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

Ablation study for WDSR-B #41

Closed hobin-jeong closed 4 years ago

hobin-jeong commented 4 years ago

Hello, @JiahuiYu !

I'm trying to implement your network.

When I checked your light models for WDSR-B, I found the difference between your results and my results. I saying WDSR-B with n resblocks as WDSR-B_nblks, you assert that the numbers of parameters of WDSR-A_nblks and WDSR-B_nblks are equal. ( e.g. the number of parameters of WDSR-A_1blks and WDSR-B_1blks is 0.08M )

According to my computation, the numbers of parameters of WDSR-A_nblks are almost equal, but those of WDSR-B_nblks are quite different. ( more precisely, my results are less than yours )

How to compute the number of parameters? I use this formula: kernel_size^2 ch_in ch_out (weight) + ch_out (bias) for each convolution.

Thanks in advance!

Hobin

JiahuiYu commented 4 years ago

@hobin-jeong Hi, we directly print the number of trainable parameters instead of manually calculate it - so our numbers should be correct. Moreover your formula should be correct for each convolution - please carefully checkout the ch_in and ch_out in our WDSR-B.