知识卡片

普通读书笔记卡 · 6604.e · 生成模型:GPT-6

应用方案以质量和资源共同验收

理解等级:中级

生产方案除了回答是否正确,还需满足延迟、成本、可控输出和可靠运行等约束。选择模型、微调、缓存或量化时应明确改变哪项指标以及代价;单次演示成功不能替代目标负载和失败情形下的验收。

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

      nonfunctional requirements, including minimizing latency (the time the user needs to wait for the model response) and cost. In Chapter 9, we will discuss techniques for inference optimization, including caching, distillation, and quantization.

    参考来源

    - 位置:《Designing Large Language Model Applications》第01章《大语言模型入门》(来源项目:a19c7933-4eda-4176-aebd-30fe7e1ad292;源文件:extracted/Designing Large Language Model Applications/_epub-src/OEBPS/ch01.html) - 结论依据:摘录中关于应用方案以质量和资源共同验收的定义、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:nonfunctional requirements, including minimizing latency (the time the user needs to wait for the model response) and cost. In Chapter 9, we will discuss techniques for inference optimization, including caching, distillation, and quantization.