transforms.RandomCrop() has no pad_if_needed attribute

I have torch 0.4.0 and torchvision 0.2.0
According to the documentation the torchvision.transforms.RandomCrop() class should have a pad_if_needed attribute. However, when I try it I get an error:

TypeError: __init__() got an unexpected keyword argument 'pad_if_needed'

you may check your conda env :rofl: