[Fix] Fix PyTorch1.11 Dist Remove _sync_params

[teamwong111]

Motivation

Fix PyTorch1.11 Dist Remove _sync_params.

Modification

As above.

Checklist

Before PR:

• I have read and followed the workflow indicated in the CONTRIBUTING.md to create this PR.
• Pre-commit or linting tools indicated in CONTRIBUTING.md are used to fix the potential lint issues.
• Bug fixes are covered by unit tests, the case that causes the bug should be added in the unit tests.
• New functionalities are covered by complete unit tests. If not, please add more unit test to ensure the correctness.
• The documentation has been modified accordingly, including docstring or example tutorials.

After PR:

• If the modification has potential influence on downstream or other related projects, this PR should be tested with some of those projects, like MMDet or MMCls.
• CLA has been signed and all committers have signed the CLA in this PR.

[ZwwWayne]

Should involve downstream repos to test with this PR.

[zhouzaida]

https://github.com/pytorch/pytorch/blob/50c90a22be3ee6a547ad0222951f2c9f50c02b50/torch/nn/parallel/distributed.py#L277

require_forward_param_sync is a attribute of DDP since torch1.2.0 so do we still need to use getattr to get it rather than self.require_forward_param_sync?

[zhouzaida]

Please @luopeichao have a look.

[luopeichao]

It works in parrots. LGTM.

[luopeichao]

And parrots also support self.require_forward_param_sync.

[zhouzaida]

val_step should also be modified.

mmcv/mmcv/parallel/distributed.py

Line 72 in 1a2f174

def val_step(self, *inputs, **kwargs):