TrainOutput(global_step=216, training_loss=1.1122471597735528, metrics={'train_runtime': 1291.5522, 'train_samples_per_second': 0.671, 'train_steps_per_second': 0.167, 'total_flos': 1466171059691520.0, 'train_loss': 1.1122471597735528, 'epoch': 2.99})