likesum / bpn

likesum/bpn Issues

The training cost
Updated 2 years ago
There is a big difference between the flops of network and the paper
Updated 2 years ago
Hello，this is a amazing work. And I have run the train.py. But I try to save the image, then many problems appeared. Why tensor can convert to numpy dtype? Or could you show me how to save the denoised answer.
Updated 2 years ago
OOM...
Closed 3 years ago1
Pretrained model is missing
Closed 3 years ago1