bookman233/TADAT

★ 1Forks 0PythonGitHub ↗Compare

README

Rethinking Frequency Modeling: Tail-Aware Dynamic Adversarial Training for Long-Tailed Robustness

Dependencies

  1. Install PyTorch.
  2. Install required python packages using the following command:
    pip install -r requirements.txt

Datasets

  • Training and Evaluation data:
    • CIFAR10, CIFAR100, and Tiny ImageNet

Training

Run the following command:

CUDA_VISIBLE_DEVICES=0 python TADAT.py --eta 4.0 --lamda 0.4

Evaluations

Run the following command:

CUDA_VISIBLE_DEVICES=0 python eval_AA.py --arch '' --dataset '' --model_path '' --model_name '' --save_file_name ''

Acknowledgements

The code refers to BSL-AT.

We thank the authors for sharing sincerely.

License

This project is released under the MIT License.

Contributors

boookman

Issues