[big]
hidden_size = 10
"""
from __future__ import division, print_function, unicode_literals
from mlizard.caches import ShelveCache
from mlizard.experiment import createExperiment
from datasets import load_iris
from sklearn.preprocessing import LabelBinarizer
from neural_nets.connections import FullConnectionWithBias, SigmoidLayer
from neural_nets.fann import FANN
import matplotlib.pyplot as plt
cache = ShelveCache("iris.shelve")
ex = createExperiment("Iris", config_string=__doc__, cache=cache)
@ex.stage
def binarize_labels(y):
lb = LabelBinarizer()
return lb.fit_transform(y), lb
@ex.stage
def create_neural_network(in_size, hidden_size, out_size, rnd, logger):
logger.info(
"Creating a NN with {} inputs, {} hidden units, and {} output units.".
format(in_size, hidden_size, out_size))
c0 = FullConnectionWithBias(in_size, hidden_size)
s0 = SigmoidLayer(hidden_size)
# contain the configuration file.
seed = 1234567
hidden_units = 10
iterations = 1000
learning_rate = 0.01
"""
from __future__ import division, print_function, unicode_literals
from mlizard.experiment import createExperiment
from neural_nets.connections import FullConnectionWithBias, SigmoidLayer
from neural_nets.fann import FANN
from datasets import load_iris
from sklearn.preprocessing import LabelBinarizer
ex = createExperiment("demo", config_string=__doc__)
@ex.stage
def binarize_labels(y):
lb = LabelBinarizer()
return lb, lb.fit_transform(y)
@ex.stage
def build_nn(hidden_units):
l0 = FullConnectionWithBias(4, hidden_units)
s0 = SigmoidLayer(hidden_units)
l1 = FullConnectionWithBias(hidden_units, 3)
s1 = SigmoidLayer(3)
return FANN([l0, s0, l1, s1])
# contain the configuration file.
seed = 1234567
hidden_units = 10
iterations = 1000
learning_rate = 0.01
"""
from __future__ import division, print_function, unicode_literals
from mlizard.experiment import createExperiment
from neural_nets.connections import FullConnectionWithBias, SigmoidLayer
from neural_nets.fann import FANN
from datasets import load_iris
from sklearn.preprocessing import LabelBinarizer
ex = createExperiment("demo", config_string=__doc__)
@ex.stage
def binarize_labels(y):
lb = LabelBinarizer()
return lb, lb.fit_transform(y)
@ex.stage
def build_nn(hidden_units):
l0 = FullConnectionWithBias(4, hidden_units)
s0 = SigmoidLayer(hidden_units)
l1 = FullConnectionWithBias(hidden_units, 3)
s1 = SigmoidLayer(3)
return FANN([l0, s0, l1, s1])
@ex.stage
[big]
hidden_size = 10
"""
from __future__ import division, print_function, unicode_literals
from mlizard.caches import ShelveCache
from mlizard.experiment import createExperiment
from datasets import load_iris
from sklearn.preprocessing import LabelBinarizer
from neural_nets.connections import FullConnectionWithBias, SigmoidLayer
from neural_nets.fann import FANN
import matplotlib.pyplot as plt
cache = ShelveCache("iris.shelve")
ex = createExperiment("Iris", config_string=__doc__, cache=cache)
@ex.stage
def binarize_labels(y):
lb = LabelBinarizer()
return lb.fit_transform(y), lb
@ex.stage
def create_neural_network(in_size, hidden_size, out_size, rnd, logger):
logger.info("Creating a NN with {} inputs, {} hidden units, and {} output units.".format(in_size, hidden_size, out_size))
c0 = FullConnectionWithBias(in_size, hidden_size)
s0 = SigmoidLayer(hidden_size)
c1 = FullConnectionWithBias(hidden_size, out_size)
s1 = SigmoidLayer(out_size)
nn = FANN([c0, s0, c1, s1])
theta = rnd.randn(nn.get_param_dim())
