SeaCache: Spectral-Evolution-Aware Cache for Accelerating Diffusion Models

TL;DR

SeaCache introduces a spectral-evolution-aware caching strategy for diffusion models, significantly reducing inference latency while maintaining high output quality by leveraging spectral priors and dynamic reuse schedules.

Contribution

It proposes a novel spectral-evolution-aware cache that improves sampling speed in diffusion models without retraining, based on spectral analysis and dynamic scheduling.

Findings

Achieves state-of-the-art latency-quality trade-offs.

Effective across diverse visual generative models.

Spectral filtering preserves content while reducing noise.

Abstract

Diffusion models are a strong backbone for visual generation, but their inherently sequential denoising process leads to slow inference. Previous methods accelerate sampling by caching and reusing intermediate outputs based on feature distances between adjacent timesteps. However, existing caching strategies typically rely on raw feature differences that entangle content and noise. This design overlooks spectral evolution, where low-frequency structure appears early and high-frequency detail is refined later. We introduce Spectral-Evolution-Aware Cache (SeaCache), a training-free cache schedule that bases reuse decisions on a spectrally aligned representation. Through theoretical and empirical analysis, we derive a Spectral-Evolution-Aware (SEA) filter that preserves content-relevant components while suppressing noise. Employing SEA-filtered input features to estimate redundancy leads to dynamic schedules that adapt to content while respecting the spectral priors underlying the diffusion model. Extensive experiments on diverse visual generative models and the baselines show that SeaCache achieves state-of-the-art latency-quality trade-offs.