def get_gmm(self):
from sklearn.mixture import GMM as skGMM
from gmmset import GMM as pyGMM
if GMM == skGMM:
print 'using GMM from sklearn'
return GMM(self.nr_mixture)
else:
print 'using pyGMM'
return GMM(nr_mixture=self.nr_mixture,
nr_iteration=500,
init_with_kmeans=0,
concurrency=8,
threshold=1e-15,
verbosity=2)
def get_gmm():
# from sklearn.mixture import GMM as skGMM
from sklearn.mixture import GaussianMixture as skGMM
from gmmset import GMM as pyGMM
if GMM == skGMM:
print('using GMM from sklearn')
return GMM(nr_mixture)
else:
print('using pyGMM')
return GMM(nr_mixture=nr_mixture,
nr_iteration=500,
init_with_kmeans=0,
concurrency=8,
threshold=1e-15,
verbosity=2)
def test_ubm_var_channel():
ubm = GMM.load('model/ubm.mixture-32.person-20.immature.model')
train_duration = 8.
nr_test = 5
test_duration = 3.
audio_files = ['xinyu.vad.wav', 'wyx.wav']
X_train, y_train, X_test, y_test = [], [], [], []
for audio_file in audio_files:
fs, signal = wavfile.read(audio_file)
signal = monotize_signal(signal)
train_len = int(fs * train_duration)
test_len = int(fs * test_duration)
X_train.append(mix_feature((fs, signal[:train_len])))
y_train.append(audio_file)
for i in range(nr_test):
start = random.randint(train_len, len(signal) - test_len)
X_test.append(mix_feature((fs, signal[start:start + train_len])))
y_test.append(audio_file)
gmmset = GMMSet(32, ubm=ubm)
gmmset.fit(X_train, y_train)
y_pred = gmmset.predict_with_reject(X_test)
for i in xrange(len(y_pred)):
print y_test[i], y_pred[i], '' if y_test[i] == y_pred[i] else 'wrong'
for imposter_audio_file in map(lambda x: 'test-{}.wav'.format(x),
range(5)):
fs, signal = wavfile.read(imposter_audio_file)
signal = monotize_signal(signal)
imposter_x = mix_feature((fs, signal))
print gmmset.predict_one_with_rejection(imposter_x)
def test_ubm_var_channel():
ubm = GMM.load('model/ubm.mixture-32.person-20.immature.model')
train_duration = 8.
nr_test = 5
test_duration = 3.
audio_files = ['xinyu.vad.wav', 'wyx.wav']
X_train, y_train, X_test, y_test = [], [], [], []
for audio_file in audio_files:
fs, signal = wavfile.read(audio_file)
signal = monotize_signal(signal)
train_len = int(fs * train_duration)
test_len = int(fs * test_duration)
X_train.append(mix_feature((fs, signal[:train_len])))
y_train.append(audio_file)
for i in range(nr_test):
start = random.randint(train_len, len(signal) - test_len)
X_test.append(mix_feature((fs, signal[start:start+train_len])))
y_test.append(audio_file)
gmmset = GMMSet(32, ubm=ubm)
gmmset.fit(X_train, y_train)
y_pred = gmmset.predict_with_reject(X_test)
for i in xrange(len(y_pred)):
print y_test[i], y_pred[i], '' if y_test[i] == y_pred[i] else 'wrong'
for imposter_audio_file in map(
lambda x: 'test-{}.wav'.format(x), range(5)):
fs, signal = wavfile.read(imposter_audio_file)
signal = monotize_signal(signal)
imposter_x = mix_feature((fs, signal))
print gmmset.predict_one_with_rejection(imposter_x)
def load(fname):
""" load from a dumped model file"""
with open(fname, 'r') as f:
R = pickle.load(f)
R.gmmset.after_pickle()
ubm = GMM.load(ModelInterface.UBM_MODEL_FILE)
R.gmmset.ubm = ubm
return R
def main():
nr_person = 50
fpaths = get_training_data_fpaths()
X_train, y_train, X_test, y_test = datautil.read_data(fpaths, nr_person)
ubm = GMM.load('model/ubm-32.model')
for x, y in zip(X_train, y_train):
gmm = GMM(concurrency=8, threshold=0.01, nr_iteration=100, verbosity=1)
gmm.fit(x, ubm=ubm)
gmm.dump("model/" + y + ".32.model")
def _get_gmm_set(self):
if os.path.isfile(self.ubm_model_file):
try:
from gmmset import GMMSetPyGMM
if GMMSet is GMMSetPyGMM:
return GMMSet(ubm=GMM.load(self.ubm_model_file))
except Exception as e:
print "Warning: failed to import gmmset. You may forget to compile gmm:"
print e
print "Try running `make -C src/gmm` to compile gmm module."
print "But gmm from sklearn will work as well! Using it now!"
return GMMSet()
return GMMSet()
def _get_gmm_set(self):
if self.UBM_MODEL_FILE and os.path.isfile(self.UBM_MODEL_FILE):
try:
from gmmset import GMMSetPyGMM
if GMMSet is GMMSetPyGMM:
return GMMSet(ubm=GMM.load(self.UBM_MODEL_FILE))
except Exception as e:
print("Warning: failed to import gmmset. You may forget to compile gmm:")
print(e)
print("Try running `make -C src/gmm` to compile gmm module.")
print("But gmm from sklearn will work as well! Using it now!")
return GMMSet()
return GMMSet()
def _get_gmm_set(self):
if os.path.isfile(self.UBM_MODEL_FILE):
try:
from gmmset import GMMSetPyGMM
if GMMSet is GMMSetPyGMM:
return GMMSet(ubm=GMM.load(self.UBM_MODEL_FILE))
except Exception as e:
print "Warning: failed to import gmmset. You may forget to compile gmm:"
print e
print "Try running `make -C src/gmm` to compile gmm module."
print "But gmm from sklearn will work as well! Using it now!"
return GMMSet()
return GMMSet()
def main():
nr_person = 50
fpaths = get_training_data_fpaths()
X_train, y_train, X_test, y_test = datautil.read_data(fpaths, nr_person)
ubm = GMM.load("model/ubm-32.model")
for x, y in zip(X_train, y_train):
gmm = GMM(concurrency=8, threshold=0.01, nr_iteration=100, verbosity=1)
gmm.fit(x, ubm=ubm)
gmm.dump("model/" + y + ".32.model")
def predict(self):
""" return a label (name)"""
try:
'''
mfcc = MFCC.extract(self.FS, self.signal)
lpc = LPC.extract(self.FS, self.signal)
features = np.concatenate((mfcc, lpc), axis=1)
'''
features = self.mix_feature()
except Exception as e:
print tb.format_exc()
return None
self.gmmset.ubm = GMM.load(self.ubm_model_file)
#print 'The registered users are :', len(self.gmmset.y)
#print self.gmmset.y
#print [y for y in self.gmmset.y]
return self.gmmset.predict_one_with_rejection(features)
for i in xrange(len(y_pred)):
print y_test[i], y_pred[i], '' if y_test[i] == y_pred[i] else 'wrong'
for imposter_audio_file in map(lambda x: 'test-{}.wav'.format(x),
range(5)):
fs, signal = wavfile.read(imposter_audio_file)
signal = monotize_signal(signal)
imposter_x = mix_feature((fs, signal))
print gmmset.predict_one_with_rejection(imposter_x)
test_ubm_var_channel()
import sys
sys.exit(0)
ubm = GMM.load('model/ubm.mixture-32.person-20.immature.model')
gmm = GMM(32, verbosity=1)
#audio_file = 'test-data/corpus.silence-removed/Style_Reading/f_001_03.wav'
fs, signal = wavfile.read(audio_file)
signal = monotize_signal(signal)
X = mix_feature((fs, signal))
ubm = GMM.load('model/ubm.mixture-32.person-20.immature.model')
gmm = GMM(32, verbosity=1)
X = X[:1000]
gmm.fit(X, ubm=ubm)
gmm.dump('xinyu.model')
y_pred = gmmset.predict_with_reject(X_test)
for i in xrange(len(y_pred)):
print y_test[i], y_pred[i], '' if y_test[i] == y_pred[i] else 'wrong'
for imposter_audio_file in map(
lambda x: 'test-{}.wav'.format(x), range(5)):
fs, signal = wavfile.read(imposter_audio_file)
signal = monotize_signal(signal)
imposter_x = mix_feature((fs, signal))
print gmmset.predict_one_with_rejection(imposter_x)
test_ubm_var_channel()
import sys
sys.exit(0)
ubm = GMM.load('model/ubm.mixture-32.person-20.immature.model')
gmm = GMM(32, verbosity=1)
#audio_file = 'test-data/corpus.silence-removed/Style_Reading/f_001_03.wav'
fs, signal = wavfile.read(audio_file)
signal = monotize_signal(signal)
X = mix_feature((fs, signal))
ubm = GMM.load('model/ubm.mixture-32.person-20.immature.model')
gmm = GMM(32, verbosity=1)
X = X[:1000]
gmm.fit(X, ubm=ubm)
gmm.dump('xinyu.model')