NbAiLab
/

nb-wav2vec2-300m-nynorsk

Automatic Speech Recognition

Norwegian Nynorsk

Inference Endpoints

Model card Files Files and versions Metrics Training metrics Community

versae commited on Jan 31, 2023

Commit

7ca8c95

•

1 Parent(s): 9e1f9cf

Update eval.py

Files changed (1) hide show

eval.py +3 -3

eval.py CHANGED Viewed

@@ -130,9 +130,9 @@ def normalize_text(original_text: str, dataset: str) -> str:
         text = re.sub('[úùüû]', 'u', text)
         text = re.sub('[«»]', '', text)
         text = re.sub('\s+', ' ', text)
-    text = re.sub('<ee>', 'eee', text)
-    text = re.sub('<qq>', 'qqq', text)
-    text = re.sub('<mm>', 'mmm', text)
     text = re.sub('<inaudible>', 'xxx', text)
     text = re.sub('[<>]', '', text)

         text = re.sub('[úùüû]', 'u', text)
         text = re.sub('[«»]', '', text)
         text = re.sub('\s+', ' ', text)
+    text = re.sub('<e+h?>', 'eee', text)
+    text = re.sub('<m+>', 'mmm', text)
+    text = re.sub('<q+>', 'qqq', text)
     text = re.sub('<inaudible>', 'xxx', text)
     text = re.sub('[<>]', '', text)