v5 update to match TokenizersBackend to AutoTokenizer

tokenizer_config.json

@@ -309,6 +309,6 @@
   "pad_token": "<pad>",
   "sp_model_kwargs": {},
   "spaces_between_special_tokens": false,
-  "tokenizer_class": "T5Tokenizer",
+  "tokenizer_class": "PreTrainedTokenizerFast",
   "unk_token": "<unk>"
 }