[None][chroe] Polish qwen3-next modeling code.#8902

nv-guomingz merged 1 commit intoNVIDIA:mainNVIDIA/TensorRT-LLM:mainfrom

nv-guomingz:user/guomingz/clean_qwen3_nextnv-guomingz/TensorRT-LLM:user/guomingz/clean_qwen3_nextCopy head branch name to clipboard