-
Hi,
I use some attacking method but the val acc of data is still high. I feel confused. I compare my codes with the tutorial examples. But I cannot find any reason. I use cifar10.
```
#mode is trai…
-
I run the code in 'cleverhans/examples/madry_lab_challenges/cifar10/attack_model.py' with default parameter settings to attack target model with 'models/adv_trained' checkpoints. And I get the results…
-
FGSM attack: When changing the max_epsilon from 1 to 0.01 is gives us this error. Is there a limit for how small the max_epsilon can be?
![whatsapp image 2018-12-04 at 16 41 35](https://user-images…
-
When I try to run "python train.py --env Breakout --noisy --save-dir ./data/Breakout/ --attack fgsm --num-steps 20 --attack-prob 1.0" I get the following error,
Making new env: BreakoutNoFrameskip-…
-
I am working to recreate some of the results from your paper, specifically some cifar10 transfer results. I have noticed something in the tables that doesnt seem intuitive so I was wondering if you co…
-
Hi,
In your Adversarial examples - targeted FGSM.ipynb, there is some problem. It does not becomes the target class. Have you also noticed this?
-
`imagenet_tutorial_fgsm_pytorch.py[line:52] INFO CUDA Available: True cuda
pytorch.py[line:63] INFO Finish PytorchModel init
(1, 3, 224, 224)
Traceback (most recent call last):
File "imagenet_tu…
-
I want a code example to how to generate train data from clever hans' adversarial attacks.
`adv_x = fgsm.generate_np(X_test, **fgsm_params)`
This generates adversarial x data but how can I get y…
-
After training, we are left with ~80 models saved at each 1k iteration. What is your rule for selecting the "best" model to then keep and do further evaluations with? I am especially wondering because…
-
**Describe the bug**
I implement a simple resnet9 and want to test cw2 result. All is good when I set max_iter=10. But when I set max_iter=1000 that is same value as cw2 original version. The index e…