Junru Wu comments

Results 7 comments of


                                            Junru Wu

x2num.py:14] NaN or Inf found in input tensor

How did you load image to your numpy array? Did NaN or Inf occured during loading?

Problems in running under Windows

CUDA on windows is known to be slower than that on linux, what's your gpu utility rate？

RAGAN-LS loss is difference between code and paper

Hi Hong, After carefully check the code, i found it the implementation consistent with the paper. In the paper, RaLSGAN Loss is described as: In the repo, we use: https://github.com/TAMU-VITA/DeblurGANv2/blob/c5bbf780b152d5b877d9911b8323898d5f2e0a50/models/losses.py#L210-L211...

1920*1080 image It doesn't work

Please provide more detail on it, the pre-trained model works best on 1280 x 720 image.

train problem

How did you arrange you data? The error indicate there is no available image under under one of the folders. Did you make sure folder ./datasets/train/blur/ & ./datasets/train/sharp/, ./datasets/test/blur/ &...

Train slowly under window

CUDA on windows is known to be slower than that on linux, what's your gpu utility rate？

Instructions to obtain the pre-processed NAS-Bench-201 xxx Subset result?

Hi detectRecog, Thanks for your interest in our paper! I just fix the AssertionError bug in the latest commits, can you try it again? Regarding pre-processed subset: For the NASBench...