Good to know. It also looks like you can host TabbyML as an on-premise server with docker and serve requests over a private network. Interesting to think that a

Good to know. It also looks like you can host TabbyML as an on-premise server with docker and serve requests over a private network. Interesting to think that a self-hosted GPU server might become a thing.