Make compatible with newer version of sklearn

fastdatascience · Feb 14, 2024 · 80f37fb · 80f37fb
1 parent 72530c5
commit 80f37fb
Showing 1 changed file with 1 addition and 1 deletion.
diff --git a/train/train_sap_classifier.py b/train/train_sap_classifier.py
@@ -88,7 +88,7 @@ def upsample(features, labels):
 
 
 # Specially engineered regex to include 95%, 95%ci, etc
-vectoriser = CountVectorizer(lowercase=True, stop_words=stops, min_df=5, max_features=NUM_FEATURES,
+vectoriser = CountVectorizer(lowercase=True, stop_words=list(stops), min_df=5, max_features=NUM_FEATURES,
                              token_pattern=r'[59][05]%?(?:ci)?|[a-z][a-z]+')
 transformer = TfidfTransformer()