| { | |
| "best_metric": null, | |
| "best_model_checkpoint": null, | |
| "epoch": 0.05855855855855856, | |
| "eval_steps": 13, | |
| "global_step": 39, | |
| "is_hyper_param_search": false, | |
| "is_local_process_zero": true, | |
| "is_world_process_zero": true, | |
| "log_history": [ | |
| { | |
| "epoch": 0.0015015015015015015, | |
| "eval_loss": 2.496234178543091, | |
| "eval_runtime": 9.2514, | |
| "eval_samples_per_second": 30.374, | |
| "eval_steps_per_second": 15.241, | |
| "step": 1 | |
| }, | |
| { | |
| "epoch": 0.015015015015015015, | |
| "grad_norm": 2.077500104904175, | |
| "learning_rate": 0.0002, | |
| "loss": 2.109, | |
| "step": 10 | |
| }, | |
| { | |
| "epoch": 0.01951951951951952, | |
| "eval_loss": 0.898557186126709, | |
| "eval_runtime": 7.9856, | |
| "eval_samples_per_second": 35.188, | |
| "eval_steps_per_second": 17.657, | |
| "step": 13 | |
| }, | |
| { | |
| "epoch": 0.03003003003003003, | |
| "grad_norm": 1.4663447141647339, | |
| "learning_rate": 0.00017071067811865476, | |
| "loss": 0.904, | |
| "step": 20 | |
| }, | |
| { | |
| "epoch": 0.03903903903903904, | |
| "eval_loss": 0.6685091853141785, | |
| "eval_runtime": 8.0321, | |
| "eval_samples_per_second": 34.985, | |
| "eval_steps_per_second": 17.555, | |
| "step": 26 | |
| }, | |
| { | |
| "epoch": 0.04504504504504504, | |
| "grad_norm": 1.526362657546997, | |
| "learning_rate": 0.0001, | |
| "loss": 0.7679, | |
| "step": 30 | |
| }, | |
| { | |
| "epoch": 0.05855855855855856, | |
| "eval_loss": 0.6266762018203735, | |
| "eval_runtime": 8.0081, | |
| "eval_samples_per_second": 35.089, | |
| "eval_steps_per_second": 17.607, | |
| "step": 39 | |
| } | |
| ], | |
| "logging_steps": 10, | |
| "max_steps": 50, | |
| "num_input_tokens_seen": 0, | |
| "num_train_epochs": 1, | |
| "save_steps": 13, | |
| "stateful_callbacks": { | |
| "TrainerControl": { | |
| "args": { | |
| "should_epoch_stop": false, | |
| "should_evaluate": false, | |
| "should_log": false, | |
| "should_save": true, | |
| "should_training_stop": false | |
| }, | |
| "attributes": {} | |
| } | |
| }, | |
| "total_flos": 6352059230060544.0, | |
| "train_batch_size": 2, | |
| "trial_name": null, | |
| "trial_params": null | |
| } | |