| { | |
| "best_metric": null, | |
| "best_model_checkpoint": null, | |
| "epoch": 2.469461817620261, | |
| "eval_steps": 500, | |
| "global_step": 2000, | |
| "is_hyper_param_search": false, | |
| "is_local_process_zero": true, | |
| "is_world_process_zero": true, | |
| "log_history": [ | |
| { | |
| "epoch": 0.617712917921396, | |
| "grad_norm": 1.8864717483520508, | |
| "learning_rate": 7.939843428100536e-05, | |
| "loss": 70.071, | |
| "step": 500 | |
| }, | |
| { | |
| "epoch": 1.2347309088101306, | |
| "grad_norm": 0.953060507774353, | |
| "learning_rate": 5.879686856201072e-05, | |
| "loss": 22.2608, | |
| "step": 1000 | |
| }, | |
| { | |
| "epoch": 1.8524438267315264, | |
| "grad_norm": 0.8967470526695251, | |
| "learning_rate": 3.8195302843016075e-05, | |
| "loss": 8.2034, | |
| "step": 1500 | |
| }, | |
| { | |
| "epoch": 2.469461817620261, | |
| "grad_norm": 0.8646658062934875, | |
| "learning_rate": 1.7593737124021424e-05, | |
| "loss": 3.6366, | |
| "step": 2000 | |
| } | |
| ], | |
| "logging_steps": 500, | |
| "max_steps": 2427, | |
| "num_input_tokens_seen": 0, | |
| "num_train_epochs": 3, | |
| "save_steps": 500, | |
| "stateful_callbacks": { | |
| "TrainerControl": { | |
| "args": { | |
| "should_epoch_stop": false, | |
| "should_evaluate": false, | |
| "should_log": false, | |
| "should_save": true, | |
| "should_training_stop": false | |
| }, | |
| "attributes": {} | |
| } | |
| }, | |
| "total_flos": 9.893135668988895e+17, | |
| "train_batch_size": 1, | |
| "trial_name": null, | |
| "trial_params": null | |
| } | |