Each decoding method's final answer is printed as soon as it is done. Model used: HuggingFaceTB/SmolLM2-360M-Instruct.