def test_initialization(self):
instance = nn.binaryClassifier(2, 2)
# Test that the first "input" has 3 sigmoids
assert len(instance.layers[0]) == 3
# Test that the sigmoids in the first layer have 3 weights
assert len(instance.layers[0][0]) == 3
# Test that the second layer has 3 sigmoids
assert len(instance.layers[1]) == 3
# Test that sigmoids in the second layer have 4 weights
assert len(instance.layers[1][0]) == 4
def test_randomNudge(self):
instance = nn.binaryClassifier(2, 2)
newNN = instance.randomNudge()
numberDifferent = 0
for layer in range(instance.numLayers):
for sigmoid in range(len(instance.layers[layer])):
for weight in range(len(instance.layers[layer][sigmoid])):
if not instance.layers[layer][sigmoid][weight] == newNN.layers[layer][sigmoid][weight]:
numberDifferent += 1
assert numberDifferent == 1
def test_sigmoid(self):
assert nn.binaryClassifier(2, 2).sigmoid(0) == 0.5
def test_randomArray(self):
instance = nn.binaryClassifier(2, 2).randomArray(1)
assert instance[0] <= 1
assert instance[0] >= -1