Skip to content

Latest commit

 

History

History
43 lines (31 loc) · 1.39 KB

File metadata and controls

43 lines (31 loc) · 1.39 KB

tinyvector

Category: Vector Database Engines
Website/Source: https://github.com/0hq/tinyvector

Description

tinyvector is a minimal vector database / approximate nearest neighbor (ANN) engine built with SQLite and PyTorch. It is focused on simplicity and compact implementation for educational use and small-scale similarity search applications. The project is currently in pre-release and still under active development.

Features

  • Minimal vector database / ANN engine
  • Designed for educational and small-scale similarity search use cases
  • Built on top of SQLite for storage
  • Uses PyTorch in its implementation
  • Provides a server that can be run manually
  • Test suite available for validation and development
  • Multi-language ecosystem:
    • Python implementation: tinyvector (this repository)
    • Rust implementation: tinyvector-rs

Development status

  • Pre-release, not yet production-ready
  • Aiming for production readiness in a future release (timeline indicated as late July in the repository, subject to change)

Usage

Basic commands from the repository:

# Run the server manually
pip install -r requirements.txt
python -m server

# Run tests
pip install pytest pytest-mock
pytest

Pricing

  • Open-source project (license present in repository).
    No paid pricing plans are listed in the available content.