| { | |
| "best_metric": null, | |
| "best_model_checkpoint": null, | |
| "epoch": 3.0, | |
| "global_step": 6729, | |
| "is_hyper_param_search": false, | |
| "is_local_process_zero": true, | |
| "is_world_process_zero": true, | |
| "log_history": [ | |
| { | |
| "epoch": 0.22, | |
| "learning_rate": 9.25694754049636e-06, | |
| "loss": 2.6361, | |
| "step": 500 | |
| }, | |
| { | |
| "epoch": 0.45, | |
| "learning_rate": 8.513895080992719e-06, | |
| "loss": 1.8094, | |
| "step": 1000 | |
| }, | |
| { | |
| "epoch": 0.67, | |
| "learning_rate": 7.770842621489078e-06, | |
| "loss": 1.1517, | |
| "step": 1500 | |
| }, | |
| { | |
| "epoch": 0.89, | |
| "learning_rate": 7.027790161985437e-06, | |
| "loss": 0.6839, | |
| "step": 2000 | |
| }, | |
| { | |
| "epoch": 1.11, | |
| "learning_rate": 6.284737702481796e-06, | |
| "loss": 0.4235, | |
| "step": 2500 | |
| }, | |
| { | |
| "epoch": 1.34, | |
| "learning_rate": 5.541685242978154e-06, | |
| "loss": 0.3156, | |
| "step": 3000 | |
| }, | |
| { | |
| "epoch": 1.56, | |
| "learning_rate": 4.798632783474514e-06, | |
| "loss": 0.2609, | |
| "step": 3500 | |
| }, | |
| { | |
| "epoch": 1.78, | |
| "learning_rate": 4.055580323970872e-06, | |
| "loss": 0.2301, | |
| "step": 4000 | |
| }, | |
| { | |
| "epoch": 2.01, | |
| "learning_rate": 3.3125278644672313e-06, | |
| "loss": 0.2043, | |
| "step": 4500 | |
| }, | |
| { | |
| "epoch": 2.23, | |
| "learning_rate": 2.5694754049635905e-06, | |
| "loss": 0.1912, | |
| "step": 5000 | |
| }, | |
| { | |
| "epoch": 2.45, | |
| "learning_rate": 1.8264229454599496e-06, | |
| "loss": 0.1811, | |
| "step": 5500 | |
| }, | |
| { | |
| "epoch": 2.67, | |
| "learning_rate": 1.0833704859563087e-06, | |
| "loss": 0.1756, | |
| "step": 6000 | |
| }, | |
| { | |
| "epoch": 2.9, | |
| "learning_rate": 3.403180264526676e-07, | |
| "loss": 0.1709, | |
| "step": 6500 | |
| }, | |
| { | |
| "epoch": 3.0, | |
| "step": 6729, | |
| "total_flos": 1.171282836455424e+17, | |
| "train_loss": 0.6325470058006347, | |
| "train_runtime": 19128.0107, | |
| "train_samples_per_second": 5.628, | |
| "train_steps_per_second": 0.352 | |
| } | |
| ], | |
| "max_steps": 6729, | |
| "num_train_epochs": 3, | |
| "total_flos": 1.171282836455424e+17, | |
| "trial_name": null, | |
| "trial_params": null | |
| } | |