supporting models so ollama can then 'support' them too
if you use llama.cpp server, it's quite a nice experience. you can even directly download stuff from Huggingface.