Y
Hacker News
new
|
ask
|
show
|
jobs
by
Dkuku
873 days ago
In llama.cpp You can offload some of the layers to gpu with -ngl X. Where x is the number of layers