llama.cpp

convert.py: Fix loading safetensors and ggml format on Windows

#991

Merged