InfoMax

Data Pruning via Information Maximization

• Any question please contact Haoru Tan via personal email (tanhr2014@163.com) or personal Wechat: TanHR_100
• Paper-link: arXiv
• This repository is the official implementation of InfoMax. And it is based on the D2Pruning codebase (We appreciate D2Pruning's contributions to the open-source community and academia.).

Environment requirements.

scipy
transformer 
torch
torchvision
scikit-learn
numpy

Acknowledgements

Thanks to the authors of D2 Pruning: Message Passing for Balancing Diversity and Difficulty in Data Pruning for releasing their code. Much of this codebase has been adapted from their code.

Reference

Any question please contact Haoru Tan via personal email (tanhr2014@163.com) or personal Wechat: TanHR_100

@article{tan2025data,
  title={Data pruning by information maximization},
  author={Tan, Haoru and et.al},
  journal={arXiv preprint arXiv:2506.01701},
  year={2025}
}