llama.cpp

5ffc2a02 - use another model