| { | |
| "best_global_step": null, | |
| "best_metric": null, | |
| "best_model_checkpoint": null, | |
| "epoch": 3.0, | |
| "eval_steps": 500, | |
| "global_step": 105, | |
| "is_hyper_param_search": false, | |
| "is_local_process_zero": true, | |
| "is_world_process_zero": true, | |
| "log_history": [ | |
| { | |
| "epoch": 0.2857142857142857, | |
| "grad_norm": 6.8948283195495605, | |
| "learning_rate": 2.7428571428571428e-05, | |
| "loss": 1.44, | |
| "step": 10 | |
| }, | |
| { | |
| "epoch": 0.5714285714285714, | |
| "grad_norm": 5.060886383056641, | |
| "learning_rate": 2.4571428571428572e-05, | |
| "loss": 0.6458, | |
| "step": 20 | |
| }, | |
| { | |
| "epoch": 0.8571428571428571, | |
| "grad_norm": 5.317773342132568, | |
| "learning_rate": 2.1714285714285715e-05, | |
| "loss": 0.5486, | |
| "step": 30 | |
| }, | |
| { | |
| "epoch": 1.1428571428571428, | |
| "grad_norm": 1.5327340364456177, | |
| "learning_rate": 1.8857142857142856e-05, | |
| "loss": 0.2922, | |
| "step": 40 | |
| }, | |
| { | |
| "epoch": 1.4285714285714286, | |
| "grad_norm": 5.422910213470459, | |
| "learning_rate": 1.6e-05, | |
| "loss": 0.1456, | |
| "step": 50 | |
| }, | |
| { | |
| "epoch": 1.7142857142857144, | |
| "grad_norm": 6.254992961883545, | |
| "learning_rate": 1.3142857142857143e-05, | |
| "loss": 0.1324, | |
| "step": 60 | |
| }, | |
| { | |
| "epoch": 2.0, | |
| "grad_norm": 1.4883731603622437, | |
| "learning_rate": 1.0285714285714286e-05, | |
| "loss": 0.1583, | |
| "step": 70 | |
| }, | |
| { | |
| "epoch": 2.2857142857142856, | |
| "grad_norm": 0.4062122404575348, | |
| "learning_rate": 7.428571428571429e-06, | |
| "loss": 0.0331, | |
| "step": 80 | |
| }, | |
| { | |
| "epoch": 2.571428571428571, | |
| "grad_norm": 0.2686053514480591, | |
| "learning_rate": 4.571428571428572e-06, | |
| "loss": 0.0317, | |
| "step": 90 | |
| }, | |
| { | |
| "epoch": 2.857142857142857, | |
| "grad_norm": 0.3767339885234833, | |
| "learning_rate": 1.7142857142857143e-06, | |
| "loss": 0.0101, | |
| "step": 100 | |
| } | |
| ], | |
| "logging_steps": 10, | |
| "max_steps": 105, | |
| "num_input_tokens_seen": 0, | |
| "num_train_epochs": 3, | |
| "save_steps": 500, | |
| "stateful_callbacks": { | |
| "TrainerControl": { | |
| "args": { | |
| "should_epoch_stop": false, | |
| "should_evaluate": false, | |
| "should_log": false, | |
| "should_save": true, | |
| "should_training_stop": true | |
| }, | |
| "attributes": {} | |
| } | |
| }, | |
| "total_flos": 6.369841546319462e+16, | |
| "train_batch_size": 8, | |
| "trial_name": null, | |
| "trial_params": null | |
| } | |