DSA: 用于快速自回归视频生成的动态步数分配
DSA: Dynamic Step Allocation for Fast Autoregressive Video Generation
AI总结：
提出一种置信度引导的自适应计算框架DSA，通过轻量级置信度头动态调整每帧去噪步数，在保持视频质量的同时实现实时自回归视频生成。
摘要：
视频扩散变压器已实现最先进的视觉质量，但其高推理成本仍是实时应用的主要瓶颈。最近的蒸馏框架产生了具有降低延迟的自回归视频扩散模型，但这些模型仍然每帧使用固定数量的去噪步数，在可预测帧上浪费计算，而在具有挑战性的帧上精炼不足。我们提出了DSA，一种用于自回归视频扩散的置信度引导自适应计算框架。DSA引入了一个轻量级置信度头，在分布匹配蒸馏目标下与生成器联合训练，以估计每帧去噪可靠性。在推理时，该置信度信号动态调整扩散步数：简单帧提前终止以提高速度，而复杂帧获得额外精炼。我们的方法不需要额外的视频数据、启发式规则，且几乎不需要架构修改。实验表明，DSA实现了实时自回归视频生成，在H100 GPU上达到22.63 FPS，延迟低于1秒，同时与最近的自回归和双向视频扩散模型相比，保持了有竞争力或更优的VBench质量。我们的结果表明，置信度引导的自适应采样为交互式视频生成提供了一条有效且实用的路径。
英文摘要：
Video diffusion transformers have achieved state-of-the-art visual quality, but their high inference cost remains a major bottleneck for real-time applications. Recent distillation frameworks produce autoregressive video diffusion models with reduced latency, yet these models still use a fixed number of denoising steps per frame, wasting computation on predictable frames and under-refining challenging ones. We present DSA, a confidence-guided adaptive computation framework for AR video diffusion. DSA introduces a lightweight confidence head, trained jointly with the generator under a distribution-matching distillation objective, to estimate per-frame denoising reliability. At inference, this confidence signal dynamically adjusts the number of diffusion steps: simple frames terminate early for speed, while complex frames receive additional refinement. Our method requires no extra video data, no heuristics, and little architectural modification. Experiments show that DSA achieves real-time autoregressive video generation, reaching 22.63 FPS with sub-second latency on H100 GPUs, while maintaining competitive or superior VBench quality compared to recent autoregressive and bidirectional video diffusion models. Our results demonstrate that confidence-guided adaptive sampling provides an effective and practical path toward interactive 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。
