| { | |
| "best_metric": null, | |
| "best_model_checkpoint": null, | |
| "epoch": 5.0, | |
| "eval_steps": 500, | |
| "global_step": 755, | |
| "is_hyper_param_search": false, | |
| "is_local_process_zero": true, | |
| "is_world_process_zero": true, | |
| "log_history": [ | |
| { | |
| "epoch": 1.0, | |
| "eval_loss": 0.3764611894362851, | |
| "las": 91.07359765783163, | |
| "step": 151, | |
| "uas": 93.05135951661632 | |
| }, | |
| { | |
| "epoch": 2.0, | |
| "eval_loss": 0.3381248189038352, | |
| "las": 92.36615068365154, | |
| "step": 302, | |
| "uas": 93.98573519793192 | |
| }, | |
| { | |
| "epoch": 3.0, | |
| "eval_loss": 0.36146247798674985, | |
| "las": 92.64957797365061, | |
| "step": 453, | |
| "uas": 94.15080823496433 | |
| }, | |
| { | |
| "epoch": 3.3112582781456954, | |
| "grad_norm": 1.252389669418335, | |
| "learning_rate": 6.75496688741722e-05, | |
| "loss": 0.4479, | |
| "step": 500 | |
| }, | |
| { | |
| "epoch": 4.0, | |
| "eval_loss": 0.3959472978389577, | |
| "las": 93.16971376958296, | |
| "step": 604, | |
| "uas": 94.63356900364407 | |
| } | |
| ], | |
| "logging_steps": 500, | |
| "max_steps": 755, | |
| "num_input_tokens_seen": 0, | |
| "num_train_epochs": 5, | |
| "save_steps": 500, | |
| "stateful_callbacks": { | |
| "TrainerControl": { | |
| "args": { | |
| "should_epoch_stop": false, | |
| "should_evaluate": false, | |
| "should_log": false, | |
| "should_save": true, | |
| "should_training_stop": true | |
| }, | |
| "attributes": {} | |
| } | |
| }, | |
| "total_flos": 8254637505907200.0, | |
| "train_batch_size": 64, | |
| "trial_name": null, | |
| "trial_params": null | |
| } | |