Bee Lim

@qingchuanhuajuan Check this link from the SRGAN paper: https://twitter.app.box.com/s/lcue6vlrd01ljkdtdkhmfvk7vtjhetog

I'm not sure why it prints the segmentation fault message, but the experiment is done successfully. Trained models are saved at experiment/

Try nResBlock=32 instead of 36, if you're using TitanX. We used 32 residual blocks when writing a paper since sometimes 12GB of GPU memory is not enough for 36 resblocks.

`/var/tmp/dataset/DIV2K/DIV2K_train_LR_bicubic/X4/0045x4.png: No such file or directory` This line indicates that the training image is not located correctly, or you didn't specify the location of your dataset. You should first define...

The number of channels is set to 256, and patch size is 96. https://github.com/limbee/NTIRE2017/blob/db34606c2844e89317aac8728a2de562ef1f8aba/code/training.sh#L1-L2 This setting is suited for GPUs with 12GB of memory, so other GPUs with less than...

Hi, maybe you can validate your result by comparing the final scores with our paper. I think the warnings are not that critical.

I'm sorry I couldn't find the cause of this error. It has never seen before in our numerous experiments. If you find the solution, please let us know. Thank you.