Similarity Search - Returns the same chunk

[ss323]

I am creating a pdf summarizer, for each query, first I search for the relevant chunks of data whose embedding is already stored in ChromaDB.

My problem is that I am getting the same chunk four times rather than four different chunks of data which are most related to the query.

This happens even when my query prompt is "summarize the document/ case".

Is there a way where I can get four different chunks of data while using similarity search?

Let me know if any other info is required to help with this question.

[mohit19906]

I am just a newbie so forgive me if I don't make any sense. I would suggest you to try FAISS once and see if you get better results.

[ss323]

I solved it by saving the DB to machine rather than using it on the go from the memory. FAISS gives the same issue if run directly without saving to the disc.
Thanks for the help.

[ss323]

Solved by saving the Database to local disc.