code looks good to me, I'm not sure about the class name RandomOrder, to me the operation is similar with np.random.shuffle https://numpy.org/doc/stable/reference/random/generated/numpy.random.shuffle.html

Yes, actually I'm referring to the name of torchvision.
https://pytorch.org/vision/stable/_modules/torchvision/transforms/transforms.html#RandomOrder

code looks good to me, I'm not sure about the class name RandomOrder, to me the operation is similar with np.random.shuffle https://numpy.org/doc/stable/reference/random/generated/numpy.random.shuffle.html

Yes, actually I'm referring to the name of torchvision. https://pytorch.org/vision/stable/_modules/torchvision/transforms/transforms.html#RandomOrder

Ok, that's also implemented with a 'shuffle'. Any particular reason that permutation is used in this implementation?

code looks good to me, I'm not sure about the class name RandomOrder, to me the operation is similar with np.random.shuffle https://numpy.org/doc/stable/reference/random/generated/numpy.random.shuffle.html

Yes, actually I'm referring to the name of torchvision. https://pytorch.org/vision/stable/_modules/torchvision/transforms/transforms.html#RandomOrder

Ok, that's also implemented with a 'shuffle'. Any particular reason that permutation is used in this implementation?

Yes, since self. transforms is tuple object, which does not support item assignment.

code looks good to me, I'm not sure about the class name RandomOrder, to me the operation is similar with np.random.shuffle https://numpy.org/doc/stable/reference/random/generated/numpy.random.shuffle.html

Yes, actually I'm referring to the name of torchvision. https://pytorch.org/vision/stable/_modules/torchvision/transforms/transforms.html#RandomOrder

Ok, that's also implemented with a 'shuffle'. Any particular reason that permutation is used in this implementation?

Yes, since self. transforms is tuple object, which does not support item assignment.

MONAI/monai/transforms/compose.py

Line 126 in 624e832

self.transforms = ensure_tuple(transforms)

Thanks!

Since the transforms are changed at every iteration, perhaps change it to a list and then shuffle the list is more efficient

/build