Skip to content

[CUDA] Add opt-in single-warp INT4 GEMV for FP16 decode - #33952

Open
Sunghoon Choi (hanbitmyths) wants to merge 4 commits into
mainfrom
sunghcho/qwen-vl-single-warp-int4-gemv-20261106
Open

Sunghoon Choi (hanbitmyths) wants to merge 4 commits into
mainfrom
sunghcho/qwen-vl-single-warp-int4-gemv-20261106

Commits

  1. Commits on Nov 6, 2026

  2. Commits on Nov 7, 2026