Skip to content

perf(mtp): fuse EAGLE decode into CUDA graphs - #1451

Closed
sufubao wants to merge 2 commits into
ModelTC:mainfrom
sufubao:perf/qwen35-mtp-inference
Closed

perf(mtp): fuse EAGLE decode into CUDA graphs#1451
sufubao wants to merge 2 commits into
ModelTC:mainfrom
sufubao:perf/qwen35-mtp-inference

Commits

Commits on Aug 9, 2026