def test_representative(self):
expected = 0.880797
actual = sigmoid(2)
self.assertAlmostEqual(expected, actual, places=6)
expected = 0.268941
actual = sigmoid(-1)
self.assertAlmostEqual(expected, actual, places=6)
def test_infinity(self):
expected = 1
actual = sigmoid(float('inf'))
self.assertEqual(expected, actual)
expected = 0
actual = sigmoid(-float('inf'))
self.assertEqual(expected, actual)
def predict(network, x):
W1, W2, W3 = network['W1'], network['W2'], network['W3']
b1, b2, b3 = network['b1'], network['b2'], network['b3']
a1 = np.dot(x, W1) + b1
z1 = sigmoid(a1)
a2 = np.dot(z1, W2) + b2
z2 = sigmoid(a2)
a3 = np.dot(z2, W3) + b3
y = softmax(a3)
return y
def test_supports_ndarray(self):
expected = np.array([0.268941, 0.5, 0.731059])
actual = sigmoid(np.array([-1, 0, 1]))
np.testing.assert_allclose(expected, actual, atol=1e-6)
def test_representative(self):
expected = 0.952574
actual = sigmoid(3)
places = len(str(expected).split('.')[1])
self.assertAlmostEqual(expected, actual, places=places)
def swish(x):
return np.nan_to_num(x * sigmoid(x),
nan=0.0,
posinf=np.inf,
neginf=-np.inf)