| { | |
| "best_metric": 0.8692870201096892, | |
| "best_model_checkpoint": "berttiny-hate_speech18-bothpretrained\\run-8\\checkpoint-480", | |
| "epoch": 4.0, | |
| "eval_steps": 500, | |
| "global_step": 480, | |
| "is_hyper_param_search": true, | |
| "is_local_process_zero": true, | |
| "is_world_process_zero": true, | |
| "log_history": [ | |
| { | |
| "epoch": 1.0, | |
| "learning_rate": 5.043236986060486e-05, | |
| "loss": 0.5087, | |
| "step": 120 | |
| }, | |
| { | |
| "epoch": 1.0, | |
| "eval_accuracy": 0.8555758683729433, | |
| "eval_loss": 0.4896513521671295, | |
| "eval_runtime": 10.857, | |
| "eval_samples_per_second": 201.529, | |
| "eval_steps_per_second": 1.658, | |
| "step": 120 | |
| }, | |
| { | |
| "epoch": 2.0, | |
| "learning_rate": 3.782427739545364e-05, | |
| "loss": 0.4749, | |
| "step": 240 | |
| }, | |
| { | |
| "epoch": 2.0, | |
| "eval_accuracy": 0.8619744058500914, | |
| "eval_loss": 0.4745013117790222, | |
| "eval_runtime": 10.9152, | |
| "eval_samples_per_second": 200.455, | |
| "eval_steps_per_second": 1.649, | |
| "step": 240 | |
| }, | |
| { | |
| "epoch": 3.0, | |
| "learning_rate": 2.521618493030243e-05, | |
| "loss": 0.4634, | |
| "step": 360 | |
| }, | |
| { | |
| "epoch": 3.0, | |
| "eval_accuracy": 0.8665447897623401, | |
| "eval_loss": 0.47145363688468933, | |
| "eval_runtime": 10.8657, | |
| "eval_samples_per_second": 201.368, | |
| "eval_steps_per_second": 1.657, | |
| "step": 360 | |
| }, | |
| { | |
| "epoch": 4.0, | |
| "learning_rate": 1.2608092465151215e-05, | |
| "loss": 0.4568, | |
| "step": 480 | |
| }, | |
| { | |
| "epoch": 4.0, | |
| "eval_accuracy": 0.8692870201096892, | |
| "eval_loss": 0.46969375014305115, | |
| "eval_runtime": 10.7546, | |
| "eval_samples_per_second": 203.448, | |
| "eval_steps_per_second": 1.674, | |
| "step": 480 | |
| } | |
| ], | |
| "logging_steps": 500, | |
| "max_steps": 600, | |
| "num_input_tokens_seen": 0, | |
| "num_train_epochs": 5, | |
| "save_steps": 500, | |
| "total_flos": 8040247658880.0, | |
| "trial_name": null, | |
| "trial_params": { | |
| "learning_rate": 6.304046232575607e-05, | |
| "num_train_epochs": 5, | |
| "per_device_train_batch_size": 64, | |
| "seed": 1 | |
| } | |
| } | |