if not os.path.exists(resultsFn):
cleanCRPfolder()
cleanNCDfolder()
startDateTime = datetime.now()
# load weights etc. if this is for a neural net run
if settingsDict['NN Type'] is not None:
weightMatrix, biases, featureOffset, featureScaling = get_NN_NCD_params(
NNtype=settingsDict['NN Type'],
featureName=settingsDict['Feature Name'],
learningRate=settingsDict['dA Learning Rate'],
learningRateBoostFactor=settingsDict[
'dA Learning Rate Boost Factor'],
corruptionLevel=settingsDict['dA Corruption Level'],
numVisible=int(settingsDict['dA Num Visible Units']),
numHidden=int(settingsDict['dA Num Hidden Units']),
batchSize=int(settingsDict['dA Batch Size']),
freqStd=bool(settingsDict['NN Frequency Standardisation']),
NNnumFolders=int(settingsDict['NN Num Folders']),
NNnumFilesPerFolder=int(
settingsDict['NN Num Files per Folder']),
NNtimeStacking=int(settingsDict['NN Time Stacking']))
else:
weightMatrix = biases = featureOffset = featureScaling = None
# Create NCD files
for key in settingsDict.keys():
print key, ':', settingsDict[key]
createNCDfiles(
existingNCDs=None,
# converting to FENS
# Get the folders (performances)
piecesPath = FFP.getRootPath(featureName)
piecesFolders = getFolderNames(piecesPath,
contains = 'mazurka',
orderAlphabetically = True) # added the contains parameter to avoid the new powerspectrum folder
if numFolders is not None:
piecesFolders = piecesFolders[: numFolders]
# Load weights and biases
if NNtype is not None:
weightMatrix, biases, featureOffset, featureScaling = get_NN_NCD_params(
NNtype, featureName, learningRate, learningRateBoostFactor,
corruptionLevel, numOriginalFeatures, numNewFeatures, batchSize,
freqStd = frequencyStandardisation, NNnumFolders = numFolders,
NNnumFilesPerFolder = numFilesPerFolder,
NNtimeStacking = timeStacking)
# Load (and optionally transform) the feature files
p = 0
featuresDataFrames = []
for piecesFolder in piecesFolders:
performancesPath = FFP.getFeatureFolderPath(piecesPath + piecesFolder + '/', featureName)
performances = getFileNames(performancesPath,
orderAlphabetically = True,
endsWith = '.csv')
if numFilesPerFolder is not None:
performances = performances[: numFilesPerFolder]
for performance in performances:
p+= 1
print '\rloading feature file %i...' % p,
while currentDateTime < stopRunningAt:
nextSettings = True
iteration += 1
while nextSettings is not None and currentDateTime < stopRunningAt:
nextSettings = opt.getNextSettings()
if nextSettings is not None:
for setting in nextSettings:
# load weights etc. if this is for a neural net run
if NNtype is not None:
weightMatrix, biases, featureOffset, featureScaling = get_NN_NCD_params(
NNtype, featureName, learningRate,
learningRateBoostFactor,
setting['dA Corruption Level'],
setting['dA Num Visible Units'],
setting['dA Num Hidden Units'], batchSize, freqStd,
numFolders, numFilesPerFolder, timeStacking)
else:
weightMatrix = biases = featureOffset = featureScaling = None
# Calculate NCDs
for key in setting.keys():
print key, ':', setting[key]
NCDlist = calculateNCDs(processPool, featureName, numFeatures,
setting['DownSample Factor'],
setting['Time Delay'],
setting['Dimension'], CRPmethod,
setting['Neighbourhood Size'],
numFolders, numFilesPerFolder,