llama.cpp

b3b67d6b - (try) adding helium model