Embeddings for Chroma for LangChain for document search

#26
by LeoArtaza - opened

I guess this is not specific to this model but I was wondering what should I use in this function

Chroma.from_documents(pages, embedding=...)

to get embeddings for the document to be used with this model. For example if we were using GPT-3.5 we would pass OpenAIEmbeddings().

I tried HuggingFaceHubEmbeddings(repo_id="tiiuae/falcon-7b")

and got

Currently only 'sentence-transformers' embedding models are supported. Got invalid 'repo_id' tiiuae/falcon-7b-instruct. (type=value_error)

which I'm not sure what it means.

LeoArtaza changed discussion status to closed
LeoArtaza changed discussion status to open

I think you can use any embeddings, maybe I was getting confused with tokenizing.

LeoArtaza changed discussion status to closed
LeoArtaza changed discussion status to open

Sign up or log in to comment