Skip to content

fix: avoid f16 overflow in Z-Image quantized matmuls on CUDA - #2074

Merged
leejet merged 2 commits into
leejet:masterfrom
losewayy:fix/zimage-quant-outproj-scale
Oct 6, 2026
Merged

leejet merged 2 commits into
leejet:masterfrom
losewayy:fix/zimage-quant-outproj-scale

apply Z-Image attention output scaling only to quantized weights

9cb258b
Select commit
Loading
Failed to load commit list.
Sign in for the full log view