Fetching the paper…

UniWorld-V1: High-Resolution Semantic Encoders for Unified Visual Understanding and Generation · Around