{ "effective_tokens_per_sec": 7162.660003920485, "epoch": 4.0, "total_flos": 4.038502240003031e+18, "train_loss": 0.7418717171225059, "train_runtime": 10936.9117, "train_samples_per_second": 5.508, "train_steps_per_second": 0.092 }