Patch dataset seems extremely inefficient in training

[Ilykuleshov]

Hello! Thank you for this incredible library! I'm loving my experience with MONAI so far.
Is your feature request related to a problem? Please describe.
But I've noticed a severe inefficiency in the way PatchDataset is written. If used with RandSpacialCropSamplesd, as it is in the example, it takes more time to sample the same number of patches, than if using RandSpacialCropd with a simple Dataset. That's due to how the patches are sampled: for every patch, first, all patches are resampled, then a patch is chosen with the corresponding index, while the other sampled patches are just thrown away. So, this creates two problems

1. First, obviously, it is unnecessarily slow, re-sampling patches every time, so it makes no sense to use it to speed up training.
2. Second, it eats up more memory than it should. One could easily implement this with generators, so that the memory footprint would be basically the same as in sampling a single patch.

Describe the solution you'd like
I wish PatchDataset would store a generator inside, which would yield patches. On StopIteration, one would store newly-sampled patches in the generator. Also, please add a flag, allowing to turn RandSpacialCropSamplesd into a generator, so that it wouldn't store all sampled patches in memory. It would be slower than storing all of them in memory, but it would greatly reduce the memory footprint (to the footprint of sampling a single patch).

[Nic-Ma]

Hi @drbeh ,

Could you please help share some comments on this question?

Thanks in advance.

[wyli]

the gridpatchdataset works with a generator style input

MONAI/monai/data/grid_dataset.py

Lines 148 to 160 in 575954c

	class GridPatchDataset(IterableDataset):
	"""
	Yields patches from data read from an image dataset.
	Typically used with `PatchIter` or `PatchIterd` so that the patches are chosen in a contiguous grid sampling scheme.
	.. code-block:: python
	import numpy as np
	from monai.data import GridPatchDataset, DataLoader, PatchIter, RandShiftIntensity
	# image-level dataset
	images = [np.arange(16, dtype=float).reshape(1, 4, 4),

you can combine it with monai.data.ShuffleBuffer for randomised sampling, please see

https://github.com/Project-MONAI/tutorials/blob/main/modules/2d_slices_from_3d_training.ipynb