llama.cpp

78b226a9 - gguf : initial model loading - not tested