| { | |
| "best_metric": 0.7602739726027398, | |
| "best_model_checkpoint": "tiny-bert-sst2-distilled/run-1/checkpoint-198", | |
| "epoch": 2.0, | |
| "eval_steps": 500, | |
| "global_step": 198, | |
| "is_hyper_param_search": true, | |
| "is_local_process_zero": true, | |
| "is_world_process_zero": true, | |
| "log_history": [ | |
| { | |
| "epoch": 1.0, | |
| "grad_norm": 0.8574622273445129, | |
| "learning_rate": 0.0005047460143242413, | |
| "loss": 0.4026, | |
| "step": 99 | |
| }, | |
| { | |
| "epoch": 1.0, | |
| "eval_accuracy": 0.5684931506849316, | |
| "eval_f1": 0.28985507246376807, | |
| "eval_loss": 0.3736521005630493, | |
| "eval_precision": 0.8181818181818182, | |
| "eval_recall": 0.1761252446183953, | |
| "eval_runtime": 30.2349, | |
| "eval_samples_per_second": 33.802, | |
| "eval_steps_per_second": 1.058, | |
| "step": 99 | |
| }, | |
| { | |
| "epoch": 2.0, | |
| "grad_norm": 1.298500657081604, | |
| "learning_rate": 0.00025237300716212063, | |
| "loss": 0.3707, | |
| "step": 198 | |
| }, | |
| { | |
| "epoch": 2.0, | |
| "eval_accuracy": 0.7602739726027398, | |
| "eval_f1": 0.7231638418079097, | |
| "eval_loss": 0.3502894341945648, | |
| "eval_precision": 0.8556149732620321, | |
| "eval_recall": 0.6262230919765166, | |
| "eval_runtime": 30.2651, | |
| "eval_samples_per_second": 33.768, | |
| "eval_steps_per_second": 1.057, | |
| "step": 198 | |
| } | |
| ], | |
| "logging_steps": 500, | |
| "max_steps": 297, | |
| "num_input_tokens_seen": 0, | |
| "num_train_epochs": 3, | |
| "save_steps": 500, | |
| "total_flos": 471390394560.0, | |
| "train_batch_size": 31, | |
| "trial_name": null, | |
| "trial_params": { | |
| "alpha": 0.571532167699631, | |
| "learning_rate": 0.000757119021486362, | |
| "num_train_epochs": 3, | |
| "per_device_train_batch_size": 31, | |
| "temperature": 19 | |
| } | |
| } | |