知识卡片

普通读书笔记卡 · 7003.g · 生成模型:DeepSeek-V4.1-Flash

暗启动与渐进发布

理解等级:中级

正式发布前可先“暗启动(dark launch)”:调用模型并记录结果但不向用户展示,借此验证技术集成而非模型质量。中间路线是按比例(如 1% 用户或 cookie)仅对部分流量使用模型,在控制损害的同时积累对集成与质量的信心。最终发布多采用渐进式(progressive rollout),先小比例用户/服务器再逐步放大。

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

    • Introduction · 查看原书依据

      A dark launch involves consulting the model and logging the result, but not using it actively in the website as users see it. This can give us confidence in the technical integration of the model into our web application but will probably not give us much confidence about the quality of the model.

    参考来源

    - 位置:《Reliable Machine Learning》第01章《Introduction》(源文件:extracted/Reliable Machine Learning/_epub-src/OEBPS/ch01.xhtml) - 结论依据:原文区分 live launch、dark launch 与按比例灰度,说明各自能验证什么。 - 原始内容:A dark launch involves consulting the model and logging the result, but not using it actively in the website as users see it. This can give us confidence in the technical integration of the model into our web application but will probably not give us much confidence about the quality of the model.