I have been trying to play around with the RAG model for QA for quite some time (few weeks).
Can someone in simple words or codes explain how to use RAG for QA?
I wanted to explore two settings:
- retrieving context passages on the go using RAG Retriever
- using pre retrieved passages to answer the questions.
I have searched everywhere, including the docs, github repo and examples as well as here, but could not find anything simple enough to understand.
BTW I have read the paper and understand the underlying theory quite well.