[bug] Pyro is not wrapping LR schedulers for pytorch 2.0

[sjfleming]

Issue Description

With pytorch 2.0, the pyro.optim PyroLRScheduler wrappers no longer seem to be wrapping the pytorch learning rate schedulers and including them in the pyro.optim namespace.

from pyro import optim
optim.ExponentialLR

Traceback (most recent call last):
  File "<stdin>", line 1, in <module>
AttributeError: module 'pyro.optim' has no attribute 'ExponentialLR'

same for

import pyro.optim
pyro.optim.ExponentialLR

Traceback (most recent call last):
  File "<stdin>", line 1, in <module>
AttributeError: module 'pyro.optim' has no attribute 'ExponentialLR'

Environment

• OS: Mac Monterey 12.6.3
• PyTorch version: 2.0.0
• Pyro version: 1.8.4
• Python 3.8 and 3.9 (did not try higher)

Code Snippet

import pyro.optim
pyro.optim.ExponentialLR

[sjfleming]

Oops, looks like @eb8680 already took care of this. Thanks for pointing that out @ordabayevy .

[sjfleming]

Closed by #3167