{ "effective_tokens_per_sec": 6353.293364414007, "epoch": 3.0, "total_flos": 2.502111532985352e+18, "train_loss": 0.8120331417161523, "train_runtime": 7663.1544, "train_samples_per_second": 5.888, "train_steps_per_second": 0.092 }