This repository contains an implementation on pytorch of the unbalaned Gromov-Wasserstein divergence presented in the paper The Unbalanced Gromov Wasserstein Distance: Conic Formulation and Relaxation. It allows to compare weighted point clouds equipped with a cost matrix.
If you use this work for your research, please cite the paper:
@article{sejourne2020unbalanced,
title={The Unbalanced Gromov Wasserstein Distance: Conic Formulation and Relaxation},
author={S{\'e}journ{\'e}, Thibault and Vialard, Fran{\c{c}}ois-Xavier and Peyr{\'e}, Gabriel},
journal={arXiv preprint arXiv:2009.04266},
year={2020}
}