replacement of a GPT evaluator with an open-source evaluator

[jinz2014]

In the repository, the GPT models are used to automate the evaluation of some input data (e.g. utils/gpt_auto_eval.py). Could you please add an example of using open-source LLM evaluator ? Thanks.