知识卡片

普通读书笔记卡 · 9259.a · 生成模型:GPT-6

池化示例均值需保留小数

理解等级:中级

按图8.9右侧2×2不重叠窗口,左上(31+15+0+100)/4=36.5,右下(7+2+45+6)/4=15.5。图印36与15未说明取整,不可当精确平均;左右输入首项还分别为29和31。

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 illustrations compare max pooling and average pooling using a 4 by 4 matrix as input, with a 2 by 2 pool size applied in both cases. Left: Max pooling is performed, where the maximum value from each 2 by 2 region is selected, resulting in a reduced 2 by 2 output matrix containing the highest values from each subregion. Right: Average pooling is demonstrated,

    参考来源

    - 位置:《Machine Learning Fundamentals: Concepts, Models, and Applications》第08章《深度学习模型》(来源项目:a19c7933-4eda-4176-aebd-30fe7e1ad292;源文件:extracted/Machine Learning Fundamentals: Concepts, Models, and Applications/_epub-src/OEBPS/xhtml/21_chapter08.xhtml) - 结论依据:摘录中关于池化示例均值需保留小数的定义、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 illustrations compare max pooling and average pooling using a 4 by 4 matrix as input, with a 2 by 2 pool size applied in both cases. Left: Max pooling is performed, where the maximum value from each 2 by 2 region is selected, resulting in a reduced 2 by 2 output matrix containing the highest values from each subregion. Right: Average pooling is demonstrated,