Update eval.py

eval.py

@@ -130,11 +130,10 @@ 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('<e+h?>', 'eee', text)
-    text = re.sub('<m+>', 'mmm', text)
-    text = re.sub('<q+>', 'qqq', 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)
 
     # # In addition, we can normalize the target text, e.g. removing new lines characters etc...
     # # note that order is important here!
@@ -143,7 +142,7 @@ def normalize_text(original_text: str, dataset: str) -> str:
     # for t in token_sequences_to_ignore:
     #     text = " ".join(text.split(t))
 
-    return text.strip()
+    return text
 
 
 def main(args):