Update channels_first parameter and docstring for PyTorch Object Detectors

[f4str]

Description

Set the channels_first parameter in PyTorchObjectDetector and PyTorchFasterRCNN to default to true instead of false and update the input_shape default parameter accordingly.

Additionally, update docstrings and the type hinting for PyTorchFasterRCNN.

Fixes #2264

Type of change

Please check all relevant options.

• Improvement (non-breaking)
• Bug fix (non-breaking)
• New feature (non-breaking)
• Breaking change (fix or feature that would cause existing functionality to not work as expected)
• This change requires a documentation update

Testing

Please describe the tests that you ran to verify your changes. Consider listing any relevant details of your test configuration.

• Test cases for PyTorchObjectDetector are unchanged
• Test cases for PyTorchFasterRCNN are unchanged
• Test cases for PyTorchYolo are unchanged

Test Configuration:

• OS
• Python version
• ART version or commit number
• TensorFlow / Keras / PyTorch / MXNet version

Checklist

• My code follows the style guidelines of this project
• I have performed a self-review of my own code
• I have commented my code
• I have made corresponding changes to the documentation
• My changes generate no new warnings
• I have added tests that prove my fix is effective or that my feature works
• New and existing unit tests pass locally with my changes

[codecov-commenter]

Codecov Report

Merging #2265 (98016c6) into dev_1.15.2 (ad0a546) will increase coverage by 3.98%.
The diff coverage is n/a.

❗ Your organization needs to install the Codecov GitHub app to enable full functionality.

@@              Coverage Diff               @@
##           dev_1.15.2    #2265      +/-   ##
==============================================
+ Coverage       81.63%   85.61%   +3.98%     
==============================================
  Files             308      308              
  Lines           27476    27476              
  Branches         5048     5048              
==============================================
+ Hits            22429    23523    +1094     
+ Misses           3817     2671    -1146     
- Partials         1230     1282      +52     

Files Changed	Coverage Δ
...estimators/object_detection/pytorch_faster_rcnn.py	100.00% <ø> (ø)
...mators/object_detection/pytorch_object_detector.py	84.65% <ø> (ø)
art/estimators/object_detection/pytorch_yolo.py	82.48% <ø> (ø)

... and 30 files with indirect coverage changes

📢 Have feedback on the report? Share it here.

[beat-buesser]

Does the Faster R-CNN model of torchvision expect this input size?

[beat-buesser]

Hi @f4str Thank you very much for the pull request! I have one question about the input shape, what do you think?

[f4str]

Hi @beat-buesser thank you for the review. As per our discussion, it does make sense to leave the input_shape parameter for the PyTorchObjectDetector and PyTorchFasterRCNN as (-1, -1, -1) since they accept any input arbitrarily sized image. I've made the corresponding change.