Paper

• Training on CIFAR-100 dataset:

$ sh sv.sh

• We recommend you to use python3.6 as the base python environment. And other dependency libraries are listed in requirements.txt.

$ pip install -r requirements.txt

• Please download the MAE buffer used for training. After download it, please put it in the root folder.

• We use the codebase from DyTox to align some data processing implementations and evaluation metrics. Thanks for their wonderful work!

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

@inproceedings{zhai2023masked,
  title={Masked autoencoders are efficient class incremental learners},
  author={Zhai, Jiang-Tian and Liu, Xialei and Bagdanov, Andrew D and Li, Ke and Cheng, Ming-Ming},
  booktitle={Proceedings of the IEEE/CVF International Conference on Computer Vision},
  pages={19104--19113},
  year={2023}
}