| { | |
| "best_metric": 0.45115517656589194, | |
| "best_model_checkpoint": "hw-01-finetuned-cola/run-0/checkpoint-3207", | |
| "epoch": 3.0, | |
| "eval_steps": 500, | |
| "global_step": 3207, | |
| "is_hyper_param_search": true, | |
| "is_local_process_zero": true, | |
| "is_world_process_zero": true, | |
| "log_history": [ | |
| { | |
| "epoch": 0.47, | |
| "learning_rate": 2.7326435592927605e-06, | |
| "loss": 0.5977, | |
| "step": 500 | |
| }, | |
| { | |
| "epoch": 0.94, | |
| "learning_rate": 2.4506370000262214e-06, | |
| "loss": 0.5335, | |
| "step": 1000 | |
| }, | |
| { | |
| "epoch": 1.0, | |
| "eval_loss": 0.5340704321861267, | |
| "eval_matthews_correlation": 0.36161681105121063, | |
| "eval_runtime": 0.7405, | |
| "eval_samples_per_second": 1408.534, | |
| "eval_steps_per_second": 89.131, | |
| "step": 1069 | |
| }, | |
| { | |
| "epoch": 1.4, | |
| "learning_rate": 2.168630440759683e-06, | |
| "loss": 0.4751, | |
| "step": 1500 | |
| }, | |
| { | |
| "epoch": 1.87, | |
| "learning_rate": 1.886623881493144e-06, | |
| "loss": 0.4579, | |
| "step": 2000 | |
| }, | |
| { | |
| "epoch": 2.0, | |
| "eval_loss": 0.5136641263961792, | |
| "eval_matthews_correlation": 0.4491534084057902, | |
| "eval_runtime": 0.906, | |
| "eval_samples_per_second": 1151.249, | |
| "eval_steps_per_second": 72.85, | |
| "step": 2138 | |
| }, | |
| { | |
| "epoch": 2.34, | |
| "learning_rate": 1.6046173222266052e-06, | |
| "loss": 0.4286, | |
| "step": 2500 | |
| }, | |
| { | |
| "epoch": 2.81, | |
| "learning_rate": 1.3226107629600667e-06, | |
| "loss": 0.4101, | |
| "step": 3000 | |
| }, | |
| { | |
| "epoch": 3.0, | |
| "eval_loss": 0.5329485535621643, | |
| "eval_matthews_correlation": 0.45115517656589194, | |
| "eval_runtime": 0.7978, | |
| "eval_samples_per_second": 1307.396, | |
| "eval_steps_per_second": 82.731, | |
| "step": 3207 | |
| } | |
| ], | |
| "logging_steps": 500, | |
| "max_steps": 5345, | |
| "num_input_tokens_seen": 0, | |
| "num_train_epochs": 5, | |
| "save_steps": 500, | |
| "total_flos": 112981754784168.0, | |
| "train_batch_size": 8, | |
| "trial_name": null, | |
| "trial_params": { | |
| "learning_rate": 3.014650118559299e-06, | |
| "num_train_epochs": 5, | |
| "per_device_train_batch_size": 8, | |
| "seed": 16 | |
| } | |
| } | |