def __init__(self, data):
# normalize
self._normalization = normalization.Normalization(data)
normalized_data = self._normalization.normalized_dataset()
# normalized_data = data
# find covariance matrix
data_matrix = sp.matrix(normalized_data)
m = data_matrix.shape[0]
covariance_matrix = data_matrix.transpose() * data_matrix
covariance_matrix /= m
# find principal components
eig_decomp = linalg.eigh(
covariance_matrix
) # sorted by eig. values (ascending - we want descending)
self._n = len(eig_decomp[0])
self._pcas = sp.zeros(
(self._n, self._n)
) # one row will be one princ. comp. (starting from most important PC)
for i in range(self._n):
self._pcas[i, :] = eig_decomp[1][:, self._n - i - 1]
self._eig_vals = list(eig_decomp[0])
self._eig_vals.reverse()
def __init__(self, owner):
self.owner = owner
self.data_concatenation = concatenation.Concatenation(self)
self.data_imputation = imputation.Imputation(self)
self.data_normalization = normalization.Normalization(self)
self.configure_parameter_set()
def startNormalize(self):
try:
self.normalizer = normalization.Normalization(
inputAudioFileName, targetLUFS)
self.normalizer.normalize(inputAudioFileName, targetLUFS)
self.printBarChar(self.histgrammPlotBefore,
self.normalizer.histogramm)
self.printBarChar(self.histgrammPlotAfter,
self.normalizer.histogramm2)
self.printChar(self.graphicsView_3,
self.normalizer.plotSamplesArray1,
self.normalizer.yArray1)
self.printChar(self.waveFormPlotAfter,
self.normalizer.plotSamplesArray2,
self.normalizer.yArray2)
except Exception as e:
widgetMB = QtWidgets.QWidget()
QtWidgets.QMessageBox.about(
widgetMB, "Ошибка",
"Не удалось нормализовать файл! Ошибка: " + str(e))
# test[i] = indices[-nb_val:] # # whole_indices = [] # train_indices = [] # test_indices = [] # for i in range(m): # # whole_indices += labels_loc[i] # train_indices += train[i] # test_indices += test[i] # np.random.shuffle(train_indices) # np.random.shuffle(test_indices) # return train_indices, test_indices data = data_UP.reshape(np.prod(data_UP.shape[:2]), np.prod(data_UP.shape[2:])) gt = gt_UP.reshape(np.prod(gt_UP.shape[:2]), ) data = normalization.Normalization(data) data_ = data.reshape(data_UP.shape[0], data_UP.shape[1], data_UP.shape[2]) # data_trans = data.transpose() # whole_pca = doPCA.dimension_PCA(data_trans, data_UP, INPUT_DIMENSION) whole_pca = data_ #raw data print(whole_pca.shape) padded_data = zeroPadding.zeroPadding_3D(whole_pca, PATCH_LENGTH) ITER = 10 #ITER = 10 CATEGORY = 9 OA = [] AA = [] TRAINING_TIME = []
import normalization as normtool
import permutation as permtool
import sqlconnection as conntool
sqlcon = conntool.SQLConnection("db/sentiment.db")
your_api_key = ""
norm = normtool.Normalization()
perm = permtool.Permutation(api_key=your_api_key,
max_meanings=3,
use_local_meanings=True)
counter_save_local_meanings = 0
while True:
progress = sqlcon.getNumberOfSamplesToExpand()
print(progress)
samplesToExpand = sqlcon.getSamplesToExpand()
if len(samplesToExpand) == 0:
break
for sample in samplesToExpand:
sampleId, message, status = sample
text = norm.transform(message)
disambiguation_set = perm.getDisambiguation(text)
meanings = perm.getMeanings(disambiguation_set)
new_samples = perm.transform(text,
t0 = t000 = float(time.clock())
data = cPickle.load(open('cifar_2class_py2.p', 'rb'))
t1 = float(time.clock())
print('Loading time is %.4f s. \n' % (t1 - t0))
newdir = './result/result_%s' % (t1)
if not os.path.exists(newdir):
os.makedirs(newdir)
newdir += '/'
f = open(newdir + '/runtime.txt', 'w')
#plt.figure()
#plt.plot([1,2,3,4,3,5],'-r')
#plt.show()
x, test_x = nor.Normalization(data['train_data'], data['test_data'])
train_y = data['train_labels']
test_y = data['test_labels']
num_epochs = 30
num_batches = 1000
#batch_list = [100] # tune
batch_list = [100, 500, 1000, 2000] # tune
hidden_units = 100 # fix learning when tuning other
#hiddenU_list = [100,1000] # tune
hiddenU_list = [10, 100, 1000] # tune
lr = 0.001 # fix learning when tuning other
#lr_list = [0.001] # tune learning_rate
lr_list = [0.005, 0.001, 0.0005, 0.0001] # tune learning_rate
mu = 0.8 # fix momentum to 0.8
ojo_dcho = detector_cara.calculate_centroide(puntos, detector_cara.RIGHT_EYE_POINTS)
ojo_izdo = detector_cara.calculate_centroide(puntos, detector_cara.LEFT_EYE_POINTS)
p1x = faces[0][0]
p1y = faces[0][1] - 20
p2x = faces[0][2]
p2y = faces[0][3]
w = p2x - p1x
h = p2y - p1y
oix = ojo_izdo[0] - p1x
oiy = ojo_izdo[1] - p1y
odx = ojo_dcho[0] - p1x
ody = ojo_dcho[1] - p1y
roi_img = img[p1y:p1y + h, p1x:p1x + w]
roi_gray = cv2.cvtColor(roi_img, cv2.COLOR_BGR2GRAY)
normalizator = normalization.Normalization()
normalizator.normalize_gray_img(roi_gray, odx, ody, oix, oiy,
normalization.Kind_wraping.HS)
img_final = cv2.resize(normalizator.norm_image, (227, 227))
cv2.imwrite('normaHS.jpeg', img_final)
img2 = cv2.imread('normaHS.jpeg')
female, male = GenderHSNet.image_gender_classifier("normaHS.jpeg")
print(os.path.join(base, file))
if male * 100 > 50:
# meter en man
os.rename(os.path.join(base, file), os.path.join(manDir, file))
print(os.path.join(manDir, file))
print('probabilidad de ser hombre', male * 100)
else:
# meter en woman