| { | |
| "best_metric": 0.7348336594911937, | |
| "best_model_checkpoint": "tiny-bert-sst2-distilled/run-8/checkpoint-2298", | |
| "epoch": 3.0, | |
| "eval_steps": 500, | |
| "global_step": 2298, | |
| "is_hyper_param_search": true, | |
| "is_local_process_zero": true, | |
| "is_world_process_zero": true, | |
| "log_history": [ | |
| { | |
| "epoch": 1.0, | |
| "grad_norm": 1.324540138244629, | |
| "learning_rate": 5.804695204967391e-05, | |
| "loss": 0.357, | |
| "step": 766 | |
| }, | |
| { | |
| "epoch": 1.0, | |
| "eval_accuracy": 0.6193737769080235, | |
| "eval_f1": 0.45133991537376583, | |
| "eval_loss": 0.33073100447654724, | |
| "eval_precision": 0.8080808080808081, | |
| "eval_recall": 0.3131115459882583, | |
| "eval_runtime": 132.7451, | |
| "eval_samples_per_second": 7.699, | |
| "eval_steps_per_second": 1.929, | |
| "step": 766 | |
| }, | |
| { | |
| "epoch": 2.0, | |
| "grad_norm": 5.969402313232422, | |
| "learning_rate": 2.9023476024836956e-05, | |
| "loss": 0.3322, | |
| "step": 1532 | |
| }, | |
| { | |
| "epoch": 2.0, | |
| "eval_accuracy": 0.687866927592955, | |
| "eval_f1": 0.5997490589711417, | |
| "eval_loss": 0.3239139914512634, | |
| "eval_precision": 0.8356643356643356, | |
| "eval_recall": 0.46771037181996084, | |
| "eval_runtime": 132.1156, | |
| "eval_samples_per_second": 7.736, | |
| "eval_steps_per_second": 1.938, | |
| "step": 1532 | |
| }, | |
| { | |
| "epoch": 3.0, | |
| "grad_norm": 2.5323688983917236, | |
| "learning_rate": 0.0, | |
| "loss": 0.3258, | |
| "step": 2298 | |
| }, | |
| { | |
| "epoch": 3.0, | |
| "eval_accuracy": 0.7348336594911937, | |
| "eval_f1": 0.6937853107344633, | |
| "eval_loss": 0.3193659484386444, | |
| "eval_precision": 0.820855614973262, | |
| "eval_recall": 0.6007827788649707, | |
| "eval_runtime": 135.0076, | |
| "eval_samples_per_second": 7.57, | |
| "eval_steps_per_second": 1.896, | |
| "step": 2298 | |
| } | |
| ], | |
| "logging_steps": 500, | |
| "max_steps": 2298, | |
| "num_input_tokens_seen": 0, | |
| "num_train_epochs": 3, | |
| "save_steps": 500, | |
| "total_flos": 707085591840.0, | |
| "train_batch_size": 4, | |
| "trial_name": null, | |
| "trial_params": { | |
| "alpha": 0.49392332530271654, | |
| "learning_rate": 8.707042807451088e-05, | |
| "num_train_epochs": 3, | |
| "temperature": 24 | |
| } | |
| } | |