UPDATE THE LLAMA.CPP

[raphaelallain]

new versions of llms not supported with the onde ampere computing llama.cpp, can you guys update or give the source so we can do it?