To create the neural network with the Shogun library, we have to start by defining the architecture of the network. We use the CNeuralLayers class in the Shogun library to do so, which is used for aggregating the network layers. It has different methods for creating layers:
- input: Creates the input layer with a specified number of dimensions
- logistic: Creates a fully connected hidden layer with the logistic (sigmoid) activation function
- linear: Creates a fully connected hidden layer with the linear activation function
- rectified_linear: Creates a fully connected hidden layer with the ReLU activation function
- leaky_rectified_linear: Creates a fully connected hidden layer with the Leaky ReLU activation function
- softmax: Creates a fully ...
