| { | |
| "best_metric": 0.8245412844036697, | |
| "best_model_checkpoint": "tiny-bert-sst2-distilled/run-0/checkpoint-1054", | |
| "epoch": 4.0, | |
| "eval_steps": 500, | |
| "global_step": 2108, | |
| "is_hyper_param_search": true, | |
| "is_local_process_zero": true, | |
| "is_world_process_zero": true, | |
| "log_history": [ | |
| { | |
| "epoch": 1.0, | |
| "learning_rate": 9.507440967713177e-05, | |
| "loss": 1.6043, | |
| "step": 527 | |
| }, | |
| { | |
| "epoch": 1.0, | |
| "eval_accuracy": 0.8107798165137615, | |
| "eval_loss": 1.3382474184036255, | |
| "eval_runtime": 7.7477, | |
| "eval_samples_per_second": 112.55, | |
| "eval_steps_per_second": 0.903, | |
| "step": 527 | |
| }, | |
| { | |
| "epoch": 2.0, | |
| "learning_rate": 6.338293978475452e-05, | |
| "loss": 0.7901, | |
| "step": 1054 | |
| }, | |
| { | |
| "epoch": 2.0, | |
| "eval_accuracy": 0.8245412844036697, | |
| "eval_loss": 1.3008999824523926, | |
| "eval_runtime": 7.5885, | |
| "eval_samples_per_second": 114.91, | |
| "eval_steps_per_second": 0.922, | |
| "step": 1054 | |
| }, | |
| { | |
| "epoch": 3.0, | |
| "learning_rate": 3.169146989237726e-05, | |
| "loss": 0.6181, | |
| "step": 1581 | |
| }, | |
| { | |
| "epoch": 3.0, | |
| "eval_accuracy": 0.8188073394495413, | |
| "eval_loss": 1.3191856145858765, | |
| "eval_runtime": 7.916, | |
| "eval_samples_per_second": 110.157, | |
| "eval_steps_per_second": 0.884, | |
| "step": 1581 | |
| }, | |
| { | |
| "epoch": 4.0, | |
| "learning_rate": 0.0, | |
| "loss": 0.5387, | |
| "step": 2108 | |
| }, | |
| { | |
| "epoch": 4.0, | |
| "eval_accuracy": 0.819954128440367, | |
| "eval_loss": 1.3426295518875122, | |
| "eval_runtime": 9.3241, | |
| "eval_samples_per_second": 93.521, | |
| "eval_steps_per_second": 0.751, | |
| "step": 2108 | |
| } | |
| ], | |
| "logging_steps": 500, | |
| "max_steps": 2108, | |
| "num_input_tokens_seen": 0, | |
| "num_train_epochs": 4, | |
| "save_steps": 500, | |
| "total_flos": 32344049138640.0, | |
| "train_batch_size": 128, | |
| "trial_name": null, | |
| "trial_params": { | |
| "alpha": 0.46638270752374766, | |
| "learning_rate": 0.00012676587956950903, | |
| "num_train_epochs": 4, | |
| "temperature": 20 | |
| } | |
| } | |