After cropping the shape has been changed.

[brad-luo]

Hi team,

After training, I tried to do the inference. I don't wanna do any cropping procedure because it changed the output shape. Here is the example.

before crop: (3, 21, 384, 384) after crop: (3, 21, 384, 383) spacing: [3.  0.5 0.5] 
no resampling necessary
no resampling necessary
before: {'spacing': array([3. , 0.5, 0.5]), 'spacing_transposed': array([3. , 0.5, 0.5]), 'data.shape (data is transposed)': (3, 21, 384, 383)} 
after:  {'spacing': array([3. , 0.5, 0.5]), 'data.shape (data is resampled)': (3, 21, 384, 383)} 

Is there any way I can set the cropping off?

Thanks.

[FabianIsensee]

Why do you not want to have the cropping step? It will not hurt the predictions. Disabling it would be a pain because it's an integral part of nnU-net that is needed for some datasets