| { | |
| "best_metric": null, | |
| "best_model_checkpoint": null, | |
| "epoch": 0.9996728819103696, | |
| "eval_steps": 100, | |
| "global_step": 1528, | |
| "is_hyper_param_search": false, | |
| "is_local_process_zero": true, | |
| "is_world_process_zero": true, | |
| "log_history": [ | |
| { | |
| "epoch": 0.0006542361792607131, | |
| "grad_norm": 7.53125, | |
| "learning_rate": 3.267973856209151e-08, | |
| "logits/chosen": -2.690979480743408, | |
| "logits/rejected": -2.4915528297424316, | |
| "logps/chosen": -306.9772644042969, | |
| "logps/rejected": -274.77850341796875, | |
| "loss": 1.3863, | |
| "rewards/accuracies": 0.0, | |
| "rewards/chosen": 0.0, | |
| "rewards/margins": 0.0, | |
| "rewards/rejected": 0.0, | |
| "step": 1 | |
| }, | |
| { | |
| "epoch": 0.006542361792607131, | |
| "grad_norm": 6.75, | |
| "learning_rate": 3.267973856209151e-07, | |
| "logits/chosen": -2.5390172004699707, | |
| "logits/rejected": -2.4508614540100098, | |
| "logps/chosen": -286.64813232421875, | |
| "logps/rejected": -271.6336669921875, | |
| "loss": 1.3872, | |
| "rewards/accuracies": 0.4305555522441864, | |
| "rewards/chosen": 0.002395547926425934, | |
| "rewards/margins": 0.0021986099891364574, | |
| "rewards/rejected": 0.0001969372679013759, | |
| "step": 10 | |
| }, | |
| { | |
| "epoch": 0.013084723585214262, | |
| "grad_norm": 7.4375, | |
| "learning_rate": 6.535947712418302e-07, | |
| "logits/chosen": -2.479840040206909, | |
| "logits/rejected": -2.3548035621643066, | |
| "logps/chosen": -268.228759765625, | |
| "logps/rejected": -197.14610290527344, | |
| "loss": 1.3853, | |
| "rewards/accuracies": 0.5249999761581421, | |
| "rewards/chosen": 0.003464195877313614, | |
| "rewards/margins": -0.0019345780601724982, | |
| "rewards/rejected": 0.005398774053901434, | |
| "step": 20 | |
| }, | |
| { | |
| "epoch": 0.019627085377821395, | |
| "grad_norm": 8.4375, | |
| "learning_rate": 9.80392156862745e-07, | |
| "logits/chosen": -2.4206929206848145, | |
| "logits/rejected": -2.365497350692749, | |
| "logps/chosen": -258.31414794921875, | |
| "logps/rejected": -255.288818359375, | |
| "loss": 1.3826, | |
| "rewards/accuracies": 0.637499988079071, | |
| "rewards/chosen": 0.025201931595802307, | |
| "rewards/margins": 0.022868353873491287, | |
| "rewards/rejected": 0.0023335753940045834, | |
| "step": 30 | |
| }, | |
| { | |
| "epoch": 0.026169447170428524, | |
| "grad_norm": 7.09375, | |
| "learning_rate": 1.3071895424836604e-06, | |
| "logits/chosen": -2.370481252670288, | |
| "logits/rejected": -2.3122944831848145, | |
| "logps/chosen": -273.23687744140625, | |
| "logps/rejected": -237.98342895507812, | |
| "loss": 1.3774, | |
| "rewards/accuracies": 0.6000000238418579, | |
| "rewards/chosen": 0.017533209174871445, | |
| "rewards/margins": 0.031263187527656555, | |
| "rewards/rejected": -0.01372998021543026, | |
| "step": 40 | |
| }, | |
| { | |
| "epoch": 0.03271180896303565, | |
| "grad_norm": 7.25, | |
| "learning_rate": 1.6339869281045753e-06, | |
| "logits/chosen": -2.4228413105010986, | |
| "logits/rejected": -2.365399122238159, | |
| "logps/chosen": -237.292724609375, | |
| "logps/rejected": -250.16238403320312, | |
| "loss": 1.3682, | |
| "rewards/accuracies": 0.7250000238418579, | |
| "rewards/chosen": 0.0359283983707428, | |
| "rewards/margins": 0.08779667317867279, | |
| "rewards/rejected": -0.0518682599067688, | |
| "step": 50 | |
| }, | |
| { | |
| "epoch": 0.03925417075564279, | |
| "grad_norm": 6.84375, | |
| "learning_rate": 1.96078431372549e-06, | |
| "logits/chosen": -2.5553746223449707, | |
| "logits/rejected": -2.249553680419922, | |
| "logps/chosen": -291.7508239746094, | |
| "logps/rejected": -209.22256469726562, | |
| "loss": 1.3681, | |
| "rewards/accuracies": 0.6875, | |
| "rewards/chosen": 0.006251047365367413, | |
| "rewards/margins": 0.1306430548429489, | |
| "rewards/rejected": -0.12439201027154922, | |
| "step": 60 | |
| }, | |
| { | |
| "epoch": 0.04579653254824992, | |
| "grad_norm": 7.15625, | |
| "learning_rate": 2.2875816993464053e-06, | |
| "logits/chosen": -2.4841151237487793, | |
| "logits/rejected": -2.4571692943573, | |
| "logps/chosen": -233.838623046875, | |
| "logps/rejected": -244.6779022216797, | |
| "loss": 1.3704, | |
| "rewards/accuracies": 0.7124999761581421, | |
| "rewards/chosen": 0.05333448573946953, | |
| "rewards/margins": 0.15633264183998108, | |
| "rewards/rejected": -0.10299815982580185, | |
| "step": 70 | |
| }, | |
| { | |
| "epoch": 0.05233889434085705, | |
| "grad_norm": 7.40625, | |
| "learning_rate": 2.6143790849673208e-06, | |
| "logits/chosen": -2.369415760040283, | |
| "logits/rejected": -2.2743725776672363, | |
| "logps/chosen": -234.2888946533203, | |
| "logps/rejected": -199.08114624023438, | |
| "loss": 1.3623, | |
| "rewards/accuracies": 0.637499988079071, | |
| "rewards/chosen": 0.011479836888611317, | |
| "rewards/margins": 0.14200374484062195, | |
| "rewards/rejected": -0.1305239051580429, | |
| "step": 80 | |
| }, | |
| { | |
| "epoch": 0.058881256133464184, | |
| "grad_norm": 8.25, | |
| "learning_rate": 2.9411764705882355e-06, | |
| "logits/chosen": -2.3975484371185303, | |
| "logits/rejected": -2.3238232135772705, | |
| "logps/chosen": -267.9956359863281, | |
| "logps/rejected": -265.9745178222656, | |
| "loss": 1.3549, | |
| "rewards/accuracies": 0.762499988079071, | |
| "rewards/chosen": 0.1009618490934372, | |
| "rewards/margins": 0.2633386254310608, | |
| "rewards/rejected": -0.16237673163414001, | |
| "step": 90 | |
| }, | |
| { | |
| "epoch": 0.0654236179260713, | |
| "grad_norm": 6.96875, | |
| "learning_rate": 3.2679738562091506e-06, | |
| "logits/chosen": -2.496324062347412, | |
| "logits/rejected": -2.4217727184295654, | |
| "logps/chosen": -258.4588317871094, | |
| "logps/rejected": -248.0790557861328, | |
| "loss": 1.3652, | |
| "rewards/accuracies": 0.6499999761581421, | |
| "rewards/chosen": 0.04371471330523491, | |
| "rewards/margins": 0.1967853307723999, | |
| "rewards/rejected": -0.1530705988407135, | |
| "step": 100 | |
| }, | |
| { | |
| "epoch": 0.0654236179260713, | |
| "eval_logits/chosen": -2.2871594429016113, | |
| "eval_logits/rejected": -2.2039709091186523, | |
| "eval_logps/chosen": -271.9973449707031, | |
| "eval_logps/rejected": -256.3067321777344, | |
| "eval_loss": 1.3585007190704346, | |
| "eval_rewards/accuracies": 0.7200000286102295, | |
| "eval_rewards/chosen": 0.01791691593825817, | |
| "eval_rewards/margins": 0.23829618096351624, | |
| "eval_rewards/rejected": -0.2203792780637741, | |
| "eval_runtime": 193.0612, | |
| "eval_samples_per_second": 10.359, | |
| "eval_steps_per_second": 0.518, | |
| "step": 100 | |
| }, | |
| { | |
| "epoch": 0.07196597971867845, | |
| "grad_norm": 7.78125, | |
| "learning_rate": 3.5947712418300657e-06, | |
| "logits/chosen": -2.355022668838501, | |
| "logits/rejected": -2.3289151191711426, | |
| "logps/chosen": -282.1571960449219, | |
| "logps/rejected": -228.05319213867188, | |
| "loss": 1.3483, | |
| "rewards/accuracies": 0.699999988079071, | |
| "rewards/chosen": 0.0913485437631607, | |
| "rewards/margins": 0.3209895193576813, | |
| "rewards/rejected": -0.229640930891037, | |
| "step": 110 | |
| }, | |
| { | |
| "epoch": 0.07850834151128558, | |
| "grad_norm": 8.4375, | |
| "learning_rate": 3.92156862745098e-06, | |
| "logits/chosen": -2.3910622596740723, | |
| "logits/rejected": -2.2716870307922363, | |
| "logps/chosen": -274.3619079589844, | |
| "logps/rejected": -280.36614990234375, | |
| "loss": 1.3505, | |
| "rewards/accuracies": 0.800000011920929, | |
| "rewards/chosen": 0.11332492530345917, | |
| "rewards/margins": 0.39289209246635437, | |
| "rewards/rejected": -0.279567152261734, | |
| "step": 120 | |
| }, | |
| { | |
| "epoch": 0.08505070330389271, | |
| "grad_norm": 6.65625, | |
| "learning_rate": 4.2483660130718954e-06, | |
| "logits/chosen": -2.4673023223876953, | |
| "logits/rejected": -2.3496053218841553, | |
| "logps/chosen": -233.6593017578125, | |
| "logps/rejected": -225.4088134765625, | |
| "loss": 1.3577, | |
| "rewards/accuracies": 0.7250000238418579, | |
| "rewards/chosen": 0.04373621195554733, | |
| "rewards/margins": 0.27300602197647095, | |
| "rewards/rejected": -0.22926978766918182, | |
| "step": 130 | |
| }, | |
| { | |
| "epoch": 0.09159306509649984, | |
| "grad_norm": 8.8125, | |
| "learning_rate": 4.5751633986928105e-06, | |
| "logits/chosen": -2.488830804824829, | |
| "logits/rejected": -2.363779067993164, | |
| "logps/chosen": -295.23980712890625, | |
| "logps/rejected": -280.0679016113281, | |
| "loss": 1.3551, | |
| "rewards/accuracies": 0.6499999761581421, | |
| "rewards/chosen": 0.03703129664063454, | |
| "rewards/margins": 0.1885422170162201, | |
| "rewards/rejected": -0.15151092410087585, | |
| "step": 140 | |
| }, | |
| { | |
| "epoch": 0.09813542688910697, | |
| "grad_norm": 7.21875, | |
| "learning_rate": 4.901960784313726e-06, | |
| "logits/chosen": -2.4635255336761475, | |
| "logits/rejected": -2.3997857570648193, | |
| "logps/chosen": -271.8360900878906, | |
| "logps/rejected": -239.2571563720703, | |
| "loss": 1.3443, | |
| "rewards/accuracies": 0.7749999761581421, | |
| "rewards/chosen": 0.13082675635814667, | |
| "rewards/margins": 0.38594740629196167, | |
| "rewards/rejected": -0.2551206648349762, | |
| "step": 150 | |
| }, | |
| { | |
| "epoch": 0.1046777886817141, | |
| "grad_norm": 8.1875, | |
| "learning_rate": 4.999680264259825e-06, | |
| "logits/chosen": -2.5628502368927, | |
| "logits/rejected": -2.309502601623535, | |
| "logps/chosen": -289.554931640625, | |
| "logps/rejected": -228.1037139892578, | |
| "loss": 1.3507, | |
| "rewards/accuracies": 0.6875, | |
| "rewards/chosen": 0.056355126202106476, | |
| "rewards/margins": 0.30590537190437317, | |
| "rewards/rejected": -0.2495502233505249, | |
| "step": 160 | |
| }, | |
| { | |
| "epoch": 0.11122015047432122, | |
| "grad_norm": 24.125, | |
| "learning_rate": 4.998114408534616e-06, | |
| "logits/chosen": -2.459836483001709, | |
| "logits/rejected": -2.241123676300049, | |
| "logps/chosen": -269.2346496582031, | |
| "logps/rejected": -245.09005737304688, | |
| "loss": 1.3566, | |
| "rewards/accuracies": 0.7250000238418579, | |
| "rewards/chosen": 0.11544966697692871, | |
| "rewards/margins": 0.3507261276245117, | |
| "rewards/rejected": -0.2352764904499054, | |
| "step": 170 | |
| }, | |
| { | |
| "epoch": 0.11776251226692837, | |
| "grad_norm": 7.28125, | |
| "learning_rate": 4.995244522215781e-06, | |
| "logits/chosen": -2.479318857192993, | |
| "logits/rejected": -2.452930450439453, | |
| "logps/chosen": -253.64907836914062, | |
| "logps/rejected": -255.054443359375, | |
| "loss": 1.3515, | |
| "rewards/accuracies": 0.6499999761581421, | |
| "rewards/chosen": 0.0358736589550972, | |
| "rewards/margins": 0.2104395180940628, | |
| "rewards/rejected": -0.174565851688385, | |
| "step": 180 | |
| }, | |
| { | |
| "epoch": 0.1243048740595355, | |
| "grad_norm": 6.78125, | |
| "learning_rate": 4.9910721034010655e-06, | |
| "logits/chosen": -2.4955029487609863, | |
| "logits/rejected": -2.3607935905456543, | |
| "logps/chosen": -259.48529052734375, | |
| "logps/rejected": -232.8231964111328, | |
| "loss": 1.3516, | |
| "rewards/accuracies": 0.6499999761581421, | |
| "rewards/chosen": 0.06281731277704239, | |
| "rewards/margins": 0.3066301941871643, | |
| "rewards/rejected": -0.24381284415721893, | |
| "step": 190 | |
| }, | |
| { | |
| "epoch": 0.1308472358521426, | |
| "grad_norm": 7.90625, | |
| "learning_rate": 4.985599330117931e-06, | |
| "logits/chosen": -2.4151864051818848, | |
| "logits/rejected": -2.3472375869750977, | |
| "logps/chosen": -251.25210571289062, | |
| "logps/rejected": -235.2193603515625, | |
| "loss": 1.3457, | |
| "rewards/accuracies": 0.7124999761581421, | |
| "rewards/chosen": 0.0700979083776474, | |
| "rewards/margins": 0.3274100422859192, | |
| "rewards/rejected": -0.25731217861175537, | |
| "step": 200 | |
| }, | |
| { | |
| "epoch": 0.1308472358521426, | |
| "eval_logits/chosen": -2.2911951541900635, | |
| "eval_logits/rejected": -2.2080154418945312, | |
| "eval_logps/chosen": -270.37738037109375, | |
| "eval_logps/rejected": -256.1176452636719, | |
| "eval_loss": 1.3529404401779175, | |
| "eval_rewards/accuracies": 0.7425000071525574, | |
| "eval_rewards/chosen": 0.17991353571414948, | |
| "eval_rewards/margins": 0.3813881278038025, | |
| "eval_rewards/rejected": -0.20147459208965302, | |
| "eval_runtime": 192.5507, | |
| "eval_samples_per_second": 10.387, | |
| "eval_steps_per_second": 0.519, | |
| "step": 200 | |
| }, | |
| { | |
| "epoch": 0.13738959764474976, | |
| "grad_norm": 6.9375, | |
| "learning_rate": 4.978829059186611e-06, | |
| "logits/chosen": -2.512892246246338, | |
| "logits/rejected": -2.4710421562194824, | |
| "logps/chosen": -278.5118713378906, | |
| "logps/rejected": -281.03924560546875, | |
| "loss": 1.3592, | |
| "rewards/accuracies": 0.737500011920929, | |
| "rewards/chosen": 0.01397608406841755, | |
| "rewards/margins": 0.33375468850135803, | |
| "rewards/rejected": -0.31977859139442444, | |
| "step": 210 | |
| }, | |
| { | |
| "epoch": 0.1439319594373569, | |
| "grad_norm": 7.8125, | |
| "learning_rate": 4.97076482472884e-06, | |
| "logits/chosen": -2.3786368370056152, | |
| "logits/rejected": -2.4024887084960938, | |
| "logps/chosen": -307.94610595703125, | |
| "logps/rejected": -310.7808532714844, | |
| "loss": 1.3364, | |
| "rewards/accuracies": 0.762499988079071, | |
| "rewards/chosen": 0.07198480516672134, | |
| "rewards/margins": 0.43232816457748413, | |
| "rewards/rejected": -0.3603433668613434, | |
| "step": 220 | |
| }, | |
| { | |
| "epoch": 0.15047432122996401, | |
| "grad_norm": 7.96875, | |
| "learning_rate": 4.961410836323014e-06, | |
| "logits/chosen": -2.5079922676086426, | |
| "logits/rejected": -2.3904035091400146, | |
| "logps/chosen": -292.0177917480469, | |
| "logps/rejected": -243.524658203125, | |
| "loss": 1.3466, | |
| "rewards/accuracies": 0.7250000238418579, | |
| "rewards/chosen": 0.0814410150051117, | |
| "rewards/margins": 0.4043430685997009, | |
| "rewards/rejected": -0.322902113199234, | |
| "step": 230 | |
| }, | |
| { | |
| "epoch": 0.15701668302257116, | |
| "grad_norm": 6.75, | |
| "learning_rate": 4.950771976806769e-06, | |
| "logits/chosen": -2.643059730529785, | |
| "logits/rejected": -2.447086811065674, | |
| "logps/chosen": -279.7219543457031, | |
| "logps/rejected": -228.05068969726562, | |
| "loss": 1.3531, | |
| "rewards/accuracies": 0.7749999761581421, | |
| "rewards/chosen": 0.1486392319202423, | |
| "rewards/margins": 0.4861406683921814, | |
| "rewards/rejected": -0.3375014662742615, | |
| "step": 240 | |
| }, | |
| { | |
| "epoch": 0.16355904481517827, | |
| "grad_norm": 7.25, | |
| "learning_rate": 4.938853799728112e-06, | |
| "logits/chosen": -2.533738613128662, | |
| "logits/rejected": -2.347072124481201, | |
| "logps/chosen": -291.59527587890625, | |
| "logps/rejected": -224.52340698242188, | |
| "loss": 1.3444, | |
| "rewards/accuracies": 0.737500011920929, | |
| "rewards/chosen": 0.14931416511535645, | |
| "rewards/margins": 0.5157926082611084, | |
| "rewards/rejected": -0.36647850275039673, | |
| "step": 250 | |
| }, | |
| { | |
| "epoch": 0.17010140660778542, | |
| "grad_norm": 7.96875, | |
| "learning_rate": 4.925662526446431e-06, | |
| "logits/chosen": -2.4311368465423584, | |
| "logits/rejected": -2.3289899826049805, | |
| "logps/chosen": -253.0860137939453, | |
| "logps/rejected": -219.7169952392578, | |
| "loss": 1.3418, | |
| "rewards/accuracies": 0.6625000238418579, | |
| "rewards/chosen": 0.03077099286019802, | |
| "rewards/margins": 0.30891746282577515, | |
| "rewards/rejected": -0.27814650535583496, | |
| "step": 260 | |
| }, | |
| { | |
| "epoch": 0.17664376840039253, | |
| "grad_norm": 6.90625, | |
| "learning_rate": 4.911205042884912e-06, | |
| "logits/chosen": -2.497413158416748, | |
| "logits/rejected": -2.3586812019348145, | |
| "logps/chosen": -294.7901916503906, | |
| "logps/rejected": -242.64291381835938, | |
| "loss": 1.3551, | |
| "rewards/accuracies": 0.625, | |
| "rewards/chosen": 0.0127085717394948, | |
| "rewards/margins": 0.21187984943389893, | |
| "rewards/rejected": -0.1991712599992752, | |
| "step": 270 | |
| }, | |
| { | |
| "epoch": 0.18318613019299967, | |
| "grad_norm": 7.15625, | |
| "learning_rate": 4.895488895936047e-06, | |
| "logits/chosen": -2.4523603916168213, | |
| "logits/rejected": -2.4156365394592285, | |
| "logps/chosen": -240.6490020751953, | |
| "logps/rejected": -219.9276580810547, | |
| "loss": 1.3329, | |
| "rewards/accuracies": 0.762499988079071, | |
| "rewards/chosen": 0.1077442541718483, | |
| "rewards/margins": 0.4334333539009094, | |
| "rewards/rejected": -0.32568907737731934, | |
| "step": 280 | |
| }, | |
| { | |
| "epoch": 0.18972849198560682, | |
| "grad_norm": 10.0, | |
| "learning_rate": 4.8785222895221075e-06, | |
| "logits/chosen": -2.671842336654663, | |
| "logits/rejected": -2.547651767730713, | |
| "logps/chosen": -322.6619567871094, | |
| "logps/rejected": -319.7884216308594, | |
| "loss": 1.3476, | |
| "rewards/accuracies": 0.6875, | |
| "rewards/chosen": 0.07628868520259857, | |
| "rewards/margins": 0.34415873885154724, | |
| "rewards/rejected": -0.26787006855010986, | |
| "step": 290 | |
| }, | |
| { | |
| "epoch": 0.19627085377821393, | |
| "grad_norm": 7.8125, | |
| "learning_rate": 4.860314080312651e-06, | |
| "logits/chosen": -2.5084714889526367, | |
| "logits/rejected": -2.4003777503967285, | |
| "logps/chosen": -274.4216003417969, | |
| "logps/rejected": -254.49710083007812, | |
| "loss": 1.3328, | |
| "rewards/accuracies": 0.737500011920929, | |
| "rewards/chosen": 0.07708136737346649, | |
| "rewards/margins": 0.42221707105636597, | |
| "rewards/rejected": -0.3451356589794159, | |
| "step": 300 | |
| }, | |
| { | |
| "epoch": 0.19627085377821393, | |
| "eval_logits/chosen": -2.3106164932250977, | |
| "eval_logits/rejected": -2.230290412902832, | |
| "eval_logps/chosen": -270.9070739746094, | |
| "eval_logps/rejected": -257.02178955078125, | |
| "eval_loss": 1.3500434160232544, | |
| "eval_rewards/accuracies": 0.7149999737739563, | |
| "eval_rewards/chosen": 0.12694190442562103, | |
| "eval_rewards/margins": 0.41882869601249695, | |
| "eval_rewards/rejected": -0.2918868362903595, | |
| "eval_runtime": 192.4649, | |
| "eval_samples_per_second": 10.392, | |
| "eval_steps_per_second": 0.52, | |
| "step": 300 | |
| }, | |
| { | |
| "epoch": 0.20281321557082108, | |
| "grad_norm": 8.125, | |
| "learning_rate": 4.840873773101287e-06, | |
| "logits/chosen": -2.4888222217559814, | |
| "logits/rejected": -2.4138476848602295, | |
| "logps/chosen": -264.6638488769531, | |
| "logps/rejected": -239.2753448486328, | |
| "loss": 1.3448, | |
| "rewards/accuracies": 0.7875000238418579, | |
| "rewards/chosen": 0.10051695257425308, | |
| "rewards/margins": 0.5004209280014038, | |
| "rewards/rejected": -0.39990395307540894, | |
| "step": 310 | |
| }, | |
| { | |
| "epoch": 0.2093555773634282, | |
| "grad_norm": 7.6875, | |
| "learning_rate": 4.820211515844116e-06, | |
| "logits/chosen": -2.500939130783081, | |
| "logits/rejected": -2.4757304191589355, | |
| "logps/chosen": -203.4287109375, | |
| "logps/rejected": -218.0804901123047, | |
| "loss": 1.3531, | |
| "rewards/accuracies": 0.675000011920929, | |
| "rewards/chosen": 0.10015901178121567, | |
| "rewards/margins": 0.3148805797100067, | |
| "rewards/rejected": -0.21472156047821045, | |
| "step": 320 | |
| }, | |
| { | |
| "epoch": 0.21589793915603533, | |
| "grad_norm": 7.96875, | |
| "learning_rate": 4.798338094362439e-06, | |
| "logits/chosen": -2.477666139602661, | |
| "logits/rejected": -2.5014617443084717, | |
| "logps/chosen": -288.17950439453125, | |
| "logps/rejected": -281.34295654296875, | |
| "loss": 1.3666, | |
| "rewards/accuracies": 0.675000011920929, | |
| "rewards/chosen": 0.07362223416566849, | |
| "rewards/margins": 0.274558961391449, | |
| "rewards/rejected": -0.2009367197751999, | |
| "step": 330 | |
| }, | |
| { | |
| "epoch": 0.22244030094864245, | |
| "grad_norm": 6.75, | |
| "learning_rate": 4.775264926712489e-06, | |
| "logits/chosen": -2.4466490745544434, | |
| "logits/rejected": -2.4048807621002197, | |
| "logps/chosen": -246.18392944335938, | |
| "logps/rejected": -259.27392578125, | |
| "loss": 1.3409, | |
| "rewards/accuracies": 0.762499988079071, | |
| "rewards/chosen": 0.11778600513935089, | |
| "rewards/margins": 0.4433900713920593, | |
| "rewards/rejected": -0.32560408115386963, | |
| "step": 340 | |
| }, | |
| { | |
| "epoch": 0.2289826627412496, | |
| "grad_norm": 6.5, | |
| "learning_rate": 4.751004057225147e-06, | |
| "logits/chosen": -2.4351608753204346, | |
| "logits/rejected": -2.3736894130706787, | |
| "logps/chosen": -275.3951721191406, | |
| "logps/rejected": -281.2572021484375, | |
| "loss": 1.344, | |
| "rewards/accuracies": 0.762499988079071, | |
| "rewards/chosen": 0.12299034744501114, | |
| "rewards/margins": 0.42316460609436035, | |
| "rewards/rejected": -0.3001742362976074, | |
| "step": 350 | |
| }, | |
| { | |
| "epoch": 0.23552502453385674, | |
| "grad_norm": 7.875, | |
| "learning_rate": 4.725568150218719e-06, | |
| "logits/chosen": -2.353898763656616, | |
| "logits/rejected": -2.3495678901672363, | |
| "logps/chosen": -247.7540283203125, | |
| "logps/rejected": -262.0528564453125, | |
| "loss": 1.339, | |
| "rewards/accuracies": 0.7749999761581421, | |
| "rewards/chosen": 0.10569906234741211, | |
| "rewards/margins": 0.41347068548202515, | |
| "rewards/rejected": -0.3077716529369354, | |
| "step": 360 | |
| }, | |
| { | |
| "epoch": 0.24206738632646385, | |
| "grad_norm": 8.6875, | |
| "learning_rate": 4.6989704833880936e-06, | |
| "logits/chosen": -2.455857515335083, | |
| "logits/rejected": -2.2684264183044434, | |
| "logps/chosen": -278.91278076171875, | |
| "logps/rejected": -248.13729858398438, | |
| "loss": 1.3587, | |
| "rewards/accuracies": 0.637499988079071, | |
| "rewards/chosen": 0.05706097558140755, | |
| "rewards/margins": 0.2867721915245056, | |
| "rewards/rejected": -0.22971120476722717, | |
| "step": 370 | |
| }, | |
| { | |
| "epoch": 0.248609748119071, | |
| "grad_norm": 7.5, | |
| "learning_rate": 4.671224940873704e-06, | |
| "logits/chosen": -2.5477356910705566, | |
| "logits/rejected": -2.4833083152770996, | |
| "logps/chosen": -284.73394775390625, | |
| "logps/rejected": -268.12493896484375, | |
| "loss": 1.3407, | |
| "rewards/accuracies": 0.762499988079071, | |
| "rewards/chosen": 0.14602813124656677, | |
| "rewards/margins": 0.4975086748600006, | |
| "rewards/rejected": -0.35148054361343384, | |
| "step": 380 | |
| }, | |
| { | |
| "epoch": 0.25515210991167814, | |
| "grad_norm": 8.0625, | |
| "learning_rate": 4.642346006013925e-06, | |
| "logits/chosen": -2.563493251800537, | |
| "logits/rejected": -2.495145320892334, | |
| "logps/chosen": -258.86444091796875, | |
| "logps/rejected": -258.3479309082031, | |
| "loss": 1.3374, | |
| "rewards/accuracies": 0.7124999761581421, | |
| "rewards/chosen": 0.12897971272468567, | |
| "rewards/margins": 0.4424809515476227, | |
| "rewards/rejected": -0.3135012090206146, | |
| "step": 390 | |
| }, | |
| { | |
| "epoch": 0.2616944717042852, | |
| "grad_norm": 7.6875, | |
| "learning_rate": 4.612348753784682e-06, | |
| "logits/chosen": -2.5260744094848633, | |
| "logits/rejected": -2.3809266090393066, | |
| "logps/chosen": -296.6476745605469, | |
| "logps/rejected": -242.137451171875, | |
| "loss": 1.3452, | |
| "rewards/accuracies": 0.699999988079071, | |
| "rewards/chosen": 0.08820326626300812, | |
| "rewards/margins": 0.4365547299385071, | |
| "rewards/rejected": -0.34835144877433777, | |
| "step": 400 | |
| }, | |
| { | |
| "epoch": 0.2616944717042852, | |
| "eval_logits/chosen": -2.3061885833740234, | |
| "eval_logits/rejected": -2.224980115890503, | |
| "eval_logps/chosen": -270.3225402832031, | |
| "eval_logps/rejected": -256.4975891113281, | |
| "eval_loss": 1.3535617589950562, | |
| "eval_rewards/accuracies": 0.7200000286102295, | |
| "eval_rewards/chosen": 0.185396209359169, | |
| "eval_rewards/margins": 0.42486390471458435, | |
| "eval_rewards/rejected": -0.23946763575077057, | |
| "eval_runtime": 192.5082, | |
| "eval_samples_per_second": 10.389, | |
| "eval_steps_per_second": 0.519, | |
| "step": 400 | |
| }, | |
| { | |
| "epoch": 0.26823683349689237, | |
| "grad_norm": 7.71875, | |
| "learning_rate": 4.5812488429302245e-06, | |
| "logits/chosen": -2.4703800678253174, | |
| "logits/rejected": -2.384880542755127, | |
| "logps/chosen": -242.77249145507812, | |
| "logps/rejected": -216.26919555664062, | |
| "loss": 1.3521, | |
| "rewards/accuracies": 0.675000011920929, | |
| "rewards/chosen": 0.08403290808200836, | |
| "rewards/margins": 0.23067112267017365, | |
| "rewards/rejected": -0.14663821458816528, | |
| "step": 410 | |
| }, | |
| { | |
| "epoch": 0.2747791952894995, | |
| "grad_norm": 7.625, | |
| "learning_rate": 4.54906250778917e-06, | |
| "logits/chosen": -2.494356632232666, | |
| "logits/rejected": -2.4741995334625244, | |
| "logps/chosen": -307.1471252441406, | |
| "logps/rejected": -280.8088073730469, | |
| "loss": 1.3253, | |
| "rewards/accuracies": 0.675000011920929, | |
| "rewards/chosen": 0.11699722707271576, | |
| "rewards/margins": 0.491929829120636, | |
| "rewards/rejected": -0.37493258714675903, | |
| "step": 420 | |
| }, | |
| { | |
| "epoch": 0.28132155708210665, | |
| "grad_norm": 6.53125, | |
| "learning_rate": 4.515806549820084e-06, | |
| "logits/chosen": -2.4651083946228027, | |
| "logits/rejected": -2.3330154418945312, | |
| "logps/chosen": -281.57110595703125, | |
| "logps/rejected": -254.07498168945312, | |
| "loss": 1.3406, | |
| "rewards/accuracies": 0.7124999761581421, | |
| "rewards/chosen": 0.16895495355129242, | |
| "rewards/margins": 0.4621726870536804, | |
| "rewards/rejected": -0.2932177186012268, | |
| "step": 430 | |
| }, | |
| { | |
| "epoch": 0.2878639188747138, | |
| "grad_norm": 8.875, | |
| "learning_rate": 4.48149832883101e-06, | |
| "logits/chosen": -2.4727940559387207, | |
| "logits/rejected": -2.314194917678833, | |
| "logps/chosen": -266.56134033203125, | |
| "logps/rejected": -236.39089965820312, | |
| "loss": 1.3585, | |
| "rewards/accuracies": 0.6000000238418579, | |
| "rewards/chosen": -0.08974994719028473, | |
| "rewards/margins": 0.26218581199645996, | |
| "rewards/rejected": -0.3519357144832611, | |
| "step": 440 | |
| }, | |
| { | |
| "epoch": 0.2944062806673209, | |
| "grad_norm": 9.125, | |
| "learning_rate": 4.446155753917559e-06, | |
| "logits/chosen": -2.4373867511749268, | |
| "logits/rejected": -2.3501367568969727, | |
| "logps/chosen": -249.449462890625, | |
| "logps/rejected": -280.8139343261719, | |
| "loss": 1.355, | |
| "rewards/accuracies": 0.7250000238418579, | |
| "rewards/chosen": 0.09939811378717422, | |
| "rewards/margins": 0.3732451796531677, | |
| "rewards/rejected": -0.2738470435142517, | |
| "step": 450 | |
| }, | |
| { | |
| "epoch": 0.30094864245992803, | |
| "grad_norm": 7.375, | |
| "learning_rate": 4.409797274114245e-06, | |
| "logits/chosen": -2.5222246646881104, | |
| "logits/rejected": -2.3621227741241455, | |
| "logps/chosen": -293.97747802734375, | |
| "logps/rejected": -264.7182312011719, | |
| "loss": 1.3389, | |
| "rewards/accuracies": 0.75, | |
| "rewards/chosen": 0.11550019681453705, | |
| "rewards/margins": 0.4369579255580902, | |
| "rewards/rejected": -0.32145771384239197, | |
| "step": 460 | |
| }, | |
| { | |
| "epoch": 0.30749100425253517, | |
| "grad_norm": 8.0, | |
| "learning_rate": 4.372441868763981e-06, | |
| "logits/chosen": -2.4279232025146484, | |
| "logits/rejected": -2.3788094520568848, | |
| "logps/chosen": -269.70782470703125, | |
| "logps/rejected": -271.08380126953125, | |
| "loss": 1.3499, | |
| "rewards/accuracies": 0.737500011920929, | |
| "rewards/chosen": 0.1701391041278839, | |
| "rewards/margins": 0.48491889238357544, | |
| "rewards/rejected": -0.31477978825569153, | |
| "step": 470 | |
| }, | |
| { | |
| "epoch": 0.3140333660451423, | |
| "grad_norm": 6.8125, | |
| "learning_rate": 4.334109037610757e-06, | |
| "logits/chosen": -2.47053599357605, | |
| "logits/rejected": -2.4681105613708496, | |
| "logps/chosen": -248.77597045898438, | |
| "logps/rejected": -264.0259704589844, | |
| "loss": 1.3588, | |
| "rewards/accuracies": 0.7124999761581421, | |
| "rewards/chosen": 0.08740128576755524, | |
| "rewards/margins": 0.4287898540496826, | |
| "rewards/rejected": -0.3413885235786438, | |
| "step": 480 | |
| }, | |
| { | |
| "epoch": 0.3205757278377494, | |
| "grad_norm": 9.0625, | |
| "learning_rate": 4.294818790620644e-06, | |
| "logits/chosen": -2.388354539871216, | |
| "logits/rejected": -2.2699804306030273, | |
| "logps/chosen": -230.1956329345703, | |
| "logps/rejected": -222.83560180664062, | |
| "loss": 1.3372, | |
| "rewards/accuracies": 0.7250000238418579, | |
| "rewards/chosen": 0.11192689836025238, | |
| "rewards/margins": 0.4855028986930847, | |
| "rewards/rejected": -0.37357598543167114, | |
| "step": 490 | |
| }, | |
| { | |
| "epoch": 0.32711808963035655, | |
| "grad_norm": 6.46875, | |
| "learning_rate": 4.2545916375364835e-06, | |
| "logits/chosen": -2.4573752880096436, | |
| "logits/rejected": -2.384742259979248, | |
| "logps/chosen": -253.0285186767578, | |
| "logps/rejected": -235.6699676513672, | |
| "loss": 1.3446, | |
| "rewards/accuracies": 0.7250000238418579, | |
| "rewards/chosen": 0.12988470494747162, | |
| "rewards/margins": 0.4586367607116699, | |
| "rewards/rejected": -0.3287521004676819, | |
| "step": 500 | |
| }, | |
| { | |
| "epoch": 0.32711808963035655, | |
| "eval_logits/chosen": -2.2817647457122803, | |
| "eval_logits/rejected": -2.1983530521392822, | |
| "eval_logps/chosen": -271.3174743652344, | |
| "eval_logps/rejected": -258.03887939453125, | |
| "eval_loss": 1.3500897884368896, | |
| "eval_rewards/accuracies": 0.7275000214576721, | |
| "eval_rewards/chosen": 0.08590395003557205, | |
| "eval_rewards/margins": 0.47949859499931335, | |
| "eval_rewards/rejected": -0.3935946226119995, | |
| "eval_runtime": 192.5361, | |
| "eval_samples_per_second": 10.388, | |
| "eval_steps_per_second": 0.519, | |
| "step": 500 | |
| }, | |
| { | |
| "epoch": 0.3336604514229637, | |
| "grad_norm": 6.875, | |
| "learning_rate": 4.213448577171676e-06, | |
| "logits/chosen": -2.4477853775024414, | |
| "logits/rejected": -2.2752411365509033, | |
| "logps/chosen": -286.5871887207031, | |
| "logps/rejected": -277.1395568847656, | |
| "loss": 1.3551, | |
| "rewards/accuracies": 0.699999988079071, | |
| "rewards/chosen": 0.09675732254981995, | |
| "rewards/margins": 0.36776870489120483, | |
| "rewards/rejected": -0.2710114121437073, | |
| "step": 510 | |
| }, | |
| { | |
| "epoch": 0.34020281321557083, | |
| "grad_norm": 6.71875, | |
| "learning_rate": 4.171411086448674e-06, | |
| "logits/chosen": -2.4915995597839355, | |
| "logits/rejected": -2.40187668800354, | |
| "logps/chosen": -264.9870910644531, | |
| "logps/rejected": -248.1427001953125, | |
| "loss": 1.3515, | |
| "rewards/accuracies": 0.7124999761581421, | |
| "rewards/chosen": 0.009617133066058159, | |
| "rewards/margins": 0.3053894639015198, | |
| "rewards/rejected": -0.29577234387397766, | |
| "step": 520 | |
| }, | |
| { | |
| "epoch": 0.346745175008178, | |
| "grad_norm": 6.84375, | |
| "learning_rate": 4.128501109187903e-06, | |
| "logits/chosen": -2.5452780723571777, | |
| "logits/rejected": -2.388065814971924, | |
| "logps/chosen": -256.1615905761719, | |
| "logps/rejected": -243.15542602539062, | |
| "loss": 1.3448, | |
| "rewards/accuracies": 0.737500011920929, | |
| "rewards/chosen": 0.09919309616088867, | |
| "rewards/margins": 0.45780619978904724, | |
| "rewards/rejected": -0.35861313343048096, | |
| "step": 530 | |
| }, | |
| { | |
| "epoch": 0.35328753680078506, | |
| "grad_norm": 8.0, | |
| "learning_rate": 4.084741044652956e-06, | |
| "logits/chosen": -2.5606534481048584, | |
| "logits/rejected": -2.3131465911865234, | |
| "logps/chosen": -287.9478759765625, | |
| "logps/rejected": -224.75527954101562, | |
| "loss": 1.3668, | |
| "rewards/accuracies": 0.737500011920929, | |
| "rewards/chosen": 0.17254912853240967, | |
| "rewards/margins": 0.42019981145858765, | |
| "rewards/rejected": -0.24765071272850037, | |
| "step": 540 | |
| }, | |
| { | |
| "epoch": 0.3598298985933922, | |
| "grad_norm": 6.90625, | |
| "learning_rate": 4.040153735858041e-06, | |
| "logits/chosen": -2.4491143226623535, | |
| "logits/rejected": -2.3129725456237793, | |
| "logps/chosen": -274.43658447265625, | |
| "logps/rejected": -268.6435241699219, | |
| "loss": 1.3476, | |
| "rewards/accuracies": 0.75, | |
| "rewards/chosen": 0.14739489555358887, | |
| "rewards/margins": 0.41326436400413513, | |
| "rewards/rejected": -0.26586946845054626, | |
| "step": 550 | |
| }, | |
| { | |
| "epoch": 0.36637226038599935, | |
| "grad_norm": 6.84375, | |
| "learning_rate": 3.9947624576437975e-06, | |
| "logits/chosen": -2.47575306892395, | |
| "logits/rejected": -2.3158023357391357, | |
| "logps/chosen": -230.30819702148438, | |
| "logps/rejected": -211.89633178710938, | |
| "loss": 1.3509, | |
| "rewards/accuracies": 0.7124999761581421, | |
| "rewards/chosen": 0.07401107251644135, | |
| "rewards/margins": 0.36348801851272583, | |
| "rewards/rejected": -0.2894769012928009, | |
| "step": 560 | |
| }, | |
| { | |
| "epoch": 0.3729146221786065, | |
| "grad_norm": 6.90625, | |
| "learning_rate": 3.948590904527689e-06, | |
| "logits/chosen": -2.395169496536255, | |
| "logits/rejected": -2.330249309539795, | |
| "logps/chosen": -243.09487915039062, | |
| "logps/rejected": -263.39227294921875, | |
| "loss": 1.3349, | |
| "rewards/accuracies": 0.7250000238418579, | |
| "rewards/chosen": -0.01647457666695118, | |
| "rewards/margins": 0.5050023794174194, | |
| "rewards/rejected": -0.5214769244194031, | |
| "step": 570 | |
| }, | |
| { | |
| "epoch": 0.37945698397121363, | |
| "grad_norm": 6.46875, | |
| "learning_rate": 3.901663178335318e-06, | |
| "logits/chosen": -2.5840940475463867, | |
| "logits/rejected": -2.4302070140838623, | |
| "logps/chosen": -292.9772644042969, | |
| "logps/rejected": -280.25726318359375, | |
| "loss": 1.3405, | |
| "rewards/accuracies": 0.675000011920929, | |
| "rewards/chosen": 0.11058641970157623, | |
| "rewards/margins": 0.506061851978302, | |
| "rewards/rejected": -0.39547544717788696, | |
| "step": 580 | |
| }, | |
| { | |
| "epoch": 0.3859993457638207, | |
| "grad_norm": 7.59375, | |
| "learning_rate": 3.854003775619142e-06, | |
| "logits/chosen": -2.3781380653381348, | |
| "logits/rejected": -2.372018337249756, | |
| "logps/chosen": -263.7474365234375, | |
| "logps/rejected": -244.7537384033203, | |
| "loss": 1.3416, | |
| "rewards/accuracies": 0.7250000238418579, | |
| "rewards/chosen": 0.06603378802537918, | |
| "rewards/margins": 0.4166053831577301, | |
| "rewards/rejected": -0.3505716323852539, | |
| "step": 590 | |
| }, | |
| { | |
| "epoch": 0.39254170755642787, | |
| "grad_norm": 8.0625, | |
| "learning_rate": 3.805637574871115e-06, | |
| "logits/chosen": -2.455042600631714, | |
| "logits/rejected": -2.3687422275543213, | |
| "logps/chosen": -266.13043212890625, | |
| "logps/rejected": -225.39614868164062, | |
| "loss": 1.333, | |
| "rewards/accuracies": 0.75, | |
| "rewards/chosen": 0.10043742507696152, | |
| "rewards/margins": 0.5269637703895569, | |
| "rewards/rejected": -0.4265263080596924, | |
| "step": 600 | |
| }, | |
| { | |
| "epoch": 0.39254170755642787, | |
| "eval_logits/chosen": -2.293658494949341, | |
| "eval_logits/rejected": -2.2107114791870117, | |
| "eval_logps/chosen": -271.6837463378906, | |
| "eval_logps/rejected": -257.95440673828125, | |
| "eval_loss": 1.3495758771896362, | |
| "eval_rewards/accuracies": 0.7450000047683716, | |
| "eval_rewards/chosen": 0.04927777126431465, | |
| "eval_rewards/margins": 0.43442490696907043, | |
| "eval_rewards/rejected": -0.38514718413352966, | |
| "eval_runtime": 192.5013, | |
| "eval_samples_per_second": 10.39, | |
| "eval_steps_per_second": 0.519, | |
| "step": 600 | |
| }, | |
| { | |
| "epoch": 0.399084069349035, | |
| "grad_norm": 6.71875, | |
| "learning_rate": 3.7565898235359717e-06, | |
| "logits/chosen": -2.531266689300537, | |
| "logits/rejected": -2.4104650020599365, | |
| "logps/chosen": -268.50482177734375, | |
| "logps/rejected": -252.98202514648438, | |
| "loss": 1.3522, | |
| "rewards/accuracies": 0.75, | |
| "rewards/chosen": 0.11455075442790985, | |
| "rewards/margins": 0.33470815420150757, | |
| "rewards/rejected": -0.2201574295759201, | |
| "step": 610 | |
| }, | |
| { | |
| "epoch": 0.40562643114164215, | |
| "grad_norm": 7.0, | |
| "learning_rate": 3.7068861248319127e-06, | |
| "logits/chosen": -2.420405626296997, | |
| "logits/rejected": -2.36627197265625, | |
| "logps/chosen": -252.94091796875, | |
| "logps/rejected": -260.6941833496094, | |
| "loss": 1.3502, | |
| "rewards/accuracies": 0.7749999761581421, | |
| "rewards/chosen": 0.004491160623729229, | |
| "rewards/margins": 0.4471139907836914, | |
| "rewards/rejected": -0.4426228106021881, | |
| "step": 620 | |
| }, | |
| { | |
| "epoch": 0.41216879293424924, | |
| "grad_norm": 8.125, | |
| "learning_rate": 3.6565524243855695e-06, | |
| "logits/chosen": -2.482597589492798, | |
| "logits/rejected": -2.381808042526245, | |
| "logps/chosen": -270.79833984375, | |
| "logps/rejected": -254.92868041992188, | |
| "loss": 1.3455, | |
| "rewards/accuracies": 0.75, | |
| "rewards/chosen": 0.18158364295959473, | |
| "rewards/margins": 0.5938194394111633, | |
| "rewards/rejected": -0.41223573684692383, | |
| "step": 630 | |
| }, | |
| { | |
| "epoch": 0.4187111547268564, | |
| "grad_norm": 8.125, | |
| "learning_rate": 3.6056149966882325e-06, | |
| "logits/chosen": -2.4741575717926025, | |
| "logits/rejected": -2.4443914890289307, | |
| "logps/chosen": -257.9027099609375, | |
| "logps/rejected": -263.8404235839844, | |
| "loss": 1.3501, | |
| "rewards/accuracies": 0.7749999761581421, | |
| "rewards/chosen": 0.12312284857034683, | |
| "rewards/margins": 0.4808879792690277, | |
| "rewards/rejected": -0.35776516795158386, | |
| "step": 640 | |
| }, | |
| { | |
| "epoch": 0.4252535165194635, | |
| "grad_norm": 7.84375, | |
| "learning_rate": 3.554100431380414e-06, | |
| "logits/chosen": -2.493617534637451, | |
| "logits/rejected": -2.3935816287994385, | |
| "logps/chosen": -249.92294311523438, | |
| "logps/rejected": -238.17404174804688, | |
| "loss": 1.3504, | |
| "rewards/accuracies": 0.75, | |
| "rewards/chosen": 0.04006017744541168, | |
| "rewards/margins": 0.48104342818260193, | |
| "rewards/rejected": -0.44098323583602905, | |
| "step": 650 | |
| }, | |
| { | |
| "epoch": 0.43179587831207067, | |
| "grad_norm": 9.0625, | |
| "learning_rate": 3.5020356193718934e-06, | |
| "logits/chosen": -2.4255988597869873, | |
| "logits/rejected": -2.3747265338897705, | |
| "logps/chosen": -217.7563934326172, | |
| "logps/rejected": -220.0769500732422, | |
| "loss": 1.3667, | |
| "rewards/accuracies": 0.7875000238418579, | |
| "rewards/chosen": 0.10521771758794785, | |
| "rewards/margins": 0.3761609196662903, | |
| "rewards/rejected": -0.27094319462776184, | |
| "step": 660 | |
| }, | |
| { | |
| "epoch": 0.4383382401046778, | |
| "grad_norm": 7.96875, | |
| "learning_rate": 3.4494477388045035e-06, | |
| "logits/chosen": -2.491364002227783, | |
| "logits/rejected": -2.235619068145752, | |
| "logps/chosen": -301.46600341796875, | |
| "logps/rejected": -238.8917999267578, | |
| "loss": 1.3419, | |
| "rewards/accuracies": 0.75, | |
| "rewards/chosen": 0.07640339434146881, | |
| "rewards/margins": 0.46867623925209045, | |
| "rewards/rejected": -0.39227285981178284, | |
| "step": 670 | |
| }, | |
| { | |
| "epoch": 0.4448806018972849, | |
| "grad_norm": 7.15625, | |
| "learning_rate": 3.3963642408649783e-06, | |
| "logits/chosen": -2.4565513134002686, | |
| "logits/rejected": -2.3826098442077637, | |
| "logps/chosen": -282.40228271484375, | |
| "logps/rejected": -249.9719696044922, | |
| "loss": 1.3423, | |
| "rewards/accuracies": 0.7749999761581421, | |
| "rewards/chosen": 0.1387428194284439, | |
| "rewards/margins": 0.49391037225723267, | |
| "rewards/rejected": -0.35516756772994995, | |
| "step": 680 | |
| }, | |
| { | |
| "epoch": 0.45142296368989204, | |
| "grad_norm": 8.5, | |
| "learning_rate": 3.3428128354552727e-06, | |
| "logits/chosen": -2.3810625076293945, | |
| "logits/rejected": -2.34759521484375, | |
| "logps/chosen": -252.3175048828125, | |
| "logps/rejected": -237.6392364501953, | |
| "loss": 1.3457, | |
| "rewards/accuracies": 0.699999988079071, | |
| "rewards/chosen": 0.10267229378223419, | |
| "rewards/margins": 0.4029284119606018, | |
| "rewards/rejected": -0.30025607347488403, | |
| "step": 690 | |
| }, | |
| { | |
| "epoch": 0.4579653254824992, | |
| "grad_norm": 6.375, | |
| "learning_rate": 3.2888214767278246e-06, | |
| "logits/chosen": -2.5027060508728027, | |
| "logits/rejected": -2.3413376808166504, | |
| "logps/chosen": -307.34686279296875, | |
| "logps/rejected": -240.319091796875, | |
| "loss": 1.3577, | |
| "rewards/accuracies": 0.675000011920929, | |
| "rewards/chosen": 0.04387027025222778, | |
| "rewards/margins": 0.33918899297714233, | |
| "rewards/rejected": -0.29531875252723694, | |
| "step": 700 | |
| }, | |
| { | |
| "epoch": 0.4579653254824992, | |
| "eval_logits/chosen": -2.2934322357177734, | |
| "eval_logits/rejected": -2.2099828720092773, | |
| "eval_logps/chosen": -270.8705749511719, | |
| "eval_logps/rejected": -256.79083251953125, | |
| "eval_loss": 1.345709204673767, | |
| "eval_rewards/accuracies": 0.7174999713897705, | |
| "eval_rewards/chosen": 0.13059137761592865, | |
| "eval_rewards/margins": 0.39938145875930786, | |
| "eval_rewards/rejected": -0.268790066242218, | |
| "eval_runtime": 192.4798, | |
| "eval_samples_per_second": 10.391, | |
| "eval_steps_per_second": 0.52, | |
| "step": 700 | |
| }, | |
| { | |
| "epoch": 0.46450768727510633, | |
| "grad_norm": 7.21875, | |
| "learning_rate": 3.2344183484933247e-06, | |
| "logits/chosen": -2.443279266357422, | |
| "logits/rejected": -2.3943705558776855, | |
| "logps/chosen": -256.82928466796875, | |
| "logps/rejected": -235.6039276123047, | |
| "loss": 1.3379, | |
| "rewards/accuracies": 0.699999988079071, | |
| "rewards/chosen": 0.02693643607199192, | |
| "rewards/margins": 0.40261945128440857, | |
| "rewards/rejected": -0.3756829798221588, | |
| "step": 710 | |
| }, | |
| { | |
| "epoch": 0.47105004906771347, | |
| "grad_norm": 9.625, | |
| "learning_rate": 3.179631849508597e-06, | |
| "logits/chosen": -2.445950984954834, | |
| "logits/rejected": -2.4440605640411377, | |
| "logps/chosen": -249.5127410888672, | |
| "logps/rejected": -268.5820007324219, | |
| "loss": 1.3417, | |
| "rewards/accuracies": 0.699999988079071, | |
| "rewards/chosen": 0.09297636151313782, | |
| "rewards/margins": 0.33975750207901, | |
| "rewards/rejected": -0.2467811554670334, | |
| "step": 720 | |
| }, | |
| { | |
| "epoch": 0.47759241086032056, | |
| "grad_norm": 7.59375, | |
| "learning_rate": 3.1244905786522796e-06, | |
| "logits/chosen": -2.4762067794799805, | |
| "logits/rejected": -2.4062716960906982, | |
| "logps/chosen": -264.5029602050781, | |
| "logps/rejected": -255.11361694335938, | |
| "loss": 1.3587, | |
| "rewards/accuracies": 0.737500011920929, | |
| "rewards/chosen": 0.11461961269378662, | |
| "rewards/margins": 0.3867848515510559, | |
| "rewards/rejected": -0.2721652388572693, | |
| "step": 730 | |
| }, | |
| { | |
| "epoch": 0.4841347726529277, | |
| "grad_norm": 7.4375, | |
| "learning_rate": 3.0690233199960393e-06, | |
| "logits/chosen": -2.3173129558563232, | |
| "logits/rejected": -2.2710585594177246, | |
| "logps/chosen": -227.1941375732422, | |
| "logps/rejected": -248.1180877685547, | |
| "loss": 1.3402, | |
| "rewards/accuracies": 0.800000011920929, | |
| "rewards/chosen": 0.1001671776175499, | |
| "rewards/margins": 0.45164409279823303, | |
| "rewards/rejected": -0.35147690773010254, | |
| "step": 740 | |
| }, | |
| { | |
| "epoch": 0.49067713444553485, | |
| "grad_norm": 7.4375, | |
| "learning_rate": 3.0132590277791163e-06, | |
| "logits/chosen": -2.524819850921631, | |
| "logits/rejected": -2.406615734100342, | |
| "logps/chosen": -263.93133544921875, | |
| "logps/rejected": -243.890625, | |
| "loss": 1.3328, | |
| "rewards/accuracies": 0.7749999761581421, | |
| "rewards/chosen": 0.14402124285697937, | |
| "rewards/margins": 0.5698949098587036, | |
| "rewards/rejected": -0.42587366700172424, | |
| "step": 750 | |
| }, | |
| { | |
| "epoch": 0.497219496238142, | |
| "grad_norm": 7.96875, | |
| "learning_rate": 2.9572268112940354e-06, | |
| "logits/chosen": -2.495116710662842, | |
| "logits/rejected": -2.4868061542510986, | |
| "logps/chosen": -254.06405639648438, | |
| "logps/rejected": -259.20379638671875, | |
| "loss": 1.3341, | |
| "rewards/accuracies": 0.7124999761581421, | |
| "rewards/chosen": 0.02839145064353943, | |
| "rewards/margins": 0.419182687997818, | |
| "rewards/rejected": -0.39079123735427856, | |
| "step": 760 | |
| }, | |
| { | |
| "epoch": 0.5037618580307491, | |
| "grad_norm": 8.1875, | |
| "learning_rate": 2.9009559196913882e-06, | |
| "logits/chosen": -2.5047004222869873, | |
| "logits/rejected": -2.386676788330078, | |
| "logps/chosen": -291.96478271484375, | |
| "logps/rejected": -252.67971801757812, | |
| "loss": 1.3422, | |
| "rewards/accuracies": 0.7250000238418579, | |
| "rewards/chosen": 0.1507759988307953, | |
| "rewards/margins": 0.5127144455909729, | |
| "rewards/rejected": -0.3619384467601776, | |
| "step": 770 | |
| }, | |
| { | |
| "epoch": 0.5103042198233563, | |
| "grad_norm": 7.84375, | |
| "learning_rate": 2.844475726711595e-06, | |
| "logits/chosen": -2.4655020236968994, | |
| "logits/rejected": -2.294151782989502, | |
| "logps/chosen": -249.9043731689453, | |
| "logps/rejected": -220.50216674804688, | |
| "loss": 1.3446, | |
| "rewards/accuracies": 0.675000011920929, | |
| "rewards/chosen": 0.06338523328304291, | |
| "rewards/margins": 0.47769251465797424, | |
| "rewards/rejected": -0.41430729627609253, | |
| "step": 780 | |
| }, | |
| { | |
| "epoch": 0.5168465816159633, | |
| "grad_norm": 6.0625, | |
| "learning_rate": 2.7878157153516446e-06, | |
| "logits/chosen": -2.4090731143951416, | |
| "logits/rejected": -2.3185999393463135, | |
| "logps/chosen": -270.53857421875, | |
| "logps/rejected": -269.4087829589844, | |
| "loss": 1.33, | |
| "rewards/accuracies": 0.7749999761581421, | |
| "rewards/chosen": 0.11772044748067856, | |
| "rewards/margins": 0.4702891707420349, | |
| "rewards/rejected": -0.35256871581077576, | |
| "step": 790 | |
| }, | |
| { | |
| "epoch": 0.5233889434085705, | |
| "grad_norm": 7.75, | |
| "learning_rate": 2.731005462474787e-06, | |
| "logits/chosen": -2.5132055282592773, | |
| "logits/rejected": -2.4322190284729004, | |
| "logps/chosen": -271.2453308105469, | |
| "logps/rejected": -266.37701416015625, | |
| "loss": 1.343, | |
| "rewards/accuracies": 0.7124999761581421, | |
| "rewards/chosen": 0.21996447443962097, | |
| "rewards/margins": 0.497943639755249, | |
| "rewards/rejected": -0.27797916531562805, | |
| "step": 800 | |
| }, | |
| { | |
| "epoch": 0.5233889434085705, | |
| "eval_logits/chosen": -2.3120627403259277, | |
| "eval_logits/rejected": -2.2312381267547607, | |
| "eval_logps/chosen": -271.3629150390625, | |
| "eval_logps/rejected": -257.9126892089844, | |
| "eval_loss": 1.3448688983917236, | |
| "eval_rewards/accuracies": 0.7149999737739563, | |
| "eval_rewards/chosen": 0.08135941624641418, | |
| "eval_rewards/margins": 0.4623354375362396, | |
| "eval_rewards/rejected": -0.38097602128982544, | |
| "eval_runtime": 192.5442, | |
| "eval_samples_per_second": 10.387, | |
| "eval_steps_per_second": 0.519, | |
| "step": 800 | |
| }, | |
| { | |
| "epoch": 0.5299313052011776, | |
| "grad_norm": 8.25, | |
| "learning_rate": 2.67407462337124e-06, | |
| "logits/chosen": -2.516996383666992, | |
| "logits/rejected": -2.399326801300049, | |
| "logps/chosen": -245.3385772705078, | |
| "logps/rejected": -241.2891082763672, | |
| "loss": 1.334, | |
| "rewards/accuracies": 0.75, | |
| "rewards/chosen": 0.0016076326137408614, | |
| "rewards/margins": 0.4354288578033447, | |
| "rewards/rejected": -0.43382126092910767, | |
| "step": 810 | |
| }, | |
| { | |
| "epoch": 0.5364736669937847, | |
| "grad_norm": 7.0, | |
| "learning_rate": 2.617052916277952e-06, | |
| "logits/chosen": -2.5313477516174316, | |
| "logits/rejected": -2.338625431060791, | |
| "logps/chosen": -253.85226440429688, | |
| "logps/rejected": -227.67898559570312, | |
| "loss": 1.3429, | |
| "rewards/accuracies": 0.762499988079071, | |
| "rewards/chosen": 0.07979673147201538, | |
| "rewards/margins": 0.5423867106437683, | |
| "rewards/rejected": -0.4625900387763977, | |
| "step": 820 | |
| }, | |
| { | |
| "epoch": 0.5430160287863919, | |
| "grad_norm": 7.1875, | |
| "learning_rate": 2.5599701068654985e-06, | |
| "logits/chosen": -2.5015134811401367, | |
| "logits/rejected": -2.4792141914367676, | |
| "logps/chosen": -250.22543334960938, | |
| "logps/rejected": -245.0906219482422, | |
| "loss": 1.3437, | |
| "rewards/accuracies": 0.762499988079071, | |
| "rewards/chosen": 0.13883474469184875, | |
| "rewards/margins": 0.5313934087753296, | |
| "rewards/rejected": -0.39255863428115845, | |
| "step": 830 | |
| }, | |
| { | |
| "epoch": 0.549558390578999, | |
| "grad_norm": 7.15625, | |
| "learning_rate": 2.5028559927002326e-06, | |
| "logits/chosen": -2.562939405441284, | |
| "logits/rejected": -2.4629690647125244, | |
| "logps/chosen": -295.15142822265625, | |
| "logps/rejected": -249.9164276123047, | |
| "loss": 1.3382, | |
| "rewards/accuracies": 0.8374999761581421, | |
| "rewards/chosen": 0.11802862584590912, | |
| "rewards/margins": 0.46983498334884644, | |
| "rewards/rejected": -0.3518063426017761, | |
| "step": 840 | |
| }, | |
| { | |
| "epoch": 0.5561007523716062, | |
| "grad_norm": 10.4375, | |
| "learning_rate": 2.4457403876897756e-06, | |
| "logits/chosen": -2.3556885719299316, | |
| "logits/rejected": -2.300297737121582, | |
| "logps/chosen": -241.96975708007812, | |
| "logps/rejected": -196.66481018066406, | |
| "loss": 1.3402, | |
| "rewards/accuracies": 0.6499999761581421, | |
| "rewards/chosen": 0.06055239588022232, | |
| "rewards/margins": 0.29153305292129517, | |
| "rewards/rejected": -0.23098066449165344, | |
| "step": 850 | |
| }, | |
| { | |
| "epoch": 0.5626431141642133, | |
| "grad_norm": 7.84375, | |
| "learning_rate": 2.388653106519975e-06, | |
| "logits/chosen": -2.504599094390869, | |
| "logits/rejected": -2.3005897998809814, | |
| "logps/chosen": -257.0250549316406, | |
| "logps/rejected": -192.78662109375, | |
| "loss": 1.3466, | |
| "rewards/accuracies": 0.699999988079071, | |
| "rewards/chosen": 0.04565539211034775, | |
| "rewards/margins": 0.4032720923423767, | |
| "rewards/rejected": -0.35761672258377075, | |
| "step": 860 | |
| }, | |
| { | |
| "epoch": 0.5691854759568205, | |
| "grad_norm": 7.21875, | |
| "learning_rate": 2.331623949091467e-06, | |
| "logits/chosen": -2.5008959770202637, | |
| "logits/rejected": -2.472198009490967, | |
| "logps/chosen": -277.51580810546875, | |
| "logps/rejected": -266.95245361328125, | |
| "loss": 1.3361, | |
| "rewards/accuracies": 0.699999988079071, | |
| "rewards/chosen": 0.13331273198127747, | |
| "rewards/margins": 0.45952072739601135, | |
| "rewards/rejected": -0.3262080252170563, | |
| "step": 870 | |
| }, | |
| { | |
| "epoch": 0.5757278377494276, | |
| "grad_norm": 7.75, | |
| "learning_rate": 2.2746826849639513e-06, | |
| "logits/chosen": -2.505166530609131, | |
| "logits/rejected": -2.3743433952331543, | |
| "logps/chosen": -281.8514709472656, | |
| "logps/rejected": -239.01443481445312, | |
| "loss": 1.3391, | |
| "rewards/accuracies": 0.699999988079071, | |
| "rewards/chosen": 0.11163483560085297, | |
| "rewards/margins": 0.4709078371524811, | |
| "rewards/rejected": -0.35927295684814453, | |
| "step": 880 | |
| }, | |
| { | |
| "epoch": 0.5822701995420346, | |
| "grad_norm": 7.1875, | |
| "learning_rate": 2.2178590378162957e-06, | |
| "logits/chosen": -2.5283544063568115, | |
| "logits/rejected": -2.402369260787964, | |
| "logps/chosen": -319.1461486816406, | |
| "logps/rejected": -279.74462890625, | |
| "loss": 1.3521, | |
| "rewards/accuracies": 0.737500011920929, | |
| "rewards/chosen": 0.02043941244482994, | |
| "rewards/margins": 0.45712727308273315, | |
| "rewards/rejected": -0.4366879463195801, | |
| "step": 890 | |
| }, | |
| { | |
| "epoch": 0.5888125613346418, | |
| "grad_norm": 7.15625, | |
| "learning_rate": 2.1611826699306104e-06, | |
| "logits/chosen": -2.4243669509887695, | |
| "logits/rejected": -2.32483172416687, | |
| "logps/chosen": -271.6533203125, | |
| "logps/rejected": -232.5511474609375, | |
| "loss": 1.3439, | |
| "rewards/accuracies": 0.699999988079071, | |
| "rewards/chosen": 0.13815268874168396, | |
| "rewards/margins": 0.49229878187179565, | |
| "rewards/rejected": -0.35414618253707886, | |
| "step": 900 | |
| }, | |
| { | |
| "epoch": 0.5888125613346418, | |
| "eval_logits/chosen": -2.313720941543579, | |
| "eval_logits/rejected": -2.2327077388763428, | |
| "eval_logps/chosen": -271.791748046875, | |
| "eval_logps/rejected": -258.1572570800781, | |
| "eval_loss": 1.3458954095840454, | |
| "eval_rewards/accuracies": 0.7250000238418579, | |
| "eval_rewards/chosen": 0.038476575165987015, | |
| "eval_rewards/margins": 0.443908154964447, | |
| "eval_rewards/rejected": -0.4054316282272339, | |
| "eval_runtime": 192.5142, | |
| "eval_samples_per_second": 10.389, | |
| "eval_steps_per_second": 0.519, | |
| "step": 900 | |
| }, | |
| { | |
| "epoch": 0.5953549231272489, | |
| "grad_norm": 8.0625, | |
| "learning_rate": 2.1046831667083483e-06, | |
| "logits/chosen": -2.4782989025115967, | |
| "logits/rejected": -2.3582215309143066, | |
| "logps/chosen": -341.8036193847656, | |
| "logps/rejected": -267.46795654296875, | |
| "loss": 1.3463, | |
| "rewards/accuracies": 0.8125, | |
| "rewards/chosen": 0.08111786842346191, | |
| "rewards/margins": 0.4356920123100281, | |
| "rewards/rejected": -0.35457414388656616, | |
| "step": 910 | |
| }, | |
| { | |
| "epoch": 0.6018972849198561, | |
| "grad_norm": 7.09375, | |
| "learning_rate": 2.048390021226559e-06, | |
| "logits/chosen": -2.445460557937622, | |
| "logits/rejected": -2.3675286769866943, | |
| "logps/chosen": -262.48907470703125, | |
| "logps/rejected": -266.7914733886719, | |
| "loss": 1.3489, | |
| "rewards/accuracies": 0.762499988079071, | |
| "rewards/chosen": 0.1052175760269165, | |
| "rewards/margins": 0.4468202590942383, | |
| "rewards/rejected": -0.341602623462677, | |
| "step": 920 | |
| }, | |
| { | |
| "epoch": 0.6084396467124632, | |
| "grad_norm": 7.25, | |
| "learning_rate": 1.9923326188423212e-06, | |
| "logits/chosen": -2.5053839683532715, | |
| "logits/rejected": -2.4001247882843018, | |
| "logps/chosen": -236.89382934570312, | |
| "logps/rejected": -205.64291381835938, | |
| "loss": 1.3523, | |
| "rewards/accuracies": 0.675000011920929, | |
| "rewards/chosen": 0.08720345795154572, | |
| "rewards/margins": 0.3279283344745636, | |
| "rewards/rejected": -0.24072487652301788, | |
| "step": 930 | |
| }, | |
| { | |
| "epoch": 0.6149820085050703, | |
| "grad_norm": 7.4375, | |
| "learning_rate": 1.936540221853415e-06, | |
| "logits/chosen": -2.4638781547546387, | |
| "logits/rejected": -2.3390696048736572, | |
| "logps/chosen": -283.2751770019531, | |
| "logps/rejected": -238.634765625, | |
| "loss": 1.3297, | |
| "rewards/accuracies": 0.75, | |
| "rewards/chosen": 0.058457307517528534, | |
| "rewards/margins": 0.6001859307289124, | |
| "rewards/rejected": -0.5417286157608032, | |
| "step": 940 | |
| }, | |
| { | |
| "epoch": 0.6215243702976775, | |
| "grad_norm": 6.71875, | |
| "learning_rate": 1.8810419542232245e-06, | |
| "logits/chosen": -2.5065360069274902, | |
| "logits/rejected": -2.3907275199890137, | |
| "logps/chosen": -273.1107177734375, | |
| "logps/rejected": -269.38140869140625, | |
| "loss": 1.3301, | |
| "rewards/accuracies": 0.762499988079071, | |
| "rewards/chosen": 0.07303529977798462, | |
| "rewards/margins": 0.37555113434791565, | |
| "rewards/rejected": -0.30251583456993103, | |
| "step": 950 | |
| }, | |
| { | |
| "epoch": 0.6280667320902846, | |
| "grad_norm": 6.71875, | |
| "learning_rate": 1.8258667863778573e-06, | |
| "logits/chosen": -2.415342092514038, | |
| "logits/rejected": -2.3436694145202637, | |
| "logps/chosen": -305.41937255859375, | |
| "logps/rejected": -254.4953155517578, | |
| "loss": 1.3252, | |
| "rewards/accuracies": 0.699999988079071, | |
| "rewards/chosen": 0.03756101056933403, | |
| "rewards/margins": 0.33726876974105835, | |
| "rewards/rejected": -0.2997077405452728, | |
| "step": 960 | |
| }, | |
| { | |
| "epoch": 0.6346090938828918, | |
| "grad_norm": 7.09375, | |
| "learning_rate": 1.7710435200834126e-06, | |
| "logits/chosen": -2.424187660217285, | |
| "logits/rejected": -2.4013800621032715, | |
| "logps/chosen": -253.3881378173828, | |
| "logps/rejected": -242.24853515625, | |
| "loss": 1.3294, | |
| "rewards/accuracies": 0.737500011920929, | |
| "rewards/chosen": 0.09904266893863678, | |
| "rewards/margins": 0.44573745131492615, | |
| "rewards/rejected": -0.3466947376728058, | |
| "step": 970 | |
| }, | |
| { | |
| "epoch": 0.6411514556754988, | |
| "grad_norm": 10.1875, | |
| "learning_rate": 1.7166007734112808e-06, | |
| "logits/chosen": -2.512207269668579, | |
| "logits/rejected": -2.3953731060028076, | |
| "logps/chosen": -282.1387939453125, | |
| "logps/rejected": -238.2812042236328, | |
| "loss": 1.3447, | |
| "rewards/accuracies": 0.75, | |
| "rewards/chosen": 0.10133460909128189, | |
| "rewards/margins": 0.4784146845340729, | |
| "rewards/rejected": -0.3770800232887268, | |
| "step": 980 | |
| }, | |
| { | |
| "epoch": 0.647693817468106, | |
| "grad_norm": 6.84375, | |
| "learning_rate": 1.6625669657993483e-06, | |
| "logits/chosen": -2.588611125946045, | |
| "logits/rejected": -2.5047097206115723, | |
| "logps/chosen": -316.20037841796875, | |
| "logps/rejected": -254.8091278076172, | |
| "loss": 1.3237, | |
| "rewards/accuracies": 0.862500011920929, | |
| "rewards/chosen": 0.18245932459831238, | |
| "rewards/margins": 0.5891641974449158, | |
| "rewards/rejected": -0.40670496225357056, | |
| "step": 990 | |
| }, | |
| { | |
| "epoch": 0.6542361792607131, | |
| "grad_norm": 6.90625, | |
| "learning_rate": 1.6089703032168736e-06, | |
| "logits/chosen": -2.417908191680908, | |
| "logits/rejected": -2.41746187210083, | |
| "logps/chosen": -254.3357391357422, | |
| "logps/rejected": -241.4535675048828, | |
| "loss": 1.3388, | |
| "rewards/accuracies": 0.7250000238418579, | |
| "rewards/chosen": 0.06749463826417923, | |
| "rewards/margins": 0.35123783349990845, | |
| "rewards/rejected": -0.2837432026863098, | |
| "step": 1000 | |
| }, | |
| { | |
| "epoch": 0.6542361792607131, | |
| "eval_logits/chosen": -2.318289041519165, | |
| "eval_logits/rejected": -2.238693952560425, | |
| "eval_logps/chosen": -270.0261535644531, | |
| "eval_logps/rejected": -256.7276611328125, | |
| "eval_loss": 1.344198226928711, | |
| "eval_rewards/accuracies": 0.7325000166893005, | |
| "eval_rewards/chosen": 0.215036541223526, | |
| "eval_rewards/margins": 0.4775146245956421, | |
| "eval_rewards/rejected": -0.2624781131744385, | |
| "eval_runtime": 192.5272, | |
| "eval_samples_per_second": 10.388, | |
| "eval_steps_per_second": 0.519, | |
| "step": 1000 | |
| }, | |
| { | |
| "epoch": 0.6607785410533202, | |
| "grad_norm": 7.8125, | |
| "learning_rate": 1.55583876344081e-06, | |
| "logits/chosen": -2.4103312492370605, | |
| "logits/rejected": -2.1900105476379395, | |
| "logps/chosen": -260.7649230957031, | |
| "logps/rejected": -209.84048461914062, | |
| "loss": 1.3325, | |
| "rewards/accuracies": 0.762499988079071, | |
| "rewards/chosen": 0.10994074493646622, | |
| "rewards/margins": 0.5784383416175842, | |
| "rewards/rejected": -0.46849751472473145, | |
| "step": 1010 | |
| }, | |
| { | |
| "epoch": 0.6673209028459274, | |
| "grad_norm": 7.3125, | |
| "learning_rate": 1.5032000814512372e-06, | |
| "logits/chosen": -2.4817819595336914, | |
| "logits/rejected": -2.4498400688171387, | |
| "logps/chosen": -240.53134155273438, | |
| "logps/rejected": -238.8572998046875, | |
| "loss": 1.3307, | |
| "rewards/accuracies": 0.6875, | |
| "rewards/chosen": -0.0427989736199379, | |
| "rewards/margins": 0.3493194580078125, | |
| "rewards/rejected": -0.3921184539794922, | |
| "step": 1020 | |
| }, | |
| { | |
| "epoch": 0.6738632646385345, | |
| "grad_norm": 5.59375, | |
| "learning_rate": 1.4510817349535323e-06, | |
| "logits/chosen": -2.453907012939453, | |
| "logits/rejected": -2.399127244949341, | |
| "logps/chosen": -262.92559814453125, | |
| "logps/rejected": -249.8006134033203, | |
| "loss": 1.3411, | |
| "rewards/accuracies": 0.762499988079071, | |
| "rewards/chosen": 0.13903044164180756, | |
| "rewards/margins": 0.5323747396469116, | |
| "rewards/rejected": -0.39334431290626526, | |
| "step": 1030 | |
| }, | |
| { | |
| "epoch": 0.6804056264311417, | |
| "grad_norm": 8.25, | |
| "learning_rate": 1.3995109300348537e-06, | |
| "logits/chosen": -2.5314221382141113, | |
| "logits/rejected": -2.424898624420166, | |
| "logps/chosen": -296.252197265625, | |
| "logps/rejected": -279.55096435546875, | |
| "loss": 1.3372, | |
| "rewards/accuracies": 0.699999988079071, | |
| "rewards/chosen": -0.001908986596390605, | |
| "rewards/margins": 0.4961877763271332, | |
| "rewards/rejected": -0.49809688329696655, | |
| "step": 1040 | |
| }, | |
| { | |
| "epoch": 0.6869479882237488, | |
| "grad_norm": 7.03125, | |
| "learning_rate": 1.348514586962389e-06, | |
| "logits/chosen": -2.543003559112549, | |
| "logits/rejected": -2.466203212738037, | |
| "logps/chosen": -244.98037719726562, | |
| "logps/rejected": -231.51449584960938, | |
| "loss": 1.3373, | |
| "rewards/accuracies": 0.762499988079071, | |
| "rewards/chosen": 0.1332973688840866, | |
| "rewards/margins": 0.5175554752349854, | |
| "rewards/rejected": -0.38425812125205994, | |
| "step": 1050 | |
| }, | |
| { | |
| "epoch": 0.693490350016356, | |
| "grad_norm": 6.625, | |
| "learning_rate": 1.2981193261308284e-06, | |
| "logits/chosen": -2.485957622528076, | |
| "logits/rejected": -2.440640687942505, | |
| "logps/chosen": -243.56118774414062, | |
| "logps/rejected": -264.2271728515625, | |
| "loss": 1.3184, | |
| "rewards/accuracies": 0.800000011920929, | |
| "rewards/chosen": 0.058649301528930664, | |
| "rewards/margins": 0.4494194984436035, | |
| "rewards/rejected": -0.39077019691467285, | |
| "step": 1060 | |
| }, | |
| { | |
| "epoch": 0.700032711808963, | |
| "grad_norm": 7.625, | |
| "learning_rate": 1.2483514541663501e-06, | |
| "logits/chosen": -2.567582607269287, | |
| "logits/rejected": -2.398489475250244, | |
| "logps/chosen": -268.79248046875, | |
| "logps/rejected": -236.4168701171875, | |
| "loss": 1.3346, | |
| "rewards/accuracies": 0.7875000238418579, | |
| "rewards/chosen": 0.1122722402215004, | |
| "rewards/margins": 0.54461270570755, | |
| "rewards/rejected": -0.43234047293663025, | |
| "step": 1070 | |
| }, | |
| { | |
| "epoch": 0.7065750736015701, | |
| "grad_norm": 9.875, | |
| "learning_rate": 1.1992369501944096e-06, | |
| "logits/chosen": -2.6160035133361816, | |
| "logits/rejected": -2.485849380493164, | |
| "logps/chosen": -301.0981140136719, | |
| "logps/rejected": -261.76226806640625, | |
| "loss": 1.3344, | |
| "rewards/accuracies": 0.7749999761581421, | |
| "rewards/chosen": 0.12587206065654755, | |
| "rewards/margins": 0.49982064962387085, | |
| "rewards/rejected": -0.3739486336708069, | |
| "step": 1080 | |
| }, | |
| { | |
| "epoch": 0.7131174353941773, | |
| "grad_norm": 5.15625, | |
| "learning_rate": 1.1508014522784803e-06, | |
| "logits/chosen": -2.5031826496124268, | |
| "logits/rejected": -2.380375862121582, | |
| "logps/chosen": -259.3973693847656, | |
| "logps/rejected": -242.9427490234375, | |
| "loss": 1.3325, | |
| "rewards/accuracies": 0.75, | |
| "rewards/chosen": 0.098679319024086, | |
| "rewards/margins": 0.5007960796356201, | |
| "rewards/rejected": -0.40211671590805054, | |
| "step": 1090 | |
| }, | |
| { | |
| "epoch": 0.7196597971867844, | |
| "grad_norm": 6.3125, | |
| "learning_rate": 1.1030702440368319e-06, | |
| "logits/chosen": -2.4038383960723877, | |
| "logits/rejected": -2.402168035507202, | |
| "logps/chosen": -282.268798828125, | |
| "logps/rejected": -266.29132080078125, | |
| "loss": 1.3186, | |
| "rewards/accuracies": 0.8125, | |
| "rewards/chosen": 0.16258536279201508, | |
| "rewards/margins": 0.7569273114204407, | |
| "rewards/rejected": -0.5943418741226196, | |
| "step": 1100 | |
| }, | |
| { | |
| "epoch": 0.7196597971867844, | |
| "eval_logits/chosen": -2.3106985092163086, | |
| "eval_logits/rejected": -2.2305588722229004, | |
| "eval_logps/chosen": -270.9344787597656, | |
| "eval_logps/rejected": -257.6895446777344, | |
| "eval_loss": 1.3422751426696777, | |
| "eval_rewards/accuracies": 0.7325000166893005, | |
| "eval_rewards/chosen": 0.12420222908258438, | |
| "eval_rewards/margins": 0.48286232352256775, | |
| "eval_rewards/rejected": -0.3586600422859192, | |
| "eval_runtime": 192.5343, | |
| "eval_samples_per_second": 10.388, | |
| "eval_steps_per_second": 0.519, | |
| "step": 1100 | |
| }, | |
| { | |
| "epoch": 0.7262021589793916, | |
| "grad_norm": 8.0625, | |
| "learning_rate": 1.0560682414443315e-06, | |
| "logits/chosen": -2.492471694946289, | |
| "logits/rejected": -2.3619418144226074, | |
| "logps/chosen": -296.5441589355469, | |
| "logps/rejected": -249.51174926757812, | |
| "loss": 1.333, | |
| "rewards/accuracies": 0.800000011920929, | |
| "rewards/chosen": 0.08190511167049408, | |
| "rewards/margins": 0.522915244102478, | |
| "rewards/rejected": -0.44101008772850037, | |
| "step": 1110 | |
| }, | |
| { | |
| "epoch": 0.7327445207719987, | |
| "grad_norm": 7.03125, | |
| "learning_rate": 1.009819979826156e-06, | |
| "logits/chosen": -2.49381947517395, | |
| "logits/rejected": -2.4158735275268555, | |
| "logps/chosen": -277.04742431640625, | |
| "logps/rejected": -277.2239074707031, | |
| "loss": 1.3209, | |
| "rewards/accuracies": 0.800000011920929, | |
| "rewards/chosen": 0.1789507120847702, | |
| "rewards/margins": 0.5798290967941284, | |
| "rewards/rejected": -0.4008784294128418, | |
| "step": 1120 | |
| }, | |
| { | |
| "epoch": 0.7392868825646058, | |
| "grad_norm": 9.875, | |
| "learning_rate": 9.643496010502054e-07, | |
| "logits/chosen": -2.383232831954956, | |
| "logits/rejected": -2.3699889183044434, | |
| "logps/chosen": -282.37872314453125, | |
| "logps/rejected": -288.27764892578125, | |
| "loss": 1.3569, | |
| "rewards/accuracies": 0.699999988079071, | |
| "rewards/chosen": 0.049896519631147385, | |
| "rewards/margins": 0.3302707076072693, | |
| "rewards/rejected": -0.280374139547348, | |
| "step": 1130 | |
| }, | |
| { | |
| "epoch": 0.745829244357213, | |
| "grad_norm": 7.75, | |
| "learning_rate": 9.196808409249086e-07, | |
| "logits/chosen": -2.4595417976379395, | |
| "logits/rejected": -2.346752882003784, | |
| "logps/chosen": -263.95172119140625, | |
| "logps/rejected": -211.8850555419922, | |
| "loss": 1.3457, | |
| "rewards/accuracies": 0.7124999761581421, | |
| "rewards/chosen": 0.07734758406877518, | |
| "rewards/margins": 0.34833234548568726, | |
| "rewards/rejected": -0.2709847092628479, | |
| "step": 1140 | |
| }, | |
| { | |
| "epoch": 0.7523716061498201, | |
| "grad_norm": 8.25, | |
| "learning_rate": 8.758370168089797e-07, | |
| "logits/chosen": -2.46266770362854, | |
| "logits/rejected": -2.4414191246032715, | |
| "logps/chosen": -258.66815185546875, | |
| "logps/rejected": -272.05242919921875, | |
| "loss": 1.3453, | |
| "rewards/accuracies": 0.75, | |
| "rewards/chosen": 0.12471766769886017, | |
| "rewards/margins": 0.5029908418655396, | |
| "rewards/rejected": -0.3782731890678406, | |
| "step": 1150 | |
| }, | |
| { | |
| "epoch": 0.7589139679424273, | |
| "grad_norm": 7.90625, | |
| "learning_rate": 8.328410154396318e-07, | |
| "logits/chosen": -2.473407030105591, | |
| "logits/rejected": -2.4644668102264404, | |
| "logps/chosen": -270.7806396484375, | |
| "logps/rejected": -228.4518280029297, | |
| "loss": 1.3378, | |
| "rewards/accuracies": 0.7124999761581421, | |
| "rewards/chosen": 0.08448322862386703, | |
| "rewards/margins": 0.4490872323513031, | |
| "rewards/rejected": -0.36460399627685547, | |
| "step": 1160 | |
| }, | |
| { | |
| "epoch": 0.7654563297350343, | |
| "grad_norm": 6.28125, | |
| "learning_rate": 7.907152809855529e-07, | |
| "logits/chosen": -2.4258522987365723, | |
| "logits/rejected": -2.3456382751464844, | |
| "logps/chosen": -260.29730224609375, | |
| "logps/rejected": -251.9761199951172, | |
| "loss": 1.3517, | |
| "rewards/accuracies": 0.625, | |
| "rewards/chosen": 0.03364313393831253, | |
| "rewards/margins": 0.29857534170150757, | |
| "rewards/rejected": -0.26493218541145325, | |
| "step": 1170 | |
| }, | |
| { | |
| "epoch": 0.7719986915276414, | |
| "grad_norm": 11.1875, | |
| "learning_rate": 7.494818033309207e-07, | |
| "logits/chosen": -2.4767394065856934, | |
| "logits/rejected": -2.3014492988586426, | |
| "logps/chosen": -238.3695526123047, | |
| "logps/rejected": -213.4428253173828, | |
| "loss": 1.3469, | |
| "rewards/accuracies": 0.7250000238418579, | |
| "rewards/chosen": 0.12686462700366974, | |
| "rewards/margins": 0.3980763554573059, | |
| "rewards/rejected": -0.27121174335479736, | |
| "step": 1180 | |
| }, | |
| { | |
| "epoch": 0.7785410533202486, | |
| "grad_norm": 7.1875, | |
| "learning_rate": 7.091621065965521e-07, | |
| "logits/chosen": -2.454749584197998, | |
| "logits/rejected": -2.4374210834503174, | |
| "logps/chosen": -276.31488037109375, | |
| "logps/rejected": -310.75, | |
| "loss": 1.3416, | |
| "rewards/accuracies": 0.6625000238418579, | |
| "rewards/chosen": 0.041704945266246796, | |
| "rewards/margins": 0.38392168283462524, | |
| "rewards/rejected": -0.34221673011779785, | |
| "step": 1190 | |
| }, | |
| { | |
| "epoch": 0.7850834151128557, | |
| "grad_norm": 7.34375, | |
| "learning_rate": 6.697772379041823e-07, | |
| "logits/chosen": -2.4566853046417236, | |
| "logits/rejected": -2.416715145111084, | |
| "logps/chosen": -245.9681396484375, | |
| "logps/rejected": -241.95858764648438, | |
| "loss": 1.3299, | |
| "rewards/accuracies": 0.6875, | |
| "rewards/chosen": 0.12603041529655457, | |
| "rewards/margins": 0.42295509576797485, | |
| "rewards/rejected": -0.2969246804714203, | |
| "step": 1200 | |
| }, | |
| { | |
| "epoch": 0.7850834151128557, | |
| "eval_logits/chosen": -2.3078200817108154, | |
| "eval_logits/rejected": -2.2274787425994873, | |
| "eval_logps/chosen": -270.70892333984375, | |
| "eval_logps/rejected": -257.372802734375, | |
| "eval_loss": 1.3416602611541748, | |
| "eval_rewards/accuracies": 0.7425000071525574, | |
| "eval_rewards/chosen": 0.14675647020339966, | |
| "eval_rewards/margins": 0.4737465977668762, | |
| "eval_rewards/rejected": -0.3269902169704437, | |
| "eval_runtime": 192.5364, | |
| "eval_samples_per_second": 10.388, | |
| "eval_steps_per_second": 0.519, | |
| "step": 1200 | |
| }, | |
| { | |
| "epoch": 0.7916257769054629, | |
| "grad_norm": 5.8125, | |
| "learning_rate": 6.313477563897466e-07, | |
| "logits/chosen": -2.516838550567627, | |
| "logits/rejected": -2.4841699600219727, | |
| "logps/chosen": -243.35372924804688, | |
| "logps/rejected": -246.02450561523438, | |
| "loss": 1.3459, | |
| "rewards/accuracies": 0.762499988079071, | |
| "rewards/chosen": 0.11812669038772583, | |
| "rewards/margins": 0.5516510009765625, | |
| "rewards/rejected": -0.43352431058883667, | |
| "step": 1210 | |
| }, | |
| { | |
| "epoch": 0.79816813869807, | |
| "grad_norm": 21.5, | |
| "learning_rate": 5.9389372247138e-07, | |
| "logits/chosen": -2.493438482284546, | |
| "logits/rejected": -2.3540115356445312, | |
| "logps/chosen": -215.5032501220703, | |
| "logps/rejected": -212.7133026123047, | |
| "loss": 1.3261, | |
| "rewards/accuracies": 0.8374999761581421, | |
| "rewards/chosen": 0.16954867541790009, | |
| "rewards/margins": 0.5659071803092957, | |
| "rewards/rejected": -0.39635851979255676, | |
| "step": 1220 | |
| }, | |
| { | |
| "epoch": 0.8047105004906772, | |
| "grad_norm": 13.0625, | |
| "learning_rate": 5.574346873777714e-07, | |
| "logits/chosen": -2.4212934970855713, | |
| "logits/rejected": -2.4491488933563232, | |
| "logps/chosen": -249.1747589111328, | |
| "logps/rejected": -282.01385498046875, | |
| "loss": 1.3411, | |
| "rewards/accuracies": 0.75, | |
| "rewards/chosen": 0.17586642503738403, | |
| "rewards/margins": 0.467965304851532, | |
| "rewards/rejected": -0.29209887981414795, | |
| "step": 1230 | |
| }, | |
| { | |
| "epoch": 0.8112528622832843, | |
| "grad_norm": 7.5625, | |
| "learning_rate": 5.219896829422927e-07, | |
| "logits/chosen": -2.491405487060547, | |
| "logits/rejected": -2.3750882148742676, | |
| "logps/chosen": -252.0865020751953, | |
| "logps/rejected": -231.3611602783203, | |
| "loss": 1.3306, | |
| "rewards/accuracies": 0.7250000238418579, | |
| "rewards/chosen": 0.12007620185613632, | |
| "rewards/margins": 0.5494218468666077, | |
| "rewards/rejected": -0.42934560775756836, | |
| "step": 1240 | |
| }, | |
| { | |
| "epoch": 0.8177952240758914, | |
| "grad_norm": 7.5, | |
| "learning_rate": 4.875772116682817e-07, | |
| "logits/chosen": -2.359609842300415, | |
| "logits/rejected": -2.25789475440979, | |
| "logps/chosen": -228.32534790039062, | |
| "logps/rejected": -250.2344512939453, | |
| "loss": 1.3465, | |
| "rewards/accuracies": 0.699999988079071, | |
| "rewards/chosen": 0.0579691156744957, | |
| "rewards/margins": 0.4779117703437805, | |
| "rewards/rejected": -0.41994261741638184, | |
| "step": 1250 | |
| }, | |
| { | |
| "epoch": 0.8243375858684985, | |
| "grad_norm": 8.125, | |
| "learning_rate": 4.542152370706149e-07, | |
| "logits/chosen": -2.395474672317505, | |
| "logits/rejected": -2.3001809120178223, | |
| "logps/chosen": -254.91268920898438, | |
| "logps/rejected": -258.6715087890625, | |
| "loss": 1.331, | |
| "rewards/accuracies": 0.675000011920929, | |
| "rewards/chosen": 0.06895018368959427, | |
| "rewards/margins": 0.4956684112548828, | |
| "rewards/rejected": -0.4267183244228363, | |
| "step": 1260 | |
| }, | |
| { | |
| "epoch": 0.8308799476611056, | |
| "grad_norm": 6.25, | |
| "learning_rate": 4.2192117429865067e-07, | |
| "logits/chosen": -2.28416109085083, | |
| "logits/rejected": -2.2375693321228027, | |
| "logps/chosen": -239.7650909423828, | |
| "logps/rejected": -263.7681579589844, | |
| "loss": 1.3323, | |
| "rewards/accuracies": 0.6625000238418579, | |
| "rewards/chosen": 0.11223635822534561, | |
| "rewards/margins": 0.42569345235824585, | |
| "rewards/rejected": -0.3134571313858032, | |
| "step": 1270 | |
| }, | |
| { | |
| "epoch": 0.8374223094537128, | |
| "grad_norm": 6.96875, | |
| "learning_rate": 3.907118810454172e-07, | |
| "logits/chosen": -2.401803970336914, | |
| "logits/rejected": -2.3998618125915527, | |
| "logps/chosen": -274.4956359863281, | |
| "logps/rejected": -298.844970703125, | |
| "loss": 1.3443, | |
| "rewards/accuracies": 0.7250000238418579, | |
| "rewards/chosen": 0.061115562915802, | |
| "rewards/margins": 0.4573536813259125, | |
| "rewards/rejected": -0.39623817801475525, | |
| "step": 1280 | |
| }, | |
| { | |
| "epoch": 0.8439646712463199, | |
| "grad_norm": 6.46875, | |
| "learning_rate": 3.6060364874779455e-07, | |
| "logits/chosen": -2.399519443511963, | |
| "logits/rejected": -2.386014461517334, | |
| "logps/chosen": -262.71044921875, | |
| "logps/rejected": -251.5861358642578, | |
| "loss": 1.351, | |
| "rewards/accuracies": 0.637499988079071, | |
| "rewards/chosen": 0.02161099575459957, | |
| "rewards/margins": 0.339490681886673, | |
| "rewards/rejected": -0.31787967681884766, | |
| "step": 1290 | |
| }, | |
| { | |
| "epoch": 0.850507033038927, | |
| "grad_norm": 6.25, | |
| "learning_rate": 3.3161219408229026e-07, | |
| "logits/chosen": -2.5390801429748535, | |
| "logits/rejected": -2.4665403366088867, | |
| "logps/chosen": -261.7703552246094, | |
| "logps/rejected": -245.41683959960938, | |
| "loss": 1.3248, | |
| "rewards/accuracies": 0.7124999761581421, | |
| "rewards/chosen": 0.11996345221996307, | |
| "rewards/margins": 0.4374065399169922, | |
| "rewards/rejected": -0.31744304299354553, | |
| "step": 1300 | |
| }, | |
| { | |
| "epoch": 0.850507033038927, | |
| "eval_logits/chosen": -2.3104758262634277, | |
| "eval_logits/rejected": -2.2305808067321777, | |
| "eval_logps/chosen": -270.62164306640625, | |
| "eval_logps/rejected": -257.2347106933594, | |
| "eval_loss": 1.341292381286621, | |
| "eval_rewards/accuracies": 0.7524999976158142, | |
| "eval_rewards/chosen": 0.15548919141292572, | |
| "eval_rewards/margins": 0.4686690866947174, | |
| "eval_rewards/rejected": -0.3131798803806305, | |
| "eval_runtime": 192.5429, | |
| "eval_samples_per_second": 10.387, | |
| "eval_steps_per_second": 0.519, | |
| "step": 1300 | |
| }, | |
| { | |
| "epoch": 0.8570493948315342, | |
| "grad_norm": 6.34375, | |
| "learning_rate": 3.0375265076083796e-07, | |
| "logits/chosen": -2.5264501571655273, | |
| "logits/rejected": -2.514683246612549, | |
| "logps/chosen": -277.8392028808594, | |
| "logps/rejected": -238.16134643554688, | |
| "loss": 1.3384, | |
| "rewards/accuracies": 0.699999988079071, | |
| "rewards/chosen": 0.1312626153230667, | |
| "rewards/margins": 0.3692128658294678, | |
| "rewards/rejected": -0.23795023560523987, | |
| "step": 1310 | |
| }, | |
| { | |
| "epoch": 0.8635917566241413, | |
| "grad_norm": 7.1875, | |
| "learning_rate": 2.7703956163091153e-07, | |
| "logits/chosen": -2.4872682094573975, | |
| "logits/rejected": -2.399378538131714, | |
| "logps/chosen": -252.8599853515625, | |
| "logps/rejected": -243.26425170898438, | |
| "loss": 1.3411, | |
| "rewards/accuracies": 0.824999988079071, | |
| "rewards/chosen": 0.09871190041303635, | |
| "rewards/margins": 0.4442734718322754, | |
| "rewards/rejected": -0.3455616533756256, | |
| "step": 1320 | |
| }, | |
| { | |
| "epoch": 0.8701341184167485, | |
| "grad_norm": 7.71875, | |
| "learning_rate": 2.514868710840723e-07, | |
| "logits/chosen": -2.482853889465332, | |
| "logits/rejected": -2.420342445373535, | |
| "logps/chosen": -310.83123779296875, | |
| "logps/rejected": -267.9164123535156, | |
| "loss": 1.3399, | |
| "rewards/accuracies": 0.625, | |
| "rewards/chosen": 0.10147074609994888, | |
| "rewards/margins": 0.3591291904449463, | |
| "rewards/rejected": -0.2576584219932556, | |
| "step": 1330 | |
| }, | |
| { | |
| "epoch": 0.8766764802093556, | |
| "grad_norm": 7.09375, | |
| "learning_rate": 2.271079177769117e-07, | |
| "logits/chosen": -2.4147746562957764, | |
| "logits/rejected": -2.420971632003784, | |
| "logps/chosen": -284.0067138671875, | |
| "logps/rejected": -307.20989990234375, | |
| "loss": 1.3301, | |
| "rewards/accuracies": 0.6875, | |
| "rewards/chosen": 0.12000395357608795, | |
| "rewards/margins": 0.36496061086654663, | |
| "rewards/rejected": -0.24495668709278107, | |
| "step": 1340 | |
| }, | |
| { | |
| "epoch": 0.8832188420019627, | |
| "grad_norm": 6.53125, | |
| "learning_rate": 2.0391542766819456e-07, | |
| "logits/chosen": -2.473135471343994, | |
| "logits/rejected": -2.308255195617676, | |
| "logps/chosen": -278.90167236328125, | |
| "logps/rejected": -262.17694091796875, | |
| "loss": 1.3217, | |
| "rewards/accuracies": 0.800000011920929, | |
| "rewards/chosen": 0.08379217237234116, | |
| "rewards/margins": 0.525702953338623, | |
| "rewards/rejected": -0.4419107437133789, | |
| "step": 1350 | |
| }, | |
| { | |
| "epoch": 0.8897612037945698, | |
| "grad_norm": 8.1875, | |
| "learning_rate": 1.8192150737583264e-07, | |
| "logits/chosen": -2.471273899078369, | |
| "logits/rejected": -2.355404853820801, | |
| "logps/chosen": -262.5338439941406, | |
| "logps/rejected": -240.73681640625, | |
| "loss": 1.3361, | |
| "rewards/accuracies": 0.675000011920929, | |
| "rewards/chosen": 0.20122098922729492, | |
| "rewards/margins": 0.45535793900489807, | |
| "rewards/rejected": -0.25413697957992554, | |
| "step": 1360 | |
| }, | |
| { | |
| "epoch": 0.8963035655871769, | |
| "grad_norm": 8.75, | |
| "learning_rate": 1.61137637857158e-07, | |
| "logits/chosen": -2.5680222511291504, | |
| "logits/rejected": -2.452162027359009, | |
| "logps/chosen": -331.45050048828125, | |
| "logps/rejected": -277.1683654785156, | |
| "loss": 1.3416, | |
| "rewards/accuracies": 0.7749999761581421, | |
| "rewards/chosen": 0.10738255083560944, | |
| "rewards/margins": 0.42853283882141113, | |
| "rewards/rejected": -0.3211502432823181, | |
| "step": 1370 | |
| }, | |
| { | |
| "epoch": 0.9028459273797841, | |
| "grad_norm": 6.6875, | |
| "learning_rate": 1.415746684157951e-07, | |
| "logits/chosen": -2.469827890396118, | |
| "logits/rejected": -2.433368444442749, | |
| "logps/chosen": -222.18994140625, | |
| "logps/rejected": -224.70907592773438, | |
| "loss": 1.3334, | |
| "rewards/accuracies": 0.75, | |
| "rewards/chosen": 0.13235194981098175, | |
| "rewards/margins": 0.4484756886959076, | |
| "rewards/rejected": -0.3161238133907318, | |
| "step": 1380 | |
| }, | |
| { | |
| "epoch": 0.9093882891723912, | |
| "grad_norm": 7.03125, | |
| "learning_rate": 1.232428110382586e-07, | |
| "logits/chosen": -2.569434642791748, | |
| "logits/rejected": -2.4922871589660645, | |
| "logps/chosen": -249.40328979492188, | |
| "logps/rejected": -248.40054321289062, | |
| "loss": 1.3386, | |
| "rewards/accuracies": 0.675000011920929, | |
| "rewards/chosen": -0.005796324461698532, | |
| "rewards/margins": 0.4979252815246582, | |
| "rewards/rejected": -0.5037215948104858, | |
| "step": 1390 | |
| }, | |
| { | |
| "epoch": 0.9159306509649984, | |
| "grad_norm": 6.9375, | |
| "learning_rate": 1.0615163506323856e-07, | |
| "logits/chosen": -2.4826645851135254, | |
| "logits/rejected": -2.3935964107513428, | |
| "logps/chosen": -256.60003662109375, | |
| "logps/rejected": -247.3723602294922, | |
| "loss": 1.3398, | |
| "rewards/accuracies": 0.6875, | |
| "rewards/chosen": 0.10640044510364532, | |
| "rewards/margins": 0.3504992723464966, | |
| "rewards/rejected": -0.24409881234169006, | |
| "step": 1400 | |
| }, | |
| { | |
| "epoch": 0.9159306509649984, | |
| "eval_logits/chosen": -2.3116648197174072, | |
| "eval_logits/rejected": -2.2316980361938477, | |
| "eval_logps/chosen": -270.7674560546875, | |
| "eval_logps/rejected": -257.3534851074219, | |
| "eval_loss": 1.3414233922958374, | |
| "eval_rewards/accuracies": 0.7475000023841858, | |
| "eval_rewards/chosen": 0.14090880751609802, | |
| "eval_rewards/margins": 0.46596458554267883, | |
| "eval_rewards/rejected": -0.3250557780265808, | |
| "eval_runtime": 192.5164, | |
| "eval_samples_per_second": 10.389, | |
| "eval_steps_per_second": 0.519, | |
| "step": 1400 | |
| }, | |
| { | |
| "epoch": 0.9224730127576055, | |
| "grad_norm": 6.5625, | |
| "learning_rate": 9.031006218634342e-08, | |
| "logits/chosen": -2.4582056999206543, | |
| "logits/rejected": -2.4208710193634033, | |
| "logps/chosen": -294.5101623535156, | |
| "logps/rejected": -322.93707275390625, | |
| "loss": 1.3372, | |
| "rewards/accuracies": 0.7124999761581421, | |
| "rewards/chosen": 0.17300447821617126, | |
| "rewards/margins": 0.528673529624939, | |
| "rewards/rejected": -0.3556690812110901, | |
| "step": 1410 | |
| }, | |
| { | |
| "epoch": 0.9290153745502127, | |
| "grad_norm": 7.46875, | |
| "learning_rate": 7.572636180292831e-08, | |
| "logits/chosen": -2.4200243949890137, | |
| "logits/rejected": -2.357060194015503, | |
| "logps/chosen": -254.03231811523438, | |
| "logps/rejected": -243.3819580078125, | |
| "loss": 1.3276, | |
| "rewards/accuracies": 0.7250000238418579, | |
| "rewards/chosen": 0.08049002289772034, | |
| "rewards/margins": 0.46086424589157104, | |
| "rewards/rejected": -0.3803742825984955, | |
| "step": 1420 | |
| }, | |
| { | |
| "epoch": 0.9355577363428198, | |
| "grad_norm": 7.4375, | |
| "learning_rate": 6.240814669141559e-08, | |
| "logits/chosen": -2.4541358947753906, | |
| "logits/rejected": -2.4223341941833496, | |
| "logps/chosen": -266.42156982421875, | |
| "logps/rejected": -304.84600830078125, | |
| "loss": 1.3296, | |
| "rewards/accuracies": 0.7250000238418579, | |
| "rewards/chosen": 0.07571353018283844, | |
| "rewards/margins": 0.49027562141418457, | |
| "rewards/rejected": -0.4145621359348297, | |
| "step": 1430 | |
| }, | |
| { | |
| "epoch": 0.9421000981354269, | |
| "grad_norm": 8.25, | |
| "learning_rate": 5.036236903938285e-08, | |
| "logits/chosen": -2.5123915672302246, | |
| "logits/rejected": -2.4232895374298096, | |
| "logps/chosen": -289.10601806640625, | |
| "logps/rejected": -232.24868774414062, | |
| "loss": 1.3312, | |
| "rewards/accuracies": 0.7250000238418579, | |
| "rewards/chosen": 0.08486177772283554, | |
| "rewards/margins": 0.33679136633872986, | |
| "rewards/rejected": -0.2519296109676361, | |
| "step": 1440 | |
| }, | |
| { | |
| "epoch": 0.948642459928034, | |
| "grad_norm": 6.90625, | |
| "learning_rate": 3.959531681447859e-08, | |
| "logits/chosen": -2.3544886112213135, | |
| "logits/rejected": -2.306253433227539, | |
| "logps/chosen": -259.6851501464844, | |
| "logps/rejected": -238.7523956298828, | |
| "loss": 1.3431, | |
| "rewards/accuracies": 0.699999988079071, | |
| "rewards/chosen": 0.04705673083662987, | |
| "rewards/margins": 0.3753374218940735, | |
| "rewards/rejected": -0.3282806873321533, | |
| "step": 1450 | |
| }, | |
| { | |
| "epoch": 0.9551848217206411, | |
| "grad_norm": 7.03125, | |
| "learning_rate": 3.0112610482064544e-08, | |
| "logits/chosen": -2.586221933364868, | |
| "logits/rejected": -2.4770185947418213, | |
| "logps/chosen": -299.3462829589844, | |
| "logps/rejected": -263.6593017578125, | |
| "loss": 1.342, | |
| "rewards/accuracies": 0.7124999761581421, | |
| "rewards/chosen": 0.11202242225408554, | |
| "rewards/margins": 0.4375079572200775, | |
| "rewards/rejected": -0.3254855275154114, | |
| "step": 1460 | |
| }, | |
| { | |
| "epoch": 0.9617271835132483, | |
| "grad_norm": 6.96875, | |
| "learning_rate": 2.1919200071301715e-08, | |
| "logits/chosen": -2.5899202823638916, | |
| "logits/rejected": -2.45597767829895, | |
| "logps/chosen": -292.31683349609375, | |
| "logps/rejected": -233.8472900390625, | |
| "loss": 1.3513, | |
| "rewards/accuracies": 0.7250000238418579, | |
| "rewards/chosen": 0.09856925159692764, | |
| "rewards/margins": 0.4128998816013336, | |
| "rewards/rejected": -0.31433066725730896, | |
| "step": 1470 | |
| }, | |
| { | |
| "epoch": 0.9682695453058554, | |
| "grad_norm": 6.78125, | |
| "learning_rate": 1.50193625912029e-08, | |
| "logits/chosen": -2.3694183826446533, | |
| "logits/rejected": -2.3877651691436768, | |
| "logps/chosen": -229.5559539794922, | |
| "logps/rejected": -265.0757141113281, | |
| "loss": 1.3336, | |
| "rewards/accuracies": 0.737500011920929, | |
| "rewards/chosen": -0.014900955371558666, | |
| "rewards/margins": 0.37212496995925903, | |
| "rewards/rejected": -0.38702592253685, | |
| "step": 1480 | |
| }, | |
| { | |
| "epoch": 0.9748119070984625, | |
| "grad_norm": 8.25, | |
| "learning_rate": 9.416699798010521e-09, | |
| "logits/chosen": -2.3672139644622803, | |
| "logits/rejected": -2.2716662883758545, | |
| "logps/chosen": -253.46322631835938, | |
| "logps/rejected": -208.71768188476562, | |
| "loss": 1.3194, | |
| "rewards/accuracies": 0.7875000238418579, | |
| "rewards/chosen": 0.11514721810817719, | |
| "rewards/margins": 0.5433107018470764, | |
| "rewards/rejected": -0.42816343903541565, | |
| "step": 1490 | |
| }, | |
| { | |
| "epoch": 0.9813542688910697, | |
| "grad_norm": 6.5, | |
| "learning_rate": 5.114136315058083e-09, | |
| "logits/chosen": -2.3912510871887207, | |
| "logits/rejected": -2.3454089164733887, | |
| "logps/chosen": -262.8589782714844, | |
| "logps/rejected": -278.1678161621094, | |
| "loss": 1.325, | |
| "rewards/accuracies": 0.800000011920929, | |
| "rewards/chosen": 0.12093720585107803, | |
| "rewards/margins": 0.5334186553955078, | |
| "rewards/rejected": -0.41248148679733276, | |
| "step": 1500 | |
| }, | |
| { | |
| "epoch": 0.9813542688910697, | |
| "eval_logits/chosen": -2.313671827316284, | |
| "eval_logits/rejected": -2.23390793800354, | |
| "eval_logps/chosen": -270.74359130859375, | |
| "eval_logps/rejected": -257.3707275390625, | |
| "eval_loss": 1.3408766984939575, | |
| "eval_rewards/accuracies": 0.7475000023841858, | |
| "eval_rewards/chosen": 0.14329300820827484, | |
| "eval_rewards/margins": 0.47007519006729126, | |
| "eval_rewards/rejected": -0.3267821967601776, | |
| "eval_runtime": 192.5128, | |
| "eval_samples_per_second": 10.389, | |
| "eval_steps_per_second": 0.519, | |
| "step": 1500 | |
| }, | |
| { | |
| "epoch": 0.9878966306836768, | |
| "grad_norm": 7.09375, | |
| "learning_rate": 2.113918106098345e-09, | |
| "logits/chosen": -2.456124782562256, | |
| "logits/rejected": -2.4603703022003174, | |
| "logps/chosen": -279.5404052734375, | |
| "logps/rejected": -286.7142333984375, | |
| "loss": 1.3363, | |
| "rewards/accuracies": 0.675000011920929, | |
| "rewards/chosen": 0.1024564728140831, | |
| "rewards/margins": 0.3759761452674866, | |
| "rewards/rejected": -0.2735196650028229, | |
| "step": 1510 | |
| }, | |
| { | |
| "epoch": 0.994438992476284, | |
| "grad_norm": 8.6875, | |
| "learning_rate": 4.176113028983575e-10, | |
| "logits/chosen": -2.407866954803467, | |
| "logits/rejected": -2.355156421661377, | |
| "logps/chosen": -266.18231201171875, | |
| "logps/rejected": -265.04022216796875, | |
| "loss": 1.337, | |
| "rewards/accuracies": 0.675000011920929, | |
| "rewards/chosen": 0.07989536225795746, | |
| "rewards/margins": 0.4557567238807678, | |
| "rewards/rejected": -0.37586134672164917, | |
| "step": 1520 | |
| }, | |
| { | |
| "epoch": 0.9996728819103696, | |
| "step": 1528, | |
| "total_flos": 0.0, | |
| "train_loss": 1.343712306771603, | |
| "train_runtime": 13958.9948, | |
| "train_samples_per_second": 4.38, | |
| "train_steps_per_second": 0.109 | |
| } | |
| ], | |
| "logging_steps": 10, | |
| "max_steps": 1528, | |
| "num_input_tokens_seen": 0, | |
| "num_train_epochs": 1, | |
| "save_steps": 100, | |
| "stateful_callbacks": { | |
| "TrainerControl": { | |
| "args": { | |
| "should_epoch_stop": false, | |
| "should_evaluate": false, | |
| "should_log": false, | |
| "should_save": true, | |
| "should_training_stop": true | |
| }, | |
| "attributes": {} | |
| } | |
| }, | |
| "total_flos": 0.0, | |
| "train_batch_size": 2, | |
| "trial_name": null, | |
| "trial_params": null | |
| } | |