Fetching the paper…

Lumen: Unleashing Versatile Vision-Centric Capabilities of Large Multimodal Models · Around