| { | |
| "best_metric": 0.6102677583694458, | |
| "best_model_checkpoint": "BERT-WMM/run-15/checkpoint-1066", | |
| "epoch": 5.0, | |
| "eval_steps": 500, | |
| "global_step": 2665, | |
| "is_hyper_param_search": true, | |
| "is_local_process_zero": true, | |
| "is_world_process_zero": true, | |
| "log_history": [ | |
| { | |
| "epoch": 0.94, | |
| "grad_norm": 8.692239761352539, | |
| "learning_rate": 1.3731493717417162e-05, | |
| "loss": 0.7076, | |
| "step": 500 | |
| }, | |
| { | |
| "epoch": 1.0, | |
| "eval_accuracy": 0.7530516431924883, | |
| "eval_loss": 0.6211019158363342, | |
| "eval_runtime": 1.9926, | |
| "eval_samples_per_second": 1068.944, | |
| "eval_steps_per_second": 67.248, | |
| "step": 533 | |
| }, | |
| { | |
| "epoch": 1.88, | |
| "grad_norm": 11.677574157714844, | |
| "learning_rate": 1.0560248055196108e-05, | |
| "loss": 0.4535, | |
| "step": 1000 | |
| }, | |
| { | |
| "epoch": 2.0, | |
| "eval_accuracy": 0.7680751173708921, | |
| "eval_loss": 0.6102677583694458, | |
| "eval_runtime": 2.2362, | |
| "eval_samples_per_second": 952.498, | |
| "eval_steps_per_second": 59.922, | |
| "step": 1066 | |
| }, | |
| { | |
| "epoch": 2.81, | |
| "grad_norm": 33.26390838623047, | |
| "learning_rate": 7.3890023929750545e-06, | |
| "loss": 0.2958, | |
| "step": 1500 | |
| }, | |
| { | |
| "epoch": 3.0, | |
| "eval_accuracy": 0.7751173708920188, | |
| "eval_loss": 0.7481958866119385, | |
| "eval_runtime": 2.135, | |
| "eval_samples_per_second": 997.66, | |
| "eval_steps_per_second": 62.764, | |
| "step": 1599 | |
| }, | |
| { | |
| "epoch": 3.75, | |
| "grad_norm": 14.7713041305542, | |
| "learning_rate": 4.217756730754001e-06, | |
| "loss": 0.1874, | |
| "step": 2000 | |
| }, | |
| { | |
| "epoch": 4.0, | |
| "eval_accuracy": 0.7708920187793428, | |
| "eval_loss": 1.0311119556427002, | |
| "eval_runtime": 2.0623, | |
| "eval_samples_per_second": 1032.824, | |
| "eval_steps_per_second": 64.976, | |
| "step": 2132 | |
| }, | |
| { | |
| "epoch": 4.69, | |
| "grad_norm": 1.1739749908447266, | |
| "learning_rate": 1.0465110685329476e-06, | |
| "loss": 0.1051, | |
| "step": 2500 | |
| }, | |
| { | |
| "epoch": 5.0, | |
| "eval_accuracy": 0.7727699530516432, | |
| "eval_loss": 1.1397253274917603, | |
| "eval_runtime": 2.1823, | |
| "eval_samples_per_second": 976.021, | |
| "eval_steps_per_second": 61.402, | |
| "step": 2665 | |
| } | |
| ], | |
| "logging_steps": 500, | |
| "max_steps": 2665, | |
| "num_input_tokens_seen": 0, | |
| "num_train_epochs": 5, | |
| "save_steps": 500, | |
| "total_flos": 856808289095472.0, | |
| "train_batch_size": 16, | |
| "trial_name": null, | |
| "trial_params": { | |
| "learning_rate": 1.6902739379638215e-05, | |
| "num_train_epochs": 5, | |
| "per_device_train_batch_size": 16, | |
| "weight_decay": 0.019154977870537472 | |
| } | |
| } | |