llama.cpp

45eba936 - build : use std::make_tuple() for compatibility with older GCC versions (#3488)