Congratulations on your excellent work. I have a question about the robustness experiment. In the paper, you claimed that "For the Diff Attack, we adopt a method similar to DiffPure [24], which uses a pre-trained diffusion model to purify our watermarked images. In our experiments, we set the number of diffusion and denoising steps to 20 and utilize the original DDPM sampler for this attack".
Could you please provide the code for this part? Thank you very much.
Congratulations on your excellent work. I have a question about the robustness experiment. In the paper, you claimed that "For the Diff Attack, we adopt a method similar to DiffPure [24], which uses a pre-trained diffusion model to purify our watermarked images. In our experiments, we set the number of diffusion and denoising steps to 20 and utilize the original DDPM sampler for this attack".
Could you please provide the code for this part? Thank you very much.