| { | |
| "best_global_step": null, | |
| "best_metric": null, | |
| "best_model_checkpoint": null, | |
| "epoch": 2.0964360587002098, | |
| "eval_steps": 500, | |
| "global_step": 1000, | |
| "is_hyper_param_search": true, | |
| "is_local_process_zero": true, | |
| "is_world_process_zero": true, | |
| "log_history": [ | |
| { | |
| "epoch": 1.0, | |
| "eval_accuracy": 0.8709677419354839, | |
| "eval_loss": 2.2650985717773438, | |
| "eval_runtime": 17.8499, | |
| "eval_samples_per_second": 173.671, | |
| "eval_steps_per_second": 5.434, | |
| "step": 477 | |
| }, | |
| { | |
| "epoch": 1.0482180293501049, | |
| "grad_norm": 14.103622436523438, | |
| "learning_rate": 1.302585604472397e-05, | |
| "loss": 6.1632, | |
| "step": 500 | |
| }, | |
| { | |
| "epoch": 2.0, | |
| "eval_accuracy": 0.9187096774193548, | |
| "eval_loss": 1.4211522340774536, | |
| "eval_runtime": 24.025, | |
| "eval_samples_per_second": 129.032, | |
| "eval_steps_per_second": 4.037, | |
| "step": 954 | |
| }, | |
| { | |
| "epoch": 2.0964360587002098, | |
| "grad_norm": 6.999321937561035, | |
| "learning_rate": 6.037735849056604e-06, | |
| "loss": 1.5867, | |
| "step": 1000 | |
| } | |
| ], | |
| "logging_steps": 500, | |
| "max_steps": 1431, | |
| "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": 448725995623620.0, | |
| "train_batch_size": 32, | |
| "trial_name": null, | |
| "trial_params": { | |
| "alpha": 0.8329290526522004, | |
| "num_train_epochs": 3, | |
| "temperature": 4.156866262168391 | |
| } | |
| } | |