add device_mesh support for FSDP

[winglian]

What does this PR do?

Adds support for multidimensional parallelism with FSDP.

@muellerzr

Fixes # (issue)

Before submitting

• This PR fixes a typo or improves the docs (you can dismiss the other checks if that's the case).
• Did you read the contributor guideline,
  Pull Request section?
• Was this discussed/approved via a Github issue or the forum? Please add a link
  to it if that's the case.
• Did you make sure to update the documentation with your changes? Here are the
  documentation guidelines, and
  here are tips on formatting docstrings.
• Did you write any new necessary tests?

Who can review?

Anyone in the community is free to review the PR once the tests have passed. Feel free to tag
members/contributors who may be interested in your PR.

[HuggingFaceDocBuilderDev]

The docs for this PR live here. All of your documentation changes will be reflected on that endpoint. The docs are available until 30 days after the last update.

[BenjaminBossan]

Suggested change

	from torch.distributed._tensor import DeviceMesh
	from torch.distributed import DeviceMesh

should also work. We should also check when DeviceMesh was added in torch. I tried to figure this out from the code but it's not quite clear to me, apparently it moved around a bit. Since it's a top level import, it has to be importable for all supported torch versions. Otherwise, we need some kind of guard.

[muellerzr]

I'd rather we do from torch.distributed.device_mesh import DeviceMesh.

2.3.0+ has this import, 2.2.0 has it under _device_mesh. Anything lower and we should raise a NotImplementedError.

[github-actions]

This issue has been automatically marked as stale because it has not had recent activity. If you think this still needs to be addressed please comment on this thread.

Please note that issues that do not follow the contributing guidelines are likely to be ignored.