arXiv：cs.LG（机器学习，全量分类）· Tianshuo Qiao, Naiqian Zheng, Xiaopeng Liu, Shuguang Wang, Diandian Gu, Xuanzhe Liu, Xin Jin·· 14 小时前AI 评分40
FastCI：面向 LLM 训练框架的高效 GPU 密集型 CI
FastCI: Efficient GPU-Intensive CI for LLM Training Frameworks
AI 导读
字节跳动提出 FastCI，通过运行时证据筛选受影响测试、剪枝等价上下文中的测试并优先执行高风险测试，将 LLM 训练框架的 CI 延迟降低 77.5%、GPU 资源占用减少 63.9%，同时修改代码覆盖率保留率提升 3.2%。该框架已集成到字节跳动 LLM 训练框架的 CI 流水线中。
正文
Abstract:As large language models (LLMs) keep growing in size and complexity, their training frameworks evolve at a rapid pace as well. Therefore, continuous integration (CI) is critical for maintaining the quality and stability of these frameworks. However, unlike traditional software, CI for LLM training frameworks relies on GPU-intensive tests, which usually involve complete model training or evaluation. This leads CI itself to become a new bottleneck for fast-paced development. In this paper, we introduce FastCI, a framework that improves the efficiency of CI for LLM training frameworks. FastCI leverages runtime evidence to select affected tests and prune tests that execute changed code in equivalent contexts. Then FastCI prioritizes high-risk tests to expose potential failures earlier, and optimizes test workloads along dimensions outside the intended validation scope of each test. Evaluated on the CI workload of our LLM training framework, FastCI reduces the CI latency by 77.5% and the GPU resource usage by 63.9%, while improving the modified code coverage retention by 3.2%, compared with the currently deployed CI pipelines. FastCI has now been integrated into the CI pipelines of our LLM training framework at ByteDance.
| Subjects: | Machine Learning (cs.LG); Software Engineering (cs.SE) |
| Cite as: | arXiv:2610.01967 [cs.LG] |
| (or arXiv:2610.01967v1 [cs.LG] for this version) | |
| https://doi.org/10.48550/arXiv.2610.01967 arXiv-issued DOI via DataCite (pending registration) |
Submission history
From: Tianshuo Qiao [view email]
[v1]
Thu, 1 Oct 2026 16:18:14 UTC (1,043 KB)
来源：arXiv：cs.LG（机器学习，全量分类） · arxiv.org