{ "effective_tokens_per_sec": 8375.265574307812, "epoch": 4.0, "total_flos": 3.5948540672197263e+18, "train_loss": 0.8366947202287017, "train_runtime": 8313.2571, "train_samples_per_second": 6.842, "train_steps_per_second": 0.228 }