The requirement.txt file is friendly provided. You can prepare the environment by running pip install -r requirements.txt
in your command line with python=3.8.
To avoid unpredictable mistakes, you can also install python=3.8, torch=1.13.1 and other packages separately.
You can run the following codes to generate images using the checkpoint placed in ./ckpts.
cd codes
python Main.py --state eval --ckpt ckpt_best.pt
The generated images are saved in ./codes/SampledImgs. Note that it may take several seconds for the model to complete the denoising process.
Run the following codes to train your own DDPM.
cd codes
python Main.py --state train
The checkpoints will be saved in ./ckpts.
Denoising Diffusion Probabilistic Models