Fetching the paper…

UniToken: Harmonizing Multimodal Understanding and Generation through Unified Visual Encoding · Around