FlowBudget: Training-Free Sampling-Budget Estimation for Flow Matching Models

查看原文 / View Original 编辑 / Edit

Computer Science > Machine Learning arXiv:2610.20636 (cs) [Submitted on 1 Oct 2026] Title: FlowBudget: Training-Free Sampling-Budget Estimation for Flow Matching Models Authors: Zahra Dehghani, Pablo Piantanida, Mohammadhadi Shateri.

计算机科学 > 机器学习 arXiv:2610.20636 (cs) [提交于 2026 年 10 月 1 日] 标题：FlowBudget：流匹配模型的免训练采样预算估计作者：Zahra Dehghani, Pablo Piantanida, Mohammadhadi Shateri。

Abstract: Sampling from continuous flow-based generators such as flow matching models requires numerical ODE or SDE integration, making inference cost depend on the number of function evaluations (NFEs). In practice, the sampling budget is inherited from prior work or selected through expensive FID-versus-NFE sweeps, using FID to assess image-generation quality, requiring sample generation and evaluation at multiple budgets. This makes sampling-budget selection dependent on repeated empirical evaluation.

摘要：从流匹配模型等连续流生成器中进行采样需要数值 ODE 或 SDE 积分，这使得推理成本取决于函数评估次数 (NFE)。在实践中，采样预算通常继承自先前的工作，或是通过昂贵的 FID 与 NFE 扫描来选择——即使用 FID 评估图像生成质量，这需要针对多个预算进行样本生成和评估。这使得采样预算的选择依赖于重复的经验评估。

We therefore ask a complementary question to solver design and timestep allocation: for a fixed pretrained model and solver, what numerical resolution is sufficient? We introduce FLOWBUDGET, a training-free estimator of solver-relative sampling complexity. Using simulation-free interpolation probes of the learned dynamics, FLOWBUDGET normalizes the leading one-step discretization error by the motion, estimates its coefficient from interpolation probes, and maps a numerical tolerance to a step budget using the solver order.

因此，我们提出了一个与求解器设计和时间步长分配互补的问题：对于固定的预训练模型和求解器，什么样的数值分辨率是足够的？我们引入了 FLOWBUDGET，这是一种针对求解器相关采样复杂度的免训练估计器。通过对学习到的动力学进行无需模拟的插值探测，FLOWBUDGET 利用运动对领先的一步离散化误差进行归一化，从插值探测中估计其系数，并利用求解器阶数将数值容差映射为步长预算。

We derive estimators for numerical solvers in flow matching sampling, namely Euler and Heun for ODE sampling and additive-noise Euler—Maruyama for SDE sampling, and connect the ODE coefficients to terminal integration error. Experiments on SiT, FlowDCN and Rectified Flow models, spanning flow-model architectures and formulations, test whether FLOWBUDGET can estimate the NFE budget without generating samples at multiple budgets or performing an FID-versus-NFE sweep.

我们推导了流匹配采样中数值求解器的估计器，即用于 ODE 采样的 Euler 和 Heun 方法，以及用于 SDE 采样的加性噪声 Euler-Maruyama 方法，并将 ODE 系数与最终积分误差联系起来。在涵盖多种流模型架构和公式的 SiT、FlowDCN 和 Rectified Flow 模型上进行的实验测试了 FLOWBUDGET 是否能够在无需在多个预算下生成样本或执行 FID 与 NFE 扫描的情况下估计 NFE 预算。