@ushareng has the VideoSwin model been added yet? The video classifier should be added once we have the VideoSwin backbone in. That will help us verify the implementation.
Going to close this. The commit history got crazy. We can still add a video classifier, but probably should add it along with the model that will use the task, rather than adding the task first.
@ushareng has the VideoSwin model been added yet? The video classifier should be added once we have the VideoSwin backbone in. That will help us verify the implementation.