Repository navigation
v0.1.5
π Release: v0.1.5 β Initial Public Release
π¦ FastEmbed VectorStore
A high-performance, Rust-based in-memory vector store with seamless Python integration via FastEmbed.
π [π View on PyPI](https://pypi.org/project/fastembed-vectorstore/0.1.5)
π [π Source on GitHub](https://github.com/sauravniraula/fastembed_vectorstore/commits/v0.1.5)
π Highlights
- Rust-based core with blazing-fast performance and minimal memory overhead
- Python bindings via PyO3 for easy integration with your Python applications
- Supports 30+ Embedding Models including BGE, GTE, Nomic, Multilingual E5, and more
- In-memory vector store with JSON-based persistence and restore
- Cosine similarity search with ranked results
π§ Features
- π§ Embedding Model Support: Choose from a wide range of pre-trained models including quantized variants
- πΎ Persistence: Save and load vector stores to/from disk (
.json) - π Search: Cosine similarity-based nearest neighbor search
- βοΈ Simple API:
π License
Licensed under Apache 2.0
Author: Saurav Niraula (π§ [developmentsaurav@gmail.com](mailto:developmentsaurav@gmail.com))