The TMU repository is a collection of Tsetlin Machine implementations, namely:
- Tsetlin Machine (https://arxiv.org/abs/1804.01508)
- Coalesced Tsetlin Machine (https://arxiv.org/abs/2108.07594)
- Convolutional Tsetlin Machine (https://arxiv.org/abs/1905.09688)
- Regression Tsetlin Machine (https://royalsocietypublishing.org/doi/full/10.1098/rsta.2019.0165)
- Weighted Tsetlin Machine (https://ieeexplore.ieee.org/document/9316190)
Further, we implement many TM features, including:
- Support for continuous features (https://arxiv.org/abs/1905.04199)
- Drop clause (https://arxiv.org/abs/2105.14506)
- Type III Feedback (to be published)
- Focused negative sampling (https://ieeexplore.ieee.org/document/9923859)
- Multi-task classifier (to be published)
- Autoencoder (https://arxiv.org/abs/2301.00709)
- Literal budget (to be published)
- Incremental clause evaluation (to be published)
- One-vs-one multi-class classifier (to be published).
TMU is written in Python with wrappers for C and CUDA-based clause evaluation and updating.
To install on windows, you will need the MSVC build tools, (https://visualstudio.microsoft.com/visual-cpp-build-tools/
)[found here]. When prompted, select the Workloads → Desktop development with C++ package,
which is roughly 6-7GB of size, install it and you should be able to compile the cffi modules.
pip install git+https://github.com/cair/tmu.git