| { | |
| "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": 19.247772216796875, | |
| "learning_rate": 2.5e-07, | |
| "log_odds_chosen": 1.202063798904419, | |
| "log_odds_ratio": -0.3932521939277649, | |
| "logps/chosen": -0.4192759096622467, | |
| "logps/rejected": -0.8378359079360962, | |
| "loss": 11.1199, | |
| "nll_loss": 0.6684929728507996, | |
| "rewards/accuracies": 0.828125, | |
| "rewards/chosen": -0.020963797345757484, | |
| "rewards/margins": 0.020928001031279564, | |
| "rewards/rejected": -0.04189179837703705, | |
| "step": 5 | |
| }, | |
| { | |
| "epoch": 0.137221269296741, | |
| "grad_norm": 16.6769962310791, | |
| "learning_rate": 4.996988640512931e-07, | |
| "log_odds_chosen": 1.2801547050476074, | |
| "log_odds_ratio": -0.3870637118816376, | |
| "logps/chosen": -0.39408302307128906, | |
| "logps/rejected": -0.8230869174003601, | |
| "loss": 10.5716, | |
| "nll_loss": 0.6343531608581543, | |
| "rewards/accuracies": 0.84375, | |
| "rewards/chosen": -0.019704151898622513, | |
| "rewards/margins": 0.021450195461511612, | |
| "rewards/rejected": -0.041154347360134125, | |
| "step": 10 | |
| }, | |
| { | |
| "epoch": 0.2058319039451115, | |
| "grad_norm": 18.085586547851562, | |
| "learning_rate": 4.892350839330522e-07, | |
| "log_odds_chosen": 1.465559959411621, | |
| "log_odds_ratio": -0.3448963165283203, | |
| "logps/chosen": -0.3543047606945038, | |
| "logps/rejected": -0.8525373339653015, | |
| "loss": 9.6005, | |
| "nll_loss": 0.5749582052230835, | |
| "rewards/accuracies": 0.878125011920929, | |
| "rewards/chosen": -0.01771523989737034, | |
| "rewards/margins": 0.024911630898714066, | |
| "rewards/rejected": -0.04262687265872955, | |
| "step": 15 | |
| }, | |
| { | |
| "epoch": 0.274442538593482, | |
| "grad_norm": 15.663309097290039, | |
| "learning_rate": 4.64432152500068e-07, | |
| "log_odds_chosen": 1.7299188375473022, | |
| "log_odds_ratio": -0.30832645297050476, | |
| "logps/chosen": -0.3413788974285126, | |
| "logps/rejected": -0.897726833820343, | |
| "loss": 8.7406, | |
| "nll_loss": 0.5222011804580688, | |
| "rewards/accuracies": 0.887499988079071, | |
| "rewards/chosen": -0.017068946734070778, | |
| "rewards/margins": 0.027817392721772194, | |
| "rewards/rejected": -0.04488633945584297, | |
| "step": 20 | |
| }, | |
| { | |
| "epoch": 0.34305317324185247, | |
| "grad_norm": 18.158308029174805, | |
| "learning_rate": 4.2677669529663686e-07, | |
| "log_odds_chosen": 1.4837188720703125, | |
| "log_odds_ratio": -0.3287079334259033, | |
| "logps/chosen": -0.3441501259803772, | |
| "logps/rejected": -0.8242276906967163, | |
| "loss": 9.7102, | |
| "nll_loss": 0.5818053483963013, | |
| "rewards/accuracies": 0.8968750238418579, | |
| "rewards/chosen": -0.01720750704407692, | |
| "rewards/margins": 0.024003881961107254, | |
| "rewards/rejected": -0.041211389005184174, | |
| "step": 25 | |
| }, | |
| { | |
| "epoch": 0.411663807890223, | |
| "grad_norm": 16.503480911254883, | |
| "learning_rate": 3.7852568604830535e-07, | |
| "log_odds_chosen": 1.5456057786941528, | |
| "log_odds_ratio": -0.3088548183441162, | |
| "logps/chosen": -0.3460735082626343, | |
| "logps/rejected": -0.8717023134231567, | |
| "loss": 9.158, | |
| "nll_loss": 0.5480188727378845, | |
| "rewards/accuracies": 0.887499988079071, | |
| "rewards/chosen": -0.017303675413131714, | |
| "rewards/margins": 0.026281436905264854, | |
| "rewards/rejected": -0.04358511418104172, | |
| "step": 30 | |
| }, | |
| { | |
| "epoch": 0.48027444253859347, | |
| "grad_norm": 16.41570472717285, | |
| "learning_rate": 3.2257116931361555e-07, | |
| "log_odds_chosen": 1.6255455017089844, | |
| "log_odds_ratio": -0.2997768521308899, | |
| "logps/chosen": -0.3172294795513153, | |
| "logps/rejected": -0.8485333323478699, | |
| "loss": 8.8971, | |
| "nll_loss": 0.531952977180481, | |
| "rewards/accuracies": 0.903124988079071, | |
| "rewards/chosen": -0.015861475840210915, | |
| "rewards/margins": 0.026565194129943848, | |
| "rewards/rejected": -0.04242666810750961, | |
| "step": 35 | |
| }, | |
| { | |
| "epoch": 0.548885077186964, | |
| "grad_norm": 16.295194625854492, | |
| "learning_rate": 2.6226691858185454e-07, | |
| "log_odds_chosen": 1.7775678634643555, | |
| "log_odds_ratio": -0.2819564938545227, | |
| "logps/chosen": -0.29198122024536133, | |
| "logps/rejected": -0.8509271740913391, | |
| "loss": 8.6232, | |
| "nll_loss": 0.5151973962783813, | |
| "rewards/accuracies": 0.9125000238418579, | |
| "rewards/chosen": -0.014599060639739037, | |
| "rewards/margins": 0.027947301045060158, | |
| "rewards/rejected": -0.042546361684799194, | |
| "step": 40 | |
| }, | |
| { | |
| "epoch": 0.6174957118353345, | |
| "grad_norm": 18.461917877197266, | |
| "learning_rate": 2.0122741949596793e-07, | |
| "log_odds_chosen": 1.683835744857788, | |
| "log_odds_ratio": -0.30903160572052, | |
| "logps/chosen": -0.31781280040740967, | |
| "logps/rejected": -0.8516524434089661, | |
| "loss": 8.6307, | |
| "nll_loss": 0.5150163769721985, | |
| "rewards/accuracies": 0.8812500238418579, | |
| "rewards/chosen": -0.015890639275312424, | |
| "rewards/margins": 0.02669198252260685, | |
| "rewards/rejected": -0.04258262366056442, | |
| "step": 45 | |
| }, | |
| { | |
| "epoch": 0.6861063464837049, | |
| "grad_norm": 17.289121627807617, | |
| "learning_rate": 1.4311122664242953e-07, | |
| "log_odds_chosen": 1.7723395824432373, | |
| "log_odds_ratio": -0.28731250762939453, | |
| "logps/chosen": -0.31122511625289917, | |
| "logps/rejected": -0.876534640789032, | |
| "loss": 8.8422, | |
| "nll_loss": 0.5288792848587036, | |
| "rewards/accuracies": 0.8999999761581421, | |
| "rewards/chosen": -0.015561257489025593, | |
| "rewards/margins": 0.02826547622680664, | |
| "rewards/rejected": -0.04382672905921936, | |
| "step": 50 | |
| }, | |
| { | |
| "epoch": 0.7547169811320755, | |
| "grad_norm": 17.134496688842773, | |
| "learning_rate": 9.140167895908865e-08, | |
| "log_odds_chosen": 1.7291252613067627, | |
| "log_odds_ratio": -0.28900426626205444, | |
| "logps/chosen": -0.3234766721725464, | |
| "logps/rejected": -0.8822552561759949, | |
| "loss": 8.9071, | |
| "nll_loss": 0.5326792001724243, | |
| "rewards/accuracies": 0.909375011920929, | |
| "rewards/chosen": -0.0161738358438015, | |
| "rewards/margins": 0.027938928455114365, | |
| "rewards/rejected": -0.04411276429891586, | |
| "step": 55 | |
| }, | |
| { | |
| "epoch": 0.823327615780446, | |
| "grad_norm": 16.877124786376953, | |
| "learning_rate": 4.919811712983879e-08, | |
| "log_odds_chosen": 1.698340654373169, | |
| "log_odds_ratio": -0.29392507672309875, | |
| "logps/chosen": -0.3388321101665497, | |
| "logps/rejected": -0.8975754976272583, | |
| "loss": 9.0244, | |
| "nll_loss": 0.5402450561523438, | |
| "rewards/accuracies": 0.893750011920929, | |
| "rewards/chosen": -0.016941606998443604, | |
| "rewards/margins": 0.02793716825544834, | |
| "rewards/rejected": -0.044878773391246796, | |
| "step": 60 | |
| }, | |
| { | |
| "epoch": 0.8919382504288165, | |
| "grad_norm": 18.37203598022461, | |
| "learning_rate": 1.9030116872178314e-08, | |
| "log_odds_chosen": 1.82688307762146, | |
| "log_odds_ratio": -0.2703757882118225, | |
| "logps/chosen": -0.30324336886405945, | |
| "logps/rejected": -0.8961852788925171, | |
| "loss": 8.723, | |
| "nll_loss": 0.5218873023986816, | |
| "rewards/accuracies": 0.9156249761581421, | |
| "rewards/chosen": -0.015162169933319092, | |
| "rewards/margins": 0.029647093266248703, | |
| "rewards/rejected": -0.044809263199567795, | |
| "step": 65 | |
| }, | |
| { | |
| "epoch": 0.9605488850771869, | |
| "grad_norm": 17.195167541503906, | |
| "learning_rate": 2.7058725088047464e-09, | |
| "log_odds_chosen": 1.7199735641479492, | |
| "log_odds_ratio": -0.28608742356300354, | |
| "logps/chosen": -0.32787787914276123, | |
| "logps/rejected": -0.8620197176933289, | |
| "loss": 9.1848, | |
| "nll_loss": 0.5500085949897766, | |
| "rewards/accuracies": 0.918749988079071, | |
| "rewards/chosen": -0.01639389432966709, | |
| "rewards/margins": 0.02670709416270256, | |
| "rewards/rejected": -0.0431009940803051, | |
| "step": 70 | |
| }, | |
| { | |
| "epoch": 0.9879931389365352, | |
| "step": 72, | |
| "total_flos": 0.0, | |
| "train_loss": 9.270023425420126, | |
| "train_runtime": 759.6158, | |
| "train_samples_per_second": 6.137, | |
| "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 | |
| } | |