| { |
| "best_global_step": null, |
| "best_metric": null, |
| "best_model_checkpoint": null, |
| "epoch": 1.0, |
| "eval_steps": 500, |
| "global_step": 75, |
| "is_hyper_param_search": false, |
| "is_local_process_zero": true, |
| "is_world_process_zero": true, |
| "log_history": [ |
| { |
| "epoch": 0.13333333333333333, |
| "grad_norm": 55.11389911931987, |
| "learning_rate": 9.994504457428556e-07, |
| "loss": 1.1211, |
| "step": 10 |
| }, |
| { |
| "epoch": 0.26666666666666666, |
| "grad_norm": 9.026758369549915, |
| "learning_rate": 9.349531862043951e-07, |
| "loss": 0.618, |
| "step": 20 |
| }, |
| { |
| "epoch": 0.4, |
| "grad_norm": 13.990650442868638, |
| "learning_rate": 7.765655770625996e-07, |
| "loss": 0.475, |
| "step": 30 |
| }, |
| { |
| "epoch": 0.5333333333333333, |
| "grad_norm": 6.730754307971874, |
| "learning_rate": 5.584776609860413e-07, |
| "loss": 0.4135, |
| "step": 40 |
| }, |
| { |
| "epoch": 0.6666666666666666, |
| "grad_norm": 8.787459788364407, |
| "learning_rate": 3.2776657485783356e-07, |
| "loss": 0.3713, |
| "step": 50 |
| }, |
| { |
| "epoch": 0.8, |
| "grad_norm": 10.319215173656126, |
| "learning_rate": 1.3423433236816562e-07, |
| "loss": 0.432, |
| "step": 60 |
| }, |
| { |
| "epoch": 0.9333333333333333, |
| "grad_norm": 8.614158289652226, |
| "learning_rate": 1.9657406211579962e-08, |
| "loss": 0.3647, |
| "step": 70 |
| }, |
| { |
| "epoch": 1.0, |
| "step": 75, |
| "total_flos": 201222733824.0, |
| "train_loss": 0.5335471979777018, |
| "train_runtime": 121.1204, |
| "train_samples_per_second": 1.238, |
| "train_steps_per_second": 0.619 |
| } |
| ], |
| "logging_steps": 10, |
| "max_steps": 75, |
| "num_input_tokens_seen": 0, |
| "num_train_epochs": 1, |
| "save_steps": 4000000, |
| "stateful_callbacks": { |
| "TrainerControl": { |
| "args": { |
| "should_epoch_stop": false, |
| "should_evaluate": false, |
| "should_log": false, |
| "should_save": true, |
| "should_training_stop": true |
| }, |
| "attributes": {} |
| } |
| }, |
| "total_flos": 201222733824.0, |
| "train_batch_size": 1, |
| "trial_name": null, |
| "trial_params": null |
| } |
|
|
