| { | |
| "best_global_step": null, | |
| "best_metric": null, | |
| "best_model_checkpoint": null, | |
| "epoch": 4.2949308755760365, | |
| "eval_steps": 500, | |
| "global_step": 30, | |
| "is_hyper_param_search": false, | |
| "is_local_process_zero": true, | |
| "is_world_process_zero": true, | |
| "log_history": [ | |
| { | |
| "epoch": 0.7373271889400922, | |
| "grad_norm": 1.258885555564063, | |
| "learning_rate": 6.989700043360187e-06, | |
| "loss": 0.2967, | |
| "step": 5 | |
| }, | |
| { | |
| "epoch": 1.4423963133640554, | |
| "grad_norm": 0.8498651909394829, | |
| "learning_rate": 9.999999999999999e-06, | |
| "loss": 0.257, | |
| "step": 10 | |
| }, | |
| { | |
| "epoch": 2.1474654377880182, | |
| "grad_norm": 0.6656377186139756, | |
| "learning_rate": 1.1760912590556813e-05, | |
| "loss": 0.2261, | |
| "step": 15 | |
| }, | |
| { | |
| "epoch": 2.8847926267281108, | |
| "grad_norm": 0.6127831891899057, | |
| "learning_rate": 1.301029995663981e-05, | |
| "loss": 0.2073, | |
| "step": 20 | |
| }, | |
| { | |
| "epoch": 3.589861751152074, | |
| "grad_norm": 0.5549379002562179, | |
| "learning_rate": 1.3979400086720374e-05, | |
| "loss": 0.1842, | |
| "step": 25 | |
| }, | |
| { | |
| "epoch": 4.2949308755760365, | |
| "grad_norm": 0.4947135885932024, | |
| "learning_rate": 1.4771212547196623e-05, | |
| "loss": 0.1678, | |
| "step": 30 | |
| } | |
| ], | |
| "logging_steps": 5, | |
| "max_steps": 35, | |
| "num_input_tokens_seen": 0, | |
| "num_train_epochs": 5, | |
| "save_steps": 5, | |
| "stateful_callbacks": { | |
| "TrainerControl": { | |
| "args": { | |
| "should_epoch_stop": false, | |
| "should_evaluate": false, | |
| "should_log": false, | |
| "should_save": true, | |
| "should_training_stop": false | |
| }, | |
| "attributes": {} | |
| } | |
| }, | |
| "total_flos": 481142620815360.0, | |
| "train_batch_size": 32, | |
| "trial_name": null, | |
| "trial_params": null | |
| } | |