NMSLIB (Non-Metric Space Library)

NMSLIB (Non-Metric Space Library) is an efficient cross-platform similarity search library with Python bindings, designed for approximate nearest neighbor search in high-dimensional and non-metric spaces. It is widely used for vector similarity tasks, often in conjunction with embeddings from libraries like Gensim. The current version is 2.1.2, and the library receives updates focused on performance, new features, and broader platform/Python support.