all-MiniLM-L6-v2

all-MiniLM-L6-v2 is a compact, efficient sentence-transformers model developed by the Microsoft Research team and distributed via the Hugging Face platform. It is part of the MiniLM family, which focuses on distilling large language models into smaller, high-performance variants suitable for embedding generation and semantic search tasks. The model maps sentences and paragraphs to a 384-dimensional dense vector space, enabling applications such as clustering, semantic similarity, and information retrieval. It is designed for low-latency, high-throughput environments and is widely used in production systems due to its balance of speed and accuracy.