Practical convolution correlation trick

One .  Convolution kernel channel number and input layer 、 Relationship between output layers

The number of input channels of convolution kernel = The number of channels when a picture is used as the input layer

The number of convolution kernels = When a picture becomes an output layer after convolution , The number of channels in this output layer

For the above, please refer to the blog ：

[1.] How to determine the CNN The number of convolution kernel channels and the number of convolution output channels ( depth )_ Cheap little Du's blog -CSDN Blog _cnn How to determine the number of convolution kernels

Two . Calculation formula of convolution output layer size