This may be the simplest implement of DDPM. You can directly run Main.py to train the UNet on CIFAR-10 dataset and see the amazing process of denoising.
MIT License
1.48k
stars
156
forks
source link
Where is the distribution of what is learned through ddpm #28
I know that using the network can predict the noise, but I still don't know where I learned the distribution of the original image.