search

fra31 / auto-attack

Code relative to "Reliable evaluation of adversarial robustness with an ensemble of diverse parameter-free attacks"
https://arxiv.org/abs/2003.01690
MIT License
639 stars 111 forks source link

fix hard coding #101

Closed s-kumano closed 1 year ago

s-kumano commented 1 year ago

First of all, thank you for your excellent research and code!

I have fixed hard coding for n_fts in autopgd_base.py.

I understand that AutoAttack mainly focuses on images, but I think this code is intended for arbitrary input formats, as indicated by https://github.com/fra31/auto-attack/blob/master/autoattack/autopgd_base.py#L301.

fra31 commented 1 year ago

Hi,

thanks for the contribution! I agree, some of the attacks can be used even with other input formats.