Thanks for sharing your code!
In your paper, I can find SoTA performance in many tasks, including classification, action recognition, detection, and semantic segmentation. But for now (Apr.24, 24), classification code is only available.
Do you have plans to release the remaining code?