This project fine-tunes large language models (LLMs) using personal WeChat conversation records to create a personalized digital twin. It provides a full chain solution from chat data to model fine-tuning, from text generation to voice cloning, and from training to deployment. Your digital twin will not only 'speak your words' but also 'sound like you'.