sanghyun-son / EDSR-PyTorch

PyTorch version of the paper 'Enhanced Deep Residual Networks for Single Image Super-Resolution' (CVPRW 2017)
MIT License
2.44k stars 671 forks source link

get a wrong output image when I only run test() function #231

Closed tangeroo closed 4 years ago

tangeroo commented 4 years ago

@thstkdgus35 Hi, Thanks for your work.I ran this code on my own data.In your main.py, data testing runs after data training as below, but I want to input a trained model and test only,so I removed t.train() and I got a definitely wrong output image like below, while test output image when training is right.I really can't find the problem.Do you have any idea? Looking forward to your reply. Thank you. image image image