Training details

[vidit98]

Hi, great work the results look really good. I had few questions regarding training

1. How long did you train the BSRNet?
2. Did you apply the proposed data degradation pipeline on the fly as a data-augmentation technique or you applied it on images beforehand and stored them?

[iamwwn]

I also want to know.
On the fly is 10 time slow at least when I try the code.

[lucky-zwx]

大家在训练的时候有没有出现GPU间断运行的情况？

[HodaKhsh]

I also want to know. On the fly is 10 time slow at least when I try the code.

@iamadai Did you train BSRGAN with your own data?