def gen_regression():
base_path = g_base_data_path + '{}emodb_regression_train'.format(
g_base_dir_delimiter)
aT.featureAndTrainRegression(base_path, 1, 1, aT.shortTermWindow,
aT.shortTermStep, "svm",
g_regression_model_path, False)
def trainRegressionWrapper(method, beat_feats, dirName, model_name):
aT.featureAndTrainRegression(dirName,
1,
1,
aT.shortTermWindow,
aT.shortTermStep,
method.lower(),
model_name,
compute_beat=beat_feats)
def training_regression(request):
try:
print('--- Start handling ...')
aT.featureAndTrainRegression(
"/home/dinhkhoi1/data/songs/train/", 1, 1, aT.shortTermWindow,
aT.shortTermStep, "svm",
"/home/dinhkhoi1/data/songs/train/svm_model", False)
return Response(status=status.HTTP_201_CREATED)
except Exception as ex:
sys.stderr(ex)
return Response(ex.__cause__, status=status.HTTP_400_BAD_REQUEST)
def train(self, audio_files_path, output_path, model_name):
afp = Path(audio_files_path)
op = Path(output_path)
model = str(op / model_name)
if not os.path.exists(afp):
print('The audio files path does not exist')
return
if not os.path.exists(op):
os.mkdir(op)
return
return aT.featureAndTrainRegression(audio_files_path, 1.0, 1.0, aT.shortTermWindow, aT.shortTermStep, "svm",
model, True)
Regression
'''
'''
Train Regression Models for Audio Segments
featureAndTrainRegression(). It calls functions dirsWavFeatureExtraction() to extract the audio features and then
repetivelly calls evaluateRegression() to extract one optimized regression model for each learned parameter.
'''
import subprocess
from pyAudioAnalysis import audioTrainTest as aT
aT.featureAndTrainRegression("pyAudioAnalysis/data/speechEmotion/", 1, 1,
aT.shortTermWindow, aT.shortTermStep, "svm",
"output/svmSpeechEmotion", False)
'''
Regression task valence
Param MSE T-MSE R-MSE
0.0010 0.37 0.45 0.42
0.0050 0.30 0.32 0.43
0.0100 0.29 0.28 0.43 best
0.0500 0.30 0.20 0.44
0.1000 0.33 0.18 0.44
0.2500 0.46 0.16 0.42
0.5000 0.62 0.15 0.44
1.0000 0.86 0.14 0.41
5.0000 1.70 0.13 0.44
10.0000 2.41 0.12 0.42
Selected params: 0.01000
from pyAudioAnalysis import audioTrainTest as att
att.featureAndTrainRegression('wav', 1.0, 1.0, att.shortTermWindow,
att.shortTermStep, 'svm', 'mesto', True)
def trainRegressionWrapper(method, beat_feats, dirName, model_name):
aT.featureAndTrainRegression(dirName, 1, 1, aT.shortTermWindow,
aT.shortTermStep, method.lower(), model_name,
compute_beat=beat_feats)
root_data_path + 'radioFinal/train/bbc4A.segments',
'hmmTemp1', 1.0, 1.0)
aS.trainHMM_fromDir(root_data_path + 'radioFinal/small', 'hmmTemp2', 1.0, 1.0)
aS.hmmSegmentation(root_data_path + 'pyAudioAnalysis/data//scottish.wav',
'hmmTemp1', True, root_data_path +
'pyAudioAnalysis/data//scottish.segments') # test 1
aS.hmmSegmentation(root_data_path + 'pyAudioAnalysis/data//scottish.wav',
'hmmTemp2', True, root_data_path +
'pyAudioAnalysis/data//scottish.segments') # test 2
print("\n\n\n * * * TEST 7 * * * \n\n\n")
aT.featureAndTrainRegression(root_data_path +
"pyAudioAnalysis/data/speechEmotion",
1,
1,
0.050,
0.050,
"svm_rbf",
"temp.mod",
compute_beat=False)
print(
aT.fileRegression(
root_data_path + "pyAudioAnalysis/data/speechEmotion/01.wav",
"temp.mod", "svm_rbf"))
print("\n\n\n * * * TEST 8 * * * \n\n\n")
aT.featureAndTrainRegression(root_data_path +
"pyAudioAnalysis/data/speechEmotion",
1,
1,
0.050,
print("\n\n\n * * * TEST 3 * * * \n\n\n")
[Fs, x] = audioBasicIO.readAudioFile(root_data_path + "pyAudioAnalysis/data/doremi.wav")
x = audioBasicIO.stereo2mono(x)
specgram, TimeAxis, FreqAxis = audioFeatureExtraction.stChromagram(x, Fs, round(Fs * 0.040), round(Fs * 0.040), True)
print("\n\n\n * * * TEST 4 * * * \n\n\n")
aT.featureAndTrain([root_data_path +"SM/speech",root_data_path + "SM/music"], 1.0, 1.0, 0.2, 0.2, "svm", "temp", True)
print("\n\n\n * * * TEST 5 * * * \n\n\n")
[flagsInd, classesAll, acc, CM] = aS.mtFileClassification(root_data_path + "pyAudioAnalysis/data//scottish.wav", root_data_path + "pyAudioAnalysis/data/svmSM", "svm", True, root_data_path + 'pyAudioAnalysis/data/scottish.segments')
print("\n\n\n * * * TEST 6 * * * \n\n\n")
aS.trainHMM_fromFile(root_data_path + 'radioFinal/train/bbc4A.wav', root_data_path + 'radioFinal/train/bbc4A.segments', 'hmmTemp1', 1.0, 1.0)
aS.trainHMM_fromDir(root_data_path + 'radioFinal/small', 'hmmTemp2', 1.0, 1.0)
aS.hmmSegmentation(root_data_path + 'pyAudioAnalysis/data//scottish.wav', 'hmmTemp1', True, root_data_path + 'pyAudioAnalysis/data//scottish.segments') # test 1
aS.hmmSegmentation(root_data_path + 'pyAudioAnalysis/data//scottish.wav', 'hmmTemp2', True, root_data_path + 'pyAudioAnalysis/data//scottish.segments') # test 2
print("\n\n\n * * * TEST 7 * * * \n\n\n")
aT.featureAndTrainRegression(root_data_path + "pyAudioAnalysis/data/speechEmotion", 1, 1, 0.050, 0.050, "svm_rbf", "temp.mod", compute_beat=False)
print(aT.fileRegression(root_data_path + "pyAudioAnalysis/data/speechEmotion/01.wav", "temp.mod", "svm_rbf"))
print("\n\n\n * * * TEST 8 * * * \n\n\n")
aT.featureAndTrainRegression(root_data_path + "pyAudioAnalysis/data/speechEmotion", 1, 1, 0.050, 0.050, "svm", "temp.mod", compute_beat=False)
print(aT.fileRegression(root_data_path + "pyAudioAnalysis/data/speechEmotion/01.wav", "temp.mod", "svm"))
print("\n\n\n * * * TEST 9 * * * \n\n\n")
aT.featureAndTrainRegression(root_data_path + "pyAudioAnalysis/data/speechEmotion", 1, 1, 0.050, 0.050, "randomforest", "temp.mod", compute_beat=False)
print(aT.fileRegression(root_data_path + "pyAudioAnalysis/data/speechEmotion/01.wav", "temp.mod", "randomforest"))
