This repository is the official Pytorch implementation of "When Optimizing f-Divergence is Robust with Label noise" accepted by ICLR2021.
Supported OS: Windows, Linux, Mac OS X; Python: 3.6/3.7;
Deep Learning Library: PyTorch (GPU required)
Required Packages: Numpy, Pandas, random, sklearn, tqdm, csv, torch (Keras is required if you want to estimate the noise transition matrix).
Details of reproducing our experiment results on MNIST, Fashion MNIST, CIFAR-10, CIFAR-100, Clothing 1M are mentioned in the README.md
file in each folder.
We repeat Figure 1 in our paper here:
If you use our code, please cite the following paper:
@article{wei2020optimizing,
title={When Optimizing $ f $-divergence is Robust with Label Noise},
author={Wei, Jiaheng and Liu, Yang},
journal={arXiv preprint arXiv:2011.03687},
year={2020}
}
📋 The code about estimating the noise transition matrix is based on https://github.com/giorgiop/loss-correction