| { | |
| "best_metric": null, | |
| "best_model_checkpoint": null, | |
| "epoch": 9.674418604651162, | |
| "eval_steps": 500, | |
| "global_step": 260, | |
| "is_hyper_param_search": false, | |
| "is_local_process_zero": true, | |
| "is_world_process_zero": true, | |
| "log_history": [ | |
| { | |
| "epoch": 0.37209302325581395, | |
| "grad_norm": 0.9763292670249939, | |
| "learning_rate": 1.9230769230769234e-06, | |
| "logits/chosen": 1.7089074850082397, | |
| "logits/rejected": 1.7024719715118408, | |
| "logps/chosen": -80.31088256835938, | |
| "logps/rejected": -91.45267486572266, | |
| "loss": 0.692, | |
| "rewards/accuracies": 0.38749998807907104, | |
| "rewards/chosen": -0.0024384786374866962, | |
| "rewards/margins": -0.0009521525353193283, | |
| "rewards/rejected": -0.001486325403675437, | |
| "step": 10 | |
| }, | |
| { | |
| "epoch": 0.7441860465116279, | |
| "grad_norm": 0.5258393883705139, | |
| "learning_rate": 3.846153846153847e-06, | |
| "logits/chosen": 1.8638836145401, | |
| "logits/rejected": 1.8883317708969116, | |
| "logps/chosen": -81.79865264892578, | |
| "logps/rejected": -76.99979400634766, | |
| "loss": 0.6928, | |
| "rewards/accuracies": 0.5375000238418579, | |
| "rewards/chosen": 0.00844244472682476, | |
| "rewards/margins": 0.0048658037558197975, | |
| "rewards/rejected": 0.0035766414366662502, | |
| "step": 20 | |
| }, | |
| { | |
| "epoch": 1.1162790697674418, | |
| "grad_norm": 0.5254527926445007, | |
| "learning_rate": 4.996395926410354e-06, | |
| "logits/chosen": 1.7035324573516846, | |
| "logits/rejected": 1.731885552406311, | |
| "logps/chosen": -80.90572357177734, | |
| "logps/rejected": -82.523193359375, | |
| "loss": 0.6945, | |
| "rewards/accuracies": 0.48750001192092896, | |
| "rewards/chosen": -0.003143525216728449, | |
| "rewards/margins": -0.0011914735659956932, | |
| "rewards/rejected": -0.0019520517671480775, | |
| "step": 30 | |
| }, | |
| { | |
| "epoch": 1.4883720930232558, | |
| "grad_norm": 0.5585340261459351, | |
| "learning_rate": 4.955969343539162e-06, | |
| "logits/chosen": 1.7126038074493408, | |
| "logits/rejected": 1.7729876041412354, | |
| "logps/chosen": -82.4004898071289, | |
| "logps/rejected": -86.97776794433594, | |
| "loss": 0.6942, | |
| "rewards/accuracies": 0.36250001192092896, | |
| "rewards/chosen": -0.005664472468197346, | |
| "rewards/margins": -0.014465728774666786, | |
| "rewards/rejected": 0.00880125630646944, | |
| "step": 40 | |
| }, | |
| { | |
| "epoch": 1.8604651162790697, | |
| "grad_norm": 0.46961840987205505, | |
| "learning_rate": 4.8713411048678635e-06, | |
| "logits/chosen": 1.7948715686798096, | |
| "logits/rejected": 1.836954116821289, | |
| "logps/chosen": -87.7071304321289, | |
| "logps/rejected": -75.44202423095703, | |
| "loss": 0.693, | |
| "rewards/accuracies": 0.4625000059604645, | |
| "rewards/chosen": 0.0024648040998727083, | |
| "rewards/margins": -0.004807753954082727, | |
| "rewards/rejected": 0.007272557821124792, | |
| "step": 50 | |
| }, | |
| { | |
| "epoch": 2.2325581395348837, | |
| "grad_norm": 0.48951366543769836, | |
| "learning_rate": 4.744034319097536e-06, | |
| "logits/chosen": 1.8686463832855225, | |
| "logits/rejected": 1.831539511680603, | |
| "logps/chosen": -97.39242553710938, | |
| "logps/rejected": -83.82582092285156, | |
| "loss": 0.6931, | |
| "rewards/accuracies": 0.512499988079071, | |
| "rewards/chosen": 0.006232857704162598, | |
| "rewards/margins": 0.006834377534687519, | |
| "rewards/rejected": -0.0006015209364704788, | |
| "step": 60 | |
| }, | |
| { | |
| "epoch": 2.604651162790698, | |
| "grad_norm": 0.5472654700279236, | |
| "learning_rate": 4.5763402081200295e-06, | |
| "logits/chosen": 1.7030988931655884, | |
| "logits/rejected": 1.7544777393341064, | |
| "logps/chosen": -76.77324676513672, | |
| "logps/rejected": -74.05345153808594, | |
| "loss": 0.6906, | |
| "rewards/accuracies": 0.512499988079071, | |
| "rewards/chosen": -0.005271740257740021, | |
| "rewards/margins": -0.0019098047632724047, | |
| "rewards/rejected": -0.0033619359601289034, | |
| "step": 70 | |
| }, | |
| { | |
| "epoch": 2.9767441860465116, | |
| "grad_norm": 0.5428044199943542, | |
| "learning_rate": 4.3712768704277535e-06, | |
| "logits/chosen": 1.7908856868743896, | |
| "logits/rejected": 1.8412494659423828, | |
| "logps/chosen": -81.26551818847656, | |
| "logps/rejected": -90.65868377685547, | |
| "loss": 0.6903, | |
| "rewards/accuracies": 0.4124999940395355, | |
| "rewards/chosen": -0.0007023714715614915, | |
| "rewards/margins": -0.004734506830573082, | |
| "rewards/rejected": 0.004032135009765625, | |
| "step": 80 | |
| }, | |
| { | |
| "epoch": 3.3488372093023258, | |
| "grad_norm": 0.5832298994064331, | |
| "learning_rate": 4.1325349624589625e-06, | |
| "logits/chosen": 1.6744792461395264, | |
| "logits/rejected": 1.7848085165023804, | |
| "logps/chosen": -69.33172607421875, | |
| "logps/rejected": -89.09760284423828, | |
| "loss": 0.6887, | |
| "rewards/accuracies": 0.612500011920929, | |
| "rewards/chosen": 0.001944942632690072, | |
| "rewards/margins": 0.011368460953235626, | |
| "rewards/rejected": -0.009423518553376198, | |
| "step": 90 | |
| }, | |
| { | |
| "epoch": 3.7209302325581395, | |
| "grad_norm": 0.5550350546836853, | |
| "learning_rate": 3.8644112754862614e-06, | |
| "logits/chosen": 1.8829429149627686, | |
| "logits/rejected": 1.8508192300796509, | |
| "logps/chosen": -87.32453918457031, | |
| "logps/rejected": -90.84036254882812, | |
| "loss": 0.6894, | |
| "rewards/accuracies": 0.5375000238418579, | |
| "rewards/chosen": -0.0012000806163996458, | |
| "rewards/margins": 0.0026218085549771786, | |
| "rewards/rejected": -0.0038218882400542498, | |
| "step": 100 | |
| }, | |
| { | |
| "epoch": 4.093023255813954, | |
| "grad_norm": 3.977888584136963, | |
| "learning_rate": 3.5717314035076355e-06, | |
| "logits/chosen": 1.787570595741272, | |
| "logits/rejected": 1.7879459857940674, | |
| "logps/chosen": -99.02583312988281, | |
| "logps/rejected": -94.5812759399414, | |
| "loss": 0.6865, | |
| "rewards/accuracies": 0.612500011920929, | |
| "rewards/chosen": 0.006580929271876812, | |
| "rewards/margins": 0.021825337782502174, | |
| "rewards/rejected": -0.015244407579302788, | |
| "step": 110 | |
| }, | |
| { | |
| "epoch": 4.465116279069767, | |
| "grad_norm": 0.6172447204589844, | |
| "learning_rate": 3.2597628939356174e-06, | |
| "logits/chosen": 1.7528693675994873, | |
| "logits/rejected": 1.7908378839492798, | |
| "logps/chosen": -79.02194213867188, | |
| "logps/rejected": -88.16065216064453, | |
| "loss": 0.689, | |
| "rewards/accuracies": 0.512499988079071, | |
| "rewards/chosen": -0.010873542167246342, | |
| "rewards/margins": 0.001626663259230554, | |
| "rewards/rejected": -0.012500204145908356, | |
| "step": 120 | |
| }, | |
| { | |
| "epoch": 4.837209302325581, | |
| "grad_norm": 0.7959902286529541, | |
| "learning_rate": 2.9341204441673267e-06, | |
| "logits/chosen": 1.8172928094863892, | |
| "logits/rejected": 1.713372826576233, | |
| "logps/chosen": -104.71771240234375, | |
| "logps/rejected": -74.14459228515625, | |
| "loss": 0.6807, | |
| "rewards/accuracies": 0.637499988079071, | |
| "rewards/chosen": 0.01367110200226307, | |
| "rewards/margins": 0.0376802422106266, | |
| "rewards/rejected": -0.024009134620428085, | |
| "step": 130 | |
| }, | |
| { | |
| "epoch": 5.209302325581396, | |
| "grad_norm": 0.7148152589797974, | |
| "learning_rate": 2.6006648502735384e-06, | |
| "logits/chosen": 1.8063843250274658, | |
| "logits/rejected": 1.7726043462753296, | |
| "logps/chosen": -94.35012817382812, | |
| "logps/rejected": -74.70433044433594, | |
| "loss": 0.679, | |
| "rewards/accuracies": 0.675000011920929, | |
| "rewards/chosen": 0.0016125829424709082, | |
| "rewards/margins": 0.048104941844940186, | |
| "rewards/rejected": -0.04649236053228378, | |
| "step": 140 | |
| }, | |
| { | |
| "epoch": 5.5813953488372094, | |
| "grad_norm": 0.7632599472999573, | |
| "learning_rate": 2.265397526492052e-06, | |
| "logits/chosen": 1.7107969522476196, | |
| "logits/rejected": 1.7916028499603271, | |
| "logps/chosen": -82.15348815917969, | |
| "logps/rejected": -72.45692443847656, | |
| "loss": 0.6803, | |
| "rewards/accuracies": 0.762499988079071, | |
| "rewards/chosen": 0.010645274072885513, | |
| "rewards/margins": 0.05290994048118591, | |
| "rewards/rejected": -0.0422646626830101, | |
| "step": 150 | |
| }, | |
| { | |
| "epoch": 5.953488372093023, | |
| "grad_norm": 0.7417441606521606, | |
| "learning_rate": 1.934352493925695e-06, | |
| "logits/chosen": 1.805847406387329, | |
| "logits/rejected": 1.853757619857788, | |
| "logps/chosen": -80.99852752685547, | |
| "logps/rejected": -85.37675476074219, | |
| "loss": 0.6804, | |
| "rewards/accuracies": 0.6000000238418579, | |
| "rewards/chosen": -0.018115444108843803, | |
| "rewards/margins": 0.026529336348176003, | |
| "rewards/rejected": -0.044644780457019806, | |
| "step": 160 | |
| }, | |
| { | |
| "epoch": 6.325581395348837, | |
| "grad_norm": 0.6947072744369507, | |
| "learning_rate": 1.613487782393661e-06, | |
| "logits/chosen": 1.78365159034729, | |
| "logits/rejected": 1.7505042552947998, | |
| "logps/chosen": -79.20752716064453, | |
| "logps/rejected": -76.4256362915039, | |
| "loss": 0.6756, | |
| "rewards/accuracies": 0.6499999761581421, | |
| "rewards/chosen": -0.028582002967596054, | |
| "rewards/margins": 0.0369187593460083, | |
| "rewards/rejected": -0.06550076603889465, | |
| "step": 170 | |
| }, | |
| { | |
| "epoch": 6.6976744186046515, | |
| "grad_norm": 0.708878755569458, | |
| "learning_rate": 1.3085781999467303e-06, | |
| "logits/chosen": 1.8475749492645264, | |
| "logits/rejected": 1.8639358282089233, | |
| "logps/chosen": -86.50167083740234, | |
| "logps/rejected": -86.65702819824219, | |
| "loss": 0.6744, | |
| "rewards/accuracies": 0.6499999761581421, | |
| "rewards/chosen": -0.020427193492650986, | |
| "rewards/margins": 0.03602122887969017, | |
| "rewards/rejected": -0.056448422372341156, | |
| "step": 180 | |
| }, | |
| { | |
| "epoch": 7.069767441860465, | |
| "grad_norm": 0.7321914434432983, | |
| "learning_rate": 1.0251113999421936e-06, | |
| "logits/chosen": 1.7573623657226562, | |
| "logits/rejected": 1.8096891641616821, | |
| "logps/chosen": -73.01332092285156, | |
| "logps/rejected": -84.01178741455078, | |
| "loss": 0.6703, | |
| "rewards/accuracies": 0.5874999761581421, | |
| "rewards/chosen": -0.03601118177175522, | |
| "rewards/margins": 0.02739904262125492, | |
| "rewards/rejected": -0.06341022253036499, | |
| "step": 190 | |
| }, | |
| { | |
| "epoch": 7.441860465116279, | |
| "grad_norm": 0.7363404035568237, | |
| "learning_rate": 7.681891162260016e-07, | |
| "logits/chosen": 1.7942943572998047, | |
| "logits/rejected": 1.8757835626602173, | |
| "logps/chosen": -71.05104064941406, | |
| "logps/rejected": -97.1144790649414, | |
| "loss": 0.6705, | |
| "rewards/accuracies": 0.6499999761581421, | |
| "rewards/chosen": -0.04499458894133568, | |
| "rewards/margins": 0.023196522146463394, | |
| "rewards/rejected": -0.06819111108779907, | |
| "step": 200 | |
| }, | |
| { | |
| "epoch": 7.813953488372093, | |
| "grad_norm": 0.751647412776947, | |
| "learning_rate": 5.424353439559446e-07, | |
| "logits/chosen": 1.8643348217010498, | |
| "logits/rejected": 1.852317452430725, | |
| "logps/chosen": -88.37493896484375, | |
| "logps/rejected": -86.77766418457031, | |
| "loss": 0.675, | |
| "rewards/accuracies": 0.637499988079071, | |
| "rewards/chosen": -0.017291061580181122, | |
| "rewards/margins": 0.036708246916532516, | |
| "rewards/rejected": -0.053999315947294235, | |
| "step": 210 | |
| }, | |
| { | |
| "epoch": 8.186046511627907, | |
| "grad_norm": 0.7266362309455872, | |
| "learning_rate": 3.51913118594458e-07, | |
| "logits/chosen": 1.7909473180770874, | |
| "logits/rejected": 1.8149019479751587, | |
| "logps/chosen": -87.76082611083984, | |
| "logps/rejected": -94.9640884399414, | |
| "loss": 0.6657, | |
| "rewards/accuracies": 0.6875, | |
| "rewards/chosen": -0.016025662422180176, | |
| "rewards/margins": 0.05901428312063217, | |
| "rewards/rejected": -0.07503994554281235, | |
| "step": 220 | |
| }, | |
| { | |
| "epoch": 8.55813953488372, | |
| "grad_norm": 0.7438246011734009, | |
| "learning_rate": 2.0005139085293945e-07, | |
| "logits/chosen": 1.805499792098999, | |
| "logits/rejected": 1.8353807926177979, | |
| "logps/chosen": -86.86074829101562, | |
| "logps/rejected": -97.83507537841797, | |
| "loss": 0.6736, | |
| "rewards/accuracies": 0.6625000238418579, | |
| "rewards/chosen": -0.034867025911808014, | |
| "rewards/margins": 0.06159989908337593, | |
| "rewards/rejected": -0.09646693617105484, | |
| "step": 230 | |
| }, | |
| { | |
| "epoch": 8.930232558139535, | |
| "grad_norm": 0.726691484451294, | |
| "learning_rate": 8.958331366609424e-08, | |
| "logits/chosen": 1.8834251165390015, | |
| "logits/rejected": 1.7817695140838623, | |
| "logps/chosen": -92.15116119384766, | |
| "logps/rejected": -71.86451721191406, | |
| "loss": 0.6683, | |
| "rewards/accuracies": 0.7250000238418579, | |
| "rewards/chosen": -0.0010798743460327387, | |
| "rewards/margins": 0.08708261698484421, | |
| "rewards/rejected": -0.08816249668598175, | |
| "step": 240 | |
| }, | |
| { | |
| "epoch": 9.30232558139535, | |
| "grad_norm": 0.8213956952095032, | |
| "learning_rate": 2.2497051885228825e-08, | |
| "logits/chosen": 1.5768067836761475, | |
| "logits/rejected": 1.6853668689727783, | |
| "logps/chosen": -83.7536849975586, | |
| "logps/rejected": -89.20189666748047, | |
| "loss": 0.6717, | |
| "rewards/accuracies": 0.550000011920929, | |
| "rewards/chosen": -0.05014566332101822, | |
| "rewards/margins": 0.031247947365045547, | |
| "rewards/rejected": -0.08139361441135406, | |
| "step": 250 | |
| }, | |
| { | |
| "epoch": 9.674418604651162, | |
| "grad_norm": 2.622415781021118, | |
| "learning_rate": 0.0, | |
| "logits/chosen": 1.8887628316879272, | |
| "logits/rejected": 1.8104219436645508, | |
| "logps/chosen": -98.27021789550781, | |
| "logps/rejected": -68.67156219482422, | |
| "loss": 0.6705, | |
| "rewards/accuracies": 0.6875, | |
| "rewards/chosen": -0.009171323850750923, | |
| "rewards/margins": 0.09182025492191315, | |
| "rewards/rejected": -0.10099156945943832, | |
| "step": 260 | |
| }, | |
| { | |
| "epoch": 9.674418604651162, | |
| "step": 260, | |
| "total_flos": 8.856435014909297e+17, | |
| "train_loss": 0.6819301917002751, | |
| "train_runtime": 1908.5849, | |
| "train_samples_per_second": 8.996, | |
| "train_steps_per_second": 0.136 | |
| } | |
| ], | |
| "logging_steps": 10, | |
| "max_steps": 260, | |
| "num_input_tokens_seen": 0, | |
| "num_train_epochs": 10, | |
| "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": 8.856435014909297e+17, | |
| "train_batch_size": 1, | |
| "trial_name": null, | |
| "trial_params": null | |
| } | |