Home >
other > Convolution of the neural network connection layer all 4096 what is the purpose of the hidden units
Convolution of the neural network connection layer all 4096 what is the purpose of the hidden units
After the biggest pooling layer is drawn into a one dimensional vector data input into the full connection layer, there are three full connection layers respectively 4096102, 4128 hidden units, what's the function of the hidden units?
