知识卡片

普通读书笔记卡 · 8789.c · 生成模型:GPT-6

特性分支需要承认集成代价

理解等级:中级

长训练可能使团队选择特性分支,但这仍是对延迟集成的取舍,不应因有自动流水线就称已做到持续集成。保留结对、快速测试、及时审阅和相同发布门禁,尽可能缩短分支寿命。书中两三周是案例上限,不是短分支的通行标准。

这张卡片看不懂?

已有反馈 · 0 条

    还没有人提交反馈。

    在整书笔记中的位置

    原书出处 · 1 处

    • · 查看原书依据

      feature branching and pull requests (i.e., not doing trunk-based development and CI) may be an acceptable trade-off. But ML practitioners must still execute fast-running tests (e.g., training smoke tests as described in Chapter 6) locally and on the CI pipeline, before long and costly training runs. This affords ML practitioners fast feedback on the quality of their changes as they make small and frequent commits

    参考来源

    - 位置:《Effective Machine Learning Teams (David Tan, Ada Leung, David Colls)》第09章《MLOps与持续交付》(来源项目:a19c7933-4eda-4176-aebd-30fe7e1ad292;源文件:extracted/Effective Machine Learning Teams (David Tan, Ada Leung, David Colls)/_epub-src/OEBPS/ch09.html) - 结论依据:摘录中关于特性分支需要承认集成代价的定义、比较或条件支撑卡片;原书表述不严谨的范围已在覆盖记录逐项限定,不把案例结果外推为普遍结论。 - 原始内容:feature branching and pull requests (i.e., not doing trunk-based development and CI) may be an acceptable trade-off. But ML practitioners must still execute fast-running tests (e.g., training smoke tests as described in Chapter 6) locally and on the CI pipeline, before long and costly training runs. This affords ML practitioners fast feedback on the quality of their changes as they make small and frequent commits