[Feature] Support random degradations during training

[ckkelvinchan]

Motivation

Recent works on face restoration or real-world super-resolution rely heavily on applying random degradations during training.

Modification

This PR adds support to generate random degradations in the data processing pipeline. It currently supports

1. Random resize
2. Gaussian blur
3. Sinc filter
4. Gaussian noise
5. Poisson noise
6. JPEG compression
7. Random shuffling of degradations

This is a generic class. Supporting other degradations can be done by adding corresponding operations in this class.

[codecov]

Codecov Report

Merging #504 (8246d99) into master (84d2eeb) will increase coverage by 0.13%.
The diff coverage is 84.29%.

❗ Current head 8246d99 differs from pull request most recent head c68f54c. Consider uploading reports for the commit c68f54c to get more accurate results

@@            Coverage Diff             @@
##           master     #504      +/-   ##
==========================================
+ Coverage   80.41%   80.54%   +0.13%     
==========================================
  Files         191      193       +2     
  Lines       10378    10759     +381     
  Branches     1553     1631      +78     
==========================================
+ Hits         8345     8666     +321     
- Misses       1796     1835      +39     
- Partials      237      258      +21     

Flag	Coverage Δ
unittests	80.54% <84.29%> (+0.13%)	⬆️

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

Impacted Files	Coverage Δ
mmedit/datasets/pipelines/augmentation.py	91.22% <27.77%> (-5.76%)	⬇️
mmedit/datasets/pipelines/blur_kernels.py	84.55% <84.55%> (ø)
mmedit/datasets/pipelines/random_degradations.py	93.75% <93.75%> (ø)
mmedit/datasets/pipelines/__init__.py	100.00% <100.00%> (ø)
...ls/components/stylegan2/generator_discriminator.py	84.93% <0.00%> (ø)

---

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 84d2eeb...c68f54c. Read the comment docs.

[innerlee]

A general note: for augmentations, please add the added/modified keys in docstring.

[innerlee]

How about input_ -> input?