Fix metric docstring (#5475)

* fix docstring * changelog * Update allennlp/training/metrics/fbeta_multi_label_measure.py Co-authored-by: Pete <petew@allenai.org> Co-authored-by: Pete <petew@allenai.org>
allenai · Nov 23, 2021 · 0a794c6 · 0a794c6
1 parent f86ff9f
commit 0a794c6
Show file tree

Hide file tree

Showing 2 changed files with 7 additions and 1 deletion.
diff --git a/CHANGELOG.md b/CHANGELOG.md
@@ -7,6 +7,10 @@ and this project adheres to [Semantic Versioning](https://semver.org/spec/v2.0.0
 
 ## Unreleased
 
+### Fixed
+
+- Fixed the docstring information for the `FBetaMultiLabelMeasure` metric.
+
 ## [v2.8.0](https://github.com/allenai/allennlp/releases/tag/v2.8.0) - 2021-11-01
 
 ### Added

diff --git a/allennlp/training/metrics/fbeta_multi_label_measure.py b/allennlp/training/metrics/fbeta_multi_label_measure.py
@@ -61,7 +61,9 @@ class FBetaMultiLabelMeasure(FBetaMeasure):
         in the data will result in 0 components in a macro or weighted average.
 
     threshold: `float`, optional (default = `0.5`)
-        Logits over this threshold will be considered predictions for the corresponding class.
+        Probabilities over this threshold will be considered predictions for the corresponding class.
+        Note that you can also use this metric with logits, in which case it would make more sense to set
+        the `threshold` value to `0.0`.
 
     """