一句话：提出缓存式服务系统 MoDM，通过混合大小扩散模型在延迟与图像质量间动态权衡，服务时间降低 2.5 倍。
扩散式文本到图像生成模型面临延迟与质量的权衡：小模型推理快但图像质量低，大模型质量高但速度慢。本文提出 MoDM，一个基于缓存的扩散模型服务系统，通过混合多个扩散模型动态平衡延迟与质量。与依赖模型内部特征的已有工作不同，MoDM 直接缓存最终生成图像，因而可在不同扩散模型族之间无缝检索与复用。系统对小模型处理的缓存命中请求复用缓存图像以保持质量，仅将大模型留给缓存未命中请求以降低延迟；同时设计全局监控器，最优分配 GPU 资源并平衡推理负载，在满足服务级目标（SLO）的同时保证高吞吐。实验表明，MoDM 在保持图像质量的前提下将平均服务时间降低 2.5 倍，适用于可扩展且资源高效的模型部署。
Diffusion-based text-to-image generation models trade latency for quality: small models are fast but generate lower quality images, while large models produce better images but are slow. We present MoDM, a novel caching-based serving system for diffusion models that dynamically balances latency and quality through a mixture of diffusion models. Unlike prior approaches that rely on model-specific internal features, MoDM caches final images, allowing seamless retrieval and reuse across multiple diffusion model families. This design enables adaptive serving by dynamically balancing latency and image quality: using smaller models for cache-hit requests to reduce latency while reserving larger models for cache-miss requests to maintain quality. Small model image quality is preserved using retrieved cached images. We design a global monitor that optimally allocates GPU resources and balances inference workload, ensuring high throughput while meeting Service-Level Objectives (SLOs) under varying request rates. Our evaluations show that MoDM significantly reduces average serving time by 2.5× while retaining image quality, making it a practical solution for scalable and resource-efficient model deployment.
DOI 原文 · arXiv · PDF（开放获取） · DBLP
@inproceedings{abs-2503-11972,
title = {MoDM: Efficient Serving for Image Generation via
<u>M</u>
ixture-
<u>o</u>
f-
<u>D</u>
iffusion
<u>M</u>
odels},
author = {Yuchen Xia and Divyam Sharma and Yichao Yuan and Souvik Kundu and Nishil Talati},
booktitle = {ASPLOS},
year = {2025},
doi = {10.1145/3760250.3762220}
}