发表机构
Grand Valley State University; University of North Carolina Wilmington; Oakland University; DePaul University（格兰德维州大学; 北卡罗来纳州温斯洛大学; 奥克兰大学; 德保罗大学）
机构由 AI 辅助整理，请以论文原文为准。AI 中文总结
研究开源机器学习项目中容器化实践，通过对1993个相关Dockerfile进行大规模实证研究，结合定量与定性分析，揭示容器在不同环节作用，指出其构建问题及浪费情况，并识别出7种提高构建效率和减少容器大小的重构模式。
AI 中文摘要
容器化在机器学习领域变得愈发重要，能提供可重复性、可移植性和环境一致性。此前研究未深入考察机器学习工作流的迭代特性对容器大小、构建性能和缓存行为的影响。本文对1993个与机器学习相关的Dockerfile进行了首次大规模实证研究，结合定量与定性分析。结果表明容器在训练、推理和基础设施中作用不同，通常较大且构建时间长，多数提交会触发重建，虽有部分缓存重用，但大量工作浪费于冗余计算，还识别出7种可提高构建效率和减少容器大小的重构模式。
英文摘要
Containerization has become increasingly essential in the machine learning (ML) domain, providing reproducibility, portability, and environment consistency. While prior studies have analyzed Dockerfile structures and best practices, none have examined ML projects in depth to reveal how the iterative nature of ML workflows influences container footprint, build performance, and caching behavior. We present the first large scale empirical study of 1,993 ML related Dockerfiles, combining quantitative analysis of container roles in ML projects and build dynamics with a qualitative investigation of refactoring practices. Results show that containers serve distinct roles across training, inference, and infrastructure. Containers are typically large, averaging 10.27 GB in size, and require long build times of about 8.84 minutes. We find that 44.4% of commits trigger rebuilds, primarily due to context file changes (96.4%), with experimentation being the main motive behind those commits that initiate rebuilds. Despite partial cache reuse, 71% of rebuild work is wasted on redundant computation. From stable projects, we identify 7 recurring ML-specific Dockerfile refactoring patterns that improve build efficiency and reduce container footprint.
Journal refProceedings of MSR '26: 23rd International Conference on Mining Software Repositories, April 13-14, 2026, Rio de Janeiro, Brazil. ACM, 2026
