Zijian Zhang

Thanks for your attention. A step means a step of optimization, and `training_samples = batch_size * step` . We don't explicitly have a parameter to stop the training process and...

Thanks for your attention! Dou you mean the zero_module here? https://github.com/ckczzj/PDAE/blob/fbba0355634861196aed8b80b9ba4948ed210ab9/model/module/module.py#L362-L364 It is just a zero-initialization of the output conv layer. The zero-initialization makes the residual block work like an...

If no trained_ddpm_checkpoint exists for your data, you may need to train it by yourself using [UNet](https://github.com/ckczzj/PDAE/blob/master/model/unet/unet.py). We will update the code to support regular ddpm training later.

We have updated the regular ddpm training code. To train your own ddpm, you should add or modify following files: - dataset/your_dataset.py - config/your_dataset_regular.yaml - model/denoise_fn/your_dataset.py - model/denoise_fn/__init__.py - trainer/train_regular.py...