| { | |
| "best_metric": 0.8788848263254113, | |
| "best_model_checkpoint": "berttiny-hate_speech18-bothpretrained\\run-38\\checkpoint-3832", | |
| "epoch": 4.0, | |
| "eval_steps": 500, | |
| "global_step": 3832, | |
| "is_hyper_param_search": true, | |
| "is_local_process_zero": true, | |
| "is_world_process_zero": true, | |
| "log_history": [ | |
| { | |
| "epoch": 1.0, | |
| "learning_rate": 5.391577145227344e-05, | |
| "loss": 0.4771, | |
| "step": 958 | |
| }, | |
| { | |
| "epoch": 1.0, | |
| "eval_accuracy": 0.8656307129798904, | |
| "eval_loss": 0.4757784307003021, | |
| "eval_runtime": 10.7985, | |
| "eval_samples_per_second": 202.621, | |
| "eval_steps_per_second": 1.667, | |
| "step": 958 | |
| }, | |
| { | |
| "epoch": 2.0, | |
| "learning_rate": 3.594384763484896e-05, | |
| "loss": 0.4555, | |
| "step": 1916 | |
| }, | |
| { | |
| "epoch": 2.0, | |
| "eval_accuracy": 0.8747714808043876, | |
| "eval_loss": 0.468022882938385, | |
| "eval_runtime": 10.8634, | |
| "eval_samples_per_second": 201.41, | |
| "eval_steps_per_second": 1.657, | |
| "step": 1916 | |
| }, | |
| { | |
| "epoch": 3.0, | |
| "learning_rate": 1.797192381742448e-05, | |
| "loss": 0.4439, | |
| "step": 2874 | |
| }, | |
| { | |
| "epoch": 3.0, | |
| "eval_accuracy": 0.8775137111517367, | |
| "eval_loss": 0.47030702233314514, | |
| "eval_runtime": 10.7413, | |
| "eval_samples_per_second": 203.699, | |
| "eval_steps_per_second": 1.676, | |
| "step": 2874 | |
| }, | |
| { | |
| "epoch": 4.0, | |
| "learning_rate": 0.0, | |
| "loss": 0.4368, | |
| "step": 3832 | |
| }, | |
| { | |
| "epoch": 4.0, | |
| "eval_accuracy": 0.8788848263254113, | |
| "eval_loss": 0.4700613021850586, | |
| "eval_runtime": 10.6938, | |
| "eval_samples_per_second": 204.605, | |
| "eval_steps_per_second": 1.683, | |
| "step": 3832 | |
| } | |
| ], | |
| "logging_steps": 500, | |
| "max_steps": 3832, | |
| "num_input_tokens_seen": 0, | |
| "num_train_epochs": 4, | |
| "save_steps": 500, | |
| "total_flos": 3932355876576.0, | |
| "trial_name": null, | |
| "trial_params": { | |
| "learning_rate": 7.188769526969792e-05, | |
| "num_train_epochs": 4, | |
| "per_device_train_batch_size": 8, | |
| "seed": 34 | |
| } | |
| } | |