| { | |
| "best_metric": 0.8030850472484714, | |
| "best_model_checkpoint": "tinybert-TG-HS-HX-parentpretrained\\run-9\\checkpoint-788", | |
| "epoch": 4.0, | |
| "eval_steps": 500, | |
| "global_step": 788, | |
| "is_hyper_param_search": true, | |
| "is_local_process_zero": true, | |
| "is_world_process_zero": true, | |
| "log_history": [ | |
| { | |
| "epoch": 1.0, | |
| "learning_rate": 1.1884492449582221e-05, | |
| "loss": 0.3771, | |
| "step": 197 | |
| }, | |
| { | |
| "epoch": 1.0, | |
| "eval_accuracy": 0.7761256253474152, | |
| "eval_loss": 0.35103824734687805, | |
| "eval_runtime": 125.0972, | |
| "eval_samples_per_second": 57.523, | |
| "eval_steps_per_second": 0.456, | |
| "step": 197 | |
| }, | |
| { | |
| "epoch": 2.0, | |
| "learning_rate": 1.0186707813927618e-05, | |
| "loss": 0.356, | |
| "step": 394 | |
| }, | |
| { | |
| "epoch": 2.0, | |
| "eval_accuracy": 0.7898832684824902, | |
| "eval_loss": 0.3439778685569763, | |
| "eval_runtime": 125.0439, | |
| "eval_samples_per_second": 57.548, | |
| "eval_steps_per_second": 0.456, | |
| "step": 394 | |
| }, | |
| { | |
| "epoch": 3.0, | |
| "learning_rate": 8.488923178273016e-06, | |
| "loss": 0.3473, | |
| "step": 591 | |
| }, | |
| { | |
| "epoch": 3.0, | |
| "eval_accuracy": 0.797526403557532, | |
| "eval_loss": 0.34075456857681274, | |
| "eval_runtime": 125.2215, | |
| "eval_samples_per_second": 57.466, | |
| "eval_steps_per_second": 0.455, | |
| "step": 591 | |
| }, | |
| { | |
| "epoch": 4.0, | |
| "learning_rate": 6.791138542618412e-06, | |
| "loss": 0.3428, | |
| "step": 788 | |
| }, | |
| { | |
| "epoch": 4.0, | |
| "eval_accuracy": 0.8030850472484714, | |
| "eval_loss": 0.33721399307250977, | |
| "eval_runtime": 124.8586, | |
| "eval_samples_per_second": 57.633, | |
| "eval_steps_per_second": 0.457, | |
| "step": 788 | |
| } | |
| ], | |
| "logging_steps": 500, | |
| "max_steps": 1576, | |
| "num_input_tokens_seen": 0, | |
| "num_train_epochs": 8, | |
| "save_steps": 500, | |
| "total_flos": 177030409545900.0, | |
| "trial_name": null, | |
| "trial_params": { | |
| "alpha": 0.625656755680376, | |
| "learning_rate": 1.3582277085236825e-05, | |
| "num_train_epochs": 8, | |
| "temperature": 10 | |
| } | |
| } | |