llama.cpp

gguf-py : do not use internal numpy types

#7472

Merged