Skip to content

[TRTLLM-11484][feat] VisualGen fmha.py support per-channel V scaling - #18020

Closed
xrq-phys wants to merge 3 commits into
NVIDIA:mainfrom
xrq-phys:ruqingx/feat/qk16pv8_scale_v_ch
Closed

xrq-phys wants to merge 3 commits into
NVIDIA:mainfrom
xrq-phys:ruqingx/feat/qk16pv8_scale_v_ch