def buildNet_big_layer():
substrate = NEAT.Substrate([(-1, -1), (-1, 1)],
[(0, i - (10 - 1) / 2) for i in xrange(10)],
[(1, 0), (1, 1)])
substrate.m_hidden_nodes_activation = NEAT.ActivationFunction.TANH
substrate.m_output_nodes_activation = NEAT.ActivationFunction.LINEAR
return substrate
def create_substrate(dim):
"""
The function to create two-sheets substrate configuration with specified
dimensions of each sheet.
Arguments:
dim: The dimensions accross X, Y axis of the sheet
"""
# Building sheet configurations of inputs and outputs
inputs = create_sheet_space(-1, 1, dim, -1)
outputs = create_sheet_space(-1, 1, dim, 0)
substrate = NEAT.Substrate( inputs,
[], # hidden
outputs)
substrate.m_allow_input_output_links = True
substrate.m_allow_input_hidden_links = False
substrate.m_allow_hidden_hidden_links = False
substrate.m_allow_hidden_output_links = False
substrate.m_allow_output_hidden_links = False
substrate.m_allow_output_output_links = False
substrate.m_allow_looped_hidden_links = False
substrate.m_allow_looped_output_links = False
substrate.m_hidden_nodes_activation = NEAT.ActivationFunction.SIGNED_SIGMOID
substrate.m_output_nodes_activation = NEAT.ActivationFunction.UNSIGNED_SIGMOID
substrate.m_with_distance = True
substrate.m_max_weight_and_bias = 3.0
return substrate
def buildNet_perceptron_mnist():
substrate = NEAT.Substrate([(-1, x, y) for x in xrange(28)
for y in xrange(28)], [],
[(1, 0, x) for x in xrange(10)])
substrate.m_allow_input_output_links = True
substrate.m_hidden_nodes_activation = NEAT.ActivationFunction.TANH
substrate.m_output_nodes_activation = NEAT.ActivationFunction.LINEAR
return substrate
def buildNet_big_layer_mnist():
substrate = NEAT.Substrate([(-1, x, y) for x in xrange(28)
for y in xrange(28)], [(0, x, y)
for x in xrange(20)
for y in xrange(20)],
[(1, 0, x) for x in xrange(10)])
substrate.m_hidden_nodes_activation = NEAT.ActivationFunction.TANH
substrate.m_output_nodes_activation = NEAT.ActivationFunction.LINEAR
return substrate
def grid2d_substrate(inputs: int, hidden_layers: int, nodes_per_layer: [int], outputs: int, leaky=True):
input_nodes = get_2d_point_line(inputs, (-1.0, -1.0), (-1.0, 1.0))
hidden_nodes = []
layers = get_2d_point_line(hidden_layers + 2, (-1, 0), (1, 0))
for i in range(hidden_layers):
layer_x = layers[i + 1][0]
hidden_nodes += get_2d_point_line(nodes_per_layer[i], (layer_x, -1.0), (layer_x, 1.0))
output_nodes = get_2d_point_line(outputs, (1.0, -1.0), (1.0, 1.0))
subst = neat.Substrate(input_nodes, hidden_nodes, output_nodes)
subst.m_query_weights_only = True
return subst
def create_substrate():
"""
The function to create appropriate substrate configuration with 16 inputs and 2 outputs.
"""
# The input layer
x_space = np.linspace(-1.0, 1.0, num=4)
inputs = [
(x_space[0], 0.0, 1.0),
(x_space[1], 0.0, 1.0),
(x_space[0], 0.0, -1.0),
(x_space[1], 0.0, -1.0), # the left side
(x_space[2], 0.0, 1.0),
(x_space[3], 0.0, 1.0),
(x_space[2], 0.0, -1.0),
(x_space[3], 0.0, -1.0), # the right side
(0, 0, 0) # the bias
]
# The output layer
outputs = [(-1.0, 1.0, 0.0), (1.0, 1.0, 0.0)]
substrate = NEAT.Substrate(
inputs,
[], # hidden
outputs)
# Allow connections: input-to-hidden, hidden-to-output, and hidden-to-hidden
substrate.m_allow_input_hidden_links = True
substrate.m_allow_hidden_output_links = True
substrate.m_allow_hidden_hidden_links = True
substrate.m_allow_input_output_links = False
substrate.m_allow_output_hidden_links = False
substrate.m_allow_output_output_links = False
substrate.m_allow_looped_hidden_links = False
substrate.m_allow_looped_output_links = False
substrate.m_hidden_nodes_activation = NEAT.ActivationFunction.SIGNED_SIGMOID
substrate.m_output_nodes_activation = NEAT.ActivationFunction.UNSIGNED_SIGMOID
substrate.m_with_distance = True # send connection length to the CPPN as a parameter
substrate.m_max_weight_and_bias = 8.0
return substrate
#!/usr/bin/env python3
import gym
import sys
import MultiNEAT as NEAT
import MultiNEAT.viz as viz
import random as rnd
import pickle
import numpy as np
from tqdm import tqdm
import cv2
substrate = NEAT.Substrate([(-1, -1), (-1, 0), (-1, 1)],
[(0, -1), (0, 0), (0, 1)],
[(1, 0)])
substrate.m_allow_input_hidden_links = False
substrate.m_allow_input_output_links = False
substrate.m_allow_hidden_hidden_links = False
substrate.m_allow_hidden_output_links = False
substrate.m_allow_output_hidden_links = False
substrate.m_allow_output_output_links = False
substrate.m_allow_looped_hidden_links = False
substrate.m_allow_looped_output_links = False
substrate.m_allow_input_hidden_links = True
substrate.m_allow_input_output_links = False
substrate.m_allow_hidden_output_links = True
substrate.m_allow_hidden_hidden_links = False
- Horizontal displacement
'''
genome = NEAT.Genome(0, 310, 30, 2, False,
NEAT.ActivationFunction.UNSIGNED_SIGMOID,
NEAT.ActivationFunction.UNSIGNED_SIGMOID, 0, params)
params = NEAT.Parameters()
params.PopulationSize = 20
params.DynamicCompatibility = True
pop = NEAT.Population(genome, params, True, 1.0, 0)
substrate = [((i / width) / height, (i % width) / width, 0.0)
for i in range(width * height)]
piecearray = [(i / 25, (i % 5) / 5, 1.0) for i in range(25)]
substrate.extend(piecearray * 4)
substrate = NEAT.Substrate(substrate, [], [((i - 7) / 7, ) for i in range(14)])
def evaluate(genome):
# create a neural network for the genome
net = NEAT.NeuralNetwork()
genome.BuildESHyperNEATPhenotype(net, substrate, params)
# create a board
board = Board(width, height)
canAddPiece = 1
current = 0
line = 0
maxPieces = 300
# while the game is not over
params.Width = 1.
params.Height = 1.
params.Elitism = 0.1
params.CrossoverRate = 0.5
params.MutateWeightsSevereProb = 0.01
params.MutateNeuronTraitsProb = 0
params.MutateLinkTraitsProb = 0
rng = NEAT.RNG()
rng.TimeSeed()
substrate = NEAT.Substrate([
(-1., -1., 0.0),
(-.5, -1., 0.0),
(0.0, -1., 0.0),
(.5, -1., 0.0),
(1.0, -1., 0.0),
(0.0, -1.0, -1.0),
], [], [(-1., 1., 0.0), (1.0, 1.0, 0.0)])
substrate.m_allow_input_hidden_links = False
substrate.m_allow_input_output_links = False
substrate.m_allow_hidden_hidden_links = False
substrate.m_allow_hidden_output_links = False
substrate.m_allow_output_hidden_links = False
substrate.m_allow_output_output_links = False
substrate.m_allow_looped_hidden_links = False
substrate.m_allow_looped_output_links = False
# let's set the activation functions
substrate.m_hidden_nodes_activation = NEAT.ActivationFunction.SIGNED_SIGMOID
params.BandThreshold = 0.3
params.InitialDepth = 3
params.MaxDepth = 3
params.IterationLevel = 1
params.Leo = True
params.LeoSeed = False
params.LeoThreshold = 0.3
params.CPPN_Bias = 1.0
params.Qtree_X = 0.0
params.Qtree_Y = 0.0
params.Width = 2.0
rng = NEAT.RNG()
rng.TimeSeed()
substrate = NEAT.Substrate([(-1., -1., 0.0), (1., -1., 0.0), (0., -1., 0.0)],
[], [(0., 1., 0.0)])
substrate.m_allow_input_hidden_links = False
substrate.m_allow_input_output_links = False
substrate.m_allow_hidden_hidden_links = False
substrate.m_allow_hidden_output_links = False
substrate.m_allow_output_hidden_links = False
substrate.m_allow_output_output_links = False
substrate.m_allow_looped_hidden_links = False
substrate.m_allow_looped_output_links = False
# let's set the activation functions
substrate.m_hidden_nodes_activation = NEAT.ActivationFunction.SIGNED_SIGMOID
substrate.m_output_nodes_activation = NEAT.ActivationFunction.SIGNED_SIGMOID
# when to output a link and max weight
# elif i>11 and i<22:
# meh.append(((i-11)/9,-1.,1.))
meh.append((0,0.5,-1))
meh.append((-0.5,0.5,-1))
meh.append((0.5,0.5,-1))
meh.append((0,-1,-1))
# else:
# meh.append((i/11.,-1.,1.))
# meh.append(((i-0)/41,-1,0))
# elif i >10 and i<14:
# meh.append(((i-0)/41,-1,0))
# else:
# meh.append((i/41,-1.,0.))
print(meh)
substrate = NEAT.Substrate(meh,#(1., -1, 1.),(0., -1, 1.)],
[(0., 0, 1.),(-1.,0,1.0),(1.,0,1.0)],
[(-1.9, 1., 1.),(1.9, 1., 1.)])#(-1.9, 1., 1.),(-1.3, 1., 1.),(-0.7, 1., 1.),(0, 1., 1.),(0.7, 1., 1.),(1.3, 1., 1.),(1.9, 1., 1.),])
substrate.m_allow_input_hidden_links = True
substrate.m_allow_input_output_links = True
substrate.m_allow_hidden_hidden_links = True
substrate.m_allow_hidden_output_links = True
substrate.m_allow_output_hidden_links = True
substrate.m_allow_output_output_links = True
substrate.m_allow_looped_hidden_links = True
substrate.m_allow_looped_output_links = True
substrate.m_allow_input_hidden_links = True
substrate.m_allow_input_output_links = True
substrate.m_allow_hidden_output_links = True
substrate.m_allow_hidden_hidden_links = True
cm = ax.contourf(pattern,
200,
cmap='autumn',
origin='lower',
extent=[-1, 1, -1, 1])
#for point in points:
# ax.add_patch(plt.Circle((point[0], point[1]), 0.04, fc='red'))
plt.show()
return
################################################################
substrate = NEAT.Substrate([(-1, -1, -1.0), (-0.66, -1.0, -1.0),
(0.66, -1.0, 1.0), (1., -1.0, 1.0), (0., -1., 0.)],
[], [(-0.5, 1, 0.), (0.5, 1, 0.0)])
'''
substrate = NEAT.Substrate([(-1., -1., 0.0), (-1., 1., 0.0), (-1., 0., 0.0)],
[],
[(1., 0., 0.0)])'''
def Scale(a, a_min, a_max, a_tr_min, a_tr_max):
t_a_r = a_max - a_min
if t_a_r == 0:
return a_max
t_r = a_tr_max - a_tr_min
rel_a = (a - a_min) / t_a_r
return a_tr_min + t_r * rel_a
import sys
sys.path.append("/home/peter")
sys.path.append("/home/peter/Desktop")
sys.path.append("/home/peter/Desktop/projects")
import time
import random as rnd
import commands as comm
import cv2
import numpy as np
import cPickle as pickle
import MultiNEAT as NEAT
import multiprocessing as mpc
# the simple 3D substrate with 3 input points, 2 hidden and 1 output for XOR
substrate = NEAT.Substrate([(-1, 1, 0), (1, 0, 0), (0, 1, 0)],
[(0.5, 0.5, 0.5), (-0.5, 1.5, 0.5)], [(0, 0, 1)])
# let's configure it a bit to avoid recurrence in the substrate
substrate.m_allow_hidden_hidden_links = False
substrate.m_allow_hidden_output_links = False
substrate.m_allow_looped_hidden_links = False
substrate.m_allow_looped_output_links = False
# let's set the activation functions
substrate.m_hidden_nodes_activation = NEAT.ActivationFunction.UNSIGNED_SIGMOID
substrate.m_outputs_nodes_activation = NEAT.ActivationFunction.UNSIGNED_SIGMOID
# when to output a link and max weight
substrate.m_link_threshold = 0.2
substrate.m_max_weight = 5.0
def buildNet_perceptron():
substrate = NEAT.Substrate([(-1, -1), (-1, 1)], [], [(1, 0), (1, 1)])
substrate.m_allow_input_output_links = True
substrate.m_hidden_nodes_activation = NEAT.ActivationFunction.TANH
substrate.m_output_nodes_activation = NEAT.ActivationFunction.LINEAR
return substrate
