Fix composite compliance

[zou3519]

Aka, pytorch/pytorch#69991

Motivation

We should make sure all PyTorch operations are "Composite Compliant". This condition is necessary for operators that look out-of-place to preserve the Tensor Subclass when passed in a Tensor Subclass.

Consider the following hypothetical out-of-place operation:

def my_add(x, y):
  result = x.clone()
  result.add_(y)
  return result

You may expect this to work the same as torch.add. However, if x is not a Tensor Subclass, but y is a Tensor subclass, then this returns us a regular Tensor, NOT a Tensor subclass!