Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

It seems that the understanding of the surrogate process might be incorrect. #3

Open
892446631 opened this issue Aug 29, 2024 · 0 comments

Comments

@892446631
Copy link

The ICCV original paper describes the surrogate process as follows:”The surrogate process utilizes the fact that given the total amount of noise, we can denoise the same amount of noise with different numbers of denoising steps [33].“
However,your code is as follows:
# Set diffusion process for attack and defense
attack_forward = PurificationForward(
clf, diffusion, att_max_timesteps, att_diffusion_steps, args.att_sampling_method, is_imagenet, device)
defense_forward = PurificationForward_mimic(
clf, diffusion, def_max_timesteps, def_diffusion_steps, args.def_sampling_method, is_imagenet, device)
I think that's completely wrong, right?

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
None yet
Projects
None yet
Development

No branches or pull requests

1 participant