What I want to do is upload a PDF file, split them into chunks, use OpenAI embeddings and then store in Vectara vector store.
I know Vectara has its own embeddings and chunking method, so I do not have to the tasks. But I just want to compare some results. Is this possible?
I remeber reading a document on LangChain JS where it says even if I set to use OpenAI embeddings, Vectara will ignore any other embeddings other than its own.