Ask HN: Help for Choosing an Embedding Model?
I've tried looking and the MTEB leaderboard and tested a few, but I'm curious about your experiences!
I'm looking for a model that would optimize for accuracy (finding relevant results), language support (as many as possible, English only is a no-no) and efficiency so I could potentially run it locally or that there's a cheap API for it.
OpenAI embedding API gets expensive real quick when generating embeddings for 10^5 documents and more.