| { |
| "best_metric": 0.34402796626091003, |
| "best_model_checkpoint": "moments-7_lr0.00001_beta0.999/checkpoint-1125", |
| "epoch": 5.0, |
| "eval_steps": 500, |
| "global_step": 1125, |
| "is_hyper_param_search": false, |
| "is_local_process_zero": true, |
| "is_world_process_zero": true, |
| "log_history": [ |
| { |
| "epoch": 1.0, |
| "eval_f1": 0.0, |
| "eval_loss": 0.6633961796760559, |
| "eval_runtime": 3.1235, |
| "eval_samples_per_second": 96.045, |
| "eval_steps_per_second": 12.166, |
| "step": 225 |
| }, |
| { |
| "epoch": 2.0, |
| "eval_f1": 0.0, |
| "eval_loss": 0.3961953818798065, |
| "eval_runtime": 3.134, |
| "eval_samples_per_second": 95.724, |
| "eval_steps_per_second": 12.125, |
| "step": 450 |
| }, |
| { |
| "epoch": 2.2222222222222223, |
| "grad_norm": 2.4490463733673096, |
| "learning_rate": 8.888888888888888e-06, |
| "loss": 0.845, |
| "step": 500 |
| }, |
| { |
| "epoch": 3.0, |
| "eval_f1": 0.0, |
| "eval_loss": 0.5220192670822144, |
| "eval_runtime": 3.1258, |
| "eval_samples_per_second": 95.977, |
| "eval_steps_per_second": 12.157, |
| "step": 675 |
| }, |
| { |
| "epoch": 4.0, |
| "eval_f1": 0.41025641025641024, |
| "eval_loss": 0.36355268955230713, |
| "eval_runtime": 3.1229, |
| "eval_samples_per_second": 96.065, |
| "eval_steps_per_second": 12.168, |
| "step": 900 |
| }, |
| { |
| "epoch": 4.444444444444445, |
| "grad_norm": 0.07460984587669373, |
| "learning_rate": 7.77777777777778e-06, |
| "loss": 0.394, |
| "step": 1000 |
| }, |
| { |
| "epoch": 5.0, |
| "eval_f1": 0.56, |
| "eval_loss": 0.34402796626091003, |
| "eval_runtime": 3.1356, |
| "eval_samples_per_second": 95.676, |
| "eval_steps_per_second": 12.119, |
| "step": 1125 |
| } |
| ], |
| "logging_steps": 500, |
| "max_steps": 4500, |
| "num_input_tokens_seen": 0, |
| "num_train_epochs": 20, |
| "save_steps": 500, |
| "stateful_callbacks": { |
| "TrainerControl": { |
| "args": { |
| "should_epoch_stop": false, |
| "should_evaluate": false, |
| "should_log": false, |
| "should_save": true, |
| "should_training_stop": false |
| }, |
| "attributes": {} |
| } |
| }, |
| "total_flos": 2885015596951992.0, |
| "train_batch_size": 8, |
| "trial_name": null, |
| "trial_params": null |
| } |
|
|
