通过轨迹几何检测冗余步骤，实现无需训练的高效扩散模型采样优化。
GeoSPRINT: Geometric Redundancy-Aware Step Pruning for Inference in Diffusion Trajectories
- 基于隐空间超平面性测试，动态识别冗余采样步。
- 在相同计算量下，FID性能优于统一步长方法，最高提升1.93。
- 适用于追求效率的扩散模型部署，无需重新训练。
扩散模型虽生成质量高，但推理成本昂贵，因采样需大量顺序神经函数评估（NFE）。现有加速方法或采用固定跳步策略，或基于局部数值误差调整步长，或需额外训练。本文提出GeoSPRINT（几何冗余感知步长剪枝），一种无需训练的框架，从去噪轨迹的几何结构构建非均匀采样调度。GeoSPRINT利用隐空间中的超平面性测试检测几何冗余，通过QR分解高效实现，并将冗余分布转化为采样调度，使更多步骤集中在轨迹高曲率区域。同时引入轨迹投影评分α_traj，一种残差方差指标，量化轨迹直线度，作为无模型的修正流质量诊断工具。在CIFAR-10（32×32）、LSUN Church（256×256）和Stable Diffusion v1.5（512×512隐空间）上，GeoSPRINT在匹配的NFE预算下持续优于统一步长的DDIM。在CIFAR-10上，49–89 NFE下FID提升0.7–1.1，且在NFE≥30时超越DPM-Solver++，即使仅使用一阶DDIM求解器。在LSUN Church上，52步时FID从1.48降至1.26；在Stable Diffusion v1.5上，相比DDIM最多提升1.93 FID。结果表明，轨迹几何可作为分配推理步数的有用全局信号，且无需重训练即可显著提升采样效率。
原文摘要 · Abstract (English)
Diffusion models achieve high sample quality but remain expensive at inference time because sampling requires many sequential neural function evaluations (NFEs). Existing acceleration methods either use fixed step-skipping schedules, adapt step sizes based on local numerical error, or require additional training. We introduce GeoSPRINT (Geometric Step Pruning for Inference in Trajectories), a training-free framework for constructing non-uniform sampling schedules from the geometry of denoising trajectories. GeoSPRINT detects geometrically redundant steps using a hyperplanarity test in latent space, implemented efficiently via QR factorization, and converts the resulting redundancy profile into a sampling schedule that allocates more steps to high-curvature regions of the trajectory. In addition, we introduce the trajectory projection score $\alpha_{\mathrm{traj}}$, a residual-variance metric that quantifies trajectory straightness and serves as a model-free diagnostic for rectified flow quality. Across CIFAR-10 ($32{\times}32$), LSUN Church ($256{\times}256$), and Stable Diffusion v1.5 ($512{\times}512$ latent), GeoSPRINT consistently improves over uniform DDIM (Denoising Diffusion Implicit Models) schedules at matched NFE budgets. On CIFAR-10, GeoSPRINT improves FID (Fr\'echet Inception Distance) by 0.7-1.1 over DDIM across 49-89 NFEs and surpasses DPM-Solver++ at NFE${\geq}30$ despite using a first-order DDIM solver. On LSUN Church, it reduces FID from 1.48 to 1.26 at 52 steps, and on Stable Diffusion v1.5 it achieves up to 1.93 FID improvement over DDIM. These results show that trajectory geometry provides a useful global signal for allocating inference steps and that schedule quality can substantially improve diffusion sampling efficiency without retraining.
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。