Table of Contents  ▸  List of Commands  ▸  Neural Networks  ▸  nn_layer_conv3dnnl

◀  nn_layer_conv3dnl    |    nn_layer_crop  ▶

nn_layer_conv3dnnl

Arguments:

• out,in,nb_channels>0,_kernel_size>0,_stride>0,_dilation>0,_activation,0<=_learning_mode<=3

Description:

Add a conv3dnnl (3D convolutional layer followed by a normalization layer, then a non-linearity) to the network.

learning_mode can be { 0=no learning | 1=weights only | 2=biases only | 3=weights+biases }.

Default values:

kernel_size=3, stride=1, dilation=1, activation=leakyrelu and learning_mode=3.