| { | |
| "best_global_step": null, | |
| "best_metric": null, | |
| "best_model_checkpoint": null, | |
| "epoch": 0.9879931389365352, | |
| "eval_steps": 500, | |
| "global_step": 72, | |
| "is_hyper_param_search": false, | |
| "is_local_process_zero": true, | |
| "is_world_process_zero": true, | |
| "log_history": [ | |
| { | |
| "epoch": 0.0686106346483705, | |
| "grad_norm": 19.744274139404297, | |
| "learning_rate": 2.5e-07, | |
| "log_odds_chosen": 1.201342225074768, | |
| "log_odds_ratio": -0.3934743106365204, | |
| "logps/chosen": -0.4193040728569031, | |
| "logps/rejected": -0.8378152847290039, | |
| "loss": 12.1235, | |
| "nll_loss": 0.668493390083313, | |
| "rewards/accuracies": 0.828125, | |
| "rewards/chosen": -0.020965203642845154, | |
| "rewards/margins": 0.02092556282877922, | |
| "rewards/rejected": -0.041890766471624374, | |
| "step": 5 | |
| }, | |
| { | |
| "epoch": 0.137221269296741, | |
| "grad_norm": 17.17472267150879, | |
| "learning_rate": 4.996988640512931e-07, | |
| "log_odds_chosen": 1.2820667028427124, | |
| "log_odds_ratio": -0.38662344217300415, | |
| "logps/chosen": -0.3940594494342804, | |
| "logps/rejected": -0.8234702944755554, | |
| "loss": 11.5715, | |
| "nll_loss": 0.6343871355056763, | |
| "rewards/accuracies": 0.84375, | |
| "rewards/chosen": -0.0197029747068882, | |
| "rewards/margins": 0.02147054485976696, | |
| "rewards/rejected": -0.04117351397871971, | |
| "step": 10 | |
| }, | |
| { | |
| "epoch": 0.2058319039451115, | |
| "grad_norm": 18.691057205200195, | |
| "learning_rate": 4.892350839330522e-07, | |
| "log_odds_chosen": 1.4684958457946777, | |
| "log_odds_ratio": -0.34438949823379517, | |
| "logps/chosen": -0.35416024923324585, | |
| "logps/rejected": -0.853132426738739, | |
| "loss": 10.573, | |
| "nll_loss": 0.5747630000114441, | |
| "rewards/accuracies": 0.8812500238418579, | |
| "rewards/chosen": -0.01770801469683647, | |
| "rewards/margins": 0.024948609992861748, | |
| "rewards/rejected": -0.04265662282705307, | |
| "step": 15 | |
| }, | |
| { | |
| "epoch": 0.274442538593482, | |
| "grad_norm": 16.21430206298828, | |
| "learning_rate": 4.64432152500068e-07, | |
| "log_odds_chosen": 1.7324159145355225, | |
| "log_odds_ratio": -0.307681143283844, | |
| "logps/chosen": -0.3413434624671936, | |
| "logps/rejected": -0.8989057540893555, | |
| "loss": 9.6949, | |
| "nll_loss": 0.5221809148788452, | |
| "rewards/accuracies": 0.8843749761581421, | |
| "rewards/chosen": -0.01706717535853386, | |
| "rewards/margins": 0.027878114953637123, | |
| "rewards/rejected": -0.04494528844952583, | |
| "step": 20 | |
| }, | |
| { | |
| "epoch": 0.34305317324185247, | |
| "grad_norm": 18.64670181274414, | |
| "learning_rate": 4.2677669529663686e-07, | |
| "log_odds_chosen": 1.4884607791900635, | |
| "log_odds_ratio": -0.32770583033561707, | |
| "logps/chosen": -0.3440612554550171, | |
| "logps/rejected": -0.826368510723114, | |
| "loss": 10.6914, | |
| "nll_loss": 0.5817240476608276, | |
| "rewards/accuracies": 0.8968750238418579, | |
| "rewards/chosen": -0.017203064635396004, | |
| "rewards/margins": 0.024115364998579025, | |
| "rewards/rejected": -0.04131842777132988, | |
| "step": 25 | |
| }, | |
| { | |
| "epoch": 0.411663807890223, | |
| "grad_norm": 17.069456100463867, | |
| "learning_rate": 3.7852568604830535e-07, | |
| "log_odds_chosen": 1.5529407262802124, | |
| "log_odds_ratio": -0.30767717957496643, | |
| "logps/chosen": -0.34597113728523254, | |
| "logps/rejected": -0.8747011423110962, | |
| "loss": 10.1225, | |
| "nll_loss": 0.5479608774185181, | |
| "rewards/accuracies": 0.890625, | |
| "rewards/chosen": -0.017298556864261627, | |
| "rewards/margins": 0.02643650211393833, | |
| "rewards/rejected": -0.04373506084084511, | |
| "step": 30 | |
| }, | |
| { | |
| "epoch": 0.48027444253859347, | |
| "grad_norm": 16.975975036621094, | |
| "learning_rate": 3.2257116931361555e-07, | |
| "log_odds_chosen": 1.6338489055633545, | |
| "log_odds_ratio": -0.2982104420661926, | |
| "logps/chosen": -0.317135751247406, | |
| "logps/rejected": -0.8524333834648132, | |
| "loss": 9.8603, | |
| "nll_loss": 0.5319342613220215, | |
| "rewards/accuracies": 0.903124988079071, | |
| "rewards/chosen": -0.0158567875623703, | |
| "rewards/margins": 0.0267648808658123, | |
| "rewards/rejected": -0.0426216684281826, | |
| "step": 35 | |
| }, | |
| { | |
| "epoch": 0.548885077186964, | |
| "grad_norm": 16.777755737304688, | |
| "learning_rate": 2.6226691858185454e-07, | |
| "log_odds_chosen": 1.7881925106048584, | |
| "log_odds_ratio": -0.2799718976020813, | |
| "logps/chosen": -0.2918943166732788, | |
| "logps/rejected": -0.8554312586784363, | |
| "loss": 9.5767, | |
| "nll_loss": 0.5152069330215454, | |
| "rewards/accuracies": 0.9156249761581421, | |
| "rewards/chosen": -0.014594716019928455, | |
| "rewards/margins": 0.028176847845315933, | |
| "rewards/rejected": -0.042771559208631516, | |
| "step": 40 | |
| }, | |
| { | |
| "epoch": 0.6174957118353345, | |
| "grad_norm": 19.050371170043945, | |
| "learning_rate": 2.0122741949596793e-07, | |
| "log_odds_chosen": 1.693018913269043, | |
| "log_odds_ratio": -0.30690881609916687, | |
| "logps/chosen": -0.31784263253211975, | |
| "logps/rejected": -0.8562124371528625, | |
| "loss": 9.5933, | |
| "nll_loss": 0.5150309205055237, | |
| "rewards/accuracies": 0.8843749761581421, | |
| "rewards/chosen": -0.015892133116722107, | |
| "rewards/margins": 0.026918485760688782, | |
| "rewards/rejected": -0.04281061887741089, | |
| "step": 45 | |
| }, | |
| { | |
| "epoch": 0.6861063464837049, | |
| "grad_norm": 17.89000129699707, | |
| "learning_rate": 1.4311122664242953e-07, | |
| "log_odds_chosen": 1.7827718257904053, | |
| "log_odds_ratio": -0.2854536771774292, | |
| "logps/chosen": -0.3112075924873352, | |
| "logps/rejected": -0.8817294836044312, | |
| "loss": 9.7932, | |
| "nll_loss": 0.5288890600204468, | |
| "rewards/accuracies": 0.8999999761581421, | |
| "rewards/chosen": -0.01556038111448288, | |
| "rewards/margins": 0.028526093810796738, | |
| "rewards/rejected": -0.044086478650569916, | |
| "step": 50 | |
| }, | |
| { | |
| "epoch": 0.7547169811320755, | |
| "grad_norm": 17.733768463134766, | |
| "learning_rate": 9.140167895908865e-08, | |
| "log_odds_chosen": 1.7418218851089478, | |
| "log_odds_ratio": -0.28721481561660767, | |
| "logps/chosen": -0.3235136866569519, | |
| "logps/rejected": -0.8874639272689819, | |
| "loss": 9.8656, | |
| "nll_loss": 0.5327537655830383, | |
| "rewards/accuracies": 0.909375011920929, | |
| "rewards/chosen": -0.016175683587789536, | |
| "rewards/margins": 0.028197508305311203, | |
| "rewards/rejected": -0.04437319189310074, | |
| "step": 55 | |
| }, | |
| { | |
| "epoch": 0.823327615780446, | |
| "grad_norm": 17.464258193969727, | |
| "learning_rate": 4.919811712983879e-08, | |
| "log_odds_chosen": 1.7085039615631104, | |
| "log_odds_ratio": -0.2921965420246124, | |
| "logps/chosen": -0.33900579810142517, | |
| "logps/rejected": -0.9029625654220581, | |
| "loss": 9.9775, | |
| "nll_loss": 0.5402873754501343, | |
| "rewards/accuracies": 0.893750011920929, | |
| "rewards/chosen": -0.016950294375419617, | |
| "rewards/margins": 0.028197839856147766, | |
| "rewards/rejected": -0.045148126780986786, | |
| "step": 60 | |
| }, | |
| { | |
| "epoch": 0.8919382504288165, | |
| "grad_norm": 19.061491012573242, | |
| "learning_rate": 1.9030116872178314e-08, | |
| "log_odds_chosen": 1.8424354791641235, | |
| "log_odds_ratio": -0.26801738142967224, | |
| "logps/chosen": -0.30308759212493896, | |
| "logps/rejected": -0.9023351669311523, | |
| "loss": 9.665, | |
| "nll_loss": 0.5219208598136902, | |
| "rewards/accuracies": 0.9156249761581421, | |
| "rewards/chosen": -0.015154379419982433, | |
| "rewards/margins": 0.02996237948536873, | |
| "rewards/rejected": -0.04511675983667374, | |
| "step": 65 | |
| }, | |
| { | |
| "epoch": 0.9605488850771869, | |
| "grad_norm": 17.77915382385254, | |
| "learning_rate": 2.7058725088047464e-09, | |
| "log_odds_chosen": 1.7313966751098633, | |
| "log_odds_ratio": -0.28424638509750366, | |
| "logps/chosen": -0.3278297483921051, | |
| "logps/rejected": -0.8673001527786255, | |
| "loss": 10.1477, | |
| "nll_loss": 0.5500699877738953, | |
| "rewards/accuracies": 0.9156249761581421, | |
| "rewards/chosen": -0.016391487792134285, | |
| "rewards/margins": 0.02697352133691311, | |
| "rewards/rejected": -0.043365009129047394, | |
| "step": 70 | |
| }, | |
| { | |
| "epoch": 0.9879931389365352, | |
| "step": 72, | |
| "total_flos": 0.0, | |
| "train_loss": 10.236132118437025, | |
| "train_runtime": 760.2628, | |
| "train_samples_per_second": 6.132, | |
| "train_steps_per_second": 0.095 | |
| } | |
| ], | |
| "logging_steps": 5, | |
| "max_steps": 72, | |
| "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": 1, | |
| "trial_name": null, | |
| "trial_params": null | |
| } | |