Minor Bug Fix for Running Roberta on Glue (#3240)

* added return_token_type_ids argument for tokenizers which do not generate return_type_ids by default

* fixed styling

* Style

Co-authored-by: LysandreJik <lysandre.debut@reseau.eseo.fr>

src/transformers/data/processors/glue.py

@@ -90,7 +90,9 @@ def glue_convert_examples_to_features(
         if ex_index % 10000 == 0:
             logger.info("Writing example %d/%d" % (ex_index, len_examples))
 
-        inputs = tokenizer.encode_plus(example.text_a, example.text_b, add_special_tokens=True, max_length=max_length,)
+        inputs = tokenizer.encode_plus(
+            example.text_a, example.text_b, add_special_tokens=True, max_length=max_length, return_token_type_ids=True,
+        )
         input_ids, token_type_ids = inputs["input_ids"], inputs["token_type_ids"]
 
         # The mask has 1 for real tokens and 0 for padding tokens. Only real