| { | |
| "best_metric": 0.7954990215264188, | |
| "best_model_checkpoint": "tiny-bert-sst2-distilled/run-20/checkpoint-96", | |
| "epoch": 2.0, | |
| "eval_steps": 500, | |
| "global_step": 96, | |
| "is_hyper_param_search": true, | |
| "is_local_process_zero": true, | |
| "is_world_process_zero": true, | |
| "log_history": [ | |
| { | |
| "epoch": 1.0, | |
| "grad_norm": 1.460697889328003, | |
| "learning_rate": 0.00025858582035056794, | |
| "loss": 0.5513, | |
| "step": 48 | |
| }, | |
| { | |
| "epoch": 1.0, | |
| "eval_accuracy": 0.773972602739726, | |
| "eval_f1": 0.8053917438921651, | |
| "eval_loss": 0.46848487854003906, | |
| "eval_precision": 0.7071005917159763, | |
| "eval_recall": 0.9354207436399217, | |
| "eval_runtime": 28.0493, | |
| "eval_samples_per_second": 36.436, | |
| "eval_steps_per_second": 1.141, | |
| "step": 48 | |
| }, | |
| { | |
| "epoch": 2.0, | |
| "grad_norm": 1.8203953504562378, | |
| "learning_rate": 0.0, | |
| "loss": 0.4771, | |
| "step": 96 | |
| }, | |
| { | |
| "epoch": 2.0, | |
| "eval_accuracy": 0.7954990215264188, | |
| "eval_f1": 0.8184187662901825, | |
| "eval_loss": 0.4446437954902649, | |
| "eval_precision": 0.7359375, | |
| "eval_recall": 0.9217221135029354, | |
| "eval_runtime": 28.0332, | |
| "eval_samples_per_second": 36.457, | |
| "eval_steps_per_second": 1.142, | |
| "step": 96 | |
| } | |
| ], | |
| "logging_steps": 500, | |
| "max_steps": 96, | |
| "num_input_tokens_seen": 0, | |
| "num_train_epochs": 2, | |
| "save_steps": 500, | |
| "total_flos": 471390394560.0, | |
| "train_batch_size": 64, | |
| "trial_name": null, | |
| "trial_params": { | |
| "alpha": 0.8864140050782419, | |
| "learning_rate": 0.0005171716407011359, | |
| "num_train_epochs": 2, | |
| "per_device_train_batch_size": 64, | |
| "temperature": 26 | |
| } | |
| } | |