知识卡片

普通读书笔记卡 · 8768.c · 生成模型:GPT-6

生产监控区分三类反馈

理解等级:中级

应用监控观察延迟、吞吐和错误,数据监控看输入变化,模型监控用新标签衡量预测质量。前两类可较实时,第三类受标签延迟限制;输入漂移不必然说明质量下降,服务无错也不说明模型正确。

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

      There are three aspects to monitoring ML models: Application monitoring Metrics such as throughput, latency, and error rates give you feedback on how your application is behaving in production. Is everything going well? Is it crawling to a near-stop? Is everything in flames? Any errors you observe, complemented with useful application logs, provide you with the prompt and the information needed to reproduce the error

    参考来源

    - 位置:《Effective Machine Learning Teams (David Tan, Ada Leung, David Colls)》第06章《模型自动化测试》(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/ch06.html) - 结论依据:摘录中关于生产监控区分三类反馈的定义、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:There are three aspects to monitoring ML models: Application monitoring Metrics such as throughput, latency, and error rates give you feedback on how your application is behaving in production. Is everything going well? Is it crawling to a near-stop? Is everything in flames? Any errors you observe, complemented with useful application logs, provide you with the prompt and the information needed to reproduce the error