Although it is open, its really just all code borrowed from llama.cpp.

If you want to see where the actual developers do the actual hard work, go use llama.cpp instead.