Find answers from the community

Updated 4 months ago

here is a snapshot ...i have over 20

here is a snapshot ...i have over 20 files in that directory
Attachment
image.png
L
h
9 comments
Some of the default readers will download models for image captioning

Specifically
  • PptxReader
  • VideoAudioReader
If you don't want to use those readers, either remove those file types from the directory, exclude those file types in the in simple directory reader constructor, or pass in another reader to override the defaults in the simple directory reader constructor (would likely be a custom reader, or something from http://llamahub.ai)
Attachment
image.png
ok i just realised that some of the document have images in them hence it is trying to use OCR
Welldone for ansa
please are you quite conversant with reranking?
yea, what question about reranking do you have?
before we get there what type of embedding do you think i should use?
i need opensource
personally I think the BGE models are pretty good, mixedbread as well
@Logan M are u available for a call i wand to show u my code?
Add a reply
Sign up and join the conversation on Discord