Fix PyTorch Object Detection Estimators Missing Gradients Bug

[f4str]

Description

When a non-leaf tensor is moved to another device, the gradient retention property is lost and needs to be manually set again. This PR adds this minor change to both the PyTorchObjectDetector and PyTorchYolo estimators.

Fixes #2248

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.

• Tests for PyTorchObjectDetector are unchanged
• Tests for PyTorchFasterRCNN 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 #2249 (4650875) into dev_1.15.1 (52c240a) will decrease coverage by 0.01%.
The diff coverage is 66.66%.

❗ Your organization is not using the GitHub App Integration. As a result you may experience degraded service beginning May 15th. Please install the Github App Integration for your organization. Read more.

@@              Coverage Diff               @@
##           dev_1.15.1    #2249      +/-   ##
==============================================
- Coverage       85.62%   85.62%   -0.01%     
==============================================
  Files             308      308              
  Lines           27470    27476       +6     
  Branches         5046     5048       +2     
==============================================
+ Hits            23522    23526       +4     
- Misses           2668     2669       +1     
- Partials         1280     1281       +1     

Files Changed	Coverage Δ
...mators/object_detection/pytorch_object_detector.py	84.65% <33.33%> (-0.90%)	⬇️
art/estimators/object_detection/pytorch_yolo.py	82.48% <100.00%> (+0.24%)	⬆️

[beat-buesser]

Hi @f4str Thank you for your pull request. In #2238 we made the same change also in method PyTorchObjectDetector._preprocess_and_convert_inputs. Is the same change also needed in PyTorchYolo._preprocess_and_convert_inputs?

[f4str]

Hi @beat-buesser the previous change in #2238 addressed a different bug that only occured for the PyTorchObjectDetector that was bug was already addressed for the PyTorchYolo in #2169. The PR applied the same fix.

This is a separate bug-fix independent from both #2238 and #2169 that affects both the PyTorchObjectDetector and PyTorchYolo. The code solution looks similar, but it's a different fix.

[beat-buesser]

Hi @f4str Thank you very much for noticing and fixing this bug!