Photo-Realistic-Super-Resoluton
Photo-Realistic-Super-Resoluton copied to clipboard
Can u gave some model to test?
Sure. I will release the trained model soon.
Hi,
I have two questions and hope you can help me:
- In SRGAN, the G network is supervised by the high-resolution images. Do I understand correctly?
- How many epochs does this network need to generate good super-resolved images?
BTW, there's a typo in line 41 of run_test.lua, where a comma is missed.
- Yes it is correct.
- For me, I trained 3 epochs on ImageNet and it looks good already.
@leehomyc can you gave some models to test and compare?
@leehomyc - 3 epochs on image net, but how many total images were you training on?