| { | |
| "best_metric": 0.3881785571575165, | |
| "best_model_checkpoint": "mikhail-panzo/zlm_b64_le5_s8000/checkpoint-4500", | |
| "epoch": 3.769633507853403, | |
| "eval_steps": 500, | |
| "global_step": 4500, | |
| "is_hyper_param_search": false, | |
| "is_local_process_zero": true, | |
| "is_world_process_zero": true, | |
| "log_history": [ | |
| { | |
| "epoch": 0.041884816753926704, | |
| "grad_norm": 11.218673706054688, | |
| "learning_rate": 2.4500000000000004e-07, | |
| "loss": 1.1196, | |
| "step": 50 | |
| }, | |
| { | |
| "epoch": 0.08376963350785341, | |
| "grad_norm": 5.423076629638672, | |
| "learning_rate": 4.95e-07, | |
| "loss": 1.057, | |
| "step": 100 | |
| }, | |
| { | |
| "epoch": 0.1256544502617801, | |
| "grad_norm": 12.713382720947266, | |
| "learning_rate": 7.4e-07, | |
| "loss": 1.0037, | |
| "step": 150 | |
| }, | |
| { | |
| "epoch": 0.16753926701570682, | |
| "grad_norm": 4.434875965118408, | |
| "learning_rate": 9.9e-07, | |
| "loss": 0.9381, | |
| "step": 200 | |
| }, | |
| { | |
| "epoch": 0.2094240837696335, | |
| "grad_norm": 3.668553590774536, | |
| "learning_rate": 1.2400000000000002e-06, | |
| "loss": 0.8591, | |
| "step": 250 | |
| }, | |
| { | |
| "epoch": 0.2513089005235602, | |
| "grad_norm": 2.2456483840942383, | |
| "learning_rate": 1.4900000000000001e-06, | |
| "loss": 0.8262, | |
| "step": 300 | |
| }, | |
| { | |
| "epoch": 0.2931937172774869, | |
| "grad_norm": 2.70170259475708, | |
| "learning_rate": 1.74e-06, | |
| "loss": 0.7634, | |
| "step": 350 | |
| }, | |
| { | |
| "epoch": 0.33507853403141363, | |
| "grad_norm": 2.9946064949035645, | |
| "learning_rate": 1.9900000000000004e-06, | |
| "loss": 0.7336, | |
| "step": 400 | |
| }, | |
| { | |
| "epoch": 0.3769633507853403, | |
| "grad_norm": 2.2950398921966553, | |
| "learning_rate": 2.24e-06, | |
| "loss": 0.7121, | |
| "step": 450 | |
| }, | |
| { | |
| "epoch": 0.418848167539267, | |
| "grad_norm": 8.007049560546875, | |
| "learning_rate": 2.4900000000000003e-06, | |
| "loss": 0.6879, | |
| "step": 500 | |
| }, | |
| { | |
| "epoch": 0.418848167539267, | |
| "eval_loss": 0.5903987884521484, | |
| "eval_runtime": 258.509, | |
| "eval_samples_per_second": 32.838, | |
| "eval_steps_per_second": 4.108, | |
| "step": 500 | |
| }, | |
| { | |
| "epoch": 0.4607329842931937, | |
| "grad_norm": 3.0767672061920166, | |
| "learning_rate": 2.7400000000000004e-06, | |
| "loss": 0.6734, | |
| "step": 550 | |
| }, | |
| { | |
| "epoch": 0.5026178010471204, | |
| "grad_norm": 8.150467872619629, | |
| "learning_rate": 2.99e-06, | |
| "loss": 0.6539, | |
| "step": 600 | |
| }, | |
| { | |
| "epoch": 0.5445026178010471, | |
| "grad_norm": 2.2473220825195312, | |
| "learning_rate": 3.2400000000000003e-06, | |
| "loss": 0.657, | |
| "step": 650 | |
| }, | |
| { | |
| "epoch": 0.5863874345549738, | |
| "grad_norm": 2.448143482208252, | |
| "learning_rate": 3.49e-06, | |
| "loss": 0.6258, | |
| "step": 700 | |
| }, | |
| { | |
| "epoch": 0.6282722513089005, | |
| "grad_norm": 4.810636520385742, | |
| "learning_rate": 3.74e-06, | |
| "loss": 0.6155, | |
| "step": 750 | |
| }, | |
| { | |
| "epoch": 0.6701570680628273, | |
| "grad_norm": 2.5927581787109375, | |
| "learning_rate": 3.990000000000001e-06, | |
| "loss": 0.6073, | |
| "step": 800 | |
| }, | |
| { | |
| "epoch": 0.7120418848167539, | |
| "grad_norm": 1.913765549659729, | |
| "learning_rate": 4.24e-06, | |
| "loss": 0.6118, | |
| "step": 850 | |
| }, | |
| { | |
| "epoch": 0.7539267015706806, | |
| "grad_norm": 2.210376262664795, | |
| "learning_rate": 4.49e-06, | |
| "loss": 0.5807, | |
| "step": 900 | |
| }, | |
| { | |
| "epoch": 0.7958115183246073, | |
| "grad_norm": 3.323096752166748, | |
| "learning_rate": 4.74e-06, | |
| "loss": 0.5915, | |
| "step": 950 | |
| }, | |
| { | |
| "epoch": 0.837696335078534, | |
| "grad_norm": 2.1455421447753906, | |
| "learning_rate": 4.9900000000000005e-06, | |
| "loss": 0.5606, | |
| "step": 1000 | |
| }, | |
| { | |
| "epoch": 0.837696335078534, | |
| "eval_loss": 0.49923568964004517, | |
| "eval_runtime": 260.1008, | |
| "eval_samples_per_second": 32.637, | |
| "eval_steps_per_second": 4.083, | |
| "step": 1000 | |
| }, | |
| { | |
| "epoch": 0.8795811518324608, | |
| "grad_norm": 2.7230141162872314, | |
| "learning_rate": 5.240000000000001e-06, | |
| "loss": 0.5731, | |
| "step": 1050 | |
| }, | |
| { | |
| "epoch": 0.9214659685863874, | |
| "grad_norm": 2.0273685455322266, | |
| "learning_rate": 5.490000000000001e-06, | |
| "loss": 0.5361, | |
| "step": 1100 | |
| }, | |
| { | |
| "epoch": 0.9633507853403142, | |
| "grad_norm": 3.9353837966918945, | |
| "learning_rate": 5.74e-06, | |
| "loss": 0.5476, | |
| "step": 1150 | |
| }, | |
| { | |
| "epoch": 1.0052356020942408, | |
| "grad_norm": 2.7838985919952393, | |
| "learning_rate": 5.99e-06, | |
| "loss": 0.5356, | |
| "step": 1200 | |
| }, | |
| { | |
| "epoch": 1.0471204188481675, | |
| "grad_norm": 2.8621716499328613, | |
| "learning_rate": 6.24e-06, | |
| "loss": 0.5413, | |
| "step": 1250 | |
| }, | |
| { | |
| "epoch": 1.0890052356020943, | |
| "grad_norm": 2.490875005722046, | |
| "learning_rate": 6.4900000000000005e-06, | |
| "loss": 0.5375, | |
| "step": 1300 | |
| }, | |
| { | |
| "epoch": 1.130890052356021, | |
| "grad_norm": 3.052311897277832, | |
| "learning_rate": 6.740000000000001e-06, | |
| "loss": 0.5415, | |
| "step": 1350 | |
| }, | |
| { | |
| "epoch": 1.1727748691099475, | |
| "grad_norm": 2.5695457458496094, | |
| "learning_rate": 6.99e-06, | |
| "loss": 0.5299, | |
| "step": 1400 | |
| }, | |
| { | |
| "epoch": 1.2146596858638743, | |
| "grad_norm": 2.5093400478363037, | |
| "learning_rate": 7.24e-06, | |
| "loss": 0.5175, | |
| "step": 1450 | |
| }, | |
| { | |
| "epoch": 1.256544502617801, | |
| "grad_norm": 3.005453586578369, | |
| "learning_rate": 7.49e-06, | |
| "loss": 0.5203, | |
| "step": 1500 | |
| }, | |
| { | |
| "epoch": 1.256544502617801, | |
| "eval_loss": 0.45439228415489197, | |
| "eval_runtime": 266.4154, | |
| "eval_samples_per_second": 31.864, | |
| "eval_steps_per_second": 3.986, | |
| "step": 1500 | |
| }, | |
| { | |
| "epoch": 1.2984293193717278, | |
| "grad_norm": 3.572266101837158, | |
| "learning_rate": 7.74e-06, | |
| "loss": 0.5192, | |
| "step": 1550 | |
| }, | |
| { | |
| "epoch": 1.3403141361256545, | |
| "grad_norm": 2.545226573944092, | |
| "learning_rate": 7.990000000000001e-06, | |
| "loss": 0.5128, | |
| "step": 1600 | |
| }, | |
| { | |
| "epoch": 1.3821989528795813, | |
| "grad_norm": 2.4371402263641357, | |
| "learning_rate": 8.24e-06, | |
| "loss": 0.5118, | |
| "step": 1650 | |
| }, | |
| { | |
| "epoch": 1.4240837696335078, | |
| "grad_norm": 2.7607970237731934, | |
| "learning_rate": 8.49e-06, | |
| "loss": 0.5059, | |
| "step": 1700 | |
| }, | |
| { | |
| "epoch": 1.4659685863874345, | |
| "grad_norm": 2.4732346534729004, | |
| "learning_rate": 8.740000000000001e-06, | |
| "loss": 0.5129, | |
| "step": 1750 | |
| }, | |
| { | |
| "epoch": 1.5078534031413613, | |
| "grad_norm": 2.1999361515045166, | |
| "learning_rate": 8.99e-06, | |
| "loss": 0.4977, | |
| "step": 1800 | |
| }, | |
| { | |
| "epoch": 1.5497382198952878, | |
| "grad_norm": 2.2437193393707275, | |
| "learning_rate": 9.240000000000001e-06, | |
| "loss": 0.4964, | |
| "step": 1850 | |
| }, | |
| { | |
| "epoch": 1.5916230366492146, | |
| "grad_norm": 1.9161577224731445, | |
| "learning_rate": 9.49e-06, | |
| "loss": 0.4914, | |
| "step": 1900 | |
| }, | |
| { | |
| "epoch": 1.6335078534031413, | |
| "grad_norm": 3.3452305793762207, | |
| "learning_rate": 9.74e-06, | |
| "loss": 0.4971, | |
| "step": 1950 | |
| }, | |
| { | |
| "epoch": 1.675392670157068, | |
| "grad_norm": 1.818142294883728, | |
| "learning_rate": 9.990000000000001e-06, | |
| "loss": 0.4873, | |
| "step": 2000 | |
| }, | |
| { | |
| "epoch": 1.675392670157068, | |
| "eval_loss": 0.43226656317710876, | |
| "eval_runtime": 264.8191, | |
| "eval_samples_per_second": 32.056, | |
| "eval_steps_per_second": 4.01, | |
| "step": 2000 | |
| }, | |
| { | |
| "epoch": 1.7172774869109948, | |
| "grad_norm": 2.1289865970611572, | |
| "learning_rate": 9.920000000000002e-06, | |
| "loss": 0.4809, | |
| "step": 2050 | |
| }, | |
| { | |
| "epoch": 1.7591623036649215, | |
| "grad_norm": 3.333080768585205, | |
| "learning_rate": 9.836666666666668e-06, | |
| "loss": 0.4853, | |
| "step": 2100 | |
| }, | |
| { | |
| "epoch": 1.8010471204188483, | |
| "grad_norm": 3.106902599334717, | |
| "learning_rate": 9.753333333333335e-06, | |
| "loss": 0.4841, | |
| "step": 2150 | |
| }, | |
| { | |
| "epoch": 1.8429319371727748, | |
| "grad_norm": 3.021559238433838, | |
| "learning_rate": 9.67e-06, | |
| "loss": 0.4795, | |
| "step": 2200 | |
| }, | |
| { | |
| "epoch": 1.8848167539267016, | |
| "grad_norm": 2.128567695617676, | |
| "learning_rate": 9.586666666666667e-06, | |
| "loss": 0.4784, | |
| "step": 2250 | |
| }, | |
| { | |
| "epoch": 1.9267015706806283, | |
| "grad_norm": 2.3653769493103027, | |
| "learning_rate": 9.503333333333334e-06, | |
| "loss": 0.4792, | |
| "step": 2300 | |
| }, | |
| { | |
| "epoch": 1.9685863874345548, | |
| "grad_norm": 2.409151792526245, | |
| "learning_rate": 9.42e-06, | |
| "loss": 0.4727, | |
| "step": 2350 | |
| }, | |
| { | |
| "epoch": 2.0104712041884816, | |
| "grad_norm": 2.9087882041931152, | |
| "learning_rate": 9.336666666666666e-06, | |
| "loss": 0.4645, | |
| "step": 2400 | |
| }, | |
| { | |
| "epoch": 2.0523560209424083, | |
| "grad_norm": 2.5785608291625977, | |
| "learning_rate": 9.253333333333333e-06, | |
| "loss": 0.4696, | |
| "step": 2450 | |
| }, | |
| { | |
| "epoch": 2.094240837696335, | |
| "grad_norm": 2.625706672668457, | |
| "learning_rate": 9.17e-06, | |
| "loss": 0.4688, | |
| "step": 2500 | |
| }, | |
| { | |
| "epoch": 2.094240837696335, | |
| "eval_loss": 0.41558554768562317, | |
| "eval_runtime": 267.455, | |
| "eval_samples_per_second": 31.74, | |
| "eval_steps_per_second": 3.971, | |
| "step": 2500 | |
| }, | |
| { | |
| "epoch": 2.136125654450262, | |
| "grad_norm": 3.3157095909118652, | |
| "learning_rate": 9.086666666666667e-06, | |
| "loss": 0.4684, | |
| "step": 2550 | |
| }, | |
| { | |
| "epoch": 2.1780104712041886, | |
| "grad_norm": 1.7106587886810303, | |
| "learning_rate": 9.003333333333334e-06, | |
| "loss": 0.4721, | |
| "step": 2600 | |
| }, | |
| { | |
| "epoch": 2.2198952879581153, | |
| "grad_norm": 2.8600311279296875, | |
| "learning_rate": 8.920000000000001e-06, | |
| "loss": 0.463, | |
| "step": 2650 | |
| }, | |
| { | |
| "epoch": 2.261780104712042, | |
| "grad_norm": 2.0305252075195312, | |
| "learning_rate": 8.836666666666668e-06, | |
| "loss": 0.4648, | |
| "step": 2700 | |
| }, | |
| { | |
| "epoch": 2.303664921465969, | |
| "grad_norm": 3.124579906463623, | |
| "learning_rate": 8.753333333333333e-06, | |
| "loss": 0.4664, | |
| "step": 2750 | |
| }, | |
| { | |
| "epoch": 2.345549738219895, | |
| "grad_norm": 2.565577268600464, | |
| "learning_rate": 8.67e-06, | |
| "loss": 0.456, | |
| "step": 2800 | |
| }, | |
| { | |
| "epoch": 2.387434554973822, | |
| "grad_norm": 1.6280083656311035, | |
| "learning_rate": 8.588333333333334e-06, | |
| "loss": 0.4615, | |
| "step": 2850 | |
| }, | |
| { | |
| "epoch": 2.4293193717277486, | |
| "grad_norm": 2.195096969604492, | |
| "learning_rate": 8.505e-06, | |
| "loss": 0.4712, | |
| "step": 2900 | |
| }, | |
| { | |
| "epoch": 2.4712041884816753, | |
| "grad_norm": 2.9661295413970947, | |
| "learning_rate": 8.421666666666668e-06, | |
| "loss": 0.4607, | |
| "step": 2950 | |
| }, | |
| { | |
| "epoch": 2.513089005235602, | |
| "grad_norm": 1.7868503332138062, | |
| "learning_rate": 8.338333333333335e-06, | |
| "loss": 0.4613, | |
| "step": 3000 | |
| }, | |
| { | |
| "epoch": 2.513089005235602, | |
| "eval_loss": 0.405597448348999, | |
| "eval_runtime": 267.2992, | |
| "eval_samples_per_second": 31.758, | |
| "eval_steps_per_second": 3.973, | |
| "step": 3000 | |
| }, | |
| { | |
| "epoch": 2.554973821989529, | |
| "grad_norm": 2.4901626110076904, | |
| "learning_rate": 8.255000000000001e-06, | |
| "loss": 0.4522, | |
| "step": 3050 | |
| }, | |
| { | |
| "epoch": 2.5968586387434556, | |
| "grad_norm": 2.0079450607299805, | |
| "learning_rate": 8.171666666666668e-06, | |
| "loss": 0.4503, | |
| "step": 3100 | |
| }, | |
| { | |
| "epoch": 2.6387434554973823, | |
| "grad_norm": 2.885077714920044, | |
| "learning_rate": 8.088333333333334e-06, | |
| "loss": 0.4487, | |
| "step": 3150 | |
| }, | |
| { | |
| "epoch": 2.680628272251309, | |
| "grad_norm": 1.8914357423782349, | |
| "learning_rate": 8.005e-06, | |
| "loss": 0.4544, | |
| "step": 3200 | |
| }, | |
| { | |
| "epoch": 2.7225130890052354, | |
| "grad_norm": 1.7431243658065796, | |
| "learning_rate": 7.921666666666667e-06, | |
| "loss": 0.4491, | |
| "step": 3250 | |
| }, | |
| { | |
| "epoch": 2.7643979057591626, | |
| "grad_norm": 2.154836893081665, | |
| "learning_rate": 7.838333333333334e-06, | |
| "loss": 0.4488, | |
| "step": 3300 | |
| }, | |
| { | |
| "epoch": 2.806282722513089, | |
| "grad_norm": 3.063981056213379, | |
| "learning_rate": 7.755000000000001e-06, | |
| "loss": 0.4501, | |
| "step": 3350 | |
| }, | |
| { | |
| "epoch": 2.8481675392670156, | |
| "grad_norm": 1.9389702081680298, | |
| "learning_rate": 7.671666666666668e-06, | |
| "loss": 0.449, | |
| "step": 3400 | |
| }, | |
| { | |
| "epoch": 2.8900523560209423, | |
| "grad_norm": 1.8112986087799072, | |
| "learning_rate": 7.588333333333334e-06, | |
| "loss": 0.4482, | |
| "step": 3450 | |
| }, | |
| { | |
| "epoch": 2.931937172774869, | |
| "grad_norm": 2.060755729675293, | |
| "learning_rate": 7.505e-06, | |
| "loss": 0.4454, | |
| "step": 3500 | |
| }, | |
| { | |
| "epoch": 2.931937172774869, | |
| "eval_loss": 0.39764249324798584, | |
| "eval_runtime": 270.8158, | |
| "eval_samples_per_second": 31.346, | |
| "eval_steps_per_second": 3.921, | |
| "step": 3500 | |
| }, | |
| { | |
| "epoch": 2.973821989528796, | |
| "grad_norm": 1.6399081945419312, | |
| "learning_rate": 7.421666666666667e-06, | |
| "loss": 0.455, | |
| "step": 3550 | |
| }, | |
| { | |
| "epoch": 3.0157068062827226, | |
| "grad_norm": 1.8340182304382324, | |
| "learning_rate": 7.338333333333334e-06, | |
| "loss": 0.4442, | |
| "step": 3600 | |
| }, | |
| { | |
| "epoch": 3.0575916230366493, | |
| "grad_norm": 2.0830767154693604, | |
| "learning_rate": 7.255000000000001e-06, | |
| "loss": 0.443, | |
| "step": 3650 | |
| }, | |
| { | |
| "epoch": 3.099476439790576, | |
| "grad_norm": 1.8929067850112915, | |
| "learning_rate": 7.171666666666667e-06, | |
| "loss": 0.4424, | |
| "step": 3700 | |
| }, | |
| { | |
| "epoch": 3.141361256544503, | |
| "grad_norm": 1.7542285919189453, | |
| "learning_rate": 7.088333333333334e-06, | |
| "loss": 0.4348, | |
| "step": 3750 | |
| }, | |
| { | |
| "epoch": 3.183246073298429, | |
| "grad_norm": 1.7470309734344482, | |
| "learning_rate": 7.005000000000001e-06, | |
| "loss": 0.4407, | |
| "step": 3800 | |
| }, | |
| { | |
| "epoch": 3.225130890052356, | |
| "grad_norm": 1.6802048683166504, | |
| "learning_rate": 6.921666666666668e-06, | |
| "loss": 0.4372, | |
| "step": 3850 | |
| }, | |
| { | |
| "epoch": 3.2670157068062826, | |
| "grad_norm": 1.4893133640289307, | |
| "learning_rate": 6.838333333333334e-06, | |
| "loss": 0.4348, | |
| "step": 3900 | |
| }, | |
| { | |
| "epoch": 3.3089005235602094, | |
| "grad_norm": 3.0007338523864746, | |
| "learning_rate": 6.7550000000000005e-06, | |
| "loss": 0.4427, | |
| "step": 3950 | |
| }, | |
| { | |
| "epoch": 3.350785340314136, | |
| "grad_norm": 1.806954264640808, | |
| "learning_rate": 6.6716666666666674e-06, | |
| "loss": 0.4299, | |
| "step": 4000 | |
| }, | |
| { | |
| "epoch": 3.350785340314136, | |
| "eval_loss": 0.3921334147453308, | |
| "eval_runtime": 264.7584, | |
| "eval_samples_per_second": 32.063, | |
| "eval_steps_per_second": 4.011, | |
| "step": 4000 | |
| }, | |
| { | |
| "epoch": 3.392670157068063, | |
| "grad_norm": 2.3148508071899414, | |
| "learning_rate": 6.588333333333334e-06, | |
| "loss": 0.4321, | |
| "step": 4050 | |
| }, | |
| { | |
| "epoch": 3.4345549738219896, | |
| "grad_norm": 1.546831488609314, | |
| "learning_rate": 6.505e-06, | |
| "loss": 0.4331, | |
| "step": 4100 | |
| }, | |
| { | |
| "epoch": 3.4764397905759163, | |
| "grad_norm": 2.161104679107666, | |
| "learning_rate": 6.421666666666667e-06, | |
| "loss": 0.4388, | |
| "step": 4150 | |
| }, | |
| { | |
| "epoch": 3.518324607329843, | |
| "grad_norm": 1.6207386255264282, | |
| "learning_rate": 6.338333333333334e-06, | |
| "loss": 0.4352, | |
| "step": 4200 | |
| }, | |
| { | |
| "epoch": 3.5602094240837694, | |
| "grad_norm": 2.1468095779418945, | |
| "learning_rate": 6.255e-06, | |
| "loss": 0.4358, | |
| "step": 4250 | |
| }, | |
| { | |
| "epoch": 3.6020942408376966, | |
| "grad_norm": 2.104029417037964, | |
| "learning_rate": 6.171666666666667e-06, | |
| "loss": 0.4314, | |
| "step": 4300 | |
| }, | |
| { | |
| "epoch": 3.643979057591623, | |
| "grad_norm": 3.058011293411255, | |
| "learning_rate": 6.088333333333334e-06, | |
| "loss": 0.4298, | |
| "step": 4350 | |
| }, | |
| { | |
| "epoch": 3.6858638743455496, | |
| "grad_norm": 1.4571338891983032, | |
| "learning_rate": 6.005000000000001e-06, | |
| "loss": 0.4364, | |
| "step": 4400 | |
| }, | |
| { | |
| "epoch": 3.7277486910994764, | |
| "grad_norm": 1.5542718172073364, | |
| "learning_rate": 5.921666666666667e-06, | |
| "loss": 0.432, | |
| "step": 4450 | |
| }, | |
| { | |
| "epoch": 3.769633507853403, | |
| "grad_norm": 2.5827696323394775, | |
| "learning_rate": 5.838333333333334e-06, | |
| "loss": 0.436, | |
| "step": 4500 | |
| }, | |
| { | |
| "epoch": 3.769633507853403, | |
| "eval_loss": 0.3881785571575165, | |
| "eval_runtime": 272.3698, | |
| "eval_samples_per_second": 31.167, | |
| "eval_steps_per_second": 3.899, | |
| "step": 4500 | |
| } | |
| ], | |
| "logging_steps": 50, | |
| "max_steps": 8000, | |
| "num_input_tokens_seen": 0, | |
| "num_train_epochs": 7, | |
| "save_steps": 500, | |
| "stateful_callbacks": { | |
| "TrainerControl": { | |
| "args": { | |
| "should_epoch_stop": false, | |
| "should_evaluate": false, | |
| "should_log": false, | |
| "should_save": true, | |
| "should_training_stop": false | |
| }, | |
| "attributes": {} | |
| } | |
| }, | |
| "total_flos": 4.028518340583264e+16, | |
| "train_batch_size": 16, | |
| "trial_name": null, | |
| "trial_params": null | |
| } | |