Skip to content

howardheaton/feasibility_fixed_point_networks

Repository files navigation

Feasibility-based Fixed Point Networks (F-FPNs)

This repo provides the code for the paper Feasibility-based Fixed Point Networks (preprint available here).

Please cite as

@article{heaton2021feasibility,
    title={Feasibility-based Fixed Point Networks},
    author={Heaton, Howard and Fung, Samy Wu and Gibali, Aviv and Yin, Wotao},
    journal={arXiv preprint arXiv:2104.14090},
    year={2021}}

Compatible with python 3.5+ and known to work with pytorch 1.4, torchvision 0.5, and numpy 1.18. Can install with pip install -r requirements.txt.

All codes for this project were run on Google Colab. Jupyter notebook files may be downloaded directly or may be run on Google Colab via the links below.

F-FPN Colab Notebook - Ellipses

F-FPN Colab Notebook - LoDoPab

TV Minimization Colab Notebook

TV Superiorization Colab Notebook

Download Data

About

Feasibility algorithms (projection methods) with data-driven regularization

Resources

License

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published