[WIP] Add swinv2 upernet

[rznas]

trafficstars

Motivation

Add SWinV2 backbone for UPerNet implementation #1778

Because of the minor changes introduce by the paper, It's preferred to inherit SWinV2 from SWin implementation.

Swin Transformer V2: Scaling Up Capacity and Resolution

Modification

• Edit SWin backbone for compatibility and use as abstract class (diff is somehow irrelevant, please compare two version side-by-side 😅)
• Add SWinV2 class based on SWin (The code added is based on Official Repo)

BC-breaking (Optional)

No

Checklist

• [x] Pre-commit or other linting tools are used to fix the potential lint issues.
• [ ] The modification is covered by complete unit tests. If not, please add more unit test to ensure the correctness.
• [ ] If the modification has potential influence on downstream projects, this PR should be tested with downstream projects, like MMDet or MMDet3D.
• [ ] The documentation has been modified accordingly, like docstring or example tutorials.

[CLAassistant]

All committers have signed the CLA.

[codecov[bot]]

Codecov Report

Base: 88.97% // Head: 88.05% // Decreases project coverage by -0.92% :warning:

Coverage data is based on head (4134e51) compared to base (77dbecd). Patch coverage: 13.15% of modified lines in pull request are covered.

:exclamation: Current head 4134e51 differs from pull request most recent head 67781d9. Consider uploading reports for the commit 67781d9 to get more accurate results

Additional details and impacted files

@@            Coverage Diff             @@
##           master    #1789      +/-   ##
==========================================
- Coverage   88.97%   88.05%   -0.93%     
==========================================
  Files         146      145       -1     
  Lines        8747     8739       -8     
  Branches     1474     1473       -1     
==========================================
- Hits         7783     7695      -88     
- Misses        722      804      +82     
+ Partials      242      240       -2     

Flag	Coverage Δ
unittests	88.05% <13.15%> (-0.93%)	:arrow_down:

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

Impacted Files	Coverage Δ
mmseg/models/backbones/swinv2.py	0.00% <0.00%> (ø)
mmseg/models/backbones/swin.py	83.90% <93.75%> (+0.28%)	:arrow_up:
mmseg/apis/inference.py	61.66% <0.00%> (-2.40%)	:arrow_down:
mmseg/models/decode_heads/decode_head.py	92.63% <0.00%> (-0.83%)	:arrow_down:
mmseg/datasets/pipelines/transforms.py	98.06% <0.00%> (-0.78%)	:arrow_down:
mmseg/models/segmentors/cascade_encoder_decoder.py	94.59% <0.00%> (-0.15%)	:arrow_down:
mmseg/core/evaluation/metrics.py	93.61% <0.00%> (-0.07%)	:arrow_down:
mmseg/models/decode_heads/knet_head.py	93.71% <0.00%> (-0.04%)	:arrow_down:
mmseg/datasets/__init__.py	100.00% <0.00%> (ø)
mmseg/models/losses/__init__.py	100.00% <0.00%> (ø)
... and 5 more

Help us with your feedback. Take ten seconds to tell us how you rate us. Have a feature suggestion? Share it here.

:umbrella: View full report at Codecov.
:loudspeaker: Do you have feedback about the report comment? Let us know in this issue.