Emu2

Second-generation multimodal model from BAAI, advancing vision-language generation with improved architecture and training. Supports any-to-any generation across text and images.