| { | |
| "best_metric": 0.6096339821815491, | |
| "best_model_checkpoint": "BERT-WMM/run-1/checkpoint-1066", | |
| "epoch": 3.0, | |
| "eval_steps": 500, | |
| "global_step": 1599, | |
| "is_hyper_param_search": true, | |
| "is_local_process_zero": true, | |
| "is_world_process_zero": true, | |
| "log_history": [ | |
| { | |
| "epoch": 0.94, | |
| "grad_norm": 13.094008445739746, | |
| "learning_rate": 6.009687524192743e-06, | |
| "loss": 0.7652, | |
| "step": 500 | |
| }, | |
| { | |
| "epoch": 1.0, | |
| "eval_accuracy": 0.7244131455399061, | |
| "eval_loss": 0.6637173295021057, | |
| "eval_runtime": 2.0459, | |
| "eval_samples_per_second": 1041.116, | |
| "eval_steps_per_second": 65.497, | |
| "step": 533 | |
| }, | |
| { | |
| "epoch": 1.88, | |
| "grad_norm": 24.103755950927734, | |
| "learning_rate": 3.27552577524245e-06, | |
| "loss": 0.543, | |
| "step": 1000 | |
| }, | |
| { | |
| "epoch": 2.0, | |
| "eval_accuracy": 0.7615023474178404, | |
| "eval_loss": 0.6096339821815491, | |
| "eval_runtime": 2.2188, | |
| "eval_samples_per_second": 959.966, | |
| "eval_steps_per_second": 60.392, | |
| "step": 1066 | |
| }, | |
| { | |
| "epoch": 2.81, | |
| "grad_norm": 58.51288986206055, | |
| "learning_rate": 5.413640262921578e-07, | |
| "loss": 0.4444, | |
| "step": 1500 | |
| }, | |
| { | |
| "epoch": 3.0, | |
| "eval_accuracy": 0.7596244131455399, | |
| "eval_loss": 0.6346573233604431, | |
| "eval_runtime": 2.0157, | |
| "eval_samples_per_second": 1056.709, | |
| "eval_steps_per_second": 66.478, | |
| "step": 1599 | |
| } | |
| ], | |
| "logging_steps": 500, | |
| "max_steps": 1599, | |
| "num_input_tokens_seen": 0, | |
| "num_train_epochs": 3, | |
| "save_steps": 500, | |
| "total_flos": 507646505902536.0, | |
| "train_batch_size": 16, | |
| "trial_name": null, | |
| "trial_params": { | |
| "learning_rate": 8.743849273143035e-06 | |
| } | |
| } | |