GWU_Network

class gwu_nn.gwu_network.GWUNetwork

The GWUNetwork class is the core class of the library that provies a foundation to build a network by iteratively adding layers

add(layer)

A network is comprised of a series of layers connected together. The add method provides a means to add a layer to a network

Args:

Layer (Layer): A Layer object to add to the network

compile(loss, lr)

Compile sets a model’s loss function and learning rate, preparing the model for training

Args:

loss (LossFunction): The loss function used for the network lr (float): The learning rate for the network

fit(x_train, y_train, epochs, batch_size=None)

Fit is the trianing loop for the model/network

Args:

x_train (np.array): Inputs for the network to train on y_train (np.array): Expected outputs for the network epochs (int): Number of training cycles to run through batch_size (int): Number of records to train on at a time

get_weights()

Get the weights for the model

Returns:

np.array: weights of the model

predict(input_data)

Predict produces predictions for the provided input data

Args:

input_data (np.array): Input data to inference

Returns:

np.array: the predictions for the given model