← 返回刚才阅读的位置

知识卡片

普通读书笔记卡 · 6790.c · 生成模型:GPT-6

训练样例要覆盖表达差异与反例

理解等级:中级

每个意图需有代表性措辞,并与容易混淆的类别比较。七到十条例子只是起步建议,不保证泛化;增加近重复句不等于增加覆盖,评估应使用独立用户或表达变式。

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

      Pretty straightforward, right? Rasa will warn if you have too few examples for a particular intent and recommends at least 7–10 utterance examples per intent. The next file you should look at is domain.yml, in the main directory. Its first section is quite straightforward: it defines the intents from the nlu.yml file that the chatbot should be able to understand. Let’s add the intents we

    参考来源

    - 位置:《Natural Language Processing in Action, Second Edition》第12章《对话引擎与会话设计》(来源项目:a19c7933-4eda-4176-aebd-30fe7e1ad292;源文件:extracted/Natural Language Processing in Action, Second Edition/_epub-src/OEBPS/Text/chapter-12.html) - 结论依据:摘录中关于训练样例要覆盖表达差异与反例的定义、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:Pretty straightforward, right? Rasa will warn if you have too few examples for a particular intent and recommends at least 7–10 utterance examples per intent. The next file you should look at is domain.yml, in the main directory. Its first section is quite straightforward: it defines the intents from the nlu.yml file that the chatbot should be able to understand. Let’s add the intents we