There is currently a somewhat implicit expectation that annotate(x, annotator) is called with x being a 2D image. It would be natural for annotate to also be callable on a 3D image tensor (a batch of images), and a list of 2D images.
Likely not breaking...
There is currently a somewhat implicit expectation that
annotate(x, annotator)is called withxbeing a 2D image. It would be natural for annotate to also be callable on a 3D image tensor (a batch of images), and a list of 2D images.Likely not breaking...