解决流匹配生成中速度不足问题，提升采样效率与质量
The Velocity Deficit: Initial Energy Injection for Flow Matching
- 通过初始能量注入纠正速度低估，缓解轨迹起点动能停滞
- 无需重训练，单行代码实现5倍加速，ImageNet上FID降低44.6%
- 适用于文本到图像、高分辨率生成，通用性强且效果显著
尽管流匹配理论上保证恒定速度轨迹，但在高维实践中存在关键缺陷：速度低估现象（称为速度赤字）。我们发现，均方误差目标系统性低估速度幅度，导致生成样本无法抵达数据流形，即积分滞后。为此，提出两种互补方法：基于训练的幅度感知流匹配（MAFM）和免训练的尺度调度修正器（SSC）。二者基于一个关键发现：速度收缩在轨迹起始导致有害动能停滞，而在末端则具有有益去噪作用。实验表明，SSC无需重训练，仅需一行代码即可带来显著效率提升。在ImageNet-1k（256×256）上，FID从13.68降至7.58（改善44.6%），实现5倍加速；50步生成器（FID 7.58）超越250步基线（FID 8.65）。方法还推广至文本到图像任务与高分辨率生成，在MS-COCO上FID改善约22%。
原文摘要 · Abstract (English)
While Flow Matching theoretically guarantees constant-velocity trajectories, we identify a critical breakdown in high-dimensional practice: the Velocity Deficit. We show that the MSE objective systematically underestimates velocity magnitude, causing generated samples to fail to reach the data manifold-a phenomenon we term Integration Lag. To rectify this, we propose Initial Energy Injection, instantiated via two complementary methods: the training-based Magnitude-Aware Flow Matching (MAFM) and the training-free Scale Schedule Corrector (SSC). Both are grounded in our discovery of a crucial asymmetry: velocity contraction causes harmful kinetic stagnation at the trajectory's start, yet acts as a beneficial denoising mechanism at its end. Empirically, SSC yields significant efficiency gains with zero retraining and just one line of code. On ImageNet-1k (256x256), it improves FID by 44.6% (from 13.68 to 7.58) and achieves a 5x speedup, enabling a 50-step generator (FID 7.58) to beat a 250-step baseline (FID 8.65). Furthermore, our methods generalize to Text-to-Image tasks and high-resolution generation, improving FID on MS-COCO by ~22%.
Thank you to arXiv for use of its open access interoperability. PaperDance 不是 arXiv 官方产品；中文卡片由大模型生成，请以原文为准。