[Fix] ZeroRedundancyOptimizer ambiguous error with param groups when pytorch < 1.12.0

[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

See issue #778

Modification

Add check for param groups and torch version, and give a better instruction to users.

BC-breaking (Optional)

No

Use cases (Optional)

Same as before

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.

[CLAassistant]

All committers have signed the CLA.

[codecov]

Codecov Report

Base: 78.66% // Head: 78.86% // Increases project coverage by +0.19% 🎉

Coverage data is based on head (f8c57b5) compared to base (fe26c65).
Patch coverage: 26.82% of modified lines in pull request are covered.

❗ Current head f8c57b5 differs from pull request most recent head 1228931. Consider uploading reports for the commit 1228931 to get more accurate results

Additional details and impacted files

@@            Coverage Diff             @@
##             main     #818      +/-   ##
==========================================
+ Coverage   78.66%   78.86%   +0.19%     
==========================================
  Files         128      128              
  Lines        9348     9368      +20     
  Branches     1848     1857       +9     
==========================================
+ Hits         7354     7388      +34     
+ Misses       1679     1666      -13     
+ Partials      315      314       -1     

Flag	Coverage Δ
unittests	78.86% <26.82%> (+0.19%)	⬆️

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

Impacted Files	Coverage Δ
mmengine/optim/optimizer/default_constructor.py	98.23% <ø> (ø)
mmengine/optim/optimizer/zero_optimizer.py	52.17% <0.00%> (-4.97%)	⬇️
mmengine/registry/build_functions.py	80.20% <ø> (ø)
mmengine/registry/default_scope.py	100.00% <ø> (ø)
mmengine/visualization/vis_backend.py	84.49% <ø> (ø)
mmengine/visualization/visualizer.py	93.19% <ø> (ø)
mmengine/runner/checkpoint.py	52.07% <11.11%> (+8.11%)	⬆️
mmengine/model/weight_init.py	35.63% <33.33%> (+0.12%)	⬆️
mmengine/optim/optimizer/optimizer_wrapper.py	62.06% <71.42%> (-0.44%)	⬇️
mmengine/optim/optimizer/optimizer_wrapper_dict.py	98.50% <100.00%> (ø)
... and 1 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 at Codecov.
📢 Do you have feedback about the report comment? Let us know in this issue.

[zhouzaida]

Please update unit tests in

mmengine/tests/test_optim/test_optimizer/test_optimizer.py

Line 727 in bf5c265

class TestZeroOptimizer(MultiProcessTestCase):

[nijkah]

LGTM

[HAOCHENYE]

We should also test has the overriden function state_dict worked