Is your feature request related to a problem? Please describe.
For digital pathology training pipeline to use torch.Tensor, there are still two transform that are not supporting this input types RandRotate90 and RandZoomD. I know that there have been efforts on this front, so I am wondering if these transforms are in the pipeline? If so, is it possible to prioritize them? If not, is there anything that holding us from doing that besides time?
Is your feature request related to a problem? Please describe.
For digital pathology training pipeline to use
torch.Tensor, there are still two transform that are not supporting this input typesRandRotate90andRandZoomD. I know that there have been efforts on this front, so I am wondering if these transforms are in the pipeline? If so, is it possible to prioritize them? If not, is there anything that holding us from doing that besides time?