Image Diffusion Preview with Consistency Solver

TL;DR

This paper introduces Diffusion Preview with a ConsistencySolver to enable fast, high-quality image diffusion previews, reducing user interaction time by nearly 50% while maintaining output quality.

Contribution

It proposes a novel high-order, reinforcement learning-optimized solver that improves preview quality and consistency in low-step diffusion sampling.

Findings

Achieves FID scores comparable to multistep methods with 47% fewer steps.

Significantly improves preview quality and consistency.

Reduces user interaction time by nearly 50%.

Abstract

The slow inference process of image diffusion models significantly degrades interactive user experiences. To address this, we introduce Diffusion Preview, a novel paradigm employing rapid, low-step sampling to generate preliminary outputs for user evaluation, deferring full-step refinement until the preview is deemed satisfactory. Existing acceleration methods, including training-free solvers and post-training distillation, struggle to deliver high-quality previews or ensure consistency between previews and final outputs. We propose ConsistencySolver derived from general linear multistep methods, a lightweight, trainable high-order solver optimized via Reinforcement Learning, that enhances preview quality and consistency. Experimental results demonstrate that ConsistencySolver significantly improves generation quality and consistency in low-step scenarios, making it ideal for efficient preview-and-refine workflows. Notably, it achieves FID scores on-par with Multistep DPM-Solver using 47% fewer steps, while outperforming distillation baselines. Furthermore, user studies indicate our approach reduces overall user interaction time by nearly 50% while maintaining generation quality. Code is available at https://github.com/G-U-N/consolver.