All notable changes to this project will be documented in this file.
The format is based on Keep a Changelog, and this project adheres to Semantic Versioning.
- Core TurboQuant implementation with PolarQuant (Stage 1) and QJL (Stage 2)
- Lloyd-Max quantizer for Beta(d/2, d/2) distribution
- Random orthogonal rotation matrix generation
TurboQuantclass with compress, decompress, and inner product estimationTurboQuantKVCacheclass for KV cache simulation- Self-test and demo suite (
python turboquant.py) - Test suite with compression, inner product, and KV cache tests
- Basic usage and KV cache demo examples
- Full documentation and API reference