An unofficial code for paper "Learning Frequency-aware Dynamic Network for Efficient Super-Resolution". I've implemented the learnable frequency mask part (the dilated frequency mask on the highest frequency branch is not implemented). And there is no specific threshold for manual DCT frequency masks so that part is not implemented. The code is based on https://github.com/sanghyun-son/EDSR-PyTorch Now the implemented model's performance seems poor.
-
Notifications
You must be signed in to change notification settings - Fork 1
songyonger/FADN
About
An unofficial code for paper "Learning Frequency-aware Dynamic Network for Efficient Super-Resolution"
Resources
Stars
Watchers
Forks
Releases
No releases published
Packages 0
No packages published