def test_minimal():
# sample fitness function
def eval_fitness(population):
for individual in population:
individual.fitness = 1.0
# creates the population
local_dir = os.path.dirname(__file__)
config = Config(os.path.join(local_dir, 'test_configuration'))
config.save_best = True
pop = Population(config)
# run the simulation for up to 20 generations
pop.run(eval_fitness, 20)
# Test save_checkpoint with defaults
pop.save_checkpoint()
# get statistics
best_unique = pop.statistics.best_unique_genomes(1)
best_unique = pop.statistics.best_unique_genomes(2)
avg_fitness = pop.statistics.get_average_fitness()
assert len(avg_fitness) == 1
assert all(f == 1 for f in avg_fitness)
# Change fitness threshold and do another run.
config.max_fitness_threshold = 1.1
pop = Population(config)
# runs the simulation for 20 generations
pop.run(eval_fitness, 20)
# get statistics
avg_fitness = pop.statistics.get_average_fitness()
assert len(avg_fitness) == 20
assert all(f == 1 for f in avg_fitness)
def main():
print("main")
config = Config(config_path)
# config.tcp_port += np.random.randint(0, 1000)
config.cluster_evaluation = False
# config.cluster_nodes_loc = "/home/adam/Workspace/pycharm/neat/nodes/"
# config.cluster_main_loc = "/home/adam/Workspace/pycharm/neat/nodes/"
# config.work_dir = "/home/adam/Results/local_0/"
# config.ray_info_loc = "/home/cec0113/ray_info/info"
# config.cluster_evaluation = True
# ray.init()
if os.path.isfile(config.work_dir + "autosave"):
config = Neat.open(config.work_dir + "autosave").config
print("Starting neat")
if os.path.isfile(config.work_dir + "autosave"):
print("opening")
neat = Neat.open(config.work_dir + "autosave")
else:
print("initing")
neat = Neat(config)
neat.add_observer(CsvObserver(neat.config))
neat.add_observer(ConsoleObserver(neat.config.work_dir, neat.config))
neat.add_observer(PlotObserver(neat.config.work_dir, 25))
neat.add_observer(RenderObserver(25))
# neat.add_observer(AutosaveObserver(neat.config.work_dir, neat.config.walltime_sec, 25))
print("start")
print(neat.config.dataset_name)
neat.start()
def __init__(self, config, initial_population=None):
"""
:param config: Either a config.Config object or path to a configuration file.
:param initial_population: Either an initial set of Genome instances to be used
as the initial population, or None, in which case a randomized set of Genomes
will be created automatically based on the configuration parameters.
"""
# If config is not a Config object, assume it is a path to the config file.
if not isinstance(config, Config):
config = Config(config)
# Configure statistics and reporting as requested by the user.
self.reporters = ReporterSet()
if config.collect_statistics:
self.statistics = StatisticsReporter()
self.add_reporter(self.statistics)
else:
self.statistics = None
if config.report:
self.add_reporter(StdOutReporter())
self.config = config
self.reproduction = config.reproduction_type(config, self.reporters)
self.species = SpeciesSet(config)
self.generation = -1
self.total_evaluations = 0
# Create a population if one is not given, then partition into species.
if initial_population is None:
initial_population = self.reproduction.create_new(config.pop_size)
self.species.speciate(initial_population)
def run():
# load settings file
local_dir = os.path.dirname(__file__)
config = Config(os.path.join(local_dir, 'dpole_config'))
# change the number of inputs accordingly to the type
# of experiment: markov (6) or non-markov (3)
# you can also set the configs in dpole_config as long
# as you have two config files for each type of experiment
config.input_nodes = 3
pop = population.Population(config)
pop.epoch(evaluate_population, 200, report=1, save_best=0)
winner = pop.most_fit_genomes[-1]
print('Number of evaluations: {0:d}'.format(winner.ID))
print('Winner fitness: {0:f}'.format(winner.fitness))
# save the winner
with open('winner_chromosome', 'w') as f:
cPickle.dump(winner, f)
# Plots the evolution of the best/average fitness
visualize.plot_stats(pop, ylog=True)
# Visualizes speciation
visualize.plot_species(pop)
# visualize the best topology
visualize.draw_net(winner, view=True)
def __init__(self, config, initial_population=None):
"""
:param config: Either a config.Config object or path to a configuration file.
:param initial_population: Either an initial set of Genome instances to be used
as the initial population, or None, in which case a randomized set of Genomes
will be created automatically based on the configuration parameters.
"""
# If config is not a Config object, assume it is a path to the config file.
if not isinstance(config, Config):
config = Config(config)
# Configure statistics and reporting as requested by the user.
self.reporters = ReporterSet()
if config.collect_statistics:
self.statistics = StatisticsReporter()
self.add_reporter(self.statistics)
else:
self.statistics = None
if config.report:
self.add_reporter(StdOutReporter())
self.config = config
self.reproduction = config.reproduction_type(config, self.reporters)
self.species = SpeciesSet(config)
self.generation = -1
self.total_evaluations = 0
# Create a population if one is not given, then partition into species.
if initial_population is None:
initial_population = self.reproduction.create_new(config.pop_size)
self.species.speciate(initial_population)
def __init__(self, config: Config, initial_state=None):
self.reporters: ReporterSet = ReporterSet()
self.config = config
stagnation: DefaultStagnation = config.stagnation_type(
config.stagnation_config, self.reporters)
self.reproduction: DefaultReproduction = config.reproduction_type(
config.reproduction_config, self.reporters, stagnation)
if config.fitness_criterion == "max":
self.fitness_criterion = max
elif config.fitness_criterion == "min":
self.fitness_criterion = min
elif config.fitness_criterion == "mean":
self.fitness_criterion = mean
elif not config.no_fitness_termination:
raise RuntimeError("Unexpected fitness_criterion: {0!r}".format(
config.fitness_criterion))
if initial_state is None:
# Create a population from scratch, then partition into species.
self.population: Dict[
int, DefaultGenome] = self.reproduction.create_new(
config.genome_type, config.genome_config, config.pop_size)
self.species_set: DefaultSpeciesSet = config.species_set_type(
config.species_set_config, self.reporters)
self.generation: int = 0
self.species_set.speciate(config, self.population, self.generation)
else:
self.population, self.species_set, self.generation = initial_state
self.best_genome: Optional[DefaultGenome] = None
def __init__(self, config, initial_population=None):
"""
:param config: Either a config.Config object or path to a configuration file.
:param initial_population:
"""
# If config is not a Config object, assume it is a path to the config file.
if not isinstance(config, Config):
config = Config(config)
self.config = config
self.diversity = config.diversity_type(self.config)
self.species_indexer = Indexer(1)
self.genome_indexer = Indexer(1)
self.species = []
self.generation_statistics = []
self.most_fit_genomes = []
self.generation = -1
self.total_evaluations = 0
if initial_population is None:
initial_population = self._create_population()
# Partition the population into species based on current configuration.
self._speciate(initial_population)
def run():
# load settings file
local_dir = os.path.dirname(__file__)
config = Config(os.path.join(local_dir, 'dpole_config'))
# change the number of inputs accordingly to the type
# of experiment: markov (6) or non-markov (3)
# you can also set the configs in dpole_config as long
# as you have two config files for each type of experiment
config.input_nodes = 3
pop = population.Population(config)
pop.epoch(evaluate_population, 200, report=1, save_best=0)
winner = pop.most_fit_genomes[-1]
print('Number of evaluations: {0:d}'.format(winner.ID))
print('Winner fitness: {0:f}'.format(winner.fitness))
# save the winner
with open('winner_chromosome', 'w') as f:
cPickle.dump(winner, f)
# Plots the evolution of the best/average fitness
visualize.plot_stats(pop, ylog=True)
# Visualizes speciation
visualize.plot_species(pop)
# visualize the best topology
visualize.draw_net(winner, view=True)
def test_minimal():
# sample fitness function
def eval_fitness(population):
for individual in population:
individual.fitness = 1.0
# creates the population
local_dir = os.path.dirname(__file__)
config = Config(os.path.join(local_dir, 'test_configuration'))
pop = Population(config)
# run the simulation for up to 20 generations
pop.run(eval_fitness, 20)
# get statistics
avg_fitness = pop.statistics.get_average_fitness()
assert len(avg_fitness) == 1
assert all(f == 1 for f in avg_fitness)
# Change fitness threshold and do another run.
config.max_fitness_threshold = 1.1
pop = Population(config)
# runs the simulation for 20 generations
pop.run(eval_fitness, 20)
# get statistics
avg_fitness = pop.statistics.get_average_fitness()
assert len(avg_fitness) == 20
assert all(f == 1 for f in avg_fitness)
def neat_start(run: int, p_call: bool):
print("main")
config = Config(config_path)
config.work_dir = config.work_dir[:-1] + "_" + str(run) + "/"
neat = None
try:
if os.path.isfile(config.work_dir + "autosave"):
neat = Neat.open(config.work_dir + "autosave")
except EOFError:
if os.path.isfile(config.work_dir + "autosave_prev"):
neat = Neat.open(config.work_dir + "autosave_prev")
if neat is not None:
config = neat.config
if config.done:
neat_start(run + 1, p_call)
files = os.listdir(config.cluster_main_loc)
hostname = socket.gethostname()
if any(hostname in h for h in files):
if p_call:
if config.cluster_evaluation:
print("Starting node client")
config.mp_max_proc = config.cluster_main_max_load
ClusterNode(config, True)
time.sleep(5) # Wait for other cluster nodes to startup...
subprocess.call([
"qsub", "-q", "qexp", "-l", "select=8:ncpus=24",
"/home/cec0113/myjob_multi"
])
else:
subprocess.call([
"qsub", "-q", "qexp", "-l", "select=1:ncpus=24",
"/home/cec0113/myjob_multi"
])
p_call = False
print("Starting neat")
if neat is None:
print("initing")
neat = Neat(config)
neat.add_observer(CsvObserver(neat.config))
neat.add_observer(
ConsoleObserver(neat.config.work_dir, neat.config))
neat.add_observer(PlotObserver(neat.config.work_dir, 25))
neat.add_observer(
AutosaveObserver(neat.config.work_dir,
neat.config.walltime_sec, 25))
print("run")
print("Start")
neat.start()
neat_start(run + 1, p_call)
else:
print("Starting node")
ClusterNode(config).run()
def run():
# Load the config file, which is assumed to live in
# the same directory as this script.
local_dir = os.path.dirname(__file__)
config = Config(os.path.join(local_dir, 'xor2_config'))
# For spiking networks
config.output_nodes = 2
pop = population.Population(config)
pop.epoch(eval_fitness, 500)
winner = pop.most_fit_genomes[-1]
print 'Number of evaluations: %d' % winner.ID
# Verify network output against training data.
print '\nBest network output:'
net = iznn.create_phenotype(winner)
for inputData, outputData in zip(INPUTS, OUTPUTS):
for j in range(MAX_TIME):
output = net.advance([x * 10 for x in inputData])
if output != [False, False]:
break
if output[0] and not output[1]: # Network answered 1
print "%r expected %d got 1" % (inputData, outputData)
elif not output[0] and output[1]: # Network answered 0
print "%r expected %d got 0" % (inputData, outputData)
else:
print "%r expected %d got ?" % (inputData, outputData)
# Visualize the winner network and plot statistics.
visualize.plot_stats(pop.most_fit_genomes, pop.avg_fitness_scores)
visualize.plot_species(pop.species_log)
visualize.draw_net(winner, view=True)
def run():
# Load the config file, which is assumed to live in
# the same directory as this script.
local_dir = os.path.dirname(__file__)
config = Config(os.path.join(local_dir, 'xor2_config'))
# For this network, we use two output neurons and use the difference between
# the "time to first spike" to determine the network response. There are
# probably a great many different choices one could make for an output encoding,
# and this choice may not be the best for tackling a real problem.
config.output_nodes = 2
pop = population.Population(config)
pop.run(eval_fitness, 200)
print('Number of evaluations: {0}'.format(pop.total_evaluations))
# Display the winning genome.
winner = pop.statistics.best_genome()
print('\nBest genome:\n{!s}'.format(winner))
# Show output of the most fit genome against training data.
winner = pop.statistics.best_genome()
print('\nBest genome:\n{!s}'.format(winner))
print('\nBest network output:')
net = iznn.create_phenotype(winner, *iz_params)
dt = iz_params[-1]
for inputData, outputData in zip(xor_inputs, xor_outputs):
neuron_data = {}
for i, n in net.neurons.items():
neuron_data[i] = []
# Reset the network, apply the XOR inputs, and run for the maximum allowed time.
net.reset()
net.set_inputs(inputData)
t0 = None
t1 = None
v0 = None
v1 = None
num_steps = int(max_time / dt)
for j in range(num_steps):
t = dt * j
output = net.advance()
# Capture the time and neuron membrane potential for later use if desired.
for i, n in net.neurons.items():
neuron_data[i].append((t, n.v))
# Remember time and value of the first output spikes from each neuron.
if t0 is None and output[0] > 0:
t0, v0 = neuron_data[net.outputs[0]][-2]
if t1 is None and output[1] > 0:
t1, v1 = neuron_data[net.outputs[1]][-2]
response = compute_output(t0, t1)
print(inputData, response)
def run():
# Load the config file, which is assumed to live in
# the same directory as this script.
local_dir = os.path.dirname(__file__)
config = Config(os.path.join(local_dir, 'xor2_config'))
# For this network, we use two output neurons and use the difference between
# the "time to first spike" to determine the network response. There are
# probably a great many different choices one could make for an output encoding,
# and this choice may not be the best for tackling a real problem.
config.output_nodes = 2
pop = population.Population(config)
pop.run(eval_fitness, 200)
print('Number of evaluations: {0}'.format(pop.total_evaluations))
# Visualize the winner network and plot statistics.
winner = pop.statistics.best_genome()
node_names = {0: 'A', 1: 'B', 2: 'Out1', 3: 'Out2'}
visualize.draw_net(winner, view=True, node_names=node_names)
visualize.plot_stats(pop.statistics)
visualize.plot_species(pop.statistics)
# Verify network output against training data.
print('\nBest network output:')
net = iznn.create_phenotype(winner, *iz_params)
sum_square_error, simulated = simulate(winner)
# Create a plot of the traces out to the max time for each set of inputs.
plt.figure(figsize=(12, 12))
for r, (inputData, outputData, t0, t1, v0, v1, neuron_data) in enumerate(simulated):
response = compute_output(t0, t1)
print("{0!r} expected {1:.3f} got {2:.3f}".format(inputData, outputData, response))
axes = plt.subplot(4, 1, r + 1)
plt.title("Traces for XOR input {{{0:.1f}, {1:.1f}}}".format(*inputData), fontsize=12)
for i, s in neuron_data.items():
if i in net.outputs:
t, v = zip(*s)
plt.plot(t, v, "-", label="neuron {0:d}".format(i))
# Circle the first peak of each output.
circle0 = patches.Ellipse((t0, v0), 1.0, 10.0, color='r', fill=False)
circle1 = patches.Ellipse((t1, v1), 1.0, 10.0, color='r', fill=False)
axes.add_artist(circle0)
axes.add_artist(circle1)
plt.ylabel("Potential (mv)", fontsize=10)
plt.ylim(-100, 50)
plt.tick_params(labelsize=8)
plt.grid()
plt.xlabel("Time (in ms)", fontsize=10)
plt.tight_layout(pad=0.4, w_pad=0.5, h_pad=1.0)
plt.savefig("traces.png", dpi=90)
plt.show()
def run():
# Load the config file, which is assumed to live in
# the same directory as this script.
local_dir = os.path.dirname(__file__)
config = Config(os.path.join(local_dir, 'xor2_config'))
# For this network, we use two output neurons and use the difference between
# the "time to first spike" to determine the network response. There are
# probably a great many different choices one could make for an output encoding,
# and this choice may not be the best for tackling a real problem.
config.output_nodes = 2
pop = population.Population(config)
pop.run(eval_fitness, 200)
print('Number of evaluations: {0}'.format(pop.total_evaluations))
# Display the winning genome.
winner = pop.statistics.best_genome()
print('\nBest genome:\n{!s}'.format(winner))
# Show output of the most fit genome against training data.
winner = pop.statistics.best_genome()
print('\nBest genome:\n{!s}'.format(winner))
print('\nBest network output:')
# Create a network of "fast spiking" Izhikevich neurons, simulation time step 0.25 millisecond.
net = iznn.create_phenotype(winner, **neuron_params)
for inputData, outputData in zip(xor_inputs, xor_outputs):
neuron_data = {}
for i, n in net.neurons.items():
neuron_data[i] = []
# Reset the network, apply the XOR inputs, and run for the maximum allowed time.
net.reset()
net.set_inputs(inputData)
t0 = None
t1 = None
num_steps = int(max_time / dt)
for j in range(num_steps):
t = dt * j
output = net.advance(dt)
# Capture the time and neuron membrane potential for later use if desired.
for i, n in net.neurons.items():
neuron_data[i].append((t, n.v))
# Remember time and value of the first output spikes from each neuron.
if t0 is None and output[0] > 0:
t0, v0 = neuron_data[net.outputs[0]][-2]
if t1 is None and output[1] > 0:
t1, v1 = neuron_data[net.outputs[1]][-2]
response = compute_output(t0, t1)
print(inputData, response)
def setUp(self):
local_dir = os.path.dirname(__file__)
indexer = InnovationIndexer(0)
config = Config(os.path.join(local_dir, 'config.txt'))
self.genome_config = {
'inputs': ['in0', 'in1'],
'outputs': [('out0', 'sigmoid')],
'modules': []
}
config.genome_config = self.genome_config
self.genome = CellGenome.create_unconnected(1, config)
def run():
# Load the config file, which is assumed to live in
# the same directory as this script.
local_dir = os.path.dirname(__file__)
config = Config(os.path.join(local_dir, 'xor2_config'))
# For this network, we use two output neurons and use the difference between
# the "time to first spike" to determine the network response. There are
# probably a great many different choices one could make for an output encoding,
# and this choice may not be the best for tackling a real problem.
config.output_nodes = 2
pop = population.Population(config)
pop.epoch(eval_fitness, 200)
winner = pop.most_fit_genomes[-1]
print('Number of evaluations: %d' % winner.ID)
# Visualize the winner network and plot statistics.
visualize.plot_stats(pop.most_fit_genomes, pop.avg_fitness_scores)
visualize.plot_species(pop.species_log)
visualize.draw_net(winner, view=True)
# Verify network output against training data.
print('\nBest network output:')
net = iznn.create_phenotype(winner, *iz_params)
error, simulated = simulate(winner)
# Create a plot of the traces out to the max time for each set of inputs.
plt.figure(figsize=(12,12))
for r, (inputData, outputData, t0, t1, neuron_data) in enumerate(simulated):
response = compute_output(t0, t1)
print("%r expected %.3f got %.3f" % (inputData, outputData, response))
plt.subplot(4, 1, r + 1)
plt.title("Traces for XOR input {%.1f, %.1f}" % inputData, fontsize=12)
for i, s in neuron_data.items():
if i not in net.inputs:
t, v = zip(*s)
plt.plot(t, v, "-", label="neuron %d" % i)
plt.ylabel("Potential (mv)", fontsize=10)
plt.ylim(-100, 50)
plt.tick_params(axis='both', labelsize=8)
plt.grid()
plt.xlabel("Time (in ms)", fontsize=10)
plt.tight_layout(pad=0.4, w_pad=0.5, h_pad=1.0)
plt.savefig("traces.png", dpi=90)
plt.show()
def getAction(self, gameState):
legalActions = gameState.getLegalActions(0)
if Directions.STOP in legalActions:
legalActions.remove(Directions.STOP)
inputs = []
outputs = legalActions
# Load the config file, which is assumed to live in
# the same directory as this script.
local_dir = os.path.dirname(__file__)
config = Config(os.path.join(local_dir, 'nn_config'))
pop = population.Population(config)
pop.epoch(fitness_function, 1000)
def eval_fitness(genomes):
for g in genomes:
net = nn.create_feed_forward_phenotype(g)
g.fitness = gamestate.score
output = net.serial_activate(intputs)
"*** YOUR CODE HERE ***"
# util.raiseNotDefined()
expectimax = self.expectimax(gameState, self.depth, 0)
# print "minimax({0},{1})" .format(expectimax[0], expectimax[1])
# print "GameState :\n{0}\n\n" .format(gameState)
return expectimax[1]
def test_run():
xor_inputs = [[0, 0], [0, 1], [1, 0], [1, 1]]
xor_outputs = [0, 1, 1, 0]
def eval_fitness(genomes):
for g in genomes:
net = nn.create_feed_forward_phenotype(g)
error = 0.0
for inputs, expected in zip(xor_inputs, xor_outputs):
# Serial activation propagates the inputs through the entire network.
output = net.serial_activate(inputs)
error += (output[0] - expected)**2
# When the output matches expected for all inputs, fitness will reach
# its maximum value of 1.0.
g.fitness = 1 - error
local_dir = os.path.dirname(__file__)
config = Config(os.path.join(local_dir, 'test_configuration'))
pop = population.Population(config)
pop.run(eval_fitness, 10)
winner = pop.statistics.best_genome()
# Validate winner.
for g in pop.statistics.most_fit_genomes:
assert winner.fitness >= g.fitness
statistics.save_stats(pop.statistics)
statistics.save_species_count(pop.statistics)
statistics.save_species_fitness(pop.statistics)
def check_self_crossover(genome_type):
# Check that self-crossover produces a genetically identical child (with a different ID).
local_dir = os.path.dirname(__file__)
config = Config(os.path.join(local_dir, 'test_configuration'))
c = genome_type.create_unconnected(1, config)
c.connect_full()
c.fitness = 0.0
cnew = c.crossover(c, 2)
assert cnew.ID != c.ID
assert len(cnew.conn_genes) == len(c.conn_genes)
for kold, vold in cnew.conn_genes.items():
print(kold, vold)
print(c.conn_genes)
assert kold in c.conn_genes
vnew = c.conn_genes[kold]
assert vold.is_same_innov(vnew)
assert vnew.weight == vold.weight
assert vnew.in_node_id == vold.in_node_id
assert vnew.out_node_id == vold.out_node_id
assert vnew.enabled == vold.enabled
assert len(cnew.node_genes) == len(c.node_genes)
for kold, vold in cnew.node_genes.items():
assert kold in c.node_genes
vnew = c.node_genes[kold]
assert vnew.ID == vold.ID
assert vnew.type == vold.type
assert vnew.bias == vold.bias
assert vnew.response == vold.response
assert vnew.activation_type == vold.activation_type
def main():
local_dir = os.path.dirname(__file__)
config = Config(neat.DefaultGenome, neat.DefaultReproduction,
neat.DefaultSpeciesSet, neat.DefaultStagnation,
os.path.join(local_dir, 'train_config.txt'))
config.save_best = True
config.checkpoint_time_interval = 3
pop = population.Population(config)
stats = neat.StatisticsReporter()
pop.add_reporter(stats)
pop.add_reporter(neat.StdOutReporter(True))
pop.add_reporter(neat.StatisticsReporter())
pop.add_reporter(neat.Checkpointer(2))
winner = pop.run(eval_fitness, 100)
with open('winner.pkl', 'wb') as f:
pickle.dump(winner, f)
def test_nonexistent_config():
"""Check that attempting to open a non-existent config file raises
an Exception with appropriate message."""
passed = False
try:
c = Config('wubba-lubba-dub-dub')
except Exception as e:
passed = 'No such config file' in str(e)
assert passed
def test_randomly_created_genome(self):
config = self.config
config.num_outputs = 3
config.num_inputs = 3
genome = Genome(config)
self.config = Config('../tests/conf_tester.conf')
assert len(genome.output_genes) == 3, \
"Genome should have 3 outputs, has %i" % len(genome.output_genes)
assert len(genome.input_genes) == 3, \
"Genome should have 3 inputs, has %i" % len(genome.input_genes)
def fit(self, input_matrix, target_vector, metric, verbose=False):
self.input_matrix = input_matrix
self.target_vector = target_vector
self.metric = metric
self._write_configuration()
self.configuration = Config(DefaultGenome, DefaultReproduction, DefaultSpeciesSet, DefaultStagnation, get_configuration_path())
remove_configuration()
self._run(verbose)
self.input_matrix = None
self.target_vector = None
def test_bad_config_activation():
"""Check that an unknown activation function raises an Exception with
the appropriate message."""
passed = False
try:
local_dir = os.path.dirname(__file__)
c = Config(os.path.join(local_dir, 'bad_configuration1'))
except Exception as e:
passed = 'Invalid activation function name' in str(e)
assert passed
def setUp(self):
self.config = Config('../tests/conf_tester.conf')
NodeGene.counter = 1
LinkGene.innovation_counter = 0
self.test_node_genes = [NodeGene(node_type='INPUT',),
NodeGene(node_type='INPUT'),
NodeGene(node_type='OUTPUT'),
NodeGene(node_type='OUTPUT')]
self.test_link_genes = [LinkGene(2, 3, weight=1), LinkGene(1, 3, weight=1)]
self.genome = Genome(self.config, node_genes=self.test_node_genes, link_genes=self.test_link_genes)
def test_config_options():
# sample fitness function
def eval_fitness(population):
for individual in population:
individual.fitness = 1.0
local_dir = os.path.dirname(__file__)
config = Config(os.path.join(local_dir, 'test_configuration'))
for hn in (0, 1, 2):
config.hidden_nodes = hn
for fc in (0, 1):
config.fully_connected = fc
for act in activation_functions.functions.keys():
config.allowed_activation = [act]
for ff in (0, 1):
config.feedforward = ff
pop = Population(config)
pop.run(eval_fitness, 250)
def main():
local_dir = os.path.dirname(__file__)
config = Config(neat.DefaultGenome, neat.DefaultReproduction,
neat.DefaultSpeciesSet, neat.DefaultStagnation,
os.path.join(local_dir, 'train_config.txt'))
with open('winner.pkl', 'rb') as f:
winner = pickle.load(f)
print('\nBest genome:\n{!s}'.format(winner))
print('\nOutput:')
winner_net = neat.nn.FeedForwardNetwork.create(winner, config)
print('Score:', dino_api.play_game(GetCommand(winner_net)))
def check_add_connection(genome_type, feed_forward):
local_dir = os.path.dirname(__file__)
config = Config(os.path.join(local_dir, 'test_configuration'))
config.input_nodes = 3
config.output_nodes = 4
config.hidden_nodes = 5
config.feedforward = feed_forward
N = config.input_nodes + config.hidden_nodes + config.output_nodes
connections = {}
for a in range(100):
g = genome_type.create_unconnected(a, config)
g.add_hidden_nodes(config.hidden_nodes)
for b in range(1000):
g.mutate_add_connection()
for c in g.conn_genes.values():
connections[c.key] = connections.get(c.key, 0) + 1
# TODO: The connections should be returned to the caller and checked
# against the constraints/assumptions particular to the network type.
for i in range(N):
values = []
for j in range(N):
values.append(connections.get((i, j), 0))
print("{0:2d}: {1}".format(i, " ".join("{0:3d}".format(x) for x in values)))
def main():
local_dir = os.path.dirname(__file__)
config = Config(neat.DefaultGenome, neat.DefaultReproduction,
neat.DefaultSpeciesSet, neat.DefaultStagnation,
os.path.join(local_dir, 'train_config.txt'))
config.save_best = True
config.checkpoint_time_interval = 3
initial_pop = []
# initial_pop = [138, 74, 40, 84, 97, 133, 127, 60, 102, 70, 0, 1, 2]
for root, dirs, files in os.walk('best'):
for file_name in files:
with open(os.path.join('best', file_name), 'rb') as f:
initial_pop.append(pickle.load(f))
if not initial_pop:
initial_pop = None
pop = population.Population(config, initial_pop)
stats = neat.StatisticsReporter()
pop.add_reporter(stats)
pop.add_reporter(neat.Checkpointer(5))
pop.run(eval_fitness, 100)
def check_simple(genome_type):
indexer = InnovationIndexer(0)
local_dir = os.path.dirname(__file__)
config = Config(os.path.join(local_dir, 'test_configuration'))
c1 = genome_type.create_unconnected(1, config)
c1.connect_full(indexer)
# add two hidden nodes
c1.add_hidden_nodes(2)
# apply some mutations
c1.mutate_add_node(indexer)
c1.mutate_add_connection(indexer)
def run():
# Load the config file, which is assumed to live in
# the same directory as this script.
local_dir = os.path.dirname(__file__)
config = Config(os.path.join(local_dir, 'xor2_config'))
config.report = False
pop = population.Population(config)
pop.run(eval_fitness, 1000)
winner = pop.statistics.best_genome()
num_hidden = 0
for ng in winner.node_genes.values():
if ng.type == 'HIDDEN':
num_hidden += 1
num_connections = len(winner.conn_genes)
num_enabled = 0
for cg in winner.conn_genes.values():
if cg.enabled:
num_enabled += 1
return pop.generation, num_hidden, num_connections, num_enabled
def check_simple(genome_type):
local_dir = os.path.dirname(__file__)
config = Config(os.path.join(local_dir, 'test_configuration'))
c1 = genome_type.create_unconnected(1, config)
c1.connect_full()
repr(c1)
str(c1)
# add two hidden nodes
c1.add_hidden_nodes(2)
repr(c1)
str(c1)
# apply some mutations
c1.mutate_add_node()
c1.mutate_add_connection()
repr(c1)
str(c1)
def test_checkpoint():
# sample fitness function
def eval_fitness(population):
for individual in population:
individual.fitness = 0.99
# creates the population
local_dir = os.path.dirname(__file__)
config = Config(os.path.join(local_dir, 'test_configuration'))
pop = Population(config)
pop.run(eval_fitness, 20)
t = tempfile.NamedTemporaryFile(delete=False)
t.close()
pop.save_checkpoint(t.name)
pop2 = Population(config)
pop2.load_checkpoint(t.name)
def test_evolve():
test_values = [random.random() for _ in range(10)]
def evaluate_genome(genomes):
for g in genomes:
net = ctrnn.create_phenotype(g)
fitness = 0.0
for t in test_values:
net.reset()
output = net.serial_activate([t])
expected = t ** 2
error = output[0] - expected
fitness -= error ** 2
g.fitness = fitness
# Load the config file, which is assumed to live in
# the same directory as this script.
local_dir = os.path.dirname(__file__)
config = Config(os.path.join(local_dir, 'ctrnn_config'))
config.node_gene_type = ctrnn.CTNodeGene
config.prob_mutate_time_constant = 0.1
config.checkpoint_time_interval = 0.1
config.checkpoint_gen_interval = 1
pop = population.Population(config)
pop.run(evaluate_genome, 10)
# Save the winner.
print('Number of evaluations: {0:d}'.format(pop.total_evaluations))
winner = pop.statistics.best_genome()
with open('winner_genome', 'wb') as f:
pickle.dump(winner, f)
repr(winner)
str(winner)
for g in winner.node_genes:
repr(g)
str(g)
for g in winner.conn_genes:
repr(g)
str(g)
def setUp(self):
self.config = Config('../tests/conf_tester.conf')
NodeGene.counter = 1
self.test_node_genes = [
NodeGene(node_type='INPUT', activation='identity'),
NodeGene(node_type='INPUT', activation='identity'),
NodeGene(node_type='OUTPUT', activation='identity'),
NodeGene(node_type='OUTPUT', activation='identity')
]
self.test_link_genes = [
LinkGene(1, 3, weight=1),
LinkGene(2, 3, weight=1),
LinkGene(1, 4, weight=1),
LinkGene(2, 4, weight=1),
LinkGene(0, 3, weight=1),
LinkGene(0, 4, weight=1)
]
self.genome = Genome(
self.config,
node_genes=self.test_node_genes,
link_genes=self.test_link_genes,
)
self.phenome = FeedForwardPhenome(self.genome, self.config)
def setUp(self):
local_dir = os.path.dirname(__file__)
self.indexer = InnovationIndexer(0)
self.config = Config(os.path.join(local_dir, 'config.txt'))
self.prob_add = 0
self.prob_remove = 0
self.min_genes = 2
self.max_genes = 3
self.genome_config = {
'inputs': ['in0', 'in1'],
'outputs': [('out0', 'sigmoid')],
'modules': [(Signal0Module, {
'prob_add': self.prob_add,
'prob_remove': self.prob_remove,
'min_genes': self.min_genes,
'max_genes': self.max_genes
})]
}
self.config.genome_config = self.genome_config
self.genome = CellGenome.create_unconnected(1, self.config)
self.genome.connect_full(self.indexer)
def test_config_options():
# sample fitness function
def eval_fitness(population):
for individual in population:
individual.fitness = 1.0
local_dir = os.path.dirname(__file__)
config = Config(os.path.join(local_dir, 'test_configuration'))
for hn in (0, 1, 2):
config.hidden_nodes = hn
for fc in (0, 1):
config.fully_connected = fc
for act in activation_functions.functions.keys():
config.allowed_activation = [act]
for ff in (0, 1):
config.feedforward = ff
pop = Population(config)
pop.run(eval_fitness, 250)
return pop.statistics, out
# return winner.fitness, out
def evolve_cppn(config, pattern, generations):
return evolve(config, express_cppn, pattern, generations)
def evolve_recurrent_cppn(config, pattern, max_steps, generations):
def eval_func(net, w, h):
return express_recurrent_cppn(net, w, h, max_steps)
return evolve(config, eval_func, pattern, generations)
if __name__ == '__main__':
import numpy as np
from neat.config import Config # Python neat library
from neat import nn
import patterns
target = patterns.checkerboard(16, 16, 4)
config = Config('config.txt')
config.pop_size = 50
config.hidden_nodes = 0
config.initial_connection = 'fully_connected'
config.input_nodes = 7
evolve_recurrent_cppn(config, target, 16, 50)
# without exceeding these limits.
if abs(sim.x) >= sim.position_limit or abs(sim.theta) >= sim.angle_limit_radians:
break
fitness += 1.0
fitnesses.append(fitness)
# The genome's fitness is its worst performance across all runs.
return min(fitnesses)
# Load the config file, which is assumed to live in
# the same directory as this script.
local_dir = os.path.dirname(__file__)
config = Config(os.path.join(local_dir, 'ctrnn_config'))
config.node_gene_type = ctrnn.CTNodeGene
pop = population.Population(config)
pe = parallel.ParallelEvaluator(4, evaluate_genome)
pop.run(pe.evaluate, 2000)
# Save the winner.
print('Number of evaluations: {0:d}'.format(pop.total_evaluations))
winner = pop.statistics.best_genome()
with open('ctrnn_winner_genome', 'wb') as f:
pickle.dump(winner, f)
print(winner)
# Plot the evolution of the best/average fitness.
def run():
# Load the config file, which is assumed to live in
# the same directory as this script.
local_dir = os.path.dirname(__file__)
config = Config(os.path.join(local_dir, 'xor2_config'))
# For this network, we use two output neurons and use the difference between
# the "time to first spike" to determine the network response. There are
# probably a great many different choices one could make for an output encoding,
# and this choice may not be the best for tackling a real problem.
config.output_nodes = 2
pop = population.Population(config)
pop.run(eval_fitness, 200)
print('Number of evaluations: {0}'.format(pop.total_evaluations))
# Visualize the winner network and plot statistics.
winner = pop.statistics.best_genome()
node_names = {0: 'A', 1: 'B', 2: 'Out1', 3: 'Out2'}
visualize.draw_net(winner, view=True, node_names=node_names)
visualize.plot_stats(pop.statistics)
visualize.plot_species(pop.statistics)
# Verify network output against training data.
print('\nBest network output:')
net = iznn.create_phenotype(winner, *iz_params)
sum_square_error, simulated = simulate(winner)
# Create a plot of the traces out to the max time for each set of inputs.
plt.figure(figsize=(12, 12))
for r, (inputData, outputData, t0, t1, v0, v1,
neuron_data) in enumerate(simulated):
response = compute_output(t0, t1)
print("{0!r} expected {1:.3f} got {2:.3f}".format(
inputData, outputData, response))
axes = plt.subplot(4, 1, r + 1)
plt.title(
"Traces for XOR input {{{0:.1f}, {1:.1f}}}".format(*inputData),
fontsize=12)
for i, s in neuron_data.items():
if i in net.outputs:
t, v = zip(*s)
plt.plot(t, v, "-", label="neuron {0:d}".format(i))
# Circle the first peak of each output.
circle0 = patches.Ellipse((t0, v0), 1.0, 10.0, color='r', fill=False)
circle1 = patches.Ellipse((t1, v1), 1.0, 10.0, color='r', fill=False)
axes.add_artist(circle0)
axes.add_artist(circle1)
plt.ylabel("Potential (mv)", fontsize=10)
plt.ylim(-100, 50)
plt.tick_params(labelsize=8)
plt.grid()
plt.xlabel("Time (in ms)", fontsize=10)
plt.tight_layout(pad=0.4, w_pad=0.5, h_pad=1.0)
plt.savefig("traces.png", dpi=90)
plt.show()
def test_iznn_evolve():
"""This is a stripped-down copy of the XOR2 spiking example."""
# Network inputs and expected outputs.
xor_inputs = ((0, 0), (0, 1), (1, 0), (1, 1))
xor_outputs = (0, 1, 1, 0)
# Maximum amount of simulated time (in milliseconds) to wait for the network to produce an output.
max_time = 50.0
def compute_output(t0, t1):
'''Compute the network's output based on the "time to first spike" of the two output neurons.'''
if t0 is None or t1 is None:
# If one of the output neurons failed to fire within the allotted time,
# give a response which produces a large error.
return -1.0
else:
# If the output neurons fire within 1.0 milliseconds of each other,
# the output is 1, and if they fire more than 11 milliseconds apart,
# the output is 0, with linear interpolation between 1 and 11 milliseconds.
response = 1.1 - 0.1 * abs(t0 - t1)
return max(0.0, min(1.0, response))
def simulate(genome):
# Create a network of Izhikevich neurons based on the given genome.
net = iznn.create_phenotype(genome, **iznn.THALAMO_CORTICAL_PARAMS)
dt = 0.25
sum_square_error = 0.0
simulated = []
for inputData, outputData in zip(xor_inputs, xor_outputs):
neuron_data = {}
for i, n in net.neurons.items():
neuron_data[i] = []
# Reset the network, apply the XOR inputs, and run for the maximum allowed time.
net.reset()
net.set_inputs(inputData)
t0 = None
t1 = None
v0 = None
v1 = None
num_steps = int(max_time / dt)
for j in range(num_steps):
t = dt * j
output = net.advance(dt)
# Capture the time and neuron membrane potential for later use if desired.
for i, n in net.neurons.items():
neuron_data[i].append((t, n.v))
# Remember time and value of the first output spikes from each neuron.
if t0 is None and output[0] > 0:
t0, v0 = neuron_data[net.outputs[0]][-2]
if t1 is None and output[1] > 0:
t1, v1 = neuron_data[net.outputs[1]][-2]
response = compute_output(t0, t1)
sum_square_error += (response - outputData) ** 2
simulated.append(
(inputData, outputData, t0, t1, v0, v1, neuron_data))
return sum_square_error, simulated
def eval_fitness(genomes):
for genome in genomes:
sum_square_error, simulated = simulate(genome)
genome.fitness = 1 - sum_square_error
# Load the config file, which is assumed to live in
# the same directory as this script.
local_dir = os.path.dirname(__file__)
config = Config(os.path.join(local_dir, 'test_configuration'))
# TODO: This is a little hackish, but will a user ever want to do it?
# If so, provide a convenience method on Config for it.
for i, tc in enumerate(config.type_config['DefaultStagnation']):
if tc[0] == 'species_fitness_func':
config.type_config['DefaultStagnation'][i] = (tc[0], 'median')
# For this network, we use two output neurons and use the difference between
# the "time to first spike" to determine the network response. There are
# probably a great many different choices one could make for an output encoding,
# and this choice may not be the best for tackling a real problem.
config.output_nodes = 2
pop = population.Population(config)
pop.run(eval_fitness, 10)
print('Number of evaluations: {0}'.format(pop.total_evaluations))
# Visualize the winner network and plot statistics.
winner = pop.statistics.best_genome()
# Verify network output against training data.
print('\nBest network output:')
net = iznn.create_phenotype(winner, **iznn.RESONATOR_PARAMS)
sum_square_error, simulated = simulate(winner)
repr(winner)
str(winner)
for g in winner.node_genes:
repr(g)
str(g)
for g in winner.conn_genes:
repr(g)
str(g)