PyTorch implementation of TurboQuant. Near-optimal vector quantization for KV cache compression and vector search. 3-bit with zero accuracy loss.
Stars
10
Forks
0
Watchers
10
Open Issues
1
Overall repository health assessment
No package.json found
This might not be a Node.js project
33
commits
6
commits
Release v0.3.0: outlier routing, adaptive allocation, model compat (#8)
d67770bView on GitHubfeat: add HuggingFace demo, benchmarks, visual cards, fix PyPI README
a60a77dView on GitHubdocs: update CLAUDE.md with versioning and release flow
1028c96View on GitHubfix: add fetch-depth 0 to all checkouts for setuptools-scm
0d1c1a1View on GitHubbuild: use setuptools-scm for automatic versioning from git tags
4295bc1View on GitHubci: replace pip with uv for faster dependency installs
3131d9bView on GitHub