Morph MORPH
SIGN IN SIGN UP
dhruvhead / llama.cpp UNCLAIMED
0 0 0 C++

CUDA: use MMQ instead of cuBLAS by default (#8075)

J
Johannes Gäßler committed
a818f3028d1497a51cb2b8eb7d993ad58784940e
Parent: d62e4aa
Committed by GitHub <noreply@github.com> on 6/24/2024, 3:43:42 PM