| { | |
| "best_metric": null, | |
| "best_model_checkpoint": null, | |
| "epoch": 19.45945945945946, | |
| "eval_steps": 500, | |
| "global_step": 540, | |
| "is_hyper_param_search": false, | |
| "is_local_process_zero": true, | |
| "is_world_process_zero": true, | |
| "log_history": [ | |
| { | |
| "epoch": 0.36036036036036034, | |
| "grad_norm": 0.5300200581550598, | |
| "learning_rate": 9.259259259259259e-07, | |
| "logits/chosen": 1.7405741214752197, | |
| "logits/rejected": 1.663368582725525, | |
| "logps/chosen": -97.57881164550781, | |
| "logps/rejected": -70.59793853759766, | |
| "loss": 0.6938, | |
| "rewards/accuracies": 0.4375, | |
| "rewards/chosen": 0.0018835498485714197, | |
| "rewards/margins": -0.0008800366776995361, | |
| "rewards/rejected": 0.0027635858859866858, | |
| "step": 10 | |
| }, | |
| { | |
| "epoch": 0.7207207207207207, | |
| "grad_norm": 0.4883837401866913, | |
| "learning_rate": 1.8518518518518519e-06, | |
| "logits/chosen": 1.8280715942382812, | |
| "logits/rejected": 1.8215343952178955, | |
| "logps/chosen": -90.60624694824219, | |
| "logps/rejected": -79.04981994628906, | |
| "loss": 0.6945, | |
| "rewards/accuracies": 0.4749999940395355, | |
| "rewards/chosen": 0.0014454321935772896, | |
| "rewards/margins": -0.004174981266260147, | |
| "rewards/rejected": 0.005620413459837437, | |
| "step": 20 | |
| }, | |
| { | |
| "epoch": 1.0810810810810811, | |
| "grad_norm": 0.5554109215736389, | |
| "learning_rate": 2.7777777777777783e-06, | |
| "logits/chosen": 1.7980501651763916, | |
| "logits/rejected": 1.841104507446289, | |
| "logps/chosen": -80.78228759765625, | |
| "logps/rejected": -85.08882141113281, | |
| "loss": 0.693, | |
| "rewards/accuracies": 0.4749999940395355, | |
| "rewards/chosen": 0.0042568682692945, | |
| "rewards/margins": -0.0038191028870642185, | |
| "rewards/rejected": 0.008075973019003868, | |
| "step": 30 | |
| }, | |
| { | |
| "epoch": 1.4414414414414414, | |
| "grad_norm": 0.537497341632843, | |
| "learning_rate": 3.7037037037037037e-06, | |
| "logits/chosen": 1.7320470809936523, | |
| "logits/rejected": 1.7411377429962158, | |
| "logps/chosen": -82.24813842773438, | |
| "logps/rejected": -80.90709686279297, | |
| "loss": 0.6952, | |
| "rewards/accuracies": 0.42500001192092896, | |
| "rewards/chosen": -0.00350201572291553, | |
| "rewards/margins": -0.012091752141714096, | |
| "rewards/rejected": 0.008589735254645348, | |
| "step": 40 | |
| }, | |
| { | |
| "epoch": 1.8018018018018018, | |
| "grad_norm": 0.5023094415664673, | |
| "learning_rate": 4.62962962962963e-06, | |
| "logits/chosen": 1.8292573690414429, | |
| "logits/rejected": 1.8632844686508179, | |
| "logps/chosen": -85.98481750488281, | |
| "logps/rejected": -86.14008331298828, | |
| "loss": 0.6928, | |
| "rewards/accuracies": 0.4124999940395355, | |
| "rewards/chosen": 0.00168000184930861, | |
| "rewards/margins": -0.011136507615447044, | |
| "rewards/rejected": 0.01281650923192501, | |
| "step": 50 | |
| }, | |
| { | |
| "epoch": 2.1621621621621623, | |
| "grad_norm": 0.4981901943683624, | |
| "learning_rate": 4.998119881260576e-06, | |
| "logits/chosen": 1.8536121845245361, | |
| "logits/rejected": 1.7948782444000244, | |
| "logps/chosen": -90.0439682006836, | |
| "logps/rejected": -79.8309555053711, | |
| "loss": 0.6914, | |
| "rewards/accuracies": 0.48750001192092896, | |
| "rewards/chosen": 0.0025153260212391615, | |
| "rewards/margins": -0.0021209525875747204, | |
| "rewards/rejected": 0.004636278375983238, | |
| "step": 60 | |
| }, | |
| { | |
| "epoch": 2.5225225225225225, | |
| "grad_norm": 0.6208468079566956, | |
| "learning_rate": 4.9866405060165044e-06, | |
| "logits/chosen": 1.7845121622085571, | |
| "logits/rejected": 1.8721017837524414, | |
| "logps/chosen": -74.6461181640625, | |
| "logps/rejected": -98.91078186035156, | |
| "loss": 0.694, | |
| "rewards/accuracies": 0.5249999761581421, | |
| "rewards/chosen": 0.00022823773906566203, | |
| "rewards/margins": 0.0024594543501734734, | |
| "rewards/rejected": -0.00223121652379632, | |
| "step": 70 | |
| }, | |
| { | |
| "epoch": 2.8828828828828827, | |
| "grad_norm": 0.5074446797370911, | |
| "learning_rate": 4.964774158361991e-06, | |
| "logits/chosen": 1.8522154092788696, | |
| "logits/rejected": 1.8161392211914062, | |
| "logps/chosen": -89.16864013671875, | |
| "logps/rejected": -81.97349548339844, | |
| "loss": 0.6911, | |
| "rewards/accuracies": 0.5249999761581421, | |
| "rewards/chosen": 0.006215038243681192, | |
| "rewards/margins": 0.005521883722394705, | |
| "rewards/rejected": 0.0006931538810022175, | |
| "step": 80 | |
| }, | |
| { | |
| "epoch": 3.2432432432432434, | |
| "grad_norm": 0.6148263812065125, | |
| "learning_rate": 4.93261217644956e-06, | |
| "logits/chosen": 1.7915083169937134, | |
| "logits/rejected": 1.7579513788223267, | |
| "logps/chosen": -88.66618347167969, | |
| "logps/rejected": -80.2238998413086, | |
| "loss": 0.6918, | |
| "rewards/accuracies": 0.574999988079071, | |
| "rewards/chosen": 0.005774274934083223, | |
| "rewards/margins": 0.016623441129922867, | |
| "rewards/rejected": -0.010849165730178356, | |
| "step": 90 | |
| }, | |
| { | |
| "epoch": 3.6036036036036037, | |
| "grad_norm": 0.6143240332603455, | |
| "learning_rate": 4.8902889044347e-06, | |
| "logits/chosen": 1.6886920928955078, | |
| "logits/rejected": 1.8121178150177002, | |
| "logps/chosen": -72.908203125, | |
| "logps/rejected": -92.44017028808594, | |
| "loss": 0.6911, | |
| "rewards/accuracies": 0.48750001192092896, | |
| "rewards/chosen": -0.0015121791511774063, | |
| "rewards/margins": 0.0036420777905732393, | |
| "rewards/rejected": -0.005154256708920002, | |
| "step": 100 | |
| }, | |
| { | |
| "epoch": 3.963963963963964, | |
| "grad_norm": 0.7140340209007263, | |
| "learning_rate": 4.837981131305475e-06, | |
| "logits/chosen": 1.7481105327606201, | |
| "logits/rejected": 1.723141074180603, | |
| "logps/chosen": -80.63452911376953, | |
| "logps/rejected": -72.64492797851562, | |
| "loss": 0.6894, | |
| "rewards/accuracies": 0.48750001192092896, | |
| "rewards/chosen": -0.006246576085686684, | |
| "rewards/margins": -0.0035218377597630024, | |
| "rewards/rejected": -0.0027247383259236813, | |
| "step": 110 | |
| }, | |
| { | |
| "epoch": 4.324324324324325, | |
| "grad_norm": 1.1645854711532593, | |
| "learning_rate": 4.775907352415367e-06, | |
| "logits/chosen": 1.7416290044784546, | |
| "logits/rejected": 1.8237574100494385, | |
| "logps/chosen": -89.01248931884766, | |
| "logps/rejected": -92.31901550292969, | |
| "loss": 0.6869, | |
| "rewards/accuracies": 0.612500011920929, | |
| "rewards/chosen": -0.008855113759636879, | |
| "rewards/margins": 0.013151508755981922, | |
| "rewards/rejected": -0.022006623446941376, | |
| "step": 120 | |
| }, | |
| { | |
| "epoch": 4.684684684684685, | |
| "grad_norm": 0.6638664603233337, | |
| "learning_rate": 4.70432685680402e-06, | |
| "logits/chosen": 1.7124770879745483, | |
| "logits/rejected": 1.7777938842773438, | |
| "logps/chosen": -84.87271881103516, | |
| "logps/rejected": -92.4839096069336, | |
| "loss": 0.6855, | |
| "rewards/accuracies": 0.574999988079071, | |
| "rewards/chosen": -0.004437069408595562, | |
| "rewards/margins": 0.019811339676380157, | |
| "rewards/rejected": -0.024248410016298294, | |
| "step": 130 | |
| }, | |
| { | |
| "epoch": 5.045045045045045, | |
| "grad_norm": 0.785169780254364, | |
| "learning_rate": 4.623538644118244e-06, | |
| "logits/chosen": 1.7838690280914307, | |
| "logits/rejected": 1.8052574396133423, | |
| "logps/chosen": -94.57842254638672, | |
| "logps/rejected": -80.6390609741211, | |
| "loss": 0.6836, | |
| "rewards/accuracies": 0.48750001192092896, | |
| "rewards/chosen": -0.026147600263357162, | |
| "rewards/margins": 0.010936126112937927, | |
| "rewards/rejected": -0.03708372637629509, | |
| "step": 140 | |
| }, | |
| { | |
| "epoch": 5.405405405405405, | |
| "grad_norm": 0.745952844619751, | |
| "learning_rate": 4.533880175657419e-06, | |
| "logits/chosen": 1.7925735712051392, | |
| "logits/rejected": 1.7987396717071533, | |
| "logps/chosen": -82.4854507446289, | |
| "logps/rejected": -77.50787353515625, | |
| "loss": 0.6767, | |
| "rewards/accuracies": 0.625, | |
| "rewards/chosen": -0.013469865545630455, | |
| "rewards/margins": 0.04023212194442749, | |
| "rewards/rejected": -0.053701985627412796, | |
| "step": 150 | |
| }, | |
| { | |
| "epoch": 5.7657657657657655, | |
| "grad_norm": 0.7885063290596008, | |
| "learning_rate": 4.435725964760331e-06, | |
| "logits/chosen": 1.7725191116333008, | |
| "logits/rejected": 1.8345234394073486, | |
| "logps/chosen": -78.7154541015625, | |
| "logps/rejected": -80.75708770751953, | |
| "loss": 0.6721, | |
| "rewards/accuracies": 0.675000011920929, | |
| "rewards/chosen": -0.018351273611187935, | |
| "rewards/margins": 0.05368901416659355, | |
| "rewards/rejected": -0.07204028964042664, | |
| "step": 160 | |
| }, | |
| { | |
| "epoch": 6.126126126126126, | |
| "grad_norm": 0.7734187245368958, | |
| "learning_rate": 4.329486012421531e-06, | |
| "logits/chosen": 1.827528715133667, | |
| "logits/rejected": 1.7939176559448242, | |
| "logps/chosen": -75.33561706542969, | |
| "logps/rejected": -71.34326934814453, | |
| "loss": 0.6775, | |
| "rewards/accuracies": 0.612500011920929, | |
| "rewards/chosen": -0.038218818604946136, | |
| "rewards/margins": 0.06159939616918564, | |
| "rewards/rejected": -0.09981821477413177, | |
| "step": 170 | |
| }, | |
| { | |
| "epoch": 6.486486486486487, | |
| "grad_norm": 0.7796682715415955, | |
| "learning_rate": 4.215604094671835e-06, | |
| "logits/chosen": 1.729288101196289, | |
| "logits/rejected": 1.7497504949569702, | |
| "logps/chosen": -83.02750396728516, | |
| "logps/rejected": -78.53959655761719, | |
| "loss": 0.6632, | |
| "rewards/accuracies": 0.699999988079071, | |
| "rewards/chosen": -0.03207743167877197, | |
| "rewards/margins": 0.06858544796705246, | |
| "rewards/rejected": -0.10066288709640503, | |
| "step": 180 | |
| }, | |
| { | |
| "epoch": 6.846846846846847, | |
| "grad_norm": 0.8168752193450928, | |
| "learning_rate": 4.094555908876765e-06, | |
| "logits/chosen": 1.747865080833435, | |
| "logits/rejected": 1.7993282079696655, | |
| "logps/chosen": -78.75782775878906, | |
| "logps/rejected": -85.8946304321289, | |
| "loss": 0.668, | |
| "rewards/accuracies": 0.612500011920929, | |
| "rewards/chosen": -0.0766468197107315, | |
| "rewards/margins": 0.045263100415468216, | |
| "rewards/rejected": -0.12190990149974823, | |
| "step": 190 | |
| }, | |
| { | |
| "epoch": 7.207207207207207, | |
| "grad_norm": 0.9816763997077942, | |
| "learning_rate": 3.966847086696045e-06, | |
| "logits/chosen": 1.792106032371521, | |
| "logits/rejected": 1.7798125743865967, | |
| "logps/chosen": -93.53439331054688, | |
| "logps/rejected": -73.8904037475586, | |
| "loss": 0.6582, | |
| "rewards/accuracies": 0.6875, | |
| "rewards/chosen": -0.06042450666427612, | |
| "rewards/margins": 0.10538975894451141, | |
| "rewards/rejected": -0.16581428050994873, | |
| "step": 200 | |
| }, | |
| { | |
| "epoch": 7.5675675675675675, | |
| "grad_norm": 0.9053374528884888, | |
| "learning_rate": 3.833011082004229e-06, | |
| "logits/chosen": 1.7648969888687134, | |
| "logits/rejected": 1.7725557088851929, | |
| "logps/chosen": -79.61383056640625, | |
| "logps/rejected": -71.8048095703125, | |
| "loss": 0.6551, | |
| "rewards/accuracies": 0.6875, | |
| "rewards/chosen": -0.040038011968135834, | |
| "rewards/margins": 0.11104954779148102, | |
| "rewards/rejected": -0.15108755230903625, | |
| "step": 210 | |
| }, | |
| { | |
| "epoch": 7.927927927927928, | |
| "grad_norm": 1.1332669258117676, | |
| "learning_rate": 3.693606942594873e-06, | |
| "logits/chosen": 1.8605678081512451, | |
| "logits/rejected": 1.91280996799469, | |
| "logps/chosen": -74.39783477783203, | |
| "logps/rejected": -90.66893005371094, | |
| "loss": 0.6565, | |
| "rewards/accuracies": 0.637499988079071, | |
| "rewards/chosen": -0.08056517690420151, | |
| "rewards/margins": 0.0968993678689003, | |
| "rewards/rejected": -0.1774645447731018, | |
| "step": 220 | |
| }, | |
| { | |
| "epoch": 8.288288288288289, | |
| "grad_norm": 1.0496221780776978, | |
| "learning_rate": 3.549216974976073e-06, | |
| "logits/chosen": 1.8636070489883423, | |
| "logits/rejected": 1.7996151447296143, | |
| "logps/chosen": -89.9018783569336, | |
| "logps/rejected": -74.14125061035156, | |
| "loss": 0.6423, | |
| "rewards/accuracies": 0.6875, | |
| "rewards/chosen": -0.05598217993974686, | |
| "rewards/margins": 0.14807265996932983, | |
| "rewards/rejected": -0.2040548026561737, | |
| "step": 230 | |
| }, | |
| { | |
| "epoch": 8.64864864864865, | |
| "grad_norm": 0.8506491780281067, | |
| "learning_rate": 3.400444312011776e-06, | |
| "logits/chosen": 1.7753232717514038, | |
| "logits/rejected": 1.7647806406021118, | |
| "logps/chosen": -88.50201416015625, | |
| "logps/rejected": -81.80894470214844, | |
| "loss": 0.6358, | |
| "rewards/accuracies": 0.6625000238418579, | |
| "rewards/chosen": -0.09656485170125961, | |
| "rewards/margins": 0.11554913222789764, | |
| "rewards/rejected": -0.21211397647857666, | |
| "step": 240 | |
| }, | |
| { | |
| "epoch": 9.00900900900901, | |
| "grad_norm": 0.9554071426391602, | |
| "learning_rate": 3.2479103935691047e-06, | |
| "logits/chosen": 1.7528200149536133, | |
| "logits/rejected": 1.7827808856964111, | |
| "logps/chosen": -90.01268005371094, | |
| "logps/rejected": -89.0714340209961, | |
| "loss": 0.6374, | |
| "rewards/accuracies": 0.675000011920929, | |
| "rewards/chosen": -0.07412372529506683, | |
| "rewards/margins": 0.1321595013141632, | |
| "rewards/rejected": -0.20628324151039124, | |
| "step": 250 | |
| }, | |
| { | |
| "epoch": 9.36936936936937, | |
| "grad_norm": 0.9363342523574829, | |
| "learning_rate": 3.092252370695298e-06, | |
| "logits/chosen": 1.6990101337432861, | |
| "logits/rejected": 1.8281656503677368, | |
| "logps/chosen": -71.15428161621094, | |
| "logps/rejected": -79.72063446044922, | |
| "loss": 0.6303, | |
| "rewards/accuracies": 0.762499988079071, | |
| "rewards/chosen": -0.08240491151809692, | |
| "rewards/margins": 0.16954098641872406, | |
| "rewards/rejected": -0.2519459128379822, | |
| "step": 260 | |
| }, | |
| { | |
| "epoch": 9.72972972972973, | |
| "grad_norm": 1.023535966873169, | |
| "learning_rate": 2.9341204441673267e-06, | |
| "logits/chosen": 1.7418006658554077, | |
| "logits/rejected": 1.7500028610229492, | |
| "logps/chosen": -82.24595642089844, | |
| "logps/rejected": -91.16307067871094, | |
| "loss": 0.6321, | |
| "rewards/accuracies": 0.6875, | |
| "rewards/chosen": -0.13560767471790314, | |
| "rewards/margins": 0.1204828992486, | |
| "rewards/rejected": -0.25609058141708374, | |
| "step": 270 | |
| }, | |
| { | |
| "epoch": 10.09009009009009, | |
| "grad_norm": 1.0393471717834473, | |
| "learning_rate": 2.7741751485313295e-06, | |
| "logits/chosen": 1.8437154293060303, | |
| "logits/rejected": 1.8877627849578857, | |
| "logps/chosen": -68.61182403564453, | |
| "logps/rejected": -77.12557220458984, | |
| "loss": 0.6109, | |
| "rewards/accuracies": 0.737500011920929, | |
| "rewards/chosen": -0.11025551706552505, | |
| "rewards/margins": 0.14969798922538757, | |
| "rewards/rejected": -0.25995349884033203, | |
| "step": 280 | |
| }, | |
| { | |
| "epoch": 10.45045045045045, | |
| "grad_norm": 0.9554691910743713, | |
| "learning_rate": 2.6130845929767662e-06, | |
| "logits/chosen": 1.7998764514923096, | |
| "logits/rejected": 1.796451210975647, | |
| "logps/chosen": -72.6626205444336, | |
| "logps/rejected": -74.15937042236328, | |
| "loss": 0.6123, | |
| "rewards/accuracies": 0.737500011920929, | |
| "rewards/chosen": -0.13948099315166473, | |
| "rewards/margins": 0.18039894104003906, | |
| "rewards/rejected": -0.319879949092865, | |
| "step": 290 | |
| }, | |
| { | |
| "epoch": 10.81081081081081, | |
| "grad_norm": 1.1667226552963257, | |
| "learning_rate": 2.4515216705704396e-06, | |
| "logits/chosen": 1.8437564373016357, | |
| "logits/rejected": 1.9043811559677124, | |
| "logps/chosen": -78.18916320800781, | |
| "logps/rejected": -90.88524627685547, | |
| "loss": 0.6218, | |
| "rewards/accuracies": 0.7875000238418579, | |
| "rewards/chosen": -0.12091793864965439, | |
| "rewards/margins": 0.20490169525146484, | |
| "rewards/rejected": -0.32581964135169983, | |
| "step": 300 | |
| }, | |
| { | |
| "epoch": 11.17117117117117, | |
| "grad_norm": 1.2568660974502563, | |
| "learning_rate": 2.290161247507733e-06, | |
| "logits/chosen": 1.7295608520507812, | |
| "logits/rejected": 1.8652616739273071, | |
| "logps/chosen": -79.8567886352539, | |
| "logps/rejected": -87.59798431396484, | |
| "loss": 0.6067, | |
| "rewards/accuracies": 0.8374999761581421, | |
| "rewards/chosen": -0.10961383581161499, | |
| "rewards/margins": 0.2835196554660797, | |
| "rewards/rejected": -0.3931335210800171, | |
| "step": 310 | |
| }, | |
| { | |
| "epoch": 11.531531531531531, | |
| "grad_norm": 1.115491271018982, | |
| "learning_rate": 2.129677344121879e-06, | |
| "logits/chosen": 1.803063988685608, | |
| "logits/rejected": 1.820227861404419, | |
| "logps/chosen": -74.64582061767578, | |
| "logps/rejected": -72.14736938476562, | |
| "loss": 0.6113, | |
| "rewards/accuracies": 0.75, | |
| "rewards/chosen": -0.09987158328294754, | |
| "rewards/margins": 0.2059938907623291, | |
| "rewards/rejected": -0.30586546659469604, | |
| "step": 320 | |
| }, | |
| { | |
| "epoch": 11.891891891891891, | |
| "grad_norm": 0.9227492213249207, | |
| "learning_rate": 1.970740319426474e-06, | |
| "logits/chosen": 1.7561572790145874, | |
| "logits/rejected": 1.789878249168396, | |
| "logps/chosen": -85.27911376953125, | |
| "logps/rejected": -77.4220962524414, | |
| "loss": 0.5891, | |
| "rewards/accuracies": 0.7875000238418579, | |
| "rewards/chosen": -0.14200571179389954, | |
| "rewards/margins": 0.22272682189941406, | |
| "rewards/rejected": -0.3647325336933136, | |
| "step": 330 | |
| }, | |
| { | |
| "epoch": 12.252252252252251, | |
| "grad_norm": 1.541570782661438, | |
| "learning_rate": 1.8140140709517467e-06, | |
| "logits/chosen": 1.7364771366119385, | |
| "logits/rejected": 1.7182557582855225, | |
| "logps/chosen": -76.05533599853516, | |
| "logps/rejected": -81.50830078125, | |
| "loss": 0.5994, | |
| "rewards/accuracies": 0.75, | |
| "rewards/chosen": -0.12165775150060654, | |
| "rewards/margins": 0.19019995629787445, | |
| "rewards/rejected": -0.3118577301502228, | |
| "step": 340 | |
| }, | |
| { | |
| "epoch": 12.612612612612612, | |
| "grad_norm": 1.1125155687332153, | |
| "learning_rate": 1.6601532615711452e-06, | |
| "logits/chosen": 1.7223026752471924, | |
| "logits/rejected": 1.8163812160491943, | |
| "logps/chosen": -71.94869232177734, | |
| "logps/rejected": -90.12442016601562, | |
| "loss": 0.6013, | |
| "rewards/accuracies": 0.7124999761581421, | |
| "rewards/chosen": -0.16840626299381256, | |
| "rewards/margins": 0.1776614487171173, | |
| "rewards/rejected": -0.3460676968097687, | |
| "step": 350 | |
| }, | |
| { | |
| "epoch": 12.972972972972974, | |
| "grad_norm": 1.0530396699905396, | |
| "learning_rate": 1.509800584902108e-06, | |
| "logits/chosen": 1.795819878578186, | |
| "logits/rejected": 1.8446025848388672, | |
| "logps/chosen": -82.24161529541016, | |
| "logps/rejected": -79.51258850097656, | |
| "loss": 0.5857, | |
| "rewards/accuracies": 0.800000011920929, | |
| "rewards/chosen": -0.13156726956367493, | |
| "rewards/margins": 0.2468404471874237, | |
| "rewards/rejected": -0.37840771675109863, | |
| "step": 360 | |
| }, | |
| { | |
| "epoch": 13.333333333333334, | |
| "grad_norm": 1.0298306941986084, | |
| "learning_rate": 1.3635840807037487e-06, | |
| "logits/chosen": 1.8054157495498657, | |
| "logits/rejected": 1.8246568441390991, | |
| "logps/chosen": -85.40760803222656, | |
| "logps/rejected": -78.77482604980469, | |
| "loss": 0.5943, | |
| "rewards/accuracies": 0.762499988079071, | |
| "rewards/chosen": -0.10383795201778412, | |
| "rewards/margins": 0.2493254393339157, | |
| "rewards/rejected": -0.35316339135169983, | |
| "step": 370 | |
| }, | |
| { | |
| "epoch": 13.693693693693694, | |
| "grad_norm": 1.0601342916488647, | |
| "learning_rate": 1.2221145114853172e-06, | |
| "logits/chosen": 1.6507200002670288, | |
| "logits/rejected": 1.720569372177124, | |
| "logps/chosen": -74.90982055664062, | |
| "logps/rejected": -92.51995849609375, | |
| "loss": 0.5789, | |
| "rewards/accuracies": 0.7875000238418579, | |
| "rewards/chosen": -0.13705268502235413, | |
| "rewards/margins": 0.2740176022052765, | |
| "rewards/rejected": -0.4110702872276306, | |
| "step": 380 | |
| }, | |
| { | |
| "epoch": 14.054054054054054, | |
| "grad_norm": 1.1135063171386719, | |
| "learning_rate": 1.085982811283654e-06, | |
| "logits/chosen": 1.7133989334106445, | |
| "logits/rejected": 1.7338130474090576, | |
| "logps/chosen": -80.55327606201172, | |
| "logps/rejected": -82.32875061035156, | |
| "loss": 0.5847, | |
| "rewards/accuracies": 0.8500000238418579, | |
| "rewards/chosen": -0.12969203293323517, | |
| "rewards/margins": 0.26957422494888306, | |
| "rewards/rejected": -0.39926621317863464, | |
| "step": 390 | |
| }, | |
| { | |
| "epoch": 14.414414414414415, | |
| "grad_norm": 1.0821375846862793, | |
| "learning_rate": 9.557576172663577e-07, | |
| "logits/chosen": 1.7035369873046875, | |
| "logits/rejected": 1.7184028625488281, | |
| "logps/chosen": -80.99266052246094, | |
| "logps/rejected": -85.05413818359375, | |
| "loss": 0.583, | |
| "rewards/accuracies": 0.824999988079071, | |
| "rewards/chosen": -0.1366995871067047, | |
| "rewards/margins": 0.268393874168396, | |
| "rewards/rejected": -0.4050934910774231, | |
| "step": 400 | |
| }, | |
| { | |
| "epoch": 14.774774774774775, | |
| "grad_norm": 1.1167497634887695, | |
| "learning_rate": 8.319828944714508e-07, | |
| "logits/chosen": 1.8293631076812744, | |
| "logits/rejected": 1.7202396392822266, | |
| "logps/chosen": -96.96324157714844, | |
| "logps/rejected": -80.58685302734375, | |
| "loss": 0.5768, | |
| "rewards/accuracies": 0.7875000238418579, | |
| "rewards/chosen": -0.14515718817710876, | |
| "rewards/margins": 0.27918386459350586, | |
| "rewards/rejected": -0.4243410527706146, | |
| "step": 410 | |
| }, | |
| { | |
| "epoch": 15.135135135135135, | |
| "grad_norm": 1.0140223503112793, | |
| "learning_rate": 7.151756636052529e-07, | |
| "logits/chosen": 1.7751014232635498, | |
| "logits/rejected": 1.7394100427627563, | |
| "logps/chosen": -104.48625183105469, | |
| "logps/rejected": -90.3105239868164, | |
| "loss": 0.5804, | |
| "rewards/accuracies": 0.7749999761581421, | |
| "rewards/chosen": -0.14849844574928284, | |
| "rewards/margins": 0.2261633574962616, | |
| "rewards/rejected": -0.37466174364089966, | |
| "step": 420 | |
| }, | |
| { | |
| "epoch": 15.495495495495495, | |
| "grad_norm": 1.0056337118148804, | |
| "learning_rate": 6.058238413897052e-07, | |
| "logits/chosen": 1.7368510961532593, | |
| "logits/rejected": 1.844665288925171, | |
| "logps/chosen": -88.00404357910156, | |
| "logps/rejected": -103.71427917480469, | |
| "loss": 0.5786, | |
| "rewards/accuracies": 0.7875000238418579, | |
| "rewards/chosen": -0.2264096438884735, | |
| "rewards/margins": 0.2332192212343216, | |
| "rewards/rejected": -0.4596289098262787, | |
| "step": 430 | |
| }, | |
| { | |
| "epoch": 15.855855855855856, | |
| "grad_norm": 1.0402743816375732, | |
| "learning_rate": 5.043842024802675e-07, | |
| "logits/chosen": 1.835862398147583, | |
| "logits/rejected": 1.8621985912322998, | |
| "logps/chosen": -80.8730697631836, | |
| "logps/rejected": -77.32243347167969, | |
| "loss": 0.5737, | |
| "rewards/accuracies": 0.762499988079071, | |
| "rewards/chosen": -0.10934928804636002, | |
| "rewards/margins": 0.26070791482925415, | |
| "rewards/rejected": -0.3700571656227112, | |
| "step": 440 | |
| }, | |
| { | |
| "epoch": 16.216216216216218, | |
| "grad_norm": 1.0037925243377686, | |
| "learning_rate": 4.1128047146765936e-07, | |
| "logits/chosen": 1.6536343097686768, | |
| "logits/rejected": 1.748152494430542, | |
| "logps/chosen": -85.28504180908203, | |
| "logps/rejected": -83.04313659667969, | |
| "loss": 0.5869, | |
| "rewards/accuracies": 0.8125, | |
| "rewards/chosen": -0.1468530297279358, | |
| "rewards/margins": 0.24461284279823303, | |
| "rewards/rejected": -0.39146584272384644, | |
| "step": 450 | |
| }, | |
| { | |
| "epoch": 16.576576576576578, | |
| "grad_norm": 1.0473060607910156, | |
| "learning_rate": 3.269015529333805e-07, | |
| "logits/chosen": 1.7145726680755615, | |
| "logits/rejected": 1.7245705127716064, | |
| "logps/chosen": -68.06051635742188, | |
| "logps/rejected": -77.09696960449219, | |
| "loss": 0.5679, | |
| "rewards/accuracies": 0.7749999761581421, | |
| "rewards/chosen": -0.16770952939987183, | |
| "rewards/margins": 0.23443298041820526, | |
| "rewards/rejected": -0.4021424651145935, | |
| "step": 460 | |
| }, | |
| { | |
| "epoch": 16.936936936936938, | |
| "grad_norm": 1.0985324382781982, | |
| "learning_rate": 2.515999069522676e-07, | |
| "logits/chosen": 1.7093912363052368, | |
| "logits/rejected": 1.6903793811798096, | |
| "logps/chosen": -79.38801574707031, | |
| "logps/rejected": -74.43864440917969, | |
| "loss": 0.5754, | |
| "rewards/accuracies": 0.8125, | |
| "rewards/chosen": -0.14471781253814697, | |
| "rewards/margins": 0.24466648697853088, | |
| "rewards/rejected": -0.38938429951667786, | |
| "step": 470 | |
| }, | |
| { | |
| "epoch": 17.2972972972973, | |
| "grad_norm": 1.1457403898239136, | |
| "learning_rate": 1.8569007682777417e-07, | |
| "logits/chosen": 1.6642320156097412, | |
| "logits/rejected": 1.7295513153076172, | |
| "logps/chosen": -80.79732513427734, | |
| "logps/rejected": -80.07367706298828, | |
| "loss": 0.5749, | |
| "rewards/accuracies": 0.7250000238418579, | |
| "rewards/chosen": -0.16041259467601776, | |
| "rewards/margins": 0.26678892970085144, | |
| "rewards/rejected": -0.4272015690803528, | |
| "step": 480 | |
| }, | |
| { | |
| "epoch": 17.65765765765766, | |
| "grad_norm": 1.04747474193573, | |
| "learning_rate": 1.2944737520980883e-07, | |
| "logits/chosen": 1.699061632156372, | |
| "logits/rejected": 1.837684988975525, | |
| "logps/chosen": -73.55027770996094, | |
| "logps/rejected": -93.25712585449219, | |
| "loss": 0.5732, | |
| "rewards/accuracies": 0.7749999761581421, | |
| "rewards/chosen": -0.20150330662727356, | |
| "rewards/margins": 0.2376147210597992, | |
| "rewards/rejected": -0.43911799788475037, | |
| "step": 490 | |
| }, | |
| { | |
| "epoch": 18.01801801801802, | |
| "grad_norm": 1.033074140548706, | |
| "learning_rate": 8.310673408334496e-08, | |
| "logits/chosen": 1.655987024307251, | |
| "logits/rejected": 1.6537258625030518, | |
| "logps/chosen": -103.11566162109375, | |
| "logps/rejected": -96.29421997070312, | |
| "loss": 0.5747, | |
| "rewards/accuracies": 0.800000011920929, | |
| "rewards/chosen": -0.15963700413703918, | |
| "rewards/margins": 0.30574050545692444, | |
| "rewards/rejected": -0.4653775095939636, | |
| "step": 500 | |
| }, | |
| { | |
| "epoch": 18.37837837837838, | |
| "grad_norm": 1.0928398370742798, | |
| "learning_rate": 4.6861723431538273e-08, | |
| "logits/chosen": 1.7491756677627563, | |
| "logits/rejected": 1.7941337823867798, | |
| "logps/chosen": -79.0124282836914, | |
| "logps/rejected": -94.39396667480469, | |
| "loss": 0.5745, | |
| "rewards/accuracies": 0.824999988079071, | |
| "rewards/chosen": -0.14681483805179596, | |
| "rewards/margins": 0.31730785965919495, | |
| "rewards/rejected": -0.4641226828098297, | |
| "step": 510 | |
| }, | |
| { | |
| "epoch": 18.73873873873874, | |
| "grad_norm": 1.0533796548843384, | |
| "learning_rate": 2.0863742672497244e-08, | |
| "logits/chosen": 1.7281211614608765, | |
| "logits/rejected": 1.7764543294906616, | |
| "logps/chosen": -91.92194366455078, | |
| "logps/rejected": -85.10621643066406, | |
| "loss": 0.5726, | |
| "rewards/accuracies": 0.824999988079071, | |
| "rewards/chosen": -0.15819533169269562, | |
| "rewards/margins": 0.2960740923881531, | |
| "rewards/rejected": -0.4542694687843323, | |
| "step": 520 | |
| }, | |
| { | |
| "epoch": 19.0990990990991, | |
| "grad_norm": 1.1859172582626343, | |
| "learning_rate": 5.221388247169945e-09, | |
| "logits/chosen": 1.7769553661346436, | |
| "logits/rejected": 1.7326488494873047, | |
| "logps/chosen": -85.6234359741211, | |
| "logps/rejected": -75.64092254638672, | |
| "loss": 0.5688, | |
| "rewards/accuracies": 0.875, | |
| "rewards/chosen": -0.15740445256233215, | |
| "rewards/margins": 0.31495219469070435, | |
| "rewards/rejected": -0.4723566472530365, | |
| "step": 530 | |
| }, | |
| { | |
| "epoch": 19.45945945945946, | |
| "grad_norm": 1.1295475959777832, | |
| "learning_rate": 0.0, | |
| "logits/chosen": 1.7790734767913818, | |
| "logits/rejected": 1.8196531534194946, | |
| "logps/chosen": -83.07762145996094, | |
| "logps/rejected": -77.76747131347656, | |
| "loss": 0.575, | |
| "rewards/accuracies": 0.7875000238418579, | |
| "rewards/chosen": -0.1493213176727295, | |
| "rewards/margins": 0.2750301659107208, | |
| "rewards/rejected": -0.42435145378112793, | |
| "step": 540 | |
| }, | |
| { | |
| "epoch": 19.45945945945946, | |
| "step": 540, | |
| "total_flos": 1.8293058049230766e+18, | |
| "train_loss": 0.6298594880987096, | |
| "train_runtime": 3921.3478, | |
| "train_samples_per_second": 9.033, | |
| "train_steps_per_second": 0.138 | |
| } | |
| ], | |
| "logging_steps": 10, | |
| "max_steps": 540, | |
| "num_input_tokens_seen": 0, | |
| "num_train_epochs": 20, | |
| "save_steps": 50, | |
| "stateful_callbacks": { | |
| "TrainerControl": { | |
| "args": { | |
| "should_epoch_stop": false, | |
| "should_evaluate": false, | |
| "should_log": false, | |
| "should_save": true, | |
| "should_training_stop": true | |
| }, | |
| "attributes": {} | |
| } | |
| }, | |
| "total_flos": 1.8293058049230766e+18, | |
| "train_batch_size": 1, | |
| "trial_name": null, | |
| "trial_params": null | |
| } | |