Update multiclass_specificity_weighted.py

multiclass_specificity_weighted.py

@@ -86,19 +86,19 @@ class multiclass_specificity_weighted(evaluate.Metric):
         # TODO: Download external resources if needed
         pass
 
-    def _compute(self, y_pred, y_true):
+    def _compute(self, predictions, references):
         import numpy as np
         """Returns the scores"""
         # TODO: Compute the different scores of the module
-        unique_classes = np.unique(y_true)
+        unique_classes = np.unique(predictions)
         num_classes = len(unique_classes)
         specificity = np.zeros(num_classes)
-        class_counts = np.bincount(y_true)
-        total_samples = len(y_true)
+        class_counts = np.bincount(predictions)
+        total_samples = len(predictions)
 
         for i, class_label in enumerate(unique_classes):
-            true_negative = sum((y_true != class_label) & (y_pred != class_label))
-            total_negative = sum(y_true != class_label)
+            true_negative = sum((predictions != class_label) & (references != class_label))
+            total_negative = sum(predictions != class_label)
             
             if total_negative != 0:
                 specificity[i] = true_negative / total_negative
@@ -106,7 +106,6 @@ class multiclass_specificity_weighted(evaluate.Metric):
                 specificity[i] = 0.0
 
         weighted_specificity = np.sum(specificity * (class_counts / total_samples))
-        macro_specificity = np.mean(specificity)
 
         return {
             "weighted_specificity": weighted_specificity,