Introduced percentile normalization for synthesis challenge metrics

[FelixSteinbauer]

Proposed Changes

• Added __percentile_clip function that normalizes input tensors based on percentiles and clips values that are blow/above. This is relevant as we notices that the image data we are working with has few but very strong outliers. These outliers are removed and the metric computation (especially metrics like PSNR which depend on the value range) gets more robust/reliable.
• Also, as we now know for sure that all values will be between 0 and 1 we can fix the value range for PSNR to 1.0 (as soon as the other pull request which introduces the data_range parameter is approved. This is why there is a "#TODO" comment in the code)

Note: I did test the __percentile_clip() function.

Checklist

• I have read the CONTRIBUTING guide.
• My PR is based from the current GaNDLF master .
• Non-breaking change (does not break existing functionality): provide as many details as possible for any breaking change.
• Function/class source code documentation added/updated.
• Code has been blacked for style consistency.
• If applicable, version information has been updated in GANDLF/version.py.
• If adding a git submodule, add to list of exceptions for black styling in pyproject.toml file.
• Usage documentation has been updated, if appropriate.
• Tests added or modified to cover the changes; if coverage is reduced, please give explanation.
• If customized dependency installation is required (i.e., a separate pip install step is needed for PR to be functional), please ensure it is reflected in all the files that control the CI, namely: python-test.yml, and all docker files [1,2,3].

[github-actions]

MLCommons CLA bot All contributors have signed the MLCommons CLA ✍️ ✅

[codecov]

Codecov Report

Merging #700 (28fc7ea) into master (44e6748) will decrease coverage by 0.01%.
The diff coverage is 100.00%.

❗ Current head 28fc7ea differs from pull request most recent head 3c31ab8. Consider uploading reports for the commit 3c31ab8 to get more accurate results

@@            Coverage Diff             @@
##           master     #700      +/-   ##
==========================================
- Coverage   94.70%   94.69%   -0.01%     
==========================================
  Files         117      117              
  Lines        8208     8207       -1     
==========================================
- Hits         7773     7772       -1     
  Misses        435      435              

Flag	Coverage Δ
unittests	94.69% <100.00%> (-0.01%)	⬇️

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

Files Changed	Coverage Δ
GANDLF/cli/generate_metrics.py	100.00% <100.00%> (ø)

... and 1 file with indirect coverage changes

📣 We’re building smart automated test selection to slash your CI/CD build times. Learn more

[sarthakpati]

Minor changes to increase code coverage. Looks good otherwise. 👍🏽

[FelixSteinbauer]

@sarthakpati This PR seems to have the same issue with the testing pipeline I reported for our other PR

[sarthakpati]

Apologies, but due to an unforeseen issue, OpenFL tests are failing [ref]. Please hang on tight while we are coordinating a fix, and we will re-initiate the failing run automatically.

[sarthakpati]

LGTM!

[sarthakpati]

LGTM

Edit - mistakenly requested changes.

[sarthakpati]

LGTM