>>1289411
>denoising a random gaussian distribution
Do you mean taking an image of a real child and the de-noising diffusion models so that the latent is sampled with a unit normal distribution, and then the sample (e.g. image) is generated by iteratively removing noise during the backwards process. Whereas in the diffusion (forward) process, the random Gaussian latent is predicted by iteratively adding Gaussian noise to the original image. So is the implication of your comment that this iterative addition of gaussian noise to the image (in the forward process) eventually leads back to an approximately unit gaussian distribution for the resulting random variable?

Back | Go to board

Board:

/news/

Thread:

AI-generated child pornography threatens to overwhelm reporting system