I think this doc would help.
Tensor are treated in a element-wise way, so if you code
A = torch.torch.randn(10000, 1, 1024)
B = A / 255
print(B)
print(A)
Even though, there is some specific documentation about lib torchvision.transforms that include normalize
https://pytorch.org/docs/stable/_modules/torchvision/transforms/transforms.html#Normalize
I hope this helps 