LoL: 长于更长，将视频生成扩展至小时级
LoL: Longer than Longer, Scaling Video Generation to Hour
AI总结：
LoL提出一种无需训练的方法，通过多头RoPE抖动缓解视频生成中的sink-collapse问题，实现长时长、高质量的流式视频生成。
摘要：
近期关于长视频生成的研究从双向模型转向自回归模型，但这些方法通常会遭受误差累积和长期一致性丧失的问题。尽管引入了注意力sink帧来缓解这一性能下降，但它们往往导致我们称之为sink-collapse的临界故障模式：生成的内容反复回到sink帧，导致场景重置和循环运动模式。我们的分析表明，sink-collapse源于旋转位置嵌入（RoPE）的周期性结构与当前生成模型中普遍存在的多头注意力机制之间的内在冲突。为了解决这个问题，我们提出了一种轻量且无需训练的方法，通过引入多头RoPE抖动来抑制这种行为，打破头间注意力同质化并缓解长周期崩溃。大量实验显示，我们的方法有效缓解了sink-collapse，同时保持了生成质量。据我们所知，这项工作实现了首次实时、流式和无限长度视频生成的演示，且质量衰减很小。作为这种鲁棒性的示例，我们生成了长达12小时的连续视频，这在公开展示的流式视频生成结果中属于最长的之一。
英文摘要：
Recent research in long-form video generation has shifted from bidirectional to autoregressive models, yet these methods commonly suffer from error accumulation and a loss of long-term coherence. While attention sink frames have been introduced to mitigate this performance decay, they often induce a critical failure mode we term sink-collapse: the generated content repeatedly reverts to the sink frame, resulting in abrupt scene resets and cyclic motion patterns. Our analysis reveals that sink-collapse originates from an inherent conflict between the periodic structure of Rotary Position Embedding (RoPE) and the multi-head attention mechanisms prevalent in current generative models. To address it, we propose a lightweight, training-free approach that effectively suppresses this behavior by introducing multi-head RoPE jitter that breaks inter-head attention homogenization and mitigates long-horizon collapse. Extensive experiments show that our method successfully alleviates sink-collapse while preserving generation quality. To the best of our knowledge, this work achieves the first demonstration of real-time, streaming, and infinite-length video generation with little quality decay. As an illustration of this robustness, we generate continuous videos up to 12 hours in length, which, to our knowledge, is among the longest publicly demonstrated results in streaming video generation.
机构、总结、中文摘要均由AI分析而来，可能存在错误，请以论文正文为准。
