search

gongzhitaao / adversarial-classifier

It turns out that adversarial and clean data are not twins, not at all.
https://arxiv.org/abs/1704.04960
MIT License
20 stars 6 forks source link

May I use it? #2

Closed EmotionalXX closed 4 years ago

EmotionalXX commented 4 years ago

Hello,sir. May I need to train the binary classfier? Can not I use it to detect adversarial samples？

gongzhitaao commented 4 years ago

Sure. Feel feel to grab snippets of the code for your project. It is MIT-licence, so you are free to do anything :smile: The only thing I'm worried about is the version. The classifier was written with tf1.4, which is already deprecated. You might need to migrate some of the code to tf2. Good luck!

EmotionalXX commented 4 years ago

Oh,that is good.Thank you!And is there code for training the binary classfier？

gongzhitaao commented 4 years ago

adversarial-classifier/src/table_1_mnist.py But it is just standard binary classifier, you could use whatever framework you like, e.g., tf2, pytorch, jax, etc.