UnboundLocalError: local variable 'e' referenced before assignment

[rcremese]

Hello, First thanks for the good job of implementing this segmentation pipeline and to take time to fixe this issue. As described in the documentation, I trained a 2d model on the 5 folds of a segmentation task (task_id=500) without saving the soft_max result. After doing so, I needed to compare the trained models so I reffered to the Model Training section of your documentation and ran the following command : nnUNet_train 2d nnUNetTrainerV2 500 1 -val --npz

And I get this unexpected error after some preprocessing : File "~/segmentation-pipeline/nnUNet/nnunet/training/network_training/nnUNetTrainer.py", line 678, in validate if e is not None: UnboundLocalError: local variable 'e' referenced before assignment

The chunk of code that raises an error from l.665 to l.679 :

for f in subfiles(self.gt_niftis_folder, suffix=".nii.gz"):
    success = False
    attempts = 0
    e = None
    while not success and attempts < 10:
        try:
            shutil.copy(f, gt_nifti_folder)
            success = True
        except OSError as e:
            attempts += 1
            sleep(1)
    if not success:
        print("Could not copy gt nifti file %s into folder %s" % (f, gt_nifti_folder))
        if e is not None:
            raise e

The solution I propose, knowing I mostly use your package as a blackbox :

for f in subfiles(self.gt_niftis_folder, suffix=".nii.gz"):
    success = False
    attempts = 0
    while not success and attempts < 10:
        try:
            shutil.copy(f, gt_nifti_folder)
            success = True
        except OSError:
            attempts += 1
            sleep(1)
    if not success:
        print("Could not copy gt nifti file %s into folder %s" % (f, gt_nifti_folder))
        raise OSError("Could not copy gt nifti file %s into folder %s" % (f, gt_nifti_folder))