Onawa

Greetings all! I am looking for a solution to dynamically serve and cache multiple LLMs for inference. I swear that I saw a solution for this previously and bookmarked it, but after searching for... probably over a week…