Using pre-trained model on different image resolution (512x512 -> 1024x1024)

[KanielDatz]

Hi, I trained an nnUNet segmentation model on data from Hospital A with images sized 512x512. I now have a new dataset from Hospital B where the image resolution is 1024x1024.

I would like to use the pre-trained weights from Hospital A as a starting point.

Is it possible to load those weights directly even though the input resolution is different?

If not, is there a recommended way to use the pre-trained model (e.g., resizing, fine-tuning, transfer learning) when training on the Hospital B dataset?

Thanks in advance!