fixing PYTORCH_PRETRAINED_BERT_CACHE use in examples

examples/run_classifier.py

@@ -495,7 +495,7 @@ def main():
             num_train_optimization_steps = num_train_optimization_steps // torch.distributed.get_world_size()
 
     # Prepare model
-    cache_dir = args.cache_dir if args.cache_dir else os.path.join(PYTORCH_PRETRAINED_BERT_CACHE, 'distributed_{}'.format(args.local_rank))
+    cache_dir = args.cache_dir if args.cache_dir else os.path.join(str(PYTORCH_PRETRAINED_BERT_CACHE), 'distributed_{}'.format(args.local_rank))
     model = BertForSequenceClassification.from_pretrained(args.bert_model,
               cache_dir=cache_dir,
               num_labels = num_labels)

examples/run_squad.py

@@ -894,7 +894,7 @@ def main():
 
     # Prepare model
     model = BertForQuestionAnswering.from_pretrained(args.bert_model,
-                cache_dir=os.path.join(PYTORCH_PRETRAINED_BERT_CACHE, 'distributed_{}'.format(args.local_rank)))
+                cache_dir=os.path.join(str(PYTORCH_PRETRAINED_BERT_CACHE), 'distributed_{}'.format(args.local_rank)))
 
     if args.fp16:
         model.half()

examples/run_swag.py

@@ -367,7 +367,7 @@ def main():
 
     # Prepare model
     model = BertForMultipleChoice.from_pretrained(args.bert_model,
-        cache_dir=os.path.join(PYTORCH_PRETRAINED_BERT_CACHE, 'distributed_{}'.format(args.local_rank)),
+        cache_dir=os.path.join(str(PYTORCH_PRETRAINED_BERT_CACHE), 'distributed_{}'.format(args.local_rank)),
         num_choices=4)
     if args.fp16:
         model.half()