Please make sure to update to the latest llama.cpp version