[Feature]: SentenceTransformer replacement solution

[Aisuko]

Contact Details(optional)

No response

What feature are you requesting?

Currently, we use sentence transformers as encoder. And it isn't efficient on CPU(of course it is). So, let's make a solution based llama.cpp to solve this.

For embeddings see

• https://platform.openai.com/docs/api-reference/embeddings

For vector-db

We are using v2.3.x. So, we can remove sentence transformer

• https://milvus.io/blog/2022-08-08-How-to-use-string-data-to-empower-your-similarity-search-applications.md#Insert-data
• https://github.com/milvus-io/pymilvus/blob/master/examples/example_str.py

For cosine similarity score

• https://github.com/ggerganov/llama.cpp/pull/6193/files