"How to train CNN with images of different sizes? "


Question here:
I see that this CIFAR dataset has all images with 3X32X32 .
hence we are giving -
input_size = 3 32 32

what if we have images with different sizes and how do we give the input image size. How do we train them? Please let me know