{ "effective_tokens_per_sec": 8228.629439152106, "epoch": 3.0, "total_flos": 4.957143256761631e+18, "train_loss": 0.9743868487671642, "train_runtime": 11662.4045, "train_samples_per_second": 5.649, "train_steps_per_second": 0.094 }