| { |
| "best_global_step": 60, |
| "best_metric": 0.7289719626168224, |
| "best_model_checkpoint": "/content/gemma_lora_imb/checkpoint-60", |
| "epoch": 1.1764705882352942, |
| "eval_steps": 20, |
| "global_step": 60, |
| "is_hyper_param_search": false, |
| "is_local_process_zero": true, |
| "is_world_process_zero": true, |
| "log_history": [ |
| { |
| "epoch": 0.39215686274509803, |
| "grad_norm": 60.42306137084961, |
| "learning_rate": 8.627450980392157e-06, |
| "loss": 1.610392189025879, |
| "step": 20 |
| }, |
| { |
| "epoch": 0.39215686274509803, |
| "eval_accuracy": 0.6305418719211823, |
| "eval_f1": 0.6781115879828327, |
| "eval_loss": 1.1058138608932495, |
| "eval_runtime": 10.907, |
| "eval_samples_per_second": 18.612, |
| "eval_steps_per_second": 0.642, |
| "step": 20 |
| }, |
| { |
| "epoch": 0.7843137254901961, |
| "grad_norm": 29.146869659423828, |
| "learning_rate": 6.666666666666667e-06, |
| "loss": 0.8310173988342285, |
| "step": 40 |
| }, |
| { |
| "epoch": 0.7843137254901961, |
| "eval_accuracy": 0.6896551724137931, |
| "eval_f1": 0.7149321266968326, |
| "eval_loss": 0.8462706208229065, |
| "eval_runtime": 10.8418, |
| "eval_samples_per_second": 18.724, |
| "eval_steps_per_second": 0.646, |
| "step": 40 |
| }, |
| { |
| "epoch": 1.1764705882352942, |
| "grad_norm": 52.65027618408203, |
| "learning_rate": 4.705882352941177e-06, |
| "loss": 0.5961102485656739, |
| "step": 60 |
| }, |
| { |
| "epoch": 1.1764705882352942, |
| "eval_accuracy": 0.7142857142857143, |
| "eval_f1": 0.7289719626168224, |
| "eval_loss": 0.8570315837860107, |
| "eval_runtime": 10.8268, |
| "eval_samples_per_second": 18.75, |
| "eval_steps_per_second": 0.647, |
| "step": 60 |
| } |
| ], |
| "logging_steps": 20, |
| "max_steps": 102, |
| "num_input_tokens_seen": 0, |
| "num_train_epochs": 2, |
| "save_steps": 20, |
| "stateful_callbacks": { |
| "TrainerControl": { |
| "args": { |
| "should_epoch_stop": false, |
| "should_evaluate": false, |
| "should_log": false, |
| "should_save": true, |
| "should_training_stop": false |
| }, |
| "attributes": {} |
| } |
| }, |
| "total_flos": 3346006641030144.0, |
| "train_batch_size": 32, |
| "trial_name": null, |
| "trial_params": null |
| } |
|
|
