llama.cpp
CUDA: fix padding of GQA to power of 2 in FA
#19115
Merged
Overview
Commits
1
Changes
View On GitHub
FAQ Terms Privacy Refunds Impressum
Loading