"""
Utility used by the Network class to actually train.
Based on:
https://github.com/fchollet/keras/blob/master/examples/mnist_mlp.py
"""
from keras.datasets import mnist, cifar10
from keras.models import Sequential
from keras.layers import Dense, Dropout
from keras.utils.np_utils import to_categorical
from keras.callbacks import EarlyStopping
# Helper: Early stopping.
early_stopper = EarlyStopping(patience=5)
def get_cifar10():
"""Retrieve the CIFAR dataset and process the data."""
# Set defaults.
nb_classes = 10
batch_size = 64
input_shape = (3072,)
# Get the data.
(x_train, y_train), (x_test, y_test) = cifar10.load_data()
x_train = x_train.reshape(50000, 3072)
x_test = x_test.reshape(10000, 3072)
x_train = x_train.astype('float32')
x_test = x_test.astype('float32')
x_train /= 255
x_test /= 255
# convert class vectors to binary class matrices
y_train = to_categorical(y_train, nb_classes)
y_test = to_categorical(y_test, nb_classes)
return (nb_classes, batch_size, input_shape, x_train, x_test, y_train, y_test)
def get_mnist():
"""Retrieve the MNIST dataset and process the data."""
# Set defaults.
nb_classes = 10
batch_size = 128
input_shape = (784,)
# Get the data.
(x_train, y_train), (x_test, y_test) = mnist.load_data()
x_train = x_train.reshape(60000, 784)
x_test = x_test.reshape(10000, 784)
x_train = x_train.astype('float32')
x_test = x_test.astype('float32')
x_train /= 255
x_test /= 255
# convert class vectors to binary class matrices
y_train = to_categorical(y_train, nb_classes)
y_test = to_categorical(y_test, nb_classes)
return (nb_classes, batch_size, input_shape, x_train, x_test, y_train, y_test)
def compile_model(network, nb_classes, input_shape):
"""Compile a sequential model.
Args:
network (dict): the parameters of the network
Returns:
a compiled network.
"""
# Get our network parameters.
nb_layers = network['nb_layers']
nb_neurons = network['nb_neurons']
activation = network['activation']
optimizer = network['optimizer']
model = Sequential()
# Add each layer.
for i in range(nb_layers):
# Need input shape for first layer.
if i == 0:
model.add(Dense(nb_neurons, activation=activation, input_shape=input_shape))
else:
model.add(Dense(nb_neurons, activation=activation))
model.add(Dropout(0.2)) # hard-coded dropout
# Output layer.
model.add(Dense(nb_classes, activation='softmax'))
model.compile(loss='categorical_crossentropy', optimizer=optimizer,
metrics=['accuracy'])
return model
def train_and_score(network, dataset):
"""Train the model, return test loss.
Args:
network (dict): the parameters of the network
dataset (str): Dataset to use for training/evaluating
"""
if dataset == 'cifar10':
nb_classes, batch_size, input_shape, x_train, \
x_test, y_train, y_test = get_cifar10()
elif dataset == 'mnist':
nb_classes, batch_size, input_shape, x_train, \
x_test, y_train, y_test = get_mnist()
model = compile_model(network, nb_classes, input_shape)
model.fit(x_train, y_train,
batch_size=batch_size,
epochs=10000, # using early stopping, so no real limit
verbose=0,
validation_data=(x_test, y_test),
callbacks=[early_stopper])
score = model.evaluate(x_test, y_test, verbose=0)
return score[1] # 1 is accuracy. 0 is loss.
