def hand_crafted_xor(i1, i2):
s1, s2, bias = SensorNeuron(lambda: i1, 's1'), SensorNeuron(lambda: i2, 's2'), BiasNeuron('b1')
nand = Neuron('nand')
orr = Neuron('or')
and_result = Neuron('and=result')
net = Network((s1, s2, bias, nand, orr, and_result), lambda x: round(x))
synapses = ((s1, nand, -1), (s1, orr, 1),
(s2, nand, -1), (s2, orr, 1),
(bias, nand, 1.5), (bias, orr, -1), (bias, and_result, -1.25),
(nand, and_result, 1), (orr, and_result, 1))
for axon, dendrite, weight in synapses:
net.splice(axon, dendrite, weight)
return net.activate()
def should_fire_until_all_nodes_relaxed(self):
sensor = SensorNeuron(lambda: 10)
output = Neuron()
net = Network((sensor, output))
net.activate()
expect((sensor.fire(), output.fire())) == (RELAXED, RELAXED)