| { | |
| "best_metric": null, | |
| "best_model_checkpoint": null, | |
| "epoch": 0.9994767137624281, | |
| "eval_steps": 100, | |
| "global_step": 955, | |
| "is_hyper_param_search": false, | |
| "is_local_process_zero": true, | |
| "is_world_process_zero": true, | |
| "log_history": [ | |
| { | |
| "epoch": 0.0010465724751439038, | |
| "grad_norm": 21.102116873134612, | |
| "learning_rate": 5.208333333333333e-09, | |
| "logits/chosen": -2.924262046813965, | |
| "logits/rejected": -2.7925047874450684, | |
| "logps/chosen": -380.8447570800781, | |
| "logps/rejected": -358.51123046875, | |
| "loss": 4.6506, | |
| "rewards/accuracies": 0.0, | |
| "rewards/chosen": 0.0, | |
| "rewards/margins": 0.0, | |
| "rewards/rejected": 0.0, | |
| "step": 1 | |
| }, | |
| { | |
| "epoch": 0.010465724751439037, | |
| "grad_norm": 15.822543074567085, | |
| "learning_rate": 5.208333333333333e-08, | |
| "logits/chosen": -2.595761299133301, | |
| "logits/rejected": -2.569227457046509, | |
| "logps/chosen": -256.6064453125, | |
| "logps/rejected": -234.93408203125, | |
| "loss": 4.5621, | |
| "rewards/accuracies": 0.5, | |
| "rewards/chosen": 0.00042897689854726195, | |
| "rewards/margins": 0.0009927540086209774, | |
| "rewards/rejected": -0.0005637770518660545, | |
| "step": 10 | |
| }, | |
| { | |
| "epoch": 0.020931449502878074, | |
| "grad_norm": 18.010820015079055, | |
| "learning_rate": 1.0416666666666667e-07, | |
| "logits/chosen": -2.613164186477661, | |
| "logits/rejected": -2.5756287574768066, | |
| "logps/chosen": -283.0158996582031, | |
| "logps/rejected": -282.265869140625, | |
| "loss": 4.4053, | |
| "rewards/accuracies": 0.543749988079071, | |
| "rewards/chosen": 0.0006733193295076489, | |
| "rewards/margins": 0.0005819452926516533, | |
| "rewards/rejected": 9.137402230408043e-05, | |
| "step": 20 | |
| }, | |
| { | |
| "epoch": 0.03139717425431711, | |
| "grad_norm": 21.44807572026145, | |
| "learning_rate": 1.5624999999999999e-07, | |
| "logits/chosen": -2.691143035888672, | |
| "logits/rejected": -2.6666667461395264, | |
| "logps/chosen": -269.9042053222656, | |
| "logps/rejected": -276.4795837402344, | |
| "loss": 5.105, | |
| "rewards/accuracies": 0.574999988079071, | |
| "rewards/chosen": 0.0037794082891196012, | |
| "rewards/margins": 0.0018267262494191527, | |
| "rewards/rejected": 0.0019526820397004485, | |
| "step": 30 | |
| }, | |
| { | |
| "epoch": 0.04186289900575615, | |
| "grad_norm": 17.302023991146115, | |
| "learning_rate": 2.0833333333333333e-07, | |
| "logits/chosen": -2.6577816009521484, | |
| "logits/rejected": -2.5818943977355957, | |
| "logps/chosen": -288.9285888671875, | |
| "logps/rejected": -280.9770202636719, | |
| "loss": 4.9032, | |
| "rewards/accuracies": 0.606249988079071, | |
| "rewards/chosen": 0.020702064037322998, | |
| "rewards/margins": 0.009830506518483162, | |
| "rewards/rejected": 0.01087155845016241, | |
| "step": 40 | |
| }, | |
| { | |
| "epoch": 0.052328623757195186, | |
| "grad_norm": 22.46337927130885, | |
| "learning_rate": 2.604166666666667e-07, | |
| "logits/chosen": -2.6507585048675537, | |
| "logits/rejected": -2.5627222061157227, | |
| "logps/chosen": -263.1905212402344, | |
| "logps/rejected": -234.9305419921875, | |
| "loss": 4.8274, | |
| "rewards/accuracies": 0.606249988079071, | |
| "rewards/chosen": 0.044054824858903885, | |
| "rewards/margins": 0.02749818004667759, | |
| "rewards/rejected": 0.016556641086935997, | |
| "step": 50 | |
| }, | |
| { | |
| "epoch": 0.06279434850863422, | |
| "grad_norm": 18.98737987603255, | |
| "learning_rate": 3.1249999999999997e-07, | |
| "logits/chosen": -2.5976526737213135, | |
| "logits/rejected": -2.5587098598480225, | |
| "logps/chosen": -299.9574890136719, | |
| "logps/rejected": -276.1783142089844, | |
| "loss": 4.5279, | |
| "rewards/accuracies": 0.6312500238418579, | |
| "rewards/chosen": 0.040667824447155, | |
| "rewards/margins": 0.04492232948541641, | |
| "rewards/rejected": -0.004254504106938839, | |
| "step": 60 | |
| }, | |
| { | |
| "epoch": 0.07326007326007326, | |
| "grad_norm": 20.501382800234886, | |
| "learning_rate": 3.645833333333333e-07, | |
| "logits/chosen": -2.469130039215088, | |
| "logits/rejected": -2.452857732772827, | |
| "logps/chosen": -265.96978759765625, | |
| "logps/rejected": -271.6788330078125, | |
| "loss": 4.6703, | |
| "rewards/accuracies": 0.5625, | |
| "rewards/chosen": -0.0633089542388916, | |
| "rewards/margins": 0.07126398384571075, | |
| "rewards/rejected": -0.13457295298576355, | |
| "step": 70 | |
| }, | |
| { | |
| "epoch": 0.0837257980115123, | |
| "grad_norm": 25.49997843488533, | |
| "learning_rate": 4.1666666666666667e-07, | |
| "logits/chosen": -2.4551777839660645, | |
| "logits/rejected": -2.3624327182769775, | |
| "logps/chosen": -285.5320739746094, | |
| "logps/rejected": -276.4596252441406, | |
| "loss": 4.5605, | |
| "rewards/accuracies": 0.6875, | |
| "rewards/chosen": -0.09489366412162781, | |
| "rewards/margins": 0.15657536685466766, | |
| "rewards/rejected": -0.2514690160751343, | |
| "step": 80 | |
| }, | |
| { | |
| "epoch": 0.09419152276295134, | |
| "grad_norm": 30.61647338954573, | |
| "learning_rate": 4.6874999999999996e-07, | |
| "logits/chosen": -2.3756256103515625, | |
| "logits/rejected": -2.332918882369995, | |
| "logps/chosen": -277.46014404296875, | |
| "logps/rejected": -290.0049743652344, | |
| "loss": 4.1231, | |
| "rewards/accuracies": 0.5562499761581421, | |
| "rewards/chosen": -0.21862252056598663, | |
| "rewards/margins": 0.033695660531520844, | |
| "rewards/rejected": -0.25231820344924927, | |
| "step": 90 | |
| }, | |
| { | |
| "epoch": 0.10465724751439037, | |
| "grad_norm": 38.124561793065574, | |
| "learning_rate": 4.999732492681437e-07, | |
| "logits/chosen": -2.332035779953003, | |
| "logits/rejected": -2.2253689765930176, | |
| "logps/chosen": -314.4341125488281, | |
| "logps/rejected": -317.18695068359375, | |
| "loss": 4.5854, | |
| "rewards/accuracies": 0.637499988079071, | |
| "rewards/chosen": -0.1460995227098465, | |
| "rewards/margins": 0.22573721408843994, | |
| "rewards/rejected": -0.37183672189712524, | |
| "step": 100 | |
| }, | |
| { | |
| "epoch": 0.10465724751439037, | |
| "eval_logits/chosen": -2.2812609672546387, | |
| "eval_logits/rejected": -2.192293167114258, | |
| "eval_logps/chosen": -309.1551818847656, | |
| "eval_logps/rejected": -310.1242370605469, | |
| "eval_loss": 4.381103515625, | |
| "eval_rewards/accuracies": 0.648809552192688, | |
| "eval_rewards/chosen": -0.2718724012374878, | |
| "eval_rewards/margins": 0.2273014634847641, | |
| "eval_rewards/rejected": -0.4991738498210907, | |
| "eval_runtime": 176.2372, | |
| "eval_samples_per_second": 11.348, | |
| "eval_steps_per_second": 0.357, | |
| "step": 100 | |
| }, | |
| { | |
| "epoch": 0.1151229722658294, | |
| "grad_norm": 47.336977780094564, | |
| "learning_rate": 4.996723692767926e-07, | |
| "logits/chosen": -2.0436112880706787, | |
| "logits/rejected": -1.9534924030303955, | |
| "logps/chosen": -310.6973571777344, | |
| "logps/rejected": -324.1681823730469, | |
| "loss": 3.758, | |
| "rewards/accuracies": 0.574999988079071, | |
| "rewards/chosen": -0.6924275755882263, | |
| "rewards/margins": 0.17653007805347443, | |
| "rewards/rejected": -0.8689576387405396, | |
| "step": 110 | |
| }, | |
| { | |
| "epoch": 0.12558869701726844, | |
| "grad_norm": 109.43376131471078, | |
| "learning_rate": 4.990375746213598e-07, | |
| "logits/chosen": -0.08515436947345734, | |
| "logits/rejected": 0.34949326515197754, | |
| "logps/chosen": -343.26495361328125, | |
| "logps/rejected": -412.98577880859375, | |
| "loss": 4.0333, | |
| "rewards/accuracies": 0.65625, | |
| "rewards/chosen": -0.8068662881851196, | |
| "rewards/margins": 0.438527911901474, | |
| "rewards/rejected": -1.2453943490982056, | |
| "step": 120 | |
| }, | |
| { | |
| "epoch": 0.1360544217687075, | |
| "grad_norm": 95.04671304091885, | |
| "learning_rate": 4.980697142834314e-07, | |
| "logits/chosen": 0.396954745054245, | |
| "logits/rejected": 1.0232269763946533, | |
| "logps/chosen": -406.28521728515625, | |
| "logps/rejected": -430.10760498046875, | |
| "loss": 4.2005, | |
| "rewards/accuracies": 0.6312500238418579, | |
| "rewards/chosen": -1.2501262426376343, | |
| "rewards/margins": 0.5063079595565796, | |
| "rewards/rejected": -1.7564342021942139, | |
| "step": 130 | |
| }, | |
| { | |
| "epoch": 0.14652014652014653, | |
| "grad_norm": 144.39035434160894, | |
| "learning_rate": 4.967700826904229e-07, | |
| "logits/chosen": -0.1560676395893097, | |
| "logits/rejected": 0.6105406880378723, | |
| "logps/chosen": -416.2538146972656, | |
| "logps/rejected": -463.2472229003906, | |
| "loss": 3.7876, | |
| "rewards/accuracies": 0.668749988079071, | |
| "rewards/chosen": -1.2516638040542603, | |
| "rewards/margins": 0.49105915427207947, | |
| "rewards/rejected": -1.7427231073379517, | |
| "step": 140 | |
| }, | |
| { | |
| "epoch": 0.15698587127158556, | |
| "grad_norm": 125.21681673589694, | |
| "learning_rate": 4.951404179843962e-07, | |
| "logits/chosen": 2.0407581329345703, | |
| "logits/rejected": 2.8481547832489014, | |
| "logps/chosen": -510.521484375, | |
| "logps/rejected": -534.6341552734375, | |
| "loss": 3.898, | |
| "rewards/accuracies": 0.625, | |
| "rewards/chosen": -2.226250648498535, | |
| "rewards/margins": 0.6501102447509766, | |
| "rewards/rejected": -2.876361131668091, | |
| "step": 150 | |
| }, | |
| { | |
| "epoch": 0.1674515960230246, | |
| "grad_norm": 66.88313091855639, | |
| "learning_rate": 4.931828996974498e-07, | |
| "logits/chosen": 2.163175106048584, | |
| "logits/rejected": 3.5420451164245605, | |
| "logps/chosen": -585.4688720703125, | |
| "logps/rejected": -635.2697143554688, | |
| "loss": 3.9393, | |
| "rewards/accuracies": 0.675000011920929, | |
| "rewards/chosen": -2.89656138420105, | |
| "rewards/margins": 0.8964195251464844, | |
| "rewards/rejected": -3.792980909347534, | |
| "step": 160 | |
| }, | |
| { | |
| "epoch": 0.17791732077446362, | |
| "grad_norm": 188.98325062900707, | |
| "learning_rate": 4.909001458367866e-07, | |
| "logits/chosen": 0.49319368600845337, | |
| "logits/rejected": 1.3766599893569946, | |
| "logps/chosen": -599.5331420898438, | |
| "logps/rejected": -654.1383056640625, | |
| "loss": 3.9922, | |
| "rewards/accuracies": 0.5874999761581421, | |
| "rewards/chosen": -3.210639476776123, | |
| "rewards/margins": 0.922272801399231, | |
| "rewards/rejected": -4.132911682128906, | |
| "step": 170 | |
| }, | |
| { | |
| "epoch": 0.18838304552590268, | |
| "grad_norm": 320.6202106283321, | |
| "learning_rate": 4.882952093833627e-07, | |
| "logits/chosen": 0.6820823550224304, | |
| "logits/rejected": 1.588409185409546, | |
| "logps/chosen": -1040.5491943359375, | |
| "logps/rejected": -1233.1207275390625, | |
| "loss": 3.36, | |
| "rewards/accuracies": 0.706250011920929, | |
| "rewards/chosen": -8.006011009216309, | |
| "rewards/margins": 1.8573882579803467, | |
| "rewards/rejected": -9.86340045928955, | |
| "step": 180 | |
| }, | |
| { | |
| "epoch": 0.1988487702773417, | |
| "grad_norm": 157.79546381015746, | |
| "learning_rate": 4.853715742087946e-07, | |
| "logits/chosen": 3.3087031841278076, | |
| "logits/rejected": 4.11985445022583, | |
| "logps/chosen": -1690.8167724609375, | |
| "logps/rejected": -1890.6634521484375, | |
| "loss": 2.6799, | |
| "rewards/accuracies": 0.574999988079071, | |
| "rewards/chosen": -14.548372268676758, | |
| "rewards/margins": 1.8595011234283447, | |
| "rewards/rejected": -16.407875061035156, | |
| "step": 190 | |
| }, | |
| { | |
| "epoch": 0.20931449502878074, | |
| "grad_norm": 178.97245767319544, | |
| "learning_rate": 4.821331504159906e-07, | |
| "logits/chosen": 0.3337511122226715, | |
| "logits/rejected": 1.9961885213851929, | |
| "logps/chosen": -1578.712158203125, | |
| "logps/rejected": -1801.65625, | |
| "loss": 2.6464, | |
| "rewards/accuracies": 0.637499988079071, | |
| "rewards/chosen": -12.680551528930664, | |
| "rewards/margins": 2.764849901199341, | |
| "rewards/rejected": -15.445402145385742, | |
| "step": 200 | |
| }, | |
| { | |
| "epoch": 0.20931449502878074, | |
| "eval_logits/chosen": -0.35622134804725647, | |
| "eval_logits/rejected": 0.6981890797615051, | |
| "eval_logps/chosen": -1244.43603515625, | |
| "eval_logps/rejected": -1423.3580322265625, | |
| "eval_loss": 2.606262683868408, | |
| "eval_rewards/accuracies": 0.625, | |
| "eval_rewards/chosen": -9.624680519104004, | |
| "eval_rewards/margins": 2.0068302154541016, | |
| "eval_rewards/rejected": -11.631510734558105, | |
| "eval_runtime": 177.3795, | |
| "eval_samples_per_second": 11.275, | |
| "eval_steps_per_second": 0.355, | |
| "step": 200 | |
| }, | |
| { | |
| "epoch": 0.21978021978021978, | |
| "grad_norm": 221.39959720400535, | |
| "learning_rate": 4.785842691097342e-07, | |
| "logits/chosen": 0.43124809861183167, | |
| "logits/rejected": 1.6196168661117554, | |
| "logps/chosen": -1394.329345703125, | |
| "logps/rejected": -1612.8701171875, | |
| "loss": 2.2192, | |
| "rewards/accuracies": 0.6000000238418579, | |
| "rewards/chosen": -11.115188598632812, | |
| "rewards/margins": 2.29093337059021, | |
| "rewards/rejected": -13.406122207641602, | |
| "step": 210 | |
| }, | |
| { | |
| "epoch": 0.2302459445316588, | |
| "grad_norm": 107.97254065213261, | |
| "learning_rate": 4.7472967660421603e-07, | |
| "logits/chosen": 0.5400440096855164, | |
| "logits/rejected": 1.9760030508041382, | |
| "logps/chosen": -1507.001220703125, | |
| "logps/rejected": -1713.616455078125, | |
| "loss": 2.018, | |
| "rewards/accuracies": 0.6499999761581421, | |
| "rewards/chosen": -12.197932243347168, | |
| "rewards/margins": 2.5764663219451904, | |
| "rewards/rejected": -14.774396896362305, | |
| "step": 220 | |
| }, | |
| { | |
| "epoch": 0.24071166928309787, | |
| "grad_norm": 217.88193736039008, | |
| "learning_rate": 4.705745280752585e-07, | |
| "logits/chosen": 1.4225207567214966, | |
| "logits/rejected": 2.4756038188934326, | |
| "logps/chosen": -1726.320068359375, | |
| "logps/rejected": -2005.7041015625, | |
| "loss": 1.9719, | |
| "rewards/accuracies": 0.612500011920929, | |
| "rewards/chosen": -14.375930786132812, | |
| "rewards/margins": 2.995753526687622, | |
| "rewards/rejected": -17.37168312072754, | |
| "step": 230 | |
| }, | |
| { | |
| "epoch": 0.25117739403453687, | |
| "grad_norm": 109.77258728949327, | |
| "learning_rate": 4.6612438066572555e-07, | |
| "logits/chosen": 2.2113587856292725, | |
| "logits/rejected": 3.125591993331909, | |
| "logps/chosen": -1894.770751953125, | |
| "logps/rejected": -2110.86376953125, | |
| "loss": 1.9847, | |
| "rewards/accuracies": 0.59375, | |
| "rewards/chosen": -16.27196502685547, | |
| "rewards/margins": 2.294943332672119, | |
| "rewards/rejected": -18.56690788269043, | |
| "step": 240 | |
| }, | |
| { | |
| "epoch": 0.2616431187859759, | |
| "grad_norm": 276.53415343052893, | |
| "learning_rate": 4.6138518605333664e-07, | |
| "logits/chosen": 1.203977108001709, | |
| "logits/rejected": 1.9225616455078125, | |
| "logps/chosen": -1561.0047607421875, | |
| "logps/rejected": -1763.075439453125, | |
| "loss": 2.257, | |
| "rewards/accuracies": 0.6499999761581421, | |
| "rewards/chosen": -12.832204818725586, | |
| "rewards/margins": 2.222775936126709, | |
| "rewards/rejected": -15.05497932434082, | |
| "step": 250 | |
| }, | |
| { | |
| "epoch": 0.272108843537415, | |
| "grad_norm": 159.14963627253198, | |
| "learning_rate": 4.5636328249082514e-07, | |
| "logits/chosen": 1.134037733078003, | |
| "logits/rejected": 2.1568219661712646, | |
| "logps/chosen": -1608.8623046875, | |
| "logps/rejected": -1763.599853515625, | |
| "loss": 2.2606, | |
| "rewards/accuracies": 0.5874999761581421, | |
| "rewards/chosen": -13.198956489562988, | |
| "rewards/margins": 1.6145280599594116, | |
| "rewards/rejected": -14.813486099243164, | |
| "step": 260 | |
| }, | |
| { | |
| "epoch": 0.282574568288854, | |
| "grad_norm": 199.45417630865836, | |
| "learning_rate": 4.510653863290871e-07, | |
| "logits/chosen": 0.3547247350215912, | |
| "logits/rejected": 1.2751286029815674, | |
| "logps/chosen": -1781.0726318359375, | |
| "logps/rejected": -2089.05615234375, | |
| "loss": 1.7211, | |
| "rewards/accuracies": 0.6312500238418579, | |
| "rewards/chosen": -14.814226150512695, | |
| "rewards/margins": 3.540767192840576, | |
| "rewards/rejected": -18.354991912841797, | |
| "step": 270 | |
| }, | |
| { | |
| "epoch": 0.29304029304029305, | |
| "grad_norm": 162.5497817330968, | |
| "learning_rate": 4.4549858303465737e-07, | |
| "logits/chosen": 0.21130748093128204, | |
| "logits/rejected": 1.2269564867019653, | |
| "logps/chosen": -1743.0787353515625, | |
| "logps/rejected": -2033.669921875, | |
| "loss": 1.9445, | |
| "rewards/accuracies": 0.612500011920929, | |
| "rewards/chosen": -14.468345642089844, | |
| "rewards/margins": 3.0355849266052246, | |
| "rewards/rejected": -17.50392723083496, | |
| "step": 280 | |
| }, | |
| { | |
| "epoch": 0.3035060177917321, | |
| "grad_norm": 307.15808847538113, | |
| "learning_rate": 4.396703177135261e-07, | |
| "logits/chosen": 0.7419403791427612, | |
| "logits/rejected": 1.9202260971069336, | |
| "logps/chosen": -1948.1787109375, | |
| "logps/rejected": -2273.5205078125, | |
| "loss": 1.9864, | |
| "rewards/accuracies": 0.668749988079071, | |
| "rewards/chosen": -16.753948211669922, | |
| "rewards/margins": 3.6111111640930176, | |
| "rewards/rejected": -20.36505699157715, | |
| "step": 290 | |
| }, | |
| { | |
| "epoch": 0.3139717425431711, | |
| "grad_norm": 90.00202577382801, | |
| "learning_rate": 4.335883851539693e-07, | |
| "logits/chosen": 0.30849236249923706, | |
| "logits/rejected": 1.1072229146957397, | |
| "logps/chosen": -1431.3275146484375, | |
| "logps/rejected": -1653.4029541015625, | |
| "loss": 1.9069, | |
| "rewards/accuracies": 0.612500011920929, | |
| "rewards/chosen": -11.623054504394531, | |
| "rewards/margins": 2.2337088584899902, | |
| "rewards/rejected": -13.856762886047363, | |
| "step": 300 | |
| }, | |
| { | |
| "epoch": 0.3139717425431711, | |
| "eval_logits/chosen": 0.45899611711502075, | |
| "eval_logits/rejected": 1.5569082498550415, | |
| "eval_logps/chosen": -1266.6490478515625, | |
| "eval_logps/rejected": -1452.7674560546875, | |
| "eval_loss": 2.262396812438965, | |
| "eval_rewards/accuracies": 0.6329365372657776, | |
| "eval_rewards/chosen": -9.846811294555664, | |
| "eval_rewards/margins": 2.0787949562072754, | |
| "eval_rewards/rejected": -11.925606727600098, | |
| "eval_runtime": 176.5188, | |
| "eval_samples_per_second": 11.33, | |
| "eval_steps_per_second": 0.357, | |
| "step": 300 | |
| }, | |
| { | |
| "epoch": 0.32443746729461015, | |
| "grad_norm": 177.8700388917398, | |
| "learning_rate": 4.272609194017105e-07, | |
| "logits/chosen": 0.647371768951416, | |
| "logits/rejected": 2.9104599952697754, | |
| "logps/chosen": -1395.496826171875, | |
| "logps/rejected": -1711.9573974609375, | |
| "loss": 2.3095, | |
| "rewards/accuracies": 0.75, | |
| "rewards/chosen": -11.117349624633789, | |
| "rewards/margins": 3.674748182296753, | |
| "rewards/rejected": -14.792098999023438, | |
| "step": 310 | |
| }, | |
| { | |
| "epoch": 0.3349031920460492, | |
| "grad_norm": 180.92515200199898, | |
| "learning_rate": 4.2069638288135547e-07, | |
| "logits/chosen": 0.9543863534927368, | |
| "logits/rejected": 1.7447538375854492, | |
| "logps/chosen": -1926.299560546875, | |
| "logps/rejected": -2217.88037109375, | |
| "loss": 2.1724, | |
| "rewards/accuracies": 0.6312500238418579, | |
| "rewards/chosen": -16.73282814025879, | |
| "rewards/margins": 2.939984083175659, | |
| "rewards/rejected": -19.672813415527344, | |
| "step": 320 | |
| }, | |
| { | |
| "epoch": 0.3453689167974882, | |
| "grad_norm": 145.6894284610869, | |
| "learning_rate": 4.139035550786494e-07, | |
| "logits/chosen": -0.039321091026067734, | |
| "logits/rejected": 0.5018073320388794, | |
| "logps/chosen": -1734.091796875, | |
| "logps/rejected": -1908.339111328125, | |
| "loss": 1.716, | |
| "rewards/accuracies": 0.574999988079071, | |
| "rewards/chosen": -14.77747917175293, | |
| "rewards/margins": 1.9848415851593018, | |
| "rewards/rejected": -16.76232147216797, | |
| "step": 330 | |
| }, | |
| { | |
| "epoch": 0.35583464154892724, | |
| "grad_norm": 183.78050890033984, | |
| "learning_rate": 4.0689152079869306e-07, | |
| "logits/chosen": -0.5724295377731323, | |
| "logits/rejected": 0.023262571543455124, | |
| "logps/chosen": -1660.732177734375, | |
| "logps/rejected": -1876.1025390625, | |
| "loss": 1.8439, | |
| "rewards/accuracies": 0.606249988079071, | |
| "rewards/chosen": -13.912447929382324, | |
| "rewards/margins": 2.488671064376831, | |
| "rewards/rejected": -16.401119232177734, | |
| "step": 340 | |
| }, | |
| { | |
| "epoch": 0.3663003663003663, | |
| "grad_norm": 149.28700648360655, | |
| "learning_rate": 3.99669658015821e-07, | |
| "logits/chosen": 0.006322336383163929, | |
| "logits/rejected": 0.6332755088806152, | |
| "logps/chosen": -1966.5765380859375, | |
| "logps/rejected": -2201.843505859375, | |
| "loss": 1.6671, | |
| "rewards/accuracies": 0.6000000238418579, | |
| "rewards/chosen": -16.753414154052734, | |
| "rewards/margins": 2.7041499614715576, | |
| "rewards/rejected": -19.457565307617188, | |
| "step": 350 | |
| }, | |
| { | |
| "epoch": 0.37676609105180536, | |
| "grad_norm": 237.65668361495474, | |
| "learning_rate": 3.92247625331392e-07, | |
| "logits/chosen": -0.21500203013420105, | |
| "logits/rejected": 0.6255682110786438, | |
| "logps/chosen": -1989.7509765625, | |
| "logps/rejected": -2207.83349609375, | |
| "loss": 1.6927, | |
| "rewards/accuracies": 0.625, | |
| "rewards/chosen": -17.038667678833008, | |
| "rewards/margins": 2.4120330810546875, | |
| "rewards/rejected": -19.450698852539062, | |
| "step": 360 | |
| }, | |
| { | |
| "epoch": 0.3872318158032444, | |
| "grad_norm": 152.55773033990448, | |
| "learning_rate": 3.846353490562664e-07, | |
| "logits/chosen": -0.39199286699295044, | |
| "logits/rejected": -0.043508779257535934, | |
| "logps/chosen": -1889.5286865234375, | |
| "logps/rejected": -2139.589111328125, | |
| "loss": 1.7098, | |
| "rewards/accuracies": 0.59375, | |
| "rewards/chosen": -16.262523651123047, | |
| "rewards/margins": 2.7768733501434326, | |
| "rewards/rejected": -19.039398193359375, | |
| "step": 370 | |
| }, | |
| { | |
| "epoch": 0.3976975405546834, | |
| "grad_norm": 239.86422108427834, | |
| "learning_rate": 3.768430099352445e-07, | |
| "logits/chosen": -0.5338395833969116, | |
| "logits/rejected": -0.10323655605316162, | |
| "logps/chosen": -1830.7080078125, | |
| "logps/rejected": -2104.773681640625, | |
| "loss": 1.786, | |
| "rewards/accuracies": 0.6187499761581421, | |
| "rewards/chosen": -15.76060962677002, | |
| "rewards/margins": 2.8237688541412354, | |
| "rewards/rejected": -18.58437728881836, | |
| "step": 380 | |
| }, | |
| { | |
| "epoch": 0.40816326530612246, | |
| "grad_norm": 137.89263121746114, | |
| "learning_rate": 3.6888102953122304e-07, | |
| "logits/chosen": -0.3421451449394226, | |
| "logits/rejected": 0.2877078056335449, | |
| "logps/chosen": -1774.384765625, | |
| "logps/rejected": -2007.7366943359375, | |
| "loss": 1.9274, | |
| "rewards/accuracies": 0.574999988079071, | |
| "rewards/chosen": -14.92773151397705, | |
| "rewards/margins": 2.465951442718506, | |
| "rewards/rejected": -17.393680572509766, | |
| "step": 390 | |
| }, | |
| { | |
| "epoch": 0.4186289900575615, | |
| "grad_norm": 164.86784545063486, | |
| "learning_rate": 3.607600562872785e-07, | |
| "logits/chosen": -0.7335325479507446, | |
| "logits/rejected": -0.33919858932495117, | |
| "logps/chosen": -1733.375244140625, | |
| "logps/rejected": -1963.0279541015625, | |
| "loss": 1.6642, | |
| "rewards/accuracies": 0.5625, | |
| "rewards/chosen": -14.63383960723877, | |
| "rewards/margins": 2.319460391998291, | |
| "rewards/rejected": -16.95330047607422, | |
| "step": 400 | |
| }, | |
| { | |
| "epoch": 0.4186289900575615, | |
| "eval_logits/chosen": -0.7751028537750244, | |
| "eval_logits/rejected": -0.08748837560415268, | |
| "eval_logps/chosen": -1731.152587890625, | |
| "eval_logps/rejected": -2045.1492919921875, | |
| "eval_loss": 1.6421091556549072, | |
| "eval_rewards/accuracies": 0.625, | |
| "eval_rewards/chosen": -14.491846084594727, | |
| "eval_rewards/margins": 3.3575782775878906, | |
| "eval_rewards/rejected": -17.849422454833984, | |
| "eval_runtime": 176.0651, | |
| "eval_samples_per_second": 11.359, | |
| "eval_steps_per_second": 0.358, | |
| "step": 400 | |
| }, | |
| { | |
| "epoch": 0.4290947148090005, | |
| "grad_norm": 128.91689311765836, | |
| "learning_rate": 3.5249095128531856e-07, | |
| "logits/chosen": -0.10633065551519394, | |
| "logits/rejected": 0.350477933883667, | |
| "logps/chosen": -1862.1099853515625, | |
| "logps/rejected": -2067.15673828125, | |
| "loss": 1.7556, | |
| "rewards/accuracies": 0.59375, | |
| "rewards/chosen": -15.903741836547852, | |
| "rewards/margins": 2.2993245124816895, | |
| "rewards/rejected": -18.203065872192383, | |
| "step": 410 | |
| }, | |
| { | |
| "epoch": 0.43956043956043955, | |
| "grad_norm": 187.2282869549343, | |
| "learning_rate": 3.4408477372034736e-07, | |
| "logits/chosen": -0.2209610939025879, | |
| "logits/rejected": 0.7663095593452454, | |
| "logps/chosen": -1825.959228515625, | |
| "logps/rejected": -2182.580810546875, | |
| "loss": 1.8542, | |
| "rewards/accuracies": 0.6312500238418579, | |
| "rewards/chosen": -15.437875747680664, | |
| "rewards/margins": 3.867755174636841, | |
| "rewards/rejected": -19.30562973022461, | |
| "step": 420 | |
| }, | |
| { | |
| "epoch": 0.4500261643118786, | |
| "grad_norm": 150.13979068919696, | |
| "learning_rate": 3.3555276610977276e-07, | |
| "logits/chosen": -1.128701090812683, | |
| "logits/rejected": -0.5558885335922241, | |
| "logps/chosen": -1832.6103515625, | |
| "logps/rejected": -2176.197265625, | |
| "loss": 1.5079, | |
| "rewards/accuracies": 0.612500011920929, | |
| "rewards/chosen": -15.788568496704102, | |
| "rewards/margins": 3.47161602973938, | |
| "rewards/rejected": -19.26018714904785, | |
| "step": 430 | |
| }, | |
| { | |
| "epoch": 0.4604918890633176, | |
| "grad_norm": 163.41066719667168, | |
| "learning_rate": 3.269063392575352e-07, | |
| "logits/chosen": -0.6949409246444702, | |
| "logits/rejected": -0.05746125057339668, | |
| "logps/chosen": -1597.5341796875, | |
| "logps/rejected": -1821.0198974609375, | |
| "loss": 1.4868, | |
| "rewards/accuracies": 0.6312500238418579, | |
| "rewards/chosen": -13.023595809936523, | |
| "rewards/margins": 2.632272481918335, | |
| "rewards/rejected": -15.655868530273438, | |
| "step": 440 | |
| }, | |
| { | |
| "epoch": 0.47095761381475665, | |
| "grad_norm": 133.46596474594617, | |
| "learning_rate": 3.1815705699316964e-07, | |
| "logits/chosen": -0.4808398187160492, | |
| "logits/rejected": 0.3264926075935364, | |
| "logps/chosen": -1599.6370849609375, | |
| "logps/rejected": -1936.6884765625, | |
| "loss": 1.5413, | |
| "rewards/accuracies": 0.625, | |
| "rewards/chosen": -13.172491073608398, | |
| "rewards/margins": 3.4112179279327393, | |
| "rewards/rejected": -16.583707809448242, | |
| "step": 450 | |
| }, | |
| { | |
| "epoch": 0.48142333856619574, | |
| "grad_norm": 155.84007478164062, | |
| "learning_rate": 3.0931662070620794e-07, | |
| "logits/chosen": -0.719369113445282, | |
| "logits/rejected": -0.06152462959289551, | |
| "logps/chosen": -1643.2447509765625, | |
| "logps/rejected": -1872.9976806640625, | |
| "loss": 1.7906, | |
| "rewards/accuracies": 0.612500011920929, | |
| "rewards/chosen": -13.85786247253418, | |
| "rewards/margins": 2.4219117164611816, | |
| "rewards/rejected": -16.279773712158203, | |
| "step": 460 | |
| }, | |
| { | |
| "epoch": 0.49188906331763477, | |
| "grad_norm": 203.3322056694353, | |
| "learning_rate": 3.003968536966078e-07, | |
| "logits/chosen": -0.4609583020210266, | |
| "logits/rejected": -0.09374441206455231, | |
| "logps/chosen": -1654.1614990234375, | |
| "logps/rejected": -1845.5618896484375, | |
| "loss": 1.7718, | |
| "rewards/accuracies": 0.59375, | |
| "rewards/chosen": -13.703729629516602, | |
| "rewards/margins": 2.2525086402893066, | |
| "rewards/rejected": -15.956239700317383, | |
| "step": 470 | |
| }, | |
| { | |
| "epoch": 0.5023547880690737, | |
| "grad_norm": 156.4799546194198, | |
| "learning_rate": 2.9140968536213693e-07, | |
| "logits/chosen": -0.2353781908750534, | |
| "logits/rejected": 0.5946909785270691, | |
| "logps/chosen": -1859.3265380859375, | |
| "logps/rejected": -2325.88134765625, | |
| "loss": 1.3829, | |
| "rewards/accuracies": 0.6625000238418579, | |
| "rewards/chosen": -15.921140670776367, | |
| "rewards/margins": 4.824706077575684, | |
| "rewards/rejected": -20.745845794677734, | |
| "step": 480 | |
| }, | |
| { | |
| "epoch": 0.5128205128205128, | |
| "grad_norm": 160.19325879757844, | |
| "learning_rate": 2.823671352438608e-07, | |
| "logits/chosen": -0.9654836654663086, | |
| "logits/rejected": -0.002035105135291815, | |
| "logps/chosen": -1637.873291015625, | |
| "logps/rejected": -2143.010986328125, | |
| "loss": 1.6206, | |
| "rewards/accuracies": 0.7250000238418579, | |
| "rewards/chosen": -13.259417533874512, | |
| "rewards/margins": 5.606515407562256, | |
| "rewards/rejected": -18.86593246459961, | |
| "step": 490 | |
| }, | |
| { | |
| "epoch": 0.5232862375719518, | |
| "grad_norm": 221.83952267135834, | |
| "learning_rate": 2.73281296951072e-07, | |
| "logits/chosen": -0.6597784161567688, | |
| "logits/rejected": -0.14649493992328644, | |
| "logps/chosen": -1530.5738525390625, | |
| "logps/rejected": -1781.8070068359375, | |
| "loss": 1.6328, | |
| "rewards/accuracies": 0.6625000238418579, | |
| "rewards/chosen": -12.7192964553833, | |
| "rewards/margins": 2.8244967460632324, | |
| "rewards/rejected": -15.543792724609375, | |
| "step": 500 | |
| }, | |
| { | |
| "epoch": 0.5232862375719518, | |
| "eval_logits/chosen": -0.6590258479118347, | |
| "eval_logits/rejected": -0.091790109872818, | |
| "eval_logps/chosen": -1589.3370361328125, | |
| "eval_logps/rejected": -1890.562255859375, | |
| "eval_loss": 1.5119922161102295, | |
| "eval_rewards/accuracies": 0.6388888955116272, | |
| "eval_rewards/chosen": -13.073691368103027, | |
| "eval_rewards/margins": 3.229863166809082, | |
| "eval_rewards/rejected": -16.303556442260742, | |
| "eval_runtime": 177.8158, | |
| "eval_samples_per_second": 11.248, | |
| "eval_steps_per_second": 0.354, | |
| "step": 500 | |
| }, | |
| { | |
| "epoch": 0.533751962323391, | |
| "grad_norm": 187.4336485549293, | |
| "learning_rate": 2.641643219871597e-07, | |
| "logits/chosen": -0.5598984360694885, | |
| "logits/rejected": -0.2727218270301819, | |
| "logps/chosen": -1694.568359375, | |
| "logps/rejected": -2086.98193359375, | |
| "loss": 1.4069, | |
| "rewards/accuracies": 0.606249988079071, | |
| "rewards/chosen": -14.106300354003906, | |
| "rewards/margins": 4.125433921813965, | |
| "rewards/rejected": -18.231733322143555, | |
| "step": 510 | |
| }, | |
| { | |
| "epoch": 0.54421768707483, | |
| "grad_norm": 180.24950333654212, | |
| "learning_rate": 2.550284034980507e-07, | |
| "logits/chosen": -0.652435302734375, | |
| "logits/rejected": -0.25857192277908325, | |
| "logps/chosen": -1941.6849365234375, | |
| "logps/rejected": -2231.46337890625, | |
| "loss": 1.6022, | |
| "rewards/accuracies": 0.5874999761581421, | |
| "rewards/chosen": -16.69613265991211, | |
| "rewards/margins": 3.1190426349639893, | |
| "rewards/rejected": -19.815174102783203, | |
| "step": 520 | |
| }, | |
| { | |
| "epoch": 0.554683411826269, | |
| "grad_norm": 147.71519410172087, | |
| "learning_rate": 2.4588575996495794e-07, | |
| "logits/chosen": -0.6198351979255676, | |
| "logits/rejected": -0.19036616384983063, | |
| "logps/chosen": -1601.6470947265625, | |
| "logps/rejected": -1820.4556884765625, | |
| "loss": 1.5136, | |
| "rewards/accuracies": 0.5062500238418579, | |
| "rewards/chosen": -13.329002380371094, | |
| "rewards/margins": 2.3253164291381836, | |
| "rewards/rejected": -15.654316902160645, | |
| "step": 530 | |
| }, | |
| { | |
| "epoch": 0.565149136577708, | |
| "grad_norm": 146.6770433780799, | |
| "learning_rate": 2.367486188632446e-07, | |
| "logits/chosen": -0.7303057909011841, | |
| "logits/rejected": 0.15564236044883728, | |
| "logps/chosen": -1670.916015625, | |
| "logps/rejected": -2011.5406494140625, | |
| "loss": 1.5458, | |
| "rewards/accuracies": 0.625, | |
| "rewards/chosen": -13.805659294128418, | |
| "rewards/margins": 3.703829288482666, | |
| "rewards/rejected": -17.509489059448242, | |
| "step": 540 | |
| }, | |
| { | |
| "epoch": 0.5756148613291471, | |
| "grad_norm": 206.94359776232758, | |
| "learning_rate": 2.276292003092593e-07, | |
| "logits/chosen": -0.22513580322265625, | |
| "logits/rejected": 0.4895138740539551, | |
| "logps/chosen": -1914.7532958984375, | |
| "logps/rejected": -2300.30322265625, | |
| "loss": 1.6801, | |
| "rewards/accuracies": 0.6625000238418579, | |
| "rewards/chosen": -16.434810638427734, | |
| "rewards/margins": 4.275403022766113, | |
| "rewards/rejected": -20.71021270751953, | |
| "step": 550 | |
| }, | |
| { | |
| "epoch": 0.5860805860805861, | |
| "grad_norm": 175.41735239090949, | |
| "learning_rate": 2.185397007170141e-07, | |
| "logits/chosen": -0.1453290730714798, | |
| "logits/rejected": 0.3121495842933655, | |
| "logps/chosen": -1876.300537109375, | |
| "logps/rejected": -2229.38134765625, | |
| "loss": 1.3878, | |
| "rewards/accuracies": 0.6187499761581421, | |
| "rewards/chosen": -15.934832572937012, | |
| "rewards/margins": 3.8099570274353027, | |
| "rewards/rejected": -19.744789123535156, | |
| "step": 560 | |
| }, | |
| { | |
| "epoch": 0.5965463108320251, | |
| "grad_norm": 142.79294258337345, | |
| "learning_rate": 2.094922764865619e-07, | |
| "logits/chosen": -0.276650995016098, | |
| "logits/rejected": 0.13945253193378448, | |
| "logps/chosen": -1827.0634765625, | |
| "logps/rejected": -2034.280517578125, | |
| "loss": 1.4902, | |
| "rewards/accuracies": 0.637499988079071, | |
| "rewards/chosen": -15.475687980651855, | |
| "rewards/margins": 2.2050392627716064, | |
| "rewards/rejected": -17.680728912353516, | |
| "step": 570 | |
| }, | |
| { | |
| "epoch": 0.6070120355834642, | |
| "grad_norm": 245.80968468908674, | |
| "learning_rate": 2.0049902774588797e-07, | |
| "logits/chosen": -0.011815989390015602, | |
| "logits/rejected": 0.42436084151268005, | |
| "logps/chosen": -1794.5543212890625, | |
| "logps/rejected": -2061.93310546875, | |
| "loss": 1.4461, | |
| "rewards/accuracies": 0.643750011920929, | |
| "rewards/chosen": -15.412150382995605, | |
| "rewards/margins": 3.078895330429077, | |
| "rewards/rejected": -18.491044998168945, | |
| "step": 580 | |
| }, | |
| { | |
| "epoch": 0.6174777603349032, | |
| "grad_norm": 175.38280547329734, | |
| "learning_rate": 1.9157198216806238e-07, | |
| "logits/chosen": -0.3044319152832031, | |
| "logits/rejected": 0.3406422734260559, | |
| "logps/chosen": -1649.8509521484375, | |
| "logps/rejected": -2006.366455078125, | |
| "loss": 1.5446, | |
| "rewards/accuracies": 0.6937500238418579, | |
| "rewards/chosen": -13.71898365020752, | |
| "rewards/margins": 3.657163143157959, | |
| "rewards/rejected": -17.376148223876953, | |
| "step": 590 | |
| }, | |
| { | |
| "epoch": 0.6279434850863422, | |
| "grad_norm": 203.04339818262545, | |
| "learning_rate": 1.8272307888529274e-07, | |
| "logits/chosen": 0.16477735340595245, | |
| "logits/rejected": 0.6171606183052063, | |
| "logps/chosen": -1870.41015625, | |
| "logps/rejected": -2165.638427734375, | |
| "loss": 1.6032, | |
| "rewards/accuracies": 0.5687500238418579, | |
| "rewards/chosen": -15.878863334655762, | |
| "rewards/margins": 3.1766743659973145, | |
| "rewards/rejected": -19.055538177490234, | |
| "step": 600 | |
| }, | |
| { | |
| "epoch": 0.6279434850863422, | |
| "eval_logits/chosen": 0.01903720200061798, | |
| "eval_logits/rejected": 0.6402472853660583, | |
| "eval_logps/chosen": -2015.7071533203125, | |
| "eval_logps/rejected": -2402.58447265625, | |
| "eval_loss": 1.4751698970794678, | |
| "eval_rewards/accuracies": 0.6230158805847168, | |
| "eval_rewards/chosen": -17.33738899230957, | |
| "eval_rewards/margins": 4.086385250091553, | |
| "eval_rewards/rejected": -21.42377471923828, | |
| "eval_runtime": 176.4506, | |
| "eval_samples_per_second": 11.335, | |
| "eval_steps_per_second": 0.357, | |
| "step": 600 | |
| }, | |
| { | |
| "epoch": 0.6384092098377813, | |
| "grad_norm": 184.64896406440843, | |
| "learning_rate": 1.7396415252139288e-07, | |
| "logits/chosen": 0.0034784465096890926, | |
| "logits/rejected": 0.6044633388519287, | |
| "logps/chosen": -2050.113037109375, | |
| "logps/rejected": -2622.564453125, | |
| "loss": 1.5229, | |
| "rewards/accuracies": 0.6812499761581421, | |
| "rewards/chosen": -17.44542694091797, | |
| "rewards/margins": 6.333140850067139, | |
| "rewards/rejected": -23.778566360473633, | |
| "step": 610 | |
| }, | |
| { | |
| "epoch": 0.6488749345892203, | |
| "grad_norm": 150.92780625161797, | |
| "learning_rate": 1.6530691736402316e-07, | |
| "logits/chosen": -0.05873150750994682, | |
| "logits/rejected": 0.2572210133075714, | |
| "logps/chosen": -1822.690185546875, | |
| "logps/rejected": -2140.002685546875, | |
| "loss": 1.3047, | |
| "rewards/accuracies": 0.668749988079071, | |
| "rewards/chosen": -15.516670227050781, | |
| "rewards/margins": 3.519291400909424, | |
| "rewards/rejected": -19.035961151123047, | |
| "step": 620 | |
| }, | |
| { | |
| "epoch": 0.6593406593406593, | |
| "grad_norm": 158.62413320623054, | |
| "learning_rate": 1.5676295169786864e-07, | |
| "logits/chosen": -0.5535549521446228, | |
| "logits/rejected": -0.16974008083343506, | |
| "logps/chosen": -1799.411376953125, | |
| "logps/rejected": -2184.095458984375, | |
| "loss": 1.4004, | |
| "rewards/accuracies": 0.643750011920929, | |
| "rewards/chosen": -15.247261047363281, | |
| "rewards/margins": 4.233900547027588, | |
| "rewards/rejected": -19.481159210205078, | |
| "step": 630 | |
| }, | |
| { | |
| "epoch": 0.6698063840920984, | |
| "grad_norm": 174.63990723873954, | |
| "learning_rate": 1.483436823197092e-07, | |
| "logits/chosen": -0.49727511405944824, | |
| "logits/rejected": -0.09024439752101898, | |
| "logps/chosen": -1910.181396484375, | |
| "logps/rejected": -2272.175537109375, | |
| "loss": 1.2582, | |
| "rewards/accuracies": 0.668749988079071, | |
| "rewards/chosen": -16.431602478027344, | |
| "rewards/margins": 4.003415107727051, | |
| "rewards/rejected": -20.43501853942871, | |
| "step": 640 | |
| }, | |
| { | |
| "epoch": 0.6802721088435374, | |
| "grad_norm": 212.30897956956616, | |
| "learning_rate": 1.4006036925609243e-07, | |
| "logits/chosen": -0.5441917777061462, | |
| "logits/rejected": -0.3759006857872009, | |
| "logps/chosen": -1762.1038818359375, | |
| "logps/rejected": -1993.853515625, | |
| "loss": 1.3183, | |
| "rewards/accuracies": 0.574999988079071, | |
| "rewards/chosen": -14.94421100616455, | |
| "rewards/margins": 2.2772457599639893, | |
| "rewards/rejected": -17.221454620361328, | |
| "step": 650 | |
| }, | |
| { | |
| "epoch": 0.6907378335949764, | |
| "grad_norm": 122.40725726992933, | |
| "learning_rate": 1.319240907040458e-07, | |
| "logits/chosen": -0.578727126121521, | |
| "logits/rejected": -0.15290720760822296, | |
| "logps/chosen": -1786.3648681640625, | |
| "logps/rejected": -2103.92919921875, | |
| "loss": 1.5482, | |
| "rewards/accuracies": 0.6312500238418579, | |
| "rewards/chosen": -15.042015075683594, | |
| "rewards/margins": 3.422727584838867, | |
| "rewards/rejected": -18.46474266052246, | |
| "step": 660 | |
| }, | |
| { | |
| "epoch": 0.7012035583464155, | |
| "grad_norm": 273.40146184819037, | |
| "learning_rate": 1.239457282149695e-07, | |
| "logits/chosen": -0.6542818546295166, | |
| "logits/rejected": -0.6405806541442871, | |
| "logps/chosen": -1718.8697509765625, | |
| "logps/rejected": -2025.167236328125, | |
| "loss": 1.1528, | |
| "rewards/accuracies": 0.5874999761581421, | |
| "rewards/chosen": -14.303213119506836, | |
| "rewards/margins": 3.1175124645233154, | |
| "rewards/rejected": -17.420726776123047, | |
| "step": 670 | |
| }, | |
| { | |
| "epoch": 0.7116692830978545, | |
| "grad_norm": 246.28508875936285, | |
| "learning_rate": 1.1613595214152711e-07, | |
| "logits/chosen": -0.6755629777908325, | |
| "logits/rejected": -0.26193898916244507, | |
| "logps/chosen": -1862.4000244140625, | |
| "logps/rejected": -2191.969482421875, | |
| "loss": 1.3671, | |
| "rewards/accuracies": 0.612500011920929, | |
| "rewards/chosen": -15.507433891296387, | |
| "rewards/margins": 3.5333023071289062, | |
| "rewards/rejected": -19.04073715209961, | |
| "step": 680 | |
| }, | |
| { | |
| "epoch": 0.7221350078492935, | |
| "grad_norm": 216.14843384209277, | |
| "learning_rate": 1.0850520736699362e-07, | |
| "logits/chosen": -0.6002136468887329, | |
| "logits/rejected": 0.03606845811009407, | |
| "logps/chosen": -1838.6982421875, | |
| "logps/rejected": -2214.07470703125, | |
| "loss": 1.3895, | |
| "rewards/accuracies": 0.7250000238418579, | |
| "rewards/chosen": -15.680435180664062, | |
| "rewards/margins": 4.133326053619385, | |
| "rewards/rejected": -19.813762664794922, | |
| "step": 690 | |
| }, | |
| { | |
| "epoch": 0.7326007326007326, | |
| "grad_norm": 162.01079027631573, | |
| "learning_rate": 1.0106369933615042e-07, | |
| "logits/chosen": -0.7846351861953735, | |
| "logits/rejected": -0.5915166735649109, | |
| "logps/chosen": -1752.784423828125, | |
| "logps/rejected": -2021.7802734375, | |
| "loss": 1.5039, | |
| "rewards/accuracies": 0.581250011920929, | |
| "rewards/chosen": -14.986889839172363, | |
| "rewards/margins": 2.845428943634033, | |
| "rewards/rejected": -17.832317352294922, | |
| "step": 700 | |
| }, | |
| { | |
| "epoch": 0.7326007326007326, | |
| "eval_logits/chosen": -0.8898468017578125, | |
| "eval_logits/rejected": -0.4967605769634247, | |
| "eval_logps/chosen": -1694.96240234375, | |
| "eval_logps/rejected": -2016.4490966796875, | |
| "eval_loss": 1.3852962255477905, | |
| "eval_rewards/accuracies": 0.6527777910232544, | |
| "eval_rewards/chosen": -14.129942893981934, | |
| "eval_rewards/margins": 3.432478666305542, | |
| "eval_rewards/rejected": -17.562421798706055, | |
| "eval_runtime": 176.0679, | |
| "eval_samples_per_second": 11.359, | |
| "eval_steps_per_second": 0.358, | |
| "step": 700 | |
| }, | |
| { | |
| "epoch": 0.7430664573521716, | |
| "grad_norm": 177.45761000957364, | |
| "learning_rate": 9.382138040640714e-08, | |
| "logits/chosen": -1.012629747390747, | |
| "logits/rejected": -0.6268833875656128, | |
| "logps/chosen": -1776.499755859375, | |
| "logps/rejected": -2017.539794921875, | |
| "loss": 1.5264, | |
| "rewards/accuracies": 0.637499988079071, | |
| "rewards/chosen": -15.180249214172363, | |
| "rewards/margins": 2.565770387649536, | |
| "rewards/rejected": -17.74601936340332, | |
| "step": 710 | |
| }, | |
| { | |
| "epoch": 0.7535321821036107, | |
| "grad_norm": 140.94359920373847, | |
| "learning_rate": 8.678793653740632e-08, | |
| "logits/chosen": -0.9271895289421082, | |
| "logits/rejected": -0.47789469361305237, | |
| "logps/chosen": -1664.4437255859375, | |
| "logps/rejected": -1977.908447265625, | |
| "loss": 1.3295, | |
| "rewards/accuracies": 0.6812499761581421, | |
| "rewards/chosen": -13.8070650100708, | |
| "rewards/margins": 3.4972636699676514, | |
| "rewards/rejected": -17.3043270111084, | |
| "step": 720 | |
| }, | |
| { | |
| "epoch": 0.7639979068550498, | |
| "grad_norm": 190.75937551525504, | |
| "learning_rate": 7.997277433690983e-08, | |
| "logits/chosen": -0.8303499221801758, | |
| "logits/rejected": -0.2948521077632904, | |
| "logps/chosen": -1813.2340087890625, | |
| "logps/rejected": -2049.240234375, | |
| "loss": 1.4631, | |
| "rewards/accuracies": 0.574999988079071, | |
| "rewards/chosen": -15.478715896606445, | |
| "rewards/margins": 2.6438088417053223, | |
| "rewards/rejected": -18.12252426147461, | |
| "step": 730 | |
| }, | |
| { | |
| "epoch": 0.7744636316064888, | |
| "grad_norm": 164.74206538760382, | |
| "learning_rate": 7.338500848029602e-08, | |
| "logits/chosen": -0.6835179924964905, | |
| "logits/rejected": -0.42263850569725037, | |
| "logps/chosen": -1808.6490478515625, | |
| "logps/rejected": -2096.81396484375, | |
| "loss": 1.2242, | |
| "rewards/accuracies": 0.6312500238418579, | |
| "rewards/chosen": -15.557826042175293, | |
| "rewards/margins": 3.0879101753234863, | |
| "rewards/rejected": -18.645736694335938, | |
| "step": 740 | |
| }, | |
| { | |
| "epoch": 0.7849293563579278, | |
| "grad_norm": 135.0757551116068, | |
| "learning_rate": 6.70334495204884e-08, | |
| "logits/chosen": -0.5583680868148804, | |
| "logits/rejected": -0.36530551314353943, | |
| "logps/chosen": -1854.912353515625, | |
| "logps/rejected": -2177.262451171875, | |
| "loss": 1.3344, | |
| "rewards/accuracies": 0.5562499761581421, | |
| "rewards/chosen": -16.08046531677246, | |
| "rewards/margins": 3.2556281089782715, | |
| "rewards/rejected": -19.33609390258789, | |
| "step": 750 | |
| }, | |
| { | |
| "epoch": 0.7953950811093669, | |
| "grad_norm": 158.01784405358154, | |
| "learning_rate": 6.092659210462231e-08, | |
| "logits/chosen": -0.653573215007782, | |
| "logits/rejected": -0.4876467287540436, | |
| "logps/chosen": -1903.880615234375, | |
| "logps/rejected": -2182.48291015625, | |
| "loss": 1.4038, | |
| "rewards/accuracies": 0.5625, | |
| "rewards/chosen": -16.625337600708008, | |
| "rewards/margins": 2.7693800926208496, | |
| "rewards/rejected": -19.394718170166016, | |
| "step": 760 | |
| }, | |
| { | |
| "epoch": 0.8058608058608059, | |
| "grad_norm": 169.97964049682443, | |
| "learning_rate": 5.507260361320737e-08, | |
| "logits/chosen": -1.0366981029510498, | |
| "logits/rejected": -0.9037246704101562, | |
| "logps/chosen": -1879.755126953125, | |
| "logps/rejected": -2001.697265625, | |
| "loss": 1.286, | |
| "rewards/accuracies": 0.5375000238418579, | |
| "rewards/chosen": -15.764042854309082, | |
| "rewards/margins": 1.033178687095642, | |
| "rewards/rejected": -16.797222137451172, | |
| "step": 770 | |
| }, | |
| { | |
| "epoch": 0.8163265306122449, | |
| "grad_norm": 162.02338031146334, | |
| "learning_rate": 4.947931323697982e-08, | |
| "logits/chosen": -1.0304605960845947, | |
| "logits/rejected": -0.9400796890258789, | |
| "logps/chosen": -1669.2073974609375, | |
| "logps/rejected": -2004.525390625, | |
| "loss": 1.5927, | |
| "rewards/accuracies": 0.643750011920929, | |
| "rewards/chosen": -13.502462387084961, | |
| "rewards/margins": 3.8671538829803467, | |
| "rewards/rejected": -17.369617462158203, | |
| "step": 780 | |
| }, | |
| { | |
| "epoch": 0.826792255363684, | |
| "grad_norm": 140.7368428333841, | |
| "learning_rate": 4.415420150605398e-08, | |
| "logits/chosen": -1.0811887979507446, | |
| "logits/rejected": -0.5253428220748901, | |
| "logps/chosen": -1726.182373046875, | |
| "logps/rejected": -2063.27099609375, | |
| "loss": 1.4648, | |
| "rewards/accuracies": 0.581250011920929, | |
| "rewards/chosen": -14.476069450378418, | |
| "rewards/margins": 3.7039875984191895, | |
| "rewards/rejected": -18.180057525634766, | |
| "step": 790 | |
| }, | |
| { | |
| "epoch": 0.837257980115123, | |
| "grad_norm": 202.82775780509928, | |
| "learning_rate": 3.9104390285376374e-08, | |
| "logits/chosen": -0.835501492023468, | |
| "logits/rejected": -0.5900505781173706, | |
| "logps/chosen": -1749.853759765625, | |
| "logps/rejected": -1951.329345703125, | |
| "loss": 1.3527, | |
| "rewards/accuracies": 0.5874999761581421, | |
| "rewards/chosen": -14.83747673034668, | |
| "rewards/margins": 2.0425891876220703, | |
| "rewards/rejected": -16.88006591796875, | |
| "step": 800 | |
| }, | |
| { | |
| "epoch": 0.837257980115123, | |
| "eval_logits/chosen": -1.0374784469604492, | |
| "eval_logits/rejected": -0.6750361919403076, | |
| "eval_logps/chosen": -1672.130615234375, | |
| "eval_logps/rejected": -1986.035888671875, | |
| "eval_loss": 1.366306185722351, | |
| "eval_rewards/accuracies": 0.6448412537574768, | |
| "eval_rewards/chosen": -13.901623725891113, | |
| "eval_rewards/margins": 3.3566668033599854, | |
| "eval_rewards/rejected": -17.25829315185547, | |
| "eval_runtime": 176.0547, | |
| "eval_samples_per_second": 11.36, | |
| "eval_steps_per_second": 0.358, | |
| "step": 800 | |
| }, | |
| { | |
| "epoch": 0.847723704866562, | |
| "grad_norm": 218.80895490878117, | |
| "learning_rate": 3.433663324986208e-08, | |
| "logits/chosen": -1.2597501277923584, | |
| "logits/rejected": -0.7243804931640625, | |
| "logps/chosen": -1665.3489990234375, | |
| "logps/rejected": -2045.541259765625, | |
| "loss": 1.4186, | |
| "rewards/accuracies": 0.706250011920929, | |
| "rewards/chosen": -13.741106986999512, | |
| "rewards/margins": 4.249786853790283, | |
| "rewards/rejected": -17.990894317626953, | |
| "step": 810 | |
| }, | |
| { | |
| "epoch": 0.858189429618001, | |
| "grad_norm": 175.67069527310957, | |
| "learning_rate": 2.9857306851953897e-08, | |
| "logits/chosen": -1.075448751449585, | |
| "logits/rejected": -0.8459098935127258, | |
| "logps/chosen": -1705.802734375, | |
| "logps/rejected": -1971.207275390625, | |
| "loss": 1.1819, | |
| "rewards/accuracies": 0.606249988079071, | |
| "rewards/chosen": -14.010282516479492, | |
| "rewards/margins": 2.8707687854766846, | |
| "rewards/rejected": -16.881052017211914, | |
| "step": 820 | |
| }, | |
| { | |
| "epoch": 0.8686551543694401, | |
| "grad_norm": 150.14969837730865, | |
| "learning_rate": 2.567240179368185e-08, | |
| "logits/chosen": -0.8211779594421387, | |
| "logits/rejected": -0.672277569770813, | |
| "logps/chosen": -1724.1959228515625, | |
| "logps/rejected": -1975.289306640625, | |
| "loss": 1.3771, | |
| "rewards/accuracies": 0.6187499761581421, | |
| "rewards/chosen": -14.66050910949707, | |
| "rewards/margins": 2.5944151878356934, | |
| "rewards/rejected": -17.25492286682129, | |
| "step": 830 | |
| }, | |
| { | |
| "epoch": 0.8791208791208791, | |
| "grad_norm": 143.51050018041488, | |
| "learning_rate": 2.1787515014630357e-08, | |
| "logits/chosen": -0.9592329263687134, | |
| "logits/rejected": -0.6304475665092468, | |
| "logps/chosen": -1664.050048828125, | |
| "logps/rejected": -2090.85107421875, | |
| "loss": 1.1841, | |
| "rewards/accuracies": 0.643750011920929, | |
| "rewards/chosen": -13.817761421203613, | |
| "rewards/margins": 4.430028915405273, | |
| "rewards/rejected": -18.247791290283203, | |
| "step": 840 | |
| }, | |
| { | |
| "epoch": 0.8895866038723181, | |
| "grad_norm": 160.09590738302992, | |
| "learning_rate": 1.820784220652766e-08, | |
| "logits/chosen": -0.8976573944091797, | |
| "logits/rejected": -0.619744598865509, | |
| "logps/chosen": -1732.6185302734375, | |
| "logps/rejected": -2009.6126708984375, | |
| "loss": 1.3946, | |
| "rewards/accuracies": 0.612500011920929, | |
| "rewards/chosen": -14.51783561706543, | |
| "rewards/margins": 3.2437214851379395, | |
| "rewards/rejected": -17.761554718017578, | |
| "step": 850 | |
| }, | |
| { | |
| "epoch": 0.9000523286237572, | |
| "grad_norm": 140.45079725700174, | |
| "learning_rate": 1.4938170864468636e-08, | |
| "logits/chosen": -1.2183126211166382, | |
| "logits/rejected": -0.7451462149620056, | |
| "logps/chosen": -1663.8861083984375, | |
| "logps/rejected": -2030.5501708984375, | |
| "loss": 1.403, | |
| "rewards/accuracies": 0.6625000238418579, | |
| "rewards/chosen": -13.755941390991211, | |
| "rewards/margins": 3.9360270500183105, | |
| "rewards/rejected": -17.691970825195312, | |
| "step": 860 | |
| }, | |
| { | |
| "epoch": 0.9105180533751962, | |
| "grad_norm": 177.87764974909854, | |
| "learning_rate": 1.1982873884064465e-08, | |
| "logits/chosen": -1.142114281654358, | |
| "logits/rejected": -0.8570957183837891, | |
| "logps/chosen": -1702.1165771484375, | |
| "logps/rejected": -2053.07568359375, | |
| "loss": 1.364, | |
| "rewards/accuracies": 0.6499999761581421, | |
| "rewards/chosen": -14.30346393585205, | |
| "rewards/margins": 3.762028932571411, | |
| "rewards/rejected": -18.06549072265625, | |
| "step": 870 | |
| }, | |
| { | |
| "epoch": 0.9209837781266352, | |
| "grad_norm": 138.3301348415624, | |
| "learning_rate": 9.345903713082304e-09, | |
| "logits/chosen": -1.0760080814361572, | |
| "logits/rejected": -0.866096019744873, | |
| "logps/chosen": -1735.3382568359375, | |
| "logps/rejected": -2023.660888671875, | |
| "loss": 1.355, | |
| "rewards/accuracies": 0.5625, | |
| "rewards/chosen": -14.312044143676758, | |
| "rewards/margins": 2.978205680847168, | |
| "rewards/rejected": -17.290246963500977, | |
| "step": 880 | |
| }, | |
| { | |
| "epoch": 0.9314495028780743, | |
| "grad_norm": 179.16273994251034, | |
| "learning_rate": 7.030787065396865e-09, | |
| "logits/chosen": -1.0234577655792236, | |
| "logits/rejected": -0.9720734357833862, | |
| "logps/chosen": -1736.5269775390625, | |
| "logps/rejected": -2083.37939453125, | |
| "loss": 1.4332, | |
| "rewards/accuracies": 0.581250011920929, | |
| "rewards/chosen": -14.488537788391113, | |
| "rewards/margins": 3.543290376663208, | |
| "rewards/rejected": -18.031827926635742, | |
| "step": 890 | |
| }, | |
| { | |
| "epoch": 0.9419152276295133, | |
| "grad_norm": 163.4835379161221, | |
| "learning_rate": 5.04062020432286e-09, | |
| "logits/chosen": -0.8189510107040405, | |
| "logits/rejected": -0.8584410548210144, | |
| "logps/chosen": -1706.8818359375, | |
| "logps/rejected": -1968.8441162109375, | |
| "loss": 1.5137, | |
| "rewards/accuracies": 0.6000000238418579, | |
| "rewards/chosen": -14.264904975891113, | |
| "rewards/margins": 2.609503984451294, | |
| "rewards/rejected": -16.874408721923828, | |
| "step": 900 | |
| }, | |
| { | |
| "epoch": 0.9419152276295133, | |
| "eval_logits/chosen": -1.001752495765686, | |
| "eval_logits/rejected": -0.673967182636261, | |
| "eval_logps/chosen": -1735.9151611328125, | |
| "eval_logps/rejected": -2073.3388671875, | |
| "eval_loss": 1.3373700380325317, | |
| "eval_rewards/accuracies": 0.6408730149269104, | |
| "eval_rewards/chosen": -14.539473533630371, | |
| "eval_rewards/margins": 3.5918467044830322, | |
| "eval_rewards/rejected": -18.13132095336914, | |
| "eval_runtime": 176.3334, | |
| "eval_samples_per_second": 11.342, | |
| "eval_steps_per_second": 0.357, | |
| "step": 900 | |
| }, | |
| { | |
| "epoch": 0.9523809523809523, | |
| "grad_norm": 190.32378571700949, | |
| "learning_rate": 3.3780648016376866e-09, | |
| "logits/chosen": -0.9321626424789429, | |
| "logits/rejected": -0.5902298092842102, | |
| "logps/chosen": -1696.779296875, | |
| "logps/rejected": -1922.1607666015625, | |
| "loss": 1.4578, | |
| "rewards/accuracies": 0.625, | |
| "rewards/chosen": -14.7335786819458, | |
| "rewards/margins": 2.3592441082000732, | |
| "rewards/rejected": -17.092823028564453, | |
| "step": 910 | |
| }, | |
| { | |
| "epoch": 0.9628466771323915, | |
| "grad_norm": 183.98567167006505, | |
| "learning_rate": 2.0453443778310766e-09, | |
| "logits/chosen": -1.0600922107696533, | |
| "logits/rejected": -0.7931039929389954, | |
| "logps/chosen": -1763.392822265625, | |
| "logps/rejected": -2107.805419921875, | |
| "loss": 1.3202, | |
| "rewards/accuracies": 0.6312500238418579, | |
| "rewards/chosen": -14.767558097839355, | |
| "rewards/margins": 3.794466495513916, | |
| "rewards/rejected": -18.562023162841797, | |
| "step": 920 | |
| }, | |
| { | |
| "epoch": 0.9733124018838305, | |
| "grad_norm": 181.56437725274117, | |
| "learning_rate": 1.0442413283435758e-09, | |
| "logits/chosen": -1.1890182495117188, | |
| "logits/rejected": -0.5295430421829224, | |
| "logps/chosen": -1729.0921630859375, | |
| "logps/rejected": -1985.2783203125, | |
| "loss": 1.5669, | |
| "rewards/accuracies": 0.643750011920929, | |
| "rewards/chosen": -14.14258098602295, | |
| "rewards/margins": 3.068037748336792, | |
| "rewards/rejected": -17.210617065429688, | |
| "step": 930 | |
| }, | |
| { | |
| "epoch": 0.9837781266352695, | |
| "grad_norm": 173.28786175289625, | |
| "learning_rate": 3.760945397705828e-10, | |
| "logits/chosen": -0.856045126914978, | |
| "logits/rejected": -0.7398639917373657, | |
| "logps/chosen": -1713.3883056640625, | |
| "logps/rejected": -2039.740966796875, | |
| "loss": 1.266, | |
| "rewards/accuracies": 0.6000000238418579, | |
| "rewards/chosen": -14.502069473266602, | |
| "rewards/margins": 3.2981293201446533, | |
| "rewards/rejected": -17.800199508666992, | |
| "step": 940 | |
| }, | |
| { | |
| "epoch": 0.9942438513867086, | |
| "grad_norm": 188.65879146663107, | |
| "learning_rate": 4.17975992204056e-11, | |
| "logits/chosen": -1.168084740638733, | |
| "logits/rejected": -0.8855546116828918, | |
| "logps/chosen": -1736.102783203125, | |
| "logps/rejected": -1955.3255615234375, | |
| "loss": 1.4604, | |
| "rewards/accuracies": 0.612500011920929, | |
| "rewards/chosen": -14.33259391784668, | |
| "rewards/margins": 2.370856761932373, | |
| "rewards/rejected": -16.70345115661621, | |
| "step": 950 | |
| }, | |
| { | |
| "epoch": 0.9994767137624281, | |
| "step": 955, | |
| "total_flos": 0.0, | |
| "train_loss": 2.1165736393154604, | |
| "train_runtime": 18133.1885, | |
| "train_samples_per_second": 3.371, | |
| "train_steps_per_second": 0.053 | |
| } | |
| ], | |
| "logging_steps": 10, | |
| "max_steps": 955, | |
| "num_input_tokens_seen": 0, | |
| "num_train_epochs": 1, | |
| "save_steps": 100, | |
| "total_flos": 0.0, | |
| "train_batch_size": 8, | |
| "trial_name": null, | |
| "trial_params": null | |
| } | |