You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
In the inference stage, you only need to input the height and width into the pipeline. Note that the height and width should be n*4.
In the training stage, the model architecture supports variable resolution, but the data loader doesn't support variable resolution because of unbalanced GPU utilization. We are designing a more flexible training approach.
No description provided.
The text was updated successfully, but these errors were encountered: