Hello,
I think including a few things in the Vector Stores documentation would be helpful.
- Limits - Provide an explicit mention that uploading files to the vector store counts towards the batch TPD limit (of the embedding model?).
- Embedding model—I assume that the vector store uses
text-embedding-3-large
, as mentioned in thefile_search
How It Works section. However, this information seems more for inference/retrieval than for storing the vectors themselves.
I think being a bit more explicit with these would help.
Thanks:)