Find answers from the community

Updated 8 months ago

New

Hello guys, I have a data source from where I embedd the files. But at the moment it embedds everything whenever a new file is added in the folder. How can I only embedd the new file push the new data into current vector store.
And one more thing. I am embedding pdfs and doc files. They have images too. How can embedd the images as well. Is it possible to embedd everything in one single vector store. So I can query about the images as well?
L
A
3 comments
index.insert(document) or index.insert_nodes(nodes) if you want to put into an existing index

You can't embed everything in a single store, since image embeddings typically use a model optimized for images (like clip), and the vectors for images aren't comparable to text
Clip can embed text, but using it to retrieve other texts is not great in terms of quality
Got it. Thanks for the help.
Add a reply
Sign up and join the conversation on Discord