After deploying a workaround, I can finally have peace of mind.
The local AI inference engine llama.cpp and its tensor library ggml have hit their first major release. We look at the public ...