def iterRbms(self):
"""Yield every two layers as an rbm."""
layers = [i for i in self.net.modulesSorted if isinstance(i, NeuronLayer) and not isinstance(i, BiasUnit)]
# There will be a single bias.
bias = [i for i in self.net.modulesSorted if isinstance(i, BiasUnit)][0]
layercons = (self.net.connections[i][0] for i in layers)
# The biascons will not be sorted; we have to sort them to zip nicely
# with the corresponding layers.
biascons = self.net.connections[bias]
biascons.sort(key=lambda c: layers.index(c.outmod))
modules = list(zip(layers, layers[1:], layercons, biascons))
for visible, hidden, layercon, biascon in modules:
rbm = Rbm.fromModules(visible, hidden, bias, layercon, biascon)
yield rbm
def iterRbms(self):
"""Yield every two layers as an rbm."""
layers = [i for i in self.net.modulesSorted
if isinstance(i, NeuronLayer) and not isinstance(i, BiasUnit)]
# There will be a single bias.
bias = [i for i in self.net.modulesSorted if isinstance(i, BiasUnit)][0]
layercons = (self.net.connections[i][0] for i in layers)
# The biascons will not be sorted; we have to sort them to zip nicely
# with the corresponding layers.
biascons = self.net.connections[bias]
biascons.sort(key=lambda c: layers.index(c.outmod))
modules = zip(layers, layers[1:], layercons, biascons)
for visible, hidden, layercon, biascon in modules:
rbm = Rbm.fromModules(visible, hidden, bias,
layercon, biascon)
yield rbm
from __future__ import print_function #!/usr/bin/env python """ Miniscule restricted Boltzmann machine usage example """ __author__ = 'Justin S Bayer, [email protected]' from pybrain.structure.networks.rbm import Rbm from pybrain.unsupervised.trainers.rbm import (RbmGibbsTrainerConfig, RbmBernoulliTrainer) from pybrain.datasets import UnsupervisedDataSet ds = UnsupervisedDataSet(6) ds.addSample([0, 1] * 3) ds.addSample([1, 0] * 3) cfg = RbmGibbsTrainerConfig() cfg.maxIter = 3 rbm = Rbm.fromDims(6, 1) trainer = RbmBernoulliTrainer(rbm, ds, cfg) print(rbm.params, rbm.biasParams) for _ in range(50): trainer.train() print(rbm.params, rbm.biasParams)
def __init__(self, in_dims, out_dims): self.dataset = UnsupervisedDataSet(in_dims) cfg = RbmGibbsTrainerConfig() cfg.maxIter = 5 self.model = Rbm.fromDims(in_dims, out_dims) self.trainer = RbmBernoulliTrainer(self.model, self.dataset, cfg)
from __future__ import print_function #!/usr/bin/env python """ Miniscule restricted Boltzmann machine usage example """ __author__ = 'Justin S Bayer, [email protected]' from pybrain.structure.networks.rbm import Rbm from pybrain.unsupervised.trainers.rbm import (RbmGibbsTrainerConfig, RbmBernoulliTrainer) from pybrain.datasets import UnsupervisedDataSet ds = UnsupervisedDataSet(6) ds.addSample([0, 1] * 3) ds.addSample([1, 0] * 3) cfg = RbmGibbsTrainerConfig() cfg.maxIter = 3 rbm = Rbm.fromDims(6, 1) trainer = RbmBernoulliTrainer(rbm, ds, cfg) print(rbm.params, rbm.biasParams) for _ in range(50): trainer.train() print(rbm.params, rbm.biasParams)
# Sort modules topologically and initialize
net.sortModules()
with open('vsample.csv','rb') as f:
reader = csv.reader(f)
for row in reader:
d_input = map(float,row[1:10])
output = map(float, row[10])
n_input = d_input/numpy.linalg.norm(d_input)
ds.addSample(d_input,output)
temp_ds.addSample(d_input)
#print ds
cfg = RbmGibbsTrainerConfig()
cfg.maxIter = 3
rbm = Rbm.fromDims(9,5)
trainer = BackpropTrainer(net, dataset=ds, learningrate= 0.001, weightdecay=0.01, verbose=True)
#trainer = DeepBeliefTrainer(net, dataset=temp_ds)
#trainer = RbmBernoulliTrainer(rbm, temp_ds, cfg)
for i in range(30):
trainer.trainEpochs(30)
print 'Expected:1 [FRAUD] ', net.activate([49,2.6,0.98,4.3,1.48,10,2.5,6,67])
print 'Expected:0 [NOT FRAUD] ', net.activate([78,5,4.4,4.5,2.99,3,1.3,10,59])
print 'Expected:1 [FRAUD] ', net.activate([57,2,0.1,1.15,0.47,7,1.8,6,73])
print 'Expected:0 [NOT FRAUD] ', net.activate([65,3,11.1,1.8,0.6,4,4,4.5,90])
print 'Expected:1 [FRAUD] ', net.activate([55,2,0.23,3.2,0.55,9,1.9,5.5,60])
print 'Expected:0 [NOT FRAUD] ', net.activate([39,5,0.07,0.5,0.17,3,3.8,3,32])
print 'Expected:1 [FRAUD] ', net.activate([63,2.5,0.25,1.23,0.3,7,1.45,4.75,35])
# Sort modules topologically and initialize
net.sortModules()
with open('vsample.csv', 'rb') as f:
reader = csv.reader(f)
for row in reader:
d_input = map(float, row[1:10])
output = map(float, row[10])
n_input = d_input / numpy.linalg.norm(d_input)
ds.addSample(d_input, output)
temp_ds.addSample(d_input)
#print ds
cfg = RbmGibbsTrainerConfig()
cfg.maxIter = 3
rbm = Rbm.fromDims(9, 5)
trainer = BackpropTrainer(net,
dataset=ds,
learningrate=0.001,
weightdecay=0.01,
verbose=True)
#trainer = DeepBeliefTrainer(net, dataset=temp_ds)
#trainer = RbmBernoulliTrainer(rbm, temp_ds, cfg)
for i in range(30):
trainer.trainEpochs(30)
print 'Expected:1 [FRAUD] ', net.activate(
[49, 2.6, 0.98, 4.3, 1.48, 10, 2.5, 6, 67])
print 'Expected:0 [NOT FRAUD] ', net.activate(
[78, 5, 4.4, 4.5, 2.99, 3, 1.3, 10, 59])
print 'Expected:1 [FRAUD] ', net.activate(