-
I just cloned this repo and try to run codes with provided instructions. (the code is not modified.)
Environment: `cuda 11.3, python 3.9.6, pytorch 1.9.0, torchvision 0.10.0`, installed via minicon…
-
For Example:
- Test FGSM
- Test C & W
Just test some attacks
Test if the implementation works and document how it works.
https://foolbox.readthedocs.io/en/stable/index.html
https://foolbox.r…
-
您好，我在认真阅读了您的论文后尝试训练一个在MNIST数据集上、替代模型是中型网络、目标模型是大型网络的替代模型（您的代码中给出的）。超参数的选择我均采用了代码中给定的值，但是训练到280epoch后acc和att便很难提升。这时att的终端输出值只有27%左右。可否请问一下，您在训练的过程中，优化器的参数是否进行了调整呢，如果调整的话进行了哪些调整呢？
-
## 🐛 Bug
Even with `@torch.enable_grad `, gradients aren't enabled during validation
### To Reproduce
```py
import os
import numpy as np
import torch
from torch.utils.data import DataLo…
-
In short, I use squeezenet to train a model in mnist dataset which acc is 99.71.
Then, I use `torchattacks.FGSM(model)` to generate attack data, acc is 97.38, is it normally ?
-
Why your code about PI-FGSM is not the same as the theory in the pape
-
Hi, @fra31 , thanks for releasing the code for evaluating various defense method.
However, I am curious about the defense in your released table (Rice et al. 2020 overfitting in adversarial robust d…
-
my dataloader:
```py
elif name == "cifar10":
transform = torchvision.transforms.Compose([
torchvision.transforms.RandomCrop(32, padding=4),
torchvision.transforms.RandomHori…
-
#44 is EPIC
- Read the paper thoroughly
- Document all necessary variables
- Document the architecture
- Document all functions
- Create the logic in "pseudo code"
-
Hi,
There are four padding attacks implemented in the library that add in bytes: c_padding_evasion, c_kreuk_evasion, c_fast_gradient_sign_evasion, c_suciu_evasion. I know that c_kreuk_evasion uses…