feat: Add w4a8_mxfp4_fp8 quantization recipe.#4867

Tracin merged 3 commits intoNVIDIA:mainNVIDIA/TensorRT-LLM:mainfrom

Tracin:qmodeTracin/TensorRT-LLM:qmodeCopy head branch name to clipboard