Qianfan-VL: Domain-Enhanced Universal Vision-Language Models - View it on GitHub