talk-llama : add new example + sync ggml from llama.cpp (#664) a8c74e6 unverified ggerganov commited on Mar 27, 2023