[Feature] Support dark dataset test

[xiexinch]

Motivation

Support Dark Zurich and Nighttime Driving datasets test.

Dark Zurich

Method	Backbone	mIoU
PSPNet	R-50-D8	10.91
PSPNet	R-101-D8	10.16
PSPNet	R-101b-D8	15.54

Nighttime Driving

Method	Backbone	mIoU
PSPNet	R-50-D8	23.02
PSPNet	R-101-D8	20.25
PSPNet	R-101b-D8	22.25

dependency

#750

[codecov]

Codecov Report

Merging #815 (6adf08b) into master (e235c1a) will increase coverage by 1.12%.
The diff coverage is 86.36%.

❗ Current head 6adf08b differs from pull request most recent head be5a3b1. Consider uploading reports for the commit be5a3b1 to get more accurate results

@@            Coverage Diff             @@
##           master     #815      +/-   ##
==========================================
+ Coverage   87.64%   88.76%   +1.12%     
==========================================
  Files         108      109       +1     
  Lines        5886     5823      -63     
  Branches      958      935      -23     
==========================================
+ Hits         5159     5169      +10     
+ Misses        535      463      -72     
+ Partials      192      191       -1     

Flag	Coverage Δ
unittests	88.75% <86.36%> (+1.11%)	⬆️

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

Impacted Files	Coverage Δ
mmseg/models/utils/__init__.py	100.00% <ø> (ø)
mmseg/models/backbones/hrnet.py	84.37% <80.00%> (+1.04%)	⬆️
mmseg/datasets/dark_zurich.py	83.33% <83.33%> (ø)
mmseg/datasets/night_driving.py	83.33% <83.33%> (ø)
mmseg/datasets/__init__.py	100.00% <100.00%> (ø)
mmseg/datasets/cityscapes.py	85.00% <100.00%> (ø)
mmseg/models/backbones/swin.py	87.41% <100.00%> (+0.51%)	⬆️
mmseg/models/backbones/vit.py	86.79% <100.00%> (+1.94%)	⬆️
... and 1 more

---

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 e235c1a...be5a3b1. Read the comment docs.

[Junjun2016]

Could you please provide some baseline results in the corresponding paper?
Add to PR message.

[Junjun2016]

Hi @xiexinch
Please

• Add data information to dataset_prepare https://github.com/open-mmlab/mmsegmentation/blob/master/docs/dataset_prepare.md (Chinese and English).
• Add Supported datasets: to README (Chinese and English).
• Add descript to https://github.com/open-mmlab/mmsegmentation/blob/master/configs/pspnet/README.md

[xiexinch]

Hi @xiexinch
Please

• Add data information to dataset_prepare https://github.com/open-mmlab/mmsegmentation/blob/master/docs/dataset_prepare.md (Chinese and English).
• Add Supported methods: to README (Chinese and English).
• Add descript to https://github.com/open-mmlab/mmsegmentation/blob/master/configs/pspnet/README.md

What information should be added to README.md at the root? I didn't implement any new methods.

[Junjun2016]

Hi @xiexinch
Please

• Add data information to dataset_prepare https://github.com/open-mmlab/mmsegmentation/blob/master/docs/dataset_prepare.md (Chinese and English).
• Add Supported methods: to README (Chinese and English).
• Add descript to https://github.com/open-mmlab/mmsegmentation/blob/master/configs/pspnet/README.md

What information should be added to README.md at the root? I didn't implement any new methods.

Typo: Supported methods --> Supported datasets
We can add an item named Supported datasets.

[MengzhangLI]

Hi, @xiexinch

I just changed your table in PSPNet config, please check it out.

Best,

[Junjun2016]

LGTM