Infinity-RoPE: 自动回归自回滚中涌现的无限视频生成
Infinity-RoPE: Action-Controllable Infinite Video Generation Emerges From Autoregressive Self-Rollout
AI总结：
本文提出Infinity-RoPE，通过Block-Relativistic RoPE、KV Flush和RoPE Cut三个组件解决自回归视频扩散模型的三个瓶颈，实现无限时域、可控和电影级视频生成。
摘要：
当前自回归视频扩散模型受限于三个核心瓶颈：（i）基础模型3D-RoPE imposed的有限时间跨度；（ii）在长形式回滚中维持细粒度动作控制的缓慢响应；（iii）无法在单次生成流中实现不连续电影过渡。我们引入$\infty$-RoPE，一个统一的推理时间框架，通过三个相互关联的组件解决这三个限制：Block-Relativistic RoPE、KV Flush和RoPE Cut。Block-Relativistic RoPE将时间编码重新表述为移动的局部参考系，其中每个新生成的潜在块相对于基础模型的最大帧跨度旋转，而早期块向后旋转以保持相对时间几何。这种相对论式表述消除了固定时间位置，使视频生成能够远超基础位置限制。为获得细粒度动作控制而无需重新编码，KV Flush通过保留两个潜在帧——全局sink和最后生成的潜在帧——刷新KV缓存，从而确保即时提示响应。最后，RoPE Cut在时间RoPE坐标中引入受控的不连续性，允许在单次连续回滚中进行多切场景过渡。这些组件共同建立了$\infty$-RoPE作为无限时域、可控和电影级视频扩散的训练免费基础。全面实验表明，$\infty$-RoPE在总体VBench分数上始终优于先前的自回归模型。
英文摘要：
Current autoregressive video diffusion models are constrained by three core bottlenecks: (i) the finite temporal horizon imposed by the base model's 3D Rotary Positional Embedding (3D-RoPE), (ii) slow prompt responsiveness in maintaining fine-grained action control during long-form rollouts, and (iii) the inability to realize discontinuous cinematic transitions within a single generation stream. We introduce $\infty$-RoPE, a unified inference-time framework that addresses all three limitations through three interconnected components: Block-Relativistic RoPE, KV Flush, and RoPE Cut. Block-Relativistic RoPE reformulates temporal encoding as a moving local reference frame, where each newly generated latent block is rotated relative to the base model's maximum frame horizon while earlier blocks are rotated backward to preserve relative temporal geometry. This relativistic formulation eliminates fixed temporal positions, enabling continuous video generation far beyond the base positional limits. To obtain fine-grained action control without re-encoding, KV Flush renews the KV cache by retaining only two latent frames, the global sink and the last generated latent frame, thereby ensuring immediate prompt responsiveness. Finally, RoPE Cut introduces controlled discontinuities in temporal RoPE coordinates, enabling multi-cut scene transitions within a single continuous rollout. Together, these components establish $\infty$-RoPE as a training-free foundation for infinite-horizon, controllable, and cinematic video diffusion. Comprehensive experiments show that $\infty$-RoPE consistently surpasses previous autoregressive models in overall VBench scores.
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。
