发表机构
Seoul National University; Korea University; Sungkyunkwan University; Georgia Institute of Technology（首尔大学; 高丽大学; 成均馆大学; 佐治亚理工学院）
机构由 AI 辅助整理，请以论文原文为准。AI 中文总结
针对长视频生成中的漂移问题，提出ID-Forcing测试时框架，通过自缓存机制确保KV条目分布内，无缝扩展短时域模型至分钟级视频，显著缓解漂移。
AI 中文摘要
现代自回归（AR）视频扩散模型擅长短时域视频生成，但生成长时间视频仍具挑战性，原因在于漂移现象，即颜色和纹理发生变化，以及运动动态衰减。现有工作主要依赖KV条件化，通过选择或修改缓存的键值（KV）条目来缓解漂移。然而，我们观察到仅KV条件化是不够的，因为它假设缓存的KV条目保持在分布内。这一假设在训练时域之外失效：在生成过程中没有任何约束KV条目的构建，从而引发KV来源问题，即缓存的条目本身变得分布外（OOD）。为解决此问题，我们提出分布内强制（ID-Forcing），一种测试时框架，使KV缓存和KV条件化与训练配置对齐。其关键机制——自缓存，从源头防止OOD KV条目。每个块在缓存时不参考先前的KV条目，使滚动窗口精确保持在分布内。因此，ID-Forcing无缝地将短时域模型扩展到分钟级视频生成。广泛评估表明，我们的方法在标准视频生成基准上保持竞争力，同时在缓解漂移方面大幅优于先前工作，这通过我们的漂移指标和用户研究得到验证。
英文摘要
Modern autoregressive (AR) video diffusion models excel at short-horizon video generation, yet generating long videos remains challenging due to drifting, where colors and textures shift, and motion dynamics decay. Existing works primarily rely on KV conditioning, which selects or modifies cached key-value (KV) entries to mitigate drifting. However, we observe that KV conditioning alone is insufficient as it assumes cached KV entries remain in-distribution. This assumption fails beyond the training horizon: nothing constrains the construction of KV entries during rollout, giving rise to the KV-provenance problem where cached entries themselves become out-of-distribution (OOD). To address this, we propose In-Distribution Forcing (ID-Forcing), a test-time framework that aligns both KV caching and KV conditioning with training configurations. Its key mechanism, self-caching, prevents OOD KV entries at their source. Each chunk is cached without attending to prior KV entry, keeping the rolling window exactly in-distribution. Consequently, ID-Forcing seamlessly extends short-horizon models to minute-scale video generation. Extensive evaluations show that our method remains competitive on standard video generation benchmark while substantially outperforming prior work in mitigating drifting, as validated by both our drift metrics and a user study.
Commentsproject page: https://in-distribution-forcing.github.io/
