| { | |
| "best_metric": null, | |
| "best_model_checkpoint": null, | |
| "epoch": 3.0, | |
| "global_step": 6999, | |
| "is_hyper_param_search": false, | |
| "is_local_process_zero": true, | |
| "is_world_process_zero": true, | |
| "log_history": [ | |
| { | |
| "epoch": 0.21, | |
| "learning_rate": 4.642806115159309e-05, | |
| "loss": 2.6203, | |
| "step": 500 | |
| }, | |
| { | |
| "epoch": 0.43, | |
| "learning_rate": 4.285612230318617e-05, | |
| "loss": 2.4108, | |
| "step": 1000 | |
| }, | |
| { | |
| "epoch": 0.64, | |
| "learning_rate": 3.928418345477925e-05, | |
| "loss": 2.3186, | |
| "step": 1500 | |
| }, | |
| { | |
| "epoch": 0.86, | |
| "learning_rate": 3.571224460637234e-05, | |
| "loss": 2.2361, | |
| "step": 2000 | |
| }, | |
| { | |
| "epoch": 1.07, | |
| "learning_rate": 3.214030575796543e-05, | |
| "loss": 2.0706, | |
| "step": 2500 | |
| }, | |
| { | |
| "epoch": 1.29, | |
| "learning_rate": 2.856836690955851e-05, | |
| "loss": 1.7984, | |
| "step": 3000 | |
| }, | |
| { | |
| "epoch": 1.5, | |
| "learning_rate": 2.4996428061151594e-05, | |
| "loss": 1.8225, | |
| "step": 3500 | |
| }, | |
| { | |
| "epoch": 1.71, | |
| "learning_rate": 2.142448921274468e-05, | |
| "loss": 1.7665, | |
| "step": 4000 | |
| }, | |
| { | |
| "epoch": 1.93, | |
| "learning_rate": 1.7852550364337764e-05, | |
| "loss": 1.7577, | |
| "step": 4500 | |
| }, | |
| { | |
| "epoch": 2.14, | |
| "learning_rate": 1.4280611515930847e-05, | |
| "loss": 1.5149, | |
| "step": 5000 | |
| }, | |
| { | |
| "epoch": 2.36, | |
| "learning_rate": 1.0708672667523933e-05, | |
| "loss": 1.4315, | |
| "step": 5500 | |
| }, | |
| { | |
| "epoch": 2.57, | |
| "learning_rate": 7.1367338191170165e-06, | |
| "loss": 1.4296, | |
| "step": 6000 | |
| }, | |
| { | |
| "epoch": 2.79, | |
| "learning_rate": 3.5647949707101015e-06, | |
| "loss": 1.4128, | |
| "step": 6500 | |
| }, | |
| { | |
| "epoch": 3.0, | |
| "step": 6999, | |
| "total_flos": 7003986897076224.0, | |
| "train_loss": 1.8582639229571858, | |
| "train_runtime": 1597.9104, | |
| "train_samples_per_second": 43.784, | |
| "train_steps_per_second": 4.38 | |
| } | |
| ], | |
| "max_steps": 6999, | |
| "num_train_epochs": 3, | |
| "total_flos": 7003986897076224.0, | |
| "trial_name": null, | |
| "trial_params": null | |
| } | |