Comment by BoredomIsFun

I'd stay away from ollana, just use llama.cpp; it is more up date, better performing and more flexible.