torch_sinkhorn

Implementation of Sinkhorn algorithms in Torch.

Installation

Simply activate your conda/Python environment, navigate to torch_sinkhorn root directory and run

pip install -e .

Acknowledgement

This repository is minimally adapted and reimplemented from OTT-JAX in PyTorch.

Citation

If you have found this work useful, please consider citing these references:

@article{cuturi2022optimal,
  title={Optimal Transport Tools (OTT): A JAX Toolbox for all things Wasserstein},
  author={Cuturi, Marco and Meng-Papaxanthos, Laetitia and Tian, Yingtao and Bunne, Charlotte and
          Davis, Geoff and Teboul, Olivier},
  journal={arXiv preprint arXiv:2201.12324},
  year={2022}
}

@misc{le2024torch_sinkhorn,
  author      = {An T. Le},
  title       = {torch_sinkhorn: Implementation of Sinkhorn-like algorithms in Torch},
  url         = {https://github.com/anindex/torch_sinkhorn},
  year        = {2024},
}