用快速采样生成高质量预览，提升图像生成交互效率。
Image Diffusion Preview with Consistency Solver
- 提出ConsistencySolver，基于强化学习优化的轻量级高阶求解器。
- 47%更少步数达到与多步DPM-Solver相当的FID分数。
- 用户实验显示交互时间减少近50%，适合实时预览场景。
图像扩散模型的慢速推理严重影响交互体验。为此，我们提出Diffusion Preview新范式：通过快速低步数采样生成初步结果供用户评估，待满意后再进行全步数精炼。现有加速方法（如训练无关求解器、后训练蒸馏）难以兼顾预览质量与最终输出一致性。我们提出的ConsistencySolver源自通用线性多步法，是一个可训练的轻量级高阶求解器，通过强化学习优化，显著提升低步数下的生成质量和一致性。实验表明，ConsistencySolver在低步数场景下大幅改善生成质量与一致性，使预览-精炼流程更高效。特别地，其在仅使用47%步数时，FID得分与多步DPM-Solver相当，优于蒸馏基线。用户研究显示，该方法将整体交互时间减少近50%，同时保持生成质量。代码已开源：https://github.com/G-U-N/consolver。
原文摘要 · Abstract (English)
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.
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。