| { | |
| "best_metric": 0.25548699498176575, | |
| "best_model_checkpoint": "uk-mt5-small-gec/checkpoint-400", | |
| "epoch": 1.5209125475285172, | |
| "global_step": 400, | |
| "is_hyper_param_search": false, | |
| "is_local_process_zero": true, | |
| "is_world_process_zero": true, | |
| "log_history": [ | |
| { | |
| "epoch": 0.76, | |
| "learning_rate": 0.0009239543726235742, | |
| "loss": 0.8634, | |
| "step": 200 | |
| }, | |
| { | |
| "epoch": 0.76, | |
| "eval_google_bleu": 0.42247763419483103, | |
| "eval_loss": 0.30603110790252686, | |
| "eval_runtime": 109.6559, | |
| "eval_samples_per_second": 17.053, | |
| "eval_steps_per_second": 0.538, | |
| "step": 200 | |
| }, | |
| { | |
| "epoch": 1.52, | |
| "learning_rate": 0.0008479087452471484, | |
| "loss": 0.2909, | |
| "step": 400 | |
| }, | |
| { | |
| "epoch": 1.52, | |
| "eval_google_bleu": 0.4226960400591444, | |
| "eval_loss": 0.25548699498176575, | |
| "eval_runtime": 109.6393, | |
| "eval_samples_per_second": 17.056, | |
| "eval_steps_per_second": 0.538, | |
| "step": 400 | |
| } | |
| ], | |
| "max_steps": 2630, | |
| "num_train_epochs": 10, | |
| "total_flos": 3106659908413440.0, | |
| "trial_name": null, | |
| "trial_params": null | |
| } | |