知识卡片

普通读书笔记卡 · 6698.d · 生成模型:GPT-6

同一输入输出同一模型分布

理解等级:中级

固定参数的skip-gram对同一Monet输入输出同一分布,训练对分别将Claude和painted作目标来共同调整该分布。书中两幅目标各近1的输出可看作不同理想化教学情形,不能理解为同一步模型已同时给两词各约98%。

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

      Figure 6.7 shows the numerical network input and output for the first two surrounding words. In this case, the input word is Monet, and the expected output of the network is either Claude or painted, depending on the training pair. In this figure, you can see some idealized example values for a Word2Vec neural network being trained using the skip-gram approach on the pair of

    参考来源

    - 位置:《Natural Language Processing in Action, Second Edition》第06章《用词嵌入推理》(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-6.html) - 结论依据:摘录中关于同一输入输出同一模型分布的定义、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:Figure 6.7 shows the numerical network input and output for the first two surrounding words. In this case, the input word is Monet, and the expected output of the network is either Claude or painted, depending on the training pair. In this figure, you can see some idealized example values for a Word2Vec neural network being trained using the skip-gram approach on the pair of