Hi, I’m curious if including a dense layer in between two convolutional layers have any sense.
You do put dense layers at the end of CNN structure to get the output but i doubt there is any use of putting those in between the CNN .
Firstly it would be a lot of neurals to match the output shape of the previous CNN or MaxPooling layer
Also I dont see how this is advantageous
You can try it out and see for yourself…just be careful with the shapes
1 Like