提出最优步长调度方法，实现高效高质图像生成。
Optimal Stepsize for Diffusion Sampling
- 基于动态规划构建步长优化框架，通过参考轨迹知识蒸馏获得最优调度。
- 实现10倍加速文本到图像生成，性能保留99.4%（GenEval）。
- 对模型架构、求解器和噪声调度均具强鲁棒性，适合高效生成场景。
扩散模型虽生成质量优异，但因步长离散化不优导致采样计算开销大。现有工作多聚焦于优化去噪方向，本文则关注步长调度的理论设计。提出最优步长蒸馏方法，利用动态规划框架从参考轨迹中提取理论上最优的调度策略。将步长优化重新建模为递归误差最小化问题，通过最优子结构利用确保全局离散化误差边界。关键优势在于，所获调度在不同模型架构、ODE求解器及噪声调度下均表现出强鲁棒性。实验表明，该方法可实现10倍加速的文本到图像生成，同时在GenEval上保持99.4%的性能。代码已开源：https://github.com/bebebe666/OptimalSteps。
原文摘要 · Abstract (English)
Diffusion models achieve remarkable generation quality but suffer from computational intensive sampling due to suboptimal step discretization. While existing works focus on optimizing denoising directions, we address the principled design of stepsize schedules. This paper proposes Optimal Stepsize Distillation, a dynamic programming framework that extracts theoretically optimal schedules by distilling knowledge from reference trajectories. By reformulating stepsize optimization as recursive error minimization, our method guarantees global discretization bounds through optimal substructure exploitation. Crucially, the distilled schedules demonstrate strong robustness across architectures, ODE solvers, and noise schedules. Experiments show 10x accelerated text-to-image generation while preserving 99.4% performance on GenEval. Our code is available at https://github.com/bebebe666/OptimalSteps.
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。