[Feature] Add configurations to support torch.compile in Runner

[C1rN09]

Thanks for your contribution and we appreciate it a lot. The following instructions would make your pull request more healthy and more easily get feedback. If you do not understand some items, don't worry, just make the pull request and seek help from maintainers.

Motivation

Support torch.compile introduced in PyTorch 2.0

Modification

Check for compile keyword in config dict. If not None, will call torch.compile according to its options.

BC-breaking (Optional)

No

Use cases (Optional)

Add a single line (or a few lines) in your config file.

Basic Use Cases

# recommended
compile = True
# equivalent to the following code
# self.model.train_step = torch.compile(self.model.train_step)

Advanced Use Cases

You can pass arguments to torch.compile, as described in PyTorch Documentation

compile = dict(
    backend='inductor',
    mode='reduce-overhead'
)
# equivalent to the following code
# self.model.train_step = torch.compile(
#     self.model.train_step,
#     backend='eager',
#     mode='reduce-overhead'
# )

Due to a PyTorch mkldnn issue, we will not compile val_step and test_step unless the issue has been fixed.

Checklist

1. Pre-commit or other linting tools are used to fix the potential lint issues.
2. The modification is covered by complete unit tests. If not, please add more unit test to ensure the correctness.
3. If the modification has potential influence on downstream projects, this PR should be tested with downstream projects, like MMDet or MMCls.
4. The documentation has been modified accordingly, like docstring or example tutorials.

[codecov]

Codecov Report

❗ No coverage uploaded for pull request base (main@b1b1f53). Click here to learn what that means.
Patch has no changes to coverable lines.

❗ Current head abe4283 differs from pull request most recent head 0c79866. Consider uploading reports for the commit 0c79866 to get more accurate results

Additional details and impacted files

@@           Coverage Diff           @@
##             main     #976   +/-   ##
=======================================
  Coverage        ?   76.61%           
=======================================
  Files           ?      139           
  Lines           ?    10954           
  Branches        ?     2190           
=======================================
  Hits            ?     8392           
  Misses          ?     2201           
  Partials        ?      361           

Flag	Coverage Δ
unittests	76.61% <0.00%> (?)

Flags with carried forward coverage won't be shown. Click here to find out more.

Help us with your feedback. Take ten seconds to tell us how you rate us. Have a feature suggestion? Share it here.

☔ View full report in Codecov by Sentry.
📢 Do you have feedback about the report comment? Let us know in this issue.