search

JiahuiYu / wdsr_ntire2018

Code of our winning entry to NTIRE super-resolution challenge, CVPR 2018
http://www.vision.ee.ethz.ch/ntire18/
598 stars 123 forks source link

WDSR-A vs. WDSR-B #4

Closed DanChen001 closed 5 years ago

DanChen001 commented 5 years ago

Hi, Jiahui,

Thank you for sharing the code.

Did you compare the performance of WDSR-A and WDSR-B when the number of resiudal blocks is larger than 8？

It seems that the WDSR-B is more effective than WDSR-A when the resiudal blocks are very few.

Thanks

JiahuiYu commented 5 years ago

@DanChen001 Hi, first thank you for your interest and investigation on our work. We mainly aim at efficient models for SR, for which we propose WDSR-A and WDSR-B. For accuracy (PSNR) only, I guess the model architecture does not matter a lot, as long as one stacks many many residual blocks (e.g., a very large EDSR may perform similarly to large WDSR-A or WDSR-B).

DanChen001 commented 5 years ago

@JiahuiYu Thank you for the reply.