Replace FakeTensorUpdater copy with Torch import

[benjaminglass1]

Description

#4017 added a copy of torch._inductor.fx_utils.FakeTensorUpdater to the repo, with some minor changes. The original class is being updated to handle subgraphs and HOPs, and fix corner cases, in pytorch/pytorch#159523. Rather than once again copying all the new code over, I've changed to subclassing from FakeTensorUpdater and only implementing the interface changes.

Type of change

• Port of PyTorch feature updates

Checklist:

• My code follows the style guidelines of this project (You can use the linters)
• I have performed a self-review of my own code
• I have commented my code, particularly in hard-to-understand areas and hacks
• I have made corresponding changes to the documentation
• I have added tests to verify my fix or my feature
• New and existing unit tests pass locally with my changes
• I have added the relevant labels to my PR in so that relevant reviewers are notified

[benjaminglass1]

Updated with a compatibility patch, to handle the fact that this code is not in a release yet.

[narendasan]

@cehongwang Can you review when you get back?