Skip to content

mllab-adv-attack/lazy-attack

Repository files navigation

Lazy-attack

Ours implementations

lls-imagenet

ImageNet version code of 'ours' method, paper version. (admm versions to be implemented here)

Use this for ours on ImageNet.

lls-cifar10

Cifar-10 version code of 'ours' method, paper version. (admm versions to be implemented here)

Use this for ours on Cifar-10.

l2-imagenet

L2 version of ours. Not used right now.

Baseline implementations

blackbox-attacks-bandits-priors

Code for baseline black-box attack 'Bandit' (ICLR 2019, https://arxiv.org/abs/1807.07978). Adopted from https://github.com/MadryLab/blackbox-bandits.

Use this for Bandit results in ImageNet.

bandit-imagenet

Code of blackbox-attacks-bandits-priors changed to use Tensorflow and match format of 'ours' code.

nes-imagenet

Code for baseline block-box attack 'NES' (ICML 2018, https://arxiv.org/abs/1804.08598). Adopted from https://github.com/labsix/limited-blackbox-attacks.

Use this for NES results in ImageNet.

Others

gaon_imagenet

Contains works before merging to a git repo. Includes ImageNet PGD attack implementation adopted from https://github.com/MadryLab/cifar10_challenge.

Use this for PGD on ImageNet.

gaon_cifar

Contains works before merging to a git repo. Includes Cifar10 PGD attack implementation from https://github.com/MadryLab/cifar10_challenge.

Use this for PGD on Cifar-10.

Downloading data

Cifar-10: cp from ~/../gaon/lazy-attack/cifar10_data to this folder.

ImageNet: cp from ~/../gaon/lazy-attack/imagenet_data to this folder.

Downloading models

Cifar-10: cp from ~/../gaon/lazy-attack/lls_cifar10/models to your corresponding Cifar folders.

Or follow https://github.com/MadryLab/cifar10_challenge.

ImageNet: cp from ~/../gaon/lazy-attack/lls-imagenet/tools/data to your corresponding ImageNet folders.

Or download from Tensorflow (https://github.com/tensorflow/models/tree/master/research/slim).

About

No description, website, or topics provided.

Resources

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published

Contributors 4

  •  
  •  
  •  
  •  

Languages