| { | |
| "best_metric": 0.8101761252446184, | |
| "best_model_checkpoint": "tiny-bert-sst2-distilled/run-2/checkpoint-288", | |
| "epoch": 3.0, | |
| "eval_steps": 500, | |
| "global_step": 288, | |
| "is_hyper_param_search": true, | |
| "is_local_process_zero": true, | |
| "is_world_process_zero": true, | |
| "log_history": [ | |
| { | |
| "epoch": 1.0, | |
| "grad_norm": 1.613167643547058, | |
| "learning_rate": 0.00027081825418014463, | |
| "loss": 0.516, | |
| "step": 96 | |
| }, | |
| { | |
| "epoch": 1.0, | |
| "eval_accuracy": 0.7837573385518591, | |
| "eval_f1": 0.8073234524847429, | |
| "eval_loss": 0.44109782576560974, | |
| "eval_precision": 0.7279874213836478, | |
| "eval_recall": 0.9060665362035225, | |
| "eval_runtime": 31.7637, | |
| "eval_samples_per_second": 32.175, | |
| "eval_steps_per_second": 1.007, | |
| "step": 96 | |
| }, | |
| { | |
| "epoch": 2.0, | |
| "grad_norm": 2.826476812362671, | |
| "learning_rate": 0.00013540912709007232, | |
| "loss": 0.4436, | |
| "step": 192 | |
| }, | |
| { | |
| "epoch": 2.0, | |
| "eval_accuracy": 0.7964774951076321, | |
| "eval_f1": 0.820069204152249, | |
| "eval_loss": 0.42240670323371887, | |
| "eval_precision": 0.7348837209302326, | |
| "eval_recall": 0.9275929549902152, | |
| "eval_runtime": 30.3967, | |
| "eval_samples_per_second": 33.622, | |
| "eval_steps_per_second": 1.053, | |
| "step": 192 | |
| }, | |
| { | |
| "epoch": 3.0, | |
| "grad_norm": 2.404919147491455, | |
| "learning_rate": 0.0, | |
| "loss": 0.4233, | |
| "step": 288 | |
| }, | |
| { | |
| "epoch": 3.0, | |
| "eval_accuracy": 0.8101761252446184, | |
| "eval_f1": 0.8264758497316635, | |
| "eval_loss": 0.41113388538360596, | |
| "eval_precision": 0.7611202635914333, | |
| "eval_recall": 0.9041095890410958, | |
| "eval_runtime": 30.8837, | |
| "eval_samples_per_second": 33.092, | |
| "eval_steps_per_second": 1.036, | |
| "step": 288 | |
| } | |
| ], | |
| "logging_steps": 500, | |
| "max_steps": 288, | |
| "num_input_tokens_seen": 0, | |
| "num_train_epochs": 3, | |
| "save_steps": 500, | |
| "total_flos": 707085591840.0, | |
| "train_batch_size": 32, | |
| "trial_name": null, | |
| "trial_params": { | |
| "alpha": 0.7853139256228429, | |
| "learning_rate": 0.00040622738127021695, | |
| "num_train_epochs": 3, | |
| "temperature": 5 | |
| } | |
| } | |