Mitigating Memorization of Noisy Labels by Clipping the Model Prediction

ICML 2023: This repository is the official implementation of LogitClip.

Requirements

To install requirements:

pip install -r requirements.txt

Training

To train the model(s) in the paper, run this command:

python train.py cifar10 --alg clip -m resnet34 --noise_type symmetric --noise_rate 0.5 --exp_name test --gpu 0 --temp 1.5

Citation

If you find this useful in your research, please consider citing:

@inproceedings{wei2023logitclip,
  title={Mitigating Memorization of Noisy Labels by Clipping the Model Prediction},
  author={Wei, Hongxin and Zhuang, Huiping and Xie, Renchunzi and Feng, Lei and Niu, Gang and An, Bo and Li, Yixuan},
  booktitle={International Conference on Machine Learning},
  year={2023},
  organization={PMLR}
}