| { | |
| "best_metric": null, | |
| "best_model_checkpoint": null, | |
| "epoch": 7.174887892376682, | |
| "eval_steps": 500, | |
| "global_step": 200, | |
| "is_hyper_param_search": false, | |
| "is_local_process_zero": true, | |
| "is_world_process_zero": true, | |
| "log_history": [ | |
| { | |
| "epoch": 0.35874439461883406, | |
| "grad_norm": 5.573123455047607, | |
| "learning_rate": 9.259259259259259e-07, | |
| "logits/chosen": -2.3576977252960205, | |
| "logits/rejected": -2.3553850650787354, | |
| "logps/chosen": -61.0846061706543, | |
| "logps/rejected": -126.1152572631836, | |
| "loss": 0.6942, | |
| "rewards/accuracies": 0.36250001192092896, | |
| "rewards/chosen": -0.0012381849810481071, | |
| "rewards/margins": -0.010831715539097786, | |
| "rewards/rejected": 0.009593529626727104, | |
| "step": 10 | |
| }, | |
| { | |
| "epoch": 0.7174887892376681, | |
| "grad_norm": 3.8175244331359863, | |
| "learning_rate": 1.8518518518518519e-06, | |
| "logits/chosen": -2.3618054389953613, | |
| "logits/rejected": -2.384873867034912, | |
| "logps/chosen": -63.2865104675293, | |
| "logps/rejected": -81.8724365234375, | |
| "loss": 0.6925, | |
| "rewards/accuracies": 0.4749999940395355, | |
| "rewards/chosen": -0.002874236088246107, | |
| "rewards/margins": -0.003056168556213379, | |
| "rewards/rejected": 0.00018193255527876318, | |
| "step": 20 | |
| }, | |
| { | |
| "epoch": 1.0762331838565022, | |
| "grad_norm": 4.383662700653076, | |
| "learning_rate": 2.7777777777777783e-06, | |
| "logits/chosen": -2.3333091735839844, | |
| "logits/rejected": -2.3545360565185547, | |
| "logps/chosen": -76.02193450927734, | |
| "logps/rejected": -88.19218444824219, | |
| "loss": 0.6932, | |
| "rewards/accuracies": 0.5249999761581421, | |
| "rewards/chosen": -0.008687756024301052, | |
| "rewards/margins": -0.0016246589366346598, | |
| "rewards/rejected": -0.007063096854835749, | |
| "step": 30 | |
| }, | |
| { | |
| "epoch": 1.4349775784753362, | |
| "grad_norm": 4.095139503479004, | |
| "learning_rate": 3.7037037037037037e-06, | |
| "logits/chosen": -2.3483197689056396, | |
| "logits/rejected": -2.363131523132324, | |
| "logps/chosen": -66.08761596679688, | |
| "logps/rejected": -78.22222900390625, | |
| "loss": 0.6889, | |
| "rewards/accuracies": 0.5249999761581421, | |
| "rewards/chosen": -0.014177674427628517, | |
| "rewards/margins": 0.012472175993025303, | |
| "rewards/rejected": -0.026649847626686096, | |
| "step": 40 | |
| }, | |
| { | |
| "epoch": 1.7937219730941703, | |
| "grad_norm": 4.034746170043945, | |
| "learning_rate": 4.62962962962963e-06, | |
| "logits/chosen": -2.31459903717041, | |
| "logits/rejected": -2.335706949234009, | |
| "logps/chosen": -69.31814575195312, | |
| "logps/rejected": -77.91527557373047, | |
| "loss": 0.6823, | |
| "rewards/accuracies": 0.4749999940395355, | |
| "rewards/chosen": -0.04294499754905701, | |
| "rewards/margins": 0.007987136952579021, | |
| "rewards/rejected": -0.050932131707668304, | |
| "step": 50 | |
| }, | |
| { | |
| "epoch": 2.1524663677130045, | |
| "grad_norm": 4.620233535766602, | |
| "learning_rate": 4.998119881260576e-06, | |
| "logits/chosen": -2.364583969116211, | |
| "logits/rejected": -2.3598217964172363, | |
| "logps/chosen": -69.84207153320312, | |
| "logps/rejected": -79.67817687988281, | |
| "loss": 0.6717, | |
| "rewards/accuracies": 0.699999988079071, | |
| "rewards/chosen": -0.06739393621683121, | |
| "rewards/margins": 0.050740111619234085, | |
| "rewards/rejected": -0.118134044110775, | |
| "step": 60 | |
| }, | |
| { | |
| "epoch": 2.5112107623318387, | |
| "grad_norm": 4.450082302093506, | |
| "learning_rate": 4.9866405060165044e-06, | |
| "logits/chosen": -2.298379421234131, | |
| "logits/rejected": -2.3169469833374023, | |
| "logps/chosen": -79.13768005371094, | |
| "logps/rejected": -79.24192810058594, | |
| "loss": 0.6499, | |
| "rewards/accuracies": 0.7124999761581421, | |
| "rewards/chosen": -0.1282506138086319, | |
| "rewards/margins": 0.08815944194793701, | |
| "rewards/rejected": -0.21641004085540771, | |
| "step": 70 | |
| }, | |
| { | |
| "epoch": 2.8699551569506725, | |
| "grad_norm": 4.7328877449035645, | |
| "learning_rate": 4.964774158361991e-06, | |
| "logits/chosen": -2.3363795280456543, | |
| "logits/rejected": -2.3331849575042725, | |
| "logps/chosen": -77.52214050292969, | |
| "logps/rejected": -86.90077209472656, | |
| "loss": 0.6346, | |
| "rewards/accuracies": 0.75, | |
| "rewards/chosen": -0.22365322709083557, | |
| "rewards/margins": 0.1937221735715866, | |
| "rewards/rejected": -0.4173754155635834, | |
| "step": 80 | |
| }, | |
| { | |
| "epoch": 3.2286995515695067, | |
| "grad_norm": 4.016219615936279, | |
| "learning_rate": 4.93261217644956e-06, | |
| "logits/chosen": -2.30871844291687, | |
| "logits/rejected": -2.3049044609069824, | |
| "logps/chosen": -65.6957778930664, | |
| "logps/rejected": -77.02860260009766, | |
| "loss": 0.6004, | |
| "rewards/accuracies": 0.762499988079071, | |
| "rewards/chosen": -0.23905189335346222, | |
| "rewards/margins": 0.2932591438293457, | |
| "rewards/rejected": -0.5323110222816467, | |
| "step": 90 | |
| }, | |
| { | |
| "epoch": 3.587443946188341, | |
| "grad_norm": 4.553714752197266, | |
| "learning_rate": 4.8902889044347e-06, | |
| "logits/chosen": -2.3068125247955322, | |
| "logits/rejected": -2.294462203979492, | |
| "logps/chosen": -85.46375274658203, | |
| "logps/rejected": -79.72854614257812, | |
| "loss": 0.5816, | |
| "rewards/accuracies": 0.6875, | |
| "rewards/chosen": -0.3506511151790619, | |
| "rewards/margins": 0.2997886538505554, | |
| "rewards/rejected": -0.6504397392272949, | |
| "step": 100 | |
| }, | |
| { | |
| "epoch": 3.9461883408071747, | |
| "grad_norm": 4.804734230041504, | |
| "learning_rate": 4.837981131305475e-06, | |
| "logits/chosen": -2.3317933082580566, | |
| "logits/rejected": -2.3493664264678955, | |
| "logps/chosen": -68.45785522460938, | |
| "logps/rejected": -88.28421020507812, | |
| "loss": 0.5672, | |
| "rewards/accuracies": 0.6875, | |
| "rewards/chosen": -0.4280410706996918, | |
| "rewards/margins": 0.3461475670337677, | |
| "rewards/rejected": -0.7741886377334595, | |
| "step": 110 | |
| }, | |
| { | |
| "epoch": 4.304932735426009, | |
| "grad_norm": 4.650998592376709, | |
| "learning_rate": 4.775907352415367e-06, | |
| "logits/chosen": -2.3039748668670654, | |
| "logits/rejected": -2.301835775375366, | |
| "logps/chosen": -76.26728820800781, | |
| "logps/rejected": -87.93345642089844, | |
| "loss": 0.5258, | |
| "rewards/accuracies": 0.8999999761581421, | |
| "rewards/chosen": -0.4848151206970215, | |
| "rewards/margins": 0.6179562211036682, | |
| "rewards/rejected": -1.102771282196045, | |
| "step": 120 | |
| }, | |
| { | |
| "epoch": 4.663677130044843, | |
| "grad_norm": 4.17809534072876, | |
| "learning_rate": 4.70432685680402e-06, | |
| "logits/chosen": -2.299182653427124, | |
| "logits/rejected": -2.291717290878296, | |
| "logps/chosen": -77.27394104003906, | |
| "logps/rejected": -76.96590423583984, | |
| "loss": 0.5141, | |
| "rewards/accuracies": 0.7749999761581421, | |
| "rewards/chosen": -0.5012012720108032, | |
| "rewards/margins": 0.39593780040740967, | |
| "rewards/rejected": -0.8971391916275024, | |
| "step": 130 | |
| }, | |
| { | |
| "epoch": 5.022421524663677, | |
| "grad_norm": 4.4158735275268555, | |
| "learning_rate": 4.623538644118244e-06, | |
| "logits/chosen": -2.267500400543213, | |
| "logits/rejected": -2.271239995956421, | |
| "logps/chosen": -77.53614807128906, | |
| "logps/rejected": -105.53443908691406, | |
| "loss": 0.4929, | |
| "rewards/accuracies": 0.8999999761581421, | |
| "rewards/chosen": -0.47719526290893555, | |
| "rewards/margins": 0.7529892921447754, | |
| "rewards/rejected": -1.230184555053711, | |
| "step": 140 | |
| }, | |
| { | |
| "epoch": 5.381165919282511, | |
| "grad_norm": 4.5556640625, | |
| "learning_rate": 4.533880175657419e-06, | |
| "logits/chosen": -2.2397818565368652, | |
| "logits/rejected": -2.2796878814697266, | |
| "logps/chosen": -75.30777740478516, | |
| "logps/rejected": -97.59407043457031, | |
| "loss": 0.4545, | |
| "rewards/accuracies": 0.9125000238418579, | |
| "rewards/chosen": -0.3897199034690857, | |
| "rewards/margins": 0.8326279520988464, | |
| "rewards/rejected": -1.2223479747772217, | |
| "step": 150 | |
| }, | |
| { | |
| "epoch": 5.739910313901345, | |
| "grad_norm": 5.067012786865234, | |
| "learning_rate": 4.435725964760331e-06, | |
| "logits/chosen": -2.284294843673706, | |
| "logits/rejected": -2.264681339263916, | |
| "logps/chosen": -87.51654052734375, | |
| "logps/rejected": -82.73191833496094, | |
| "loss": 0.4392, | |
| "rewards/accuracies": 0.8374999761581421, | |
| "rewards/chosen": -0.42664116621017456, | |
| "rewards/margins": 0.7053943872451782, | |
| "rewards/rejected": -1.132035493850708, | |
| "step": 160 | |
| }, | |
| { | |
| "epoch": 6.098654708520179, | |
| "grad_norm": 4.425904273986816, | |
| "learning_rate": 4.329486012421531e-06, | |
| "logits/chosen": -2.296356201171875, | |
| "logits/rejected": -2.2917532920837402, | |
| "logps/chosen": -76.75779724121094, | |
| "logps/rejected": -90.15029907226562, | |
| "loss": 0.4153, | |
| "rewards/accuracies": 0.875, | |
| "rewards/chosen": -0.5271193385124207, | |
| "rewards/margins": 0.7608373761177063, | |
| "rewards/rejected": -1.2879568338394165, | |
| "step": 170 | |
| }, | |
| { | |
| "epoch": 6.457399103139013, | |
| "grad_norm": 4.75029182434082, | |
| "learning_rate": 4.215604094671835e-06, | |
| "logits/chosen": -2.2561912536621094, | |
| "logits/rejected": -2.253948450088501, | |
| "logps/chosen": -68.14691162109375, | |
| "logps/rejected": -85.39097595214844, | |
| "loss": 0.3785, | |
| "rewards/accuracies": 0.925000011920929, | |
| "rewards/chosen": -0.5065070390701294, | |
| "rewards/margins": 1.024714708328247, | |
| "rewards/rejected": -1.531221628189087, | |
| "step": 180 | |
| }, | |
| { | |
| "epoch": 6.816143497757848, | |
| "grad_norm": 5.048355579376221, | |
| "learning_rate": 4.094555908876765e-06, | |
| "logits/chosen": -2.323347568511963, | |
| "logits/rejected": -2.284219741821289, | |
| "logps/chosen": -74.8355484008789, | |
| "logps/rejected": -85.9789047241211, | |
| "loss": 0.3697, | |
| "rewards/accuracies": 0.9125000238418579, | |
| "rewards/chosen": -0.533603310585022, | |
| "rewards/margins": 1.2120137214660645, | |
| "rewards/rejected": -1.7456169128417969, | |
| "step": 190 | |
| }, | |
| { | |
| "epoch": 7.174887892376682, | |
| "grad_norm": 5.01900053024292, | |
| "learning_rate": 3.966847086696045e-06, | |
| "logits/chosen": -2.300300121307373, | |
| "logits/rejected": -2.3180909156799316, | |
| "logps/chosen": -86.38873291015625, | |
| "logps/rejected": -89.72538757324219, | |
| "loss": 0.3611, | |
| "rewards/accuracies": 0.8374999761581421, | |
| "rewards/chosen": -0.7510987520217896, | |
| "rewards/margins": 0.9809296727180481, | |
| "rewards/rejected": -1.7320283651351929, | |
| "step": 200 | |
| } | |
| ], | |
| "logging_steps": 10, | |
| "max_steps": 540, | |
| "num_input_tokens_seen": 0, | |
| "num_train_epochs": 20, | |
| "save_steps": 100, | |
| "stateful_callbacks": { | |
| "TrainerControl": { | |
| "args": { | |
| "should_epoch_stop": false, | |
| "should_evaluate": false, | |
| "should_log": false, | |
| "should_save": true, | |
| "should_training_stop": false | |
| }, | |
| "attributes": {} | |
| } | |
| }, | |
| "total_flos": 7.331387755546542e+17, | |
| "train_batch_size": 1, | |
| "trial_name": null, | |
| "trial_params": null | |
| } | |