Problem getting it to use my file(s)

[ModernMagus]

Hi! I have just started using LightRAG, and everything (including document upload and graph generation) seems to be working. When testing my document on the retrieval page, however, I cannot seem to get the LLM to use my files (it typically pull a lot of knowledge from it's general training). I suspect this has to do either with the embedding process or the settings I have for the retrieval page (see pic 1). Any ideas?

[windshetty]

I have a similar issue too. I uploaded a PDF, pipeline ran to completion, Knowledge Graph is created. But when you query using "Retrieval" tab, the responses seem to be from the knowledge from its general training. Tried all the "Query Modes" and different "User Prompts". I get slightly different responses, but not from the PDF I uploaded. I too am curious to know if this is an embedding process or the settings or something else issue.

[apiipc]

i see in this part in query text "response_type": "string", not sure if i can change it to "multible"