We have trained a Xiaoyi dialogue robot model for the driving training field based on the Qwen1.5-7b model, using the method of QLora fine-tuning.