| { | |
| "best_metric": null, | |
| "best_model_checkpoint": null, | |
| "epoch": 2.0, | |
| "global_step": 5066, | |
| "is_hyper_param_search": false, | |
| "is_local_process_zero": true, | |
| "is_world_process_zero": true, | |
| "log_history": [ | |
| { | |
| "epoch": 0.2, | |
| "learning_rate": 2.7039084090011843e-05, | |
| "loss": 1.5542, | |
| "step": 500 | |
| }, | |
| { | |
| "epoch": 0.39, | |
| "learning_rate": 2.407816818002369e-05, | |
| "loss": 0.5455, | |
| "step": 1000 | |
| }, | |
| { | |
| "epoch": 0.59, | |
| "learning_rate": 2.111725227003553e-05, | |
| "loss": 0.4854, | |
| "step": 1500 | |
| }, | |
| { | |
| "epoch": 0.79, | |
| "learning_rate": 1.8156336360047373e-05, | |
| "loss": 0.4279, | |
| "step": 2000 | |
| }, | |
| { | |
| "epoch": 0.99, | |
| "learning_rate": 1.5195420450059218e-05, | |
| "loss": 0.4169, | |
| "step": 2500 | |
| }, | |
| { | |
| "epoch": 1.18, | |
| "learning_rate": 1.2234504540071062e-05, | |
| "loss": 0.1873, | |
| "step": 3000 | |
| }, | |
| { | |
| "epoch": 1.38, | |
| "learning_rate": 9.273588630082906e-06, | |
| "loss": 0.2061, | |
| "step": 3500 | |
| }, | |
| { | |
| "epoch": 1.58, | |
| "learning_rate": 6.312672720094749e-06, | |
| "loss": 0.1751, | |
| "step": 4000 | |
| }, | |
| { | |
| "epoch": 1.78, | |
| "learning_rate": 3.3517568101065932e-06, | |
| "loss": 0.1631, | |
| "step": 4500 | |
| }, | |
| { | |
| "epoch": 1.97, | |
| "learning_rate": 3.9084090011843665e-07, | |
| "loss": 0.1665, | |
| "step": 5000 | |
| }, | |
| { | |
| "epoch": 2.0, | |
| "step": 5066, | |
| "total_flos": 1.0588267176456192e+16, | |
| "train_loss": 0.42857344546809384, | |
| "train_runtime": 460.898, | |
| "train_samples_per_second": 87.92, | |
| "train_steps_per_second": 10.992 | |
| } | |
| ], | |
| "max_steps": 5066, | |
| "num_train_epochs": 2, | |
| "total_flos": 1.0588267176456192e+16, | |
| "trial_name": null, | |
| "trial_params": null | |
| } | |