Code For CVPR2022 paper "Not Just Selection, but Exploration: Online Class-Incremental Continual Learning via Dual View Consistency"
requirements.txt
- CIFAR10 & CIFAR100 will be downloaded during the first run. (datasets/cifar10;/datasets/cifar100)
- Mini-ImageNet: Download from https://www.kaggle.com/whitemoon/miniimagenet/download, and place it in datasets/mini_imagenet/
python general_main.py --data cifar100 --cl_type nc --agent ER_DVC --retrieve MGI --update random --mem_size 1000 --dl_weight 4.0 python general_main.py --data cifar10 --cl_type nc --agent ER_DVC --retrieve MGI --update random --mem_size 200 --dl_weight 2.0 --num_tasks 5python general_main.py --data mini_imagenet --cl_type nc --agent ER_DVC --retrieve MGI --update random --mem_size 1000 --dl_weight 4.0If our code or models help your work, please cite our paper:
@InProceedings{Gu_2022_CVPR,
author = {Gu, Yanan and Yang, Xu and Wei, Kun and Deng, Cheng},
title = {Not Just Selection, but Exploration: Online Class-Incremental Continual Learning via Dual View Consistency},
booktitle = {Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR)},
month = {June},
year = {2022},
pages = {7442-7451}
}