Creating a Neural Network

Hula has a multitude of ways to initialize a Neural Network.

The two modules you will need to know are the DeepR module and the rlutils module

the DeepR module has the classes required to initialize a Neural Network, while the rlutils module has activation functions to assign to each layer.

from hula.DeepR import Net, FeedForwardLayer, RecurrentLayer
from hula.rlutils import sigmoid, softplus, tanh, softmax

The FeedForwardLayer class is used to create a single layer in your Neural Network

ExampleLayer = FeedForwardLayer(2, 1, sigmoid)
ExampleLayer.activate([0, 1])

array([0.99988349])

All outputs are numpy array objects.