Please be aware that the current `attention_mask` handling in AutoRound may require additional refinement. https://github.com/vllm-project/llm-compressor/issues/2076