发表机构
EPFL; NVIDIA; Stanford University（洛桑联邦理工学院; 英伟达公司; 斯坦福大学）
机构由 AI 辅助整理，请以论文原文为准。AI 中文总结
针对少步视频生成的高计算成本问题，提出参数化轨迹蒸馏（PTD）方法，在Wan2.1-14B和MiniMax-H3数据集上均优于现有SOTA方法，获人类投票更高偏好。
AI 中文摘要
视频扩散模型和流模型需要多次顺序评估，导致生成计算成本高昂。少步蒸馏可降低该成本，但会引发容量分配问题：学生模型需在更少的顺序计算量下匹配教师模型的迭代生成。现有轨迹方法要求学生模型在高噪声下复现教师模型高度弯曲的转换，这可能超出其容量并导致精细细节退化。我们提出参数化轨迹蒸馏（Parametric Trajectory Distillation, PTD），该方法让学生模型将教师模型的轨迹段参数化为多项式，并沿自身预测路径学习教师模型的指导。PTD旨在让学习到的曲率适配主干网络的预测容量，从而保留运动和多样性。曲率头仅在训练时使用，推理阶段保留原始主干网络架构。在Wan2.1-14B模型上，四步PTD在轨迹蒸馏任务中达到新的SOTA，在相同训练设置下，相比该模型上表现最佳的纯轨迹方法PDD，显著提升了动态质量和自然度。在33B音视频MiniMax-H3数据集上，经LoRA训练的PTD相比SOTA方法LightX2V Turbo，显著提升了多样性和自然度。盲测人类投票显示，PTD对PDD和LightX2V Turbo的偏好率分别为55.1%和63.4%。项目页面：this https URL。
英文摘要
Video diffusion and flow models require many sequential evaluations, making generation computationally expensive. Few-step distillation reduces this cost but poses a capacity allocation problem: a student must match the teacher's iterative generation with far less sequential computation. Existing trajectory methods ask the student to reproduce teacher transitions that are highly curved at high noise, which can exceed its capacity and degrade fine detail. We introduce Parametric Trajectory Distillation (PTD), which lets the student parameterize teacher trajectory segments as polynomials and learn from teacher guidance along its own predicted path. PTD is designed to let the learned curvature adapt to the backbone's predictive capacity, preserving motion and diversity. The curvature head is used only in training; inference keeps the original backbone architecture. On Wan2.1-14B, four-step PTD sets a new state of the art for trajectory distillation, significantly improving dynamic quality and naturalness over PDD, the best-performing trajectory-only method on this model, under the same training setting. On the 33B audio-video MiniMax-H3, LoRA-trained PTD significantly improves diversity and naturalness over the state-of-the-art LightX2V Turbo. Blinded human votes give PTD 55.1% and 63.4% preference shares against PDD and LightX2V Turbo. Project page: https://alan-lanfeng.github.io/PTD/.
