Update tokenization_xgen.py

tokenization_xgen.py

@@ -227,4 +227,8 @@ class XgenTokenizer(PreTrainedTokenizer):
         if token_ids_1 is not None:
             output += [1] * len(token_ids_1 + eos_token_id)
 
-        return output
+        return output
+
+    # has no vocab file
+    def save_vocabulary(self, save_directory: str, filename_prefix: Optional[str] = None):
+        return ()