Add get_vocab method to PretrainedTokenizer

src/transformers/tokenization_ctrl.py

@@ -147,6 +147,9 @@ class CTRLTokenizer(PreTrainedTokenizer):
     def vocab_size(self):
         return len(self.encoder)
 
+    def get_vocab(self):
+        return dict(self.encoder, **self.added_tokens_encoder)
+
     def bpe(self, token):
         if token in self.cache:
             return self.cache[token]