breidablik.interpolate.load

Classes

FFNN	Feed Forward Neural Network

Functions

ReLU(x)
linear(x, weights, bias)

Module Contents

breidablik.interpolate.load.ReLU(x)

breidablik.interpolate.load.linear(x, weights, bias)

class breidablik.interpolate.load.FFNN(model)

Feed Forward Neural Network

forward(x)

Forward pass

Parameters:

x (ndarray) – The features matrix to predict results on.

Returns:

y – THe predicted results.

Return type:

ndarray

__call__(x)

load(path, model='ffnn.json')

Load model from json file. Assumes that the model is saved with keys weight_x, bias_x, where x is the layer number.

Parameters:

path (str) – Path to load the model from. Needs to be a folder.