limbee / NTIRE2017

Torch implementation of "Enhanced Deep Residual Networks for Single Image Super-Resolution"
652 stars 146 forks source link

the result on BSD100 #34

Closed liwanlan closed 5 years ago

liwanlan commented 6 years ago

Hi~ I want to know why the PSNR and SSIM values on BSD100 when I run the pre-trained code are worse than your published a lot . Thank you.

limbee commented 5 years ago

What kind of code did you use for evaluation?

liwanlan commented 5 years ago

the PyTorch version of EDSR(×4),the results on Set5 and Set14 are well.

limbee commented 5 years ago

Then you can ask @thstkdgus35 about this

liwanlan commented 5 years ago

ok, thank you.