Update README.md

README.md

@@ -50,6 +50,7 @@ You can use the model directly with a pipeline for masked language modeling:
 Or you can load the model and tokenizer and do what you need to do:
 
 ```python
+>>> from transformers import AutoTokenizer, AutoModelForMaskedLM
 >>> tokenizer = AutoTokenizer.from_pretrained("DmitryPogrebnoy/MedRuRobertaLarge")
 >>> model = AutoModelForMaskedLM.from_pretrained("DmitryPogrebnoy/MedRuRobertaLarge")
 ```