Update README.md

README.md

@@ -48,10 +48,10 @@ For a list of individual model parameters, refer to the `config.json` file in th
 ## Using the models
 
 ```python
-from transformers import T5Tokenizer, T5ForConditionalGeneration
+from transformers import AutoTokenizer, AutoModelForSeq2SeqLM
 
-tokenizer = T5Tokenizer.from_pretrained("gsarti/it5-base")
-model = T5ForConditionalGeneration.from_pretrained("gsarti/it5-base")
+tokenizer = AutoTokenizer.from_pretrained("gsarti/it5-base")
+model = AutoModelForSeq2SeqLM.from_pretrained("gsarti/it5-base")
 ```
 
 *Note: You will need to fine-tune the model on your downstream seq2seq task to use it. See an example [here](https://huggingface.co/gsarti/it5-base-nli).*