def Test_multipleNetworks():
networks = []
dataGen = Data_generator.Data_gen()
dataGen.S = 200
dataGen.gen_data()
size = dataGen.size
ks = [1, 3, 5, 8, 10]
for i in range(5):
networks.append(nw.Network([size[0], size[1], ks[i]]))
for k in range(500):
for i in range(len(networks)):
networks[i].Training(data=dataGen.Data, dt=0.001, p=1)
print("value network #", str(i), ": ", networks[i].total_value)
for j in range(len(networks)):
networkName = "network-" + str(j) + "-map"
Inter.trak(networks[i], dataGen.A, networkName)
def create_objects(status):
status.Data_gen=dgen.Data_gen()
status.Data_gen.S=status.S
status.Data_gen.Comp=status.Comp
status.Data_gen.gen_data()
def add_node(g,i):
node=nd.Node()
q=qu.Quadrant(i)
p=tplane.tangent_plane()
node.objects.append(q)
q.objects.append(p)
g.add_node(i,node)
#status.objects.append(node)
#Initializes graph
g=gr.Graph()
add_node(g,0)
k=0
while k<status.dx:
add_node(g,k+1)
g.add_edges(k,[k+1])
k=k+1
k=0
status.objects=list(g.key2node.values())
node=status.objects[0]
sf.safe_update(status.nets,0,
nw.Network([status.Data_gen.size[0],
status.Data_gen.size[1],2]))
p=node_plane(node)
#Initializes particles
while k<status.n:
par=particle()
par.position.append(node)
par.velocity.append(node)
#print(status.Data_gen.size)
par.objects.append(status.nets[0])
p.particles.append(par)
p.num_particles+=1
k=k+1
#Initializes conectivity radius
attach_balls(status,status.r)
#network.addFilters()
print("Entrenando la red mutada \n")
network.Training(data=data, dt=0.001, p=100)
print("mutando la red: Eliminando Filtro \n")
network.deleteFilters()
print("Entrenando la red mutada \n")
network.Training(data=data, dt=0.001, p=1000)
x = 10
y = 10
k = 3
objects = Functions.np.full((3), (x, y, k))
network = nw.Network([x, y, k])
data = []
generateData(data, objects, 100)
Test_modifyNetwork(network, data)
"""
print('testing node 3')
Test_node_3(network)
print('testing node 2 label=c')
Test_node_2(network)
print('testing node 2 label=n')
Test_node_2(network,"n")
print('testing node 1')
#data=Op.Sample((size[0],size[1]),A,S)
#data.insert(0,x)
data = Op.SampleVer2((size[0], size[1]), A, S, "n")
imageTarget = []
imageTarget.append(x)
imageTarget.append("c")
data.insert(0, imageTarget)
print('Training Net')
#Net=Net0.Network((size[0],size[1]))
networkParameters = np.full((3), (size[0], size[1], k))
Net = network.Network(networkParameters)
l = ['C']
i = 0
while i < S:
l.append('N')
i = i + 1
Net.Training(data=data, dt=dt, p=p)
print("finish training")
np.save(Net_name + ' w-node', Net.nodes[0].objects[0].value)
print('Scaning Image')
Inter.trak(Net, A, Net_name + ' map')
"""x=np.load('data.npy')
Shap=np.shape(x)
l=['C']
i=0
def generateImageRandom(objects):
image = Functions.np.zeros((objects[0], objects[1], 3), dtype=float)
for i in range(objects[0]):
for j in range(objects[1]):
image[i, j] = [
Functions.random.randint(1, 1),
Functions.random.randint(1, 1),
Functions.random.randint(1, 1)
]
return image
x = 2
y = 2
k = 50
objects = Functions.np.full((3), (x, y, k))
network = nw.Network(objects)
data = []
generateData(data, objects, 100)
network.Training(data=data, dt=0.01, p=0.9)
#print("valor 50: ",network.Predict(data[50]))