Vector Similarity library v0.2.0

This is a new version of the vector similarity library.

Headlines

This library introduces a collection of several vector indexing methods. This version includes a flat (brute force) indexing method and HNSW with real-time indexing and deleting operations. All indexing methods are implemented in a single-threaded manner.
Vector comparison metrics introduced in this release are Squared Euclidian distance (squared L2), internal product (`IP), and cosine distance. All metrics support x86 instruction set optimizations: SSE, AVX, AVX512.

Contributors

We'd like to thank all the contributors who worked on this release!

@DvirDukhan, @GuyAv46, @alonre24, @ashtul, @chayim, @gkorland and @rafie

Notes:

This is not a GA version of this library. This library is under active development