Fast sketching algorithms for computing Tensor Train decompositions of a variety of tensorial data.
This software implements the algorithms discussed in the preprint arXiv:2208.02600.
The tt-sketch package is available on PyPI, and can be installed
using pip by running
pip install tt-sketchAlternatively you can install it by first cloning this repository:
git clone [email protected]:RikVoorhaar/tt-sketch.git
cd tt-sketch
pip install .All numerical experiments in the preprint can be reproduced using the scripts starting with plot_ in the scripts directory. All experiments were produced using version 1.1 of this software. The dependencies for running these scripts, as well as running the tests or building the documentation, are listed in environment.yml.
The documentation for this project lives here: tt-sketch.readthedocs.io.
All code for this project is written by Rik Voorhaar, in a joint project with Daniel Kressner and Bart Vandereycken. This work was supported by the Swiss National Science Foundation under research project 192363.
This software is free to use and edit. When using this software for academic purposes, please cite the following preprint:
@article{
title = {Streaming tensor train approximation},
journal = {arXiv:2208.02600},
author = {Kressner, Daniel and Vandereycken, Bart and Voorhaar, Rik},
doi = {10.48550/arXiv.2208.02600},
year = {2022},
}
All contributions or suggestions are welcome. Feel free to open an issue with suggestions, or submit a pull request.