{ "best_metric": null, "best_model_checkpoint": null, "epoch": 1.4298480786416443, "eval_steps": 200, "global_step": 800, "is_hyper_param_search": false, "is_local_process_zero": true, "is_world_process_zero": true, "log_history": [ { "epoch": 0.18, "learning_rate": 2e-05, "loss": 0.5179, "step": 100 }, { "epoch": 0.36, "learning_rate": 4e-05, "loss": 0.4435, "step": 200 }, { "epoch": 0.36, "eval_f1_score": 0.4513274336283186, "eval_label_f1": 0.7168141592920354, "eval_loss": 0.4356614947319031, "eval_runtime": 45.8378, "eval_samples_per_second": 2.182, "eval_steps_per_second": 0.087, "eval_wer": 0.059877955758962625, "step": 200 }, { "epoch": 0.54, "learning_rate": 6e-05, "loss": 0.4296, "step": 300 }, { "epoch": 0.71, "learning_rate": 8e-05, "loss": 0.4309, "step": 400 }, { "epoch": 0.71, "eval_f1_score": 0.6751054852320675, "eval_label_f1": 0.8354430379746836, "eval_loss": 0.43058258295059204, "eval_runtime": 45.2842, "eval_samples_per_second": 2.208, "eval_steps_per_second": 0.088, "eval_wer": 0.059877955758962625, "step": 400 }, { "epoch": 0.89, "learning_rate": 0.0001, "loss": 0.4262, "step": 500 }, { "epoch": 1.07, "learning_rate": 9.987820251299122e-05, "loss": 0.4235, "step": 600 }, { "epoch": 1.07, "eval_f1_score": 0.6721991701244813, "eval_label_f1": 0.8547717842323652, "eval_loss": 0.42819398641586304, "eval_runtime": 45.2307, "eval_samples_per_second": 2.211, "eval_steps_per_second": 0.088, "eval_wer": 0.059877955758962625, "step": 600 }, { "epoch": 1.25, "learning_rate": 9.951340343707852e-05, "loss": 0.423, "step": 700 }, { "epoch": 1.43, "learning_rate": 9.890738003669029e-05, "loss": 0.4267, "step": 800 }, { "epoch": 1.43, "eval_f1_score": 0.7073170731707317, "eval_label_f1": 0.845528455284553, "eval_loss": 0.4269372224807739, "eval_runtime": 45.2446, "eval_samples_per_second": 2.21, "eval_steps_per_second": 0.088, "eval_wer": 0.059877955758962625, "step": 800 } ], "logging_steps": 100, "max_steps": 5000, "num_input_tokens_seen": 0, "num_train_epochs": 9, "save_steps": 200, "total_flos": 2.191867952859288e+20, "train_batch_size": 64, "trial_name": null, "trial_params": null }