + "loss": 1.3016, + "theoretical_loss": 3.335397404466759, + "tokens_seen": 3168337920 + }, + { + "epoch": 0.96, + "learning_rate": 3.24653961885657e-05, + "loss": 1.2762, + "theoretical_loss": 3.335392203397922, + "tokens_seen": 3168403456 + }, + { + "epoch": 0.96, + "learning_rate": 3.24493480441324e-05, + "loss": 1.3137, + "theoretical_loss": 3.3353870024667858, + "tokens_seen": 3168468992 + }, + { + "epoch": 0.96, + "learning_rate": 3.24332998996991e-05, + "loss": 1.2341, + "theoretical_loss": 3.335381801673344, + "tokens_seen": 3168534528 + }, + { + "epoch": 0.96, + "learning_rate": 3.2417251755265796e-05, + "loss": 1.2267, + "theoretical_loss": 3.33537660101759, + "tokens_seen": 3168600064 + }, + { + "epoch": 0.96, + "objective/train/docs_used": 1151398, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.296938419342041, + "objective/train/theoretical_loss": 3.3353714004995174, + "objective/train/tokens_used": 3189125600, + "theoretical_loss": 3.3353714004995174, + "tokens_seen": 3168665600 + }, + { + "epoch": 0.96, + "learning_rate": 3.2401203610832496e-05, + "loss": 1.2799, + "theoretical_loss": 3.3353714004995174, + "tokens_seen": 3168665600 + }, + { + "epoch": 0.96, + "learning_rate": 3.23851554663992e-05, + "loss": 1.2078, + "theoretical_loss": 3.335366200119119, + "tokens_seen": 3168731136 + }, + { + "epoch": 0.96, + "learning_rate": 3.23691073219659e-05, + "loss": 1.3537, + "theoretical_loss": 3.3353609998763893, + "tokens_seen": 3168796672 + }, + { + "epoch": 0.96, + "learning_rate": 3.2353059177532605e-05, + "loss": 1.1238, + "theoretical_loss": 3.3353557997713215, + "tokens_seen": 3168862208 + }, + { + "epoch": 0.96, + "learning_rate": 3.23370110330993e-05, + "loss": 1.2711, + "theoretical_loss": 3.335350599803909, + "tokens_seen": 3168927744 + }, + { + "epoch": 0.96, + "learning_rate": 3.2320962888666e-05, + "loss": 1.1729, + "theoretical_loss": 3.3353453999741447, + "tokens_seen": 3168993280 + }, + { + "epoch": 0.96, + "learning_rate": 3.23049147442327e-05, + "loss": 1.2447, + "theoretical_loss": 3.3353402002820234, + "tokens_seen": 3169058816 + }, + { + "epoch": 0.96, + "learning_rate": 3.22888665997994e-05, + "loss": 1.2472, + "theoretical_loss": 3.3353350007275377, + "tokens_seen": 3169124352 + }, + { + "epoch": 0.96, + "learning_rate": 3.22728184553661e-05, + "loss": 1.1887, + "theoretical_loss": 3.335329801310681, + "tokens_seen": 3169189888 + }, + { + "epoch": 0.96, + "learning_rate": 3.22567703109328e-05, + "loss": 1.2937, + "theoretical_loss": 3.3353246020314478, + "tokens_seen": 3169255424 + }, + { + "epoch": 0.96, + "learning_rate": 3.2240722166499496e-05, + "loss": 1.2046, + "theoretical_loss": 3.33531940288983, + "tokens_seen": 3169320960 + }, + { + "epoch": 0.96, + "learning_rate": 3.2224674022066196e-05, + "loss": 1.1957, + "theoretical_loss": 3.3353142038858232, + "tokens_seen": 3169386496 + }, + { + "epoch": 0.96, + "learning_rate": 3.22086258776329e-05, + "loss": 1.2044, + "theoretical_loss": 3.335309005019419, + "tokens_seen": 3169452032 + }, + { + "epoch": 0.96, + "learning_rate": 3.2192577733199604e-05, + "loss": 1.3168, + "theoretical_loss": 3.3353038062906117, + "tokens_seen": 3169517568 + }, + { + "epoch": 0.96, + "learning_rate": 3.2176529588766305e-05, + "loss": 1.3041, + "theoretical_loss": 3.335298607699395, + "tokens_seen": 3169583104 + }, + { + "epoch": 0.96, + "learning_rate": 3.2160481444333e-05, + "loss": 1.3418, + "theoretical_loss": 3.3352934092457627, + "tokens_seen": 3169648640 + }, + { + "epoch": 0.96, + "learning_rate": 3.21444332998997e-05, + "loss": 1.2397, + "theoretical_loss": 3.3352882109297073, + "tokens_seen": 3169714176 + }, + { + "epoch": 0.96, + "learning_rate": 3.21283851554664e-05, + "loss": 1.2554, + "theoretical_loss": 3.3352830127512227, + "tokens_seen": 3169779712 + }, + { + "epoch": 0.96, + "learning_rate": 3.21123370110331e-05, + "loss": 1.1771, + "theoretical_loss": 3.335277814710303, + "tokens_seen": 3169845248 + }, + { + "epoch": 0.96, + "learning_rate": 3.20962888665998e-05, + "loss": 1.2864, + "theoretical_loss": 3.335272616806941, + "tokens_seen": 3169910784 + }, + { + "epoch": 0.96, + "learning_rate": 3.20802407221665e-05, + "loss": 1.2376, + "theoretical_loss": 3.3352674190411307, + "tokens_seen": 3169976320 + }, + { + "epoch": 0.96, + "learning_rate": 3.2064192577733196e-05, + "loss": 1.2755, + "theoretical_loss": 3.335262221412865, + "tokens_seen": 3170041856 + }, + { + "epoch": 0.96, + "learning_rate": 3.2048144433299903e-05, + "loss": 1.2862, + "theoretical_loss": 3.3352570239221384, + "tokens_seen": 3170107392 + }, + { + "epoch": 0.96, + "learning_rate": 3.2032096288866604e-05, + "loss": 1.3015, + "theoretical_loss": 3.3352518265689435, + "tokens_seen": 3170172928 + }, + { + "epoch": 0.96, + "learning_rate": 3.2016048144433305e-05, + "loss": 1.3318, + "theoretical_loss": 3.335246629353274, + "tokens_seen": 3170238464 + }, + { + "epoch": 0.96, + "learning_rate": 3.2000000000000005e-05, + "loss": 1.1851, + "theoretical_loss": 3.335241432275124, + "tokens_seen": 3170304000 + }, + { + "epoch": 0.96, + "learning_rate": 3.19839518555667e-05, + "loss": 1.2015, + "theoretical_loss": 3.335236235334486, + "tokens_seen": 3170369536 + }, + { + "epoch": 0.96, + "learning_rate": 3.19679037111334e-05, + "loss": 1.221, + "theoretical_loss": 3.335231038531355, + "tokens_seen": 3170435072 + }, + { + "epoch": 0.96, + "learning_rate": 3.19518555667001e-05, + "loss": 1.2579, + "theoretical_loss": 3.335225841865723, + "tokens_seen": 3170500608 + }, + { + "epoch": 0.96, + "learning_rate": 3.19358074222668e-05, + "loss": 1.1365, + "theoretical_loss": 3.335220645337584, + "tokens_seen": 3170566144 + }, + { + "epoch": 0.96, + "learning_rate": 3.19197592778335e-05, + "loss": 1.2064, + "theoretical_loss": 3.3352154489469323, + "tokens_seen": 3170631680 + }, + { + "epoch": 0.96, + "learning_rate": 3.19037111334002e-05, + "loss": 1.3005, + "theoretical_loss": 3.33521025269376, + "tokens_seen": 3170697216 + }, + { + "epoch": 0.96, + "learning_rate": 3.18876629889669e-05, + "loss": 1.2992, + "theoretical_loss": 3.3352050565780624, + "tokens_seen": 3170762752 + }, + { + "epoch": 0.96, + "learning_rate": 3.1871614844533604e-05, + "loss": 1.1501, + "theoretical_loss": 3.3351998605998316, + "tokens_seen": 3170828288 + }, + { + "epoch": 0.96, + "learning_rate": 3.1855566700100304e-05, + "loss": 1.2444, + "theoretical_loss": 3.3351946647590616, + "tokens_seen": 3170893824 + }, + { + "epoch": 0.96, + "learning_rate": 3.1839518555667005e-05, + "loss": 1.3735, + "theoretical_loss": 3.3351894690557455, + "tokens_seen": 3170959360 + }, + { + "epoch": 0.96, + "learning_rate": 3.1823470411233706e-05, + "loss": 1.1888, + "theoretical_loss": 3.3351842734898773, + "tokens_seen": 3171024896 + }, + { + "epoch": 0.96, + "learning_rate": 3.18074222668004e-05, + "loss": 1.2522, + "theoretical_loss": 3.3351790780614508, + "tokens_seen": 3171090432 + }, + { + "epoch": 0.96, + "learning_rate": 3.17913741223671e-05, + "loss": 1.2208, + "theoretical_loss": 3.335173882770459, + "tokens_seen": 3171155968 + }, + { + "epoch": 0.96, + "learning_rate": 3.17753259779338e-05, + "loss": 1.2787, + "theoretical_loss": 3.3351686876168958, + "tokens_seen": 3171221504 + }, + { + "epoch": 0.96, + "learning_rate": 3.17592778335005e-05, + "loss": 1.2826, + "theoretical_loss": 3.335163492600754, + "tokens_seen": 3171287040 + }, + { + "epoch": 0.96, + "learning_rate": 3.174322968906721e-05, + "loss": 1.2684, + "theoretical_loss": 3.335158297722028, + "tokens_seen": 3171352576 + }, + { + "epoch": 0.96, + "learning_rate": 3.17271815446339e-05, + "loss": 1.299, + "theoretical_loss": 3.335153102980711, + "tokens_seen": 3171418112 + }, + { + "epoch": 0.96, + "learning_rate": 3.17111334002006e-05, + "loss": 1.3104, + "theoretical_loss": 3.335147908376796, + "tokens_seen": 3171483648 + }, + { + "epoch": 0.96, + "learning_rate": 3.1695085255767304e-05, + "loss": 1.2621, + "theoretical_loss": 3.335142713910278, + "tokens_seen": 3171549184 + }, + { + "epoch": 0.96, + "learning_rate": 3.1679037111334004e-05, + "loss": 1.2446, + "theoretical_loss": 3.3351375195811483, + "tokens_seen": 3171614720 + }, + { + "epoch": 0.96, + "learning_rate": 3.1662988966900705e-05, + "loss": 1.2099, + "theoretical_loss": 3.3351323253894027, + "tokens_seen": 3171680256 + }, + { + "epoch": 0.96, + "learning_rate": 3.1646940822467406e-05, + "loss": 1.2681, + "theoretical_loss": 3.335127131335033, + "tokens_seen": 3171745792 + }, + { + "epoch": 0.96, + "learning_rate": 3.16308926780341e-05, + "loss": 1.2788, + "theoretical_loss": 3.335121937418034, + "tokens_seen": 3171811328 + }, + { + "epoch": 0.96, + "learning_rate": 3.16148445336008e-05, + "loss": 1.2667, + "theoretical_loss": 3.335116743638398, + "tokens_seen": 3171876864 + }, + { + "epoch": 0.96, + "objective/train/docs_used": 1152731, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.3690487146377563, + "objective/train/theoretical_loss": 3.3351115499961197, + "objective/train/tokens_used": 3192402400, + "theoretical_loss": 3.3351115499961197, + "tokens_seen": 3171942400 + }, + { + "epoch": 0.96, + "learning_rate": 3.159879638916751e-05, + "loss": 1.2485, + "theoretical_loss": 3.3351115499961197, + "tokens_seen": 3171942400 + }, + { + "epoch": 0.96, + "learning_rate": 3.158274824473421e-05, + "loss": 1.228, + "theoretical_loss": 3.335106356491192, + "tokens_seen": 3172007936 + }, + { + "epoch": 0.96, + "learning_rate": 3.156670010030091e-05, + "loss": 1.2613, + "theoretical_loss": 3.3351011631236083, + "tokens_seen": 3172073472 + }, + { + "epoch": 0.96, + "learning_rate": 3.15506519558676e-05, + "loss": 1.2143, + "theoretical_loss": 3.3350959698933624, + "tokens_seen": 3172139008 + }, + { + "epoch": 0.96, + "learning_rate": 3.1534603811434303e-05, + "loss": 1.3118, + "theoretical_loss": 3.335090776800448, + "tokens_seen": 3172204544 + }, + { + "epoch": 0.96, + "learning_rate": 3.1518555667001004e-05, + "loss": 1.39, + "theoretical_loss": 3.3350855838448585, + "tokens_seen": 3172270080 + }, + { + "epoch": 0.96, + "learning_rate": 3.1502507522567705e-05, + "loss": 1.1793, + "theoretical_loss": 3.335080391026587, + "tokens_seen": 3172335616 + }, + { + "epoch": 0.96, + "learning_rate": 3.1486459378134405e-05, + "loss": 1.1999, + "theoretical_loss": 3.3350751983456277, + "tokens_seen": 3172401152 + }, + { + "epoch": 0.96, + "learning_rate": 3.1470411233701106e-05, + "loss": 1.2398, + "theoretical_loss": 3.3350700058019735, + "tokens_seen": 3172466688 + }, + { + "epoch": 0.96, + "learning_rate": 3.1454363089267807e-05, + "loss": 1.2698, + "theoretical_loss": 3.3350648133956184, + "tokens_seen": 3172532224 + }, + { + "epoch": 0.96, + "learning_rate": 3.143831494483451e-05, + "loss": 1.3485, + "theoretical_loss": 3.335059621126556, + "tokens_seen": 3172597760 + }, + { + "epoch": 0.96, + "learning_rate": 3.142226680040121e-05, + "loss": 1.27, + "theoretical_loss": 3.3350544289947797, + "tokens_seen": 3172663296 + }, + { + "epoch": 0.96, + "learning_rate": 3.140621865596791e-05, + "loss": 1.2607, + "theoretical_loss": 3.3350492370002827, + "tokens_seen": 3172728832 + }, + { + "epoch": 0.96, + "learning_rate": 3.13901705115346e-05, + "loss": 1.1454, + "theoretical_loss": 3.335044045143059, + "tokens_seen": 3172794368 + }, + { + "epoch": 0.96, + "learning_rate": 3.13741223671013e-05, + "loss": 1.1844, + "theoretical_loss": 3.3350388534231015, + "tokens_seen": 3172859904 + }, + { + "epoch": 0.96, + "learning_rate": 3.1358074222668004e-05, + "loss": 1.2177, + "theoretical_loss": 3.335033661840405, + "tokens_seen": 3172925440 + }, + { + "epoch": 0.96, + "learning_rate": 3.1342026078234704e-05, + "loss": 1.305, + "theoretical_loss": 3.335028470394961, + "tokens_seen": 3172990976 + }, + { + "epoch": 0.96, + "learning_rate": 3.1325977933801405e-05, + "loss": 1.3288, + "theoretical_loss": 3.3350232790867653, + "tokens_seen": 3173056512 + }, + { + "epoch": 0.96, + "learning_rate": 3.1309929789368106e-05, + "loss": 1.3852, + "theoretical_loss": 3.3350180879158096, + "tokens_seen": 3173122048 + }, + { + "epoch": 0.96, + "learning_rate": 3.1293881644934806e-05, + "loss": 1.2384, + "theoretical_loss": 3.335012896882089, + "tokens_seen": 3173187584 + }, + { + "epoch": 0.96, + "learning_rate": 3.127783350050151e-05, + "loss": 1.2851, + "theoretical_loss": 3.335007705985596, + "tokens_seen": 3173253120 + }, + { + "epoch": 0.96, + "learning_rate": 3.126178535606821e-05, + "loss": 1.3004, + "theoretical_loss": 3.335002515226324, + "tokens_seen": 3173318656 + }, + { + "epoch": 0.96, + "learning_rate": 3.124573721163491e-05, + "loss": 1.2597, + "theoretical_loss": 3.334997324604267, + "tokens_seen": 3173384192 + }, + { + "epoch": 0.96, + "learning_rate": 3.122968906720161e-05, + "loss": 1.2198, + "theoretical_loss": 3.3349921341194184, + "tokens_seen": 3173449728 + }, + { + "epoch": 0.96, + "learning_rate": 3.12136409227683e-05, + "loss": 1.2326, + "theoretical_loss": 3.3349869437717725, + "tokens_seen": 3173515264 + }, + { + "epoch": 0.96, + "learning_rate": 3.1197592778335e-05, + "loss": 1.1745, + "theoretical_loss": 3.3349817535613218, + "tokens_seen": 3173580800 + }, + { + "epoch": 0.96, + "learning_rate": 3.1181544633901704e-05, + "loss": 1.2016, + "theoretical_loss": 3.33497656348806, + "tokens_seen": 3173646336 + }, + { + "epoch": 0.96, + "learning_rate": 3.1165496489468404e-05, + "loss": 1.184, + "theoretical_loss": 3.334971373551981, + "tokens_seen": 3173711872 + }, + { + "epoch": 0.96, + "learning_rate": 3.114944834503511e-05, + "loss": 1.3535, + "theoretical_loss": 3.3349661837530777, + "tokens_seen": 3173777408 + }, + { + "epoch": 0.96, + "learning_rate": 3.1133400200601806e-05, + "loss": 1.2635, + "theoretical_loss": 3.3349609940913445, + "tokens_seen": 3173842944 + }, + { + "epoch": 0.96, + "learning_rate": 3.1117352056168506e-05, + "loss": 1.2959, + "theoretical_loss": 3.3349558045667744, + "tokens_seen": 3173908480 + }, + { + "epoch": 0.96, + "learning_rate": 3.110130391173521e-05, + "loss": 1.2454, + "theoretical_loss": 3.3349506151793613, + "tokens_seen": 3173974016 + }, + { + "epoch": 0.96, + "learning_rate": 3.108525576730191e-05, + "loss": 1.2189, + "theoretical_loss": 3.3349454259290985, + "tokens_seen": 3174039552 + }, + { + "epoch": 0.96, + "learning_rate": 3.106920762286861e-05, + "loss": 1.1698, + "theoretical_loss": 3.3349402368159797, + "tokens_seen": 3174105088 + }, + { + "epoch": 0.96, + "learning_rate": 3.105315947843531e-05, + "loss": 1.3171, + "theoretical_loss": 3.334935047839998, + "tokens_seen": 3174170624 + }, + { + "epoch": 0.96, + "learning_rate": 3.1037111334002e-05, + "loss": 1.2061, + "theoretical_loss": 3.3349298590011474, + "tokens_seen": 3174236160 + }, + { + "epoch": 0.96, + "learning_rate": 3.10210631895687e-05, + "loss": 1.2213, + "theoretical_loss": 3.3349246702994213, + "tokens_seen": 3174301696 + }, + { + "epoch": 0.96, + "learning_rate": 3.100501504513541e-05, + "loss": 1.315, + "theoretical_loss": 3.3349194817348136, + "tokens_seen": 3174367232 + }, + { + "epoch": 0.96, + "learning_rate": 3.098896690070211e-05, + "loss": 1.2514, + "theoretical_loss": 3.334914293307317, + "tokens_seen": 3174432768 + }, + { + "epoch": 0.96, + "learning_rate": 3.097291875626881e-05, + "loss": 1.2947, + "theoretical_loss": 3.334909105016926, + "tokens_seen": 3174498304 + }, + { + "epoch": 0.96, + "learning_rate": 3.0956870611835506e-05, + "loss": 1.3201, + "theoretical_loss": 3.334903916863633, + "tokens_seen": 3174563840 + }, + { + "epoch": 0.96, + "learning_rate": 3.0940822467402207e-05, + "loss": 1.267, + "theoretical_loss": 3.334898728847433, + "tokens_seen": 3174629376 + }, + { + "epoch": 0.96, + "learning_rate": 3.092477432296891e-05, + "loss": 1.2822, + "theoretical_loss": 3.3348935409683187, + "tokens_seen": 3174694912 + }, + { + "epoch": 0.96, + "learning_rate": 3.090872617853561e-05, + "loss": 1.1936, + "theoretical_loss": 3.3348883532262836, + "tokens_seen": 3174760448 + }, + { + "epoch": 0.96, + "learning_rate": 3.089267803410231e-05, + "loss": 1.2874, + "theoretical_loss": 3.334883165621321, + "tokens_seen": 3174825984 + }, + { + "epoch": 0.96, + "learning_rate": 3.087662988966901e-05, + "loss": 1.2471, + "theoretical_loss": 3.3348779781534255, + "tokens_seen": 3174891520 + }, + { + "epoch": 0.96, + "learning_rate": 3.08605817452357e-05, + "loss": 1.3368, + "theoretical_loss": 3.3348727908225895, + "tokens_seen": 3174957056 + }, + { + "epoch": 0.96, + "learning_rate": 3.084453360080241e-05, + "loss": 1.2188, + "theoretical_loss": 3.3348676036288074, + "tokens_seen": 3175022592 + }, + { + "epoch": 0.96, + "learning_rate": 3.082848545636911e-05, + "loss": 1.2733, + "theoretical_loss": 3.334862416572072, + "tokens_seen": 3175088128 + }, + { + "epoch": 0.96, + "learning_rate": 3.081243731193581e-05, + "loss": 1.3299, + "theoretical_loss": 3.3348572296523775, + "tokens_seen": 3175153664 + }, + { + "epoch": 0.96, + "objective/train/docs_used": 1153914, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.20263671875, + "objective/train/theoretical_loss": 3.3348520428697173, + "objective/train/tokens_used": 3195679200, + "theoretical_loss": 3.3348520428697173, + "tokens_seen": 3175219200 + }, + { + "epoch": 0.96, + "learning_rate": 3.079638916750251e-05, + "loss": 1.2237, + "theoretical_loss": 3.3348520428697173, + "tokens_seen": 3175219200 + }, + { + "epoch": 0.96, + "learning_rate": 3.0780341023069206e-05, + "loss": 1.2104, + "theoretical_loss": 3.3348468562240847, + "tokens_seen": 3175284736 + }, + { + "epoch": 0.96, + "learning_rate": 3.076429287863591e-05, + "loss": 1.2477, + "theoretical_loss": 3.3348416697154732, + "tokens_seen": 3175350272 + }, + { + "epoch": 0.96, + "learning_rate": 3.074824473420261e-05, + "loss": 1.2533, + "theoretical_loss": 3.334836483343877, + "tokens_seen": 3175415808 + }, + { + "epoch": 0.96, + "learning_rate": 3.073219658976931e-05, + "loss": 1.2535, + "theoretical_loss": 3.334831297109289, + "tokens_seen": 3175481344 + }, + { + "epoch": 0.96, + "learning_rate": 3.071614844533601e-05, + "loss": 1.2636, + "theoretical_loss": 3.3348261110117026, + "tokens_seen": 3175546880 + }, + { + "epoch": 0.96, + "learning_rate": 3.070010030090271e-05, + "loss": 1.1779, + "theoretical_loss": 3.3348209250511123, + "tokens_seen": 3175612416 + }, + { + "epoch": 0.96, + "learning_rate": 3.068405215646941e-05, + "loss": 1.2216, + "theoretical_loss": 3.3348157392275106, + "tokens_seen": 3175677952 + }, + { + "epoch": 0.96, + "learning_rate": 3.066800401203611e-05, + "loss": 1.2449, + "theoretical_loss": 3.334810553540892, + "tokens_seen": 3175743488 + }, + { + "epoch": 0.96, + "learning_rate": 3.065195586760281e-05, + "loss": 1.3231, + "theoretical_loss": 3.3348053679912493, + "tokens_seen": 3175809024 + }, + { + "epoch": 0.96, + "learning_rate": 3.063590772316951e-05, + "loss": 1.3719, + "theoretical_loss": 3.334800182578576, + "tokens_seen": 3175874560 + }, + { + "epoch": 0.96, + "learning_rate": 3.061985957873621e-05, + "loss": 1.2896, + "theoretical_loss": 3.334794997302866, + "tokens_seen": 3175940096 + }, + { + "epoch": 0.96, + "learning_rate": 3.0603811434302906e-05, + "loss": 1.2895, + "theoretical_loss": 3.3347898121641135, + "tokens_seen": 3176005632 + }, + { + "epoch": 0.96, + "learning_rate": 3.058776328986961e-05, + "loss": 1.2695, + "theoretical_loss": 3.334784627162311, + "tokens_seen": 3176071168 + }, + { + "epoch": 0.96, + "learning_rate": 3.057171514543631e-05, + "loss": 1.227, + "theoretical_loss": 3.334779442297452, + "tokens_seen": 3176136704 + }, + { + "epoch": 0.96, + "learning_rate": 3.055566700100301e-05, + "loss": 1.3284, + "theoretical_loss": 3.3347742575695314, + "tokens_seen": 3176202240 + }, + { + "epoch": 0.96, + "learning_rate": 3.0539618856569716e-05, + "loss": 1.3306, + "theoretical_loss": 3.334769072978541, + "tokens_seen": 3176267776 + }, + { + "epoch": 0.96, + "learning_rate": 3.052357071213641e-05, + "loss": 1.239, + "theoretical_loss": 3.334763888524476, + "tokens_seen": 3176333312 + }, + { + "epoch": 0.96, + "learning_rate": 3.050752256770311e-05, + "loss": 1.3183, + "theoretical_loss": 3.334758704207329, + "tokens_seen": 3176398848 + }, + { + "epoch": 0.96, + "learning_rate": 3.049147442326981e-05, + "loss": 1.3481, + "theoretical_loss": 3.334753520027093, + "tokens_seen": 3176464384 + }, + { + "epoch": 0.96, + "learning_rate": 3.047542627883651e-05, + "loss": 1.2351, + "theoretical_loss": 3.334748335983763, + "tokens_seen": 3176529920 + }, + { + "epoch": 0.96, + "learning_rate": 3.045937813440321e-05, + "loss": 1.3278, + "theoretical_loss": 3.3347431520773316, + "tokens_seen": 3176595456 + }, + { + "epoch": 0.96, + "learning_rate": 3.044332998996991e-05, + "loss": 1.3273, + "theoretical_loss": 3.3347379683077927, + "tokens_seen": 3176660992 + }, + { + "epoch": 0.96, + "learning_rate": 3.042728184553661e-05, + "loss": 1.3162, + "theoretical_loss": 3.33473278467514, + "tokens_seen": 3176726528 + }, + { + "epoch": 0.96, + "learning_rate": 3.0411233701103307e-05, + "loss": 1.2567, + "theoretical_loss": 3.3347276011793667, + "tokens_seen": 3176792064 + }, + { + "epoch": 0.96, + "learning_rate": 3.0395185556670015e-05, + "loss": 1.1876, + "theoretical_loss": 3.3347224178204664, + "tokens_seen": 3176857600 + }, + { + "epoch": 0.96, + "learning_rate": 3.0379137412236712e-05, + "loss": 1.1887, + "theoretical_loss": 3.3347172345984326, + "tokens_seen": 3176923136 + }, + { + "epoch": 0.96, + "learning_rate": 3.0363089267803412e-05, + "loss": 1.2624, + "theoretical_loss": 3.3347120515132596, + "tokens_seen": 3176988672 + }, + { + "epoch": 0.96, + "learning_rate": 3.0347041123370113e-05, + "loss": 1.1924, + "theoretical_loss": 3.3347068685649397, + "tokens_seen": 3177054208 + }, + { + "epoch": 0.96, + "learning_rate": 3.033099297893681e-05, + "loss": 1.2541, + "theoretical_loss": 3.3347016857534677, + "tokens_seen": 3177119744 + }, + { + "epoch": 0.96, + "learning_rate": 3.031494483450351e-05, + "loss": 1.2346, + "theoretical_loss": 3.3346965030788365, + "tokens_seen": 3177185280 + }, + { + "epoch": 0.96, + "learning_rate": 3.029889669007021e-05, + "loss": 1.3022, + "theoretical_loss": 3.33469132054104, + "tokens_seen": 3177250816 + }, + { + "epoch": 0.96, + "learning_rate": 3.028284854563691e-05, + "loss": 1.2689, + "theoretical_loss": 3.334686138140071, + "tokens_seen": 3177316352 + }, + { + "epoch": 0.96, + "learning_rate": 3.026680040120361e-05, + "loss": 1.2902, + "theoretical_loss": 3.334680955875924, + "tokens_seen": 3177381888 + }, + { + "epoch": 0.96, + "learning_rate": 3.0250752256770314e-05, + "loss": 1.2211, + "theoretical_loss": 3.334675773748592, + "tokens_seen": 3177447424 + }, + { + "epoch": 0.96, + "learning_rate": 3.0234704112337014e-05, + "loss": 1.3033, + "theoretical_loss": 3.334670591758069, + "tokens_seen": 3177512960 + }, + { + "epoch": 0.96, + "learning_rate": 3.0218655967903715e-05, + "loss": 1.2829, + "theoretical_loss": 3.334665409904348, + "tokens_seen": 3177578496 + }, + { + "epoch": 0.96, + "learning_rate": 3.0202607823470412e-05, + "loss": 1.2598, + "theoretical_loss": 3.334660228187423, + "tokens_seen": 3177644032 + }, + { + "epoch": 0.96, + "learning_rate": 3.0186559679037113e-05, + "loss": 1.1741, + "theoretical_loss": 3.3346550466072875, + "tokens_seen": 3177709568 + }, + { + "epoch": 0.96, + "learning_rate": 3.0170511534603813e-05, + "loss": 1.2638, + "theoretical_loss": 3.334649865163935, + "tokens_seen": 3177775104 + }, + { + "epoch": 0.96, + "learning_rate": 3.015446339017051e-05, + "loss": 1.2286, + "theoretical_loss": 3.334644683857359, + "tokens_seen": 3177840640 + }, + { + "epoch": 0.96, + "learning_rate": 3.013841524573721e-05, + "loss": 1.2374, + "theoretical_loss": 3.3346395026875535, + "tokens_seen": 3177906176 + }, + { + "epoch": 0.96, + "learning_rate": 3.0122367101303912e-05, + "loss": 1.2386, + "theoretical_loss": 3.3346343216545113, + "tokens_seen": 3177971712 + }, + { + "epoch": 0.96, + "learning_rate": 3.010631895687061e-05, + "loss": 1.2926, + "theoretical_loss": 3.334629140758227, + "tokens_seen": 3178037248 + }, + { + "epoch": 0.96, + "learning_rate": 3.0090270812437316e-05, + "loss": 1.2085, + "theoretical_loss": 3.3346239599986927, + "tokens_seen": 3178102784 + }, + { + "epoch": 0.96, + "learning_rate": 3.0074222668004014e-05, + "loss": 1.1796, + "theoretical_loss": 3.3346187793759032, + "tokens_seen": 3178168320 + }, + { + "epoch": 0.96, + "learning_rate": 3.0058174523570714e-05, + "loss": 1.2748, + "theoretical_loss": 3.334613598889852, + "tokens_seen": 3178233856 + }, + { + "epoch": 0.96, + "learning_rate": 3.0042126379137415e-05, + "loss": 1.1449, + "theoretical_loss": 3.3346084185405322, + "tokens_seen": 3178299392 + }, + { + "epoch": 0.96, + "learning_rate": 3.0026078234704112e-05, + "loss": 1.2533, + "theoretical_loss": 3.3346032383279374, + "tokens_seen": 3178364928 + }, + { + "epoch": 0.96, + "learning_rate": 3.0010030090270813e-05, + "loss": 1.2335, + "theoretical_loss": 3.3345980582520616, + "tokens_seen": 3178430464 + }, + { + "epoch": 0.96, + "objective/train/docs_used": 1155119, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.3053851127624512, + "objective/train/theoretical_loss": 3.3345928783128977, + "objective/train/tokens_used": 3198956000, + "theoretical_loss": 3.3345928783128977, + "tokens_seen": 3178496000 + }, + { + "epoch": 0.96, + "learning_rate": 2.9993981945837513e-05, + "loss": 1.2844, + "theoretical_loss": 3.3345928783128977, + "tokens_seen": 3178496000 + }, + { + "epoch": 0.96, + "learning_rate": 2.997793380140421e-05, + "loss": 1.2557, + "theoretical_loss": 3.33458769851044, + "tokens_seen": 3178561536 + }, + { + "epoch": 0.96, + "learning_rate": 2.996188565697091e-05, + "loss": 1.3043, + "theoretical_loss": 3.3345825188446816, + "tokens_seen": 3178627072 + }, + { + "epoch": 0.96, + "learning_rate": 2.9945837512537615e-05, + "loss": 1.2585, + "theoretical_loss": 3.3345773393156164, + "tokens_seen": 3178692608 + }, + { + "epoch": 0.96, + "learning_rate": 2.9929789368104316e-05, + "loss": 1.2789, + "theoretical_loss": 3.334572159923238, + "tokens_seen": 3178758144 + }, + { + "epoch": 0.96, + "learning_rate": 2.9913741223671017e-05, + "loss": 1.2758, + "theoretical_loss": 3.334566980667539, + "tokens_seen": 3178823680 + }, + { + "epoch": 0.96, + "learning_rate": 2.9897693079237714e-05, + "loss": 1.3357, + "theoretical_loss": 3.334561801548514, + "tokens_seen": 3178889216 + }, + { + "epoch": 0.96, + "learning_rate": 2.9881644934804415e-05, + "loss": 1.2189, + "theoretical_loss": 3.334556622566157, + "tokens_seen": 3178954752 + }, + { + "epoch": 0.96, + "learning_rate": 2.9865596790371115e-05, + "loss": 1.1841, + "theoretical_loss": 3.33455144372046, + "tokens_seen": 3179020288 + }, + { + "epoch": 0.96, + "learning_rate": 2.9849548645937812e-05, + "loss": 1.2712, + "theoretical_loss": 3.3345462650114177, + "tokens_seen": 3179085824 + }, + { + "epoch": 0.96, + "learning_rate": 2.9833500501504513e-05, + "loss": 1.297, + "theoretical_loss": 3.3345410864390237, + "tokens_seen": 3179151360 + }, + { + "epoch": 0.96, + "learning_rate": 2.9817452357071214e-05, + "loss": 1.2496, + "theoretical_loss": 3.3345359080032715, + "tokens_seen": 3179216896 + }, + { + "epoch": 0.96, + "learning_rate": 2.9801404212637918e-05, + "loss": 1.1907, + "theoretical_loss": 3.334530729704154, + "tokens_seen": 3179282432 + }, + { + "epoch": 0.96, + "learning_rate": 2.978535606820462e-05, + "loss": 1.2093, + "theoretical_loss": 3.3345255515416654, + "tokens_seen": 3179347968 + }, + { + "epoch": 0.96, + "learning_rate": 2.9769307923771316e-05, + "loss": 1.2586, + "theoretical_loss": 3.334520373515799, + "tokens_seen": 3179413504 + }, + { + "epoch": 0.96, + "learning_rate": 2.9753259779338016e-05, + "loss": 1.2367, + "theoretical_loss": 3.3345151956265493, + "tokens_seen": 3179479040 + }, + { + "epoch": 0.96, + "learning_rate": 2.9737211634904717e-05, + "loss": 1.3303, + "theoretical_loss": 3.3345100178739084, + "tokens_seen": 3179544576 + }, + { + "epoch": 0.96, + "learning_rate": 2.9721163490471414e-05, + "loss": 1.2625, + "theoretical_loss": 3.3345048402578708, + "tokens_seen": 3179610112 + }, + { + "epoch": 0.96, + "learning_rate": 2.9705115346038115e-05, + "loss": 1.1992, + "theoretical_loss": 3.33449966277843, + "tokens_seen": 3179675648 + }, + { + "epoch": 0.96, + "learning_rate": 2.9689067201604815e-05, + "loss": 1.2814, + "theoretical_loss": 3.334494485435579, + "tokens_seen": 3179741184 + }, + { + "epoch": 0.96, + "learning_rate": 2.9673019057171513e-05, + "loss": 1.3177, + "theoretical_loss": 3.3344893082293123, + "tokens_seen": 3179806720 + }, + { + "epoch": 0.96, + "learning_rate": 2.9656970912738213e-05, + "loss": 1.2821, + "theoretical_loss": 3.334484131159623, + "tokens_seen": 3179872256 + }, + { + "epoch": 0.96, + "learning_rate": 2.9640922768304917e-05, + "loss": 1.2357, + "theoretical_loss": 3.3344789542265048, + "tokens_seen": 3179937792 + }, + { + "epoch": 0.96, + "learning_rate": 2.9624874623871618e-05, + "loss": 1.1708, + "theoretical_loss": 3.334473777429951, + "tokens_seen": 3180003328 + }, + { + "epoch": 0.96, + "learning_rate": 2.960882647943832e-05, + "loss": 1.2739, + "theoretical_loss": 3.3344686007699553, + "tokens_seen": 3180068864 + }, + { + "epoch": 0.96, + "learning_rate": 2.9592778335005016e-05, + "loss": 1.2728, + "theoretical_loss": 3.3344634242465117, + "tokens_seen": 3180134400 + }, + { + "epoch": 0.96, + "learning_rate": 2.9576730190571716e-05, + "loss": 1.3774, + "theoretical_loss": 3.334458247859613, + "tokens_seen": 3180199936 + }, + { + "epoch": 0.96, + "learning_rate": 2.9560682046138417e-05, + "loss": 1.2862, + "theoretical_loss": 3.3344530716092535, + "tokens_seen": 3180265472 + }, + { + "epoch": 0.96, + "learning_rate": 2.9544633901705114e-05, + "loss": 1.3401, + "theoretical_loss": 3.334447895495426, + "tokens_seen": 3180331008 + }, + { + "epoch": 0.96, + "learning_rate": 2.9528585757271815e-05, + "loss": 1.3183, + "theoretical_loss": 3.334442719518125, + "tokens_seen": 3180396544 + }, + { + "epoch": 0.96, + "learning_rate": 2.9512537612838516e-05, + "loss": 1.1118, + "theoretical_loss": 3.334437543677344, + "tokens_seen": 3180462080 + }, + { + "epoch": 0.96, + "learning_rate": 2.949648946840522e-05, + "loss": 1.2515, + "theoretical_loss": 3.334432367973076, + "tokens_seen": 3180527616 + }, + { + "epoch": 0.96, + "learning_rate": 2.948044132397192e-05, + "loss": 1.3065, + "theoretical_loss": 3.3344271924053146, + "tokens_seen": 3180593152 + }, + { + "epoch": 0.96, + "learning_rate": 2.9464393179538617e-05, + "loss": 1.295, + "theoretical_loss": 3.334422016974054, + "tokens_seen": 3180658688 + }, + { + "epoch": 0.96, + "learning_rate": 2.9448345035105318e-05, + "loss": 1.3169, + "theoretical_loss": 3.334416841679287, + "tokens_seen": 3180724224 + }, + { + "epoch": 0.96, + "learning_rate": 2.943229689067202e-05, + "loss": 1.1871, + "theoretical_loss": 3.3344116665210084, + "tokens_seen": 3180789760 + }, + { + "epoch": 0.96, + "learning_rate": 2.9416248746238716e-05, + "loss": 1.2961, + "theoretical_loss": 3.33440649149921, + "tokens_seen": 3180855296 + }, + { + "epoch": 0.96, + "learning_rate": 2.9400200601805417e-05, + "loss": 1.2962, + "theoretical_loss": 3.3344013166138873, + "tokens_seen": 3180920832 + }, + { + "epoch": 0.96, + "learning_rate": 2.9384152457372117e-05, + "loss": 1.2861, + "theoretical_loss": 3.3343961418650325, + "tokens_seen": 3180986368 + }, + { + "epoch": 0.96, + "learning_rate": 2.9368104312938815e-05, + "loss": 1.2495, + "theoretical_loss": 3.3343909672526397, + "tokens_seen": 3181051904 + }, + { + "epoch": 0.96, + "learning_rate": 2.9352056168505522e-05, + "loss": 1.2952, + "theoretical_loss": 3.334385792776702, + "tokens_seen": 3181117440 + }, + { + "epoch": 0.96, + "learning_rate": 2.933600802407222e-05, + "loss": 1.2544, + "theoretical_loss": 3.3343806184372147, + "tokens_seen": 3181182976 + }, + { + "epoch": 0.96, + "learning_rate": 2.931995987963892e-05, + "loss": 1.1612, + "theoretical_loss": 3.334375444234169, + "tokens_seen": 3181248512 + }, + { + "epoch": 0.96, + "learning_rate": 2.930391173520562e-05, + "loss": 1.3868, + "theoretical_loss": 3.33437027016756, + "tokens_seen": 3181314048 + }, + { + "epoch": 0.96, + "learning_rate": 2.9287863590772318e-05, + "loss": 1.1915, + "theoretical_loss": 3.334365096237381, + "tokens_seen": 3181379584 + }, + { + "epoch": 0.96, + "learning_rate": 2.927181544633902e-05, + "loss": 1.3675, + "theoretical_loss": 3.3343599224436256, + "tokens_seen": 3181445120 + }, + { + "epoch": 0.96, + "learning_rate": 2.925576730190572e-05, + "loss": 1.2735, + "theoretical_loss": 3.3343547487862875, + "tokens_seen": 3181510656 + }, + { + "epoch": 0.96, + "learning_rate": 2.9239719157472416e-05, + "loss": 1.37, + "theoretical_loss": 3.3343495752653594, + "tokens_seen": 3181576192 + }, + { + "epoch": 0.96, + "learning_rate": 2.9223671013039117e-05, + "loss": 1.3446, + "theoretical_loss": 3.334344401880836, + "tokens_seen": 3181641728 + }, + { + "epoch": 0.96, + "learning_rate": 2.9207622868605817e-05, + "loss": 1.2487, + "theoretical_loss": 3.334339228632711, + "tokens_seen": 3181707264 + }, + { + "epoch": 0.96, + "objective/train/docs_used": 1156307, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.2672847509384155, + "objective/train/theoretical_loss": 3.3343340555209764, + "objective/train/tokens_used": 3202232800, + "theoretical_loss": 3.3343340555209764, + "tokens_seen": 3181772800 + }, + { + "epoch": 0.96, + "learning_rate": 2.919157472417252e-05, + "loss": 1.2094, + "theoretical_loss": 3.3343340555209764, + "tokens_seen": 3181772800 + }, + { + "epoch": 0.96, + "learning_rate": 2.9175526579739222e-05, + "loss": 1.3521, + "theoretical_loss": 3.334328882545628, + "tokens_seen": 3181838336 + }, + { + "epoch": 0.96, + "learning_rate": 2.915947843530592e-05, + "loss": 1.3186, + "theoretical_loss": 3.3343237097066574, + "tokens_seen": 3181903872 + }, + { + "epoch": 0.96, + "learning_rate": 2.914343029087262e-05, + "loss": 1.2364, + "theoretical_loss": 3.3343185370040596, + "tokens_seen": 3181969408 + }, + { + "epoch": 0.96, + "learning_rate": 2.912738214643932e-05, + "loss": 1.341, + "theoretical_loss": 3.3343133644378273, + "tokens_seen": 3182034944 + }, + { + "epoch": 0.96, + "learning_rate": 2.9111334002006018e-05, + "loss": 1.2543, + "theoretical_loss": 3.3343081920079545, + "tokens_seen": 3182100480 + }, + { + "epoch": 0.96, + "learning_rate": 2.909528585757272e-05, + "loss": 1.3286, + "theoretical_loss": 3.334303019714435, + "tokens_seen": 3182166016 + }, + { + "epoch": 0.96, + "learning_rate": 2.907923771313942e-05, + "loss": 1.3065, + "theoretical_loss": 3.334297847557262, + "tokens_seen": 3182231552 + }, + { + "epoch": 0.96, + "learning_rate": 2.9063189568706116e-05, + "loss": 1.389, + "theoretical_loss": 3.3342926755364295, + "tokens_seen": 3182297088 + }, + { + "epoch": 0.96, + "learning_rate": 2.9047141424272824e-05, + "loss": 1.2564, + "theoretical_loss": 3.3342875036519306, + "tokens_seen": 3182362624 + }, + { + "epoch": 0.96, + "learning_rate": 2.903109327983952e-05, + "loss": 1.2425, + "theoretical_loss": 3.3342823319037596, + "tokens_seen": 3182428160 + }, + { + "epoch": 0.96, + "learning_rate": 2.9015045135406222e-05, + "loss": 1.171, + "theoretical_loss": 3.334277160291909, + "tokens_seen": 3182493696 + }, + { + "epoch": 0.96, + "learning_rate": 2.8998996990972922e-05, + "loss": 1.2927, + "theoretical_loss": 3.3342719888163734, + "tokens_seen": 3182559232 + }, + { + "epoch": 0.96, + "learning_rate": 2.898294884653962e-05, + "loss": 1.3446, + "theoretical_loss": 3.334266817477146, + "tokens_seen": 3182624768 + }, + { + "epoch": 0.96, + "learning_rate": 2.896690070210632e-05, + "loss": 1.2244, + "theoretical_loss": 3.3342616462742205, + "tokens_seen": 3182690304 + }, + { + "epoch": 0.96, + "learning_rate": 2.895085255767302e-05, + "loss": 1.3249, + "theoretical_loss": 3.3342564752075905, + "tokens_seen": 3182755840 + }, + { + "epoch": 0.96, + "learning_rate": 2.8934804413239718e-05, + "loss": 1.2931, + "theoretical_loss": 3.3342513042772493, + "tokens_seen": 3182821376 + }, + { + "epoch": 0.96, + "learning_rate": 2.891875626880642e-05, + "loss": 1.2295, + "theoretical_loss": 3.334246133483191, + "tokens_seen": 3182886912 + }, + { + "epoch": 0.96, + "learning_rate": 2.8902708124373123e-05, + "loss": 1.3214, + "theoretical_loss": 3.334240962825409, + "tokens_seen": 3182952448 + }, + { + "epoch": 0.96, + "learning_rate": 2.8886659979939823e-05, + "loss": 1.1204, + "theoretical_loss": 3.3342357923038968, + "tokens_seen": 3183017984 + }, + { + "epoch": 0.96, + "learning_rate": 2.8870611835506524e-05, + "loss": 1.331, + "theoretical_loss": 3.334230621918648, + "tokens_seen": 3183083520 + }, + { + "epoch": 0.96, + "learning_rate": 2.885456369107322e-05, + "loss": 1.2219, + "theoretical_loss": 3.3342254516696563, + "tokens_seen": 3183149056 + }, + { + "epoch": 0.96, + "learning_rate": 2.8838515546639922e-05, + "loss": 1.2931, + "theoretical_loss": 3.3342202815569153, + "tokens_seen": 3183214592 + }, + { + "epoch": 0.96, + "learning_rate": 2.8822467402206623e-05, + "loss": 1.2642, + "theoretical_loss": 3.334215111580418, + "tokens_seen": 3183280128 + }, + { + "epoch": 0.96, + "learning_rate": 2.880641925777332e-05, + "loss": 1.3169, + "theoretical_loss": 3.3342099417401596, + "tokens_seen": 3183345664 + }, + { + "epoch": 0.96, + "learning_rate": 2.879037111334002e-05, + "loss": 1.3135, + "theoretical_loss": 3.334204772036132, + "tokens_seen": 3183411200 + }, + { + "epoch": 0.96, + "learning_rate": 2.877432296890672e-05, + "loss": 1.3516, + "theoretical_loss": 3.33419960246833, + "tokens_seen": 3183476736 + }, + { + "epoch": 0.96, + "learning_rate": 2.875827482447342e-05, + "loss": 1.3811, + "theoretical_loss": 3.334194433036746, + "tokens_seen": 3183542272 + }, + { + "epoch": 0.96, + "learning_rate": 2.8742226680040126e-05, + "loss": 1.2587, + "theoretical_loss": 3.334189263741375, + "tokens_seen": 3183607808 + }, + { + "epoch": 0.96, + "learning_rate": 2.8726178535606823e-05, + "loss": 1.3607, + "theoretical_loss": 3.3341840945822097, + "tokens_seen": 3183673344 + }, + { + "epoch": 0.96, + "learning_rate": 2.8710130391173524e-05, + "loss": 1.2642, + "theoretical_loss": 3.334178925559244, + "tokens_seen": 3183738880 + }, + { + "epoch": 0.96, + "learning_rate": 2.8694082246740224e-05, + "loss": 1.3357, + "theoretical_loss": 3.334173756672471, + "tokens_seen": 3183804416 + }, + { + "epoch": 0.96, + "learning_rate": 2.867803410230692e-05, + "loss": 1.2383, + "theoretical_loss": 3.3341685879218854, + "tokens_seen": 3183869952 + }, + { + "epoch": 0.96, + "learning_rate": 2.8661985957873622e-05, + "loss": 1.2065, + "theoretical_loss": 3.3341634193074796, + "tokens_seen": 3183935488 + }, + { + "epoch": 0.96, + "learning_rate": 2.8645937813440323e-05, + "loss": 1.2325, + "theoretical_loss": 3.334158250829248, + "tokens_seen": 3184001024 + }, + { + "epoch": 0.96, + "learning_rate": 2.862988966900702e-05, + "loss": 1.3136, + "theoretical_loss": 3.334153082487184, + "tokens_seen": 3184066560 + }, + { + "epoch": 0.96, + "learning_rate": 2.861384152457372e-05, + "loss": 1.2803, + "theoretical_loss": 3.3341479142812807, + "tokens_seen": 3184132096 + }, + { + "epoch": 0.96, + "learning_rate": 2.8597793380140425e-05, + "loss": 1.2661, + "theoretical_loss": 3.3341427462115325, + "tokens_seen": 3184197632 + }, + { + "epoch": 0.96, + "learning_rate": 2.8581745235707125e-05, + "loss": 1.3322, + "theoretical_loss": 3.334137578277933, + "tokens_seen": 3184263168 + }, + { + "epoch": 0.96, + "learning_rate": 2.8565697091273823e-05, + "loss": 1.2385, + "theoretical_loss": 3.3341324104804753, + "tokens_seen": 3184328704 + }, + { + "epoch": 0.96, + "learning_rate": 2.8549648946840523e-05, + "loss": 1.3261, + "theoretical_loss": 3.334127242819153, + "tokens_seen": 3184394240 + }, + { + "epoch": 0.96, + "learning_rate": 2.8533600802407224e-05, + "loss": 1.2814, + "theoretical_loss": 3.33412207529396, + "tokens_seen": 3184459776 + }, + { + "epoch": 0.97, + "learning_rate": 2.851755265797392e-05, + "loss": 1.2569, + "theoretical_loss": 3.3341169079048902, + "tokens_seen": 3184525312 + }, + { + "epoch": 0.97, + "learning_rate": 2.850150451354062e-05, + "loss": 1.2848, + "theoretical_loss": 3.3341117406519367, + "tokens_seen": 3184590848 + }, + { + "epoch": 0.97, + "learning_rate": 2.8485456369107322e-05, + "loss": 1.1718, + "theoretical_loss": 3.334106573535093, + "tokens_seen": 3184656384 + }, + { + "epoch": 0.97, + "learning_rate": 2.846940822467402e-05, + "loss": 1.2943, + "theoretical_loss": 3.3341014065543533, + "tokens_seen": 3184721920 + }, + { + "epoch": 0.97, + "learning_rate": 2.8453360080240727e-05, + "loss": 1.3841, + "theoretical_loss": 3.3340962397097105, + "tokens_seen": 3184787456 + }, + { + "epoch": 0.97, + "learning_rate": 2.8437311935807424e-05, + "loss": 1.1851, + "theoretical_loss": 3.334091073001159, + "tokens_seen": 3184852992 + }, + { + "epoch": 0.97, + "learning_rate": 2.8421263791374125e-05, + "loss": 1.2349, + "theoretical_loss": 3.334085906428692, + "tokens_seen": 3184918528 + }, + { + "epoch": 0.97, + "learning_rate": 2.8405215646940825e-05, + "loss": 1.2432, + "theoretical_loss": 3.334080739992303, + "tokens_seen": 3184984064 + }, + { + "epoch": 0.97, + "objective/train/docs_used": 1157476, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.1351220607757568, + "objective/train/theoretical_loss": 3.334075573691986, + "objective/train/tokens_used": 3205509600, + "theoretical_loss": 3.334075573691986, + "tokens_seen": 3185049600 + }, + { + "epoch": 0.97, + "learning_rate": 2.8389167502507523e-05, + "loss": 1.2341, + "theoretical_loss": 3.334075573691986, + "tokens_seen": 3185049600 + }, + { + "epoch": 0.97, + "learning_rate": 2.8373119358074223e-05, + "loss": 1.3053, + "theoretical_loss": 3.334070407527734, + "tokens_seen": 3185115136 + }, + { + "epoch": 0.97, + "learning_rate": 2.8357071213640924e-05, + "loss": 1.2095, + "theoretical_loss": 3.334065241499541, + "tokens_seen": 3185180672 + }, + { + "epoch": 0.97, + "learning_rate": 2.834102306920762e-05, + "loss": 1.3699, + "theoretical_loss": 3.3340600756074013, + "tokens_seen": 3185246208 + }, + { + "epoch": 0.97, + "learning_rate": 2.8324974924774322e-05, + "loss": 1.3627, + "theoretical_loss": 3.3340549098513073, + "tokens_seen": 3185311744 + }, + { + "epoch": 0.97, + "learning_rate": 2.8308926780341023e-05, + "loss": 1.3391, + "theoretical_loss": 3.334049744231253, + "tokens_seen": 3185377280 + }, + { + "epoch": 0.97, + "learning_rate": 2.8292878635907727e-05, + "loss": 1.3048, + "theoretical_loss": 3.3340445787472324, + "tokens_seen": 3185442816 + }, + { + "epoch": 0.97, + "learning_rate": 2.8276830491474427e-05, + "loss": 1.3113, + "theoretical_loss": 3.3340394133992386, + "tokens_seen": 3185508352 + }, + { + "epoch": 0.97, + "learning_rate": 2.8260782347041124e-05, + "loss": 1.3013, + "theoretical_loss": 3.334034248187266, + "tokens_seen": 3185573888 + }, + { + "epoch": 0.97, + "learning_rate": 2.8244734202607825e-05, + "loss": 1.2532, + "theoretical_loss": 3.3340290831113073, + "tokens_seen": 3185639424 + }, + { + "epoch": 0.97, + "learning_rate": 2.8228686058174526e-05, + "loss": 1.2162, + "theoretical_loss": 3.334023918171357, + "tokens_seen": 3185704960 + }, + { + "epoch": 0.97, + "learning_rate": 2.8212637913741223e-05, + "loss": 1.2124, + "theoretical_loss": 3.3340187533674075, + "tokens_seen": 3185770496 + }, + { + "epoch": 0.97, + "learning_rate": 2.8196589769307924e-05, + "loss": 1.2504, + "theoretical_loss": 3.3340135886994537, + "tokens_seen": 3185836032 + }, + { + "epoch": 0.97, + "learning_rate": 2.8180541624874624e-05, + "loss": 1.3266, + "theoretical_loss": 3.3340084241674885, + "tokens_seen": 3185901568 + }, + { + "epoch": 0.97, + "learning_rate": 2.816449348044132e-05, + "loss": 1.2981, + "theoretical_loss": 3.334003259771506, + "tokens_seen": 3185967104 + }, + { + "epoch": 0.97, + "learning_rate": 2.814844533600803e-05, + "loss": 1.3091, + "theoretical_loss": 3.3339980955114994, + "tokens_seen": 3186032640 + }, + { + "epoch": 0.97, + "learning_rate": 2.8132397191574726e-05, + "loss": 1.2349, + "theoretical_loss": 3.3339929313874626, + "tokens_seen": 3186098176 + }, + { + "epoch": 0.97, + "learning_rate": 2.8116349047141427e-05, + "loss": 1.2159, + "theoretical_loss": 3.333987767399389, + "tokens_seen": 3186163712 + }, + { + "epoch": 0.97, + "learning_rate": 2.8100300902708127e-05, + "loss": 1.4096, + "theoretical_loss": 3.3339826035472724, + "tokens_seen": 3186229248 + }, + { + "epoch": 0.97, + "learning_rate": 2.8084252758274825e-05, + "loss": 1.2171, + "theoretical_loss": 3.333977439831106, + "tokens_seen": 3186294784 + }, + { + "epoch": 0.97, + "learning_rate": 2.8068204613841525e-05, + "loss": 1.3605, + "theoretical_loss": 3.3339722762508845, + "tokens_seen": 3186360320 + }, + { + "epoch": 0.97, + "learning_rate": 2.8052156469408226e-05, + "loss": 1.261, + "theoretical_loss": 3.3339671128066004, + "tokens_seen": 3186425856 + }, + { + "epoch": 0.97, + "learning_rate": 2.8036108324974923e-05, + "loss": 1.3986, + "theoretical_loss": 3.3339619494982475, + "tokens_seen": 3186491392 + }, + { + "epoch": 0.97, + "learning_rate": 2.8020060180541624e-05, + "loss": 1.2832, + "theoretical_loss": 3.3339567863258197, + "tokens_seen": 3186556928 + }, + { + "epoch": 0.97, + "learning_rate": 2.8004012036108324e-05, + "loss": 1.2722, + "theoretical_loss": 3.3339516232893107, + "tokens_seen": 3186622464 + }, + { + "epoch": 0.97, + "learning_rate": 2.798796389167503e-05, + "loss": 1.3938, + "theoretical_loss": 3.3339464603887143, + "tokens_seen": 3186688000 + }, + { + "epoch": 0.97, + "learning_rate": 2.797191574724173e-05, + "loss": 1.23, + "theoretical_loss": 3.3339412976240235, + "tokens_seen": 3186753536 + }, + { + "epoch": 0.97, + "learning_rate": 2.7955867602808426e-05, + "loss": 1.2797, + "theoretical_loss": 3.3339361349952323, + "tokens_seen": 3186819072 + }, + { + "epoch": 0.97, + "learning_rate": 2.7939819458375127e-05, + "loss": 1.2482, + "theoretical_loss": 3.3339309725023343, + "tokens_seen": 3186884608 + }, + { + "epoch": 0.97, + "learning_rate": 2.7923771313941828e-05, + "loss": 1.295, + "theoretical_loss": 3.333925810145323, + "tokens_seen": 3186950144 + }, + { + "epoch": 0.97, + "learning_rate": 2.7907723169508525e-05, + "loss": 1.2629, + "theoretical_loss": 3.333920647924192, + "tokens_seen": 3187015680 + }, + { + "epoch": 0.97, + "learning_rate": 2.7891675025075225e-05, + "loss": 1.2884, + "theoretical_loss": 3.3339154858389355, + "tokens_seen": 3187081216 + }, + { + "epoch": 0.97, + "learning_rate": 2.7875626880641926e-05, + "loss": 1.2429, + "theoretical_loss": 3.3339103238895467, + "tokens_seen": 3187146752 + }, + { + "epoch": 0.97, + "learning_rate": 2.7859578736208623e-05, + "loss": 1.2115, + "theoretical_loss": 3.333905162076019, + "tokens_seen": 3187212288 + }, + { + "epoch": 0.97, + "learning_rate": 2.784353059177533e-05, + "loss": 1.2372, + "theoretical_loss": 3.333900000398346, + "tokens_seen": 3187277824 + }, + { + "epoch": 0.97, + "learning_rate": 2.7827482447342028e-05, + "loss": 1.2165, + "theoretical_loss": 3.333894838856522, + "tokens_seen": 3187343360 + }, + { + "epoch": 0.97, + "learning_rate": 2.781143430290873e-05, + "loss": 1.2491, + "theoretical_loss": 3.33388967745054, + "tokens_seen": 3187408896 + }, + { + "epoch": 0.97, + "learning_rate": 2.779538615847543e-05, + "loss": 1.2014, + "theoretical_loss": 3.333884516180394, + "tokens_seen": 3187474432 + }, + { + "epoch": 0.97, + "learning_rate": 2.7779338014042127e-05, + "loss": 1.3245, + "theoretical_loss": 3.3338793550460775, + "tokens_seen": 3187539968 + }, + { + "epoch": 0.97, + "learning_rate": 2.7763289869608827e-05, + "loss": 1.2306, + "theoretical_loss": 3.333874194047584, + "tokens_seen": 3187605504 + }, + { + "epoch": 0.97, + "learning_rate": 2.7747241725175528e-05, + "loss": 1.2848, + "theoretical_loss": 3.3338690331849072, + "tokens_seen": 3187671040 + }, + { + "epoch": 0.97, + "learning_rate": 2.7731193580742225e-05, + "loss": 1.2833, + "theoretical_loss": 3.3338638724580414, + "tokens_seen": 3187736576 + }, + { + "epoch": 0.97, + "learning_rate": 2.7715145436308926e-05, + "loss": 1.1912, + "theoretical_loss": 3.333858711866979, + "tokens_seen": 3187802112 + }, + { + "epoch": 0.97, + "learning_rate": 2.769909729187563e-05, + "loss": 1.2817, + "theoretical_loss": 3.3338535514117145, + "tokens_seen": 3187867648 + }, + { + "epoch": 0.97, + "learning_rate": 2.768304914744233e-05, + "loss": 1.3379, + "theoretical_loss": 3.333848391092241, + "tokens_seen": 3187933184 + }, + { + "epoch": 0.97, + "learning_rate": 2.766700100300903e-05, + "loss": 1.271, + "theoretical_loss": 3.3338432309085526, + "tokens_seen": 3187998720 + }, + { + "epoch": 0.97, + "learning_rate": 2.7650952858575728e-05, + "loss": 1.2286, + "theoretical_loss": 3.3338380708606428, + "tokens_seen": 3188064256 + }, + { + "epoch": 0.97, + "learning_rate": 2.763490471414243e-05, + "loss": 1.281, + "theoretical_loss": 3.333832910948505, + "tokens_seen": 3188129792 + }, + { + "epoch": 0.97, + "learning_rate": 2.761885656970913e-05, + "loss": 1.2034, + "theoretical_loss": 3.3338277511721337, + "tokens_seen": 3188195328 + }, + { + "epoch": 0.97, + "learning_rate": 2.7602808425275827e-05, + "loss": 1.2052, + "theoretical_loss": 3.333822591531521, + "tokens_seen": 3188260864 + }, + { + "epoch": 0.97, + "objective/train/docs_used": 1158745, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.2742055654525757, + "objective/train/theoretical_loss": 3.3338174320266623, + "objective/train/tokens_used": 3208786400, + "theoretical_loss": 3.3338174320266623, + "tokens_seen": 3188326400 + }, + { + "epoch": 0.97, + "learning_rate": 2.7586760280842527e-05, + "loss": 1.3044, + "theoretical_loss": 3.3338174320266623, + "tokens_seen": 3188326400 + }, + { + "epoch": 0.97, + "learning_rate": 2.7570712136409228e-05, + "loss": 1.2792, + "theoretical_loss": 3.3338122726575494, + "tokens_seen": 3188391936 + }, + { + "epoch": 0.97, + "learning_rate": 2.7554663991975925e-05, + "loss": 1.3404, + "theoretical_loss": 3.3338071134241773, + "tokens_seen": 3188457472 + }, + { + "epoch": 0.97, + "learning_rate": 2.7538615847542633e-05, + "loss": 1.3076, + "theoretical_loss": 3.3338019543265394, + "tokens_seen": 3188523008 + }, + { + "epoch": 0.97, + "learning_rate": 2.752256770310933e-05, + "loss": 1.3167, + "theoretical_loss": 3.333796795364629, + "tokens_seen": 3188588544 + }, + { + "epoch": 0.97, + "learning_rate": 2.750651955867603e-05, + "loss": 1.412, + "theoretical_loss": 3.33379163653844, + "tokens_seen": 3188654080 + }, + { + "epoch": 0.97, + "learning_rate": 2.749047141424273e-05, + "loss": 1.2487, + "theoretical_loss": 3.333786477847966, + "tokens_seen": 3188719616 + }, + { + "epoch": 0.97, + "learning_rate": 2.747442326980943e-05, + "loss": 1.1634, + "theoretical_loss": 3.3337813192932004, + "tokens_seen": 3188785152 + }, + { + "epoch": 0.97, + "learning_rate": 2.745837512537613e-05, + "loss": 1.3215, + "theoretical_loss": 3.333776160874137, + "tokens_seen": 3188850688 + }, + { + "epoch": 0.97, + "learning_rate": 2.744232698094283e-05, + "loss": 1.3198, + "theoretical_loss": 3.33377100259077, + "tokens_seen": 3188916224 + }, + { + "epoch": 0.97, + "learning_rate": 2.7426278836509527e-05, + "loss": 1.2104, + "theoretical_loss": 3.333765844443092, + "tokens_seen": 3188981760 + }, + { + "epoch": 0.97, + "learning_rate": 2.7410230692076228e-05, + "loss": 1.2311, + "theoretical_loss": 3.333760686431097, + "tokens_seen": 3189047296 + }, + { + "epoch": 0.97, + "learning_rate": 2.739418254764293e-05, + "loss": 1.2224, + "theoretical_loss": 3.333755528554779, + "tokens_seen": 3189112832 + }, + { + "epoch": 0.97, + "learning_rate": 2.7378134403209632e-05, + "loss": 1.203, + "theoretical_loss": 3.3337503708141316, + "tokens_seen": 3189178368 + }, + { + "epoch": 0.97, + "learning_rate": 2.7362086258776333e-05, + "loss": 1.212, + "theoretical_loss": 3.333745213209148, + "tokens_seen": 3189243904 + }, + { + "epoch": 0.97, + "learning_rate": 2.734603811434303e-05, + "loss": 1.2421, + "theoretical_loss": 3.3337400557398222, + "tokens_seen": 3189309440 + }, + { + "epoch": 0.97, + "learning_rate": 2.732998996990973e-05, + "loss": 1.2122, + "theoretical_loss": 3.333734898406148, + "tokens_seen": 3189374976 + }, + { + "epoch": 0.97, + "learning_rate": 2.731394182547643e-05, + "loss": 1.2399, + "theoretical_loss": 3.333729741208119, + "tokens_seen": 3189440512 + }, + { + "epoch": 0.97, + "learning_rate": 2.729789368104313e-05, + "loss": 1.2426, + "theoretical_loss": 3.333724584145728, + "tokens_seen": 3189506048 + }, + { + "epoch": 0.97, + "learning_rate": 2.728184553660983e-05, + "loss": 1.2365, + "theoretical_loss": 3.3337194272189694, + "tokens_seen": 3189571584 + }, + { + "epoch": 0.97, + "learning_rate": 2.726579739217653e-05, + "loss": 1.2618, + "theoretical_loss": 3.333714270427837, + "tokens_seen": 3189637120 + }, + { + "epoch": 0.97, + "learning_rate": 2.7249749247743234e-05, + "loss": 1.296, + "theoretical_loss": 3.3337091137723243, + "tokens_seen": 3189702656 + }, + { + "epoch": 0.97, + "learning_rate": 2.7233701103309935e-05, + "loss": 1.2178, + "theoretical_loss": 3.3337039572524247, + "tokens_seen": 3189768192 + }, + { + "epoch": 0.97, + "learning_rate": 2.7217652958876632e-05, + "loss": 1.2187, + "theoretical_loss": 3.333698800868132, + "tokens_seen": 3189833728 + }, + { + "epoch": 0.97, + "learning_rate": 2.7201604814443332e-05, + "loss": 1.2358, + "theoretical_loss": 3.33369364461944, + "tokens_seen": 3189899264 + }, + { + "epoch": 0.97, + "learning_rate": 2.7185556670010033e-05, + "loss": 1.2031, + "theoretical_loss": 3.3336884885063416, + "tokens_seen": 3189964800 + }, + { + "epoch": 0.97, + "learning_rate": 2.716950852557673e-05, + "loss": 1.2323, + "theoretical_loss": 3.3336833325288318, + "tokens_seen": 3190030336 + }, + { + "epoch": 0.97, + "learning_rate": 2.715346038114343e-05, + "loss": 1.1975, + "theoretical_loss": 3.3336781766869032, + "tokens_seen": 3190095872 + }, + { + "epoch": 0.97, + "learning_rate": 2.713741223671013e-05, + "loss": 1.2548, + "theoretical_loss": 3.33367302098055, + "tokens_seen": 3190161408 + }, + { + "epoch": 0.97, + "learning_rate": 2.712136409227683e-05, + "loss": 1.3309, + "theoretical_loss": 3.333667865409765, + "tokens_seen": 3190226944 + }, + { + "epoch": 0.97, + "learning_rate": 2.710531594784353e-05, + "loss": 1.2518, + "theoretical_loss": 3.3336627099745426, + "tokens_seen": 3190292480 + }, + { + "epoch": 0.97, + "learning_rate": 2.7089267803410233e-05, + "loss": 1.2457, + "theoretical_loss": 3.3336575546748763, + "tokens_seen": 3190358016 + }, + { + "epoch": 0.97, + "learning_rate": 2.7073219658976934e-05, + "loss": 1.1676, + "theoretical_loss": 3.33365239951076, + "tokens_seen": 3190423552 + }, + { + "epoch": 0.97, + "learning_rate": 2.7057171514543635e-05, + "loss": 1.2842, + "theoretical_loss": 3.333647244482187, + "tokens_seen": 3190489088 + }, + { + "epoch": 0.97, + "learning_rate": 2.7041123370110332e-05, + "loss": 1.2764, + "theoretical_loss": 3.333642089589151, + "tokens_seen": 3190554624 + }, + { + "epoch": 0.97, + "learning_rate": 2.7025075225677033e-05, + "loss": 1.222, + "theoretical_loss": 3.3336369348316457, + "tokens_seen": 3190620160 + }, + { + "epoch": 0.97, + "learning_rate": 2.7009027081243733e-05, + "loss": 1.3769, + "theoretical_loss": 3.3336317802096644, + "tokens_seen": 3190685696 + }, + { + "epoch": 0.97, + "learning_rate": 2.699297893681043e-05, + "loss": 1.2362, + "theoretical_loss": 3.3336266257232015, + "tokens_seen": 3190751232 + }, + { + "epoch": 0.97, + "learning_rate": 2.697693079237713e-05, + "loss": 1.2508, + "theoretical_loss": 3.3336214713722505, + "tokens_seen": 3190816768 + }, + { + "epoch": 0.97, + "learning_rate": 2.6960882647943832e-05, + "loss": 1.3239, + "theoretical_loss": 3.333616317156804, + "tokens_seen": 3190882304 + }, + { + "epoch": 0.97, + "learning_rate": 2.6944834503510536e-05, + "loss": 1.2332, + "theoretical_loss": 3.3336111630768572, + "tokens_seen": 3190947840 + }, + { + "epoch": 0.97, + "learning_rate": 2.6928786359077236e-05, + "loss": 1.242, + "theoretical_loss": 3.333606009132403, + "tokens_seen": 3191013376 + }, + { + "epoch": 0.97, + "learning_rate": 2.6912738214643934e-05, + "loss": 1.2779, + "theoretical_loss": 3.333600855323435, + "tokens_seen": 3191078912 + }, + { + "epoch": 0.97, + "learning_rate": 2.6896690070210634e-05, + "loss": 1.379, + "theoretical_loss": 3.3335957016499465, + "tokens_seen": 3191144448 + }, + { + "epoch": 0.97, + "learning_rate": 2.6880641925777335e-05, + "loss": 1.2599, + "theoretical_loss": 3.333590548111932, + "tokens_seen": 3191209984 + }, + { + "epoch": 0.97, + "learning_rate": 2.6864593781344032e-05, + "loss": 1.1843, + "theoretical_loss": 3.3335853947093845, + "tokens_seen": 3191275520 + }, + { + "epoch": 0.97, + "learning_rate": 2.6848545636910733e-05, + "loss": 1.2856, + "theoretical_loss": 3.3335802414422977, + "tokens_seen": 3191341056 + }, + { + "epoch": 0.97, + "learning_rate": 2.6832497492477433e-05, + "loss": 1.2789, + "theoretical_loss": 3.333575088310666, + "tokens_seen": 3191406592 + }, + { + "epoch": 0.97, + "learning_rate": 2.681644934804413e-05, + "loss": 1.3227, + "theoretical_loss": 3.333569935314482, + "tokens_seen": 3191472128 + }, + { + "epoch": 0.97, + "learning_rate": 2.6800401203610838e-05, + "loss": 1.2287, + "theoretical_loss": 3.3335647824537404, + "tokens_seen": 3191537664 + }, + { + "epoch": 0.97, + "objective/train/docs_used": 1159894, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.2222487926483154, + "objective/train/theoretical_loss": 3.333559629728434, + "objective/train/tokens_used": 3212063200, + "theoretical_loss": 3.333559629728434, + "tokens_seen": 3191603200 + }, + { + "epoch": 0.97, + "learning_rate": 2.6784353059177535e-05, + "loss": 1.2546, + "theoretical_loss": 3.333559629728434, + "tokens_seen": 3191603200 + }, + { + "epoch": 0.97, + "learning_rate": 2.6768304914744236e-05, + "loss": 1.3386, + "theoretical_loss": 3.333554477138557, + "tokens_seen": 3191668736 + }, + { + "epoch": 0.97, + "learning_rate": 2.6752256770310937e-05, + "loss": 1.1693, + "theoretical_loss": 3.333549324684103, + "tokens_seen": 3191734272 + }, + { + "epoch": 0.97, + "learning_rate": 2.6736208625877634e-05, + "loss": 1.2095, + "theoretical_loss": 3.3335441723650647, + "tokens_seen": 3191799808 + }, + { + "epoch": 0.97, + "learning_rate": 2.6720160481444335e-05, + "loss": 1.1953, + "theoretical_loss": 3.3335390201814374, + "tokens_seen": 3191865344 + }, + { + "epoch": 0.97, + "learning_rate": 2.6704112337011035e-05, + "loss": 1.1654, + "theoretical_loss": 3.3335338681332134, + "tokens_seen": 3191930880 + }, + { + "epoch": 0.97, + "learning_rate": 2.6688064192577732e-05, + "loss": 1.2136, + "theoretical_loss": 3.3335287162203873, + "tokens_seen": 3191996416 + }, + { + "epoch": 0.97, + "learning_rate": 2.6672016048144433e-05, + "loss": 1.2393, + "theoretical_loss": 3.3335235644429524, + "tokens_seen": 3192061952 + }, + { + "epoch": 0.97, + "learning_rate": 2.6655967903711134e-05, + "loss": 1.3506, + "theoretical_loss": 3.333518412800902, + "tokens_seen": 3192127488 + }, + { + "epoch": 0.97, + "learning_rate": 2.6639919759277838e-05, + "loss": 1.2416, + "theoretical_loss": 3.33351326129423, + "tokens_seen": 3192193024 + }, + { + "epoch": 0.97, + "learning_rate": 2.6623871614844535e-05, + "loss": 1.2298, + "theoretical_loss": 3.3335081099229305, + "tokens_seen": 3192258560 + }, + { + "epoch": 0.97, + "learning_rate": 2.6607823470411236e-05, + "loss": 1.316, + "theoretical_loss": 3.3335029586869966, + "tokens_seen": 3192324096 + }, + { + "epoch": 0.97, + "learning_rate": 2.6591775325977936e-05, + "loss": 1.2108, + "theoretical_loss": 3.3334978075864226, + "tokens_seen": 3192389632 + }, + { + "epoch": 0.97, + "learning_rate": 2.6575727181544633e-05, + "loss": 1.2352, + "theoretical_loss": 3.333492656621201, + "tokens_seen": 3192455168 + }, + { + "epoch": 0.97, + "learning_rate": 2.6559679037111334e-05, + "loss": 1.2955, + "theoretical_loss": 3.333487505791327, + "tokens_seen": 3192520704 + }, + { + "epoch": 0.97, + "learning_rate": 2.6543630892678035e-05, + "loss": 1.2433, + "theoretical_loss": 3.333482355096793, + "tokens_seen": 3192586240 + }, + { + "epoch": 0.97, + "learning_rate": 2.6527582748244732e-05, + "loss": 1.2092, + "theoretical_loss": 3.3334772045375933, + "tokens_seen": 3192651776 + }, + { + "epoch": 0.97, + "learning_rate": 2.6511534603811433e-05, + "loss": 1.26, + "theoretical_loss": 3.333472054113721, + "tokens_seen": 3192717312 + }, + { + "epoch": 0.97, + "learning_rate": 2.6495486459378137e-05, + "loss": 1.2901, + "theoretical_loss": 3.3334669038251707, + "tokens_seen": 3192782848 + }, + { + "epoch": 0.97, + "learning_rate": 2.6479438314944837e-05, + "loss": 1.3249, + "theoretical_loss": 3.3334617536719353, + "tokens_seen": 3192848384 + }, + { + "epoch": 0.97, + "learning_rate": 2.6463390170511538e-05, + "loss": 1.2732, + "theoretical_loss": 3.3334566036540085, + "tokens_seen": 3192913920 + }, + { + "epoch": 0.97, + "learning_rate": 2.6447342026078235e-05, + "loss": 1.2611, + "theoretical_loss": 3.333451453771384, + "tokens_seen": 3192979456 + }, + { + "epoch": 0.97, + "learning_rate": 2.6431293881644936e-05, + "loss": 1.1605, + "theoretical_loss": 3.3334463040240565, + "tokens_seen": 3193044992 + }, + { + "epoch": 0.97, + "learning_rate": 2.6415245737211636e-05, + "loss": 1.3901, + "theoretical_loss": 3.3334411544120184, + "tokens_seen": 3193110528 + }, + { + "epoch": 0.97, + "learning_rate": 2.6399197592778334e-05, + "loss": 1.283, + "theoretical_loss": 3.3334360049352636, + "tokens_seen": 3193176064 + }, + { + "epoch": 0.97, + "learning_rate": 2.6383149448345034e-05, + "loss": 1.3724, + "theoretical_loss": 3.333430855593786, + "tokens_seen": 3193241600 + }, + { + "epoch": 0.97, + "learning_rate": 2.6367101303911735e-05, + "loss": 1.2516, + "theoretical_loss": 3.333425706387579, + "tokens_seen": 3193307136 + }, + { + "epoch": 0.97, + "learning_rate": 2.635105315947844e-05, + "loss": 1.1823, + "theoretical_loss": 3.333420557316637, + "tokens_seen": 3193372672 + }, + { + "epoch": 0.97, + "learning_rate": 2.633500501504514e-05, + "loss": 1.2631, + "theoretical_loss": 3.3334154083809526, + "tokens_seen": 3193438208 + }, + { + "epoch": 0.97, + "learning_rate": 2.6318956870611837e-05, + "loss": 1.3453, + "theoretical_loss": 3.3334102595805204, + "tokens_seen": 3193503744 + }, + { + "epoch": 0.97, + "learning_rate": 2.6302908726178537e-05, + "loss": 1.2044, + "theoretical_loss": 3.3334051109153338, + "tokens_seen": 3193569280 + }, + { + "epoch": 0.97, + "learning_rate": 2.6286860581745238e-05, + "loss": 1.2793, + "theoretical_loss": 3.3333999623853865, + "tokens_seen": 3193634816 + }, + { + "epoch": 0.97, + "learning_rate": 2.6270812437311935e-05, + "loss": 1.323, + "theoretical_loss": 3.3333948139906715, + "tokens_seen": 3193700352 + }, + { + "epoch": 0.97, + "learning_rate": 2.6254764292878636e-05, + "loss": 1.2198, + "theoretical_loss": 3.3333896657311834, + "tokens_seen": 3193765888 + }, + { + "epoch": 0.97, + "learning_rate": 2.6238716148445337e-05, + "loss": 1.2397, + "theoretical_loss": 3.3333845176069152, + "tokens_seen": 3193831424 + }, + { + "epoch": 0.97, + "learning_rate": 2.6222668004012034e-05, + "loss": 1.2499, + "theoretical_loss": 3.333379369617861, + "tokens_seen": 3193896960 + }, + { + "epoch": 0.97, + "learning_rate": 2.6206619859578735e-05, + "loss": 1.3268, + "theoretical_loss": 3.3333742217640143, + "tokens_seen": 3193962496 + }, + { + "epoch": 0.97, + "learning_rate": 2.619057171514544e-05, + "loss": 1.3832, + "theoretical_loss": 3.333369074045369, + "tokens_seen": 3194028032 + }, + { + "epoch": 0.97, + "learning_rate": 2.617452357071214e-05, + "loss": 1.365, + "theoretical_loss": 3.3333639264619186, + "tokens_seen": 3194093568 + }, + { + "epoch": 0.97, + "learning_rate": 2.615847542627884e-05, + "loss": 1.2781, + "theoretical_loss": 3.3333587790136567, + "tokens_seen": 3194159104 + }, + { + "epoch": 0.97, + "learning_rate": 2.6142427281845537e-05, + "loss": 1.3317, + "theoretical_loss": 3.333353631700577, + "tokens_seen": 3194224640 + }, + { + "epoch": 0.97, + "learning_rate": 2.6126379137412238e-05, + "loss": 1.3227, + "theoretical_loss": 3.333348484522673, + "tokens_seen": 3194290176 + }, + { + "epoch": 0.97, + "learning_rate": 2.611033099297894e-05, + "loss": 1.2753, + "theoretical_loss": 3.333343337479939, + "tokens_seen": 3194355712 + }, + { + "epoch": 0.97, + "learning_rate": 2.6094282848545636e-05, + "loss": 1.2176, + "theoretical_loss": 3.333338190572368, + "tokens_seen": 3194421248 + }, + { + "epoch": 0.97, + "learning_rate": 2.6078234704112336e-05, + "loss": 1.2533, + "theoretical_loss": 3.333333043799954, + "tokens_seen": 3194486784 + }, + { + "epoch": 0.97, + "learning_rate": 2.6062186559679037e-05, + "loss": 1.1023, + "theoretical_loss": 3.3333278971626905, + "tokens_seen": 3194552320 + }, + { + "epoch": 0.97, + "learning_rate": 2.604613841524574e-05, + "loss": 1.2868, + "theoretical_loss": 3.3333227506605714, + "tokens_seen": 3194617856 + }, + { + "epoch": 0.97, + "learning_rate": 2.603009027081244e-05, + "loss": 1.2359, + "theoretical_loss": 3.3333176042935904, + "tokens_seen": 3194683392 + }, + { + "epoch": 0.97, + "learning_rate": 2.601404212637914e-05, + "loss": 1.2705, + "theoretical_loss": 3.333312458061741, + "tokens_seen": 3194748928 + }, + { + "epoch": 0.97, + "learning_rate": 2.599799398194584e-05, + "loss": 1.3323, + "theoretical_loss": 3.333307311965017, + "tokens_seen": 3194814464 + }, + { + "epoch": 0.97, + "objective/train/docs_used": 1160990, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.2455697059631348, + "objective/train/theoretical_loss": 3.3333021660034117, + "objective/train/tokens_used": 3215340000, + "theoretical_loss": 3.3333021660034117, + "tokens_seen": 3194880000 + }, + { + "epoch": 0.97, + "learning_rate": 2.598194583751254e-05, + "loss": 1.2555, + "theoretical_loss": 3.3333021660034117, + "tokens_seen": 3194880000 + }, + { + "epoch": 0.97, + "learning_rate": 2.5965897693079237e-05, + "loss": 1.1699, + "theoretical_loss": 3.3332970201769196, + "tokens_seen": 3194945536 + }, + { + "epoch": 0.97, + "learning_rate": 2.5949849548645938e-05, + "loss": 1.2471, + "theoretical_loss": 3.3332918744855338, + "tokens_seen": 3195011072 + }, + { + "epoch": 0.97, + "learning_rate": 2.593380140421264e-05, + "loss": 1.2159, + "theoretical_loss": 3.333286728929248, + "tokens_seen": 3195076608 + }, + { + "epoch": 0.97, + "learning_rate": 2.5917753259779336e-05, + "loss": 1.2818, + "theoretical_loss": 3.3332815835080556, + "tokens_seen": 3195142144 + }, + { + "epoch": 0.97, + "learning_rate": 2.5901705115346036e-05, + "loss": 1.2464, + "theoretical_loss": 3.333276438221951, + "tokens_seen": 3195207680 + }, + { + "epoch": 0.97, + "learning_rate": 2.588565697091274e-05, + "loss": 1.1699, + "theoretical_loss": 3.3332712930709274, + "tokens_seen": 3195273216 + }, + { + "epoch": 0.97, + "learning_rate": 2.586960882647944e-05, + "loss": 1.2488, + "theoretical_loss": 3.3332661480549786, + "tokens_seen": 3195338752 + }, + { + "epoch": 0.97, + "learning_rate": 2.585356068204614e-05, + "loss": 1.2755, + "theoretical_loss": 3.333261003174098, + "tokens_seen": 3195404288 + }, + { + "epoch": 0.97, + "learning_rate": 2.583751253761284e-05, + "loss": 1.2781, + "theoretical_loss": 3.3332558584282803, + "tokens_seen": 3195469824 + }, + { + "epoch": 0.97, + "learning_rate": 2.582146439317954e-05, + "loss": 1.2469, + "theoretical_loss": 3.333250713817518, + "tokens_seen": 3195535360 + }, + { + "epoch": 0.97, + "learning_rate": 2.580541624874624e-05, + "loss": 1.2495, + "theoretical_loss": 3.333245569341805, + "tokens_seen": 3195600896 + }, + { + "epoch": 0.97, + "learning_rate": 2.5789368104312937e-05, + "loss": 1.2329, + "theoretical_loss": 3.3332404250011356, + "tokens_seen": 3195666432 + }, + { + "epoch": 0.97, + "learning_rate": 2.5773319959879638e-05, + "loss": 1.3328, + "theoretical_loss": 3.333235280795503, + "tokens_seen": 3195731968 + }, + { + "epoch": 0.97, + "learning_rate": 2.575727181544634e-05, + "loss": 1.2482, + "theoretical_loss": 3.333230136724901, + "tokens_seen": 3195797504 + }, + { + "epoch": 0.97, + "learning_rate": 2.5741223671013043e-05, + "loss": 1.1533, + "theoretical_loss": 3.333224992789323, + "tokens_seen": 3195863040 + }, + { + "epoch": 0.97, + "learning_rate": 2.5725175526579743e-05, + "loss": 1.278, + "theoretical_loss": 3.3332198489887634, + "tokens_seen": 3195928576 + }, + { + "epoch": 0.97, + "learning_rate": 2.570912738214644e-05, + "loss": 1.3266, + "theoretical_loss": 3.3332147053232153, + "tokens_seen": 3195994112 + }, + { + "epoch": 0.97, + "learning_rate": 2.569307923771314e-05, + "loss": 1.2088, + "theoretical_loss": 3.3332095617926725, + "tokens_seen": 3196059648 + }, + { + "epoch": 0.97, + "learning_rate": 2.5677031093279842e-05, + "loss": 1.1763, + "theoretical_loss": 3.3332044183971288, + "tokens_seen": 3196125184 + }, + { + "epoch": 0.97, + "learning_rate": 2.566098294884654e-05, + "loss": 1.3406, + "theoretical_loss": 3.333199275136578, + "tokens_seen": 3196190720 + }, + { + "epoch": 0.97, + "learning_rate": 2.564493480441324e-05, + "loss": 1.2307, + "theoretical_loss": 3.333194132011013, + "tokens_seen": 3196256256 + }, + { + "epoch": 0.97, + "learning_rate": 2.562888665997994e-05, + "loss": 1.2151, + "theoretical_loss": 3.3331889890204285, + "tokens_seen": 3196321792 + }, + { + "epoch": 0.97, + "learning_rate": 2.5612838515546638e-05, + "loss": 1.1535, + "theoretical_loss": 3.3331838461648178, + "tokens_seen": 3196387328 + }, + { + "epoch": 0.97, + "learning_rate": 2.5596790371113345e-05, + "loss": 1.1645, + "theoretical_loss": 3.3331787034441747, + "tokens_seen": 3196452864 + }, + { + "epoch": 0.97, + "learning_rate": 2.5580742226680042e-05, + "loss": 1.2356, + "theoretical_loss": 3.3331735608584925, + "tokens_seen": 3196518400 + }, + { + "epoch": 0.97, + "learning_rate": 2.5564694082246743e-05, + "loss": 1.2955, + "theoretical_loss": 3.333168418407765, + "tokens_seen": 3196583936 + }, + { + "epoch": 0.97, + "learning_rate": 2.5548645937813444e-05, + "loss": 1.2056, + "theoretical_loss": 3.3331632760919865, + "tokens_seen": 3196649472 + }, + { + "epoch": 0.97, + "learning_rate": 2.553259779338014e-05, + "loss": 1.275, + "theoretical_loss": 3.3331581339111502, + "tokens_seen": 3196715008 + }, + { + "epoch": 0.97, + "learning_rate": 2.551654964894684e-05, + "loss": 1.2291, + "theoretical_loss": 3.3331529918652496, + "tokens_seen": 3196780544 + }, + { + "epoch": 0.97, + "learning_rate": 2.5500501504513542e-05, + "loss": 1.2577, + "theoretical_loss": 3.333147849954279, + "tokens_seen": 3196846080 + }, + { + "epoch": 0.97, + "learning_rate": 2.548445336008024e-05, + "loss": 1.2079, + "theoretical_loss": 3.333142708178231, + "tokens_seen": 3196911616 + }, + { + "epoch": 0.97, + "learning_rate": 2.546840521564694e-05, + "loss": 1.1685, + "theoretical_loss": 3.3331375665371006, + "tokens_seen": 3196977152 + }, + { + "epoch": 0.97, + "learning_rate": 2.545235707121364e-05, + "loss": 1.2482, + "theoretical_loss": 3.3331324250308807, + "tokens_seen": 3197042688 + }, + { + "epoch": 0.97, + "learning_rate": 2.5436308926780345e-05, + "loss": 1.3399, + "theoretical_loss": 3.3331272836595653, + "tokens_seen": 3197108224 + }, + { + "epoch": 0.97, + "learning_rate": 2.5420260782347045e-05, + "loss": 1.2341, + "theoretical_loss": 3.3331221424231483, + "tokens_seen": 3197173760 + }, + { + "epoch": 0.97, + "learning_rate": 2.5404212637913743e-05, + "loss": 1.1345, + "theoretical_loss": 3.3331170013216225, + "tokens_seen": 3197239296 + }, + { + "epoch": 0.97, + "learning_rate": 2.5388164493480443e-05, + "loss": 1.3422, + "theoretical_loss": 3.3331118603549825, + "tokens_seen": 3197304832 + }, + { + "epoch": 0.97, + "learning_rate": 2.5372116349047144e-05, + "loss": 1.2363, + "theoretical_loss": 3.333106719523222, + "tokens_seen": 3197370368 + }, + { + "epoch": 0.97, + "learning_rate": 2.535606820461384e-05, + "loss": 1.2155, + "theoretical_loss": 3.3331015788263336, + "tokens_seen": 3197435904 + }, + { + "epoch": 0.97, + "learning_rate": 2.534002006018054e-05, + "loss": 1.3129, + "theoretical_loss": 3.3330964382643122, + "tokens_seen": 3197501440 + }, + { + "epoch": 0.97, + "learning_rate": 2.5323971915747242e-05, + "loss": 1.2887, + "theoretical_loss": 3.3330912978371514, + "tokens_seen": 3197566976 + }, + { + "epoch": 0.97, + "learning_rate": 2.530792377131394e-05, + "loss": 1.27, + "theoretical_loss": 3.333086157544844, + "tokens_seen": 3197632512 + }, + { + "epoch": 0.97, + "learning_rate": 2.5291875626880647e-05, + "loss": 1.1621, + "theoretical_loss": 3.333081017387385, + "tokens_seen": 3197698048 + }, + { + "epoch": 0.97, + "learning_rate": 2.5275827482447344e-05, + "loss": 1.2545, + "theoretical_loss": 3.3330758773647666, + "tokens_seen": 3197763584 + }, + { + "epoch": 0.97, + "learning_rate": 2.5259779338014045e-05, + "loss": 1.2871, + "theoretical_loss": 3.3330707374769837, + "tokens_seen": 3197829120 + }, + { + "epoch": 0.97, + "learning_rate": 2.5243731193580745e-05, + "loss": 1.1762, + "theoretical_loss": 3.3330655977240293, + "tokens_seen": 3197894656 + }, + { + "epoch": 0.97, + "learning_rate": 2.5227683049147443e-05, + "loss": 1.3516, + "theoretical_loss": 3.333060458105898, + "tokens_seen": 3197960192 + }, + { + "epoch": 0.97, + "learning_rate": 2.5211634904714143e-05, + "loss": 1.2484, + "theoretical_loss": 3.333055318622582, + "tokens_seen": 3198025728 + }, + { + "epoch": 0.97, + "learning_rate": 2.5195586760280844e-05, + "loss": 1.2635, + "theoretical_loss": 3.3330501792740765, + "tokens_seen": 3198091264 + }, + { + "epoch": 0.97, + "objective/train/docs_used": 1162228, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.161818504333496, + "objective/train/theoretical_loss": 3.3330450400603744, + "objective/train/tokens_used": 3218616800, + "theoretical_loss": 3.3330450400603744, + "tokens_seen": 3198156800 + }, + { + "epoch": 0.97, + "learning_rate": 2.517953861584754e-05, + "loss": 1.1715, + "theoretical_loss": 3.3330450400603744, + "tokens_seen": 3198156800 + }, + { + "epoch": 0.97, + "learning_rate": 2.5163490471414242e-05, + "loss": 1.3641, + "theoretical_loss": 3.3330399009814693, + "tokens_seen": 3198222336 + }, + { + "epoch": 0.97, + "learning_rate": 2.5147442326980946e-05, + "loss": 1.2986, + "theoretical_loss": 3.3330347620373555, + "tokens_seen": 3198287872 + }, + { + "epoch": 0.97, + "learning_rate": 2.5131394182547647e-05, + "loss": 1.2287, + "theoretical_loss": 3.3330296232280263, + "tokens_seen": 3198353408 + }, + { + "epoch": 0.97, + "learning_rate": 2.5115346038114347e-05, + "loss": 1.2095, + "theoretical_loss": 3.3330244845534756, + "tokens_seen": 3198418944 + }, + { + "epoch": 0.97, + "learning_rate": 2.5099297893681044e-05, + "loss": 1.1856, + "theoretical_loss": 3.333019346013697, + "tokens_seen": 3198484480 + }, + { + "epoch": 0.97, + "learning_rate": 2.5083249749247745e-05, + "loss": 1.2976, + "theoretical_loss": 3.3330142076086844, + "tokens_seen": 3198550016 + }, + { + "epoch": 0.97, + "learning_rate": 2.5067201604814446e-05, + "loss": 1.2312, + "theoretical_loss": 3.3330090693384307, + "tokens_seen": 3198615552 + }, + { + "epoch": 0.97, + "learning_rate": 2.5051153460381143e-05, + "loss": 1.3356, + "theoretical_loss": 3.3330039312029305, + "tokens_seen": 3198681088 + }, + { + "epoch": 0.97, + "learning_rate": 2.5035105315947844e-05, + "loss": 1.2296, + "theoretical_loss": 3.3329987932021776, + "tokens_seen": 3198746624 + }, + { + "epoch": 0.97, + "learning_rate": 2.5019057171514544e-05, + "loss": 1.1736, + "theoretical_loss": 3.332993655336165, + "tokens_seen": 3198812160 + }, + { + "epoch": 0.97, + "learning_rate": 2.500300902708124e-05, + "loss": 1.3045, + "theoretical_loss": 3.3329885176048863, + "tokens_seen": 3198877696 + }, + { + "epoch": 0.97, + "learning_rate": 2.4986960882647945e-05, + "loss": 1.259, + "theoretical_loss": 3.332983380008336, + "tokens_seen": 3198943232 + }, + { + "epoch": 0.97, + "learning_rate": 2.4970912738214643e-05, + "loss": 1.1786, + "theoretical_loss": 3.3329782425465075, + "tokens_seen": 3199008768 + }, + { + "epoch": 0.97, + "learning_rate": 2.4954864593781347e-05, + "loss": 1.277, + "theoretical_loss": 3.3329731052193945, + "tokens_seen": 3199074304 + }, + { + "epoch": 0.97, + "learning_rate": 2.4938816449348047e-05, + "loss": 1.1751, + "theoretical_loss": 3.3329679680269906, + "tokens_seen": 3199139840 + }, + { + "epoch": 0.97, + "learning_rate": 2.4922768304914745e-05, + "loss": 1.2359, + "theoretical_loss": 3.332962830969289, + "tokens_seen": 3199205376 + }, + { + "epoch": 0.97, + "learning_rate": 2.4906720160481445e-05, + "loss": 1.2476, + "theoretical_loss": 3.332957694046285, + "tokens_seen": 3199270912 + }, + { + "epoch": 0.97, + "learning_rate": 2.4890672016048146e-05, + "loss": 1.253, + "theoretical_loss": 3.3329525572579706, + "tokens_seen": 3199336448 + }, + { + "epoch": 0.97, + "learning_rate": 2.4874623871614847e-05, + "loss": 1.248, + "theoretical_loss": 3.33294742060434, + "tokens_seen": 3199401984 + }, + { + "epoch": 0.97, + "learning_rate": 2.4858575727181547e-05, + "loss": 1.2726, + "theoretical_loss": 3.3329422840853877, + "tokens_seen": 3199467520 + }, + { + "epoch": 0.97, + "learning_rate": 2.4842527582748244e-05, + "loss": 1.2043, + "theoretical_loss": 3.3329371477011067, + "tokens_seen": 3199533056 + }, + { + "epoch": 0.97, + "learning_rate": 2.4826479438314945e-05, + "loss": 1.2221, + "theoretical_loss": 3.3329320114514904, + "tokens_seen": 3199598592 + }, + { + "epoch": 0.97, + "learning_rate": 2.481043129388165e-05, + "loss": 1.2975, + "theoretical_loss": 3.332926875336533, + "tokens_seen": 3199664128 + }, + { + "epoch": 0.97, + "learning_rate": 2.4794383149448346e-05, + "loss": 1.2084, + "theoretical_loss": 3.3329217393562285, + "tokens_seen": 3199729664 + }, + { + "epoch": 0.97, + "learning_rate": 2.4778335005015047e-05, + "loss": 1.2873, + "theoretical_loss": 3.33291660351057, + "tokens_seen": 3199795200 + }, + { + "epoch": 0.97, + "learning_rate": 2.4762286860581748e-05, + "loss": 1.3478, + "theoretical_loss": 3.3329114677995517, + "tokens_seen": 3199860736 + }, + { + "epoch": 0.97, + "learning_rate": 2.4746238716148445e-05, + "loss": 1.2941, + "theoretical_loss": 3.3329063322231667, + "tokens_seen": 3199926272 + }, + { + "epoch": 0.97, + "learning_rate": 2.473019057171515e-05, + "loss": 1.1668, + "theoretical_loss": 3.3329011967814095, + "tokens_seen": 3199991808 + }, + { + "epoch": 0.97, + "learning_rate": 2.4714142427281846e-05, + "loss": 1.2024, + "theoretical_loss": 3.332896061474273, + "tokens_seen": 3200057344 + }, + { + "epoch": 0.97, + "learning_rate": 2.4698094282848547e-05, + "loss": 1.2936, + "theoretical_loss": 3.3328909263017517, + "tokens_seen": 3200122880 + }, + { + "epoch": 0.97, + "learning_rate": 2.4682046138415247e-05, + "loss": 1.3206, + "theoretical_loss": 3.3328857912638385, + "tokens_seen": 3200188416 + }, + { + "epoch": 0.97, + "learning_rate": 2.4665997993981945e-05, + "loss": 1.339, + "theoretical_loss": 3.332880656360528, + "tokens_seen": 3200253952 + }, + { + "epoch": 0.97, + "learning_rate": 2.464994984954865e-05, + "loss": 1.2075, + "theoretical_loss": 3.3328755215918133, + "tokens_seen": 3200319488 + }, + { + "epoch": 0.97, + "learning_rate": 2.4633901705115346e-05, + "loss": 1.2432, + "theoretical_loss": 3.3328703869576883, + "tokens_seen": 3200385024 + }, + { + "epoch": 0.97, + "learning_rate": 2.4617853560682047e-05, + "loss": 1.25, + "theoretical_loss": 3.3328652524581464, + "tokens_seen": 3200450560 + }, + { + "epoch": 0.97, + "learning_rate": 2.4601805416248747e-05, + "loss": 1.2812, + "theoretical_loss": 3.332860118093182, + "tokens_seen": 3200516096 + }, + { + "epoch": 0.97, + "learning_rate": 2.4585757271815448e-05, + "loss": 1.2165, + "theoretical_loss": 3.3328549838627883, + "tokens_seen": 3200581632 + }, + { + "epoch": 0.97, + "learning_rate": 2.456970912738215e-05, + "loss": 1.3005, + "theoretical_loss": 3.332849849766959, + "tokens_seen": 3200647168 + }, + { + "epoch": 0.97, + "learning_rate": 2.455366098294885e-05, + "loss": 1.2097, + "theoretical_loss": 3.332844715805688, + "tokens_seen": 3200712704 + }, + { + "epoch": 0.97, + "learning_rate": 2.4537612838515546e-05, + "loss": 1.1998, + "theoretical_loss": 3.3328395819789693, + "tokens_seen": 3200778240 + }, + { + "epoch": 0.97, + "learning_rate": 2.4521564694082247e-05, + "loss": 1.2648, + "theoretical_loss": 3.332834448286796, + "tokens_seen": 3200843776 + }, + { + "epoch": 0.97, + "learning_rate": 2.4505516549648948e-05, + "loss": 1.317, + "theoretical_loss": 3.332829314729162, + "tokens_seen": 3200909312 + }, + { + "epoch": 0.97, + "learning_rate": 2.4489468405215648e-05, + "loss": 1.2089, + "theoretical_loss": 3.3328241813060617, + "tokens_seen": 3200974848 + }, + { + "epoch": 0.97, + "learning_rate": 2.447342026078235e-05, + "loss": 1.2715, + "theoretical_loss": 3.3328190480174875, + "tokens_seen": 3201040384 + }, + { + "epoch": 0.97, + "learning_rate": 2.4457372116349046e-05, + "loss": 1.1965, + "theoretical_loss": 3.3328139148634346, + "tokens_seen": 3201105920 + }, + { + "epoch": 0.97, + "learning_rate": 2.4441323971915747e-05, + "loss": 1.2031, + "theoretical_loss": 3.332808781843896, + "tokens_seen": 3201171456 + }, + { + "epoch": 0.97, + "learning_rate": 2.442527582748245e-05, + "loss": 1.1807, + "theoretical_loss": 3.3328036489588646, + "tokens_seen": 3201236992 + }, + { + "epoch": 0.97, + "learning_rate": 2.4409227683049148e-05, + "loss": 1.3255, + "theoretical_loss": 3.3327985162083356, + "tokens_seen": 3201302528 + }, + { + "epoch": 0.97, + "learning_rate": 2.439317953861585e-05, + "loss": 1.3014, + "theoretical_loss": 3.332793383592302, + "tokens_seen": 3201368064 + }, + { + "epoch": 0.97, + "objective/train/docs_used": 1163323, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.2866604328155518, + "objective/train/theoretical_loss": 3.3327882511107574, + "objective/train/tokens_used": 3221893600, + "theoretical_loss": 3.3327882511107574, + "tokens_seen": 3201433600 + }, + { + "epoch": 0.97, + "learning_rate": 2.437713139418255e-05, + "loss": 1.2781, + "theoretical_loss": 3.3327882511107574, + "tokens_seen": 3201433600 + }, + { + "epoch": 0.97, + "learning_rate": 2.436108324974925e-05, + "loss": 1.1933, + "theoretical_loss": 3.332783118763696, + "tokens_seen": 3201499136 + }, + { + "epoch": 0.97, + "learning_rate": 2.434503510531595e-05, + "loss": 1.2138, + "theoretical_loss": 3.332777986551111, + "tokens_seen": 3201564672 + }, + { + "epoch": 0.97, + "learning_rate": 2.4328986960882648e-05, + "loss": 1.2629, + "theoretical_loss": 3.332772854472996, + "tokens_seen": 3201630208 + }, + { + "epoch": 0.97, + "learning_rate": 2.431293881644935e-05, + "loss": 1.2663, + "theoretical_loss": 3.3327677225293457, + "tokens_seen": 3201695744 + }, + { + "epoch": 0.97, + "learning_rate": 2.429689067201605e-05, + "loss": 1.1803, + "theoretical_loss": 3.332762590720153, + "tokens_seen": 3201761280 + }, + { + "epoch": 0.97, + "learning_rate": 2.428084252758275e-05, + "loss": 1.2629, + "theoretical_loss": 3.332757459045412, + "tokens_seen": 3201826816 + }, + { + "epoch": 0.97, + "learning_rate": 2.426479438314945e-05, + "loss": 1.2768, + "theoretical_loss": 3.332752327505116, + "tokens_seen": 3201892352 + }, + { + "epoch": 0.97, + "learning_rate": 2.424874623871615e-05, + "loss": 1.2096, + "theoretical_loss": 3.3327471960992594, + "tokens_seen": 3201957888 + }, + { + "epoch": 0.97, + "learning_rate": 2.4232698094282848e-05, + "loss": 1.3084, + "theoretical_loss": 3.332742064827835, + "tokens_seen": 3202023424 + }, + { + "epoch": 0.97, + "learning_rate": 2.421664994984955e-05, + "loss": 1.2599, + "theoretical_loss": 3.3327369336908372, + "tokens_seen": 3202088960 + }, + { + "epoch": 0.97, + "learning_rate": 2.420060180541625e-05, + "loss": 1.1355, + "theoretical_loss": 3.3327318026882597, + "tokens_seen": 3202154496 + }, + { + "epoch": 0.97, + "learning_rate": 2.418455366098295e-05, + "loss": 1.2811, + "theoretical_loss": 3.332726671820096, + "tokens_seen": 3202220032 + }, + { + "epoch": 0.97, + "learning_rate": 2.416850551654965e-05, + "loss": 1.3048, + "theoretical_loss": 3.33272154108634, + "tokens_seen": 3202285568 + }, + { + "epoch": 0.97, + "learning_rate": 2.4152457372116348e-05, + "loss": 1.2795, + "theoretical_loss": 3.3327164104869853, + "tokens_seen": 3202351104 + }, + { + "epoch": 0.97, + "learning_rate": 2.4136409227683052e-05, + "loss": 1.2607, + "theoretical_loss": 3.3327112800220258, + "tokens_seen": 3202416640 + }, + { + "epoch": 0.97, + "learning_rate": 2.4120361083249753e-05, + "loss": 1.2468, + "theoretical_loss": 3.3327061496914547, + "tokens_seen": 3202482176 + }, + { + "epoch": 0.97, + "learning_rate": 2.410431293881645e-05, + "loss": 1.3203, + "theoretical_loss": 3.3327010194952664, + "tokens_seen": 3202547712 + }, + { + "epoch": 0.97, + "learning_rate": 2.408826479438315e-05, + "loss": 1.0833, + "theoretical_loss": 3.332695889433454, + "tokens_seen": 3202613248 + }, + { + "epoch": 0.97, + "learning_rate": 2.407221664994985e-05, + "loss": 1.1984, + "theoretical_loss": 3.3326907595060122, + "tokens_seen": 3202678784 + }, + { + "epoch": 0.97, + "learning_rate": 2.4056168505516552e-05, + "loss": 1.1915, + "theoretical_loss": 3.332685629712934, + "tokens_seen": 3202744320 + }, + { + "epoch": 0.97, + "learning_rate": 2.4040120361083252e-05, + "loss": 1.2629, + "theoretical_loss": 3.3326805000542135, + "tokens_seen": 3202809856 + }, + { + "epoch": 0.97, + "learning_rate": 2.402407221664995e-05, + "loss": 1.1815, + "theoretical_loss": 3.3326753705298438, + "tokens_seen": 3202875392 + }, + { + "epoch": 0.97, + "learning_rate": 2.400802407221665e-05, + "loss": 1.2461, + "theoretical_loss": 3.332670241139819, + "tokens_seen": 3202940928 + }, + { + "epoch": 0.97, + "learning_rate": 2.399197592778335e-05, + "loss": 1.2662, + "theoretical_loss": 3.332665111884133, + "tokens_seen": 3203006464 + }, + { + "epoch": 0.97, + "learning_rate": 2.397592778335005e-05, + "loss": 1.2781, + "theoretical_loss": 3.332659982762779, + "tokens_seen": 3203072000 + }, + { + "epoch": 0.97, + "learning_rate": 2.3959879638916752e-05, + "loss": 1.2436, + "theoretical_loss": 3.3326548537757517, + "tokens_seen": 3203137536 + }, + { + "epoch": 0.97, + "learning_rate": 2.3943831494483453e-05, + "loss": 1.2119, + "theoretical_loss": 3.3326497249230442, + "tokens_seen": 3203203072 + }, + { + "epoch": 0.97, + "learning_rate": 2.392778335005015e-05, + "loss": 1.2612, + "theoretical_loss": 3.33264459620465, + "tokens_seen": 3203268608 + }, + { + "epoch": 0.97, + "learning_rate": 2.391173520561685e-05, + "loss": 1.2387, + "theoretical_loss": 3.332639467620563, + "tokens_seen": 3203334144 + }, + { + "epoch": 0.97, + "learning_rate": 2.389568706118355e-05, + "loss": 1.2251, + "theoretical_loss": 3.3326343391707773, + "tokens_seen": 3203399680 + }, + { + "epoch": 0.97, + "learning_rate": 2.3879638916750252e-05, + "loss": 1.1516, + "theoretical_loss": 3.3326292108552864, + "tokens_seen": 3203465216 + }, + { + "epoch": 0.97, + "learning_rate": 2.3863590772316953e-05, + "loss": 1.3299, + "theoretical_loss": 3.332624082674084, + "tokens_seen": 3203530752 + }, + { + "epoch": 0.97, + "learning_rate": 2.384754262788365e-05, + "loss": 1.2755, + "theoretical_loss": 3.332618954627164, + "tokens_seen": 3203596288 + }, + { + "epoch": 0.97, + "learning_rate": 2.3831494483450354e-05, + "loss": 1.2379, + "theoretical_loss": 3.3326138267145202, + "tokens_seen": 3203661824 + }, + { + "epoch": 0.97, + "learning_rate": 2.3815446339017055e-05, + "loss": 1.2042, + "theoretical_loss": 3.332608698936146, + "tokens_seen": 3203727360 + }, + { + "epoch": 0.97, + "learning_rate": 2.3799398194583752e-05, + "loss": 1.211, + "theoretical_loss": 3.3326035712920348, + "tokens_seen": 3203792896 + }, + { + "epoch": 0.97, + "learning_rate": 2.3783350050150452e-05, + "loss": 1.4313, + "theoretical_loss": 3.332598443782181, + "tokens_seen": 3203858432 + }, + { + "epoch": 0.97, + "learning_rate": 2.3767301905717153e-05, + "loss": 1.2529, + "theoretical_loss": 3.3325933164065784, + "tokens_seen": 3203923968 + }, + { + "epoch": 0.97, + "learning_rate": 2.3751253761283854e-05, + "loss": 1.1989, + "theoretical_loss": 3.3325881891652207, + "tokens_seen": 3203989504 + }, + { + "epoch": 0.97, + "learning_rate": 2.3735205616850554e-05, + "loss": 1.217, + "theoretical_loss": 3.332583062058101, + "tokens_seen": 3204055040 + }, + { + "epoch": 0.97, + "learning_rate": 2.371915747241725e-05, + "loss": 1.2476, + "theoretical_loss": 3.332577935085214, + "tokens_seen": 3204120576 + }, + { + "epoch": 0.97, + "learning_rate": 2.3703109327983952e-05, + "loss": 1.3149, + "theoretical_loss": 3.332572808246552, + "tokens_seen": 3204186112 + }, + { + "epoch": 0.97, + "learning_rate": 2.3687061183550653e-05, + "loss": 1.1341, + "theoretical_loss": 3.332567681542111, + "tokens_seen": 3204251648 + }, + { + "epoch": 0.97, + "learning_rate": 2.3671013039117353e-05, + "loss": 1.1981, + "theoretical_loss": 3.332562554971882, + "tokens_seen": 3204317184 + }, + { + "epoch": 0.97, + "learning_rate": 2.3654964894684054e-05, + "loss": 1.1157, + "theoretical_loss": 3.332557428535861, + "tokens_seen": 3204382720 + }, + { + "epoch": 0.97, + "learning_rate": 2.3638916750250755e-05, + "loss": 1.2044, + "theoretical_loss": 3.3325523022340406, + "tokens_seen": 3204448256 + }, + { + "epoch": 0.97, + "learning_rate": 2.3622868605817452e-05, + "loss": 1.3562, + "theoretical_loss": 3.3325471760664147, + "tokens_seen": 3204513792 + }, + { + "epoch": 0.97, + "learning_rate": 2.3606820461384156e-05, + "loss": 1.1986, + "theoretical_loss": 3.3325420500329774, + "tokens_seen": 3204579328 + }, + { + "epoch": 0.97, + "learning_rate": 2.3590772316950853e-05, + "loss": 1.2994, + "theoretical_loss": 3.332536924133722, + "tokens_seen": 3204644864 + }, + { + "epoch": 0.97, + "objective/train/docs_used": 1164453, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.3510029315948486, + "objective/train/theoretical_loss": 3.3325317983686427, + "objective/train/tokens_used": 3225170400, + "theoretical_loss": 3.3325317983686427, + "tokens_seen": 3204710400 + }, + { + "epoch": 0.97, + "learning_rate": 2.3574724172517554e-05, + "loss": 1.2973, + "theoretical_loss": 3.3325317983686427, + "tokens_seen": 3204710400 + }, + { + "epoch": 0.97, + "learning_rate": 2.3558676028084255e-05, + "loss": 1.3135, + "theoretical_loss": 3.332526672737733, + "tokens_seen": 3204775936 + }, + { + "epoch": 0.97, + "learning_rate": 2.3542627883650952e-05, + "loss": 1.1061, + "theoretical_loss": 3.3325215472409866, + "tokens_seen": 3204841472 + }, + { + "epoch": 0.97, + "learning_rate": 2.3526579739217656e-05, + "loss": 1.1767, + "theoretical_loss": 3.332516421878397, + "tokens_seen": 3204907008 + }, + { + "epoch": 0.97, + "learning_rate": 2.3510531594784353e-05, + "loss": 1.2312, + "theoretical_loss": 3.332511296649959, + "tokens_seen": 3204972544 + }, + { + "epoch": 0.97, + "learning_rate": 2.3494483450351054e-05, + "loss": 1.1908, + "theoretical_loss": 3.3325061715556648, + "tokens_seen": 3205038080 + }, + { + "epoch": 0.97, + "learning_rate": 2.3478435305917754e-05, + "loss": 1.1986, + "theoretical_loss": 3.3325010465955094, + "tokens_seen": 3205103616 + }, + { + "epoch": 0.97, + "learning_rate": 2.346238716148445e-05, + "loss": 1.1258, + "theoretical_loss": 3.3324959217694854, + "tokens_seen": 3205169152 + }, + { + "epoch": 0.97, + "learning_rate": 2.3446339017051156e-05, + "loss": 1.1876, + "theoretical_loss": 3.332490797077588, + "tokens_seen": 3205234688 + }, + { + "epoch": 0.97, + "learning_rate": 2.3430290872617856e-05, + "loss": 1.2394, + "theoretical_loss": 3.3324856725198098, + "tokens_seen": 3205300224 + }, + { + "epoch": 0.97, + "learning_rate": 2.3414242728184553e-05, + "loss": 1.2693, + "theoretical_loss": 3.332480548096145, + "tokens_seen": 3205365760 + }, + { + "epoch": 0.97, + "learning_rate": 2.3398194583751254e-05, + "loss": 1.0874, + "theoretical_loss": 3.3324754238065872, + "tokens_seen": 3205431296 + }, + { + "epoch": 0.97, + "learning_rate": 2.3382146439317955e-05, + "loss": 1.263, + "theoretical_loss": 3.3324702996511304, + "tokens_seen": 3205496832 + }, + { + "epoch": 0.97, + "learning_rate": 2.3366098294884655e-05, + "loss": 1.2967, + "theoretical_loss": 3.3324651756297676, + "tokens_seen": 3205562368 + }, + { + "epoch": 0.97, + "learning_rate": 2.3350050150451356e-05, + "loss": 1.236, + "theoretical_loss": 3.332460051742494, + "tokens_seen": 3205627904 + }, + { + "epoch": 0.97, + "learning_rate": 2.3334002006018053e-05, + "loss": 1.1824, + "theoretical_loss": 3.3324549279893017, + "tokens_seen": 3205693440 + }, + { + "epoch": 0.97, + "learning_rate": 2.3317953861584754e-05, + "loss": 1.2327, + "theoretical_loss": 3.3324498043701856, + "tokens_seen": 3205758976 + }, + { + "epoch": 0.97, + "learning_rate": 2.3301905717151458e-05, + "loss": 1.2624, + "theoretical_loss": 3.3324446808851387, + "tokens_seen": 3205824512 + }, + { + "epoch": 0.97, + "learning_rate": 2.3285857572718155e-05, + "loss": 1.2901, + "theoretical_loss": 3.3324395575341557, + "tokens_seen": 3205890048 + }, + { + "epoch": 0.97, + "learning_rate": 2.3269809428284856e-05, + "loss": 1.2704, + "theoretical_loss": 3.3324344343172294, + "tokens_seen": 3205955584 + }, + { + "epoch": 0.97, + "learning_rate": 2.3253761283851556e-05, + "loss": 1.253, + "theoretical_loss": 3.3324293112343537, + "tokens_seen": 3206021120 + }, + { + "epoch": 0.97, + "learning_rate": 2.3237713139418254e-05, + "loss": 1.1285, + "theoretical_loss": 3.332424188285523, + "tokens_seen": 3206086656 + }, + { + "epoch": 0.97, + "learning_rate": 2.3221664994984958e-05, + "loss": 1.3294, + "theoretical_loss": 3.3324190654707304, + "tokens_seen": 3206152192 + }, + { + "epoch": 0.97, + "learning_rate": 2.3205616850551655e-05, + "loss": 1.1788, + "theoretical_loss": 3.33241394278997, + "tokens_seen": 3206217728 + }, + { + "epoch": 0.97, + "learning_rate": 2.3189568706118356e-05, + "loss": 1.289, + "theoretical_loss": 3.332408820243235, + "tokens_seen": 3206283264 + }, + { + "epoch": 0.97, + "learning_rate": 2.3173520561685056e-05, + "loss": 1.3121, + "theoretical_loss": 3.33240369783052, + "tokens_seen": 3206348800 + }, + { + "epoch": 0.97, + "learning_rate": 2.3157472417251757e-05, + "loss": 1.2717, + "theoretical_loss": 3.332398575551818, + "tokens_seen": 3206414336 + }, + { + "epoch": 0.97, + "learning_rate": 2.3141424272818457e-05, + "loss": 1.2371, + "theoretical_loss": 3.3323934534071236, + "tokens_seen": 3206479872 + }, + { + "epoch": 0.97, + "learning_rate": 2.3125376128385158e-05, + "loss": 1.264, + "theoretical_loss": 3.3323883313964298, + "tokens_seen": 3206545408 + }, + { + "epoch": 0.97, + "learning_rate": 2.3109327983951855e-05, + "loss": 1.2293, + "theoretical_loss": 3.3323832095197305, + "tokens_seen": 3206610944 + }, + { + "epoch": 0.97, + "learning_rate": 2.3093279839518556e-05, + "loss": 1.2369, + "theoretical_loss": 3.3323780877770197, + "tokens_seen": 3206676480 + }, + { + "epoch": 0.97, + "learning_rate": 2.3077231695085257e-05, + "loss": 1.2255, + "theoretical_loss": 3.332372966168291, + "tokens_seen": 3206742016 + }, + { + "epoch": 0.97, + "learning_rate": 2.3061183550651957e-05, + "loss": 1.1979, + "theoretical_loss": 3.332367844693538, + "tokens_seen": 3206807552 + }, + { + "epoch": 0.97, + "learning_rate": 2.3045135406218658e-05, + "loss": 1.3483, + "theoretical_loss": 3.332362723352755, + "tokens_seen": 3206873088 + }, + { + "epoch": 0.97, + "learning_rate": 2.3029087261785355e-05, + "loss": 1.3336, + "theoretical_loss": 3.332357602145935, + "tokens_seen": 3206938624 + }, + { + "epoch": 0.97, + "learning_rate": 2.3013039117352056e-05, + "loss": 1.1546, + "theoretical_loss": 3.3323524810730722, + "tokens_seen": 3207004160 + }, + { + "epoch": 0.97, + "learning_rate": 2.299699097291876e-05, + "loss": 1.2657, + "theoretical_loss": 3.3323473601341607, + "tokens_seen": 3207069696 + }, + { + "epoch": 0.97, + "learning_rate": 2.2980942828485457e-05, + "loss": 1.2058, + "theoretical_loss": 3.3323422393291935, + "tokens_seen": 3207135232 + }, + { + "epoch": 0.97, + "learning_rate": 2.2964894684052158e-05, + "loss": 1.2408, + "theoretical_loss": 3.332337118658165, + "tokens_seen": 3207200768 + }, + { + "epoch": 0.97, + "learning_rate": 2.2948846539618858e-05, + "loss": 1.1477, + "theoretical_loss": 3.3323319981210684, + "tokens_seen": 3207266304 + }, + { + "epoch": 0.97, + "learning_rate": 2.293279839518556e-05, + "loss": 1.2601, + "theoretical_loss": 3.3323268777178976, + "tokens_seen": 3207331840 + }, + { + "epoch": 0.97, + "learning_rate": 2.291675025075226e-05, + "loss": 1.2506, + "theoretical_loss": 3.332321757448647, + "tokens_seen": 3207397376 + }, + { + "epoch": 0.97, + "learning_rate": 2.2900702106318957e-05, + "loss": 1.1642, + "theoretical_loss": 3.3323166373133093, + "tokens_seen": 3207462912 + }, + { + "epoch": 0.97, + "learning_rate": 2.2884653961885657e-05, + "loss": 1.2778, + "theoretical_loss": 3.3323115173118794, + "tokens_seen": 3207528448 + }, + { + "epoch": 0.97, + "learning_rate": 2.2868605817452358e-05, + "loss": 1.3454, + "theoretical_loss": 3.3323063974443503, + "tokens_seen": 3207593984 + }, + { + "epoch": 0.97, + "learning_rate": 2.285255767301906e-05, + "loss": 1.3371, + "theoretical_loss": 3.3323012777107155, + "tokens_seen": 3207659520 + }, + { + "epoch": 0.97, + "learning_rate": 2.283650952858576e-05, + "loss": 1.1826, + "theoretical_loss": 3.3322961581109696, + "tokens_seen": 3207725056 + }, + { + "epoch": 0.97, + "learning_rate": 2.282046138415246e-05, + "loss": 1.2038, + "theoretical_loss": 3.332291038645106, + "tokens_seen": 3207790592 + }, + { + "epoch": 0.97, + "learning_rate": 2.2804413239719157e-05, + "loss": 1.2359, + "theoretical_loss": 3.3322859193131187, + "tokens_seen": 3207856128 + }, + { + "epoch": 0.97, + "learning_rate": 2.2788365095285858e-05, + "loss": 1.1958, + "theoretical_loss": 3.332280800115001, + "tokens_seen": 3207921664 + }, + { + "epoch": 0.97, + "objective/train/docs_used": 1165709, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.218569278717041, + "objective/train/theoretical_loss": 3.3322756810507466, + "objective/train/tokens_used": 3228447200, + "theoretical_loss": 3.3322756810507466, + "tokens_seen": 3207987200 + }, + { + "epoch": 0.97, + "learning_rate": 2.277231695085256e-05, + "loss": 1.1767, + "theoretical_loss": 3.3322756810507466, + "tokens_seen": 3207987200 + }, + { + "epoch": 0.97, + "learning_rate": 2.275626880641926e-05, + "loss": 1.2534, + "theoretical_loss": 3.3322705621203497, + "tokens_seen": 3208052736 + }, + { + "epoch": 0.97, + "learning_rate": 2.274022066198596e-05, + "loss": 1.2982, + "theoretical_loss": 3.332265443323804, + "tokens_seen": 3208118272 + }, + { + "epoch": 0.97, + "learning_rate": 2.2724172517552657e-05, + "loss": 1.309, + "theoretical_loss": 3.3322603246611027, + "tokens_seen": 3208183808 + }, + { + "epoch": 0.97, + "learning_rate": 2.270812437311936e-05, + "loss": 1.3152, + "theoretical_loss": 3.3322552061322406, + "tokens_seen": 3208249344 + }, + { + "epoch": 0.97, + "learning_rate": 2.269207622868606e-05, + "loss": 1.2758, + "theoretical_loss": 3.3322500877372105, + "tokens_seen": 3208314880 + }, + { + "epoch": 0.97, + "learning_rate": 2.267602808425276e-05, + "loss": 1.2711, + "theoretical_loss": 3.332244969476007, + "tokens_seen": 3208380416 + }, + { + "epoch": 0.97, + "learning_rate": 2.265997993981946e-05, + "loss": 1.1732, + "theoretical_loss": 3.332239851348623, + "tokens_seen": 3208445952 + }, + { + "epoch": 0.97, + "learning_rate": 2.264393179538616e-05, + "loss": 1.2568, + "theoretical_loss": 3.332234733355053, + "tokens_seen": 3208511488 + }, + { + "epoch": 0.97, + "learning_rate": 2.262788365095286e-05, + "loss": 1.2279, + "theoretical_loss": 3.33222961549529, + "tokens_seen": 3208577024 + }, + { + "epoch": 0.97, + "learning_rate": 2.261183550651956e-05, + "loss": 1.2396, + "theoretical_loss": 3.332224497769329, + "tokens_seen": 3208642560 + }, + { + "epoch": 0.97, + "learning_rate": 2.259578736208626e-05, + "loss": 1.2814, + "theoretical_loss": 3.332219380177163, + "tokens_seen": 3208708096 + }, + { + "epoch": 0.97, + "learning_rate": 2.257973921765296e-05, + "loss": 1.2295, + "theoretical_loss": 3.332214262718785, + "tokens_seen": 3208773632 + }, + { + "epoch": 0.97, + "learning_rate": 2.256369107321966e-05, + "loss": 1.2442, + "theoretical_loss": 3.3322091453941898, + "tokens_seen": 3208839168 + }, + { + "epoch": 0.97, + "learning_rate": 2.254764292878636e-05, + "loss": 1.2314, + "theoretical_loss": 3.332204028203371, + "tokens_seen": 3208904704 + }, + { + "epoch": 0.97, + "learning_rate": 2.253159478435306e-05, + "loss": 1.2152, + "theoretical_loss": 3.332198911146323, + "tokens_seen": 3208970240 + }, + { + "epoch": 0.97, + "learning_rate": 2.251554663991976e-05, + "loss": 1.2946, + "theoretical_loss": 3.3321937942230377, + "tokens_seen": 3209035776 + }, + { + "epoch": 0.97, + "learning_rate": 2.249949849548646e-05, + "loss": 1.3035, + "theoretical_loss": 3.3321886774335105, + "tokens_seen": 3209101312 + }, + { + "epoch": 0.97, + "learning_rate": 2.2483450351053163e-05, + "loss": 1.2072, + "theoretical_loss": 3.332183560777735, + "tokens_seen": 3209166848 + }, + { + "epoch": 0.97, + "learning_rate": 2.246740220661986e-05, + "loss": 1.3215, + "theoretical_loss": 3.3321784442557045, + "tokens_seen": 3209232384 + }, + { + "epoch": 0.97, + "learning_rate": 2.245135406218656e-05, + "loss": 1.2463, + "theoretical_loss": 3.332173327867413, + "tokens_seen": 3209297920 + }, + { + "epoch": 0.97, + "learning_rate": 2.243530591775326e-05, + "loss": 1.2059, + "theoretical_loss": 3.3321682116128537, + "tokens_seen": 3209363456 + }, + { + "epoch": 0.97, + "learning_rate": 2.241925777331996e-05, + "loss": 1.1809, + "theoretical_loss": 3.3321630954920214, + "tokens_seen": 3209428992 + }, + { + "epoch": 0.97, + "learning_rate": 2.2403209628886663e-05, + "loss": 1.2274, + "theoretical_loss": 3.3321579795049097, + "tokens_seen": 3209494528 + }, + { + "epoch": 0.97, + "learning_rate": 2.238716148445336e-05, + "loss": 1.3062, + "theoretical_loss": 3.3321528636515114, + "tokens_seen": 3209560064 + }, + { + "epoch": 0.97, + "learning_rate": 2.237111334002006e-05, + "loss": 1.168, + "theoretical_loss": 3.3321477479318213, + "tokens_seen": 3209625600 + }, + { + "epoch": 0.97, + "learning_rate": 2.235506519558676e-05, + "loss": 1.2105, + "theoretical_loss": 3.3321426323458323, + "tokens_seen": 3209691136 + }, + { + "epoch": 0.97, + "learning_rate": 2.233901705115346e-05, + "loss": 1.2365, + "theoretical_loss": 3.332137516893539, + "tokens_seen": 3209756672 + }, + { + "epoch": 0.97, + "learning_rate": 2.2322968906720163e-05, + "loss": 1.2449, + "theoretical_loss": 3.3321324015749347, + "tokens_seen": 3209822208 + }, + { + "epoch": 0.97, + "learning_rate": 2.2306920762286863e-05, + "loss": 1.2916, + "theoretical_loss": 3.3321272863900138, + "tokens_seen": 3209887744 + }, + { + "epoch": 0.97, + "learning_rate": 2.229087261785356e-05, + "loss": 1.2481, + "theoretical_loss": 3.332122171338769, + "tokens_seen": 3209953280 + }, + { + "epoch": 0.97, + "learning_rate": 2.227482447342026e-05, + "loss": 1.1733, + "theoretical_loss": 3.332117056421195, + "tokens_seen": 3210018816 + }, + { + "epoch": 0.97, + "learning_rate": 2.2258776328986962e-05, + "loss": 1.2413, + "theoretical_loss": 3.3321119416372853, + "tokens_seen": 3210084352 + }, + { + "epoch": 0.97, + "learning_rate": 2.2242728184553662e-05, + "loss": 1.2518, + "theoretical_loss": 3.3321068269870335, + "tokens_seen": 3210149888 + }, + { + "epoch": 0.97, + "learning_rate": 2.2226680040120363e-05, + "loss": 1.3908, + "theoretical_loss": 3.3321017124704335, + "tokens_seen": 3210215424 + }, + { + "epoch": 0.97, + "learning_rate": 2.221063189568706e-05, + "loss": 1.1688, + "theoretical_loss": 3.3320965980874795, + "tokens_seen": 3210280960 + }, + { + "epoch": 0.97, + "learning_rate": 2.219458375125376e-05, + "loss": 1.2634, + "theoretical_loss": 3.3320914838381643, + "tokens_seen": 3210346496 + }, + { + "epoch": 0.97, + "learning_rate": 2.2178535606820465e-05, + "loss": 1.3135, + "theoretical_loss": 3.3320863697224823, + "tokens_seen": 3210412032 + }, + { + "epoch": 0.97, + "learning_rate": 2.2162487462387162e-05, + "loss": 1.2486, + "theoretical_loss": 3.3320812557404276, + "tokens_seen": 3210477568 + }, + { + "epoch": 0.97, + "learning_rate": 2.2146439317953863e-05, + "loss": 1.3023, + "theoretical_loss": 3.3320761418919935, + "tokens_seen": 3210543104 + }, + { + "epoch": 0.97, + "learning_rate": 2.2130391173520564e-05, + "loss": 1.1616, + "theoretical_loss": 3.332071028177174, + "tokens_seen": 3210608640 + }, + { + "epoch": 0.97, + "learning_rate": 2.211434302908726e-05, + "loss": 1.163, + "theoretical_loss": 3.3320659145959626, + "tokens_seen": 3210674176 + }, + { + "epoch": 0.97, + "learning_rate": 2.2098294884653965e-05, + "loss": 1.174, + "theoretical_loss": 3.3320608011483532, + "tokens_seen": 3210739712 + }, + { + "epoch": 0.97, + "learning_rate": 2.2082246740220662e-05, + "loss": 1.1538, + "theoretical_loss": 3.3320556878343397, + "tokens_seen": 3210805248 + }, + { + "epoch": 0.97, + "learning_rate": 2.2066198595787363e-05, + "loss": 1.2604, + "theoretical_loss": 3.3320505746539157, + "tokens_seen": 3210870784 + }, + { + "epoch": 0.97, + "learning_rate": 2.2050150451354063e-05, + "loss": 1.2437, + "theoretical_loss": 3.332045461607075, + "tokens_seen": 3210936320 + }, + { + "epoch": 0.97, + "learning_rate": 2.2034102306920764e-05, + "loss": 1.2338, + "theoretical_loss": 3.3320403486938117, + "tokens_seen": 3211001856 + }, + { + "epoch": 0.97, + "learning_rate": 2.2018054162487465e-05, + "loss": 1.2535, + "theoretical_loss": 3.3320352359141197, + "tokens_seen": 3211067392 + }, + { + "epoch": 0.97, + "learning_rate": 2.2002006018054165e-05, + "loss": 1.1657, + "theoretical_loss": 3.332030123267992, + "tokens_seen": 3211132928 + }, + { + "epoch": 0.97, + "learning_rate": 2.1985957873620862e-05, + "loss": 1.1866, + "theoretical_loss": 3.332025010755423, + "tokens_seen": 3211198464 + }, + { + "epoch": 0.97, + "objective/train/docs_used": 1166890, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.1292212009429932, + "objective/train/theoretical_loss": 3.3320198983764064, + "objective/train/tokens_used": 3231724000, + "theoretical_loss": 3.3320198983764064, + "tokens_seen": 3211264000 + }, + { + "epoch": 0.97, + "learning_rate": 2.1969909729187563e-05, + "loss": 1.2784, + "theoretical_loss": 3.3320198983764064, + "tokens_seen": 3211264000 + }, + { + "epoch": 0.97, + "learning_rate": 2.1953861584754264e-05, + "loss": 1.2726, + "theoretical_loss": 3.332014786130936, + "tokens_seen": 3211329536 + }, + { + "epoch": 0.97, + "learning_rate": 2.1937813440320964e-05, + "loss": 1.2506, + "theoretical_loss": 3.3320096740190053, + "tokens_seen": 3211395072 + }, + { + "epoch": 0.97, + "learning_rate": 2.1921765295887665e-05, + "loss": 1.2809, + "theoretical_loss": 3.3320045620406082, + "tokens_seen": 3211460608 + }, + { + "epoch": 0.97, + "learning_rate": 2.1905717151454362e-05, + "loss": 1.3177, + "theoretical_loss": 3.3319994501957386, + "tokens_seen": 3211526144 + }, + { + "epoch": 0.97, + "learning_rate": 2.1889669007021063e-05, + "loss": 1.1871, + "theoretical_loss": 3.3319943384843906, + "tokens_seen": 3211591680 + }, + { + "epoch": 0.97, + "learning_rate": 2.1873620862587767e-05, + "loss": 1.1937, + "theoretical_loss": 3.331989226906557, + "tokens_seen": 3211657216 + }, + { + "epoch": 0.97, + "learning_rate": 2.1857572718154464e-05, + "loss": 1.1576, + "theoretical_loss": 3.331984115462233, + "tokens_seen": 3211722752 + }, + { + "epoch": 0.97, + "learning_rate": 2.1841524573721165e-05, + "loss": 1.1988, + "theoretical_loss": 3.3319790041514112, + "tokens_seen": 3211788288 + }, + { + "epoch": 0.97, + "learning_rate": 2.1825476429287865e-05, + "loss": 1.2988, + "theoretical_loss": 3.331973892974086, + "tokens_seen": 3211853824 + }, + { + "epoch": 0.97, + "learning_rate": 2.1809428284854563e-05, + "loss": 1.3071, + "theoretical_loss": 3.331968781930251, + "tokens_seen": 3211919360 + }, + { + "epoch": 0.97, + "learning_rate": 2.1793380140421267e-05, + "loss": 1.3549, + "theoretical_loss": 3.3319636710198997, + "tokens_seen": 3211984896 + }, + { + "epoch": 0.97, + "learning_rate": 2.1777331995987964e-05, + "loss": 1.242, + "theoretical_loss": 3.3319585602430264, + "tokens_seen": 3212050432 + }, + { + "epoch": 0.97, + "learning_rate": 2.1761283851554665e-05, + "loss": 1.3242, + "theoretical_loss": 3.3319534495996246, + "tokens_seen": 3212115968 + }, + { + "epoch": 0.97, + "learning_rate": 2.1745235707121365e-05, + "loss": 1.358, + "theoretical_loss": 3.3319483390896885, + "tokens_seen": 3212181504 + }, + { + "epoch": 0.97, + "learning_rate": 2.1729187562688066e-05, + "loss": 1.2546, + "theoretical_loss": 3.3319432287132114, + "tokens_seen": 3212247040 + }, + { + "epoch": 0.97, + "learning_rate": 2.1713139418254766e-05, + "loss": 1.2426, + "theoretical_loss": 3.331938118470187, + "tokens_seen": 3212312576 + }, + { + "epoch": 0.97, + "learning_rate": 2.1697091273821467e-05, + "loss": 1.2918, + "theoretical_loss": 3.3319330083606093, + "tokens_seen": 3212378112 + }, + { + "epoch": 0.97, + "learning_rate": 2.1681043129388164e-05, + "loss": 1.1261, + "theoretical_loss": 3.3319278983844725, + "tokens_seen": 3212443648 + }, + { + "epoch": 0.97, + "learning_rate": 2.1664994984954865e-05, + "loss": 1.2568, + "theoretical_loss": 3.33192278854177, + "tokens_seen": 3212509184 + }, + { + "epoch": 0.97, + "learning_rate": 2.1648946840521566e-05, + "loss": 1.2715, + "theoretical_loss": 3.3319176788324953, + "tokens_seen": 3212574720 + }, + { + "epoch": 0.97, + "learning_rate": 2.1632898696088266e-05, + "loss": 1.3224, + "theoretical_loss": 3.3319125692566427, + "tokens_seen": 3212640256 + }, + { + "epoch": 0.97, + "learning_rate": 2.1616850551654967e-05, + "loss": 1.264, + "theoretical_loss": 3.331907459814206, + "tokens_seen": 3212705792 + }, + { + "epoch": 0.97, + "learning_rate": 2.1600802407221664e-05, + "loss": 1.2567, + "theoretical_loss": 3.3319023505051786, + "tokens_seen": 3212771328 + }, + { + "epoch": 0.97, + "learning_rate": 2.1584754262788365e-05, + "loss": 1.3216, + "theoretical_loss": 3.3318972413295547, + "tokens_seen": 3212836864 + }, + { + "epoch": 0.97, + "learning_rate": 2.1568706118355065e-05, + "loss": 1.1803, + "theoretical_loss": 3.3318921322873276, + "tokens_seen": 3212902400 + }, + { + "epoch": 0.97, + "learning_rate": 2.1552657973921766e-05, + "loss": 1.1499, + "theoretical_loss": 3.3318870233784916, + "tokens_seen": 3212967936 + }, + { + "epoch": 0.97, + "learning_rate": 2.1536609829488467e-05, + "loss": 1.2425, + "theoretical_loss": 3.33188191460304, + "tokens_seen": 3213033472 + }, + { + "epoch": 0.97, + "learning_rate": 2.1520561685055164e-05, + "loss": 1.3143, + "theoretical_loss": 3.3318768059609676, + "tokens_seen": 3213099008 + }, + { + "epoch": 0.97, + "learning_rate": 2.1504513540621868e-05, + "loss": 1.2771, + "theoretical_loss": 3.3318716974522666, + "tokens_seen": 3213164544 + }, + { + "epoch": 0.97, + "learning_rate": 2.148846539618857e-05, + "loss": 1.1866, + "theoretical_loss": 3.331866589076932, + "tokens_seen": 3213230080 + }, + { + "epoch": 0.97, + "learning_rate": 2.1472417251755266e-05, + "loss": 1.1963, + "theoretical_loss": 3.3318614808349576, + "tokens_seen": 3213295616 + }, + { + "epoch": 0.97, + "learning_rate": 2.1456369107321966e-05, + "loss": 1.2005, + "theoretical_loss": 3.3318563727263366, + "tokens_seen": 3213361152 + }, + { + "epoch": 0.97, + "learning_rate": 2.1440320962888667e-05, + "loss": 1.2089, + "theoretical_loss": 3.331851264751063, + "tokens_seen": 3213426688 + }, + { + "epoch": 0.97, + "learning_rate": 2.1424272818455368e-05, + "loss": 1.2868, + "theoretical_loss": 3.331846156909131, + "tokens_seen": 3213492224 + }, + { + "epoch": 0.97, + "learning_rate": 2.140822467402207e-05, + "loss": 1.2332, + "theoretical_loss": 3.331841049200534, + "tokens_seen": 3213557760 + }, + { + "epoch": 0.97, + "learning_rate": 2.1392176529588766e-05, + "loss": 1.2341, + "theoretical_loss": 3.3318359416252656, + "tokens_seen": 3213623296 + }, + { + "epoch": 0.97, + "learning_rate": 2.1376128385155466e-05, + "loss": 1.2738, + "theoretical_loss": 3.3318308341833203, + "tokens_seen": 3213688832 + }, + { + "epoch": 0.97, + "learning_rate": 2.1360080240722167e-05, + "loss": 1.3014, + "theoretical_loss": 3.3318257268746914, + "tokens_seen": 3213754368 + }, + { + "epoch": 0.97, + "learning_rate": 2.1344032096288868e-05, + "loss": 1.2744, + "theoretical_loss": 3.3318206196993723, + "tokens_seen": 3213819904 + }, + { + "epoch": 0.97, + "learning_rate": 2.1327983951855568e-05, + "loss": 1.2086, + "theoretical_loss": 3.3318155126573576, + "tokens_seen": 3213885440 + }, + { + "epoch": 0.97, + "learning_rate": 2.131193580742227e-05, + "loss": 1.2117, + "theoretical_loss": 3.331810405748641, + "tokens_seen": 3213950976 + }, + { + "epoch": 0.97, + "learning_rate": 2.1295887662988966e-05, + "loss": 1.3266, + "theoretical_loss": 3.3318052989732156, + "tokens_seen": 3214016512 + }, + { + "epoch": 0.97, + "learning_rate": 2.127983951855567e-05, + "loss": 1.1586, + "theoretical_loss": 3.3318001923310763, + "tokens_seen": 3214082048 + }, + { + "epoch": 0.97, + "learning_rate": 2.1263791374122367e-05, + "loss": 1.1946, + "theoretical_loss": 3.3317950858222156, + "tokens_seen": 3214147584 + }, + { + "epoch": 0.97, + "learning_rate": 2.1247743229689068e-05, + "loss": 1.1592, + "theoretical_loss": 3.331789979446629, + "tokens_seen": 3214213120 + }, + { + "epoch": 0.97, + "learning_rate": 2.123169508525577e-05, + "loss": 1.1863, + "theoretical_loss": 3.3317848732043083, + "tokens_seen": 3214278656 + }, + { + "epoch": 0.97, + "learning_rate": 2.1215646940822466e-05, + "loss": 1.3025, + "theoretical_loss": 3.3317797670952487, + "tokens_seen": 3214344192 + }, + { + "epoch": 0.97, + "learning_rate": 2.119959879638917e-05, + "loss": 1.1914, + "theoretical_loss": 3.331774661119444, + "tokens_seen": 3214409728 + }, + { + "epoch": 0.97, + "learning_rate": 2.118355065195587e-05, + "loss": 1.2333, + "theoretical_loss": 3.331769555276887, + "tokens_seen": 3214475264 + }, + { + "epoch": 0.97, + "objective/train/docs_used": 1168047, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.2495125532150269, + "objective/train/theoretical_loss": 3.3317644495675722, + "objective/train/tokens_used": 3235000800, + "theoretical_loss": 3.3317644495675722, + "tokens_seen": 3214540800 + }, + { + "epoch": 0.97, + "learning_rate": 2.1167502507522568e-05, + "loss": 1.1715, + "theoretical_loss": 3.3317644495675722, + "tokens_seen": 3214540800 + }, + { + "epoch": 0.97, + "learning_rate": 2.115145436308927e-05, + "loss": 1.1356, + "theoretical_loss": 3.331759343991494, + "tokens_seen": 3214606336 + }, + { + "epoch": 0.97, + "learning_rate": 2.113540621865597e-05, + "loss": 1.1807, + "theoretical_loss": 3.331754238548645, + "tokens_seen": 3214671872 + }, + { + "epoch": 0.97, + "learning_rate": 2.111935807422267e-05, + "loss": 1.1526, + "theoretical_loss": 3.3317491332390192, + "tokens_seen": 3214737408 + }, + { + "epoch": 0.97, + "learning_rate": 2.110330992978937e-05, + "loss": 1.2967, + "theoretical_loss": 3.3317440280626114, + "tokens_seen": 3214802944 + }, + { + "epoch": 0.97, + "learning_rate": 2.1087261785356068e-05, + "loss": 1.3126, + "theoretical_loss": 3.3317389230194143, + "tokens_seen": 3214868480 + }, + { + "epoch": 0.97, + "learning_rate": 2.1071213640922768e-05, + "loss": 1.1255, + "theoretical_loss": 3.3317338181094227, + "tokens_seen": 3214934016 + }, + { + "epoch": 0.97, + "learning_rate": 2.1055165496489472e-05, + "loss": 1.2794, + "theoretical_loss": 3.3317287133326294, + "tokens_seen": 3214999552 + }, + { + "epoch": 0.97, + "learning_rate": 2.103911735205617e-05, + "loss": 1.1921, + "theoretical_loss": 3.331723608689029, + "tokens_seen": 3215065088 + }, + { + "epoch": 0.97, + "learning_rate": 2.102306920762287e-05, + "loss": 1.2608, + "theoretical_loss": 3.331718504178615, + "tokens_seen": 3215130624 + }, + { + "epoch": 0.97, + "learning_rate": 2.100702106318957e-05, + "loss": 1.3511, + "theoretical_loss": 3.331713399801381, + "tokens_seen": 3215196160 + }, + { + "epoch": 0.97, + "learning_rate": 2.0990972918756268e-05, + "loss": 1.2047, + "theoretical_loss": 3.331708295557321, + "tokens_seen": 3215261696 + }, + { + "epoch": 0.97, + "learning_rate": 2.0974924774322972e-05, + "loss": 1.2083, + "theoretical_loss": 3.331703191446429, + "tokens_seen": 3215327232 + }, + { + "epoch": 0.97, + "learning_rate": 2.095887662988967e-05, + "loss": 1.2892, + "theoretical_loss": 3.3316980874686988, + "tokens_seen": 3215392768 + }, + { + "epoch": 0.97, + "learning_rate": 2.094282848545637e-05, + "loss": 1.2066, + "theoretical_loss": 3.331692983624124, + "tokens_seen": 3215458304 + }, + { + "epoch": 0.97, + "learning_rate": 2.092678034102307e-05, + "loss": 1.1865, + "theoretical_loss": 3.331687879912698, + "tokens_seen": 3215523840 + }, + { + "epoch": 0.97, + "learning_rate": 2.0910732196589768e-05, + "loss": 1.2768, + "theoretical_loss": 3.3316827763344157, + "tokens_seen": 3215589376 + }, + { + "epoch": 0.97, + "learning_rate": 2.0894684052156472e-05, + "loss": 1.3557, + "theoretical_loss": 3.3316776728892696, + "tokens_seen": 3215654912 + }, + { + "epoch": 0.97, + "learning_rate": 2.0878635907723172e-05, + "loss": 1.1969, + "theoretical_loss": 3.331672569577255, + "tokens_seen": 3215720448 + }, + { + "epoch": 0.97, + "learning_rate": 2.086258776328987e-05, + "loss": 1.3039, + "theoretical_loss": 3.3316674663983648, + "tokens_seen": 3215785984 + }, + { + "epoch": 0.97, + "learning_rate": 2.084653961885657e-05, + "loss": 1.1595, + "theoretical_loss": 3.331662363352592, + "tokens_seen": 3215851520 + }, + { + "epoch": 0.97, + "learning_rate": 2.083049147442327e-05, + "loss": 1.2326, + "theoretical_loss": 3.3316572604399326, + "tokens_seen": 3215917056 + }, + { + "epoch": 0.97, + "learning_rate": 2.081444332998997e-05, + "loss": 1.1669, + "theoretical_loss": 3.3316521576603786, + "tokens_seen": 3215982592 + }, + { + "epoch": 0.97, + "learning_rate": 2.0798395185556672e-05, + "loss": 1.2894, + "theoretical_loss": 3.3316470550139243, + "tokens_seen": 3216048128 + }, + { + "epoch": 0.97, + "learning_rate": 2.078234704112337e-05, + "loss": 1.2644, + "theoretical_loss": 3.3316419525005636, + "tokens_seen": 3216113664 + }, + { + "epoch": 0.97, + "learning_rate": 2.076629889669007e-05, + "loss": 1.2901, + "theoretical_loss": 3.33163685012029, + "tokens_seen": 3216179200 + }, + { + "epoch": 0.97, + "learning_rate": 2.0750250752256774e-05, + "loss": 1.1793, + "theoretical_loss": 3.3316317478730983, + "tokens_seen": 3216244736 + }, + { + "epoch": 0.97, + "learning_rate": 2.073420260782347e-05, + "loss": 1.1592, + "theoretical_loss": 3.3316266457589814, + "tokens_seen": 3216310272 + }, + { + "epoch": 0.97, + "learning_rate": 2.0718154463390172e-05, + "loss": 1.3084, + "theoretical_loss": 3.3316215437779335, + "tokens_seen": 3216375808 + }, + { + "epoch": 0.97, + "learning_rate": 2.0702106318956873e-05, + "loss": 1.2306, + "theoretical_loss": 3.331616441929948, + "tokens_seen": 3216441344 + }, + { + "epoch": 0.97, + "learning_rate": 2.068605817452357e-05, + "loss": 1.2338, + "theoretical_loss": 3.3316113402150194, + "tokens_seen": 3216506880 + }, + { + "epoch": 0.97, + "learning_rate": 2.0670010030090274e-05, + "loss": 1.2687, + "theoretical_loss": 3.3316062386331406, + "tokens_seen": 3216572416 + }, + { + "epoch": 0.97, + "learning_rate": 2.065396188565697e-05, + "loss": 1.2227, + "theoretical_loss": 3.3316011371843066, + "tokens_seen": 3216637952 + }, + { + "epoch": 0.97, + "learning_rate": 2.0637913741223672e-05, + "loss": 1.3252, + "theoretical_loss": 3.33159603586851, + "tokens_seen": 3216703488 + }, + { + "epoch": 0.97, + "learning_rate": 2.0621865596790372e-05, + "loss": 1.2055, + "theoretical_loss": 3.3315909346857455, + "tokens_seen": 3216769024 + }, + { + "epoch": 0.97, + "learning_rate": 2.0605817452357073e-05, + "loss": 1.1364, + "theoretical_loss": 3.3315858336360065, + "tokens_seen": 3216834560 + }, + { + "epoch": 0.97, + "learning_rate": 2.0589769307923774e-05, + "loss": 1.3096, + "theoretical_loss": 3.331580732719287, + "tokens_seen": 3216900096 + }, + { + "epoch": 0.97, + "learning_rate": 2.057372116349047e-05, + "loss": 1.2758, + "theoretical_loss": 3.3315756319355807, + "tokens_seen": 3216965632 + }, + { + "epoch": 0.97, + "learning_rate": 2.055767301905717e-05, + "loss": 1.367, + "theoretical_loss": 3.3315705312848816, + "tokens_seen": 3217031168 + }, + { + "epoch": 0.97, + "learning_rate": 2.0541624874623872e-05, + "loss": 1.1927, + "theoretical_loss": 3.331565430767183, + "tokens_seen": 3217096704 + }, + { + "epoch": 0.97, + "learning_rate": 2.0525576730190573e-05, + "loss": 1.2792, + "theoretical_loss": 3.3315603303824792, + "tokens_seen": 3217162240 + }, + { + "epoch": 0.97, + "learning_rate": 2.0509528585757273e-05, + "loss": 1.1794, + "theoretical_loss": 3.331555230130764, + "tokens_seen": 3217227776 + }, + { + "epoch": 0.97, + "learning_rate": 2.0493480441323974e-05, + "loss": 1.1659, + "theoretical_loss": 3.3315501300120314, + "tokens_seen": 3217293312 + }, + { + "epoch": 0.97, + "learning_rate": 2.047743229689067e-05, + "loss": 1.2893, + "theoretical_loss": 3.331545030026275, + "tokens_seen": 3217358848 + }, + { + "epoch": 0.97, + "learning_rate": 2.0461384152457372e-05, + "loss": 1.277, + "theoretical_loss": 3.3315399301734887, + "tokens_seen": 3217424384 + }, + { + "epoch": 0.97, + "learning_rate": 2.0445336008024073e-05, + "loss": 1.1995, + "theoretical_loss": 3.3315348304536654, + "tokens_seen": 3217489920 + }, + { + "epoch": 0.98, + "learning_rate": 2.0429287863590773e-05, + "loss": 1.3742, + "theoretical_loss": 3.3315297308668006, + "tokens_seen": 3217555456 + }, + { + "epoch": 0.98, + "learning_rate": 2.0413239719157474e-05, + "loss": 1.2806, + "theoretical_loss": 3.3315246314128872, + "tokens_seen": 3217620992 + }, + { + "epoch": 0.98, + "learning_rate": 2.039719157472417e-05, + "loss": 1.3325, + "theoretical_loss": 3.3315195320919186, + "tokens_seen": 3217686528 + }, + { + "epoch": 0.98, + "learning_rate": 2.0381143430290875e-05, + "loss": 1.3254, + "theoretical_loss": 3.3315144329038895, + "tokens_seen": 3217752064 + }, + { + "epoch": 0.98, + "objective/train/docs_used": 1169144, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.353203296661377, + "objective/train/theoretical_loss": 3.3315093338487936, + "objective/train/tokens_used": 3238277600, + "theoretical_loss": 3.3315093338487936, + "tokens_seen": 3217817600 + }, + { + "epoch": 0.98, + "learning_rate": 2.0365095285857576e-05, + "loss": 1.3461, + "theoretical_loss": 3.3315093338487936, + "tokens_seen": 3217817600 + }, + { + "epoch": 0.98, + "learning_rate": 2.0349047141424273e-05, + "loss": 1.2209, + "theoretical_loss": 3.331504234926624, + "tokens_seen": 3217883136 + }, + { + "epoch": 0.98, + "learning_rate": 2.0332998996990974e-05, + "loss": 1.3161, + "theoretical_loss": 3.331499136137375, + "tokens_seen": 3217948672 + }, + { + "epoch": 0.98, + "learning_rate": 2.0316950852557674e-05, + "loss": 1.2968, + "theoretical_loss": 3.331494037481041, + "tokens_seen": 3218014208 + }, + { + "epoch": 0.98, + "learning_rate": 2.0300902708124375e-05, + "loss": 1.2717, + "theoretical_loss": 3.331488938957615, + "tokens_seen": 3218079744 + }, + { + "epoch": 0.98, + "learning_rate": 2.0284854563691076e-05, + "loss": 1.2713, + "theoretical_loss": 3.331483840567091, + "tokens_seen": 3218145280 + }, + { + "epoch": 0.98, + "learning_rate": 2.0268806419257773e-05, + "loss": 1.2952, + "theoretical_loss": 3.331478742309463, + "tokens_seen": 3218210816 + }, + { + "epoch": 0.98, + "learning_rate": 2.0252758274824473e-05, + "loss": 1.2576, + "theoretical_loss": 3.331473644184725, + "tokens_seen": 3218276352 + }, + { + "epoch": 0.98, + "learning_rate": 2.0236710130391174e-05, + "loss": 1.3139, + "theoretical_loss": 3.33146854619287, + "tokens_seen": 3218341888 + }, + { + "epoch": 0.98, + "learning_rate": 2.0220661985957875e-05, + "loss": 1.2555, + "theoretical_loss": 3.331463448333893, + "tokens_seen": 3218407424 + }, + { + "epoch": 0.98, + "learning_rate": 2.0204613841524575e-05, + "loss": 1.1485, + "theoretical_loss": 3.331458350607787, + "tokens_seen": 3218472960 + }, + { + "epoch": 0.98, + "learning_rate": 2.0188565697091276e-05, + "loss": 1.1762, + "theoretical_loss": 3.331453253014546, + "tokens_seen": 3218538496 + }, + { + "epoch": 0.98, + "learning_rate": 2.0172517552657973e-05, + "loss": 1.1688, + "theoretical_loss": 3.331448155554164, + "tokens_seen": 3218604032 + }, + { + "epoch": 0.98, + "learning_rate": 2.0156469408224677e-05, + "loss": 1.2684, + "theoretical_loss": 3.3314430582266352, + "tokens_seen": 3218669568 + }, + { + "epoch": 0.98, + "learning_rate": 2.0140421263791374e-05, + "loss": 1.2538, + "theoretical_loss": 3.3314379610319524, + "tokens_seen": 3218735104 + }, + { + "epoch": 0.98, + "learning_rate": 2.0124373119358075e-05, + "loss": 1.2983, + "theoretical_loss": 3.33143286397011, + "tokens_seen": 3218800640 + }, + { + "epoch": 0.98, + "learning_rate": 2.0108324974924776e-05, + "loss": 1.2236, + "theoretical_loss": 3.3314277670411023, + "tokens_seen": 3218866176 + }, + { + "epoch": 0.98, + "learning_rate": 2.0092276830491473e-05, + "loss": 1.1259, + "theoretical_loss": 3.331422670244922, + "tokens_seen": 3218931712 + }, + { + "epoch": 0.98, + "learning_rate": 2.0076228686058177e-05, + "loss": 1.1878, + "theoretical_loss": 3.3314175735815645, + "tokens_seen": 3218997248 + }, + { + "epoch": 0.98, + "learning_rate": 2.0060180541624878e-05, + "loss": 1.2123, + "theoretical_loss": 3.331412477051022, + "tokens_seen": 3219062784 + }, + { + "epoch": 0.98, + "learning_rate": 2.0044132397191575e-05, + "loss": 1.2934, + "theoretical_loss": 3.3314073806532893, + "tokens_seen": 3219128320 + }, + { + "epoch": 0.98, + "learning_rate": 2.0028084252758276e-05, + "loss": 1.2128, + "theoretical_loss": 3.33140228438836, + "tokens_seen": 3219193856 + }, + { + "epoch": 0.98, + "learning_rate": 2.0012036108324976e-05, + "loss": 1.3281, + "theoretical_loss": 3.3313971882562283, + "tokens_seen": 3219259392 + }, + { + "epoch": 0.98, + "learning_rate": 1.9995987963891677e-05, + "loss": 1.2008, + "theoretical_loss": 3.3313920922568876, + "tokens_seen": 3219324928 + }, + { + "epoch": 0.98, + "learning_rate": 1.9979939819458377e-05, + "loss": 1.3318, + "theoretical_loss": 3.3313869963903318, + "tokens_seen": 3219390464 + }, + { + "epoch": 0.98, + "learning_rate": 1.9963891675025075e-05, + "loss": 1.1783, + "theoretical_loss": 3.3313819006565546, + "tokens_seen": 3219456000 + }, + { + "epoch": 0.98, + "learning_rate": 1.9947843530591775e-05, + "loss": 1.1917, + "theoretical_loss": 3.33137680505555, + "tokens_seen": 3219521536 + }, + { + "epoch": 0.98, + "learning_rate": 1.993179538615848e-05, + "loss": 1.3638, + "theoretical_loss": 3.331371709587312, + "tokens_seen": 3219587072 + }, + { + "epoch": 0.98, + "learning_rate": 1.9915747241725177e-05, + "loss": 1.1616, + "theoretical_loss": 3.331366614251834, + "tokens_seen": 3219652608 + }, + { + "epoch": 0.98, + "learning_rate": 1.9899699097291877e-05, + "loss": 1.2, + "theoretical_loss": 3.33136151904911, + "tokens_seen": 3219718144 + }, + { + "epoch": 0.98, + "learning_rate": 1.9883650952858578e-05, + "loss": 1.2396, + "theoretical_loss": 3.3313564239791345, + "tokens_seen": 3219783680 + }, + { + "epoch": 0.98, + "learning_rate": 1.9867602808425275e-05, + "loss": 1.268, + "theoretical_loss": 3.3313513290419006, + "tokens_seen": 3219849216 + }, + { + "epoch": 0.98, + "learning_rate": 1.985155466399198e-05, + "loss": 1.3812, + "theoretical_loss": 3.3313462342374023, + "tokens_seen": 3219914752 + }, + { + "epoch": 0.98, + "learning_rate": 1.9835506519558676e-05, + "loss": 1.1922, + "theoretical_loss": 3.3313411395656334, + "tokens_seen": 3219980288 + }, + { + "epoch": 0.98, + "learning_rate": 1.9819458375125377e-05, + "loss": 1.2962, + "theoretical_loss": 3.331336045026588, + "tokens_seen": 3220045824 + }, + { + "epoch": 0.98, + "learning_rate": 1.9803410230692078e-05, + "loss": 1.3031, + "theoretical_loss": 3.3313309506202597, + "tokens_seen": 3220111360 + }, + { + "epoch": 0.98, + "learning_rate": 1.9787362086258775e-05, + "loss": 1.2333, + "theoretical_loss": 3.331325856346642, + "tokens_seen": 3220176896 + }, + { + "epoch": 0.98, + "learning_rate": 1.977131394182548e-05, + "loss": 1.2731, + "theoretical_loss": 3.33132076220573, + "tokens_seen": 3220242432 + }, + { + "epoch": 0.98, + "learning_rate": 1.975526579739218e-05, + "loss": 1.2331, + "theoretical_loss": 3.331315668197516, + "tokens_seen": 3220307968 + }, + { + "epoch": 0.98, + "learning_rate": 1.9739217652958877e-05, + "loss": 1.2554, + "theoretical_loss": 3.3313105743219946, + "tokens_seen": 3220373504 + }, + { + "epoch": 0.98, + "learning_rate": 1.9723169508525577e-05, + "loss": 1.267, + "theoretical_loss": 3.33130548057916, + "tokens_seen": 3220439040 + }, + { + "epoch": 0.98, + "learning_rate": 1.9707121364092278e-05, + "loss": 1.2762, + "theoretical_loss": 3.331300386969005, + "tokens_seen": 3220504576 + }, + { + "epoch": 0.98, + "learning_rate": 1.969107321965898e-05, + "loss": 1.3897, + "theoretical_loss": 3.3312952934915248, + "tokens_seen": 3220570112 + }, + { + "epoch": 0.98, + "learning_rate": 1.967502507522568e-05, + "loss": 1.2704, + "theoretical_loss": 3.3312902001467117, + "tokens_seen": 3220635648 + }, + { + "epoch": 0.98, + "learning_rate": 1.9658976930792377e-05, + "loss": 1.2469, + "theoretical_loss": 3.331285106934561, + "tokens_seen": 3220701184 + }, + { + "epoch": 0.98, + "learning_rate": 1.9642928786359077e-05, + "loss": 1.2951, + "theoretical_loss": 3.3312800138550656, + "tokens_seen": 3220766720 + }, + { + "epoch": 0.98, + "learning_rate": 1.9626880641925778e-05, + "loss": 1.1686, + "theoretical_loss": 3.3312749209082195, + "tokens_seen": 3220832256 + }, + { + "epoch": 0.98, + "learning_rate": 1.961083249749248e-05, + "loss": 1.344, + "theoretical_loss": 3.331269828094017, + "tokens_seen": 3220897792 + }, + { + "epoch": 0.98, + "learning_rate": 1.959478435305918e-05, + "loss": 1.2698, + "theoretical_loss": 3.331264735412452, + "tokens_seen": 3220963328 + }, + { + "epoch": 0.98, + "learning_rate": 1.9578736208625876e-05, + "loss": 1.2519, + "theoretical_loss": 3.3312596428635173, + "tokens_seen": 3221028864 + }, + { + "epoch": 0.98, + "objective/train/docs_used": 1170443, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.2824424505233765, + "objective/train/theoretical_loss": 3.3312545504472073, + "objective/train/tokens_used": 3241554400, + "theoretical_loss": 3.3312545504472073, + "tokens_seen": 3221094400 + }, + { + "epoch": 0.98, + "learning_rate": 1.9562688064192577e-05, + "loss": 1.2653, + "theoretical_loss": 3.3312545504472073, + "tokens_seen": 3221094400 + }, + { + "epoch": 0.98, + "learning_rate": 1.954663991975928e-05, + "loss": 1.1993, + "theoretical_loss": 3.3312494581635166, + "tokens_seen": 3221159936 + }, + { + "epoch": 0.98, + "learning_rate": 1.9530591775325978e-05, + "loss": 1.2435, + "theoretical_loss": 3.331244366012438, + "tokens_seen": 3221225472 + }, + { + "epoch": 0.98, + "learning_rate": 1.951454363089268e-05, + "loss": 1.2078, + "theoretical_loss": 3.331239273993966, + "tokens_seen": 3221291008 + }, + { + "epoch": 0.98, + "learning_rate": 1.949849548645938e-05, + "loss": 1.3085, + "theoretical_loss": 3.331234182108094, + "tokens_seen": 3221356544 + }, + { + "epoch": 0.98, + "learning_rate": 1.9482447342026077e-05, + "loss": 1.2319, + "theoretical_loss": 3.331229090354816, + "tokens_seen": 3221422080 + }, + { + "epoch": 0.98, + "learning_rate": 1.946639919759278e-05, + "loss": 1.2608, + "theoretical_loss": 3.331223998734126, + "tokens_seen": 3221487616 + }, + { + "epoch": 0.98, + "learning_rate": 1.9450351053159478e-05, + "loss": 1.247, + "theoretical_loss": 3.331218907246018, + "tokens_seen": 3221553152 + }, + { + "epoch": 0.98, + "learning_rate": 1.943430290872618e-05, + "loss": 1.1874, + "theoretical_loss": 3.331213815890486, + "tokens_seen": 3221618688 + }, + { + "epoch": 0.98, + "learning_rate": 1.941825476429288e-05, + "loss": 1.3824, + "theoretical_loss": 3.3312087246675226, + "tokens_seen": 3221684224 + }, + { + "epoch": 0.98, + "learning_rate": 1.940220661985958e-05, + "loss": 1.2797, + "theoretical_loss": 3.331203633577123, + "tokens_seen": 3221749760 + }, + { + "epoch": 0.98, + "learning_rate": 1.938615847542628e-05, + "loss": 1.2631, + "theoretical_loss": 3.3311985426192807, + "tokens_seen": 3221815296 + }, + { + "epoch": 0.98, + "learning_rate": 1.937011033099298e-05, + "loss": 1.149, + "theoretical_loss": 3.331193451793989, + "tokens_seen": 3221880832 + }, + { + "epoch": 0.98, + "learning_rate": 1.935406218655968e-05, + "loss": 1.2256, + "theoretical_loss": 3.3311883611012423, + "tokens_seen": 3221946368 + }, + { + "epoch": 0.98, + "learning_rate": 1.933801404212638e-05, + "loss": 1.1534, + "theoretical_loss": 3.3311832705410347, + "tokens_seen": 3222011904 + }, + { + "epoch": 0.98, + "learning_rate": 1.932196589769308e-05, + "loss": 1.225, + "theoretical_loss": 3.3311781801133593, + "tokens_seen": 3222077440 + }, + { + "epoch": 0.98, + "learning_rate": 1.930591775325978e-05, + "loss": 1.3396, + "theoretical_loss": 3.3311730898182104, + "tokens_seen": 3222142976 + }, + { + "epoch": 0.98, + "learning_rate": 1.928986960882648e-05, + "loss": 1.2975, + "theoretical_loss": 3.331167999655582, + "tokens_seen": 3222208512 + }, + { + "epoch": 0.98, + "learning_rate": 1.9273821464393178e-05, + "loss": 1.0697, + "theoretical_loss": 3.3311629096254673, + "tokens_seen": 3222274048 + }, + { + "epoch": 0.98, + "learning_rate": 1.925777331995988e-05, + "loss": 1.297, + "theoretical_loss": 3.3311578197278613, + "tokens_seen": 3222339584 + }, + { + "epoch": 0.98, + "learning_rate": 1.9241725175526583e-05, + "loss": 1.2146, + "theoretical_loss": 3.331152729962757, + "tokens_seen": 3222405120 + }, + { + "epoch": 0.98, + "learning_rate": 1.922567703109328e-05, + "loss": 1.235, + "theoretical_loss": 3.331147640330148, + "tokens_seen": 3222470656 + }, + { + "epoch": 0.98, + "learning_rate": 1.920962888665998e-05, + "loss": 1.2584, + "theoretical_loss": 3.3311425508300285, + "tokens_seen": 3222536192 + }, + { + "epoch": 0.98, + "learning_rate": 1.919358074222668e-05, + "loss": 1.2665, + "theoretical_loss": 3.3311374614623928, + "tokens_seen": 3222601728 + }, + { + "epoch": 0.98, + "learning_rate": 1.9177532597793382e-05, + "loss": 1.3206, + "theoretical_loss": 3.3311323722272346, + "tokens_seen": 3222667264 + }, + { + "epoch": 0.98, + "learning_rate": 1.9161484453360083e-05, + "loss": 1.2161, + "theoretical_loss": 3.3311272831245473, + "tokens_seen": 3222732800 + }, + { + "epoch": 0.98, + "learning_rate": 1.914543630892678e-05, + "loss": 1.1648, + "theoretical_loss": 3.3311221941543248, + "tokens_seen": 3222798336 + }, + { + "epoch": 0.98, + "learning_rate": 1.912938816449348e-05, + "loss": 1.2979, + "theoretical_loss": 3.3311171053165616, + "tokens_seen": 3222863872 + }, + { + "epoch": 0.98, + "learning_rate": 1.911334002006018e-05, + "loss": 1.2288, + "theoretical_loss": 3.3311120166112507, + "tokens_seen": 3222929408 + }, + { + "epoch": 0.98, + "learning_rate": 1.9097291875626882e-05, + "loss": 1.2532, + "theoretical_loss": 3.3311069280383867, + "tokens_seen": 3222994944 + }, + { + "epoch": 0.98, + "learning_rate": 1.9081243731193582e-05, + "loss": 1.2738, + "theoretical_loss": 3.331101839597963, + "tokens_seen": 3223060480 + }, + { + "epoch": 0.98, + "learning_rate": 1.9065195586760283e-05, + "loss": 1.2385, + "theoretical_loss": 3.3310967512899734, + "tokens_seen": 3223126016 + }, + { + "epoch": 0.98, + "learning_rate": 1.904914744232698e-05, + "loss": 1.2094, + "theoretical_loss": 3.331091663114412, + "tokens_seen": 3223191552 + }, + { + "epoch": 0.98, + "learning_rate": 1.903309929789368e-05, + "loss": 1.2508, + "theoretical_loss": 3.331086575071273, + "tokens_seen": 3223257088 + }, + { + "epoch": 0.98, + "learning_rate": 1.901705115346038e-05, + "loss": 1.2719, + "theoretical_loss": 3.3310814871605494, + "tokens_seen": 3223322624 + }, + { + "epoch": 0.98, + "learning_rate": 1.9001003009027082e-05, + "loss": 1.2629, + "theoretical_loss": 3.331076399382236, + "tokens_seen": 3223388160 + }, + { + "epoch": 0.98, + "learning_rate": 1.8984954864593783e-05, + "loss": 1.3517, + "theoretical_loss": 3.331071311736326, + "tokens_seen": 3223453696 + }, + { + "epoch": 0.98, + "learning_rate": 1.896890672016048e-05, + "loss": 1.2998, + "theoretical_loss": 3.3310662242228135, + "tokens_seen": 3223519232 + }, + { + "epoch": 0.98, + "learning_rate": 1.8952858575727184e-05, + "loss": 1.256, + "theoretical_loss": 3.3310611368416922, + "tokens_seen": 3223584768 + }, + { + "epoch": 0.98, + "learning_rate": 1.8936810431293885e-05, + "loss": 1.2158, + "theoretical_loss": 3.331056049592956, + "tokens_seen": 3223650304 + }, + { + "epoch": 0.98, + "learning_rate": 1.8920762286860582e-05, + "loss": 1.1126, + "theoretical_loss": 3.331050962476599, + "tokens_seen": 3223715840 + }, + { + "epoch": 0.98, + "learning_rate": 1.8904714142427283e-05, + "loss": 1.2709, + "theoretical_loss": 3.3310458754926153, + "tokens_seen": 3223781376 + }, + { + "epoch": 0.98, + "learning_rate": 1.8888665997993983e-05, + "loss": 1.2334, + "theoretical_loss": 3.331040788640998, + "tokens_seen": 3223846912 + }, + { + "epoch": 0.98, + "learning_rate": 1.8872617853560684e-05, + "loss": 1.279, + "theoretical_loss": 3.3310357019217416, + "tokens_seen": 3223912448 + }, + { + "epoch": 0.98, + "learning_rate": 1.8856569709127385e-05, + "loss": 1.2808, + "theoretical_loss": 3.3310306153348397, + "tokens_seen": 3223977984 + }, + { + "epoch": 0.98, + "learning_rate": 1.8840521564694082e-05, + "loss": 1.3078, + "theoretical_loss": 3.331025528880286, + "tokens_seen": 3224043520 + }, + { + "epoch": 0.98, + "learning_rate": 1.8824473420260782e-05, + "loss": 1.2273, + "theoretical_loss": 3.3310204425580747, + "tokens_seen": 3224109056 + }, + { + "epoch": 0.98, + "learning_rate": 1.8808425275827483e-05, + "loss": 1.2572, + "theoretical_loss": 3.3310153563681992, + "tokens_seen": 3224174592 + }, + { + "epoch": 0.98, + "learning_rate": 1.8792377131394184e-05, + "loss": 1.3261, + "theoretical_loss": 3.331010270310654, + "tokens_seen": 3224240128 + }, + { + "epoch": 0.98, + "learning_rate": 1.8776328986960884e-05, + "loss": 1.2594, + "theoretical_loss": 3.3310051843854325, + "tokens_seen": 3224305664 + }, + { + "epoch": 0.98, + "objective/train/docs_used": 1171684, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.357993721961975, + "objective/train/theoretical_loss": 3.331000098592529, + "objective/train/tokens_used": 3244831200, + "theoretical_loss": 3.331000098592529, + "tokens_seen": 3224371200 + }, + { + "epoch": 0.98, + "learning_rate": 1.8760280842527585e-05, + "loss": 1.3039, + "theoretical_loss": 3.331000098592529, + "tokens_seen": 3224371200 + }, + { + "epoch": 0.98, + "learning_rate": 1.8744232698094282e-05, + "loss": 1.1692, + "theoretical_loss": 3.330995012931937, + "tokens_seen": 3224436736 + }, + { + "epoch": 0.98, + "learning_rate": 1.8728184553660986e-05, + "loss": 1.3145, + "theoretical_loss": 3.33098992740365, + "tokens_seen": 3224502272 + }, + { + "epoch": 0.98, + "learning_rate": 1.8712136409227684e-05, + "loss": 1.3052, + "theoretical_loss": 3.3309848420076626, + "tokens_seen": 3224567808 + }, + { + "epoch": 0.98, + "learning_rate": 1.8696088264794384e-05, + "loss": 1.3201, + "theoretical_loss": 3.3309797567439685, + "tokens_seen": 3224633344 + }, + { + "epoch": 0.98, + "learning_rate": 1.8680040120361085e-05, + "loss": 1.251, + "theoretical_loss": 3.3309746716125614, + "tokens_seen": 3224698880 + }, + { + "epoch": 0.98, + "learning_rate": 1.8663991975927782e-05, + "loss": 1.2074, + "theoretical_loss": 3.3309695866134357, + "tokens_seen": 3224764416 + }, + { + "epoch": 0.98, + "learning_rate": 1.8647943831494486e-05, + "loss": 1.2383, + "theoretical_loss": 3.330964501746584, + "tokens_seen": 3224829952 + }, + { + "epoch": 0.98, + "learning_rate": 1.8631895687061183e-05, + "loss": 1.2576, + "theoretical_loss": 3.330959417012002, + "tokens_seen": 3224895488 + }, + { + "epoch": 0.98, + "learning_rate": 1.8615847542627884e-05, + "loss": 1.1574, + "theoretical_loss": 3.3309543324096813, + "tokens_seen": 3224961024 + }, + { + "epoch": 0.98, + "learning_rate": 1.8599799398194585e-05, + "loss": 1.2789, + "theoretical_loss": 3.3309492479396177, + "tokens_seen": 3225026560 + }, + { + "epoch": 0.98, + "learning_rate": 1.8583751253761282e-05, + "loss": 1.2057, + "theoretical_loss": 3.3309441636018047, + "tokens_seen": 3225092096 + }, + { + "epoch": 0.98, + "learning_rate": 1.8567703109327986e-05, + "loss": 1.2325, + "theoretical_loss": 3.3309390793962352, + "tokens_seen": 3225157632 + }, + { + "epoch": 0.98, + "learning_rate": 1.8551654964894686e-05, + "loss": 1.1481, + "theoretical_loss": 3.330933995322904, + "tokens_seen": 3225223168 + }, + { + "epoch": 0.98, + "learning_rate": 1.8535606820461384e-05, + "loss": 1.2653, + "theoretical_loss": 3.330928911381805, + "tokens_seen": 3225288704 + }, + { + "epoch": 0.98, + "learning_rate": 1.8519558676028084e-05, + "loss": 1.327, + "theoretical_loss": 3.3309238275729314, + "tokens_seen": 3225354240 + }, + { + "epoch": 0.98, + "learning_rate": 1.8503510531594785e-05, + "loss": 1.2541, + "theoretical_loss": 3.3309187438962775, + "tokens_seen": 3225419776 + }, + { + "epoch": 0.98, + "learning_rate": 1.8487462387161486e-05, + "loss": 1.2807, + "theoretical_loss": 3.3309136603518374, + "tokens_seen": 3225485312 + }, + { + "epoch": 0.98, + "learning_rate": 1.8471414242728186e-05, + "loss": 1.2754, + "theoretical_loss": 3.3309085769396045, + "tokens_seen": 3225550848 + }, + { + "epoch": 0.98, + "learning_rate": 1.8455366098294884e-05, + "loss": 1.1962, + "theoretical_loss": 3.330903493659573, + "tokens_seen": 3225616384 + }, + { + "epoch": 0.98, + "learning_rate": 1.8439317953861584e-05, + "loss": 1.2847, + "theoretical_loss": 3.3308984105117365, + "tokens_seen": 3225681920 + }, + { + "epoch": 0.98, + "learning_rate": 1.8423269809428288e-05, + "loss": 1.2998, + "theoretical_loss": 3.3308933274960895, + "tokens_seen": 3225747456 + }, + { + "epoch": 0.98, + "learning_rate": 1.8407221664994985e-05, + "loss": 1.2754, + "theoretical_loss": 3.3308882446126247, + "tokens_seen": 3225812992 + }, + { + "epoch": 0.98, + "learning_rate": 1.8391173520561686e-05, + "loss": 1.374, + "theoretical_loss": 3.3308831618613373, + "tokens_seen": 3225878528 + }, + { + "epoch": 0.98, + "learning_rate": 1.8375125376128387e-05, + "loss": 1.1998, + "theoretical_loss": 3.33087807924222, + "tokens_seen": 3225944064 + }, + { + "epoch": 0.98, + "learning_rate": 1.8359077231695084e-05, + "loss": 1.2668, + "theoretical_loss": 3.330872996755268, + "tokens_seen": 3226009600 + }, + { + "epoch": 0.98, + "learning_rate": 1.8343029087261788e-05, + "loss": 1.2356, + "theoretical_loss": 3.330867914400474, + "tokens_seen": 3226075136 + }, + { + "epoch": 0.98, + "learning_rate": 1.8326980942828485e-05, + "loss": 1.1559, + "theoretical_loss": 3.330862832177832, + "tokens_seen": 3226140672 + }, + { + "epoch": 0.98, + "learning_rate": 1.8310932798395186e-05, + "loss": 1.2134, + "theoretical_loss": 3.3308577500873366, + "tokens_seen": 3226206208 + }, + { + "epoch": 0.98, + "learning_rate": 1.8294884653961886e-05, + "loss": 1.2125, + "theoretical_loss": 3.330852668128981, + "tokens_seen": 3226271744 + }, + { + "epoch": 0.98, + "learning_rate": 1.8278836509528587e-05, + "loss": 1.2363, + "theoretical_loss": 3.3308475863027596, + "tokens_seen": 3226337280 + }, + { + "epoch": 0.98, + "learning_rate": 1.8262788365095288e-05, + "loss": 1.2904, + "theoretical_loss": 3.330842504608666, + "tokens_seen": 3226402816 + }, + { + "epoch": 0.98, + "learning_rate": 1.824674022066199e-05, + "loss": 1.2675, + "theoretical_loss": 3.3308374230466944, + "tokens_seen": 3226468352 + }, + { + "epoch": 0.98, + "learning_rate": 1.8230692076228686e-05, + "loss": 1.2763, + "theoretical_loss": 3.330832341616838, + "tokens_seen": 3226533888 + }, + { + "epoch": 0.98, + "learning_rate": 1.8214643931795386e-05, + "loss": 1.2247, + "theoretical_loss": 3.330827260319091, + "tokens_seen": 3226599424 + }, + { + "epoch": 0.98, + "learning_rate": 1.8198595787362087e-05, + "loss": 1.3128, + "theoretical_loss": 3.3308221791534476, + "tokens_seen": 3226664960 + }, + { + "epoch": 0.98, + "learning_rate": 1.8182547642928788e-05, + "loss": 1.2505, + "theoretical_loss": 3.3308170981199012, + "tokens_seen": 3226730496 + }, + { + "epoch": 0.98, + "learning_rate": 1.8166499498495488e-05, + "loss": 1.3059, + "theoretical_loss": 3.330812017218446, + "tokens_seen": 3226796032 + }, + { + "epoch": 0.98, + "learning_rate": 1.8150451354062185e-05, + "loss": 1.2448, + "theoretical_loss": 3.330806936449076, + "tokens_seen": 3226861568 + }, + { + "epoch": 0.98, + "learning_rate": 1.8134403209628886e-05, + "loss": 1.1346, + "theoretical_loss": 3.3308018558117847, + "tokens_seen": 3226927104 + }, + { + "epoch": 0.98, + "learning_rate": 1.811835506519559e-05, + "loss": 1.1432, + "theoretical_loss": 3.3307967753065664, + "tokens_seen": 3226992640 + }, + { + "epoch": 0.98, + "learning_rate": 1.8102306920762287e-05, + "loss": 1.1575, + "theoretical_loss": 3.3307916949334144, + "tokens_seen": 3227058176 + }, + { + "epoch": 0.98, + "learning_rate": 1.8086258776328988e-05, + "loss": 1.3102, + "theoretical_loss": 3.330786614692323, + "tokens_seen": 3227123712 + }, + { + "epoch": 0.98, + "learning_rate": 1.807021063189569e-05, + "loss": 1.2249, + "theoretical_loss": 3.330781534583286, + "tokens_seen": 3227189248 + }, + { + "epoch": 0.98, + "learning_rate": 1.805416248746239e-05, + "loss": 1.2692, + "theoretical_loss": 3.3307764546062977, + "tokens_seen": 3227254784 + }, + { + "epoch": 0.98, + "learning_rate": 1.803811434302909e-05, + "loss": 1.1943, + "theoretical_loss": 3.330771374761351, + "tokens_seen": 3227320320 + }, + { + "epoch": 0.98, + "learning_rate": 1.8022066198595787e-05, + "loss": 1.2587, + "theoretical_loss": 3.3307662950484413, + "tokens_seen": 3227385856 + }, + { + "epoch": 0.98, + "learning_rate": 1.8006018054162488e-05, + "loss": 1.1328, + "theoretical_loss": 3.330761215467561, + "tokens_seen": 3227451392 + }, + { + "epoch": 0.98, + "learning_rate": 1.798996990972919e-05, + "loss": 1.1701, + "theoretical_loss": 3.3307561360187043, + "tokens_seen": 3227516928 + }, + { + "epoch": 0.98, + "learning_rate": 1.797392176529589e-05, + "loss": 1.2655, + "theoretical_loss": 3.3307510567018657, + "tokens_seen": 3227582464 + }, + { + "epoch": 0.98, + "objective/train/docs_used": 1172972, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.2853286266326904, + "objective/train/theoretical_loss": 3.3307459775170387, + "objective/train/tokens_used": 3248108000, + "theoretical_loss": 3.3307459775170387, + "tokens_seen": 3227648000 + }, + { + "epoch": 0.98, + "learning_rate": 1.795787362086259e-05, + "loss": 1.2748, + "theoretical_loss": 3.3307459775170387, + "tokens_seen": 3227648000 + }, + { + "epoch": 0.98, + "learning_rate": 1.794182547642929e-05, + "loss": 1.3193, + "theoretical_loss": 3.3307408984642173, + "tokens_seen": 3227713536 + }, + { + "epoch": 0.98, + "learning_rate": 1.7925777331995988e-05, + "loss": 1.3069, + "theoretical_loss": 3.330735819543395, + "tokens_seen": 3227779072 + }, + { + "epoch": 0.98, + "learning_rate": 1.7909729187562688e-05, + "loss": 1.3388, + "theoretical_loss": 3.330730740754566, + "tokens_seen": 3227844608 + }, + { + "epoch": 0.98, + "learning_rate": 1.789368104312939e-05, + "loss": 1.1795, + "theoretical_loss": 3.330725662097725, + "tokens_seen": 3227910144 + }, + { + "epoch": 0.98, + "learning_rate": 1.787763289869609e-05, + "loss": 1.2148, + "theoretical_loss": 3.330720583572864, + "tokens_seen": 3227975680 + }, + { + "epoch": 0.98, + "learning_rate": 1.786158475426279e-05, + "loss": 1.332, + "theoretical_loss": 3.3307155051799784, + "tokens_seen": 3228041216 + }, + { + "epoch": 0.98, + "learning_rate": 1.7845536609829487e-05, + "loss": 1.3742, + "theoretical_loss": 3.3307104269190617, + "tokens_seen": 3228106752 + }, + { + "epoch": 0.98, + "learning_rate": 1.782948846539619e-05, + "loss": 1.2966, + "theoretical_loss": 3.330705348790108, + "tokens_seen": 3228172288 + }, + { + "epoch": 0.98, + "learning_rate": 1.7813440320962892e-05, + "loss": 1.2029, + "theoretical_loss": 3.330700270793111, + "tokens_seen": 3228237824 + }, + { + "epoch": 0.98, + "learning_rate": 1.779739217652959e-05, + "loss": 1.2335, + "theoretical_loss": 3.330695192928064, + "tokens_seen": 3228303360 + }, + { + "epoch": 0.98, + "learning_rate": 1.778134403209629e-05, + "loss": 1.2834, + "theoretical_loss": 3.3306901151949617, + "tokens_seen": 3228368896 + }, + { + "epoch": 0.98, + "learning_rate": 1.776529588766299e-05, + "loss": 1.3025, + "theoretical_loss": 3.3306850375937977, + "tokens_seen": 3228434432 + }, + { + "epoch": 0.98, + "learning_rate": 1.774924774322969e-05, + "loss": 1.2511, + "theoretical_loss": 3.330679960124566, + "tokens_seen": 3228499968 + }, + { + "epoch": 0.98, + "learning_rate": 1.7733199598796392e-05, + "loss": 1.3627, + "theoretical_loss": 3.3306748827872603, + "tokens_seen": 3228565504 + }, + { + "epoch": 0.98, + "learning_rate": 1.771715145436309e-05, + "loss": 1.2492, + "theoretical_loss": 3.3306698055818744, + "tokens_seen": 3228631040 + }, + { + "epoch": 0.98, + "learning_rate": 1.770110330992979e-05, + "loss": 1.1795, + "theoretical_loss": 3.330664728508403, + "tokens_seen": 3228696576 + }, + { + "epoch": 0.98, + "learning_rate": 1.768505516549649e-05, + "loss": 1.2187, + "theoretical_loss": 3.330659651566839, + "tokens_seen": 3228762112 + }, + { + "epoch": 0.98, + "learning_rate": 1.766900702106319e-05, + "loss": 1.2991, + "theoretical_loss": 3.330654574757177, + "tokens_seen": 3228827648 + }, + { + "epoch": 0.98, + "learning_rate": 1.765295887662989e-05, + "loss": 1.3554, + "theoretical_loss": 3.33064949807941, + "tokens_seen": 3228893184 + }, + { + "epoch": 0.98, + "learning_rate": 1.763691073219659e-05, + "loss": 1.2521, + "theoretical_loss": 3.330644421533533, + "tokens_seen": 3228958720 + }, + { + "epoch": 0.98, + "learning_rate": 1.762086258776329e-05, + "loss": 1.2817, + "theoretical_loss": 3.3306393451195393, + "tokens_seen": 3229024256 + }, + { + "epoch": 0.98, + "learning_rate": 1.7604814443329993e-05, + "loss": 1.3376, + "theoretical_loss": 3.330634268837423, + "tokens_seen": 3229089792 + }, + { + "epoch": 0.98, + "learning_rate": 1.758876629889669e-05, + "loss": 1.2397, + "theoretical_loss": 3.3306291926871774, + "tokens_seen": 3229155328 + }, + { + "epoch": 0.98, + "learning_rate": 1.757271815446339e-05, + "loss": 1.2029, + "theoretical_loss": 3.330624116668797, + "tokens_seen": 3229220864 + }, + { + "epoch": 0.98, + "learning_rate": 1.7556670010030092e-05, + "loss": 1.3422, + "theoretical_loss": 3.3306190407822758, + "tokens_seen": 3229286400 + }, + { + "epoch": 0.98, + "learning_rate": 1.754062186559679e-05, + "loss": 1.2365, + "theoretical_loss": 3.3306139650276076, + "tokens_seen": 3229351936 + }, + { + "epoch": 0.98, + "learning_rate": 1.7524573721163493e-05, + "loss": 1.2478, + "theoretical_loss": 3.330608889404786, + "tokens_seen": 3229417472 + }, + { + "epoch": 0.98, + "learning_rate": 1.750852557673019e-05, + "loss": 1.2513, + "theoretical_loss": 3.330603813913805, + "tokens_seen": 3229483008 + }, + { + "epoch": 0.98, + "learning_rate": 1.749247743229689e-05, + "loss": 1.2155, + "theoretical_loss": 3.3305987385546585, + "tokens_seen": 3229548544 + }, + { + "epoch": 0.98, + "learning_rate": 1.7476429287863592e-05, + "loss": 1.2499, + "theoretical_loss": 3.330593663327341, + "tokens_seen": 3229614080 + }, + { + "epoch": 0.98, + "learning_rate": 1.746038114343029e-05, + "loss": 1.2812, + "theoretical_loss": 3.330588588231845, + "tokens_seen": 3229679616 + }, + { + "epoch": 0.98, + "learning_rate": 1.7444332998996993e-05, + "loss": 1.2945, + "theoretical_loss": 3.330583513268166, + "tokens_seen": 3229745152 + }, + { + "epoch": 0.98, + "learning_rate": 1.7428284854563694e-05, + "loss": 1.2308, + "theoretical_loss": 3.330578438436297, + "tokens_seen": 3229810688 + }, + { + "epoch": 0.98, + "learning_rate": 1.741223671013039e-05, + "loss": 1.2464, + "theoretical_loss": 3.3305733637362316, + "tokens_seen": 3229876224 + }, + { + "epoch": 0.98, + "learning_rate": 1.739618856569709e-05, + "loss": 1.3171, + "theoretical_loss": 3.3305682891679647, + "tokens_seen": 3229941760 + }, + { + "epoch": 0.98, + "learning_rate": 1.7380140421263792e-05, + "loss": 1.2536, + "theoretical_loss": 3.3305632147314896, + "tokens_seen": 3230007296 + }, + { + "epoch": 0.98, + "learning_rate": 1.7364092276830493e-05, + "loss": 1.2191, + "theoretical_loss": 3.3305581404268003, + "tokens_seen": 3230072832 + }, + { + "epoch": 0.98, + "learning_rate": 1.7348044132397193e-05, + "loss": 1.0923, + "theoretical_loss": 3.330553066253891, + "tokens_seen": 3230138368 + }, + { + "epoch": 0.98, + "learning_rate": 1.733199598796389e-05, + "loss": 1.2387, + "theoretical_loss": 3.3305479922127548, + "tokens_seen": 3230203904 + }, + { + "epoch": 0.98, + "learning_rate": 1.731594784353059e-05, + "loss": 1.2161, + "theoretical_loss": 3.330542918303386, + "tokens_seen": 3230269440 + }, + { + "epoch": 0.98, + "learning_rate": 1.7299899699097295e-05, + "loss": 1.1234, + "theoretical_loss": 3.3305378445257787, + "tokens_seen": 3230334976 + }, + { + "epoch": 0.98, + "learning_rate": 1.7283851554663993e-05, + "loss": 1.2433, + "theoretical_loss": 3.3305327708799273, + "tokens_seen": 3230400512 + }, + { + "epoch": 0.98, + "learning_rate": 1.7267803410230693e-05, + "loss": 1.3307, + "theoretical_loss": 3.3305276973658247, + "tokens_seen": 3230466048 + }, + { + "epoch": 0.98, + "learning_rate": 1.7251755265797394e-05, + "loss": 1.2117, + "theoretical_loss": 3.330522623983465, + "tokens_seen": 3230531584 + }, + { + "epoch": 0.98, + "learning_rate": 1.723570712136409e-05, + "loss": 1.3011, + "theoretical_loss": 3.3305175507328424, + "tokens_seen": 3230597120 + }, + { + "epoch": 0.98, + "learning_rate": 1.7219658976930795e-05, + "loss": 1.2684, + "theoretical_loss": 3.330512477613951, + "tokens_seen": 3230662656 + }, + { + "epoch": 0.98, + "learning_rate": 1.7203610832497492e-05, + "loss": 1.2227, + "theoretical_loss": 3.3305074046267844, + "tokens_seen": 3230728192 + }, + { + "epoch": 0.98, + "learning_rate": 1.7187562688064193e-05, + "loss": 1.2668, + "theoretical_loss": 3.3305023317713363, + "tokens_seen": 3230793728 + }, + { + "epoch": 0.98, + "learning_rate": 1.7171514543630894e-05, + "loss": 1.3155, + "theoretical_loss": 3.3304972590476014, + "tokens_seen": 3230859264 + }, + { + "epoch": 0.98, + "objective/train/docs_used": 1174146, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.5498201847076416, + "objective/train/theoretical_loss": 3.3304921864555723, + "objective/train/tokens_used": 3251384800, + "theoretical_loss": 3.3304921864555723, + "tokens_seen": 3230924800 + }, + { + "epoch": 0.98, + "learning_rate": 1.715546639919759e-05, + "loss": 1.4177, + "theoretical_loss": 3.3304921864555723, + "tokens_seen": 3230924800 + }, + { + "epoch": 0.98, + "learning_rate": 1.7139418254764295e-05, + "loss": 1.267, + "theoretical_loss": 3.330487113995244, + "tokens_seen": 3230990336 + }, + { + "epoch": 0.98, + "learning_rate": 1.7123370110330996e-05, + "loss": 1.3137, + "theoretical_loss": 3.3304820416666105, + "tokens_seen": 3231055872 + }, + { + "epoch": 0.98, + "learning_rate": 1.7107321965897693e-05, + "loss": 1.2764, + "theoretical_loss": 3.330476969469665, + "tokens_seen": 3231121408 + }, + { + "epoch": 0.98, + "learning_rate": 1.7091273821464393e-05, + "loss": 1.2269, + "theoretical_loss": 3.3304718974044016, + "tokens_seen": 3231186944 + }, + { + "epoch": 0.98, + "learning_rate": 1.7075225677031094e-05, + "loss": 1.2504, + "theoretical_loss": 3.3304668254708143, + "tokens_seen": 3231252480 + }, + { + "epoch": 0.98, + "learning_rate": 1.7059177532597795e-05, + "loss": 1.193, + "theoretical_loss": 3.3304617536688967, + "tokens_seen": 3231318016 + }, + { + "epoch": 0.98, + "learning_rate": 1.7043129388164495e-05, + "loss": 1.2294, + "theoretical_loss": 3.3304566819986436, + "tokens_seen": 3231383552 + }, + { + "epoch": 0.98, + "learning_rate": 1.7027081243731193e-05, + "loss": 1.3378, + "theoretical_loss": 3.3304516104600483, + "tokens_seen": 3231449088 + }, + { + "epoch": 0.98, + "learning_rate": 1.7011033099297893e-05, + "loss": 1.2396, + "theoretical_loss": 3.3304465390531046, + "tokens_seen": 3231514624 + }, + { + "epoch": 0.98, + "learning_rate": 1.6994984954864597e-05, + "loss": 1.1872, + "theoretical_loss": 3.3304414677778067, + "tokens_seen": 3231580160 + }, + { + "epoch": 0.98, + "learning_rate": 1.6978936810431294e-05, + "loss": 1.2533, + "theoretical_loss": 3.3304363966341484, + "tokens_seen": 3231645696 + }, + { + "epoch": 0.98, + "learning_rate": 1.6962888665997995e-05, + "loss": 1.2959, + "theoretical_loss": 3.3304313256221234, + "tokens_seen": 3231711232 + }, + { + "epoch": 0.98, + "learning_rate": 1.6946840521564696e-05, + "loss": 1.266, + "theoretical_loss": 3.330426254741726, + "tokens_seen": 3231776768 + }, + { + "epoch": 0.98, + "learning_rate": 1.6930792377131393e-05, + "loss": 1.3311, + "theoretical_loss": 3.33042118399295, + "tokens_seen": 3231842304 + }, + { + "epoch": 0.98, + "learning_rate": 1.6914744232698097e-05, + "loss": 1.2223, + "theoretical_loss": 3.3304161133757892, + "tokens_seen": 3231907840 + }, + { + "epoch": 0.98, + "learning_rate": 1.6898696088264794e-05, + "loss": 1.211, + "theoretical_loss": 3.3304110428902374, + "tokens_seen": 3231973376 + }, + { + "epoch": 0.98, + "learning_rate": 1.6882647943831495e-05, + "loss": 1.3041, + "theoretical_loss": 3.3304059725362887, + "tokens_seen": 3232038912 + }, + { + "epoch": 0.98, + "learning_rate": 1.6866599799398196e-05, + "loss": 1.3227, + "theoretical_loss": 3.330400902313937, + "tokens_seen": 3232104448 + }, + { + "epoch": 0.98, + "learning_rate": 1.6850551654964896e-05, + "loss": 1.3443, + "theoretical_loss": 3.3303958322231764, + "tokens_seen": 3232169984 + }, + { + "epoch": 0.98, + "learning_rate": 1.6834503510531597e-05, + "loss": 1.3267, + "theoretical_loss": 3.3303907622640003, + "tokens_seen": 3232235520 + }, + { + "epoch": 0.98, + "learning_rate": 1.6818455366098297e-05, + "loss": 1.255, + "theoretical_loss": 3.330385692436403, + "tokens_seen": 3232301056 + }, + { + "epoch": 0.98, + "learning_rate": 1.6802407221664995e-05, + "loss": 1.3321, + "theoretical_loss": 3.3303806227403783, + "tokens_seen": 3232366592 + }, + { + "epoch": 0.98, + "learning_rate": 1.6786359077231695e-05, + "loss": 1.2909, + "theoretical_loss": 3.3303755531759207, + "tokens_seen": 3232432128 + }, + { + "epoch": 0.98, + "learning_rate": 1.6770310932798396e-05, + "loss": 1.2492, + "theoretical_loss": 3.3303704837430232, + "tokens_seen": 3232497664 + }, + { + "epoch": 0.98, + "learning_rate": 1.6754262788365097e-05, + "loss": 1.2379, + "theoretical_loss": 3.33036541444168, + "tokens_seen": 3232563200 + }, + { + "epoch": 0.98, + "learning_rate": 1.6738214643931797e-05, + "loss": 1.2038, + "theoretical_loss": 3.3303603452718855, + "tokens_seen": 3232628736 + }, + { + "epoch": 0.98, + "learning_rate": 1.6722166499498494e-05, + "loss": 1.2398, + "theoretical_loss": 3.3303552762336333, + "tokens_seen": 3232694272 + }, + { + "epoch": 0.98, + "learning_rate": 1.6706118355065195e-05, + "loss": 1.2206, + "theoretical_loss": 3.3303502073269167, + "tokens_seen": 3232759808 + }, + { + "epoch": 0.98, + "learning_rate": 1.6690070210631896e-05, + "loss": 1.2918, + "theoretical_loss": 3.3303451385517304, + "tokens_seen": 3232825344 + }, + { + "epoch": 0.98, + "learning_rate": 1.6674022066198596e-05, + "loss": 1.2191, + "theoretical_loss": 3.3303400699080683, + "tokens_seen": 3232890880 + }, + { + "epoch": 0.98, + "learning_rate": 1.6657973921765297e-05, + "loss": 1.2182, + "theoretical_loss": 3.330335001395924, + "tokens_seen": 3232956416 + }, + { + "epoch": 0.98, + "learning_rate": 1.6641925777331994e-05, + "loss": 1.2473, + "theoretical_loss": 3.3303299330152916, + "tokens_seen": 3233021952 + }, + { + "epoch": 0.98, + "learning_rate": 1.6625877632898698e-05, + "loss": 1.2734, + "theoretical_loss": 3.330324864766165, + "tokens_seen": 3233087488 + }, + { + "epoch": 0.98, + "learning_rate": 1.66098294884654e-05, + "loss": 1.1507, + "theoretical_loss": 3.330319796648538, + "tokens_seen": 3233153024 + }, + { + "epoch": 0.98, + "learning_rate": 1.6593781344032096e-05, + "loss": 1.2389, + "theoretical_loss": 3.330314728662405, + "tokens_seen": 3233218560 + }, + { + "epoch": 0.98, + "learning_rate": 1.6577733199598797e-05, + "loss": 1.2153, + "theoretical_loss": 3.3303096608077594, + "tokens_seen": 3233284096 + }, + { + "epoch": 0.98, + "learning_rate": 1.6561685055165497e-05, + "loss": 1.3025, + "theoretical_loss": 3.3303045930845947, + "tokens_seen": 3233349632 + }, + { + "epoch": 0.98, + "learning_rate": 1.6545636910732198e-05, + "loss": 1.1615, + "theoretical_loss": 3.330299525492906, + "tokens_seen": 3233415168 + }, + { + "epoch": 0.98, + "learning_rate": 1.65295887662989e-05, + "loss": 1.3338, + "theoretical_loss": 3.3302944580326868, + "tokens_seen": 3233480704 + }, + { + "epoch": 0.98, + "learning_rate": 1.6513540621865596e-05, + "loss": 1.3235, + "theoretical_loss": 3.33028939070393, + "tokens_seen": 3233546240 + }, + { + "epoch": 0.98, + "learning_rate": 1.6497492477432297e-05, + "loss": 1.1929, + "theoretical_loss": 3.330284323506631, + "tokens_seen": 3233611776 + }, + { + "epoch": 0.98, + "learning_rate": 1.6481444332998997e-05, + "loss": 1.3771, + "theoretical_loss": 3.3302792564407833, + "tokens_seen": 3233677312 + }, + { + "epoch": 0.98, + "learning_rate": 1.6465396188565698e-05, + "loss": 1.2397, + "theoretical_loss": 3.3302741895063797, + "tokens_seen": 3233742848 + }, + { + "epoch": 0.98, + "learning_rate": 1.64493480441324e-05, + "loss": 1.2167, + "theoretical_loss": 3.330269122703416, + "tokens_seen": 3233808384 + }, + { + "epoch": 0.98, + "learning_rate": 1.64332998996991e-05, + "loss": 1.2848, + "theoretical_loss": 3.3302640560318846, + "tokens_seen": 3233873920 + }, + { + "epoch": 0.98, + "learning_rate": 1.6417251755265796e-05, + "loss": 1.3411, + "theoretical_loss": 3.3302589894917802, + "tokens_seen": 3233939456 + }, + { + "epoch": 0.98, + "learning_rate": 1.64012036108325e-05, + "loss": 1.2569, + "theoretical_loss": 3.3302539230830965, + "tokens_seen": 3234004992 + }, + { + "epoch": 0.98, + "learning_rate": 1.6385155466399198e-05, + "loss": 1.2126, + "theoretical_loss": 3.3302488568058277, + "tokens_seen": 3234070528 + }, + { + "epoch": 0.98, + "learning_rate": 1.6369107321965898e-05, + "loss": 1.2159, + "theoretical_loss": 3.330243790659967, + "tokens_seen": 3234136064 + }, + { + "epoch": 0.98, + "objective/train/docs_used": 1175262, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.2757314443588257, + "objective/train/theoretical_loss": 3.3302387246455094, + "objective/train/tokens_used": 3254661600, + "theoretical_loss": 3.3302387246455094, + "tokens_seen": 3234201600 + }, + { + "epoch": 0.98, + "learning_rate": 1.63530591775326e-05, + "loss": 1.2054, + "theoretical_loss": 3.3302387246455094, + "tokens_seen": 3234201600 + }, + { + "epoch": 0.98, + "learning_rate": 1.6337011033099296e-05, + "loss": 1.2344, + "theoretical_loss": 3.330233658762448, + "tokens_seen": 3234267136 + }, + { + "epoch": 0.98, + "learning_rate": 1.6320962888666e-05, + "loss": 1.2848, + "theoretical_loss": 3.3302285930107765, + "tokens_seen": 3234332672 + }, + { + "epoch": 0.98, + "learning_rate": 1.63049147442327e-05, + "loss": 1.193, + "theoretical_loss": 3.3302235273904897, + "tokens_seen": 3234398208 + }, + { + "epoch": 0.98, + "learning_rate": 1.6288866599799398e-05, + "loss": 1.3144, + "theoretical_loss": 3.3302184619015813, + "tokens_seen": 3234463744 + }, + { + "epoch": 0.98, + "learning_rate": 1.62728184553661e-05, + "loss": 1.2829, + "theoretical_loss": 3.330213396544045, + "tokens_seen": 3234529280 + }, + { + "epoch": 0.98, + "learning_rate": 1.62567703109328e-05, + "loss": 1.2295, + "theoretical_loss": 3.3302083313178743, + "tokens_seen": 3234594816 + }, + { + "epoch": 0.98, + "learning_rate": 1.62407221664995e-05, + "loss": 1.266, + "theoretical_loss": 3.330203266223064, + "tokens_seen": 3234660352 + }, + { + "epoch": 0.98, + "learning_rate": 1.62246740220662e-05, + "loss": 1.3462, + "theoretical_loss": 3.330198201259608, + "tokens_seen": 3234725888 + }, + { + "epoch": 0.98, + "learning_rate": 1.6208625877632898e-05, + "loss": 1.2194, + "theoretical_loss": 3.3301931364274995, + "tokens_seen": 3234791424 + }, + { + "epoch": 0.98, + "learning_rate": 1.61925777331996e-05, + "loss": 1.2578, + "theoretical_loss": 3.3301880717267327, + "tokens_seen": 3234856960 + }, + { + "epoch": 0.98, + "learning_rate": 1.6176529588766302e-05, + "loss": 1.158, + "theoretical_loss": 3.3301830071573018, + "tokens_seen": 3234922496 + }, + { + "epoch": 0.98, + "learning_rate": 1.6160481444333e-05, + "loss": 1.3064, + "theoretical_loss": 3.3301779427192004, + "tokens_seen": 3234988032 + }, + { + "epoch": 0.98, + "learning_rate": 1.61444332998997e-05, + "loss": 1.1236, + "theoretical_loss": 3.330172878412423, + "tokens_seen": 3235053568 + }, + { + "epoch": 0.98, + "learning_rate": 1.61283851554664e-05, + "loss": 1.2413, + "theoretical_loss": 3.330167814236963, + "tokens_seen": 3235119104 + }, + { + "epoch": 0.98, + "learning_rate": 1.6112337011033098e-05, + "loss": 1.2492, + "theoretical_loss": 3.3301627501928146, + "tokens_seen": 3235184640 + }, + { + "epoch": 0.98, + "learning_rate": 1.6096288866599802e-05, + "loss": 1.2936, + "theoretical_loss": 3.3301576862799713, + "tokens_seen": 3235250176 + }, + { + "epoch": 0.98, + "learning_rate": 1.60802407221665e-05, + "loss": 1.3319, + "theoretical_loss": 3.330152622498428, + "tokens_seen": 3235315712 + }, + { + "epoch": 0.98, + "learning_rate": 1.60641925777332e-05, + "loss": 1.2732, + "theoretical_loss": 3.3301475588481773, + "tokens_seen": 3235381248 + }, + { + "epoch": 0.98, + "learning_rate": 1.60481444332999e-05, + "loss": 1.2171, + "theoretical_loss": 3.3301424953292145, + "tokens_seen": 3235446784 + }, + { + "epoch": 0.98, + "learning_rate": 1.6032096288866598e-05, + "loss": 1.2086, + "theoretical_loss": 3.330137431941532, + "tokens_seen": 3235512320 + }, + { + "epoch": 0.98, + "learning_rate": 1.6016048144433302e-05, + "loss": 1.3165, + "theoretical_loss": 3.3301323686851254, + "tokens_seen": 3235577856 + }, + { + "epoch": 0.98, + "learning_rate": 1.6000000000000003e-05, + "loss": 1.2589, + "theoretical_loss": 3.3301273055599876, + "tokens_seen": 3235643392 + }, + { + "epoch": 0.98, + "learning_rate": 1.59839518555667e-05, + "loss": 1.183, + "theoretical_loss": 3.3301222425661123, + "tokens_seen": 3235708928 + }, + { + "epoch": 0.98, + "learning_rate": 1.59679037111334e-05, + "loss": 1.2083, + "theoretical_loss": 3.3301171797034947, + "tokens_seen": 3235774464 + }, + { + "epoch": 0.98, + "learning_rate": 1.59518555667001e-05, + "loss": 1.3275, + "theoretical_loss": 3.3301121169721277, + "tokens_seen": 3235840000 + }, + { + "epoch": 0.98, + "learning_rate": 1.5935807422266802e-05, + "loss": 1.3056, + "theoretical_loss": 3.3301070543720055, + "tokens_seen": 3235905536 + }, + { + "epoch": 0.98, + "learning_rate": 1.5919759277833502e-05, + "loss": 1.2045, + "theoretical_loss": 3.330101991903122, + "tokens_seen": 3235971072 + }, + { + "epoch": 0.98, + "learning_rate": 1.59037111334002e-05, + "loss": 1.3421, + "theoretical_loss": 3.330096929565471, + "tokens_seen": 3236036608 + }, + { + "epoch": 0.98, + "learning_rate": 1.58876629889669e-05, + "loss": 1.24, + "theoretical_loss": 3.3300918673590467, + "tokens_seen": 3236102144 + }, + { + "epoch": 0.98, + "learning_rate": 1.5871614844533604e-05, + "loss": 1.3365, + "theoretical_loss": 3.3300868052838433, + "tokens_seen": 3236167680 + }, + { + "epoch": 0.98, + "learning_rate": 1.58555667001003e-05, + "loss": 1.1865, + "theoretical_loss": 3.330081743339854, + "tokens_seen": 3236233216 + }, + { + "epoch": 0.98, + "learning_rate": 1.5839518555667002e-05, + "loss": 1.2787, + "theoretical_loss": 3.330076681527073, + "tokens_seen": 3236298752 + }, + { + "epoch": 0.98, + "learning_rate": 1.5823470411233703e-05, + "loss": 1.1955, + "theoretical_loss": 3.3300716198454947, + "tokens_seen": 3236364288 + }, + { + "epoch": 0.98, + "learning_rate": 1.58074222668004e-05, + "loss": 1.1736, + "theoretical_loss": 3.3300665582951128, + "tokens_seen": 3236429824 + }, + { + "epoch": 0.98, + "learning_rate": 1.5791374122367104e-05, + "loss": 1.3114, + "theoretical_loss": 3.3300614968759215, + "tokens_seen": 3236495360 + }, + { + "epoch": 0.98, + "learning_rate": 1.57753259779338e-05, + "loss": 1.1554, + "theoretical_loss": 3.3300564355879136, + "tokens_seen": 3236560896 + }, + { + "epoch": 0.98, + "learning_rate": 1.5759277833500502e-05, + "loss": 1.3466, + "theoretical_loss": 3.3300513744310845, + "tokens_seen": 3236626432 + }, + { + "epoch": 0.98, + "learning_rate": 1.5743229689067203e-05, + "loss": 1.2731, + "theoretical_loss": 3.3300463134054272, + "tokens_seen": 3236691968 + }, + { + "epoch": 0.98, + "learning_rate": 1.5727181544633903e-05, + "loss": 1.2698, + "theoretical_loss": 3.330041252510936, + "tokens_seen": 3236757504 + }, + { + "epoch": 0.98, + "learning_rate": 1.5711133400200604e-05, + "loss": 1.3222, + "theoretical_loss": 3.330036191747605, + "tokens_seen": 3236823040 + }, + { + "epoch": 0.98, + "learning_rate": 1.56950852557673e-05, + "loss": 1.286, + "theoretical_loss": 3.330031131115428, + "tokens_seen": 3236888576 + }, + { + "epoch": 0.98, + "learning_rate": 1.5679037111334002e-05, + "loss": 1.2262, + "theoretical_loss": 3.330026070614398, + "tokens_seen": 3236954112 + }, + { + "epoch": 0.98, + "learning_rate": 1.5662988966900702e-05, + "loss": 1.307, + "theoretical_loss": 3.3300210102445105, + "tokens_seen": 3237019648 + }, + { + "epoch": 0.98, + "learning_rate": 1.5646940822467403e-05, + "loss": 1.1672, + "theoretical_loss": 3.330015950005759, + "tokens_seen": 3237085184 + }, + { + "epoch": 0.98, + "learning_rate": 1.5630892678034104e-05, + "loss": 1.247, + "theoretical_loss": 3.330010889898137, + "tokens_seen": 3237150720 + }, + { + "epoch": 0.98, + "learning_rate": 1.5614844533600804e-05, + "loss": 1.2895, + "theoretical_loss": 3.330005829921639, + "tokens_seen": 3237216256 + }, + { + "epoch": 0.98, + "learning_rate": 1.55987963891675e-05, + "loss": 1.2666, + "theoretical_loss": 3.330000770076258, + "tokens_seen": 3237281792 + }, + { + "epoch": 0.98, + "learning_rate": 1.5582748244734202e-05, + "loss": 1.1635, + "theoretical_loss": 3.3299957103619886, + "tokens_seen": 3237347328 + }, + { + "epoch": 0.98, + "learning_rate": 1.5566700100300903e-05, + "loss": 1.3021, + "theoretical_loss": 3.329990650778825, + "tokens_seen": 3237412864 + }, + { + "epoch": 0.98, + "objective/train/docs_used": 1176230, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.2090409994125366, + "objective/train/theoretical_loss": 3.329985591326761, + "objective/train/tokens_used": 3257938400, + "theoretical_loss": 3.329985591326761, + "tokens_seen": 3237478400 + }, + { + "epoch": 0.98, + "learning_rate": 1.5550651955867603e-05, + "loss": 1.2107, + "theoretical_loss": 3.329985591326761, + "tokens_seen": 3237478400 + }, + { + "epoch": 0.98, + "learning_rate": 1.5534603811434304e-05, + "loss": 1.357, + "theoretical_loss": 3.3299805320057905, + "tokens_seen": 3237543936 + }, + { + "epoch": 0.98, + "learning_rate": 1.5518555667001e-05, + "loss": 1.2706, + "theoretical_loss": 3.329975472815907, + "tokens_seen": 3237609472 + }, + { + "epoch": 0.98, + "learning_rate": 1.5502507522567705e-05, + "loss": 1.2713, + "theoretical_loss": 3.329970413757105, + "tokens_seen": 3237675008 + }, + { + "epoch": 0.98, + "learning_rate": 1.5486459378134406e-05, + "loss": 1.3048, + "theoretical_loss": 3.329965354829378, + "tokens_seen": 3237740544 + }, + { + "epoch": 0.98, + "learning_rate": 1.5470411233701103e-05, + "loss": 1.245, + "theoretical_loss": 3.3299602960327204, + "tokens_seen": 3237806080 + }, + { + "epoch": 0.98, + "learning_rate": 1.5454363089267804e-05, + "loss": 1.1797, + "theoretical_loss": 3.329955237367126, + "tokens_seen": 3237871616 + }, + { + "epoch": 0.98, + "learning_rate": 1.5438314944834505e-05, + "loss": 1.2069, + "theoretical_loss": 3.329950178832589, + "tokens_seen": 3237937152 + }, + { + "epoch": 0.98, + "learning_rate": 1.5422266800401205e-05, + "loss": 1.2867, + "theoretical_loss": 3.3299451204291026, + "tokens_seen": 3238002688 + }, + { + "epoch": 0.98, + "learning_rate": 1.5406218655967906e-05, + "loss": 1.2422, + "theoretical_loss": 3.3299400621566613, + "tokens_seen": 3238068224 + }, + { + "epoch": 0.98, + "learning_rate": 1.5390170511534603e-05, + "loss": 1.1756, + "theoretical_loss": 3.329935004015259, + "tokens_seen": 3238133760 + }, + { + "epoch": 0.98, + "learning_rate": 1.5374122367101304e-05, + "loss": 1.194, + "theoretical_loss": 3.32992994600489, + "tokens_seen": 3238199296 + }, + { + "epoch": 0.98, + "learning_rate": 1.5358074222668004e-05, + "loss": 1.1536, + "theoretical_loss": 3.3299248881255474, + "tokens_seen": 3238264832 + }, + { + "epoch": 0.98, + "learning_rate": 1.5342026078234705e-05, + "loss": 1.1957, + "theoretical_loss": 3.329919830377226, + "tokens_seen": 3238330368 + }, + { + "epoch": 0.98, + "learning_rate": 1.5325977933801406e-05, + "loss": 1.1794, + "theoretical_loss": 3.3299147727599188, + "tokens_seen": 3238395904 + }, + { + "epoch": 0.98, + "learning_rate": 1.5309929789368106e-05, + "loss": 1.1457, + "theoretical_loss": 3.329909715273621, + "tokens_seen": 3238461440 + }, + { + "epoch": 0.98, + "learning_rate": 1.5293881644934803e-05, + "loss": 1.2996, + "theoretical_loss": 3.3299046579183256, + "tokens_seen": 3238526976 + }, + { + "epoch": 0.98, + "learning_rate": 1.5277833500501504e-05, + "loss": 1.1711, + "theoretical_loss": 3.329899600694027, + "tokens_seen": 3238592512 + }, + { + "epoch": 0.98, + "learning_rate": 1.5261785356068205e-05, + "loss": 1.185, + "theoretical_loss": 3.329894543600719, + "tokens_seen": 3238658048 + }, + { + "epoch": 0.98, + "learning_rate": 1.5245737211634905e-05, + "loss": 1.1922, + "theoretical_loss": 3.3298894866383955, + "tokens_seen": 3238723584 + }, + { + "epoch": 0.98, + "learning_rate": 1.5229689067201604e-05, + "loss": 1.3295, + "theoretical_loss": 3.3298844298070507, + "tokens_seen": 3238789120 + }, + { + "epoch": 0.98, + "learning_rate": 1.5213640922768305e-05, + "loss": 1.2325, + "theoretical_loss": 3.329879373106678, + "tokens_seen": 3238854656 + }, + { + "epoch": 0.98, + "learning_rate": 1.5197592778335007e-05, + "loss": 1.2239, + "theoretical_loss": 3.329874316537272, + "tokens_seen": 3238920192 + }, + { + "epoch": 0.98, + "learning_rate": 1.5181544633901706e-05, + "loss": 1.2336, + "theoretical_loss": 3.3298692600988264, + "tokens_seen": 3238985728 + }, + { + "epoch": 0.98, + "learning_rate": 1.5165496489468405e-05, + "loss": 1.3155, + "theoretical_loss": 3.3298642037913355, + "tokens_seen": 3239051264 + }, + { + "epoch": 0.98, + "learning_rate": 1.5149448345035106e-05, + "loss": 1.2498, + "theoretical_loss": 3.3298591476147923, + "tokens_seen": 3239116800 + }, + { + "epoch": 0.98, + "learning_rate": 1.5133400200601805e-05, + "loss": 1.2003, + "theoretical_loss": 3.3298540915691914, + "tokens_seen": 3239182336 + }, + { + "epoch": 0.98, + "learning_rate": 1.5117352056168507e-05, + "loss": 1.1802, + "theoretical_loss": 3.3298490356545276, + "tokens_seen": 3239247872 + }, + { + "epoch": 0.98, + "learning_rate": 1.5101303911735206e-05, + "loss": 1.2528, + "theoretical_loss": 3.3298439798707933, + "tokens_seen": 3239313408 + }, + { + "epoch": 0.98, + "learning_rate": 1.5085255767301907e-05, + "loss": 1.298, + "theoretical_loss": 3.329838924217983, + "tokens_seen": 3239378944 + }, + { + "epoch": 0.98, + "learning_rate": 1.5069207622868606e-05, + "loss": 1.2675, + "theoretical_loss": 3.3298338686960913, + "tokens_seen": 3239444480 + }, + { + "epoch": 0.98, + "learning_rate": 1.5053159478435305e-05, + "loss": 1.2315, + "theoretical_loss": 3.3298288133051113, + "tokens_seen": 3239510016 + }, + { + "epoch": 0.98, + "learning_rate": 1.5037111334002007e-05, + "loss": 1.2299, + "theoretical_loss": 3.3298237580450376, + "tokens_seen": 3239575552 + }, + { + "epoch": 0.98, + "learning_rate": 1.5021063189568707e-05, + "loss": 1.2499, + "theoretical_loss": 3.3298187029158637, + "tokens_seen": 3239641088 + }, + { + "epoch": 0.98, + "learning_rate": 1.5005015045135406e-05, + "loss": 1.1562, + "theoretical_loss": 3.329813647917584, + "tokens_seen": 3239706624 + }, + { + "epoch": 0.98, + "learning_rate": 1.4988966900702105e-05, + "loss": 1.2941, + "theoretical_loss": 3.329808593050192, + "tokens_seen": 3239772160 + }, + { + "epoch": 0.98, + "learning_rate": 1.4972918756268808e-05, + "loss": 1.2819, + "theoretical_loss": 3.329803538313682, + "tokens_seen": 3239837696 + }, + { + "epoch": 0.98, + "learning_rate": 1.4956870611835508e-05, + "loss": 1.1346, + "theoretical_loss": 3.329798483708048, + "tokens_seen": 3239903232 + }, + { + "epoch": 0.98, + "learning_rate": 1.4940822467402207e-05, + "loss": 1.2332, + "theoretical_loss": 3.3297934292332836, + "tokens_seen": 3239968768 + }, + { + "epoch": 0.98, + "learning_rate": 1.4924774322968906e-05, + "loss": 1.2684, + "theoretical_loss": 3.3297883748893833, + "tokens_seen": 3240034304 + }, + { + "epoch": 0.98, + "learning_rate": 1.4908726178535607e-05, + "loss": 1.2408, + "theoretical_loss": 3.3297833206763405, + "tokens_seen": 3240099840 + }, + { + "epoch": 0.98, + "learning_rate": 1.489267803410231e-05, + "loss": 1.1948, + "theoretical_loss": 3.3297782665941495, + "tokens_seen": 3240165376 + }, + { + "epoch": 0.98, + "learning_rate": 1.4876629889669008e-05, + "loss": 1.2881, + "theoretical_loss": 3.329773212642804, + "tokens_seen": 3240230912 + }, + { + "epoch": 0.98, + "learning_rate": 1.4860581745235707e-05, + "loss": 1.2297, + "theoretical_loss": 3.329768158822298, + "tokens_seen": 3240296448 + }, + { + "epoch": 0.98, + "learning_rate": 1.4844533600802408e-05, + "loss": 1.2907, + "theoretical_loss": 3.329763105132626, + "tokens_seen": 3240361984 + }, + { + "epoch": 0.98, + "learning_rate": 1.4828485456369107e-05, + "loss": 1.2282, + "theoretical_loss": 3.3297580515737817, + "tokens_seen": 3240427520 + }, + { + "epoch": 0.98, + "learning_rate": 1.4812437311935809e-05, + "loss": 1.1955, + "theoretical_loss": 3.3297529981457585, + "tokens_seen": 3240493056 + }, + { + "epoch": 0.98, + "learning_rate": 1.4796389167502508e-05, + "loss": 1.3768, + "theoretical_loss": 3.329747944848551, + "tokens_seen": 3240558592 + }, + { + "epoch": 0.98, + "learning_rate": 1.4780341023069209e-05, + "loss": 1.1965, + "theoretical_loss": 3.3297428916821534, + "tokens_seen": 3240624128 + }, + { + "epoch": 0.98, + "learning_rate": 1.4764292878635907e-05, + "loss": 1.223, + "theoretical_loss": 3.3297378386465586, + "tokens_seen": 3240689664 + }, + { + "epoch": 0.98, + "objective/train/docs_used": 1177395, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.2961466312408447, + "objective/train/theoretical_loss": 3.329732785741762, + "objective/train/tokens_used": 3261215200, + "theoretical_loss": 3.329732785741762, + "tokens_seen": 3240755200 + }, + { + "epoch": 0.98, + "learning_rate": 1.474824473420261e-05, + "loss": 1.2662, + "theoretical_loss": 3.329732785741762, + "tokens_seen": 3240755200 + }, + { + "epoch": 0.98, + "learning_rate": 1.4732196589769309e-05, + "loss": 1.2394, + "theoretical_loss": 3.329727732967756, + "tokens_seen": 3240820736 + }, + { + "epoch": 0.98, + "learning_rate": 1.471614844533601e-05, + "loss": 1.2731, + "theoretical_loss": 3.3297226803245357, + "tokens_seen": 3240886272 + }, + { + "epoch": 0.98, + "learning_rate": 1.4700100300902708e-05, + "loss": 1.2819, + "theoretical_loss": 3.329717627812095, + "tokens_seen": 3240951808 + }, + { + "epoch": 0.98, + "learning_rate": 1.4684052156469407e-05, + "loss": 1.3096, + "theoretical_loss": 3.329712575430427, + "tokens_seen": 3241017344 + }, + { + "epoch": 0.98, + "learning_rate": 1.466800401203611e-05, + "loss": 1.3284, + "theoretical_loss": 3.3297075231795263, + "tokens_seen": 3241082880 + }, + { + "epoch": 0.98, + "learning_rate": 1.465195586760281e-05, + "loss": 1.1686, + "theoretical_loss": 3.329702471059387, + "tokens_seen": 3241148416 + }, + { + "epoch": 0.98, + "learning_rate": 1.463590772316951e-05, + "loss": 1.3816, + "theoretical_loss": 3.3296974190700035, + "tokens_seen": 3241213952 + }, + { + "epoch": 0.98, + "learning_rate": 1.4619859578736208e-05, + "loss": 1.2854, + "theoretical_loss": 3.3296923672113685, + "tokens_seen": 3241279488 + }, + { + "epoch": 0.98, + "learning_rate": 1.4603811434302909e-05, + "loss": 1.1777, + "theoretical_loss": 3.329687315483477, + "tokens_seen": 3241345024 + }, + { + "epoch": 0.98, + "learning_rate": 1.4587763289869611e-05, + "loss": 1.3208, + "theoretical_loss": 3.3296822638863226, + "tokens_seen": 3241410560 + }, + { + "epoch": 0.98, + "learning_rate": 1.457171514543631e-05, + "loss": 1.2909, + "theoretical_loss": 3.3296772124198992, + "tokens_seen": 3241476096 + }, + { + "epoch": 0.98, + "learning_rate": 1.4555667001003009e-05, + "loss": 1.1845, + "theoretical_loss": 3.329672161084201, + "tokens_seen": 3241541632 + }, + { + "epoch": 0.98, + "learning_rate": 1.453961885656971e-05, + "loss": 1.3783, + "theoretical_loss": 3.329667109879222, + "tokens_seen": 3241607168 + }, + { + "epoch": 0.98, + "learning_rate": 1.4523570712136412e-05, + "loss": 1.3415, + "theoretical_loss": 3.3296620588049555, + "tokens_seen": 3241672704 + }, + { + "epoch": 0.98, + "learning_rate": 1.4507522567703111e-05, + "loss": 1.2591, + "theoretical_loss": 3.3296570078613965, + "tokens_seen": 3241738240 + }, + { + "epoch": 0.98, + "learning_rate": 1.449147442326981e-05, + "loss": 1.1963, + "theoretical_loss": 3.3296519570485383, + "tokens_seen": 3241803776 + }, + { + "epoch": 0.98, + "learning_rate": 1.447542627883651e-05, + "loss": 1.2944, + "theoretical_loss": 3.329646906366375, + "tokens_seen": 3241869312 + }, + { + "epoch": 0.98, + "learning_rate": 1.445937813440321e-05, + "loss": 1.1436, + "theoretical_loss": 3.329641855814901, + "tokens_seen": 3241934848 + }, + { + "epoch": 0.98, + "learning_rate": 1.4443329989969912e-05, + "loss": 1.1443, + "theoretical_loss": 3.3296368053941094, + "tokens_seen": 3242000384 + }, + { + "epoch": 0.98, + "learning_rate": 1.442728184553661e-05, + "loss": 1.2738, + "theoretical_loss": 3.329631755103995, + "tokens_seen": 3242065920 + }, + { + "epoch": 0.98, + "learning_rate": 1.4411233701103311e-05, + "loss": 1.2667, + "theoretical_loss": 3.3296267049445514, + "tokens_seen": 3242131456 + }, + { + "epoch": 0.98, + "learning_rate": 1.439518555667001e-05, + "loss": 1.3523, + "theoretical_loss": 3.3296216549157727, + "tokens_seen": 3242196992 + }, + { + "epoch": 0.98, + "learning_rate": 1.437913741223671e-05, + "loss": 1.1717, + "theoretical_loss": 3.3296166050176526, + "tokens_seen": 3242262528 + }, + { + "epoch": 0.98, + "learning_rate": 1.4363089267803411e-05, + "loss": 1.2679, + "theoretical_loss": 3.3296115552501853, + "tokens_seen": 3242328064 + }, + { + "epoch": 0.98, + "learning_rate": 1.4347041123370112e-05, + "loss": 1.1877, + "theoretical_loss": 3.329606505613365, + "tokens_seen": 3242393600 + }, + { + "epoch": 0.98, + "learning_rate": 1.4330992978936811e-05, + "loss": 1.2554, + "theoretical_loss": 3.3296014561071856, + "tokens_seen": 3242459136 + }, + { + "epoch": 0.98, + "learning_rate": 1.431494483450351e-05, + "loss": 1.2429, + "theoretical_loss": 3.329596406731641, + "tokens_seen": 3242524672 + }, + { + "epoch": 0.98, + "learning_rate": 1.4298896690070212e-05, + "loss": 1.2616, + "theoretical_loss": 3.3295913574867244, + "tokens_seen": 3242590208 + }, + { + "epoch": 0.98, + "learning_rate": 1.4282848545636911e-05, + "loss": 1.2225, + "theoretical_loss": 3.329586308372431, + "tokens_seen": 3242655744 + }, + { + "epoch": 0.98, + "learning_rate": 1.4266800401203612e-05, + "loss": 1.3227, + "theoretical_loss": 3.329581259388754, + "tokens_seen": 3242721280 + }, + { + "epoch": 0.98, + "learning_rate": 1.425075225677031e-05, + "loss": 1.3191, + "theoretical_loss": 3.329576210535688, + "tokens_seen": 3242786816 + }, + { + "epoch": 0.98, + "learning_rate": 1.423470411233701e-05, + "loss": 1.2136, + "theoretical_loss": 3.3295711618132264, + "tokens_seen": 3242852352 + }, + { + "epoch": 0.98, + "learning_rate": 1.4218655967903712e-05, + "loss": 1.1407, + "theoretical_loss": 3.3295661132213636, + "tokens_seen": 3242917888 + }, + { + "epoch": 0.98, + "learning_rate": 1.4202607823470413e-05, + "loss": 1.2297, + "theoretical_loss": 3.3295610647600933, + "tokens_seen": 3242983424 + }, + { + "epoch": 0.98, + "learning_rate": 1.4186559679037112e-05, + "loss": 1.2306, + "theoretical_loss": 3.32955601642941, + "tokens_seen": 3243048960 + }, + { + "epoch": 0.98, + "learning_rate": 1.417051153460381e-05, + "loss": 1.1691, + "theoretical_loss": 3.3295509682293067, + "tokens_seen": 3243114496 + }, + { + "epoch": 0.98, + "learning_rate": 1.4154463390170511e-05, + "loss": 1.3287, + "theoretical_loss": 3.329545920159778, + "tokens_seen": 3243180032 + }, + { + "epoch": 0.98, + "learning_rate": 1.4138415245737214e-05, + "loss": 1.1743, + "theoretical_loss": 3.329540872220818, + "tokens_seen": 3243245568 + }, + { + "epoch": 0.98, + "learning_rate": 1.4122367101303913e-05, + "loss": 1.1938, + "theoretical_loss": 3.329535824412421, + "tokens_seen": 3243311104 + }, + { + "epoch": 0.98, + "learning_rate": 1.4106318956870611e-05, + "loss": 1.2305, + "theoretical_loss": 3.32953077673458, + "tokens_seen": 3243376640 + }, + { + "epoch": 0.98, + "learning_rate": 1.4090270812437312e-05, + "loss": 1.2807, + "theoretical_loss": 3.3295257291872895, + "tokens_seen": 3243442176 + }, + { + "epoch": 0.98, + "learning_rate": 1.4074222668004014e-05, + "loss": 1.256, + "theoretical_loss": 3.3295206817705436, + "tokens_seen": 3243507712 + }, + { + "epoch": 0.98, + "learning_rate": 1.4058174523570713e-05, + "loss": 1.2633, + "theoretical_loss": 3.329515634484336, + "tokens_seen": 3243573248 + }, + { + "epoch": 0.98, + "learning_rate": 1.4042126379137412e-05, + "loss": 1.2764, + "theoretical_loss": 3.329510587328661, + "tokens_seen": 3243638784 + }, + { + "epoch": 0.98, + "learning_rate": 1.4026078234704113e-05, + "loss": 1.13, + "theoretical_loss": 3.3295055403035123, + "tokens_seen": 3243704320 + }, + { + "epoch": 0.98, + "learning_rate": 1.4010030090270812e-05, + "loss": 1.2451, + "theoretical_loss": 3.329500493408884, + "tokens_seen": 3243769856 + }, + { + "epoch": 0.98, + "learning_rate": 1.3993981945837514e-05, + "loss": 1.2495, + "theoretical_loss": 3.3294954466447706, + "tokens_seen": 3243835392 + }, + { + "epoch": 0.98, + "learning_rate": 1.3977933801404213e-05, + "loss": 1.239, + "theoretical_loss": 3.329490400011165, + "tokens_seen": 3243900928 + }, + { + "epoch": 0.98, + "learning_rate": 1.3961885656970914e-05, + "loss": 1.2829, + "theoretical_loss": 3.329485353508062, + "tokens_seen": 3243966464 + }, + { + "epoch": 0.98, + "objective/train/docs_used": 1178511, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.2130179405212402, + "objective/train/theoretical_loss": 3.3294803071354555, + "objective/train/tokens_used": 3264492000, + "theoretical_loss": 3.3294803071354555, + "tokens_seen": 3244032000 + }, + { + "epoch": 0.98, + "learning_rate": 1.3945837512537613e-05, + "loss": 1.2614, + "theoretical_loss": 3.3294803071354555, + "tokens_seen": 3244032000 + }, + { + "epoch": 0.98, + "learning_rate": 1.3929789368104312e-05, + "loss": 1.2557, + "theoretical_loss": 3.3294752608933393, + "tokens_seen": 3244097536 + }, + { + "epoch": 0.98, + "learning_rate": 1.3913741223671014e-05, + "loss": 1.3066, + "theoretical_loss": 3.3294702147817077, + "tokens_seen": 3244163072 + }, + { + "epoch": 0.98, + "learning_rate": 1.3897693079237715e-05, + "loss": 1.1614, + "theoretical_loss": 3.329465168800554, + "tokens_seen": 3244228608 + }, + { + "epoch": 0.98, + "learning_rate": 1.3881644934804414e-05, + "loss": 1.2714, + "theoretical_loss": 3.3294601229498726, + "tokens_seen": 3244294144 + }, + { + "epoch": 0.98, + "learning_rate": 1.3865596790371113e-05, + "loss": 1.2508, + "theoretical_loss": 3.329455077229658, + "tokens_seen": 3244359680 + }, + { + "epoch": 0.98, + "learning_rate": 1.3849548645937815e-05, + "loss": 1.2469, + "theoretical_loss": 3.3294500316399036, + "tokens_seen": 3244425216 + }, + { + "epoch": 0.98, + "learning_rate": 1.3833500501504515e-05, + "loss": 1.2953, + "theoretical_loss": 3.3294449861806035, + "tokens_seen": 3244490752 + }, + { + "epoch": 0.98, + "learning_rate": 1.3817452357071214e-05, + "loss": 1.1765, + "theoretical_loss": 3.3294399408517514, + "tokens_seen": 3244556288 + }, + { + "epoch": 0.98, + "learning_rate": 1.3801404212637913e-05, + "loss": 1.2879, + "theoretical_loss": 3.329434895653342, + "tokens_seen": 3244621824 + }, + { + "epoch": 0.98, + "learning_rate": 1.3785356068204614e-05, + "loss": 1.2111, + "theoretical_loss": 3.3294298505853686, + "tokens_seen": 3244687360 + }, + { + "epoch": 0.98, + "learning_rate": 1.3769307923771316e-05, + "loss": 1.231, + "theoretical_loss": 3.3294248056478257, + "tokens_seen": 3244752896 + }, + { + "epoch": 0.98, + "learning_rate": 1.3753259779338015e-05, + "loss": 1.1969, + "theoretical_loss": 3.329419760840707, + "tokens_seen": 3244818432 + }, + { + "epoch": 0.98, + "learning_rate": 1.3737211634904714e-05, + "loss": 1.2104, + "theoretical_loss": 3.3294147161640066, + "tokens_seen": 3244883968 + }, + { + "epoch": 0.98, + "learning_rate": 1.3721163490471415e-05, + "loss": 1.1433, + "theoretical_loss": 3.3294096716177184, + "tokens_seen": 3244949504 + }, + { + "epoch": 0.98, + "learning_rate": 1.3705115346038114e-05, + "loss": 1.1899, + "theoretical_loss": 3.3294046272018365, + "tokens_seen": 3245015040 + }, + { + "epoch": 0.98, + "learning_rate": 1.3689067201604816e-05, + "loss": 1.2584, + "theoretical_loss": 3.3293995829163547, + "tokens_seen": 3245080576 + }, + { + "epoch": 0.98, + "learning_rate": 1.3673019057171515e-05, + "loss": 1.2992, + "theoretical_loss": 3.3293945387612673, + "tokens_seen": 3245146112 + }, + { + "epoch": 0.98, + "learning_rate": 1.3656970912738216e-05, + "loss": 1.2533, + "theoretical_loss": 3.3293894947365685, + "tokens_seen": 3245211648 + }, + { + "epoch": 0.98, + "learning_rate": 1.3640922768304915e-05, + "loss": 1.2078, + "theoretical_loss": 3.3293844508422517, + "tokens_seen": 3245277184 + }, + { + "epoch": 0.98, + "learning_rate": 1.3624874623871617e-05, + "loss": 1.2313, + "theoretical_loss": 3.329379407078311, + "tokens_seen": 3245342720 + }, + { + "epoch": 0.98, + "learning_rate": 1.3608826479438316e-05, + "loss": 1.3069, + "theoretical_loss": 3.3293743634447406, + "tokens_seen": 3245408256 + }, + { + "epoch": 0.98, + "learning_rate": 1.3592778335005017e-05, + "loss": 1.2655, + "theoretical_loss": 3.3293693199415344, + "tokens_seen": 3245473792 + }, + { + "epoch": 0.98, + "learning_rate": 1.3576730190571715e-05, + "loss": 1.2238, + "theoretical_loss": 3.3293642765686866, + "tokens_seen": 3245539328 + }, + { + "epoch": 0.98, + "learning_rate": 1.3560682046138414e-05, + "loss": 1.1893, + "theoretical_loss": 3.329359233326191, + "tokens_seen": 3245604864 + }, + { + "epoch": 0.98, + "learning_rate": 1.3544633901705117e-05, + "loss": 1.2191, + "theoretical_loss": 3.3293541902140413, + "tokens_seen": 3245670400 + }, + { + "epoch": 0.98, + "learning_rate": 1.3528585757271817e-05, + "loss": 1.2107, + "theoretical_loss": 3.3293491472322323, + "tokens_seen": 3245735936 + }, + { + "epoch": 0.98, + "learning_rate": 1.3512537612838516e-05, + "loss": 1.247, + "theoretical_loss": 3.3293441043807572, + "tokens_seen": 3245801472 + }, + { + "epoch": 0.98, + "learning_rate": 1.3496489468405215e-05, + "loss": 1.2498, + "theoretical_loss": 3.329339061659611, + "tokens_seen": 3245867008 + }, + { + "epoch": 0.98, + "learning_rate": 1.3480441323971916e-05, + "loss": 1.2145, + "theoretical_loss": 3.329334019068786, + "tokens_seen": 3245932544 + }, + { + "epoch": 0.98, + "learning_rate": 1.3464393179538618e-05, + "loss": 1.2024, + "theoretical_loss": 3.329328976608278, + "tokens_seen": 3245998080 + }, + { + "epoch": 0.98, + "learning_rate": 1.3448345035105317e-05, + "loss": 1.3035, + "theoretical_loss": 3.32932393427808, + "tokens_seen": 3246063616 + }, + { + "epoch": 0.98, + "learning_rate": 1.3432296890672016e-05, + "loss": 1.3633, + "theoretical_loss": 3.329318892078186, + "tokens_seen": 3246129152 + }, + { + "epoch": 0.98, + "learning_rate": 1.3416248746238717e-05, + "loss": 1.2666, + "theoretical_loss": 3.3293138500085906, + "tokens_seen": 3246194688 + }, + { + "epoch": 0.98, + "learning_rate": 1.3400200601805419e-05, + "loss": 1.2614, + "theoretical_loss": 3.329308808069287, + "tokens_seen": 3246260224 + }, + { + "epoch": 0.98, + "learning_rate": 1.3384152457372118e-05, + "loss": 1.2407, + "theoretical_loss": 3.3293037662602702, + "tokens_seen": 3246325760 + }, + { + "epoch": 0.98, + "learning_rate": 1.3368104312938817e-05, + "loss": 1.3561, + "theoretical_loss": 3.3292987245815335, + "tokens_seen": 3246391296 + }, + { + "epoch": 0.98, + "learning_rate": 1.3352056168505518e-05, + "loss": 1.3021, + "theoretical_loss": 3.329293683033071, + "tokens_seen": 3246456832 + }, + { + "epoch": 0.98, + "learning_rate": 1.3336008024072217e-05, + "loss": 1.1768, + "theoretical_loss": 3.329288641614877, + "tokens_seen": 3246522368 + }, + { + "epoch": 0.98, + "learning_rate": 1.3319959879638919e-05, + "loss": 1.2613, + "theoretical_loss": 3.3292836003269444, + "tokens_seen": 3246587904 + }, + { + "epoch": 0.98, + "learning_rate": 1.3303911735205618e-05, + "loss": 1.2543, + "theoretical_loss": 3.3292785591692686, + "tokens_seen": 3246653440 + }, + { + "epoch": 0.98, + "learning_rate": 1.3287863590772317e-05, + "loss": 1.2371, + "theoretical_loss": 3.3292735181418434, + "tokens_seen": 3246718976 + }, + { + "epoch": 0.98, + "learning_rate": 1.3271815446339017e-05, + "loss": 1.342, + "theoretical_loss": 3.329268477244662, + "tokens_seen": 3246784512 + }, + { + "epoch": 0.98, + "learning_rate": 1.3255767301905716e-05, + "loss": 1.2127, + "theoretical_loss": 3.329263436477719, + "tokens_seen": 3246850048 + }, + { + "epoch": 0.98, + "learning_rate": 1.3239719157472419e-05, + "loss": 1.2134, + "theoretical_loss": 3.3292583958410082, + "tokens_seen": 3246915584 + }, + { + "epoch": 0.98, + "learning_rate": 1.3223671013039118e-05, + "loss": 1.36, + "theoretical_loss": 3.329253355334524, + "tokens_seen": 3246981120 + }, + { + "epoch": 0.98, + "learning_rate": 1.3207622868605818e-05, + "loss": 1.151, + "theoretical_loss": 3.32924831495826, + "tokens_seen": 3247046656 + }, + { + "epoch": 0.98, + "learning_rate": 1.3191574724172517e-05, + "loss": 1.2642, + "theoretical_loss": 3.3292432747122103, + "tokens_seen": 3247112192 + }, + { + "epoch": 0.98, + "learning_rate": 1.317552657973922e-05, + "loss": 1.3224, + "theoretical_loss": 3.329238234596369, + "tokens_seen": 3247177728 + }, + { + "epoch": 0.98, + "learning_rate": 1.3159478435305918e-05, + "loss": 1.321, + "theoretical_loss": 3.3292331946107296, + "tokens_seen": 3247243264 + }, + { + "epoch": 0.98, + "objective/train/docs_used": 1179779, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.2410207986831665, + "objective/train/theoretical_loss": 3.3292281547552873, + "objective/train/tokens_used": 3267768800, + "theoretical_loss": 3.3292281547552873, + "tokens_seen": 3247308800 + }, + { + "epoch": 0.98, + "learning_rate": 1.3143430290872619e-05, + "loss": 1.2665, + "theoretical_loss": 3.3292281547552873, + "tokens_seen": 3247308800 + }, + { + "epoch": 0.98, + "learning_rate": 1.3127382146439318e-05, + "loss": 1.268, + "theoretical_loss": 3.3292231150300347, + "tokens_seen": 3247374336 + }, + { + "epoch": 0.98, + "learning_rate": 1.3111334002006017e-05, + "loss": 1.254, + "theoretical_loss": 3.3292180754349667, + "tokens_seen": 3247439872 + }, + { + "epoch": 0.98, + "learning_rate": 1.309528585757272e-05, + "loss": 1.3312, + "theoretical_loss": 3.329213035970077, + "tokens_seen": 3247505408 + }, + { + "epoch": 0.98, + "learning_rate": 1.307923771313942e-05, + "loss": 1.3022, + "theoretical_loss": 3.32920799663536, + "tokens_seen": 3247570944 + }, + { + "epoch": 0.98, + "learning_rate": 1.3063189568706119e-05, + "loss": 1.2838, + "theoretical_loss": 3.329202957430809, + "tokens_seen": 3247636480 + }, + { + "epoch": 0.98, + "learning_rate": 1.3047141424272818e-05, + "loss": 1.2098, + "theoretical_loss": 3.329197918356418, + "tokens_seen": 3247702016 + }, + { + "epoch": 0.98, + "learning_rate": 1.3031093279839518e-05, + "loss": 1.2179, + "theoretical_loss": 3.329192879412182, + "tokens_seen": 3247767552 + }, + { + "epoch": 0.98, + "learning_rate": 1.301504513540622e-05, + "loss": 1.3787, + "theoretical_loss": 3.3291878405980944, + "tokens_seen": 3247833088 + }, + { + "epoch": 0.98, + "learning_rate": 1.299899699097292e-05, + "loss": 1.2526, + "theoretical_loss": 3.3291828019141496, + "tokens_seen": 3247898624 + }, + { + "epoch": 0.98, + "learning_rate": 1.2982948846539619e-05, + "loss": 1.261, + "theoretical_loss": 3.3291777633603408, + "tokens_seen": 3247964160 + }, + { + "epoch": 0.98, + "learning_rate": 1.296690070210632e-05, + "loss": 1.2464, + "theoretical_loss": 3.329172724936662, + "tokens_seen": 3248029696 + }, + { + "epoch": 0.98, + "learning_rate": 1.2950852557673018e-05, + "loss": 1.2967, + "theoretical_loss": 3.3291676866431086, + "tokens_seen": 3248095232 + }, + { + "epoch": 0.98, + "learning_rate": 1.293480441323972e-05, + "loss": 1.2742, + "theoretical_loss": 3.329162648479673, + "tokens_seen": 3248160768 + }, + { + "epoch": 0.98, + "learning_rate": 1.291875626880642e-05, + "loss": 1.2485, + "theoretical_loss": 3.32915761044635, + "tokens_seen": 3248226304 + }, + { + "epoch": 0.98, + "learning_rate": 1.290270812437312e-05, + "loss": 1.3094, + "theoretical_loss": 3.3291525725431335, + "tokens_seen": 3248291840 + }, + { + "epoch": 0.98, + "learning_rate": 1.2886659979939819e-05, + "loss": 1.2649, + "theoretical_loss": 3.329147534770018, + "tokens_seen": 3248357376 + }, + { + "epoch": 0.98, + "learning_rate": 1.2870611835506521e-05, + "loss": 1.2651, + "theoretical_loss": 3.3291424971269965, + "tokens_seen": 3248422912 + }, + { + "epoch": 0.98, + "learning_rate": 1.285456369107322e-05, + "loss": 1.2595, + "theoretical_loss": 3.329137459614064, + "tokens_seen": 3248488448 + }, + { + "epoch": 0.98, + "learning_rate": 1.2838515546639921e-05, + "loss": 1.336, + "theoretical_loss": 3.329132422231214, + "tokens_seen": 3248553984 + }, + { + "epoch": 0.98, + "learning_rate": 1.282246740220662e-05, + "loss": 1.1514, + "theoretical_loss": 3.32912738497844, + "tokens_seen": 3248619520 + }, + { + "epoch": 0.98, + "learning_rate": 1.2806419257773319e-05, + "loss": 1.3258, + "theoretical_loss": 3.3291223478557375, + "tokens_seen": 3248685056 + }, + { + "epoch": 0.98, + "learning_rate": 1.2790371113340021e-05, + "loss": 1.3043, + "theoretical_loss": 3.329117310863099, + "tokens_seen": 3248750592 + }, + { + "epoch": 0.98, + "learning_rate": 1.2774322968906722e-05, + "loss": 1.2978, + "theoretical_loss": 3.3291122740005195, + "tokens_seen": 3248816128 + }, + { + "epoch": 0.98, + "learning_rate": 1.275827482447342e-05, + "loss": 1.3611, + "theoretical_loss": 3.3291072372679924, + "tokens_seen": 3248881664 + }, + { + "epoch": 0.98, + "learning_rate": 1.274222668004012e-05, + "loss": 1.4028, + "theoretical_loss": 3.329102200665512, + "tokens_seen": 3248947200 + }, + { + "epoch": 0.98, + "learning_rate": 1.272617853560682e-05, + "loss": 1.1512, + "theoretical_loss": 3.3290971641930724, + "tokens_seen": 3249012736 + }, + { + "epoch": 0.98, + "learning_rate": 1.2710130391173523e-05, + "loss": 1.3156, + "theoretical_loss": 3.3290921278506675, + "tokens_seen": 3249078272 + }, + { + "epoch": 0.98, + "learning_rate": 1.2694082246740222e-05, + "loss": 1.31, + "theoretical_loss": 3.3290870916382915, + "tokens_seen": 3249143808 + }, + { + "epoch": 0.98, + "learning_rate": 1.267803410230692e-05, + "loss": 1.2637, + "theoretical_loss": 3.329082055555938, + "tokens_seen": 3249209344 + }, + { + "epoch": 0.98, + "learning_rate": 1.2661985957873621e-05, + "loss": 1.2044, + "theoretical_loss": 3.3290770196036013, + "tokens_seen": 3249274880 + }, + { + "epoch": 0.98, + "learning_rate": 1.2645937813440323e-05, + "loss": 1.2764, + "theoretical_loss": 3.329071983781276, + "tokens_seen": 3249340416 + }, + { + "epoch": 0.98, + "learning_rate": 1.2629889669007022e-05, + "loss": 1.2767, + "theoretical_loss": 3.329066948088955, + "tokens_seen": 3249405952 + }, + { + "epoch": 0.98, + "learning_rate": 1.2613841524573721e-05, + "loss": 1.2979, + "theoretical_loss": 3.3290619125266327, + "tokens_seen": 3249471488 + }, + { + "epoch": 0.98, + "learning_rate": 1.2597793380140422e-05, + "loss": 1.2473, + "theoretical_loss": 3.3290568770943034, + "tokens_seen": 3249537024 + }, + { + "epoch": 0.98, + "learning_rate": 1.2581745235707121e-05, + "loss": 1.2318, + "theoretical_loss": 3.3290518417919612, + "tokens_seen": 3249602560 + }, + { + "epoch": 0.98, + "learning_rate": 1.2565697091273823e-05, + "loss": 1.2188, + "theoretical_loss": 3.3290468066195995, + "tokens_seen": 3249668096 + }, + { + "epoch": 0.98, + "learning_rate": 1.2549648946840522e-05, + "loss": 1.2696, + "theoretical_loss": 3.3290417715772134, + "tokens_seen": 3249733632 + }, + { + "epoch": 0.98, + "learning_rate": 1.2533600802407223e-05, + "loss": 1.3374, + "theoretical_loss": 3.3290367366647957, + "tokens_seen": 3249799168 + }, + { + "epoch": 0.98, + "learning_rate": 1.2517552657973922e-05, + "loss": 1.2381, + "theoretical_loss": 3.329031701882341, + "tokens_seen": 3249864704 + }, + { + "epoch": 0.98, + "learning_rate": 1.250150451354062e-05, + "loss": 1.2534, + "theoretical_loss": 3.3290266672298436, + "tokens_seen": 3249930240 + }, + { + "epoch": 0.98, + "learning_rate": 1.2485456369107321e-05, + "loss": 1.2218, + "theoretical_loss": 3.3290216327072972, + "tokens_seen": 3249995776 + }, + { + "epoch": 0.98, + "learning_rate": 1.2469408224674024e-05, + "loss": 1.3755, + "theoretical_loss": 3.3290165983146958, + "tokens_seen": 3250061312 + }, + { + "epoch": 0.98, + "learning_rate": 1.2453360080240723e-05, + "loss": 1.1197, + "theoretical_loss": 3.329011564052034, + "tokens_seen": 3250126848 + }, + { + "epoch": 0.98, + "learning_rate": 1.2437311935807423e-05, + "loss": 1.2605, + "theoretical_loss": 3.3290065299193046, + "tokens_seen": 3250192384 + }, + { + "epoch": 0.98, + "learning_rate": 1.2421263791374122e-05, + "loss": 1.2146, + "theoretical_loss": 3.3290014959165024, + "tokens_seen": 3250257920 + }, + { + "epoch": 0.98, + "learning_rate": 1.2405215646940825e-05, + "loss": 1.1611, + "theoretical_loss": 3.3289964620436217, + "tokens_seen": 3250323456 + }, + { + "epoch": 0.98, + "learning_rate": 1.2389167502507523e-05, + "loss": 1.3268, + "theoretical_loss": 3.3289914283006556, + "tokens_seen": 3250388992 + }, + { + "epoch": 0.98, + "learning_rate": 1.2373119358074222e-05, + "loss": 1.1776, + "theoretical_loss": 3.3289863946875995, + "tokens_seen": 3250454528 + }, + { + "epoch": 0.99, + "learning_rate": 1.2357071213640923e-05, + "loss": 1.2636, + "theoretical_loss": 3.3289813612044465, + "tokens_seen": 3250520064 + }, + { + "epoch": 0.99, + "objective/train/docs_used": 1181030, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.1074275970458984, + "objective/train/theoretical_loss": 3.3289763278511906, + "objective/train/tokens_used": 3271045600, + "theoretical_loss": 3.3289763278511906, + "tokens_seen": 3250585600 + }, + { + "epoch": 0.99, + "learning_rate": 1.2341023069207624e-05, + "loss": 1.2211, + "theoretical_loss": 3.3289763278511906, + "tokens_seen": 3250585600 + }, + { + "epoch": 0.99, + "learning_rate": 1.2324974924774324e-05, + "loss": 1.1822, + "theoretical_loss": 3.328971294627826, + "tokens_seen": 3250651136 + }, + { + "epoch": 0.99, + "learning_rate": 1.2308926780341023e-05, + "loss": 1.2073, + "theoretical_loss": 3.3289662615343465, + "tokens_seen": 3250716672 + }, + { + "epoch": 0.99, + "learning_rate": 1.2292878635907724e-05, + "loss": 1.2669, + "theoretical_loss": 3.328961228570747, + "tokens_seen": 3250782208 + }, + { + "epoch": 0.99, + "learning_rate": 1.2276830491474425e-05, + "loss": 1.3119, + "theoretical_loss": 3.32895619573702, + "tokens_seen": 3250847744 + }, + { + "epoch": 0.99, + "learning_rate": 1.2260782347041123e-05, + "loss": 1.2496, + "theoretical_loss": 3.328951163033161, + "tokens_seen": 3250913280 + }, + { + "epoch": 0.99, + "learning_rate": 1.2244734202607824e-05, + "loss": 1.2807, + "theoretical_loss": 3.3289461304591637, + "tokens_seen": 3250978816 + }, + { + "epoch": 0.99, + "learning_rate": 1.2228686058174523e-05, + "loss": 1.1264, + "theoretical_loss": 3.3289410980150214, + "tokens_seen": 3251044352 + }, + { + "epoch": 0.99, + "learning_rate": 1.2212637913741225e-05, + "loss": 1.2829, + "theoretical_loss": 3.328936065700729, + "tokens_seen": 3251109888 + }, + { + "epoch": 0.99, + "learning_rate": 1.2196589769307924e-05, + "loss": 1.2003, + "theoretical_loss": 3.32893103351628, + "tokens_seen": 3251175424 + }, + { + "epoch": 0.99, + "learning_rate": 1.2180541624874625e-05, + "loss": 1.2567, + "theoretical_loss": 3.328926001461668, + "tokens_seen": 3251240960 + }, + { + "epoch": 0.99, + "learning_rate": 1.2164493480441324e-05, + "loss": 1.3119, + "theoretical_loss": 3.3289209695368887, + "tokens_seen": 3251306496 + }, + { + "epoch": 0.99, + "learning_rate": 1.2148445336008025e-05, + "loss": 1.2544, + "theoretical_loss": 3.328915937741934, + "tokens_seen": 3251372032 + }, + { + "epoch": 0.99, + "learning_rate": 1.2132397191574725e-05, + "loss": 1.2778, + "theoretical_loss": 3.3289109060768, + "tokens_seen": 3251437568 + }, + { + "epoch": 0.99, + "learning_rate": 1.2116349047141424e-05, + "loss": 1.1332, + "theoretical_loss": 3.328905874541479, + "tokens_seen": 3251503104 + }, + { + "epoch": 0.99, + "learning_rate": 1.2100300902708125e-05, + "loss": 1.1926, + "theoretical_loss": 3.3289008431359663, + "tokens_seen": 3251568640 + }, + { + "epoch": 0.99, + "learning_rate": 1.2084252758274825e-05, + "loss": 1.3423, + "theoretical_loss": 3.328895811860255, + "tokens_seen": 3251634176 + }, + { + "epoch": 0.99, + "learning_rate": 1.2068204613841526e-05, + "loss": 1.3134, + "theoretical_loss": 3.3288907807143397, + "tokens_seen": 3251699712 + }, + { + "epoch": 0.99, + "learning_rate": 1.2052156469408225e-05, + "loss": 1.1475, + "theoretical_loss": 3.3288857496982143, + "tokens_seen": 3251765248 + }, + { + "epoch": 0.99, + "learning_rate": 1.2036108324974926e-05, + "loss": 1.1434, + "theoretical_loss": 3.3288807188118725, + "tokens_seen": 3251830784 + }, + { + "epoch": 0.99, + "learning_rate": 1.2020060180541626e-05, + "loss": 1.2545, + "theoretical_loss": 3.3288756880553088, + "tokens_seen": 3251896320 + }, + { + "epoch": 0.99, + "learning_rate": 1.2004012036108325e-05, + "loss": 1.1592, + "theoretical_loss": 3.328870657428517, + "tokens_seen": 3251961856 + }, + { + "epoch": 0.99, + "learning_rate": 1.1987963891675026e-05, + "loss": 1.224, + "theoretical_loss": 3.3288656269314916, + "tokens_seen": 3252027392 + }, + { + "epoch": 0.99, + "learning_rate": 1.1971915747241726e-05, + "loss": 1.2558, + "theoretical_loss": 3.328860596564226, + "tokens_seen": 3252092928 + }, + { + "epoch": 0.99, + "learning_rate": 1.1955867602808425e-05, + "loss": 1.2839, + "theoretical_loss": 3.3288555663267143, + "tokens_seen": 3252158464 + }, + { + "epoch": 0.99, + "learning_rate": 1.1939819458375126e-05, + "loss": 1.2464, + "theoretical_loss": 3.328850536218951, + "tokens_seen": 3252224000 + }, + { + "epoch": 0.99, + "learning_rate": 1.1923771313941825e-05, + "loss": 1.288, + "theoretical_loss": 3.32884550624093, + "tokens_seen": 3252289536 + }, + { + "epoch": 0.99, + "learning_rate": 1.1907723169508527e-05, + "loss": 1.3004, + "theoretical_loss": 3.328840476392645, + "tokens_seen": 3252355072 + }, + { + "epoch": 0.99, + "learning_rate": 1.1891675025075226e-05, + "loss": 1.3386, + "theoretical_loss": 3.32883544667409, + "tokens_seen": 3252420608 + }, + { + "epoch": 0.99, + "learning_rate": 1.1875626880641927e-05, + "loss": 1.1877, + "theoretical_loss": 3.3288304170852596, + "tokens_seen": 3252486144 + }, + { + "epoch": 0.99, + "learning_rate": 1.1859578736208626e-05, + "loss": 1.2331, + "theoretical_loss": 3.3288253876261473, + "tokens_seen": 3252551680 + }, + { + "epoch": 0.99, + "learning_rate": 1.1843530591775326e-05, + "loss": 1.3823, + "theoretical_loss": 3.3288203582967473, + "tokens_seen": 3252617216 + }, + { + "epoch": 0.99, + "learning_rate": 1.1827482447342027e-05, + "loss": 1.2526, + "theoretical_loss": 3.328815329097054, + "tokens_seen": 3252682752 + }, + { + "epoch": 0.99, + "learning_rate": 1.1811434302908726e-05, + "loss": 1.199, + "theoretical_loss": 3.328810300027061, + "tokens_seen": 3252748288 + }, + { + "epoch": 0.99, + "learning_rate": 1.1795386158475427e-05, + "loss": 1.174, + "theoretical_loss": 3.328805271086763, + "tokens_seen": 3252813824 + }, + { + "epoch": 0.99, + "learning_rate": 1.1779338014042127e-05, + "loss": 1.1911, + "theoretical_loss": 3.3288002422761527, + "tokens_seen": 3252879360 + }, + { + "epoch": 0.99, + "learning_rate": 1.1763289869608828e-05, + "loss": 1.291, + "theoretical_loss": 3.3287952135952255, + "tokens_seen": 3252944896 + }, + { + "epoch": 0.99, + "learning_rate": 1.1747241725175527e-05, + "loss": 1.2308, + "theoretical_loss": 3.3287901850439745, + "tokens_seen": 3253010432 + }, + { + "epoch": 0.99, + "learning_rate": 1.1731193580742226e-05, + "loss": 1.1166, + "theoretical_loss": 3.3287851566223945, + "tokens_seen": 3253075968 + }, + { + "epoch": 0.99, + "learning_rate": 1.1715145436308928e-05, + "loss": 1.2137, + "theoretical_loss": 3.3287801283304788, + "tokens_seen": 3253141504 + }, + { + "epoch": 0.99, + "learning_rate": 1.1699097291875627e-05, + "loss": 1.2832, + "theoretical_loss": 3.3287751001682224, + "tokens_seen": 3253207040 + }, + { + "epoch": 0.99, + "learning_rate": 1.1683049147442328e-05, + "loss": 1.2662, + "theoretical_loss": 3.3287700721356184, + "tokens_seen": 3253272576 + }, + { + "epoch": 0.99, + "learning_rate": 1.1667001003009027e-05, + "loss": 1.2935, + "theoretical_loss": 3.3287650442326617, + "tokens_seen": 3253338112 + }, + { + "epoch": 0.99, + "learning_rate": 1.1650952858575729e-05, + "loss": 1.256, + "theoretical_loss": 3.3287600164593454, + "tokens_seen": 3253403648 + }, + { + "epoch": 0.99, + "learning_rate": 1.1634904714142428e-05, + "loss": 1.3406, + "theoretical_loss": 3.328754988815664, + "tokens_seen": 3253469184 + }, + { + "epoch": 0.99, + "learning_rate": 1.1618856569709127e-05, + "loss": 1.2123, + "theoretical_loss": 3.328749961301612, + "tokens_seen": 3253534720 + }, + { + "epoch": 0.99, + "learning_rate": 1.1602808425275827e-05, + "loss": 1.2932, + "theoretical_loss": 3.3287449339171826, + "tokens_seen": 3253600256 + }, + { + "epoch": 0.99, + "learning_rate": 1.1586760280842528e-05, + "loss": 1.2425, + "theoretical_loss": 3.3287399066623706, + "tokens_seen": 3253665792 + }, + { + "epoch": 0.99, + "learning_rate": 1.1570712136409229e-05, + "loss": 1.2692, + "theoretical_loss": 3.3287348795371696, + "tokens_seen": 3253731328 + }, + { + "epoch": 0.99, + "learning_rate": 1.1554663991975928e-05, + "loss": 1.1545, + "theoretical_loss": 3.328729852541574, + "tokens_seen": 3253796864 + }, + { + "epoch": 0.99, + "objective/train/docs_used": 1182213, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.1620917320251465, + "objective/train/theoretical_loss": 3.328724825675577, + "objective/train/tokens_used": 3274322400, + "theoretical_loss": 3.328724825675577, + "tokens_seen": 3253862400 + }, + { + "epoch": 0.99, + "learning_rate": 1.1538615847542628e-05, + "loss": 1.1779, + "theoretical_loss": 3.328724825675577, + "tokens_seen": 3253862400 + }, + { + "epoch": 0.99, + "learning_rate": 1.1522567703109329e-05, + "loss": 1.3017, + "theoretical_loss": 3.328719798939174, + "tokens_seen": 3253927936 + }, + { + "epoch": 0.99, + "learning_rate": 1.1506519558676028e-05, + "loss": 1.2926, + "theoretical_loss": 3.328714772332358, + "tokens_seen": 3253993472 + }, + { + "epoch": 0.99, + "learning_rate": 1.1490471414242729e-05, + "loss": 1.2289, + "theoretical_loss": 3.328709745855123, + "tokens_seen": 3254059008 + }, + { + "epoch": 0.99, + "learning_rate": 1.1474423269809429e-05, + "loss": 1.3129, + "theoretical_loss": 3.328704719507464, + "tokens_seen": 3254124544 + }, + { + "epoch": 0.99, + "learning_rate": 1.145837512537613e-05, + "loss": 1.3424, + "theoretical_loss": 3.3286996932893738, + "tokens_seen": 3254190080 + }, + { + "epoch": 0.99, + "learning_rate": 1.1442326980942829e-05, + "loss": 1.262, + "theoretical_loss": 3.328694667200848, + "tokens_seen": 3254255616 + }, + { + "epoch": 0.99, + "learning_rate": 1.142627883650953e-05, + "loss": 1.1659, + "theoretical_loss": 3.328689641241879, + "tokens_seen": 3254321152 + }, + { + "epoch": 0.99, + "learning_rate": 1.141023069207623e-05, + "loss": 1.1681, + "theoretical_loss": 3.328684615412462, + "tokens_seen": 3254386688 + }, + { + "epoch": 0.99, + "learning_rate": 1.1394182547642929e-05, + "loss": 1.1977, + "theoretical_loss": 3.3286795897125905, + "tokens_seen": 3254452224 + }, + { + "epoch": 0.99, + "learning_rate": 1.137813440320963e-05, + "loss": 1.2007, + "theoretical_loss": 3.3286745641422586, + "tokens_seen": 3254517760 + }, + { + "epoch": 0.99, + "learning_rate": 1.1362086258776329e-05, + "loss": 1.2248, + "theoretical_loss": 3.328669538701461, + "tokens_seen": 3254583296 + }, + { + "epoch": 0.99, + "learning_rate": 1.134603811434303e-05, + "loss": 1.202, + "theoretical_loss": 3.3286645133901906, + "tokens_seen": 3254648832 + }, + { + "epoch": 0.99, + "learning_rate": 1.132998996990973e-05, + "loss": 1.2395, + "theoretical_loss": 3.3286594882084426, + "tokens_seen": 3254714368 + }, + { + "epoch": 0.99, + "learning_rate": 1.131394182547643e-05, + "loss": 1.2239, + "theoretical_loss": 3.32865446315621, + "tokens_seen": 3254779904 + }, + { + "epoch": 0.99, + "learning_rate": 1.129789368104313e-05, + "loss": 1.1751, + "theoretical_loss": 3.3286494382334877, + "tokens_seen": 3254845440 + }, + { + "epoch": 0.99, + "learning_rate": 1.128184553660983e-05, + "loss": 1.3095, + "theoretical_loss": 3.3286444134402693, + "tokens_seen": 3254910976 + }, + { + "epoch": 0.99, + "learning_rate": 1.126579739217653e-05, + "loss": 1.29, + "theoretical_loss": 3.3286393887765495, + "tokens_seen": 3254976512 + }, + { + "epoch": 0.99, + "learning_rate": 1.124974924774323e-05, + "loss": 1.1689, + "theoretical_loss": 3.3286343642423213, + "tokens_seen": 3255042048 + }, + { + "epoch": 0.99, + "learning_rate": 1.123370110330993e-05, + "loss": 1.2404, + "theoretical_loss": 3.3286293398375792, + "tokens_seen": 3255107584 + }, + { + "epoch": 0.99, + "learning_rate": 1.121765295887663e-05, + "loss": 1.3008, + "theoretical_loss": 3.328624315562318, + "tokens_seen": 3255173120 + }, + { + "epoch": 0.99, + "learning_rate": 1.1201604814443331e-05, + "loss": 1.245, + "theoretical_loss": 3.3286192914165307, + "tokens_seen": 3255238656 + }, + { + "epoch": 0.99, + "learning_rate": 1.118555667001003e-05, + "loss": 1.3381, + "theoretical_loss": 3.3286142674002117, + "tokens_seen": 3255304192 + }, + { + "epoch": 0.99, + "learning_rate": 1.116950852557673e-05, + "loss": 1.2221, + "theoretical_loss": 3.3286092435133554, + "tokens_seen": 3255369728 + }, + { + "epoch": 0.99, + "learning_rate": 1.1153460381143432e-05, + "loss": 1.252, + "theoretical_loss": 3.328604219755955, + "tokens_seen": 3255435264 + }, + { + "epoch": 0.99, + "learning_rate": 1.113741223671013e-05, + "loss": 1.1503, + "theoretical_loss": 3.328599196128006, + "tokens_seen": 3255500800 + }, + { + "epoch": 0.99, + "learning_rate": 1.1121364092276831e-05, + "loss": 1.1071, + "theoretical_loss": 3.328594172629501, + "tokens_seen": 3255566336 + }, + { + "epoch": 0.99, + "learning_rate": 1.110531594784353e-05, + "loss": 1.232, + "theoretical_loss": 3.328589149260435, + "tokens_seen": 3255631872 + }, + { + "epoch": 0.99, + "learning_rate": 1.1089267803410233e-05, + "loss": 1.1373, + "theoretical_loss": 3.3285841260208016, + "tokens_seen": 3255697408 + }, + { + "epoch": 0.99, + "learning_rate": 1.1073219658976931e-05, + "loss": 1.2705, + "theoretical_loss": 3.328579102910595, + "tokens_seen": 3255762944 + }, + { + "epoch": 0.99, + "learning_rate": 1.105717151454363e-05, + "loss": 1.3043, + "theoretical_loss": 3.328574079929809, + "tokens_seen": 3255828480 + }, + { + "epoch": 0.99, + "learning_rate": 1.1041123370110331e-05, + "loss": 1.1912, + "theoretical_loss": 3.3285690570784383, + "tokens_seen": 3255894016 + }, + { + "epoch": 0.99, + "learning_rate": 1.1025075225677032e-05, + "loss": 1.2627, + "theoretical_loss": 3.3285640343564764, + "tokens_seen": 3255959552 + }, + { + "epoch": 0.99, + "learning_rate": 1.1009027081243732e-05, + "loss": 1.1963, + "theoretical_loss": 3.3285590117639177, + "tokens_seen": 3256025088 + }, + { + "epoch": 0.99, + "learning_rate": 1.0992978936810431e-05, + "loss": 1.2973, + "theoretical_loss": 3.328553989300756, + "tokens_seen": 3256090624 + }, + { + "epoch": 0.99, + "learning_rate": 1.0976930792377132e-05, + "loss": 1.258, + "theoretical_loss": 3.328548966966985, + "tokens_seen": 3256156160 + }, + { + "epoch": 0.99, + "learning_rate": 1.0960882647943833e-05, + "loss": 1.1387, + "theoretical_loss": 3.3285439447626, + "tokens_seen": 3256221696 + }, + { + "epoch": 0.99, + "learning_rate": 1.0944834503510531e-05, + "loss": 1.2931, + "theoretical_loss": 3.328538922687594, + "tokens_seen": 3256287232 + }, + { + "epoch": 0.99, + "learning_rate": 1.0928786359077232e-05, + "loss": 1.2673, + "theoretical_loss": 3.3285339007419608, + "tokens_seen": 3256352768 + }, + { + "epoch": 0.99, + "learning_rate": 1.0912738214643933e-05, + "loss": 1.2852, + "theoretical_loss": 3.3285288789256957, + "tokens_seen": 3256418304 + }, + { + "epoch": 0.99, + "learning_rate": 1.0896690070210633e-05, + "loss": 1.2528, + "theoretical_loss": 3.3285238572387916, + "tokens_seen": 3256483840 + }, + { + "epoch": 0.99, + "learning_rate": 1.0880641925777332e-05, + "loss": 1.3107, + "theoretical_loss": 3.328518835681243, + "tokens_seen": 3256549376 + }, + { + "epoch": 0.99, + "learning_rate": 1.0864593781344033e-05, + "loss": 1.3469, + "theoretical_loss": 3.3285138142530446, + "tokens_seen": 3256614912 + }, + { + "epoch": 0.99, + "learning_rate": 1.0848545636910734e-05, + "loss": 1.3493, + "theoretical_loss": 3.328508792954189, + "tokens_seen": 3256680448 + }, + { + "epoch": 0.99, + "learning_rate": 1.0832497492477433e-05, + "loss": 1.2448, + "theoretical_loss": 3.328503771784672, + "tokens_seen": 3256745984 + }, + { + "epoch": 0.99, + "learning_rate": 1.0816449348044133e-05, + "loss": 1.2017, + "theoretical_loss": 3.3284987507444863, + "tokens_seen": 3256811520 + }, + { + "epoch": 0.99, + "learning_rate": 1.0800401203610832e-05, + "loss": 1.3273, + "theoretical_loss": 3.3284937298336263, + "tokens_seen": 3256877056 + }, + { + "epoch": 0.99, + "learning_rate": 1.0784353059177533e-05, + "loss": 1.3161, + "theoretical_loss": 3.328488709052087, + "tokens_seen": 3256942592 + }, + { + "epoch": 0.99, + "learning_rate": 1.0768304914744233e-05, + "loss": 1.2937, + "theoretical_loss": 3.328483688399861, + "tokens_seen": 3257008128 + }, + { + "epoch": 0.99, + "learning_rate": 1.0752256770310934e-05, + "loss": 1.2832, + "theoretical_loss": 3.328478667876943, + "tokens_seen": 3257073664 + }, + { + "epoch": 0.99, + "objective/train/docs_used": 1183352, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.2485523223876953, + "objective/train/theoretical_loss": 3.3284736474833276, + "objective/train/tokens_used": 3277599200, + "theoretical_loss": 3.3284736474833276, + "tokens_seen": 3257139200 + }, + { + "epoch": 0.99, + "learning_rate": 1.0736208625877633e-05, + "loss": 1.2338, + "theoretical_loss": 3.3284736474833276, + "tokens_seen": 3257139200 + }, + { + "epoch": 0.99, + "learning_rate": 1.0720160481444334e-05, + "loss": 1.1105, + "theoretical_loss": 3.3284686272190083, + "tokens_seen": 3257204736 + }, + { + "epoch": 0.99, + "learning_rate": 1.0704112337011034e-05, + "loss": 1.2289, + "theoretical_loss": 3.328463607083979, + "tokens_seen": 3257270272 + }, + { + "epoch": 0.99, + "learning_rate": 1.0688064192577733e-05, + "loss": 1.1272, + "theoretical_loss": 3.3284585870782344, + "tokens_seen": 3257335808 + }, + { + "epoch": 0.99, + "learning_rate": 1.0672016048144434e-05, + "loss": 1.3429, + "theoretical_loss": 3.328453567201768, + "tokens_seen": 3257401344 + }, + { + "epoch": 0.99, + "learning_rate": 1.0655967903711134e-05, + "loss": 1.2527, + "theoretical_loss": 3.328448547454574, + "tokens_seen": 3257466880 + }, + { + "epoch": 0.99, + "learning_rate": 1.0639919759277835e-05, + "loss": 1.1909, + "theoretical_loss": 3.3284435278366464, + "tokens_seen": 3257532416 + }, + { + "epoch": 0.99, + "learning_rate": 1.0623871614844534e-05, + "loss": 1.2648, + "theoretical_loss": 3.32843850834798, + "tokens_seen": 3257597952 + }, + { + "epoch": 0.99, + "learning_rate": 1.0607823470411233e-05, + "loss": 1.2299, + "theoretical_loss": 3.328433488988568, + "tokens_seen": 3257663488 + }, + { + "epoch": 0.99, + "learning_rate": 1.0591775325977935e-05, + "loss": 1.2948, + "theoretical_loss": 3.3284284697584043, + "tokens_seen": 3257729024 + }, + { + "epoch": 0.99, + "learning_rate": 1.0575727181544634e-05, + "loss": 1.2494, + "theoretical_loss": 3.3284234506574837, + "tokens_seen": 3257794560 + }, + { + "epoch": 0.99, + "learning_rate": 1.0559679037111335e-05, + "loss": 1.297, + "theoretical_loss": 3.3284184316858, + "tokens_seen": 3257860096 + }, + { + "epoch": 0.99, + "learning_rate": 1.0543630892678034e-05, + "loss": 1.2043, + "theoretical_loss": 3.328413412843348, + "tokens_seen": 3257925632 + }, + { + "epoch": 0.99, + "learning_rate": 1.0527582748244736e-05, + "loss": 1.2716, + "theoretical_loss": 3.32840839413012, + "tokens_seen": 3257991168 + }, + { + "epoch": 0.99, + "learning_rate": 1.0511534603811435e-05, + "loss": 1.1212, + "theoretical_loss": 3.3284033755461118, + "tokens_seen": 3258056704 + }, + { + "epoch": 0.99, + "learning_rate": 1.0495486459378134e-05, + "loss": 1.1996, + "theoretical_loss": 3.3283983570913165, + "tokens_seen": 3258122240 + }, + { + "epoch": 0.99, + "learning_rate": 1.0479438314944835e-05, + "loss": 1.0803, + "theoretical_loss": 3.3283933387657285, + "tokens_seen": 3258187776 + }, + { + "epoch": 0.99, + "learning_rate": 1.0463390170511535e-05, + "loss": 1.2113, + "theoretical_loss": 3.328388320569342, + "tokens_seen": 3258253312 + }, + { + "epoch": 0.99, + "learning_rate": 1.0447342026078236e-05, + "loss": 1.2834, + "theoretical_loss": 3.3283833025021505, + "tokens_seen": 3258318848 + }, + { + "epoch": 0.99, + "learning_rate": 1.0431293881644935e-05, + "loss": 1.1877, + "theoretical_loss": 3.328378284564149, + "tokens_seen": 3258384384 + }, + { + "epoch": 0.99, + "learning_rate": 1.0415245737211635e-05, + "loss": 1.167, + "theoretical_loss": 3.3283732667553307, + "tokens_seen": 3258449920 + }, + { + "epoch": 0.99, + "learning_rate": 1.0399197592778336e-05, + "loss": 1.2273, + "theoretical_loss": 3.32836824907569, + "tokens_seen": 3258515456 + }, + { + "epoch": 0.99, + "learning_rate": 1.0383149448345035e-05, + "loss": 1.1746, + "theoretical_loss": 3.3283632315252216, + "tokens_seen": 3258580992 + }, + { + "epoch": 0.99, + "learning_rate": 1.0367101303911736e-05, + "loss": 1.1769, + "theoretical_loss": 3.3283582141039187, + "tokens_seen": 3258646528 + }, + { + "epoch": 0.99, + "learning_rate": 1.0351053159478436e-05, + "loss": 1.2326, + "theoretical_loss": 3.3283531968117757, + "tokens_seen": 3258712064 + }, + { + "epoch": 0.99, + "learning_rate": 1.0335005015045137e-05, + "loss": 1.2753, + "theoretical_loss": 3.3283481796487866, + "tokens_seen": 3258777600 + }, + { + "epoch": 0.99, + "learning_rate": 1.0318956870611836e-05, + "loss": 1.2657, + "theoretical_loss": 3.3283431626149453, + "tokens_seen": 3258843136 + }, + { + "epoch": 0.99, + "learning_rate": 1.0302908726178537e-05, + "loss": 1.1411, + "theoretical_loss": 3.328338145710246, + "tokens_seen": 3258908672 + }, + { + "epoch": 0.99, + "learning_rate": 1.0286860581745235e-05, + "loss": 1.2409, + "theoretical_loss": 3.3283331289346836, + "tokens_seen": 3258974208 + }, + { + "epoch": 0.99, + "learning_rate": 1.0270812437311936e-05, + "loss": 1.2032, + "theoretical_loss": 3.328328112288251, + "tokens_seen": 3259039744 + }, + { + "epoch": 0.99, + "learning_rate": 1.0254764292878637e-05, + "loss": 1.4479, + "theoretical_loss": 3.328323095770943, + "tokens_seen": 3259105280 + }, + { + "epoch": 0.99, + "learning_rate": 1.0238716148445336e-05, + "loss": 1.2473, + "theoretical_loss": 3.328318079382753, + "tokens_seen": 3259170816 + }, + { + "epoch": 0.99, + "learning_rate": 1.0222668004012036e-05, + "loss": 1.2163, + "theoretical_loss": 3.3283130631236757, + "tokens_seen": 3259236352 + }, + { + "epoch": 0.99, + "learning_rate": 1.0206619859578737e-05, + "loss": 1.2164, + "theoretical_loss": 3.328308046993705, + "tokens_seen": 3259301888 + }, + { + "epoch": 0.99, + "learning_rate": 1.0190571715145438e-05, + "loss": 1.3258, + "theoretical_loss": 3.328303030992835, + "tokens_seen": 3259367424 + }, + { + "epoch": 0.99, + "learning_rate": 1.0174523570712136e-05, + "loss": 1.1931, + "theoretical_loss": 3.32829801512106, + "tokens_seen": 3259432960 + }, + { + "epoch": 0.99, + "learning_rate": 1.0158475426278837e-05, + "loss": 1.2673, + "theoretical_loss": 3.3282929993783736, + "tokens_seen": 3259498496 + }, + { + "epoch": 0.99, + "learning_rate": 1.0142427281845538e-05, + "loss": 1.2179, + "theoretical_loss": 3.32828798376477, + "tokens_seen": 3259564032 + }, + { + "epoch": 0.99, + "learning_rate": 1.0126379137412237e-05, + "loss": 1.2354, + "theoretical_loss": 3.3282829682802437, + "tokens_seen": 3259629568 + }, + { + "epoch": 0.99, + "learning_rate": 1.0110330992978937e-05, + "loss": 1.2268, + "theoretical_loss": 3.3282779529247883, + "tokens_seen": 3259695104 + }, + { + "epoch": 0.99, + "learning_rate": 1.0094282848545638e-05, + "loss": 1.1759, + "theoretical_loss": 3.328272937698398, + "tokens_seen": 3259760640 + }, + { + "epoch": 0.99, + "learning_rate": 1.0078234704112339e-05, + "loss": 1.1356, + "theoretical_loss": 3.328267922601067, + "tokens_seen": 3259826176 + }, + { + "epoch": 0.99, + "learning_rate": 1.0062186559679038e-05, + "loss": 1.3643, + "theoretical_loss": 3.3282629076327894, + "tokens_seen": 3259891712 + }, + { + "epoch": 0.99, + "learning_rate": 1.0046138415245736e-05, + "loss": 1.2441, + "theoretical_loss": 3.3282578927935593, + "tokens_seen": 3259957248 + }, + { + "epoch": 0.99, + "learning_rate": 1.0030090270812439e-05, + "loss": 1.1813, + "theoretical_loss": 3.3282528780833704, + "tokens_seen": 3260022784 + }, + { + "epoch": 0.99, + "learning_rate": 1.0014042126379138e-05, + "loss": 1.2437, + "theoretical_loss": 3.3282478635022175, + "tokens_seen": 3260088320 + }, + { + "epoch": 0.99, + "learning_rate": 9.997993981945838e-06, + "loss": 1.1897, + "theoretical_loss": 3.328242849050094, + "tokens_seen": 3260153856 + }, + { + "epoch": 0.99, + "learning_rate": 9.981945837512537e-06, + "loss": 1.2375, + "theoretical_loss": 3.3282378347269943, + "tokens_seen": 3260219392 + }, + { + "epoch": 0.99, + "learning_rate": 9.96589769307924e-06, + "loss": 1.1877, + "theoretical_loss": 3.3282328205329126, + "tokens_seen": 3260284928 + }, + { + "epoch": 0.99, + "learning_rate": 9.949849548645939e-06, + "loss": 1.1774, + "theoretical_loss": 3.3282278064678428, + "tokens_seen": 3260350464 + }, + { + "epoch": 0.99, + "objective/train/docs_used": 1184523, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.236487627029419, + "objective/train/theoretical_loss": 3.3282227925317787, + "objective/train/tokens_used": 3280876000, + "theoretical_loss": 3.3282227925317787, + "tokens_seen": 3260416000 + }, + { + "epoch": 0.99, + "learning_rate": 9.933801404212638e-06, + "loss": 1.1996, + "theoretical_loss": 3.3282227925317787, + "tokens_seen": 3260416000 + }, + { + "epoch": 0.99, + "learning_rate": 9.917753259779338e-06, + "loss": 1.2728, + "theoretical_loss": 3.328217778724715, + "tokens_seen": 3260481536 + }, + { + "epoch": 0.99, + "learning_rate": 9.901705115346039e-06, + "loss": 1.2027, + "theoretical_loss": 3.3282127650466458, + "tokens_seen": 3260547072 + }, + { + "epoch": 0.99, + "learning_rate": 9.88565697091274e-06, + "loss": 1.2147, + "theoretical_loss": 3.3282077514975645, + "tokens_seen": 3260612608 + }, + { + "epoch": 0.99, + "learning_rate": 9.869608826479438e-06, + "loss": 1.191, + "theoretical_loss": 3.3282027380774655, + "tokens_seen": 3260678144 + }, + { + "epoch": 0.99, + "learning_rate": 9.853560682046139e-06, + "loss": 1.1858, + "theoretical_loss": 3.3281977247863432, + "tokens_seen": 3260743680 + }, + { + "epoch": 0.99, + "learning_rate": 9.83751253761284e-06, + "loss": 1.3422, + "theoretical_loss": 3.328192711624191, + "tokens_seen": 3260809216 + }, + { + "epoch": 0.99, + "learning_rate": 9.821464393179539e-06, + "loss": 1.2277, + "theoretical_loss": 3.328187698591004, + "tokens_seen": 3260874752 + }, + { + "epoch": 0.99, + "learning_rate": 9.80541624874624e-06, + "loss": 1.2551, + "theoretical_loss": 3.3281826856867758, + "tokens_seen": 3260940288 + }, + { + "epoch": 0.99, + "learning_rate": 9.789368104312938e-06, + "loss": 1.1473, + "theoretical_loss": 3.3281776729115, + "tokens_seen": 3261005824 + }, + { + "epoch": 0.99, + "learning_rate": 9.77331995987964e-06, + "loss": 1.227, + "theoretical_loss": 3.3281726602651713, + "tokens_seen": 3261071360 + }, + { + "epoch": 0.99, + "learning_rate": 9.75727181544634e-06, + "loss": 1.2474, + "theoretical_loss": 3.3281676477477835, + "tokens_seen": 3261136896 + }, + { + "epoch": 0.99, + "learning_rate": 9.741223671013038e-06, + "loss": 1.1193, + "theoretical_loss": 3.3281626353593308, + "tokens_seen": 3261202432 + }, + { + "epoch": 0.99, + "learning_rate": 9.725175526579739e-06, + "loss": 1.236, + "theoretical_loss": 3.3281576230998073, + "tokens_seen": 3261267968 + }, + { + "epoch": 0.99, + "learning_rate": 9.70912738214644e-06, + "loss": 1.2035, + "theoretical_loss": 3.3281526109692074, + "tokens_seen": 3261333504 + }, + { + "epoch": 0.99, + "learning_rate": 9.69307923771314e-06, + "loss": 1.1197, + "theoretical_loss": 3.328147598967525, + "tokens_seen": 3261399040 + }, + { + "epoch": 0.99, + "learning_rate": 9.67703109327984e-06, + "loss": 1.2359, + "theoretical_loss": 3.3281425870947534, + "tokens_seen": 3261464576 + }, + { + "epoch": 0.99, + "learning_rate": 9.66098294884654e-06, + "loss": 1.2738, + "theoretical_loss": 3.3281375753508877, + "tokens_seen": 3261530112 + }, + { + "epoch": 0.99, + "learning_rate": 9.64493480441324e-06, + "loss": 1.2287, + "theoretical_loss": 3.3281325637359216, + "tokens_seen": 3261595648 + }, + { + "epoch": 0.99, + "learning_rate": 9.62888665997994e-06, + "loss": 1.2493, + "theoretical_loss": 3.3281275522498497, + "tokens_seen": 3261661184 + }, + { + "epoch": 0.99, + "learning_rate": 9.61283851554664e-06, + "loss": 1.2941, + "theoretical_loss": 3.3281225408926653, + "tokens_seen": 3261726720 + }, + { + "epoch": 0.99, + "learning_rate": 9.59679037111334e-06, + "loss": 1.2025, + "theoretical_loss": 3.328117529664363, + "tokens_seen": 3261792256 + }, + { + "epoch": 0.99, + "learning_rate": 9.580742226680041e-06, + "loss": 1.1648, + "theoretical_loss": 3.3281125185649363, + "tokens_seen": 3261857792 + }, + { + "epoch": 0.99, + "learning_rate": 9.56469408224674e-06, + "loss": 1.1955, + "theoretical_loss": 3.32810750759438, + "tokens_seen": 3261923328 + }, + { + "epoch": 0.99, + "learning_rate": 9.548645937813441e-06, + "loss": 1.2509, + "theoretical_loss": 3.328102496752688, + "tokens_seen": 3261988864 + }, + { + "epoch": 0.99, + "learning_rate": 9.532597793380142e-06, + "loss": 1.217, + "theoretical_loss": 3.3280974860398542, + "tokens_seen": 3262054400 + }, + { + "epoch": 0.99, + "learning_rate": 9.51654964894684e-06, + "loss": 1.2422, + "theoretical_loss": 3.328092475455873, + "tokens_seen": 3262119936 + }, + { + "epoch": 0.99, + "learning_rate": 9.500501504513541e-06, + "loss": 1.1359, + "theoretical_loss": 3.3280874650007384, + "tokens_seen": 3262185472 + }, + { + "epoch": 0.99, + "learning_rate": 9.48445336008024e-06, + "loss": 1.1987, + "theoretical_loss": 3.3280824546744445, + "tokens_seen": 3262251008 + }, + { + "epoch": 0.99, + "learning_rate": 9.468405215646942e-06, + "loss": 1.2594, + "theoretical_loss": 3.328077444476985, + "tokens_seen": 3262316544 + }, + { + "epoch": 0.99, + "learning_rate": 9.452357071213641e-06, + "loss": 1.259, + "theoretical_loss": 3.3280724344083543, + "tokens_seen": 3262382080 + }, + { + "epoch": 0.99, + "learning_rate": 9.436308926780342e-06, + "loss": 1.2777, + "theoretical_loss": 3.328067424468547, + "tokens_seen": 3262447616 + }, + { + "epoch": 0.99, + "learning_rate": 9.420260782347041e-06, + "loss": 1.1355, + "theoretical_loss": 3.3280624146575564, + "tokens_seen": 3262513152 + }, + { + "epoch": 0.99, + "learning_rate": 9.404212637913742e-06, + "loss": 1.1828, + "theoretical_loss": 3.328057404975377, + "tokens_seen": 3262578688 + }, + { + "epoch": 0.99, + "learning_rate": 9.388164493480442e-06, + "loss": 1.2906, + "theoretical_loss": 3.3280523954220027, + "tokens_seen": 3262644224 + }, + { + "epoch": 0.99, + "learning_rate": 9.372116349047141e-06, + "loss": 1.1591, + "theoretical_loss": 3.3280473859974276, + "tokens_seen": 3262709760 + }, + { + "epoch": 0.99, + "learning_rate": 9.356068204613842e-06, + "loss": 1.1749, + "theoretical_loss": 3.3280423767016467, + "tokens_seen": 3262775296 + }, + { + "epoch": 0.99, + "learning_rate": 9.340020060180542e-06, + "loss": 1.1929, + "theoretical_loss": 3.3280373675346526, + "tokens_seen": 3262840832 + }, + { + "epoch": 0.99, + "learning_rate": 9.323971915747243e-06, + "loss": 1.1879, + "theoretical_loss": 3.3280323584964404, + "tokens_seen": 3262906368 + }, + { + "epoch": 0.99, + "learning_rate": 9.307923771313942e-06, + "loss": 1.338, + "theoretical_loss": 3.328027349587004, + "tokens_seen": 3262971904 + }, + { + "epoch": 0.99, + "learning_rate": 9.291875626880641e-06, + "loss": 1.1128, + "theoretical_loss": 3.328022340806337, + "tokens_seen": 3263037440 + }, + { + "epoch": 0.99, + "learning_rate": 9.275827482447343e-06, + "loss": 1.3092, + "theoretical_loss": 3.328017332154434, + "tokens_seen": 3263102976 + }, + { + "epoch": 0.99, + "learning_rate": 9.259779338014042e-06, + "loss": 1.3132, + "theoretical_loss": 3.3280123236312895, + "tokens_seen": 3263168512 + }, + { + "epoch": 0.99, + "learning_rate": 9.243731193580743e-06, + "loss": 1.2128, + "theoretical_loss": 3.328007315236897, + "tokens_seen": 3263234048 + }, + { + "epoch": 0.99, + "learning_rate": 9.227683049147442e-06, + "loss": 1.1513, + "theoretical_loss": 3.3280023069712508, + "tokens_seen": 3263299584 + }, + { + "epoch": 0.99, + "learning_rate": 9.211634904714144e-06, + "loss": 1.2481, + "theoretical_loss": 3.3279972988343447, + "tokens_seen": 3263365120 + }, + { + "epoch": 0.99, + "learning_rate": 9.195586760280843e-06, + "loss": 1.1989, + "theoretical_loss": 3.3279922908261734, + "tokens_seen": 3263430656 + }, + { + "epoch": 0.99, + "learning_rate": 9.179538615847542e-06, + "loss": 1.2789, + "theoretical_loss": 3.3279872829467307, + "tokens_seen": 3263496192 + }, + { + "epoch": 0.99, + "learning_rate": 9.163490471414243e-06, + "loss": 1.1848, + "theoretical_loss": 3.3279822751960104, + "tokens_seen": 3263561728 + }, + { + "epoch": 0.99, + "learning_rate": 9.147442326980943e-06, + "loss": 1.2381, + "theoretical_loss": 3.327977267574007, + "tokens_seen": 3263627264 + }, + { + "epoch": 0.99, + "objective/train/docs_used": 1185836, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.3209866285324097, + "objective/train/theoretical_loss": 3.3279722600807142, + "objective/train/tokens_used": 3284152800, + "theoretical_loss": 3.3279722600807142, + "tokens_seen": 3263692800 + }, + { + "epoch": 0.99, + "learning_rate": 9.131394182547644e-06, + "loss": 1.4255, + "theoretical_loss": 3.3279722600807142, + "tokens_seen": 3263692800 + }, + { + "epoch": 0.99, + "learning_rate": 9.115346038114343e-06, + "loss": 1.1818, + "theoretical_loss": 3.327967252716127, + "tokens_seen": 3263758336 + }, + { + "epoch": 0.99, + "learning_rate": 9.099297893681043e-06, + "loss": 1.2603, + "theoretical_loss": 3.3279622454802382, + "tokens_seen": 3263823872 + }, + { + "epoch": 0.99, + "learning_rate": 9.083249749247744e-06, + "loss": 1.2175, + "theoretical_loss": 3.3279572383730436, + "tokens_seen": 3263889408 + }, + { + "epoch": 0.99, + "learning_rate": 9.067201604814443e-06, + "loss": 1.235, + "theoretical_loss": 3.3279522313945353, + "tokens_seen": 3263954944 + }, + { + "epoch": 0.99, + "learning_rate": 9.051153460381144e-06, + "loss": 1.2451, + "theoretical_loss": 3.327947224544709, + "tokens_seen": 3264020480 + }, + { + "epoch": 0.99, + "learning_rate": 9.035105315947844e-06, + "loss": 1.2879, + "theoretical_loss": 3.3279422178235585, + "tokens_seen": 3264086016 + }, + { + "epoch": 0.99, + "learning_rate": 9.019057171514545e-06, + "loss": 1.3214, + "theoretical_loss": 3.327937211231077, + "tokens_seen": 3264151552 + }, + { + "epoch": 0.99, + "learning_rate": 9.003009027081244e-06, + "loss": 1.0906, + "theoretical_loss": 3.3279322047672597, + "tokens_seen": 3264217088 + }, + { + "epoch": 0.99, + "learning_rate": 8.986960882647944e-06, + "loss": 1.1837, + "theoretical_loss": 3.3279271984321, + "tokens_seen": 3264282624 + }, + { + "epoch": 0.99, + "learning_rate": 8.970912738214645e-06, + "loss": 1.1903, + "theoretical_loss": 3.3279221922255924, + "tokens_seen": 3264348160 + }, + { + "epoch": 0.99, + "learning_rate": 8.954864593781344e-06, + "loss": 1.1543, + "theoretical_loss": 3.327917186147731, + "tokens_seen": 3264413696 + }, + { + "epoch": 0.99, + "learning_rate": 8.938816449348045e-06, + "loss": 1.3292, + "theoretical_loss": 3.3279121801985094, + "tokens_seen": 3264479232 + }, + { + "epoch": 0.99, + "learning_rate": 8.922768304914744e-06, + "loss": 1.3748, + "theoretical_loss": 3.3279071743779225, + "tokens_seen": 3264544768 + }, + { + "epoch": 0.99, + "learning_rate": 8.906720160481446e-06, + "loss": 1.159, + "theoretical_loss": 3.327902168685964, + "tokens_seen": 3264610304 + }, + { + "epoch": 0.99, + "learning_rate": 8.890672016048145e-06, + "loss": 1.2403, + "theoretical_loss": 3.327897163122628, + "tokens_seen": 3264675840 + }, + { + "epoch": 0.99, + "learning_rate": 8.874623871614846e-06, + "loss": 1.282, + "theoretical_loss": 3.3278921576879084, + "tokens_seen": 3264741376 + }, + { + "epoch": 0.99, + "learning_rate": 8.858575727181544e-06, + "loss": 1.1639, + "theoretical_loss": 3.3278871523818, + "tokens_seen": 3264806912 + }, + { + "epoch": 0.99, + "learning_rate": 8.842527582748245e-06, + "loss": 1.199, + "theoretical_loss": 3.3278821472042965, + "tokens_seen": 3264872448 + }, + { + "epoch": 0.99, + "learning_rate": 8.826479438314946e-06, + "loss": 1.3232, + "theoretical_loss": 3.3278771421553914, + "tokens_seen": 3264937984 + }, + { + "epoch": 0.99, + "learning_rate": 8.810431293881645e-06, + "loss": 1.254, + "theoretical_loss": 3.32787213723508, + "tokens_seen": 3265003520 + }, + { + "epoch": 0.99, + "learning_rate": 8.794383149448345e-06, + "loss": 1.2716, + "theoretical_loss": 3.3278671324433557, + "tokens_seen": 3265069056 + }, + { + "epoch": 0.99, + "learning_rate": 8.778335005015046e-06, + "loss": 1.2335, + "theoretical_loss": 3.3278621277802127, + "tokens_seen": 3265134592 + }, + { + "epoch": 0.99, + "learning_rate": 8.762286860581747e-06, + "loss": 1.2962, + "theoretical_loss": 3.327857123245645, + "tokens_seen": 3265200128 + }, + { + "epoch": 0.99, + "learning_rate": 8.746238716148446e-06, + "loss": 1.2317, + "theoretical_loss": 3.327852118839647, + "tokens_seen": 3265265664 + }, + { + "epoch": 0.99, + "learning_rate": 8.730190571715144e-06, + "loss": 1.2313, + "theoretical_loss": 3.327847114562213, + "tokens_seen": 3265331200 + }, + { + "epoch": 0.99, + "learning_rate": 8.714142427281847e-06, + "loss": 1.2747, + "theoretical_loss": 3.327842110413336, + "tokens_seen": 3265396736 + }, + { + "epoch": 0.99, + "learning_rate": 8.698094282848546e-06, + "loss": 1.2287, + "theoretical_loss": 3.3278371063930114, + "tokens_seen": 3265462272 + }, + { + "epoch": 0.99, + "learning_rate": 8.682046138415246e-06, + "loss": 1.2489, + "theoretical_loss": 3.327832102501233, + "tokens_seen": 3265527808 + }, + { + "epoch": 0.99, + "learning_rate": 8.665997993981945e-06, + "loss": 1.1378, + "theoretical_loss": 3.3278270987379948, + "tokens_seen": 3265593344 + }, + { + "epoch": 0.99, + "learning_rate": 8.649949849548648e-06, + "loss": 1.1708, + "theoretical_loss": 3.327822095103291, + "tokens_seen": 3265658880 + }, + { + "epoch": 0.99, + "learning_rate": 8.633901705115347e-06, + "loss": 1.303, + "theoretical_loss": 3.3278170915971153, + "tokens_seen": 3265724416 + }, + { + "epoch": 0.99, + "learning_rate": 8.617853560682046e-06, + "loss": 1.2602, + "theoretical_loss": 3.3278120882194617, + "tokens_seen": 3265789952 + }, + { + "epoch": 0.99, + "learning_rate": 8.601805416248746e-06, + "loss": 1.2525, + "theoretical_loss": 3.3278070849703254, + "tokens_seen": 3265855488 + }, + { + "epoch": 0.99, + "learning_rate": 8.585757271815447e-06, + "loss": 1.172, + "theoretical_loss": 3.3278020818496996, + "tokens_seen": 3265921024 + }, + { + "epoch": 0.99, + "learning_rate": 8.569709127382147e-06, + "loss": 1.1973, + "theoretical_loss": 3.3277970788575786, + "tokens_seen": 3265986560 + }, + { + "epoch": 0.99, + "learning_rate": 8.553660982948846e-06, + "loss": 1.2104, + "theoretical_loss": 3.327792075993957, + "tokens_seen": 3266052096 + }, + { + "epoch": 0.99, + "learning_rate": 8.537612838515547e-06, + "loss": 1.2482, + "theoretical_loss": 3.327787073258828, + "tokens_seen": 3266117632 + }, + { + "epoch": 0.99, + "learning_rate": 8.521564694082248e-06, + "loss": 1.2686, + "theoretical_loss": 3.327782070652187, + "tokens_seen": 3266183168 + }, + { + "epoch": 0.99, + "learning_rate": 8.505516549648947e-06, + "loss": 1.1146, + "theoretical_loss": 3.3277770681740266, + "tokens_seen": 3266248704 + }, + { + "epoch": 0.99, + "learning_rate": 8.489468405215647e-06, + "loss": 1.2274, + "theoretical_loss": 3.327772065824342, + "tokens_seen": 3266314240 + }, + { + "epoch": 0.99, + "learning_rate": 8.473420260782348e-06, + "loss": 1.2974, + "theoretical_loss": 3.327767063603127, + "tokens_seen": 3266379776 + }, + { + "epoch": 0.99, + "learning_rate": 8.457372116349048e-06, + "loss": 1.2646, + "theoretical_loss": 3.3277620615103753, + "tokens_seen": 3266445312 + }, + { + "epoch": 0.99, + "learning_rate": 8.441323971915747e-06, + "loss": 1.1513, + "theoretical_loss": 3.327757059546082, + "tokens_seen": 3266510848 + }, + { + "epoch": 0.99, + "learning_rate": 8.425275827482448e-06, + "loss": 1.2181, + "theoretical_loss": 3.3277520577102404, + "tokens_seen": 3266576384 + }, + { + "epoch": 0.99, + "learning_rate": 8.409227683049149e-06, + "loss": 1.2548, + "theoretical_loss": 3.327747056002845, + "tokens_seen": 3266641920 + }, + { + "epoch": 0.99, + "learning_rate": 8.393179538615848e-06, + "loss": 1.1918, + "theoretical_loss": 3.32774205442389, + "tokens_seen": 3266707456 + }, + { + "epoch": 0.99, + "learning_rate": 8.377131394182548e-06, + "loss": 1.3435, + "theoretical_loss": 3.327737052973369, + "tokens_seen": 3266772992 + }, + { + "epoch": 0.99, + "learning_rate": 8.361083249749247e-06, + "loss": 1.214, + "theoretical_loss": 3.327732051651277, + "tokens_seen": 3266838528 + }, + { + "epoch": 0.99, + "learning_rate": 8.345035105315948e-06, + "loss": 1.254, + "theoretical_loss": 3.3277270504576073, + "tokens_seen": 3266904064 + }, + { + "epoch": 0.99, + "objective/train/docs_used": 1187115, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.265384554862976, + "objective/train/theoretical_loss": 3.3277220493923543, + "objective/train/tokens_used": 3287429600, + "theoretical_loss": 3.3277220493923543, + "tokens_seen": 3266969600 + }, + { + "epoch": 0.99, + "learning_rate": 8.328986960882648e-06, + "loss": 1.255, + "theoretical_loss": 3.3277220493923543, + "tokens_seen": 3266969600 + }, + { + "epoch": 0.99, + "learning_rate": 8.312938816449349e-06, + "loss": 1.3572, + "theoretical_loss": 3.3277170484555123, + "tokens_seen": 3267035136 + }, + { + "epoch": 0.99, + "learning_rate": 8.296890672016048e-06, + "loss": 1.3502, + "theoretical_loss": 3.327712047647075, + "tokens_seen": 3267100672 + }, + { + "epoch": 0.99, + "learning_rate": 8.280842527582749e-06, + "loss": 1.3192, + "theoretical_loss": 3.327707046967037, + "tokens_seen": 3267166208 + }, + { + "epoch": 0.99, + "learning_rate": 8.26479438314945e-06, + "loss": 1.2856, + "theoretical_loss": 3.3277020464153924, + "tokens_seen": 3267231744 + }, + { + "epoch": 0.99, + "learning_rate": 8.248746238716148e-06, + "loss": 1.2598, + "theoretical_loss": 3.327697045992135, + "tokens_seen": 3267297280 + }, + { + "epoch": 0.99, + "learning_rate": 8.232698094282849e-06, + "loss": 1.283, + "theoretical_loss": 3.327692045697259, + "tokens_seen": 3267362816 + }, + { + "epoch": 0.99, + "learning_rate": 8.21664994984955e-06, + "loss": 1.2265, + "theoretical_loss": 3.327687045530759, + "tokens_seen": 3267428352 + }, + { + "epoch": 0.99, + "learning_rate": 8.20060180541625e-06, + "loss": 1.164, + "theoretical_loss": 3.327682045492628, + "tokens_seen": 3267493888 + }, + { + "epoch": 0.99, + "learning_rate": 8.184553660982949e-06, + "loss": 1.2393, + "theoretical_loss": 3.3276770455828615, + "tokens_seen": 3267559424 + }, + { + "epoch": 0.99, + "learning_rate": 8.168505516549648e-06, + "loss": 1.2371, + "theoretical_loss": 3.327672045801453, + "tokens_seen": 3267624960 + }, + { + "epoch": 0.99, + "learning_rate": 8.15245737211635e-06, + "loss": 1.3585, + "theoretical_loss": 3.3276670461483966, + "tokens_seen": 3267690496 + }, + { + "epoch": 0.99, + "learning_rate": 8.13640922768305e-06, + "loss": 1.2313, + "theoretical_loss": 3.3276620466236864, + "tokens_seen": 3267756032 + }, + { + "epoch": 0.99, + "learning_rate": 8.12036108324975e-06, + "loss": 1.21, + "theoretical_loss": 3.3276570472273166, + "tokens_seen": 3267821568 + }, + { + "epoch": 0.99, + "learning_rate": 8.104312938816449e-06, + "loss": 1.0994, + "theoretical_loss": 3.3276520479592815, + "tokens_seen": 3267887104 + }, + { + "epoch": 0.99, + "learning_rate": 8.088264794383151e-06, + "loss": 1.2642, + "theoretical_loss": 3.327647048819575, + "tokens_seen": 3267952640 + }, + { + "epoch": 0.99, + "learning_rate": 8.07221664994985e-06, + "loss": 1.2935, + "theoretical_loss": 3.327642049808191, + "tokens_seen": 3268018176 + }, + { + "epoch": 0.99, + "learning_rate": 8.056168505516549e-06, + "loss": 1.2232, + "theoretical_loss": 3.3276370509251243, + "tokens_seen": 3268083712 + }, + { + "epoch": 0.99, + "learning_rate": 8.04012036108325e-06, + "loss": 1.2246, + "theoretical_loss": 3.3276320521703684, + "tokens_seen": 3268149248 + }, + { + "epoch": 0.99, + "learning_rate": 8.02407221664995e-06, + "loss": 1.1163, + "theoretical_loss": 3.327627053543918, + "tokens_seen": 3268214784 + }, + { + "epoch": 0.99, + "learning_rate": 8.008024072216651e-06, + "loss": 1.2359, + "theoretical_loss": 3.327622055045767, + "tokens_seen": 3268280320 + }, + { + "epoch": 0.99, + "learning_rate": 7.99197592778335e-06, + "loss": 1.1851, + "theoretical_loss": 3.3276170566759093, + "tokens_seen": 3268345856 + }, + { + "epoch": 0.99, + "learning_rate": 7.97592778335005e-06, + "loss": 1.2623, + "theoretical_loss": 3.327612058434339, + "tokens_seen": 3268411392 + }, + { + "epoch": 0.99, + "learning_rate": 7.959879638916751e-06, + "loss": 1.3516, + "theoretical_loss": 3.327607060321051, + "tokens_seen": 3268476928 + }, + { + "epoch": 0.99, + "learning_rate": 7.94383149448345e-06, + "loss": 1.3215, + "theoretical_loss": 3.3276020623360383, + "tokens_seen": 3268542464 + }, + { + "epoch": 0.99, + "learning_rate": 7.92778335005015e-06, + "loss": 1.2348, + "theoretical_loss": 3.3275970644792956, + "tokens_seen": 3268608000 + }, + { + "epoch": 0.99, + "learning_rate": 7.911735205616851e-06, + "loss": 1.1207, + "theoretical_loss": 3.3275920667508174, + "tokens_seen": 3268673536 + }, + { + "epoch": 0.99, + "learning_rate": 7.895687061183552e-06, + "loss": 1.2865, + "theoretical_loss": 3.327587069150597, + "tokens_seen": 3268739072 + }, + { + "epoch": 0.99, + "learning_rate": 7.879638916750251e-06, + "loss": 1.2215, + "theoretical_loss": 3.3275820716786297, + "tokens_seen": 3268804608 + }, + { + "epoch": 0.99, + "learning_rate": 7.863590772316952e-06, + "loss": 1.2288, + "theoretical_loss": 3.3275770743349087, + "tokens_seen": 3268870144 + }, + { + "epoch": 0.99, + "learning_rate": 7.84754262788365e-06, + "loss": 1.1206, + "theoretical_loss": 3.327572077119428, + "tokens_seen": 3268935680 + }, + { + "epoch": 0.99, + "learning_rate": 7.831494483450351e-06, + "loss": 1.2303, + "theoretical_loss": 3.327567080032183, + "tokens_seen": 3269001216 + }, + { + "epoch": 0.99, + "learning_rate": 7.815446339017052e-06, + "loss": 1.2835, + "theoretical_loss": 3.327562083073166, + "tokens_seen": 3269066752 + }, + { + "epoch": 0.99, + "learning_rate": 7.79939819458375e-06, + "loss": 1.4483, + "theoretical_loss": 3.327557086242373, + "tokens_seen": 3269132288 + }, + { + "epoch": 0.99, + "learning_rate": 7.783350050150451e-06, + "loss": 1.2028, + "theoretical_loss": 3.3275520895397968, + "tokens_seen": 3269197824 + }, + { + "epoch": 0.99, + "learning_rate": 7.767301905717152e-06, + "loss": 1.1839, + "theoretical_loss": 3.327547092965432, + "tokens_seen": 3269263360 + }, + { + "epoch": 0.99, + "learning_rate": 7.751253761283853e-06, + "loss": 1.137, + "theoretical_loss": 3.3275420965192724, + "tokens_seen": 3269328896 + }, + { + "epoch": 0.99, + "learning_rate": 7.735205616850552e-06, + "loss": 1.2537, + "theoretical_loss": 3.327537100201313, + "tokens_seen": 3269394432 + }, + { + "epoch": 0.99, + "learning_rate": 7.719157472417252e-06, + "loss": 1.2281, + "theoretical_loss": 3.3275321040115475, + "tokens_seen": 3269459968 + }, + { + "epoch": 0.99, + "learning_rate": 7.703109327983953e-06, + "loss": 1.2421, + "theoretical_loss": 3.3275271079499693, + "tokens_seen": 3269525504 + }, + { + "epoch": 0.99, + "learning_rate": 7.687061183550652e-06, + "loss": 1.3165, + "theoretical_loss": 3.327522112016574, + "tokens_seen": 3269591040 + }, + { + "epoch": 0.99, + "learning_rate": 7.671013039117352e-06, + "loss": 1.2555, + "theoretical_loss": 3.3275171162113546, + "tokens_seen": 3269656576 + }, + { + "epoch": 0.99, + "learning_rate": 7.654964894684053e-06, + "loss": 1.1418, + "theoretical_loss": 3.3275121205343052, + "tokens_seen": 3269722112 + }, + { + "epoch": 0.99, + "learning_rate": 7.638916750250752e-06, + "loss": 1.2099, + "theoretical_loss": 3.327507124985421, + "tokens_seen": 3269787648 + }, + { + "epoch": 0.99, + "learning_rate": 7.622868605817453e-06, + "loss": 1.2278, + "theoretical_loss": 3.327502129564695, + "tokens_seen": 3269853184 + }, + { + "epoch": 0.99, + "learning_rate": 7.6068204613841525e-06, + "loss": 1.233, + "theoretical_loss": 3.327497134272122, + "tokens_seen": 3269918720 + }, + { + "epoch": 0.99, + "learning_rate": 7.590772316950853e-06, + "loss": 1.3063, + "theoretical_loss": 3.3274921391076955, + "tokens_seen": 3269984256 + }, + { + "epoch": 0.99, + "learning_rate": 7.574724172517553e-06, + "loss": 1.2059, + "theoretical_loss": 3.327487144071411, + "tokens_seen": 3270049792 + }, + { + "epoch": 0.99, + "learning_rate": 7.5586760280842535e-06, + "loss": 1.2072, + "theoretical_loss": 3.327482149163261, + "tokens_seen": 3270115328 + }, + { + "epoch": 0.99, + "learning_rate": 7.542627883650953e-06, + "loss": 1.2754, + "theoretical_loss": 3.3274771543832404, + "tokens_seen": 3270180864 + }, + { + "epoch": 0.99, + "objective/train/docs_used": 1188255, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.3096064329147339, + "objective/train/theoretical_loss": 3.327472159731344, + "objective/train/tokens_used": 3290706400, + "theoretical_loss": 3.327472159731344, + "tokens_seen": 3270246400 + }, + { + "epoch": 0.99, + "learning_rate": 7.526579739217652e-06, + "loss": 1.295, + "theoretical_loss": 3.327472159731344, + "tokens_seen": 3270246400 + }, + { + "epoch": 0.99, + "learning_rate": 7.510531594784354e-06, + "loss": 1.1607, + "theoretical_loss": 3.327467165207565, + "tokens_seen": 3270311936 + }, + { + "epoch": 0.99, + "learning_rate": 7.494483450351053e-06, + "loss": 1.1937, + "theoretical_loss": 3.3274621708118977, + "tokens_seen": 3270377472 + }, + { + "epoch": 0.99, + "learning_rate": 7.478435305917754e-06, + "loss": 1.2118, + "theoretical_loss": 3.3274571765443364, + "tokens_seen": 3270443008 + }, + { + "epoch": 0.99, + "learning_rate": 7.462387161484453e-06, + "loss": 1.2424, + "theoretical_loss": 3.327452182404875, + "tokens_seen": 3270508544 + }, + { + "epoch": 0.99, + "learning_rate": 7.446339017051155e-06, + "loss": 1.2157, + "theoretical_loss": 3.327447188393508, + "tokens_seen": 3270574080 + }, + { + "epoch": 0.99, + "learning_rate": 7.4302908726178535e-06, + "loss": 1.2629, + "theoretical_loss": 3.3274421945102297, + "tokens_seen": 3270639616 + }, + { + "epoch": 0.99, + "learning_rate": 7.414242728184553e-06, + "loss": 1.3776, + "theoretical_loss": 3.3274372007550337, + "tokens_seen": 3270705152 + }, + { + "epoch": 0.99, + "learning_rate": 7.398194583751254e-06, + "loss": 1.2025, + "theoretical_loss": 3.3274322071279143, + "tokens_seen": 3270770688 + }, + { + "epoch": 0.99, + "learning_rate": 7.382146439317954e-06, + "loss": 1.2478, + "theoretical_loss": 3.327427213628866, + "tokens_seen": 3270836224 + }, + { + "epoch": 0.99, + "learning_rate": 7.366098294884654e-06, + "loss": 1.2052, + "theoretical_loss": 3.3274222202578825, + "tokens_seen": 3270901760 + }, + { + "epoch": 0.99, + "learning_rate": 7.350050150451354e-06, + "loss": 1.2104, + "theoretical_loss": 3.3274172270149585, + "tokens_seen": 3270967296 + }, + { + "epoch": 0.99, + "learning_rate": 7.334002006018055e-06, + "loss": 1.2492, + "theoretical_loss": 3.3274122339000876, + "tokens_seen": 3271032832 + }, + { + "epoch": 0.99, + "learning_rate": 7.317953861584755e-06, + "loss": 1.218, + "theoretical_loss": 3.3274072409132645, + "tokens_seen": 3271098368 + }, + { + "epoch": 0.99, + "learning_rate": 7.301905717151454e-06, + "loss": 1.1816, + "theoretical_loss": 3.3274022480544825, + "tokens_seen": 3271163904 + }, + { + "epoch": 0.99, + "learning_rate": 7.285857572718155e-06, + "loss": 1.2526, + "theoretical_loss": 3.3273972553237363, + "tokens_seen": 3271229440 + }, + { + "epoch": 0.99, + "learning_rate": 7.269809428284855e-06, + "loss": 1.2949, + "theoretical_loss": 3.3273922627210206, + "tokens_seen": 3271294976 + }, + { + "epoch": 0.99, + "learning_rate": 7.2537612838515554e-06, + "loss": 1.1914, + "theoretical_loss": 3.3273872702463283, + "tokens_seen": 3271360512 + }, + { + "epoch": 0.99, + "learning_rate": 7.237713139418255e-06, + "loss": 1.2187, + "theoretical_loss": 3.3273822778996545, + "tokens_seen": 3271426048 + }, + { + "epoch": 0.99, + "learning_rate": 7.221664994984956e-06, + "loss": 1.2274, + "theoretical_loss": 3.327377285680993, + "tokens_seen": 3271491584 + }, + { + "epoch": 0.99, + "learning_rate": 7.205616850551656e-06, + "loss": 1.3006, + "theoretical_loss": 3.3273722935903383, + "tokens_seen": 3271557120 + }, + { + "epoch": 0.99, + "learning_rate": 7.189568706118355e-06, + "loss": 1.2668, + "theoretical_loss": 3.327367301627684, + "tokens_seen": 3271622656 + }, + { + "epoch": 0.99, + "learning_rate": 7.173520561685056e-06, + "loss": 1.3006, + "theoretical_loss": 3.327362309793025, + "tokens_seen": 3271688192 + }, + { + "epoch": 0.99, + "learning_rate": 7.157472417251755e-06, + "loss": 1.2081, + "theoretical_loss": 3.3273573180863543, + "tokens_seen": 3271753728 + }, + { + "epoch": 0.99, + "learning_rate": 7.141424272818456e-06, + "loss": 1.2376, + "theoretical_loss": 3.327352326507667, + "tokens_seen": 3271819264 + }, + { + "epoch": 0.99, + "learning_rate": 7.125376128385155e-06, + "loss": 1.2256, + "theoretical_loss": 3.327347335056957, + "tokens_seen": 3271884800 + }, + { + "epoch": 0.99, + "learning_rate": 7.109327983951856e-06, + "loss": 1.1661, + "theoretical_loss": 3.3273423437342187, + "tokens_seen": 3271950336 + }, + { + "epoch": 0.99, + "learning_rate": 7.093279839518556e-06, + "loss": 1.2634, + "theoretical_loss": 3.3273373525394456, + "tokens_seen": 3272015872 + }, + { + "epoch": 0.99, + "learning_rate": 7.077231695085256e-06, + "loss": 1.2369, + "theoretical_loss": 3.327332361472633, + "tokens_seen": 3272081408 + }, + { + "epoch": 0.99, + "learning_rate": 7.061183550651956e-06, + "loss": 1.1223, + "theoretical_loss": 3.327327370533774, + "tokens_seen": 3272146944 + }, + { + "epoch": 0.99, + "learning_rate": 7.045135406218656e-06, + "loss": 1.3161, + "theoretical_loss": 3.3273223797228626, + "tokens_seen": 3272212480 + }, + { + "epoch": 0.99, + "learning_rate": 7.029087261785357e-06, + "loss": 1.2637, + "theoretical_loss": 3.3273173890398935, + "tokens_seen": 3272278016 + }, + { + "epoch": 0.99, + "learning_rate": 7.0130391173520565e-06, + "loss": 1.2225, + "theoretical_loss": 3.327312398484861, + "tokens_seen": 3272343552 + }, + { + "epoch": 0.99, + "learning_rate": 6.996990972918757e-06, + "loss": 1.2938, + "theoretical_loss": 3.3273074080577594, + "tokens_seen": 3272409088 + }, + { + "epoch": 0.99, + "learning_rate": 6.980942828485457e-06, + "loss": 1.2213, + "theoretical_loss": 3.3273024177585824, + "tokens_seen": 3272474624 + }, + { + "epoch": 0.99, + "learning_rate": 6.964894684052156e-06, + "loss": 1.1367, + "theoretical_loss": 3.327297427587324, + "tokens_seen": 3272540160 + }, + { + "epoch": 0.99, + "learning_rate": 6.948846539618857e-06, + "loss": 1.2852, + "theoretical_loss": 3.3272924375439787, + "tokens_seen": 3272605696 + }, + { + "epoch": 0.99, + "learning_rate": 6.932798395185556e-06, + "loss": 1.1791, + "theoretical_loss": 3.3272874476285406, + "tokens_seen": 3272671232 + }, + { + "epoch": 0.99, + "learning_rate": 6.916750250752258e-06, + "loss": 1.2644, + "theoretical_loss": 3.3272824578410036, + "tokens_seen": 3272736768 + }, + { + "epoch": 0.99, + "learning_rate": 6.900702106318957e-06, + "loss": 1.2294, + "theoretical_loss": 3.3272774681813626, + "tokens_seen": 3272802304 + }, + { + "epoch": 0.99, + "learning_rate": 6.884653961885658e-06, + "loss": 1.25, + "theoretical_loss": 3.3272724786496113, + "tokens_seen": 3272867840 + }, + { + "epoch": 0.99, + "learning_rate": 6.868605817452357e-06, + "loss": 1.1686, + "theoretical_loss": 3.3272674892457434, + "tokens_seen": 3272933376 + }, + { + "epoch": 0.99, + "learning_rate": 6.852557673019057e-06, + "loss": 1.2191, + "theoretical_loss": 3.3272624999697538, + "tokens_seen": 3272998912 + }, + { + "epoch": 0.99, + "learning_rate": 6.8365095285857575e-06, + "loss": 1.2562, + "theoretical_loss": 3.3272575108216365, + "tokens_seen": 3273064448 + }, + { + "epoch": 0.99, + "learning_rate": 6.820461384152457e-06, + "loss": 1.2781, + "theoretical_loss": 3.327252521801385, + "tokens_seen": 3273129984 + }, + { + "epoch": 0.99, + "learning_rate": 6.804413239719158e-06, + "loss": 1.2193, + "theoretical_loss": 3.3272475329089946, + "tokens_seen": 3273195520 + }, + { + "epoch": 0.99, + "learning_rate": 6.788365095285858e-06, + "loss": 1.2361, + "theoretical_loss": 3.3272425441444584, + "tokens_seen": 3273261056 + }, + { + "epoch": 0.99, + "learning_rate": 6.772316950852558e-06, + "loss": 1.2735, + "theoretical_loss": 3.327237555507771, + "tokens_seen": 3273326592 + }, + { + "epoch": 0.99, + "learning_rate": 6.756268806419258e-06, + "loss": 1.2941, + "theoretical_loss": 3.327232566998927, + "tokens_seen": 3273392128 + }, + { + "epoch": 0.99, + "learning_rate": 6.740220661985958e-06, + "loss": 1.1933, + "theoretical_loss": 3.3272275786179195, + "tokens_seen": 3273457664 + }, + { + "epoch": 0.99, + "objective/train/docs_used": 1189374, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.217599630355835, + "objective/train/theoretical_loss": 3.3272225903647437, + "objective/train/tokens_used": 3293983200, + "theoretical_loss": 3.3272225903647437, + "tokens_seen": 3273523200 + }, + { + "epoch": 0.99, + "learning_rate": 6.724172517552659e-06, + "loss": 1.2898, + "theoretical_loss": 3.3272225903647437, + "tokens_seen": 3273523200 + }, + { + "epoch": 0.99, + "learning_rate": 6.708124373119358e-06, + "loss": 1.2034, + "theoretical_loss": 3.3272176022393936, + "tokens_seen": 3273588736 + }, + { + "epoch": 0.99, + "learning_rate": 6.692076228686059e-06, + "loss": 1.2652, + "theoretical_loss": 3.327212614241863, + "tokens_seen": 3273654272 + }, + { + "epoch": 0.99, + "learning_rate": 6.676028084252759e-06, + "loss": 1.3307, + "theoretical_loss": 3.3272076263721457, + "tokens_seen": 3273719808 + }, + { + "epoch": 0.99, + "learning_rate": 6.6599799398194594e-06, + "loss": 1.2282, + "theoretical_loss": 3.327202638630237, + "tokens_seen": 3273785344 + }, + { + "epoch": 0.99, + "learning_rate": 6.643931795386158e-06, + "loss": 1.3026, + "theoretical_loss": 3.32719765101613, + "tokens_seen": 3273850880 + }, + { + "epoch": 0.99, + "learning_rate": 6.627883650952858e-06, + "loss": 1.1658, + "theoretical_loss": 3.3271926635298197, + "tokens_seen": 3273916416 + }, + { + "epoch": 0.99, + "learning_rate": 6.611835506519559e-06, + "loss": 1.2477, + "theoretical_loss": 3.3271876761712993, + "tokens_seen": 3273981952 + }, + { + "epoch": 0.99, + "learning_rate": 6.5957873620862586e-06, + "loss": 1.0766, + "theoretical_loss": 3.3271826889405642, + "tokens_seen": 3274047488 + }, + { + "epoch": 0.99, + "learning_rate": 6.579739217652959e-06, + "loss": 1.2523, + "theoretical_loss": 3.3271777018376074, + "tokens_seen": 3274113024 + }, + { + "epoch": 0.99, + "learning_rate": 6.563691073219659e-06, + "loss": 1.2815, + "theoretical_loss": 3.327172714862424, + "tokens_seen": 3274178560 + }, + { + "epoch": 0.99, + "learning_rate": 6.54764292878636e-06, + "loss": 1.2557, + "theoretical_loss": 3.3271677280150076, + "tokens_seen": 3274244096 + }, + { + "epoch": 0.99, + "learning_rate": 6.531594784353059e-06, + "loss": 1.1911, + "theoretical_loss": 3.327162741295352, + "tokens_seen": 3274309632 + }, + { + "epoch": 0.99, + "learning_rate": 6.515546639919759e-06, + "loss": 1.1772, + "theoretical_loss": 3.3271577547034523, + "tokens_seen": 3274375168 + }, + { + "epoch": 0.99, + "learning_rate": 6.49949849548646e-06, + "loss": 1.3168, + "theoretical_loss": 3.3271527682393023, + "tokens_seen": 3274440704 + }, + { + "epoch": 0.99, + "learning_rate": 6.48345035105316e-06, + "loss": 1.2517, + "theoretical_loss": 3.3271477819028963, + "tokens_seen": 3274506240 + }, + { + "epoch": 0.99, + "learning_rate": 6.46740220661986e-06, + "loss": 1.1399, + "theoretical_loss": 3.327142795694228, + "tokens_seen": 3274571776 + }, + { + "epoch": 0.99, + "learning_rate": 6.45135406218656e-06, + "loss": 1.2927, + "theoretical_loss": 3.327137809613292, + "tokens_seen": 3274637312 + }, + { + "epoch": 0.99, + "learning_rate": 6.435305917753261e-06, + "loss": 1.1686, + "theoretical_loss": 3.327132823660082, + "tokens_seen": 3274702848 + }, + { + "epoch": 0.99, + "learning_rate": 6.4192577733199605e-06, + "loss": 1.2163, + "theoretical_loss": 3.3271278378345928, + "tokens_seen": 3274768384 + }, + { + "epoch": 0.99, + "learning_rate": 6.403209628886659e-06, + "loss": 1.2162, + "theoretical_loss": 3.3271228521368186, + "tokens_seen": 3274833920 + }, + { + "epoch": 0.99, + "learning_rate": 6.387161484453361e-06, + "loss": 1.229, + "theoretical_loss": 3.327117866566753, + "tokens_seen": 3274899456 + }, + { + "epoch": 0.99, + "learning_rate": 6.37111334002006e-06, + "loss": 1.319, + "theoretical_loss": 3.32711288112439, + "tokens_seen": 3274964992 + }, + { + "epoch": 0.99, + "learning_rate": 6.355065195586761e-06, + "loss": 1.1764, + "theoretical_loss": 3.327107895809725, + "tokens_seen": 3275030528 + }, + { + "epoch": 0.99, + "learning_rate": 6.33901705115346e-06, + "loss": 1.2455, + "theoretical_loss": 3.3271029106227505, + "tokens_seen": 3275096064 + }, + { + "epoch": 0.99, + "learning_rate": 6.322968906720162e-06, + "loss": 1.2458, + "theoretical_loss": 3.327097925563462, + "tokens_seen": 3275161600 + }, + { + "epoch": 0.99, + "learning_rate": 6.306920762286861e-06, + "loss": 1.121, + "theoretical_loss": 3.3270929406318532, + "tokens_seen": 3275227136 + }, + { + "epoch": 0.99, + "learning_rate": 6.2908726178535605e-06, + "loss": 1.3003, + "theoretical_loss": 3.3270879558279183, + "tokens_seen": 3275292672 + }, + { + "epoch": 0.99, + "learning_rate": 6.274824473420261e-06, + "loss": 1.0635, + "theoretical_loss": 3.3270829711516514, + "tokens_seen": 3275358208 + }, + { + "epoch": 0.99, + "learning_rate": 6.258776328986961e-06, + "loss": 1.2481, + "theoretical_loss": 3.327077986603047, + "tokens_seen": 3275423744 + }, + { + "epoch": 0.99, + "learning_rate": 6.242728184553661e-06, + "loss": 1.1576, + "theoretical_loss": 3.3270730021820984, + "tokens_seen": 3275489280 + }, + { + "epoch": 0.99, + "learning_rate": 6.226680040120361e-06, + "loss": 1.2316, + "theoretical_loss": 3.3270680178888012, + "tokens_seen": 3275554816 + }, + { + "epoch": 0.99, + "learning_rate": 6.210631895687061e-06, + "loss": 1.2617, + "theoretical_loss": 3.3270630337231486, + "tokens_seen": 3275620352 + }, + { + "epoch": 0.99, + "learning_rate": 6.194583751253762e-06, + "loss": 1.1779, + "theoretical_loss": 3.3270580496851347, + "tokens_seen": 3275685888 + }, + { + "epoch": 0.99, + "learning_rate": 6.1785356068204615e-06, + "loss": 1.2305, + "theoretical_loss": 3.327053065774754, + "tokens_seen": 3275751424 + }, + { + "epoch": 0.99, + "learning_rate": 6.162487462387162e-06, + "loss": 1.0387, + "theoretical_loss": 3.3270480819920003, + "tokens_seen": 3275816960 + }, + { + "epoch": 0.99, + "learning_rate": 6.146439317953862e-06, + "loss": 1.309, + "theoretical_loss": 3.3270430983368686, + "tokens_seen": 3275882496 + }, + { + "epoch": 0.99, + "learning_rate": 6.130391173520562e-06, + "loss": 1.1578, + "theoretical_loss": 3.327038114809352, + "tokens_seen": 3275948032 + }, + { + "epoch": 0.99, + "learning_rate": 6.1143430290872615e-06, + "loss": 1.2878, + "theoretical_loss": 3.327033131409446, + "tokens_seen": 3276013568 + }, + { + "epoch": 0.99, + "learning_rate": 6.098294884653962e-06, + "loss": 1.3262, + "theoretical_loss": 3.3270281481371438, + "tokens_seen": 3276079104 + }, + { + "epoch": 0.99, + "learning_rate": 6.082246740220662e-06, + "loss": 1.2777, + "theoretical_loss": 3.3270231649924398, + "tokens_seen": 3276144640 + }, + { + "epoch": 0.99, + "learning_rate": 6.0661985957873626e-06, + "loss": 1.2108, + "theoretical_loss": 3.327018181975328, + "tokens_seen": 3276210176 + }, + { + "epoch": 0.99, + "learning_rate": 6.050150451354062e-06, + "loss": 1.2142, + "theoretical_loss": 3.3270131990858025, + "tokens_seen": 3276275712 + }, + { + "epoch": 0.99, + "learning_rate": 6.034102306920763e-06, + "loss": 1.2676, + "theoretical_loss": 3.3270082163238586, + "tokens_seen": 3276341248 + }, + { + "epoch": 0.99, + "learning_rate": 6.018054162487463e-06, + "loss": 1.2728, + "theoretical_loss": 3.327003233689489, + "tokens_seen": 3276406784 + }, + { + "epoch": 0.99, + "learning_rate": 6.0020060180541626e-06, + "loss": 1.1943, + "theoretical_loss": 3.3269982511826885, + "tokens_seen": 3276472320 + }, + { + "epoch": 0.99, + "learning_rate": 5.985957873620863e-06, + "loss": 1.3003, + "theoretical_loss": 3.3269932688034514, + "tokens_seen": 3276537856 + }, + { + "epoch": 0.99, + "learning_rate": 5.969909729187563e-06, + "loss": 1.1686, + "theoretical_loss": 3.3269882865517717, + "tokens_seen": 3276603392 + }, + { + "epoch": 0.99, + "learning_rate": 5.953861584754264e-06, + "loss": 1.2165, + "theoretical_loss": 3.326983304427644, + "tokens_seen": 3276668928 + }, + { + "epoch": 0.99, + "learning_rate": 5.937813440320963e-06, + "loss": 1.2002, + "theoretical_loss": 3.326978322431062, + "tokens_seen": 3276734464 + }, + { + "debugging/Compilability": 1.0, + "debugging/distinct-1-grams": 0.7558328814763761, + "debugging/entropy-1-grams": 5.085795856196755, + "debugging/length": 437.0, + "debugging/num_segments": 8, + "debugging/score": 0.011514534081995126, + "debugging/score_std": 0.01330564678502946, + "epoch": 0.99, + "objective/train/docs_used": 1190634, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.2391278743743896, + "objective/train/theoretical_loss": 3.3269733405620197, + "objective/train/tokens_used": 3297260000, + "theoretical_loss": 3.3269733405620197, + "tokens_seen": 3276800000 + }, + { + "epoch": 0.99, + "learning_rate": 5.921765295887663e-06, + "loss": 1.2761, + "theoretical_loss": 3.3269733405620197, + "tokens_seen": 3276800000 + }, + { + "epoch": 0.99, + "learning_rate": 5.905717151454363e-06, + "loss": 1.2402, + "theoretical_loss": 3.326968358820512, + "tokens_seen": 3276865536 + }, + { + "epoch": 0.99, + "learning_rate": 5.889669007021064e-06, + "loss": 1.246, + "theoretical_loss": 3.326963377206533, + "tokens_seen": 3276931072 + }, + { + "epoch": 0.99, + "learning_rate": 5.873620862587763e-06, + "loss": 1.1395, + "theoretical_loss": 3.3269583957200757, + "tokens_seen": 3276996608 + }, + { + "epoch": 0.99, + "learning_rate": 5.857572718154464e-06, + "loss": 1.3316, + "theoretical_loss": 3.3269534143611357, + "tokens_seen": 3277062144 + }, + { + "epoch": 0.99, + "learning_rate": 5.841524573721164e-06, + "loss": 1.2407, + "theoretical_loss": 3.3269484331297066, + "tokens_seen": 3277127680 + }, + { + "epoch": 0.99, + "learning_rate": 5.8254764292878645e-06, + "loss": 1.3167, + "theoretical_loss": 3.3269434520257826, + "tokens_seen": 3277193216 + }, + { + "epoch": 0.99, + "learning_rate": 5.809428284854563e-06, + "loss": 1.4143, + "theoretical_loss": 3.326938471049358, + "tokens_seen": 3277258752 + }, + { + "epoch": 0.99, + "learning_rate": 5.793380140421264e-06, + "loss": 1.2422, + "theoretical_loss": 3.326933490200427, + "tokens_seen": 3277324288 + }, + { + "epoch": 0.99, + "learning_rate": 5.777331995987964e-06, + "loss": 1.1941, + "theoretical_loss": 3.326928509478983, + "tokens_seen": 3277389824 + }, + { + "epoch": 0.99, + "learning_rate": 5.7612838515546645e-06, + "loss": 1.2231, + "theoretical_loss": 3.326923528885022, + "tokens_seen": 3277455360 + }, + { + "epoch": 0.99, + "learning_rate": 5.745235707121364e-06, + "loss": 1.164, + "theoretical_loss": 3.3269185484185364, + "tokens_seen": 3277520896 + }, + { + "epoch": 0.99, + "learning_rate": 5.729187562688065e-06, + "loss": 1.3381, + "theoretical_loss": 3.326913568079521, + "tokens_seen": 3277586432 + }, + { + "epoch": 0.99, + "learning_rate": 5.713139418254765e-06, + "loss": 1.2554, + "theoretical_loss": 3.3269085878679703, + "tokens_seen": 3277651968 + }, + { + "epoch": 0.99, + "learning_rate": 5.6970912738214645e-06, + "loss": 1.288, + "theoretical_loss": 3.326903607783878, + "tokens_seen": 3277717504 + }, + { + "epoch": 0.99, + "learning_rate": 5.681043129388164e-06, + "loss": 1.2332, + "theoretical_loss": 3.326898627827239, + "tokens_seen": 3277783040 + }, + { + "epoch": 0.99, + "learning_rate": 5.664994984954865e-06, + "loss": 1.1116, + "theoretical_loss": 3.326893647998047, + "tokens_seen": 3277848576 + }, + { + "epoch": 0.99, + "learning_rate": 5.648946840521565e-06, + "loss": 1.1847, + "theoretical_loss": 3.326888668296296, + "tokens_seen": 3277914112 + }, + { + "epoch": 0.99, + "learning_rate": 5.632898696088265e-06, + "loss": 1.2665, + "theoretical_loss": 3.3268836887219804, + "tokens_seen": 3277979648 + }, + { + "epoch": 0.99, + "learning_rate": 5.616850551654965e-06, + "loss": 1.241, + "theoretical_loss": 3.326878709275094, + "tokens_seen": 3278045184 + }, + { + "epoch": 0.99, + "learning_rate": 5.600802407221666e-06, + "loss": 1.1802, + "theoretical_loss": 3.326873729955632, + "tokens_seen": 3278110720 + }, + { + "epoch": 0.99, + "learning_rate": 5.584754262788365e-06, + "loss": 1.2655, + "theoretical_loss": 3.3268687507635875, + "tokens_seen": 3278176256 + }, + { + "epoch": 0.99, + "learning_rate": 5.568706118355065e-06, + "loss": 1.2476, + "theoretical_loss": 3.3268637716989558, + "tokens_seen": 3278241792 + }, + { + "epoch": 0.99, + "learning_rate": 5.552657973921765e-06, + "loss": 1.2577, + "theoretical_loss": 3.3268587927617297, + "tokens_seen": 3278307328 + }, + { + "epoch": 0.99, + "learning_rate": 5.536609829488466e-06, + "loss": 1.2733, + "theoretical_loss": 3.326853813951905, + "tokens_seen": 3278372864 + }, + { + "epoch": 0.99, + "learning_rate": 5.5205616850551655e-06, + "loss": 1.3158, + "theoretical_loss": 3.326848835269474, + "tokens_seen": 3278438400 + }, + { + "epoch": 0.99, + "learning_rate": 5.504513540621866e-06, + "loss": 1.2301, + "theoretical_loss": 3.326843856714433, + "tokens_seen": 3278503936 + }, + { + "epoch": 0.99, + "learning_rate": 5.488465396188566e-06, + "loss": 1.3075, + "theoretical_loss": 3.3268388782867744, + "tokens_seen": 3278569472 + }, + { + "epoch": 0.99, + "learning_rate": 5.472417251755266e-06, + "loss": 1.222, + "theoretical_loss": 3.3268338999864935, + "tokens_seen": 3278635008 + }, + { + "epoch": 0.99, + "learning_rate": 5.456369107321966e-06, + "loss": 1.1603, + "theoretical_loss": 3.326828921813584, + "tokens_seen": 3278700544 + }, + { + "epoch": 0.99, + "learning_rate": 5.440320962888666e-06, + "loss": 1.2302, + "theoretical_loss": 3.3268239437680402, + "tokens_seen": 3278766080 + }, + { + "epoch": 0.99, + "learning_rate": 5.424272818455367e-06, + "loss": 1.1997, + "theoretical_loss": 3.3268189658498564, + "tokens_seen": 3278831616 + }, + { + "epoch": 0.99, + "learning_rate": 5.4082246740220666e-06, + "loss": 1.2304, + "theoretical_loss": 3.326813988059027, + "tokens_seen": 3278897152 + }, + { + "epoch": 0.99, + "learning_rate": 5.392176529588766e-06, + "loss": 1.2396, + "theoretical_loss": 3.3268090103955457, + "tokens_seen": 3278962688 + }, + { + "epoch": 0.99, + "learning_rate": 5.376128385155467e-06, + "loss": 1.3339, + "theoretical_loss": 3.326804032859407, + "tokens_seen": 3279028224 + }, + { + "epoch": 0.99, + "learning_rate": 5.360080240722167e-06, + "loss": 1.1959, + "theoretical_loss": 3.3267990554506044, + "tokens_seen": 3279093760 + }, + { + "epoch": 0.99, + "learning_rate": 5.3440320962888666e-06, + "loss": 1.1397, + "theoretical_loss": 3.326794078169133, + "tokens_seen": 3279159296 + }, + { + "epoch": 0.99, + "learning_rate": 5.327983951855567e-06, + "loss": 1.2099, + "theoretical_loss": 3.3267891010149873, + "tokens_seen": 3279224832 + }, + { + "epoch": 0.99, + "learning_rate": 5.311935807422267e-06, + "loss": 1.1288, + "theoretical_loss": 3.32678412398816, + "tokens_seen": 3279290368 + }, + { + "epoch": 0.99, + "learning_rate": 5.295887662988968e-06, + "loss": 1.2803, + "theoretical_loss": 3.3267791470886467, + "tokens_seen": 3279355904 + }, + { + "epoch": 0.99, + "learning_rate": 5.279839518555667e-06, + "loss": 1.1445, + "theoretical_loss": 3.3267741703164413, + "tokens_seen": 3279421440 + }, + { + "epoch": 0.99, + "learning_rate": 5.263791374122368e-06, + "loss": 1.1829, + "theoretical_loss": 3.326769193671537, + "tokens_seen": 3279486976 + }, + { + "epoch": 0.99, + "learning_rate": 5.247743229689067e-06, + "loss": 1.2885, + "theoretical_loss": 3.32676421715393, + "tokens_seen": 3279552512 + }, + { + "epoch": 0.99, + "learning_rate": 5.231695085255768e-06, + "loss": 1.296, + "theoretical_loss": 3.326759240763612, + "tokens_seen": 3279618048 + }, + { + "epoch": 0.99, + "learning_rate": 5.215646940822467e-06, + "loss": 1.2789, + "theoretical_loss": 3.3267542645005794, + "tokens_seen": 3279683584 + }, + { + "epoch": 0.99, + "learning_rate": 5.199598796389168e-06, + "loss": 1.2449, + "theoretical_loss": 3.3267492883648253, + "tokens_seen": 3279749120 + }, + { + "epoch": 0.99, + "learning_rate": 5.183550651955868e-06, + "loss": 1.2648, + "theoretical_loss": 3.326744312356344, + "tokens_seen": 3279814656 + }, + { + "epoch": 0.99, + "learning_rate": 5.1675025075225685e-06, + "loss": 1.2887, + "theoretical_loss": 3.32673933647513, + "tokens_seen": 3279880192 + }, + { + "epoch": 0.99, + "learning_rate": 5.151454363089268e-06, + "loss": 1.174, + "theoretical_loss": 3.326734360721177, + "tokens_seen": 3279945728 + }, + { + "epoch": 0.99, + "learning_rate": 5.135406218655968e-06, + "loss": 1.3037, + "theoretical_loss": 3.3267293850944797, + "tokens_seen": 3280011264 + }, + { + "epoch": 0.99, + "objective/train/docs_used": 1191833, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.205444574356079, + "objective/train/theoretical_loss": 3.326724409595032, + "objective/train/tokens_used": 3300536800, + "theoretical_loss": 3.326724409595032, + "tokens_seen": 3280076800 + }, + { + "epoch": 0.99, + "learning_rate": 5.119358074222668e-06, + "loss": 1.271, + "theoretical_loss": 3.326724409595032, + "tokens_seen": 3280076800 + }, + { + "epoch": 0.99, + "learning_rate": 5.1033099297893685e-06, + "loss": 1.2696, + "theoretical_loss": 3.326719434222828, + "tokens_seen": 3280142336 + }, + { + "epoch": 0.99, + "learning_rate": 5.087261785356068e-06, + "loss": 1.3046, + "theoretical_loss": 3.3267144589778628, + "tokens_seen": 3280207872 + }, + { + "epoch": 0.99, + "learning_rate": 5.071213640922769e-06, + "loss": 1.2702, + "theoretical_loss": 3.326709483860129, + "tokens_seen": 3280273408 + }, + { + "epoch": 0.99, + "learning_rate": 5.055165496489469e-06, + "loss": 1.2344, + "theoretical_loss": 3.3267045088696223, + "tokens_seen": 3280338944 + }, + { + "epoch": 0.99, + "learning_rate": 5.039117352056169e-06, + "loss": 1.302, + "theoretical_loss": 3.3266995340063366, + "tokens_seen": 3280404480 + }, + { + "epoch": 0.99, + "learning_rate": 5.023069207622868e-06, + "loss": 1.2833, + "theoretical_loss": 3.3266945592702655, + "tokens_seen": 3280470016 + }, + { + "epoch": 0.99, + "learning_rate": 5.007021063189569e-06, + "loss": 1.2671, + "theoretical_loss": 3.3266895846614033, + "tokens_seen": 3280535552 + }, + { + "epoch": 0.99, + "learning_rate": 4.990972918756269e-06, + "loss": 1.2016, + "theoretical_loss": 3.3266846101797447, + "tokens_seen": 3280601088 + }, + { + "epoch": 0.99, + "learning_rate": 4.974924774322969e-06, + "loss": 1.2973, + "theoretical_loss": 3.3266796358252835, + "tokens_seen": 3280666624 + }, + { + "epoch": 0.99, + "learning_rate": 4.958876629889669e-06, + "loss": 1.3062, + "theoretical_loss": 3.3266746615980143, + "tokens_seen": 3280732160 + }, + { + "epoch": 0.99, + "learning_rate": 4.94282848545637e-06, + "loss": 1.245, + "theoretical_loss": 3.3266696874979313, + "tokens_seen": 3280797696 + }, + { + "epoch": 0.99, + "learning_rate": 4.9267803410230695e-06, + "loss": 1.243, + "theoretical_loss": 3.326664713525028, + "tokens_seen": 3280863232 + }, + { + "epoch": 0.99, + "learning_rate": 4.910732196589769e-06, + "loss": 1.1718, + "theoretical_loss": 3.326659739679299, + "tokens_seen": 3280928768 + }, + { + "epoch": 0.99, + "learning_rate": 4.894684052156469e-06, + "loss": 1.0824, + "theoretical_loss": 3.326654765960739, + "tokens_seen": 3280994304 + }, + { + "epoch": 0.99, + "learning_rate": 4.87863590772317e-06, + "loss": 1.1227, + "theoretical_loss": 3.3266497923693414, + "tokens_seen": 3281059840 + }, + { + "epoch": 0.99, + "learning_rate": 4.8625877632898695e-06, + "loss": 1.1825, + "theoretical_loss": 3.326644818905101, + "tokens_seen": 3281125376 + }, + { + "epoch": 0.99, + "learning_rate": 4.84653961885657e-06, + "loss": 1.1776, + "theoretical_loss": 3.326639845568012, + "tokens_seen": 3281190912 + }, + { + "epoch": 0.99, + "learning_rate": 4.83049147442327e-06, + "loss": 1.2639, + "theoretical_loss": 3.3266348723580683, + "tokens_seen": 3281256448 + }, + { + "epoch": 0.99, + "learning_rate": 4.81444332998997e-06, + "loss": 1.2652, + "theoretical_loss": 3.326629899275264, + "tokens_seen": 3281321984 + }, + { + "epoch": 0.99, + "learning_rate": 4.79839518555667e-06, + "loss": 1.2815, + "theoretical_loss": 3.326624926319594, + "tokens_seen": 3281387520 + }, + { + "epoch": 0.99, + "learning_rate": 4.78234704112337e-06, + "loss": 1.2109, + "theoretical_loss": 3.326619953491052, + "tokens_seen": 3281453056 + }, + { + "epoch": 0.99, + "learning_rate": 4.766298896690071e-06, + "loss": 1.1776, + "theoretical_loss": 3.326614980789632, + "tokens_seen": 3281518592 + }, + { + "epoch": 0.99, + "learning_rate": 4.7502507522567706e-06, + "loss": 1.2741, + "theoretical_loss": 3.3266100082153285, + "tokens_seen": 3281584128 + }, + { + "epoch": 0.99, + "learning_rate": 4.734202607823471e-06, + "loss": 1.3535, + "theoretical_loss": 3.326605035768136, + "tokens_seen": 3281649664 + }, + { + "epoch": 0.99, + "learning_rate": 4.718154463390171e-06, + "loss": 1.1225, + "theoretical_loss": 3.3266000634480477, + "tokens_seen": 3281715200 + }, + { + "epoch": 0.99, + "learning_rate": 4.702106318956871e-06, + "loss": 1.3081, + "theoretical_loss": 3.326595091255059, + "tokens_seen": 3281780736 + }, + { + "epoch": 0.99, + "learning_rate": 4.6860581745235706e-06, + "loss": 1.2643, + "theoretical_loss": 3.326590119189164, + "tokens_seen": 3281846272 + }, + { + "epoch": 0.99, + "learning_rate": 4.670010030090271e-06, + "loss": 1.2745, + "theoretical_loss": 3.3265851472503556, + "tokens_seen": 3281911808 + }, + { + "epoch": 0.99, + "learning_rate": 4.653961885656971e-06, + "loss": 1.219, + "theoretical_loss": 3.3265801754386297, + "tokens_seen": 3281977344 + }, + { + "epoch": 0.99, + "learning_rate": 4.637913741223672e-06, + "loss": 1.2153, + "theoretical_loss": 3.3265752037539795, + "tokens_seen": 3282042880 + }, + { + "epoch": 0.99, + "learning_rate": 4.621865596790371e-06, + "loss": 1.2392, + "theoretical_loss": 3.3265702321963997, + "tokens_seen": 3282108416 + }, + { + "epoch": 0.99, + "learning_rate": 4.605817452357072e-06, + "loss": 1.1867, + "theoretical_loss": 3.326565260765884, + "tokens_seen": 3282173952 + }, + { + "epoch": 0.99, + "learning_rate": 4.589769307923771e-06, + "loss": 1.187, + "theoretical_loss": 3.326560289462427, + "tokens_seen": 3282239488 + }, + { + "epoch": 0.99, + "learning_rate": 4.573721163490472e-06, + "loss": 1.3118, + "theoretical_loss": 3.3265553182860232, + "tokens_seen": 3282305024 + }, + { + "epoch": 0.99, + "learning_rate": 4.557673019057171e-06, + "loss": 1.141, + "theoretical_loss": 3.326550347236666, + "tokens_seen": 3282370560 + }, + { + "epoch": 0.99, + "learning_rate": 4.541624874623872e-06, + "loss": 1.1874, + "theoretical_loss": 3.32654537631435, + "tokens_seen": 3282436096 + }, + { + "epoch": 0.99, + "learning_rate": 4.525576730190572e-06, + "loss": 1.195, + "theoretical_loss": 3.3265404055190704, + "tokens_seen": 3282501632 + }, + { + "epoch": 0.99, + "learning_rate": 4.5095285857572725e-06, + "loss": 1.2297, + "theoretical_loss": 3.3265354348508196, + "tokens_seen": 3282567168 + }, + { + "epoch": 0.99, + "learning_rate": 4.493480441323972e-06, + "loss": 1.2395, + "theoretical_loss": 3.326530464309593, + "tokens_seen": 3282632704 + }, + { + "epoch": 0.99, + "learning_rate": 4.477432296890672e-06, + "loss": 1.2075, + "theoretical_loss": 3.3265254938953843, + "tokens_seen": 3282698240 + }, + { + "epoch": 0.99, + "learning_rate": 4.461384152457372e-06, + "loss": 1.3061, + "theoretical_loss": 3.3265205236081883, + "tokens_seen": 3282763776 + }, + { + "epoch": 0.99, + "learning_rate": 4.4453360080240725e-06, + "loss": 1.2243, + "theoretical_loss": 3.3265155534479987, + "tokens_seen": 3282829312 + }, + { + "epoch": 0.99, + "learning_rate": 4.429287863590772e-06, + "loss": 1.2779, + "theoretical_loss": 3.3265105834148097, + "tokens_seen": 3282894848 + }, + { + "epoch": 0.99, + "learning_rate": 4.413239719157473e-06, + "loss": 1.1167, + "theoretical_loss": 3.326505613508616, + "tokens_seen": 3282960384 + }, + { + "epoch": 0.99, + "learning_rate": 4.397191574724173e-06, + "loss": 1.1832, + "theoretical_loss": 3.3265006437294113, + "tokens_seen": 3283025920 + }, + { + "epoch": 0.99, + "learning_rate": 4.381143430290873e-06, + "loss": 1.3176, + "theoretical_loss": 3.3264956740771905, + "tokens_seen": 3283091456 + }, + { + "epoch": 0.99, + "learning_rate": 4.365095285857572e-06, + "loss": 1.2342, + "theoretical_loss": 3.326490704551947, + "tokens_seen": 3283156992 + }, + { + "epoch": 0.99, + "learning_rate": 4.349047141424273e-06, + "loss": 1.1386, + "theoretical_loss": 3.3264857351536756, + "tokens_seen": 3283222528 + }, + { + "epoch": 0.99, + "learning_rate": 4.332998996990973e-06, + "loss": 1.2352, + "theoretical_loss": 3.32648076588237, + "tokens_seen": 3283288064 + }, + { + "epoch": 0.99, + "objective/train/docs_used": 1192893, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.1777962446212769, + "objective/train/theoretical_loss": 3.326475796738025, + "objective/train/tokens_used": 3303813600, + "theoretical_loss": 3.326475796738025, + "tokens_seen": 3283353600 + }, + { + "epoch": 0.99, + "learning_rate": 4.316950852557673e-06, + "loss": 1.2039, + "theoretical_loss": 3.326475796738025, + "tokens_seen": 3283353600 + }, + { + "epoch": 0.99, + "learning_rate": 4.300902708124373e-06, + "loss": 1.2548, + "theoretical_loss": 3.3264708277206347, + "tokens_seen": 3283419136 + }, + { + "epoch": 0.99, + "learning_rate": 4.284854563691074e-06, + "loss": 1.2419, + "theoretical_loss": 3.326465858830193, + "tokens_seen": 3283484672 + }, + { + "epoch": 1.0, + "learning_rate": 4.2688064192577735e-06, + "loss": 1.0857, + "theoretical_loss": 3.326460890066694, + "tokens_seen": 3283550208 + }, + { + "epoch": 1.0, + "learning_rate": 4.252758274824473e-06, + "loss": 1.3778, + "theoretical_loss": 3.3264559214301324, + "tokens_seen": 3283615744 + }, + { + "epoch": 1.0, + "learning_rate": 4.236710130391174e-06, + "loss": 1.2456, + "theoretical_loss": 3.3264509529205024, + "tokens_seen": 3283681280 + }, + { + "epoch": 1.0, + "learning_rate": 4.220661985957874e-06, + "loss": 1.2141, + "theoretical_loss": 3.3264459845377985, + "tokens_seen": 3283746816 + }, + { + "epoch": 1.0, + "learning_rate": 4.204613841524574e-06, + "loss": 1.3141, + "theoretical_loss": 3.326441016282014, + "tokens_seen": 3283812352 + }, + { + "epoch": 1.0, + "learning_rate": 4.188565697091274e-06, + "loss": 1.1872, + "theoretical_loss": 3.3264360481531434, + "tokens_seen": 3283877888 + }, + { + "epoch": 1.0, + "learning_rate": 4.172517552657974e-06, + "loss": 1.3028, + "theoretical_loss": 3.3264310801511816, + "tokens_seen": 3283943424 + }, + { + "epoch": 1.0, + "learning_rate": 4.1564694082246746e-06, + "loss": 1.2624, + "theoretical_loss": 3.3264261122761223, + "tokens_seen": 3284008960 + }, + { + "epoch": 1.0, + "learning_rate": 4.140421263791374e-06, + "loss": 1.2261, + "theoretical_loss": 3.3264211445279592, + "tokens_seen": 3284074496 + }, + { + "epoch": 1.0, + "learning_rate": 4.124373119358074e-06, + "loss": 1.1451, + "theoretical_loss": 3.326416176906688, + "tokens_seen": 3284140032 + }, + { + "epoch": 1.0, + "learning_rate": 4.108324974924775e-06, + "loss": 1.1948, + "theoretical_loss": 3.3264112094123015, + "tokens_seen": 3284205568 + }, + { + "epoch": 1.0, + "learning_rate": 4.0922768304914746e-06, + "loss": 1.2119, + "theoretical_loss": 3.326406242044795, + "tokens_seen": 3284271104 + }, + { + "epoch": 1.0, + "learning_rate": 4.076228686058175e-06, + "loss": 1.235, + "theoretical_loss": 3.326401274804162, + "tokens_seen": 3284336640 + }, + { + "epoch": 1.0, + "learning_rate": 4.060180541624875e-06, + "loss": 1.1581, + "theoretical_loss": 3.3263963076903966, + "tokens_seen": 3284402176 + }, + { + "epoch": 1.0, + "learning_rate": 4.044132397191576e-06, + "loss": 1.3587, + "theoretical_loss": 3.3263913407034935, + "tokens_seen": 3284467712 + }, + { + "epoch": 1.0, + "learning_rate": 4.0280842527582746e-06, + "loss": 1.2422, + "theoretical_loss": 3.326386373843447, + "tokens_seen": 3284533248 + }, + { + "epoch": 1.0, + "learning_rate": 4.012036108324975e-06, + "loss": 1.257, + "theoretical_loss": 3.326381407110251, + "tokens_seen": 3284598784 + }, + { + "epoch": 1.0, + "learning_rate": 3.995987963891675e-06, + "loss": 1.2127, + "theoretical_loss": 3.3263764405039, + "tokens_seen": 3284664320 + }, + { + "epoch": 1.0, + "learning_rate": 3.979939819458376e-06, + "loss": 1.2098, + "theoretical_loss": 3.3263714740243877, + "tokens_seen": 3284729856 + }, + { + "epoch": 1.0, + "learning_rate": 3.963891675025075e-06, + "loss": 1.2988, + "theoretical_loss": 3.3263665076717093, + "tokens_seen": 3284795392 + }, + { + "epoch": 1.0, + "learning_rate": 3.947843530591776e-06, + "loss": 1.244, + "theoretical_loss": 3.326361541445858, + "tokens_seen": 3284860928 + }, + { + "epoch": 1.0, + "learning_rate": 3.931795386158476e-06, + "loss": 1.2267, + "theoretical_loss": 3.3263565753468285, + "tokens_seen": 3284926464 + }, + { + "epoch": 1.0, + "learning_rate": 3.915747241725176e-06, + "loss": 1.2259, + "theoretical_loss": 3.3263516093746155, + "tokens_seen": 3284992000 + }, + { + "epoch": 1.0, + "learning_rate": 3.899699097291875e-06, + "loss": 1.2309, + "theoretical_loss": 3.326346643529212, + "tokens_seen": 3285057536 + }, + { + "epoch": 1.0, + "learning_rate": 3.883650952858576e-06, + "loss": 1.2536, + "theoretical_loss": 3.3263416778106136, + "tokens_seen": 3285123072 + }, + { + "epoch": 1.0, + "learning_rate": 3.867602808425276e-06, + "loss": 1.3259, + "theoretical_loss": 3.3263367122188137, + "tokens_seen": 3285188608 + }, + { + "epoch": 1.0, + "learning_rate": 3.8515546639919765e-06, + "loss": 1.1436, + "theoretical_loss": 3.3263317467538065, + "tokens_seen": 3285254144 + }, + { + "epoch": 1.0, + "learning_rate": 3.835506519558676e-06, + "loss": 1.2593, + "theoretical_loss": 3.3263267814155864, + "tokens_seen": 3285319680 + }, + { + "epoch": 1.0, + "learning_rate": 3.819458375125376e-06, + "loss": 1.2656, + "theoretical_loss": 3.3263218162041484, + "tokens_seen": 3285385216 + }, + { + "epoch": 1.0, + "learning_rate": 3.8034102306920762e-06, + "loss": 1.3294, + "theoretical_loss": 3.3263168511194854, + "tokens_seen": 3285450752 + }, + { + "epoch": 1.0, + "learning_rate": 3.7873620862587765e-06, + "loss": 1.1859, + "theoretical_loss": 3.3263118861615926, + "tokens_seen": 3285516288 + }, + { + "epoch": 1.0, + "learning_rate": 3.7713139418254767e-06, + "loss": 1.1448, + "theoretical_loss": 3.3263069213304637, + "tokens_seen": 3285581824 + }, + { + "epoch": 1.0, + "learning_rate": 3.755265797392177e-06, + "loss": 1.2222, + "theoretical_loss": 3.326301956626093, + "tokens_seen": 3285647360 + }, + { + "epoch": 1.0, + "learning_rate": 3.739217652958877e-06, + "loss": 1.2595, + "theoretical_loss": 3.326296992048475, + "tokens_seen": 3285712896 + }, + { + "epoch": 1.0, + "learning_rate": 3.7231695085255773e-06, + "loss": 1.2358, + "theoretical_loss": 3.326292027597604, + "tokens_seen": 3285778432 + }, + { + "epoch": 1.0, + "learning_rate": 3.7071213640922767e-06, + "loss": 1.2357, + "theoretical_loss": 3.3262870632734742, + "tokens_seen": 3285843968 + }, + { + "epoch": 1.0, + "learning_rate": 3.691073219658977e-06, + "loss": 1.1519, + "theoretical_loss": 3.3262820990760793, + "tokens_seen": 3285909504 + }, + { + "epoch": 1.0, + "learning_rate": 3.675025075225677e-06, + "loss": 1.2545, + "theoretical_loss": 3.326277135005414, + "tokens_seen": 3285975040 + }, + { + "epoch": 1.0, + "learning_rate": 3.6589769307923773e-06, + "loss": 1.1948, + "theoretical_loss": 3.3262721710614724, + "tokens_seen": 3286040576 + }, + { + "epoch": 1.0, + "learning_rate": 3.6429287863590775e-06, + "loss": 1.2707, + "theoretical_loss": 3.326267207244249, + "tokens_seen": 3286106112 + }, + { + "epoch": 1.0, + "learning_rate": 3.6268806419257777e-06, + "loss": 1.2246, + "theoretical_loss": 3.326262243553738, + "tokens_seen": 3286171648 + }, + { + "epoch": 1.0, + "learning_rate": 3.610832497492478e-06, + "loss": 1.1964, + "theoretical_loss": 3.326257279989933, + "tokens_seen": 3286237184 + }, + { + "epoch": 1.0, + "learning_rate": 3.5947843530591773e-06, + "loss": 1.1525, + "theoretical_loss": 3.326252316552829, + "tokens_seen": 3286302720 + }, + { + "epoch": 1.0, + "learning_rate": 3.5787362086258775e-06, + "loss": 1.273, + "theoretical_loss": 3.3262473532424197, + "tokens_seen": 3286368256 + }, + { + "epoch": 1.0, + "learning_rate": 3.5626880641925777e-06, + "loss": 1.2804, + "theoretical_loss": 3.3262423900586997, + "tokens_seen": 3286433792 + }, + { + "epoch": 1.0, + "learning_rate": 3.546639919759278e-06, + "loss": 1.2499, + "theoretical_loss": 3.3262374270016632, + "tokens_seen": 3286499328 + }, + { + "epoch": 1.0, + "learning_rate": 3.530591775325978e-06, + "loss": 1.219, + "theoretical_loss": 3.3262324640713046, + "tokens_seen": 3286564864 + }, + { + "epoch": 1.0, + "objective/train/docs_used": 1194104, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.2089579105377197, + "objective/train/theoretical_loss": 3.326227501267618, + "objective/train/tokens_used": 3307090400, + "theoretical_loss": 3.326227501267618, + "tokens_seen": 3286630400 + }, + { + "epoch": 1.0, + "learning_rate": 3.5145436308926783e-06, + "loss": 1.2672, + "theoretical_loss": 3.326227501267618, + "tokens_seen": 3286630400 + }, + { + "epoch": 1.0, + "learning_rate": 3.4984954864593786e-06, + "loss": 1.2856, + "theoretical_loss": 3.326222538590597, + "tokens_seen": 3286695936 + }, + { + "epoch": 1.0, + "learning_rate": 3.482447342026078e-06, + "loss": 1.1723, + "theoretical_loss": 3.3262175760402366, + "tokens_seen": 3286761472 + }, + { + "epoch": 1.0, + "learning_rate": 3.466399197592778e-06, + "loss": 1.1657, + "theoretical_loss": 3.326212613616531, + "tokens_seen": 3286827008 + }, + { + "epoch": 1.0, + "learning_rate": 3.4503510531594783e-06, + "loss": 1.1772, + "theoretical_loss": 3.326207651319474, + "tokens_seen": 3286892544 + }, + { + "epoch": 1.0, + "learning_rate": 3.4343029087261786e-06, + "loss": 1.1046, + "theoretical_loss": 3.3262026891490604, + "tokens_seen": 3286958080 + }, + { + "epoch": 1.0, + "learning_rate": 3.4182547642928788e-06, + "loss": 1.2781, + "theoretical_loss": 3.326197727105284, + "tokens_seen": 3287023616 + }, + { + "epoch": 1.0, + "learning_rate": 3.402206619859579e-06, + "loss": 1.2735, + "theoretical_loss": 3.3261927651881393, + "tokens_seen": 3287089152 + }, + { + "epoch": 1.0, + "learning_rate": 3.386158475426279e-06, + "loss": 1.2595, + "theoretical_loss": 3.3261878033976204, + "tokens_seen": 3287154688 + }, + { + "epoch": 1.0, + "learning_rate": 3.370110330992979e-06, + "loss": 1.239, + "theoretical_loss": 3.3261828417337216, + "tokens_seen": 3287220224 + }, + { + "epoch": 1.0, + "learning_rate": 3.354062186559679e-06, + "loss": 1.2508, + "theoretical_loss": 3.326177880196437, + "tokens_seen": 3287285760 + }, + { + "epoch": 1.0, + "learning_rate": 3.3380140421263794e-06, + "loss": 1.296, + "theoretical_loss": 3.3261729187857614, + "tokens_seen": 3287351296 + }, + { + "epoch": 1.0, + "learning_rate": 3.321965897693079e-06, + "loss": 1.2672, + "theoretical_loss": 3.3261679575016885, + "tokens_seen": 3287416832 + }, + { + "epoch": 1.0, + "learning_rate": 3.3059177532597794e-06, + "loss": 1.3235, + "theoretical_loss": 3.3261629963442125, + "tokens_seen": 3287482368 + }, + { + "epoch": 1.0, + "learning_rate": 3.2898696088264796e-06, + "loss": 1.2508, + "theoretical_loss": 3.3261580353133278, + "tokens_seen": 3287547904 + }, + { + "epoch": 1.0, + "learning_rate": 3.27382146439318e-06, + "loss": 1.2195, + "theoretical_loss": 3.326153074409029, + "tokens_seen": 3287613440 + }, + { + "epoch": 1.0, + "learning_rate": 3.2577733199598796e-06, + "loss": 1.2637, + "theoretical_loss": 3.3261481136313096, + "tokens_seen": 3287678976 + }, + { + "epoch": 1.0, + "learning_rate": 3.24172517552658e-06, + "loss": 1.2952, + "theoretical_loss": 3.326143152980164, + "tokens_seen": 3287744512 + }, + { + "epoch": 1.0, + "learning_rate": 3.22567703109328e-06, + "loss": 1.2493, + "theoretical_loss": 3.3261381924555877, + "tokens_seen": 3287810048 + }, + { + "epoch": 1.0, + "learning_rate": 3.2096288866599802e-06, + "loss": 1.289, + "theoretical_loss": 3.326133232057573, + "tokens_seen": 3287875584 + }, + { + "epoch": 1.0, + "learning_rate": 3.1935807422266804e-06, + "loss": 1.2723, + "theoretical_loss": 3.3261282717861156, + "tokens_seen": 3287941120 + }, + { + "epoch": 1.0, + "learning_rate": 3.1775325977933807e-06, + "loss": 1.3254, + "theoretical_loss": 3.3261233116412092, + "tokens_seen": 3288006656 + }, + { + "epoch": 1.0, + "learning_rate": 3.161484453360081e-06, + "loss": 1.1102, + "theoretical_loss": 3.3261183516228483, + "tokens_seen": 3288072192 + }, + { + "epoch": 1.0, + "learning_rate": 3.1454363089267802e-06, + "loss": 1.2166, + "theoretical_loss": 3.3261133917310266, + "tokens_seen": 3288137728 + }, + { + "epoch": 1.0, + "learning_rate": 3.1293881644934804e-06, + "loss": 1.2823, + "theoretical_loss": 3.326108431965739, + "tokens_seen": 3288203264 + }, + { + "epoch": 1.0, + "learning_rate": 3.1133400200601807e-06, + "loss": 1.1507, + "theoretical_loss": 3.326103472326979, + "tokens_seen": 3288268800 + }, + { + "epoch": 1.0, + "learning_rate": 3.097291875626881e-06, + "loss": 1.2547, + "theoretical_loss": 3.326098512814742, + "tokens_seen": 3288334336 + }, + { + "epoch": 1.0, + "learning_rate": 3.081243731193581e-06, + "loss": 1.1009, + "theoretical_loss": 3.3260935534290206, + "tokens_seen": 3288399872 + }, + { + "epoch": 1.0, + "learning_rate": 3.065195586760281e-06, + "loss": 1.3518, + "theoretical_loss": 3.3260885941698106, + "tokens_seen": 3288465408 + }, + { + "epoch": 1.0, + "learning_rate": 3.049147442326981e-06, + "loss": 1.1823, + "theoretical_loss": 3.3260836350371057, + "tokens_seen": 3288530944 + }, + { + "epoch": 1.0, + "learning_rate": 3.0330992978936813e-06, + "loss": 1.2411, + "theoretical_loss": 3.3260786760309, + "tokens_seen": 3288596480 + }, + { + "epoch": 1.0, + "learning_rate": 3.0170511534603815e-06, + "loss": 1.0912, + "theoretical_loss": 3.326073717151188, + "tokens_seen": 3288662016 + }, + { + "epoch": 1.0, + "learning_rate": 3.0010030090270813e-06, + "loss": 1.2766, + "theoretical_loss": 3.326068758397964, + "tokens_seen": 3288727552 + }, + { + "epoch": 1.0, + "learning_rate": 2.9849548645937815e-06, + "loss": 1.3357, + "theoretical_loss": 3.3260637997712212, + "tokens_seen": 3288793088 + }, + { + "epoch": 1.0, + "learning_rate": 2.9689067201604817e-06, + "loss": 1.2194, + "theoretical_loss": 3.3260588412709553, + "tokens_seen": 3288858624 + }, + { + "epoch": 1.0, + "learning_rate": 2.9528585757271815e-06, + "loss": 1.2122, + "theoretical_loss": 3.32605388289716, + "tokens_seen": 3288924160 + }, + { + "epoch": 1.0, + "learning_rate": 2.9368104312938817e-06, + "loss": 1.2083, + "theoretical_loss": 3.3260489246498297, + "tokens_seen": 3288989696 + }, + { + "epoch": 1.0, + "learning_rate": 2.920762286860582e-06, + "loss": 1.1949, + "theoretical_loss": 3.326043966528958, + "tokens_seen": 3289055232 + }, + { + "epoch": 1.0, + "learning_rate": 2.9047141424272817e-06, + "loss": 1.3417, + "theoretical_loss": 3.32603900853454, + "tokens_seen": 3289120768 + }, + { + "epoch": 1.0, + "learning_rate": 2.888665997993982e-06, + "loss": 1.1793, + "theoretical_loss": 3.3260340506665695, + "tokens_seen": 3289186304 + }, + { + "epoch": 1.0, + "learning_rate": 2.872617853560682e-06, + "loss": 1.1348, + "theoretical_loss": 3.326029092925041, + "tokens_seen": 3289251840 + }, + { + "epoch": 1.0, + "learning_rate": 2.8565697091273823e-06, + "loss": 1.2862, + "theoretical_loss": 3.3260241353099484, + "tokens_seen": 3289317376 + }, + { + "epoch": 1.0, + "learning_rate": 2.840521564694082e-06, + "loss": 1.1006, + "theoretical_loss": 3.326019177821286, + "tokens_seen": 3289382912 + }, + { + "epoch": 1.0, + "learning_rate": 2.8244734202607823e-06, + "loss": 1.274, + "theoretical_loss": 3.3260142204590486, + "tokens_seen": 3289448448 + }, + { + "epoch": 1.0, + "learning_rate": 2.8084252758274825e-06, + "loss": 1.2104, + "theoretical_loss": 3.32600926322323, + "tokens_seen": 3289513984 + }, + { + "epoch": 1.0, + "learning_rate": 2.7923771313941823e-06, + "loss": 1.1799, + "theoretical_loss": 3.3260043061138247, + "tokens_seen": 3289579520 + }, + { + "epoch": 1.0, + "learning_rate": 2.7763289869608825e-06, + "loss": 1.0651, + "theoretical_loss": 3.3259993491308264, + "tokens_seen": 3289645056 + }, + { + "epoch": 1.0, + "learning_rate": 2.7602808425275828e-06, + "loss": 1.2037, + "theoretical_loss": 3.32599439227423, + "tokens_seen": 3289710592 + }, + { + "epoch": 1.0, + "learning_rate": 2.744232698094283e-06, + "loss": 1.2698, + "theoretical_loss": 3.325989435544029, + "tokens_seen": 3289776128 + }, + { + "epoch": 1.0, + "learning_rate": 2.728184553660983e-06, + "loss": 1.2331, + "theoretical_loss": 3.325984478940219, + "tokens_seen": 3289841664 + }, + { + "epoch": 1.0, + "objective/train/docs_used": 1195204, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.2326114177703857, + "objective/train/theoretical_loss": 3.3259795224627933, + "objective/train/tokens_used": 3310367200, + "theoretical_loss": 3.3259795224627933, + "tokens_seen": 3289907200 + }, + { + "epoch": 1.0, + "learning_rate": 2.7121364092276834e-06, + "loss": 1.2194, + "theoretical_loss": 3.3259795224627933, + "tokens_seen": 3289907200 + }, + { + "epoch": 1.0, + "learning_rate": 2.696088264794383e-06, + "loss": 1.2689, + "theoretical_loss": 3.325974566111746, + "tokens_seen": 3289972736 + }, + { + "epoch": 1.0, + "learning_rate": 2.6800401203610834e-06, + "loss": 1.2149, + "theoretical_loss": 3.3259696098870712, + "tokens_seen": 3290038272 + }, + { + "epoch": 1.0, + "learning_rate": 2.6639919759277836e-06, + "loss": 1.3132, + "theoretical_loss": 3.3259646537887644, + "tokens_seen": 3290103808 + }, + { + "epoch": 1.0, + "learning_rate": 2.647943831494484e-06, + "loss": 1.2662, + "theoretical_loss": 3.3259596978168187, + "tokens_seen": 3290169344 + }, + { + "epoch": 1.0, + "learning_rate": 2.631895687061184e-06, + "loss": 1.1818, + "theoretical_loss": 3.3259547419712288, + "tokens_seen": 3290234880 + }, + { + "epoch": 1.0, + "learning_rate": 2.615847542627884e-06, + "loss": 1.2238, + "theoretical_loss": 3.325949786251989, + "tokens_seen": 3290300416 + }, + { + "epoch": 1.0, + "learning_rate": 2.599799398194584e-06, + "loss": 1.2035, + "theoretical_loss": 3.3259448306590933, + "tokens_seen": 3290365952 + }, + { + "epoch": 1.0, + "learning_rate": 2.5837512537612842e-06, + "loss": 1.2534, + "theoretical_loss": 3.325939875192536, + "tokens_seen": 3290431488 + }, + { + "epoch": 1.0, + "learning_rate": 2.567703109327984e-06, + "loss": 1.2505, + "theoretical_loss": 3.325934919852312, + "tokens_seen": 3290497024 + }, + { + "epoch": 1.0, + "learning_rate": 2.5516549648946842e-06, + "loss": 1.1283, + "theoretical_loss": 3.3259299646384144, + "tokens_seen": 3290562560 + }, + { + "epoch": 1.0, + "learning_rate": 2.5356068204613844e-06, + "loss": 1.1386, + "theoretical_loss": 3.325925009550838, + "tokens_seen": 3290628096 + }, + { + "epoch": 1.0, + "learning_rate": 2.5195586760280847e-06, + "loss": 1.2694, + "theoretical_loss": 3.3259200545895777, + "tokens_seen": 3290693632 + }, + { + "epoch": 1.0, + "learning_rate": 2.5035105315947844e-06, + "loss": 1.2238, + "theoretical_loss": 3.3259150997546274, + "tokens_seen": 3290759168 + }, + { + "epoch": 1.0, + "learning_rate": 2.4874623871614847e-06, + "loss": 1.2229, + "theoretical_loss": 3.325910145045981, + "tokens_seen": 3290824704 + }, + { + "epoch": 1.0, + "learning_rate": 2.471414242728185e-06, + "loss": 1.2043, + "theoretical_loss": 3.3259051904636325, + "tokens_seen": 3290890240 + }, + { + "epoch": 1.0, + "learning_rate": 2.4553660982948847e-06, + "loss": 1.1641, + "theoretical_loss": 3.3259002360075773, + "tokens_seen": 3290955776 + }, + { + "epoch": 1.0, + "learning_rate": 2.439317953861585e-06, + "loss": 1.2592, + "theoretical_loss": 3.3258952816778087, + "tokens_seen": 3291021312 + }, + { + "epoch": 1.0, + "learning_rate": 2.423269809428285e-06, + "loss": 1.2307, + "theoretical_loss": 3.3258903274743212, + "tokens_seen": 3291086848 + }, + { + "epoch": 1.0, + "learning_rate": 2.407221664994985e-06, + "loss": 1.1392, + "theoretical_loss": 3.325885373397109, + "tokens_seen": 3291152384 + }, + { + "epoch": 1.0, + "learning_rate": 2.391173520561685e-06, + "loss": 1.2364, + "theoretical_loss": 3.325880419446167, + "tokens_seen": 3291217920 + }, + { + "epoch": 1.0, + "learning_rate": 2.3751253761283853e-06, + "loss": 1.2734, + "theoretical_loss": 3.325875465621489, + "tokens_seen": 3291283456 + }, + { + "epoch": 1.0, + "learning_rate": 2.3590772316950855e-06, + "loss": 1.2957, + "theoretical_loss": 3.3258705119230685, + "tokens_seen": 3291348992 + }, + { + "epoch": 1.0, + "learning_rate": 2.3430290872617853e-06, + "loss": 1.1264, + "theoretical_loss": 3.325865558350901, + "tokens_seen": 3291414528 + }, + { + "epoch": 1.0, + "learning_rate": 2.3269809428284855e-06, + "loss": 1.2891, + "theoretical_loss": 3.32586060490498, + "tokens_seen": 3291480064 + }, + { + "epoch": 1.0, + "learning_rate": 2.3109327983951857e-06, + "loss": 1.1236, + "theoretical_loss": 3.3258556515853, + "tokens_seen": 3291545600 + }, + { + "epoch": 1.0, + "learning_rate": 2.2948846539618855e-06, + "loss": 1.25, + "theoretical_loss": 3.325850698391856, + "tokens_seen": 3291611136 + }, + { + "epoch": 1.0, + "learning_rate": 2.2788365095285857e-06, + "loss": 1.3026, + "theoretical_loss": 3.325845745324641, + "tokens_seen": 3291676672 + }, + { + "epoch": 1.0, + "learning_rate": 2.262788365095286e-06, + "loss": 1.2211, + "theoretical_loss": 3.3258407923836497, + "tokens_seen": 3291742208 + }, + { + "epoch": 1.0, + "learning_rate": 2.246740220661986e-06, + "loss": 1.2273, + "theoretical_loss": 3.325835839568877, + "tokens_seen": 3291807744 + }, + { + "epoch": 1.0, + "learning_rate": 2.230692076228686e-06, + "loss": 1.2508, + "theoretical_loss": 3.3258308868803166, + "tokens_seen": 3291873280 + }, + { + "epoch": 1.0, + "learning_rate": 2.214643931795386e-06, + "loss": 1.2742, + "theoretical_loss": 3.3258259343179626, + "tokens_seen": 3291938816 + }, + { + "epoch": 1.0, + "learning_rate": 2.1985957873620863e-06, + "loss": 1.2293, + "theoretical_loss": 3.3258209818818094, + "tokens_seen": 3292004352 + }, + { + "epoch": 1.0, + "learning_rate": 2.182547642928786e-06, + "loss": 1.1976, + "theoretical_loss": 3.3258160295718517, + "tokens_seen": 3292069888 + }, + { + "epoch": 1.0, + "learning_rate": 2.1664994984954863e-06, + "loss": 1.2143, + "theoretical_loss": 3.3258110773880833, + "tokens_seen": 3292135424 + }, + { + "epoch": 1.0, + "learning_rate": 2.1504513540621865e-06, + "loss": 1.1694, + "theoretical_loss": 3.325806125330499, + "tokens_seen": 3292200960 + }, + { + "epoch": 1.0, + "learning_rate": 2.1344032096288868e-06, + "loss": 1.1488, + "theoretical_loss": 3.3258011733990926, + "tokens_seen": 3292266496 + }, + { + "epoch": 1.0, + "learning_rate": 2.118355065195587e-06, + "loss": 1.2816, + "theoretical_loss": 3.3257962215938583, + "tokens_seen": 3292332032 + }, + { + "epoch": 1.0, + "learning_rate": 2.102306920762287e-06, + "loss": 1.1937, + "theoretical_loss": 3.3257912699147907, + "tokens_seen": 3292397568 + }, + { + "epoch": 1.0, + "learning_rate": 2.086258776328987e-06, + "loss": 1.2201, + "theoretical_loss": 3.325786318361884, + "tokens_seen": 3292463104 + }, + { + "epoch": 1.0, + "learning_rate": 2.070210631895687e-06, + "loss": 1.2017, + "theoretical_loss": 3.325781366935132, + "tokens_seen": 3292528640 + }, + { + "epoch": 1.0, + "learning_rate": 2.0541624874623874e-06, + "loss": 1.215, + "theoretical_loss": 3.32577641563453, + "tokens_seen": 3292594176 + }, + { + "epoch": 1.0, + "learning_rate": 2.0381143430290876e-06, + "loss": 1.1693, + "theoretical_loss": 3.3257714644600713, + "tokens_seen": 3292659712 + }, + { + "epoch": 1.0, + "learning_rate": 2.022066198595788e-06, + "loss": 1.3294, + "theoretical_loss": 3.3257665134117507, + "tokens_seen": 3292725248 + }, + { + "epoch": 1.0, + "learning_rate": 2.0060180541624876e-06, + "loss": 1.1773, + "theoretical_loss": 3.3257615624895625, + "tokens_seen": 3292790784 + }, + { + "epoch": 1.0, + "learning_rate": 1.989969909729188e-06, + "loss": 1.2179, + "theoretical_loss": 3.3257566116935005, + "tokens_seen": 3292856320 + }, + { + "epoch": 1.0, + "learning_rate": 1.973921765295888e-06, + "loss": 1.2665, + "theoretical_loss": 3.3257516610235593, + "tokens_seen": 3292921856 + }, + { + "epoch": 1.0, + "learning_rate": 1.957873620862588e-06, + "loss": 1.3174, + "theoretical_loss": 3.325746710479733, + "tokens_seen": 3292987392 + }, + { + "epoch": 1.0, + "learning_rate": 1.941825476429288e-06, + "loss": 1.1764, + "theoretical_loss": 3.3257417600620163, + "tokens_seen": 3293052928 + }, + { + "epoch": 1.0, + "learning_rate": 1.9257773319959882e-06, + "loss": 1.2377, + "theoretical_loss": 3.325736809770403, + "tokens_seen": 3293118464 + }, + { + "epoch": 1.0, + "objective/train/docs_used": 1196393, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.1790024042129517, + "objective/train/theoretical_loss": 3.325731859604888, + "objective/train/tokens_used": 3313644000, + "theoretical_loss": 3.325731859604888, + "tokens_seen": 3293184000 + }, + { + "epoch": 1.0, + "learning_rate": 1.909729187562688e-06, + "loss": 1.1784, + "theoretical_loss": 3.325731859604888, + "tokens_seen": 3293184000 + }, + { + "epoch": 1.0, + "learning_rate": 1.8936810431293882e-06, + "loss": 1.2259, + "theoretical_loss": 3.3257269095654647, + "tokens_seen": 3293249536 + }, + { + "epoch": 1.0, + "learning_rate": 1.8776328986960884e-06, + "loss": 1.3083, + "theoretical_loss": 3.325721959652128, + "tokens_seen": 3293315072 + }, + { + "epoch": 1.0, + "learning_rate": 1.8615847542627886e-06, + "loss": 1.153, + "theoretical_loss": 3.325717009864872, + "tokens_seen": 3293380608 + }, + { + "epoch": 1.0, + "learning_rate": 1.8455366098294884e-06, + "loss": 1.2404, + "theoretical_loss": 3.325712060203691, + "tokens_seen": 3293446144 + }, + { + "epoch": 1.0, + "learning_rate": 1.8294884653961886e-06, + "loss": 1.1712, + "theoretical_loss": 3.325707110668579, + "tokens_seen": 3293511680 + }, + { + "epoch": 1.0, + "learning_rate": 1.8134403209628889e-06, + "loss": 1.2986, + "theoretical_loss": 3.325702161259531, + "tokens_seen": 3293577216 + }, + { + "epoch": 1.0, + "learning_rate": 1.7973921765295886e-06, + "loss": 1.247, + "theoretical_loss": 3.325697211976541, + "tokens_seen": 3293642752 + }, + { + "epoch": 1.0, + "learning_rate": 1.7813440320962889e-06, + "loss": 1.2762, + "theoretical_loss": 3.3256922628196026, + "tokens_seen": 3293708288 + }, + { + "epoch": 1.0, + "learning_rate": 1.765295887662989e-06, + "loss": 1.2061, + "theoretical_loss": 3.325687313788711, + "tokens_seen": 3293773824 + }, + { + "epoch": 1.0, + "learning_rate": 1.7492477432296893e-06, + "loss": 1.155, + "theoretical_loss": 3.32568236488386, + "tokens_seen": 3293839360 + }, + { + "epoch": 1.0, + "learning_rate": 1.733199598796389e-06, + "loss": 1.2772, + "theoretical_loss": 3.325677416105044, + "tokens_seen": 3293904896 + }, + { + "epoch": 1.0, + "learning_rate": 1.7171514543630893e-06, + "loss": 1.1915, + "theoretical_loss": 3.325672467452257, + "tokens_seen": 3293970432 + }, + { + "epoch": 1.0, + "learning_rate": 1.7011033099297895e-06, + "loss": 1.2178, + "theoretical_loss": 3.325667518925494, + "tokens_seen": 3294035968 + }, + { + "epoch": 1.0, + "learning_rate": 1.6850551654964895e-06, + "loss": 1.3144, + "theoretical_loss": 3.3256625705247482, + "tokens_seen": 3294101504 + }, + { + "epoch": 1.0, + "learning_rate": 1.6690070210631897e-06, + "loss": 1.2402, + "theoretical_loss": 3.3256576222500147, + "tokens_seen": 3294167040 + }, + { + "epoch": 1.0, + "learning_rate": 1.6529588766298897e-06, + "loss": 1.3451, + "theoretical_loss": 3.3256526741012875, + "tokens_seen": 3294232576 + }, + { + "epoch": 1.0, + "learning_rate": 1.63691073219659e-06, + "loss": 1.2386, + "theoretical_loss": 3.3256477260785613, + "tokens_seen": 3294298112 + }, + { + "epoch": 1.0, + "learning_rate": 1.62086258776329e-06, + "loss": 1.2458, + "theoretical_loss": 3.3256427781818303, + "tokens_seen": 3294363648 + }, + { + "epoch": 1.0, + "learning_rate": 1.6048144433299901e-06, + "loss": 1.2584, + "theoretical_loss": 3.325637830411088, + "tokens_seen": 3294429184 + }, + { + "epoch": 1.0, + "learning_rate": 1.5887662988966903e-06, + "loss": 1.2575, + "theoretical_loss": 3.325632882766329, + "tokens_seen": 3294494720 + }, + { + "epoch": 1.0, + "learning_rate": 1.5727181544633901e-06, + "loss": 1.1768, + "theoretical_loss": 3.3256279352475486, + "tokens_seen": 3294560256 + }, + { + "epoch": 1.0, + "learning_rate": 1.5566700100300903e-06, + "loss": 1.2292, + "theoretical_loss": 3.32562298785474, + "tokens_seen": 3294625792 + }, + { + "epoch": 1.0, + "learning_rate": 1.5406218655967905e-06, + "loss": 1.2129, + "theoretical_loss": 3.3256180405878975, + "tokens_seen": 3294691328 + }, + { + "epoch": 1.0, + "learning_rate": 1.5245737211634905e-06, + "loss": 1.294, + "theoretical_loss": 3.3256130934470156, + "tokens_seen": 3294756864 + }, + { + "epoch": 1.0, + "learning_rate": 1.5085255767301907e-06, + "loss": 1.3106, + "theoretical_loss": 3.325608146432089, + "tokens_seen": 3294822400 + }, + { + "epoch": 1.0, + "learning_rate": 1.4924774322968907e-06, + "loss": 1.2252, + "theoretical_loss": 3.3256031995431115, + "tokens_seen": 3294887936 + }, + { + "epoch": 1.0, + "learning_rate": 1.4764292878635907e-06, + "loss": 1.1815, + "theoretical_loss": 3.325598252780078, + "tokens_seen": 3294953472 + }, + { + "epoch": 1.0, + "learning_rate": 1.460381143430291e-06, + "loss": 1.2726, + "theoretical_loss": 3.3255933061429817, + "tokens_seen": 3295019008 + }, + { + "epoch": 1.0, + "learning_rate": 1.444332998996991e-06, + "loss": 1.2207, + "theoretical_loss": 3.325588359631818, + "tokens_seen": 3295084544 + }, + { + "epoch": 1.0, + "learning_rate": 1.4282848545636912e-06, + "loss": 1.2519, + "theoretical_loss": 3.32558341324658, + "tokens_seen": 3295150080 + }, + { + "epoch": 1.0, + "learning_rate": 1.4122367101303912e-06, + "loss": 1.1789, + "theoretical_loss": 3.3255784669872632, + "tokens_seen": 3295215616 + }, + { + "epoch": 1.0, + "learning_rate": 1.3961885656970912e-06, + "loss": 1.292, + "theoretical_loss": 3.3255735208538613, + "tokens_seen": 3295281152 + }, + { + "epoch": 1.0, + "learning_rate": 1.3801404212637914e-06, + "loss": 1.2381, + "theoretical_loss": 3.325568574846369, + "tokens_seen": 3295346688 + }, + { + "epoch": 1.0, + "learning_rate": 1.3640922768304916e-06, + "loss": 1.1676, + "theoretical_loss": 3.3255636289647796, + "tokens_seen": 3295412224 + }, + { + "epoch": 1.0, + "learning_rate": 1.3480441323971916e-06, + "loss": 1.3061, + "theoretical_loss": 3.3255586832090884, + "tokens_seen": 3295477760 + }, + { + "epoch": 1.0, + "learning_rate": 1.3319959879638918e-06, + "loss": 1.2982, + "theoretical_loss": 3.3255537375792894, + "tokens_seen": 3295543296 + }, + { + "epoch": 1.0, + "learning_rate": 1.315947843530592e-06, + "loss": 1.2385, + "theoretical_loss": 3.3255487920753772, + "tokens_seen": 3295608832 + }, + { + "epoch": 1.0, + "learning_rate": 1.299899699097292e-06, + "loss": 1.2924, + "theoretical_loss": 3.3255438466973453, + "tokens_seen": 3295674368 + }, + { + "epoch": 1.0, + "learning_rate": 1.283851554663992e-06, + "loss": 1.213, + "theoretical_loss": 3.325538901445188, + "tokens_seen": 3295739904 + }, + { + "epoch": 1.0, + "learning_rate": 1.2678034102306922e-06, + "loss": 1.1299, + "theoretical_loss": 3.3255339563189006, + "tokens_seen": 3295805440 + }, + { + "epoch": 1.0, + "learning_rate": 1.2517552657973922e-06, + "loss": 1.2425, + "theoretical_loss": 3.325529011318477, + "tokens_seen": 3295870976 + }, + { + "epoch": 1.0, + "learning_rate": 1.2357071213640924e-06, + "loss": 1.1965, + "theoretical_loss": 3.325524066443911, + "tokens_seen": 3295936512 + }, + { + "epoch": 1.0, + "learning_rate": 1.2196589769307924e-06, + "loss": 1.1964, + "theoretical_loss": 3.325519121695197, + "tokens_seen": 3296002048 + }, + { + "epoch": 1.0, + "learning_rate": 1.2036108324974924e-06, + "loss": 1.2224, + "theoretical_loss": 3.3255141770723298, + "tokens_seen": 3296067584 + }, + { + "epoch": 1.0, + "learning_rate": 1.1875626880641926e-06, + "loss": 1.233, + "theoretical_loss": 3.3255092325753033, + "tokens_seen": 3296133120 + }, + { + "epoch": 1.0, + "learning_rate": 1.1715145436308926e-06, + "loss": 1.316, + "theoretical_loss": 3.3255042882041117, + "tokens_seen": 3296198656 + }, + { + "epoch": 1.0, + "learning_rate": 1.1554663991975929e-06, + "loss": 1.2457, + "theoretical_loss": 3.32549934395875, + "tokens_seen": 3296264192 + }, + { + "epoch": 1.0, + "learning_rate": 1.1394182547642929e-06, + "loss": 1.2389, + "theoretical_loss": 3.3254943998392115, + "tokens_seen": 3296329728 + }, + { + "epoch": 1.0, + "learning_rate": 1.123370110330993e-06, + "loss": 1.2048, + "theoretical_loss": 3.3254894558454913, + "tokens_seen": 3296395264 + }, + { + "epoch": 1.0, + "objective/train/docs_used": 1197549, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.2782599925994873, + "objective/train/theoretical_loss": 3.3254845119775833, + "objective/train/tokens_used": 3316920800, + "theoretical_loss": 3.3254845119775833, + "tokens_seen": 3296460800 + }, + { + "epoch": 1.0, + "learning_rate": 1.107321965897693e-06, + "loss": 1.2401, + "theoretical_loss": 3.3254845119775833, + "tokens_seen": 3296460800 + }, + { + "epoch": 1.0, + "learning_rate": 1.091273821464393e-06, + "loss": 1.2726, + "theoretical_loss": 3.3254795682354814, + "tokens_seen": 3296526336 + }, + { + "epoch": 1.0, + "learning_rate": 1.0752256770310933e-06, + "loss": 1.2689, + "theoretical_loss": 3.325474624619181, + "tokens_seen": 3296591872 + }, + { + "epoch": 1.0, + "learning_rate": 1.0591775325977935e-06, + "loss": 1.2887, + "theoretical_loss": 3.3254696811286752, + "tokens_seen": 3296657408 + }, + { + "epoch": 1.0, + "learning_rate": 1.0431293881644935e-06, + "loss": 1.3671, + "theoretical_loss": 3.3254647377639595, + "tokens_seen": 3296722944 + }, + { + "epoch": 1.0, + "learning_rate": 1.0270812437311937e-06, + "loss": 1.1485, + "theoretical_loss": 3.325459794525027, + "tokens_seen": 3296788480 + }, + { + "epoch": 1.0, + "learning_rate": 1.011033099297894e-06, + "loss": 1.1765, + "theoretical_loss": 3.3254548514118727, + "tokens_seen": 3296854016 + }, + { + "epoch": 1.0, + "learning_rate": 9.94984954864594e-07, + "loss": 1.3379, + "theoretical_loss": 3.325449908424491, + "tokens_seen": 3296919552 + }, + { + "epoch": 1.0, + "learning_rate": 9.78936810431294e-07, + "loss": 1.2133, + "theoretical_loss": 3.325444965562876, + "tokens_seen": 3296985088 + }, + { + "epoch": 1.0, + "learning_rate": 9.628886659979941e-07, + "loss": 1.3079, + "theoretical_loss": 3.3254400228270216, + "tokens_seen": 3297050624 + }, + { + "epoch": 1.0, + "learning_rate": 9.468405215646941e-07, + "loss": 1.3119, + "theoretical_loss": 3.3254350802169226, + "tokens_seen": 3297116160 + }, + { + "epoch": 1.0, + "learning_rate": 9.307923771313943e-07, + "loss": 1.4301, + "theoretical_loss": 3.3254301377325732, + "tokens_seen": 3297181696 + }, + { + "epoch": 1.0, + "learning_rate": 9.147442326980943e-07, + "loss": 1.2818, + "theoretical_loss": 3.3254251953739677, + "tokens_seen": 3297247232 + }, + { + "epoch": 1.0, + "learning_rate": 8.986960882647943e-07, + "loss": 1.2937, + "theoretical_loss": 3.3254202531411003, + "tokens_seen": 3297312768 + }, + { + "epoch": 1.0, + "learning_rate": 8.826479438314945e-07, + "loss": 1.2245, + "theoretical_loss": 3.3254153110339653, + "tokens_seen": 3297378304 + }, + { + "epoch": 1.0, + "learning_rate": 8.665997993981945e-07, + "loss": 1.0888, + "theoretical_loss": 3.325410369052557, + "tokens_seen": 3297443840 + }, + { + "epoch": 1.0, + "learning_rate": 8.505516549648947e-07, + "loss": 1.23, + "theoretical_loss": 3.32540542719687, + "tokens_seen": 3297509376 + }, + { + "epoch": 1.0, + "learning_rate": 8.345035105315948e-07, + "loss": 1.191, + "theoretical_loss": 3.3254004854668984, + "tokens_seen": 3297574912 + }, + { + "epoch": 1.0, + "learning_rate": 8.18455366098295e-07, + "loss": 1.348, + "theoretical_loss": 3.325395543862636, + "tokens_seen": 3297640448 + }, + { + "epoch": 1.0, + "learning_rate": 8.024072216649951e-07, + "loss": 1.2377, + "theoretical_loss": 3.3253906023840782, + "tokens_seen": 3297705984 + }, + { + "epoch": 1.0, + "learning_rate": 7.863590772316951e-07, + "loss": 1.2487, + "theoretical_loss": 3.3253856610312185, + "tokens_seen": 3297771520 + }, + { + "epoch": 1.0, + "learning_rate": 7.703109327983953e-07, + "loss": 1.2737, + "theoretical_loss": 3.3253807198040515, + "tokens_seen": 3297837056 + }, + { + "epoch": 1.0, + "learning_rate": 7.542627883650954e-07, + "loss": 1.2192, + "theoretical_loss": 3.325375778702571, + "tokens_seen": 3297902592 + }, + { + "epoch": 1.0, + "learning_rate": 7.382146439317954e-07, + "loss": 1.2779, + "theoretical_loss": 3.325370837726772, + "tokens_seen": 3297968128 + }, + { + "epoch": 1.0, + "learning_rate": 7.221664994984955e-07, + "loss": 1.2166, + "theoretical_loss": 3.325365896876648, + "tokens_seen": 3298033664 + }, + { + "epoch": 1.0, + "learning_rate": 7.061183550651956e-07, + "loss": 1.3456, + "theoretical_loss": 3.3253609561521946, + "tokens_seen": 3298099200 + }, + { + "epoch": 1.0, + "learning_rate": 6.900702106318957e-07, + "loss": 1.3369, + "theoretical_loss": 3.325356015553405, + "tokens_seen": 3298164736 + }, + { + "epoch": 1.0, + "learning_rate": 6.740220661985958e-07, + "loss": 1.2551, + "theoretical_loss": 3.325351075080274, + "tokens_seen": 3298230272 + }, + { + "epoch": 1.0, + "learning_rate": 6.57973921765296e-07, + "loss": 1.3545, + "theoretical_loss": 3.3253461347327953, + "tokens_seen": 3298295808 + }, + { + "epoch": 1.0, + "learning_rate": 6.41925777331996e-07, + "loss": 1.2866, + "theoretical_loss": 3.325341194510964, + "tokens_seen": 3298361344 + }, + { + "epoch": 1.0, + "learning_rate": 6.258776328986961e-07, + "loss": 1.2303, + "theoretical_loss": 3.3253362544147738, + "tokens_seen": 3298426880 + }, + { + "epoch": 1.0, + "learning_rate": 6.098294884653962e-07, + "loss": 1.2972, + "theoretical_loss": 3.3253313144442194, + "tokens_seen": 3298492416 + }, + { + "epoch": 1.0, + "learning_rate": 5.937813440320963e-07, + "loss": 1.2552, + "theoretical_loss": 3.325326374599295, + "tokens_seen": 3298557952 + }, + { + "epoch": 1.0, + "learning_rate": 5.777331995987964e-07, + "loss": 1.2696, + "theoretical_loss": 3.325321434879995, + "tokens_seen": 3298623488 + }, + { + "epoch": 1.0, + "learning_rate": 5.616850551654965e-07, + "loss": 1.1701, + "theoretical_loss": 3.3253164952863132, + "tokens_seen": 3298689024 + }, + { + "epoch": 1.0, + "learning_rate": 5.456369107321965e-07, + "loss": 1.2002, + "theoretical_loss": 3.325311555818245, + "tokens_seen": 3298754560 + }, + { + "epoch": 1.0, + "learning_rate": 5.295887662988967e-07, + "loss": 1.1596, + "theoretical_loss": 3.325306616475783, + "tokens_seen": 3298820096 + }, + { + "epoch": 1.0, + "learning_rate": 5.135406218655968e-07, + "loss": 1.2648, + "theoretical_loss": 3.3253016772589232, + "tokens_seen": 3298885632 + }, + { + "epoch": 1.0, + "learning_rate": 4.97492477432297e-07, + "loss": 1.1689, + "theoretical_loss": 3.3252967381676592, + "tokens_seen": 3298951168 + }, + { + "epoch": 1.0, + "learning_rate": 4.814443329989971e-07, + "loss": 1.1732, + "theoretical_loss": 3.325291799201985, + "tokens_seen": 3299016704 + }, + { + "epoch": 1.0, + "learning_rate": 4.6539618856569716e-07, + "loss": 1.1415, + "theoretical_loss": 3.3252868603618957, + "tokens_seen": 3299082240 + }, + { + "epoch": 1.0, + "learning_rate": 4.4934804413239716e-07, + "loss": 1.261, + "theoretical_loss": 3.3252819216473846, + "tokens_seen": 3299147776 + }, + { + "epoch": 1.0, + "learning_rate": 4.3329989969909727e-07, + "loss": 1.1689, + "theoretical_loss": 3.325276983058447, + "tokens_seen": 3299213312 + }, + { + "epoch": 1.0, + "learning_rate": 4.172517552657974e-07, + "loss": 1.2168, + "theoretical_loss": 3.3252720445950765, + "tokens_seen": 3299278848 + }, + { + "epoch": 1.0, + "learning_rate": 4.0120361083249753e-07, + "loss": 1.1963, + "theoretical_loss": 3.325267106257268, + "tokens_seen": 3299344384 + }, + { + "epoch": 1.0, + "learning_rate": 3.8515546639919763e-07, + "loss": 1.2091, + "theoretical_loss": 3.3252621680450156, + "tokens_seen": 3299409920 + }, + { + "epoch": 1.0, + "learning_rate": 3.691073219658977e-07, + "loss": 1.2383, + "theoretical_loss": 3.3252572299583134, + "tokens_seen": 3299475456 + }, + { + "epoch": 1.0, + "learning_rate": 3.530591775325978e-07, + "loss": 1.2146, + "theoretical_loss": 3.3252522919971557, + "tokens_seen": 3299540992 + }, + { + "epoch": 1.0, + "learning_rate": 3.370110330992979e-07, + "loss": 1.1685, + "theoretical_loss": 3.325247354161537, + "tokens_seen": 3299606528 + }, + { + "epoch": 1.0, + "learning_rate": 3.20962888665998e-07, + "loss": 1.1955, + "theoretical_loss": 3.3252424164514514, + "tokens_seen": 3299672064 + }, + { + "epoch": 1.0, + "objective/train/docs_used": 1198719, + "objective/train/instantaneous_batch_size": 32, + "objective/train/instantaneous_microbatch_size": 32768, + "objective/train/original_loss": 1.1255238056182861, + "objective/train/theoretical_loss": 3.3252374788668937, + "objective/train/tokens_used": 3320197600, + "theoretical_loss": 3.3252374788668937, + "tokens_seen": 3299737600 + }, + { + "epoch": 1.0, + "learning_rate": 3.049147442326981e-07, + "loss": 1.1318, + "theoretical_loss": 3.3252374788668937, + "tokens_seen": 3299737600 + }, + { + "epoch": 1.0, + "learning_rate": 2.888665997993982e-07, + "loss": 1.2515, + "theoretical_loss": 3.325232541407858, + "tokens_seen": 3299803136 + }, + { + "epoch": 1.0, + "learning_rate": 2.7281845536609827e-07, + "loss": 1.1338, + "theoretical_loss": 3.325227604074338, + "tokens_seen": 3299868672 + }, + { + "epoch": 1.0, + "learning_rate": 2.567703109327984e-07, + "loss": 1.1679, + "theoretical_loss": 3.325222666866329, + "tokens_seen": 3299934208 + }, + { + "epoch": 1.0, + "learning_rate": 2.4072216649949853e-07, + "loss": 1.1929, + "theoretical_loss": 3.325217729783825, + "tokens_seen": 3299999744 + } + ], + "max_steps": 50354, + "num_train_epochs": 9223372036854775807, + "total_flos": 1.684108669353984e+18, + "trial_name": null, + "trial_params": null +}