| { | |
| "best_metric": null, | |
| "best_model_checkpoint": null, | |
| "epoch": 3.0, | |
| "eval_steps": 500, | |
| "global_step": 2712, | |
| "is_hyper_param_search": false, | |
| "is_local_process_zero": true, | |
| "is_world_process_zero": true, | |
| "log_history": [ | |
| { | |
| "epoch": 0.2, | |
| "learning_rate": 1.9852941176470586e-05, | |
| "loss": 1.1651, | |
| "step": 180 | |
| }, | |
| { | |
| "epoch": 0.4, | |
| "learning_rate": 2.8918032786885247e-05, | |
| "loss": 1.0203, | |
| "step": 360 | |
| }, | |
| { | |
| "epoch": 0.6, | |
| "learning_rate": 2.670491803278689e-05, | |
| "loss": 0.9511, | |
| "step": 540 | |
| }, | |
| { | |
| "epoch": 0.8, | |
| "learning_rate": 2.4491803278688524e-05, | |
| "loss": 0.9128, | |
| "step": 720 | |
| }, | |
| { | |
| "epoch": 1.0, | |
| "learning_rate": 2.2278688524590165e-05, | |
| "loss": 0.9155, | |
| "step": 900 | |
| }, | |
| { | |
| "epoch": 1.19, | |
| "learning_rate": 2.0065573770491804e-05, | |
| "loss": 0.8939, | |
| "step": 1080 | |
| }, | |
| { | |
| "epoch": 1.39, | |
| "learning_rate": 1.7852459016393445e-05, | |
| "loss": 0.8317, | |
| "step": 1260 | |
| }, | |
| { | |
| "epoch": 1.59, | |
| "learning_rate": 1.563934426229508e-05, | |
| "loss": 0.8674, | |
| "step": 1440 | |
| }, | |
| { | |
| "epoch": 1.79, | |
| "learning_rate": 1.3426229508196722e-05, | |
| "loss": 0.8924, | |
| "step": 1620 | |
| }, | |
| { | |
| "epoch": 1.99, | |
| "learning_rate": 1.1213114754098362e-05, | |
| "loss": 0.8266, | |
| "step": 1800 | |
| }, | |
| { | |
| "epoch": 2.19, | |
| "learning_rate": 9e-06, | |
| "loss": 0.8614, | |
| "step": 1980 | |
| }, | |
| { | |
| "epoch": 2.39, | |
| "learning_rate": 6.786885245901639e-06, | |
| "loss": 0.885, | |
| "step": 2160 | |
| }, | |
| { | |
| "epoch": 2.59, | |
| "learning_rate": 4.586065573770491e-06, | |
| "loss": 0.8276, | |
| "step": 2340 | |
| }, | |
| { | |
| "epoch": 2.79, | |
| "learning_rate": 2.3729508196721312e-06, | |
| "loss": 0.8301, | |
| "step": 2520 | |
| }, | |
| { | |
| "epoch": 2.99, | |
| "learning_rate": 1.5983606557377048e-07, | |
| "loss": 0.8375, | |
| "step": 2700 | |
| } | |
| ], | |
| "logging_steps": 180, | |
| "max_steps": 2712, | |
| "num_train_epochs": 3, | |
| "save_steps": 500, | |
| "total_flos": 1.1044941545550643e+17, | |
| "trial_name": null, | |
| "trial_params": null | |
| } | |