← 返回起始笔记位置

知识卡片

普通读书笔记卡 · 8786.a · 生成模型:GPT-6

无测试让人工验证成为瓶颈

理解等级:中级

没有自动检查时,每次提交都依赖人工回归,或者冒险跳过验证。即使部署已很快,等待测试仍拖长整个周期;团队于是延迟合并,积累冲突和返工。优化局部部署速度之前,应先找端到端最慢的反馈环节。

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

      we end up wasting significant amounts of time testing for or fixing errors. Third, even if we’ve optimized other parts of our CI/CD pipeline (e.g., model deployment in 30 seconds), the testing and quality assurance step will remain a bottleneck (e.g., manual testing takes a few hours, or whenever we have time to do it, which then stretches the bottleneck out to days or weeks).

    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:摘录中关于无测试让人工验证成为瓶颈的定义、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:we end up wasting significant amounts of time testing for or fixing errors. Third, even if we’ve optimized other parts of our CI/CD pipeline (e.g., model deployment in 30 seconds), the testing and quality assurance step will remain a bottleneck (e.g., manual testing takes a few hours, or whenever we have time to do it, which then stretches the bottleneck out to days or weeks).