| { | |
| "best_metric": 0.8788848263254113, | |
| "best_model_checkpoint": "berttiny-hate_speech18-bothpretrained\\run-44\\checkpoint-1437", | |
| "epoch": 4.0, | |
| "eval_steps": 500, | |
| "global_step": 1916, | |
| "is_hyper_param_search": true, | |
| "is_local_process_zero": true, | |
| "is_world_process_zero": true, | |
| "log_history": [ | |
| { | |
| "epoch": 1.0, | |
| "learning_rate": 4.5410240915591696e-05, | |
| "loss": 0.4851, | |
| "step": 479 | |
| }, | |
| { | |
| "epoch": 1.0, | |
| "eval_accuracy": 0.8596892138939671, | |
| "eval_loss": 0.47658035159111023, | |
| "eval_runtime": 10.8435, | |
| "eval_samples_per_second": 201.78, | |
| "eval_steps_per_second": 1.66, | |
| "step": 479 | |
| }, | |
| { | |
| "epoch": 2.0, | |
| "learning_rate": 3.0273493943727797e-05, | |
| "loss": 0.4611, | |
| "step": 958 | |
| }, | |
| { | |
| "epoch": 2.0, | |
| "eval_accuracy": 0.8747714808043876, | |
| "eval_loss": 0.4699705243110657, | |
| "eval_runtime": 11.0102, | |
| "eval_samples_per_second": 198.725, | |
| "eval_steps_per_second": 1.635, | |
| "step": 958 | |
| }, | |
| { | |
| "epoch": 3.0, | |
| "learning_rate": 1.5136746971863899e-05, | |
| "loss": 0.4512, | |
| "step": 1437 | |
| }, | |
| { | |
| "epoch": 3.0, | |
| "eval_accuracy": 0.8788848263254113, | |
| "eval_loss": 0.4677220582962036, | |
| "eval_runtime": 10.9155, | |
| "eval_samples_per_second": 200.449, | |
| "eval_steps_per_second": 1.649, | |
| "step": 1437 | |
| }, | |
| { | |
| "epoch": 4.0, | |
| "learning_rate": 0.0, | |
| "loss": 0.4463, | |
| "step": 1916 | |
| }, | |
| { | |
| "epoch": 4.0, | |
| "eval_accuracy": 0.8788848263254113, | |
| "eval_loss": 0.4686000347137451, | |
| "eval_runtime": 10.9127, | |
| "eval_samples_per_second": 200.501, | |
| "eval_steps_per_second": 1.649, | |
| "step": 1916 | |
| } | |
| ], | |
| "logging_steps": 500, | |
| "max_steps": 1916, | |
| "num_input_tokens_seen": 0, | |
| "num_train_epochs": 4, | |
| "save_steps": 500, | |
| "total_flos": 4969223449632.0, | |
| "trial_name": null, | |
| "trial_params": { | |
| "learning_rate": 6.0546987887455595e-05, | |
| "num_train_epochs": 4, | |
| "per_device_train_batch_size": 16, | |
| "seed": 34 | |
| } | |
| } | |