hello, I was wondering if there is any way to take a model from AutoModelForCausalLM and fine-tune it in a Seq2Seq task