def basicTest(self):
"""Basic test (creation, pickling, basic run of learning and inference)"""
# Create TP object
tp = TP10X2(numberOfCols=10,
cellsPerColumn=3,
initialPerm=.2,
connectedPerm=0.8,
minThreshold=2,
newSynapseCount=5,
permanenceInc=.1,
permanenceDec=.05,
permanenceMax=1,
globalDecay=.05,
activationThreshold=4,
doPooling=False,
segUpdateValidDuration=5,
seed=SEED,
verbosity=VERBOSITY)
tp.retrieveLearningStates = True
# Save and reload
tp.makeCells4Ephemeral = False
pickle.dump(tp, open("test_tp10x.pkl", "wb"))
tp2 = pickle.load(open("test_tp10x.pkl"))
self.assertTrue(fdrutils.tpDiff2(tp, tp2, VERBOSITY,
checkStates=False))
# Learn
for i in xrange(5):
x = numpy.zeros(tp.numberOfCols, dtype='uint32')
_RGEN.initializeUInt32Array(x, 2)
tp.learn(x)
# Save and reload after learning
tp.reset()
tp.makeCells4Ephemeral = False
pickle.dump(tp, open("test_tp10x.pkl", "wb"))
tp2 = pickle.load(open("test_tp10x.pkl"))
self.assertTrue(fdrutils.tpDiff2(tp, tp2, VERBOSITY))
## Infer
patterns = numpy.zeros((4, tp.numberOfCols), dtype='uint32')
for i in xrange(4):
_RGEN.initializeUInt32Array(patterns[i], 2)
for i in xrange(10):
x = numpy.zeros(tp.numberOfCols, dtype='uint32')
_RGEN.initializeUInt32Array(x, 2)
tp.infer(x)
if i > 0:
tp.checkPrediction2(patterns)
def _createTPs(numCols, cellsPerColumn=4, checkSynapseConsistency=True):
"""Create TP and TP10X instances with identical parameters. """
# Keep these fixed for both TP's:
minThreshold = 4
activationThreshold = 4
newSynapseCount = 5
initialPerm = 0.6
connectedPerm = 0.5
permanenceInc = 0.1
permanenceDec = 0.001
globalDecay = 0.0
if VERBOSITY > 1:
print "Creating TP10X instance"
cppTp = TP10X2(numberOfCols=numCols,
cellsPerColumn=cellsPerColumn,
initialPerm=initialPerm,
connectedPerm=connectedPerm,
minThreshold=minThreshold,
newSynapseCount=newSynapseCount,
permanenceInc=permanenceInc,
permanenceDec=permanenceDec,
activationThreshold=activationThreshold,
globalDecay=globalDecay,
burnIn=1,
seed=SEED,
verbosity=VERBOSITY,
checkSynapseConsistency=checkSynapseConsistency,
pamLength=1000)
if VERBOSITY > 1:
print "Creating PY TP instance"
pyTp = TP(numberOfCols=numCols,
cellsPerColumn=cellsPerColumn,
initialPerm=initialPerm,
connectedPerm=connectedPerm,
minThreshold=minThreshold,
newSynapseCount=newSynapseCount,
permanenceInc=permanenceInc,
permanenceDec=permanenceDec,
activationThreshold=activationThreshold,
globalDecay=globalDecay,
burnIn=1,
seed=SEED,
verbosity=VERBOSITY,
pamLength=1000)
return cppTp, pyTp
def setUp(self):
self.tmPy = TemporalMemoryPy(columnDimensions=[2048],
cellsPerColumn=32,
initialPermanence=0.5,
connectedPermanence=0.8,
minThreshold=10,
maxNewSynapseCount=12,
permanenceIncrement=0.1,
permanenceDecrement=0.05,
activationThreshold=15)
self.tmCPP = TemporalMemoryCPP(columnDimensions=[2048],
cellsPerColumn=32,
initialPermanence=0.5,
connectedPermanence=0.8,
minThreshold=10,
maxNewSynapseCount=12,
permanenceIncrement=0.1,
permanenceDecrement=0.05,
activationThreshold=15)
self.tp = TP(numberOfCols=2048,
cellsPerColumn=32,
initialPerm=0.5,
connectedPerm=0.8,
minThreshold=10,
newSynapseCount=12,
permanenceInc=0.1,
permanenceDec=0.05,
activationThreshold=15,
globalDecay=0,
burnIn=1,
checkSynapseConsistency=False,
pamLength=1)
self.tp10x2 = TP10X2(numberOfCols=2048,
cellsPerColumn=32,
initialPerm=0.5,
connectedPerm=0.8,
minThreshold=10,
newSynapseCount=12,
permanenceInc=0.1,
permanenceDec=0.05,
activationThreshold=15,
globalDecay=0,
burnIn=1,
checkSynapseConsistency=False,
pamLength=1)
self.patternMachine = PatternMachine(2048, 40, num=100)
self.sequenceMachine = SequenceMachine(self.patternMachine)
def _createTps(numCols):
"""Create two instances of temporal poolers (TP.py and TP10X2.py) with
identical parameter settings."""
# Keep these fixed:
minThreshold = 4
activationThreshold = 5
newSynapseCount = 7
initialPerm = 0.3
connectedPerm = 0.5
permanenceInc = 0.1
permanenceDec = 0.05
globalDecay = 0
cellsPerColumn = 1
cppTp = TP10X2(numberOfCols=numCols,
cellsPerColumn=cellsPerColumn,
initialPerm=initialPerm,
connectedPerm=connectedPerm,
minThreshold=minThreshold,
newSynapseCount=newSynapseCount,
permanenceInc=permanenceInc,
permanenceDec=permanenceDec,
activationThreshold=activationThreshold,
globalDecay=globalDecay,
burnIn=1,
seed=SEED,
verbosity=VERBOSITY,
checkSynapseConsistency=True,
pamLength=1000)
# Ensure we are copying over learning states for TPDiff
cppTp.retrieveLearningStates = True
pyTp = TP(numberOfCols=numCols,
cellsPerColumn=cellsPerColumn,
initialPerm=initialPerm,
connectedPerm=connectedPerm,
minThreshold=minThreshold,
newSynapseCount=newSynapseCount,
permanenceInc=permanenceInc,
permanenceDec=permanenceDec,
activationThreshold=activationThreshold,
globalDecay=globalDecay,
burnIn=1,
seed=SEED,
verbosity=VERBOSITY,
pamLength=1000)
return cppTp, pyTp
def setUp(self):
self.tmPy = TemporalMemoryPy(columnDimensions=[2048],
cellsPerColumn=32,
initialPermanence=0.5,
connectedPermanence=0.8,
minThreshold=10,
maxNewSynapseCount=12,
permanenceIncrement=0.1,
permanenceDecrement=0.05,
activationThreshold=15)
self.tmCPP = TemporalMemoryCPP(columnDimensions=[2048],
cellsPerColumn=32,
initialPermanence=0.5,
connectedPermanence=0.8,
minThreshold=10,
maxNewSynapseCount=12,
permanenceIncrement=0.1,
permanenceDecrement=0.05,
activationThreshold=15)
self.tp = TP(numberOfCols=2048,
cellsPerColumn=32,
initialPerm=0.5,
connectedPerm=0.8,
minThreshold=10,
newSynapseCount=12,
permanenceInc=0.1,
permanenceDec=0.05,
activationThreshold=15,
globalDecay=0,
burnIn=1,
checkSynapseConsistency=False,
pamLength=1)
self.tp10x2 = TP10X2(numberOfCols=2048,
cellsPerColumn=32,
initialPerm=0.5,
connectedPerm=0.8,
minThreshold=10,
newSynapseCount=12,
permanenceInc=0.1,
permanenceDec=0.05,
activationThreshold=15,
globalDecay=0,
burnIn=1,
checkSynapseConsistency=False,
pamLength=1)
self.scalarEncoder = RandomDistributedScalarEncoder(0.88)
s += ' '
s += str(x[c])
s += ' '
return s
#######################################################################
#
# Step 1: create Temporal Pooler instance with appropriate parameters
tp = TP10X2(numberOfCols=50,
cellsPerColumn=1,
initialPerm=0.5,
connectedPerm=0.5,
minThreshold=10,
newSynapseCount=10,
permanenceInc=0.1,
permanenceDec=0.0,
activationThreshold=8,
globalDecay=0,
burnIn=1,
checkSynapseConsistency=False,
pamLength=10)
#######################################################################
#
# Step 2: create input vectors to feed to the temporal pooler. Each input vector
# must be numberOfCols wide. Here we create a simple sequence of 5 vectors
# representing the sequence A -> B -> C -> D -> E
x = numpy.zeros((5, tp.numberOfCols), dtype="uint32")
x[0, 0:10] = 1 # Input SDR representing "A"
x[1, 10:20] = 1 # Input SDR representing "B"
def testTPs(self, short=True):
"""Call basicTest2 with multiple parameter settings and ensure the C++ and
PY versions are identical throughout."""
if short == True:
print "Testing short version"
else:
print "Testing long version"
if short:
print "\nTesting with fixed resource CLA - test max segment and synapses"
tp = TP10X2(numberOfCols=30,
cellsPerColumn=5,
initialPerm=.5,
connectedPerm=0.5,
permanenceMax=1,
minThreshold=8,
newSynapseCount=10,
permanenceInc=0.1,
permanenceDec=0.01,
globalDecay=.0,
activationThreshold=8,
doPooling=False,
segUpdateValidDuration=5,
seed=SEED,
verbosity=VERBOSITY,
maxAge=0,
maxSegmentsPerCell=2,
maxSynapsesPerSegment=10,
checkSynapseConsistency=True)
tp.cells4.setCellSegmentOrder(True)
self.basicTest2(tp, numPatterns=15, numRepetitions=1)
if not short:
print "\nTesting with fixed resource CLA - test max segment and synapses"
tp = TP10X2(numberOfCols=30,
cellsPerColumn=5,
initialPerm=.5,
connectedPerm=0.5,
permanenceMax=1,
minThreshold=8,
newSynapseCount=10,
permanenceInc=.1,
permanenceDec=.01,
globalDecay=.0,
activationThreshold=8,
doPooling=False,
segUpdateValidDuration=5,
seed=SEED,
verbosity=VERBOSITY,
maxAge=0,
maxSegmentsPerCell=2,
maxSynapsesPerSegment=10,
checkSynapseConsistency=True)
tp.cells4.setCellSegmentOrder(1)
self.basicTest2(tp, numPatterns=30, numRepetitions=2)
print "\nTesting with permanenceInc = 0 and Dec = 0"
tp = TP10X2(numberOfCols=30,
cellsPerColumn=5,
initialPerm=.5,
connectedPerm=0.5,
minThreshold=3,
newSynapseCount=3,
permanenceInc=0.0,
permanenceDec=0.00,
permanenceMax=1,
globalDecay=.0,
activationThreshold=3,
doPooling=False,
segUpdateValidDuration=5,
seed=SEED,
verbosity=VERBOSITY,
checkSynapseConsistency=False)
tp.printParameters()
self.basicTest2(tp, numPatterns=30, numRepetitions=3)
print "Testing with permanenceInc = 0 and Dec = 0 and 1 cell per column"
tp = TP10X2(numberOfCols=30,
cellsPerColumn=1,
initialPerm=.5,
connectedPerm=0.5,
minThreshold=3,
newSynapseCount=3,
permanenceInc=0.0,
permanenceDec=0.0,
permanenceMax=1,
globalDecay=.0,
activationThreshold=3,
doPooling=False,
segUpdateValidDuration=5,
seed=SEED,
verbosity=VERBOSITY,
checkSynapseConsistency=False)
self.basicTest2(tp)
print "Testing with permanenceInc = 0.1 and Dec = .0"
tp = TP10X2(numberOfCols=30,
cellsPerColumn=5,
initialPerm=.5,
connectedPerm=0.5,
minThreshold=3,
newSynapseCount=3,
permanenceInc=.1,
permanenceDec=.0,
permanenceMax=1,
globalDecay=.0,
activationThreshold=3,
doPooling=False,
segUpdateValidDuration=5,
seed=SEED,
verbosity=VERBOSITY,
checkSynapseConsistency=False)
self.basicTest2(tp)
print(
"Testing with permanenceInc = 0.1, Dec = .01 and higher synapse "
"count")
tp = TP10X2(numberOfCols=30,
cellsPerColumn=2,
initialPerm=.5,
connectedPerm=0.5,
minThreshold=3,
newSynapseCount=5,
permanenceInc=.1,
permanenceDec=.01,
permanenceMax=1,
globalDecay=.0,
activationThreshold=3,
doPooling=False,
segUpdateValidDuration=5,
seed=SEED,
verbosity=VERBOSITY,
checkSynapseConsistency=True)
self.basicTest2(tp, numPatterns=10, numRepetitions=2)
print "Testing age based global decay"
tp = TP10X2(numberOfCols=30,
cellsPerColumn=5,
initialPerm=.4,
connectedPerm=0.5,
minThreshold=3,
newSynapseCount=3,
permanenceInc=0.1,
permanenceDec=0.1,
permanenceMax=1,
globalDecay=.25,
activationThreshold=3,
doPooling=False,
segUpdateValidDuration=5,
pamLength=2,
maxAge=20,
seed=SEED,
verbosity=VERBOSITY,
checkSynapseConsistency=True)
tp.cells4.setCellSegmentOrder(1)
self.basicTest2(tp)
print "\nTesting with fixed size CLA, max segments per cell"
tp = TP10X2(numberOfCols=30,
cellsPerColumn=5,
initialPerm=.5,
connectedPerm=0.5,
permanenceMax=1,
minThreshold=8,
newSynapseCount=10,
permanenceInc=.1,
permanenceDec=.01,
globalDecay=.0,
activationThreshold=8,
doPooling=False,
segUpdateValidDuration=5,
seed=SEED,
verbosity=VERBOSITY,
maxAge=0,
maxSegmentsPerCell=2,
maxSynapsesPerSegment=100,
checkSynapseConsistency=True)
tp.cells4.setCellSegmentOrder(1)
self.basicTest2(tp, numPatterns=30, numRepetitions=2)
def createTPs(includeCPP=True,
includePy=True,
numCols=100,
cellsPerCol=4,
activationThreshold=3,
minThreshold=3,
newSynapseCount=3,
initialPerm=0.6,
permanenceInc=0.1,
permanenceDec=0.0,
globalDecay=0.0,
pamLength=0,
checkSynapseConsistency=True,
maxInfBacktrack=0,
maxLrnBacktrack=0,
**kwargs):
"""Create one or more TP instances, placing each into a dict keyed by
name.
Parameters:
------------------------------------------------------------------
retval: tps - dict of TP instances
"""
# Keep these fixed:
connectedPerm = 0.5
tps = dict()
if includeCPP:
if VERBOSITY >= 2:
print "Creating TP10X2 instance"
cpp_tp = TP10X2(
numberOfCols=numCols,
cellsPerColumn=cellsPerCol,
initialPerm=initialPerm,
connectedPerm=connectedPerm,
minThreshold=minThreshold,
newSynapseCount=newSynapseCount,
permanenceInc=permanenceInc,
permanenceDec=permanenceDec,
activationThreshold=activationThreshold,
globalDecay=globalDecay,
burnIn=1,
seed=SEED,
verbosity=VERBOSITY,
checkSynapseConsistency=checkSynapseConsistency,
collectStats=True,
pamLength=pamLength,
maxInfBacktrack=maxInfBacktrack,
maxLrnBacktrack=maxLrnBacktrack,
)
# Ensure we are copying over learning states for TPDiff
cpp_tp.retrieveLearningStates = True
tps['CPP'] = cpp_tp
if includePy:
if VERBOSITY >= 2:
print "Creating PY TP instance"
py_tp = TP(
numberOfCols=numCols,
cellsPerColumn=cellsPerCol,
initialPerm=initialPerm,
connectedPerm=connectedPerm,
minThreshold=minThreshold,
newSynapseCount=newSynapseCount,
permanenceInc=permanenceInc,
permanenceDec=permanenceDec,
activationThreshold=activationThreshold,
globalDecay=globalDecay,
burnIn=1,
seed=SEED,
verbosity=VERBOSITY,
collectStats=True,
pamLength=pamLength,
maxInfBacktrack=maxInfBacktrack,
maxLrnBacktrack=maxLrnBacktrack,
)
tps['PY '] = py_tp
return tps
def _createTPs(self,
numCols,
fixedResources=False,
checkSynapseConsistency=True):
"""Create an instance of the appropriate temporal pooler. We isolate
all parameters as constants specified here."""
# Keep these fixed:
minThreshold = 4
activationThreshold = 8
newSynapseCount = 15
initialPerm = 0.3
connectedPerm = 0.5
permanenceInc = 0.1
permanenceDec = 0.05
if fixedResources:
permanenceDec = 0.1
maxSegmentsPerCell = 5
maxSynapsesPerSegment = 15
globalDecay = 0
maxAge = 0
else:
permanenceDec = 0.05
maxSegmentsPerCell = -1
maxSynapsesPerSegment = -1
globalDecay = 0.0001
maxAge = 1
if g_testCPPTP:
if g_options.verbosity > 1:
print "Creating TP10X2 instance"
cppTP = TP10X2(
numberOfCols=numCols,
cellsPerColumn=4,
initialPerm=initialPerm,
connectedPerm=connectedPerm,
minThreshold=minThreshold,
newSynapseCount=newSynapseCount,
permanenceInc=permanenceInc,
permanenceDec=permanenceDec,
activationThreshold=activationThreshold,
globalDecay=globalDecay,
maxAge=maxAge,
burnIn=1,
seed=g_options.seed,
verbosity=g_options.verbosity,
checkSynapseConsistency=checkSynapseConsistency,
pamLength=1000,
maxSegmentsPerCell=maxSegmentsPerCell,
maxSynapsesPerSegment=maxSynapsesPerSegment,
)
# Ensure we are copying over learning states for TPDiff
cppTP.retrieveLearningStates = True
else:
cppTP = None
if g_options.verbosity > 1:
print "Creating PY TP instance"
pyTP = TP(
numberOfCols=numCols,
cellsPerColumn=4,
initialPerm=initialPerm,
connectedPerm=connectedPerm,
minThreshold=minThreshold,
newSynapseCount=newSynapseCount,
permanenceInc=permanenceInc,
permanenceDec=permanenceDec,
activationThreshold=activationThreshold,
globalDecay=globalDecay,
maxAge=maxAge,
burnIn=1,
seed=g_options.seed,
verbosity=g_options.verbosity,
pamLength=1000,
maxSegmentsPerCell=maxSegmentsPerCell,
maxSynapsesPerSegment=maxSynapsesPerSegment,
)
return cppTP, pyTP
def setUp(cls):
tmPy = TemporalMemoryPy(columnDimensions=[2048],
cellsPerColumn=32,
initialPermanence=0.5,
connectedPermanence=0.8,
minThreshold=10,
maxNewSynapseCount=12,
permanenceIncrement=0.1,
permanenceDecrement=0.05,
activationThreshold=15)
tmCPP = TemporalMemoryCPP(columnDimensions=[2048],
cellsPerColumn=32,
initialPermanence=0.5,
connectedPermanence=0.8,
minThreshold=10,
maxNewSynapseCount=12,
permanenceIncrement=0.1,
permanenceDecrement=0.05,
activationThreshold=15)
tp = TP(numberOfCols=2048,
cellsPerColumn=32,
initialPerm=0.5,
connectedPerm=0.8,
minThreshold=10,
newSynapseCount=12,
permanenceInc=0.1,
permanenceDec=0.05,
activationThreshold=15,
globalDecay=0,
burnIn=1,
checkSynapseConsistency=False,
pamLength=1)
tp10x2 = TP10X2(numberOfCols=2048,
cellsPerColumn=32,
initialPerm=0.5,
connectedPerm=0.8,
minThreshold=10,
newSynapseCount=12,
permanenceInc=0.1,
permanenceDec=0.05,
activationThreshold=15,
globalDecay=0,
burnIn=1,
checkSynapseConsistency=False,
pamLength=1)
def tmComputeFn(pattern, instance):
instance.compute(pattern, True)
def tpComputeFn(pattern, instance):
array = cls._patternToNumpyArray(pattern)
instance.compute(array, enableLearn=True, computeInfOutput=True)
return (
("TM (py)", tmPy, tmComputeFn),
("TM (C++)", tmCPP, tmComputeFn),
("TP", tp, tpComputeFn),
("TP10X2", tp10x2, tpComputeFn),
)
