直接优化微分方程求解误差，反而让一致性模型生成效果变差。
Inconsistencies In Consistency Models: Better ODE Solving Does Not Imply Better Samples
- 直接最小化概率流微分方程求解误差，提升数值精度。
- 数值误差降低但样本质量显著下降，与直觉相反。
- 揭示了当前一致性模型的隐性工作机制，适合研究生成机制者阅读。
尽管扩散模型能生成高质量样本，但其迭代采样过程成本高昂。一致性模型（CMs）作为扩散模型的蒸馏方法，通过少量迭代即可生成高保真样本，显著降低采样开销。其核心是求解现有扩散模型定义的概率流常微分方程（ODE）。然而，CMs并未直接训练以最小化与ODE求解器的误差，而是采用更易计算的目标函数。为探究CMs对概率流ODE的求解能力及其误差对样本质量的影响，本文提出直接一致性模型（Direct CMs），直接最小化该误差。令人意外的是，虽然Direct CMs的ODE求解误差低于原始CMs，但生成样本质量却明显更差，这挑战了我们对一致性模型为何有效的理解。完整代码已公开于：https://github.com/layer6ai-labs/direct-cms。
原文摘要 · Abstract (English)
Although diffusion models can generate remarkably high-quality samples, they are intrinsically bottlenecked by their expensive iterative sampling procedure. Consistency models (CMs) have recently emerged as a promising diffusion model distillation method, reducing the cost of sampling by generating high-fidelity samples in just a few iterations. Consistency model distillation aims to solve the probability flow ordinary differential equation (ODE) defined by an existing diffusion model. CMs are not directly trained to minimize error against an ODE solver, rather they use a more computationally tractable objective. As a way to study how effectively CMs solve the probability flow ODE, and the effect that any induced error has on the quality of generated samples, we introduce Direct CMs, which \textit{directly} minimize this error. Intriguingly, we find that Direct CMs reduce the ODE solving error compared to CMs but also result in significantly worse sample quality, calling into question why exactly CMs work well in the first place. Full code is available at: https://github.com/layer6ai-labs/direct-cms.
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。