Fine-tuning of Large Language Models for Constituency Parsing Using a Sequence to Sequence Approach