tokenizer: fix decode() to handle torch tensor input via .tolist()

tokenizer.py

@@ -345,6 +345,8 @@ class HybridDNATokenizer(PreTrainedTokenizer):
         skip_special_tokens: bool = False,
         **kwargs
     ) -> str:
+        if hasattr(token_ids, 'tolist'):
+            token_ids = token_ids.tolist()
         if isinstance(token_ids, int):
             token_ids = [token_ids]