知识卡片
编程框架分离流程与可调变量
DSPy把模块连接和控制流与提示、示例及模型参数的优化分开,使迭代更可重复。抽象不能省去任务定义、数据和指标选择,自动编译也不证明得到业务最优或无幻觉系统。
这张卡片看不懂?
这张卡片已通过人工审查,不再接受失败反馈。
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
To this end, in this chapter we will discuss various techniques to compose a production-level LLM system that can power useful applications. We will explore how to leverage multi-LLM architectures to balance cost and performance. Finally, we will look into software frameworks like DSPy that integrate LLM application development into the conventional software programming paradigm.
参考来源
- 位置:《Designing Large Language Model Applications》第13章《设计模式与系统架构》(来源项目:a19c7933-4eda-4176-aebd-30fe7e1ad292;源文件:extracted/Designing Large Language Model Applications/_epub-src/OEBPS/ch13.html)
- 结论依据:摘录中关于编程框架分离流程与可调变量的定义、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:To this end, in this chapter we will discuss various techniques to compose a production-level LLM system that can power useful applications. We will explore how to leverage multi-LLM architectures to balance cost and performance. Finally, we will look into software frameworks like DSPy that integrate LLM application development into the conventional software programming paradigm.