MemRoPE：通过演化记忆标记实现无训练的无限视频生成
MemRoPE: Training-Free Infinite Video Generation via Evolving Memory Tokens
AI总结：
MemRoPE通过演化记忆标记和在线RoPE索引机制，解决长视频生成中的时间一致性、视觉保真度和主体一致性问题，实现无训练的无限视频生成。
摘要：
自回归扩散模型能够实现实时帧流，但现有滑动窗口缓存丢弃过去上下文，导致保真度下降、身份漂移和运动停滞。当前方法保留固定数量的早期标记作为注意力 sinks，但这种静态锚点无法反映视频内容的演变。我们引入MemRoPE，一个无训练框架，包含两个协同设计的组件。记忆标记通过指数移动平均持续压缩所有过去的键，形成双流长短期流，保持全局身份和近期动态在固定大小的缓存中。在线RoPE索引缓存未旋转的键并在注意力时间动态应用位置嵌入，确保聚合过程不受冲突的位置相位影响。这两种机制相互促进：位置解耦使时间聚合定义明确，而聚合使固定大小缓存可用于无限生成。大量实验验证MemRoPE在分钟到小时尺度的生成中在时间一致性、视觉保真度和主体一致性方面均优于现有方法。
英文摘要：
Autoregressive diffusion enables real-time frame streaming, yet existing sliding-window caches discard past context, causing fidelity degradation, identity drift, and motion stagnation over long horizons. Current approaches preserve a fixed set of early tokens as attention sinks, but this static anchor cannot reflect the evolving content of a growing video. We introduce MemRoPE, a training-free framework with two co-designed components. Memory Tokens continuously compress all past keys into dual long-term and short-term streams via exponential moving averages, maintaining both global identity and recent dynamics within a fixed-size cache. Online RoPE Indexing caches unrotated keys and applies positional embeddings dynamically at attention time, ensuring the aggregation is free of conflicting positional phases. These two mechanisms are mutually enabling: positional decoupling makes temporal aggregation well-defined, while aggregation makes fixed-size caching viable for unbounded generation. Extensive experiments validate that MemRoPE outperforms existing methods in temporal coherence, visual fidelity, and subject consistency across minute- to hour-scale 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。
