Denoising Diffusion Probabilistic Models

This is a pytorch implementation of DDPM. The original paper is here https://arxiv.org/abs/2006.11239 .

If you want to know more about the framwork of DDPM, these two blogs may help you:

• https://zhuanlan.zhihu.com/p/563661713
• https://lilianweng.github.io/posts/2021-07-11-diffusion-models/#:~:text=Diffusion%20models%20are%20inspired%20by,data%20samples%20from%20the%20noise

how to use

Almost all the parameters that can be modified are listed in the config.yml file. You can modify the relevant parameters as needed, and then run the train.py file to start training.

After training, run the generate.py file to generate the results. These are the parameters of generate.py :

• -cp : the path of checkpoint.
• --device : device used. 'cuda' (default) or 'cpu'.
• -bs : how many images to generate at once. Default 16.
• --result_only : whether to output only the generated results. Default False.
• --interval : extract an image every how many steps. Only valid without the result_only parameter. Default 50.
• --nrow : how many images are displayed in a row. Only valid with the result_only parameter. Default 4.
• --show : whether to display the result image. Default False.
• -sp : save path of the result image. Default None.
• --to_grayscale : convert images to grayscale. Default False.

Some generated images

• MNIST，click to download checkpoint

python generate.py -cp "checkpoint/mnist.pth" -bs 16 --interval 10 --show -sp "data/result/mnist_sampler.png"

python generate.py -cp "checkpoint/mnist.pth" -bs 256 --show -sp "data/result/mnist_result.png" --nrow 16 --result_only

• CIFAR10, click to download checkpoint

python generate.py -cp "checkpoint/cifar10.pth" -bs 16 --interval 50 --show -sp "data/result/cifar10_sampler.png"

python generate.py -cp "checkpoint/cifar10.pth" -bs 256 --show -sp "data/result/cifar10_result.png" --nrow 16 --result_only

• celeba_hq , click to download checkpoint

python generate.py -cp "checkpoint/celeba_hq.pth" --interval 30 -bs 16 -sp "data/result/celeba_hq_sampler.png" --show

python generate.py -cp "checkpoint/celeba_hq.pth" -bs 16 -sp "data/result/celeba_hq_result.png" --show --nrow 4 --result_only