[Quantization][MoE] Route compressed-tensors W4A8-INT MoE to WNA16 Marlin on CUDA - #58326

[Quantization][MoE] Route compressed-tensors W4A8-INT MoE to WNA16 Marlin on CUDA#58326

tangshiwen wants to merge 5 commits into

vllm-project:mainvllm-project/vllm:mainfrom

tangshiwen:maintangshiwen/vllm:mainCopy head branch name to clipboard