知识卡片

普通读书笔记卡 · 9378.c · 生成模型:GPT-6

一维转置不会自动增加行列轴

理解等级:中级

二维数组转置交换行列,一维shape为(n,)的数组转置仍是(n,)。需要列向量时应显式增加轴或reshape;高维.T还会改变多轴次序,不宜把二维直觉套到所有张量。

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

      The transpose of a matrix is a corresponding matrix with the rows and columns interchanged. If you imagine writing down the matrix of numbers on a piece of graph paper, you could flip the piece of paper along its diagonal, so it is face down on the table. If you could see through the piece of paper, you would see the numbers in the correct

    参考来源

    - 位置:《Natural Language Processing in Action, Second Edition》附录C《附录C:向量与线性代数》(来源项目:a19c7933-4eda-4176-aebd-30fe7e1ad292;源文件:extracted/Natural Language Processing in Action, Second Edition/_epub-src/OEBPS/Text/appendix-c.html) - 结论依据:摘录中关于一维转置不会自动增加行列轴的定义、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:The transpose of a matrix is a corresponding matrix with the rows and columns interchanged. If you imagine writing down the matrix of numbers on a piece of graph paper, you could flip the piece of paper along its diagonal, so it is face down on the table. If you could see through the piece of paper, you would see the numbers in the correct