#numActiveColumnsPerInhArea = -1, # Using percentage instead
numActiveColumnsPerInhArea=1, # Only one feature active at a time
# All input activity can contribute to feature output
stimulusThreshold=0,
synPermInactiveDec=0.3,
synPermActiveInc=0.3,
synPermConnected=0.3, # Connected threshold
maxBoost=2,
seed=1956, # The seed that Grok uses
spVerbosity=1)
# Instantiate the spatial pooler test bench.
tb = VisionTestBench(sp)
# Instantiate the classifier
clf = exactMatch()
# Train the spatial pooler on trainingVectors.
numCycles = tb.train(trainingVectors, trainingTags, clf, maxTrainingCycles)
# View the permanences and connections after training.
tb.showPermsAndConns()
#tb.savePermsAndConns('perms_and_conns.jpg')
# Get testing images and convert them to vectors.
testingImages, testingTags = data.getImagesAndTags(testingDataset)
testingVectors = encoder.imagesToVectors(testingImages)
# Test the spatial pooler on testingVectors.
accuracy = tb.test(testingVectors, testingTags, clf, verbose=1)
#numActiveColumnsPerInhArea = -1, # Using percentage instead
numActiveColumnsPerInhArea = 1, # Only one feature active at a time
# All input activity can contribute to feature output
stimulusThreshold = 0,
synPermInactiveDec = 0.3,
synPermActiveInc = 0.3,
synPermConnected = 0.3, # Connected threshold
maxBoost = 2,
seed = 1956, # The seed that Grok uses
spVerbosity = 1)
# Instantiate the spatial pooler test bench.
tb = VisionTestBench(sp)
# Instantiate the classifier
clf = exactMatch()
# Train the spatial pooler on trainingVectors.
numCycles = tb.train(trainingVectors, trainingTags, clf, maxTrainingCycles)
# View the permanences and connections after training.
tb.showPermsAndConns()
#tb.savePermsAndConns('perms_and_conns.jpg')
# Get testing images and convert them to vectors.
testingImages, testingTags = data.getImagesAndTags(testingDataset)
testingVectors = encoder.imagesToVectors(testingImages)
# Test the spatial pooler on testingVectors.
accuracy = tb.test(testingVectors, testingTags, clf, verbose=1)
clf = KNNClassifier()
# Train the spatial pooler on trainingVectors.
numCycles = tb.train(trainingVectors, trainingTags, clf, maxTrainingCycles,
minAccuracy)
# Save the permanences and connections after training.
#tb.savePermanences('perms.jpg')
#tb.showPermanences()
#tb.showConnections()
# Get testing images and convert them to vectors.
testingImages, testingTags = data.getImagesAndTags(testingDataset)
testingVectors = encoder.imagesToVectors(testingImages)
# Reverse the order of the vectors and tags for testing
testingTags = [testingTag for testingTag in reversed(testingTags)]
testingVectors = [
testingVector for testingVector in reversed(testingVectors)
]
# Test the spatial pooler on testingVectors.
accuracy = tb.test(testingVectors, testingTags, clf)
# Add results to the list
parameters.appendResults([accuracy, numCycles])
parameters.printResults(["Percent Accuracy", "Training Cycles"],
[", %.2f", ", %d"])
print "The maximum number of training cycles is set to:", maxTrainingCycles
# Instantiate the classifier
clf = KNNClassifier()
# Get training images and convert them to vectors.
trainingImages, trainingTags = data.getImagesAndTags(trainingDataset)
trainingVectors = encoder.imagesToVectors(trainingImages)
# Train the spatial pooler on trainingVectors.
numCycles = tb.train(trainingVectors, trainingTags, clf, maxTrainingCycles,
minAccuracy)
# Save the permanences and connections after training.
#tb.savePermanences('perms.jpg')
#tb.showPermanences()
#tb.showConnections()
# Get testing images and convert them to vectors.
testingImages, testingTags = data.getImagesAndTags(testingDataset)
testingVectors = encoder.imagesToVectors(testingImages)
# Reverse the order of the vectors and tags for testing
testingTags = [testingTag for testingTag in reversed(testingTags)]
testingVectors = [testingVector for testingVector in reversed(testingVectors)]
# Test the spatial pooler on testingVectors.
accuracy = tb.test(testingVectors, testingTags, clf, learn=True)
print "Number of training cycles:", numCycles
# Instantiate the classifier
clf = KNNClassifier()
# Train the spatial pooler on trainingVectors.
numCycles = tb.train(trainingVectors, trainingTags, clf, maxTrainingCycles,
minAccuracy)
# Save the permanences and connections after training.
#tb.savePermanences('perms.jpg')
#tb.showPermanences()
#tb.showConnections()
# Get testing images and convert them to vectors.
testingImages, testingTags = data.getImagesAndTags(testingDataset)
testingVectors = encoder.imagesToVectors(testingImages)
# Reverse the order of the vectors and tags for testing
testingTags = [testingTag for testingTag in reversed(testingTags)]
testingVectors = [testingVector for testingVector in reversed(testingVectors)]
# Test the spatial pooler on testingVectors.
accuracy = tb.test(testingVectors, testingTags, clf)
# Add results to the list
parameters.appendResults([accuracy, numCycles])
parameters.printResults(["Percent Accuracy", "Training Cycles"], [", %.2f", ", %d"])
print "The maximum number of training cycles is set to:", maxTrainingCycles
tb = VisionTestBench(sp)
# Train the spatial pooler on trainingVectors.
trainSDRIs, numCycles = tb.train(trainingVectors, trainingTags,
maxTrainingCycles, usePPM=False)
# Save the permanences and connections after training.
tb.savePermsAndConns('perms_and_conns.jpg')
#tb.showPermsAndConns()
# Get testing images and convert them to vectors.
testingImages, testingTags = data.getImagesAndTags(testingDataset)
testingVectors = encoder.imagesToVectors(testingImages)
# Test the spatial pooler on testingVectors.
testSDRIs = tb.test(testingVectors, testingTags)
if testSDRIs != trainSDRIs:
print "Yo! SDRs don't match!"
#for i in range(len(testSDRIs)):
#if testSDRIs[i] != trainSDRIs[i]:
#print "%6s %6s %6s" % (i, trainSDRIs[i], testSDRIs[i])
#tb.printSDR(trainSDRIs[i])
#print
#tb.printSDR(testSDRIs[i])
#junk = raw_input()
# Classifier Hack, uses the testing image tags along with the SDRs from the
# last training cycle to interpret the SDRs from testing.
testResults = []
[testResults.append('') for i in range(len(testSDRIs))]
