Code for our paper "On the Adversarial Robustness of Visual Transformers"

Paper link: https://arxiv.org/abs/2103.15670

Install dependencies:

pip install -r requirements.txt

To test adversarial robustness under white-box attack

python white_box_test.py --data_dir $DATA_DIR --mode foolbox --model vit_small_patch16_224

• The --mode flag decides the evaluation approach:
  • --mode foolbox applies a single attack (default: LinfPGD) for evalution.
  • --mode auto applies AutoAttack for evaluation.
  • --mode foolbox-filter applies the frequency-based attack for evaluation.
  • --mode evaluate evaluates the clean accuracy.
  • --mode count counts the number of parameters.

To test the transferability of adversarial examples generated by different models

python black_box_test.py --data_dir $DATA_DIR

Go to the training code:

cd training

Install dependencies:

pip install -r requirements.txt

Run:

python train.py --dir {OUTPUT_DIR} --model vit_base_patch16_224_in21k --method {pgd|trades}

You may set `--accum-steps {N}' for gradient accumulation in case that GPU memory is not enough.