Add YOLACT support from mmdet

[dnth]

This PR adds YOLACT: Real-time Instance Segmentation, a single stage instance segmentation model, into IceVision.

Paper: https://arxiv.org/abs/1904.02689
Configs/models: https://github.com/open-mmlab/mmdetection/blob/master/configs/yolact/README.md

[review-notebook-app]

Check out this pull request on 

See visual diffs & provide feedback on Jupyter Notebooks.

---

Powered by ReviewNB

[codecov]

Codecov Report

Merging #1046 (6f4658e) into master (bfbed41) will decrease coverage by 0.03%.
The diff coverage is 80.48%.

@@            Coverage Diff             @@
##           master    #1046      +/-   ##
==========================================
- Coverage   85.96%   85.92%   -0.04%     
==========================================
  Files         272      277       +5     
  Lines        5794     5835      +41     
==========================================
+ Hits         4981     5014      +33     
- Misses        813      821       +8     

Flag	Coverage Δ
unittests	85.92% <80.48%> (-0.04%)	⬇️

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

Impacted Files	Coverage Δ
icevision/models/mmdet/utils.py	57.81% <0.00%> (-3.86%)	⬇️
icevision/models/mmdet/common/utils.py	81.81% <25.00%> (-5.69%)	⬇️
...ion/models/mmdet/common/mask/single_stage/model.py	90.90% <90.90%> (ø)
.../models/mmdet/common/mask/single_stage/__init__.py	100.00% <100.00%> (ø)
icevision/models/mmdet/models/__init__.py	100.00% <100.00%> (ø)
icevision/models/mmdet/models/yolact/__init__.py	100.00% <100.00%> (ø)
...n/models/mmdet/models/yolact/backbones/__init__.py	100.00% <100.00%> (ø)
...models/mmdet/models/yolact/backbones/resnet_fpn.py	100.00% <100.00%> (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 bfbed41...6f4658e. Read the comment docs.

[ai-fast-track]

Thanks a lot @dnth!
Did you test it in the getting_started_instance_segmentation notebook?
What's the COCO metric score?

[dnth]

Thanks a lot @dnth! Did you test it in the getting_started_instance_segmentation notebook? What's the COCO metric score?

Yes! I tested the model with the getting started notebook on my machine. The highest score was 0.76 as shown in the screenshot.

I have also tested on Colab. Here is the link to the notebook.
https://colab.research.google.com/drive/10PUIfqvI5R9dydmn15qPTVNFdhLHhChR?usp=sharing