| { | |
| "best_metric": null, | |
| "best_model_checkpoint": null, | |
| "epoch": 3.0, | |
| "eval_steps": 500, | |
| "global_step": 27, | |
| "is_hyper_param_search": false, | |
| "is_local_process_zero": true, | |
| "is_world_process_zero": true, | |
| "log_history": [ | |
| { | |
| "epoch": 1.0, | |
| "eval_loss": 0.9566018581390381, | |
| "eval_runtime": 6.459, | |
| "eval_samples_per_second": 37.313, | |
| "eval_steps_per_second": 0.619, | |
| "step": 9 | |
| }, | |
| { | |
| "epoch": 1.1111111111111112, | |
| "grad_norm": 10.51255522349344, | |
| "learning_rate": 5e-06, | |
| "loss": 1.0205, | |
| "step": 10 | |
| }, | |
| { | |
| "epoch": 2.0, | |
| "eval_loss": 0.8994330763816833, | |
| "eval_runtime": 6.3624, | |
| "eval_samples_per_second": 37.879, | |
| "eval_steps_per_second": 0.629, | |
| "step": 18 | |
| }, | |
| { | |
| "epoch": 2.2222222222222223, | |
| "grad_norm": 2.3618296130194647, | |
| "learning_rate": 5e-06, | |
| "loss": 0.8881, | |
| "step": 20 | |
| }, | |
| { | |
| "epoch": 3.0, | |
| "eval_loss": 0.8868955373764038, | |
| "eval_runtime": 5.256, | |
| "eval_samples_per_second": 45.852, | |
| "eval_steps_per_second": 0.761, | |
| "step": 27 | |
| }, | |
| { | |
| "epoch": 3.0, | |
| "step": 27, | |
| "total_flos": 45016625971200.0, | |
| "train_loss": 0.925646322744864, | |
| "train_runtime": 1179.1398, | |
| "train_samples_per_second": 11.632, | |
| "train_steps_per_second": 0.023 | |
| } | |
| ], | |
| "logging_steps": 10, | |
| "max_steps": 27, | |
| "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": 45016625971200.0, | |
| "train_batch_size": 8, | |
| "trial_name": null, | |
| "trial_params": null | |
| } | |