This repository contains PyTorch evaluation code, training code and pretrained models for ReSSL.

For details see ReSSL: Relational Self-Supervised Learning with Weak Augmentation by Mingkai Zheng, Shan You, Fei Wang, Chen Qian, Changshui Zhang, Xiaogang Wang and Chang Xu

This repository is based on ImageNet dataset, We also provide the training code and pretrained model for cifar10/100, STL10 and TinyImageNet, please download it from this link.

To run the code, you probably need to change the Dataset setting (dataset/imagenet.py), and Pytorch DDP setting (util/dist_init.py) for your own server enviroments.

The distribued training of this code is base on slurm enviroments, we have provide the training scrips under the script folder.

We also provide the pretrained model for ResNet50 (single crop and 5 crops)

If you want to test the pretained model, please download the weights from the link above, and move it to the checkpoints folder (create one if you don't have .checkpoints/ directory). The evaluation scripts also has been provided in script/train.sh

If you find that ReSSL interesting and help your research, please consider citing it:

@inproceedings{
      zheng2021ressl,
      title={Re{SSL}: Relational Self-Supervised Learning with Weak Augmentation},
      author={Mingkai Zheng and Shan You and Fei Wang and Chen Qian and Changshui Zhang and Xiaogang Wang and Chang Xu},
      booktitle={Thirty-Fifth Conference on Neural Information Processing Systems},
      year={2021},
      url={https://openreview.net/forum?id=ErivP29kYnx}
}