Enhance w4afp8 performance: implement per-token w4afp8 CUTLASS MoE GEMM for FP8 dispatch, improve performance with w4afp8 moe gemm#18144

Wangzheee wants to merge 0 commit intosgl-project:mainsgl-project/sglang:mainfrom

Wangzheee:w4afp8_per-token-kernelWangzheee/sglang:w4afp8_per-token-kernelCopy head branch name to clipboard