Find answers from the community

Updated last year

Postgress

At a glance

I am currently having Postgres Vector store, I can see that llama_index creates these vectors as tables, but langchain manages the metadata and stores details regarding the collections.
1.Is there any way we can have cross compatibility irrespective of how the vectorstores are created?

Is there anyway to method to keep a track of which embedding model i sused to create a certain collection and the embedding size? Having this info will make the re-use of data easier for large audience

@Logan M

1 comment

LLogan M

Tbh I'm a little confused why you need both lamgchain and llama index interacting with the vector store. In any case, I don't think it's possible, there needs to be a PR to add a text field and metadata field to the pgvector store class

For keeping track of embeddings, I would just put the model name in the metadata no?

Add a reply