Skip to content

perf(moe): fuse router weight multiply with Triton kernel - #3800

Open
piyushumate wants to merge 1 commit into
NVIDIA-NeMo:mainfrom
piyushumate:piyushumate/perf/fused-moe-router-weight-mul
Open

perf(moe): fuse router weight multiply with Triton kernel#3800
piyushumate wants to merge 1 commit into
NVIDIA-NeMo:mainfrom
piyushumate:piyushumate/perf/fused-moe-router-weight-mul