Learning from Noisy Data with Robust Representation Learning (ICCV 2021)
This is the PyTorch implementation of the ICCV paper [link].
Requirements:
- PyTorch = 1.4
- pip install tensorboard_logger torchnet faiss-gpu
Configuration:
Hyper-parameters and model configurations are located in ./config
Dataset:
In order to run experiments, please download the corresponding dataset and place it at the location specified in the config file.
Execution:
python main.py --exp [config_file]
For example, run the following command to reproduce the paper's result on CIFAR-10:
- 50% symmetric noise:
python main.py --exp cifar10_sym
- 40% asymmetric noise:
python main.py --exp cifar10_asym
Citation
If you find this code to be useful for your research, please consider citing.
@inproceedings{ALBEF,
title={Learning from Noisy Data with Robust Representation Learning},
author={Junnan Li and Caiming Xiong and Steven Hoi},
year={2021},
booktitle = {{ICCV}},
}