| { | |
| "best_metric": null, | |
| "best_model_checkpoint": null, | |
| "epoch": 0.009687575684185032, | |
| "eval_steps": 3, | |
| "global_step": 10, | |
| "is_hyper_param_search": false, | |
| "is_local_process_zero": true, | |
| "is_world_process_zero": true, | |
| "log_history": [ | |
| { | |
| "epoch": 0.0009687575684185032, | |
| "grad_norm": 10.34457778930664, | |
| "learning_rate": 2e-05, | |
| "loss": 11.3937, | |
| "step": 1 | |
| }, | |
| { | |
| "epoch": 0.0009687575684185032, | |
| "eval_loss": 2.8135902881622314, | |
| "eval_runtime": 9.534, | |
| "eval_samples_per_second": 45.626, | |
| "eval_steps_per_second": 22.866, | |
| "step": 1 | |
| }, | |
| { | |
| "epoch": 0.0019375151368370065, | |
| "grad_norm": 8.082938194274902, | |
| "learning_rate": 4e-05, | |
| "loss": 10.2391, | |
| "step": 2 | |
| }, | |
| { | |
| "epoch": 0.00290627270525551, | |
| "grad_norm": 10.333882331848145, | |
| "learning_rate": 6e-05, | |
| "loss": 10.9411, | |
| "step": 3 | |
| }, | |
| { | |
| "epoch": 0.00290627270525551, | |
| "eval_loss": 2.7852697372436523, | |
| "eval_runtime": 8.083, | |
| "eval_samples_per_second": 53.816, | |
| "eval_steps_per_second": 26.97, | |
| "step": 3 | |
| }, | |
| { | |
| "epoch": 0.003875030273674013, | |
| "grad_norm": 10.25482177734375, | |
| "learning_rate": 8e-05, | |
| "loss": 10.3708, | |
| "step": 4 | |
| }, | |
| { | |
| "epoch": 0.004843787842092516, | |
| "grad_norm": 11.089507102966309, | |
| "learning_rate": 0.0001, | |
| "loss": 11.1691, | |
| "step": 5 | |
| }, | |
| { | |
| "epoch": 0.00581254541051102, | |
| "grad_norm": 9.932144165039062, | |
| "learning_rate": 0.00012, | |
| "loss": 10.6171, | |
| "step": 6 | |
| }, | |
| { | |
| "epoch": 0.00581254541051102, | |
| "eval_loss": 2.4991540908813477, | |
| "eval_runtime": 7.955, | |
| "eval_samples_per_second": 54.682, | |
| "eval_steps_per_second": 27.404, | |
| "step": 6 | |
| }, | |
| { | |
| "epoch": 0.006781302978929523, | |
| "grad_norm": 12.2509126663208, | |
| "learning_rate": 0.00014, | |
| "loss": 9.7706, | |
| "step": 7 | |
| }, | |
| { | |
| "epoch": 0.007750060547348026, | |
| "grad_norm": 9.277876853942871, | |
| "learning_rate": 0.00016, | |
| "loss": 9.0295, | |
| "step": 8 | |
| }, | |
| { | |
| "epoch": 0.00871881811576653, | |
| "grad_norm": 8.91284465789795, | |
| "learning_rate": 0.00018, | |
| "loss": 8.7262, | |
| "step": 9 | |
| }, | |
| { | |
| "epoch": 0.00871881811576653, | |
| "eval_loss": 1.898510217666626, | |
| "eval_runtime": 8.0217, | |
| "eval_samples_per_second": 54.228, | |
| "eval_steps_per_second": 27.176, | |
| "step": 9 | |
| }, | |
| { | |
| "epoch": 0.009687575684185032, | |
| "grad_norm": 7.864237308502197, | |
| "learning_rate": 0.0002, | |
| "loss": 7.9215, | |
| "step": 10 | |
| } | |
| ], | |
| "logging_steps": 1, | |
| "max_steps": 10, | |
| "num_input_tokens_seen": 0, | |
| "num_train_epochs": 1, | |
| "save_steps": 3, | |
| "stateful_callbacks": { | |
| "TrainerControl": { | |
| "args": { | |
| "should_epoch_stop": false, | |
| "should_evaluate": false, | |
| "should_log": false, | |
| "should_save": true, | |
| "should_training_stop": true | |
| }, | |
| "attributes": {} | |
| } | |
| }, | |
| "total_flos": 750891722342400.0, | |
| "train_batch_size": 2, | |
| "trial_name": null, | |
| "trial_params": null | |
| } | |