Support docker install and Ollama

[chllei]

Thank you very much for developing such a successful and practical educational app.

I am particularly concerned about privacy. I would like to know if it is possible to use smaller models (for example 70B or even 8B) through Ollama as the inference backend on local Docker containers to set up this tutor?

[namanbnsl]

I think this might require a writeup of some part of the code because as far as I understand, the website is directly interacting with the Together API. The Together API is OpenAI compatible and so is Ollama. Maybe you can try to change API calls to the ollama openai endpoint.