Add MetricInputTransformer
#2392
Conversation
lgienapp
requested review from
lantiga,
SkafteNicki,
Borda,
justusschock and
stancld
as code owners
github-actions
bot
added
the
documentation
for more information, see https://pre-commit.ci
Borda
changed the title
MetricInputTransformer
* origin/master: [pre-commit.ci] auto fixes from pre-commit.com hooks # Conflicts: # src/torchmetrics/wrappers/transformations.py # tests/unittests/wrappers/test_transformations.py
for more information, see https://pre-commit.ci
Codecov ReportAttention: Patch coverage is
Additional details and impacted files@@ Coverage Diff @@
## master #2392 +/- ##
========================================
- Coverage 69% 35% -35%
========================================
Files 313 314 +1
Lines 17615 17664 +49
========================================
- Hits 12161 6096 -6065
- Misses 5454 11568 +6114 |
* origin/master: [pre-commit.ci] auto fixes from pre-commit.com hooks
|
Unsure what to do about the failing mypy test - assigning the lambdas to overwrite the original class methods seems to be the most straightforward solution to supplying user-defined behaviour here; I would add the file to |
Co-authored-by: Bas Krahmer <baskrahmer@gmail.com>
Co-authored-by: Bas Krahmer <baskrahmer@gmail.com>
for more information, see https://pre-commit.ci
|
@lgienapp I just fixed some typing, but suddenly, all tests are failing, Do you mind checking what is wrong... 🦩 |
|
Some unit tests check for error messages via regexes, and these had to be adapted to the typo fixes as well. Should all pass now. |
Co-authored-by: Daniel Stancl <46073029+stancld@users.noreply.github.com>
for more information, see https://pre-commit.ci
What does this PR do?
Fixes #2371
This PR adds a new kind of metric wrapper for data transformations. Transformations allow for modifications to the input a metric receives by wrapping its
predandtargetarguments. Transformations can be implemented by either subclassing theMetricInputTransformerbase class and overriding the.transform_pred()and/or.transform_target()functions, or by supplying a lambda function via theLambdaInputTransformer. ABinaryTargetTransformerwhich casts target labels to 0/1 given a threshold is provided for convenience.Before submitting
PR review
Anyone in the community is free to review the PR once the tests have passed.
If we didn't discuss your PR in Github issues there's a high chance it will not be merged.
📚 Documentation preview 📚: https://torchmetrics--2392.org.readthedocs.build/en/2392/