Morph MORPH
SIGN IN SIGN UP
dhruvhead / llama.cpp UNCLAIMED
0 0 1 C++

CUDA: revise q8_1 data layout for mul_mat_q (#7824)

J
Johannes Gäßler committed
42b53d192f4e3abf1b7c8e424628424504ea5dc5
Parent: 2decf57
Committed by GitHub <noreply@github.com> on 6/9/2024, 7:42:25 AM