ETC: training-free diffusion models acceleration with Error-aware Trend Consistency

TL;DR

ETC is a training-free framework that accelerates diffusion models by ensuring denoising trend consistency and error control, significantly reducing sampling time with minimal quality loss.

Contribution

Introduces Error-aware Trend Consistency (ETC), a novel method that maintains trajectory stability and error tolerance in diffusion model acceleration without retraining.

Findings

Achieves 2.65x faster sampling compared to FLUX.

Maintains high consistency with only slight SSIM score degradation.

Demonstrates effective error control and trend prediction in diffusion trajectories.

Abstract

Diffusion models have achieved remarkable generative quality but remain bottlenecked by costly iterative sampling. Recent training-free methods accelerate diffusion process by reusing model outputs. However, these methods ignore denoising trends and lack error control for model-specific tolerance, leading to trajectory deviations under multi-step reuse and exacerbating inconsistencies in the generated results. To address these issues, we introduce Error-aware Trend Consistency (ETC), a framework that (1) introduces a consistent trend predictor that leverages the smooth continuity of diffusion trajectories, projecting historical denoising patterns into stable future directions and progressively distributing them across multiple approximation steps to achieve acceleration without deviating; (2) proposes a model-specific error tolerance search mechanism that derives corrective thresholds by identifying transition points from volatile semantic planning to stable quality refinement. Experiments show that ETC achieves a 2.65x acceleration over FLUX with negligible (-0.074 SSIM score) degradation of consistency.