Remove in-place operations

[beat-buesser]

Description

This pull request removes in-place operations that overwrite the input data.

Fixes #2262

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

[codecov-commenter]

Codecov Report

All modified and coverable lines are covered by tests ✅

Project coverage is 85.62%. Comparing base (8d9de16) to head (52f64a8).
Report is 25 commits behind head on dev_1.15.2.

Additional details and impacted files

@@             Coverage Diff             @@
##           dev_1.15.2    #2263   +/-   ##
===========================================
  Coverage       85.62%   85.62%           
===========================================
  Files             308      308           
  Lines           27476    27476           
  Branches         5048     5048           
===========================================
  Hits            23526    23526           
  Misses           2669     2669           
  Partials         1281     1281           

Flag	Coverage Δ
	19.04% <0.00%> (?)

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

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

[f4str]

Hi @beat-buesser thank you fixing this issue! I've tested it with the attack_adversarial_patch_pytorch_yolo.ipynb demo notebook and it is working once again without issues.

One minor comment that the change in translate_predictions_xcycwh_to_x1y1x2y2 is not needed, but functionality-wise its the same. Feel free to keep the change or revert it back.

[f4str]

This isn't necessary since labels is a newly created tensor. But the functionality is still the same.