I’m pretty sure you can’t use it without connecting to the private model binary server.
It’s a very small step to a paid docker hub, cough sorry, ollama hub.
It does not just magically conjure LLM model files out of thin air.
Where do those models come from?
https://github.com/ollama/ollama/issues/2390
The registry is not open source.
You think I’m being unfair?
https://github.com/ollama/ollama/issues/914#issuecomment-195...
(Paraphrased)
>> How do I run my own registry?
> email us, let’s talk.