similarity Search Issue

[mohitraj]

We are using Chroma for storing the records in vector form. When searching the query, the return documents do not give accurate results.
c1 = Chroma('langchain', embedding, persist_directory)
qa = ChatVectorDBChain(vectorstore=c1, combine_docs_chain=doc_chain, question_generator=question_generator,top_k_docs_for_context=12, return_source_documents=True)

What is the solution to get accurate results?

[ghost]

can tuke chucksize and overlaping paramter when you splitting the text and see will it improve acc. In my case it actually work

[mohitraj]

What is your chunk size and overlapping parameter?

[khimaros]

for me, when using LlamaCppEmbedding, chunk and overlap was not helpful. the results returned are almost in reverse order of what they should be with the best results almost dead last.

[dosubot]

Hi, @mohitraj! I'm Dosu, and I'm here to help the LangChain team manage their backlog. I wanted to let you know that we are marking this issue as stale.

From what I understand, the issue you raised is about the return documents from a similarity search using Chroma not giving accurate results. In the comments, there were suggestions to try different chunk sizes and overlapping parameters, but it seems that these parameters did not help in improving the accuracy of the search. Unfortunately, there doesn't appear to be a resolution to this issue at the moment.

Before we close this issue, we wanted to check with you if it is still relevant to the latest version of the LangChain repository. If it is, please let us know by commenting on the issue. Otherwise, feel free to close the issue yourself, or it will be automatically closed in 7 days.

Thank you for your understanding and contribution to the LangChain project. If you have any further questions or concerns, please don't hesitate to reach out.