| { | |
| "best_metric": null, | |
| "best_model_checkpoint": null, | |
| "epoch": 0.999508116084604, | |
| "eval_steps": 100, | |
| "global_step": 1016, | |
| "is_hyper_param_search": false, | |
| "is_local_process_zero": true, | |
| "is_world_process_zero": true, | |
| "log_history": [ | |
| { | |
| "epoch": 0.0, | |
| "learning_rate": 4.901960784313726e-08, | |
| "logits/chosen": -2.493000030517578, | |
| "logits/rejected": -2.5710701942443848, | |
| "logps/chosen": -136.35806274414062, | |
| "logps/rejected": -147.18719482421875, | |
| "loss": 0.6931, | |
| "rewards/accuracies": 0.0, | |
| "rewards/chosen": 0.0, | |
| "rewards/margins": 0.0, | |
| "rewards/rejected": 0.0, | |
| "step": 1 | |
| }, | |
| { | |
| "epoch": 0.01, | |
| "learning_rate": 4.901960784313725e-07, | |
| "logits/chosen": -2.5276925563812256, | |
| "logits/rejected": -2.4688756465911865, | |
| "logps/chosen": -160.81719970703125, | |
| "logps/rejected": -124.57797241210938, | |
| "loss": 0.6933, | |
| "rewards/accuracies": 0.3888888955116272, | |
| "rewards/chosen": -0.00038504067924804986, | |
| "rewards/margins": -0.0003390010679140687, | |
| "rewards/rejected": -4.6039687731536105e-05, | |
| "step": 10 | |
| }, | |
| { | |
| "epoch": 0.02, | |
| "learning_rate": 9.80392156862745e-07, | |
| "logits/chosen": -2.5096638202667236, | |
| "logits/rejected": -2.4554123878479004, | |
| "logps/chosen": -169.54324340820312, | |
| "logps/rejected": -142.311767578125, | |
| "loss": 0.6933, | |
| "rewards/accuracies": 0.4375, | |
| "rewards/chosen": -0.00011383053060853854, | |
| "rewards/margins": -0.00031454654526896775, | |
| "rewards/rejected": 0.0002007160073844716, | |
| "step": 20 | |
| }, | |
| { | |
| "epoch": 0.03, | |
| "learning_rate": 1.4705882352941177e-06, | |
| "logits/chosen": -2.5936665534973145, | |
| "logits/rejected": -2.5427086353302, | |
| "logps/chosen": -158.86428833007812, | |
| "logps/rejected": -142.90164184570312, | |
| "loss": 0.6932, | |
| "rewards/accuracies": 0.5375000238418579, | |
| "rewards/chosen": 0.0005116682150401175, | |
| "rewards/margins": -0.0001444850640837103, | |
| "rewards/rejected": 0.0006561531918123364, | |
| "step": 30 | |
| }, | |
| { | |
| "epoch": 0.04, | |
| "learning_rate": 1.96078431372549e-06, | |
| "logits/chosen": -2.5783634185791016, | |
| "logits/rejected": -2.514772415161133, | |
| "logps/chosen": -163.1798858642578, | |
| "logps/rejected": -140.1341094970703, | |
| "loss": 0.6928, | |
| "rewards/accuracies": 0.5874999761581421, | |
| "rewards/chosen": 0.0009201603243127465, | |
| "rewards/margins": 0.0006632919539697468, | |
| "rewards/rejected": 0.0002568683703429997, | |
| "step": 40 | |
| }, | |
| { | |
| "epoch": 0.05, | |
| "learning_rate": 2.450980392156863e-06, | |
| "logits/chosen": -2.593667984008789, | |
| "logits/rejected": -2.540072202682495, | |
| "logps/chosen": -162.23658752441406, | |
| "logps/rejected": -134.82162475585938, | |
| "loss": 0.6925, | |
| "rewards/accuracies": 0.6499999761581421, | |
| "rewards/chosen": 0.0019953851588070393, | |
| "rewards/margins": 0.0013176979264244437, | |
| "rewards/rejected": 0.0006776872323825955, | |
| "step": 50 | |
| }, | |
| { | |
| "epoch": 0.06, | |
| "learning_rate": 2.9411764705882355e-06, | |
| "logits/chosen": -2.5982372760772705, | |
| "logits/rejected": -2.565171718597412, | |
| "logps/chosen": -167.44467163085938, | |
| "logps/rejected": -147.24746704101562, | |
| "loss": 0.6927, | |
| "rewards/accuracies": 0.612500011920929, | |
| "rewards/chosen": 0.0029606518801301718, | |
| "rewards/margins": 0.0008354202145710588, | |
| "rewards/rejected": 0.002125231781974435, | |
| "step": 60 | |
| }, | |
| { | |
| "epoch": 0.07, | |
| "learning_rate": 3.431372549019608e-06, | |
| "logits/chosen": -2.6009891033172607, | |
| "logits/rejected": -2.564056873321533, | |
| "logps/chosen": -155.53407287597656, | |
| "logps/rejected": -126.87095642089844, | |
| "loss": 0.6924, | |
| "rewards/accuracies": 0.625, | |
| "rewards/chosen": 0.003734055906534195, | |
| "rewards/margins": 0.0015770606696605682, | |
| "rewards/rejected": 0.0021569952368736267, | |
| "step": 70 | |
| }, | |
| { | |
| "epoch": 0.08, | |
| "learning_rate": 3.92156862745098e-06, | |
| "logits/chosen": -2.549417018890381, | |
| "logits/rejected": -2.4943110942840576, | |
| "logps/chosen": -169.3980712890625, | |
| "logps/rejected": -143.16995239257812, | |
| "loss": 0.6915, | |
| "rewards/accuracies": 0.637499988079071, | |
| "rewards/chosen": 0.008090810850262642, | |
| "rewards/margins": 0.003366248682141304, | |
| "rewards/rejected": 0.004724562633782625, | |
| "step": 80 | |
| }, | |
| { | |
| "epoch": 0.09, | |
| "learning_rate": 4.411764705882353e-06, | |
| "logits/chosen": -2.497745990753174, | |
| "logits/rejected": -2.445770740509033, | |
| "logps/chosen": -143.51312255859375, | |
| "logps/rejected": -118.7002944946289, | |
| "loss": 0.6909, | |
| "rewards/accuracies": 0.5625, | |
| "rewards/chosen": 0.009854977950453758, | |
| "rewards/margins": 0.004432962276041508, | |
| "rewards/rejected": 0.0054220156744122505, | |
| "step": 90 | |
| }, | |
| { | |
| "epoch": 0.1, | |
| "learning_rate": 4.901960784313726e-06, | |
| "logits/chosen": -2.501483917236328, | |
| "logits/rejected": -2.4084057807922363, | |
| "logps/chosen": -180.1342010498047, | |
| "logps/rejected": -141.8975372314453, | |
| "loss": 0.6874, | |
| "rewards/accuracies": 0.7749999761581421, | |
| "rewards/chosen": 0.021252702921628952, | |
| "rewards/margins": 0.01164860837161541, | |
| "rewards/rejected": 0.009604096412658691, | |
| "step": 100 | |
| }, | |
| { | |
| "epoch": 0.1, | |
| "eval_logits/chosen": -2.484314203262329, | |
| "eval_logits/rejected": -2.4462974071502686, | |
| "eval_logps/chosen": -157.93946838378906, | |
| "eval_logps/rejected": -132.1924285888672, | |
| "eval_loss": 0.6892239451408386, | |
| "eval_rewards/accuracies": 0.6698113083839417, | |
| "eval_rewards/chosen": 0.021303629502654076, | |
| "eval_rewards/margins": 0.008022623136639595, | |
| "eval_rewards/rejected": 0.013281003572046757, | |
| "eval_runtime": 285.1988, | |
| "eval_samples_per_second": 1.466, | |
| "eval_steps_per_second": 0.186, | |
| "step": 100 | |
| }, | |
| { | |
| "epoch": 0.11, | |
| "learning_rate": 4.9990549169459415e-06, | |
| "logits/chosen": -2.605736494064331, | |
| "logits/rejected": -2.551860809326172, | |
| "logps/chosen": -160.9684600830078, | |
| "logps/rejected": -134.45443725585938, | |
| "loss": 0.6862, | |
| "rewards/accuracies": 0.7875000238418579, | |
| "rewards/chosen": 0.029262090101838112, | |
| "rewards/margins": 0.014033245854079723, | |
| "rewards/rejected": 0.015228847041726112, | |
| "step": 110 | |
| }, | |
| { | |
| "epoch": 0.12, | |
| "learning_rate": 4.995216741642263e-06, | |
| "logits/chosen": -2.6044211387634277, | |
| "logits/rejected": -2.5780460834503174, | |
| "logps/chosen": -152.76980590820312, | |
| "logps/rejected": -132.57339477539062, | |
| "loss": 0.6886, | |
| "rewards/accuracies": 0.550000011920929, | |
| "rewards/chosen": 0.0384935587644577, | |
| "rewards/margins": 0.00926611851900816, | |
| "rewards/rejected": 0.02922743931412697, | |
| "step": 120 | |
| }, | |
| { | |
| "epoch": 0.13, | |
| "learning_rate": 4.988430936991089e-06, | |
| "logits/chosen": -2.594087600708008, | |
| "logits/rejected": -2.5282979011535645, | |
| "logps/chosen": -162.58688354492188, | |
| "logps/rejected": -133.3597869873047, | |
| "loss": 0.6804, | |
| "rewards/accuracies": 0.7124999761581421, | |
| "rewards/chosen": 0.06138850376009941, | |
| "rewards/margins": 0.02613685093820095, | |
| "rewards/rejected": 0.03525165468454361, | |
| "step": 130 | |
| }, | |
| { | |
| "epoch": 0.14, | |
| "learning_rate": 4.978705519144525e-06, | |
| "logits/chosen": -2.5525660514831543, | |
| "logits/rejected": -2.4592275619506836, | |
| "logps/chosen": -174.51707458496094, | |
| "logps/rejected": -135.41912841796875, | |
| "loss": 0.6759, | |
| "rewards/accuracies": 0.7250000238418579, | |
| "rewards/chosen": 0.06431619822978973, | |
| "rewards/margins": 0.03535788506269455, | |
| "rewards/rejected": 0.028958311304450035, | |
| "step": 140 | |
| }, | |
| { | |
| "epoch": 0.15, | |
| "learning_rate": 4.966051976854862e-06, | |
| "logits/chosen": -2.6359076499938965, | |
| "logits/rejected": -2.5874571800231934, | |
| "logps/chosen": -144.5331268310547, | |
| "logps/rejected": -122.17633056640625, | |
| "loss": 0.6729, | |
| "rewards/accuracies": 0.7250000238418579, | |
| "rewards/chosen": 0.08950608223676682, | |
| "rewards/margins": 0.042167335748672485, | |
| "rewards/rejected": 0.04733874648809433, | |
| "step": 150 | |
| }, | |
| { | |
| "epoch": 0.16, | |
| "learning_rate": 4.950485257902782e-06, | |
| "logits/chosen": -2.555276393890381, | |
| "logits/rejected": -2.523134708404541, | |
| "logps/chosen": -152.33584594726562, | |
| "logps/rejected": -128.93971252441406, | |
| "loss": 0.671, | |
| "rewards/accuracies": 0.762499988079071, | |
| "rewards/chosen": 0.10823699086904526, | |
| "rewards/margins": 0.04584677889943123, | |
| "rewards/rejected": 0.06239022687077522, | |
| "step": 160 | |
| }, | |
| { | |
| "epoch": 0.17, | |
| "learning_rate": 4.932023751439358e-06, | |
| "logits/chosen": -2.6216721534729004, | |
| "logits/rejected": -2.5455801486968994, | |
| "logps/chosen": -162.0784454345703, | |
| "logps/rejected": -133.4407501220703, | |
| "loss": 0.6717, | |
| "rewards/accuracies": 0.675000011920929, | |
| "rewards/chosen": 0.11643824726343155, | |
| "rewards/margins": 0.0462535098195076, | |
| "rewards/rejected": 0.07018472254276276, | |
| "step": 170 | |
| }, | |
| { | |
| "epoch": 0.18, | |
| "learning_rate": 4.9106892662627395e-06, | |
| "logits/chosen": -2.6522557735443115, | |
| "logits/rejected": -2.5888028144836426, | |
| "logps/chosen": -158.68748474121094, | |
| "logps/rejected": -135.9111785888672, | |
| "loss": 0.6686, | |
| "rewards/accuracies": 0.6875, | |
| "rewards/chosen": 0.11335412412881851, | |
| "rewards/margins": 0.05232863873243332, | |
| "rewards/rejected": 0.06102548912167549, | |
| "step": 180 | |
| }, | |
| { | |
| "epoch": 0.19, | |
| "learning_rate": 4.886507005055149e-06, | |
| "logits/chosen": -2.640076160430908, | |
| "logits/rejected": -2.545991897583008, | |
| "logps/chosen": -167.3929443359375, | |
| "logps/rejected": -137.95436096191406, | |
| "loss": 0.6509, | |
| "rewards/accuracies": 0.7124999761581421, | |
| "rewards/chosen": 0.10093629360198975, | |
| "rewards/margins": 0.09009988605976105, | |
| "rewards/rejected": 0.010836412198841572, | |
| "step": 190 | |
| }, | |
| { | |
| "epoch": 0.2, | |
| "learning_rate": 4.859505534610658e-06, | |
| "logits/chosen": -2.614877700805664, | |
| "logits/rejected": -2.576866865158081, | |
| "logps/chosen": -154.33758544921875, | |
| "logps/rejected": -138.316650390625, | |
| "loss": 0.6592, | |
| "rewards/accuracies": 0.625, | |
| "rewards/chosen": 0.017874887213110924, | |
| "rewards/margins": 0.0757044330239296, | |
| "rewards/rejected": -0.05782955139875412, | |
| "step": 200 | |
| }, | |
| { | |
| "epoch": 0.2, | |
| "eval_logits/chosen": -2.5370168685913086, | |
| "eval_logits/rejected": -2.4922032356262207, | |
| "eval_logps/chosen": -159.51800537109375, | |
| "eval_logps/rejected": -140.55880737304688, | |
| "eval_loss": 0.6593995690345764, | |
| "eval_rewards/accuracies": 0.6698113083839417, | |
| "eval_rewards/chosen": 0.00551834749057889, | |
| "eval_rewards/margins": 0.07590115815401077, | |
| "eval_rewards/rejected": -0.07038279622793198, | |
| "eval_runtime": 284.6148, | |
| "eval_samples_per_second": 1.469, | |
| "eval_steps_per_second": 0.186, | |
| "step": 200 | |
| }, | |
| { | |
| "epoch": 0.21, | |
| "learning_rate": 4.829716752088893e-06, | |
| "logits/chosen": -2.55920147895813, | |
| "logits/rejected": -2.5236589908599854, | |
| "logps/chosen": -148.85671997070312, | |
| "logps/rejected": -144.40414428710938, | |
| "loss": 0.6669, | |
| "rewards/accuracies": 0.6625000238418579, | |
| "rewards/chosen": -0.021094869822263718, | |
| "rewards/margins": 0.05866061523556709, | |
| "rewards/rejected": -0.07975548505783081, | |
| "step": 210 | |
| }, | |
| { | |
| "epoch": 0.22, | |
| "learning_rate": 4.797175847334535e-06, | |
| "logits/chosen": -2.614175319671631, | |
| "logits/rejected": -2.5591626167297363, | |
| "logps/chosen": -173.6731414794922, | |
| "logps/rejected": -153.340087890625, | |
| "loss": 0.6663, | |
| "rewards/accuracies": 0.637499988079071, | |
| "rewards/chosen": -0.06277570873498917, | |
| "rewards/margins": 0.06162955239415169, | |
| "rewards/rejected": -0.12440524995326996, | |
| "step": 220 | |
| }, | |
| { | |
| "epoch": 0.23, | |
| "learning_rate": 4.761921261307143e-06, | |
| "logits/chosen": -2.6242852210998535, | |
| "logits/rejected": -2.5607199668884277, | |
| "logps/chosen": -167.7920684814453, | |
| "logps/rejected": -153.7227020263672, | |
| "loss": 0.6412, | |
| "rewards/accuracies": 0.699999988079071, | |
| "rewards/chosen": -0.058848392218351364, | |
| "rewards/margins": 0.12100015580654144, | |
| "rewards/rejected": -0.1798485666513443, | |
| "step": 230 | |
| }, | |
| { | |
| "epoch": 0.24, | |
| "learning_rate": 4.723994640670377e-06, | |
| "logits/chosen": -2.6291072368621826, | |
| "logits/rejected": -2.528074026107788, | |
| "logps/chosen": -179.40382385253906, | |
| "logps/rejected": -153.91192626953125, | |
| "loss": 0.6197, | |
| "rewards/accuracies": 0.824999988079071, | |
| "rewards/chosen": -0.051289211958646774, | |
| "rewards/margins": 0.1633162945508957, | |
| "rewards/rejected": -0.21460552513599396, | |
| "step": 240 | |
| }, | |
| { | |
| "epoch": 0.25, | |
| "learning_rate": 4.68344078859431e-06, | |
| "logits/chosen": -2.5995261669158936, | |
| "logits/rejected": -2.5724520683288574, | |
| "logps/chosen": -163.8385009765625, | |
| "logps/rejected": -167.84129333496094, | |
| "loss": 0.6333, | |
| "rewards/accuracies": 0.675000011920929, | |
| "rewards/chosen": -0.11021089553833008, | |
| "rewards/margins": 0.13913698494434357, | |
| "rewards/rejected": -0.24934788048267365, | |
| "step": 250 | |
| }, | |
| { | |
| "epoch": 0.26, | |
| "learning_rate": 4.6403076118289006e-06, | |
| "logits/chosen": -2.5930025577545166, | |
| "logits/rejected": -2.5077426433563232, | |
| "logps/chosen": -182.69297790527344, | |
| "logps/rejected": -169.84852600097656, | |
| "loss": 0.6071, | |
| "rewards/accuracies": 0.8125, | |
| "rewards/chosen": -0.1480914056301117, | |
| "rewards/margins": 0.20134063065052032, | |
| "rewards/rejected": -0.3494320213794708, | |
| "step": 260 | |
| }, | |
| { | |
| "epoch": 0.27, | |
| "learning_rate": 4.5946460641111776e-06, | |
| "logits/chosen": -2.7093770503997803, | |
| "logits/rejected": -2.577432632446289, | |
| "logps/chosen": -188.6643524169922, | |
| "logps/rejected": -187.13156127929688, | |
| "loss": 0.5669, | |
| "rewards/accuracies": 0.8125, | |
| "rewards/chosen": -0.23154917359352112, | |
| "rewards/margins": 0.31018704175949097, | |
| "rewards/rejected": -0.5417362451553345, | |
| "step": 270 | |
| }, | |
| { | |
| "epoch": 0.28, | |
| "learning_rate": 4.546510085972983e-06, | |
| "logits/chosen": -2.679447650909424, | |
| "logits/rejected": -2.6179795265197754, | |
| "logps/chosen": -211.7890625, | |
| "logps/rejected": -199.5372772216797, | |
| "loss": 0.6167, | |
| "rewards/accuracies": 0.699999988079071, | |
| "rewards/chosen": -0.38407862186431885, | |
| "rewards/margins": 0.20160210132598877, | |
| "rewards/rejected": -0.5856807231903076, | |
| "step": 280 | |
| }, | |
| { | |
| "epoch": 0.29, | |
| "learning_rate": 4.495956541020376e-06, | |
| "logits/chosen": -2.7100636959075928, | |
| "logits/rejected": -2.6252739429473877, | |
| "logps/chosen": -222.813720703125, | |
| "logps/rejected": -226.9723663330078, | |
| "loss": 0.5781, | |
| "rewards/accuracies": 0.75, | |
| "rewards/chosen": -0.45635899901390076, | |
| "rewards/margins": 0.3029539883136749, | |
| "rewards/rejected": -0.7593129873275757, | |
| "step": 290 | |
| }, | |
| { | |
| "epoch": 0.3, | |
| "learning_rate": 4.443045148759978e-06, | |
| "logits/chosen": -2.6889595985412598, | |
| "logits/rejected": -2.5976881980895996, | |
| "logps/chosen": -219.4752960205078, | |
| "logps/rejected": -215.431396484375, | |
| "loss": 0.5451, | |
| "rewards/accuracies": 0.7124999761581421, | |
| "rewards/chosen": -0.3557833135128021, | |
| "rewards/margins": 0.4194447100162506, | |
| "rewards/rejected": -0.7752280831336975, | |
| "step": 300 | |
| }, | |
| { | |
| "epoch": 0.3, | |
| "eval_logits/chosen": -2.561957359313965, | |
| "eval_logits/rejected": -2.512842893600464, | |
| "eval_logps/chosen": -204.9713134765625, | |
| "eval_logps/rejected": -209.39376831054688, | |
| "eval_loss": 0.5867109298706055, | |
| "eval_rewards/accuracies": 0.6863207817077637, | |
| "eval_rewards/chosen": -0.44901466369628906, | |
| "eval_rewards/margins": 0.30971792340278625, | |
| "eval_rewards/rejected": -0.7587326169013977, | |
| "eval_runtime": 286.3596, | |
| "eval_samples_per_second": 1.46, | |
| "eval_steps_per_second": 0.185, | |
| "step": 300 | |
| }, | |
| { | |
| "epoch": 0.3, | |
| "learning_rate": 4.3878384140516025e-06, | |
| "logits/chosen": -2.6861400604248047, | |
| "logits/rejected": -2.6122899055480957, | |
| "logps/chosen": -209.1049346923828, | |
| "logps/rejected": -223.94308471679688, | |
| "loss": 0.5685, | |
| "rewards/accuracies": 0.675000011920929, | |
| "rewards/chosen": -0.4111746847629547, | |
| "rewards/margins": 0.355561226606369, | |
| "rewards/rejected": -0.7667359113693237, | |
| "step": 310 | |
| }, | |
| { | |
| "epoch": 0.31, | |
| "learning_rate": 4.330401553270522e-06, | |
| "logits/chosen": -2.649998188018799, | |
| "logits/rejected": -2.590202569961548, | |
| "logps/chosen": -201.97621154785156, | |
| "logps/rejected": -211.2521209716797, | |
| "loss": 0.5949, | |
| "rewards/accuracies": 0.6875, | |
| "rewards/chosen": -0.3732680678367615, | |
| "rewards/margins": 0.3101999759674072, | |
| "rewards/rejected": -0.6834679841995239, | |
| "step": 320 | |
| }, | |
| { | |
| "epoch": 0.32, | |
| "learning_rate": 4.2708024172665795e-06, | |
| "logits/chosen": -2.7253522872924805, | |
| "logits/rejected": -2.6092331409454346, | |
| "logps/chosen": -201.93325805664062, | |
| "logps/rejected": -203.0877685546875, | |
| "loss": 0.5527, | |
| "rewards/accuracies": 0.737500011920929, | |
| "rewards/chosen": -0.35112464427948, | |
| "rewards/margins": 0.431334912776947, | |
| "rewards/rejected": -0.7824596166610718, | |
| "step": 330 | |
| }, | |
| { | |
| "epoch": 0.33, | |
| "learning_rate": 4.209111411211174e-06, | |
| "logits/chosen": -2.7140471935272217, | |
| "logits/rejected": -2.611151933670044, | |
| "logps/chosen": -198.9755096435547, | |
| "logps/rejected": -215.7385711669922, | |
| "loss": 0.5351, | |
| "rewards/accuracies": 0.75, | |
| "rewards/chosen": -0.41874709725379944, | |
| "rewards/margins": 0.46593284606933594, | |
| "rewards/rejected": -0.8846799731254578, | |
| "step": 340 | |
| }, | |
| { | |
| "epoch": 0.34, | |
| "learning_rate": 4.145401411426788e-06, | |
| "logits/chosen": -2.7166478633880615, | |
| "logits/rejected": -2.650458574295044, | |
| "logps/chosen": -201.68899536132812, | |
| "logps/rejected": -224.9377899169922, | |
| "loss": 0.5406, | |
| "rewards/accuracies": 0.762499988079071, | |
| "rewards/chosen": -0.44746604561805725, | |
| "rewards/margins": 0.4966231882572174, | |
| "rewards/rejected": -0.9440892934799194, | |
| "step": 350 | |
| }, | |
| { | |
| "epoch": 0.35, | |
| "learning_rate": 4.079747679297314e-06, | |
| "logits/chosen": -2.631408214569092, | |
| "logits/rejected": -2.56895112991333, | |
| "logps/chosen": -220.35415649414062, | |
| "logps/rejected": -229.13998413085938, | |
| "loss": 0.5639, | |
| "rewards/accuracies": 0.6875, | |
| "rewards/chosen": -0.5187793970108032, | |
| "rewards/margins": 0.4147488474845886, | |
| "rewards/rejected": -0.9335282444953918, | |
| "step": 360 | |
| }, | |
| { | |
| "epoch": 0.36, | |
| "learning_rate": 4.012227772360889e-06, | |
| "logits/chosen": -2.64164662361145, | |
| "logits/rejected": -2.554852247238159, | |
| "logps/chosen": -224.2294158935547, | |
| "logps/rejected": -246.3839111328125, | |
| "loss": 0.5202, | |
| "rewards/accuracies": 0.7124999761581421, | |
| "rewards/chosen": -0.502095103263855, | |
| "rewards/margins": 0.5576826930046082, | |
| "rewards/rejected": -1.0597779750823975, | |
| "step": 370 | |
| }, | |
| { | |
| "epoch": 0.37, | |
| "learning_rate": 3.942921452690245e-06, | |
| "logits/chosen": -2.6596052646636963, | |
| "logits/rejected": -2.597478151321411, | |
| "logps/chosen": -226.8369140625, | |
| "logps/rejected": -239.9619140625, | |
| "loss": 0.6339, | |
| "rewards/accuracies": 0.6625000238418579, | |
| "rewards/chosen": -0.5941166877746582, | |
| "rewards/margins": 0.3102978467941284, | |
| "rewards/rejected": -0.9044146537780762, | |
| "step": 380 | |
| }, | |
| { | |
| "epoch": 0.38, | |
| "learning_rate": 3.871910592668817e-06, | |
| "logits/chosen": -2.70902156829834, | |
| "logits/rejected": -2.634817123413086, | |
| "logps/chosen": -222.9804229736328, | |
| "logps/rejected": -253.72097778320312, | |
| "loss": 0.5281, | |
| "rewards/accuracies": 0.7875000238418579, | |
| "rewards/chosen": -0.6194333434104919, | |
| "rewards/margins": 0.5652093887329102, | |
| "rewards/rejected": -1.1846425533294678, | |
| "step": 390 | |
| }, | |
| { | |
| "epoch": 0.39, | |
| "learning_rate": 3.799279078273921e-06, | |
| "logits/chosen": -2.6576032638549805, | |
| "logits/rejected": -2.552799940109253, | |
| "logps/chosen": -214.78018188476562, | |
| "logps/rejected": -235.85055541992188, | |
| "loss": 0.4933, | |
| "rewards/accuracies": 0.7124999761581421, | |
| "rewards/chosen": -0.48761433362960815, | |
| "rewards/margins": 0.6489143371582031, | |
| "rewards/rejected": -1.136528730392456, | |
| "step": 400 | |
| }, | |
| { | |
| "epoch": 0.39, | |
| "eval_logits/chosen": -2.5385656356811523, | |
| "eval_logits/rejected": -2.486776113510132, | |
| "eval_logps/chosen": -220.6713409423828, | |
| "eval_logps/rejected": -243.80616760253906, | |
| "eval_loss": 0.5590807795524597, | |
| "eval_rewards/accuracies": 0.7146226167678833, | |
| "eval_rewards/chosen": -0.6060149669647217, | |
| "eval_rewards/margins": 0.4968414306640625, | |
| "eval_rewards/rejected": -1.1028562784194946, | |
| "eval_runtime": 285.6623, | |
| "eval_samples_per_second": 1.463, | |
| "eval_steps_per_second": 0.186, | |
| "step": 400 | |
| }, | |
| { | |
| "epoch": 0.4, | |
| "learning_rate": 3.725112709981249e-06, | |
| "logits/chosen": -2.5486836433410645, | |
| "logits/rejected": -2.4727280139923096, | |
| "logps/chosen": -231.33547973632812, | |
| "logps/rejected": -252.45864868164062, | |
| "loss": 0.5563, | |
| "rewards/accuracies": 0.7250000238418579, | |
| "rewards/chosen": -0.5974920988082886, | |
| "rewards/margins": 0.4878315329551697, | |
| "rewards/rejected": -1.0853235721588135, | |
| "step": 410 | |
| }, | |
| { | |
| "epoch": 0.41, | |
| "learning_rate": 3.649499101407737e-06, | |
| "logits/chosen": -2.651151180267334, | |
| "logits/rejected": -2.556567430496216, | |
| "logps/chosen": -241.06167602539062, | |
| "logps/rejected": -253.01986694335938, | |
| "loss": 0.6047, | |
| "rewards/accuracies": 0.6875, | |
| "rewards/chosen": -0.6971057057380676, | |
| "rewards/margins": 0.3935045599937439, | |
| "rewards/rejected": -1.0906102657318115, | |
| "step": 420 | |
| }, | |
| { | |
| "epoch": 0.42, | |
| "learning_rate": 3.5725275758125564e-06, | |
| "logits/chosen": -2.616089105606079, | |
| "logits/rejected": -2.516278028488159, | |
| "logps/chosen": -213.630859375, | |
| "logps/rejected": -243.5413818359375, | |
| "loss": 0.5115, | |
| "rewards/accuracies": 0.7875000238418579, | |
| "rewards/chosen": -0.6288307905197144, | |
| "rewards/margins": 0.6640511751174927, | |
| "rewards/rejected": -1.2928818464279175, | |
| "step": 430 | |
| }, | |
| { | |
| "epoch": 0.43, | |
| "learning_rate": 3.494289060578478e-06, | |
| "logits/chosen": -2.6611900329589844, | |
| "logits/rejected": -2.609663486480713, | |
| "logps/chosen": -239.8604736328125, | |
| "logps/rejected": -275.2757873535156, | |
| "loss": 0.5247, | |
| "rewards/accuracies": 0.75, | |
| "rewards/chosen": -0.8715305328369141, | |
| "rewards/margins": 0.5755683779716492, | |
| "rewards/rejected": -1.447098970413208, | |
| "step": 440 | |
| }, | |
| { | |
| "epoch": 0.44, | |
| "learning_rate": 3.414875979798272e-06, | |
| "logits/chosen": -2.640732526779175, | |
| "logits/rejected": -2.5787458419799805, | |
| "logps/chosen": -237.64920043945312, | |
| "logps/rejected": -276.0288391113281, | |
| "loss": 0.4854, | |
| "rewards/accuracies": 0.699999988079071, | |
| "rewards/chosen": -0.5312285423278809, | |
| "rewards/margins": 0.7248728275299072, | |
| "rewards/rejected": -1.256101369857788, | |
| "step": 450 | |
| }, | |
| { | |
| "epoch": 0.45, | |
| "learning_rate": 3.3343821450930196e-06, | |
| "logits/chosen": -2.7021923065185547, | |
| "logits/rejected": -2.6052348613739014, | |
| "logps/chosen": -231.91165161132812, | |
| "logps/rejected": -256.6305847167969, | |
| "loss": 0.5014, | |
| "rewards/accuracies": 0.762499988079071, | |
| "rewards/chosen": -0.6199955940246582, | |
| "rewards/margins": 0.6785587072372437, | |
| "rewards/rejected": -1.2985544204711914, | |
| "step": 460 | |
| }, | |
| { | |
| "epoch": 0.46, | |
| "learning_rate": 3.252902644791325e-06, | |
| "logits/chosen": -2.5665388107299805, | |
| "logits/rejected": -2.487617015838623, | |
| "logps/chosen": -232.90200805664062, | |
| "logps/rejected": -276.9921569824219, | |
| "loss": 0.4943, | |
| "rewards/accuracies": 0.8374999761581421, | |
| "rewards/chosen": -0.6983159780502319, | |
| "rewards/margins": 0.7069972157478333, | |
| "rewards/rejected": -1.40531325340271, | |
| "step": 470 | |
| }, | |
| { | |
| "epoch": 0.47, | |
| "learning_rate": 3.170533731600339e-06, | |
| "logits/chosen": -2.618633985519409, | |
| "logits/rejected": -2.5317630767822266, | |
| "logps/chosen": -239.92919921875, | |
| "logps/rejected": -261.753173828125, | |
| "loss": 0.543, | |
| "rewards/accuracies": 0.699999988079071, | |
| "rewards/chosen": -0.7530533075332642, | |
| "rewards/margins": 0.5510448217391968, | |
| "rewards/rejected": -1.304098129272461, | |
| "step": 480 | |
| }, | |
| { | |
| "epoch": 0.48, | |
| "learning_rate": 3.0873727089012816e-06, | |
| "logits/chosen": -2.7126545906066895, | |
| "logits/rejected": -2.6558868885040283, | |
| "logps/chosen": -242.5832977294922, | |
| "logps/rejected": -258.93365478515625, | |
| "loss": 0.5719, | |
| "rewards/accuracies": 0.7250000238418579, | |
| "rewards/chosen": -0.7276023626327515, | |
| "rewards/margins": 0.48720917105674744, | |
| "rewards/rejected": -1.2148113250732422, | |
| "step": 490 | |
| }, | |
| { | |
| "epoch": 0.49, | |
| "learning_rate": 3.0035178158038026e-06, | |
| "logits/chosen": -2.594712018966675, | |
| "logits/rejected": -2.4889943599700928, | |
| "logps/chosen": -238.0956268310547, | |
| "logps/rejected": -263.0235290527344, | |
| "loss": 0.5271, | |
| "rewards/accuracies": 0.75, | |
| "rewards/chosen": -0.6214576959609985, | |
| "rewards/margins": 0.6590222716331482, | |
| "rewards/rejected": -1.280479907989502, | |
| "step": 500 | |
| }, | |
| { | |
| "epoch": 0.49, | |
| "eval_logits/chosen": -2.531158924102783, | |
| "eval_logits/rejected": -2.4784035682678223, | |
| "eval_logps/chosen": -227.1888885498047, | |
| "eval_logps/rejected": -260.8957824707031, | |
| "eval_loss": 0.5488241314888, | |
| "eval_rewards/accuracies": 0.7193396091461182, | |
| "eval_rewards/chosen": -0.6711904406547546, | |
| "eval_rewards/margins": 0.6025623083114624, | |
| "eval_rewards/rejected": -1.2737528085708618, | |
| "eval_runtime": 286.5537, | |
| "eval_samples_per_second": 1.459, | |
| "eval_steps_per_second": 0.185, | |
| "step": 500 | |
| }, | |
| { | |
| "epoch": 0.5, | |
| "learning_rate": 2.919068111094937e-06, | |
| "logits/chosen": -2.6373202800750732, | |
| "logits/rejected": -2.5553860664367676, | |
| "logps/chosen": -229.17578125, | |
| "logps/rejected": -268.92559814453125, | |
| "loss": 0.5683, | |
| "rewards/accuracies": 0.7124999761581421, | |
| "rewards/chosen": -0.7968514561653137, | |
| "rewards/margins": 0.5847809314727783, | |
| "rewards/rejected": -1.3816323280334473, | |
| "step": 510 | |
| }, | |
| { | |
| "epoch": 0.51, | |
| "learning_rate": 2.8341233562197895e-06, | |
| "logits/chosen": -2.64227032661438, | |
| "logits/rejected": -2.599839687347412, | |
| "logps/chosen": -215.4423828125, | |
| "logps/rejected": -253.928955078125, | |
| "loss": 0.5432, | |
| "rewards/accuracies": 0.7749999761581421, | |
| "rewards/chosen": -0.6491907835006714, | |
| "rewards/margins": 0.5866181254386902, | |
| "rewards/rejected": -1.235809087753296, | |
| "step": 520 | |
| }, | |
| { | |
| "epoch": 0.52, | |
| "learning_rate": 2.7487838974321352e-06, | |
| "logits/chosen": -2.613891124725342, | |
| "logits/rejected": -2.5329458713531494, | |
| "logps/chosen": -212.8946533203125, | |
| "logps/rejected": -254.96609497070312, | |
| "loss": 0.5181, | |
| "rewards/accuracies": 0.824999988079071, | |
| "rewards/chosen": -0.5764448046684265, | |
| "rewards/margins": 0.6870908737182617, | |
| "rewards/rejected": -1.263535737991333, | |
| "step": 530 | |
| }, | |
| { | |
| "epoch": 0.53, | |
| "learning_rate": 2.6631505472541997e-06, | |
| "logits/chosen": -2.60998272895813, | |
| "logits/rejected": -2.516542911529541, | |
| "logps/chosen": -222.1146697998047, | |
| "logps/rejected": -263.4886779785156, | |
| "loss": 0.492, | |
| "rewards/accuracies": 0.800000011920929, | |
| "rewards/chosen": -0.549847424030304, | |
| "rewards/margins": 0.7145514488220215, | |
| "rewards/rejected": -1.2643989324569702, | |
| "step": 540 | |
| }, | |
| { | |
| "epoch": 0.54, | |
| "learning_rate": 2.5773244653856173e-06, | |
| "logits/chosen": -2.579040050506592, | |
| "logits/rejected": -2.4950509071350098, | |
| "logps/chosen": -228.0793914794922, | |
| "logps/rejected": -273.71649169921875, | |
| "loss": 0.4673, | |
| "rewards/accuracies": 0.75, | |
| "rewards/chosen": -0.5778298377990723, | |
| "rewards/margins": 0.7735764980316162, | |
| "rewards/rejected": -1.3514063358306885, | |
| "step": 550 | |
| }, | |
| { | |
| "epoch": 0.55, | |
| "learning_rate": 2.4914070392022717e-06, | |
| "logits/chosen": -2.612720251083374, | |
| "logits/rejected": -2.5363128185272217, | |
| "logps/chosen": -247.00271606445312, | |
| "logps/rejected": -277.1578369140625, | |
| "loss": 0.5677, | |
| "rewards/accuracies": 0.737500011920929, | |
| "rewards/chosen": -0.7081204652786255, | |
| "rewards/margins": 0.5781086087226868, | |
| "rewards/rejected": -1.286229133605957, | |
| "step": 560 | |
| }, | |
| { | |
| "epoch": 0.56, | |
| "learning_rate": 2.4054997639861778e-06, | |
| "logits/chosen": -2.5522828102111816, | |
| "logits/rejected": -2.4502105712890625, | |
| "logps/chosen": -244.18655395507812, | |
| "logps/rejected": -291.7619934082031, | |
| "loss": 0.4671, | |
| "rewards/accuracies": 0.824999988079071, | |
| "rewards/chosen": -0.685632050037384, | |
| "rewards/margins": 0.871982753276825, | |
| "rewards/rejected": -1.557614803314209, | |
| "step": 570 | |
| }, | |
| { | |
| "epoch": 0.57, | |
| "learning_rate": 2.3197041230278905e-06, | |
| "logits/chosen": -2.607151508331299, | |
| "logits/rejected": -2.5367910861968994, | |
| "logps/chosen": -242.89242553710938, | |
| "logps/rejected": -303.93426513671875, | |
| "loss": 0.4696, | |
| "rewards/accuracies": 0.800000011920929, | |
| "rewards/chosen": -0.7322829961776733, | |
| "rewards/margins": 0.8701601028442383, | |
| "rewards/rejected": -1.6024430990219116, | |
| "step": 580 | |
| }, | |
| { | |
| "epoch": 0.58, | |
| "learning_rate": 2.234121467743082e-06, | |
| "logits/chosen": -2.630315065383911, | |
| "logits/rejected": -2.5930495262145996, | |
| "logps/chosen": -234.6758270263672, | |
| "logps/rejected": -278.23175048828125, | |
| "loss": 0.5646, | |
| "rewards/accuracies": 0.75, | |
| "rewards/chosen": -0.7757617235183716, | |
| "rewards/margins": 0.6909924149513245, | |
| "rewards/rejected": -1.4667540788650513, | |
| "step": 590 | |
| }, | |
| { | |
| "epoch": 0.59, | |
| "learning_rate": 2.148852897944905e-06, | |
| "logits/chosen": -2.6809496879577637, | |
| "logits/rejected": -2.5784506797790527, | |
| "logps/chosen": -226.6197052001953, | |
| "logps/rejected": -279.7991027832031, | |
| "loss": 0.4594, | |
| "rewards/accuracies": 0.737500011920929, | |
| "rewards/chosen": -0.6655336022377014, | |
| "rewards/margins": 0.8101358413696289, | |
| "rewards/rejected": -1.475669503211975, | |
| "step": 600 | |
| }, | |
| { | |
| "epoch": 0.59, | |
| "eval_logits/chosen": -2.519953966140747, | |
| "eval_logits/rejected": -2.4671545028686523, | |
| "eval_logps/chosen": -239.843017578125, | |
| "eval_logps/rejected": -280.2419738769531, | |
| "eval_loss": 0.5418068170547485, | |
| "eval_rewards/accuracies": 0.7311320900917053, | |
| "eval_rewards/chosen": -0.7977315783500671, | |
| "eval_rewards/margins": 0.6694828271865845, | |
| "eval_rewards/rejected": -1.4672144651412964, | |
| "eval_runtime": 286.7576, | |
| "eval_samples_per_second": 1.458, | |
| "eval_steps_per_second": 0.185, | |
| "step": 600 | |
| }, | |
| { | |
| "epoch": 0.6, | |
| "learning_rate": 2.063999142413574e-06, | |
| "logits/chosen": -2.5949954986572266, | |
| "logits/rejected": -2.512058734893799, | |
| "logps/chosen": -240.7419891357422, | |
| "logps/rejected": -261.2515869140625, | |
| "loss": 0.5227, | |
| "rewards/accuracies": 0.762499988079071, | |
| "rewards/chosen": -0.7232692241668701, | |
| "rewards/margins": 0.6376546621322632, | |
| "rewards/rejected": -1.3609240055084229, | |
| "step": 610 | |
| }, | |
| { | |
| "epoch": 0.61, | |
| "learning_rate": 1.9796604399042547e-06, | |
| "logits/chosen": -2.6472525596618652, | |
| "logits/rejected": -2.5546352863311768, | |
| "logps/chosen": -257.9134521484375, | |
| "logps/rejected": -305.9988098144531, | |
| "loss": 0.4877, | |
| "rewards/accuracies": 0.737500011920929, | |
| "rewards/chosen": -0.944113552570343, | |
| "rewards/margins": 0.8468207120895386, | |
| "rewards/rejected": -1.7909342050552368, | |
| "step": 620 | |
| }, | |
| { | |
| "epoch": 0.62, | |
| "learning_rate": 1.8959364207338216e-06, | |
| "logits/chosen": -2.6289591789245605, | |
| "logits/rejected": -2.536094903945923, | |
| "logps/chosen": -235.07626342773438, | |
| "logps/rejected": -293.07818603515625, | |
| "loss": 0.499, | |
| "rewards/accuracies": 0.75, | |
| "rewards/chosen": -0.8861194849014282, | |
| "rewards/margins": 0.8537158966064453, | |
| "rewards/rejected": -1.7398353815078735, | |
| "step": 630 | |
| }, | |
| { | |
| "epoch": 0.63, | |
| "learning_rate": 1.8129259890863825e-06, | |
| "logits/chosen": -2.599257469177246, | |
| "logits/rejected": -2.542430877685547, | |
| "logps/chosen": -262.864013671875, | |
| "logps/rejected": -305.956787109375, | |
| "loss": 0.5611, | |
| "rewards/accuracies": 0.7250000238418579, | |
| "rewards/chosen": -0.9360750317573547, | |
| "rewards/margins": 0.6403096318244934, | |
| "rewards/rejected": -1.5763845443725586, | |
| "step": 640 | |
| }, | |
| { | |
| "epoch": 0.64, | |
| "learning_rate": 1.7307272061765738e-06, | |
| "logits/chosen": -2.641481876373291, | |
| "logits/rejected": -2.58241605758667, | |
| "logps/chosen": -251.8070068359375, | |
| "logps/rejected": -303.80462646484375, | |
| "loss": 0.5001, | |
| "rewards/accuracies": 0.737500011920929, | |
| "rewards/chosen": -0.8562856912612915, | |
| "rewards/margins": 0.783263087272644, | |
| "rewards/rejected": -1.639548659324646, | |
| "step": 650 | |
| }, | |
| { | |
| "epoch": 0.65, | |
| "learning_rate": 1.649437174408685e-06, | |
| "logits/chosen": -2.597597599029541, | |
| "logits/rejected": -2.541797637939453, | |
| "logps/chosen": -235.8343505859375, | |
| "logps/rejected": -286.0321350097656, | |
| "loss": 0.5432, | |
| "rewards/accuracies": 0.75, | |
| "rewards/chosen": -0.6598041653633118, | |
| "rewards/margins": 0.7441025972366333, | |
| "rewards/rejected": -1.4039068222045898, | |
| "step": 660 | |
| }, | |
| { | |
| "epoch": 0.66, | |
| "learning_rate": 1.569151922668422e-06, | |
| "logits/chosen": -2.638943910598755, | |
| "logits/rejected": -2.5677754878997803, | |
| "logps/chosen": -246.6320037841797, | |
| "logps/rejected": -297.98944091796875, | |
| "loss": 0.5573, | |
| "rewards/accuracies": 0.699999988079071, | |
| "rewards/chosen": -0.9078775644302368, | |
| "rewards/margins": 0.736061692237854, | |
| "rewards/rejected": -1.6439392566680908, | |
| "step": 670 | |
| }, | |
| { | |
| "epoch": 0.67, | |
| "learning_rate": 1.4899662928828428e-06, | |
| "logits/chosen": -2.6092512607574463, | |
| "logits/rejected": -2.5140151977539062, | |
| "logps/chosen": -229.247802734375, | |
| "logps/rejected": -278.5384521484375, | |
| "loss": 0.5307, | |
| "rewards/accuracies": 0.800000011920929, | |
| "rewards/chosen": -0.7807116508483887, | |
| "rewards/margins": 0.7786010503768921, | |
| "rewards/rejected": -1.5593125820159912, | |
| "step": 680 | |
| }, | |
| { | |
| "epoch": 0.68, | |
| "learning_rate": 1.4119738279824507e-06, | |
| "logits/chosen": -2.5588324069976807, | |
| "logits/rejected": -2.4750542640686035, | |
| "logps/chosen": -229.9861602783203, | |
| "logps/rejected": -279.2628479003906, | |
| "loss": 0.5406, | |
| "rewards/accuracies": 0.7250000238418579, | |
| "rewards/chosen": -0.7474340200424194, | |
| "rewards/margins": 0.7327946424484253, | |
| "rewards/rejected": -1.4802285432815552, | |
| "step": 690 | |
| }, | |
| { | |
| "epoch": 0.69, | |
| "learning_rate": 1.3352666613978152e-06, | |
| "logits/chosen": -2.522113084793091, | |
| "logits/rejected": -2.478847026824951, | |
| "logps/chosen": -247.2416229248047, | |
| "logps/rejected": -284.3504943847656, | |
| "loss": 0.5444, | |
| "rewards/accuracies": 0.7124999761581421, | |
| "rewards/chosen": -0.8833193778991699, | |
| "rewards/margins": 0.644477367401123, | |
| "rewards/rejected": -1.5277965068817139, | |
| "step": 700 | |
| }, | |
| { | |
| "epoch": 0.69, | |
| "eval_logits/chosen": -2.5126736164093018, | |
| "eval_logits/rejected": -2.459439277648926, | |
| "eval_logps/chosen": -236.95314025878906, | |
| "eval_logps/rejected": -278.8013610839844, | |
| "eval_loss": 0.5358262062072754, | |
| "eval_rewards/accuracies": 0.7334905862808228, | |
| "eval_rewards/chosen": -0.7688332200050354, | |
| "eval_rewards/margins": 0.683975100517273, | |
| "eval_rewards/rejected": -1.4528082609176636, | |
| "eval_runtime": 286.9906, | |
| "eval_samples_per_second": 1.456, | |
| "eval_steps_per_second": 0.185, | |
| "step": 700 | |
| }, | |
| { | |
| "epoch": 0.7, | |
| "learning_rate": 1.2599354082212523e-06, | |
| "logits/chosen": -2.5635337829589844, | |
| "logits/rejected": -2.475813388824463, | |
| "logps/chosen": -212.53189086914062, | |
| "logps/rejected": -260.5958557128906, | |
| "loss": 0.5358, | |
| "rewards/accuracies": 0.762499988079071, | |
| "rewards/chosen": -0.5574060678482056, | |
| "rewards/margins": 0.7405737042427063, | |
| "rewards/rejected": -1.297979712486267, | |
| "step": 710 | |
| }, | |
| { | |
| "epoch": 0.71, | |
| "learning_rate": 1.186069058162127e-06, | |
| "logits/chosen": -2.6438024044036865, | |
| "logits/rejected": -2.565239906311035, | |
| "logps/chosen": -234.189697265625, | |
| "logps/rejected": -277.8279724121094, | |
| "loss": 0.5251, | |
| "rewards/accuracies": 0.737500011920929, | |
| "rewards/chosen": -0.8120694160461426, | |
| "rewards/margins": 0.6847533583641052, | |
| "rewards/rejected": -1.496822714805603, | |
| "step": 720 | |
| }, | |
| { | |
| "epoch": 0.72, | |
| "learning_rate": 1.113754870422254e-06, | |
| "logits/chosen": -2.6735897064208984, | |
| "logits/rejected": -2.599342107772827, | |
| "logps/chosen": -255.75906372070312, | |
| "logps/rejected": -297.49774169921875, | |
| "loss": 0.5672, | |
| "rewards/accuracies": 0.6625000238418579, | |
| "rewards/chosen": -0.9736843109130859, | |
| "rewards/margins": 0.5445233583450317, | |
| "rewards/rejected": -1.5182076692581177, | |
| "step": 730 | |
| }, | |
| { | |
| "epoch": 0.73, | |
| "learning_rate": 1.0430782706155545e-06, | |
| "logits/chosen": -2.6966583728790283, | |
| "logits/rejected": -2.6031241416931152, | |
| "logps/chosen": -255.4834747314453, | |
| "logps/rejected": -305.7486267089844, | |
| "loss": 0.4757, | |
| "rewards/accuracies": 0.8125, | |
| "rewards/chosen": -0.8519601821899414, | |
| "rewards/margins": 0.8620556592941284, | |
| "rewards/rejected": -1.7140159606933594, | |
| "step": 740 | |
| }, | |
| { | |
| "epoch": 0.74, | |
| "learning_rate": 9.741227498537615e-07, | |
| "logits/chosen": -2.7119975090026855, | |
| "logits/rejected": -2.626258373260498, | |
| "logps/chosen": -239.32894897460938, | |
| "logps/rejected": -292.5682678222656, | |
| "loss": 0.4648, | |
| "rewards/accuracies": 0.7875000238418579, | |
| "rewards/chosen": -0.9102465510368347, | |
| "rewards/margins": 0.7873125076293945, | |
| "rewards/rejected": -1.697559118270874, | |
| "step": 750 | |
| }, | |
| { | |
| "epoch": 0.75, | |
| "learning_rate": 9.069697661173668e-07, | |
| "logits/chosen": -2.6146960258483887, | |
| "logits/rejected": -2.5526528358459473, | |
| "logps/chosen": -253.11514282226562, | |
| "logps/rejected": -305.19647216796875, | |
| "loss": 0.5725, | |
| "rewards/accuracies": 0.637499988079071, | |
| "rewards/chosen": -0.9893601536750793, | |
| "rewards/margins": 0.6918299794197083, | |
| "rewards/rejected": -1.6811901330947876, | |
| "step": 760 | |
| }, | |
| { | |
| "epoch": 0.76, | |
| "learning_rate": 8.416986480283434e-07, | |
| "logits/chosen": -2.593777894973755, | |
| "logits/rejected": -2.497063636779785, | |
| "logps/chosen": -259.0460510253906, | |
| "logps/rejected": -318.4884033203125, | |
| "loss": 0.5069, | |
| "rewards/accuracies": 0.7749999761581421, | |
| "rewards/chosen": -1.0463413000106812, | |
| "rewards/margins": 0.8268297910690308, | |
| "rewards/rejected": -1.8731712102890015, | |
| "step": 770 | |
| }, | |
| { | |
| "epoch": 0.77, | |
| "learning_rate": 7.783865011382876e-07, | |
| "logits/chosen": -2.6105198860168457, | |
| "logits/rejected": -2.5155270099639893, | |
| "logps/chosen": -267.23785400390625, | |
| "logps/rejected": -316.72027587890625, | |
| "loss": 0.4906, | |
| "rewards/accuracies": 0.75, | |
| "rewards/chosen": -0.9409424662590027, | |
| "rewards/margins": 0.8792687654495239, | |
| "rewards/rejected": -1.8202111721038818, | |
| "step": 780 | |
| }, | |
| { | |
| "epoch": 0.78, | |
| "learning_rate": 7.171081168427205e-07, | |
| "logits/chosen": -2.642396926879883, | |
| "logits/rejected": -2.5414986610412598, | |
| "logps/chosen": -278.53076171875, | |
| "logps/rejected": -327.7699279785156, | |
| "loss": 0.4906, | |
| "rewards/accuracies": 0.7250000238418579, | |
| "rewards/chosen": -0.944370448589325, | |
| "rewards/margins": 0.8692710995674133, | |
| "rewards/rejected": -1.8136417865753174, | |
| "step": 790 | |
| }, | |
| { | |
| "epoch": 0.79, | |
| "learning_rate": 6.579358840291064e-07, | |
| "logits/chosen": -2.619293451309204, | |
| "logits/rejected": -2.5392627716064453, | |
| "logps/chosen": -275.8484802246094, | |
| "logps/rejected": -315.7076721191406, | |
| "loss": 0.5755, | |
| "rewards/accuracies": 0.699999988079071, | |
| "rewards/chosen": -1.142333745956421, | |
| "rewards/margins": 0.5214550495147705, | |
| "rewards/rejected": -1.6637885570526123, | |
| "step": 800 | |
| }, | |
| { | |
| "epoch": 0.79, | |
| "eval_logits/chosen": -2.511751890182495, | |
| "eval_logits/rejected": -2.458522081375122, | |
| "eval_logps/chosen": -266.7905578613281, | |
| "eval_logps/rejected": -309.8293151855469, | |
| "eval_loss": 0.5405105352401733, | |
| "eval_rewards/accuracies": 0.7311320900917053, | |
| "eval_rewards/chosen": -1.0672070980072021, | |
| "eval_rewards/margins": 0.695880651473999, | |
| "eval_rewards/rejected": -1.7630878686904907, | |
| "eval_runtime": 287.1767, | |
| "eval_samples_per_second": 1.456, | |
| "eval_steps_per_second": 0.185, | |
| "step": 800 | |
| }, | |
| { | |
| "epoch": 0.8, | |
| "learning_rate": 6.00939703563006e-07, | |
| "logits/chosen": -2.647855281829834, | |
| "logits/rejected": -2.6014881134033203, | |
| "logps/chosen": -251.8603057861328, | |
| "logps/rejected": -300.19500732421875, | |
| "loss": 0.5697, | |
| "rewards/accuracies": 0.6875, | |
| "rewards/chosen": -1.0764825344085693, | |
| "rewards/margins": 0.6206298470497131, | |
| "rewards/rejected": -1.6971123218536377, | |
| "step": 810 | |
| }, | |
| { | |
| "epoch": 0.81, | |
| "learning_rate": 5.461869057133412e-07, | |
| "logits/chosen": -2.6618542671203613, | |
| "logits/rejected": -2.6055150032043457, | |
| "logps/chosen": -266.5983581542969, | |
| "logps/rejected": -328.2469177246094, | |
| "loss": 0.498, | |
| "rewards/accuracies": 0.737500011920929, | |
| "rewards/chosen": -1.0761178731918335, | |
| "rewards/margins": 0.7884665727615356, | |
| "rewards/rejected": -1.8645845651626587, | |
| "step": 820 | |
| }, | |
| { | |
| "epoch": 0.82, | |
| "learning_rate": 4.937421706143497e-07, | |
| "logits/chosen": -2.5800223350524902, | |
| "logits/rejected": -2.5038163661956787, | |
| "logps/chosen": -276.6379699707031, | |
| "logps/rejected": -326.5852355957031, | |
| "loss": 0.5034, | |
| "rewards/accuracies": 0.762499988079071, | |
| "rewards/chosen": -1.1193444728851318, | |
| "rewards/margins": 0.8236180543899536, | |
| "rewards/rejected": -1.9429620504379272, | |
| "step": 830 | |
| }, | |
| { | |
| "epoch": 0.83, | |
| "learning_rate": 4.43667451858166e-07, | |
| "logits/chosen": -2.5895838737487793, | |
| "logits/rejected": -2.4980013370513916, | |
| "logps/chosen": -245.9558563232422, | |
| "logps/rejected": -303.1318054199219, | |
| "loss": 0.4828, | |
| "rewards/accuracies": 0.8125, | |
| "rewards/chosen": -0.9489553570747375, | |
| "rewards/margins": 0.8449093103408813, | |
| "rewards/rejected": -1.7938644886016846, | |
| "step": 840 | |
| }, | |
| { | |
| "epoch": 0.84, | |
| "learning_rate": 3.9602190330830484e-07, | |
| "logits/chosen": -2.5641655921936035, | |
| "logits/rejected": -2.5129027366638184, | |
| "logps/chosen": -265.5952453613281, | |
| "logps/rejected": -319.8917236328125, | |
| "loss": 0.5089, | |
| "rewards/accuracies": 0.737500011920929, | |
| "rewards/chosen": -1.1247035264968872, | |
| "rewards/margins": 0.7079113721847534, | |
| "rewards/rejected": -1.8326148986816406, | |
| "step": 850 | |
| }, | |
| { | |
| "epoch": 0.85, | |
| "learning_rate": 3.5086180922049295e-07, | |
| "logits/chosen": -2.653895854949951, | |
| "logits/rejected": -2.570190906524658, | |
| "logps/chosen": -277.4553527832031, | |
| "logps/rejected": -311.83050537109375, | |
| "loss": 0.4957, | |
| "rewards/accuracies": 0.75, | |
| "rewards/chosen": -0.9871084094047546, | |
| "rewards/margins": 0.7766774892807007, | |
| "rewards/rejected": -1.7637859582901, | |
| "step": 860 | |
| }, | |
| { | |
| "epoch": 0.86, | |
| "learning_rate": 3.0824051775340895e-07, | |
| "logits/chosen": -2.6221535205841064, | |
| "logits/rejected": -2.560570001602173, | |
| "logps/chosen": -223.2664337158203, | |
| "logps/rejected": -278.30316162109375, | |
| "loss": 0.5322, | |
| "rewards/accuracies": 0.7124999761581421, | |
| "rewards/chosen": -0.8692310452461243, | |
| "rewards/margins": 0.6867448091506958, | |
| "rewards/rejected": -1.5559759140014648, | |
| "step": 870 | |
| }, | |
| { | |
| "epoch": 0.87, | |
| "learning_rate": 2.6820837794786336e-07, | |
| "logits/chosen": -2.5683445930480957, | |
| "logits/rejected": -2.54264497756958, | |
| "logps/chosen": -264.15106201171875, | |
| "logps/rejected": -287.0775451660156, | |
| "loss": 0.6567, | |
| "rewards/accuracies": 0.6000000238418579, | |
| "rewards/chosen": -1.0795668363571167, | |
| "rewards/margins": 0.3045983612537384, | |
| "rewards/rejected": -1.3841651678085327, | |
| "step": 880 | |
| }, | |
| { | |
| "epoch": 0.88, | |
| "learning_rate": 2.3081268024887694e-07, | |
| "logits/chosen": -2.5630886554718018, | |
| "logits/rejected": -2.45955228805542, | |
| "logps/chosen": -250.3737335205078, | |
| "logps/rejected": -303.13177490234375, | |
| "loss": 0.4701, | |
| "rewards/accuracies": 0.7250000238418579, | |
| "rewards/chosen": -0.9046028852462769, | |
| "rewards/margins": 0.8787680864334106, | |
| "rewards/rejected": -1.7833709716796875, | |
| "step": 890 | |
| }, | |
| { | |
| "epoch": 0.89, | |
| "learning_rate": 1.9609760064091044e-07, | |
| "logits/chosen": -2.6273365020751953, | |
| "logits/rejected": -2.597182512283325, | |
| "logps/chosen": -255.65463256835938, | |
| "logps/rejected": -284.71759033203125, | |
| "loss": 0.5495, | |
| "rewards/accuracies": 0.7250000238418579, | |
| "rewards/chosen": -0.941402792930603, | |
| "rewards/margins": 0.628917396068573, | |
| "rewards/rejected": -1.5703203678131104, | |
| "step": 900 | |
| }, | |
| { | |
| "epoch": 0.89, | |
| "eval_logits/chosen": -2.509645938873291, | |
| "eval_logits/rejected": -2.4557979106903076, | |
| "eval_logps/chosen": -253.28036499023438, | |
| "eval_logps/rejected": -298.0242004394531, | |
| "eval_loss": 0.5370938777923584, | |
| "eval_rewards/accuracies": 0.7287735939025879, | |
| "eval_rewards/chosen": -0.932105302810669, | |
| "eval_rewards/margins": 0.7129315733909607, | |
| "eval_rewards/rejected": -1.645037055015564, | |
| "eval_runtime": 286.9177, | |
| "eval_samples_per_second": 1.457, | |
| "eval_steps_per_second": 0.185, | |
| "step": 900 | |
| }, | |
| { | |
| "epoch": 0.9, | |
| "learning_rate": 1.6410414846224992e-07, | |
| "logits/chosen": -2.5519988536834717, | |
| "logits/rejected": -2.4820520877838135, | |
| "logps/chosen": -240.66830444335938, | |
| "logps/rejected": -293.9994812011719, | |
| "loss": 0.4861, | |
| "rewards/accuracies": 0.7875000238418579, | |
| "rewards/chosen": -0.8613694906234741, | |
| "rewards/margins": 0.8020180463790894, | |
| "rewards/rejected": -1.6633875370025635, | |
| "step": 910 | |
| }, | |
| { | |
| "epoch": 0.91, | |
| "learning_rate": 1.348701179601819e-07, | |
| "logits/chosen": -2.6739838123321533, | |
| "logits/rejected": -2.5836825370788574, | |
| "logps/chosen": -261.11077880859375, | |
| "logps/rejected": -305.97698974609375, | |
| "loss": 0.5087, | |
| "rewards/accuracies": 0.7875000238418579, | |
| "rewards/chosen": -0.8714740872383118, | |
| "rewards/margins": 0.7631353139877319, | |
| "rewards/rejected": -1.6346092224121094, | |
| "step": 920 | |
| }, | |
| { | |
| "epoch": 0.91, | |
| "learning_rate": 1.0843004364420151e-07, | |
| "logits/chosen": -2.5310449600219727, | |
| "logits/rejected": -2.465944766998291, | |
| "logps/chosen": -262.35516357421875, | |
| "logps/rejected": -307.0565185546875, | |
| "loss": 0.5616, | |
| "rewards/accuracies": 0.75, | |
| "rewards/chosen": -0.8902053833007812, | |
| "rewards/margins": 0.6049326658248901, | |
| "rewards/rejected": -1.4951379299163818, | |
| "step": 930 | |
| }, | |
| { | |
| "epoch": 0.92, | |
| "learning_rate": 8.481515948997931e-08, | |
| "logits/chosen": -2.6589980125427246, | |
| "logits/rejected": -2.611510753631592, | |
| "logps/chosen": -270.62567138671875, | |
| "logps/rejected": -293.7890319824219, | |
| "loss": 0.6245, | |
| "rewards/accuracies": 0.6499999761581421, | |
| "rewards/chosen": -0.9786608815193176, | |
| "rewards/margins": 0.5622497797012329, | |
| "rewards/rejected": -1.5409104824066162, | |
| "step": 940 | |
| }, | |
| { | |
| "epoch": 0.93, | |
| "learning_rate": 6.4053362042297e-08, | |
| "logits/chosen": -2.5867984294891357, | |
| "logits/rejected": -2.4980480670928955, | |
| "logps/chosen": -251.9007568359375, | |
| "logps/rejected": -304.0977478027344, | |
| "loss": 0.4937, | |
| "rewards/accuracies": 0.762499988079071, | |
| "rewards/chosen": -0.8755489587783813, | |
| "rewards/margins": 0.829258918762207, | |
| "rewards/rejected": -1.7048076391220093, | |
| "step": 950 | |
| }, | |
| { | |
| "epoch": 0.94, | |
| "learning_rate": 4.616917746052163e-08, | |
| "logits/chosen": -2.6321659088134766, | |
| "logits/rejected": -2.556123971939087, | |
| "logps/chosen": -261.781982421875, | |
| "logps/rejected": -302.46527099609375, | |
| "loss": 0.5602, | |
| "rewards/accuracies": 0.699999988079071, | |
| "rewards/chosen": -1.0545774698257446, | |
| "rewards/margins": 0.6820313334465027, | |
| "rewards/rejected": -1.7366087436676025, | |
| "step": 960 | |
| }, | |
| { | |
| "epoch": 0.95, | |
| "learning_rate": 3.118373254556412e-08, | |
| "logits/chosen": -2.625166416168213, | |
| "logits/rejected": -2.547811269760132, | |
| "logps/chosen": -235.4457550048828, | |
| "logps/rejected": -285.9036560058594, | |
| "loss": 0.4703, | |
| "rewards/accuracies": 0.762499988079071, | |
| "rewards/chosen": -0.8371899724006653, | |
| "rewards/margins": 0.8404737710952759, | |
| "rewards/rejected": -1.677663803100586, | |
| "step": 970 | |
| }, | |
| { | |
| "epoch": 0.96, | |
| "learning_rate": 1.9114729782535037e-08, | |
| "logits/chosen": -2.6821470260620117, | |
| "logits/rejected": -2.605581045150757, | |
| "logps/chosen": -240.4625701904297, | |
| "logps/rejected": -286.89263916015625, | |
| "loss": 0.5037, | |
| "rewards/accuracies": 0.637499988079071, | |
| "rewards/chosen": -0.86335289478302, | |
| "rewards/margins": 0.7098163366317749, | |
| "rewards/rejected": -1.5731691122055054, | |
| "step": 980 | |
| }, | |
| { | |
| "epoch": 0.97, | |
| "learning_rate": 9.97642642858815e-09, | |
| "logits/chosen": -2.5950284004211426, | |
| "logits/rejected": -2.50787615776062, | |
| "logps/chosen": -258.5409851074219, | |
| "logps/rejected": -290.9730224609375, | |
| "loss": 0.5398, | |
| "rewards/accuracies": 0.737500011920929, | |
| "rewards/chosen": -0.9514690637588501, | |
| "rewards/margins": 0.6696909666061401, | |
| "rewards/rejected": -1.6211597919464111, | |
| "step": 990 | |
| }, | |
| { | |
| "epoch": 0.98, | |
| "learning_rate": 3.779617670651436e-09, | |
| "logits/chosen": -2.56034517288208, | |
| "logits/rejected": -2.5073485374450684, | |
| "logps/chosen": -267.876953125, | |
| "logps/rejected": -305.6854248046875, | |
| "loss": 0.5948, | |
| "rewards/accuracies": 0.6625000238418579, | |
| "rewards/chosen": -1.0524874925613403, | |
| "rewards/margins": 0.6285669207572937, | |
| "rewards/rejected": -1.6810543537139893, | |
| "step": 1000 | |
| }, | |
| { | |
| "epoch": 0.98, | |
| "eval_logits/chosen": -2.5093116760253906, | |
| "eval_logits/rejected": -2.4554219245910645, | |
| "eval_logps/chosen": -253.41494750976562, | |
| "eval_logps/rejected": -298.073486328125, | |
| "eval_loss": 0.5370525121688843, | |
| "eval_rewards/accuracies": 0.7264150977134705, | |
| "eval_rewards/chosen": -0.9334509372711182, | |
| "eval_rewards/margins": 0.7120789289474487, | |
| "eval_rewards/rejected": -1.6455297470092773, | |
| "eval_runtime": 287.936, | |
| "eval_samples_per_second": 1.452, | |
| "eval_steps_per_second": 0.184, | |
| "step": 1000 | |
| }, | |
| { | |
| "epoch": 0.99, | |
| "learning_rate": 5.316238729444201e-10, | |
| "logits/chosen": -2.580871820449829, | |
| "logits/rejected": -2.4882521629333496, | |
| "logps/chosen": -263.0270690917969, | |
| "logps/rejected": -309.5653076171875, | |
| "loss": 0.4902, | |
| "rewards/accuracies": 0.737500011920929, | |
| "rewards/chosen": -1.025237798690796, | |
| "rewards/margins": 0.790520191192627, | |
| "rewards/rejected": -1.8157583475112915, | |
| "step": 1010 | |
| }, | |
| { | |
| "epoch": 1.0, | |
| "step": 1016, | |
| "total_flos": 0.0, | |
| "train_loss": 0.5672297686573089, | |
| "train_runtime": 12946.9794, | |
| "train_samples_per_second": 0.628, | |
| "train_steps_per_second": 0.078 | |
| } | |
| ], | |
| "logging_steps": 10, | |
| "max_steps": 1016, | |
| "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 | |
| } | |