fix input in run_glue for distilbert

examples/run_glue.py

@@ -134,8 +134,9 @@ def train(args, train_dataset, model, tokenizer):
             batch = tuple(t.to(args.device) for t in batch)
             inputs = {'input_ids':      batch[0],
                       'attention_mask': batch[1],
-                      'token_type_ids': batch[2] if args.model_type in ['bert', 'xlnet'] else None,  # XLM, DistilBERT and RoBERTa don't use segment_ids
                       'labels':         batch[3]}
+            if args.model_type != 'distilbert':
+                inputs['token_type_ids'] = batch[2] if args.model_type in ['bert', 'xlnet'] else None  # XLM, DistilBERT and RoBERTa don't use segment_ids
             outputs = model(**inputs)
             loss = outputs[0]  # model outputs are always tuple in transformers (see doc)
 
@@ -224,8 +225,9 @@ def evaluate(args, model, tokenizer, prefix=""):
             with torch.no_grad():
                 inputs = {'input_ids':      batch[0],
                           'attention_mask': batch[1],
-                          'token_type_ids': batch[2] if args.model_type in ['bert', 'xlnet'] else None,  # XLM, DistilBERT and RoBERTa don't use segment_ids
                           'labels':         batch[3]}
+                if args.model_type != 'distilbert':
+                    inputs['token_type_ids'] = batch[2] if args.model_type in ['bert', 'xlnet'] else None  # XLM, DistilBERT and RoBERTa don't use segment_ids
                 outputs = model(**inputs)
                 tmp_eval_loss, logits = outputs[:2]