Intern Lumina U2：多码本扩散语言模型统一全视觉理解与生成
原标题：Paper page - Intern Lumina U2: A Multi-Codebook Diffusion Large Language Model for Omni-Visual Understanding and Image Generation
AI 摘要
研究团队发布 Intern Lumina U2，一个基于多码本离散视觉 token 的扩散语言模型，在单一骨干网络中统一图像、视频、3D 的理解、生成与编辑。该模型为 16.9B MoE 架构，仅激活 1.4B 参数，是首个在全离散空间建模多码本视觉 token 的扩散语言模型。通过 Plateau-Halton 采样与 V-verify 策略，图像生成端到端加速 4.45 倍，并在多项理解、生成与编辑基准上取得领先表现，同时在 Ascend 910B 与 NVIDIA 平台完成全栈验证。
正文节选
Intern Lumina U2: A Multi-Codebook Diffusion Large Language Model for Omni-Visual Understanding and Image Generation Abstract Intern Lumina U2 is a diffusion language model that uses multi-codebook discrete visual tokens to unify image, video, and 3D generation with text in a single backbone, achieving high performance and faster inference. Unified multimodal generation covers discrete, continuous, and hybrid paradigms. Discrete approaches with single-codebook visual tokenizers align naturally w