| { | |
| "best_metric": null, | |
| "best_model_checkpoint": null, | |
| "epoch": 2.757986669731096, | |
| "eval_steps": 500, | |
| "global_step": 3000, | |
| "is_hyper_param_search": false, | |
| "is_local_process_zero": true, | |
| "is_world_process_zero": true, | |
| "log_history": [ | |
| { | |
| "epoch": 0.09193288899103655, | |
| "grad_norm": 0.09970999509096146, | |
| "learning_rate": 0.00019988036932954271, | |
| "loss": 0.7687, | |
| "step": 100 | |
| }, | |
| { | |
| "epoch": 0.1838657779820731, | |
| "grad_norm": 0.11836739629507065, | |
| "learning_rate": 0.00019892504066072438, | |
| "loss": 0.5693, | |
| "step": 200 | |
| }, | |
| { | |
| "epoch": 0.2757986669731096, | |
| "grad_norm": 0.1443864107131958, | |
| "learning_rate": 0.0001970235207829469, | |
| "loss": 0.5042, | |
| "step": 300 | |
| }, | |
| { | |
| "epoch": 0.3677315559641462, | |
| "grad_norm": 0.1083601862192154, | |
| "learning_rate": 0.0001941939972186009, | |
| "loss": 0.4697, | |
| "step": 400 | |
| }, | |
| { | |
| "epoch": 0.4596644449551827, | |
| "grad_norm": 0.10843081027269363, | |
| "learning_rate": 0.00019046353359388504, | |
| "loss": 0.4447, | |
| "step": 500 | |
| }, | |
| { | |
| "epoch": 0.5515973339462192, | |
| "grad_norm": 0.11222769320011139, | |
| "learning_rate": 0.00018586781078255458, | |
| "loss": 0.4265, | |
| "step": 600 | |
| }, | |
| { | |
| "epoch": 0.6435302229372558, | |
| "grad_norm": 0.11036638170480728, | |
| "learning_rate": 0.00018045078562803203, | |
| "loss": 0.4062, | |
| "step": 700 | |
| }, | |
| { | |
| "epoch": 0.7354631119282924, | |
| "grad_norm": 0.1659434586763382, | |
| "learning_rate": 0.0001742642705081106, | |
| "loss": 0.3864, | |
| "step": 800 | |
| }, | |
| { | |
| "epoch": 0.8273960009193289, | |
| "grad_norm": 0.14999587833881378, | |
| "learning_rate": 0.00016736743776359978, | |
| "loss": 0.3771, | |
| "step": 900 | |
| }, | |
| { | |
| "epoch": 0.9193288899103654, | |
| "grad_norm": 0.10731250792741776, | |
| "learning_rate": 0.00015982625373091875, | |
| "loss": 0.3602, | |
| "step": 1000 | |
| }, | |
| { | |
| "epoch": 1.011261778901402, | |
| "grad_norm": 0.15278048813343048, | |
| "learning_rate": 0.00015171284779196334, | |
| "loss": 0.357, | |
| "step": 1100 | |
| }, | |
| { | |
| "epoch": 1.1031946678924385, | |
| "grad_norm": 0.08481089770793915, | |
| "learning_rate": 0.00014310482247611208, | |
| "loss": 0.3405, | |
| "step": 1200 | |
| }, | |
| { | |
| "epoch": 1.1951275568834752, | |
| "grad_norm": 0.09210831671953201, | |
| "learning_rate": 0.00013408451121306046, | |
| "loss": 0.3361, | |
| "step": 1300 | |
| }, | |
| { | |
| "epoch": 1.2870604458745116, | |
| "grad_norm": 0.09653517603874207, | |
| "learning_rate": 0.0001247381908358749, | |
| "loss": 0.3308, | |
| "step": 1400 | |
| }, | |
| { | |
| "epoch": 1.378993334865548, | |
| "grad_norm": 0.11181017011404037, | |
| "learning_rate": 0.00011515525636646231, | |
| "loss": 0.3277, | |
| "step": 1500 | |
| }, | |
| { | |
| "epoch": 1.4709262238565848, | |
| "grad_norm": 0.08782528340816498, | |
| "learning_rate": 0.00010542736597640826, | |
| "loss": 0.3305, | |
| "step": 1600 | |
| }, | |
| { | |
| "epoch": 1.5628591128476212, | |
| "grad_norm": 0.0979943498969078, | |
| "learning_rate": 9.564756430140164e-05, | |
| "loss": 0.3246, | |
| "step": 1700 | |
| }, | |
| { | |
| "epoch": 1.6547920018386577, | |
| "grad_norm": 0.08940441906452179, | |
| "learning_rate": 8.590939249450595e-05, | |
| "loss": 0.3229, | |
| "step": 1800 | |
| }, | |
| { | |
| "epoch": 1.7467248908296944, | |
| "grad_norm": 0.08743057399988174, | |
| "learning_rate": 7.630599353037633e-05, | |
| "loss": 0.3235, | |
| "step": 1900 | |
| }, | |
| { | |
| "epoch": 1.8386577798207309, | |
| "grad_norm": 0.12237745523452759, | |
| "learning_rate": 6.692922131794517e-05, | |
| "loss": 0.3219, | |
| "step": 2000 | |
| }, | |
| { | |
| "epoch": 1.9305906688117673, | |
| "grad_norm": 0.08483380824327469, | |
| "learning_rate": 5.7868762142672204e-05, | |
| "loss": 0.3206, | |
| "step": 2100 | |
| }, | |
| { | |
| "epoch": 2.022523557802804, | |
| "grad_norm": 0.08572427928447723, | |
| "learning_rate": 4.9211276841525744e-05, | |
| "loss": 0.3117, | |
| "step": 2200 | |
| }, | |
| { | |
| "epoch": 2.1144564467938407, | |
| "grad_norm": 0.09082117676734924, | |
| "learning_rate": 4.103957191555876e-05, | |
| "loss": 0.3008, | |
| "step": 2300 | |
| }, | |
| { | |
| "epoch": 2.206389335784877, | |
| "grad_norm": 0.09148624539375305, | |
| "learning_rate": 3.343180750816377e-05, | |
| "loss": 0.3023, | |
| "step": 2400 | |
| }, | |
| { | |
| "epoch": 2.2983222247759136, | |
| "grad_norm": 0.09966079890727997, | |
| "learning_rate": 2.6460749824479912e-05, | |
| "loss": 0.3044, | |
| "step": 2500 | |
| }, | |
| { | |
| "epoch": 2.3902551137669503, | |
| "grad_norm": 0.09154877066612244, | |
| "learning_rate": 2.019307514235741e-05, | |
| "loss": 0.3015, | |
| "step": 2600 | |
| }, | |
| { | |
| "epoch": 2.4821880027579866, | |
| "grad_norm": 0.08882308751344681, | |
| "learning_rate": 1.4688732071827094e-05, | |
| "loss": 0.3011, | |
| "step": 2700 | |
| }, | |
| { | |
| "epoch": 2.5741208917490233, | |
| "grad_norm": 0.09436959028244019, | |
| "learning_rate": 1.0000368162888795e-05, | |
| "loss": 0.2982, | |
| "step": 2800 | |
| }, | |
| { | |
| "epoch": 2.66605378074006, | |
| "grad_norm": 0.10637196153402328, | |
| "learning_rate": 6.1728263459614796e-06, | |
| "loss": 0.2993, | |
| "step": 2900 | |
| }, | |
| { | |
| "epoch": 2.757986669731096, | |
| "grad_norm": 0.10626891255378723, | |
| "learning_rate": 3.2427160214043793e-06, | |
| "loss": 0.304, | |
| "step": 3000 | |
| } | |
| ], | |
| "logging_steps": 100, | |
| "max_steps": 3261, | |
| "num_input_tokens_seen": 0, | |
| "num_train_epochs": 3, | |
| "save_steps": 1000, | |
| "stateful_callbacks": { | |
| "TrainerControl": { | |
| "args": { | |
| "should_epoch_stop": false, | |
| "should_evaluate": false, | |
| "should_log": false, | |
| "should_save": true, | |
| "should_training_stop": false | |
| }, | |
| "attributes": {} | |
| } | |
| }, | |
| "total_flos": 8.528579886191739e+17, | |
| "train_batch_size": 2, | |
| "trial_name": null, | |
| "trial_params": null | |
| } | |