What does torchvision.transforms take as an input? Is it an image array or the tuple of the image and its label

torchvision.transforms usualy take a single input, not a list.
What transformation are you looking for?

I have an image as numpy array of the shape (1024,1248,3). I want to apply transformations of resize and ToTensor.

transforms.Reize as well as ToTensor take a PIL.Image as their input, so you would need to transform the numpy array to a PIL.Image first.

Thanks a lot. I will just look to that