| { | |
| "best_metric": null, | |
| "best_model_checkpoint": null, | |
| "epoch": 0.9984, | |
| "eval_steps": 500, | |
| "global_step": 468, | |
| "is_hyper_param_search": false, | |
| "is_local_process_zero": true, | |
| "is_world_process_zero": true, | |
| "log_history": [ | |
| { | |
| "epoch": 0.0, | |
| "learning_rate": 1.0638297872340426e-07, | |
| "logits/chosen": 0.1359557956457138, | |
| "logits/rejected": 0.030706744641065598, | |
| "logps/chosen": -736.0869140625, | |
| "logps/rejected": -613.6344604492188, | |
| "loss": 2.0331, | |
| "rewards/accuracies": 0.0, | |
| "rewards/chosen": 0.0, | |
| "rewards/margins": 0.0, | |
| "rewards/rejected": 0.0, | |
| "step": 1 | |
| }, | |
| { | |
| "epoch": 0.02, | |
| "learning_rate": 1.0638297872340427e-06, | |
| "logits/chosen": 0.11667777597904205, | |
| "logits/rejected": 0.26604601740837097, | |
| "logps/chosen": -546.5281982421875, | |
| "logps/rejected": -597.5736083984375, | |
| "loss": 2.1592, | |
| "rewards/accuracies": 0.4583333432674408, | |
| "rewards/chosen": 0.0007250224007293582, | |
| "rewards/margins": 0.00040180076030083, | |
| "rewards/rejected": 0.0003232216986361891, | |
| "step": 10 | |
| }, | |
| { | |
| "epoch": 0.04, | |
| "learning_rate": 2.1276595744680853e-06, | |
| "logits/chosen": 0.16373148560523987, | |
| "logits/rejected": 0.2677033543586731, | |
| "logps/chosen": -604.6590576171875, | |
| "logps/rejected": -649.482177734375, | |
| "loss": 2.0972, | |
| "rewards/accuracies": 0.4375, | |
| "rewards/chosen": -0.0005862273974344134, | |
| "rewards/margins": -0.0003054165281355381, | |
| "rewards/rejected": -0.0002808108984027058, | |
| "step": 20 | |
| }, | |
| { | |
| "epoch": 0.06, | |
| "learning_rate": 3.191489361702128e-06, | |
| "logits/chosen": 0.14978544414043427, | |
| "logits/rejected": 0.1915779411792755, | |
| "logps/chosen": -594.8548583984375, | |
| "logps/rejected": -588.2429809570312, | |
| "loss": 2.122, | |
| "rewards/accuracies": 0.512499988079071, | |
| "rewards/chosen": -0.004188057966530323, | |
| "rewards/margins": 0.0009490737575106323, | |
| "rewards/rejected": -0.0051371315494179726, | |
| "step": 30 | |
| }, | |
| { | |
| "epoch": 0.09, | |
| "learning_rate": 4.255319148936171e-06, | |
| "logits/chosen": 0.16862796247005463, | |
| "logits/rejected": 0.23586151003837585, | |
| "logps/chosen": -574.7235107421875, | |
| "logps/rejected": -631.8544921875, | |
| "loss": 2.1863, | |
| "rewards/accuracies": 0.5375000238418579, | |
| "rewards/chosen": -0.01229151152074337, | |
| "rewards/margins": 0.005582691170275211, | |
| "rewards/rejected": -0.017874203622341156, | |
| "step": 40 | |
| }, | |
| { | |
| "epoch": 0.11, | |
| "learning_rate": 4.999373573764188e-06, | |
| "logits/chosen": 0.1411871314048767, | |
| "logits/rejected": 0.2258455753326416, | |
| "logps/chosen": -612.8582763671875, | |
| "logps/rejected": -636.5026245117188, | |
| "loss": 2.1508, | |
| "rewards/accuracies": 0.5375000238418579, | |
| "rewards/chosen": -0.04941480979323387, | |
| "rewards/margins": 0.019247086718678474, | |
| "rewards/rejected": -0.0686618983745575, | |
| "step": 50 | |
| }, | |
| { | |
| "epoch": 0.13, | |
| "learning_rate": 4.988245838331339e-06, | |
| "logits/chosen": 0.17244111001491547, | |
| "logits/rejected": 0.17342150211334229, | |
| "logps/chosen": -634.6348266601562, | |
| "logps/rejected": -667.5384521484375, | |
| "loss": 2.0758, | |
| "rewards/accuracies": 0.581250011920929, | |
| "rewards/chosen": -0.13025899231433868, | |
| "rewards/margins": 0.05111612752079964, | |
| "rewards/rejected": -0.18137511610984802, | |
| "step": 60 | |
| }, | |
| { | |
| "epoch": 0.15, | |
| "learning_rate": 4.963268819535228e-06, | |
| "logits/chosen": 0.12650486826896667, | |
| "logits/rejected": 0.14093999564647675, | |
| "logps/chosen": -608.5107421875, | |
| "logps/rejected": -702.1578369140625, | |
| "loss": 2.0556, | |
| "rewards/accuracies": 0.5625, | |
| "rewards/chosen": -0.21341009438037872, | |
| "rewards/margins": 0.09893321990966797, | |
| "rewards/rejected": -0.3123432993888855, | |
| "step": 70 | |
| }, | |
| { | |
| "epoch": 0.17, | |
| "learning_rate": 4.9245815365216115e-06, | |
| "logits/chosen": 0.19184628129005432, | |
| "logits/rejected": 0.2408786565065384, | |
| "logps/chosen": -679.4183349609375, | |
| "logps/rejected": -609.7093505859375, | |
| "loss": 2.1137, | |
| "rewards/accuracies": 0.5562499761581421, | |
| "rewards/chosen": -0.21463651955127716, | |
| "rewards/margins": 0.05772104859352112, | |
| "rewards/rejected": -0.2723575234413147, | |
| "step": 80 | |
| }, | |
| { | |
| "epoch": 0.19, | |
| "learning_rate": 4.872399318152594e-06, | |
| "logits/chosen": 0.1250939965248108, | |
| "logits/rejected": 0.18045032024383545, | |
| "logps/chosen": -622.2333374023438, | |
| "logps/rejected": -655.4575805664062, | |
| "loss": 2.0044, | |
| "rewards/accuracies": 0.606249988079071, | |
| "rewards/chosen": -0.1839352548122406, | |
| "rewards/margins": 0.10977420955896378, | |
| "rewards/rejected": -0.2937094569206238, | |
| "step": 90 | |
| }, | |
| { | |
| "epoch": 0.21, | |
| "learning_rate": 4.807012604511542e-06, | |
| "logits/chosen": 0.18265239894390106, | |
| "logits/rejected": 0.2614283859729767, | |
| "logps/chosen": -649.8997802734375, | |
| "logps/rejected": -658.8975830078125, | |
| "loss": 1.9995, | |
| "rewards/accuracies": 0.643750011920929, | |
| "rewards/chosen": -0.16064395010471344, | |
| "rewards/margins": 0.08805385231971741, | |
| "rewards/rejected": -0.24869783222675323, | |
| "step": 100 | |
| }, | |
| { | |
| "epoch": 0.23, | |
| "learning_rate": 4.728785330347771e-06, | |
| "logits/chosen": 0.2479465901851654, | |
| "logits/rejected": 0.2932817339897156, | |
| "logps/chosen": -674.0836181640625, | |
| "logps/rejected": -645.6417236328125, | |
| "loss": 1.895, | |
| "rewards/accuracies": 0.606249988079071, | |
| "rewards/chosen": -0.12688389420509338, | |
| "rewards/margins": 0.08782283961772919, | |
| "rewards/rejected": -0.21470670402050018, | |
| "step": 110 | |
| }, | |
| { | |
| "epoch": 0.26, | |
| "learning_rate": 4.63815289945858e-06, | |
| "logits/chosen": 0.19643843173980713, | |
| "logits/rejected": 0.2974274456501007, | |
| "logps/chosen": -573.49658203125, | |
| "logps/rejected": -666.606689453125, | |
| "loss": 1.89, | |
| "rewards/accuracies": 0.6812499761581421, | |
| "rewards/chosen": -0.09824337065219879, | |
| "rewards/margins": 0.13982543349266052, | |
| "rewards/rejected": -0.2380688190460205, | |
| "step": 120 | |
| }, | |
| { | |
| "epoch": 0.28, | |
| "learning_rate": 4.535619761282989e-06, | |
| "logits/chosen": 0.23821644484996796, | |
| "logits/rejected": 0.288485586643219, | |
| "logps/chosen": -590.9158935546875, | |
| "logps/rejected": -623.23974609375, | |
| "loss": 1.9389, | |
| "rewards/accuracies": 0.612500011920929, | |
| "rewards/chosen": -0.14589470624923706, | |
| "rewards/margins": 0.12624357640743256, | |
| "rewards/rejected": -0.2721382975578308, | |
| "step": 130 | |
| }, | |
| { | |
| "epoch": 0.3, | |
| "learning_rate": 4.42175660319555e-06, | |
| "logits/chosen": 0.2631734013557434, | |
| "logits/rejected": 0.2810806632041931, | |
| "logps/chosen": -645.8680419921875, | |
| "logps/rejected": -654.8004760742188, | |
| "loss": 1.8203, | |
| "rewards/accuracies": 0.6187499761581421, | |
| "rewards/chosen": -0.20492109656333923, | |
| "rewards/margins": 0.20386295020580292, | |
| "rewards/rejected": -0.40878406167030334, | |
| "step": 140 | |
| }, | |
| { | |
| "epoch": 0.32, | |
| "learning_rate": 4.297197174127619e-06, | |
| "logits/chosen": 0.2586398422718048, | |
| "logits/rejected": 0.3086986839771271, | |
| "logps/chosen": -619.4220581054688, | |
| "logps/rejected": -697.2005615234375, | |
| "loss": 1.7553, | |
| "rewards/accuracies": 0.6187499761581421, | |
| "rewards/chosen": -0.24971242249011993, | |
| "rewards/margins": 0.2221045196056366, | |
| "rewards/rejected": -0.4718169569969177, | |
| "step": 150 | |
| }, | |
| { | |
| "epoch": 0.34, | |
| "learning_rate": 4.162634757195418e-06, | |
| "logits/chosen": 0.2681664526462555, | |
| "logits/rejected": 0.2807798683643341, | |
| "logps/chosen": -630.39306640625, | |
| "logps/rejected": -645.6117553710938, | |
| "loss": 1.8404, | |
| "rewards/accuracies": 0.612500011920929, | |
| "rewards/chosen": -0.19384464621543884, | |
| "rewards/margins": 0.1983100175857544, | |
| "rewards/rejected": -0.39215466380119324, | |
| "step": 160 | |
| }, | |
| { | |
| "epoch": 0.36, | |
| "learning_rate": 4.018818310967843e-06, | |
| "logits/chosen": 0.27496370673179626, | |
| "logits/rejected": 0.30781346559524536, | |
| "logps/chosen": -559.2887573242188, | |
| "logps/rejected": -601.3917846679688, | |
| "loss": 1.8382, | |
| "rewards/accuracies": 0.637499988079071, | |
| "rewards/chosen": -0.0747746005654335, | |
| "rewards/margins": 0.19791939854621887, | |
| "rewards/rejected": -0.2726939916610718, | |
| "step": 170 | |
| }, | |
| { | |
| "epoch": 0.38, | |
| "learning_rate": 3.866548300851254e-06, | |
| "logits/chosen": 0.2482290267944336, | |
| "logits/rejected": 0.2852781414985657, | |
| "logps/chosen": -620.8068237304688, | |
| "logps/rejected": -665.9005737304688, | |
| "loss": 1.8229, | |
| "rewards/accuracies": 0.625, | |
| "rewards/chosen": -0.10497160255908966, | |
| "rewards/margins": 0.20543234050273895, | |
| "rewards/rejected": -0.3104039430618286, | |
| "step": 180 | |
| }, | |
| { | |
| "epoch": 0.41, | |
| "learning_rate": 3.706672243793271e-06, | |
| "logits/chosen": 0.2958913743495941, | |
| "logits/rejected": 0.3795389235019684, | |
| "logps/chosen": -611.8587646484375, | |
| "logps/rejected": -658.9635009765625, | |
| "loss": 1.7752, | |
| "rewards/accuracies": 0.668749988079071, | |
| "rewards/chosen": -0.0870656967163086, | |
| "rewards/margins": 0.23995642364025116, | |
| "rewards/rejected": -0.32702213525772095, | |
| "step": 190 | |
| }, | |
| { | |
| "epoch": 0.43, | |
| "learning_rate": 3.5400799911032357e-06, | |
| "logits/chosen": 0.2935205101966858, | |
| "logits/rejected": 0.3416239321231842, | |
| "logps/chosen": -660.2877197265625, | |
| "logps/rejected": -730.04541015625, | |
| "loss": 1.7351, | |
| "rewards/accuracies": 0.668749988079071, | |
| "rewards/chosen": -0.1828436255455017, | |
| "rewards/margins": 0.3010478913784027, | |
| "rewards/rejected": -0.4838915765285492, | |
| "step": 200 | |
| }, | |
| { | |
| "epoch": 0.45, | |
| "learning_rate": 3.3676987756445894e-06, | |
| "logits/chosen": 0.24807122349739075, | |
| "logits/rejected": 0.32862648367881775, | |
| "logps/chosen": -605.8773193359375, | |
| "logps/rejected": -641.6677856445312, | |
| "loss": 1.8245, | |
| "rewards/accuracies": 0.7124999761581421, | |
| "rewards/chosen": -0.13868902623653412, | |
| "rewards/margins": 0.2735101878643036, | |
| "rewards/rejected": -0.4121991991996765, | |
| "step": 210 | |
| }, | |
| { | |
| "epoch": 0.47, | |
| "learning_rate": 3.1904880509659397e-06, | |
| "logits/chosen": 0.270724892616272, | |
| "logits/rejected": 0.3151053786277771, | |
| "logps/chosen": -650.7314453125, | |
| "logps/rejected": -708.2312622070312, | |
| "loss": 1.735, | |
| "rewards/accuracies": 0.637499988079071, | |
| "rewards/chosen": -0.16527561843395233, | |
| "rewards/margins": 0.2484448254108429, | |
| "rewards/rejected": -0.4137204587459564, | |
| "step": 220 | |
| }, | |
| { | |
| "epoch": 0.49, | |
| "learning_rate": 3.0094341510955697e-06, | |
| "logits/chosen": 0.19233042001724243, | |
| "logits/rejected": 0.29483872652053833, | |
| "logps/chosen": -663.5474243164062, | |
| "logps/rejected": -743.0173950195312, | |
| "loss": 1.7378, | |
| "rewards/accuracies": 0.7562500238418579, | |
| "rewards/chosen": -0.14797742664813995, | |
| "rewards/margins": 0.3706679344177246, | |
| "rewards/rejected": -0.5186454057693481, | |
| "step": 230 | |
| }, | |
| { | |
| "epoch": 0.51, | |
| "learning_rate": 2.825544800722376e-06, | |
| "logits/chosen": 0.2124979943037033, | |
| "logits/rejected": 0.3365432620048523, | |
| "logps/chosen": -619.9740600585938, | |
| "logps/rejected": -700.7166748046875, | |
| "loss": 1.8168, | |
| "rewards/accuracies": 0.5874999761581421, | |
| "rewards/chosen": -0.1956629902124405, | |
| "rewards/margins": 0.2987174093723297, | |
| "rewards/rejected": -0.494380384683609, | |
| "step": 240 | |
| }, | |
| { | |
| "epoch": 0.53, | |
| "learning_rate": 2.639843506318899e-06, | |
| "logits/chosen": 0.2796134054660797, | |
| "logits/rejected": 0.2740449607372284, | |
| "logps/chosen": -582.3416748046875, | |
| "logps/rejected": -674.327880859375, | |
| "loss": 1.8901, | |
| "rewards/accuracies": 0.6187499761581421, | |
| "rewards/chosen": -0.19822832942008972, | |
| "rewards/margins": 0.19228845834732056, | |
| "rewards/rejected": -0.3905167877674103, | |
| "step": 250 | |
| }, | |
| { | |
| "epoch": 0.55, | |
| "learning_rate": 2.4533638594248094e-06, | |
| "logits/chosen": 0.25897207856178284, | |
| "logits/rejected": 0.31485193967819214, | |
| "logps/chosen": -604.8118896484375, | |
| "logps/rejected": -667.9144897460938, | |
| "loss": 1.8606, | |
| "rewards/accuracies": 0.6499999761581421, | |
| "rewards/chosen": -0.14205999672412872, | |
| "rewards/margins": 0.28450149297714233, | |
| "rewards/rejected": -0.42656150460243225, | |
| "step": 260 | |
| }, | |
| { | |
| "epoch": 0.58, | |
| "learning_rate": 2.2671437837980943e-06, | |
| "logits/chosen": 0.22259187698364258, | |
| "logits/rejected": 0.22855930030345917, | |
| "logps/chosen": -593.6612548828125, | |
| "logps/rejected": -673.6566162109375, | |
| "loss": 1.7486, | |
| "rewards/accuracies": 0.6499999761581421, | |
| "rewards/chosen": -0.14823240041732788, | |
| "rewards/margins": 0.2802043557167053, | |
| "rewards/rejected": -0.4284366965293884, | |
| "step": 270 | |
| }, | |
| { | |
| "epoch": 0.6, | |
| "learning_rate": 2.082219758453629e-06, | |
| "logits/chosen": 0.2169434130191803, | |
| "logits/rejected": 0.2703471779823303, | |
| "logps/chosen": -611.6048583984375, | |
| "logps/rejected": -682.5806884765625, | |
| "loss": 1.6556, | |
| "rewards/accuracies": 0.6625000238418579, | |
| "rewards/chosen": -0.11671599000692368, | |
| "rewards/margins": 0.26952022314071655, | |
| "rewards/rejected": -0.3862362205982208, | |
| "step": 280 | |
| }, | |
| { | |
| "epoch": 0.62, | |
| "learning_rate": 1.899621048743019e-06, | |
| "logits/chosen": 0.22146745026111603, | |
| "logits/rejected": 0.34733515977859497, | |
| "logps/chosen": -603.9933471679688, | |
| "logps/rejected": -673.3649291992188, | |
| "loss": 1.7238, | |
| "rewards/accuracies": 0.65625, | |
| "rewards/chosen": -0.20439250767230988, | |
| "rewards/margins": 0.2682177424430847, | |
| "rewards/rejected": -0.4726102352142334, | |
| "step": 290 | |
| }, | |
| { | |
| "epoch": 0.64, | |
| "learning_rate": 1.7203639775848423e-06, | |
| "logits/chosen": 0.19099445641040802, | |
| "logits/rejected": 0.3011043667793274, | |
| "logps/chosen": -606.6263427734375, | |
| "logps/rejected": -639.6136474609375, | |
| "loss": 1.8381, | |
| "rewards/accuracies": 0.643750011920929, | |
| "rewards/chosen": -0.17796705663204193, | |
| "rewards/margins": 0.23042461276054382, | |
| "rewards/rejected": -0.40839165449142456, | |
| "step": 300 | |
| }, | |
| { | |
| "epoch": 0.66, | |
| "learning_rate": 1.5454462687309445e-06, | |
| "logits/chosen": 0.2036764919757843, | |
| "logits/rejected": 0.26239025592803955, | |
| "logps/chosen": -602.3845825195312, | |
| "logps/rejected": -666.4627075195312, | |
| "loss": 1.8042, | |
| "rewards/accuracies": 0.6499999761581421, | |
| "rewards/chosen": -0.1518932580947876, | |
| "rewards/margins": 0.2536298632621765, | |
| "rewards/rejected": -0.4055231511592865, | |
| "step": 310 | |
| }, | |
| { | |
| "epoch": 0.68, | |
| "learning_rate": 1.3758414935535147e-06, | |
| "logits/chosen": 0.21739721298217773, | |
| "logits/rejected": 0.2840099334716797, | |
| "logps/chosen": -636.0455322265625, | |
| "logps/rejected": -709.1137084960938, | |
| "loss": 1.65, | |
| "rewards/accuracies": 0.65625, | |
| "rewards/chosen": -0.16815349459648132, | |
| "rewards/margins": 0.29733169078826904, | |
| "rewards/rejected": -0.465485155582428, | |
| "step": 320 | |
| }, | |
| { | |
| "epoch": 0.7, | |
| "learning_rate": 1.2124936522614622e-06, | |
| "logits/chosen": 0.20938508212566376, | |
| "logits/rejected": 0.22490420937538147, | |
| "logps/chosen": -615.7994995117188, | |
| "logps/rejected": -669.2200927734375, | |
| "loss": 1.7098, | |
| "rewards/accuracies": 0.6937500238418579, | |
| "rewards/chosen": -0.18394342064857483, | |
| "rewards/margins": 0.31033387780189514, | |
| "rewards/rejected": -0.49427732825279236, | |
| "step": 330 | |
| }, | |
| { | |
| "epoch": 0.73, | |
| "learning_rate": 1.0563119197063934e-06, | |
| "logits/chosen": 0.23827771842479706, | |
| "logits/rejected": 0.2663131356239319, | |
| "logps/chosen": -612.7750244140625, | |
| "logps/rejected": -685.60107421875, | |
| "loss": 1.7109, | |
| "rewards/accuracies": 0.637499988079071, | |
| "rewards/chosen": -0.19161880016326904, | |
| "rewards/margins": 0.26392242312431335, | |
| "rewards/rejected": -0.4555412232875824, | |
| "step": 340 | |
| }, | |
| { | |
| "epoch": 0.75, | |
| "learning_rate": 9.081655850224449e-07, | |
| "logits/chosen": 0.19827114045619965, | |
| "logits/rejected": 0.2343660295009613, | |
| "logps/chosen": -628.5892333984375, | |
| "logps/rejected": -699.3311767578125, | |
| "loss": 1.6981, | |
| "rewards/accuracies": 0.675000011920929, | |
| "rewards/chosen": -0.23514249920845032, | |
| "rewards/margins": 0.30311545729637146, | |
| "rewards/rejected": -0.5382579565048218, | |
| "step": 350 | |
| }, | |
| { | |
| "epoch": 0.77, | |
| "learning_rate": 7.688792132653111e-07, | |
| "logits/chosen": 0.19120459258556366, | |
| "logits/rejected": 0.2861759066581726, | |
| "logps/chosen": -659.7528076171875, | |
| "logps/rejected": -748.490234375, | |
| "loss": 1.6967, | |
| "rewards/accuracies": 0.675000011920929, | |
| "rewards/chosen": -0.19031907618045807, | |
| "rewards/margins": 0.34352895617485046, | |
| "rewards/rejected": -0.533847987651825, | |
| "step": 360 | |
| }, | |
| { | |
| "epoch": 0.79, | |
| "learning_rate": 6.392280559802341e-07, | |
| "logits/chosen": 0.2406836450099945, | |
| "logits/rejected": 0.23908407986164093, | |
| "logps/chosen": -658.35400390625, | |
| "logps/rejected": -720.8883666992188, | |
| "loss": 1.7368, | |
| "rewards/accuracies": 0.625, | |
| "rewards/chosen": -0.23702308535575867, | |
| "rewards/margins": 0.24957367777824402, | |
| "rewards/rejected": -0.48659682273864746, | |
| "step": 370 | |
| }, | |
| { | |
| "epoch": 0.81, | |
| "learning_rate": 5.199337362431792e-07, | |
| "logits/chosen": 0.26719361543655396, | |
| "logits/rejected": 0.1743316501379013, | |
| "logps/chosen": -621.3897094726562, | |
| "logps/rejected": -680.0, | |
| "loss": 1.7425, | |
| "rewards/accuracies": 0.6499999761581421, | |
| "rewards/chosen": -0.22942551970481873, | |
| "rewards/margins": 0.26667481660842896, | |
| "rewards/rejected": -0.49610036611557007, | |
| "step": 380 | |
| }, | |
| { | |
| "epoch": 0.83, | |
| "learning_rate": 4.1166023219176176e-07, | |
| "logits/chosen": 0.21561181545257568, | |
| "logits/rejected": 0.286629855632782, | |
| "logps/chosen": -654.0867919921875, | |
| "logps/rejected": -668.467529296875, | |
| "loss": 1.6798, | |
| "rewards/accuracies": 0.699999988079071, | |
| "rewards/chosen": -0.21663355827331543, | |
| "rewards/margins": 0.2600599527359009, | |
| "rewards/rejected": -0.4766935408115387, | |
| "step": 390 | |
| }, | |
| { | |
| "epoch": 0.85, | |
| "learning_rate": 3.150101814011136e-07, | |
| "logits/chosen": 0.16323356330394745, | |
| "logits/rejected": 0.21500280499458313, | |
| "logps/chosen": -600.4713134765625, | |
| "logps/rejected": -730.5057983398438, | |
| "loss": 1.7084, | |
| "rewards/accuracies": 0.6499999761581421, | |
| "rewards/chosen": -0.19050315022468567, | |
| "rewards/margins": 0.28324562311172485, | |
| "rewards/rejected": -0.47374874353408813, | |
| "step": 400 | |
| }, | |
| { | |
| "epoch": 0.87, | |
| "learning_rate": 2.3052152667409289e-07, | |
| "logits/chosen": 0.1962326616048813, | |
| "logits/rejected": 0.22506949305534363, | |
| "logps/chosen": -614.2760009765625, | |
| "logps/rejected": -675.3383178710938, | |
| "loss": 1.7679, | |
| "rewards/accuracies": 0.706250011920929, | |
| "rewards/chosen": -0.1459427773952484, | |
| "rewards/margins": 0.3252793252468109, | |
| "rewards/rejected": -0.4712221026420593, | |
| "step": 410 | |
| }, | |
| { | |
| "epoch": 0.9, | |
| "learning_rate": 1.5866452191498488e-07, | |
| "logits/chosen": 0.20015636086463928, | |
| "logits/rejected": 0.25162121653556824, | |
| "logps/chosen": -651.9236450195312, | |
| "logps/rejected": -707.2882080078125, | |
| "loss": 1.7514, | |
| "rewards/accuracies": 0.59375, | |
| "rewards/chosen": -0.23218846321105957, | |
| "rewards/margins": 0.2290785312652588, | |
| "rewards/rejected": -0.46126699447631836, | |
| "step": 420 | |
| }, | |
| { | |
| "epoch": 0.92, | |
| "learning_rate": 9.983911475163727e-08, | |
| "logits/chosen": 0.16698592901229858, | |
| "logits/rejected": 0.2591376304626465, | |
| "logps/chosen": -590.045166015625, | |
| "logps/rejected": -642.6705322265625, | |
| "loss": 1.8093, | |
| "rewards/accuracies": 0.6187499761581421, | |
| "rewards/chosen": -0.17988340556621552, | |
| "rewards/margins": 0.23005299270153046, | |
| "rewards/rejected": -0.4099363684654236, | |
| "step": 430 | |
| }, | |
| { | |
| "epoch": 0.94, | |
| "learning_rate": 5.437272047405712e-08, | |
| "logits/chosen": 0.1858983337879181, | |
| "logits/rejected": 0.3158418536186218, | |
| "logps/chosen": -559.8682250976562, | |
| "logps/rejected": -648.7040405273438, | |
| "loss": 1.7686, | |
| "rewards/accuracies": 0.6312500238418579, | |
| "rewards/chosen": -0.19649046659469604, | |
| "rewards/margins": 0.26454511284828186, | |
| "rewards/rejected": -0.4610355794429779, | |
| "step": 440 | |
| }, | |
| { | |
| "epoch": 0.96, | |
| "learning_rate": 2.251839967945535e-08, | |
| "logits/chosen": 0.13786078989505768, | |
| "logits/rejected": 0.2333669662475586, | |
| "logps/chosen": -645.2703857421875, | |
| "logps/rejected": -707.0418090820312, | |
| "loss": 1.6172, | |
| "rewards/accuracies": 0.6625000238418579, | |
| "rewards/chosen": -0.17300908267498016, | |
| "rewards/margins": 0.3292023241519928, | |
| "rewards/rejected": -0.5022113919258118, | |
| "step": 450 | |
| }, | |
| { | |
| "epoch": 0.98, | |
| "learning_rate": 4.453449766758933e-09, | |
| "logits/chosen": 0.1742466688156128, | |
| "logits/rejected": 0.2268284559249878, | |
| "logps/chosen": -576.7985229492188, | |
| "logps/rejected": -652.7803344726562, | |
| "loss": 1.7297, | |
| "rewards/accuracies": 0.625, | |
| "rewards/chosen": -0.24071533977985382, | |
| "rewards/margins": 0.18981412053108215, | |
| "rewards/rejected": -0.4305294454097748, | |
| "step": 460 | |
| }, | |
| { | |
| "epoch": 1.0, | |
| "step": 468, | |
| "total_flos": 0.0, | |
| "train_loss": 1.8394590188295414, | |
| "train_runtime": 15861.5475, | |
| "train_samples_per_second": 1.891, | |
| "train_steps_per_second": 0.03 | |
| } | |
| ], | |
| "logging_steps": 10, | |
| "max_steps": 468, | |
| "num_input_tokens_seen": 0, | |
| "num_train_epochs": 1, | |
| "save_steps": 100, | |
| "total_flos": 0.0, | |
| "train_batch_size": 4, | |
| "trial_name": null, | |
| "trial_params": null | |
| } | |