These kinds of operations are quite common in machine learning contexts (e.g. used in PyTorch). Are they implemented in the same way on PyTorch as we'd be describing in class?