{ "best_global_step": 1866, "best_metric": 0.8531271390639439, "best_model_checkpoint": "./results-PR/bert-base-multilingual-cased_fold_4/checkpoint-1866", "epoch": 3.0, "eval_steps": 500, "global_step": 1866, "is_hyper_param_search": false, "is_local_process_zero": true, "is_world_process_zero": true, "log_history": [ { "epoch": 0.8038585209003215, "grad_norm": 6.886026382446289, "learning_rate": 2.0056270096463023e-05, "loss": 0.5771, "step": 500 }, { "epoch": 1.0, "eval_accuracy": 0.8534621578099839, "eval_f1": 0.8530062391704848, "eval_loss": 0.5684341788291931, "eval_precision": 0.8526705892546296, "eval_recall": 0.8534621578099839, "eval_runtime": 5.0808, "eval_samples_per_second": 244.449, "eval_steps_per_second": 30.704, "step": 622 }, { "epoch": 1.607717041800643, "grad_norm": 5.165545463562012, "learning_rate": 4.0152733118971066e-05, "loss": 0.4759, "step": 1000 }, { "epoch": 2.0, "eval_accuracy": 0.3333333333333333, "eval_f1": 0.16666666666666666, "eval_loss": 0.6854161620140076, "eval_precision": 0.1111111111111111, "eval_recall": 0.3333333333333333, "eval_runtime": 5.0091, "eval_samples_per_second": 247.95, "eval_steps_per_second": 31.143, "step": 1244 }, { "epoch": 2.4115755627009645, "grad_norm": 3.9097988605499268, "learning_rate": 4.993602960395657e-05, "loss": 0.5325, "step": 1500 }, { "epoch": 3.0, "eval_accuracy": 0.8510466988727858, "eval_f1": 0.8531271390639439, "eval_loss": 0.4922334849834442, "eval_precision": 0.858653507550034, "eval_recall": 0.8510466988727858, "eval_runtime": 4.9847, "eval_samples_per_second": 249.161, "eval_steps_per_second": 31.296, "step": 1866 } ], "logging_steps": 500, "max_steps": 12440, "num_input_tokens_seen": 0, "num_train_epochs": 20, "save_steps": 500, "stateful_callbacks": { "EarlyStoppingCallback": { "args": { "early_stopping_patience": 3, "early_stopping_threshold": 0.0 }, "attributes": { "early_stopping_patience_counter": 0 } }, "TrainerControl": { "args": { "should_epoch_stop": false, "should_evaluate": false, "should_log": false, "should_save": true, "should_training_stop": false }, "attributes": {} } }, "total_flos": 3923775168583680.0, "train_batch_size": 8, "trial_name": null, "trial_params": null }