通过共享相似提示的早期生成计算，大幅降低图像集生成成本
Reusing Computation in Text-to-Image Diffusion for Efficient Generation of Image Sets
- 基于语义相似性聚类提示，共享扩散模型早期去噪步骤
- 在相同质量下，计算量减少30%以上，且图像质量反而提升
- 无需训练，可直接接入现有流程，适合大规模图像生成场景
文本到图像扩散模型虽能生成高质量图像，但计算开销巨大。现有工作多聚焦单次推理优化，本文提出一种新思路：减少相关提示间的冗余。利用扩散模型从粗到精的特性，早期去噪步骤可捕捉相似提示间的共性结构。我们提出一种无需训练的方法，通过语义相似性聚类提示，并在早期扩散步骤中共享计算。实验表明，对基于图像嵌入条件训练的模型，该方法显著降低计算成本，同时提升图像质量。结合UnClip的文本到图像先验，进一步优化扩散步数分配，提升效率。方法可无缝集成至现有流程，随提示集规模扩展，有效减轻大规模文本到图像生成的环境与经济负担。
原文摘要 · Abstract (English)
Text-to-image diffusion models enable high-quality image generation but are computationally expensive. While prior work optimizes per-inference efficiency, we explore an orthogonal approach: reducing redundancy across correlated prompts. Our method leverages the coarse-to-fine nature of diffusion models, where early denoising steps capture shared structures among similar prompts. We propose a training-free approach that clusters prompts based on semantic similarity and shares computation in early diffusion steps. Experiments show that for models trained conditioned on image embeddings, our approach significantly reduces compute cost while improving image quality. By leveraging UnClip's text-to-image prior, we enhance diffusion step allocation for greater efficiency. Our method seamlessly integrates with existing pipelines, scales with prompt sets, and reduces the environmental and financial burden of large-scale text-to-image generation. Project page: https://ddecatur.github.io/hierarchical-diffusion/
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。