Shuyang Gu

Results on Vimeo-90K-T

2

Hi, can your model EDVR_Vimeo90K_SR_L get the results fo 35.79/0.9374 on Vimeo-90K-T datasets? If not, how to get it?

Could you release your generative model for fast test?

Learning rate and ngpus

2

The default learning rate is 0.002 when I use 4 GPUs, each of them adopts the batch size number as 8. When I use 8 GPUs, and each of them...

What's the meaning of augment_p?

2

It seems that the augment comes from "Training Generative Adversarial Networks with Limited Data", But I am confused what 'augment_p' and 'ada_target' and 'ada_length' means. Can I use these augmentation...

image size and the differences with official implementation

20

thanks for this work! But i still have two question: 1. the number of styledconvblock is 6, so the output image size is no more than 128, am i right?...