Jean-Baptiste-Lasselle
commented
11 months ago - to train / fine tune a model: quivr is based on ollama , sototrain/fune tune a model, we wilhave to do that with the ollama docs
- what is insteresting with quivr, is tht it brings in lots of tooling ,ui , even chat integrations (like telegram), to build new things. among those, there is a made-easier concept called chains, which is based on the langchain framework (a utility for LLMs) https://brain.quivr.app/docs/Developers/contribution/chains/qa
- quivr uses chains to perform most of the interactions qith ollama
let's experiment https://github.com/StanGirard/quivr