番外 CNN

本次学习参考 stanford的cs231n 的letcture5
Pasted image 20240909155831.png
这个图很好
Pasted image 20240909161034.png

Pasted image 20240909161218.png

Pasted image 20240909161320.png

Pasted image 20240909161615.png
下面的stride3对应上面的红字
Pasted image 20240909161540.png
如果有padding
Pasted image 20240909161735.png

注意下面这个例子，bias是根据卷积核确定的！！！
Pasted image 20240909162250.png
这个概念看一下，感受视野
Pasted image 20240909162430.png
红色方框和绿色方框给出了两个知识点。红色计算感受视野，绿色给出下采样的原因
Pasted image 20240909162747.png

Pasted image 20240909163034.png

在udlbook的CNN章节对应的练习# 10.1: 1D Convolution中提到了下面这段话

Note

If you have the first edition of the printed book, it mistakenly refers to a convolutional filter with no spaces between the elements (i.e. a normal filter without dilation) as having dilation zero. Actually, the convention is (weirdly) that this has dilation one. And when there is one space between the elements, this is dilation two.

下面是一段解释。
Pasted image 20240911093042.png