| { | |
| "best_metric": 0.8170254403131115, | |
| "best_model_checkpoint": "tiny-bert-sst2-distilled/run-12/checkpoint-440", | |
| "epoch": 4.0, | |
| "eval_steps": 500, | |
| "global_step": 440, | |
| "is_hyper_param_search": true, | |
| "is_local_process_zero": true, | |
| "is_world_process_zero": true, | |
| "log_history": [ | |
| { | |
| "epoch": 1.0, | |
| "grad_norm": 3.2588086128234863, | |
| "learning_rate": 8.732687275158285e-05, | |
| "loss": 0.5286, | |
| "step": 110 | |
| }, | |
| { | |
| "epoch": 1.0, | |
| "eval_accuracy": 0.764187866927593, | |
| "eval_f1": 0.7826871055004507, | |
| "eval_loss": 0.45278048515319824, | |
| "eval_precision": 0.725752508361204, | |
| "eval_recall": 0.8493150684931506, | |
| "eval_runtime": 27.9751, | |
| "eval_samples_per_second": 36.532, | |
| "eval_steps_per_second": 1.144, | |
| "step": 110 | |
| }, | |
| { | |
| "epoch": 2.0, | |
| "grad_norm": 4.785890579223633, | |
| "learning_rate": 5.821791516772189e-05, | |
| "loss": 0.4524, | |
| "step": 220 | |
| }, | |
| { | |
| "epoch": 2.0, | |
| "eval_accuracy": 0.8091976516634051, | |
| "eval_f1": 0.823529411764706, | |
| "eval_loss": 0.4230343699455261, | |
| "eval_precision": 0.765993265993266, | |
| "eval_recall": 0.8904109589041096, | |
| "eval_runtime": 28.1954, | |
| "eval_samples_per_second": 36.247, | |
| "eval_steps_per_second": 1.135, | |
| "step": 220 | |
| }, | |
| { | |
| "epoch": 3.0, | |
| "grad_norm": 3.3008711338043213, | |
| "learning_rate": 2.9108957583860947e-05, | |
| "loss": 0.4339, | |
| "step": 330 | |
| }, | |
| { | |
| "epoch": 3.0, | |
| "eval_accuracy": 0.8062622309197651, | |
| "eval_f1": 0.8173431734317342, | |
| "eval_loss": 0.41464948654174805, | |
| "eval_precision": 0.7731239092495636, | |
| "eval_recall": 0.8669275929549902, | |
| "eval_runtime": 28.0059, | |
| "eval_samples_per_second": 36.492, | |
| "eval_steps_per_second": 1.143, | |
| "step": 330 | |
| }, | |
| { | |
| "epoch": 4.0, | |
| "grad_norm": 5.163329124450684, | |
| "learning_rate": 0.0, | |
| "loss": 0.4235, | |
| "step": 440 | |
| }, | |
| { | |
| "epoch": 4.0, | |
| "eval_accuracy": 0.8170254403131115, | |
| "eval_f1": 0.8316831683168315, | |
| "eval_loss": 0.4148567318916321, | |
| "eval_precision": 0.77, | |
| "eval_recall": 0.9041095890410958, | |
| "eval_runtime": 28.087, | |
| "eval_samples_per_second": 36.387, | |
| "eval_steps_per_second": 1.139, | |
| "step": 440 | |
| } | |
| ], | |
| "logging_steps": 500, | |
| "max_steps": 440, | |
| "num_input_tokens_seen": 0, | |
| "num_train_epochs": 4, | |
| "save_steps": 500, | |
| "total_flos": 942780789120.0, | |
| "train_batch_size": 28, | |
| "trial_name": null, | |
| "trial_params": { | |
| "alpha": 0.7727775292077371, | |
| "learning_rate": 0.00011643583033544379, | |
| "num_train_epochs": 4, | |
| "per_device_train_batch_size": 28, | |
| "temperature": 18 | |
| } | |
| } | |