Public release of sparse binary hypervectors and associated learners, powered by the Rust-backed kongming-rs-hv package.
pip install kongming-rs-hvSupports Linux, macOS, and Windows on Python 3.10–3.14.
Full documentation is available at yangzh.github.io/hv, including:
- Concepts — hypervectors, models, composites, operators, constant-time near-neighbor search
- API Reference — Python, Go, and Rust side by side
- Python Quick Start — installation, examples, and notebooks
- Notebook Quick Start — cell-by-cell Jupyter walkthrough
- PDF download
The published site is deployed from release tags (
v*) and tracks the latestkongming-rs-hvrelease on PyPI. Themainbranch of this repo is the working head — it may describe APIs or examples that haven't been released yet.
| Platform | Link |
|---|---|
| Colab — tutorial walkthrough | |
| Colab — hypervector storage | |
| Binder |
Runnable scripts under examples/ including:
- Mexican Dollar: analogical reasoning;
- Word Indexer: suffix-queryable word indexing;
- Bulk Storage: storage benchmarks;
- Operators from Scratch: the math underneath the library;
- LISP Interpreter: where every value is a hypervector.
See the examples index for walkthroughs.
Questions, ideas, or feedback? Visit GitHub Discussions for announcements, Q&A, and general conversation. For bugs, please use Issues.
For private inquiries, use the contact form.
Yang, Zhonghao (2023). Cognitive modeling and learning with sparse binary hypervectors. arXiv:2310.18316 [cs.AI]
The Python source code, examples, and documentation in this repository are licensed under the MIT License.
The underlying engine distributed via PyPI (kongming-rs-hv), however, is proprietary.