Source Code for PCL

Prototypical Classifier for Robust Class-Imbalanced Learning

Authors: Tong Wei, Jiang-Xin Shi, Yu-Feng Li, and Min-Ling Zhang

[arXiv] [slides] [bibtex]

This paper won the Best Paper Award at PAKDD 2022!

Installation

Requirements

Python 3.6
torchvision 0.9.0
Pytorch 1.8.0

Dataset Preparation

CIFAR-10 & CIFAR-100

Training

python main.py --exp cifar10_imb

python main.py --exp cifar100_imb

The saved folder (including logs and checkpoints) is organized as follows.

PCL
├── experiment
│   ├── cifar10_exp_0.1_imb_0.2
│   │   ├── model_best.pth.tar
│   │   └── logs
│   │       └── modelname.txt
│   ...

Citation

Please consider citing PCL in your publications if it helps your research. :)

@inproceedings{wei2022pcl,
    title={Prototypical Classifier for Robust Class-Imbalanced Learning},
    author={Tong Wei, Jiang-Xin Shi, Yu-Feng Li, and Min-Ling Zhang},
    booktitle={Pacific-Asia Conference on Knowledge Discovery and Data Mining},
    year={2022},
}

Contact

If you have any questions about our work, feel free to contact us through email (Tong Wei: weit@lamda.nju.edu.cn) or Github issues.

Stomach-ache / PCL