def setUp(self):
tf.reset_default_graph()
if self.single_thread:
sess = tensorflow_session()
sess.close()
config = tf.ConfigProto(
allow_soft_placement=True,
intra_op_parallelism_threads=1,
inter_op_parallelism_threads=1,
)
tensorflow_session.cache = tf.Session(config=config)
if not self.verbose:
logging.disable(logging.CRITICAL)
# Clean identifiers map for each test
if hasattr(format_name_if_specified_as_pattern, 'counters'):
del format_name_if_specified_as_pattern.counters
utils.reproducible(seed=self.random_seed)
def reproducible_network_train(seed=0, epochs=500, **additional_params):
"""
Make a reproducible train for Gradient Descent based neural
network with a XOR problem and return trained network.
Parameters
----------
seed : int
Random State seed number for reproducibility. Defaults to ``0``.
epochs : int
Number of epochs for training. Defaults to ``500``.
**additional_params
Aditional parameters for Neural Network.
Returns
-------
GradientDescent instance
Returns trained network.
"""
utils.reproducible(seed)
xor_x_train = np.array([[-1, -1], [-1, 1], [1, -1], [1, 1]])
xor_y_train = np.array([[1, -1, -1, 1]]).T
xavier_normal = init.XavierNormal()
tanh_weight1 = xavier_normal.sample((2, 5), return_array=True)
tanh_weight2 = xavier_normal.sample((5, 1), return_array=True)
network = algorithms.GradientDescent(network=[
layers.Input(2),
layers.Tanh(5, weight=tanh_weight1),
layers.Tanh(1, weight=tanh_weight2),
],
batch_size=None,
**additional_params)
network.train(xor_x_train, xor_y_train, epochs=epochs)
return network
import matplotlib.pyplot as plt
from sklearn import datasets
from neupy import algorithms, utils
from utils import plot_2d_grid
plt.style.use('ggplot')
utils.reproducible()
if __name__ == '__main__':
GRID_WIDTH = 20
GRID_HEIGHT = 1
data, targets = datasets.make_moons(n_samples=400, noise=0.1)
data = data[targets == 1]
sofm = algorithms.SOFM(
n_inputs=2,
features_grid=(GRID_HEIGHT, GRID_WIDTH),
verbose=True,
shuffle_data=True,
# The winning neuron will be selected based on the
# Euclidean distance. For this task it's important
# that distance is Euclidean. Other distances will
# not give us the same results.
distance='euclid',
learning_radius=2,
# Reduce learning radius by 1 after every 20 epochs.
# Learning radius will be equal to 2 during first
# 20 epochs and on the 21st epoch it will be equal to 1.
