| { | |
| "best_metric": null, | |
| "best_model_checkpoint": null, | |
| "epoch": 5.0, | |
| "eval_steps": 500, | |
| "global_step": 1110, | |
| "is_hyper_param_search": false, | |
| "is_local_process_zero": true, | |
| "is_world_process_zero": true, | |
| "log_history": [ | |
| { | |
| "epoch": 1.0, | |
| "eval_loss": 0.8045559507179059, | |
| "las": 79.02982532963557, | |
| "step": 222, | |
| "uas": 85.29168491217827 | |
| }, | |
| { | |
| "epoch": 2.0, | |
| "eval_loss": 0.6223693984418602, | |
| "las": 83.65202160268575, | |
| "step": 444, | |
| "uas": 88.52722230331338 | |
| }, | |
| { | |
| "epoch": 2.2522522522522523, | |
| "grad_norm": 2.5357823371887207, | |
| "learning_rate": 0.00010990990990990993, | |
| "loss": 1.1667, | |
| "step": 500 | |
| }, | |
| { | |
| "epoch": 3.0, | |
| "eval_loss": 0.5744358095324645, | |
| "las": 85.34033960978932, | |
| "step": 666, | |
| "uas": 89.64628034836764 | |
| }, | |
| { | |
| "epoch": 4.0, | |
| "eval_loss": 0.5734678677172732, | |
| "las": 86.92648275190969, | |
| "step": 888, | |
| "uas": 90.83345497007737 | |
| }, | |
| { | |
| "epoch": 4.504504504504505, | |
| "grad_norm": 2.42525315284729, | |
| "learning_rate": 1.981981981981982e-05, | |
| "loss": 0.2744, | |
| "step": 1000 | |
| } | |
| ], | |
| "logging_steps": 500, | |
| "max_steps": 1110, | |
| "num_input_tokens_seen": 0, | |
| "num_train_epochs": 5, | |
| "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": 1.215058602403584e+16, | |
| "train_batch_size": 64, | |
| "trial_name": null, | |
| "trial_params": null | |
| } | |