llama.cpp

Refactoring `convert-pth-to-ggml.py`: more concise and readable

#109

Merged