本文已参与「新人创作礼」活动,一起开启掘金创作之路。
8. 卷积神经网络的一个完整例子
一个完整的卷积神经网络模型:
9. Pooling layer(池化层)
Max pooling 没有参数,因此不需要通过反向传播来学习参数
多通道图片时,对每个通道分别进行max pooling
pooling层的作用:平移不变性,减少图片尺寸pooling层的作用-知乎
The pooling (POOL) layer reduces the height and width of the input. It helps reduce computation, as well as helps make feature detectors more invariant to its position in the input. max pooling比较常用,average pooling不常用
hyperparameters:
- f: filter size (f=2,s=2,used quite often)
- s: stride
- Max or average pooling
10. 卷积神经网络经典模型
10.1 Outline
经典的网络结构:
- LeNet-5
- AlexNet
- VGG
- ResNet
- Inception
10.1 LeNet-5
在计算层数时通常把有参数的算作一层,因为pooling层没有参数,因此conv层和pooling层放在一起当作一层. LetNet-5结构图:
![](https://github.com/lankuohsing/Markdown-Images/raw/master/%E6%B7%B1%E5%BA%A6%E5%AD%A6%E4%B9%A0%E5%85%A5%E9%97%A8%E6%95%99%E7%A8%8B/%E5%8D%B7%E7%A7%AF%E7%A5%9E%E7%BB%8F%E7%BD%91%E7%BB%9C/LeNet-5.png)
![](https://github.com/lankuohsing/Markdown-Images/raw/master/%E6%B7%B1%E5%BA%A6%E5%AD%A6%E4%B9%A0%E5%85%A5%E9%97%A8%E6%95%99%E7%A8%8B/%E5%8D%B7%E7%A7%AF%E7%A5%9E%E7%BB%8F%E7%BD%91%E7%BB%9C/AlexNet.png)
![](https://github.com/lankuohsing/Markdown-Images/raw/master/%E6%B7%B1%E5%BA%A6%E5%AD%A6%E4%B9%A0%E5%85%A5%E9%97%A8%E6%95%99%E7%A8%8B/%E5%8D%B7%E7%A7%AF%E7%A5%9E%E7%BB%8F%E7%BD%91%E7%BB%9C/VGG-16.png)