{ "effective_tokens_per_sec": 904.370374320899, "epoch": 3.0, "total_flos": 6.574576867100066e+18, "train_loss": 0.8942287936473116, "train_runtime": 42713.458, "train_samples_per_second": 0.92, "train_steps_per_second": 0.058 }