[Enhance] LR and Momentum Visualizer

[teamwong111]

Motivation

This PR can be used to check curves of lr & momentum in actual training.
Besides, a web demo on Hugging face spaces likes this can be provided to make schedulers accessible. The prototype is here

Modification

Add a script file.

Use cases (Optional)

I have tested it with resnet18_8xb16_cifar10.py in mmcls.

Downstream repo can use

from mmengine.utils.visualize import fake_run
fake_run(cfg)

to replace

runner = Runner.from_cfg(cfg)
runner.train()

in tools/train.py

After training, you can start tensorboard to see curves of lr & momentum.

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.

[HAOCHENYE]

Why is it here?

[teamwong111]

Why is it here?

Because the momentum logged at iter-0 is updated.

[teamwong111]

So our purpose is to make the scheduler accessible or be useful in the training process? If the former, I think this form is more suitable.

[Ezra-Yu]

Can we plot the curve by using mmengine.Visualizer?

[Ezra-Yu]

LGTM.