Donghee Son
Results
2
issues of
Donghee Son
Thanks for your great work! I have a question about training using CelebA dataset In SRFlow paper, we pre-process the dataset by cropping aligned patches, which are resized to the...
I used the pre-trained RCAN_BIX2 to train RCAB_BIX3 or RCAB_BIX4, as written in your TrainRCAN_scripts.sh . However, the training loss is too large. If I use pre-trained RCAB_BIX2 to train...