-
**Describe the bug**
In PGD (https://github.com/tensorflow/cleverhans/blob/master/cleverhans/attacks/projected_gradient_descent.py), line 82: adv_x = x + eta, which initialize adv_x with a random per…
ywu36 updated
4 years ago
-
at least the main_free.py, utils and validations
could you please share the missing files?
-
hello，sir：
the line:from attacks.fgsm import fgsm
I can't find this file for attacks.fgsm,can you help me?
-
Hi,
From the pictures you showed, We can see that TI-FGSM can achieve good results. In you paper, you said the step size is1.6. But if I set the number of iteration as 10, the result are not good eno…
-
I am building a model with two outputs with custom loss function. Then I am trying to apply an attack on the model using KerasClassifier. but I have the following error (I guess that Kera wrapper coud…
-
Dear teacher, I am a student studying adversarial samples, may I ask that the adversarial examples generated by your library do not have the ability to generalize, what is the reason, and is it my fau…
-
Hi Yinpeng,
Happy new year! I run your code to test the result of TI-DIM, but my result is really bad!
For TI-DIM, I use single model "InceptionV3" and I set num_iter=10, prob=0.7, kernel_legth=…
-
There are only 4 tasks' hyper parameters in this [file](https://github.com/zhuchen03/FreeLB/blob/master/huggingface-transformers/launch/run_glue.sh), would you please release others?
-
I have thought about re-implementing "universal adversarial perturbation" using a trained model and cleverhans. As the previous stage, I searched for a examples that uses trained model to output somet…
-
hi, I generated respectively my adversarial examples using fgsm and deepfool of deepbox, and I saved it in pickle format. Then I tested them using original, the prob was near 0.5(my model is a binary …