I want to do this for some cut of Wikipedia. If there is already a RAG for Wikipedia, that would be cool. I want my own ... I don't know ... is it the Encyclopedia Galactica? (Or Hitchhiker's Guide to Earth and its Inhabitants?)
Here's a simple example using a part of a DBpedia dataset, but you can download other Wikipedia datasets and import those: https://github.com/philippgille/chromem-go/tree/v0.7.0/examp...
(Disclosure: I'm the author)
Very cool — thanks. I'll play with it.
Neat. Does ollama have native support for storing documents ? I dont think the author wrote anything custom here (and I mean it in a good way)
The Open-WebUI has a built-in vector database and document storage.
Examples would be good - I'm not going to run all that locally just to check whether it's useful. (LLM helpers rarely are)
Yes, I will add some examples within the next days.
I hope now I'll finally figure out how to patch KDE2 for FreeBSD
In that case the RAG should be based on an anime dataset