def backprop(self, x, y):
nabla_b = [np.zeros(b.shape) for b in self.net.biases]
nabla_w = [np.zeros(w.shape) for w in self.net.weights]
# feedforward
activation = x
activations = [x] # list to store all the activations, layer by layer
zs = list() # list to store all the z vectors, layer by layer
for b, w in zip(self.net.biases, self.net.weights):
z = np.dot(w, activation)+b
zs.append(z)
activation = sigmoid(z)
activations.append(activation)
# backward pass
delta = (activations[-1]-y)*sigmoid_prime(zs[-1])
nabla_b[-1] = delta
nabla_w[-1] = np.dot(delta, activations[-2].transpose())
for l in xrange(2, len(self.net.n_neurons)):
z = zs[-l]
sp = sigmoid_prime(z)
delta = np.dot(self.net.weights[-l+1].transpose(), delta) * sp
nabla_b[-l] = delta
nabla_w[-l] = np.dot(delta, activations[-l-1].transpose())
return nabla_b, nabla_w
def feed_forward_fast(self, a):
for b, w in zip(self.biases, self.weights):
a = sigmoid(np.dot(w, a)+b)
return a
def activate(self):
self.z = sum([synapse.neuron_from.activity*synapse.weight for synapse in self.synapses_in]) + self.bias
self.activity = sigmoid(self.z)
