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

CUDA: use only 1 thread if fully offloaded (#2915)

J
Johannes Gäßler committed
8185710a80531e9ee0c0cb99d3a9c9af1019ab67
Parent: 7eb4117
Committed by GitHub <noreply@github.com> on 9/21/2023, 8:43:53 AM