tensorlayer / SRGAN

Photo-Realistic Single Image Super-Resolution Using a Generative Adversarial Network
https://github.com/tensorlayer/tensorlayerx
3.29k stars 810 forks source link

How to test on custom image? #171

Closed kapilsarwat closed 5 years ago

kapilsarwat commented 5 years ago

Python train.py --mode=evaluate is running only for some images in valid_hr and valid_lr directories. Is there any way to run the model for test_lr? @zsdonghao

ChrisWu1997 commented 5 years ago

I think it's easy to modify a few line of the evaluate function in train.py to enable testing on customized image. https://github.com/tensorlayer/srgan/blob/aeca83bc85adabdefa0e75db470379b878e54051/train.py#L167

kapilsarwat commented 5 years ago

Ok thanks