def setUp(self):
super(LSTMTestCase, self).setUp()
data, labels = reber.make_reber_classification(n_samples=100,
return_indeces=True)
data = add_padding(data + 1) # +1 to shift indeces
self.data = train_test_split(data, labels, test_size=0.2)
self.n_categories = len(reber.avaliable_letters) + 1
self.n_time_steps = self.data[0].shape[1]
data_matrix[i, -len(sample):] = sample
return data_matrix
# An example of possible values for the `data` and `labels`
# variables
#
# >>> data
# array([array([1, 3, 1, 4]),
# array([0, 3, 0, 3, 0, 4, 3, 0, 4, 4]),
# array([0, 3, 0, 0, 3, 0, 4, 2, 4, 1, 0, 4, 0])], dtype=object)
# >>>
# >>> labels
# array([1, 0, 0])
data, labels = reber.make_reber_classification(n_samples=10000,
return_indeces=True)
# Shift all indeces by 1. In the next row we will add zero
# paddings, so we need to make sure that we will not confuse
# paddings with zero indeces.
data = data + 1
# Add paddings at the beggining of each vector to make sure
# that all samples has the same length. This trick allows to
# train network with multiple independent samples.
data = add_padding(data)
x_train, x_test, y_train, y_test = train_test_split(data,
labels,
train_size=0.8)
