class MLP:
def __init__(self, input_size, hidden_size, output_size):
self.input_layer = InputLayer(input_size, hidden_size, Sigmoid)
self.output_layer = SigmoidOutputLayer(hidden_size, output_size)
def predict(self, x):
x = self.input_layer.forward(x)
pred = self.output_layer.predict(x)
pred[pred >= 0.5] = 1
pred[pred < 0.5] = 0
return pred
def loss(self, x, y):
x = self.input_layer.forward(x)
loss = self.output_layer.forward(x, y)
return loss
def gradient(self):
d_prev = 1
d_prev = self.output_layer.backward(d_prev=d_prev)
self.input_layer.backward(d_prev)
def update(self, learning_rate):
self.input_layer.W -= self.input_layer.dW * learning_rate
self.input_layer.b -= self.input_layer.db * learning_rate
self.output_layer.W -= self.output_layer.dW * learning_rate
self.output_layer.b -= self.output_layer.db * learning_rate
Predict:
[[0.90437415]
[0.00750045]]
Forward:
0.05402041769476528
Backward:
[[-0.05371304 0.10152596 -0.14933889]
[ 0.004213 -0.00796322 0.01171345]]
dW:
[[-0.05306324]
[ 0.0574135 ]
[ 0.00375022]]
db:
[-0.0440627]
"""
sigmoid_output_layer = SigmoidOutputLayer(3, 1)
sigmoid_output_layer.W = np.array([[1.1234, -2.1234, 3.1234]]).T
sigmoid_output_layer.b = np.array([-1])
temp_y_hat = np.array([0.0, 0.5, 1.0])
temp_y = np.array([1, 1, 0])
print('Binary Cross-entropy Loss: \n',
sigmoid_output_layer.binary_ce_loss(temp_y_hat, temp_y))
temp8 = np.array([[1, -1, 0], [-1.4, 2.56, 1]])
print('Predict: \n', sigmoid_output_layer.predict(temp8))
temp_y = np.array([[1], [0]])
print('Forward: \n', sigmoid_output_layer.forward(temp8, temp_y))
print('Backward: \n', sigmoid_output_layer.backward(d_prev=1))
print('dW: \n', sigmoid_output_layer.dW)
def __init__(self, input_size, hidden_size, output_size):
self.input_layer = InputLayer(input_size, hidden_size, Sigmoid)
self.output_layer = SigmoidOutputLayer(hidden_size, output_size)