🚀 Feature

Add aliases for functions which are present in Numpy and perform similar (same?) functionality to PyTorch.

Motivation

Too often one ends up jumping between PyTorch and Numpy (for plotting purposes, but also for scipy and other things). It's more than frequent that one encounters AttributeError: 'Tensor' object has no attribute 'clip', when by inaccuracy applies function of the wrong name. It's understandable, that naming convention for PyTorch, perhaps, has a good motivation behind it, but adding convenience aliases could alleviate some of unnecessary pain.

Pitch

Introducing aliases for some functionality found in NumPy. NumPy has 1+ version and a stable API, which is familiar to most Python users who work with numerical data.

Alternatives

Adding a wrapper library, creating a library to monkey-patch the repo, or forking the repo. None of those are desirable, since the risk of breaking things would overweight any kind of convenience benefits.

Additional context

I understand that this is unlikely to be done, but simply wanted to see community opinion about this issue, and if, perhaps, I'm missing some understanding behind those differences in naming.

cc @mruberry @rgommers