Tencent Releases WeMM-Embedding for Multimodal Retrieval

Tencent’s WeChat Vision team releases three WeMM-Embedding models for text, image, video and document retrieval, with evaluation and serving code.