def backprop(self, a, y):
gradients = neur.backprop(a, y, [self.encode_weights, self.decode_weights], 0)
if self.contractive:
reg = 0.01 * self.contractive_reg_gradient(a[1][:,1:])
gradients[0][:,1:] += reg[:,1:]
gradients[1][:,1:] += reg[:,1:].transpose()
return gradients
def test_backprop(self):
X = np.array([[1, 2, 3], [2, 3, 4]])
y = np.array([[0, 0, 1], [0, 0, 1]])
res, a = neur.forward_prop(X, [self.theta, self.theta2])
grads = neur.backprop(a, y, [self.theta, self.theta2], 0)
grads_check = neur.gradient_check(X, y, [self.theta, self.theta2], neur.logistic_squared_cost_function)
self.equalish(grads[0], grads_check[0])
self.equalish(grads[1], grads_check[1])
def test_backprop(self):
X = np.array([[1, 2, 3], [2, 3, 4]])
y = np.array([[0, 0, 1], [0, 0, 1]])
res, a = neur.forward_prop(X, [self.theta, self.theta2])
grads = neur.backprop(a, y, [self.theta, self.theta2], 0)
grads_check = neur.gradient_check(X, y, [self.theta, self.theta2],
neur.logistic_squared_cost_function)
self.equalish(grads[0], grads_check[0])
self.equalish(grads[1], grads_check[1])
def backprop(self, a, y):
return neur.backprop(a, y, [self.encode_weights, self.decode_weights], 0)
def backprop(self, a, y):
return neur.backprop(a, y, [self.encode_weights, self.decode_weights],
0)
