Try writing it to an xml file, reread it and determine if it looks the same:
>>> from pybrain.tests import xmlInvariance
>>> xmlInvariance(n)
Same representation
Same function
Same class
"""
__author__ = 'Tom Schaul, [email protected]'
from scipy import ones
from pybrain.structure.networks import NeuronDecomposableNetwork
from pybrain.tools.shortcuts import buildNetwork
from pybrain.tests import runModuleTestSuite
def buildDecomposableNetwork():
""" three hidden neurons, with 2 in- and 2 outconnections each. """
n = buildNetwork(2, 3, 2, bias = False)
ndc = NeuronDecomposableNetwork.convertNormalNetwork(n)
# set all the weights to 1
ndc._setParameters(ones(12))
return ndc
if __name__ == "__main__":
runModuleTestSuite(__import__('__main__'))
>>> t = BackpropTrainer(n, learningrate = 0.01, momentum = 0.99, verbose = True)
>>> t.trainOnDataset(ds, 4)
Total error: 2.44696473875
Total error: 1.97570498879
Total error: 1.23940309483
Total error: 0.546129967878
>>> abs(n.params[10:15] - array([ -0.53868206, -0.54185834, 0.26726394, -1.90008234, -1.12114946])).round(5)
array([ 0., 0., 0., 0., 0.])
Now the same for RPROP
>>> t = RPropMinusTrainer(n, verbose = True)
>>> t.trainOnDataset(ds, 4)
epoch 0 total error 0.16818 avg weight 0.92638
epoch 1 total error 0.15007 avg weight 0.92202
epoch 2 total error 0.15572 avg weight 0.92684
epoch 3 total error 0.13036 avg weight 0.92604
>>> abs(n.params[5:10] - array([ -0.19241111, 1.43404022, 0.23062397, -0.40105413, 0.62100109])).round(5)
array([ 0., 0., 0., 0., 0.])
"""
__author__ = "Martin Felder, [email protected]"
from pybrain.tests import runModuleTestSuite
if __name__ == "__main__":
runModuleTestSuite(__import__("__main__"))
def buildSimpleMDLSTMNetwork(peepholes=False):
N = RecurrentNetwork('simpleMDLstmNet')
i = LinearLayer(1, name='i')
dim = 1
h = MDLSTMLayer(dim, peepholes=peepholes, name='MDlstm')
o = LinearLayer(1, name='o')
b = BiasUnit('bias')
N.addModule(b)
N.addOutputModule(o)
N.addInputModule(i)
N.addModule(h)
N.addConnection(FullConnection(i, h, outSliceTo=4 * dim, name='f1'))
N.addConnection(FullConnection(b, h, outSliceTo=4 * dim, name='f2'))
N.addRecurrentConnection(
FullConnection(h, h, inSliceTo=dim, outSliceTo=4 * dim, name='r1'))
N.addRecurrentConnection(
IdentityConnection(h,
h,
inSliceFrom=dim,
outSliceFrom=4 * dim,
name='rstate'))
N.addConnection(FullConnection(h, o, inSliceTo=dim, name='f3'))
N.sortModules()
return N
if __name__ == "__main__":
runModuleTestSuite(__import__('__main__'))
self.w[:] = self.f
if fullUpdate and self.hasLearnableConn():
for c in self.conns:
c.updateLearnParams()
def _forwardImplementation(self, inbuf, outbuf):
extin = inbuf[:self.dim]
# update frequencies
if self.hasLearnableConn():
z, fr, conns = zfcrk4(self.t, self.dt, self, extin)
for i in range(len(self.conns)):
# self.conns[i].c[:] = limitC(np.reshape(conns[i], self.conns[i].c.shape), self.conns[i].roote)
self.conns[i].c[:] = conns[i]
else:
z, fr = zfrk4(self.t, self.dt, self, extin)
self.fr[:] = np.minimum(np.maximum(fr, self.fr_min), self.fr_max)
self.f[:] = self.fr / TWO_PI
self.updateOscParams(self.offset % self.fupdate == 0)
self.t += self.dt
outbuf[:] = z
if __name__ == "__main__":
from pybrain.tests import runModuleTestSuite
import pygfnn.tests.unittests.structure.modules.test_simple_gfnn_network as test
runModuleTestSuite(test)
