doudou

Also curious about what if you actually have some non-linearity in hypernet

I think the hypernets from NovelAI are probably trained as in the code, so to also support existing NovelAI pipeline one should either keep the original implementation or perform merging...

Hi, I tested the script on my local Ubuntu system, got the same output, but the images were correctly generated. If it outputs a bunch of zeros followed by server...

Thanks for the contribution. As for the BIM evaluation, how many iterative steps did you run to optimize the adversarial images? I found the hyper-parameters missing in your paper.

> We use the foolbox to achieve the BIM attack method. And the iterations=10. You can refer to the URL(https://foolbox.readthedocs.io/en/latest/modules/attacks/gradient.html#foolbox.attacks.BIM) Thanks for the reply. I'm also wondering did you perform...

> When we use the 10 steps to attack the welled-trained classifier, it can achieve a percent attack success rate. In my opinion, the 10 steps can get the perceptible...