VLE: Vision-Language Encoder (VLE: 视觉-语言多模态预训练模型) - View it on GitHub