Correctly report AA precision and recall during validation

Fixes #252.
Noble-Lab · Oct 13, 2023 · 16c95dd · 16c95dd
1 parent 86630e3
commit 16c95dd
Show file tree

Hide file tree

Showing 2 changed files with 2 additions and 1 deletion.
diff --git a/CHANGELOG.md b/CHANGELOG.md
@@ -30,6 +30,7 @@ The format is based on [Keep a Changelog](https://keepachangelog.com/en/1.0.0/),
 - Casanovo runs on CPU and can passes all tests.
 - Enable gradients during prediction and validation to avoid NaNs from occuring as a temporary workaround until a new Pytorch version is available.
 - Upgrade to depthcharge v0.2.3 for `PeptideTransformerDecoder` hotfix.
+- Correctly report amino acid precision and recall during validation.
 
 ### Removed
 

diff --git a/casanovo/denovo/model.py b/casanovo/denovo/model.py
@@ -773,8 +773,8 @@ def validation_step(
 
         aa_precision, _, pep_precision = evaluate.aa_match_metrics(
             *evaluate.aa_match_batch(
-                peptides_pred,
                 peptides_true,
+                peptides_pred,
                 self.decoder._peptide_mass.masses,
             )
         )