CaBRNet is an open source library that offers an API to use state-of-the-art prototype-based architectures (also called case-based reasoning models), or easily add a new one.

Currently, CaBRNet supports the following architectures:

• ProtoPNet, as described in Chaofan Chen, Oscar Li, Chaofan Tao, Alina Jade Barnett, Jonathan Su and Cynthia Rudin. This Looks like That: Deep Learning for Interpretable Image Recognition. Proceedings of the 33rd International Conference on Neural Information Processing Systems, page 8930–8941, 2019.
• ProtoTree, as described in Meike Nauta, Ron van Bree and Christin Seifert. Neural Prototype Trees for Interpretable Fine-grained Image Recognition. 2021 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR), pages 14928–14938, 2021.
• ProtoPool, as described in Dawid Rymarczyk, Lukasz Struski, Michal Gorszczak, Koryna Lewandowska, Jacek Tabor and Bartosz Zielinski. Interpretable Image Classification with Differentiable Prototypes Assignment. 2021 European Conference on Computer Vision (ECCV).

Install

• To install the package:

python3 -m pip install --upgrade cabrnet

• To install development related dependencies

python3 -m pip install --upgrade cabrnet[dev]

• To install documentation related dependencies

python3 -m pip install --upgrade cabrnet[doc]

• To install legacy testing related dependencies

python3 -m pip install --upgrade cabrnet[legacy]

Links

• Documentation
• GitHub Repository

Authors

This library is collaboratively maintained by members of CEA-LIST. The current point of contact is Romain Xu-Darme. The following authors contributed in a significant manner to the code base and/or the documentation of the library:

• Romain Xu-Darme (CEA-LIST)
• Aymeric Varasse (CEA-LIST)
• Alban Grastien (CEA-LIST)
• Julien Girard-Satabin (CEA-LIST)

The following authors contributed in a significant manner to the experiments and the publication of trained models:

• Jules Soria (CEA-LIST)
• Alban Grastien (CEA-LIST)
• Romain Xu-Darme (CEA-LIST)

Reference and Citation

Please refer to our work when using CaBRNet:

Romain Xu-Darme, Aymeric Varasse, Alban Grastien, Julien Girard-Satabin, Zakaria Chihani. "CaBRNet, an open-source library for developing and evaluating Case-Based Reasoning Models", xAI-2024 Late-breaking Work, Demos and Doctoral Consortium at the 2nd World Conference on eXplainable Artificial Intelligence.

BibTex citation:

@article{xudarme2024cabrnet,
  title={CaBRNet, an open-source library for developing and evaluating Case-Based Reasoning Models},
  author={Romain Xu-Darme and Aymeric Varasse and Alban Grastien and Julien Girard and Zakaria Chihani},
  booktitle={Proceedings of the xAI-2024 Late-breaking Work, Demos and Doctoral Consortium at the 2nd World Conference on eXplainable Artificial Intelligence},
  year={2024},
}

License

This project is licensed under the LGPL-2.1 license.