3 ms·
Semantic search with embeddings: index anything
- wrnr 5y agoReminds me a bit of Weaviate: https://www.semi.technology/developers/weaviate/current/ https://www.semi.technology/developers/weaviate/current/
- gk1 5y agoThis is from 2020. These days you can do most of this with a few API calls using Pinecone (https://www.pinecone.io https://www.pinecone.io). Take your text/image/audio/timeseries/whatever embeddings and upsert(), then 2s later take a query embedding to search with and query().