3 ms·
This is a rehash (pardon me) of this post from 2021: https://www.search.io/blog/vectors-versus-hashes https://www.search.io/blog/vectors-versus-hashes The dema
by gk1 4y ago
This is a rehash (pardon me) of this post from 2021: https://www.search.io/blog/vectors-versus-hashes https://www.search.io/blog/vectors-versus-hashes
The demand for vector embedding models (like those released by OpenAI, Cohere, HuggingFace, etc) and vector databases (like https://pinecone.io https://pinecone.io -- disclosure: I work there) has only grown since then. The market has decided that vectors are not, in fact, over.
- packetlost 4y agoPineCone seems interesting. Is the storage backend open source? I've been working on a persistent hashmap database that's somewhat similar (albeit not done) that should have less RAM requirements than bitcask (ie. larger than RAM keysets)
- gk1 4y agoAlthough we may open-source parts in the future, currently no part of Pinecone is open-sourced. Instead, there are several proprietary index types available, packaged along with hardware/compute resources into what we call “pods.”
- psteitz 4y agoHave a look at Milvus (BSD license) and Weviate (Apache 2)