A from-scratch implementation of a feedforward neural network. Written in Python with NumPy.
Run this test to train a neural net on the MNIST database.
Create and train a network using:
net = neural_net.Network([input_layer_size, hidden_layer_one_size, ..., output_layer_size])
net.train(training_inputs, training_outputs, epochs, batch_size, learning_rate, test_inputs, test_outputs)Where training_inputs is a numpy ndarray of dimensions [training set size, input layer size, 1], and
training_outputs is a list of numpy ndarrays of dimensions [output layer size, 1].
test_inputs and test_outputs are optional. If specified, they are used to evaluate and print the performance of the
network after each epoch.
Evaluate the network against a test set:
net.percentage_correct(test_inputs, test_outputs)Classify new samples using the trained network with:
net.classify(sample)Pull requests are welcome. For major changes, please open an issue first to discuss what you would like to change.
Please make sure to update tests as appropriate.