Self-Sustaining Representation Expansion for Non-Exemplar Class-Incremental Learning

This is the implementation of the paper "Self-Sustaining Representation Expansion forNon-Exemplar Class-Incremental Learning" (accepted to CVPR2022).

For more information, check out the paper on [arXiv].

Requirements

• Python 3.8
• PyTorch 1.8.1 (>1.1.0)
• cuda 11.2

Preparing Datasets

Download following datasets:

1. CIFAR-100

2. Tiny-ImageNet

3. ImageNet

Locate the above three datasets under ./data directory.

Incremental Training.

1. Download pretrained models to the 'pre' folder.

Pretrained models are available on our [Google Drive].

2. Training

sh train_cvpr.sh 

Base Training

Coming soon.

Requirements

We thank the following repos providing helpful components/functions in our work.

• PASS
• RepVGG

BibTeX

If you use this code for your research, please consider citing:

@article{zhu2022self,
  title={Self-Sustaining Representation Expansion for Non-Exemplar Class-Incremental Learning},
  author={Zhu, Kai and Zhai, Wei and Cao, Yang and Luo, Jiebo and Zha, Zheng-Jun},
  journal={arXiv preprint arXiv:2203.06359},
  year={2022}
}