lijun2005

I think the trainable params of DWT and IDWT should be 0, but I have tried thop and torchinfo two libraries which indicate params is not 0.

python3.10使用不了

1

python3.10使用不了

In many researches, dwt and idwt often go hand in hand. Could you please tell me why your paper uses transposed convolution instead of idwt for upsampling？

About difftopk and torch.topk

1

Great job! This is the first time I've seen a differentiable top-k selection in an engineering project so well-implemented. But I have a question for the author: What are the...

关于测试尺寸

5

作者您好，看代码中测试也是针对每一个patch做的吗？在测试的时候为啥不对整张大图进行推理计算大图的psnr呢？

Support torch.compile()?

6

Hi, thank you for providing this toolbox. Is there any future work planned toward allowing torch.compile() for ptwt?