Could you provide more details on generating rank dataset

XPixelGroup / RankSRGAN

ICCV 2019 (oral) RankSRGAN: Generative Adversarial Networks with Ranker for Image Super-Resolution. PyTorch implementation

269 stars 60 forks source link

Could you provide more details on generating rank dataset #6

Open xmengli opened 5 years ago

xmengli commented 5 years ago

When generating rank dataset, what should change in test_RankSRGAN.yml? For example, if generating DIV2K_train_ESRGAN rank dataset, which_model_G in test_RankSRGAN.yml should I set?

WenlongZhang0517 commented 5 years ago

This is just like to do testing by a pretrain model. You could set the which_model_G in test RankSRGAN.yml to ESRGAN/SRGAN/SRResNet. I mean that you could download the ESRGAN pretrain model and set the path for which_model_G. Then, you can use DIV2K LR image to generate three level images and crop them to get the pair-wise data using matlab code.

myycurs commented 4 years ago

@WenlongZhang0517 but I found only 'MSRResNet', 'RRDBNet','SRResNet' model name can be choosed in function codes/model/networks.define_G? which receive which_model_G parameter.