[Feature] Support ReduceLrUpdaterHook

[Yshuo-Li]

No description provided.

[codecov]

Codecov Report

Merging #860 (6c6a715) into master (fa272ac) will increase coverage by 0.22%.
The diff coverage is 92.25%.

@@            Coverage Diff             @@
##           master     #860      +/-   ##
==========================================
+ Coverage   83.26%   83.48%   +0.22%     
==========================================
  Files         222      222              
  Lines       12669    12808     +139     
  Branches     2054     2093      +39     
==========================================
+ Hits        10549    10693     +144     
+ Misses       1795     1784      -11     
- Partials      325      331       +6     

Flag	Coverage Δ
unittests	83.44% <92.25%> (+0.21%)	⬆️

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

Impacted Files	Coverage Δ
mmedit/core/scheduler/lr_updater.py	93.03% <92.08%> (+66.72%)	⬆️
mmedit/core/__init__.py	100.00% <100.00%> (ø)
mmedit/core/scheduler/__init__.py	100.00% <100.00%> (ø)
...ls/components/stylegan2/generator_discriminator.py	86.14% <0.00%> (+1.20%)	⬆️

---

Continue to review full report at Codecov.

Legend - Click here to learn more
Δ = absolute <relative> (impact), ø = not affected, ? = missing data
Powered by Codecov. Last update fa272ac...6c6a715. Read the comment docs.

[wangruohui]

We need to at least assert something in the test, e.g. if the learning rate reduces to a desired value.
Same for test_reduce_lr_updater_hook

[wangruohui]

Is this another feature irrelevant to reducing on the plateau?

[Yshuo-Li]

Is this another feature irrelevant to reducing on the plateau?

Yes, it is another feature irrelevant to reducing on the plateau, which is in lack of unit test

[wangruohui]

We still need to assert the lr get from lr_updater.get_lr(fake_runner, 1) is expected. They are missing currently.