FAISS
FAISS: A library for vector similarity and clustering that implements efficient indexes for large collections. It offers flat, IVF and PQ indexes, with CPU or GPU execution depending on the build. It is not a complete database: persistence, metadata and concurrency belong to the application or an external system.
In italiano: FAISSLegge I · Embedding, feature e spazi vettorialiIt offers flat, IVF and PQ indexes, with CPU or GPU execution depending on the build. It is not a complete database: persistence, metadata and concurrency belong to the application or an external system.
Examples
- FlatIP index for cosine similarity
- IVF-PQ search
- Local retrieval prototype