supported embedding models

Question

Closed this issue 9 months ago · 2 comments

fastembed only support very limited models, will you change implementation to support all huggingface embedding models?

Answer 1 · 2024-01-14T07:19:56.000Z

I've hardcoded huggingface.py to replace embedding model, not elegant though, better set model name as parameter.

Answer 2 · 2024-01-14T10:08:36.000Z

@Jeru2023 we have recently added a HuggingFaceEncoder to enable support for embedding models on Hugging Face in PR #90. You can use it as follows:

from semantic_router.encoders import HuggingFaceEncoder

encoder = HuggingFaceEncoder(name="sentence-transformers/all-MiniLM-L6-v2")