def __load_data(self):
[
self.trainData, self.trainDataSymbols, self.trainTargets,
self.trainTargetsSymbols, self.trainTargetClasses, self.testData,
self.testDataSymbols, self.testTargets, self.testTargetsSymbols,
self.validationData, self.validationDataSymbols,
self.validationTargets, self.validationTargetsSymbols,
self.validationTargetClasses
] = data.loadDataSet()
self._data_loaded = True
from framework.models import Sequential
from framework.activations import ReLU, Sigmoid
from framework.layers import Linear
from framework.losses import MSE
from framework.optimizers import SGD
from framework import FloatTensor as Tensor
import matplotlib.pyplot as plt
import utilites
from data import loadDataSet
if __name__ == "__main__":
plt.ion()
simple = False
data = loadDataSet(one_hot=False)
# Create simple linear calssifier model
lin1 = Linear(2, 10)
relu1 = ReLU()
lin2 = Linear(10, 10)
relu2 = ReLU()
lin3 = Linear(10, 1)
sig1 = Sigmoid()
model = Sequential([lin1, relu1, lin2, relu2, lin3, sig1])
# Training model
epochs_per_step = 1
for e in range(0, 100, epochs_per_step):
# Testing model on training data
pred_labels = model.predict(x=Tensor(data.test.data))
def getDontCare(size=1568):
ret = []
for i in range(size):
ret.append(0.5)
return ret
# Load Datasets
[
trainData, trainDataSymbols, trainTargets, trainTargetsSymbols,
trainTargetClasses, testData, testDataSymbols, testTargets,
testTargetsSymbols, validationData, validationDataSymbols,
validationTargets, validationTargetsSymbols, validationTargetClasses
] = data.loadDataSet()
trainDataCombined = []
trainTargetsCombined = []
testDataCombined = []
validationDataCombined = []
validationTargetsCombined = []
for index, item in enumerate(trainData):
trainDataCombined.append(
np.concatenate((item, trainDataSymbols[index]), axis=0))
for i in range(70):
for index, item in enumerate(trainData):
trainDataCombined.append(
np.concatenate((getDontCare(), trainDataSymbols[index]), axis=0))
for index, item in enumerate(trainTargets):
for i, indv in enumerate(parents):
if random.random() > pc:
while True:
tmp = copy.deepcopy(indv)
chromosome = exchange_chromosome(nodes, tmp.chromosome)
if nodes.is_feasible(chromosome):
break
offspring = Individual(i + 1, chromosome)
else:
offspring = Individual(i + 1, indv.chromosome)
offsprings.append(offspring)
return offsprings
if __name__ == '__main__':
nodes = loadDataSet('data5.xml')
chromosome1 = [{
1: 1,
2: 4,
3: 7,
4: 10,
5: 6,
6: 12,
7: 5,
8: 11
}, {
1: 2,
2: 8,
3: 3,
4: 9
}]
from framework.models import Sequential
from framework.activations import ReLU, Sigmoid
from framework.layers import Linear
from framework.losses import MSE
from framework.optimizers import Adam # SGD
import utilites
from data import loadDataSet
if __name__ == "__main__":
data = loadDataSet(one_hot=True)
# Create simple linear calssifier model
lin1 = Linear(2, 25, std=0.1)
relu1 = ReLU()
lin2 = Linear(25, 25, std=0.1)
relu2 = ReLU()
lin3 = Linear(25, 25, std=0.1)
relu3 = ReLU()
lin4 = Linear(25, 2, std=0.1)
sig1 = Sigmoid()
model = Sequential([lin1, relu1, lin2, relu2, lin3, relu3, lin4, sig1])
loss = MSE()
optimizer = Adam()
# Training model
model.train(x=data.train.data,
y=data.train.labels,
batch_size=8,
optimizer=optimizer,
loss=loss,