| { | |
| "best_metric": 0.6166056394577026, | |
| "best_model_checkpoint": "BERT-WMM/run-2/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": 11.24305248260498, | |
| "learning_rate": 2.922625428602704e-06, | |
| "loss": 0.7917, | |
| "step": 500 | |
| }, | |
| { | |
| "epoch": 1.0, | |
| "eval_accuracy": 0.7352112676056338, | |
| "eval_loss": 0.647247314453125, | |
| "eval_runtime": 2.1238, | |
| "eval_samples_per_second": 1002.939, | |
| "eval_steps_per_second": 63.096, | |
| "step": 533 | |
| }, | |
| { | |
| "epoch": 1.88, | |
| "grad_norm": 23.33644676208496, | |
| "learning_rate": 1.5929505293294082e-06, | |
| "loss": 0.5783, | |
| "step": 1000 | |
| }, | |
| { | |
| "epoch": 2.0, | |
| "eval_accuracy": 0.7511737089201878, | |
| "eval_loss": 0.6166056394577026, | |
| "eval_runtime": 2.7158, | |
| "eval_samples_per_second": 784.299, | |
| "eval_steps_per_second": 49.341, | |
| "step": 1066 | |
| }, | |
| { | |
| "epoch": 2.81, | |
| "grad_norm": 19.725927352905273, | |
| "learning_rate": 2.6327563005611253e-07, | |
| "loss": 0.5084, | |
| "step": 1500 | |
| }, | |
| { | |
| "epoch": 3.0, | |
| "eval_accuracy": 0.7549295774647887, | |
| "eval_loss": 0.6294087171554565, | |
| "eval_runtime": 2.0645, | |
| "eval_samples_per_second": 1031.734, | |
| "eval_steps_per_second": 64.907, | |
| "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": 4.252300327876e-06 | |
| } | |
| } | |