| { | |
| "best_metric": null, | |
| "best_model_checkpoint": null, | |
| "epoch": 0.998766954377312, | |
| "eval_steps": 100, | |
| "global_step": 405, | |
| "is_hyper_param_search": false, | |
| "is_local_process_zero": true, | |
| "is_world_process_zero": true, | |
| "log_history": [ | |
| { | |
| "epoch": 0.0, | |
| "learning_rate": 1.2195121951219512e-08, | |
| "logits/chosen": -2.8695335388183594, | |
| "logits/rejected": -2.8522377014160156, | |
| "logps/chosen": -537.80126953125, | |
| "logps/rejected": -108.91968536376953, | |
| "loss": 0.3287, | |
| "rewards/accuracies": 0.0, | |
| "rewards/chosen": 0.0, | |
| "rewards/margins": 0.0, | |
| "rewards/rejected": 0.0, | |
| "step": 1 | |
| }, | |
| { | |
| "epoch": 0.02, | |
| "learning_rate": 1.219512195121951e-07, | |
| "logits/chosen": -2.8006999492645264, | |
| "logits/rejected": -2.7513413429260254, | |
| "logps/chosen": -339.1315612792969, | |
| "logps/rejected": -113.41014862060547, | |
| "loss": 0.3429, | |
| "rewards/accuracies": 0.5625, | |
| "rewards/chosen": 0.0010660986881703138, | |
| "rewards/margins": 0.0017727299127727747, | |
| "rewards/rejected": -0.0007066310499794781, | |
| "step": 10 | |
| }, | |
| { | |
| "epoch": 0.05, | |
| "learning_rate": 2.439024390243902e-07, | |
| "logits/chosen": -2.8162312507629395, | |
| "logits/rejected": -2.8079066276550293, | |
| "logps/chosen": -435.261962890625, | |
| "logps/rejected": -116.0378189086914, | |
| "loss": 0.3332, | |
| "rewards/accuracies": 0.7875000238418579, | |
| "rewards/chosen": 0.021241962909698486, | |
| "rewards/margins": 0.038096584379673004, | |
| "rewards/rejected": -0.016854625195264816, | |
| "step": 20 | |
| }, | |
| { | |
| "epoch": 0.07, | |
| "learning_rate": 3.6585365853658536e-07, | |
| "logits/chosen": -2.7254586219787598, | |
| "logits/rejected": -2.688169002532959, | |
| "logps/chosen": -437.81072998046875, | |
| "logps/rejected": -141.1324920654297, | |
| "loss": 0.309, | |
| "rewards/accuracies": 0.8500000238418579, | |
| "rewards/chosen": 0.08655615150928497, | |
| "rewards/margins": 0.20261511206626892, | |
| "rewards/rejected": -0.11605894565582275, | |
| "step": 30 | |
| }, | |
| { | |
| "epoch": 0.1, | |
| "learning_rate": 4.878048780487804e-07, | |
| "logits/chosen": -2.5956244468688965, | |
| "logits/rejected": -2.5760390758514404, | |
| "logps/chosen": -414.177490234375, | |
| "logps/rejected": -168.2574005126953, | |
| "loss": 0.2577, | |
| "rewards/accuracies": 0.7749999761581421, | |
| "rewards/chosen": 0.02355712652206421, | |
| "rewards/margins": 0.46499890089035034, | |
| "rewards/rejected": -0.4414418339729309, | |
| "step": 40 | |
| }, | |
| { | |
| "epoch": 0.12, | |
| "learning_rate": 4.992461696250783e-07, | |
| "logits/chosen": -2.474365711212158, | |
| "logits/rejected": -2.461667537689209, | |
| "logps/chosen": -420.9219665527344, | |
| "logps/rejected": -210.11111450195312, | |
| "loss": 0.1873, | |
| "rewards/accuracies": 0.7250000238418579, | |
| "rewards/chosen": -0.11808023601770401, | |
| "rewards/margins": 0.785495400428772, | |
| "rewards/rejected": -0.903575599193573, | |
| "step": 50 | |
| }, | |
| { | |
| "epoch": 0.15, | |
| "learning_rate": 4.966461721767899e-07, | |
| "logits/chosen": -2.4058423042297363, | |
| "logits/rejected": -2.3743529319763184, | |
| "logps/chosen": -377.65484619140625, | |
| "logps/rejected": -207.171142578125, | |
| "loss": 0.1202, | |
| "rewards/accuracies": 0.762499988079071, | |
| "rewards/chosen": -0.32305708527565, | |
| "rewards/margins": 0.8489507436752319, | |
| "rewards/rejected": -1.1720077991485596, | |
| "step": 60 | |
| }, | |
| { | |
| "epoch": 0.17, | |
| "learning_rate": 4.922100518015975e-07, | |
| "logits/chosen": -2.4014129638671875, | |
| "logits/rejected": -2.365219831466675, | |
| "logps/chosen": -448.73516845703125, | |
| "logps/rejected": -269.4298400878906, | |
| "loss": 0.0894, | |
| "rewards/accuracies": 0.75, | |
| "rewards/chosen": -0.4889157712459564, | |
| "rewards/margins": 1.191009759902954, | |
| "rewards/rejected": -1.679925560951233, | |
| "step": 70 | |
| }, | |
| { | |
| "epoch": 0.2, | |
| "learning_rate": 4.859708325770919e-07, | |
| "logits/chosen": -2.3665881156921387, | |
| "logits/rejected": -2.313610792160034, | |
| "logps/chosen": -442.3289489746094, | |
| "logps/rejected": -284.679931640625, | |
| "loss": 0.0863, | |
| "rewards/accuracies": 0.8374999761581421, | |
| "rewards/chosen": -0.31383419036865234, | |
| "rewards/margins": 1.4407079219818115, | |
| "rewards/rejected": -1.7545421123504639, | |
| "step": 80 | |
| }, | |
| { | |
| "epoch": 0.22, | |
| "learning_rate": 4.779749614980225e-07, | |
| "logits/chosen": -2.3862123489379883, | |
| "logits/rejected": -2.336027145385742, | |
| "logps/chosen": -456.930419921875, | |
| "logps/rejected": -332.8990478515625, | |
| "loss": 0.0628, | |
| "rewards/accuracies": 0.8062499761581421, | |
| "rewards/chosen": -0.7161350846290588, | |
| "rewards/margins": 1.5251940488815308, | |
| "rewards/rejected": -2.2413289546966553, | |
| "step": 90 | |
| }, | |
| { | |
| "epoch": 0.25, | |
| "learning_rate": 4.682819627081427e-07, | |
| "logits/chosen": -2.3527557849884033, | |
| "logits/rejected": -2.2716236114501953, | |
| "logps/chosen": -491.0079650878906, | |
| "logps/rejected": -360.1409912109375, | |
| "loss": 0.067, | |
| "rewards/accuracies": 0.7437499761581421, | |
| "rewards/chosen": -0.676110029220581, | |
| "rewards/margins": 1.8081252574920654, | |
| "rewards/rejected": -2.4842352867126465, | |
| "step": 100 | |
| }, | |
| { | |
| "epoch": 0.25, | |
| "eval_logits/chosen": -2.340771198272705, | |
| "eval_logits/rejected": -2.3243579864501953, | |
| "eval_logps/chosen": -405.76531982421875, | |
| "eval_logps/rejected": -430.49688720703125, | |
| "eval_loss": 0.11736096441745758, | |
| "eval_rewards/accuracies": 0.61328125, | |
| "eval_rewards/chosen": -1.4872568845748901, | |
| "eval_rewards/margins": 0.2441793829202652, | |
| "eval_rewards/rejected": -1.7314363718032837, | |
| "eval_runtime": 53.3203, | |
| "eval_samples_per_second": 37.509, | |
| "eval_steps_per_second": 0.6, | |
| "step": 100 | |
| }, | |
| { | |
| "epoch": 0.27, | |
| "learning_rate": 4.569639943810477e-07, | |
| "logits/chosen": -2.361056089401245, | |
| "logits/rejected": -2.282515287399292, | |
| "logps/chosen": -496.8377380371094, | |
| "logps/rejected": -344.2028503417969, | |
| "loss": 0.0656, | |
| "rewards/accuracies": 0.8062499761581421, | |
| "rewards/chosen": -0.5422588586807251, | |
| "rewards/margins": 1.8025261163711548, | |
| "rewards/rejected": -2.344785213470459, | |
| "step": 110 | |
| }, | |
| { | |
| "epoch": 0.3, | |
| "learning_rate": 4.4410531154874543e-07, | |
| "logits/chosen": -2.31123685836792, | |
| "logits/rejected": -2.2176012992858887, | |
| "logps/chosen": -541.8471069335938, | |
| "logps/rejected": -430.7569885253906, | |
| "loss": 0.0355, | |
| "rewards/accuracies": 0.8125, | |
| "rewards/chosen": -1.0784400701522827, | |
| "rewards/margins": 2.053567409515381, | |
| "rewards/rejected": -3.132007360458374, | |
| "step": 120 | |
| }, | |
| { | |
| "epoch": 0.32, | |
| "learning_rate": 4.298016388768561e-07, | |
| "logits/chosen": -2.288428783416748, | |
| "logits/rejected": -2.175412178039551, | |
| "logps/chosen": -530.9747314453125, | |
| "logps/rejected": -429.48785400390625, | |
| "loss": 0.0441, | |
| "rewards/accuracies": 0.8125, | |
| "rewards/chosen": -1.0411689281463623, | |
| "rewards/margins": 1.9912636280059814, | |
| "rewards/rejected": -3.0324320793151855, | |
| "step": 130 | |
| }, | |
| { | |
| "epoch": 0.35, | |
| "learning_rate": 4.1415945805573005e-07, | |
| "logits/chosen": -2.3122925758361816, | |
| "logits/rejected": -2.227431058883667, | |
| "logps/chosen": -509.9908142089844, | |
| "logps/rejected": -396.0080871582031, | |
| "loss": 0.0515, | |
| "rewards/accuracies": 0.856249988079071, | |
| "rewards/chosen": -0.6500633358955383, | |
| "rewards/margins": 2.154536724090576, | |
| "rewards/rejected": -2.804600238800049, | |
| "step": 140 | |
| }, | |
| { | |
| "epoch": 0.37, | |
| "learning_rate": 3.972952151123984e-07, | |
| "logits/chosen": -2.2672953605651855, | |
| "logits/rejected": -2.170297384262085, | |
| "logps/chosen": -522.5911254882812, | |
| "logps/rejected": -445.1206970214844, | |
| "loss": 0.0433, | |
| "rewards/accuracies": 0.84375, | |
| "rewards/chosen": -0.9818037748336792, | |
| "rewards/margins": 2.1947696208953857, | |
| "rewards/rejected": -3.1765732765197754, | |
| "step": 150 | |
| }, | |
| { | |
| "epoch": 0.39, | |
| "learning_rate": 3.793344535444142e-07, | |
| "logits/chosen": -2.3276212215423584, | |
| "logits/rejected": -2.2313549518585205, | |
| "logps/chosen": -537.3373413085938, | |
| "logps/rejected": -405.67559814453125, | |
| "loss": 0.0465, | |
| "rewards/accuracies": 0.8687499761581421, | |
| "rewards/chosen": -0.7096673250198364, | |
| "rewards/margins": 2.1882476806640625, | |
| "rewards/rejected": -2.8979151248931885, | |
| "step": 160 | |
| }, | |
| { | |
| "epoch": 0.42, | |
| "learning_rate": 3.604108797288461e-07, | |
| "logits/chosen": -2.2612414360046387, | |
| "logits/rejected": -2.178300142288208, | |
| "logps/chosen": -512.8599853515625, | |
| "logps/rejected": -435.66693115234375, | |
| "loss": 0.0375, | |
| "rewards/accuracies": 0.831250011920929, | |
| "rewards/chosen": -1.0879271030426025, | |
| "rewards/margins": 2.0758705139160156, | |
| "rewards/rejected": -3.163797616958618, | |
| "step": 170 | |
| }, | |
| { | |
| "epoch": 0.44, | |
| "learning_rate": 3.40665367563858e-07, | |
| "logits/chosen": -2.242908477783203, | |
| "logits/rejected": -2.1256449222564697, | |
| "logps/chosen": -503.54547119140625, | |
| "logps/rejected": -403.26593017578125, | |
| "loss": 0.037, | |
| "rewards/accuracies": 0.762499988079071, | |
| "rewards/chosen": -1.299680471420288, | |
| "rewards/margins": 1.7938661575317383, | |
| "rewards/rejected": -3.0935468673706055, | |
| "step": 180 | |
| }, | |
| { | |
| "epoch": 0.47, | |
| "learning_rate": 3.202449097526798e-07, | |
| "logits/chosen": -2.2713847160339355, | |
| "logits/rejected": -2.163483142852783, | |
| "logps/chosen": -501.21185302734375, | |
| "logps/rejected": -417.178955078125, | |
| "loss": 0.0478, | |
| "rewards/accuracies": 0.800000011920929, | |
| "rewards/chosen": -1.1433241367340088, | |
| "rewards/margins": 1.9010562896728516, | |
| "rewards/rejected": -3.0443806648254395, | |
| "step": 190 | |
| }, | |
| { | |
| "epoch": 0.49, | |
| "learning_rate": 2.993015235369905e-07, | |
| "logits/chosen": -2.2553932666778564, | |
| "logits/rejected": -2.1447877883911133, | |
| "logps/chosen": -530.4608154296875, | |
| "logps/rejected": -431.3677673339844, | |
| "loss": 0.0435, | |
| "rewards/accuracies": 0.862500011920929, | |
| "rewards/chosen": -0.9609147906303406, | |
| "rewards/margins": 2.1664257049560547, | |
| "rewards/rejected": -3.12734055519104, | |
| "step": 200 | |
| }, | |
| { | |
| "epoch": 0.49, | |
| "eval_logits/chosen": -2.1734278202056885, | |
| "eval_logits/rejected": -2.142083168029785, | |
| "eval_logps/chosen": -475.05853271484375, | |
| "eval_logps/rejected": -512.2731323242188, | |
| "eval_loss": 0.07987947016954422, | |
| "eval_rewards/accuracies": 0.62109375, | |
| "eval_rewards/chosen": -2.1801888942718506, | |
| "eval_rewards/margins": 0.36900976300239563, | |
| "eval_rewards/rejected": -2.549198627471924, | |
| "eval_runtime": 53.3455, | |
| "eval_samples_per_second": 37.491, | |
| "eval_steps_per_second": 0.6, | |
| "step": 200 | |
| }, | |
| { | |
| "epoch": 0.52, | |
| "learning_rate": 2.7799111902582693e-07, | |
| "logits/chosen": -2.1741156578063965, | |
| "logits/rejected": -2.0635221004486084, | |
| "logps/chosen": -524.964111328125, | |
| "logps/rejected": -450.088623046875, | |
| "loss": 0.0342, | |
| "rewards/accuracies": 0.793749988079071, | |
| "rewards/chosen": -1.4079291820526123, | |
| "rewards/margins": 1.9792110919952393, | |
| "rewards/rejected": -3.3871402740478516, | |
| "step": 210 | |
| }, | |
| { | |
| "epoch": 0.54, | |
| "learning_rate": 2.564723385445869e-07, | |
| "logits/chosen": -2.1615562438964844, | |
| "logits/rejected": -2.0492231845855713, | |
| "logps/chosen": -561.4410400390625, | |
| "logps/rejected": -483.4110412597656, | |
| "loss": 0.0356, | |
| "rewards/accuracies": 0.824999988079071, | |
| "rewards/chosen": -1.2221920490264893, | |
| "rewards/margins": 2.281812906265259, | |
| "rewards/rejected": -3.504004955291748, | |
| "step": 220 | |
| }, | |
| { | |
| "epoch": 0.57, | |
| "learning_rate": 2.3490537564442845e-07, | |
| "logits/chosen": -2.1855053901672363, | |
| "logits/rejected": -2.075396776199341, | |
| "logps/chosen": -530.4036254882812, | |
| "logps/rejected": -457.735107421875, | |
| "loss": 0.0301, | |
| "rewards/accuracies": 0.8125, | |
| "rewards/chosen": -1.2399654388427734, | |
| "rewards/margins": 2.154582977294922, | |
| "rewards/rejected": -3.3945488929748535, | |
| "step": 230 | |
| }, | |
| { | |
| "epoch": 0.59, | |
| "learning_rate": 2.1345078256378801e-07, | |
| "logits/chosen": -2.187288284301758, | |
| "logits/rejected": -2.0610218048095703, | |
| "logps/chosen": -533.1801147460938, | |
| "logps/rejected": -487.4967346191406, | |
| "loss": 0.0244, | |
| "rewards/accuracies": 0.8374999761581421, | |
| "rewards/chosen": -1.2845171689987183, | |
| "rewards/margins": 2.3898768424987793, | |
| "rewards/rejected": -3.674394130706787, | |
| "step": 240 | |
| }, | |
| { | |
| "epoch": 0.62, | |
| "learning_rate": 1.9226827501969865e-07, | |
| "logits/chosen": -2.166536331176758, | |
| "logits/rejected": -2.0207314491271973, | |
| "logps/chosen": -566.1360473632812, | |
| "logps/rejected": -520.3924560546875, | |
| "loss": 0.0354, | |
| "rewards/accuracies": 0.862500011920929, | |
| "rewards/chosen": -1.3790395259857178, | |
| "rewards/margins": 2.647700786590576, | |
| "rewards/rejected": -4.026740550994873, | |
| "step": 250 | |
| }, | |
| { | |
| "epoch": 0.64, | |
| "learning_rate": 1.715155432264775e-07, | |
| "logits/chosen": -2.160069704055786, | |
| "logits/rejected": -2.0387001037597656, | |
| "logps/chosen": -520.0031127929688, | |
| "logps/rejected": -447.298828125, | |
| "loss": 0.0421, | |
| "rewards/accuracies": 0.84375, | |
| "rewards/chosen": -1.3448396921157837, | |
| "rewards/margins": 2.034497022628784, | |
| "rewards/rejected": -3.3793368339538574, | |
| "step": 260 | |
| }, | |
| { | |
| "epoch": 0.67, | |
| "learning_rate": 1.51347077992983e-07, | |
| "logits/chosen": -2.201711416244507, | |
| "logits/rejected": -2.0760598182678223, | |
| "logps/chosen": -518.4276733398438, | |
| "logps/rejected": -448.0538024902344, | |
| "loss": 0.0352, | |
| "rewards/accuracies": 0.84375, | |
| "rewards/chosen": -1.2280199527740479, | |
| "rewards/margins": 2.027623176574707, | |
| "rewards/rejected": -3.255643129348755, | |
| "step": 270 | |
| }, | |
| { | |
| "epoch": 0.69, | |
| "learning_rate": 1.3191302063739906e-07, | |
| "logits/chosen": -2.2061142921447754, | |
| "logits/rejected": -2.08331298828125, | |
| "logps/chosen": -536.6764526367188, | |
| "logps/rejected": -460.00250244140625, | |
| "loss": 0.0335, | |
| "rewards/accuracies": 0.8187500238418579, | |
| "rewards/chosen": -1.3899211883544922, | |
| "rewards/margins": 2.058790445327759, | |
| "rewards/rejected": -3.44871187210083, | |
| "step": 280 | |
| }, | |
| { | |
| "epoch": 0.72, | |
| "learning_rate": 1.1335804528119475e-07, | |
| "logits/chosen": -2.194516658782959, | |
| "logits/rejected": -2.0347750186920166, | |
| "logps/chosen": -583.8428955078125, | |
| "logps/rejected": -485.03057861328125, | |
| "loss": 0.0266, | |
| "rewards/accuracies": 0.8500000238418579, | |
| "rewards/chosen": -1.3776942491531372, | |
| "rewards/margins": 2.384459972381592, | |
| "rewards/rejected": -3.7621541023254395, | |
| "step": 290 | |
| }, | |
| { | |
| "epoch": 0.74, | |
| "learning_rate": 9.582028184286423e-08, | |
| "logits/chosen": -2.1787655353546143, | |
| "logits/rejected": -2.075291156768799, | |
| "logps/chosen": -514.74072265625, | |
| "logps/rejected": -461.1167907714844, | |
| "loss": 0.0288, | |
| "rewards/accuracies": 0.768750011920929, | |
| "rewards/chosen": -1.5570707321166992, | |
| "rewards/margins": 1.984126091003418, | |
| "rewards/rejected": -3.541196823120117, | |
| "step": 300 | |
| }, | |
| { | |
| "epoch": 0.74, | |
| "eval_logits/chosen": -2.1675052642822266, | |
| "eval_logits/rejected": -2.133922576904297, | |
| "eval_logps/chosen": -500.8697204589844, | |
| "eval_logps/rejected": -548.4016723632812, | |
| "eval_loss": 0.07103094458580017, | |
| "eval_rewards/accuracies": 0.6171875, | |
| "eval_rewards/chosen": -2.438300848007202, | |
| "eval_rewards/margins": 0.4721827805042267, | |
| "eval_rewards/rejected": -2.9104835987091064, | |
| "eval_runtime": 53.3323, | |
| "eval_samples_per_second": 37.501, | |
| "eval_steps_per_second": 0.6, | |
| "step": 300 | |
| }, | |
| { | |
| "epoch": 0.76, | |
| "learning_rate": 7.943028774907065e-08, | |
| "logits/chosen": -2.1723272800445557, | |
| "logits/rejected": -2.0614748001098633, | |
| "logps/chosen": -535.1619873046875, | |
| "logps/rejected": -455.0377502441406, | |
| "loss": 0.0292, | |
| "rewards/accuracies": 0.793749988079071, | |
| "rewards/chosen": -1.294924020767212, | |
| "rewards/margins": 2.182978630065918, | |
| "rewards/rejected": -3.47790265083313, | |
| "step": 310 | |
| }, | |
| { | |
| "epoch": 0.79, | |
| "learning_rate": 6.431007601814637e-08, | |
| "logits/chosen": -2.2771997451782227, | |
| "logits/rejected": -2.146629810333252, | |
| "logps/chosen": -576.0748291015625, | |
| "logps/rejected": -492.65167236328125, | |
| "loss": 0.031, | |
| "rewards/accuracies": 0.78125, | |
| "rewards/chosen": -1.481297254562378, | |
| "rewards/margins": 2.240931510925293, | |
| "rewards/rejected": -3.722228527069092, | |
| "step": 320 | |
| }, | |
| { | |
| "epoch": 0.81, | |
| "learning_rate": 5.0572206951246e-08, | |
| "logits/chosen": -2.2158615589141846, | |
| "logits/rejected": -2.0619277954101562, | |
| "logps/chosen": -581.7293701171875, | |
| "logps/rejected": -495.04638671875, | |
| "loss": 0.0317, | |
| "rewards/accuracies": 0.8812500238418579, | |
| "rewards/chosen": -1.2919968366622925, | |
| "rewards/margins": 2.5048012733459473, | |
| "rewards/rejected": -3.7967982292175293, | |
| "step": 330 | |
| }, | |
| { | |
| "epoch": 0.84, | |
| "learning_rate": 3.831895019292897e-08, | |
| "logits/chosen": -2.2480924129486084, | |
| "logits/rejected": -2.110482692718506, | |
| "logps/chosen": -597.642333984375, | |
| "logps/rejected": -521.8727416992188, | |
| "loss": 0.0281, | |
| "rewards/accuracies": 0.8500000238418579, | |
| "rewards/chosen": -1.3266541957855225, | |
| "rewards/margins": 2.570845127105713, | |
| "rewards/rejected": -3.8974990844726562, | |
| "step": 340 | |
| }, | |
| { | |
| "epoch": 0.86, | |
| "learning_rate": 2.764152339909756e-08, | |
| "logits/chosen": -2.1906208992004395, | |
| "logits/rejected": -2.0595510005950928, | |
| "logps/chosen": -539.5335083007812, | |
| "logps/rejected": -462.76025390625, | |
| "loss": 0.034, | |
| "rewards/accuracies": 0.7875000238418579, | |
| "rewards/chosen": -1.460292935371399, | |
| "rewards/margins": 2.1251232624053955, | |
| "rewards/rejected": -3.585416078567505, | |
| "step": 350 | |
| }, | |
| { | |
| "epoch": 0.89, | |
| "learning_rate": 1.861941317991664e-08, | |
| "logits/chosen": -2.1756224632263184, | |
| "logits/rejected": -2.0511136054992676, | |
| "logps/chosen": -544.6201171875, | |
| "logps/rejected": -468.48760986328125, | |
| "loss": 0.0303, | |
| "rewards/accuracies": 0.762499988079071, | |
| "rewards/chosen": -1.5998585224151611, | |
| "rewards/margins": 1.8565304279327393, | |
| "rewards/rejected": -3.4563891887664795, | |
| "step": 360 | |
| }, | |
| { | |
| "epoch": 0.91, | |
| "learning_rate": 1.13197833728636e-08, | |
| "logits/chosen": -2.1858394145965576, | |
| "logits/rejected": -2.0600364208221436, | |
| "logps/chosen": -554.5750732421875, | |
| "logps/rejected": -469.27484130859375, | |
| "loss": 0.0296, | |
| "rewards/accuracies": 0.793749988079071, | |
| "rewards/chosen": -1.3729755878448486, | |
| "rewards/margins": 2.1774544715881348, | |
| "rewards/rejected": -3.5504302978515625, | |
| "step": 370 | |
| }, | |
| { | |
| "epoch": 0.94, | |
| "learning_rate": 5.79697505093521e-09, | |
| "logits/chosen": -2.161980628967285, | |
| "logits/rejected": -2.0485546588897705, | |
| "logps/chosen": -514.3355712890625, | |
| "logps/rejected": -432.40704345703125, | |
| "loss": 0.0282, | |
| "rewards/accuracies": 0.8062499761581421, | |
| "rewards/chosen": -1.297263741493225, | |
| "rewards/margins": 1.9683955907821655, | |
| "rewards/rejected": -3.2656593322753906, | |
| "step": 380 | |
| }, | |
| { | |
| "epoch": 0.96, | |
| "learning_rate": 2.092101988131256e-09, | |
| "logits/chosen": -2.2410452365875244, | |
| "logits/rejected": -2.109318256378174, | |
| "logps/chosen": -566.7930297851562, | |
| "logps/rejected": -473.2035217285156, | |
| "loss": 0.0274, | |
| "rewards/accuracies": 0.762499988079071, | |
| "rewards/chosen": -1.3028382062911987, | |
| "rewards/margins": 2.2111973762512207, | |
| "rewards/rejected": -3.514035701751709, | |
| "step": 390 | |
| }, | |
| { | |
| "epoch": 0.99, | |
| "learning_rate": 2.327445937151673e-10, | |
| "logits/chosen": -2.168402910232544, | |
| "logits/rejected": -2.057384967803955, | |
| "logps/chosen": -555.5765380859375, | |
| "logps/rejected": -468.2821350097656, | |
| "loss": 0.032, | |
| "rewards/accuracies": 0.831250011920929, | |
| "rewards/chosen": -1.3704006671905518, | |
| "rewards/margins": 2.0777838230133057, | |
| "rewards/rejected": -3.4481842517852783, | |
| "step": 400 | |
| }, | |
| { | |
| "epoch": 0.99, | |
| "eval_logits/chosen": -2.163764476776123, | |
| "eval_logits/rejected": -2.129502534866333, | |
| "eval_logps/chosen": -504.9170227050781, | |
| "eval_logps/rejected": -552.4011840820312, | |
| "eval_loss": 0.06752217561006546, | |
| "eval_rewards/accuracies": 0.640625, | |
| "eval_rewards/chosen": -2.478773355484009, | |
| "eval_rewards/margins": 0.47170597314834595, | |
| "eval_rewards/rejected": -2.950479507446289, | |
| "eval_runtime": 53.3856, | |
| "eval_samples_per_second": 37.463, | |
| "eval_steps_per_second": 0.599, | |
| "step": 400 | |
| }, | |
| { | |
| "epoch": 1.0, | |
| "step": 405, | |
| "total_flos": 0.0, | |
| "train_loss": 0.07294640629379838, | |
| "train_runtime": 3765.6331, | |
| "train_samples_per_second": 13.781, | |
| "train_steps_per_second": 0.108 | |
| } | |
| ], | |
| "logging_steps": 10, | |
| "max_steps": 405, | |
| "num_train_epochs": 1, | |
| "save_steps": 100, | |
| "total_flos": 0.0, | |
| "trial_name": null, | |
| "trial_params": null | |
| } | |