DVAR：面向图像超分的动态视觉自回归建模
DVAR: Dynamic Visual Autoregressive Modeling for Image Super-Resolution
把 VAR 的绝对尺度调度换成与目标尺寸解耦的相对缩放动态，使单一自回归模型可服务任意分辨率超分；其动态采样调度利用码本中视觉 token 的几何邻近性模拟预测误差分布，可迁移到其他自回归生成模型的训练-推理鸿沟缓解。
原文依据 DVAR learns a canonical scaling dynamic. This dynamic effectively decouples the logic of relative scaling from the absolute target size
摘要译文
由模型翻译 · 术语统一 · 请对照原文核对关键结论下一尺度预测范式的视觉自回归（VAR）模型在图像超分辨率任务上展现出显著潜力。然而，其僵化的、与尺寸绑定的设计限制了实际应用。这一局限源于模型依赖对固定的绝对缩放调度的记忆，从而必须为每个目标分辨率单独训练一个模型。我们提出 DVAR，一个动态视觉自回归框架，用以突破这一根本性瓶颈。DVAR 不再记忆这些僵化的调度，而是学习一种规范的缩放动态。该动态有效地将相对缩放的逻辑与绝对目标尺寸解耦，从而在生成步骤之间保持一组统一的比例关系，并可一致地应用于任意尺寸。此外，我们引入动态采样调度器，以可忽略的计算开销缓解教师强制问题。借助码本中视觉 token 的几何邻近性，该调度器高效地模拟模型的预测误差分布，从而弥合训练与推理之间的差距。据我们所知，DVAR 是首个赋予 VAR 模型尺寸灵活性的框架，打破了其对固定分辨率的一对一依赖。大量评测表明，DVAR 相比现有真实世界图像超分方法取得了更优的视觉质量，证明灵活的、纯自回归的方法是通向最优图像超分辨率的一条可行路径。代码见 https://github.com/YuZheng9/DVAR。
英文原文摘要 1513 字符
Next-scale prediction paradigm visual autoregressive (VAR) models have demonstrated significant potential for image super-resolution. However, their practical application is constrained by a rigid, size-specific design. This limitation stems from their reliance on memorizing fixed, absolute scaling schedules, which necessitates a distinct model for each target resolution. We introduce DVAR, a Dynamic Visual AutoRegressive framework that overcomes this fundamental bottleneck. Instead of memorizing these rigid schedules, DVAR learns a canonical scaling dynamic. This dynamic effectively decouples the logic of relative scaling from the absolute target size, thereby preserving a single set of proportions between generative steps that can be applied uniformly to any size. Furthermore, we introduce a dynamic sampling scheduler to mitigate the teacher-forcing problem with negligible computational overhead. By leveraging the geometric proximity of visual tokens in the codebook, it efficiently simulates the model's predictive error distribution to bridge the training-inference gap. To our knowledge, DVAR is the first framework to grant VAR models size-flexibility, breaking their one-to-one dependency on a fixed resolution. Extensive evaluations demonstrate that DVAR achieves superior visual quality over existing Real-ISR methods, proving that a flexible, purely autoregressive approach is a viable path to state-of-the-art image super-resolution. Code is available at https://github.com/YuZheng9/DVAR.