无需训练的高效图像生成方法，采样速度提升5倍且质量更高
DBMSolver: A Training-free Diffusion Bridge Sampler for High-Quality Image-to-Image Translation
- 利用扩散桥模型的半线性结构，通过指数积分器实现快速采样
- 20次函数评估下FID降低53%，在多种任务中达到新最优效率-质量平衡
- 适合需要高速高质图像生成的实时应用，如图像修复与风格迁移
基于扩散的图像到图像翻译在高保真生成方面表现优异，但现有扩散桥模型（DBMs）采样速度慢，通常需数十次函数评估（NFE）。本文提出DBMSolver，一种无需训练的采样器，利用DBM底层随机微分方程（SDE）与常微分方程（ODE）的半线性结构，采用指数积分器获得一阶与二阶高效解法。该方法将NFE减少最多5倍，同时提升生成质量（例如在DIODE数据集上20 NFE时FID下降53%）。在修复、风格迁移和语义到图像等任务中，跨分辨率至256×256的实验表明，DBMSolver实现了新的效率-质量权衡，推动了实际应用落地。代码已开源于https://github.com/snumprlab/dbmsolver。
原文摘要 · Abstract (English)
Diffusion-based image-to-image (I2I) translation excels in high-fidelity generation but suffers from slow sampling in state-of-the-art Diffusion Bridge Models (DBMs), often requiring dozens of function evaluations (NFEs). We introduce DBMSolver, a training-free sampler that exploits the semi-linear structure of DBM's underlying SDE and ODE via exponential integrators, yielding highly-efficient 1st- and 2nd-order solutions. This reduces NFEs by up to 5x while boosting quality (e.g., FID drops 53% on DIODE at 20 NFEs vs. 2nd-order baseline). Experiments on inpainting, stylization, and semantics-to-image tasks across resolutions up to 256x256 show DBMSolver sets new SOTA efficiency-quality tradeoffs, enabling real-world applicability. Our code is publicly available at https://github.com/snumprlab/dbmsolver.
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。