不训练即可加速扩散模型，保持图像质量不变
Training-Free Adaptive Diffusion with Bounded Difference Approximation Strategy
- 通过三阶潜在差分判断步骤稳定性，决定跳过哪些噪声预测
- 实测图像与视频生成平均提速2~5倍，结果与原过程完全一致
- 无需重新训练，适合对响应速度敏感的交互式应用
扩散模型在高质量图像和视频生成方面取得显著进展。然而，现有去噪技术依赖逐步噪声预测，计算开销大，导致交互应用延迟过高。本文提出AdaptiveDiffusion，通过在去噪过程中自适应减少噪声预测步骤来缓解该瓶颈。方法基于三阶潜在差分判断时间步间的稳定性，从而决定可跳过的步骤数量，同时保留先前预测结果以提升效率。在图像与视频扩散模型上的大量实验表明，该方法可在保持原始结果完全一致的前提下显著加速去噪过程，平均实现2~5倍提速，且无质量损失。
原文摘要 · Abstract (English)
Diffusion models have recently achieved great success in the synthesis of high-quality images and videos. However, the existing denoising techniques in diffusion models are commonly based on step-by-step noise predictions, which suffers from high computation cost, resulting in a prohibitive latency for interactive applications. In this paper, we propose AdaptiveDiffusion to relieve this bottleneck by adaptively reducing the noise prediction steps during the denoising process. Our method considers the potential of skipping as many noise prediction steps as possible while keeping the final denoised results identical to the original full-step ones. Specifically, the skipping strategy is guided by the third-order latent difference that indicates the stability between timesteps during the denoising process, which benefits the reusing of previous noise prediction results. Extensive experiments on image and video diffusion models demonstrate that our method can significantly speed up the denoising process while generating identical results to the original process, achieving up to an average 2~5x speedup without quality degradation.
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。