llama.cpp

7d0e23d7 - gguf-py : do not use internal numpy types (#7472)