通过共享子空间加速扩散模型得分计算，显著提升生成效率。
Efficient Score Pre-computation for Diffusion Models via Cross-Matrix Krylov Projection
- 用跨矩阵克雷洛夫投影共享子空间，复用计算资源
- 相比标准稀疏求解器提速15.8%至43.7%，最高达115倍
- 适合算力受限下高质量图像生成，实用性强
本文提出一种加速基于得分的扩散模型的新框架。将标准稳定扩散模型转换为福克-普朗克形式后，需对每张图像求解大型线性系统，训练多图时计算成本高。核心创新是跨矩阵克雷洛夫投影方法，利用矩阵间的数学相似性，通过一组‘种子’矩阵构建共享子空间，快速求解后续的‘目标’矩阵。实验表明，该技术相比标准稀疏求解器可实现15.8%至43.7%的耗时降低；在去噪任务中，相较DDPM基线最高提速115倍。在固定计算预算下，本方法仍能生成高质量图像，而DDPM无法产出可识别内容，证明其在资源受限场景下的实用性。
原文摘要 · Abstract (English)
This paper presents a novel framework to accelerate score-based diffusion models. It first converts the standard stable diffusion model into the Fokker-Planck formulation which results in solving large linear systems for each image. For training involving many images, it can lead to a high computational cost. The core innovation is a cross-matrix Krylov projection method that exploits mathematical similarities between matrices, using a shared subspace built from ``seed" matrices to rapidly solve for subsequent ``target" matrices. Our experiments show that this technique achieves a 15.8\% to 43.7\% time reduction over standard sparse solvers. Additionally, we compare our method against DDPM baselines in denoising tasks, showing a speedup of up to 115$\times$. Furthermore, under a fixed computational budget, our model is able to produce high-quality images while DDPM fails to generate recognizable content, illustrating our approach is a practical method for efficient generation in resource-limited settings.
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。