Embedding and Index with Vertex AI

[nacartwright]

I'm attempting to make a Q&A bot with Vertex (PaLM + Matching Engine). I have the embeddings down but I'm confused on the implementation of matching engine.

I have a batch of embeddings that I want to add to the Matching Engine index. I'm a little confused on the add_texts implementation. Is meant to pass a single vector at a time? It looks like the add_texts is doing embedding as well?

Or is add_texts supposed to your chunks?

I have this working but using Pinecone for the vector storage. Just confused on the Matching Engine portion.