def download_TransformedFeature_histograms(
featureName, NNtype, numVisible, numHidden, numFolders,
numFilesPerFolder, batchSize, learningRate, learningRateBoostFactor,
corruptionLevel, timeStacking, frequencyStandardisation):
# Load weights, biases, feature offset and feature scaling
weights, biases = get_NN_WeightsAndBiases(NNtype, featureName, batchSize,
learningRate,
learningRateBoostFactor,
corruptionLevel, timeStacking,
frequencyStandardisation,
numVisible, numHidden)
featureOffset, featureScaling = get_NN_featureOffsetAndScaling(
featureName, numFolders, numFilesPerFolder, timeStacking, numVisible,
frequencyStandardisation)
# Get histogram of transformed features
hist = transformedFeatureFrequenciesHistogram(featureName,
weights,
biases,
featureOffset,
featureScaling,
timeStacking,
numFolders,
numFilesPerFolder,
numBins=100)
# Save histogram
resultsPath = transferPath + 'feature histograms/'
createPath(resultsPath)
pickle.dump(hist, open(resultsPath + featureName +
'_crpt_%i_nin_%i_nhdn_%i_basz_%i_lnrt_%0.2f' \
% (corruptionLevel, numVisible, numHidden, batchSize, learningRate) +
'_transformed_features_frequency_histograms.pkl', 'wb'))
def getWeightsAndBiases(self):
if self.weights is None:
self.weights, self.biases = get_NN_WeightsAndBiases(
self._NNtype, self._featureName, self._batchSize,
self._learningRate, self._learningRateBoostFactor,
self._corruptionLevel, self._timeStacking,
self._frequencyStandardisation, self._numVisible,
self._numHidden)
return self.weights, self.biases
'CRP Time Delay': settings[12],
'NCD Sequence Length': settings[13]}
# Load feature file dict if required
if len(featureNames) != 1:
featureFileDict, pieceIds = loadFeatureFileDict(featureNames[0], numFolders, numFilesPerFolder)
# load weights and biases if this is for a neural net run
weightMatrix = None
biases = None
featureOffset = None
featureScaling = None
if settingsDict['NN Type'] is not None:
weightMatrix, biases = get_NN_WeightsAndBiases(settingsDict['NN Type'], settingsDict['Feature Name'],
numFolders, numFilesPerFolder,
settingsDict['NN Batch Size'],
settingsDict['NN Learning Rate'], settingsDict['NN Learning Rate Boost Factor'],
settingsDict['NN Corruption Level'], settingsDict['NN Time Stacking'], frequencyStandardisation,
numFeatures, settingsDict['NN # Hidden Units'])
featureOffset, featureScaling = get_NN_featureOffsetAndScaling(settingsDict['Feature Name'], numFolders, numFilesPerFolder,
settingsDict['NN Time Stacking'],
numFeatures, frequencyStandardisation)
print 'Generating transformed features...'
transformedFeatureFileDict = FFP.generateTransformedFeatureFileDict(
featureFileDict, numFeatures, settingsDict['Feature Name'],
settingsDict['NN Type'], numFeatures, settingsDict['NN # Hidden Units'],
weightMatrix, biases, featureOffset, featureScaling, settingsDict['NN Time Stacking'],
processPool = processPool)
else:
try:
featureOffset = FFP.featureOffset[settingsDict['Feature Name']]
featureScaling = FFP.featureScaling[settingsDict['Feature Name']]