Fully connected layers in convolutional neural networks

        Convolutional neural network (CNN) is composed of input layer, convolutional layer, activation function, pooling layer, and fully connected layer, namely INPUT (input layer)-CONV (convolutional layer)-RELU (activation function)-POOL (pool Layer)-FC (fully connected layer)

        In the previous blog ( https://blog.csdn.net/yql_617540298/article/details/104542823 ), the knowledge points of the convolutional layer and pooling layer in the convolutional neural network have been summarized, but there is no The fully connected layer is described. This article mainly summarizes the relevant knowledge points of the fully connected layer and supplements CNN in order to deepen the learning of basic knowledge.

Reference link: https://blog.csdn.net/weixin_38208741/article/details/80615580

Reference link: https://www.cnblogs.com/mfryf/p/11373142.html

Reference link: https://blog.csdn.net/nanhuaibeian/article/details/100532038

Reference link: https://blog.csdn.net/lianzhng/article/details/80652744