提出对比正则化方法，让生成模型更稳更准，尤其适合低步数场景。
VeCoR -- Velocity Contrastive Regularization for Flow Matching
- 引入正负双向监督，指导生成方向并避开错误路径
- 在ImageNet上降低35%的FID，MS-COCO文本生成提升32%
- 特别适合轻量级和少步数生成任务，提升图像质量与稳定性
流匹配（Flow Matching, FM）是扩散模型的高效替代方案。标准FM仅引导速度场朝向目标方向，但易积累误差，导致样本偏离数据流形，尤其在轻量或低步数设置下出现感知退化。为此，本文提出速度对比正则化（VeCoR），在标准FM目标基础上增加对比性双向监督：一方面对齐预测速度与稳定参考方向（正样本），另一方面推开不一致、离流形的方向（负样本）。该机制将单一吸引式目标转化为双侧正则信号，有效规范轨迹演化，提升跨数据集与模型的生成稳定性与保真度。在ImageNet-1K 256×256上，使用SiT-XL/2与REPA-SiT-XL/2骨干网络时，分别实现35%和22%的相对FID下降；在MS-COCO文生图任务中进一步取得32%相对提升，验证了其在低步数与轻量配置下的持续优势。
原文摘要 · Abstract (English)
Flow Matching (FM) has recently emerged as a principled and efficient alternative to diffusion models. Standard FM encourages the learned velocity field to follow a target direction; however, it may accumulate errors along the trajectory and drive samples off the data manifold, leading to perceptual degradation, especially in lightweight or low-step configurations. To enhance stability and generalization, we extend FM into a balanced attract-repel scheme that provides explicit guidance on both "where to go" and "where not to go." To be formal, we propose \textbf{Velocity Contrastive Regularization (VeCoR)}, a complementary training scheme for flow-based generative modeling that augments the standard FM objective with contrastive, two-sided supervision. VeCoR not only aligns the predicted velocity with a stable reference direction (positive supervision) but also pushes it away from inconsistent, off-manifold directions (negative supervision). This contrastive formulation transforms FM from a purely attractive, one-sided objective into a two-sided training signal, regularizing trajectory evolution and improving perceptual fidelity across datasets and backbones. On ImageNet-1K 256$\times$256, VeCoR yields 22\% and 35\% relative FID reductions on SiT-XL/2 and REPA-SiT-XL/2 backbones, respectively, and achieves further FID gains (32\% relative) on MS-COCO text-to-image generation, demonstrating consistent improvements in stability, convergence, and image quality, particularly in low-step and lightweight settings. Project page: https://p458732.github.io/VeCoR_Project_Page/
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。