← 返回刚才阅读的位置

知识卡片

普通读书笔记卡 · 8803.a · 生成模型:GPT-6

长任务可降低人工占用比例

理解等级:高级

无法明显缩短长训练时,可以自动执行并记录结果,降低需要人工持续照看的时间,使人更容易离开并恢复上下文。此时周期时长仍长,不能宣称反馈已加快。并行填满所有空档又可能增加在制品和等待,应定期检查实际流动效率。

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

      reduce your duty cycle so that long-running jobs are mainly hands-off, and easy to context-switch in and out of. This allows you to run multiple jobs—from one product or multiple products—in parallel. The same techniques that rapidly brought you to this stage of product development—such as clear definitions of done and heavy automation of processes—will help you shift smoothly from minimizing feedback cycle time to

    参考来源

    - 位置:《Effective Machine Learning Teams (David Tan, Ada Leung, David Colls)》第10章《高效团队的基础》(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/ch10.html) - 结论依据:摘录中关于长任务可降低人工占用比例的定义、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:reduce your duty cycle so that long-running jobs are mainly hands-off, and easy to context-switch in and out of. This allows you to run multiple jobs—from one product or multiple products—in parallel. The same techniques that rapidly brought you to this stage of product development—such as clear definitions of done and heavy automation of processes—will help you shift smoothly from minimizing feedback cycle time to