llama.cpp

2d43387d - ggml : fix builds, add ggml-quants-k.o (close #1712, close #1710)