Requirements

• Python 3
• PyTorch (version == 1.7.1)
• yaml
• higher(version == 0.2.1)

Installation

shell
git clone https://github.com/MyCurveNet/CurveNet.git
cd CurveNet
cd torchlight
python setup.py install
cd ..

Training

Train our model

• In the probing stage, obtain the

python main.py recognition_cycle -c config/cycle/cifar10_unif/train_c0.4_imb0.05.yaml

• In the allocating stage

python main.py recognition_curve -c config/aug/cifar10/train_c0.4_imb0.05.yaml

Visualization

After training, you can visualize the weight.

python python tools/gen_fig1_weight_cls.py work_dir/recognition_cycle/cifar10_unif/cor0.4_imb0.05 work_dir/recognition_aug_cycle/cifar10_unif/cor0.4_imb0.05

Citation

If CurveNet is useful for your research, please consider citing:

@inproceedings{jiang2022delving,
  title={Delving into Sample Loss Curve to Embrace Noisy and Imbalanced Data},
  author={Jiang, Shenwang and Li, Jianan and Wang, Ying and Huang, Bo and Zhang, Zhang and Xu, Tingfa},
  booktitle={Proceedings of the AAAI Conference on Artificial Intelligence},
  year={2022}
}

Reference

• The code is based on Meta-weight-net.
• ResNet-32 is from MLC