无需训练即可加速扩散模型，保持生成一致性。
ETC: training-free diffusion models acceleration with Error-aware Trend Consistency
- 利用去噪趋势预测实现多步重用，避免轨迹偏移。
- 在FLUX上实现2.65倍加速，保真度仅下降0.074 SSIM。
- 适合追求高效生成且需稳定输出的场景。
扩散模型虽生成质量优异，但受限于昂贵的迭代采样过程。近期无训练方法通过复用模型输出加速生成，但忽略去噪趋势且缺乏针对模型特性的误差控制，导致多步复用时轨迹偏移，生成结果不一致。为此，我们提出误差感知趋势一致性（ETC）框架：（1）引入趋势预测器，利用扩散轨迹的平滑连续性，将历史去噪模式投影为稳定未来方向，并逐步分配至多个近似步骤，实现加速且不偏离；（2）提出模型特异性误差容忍搜索机制，通过识别从波动语义规划到稳定质量优化的转折点，确定校正阈值。实验表明，ETC在FLUX上实现2.65倍加速，一致性仅下降0.074 SSIM。
原文摘要 · Abstract (English)
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.
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。